diff --git "a/pubmed-central/test.jsonl" "b/pubmed-central/test.jsonl"
new file mode 100644--- /dev/null
+++ "b/pubmed-central/test.jsonl"
@@ -0,0 +1,584 @@
+{"text": "Introduction\n============\n\nPatients with cholangiocarcinoma (CAA) have extremely poor prognosis with an average 5-year survival rate of 5%--10%.[@b1-cmo-2009-121] Among the clinicopathologic factors influencing the survival after surgical treatment, curative surgery, lymph node metastases, tumor size and cancer-free margin are the most predictive factors.[@b2-cmo-2009-121]--[@b8-cmo-2009-121] Recently, in a large series of CCA, multivariate analysis showed that EGFR expression was a risk factor for recurrence of intrahepatic CCA[@b9-cmo-2009-121] with a five-year survival rate at around 20% and in hilar bile duct cancer, after curative resection, 40% of the patients had disease recurrence.[@b8-cmo-2009-121] In these patients with high risk factors, adjuvant therapy may play a role in prolonging survival. The aim of this pilot study was to assess the efficacy and safety of adjuvant chemotherapy (GEMOX) in high-risk patients with intrahepatic and hilar CAA after curative surgery.\n\nPatients and Methods\n====================\n\nPatients and treatment\n----------------------\n\nPatients were eligible for entry into this pilot study if they fulfilled the following criteria: age \u226518 years, Karnofsky Performance Status (KPS) \u2265 80%, histologically confirmed diagnosis of CCA tumor tissue available for immunohistochemical EGFR detection and curative surgery. Laboratory acceptance parameters included an absolute neutrophil count of \u22651,500 cells/\u03bcL, a platelet count of \u2265100,000 cells/\u03bcL, a serum creatinine level \\< 130 \u03bcmol/L and a total serum bilirubin level \\< 3 \u00d7 the upper normal limit. The protocol was approved by the institutional review board (Centre Hepato-Biliaire, Villejuif, France) and was conducted according to the principles of the Declaration of Helsinski and the rules of good clinical practices. Informed consent was obtained from all patients. Twenty two Caucasian patients (10 women, 12 men, mean age: 57 years old, range: 30--73, KPS \\> 80%) were treated in our institution for an intrahepatic (n = 10) or hilar CCA (n = 12). They had one or more dismal prognostic factors: lymph node involvement, positive histologic margins, pernervous and/or vascular tumoral embols. All patients received 6 cycles of GEMOX for 5 weeks after they had undergone curative surgery. All patients had histologically proven EGFR expressing CCA. Gemcitabine was given at an initial dose of 1000 mg/m^2^ as a 10 mg/m^2^/min infusion on day 1 and oxaliplatin 85 mg/m^2^ as a 4-h infusion on day 2. A prophylactic antiemetic treatment comprising 5-hydroxytryptamine type 3 receptor antagonists and dexamethasone was given. GEMOX was repeated every 3 weeks if the neutrophil count was \\>1500 cells/\u03bcL and the platelet count was \\>100 000/\u03bcL.\n\nImmunohistochemistry\n--------------------\n\nImmunohistochemical detection of EGFR was performed using EGFR antibodies (31G7 clone and/or 2-18C9 clone) on 4 \u03bcm thick deparaffinized tumor sections before any chemotherapy. Only cell membrane staining was considered to be specific. EGFR status was considered positive when \\> = 1% of tumor cells showed complete membranous staining. The percentage of tumor cells expressing EGF-R was semiquantitatively assessed, and the intensity of staining was scored as follows: 0: no staining; 1+: weak, 2+: moderate; 3+: strong. When staining intensity was heterogeneous, the highest intensity was retained as the score. Hepatocytes and peripheral nerves served as positive internal controls and positive (HT29 cell line) and negative (CAM-1 cell line) external controls were included ([Fig. 3](#f3-cmo-2009-121){ref-type=\"fig\"}).\n\nAssessment of efficacy and toxicity\n-----------------------------------\n\nThe primary endpoint of the study was disease free survival. The secondary endpoint was overall survival. CT scans were performed at base line and every three cycles. After the treatment, CT scans were performed every three months until progression. Toxicity was graded according to the National Cancer Institute-Common Toxicity Criteria (NCI CTC) (version 2.0) and evaluated at each cycle.\n\nStatistical analysis\n--------------------\n\nThe Kaplan-Meier method was used to estimate overall and progression-free survival outcomes. Time to progression was calculated from the first day of GEMOX administration until the date when first progression of disease was observed. Overall survival was calculated from the first day of GEMOX administration until the date of death or last follow-up.\n\nResults\n=======\n\nEfficacy\n--------\n\nBetween 2000 and 2005, 22 patients with histologically confirmed CCA underwent curative surgery and lymphadenectomy. Patient characteristics are summarized in [Table 1](#t1-cmo-2009-121){ref-type=\"table\"}. The different surgical procedures included 7 right hepatectomies extended to segments 4 and 1; 8 left hepatectomies extended to segment 1; 5 left hepatectomies extended to segments 4 and 1; 1 right hepatectomy extended to segments 4; 1 left hepatectomy extended to segment 1, 4. Hepatic resection was considered as curative (R0) when there was no evidence of microscopic disease (8 patients). Resection was R1 (only microscopic disease) for 14 patients.\n\nThe men/women ratio was 12/10. The median age of the patients was 57 years (range 30--73). The median follow-up period was 30 months (range 8--66). Immunohistochemical results were available for 21 of the 22 patients according to the results of positive controls ([Table 2](#t2-cmo-2009-121){ref-type=\"table\"}). For the latest tumor, positive internal controls were always negative in spite of several tests. EGFR membranous expression was present in 20 CCA (95%) and one tumor was negative. If a 10% cutoff was chosen, 15 tumors out of 21 (76%) were EGFR-positive. EGFR was mainly overexpressed in the moderately and/or poorly differentiated tumors, whereas only two well-differentiated cases showed EGFR overexpression.\n\nAll 22 patients completed 6 cycles of adjuvant GEMOX. The 5-year survival rate was 56% (CI 95%: 25.7--85.4) ([Fig. 1](#f1-cmo-2009-121){ref-type=\"fig\"}) and the 2-year disease free survival rate was 28% (CI 95%: 3.4--52.6) ([Fig. 2](#f2-cmo-2009-121){ref-type=\"fig\"}). The median time to progression was 15 months: 10 months for hilar CCA (n = 12), and 15 months for intrahepatic CCA (n = 10). Eight patients (hilar = 4, intrahepatic n = 4) had no progression of the disease. These 8 patients had no lymph node metastases but 5 had a positive histologic margin (R1). The rate of recurrence after surgery and chemotherapy was 63% (14/22). The recurrences occurred in the liver, peritoneum and metastatic lymph nodes (hilar and celiac lymph nodes). Two patients died 13 and 29 months after the beginning of the treatment due to disease progression. After relapse, a second line of chemotherapy adding cetuximab to GEMOX was performed for 12 patients. Six patients died after 12 months on average (range: 9--48 months) after cetuximab onset. A third line of chemotherapy was administered to 6 patients, 3 of whom are still alive. ([Table 3](#t3-cmo-2009-121){ref-type=\"table\"}).\n\nSafety\n------\n\nAll 22 patients completed 6 cycles of GEMOX. Treatment was well tolerated. No reductions in gemcitabine and oxaliplatin doses or treatment delays were necessary. All patients were assessable for toxicity. There was no treatment-related death, anemia, neutropenia, thrombocytopenia, or neurotoxicity. Acne-like rash was observed in all patients treated with cetuximab and occurred within the first three weeks of the treatment.\n\nDiscussion\n==========\n\nOur study analyses the impact of adjuvant chemotherapy in hilar and intrahepatic CCA after curative surgery. All patients had 1 or more factors of poor prognosis such as positive margin, lymph node metastases, lymphatic vessel invasion or perineural invasion and 95% of tumors had membranous expression of EGFR. In our study, we used GEMOX after curative surgery since regimens including gemcitabine were found to be active and safe for the patients with advanced biliary tract carcinoma.[@b10-cmo-2009-121]--[@b15-cmo-2009-121] Even if well tolerated, the adjuvant GEMOX regimen alone does not seem to be very effective in our high-risk patients reaching a 2-year progression free survival rate of 28%. In comparison, a 3-year disease-free survival rate of 43% has been reported in resected patients with hilar or intrahepatic CCA[@b4-cmo-2009-121],[@b8-cmo-2009-121] and the five-year survival rate for intrahepatic CCA patients with EGFR-positive tumors was low at around 20%.[@b9-cmo-2009-121] A possible limited benefit of GEMOX may be due to EGFR expression that has been associated with increasing resistance to chemotherapy.[@b16-cmo-2009-121],[@b17-cmo-2009-121] However, there were 8 patients (4 intrahepatic and 4 hilar CCA) with no tumor progression, probably due to the fact that these patients had no lymph node metastasis which is a significant risk factor of tumor recurrence[@b7-cmo-2009-121] Furthermore, in hilar CCA without lymph node involvement, major hepatectomy can offer long term survival even in the case of R1 resection.[@b8-cmo-2009-121]\n\nThe median survival is not reached in our study because 12 patients received additional treatment with cetuximab/GEMOX at the time of relapse. Since adding cetuximab circumvents tumor resistance to chemotherapy, the tumor may respond to a therapy on which it had previously progressed. This mechanism has previously been documented in metastatic colorectal cancer or in recurrent head and neck cancer.[@b16-cmo-2009-121],[@b18-cmo-2009-121] In human CCA cell lines, the EGFR kinase inhibitors AG1478 or ZD1839 significantly suppress CCA cell growth.[@b19-cmo-2009-121] Recently, cetuximab or inhibitors of tyrosine kinase have been used clinically in intrahepatic CCA.[@b20-cmo-2009-121],[@b21-cmo-2009-121] Some case reports have been reported[@b22-cmo-2009-121],[@b23-cmo-2009-121] and in patients with refractory advanced intra hepatic CCA, adding cetuximab to GEMOX circumvented tumor resistance to chemotherapy in some patients.[@b21-cmo-2009-121] These studies suggest a clinical applicability of EGFR inhibitors in CCA. For improving the use of adjuvant GEMOX in EGFR positive CCA, drugs that target tumor cell---associated receptor tyrosine kinase might be useful in patients classified as high-risk.\n\nConclusion\n==========\n\nGiven the results of the present study, we consider that six cycles of adjuvant GEMOX is not the optimal chemotherapy in patients with high risk factors. The limited benefit of GEMOX is possibly due to EGFR expression, because EGFR overexpression has been associated with more aggressive disease and increased resistance to chemotherapy. Our data may indicate the need for additional studies regarding the role of the target therapy in adjuvant treatment among certain subsets of CCA including those with EGFR-positive tumors.\n\nThis article is available from <http://www.la-press.com>.\n\n**Disclosures**\n\nThe authors report no conflicts of interest.\n\n![Overall survival in high risk patients.](cmo-2009-121f1){#f1-cmo-2009-121}\n\n![Progression free survival in high risk patients.](cmo-2009-121f2){#f2-cmo-2009-121}\n\n![Anti-EGFR immunostaining (31G7 monoclonal antibody---LSAB technique) with +++ positivity \u00d7 400.](cmo-2009-121f3){#f3-cmo-2009-121}\n\n###### \n\nPatient characteristics.\n\n  **Characteristics**                     \n  --------------------------------------- -------------\n  Number of patients                      22\n  **Age (years)**                         \n  \u2003\u2003Median (range)                        57 (30--73)\n  **Gender**                              \n  \u2003\u2003Men/Women                             12/10\n  **Performance status (Karnofsky %)**    90\n  **Cholangiocarcinoma**                  \n  \u2003\u2003Extra hepatic (Klatskin)              12\n  \u2003\u2003Intrahepatic                          10\n  **TNM stage**                           \n  \u2003\u2003pT2                                   9\n  \u2003\u2003pT3                                   9\n  \u2003\u2003pT4                                   4\n  **Surgical treatment**                  \n  \u2003\u2003Partial hepatectomy right/left        10/12\n  **Indications of adjuvant treatment**   \n  \u2003\u2003Positive margins                      14\n  \u2003\u2003Negative margins                      8\n  \u2003\u2003N+                                    7\n  \u2003\u2003Tumoral embol: perinervous/vascular   10/10\n  Complete response                       8\n  Progression disease                     14\n\nOne patient could have more than one indication for adjuvant treatment.\n\n###### \n\nEGFR protein expression.\n\n  **Patient N\u00b0**   **EGFR protein expression(% of positive tumor cells/intensity)**   **Grade of differentiation**\n  ---------------- ------------------------------------------------------------------ ------------------------------\n  1                50%/2+                                                             Moderate\n  2                20%/2+                                                             Moderate\n  3                100%/2+                                                            Poor\n  4                30%/1+                                                             Poor\n  5                40%/2+                                                             Poor\n  6                0%                                                                 Moderate\n  7                100%/2+                                                            Well\n  8                100%/2+                                                            Poor\n  9                5%/1+                                                              Poor\n  10               70%/2+                                                             Well\n  11               1%/2+                                                              Moderate\n  12               5%/1+                                                              Moderate\n  13               1%/1+                                                              Well\n  14               1%/2+                                                              Well\n  15               30%/1+                                                             Moderate\n  16               20%/2+                                                             Moderate\n  17               70%/2+                                                             Moderate\n  18               30%/3+                                                             Poor\n  19               70%/2+                                                             Poor\n  20               NE                                                                 Moderate\n  21               30%/2+                                                             Poor\n  22               30%/2+                                                             Well\n\n**Abbreviation:** NE, not evaluable.\n\n###### \n\nOutcome of the patients.\n\n  --------------------------------------------------- ----\n  Disease related death                               8\n  Number of recurrent case                            14\n  Liver                                               7\n  Peritoneum                                          1\n  Lymph node                                          6\n  Alive after second and third line of chemotherapy   3\n  --------------------------------------------------- ----\n"}
+{"text": "1. Introduction\n===============\n\n*Helicobacter pylori* treatment still remains a challenge.^\\[[@R1]--[@R4]\\]^ Vaccination is the best option to *H pylori* but now we do not have it! Thus antibiotic therapy is preferable than other options.^\\[[@R5]--[@R10]\\]^ Despite initial successes, there has been an unacceptable level in *H pylori* triple eradication therapies currently due to increased antibiotic resistance, especially that to clarithromycin, metronidazole, and levofloxacin.^\\[[@R1]--[@R10]\\]^ Therefore, it is crucial to use *H pylori* eradication regimens with high efficacy and less adverse events.\n\nBismuth quadruple therapy (BQT), consisting of a proton pump inhibitor (PPI), bismuth, and 2 antibiotics (amoxicillin and clarithromycin or metronidazole et al), has been recommended in most current *H pylori* treatment guidelines as a first-line regimen.^\\[[@R2],[@R7]--[@R10]\\]^ For example, both BQT and non-bismuth quadruple therapies were recommended as first-line strategies for *H pylori* infection by Maastricht V/Florence Consensus Report guidelines and the Toronto Consensus for the Treatment of *H pylori* Infection in Adults.^\\[[@R2],[@R8]\\]^ BQT was recommended as 1 common solution in patients with a penicillin allergy by *H pylori* Management in ASEAN: the Bangkok Consensus Report.^\\[[@R7]\\]^ The Fifth Chinese National Consensus Report on the management of *H pylori* infection has also recommended BQT as the main empirical therapy for *H pylori* eradication.^\\[[@R10]\\]^ Furthermore, Maastricht-V Consensus Report recommended BQT with no need for drug-sensitive test.^\\[[@R8]\\]^ However, in some regions, BQT is not available.^\\[[@R2]\\]^ Moreover, BQT has relatively high side effects.\n\nUp to now, the global prevalence of primary and the acquired *H pylori* resistance to amoxicillin are still generally rare.^\\[[@R11]--[@R14]\\]^ The actual efficacy of amoxicillin/PPI dual therapy that has been used in several areas remains controversial partly because of differences in doses and dosing frequency.^\\[[@R13],[@R15]--[@R35]\\]^ Actually, high-dose dual therapy (HDDT), defined as amoxicillin \u22652.0\u200ag/day, amoxicillin or PPI gave 3 or 4 times daily, or administration of both amoxicillin and PPI 4 times daily for 14 days, has resulted in greater efficacy (i.e., over 90%).^\\[[@R32],[@R34]--[@R38]\\]^ Several clinical trials have reported the *H pylori* eradication rates of the HDDT compared with BQT.^\\[[@R27],[@R35],[@R39]--[@R41]\\]^\n\nIn this study, we performed this meta-analysis to compare the efficacy and safety of 2 *H pylori* eradication regimens, HDDT and BQT.\n\n2. Materials and methods\n========================\n\n2.1. Eligibility criteria\n-------------------------\n\nStudies included in the meta-analysis met the following criteria:\n\n1.  studies designed as randomized controlled trials or controlled clinical trials;\n\n2.  studies enrolling diagnosed *H pylori* infection patients, regardless of gender, age, or race;\n\n3.  studies comparing HDDT and BQT, not necessarily in a blind manner; and\n\n4.  studies with similar end-points of interests, including efficacy (intention-to-treat (ITT) eradication rate, per-protocol (PP) eradication rate and adherence) and drug-related toxicity (incidence of side effects), and reported relative risk (RR) with corresponding 95% confidence intervals (CIs).\n\n2.2. Exclusion criteria\n-----------------------\n\nThe following exclusion criteria were set:\n\n1.  studies not comparing HDDT and BQT;\n\n2.  Randomized clinical trials (RCTs) in which patients received either HDDT or BQT in combination with other drugs;\n\n3.  studies with inappropriate statistical methods or duplicated or overlapped data in multiple reports;\n\n4.  studies from which meaningful statistical data could not be extracted; and\n\n5.  studies that were animal studies, non-clinical studies, case reports, reviews, or letters.\n\n2.3. Search strategy\n--------------------\n\nPubMed, Embase, Cochrane library, CNKI, and Wanfang databases in Chinese were searched up to March 2018 to identify studies comparing HDDT with BQT for *H pylori*-infected patients. The search strategy included terms: *Helicobacter pylori* or *H pylori*, amoxicillin, dual therapy, bismuth, and quadruple therapy. No limitation was used during the literature search. The references of eligible studies were reviewed for additional studies. The reporting of this study follows the PRISMA guidelines.^\\[[@R42]\\]^\n\n2.4. Study selection and data extraction\n----------------------------------------\n\nTwo authors independently extracted the relevant data from each included study. Disagreement was resolved by consensus, and then the accuracy was checked by the third author. We extracted the following information from included studies:\n\n1.  name of the first author, year of publication, and trials types;\n\n2.  methods used to confirm *H pylori* infection and eradication;\n\n3.  number of subjects, therapeutic regimens; and\n\n4.  main outcomes including ITT eradication rate, PP eradication rate, adherence, and side effects.\n\n2.5. Risk of bias\n-----------------\n\nTwo investigators separately rated the quality of retrieved studies. The quality of RCTs was assessed by Jadad quality scale.^\\[[@R43]\\]^ Funnel plots were constructed to evaluate the risk of publication bias.\n\n2.6. Statistical analysis\n-------------------------\n\nThe endpoints of interest in the pooled analysis were eradication rate, compliance, and side effects. A sensitivity analysis was also performed to examine the impact on the overall results, depending on the heterogeneity across the included studies. Between-study heterogeneity was evaluated using I^2^ statistic.^\\[[@R44]\\]^ I^2^ value larger than 50% suggested high degree of heterogeneity, less than 50% means low or moderate degree of heterogeneity.^\\[[@R45]\\]^ Both ITT and PP analyses were used for clinical outcomes. Summary RRs were calculated by using random-effect models when there was high heterogeneity among studies. Otherwise, fixed-effect models were used. *P* values less than .05 were considered to be statistically significant. Statistical analyses were conducted using Review Manager Version 5.3 software (Revman; The Cochrane collaboration Oxford, United Kingdom). Findings of our meta-analysis were shown in forest plots. Publication bias was assessed using funnel plot.\n\n3. Results\n==========\n\n3.1. Study selection\n--------------------\n\nIn all, 159 studies were obtained from the original search algorithm, of which 128 were excluded because they were not RCTs, duplicated, or irrelevant to the current analysis. Thirty-one studies were evaluated, of which 26 were further excluded because amoxicillin or PPI was given less than 3 times per day or the given dose of amoxicillin was \\<2.0\u200ag/day in the dual therapy, and 4 studies did not use BQT as control. Furthermore, 1 study in which amoxicillin or PPI did not be given 4 times daily for 14 days in HDDT was excluded.\n\nFinally, 4 prospective RCTs (without publication bias, Figure S1) including 829 participants (455 were treated with HDDT and 374 with BQT), met the inclusion criteria and were included in the meta-analysis. The flowchart of study selection is shown in Figure [1](#F1){ref-type=\"fig\"}. The characteristics of the included studies are summarized in Table [1](#T1){ref-type=\"table\"}. Gao et al.\\'s study^\\[[@R39]\\]^ and Hu JL et al\\'s study^\\[[@R40]\\]^ had a high risk for bias and the other 2 studies^\\[[@R26],[@R35]\\]^ had an unclear risk for bias (Figure S2, S3).\n\n![Study selection.](medi-98-e14396-g001){#F1}\n\n###### \n\nStudy characteristics.\n\n![](medi-98-e14396-g002)\n\n3.2. Meta-analysis results\n--------------------------\n\n### 3.2.1. Overall eradication rate\n\nA fixed- effects model was used to pool the ITT eradication rate data since the heterogeneity across the 4 studies was low (\u03c7^2^\u200a=\u200a3.38, *P*\u200a=\u200a.34, I^2^\u200a=\u200a11%). The pooled eradication rate was 85.5% (95% CI 82.3%--88.7%) in HDDT groups compared to 87.2% (95% CI 83.8%--90.6%) in BQT groups. The pooling data did not achieve advantage in the HDDT or BQT groups (RR\u200a=\u200a1.01, 95%CI\u200a=\u200a0.96--1.06, *P*\u200a=\u200a.63) (Fig. [2](#F2){ref-type=\"fig\"}).\n\n![Forest plot for comparison of HDDT with BQT. Outcome: *Helicobacter pylori* eradication rate (intention-to-treat). BQT\u200a=\u200abismuth quadruple therapy, HDDT\u200a=\u200ahigh-dose dual therapy.](medi-98-e14396-g003){#F2}\n\nA fixed- effects model was also used to pool the PP eradication rate data, since the heterogeneity across the 4 studies was low (\u03c7^2^\u200a=\u200a4.85, *P*\u200a=\u200a.18, I^2^\u200a=\u200a38%). The pooled eradication rate was 88.4% (95% CI 85.4%--91.4%) in HDDT groups compared to 91.5% (95% CI 88.7%--94.4%) in BQT groups. Results showed that there were no significant differences between HDDT and BQT were observed (RR\u200a=\u200a1.00, 95% CI: 0.96--1.04, *P*\u200a=\u200a.99; Fig. [3](#F3){ref-type=\"fig\"}).\n\n![Forest plot for comparison of HDDT with BQT. Outcome: *Helicobacter pylori* eradication rate (per-protocol). BQT\u200a=\u200abismuth quadruple therapy, HDDT\u200a=\u200ahigh-dose dual therapy.](medi-98-e14396-g004){#F3}\n\n### 3.2.2. Compliance\n\nBoth therapies showed a high compliance rate, with 96.7% (95% CI 95.1%--98.3%) for HDDT and 94.9% (95% CI 92.7%--97.2%) for BQT. No significant difference was observed (RR\u200a=\u200a1.01, 95% CI 0.99--1.04, *P*\u200a=\u200a.32; Fig. [4](#F4){ref-type=\"fig\"}).\n\n![Forest plot for comparison of HDDT with BQT. Outcome: compliance. BQT\u200a=\u200abismuth quadruple therapy, HDDT\u200a=\u200ahigh-dose dual therapy.](medi-98-e14396-g005){#F4}\n\n### 3.2.3. Side effects\n\nThe overall side effect rate was 14.4% (95% CI 11.0%--17.8%) for HDDT and 40.4% (95% CI 35.0%--45.8%) for BQT. The pooling side-effects data did achieve advantage in the HDDT therapy (RR\u200a=\u200a0.42, 95% CI\u200a=\u200a0.32--0.54, *P*\u200a\\<.00001) without significant statistical heterogeneity (\u03c7^2^\u200a=\u200a1.60, *P*\u200a=\u200a.45, I^2^\u200a=\u200a0%). In other words, HDDT therapy compared to BQT therapy did reduce the rate of side-effects (Fig. [5](#F5){ref-type=\"fig\"}).\n\n![Forest plot for comparison of HDDT with BQT. Outcome: side effects. BQT\u200a=\u200abismuth quadruple therapy, HDDT\u200a=\u200ahigh-dose dual therapy.](medi-98-e14396-g006){#F5}\n\n3.3. Sensitivity analysis\n-------------------------\n\nIn the sensitivity analysis, by removing 1 study at a time, the statistical significance of the pooled RR was not changed.\n\n4. Discussion\n=============\n\n*H pylori* infection is one of the most common chronic bacterial infections in humans and causes chronic progressive gastric inflammation and a variety of diseases, including peptic ulcer disease and gastric cancer.^\\[[@R2],[@R5]--[@R10]\\]^ Antibiotic resistance is thought to be the key element to consider in *H pylori* treatment.^\\[[@R1]--[@R7],[@R12],[@R14]\\]^ Resistance occurs by mutations which are errors during the replication of *H pylori* DNA. This can occur during the treatment of *H pylori* infection, but also when taking the antibiotic for another infection.^\\[[@R46]\\]^ So far, resistance rates of *H pylori* to antibiotics, such as clarithromycin, metronidazole, and levofloxacin, have reached alarming levels worldwide.^\\[[@R1],[@R3],[@R4],[@R11],[@R12],[@R14]\\]^\n\nDue to the global prevalence of primary and the acquired *H pylori* resistance to amoxicillin are still generally rare,^\\[[@R11]--[@R14]\\]^ a combination of PPI and amoxicillin dual therapy have been evaluated to treat *H pylori* infection for many areas, while the efficacy of this dual therapy is under controversial.^\\[[@R13],[@R15]--[@R36]\\]^ Many studies in which *H pylori*-infected patients were treated with standard-dose amoxicillin (2\u200ag/d or less) and PPI once or twice daily did not achieve satisfying results of *H pylori* eradication rate compared with the standard triple therapy.^\\[[@R13],[@R17]--[@R21],[@R28],[@R31],[@R33],[@R47]--[@R50]\\]^\n\nThere are several explanations for the decrease in the efficacy of *H pylori* eradication. On 1 hand, amoxicillin is time-dependent semi-synthetic penicillin and could be much better absorbed to blood after oral administration. Plasma levels in excess of the minimum inhibitory concentrations were maintained for 6 to 8\u200ahours, so plasma concentrations of amoxicillin cannot be achieved by only a single oral dose of 1\u200ag amoxicillin twice daily.^\\[[@R51]\\]^ On the other hand, responding to amoxicillin is strongly affected by gastric pH value. *H pylori* are much more likely sensitive to amoxicillin when gastric pH value is high (pH \\>6). Gastric pH value is associated with the dose of PPI, dosing frequency and drug type.^\\[[@R38],[@R52]\\]^\n\nAccordingly, researches from different areas of the world have investigated the optimal designs of the dual (PPI and amoxicillin) therapy.\n\n1.  Increasing the dose and dosing frequency of amoxicillin alone. For example, Schwartz et al compared a triple therapy with PPI and amoxicillin dual therapy. They found that when lansoprazole was given 30\u200amg twice daily in combination with amoxicillin 1\u200ag 3 times daily, the cure rate of *H pylori* was only 53% with ITT analysis.^\\[[@R15]\\]^\n\n2.  Modifying the dose and dosing frequency of PPI, and increasing gastric pH alone. For example, Attumi et al treated *H pylori* infected patients with high dose extended-release lansoprazole 120\u200amg twice daily in combination with amoxicillin 1\u200ag twice daily for 14 days. They found that the success rates of both PP and ITT treatment were only 53.8%.^\\[[@R18]\\]^\n\n3.  Increasing PPI and amoxicillin simultaneously, and achieving satisfying effect.^\\[[@R32],[@R34],[@R35],[@R37],[@R39]\\]^ A previous multi-center RCT in 1995 has demonstrated that the ITT eradication rate was 91% treated with 40\u200amg omeprazole 3 times a day and 750\u200amg amoxicillin 3 times a day.^\\[[@R37]\\]^ Yang et al showed that the eradication rates of *H pylori* were 96.6% (PP) and 95.3% (ITT) using HDDT (rabeprazole 10\u200amg and amoxicillin 750\u200amg, 4\u200atimes/day for 14 days).^\\[[@R34]\\]^ While, Hu et al demonstrated that the eradication rate of *H pylori* was 96.4% (PP) and 94.7% (ITT), respectively, treated with HDDT (rabeprazole 20\u200amg and amoxicillin 750\u200amg, 4\u200atimes/day for 14 days).^\\[[@R35]\\]^\n\nHDDT means increasing PPI and amoxicillin simultaneously to achieve satisfied eradication rate of *H pylori* infection. Up to now, there is no standard HDDT, so different researchers have adopted different specific schemes. In Yang\\'s study, both amoxicillin and PPI were given 4 times daily for 14 days.^\\[[@R34]\\]^ A recent meta-analysis defined HDDT as taking amoxicillin \u22652.0\u200ag/day, amoxicillin or PPI 3 or 4 times daily.^\\[[@R36]\\]^\n\nIn our meta-analysis, we treated both amoxicillin and PPI be given 4 times daily for 14 days (Yang\\'s criteria) as the standard, running RCT comparing HDDT with BQT. We found 4 RCTs met Yang\\'s criteria,^\\[[@R34]\\]^ the efficacy of *H pylori* eradication show ITT and PP eradication rates were 75.6% to 94.7% and 81.2% to 96.4%, and the combined eradication rates are 85.5% (ITT) and 88.4% (PP) respectively, no significant differences compared HDDT with BQT were observed (Figs. [2](#F2){ref-type=\"fig\"} and [3](#F3){ref-type=\"fig\"}). We found that both HDDT and BQT can achieve similar eradication rates for *H pylori* infection.\n\n*H pylori* resistance to antibiotics plays a key role in the failure of the treatment. Bismuth can improve eradication rates without resistance, and is safe for short-term used.^\\[[@R53]\\]^ There is synergistic effect of bismuth combined with antibiotics, so the addition of bismuth might enhance the effectiveness of triple therapies.^\\[[@R54]\\]^ Bismuth can also partly overcome *H pylori* resistance to clarithromycin,^\\[[@R55]--[@R57]\\]^ metronidazole,^\\[[@R58],[@R59]\\]^ and levofloxacin.^\\[[@R60],[@R61]\\]^ On the other hand, unlike clarithromycin-containing triple therapy or levofloxacin-containing triple therapy, the efficacy of HDDT will not gradually decrease with the use of amoxicillin in terms of that *H pylori* resistance to amoxicillin, both primary and acquired, is rare.^\\[[@R11]--[@R14]\\]^ Therefore we included studies regardless of first-line or rescue therapy for *H pylori* infection in this meta-analysis.\n\nBQTs are believed to perform better than other *H pylori* eradication therapies.^\\[[@R54]--[@R62]\\]^ Currently, strong consensus was reached that classic BQT (PPI-bismuth-tetracycline-metronidazole) has been recommended for *H pylori* infection.^\\[[@R2],[@R7],[@R9],[@R10]\\]^ Despite Zhang et al^\\[[@R63]\\]^ and Wu et al^\\[[@R64]\\]^ have conducted meta-analysis of studies comparing BQT with quinolone-based (moxifloxacin, levofloxacin) triple therapy for *H pylori* eradication, and they found that quinolone-based triple regimen is more effective and well tolerated than BQT in the treatment. Recently, resistance rates of *H pylori* to quinolone were over 30%, and the rescue therapy with PPI-amoxicillin-levofloxacin still failed in \\>20% of patients.^\\[[@R65],[@R66]\\]^\n\nIn addition, side effects were more likely in BQT than in HDDT (Fig. [5](#F5){ref-type=\"fig\"}). A high rate of adverse events with BQT may decrease its compliance. Bismuth agents and tetracycline are also not available in some geographic areas. Moreover, *H pylori* showed higher risk of secondary resistance after treatment failure of the BQT than HDDT, which makes it harder with rescue treatment for *H pylori* eradication. Therefore, HDDT is an effective therapy in patients who are not allergic to penicillin; while for patients who are allergic to penicillin, BQT is a good treatment option for *H pylori* infection.\n\nThere are some limitations of our study. First, the patients were enrolled in RCTs are mainly from Asia. The response to HDDT might be affected by CYP2C19 polymorphisms, so further research needs to be taken among different people. Second, all included trials were not carried out in a blind manner, which may lead to heterogeneity among included studies. Finally, all RCTs in this meta-analysis used the different kinds of PPIs, and 1 study had 2 PPI dosage in HDDT group, and different dosage and antibiotics in BQT group, resulting in publication bias.\n\nIn conclusion, our findings showed that HDDT was comparable to BQT for *H pylori* infection. HDDT is effective and safe. In geographical areas with high antibiotic resistance, empirical treatment with HDDT would potentially achieve higher eradication rates (for non*-*penicillin*-*allergic patients) because the overall rate of amoxicillin resistance is low worldwide. Future research should be directed in comparing the 2 therapies, also in terms of different antibiotics composition and therapies based on antibiotic susceptibility testing.\n\nAuthor contributions\n====================\n\nYang X et al high dose dual therapy versus bismuth quadruple therapy for *H pylori*: a meta-analysis;\n\nYang X, Wang JX, and Gao CP acquired, analyzed and interpreted the data, and drafted the article;\n\nGao CP contributed to conception and design of the study;\n\nHan SX critically revised the manuscript; and all authors approved the final version.\n\n**Conceptualization:** Xue Yang, Cai-Ping Gao.\n\n**Data curation:** Jin-xia Wang, Sheng-Xi Han.\n\n**Formal analysis:** Jin-xia Wang, Sheng-Xi Han.\n\n**Funding acquisition:** Cai-Ping Gao.\n\n**Investigation:** Xue Yang, Cai-Ping Gao.\n\n**Methodology:** Sheng-Xi Han.\n\n**Software:** Jin-xia Wang, Sheng-Xi Han.\n\n**Writing -- original draft:** Xue Yang.\n\n**Writing -- review & editing:** Xue Yang, Cai-Ping Gao.\n\nSupplementary Material\n======================\n\n###### Supplemental Digital Content\n\nAbbreviations: BQT = bismuth quadruple therapy, CIs = confidence intervals, *H pylori* = *Helicobacter pylori*, HDDT = high-dose dual therapy, ITT = intention-to-treat, PP = per-protocol, RCTs = randomized clinical trials.\n\nNational Natural Science Foundation of China (81001083);\n\nSichuan Academy of Science & Sichuan Provincial People\\'s Hospital (2016LY06).\n\n*Helicobacter pylori* (*H pylori*) treatment still remains a challenge. Currently, bismuth quadruple therapy (BQT) has been widely used to eradicate *H pylori*. High dose dual therapy (HDDT) is an alternative treatment with high efficacy. Our meta-analysis revealed that both BQT and HDDT can achieve similar eradication rate and adherence, and generally HDDT causes fewer side effects.\n\nThe authors deny any conflict of interest.\n\nSupplemental Digital Content is available for this article.\n"}
+{"text": "Introduction {#H1-1-ZLD200107}\n============\n\nHigh-quality evidence generated by appropriately powered and controlled trials is needed to advance care for patients with coronavirus disease 2019 (COVID-19) and those who are susceptible to it.^[@zld200107r1],[@zld200107r2]^ In the midst of the COVID-19 pandemic, multiple similar therapeutic trials are being conducted in parallel, potentially reducing participant accrual across trials. In this systematic review, we characterize the landscape of current COVID-19 trials to better quantify these potential issues.\n\nMethods {#H1-2-ZLD200107}\n=======\n\nInstitutional review board approval of this study was waived because it exclusively used publicly available data without any protected health information. Screening and trial selection adhered to the Preferred Reporting Items for Systematic Reviews and Meta-analyses ([PRISMA](http://www.equator-network.org/reporting-guidelines/prisma/)) reporting guideline.^[@zld200107r3]^\n\nWe performed a data query of the ClinicalTrials.gov registry for interventional trials in any clinical phase regarding COVID-19 on June 8, 2020. Advanced search parameters included *COVID-19*, *SARS-CoV-2*, *2019-nCoV*, *2019 novel coronavirus*, and *severe acute respiratory syndrome coronavirus 2*. Data were analyzed using SPSS statistical software version 26 (IBM Corp). Data analysis was performed in June 2020.\n\nResults {#H1-3-ZLD200107}\n=======\n\nOur search yielded 674 trials after removing suspended and halted trials ([Figure](#zld200107f1){ref-type=\"fig\"}). Most were randomized multigroup studies (562 of 674 trials \\[83.4%\\]). Only 479 of 674 randomized trials (71.1%) included a control group deemed to be valid at the time of data curation (including either standard of care or placebo as the control group). Most of the trials assessed treatment of COVID-19 (570 of 674 trials \\[84.6%\\]) rather than its prevention (104 of 674 trials \\[15.4%\\]). Of randomized studies, only 201 (35.8%) were multicenter trials ([Table](#zld200107t1){ref-type=\"table\"}). Chloroquines were the most commonly tested intervention (132 of 562 randomized trials \\[23.5%\\]; 143 trials total). Among the 201 trials accruing in the US alone, the total expected enrollment was 146\u2009688 participants. This included 33 COVID-19 prevention trials with a planned total accrual of 100\u2009746 participants, of which 86\u2009950 participants (86.3%) were planned to accrue to chloroquine-specific COVID-19 prevention trials. Similarly, there were 168 US-accruing COVID-19 treatment trials with a planned total accrual of 45\u2009942 participants, of which 13\u2009542 participants (29.5%) were planned to accrue to chloroquine-specific COVID-19 treatment trials. Primary end points most commonly assessed among randomized studies were time to symptom and sign resolution (212 trials \\[37.7%\\]), mortality (180 trials \\[32.0%\\]), viral clearance (124 trials \\[22.1%\\]), and mechanical ventilation (57 trials \\[10.1%\\]) ([Table](#zld200107t1){ref-type=\"table\"}).\n\n![Flowchart of Screening, Eligibility, and Inclusion of Randomized Clinical Trials for Coronavirus Disease 2019 (COVID-19)](jamanetwopen-3-e2015100-g001){#zld200107f1}\n\n###### Characteristics of Coronavirus Disease 2019 Clinical Trials\n\n  Characteristic                                                Trials, No. (%)   \n  ------------------------------------------------------------- ----------------- -----------\n  Multicenter trials                                            201 (35.8)        23 (20.5)\n  Multinational trials                                          22 (3.9)          1 (0.9)\n  Any blinding[^b^](#zld200107t1n2){ref-type=\"table-fn\"}        331 (58.9)        3 (2.7)\n  Intervention[^c^](#zld200107t1n3){ref-type=\"table-fn\"}                          \n  Chloroquines[^d^](#zld200107t1n4){ref-type=\"table-fn\"}        132 (23.5)        11 (9.8)\n  Biologicals                                                   177 (31.5)        60 (53.6)\n  Convalescent plasma                                           30 (5.4)          18 (16.1)\n  Tocilizumab                                                   21 (3.7)          6 (5.4)\n  Tyrosine kinase inhibitor                                     20 (3.6)          12 (10.7)\n  Antivirals                                                    55 (9.8)          1 (0.9)\n  Remdesevir                                                    9 (1.6)           0\n  Protease inhibitors                                           37 (6.8)          1 (0.9)\n  Antibiotics                                                   49 (8.7)          5 (4.5)\n  Azithromycin                                                  40 (7.1)          4 (3.6)\n  Primary end point[^e^](#zld200107t1n5){ref-type=\"table-fn\"}                     \n  Time to symptom and sign resolution                           212 (37.7)        51 (45.5)\n  Mortality                                                     180 (32.0)        23 (20.5)\n  Viral clearance                                               124 (22.1)        16 (14.3)\n  Need for mechanical ventilation                               57 (10.1)         5 (4.5)\n  Industry sponsorship                                          175 (31.1)        19 (17.0)\n\nNonrandomized trials included both single-group and nonrandomized multiple-group trials.\n\nBlinding included single, double, triple, and quadruple blinding.\n\nInterventions were counted independently as many trials included multiple interventions.\n\nChloroquines included hydroxychloroquine and chloroquine.\n\nPrimary end points were counted independently because many trials included multiple primary end points.\n\nDiscussion {#H1-4-ZLD200107}\n==========\n\nWe found a high rate of trial multiplicity, particularly with chloroquines, which are being tested in 143 studies. Although these overlapping trials may afford opportunities for replication and validation, the high degree of multiplicity also enhances the likelihood of finding a positive result through chance alone, potentially resulting in widespread use of an ineffective and possibly hazardous intervention.^[@zld200107r4]^\n\nThe fragmentation of efforts could also lead to unnecessary competition for participants, which potentially compromises trial accrual and statistical power for all trials. This worrisome scenario has already occurred in China.^[@zld200107r5]^ Because the projected participant accrual for US-only COVID-19 treatment trials is approximately 45\u2009942 participants (13\u2009542 to chloroquines alone), it seems unlikely that this target will be achieved given the intrinsic challenges of participant accrual during an active pandemic.\n\nNotably, our study is limited through the use of a single US-based clinical trials registry, potentially representing only a fraction of the worldwide COVID-19--related trials portfolio. Furthermore, the ever-changing landscape of COVID-19 clinical research amid this pandemic may complicate future interpretation of this report.\n\nAlthough current trials have been initiated with the best intentions, the medical community must be mindful of the potential issues of incomplete participant accrual and publication bias that are introduced by enabling dozens of similar trials simultaneously.^[@zld200107r6]^ Together, these factors endanger the capacity to rapidly produce meaningful evidence that is vital during this critical time. Avoiding these pitfalls requires coordination of efforts. This could be achieved, in part, through makeshift cooperative groups to improve participant accrual and decrease duplicative efforts. Institutional review boards and regulators (including the US Food and Drug Administration) must also work together to responsibly ease roadblocks to coordinate pooled analyses across trials. These efforts should include synchronization and standardization of end points, focusing on the most meaningful and objective outcomes (eg, all-cause mortality, intensive care unit admission, and mechanical ventilation). It is hoped that these measures will expedite generation of high-quality prospective data to guide effective treatments while maximizing resource allocation.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nCancer is one of the pathological scenarios where apoptosis is inactivated, resulting in accumulation of malignant cells that will not die. Ovarian cancer is the seventh most common cancer in women worldwide, which accounts for nearly 4% of all new cases of cancer in women. This lethal malignancy usually diagnosed at later stage with a 5 years survival rate of below 30%. Since most of the ovarian cancer cases are diagnosed after wide spreading of tumors within the peritoneal cavity, it is restraining the effectiveness of chemotherapy, and associated death is believed to be therapy-resistant metastasis^[@CR1],\\ [@CR2]^. The prognosis is usually poor for ovarian cancer patients, because the disease reaching an advanced stage before it is discovered^[@CR3]^. Despite overall declining death rates for cervical and uterus cancers, the annual report of ovarian cancer mortality has risen by 250% since 1930^[@CR4]^. Primary cytoreductive surgery followed by chemotherapy with anticancer agents is the standard treatment regimen for patients with ovarian cancer, which improved survivals. Nevertheless, most of the patients with advanced cancer will eventually relapse and die of their cancer^[@CR2],\\ [@CR5]^. A preliminary study showed improved efficacy of chemotherapy and normalized cancer biomarkers following a high-dose of antioxidant (containing coenzyme Q~10~) treatment with chemotherapy in patients with advanced ovarian cancer^[@CR6]^. Management of ovarian cancers is still in high demand for effective therapy without adverse effects caused by the therapeutic agents.\n\nDespite being the cause of problem, apoptosis or programmed cell death plays an important role in treatment of cancers, as it is a popular target of many therapeutic strategies. Mounting evidence revealed that induction of apoptosis by chemical substances or pro-apoptotic agents eventually controlled the spreading of cancer^[@CR7]--[@CR9]^. Apoptotic signals are reported to be triggered by several cellular events, including excessive production of reactive oxygen species (ROS), mitochondrial dysfunction and ER stress^[@CR8],\\ [@CR10],\\ [@CR11]^. The apoptotic signals either intrinsic or extrinsic cascades lead to activate the caspases, a class of cysteine proteases that cleave different substrates and ultimately leading to cell dismantling and DNA fragmentation^[@CR12],\\ [@CR13]^. In apoptosis, caspases can also cleave Beclin-1, an autophagic protein, and inhibit its pro-autophagic activity. Autophagy, a process of programmed cell survival or an adaptive response is activated during periods of cellular distress and extinguished during the cell-cycle^[@CR14]^. At the molecular level, autophagy is regulated by a battery of signals, such as microtubule-associated protein light chain-3 (LC3), mammalian target of rapamycin (mTOR), autophagy-related protein 7 (ATG7) and Beclin-1. In addition, formation of acidic vesicular organelles (AVOs) is a hallmark of autophagy^[@CR15],\\ [@CR16]^. Some anticancer agents that can induce apoptosis also stimulate autophagy as a death mechanism^[@CR17]^. Recent developments in cancer research suggest that autophagy could be an additional target for adjuvant anticancer treatment, and inhibition of autophagy could strengthen the therapeutic efficacy of cancer treatment^[@CR18],\\ [@CR19]^. However, whether autophagy promotes or inhibits cancer cell death in response to cellular stress is controversial. Indeed, there is a crosstalk between autophagy and apoptosis as they share common stimuli and signalling pathways^[@CR14],\\ [@CR20],\\ [@CR21]^. Thus, understanding of the complex relation between apoptosis and autophagy in cancer therapy could enhance the treatment effectiveness.\n\nCoenzyme Q (CoQ) is a well-known biomolecules comprised of a quinone nucleus and a hydrophobic side chain containing variable number of *trans*-isoprenoid units. CoQ~10~ is the major naturally occurring form of CoQ that containing 10 isoprenoid units, while CoQ~0~ is the novel analog without isoprenoid side chains^[@CR22]^. Various CoQ analogs or ubiquinone (Ub) have been reported to either increase or decrease the production of ROS, and involved in opening/closing of mitochondrial permeability transition pore (PTP), which relies on cellular context^[@CR8],\\ [@CR23]^. CoQ analogs with shorter isoprenoid side chains (CoQ~2~ and CoQ~4~) reported to induce apoptosis in mutated BALL-1 cells, but not longer isoprenoid chains (CoQ~6~ and CoQ~10~)^[@CR24]^. Among various CoQ analogs (Ub~5~ and Ub~10~), CoQ~0~, a redox active compound profoundly triggered the ROS production in Clone-9 cells, induced PTP opening in cancerous rat liver MH1C1 cells and promote cell death^[@CR23]^. CoQ~0~ has been shown as a potent cytotoxic compound towards human breast cancer cells by induction of apoptosis and cell-cycle arrest^[@CR25]^. Our recent findings showed anti-angiogenic property of CoQ~0~ in stimulated human endothelial cells^[@CR26]^.\n\nPreviously, we have shown the potent anticancer properties of *Antrodia camphorata* against human ovarian cancers via induction of apoptosis and cell-cycle arrest^[@CR27]^. Several studies showed free radical scavenging or antioxidant activity^[@CR28],\\ [@CR29]^ and anticancer properties of *Antrodia camphorata*, a well-known medicinal mushroom in Taiwan^[@CR30]--[@CR32]^. However, those studies are limited to claim the responsible bioactive compounds, and yet no study to demonstrate the anticancer properties of CoQ~0~, a key ingredient in *Antrodia camphorata*. In this study, we used CoQ~0~ that is isolated from *Antrodia camphorata*, and investigated its anticancer potentials in human ovarian cancer (SKOV-3) cells and xenografted nude mice. To distinguish the underlying molecular mechanisms, the effect of CoQ~0~ on ROS production; and the role of ROS on cancer cell survival, apoptosis and autophagy were examined by determining the key molecules involved in regulation of apoptosis and autophagy in SKOV-3 cells.\n\nResults {#Sec2}\n=======\n\nCoQ~0~ inhibits viability and growth of human ovarian carcinoma cells {#Sec3}\n---------------------------------------------------------------------\n\nPrior to explore its anticancer properties, we examined the cytotoxic effects of CoQ~0~ on human ovarian carcinoma cell lines (SKOV-3, A2780 and A2870/CP70) and normal ovarian surface epithelial (IOSE) cells. Treatment of cells with increasing concentrations of CoQ~0~ (0--40\u2009\u00b5M, 24\u2009h) dose-dependently decreased viability of SKOV-3, A2780 and A2870/CP70 cells with the IC~50~ values of 26.6, 27.3 and 28.4\u2009\u00b5M, respectively (Fig.\u00a0[1B--D](#Fig1){ref-type=\"fig\"}). The cytotoxic concentration of CoQ~0~ on normal IOSE cell lines is \\>40\u2009\u00b5M (Fig.\u00a0[1E](#Fig1){ref-type=\"fig\"}). Inhibition of SKOV-3 cell viability with CoQ~0~ was prominent event at 20\u2009\u00b5M concentration compared to A2780 and A2870/CP70 cell lines (Fig.\u00a0[1B](#Fig1){ref-type=\"fig\"}). Higher concentration of CoQ~0~ caused severe abrupt morphological changes, which were represented by cell shrinkage and decreased density of SKOV-3 cells (Fig.\u00a0[1F](#Fig1){ref-type=\"fig\"}). We assume that the cytotoxic effect of CoQ~0~ may be due to the induction of cell-cycle arrest.Figure 1CoQ~0~ inhibits growth of human ovarian carcinoma cells and induces G2/M cell-cycle arrest in SKOV-3 cells. (**A**) Structure of CoQ~0~. (**B--D**) Human ovarian carcinoma **(**SKOV-3, A2870 and A2870/CP-70) and **(E)** human ovarian surface epithelial (IOSE) cells were treated with increasing concentrations of CoQ~0~ (0--40\u2009\u00b5M) for 24\u2009h. Cell viability was determined using MTT assay. (**F**) Morphological changes in CoQ~0~-treated (0--40\u2009\u00b5M, 24\u2009h) SKOV-3 cells were examined by phase-contrast microscope (200\u2009\u00d7\u2009magnification). (**G**) SKOV-3 cells were treated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h, stained with PI and analyzed for cell-cycle phase using flow cytometry. The cellular distributions (percentage) in different phases of cell-cycle (G1, S and G2/M) were determined after CoQ~0~ treatment. Flow cytometry images shown here are from one representative analysis that was repeated three times with similar results. (**H**) SKOV-3 cells were treated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h, and cell-cycle regulatory proteins, cyclin B1, CDK1, cyclin A and CDK2 were examined using Western blot. Relative changes in protein intensities were quantified by commercially available software, and presented as histogram, control being as 1-fold. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093), and significant at \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control cells.\n\nCoQ~0~ induces G~2~/M cell-cycle arrest and reduces cell-cycle proteins in SKOV-3 cells {#Sec4}\n---------------------------------------------------------------------------------------\n\nTo address whether CoQ~0~ induces cell-cycle arrest in SKOV-3 cells, we measured the distribution of cells (%) in different phases of cell-cycle following CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h). Flow cytometry data showed that CoQ~0~ resulted in a progressive and sustained accumulation of cells in G~2~/M phase, while cells in G~1~ phase were gradually decreased in a dose-dependent fashion. The accumulation of cells in G~2~/M phase is \\~30.5% with 30\u2009\u00b5M CoQ~0~, whereas cells in G~2~/M phase are only 8.9% without CoQ~0~ treatment (Fig.\u00a0[1G](#Fig1){ref-type=\"fig\"}). Concurrently, the expressions of cell-cycle regulatory proteins, including cyclin B1, cyclin dependent kinase 1 (CDK1), cyclin A and CDK2 were dose-dependently decreased with CoQ~0~ (Fig.\u00a0[1H](#Fig1){ref-type=\"fig\"}). These results explain that CoQ~0~ considerably inhibited ovarian cancer (SKOV-3) cell proliferation through induction of G~2~/M cell-cycle arrest and reduction of cell-cycle regulatory proteins.\n\nCoQ~0~ triggers intracellular ROS levels to promote SKOV-3 cell death {#Sec5}\n---------------------------------------------------------------------\n\nExcessive generation of ROS by oxidants/chemical substances is reported to potentiate cancer cell death *via* apoptosis or autophagy mechanisms^[@CR33]^. We found treatment of SKOV-3 cells with CoQ~0~ (30\u2009\u00b5M) for 0--30\u2009min remarkably increased the intracellular ROS levels. Especially, ROS levels at 15\u2009min following CoQ~0~ treatment were significantly higher (\\~38 fold) than the control, as evidenced by increased dichlorofluorescein (DCF) (Fig.\u00a0[2A and B](#Fig2){ref-type=\"fig\"}). Interestingly, cells incubated with ROS inhibitor (N-acetylcysteine \\[NAC\\], 2\u2009mM) 1\u2009h prior to CoQ~0~ treatment (0--30\u2009\u00b5M, 15\u2009min) substantially inhibited the ROS production (Fig.\u00a0[2C and D](#Fig2){ref-type=\"fig\"}). We further demonstrated that NAC pretreatment completely reversed the CoQ~0~-induced death of SKOV-3 cells (Fig.\u00a0[2E](#Fig2){ref-type=\"fig\"}). These findings suggest that CoQ~0~ triggered intracellular ROS generation and that are possibly contribute to death of SKOV-3 cells.Figure 2CoQ~0~ induces intracellular ROS generation in SKOV-3 cells. (**A**) Cells were treated with CoQ~0~ (30\u2009\u00b5M) for 0--30\u2009min and generation of intracellular ROS were measured using fluorescent microscopy (200\u2009\u00d7\u2009magnification). The non fluorescent probe DCFH~2~-DA reacts with cellular ROS and metabolized into fluorescent DCF. (**B**) The fluorescence intensity of DCF-stained cells, represent the levels of ROS was quantified by Olympus Soft Imaging Solution, and presented as histogram. Results are significant at \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared to time 0\u2009min. (**C**) Cells were treated with CoQ~0~ (10--30\u2009\u00b5M) for 15\u2009min, and ROS generation was measured in the presence or absence of ROS inhibitor (2\u2009mM NAC, 1\u2009h prior to CoQ~0~). The levels of intracellular ROS were indicated by strong or weak fluorescence intensity. (**D**) ROS levels were quantified and expressed in bar diagram as percentage of control. (**E**) Cells were treated with ROS inhibitor (NAC, 2\u2009mM) 1\u2009h prior to CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h), and cell viability of was determined by MTT assay. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093), and significant at ^\\#^ *p*\u2009\\<\u20090.1; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nCoQ~0~ promotes LC3 accumulation and AVOs formation in SKOV-3 cells {#Sec6}\n-------------------------------------------------------------------\n\nLC3, a promising autophagy marker exists in LC3--1 (cytosolic) and LC3\u2010II (membrane bound) forms. The conversion of LC3--1 to LC3--1I or accumulation of LC3\u2010II is correlated with the extent of autophagosome formation or increased numbers of AVOs in cells^[@CR15],\\ [@CR16]^. To address whether CoQ~0~ could induce autophagy, we measured the conversion of LC3--1 to LC3--1I and AVOs formation following CoQ~0~ treatment (0--30\u2009\u00b5M). As presented in Fig.\u00a0[3A](#Fig3){ref-type=\"fig\"}, CoQ~0~ dose-dependently increased the LC3-II accumulation in SKOV-3 cells. The sequential effect of CoQ~0~ on AVOs formation was measured *via* fluorescence microscope using acridine orange (AO) stain. Arrows on images clearly indicating the increased appearance of AVOs (red fluorescence), following CoQ~0~ treatment (Fig.\u00a0[3B](#Fig3){ref-type=\"fig\"}). The high dose of CoQ~0~ (30\u2009\u00b5M) resulted in large number of AVOs (\\>10 fold) (Fig.\u00a0[3B](#Fig3){ref-type=\"fig\"}) that is corresponding to the greater accumulation of LC3-II in SKOV-3 cells.Figure 3CoQ~0~ promotes cytoprotective autophagy as a survival mechanism in SKOV-3 cells. (**A**) Cells were treated with various concentrations of CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h and then conversion of LC3-I to LC3-II was determined by Western blot. Relative changes in the intensities of protein bands were quantified by commercially available quantitative software. (**B**) CoQ~0~ induces AVOs formation. Cells were treated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h and stained with AO. Formation of AVOs, represented by red fluorescence intensity (in lysosomes) was visualized under a red filter fluorescence microscope (100\u2009\u00d7\u2009magnification). Number of AO stained cells was presented as histogram, control being as 1.0 fold. (**C**) CoQ~0~ promotes conversion of GFP-LC3. Cells were transfected with GFP-LC3 expression vector for 24\u2009h, and then treated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h. GFP-LC3 dots in cells were observed under a confocal microscope (200 \u00d7 magnification). Conversions of GFP-LC3 and endogenous LC3 were determined by Western blot. (**D**) Cells were treated with autophagy inhibitors (2\u2009mM 3-MA or 10 \u03bcM CQ) for 1\u2009h followed by CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h, and viability was assayed by MTT assay. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control, and significant at ^\\#\\#^ *p*\u2009\\<\u20090.01; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nCoQ~0~ enhances GFP-LC3 conversion in SKOV-3 cells {#Sec7}\n--------------------------------------------------\n\nTo confirm CoQ~0~-induced autophagy, GFP-LC3 plasmid was transiently transfected into the SKOV-3 cells, and conversion of GFP-LC3 and endogenous LC3 levels were monitored following CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h). Images from confocal microscopy depicts that CoQ~0~-treated cells represented by a cornucopia of green LC3 punctate dots in the cytoplasm, while the control cells showed a diffused and weak LC3 punctate dots (Fig.\u00a0[3C](#Fig3){ref-type=\"fig\"}). We demonstrated that CoQ~0~ increased both the percentage of cells with GFP-LC3 dots and the average numbers of GFP-LC3 dots per cell in a dose-dependent fashion (Fig.\u00a0[3C](#Fig3){ref-type=\"fig\"}). Western blot results further convinced that CoQ~0~ treatment significantly promoted the conversion of LC3--1 to LC3-II in SKOV-3 cells (Fig.\u00a0[3C](#Fig3){ref-type=\"fig\"}).\n\nCoQ~0~ activates autophagy as a survival mechanism in SKOV-3 cells {#Sec8}\n------------------------------------------------------------------\n\nAutophagy has been claimed to play a paradoxical role in controlling of cell death and survival in response to various stimuli^[@CR34]^. Since CoQ~0~ shown to activate autophagy in SKOV-3 cells, we wonder whether this autophagy could contribute to cell death or survival. To disclose this phenomenon, SKOV-3 cells were pretreated with autophagy inhibitors, 3-methyladenine (3-MA, inhibitor of early autophagy/LC3-II accumulation) or chloroquine (CQ, inhibitor of late autophagy/promoter of LC3-II accumulation), and cell viability was assayed following CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h). Interesting results showed that inhibition of autophagy by 3-MA (2\u2009mM) or CQ (10\u2009\u00b5M) didn't suppress the CoQ~0~-mediated cell death, instead exacerbates the cell death (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"}). These findings imply that CoQ~0~-induced autophagy is not involved in death of SKOV-3 cells, which might be a cell survival mechanism.\n\nCoQ~0~ triggers apoptotic death of SKOV-3 cells via mitochondrial and ER-stress signals {#Sec9}\n---------------------------------------------------------------------------------------\n\nApoptotic-cell death, a key strategic phenomenon in management of cancers is mediated by either mitochondrial or ER-stress signalling cascades^[@CR10],\\ [@CR11]^. To explore the patterns of apoptosis induced by CoQ~0~ in SKOV-3 cells, the key molecular proteins involved in mitochondrial and ER-stress related apoptosis were determined by Western blot. As shown in Fig.\u00a0[4A](#Fig4){ref-type=\"fig\"}, procaspase-3, a highly expressed inactive form prior to CoQ~0~ incubation, was dose-dependently cleaved to active caspase-3 following CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h). It is known that caspase-3-mediated selective proteolytic cleavage of poly (ADP-ribose) polymerase (PARP) is a hallmark of apoptosis^[@CR35]^. Thus, we measured the PARP levels in SKOV-3 cells, and found that CoQ~0~ increased the proteolytic cleavage of 116\u2009kDa PARP to 89\u2009kDa fragment (Fig.\u00a0[4A](#Fig4){ref-type=\"fig\"}). Next we investigated the effect of CoQ~0~ on ER-stress mediated apoptosis by detecting the changes of caspase-12 and heat shock protein-70 (HSP-70) that are implicated in ER-stress apoptosis^[@CR11]^. We found CoQ~0~ (0--30\u2009\u00b5M) remarkably increased caspase-12 and HSP-70 expressions in a dose-dependent manner (Fig.\u00a0[4A](#Fig4){ref-type=\"fig\"}). These findings suggest that CoQ~0~ activates both mitochondrial and ER-stress mediated apoptosis in SKOV-3 cells.Figure 4CoQ~0~ induces apoptosis through mitochondrial and ER stress pathways in SKOV-3 cells. (**A**) Cells were treated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h, and apoptotic proteins involved in mitochondrial pathway (procaspse-3, caspase-3 and PARP) and ER stress pathway (caspase-12 and Hsp70) were estimated by Western blot. Changes in protein intensities were quantified by commercially available software, and presented as histogram. **(B)** Annexin-V-FITC and PI staining was used to identify the early/late apoptosis or necrosis of cells followed by CoQ~0~ (0--30\u2009\u00b5M) treatment. Results from flow cytometry analysis in each quadrant (Q) are labeled and interpreted as follows: (Q1) PI positive, Annexin-V-FITC-negative stained cells/necrosis. (Q2) PI positive, Annexin-V-FITC-positive stained cells/late apoptosis. (Q3) cells negative for both PI and Annexin-V-FITC staining/normal live cells. (Q4) PI-negative, Annexin-V-FITC-positive stained cells/early apoptosis. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093), and significant at \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control cells.\n\nCoQ~0~ promotes apoptosis in SKOV-3 cells {#Sec10}\n-----------------------------------------\n\nTo address whether CoQ~0~ promotes apoptosis in SKOV-3 cells, next we performed Annexin-V- fluorescein isothiocyanate (FITC) and propidium iodide (PI) assay, which stain phosphatidylserine and DNA residues, respectively^[@CR36]^. The flow cytometry results showed that treatment of SKOV-3 cells with CoQ~0~ (0--30\u2009\u00b5M) dose-dependently increased the early and late apoptotic cells. The early apoptotic cells (Annexin-V-FITC-positive, PI-negative) represented in Q4 are 25.1%, 34% and 36% with 10, 20 and 30\u2009\u00b5M CoQ~0~ treatments respectively (Fig.\u00a0[4B](#Fig4){ref-type=\"fig\"}).\n\nCoQ~0~ potentiates apoptotic DNA fragmentation and inhibition of apoptosis reversed death of SKOV-3 cells {#Sec11}\n---------------------------------------------------------------------------------------------------------\n\nCoQ~0~-induced apoptotic DNA fragmentation was determined using TUNEL assay (terminal deoxynucleotidyl transferase dUTP nick end labeling). The fluorescent microscope images revealed that CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h) significantly increased the green fluorescence, TUNEL-positive cells (Fig.\u00a0[5A](#Fig5){ref-type=\"fig\"}), which denotes increased apoptotic DNA fragmentation. However, CoQ~0~-induced DNA fragmentation (\\~6-fold, 30\u2009\u00b5M) was significantly (*P*\u2009\\<\u20090.001) diminished in cells pretreated with caspase inhibitor, Z-VAD-FMK (benzyloxycarbonyl-valyl-alanyl-aspartyl-\\[O-methyl\\]-fluoromethylketone, 20 \u03bcM) (Fig.\u00a0[5A and B](#Fig5){ref-type=\"fig\"}). Furthermore, CoQ~0~-induced PARP cleavage was also suppressed in the presence of Z-VAD-FMK, which indicates CoQ~0~ activates caspase mediated apoptosis in SKOV-3 cells (Fig.\u00a0[5C](#Fig5){ref-type=\"fig\"}). To confirm that CoQ~0~-induced death of SKOV-3 cell *via* activation of apoptosis, cells were pretreated with Z-VAD-FMK (20 \u03bcM, 1\u2009h), and cell survival was assayed following CoQ~0~ treatment (0--30\u2009\u00b5M, 24\u2009h). We found that CoQ~0~-induced death of SKOV-3 cells was predominantly reversed by inhibition of apoptosis (Fig.\u00a0[5D](#Fig5){ref-type=\"fig\"}). These findings suggest that CoQ~0~ provoked apoptotic signals contribute to death of ovarian cancer cells.Figure 5CoQ~0~ triggers apoptotic DNA fragmentation and promote death of SKOV-3 cells. (**A--D**) Cells were pretreated with caspase inhibitor (Z-VAD-FMK, 20 \u03bcM) for 1\u2009h followed by CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h. **(A)** Apoptotic DNA fragmentation was determined by TUNEL assay. The green florescence indicates TUNEL-positive cells in the microscopic fields (200 \u00d7 magnification) from three separate samples. **(B)** The fold of apoptotic cells was calculated by quantifying the florescence intensity using commercially available software. **(C)** Cleavage of PARP was estimated by Western blot. Changes in protein intensities were quantified by commercially available software. **(D)** Cell viability with or without Z-VAD-FMK treatment was determined by MTT assay. Values expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control, and significant at ^\\#\\#^ *p*\u2009\\<\u20090.01; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nCoQ~0~ increases Beclin-1/Bcl-2 and Bax/Bcl-2, and inhibits HER-2/neu/AKT/mTOR signalling in SKOV-3 cells {#Sec12}\n---------------------------------------------------------------------------------------------------------\n\nSince CoQ~0~ reported to induce apoptosis/autophagy in SKOV-3 cells, we studied the effect of CoQ~0~ (0--30\u2009\u00b5M) on Bcl-2, and its role on Bax (pro-apoptotic) and Beclin-1 (pro-autophagic) expressions. Bcl-2 is reported to decrease the pro-autophagic property of Beclin-1, but Beclin-1 unable to neutralize the apoptotic function of Bcl-2^[@CR14]^. In our study, we demonstrated that CoQ~0~ treatment substantially decreased the Bcl-2 expression, while increased the Beclin-1 and Bax expressions in a dose-dependent fashion (Fig.\u00a0[6A](#Fig6){ref-type=\"fig\"}). Degradation of Bcl-2 probably activates apoptosis in SKOV-3 cells. The dose-dependent increase of Bax/Bcl-2 ratio with CoQ~0~ represents the propagation of apoptotic mechanism in SKOV-3 cells (Fig.\u00a0[6B and C](#Fig6){ref-type=\"fig\"}).Figure 6CoQ~0~ increases Beclin-1/Bcl-2 and Bax/Bcl-2 ratios, and inhibits HER-2/*neu*/AKT/mTOR signalling in SKOV-3 cells. **(A)** Dose-dependent effect of CoQ~0~ (0--30\u2009\u00b5M, 24\u2009h) on changes in Beclin-1, Bax and Bcl-2 proteins were determined by Western blot. Relative changes in the ratio of **(B)** Beclin-1/Bcl-2 and **(C)** Bax/Bcl-2 in accordance to dose were quantified by commercially available software, and presented as a histogram, control representing as 1.0 fold. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093), and significant at \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control cells. **(D)** Time-dependent effect of CoQ~0~ (0--24\u2009h, 30\u2009\u00b5M) on phosphorylation of HER-2/*neu* (Y1221), AKT (Ser473) and mTOR (S2448), and cleavage of PARP were determined by Western blot. Results expressed as mean\u2009\u00b1\u2009SD of two independent assays (n\u2009=\u20092).\n\nHuman epidermal growth factor receptor-2 (HER-2 *neu*) is a proto-oncogene implicated in malignant transformation, and overexpression of HER-2 has been found to aggressively promote the AKT/mTOR signals, which are responsible for regulation of tumor biology, including cancer cell invasion, differentiation and survival^[@CR37]^. We found that CoQ~0~ treatment (30\u2009\u00b5M, 0--24\u2009h) downregulated the phosphorylated HER-2 (Y1221) levels. This decrease was accompanied with substantial time-dependent loss of p-AKT (Ser473) and p-mTOR (S2448) levels (Fig.\u00a0[6D](#Fig6){ref-type=\"fig\"}). CoQ~0~ enhanced the proteolytic cleavage of PARP (116 KDa to 89 KDa), which supports activated apoptotic signals in cancer cells (Fig.\u00a0[6D](#Fig6){ref-type=\"fig\"}). These results explain that CoQ~0~ may potentiate the apoptosis and/or autophagy mechanisms *via* suppression of HER-2/AKT/mTOR signalling cascades in SKOV-3 cells.\n\nInhibition of ROS production obliterates CoQ~0~-induced apoptosis, but not autophagy in SKOV-3 cells {#Sec13}\n----------------------------------------------------------------------------------------------------\n\nAberrant production of ROS involved in execution of apoptosis and/or autophagy^[@CR38],\\ [@CR39]^. Since CoQ~0~ reported to trigger the ROS production and induce apoptosis in SKOV-3 cells, we hypothesized that CoQ~0~-induced ROS could propagate the apoptosis. To address this phenomenon, cells were pretreated with ROS inhibitor (NAC, 2\u2009mM) for 1\u2009h, and then incubated with CoQ~0~ (30\u2009\u00b5M, 24\u2009h). We found that tremendously increased apoptotic DNA fragmentation with CoQ~0~ was substantially diminished in the presence of NAC. The TUNEL positive cells with NAC are almost similar to that of control (Fig.\u00a0[7A and B](#Fig7){ref-type=\"fig\"}). CoQ~0~ provoked late (31%) and early (34.7%) apoptotic-cells visualized in Q2 and Q4 were remarkably attenuated by NAC pretreatment. This was indicated by reporting only 2% (Q2) and 10.3% (Q4) of late and early apoptotic-cells following blockade of ROS production (Fig.\u00a0[7C](#Fig7){ref-type=\"fig\"}). In addition, increased proteolytic cleavage of PARP by CoQ~0~ was obliterated in NAC pretreated cells (Fig.\u00a0[7D and E](#Fig7){ref-type=\"fig\"}). These results clearly demonstrating that CoQ~0~-induced ROS are involved in execution of SKOV-3 cell apoptosis.Figure 7CoQ~0~-induced ROS involved in induction of SKOV-3 apoptosis, but not autophagy. **(A--G)** Cells were pretreated with NAC (2\u2009mM) for 1\u2009h, and then incubated with CoQ~0~ (30\u2009\u00b5M) for 24\u2009h. **(A)** Induction of apoptosis (DNA fragmentation) was determined by TUNEL assay in the presence or absence of NAC. The green florescence indicates TUNEL-positive cells in microscopic fields (200 \u00d7 magnification). **(B)** The fold of apoptotic cells was calculated by measuring the florescence intensity using commercially available software. **(C)** Annexin-V-FITC and PI staining was used to identify the early/late apoptosis or necrosis of SKOV-3 cells. **(D)** Cleavage of PARP was monitored by Western blot. **(E)** Changes in protein intensities were quantified using commercially available software. **(F)** Formation of AVOs was visualized under a red filter fluorescence microscope (100\u2009\u00d7\u2009magnification) using AO stain. **(G)** Number of AO stained cells with CoQ~0~ or NAC was presented as histogram, control being as 1.0 fold. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control, and significant at ^\\#\\#^ *p*\u2009\\<\u20090.01; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nSubsequently we examined the role of CoQ~0~-induced ROS on autophagy by detecting the AVOs in the presence or absence of NAC, following CoQ~0~ treatment. We found fascinating results that CoQ~0~-induced increased numbers of AVOs were not changed (remained same) even in the presence of NAC. This novel evidence explains that CoQ~0~-induced ROS are not involved in autophagy, or autophagy is not contributing to the death of cancer cells; despite apoptosis takes place (Fig.\u00a0[7F and G](#Fig7){ref-type=\"fig\"}).\n\nInhibition of apoptosis suppresses CoQ~0~-induced autophagy in SKOV-3 cells {#Sec14}\n---------------------------------------------------------------------------\n\nNext to address the role of apoptosis on autophagy under CoQ~0~ stimulation, SKOV-3 cells were treated with apoptosis inhibitor (Z-VAD-FMK, 20 \u03bcM) 1\u2009h prior to CoQ~0~ incubation (30\u2009\u00b5M, 24\u2009h), and changes in LC3-II accumulation, conversion of GFP-LC3 and AVOs formation were monitored. We found that increased LC3-II accumulation in CoQ~0~ treated cells was noticeably limited by apoptosis inhibition (Fig.\u00a0[8A](#Fig8){ref-type=\"fig\"}). CoQ~0~-induced increased AVOs were substantially decreased in Z-VAD-FMK pretreated cells (Fig.\u00a0[8B and C](#Fig8){ref-type=\"fig\"}). Subsequently, Z-VAD-FMK pretreatment to the GFP-LC3 plasmid transfected cells, significantly diminished the CoQ~0~-induced conversion of GFP-LC3 and endogenous LC3 levels. This was evidenced by a dearth of green LC3 punctate dots and diffused fluorescence intensity in Z-VAD-FMK pretreated cells (Fig.\u00a0[8D](#Fig8){ref-type=\"fig\"}). CoQ~0~ promoted conversion of LC3--1 to LC3-II was also impaired by apoptosis inhibitor (Fig.\u00a0[8D](#Fig8){ref-type=\"fig\"}). Our experimental evidence reveals that CoQ~0~-induced apoptosis led to evoke the autophagy in SKOV-3 cells.Figure 8Inhibition of apoptosis suppressed CoQ~0~-induced autophagy in SKOV-3 cells. (**A--C**) Cells were pretreated with caspase inhibitor (Z-VAD-FMK, 20 \u03bcM) for 1\u2009h, and then incubated with CoQ~0~ (30\u2009\u00b5M) for 24\u2009h. **(A)** LC3-I/II was determined using Western blot, and relative changes in protein intensities were quantified using commercially available software. **(B)** Formation of AVOs was visualized under a red filter fluorescence microscope (100\u2009\u00d7\u2009magnification) using AO stain. **(C)** Number of AO stained cells with CoQ~0~ or Z-VAD-FMK treatment was presented as histogram, control being as 1.0 fold. **(D)** Cells were transfected with GFP-LC3 expression vector for 24\u2009h before Z-VAD-FMK and CoQ~0~ treatment. GFP-LC3 dots in cells were observed under a confocal microscope (200 \u00d7 magnification). Conversions of GFP-LC3 and endogenous LC3 were determined by Western blot. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control, and significant at ^\\#^ *p*\u2009\\<\u20090.05; ^\\#\\#^ *p*\u2009\\<\u20090.01; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nInhibition of autophagy promotes apoptosis in CoQ~0~-treated SKOV-3 cells {#Sec15}\n-------------------------------------------------------------------------\n\nTo describe whether autophagy is able to influence CoQ~0~-induced apoptosis, SKOV-3 cells were pretreated with early autophagy inhibitor (3-MA, 2\u2009mM), and occurrence of apoptotic DNA fragmentation and cleavage of PARP was measured. CoQ~0~ treatment (30\u2009\u00b5M, 24\u2009h) increased apoptotic DNA fragmentation (\\~6-fold) as evidenced by increased TUNEL-positive cells. Intriguingly, 3-MA pretreatment further intensified the apoptotic DNA fragmentation in SKOV-3 cells (Fig.\u00a0[9A and B](#Fig9){ref-type=\"fig\"}). 3-MA pretreatment enhanced the proteolytic cleavage of PARP that was prominent when compared with CoQ~0~ alone treatment (Fig.\u00a0[9C](#Fig9){ref-type=\"fig\"}). Evidence from our study (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"}, Fig.\u00a0[9A--C](#Fig9){ref-type=\"fig\"}) emphasizes that induction of cytoprotective autophagy may renders to delay or suppress the onset of CoQ~0~-induced apoptosis or SKOV-3 cell death.Figure 9Inhibition of autophagy potentiated CoQ~0~-induced apoptosis in SKOV-3 cells. (**A--D**) Cells were pretreated with autophagy inhibitor (3-MA, 2\u2009mM) for 1\u2009h, and then incubated with CoQ~0~ (30\u2009\u00b5M) for 24\u2009h. **(A)** Induction of apoptosis (DNA fragmentation) was determined by TUNEL assay in the presence or absence of 3-MA. The green florescence indicates TUNEL-positive cells in microscopic fields (200 \u00d7 magnification). **(B)** The fold of apoptotic cells was calculated by measuring the florescence intensity using commercially available software. **(C)** Cleavage of PARP with or without 3-MA pretreatment was determined by Western blot. **(C** Changes in protein intensities were quantified using commercially available software. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\*\\*\\**p*\u2009\\<\u20090.001 compared with untreated control, and significant at ^\\#^ *p*\u2009\\<\u20090.05; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared with CoQ~0~ alone treated cells.\n\nCoQ~0~ suppresses tumor growth in SKOV-3 xenografted nude mice {#Sec16}\n--------------------------------------------------------------\n\nTo examine the *in vivo* anti-tumor activity of CoQ~0~, we used nude mice and ovarian cancer SKOV-3 cells were subcutaneously xenografted into the mice, and then treated with CoQ~0~ (1.5 and 2.5\u2009mg/kg) or vehicle for 52 days. During the experimental period no significant body weight loss was noticed among the groups, and all mice appeared healthy (Fig.\u00a0[10A](#Fig10){ref-type=\"fig\"}). Besides, no signs of the adverse effects or toxicity of CoQ~0~ was observed in nude mice, according to microscopic examination of individual organs (data not shown). The tumor volume was measured for every 4 days until 52 days, and found that both doses of CoQ~0~ (1.5 and 2.5\u2009mg/kg) inhibited the tumor volume, higher dose being more effective in a time-course (Fig.\u00a0[10B](#Fig10){ref-type=\"fig\"}). Next, to assess CoQ~0~ effect on tumor weight, all mice including control were photographed, and excised xenografted tumor tissues were weighted. Consistent to the tumor volume, tumor weight was also considerably decreased with CoQ~0~ compared with control mice (Fig.\u00a0[10C](#Fig10){ref-type=\"fig\"}). These results revealed that CoQ~0~ suppressed the tumor growth in SKOV-3 xenografted nude mice.Figure 10CoQ~0~ inhibits tumor growth in SKOV-3 xenografted nude mice. (A--C) Time-course effect of CoQ~0~ on growth of SKOV-3 xenografted nude mice was evaluated. Ovarian cancer cells (3\u2009\u00d7\u200910^6^ cells) were subcutaneously implanted into the flanks of nude mice on day 0 and inoculated for 7 days. Mice were subsequently treated with CoQ~0~ (1.5 and 2.5\u2009mg/kg) or vehicle (control) on every 4 days for 52 days. Changes in bodyweight **(A)** and tumor volume **(B)** were recorded for every 4 days, until 52 days. **(C)** On day 52, animals were photographed, and excised tumor tissue was weighed. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\**p*\u2009\\<\u20090.05; compared with vehicle treated mice.\n\nCoQ~0~ inhibits mitosis and promotes apoptosis in tumors of SKOV-3 xenografted mice {#Sec17}\n-----------------------------------------------------------------------------------\n\nTo delineate the reason why CoQ~0~ suppressed the tumor size and weight, we performed histopathological and TUNEL assays to examine the mitosis and apoptosis in tumor tissues. Microscopic images illustrated that the control tumor cells were large in size, round to oval in shape with predominant nucleoli. Tumor cells expressed high levels of cellular activity and mitotic figures, indicating proliferating cells. However, tumor sections of CoQ~0~-treated mice (1.5 and 2.5\u2009mg/kg) showed smaller cells with shrunken, condensed and pyknotic nuclei, which indicates inactivity or regression of tumor cells (Fig.\u00a0[11A](#Fig11){ref-type=\"fig\"}). The quantified mitotic-positive cells that were abundant in control tumors were notably inhibited following CoQ~0~ treatment (Fig.\u00a0[11B](#Fig11){ref-type=\"fig\"}). Conversely, CoQ~0~ significantly promoted the apoptosis in tumors of xenografted mice (Fig.\u00a0[11A and B](#Fig11){ref-type=\"fig\"}), which was confirmed by a dose-dependent increase of apoptotic DNA fragmentation. Results from TUNEL assay showed CoQ~0~ treatment enhanced the TUNEL-positive cells compared with vehicle (Fig.\u00a0[11C and D](#Fig11){ref-type=\"fig\"}). Our evidence confirmed that the anti-tumor activity of CoQ~0~ in xenografted tumors is associated with decreased proliferation and increased apoptosis.Figure 11Histopathological analyses of mitosis and apoptosis in the tumors of SKOV-3 xenografted nude mice followed by CoQ~0~ treatment. (**A,B**) Histopathological staining was performed to detect the mitotic- and apoptotic-cells in SKOV-3 xenografted tumors following CoQ~0~ (1.5 and 2.5\u2009mg/kg) treatment using light microscopy (20\u2009\u00d7\u2009and 200\u2009\u00d7\u2009magnification). Arrows indicate mitotic (tumor control) and pyknotic tumor cells (CoQ~0~). The number of mitotic-positive and apoptotic-positive cells in microscopic fields of 3 samples was averaged, and presented as histogram. (**C,D**) *In situ* detection of apoptosis was performed using TUNEL staining in tumor sections from control and experimental analogues treated with CoQ~0~ (1.5 and 2.5\u2009mg/kg). The number of TUNEL-positive cells were quantified from 3 microscopic fields, and averaged. Results expressed as mean\u2009\u00b1\u2009SD of three independent assays (n\u2009=\u20093). Significant at \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 compared with vehicle treated mice.\n\nDiscussion {#Sec18}\n==========\n\nOwing to its potent anticancer/cytotoxic properties against human breast cancer^[@CR25]^, and lung cancer cells^[@CR40]^, we assume that CoQ~0~ could be a potential candidate to treat human ovarian carcinoma cells. For the first time, we have demonstrated that CoQ~0~, isolated from *Antrodia camphorata*, potently killed the ovarian SKOV-3 cancer cells and induced cell-cycle arrest *via* decreased cell-cycle regulatory proteins. CoQ~0~ triggered the intracellular ROS production, which then contributed to activate and propagate apoptotic signals. CoQ~0~-induced apoptosis was evidenced by increased apoptotic DNA fragmentation, higher number of early/late apoptotic cells and dysregulated Bax/Bcl-2 ratio. Both mitochondrial and ER-stress mediated apoptotic signals are implicated in execution of apoptosis. The blockade of ROS production by NAC treatment diminished pro-apoptotic signals and cell death. On the other hand, activation of autophagy by CoQ~0~ was evidenced by increased LC3-II accumulation, GFP-LC-3 punctuated patters and AVOs appearance, however not contributed for autophagic-cell death. Inhibition of apoptosis by Z-VAD-FMK reversed CoQ~0~-induced autophagy, while inhibition of autophagy by 3-MA/CQ exaggerated CoQ~0~-induced apoptotic signals. These findings explain that CoQ~0~-induced apoptosis evokes cytoprotective autophagy, which may counteract the CoQ~0~-induced pro-apoptotic signals and hinder the CoQ~0~ anticancer properties. Furthermore, CoQ~0~ treatment to SKOV-3 xenografted nude mice suppressed the tumor volume. Xenografted tumors after CoQ~0~ treatment represented by significantly decreased mitotic-cells and tremendously increased apoptotic-cells and DNA fragmentation. Taken together, our findings revealed that CoQ~0~ displays both *in vitro* and *in vivo* antitumor properties. CoQ~0~-induced ROS-mediates the apoptosis and activates cytoprotective/survival autophagy in human ovarian cancer cells.\n\nInduction of apoptosis/autophagy, inhibition of cell survival and/or cell-cycle arrest by chemical substances or biological agents are the principal strategies in treatment of a variety of cancers, including ovarian carcinoma^[@CR19],\\ [@CR27],\\ [@CR41]^. In this study, we demonstrated that CoQ~0~ treatment to ovarian cancer SKOV-3 cells produced potent cytotoxic effects, as indicated by a profound increase of cell death and aberrant morphological changes. Previous study reported that among several methoxy-substituted cyclic compounds, CoQ~0~ is the strongest cytotoxic analog that inhibited the growth of MDA-MB-231 and SKBr3 human breast cancer cells. The cytotoxicity of CoQ~0~ was associated with increased proportion of cells undergoing apoptosis, and this phenomenon was addressed by the induction of G~0~/G~1~ cell-cycle arrest in MDA-MB-231 cells and S-phase arrest in SKBr3 cells^[@CR25]^. CoQ~0~ isolated from *Antrodia cinnamomea* submerged cultures has also reported to possess anticancer property by the decreased viability of A549, HepG2 and SW480 cancer cells, and increased ROS-mediated apoptosis^[@CR40]^.\n\nPreviously, we have shown that treatment of SKOV-3 cells with fermented culture broth of *Antrodia camphorata* inhibited the proliferation, and caused G~2~/M cell-cycle arrest through the downregulation of cyclin D1, cyclin A and CDK1, and upregulation of p27 expressions^[@CR27]^. The CoQ~0~ used in this study is a novel quinone derivative of *Antrodia camphorata*, which contain zero isoprenoid side chains. CoQ~0~-induced death of SKOV-3 cells in this study was also accompanied by the G~2~/M-phase cell-cycle arrest and decreased cell-cycle regulatory proteins (cyclin B1, CDK1, cyclin A and CDK2). Since overexpression of CDKs activity is often the cause of human cancer, their function is tightly regulated by cell-cycle inhibitors, such as the p21 and p27 Cip/Kip proteins. It has been stated that following anti-mitogenic signals or DNA damage, the proteins, p21 and p27 can bind to cyclin-CDK complexes to inhibit their catalytic activity and induce cell-cycle arrest^[@CR42]^. In our study, the anti-mitogenic signals and apoptotic DNA damaging actions of CoQ~0~ perhaps activates p21 and p27 expressions, which then bind to cyclin-CDK complex to inhibit their expression in SKOV2 cells. Thus, suppression of cell-cycle promoting proteins and induction of cell-cycle arrest by CoQ~0~ may contribute to inhibit the ovarian cancer cell survival.\n\nAberrant production of ROS by oxidants or chemical substances play an important role in execution of apoptosis by activating the pro-apoptotic signals^[@CR33],\\ [@CR38]^. Here we found that profoundly increased ROS production with CoQ~0~ was accompanied by an increased death of SKOV-3 cells, and blockade of ROS production by NAC treatment substantially diminished the CoQ~0~-induced cell death. Coenzyme Q exists in biomembranes act as an electron carrier and reversibly changes to either oxidized (CoQ), intermediate (CoQ^\u2022\u2212^) or reduced (CoQH~2~) form^[@CR24]^. During conversion from one form to another, short chain CoQ^\u2022\u2212^ comes into contact with cytoplasm, auto-oxidation with O~2~ occurs and generates ROS^[@CR43]^. Doughan *et al*., demonstrated that mitoquinone (MitoQ) may be the pro-oxidant and pro-apoptotic, because its quinone group can participate in redox cycling and superoxide radical production. MitoQ-induced ROS production appears to stimulate the activity of caspase-3 and endothelial cell apoptosis^[@CR8]^. Ubiquinone 0 (Ub~0~) also known as CoQ~0~ reported to dramatic increase of ROS production in culture rat liver Clone-9 cells, which is prominent than the other analogs (Ub~5~ and Ub~10~). Particularly, Ub~0~ induced PTP opening in cancerous rat liver MH1C1 cells, and promoted PTP-dependent cell death^[@CR23]^. Indeed, ROS accumulation can cause depolarization of mitochondrial membrane and results in increased release of mitochondrial proteins, including apoptosis-inducing factor (AIF) that certainly promote apoptosis^[@CR44],\\ [@CR45]^. A study on human lung carcinoma cells showed that sodium selenite induced ROS-mediated intrinsic apoptosis and mitochondrial dysfunction was blocked by ROS inhibition^[@CR33]^. Another recent study reported that CoQ~0~ provoked the ROS-mediated apoptosis of lung cancer cells, and increased apoptotic cells were noticeably inhibited with antioxidant pretreatment^[@CR40]^. These results elucidate that the intracellular ROS are critically involved in the propagation of apoptosis in SKOV-3 cells.\n\nIn line with this, we demonstrated that CoQ~0~-induced apoptotic DNA fragmentation, PARP cleavage and early/late apoptosis were substantially reversed *via* ROS inhibition in SKOV-3 cells. CoQ~0~-iduced ROS-mediated apoptosis appears to be executed by mitochondrial- and ER stress-dependent signalling cascades. It has been described that the cleavage of procaspase-3 into active caspase-3 proceeds to fragmentation of PARP, and culminating the death of cell^[@CR10],\\ [@CR12]^. PARP, an important nuclear enzyme involved in DNA repair is a substrate for activated caspase-3, and fragmentation of PARP is considered a hallmark of mitochondrial apoptosis^[@CR13]^. Cleavage of procaspase-3 is an early event in apoptosis induced by chemotherapeutic agents, and activated caspase-3 also promotes DNA fragmentation^[@CR46]^. Here, we found the marked increase of caspase-3 and PARP fragmentation in CoQ~0~-treated cells. Interestingly, pharmacological inhibition of caspase activation by Z-VAD-FMK or blockade of ROS production by NAC resulted in a diminution of CoQ~0~-iduced PARP fragmentation followed by a suppressed cell death. These findings support the occurrence of mitochondrial apoptosis in CoQ~0~-treated ovarian cancer cells. In fact, mitochondrial apoptosis is known to be stimulated by a wide range of cellular stresses, including overwhelming production of ROS, DNA damage and cytoskeletal disruption^[@CR12],\\ [@CR38]^. CoQ~0~-iduced mitochondrial apoptosis was further addressed by the increased production of intracellular ROS levels and apoptotic DNA fragmentation in SKOV-3 cells.\n\nER stress, represented by an increased accumulation of unfolded proteins due to multiple stimuli, including oxidants, triggers a conserved series of signals to ameliorate the accumulation of unfolded proteins. However, if these signalling events are severe or protracted, they can induce cell death^[@CR11]^. ER molecular chaperone, HSP70 and caspase-12 are said to be involved in the ER stress-specific apoptosis^[@CR11],\\ [@CR47]^. We found increased expressions of caspase-12 and HSP-70 proteins in cells incubated with CoQ~0~, which implies the activation of ER stress-mediated apoptosis. Caspase-12 is predominantly localized to the ER, and specifically activated by the apoptotic signals with an ER stress component^[@CR47]^. The bile salt sodium deoxycholate, an apoptosis inducer was found to activate the HSP-70 in liver cells along with other genes involved in ER stress and DNA damage. This study suggested that activated chaperone protein HSP-70 may be an additional stress response due to malfolding of proteins in ER^[@CR48]^.\n\nOur findings further demonstrated that CoQ~0~ induced early/late apoptotic death of SKOV-3 cells. Along with aberrant morphological changes, results from Annexin-V/PI assay showed the greater number of early (Annexin-V^+ve^/PI^\u2212ve^) and late (Annexin-V^+ve^/PI^+ve^) apoptotic cells following CoQ~0~ treatment. However, this phenomenon was not observed in NAC pretreated cells, which indicates the involvement of ROS in apoptosis. Treatment of HL60 cells with different CoQ structures (CoQ~1~, CoQ~2~ and CoQ~4~) reported to inhibit the proliferation, initial S-phase arrest followed by G~0~/G~1~-phase cell-cycle arrest at later time points, and induction of apoptosis^[@CR49]^. A recent study by Chung and colleagues showed that CoQ~0~ enhanced both early and late apoptosis in human lung cancer cells, and antioxidant pretreatment diminished the apoptotic-cell death^[@CR40]^. Several studies demonstrated that quinones or certain methoxy-containing analogs of CoQ with structural similarities to CoQ~0~ are able to produce cytotoxic effects on human cancer cells by the induction of apoptosis^[@CR24],\\ [@CR49]^. A study carried-out using different CoQ analogs reported the increased apoptosis, DNA fragmentation and caspase-3 activation with CoQ~2~ and CoQ~4~, but not with CoQ~6~ or CoQ~10~ in HL60 human leukemia cells^[@CR49]^. These findings imply that the anticancer or pro-apoptotic properties of CoQ analogs may be varied based on the length of isoprenyl side chain, and position of the methoxy-substitutions on the quinone nucleus. Irrespective of cell lines, CoQ~0~ that contains zero isoprenoid units inhibit ovarian cancer cell growth and induce early/late apoptosis through the ROS signals.\n\nOn the other hand, emerging evidence shows that autophagy constitutes a potential target for cancer therapy, and induction of autophagy in response to therapeutics can be viewed as a cell-death or cell-survival phenomenon that contributes to the anticancer efficacy^[@CR19]^. For the first time, we demonstrated the novel function of CoQ~0~, that is induction of autophagy in the ovarian cancer cells. This phenomenon was evidenced by the increased conversion of LC3-I to LC3-II (membrane bound), multiplied AVOs formation and elevated Beclin-1 expression following CoQ~0~ treatment. Induction of autophagy by CoQ~0~ was further confirmed by the confocal images illustrating the abundance of green GFP-LC3 puncta. The increased appearance of AVOs, a hallmark of autophagy is associated with the increased accumulation of lipidated LC3 in cells^[@CR15]^. CoQ~0~-induced higher numbers of AVOs and corresponding increased distribution of GFP-LC3-II punctuates in SKOV-3 cells are explaining the recruitment of LC3-II to autophagosomes. Antroquinonol, a ubiquinone derivative of *Antrodia camphorata* has been reported to increase the LC3-II accumulation and a certain degree of autophagic-death of pancreatic carcinoma cells^[@CR30]^. Since mTOR signalling is considered a key negative regulator of autophagy^[@CR50]^, we determined the CoQ~0~ effect on phosphorylated mTOR levels, and found severely inhibited p-mTOR (S2448) levels in SKOV-3 cells. This was accompanied by a simultaneous decrease of p-AKT levels and increase of Beclin-1/Bcl-2 ratio, which denotes activation of autophagy in SKOV-3 cells. It has been claimed that excessive production of ROS can act as intracellular messengers to trigger autophagy^[@CR39]^. To address whether CoQ~0~-induced ROS involved in activation of autophagy in SKOV-3 cells, we detected the AVOs formation in the presence and absence of NAC. It is interesting to note that inhibition of ROS production by NAC is unable to suppress the CoQ~0~-induced AVOs formation, which indicates CoQ~0~ activated autophagy is ROS-independent. Taken together, our results clearly explained that CoQ~0~-induced ROS are involved in activation of the pro-apoptotic signals, but not autophagy in ovarian cancer cells.\n\nAnother most important finding of this study is that CoQ~0~ activated autophagy in SKOV-3 cells didn't contribute for cell death; instead it might be a cytoprotective mechanism. There are confirmatory evidence explaining that induction of apoptotic cancer cell death by chemical substances is accompanied by a protective autophagy^[@CR18],\\ [@CR21],\\ [@CR33]^. Activation of autophagic pathway in response to target therapies or metabolic stress reportedly contributes to the survival of formed tumors, which might mediate resistance to the anticancer therapies^[@CR17],\\ [@CR51]^. Of note, without addressing the role of autophagy, few studies enlightened the cytotoxicity of CoQ~0~ against human lung cancer cells^[@CR40]^ and breast cancer cells^[@CR25]^ *via* induction of apoptotic-cell death. Yet it is unclear, whether CoQ~0~-induced autophagy contributes to survival or death of ovarian cancer cells. First time, we demonstrated that inhibition of CoQ~0~-induced autophagy by 3-MA or CQ didn't suppress the CoQ~0~-induced cell death. These evidence explain that CoQ~0~-induced autophagy is not involved in death of ovarian cancer cells, perhaps play a self-protective role against the cytotoxic effects of CoQ~0~. However, it can't be ruled out that activation of autophagy by anticancer drugs/oxidants act as an alternative pathway for cellular death^[@CR17]^. Although autophagy role is controversial in cancer cells, our findings suggest that autophagy may serves as an additional target for adjuvant anticancer therapy, and inhibition of such autophagy could further strengthen the therapeutic effectiveness of CoQ~0~ treatment.\n\nThe interaction between autophagy protein (Beclin-1) and anti/pro-apoptotic proteins (Bcl-2/Bcl-XL) is complex, and represents a potential important point of convergence of the apoptotic and autophagic machinery^[@CR20],\\ [@CR52]^. It has been shown that Beclin-1 cannot neutralize the Bcl-2 function, but Bcl-2 family proteins suppress the pro-autophagic function of Beclin-1^[@CR14]^. In our study, CoQ~0~ downregulated the Bcl-2 and upregulated the Beclin-1 that indicates activation of respective apoptosis and early autophagy mechanisms in SKOV-3 cells. Similar to our studies, anticancer properties of Sulforaphane were represented by the decreased Bcl-2 levels and increased caspase-3 activation in human breast cancer cells^[@CR15]^. Activated caspase-3 can cleave Beclin-1 protein at position 125 and 149. The truncated Beclin-1, which in turn disrupt its interaction with Bcl-2 then allows the release of pro-apoptotic molecules from the Bcl-2/Bcl-xL complex to initiate intrinsic apoptosis^[@CR53]^. In addition, caspase-mediated cleavage of Beclin-1 promote the crosstalk between apoptosis and autophagy^[@CR14]^. CoQ~0~-induced elevated Beclin-1/Bcl-2 ratio and Bax/Bcl-2 ratio in ovarian cancer cells suggest that CoQ~0~ promotes cell death *via* pro-apoptotic signals.\n\nHER-2/*neu* is a proto-oncoprotein belongs to the family of epidermal growth factor, overexpressed in various human cancers, including ovarian carcinoma. Overexpression of HER-2/*neu* is associated with the high risk of treatment due to highly metastatic ability of cells and resistant to drug treatments^[@CR54]^. We demonstrated that CoQ~0~ treatment to ovarian cancer cells time-dependently inhibited the HER-2/*neu* expression. The decreased HER-2/*neu* phosphorylation was accompanied by a substantial reduction of AKT and mTOR phosphorylations. It has been shown that activation of HER-2 leads to autophosphorylation of the C-terminal tyrosines of the receptor, and implicated in regulation of cell proliferation and inhibition of apoptosis^[@CR55]^. Therefore, inhibition of p-HER-2 by CoQ~0~ indicates decreased proliferation and increased apoptosis of SKOV-3 cells. In our previous study, we have shown that *Antrodia camphorata* that contains CoQ~0~ significantly decreased the basal tyrosine kinase phosphorylation and activation of HER-2/*neu* receptors in HER-2/*neu*-overexpressed SKOV-3 cells^[@CR27]^. A recent study showed that downregulation of HER-2 by Alo-emodin in breast cancer cells resulted in decreased tumor initiation, cell migration and invasion^[@CR9]^. Since overexpression of HER-2 has been linked with activation of AKT/mTOR signals that are involved in tumor biology^[@CR27],\\ [@CR37]^, the decreased AKT/mTOR phosphorylation with CoQ~0~ perhaps contributes to effective management of the ovarian cancer progression.\n\nApoptosis and autophagy shares the functions of several common regulators, and the crosstalk between apoptosis and autophagy decides the fate of cell in response to cellular stress^[@CR34]^. Regardless of their complex interrelationship, autophagy usually associated with cell survival, whereas apoptosis is invariably contributes to cell death^[@CR15],\\ [@CR56]^. Accumulating evidence suggests that under certain stress circumstances, autophagy is known to act as a partner or a promoter of apoptosis. Conversely, activation of autophagy may function to prevent the onset of apoptosis, and apoptosis can also activate autophagy^[@CR21],\\ [@CR34],\\ [@CR57]^. It has been shown that oridonin treated L929 cells exhibited both autophagy and apoptosis, where inhibition of autophagy increased the apoptotic-cell death, which means autophagy has an anti-apoptotic function^[@CR56]^. Owing to the complex interactions between apoptosis and autophagy, it is necessary for careful monitoring of the key regulatory molecules to understand the cell death phenomenon induced by chemical substances. We demonstrated that CoQ~0~-induced apoptosis contributes to death of ovarian cancer cells, rather than autophagy. Chemical inhibition of apoptosis diminished the CoQ~0~-induced cell death by inhibition of PARP cleavage and DNA fragmentation.\n\nSince occurrence of apoptosis always means to cell death^[@CR12]^, the reported autophagy in CoQ~0~-treated cancer cells does not necessarily indicates autophagic-cell death. It is worth to note that inhibition of autophagy is unable to diminish the CoQ~0~-induced cell death; despite intensified the apoptotic signals. On the other hand, inhibition of apoptosis is accompanied by a substantial diminution of CoQ~0~-induced autophagy (decreased LC3-II accumulation and AVOs formation). Cellular stresses, such as ER-stress and mitochondrial dysfunction are reported to activate the autophagy^[@CR57]^. Likewise, ginsenosides F2 induces mitochondrial dysfunction and apoptotic death of breast cancer cells was accompanied by autophagy, where inhibition of autophagy enhanced the F2-induced cell death that reveals the protective role of autophagy^[@CR21]^. Thus, CoQ~0~-induced ER-stress and increased release of mitochondrial apoptotic proteins may be resulted in activation of the autophagy mechanism to remove the damaged organelles. These novel findings addressed that CoQ~0~-induced autophagy may not participate in cell death, perhaps served as a survival mechanism against cytotoxicity and cellular stress.\n\nTo further strengthen the anticancer properties of CoQ~0~, we performed the *in vivo* studies on SKOV-3 xenografted nude mice by treating with CoQ~0~. We found that treatment of CoQ~0~ to xenografted nude mice significantly decreased the tumor volume. This antitumor activity appears to be associated with the inhibition of mitotic cells and a dramatic increase of apoptotic cells/DNA fragmentation in CoQ~0~-treated tumors. A study by Zhang and colleagues demonstrated that treatment with EB1089, an analog of vitamin D completely suppressed the growth of tumors in ovarian cancer xenografted nude mice. In addition, fewer mitotic figures and increased TUNEL-positive apoptotic cells were also observed in the tumor sections of EB1089 treated nude mice^[@CR58]^. The decreased mitotic-positive cells in the tumors of *Antrodia salmonea* treated HL60 xenografted nude mice indicating the decreased cell proliferation, which may result in decreased tumor volume in nude mice^[@CR59]^. A recent study suggests that suppression of epithelial ovarian cancer cell invasion into the omentum by EB1089 may help to improve the survival of patients with advanced ovarian cancer^[@CR3]^. CoQ~0~ treatment to xenografted nude mice contributed to decrease the cancer cell proliferation and increased apoptosis. These *in vivo* findings confirmed the potent antitumor properties of CoQ~0~ against ovarian cancer that are consistent with *in vitro* anticancer properties. Take into consideration of *in vitro* evidence, CoQ~0~-induced ROS possibly involved in propagation of apoptosis in xenografted tumors.\n\nConclusions {#Sec19}\n===========\n\nFor the first time, our findings demonstrated that CoQ~0~ induced apoptotic-death of human ovarian cancer cells through the increased production of ROS. The blockade of ROS production by NAC suppressed CoQ~0~-induced apoptotic signals and diminished cell death, but not autophagy. Chemical inhibition of apoptosis substantially retarded the CoQ~0~-induced autophagy and cell death. However, inhibition of autophagy was unable to suppress the CoQ~0~-induced apoptosis, instead accelerated the pro-apoptotic signals. Our experimental evidence provides insights into the complex relation between apoptosis and autophagy induced by CoQ~0~ in SKOV-3 cells. We further demonstrated the *in vivo* antitumor properties of CoQ~0~ in SKOV-3 xenografted nude mice. We found that CoQ~0~ treatment substantially inhibited the tumor volume through decreased mitotic-cells and increased apoptotic-cells in tumors. These findings emphasize that CoQ~0~ revealed potent anticancer properties through ROS-mediated apoptosis against human ovarian carcinoma cells. The inhibition of cytoprotective autophagy or further amplifying of intracellular ROS production could be the potential strategies to improve the effeteness of cancer treatment.\n\nMaterials and Methods {#Sec20}\n=====================\n\nChemicals and reagents {#Sec21}\n----------------------\n\nDulbecco's Modified Eagle's medium (DMEM), nutrient mixture F-12, fetal bovine serum (FBS), glutamine and penicillin/streptomycin were obtained from GIBCO BRL (Grand Island, NY, USA). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), *N*-acetylcysteine (NAC), *p*-iodonitrotetrazolium violet, fluorescein isothiocyanate (FITC), propidium iodide (PI), acridine orange (AO), 3-Methyladenine (3-MA), chloroquine (CQ) and 2\u2032,7\u2032-dihydrofluorescein-dictate (DCFH~2~-DA) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Antibodies against cyclin B1, CDK1, cyclin A, CDK2, caspase-3, Bcl-2, Bax, and \u03b2-actin were purchased from Santa Cruz Biotechnology, Inc. (Heidelberg, Germany). Antibodies against LC3-I/II, PARP, p-HER-2/*neu*, p-AKT, p-mTOR, and Bcelin-1 were obtained from Cell Signalling Technology, Inc. (Danvers, MA, USA). Antibody against GFP was purchased from Gene Tex, Inc. (Irvine, CA, USA). Antibodies against HS-70 and Caspase-12 were purchased from BD Transduction Laboratories (Hayward, CA, USA) and Millipore Corporation (Billerica, MA, USA) respectively. 4\u2032,6-Diamidino-2-phenylindole dihydrochloride (DAPI) was purchased from Calbiochem (La Jolla, CA, USA). Z-Val-Ala-Asp-fluoromethylketone (Z-VAD-FMK) was obtained from Calbiochem (San Diego, CA, USA). All other chemicals were reagent grade or HPLC grade and were supplied by either Merck & Co., Inc. (Darmstadt, Germany) or by Sigma-Aldrich.\n\nPreparation of fermented culture broth of *Antrodia camphorata* from submerged cultures {#Sec22}\n---------------------------------------------------------------------------------------\n\nThe *Antrodia camphorata* culture was inoculated onto potato dextrose agar and incubated at 30\u2009\u00b0C for 15--20 days. The whole colony was subsequently added to a flask containing 50\u2009mL sterile water. After homogenization, the fragmented mycelial suspension was used as an inoculum. The seed culture was prepared in a 20-L fermenter (BioTop Process & Eqipment, Taiwan) agitated at 150\u2009rpm with an aeration rate of 0.2 vvm at 30\u2009\u00b0C. A five-day culture of 15\u2009L mycelium inoculum was inoculated into a 250\u2009L agitated fermenter (BioTop). The fermentation conditions were the same as those used for the seed fermentation, but the aeration rate was 0.075 vvm. The fermentation product was harvested at hour 331 and poured through a non-woven fabric on a 20-mesh sieve to separate the deep-red fermented culture broth and the mycelia; the culture broth was then centrifuged at 3000 \u00d7 *g* for 10\u2009min followed by passage through a 0.22-\u03bcm filter. The culture broth was concentrated under vacuum and freeze-dried to a powder. The yield of dry matter from the culture broth was 18.4\u2009g/L. The experiments were performed with 2--4 different batches of *Antrodia camphorata* fermented culture^[@CR60]^.\n\nIsolation and characterization of CoQ~0~ from *Antrodia camphorata* {#Sec23}\n-------------------------------------------------------------------\n\nThe HPLC profile of the fermented culture broth of *Antrodia camphorata* was performed using a RP-18 column \\[COSMOSIL, 5C~18~-AR-II, Waters, 4.6\u2009\u00d7\u2009250 mm\\] at a flow rate of 1.0\u2009mL/min, detected at UV 254 and 220\u2009nm. Standard solution of the fermented culture broth from *Antrodia camphorata* was prepared by dissolving it in water (5.0\u2009mg/mL), filtered through 0.22 \u03bcm membrane filter and applied to HPLC analysis. The mobile phase consisted of (A) acetonitrile and (B) 0.05% trifluoroacetic acid (TFA) (v/v) using a gradient elution of 5--60% A at 5--40\u2009min. The flow rate was maintained as 1.0\u2009mL/min and aliquots of 20\u2009\u03bcL were injected. According to our previous results of HPLC analysis, the amounts of CoQ~0~ (Fig.\u00a0[1A](#Fig1){ref-type=\"fig\"}) in the fermented culture broth of *Antrodia camphorata* were 17.3% (254\u2009nm) and 13.5% (220\u2009nm), respectively^[@CR60]^. The purity of CoQ~0~ is \u226598%, which is similar to commercially available CoQ~0~.\n\nCell culture and treatment {#Sec24}\n--------------------------\n\nHuman ovarian cancer (SKOV-3) cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Two epithelial ovarian cell lines (A2780 and CP70) were kindly provided by Dr Cheng-I Leen (National Chung Cheng University, Taiwan). Human ovarian surface epithelial (IOSE) cells were kindly provided by Dr Michael Chan (National Chung-Cheng University, Taiwan). Cells were cultured in DMEM/F12 supplemented with 10% heat-inactivated FBS, 2\u2009mM glutamine and 1% penicillin- streptomycin-neomycin at 37\u2009\u00b0C in a humidified incubator with 5% CO~2~. Cultures were harvested and monitored for changes in cell number by counting cell suspensions using a hemocytometer (Marienfeld, Germany). Cell morphology was examined using phase contrast microscopy (200\u2009\u00d7\u2009magnification). Cells were treated with increasing concentrations of CoQ~0~ (0--40\u2009\u00b5M), the incubation time was varied depending on the assay. Wherever applicable, cells were pretreated with various pharmacological inhibitors, including NAC (2\u2009mM), Z-VAD-FMK (20 \u03bcM), 3-MA (2\u2009mM) or CQ (10 \u03bcM) for 1\u2009h, and then incubated with indicated concentration of CoQ~0~ for 24\u2009h.\n\nAssessment of cell viability by MTT assay {#Sec25}\n-----------------------------------------\n\nThe effect of CoQ~0~ on viability of human ovarian carcinoma **(**SKOV-3, A2870 and A2870/CP-70) and human ovarian surface epithelial (IOSE) cells was determined by the MTT colorimetric assay. Briefly, cells (2.5\u2009\u00d7\u200910^4^ cells/well in 24-well plates) were treated with different concentrations of CoQ~0~ (0--40\u2009\u00b5M) for 24\u2009h. After CoQ~0~ treatment, 400\u2009\u03bcL 0.5\u2009mg/mL MTT in PBS was added to each well. After incubation at 37\u2009\u00b0C for 2\u2009h, an equal volume of DMSO (400\u2009\u03bcL) was added to dissolve the MTT formazan crystals, and the absorbance was measured at 570\u2009nm (A~570~) using an ELISA microplate reader (\u00b5-Quant, Winoosky, VT, USA). The percentage (%) of cell viability was calculated as: (A~570~ of treated cells/A~570~ of untreated cells)\u2009\u00d7\u2009100. The assay was performed in triplicate at each concentration.\n\nCell-cycle analysis {#Sec26}\n-------------------\n\nCellular DNA content was determined by flow cytometry using the propidium iodide (PI)-labeling method as described previously^[@CR31]^. Briefly, cells were seeded at a density of 4\u2009\u00d7\u200910^5^ cells/dish in 10\u2009cm dishes, and the cell-cycle was synchronized by the addition of double thymidine (3\u2009mM) for 16\u2009h. Cell-cycle-synchronized cells were then washed with PBS and re-stimulated to enter the G1 phase together by the addition of fresh medium, which also contained various concentrations of CoQ~0~ (0--30\u2009\u00b5M). Cells were harvested at 24\u2009h, and the cell-cycle analysis was performed using a FAC-Scan cytometry assay kit (BD Biosciences, San Jose, CA, USA) equipped with a single argon ion laser (488\u2009nm). The DNA content of 1\u2009\u00d7\u200910^4^ cells/analysis was monitored using the FACS Calibur system. Cell-cycle profiles were analyzed with ModFit software (Verity Software House, Topsham, ME, USA).\n\nProtein isolation and immunoblotting {#Sec27}\n------------------------------------\n\nCells (1\u2009\u00d7\u200910^6^cells/10-cm dish) were incubated with CoQ~0~ (0--30\u2009\u00b5M) for the indicated time periods. After incubation, cells were harvested, pooled, washed once with PBS and suspended in 89\u2009\u03bcL of lysis buffer (10\u2009mM Tris-HCl, pH 8, 32\u2009mM sucrose, 1% Triton X-100, 5\u2009mM EDTA, 2\u2009mM DTT and 1\u2009mM phenylmethyl sulfonyflouride). The cell lysates were maintained on ice for 30\u2009min and then centrifuged at 12000\u2009rpm for 30\u2009min at 4\u2009\u00b0C. Total protein content was determined using Bio-Rad protein assay reagent (Bio-Rad, Hercules, CA, USA) with bovine serum albumin as a standard. The protein extracts were mixed with sample buffer (62\u2009mM Tris-HCl, 2% SDS, 10% glycerol, and 5% \u03b2-mercaptoethanol), and the mixture was boiled at 97\u2009\u00b0C for 5\u2009min. Equal amounts (50\u2009\u03bcg) of denatured protein samples were separated by 8--18% SDS-PAGE and then transferred onto polyvinylidene difluoride (PVDF) membranes overnight. The membranes were blocked with 5% non-fat dried milk in PBS containing 1% Tween-20 for 1\u2009h at room temperature, followed by incubation with primary antibodies for overnight. The membranes were then incubated with either a horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse antibodies for 2\u2009h prior to development using a Chemiluminescent substrate (Millipore, Billerica, MA, USA). The changes in protein intensities were digitized using the ImageQuant\u2122 LAS 4000 mini (Fujifilm). Densitometric analyses were performed using commercially available quantitative software (AlphaEase, Genetic Technology Inc. Miami, FL, USA). Each assay was performed in triplicate, with control representing 1 fold, and changes in protein intensities were presented as histograms.\n\nMeasurement of intracellular ROS production {#Sec28}\n-------------------------------------------\n\nIntracellular ROS accumulation was detected by fluorescence microscopy using the cell-permeable fluorogenic probe DCFH~2~-DA. Cells (2.5\u2009\u00d7\u200910^4^ cells/mL) were cultured in DMEM/F12 medium that had been supplemented with 10% FBS, and the culture medium was replaced when the cells had reached 80% confluence. To evaluate ROS generation in a time-dependent manner, the cells were treated with CoQ~0~ (0--30\u2009\u00b5M) for 0--30\u2009min. Then culture supernatants were removed, and incubated with non-fluorescent DCFH~2~-DA (10\u2009\u03bcM) in fresh medium at 37\u2009\u00b0C for 30\u2009min. Intracellular ROS production was quantified by measuring the intracellular accumulation of dichlorofluorescein (DCF), which is caused by the oxidation of DCFH~2~. The DCF-stained cells were captured using fluorescence microscope (200\u2009\u00d7\u2009magnification) (Olympus, Center Valley, PA, USA). The fluorescence intensity was quantified using analysis LS 5.0 soft image solution (Olympus Imaging America Inc.,). The percentage of fluorescence intensity (ROS generation) was compared with untreated control cells, which were arbitrarily assigned a value of 100%.\n\nDetection and quantification of acidic vesicular organelles (AVOs) formation {#Sec29}\n----------------------------------------------------------------------------\n\nFormation of AVOs was detected using AO stain in SKOV3 cells treated with various concentrations of CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h. Briefly, after designated treatments cells were washed with PBS twice, followed by staining with AO (1\u2009\u03bcg/mL) and dilution in PBS containing 5% FBS for 15\u2009min. After staining, cells were washed with PBS and covered with PBS containing 5% FBS. The cells were observed under a red filter fluorescence microscope and formation of AVOs in cells was visualized at 200\u2009\u00d7\u2009magnification. AO is a lysosomotropic metachromatic and weak base membrane-permeant fluorescent dye, whose fluorescence emission is concentration dependent, from red at high concentrations (in lysosomes) to green at low concentrations (in the cytosol), with yellow as intermediate in some conditions^[@CR61]^.\n\nGFP-LC3 plasmid transfection and detection of GFP-LC3 dot formation in cells {#Sec30}\n----------------------------------------------------------------------------\n\nLC3 cDNA was a kind gift from Dr. Jiunn- Liang Ko (Chung-Shan Medical University, Taiwan). GFP-LC3 fusion protein was used to make the autophagosomes visible in cells. The cells were seeded onto coverslips placed onto a 6-well plate (1.8\u2009\u00d7\u200910^5^ cells/well). After overnight culture, cells were transfected with 2.5\u2009\u03bcg GFP-LC3 expressing plasmid in each well of a 6-well plate using Lipofectamine (Invitrogen, Carlsbad, CA, USA) and incubated for 24\u2009h. The medium was removed and fresh medium containing CoQ~0~ (0--30\u2009\u00b5M) was added to the wells. At the end of CoQ~0~ treatment (24\u2009h), cells were washed twice with PBS, and expression of GFP-LC3 dots in cells were detected by a laser scanning confocal microscope at 200 \u00d7 magnification.\n\nDetermination of apoptotic cells by Annexin-V/PI staining {#Sec31}\n---------------------------------------------------------\n\nDouble staining for Annexin-V-FITC and PI (propidium iodide) was performed to estimate the apoptotic rate of SKOV-3 cells. Briefly, cells were incubated with CoQ~0~ (0--30\u2009\u00b5M) for 24\u2009h, trypsinized, washed twice with PBS, and centrifuged at 1000\u2009rpm for 5\u2009min. Then, cells (1\u2009\u00d7\u200910^6^cells/10-cm dish) were suspended in binding buffer (500\u2009\u03bcL) and double-stained with Annexin-V-FITC and PI for 15\u2009min at room temperature. Then the result green (FITC) and red (PI) fluorescence of each sample was quantitatively analyzed by FACS Caliber flow cytometer (Becton Dickinson, San Jose, CA, USA) and Cell Quest software. The obtained results were interpreted as follows: (Q3) cells negative for both PI and Annexin-V-FITC staining were considered normal live cells. (Q4) PI-negative, Annexin-V- FITC-positive stained cells were considered in early apoptosis. (Q2) PI-positive, Annexin-V-FITC-positive stained cells were considered in late apoptosis. (Q1) PI-positive, Annexin-V-FITC-negative stained cells were considered in necrosis.\n\nApoptotic DNA fragmentation by TUNEL assay {#Sec32}\n------------------------------------------\n\nDNA fragmentation in cells was measured using commercially available TUNEL assay kit (Calbiochem, San Diego, CA, USA). After treatment with designated inhibitors (1\u2009h) and CoQ~0~ (0--30\u2009\u00b5M, 24\u2009h), apoptotic cells (2\u2009\u00d7\u200910^4^ cells/well in 8-well chamber) were harvested, fixed with 4% formaldehyde and mounted on glass slides. Apoptosis was detected by labeling the 3\u2032-OH ends of fragmented DNA with biotin-dNTP using DNA I klenow fragment at 37\u2009\u00b0C for 1.5\u2009h. The slides were then incubated with horseradish peroxidase-conjugated streptavidin, followed by incubation with 3,3\u2032-diaminobenzidine and H~2~O~2~. The fragmented DNA was identified by their fluorescence nuclei under a fluorescence microscope (200 \u00d7 magnification). The green fluorescence intensity was quantified using a squared section of fluorescence-stained cells with analysis LS 5.0 soft image solution (Olympus Imaging America Inc., PA, USA). The percentage of fluorescence intensity is directly proportional to the percentage of apoptotic cells, compared to untreated control cells, which were arbitrarily assigned a value of 100%.\n\nAnimals {#Sec33}\n-------\n\nFemale athymic nude mice (BALB/*c-nu*), 5--6 weeks of age, were purchased from The National Laboratory Animal Center (Taipei, Taiwan) and were maintained in caged housing in a specifically designed pathogen-free isolation facility with a 12\u2009h/12\u2009h light/dark cycle. The mice had free access to rodent chow (Oriental Yeast Co Ltd., Tokyo, Japan) and water *ad libitum*. All animal experiments were strictly followed \"The Guidelines for the Care and Use of Laboratory Animals\" published by the Chinese Society of Animal Science, Taiwan. The entire animal study protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of China Medical University, Taichung, Taiwan.\n\nTumor cell inoculation and CoQ~0~ treatment {#Sec34}\n-------------------------------------------\n\nA total of 9 mice were randomly divided into three groups containing three animals in each group. SKOV3 cells (3\u2009\u00d7\u200910^6^ cells, 100 \u03bcL) were mixed in a 100 \u03bcL matrix gel and then injected subcutaneously in a volume of 200 \u03bcL into the right hind flanks of nude mice. The experiments were performed using cells that had been passaged fewer than 20 times. After cell inoculation for 7 days, the treatment groups received CoQ~0~ (1.5 and 2.5\u2009mg/kg b.w.) in a volume of 100 \u03bcL *via* intraperitoneal injection every 4 days for 52 days. The control group received the vehicle (PBS) only. CoQ~0~ was dissolved in PBS (containing 1% DMSO) and injections were scheduled in the morning between 10 AM and 11 AM throughout the study. To monitor drug toxicity, the body weight of each animal was measured for every 4 days. Tumor volume in mice was compared with caliper measurements of tumor length, width and depth, and then calculated every 4 days using the formula: length\u2009\u00d7\u2009width^2^\u2009\u00d7\u20091/2. On day 52, all mice were sacrificed and tumor tissues were removed and weighed. A veterinary pathologist examined the mouse organs, including the liver, lungs and kidneys.\n\nHistopathological analyses {#Sec35}\n--------------------------\n\nThe biopsied tumor tissues were embedded in paraffin and cut into 3 mm thick sections, placed in plastic cassettes and immersed in neutral buffered formalin for 24\u2009h. The fixed tissues were processed routinely and then embedded in paraffin, sectioned, deparaffinised, and rehydrated using standard techniques. The extent to which treatment shrunk tumor cells was evaluated by assessing the mitotic cell division in xenografted tumor sections using hematoxylin and eosin (H & E) staining for light microscopy. Then the number of mitotic cells or apoptotic cells in microscopic fields were averaged and presented as histogram. TUNEL apoptosis in the sections of CoQ~0~ or vehicle treated tumors was measured using *in situ* cell death detection kit (Roche, Mannheim, Germany).\n\nStatistical analyses {#Sec36}\n--------------------\n\nData from *in vitro* experiments were presented as mean and standard deviation (mean\u2009\u00b1\u2009SD). For *in vivo* experiments, mean data values are presented with standard error (mean\u2009\u00b1\u2009SE). Data from both studies were analyzed using analysis of variance followed by Dunnett's test for pair-wise comparison. Statistical significance was defined as \\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.01, \\*\\*\\**p*\u2009\\<\u20090.001 compared to untreated control cells, and significant at ^\\#^ *p*\u2009\\<\u20090.05; ^\\#\\#^ *p*\u2009\\<\u20090.01; ^\\#\\#\\#^ *p*\u2009\\<\u20090.001 compared to CoQ~0~ alone treated cells.\n\nElectronic supplementary material\n=================================\n\n {#Sec37}\n\nSupplementary Information\n\n**Electronic supplementary material**\n\n**Supplementary information** accompanies this paper at doi:10.1038/s41598-017-08659-7\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThis work was supported by the grants MOST-104-2320-B-039-040 -MY3, MOST-103-2320-B-039-038-MY3, NSC-103-2622-B-039-001-CC2, CMU103-ASIA -12, and CMU103-ASIA-09 from the Ministry of Science and Technology, National Science Council, Asia University and China Medical University, Taiwan. This study was supported by China Medical University under the Aim for Top University Plan of the Ministry of Education, Taiwan (CHM106-5-3).\n\nT.J.T., Y.C.H. and H.L.Y. performed the laboratory experiments and collected the data. Y.C.H., H.L.Y. and M.K. interpreted the data, drafted and revised the manuscript. J.Y.L., H.J.C., C.M.L. and Y.C.S. coordinated the study. Y.C.H. and H.L.Y. designed and supervised the study. All authors read and approved the final version of the manuscript.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Background\n==========\n\nDelivering a death notification is an emotionally difficult task regardless of previous experience. There are no words that can lessen the impact of the horrendous message. Every year millions of people are notified that a close relative has died. The manner in which the news is communicated may be crucial to the emotional recovery of relatives. The management of both patients and their relatives is of vital importance when one is suffering from a severe disease \\[[@B1]\\]. The emotional state of both patients and their relatives can elicit emotions in health care staff which can lead to hampered communication \\[[@B2]\\]. It is crucial for health care workers to observe and interpret these emotions in order to understand the emotions that arise within \\[[@B3]-[@B7]\\].\n\nForum Theatre, originally developed by Augusto Boal \\[[@B8]-[@B10]\\] incorporates audience participation into a performance such that audience members make suggestions to actors in order to alter the outcome of a scene. Engaging in a performance can incite questions, experiences and issues that may otherwise be overlooked. In a broad meaning forum play is a simulation education because it is a simulation of a specific event with a specific goal. Forum play may be a useful exercise in socio-analytic role playing in the teaching of delivery of death notification. We developed a unique method termed \\\"Marathon Death\\\" to illustrate and solve problems in human relations based on Forum Theatre. We have successfully used this method for preparing medical students for the delivery of death notifications. In the model, all students are engaged in a form of Forum Theatre but with a focus on a pedagogic situation which we called forum play. Marathon Death is a valuable tool for discussing difficult questions and issues pertaining to death notification, developed specifically for use as a training instrument for action and understanding of human behaviour at the Medical school at Ume\u00e5 University, in northern Sweden. Through the process, teachers become actively engaged with students, developing strong relationships and trust.\n\nDuring the fourth year of medical school students participate in a three-day retreat as part of a course in professional development. The course addresses challenging patient-physician communication issues and the students are not informed of specific exercises within the course curriculum in advance. In groups of 8-10, students have the opportunity to participate in an exercise on the communication of a death notification. All groups work simultaneously and each group is led by a teacher. The teacher plays the role of next of kin and gives a realistic introduction to the exercise by describing the events leading up to the death of his/her husband/wife as realistically and vividly as possible. The objective of the exercise is for the medical student to take on the role of medical doctors and relay the death notification to the next of kin without paraphrasing. The element of surprise is an important component for realism. After initial information about the exercise, students wait outside the classroom until their turn. The exercise is performed without an audience and is videotaped. Students are told that the exercise will start while the relative is in the waiting room anticipating news of their spouse. The teacher, playing role of next of kin, pretends to be unaware that his or her spouse has died and inquires as to the state of the relative. The exercise is stopped when the student utters the word \\'died\\' or \\'deceased\\'. The entire interaction takes place over a few minutes. Many students become emotionally affected by the experience, sometimes to the extent that they have difficulty perceiving when the teacher ends the exercise and goes out of character. Each student then has the opportunity to privately reflect on their feelings while their colleagues participate in the exercise. After all students in the group have taken part in the exercise the group gathers for discussion and reflection. Subsequently, each taped exercise is analyzed. During playback the recording is paused and participants are able to re-enact a continuation on the scene in our special model of forum playing. The novelty in the way we use forum play lays in the use of co actors, facilitators and play back that enhances the active learning through interaction and dialogue. Through the facilitator, students are helped to identify personal learning objectives and to practice behaviours through interaction with colleagues. The facilitator also draws feedback from peer observers during the playback moment enabling an added understanding from the work context.\n\nThe aim of the study was to explore reflections, attitudes and ideas toward training practices in the deliverance of death notification among medical students.\n\nMethods\n=======\n\nStudy design and participants\n-----------------------------\n\nA qualitative approach was used in the exploration of student attitudes and ideas toward the training exercise. Interviews with semi-structured questions promoting open ended content and direction were conducted. The interviews were analyzed through qualitative content analysis \\[[@B11]\\]; a method of analyzing written or verbal communication in a systematic way \\[[@B12]\\]. The method was selected based on its usefulness in the analyses of experiences, reflections, and attitudes \\[[@B13]\\].\n\nA maximum variation sample was chosen with the following variables: males and females, large age range, variety of nationalities, variety of prior experience in death notifications and participation in Marathon Death exercise groups. During the fall of 2008, 85 students were involved in the Marathon Death exercise. Of the 85 students who participated, 10 were asked to volunteer in the study. All 10 students agreed to participate in the interviews that took place within two weeks after the exercise. Pertinent background information of each student is given in Table [1](#T1){ref-type=\"table\"}.\n\n###### \n\nParticipant background information\n\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  **Emilia**35 years, from Sweden, had never been exposed to a death notification neither professionally nor private.\n  **Erik**20 years of age, from Sweden, His experience from death notification was limited. He had received one notification by telephone; it was when his brother died unexpected.\n  **Fredrik**25 years, from Germany. He had received a death notification when a relative died when he was younger, at that time his father had informed him.\n  **Kajsa**25 years, from Sweden. She experienced the trauma of having her father died unexpectedly under tragic circumstances after a longer illness and she got the notification over the telephone of the care taking staff. Kajsa usually put herself or her family in the situations she encounters during her education. When she faces a patient she always see a relative in the patient and think of how she would feel if a physician came and gave the notification that a close relative of hers had died. The thought is always with her in clinical situations.\n  **Li**25 years, from China, had experienced a death notification through a letter when a relative had died.\n  **Lina**25 years, from Sweden, had never been exposed to a death notification neither professionally nor private.\n  **Lotta**35 years, from Sweden. She had never had the experience of a death notification but had gotten a notification several years ago regarding a distant relative.\n  **Peder**20 years, from Sweden, A close relative to Peder died a few years ago and it was the parents who conveyed it over the telephone. The notification was unexpected although the relative was elderly. Peder discovered that it was rather nice at that time to get a moment to himself and just reflect over the news.\n  **Peter**20 years, from Sweden. Peter had a close relative who died unexpected and at that time it was his father who gave the notification. He also had had other relatives who had died after a period of disease. In those cases he thought that it was more of a relief for the relatives when they finally died. The horrendous of the situation was the disease and all that came before death.\n  **Sara**20 years, from Sweden, had experienced two deaths in the family. She had been there with her grandmother when she died expectedly after experienced a good, long life. Her grandmother had gotten weaker and weaker and it had been a relief when it came. It had felt nice when all family members had been able to say their farewell.\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nParticipant background information pertinent to the Marathon Death exercise.\n\nThe study design and protocol were approved by the head of the Professional Development Course faculty at Ume\u00e5 University, Sweden. The background and aim of the study was explained to all students before participation. The work was carried out in accordance with the Declaration of Helsinki, including, but not limited to there being no potential harm to participants. The students were assured of confidentiality and strict anonymity. Informed consent of students was obtained for publication.\n\nCollection and analysis of data\n-------------------------------\n\nData were collected from semi-structured interviews to capture participant experiences from a personal perspective \\[[@B14]\\]. The interview was guided by the following topics: student experience of the death notification exercise including the establishment of contact with the next of kin and the student\\'s own perception of their emotions (i.e. interest, warmth and empathy) in that situation; the amount of information the next of kin perceived, use of time during the exercise and the student\\'s interpretations of the relative\\'s feelings when the death notification was delivered. The interview guide was used throughout the study. Each interview was performed by one researcher and lasted about 60 minutes. The interviews were tape-recorded with permission from the students and then transcribed verbatim.\n\nThe interviews were analyzed by means of qualitative content analysis performed in several steps \\[[@B11]\\], using a deductive content analysis approach with a structured matrix \\[[@B15]\\]. Bloom\\'s taxonomy \\[[@B16]\\] was used as the structured matrix. The data were analyzed by a teacher (who is also a physician) of the professional development course, a pedagogue experienced in forum play and qualitative research and a physiotherapist with experience in occupational health care and qualitative research. Each author performed an individual preliminary analysis of the entire text in relation to the aims of the study. The text was divided into meaning units, each comprised of several words, sentences, or paragraphs containing interrelated aspects of content and context. Reoccurring, but independent words and phrases regarding thoughts, emotions and attitudes towards the exercise emerged. Taking context into consideration, the units were condensed and each was labelled with a code related to the three levels of Bloom\\'s taxonomy (the cognitive, affective and psychomotor domains) (See Table [2](#T2){ref-type=\"table\"}). An additional rereading of the transcripts was performed with the aim of further analyzing ways in which these categories were linked. Excerpts pertaining to identified semantic relationships between specific expressions were clustered to form thematic relationships, which in turn were analyzed for linked thematic patterns. Emerging categories and thematic patterns were synthesized and analyzed based on Bloom\\'s taxonomy \\[[@B16]\\].\n\n###### \n\nExamples of codes and categories\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Codes                                                                                                                              Categories according to Bloom\\'s taxonomy\n  ---------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------\n  *Professionalism*                                                                                                                  Cognitive Domain\n\n                                                                                                                                     \n\n  \u2003*-expressions of interest, listening, planning ahead, expectations, content of communication, how to communicate, how to react*   \n\n                                                                                                                                     \n\n  *The need for time*\\                                                                                                               \n  *A soothing environment*                                                                                                           \n\n                                                                                                                                     \n\n  *Unpleasant*\\                                                                                                                      Affective Domain\n  *Focused*\\                                                                                                                         \n  *Excitement*\\                                                                                                                      \n  *Low-spirited*\\                                                                                                                    \n  *Relieved*\\                                                                                                                        \n  *Content*\\                                                                                                                         \n  *Self-reliance*\\                                                                                                                   \n  *Sensitivity*\\                                                                                                                     \n  *Revised behaviour*                                                                                                                \n\n                                                                                                                                     \n\n  *Body language*\\                                                                                                                   Psychomotor Domain\n  *Body contact*\\                                                                                                                    \n  *Eye contact*                                                                                                                      \n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nCodes and categories describing experiences during the Marathon Death training exercise according to Bloom\\'s taxonomy.\n\nResults\n=======\n\nAnalysis of the reflections, attitudes and ideas towards training in delivering death notifications among the medical students are presented as both codes and descriptive quotations (in italic) sorted into categories according to Bloom\\'s taxonomy (Table [2](#T2){ref-type=\"table\"}).\n\nThe cognitive domain\n--------------------\n\nReflections on the learning outcome of the exercise were categorized according to **the cognitive domain**of Bloom\\'s taxonomy. Overall, a high degree of comprehension and application of knowledge was reported. Participants reported that conversations were not as anticipated. Students expected a calmer situation where the next of kin would react as they imagined people in real-life should react in the same situation. The exercise created a virtual situation where students described reflecting over their own and the next of kin\\'s reactions and emotions and their own future as a professional physician. Participants began to understand that there is no set of rules for human behaviour in this special kind of situation. Some students presented an analysis on what they could have done differently and pondered over possible alternate outcomes; *I have experienced something that will*occur *in the future, I could very well*be *the next of kin in real life and I could also be the one informing someone of the death notification*.\n\nThe students expressed the importance of *being professional*when giving notifications. Being professional included *expressing interest*while *simultaneously formulating a response. Control*of the situation was perceived as vital including *having a plan*along with having a *period of the time*before the notification in a *soothing environment*in order *to get oneself together before*delivering the news. *How to convey the message*was frequently discussed; *how much information to give*, *what to say, what not to say, how to express death (e.g. deceased, died\\...)*and how to *end the conversation. If I should start with describing the events leading up to the death of his or her relative or if I should cut right to the core saying that the relative was dead\\...//*. How to convey the message also included *setting the tone*; *how smoothly you can make it through the situation without giving (the notification) abruptly\\...*// *how to set the tone to make the next of kin realize from the start that something was not right. It would be wrong to start with a happy tone only to have to later arrive at the dreaded notification*. The students tried to imagine *what to expect*during the notification and *how to react*towards the reaction from the next of kin and some students decided to improvise as the situation progressed; *She (next of kin) was talking all the time. I didn\\'t know if I was to interrupt her or wait until she stopped talking and became quiet*.\n\nThe affective domain\n--------------------\n\nThe emotional aspects of the exercise were investigated using the **affective domain**of Bloom\\'s taxonomy. All students experienced *a tension*prior to the exercise and great *anticipation*. Some participants had positive expectations such as *excitement*to try to give a death notification and a bit of *an adrenaline kick*to *try to handle*the situation. Others were *relieved*since they had *been performing another exercise earlier that had turned out well the day before*. Yet, others perceived the exercise as *difficult*and felt *nervous*and *insecure*. There were feelings of *unpleasantness*and some students were *worried*how the experience would affect their emotions during the exercise and how they would react in the future, in their professional life.\n\nThe forum play was perceived as *artificial yet at the same time very real*. Some felt *focused*and *at ease*when the exercise started. After the exercise all students were somehow *affected*. Some participants felt *upset, shocked*and *stressed*. There were feelings of *inadequacy; I realize how incredibly inadequate I am. Normally I feel that I usually can talk my way through most things and know fairly well what to say but I just had a blackout\\... it was really unfamiliar*. At the same time the students were *content*with their *gained insight*but, *it was rather heavy*.\n\nThe psychomotor domain\n----------------------\n\n**The psychomotor domain**of Bloom\\'s taxonomy was used to express manual and physical skills. Most students reflected on their *body language*and on *physical body contact*with the next of kin and how that started the dialogue; *I took his*(next of kin) *hands and I also experienced that he searched for*mine. *I felt that I could ask questions, that we had good communication*. *I had a lot of body contact, I could see on the video afterwards. It was nothing I thought about but it was what I tried to do, partly when I took her around the shoulders, and then later when she was upset talking about the upcoming dinner*(plans that she had made), *then I took rather firmly her hand with both*(of my) *hands and said; \\\"your friend will have to wait\\\" and looked her once again in the eyes to subdue her chatter*.\n\nThe importance *of portraying the same*(emotion) *in both speech and in body language*while delivering information to the next of kin realize from the start was discussed; *When she*(next of kin) *started to chatter hysterically I didn\\'t listen and lost focus, I was rather upright in my body language and I took hold of her shoulders. When she walked around I took her around her shoulders and looked her rather determined in the eyes and said; \\\"Now, I want us to sit down.\\\" And then she*sat *right away*. Having good *eye contact*with the next of kin when giving the notification was also important; *\\...to get eye contact, to make sure we were on the same level*.\n\nThe learning process\n--------------------\n\nThe following is the reflections of one participant (Fredrik) on the learning process of the exercise.\n\nI had to decide If I would start with describing the course of events to the next of kin leading up to the death of the patient, or if I would go straight to the punch line; that the patient had died, or if I would start to unravel\\...// I started to think how I could set the tone from the start so that the next of kin receiving the information would understand from the beginning that all was not right. It would have felt wrong to have a merry tone at the start only to later come to the final information; \\\"well your husband is dead.\\\" There would be no smooth transition, so to speak, regarding the information/While there is no way to make a true smooth transition in a situation like that, it is important not to be abrupt. It was important to decide whether to talk about what had happened and what went wrong and so on, in which case the next of kin may say; \\\"well\\...is he or she alive or not?\\\" I feel that for myself I wouldn\\'t want to sit and listen to someone talking about cardiac arrest and resuscitation attempts for very long, but would rather pop the question if I didn\\'t understand what had happened\\...\n\n*After the exercise I thought it is good just to be put in this situation, to be able to prepare for it. If someone told you; \\\"you must notify someone about the death of their relative.\\\" You would start to process, how should I go about this? What strategy should I use? What can I expect will happen\\... right there and then? A lot is won just by the thought process. If instead you came to a room were someone would say; \\\"the next of kin has already been notified\\\", then you still have done your work and by that learned a whole lot. The\\...video recording was actually really good for the exercise. I was able to observe that I tried to connect with the person I spoke to, what I said, that I tried to take hold of them, to get eye contact, to make sure we were on the same level. I attempted at least to form some communication scheme\\...to get the chance to give my information. It was at least an honest attempt, although I thought that I didn\\'t succeed with it. In the aftermath I think that you should prepare for the unexpected (pertaining to) reactions from the next of kin and you shouldn\\'t take for granted that the next of kin will act in a certain way. You cannot expect that the next of kin will sit quietly and calmly waiting to be served the news, They may bombard you with information and questions before you get a chance to gather your thoughts on what to say. You need to have a plan B the whole time*.\n\nThe learning process becomes evident when using Bloom\\'s taxonomy to analyze the thoughts of the students. As exemplified in the above reflections there is a clear progression in learning. Prior to the exercise Fredrik portrayed *motivation*: he demonstrated *sensitivity*towards the individual he was about to meet and had formulated a plan on how to carry out the exercise. After the exercise he reflected over the need for balance; comparing, relating and synthesizing values. He exhibited *self-reliance*and thought of the possibilities of how to *revise his behaviour*in light of the experience. These concepts may be interpreted as components of learning in the affective domain of Bloom\\'s taxonomy; responding to and evaluating an experience. Fredrik reflected over the data and gave an interpretation of *how he would carry out the task*he was given. After the exercise he portrayed the ability to analyze the exercise and the learning outcome. He gave a synthesis of the exercise and what to *improve*on. These reflections can be construed as developments in intellectual skills (i.e. the cognitive domain of Bloom\\'s taxonomy).\n\nDiscussion\n==========\n\nThe exercise, Marathon Death, was introduced into the curricula at the medical school at Ume\u00e5 University, Sweden to meet the need for training on communication in difficult situations. Marathon Death is a useful pedagogic exercise and model, based on Forum Theatre that enables students to train on giving death notifications. The exercise was perceived as a valued training experience in a safe and controlled environment. In addition, students and teachers developed positive relationships and trust through the process. The students\\' thoughts and feelings prior to the exercise varied and did not seem to reflect gender, age, nationality, or prior experience of death notification. The students expressed the importance of having an opportunity to train before they faced this difficult task in their future professional lives. The ability of a physician to communicate death notifications to relatives in an appropriate manner can have positive effects on the grief response and subsequent resolution of the loss \\[[@B17]\\]. Despite the importance of this process, little formal training is given to health care staff during their education or in their professional life \\[[@B4],[@B18]-[@B22]\\].\n\nAttitudes, experiences and emotions are important complex learning outcomes. It is important to investigate all aspects of the learning process in the evaluation of an exercise like Marathon Death. The current study used Bloom\\'s taxonomy \\[[@B16]\\] to categorize reflections made by students during interviews. The cognitive domain explored knowledge and the development of intellectual skills. The affective domain investigated the emotional reaction of the students to the task. The psychomotor domain included reflections on body language and eye contact during the exercise but was not analyzed to any higher extent.\n\nThe attitudes elicited by the exercise were investigated. The exercise was specifically designed to bring out emotions and reflections. According to Bolton and Heathcote the performance enables the active (rather than passive) attainment of knowledge in a situation \\[[@B23]\\]. Thus, the design of the exercise contributes an additional dimension of education; the students are provided the opportunity to use the knowledge and skills that are important for their profession. Rather than reconstructing and reproducing knowledge, the students become active players in their own education; they are given the chance to put their newly acquired knowledge and skills to use. The affective domain of Bloom\\'s taxonomy \\[[@B16]\\] was used to explore the emotional responses that students expressed. All students were affected by the exercise instructions even prior to entering into the forum play. It is clear that the exercise starts an immediate emotional process even while the students reflect over the exercise instructions.\n\nThe psychomotor domain of Bloom\\'s taxonomy \\[[@B16]\\] was not the main focus of the Marathon Death exercise; focus was instead on the cognitive and affective domains. It is difficult to evaluate the psychomotor domain through conversational skills. However, most of the students reflected on the body language they observed in the videos which could be interpreted as an early stage in the learning of complex skills.\n\nThe results of the current study indicate that Marathon Death is a valuable tool for the training of medical students in delivering death notifications. Several of the students chose to give the notification in such a way that they felt that they, themselves would prefer to receive a death notification. The process of reflecting on their own preferences if the roles were reversed may have enhanced the emotional component of the exercise. It is of great value for the students to have insight into how they may act under pressure; how they may express themselves and how they may treat individuals in crisis while at the same time dealing with their own feelings. However, the exercise comes with risk. The perception of failure in such an exercise may increase tension and make job demands feel insurmountable. While the exercise comes with risk, it is clear that the feedback from peers and an experienced tutor fulfil an important function, pointing towards positive features and thereby strengthening self-confidence. Moreover, the recordings provide the opportunity to pause at any point during the interaction and then to continue to act out the rest of the scene in front of the group, enhancing learning by dividing the exercise into meaningful components. The great variety in experience and situations of each participant made the leadership role difficult for the teachers. Thus, experience and support for the teachers in the form of meetings prior to and following the exercise is essential. In addition, it is recommended that some form of support for students is provided in the event that the exercise evokes strong feelings that need some extra assistance to work through.\n\nStudents with a prior negative experience involving a death notification in their own private life were more likely to report experiencing moderate to very high stress during the exercise. These participants reflected over their private experiences and their desire to notify the next of kin in the exercise according to how they themselves would have preferred receiving a death notification. Interestingly, there were no spontaneous comments from the students of perceived stress experienced either prior to or during the exercise on the perceived usefulness of the exercise. In other words, even students who perceived stress also perceived the task as useful.\n\nInterpretation of the results is based on qualitative data. Shenton describes that qualitative data are specific to the context and the individuals investigated and that care must be taken when interpreting results \\[[@B24]\\]. The participants in the present study are medical students and caution must be used when generalizing results to any other population. It may be argued, however that there is a certain amount of transferability of these qualitative findings to students in contexts similar to those in the present study. An advantage of the study design is the detail with which the data were described and the use of Bloom\\'s taxonomy \\[[@B16]\\] for analysis of this special kind of pedagogical data. The study design can be repeated in a variety of contexts to explore whether students in similar (and different) contexts learn in similar ways \\[[@B24],[@B25]\\]. Such findings will extend the existing body of knowledge with regard to student learning.\n\nThere is a symbiotic relationship between information garnered from qualitative and quantitative data. Qualitative research is effective in the development of hypotheses in \\\"newly emerged or under-researched areas\\\" which can then be tested more effectively using quantitative measures at a later stage \\[[@B25]\\]. Further quantitative studies are needed to investigate other aspects of student learning.\n\nConclusions\n===========\n\nThe current study offers intriguing preliminary findings concerning a new valuable exercise in death notification training. Marathon Death is an effective method of encouraging self-reflection concerning the process of doctoring and the doctor-patient relationship in the process of death notification. Marathon Death appears to be a valid exercise for students in reducing the anxiety and stress inherent to death notification. Overall, forum play with a playback feature should be considered a new form of simulated learning and a useful method of enhancing and enriching training in death notification.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nAN conceptualised and implemented the study, developed study materials, collected and analysed data, and drafted the manuscript, AFW analysed and interpreted the data and drafted the manuscript, TG conceptualised the study and supervised data collection. AN and AFW helped review drafts of the manuscript. All authors have read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1472-6920/11/74/prepub>\n"}
+{"text": "1. Introduction\n===============\n\nThe relationship between hypertension and short sleep duration has raised concerns of cardiologists in recent years. Dozens of cross-sectional \\[[@B1-ijerph-12-00488],[@B2-ijerph-12-00488]\\] and longitudinal studies \\[[@B3-ijerph-12-00488],[@B4-ijerph-12-00488],[@B5-ijerph-12-00488]\\] have been conducted in populations of different ages, genders or races to investigate the potential association of short sleep duration in relation to hypertension and most of them indicated short sleep duration was related with the prevalence and incidence of hypertension. However, a few studies noted that sleep had quantitative and qualitative aspects \\[[@B6-ijerph-12-00488],[@B7-ijerph-12-00488]\\] and it may be not comprehensive to assess sleep only by sleep duration in those studies. Particularly, Bansil*et al.*'s study, which investigated the role of sleep duration and sleep quality in hypertension development separately, found that sleep duration alone failed to affect the hypertension prevalence, whereas the combination of short sheep duration and sleep disorders could influence hypertension \\[[@B7-ijerph-12-00488]\\]. Furthermore, the literature documented that poor sleep quality was associated with obesity \\[[@B8-ijerph-12-00488],[@B9-ijerph-12-00488]\\], metabolic syndrome \\[[@B10-ijerph-12-00488]\\] and glucose metabolism \\[[@B11-ijerph-12-00488]\\] which share many common risk factors with hypertension. Those findings suggested a possible link between sleep quality and hypertension which, however, has not been fully elucidated in the current literature. Therefore, it is necessary to investigate the potential relationship between prevalence of hypertension and global sleep status which contains both quantitative and qualitative aspects.\n\nPittsburgh Sleep Quality Index (PSQI) is a widely used tool for measurement of global sleep status and has acceptable internal homogeneity, test-retest reliability, and validity for clinical practice and research \\[[@B12-ijerph-12-00488]\\]. In the current study, we investigated the potential association between prevalence of hypertension and global sleep status assessed by PSQI using the data of a cross-sectional survey in Chinese adults.\n\n2. Study Design and Population\n==============================\n\nThis cross-sectional study was conducted from September to December 2013 in the Kailuan community which is located in Tangshan, a northern city of China. The Kailuan community is a functional and comprehensive community which contains ten sub-communities owned and managed by the Kailuan Group. More information about the Kailuan community can be found elsewhere \\[[@B13-ijerph-12-00488],[@B14-ijerph-12-00488]\\]. In the present study, four sub-communities, Tangshan, Fangezhuang, Lvjiatuo and Qianjiaying, were selected randomly from the Kailuan community. Subjects aged 18 years or more in those four sub-communities were invited to participate into this study. Critical exclusion criteria included: diagnosed or suspected secondary hypertension; hypertension in pregnancy; severe chronic heart failure; severe liver dysfunction; end-stage renal disease; advanced cancer; and previous diagnosis of obstructive sleep apnea syndrome (OSAS) or restless legs syndrome (RLS). In addition, subjects with a score of 1 or more on the item 10 of the PSQI, which indicated a high likelihood of comorbidity of OSAS, were excluded from the final analysis after enrollment. The contents and purposes of this study were thoroughly explained to the participants prior to the initiation of the study, and written consents were obtained from all participants.\n\nThe study protocols were in accordance with the Declaration of Helsinki and the ethical approval was obtained from the Science and Technology Committee of Tangshan City.\n\n3. Methods and Measurements\n===========================\n\n3.1. Anthropometric Measurements\n--------------------------------\n\nHeight and weight was measured to the nearest 0.1 cm or 0.1 kg when the subjects stood upright and barefoot in light clothes. Two separate measurements of height and weight for each subject were performed and the average was used for analysis. Body mass index (BMI) was calculated as the ratio of weight to height squared (kg/m^2^). Blood pressure was measured twice using standard mercury sphygmomanometers (Yuyue, China) in a seated position with a 5 min interval after a resting period of 10 min. Average of the two measurements was recorded as the final blood pressure. However, when the systolic or diastolic pressures exhibited a difference greater than 5 mmHg, a third measurement was necessary and the final blood pressure value was recorded as the average of the three measurements. Hypertension was defined according to the 7th edition report of the USA Joint National Committee on Prevention, Detection, Evaluation and Treatment of Hypertension \\[[@B15-ijerph-12-00488]\\]: as SBP \u2265 140 mmHg and/or DBP \u2265 90 mmHg on average of measurements or by current antihypertensive treatment.\n\n3.2. Blood Test\n---------------\n\nSubjects were asked to fast overnight before venous blood sample collection. Blood was collected from antecubital veins and then centrifuged at 3000 rpm for 10 min to isolate plasma. The supernatant serum were tested within 4 h in the central laboratory of Kailuan Hospital on automatic biochemical analyzers (Hitachi 717, Tyoko, Japan) for concentrations of total triglyceride (TG), total cholesterol (TC) and fasting blood glucose (FBG). Kit was provided by the Biology Institute of North China.\n\n3.3. Questionnaire Survey\n-------------------------\n\nA questionnaire survey was conducted face to face on paper to obtain demographic and behavior-associated information such as age, gender, status of smoking, status of drinking, and physical exercise habits. Status of smoking and drinking was classified as \"never\", \"former\", or \"current\" using the self-reported information. Exercise activity was evaluated from responses to questions about the type and frequency of physical activity during leisure time and was classified as \"active\" and \"inactive\". Subjects with at least 30 min physical activities for at least 5 days per week were defined as active exercisers. Sleep status of the participants was assessed by the Chinese version of the Pittsburgh Sleep Quality Index (PSQI) scale. PSQI scale consists of seven elements (subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleep medication, and daytime dysfunction) and each of them consisted of a four-grade system (*i.e*., 0, 1, 2, 3) \\[[@B12-ijerph-12-00488]\\]. The total score of PSQI is 21 and any participant with a score of 6 or more was diagnosed as having a sleep disorder. According the study of Tsai*et al.* \\[[@B16-ijerph-12-00488]\\], the Chinese version of the PSQI has good overall reliability (r = 0.82--0.83) and test--retest reliability (r = 0.77--0.85).\n\nIn addition, considering the frequent comorbidity of sleep disorders with anxiety and depression which also exert an influence on the prevalence of hypertension, status of anxiety and depression of participants were evaluated by the General Anxiety Disorder-7 (GAD-7) and Patient Health Questionnaire-9 (PHQ-9), respectively. GAD-7 is a seven-question inventory for self-assessment and is one of the most common instruments for measuring the severity of anxiety \\[[@B17-ijerph-12-00488]\\]. PHQ-9 is a widely used nine-question inventory for self-assessment of depression \\[[@B18-ijerph-12-00488]\\]. Both of the cut-off points to define anxiety and depression were \u22655 on GAD-7 or PHQ-9 \\[[@B19-ijerph-12-00488],[@B20-ijerph-12-00488]\\].\n\n4. Statistics\n=============\n\nAll statistical analyses were conducted separately by gender because significant interaction of gender and sleep status on hypertension prevalence was detected (F = 3.48, *p* \\< 0.05). Moreover, previous reports indicated that the associations of sleep in relation to hypertension varied with the ages of participants \\[[@B2-ijerph-12-00488],[@B3-ijerph-12-00488],[@B21-ijerph-12-00488]\\]; we therefore divided the participants into two groups with the age of 45 years as the cut-off value based on previous reports \\[[@B2-ijerph-12-00488]\\] and the age range of the present study (18 years to 72 years, with the average of 46.2). In addition, significant interaction between the two age classes for prevalence of hypertension was also observed in the present study (F = 2.19, *p* \\< 0.05). Continuous variables were presented as mean \u00b1 standard deviation (SD) and the comparisons between two groups were tested by *t*-tests or non-parametric tests based on distributional properties. Categorical variables were presented as frequencies and proportions and the comparisons between two groups were conducted via chi-square (\u03c7^2^) test. To explore the potential association of sleep status as well as its seven components in relation to hypertension prevalence, multivariate logistic analysis was used and other cardiovascular characteristics including age, body mass index (BMI), triglycerides (TG), total cholesterol (TC), fasting blood glucose (FBG), physical activity, status of smoking, status of drinking, score of GAD-7 and PHQ-9 were adjusted for. Considering that the limited number of subjects with the grade of 3 on the components of PSQI was not enough for multivariate logistic regression where a couple of confounders were adjusted, subjects with grades of 2 and 3 were merged. To investigate the specific association of sleep duration with hypertension prevalence, the total scores of the other six components of PSQI were further adjusted for on the basis of the adjustment of cardiovascular characteristics. For all the comparisons, the level of statistical significance was set at *p* \\< 0.05. All statistical analyses were conducted by SPSS 19.0 (IBM, Armonk, NY, USA).\n\n5. Results\n==========\n\nThere were 5980 citizens invited for the current study and 519 of them were precluded due to missing data. A total of 4076 males and 1385 females were finally enrolled into the current study. Male/female subjects were divided into two groups according to age \u226445 or \\>45 years. The basic characteristics of those subjects with or without hypertension are presented in [Table 1](#ijerph-12-00488-t001){ref-type=\"table\"}.\n\n5.1. Scores of Global PSQI and Its Components in Hypertensive and Normotensive Subjects\n---------------------------------------------------------------------------------------\n\nScores of global PSQI and each component of it in subjects with or without hypertension are presented in [Table 2](#ijerph-12-00488-t002){ref-type=\"table\"}. In male subjects aged \u226445 and \\>45 years old, hypertensive ones received significantly higher scores of global PSQI (4.30 \u00b1 0.12*vs.*3.30 \u00b1 0.07; 5.44 \u00b1 0.24*vs.* 4.16 \u00b1 0.11) and had a higher prevalence of sleep disorder (34.5%*vs.* 23.7%; 32.4%*vs.* 18.2%) than normotensive men. Out of the seven components of PSQI in males aged \u226445 years, hypertensive subjects had higher scores on six of them with the exception of sleep efficiency as compared with normotensive subjects. However, in male subjects aged \\>45 years, hypertensive men received higher scores on only five of the PSQI components and no significant statistical differences were found as to the scores of sleep efficiency and sleep latency between subjects with and without hypertension.\n\nFor female subjects aged \u226445 years, hypertensive women had significantly higher scores of global PSQI (5.73 \u00b1 0.42*vs.* 4.52 \u00b1 0.26) and higher prevalence of sleep disorders (54.3%*vs.* 29.6%) than normotensive women, which was similar to their male counterparts. In this group of participants, however, subjects with hypertension received higher scores on only one of the seven components of PSQI, the subjective sleep quality, in comparison to normotensive women. An obvious discrepancy for female subjects aged \\>45 years was observed between them and their male counterparts. With the exception of higher prevalence of sleep disorders in hypertensive subjects, no significant differences between hypertensive and normotensive subjects were found with regards to the scores of global PSQI and its seven components.\n\nijerph-12-00488-t001_Table 1\n\n###### \n\nThe basic characteristics of male and female participants.\n\n  Basic Cardiovascular Characteristics   \u226445 Years (N = 1934)   \\>45 years (N =2142)                                                                      \n  -------------------------------------- ---------------------- ---------------------- ------- ---------------- ---------------- ---------------- ------- ----------------\n  **Males**                                                                                                                                               \n                                         **N = 1433**           **N = 501**            **-**   **N = 1934**     **N = 1530**     **N = 612**      **-**   **N = 2142**\n  Age (year)                             39.45 \u00b1 5.15           40.62 \u00b1 4.89           0.00    39.76 \u00b1 5.08     50.91 \u00b1 9.26     51.29 \u00b1 3.98     0.27    51.43 \u00b1 8.03\n  BMI (kg/m^2^)                          25.32 \u00b1 3.88           26.21 \u00b1 3.87           0.00    25.54 \u00b1 3.90     24.85 \u00b1 3.36     25.77 \u00b1 3.41     0.00    25.13 \u00b1 3.40\n  SBP (mmHg)                             125.41 \u00b1 13.02         132.77 \u00b1 13.39         0.00    127.26 \u00b1 13.51   129.48 \u00b1 14.61   135.91 \u00b1 14.64   0.00    131.62 \u00b1 14.84\n  DBP (mmHg)                             83.33 \u00b1 8.85           89.26 \u00b1 9.20           0.00    84.76 \u00b1 9.31     85.06 \u00b1 9.41     89.78 \u00b1 10.11    0.00    86.47 \u00b1 9.83\n  TC (mmol/L)                            4.82 \u00b1 0.94            4.94 \u00b1 0.91            0.04    4.84 \u00b1 0.94      4.90 \u00b1 0.93      4.15 \u00b1 0.97      0.02    4.95 \u00b1 0.94\n  TG (mmol/L)                            1.95 \u00b1 2.02            2.19 \u00b1 2.38            0.03    2.01 \u00b1 2.11      1.85 \u00b1 1.70      2.11 \u00b1 2.18      0.00    1.91 \u00b1 1.79\n  FBG (mmol/L)                           5.19 \u00b1 1.05            5.38 \u00b1 1.40            0.01    5.23 \u00b1 1.15      5.54 \u00b1 1.52      5.69 \u00b1 1.75      0.03    5.58 \u00b1 1.62\n  Active exercise habits (%)             444 (31.0)             117 (23.3)             0.00    561 (29.0)       568 (37.1)       174 (28.5)       0.00    742 (34.3)\n  Current smoker (%)                     745 (52.0)             291 (58.0)             0.00    1,036 (53.5)     805 (52.6)       390 (63.8)       0.00    1,195 (55.8)\n  Current drinker (%)                    555 (38.7)             222 (44.4)             0.00    777 (40.2)       644 (42.1)       297 (48.6)       0.00    941 (43.9)\n  GAD-7                                  2.29 \u00b1 0.11            2.81 \u00b1 0.24            0.03    2.41 \u00b1 0.10      2.12 \u00b1 0.10      2.50 \u00b1 0.15      0.01    2.16 \u00b1 0.09\n  PHQ-9                                  2.55 \u00b1 0.12            3.50 \u00b1 0.27            0.00    2.78 \u00b1 0.11      2.10 \u00b1 0.10      2.62 \u00b1 0.06      0.01    2.26 \u00b1 0.09\n  **Females**                                                                                                                                             \n                                         **N = 695**            **N = 140**            **-**   **N = 835**      **N = 380**      **N = 170**      **-**   **N = 550**\n  Age (year)                             40.29 \u00b1 3.78           41.56 \u00b1 3.51           0.00    40.53 \u00b1 3.74     48.55 \u00b1 4.21     49.35 \u00b1 5.16     0.03    49.91 \u00b1 4.46\n  BMI (kg/m^2^)                          25.31 \u00b1 4.64           25.58 \u00b1 4.76           0.29    25.40 \u00b1 4.63     25.09 \u00b1 5.31     25.75 \u00b1 3.72     0.21    25.04 \u00b1 3.47\n  SBP (mmHg)                             127.03 \u00b1 17            130.85 \u00b1 13.26         0.00    127.60 \u00b1 16.59   123.72 \u00b1 16.86   130.22 \u00b1 16.97   0.00    126.21 \u00b1 16.84\n  DBP (mmHg)                             82.47 \u00b1 9.24           83.45 \u00b1 8.67           0.12    82.71 \u00b1 9.16     80.81 \u00b1 10.68    85.61 \u00b1 12.45    0.00    82.46 \u00b1 11.35\n  TC(mmol/L)                             4.88 \u00b1 0.90            5.00 \u00b1 0.79            0.13    4.90 \u00b1 0.89      5.00 \u00b1 0.93      5.04 \u00b1 1.01      0.52    4.98 \u00b1 0.96\n  TG (mmol/L)                            1.80 \u00b1 1.44            1.72 \u00b1 1.04            0.28    1.86 \u00b1 1.59      1.96 \u00b1 1.58      1.78 \u00b1 1.30      0.26    1.88 \u00b1 1.34\n  FBG (mmol/L)                           5.65 \u00b1 1.99            5.14 \u00b1 0.91            0.02    5.56 \u00b1 1.86      5.43 \u00b1 0.84      5.28 \u00b1 1.08      0.14    5.42 \u00b1 1.11\n  Active exercise habits (%)             221 (31.8)             54 (38.3)              0.12    275 (32.9)       144 (37.9)       64 (37.6)        0.96    208 (37.8)\n  Current smoker (%)                     10 (1.5)               6 (4.0)                0.00    16 (1.9)         5 (1.4)          14 (5.1)         0.00    19 (3.5)\n  Current drinker (%)                    166 (23.9)             39 (28.0)              0.33    205 (24.6)       105 (27.6)       52 (30.7)        0.50    274 (49.8)\n  GAD-7                                  3.77 \u00b1 0.44            4.39 \u00b1 0.57            0.00    3.86 \u00b1 0.41      2.38 \u00b1 0.65      3.91 \u00b1 0.70      0.00    3.37 \u00b1 0.51\n  PHQ-9                                  3.91 \u00b1 0.46            6.33 \u00b1 0.52            0.00    4.23 \u00b1 0.47      3.30 \u00b1 0.69      3.74 \u00b1 0.83      0.22    3.57 \u00b1 0.57\n\nBMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol; TG, triglycerides; FBG, fasting blood glucose; GAD-7, General Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire.\n\nijerph-12-00488-t002_Table 2\n\n###### \n\nScores of global PSQI and its component in both male and female subjects with and without hypertension.\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  PSQI and Its Components     \u226445 Years (N= 1934)   \\>45 years (N= 2142)                                                                               \n  --------------------------- --------------------- ---------------------- ---------- ----------------- ----------------- ----------------- ---------- -----------------\n  **Male**                                                                                                                                             \n\n  Global PSQI score           **3.30 \u00b1 0.07**       **4.30 \u00b1 0.12**        **0.00**   **3.61 \u00b1 0.06**   **4.16 \u00b1 0.11**   **5.44 \u00b1 0.24**   **0.00**   **4.46 \u00b1 0.11**\n\n  Sleep disorder (n, %)\\      **340 (23.7)**        **173 (34.5)**         **0.00**   **513 (26.5)**    **278 (18.2)**    **198 (32.4)**    **0.00**   **476 (22.2)**\n  (PSQI score \u2265 6)                                                                                                                                     \n\n  Subjective sleep quality    **0.53 \u00b1 0.02**       **0.74 \u00b1 0.03**        **0.00**   **0.60 \u00b1 0.02**   **0.63 \u00b1 0.02**   **0.76 \u00b1 0.05**   **0.01**   **0.66 \u00b1 0.02**\n\n  Sleep latency               **0.23 \u00b1 0.01**       **0.37 \u00b1 0.03**        **0.00**   **0.27 \u00b1 0.01**   0.85 \u00b1 0.04       1.09 \u00b1 0.07       0.52       0.90 \u00b1 0.03\n\n  Sleep duration              **0.41 \u00b1 0.02**       **0.57 \u00b1 0.03**        **0.00**   **0.46 \u00b1 0.01**   **0.39 \u00b1 0.02**   **0.53 \u00b1 0.04**   **0.01**   **0.43 \u00b1 0.02**\n\n  Habitual sleep efficiency   0.40 \u00b1 0.02           0.39 \u00b1 0.04            0.69       0.40 \u00b1 0.02       0.36 \u00b1 0.02       0.39 \u00b1 0.04       0.46       0.36 \u00b1 0.02\n\n  Sleep disturbance           **0.56 \u00b1 0.02**       **0.79 \u00b1 0.02**        **0.00**   **0.63 \u00b1 0.01**   **0.59 \u00b1 0.02**   **0.82 \u00b1 0.04**   **0.00**   **0.65 \u00b1 0.02**\n\n  Use of sleep medication     **0.74 \u00b1 0.02**       **0.89 \u00b1 0.03**        **0.00**   **0.79 \u00b1 0.01**   **0.76 \u00b1 0.02**   **0.92 \u00b1 0.04**   **0.00**   **0.80 \u00b1 0.02**\n\n  Daytime dysfunction         **0.39 \u00b1 0.02**       **0.53 \u00b1 0.03**        **0.00**   **0.43 \u00b1 0.01**   **0.75 \u00b1 0.03**   **0.98 \u00b1 0.07**   **0.00**   **0.81 \u00b1 0.03**\n\n  **Female**                                                                                                                                           \n\n  Global PSQI score           **4.52 \u00b1 0.26**       **5.73 \u00b1 0.42**        **0.01**   **4.88 \u00b1 0.22**   4.78 \u00b1 0.23       5.12 \u00b1 0.37       0.44       4.87 \u00b1 0.19\n\n  Sleep disorder (n, %)\\      **206 (29.6)**        **76 (54.3)**          **0.00**   **282 (33.8)**    **115 (30.2)**    **61 (35.9)**     **0.00**   **176 (32.0)**\n  (PSQI score \u2265 6)                                                                                                                                     \n\n  Subjective sleep quality    **0.92 \u00b1 0.05**       **1.11 \u00b1 0.08**        **0.03**   **0.98 \u00b1 0.04**   0.93 \u00b1 0.06       1.14 \u00b1 0.11       0.07       0.99 \u00b1 0.05\n\n  Sleep latency               0.71 \u00b1 0.04           0.85 \u00b1 0.07            0.07       0.75 \u00b1 0.04       0.78 \u00b1 0.05       0.93 \u00b1 0.09       0.13       0.82 \u00b1 0.04\n\n  Sleep duration              0.35 \u00b1 0.04           0.38 \u00b1 0.06            0.64       0.36 \u00b1 0.03       0.39 \u00b1 0.05       0.51 \u00b1 0.09       0.19       0.42 \u00b1 0.04\n\n  Habitual sleep efficiency   0.33 \u00b1 0.04           0.33 \u00b1 0.06            0.99       0.33 \u00b1 0.03       0.51 \u00b1 0.06       0.31 \u00b1 0.08       0.07       0.46 \u00b1 0.05\n\n  Sleep disturbance           0.35 \u00b1 0.14           0.55 \u00b1 0.20            0.43       0.41 \u00b1 0.11       0.41 \u00b1 0.04       0.23 \u00b1 0.12       0.06       0.36 \u00b1 0.04\n\n  Use of sleep medication     0.89 \u00b1 0.04           0.95 \u00b1 0.06            0.42       0.91 \u00b1 0.03       0.96 \u00b1 0.04       1.06 \u00b1 0.08       0.25       0.98 \u00b1 0.04\n\n  Daytime dysfunction         **0.83 \u00b1 0.04**       **1.00 \u00b1 0.06**        **0.02**   **0.88 \u00b1 0.03**   0.80 \u00b1 0.05       0.94 \u00b1 0.08       0.16       0.84 \u00b1 0.04\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nPSQI, Pittsburgh Sleep Quality Index.\n\n5.2. The Association between Sleep Status and Hypertension Prevalence\n---------------------------------------------------------------------\n\nThe results of multivariate logistic regression analyses about the association between sleep status (global PSQI score and its components) and hypertension prevalence are presented in [Table 3](#ijerph-12-00488-t003){ref-type=\"table\"} (for males) and [Table 4](#ijerph-12-00488-t004){ref-type=\"table\"} (for females). After adjusting for age, BMI, TC, TG, FBG, physical exercise, status of smoking and drinking, scores of GAD-7 and PHQ-9, global PSQI score was found to be associated with the prevalence of hypertension in male subjects aged \u226445 and \\>45 years and the corresponding odds ratio (OR) was 1.11 (95% confidence interval (CI), 1.07--1.15) and 1.12 (95% CI, 1.08--1.15), respectively. Sleep disorder was also found to be in relation to hypertension prevalence in male subjects aged \u226445 (OR = 1.79, 95% CI, 1.37--2.35) and \\>45 (OR = 1.86, 95% CI, 1.51--2.30). In addition, our results demonstrated that prevalence of hypertension was associated with five components of PSQI in male subjects of both age groups: subjective sleep quality, sleep latency, sleep duration, sleep disturbance, use of sleep medication and daytime dysfunction.\n\nFor female subjects aged \u226445 years, global PSQI score and sleep disorder were also significantly associated with the prevalence of hypertension (OR = 1.10, 95% CI, 1.02--1.18; OR = 1.84, 95% CI, 1.06--3.19), which was similar to their male counterparts. However, global PSQI score and sleep disorder were not associated with hypertension prevalence in female subjects aged \\>45 years. As for the association of components of PSQI in relation to hypertension prevalence in females, none of them was found to be statistically significant. In contrast to the result from their male counterparts, short sleep duration failed to be associated with hypertension prevalence in females aged either \u226445 years and \\>45 years.\n\n5.3. Association of Sleep Duration with Hypertension Prevalence after Adjustment of the Qualitative Aspect of Sleep\n-------------------------------------------------------------------------------------------------------------------\n\nTo preclude the potential interaction of sleep quality and sleep duration and investigate the specific association of sleep duration with prevalence of hypertension in male subjects, the scores of the other six components of PSQI which measured the qualitative aspect of sleep were further adjusted for addition to the basic cardiovascular characteristics in the multivariate logistic regression analysis. As shown in [Table 5](#ijerph-12-00488-t005){ref-type=\"table\"}, sleep duration was no longer significantly associated with hypertension prevalence in both age groups. The results were inconsistent with the results demonstrated in [Table 3](#ijerph-12-00488-t003){ref-type=\"table\"} where qualitative components of PSQI were not adjusted for.\n\nijerph-12-00488-t003_Table 3\n\n###### \n\nOdds ratios (OR) of scores of PSQI and its components for prevalence of hypertension in male subjects.\n\n  PSQI and Its Components           Grade     N                       \u226445 years (N= 1934)     N                       \\>45 years (N = 2142)                           \n  --------------------------------- --------- ----------------------- ----------------------- ----------------------- ----------------------- ----------------------- -----------------------\n  Global PSQI score                           **1934**                **1.11 (1.07--1.14)**   **1.11 (1.07--1.15)**   **2142**                **1.12 (1.09--1.15)**   **1.12 (1.08--1.15)**\n  Sleep disorder (PSQI score \u2265 6)   No        1421                    1.00 (Reference)        1.00 (Reference)        1666                    1.00 (Reference)        1.00 (Reference)\n  Yes                               **513**   **1.72 (1.32--2.30)**   **1.79 (1.37--2.35)**   **476**                 **1.95 (1.59--2.39)**   **1.86 (1.51--2.30)**   \n  Subjective sleep quality          0         1031                    1.00 (Reference)        1.00 (Reference)        1268                    1.00 (Reference)        1.00 (Reference)\n  1                                 619       1.04 (0.79--1.37)       1.07 (0.80--1.42)       553                     1.25 (1.02--1.52)       1.26 (1.02--1.54)       \n  2--3                              **284**   **1.61 (1.09--2.36)**   **1.62 (1.09--2.42)**   **321**                 **1.68 (1.28--2.21)**   **1.63 (1.23--2.17)**   \n  Sleep latency                     0         940                     1.00 (Reference)        1.00 (Reference)        1694                    1.00 (Reference)        1.00 (Reference)\n  1                                 777       1.11 (0.86--1.45)       1.13 (0.86--1.49)       334                     1.38 (1.10--1.73)       1.35 (1.06--1.71)       \n  2--3                              **217**   **2.15 (1.46--3.16)**   **2.28 (1.52--3.40)**   **114**                 **2.15 (1.52--3.11)**   **2.18 (1.50--3.18)**   \n  Sleep duration                    0         1,298                   1.00 (Reference)        1.00 (Reference)        1403                    1.00 (Reference)        1.00 (Reference)\n  1                                 487       1.13 (0.85--1.50)       1.18 (0.88--1.58)       536                     1.24 (1.02--1.51)       1.22 (0.99--1.50)       \n  2--3                              **149**   **2.50 (1.61--3.89)**   **2.62 (1.65--4.15)**   **203**                 **1.96 (1.44--2.67)**   **1.90 (1.38--2.62)**   \n  Habitual sleep efficiency         0         1509                    1.00 (Reference)        1.00 (Reference)        1555                    1.00 (Reference)        1.00 (Reference)\n  1                                 239       1.25 (0.87--1.80)       1.20 (0.82--1.76)       283                     1.09 (0.84--1.42)       1.12 (0.86--1.46)       \n  2--3                              186       1.00 (0.55--1.82)       1.05 (0.57--1.92)       304                     1.08 (0.76--1.54)       1.148 (0.80--1.64)      \n  Sleep disturbance                 0         837                     1.00 (Reference)        1.00 (Reference)        966                     1.00 (Reference)        1.00 (Reference)\n  1                                 **969**   **1.71 (1.31--2.24)**   **1.75 (1.32--2.31)**   **1035**                **2.04 (1.70--2.44)**   **2.01 (1.67--2.42)**   \n  2--3                              **128**   **3.30 (1.99--5.47)**   **3.53 (2.08--6.00)**   **141**                 **3.14 (2.22--4.46)**   **3.29 (2.29--4.74)**   \n  Use of sleep medication           0         698                     1.00 (Reference)        1.00 (Reference)        743                     1.00 (Reference)        1.00 (Reference)\n  1                                 **977**   **1.35 (1.03--1.77)**   **1.42 (1.07--1.88)**   **1148**                **1.35 (1.11--1.63)**   **1.34 (1.10--1.63)**   \n  2--3                              **259**   **1.66 (1.11--2.47)**   **1.82 (1.20--2.76)**   **251**                 **1.98 (1.48--2.65)**   **2.13 (1.58--2.87)**   \n  Daytime dysfunction               0         1135                    1.00 (Reference)        1.00 (Reference)        1401                    1.00 (Reference)        1.00 (Reference)\n  1                                 **594**   **1.33 (1.02--1.73)**   **1.38 (1.05--1.83)**   **568**                 **1.59 (1.31--1.93)**   **1.56 (1.28--1.90)**   \n  2--3                              **205**   **1.57 (1.06--2.32)**   **1.56 (1.03--2.36)**   **173**                 **1.66 (1.21--2.28)**   **1.60 (1.14--2.22)**   \n\nPSQI, Pittsburgh Sleep Quality Index. CI, confidence interval; **^\u2020^** Adjusting for age, BMI, TC,TG, FBG, exercise habit, status of smoking, status of drinking, score of GAD-7 and PHQ-9.\n\nijerph-12-00488-t004_Table 4\n\n###### \n\nOdds ratios (OR) of scores of PSQI and its components for prevalence of hypertension in female subjects.\n\n  PSQI and Its Components     Grade   N                       \u226445 years (N = 835)     N                       \\>45 years (N = 550)                       \n  --------------------------- ------- ----------------------- ----------------------- ----------------------- ---------------------- ------------------- -------------------\n  Global PSQI score                   **835**                 0.98 (0.88--1.10)       1.10 (1.02--1.18)       550                    0.93 (0.75--1.17)   1.03 (0.96--1.1)\n  Sleep disorder              No      **529**                 1.00 (Reference)        1.00 (Reference)        332                    1.00 (Reference)    1.00 (Reference)\n  (PSQI score \u2265 6)            Yes     **306**                 **1.76 (1.17--2.61)**   **1.84 (1.06--3.19)**   218                    1.34 (0.43--2.27)   1.36 (0.80--2.30)\n  Subjective sleep quality    0       256                     1.00 (Reference)        1.00 (Reference)        188                    1.00 (Reference)    1.00 (Reference)\n  1                           337     **1.22 (1.02--1.46)**   1.24 (0.78--1.95)       207                     1.42 (0.74--2.59)      1.39 (0.77--2.54)   \n  2--3                        242     1.30 (0.67--2.51)       1.32 (0.68--2.54)       155                     1.47 (0.45--4.95)      1.47 (0.69--3.11)   \n  Sleep latency               0       209                     1.00 (Reference)        1.00 (Reference)        233                    1.00 (Reference)    1.00 (Reference)\n  1                           232     3.72 (0.51--27.90)      3.72 (0.46--30.13)      191                     1.40 (0.70--2.81)      1.43 (0.81--2.50)   \n  2--3                        394     4.11 (0.49--35.87)      4.11 (0.48--34.84)      126                     1.53 (0.50--4.75)      1.55 (0.67--3.56)   \n  Sleep duration              0       625                     1.00 (Reference)        1.00 (Reference)        396                    1.00 (Reference)    1.00 (Reference)\n  1                           144     1.38 (0.52--3.67)       1.37 (0.83--2.26)       116                     1.64 (1.21--2.15)      1.68 (0.94--3.03)   \n  2--3                        66      0.95 (0.43--2.12)       0.97 (0.39--2.43)       48                      0.91 (0.24--2.86)      0.95 (0.30--3.01)   \n  Habitual sleep efficiency   0       670                     1.00 (Reference)        1.00 (Reference)        397                    1.00 (Reference)    1.00 (Reference)\n  1                           90      0.90 (0.19--4.35)       0.83 (0.44--1.60)       69                      0.70 (0.31--1.55)      0.64 (0.28--1.46)   \n  2--3                        75      1.64 (0.69--3.46)       1.66 (0.70--3.90)       84                      0.56 (0.02--1.02)      0.50 (0.16--1.50)   \n  Sleep disturbance           0       270                     1.00 (Reference)        1.00 (Reference)        119                    1.00 (Reference)    1.00 (Reference)\n  1                           422     1.12 (0.66--1.17)       1.14 (0.59--2.21)       331                     1.02 (0.24--3.56)      1.00 (0.29--3.44)   \n  2--3                        143     1.50 (0.50--4.53)       1.46 (0.46--4.60)       100                     1.82 (1.06--3.19)      1.78 (0.38--8.33)   \n  Use of sleep medication     0       221                     1.00 (Reference)        1.00 (Reference)        150                    1.00 (Reference)    1.00 (Reference)\n  1                           428     1.16 (0.68--1.32)       1.14 (0.72--1.81)       254                     0.95 (0.28--2.95)      0.97 (0.51--1.84)   \n  2--3                        186     1.58 (0.86--4.12)       1.51 (0.82--2.79)       146                     1.41 (0.69--2.80)      1.44 (0.66--3.16)   \n  Daytime dysfunction         0       328                     1.00 (Reference)        1.00 (Reference)        191                    1.00 (Reference)    1.00 (Reference)\n  1                           372     1.24 (0.74--1.89)       1.26 (0.80--2.00)       264                     1.96 (1.62--2.52)      1.94 (0.99--3.82)   \n  2--3                        135     1.75 (0.97--3.16)       1.79 (1.00--3.19)       95                      1.54 (0.55--4.97)      1.50 (0.67--3.36)   \n\nPSQI, Pittsburgh Sleep Quality Index. CI, confidence interval; **^\u2020^** Adjusting for age, BMI, TC,TG, FBG, exercise habit, status of smoking, status of drinking, score of GAD-7 and PHQ-9.\n\nijerph-12-00488-t005_Table 5\n\n###### \n\nOdds ratios (OR) of sleep duration for hypertension prevalence after adjusting for the components of PSQI measuring qualitative aspects of sleep.\n\n  PSQI and Its Components   Grade of Sleep Duration   N                   Unadjusted OR (95% CI)   Adjusted OR(95% CI) ^\u2020^\n  ------------------------- ------------------------- ------------------- ------------------------ -------------------------\n  Males \u2264 45 years          0                         1298                1.00 (Reference)         1.00 (Reference)\n  1                         487                       1.01 (0.75--1.36)   0.96 (0.69--1.34)        \n  2--3                      149                       1.64 (0.97--2.48)   1.60 (0.90--2.82)        \n  Males \\> 45 years         0                         1403                1.00 (Reference)         1.00 (Reference)\n  1                         536                       1.05 (0.98--1.07)   1.03 (0.82--1.29)        \n  2--3                      203                       1.34 (0.91--1.95)   1.29 (0.89--1.86)        \n\nPSQI, Pittsburgh Sleep Quality Index. CI, confidence interval; **^\u2020^** Adjusting for age, BMI, TC,TG, FBG, exercise habit, status of smoking, status of drinking, score of GAD-7, score of PHQ-9 and the total scores of six components of PSQI including subjective sleep quality, sleep latency, habitual sleep efficiency, sleep disturbance, use of sleep medication and daytime function.\n\n6. Discussion\n=============\n\nTaking both sleep quantity and quality into consideration, we used PSQI to assess the global sleep status in the present study and explored the potential association of PSQI score as well as its components with hypertension prevalence in Chinese adults. Our results demonstrated that the associations of global sleep status as well as short sleep duration with hypertension prevalence varied with age and sex. Poor sleep status was associated with hypertension prevalence in male subjects of all ages and in female subjects aged \u226445 years. Short sleep duration was found to be related to hypertension only in male subjects. However, after adjusting for the sleep qualitative aspect of PSQI on the basis of the traditional cardiovascular characteristics, the relation between short sleep duration and hypertension failed to reach a statistical significance.\n\nOne of the main findings of the present study was that sleep should be measured qualitatively and quantitatively when investigating its potential association with hypertension, and previous reports about sleep duration and hypertension may be biased by sleep quality. Dozens of studies have investigated the relationship between sleep duration and hypertension prevalence or incidence in populations of different races, genders or ages \\[[@B1-ijerph-12-00488],[@B2-ijerph-12-00488],[@B3-ijerph-12-00488],[@B4-ijerph-12-00488],[@B5-ijerph-12-00488],[@B22-ijerph-12-00488],[@B23-ijerph-12-00488],[@B24-ijerph-12-00488]\\]. Most of these studies suggest a robust relationship between sleep duration and hypertension although obvious contradictions can be observed among them. To our knowledge, the relationship between sleep duration and hypertension has never been elucidated in Chinese adults before. In the present study, male and female subjects were divided into two age groups with the age of 45 years as the demarcation point. We found that short sleep duration was associated with hypertension only in men. Our results that short sleep duration was not in relation to hypertension prevalence in women were supported by a couple of previous studies conducted in the American, Japanese, Dutch, Brazilian and Spanish populations \\[[@B25-ijerph-12-00488],[@B26-ijerph-12-00488],[@B27-ijerph-12-00488],[@B28-ijerph-12-00488]\\]. However, some results from other studies were contradictory to ours \\[[@B3-ijerph-12-00488],[@B21-ijerph-12-00488],[@B23-ijerph-12-00488],[@B29-ijerph-12-00488]\\]. Although differences in the basic characteristics of the enrolled participants and the confounder adjusted in those studies partly explains the inconsistent associations between sleep duration and hypertension prevalence in the current literature, failing to preclude the potential confounding effects of sleep quality on hypertension may be one essential underlying reason for the discrepancy. Sleep has both quantitative and qualitative aspects. Most studies about sleep and hypertension only measured sleep duration which, in our opinion, is not comprehensive. Two previous studies which paid attention to the role of sleep quality demonstrated that examining only short sleep duration without the presentation of other forms of sleep disorders failed to exert an influence on hypertension prevalence \\[[@B7-ijerph-12-00488],[@B30-ijerph-12-00488]\\]. In fact, short sleep duration is often accompanied by poor sleep quality. The analysis of the subjects enrolled in the present study has shown that the percentage of sleepers with poor subjective sleep quality kept increasing when sleep duration decreased from 8 h to less than 6 h ([Supplementary Figure S1](#ijerph-12-00488-s001){ref-type=\"supplementary-material\"}: the percentages of high-quality sleep in subjects with sleep duration of 8 h, 7 h, 6 h and less than 6 h in all participants). Furthermore, results from this study also indicated that poor sleep quality was also related to hypertension prevalence. These findings made the association of short sleep duration in relation to hypertension from previous studies become less reliable. To ensure the role of short sleep duration in hypertension prevalence, we believe it is necessary to adjust for the confounding effect of sleep quality. We made this adjustment in the present study and the significant association between sleep duration and hypertension before the adjustment of the qualitative aspect of sleep failed to reach a statistical significance.\n\nAnother finding of the present study was the association between global sleep status and hypertension in females: sleep status was not associated with hypertension prevalence in those aged more than 45 years. Although the potential mechanism under such a discrepancy is still unknown to us at the present time, it may result from the changes of endocrine hormones, as those aged \\>45 years were close to or already in the perimenopausal period and were therefore more vulnerable to major hormonal fluctuations \\[[@B31-ijerph-12-00488],[@B32-ijerph-12-00488]\\].\n\nOSAS and RLS could significantly increase the risk for hypertension \\[[@B33-ijerph-12-00488],[@B34-ijerph-12-00488]\\] and they are possible confounding factors in the relationship between sleep status and hypertension prevalence. To evaluate the specific effect of sleep status on hypertension prevalence, OSAS and RLS were excluded in this study. It has been reported that snoring has a high sensitivity (87%) for detecting OSAS in previous research \\[[@B35-ijerph-12-00488]\\]. Thus, we excluded all subjects with suspected OSAS who snored more than one day per week, as reported by himself or roommates. In addition, RLS was excluded on the basis of self-reported (or reported by roommates) symptoms, as the diagnosis of RLS was mainly based on clinical symptoms.\n\nOur study has several limitations. First of all, this study was designed as a cross-sectional one and did not allow the establishment of causality of the observed associations. Given the self-reporting of sleep status, we could neither preclude the possibility of the potential effect of hypertension on sleep habits and quality nor *vice versa*. Secondly, as a further limitation, because we did not have suitable questionnaires or other forms of measurement, we could not control the potential influence of eating habits on sleep disorders and hypertension, despite food being one important mediating factor in the relationship between them. Thirdly, although the sample size of the female participants was enough to yield sound and stable results, it is better to adopt a cautious attitude toward the sex differentiation for association of sleep quality and hypertension considering the high ratio of males to females in the current study.\n\n7. Conclusions\n==============\n\nThe association of global sleep status evaluated by the PSQI scale and hypertension prevalence in Chinese adults varies with age and sex. In addition, the interaction of sleep duration and sleep quality needs to be addressed when investigating the association between sleep duration and hypertension. Moreover, comprehensive evaluation of sleep from qualitative and quantitative aspects will achieve better accuracy.\n\nThis study was funded by grants from the 12th Five-Year Science and Technology Support Program of the Ministry of Science and Technology of China (Grant No. 2013BAI06B02). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n###### \n\nClick here for additional data file.\n\nDayi Hu and Rongjing Ding had the original idea for the study and, with all co-authors carried out the design. Kai Lu, Rongjing Ding, and Qin Tang were responsible for recruitment and follow-up of study participants. Shouling Wu was responsible for data cleaning and Kai Lu, Li Wang, and Changying Wang carried out the analyses. Jia Chen drafted the manuscript, which was revised by all authors. All authors read and approved the final manuscript.\n\nThe authors declare no conflict of interest.\n"}
+{"text": "1. Background {#sec82828}\n=============\n\nAndrogens are a class of natural or synthetic steroid hormone compounds such as testosterone, dihydrotestosterone and androstenedione ([@A18216R1]). Physiological androgens play vital roles in the regulation of anabolic metabolism and expression of male phenotype. They are actually eminent for regulation of male development particularly the male reproductive system ([@A18216R2]). The diverse effects of androgens are mediated by complex cell type-specific signaling pathways that always involve the androgen receptor (AR) ([@A18216R3]). The androgen receptor (AR) is a type of nuclear receptor that acts as a DNA-binding transcription factor to regulate gene expression ([@A18216R4]). Dysregulation of androgen/AR signaling perturbs normal reproductive development and accounts for a wide range of pathological conditions such as prostate cancer ([@A18216R2]). Recent experimental findings conform the expression of AR in androgen-dependent, androgen-independent or hormone refractory prostate cancers, where AR expression is maintained throughout prostate cancer progression ([@A18216R5]). Interestingly, a number of clinical studies suggest that prostate cancer initiation and progression is also uniquely dependent on AR expression ([@A18216R6]). This pathogenic role of AR has made it a major therapeutic target in aggressive prostate cancer ([@A18216R7]). Therefore, searching for AR inhibitors is clearly an important issue for biological research.\n\n*Hymenodictyon excelsum* is a medium-sized deciduous tree of the Rubiaceae family. It is locally known as \"Phuti-kadam\" and has an abundant distribution in Bangladesh and India. In traditional medicinal practice the bark of *H. excelsum*is used as an astringent and febrifuge and for treatment of fever and tumors, while the leaves are used to treat ulcers, sialitis, sore throat, tonsillitis and inflammatory conditions ([@A18216R8]). The stem bark of *H. excelsum* contains tannin, toxic alkaloid, hymenodictine, esculin, an apioglucoside of scopoletin andhymexelsin ([@A18216R9]). Anthraquinones, rubiadin and its methyl ether, lucidin, nordamnacanthal, damnacanthal, 2-benzylzanthopurpurin, anthragallol, soranjidiol and morindone have also been isolated from roots ([@A18216R10]). So far, little is known about the pharmacological activity of this plant. The plant has been reported to have antioxidant and anti-inflammatory properties ([@A18216R11]).\n\n2. Objectives {#sec82829}\n=============\n\nThe present study aimed to investigate *H. excelsum's*phytochemicals with anti-prostate cancer activity and to predict their mechanisms of action using a computational molecular simulation study.\n\n3. Materials and Methods {#sec82836}\n========================\n\n3.1. Selection of Receptor and Ligands {#sec82830}\n--------------------------------------\n\nThe therapeutic effects of a medicinal plant are attributed to the phytochemical constituents of that plant. Therefore, it is feasible to study individual phytochemicals present in plants when investigating the therapeutic effects of medicinal plants. From the literature search, eight phytochemicals were taken into consideration for the molecular simulation study including anthragallol, damnacanthal, esculin, lucidin, morindone, nordamnacanthal, rubiadin and soranjidiol. Besides, dihydrotestosterone was considered as the control. Canonical SMILES file of these selected ligand molecules were retrieved from the PubChem Compound server (https://pubchem.ncbi.nlm.nih.gov/). Canonical SMILES files were entered in to the Corina online demo to generate 3D Protein Data Bank (PDB) files. The human androgen receptor was the receptor used in the present study. Protein Data Bank (PDB) file of human AR ligand binding domain (PDB id: 1e3g) was downloaded from RCSB PDB database (http://www.rcsb.org/pdb/home/home.do).\n\n3.2. Drug likeness and Toxicity Evaluation of the Ligands {#sec82831}\n---------------------------------------------------------\n\nPrior to molecular simulation study the drug likeness and safety of the chosen ligands were evolved by empirical computational tools. The Molinspiration cheminformatics server was exploited to calculate the molecular properties associated with drug-likeness and to predict bioactivity of components including the nuclear receptor ligand and enzyme inhibitor ([@A18216R12]). The OSIRIS property explorer was used to evolve the toxicity of the selected ligands in term of mutagenic, tumorigenic and irritant properties ([@A18216R13]). Only safe drug-like phytochemicals were subjected to the molecular simulation study.\n\n3.3. Computational Simulation Study {#sec82835}\n-----------------------------------\n\n### 3.3.1. Binding Site Prediction {#sec82832}\n\nThe site of the receptor that interacted with the ligands was predicted by the Molegro Virtual Docker (MDV) inbuilt cavity detection algorithm and the Pocket-Finder server ([@A18216R14]). The outputs from both of the computational tools were synchronized to predict the ultimate ligand binding site and exploited in the subsequent docking study.\n\n### 3.3.2. Computational Simulation Tools and Algorithm {#sec82833}\n\nMolegro Virtual Docker (MVD) was used for the computer simulated docking study ([@A18216R15]). This tool performs flexible ligand docking with the optimization of ligand geometry during docking. Molegro Virtual Docker includes MolDock and PLANTS score for evaluating docking solutions followed by a rank for the best conformations. The MVD uses a differential evolution algorithm. The solution to the function is the sum of intermolecular interaction energy between protein and ligand with the intra-molecular interaction energy of the ligand. The docking energy scoring function is based on a modified piecewise linear potential (PLP) with new hydrogen bonding and electrostatics terms included. A lower score always indicates a higher affinity.\n\n### 3.3.3. Molecular Docking Study {#sec82834}\n\nBefore initiation of the docking operation, both protein and ligands were prepared by MVD through assigning bonds, bond orders, explicit hydrogens, charges and flexible torsions at the missing region only. Potential binding sites (cavities or active sites) were identified using the built-in cavity detection algorithm of MVD with a grid resolution (\u00c5) of 0.30. Internal electrostatic interactions, internal hydrogen bonds, and Sp2-Sp2 torsions were selected as ligand evaluation terms. The \"MolDock SE\" was set as the searching algorithm for ten runs using a maximum of 1500 iterations with a total population size of 50. Energy minimization and optimized H-bonds were enabled as post-docking steps for more accurate refinement of the docking results. Multiple poses were clustered based on the RMSD threshold of 1.0, ignoring similar poses with RMSD threshold of 1.0.\n\n4. Results {#sec82837}\n==========\n\n4.1 Result of Drug Likeness and Toxicity Evaluation of the Ligands {#sec82841}\n------------------------------------------------------------------\n\nResult of the drug likeness and toxicity evaluation of the ligand have been summarized in [Table 1](#tbl21040){ref-type=\"table\"}. The results from the Molinspiration cheminformatics server showed that all the selected ligands met the Lipinski\\'s rule of drug likeness without violation of any condition. Additionally, all the ligands fitted more or less with nuclear receptor ligands and enzyme inhibitors. The output from the OSIRIS property explorer suggested the esculin, lucidin, morindone and soranjidiol were safe ligands with no mutagenic or tumorigenic effects. Thus, the probable phytochemicals were filtered, as they were more potent physiologically active molecules without any harmless effects.\n\n###### Drug Likeness and Toxicity Evaluation of the Selected Ligands ^[a](#fn18763){ref-type=\"table-fn\"}^\n\n  Ligand                miLogP   TPSA    NV   NRL    EI     Mutagenic   Tumorigenic   Irritant\n  --------------------- -------- ------- ---- ------ ------ ----------- ------------- ----------\n  **Anthragallol**      2.61     94.83   0    0.01   0.29   Y           N             N\n  **Damnacanthal**      3.34     80.68   0    0.10   0.14   N           N             Y\n  **Esculin**           3.34     80.68   0    0.10   0.14   N           N             Y\n  **Lucidin**           2.61     94.83   0    0.19   0.38   N           N             N\n  **Morindone**         3.04     94.83   0    0.07   0.20   N           N             N\n  **Nordamnacanthal**   3.06     91.67   0    0.15   0.22   Y           N             Y\n  **Rubiadin**          3.72     74.60   0    0.07   0.24   Y           N             N\n  **Soranjidiol**       3.30     74.60   0    0.11   0.20   N           N             N\n\n^a^ Abbreviations: EI, enzyme inhibitor; miLogP, octanol-water partition coefficient developed at Molinspiration; N, No; NV, number of violation of Lipinski\\'s rule of drug likeness; TPSA, total molecular polar surface area; NRL, nuclear receptor ligand; Y, yes.\n\n4.2. Computational Simulation Study {#sec82840}\n-----------------------------------\n\n### 4.2.1. Binding Site Prediction {#sec82838}\n\nInitial prediction of the probable ligand-binding site was performed by the Pocket-Finder and MDV Cavity Search. The prediction showed an active site of 409 cubic angstroms with a maximum coordination of 8 \u00d7 42 \u00d7 12 from the Pocket-Finder output ([Figure 1](#fig16040){ref-type=\"fig\"}). The cavity searching algorithm built in the MVD tool fetched a cavity of 86.02 cubic angstroms with a maximum coordination of 0.41 \u00d7 31.67 \u00d7 4.44 and surface of 204.8 square angstroms. Output of the Pocket-Finder was synchronized with the MDV Cavity Search output for screening the prediction with higher degree of accuracy. As, these two outputs overlapped and the selected molecule had a surface of \\~95 square angstroms at best, the cavity coordination of 0.41 \u00d7 31.67 \u00d7 4.44 was set as the desired ligand interaction site or binding site.\n\n![Predicted Binding Site by the Pocket-Finder\\\nThe Pocket-Finder works by scanning the probe radius 1.6 angstroms along all gridlines of the grid resolution, 0.9 angstroms surrounding the protein. The output of the Pocket-Finder shows the highest precision (52.7) for site two, which was also energetically most favorable. The output of MVD cavity detection overlapped with that of the Pocket-finder.](jjnpp-10-01-18216-g001){#fig16040}\n\n### 4.2.2. Computational Molecular Docking {#sec82839}\n\nThe computational docking study of the ligand with the receptor (coordinate: 0.41 \u00d7 31.67 \u00d7 4.44) by MVD was evolved in terms of MolDock Score for best pose. Dihydrotestosterone exhibited the lowest docking score of -109.81. This observation is concordant with previous study with R 18-100 that has inhibitory effect on the AR. Among the studied ligands, esculin exhibited the most favorable binding to the receptor followed by morindone, anthragallol, soranjidiol, lucidin and damnacanthal. The MolDock Score also revealed that esculin's interaction with the receptor is likely to be more energetically economic than that of the dihydrotestosterone ([Table 2](#tbl21041){ref-type=\"table\"}).\n\nDuring dihydrotestosterone interaction with the receptor, the amino acid residues, including Gln 711, Arg 752 and Thr 877, were involved in hydrogen bonding while Gln 711, Met 745 and Phn 764 were involved in steric interaction ([Figure 2 A](#fig16041){ref-type=\"fig\"}).\n\nThe amino acid residues Leu 704, Asn 705, Arg 752 and Thr 877 were involved in hydrogen bonding in case of esculin's interaction receptor. Esculin exploited Leu 704, Gln 711, Leu 873, Phe 876, Met 787 and Met 895 residues at the binding site to maintain a steric interaction with the receptor ([Figure 2](#fig16041){ref-type=\"fig\"} B). Thus, the amino acid residues Arg 752 and Thr 877 were the common residues involved in the hydrogen bonding of the receptor with esculin and dihydrotestosterone.\n\n###### Molecular Docking Results from the Molegro Vertual Docker (MVD) for Ligands Bound to the Androgen Receptor\n\n  Ligands                   MolDock Score (kcal/mol)   H-Bond (kcal/mol)   Steric Energy (kcal/mol)\n  ------------------------- -------------------------- ------------------- --------------------------\n  **Damnacanthal**          -97.28                     -3.09               -123.05\n  **Lucidin**               -98.13                     -3.62               -113.96\n  **Dihydrotestosterone**   -109.81                    -4.62               -115.65\n  **Anthragallol**          -99.35                     -3.61               -106.28\n  **Soranjidiol**           -98.14                     -6.75               -109.83\n  **Morindone**             -105.49                    -7.98               -111.53\n  **Esculin**               -146.55                    -9.06               -134.47\n\n![Docking Poses of the Ligand-Androgen Receptor\\\nA) Dihydrotestosterone and androgen receptor interaction. Here, the residues of the receptor participating in hydrogen bonding with the corresponding ligand are shown in azure color. The interaction includes hydrogen bonding with Gln 711, Arg 752 and Thr 877, steric interaction with Gln 711, Met 745 and Phn 764. B) Esculin and androgen receptor interaction. Here, the residues of the receptor participating in hydrogen bonding with corresponding ligands are shown in azure color. The interaction includes hydrogen bonding with Leu 704, Asn 705, Arg 752 andThr 877steric interaction with Leu 704, Gln 711, Leu 873, Phe 876, Met 787 and Met 895.](jjnpp-10-01-18216-g002){#fig16041}\n\n5. Discussion {#sec82842}\n=============\n\nThe purpose of the present study was to evaluate potential phytochemicals with anti-prostate cancer activity from *H. excelsum* with prediction of mechanism of action using a computational molecular simulation study. This study firstly investigated the anti-prostate cancer effect of *H. excelsum* phytochemicals with a mechanistic insight.\n\nFrom the literature search, eight phytochemicals were selected as the subject ligands including anthragallol, damnacanthal, esculin, lucidin, morindone, nordamnacanthal, rubiadin, and soranjidiol while dihydrotestosterone was considered as the control. Besides, human androgen receptor (AR) ligand binding domain was selected as the receptor for the subsequent computational docking study. All the ligands were screened for drug-likeness rules, biological activity and probable toxicity. Esculin, lucidin, morindone and soranjidiol were found as more potent physiologically active molecules without any harmless effects ([Table 1](#tbl21040){ref-type=\"table\"}).\n\nUpon the generation of the 3D PDB structure, the ligands were docked to the AR ligand binding residues by Molegro Virtual Docker (MVD) tools. Results of the docking study suggest a favorable binding of esculin to the receptor with respect to dihydrotestosterone ([Table 2](#tbl21041){ref-type=\"table\"}). The literature search revealed that previous studies agree with the present findings. Several experimental data suggest the likelihood of esculin's anti-cancer effects. The coumarin glycoside esculin also scavenges superoxide radicals and decreases lipid peroxidation that is considered as a common strategy of a broad range on anti-cancer agents ([@A18216R16]). *Cichorium intybus*, an esculin containing plant, has been reported to have a modest inhibitory effects on the proliferation of prostate, breast and colorectal cells ([@A18216R17]). Analysis of amino acid residues involved in docking, hydrogen bonding energy contribution to docking score and steric interaction revealed a nearly similar ligand-receptor interaction for both of dihydrotestosterone and esculin. Still, from the present molecular simulation study it was profound that the contribution of hydrogen bonding to MolDock score was higher for esculin with respect to dihydrotestosterone ([Table 2](#tbl21041){ref-type=\"table\"}). Thus, hydrogen bonding and steric interaction patterns of esculin and dihydrotestosterone suggested a nearly similar ligand-receptor interaction pattern. Therefore, esculin has the potency to act like an antagonist of dihydrotestosterone, while it also has a potential AR inhibitory effect.\n"}
+{"text": "1. Introduction {#sec1-nutrients-11-00772}\n===============\n\nIn premature infants, human milk (HM) is associated with significant benefits on health and development. The mother's own milk (OMM) is therefore always recommended as the first nutritional choice. When OMM is unavailable, the use of donor milk (DM) rather than formula could be the first alternative for very low birth weight (VLBW) infants of less than 32 weeks \\[[@B1-nutrients-11-00772],[@B2-nutrients-11-00772],[@B3-nutrients-11-00772]\\].\n\nPreterm infants have high nutritional requirements, \\[[@B4-nutrients-11-00772],[@B5-nutrients-11-00772],[@B6-nutrients-11-00772]\\] and exclusive HM, even from infant's OMM or banked DM, will not provide intakes that reach current nutritional recommendations. Fortification is therefore recommended to improve post-natal growth \\[[@B7-nutrients-11-00772],[@B8-nutrients-11-00772]\\]. Nevertheless, the use of fortified HM could still fail to obtain qualitative and quantitative postnatal growth in the range of fetal growth \\[[@B9-nutrients-11-00772],[@B10-nutrients-11-00772],[@B11-nutrients-11-00772]\\]. That remains a concern as postnatal nutritional deficit and growth restriction during the neonatal period could be linked to altered long term health and neurodevelopment outcomes \\[[@B12-nutrients-11-00772],[@B13-nutrients-11-00772],[@B14-nutrients-11-00772]\\] in spite of the beneficial advantages associated with the early HM use \\[[@B15-nutrients-11-00772],[@B16-nutrients-11-00772],[@B17-nutrients-11-00772]\\].\n\nWorldwide, OMM and DM use in neonatal intensive care units (NICUs) increased over the last decade, but without practical and clear nutritional recommendations \\[[@B18-nutrients-11-00772]\\]. Recent studies, using infrared method, demonstrated the wide variability of protein and energy contents of either DM or OMM, suggesting that the use of theoretical composition values could induce nutritional deficiency or overload \\[[@B19-nutrients-11-00772]\\]. Studies of the impact of individualized HM fortification versus targeted or standard fortification on growth of VLBW infants are scarce \\[[@B20-nutrients-11-00772],[@B21-nutrients-11-00772],[@B22-nutrients-11-00772],[@B23-nutrients-11-00772]\\]. In addition, nutritional interests of fortified raw OMM versus pasteurized OMM or DM are still controversial \\[[@B10-nutrients-11-00772],[@B24-nutrients-11-00772],[@B25-nutrients-11-00772],[@B26-nutrients-11-00772],[@B27-nutrients-11-00772]\\]. A few studies have showed lower growth rates in infants receiving fortified DM compared to fortified OMM \\[[@B25-nutrients-11-00772],[@B26-nutrients-11-00772],[@B28-nutrients-11-00772]\\]. One frequently suggested explanation was the lower protein and fat content of DM frequently provided by mothers who delivered term infants or were in later stages of lactation \\[[@B29-nutrients-11-00772]\\]. Another explanation could be a reduction of the nutrient contents or bioavailability with the processing of DM \\[[@B27-nutrients-11-00772],[@B30-nutrients-11-00772]\\].\n\nThe primary objective of the present study is to evaluate growth in VLBW infants fed individualized fortified HM with predominant OMM (\u226575%) or predominant DM (\u226575%). We hypothesized that, using individualized fortification providing controlled similar protein and energy intakes, the use of OMM could improve growth during the early weeks of life. The secondary objective is to determine the influence of raw versus pasteurized OMM, hypothesizing that pasteurization could impair nutrients' bioavailability and therefore reduce the neonatal growth rate during the study period.\n\n2. Materials and Methods {#sec2-nutrients-11-00772}\n========================\n\n2.1. Study Population and Study Design {#sec2dot1-nutrients-11-00772}\n--------------------------------------\n\nThis is a single center prospective and non-interventional study conducted in the NICU of the University of Li\u00e8ge, Belgium evaluating growth in preterm infants fed HM with individualized fortification (IHMF). From January 1, 2007 to December 31, 2014, data on HM use, HM composition, and fortification in preterm infants born \\<32 weeks gestation (GA) were collected daily in our NICU human milk bank. From those datasets, preterm infants receiving IHMF as previously reported \\[[@B19-nutrients-11-00772],[@B20-nutrients-11-00772]\\] were included in the present study. Infants with chromosomal or congenital anomalies impacting growth and those receiving IHMF for less than 14 days where excluded.\n\nTo evaluate the respective influences of OMM and DM, growth and nutritional intakes (mean \u00b1 standard deviation (SD), during the study period, were compared in preterm infants fed predominantly OMM (\u226575%) or predominantly DM (\u226575%). In addition, the effects of HM types on growth during the study period were evaluated on the whole population, including a third group receiving a mixed HM diet ranging from 26% to 74% of OMM. Under existing Belgian law at the time of the study, the collection of anonymized data concerning clinical routine practices does not require approved of the Ethical Committee. However, the parents were informed and provided consent for donor milk use as necessary, as well as HM analysis and individualized fortification.\n\n2.2. Nutritional Practices {#sec2dot2-nutrients-11-00772}\n--------------------------\n\nGlobal nutritional management was previously reported \\[[@B31-nutrients-11-00772]\\]. According to our protocol, all VLBW infants received parenteral nutrition on the first day of life with a balanced standardized parenteral solution, designed to provide preterm infants a mean intake of 37--38 kcal/kg/day and 2.4--2.5 g/kg/day of protein on the first day of life followed by a rapid increase to a target intake of 3.8 kcal/kg/day of protein and 120 kcal/kg/day by 5 to 8 days of life \\[[@B31-nutrients-11-00772]\\]. Insulin therapy was only used in case of hyperglycemia (\\>10 mmol/L) during parenteral nutrition. Enteral nutrition (10--20 mL/kg/day) was initiated within the first hours of life with maternal colostrum or unfortified DM and progressively increased by 10 to 20 mL/kg/day until 160 to 180 mL/kg/day according to tolerance. Mothers were encouraged to breastfeed and received support from dedicated nurses in the unit. HM was expressed at the hospital or at home, by manual expression or by using an electric pump, and transported under aseptic HACCP (Hazard Analysis Critical Control Point) conditions, and mothers were provided with written instructions regarding mechanical expression, milk collection, storage, and transport. OMM provided by the mother was kept at 4 \u00b0C and used within 72 h. DM was obtained from our own NICU HM Bank. Milk donors were unpaid volunteers. Informed consent for the use of their milk for feeding preterm infants or for research purposes was obtained Most of these donors had delivered preterm. DM from the early stage of lactation (first week) was separately pooled, processed, labeled, and used during the first days of life in extremely preterm infants in the absence or as a supplement of OMM. DM was always Holder pasteurized (62.5 \u00b0C for 30 min) in batches of 5 L. OMM was used as previously described. OMM of cytomegalovirus positive mothers of infants of less than 32 weeks GA at birth was pasteurized until postconceptional age of 34 weeks. A bacteriologic count of OMM was performed after 24 h of incubation, allowing heavy contaminated OMM to be discarded or to use it directly as raw milk or pasteurized milk in case of light contamination \\[[@B32-nutrients-11-00772],[@B33-nutrients-11-00772]\\]. Supplemental parenteral nutrition was withdrawn when enteral intakes reached 100 to 120 mL/kg/day. Standard HM fortification was introduced at 25% (addition of 0.275 g of protein and 3.5 kcal in 100 mL of HM) of full fortification once preterm infants tolerated a minimum of 50 mL/kg/d enterally and was gradually increased to full fortification (addition of 1.1 g of protein and 14 kcal in 100 mL of HM). IHMF was considered when a minimum of 140 to 150 mL/kg/day was provided. As IHMF requires extra workload for the HM Bank, its prescription was left to the attending neonatologist.\n\n2.3. Individualized HM Fortification (IHMF) {#sec2dot3-nutrients-11-00772}\n-------------------------------------------\n\nFortified HM was prepared daily in the HM Bank. To allow individualized fortification, a sample of 10 mL of HM was taken from the daily pool. Macronutrient HM concentration was determined using a mid-infrared analyzer (Milkoscan minor^\u00ae^, Foss, Hiller\u00f8d, Denmark) previously validated for HM \\[[@B19-nutrients-11-00772]\\]. The Milkoscan analyzer was calibrated to provide the total protein concentration of HM similar to the total nitrogen content, including non-protein nitrogen, measured by a chemical method. HM was warmed to 37 \u00b0C and homogenized using an ultrasonic homogenizer (Sonicator^\u00ae^, Uppsala, Sweden) before analysis. Data of protein and fat contents were gathered in an excel table to calculate the needs of supplementation according to recommendations \\[[@B5-nutrients-11-00772]\\]. IHMF was performed in two steps: (1) Adjustment of fat content up to 4 g/dL by adding medium-chain triglycerides (MCTs; Liquigen^\u00ae^ Danone, The Netherlands), (2) addition of a multicomponent powdered HM fortifier (Enfamil Human Milk Fortifier powder; Mead-Johnson or Nutrilon B.M.F.; Nutricia) to finally provide 4.3 g/kg/day of protein according to the daily volume order.\n\n2.4. Data Collection and Growth Assessment {#sec2dot4-nutrients-11-00772}\n------------------------------------------\n\nDay 1 of the study was defined as the first day of IHMF. Weight, HM type (raw OMM, pasteurized OMM, and pasteurized DM), macronutrient composition of HM, MCTs, and fortifier addition and volume intakes were prospectively collected daily during all the IHMF period and used to calculate the nutritional intakes. The energy content was calculated using the Atwater factors: 4 kcal/g for protein and carbohydrate and 9 kcal/g for fat.\n\nOther clinical and demographic data were collected from the medical charts of infants, and this included prenatal complications, delivery information, and neonatal outcomes in the NICU until discharge or transfer to another hospital.\n\nInfants weight (to the nearest 1 g) was measured daily by nurses using a calibrated electronic scale. Length and head circumference (HC) were assessed weekly (both to the nearest 0.1 cm), length using a length board and HC using a non-stretch measuring tape. Weight gain velocity (grams per kilogram per day) was calculated during the IHMF period using the 2-point average method \\[[@B34-nutrients-11-00772]\\]. $$\\text{Weight\\ gain} = \\frac{1000 \\ast {({W2 - W1})}}{\\frac{W1 + W2}{2} \\ast {({d2 - d1})}}$$ where W = weight in grams; d = day; 1 = beginning of the time interval; and 2 = end of the time interval.\n\nWeight for age, length for age, and head circumference for age Z scores were calculated using Fenton reference growth charts according to corrected GA \\[[@B35-nutrients-11-00772]\\].\n\n2.5. Statistical Analysis {#sec2dot5-nutrients-11-00772}\n-------------------------\n\nNormally distributed data are reported as a mean with standard deviation and groups are compared by using *t*-tests or one-way analysis of variance (ANOVA) with Bonferroni's correction for post hoc pairwise comparisons. Non-normally distributed data are presented as a median with a range, and groups were compared by Kruskall-Wallis ANOVA tests. Categorical data are presented as numbers and percentages and groups were compared by Chi-squared tests. A *p*-value of \\<0.05 was considered as significant.\n\nStepwise multivariate analysis was performed to evaluate the respective influences of significant univariate variables and type of HM (raw OMM, pasteurized OMM, and DM) on growth parameters during the study period. The relation was presented by Pearson correlation coefficient (r or r\u00b2). A *p* \\< 0.05 was considered as significant.\n\nAll statistical analyses were performed by using Tibco Statistica software version 13 (TIBCO, Palo Alto, CA, USA).\n\n3. Results {#sec3-nutrients-11-00772}\n==========\n\n3.1. Study Population {#sec3dot1-nutrients-11-00772}\n---------------------\n\nBetween January 1, 2007 and December 31, 2014, 726 infants with gestational age of less than 32 weeks were admitted to the University of Li\u00e8ge NICU by birth or transfer, of which 665 were discharged alive. The total number of infants that received IHMF during NICU hospitalization was 204. Eighty-two were excluded as they received IHMF of less than 14 days, 12 for chromosomal or congenital anomalies impacting growth, and 9 for incomplete data, leaving 101 subjects included in the study.\n\n3.2. Clinical Variables {#sec3dot2-nutrients-11-00772}\n-----------------------\n\nOut of 101 preterm infants (BW 975 \u00b1 255 g for a GA of 27.8 \u00b1 1.9 weeks), IHMF was initiated at 19 \u00b1 8 days of life during 26 \u00b1 8 days. Thirty-seven infants were fed \u226575% of intake with OMM, 33 infants were fed \u226575% of intake with DM, and 31 with a mixed HM diet with (26%--74% OMM). Demographic and clinical characteristics according to the three HM diets (\u226575% OMM versus \u226575% DM versus 26%--74% OMM) are detailed in [Table 1](#nutrients-11-00772-t001){ref-type=\"table\"}. Demographic parameters at birth were similar in the three groups with the exception of HC being significantly lower in the DM group compared to those fed the mixed HM diet.\n\nNeonatal morbidities at study baseline were also similar in the three groups ([Table S1](#app1-nutrients-11-00772){ref-type=\"app\"}) with a trend to a higher incidence of late onset sepsis in the DM group (*p* = 0.062). However, no other significant difference in morbidities that could influence growth was reported between the three groups during and after the study period ([Table S2](#app1-nutrients-11-00772){ref-type=\"app\"}). Necrotizing enterocolitis was observed in three infants, two in the DM group after the study period, (two days after the introduction of preterm formula and the day before suggested discharge in a preterm infant fed formula for several weeks), and the last one in the intermediate group, during the study period, the day after a transfusion. Two infants in the DM category, one in OMM and three in the intermediate group presented clinical infection during or after the study: Five respiratory infections and one urinary tract infection. Insulin treatment rate was similar in all the groups and was only used in case of hyperglycemia during parenteral nutrition. No infants received insulin during the study period.\n\n3.3. Influence of OMM Versus DM {#sec3dot3-nutrients-11-00772}\n-------------------------------\n\nAccording to the primary objective of the study, nutritional intakes and growth during IHMF were compared in VLBW infants fed predominantly OMM and DM.\n\n### 3.3.1. Human Milk Composition and Nutritional Intakes {#sec3dot3dot1-nutrients-11-00772}\n\nThe contributions of the HM categories in the two groups are gathered in [Table 2](#nutrients-11-00772-t002){ref-type=\"table\"}. OMM accounted for, respectively, 95.4% and 2.2% of the HM intakes during the IHMF study. Lipid content was significantly higher in the OMM than in the DM group. Nevertheless, in both groups, fortified HM provided similar mean energy and protein intakes with low variability, accounting for, respectively, less than 5.6% and 3.6 % for energy and 3.6% and 3.4% for protein.\n\n### 3.3.2. Growth {#sec3dot3dot2-nutrients-11-00772}\n\nAs shown in [Table 3](#nutrients-11-00772-t003){ref-type=\"table\"}, weight (*p* = 0.002) and length gain (*p* = 0.020), but not HC gain (*p* = 0.120), were significantly higher in infants receiving predominantly OMM compared to those fed predominantly DM during the IHMF period. Similarly, Z-scores gains for weight (*p* \\< 0.0001), length (*p* = 0.004), and HC (*p* = 0.013) were all significantly higher in infants receiving mostly OMM than in those fed mostly DM during the IHMF period.\n\n3.4. Effects of Type of Human Milk (Raw OMM, Pasteurized OMM, and Pasteurized DM) {#sec3dot4-nutrients-11-00772}\n---------------------------------------------------------------------------------\n\n### 3.4.1. Human Milk Composition and Nutritional Intakes {#sec3dot4dot1-nutrients-11-00772}\n\nIn line with the secondary objective of the study, the whole population was evaluated according to the main HM type received during the study period, DM \\> 50% (DM), DM \u2264 50%, pasteurized \\> raw OMM (POMM), and DM \u2264 50% and raw \\> pasteurized OMM (ROMM) to evaluate the influence of OMM pasteurization on growth velocity during the study period. As shown in [Table 4](#nutrients-11-00772-t004){ref-type=\"table\"}, DM accounted to 88.5% in the DM group (*n* = 45), pasteurized OMM to 70.3% in the POMM group (*n* = 41), and raw OMM to 69.1% in the ROMM group (*n* = 15). Energy and protein intakes during the study period were similar in the three groups.\n\n### 3.4.2. Growth {#sec3dot4dot2-nutrients-11-00772}\n\nBoth weight gain and weight Z-score gain in the DM group were significantly lower than in the other two groups. In addition, weight gain, but not weight Z-score gain, was significantly higher in the ROMM versus POMM group. Length and HC gains were similar in the three groups. Nevertheless, the length and HC Z-score gains were significantly improved in the ROMM group compared to the DM group.\n\n3.5. Univariate and Multivariate Analysis on the Whole Population {#sec3dot5-nutrients-11-00772}\n-----------------------------------------------------------------\n\n### 3.5.1. Univariate Analysis {#sec3dot5dot1-nutrients-11-00772}\n\nUnivariate linear regression analysis on the whole population, showed that birthweight, gestational age, postnatal age at study day 1, as well as protein and energy intakes did not significantly influence weight and length gain during the study period.\n\nWeight gain during the IHMF period was significantly influenced by two univariate factors; study duration (r = 0.31, *p* = 0.0014) and percentage of raw OMM (r = 0.47, *p* \\< 0.00001). For length, the percentage of total OMM (r = 0.20, *p* = 0.046) was the only factor significantly influencing length gain.\n\n### 3.5.2. Multivariate Analysis {#sec3dot5dot2-nutrients-11-00772}\n\n#### *Weight Gain and Weight for Age Z-score Difference*\u00a0\n\nStepwise multivariate analysis demonstrated that weight gain (g/kg/day) was positively related to the proportion of raw OMM, proportion of pasteurized OMM, and postnatal age at the first day of study, but negatively related to study duration and birthweight. Those factors explain 22.7%, 3.7%, 3.1%, 9.8%, and 3.0% of the weight gain, respectively. It was also estimated that the weight for age Z-score difference during IHMF was related to the raw OMM proportion, gestational age, and birth weight, contributing, respectively, to 18.0%, 12.1%, and 10.7% of the Z-score difference.\n\n#### *Length Gain and Length for Age Z-score Difference*\u00a0\n\nFor length gain, only two parameters were significant; the proportion of total OMM (raw + pasteurized) and postnatal age at baseline, explaining, respectively, 4.0% and 4.4% of the length gain. Similarly, length for age Z-score difference was related to the proportion of total OMM (raw + pasteurized) and study duration, contributing, respectively, to 6.5% and 5.4% of the difference.\n\n4. Discussion {#sec4-nutrients-11-00772}\n=============\n\nThis study is the first providing daily controlled nutritional intakes in preterm infants fed HM with individualized fortification after daily determination of HM composition by a validated infrared method \\[[@B19-nutrients-11-00772],[@B20-nutrients-11-00772]\\]. Because of IHMF, protein and energy intakes were similar with very low variability ([Table 2](#nutrients-11-00772-t002){ref-type=\"table\"}) in the two groups, it adequately allows for comparisons of growth and metabolic tolerance in VLBW infants fed exclusively fortified OMM (95.4% \u00b1 7.8%) or DM (97.8% \u00b1 5.4%). This study found that weight gain velocity during IHMF was on average 1.6 g/kg/d higher in infants fed OMM than in those fed DM, with an additional benefit on length gain of around 0.18 cm/week on average, suggesting a growth specific effect of OMM in preterm infants. In addition, the use of predominant OMM (\u226575%) instead of predominant DM (\u226575%) significantly improved weight, length, and HC Z-score changes during the study period ([Table 3](#nutrients-11-00772-t003){ref-type=\"table\"}).\n\nAs shown in [Table 3](#nutrients-11-00772-t003){ref-type=\"table\"}, around two thirds of OMM was provided after Holder pasteurization and not as raw OMM. This is mostly explained by the strategy applied to reduce the risk of infectious transmission with raw milk. According to our previous study \\[[@B32-nutrients-11-00772],[@B33-nutrients-11-00772]\\], up to 20%--50% of the OMM samples were contaminated and were either pasteurized or discarded. In addition, to avoid CMV contamination or infection \\[[@B30-nutrients-11-00772]\\], OMM of CMV seropositive mothers of VLBW infants was also systematically pasteurized. The variability of the raw OMM intakes in our whole population allowed us to evaluate the respective role of raw OMM versus pasteurized OMM or DM on growth velocity in the preterm infants. We found that ROMM and POMM both have a positive effect on weight gain, contributing to an increase of +2.8 g/kg/day and +0.9 g/kg/day, respectively, compared to DM. It suggests that the major positive effect of OMM could be the result of its use as a raw product, with a mean weight gain difference of 2.0 g/kg/day compared to pasteurized OMM ([Table 4](#nutrients-11-00772-t004){ref-type=\"table\"}). Our study also suggests that the use of raw OMM also induces a significant positive effect on weight (*p* = 0.003), length (*p* = 0.013), and HC (*p* = 0.016) Z-score gains during the study period compared to DM. The benefits of POMM on DM was limited on weight gain and weight Z-score gain whereas benefits on length and HC Z-scores were not significant with *p* values of 0.2 and 0.07, respectively, contrasting with the benefits observed with ROMM. Therefore, our study suggests that the limited beneficial effect of POMM versus DM remains to be confirmed in additional studies.\n\nThe optimal reference growth chart to evaluate postnatal growth velocity in preterm infants is still debated as discussed recently by an international expert group \\[[@B34-nutrients-11-00772]\\]. From this review, it was recommended to use the average 2 points or the exponential 2 points methods to evaluate the growth velocity. Both formulas provide similar results that are highly correlated with a slightly higher value for the exponential method as shown in the [Figure 1](#nutrients-11-00772-f001){ref-type=\"fig\"}. In our study, we chose to use the average 2 points method for comparison to our previous studies \\[[@B9-nutrients-11-00772],[@B20-nutrients-11-00772],[@B31-nutrients-11-00772],[@B36-nutrients-11-00772]\\].\n\nIn addition, this review and others \\[[@B34-nutrients-11-00772],[@B37-nutrients-11-00772]\\] suggest that it is time to report growth studies in a standardized fashion. The standardized growth report is also debated and several growth charts have been proposed to evaluate postnatal Z-scores in VLBW infants. Recently, it was suggested that the Fenton revised growth charts of 2013 could be outdated by the recent INTERGROWTH-21st Postnatal Follow-up Study of preterm infants \\[[@B38-nutrients-11-00772]\\]\n\nWe agree that both the Fenton growth chart and the INTERGROWTH-21st have some limitations. The Fenton growth charts are built with a meta-analysis of cross-sectional fetal charts without take into account that postnatal growth composition differs to that of fetal growth composition. By contrast, the INTERGROWTH-21st chart is longitudinal, not cross-sectional and non-fetal. However, this contains some limitations, including the small number of very preterm infants included in the database, as well as the lack of \"gold standard nutrition\" \\[[@B39-nutrients-11-00772],[@B40-nutrients-11-00772]\\]. Indeed, the description of the feeding regimen in the INTERGROWTH study is limited. It is specified that the main feeding regimen of the included preterm infants was human milk and that the use of HMF was only added to expressed HM until a baby's weight reached 1800--2000 g and not up to discharge. The daily protein and energy intakes were not adequately controlled during the postnatal period, suggesting that some cumulative protein and energy deficits could induce relative postnatal growth restriction in the evaluated population of preterm infants. The authors of the INTERGROWTH-21 group \\[[@B39-nutrients-11-00772]\\] recommend the INTERGROWTH-21st Preterm Postnatal Growth Standards for monitoring the growth of more than 90% of preterm infants who are born at \u226532 weeks and recognize that the construction of charts for very preterm infants (\\<32 weeks' gestation) is problematic. We consider that our population, including 30% of preterm infants with a GA \\< 27 weeks and 100% at \\<32 weeks at birth, but also 80% still \\<32 weeks at baseline, is not in the optimal range of the INTERGROWTH reference. Therefore, our results were compared to the combined references growth chart of the fetus and the term infants as proposed by Fenton et al. in 2013 \\[[@B35-nutrients-11-00772]\\]. However, data of preterm infants \\>27 weeks GA were also compared to the INTERGROWTH-21st reference in [Figures S1 and S2](#app1-nutrients-11-00772){ref-type=\"app\"}.\n\nThis study demonstrates a significant positive impact of both OMM and raw OMM on growth in preterm infants fed HM. This effect seems independent of nitrogen, lipid, and carbohydrate content as this was controlled by the IHMF in this study. Nutritional and growth benefits of fortified OMM versus fortified DM is still debated and studies report controversial results regarding growth and Z-score changes in preterm infants. Thus, in two observational and one retrospective study, a weight gain benefit was reported in preterm infants fed fortified OMM. In 2011, Montjaux et al. \\[[@B25-nutrients-11-00772]\\] suggested that weight gain was directly proportional to the amount of fresh raw OMM compared to pasteurized fortified DM (*n* = 48). More recently, Madore et al. \\[[@B26-nutrients-11-00772]\\] showed a significantly higher weight gain in preterm infants fed predominantly fortified OMM compared to those fed predominantly fortified DM during the first month of life (*n* = 56). Brownell et al. \\[[@B28-nutrients-11-00772]\\], using OMM as a reference, also reported a significant decrease in mean weight and head velocity during a hospital stay for every 10% increase of the total feeding volume provided as DM (*n* = 314). By contrast, two retrospective studies did not observe any significant difference in weight gain between premature infants receiving either exclusively OMM or DM as a sole diet (*n* = 92) \\[[@B41-nutrients-11-00772]\\] or in those fed predominantly (\\>50%) fortified OMM or fortified DM (*n* = 299) \\[[@B42-nutrients-11-00772]\\]. In addition, a third retrospective study found no significant difference in weight Z-score change by HM diet (*n* = 88) (\\>75% donor vs. \\>75% OMM; *p* = 0.28) \\[[@B10-nutrients-11-00772]\\]. In contrast to our study, none of those studies precisely determined and controlled the protein and energy intakes, and the rate of pasteurization, if any, in the OMM groups was not specified. Still, the effect of pasteurization on growth velocity was recently evaluated as a secondary outcome in a randomized study of more than 300 premature infants receiving fortified OMM either raw or pasteurized. In that study, a similar growth was observed between the two groups \\[[@B43-nutrients-11-00772]\\].\n\nIn preterm infants fed fortified HM, postnatal growth restriction was frequently reported as well as loss of Z-score during the full HM fortification period \\[[@B10-nutrients-11-00772],[@B44-nutrients-11-00772]\\]. Repetitively, recommendations from various expert committees suggest that nutritional requirements are similar in VLBW infants fed fortified HM or preterm formula (PTF) \\[[@B4-nutrients-11-00772],[@B6-nutrients-11-00772]\\]. Until now, no specific guidelines have been proposed for fortified HM fed preterm infants. However, it is recognized that at similar controlled protein and energy intakes, growth velocity is significantly lower in preterm infants fed fortified HM than in those fed PTF \\[[@B9-nutrients-11-00772]\\]. Metabolic and energy balance studies show that such a difference could be the result of lower metabolized protein and energy contents of fortified HM compared to PTF \\[[@B36-nutrients-11-00772]\\]. The mean difference in nitrogen utilization (retention/intake) as well as the mean difference in energy absorption rates measured by bomb calorimetry were both about 10% less with fortified HM \\[[@B45-nutrients-11-00772]\\]. This difference could be partially due to the use of pasteurization. In addition, as shown more recently, the use of standard reference values for OMM and DM may induce an overestimation of the protein and energy content of fortified HM \\[[@B19-nutrients-11-00772],[@B46-nutrients-11-00772]\\]. While preterm OMM with its higher protein content could improve growth compared to DM, it remains insufficient to support adequate growth, especially after the first month of lactation when the OMM protein concentration decreases \\[[@B29-nutrients-11-00772]\\]. A previous study performed in our NICU found that the macronutrient and energy content of OMM was highly variable and unpredictable. Protein and energy content of DM was also significantly lower than that of OMM \\[[@B19-nutrients-11-00772]\\]. Of all the daily OMM and DM samples (*n* = 2630) measured in the present study, 67% were \\<1.5 g protein/dL and 62% were \\<67 kcal energy/dL, values commonly considered as reference values for HM composition to estimate nutrient intakes in clinical practice.\n\nBy using metabolic balance studies and indirect calorimetry, we previously showed that protein intake and the protein energy ratio were major determinants of weight gain in VLBW infants \\[[@B36-nutrients-11-00772]\\]. In a recent multicentric study \\[[@B46-nutrients-11-00772]\\], we showed that theoretical intakes of 4.46 g/kg/day of protein and 125 kcal/kg/day (not confirmed by HM content analysis) led to a stable weight Z-score during the study period in VLBW infants receiving new HMF while the weight Z-score decreased significantly in the control HMF group theoretically receiving 3.81 g/kg/day of protein and 125 kcal/kg/day. Trends in the same direction were observed for length Z-score changes. In that study, the protein intakes were not measured, but estimated according to a preterm HM reference \\[[@B47-nutrients-11-00772]\\] and were therefore probably overestimated in regard to the large use of DM and pasteurized OMM \\[[@B46-nutrients-11-00772]\\]. Based on blood urea nitrogen and urinary urea excretion, we speculated that protein utilization in the new HMF might not have been optimal due to a relative deficiency in metabolized energy intake \\[[@B46-nutrients-11-00772]\\].\n\nConsidering the variability of HM macronutrient contents and the lower bioavailability of HM, in the present study, we targeted higher mean protein and energy intakes than those generally recommended \\[[@B4-nutrients-11-00772],[@B6-nutrients-11-00772],[@B48-nutrients-11-00772]\\]. Thus, during the study period, preterm infants received controlled mean intakes of 143 kcal/kg/day and 4.2 g/kg/day of proteins between 30.5 and 34 weeks' post-menstrual age, resulting in mean positive weight and HC Z-scores changes of 0.13 and 0.59, respectively, but a limited negative mean length Z-score change of 0.25 in preterm infants fed \u226575% OMM. By contrast, negative Z-score changes for weight (0.26 on average) and length (0.59 on average) were observed in the group receiving \u226575% DM ([Table 3](#nutrients-11-00772-t003){ref-type=\"table\"}). These results suggest that such intakes are close to the minimal requirements necessary for preterm infants fed fortified OMM in such a range of post-menstrual age, but could still be limited in those fed fortified pasteurized DM. In addition, knowing that postnatal growth quality differs to that of fetal growth by an increase in fat deposition, the discrepancies between weight and length Z-scores benefits could be the result of a relative deficit in the lean body mass accretion rate during the study period. Therefore, our study also suggests that protein and energy requirements of preterm infants fed fortified HM are higher than that currently recommended \\[[@B4-nutrients-11-00772],[@B6-nutrients-11-00772],[@B48-nutrients-11-00772]\\] and that specific nutritional guidelines for HM fed preterm infants need to be designed, promoting the use of OMM, but considering the limitations of its use as raw OMM in VLBW infants.\n\nImproving HM fortification by IHMF through the use of infrared technology and extra protein and energy supplementation may be one of the strategies to optimize the nutritional composition of HM to meet the nutritional needs of preterm infants, especially when DM is used. It was demonstrated that IHMF decreases the variability linked to HM content and safely optimizes protein and energy intake \\[[@B19-nutrients-11-00772],[@B21-nutrients-11-00772],[@B49-nutrients-11-00772],[@B50-nutrients-11-00772]\\]. Premature infants fed with low macronutrient content HM benefit the most from IHMF, with improved growth outcomes. However, infrared devices, originally developed for use in the dairy industry, must be calibrated and validated for HM analysis before clinical use by following good laboratory and clinical practice, and appropriate sample preparation must be done otherwise their use can affect the growth outcomes of preterm infants \\[[@B19-nutrients-11-00772],[@B50-nutrients-11-00772],[@B51-nutrients-11-00772]\\].\n\n5. Conclusions {#sec5-nutrients-11-00772}\n==============\n\nOur study is one of the first studies showing that a daily controlled high protein and energy intakes (4.2 g of protein and 143 Kcal/kg/day) of fortified raw OMM is associated with important growth benefits in preterm infants. It also suggests that pasteurized OMM provides a limited, but significant growth benefit compared to DM, suggesting that pasteurization significantly impaired the bioavailability of protein and energy intake. The increase in protein and/or energy intakes in preterm infants receiving fortified pasteurized HM could be postulated in view of these results, but needs to be demonstrated in further studies. In addition, our study also suggests that specific and different nutritional recommendations need to be designed for preterm infants fed OMM and DM.\n\nThe authors thank the families of the infants who participated in the study, all the mothers who donated their breast milk to our HM Bank as well as the staff at HM bank who performed daily human milk analysis.\n\nThe following are available online at <https://www.mdpi.com/2072-6643/11/4/772/s1>, Table S1: Initial clinical outcomes before daily individualized HM fortification (IHMF); Table S2: Post baseline clinical outcomes. Figure S1. Weight Z-score and Weight Z-score change according to Fenton and Intergrowth during the study in all preterm infants included in the study with a GA \\> 27 weeks; Figure S2. Weight for age Z-score at day1 and at the end of the study period, and Z-score gain during the study in infants fed mostly donor (*n* = 45) versus raw OMM (*n* = 15). Comparison of FENTON and INTERGROWTH's references.\n\n###### \n\nClick here for additional data file.\n\nV.d.H. and J.R. designed the study and conducted the study with input from C.P., T.S, C.K., F.S. and V.R. C.K., F.S. and V.d.H. collected the data. V.d.H. and J.R. conducted the data analysis and interpretation and wrote the paper with input from C.P. and V.R.\n\nThis research received no external funding.\n\nThe authors declare no conflict of interest.\n\n![Evaluation of growth velocity by the exponential and the average 2 points methods.](nutrients-11-00772-g001){#nutrients-11-00772-f001}\n\nnutrients-11-00772-t001_Table 1\n\n###### \n\nInfants clinical characteristics according to human milk diet.\n\n  ----------------------------------------------------------------------------------------------------\n  m \u00b1 SD                          \u226575% OMM\\      26%--74% OMM\\   \u226575% DM\\       All Subjects\\   *p*\n                                  *n* = 37       *n* = 31        *n* = 33       *n* = 101       \n  ------------------------------- -------------- --------------- -------------- --------------- ------\n  Male sex, n (%)                 18 (49)        15 (48)         17 (52)        50 (50)         0.96\n\n  Gestational age, weeks,         27.7 \u00b1 2.1     28.2 \u00b1 1.9      27.5 \u00b1 1.8     27.8 \u00b11.9       0.26\n\n  Birth Weight, g,                983 \u00b1 244      1042 \u00b1 312      901 \u00b1 185      975 \u00b1 255       0.08\n\n  Birth Weight \\< 1000 g, n (%)   20 (54)        16 (52)         24 (73)        60 (59)         0.16\n\n  Mean Weight z score,            \u22120.19 \u00b1 0.99   \u22120.37 \u00b1 0.89    \u22120.48 \u00b1 0.82   \u22120.34 \u00b1 0.91    0.47\n\n  Birth Length, cm,               35.0 \u00b1 3.3     35.8 \u00b1 3.9      34.6 \u00b1 2.9     35.1 \u00b1 3.4      0.34\n\n  Birth HC, cm,                   24.9 \u00b1 1.9     25.8 \u00b1 2.3      24.5 \u00b1 1.7     25.0 \u00b1 2.0      0.02\n\n  Vaginal Delivery, n (%)         16 (43)        9 (29)          7 (21)         32 (32)         0.13\n\n  Twin, n (%)                     8 (22)         12 (39)         6 (18)         26 (26)         0.13\n\n  Apgar Score 1 min,              6.5 \u00b1 2.2      6.1 \u00b1 2.2       6.1 \u00b1 2.0      6.2 \u00b1 2.1       0.60\n\n  Apgar Score 5 min,              7.9 \u00b1 1.5      7.8 \u00b1 1.5       7.9 \u00b1 1.1      7.9 \u00b1 1.4       0.92\n\n  Antenatal steroids, n (%)       35 (95)        27 (87)         29 (88)        91 (90)         0.30\n\n  Study duration,                 27 \u00b1 8         27 \u00b1 8          24 \u00b1 6         26 \u00b1 8          0.14\n\n  GA age at study day 1, weeks,   30.5 \u00b1 1,5     30.8 \u00b1 1,6      30.5 \u00b1 1,5     30,6 \u00b1 1,5      0.64\n\n  GA age at study end, weeks,     34.2 \u00b1 1.4     34.7 \u00b1 1.8      33.9 \u00b1 1.5     34.3 \u00b1 1.6      0.12\n  ----------------------------------------------------------------------------------------------------\n\nOMM = own mother's milk; DM = donor milk; GA = gestational age; data are presented as n (%) for categorical variables and mean (m) \u00b1 standard deviation (SD) for continuous variables; *p* \\< 0.05 based on ANOVA for continuous variable and chi square for categorical variables.\n\nnutrients-11-00772-t002_Table 2\n\n###### \n\nHuman milk composition and nutritional intakes during study in the two groups.\n\n  --------------------------------------------------------------------------------\n                                           \u226575% OMM\\     \u226575% DM\\      *p*-Value\n                                           *n* = 37      *n* = 33      \n  ---------------------------------------- ------------- ------------- -----------\n  **Human Milk Category (%)**                                          \n\n  Raw OMM                                  31.3 \u00b1 33.6   0.5 \u00b1 3.0     \\<0.001\n\n  Pasteurized OMM                          64.1 \u00b1 33.1   1.7 \u00b1 4.7     \\<0.001\n\n  Pasteurized DM                           4.6 \u00b1 7.8     97.8 \u00b1 5.4    \\<0.001\n\n  **Human Milk Composition (Infrared)**                                \n\n  Protein, g/dL                            1.44 \u00b1 0.22   1.35 \u00b1 0.14   0.056\n\n  Lipid, g/dL                              3.87 \u00b1 0.59   3.61 \u00b1 0.23   0.022\n\n  Carbohydrates, g/dL                      6.84 \u00b1 0.22   6.86 \u00b1 0.19   0.695\n\n  **Nutritional Intakes (Units/kg/day)**                               \n\n  Volume, mL                               167 \u00b1 10      166 \u00b1 8       0.536\n\n  Energy, kcal                             143 \u00b1 8       141 \u00b1 6       0.148\n\n  Protein, g                               4.17 \u00b1 0.15   4.15 \u00b1 0.14   0.512\n  --------------------------------------------------------------------------------\n\nData are presented as mean \u00b1 SD; *p* \\< 0.05 based on *t*-test.\n\nnutrients-11-00772-t003_Table 3\n\n###### \n\nGrowth rate and Z-score gain in preterm infants fed individualized fortified with predominantly own mother's milk (OMM) or donor milk (DM.)\n\n  -----------------------------------------------------------------------------------\n                                 OMM \u2265 75%\\     DM \u2265 75%\\      Delta\\       *p*\n                                 *n* = 37       *n* = 33       OMM vs. DM   \n  ------------------------------ -------------- -------------- ------------ ---------\n  Weight gain, g/kg/day          19.8 \u00b1 2.0     18.2 \u00b1 2.2     +1.6         0.002\n\n  Length gain, cm/week           1.17 \u00b1 0.26    0.99 \u00b1 0.36    +0.18        0.020\n\n  Head circumference, cm/week    1.13 \u00b1 0.22    1.04 \u00b1 0.27    +0.09        0.120\n\n  Weight Z-score gain, g/kg/d    0.13 \u00b1 0.35    \u22120.26 \u00b1 0.41   +0.39        \\<0.001\n\n  Length Z-score gain, cm/week   \u22120.25 \u00b1 0.41   \u22120.59 \u00b1 0.52   +0.33        0.004\n\n  HC Z-score gain, cm/week       0.59 \u00b1 0.50    \u22120.24 \u00b1 0.65   +0.35        0.013\n  -----------------------------------------------------------------------------------\n\nData are presented as mean \u00b1 standard deviation; *p* \\< 0,05 based on *t*-test.\n\nnutrients-11-00772-t004_Table 4\n\n###### \n\nGrowth rate and nutritional intakes according to the main human milk type received during the study period.\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Human Milk Type\\       DM 88.5 \u00b1 16.9   POMM 70.3 \u00b1 22.6   Delta vs. DM   *p* vs. DM   ROMM 69.1 \u00b1 19.9   Delta vs. DM   *p* vs. DM   Delta vs. POMM   *p* vs. POMM\n  Volume Intake (%)                                                                                                                                      \n  ---------------------- ---------------- ------------------ -------------- ------------ ------------------ -------------- ------------ ---------------- --------------\n  ***n***                45               41                                             15                                                              \n\n  Energy, kcal/kg/day    141.3 \u00b1 6.3      142.4 \u00b1 7.3        \\-             0.432        143.7 \u00b1 6.2                       0.210        \\-               0.552\n\n  Protein, g/kg/d        4.15 \u00b1 0.14      4.19 \u00b1 0.13        \\-             0.211        4.18 \u00b1 0.15                       0.494        \\-               0.855\n\n  Weight gain, g/kg/d    18.2 \u00b1 1.9       19.1 \u00b1 1.8         +0.87          0.035        21.1 \u00b1 1.6         +2.83          \\<0.001      +1.96            \\<0.001\n\n  Length gain, cm/week   1.04 \u00b1 0.36      1.13 \u00b1 0.33        +0.10          0.193        1.17 \u00b1 0.28        +0.14          0.194        +0.04            0.697\n\n  HC gain, cm/week       1.04 \u00b1 0.24      1.10 \u00b1 0.20        +0.05          0.258        1.10 \u00b1 0.24        +0.06          0.409        +0.01            0.937\n\n  Weight Z-score gain    --0.23 \u00b1 0.39    0.09 \u00b1 0.31        +0.31          \\<0.001      0.15 \u00b1 0.44        +0.38          0.003        +0.06            0.546\n\n  Length Z-score gain    --0.53 \u00b1 0.52    --0.36 \u00b1 0.45      +0.17          0.116        --0.14 \u00b1 0.50      +0.39          0.013        +0.22            0.114\n\n  HC Z-score gain        0.28 \u00b1 0.59      0.51 \u00b1 0.56        +0.23          0.068        0.70 \u00b1 0.41        +0.41          0.016        +0.18            0.252\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nDM = donor milk; POMM = pasteurized own mother's milk; ROMM = raw own mother's milk. Data are presented as mean \u00b1 standard deviation; *p* \\< 0.05 based on *t*-test.\n"}
+{"text": "After publication of this study \\[[@CR1]\\], we realize that Dr. Sorin Boeriu was not listed among the authors. Please see the corrected author list below:\n\nAdriana Monea\u2020, Gabriela Beresescu \\*\u2020, Sorin Boeriu, Mezei Tibor, Sorin Popsor and Dragos Mihai Antonescu\n\nThe online version of the original article can be found under doi:10.1186/s12903-015-0122-7.\n"}
+{"text": "Background\n==========\n\nThe Maastricht Treaty from 1992 marked the beginning of the health mandate of the European Union (EU) as enshrined today in Article 168 of the Lisbon Treaty (TFEU, Treaty on the Functioning of the European Union) \\[[@B1]\\]. The original EU health mandate focused primarily on stimulating cooperation between member states and supporting national actions (Art. 129 (1), Treaty of the European Union (TEU)) \\[[@B2]\\]. It embodied the Union with only limited legislative powers on health matters. Although this initial mandate was enhanced through subsequent Treaties, today Article 168, still gives the EU relatively circumscribed power in areas of public health (Art. 168 (4), TFEU). Healthcare continues to remain a national competence and in this regard, the EU *\"shall respect the responsibilities of the member states for the definition of their health policy and for the organization and delivery of their health services\"* (Art. 168 (7), TFEU). Despite the restricted Treaty-based mandate for health, the EU has a relevant role to play in national public health and health systems policies and has expanded its remit in areas beyond the Treaty \\[[@B3]\\]. Areas affected by EU provisions are extensively described in the literature \\[[@B4]-[@B11]\\]. To illustrate the main developments in the area of what can be called \"EU health policy\" a timeline is illustrated in Table\u00a0 [1](#T1){ref-type=\"table\"}. However, because of its limited legal mandate, some EU legal initiatives were highly contested \\[[@B12],[@B13]\\].\n\n###### \n\nTimeline of main developments in EU health policy\n\n  \u00a0          \u00a0\n  ---------- -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  **Year**   **EU health policy developments before the introduction of a legal EU health mandate**\n  **1957**   Treaty of Rome: health is not a priority. Two aspects are considered: social security of cross-border workers and occupational health.\n  **1971**   Regulation (EEC) 1408/71 on the application of social security schemes to employed persons and their families moving within the community (accompanied by implementing Regulation (EEC) No 574/72).\n  **1985**   EC launches the action programme \"Europe against Cancer\".\n  **1993**   Maastricht treaty: the legal basis for undertaking actions in the field of public health is defined in Article 3(o) and Article 129.\n  **Year**   **EU health policy developments after the introduction of a legal EU health mandate by the Maastricht Treaty**\n  **1995**   The European Agency for the Evaluation of Medicinal Products (EMEA), now European Medicines Agency (EMA), has been formed in London.\n  **1997**   Treaty of Amsterdam: Health Impact Assessment is implemented.\n  **1999**   Directorate General for Health and Consumers (DG SANCO) is established.\n  **2000**   Lisbon Agenda recognizes health protection as a prerequisite for economic growth measured with the indicator Healthy Life Years.\n  **2002**   The European Food Safety Authority (EFSA) has been established in Parma.\n  **2002**   First Programme of Community action in the field of public health (2003-2008) launched.\n  **2003**   The Tobacco Advertising Directive 2003/ 33/EC is adopted after the first version has been annulled by the European Court of Justice.\n  **2004**   Commission decision to set up an Executive Agency for the public health programme. It has the task to manage community action in the field of public health.\n  **2005**   The European Centre for Disease Prevention and Control (ECDC) in Stockholm is operational.\n  **2007**   White paper: Together for health: a strategic approach for the EU 2008-2013.\n  **2007**   Decision for a second Programme of Community action in the field of health (2008-13).\n  **2011**   Directive 2011/24/EU on the application of patients' rights in cross-border healthcare has been adopted.\n\nTherefore, one can pose the question of what has been achieved over the last twenty years. It may be argued that, despite its narrow legislative scope, the health mandate has triggered important European actions in certain public health areas like tobacco control \\[[@B14],[@B15]\\], infectious disease control \\[[@B16],[@B17]\\], European guidelines \\[[@B18]-[@B20]\\] and the development of an EU public health infrastructure \\[[@B21]\\]. In recent years, the EC has summarized in annual reports a diverse nature of key public health achievements such as communications and recommendations, health policies, EC co-financed actions and established networks (e.g. high level groups, scientific committees, platforms) \\[[@B22]-[@B24]\\]. However, stakeholders in the field provide examples indicating that public health relevant EU policies such as single EU policy assessments on the Common Agriculture Policy (CAP) \\[[@B25],[@B26]\\], pharmaceuticals \\[[@B27]\\], or the Health in All Policy (HiAP) approach \\[[@B28]\\] do not always meet the expectations of the public health community. In these papers the authors express concerns about potentially detrimental health effects \\[[@B25],[@B26]\\] or disappointment about the support of the policies and approaches aimed at improving health in Europe \\[[@B27],[@B28]\\]. Also, EU agencies such as the European Centre for Disease Prevention and Control (ECDC) are described as agencies with a limited legal mandate, competences and resources for EU public health but, at the same time, with promising prospects to develop as a renowned international player in the field \\[[@B29]\\].\n\nIn addition, an evaluation of the EU Health Strategy acknowledges its status as a guiding framework for EC health policies and joint EC and member state actions on health but also identifies the missing impact of the Strategy on other EC policies as well as on member state health policies and actions \\[[@B30]\\]. Evaluations of the EU public health programmes, which are one of the EC's financial instruments to implement its strategic health goals, criticize missing prioritization of topics, barriers for participation in projects for some member states and ineffective dissemination of project results \\[[@B31],[@B32]\\]. Hence, the available evidence of the impact of EU health policies, infrastructure, and actions is elusive, and the identification of the value of public health relevant EU-level actions across all policies is lacking.\n\nIn this paper, we aim to explore and provide an overview of influential public health relevant EU-level policy outputs and a summary of policy outputs or actions perceived as an achievement, a failure or a missed opportunity by interviewing key experts in the field. By this, we intend to establish a qualitative indication of which EU health policies have contributed to the improvement of population health in Europe.\n\nMethods\n=======\n\nThe study focused on the evolvement of the health mandate since 1992, the year the Maastricht Treaty was signed. The study was carried out in two consecutive phases: (1) qualitative interviews, suitable to identify expert perceptions, and (2) voting on influential and public health relevant EU policy outputs and actions based on nominal group technique. The study adhered to the RATS guidelines on qualitative research \\[[@B33]\\].\n\nStudy participants\n------------------\n\nExperts were purposely selected to ensure heterogeneity of opinions. The selection was based on their individual profile and professional affiliation. We selected experts that were renowned in the field due to their current or former affiliation to specific EU-level bodies and institutions, research institutes with EU focus, or EU-level non- governmental organization. Selected experts were actively involved in public health research, policy-making, policy advice or advocacy performed at EU level, internally or externally. In addition, snowball sampling was applied until data saturation was reached. Data saturation was assumed as soon as no new EU public health policy actions and their perceptions were mentioned during the interviews.\n\nThe potential participants were contacted between December 2011 and March 2012 by a short information email to identify whether they were interested to participate in the interview study. Of 22 contacted experts, twenty participated in this study, one participant could not confirm participation due to time constraints and another did not respond to the invitation. Of those twenty experts nine belonged to the initial purposively selected sample and eleven were identified during the snowball sampling procedure based on recommendations of already interviewed experts. The majority of experts was affiliated to an institution located in Brussels (n\u2009=\u200910). The composition of the study sample in terms of represented professional affiliations is outlined in Table\u00a0 [2](#T2){ref-type=\"table\"}.\n\n###### \n\nProfessional functions of interviewed participants\n\n  ***Function***                   ***Number of participants***\n  -------------------------------- ------------------------------\n  EU/national civil servant        7\n  EU/national politician           2\n  Academia                         5\n  Public health advisor/advocate   6\n  **Total**                        **20**\n\nUpon agreement by the participant, an appointment for the interview was made and participants received an informative letter with more in-depth information about the goal of the study and an informed consent form in which the voluntary basis of the participation has been clarified and anonymized data handling was assured.\n\nQualitative interviews\n----------------------\n\nInterviews were conducted either face-to-face (n\u2009=\u20094) or via telephone, or Voice over IP (n\u2009=\u200916) in the period between January and March 2012 and were held in English, Dutch or German by one of the three principal investigators (NR, TC, KS). The interviews lasted from 35 to 90\u00a0minutes. All interviews were audio-recorded, transcribed verbatim and anonymized.\n\nThe interviews were performed using a specifically designed semi-structured interview guide. The guide was developed on the basis of previous desktop research and an internal brainstorming session of an advisory research group consisting of the three principal investigators, four senior researchers, and one junior researcher of the Department of International Health at Maastricht University to identify items relevant for investigating expert perceptions on European public health policy. During the course of this process, a list of public health relevant EU policy outputs, processes or procedures that were regarded as achievements, failures or missed opportunities were first gathered individually and then, following a group discussion, a common list was compiled. This output was used to construct the interview guide containing six guiding themes from which open-ended questions were formulated. The guiding themes included (1) a description of the individual role of the expert in European public health, (2) the individual definition of European public health, (3) the assessment of public health relevant EU-level actions as being an achievement, missed opportunity or failure, (4) the formulation of five influential European policy outputs, (5) consequences of European health policy, and (6) the policy process at the European level. The semi-structured interview guide was used as a framework for the interviews and allowed the interviewers to address other relevant topics that emerged during the interviews.\n\nAnalysis\n--------\n\nAfter completion of interviews the three principal investigators initially performed an internal analysis of each separate interview and an analysis across interviews to identify the scope of EU-level actions and experts' perceptions of these actions as achievements, failures or missed opportunities \\[[@B34]\\]. Afterwards, a directed (or deductive) content analysis approach \\[[@B35],[@B36]\\] was applied whereby the initially predefined coding scheme with the main categories of interest (achievement, failure, and missed opportunity) was used to summarize the respective topics and the reasoning that appeared during the interviews. Topics that did not fit into one of these main categories were added as new codes and were organized into new categories. The analysis was jointly performed by the principal investigators using NVivo 9 (QSR International Pty Ltd. Version 9).\n\nFurthermore, the results of the content analysis on achievements, failures and missed opportunities were grouped according to the major common themes in a table to provide an overview of the perceptions of the key informants. Prominent EU-level outputs or actions, which were discussed by almost all respondents, are described in more detail in the results section. Where applicable, we used original quotes to illustrate the views and tendencies of experts' assessments. The professional profile and study ID of respondents are indicated behind the respective quotes. Quotes which were originally given in Dutch or German were translated into English.\n\nNominal group technique\n-----------------------\n\nA slightly adapted nominal group technique \\[[@B37]\\] was used for triangulation purposes. During the interviews, the participants were asked for five influential policy outputs of European health policy-making. Following the finalization of all interviews, all participant nominations were compiled in one list and reoccurring topics were removed. To ensure comparability of policy outputs and actions, we grouped the nominations into categories under the following headlines: (a) secondary legislation and court decisions; (b) soft laws, strategies, and programmes; (c) agencies, centres, organizations; (d) networks, policy platforms, cooperation; and (e) others. Participants were asked in an online survey to select three outputs per category which were, according to their opinion, most influential. Based on the participants' nominations, a ranking in terms of a frequency distribution of selected influential policy outputs for each category was determined. The online survey was completed by 18 out of 20 participants who took part in the interviews.\n\nEnsuring quality\n----------------\n\nThe design and analysis of the study was guided by applying Guba and Lincoln's test for trustworthiness \\[[@B38]\\]. Credibility and dependability have been ensured by enlarging the sample until saturation was reached in terms of the identification of EU policy actions and their perceptions. Moreover, three researchers in the primary research group in combination with an internal advisory research group were involved with the aim of reflecting upon the study design and critically questioning the findings. Additionally, the primary research group met regularly during the interview period to exchange initial findings and experiences on the interview process. The members of the internal advisory research group were experienced in EU public health policy research or qualitative research methodologies. The confirmability was strengthened by the use of several investigators both in the data collection process and in the analysis phase, combined with the use of triangulation, where interview participants were also asked to participate in the ranking exercise.\n\nEthical considerations\n----------------------\n\nThe Medical Ethical Committee of the University Hospital Maastricht and University Maastricht declared that no ethical approval was required for this type of research. All participants were informed about their role and rights as study participant prior to their interview participation. All participants provided written or audio-recorded informed consent to be interviewed.\n\nResults\n=======\n\nOverall, respondents consistently mentioned that, during the twenty-year history of the EU health mandate, specific initiators induced change in European public health policy. The most important identified initiators included the Maastricht Treaty with its later amendments, the health-related rulings of the European Court of Justice, and the health crises such as Boviene spongiforme encefalopathie (BSE) and Severe Acute Respiratory Syndrome (SARS). In addition, the internal market provisions with the foreseen free movement of goods, people, services and capital, initiated change with both negative and positive public health impact. Additionally, a set of conditions was identified in the interviews that described and advanced the role of EU health policy as a reference point for public health. These conditions under which EU health policy made progress during the past twenty years were (i) the regulatory power at the EU-level, (ii) EU-led facilitation of cooperation and comparisons across member states; along with (iii) increased capacity building on EU issues and on EU-level (e.g. professionalization, development of interest groups, associations).\n\nInfluential EU health policy outputs\n------------------------------------\n\nThe ranking of influential policy outputs of EU-level health policy-making is provided in Table\u00a0 [3](#T3){ref-type=\"table\"}.\n\n###### \n\nRanking of influential policy outputs of EU-level public health policy, March/April 2012\n\n  **Categories**                                                                   **N**\n  -------------------------------------------------------------------------------- --------\n  **Secondary legislation and court decisions**                                    \n  \u2002**Patients\\' rights decisions by the European court of justice**                **12**\n  \u2002**Directive on patients' rights in cross-border care**                          **10**\n  \u2002**Directive on tobacco advertising**                                            **9**\n  \u2002Directive on blood safety                                                       5\n  \u2002WHO Framework Convention on Tobacco Control                                     4\n  \u2002Directive on professional qualifications                                        3\n  \u2002Regulation 1408/71                                                              3\n  \u2002Directives on tobacco products                                                  2\n  \u2002REACH regulation of chemical substances                                         2\n  **Soft laws, strategies and programmes**                                         \u00a0\n  \u2002**First and second public health programme**                                    **9**\n  \u2002**Together for health: 2008-2013**                                              **6**\n  \u2002**Framework for action in the field of public health**                          **5**\n  \u2002**Towards modern, responsive and sustainable health systems (2011/C 202/04)**   **5**\n  \u2002**Europe 2020, Europe's growth strategy**                                       **5**\n  \u2002Health in All Policies                                                          4\n  \u2002Information to Patients                                                         3\n  \u2002Open Method of Coordination                                                     3\n  \u2002Solidarity in health: reducing health inequalities                              3\n  \u2002Council recommendation on patient safety                                        **2**\n  \u2002Framework Programmes on health research                                         2\n  \u2002White paper on governance                                                       2\n  \u2002Green paper on the European workforce for health                                0\n  **Agencies, centres, organizations**                                             \u00a0\n  \u2002**European Medicine Agency**                                                    **15**\n  \u2002**European Centre for Disease Prevention and Control**                          **13**\n  \u2002**European Food Safety Authority**                                              **9**\n  \u2002**European Observatory on Health Systems and Policies**                         **6**\n  \u2002European Agency for Accreditation of Education in Public Health                 1\n  \u2002European Chemicals Agency                                                       1\n  \u2002European Foundation for the Improvement of Living and Working Conditions        1\n  \u2002European Monitoring Centre for Drugs and Drug addiction                         1\n  \u2002Social Protection Committee                                                     1\n  \u2002European Commission Food and Veterinary Office                                  0\n  \u2002European Commission Joint Research Centre institutes                            0\n  \u2002Health Intergroup at the Committee of the Regions                               0\n  **Networks, policy platforms, cooperation**                                      \u00a0\n  \u2002**European presidencies**                                                       **13**\n  \u2002**Cooperation EC with OECD and WHO**                                            **8**\n  \u2002**European Health Policy Forum**                                                **6**\n  \u2002**Network epidemiological surveillance and control of infectious diseases**     **6**\n  \u2002**Network on Health Technology Assessment**                                     **6**\n  \u2002European Platform for Action on Diet, Physical Activity and Health              3\n  \u2002eHealth Network                                                                 2\n  \u2002European involvement in GATS discussion                                         2\n  \u2002European Committee of Experts on Rare Diseases                                  2\n  \u2002European Platform against Poverty and Social Exclusion                          1\n  \u2002European Innovation Platform on Active and Healthy Ageing                       0\n  \u2002European Network of Health Promoting Schools                                    0\n  \u2002Joint Action Health Workforce Planning                                          0\n  **Others**                                                                       \u00a0\n  \u2002**Work on health determinants**                                                 **10**\n  \u2002**Best practice exchange**                                                      **8**\n  \u2002**Scientific reports that influenced European policy-making**                   **7**\n  \u2002Budget of the European Union                                                    5\n  \u2002Common Agriculture Policy reform                                                4\n  \u2002Alcohol regulation                                                              3\n  \u2002HEIDI data tool                                                                 2\n  \u2002Initiatives on Roma and excluded groups                                         1\n  \u2002Project database on DG SANCO website                                            1\n  \u2002Work on public health core competencies by ASPHER                               1\n\nASPHER: Association of Schoold for Public Health in the European Region.\n\nDG SANCO: Directorate General for Health and Consumer Protection.\n\nEC: European Commission.\n\nGATS: General Agreement on Trade in Services.\n\nHEIDI: Health in Europe: Information and Data Interface.\n\nOECD: Organization for Economic Cooperation and Development.\n\nWHO: World Health Organization.\n\nIn the category \"secondary legislation and court decisions\", the patients' rights decisions made by the European Court of Justice (n\u2009=\u200912) were chosen by most of the respondents as influential policy output, followed by the Directive on the application of patients' rights in cross-border healthcare \\[[@B39]\\] (n\u2009=\u200910) and the Directive on advertising and sponsorship of tobacco products \\[[@B14]\\] (n\u2009=\u20099). In the category entitled \"soft laws, strategies, and programmes\", the first and second EU public health programmes \\[[@B40],[@B41]\\] (n\u2009=\u20099) were selected most frequently, followed by the 2008--2013 health strategy \"Together for Health\" \\[[@B42]\\] (n\u2009=\u20096). The third rank is shared by three policy outputs: the \"Framework for action in the field of public health\" \\[[@B43]\\] (n\u2009=\u20095) which is the Commission's first proposal setting out EU-level public health after the introduction of the health mandate in the Maastricht Treaty, the Council conclusions \"Towards modern, responsive and sustainable health systems\" \\[[@B44]\\] (n\u2009=\u20095), and the current over-arching European strategy \"Europe 2020\" \\[[@B45]\\] (n\u2009=\u20095). In the third category on \"agencies, centres and organizations\", the European Medicines Agency (EMA, n\u2009=\u200915) ranked top, followed by the ECDC (n\u2009=\u200913) and the European Food Safety Authority (EFSA, n\u2009=\u20099). Among \"networks, policy platforms and collaborations\", the European presidencies (n\u2009=\u200913) were selected most often by the respondents, followed by the collaboration of the European Commission (EC), the World Health Organization Regional Office for Europe (WHO-EUR) and the Organization for Economic Co-operation and Development (OECD) (n\u2009=\u20098). Moreover, the three entries on the third rank include the EU Health Policy Forum (n\u2009=\u20096), the network on epidemiological surveillance and control of infectious diseases (n\u2009=\u20096), and the network on Health Technology Assessment (n\u2009=\u20096). The fifth category was not topic specific therefore, work on European level health determinants (n\u2009=\u200910), the exchange of best practices (n\u2009=\u20098), and published scientific reports which influenced EU policy-making (n\u2009=\u20097) were ranked on the first three positions.\n\nAchievements, missed opportunities, failures\n--------------------------------------------\n\nAt a glance, the label \"achievement\" was allocated to the public health mandate as it is laid down in the Treaties, the establishment of EU-level agencies dealing with public health topics and successes in smoking prohibition, food safety and infectious disease control. The label \"missed opportunity\" was allocated to the insufficient degree to which the HiAP approach is implemented and the ways in which health promotional aspects of alcohol and nutrition were handled. The label \"failure\" was less often assigned with the missing integration or link to social policies appearing in some interviews under this heading as well as the strength of the internal market which annulled national protective alcohol legislations in some member states. In Table\u00a0 [4](#T4){ref-type=\"table\"}, we provide the full list of EU-level outputs or actions which, based on the content analysis and the identified thematic categories, were mentioned as achievements, missed opportunities or failures by the key informants. Due to a broad and divergent spectrum of perceptions, topics almost always shared aspects of achievements, missed opportunities or failures. In the following section, we focus on those EU-level outputs or actions which were mentioned by the majority of respondents during the interviews and allowed us to draw a comprehensive picture on the breadth and the diversity of expert perceptions.\n\n###### \n\nSummary of public health relevant EU-level actions and their perception as achievement, failure or missed opportunity\n\n  **Topics**                                                                                                      **Achievement**                                                                                                                                                              **Missed opportunity**                                                                                                                            **Failure**\n  --------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------\n  **Treaty**                                                                                                      Inclusion of the health mandate as enshrined in Article 129 of the Treaty of the European Union.                                                                             Missing implementation of a connection/share of power between economic and social EU policy.                                                      \u00a0\n  **Directorate general for health and consumers (DG SANCO)**                                                     Existence and persistence of DG SANCO.                                                                                                                                       DG SANCO is not strong enough to push health in other DGs.                                                                                        DG SANCO set-up: Missing link to social policy.\n  Public health programme.                                                                                        Public health did not become a key aspect of EU policy.                                                                                                                                                                                                                                                                        \n  Sustainability development strategy: DG SANCO is not playing an active role in the marketing of the strategy.                                                                                                                                                                                                                                                                                                                                  \n  **Cooperation**                                                                                                 Cooperation between EU, WHO and OECD.                                                                                                                                        Missing connection and joint forces between EC and WHO.                                                                                           \u00a0\n  **EC agencies**                                                                                                 Development of agencies: ECDC, EFSA, EMA, EMCDDA.                                                                                                                            \u00a0                                                                                                                                                 \u00a0\n  \u2002**ECDC**                                                                                                       Legislation on infectious disease control.                                                                                                                                   ECDC mandate should include responsibilities in risk management of infectious diseases.                                                           \u00a0\n  European Programme for Intervention Epidemiology Training (EPIET).                                              ECDC profile should cover also non-communicable diseases and SDoH.                                                                                                                                                                                                                                                             \n  \u2002**EMA**                                                                                                        Coordination of the approval of efficacy, safety and quality of drugs.                                                                                                       Cost-effectiveness of pharmaceuticals is not taken into account.                                                                                  Problem of not being able to tackle pharmaceutical pricing.\n  Reversal of the approval of already approved drugs not handled on EU-level.                                                                                                                                                                                                                                                                                                                                                                    \n  \u2002**EFSA**                                                                                                       Control of health claims of food products.                                                                                                                                   EFSA mandate should include/be stronger on health promotion aspects of nutrition (e.g. regulation of advertisement of unhealthy food products).   \u00a0\n  Food safety Directive.                                                                                          \u00a0                                                                                                                                                                                                                                                                                                                              \n  **Health in All Policies (HiAP) approach**                                                                      Health mandate assures that health protection should be guaranteed in all EU policies.                                                                                       HiAP and Health Impact Assessment have never been implemented fully (tick box exercise).                                                          \u00a0\n  Leads to the discussion of health in other sectors.                                                             \u00a0                                                                                                                                                                                                                                                                                                                              \n  **Lifestyle factors**                                                                                           Common tobacco legislations in Europe (WHO Framework Convention of Tobacco Control; tobacco product-; tobacco advertising Directive).                                        The tobacco regulations could have been designed stronger (e.g. more harmonized realisation of smoking prohibition on public places).             Tobacco regulation has some aspects of failure since a strict, general ban is not reached.\n  Food safety measures and regulations on health claims.                                                          Missing political will to tackle obesity and related life style factors like unhealthy food products.                                                                        \u00a0                                                                                                                                                 \n  **Health Research Programme**                                                                                   EU health research budget and outcomes of the programme.                                                                                                                     \u00a0                                                                                                                                                 Missing integration of the research programme and EU health research outcomes in public health.\n  **EU budget**                                                                                                   Largest budget proportion shifted in the Multiannual Financial Framework 2007--2013 from agriculture financing to the funding of cohesion and sustainable growth policies.   \u00a0                                                                                                                                                 \u00a0\n  Health research budget.                                                                                                                                                                                                                                                                                                                                                                                                                        \n  The use of Structural Funds for investments in health (2007--2013).                                                                                                                                                                                                                                                                                                                                                                            \n  **Internal market provisions**                                                                                  \u00a0                                                                                                                                                                            Internal market rules as source for legislation should be more attentive to health concerns.                                                      Internal market provisions cause problems if member state regulation is more protective regarding health threats than EU regulation.\n  **Patients' rights directive**                                                                                  The patients' rights Directive in general.                                                                                                                                   Negotiations on patients' rights Directive failed to include a strong emphasis on the development of common standards.                            \u00a0\n  Effect on cross-border cooperation.                                                                                                                                                                                                                                                                                                                                                                                                            \n  Gives legal certainty to policy makers.                                                                                                                                                                                                                                                                                                                                                                                                        \n  **Common Agriculture Policy**                                                                                   Policy field which starts to recognize health, e.g. in its white paper on the CAP after 2013 (2009/2236(INI)).                                                               Unrecognized potential for health of the CAP by public health sector.                                                                             \u00a0\n  **Health information**                                                                                          Health life years as indicator in the Lisbon strategy.                                                                                                                       \u00a0                                                                                                                                                 Missing health information system.\n  Lack of morbidity data.                                                                                                                                                                                                                                                                                                                                                                                                                        \n  **Different public health topics**                                                                              \u00a0                                                                                                                                                                            \u00a0                                                                                                                                                 \u00a0\n  \u2002**health inequalities**                                                                                        EC communication: solidarity in health: reducing health inequalities in the EU.                                                                                              \u00a0                                                                                                                                                 \u00a0\n  \u2002**HTA**                                                                                                        Strengthening of the HTA approach in the EU.                                                                                                                                 Coordinating cross-country level health technology assessments.                                                                                   \u00a0\n  \u2002**Rare diseases**                                                                                              Coordinated management of rare diseases.                                                                                                                                     \u00a0                                                                                                                                                 \u00a0\n  \u2002**Tuberculosis**                                                                                               \u00a0                                                                                                                                                                            \u00a0                                                                                                                                                 Existing drug resistance of tuberculosis as indicator for lacking disease management.\n  \u2002**Health of minorities**                                                                                       Health of minorities (e.g. Roma) as part of the European agenda.                                                                                                             \u00a0                                                                                                                                                 \u00a0\n  \u2002**Social care**                                                                                                \u00a0                                                                                                                                                                            Social care is hardly seen as EU competence.                                                                                                      \u00a0\n  \u2002**Environment (and health)**                                                                                   Environmental standards set by the EU.                                                                                                                                       Missing follow-up process on the Environment and Health Action Plan (2004--2010).                                                                 \u00a0\n  \u2002**Information to patients**                                                                                    Blocking of direct to consumer advertising of prescription-only pharmaceuticals.                                                                                             \u00a0                                                                                                                                                 \u00a0\n  **Governmental issues**                                                                                         White paper on governance (2001) increased transparency.                                                                                                                     \u00a0                                                                                                                                                 \u00a0\n  More standardisation of methods (evaluation of indicators, outcomes, policies) and common language.                                                                                                                                                                                                                                                                                                                                            \n  Increased understanding of the public health community about the impact of EU policies on public health.                                                                                                                                                                                                                                                                                                                                       \n  \u2002**Industry involvement**                                                                                       \u00a0                                                                                                                                                                            \u00a0                                                                                                                                                 Cooperation with industry influences the health research agenda and policy-making.\n  Evidence-based policy-making: the interest of the industry is against public health.                                                                                                                                                                                                                                                                                                                                                           \n\nECDC: European Centre for Disease Prevention and Control.\n\nEFSA: European Food Safety Authority.\n\nEMA: European Medicines Agency.\n\nEMCDDA: European Monitoring Centre for Drugs and Drug Addiction.\n\nHTA: Health Technology Assessment.\n\nDG: Directorate General.\n\nSDoH: Social Determinants of Health.\n\nEC: European Commission.\n\nOECD: Organization for Economic Cooperation and Development.\n\nWHO: World Health Organization.\n\nEU health policy - an overview\n------------------------------\n\nAn assessment of the general value of EU-level public health actions over the last twenty years resulted in mainly ambivalent judgments. On the one hand, many relevant activities were performed at an EU-level and the existence of a health mandate contributed to an EU social model. On the other hand, its dependence on political will and economic circumstances influenced the development of EU-level public health policy and led to the perception that more should or could have been achieved within and beyond the possibilities of the current health mandate.\n\nThe role of the Directorate General for Health and Consumers\n------------------------------------------------------------\n\nThe establishment of the Directorate General for Health and Consumers (DG SANCO) in 1999 as an independent, formal structure for EU health policy was generally discussed as an achievement. The formation of DG SANCO, and thus, the political decision to separate the health dossier from DG V, the former DG with the responsibility for health policy as well as a focus on employment and social policies, was controversially perceived. The establishment of DG SANCO led, on the one hand, to a more mature health policy field.\n\n\"...the DG V was a big DG and then DG SANCO became separate from that. Health had its own commissioner, its own opportunity to protect itself and public health benefits.\" (\\#15, public health advisor/advocate)\n\nOn the other hand, aspects of failure were mentioned regarding the detachment of health and social policy at EU-level. According to the respondents being separated led to a loss of collaboration for more holistic health policies and actions in health systems and healthcare at EU-level.^a^\n\nFollowing the formation of DG SANCO, it was seen as a beneficial way forward for the DG to shift its sectorial policy approach from a focus on specific topics such as cancer, drug dependence, health monitoring, accidents and injuries, or pollution-related diseases, to a horizontal one with the formulation of the first health strategy with three cross-cutting objectives: health information, health threats and health determinants \\[[@B46]\\].\n\n\"And this was an important moment in time where the sectorial approach to AIDS, cancer and other issues has been reduced gradually and that more the integral horizontal approach, which was applied at that time already in all member states -- hence Europe was running behind in that sense, but ultimately was embraced and taken as guideline for the framing of all sorts of public health actions. \"(\\#02, EU/national civil servant)\n\nSince it fostered more visibility of the public health field and closer cooperation by financing projects, joint actions and research across Europe, the Public Health Programme of DG SANCO was commonly discussed as being supportive to the development of European public health and the mobilization of the public health community.\n\nAspects of missed opportunity became relevant when assessing the representation of health in other EU policies.\n\n\"I don't know what exactly the reason is, but they \\[DG SANCO\\] are not strong enough to push for health in \\[the other\\] DGs. The obvious example is the latest EU 2020 strategy, you cannot find reference to health anywhere it's really a disaster, because of \\[the\\] weak DG SANCO. Health is not among the headline targets, it's not among the flagships.\" (\\#15, public health advisor/advocate)\n\nWhile the cooperation with other DGs was recurrently discussed as problematic, the potential for the \"Partnership on Active and Healthy Ageing\" under the European Innovation Union appeared as a unique theme and was regarded as an achievement for strengthening health policy on the general EU policy agenda.\n\n\"...it is now coordinated under the Innovation Union \\[of Europe2020\\], and so Commissioner Neelie Kroes \\[DG Connect and Vice-Commissioner\\] is as much in the lead as Commissioner John Dalli \\[DG SANCO\\]. And that's a good sign, if we can get more of those sorts of partnerships on specific policies, then I think, we'll get a better understanding.\"(\\#04, EU/national politician)\n\nWith regard to assessing the status of DG SANCO cooperation with other international policy actors, respondents had mixed perceptions. Whereas some argued that DG SANCO's collaboration with international organizations like the WHO-EUR or the OECD is improving and therefore, can be considered as an achievement, others asserted that this collaboration was not sufficiently established and can therefore be categorized as a missed opportunity.\n\nThe role of health agencies: the example of ECDC\n------------------------------------------------\n\nThe establishment and the work of EU public health agencies like the EMA, the ECDC, the EFSA, and the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) were regarded as an achievement and as an important step forward towards the strengthening of the European dimension in health. The work and the scope of the agency mandates was a recurring topic and subject to diverging perceptions. As an example, in the case of the ECDC, its development was assessed as an achievement whilst its scope was considered a missed opportunity.\n\nThe bioterrorism attacks on the United States of America in 2001 and the SARS crisis led to calls for better international coordination of infectious disease surveillance and the establishment of ECDC in 2004 \\[[@B47]\\]. Hence, the setting up of ECDC was commonly perceived as an achievement, since it gave preceding EU actions in infectious disease control a formal structure and maintained actions in the field. Also, the close collaboration with the respective national public health agencies during outbreaks and in negotiating and developing common guidelines for infectious disease control were regarded as an important task of the ECDC.\n\nHowever, a number of respondents were critical of the scope of the ECDC mandate and thus, looked at this as a missed opportunity. Questions were raised on whether the ECDC's responsibility in surveillance, risk assessment and training are sufficient or if additional responsibilities in risk communication and management were needed to assure full stewardship during and in the prevention of health crises.\n\n\"I suppose the flu epidemic \\[...\\]. That should be put on the table not only as a missed opportunity, big failure, having put ECDC at the center of the development, but the ECDC is not authorized to risk communicational management as you know. So, in that sense, it is a failure that member states were not able to coordinate in this very important public health area and use the EU institution, either ECDC or WHO to do that.\" (\\#03, public health advisor/advocate)\n\nMoreover, interview participants reported tensions between member states and EU agencies regarding the transfer of responsibilities from national to EU-level.\n\n\"And the member states are very reluctant to hand over power regarding public health to the Commission, or to Brussels. Now if you focus on infectious diseases, that is much better because they understand that there is a need, but again it is not easy.\" (\\#30, EU/national civil servant)\n\nSince the largest burden of disease in the EU is caused by non-infectious rather than infectious diseases, a call was put forward to further increase the mandate of ECDC to all public health relevant aspects and not focus only on infectious diseases.\n\nHealth in All Policies (HiAP)\n-----------------------------\n\nThe HiAP approach was generally assessed as an achievement regarding its potential to address health determinants outside the health sector.\n\n\"\\[the article on the health mandate\\] is very important, because thereby a mandate is created that the Commissioner for Health and Consumer Affairs \\[...\\] approaches his colleagues whenever they make new legislation to ensure that the health protection dimension is guaranteed; it gives partly a mandate to break into the policy and law development in sectors which in principle do not have any links with public health. \\[...\\] this is very difficult. But its potential is very strong.\" (\\#2, EU/national civil servant)\n\nHowever, in regard to its degree of implementation participants commonly perceived HiAP as a missed opportunity. Health Impact Assessment, the implementing tool to HiAP, was regarded as a *\"tick box exercise\" (\\#03, \\#15, both public health advisor/advocate)* rather than a thorough consideration of health in other policies areas. Explanations given during the interviews demonstrated that conditions to achieve HiAP seemed not to be established yet and that there seems to be difficulty in bringing DG SANCO interests in line with the interests of other DGs without over-emphasizing the health aspect. Political assertiveness in convincing other Commissioners and DGs about the relevance for intersectoral cooperation was perceived to be lacking, even though an Inter-service group on Public Health with the participation of more than twenty EC departments was established for this purpose.\n\nApproach to life style factors\n------------------------------\n\nGenerally, the work regarding tobacco was regarded as an achievement of how European health policy-making effectively addressed a life style risk factor for health.\n\n*\"The progress around tobacco \\[Directive on tobacco advertising, Directive on tobacco products, transparency register\\], the fact that we have a piece of international law on tobacco \\[WHO Framework Convention on Tobacco Control\\] is massive and that was European led*.\" *(\\#05, public health advisor/advocate)*.\n\nThis quote echoed the perception of the majority of respondents who emphasized the leading role of the EU regarding the support and commitment to the WHO Framework Convention on Tobacco Control. Moreover, it was argued that the achievements regarding the regulation of tobacco advertising and smoking prohibition in public places would not have been achieved by single member states independently and thus this was a common achievement initiated and supported by European cooperation. Nevertheless, aspects of a missed opportunity or even failure were mentioned in this regard since some would have appreciated stronger legislative measures to achieve a more harmonized realization of smoke-free legislation across the EU.\n\nIt was considered that the achievements recognized in EU tobacco legislation were missed in the regulation of other health-related life style factors such as nutrition and alcohol. Whilst regulations in the area of food safety were generally acknowledged as an achievement by preventing food-borne health threats; a potential mandate to address the composition of food and thereby, prevent, *inter alia* obesity or non-communicable diseases seemed to be neglected and was labeled as a missed opportunity.\n\n\"...food safety has been majorly put forward over the last twenty years, in the sense that we know that the food will not be contaminated. But then it is a missed opportunity in the sense that beyond food safety there is health promotion and then one wanted the Union to have more powers to regulate issues on the content of saturated fat for instance or the percentages of sugar and so on.\" (\\#03, public health advisor/advocate)\n\nWith regard to governmental activity on these issues, the EU Platform for Action on Diet, Physical Activity and Health was named as an example of an achievement as well as a failure.\n\n\"I think the diet platform \\[...\\] can be seen as a failure and opportunity. \\[...\\] If we had not created that platform then arguably the issue wouldn't have been tackled at all. And in a way that has been really brought some issues of complexity to the political discussion around issues around marketing of food, around self-regulation, reformulation, some of the initiative like salt in diet has come as a commitment from that platform.\"(\\#05, public health advisor/advocate)\n\nThe failure aspect of the EU Platform for Action on Diet, Physical Activity and Health was related to the perception that a platform is a rather weak policy instrument and that more political will to tackle these issues with stronger EU policy or legal instruments would have had more impact.\n\nAdditionally, the lack of timely cooperation of public health professionals with other sectors such as agriculture was raised as missed opportunity. It was illustrated that agriculture policy has public health relevant links regarding affordability, accessibility, and the availability of food. However, it was also argued that this cooperation has been developed further over the recent years.\n\n\"And it is correct, that the Common Agriculture Policy has not been taken up health in the beginning, but by now they are doing this very consciously. \\[...\\] Thus, I really see an improvement; I actually do not see a situation anymore in which health was influenced really negatively \\[by the Common Agriculture Policy\\].\" (\\#26, EU/national civil servant)\n\nInternal market provisions\n--------------------------\n\nInternal market provisions were perceived as ambivalent by the respondents. The EU is based on internal market rules that also affect EU health policy.\n\n\"The engine of European health policies is still the market.\"(\\#23, academia)\n\nHowever, the influence of EU market regulations, for example on alcohol policies, was perceived as a failure when member states had more protective and stricter national legislation as was the case in the Nordic countries. Respondents claimed that EU internal market regulations that are more attentive to health issues would be appreciated in this case. Moreover, the potential given by Articles 36 and 114 of the TFEU, which put limitations on the Single Market, was mentioned and it was perceived as a missed opportunity that this potential had not been fully taken up by public health experts:\n\n\"\\[...\\] the public health aspect, which is written into Article 95 \\[now Article 114, TFEU\\] on the internal market, you can put limits on the internal market on the grounds of public safety, public morality and public health, is almost never used. What if DG Internal Market was turned into our greatest weapon?\" (\\#32, public health advisor/advocate)\n\nThis was positively exemplified by the case of tobacco control which applied internal market rules for public health purposes to assure harmonized labeling, packaging, nicotine content, etc. across the EU. However, the application of the health argument to put limitations on the internal market rules was also perceived as being negatively connoted by non-public health experts:\n\n\"If you just go to the DG Internal Market and grab the first person you see and ask them what public health means, they will tell you it's the exception member states use to defend local weird monopolies on peculiar alcohol, or something like that. It's an exception to a rule.\" (\\#19, academia)\n\nPatients' rights Directive\n--------------------------\n\nThe recent EU patients' rights Directive in cross-border healthcare \\[[@B39]\\] was mainly regarded as an important achievement. This assessment was not necessarily driven by satisfaction with the scope of the Directive but, instead, because it is the first EU secondary legislation ever enacted specifically on healthcare.\n\n\"\\[...\\] the cross-border Directive will turn out to be incredibly important. Particularly because it is so symbolic important if you like because it does represent really the first time that the EU has got any concrete in relation to healthcare as opposed to public health. The consequences of this remains to be seen.\" (\\#29, EU/national civil servant)\n\n\"Therefore, I see the patients' Directive as a true success from a legislative perspective\" (\\#23, academia)\n\nThe achievement aspect was supported by perceptions that the Directive will lead to more cross-border cooperation and will have an impact on quality of care as well as on priority setting in healthcare and the packaging of healthcare services. Thereby, it was expected that the Directive will not only influence people who seek healthcare services in other countries but also those who seek services in their home country. In this regard, some expected that the Directive would also ultimately empower patients as consumers of healthcare services.\n\n\"The cross-border Directive \\[...\\] will have consequences of more consumer empowerment, consumer rights, patient rights, more consumer participation and more literacy,...\"(\\#03, public health advisor/advocate)\n\nHowever, there were also critical voices that interpreted the Directive, as targeting a limited segment of the European population and hence, potentially increasing health inequalities. These respondents also questioned the willingness of the general population to seek healthcare treatment outside their home country. Furthermore, respondents were critical of the extent to which more EU involvement in healthcare of member states would lead to quality assurance in general:\n\n\"It is positive in the way people can be treated where they want, but it is still my point of view that we \\[...\\] want to have our own level of quality and we don't want others to decide what level it should be. Perhaps, because we have a very high quality \\[...\\]. But of course we don't mind to tell others about it, we don't mind others to come in, we don't mind to help others to get the same standard - that is cooperation, so I always say I love cooperation but I do mind the harmonization.\". (\\#18, EU/national politician)\n\nDiscussion\n==========\n\nThis study provides an overview of public health relevant EU-level actions of the past twenty years. We outlined the diverse nature of expert perceptions on key developments in the field and provided a ranking of the most influential achievements. The assessment of outputs or actions being an achievement appeared across and within interviews along with assessments of outputs or actions being a missed opportunity and less often a failure. Thereby, it turned out that the EU public health field has significantly developed its organizational structures (DG SANCO, supranational agencies dealing with public health) and incorporated public health topics like infectious disease control and tobacco control, whereas the HiAP approach still included untapped potential. This finding confirms \"The Challenge of Implementation\" \\[[@B48]\\] of the HiAP concept in the EU \\[[@B28],[@B49]\\]. Given the fact that according to Article 9 and Article 168 (1),TFEU \\[[@B1]\\], a high level of human health protection should be ensured within all EU policies and actions, it was seen as a weakness that the uptake of health consideration in the general EU policy-making process was low \\[[@B28]\\]. Ollila described the importance of communication and cooperation strategies for a successful realization of the HiAP approach \\[[@B49]\\]. The deficiency of these strategies was raised during the interviews which indicated that the performance of EU health players is perceived to be particularly poor in this regard.\n\nConcordance of interview responses with tasks formulated in the health mandate of the EU\n----------------------------------------------------------------------------------------\n\nInterestingly, the study indicated that the Treaty-based tasks such as support of cooperation between member states, development of guidelines and indicators, best practice exchange, and periodic monitoring and evaluation on EU-level public health to ensure *'a high level of human health protection'*\\[[@B1]\\] were only partially perceived as fulfilled or acknowledged by the interviewed experts. Thematic discussions on actions or policies related to the development of guidelines and indicators appeared with regard to infectious disease surveillance and management of rare diseases but were not a major theme across interviews. The EU-level task to promote best practice exchange among member states was regarded as influential, which is represented by a top position in one of the rankings presented in this paper. With regard to the task of establishing monitoring and evaluation structures, some respondents perceived the status of the EU health information system rather as a failure. This corresponds to observations in the literature indicating that although ground work such as the development of a common EU health indicator set is acknowledged \\[[@B50],[@B51]\\] further efforts are needed to implement and maintain health indicators \\[[@B51]\\] and to develop a permanent and sustainable EU public health monitoring and reporting infrastructure that supports decision making in public health on EU level \\[[@B50],[@B51]\\]. Respondents agreed that cooperation in the area of public health between member state representatives and experts as well as with other stakeholder groups has increased and has been facilitated by the EU through various projects, networks, forums, and platforms. This trend was mainly positively perceived since it supported EU-level public health policy by accumulating and exchanging knowledge, generating public support and a legitimacy to act on certain fields \\[[@B3]\\]. This finding is corroborated by the literature on the potential of new governance instruments for health- and social policy-making at EU-level \\[[@B11],[@B52]-[@B54]\\]. However, these new governance instruments can also be regarded as a rather strategic investment of the EC to keep topics on the agenda until a political window of opportunity opens but as an ineffective policy tool to enforce and implement action in due course \\[[@B6]\\]. The collaboration of a diverse set of stakeholders as it is the case for example in the EU Platform for Action on Diet, Physical Activity and Health can lead to actions that constitute rather a compromise of various interests. Consequently, the results might be disappointing from the viewpoint of public health experts \\[[@B6],[@B55]\\]. A final judgment on the impact of facilitating collaboration is to be awaited and may only be made in the long term future. It will require different ways of measuring 'impact' compared to the analysis of domestic adaptations when implementing EU hard law \\[[@B56]\\].\n\nCharacteristics of EU health policy that influenced experts' perception\n-----------------------------------------------------------------------\n\nThe assessments of EC tasks for public health policy making have been influenced by characteristics like the subsidiarity principle throughout several interviews. On the one hand some participants were in favor of more EU influence on health policies and their implementation. In their view integration and harmonization of health policy did not reach far enough and hence their perception of actions was dominated by the category 'missed opportunity'. On the other hand some experts were in favor of keeping certain health issues like health care as national responsibility which led to a perception of too much EU involvement and a negative perception of the evolvement of the health mandate.\n\nPublic health has a cross-cutting nature and cooperation across DG's often poses difficulties. Therefore, convincing evidence is required to demonstrate the health impact of policies outside the health domain and strong partnerships are needed to counter strong industrial lobbying groups \\[[@B6],[@B57],[@B58]\\]. The ease of cooperation and the potential to achieve policy coherence between DG SANCO and DGs with stronger regulatory competences like the internal market (e.g. regarding tobacco, pharmaceuticals) or agriculture policy (regarding food safety, subsidies of unhealthy versus healthy food products) represented another characteristic that influenced the individual perception of EU public health policies. Experts who assessed the value of EU health policy actions under the reality of a rather weak health mandate were more likely to perceive EU actions as achievements. This was in contrast to others who strove for more appreciation of social and health matters in EU policies and who perceived a lot of missed opportunities or failures in this regard as the power of the EU was too weak to realize change and to fulfill the objective of the health mandate to ensure human health protection for citizens in the EU.\n\nIn summary, underlying themes such as cooperation among European public health professionals, increasing institutionalization, and characteristics such as the issue of subsidiarity or the possibilities to cooperate across EU policy domains influenced experts' perceptions throughout the topics presented in this paper. These conditions and characteristics are part of what Lamping called the \"chaordic dynamics\" of European integration in the field of health policies \\[[@B3]\\]. As our study demonstrated EU health policy does not demonstrate a clear-cut success since the logic of action in the field can involve diverging interests. Nevertheless, the EU public health has quite systematically developed in terms of scope and impact beyond the original mandate.\n\nLimitations\n-----------\n\nThe ranking of influential policy outputs provided indications on important developments in EU public health policy. However, even though we categorized the outputs, they sometimes differed in character and power which might have led to imbalanced judgments.\n\nAdditionally, we received different reasons for labeling EU-level actions or policies as achievements, missed opportunities or failure for public health. Some were identified because they increased the strength or value of EU-level public health policy, whereas, others were identified because they impacted the health of the European population.\n\nThe findings of the study may not be empirically generalizable since they were closely linked to qualitative individual perceptions and the settings that participants belonged to. However, we are confident that the broad range of profiles of the experts has ensured the diversity of perceptions on the topics varying from achievement to missed opportunity and failure. Moreover, given that participants were generally active in health policy at EU-level and mainly positive about the EU, this could also have influenced the obtained results to some extent.\n\nConclusion\n==========\n\nEU public health policy is subject to divergent perceptions of how successful or unsuccessful specific topics have been tackled and how far European integration in public health policy should go. From the findings, it is unequivocal that the EU has strengthened its role over the past twenty years in supporting, coordinating, and supplementing member states' actions on public health issues as laid down in Article 168 (2), TFEU. The EU is now a recognized player in public health in Europe. However, when it comes *\"to the promotion of a high level of \\[...\\] protection of human health \\[...\\]in defining and implementing its policies and activities\"* (Article 9, TFEU), further work is needed to achieve the full potential of the EU health mandate.\n\nEndnote\n=======\n\n^a^Also several EU member states disconnected on national level the Ministry of Health from Social Affairs. At the time of writing only seven out of 28 EU member states organized health and social affairs within one ministry (Spain, France, Sweden, Finland, Estonia, Greece, the Netherlands).\n\nAbbreviations\n=============\n\nASPHER: Association of Schools for Public Health in the European Region; BSE: Boviene spongiforme encefalopathie; DG: Directorate General; DG Connect: Directorate General for Communications Networks, Content and Technology; DG SANCO: Directorate General for Health and Consumers; EC: European Commission; ECDC: European Centre for Disease Prevention and Control; EFSA: European Food Safety Authority; EMA: European Medicines Agency; EMCDDA: European Monitoring Centre for Drugs and Drug Addiction; EU: European Union; GATS: General Agreement on Trade in Services; HEIDI: Health in Europe: Information and Data Interface; HIA: Health Impact Assessment; HiAP: Health in All Policies; HTA: Health Technology Assessment; OECD: Organization for Economic Cooperation and Development; SARS: Severe acute respiratory syndrome; SDoH: Social Determinants of Health; TFEU: Treaty on the Functioning of the European Union; WHO-EUR: World Health Organization- Regional Office for Europe.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nAll authors were involved in setting up the study. NR and TC coordinated the study. NR, TC and KS carried out the interviews, performed the analysis, and interpreted the results. NR drafted the manuscript. All authors revised the manuscript and approved the final version.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2458/13/1074/prepub>\n\nAcknowledgements\n================\n\nThe paper was partly presented at a symposium held on 5th June 2012 in Brussels, Belgium, the European Public Health Conference on 10th November 2012 in Malta, and a conference on the 20th anniversary of the EU health mandate on 22nd May 2013 in Maastricht, the Netherlands. We would like to express our sincere thanks to the experts who invested their time and participated in the interviews. Their views and perceptions on European public health policy were highly valued. We would like to thank Wilco Tilburgs and Hassan el Fartakh for their support in transcribing the interviews and Ann Borg for her support and helpful recommendations during the final editing process. We also appreciate the support of our colleagues at the Department of International Health at Maastricht University; in particular Kasia Czabanowska, Matt Commers, Kai Michelsen, Christoph Aluttis and Beatrice Scholtes gave advice in setting up and designing the study, questioning the results, or reviewing the manuscript.\n"}
+{"text": "![](glasgowmedj75438-0057){#sp1 .439}\n\n![](glasgowmedj75438-0058){#sp2 .440}\n\n![](glasgowmedj75438-0059){#sp3 .441}\n"}
+{"text": "1. Introduction {#sec1-biosensors-06-00028}\n===============\n\nEllington and Szostak \\[[@B1-biosensors-06-00028]\\] developed for the first time an in vitro selection technique that allowed the discovery of oligonucleotides or peptides that bind to specific targets with high specificity and affinity. Since then, the interest in aptamers has increased significantly \\[[@B2-biosensors-06-00028],[@B3-biosensors-06-00028],[@B4-biosensors-06-00028]\\] because of their advantages compared to other biological receptors. Among these properties, we can enumerate the size, the synthesis and the stability. Additionally, the aptamer-analyte bound might be disrupted by very small changes.\n\nWhispering gallery mode based resonators (WGMR) have attracted much attention in the biomedical sensing field \\[[@B5-biosensors-06-00028],[@B6-biosensors-06-00028]\\], including aptamer based sensors for thrombin detection \\[[@B7-biosensors-06-00028],[@B8-biosensors-06-00028]\\], vascular endothelial growth factor \\[[@B8-biosensors-06-00028]\\] and toxins such as aflatoxin M1 \\[[@B9-biosensors-06-00028]\\]. WGMR have different morphologies with particular spectral features such as including narrow linewidth, high stability, and tunability. Regarding sensing applications, the most important properties are the ultrahigh quality factor *Q* and long lifetime of WGMR, where the change in *Q* or the shift resonant wavelength is used for measuring the change of parameters in the surrounding environment or binding event on the WGMR surface. WGMR are a valid alternative to other evanescent wave (EW) sensors, such as surface plasmon resonators (SPR) \\[[@B10-biosensors-06-00028]\\]. The feasibility of WGMR for detecting of single virions has been demonstrated recently \\[[@B11-biosensors-06-00028]\\]. Silica WGMR have two important advantages when compared to other EW sensors: a significantly longer penetration depth and a well established surface chemistry for silica. The combination of WGMR and covalent chemistry will be able to provide very accurate biochemical sensors.\n\nThrombin is a powerful vasoconstrictor that is involved in many diseases, like atherosclerosis and thrombosis. Vascular endothelial growth factor (VEGF) is a signaling protein that regulates angiogenesis and its normal function is the formation of new blood vessels. VEGF, when overexpressed, can also cause vascular diseases and enhance metastasis if the tumor cells can express it. Detecting these blood proteins in laboratory and clinical measurements is time consuming and costly, because there are not always available antibodies. Regarding toxins, aflatoxin M1 (AFM1) is a milk contaminant and potent carcinogen (European Commission regulation, EC No. 1881/2006) limiting the maximum allowable concentration of Aflatoxin in milk products to 50 ng/kg. Thus, aflatoxin contamination represents a serious threat to human health and, in economical terms, a loss for the dairy industry. Presently, the screening procedure involves Enzyme-LinkedImmunoSorbent Assay (ELISA) tests \\[[@B12-biosensors-06-00028]\\], and the suspicious samples need further investigations with High-Performance Liquid Chromatography (HPLC) tests \\[[@B13-biosensors-06-00028]\\], which are costly and time-consuming processes.\n\nThe focus of this review is on biochemical sensing using WGMR as transducers and aptamers as biochemical receptors. The aim of our review is also to show the feasibility of WGMR aptasensors in medical diagnosis and food safety controls. However, direct bacteria detection is quite challenging when using EW sensors due to their large dimensions \\[[@B14-biosensors-06-00028]\\].\n\nUsually, the biological recognition elements (BRE) are antibodies, streptavidin, enzymes and aptamers; these BRE bind specifically to their corresponding analytes: antigens, biotin/biotinylated proteins, amino acids, and proteins. Antibodies and antigens are highly complementary, and, as a consequence, highly specific. Streptavidin shows a high specificity only to biotin, but for sensing other analytes there is the need to biotinylate the analytes first, adding an extra chemical step. Enzymes are able to catalyze a large number of reactions, they are also able to detect substrates, inhibitors and products of the catalysis. The main advantages are specificity and catalytic activity. However, enzyme activity largely depends on the environment. Aptamers are engineered molecules, with high specificity and affinity towards the target analyte (proteins, cells, small molecules, etc.). The binding may be disrupted by very small changes. However, despite the superior performances of aptamers compared to antibodies and WGMR to other EW sensors, there is not much literature combining both types of receptors and transducers. Up to now, the WGMR that have been used are silica microspheres and silicon oxynitride (SiON) ring resonators.\n\n2. Materials and Methods {#sec2-biosensors-06-00028}\n========================\n\n2.1. Surface Functionalisation {#sec2dot1-biosensors-06-00028}\n------------------------------\n\nThe surface functionalization is the most important step in the production of accurate biosensors. The BRE layer has to be thinner than the evanescent tail and smooth for preserving the high *Q* of the WGMR . As mentioned in the introduction, there is a wide choice of BREs. Among the various techniques, covalent silane chemistry is the one used in WGMR aptasensors.\n\nThe first silica microspherical WGMR aptasensor paper was published by Fan et al. \\[[@B7-biosensors-06-00028]\\]. Silica microspheres were functionalized in dry conditions using 3-Aminopropyltrimethoxysilane (3-APS, 97%, Sigma Aldrich, St. Louis, MO, USA) whereas in the second one, published by Pasquardini et al. \\[[@B8-biosensors-06-00028]\\], the authors used 3-mercaptopropyltrimethoxysilane (MPTMS, 99%, Gelest) in wet conditions. SiON ring resonators were also functionalized in wet conditions using a 3-glycidoxypropyl methyldiethoxy silane (GPTMS) solution \\[[@B9-biosensors-06-00028]\\].\n\n### 2.1.1. Materials and Silica Microspherical WGMR Functionalisation {#sec2dot1dot1-biosensors-06-00028}\n\n#### Dry Protocol\n\n3-Aminopropyltrimethoxysilane (3-APS, 97%), 1--4 phenylene diisothiocyanate (PDC) Pyridine, *N,N*-dimethyl formamide (DMF), anhydrous ethanol, acetone and thrombin from human plasma (38 kD MW) from Sigma Aldrich are used without any further purification. The assay of all chemicals is 98% except pyridine, which is 99%.\n\nThe thrombin aptamer (5\u2032-CCAACCCAACGGTTGGTGTGGTTGG-3\u2032) \\[[@B7-biosensors-06-00028]\\] and the non-sense apatamer sequence (5\u2032-TATGAATTCAATCCGTCGAGCAGAGTT-3\u2032) were also used without further purification. Both aptamer sequences were modified at the 5\u2032 end with amino groups and were purchased from Invitrogen.\n\nAfter oxygen plasma cleaning (plasma cleaner PDC-32G, Harrick, Ossining, NY, USA) for 10 min, the WGMR were exposed to vapor silanization for 1 h inside a desiccator. 3-APS was used pure. Silanized WGMR were then annealed at 160 \u00b0C for 20 min. Afterwards, WGMR followed a two hours incubation in a 1% (*w*/*w*) PDC solution in a 10/90 (*v*/*v*) pyridine DMF solution. This step was needed for cross-linking the surface amine groups to the protein amine groups \\[[@B15-biosensors-06-00028]\\]. The last two steps were rinsing five times with anhydrous ethanol and acetone, respectively, one hour incubation were in 2 mM anti-thrombin aptamer in carbonate buffer (pH = 8.3). The storage procedure was done at room temperature.\n\n#### Wet Protocol\n\nThe following chemicals were used in the wet protocol: 3-mercaptopropyltrimethoxysilane (MPTMS, 99%, Gelest), toluene anhydrous (99.8%, Sigma Aldrich) and toluene (Sigma-Aldrich); purified Human Thrombin (0.3 mg/mL, Bioultra, THR, Sigma Aldrich), purified VEGF165 (0.3 mg/mL, Sigma Aldrich); Bovine Serum Albumin (BSA, Thermo Scientific) and Prionex^\u00ae^ solution (Branch Pentapharm, Switzerland). Two different buffer solutions were used: the Coupling Buffer (Na~2~CO~3~/NaHCO~3~ 0.5 M, pH = 9, CB) for aptamer incubation and the BioRecognition Buffer (Tris 50 mM, EDTA 1 mM, MgCl~2~ 1 mM, KCl 150 mM pH = 7.4, BRB) was used in the thrombin incubation whereas for VEGF detection the biorecogniton buffer was a Tris-HCl 20 mM buffer solution with a pH = 7.4. \u03b2-mercapto-ethanol (MP-ET, Sigma-Aldrich) was used for passivating the remaining active sites in the modified WGMR surface. For activated WGMR regeneration, a 50 mM NaOH solution was used. The aptamer sequences are HPLC purified (IDT Integrated DNA Technologies, Leuven, Belgium).\n\nThe sequences of the Thrombin Binding Aptamer (TBA) are: 5\u2032-HO-(CH~2~)~3~-S-S-(CH~2~)~3~- GGT TGG TGT GGT TGG-3\u2032 (TBA-15)5\u2032-HO-(CH~2~)~3~-S-S-(CH~2~)~3~-AG TCC GTG GTA GGG CAG GTT GGG GTG ACT-3\u2032 (TBA-29)\n\nThe sequence of the non specific aptamer is: 5\u2032-HO-(CH~2~)~3~-S-S-(CH~2~)~3~-CCG TCG AGC AGA GTT-3\u2032 (NS, Non Sense) The sequence specific for VEGF165 detection is: 5\u2032-HO-(CH~2~)~3~-S-S-(CH~2~)~3~-CC GTC TTC CAG ACA AGA GTG CAG GG-3\u2032 (hereafter called VEGF-25)\n\nMPTMS has thiol groups instead of amine groups, and it was chosen for binding the dithiol groups at 5\u2032 end of the aptamer sequences. The first step is the cleaning and oxidizing one, using piranha solution (H~2~SO~4~:H~2~O~2~ at 4:1 *v*/*v* concentration) for 3 min, followed by a washing one using MilliQ water plus a drying step in an oven at 110 \u00b0C for two hours or more. The activated WGMR were then silanized at 60 \u00b0C using a MPTMS toluene anhydrous solution (0.0025% *v*/*v*) for a total time of about 10 min. Silanized WGMR were placed in a chemical hood for toluene rinsing and then drying.\n\nThe aptamer sequences were immobilized on silanized WGMR adapting Hilliard's procedure described in \\[[@B16-biosensors-06-00028]\\]. In summary, this procedure consists in a short thermal treatment at 95 \u00b0C followed by cooling step in ice. The silanized WGMR were then incubated in an orbital shaker for 2 h in a 10 \u03bcM aptamer solution in CB buffer, followed by a washing step in CB buffer. Finally, the passivation step at room temperature was performed in 1 mM MP-ET in CB buffer for 2 h, this step reduces the aptamer density by a factor of 22. This last step enhances the target recognition, since it allows the aptamer structuration and it also blocks the residual free binding sites.\n\n### 2.1.2. Materials and SiON WGMR Functionalization {#sec2dot1dot2-biosensors-06-00028}\n\nA wet silanization protocol was used for functionalizing these sensors. First the transducer were cleaned with a Piranha solution, afterwards the samples were silanized by immersion in 0.01% *v*/*v* of GPTMS (3-glycidoxypropyl methyldiethoxy silane) solution in anhydrous toluene at 60 \u00b0C for 10 min. Then an amino-terminated DNA-aptamer (5\u2032-NH~2~-(CH~2~)~6~-GT TGG GCA CGT GTT GTC TCT CTG TGT CTC GTG CCC TTC GCT AGG CCC ACA-3\u2032) at 100 \u03bcM in phosphate buffer (50 mM, ionic strength 300 mM, pH 8) was incubated on silanized surfaces for 2 h. The aptameric anti-aflatoxin DNA sequence with a kD of 10 nM was identified by NeoVentures Biotechnology Inc. \\[[@B17-biosensors-06-00028]\\]. The aminomodified sequence is HPLC purified and was purchased from IDT Integrated DNA Technologies (Leuven, Belgium). Finally, an ethanolamine passivation at 1 mM for 30 min was applied.\n\n2.2. Experimental Set-up {#sec2dot2-biosensors-06-00028}\n------------------------\n\n### 2.2.1. Experimental Set-ups for Microspheres Based Aptasensors {#sec2dot2dot1-biosensors-06-00028}\n\nThe easiest microspherical WGMR fabrication technique is based on a fusion method. There are two techniques widely use: CO~2~ laser irradiation \\[[@B18-biosensors-06-00028],[@B19-biosensors-06-00028]\\] or melting by electrical arcs discharged by a fiber fusion splicer \\[[@B20-biosensors-06-00028],[@B21-biosensors-06-00028]\\] of a silica fiber tip. The partial melting of the silica and surface tension produce a spherical shape. The microspheres used in sensing have a diameter that can oscillate from 50 of 260 \u03bcm \\[[@B7-biosensors-06-00028],[@B22-biosensors-06-00028],[@B23-biosensors-06-00028]\\] and are stored under vacuum at room temperature, in order to avoid contamination.\n\nSample delivery is done using fluidic cells. A small fluidic cells were fabricated by using an angle-polished fiber prism \\[[@B7-biosensors-06-00028]\\] as both the cell bottom enclosure and coupling system. [Figure 1](#biosensors-06-00028-f001){ref-type=\"fig\"} shows the angle-polished fiber prism approach.\n\nThe laser light can also be coupled to the WGM resonator using a tapered fiber. [Figure 2](#biosensors-06-00028-f002){ref-type=\"fig\"} shows another microfluidic flow system \\[[@B24-biosensors-06-00028]\\]. The fluidic cell is open and it has a volume of about of 350 \u03bcL with one inlet connected to a peristaltic pump and an outlet connected to a reservoir. The fluidic loop is open for constant washing of the protein excess. The tapered fiber was placed at the bottom of the cell in parallel to the flow. The experimental flow rate was kept at 167 \u03bcL/min \\[[@B8-biosensors-06-00028]\\]. In order to avoid the non selective binding of proteins either in the cell walls or in the taper, the microfluidic flow system is passivated.\n\nA tunable narrow linewidth laser is used to excite the resonant wavelengths. Such a laser can be finely swept slowly with high accuracy. The light transmitted or reflected through the coupler-WGMR system is monitored into an oscilloscope. The resonances are spectral dips that can shift when the size and/or the refractive index of the resonators change.\n\n### 2.2.2. Experimental Set-ups for Racetrack Microring Based Aptasensors {#sec2dot2dot2-biosensors-06-00028}\n\nFor the microring fabrication, SiON films were deposited by plasma enhanced chemical vapor deposition (PECVD) on 6-in, 625 \u03bcm thick crystalline Si wafers with a 4 \u03bcm thick buffer oxide layer. The waveguides, ring resonators and directional couplers structures were fabricated using standard UV-lithography and reactive-ion etching techniques. An annealing step of 1.5 h at 1050 \u00b0C was also performed in order to remove hydrogen bonds from the material and decrease the losses of the structures. Finally, the processed wafers were covered with a 1 \u03bcm TEOS film, used as a cladding layer.\n\nIn order to create the sensors, the cladding layer was opened through a BHF wet etching. This allows the functionalization of the rings as well as the sensing measurements through their exposure to the ambient and to the molecules to be analyzed \\[[@B25-biosensors-06-00028]\\]. [Figure 3](#biosensors-06-00028-f003){ref-type=\"fig\"} shows a sketch of the sensor configuration that was coupled to a PDMS microfluidic flow cell, with a chamber of about 0.5 \u03bcL of volume, connected to a VICI M6 liquid handling pump. A detailed characterization of the racetrack microring resonators can be found at \\[[@B9-biosensors-06-00028]\\].\n\n3. Results {#sec3-biosensors-06-00028}\n==========\n\n3.1. Microspherical Aptasensors for Thrombin and VEGF {#sec3dot1-biosensors-06-00028}\n-----------------------------------------------------\n\nAfter reaching thermal equilibrium, the analyte is injected into the BRB. The wavelength first decreases (blue shift) due to thermal contraction of the microsphere, afterwards resonant wavelength increases till it saturates with time.\n\nThe following control experiments were performed, one was performed injecting a BSA solution of similar concentration into the flow cell with the right aptamer bound on the WGMR surface; the second one was performed using a non-sense aptamer bound on the spheres surface but injecting thrombin into the flow cell \\[[@B7-biosensors-06-00028],[@B8-biosensors-06-00028]\\]. The BSA spikes had no effect on the WGMR aptasensor response ([Figure 4](#biosensors-06-00028-f004){ref-type=\"fig\"}a). [Figure 4](#biosensors-06-00028-f004){ref-type=\"fig\"}b shows that thrombin spikes had no effect on the non-sense aptamer bound to the WGMR surface.\n\nZhu et al. performed one single specific bioassay for TBA-15 in buffer. Pasquardini et al. performed two specific binding assays for TBA-15 and TBA-29, both in buffer and in a complex matrix (1:10 human serum solution in buffer (*v*/*v*) without any filtering). [Figure 5](#biosensors-06-00028-f005){ref-type=\"fig\"}a shows the results for TBA-15 in buffer and in serum. The same authors repeated the same procedure for the VEGF-25 DNA-aptamer. Both authors first performed one control analysis consisting of injecting BSA of similar concentration into the flow cell with the aptamer immobilized on the WGMR surface. Similar results were obtained after spiking 5 \u03bcL of 0.3 mg/mL of VEGF165 into the flow cell, reaching equilibrium in 15 min.\n\nThe same group also tested the capability of reusing the microspheres after treatment with NaOH. The sensor regeneration was done by 50 mM NaOH solution and the WGMR aptasensors was then used for another detection cycle. The binding capacity was maintained as shown in figure 6. [Figure 6](#biosensors-06-00028-f006){ref-type=\"fig\"} shows a sensorgram of a microsphere regenerated twice with two injections of 5 \u03bcL of 0.1 mg/mL of thrombin. As expected, due to lower concentration of thrombin, equilibrium was not reached after the first injection.\n\n3.2. Racetrack Ring Aptasensors for Aflatoxin M1 {#sec3dot2-biosensors-06-00028}\n------------------------------------------------\n\nThe Aflatoxin-sensing measurements were performed in a microfluidic chamber filled with a buffer solution mainly based on 2-(Nmorpholino) ethanesulfonic acid (MES) and dimethyl sulfoxide (DSMO). The DSMO is needed to dissolve the aflatoxin. Then aflatoxin AFM1 solutions were injected into the chamber, and the resonance shift was monitored in real time. Finally, we injected again the buffer solution. All measurements were done with a flow of 3 \u03bcL/min.\n\nThree solutions with different concentrations were flown into the chamber, one at 12.5 nM, one at 25 nM and the last at 50 nM. After each Aflatoxin injection, two or three Glycine (100 mM glycine-HCl, pH 2) injections were effectuated, in order to detach the toxin from the sensor and regenerate the aptamers. The regeneration performances of Glycine concerning this aptamer were already demonstrated in \\[[@B26-biosensors-06-00028],[@B27-biosensors-06-00028]\\].\n\n[Figure 7](#biosensors-06-00028-f007){ref-type=\"fig\"} shows the sensorgrams for the three concentrations of AFM1, demonstrating the reproducibility and reusability of the sensor for AFM1. The binding effect was observed down to a minimum concentration of 1.58 nM \\[[@B25-biosensors-06-00028]\\].\n\n4. Discussion {#sec4-biosensors-06-00028}\n=============\n\nThe resonant wavelength shifts are calculated as the variation of the radius of the microsphere R and its index of refraction n~s~, using the following expression: *\u0394\u03bb/\u03bb = \u0394R/R* + *\u0394n~s~/n~s~*. However, the expression above is not valid for complex molecules binding to a surface \\[[@B18-biosensors-06-00028]\\]. In this case, the resonance shift is due to their excess of polarizability, \u03b1***~ex~***, and it can be described using the following equation: where *\u03c3* is the surface density of molecules forming a layer, and *n~s~* and *n~m~* are the refractive index of the sphere and the medium, respectively; and \u03b5~0~ is the vacuum permittivity.\n\nThrombin excess of polarizability is 4\u03c0\u03b5~0~ \u00d7 2.1 \u00d7 10^\u221221^ cm^3^, R = 130 \u03bcm (average microsphere radius of Pasquardini's experiments), n~s~ = 1.49 (silica index of refraction), n~m~ = 1.33 (buffer index of refraction) and \u03b4\u03bb = 14.6 pm \\[[@B8-biosensors-06-00028]\\], the thrombin surface density is calculated to be 3.14 \u00d7 10^12^ cm^\u22122^, which is close to a compact layer. Thrombin is a roughly spherical molecule of about 4.6 nm of diameter \\[[@B28-biosensors-06-00028]\\], so its projected area is about 17 nm^2^ \\[[@B7-biosensors-06-00028]\\] and its binding kinetics can be successfully modeled by treating thrombin as a sphere. The maximum fractional area coverage is estimated to be 0.55 for random packing of spheres \\[[@B29-biosensors-06-00028]\\] which means that \u03c3~max~ = 3.24 \u00d7 10^12^ cm^\u22122^ for thrombin. Thus, the thrombin layer is about the theoretical level. The VEGF165 calculated surface density is about 1.98 \u00d7 10^12^ cm^\u22122^. In Zhu et al., the thrombin surface density was about 5.4 \u00d7 10^12^ cm^\u22122^ for a maximum \u03b4\u03bb of about 34 pm.\n\nSince desorption was not observed in our experiments, binding of kinetics could be described using a simple differential equation that neglects the dissociation rate constant. The response of the the WGMR aptasensor is then a simple exponential \\[[@B30-biosensors-06-00028]\\]: where \u03bd~opt~ is the optical frequency, C is the concentration and *k~a~* is the association constant. It should be noted that transport phenomena and steric hindrance were also neglected \\[[@B31-biosensors-06-00028]\\]. From the experimental data, \u0394\u03bd~max~ is 7.44 \u00b1 0.22 GHz and assuming that binding is saturated with the used concentration, we obtained \u03b4R~max~ = 2.49 \u00b1 0.07 nm. This is consistent with monolayer coverage confirming the AFM measurements (not shown in here). The association constant *k~a~* is about 8.8 \u00b1 0.4 \u00d7 10^4^ M^\u22121^\u00b7s^\u22121^ for TBA-15 and about 2.2 \u00d7 10^4^ M^\u22121^\u00b7s^\u22121^ for TBA-29, which are in good accordance with published results for TBA-15 \\[[@B32-biosensors-06-00028]\\] and TBA-29 \\[[@B33-biosensors-06-00028]\\].\n\n5. Conclusions {#sec5-biosensors-06-00028}\n==============\n\nSeveral groups have demonstrated the potential of the WGMR based aptasensors in medicine and also in food safety. Microspherical WGMR have been used in the first demonstration of aptasensing. Microspheres have been used for detection of blood proteins such as thrombin (TBA-15 and TBA-29) and VEGF, whereas racetrack microrings have been used for detection of pathogens such as aflatoxin M1. WGMR aptasensors are real time sensors with high specificity and good reversibility that can be used for determining binding kinetics.\n\nL. Lunelli, L. Pasquardini and C. Pederzolli are gratefully acknowledged for help and discussions. This research study was partially supported by Italian MIUR-FIR program No. RBFR122KL1, and by Ente Cassa di Risparmio di Firenze project No. 2014.0770A2202.8861\n\nThe authors declare no conflict of interest.\n\nThe following abbreviations are used in this manuscript: WGMRWhispering gallery mode resonatorsSiONsilicon oxynitridSPRsurface plasmon resonatorsVEGFVascular endothelial growth factorAFM1aflatoxin M1ELISAEnzyme-LinkedImmunoSorbent AssayHPLCHigh-Performance Liquid ChromatographyAPSAminopropyltrimethoxysilaneMPTMSmercaptopropyltrimethoxysilaneGPTMSglycidoxypropyl methyldiethoxy silanePDCphenylene diisothiocyanateDMFdimethyl formamideCBCoupling BufferMBModification BufferBRBBioRecognition BufferTBAThrombin Binding AptamerPECVDplasma enhanced chemical vapor depositionMES2-(Nmorpholino) ethanesulfonic acidDSMOdimethyl sulfoxide\n\n![(**a**) Schematic of the microsphere and fluidic cell (side view); (**b**) top view of the fiber prism with a microsphere. Reprinted with permission from \\[[@B7-biosensors-06-00028]\\] \u00a9 2006 MDPI.](biosensors-06-00028-g001){#biosensors-06-00028-f001}\n\n![A schematic diagram of the experimental arrangement. Reprinted with permission from \\[[@B24-biosensors-06-00028]\\] \u00a9 2012 MDPI.](biosensors-06-00028-g002){#biosensors-06-00028-f002}\n\n![A schematic diagram of the microring based biosensor. **Left**: the input light is split into the four ring resonators using directional couplers with a gap of 600 nm. **Right**: zoomed in image of the ring resonator structure. The gap and coupling length between the bus waveguide and the resonator are 600 nm and 35 \u03bcm, respectively. The radius of the resonators is 100 \u03bcm. Reprinted with permission from \\[[@B25-biosensors-06-00028]\\] \u00a9 2015 Elsevier.](biosensors-06-00028-g003){#biosensors-06-00028-f003}\n\n![Control experiments: (**a**) whispering gallery mode resonator (WGMR) response of a Thrombin Binding Aptamer (TBA)-15 sensing surface for 0.3 mg/mL Bovine Serum Albumin (BSA) injected into the flow cell and (**b**) sensorgram using a non sense aptamer sensing surface for 0.3 mg/mL thrombin.](biosensors-06-00028-g004){#biosensors-06-00028-f004}\n\n![(**a**) WGMR response with TBA-15 modified resonator following the addition of thrombin in buffer (**b**) Sensorgram of thrombin binding to a WGMR immobilized with TBA-15 in 1:10 non-filtered human serum, following the addition of thrombin.](biosensors-06-00028-g005){#biosensors-06-00028-f005}\n\n![Sensorgram of thrombin binding to a regenerated WGMR immobilized with TBA-15, following the addition of thrombin (two identical additions of 5 \u03bcL of 0.1 mg/mL of thrombin).](biosensors-06-00028-g006){#biosensors-06-00028-f006}\n\n![(**left**) Sensorgram for three different concentrations of alflatoxin M1 (AFM1). The high step-like response is due to the refractive index mismatch produced by the small content of Dimethyl sulfoxide (DMSO) in the solution; (**right**) Specific binding sensorgrams obtained from the curves in (**left**) by subtracting the bulk shift induced by the DMSO content. The dashed curves are exponential fittings for the evaluation of the rate constants and of the initial slopes \\[[@B27-biosensors-06-00028]\\]. Reprinted with permission from \\[[@B9-biosensors-06-00028]\\] \u00a9 2015 MDPI.](biosensors-06-00028-g007){#biosensors-06-00028-f007}\n"}
+{"text": "INTRODUCTION {#s0}\n============\n\nIn protein synthesis, the elongation cycle comprises an elaborate sequence of steps ([@B1], [@B2]). After an aminoacyl-tRNA (aa-tRNA) binds to the ribosome, it is tested for a match between its anticodon and the current mRNA codon. When a cognate aa-tRNA is found, peptide bond formation occurs and the tRNAs and mRNA translocate through the ribosome, enabling the cycle to begin again. In bacteria, the codon recognition step is catalyzed by elongation factor Tu (EF-Tu), a GTPase. Its eukaryotic homologue is called eEF-1A ([@B3]). The translocation step is catalyzed by a second GTPase called elongation factor G (EF-G) ([@B2]).\n\nIn the standard mechanistic model of *Escherichia coli* translation ([@B1], [@B2]), aa-tRNA binds to the ribosome as a ternary complex: aa-tRNA--EF-Tu(GTP). The ternary complex is recruited to the ribosome by binding to one of four L7/L12 sites that protrude from the stalk of the ribosome, as shown schematically in [Fig.\u00a01A](#fig1){ref-type=\"fig\"} ([@B4]). L7 is identical to L12, except for an acylated N terminus. Biochemical evidence indicates that the binding interface juxtaposes the C-terminal domain of L7/L12 and domain 1 of EF-Tu ([@B5], [@B6]). The ribosomal stalk thus tethers aa-tRNA copies in close proximity to the ribosomal A site, where they can be tested for a codon match. Under good growth conditions, *E.\u00a0coli* can carry out elongation at a rate of \\~17 to 20\u00a0amino acids/s, implying that the mean time to carry out a complete elongation cycle can be as short as 50\u00a0ms ([@B7], [@B8]). Since the vast majority of aa-tRNA copies carry a noncognate or near-cognate anticodon that does not match the current mRNA codon ([@B9]), testing of individual aa-tRNAs for a codon match must be very rapid. A recent global theory of bacterial metabolism suggested that the diffusive search of EF-Tu for its ribosomal binding site is the step limiting the overall growth rate ([@B10]).\n\n![(A) Schematic diagram showing four ternary complexes bound to the four L7/L12 units on the stalk of a 70S ribosome. One of the ternary complexes is also bound to the A site for codon testing (Based on the model of reference [@B4]). We emphasize that while biochemical studies support binding of the ternary complex to L7/L12, the stalk is highly mobile in all structural studies to date ([@B2]). (B) Several single-molecule trajectories of EF-Tu--mEos2 plotted in different colors and superimposed on the phase-contrast image of the same cell. (C, top) Composite spatial distribution heat map of EF-Tu--mEos2 for 4,221 localizations from 201 *E.\u00a0coli* cells of length 4 to 5\u00a0\u00b5m. Pixels are \\~45 by 45 nm. The intensity scale shows relative counts per pixel. (Bottom) Composite spatial distribution heat map of ribosomes (30S--mEos2 labeling) for 1,967 localizations from 108 *E.\u00a0coli* cells of length 4 to 5\u00a0\u00b5m. (D) The projected axial distribution of EF-Tu--mEos2 and ribosomes (30S--mEos2) for the same sets of cells used in panels C and D. The distributions are normalized to the same area and plotted on a relative scale of \u22120.5 to +0.5 for the long axis.](mbo0011836660001){#fig1}\n\nThe sequence of events leading from the initial binding step to codon recognition and peptide bond formation has been dissected in remarkable detail by a groundbreaking series of rapid-mixing kinetics experiments carried out *in vitro* and summarized in references [@B1] and [@B9]. Single-molecule studies *in vitro* have helped to further refine the detailed sequence of mechanistic steps ([@B11], [@B12]). The inferred mechanism includes two consecutive stages of codon discrimination: initial selection and subsequent proofreading, with multiple intermediate states delineated for both stages ([@B1]). The overall mechanism enables cognate aa-tRNAs to proceed rapidly to accommodation in the A site, while rapidly rejecting noncognate and near-cognate aa-tRNAs. Most recently, a detailed set of *in vitro* transition rates has been optimally scaled to form a theoretical set of *in vivo*, codon-specific transition rates that yield the correct overall translation rate in exponentially growing *E.\u00a0coli* ([@B9]). These optimized *in vivo* transition rates were then used to predict codon-dependent translation speeds, codon-specific translation dynamics, and missense error frequencies. The good agreement of the model predictions with the experiments serves to validate the new method for transforming detailed *in vitro* rates into useful *in vivo* rates.\n\nThe *E.\u00a0coli* ribosomal stalk (schematic in [Fig.\u00a01A](#fig1){ref-type=\"fig\"}) comprises the L11 protein, which binds to rRNA and forms the base of the stalk, the protruding L10 protein, which binds to L11 via a flexible connection, and four L7/L12 copies, which bind to L10 as a pair of dimers ([@B4]). Each L7/L12 has three domains. The N terminus binds to L10, and a flexible hinge connects the N terminus to the C terminus. A compelling body of biochemical evidence detailed in reference [@B4] and summarized below indicates that the C-terminal domain of L7/L12 binds to helix D of EF-Tu within the ternary complex. The only structural evidence for L7/L12 binding to EF-Tu comes from a cryo-electron microscopy (cryo-EM) reconstruction at a 1.8-nm resolution ([@B13]). The structure suggests a bridge between domain 1 of EF-Tu (the G domain) and the L7/L12 stalk, in agreement with inferences from the biochemical data. A comprehensive model of ribosomal stalk structure and function suggested that the four highly mobile L7/L12 C-terminal domains serve to efficiently recruit ternary complexes to the ribosome and help stabilize the active GTPase conformation of EF-Tu ([@B4]). However, there is no crystal structure that reveals the molecular-level details of the initial binding step of the ternary complex to the ribosome. In all high-resolution structural studies to date, the L7/L12 stalk is highly mobile and does not yield discernible electron density ([@B2]).\n\nWe and others have used live-cell, single-molecule fluorescence methods to study the spatial distribution and diffusive properties of a variety of proteins in *E.\u00a0coli* ([@B14], [@B15]). In a typical experiment, the protein of interest is expressed from the chromosome as a fusion to a photoconvertible fluorescent protein. A weak laser at 405\u00a0nm switches the absorption and emission wavelengths of literally one or two protein copies per cell. A more powerful probe laser then enables selective excitation, localization, and tracking of the sparse photoswitched copies until they photobleach. For high-copy-number proteins, this enables the acquisition of thousands of single-molecule trajectories from each cell over tens of seconds. The spatial localization accuracy is typically a \u03c3 value of \\~40 to 80\u00a0nm, and the temporal resolution can be in the low-millisecond range ([@B16]). In favorable cases, the diffusive properties of a single copy can be related to its biochemical function at a given moment in time.\n\nHere we present a single-molecule localization and tracking study of EF-Tu in *E.\u00a0coli*. EF-Tu is labeled at the C terminus with the 26-kDa ([@B17]) photoconvertible fluorescent protein mEos2 ([@B18]). Measurement of a large number of short-lived diffusive tracks at 2\u00a0ms/frame enables an approximate decomposition of the EF-Tu population into two substates. We call these states \"slow\" (assigned as copies transiently bound to translating 70S ribosomes, including polysomes) and \"fast\" (copies not bound to ribosomes, presumably mostly EF-Tu within free ternary complexes). Accordingly, the slow copies (\\~60%) concentrate in the three ribosome-rich regions where most translation occurs, outside the nucleoids ([@B15], [@B19]).\n\nCombining the new diffusion data with copy number estimates for ribosomes and EF-Tu indicates that the four L7/L12 sites are essentially saturated with EF-Tu copies *in vivo*. This new result corroborates the earlier inference from *in vitro* kinetics measurements that all four *E.\u00a0coli* L7/L12 sites are actively engaged in recruiting ternary complexes to the ribosome ([@B4]). The time scale of binding events indicates that free ternary complexes find translating ribosomes extremely efficiently, in good quantitative agreement with the recent model of *in vivo* kinetics ([@B9]). Evidently aa-tRNA copies are tested for a match to the current codon on a time scale of 1 to 2\u00a0ms or less, in further agreement with the *in vivo* model. Simultaneous binding of four ternary complexes to each translating ribosome may greatly enhance the rate of testing ([@B4]). Finally, the results refute the main conclusion from a recent single-molecule tracking study of tRNA diffusion in *E.\u00a0coli* ([@B20]). That work inferred that most aa-tRNAs are monomeric and freely diffusing, arriving at the ribosomal A site as bare aa-tRNAs, not as ternary complexes.\n\nRESULTS {#s1}\n=======\n\nComparison of axial spatial distributions of EF-Tu and ribosomes. {#s1.1}\n-----------------------------------------------------------------\n\nEssentially identical copies of EF-Tu are expressed by two genes in *E.\u00a0coli*: *tufA* and *tufB* ([@B21]). We have fused the gene coding for the photoconvertible fluorescent protein mEos2 to the C terminus of both endogenous genes within the chromosome in the *E.\u00a0coli* strain NCM3722 and then moved the fusions to the VH1000 background strain for further study (see [Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"}\u00a0in the supplemental material).\n\n10.1128/mBio.02143-17.10\n\n(Part A) Bacterial strains used. The background strain is *E.\u00a0coli* VH1000. Also shown are the oligonucleotides used (part B) and the sources of copy number estimates relative to ribosomes (part C). Download TABLE\u00a0S1, PDF file, 0.2 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nLabeling of all copies of EF-Tu with mEos2 ensures that there is no competition with unlabeled copies. Domain 3 of EF-Tu binds to tRNA and includes the C terminus, but mEos2 is appended on the face opposite to the tRNA binding site. In \"EZ rich, defined medium\" (EZRDM), the doubling time at 30\u00b0C of the modified strain expressing EF-Tu--mEos2 from the chromosome is 60 \u00b1 3\u00a0min, compared with 45 \u00b1 2\u00a0min ([@B19]) for the unlabeled VH1000 background strain (see [Fig.\u00a0S8](#figS8){ref-type=\"supplementary-material\"} in the supplemental material). Evidently the labeling does not greatly affect the functionality of EF-Tu, an essential protein.\n\nOur goal is to use diffusive properties to distinguish ribosome-bound EF-Tu from EF-Tu not bound to ribosomes. The mass of bare EF-Tu--mEos2 is 69\u00a0kDa, 26\u00a0kDa of which is due to mEos2. The mass of a typical labeled ternary complex, including mEos2 \\[aa-tRNA--EF-Tu(GTP)-mEos2\\] is \\~95\u00a0kDa. We would expect the diffusion coefficients of free ternary complexes (not bound to ribosomes) and of free, bare EF-Tu in the cytoplasm to be similarly fast---perhaps 4 to 8\u00a0\u03bcm^2^/s ([@B22], [@B23]). Short diffusive trajectories with significant localization error will not be able to distinguish free ternary complexes from bare EF-Tu; we use \"fast EF-Tu\" to denote a composite of these two species. Below we will argue that a large majority of these fast EF-Tu copies are bound within ternary complexes. In contrast, the ribosome mass is \\~2.5\u00a0MDa ([@B24], [@B25]) and translating 70S ribosomes in exponentially growing *E.\u00a0coli* exist primarily as polysomes ([@B15], [@B19], [@B26]). The mean 70S ribosome diffusion coefficient under these fast imaging conditions is \\~0.1\u00a0\u03bcm^2^/s (supplemental material). EF-Tu copies that are bound to translating 70S ribosomes should diffuse similarly slowly.\n\nIt was previously shown that under our moderately fast exponential growth conditions, *E.\u00a0coli* exhibits strong segregation of the two major nucleoid lobes from the 70S ribosomes ([@B19]). The projected axial distribution of ribosomes within the cytoplasm typically has three peaks, with the two nucleoid lobes interleaving three \"ribosome-rich regions.\" In contrast, free 30S and 50S subunits readily penetrate the nucleoid regions ([@B15], [@B19], [@B27]). Segregation of 70S ribosomes from the chromosomal DNA may serve to enhance the efficiency of recycling of 30S and 50S subunits and also the efficiency of the search for transcription initiation sites by RNA polymerase. The slowly diffusing EF-Tu ternary complexes bound to 70S ribosomes should also exhibit a three-peaked axial distribution, while rapidly diffusing, free EF-Tu should be distributed more uniformly.\n\nWe imaged EF-Tu--mEos2 molecules in cells by photoactivating and locating fluorophores, connecting locations over multiple frames to form trajectories of individual molecules ([@B28]). Details are provided in Materials and Methods and the supplemental material. To enable efficient superresolution imaging of rapidly diffusing molecules, the exposure time was 2\u00a0ms/frame with continuous laser illumination. The number of switched-on copies per cell was limited to 0 to 2 molecules per frame to avoid spatial overlap of the single-molecule features.\n\n[Movie\u00a0S1](#movS1){ref-type=\"supplementary-material\"}\u00a0shows typical raw data, and several example trajectories from a single cell are shown in [Fig.\u00a01B](#fig1){ref-type=\"fig\"}. In constructing axial spatial distributions that combine data from many cells, we included only cells that were 4 to 5\u00a0\u00b5m in tip-to-tip length to minimize blurring of features. From 201 such cells, we obtained 4,221 EF-Tu--mEos2 trajectories that lasted at least 6 steps (7 camera frames, or a total duration of 12 ms). All localizations were included in the spatial distributions. The axial and radial cell dimensions were normalized, and the relative molecular positions were pixelated and plotted to obtain a two-dimensional heat map of the EF-Tu spatial distribution ([Fig.\u00a01C](#fig1){ref-type=\"fig\"}, top). The map shows that EF-Tu is distributed over the entire cytoplasm, but the distribution is not homogeneous. For comparison, in [Fig.\u00a01C](#fig1){ref-type=\"fig\"} (bottom), we show the heat map for ribosomes with the 30S subunit labeled by the endogenously expressed S2-mEos2 protein at the C terminus as before ([@B19]) and imaged under the same conditions used for EF-Tu. Again, trajectories of 6 steps or longer in cells 4 to 5\u00a0\u03bcm in length were included. As shown qualitatively by the heat maps of [Fig.\u00a01C](#fig1){ref-type=\"fig\"} and quantitatively in the projected axial distributions of [Fig.\u00a01D](#fig1){ref-type=\"fig\"}, ribosomes exhibit substantially greater segregation from the nucleoids than EF-Tu. The total EF-Tu distribution does exhibit three peaks, but they are less sharply defined. This indicates that at a given moment, only a fraction of EF-Tu--mEos2 copies are associated with 70S ribosomes.\n\n10.1128/mBio.02143-17.11\n\nTracking individual EF-Tu molecules tagged with photoconvertible fluorescent protein mEos2 in a single live *E.\u00a0coli* cell. The movie represents 2\u00a0ms/frame with continuous laser exposure; 300 frames of the entire movie are shown. The cell outline (cyan) was obtained from a phase-contrast image. The movie plays back at 3 frames/s. Download MOVIE\u00a0S1, AVI file, 1.5 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nDiffusion of EF-Tu. {#s1.2}\n-------------------\n\nFor the diffusion study, we used 1,912 trajectories of duration 6 steps or longer obtained from 118 different cells. Longer trajectories were truncated at 6 steps. The exposure time was 2\u00a0ms/frame. The mean diffusion coefficient, *D*~mean~, can be estimated from a plot of the two-dimensional mean-square displacement versus lag time, MSD(\u03c4), using the slope of the first two data points. This provides a population-weighted average of diffusion coefficients over the different states of the molecule. The MSD slope accounts for localization error, but does not account for confinement effects. In [Fig.\u00a02](#fig2){ref-type=\"fig\"}, we compare MSD plots for wild-type (WT) EF-Tu and ribosomes. The mean diffusion coefficients are 2.02 \u00b1 0.19\u00a0\u00b5m^2^/s for EF-Tu and 0.4 \u00b1 0.1\u00a0\u03bcm^2^/s for ribosomes. The mean value for EF-Tu is consistent with the existence of at least two diffusive states: a fast, rapidly diffusing EF-Tu state and a slow, ribosome-bound state. The intercept of the MSD plot provides an estimate of the mean localization accuracy \u03c3 value of \\~60\u00a0nm ([@B29]).\n\n![Mean square displacement (MSD) plots for WT EF-Tu, the mutant form EF-Tu^L148A^, and ribosomes under normal growth conditions and with Rif treatment as indicated. Slopes from the first two points yield population-averaged diffusion coefficient estimates as follows: WT EF-Tu, 2.02 \u00b1 0.19\u00a0\u00b5m^2^/s; WT EF-Tu after Rif treatment, 3.5 \u00b1 0.4\u00a0\u00b5m^2^/s; mutant EF-Tu^L148A^, 3.1 \u00b1 0.3\u00a0\u00b5m^2^/s; mutant EF-Tu^L148A^ after Rif treatment, 5.2 \u00b1 0.4\u00a0\u00b5m^2^/s; and ribosomes (30S--mEos2 labeling), 0.4 \u00b1 0.1\u00a0\u00b5m^2^/s.](mbo0011836660002){#fig2}\n\nIn order to quantify the fraction of ribosome-bound EF-Tu copies, the same truncated trajectories were divided into individual steps with \u0394*t* = 2\u00a0ms between camera frames. This attempts to isolate short time intervals during which EF-Tu remains in one particular diffusive state ([@B16]). The resulting distribution of experimental single-step displacements, *P*~EF-Tu~(*r*), is shown for 11,472 individual steps in [Fig.\u00a03A](#fig3){ref-type=\"fig\"}. We analyze such *P*(*r*) distributions by comparison with a large number of simulated random walk trajectories that incorporate dynamic localization error \u03c3 and confinement within a spherocylinder that mimics the dimensions of an *E.\u00a0coli* cell. Details are provided in the supplemental material. For each chosen model diffusion coefficient, *D*, and measurement error, \u03c3, the simulations provide a numerical function we call *P*~model~(*r*; *D*). We attempt to fit the experimental distribution *P*(*r*) using least squares to a single population or to a weighted average of two static populations. The goodness of each fit was judged by the reduced chi-square statistic, \u03c7~\u03bd~^2^, which should be approximately 1 for an appropriate model function ([@B30]). For a one-state model, the only fitting parameter is *D*. For unconstrained models, including two static (nonexchanging) states, the fitting function is the linear combination *P*~model~(*r*) = *f*~slow~*P*(*r*; *D*~slow~) + (1 \u2212 *f*~slow~)*P*(*r*; *D*~fast~). Here the three fitting parameters are *D*~fast~, *D*~slow~, and the fractional population *f*~slow~, which in turn fixes *f*~fast~ = (1 \u2212 *f*~slow~).\n\n![(A) The experimental distribution of single-step displacements *P*~EF-Tu~(*r*) (gray histogram) for 11,472 2-ms steps for WT EF-Tu. The solid black line shows the best-fit model using two static states: \"slow\" (blue) and \"fast\" (red). Model parameters: *f*~slow~ = 0.6, *D*~slow~ = 1.0 \u00b5m^2^/s, *f*~fast~ = 0.4, and *D*~fast~ = 4.9 \u00b5m^2^/s. (B) Axial distributions of predominantly slow (blue) and fast (red) single-step displacements of WT EF-Tu in comparison with ribosome axial distribution (30S-mEos2 labeling \\[black\\]). The cutoffs chosen to separate slow (\\<0.1-\u00b5m) and fast (\\>0.2-\u00b5m) single-step displacements are indicated by the arrows in panel A. The distributions are normalized to the same area and plotted on a relative scale of \u22120.5 to +0.5 for the long axis.](mbo0011836660003){#fig3}\n\nOne-component fits to the *P*~EF-Tu~(*r*) were poor, with minimum \u03c7~\u03bd~^2^ = 9.7 (see [Fig.\u00a0S1B](#figS1){ref-type=\"supplementary-material\"} in the supplemental material). Fits to two nonexchanging diffusive states were substantially better. The best value of \u03c7~\u03bd~^2^ was 1.24, obtained using model parameters *f*~slow~ = 0.60 \u00b1 0.05, *D*~slow~ = 1.0 \u00b1 0.2 \u00b5m^2^/s, *f*~fast~ = 0.40 \u00b1 0.05, and *D*~fast~ = 4.9 \u00b1 1.2 \u00b5m^2^/s ([Table\u00a01](#tab1){ref-type=\"table\"}). The best-fit two-state model result is plotted in [Fig.\u00a03A](#fig3){ref-type=\"fig\"} and resolved into the two separate contributions. The parameter uncertainties are based on the range of parameters that return reduced chi-square values within 0.5 units of the best value, as detailed in the supplemental material. Parameter sets with \u03c7~\u03bd~^2^ values still larger were judged by eye to be qualitatively poor. The best two-component constrained fit to *P*~EF-Tu~(*r*) with *D*~slow~ fixed at 0.1\u00a0\u03bcm^2^/s (to match the slow, 70S component of the ribosome diffusion data) has \u03c7~\u03bd~^2^ = 2.5 ([Fig.\u00a0S1A](#figS1){ref-type=\"supplementary-material\"}), which is much worse than the global best-fit value of 1.24. Our constrained search for three-component fits did not reduce \u03c7~\u03bd~^2^ significantly (supplemental material, [Fig.\u00a0S1C](#figS1){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.02143-17.2\n\nDifferent fitting models for WT EF-Tu single-step displacement probability distribution *P*~EF-Tu~(*r*). (A) Two static states with *D*~slow~ fixed at 0.1\u00a0\u00b5m^2^/s (corresponding to the estimated 70S ribosome diffusion coefficient). Best-fit *D*~fast~ and fractions as shown yield \u03c7~\u03bd~^2^ = 2.5, indicating the model is inadequate. (B) Single static state. The minimum \u03c7~\u03bd~^2^ is 9.7, corresponding to *D*~fit~ = 1.0 \u00b5m^2^/s. (C) three static states with *D*~slow~ and *D*~fast~ fixed at 0.1 and 4.9\u00a0\u00b5m^2^/s, respectively. The best-fit *D*~medium~ and fractions are shown. The minimum \u03c7~\u03bd~^2^ is 1.25, which is not an improvement over the best two-state model of [Fig.\u00a03](#fig3){ref-type=\"fig\"}. Download FIG\u00a0S1, TIF file, 18 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n###### \n\nSummary of best-fit diffusion coefficients and fractional populations\n\n  EF-Tu or ribosome type[^a^](#ngtab1.1){ref-type=\"table-fn\"}    *D*~mean~ (\u00b5m^2^/s)[^b^](#ngtab1.2){ref-type=\"table-fn\"}   *f*~slow~[^c^](#ngtab1.3){ref-type=\"table-fn\"}   *D*~slow~ (\u00b5m^2^/s)   *D*~fast~ (\u00b5m^2^/s)\n  -------------------------------------------------------------- ---------------------------------------------------------- ------------------------------------------------ --------------------- ---------------------\n  Normal growth conditions                                                                                                                                                                         \n  \u2003\u2003\u2003\u2003EF-Tu WT[^d^](#ngtab1.4){ref-type=\"table-fn\"}              2.02 \u00b1 0.19                                                0.60 \u00b1 0.05                                      1.0 \u00b1 0.2             4.9 \u00b1 1.2\n  \u2003\u2003\u2003\u2003Ribosome WT[^e^](#ngtab1.5){ref-type=\"table-fn\"}           0.4 \u00b1 0.1                                                  0.7 \u00b1 0.05                                       0.1 \u00b1 0.1             1.2 \u00b1 0.5\n  \u2003\u2003\u2003\u2003EF-Tu^L148A^ mutant[^d^](#ngtab1.4){ref-type=\"table-fn\"}   3.1 \u00b1 0.3                                                  0.3 \u00b1 0.05                                       1.2 \u00b1 0.5             4.5 \u00b1 1.0\n  After Rif treatment                                                                                                                                                                              \n  \u2003\u2003\u2003\u2003EF-Tu WT[^d^](#ngtab1.4){ref-type=\"table-fn\"}              3.5 \u00b1 0.4                                                  0.35 \u00b1 0.05                                      1.5 \u00b1 0.5             4.9 \u00b1 1.5\n  \u2003\u2003\u2003\u2003EF-Tu^L148A^ mutant[^d^](#ngtab1.4){ref-type=\"table-fn\"}   5.2 \u00b1 0.4                                                  0.1 \u00b1 0.05                                       1.9 \u00b1 1.2             5.6 \u00b1 1.2\n\nNormal growth conditions were used, except for measurements after rifampin (Rif) treatment as noted.\n\nMean diffusion coefficient estimated from first two points of MSD plot ([Fig.\u00a02](#fig2){ref-type=\"fig\"}).\n\nBest-fit fractional population of the more slowly diffusing state. The fractional population of the more rapidly diffusing state is *f*~fast~ = 1 -- *f*~slow~.\n\nC terminus labeled with mEos2.\n\n30S subunits labeled by expression of the ribosomal protein S2-mEos2.\n\nWhile the static two-state model fits the data reasonably well ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}), if it were completely adequate then a value of \u03c7~\u03bd~^2^ as large as 1.24 would be statistically highly unlikely (*P* \\~ 0.01). Here we must recognize that the true diffusive behavior of EF-Tu is surely a composite of many diffusive states: free EF-Tu and free ternary complexes (to which the fast diffusion is assigned) and EF-Tu bound to 70S ribosomes and polysomes of variable length (to which the slow diffusion is assigned). Under our fast imaging conditions, the distribution of measured step lengths for the slower population is dominated by the measurement error, not by true displacement of the tracked species. There is also the likelihood of transitions between these states on the 2-ms time scale of the single-step displacement measurements (described below).\n\nWhat is robust in the fitting results is the fraction of rapidly diffusing copies having a *D*~fast~ value of \\~4.9 \u00b5m^2^/s. The best-fit fraction *f*~fast~ is 0.40 \u00b1 0.05 in the two-state modeling and 0.35 \u00b1 0.05 in the three-state modeling. Such a fraction of fast molecules is evidently necessary to fit the long tail on the distribution *P*~EF-Tu~(*r*) ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}; [Fig.\u00a0S1C](#figS1){ref-type=\"supplementary-material\"}), and that is the part of the distribution least perturbed by measurement error. In addition, the *D*~fast~ value of \\~4.9 \u00b5m^2^/s will be confirmed below in studies of cells treated with the drug rifampin (Rif). The main conclusion of this work---that \\~60% of EF-Tu copies are not in the rapidly diffusing states over the 2-ms frame time of the measurements---appears quite robust. In what follows, we proceed with further analysis of the two-state model results under the assumption that they represent the partitioning into ribosome-bound and unbound EF-Tu copies fairly accurately. Separate axial distributions for slow and fast steps (below) will further corroborate the assignments of the fast and slow components.\n\nThe best-fit value *D*~slow~ = 1.0 \u00b1 0.2\u00a0\u03bcm^2^/s for EF-Tu is 10 times larger than the estimated diffusion coefficient of the slow component of the ribosome distribution, *P*~ribo~(*r*), which has a diffusion coefficient of 0.1 \u00b1 0.1\u00a0\u03bcm^2^/s ([Table\u00a01](#tab1){ref-type=\"table\"}; see [Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"} in the supplemental material). Importantly, fits to two-state model functions with the slow diffusion constrained to match that of the 70S ribosomes were much worse ([Fig.\u00a0S1A](#figS1){ref-type=\"supplementary-material\"}). This suggests to us that the slow component of EF-Tu diffusion is itself a composite state comprising two substates that exchange with each other during the 2-ms camera frame: EF-Tu bound to 70S ribosomes (with mean lifetime \u03c4~on~) and free EF-Tu or free ternary complexes (with mean lifetime \u03c4~off~) sequestered in the ribosome-rich regions and diffusing freely between ribosome binding events. Here \u03c4~on~ is the mean time a ternary complex spends bound to a 70S ribosome and \u03c4~off~ is the mean time a ternary complex spends searching for a ribosomal binding site, with both times referring to ternary complexes within the ribosome-rich regions. If this is essentially correct, then we can infer (\u03c4~on~ + \u03c4~off~) \u2264 2\u00a0ms. If we assume that *D*~fast~ = 4.9\u00a0\u03bcm^2^/s applies to the free EF-Tu and ternary complex components in the ribosome-rich regions, then the sequestered EF-Tu copies are spending \\~80% of the time actually bound to ribosomes and \\~20% of the time in transit between ribosome-binding sites. Those are the population fractions that yield the correct weighted average diffusion coefficient: *D*~*slow*~ = 1.0\u00a0\u03bcm^2^/s = 0.2 \u00d7 4.9\u00a0\u03bcm^2^/s + 0.8 \u00d7 0.1\u00a0\u03bcm^2^/s. The corresponding lifetime ratio is a \u03c4~on~/\u03c4~off~ value of \\~4. According to this interpretation, within the ribosome-rich region EF-Tu copies are exchanging between the ribosome-bound and free EF-Tu states so fast that our 2-ms camera frames can only report on the average diffusive behavior of the bound and free states. As discussed below, such short on and off times make good biochemical sense.\n\n10.1128/mBio.02143-17.3\n\nThe single-step displacement probability distribution *P*~ribo~(*r*) for 30S ribosomal subunits labeled by the protein S2-Eos2. The best-fit model of two static states is shown. The slow fraction (blue; *f*~slow~ = 0.7, *D*~slow~ = 0.1 \u00b5m^2^/s) is assigned to translating 70S ribosomes. The fast fraction (red; *f*~fast~ = 0.3, *D*~fast~ = 1.2 \u00b5m^2^/s) is assigned to free 30S subunits. Download FIG\u00a0S2, TIF file, 7 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02143-17.4\n\nThe axial distribution of slow and fast single steps for mutant protein EF-Tu^L148A^. Slow and fast steps are based on the same cutoffs used for WT EF-Tu. The axial coordinates are normalized to \u22120.5 to 0.5. The slow distribution (blue; *r* \\< 0.1\u00a0\u00b5m) seems fairly homogeneous, but the fast distribution (red; *r* \\> 0.2\u00a0\u00b5m) shows two distinct peaks near the DNA regions. The ribosome axial distribution (black) is also plotted for comparison. Download FIG\u00a0S3, TIF file, 7 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nTo test the assignment of the slow population to ribosome-bound EF-Tu, we plotted separate axial location distributions for the slowest (step length, *r* \\< 0.1\u00a0\u03bcm) and fastest (*r* \\> 0.2\u00a0\u03bcm) components of *P*~EF-Tu~(*r*). The arrows in [Fig.\u00a03A](#fig3){ref-type=\"fig\"} mark these cutoffs. According to the best two-state model, the slow cutoff includes steps of which \\~80% belong to the slow population, while the fast cutoff includes steps of which \\~90% belong to the fast population. The location of each step was assigned as the midpoint of the first and second locations, and the axial coordinates were scaled and normalized as before. The results are shown in [Fig.\u00a03B](#fig3){ref-type=\"fig\"} in comparison with the total ribosome axial distribution. The three-peaked distribution of slow steps extends into the end caps as the ribosomes do. The distribution of fast steps avoids the ribosome-rich end caps and is perhaps mildly concentrated in the nucleoid regions. These results are consistent with the slow population preferentially residing within the ribosome-rich regions due to transient binding to 70S and the fast population preferentially residing within the nucleoids.\n\nEffects of rifampin. {#s1.3}\n--------------------\n\nTo better characterize the diffusive properties of free EF-Tu/ternary complex, we treated exponentially growing cells with 250\u00a0\u03bcg/ml of the antibiotic rifampin (Rif) for 3\u00a0h prior to plating and imaging of EF-Tu--mEos2. Rif halts transcription and thus effectively stops mRNA production ([@B31], [@B32]). On a time scale of 10\u00a0min, the existing mRNA is degraded. Lacking mRNA to translate, the 70S polysomes dissociate into free 50S and 30S subunits. We used 792 trajectories that lasted at least 6 steps or longer from 58 cells to plot the spatial distribution of EF-Tu under Rif treatment. The selected cell lengths varied from 3 to 4\u00a0\u00b5m; after Rif treatment, the distribution of cell lengths shifts toward smaller values. The heat map shows a fairly uniform distribution of EF-Tu along the long axis of the cell, but with the end caps partially excluded (see [Fig.\u00a0S4A](#figS4){ref-type=\"supplementary-material\"} in the supplemental material). As shown earlier ([@B31]), under Rif treatment the nucleoids expand to fill the cytoplasmic volume fairly homogeneously. The 30S and 50S ribosomal subunits mix with the expanded DNA; they also occupy the cytoplasmic volume fairly uniformly. The EF-Tu distribution is similar.\n\n10.1128/mBio.02143-17.5\n\n\\(A\\) The spatial distribution heat map of EF-Tu after Rif treatment is shown for *E.\u00a0coli* cells 3 to 4 \u00b5m in length. Scale is proportional to the counts per pixel. (B) Axial distribution of EF-Tu after Rif treatment (black) along with a simulated homogeneous distribution (red). (C) *P*~EF-Tu~(*r*) distribution of 7,086 steps for EF-Tu after Rif treatment, along with the best-fit parameters for a model of two static states. Download FIG\u00a0S4, TIF file, 17.1 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nWe used 1,181 trajectories from 78 cells for the EF-Tu diffusive state analysis after Rif treatment. All trajectories of 6 steps or longer were truncated at the sixth step as before. The mean EF-Tu diffusion coefficient obtained from the MSD(\u03c4) plot increases to 3.5 \u00b1 0.4\u00a0\u00b5m^2^/s ([Fig.\u00a02](#fig2){ref-type=\"fig\"}). This is larger than that of EF-Tu in normally growing cells, 2.02 \u00b1 0.09\u00a0\u00b5m^2^/s. Accordingly, under Rif treatment, the two-state analysis of *P*~EF-Tu~(*r*) ([Fig.\u00a0S4C](#figS4){ref-type=\"supplementary-material\"}) finds *f*~slow~ = 0.35 \u00b1 0.05 of EF-Tu that moves with *D*~slow~ = 1.5 \u00b1 0.5 \u00b5m^2^/s, slightly larger than the value of *D*~slow~ = 1.0 \u00b1 0.2\u00a0\u03bcm^2^/s in untreated cells ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}). A larger fraction (*f*~fast~ = 0.65 \u00b1 0.05) of EF-Tu moves with the same *D*~*fast*~ = 4.9 \u00b1 1.5 \u00b5m^2^/s found for untreated cells. The results after Rif treatment suggest the possibility of some residual binding of EF-Tu/ternary complex to ribosomal subunits, perhaps to the same L7/L12 binding sites on 50S. This is only a suggestion, but it is supported by the results for a mutated variant of EF-Tu presented next.\n\nEF-Tu^L148A^ mutant. {#s1.4}\n--------------------\n\nRodnina and coworkers ([@B5]) studied the effects of point mutations within the C terminus of L7/L12 and within helix D of EF-Tu on the kinetics of initial binding of ternary complex to ribosomes. The mutation sites were chosen by analogy to the well-characterized structure of the EF-Ts/EF-Tu complex. The mutations that caused a substantial decrease in the association rate constant *k*~1~ were used to model the important contacts in the complex between L7/L12 and EF-Tu. The particular mutation L148A in EF-Tu decreased *k*~1~ by a factor of 5. To probe this interaction *in vivo*, we engineered a plasmid containing the same L148A mutation to EF-Tu appended to a C-terminal mEos2 label ([Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"}). The mutated protein was expressed in the same background strain, VH1000, along with WT protein expressed normally from the chromosome to enable normal cell growth.\n\nWe obtained 1,160 trajectories of 6 steps or longer from 153 cells to study the diffusion of EF-Tu^L148A^--mEos2. The mean diffusion coefficient from the MSD plot is 3.1 \u00b1 0.3\u00a0\u00b5m^2^/s ([Fig.\u00a02](#fig2){ref-type=\"fig\"}). This is larger than the mean value 2.02 \u00b1 0.19\u00a0\u00b5m^2^/s for normal EF-Tu--mEos2, consistent with a smaller degree of binding of the mutated protein to ribosomes. Accordingly, the two-component *P*(*r*) analysis of mutant protein diffusion finds *f*~slow~ = 0.30 \u00b1 0.05 (2-fold smaller than for the normal protein) with *D*~slow~ = 1.2 \u00b1 0.5 \u00b5m^2^/s and *f*~fast~ = 0.70 \u00b1 0.05 with *D*~fast~ = 4.5 \u00b1 1.0 \u00b5m^2^/s ([Fig.\u00a04C](#fig4){ref-type=\"fig\"}). The location heat map and the axial spatial distribution for the EF-Tu^L148A^ mutant ([Fig.\u00a04A](#fig4){ref-type=\"fig\"} and [B](#fig4){ref-type=\"fig\"}) show that the mutated protein is fairly uniformly distributed throughout the cell, with only a hint of three peaks. These results indicate substantially less binding of the EF-Tu^L148A^ mutant to ribosomal sites, in qualitative agreement with the mutation studies *in vitro* ([@B5]). The agreement helps to corroborate our underlying assumption that ternary complexes are binding to L7/L12 ribosomal subunits *in vivo*; see Discussion for a summary of additional biochemical evidence.\n\n![(A) Composite spatial distribution heat map of the mutant form EF-Tu^L148A^--mEos2 for 792 localizations from 123 *E.\u00a0coli* cells of length 4 to 5.5\u00a0\u00b5m. Pixels are \\~45 by 45 nm. The intensity scale shows relative counts per pixel. (B) Axial distributions of EF-Tu^L148A^ mutant (black) in comparison with ribosomes (30S-mEos2 labeling \\[red\\]). The distributions are normalized to the same area and plotted on a relative scale of \u22120.5 to +0.5 for the long axis. (C) Distribution of single-step displacements *P*(*r*) (gray histogram) for 6,960 steps of the EF-Tu^L148A^ mutant. The solid black line shows the best-fit model using two static states: \"slow\" (blue) and \"fast\" (red). Model parameters: *f*~slow~ = 0.3, *D*~slow~ = 1.2 \u00b5m^2^/s, *f*~fast~ = 0.7, and *D*~fast~ = 4.5 \u00b5m^2^/s.](mbo0011836660004){#fig4}\n\nTo control for possible effects of overexpression of the L148A mutant from the plasmid, we constructed an analogous plasmid that expresses WT EF-Tu--mEos2 and incorporated it into the same VH1000 background strain. The spatial distribution and diffusive properties of the EF-Tu--mEos2 copies expressed from the plasmid were qualitatively similar to those of EF-Tu--mEos2 expressed from the chromosome (see [Fig.\u00a0S5](#figS5){ref-type=\"supplementary-material\"} in the supplemental material).\n\n10.1128/mBio.02143-17.6\n\nSingle-step displacement probability distribution *P*~EF-Tu~(*r*) for EF-Tu expressed from a plasmid. The best-fit two-state model parameters are as follows: blue curve, *f*~slow~ = 0.7 and *D*~slow~ = 1.3 \u00b5m^2^/s; red curve, *f*~fast~ = 0.3 and *D*~*fast*~ = 4.9 \u00b5m^2^/s. The diffusion of WT EF-Tu expressed from a plasmid is similar to the diffusion of WT EF-Tu expressed from the chromosome ([Fig.\u00a03](#fig3){ref-type=\"fig\"}). Download FIG\u00a0S5, TIF file, 7.2 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02143-17.7\n\n\\(A\\) Simulated distributions of the mean of six single-step estimates of the diffusion coefficient, *P*~model~(*D*~*i*~), for two diffusion coefficients, *D*~slow~ and *D*~fast~, as indicated. These distributions were used to estimate cutoff values for slow and fast copies of EF-Tu, whose experimental distribution is shown in panel (B). (C) Separate MSD plots for slow and fast EF-Tu copies. Intercepts were used to set different values of localization error: \u03c3~slow~ = 40\u00a0nm and \u03c3~fast~ = 80\u00a0nm. Download FIG\u00a0S6, TIF file, 16.3 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02143-17.8\n\nPlanes passing through the three-dimensional grid search for the best \u03c7~\u03bd~^2^ value for the two-state fits to WT EF-Tu distribution *P*~EF-Tu~(*r*) of [Fig.\u00a03](#fig3){ref-type=\"fig\"}. Each slice shows a plane passing through the global minimum parameter set: (A) *f*~slow~ is fixed at 0.60. (B) *D*~slow~ is fixed at 1.0 \u00b5m^2^/s. (C) *D*~fast~ is fixed at 4.9 \u00b5m^2^/s. The uncertainty estimate for each parameter was chosen to enclose all values of \u03c7~\u03bd~^2^ within 0.5 U of the minimum value, as shown by the boxed-in regions of each plane. Fits with still larger values of \u03c7~\u03bd~^2^ were judged to be qualitatively poor. Download FIG\u00a0S7, TIF file, 23.5 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nTo test for possible binding of the mutant form EF-Tu^L148A^ to free 50S ribosomal subunits, we obtained 993 trajectories of 6 steps or longer from 83 cells after the 3-h Rif treatment. The slope of the MSD plot increases to 5.2 \u00b1 0.4\u00a0\u03bcm^2^/s ([Fig.\u00a02](#fig2){ref-type=\"fig\"}), compared with 3.5 \u00b1 0.4\u00a0\u00b5m^2^/s for WT EF-Tu after Rif. The *P*(*r*) distribution is fit qualitatively by a single population with *D* = 5.7 \u00b1 1.0\u00a0\u03bcm^2^/s (\u03c7~\u03bd~^2^ = 1.5). The best two-component fit yielded *f*~slow~ = 0.10 \u00b1 0.05, *D*~slow~ = 1.9 \u00b1 1.2\u00a0\u03bcm^2^/s, *f*~fast~ = 0.90 \u00b1 0.05, *D*~fast~ = 5.6 \u00b1 1.2\u00a0\u03bcm^2^/s, and \u03c7~\u03bd~^2^ = 1.2. The analysis indicates that after Rif treatment, most EF-Tu^L148A^ is diffusing essentially freely, perhaps primarily as ternary complexes. Evidently the mutant protein exhibits little or no binding to free 30S or 50S subunits.\n\nNumerical estimates and comparisons with theory. {#s1.5}\n------------------------------------------------\n\nThe present results can be combined with literature estimates for relative copy numbers of ribosomes, EF-Tu, EF-G, tRNAs, and aa-tRNA synthetases to provide semiquantitative insight into the partitioning of EF-Tu and tRNA across functional states and the time-averaged stoichiometry of the species bound to a translating ribosome. Under the same growth conditions used here (30\u00b0C in EZRDM), we previously estimated \\~50,000 30S ribosomal subunits per cell, some 80% of which (\\~40,000 copies) are engaged as translating 70S ribosomes ([@B15]). Mean copy number estimates for EF-Tu, total tRNA, EF-G, and total aa-tRNA synthetase ([Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"}) were derived from the ribosome copy number and from literature values of the ratio of each species' copy number to that of ribosomes. It was not possible to match strains, growth conditions, growth rates, and temperatures, so we chose to match only the growth rate (\\~1 doubling/h). We hope these rough estimates will help constrain future models of overall *E.\u00a0coli* translation rates. Their biological significance will be discussed further below. Details of the calculations and underlying assumptions are provided in the supplemental material; here we summarize the estimates. The primary assumption is that EF-Tu binds to translating ribosomes via contact with the C terminus of L7/L12. This is justified in the Discussion.\n\nThe time-averaged stoichiometry of EF-Tu and tRNA binding to a translating 70S ribosome can be estimated from the fraction of EF-Tu copies bound to ribosomes combined with copy number estimates from other studies. There are 61 different codons and 43 different aa-tRNA types (43 different ternary complexes) used by *E.\u00a0coli* ([@B9]). Forty-eight codons match only one type of ternary complex, 12 match two types, and one matches three types. This means that the ribosome is usually testing and rejecting noncognate or near-cognate aa-tRNAs. The A site is most frequently occupied by an aa-tRNA within its ternary complex, still tethered to L7/L12 (prior to codon selection, GTP hydrolysis, and ejection of EF-Tu) ([@B9]).\n\nUnder our growth conditions of 30\u00b0C in EZRDM, we estimate the following mean copy numbers per cell: \\~40,000 translating 70S ribosomes (concentrated in three ribosome-rich regions) ([@B15]), \\~10,000 free 30S subunits, \\~10,000 free 50S subunits, \\~350,000 total EF-Tu copies ([@B10], [@B33]), \\~350,000 total tRNA copies ([@B34]), \\~50,000 EF-G copies (which compete with EF-Tu for L7/L12 binding sites) ([@B33]), and \\~50,000\u00a0aa-tRNA synthetases ([@B33]). The new data suggest that \\~210,000 EF-Tu copies (60% \\[the \"slow\" copies\\]) are ternary complexes that occupy the ribosome-rich regions, where they are bound to 70S ribosomes \\~80% of the time (\\~170,000 ribosome-bound ternary complexes plus \\~40,000 free ternary complexes). Thus, we estimate as many as \\~170,000/40,000 = 4 ternary complexes bound to each translating ribosome. This indicates that the four L7/L12 subunits in *E.\u00a0coli* are essentially saturated with ternary complexes. However, EF-G must also bind to L7/L12 in order to drive translocation on those rare occasions when a cognate aa-tRNA is accommodated in the A site and forms a new peptide bond. In our estimate, we assume the average occupancy of the four L7/L12 units is 3.5 ternary complexes and 0.5 EF-G copy. The remaining \\~180,000 EF-Tu copies partition into \\~70,000 free ternary complexes plus \\~110,000 free (bare) EF-Tu copies. The overall partitioning of EF-Tu between ternary complexes and free EF-Tu is corroborated by an equilibrium calculation based on the aa-tRNA/EF-Tu binding constant *in vitro* ([@B35]).\n\nFor partitioning of the \\~350,000 total tRNA copies, we estimate that on average each 70S ribosome binds one tRNA at the A site (usually tethered to L7/L12 by a bridging EF-Tu), one in the P site, one-half tRNA in the E site (an average over \"2-1-2\" and \"2-3-2\" models \\[[@B36], [@B37]\\]), plus an additional \\~2.5 tRNAs bound to the other three L7/L12 sites. Recall that EF-G is assumed to take up 0.5 L7/L12 binding site. Thus, averaged over time, \\~5 tRNAs are bound to each 70S ribosome (\\~200,000 tRNAs bound to \\~40,000 translating ribosomes, comprising \\~140,000 tRNAs within ternary complexes and \\~60,000 tRNAs at the P and E sites). The remaining \\~150,000 tRNA copies not bound to 70S are estimated to partition among three states: \\~50,000 copies being recharged by aa-tRNA synthetases, \\~100,000 copies within free ternary complexes, and only \\~400 free tRNAs. These estimates are based in part on an equilibrium calculation using the *in vitro* binding constant of aa-tRNA with EF-Tu ([@B35]).\n\nIn addition, our new data are in sensible agreement with two rate constants from a model that optimally scaled a detailed set of *in vitro* rate constants to derive a set of theoretical *in vivo* rate constants describing the multistep process of the elongation cycle ([@B9]). Again, details are presented in the supplemental material. First we use the pseudo-first-order rate \u03c4~off~^\u22121^ and the 70S ribosome concentration to estimate a lower limit on the effective bimolecular association rate constant *k*~1~ for binding of a typical noncognate ternary complex to an L7/L12 subunit of a 70S ribosome within the ribosome-rich regions. The result is *k*~1~ = \u03c4~off~^\u22121^/\\[70S\\] \u2265 4.5 \u00d7 10^7^ M^\u22121^ s^\u22121^. This is remarkably fast, at least 1/6 of the calculated diffusion-limited rate constant *k*~diff~ = 3.2 \u00d7 10^8^ M^\u22121^ s^\u22121^. As suggested earlier ([@B4]), *k*~1~ (which is expressed on a per ribosome basis) may be especially large due to the four L7/L12 binding sites per ribosome and the length and flexibility of the linkages between ribosome and the C-terminal domain of L7/L12. The theoretical *in vivo* estimate for the analogous \u03ba~on~\\* (see Table\u00a02 in reference [@B9]) at 1.07 doublings/h and 37\u00b0C is 9.4 \u00d7 10^7^\u00a0M^\u22121^ s^\u22121^, 2-fold larger than our lower limit on *k*~1~.\n\nWe can also compute a lower limit on the unimolecular dissociation rate of EF-Tu (usually as part of a ternary complex) from the ribosome, *k*~\u22121~ = \u03c4~on~^\u22121^ \u2265 625\u00a0s^\u22121^ at 30\u00b0C. The value of *k*~*\u2212*1~ is temperature sensitive. If we apply an Arrhenius-based correction factor of 2.1 to our *k*~*\u2212*1~ value at 30\u00b0C (details in supplemental material), the estimated value at 37\u00b0C becomes *k*~\u22121~ \u2265 1,250\u00a0s^\u22121^. This is consistent with the theoretical *in vivo* rate constant for 1.07 doublings/h at 37\u00b0C, \u03c9~off~\\* = 1,700\u00a0s^\u22121^ (see Table\u00a02 in reference [@B9]).\n\nDisagreement with a recent tRNA tracking study. {#s1.6}\n-----------------------------------------------\n\nIn violation of the standard model of aa-tRNA recruitment, a recent single-tRNA tracking experiment from the Kapanidis lab inferred that a large majority of tRNA copies exist as free tRNA, bound neither to EF-Tu in ternary complexes nor to the aminoacyl-tRNA synthetase ([@B20]). They electroporated a small number of tRNA copies fluorescently labeled with Cy5 dye into *E.\u00a0coli* and tracked the motion of single molecules. A large fraction (70 to 90%) of the tRNA-Cy5 copies diffused very rapidly (corrected *D*~tRNA~ value of \\~8\u00a0\u03bcm^2^/s). These copies were attributed to free tRNA (not bound within ternary complexes). The conclusion was that diffusion of free aa-tRNA, not ternary complexes, must be the primary means of delivery of aa-tRNA to the ribosomal A site. The remarkably large fraction of free tRNA copies was deemed possible based on the assumption that only two tRNA copies are bound to each ribosome (one each in the A and P sites). The rationale given for the small estimated fraction of ternary complexes ([@B20]) was that EF-Tu can bind to membrane-bound MreB, as evidently occurs in both *Bacillus subtilis* and *E.\u00a0coli* ([@B38][@B39][@B40]). This would remove EF-Tu from the cytoplasm and make it less available for ternary complex formation. However, the EF-Tu copy number is about 100 times larger than that of MreB ([@B33]). In addition, we find no evidence in our EF-Tu spatial distribution of significant binding to the cytoplasmic membrane, where MreB resides. In contrast, our numerical estimates based on an average of \\~3.5 ternary complexes bound to the four L7/L12 sites indicate \\~5 bound tRNA copies per ribosome. Finally, our equilibrium calculations suggest that only \\~1% or less of total tRNA should exist as free tRNA.\n\nOne potential weakness of the electroporation method ([@B20]) is that the few labeled tRNA copies in each cell must compete with the 350,000 endogenous tRNA copies for aminoacylation, ternary complex formation, and binding and processing by the ribosome. Although the labeled tRNA-Cy5 species was shown to be functional *in vitro*, it is difficult to know how well tRNA-Cy5 copies compete with endogenous copies in each functional step *in vivo*. It seems possible that the synthetase recognizes tRNA-Cy5 poorly, aa-tRNA--Cy5 forms ternary complexes poorly *in vivo* or these complexes bind 70S ribosomes weakly, or Cy5 fluorescence is somehow quenched in ternary complexes so that they are not detected.\n\nDISCUSSION {#s2}\n==========\n\nRapid testing of aa-tRNA copies for a codon-anticodon match. {#s2.1}\n------------------------------------------------------------\n\nIn rapidly growing *E.\u00a0coli*, the mean protein elongation rate can be as fast as 20\u00a0amino acids/s. Single elongation cycles must be carried out in less than \\~50\u00a0ms ([@B7]). There are 61 different codons and 43 different aa-tRNA types (43 different ternary complexes) ([@B9]). Forty-eight codons match only one type of ternary complex, 12 match two types, and 1 matches three types. Fully 40 unique codons are used with at least 1% frequency ([@B41]). For a given mRNA codon poised at the 30S decoding site, the average chance that a particular ternary complex carries a cognate (completely matching) aa-tRNA anticodon is roughly 1 in 40. This means on average, approximately 40 different ternary complexes must be sampled before a cognate aa-tRNA is found. (See the supplemental material for the probabilistic calculation.) Sampling and testing of these complexes must occur faster than the complete elongation cycle time of 50\u00a0ms, suggesting an upper limit of \\~1\u00a0ms on the average time taken for ternary complex evaluation.\n\nSelection for cognate aa-tRNA is a two-stage process ([@B1], [@B9]). Essentially all noncognate ternary complexes and a large majority of near-cognate ternary complexes dissociate from L7/L12 in the initial recognition stage, prior to GTP hydrolysis by EF-Tu. This can be seen from the \"theoretical *in vivo*\" rate constants of Lipowsky and coworkers ([@B9]). Those events should dominate our single-molecule observations. The small fraction of near-cognate ternary complexes that pass through the initial stage is efficiently rejected in the proofreading stage, which occurs after GTP hydrolysis ([@B9]). Only cognate aa-tRNAs move forward rapidly through both stages, efficiently achieving A state accommodation.\n\nOur single-molecule tracking study provides some new insight into the spatial distribution and time scale of binding and unbinding events between EF-Tu (ternary complexes) and translating ribosomes in *E.\u00a0coli*. These methods cannot dissect binding events for cognate versus near-cognate versus noncognate ternary complexes. Instead, the measurements probe the time scale of the initial, codon-independent binding and unbinding with L7/L12. The new *in vivo* results corroborate several mechanistic inferences previously gleaned from a large body of *in vitro* kinetics measurements ([@B1]). Evidently the high concentration of ternary complexes, the segregation of 70S ribosomes in the ribosome-rich regions of the cytoplasm, the presence of four L7/L12 binding sites per 70S ribosome, and the flexible attachment of the L7/L12 binding sites to the ribosome all combine to enable extremely rapid sampling of aa-tRNA copies by the 70S ribosome.\n\nOur interpretation of *D*~slow~ = 1\u00a0\u03bcm^2^/s as arising from a composite state involving rapid exchange between 80% ribosome-bound ternary complexes (\u03c4~on~) and 20% free ternary complexes (\u03c4~off~) within the ribosome-rich regions led to the inequality (\u03c4~on~ + \u03c4~off~) \u2264 2\u00a0ms. This result is consistent with the requisite fast sampling and rejection of ternary complexes required by the predominance of noncognate and near-cognate aa-tRNAs. The estimated lower bounds on the bimolecular binding rate constant *k*~1~ and the unimolecular dissocation rate *k*~\u22121~ are consistent with recent theoretical estimates of the analogous in vivo rate constants ([@B9]). The novel method used for scaling of *in vitro* rates to find the optimal set of *in vivo* rates that match the overall *E.\u00a0coli* translation rate seems remarkably successful.\n\nRibosomal L7/L12 sites bind multiple ternary complexes simultaneously. {#s2.2}\n----------------------------------------------------------------------\n\nThe new data provide strong evidence that multiple ternary complexes bind simultaneously to the four L7/L12 sites on the 50S subunit of translating ribosomes. Our partitioning analysis suggests that the four L7/L12 sites may be saturated with ternary complexes on average. Such a high local concentration of tethered aa-tRNAs would greatly facilitate the rapid sampling required for efficient protein elongation, as previously suggested ([@B4]). The enhanced sampling rate would arise from two effects. During the same time interval in which one of the bound ternary complexes is being tested, any open L7/L12 site can be replenished with a fresh ternary complex. This saves time. In addition, when an A site comes open after a codon match and translocation or (more typically) after rejection of a noncognate aa-tRNA, the diffusive search for the open A site by a new ternary complex would be more rapid due to the high local concentration and the spatial constraints imposed by the tethering.\n\nThere is extensive biochemical evidence *in vitro* supporting our underlying assumption that aa-tRNA--EF-Tu(GTP) ternary complexes bind the ribosome via contact between L7/L12 and EF-Tu. A comprehensive summary is provided in reference [@B4]. As shown schematically in [Fig.\u00a01A](#fig1){ref-type=\"fig\"}, L7/L12 comprises an N-terminal dimerization module and a globular C-terminal domain (CTD) connected by a flexible hinge. In *E.\u00a0coli*, four copies of L7/L12 are bound to L10, which is itself flexible. An early chemical cross-linking and fluorescence study implicated L7/L12 in the binding of EF-Tu to the ribosome ([@B6]). Subsequent extraction/complementation experiments showed that the presence of L7/L12 was required for binding of both EF-Tu and EF-G to the ribosome ([@B42]). Specific point mutations in the L7/L12 CTD and in the G domain of EF-Tu affected binding of ternary complexes to the ribosome ([@B5]). In addition, there is homology between the proposed L7/L12 binding interface to EF-Tu and the well-characterized structure of the EF-Ts/EF-Tu complex. The L7/L12 subunits do not appear in crystal structures of 70S ribosomes ([@B2], [@B43]). However, the biochemical evidence is corroborated by an early reconstruction from cryo-EM data with a 1.8-nm resolution that shows density connecting the G domain of EF-Tu within a ternary complex to the L7/L12 stalk of the ribosome ([@B13]). Finally, the correspondence between the diminished binding of the mutant form EF-Tu^L148A^ *in vitro* ([@B5]) and in live *E.\u00a0coli* cells ([Fig.\u00a04](#fig4){ref-type=\"fig\"}) corroborates the assertion that we are probing ternary complex binding to L7/L12.\n\nWahl and coworkers ([@B4]) combined biochemical and additional structural evidence to propose the model of the stalk that we reproduce schematically in [Fig.\u00a01A](#fig1){ref-type=\"fig\"}. The schematic shows four ternary complexes bound to the ribosome via the four L7/L12 CTDs. One ternary complex is undergoing codon testing at the A site, while the other three are tethered and awaiting testing. The flexible attachment of the four L7/L12 CTDs to the ribosome is likely to facilitate efficient capture of ternary complexes. The long, flexible linkers may enable the CTDs to \"reach out and catch\" ternary complexes that come into near proximity of the ribosome body ([@B4], [@B44]). Although there is no detailed structural evidence supporting the simultaneous binding of four ternary complexes, the concept is supported by our stoichiometric estimates in vivo. This concept is also supported by the remarkably large bimolecular rate constant for ternary complex binding to 70S, measured earlier *in vitro* and now estimated *in vivo*.\n\nConclusions. {#s2.3}\n------------\n\nThe present work provides strong evidence that multiple ternary complexes bind the four L7/L12 initial binding sites on the 50S subunit of the 70S ribosome simultaneously. We also provide a new estimate of \\~1 to 2\u00a0ms or less for the *in vivo* time scale of binding and unbinding of noncognate ternary complexes during the initial anticodon test. Semiquantitative estimates of the partitioning of EF-Tu and tRNA among different binding states should help constrain models of translation in *E.\u00a0coli*. In future work, tracking studies of EF-G could provide an independent estimate of the fraction of EF-G bound to 70S ribosomes at a given moment in time. That would shed light on the competition *in vivo* between EF-Tu and EF-G for L7/L12 binding sites on the 70S ribosome.\n\nMATERIALS AND METHODS {#s3}\n=====================\n\nBacterial strains. {#s3.1}\n------------------\n\nWe chose 30\u00b0C for this study because the mEos2 labels fluoresce poorly at 37\u00b0C; also, 30\u00b0C matches the conditions of our earlier study of ribosome copy number, a result used here ([@B19]). The strains, doubling times, and oligonucleotides used are detailed in [Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"}. In *E.\u00a0coli*, EF-Tu is expressed from two essentially identical genes: *tufA* and *tufB*. Both of these genes were first labeled endogenously via the lambda red technique ([@B45]) in the background strain, NCM3722. The photoconvertible fluorescent protein mEos2 was covalently bound to the C terminus of EF-Tu. These genes were then transferred to the VH1000 background strain using P1 transduction. For studies of the mutated protein EF-Tu^L148A^, the *tufA* gene was point mutated from Leu to Ala at the 148th residue in a plasmid with ampicillin resistance, pASK-IBA3+. The plasmid mutation included a fusion of the same mEos2, again at the C terminus of the protein. To control for possible effects of overexpression, we also prepared a strain including a completely analogous plasmid, except that it lacked the point mutation. The 30S ribosomal subunits were labeled by expression of the protein S2-mEos2 from the chromosome.\n\nIn \"EZ rich, defined medium\" (EZRDM) at 30\u00b0C, the doubling time of the endogenously labeled *tufA* and *tufB* strain is 60 \u00b1 3\u00a0min ([Table\u00a0S1A](#tabS1){ref-type=\"supplementary-material\"}; [Fig.\u00a0S8](#figS8){ref-type=\"supplementary-material\"}). This is \\~1.3 times longer than the doubling time of the VH1000 background strain, which is 45 \u00b1 2\u00a0min, indicating that the mEos2 label enables fairly normal functionality of EF-Tu. The L148A mutant strain has a doubling time of 46 \u00b1 4\u00a0min.\n\n10.1128/mBio.02143-17.9\n\n\\(A\\) Growth curves in EZRDM at 30\u00b0C for three different strains: VH1000 (WT parent strain, no labeling), MSG195 (VH1000 with ribosomal protein S2 labeled with mEos2), and tufAB (*tufA* and *tufB* both labeled with mEos2). The exponential-phase data points are used to calculate the doubling times. (B and C) Plots of ln OD/ln2 versus time for the VH1000 and tufAB strains, respectively. The exponential-phase data points are fitted to a straight line whose inverse slope yields doubling times of 45 \u00b1 2\u00a0min for VH1000 and 60 \u00b1 3\u00a0min for tufAB. Download FIG\u00a0S8, TIF file, 10.5 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nCell growth and preparation for imaging. {#s3.2}\n----------------------------------------\n\nBulk cultures from frozen glycerol stock solution and subcultures for imaging were grown overnight at 30\u00b0C with continuous shaking in EZRDM, which is a morpholinepropanesulfonic acid (MOPS)-buffered solution with supplemental metal ions (M2130; Teknova), glucose (2\u00a0mg/ml), supplemental amino acids and vitamins (M2104; Teknova), nitrogenous bases (M2103; Teknova), 1.32\u00a0mM K~2~HPO~4~, and 76\u00a0mM NaCl. The next day, the stationary-phase culture was diluted 100-fold in fresh EZRDM and grown again to exponential phase (optical density \\[OD\\] of 0.2 to 0.5). Cells were then plated on a polylysine-coated coverslip that formed the floor of a CoverWell perfusion chamber (Invitrogen, Carlsbad, CA) with a well volume of 140\u00a0\u00b5l.\n\nFor the L148A mutant strain, when the culture reached exponential phase it was treated with anhydrous tetracycline (final concentration, 45\u00a0nM) to induce expression of EF-Tu^L148A^--mEos2 from the plasmid. Tetracycline was washed away after 5\u00a0min of induction, and the cells were grown for 30\u00a0min more in fresh medium prior to plating and imaging. To test for the effects of treatment by rifampin (Rif), cells were grown to the exponential phase, after which Rif was added to a final concentration of 250\u00a0\u00b5g/ml. The culture remained at 30\u00b0C for 3\u00a0h, after which cells were plated and imaged.\n\nSuperresolution imaging of live *E.\u00a0coli* cells. {#s3.3}\n------------------------------------------------\n\nImaging of cells began within 5\u00a0min of plating. Individual fields of view were imaged no longer than 20\u00a0s to minimize laser damage. Each prepared sample was imaged for no longer than 30\u00a0min, during which time cells continued to grow normally. Cells were imaged on an inverted microscope (model Eclipse-Ti; Nikon Instruments, Melville, NY) equipped with an oil immersion objective (CFI Plan Apo Lambda DM 100\u00d7 oil, 1.45 NA; Nikon Instruments), a 1.5\u00d7 tube lens, and the Perfect Focus system (Nikon Instruments, Melville, NY). The fluorescence images were recorded on a back-plane illuminated electron-multiplying charge-coupled device (EMCCD) camera (iXon DV-887; Andor Technology, South Windsor, CT) at the rate of 485\u00a0Hz (\\~2\u00a0ms/frame). The camera chip consisted of 128 by 128\u00a0pixels (px), each 24 by 24 \u00b5m. The fluorescent protein mEos2 was activated using a 405-nm laser (CW laser; CrystalLaser, Reno, NV); the photoswitched state was subsequently excited with a 561-nm laser (Sapphire CW laser; Coherent, Inc., Bloomingfield, CT). Both lasers illuminated the sample for the entire duration of image acquisition. Emission was collected through a 617/73 bandpass filter (bright line\u00a0617/73-25; Semrock, Rochester, NY). The 405-nm power density at the sample was \\~5 to 10\u00a0W/cm^2^, which kept the number of activated molecules at less than two in each camera frame. The 561-nm laser power density at the sample was \\~8\u00a0kW/cm^2^.\n\nSingle-molecule image analysis. {#s3.4}\n-------------------------------\n\nThe fluorescent images were analyzed using a MATLAB graphical user interface (GUI) developed in our lab ([@B23]). Noise was attenuated using 2 different digital filters. After filtering, fluorescent signals were identified using a peak-finding algorithm with a user-defined intensity threshold with pixel-level accuracy. A particle is identified if the local intensity maximum is higher than the threshold. The threshold is carefully chosen to be large enough so that the algorithm can distinguish between background and signal and small enough to avoid cutting trajectories unduly short.\n\nA centroid algorithm was used to locate the identified particles with subpixel resolution ([@B23]). Rapidly moving molecules have images that are blurred asymmetrically due to diffusion from the camera frame. Centroid fitting can locate the particles with better accuracy than Gaussian fitting. The centroid algorithm is also faster computationally. A 7- by 7-px box was drawn around the intensity maxima, and the centroid of all the pixel intensities within the box was calculated. The centroid positions from successive frames were connected to form a trajectory only if they lie within 3\u00a0px = 480\u00a0nm of each other. A modi\ufb01ed MATLAB version of the tracking program written by Crocker and Grier ([@B46]) was used.\n\nAnalysis of diffusive behavior. {#s3.5}\n-------------------------------\n\nDetails of spatial distribution, mean-square displacement plots, trajectory simulations, two-state modeling of *P*(*r*) distributions, and estimation of uncertainties in fitting parameters are provided in the supplemental material.\n\n10.1128/mBio.02143-17.1\n\nThe text file describes the data analysis and fitting procedures in detail. It also describes the probabilistic argument for the number of trials required to find an aa-tRNA whose anticodon matches the current mRNA codon and provides a detailed description of the four-ternary-complex binding model. Download TEXT\u00a0S1, PDF file, 0.3 MB.\n\nCopyright \u00a9 2018 Mustafi and Weisshaar.\n\n2018\n\nMustafi and Weisshaar\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n**Citation** Mustafi M, Weisshaar JC. 2018. Simultaneous binding of multiple EF-Tu copies to translating ribosomes in live *Escherichia coli*. mBio 9:e02143-17. <https://doi.org/10.1128/mBio.02143-17>.\n\nWe thank Suparna Sanyal (Uppsala University) and Terence Hwa (University of California, San Diego) for their work in helping to prepare the strains used in this study. In addition, T. Hwa first suggested that we study EF-TU. We also acknowledge Soni Mohapatra and Heejun Choi for many fruitful scientific discussions and for suggestions for improving the analysis of the diffusion data.\n\nThis work was supported by the National Science Foundation (MCB-1512946 to J.C.W.) and by the National Institutes of Health (NIGMS, R01-GM094510 to J.C.W.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.\n"}
+{"text": "Background {#Sec1}\n==========\n\nMedicinal plants are considered as potential source for drug development and many novel products have reached clinical trials. Scientists are investigating properties of medicinal plants in order to develop novel drugs against disease like cancer, from natural products. Medicinal herbs have profound scope and have been used to find potential anticancer compounds in them \\[[@CR1]\\]. Use of phytochemicals from the medicinal plants in cancer treatment, may reduce adverse side effects and help to treat cancer and they have in recent years, shown promising anticancer efficiency against different cancers like human mouth epidermal carcinoma (KB cell line), murine leukemia (P388 cell line), human colorectal cancer (BE cell line) and prostate cancer (PC3 cell line) by various mechanisms like up-regulation of p16INK4A, preventing inhibition of MRCK-kinase targeting multiple gene products and targeting mitotic processes \\[[@CR2]--[@CR8]\\]. Many medicinal plants found in Nepal are also found to have cytotoxic effects against different cancer cell lines. For example, Cell viability assay of plant extracts from *Berberis aristata*; showed a significant cytotoxicity to MDA-MB-231 and U-87 MG human cancer cell line; Withanone from *Withania somnifera* (Ashwagandha) has been identified to have p53-activating tumor-inhibiting property. Ashwagandha leaf powder was non-toxic and anti-tumorigenic in mice assays and caused an abrogation of mortalin-p53 interactions and reactivation of p53 function and it is also established that the different extracts of plant *Juniperus recurva* possess anticancer properties against Breast Cancer (MCF7) cells \\[[@CR9]--[@CR12]\\].\n\n*Allium wallichii*(AW) {#Sec2}\n----------------------\n\n*Allium wallichii* is a monocotyledonous plant falling in order *Asparagales,* family *Amaryllidaceae and Allioideae* subfamily. It has been used as traditional medicine and spice in Nepal. Recent research papers deal with the genetic diversity and relationship of two species of *Allium* by amplified fragment length polymorphism analysis \\[[@CR13]\\]. The phylogenetic and biogeographic investigation of *Allium sp.* based on nuclear ribosomal internal transcribed spacer and chloroplast ribosomal protein S16 (rps16) sequences implied that genus *Allium* comprises more than 800 species, which defends its place among the largest genera in monocotyledons \\[[@CR14]\\]. Use of plant resource and traditional medicine in Nepal has also been documented for millennia and they are still the most significant health care source for bulk of population \\[[@CR15]\\]. There have been reports of trnK-gene-based molecular phylogenetic studies of plants of genus *Allium* \\[[@CR16]\\]. 1, 2 bis (methylthio) ethene, 2, 4 di-Methiophene, di-Methyl disulfide and di-Methyl trisulfide were established to be the most important volatile constituents of *Allium wallichii* \\[[@CR17]\\]. Diosgenin and Tigogenin are reported as steroidal sapogenins from bulbs of *Allium wallichii.* Kunth, that shows encouraging pharmacological potentials \\[[@CR18]\\]. Neuropsychopharmacological utilization has also been conferred in the studies of the action of *Allium sp.* which shed light to its diverse use \\[[@CR19]\\]. Extracts that includes *Allium sp*. is found to exhibit analgesic and fever relieving effects and can be used for the treatment of inflammation, ulcer, viral infection, cancer, eczema, diabetes, senile gangrene, herpes zoster and most importantly Acquired Immune Deficiency Syndrome (AIDS) \\[[@CR20]\\].\n\nIn this paper we have investigated the presence of different phytochemicals in *Allium wallichii* and also established the anti -microbial, anti-oxidant activity of the crude extract of the plant. More importantly this paper is focused in establishing the anticancer potential of the extracts of *Allium wallichii* by *in-vitro* cytotoxicity assay against Prostrate, Cervical and Breast cancer cell lines and Flow-cytometric analysis for viability assay against B-lymphoma cell lines. This is the first study of its kind for the plant *Allium wallichii.*\n\nMethods {#Sec3}\n=======\n\nCollection of plant material {#Sec4}\n----------------------------\n\n*Allium wallichii* (AW) whole plants were collected from Kathmandu University campus area of Kavrepalanchowk district, Nepal. The plant material was identified by Dr. Rajendra Gyawali, Botanist, Assistant Professor of Department of Pharmacy, Kathmandu University, Nepal. A voucher herbarium specimen is deposited at the herbarium of the Department of Pharmacy, Kathmandu University, Nepal\u00a0and the voucher specimen number is DoP-H-211. The plant samples were then dried in shade, left over for 15\u00a0days and macerated and powdered with the help of kitchen grinder.\n\nPreparation of plant extract {#Sec5}\n----------------------------\n\nThe dry powder of *Allium wallichii* (100 g) was dipped in 80% Aq. Ethanol for three days stirring twice a day. The solution at the end of third day was filtered and then concentrated in high vacuum-rotary evaporator and finally all the solvent was evaporated using freeze dryer to get the semi-solid plant extract. The same procedure was repeated three times and the freeze-dried extract was mixed together. The extract was then kept in glass vial with airtight caps and stored at 4\u00a0\u00b0C.\n\nPhytochemical screening {#Sec6}\n-----------------------\n\nThe phytochemical screening was done using the standard protocols \\[[@CR21]--[@CR28]\\]. Test for Alkaloids: 5\u00a0ml of extract was concentrated to yield a\u00a0residue. Residue was dissolved in\u00a03ml of 2% (v/v) HCl, few drops of Mayer's reagent was added. Appearance of the dull white precipitate indicated the presence of basic alkaloids. Test for Coumarin: 4\u00a0ml extract solution was taken; 1--2 drops of water (hot) was added. Volume was made half (UV fluorescence). 10% NH~4~OH was added to another half volume (strong fluorescence). Presence of green fluorescence indicated the presence of Coumarin. Test for Saponins: 2\u00a0ml extract was shaken vigorously for 30\u00a0s in a test tube. Persistence of thick forth even after 30\u00a0mins indicated the presence of saponins. Test for Glycosides: 2\u00a0ml of extract was dried till 1\u00a0ml.1-2\u00a0ml NH~4~OH was added and shaken. Appearance of cherish red color indicated the presence of glycosides. Test for Reducing Sugars: 0.5\u00a0ml of extract was taken and\u00a01ml distilled water was added. 5-8 drops of Fehling's solution (hot) was added. Presence of brick red precipitation indicated the presence of reducing sugar. Test for steroids: 1\u00a0ml extract was dissolved in 10\u00a0ml chloroform. Equal volume of conc. H~2~SO~4~ was added by the side of test tube. Upper layer turned red and sulphuric acid layer turned yellow with green fluorescence. This indicated the presence of steroids. Test for Quinone: 1\u00a0ml of extract was taken.1\u00a0ml of conc. H~2~SO~4~ was added. Formation of red color indicated the presence of quinone. Test for Terpenoids: 5\u00a0ml of extract was taken and mixed with 2\u00a0ml of chloroform. 3\u00a0ml of conc. H~2~SO~4~ was added to form a layer. Reddish brown precipitate formation at the interface formed indicated the presence of terpenoids. Test for Tannins: About 0.5\u00a0g of the dried powdered samples was boiled in 20\u00a0ml of water in a test tube and then filtered. Few drops of 0.1% Ferric Chloride was added and observed for brownish green or a blue-black coloration. Test for Flavonoids: A portion of the powdered plant sample was heated with 10\u00a0ml of Ethyl Acetate over a steam bath for 3\u00a0min. The mixture was filtered and 4\u00a0ml of the filtrate was shaken with 1\u00a0ml of dilute Ammonia solution. A yellow coloration was observed indicating a positive test for flavonoids.\n\nDPPH free radical scavenging assay {#Sec7}\n----------------------------------\n\nDPPH (2, 2-diphenyl-1-picrylhydrazyl radical), is a dark-colored crystalline powder composed of stable free-radical molecules. DPPH has major application in laboratory research most notably in anti-oxidant assays. The DPPH assay is typically run by the following standard procedure \\[[@CR29]--[@CR31]\\]. The hydrogen atom or electron donating abilities of the corresponding extracts/fractions and standards were measured from the bleaching of the purple-colored methanol solution of DPPH. 10\u00a0mg DPPH was dissolved in 100\u00a0ml methanol (MeOH) to obtain a concentration of 100\u00a0\u03bcg/ml. The stock solution of the fractions/extracts was prepared by dissolving 25\u00a0mg in 50\u00a0ml MeOH. Dilutions of the stock solutions of the crude extracts were prepared to obtain concentration of 1\u00a0\u03bcg/ml, 2\u00a0\u03bcg/ml, 3\u00a0\u03bcg/ml, 4\u00a0\u03bcg/ml, 5\u00a0\u03bcg/ml and 10\u00a0\u03bcg/ml while the control was prepared by dissolving 1\u00a0ml DPPH in 4\u00a0ml MeOH (without sample). 1\u00a0ml DPPH was added to each solution and the solution were kept in dark and allowed to stand for exact 30\u00a0mins. The UV absorption of each solution was recorded at 517\u00a0nm. The reaction was allowed to progress for 30\u00a0mins at 37\u00a0\u00b0C and absorbance was monitored by microplate reader, SpectraMax340 at 517\u00a0nm. Upon reduction, the color of the solution fades (violet to pale yellow). Percent Radical Scavenging Activity (% RSA) was determined by comparison with a DMSO containing control. The concentration that causes a decrease in the initial DPPH concentration by 50% is defined as IC~50~ value. Scavenging of free radicals by DPPH as percent radical scavenging activities was calculated as follows. % RSA\u2009=\u2009Control absorbance -- extract absorbance X 100/control absorbance. The IC~50~ values of compounds were calculated by using the EZ-Fit Enzyme kinetics software program (Perrella Scientific Inc. Amherst, MA, USA). Ascorbic acid was used as the reference compound.\n\nAgar plate diffusion method for anti-microbial activity {#Sec8}\n-------------------------------------------------------\n\nMedium was dissolved and autoclaved at 121\u00a0\u00b0C for 15\u00a0min, cooled up to 45\u00a0\u00b0C and then 40-50\u00a0ml media was poured in sterile 14\u00a0cm diameter Petri plate, and then allowed to solidify and kept at room temperature as stated in the protocols \\[[@CR32]--[@CR34]\\].\n\nPreparation of plant extract {#Sec9}\n----------------------------\n\nStock solution of 400\u00a0mg/ml was prepared by weighing 200\u00a0mg of plant extract in 1.5\u00a0ml Eppendorf tube and 0.5\u00a0ml of DMSO was added by micropipette. Extract was completely dissolved by vortexing for 5--10 min. Test solution of 200\u00a0mg/ml, 100\u00a0mg/ml, 50\u00a0mg/ml and 25\u00a0mg/ml concentrations was prepared.\n\nPreparation of inoculums {#Sec10}\n------------------------\n\nEach culture to be tested was streaked onto nutrient agar plate to obtain isolated colonies. Overnight incubation was done at 37\u00a0\u00b0C. Then isolated colonies were transferred by the help of sterile loop onto Muller Hinton Broth. Overnight incubation was done at rotary shaker at 37\u00a0\u00b0C.\n\nInoculation {#Sec11}\n-----------\n\nFor inoculation, swabbing was done with the help of cotton. Sterilized filter paper discs were dipped into the desirable concentration of the plant extracts and then applied to the plates; incubated at 37\u00a0\u00b0C for 24\u00a0h. After incubation, the diameter of the Zone of inhibition was measured.\n\nMTT/cytotoxicity assay against cancer cell lines {#Sec12}\n------------------------------------------------\n\nCytotoxic activity of plant extract was evaluated in 96-well flat-bottomed micro-titer plates by using the standard MTT (3-\\[4, 5-dimethylthiazole-2-yl\\]-2, 5-diphenyl-tetrazolium bromide) colorimetric assay \\[[@CR35]\\]. For this purpose, MCF-7, PC3, HeLa cells were separately cultured in Dulbecco's Modified Eagle Medium (Sigma Aldrich, Germany), supplemented with 5% of fetal bovine serum (FBS) (Sigma Aldrich, Germany), 100\u00a0IU/ml of Penicillin and 100\u00a0\u03bcg/ml of Streptomycin in 75\u00a0cm^2^ flasks and kept in 5% CO~2~ incubator at 37\u00a0\u00b0C. Exponentially growing cells were harvested, counted with hemocytometer and diluted with a particular medium. Cell culture with the concentration of 1x10^5^ cells/ml was prepared and introduced (100\u00a0\u03bcL/well) into 96-well plates. After overnight incubation, medium was removed and 200\u00a0\u03bcL of fresh medium was added with different concentrations of plant extracts. The stock solution of the extract was first prepared (50\u00a0\u03bcg/ml) in DMSO and was serially diluted up to the concentration of 0.78125\u00a0\u03bcg/ml. After 48\u00a0h, 200\u00a0\u03bcL MTT (0.5\u00a0mg/ml) was added to each well and incubated further for 4\u00a0h. Subsequently, 100\u00a0\u03bcL of DMSO was added to each well. The extent of MTT reduction to formazan within cells was calculated by measuring the absorbance at 570\u00a0nm, using a micro plate reader (Spectra Max plus, Molecular Devices, CA, USA). The cytotoxicity was recorded as concentration causing 50% growth inhibition (IC~50~) for the cell lines used. The percent inhibition was calculated by using the following formula: % Inhibition\u2009=\u2009100-\\[(mean of O.D of test compound -- mean of O.D of negative control)/(mean of O.D of positive control -- mean of O.D of negative control) X 100\\]. The results (% Inhibition) were processed by using Soft-Max Pro software (Molecular Device, USA). IC~50~ was calculated using EZ-Fit Software.\n\nCell viability analysis using PI-On flow cytometry {#Sec13}\n--------------------------------------------------\n\nB-lymphoma cancer cell line was a generous gift from Switzerland, Raji (Burkitt's lymphoma) (ATCC Catalog. Number: CCl-86). The cell\u00a0line was maintained at 37\u00a0\u00b0C in a humidified 5% CO~2~ environment in Roswell Park Memorial Institute (RPMI) 1640 (Caisson, Oklahoma) with 1%\u00a0L-Glutamine, 1% Penicillin/Streptomycin (Invitrogen, Germany), supplemented with 10% fetal bovine serum (FBS) (PAA laboratories, Austria). The cell viability analysis using PI-On Flow cytometry was performed at Panjwani Centre for Molecular Medicine and Drug Research, International Centre for Chemical and Biological Sciences, University of Karachi, Pakistan using standard protocol \\[[@CR36]\\]. Cell cytotoxicity assay was performed using florescent dye Propidium iodide (Excitation wavelength\u2009=\u2009536\u00a0nm) (Emission wavelength\u2009=\u2009617\u00a0nm). The dye has the property that it can't cross the intact plasma membrane and as a result of this, live cells cannot be stained using this dye. However, the dead cells in which cell membrane integrity is lost, the dye penetrate inside the cell and thus intercalate the DNA of the cell as a result of that only dead cells will fluoresce and could be read in FL2 or FL3 channels/filters in flow-cytometer. In order to generate dose response curve of standard and test compounds against the cell line, cells were seeded in round bottom 96 well plate in such a way that each well contain 1.3x10^5^ cells per well with 200\u00a0\u03bcl final reaction volume. Negative control wells were given media containing cells only while in positive control wells 0.5% of the total reaction volume of DMSO was added.\n\nResults {#Sec14}\n=======\n\nThe yield percentage of plant extract from the cold extraction was calculated. Yield percentage was calculated by using following formula: Yield\u2009=\u2009(Weight of the extract obtained)/(Total weight of the sample loaded) x 100%$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ Yield\\%=\\frac{12.07}{100}\\times 100\\%=12.07\\% $$\\end{document}$$\n\nIn the phytochemical screening, as shown in Table\u00a0[1](#Tab1){ref-type=\"table\"}, we observed the presence of different phytochemicals like steroids, terpenoids, flavonoids, reducing sugars and glycosides. Anti-microbial assay was performed with the plant extract of *Allium wallichii* by agar plate diffusion method. Five different antibiotics were used as standard drugs. Eight different microorganism *B. cereus, B. thuringiensis, E.coli, P. aeruginosa, P. mirabilis, Rhizopus, A. flavus and B. subtilis* were used for the assay.Table 1Results of phytochemical screening of *Allium wallichii*Phytochemicals*Allium wallichii*Alkaloids-Tannins-Flavonoids+Reducing Sugars+Coumarin-Glycosides+Quinone-Steroids+Terpenoids+Saponins-\n\nThe anti-microbial activity was assayed by measuring the Zone of inhibition (ZOI) of different extracts on the agar disc plate. The anti-microbial test of the *Allium wallichii* was done in different concentrations of the plant used so as to study the concentration dependent toxicity against the different bacteria and fungi. The different concentrations used such were 200, 100, 50 and 25\u00a0mg/ml. The standard antibiotics used were Gentamicin (GEN10), Ciprofloxacin (CF30), Chloramphenicol (C30), Cephotaxime (CTX30) and Tetracycline (TE30) as presented in Table [2](#Tab2){ref-type=\"table\"}. The highest ZOI among the five standard antibiotics used was of CTX 30\u00a0against *Rhizopus* as stated in the Table [2](#Tab2){ref-type=\"table\"}.Table 2Zone of Inhibition shown by standard antibioticsZone of Inhibition shown by standard antibiotics*B. cereusE. coliB. thuringiensisP. mirabilisRhizopusA. flavusP. aerugenosaB. subtilis*CTX3020259203602921GEN10272930302701523CF30343535353004027C3030301272502321TE3018302092902825\n\nThe plant showed concentration dependent anti-microbial activity towards bacteria used with the highest ZOI observed being 12\u00a0mm for *B.cereus* at the highest concentration used i.e. 200\u00a0mg/ml as shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}. The percentage scavenging activity was measured at different concentrations of the extracts 1\u00a0\u03bcg/ml, 2\u00a0\u03bcg/ml, 3\u00a0\u03bcg/ml, 4\u00a0\u03bcg/ml, 5\u00a0\u03bcg/ml and 10\u00a0\u03bcg/ml. The comparison showed that the percentage scavenging activity is concentration dependent but the line plot of *Allium wallichii* is observed below the line of Ascorbic acid, which corresponds to the moderate activity as compared to the standard Ascorbic acid. The percentage scavenging activity of *Allium wallichii (AW)* was found to be 7.68, 9.79, 12.48, 14.98, 20.62 and 28.66 as compared to 8.07, 18.45, 25.81, 35.46, 44.25 and 89.25 of Standard Ascorbic acid at concentration of 1\u00a0\u03bcg/ml, 2\u00a0\u03bcg/ml, 3\u00a0\u03bcg/ml, 4\u00a0\u03bcg/ml, 5\u00a0\u03bcg/ml 10\u00a0\u03bcg/ml respectively. The IC~50~ value was calculated from the graph of regression analysis and found to be 17.87\u00a0\u03bcg/ml as shown in Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}.Table 3Results of antimicrobial assay of *Allium wallichii* against different microorganismsPlantMicrobesExtract concentration (mg/ml) and Zone of inhibition (mm)200\u00a0mg/ml100\u00a0mg/ml50\u00a0mg/ml25\u00a0mg/ml*Allium wallichiiB. cereus*121199*B. thuringiensis*9987*E. coli*7.57.56.56.5*P. aeruginosa*8866*P. mirabilis*7.5776.5*Rhizopus*7.57.576.5*A. flavus*97.576.5*B.subtilis*7776.5 Fig. 1Comparison of Percentage Scavenging of DPPH by *Allium wallichii* at the concentration range (1-10\u00a0\u03bcg/ml) as compared to Ascorbic acid\n\nStudy of cytotoxicity of *Allium wallichii* against Prostate Cancer (PC3) Fig.\u00a0[2](#Fig2){ref-type=\"fig\"} and Cervical Cancer (HeLa) Fig.\u00a0[3](#Fig3){ref-type=\"fig\"} cell lines revealed that the plant studied was moderately cytotoxic to both the cancer cell lines. The cytotoxic effect of the plant was compared with the standard (Doxorubicin). *Allium wallichii* showed 59.48\u2009\u00b1\u20092.58%, 45.32\u2009\u00b1\u20091.82%, 28.53\u2009\u00b1\u20092.59%, 22.58\u2009\u00b1\u20094.91% and 10.70\u2009\u00b1\u20095.66 of percent cytotoxicity at 100\u00a0\u03bcg/ml, 50\u00a0\u03bcg/ml, 25\u00a0\u03bcg/ml, and 12.5\u00a0\u03bcg/ml and 6.25\u00a0\u03bcg/ml concentration of the extracts, respectively, against PC3 cell lines. The IC~50~ value of *Allium wallichii* was 69.69\u00a0\u03bcg/ml for PC3. Similarly, the percent cytotoxicity for HeLa cell lines was 72.45\u2009\u00b1\u20092.12%, 52.61\u2009\u00b1\u20091.66%, 37.88\u2009\u00b1\u20091.26%, 30.24\u2009\u00b1\u20091.33% and 16.92\u2009\u00b1\u20090.27% respectively at the above mentioned concentration with calculated IC~50~ value of 46.51\u00a0\u03bcg/ml as compared to Doxorubicin (0.85\u00a0\u03bcg/ml). All IC~50~ values of the extracts were calculated using EZ-Fit Software. From these findings, moderate cytotoxic activity of *Allium wallichii* was observed towards different cancer cell lines taken for study. We observed moderate cytotoxicity of *Allium wallichii* crude extracts against Breast Cancer (MCF-7) cell lines (Fig not shown) with the percentage inhibition of 45.22% at 50\u00a0\u03bcg/ml concentration with the calculated IC~50~ value of 55.29\u00a0\u03bcg/ml.Fig. 2Percentage Inhibition of *Allium wallichi* against Prostate Cancer (PC3) Cell lines. (*n*\u2009=\u20093, Values are Mean\u2009\u00b1\u2009SD) Fig. 3Percentage Inhibition of *Allium wallichi* against Cervical Cancer (HeLa) cell lines. (*n*\u2009=\u20093, Values are Mean\u2009\u00b1\u2009SD)\n\nFlow cytometry analysis results, as seen in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"} showed that, marker M1 is placed around the events that are negative to florescence label, and Marker M2 is placed to the right of M1 to designate positive events. In our work since Propidium Iodide dye (detected at FL2 and FL3 both) is used, which holds the capability to enter into cells with compromised cell membrane (dead cells) consequently dead cells will take up the dye and will show in high florescent region M2 (left part) and viable cells will be in M1 (right half of the histogram). The Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, shows the viability assay for the total events gated by gate R1 of control and extracts at different concentrations against Burkitt's lymphoma cell lines. As shown in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, for control (a) M1 corresponds to viable 75.45% and M2 correspond to dead 21.01%. For the *Allium wallichii* at 0.31\u00a0mg/ml concentration (b) shows 53% viable and 46.10% dead cells. With further increasing concentrations to 0.62\u00a0mg/ml (c) 85% viable and 13.37% dead cells were observed. Similarly, at 1.23\u00a0mg/ml (d) 60.70% viable and 38.73% dead cells and at 2.43\u00a0mg/ml (e) 32% viable and 66.98% dead cells were observed. Analyzing all the above data, the IC~50~ value for *Allium Wallichii* against Lymphoma cells used in the flow-cytometric analysis line is found to be 3.817\u2009\u00b1\u20091.99\u00a0mg/ml which demonstrate good cytotoxicity against Lymphoma cell lines.Fig. 4The Histogram plot of the viable and dead cell count from the flow cytometric analysis for Burkitt's lymphoma cell line. Fig (**a**) Control, Fig (**b**-**e**) *Allium wallichi* at different concentrations\n\nDiscussions {#Sec15}\n===========\n\nThe yield percentage of 12.07%, for the plant shows that it has various compounds in it that could have health benefits. Phytochemicals like steroids, terpenoids, flavonoids and glycosides have been shown to have anti-cancer activity. For the anti-microbial assay, the gradual decrease in concentration showed corresponding decrease in ZOI. The microorganism like *E. coli, P. mirabilis* and *Rhizopus* were found to be less susceptible toward the action of *Allium wallichii.* We can infer from the comparison between the data from Table [2](#Tab2){ref-type=\"table\"} and Table [3](#Tab3){ref-type=\"table\"} that, the studied plant show only moderate antimicrobial activity. This can be attributed to the fact that the whole plant including the tuber of *Allium wallichii* along with leaves has been used for the study; and there must be significant difference in the phytochemicals present.\n\nThe free radical species can start chain reactions and can cause damage or death to the cell. Antioxidants terminate these chain reactions by removing free radical intermediates, and inhibit other oxidation reactions. Antioxidants compounds scavenge free radicals such as peroxide, hydrogen peroxide or lipid peroxyl and thus inhibit the oxidative mechanisms that lead to degenerative diseases. Screening of plants for this is done by measuring the antioxidant activity by various in vitro activities such as DPPH method, Nitric Oxide method and in vivo models using rats and mice. Free radicals that are commonly used to assess antioxidant activity in vitro is 2, 2-diphenyl-1-picrylhydrazyl (DPPH). So higher the percentage scavenging activity, higher the antioxidant activity and higher the anticancer property it may possess. The results in Fig. [1](#Fig1){ref-type=\"fig\"} show that, the plant have moderate anti-oxidant activity.\n\nThe extract of the plant showed moderate anti-cancer activity, when studied with cancer cell lines namely PC3, HeLa and MCF-7. The calculated IC~50~ of the plant shows that the plant can have anti-cancer activity as has been reported for other plants viz. *Caralluma adscendens (Roxb.)*, *Daedalea gibbosa, Alpinia conchigera, Chelidonium majus*, *Ocimum sanctum Linn, Potentilla erecta*, *Chamaenerium angustfolium*, *Paphia undulata, Filipendula ulmaria*, *Inula helenium* against different cell lines \\[[@CR36]--[@CR42]\\]. Analyzing the data of FACS, the IC~50~ value for *Allium Wallichii* against the Lymphoma cell used demonstrate good cytotoxicity. The moderate cytotoxicity in the studies might have been due to use of whole plant for the assay. The separation of pure compounds with bioassay guided extraction and the spectrometric analysis and subsequent cytotoxicity check would reveal the bioactive components and give a better activity. Thus, most of the research these days are directed towards separation of bioactive components.\n\nConclusions {#Sec16}\n===========\n\nFrom the experiments performed, we concluded the presence of different phytochemicals like steroids, terpenoids, flavonoids, reducing sugars and glycosides in the plant *Allium wallichii*. In the DPPH-RSA assay *Allium wallichii* showed moderate antioxidant activity; while anti-microbial assay also showed moderate activity against known pathogens. Further observations demonstrated that the plant *Allium wallichii* exhibits moderate cytotoxic activity towards Prostate cancer (PC3), cervical cancer (HeLa) and Breast Cancer (MCF-7) cell lines. Finally, the flow-cytometric analysis demonstrated a good cytotoxicity of *Allium wallichii* against Lymphoma cell lines with significant cell death in the population of the Lymphoma cell line used. These results accounts for the fact that, *Allium wallichii,* together with moderate anti-oxidant and cytotoxic activity and presence of important phytochemicals like flavonoids, steroids, glycosides and terpenoids, can be a very important candidate to be used as an anticancer agent. Moreover, use of plant parts separately and the isolation of pure compounds from the plant would certainly have better activity against cancer.\n\nC30\n\n:   Chloramphenicol\n\nCF30\n\n:   Ciprofloxacin\n\nCTX30\n\n:   Cephotaxime\n\nGEN10\n\n:   Gentamicin\n\nTE30\n\n:   Tetracycline\n\nATCC\n\n:   American type culture collection\n\nDMSO\n\n:   Dimethyl sulfoxide\n\nDPPH\n\n:   2,2-diphenyl-1-picrylhydrazyl\n\nFACS\n\n:   Fluorescence-activated cell sorting\n\nFBS\n\n:   Fetal bovine serum\n\nFL2 or FL3\n\n:   Channels/filters in flow cytometer\n\nFSC-H\n\n:   Forward scatter in flow cytometry; H refers to height\n\nHeLa\n\n:   Cervical cancer cell line\n\nIC~50~\n\n:   Half maximal inhibitory concentration\n\nMCF-7\n\n:   Human breast cancer cell line\n\nMDA-MB-231\n\n:   Human mammary cancer cell line\n\nMTT-\n\n:   3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide\n\nOD\n\n:   Optical density\n\np53\n\n:   Tumor suppressor gene p53\n\nPC3\n\n:   Human prostate cancer cell line\n\nRPMI\n\n:   Roswell park memorial institute\n\nSSC-H\n\n:   Sidewise scatter in flow cytometry; H refers to height\n\nU-87 MG\n\n:   Human primary glioblastoma cell line\n\nZOI\n\n:   Zone of inhibition\n\nWe would like to acknowledge Department of Biotechnology, Kathmandu University Nepal for the support and Panjwani Centre for Molecular Medicine and Drug Research, International Centre for Chemical and Biological Sciences, University of Karachi, Pakistan for the facilitation of Flow cytometric assay of the plant extracts. We would also like to acknowledge Dr. Achyut Adhikari for all his suggestions and providing of lab space at HEJ Laboratories for chemistry, Karachi University, Pakistan.\n\nFunding {#FPar1}\n=======\n\nThe research was funded by grant from International Foundation for Science, Sweden, grant number F/5018-1.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nPlant material deposited in herbarium of Department of Pharmacy, Kathmandu University.\n\nAuthors'contributions {#FPar3}\n=====================\n\nJB as the first author of the paper have carried out the phytochemical screening parts and the investigation of the cytotoxicity of the plant extracts against three different cancer cell lines. JB has also been inclusively involved in all the experimental design of the research; along with the literature review, prior and post of the research and the manuscript writings. BT has contributed in carrying out the Flow cells benchtop flow cytometry analysis for the viability assay and finding the IC~50~ values of the extracts against the Lymphoma cell line. PT has been involved in sample collection, extraction, phytochemical screening, anti-oxidant and anti-microbial screening. BGS as MS by Research supervisor of JB, has been involved in research design, supervision and protocol validation and optimization works. All authors have read and approved the final manuscript.\n\nCompeting interests {#FPar4}\n===================\n\nWe hereby declare that we do not have any financial or non-financial competing interests and would like to ensure, the publication of our manuscript would not, in any case, disregard any of person or organization or hamper their reputations by any means.\n\nConsent for publication {#FPar5}\n=======================\n\nVerbal consent taken from the researchers involved in the project for publication.\n\nEthics approval and consent to participate {#FPar6}\n==========================================\n\nNot applicable.\n\nAuthors' information {#FPar7}\n====================\n\nJB, MS by Research student in Cancer Biology, at Department of Biotechnology, Kathmandu University. JB has authored/co-authored few publications in field of Biotechnology. BTM, is a PhD student in Panjawani Center for Molecular Medicine and Drug Research, University of Karachi, Pakistan. He has authored/co-authored many publications in field of Biotechnology. PT is Master student at Department of Biotechnology, Kathmandu University and has research publications in similar area. BGS, PhD, is an Assistant Professor at Department of Biotechnology, Kathmandu University, Nepal and has been doing research in medicinal plants and cancer for over twelve year. He has authored/co-authored numerous publications in the field.\n"}
+{"text": "\"Neither microbial genes nor microbial names cause any harm to the body, but microbial product may do. The presence of very active microbial biological compounds in the gut may have physiological and pathophysiological consequences for the host. Tore Midtvedt (2008)\"\n\nAccording to modern scientific doctrine, the human being is a 'superorganism', the consortium of numerous *Eukarya, Bacteria, Archaea*, and *Viruses*. Various host indigenous microorganisms should be considered as essential complex extracorporal physiological systems that play a fundamental role in human health and disease. Host microbiota and other functional and metabolic systems, connected with host eukaryotic cells, are working together profitably for the whole organism and for its separated components in the concrete environment conditions. Unfortunately various unfavorable biotic or abiotic factors and stress agents (diet, age, sex, pharmaceutical, even surgical interventions, etc.) can produce microecological disorders resulting in tissue, organ and regulatory systems disturbances that can lead to an increase of a disease development risk ([@CIT0001]--[@CIT0005]). It means the modulation of microbe/microbe and microbe/host interactions is an extremely important, fundamental, and applied problem in modern biology and medicine. To date, to maintain and restore the gut microbial community, three therapeutic approaches have dominated: probiotics, prebiotics, and synbiotics ([@CIT0004], [@CIT0006]--[@CIT0010]). As defined by FAO/WHO, probiotics are non-pathogenic live microorganisms which, when administered in adequate quantities, offer a health benefit to the host ([@CIT0006]). Many therapeutics, dietary foods, and additives containing live probiotic microorganisms have been introduced in practice for human health support, acute, chronic, localized, and systemic disease prophylaxis and treatment ranging from acute or chronic diarrhea, intestinal inflammation pathology to allergy, atherosclerosis, and cancer ([@CIT0005], [@CIT0011]). In the last decade, a transplant of the distal gut microbiota (fecal bacteriotherapy) became popular in the treatment of some inflammation intestinal diseases ([@CIT0012], [@CIT0013]). The introduction of genetically engineered probiotics on the basis of recombinant live microorganisms in medical practice has also been actively discussed ([@CIT0004]).\n\nUndesired properties and adverse affects of live probiotic microorganisms {#S0001}\n=========================================================================\n\nMore than 50 years use of probiotics has shown them to be safe and beneficial, but up until now, we have not been able to define the optimal amount of bacteria for probiotic effects; there is no single mechanism of action for all probiotics. Moreover, the beneficial effects of probiotics may have short-term success, be absent or uncertain ([@CIT0003], [@CIT0007], [@CIT0014]). The latter may be explained by the low concentration of probiotic biologically active substances (bioactives) achieved in target places during the traditional application of live probiotic microorganisms. The number of these microbial bioactives produced might be inadequate to receive desirable specific effects under *in vivo* conditions ([@CIT0004], [@CIT0007]). Besides, various molecules produced in volume by live probiotic cells may interact with different receptors of indigenous microbes and host cells and simultaneously cause both beneficial and negative effects ([@CIT0015]). Modern 'omic'-technologies have revealed the substantial diversity of the gut microbiome between individuals. Only a few microbial phylotypes (species) are shared between different individuals; about 80% of human intestinal microorganisms in the adult gut are individual on the strain level ([@CIT0007], [@CIT0010], [@CIT0016]--[@CIT0019]). The experimental and clinical studies published in recent years have demonstrated that it is very difficult if not impossible to produce industrially adequate probiotics for supporting indigenous microflora at the optimal level by the simple mechanical selection of the separate or set of probiotic microorganism strains. Additional problems in the design of effective probiotics are connected with limited knowledge about the affect mechanisms of such bio-therapeutics to the host gut microbiota, physiology, and metabolism. However, little is known about the molecular mechanisms of probiotic effects ([@CIT0010], [@CIT0015], [@CIT0020]).\n\nIn recent years, our knowledge regarding the safety of probiotics has changed. It is necessary to bear in mind that any detrimental and harmful consequences may only become apparent after extended periods of probiotic use. In reality, some data now show that not all probiotic bacteria are safe even if they belong to *Lactobacillus* or *Bifidobacterium* species having no traditional genes of pathogenicity ([@CIT0021], [@CIT0022]). Medical reports have advised that lactic acid bacteria and even bifidobacteria used as microbial food cultures or probiotics may be rarely associated with human opportunistic infections (infective endocarditis, sepsis, bacteraemia, pneumonia, abdominal abscesses, peritonitis, meningitis, urological infections, rheumatic vascular diseases) especially in patients receiving antibiotic treatment or those severely immune compromised ([@CIT0021]--[@CIT0028]). There are increased incidences where these bacteria can be responsible for allergic sensitization and autoimmune disorders ([@CIT0023], [@CIT0029]--[@CIT0032]). Symbiotic microorganisms (including probiotic strains) sometimes can increase platelet aggregation aggravating hemolytic uremic syndrome ([@CIT0021], [@CIT0026]); some of them may be a source of toxic metabolites (e.g. biogenic amines) ([@CIT0022]). Probiotics found on living microbes being introduced into gastro-intestinal or vaginal tracts may cause unintended harm by gaining a competitive advantage and causing an ecological imbalance as a result of altering the microbe biodiversity and metabolic pathways. Because the vast majority of probiotic strains introduced in practice have been selected on the basis of their strong antagonistic activities against disease causative microorganisms, it seemed that many probiotics could suppress the growth and development of human gut and vaginal lactobacilli and other different indigenous microbiota also ([@CIT0033], [@CIT0034]). They may also alter intestinal metabolism due to their microbial enzymatic activities ([@CIT0035]). It is necessary also to remember that some intestinal microbial strains can participate in the transformation of some drugs modifying their activity and/or convert the prodrug to the active product ([@CIT0036]--[@CIT0038]). Unfortunately, there is practically nothing known about interactions of live probiotic microorganisms with drug function *in vitro* and *in vivo*. The situation become more complicated when new strains belonging to *Enterococcus, Streptococcus, Escherichia, Bacillus, Bacteroides*, or other microbial genera are suggested as a potential probiotic including probiotics that consist of multistrains or are constructed on the base of genetically modified microorganisms. There are data that probiotics (*Lactobacillus GG*) inside the intestinal tract can induce the expression of overran additional 400 new genes involved in immune response and inflammation, cell growth and differentiation, apoptosis, cell-to-cell signaling, and cell adhesion resulting in a wider impact on the host\\'s gene expression than thought before the era of macroarray technologies ([@CIT0039]). Oral introduction of *E. faecalis* modulates activities of 42 genes in the gut epithelial cells; these genes are involved in the regulation of the cell cycle, cell death, and signaling ([@CIT0040]). It has been also shown that during passages through the intestinal tract, some silence genes of probiotic bacteria may also be induced by host cell signals; newly formed bacterial products are poorly characterized both chemically and functionally ([@CIT0039], [@CIT0041], [@CIT0042]). For example, the induction of 72 probiotic *L. plantarum* WSFS1 genes was demonstrated in these conditions: nine genes were responsible for sugar transport and the production of different enzymes involved in their fermentation; nine genes for the synthesis of amino acids, nucleotides, cofactors, and vitamins; four genes for out cell proteins controlling resistance to specific host factors; and 46 genes for the production of non-identified proteins ([@CIT0042]). It means that the probiotic cells inside the gastrointestinal tract may be involved in the interaction with various host-specific networks (e.g. participation in the degradation and production of different nutrients by different metabolic pathways in the small intestine and colon ([@CIT0018], [@CIT0043]), or in the modification of immune response, cell differentiation, cell signaling, adhesion ([@CIT0018], [@CIT0044]); cell proliferation, tissue development, water and ion metabolism, balance of Th1/Th2 cells, etc. ([@CIT0015], [@CIT0045]). Thus, it can be expected that the expression of known or silence microbial and eukaryotic genes may lead to undesirable effects on human health. In recent years, evidence appeared that between lactic acid bacteria and enteric bacteria as well as between different lactic acid bacteria, horizontal gene transfer may take place. Natural gene transfer usually occurs via the uptake of naked DNA (transformation), viruses (transduction), or plasmids (conjugation). It is well known that the spread of antibiotic resistance is a major global health issue ([@CIT0046]). Probiotic bacteria can possess acquired antibiotic resistance genes associated with mobile genetic elements (plasmids, transposons) that permit these organisms to transfer this genetic information to strains of the same species, different species, or even different genera including both commensals and pathogens. Such recombination events might not only result in the distribution of undesired genes among intestinal microorganisms but can produce rearrangements in microbial genomes and change gene expression patterns in recipient bacteria. The gene transfer and recombination events associated with probiotics may produce long-term environmental and health negative consequences ([@CIT0021], [@CIT0022], [@CIT0026], [@CIT0046]), including chromosome rearrangement and death of recipient cells ([@CIT0047]), alteration of the genomic and epigenomic regulation of gene expression, post-translation modification of gene products, host/microbial cross talk result in the change of human metabolic and behavior reactions ([@CIT0020], [@CIT0048]). Recently, it has been shown (on the *E. coli* model) that the DNA methyltransferases represent potential threats to the epigenomic integrity of cell genomes. When the methylation system enters the cell and begins to methylate the host genome, the methylated DNA-specific microbial DNAse senses the epigenetic changes, causing cell death via chromosomal cleavage ([@CIT0049]). Traditional virulence traits should not be present in microorganisms used in food fermentation and probiotics ([@CIT0021], [@CIT0022]). Investigations made during the last decade have shown that some human symbiotic microorganisms can produce substances possessing genotoxic affects in intestinal epithelial cell DNA. Therefore, some commensals (including probiotic strain *E. coli* Nissle 1917) possess a set of genes (*pks* island) that are responsible for the production of double-strand breaks in host cell DNA. Bacteria containing the *pks* genes induce in eukaryotic cells a process called megalocytosis, in which the cell body and nucleus become enlarged and mitosis stops. Sometimes 4 hour is enough for the *pks* island carrying bacteria with eukaryotic cells resulting in the increase of DNA double-strand breaks level. A total of 34% of *E. coli* strains isolated from healthy human intestine content had such *pks* islands. A small number of bacterial cells caused minimal DNA damage in eukaryotic cells; in contrast, exposure to 100 or more bacteria per cell has broken the most nuclear DNA. Thus, the breaks in host cell DNA caused by peptide-polyketide genotoxins of some indigenous (probiotic) bacteria could trigger the various cells disorders, including intestinal cancer development ([@CIT0050], [@CIT0051]). Live cells of *Enterococcus faecalis* being inside the intestinal tract release substantial extracellular superoxide, hydroxyl radical, and H~2~O~2~ during carbohydrate fermentation and via autoxidation of membrane demethylmenaquinone. These oxidants can damage DNA and facilitate development of sporadic adenomatous polyps and colorectal cancer ([@CIT0040], [@CIT0052]). Ethanol and its first metabolite acetaldehyde have been recently classified as a class I carcinogen. The acetaldehyde concentration required for a mutagenic DNA effect increases from 100 to 500 \u00b5M. Many microbes (including lactic acid bacteria) used as microbial food cultures in fermented food products and in probiotic manufacturing can convert ethanol and/or glucose to carcinogenic acetaldehyde. The acetaldehyde levels formed may exceed the above-mentioned levels markedly. Furthermore, because living probiotic bacteria can colonize the intestinal tract for a long period of time and produce carcinogenic acetaldehyde locally, potentially they can be more dangerous than traditional dairy lactic acid bacteria because of increased total exposure to this bacterial metabolite ([@CIT0053], [@CIT0054]). Some scientists consider that 'genetic engineering of the human microbiota will enable to endow its members with new desirable functions that treat diseases or promote health' ([@CIT0010]). In spite of the scientific attractiveness of this idea, from a prolonged point of view, the massed introduction of genetic engineering live probiotic microbes in practice may have extremely dangerous ecological and medical consequences even when using such novel platform of microbial cells design as synthetic biology approaches. Thus, the above-mentioned data and discussion allow everybody to come to the major conclusion that the present state of probiotic knowledge is inadequate for reliable ecological and clinical risk assessment of short and long-term negative consequences; they often could be unpredictable. Up to now, we have no sufficient scientific knowledge to support manipulation of the human microecological, immune, and metabolic systems in an exactly predictable manner by the administration of live probiotics (including gene engineering) especially to infants and young children ([@CIT0026], [@CIT0035]). The public should have the right to participate in discussion concerning known and potentially undesired properties for probiotics made on the base of living microorganisms and have imagination regarding the unpredictable consequences of their long-term use.\n\nMetabiotic concept {#S0002}\n==================\n\nAlthough the history of live probiotic use does not give any exact identified serious concern, recent well-documented events of adverse effects and uncertainty about the level of their risk require new alternative approaches in prophylaxis and treatment of pathological conditions associated with the imbalance of host microbiota. These approaches have to retain and improve the positive accumulated experience of work with live commensal microorganisms (probiotics) and decrease a safety concern.\n\nInvestigations of the last 10--20 years have demonstrated that gut microorganisms (including probiotic strains) are able to break down and metabolize complex food nutrients and endogenous substances (saliva, gastro-intestinal juices compounds, epithelial cells, dead microbial cells, etc.) resulting in the formation of low molecular weight (LMW) bioactives that may be localized both inside and/or out of microbial cells and found in the intestinal content or passing across the intestinal epithelium barrier determined in the various human fluids, organs and tissues. These compounds derived from probiotic (symbiotic) microbes form what has been called the probiotic metabolome. Interacting with corresponding prokaryotic and eukaryotic cell targets, these biological and pharmacological active compounds may control many genetic, epigenetic, and physiological functions; biochemical and behavior reactions; and intra- and intercell exchange of information. Some commensal microbes including probiotics can secrete a variety of signaling molecules that can modify the inter-bacterial signaling (quorum quenching) and suppress the expression of virulence genes in pathogens or stimulate the growth of beneficial indigenous gut microorganisms ([@CIT0015], [@CIT0018], [@CIT0020], [@CIT0048], [@CIT0055]--[@CIT0062]). In our opinion, the probiotics commercially available now should be considered as a first generation of means directed for correction of microecological disorders. Future development of traditional probiotics will include improvements of this generation by means of the production of natural metabiotics (manufactured on the base of current probiotic strains) and synthetic (or semi-synthetic) metabiotics that will be analogies or improved copies of natural bioactives produced symbiotic microorganisms. The terms 'metabiotics' ([@CIT0004], [@CIT0020], [@CIT0063], [@CIT0064]), 'metabolic probiotics' ([@CIT0065], [@CIT0066]), 'postbiotics' ([@CIT0001]), 'biological drugs' ([@CIT0010]), or 'pharmacobiotics' ([@CIT0015]) mean small molecules that are the structural components of probiotic (symbiotic) microorganisms and/or their metabolites and/or signaling molecules with determined (known) chemical structure that can affect the microbiome and/or human metabolic and signaling pathways optimizing the indigenous microbiota composition and function and host-specific physiology, immunity, neuro-hormonbiology, regulator, and metabolic and/or behavior reactions connected with the activity of host indigenous microbiota. Different probiotic strains can become the source of hundred (thousands) of LMW bioactives (bacteriocins and other antimicrobial molecules, short chain fatty acids, various other fatty and organic acids, biosurfactants, polysaccharides, peptidoglycans, teichoic acids, lipo- and glycoproteins, vitamins, antioxidants, nucleic acids, different proteins including enzymes and lectines, peptides with various activities, amino acids, growth and coagulation factors, defensin-like molecules or their inductors in human cells, messenger (signal) molecules, plasmalogens, various co-factors, and so on) ([@CIT0002], [@CIT0007], [@CIT0010], [@CIT0015], [@CIT0018], [@CIT0020], [@CIT0048], [@CIT0055], [@CIT0056], [@CIT0067]). Various symbiotic (probiotic) strains can produce different sets of such LMW bioactive molecules, attractive candidates for metabiotics construction ([Table 1](#T0001){ref-type=\"table\"}). It should remember that the spectrum and number of bioactives of microbial origin determined in the human different biological fluids and eukaryotic cells may be also connected with the activity of the host transmembrane transporters and/or liver and other tissue enzymes that can carry or transform various microbial substances. Metabolomics analysis of plasma extracts from germ-free and conventional mice ([@CIT0068]) as well as other hosts and gut-microbial co-metabolomes revealed a significant interplay between gut microflora and mammalian metabolism ([@CIT0037]).\n\n###### \n\nSome groups of LMW compounds of symbiotic (probiotic) microbe origin that may become the basis for manufacture of potential metabiotics\n\n  ---------------------------------------------------------------\n  Bacteriocins\n  Short-chain fatty acids, other organic acids\n  Proteins, peptides, amino acids\n  Nucleic acids, nucleotides\n  Polysaccharides, peptidoglycans, other cell surface molecules\n  Plasmalogens, vitamins, antioxidants, co-factors\n  Various messenger (signal) molecules\n  ---------------------------------------------------------------\n\nSpecific representatives of these groups of LMW compounds isolated from symbiotic (probiotic) microorganisms or their cultural liquids may be used for manufacturing microbe-free food supplements, functional foods, and drugs for prophylaxis and treatment of chronic human diseases, as well as sport and anti-aging foods, etc. The above -mentioned approaches and tools for the design of new types of bio-actives on the base of LMW molecules of symbiotic microbiota origin for nutrition and medicine are already being developed in some countries ([@CIT0060]).\n\nCurrent probiotic strains in such situations may become starter strains for industry manufacturing of such microbial bioactive molecules. The knowledge of quality and quantity of LMW molecule profile of each industry used or potential probiotic strain will help researchers to design novel metabiotics with increased health effectiveness, and determine the optimal frequency, dose, and mode of their administration. Industrial production of metabiotics could become the novel prophylaxis and therapeutic approach to address human health in the near future because of their potential ability to interfere in the processes associated with stability of host genome and microbiome; modulation of epigenomic regulation of gene phenotypic expression in eukaryotic cells; improvement of the information exchange, signaling and metabolic pathways in numerous bacteria, bacteria--host and host systems that play an important role in the control for many genetic, epigenetic, physiological functions, biochemical and behavior reactions; in cell growth and host development; in supporting host health in general. We should always remember that real health effectiveness of suggested metabiotics depends on our knowledge of the physicochemical characteristics of metabiotic molecules (molecule isomerism: [l]{.smallcaps} or [d]{.smallcaps}, and \u03b1, \u03b2 or \u03b3 forms; valency and isotope state of incoming chemical elements; substance solubility; dispersiveness, ligand binding, oxido-reduction potential of metabiotic ingredient(s), its(their) interaction with other components that can be both as enhancers or inhibitors; competitive inhibitors for specific transport proteins or absorption site) as well as host physiological status, illness, constituents of the diet, medication, and so on.\n\nThe introduction of the concept 'Metabiotic' in practice permits including in biotechnology not only bifidobacteria, lactobacilli, *Escherichia*, enterococci strains but also tens and hundreds of other strains belonging to human dominant intestinal phyla (*Bacteroides, Firmicutes, Proteobacteria, Actinobacteria*, and *Archae*) for nutrition and medical aims. There is no doubt that the future metabiotics created on the base of strains belonging to the widen spectrum dominant microorganisms may be more effective at conferring health benefits than the LMW bioactives received only on the basis of classical probiotic strains (bifidobacteria and lactobacilli). The known and potential microbial starter cultures of human microbiota origin that could be used as a platform for metabiotic production should firstly be investigated on the synthesis of structural and excreted bioactive substances participating or regulating in pathways associated with carbohydrate, fat, and cholesterol metabolism; work of immune, hormone, and nervous systems; metagenome and epigenome stability; and gene expression regulation. Undoubtedly among human gut microbiota there are species and strains which are potential sources of small diffusible antimicrobial agents, regulators of host energy balance, cholesterol-lowering compounds, modifiers of immune reactions, stimulants, antidepressants, cognitive enhancers, etc. ([@CIT0010], [@CIT0060], [@CIT0063]). Development of databases of microbe-derived individual compounds or groups of compounds provides key information, which could be used toward goal-directed design metabiotics with specific medical effects. It is also important to remember the knowledge that widen rosters of LMW microbial metabolites and signal molecules produced by microorganisms and determined in the human physiological fluids can have the great diagnostic value ([@CIT0002], [@CIT0037], [@CIT0069], [@CIT0070]). Because *in vivo* production of bioactive small molecules of human and microbial origin is often connected with prebiotic-stimulated secondary metabolism ([@CIT0008], [@CIT0010]), there is a sense to design combined metabiotic/prebiotic foods and drugs for nutritional and medical purpose and prescription targeted to host microbiota or to indigenous microflora associated host functions, metabolic and behavior reactions.\n\nKnown and potential metabiotics {#S0003}\n===============================\n\nSome metabiotics on the base of natural (or artificial) bioactives of microbial origin in the last decade have already been introduced into medical practice and have proven their effectiveness in reducing infectious and metabolic diseases. For example, nowadays on the international market of drugs there is the most known metabiotic with trade mark '*Hylak Forte*' (Ratiopharm/Merckle, Germany. Code EAN: 4030096245166; N P No1497/01, 2009-05-14). This bacteria-free liquid medicine for oral use contains metabolic products of *Escherichia coli* DSM 4087 (25 g), *Streptococcus faecalis DSM 4086* (12.5 g), *Lactobacillus acidophilus DSM 4149* (12.5 g), and *L. helveticus DSM 4183* (50 g). The benefits of this sterile liquid drug are explained by the presence of SCFA, lactic acid, and some other non-identified microbial metabolites in this drug. There are data that '*Hylak Forte*' promotes the health of adults and children thanks to producing positive shifts in intestine microbiota, host acid-alkaline, water--salt metabolism, vitamins B and K balance, energy provision to intestinal epithelia and local immune cells ([@CIT0071]). Among other commercial metabiotics, it may also be named 'Zakofalk' (a combination of sodium butyrate and inulin) recommended for the treatment of mild to moderately active inflammation intestinal diseases ([@CIT0072]); *E. coli* glycoprotein with anorexic activity ([@CIT0073]); *Lactobacillus casei* (LEx) polysaccharide -- glycopeptide with antihypertension effect ([@CIT0074], [@CIT0075]); *L. helveticus* three-peptides (Val-Pro-Pro and Ile-Pro-Pro) inhibiting angeotensin convert enzyme (ACE) and resulting in the decrease of blood pressure ([@CIT0076], [@CIT0077]); lyophilized *B. subtilis* bacteria-free cultural fluid containing lysozyme, catalase, polypeptides, peptidoglycan, amino acids, and natural sorbent ceolite with immune-modulator and gut microbiota-restoring activity (\"bactistatin\") ([@CIT0078]; [www.stada.ru/products/baktistatin.html](http://www.stada.ru/products/baktistatin.html)); lactic acid bacteria SCFA and other organic acids ('Solcarmon', 'Frodo') ([@CIT0079]); Quorum sensing *Escherichia*, lactobacilli and bifidobacteria autoregulators of growth and development of indigenous intestinal microflora ([@CIT0065], [@CIT0066]). Potential metabiotics on the basis of lactobacilli, bifidobacteria, enterococci proteins, peptides, adhesions ([@CIT0055], [@CIT0056]), biosurfactants (more than 30kD) ([@CIT0080]), lectines ([@CIT0081], [@CIT0082]), nucleic acids and various cell wall molecules ([@CIT0015]) connected with microbial cells or produced in cultural liquid possessing various host effects are now under development. Special attention to my mind it should be paying to microbial membrane sterol-like compounds (e.g. plasmalogens) ([@CIT0059]) and outer membrane lipoproteins (e.g. lipocalins) ([@CIT0083], [@CIT0084]) as attractive candidate molecules for the manufacture of different types of metabiotics that are capable of inhibiting lipid and energy metabolism, immune, hormone and nervous system functions.\n\nMetabiotics, as modifiers of physiological functions, biochemical and behavior reactions, have some advantages. They have exact chemical structure, well dosed, very safe and long shelf-life. Besides, metabiotics possess the better absorption, metabolism, distribution, and excretion abilities compared with classic probiotics on the base of live microorganisms. A detailed molecular understanding can permit metabiotics to become significant specific and active contributors to the benefits derived from probiotics ([@CIT0004], [@CIT0010], [@CIT0015], [@CIT0020]). If the producer tries to label the metabiotic product with the specific therapeutic claims (prevent or treat a specific disease or improve the specific function or metabolic reaction), such metabiotics must be regulated as a drug needing approval by separate Regulatory Acts of corresponding National and International Agencies controlling Quality and Safety of Foods and Drugs.\n\nConclusion {#S0004}\n==========\n\nI now consider metabiotics not to be a myth; they are a natural evolution of the probiotic conception. In the near future on the base of probiotic LMW bioactives with proven specific benefit effect(s), a set of semi- and/or completely synthetic metabiotics will be designed that are analogies or improved copies of the natural microbial bioactives. This is a similar route of development to the traditional antibiotics that have been coming out during the last 50 years. LMW bioactives of indigenous microbiota origin being even chemically similar with the molecules of other origin (e.g. from food raw materials or food-stuffs) have advantages in principle. For many millions years of evolution, a human superorganism has been selecting prokaryotic and eukaryotic microorganisms from the environment that were functionally and metabolically the most optimal for human life and development. It means that metabiotics manufactured on the base of LMW bioactives of indigenous microbiota origin are the most beneficial and the safest.\n\nThe metabiotics of next generations will promote the further development of probiotic concept, improve effectiveness and benefit specificity of classic probiotics, and reduce environmental and health hazards of the current microecological approaches in the prevention and treatment of diseases associated with the imbalance of host microbiota.\n\nConflict of interest and funding {#S0005}\n================================\n\nThe author has declared that he has no interest which might be perceived as posing a conflict or bias.\n\n[^1]: The main areas of this paper were presented during the XXXVth Congress of the SOMED (15--17 May, 2012; Valencia, Spain).\n"}
+{"text": "Correction to: Surgery Today (2020) 50:21--29 <https://doi.org/10.1007/s00595-019-01881-y> {#Sec1}\n==========================================================================================\n\n\u00a0The article Portal flow modulation in living donor liver transplantation: review with a focus on splenectomy, written by Tomoharu Yoshizumi and Masaki Mori, was originally published Online First without Open Access. After publication in volume 50, issue 1, page 21--29 the author decided to opt for Open Choice and to make the article an Open Access publication. Therefore, the copyright of the article has been changed to \u00a9 The Author(s) 2020 and the article is forthwith distributed under the terms of the Creative Commons Attribution 4.0 International License (<https://creativecommons.org/licenses/by/4.0/>), which permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.\n\nThe original article has been corrected.\n\n**Publisher\\'s Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Comprising only 3% of all cancer cases in women in the United States, ovarian cancer is nonetheless the fifth leading cause of cancer-specific mortality.^[@bib1]^ Approximately 90% of ovarian cancers are epithelial cancers derived from ovarian surface or fallopian tube epithelium.^[@bib2]^ Serous ovarian carcinoma is the most common histologic subtype, with high-grade serous ovarian cancer (HGSOC) the most aggressive subtype, constituting 90% of these cases.^[@bib3]^ Owing to its predominance and lethal nature, HGSOC is the most widely researched type of ovarian cancer.^[@bib3]^\n\nTypical treatment of HGSOC includes initial surgical debulking and subsequent systemic or intraperitoneal carboplatin and paclitaxel. While many tumors initially respond, 60--85% of patients experience disease recurrence following primary therapy.^[@bib1],[@bib3]^ Relapse is often accompanied by disease that has acquired resistance to these drugs. One mechanism implicated in recurrence is the evasion of apoptosis, a form of programmed cell death whose loss represents an established hallmark of cancer.^[@bib4]^ Exploiting alternative cell death pathways, including necroptosis ('programmed necrosis\\'), may offer an alternative strategy to treat such recurrent disease.^[@bib5]^\n\nThe cellular inhibitor of apoptosis proteins (c-IAP1 and c-IAP2) represent promising targets for therapy, as they are overexpressed in many cancers and have important roles in both apoptotic and necroptotic death pathways.^[@bib6]^ Upon binding of tumor necrosis factor *\u03b1* (TNF*\u03b1*) to TNF*\u03b1* receptor 1, the adaptor protein TRADD (tumor necrosis factor receptor type 1-associated death domain protein) is recruited to the cytosolic death domain of TNF*\u03b1* receptor 1.^[@bib7]^ This facilitates subsequent receptor-interacting protein kinase-1 (RIPK1)^[@bib8]^ and TRAF2/5 (TNF receptor-associated factor 2/5) binding,^[@bib9]^ which leads to cellular inhibitor of apoptosis protein 1/2 (c-IAP1/2) recruitment. The formation of this TNF*\u03b1*-induced membrane-associated complex, known as complex I, results in NF-*\u03ba*B (nuclear factor-*\u03ba*-light-chain-enhancer of activated B cells) activation, apoptosis or necroptosis, depending on the proceeding intracellular signaling events.\n\nUbiquitination of RIPK1 by the E3 ligase c-IAP1/2 activates the canonical NF-*\u03ba*B pathway, resulting in cell survival.^[@bib10]^ Conversely, the deubiquitination of RIPK1 or, alternatively, the lack of ubiquitination following IAP antagonist (IAPa or I)-induced degradation of c-IAP1/2 leads to the formation of a cytosolic complex that includes FADD (Fas-associated protein with death domain) and caspase-8. This triggers *apoptosis* if caspases are active (complex IIa), or receptor-interacting serine--threonine kinase-3 (RIPK3)-dependent *necroptosis* in the presence of caspase inhibitors (complex IIb; necrosome).^[@bib11],[@bib12]^\n\nIAPa will induce apoptosis in specific triple-negative breast or ovarian cancer cell lines,^[@bib13],\\ [@bib14],\\ [@bib15]^ an observation that supports the launch of NCT01681368: *A Phase II trial of the IAP antagonist Birinipant for advanced ovarian, fallopian tube, and peritoneal cancer* (http://[www.clinicaltrials.gov](http://www.clinicaltrials.gov)). In contrast, activation of TNF*\u03b1*-dependent necroptosis via IAP antagonism has been demonstrated in a colon cancer cell line,^[@bib16]^ although not in other carcinomas.\n\nThe triggering of necroptotic signaling among ovarian cancer cells could have important therapeutic implications. Here, we demonstrate that IAP antagonism, in the presence of a caspase inhibitor, triggers necroptosis in apoptosis-resistant ovarian cancer cells, and, moreover, that patient-derived tumor cells display sensitivity to this treatment. Activation of this death pathway was dependent on the kinase activity of RIPK3, and required RIPK1, both of which we show are frequently expressed in serous ovarian cancer, as well as the cytokine TNF*\u03b1*. Notably, this death pathway occurred selectively in apoptosis-resistant cells. The results suggest a common functional necroptosis pathway that might be exploited in future therapeutic designs.\n\nResults\n=======\n\nTreatment with IAPa and the protease inhibitor zVAD induces death of RIPK3-expressing ovarian cancer cells\n----------------------------------------------------------------------------------------------------------\n\nA panel of human ovarian cancer cell lines were screened for the induction of apoptosis or necroptosis (as depicted in [Figure 1a](#fig1){ref-type=\"fig\"}), following treatment with IAPa.^[@bib14]^ Supporting the impetus to move IAPa forward in clinical trials, two of these (OVCAR4 and SKOV-3) showed some sensitivity to IAPa as a single agent ([Figure 1b](#fig1){ref-type=\"fig\"} and He *et al.*^[@bib11]^). This was at least partially blocked by the inhibitor of apoptosis z-VAD-fmk (pancaspase inhibitor, Z-Val-Ala-DL-Asp-FMK, Z) IAPa plus zVAD; IZ, which selectively, but not exclusively, inhibits caspase activity. Importantly, and perhaps not surprisingly, the majority of tumor lines tested showed resistance to the induction of apoptosis by either IAPa ([Figure 1b](#fig1){ref-type=\"fig\"}: Hey, OVCAR5 and OVCAR3; not shown: IGROV1) or cisplatin ([Supplementary Figure 1a](#sup1){ref-type=\"supplementary-material\"}). Neither caspase-8 nor FADD appeared to be limiting for apoptosis ([Figure 1c](#fig1){ref-type=\"fig\"}). Indeed, the cells with the lowest expression of FADD were the most sensitive to the induction of apoptosis by IAP antagonism.\n\nInterestingly, zVAD treatment actually promoted, rather than rescued, death in some cell lines ([Figure 1b](#fig1){ref-type=\"fig\"}, bottom panels). This raised the possibility of the induction of an alternative form of programmed cell death: necroptosis. This notion was bolstered by the observation that apoptosis-resistant but IAP antagonist plus caspase inhibitor (IZ)-sensitive lines exhibited expression of RIPK3 ([Figures 1b and c](#fig1){ref-type=\"fig\"}), a critical regulator of necrotic cell death.^[@bib11]^ Further supporting this possibility, cell death induced by IZ was not accompanied by the activation of caspases, as occurs during apoptosis ([Figure 1d](#fig1){ref-type=\"fig\"}).^[@bib6]^ As the concept that tumor cells (in particular serous ovarian tumor cells) might be sensitive to necroptosis had not been previously explored, we characterized this cell death further.\n\nFormation of the necrosome in IZ-sensitive cells\n------------------------------------------------\n\nIt remained possible that necrosis occurred as a default pathway when IAPa were ineffective at inducing the clearance of IAPs required for apoptosis.^[@bib13]^ To test this, we first evaluated the presence of two IAPs (c-IAP1 and c-IAP2) following antagonist treatment ([Figure 2a](#fig2){ref-type=\"fig\"}). As shown, IAPa treatment resulted in the complete and persistent loss of cIAPs within minutes. Thus, IAP loss is consistent with necroptotic death. However, a general loss of all IAPs was not observed, as treatment did not appear to influence the expression of X-linked inhibitor of apoptosis (XIAP) ([Figure 2b](#fig2){ref-type=\"fig\"}). A third target of IAPa, ML-IAP (a member of IAP family, containing a single copy of a baculovirus IAP repeat (BIR) as well as a RING-type zinc-finger domain), was not expressed in these cells ([Supplementary Figure 1b](#sup1){ref-type=\"supplementary-material\"}).\n\nTo evaluate whether a functional necrosome complex was indeed forming, we next immunoprecipitated RIPK3 expressed in the ovarian cancer cells and tested for the presence of associated proteins. IZ treatment resulted in the formation of a complex with abundant representation of RIPK1, but with much lower levels of FADD and caspase-8 ([Figure 2c](#fig2){ref-type=\"fig\"}). Treatment with either agent alone (I or Z) did not result in the formation of a complex ([Figure 2c](#fig2){ref-type=\"fig\"}). In contrast, mixed lineage kinase domain-like (MLKL) was constitutively associated with RIPK3 under all three conditions ([Figure 2d](#fig2){ref-type=\"fig\"}). As formation of the necrosome promotes the phosphorylation of MLKL as a downstream effector of RIPK3, we next confirmed that IZ treatment did, in fact, result in marked phosphorylation of MLKL ([Figure 2e](#fig2){ref-type=\"fig\"}). In fact, simple ectopic expression of RIPK3 led to a lower basal level of MLKL phosphorylation, although this was insufficient to induce cell death.\n\nFurther supporting a model of classical RIPK1-dependent necroptosis, IZ-induced death was rescued by the inclusion of necrostatin-1, a RIPK1 inhibitor^[@bib17]^ ([Figure 2f](#fig2){ref-type=\"fig\"}). Cell death was alternatively also blocked by inclusion of necrosulfonamide ([Figure 2f](#fig2){ref-type=\"fig\"}), which targets the death-inducing RIPK3 substrate, MLKL.^[@bib18]^ These results, together, suggest that necroptosis is induced by activation of RIPK1 followed by activation of RIPK3 and MLKL in ovarian cancer cells.\n\nRIPK1 expression is required for IZ-induced death of ovarian cancer cells\n-------------------------------------------------------------------------\n\nAll of the ovarian cancer tumor cell lines examined expressed RIPK1, and its expression was not sufficient for, or predictive of, cell death following I or IZ treatment ([Figure 1b](#fig1){ref-type=\"fig\"}). As necrostatin-1 inhibited cell death, we next silenced RIPK1 expression ([Figure 3a](#fig3){ref-type=\"fig\"}, inset) and then evaluated cell survival following exposure to IZ. As shown, the loss of RIPK1 was protective, consistent with the necrostatin sensitivity of the death pathway ([Figure 3a](#fig3){ref-type=\"fig\"}). By contrast, the knockdown of the adaptor protein FADD had a more modest effect ([Figure 3b](#fig3){ref-type=\"fig\"}), suggesting that FADD may provide some facilitating role in necroptosis in the ovarian cancer cells, yet was not critical in facilitating RIPK1 to RIPK3 signaling.\n\nRIPK3, but not an inactive mutant, promotes IZ-induced death\n------------------------------------------------------------\n\nTo more carefully characterize the requirement for RIPK3, we next used genetic approaches to reconstitute RIPK3 expression in the RIPK-deficient, death-resistant HEY cells ([Figures 1b and c](#fig1){ref-type=\"fig\"}) and to suppress RIPK3 expression in the death-inducible OVCAR3 cells. In agreement with the capacity of necrostatin to block IZ-induced death, the expression of RIPK3, but not a kinase dead RIPK3 mutant, was sufficient to confer sensitivity to IZ-induced death among HEY cells ([Figure 4a](#fig4){ref-type=\"fig\"}). The expression level of the kinase dead version was, however, somewhat lower than that of the wild-type kinase, possibly due to other recently described death-inducing effects of catalytically compromised RIPK3.^[@bib19]^ Selection of clones of HEY with much lower expression of RIPK3 similarly resulted in cell death following IZ treatment ([Figure 4b](#fig4){ref-type=\"fig\"}), suggesting that relatively low levels of RIPK3 may be sufficient to confer sensitivity. In agreement with these observations, the loss of RIPK3 expression compromised cell death in IZ-sensitive cell lines ([Figure 4c](#fig4){ref-type=\"fig\"}). Thus, RIPK3 expression is a prerequisite for this cellular death pathway.\n\nIZ treatment promotes autocrine TNF*\u03b1* production\n-------------------------------------------------\n\nThe TNF*\u03b1* receptor I must be activated by ligation and subsequent formation of the membrane-associated complex I (depicted in [Figure 1a](#fig1){ref-type=\"fig\"}) for the necrosome to form and lead to necrotic cell death. We reasoned that without exogenous death receptor ligand, IZ-sensitive ovarian cancer cells might have to produce their own extrinsic death pathway stimulus. The relative TNF*\u03b1* mRNA levels in I- or Z-only treated cells was comparable to levels in vehicle-treated cells; however, a marked upregulation of TNF*\u03b1* was observed in IZ-treated OVCAR3 cells ([Figure 5a](#fig5){ref-type=\"fig\"}), which was dependent on the expression of RIPK1 ([Figure 5b](#fig5){ref-type=\"fig\"}). Ectopic expression of RIPK3 rescued IZ-induced TNF*\u03b1* in HEY cells ([Figure 5c](#fig5){ref-type=\"fig\"}), suggesting that RIPK3 expression may enhance autocrine TNF*\u03b1* production by the ovarian cancer cells as it does in myeloid cells.^[@bib20]^ The kinase activity of RIPK3 was critical for this effect, as the expression of catalytically dead RIPK3 (K50A) did not induced the expression of TNF*\u03b1* ([Figure 5c](#fig5){ref-type=\"fig\"}). TNF*\u03b1* appeared to be required for death, as a TNF*\u03b1-*neutralizing antibody was able to rescue cell survival in ovarian carcinomas expressing endogenous ([Figure 5d](#fig5){ref-type=\"fig\"}) or ectopic ([Figure 5e](#fig5){ref-type=\"fig\"}) RIPK3.\n\nRIPK3 expression in ovarian cancer\n----------------------------------\n\nAlthough RIPK3 expression promotes necroptosis, we only noted the expression of RIPK3 in a few ovarian cancer cell lines. This raised the question as to what the relative expression of RIPK3 in ovarian tumors might be. We focused on serous ovarian cancer, which is well curated in the The Cancer Genome Atlas repository and also accounts for the greatest overall mortality among ovarian carcinoma patients. The database cataloged an enrichment for RIPK3 expression in serous ovarian cancer ([Figure 6a](#fig6){ref-type=\"fig\"}). In fact, more than three-quarters of serous ovarian tumors expressed RIPK3 levels above the mean expression established across all tumors. Almost half of tumors were enriched at least twofold over the mean expression, as opposed to only one in six tumors, which was twofold or less below the mean ([Figure 6a](#fig6){ref-type=\"fig\"}). As we had previously observed in the OVCAR3 cells, and consistent with the TCGA data set, our two cisplatin-resistant patient-derived tumor cells (OV no. 2224 and OV no. 3971) revealed RIPK3 protein expression upon immunoblot analysis ([Figure 6b](#fig6){ref-type=\"fig\"}). Moreover, the patient-derived tumor cells demonstrated IZ-induced autocrine expression of TNF*\u03b1* ([Figure 6c](#fig6){ref-type=\"fig\"}). Accordingly, we observed no induction of apoptosis with IAP antagonism alone, but IZ-induced cell death ([Figure 6d](#fig6){ref-type=\"fig\"}) that was blocked by TNF*\u03b1*-neutralizing antibody ([Figures 6e and f](#fig6){ref-type=\"fig\"}). Altogether, the results support a model in which necroptotic cell death is common following IAP antagonism.\n\nDiscussion\n==========\n\nThe poor prognosis for patients with advanced ovarian cancer, combined with the universal loss of tumor protein p53 (TP53) in an otherwise molecular heterogeneous background, has led to increasingly innovative approaches to treat the disease. Although 70--80% of disease will respond to initial chemotherapy, recurrence of recalcitrant disease is common. Interestingly, the IZ-sensitive patient tumor cells used here also exhibited quite high cisplatin resistance (half-maximal inhibitory concentration \\>10\u2009*\u03bc*M). This observation in particular would appear to support the development of agents, such as Birinipant, that antagonize IAPs as adjuvants to existing therapies to prevent the selection of cisplatin resistance in tumors.\n\nWe examined several ovarian cell lines for their ability to undergo death following IAPa treatment. This was important, as attention has recently focused on the suitability of commonly used cell lines as models for serous ovarian cancer. Domcke *et al.*^[@bib1]^ assessed the genetic similarity between commonly used ovarian cancer cell lines and serous ovarian tumor tissue samples to determine the likelihood of each cell line being of HGSOC origin. Interestingly, while the OVCAR3 cell line we study here was found to be similar to the general tumor tissue signature in this study, the other ovarian cancer cell lines that remain part of the NCI-60, such as HEY and SKOV-3, were not. Perhaps more importantly, the molecular signature of cell death proteins and the response to antagonism revealed that OVCAR3 cells were quite similar to patient-derived tumor cells typified by OV\\#2224 and OV\\#3871, and thus acted as a representative cell line for this system. Indeed, the cisplatin-resistant OVCAR3 may represent a preferred cell line for the analysis of emerging ovarian cancer treatments.\n\nWe demonstrate that TNF*\u03b1* production is required to trigger necroptosis in the ovarian cancer cells, and that RIPK3 enhanced this process. Consistent with recent reports, depletion of RIPK3 has a marked impact on the induction of necrosis,^[@bib16]^ and notably, the expression of RIPK3 was sufficient to restore sensitivity to necroptosis in the apoptosis and necroptosis-resistant HEY cells. This required the production of TNF*\u03b1*, as antibodies to TNF*\u03b1* prevented IZ-mediated killing. Nonetheless, the IZ combination elicits additional effects, as simple addition of TNF*\u03b1* to either I or Z alone was not sufficient to promote cell death (KM, DGS, unpublished observations).\n\nIn these studies, the source of TNF*\u03b1* was autocrine. However, caution is warranted in overinterpreting this requirement *in vivo*, as TNF*\u03b1* can be produced by stromal or inflammatory cells in the tumor microenvironment.^[@bib21]^ In this case, the *in vivo* use of IAPa that are thought to induce effectively apoptosis in patients might rather be effecting cell death by necroptosis. The requirements for, and the form of death may be important to understand in the clinical setting. In fact, understanding the type of death induced by an agent *in vivo* is more than just a semantic distinction. Increasingly, molecular determinants are used for the stratification of patients into preferred treatment groups. This is progressing rapidly, and the initial single gene requirements for patient stratification^[@bib22]^ are destined to become a thing of the past, yielding to more complex algorithms for personalized medicine.^[@bib23]^ Thus, the question raised by the current studies, that is, whether tumor cells undergo apoptosis or necroptosis *in vivo* following exposure to IAPs, becomes a key factor for the implementation of therapy, and should be addressed in future directed studies.\n\nMaterials and Methods\n=====================\n\nReagents\n--------\n\nIAPa were synthesized by Mitchell Vamos (Sanford-Burnham Medical Research Institute, La Jolla, CA, USA). The pancaspase inhibitor zVAD-fmk was purchased from Bachem (Torrance, CA, USA) (N-1510.0005BA). Necrostatin-1 (2324/10) and necrosulfonamide (432531-71-0) were purchased from R&D Systems (Minneapolis, MN, USA) and EMD Millipore (Temecula, CA, USA), respectively. The TNF*\u03b1-*neutralizing antibody was purchased from Cell Signaling Technology (Danvers, MA, USA; 7321). The following primary antibodies were used for western blotting: c-IAP1 (Cell Signaling Technology; 7065), c-IAP2 (Cell Signaling Technology; 3130), RIPK3 (Cell Signaling Technology; 12107), RIPK1 (Cell Signaling Technology; 3493), caspase-8 (BD Biosciences; 551242), FADD (BD Biosciences; 610399), p62 (BD Biosciences; 610832), LC3 (Novus Biologicals, Littleton, CO, USA; NB100-2220), *\u03b2*-actin (Sigma, St. Louis, MO, USA; A5441), PARP (BD Biosciences; 51-6639GR), caspase-3 (Cell Signaling Technology; 9668), caspase-7 (Cell Signaling Technology; 9492), XIAP (Cell Signaling Technology; 5471), MLKL (EMD Millipore; MAB C604), phospho-MLKL (Abcam, Vancouver, BC, USA; Ab187091) and *\u03b1*-tubulin (EMD Millipore; CP06). The following horseradish peroxidase-conjugated secondary antibodies were used for western blotting: goat anti-rabbit IgG (H+L) (Jackson ImmunoResearch, Bar Habor, ME, USA; 111-035-003), goat anti-mouse IgG (H+L) (Jackson ImmunoResearch; 115-035-003), goat anti-rabbit IgG, Fc fragment-specific (Jackson ImmunoResearch; 111-035-008) and mouse anti-rabbit IgG, light-chain-specific (Jackson ImmunoResearch; 211-032-171). The anti-mCherry antibody used for immunoprecipitation was generated in-house by immunization of rabbits with full-length mCherry and immunoaffinity purification of immune serum on a Sephadex-mCherry column.\n\nCell culture\n------------\n\nOV no. 2224 and OV no. 3971 cells were isolated from ovarian cancer patients, subcultured^[@bib12]^ and passaged through mice. Informed consent was obtained to use these cells for research purposes. The human ovarian cancer cell lines HEY, OVCAR3, OVCAR4, OVCAR5 and SKOV-3 were maintained in no-glucose RPMI-1640 (Mediatech Inc., Manassas, VA, USA) supplemented with 1\u2009g/l glucose, 10% FBS, MEM nonessential amino acids and 1\u2009mM sodium pyruvate (Mediatech Inc.). All cells were maintained in an incubator set to 37\u2009\u00b0C and 5% CO~2~.\n\nGeneration of stable cell lines\n-------------------------------\n\nAll stable cell lines were generated by using lentiviral-based RNAi or lentiviral-based gene expression systems, as described previously.^[@bib24]^ Briefly, shRIPK3 and shRIPK1 cell lines were selected for with 1\u2009*\u03bc*g/ml puromycin after being infected with lentivirus containing shRNA constructs specifically targeting RIPK3 or RIPK1 mRNA. Cell lines (OVCAR3 and HEY) stably expressing mCherry-tagged RIPK3 or a kinase dead mutant (K50A) were selected with 1\u2009*\u03bc*g/ml puromycin following lentivirus treatment with a pLV RIPK3-mCherry plasmid.\n\nIAP antagonism time-course assay\n--------------------------------\n\nCells were seeded into the wells of 6-well plates (2 \u00d7 10^6^ cells/well) and allowed to adhere for \\~16\u2009h at 37\u2009\u00b0C before the addition of drug(s). Media were replaced with fresh media containing 1\u2009*\u03bc*M IAPa and cells were incubated at 37\u2009\u00b0C for various amounts of time. At selected time points, media were aspirated and cells were washed two times with PBS and lysed with 150\u2009*\u03bc*l RIPA lysis buffer supplemented with a protease inhibitor cocktail (Sigma), 2\u2009mM sodium orthovanadate and 50\u2009mM NaF. Cells were collected using a cell lifter (Thermo Fisher, Waltham, MA, USA), transferred to an Eppendorf tube, passed through a pipet tip five times and incubated on ice for 10\u2009min. The lysate was centrifuged at 14\u2009000 \u00d7 *g* for 5\u2009min at 4\u2009\u00b0C, and the supernatant (lysate) was collected and stored at \u221220\u2009\u00b0C until western blot analysis.\n\nGrowth inhibition assay\n-----------------------\n\nCells were seeded into wells of a 96-well plate (10^4^ cells/well; 100\u2009*\u03bc*l final volume) and allowed to adhere for \\~16\u2009h at 37\u2009\u00b0C before the addition of drug(s). Treatment or vehicle was added in a total dosage volume of 11\u2009*\u03bc*l (to final concentrations of 10\u2009*\u03bc*M necrostatin, 2\u2009*\u03bc*M necrosulfonamide, 20\u2009*\u03bc*M zVAD-fmk or 1\u2009*\u03bc*M IAPa). Cells were incubated at 37\u2009\u00b0C for 48\u2009h, washed gently two times with 100\u2009*\u03bc*l PBS and stained with 75\u2009*\u03bc*l crystal violet (Sigma) aqueous dye solution (0.1% crystal violet, 150\u2009mM NaCl, 20% methanol) for 25\u2009min\u2009 at room temperature. Stain was removed and cells were washed two times with water to remove unbound dye, and then allowed to dry. Cell-bound crystal violet was resolubilized with 75\u2009*\u03bc*l methanol and *A*~600~ was measured using a PowerWave XS2 Microplate Spectrophotometer (BioTek). All treatment conditions were assayed in triplicate and plotted as mean\u00b1S.D. Except as noted, all experiments were performed three or more times with similar results. 'Cell loss\\' was calculated using the formula: 100\u2212((*A*~600~ treatment/*A*~600~ vehicle) \u00d7 100).\n\nImmunoprecipitation\n-------------------\n\nFor the immunoprecipitation of mCherry-tagged RIPK3, cells were lysed with 'NP-40 Alternative Lysis\\' (NAL) buffer (EMD Millipore; 492016) buffer (100\u2009mM Tris, pH 8.0, 150\u2009mM NaCl, 1% NP-40 Alternative) following 24\u2009h exposure to the indicated agents. Cell mixtures were centrifuged at 2500 \u00d7 *g* for 3\u2009min, then total protein concentration was determined as described and 1\u2009mg protein was diluted with lysis buffer to a final volume of 1\u2009ml. Lysates were incubated with 5\u2009*\u03bc*g rabbit anti-mCherry antibody for \\~16\u2009h at 4\u2009\u00b0C, followed by 20\u2009*\u03bc*l Protein A/G resin (G-Biosciences, St. Louis, MO, USA) for \\~3 h at 4\u2009\u00b0C. Following centrifugation at 2500 \u00d7 *g* for 3\u2009min, resin was washed four times with NP-40 Alternative wash buffer (20\u2009mM Tris, pH 7.4, 150\u2009mM NaCl, 0.1% NP-40 Alternative). Immunoprecipitates were eluted by heating in reducing sample buffer for 5\u2009min at 100\u2009\u00b0C, cooled to room temperature and briefly centrifuged to pellet beads. Supernatants were analyzed by SDS-PAGE and subjected to immunoblot analysis.\n\nImmunoblot analysis\n-------------------\n\nCell lysates were generated using NP-40 Alternative Lysis buffer, as described above, when cells reached \\~75% confluence. In treatment cases, untreated cells or cells treated following a 24\u2009h incubation with indicated agents were lysed as described above. Total protein concentration was determined using the Micro BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA), and 30\u2009*\u03bc*g lysate was loaded onto 10% SDS-PAGE gels, resolved and transferred to PVDF membrane (Bio-Rad). Following incubation with 5% milk/0.1% Tween 20 (Thermo Fisher)/TBS blocking buffer, membranes were subjected to immunoblotting. Western blots were incubated with Super Signal West Pico Chemiluminescent Substrate (Thermo Scientific) or Super Signal West Dura Chemiluminescent Substrate (Thermo Scientific) exposed to film, and the film was developed.\n\nQuantitative PCR\n----------------\n\nCells were seeded into wells of a 6-well plate (2 \u00d7 10^6^ cells/well) and allowed to adhere for \\~16\u2009h at 37\u2009\u00b0C before the addition of agents. Media were replaced with 1 ml fresh media containing treatment or vehicle, and cells were incubated at 37\u2009\u00b0C for 48\u2009h. Media were aspirated, and cells were washed with PBS and incubated with 0.5\u2009ml trypsin for several minutes at 37\u2009\u00b0C. An equal volume of media were added to each well and cells were transferred to Eppendorf tubes and centrifuged at 1000 \u00d7 *g* for 5\u2009min. The supernatant was aspirated and RNA was extracted using a NucleoSpin RNA Kit (Macherey-Nagel, Bethlehem, PA, USA) according to the manufacturer\\'s instructions. RNA concentration was determined by measuring *A*~260~, and SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) was used to synthesize cDNA from 1\u2009*\u03bc*g RNA. Quantitative PCR was performed using a 1\u2009:\u200910 dilution of generated cDNA, primers specific for human actin and TNF*\u03b1*, LightCycler 480 SYBR Green I Master (Roche, South San Francisco, CA, USA) and a LightCycler 480 Instrument (Roche).\n\nStatistics\n----------\n\nData highlighted as significant, as indicated with an asterisk (\\*) in each figure, were determined in all cases by Student\\'s *T*-test. The threshold for significance in all cases was set with a probability of \\<0.05.\n\nResearch reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Awards T32CA121938 (to KM, JD), CA107263 (to DGS) and by U54 HG00503 ( to NDPC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Additional support was provided by the Tegger foundation, Wenner-Gren foundations and the G\u00e5l\u00f6 foundation (to KB) and Grant 2013CB967200 (to RX) from the National Key Scientific Research Projects of China.\n\nIAP\n\n:   inhibitor of apoptosis protein\n\nIAPa or I\n\n:   IAP antagonist\n\nzVAD or Z\n\n:   pancaspase inhibitor; Z-Val-Ala-DL-Asp-FMKIZ, IAPa plus zVAD\n\nTNF*\u03b1*\n\n:   tumor necrosis factor-*\u03b1*\n\nRIPK3\n\n:   receptor-interacting serine--threonine kinase-3\n\nRIPK1\n\n:   receptor-interacting serine--threonine kinase-1\n\nHGSOC\n\n:   high-grade serous ovarian cancer\n\nc-IAP1/2\n\n:   cellular inhibitor of apoptosis protein 1/2\n\nTRAF2/5\n\n:   TNF receptor-associated factor 2/5\n\nNF-*\u03ba*B\n\n:   nuclear factor-*\u03ba*-light-chain-enhancer of activated B cells\n\nFADD\n\n:   Fas-associated protein with death domain\n\nXIAP\n\n:   X-linked inhibitor of apoptosis\n\nML-IAP\n\n:   a member of IAP family, containing a single copy of a baculovirus IAP repeat (BIR) as well as a RING-type zinc-finger domain\n\nMLKL\n\n:   mixed lineage kinase domain-like\n\n[Supplementary Information](#sup1){ref-type=\"supplementary-material\"} accompanies this paper on Cell Death and Disease website (http://www.nature.com/cddis)\n\nEdited by H-U Simon\n\nThe authors declare no conflict of interest.\n\nSupplementary Material {#sup1}\n======================\n\n###### \n\nClick here for additional data file.\n\n![Activation of death pathways in ovarian cancer cells. (**a**) In the presence of IAP antagonism, TNF*\u03b1* receptor-mediated signaling can lead to apoptosis, or, in the presence of inhibitors of caspases such as zVAD that block apoptosis, a necrotic death triggered by RIPK1 and RIPK3. (**b**) Ovarian cancer cell lines treated for 48\u2009h with diluent (Con), I (1\u2009*\u03bc*M), Z (20\u2009*\u03bc*M) or both (I+Z) were evaluated for cell loss (relative to Con). A representative experiment (of three or more for each cell line) is shown as mean of triplicates\u00b1S.D. \\**P*\\<0.05 *versus* control. (**c**) The expression of proteins contributing to apoptosis or necroptosis was evaluated in ovarian cancer cells as indicated by immunoblot analysis. (**d**) Representative apoptotic (OVCAR4) and necroptotic (OVCAR3) cell lines were evaluated for poly-ADP ribose polymerase (PARP) cleavage and their capacity to elicit caspase maturation following 24\u2009h treatment with I, Z or IZ as described in (**b**), above](cddis2014448f1){#fig1}\n\n![Evaluation of the necroptotic phenotype in ovarian cancer. Time course showing the effect of incubation of I (1\u2009*\u03bc*M) on (**a**) cIAP levels and (**b**) XIAP levels among ovarian tumor cells (OVCAR3) that exhibit a 'necroptotic phenotype\\'. Actin was used as a loading control as XIAP and tubulin have a similar mass. (**c**) Immunoprecipitation of a complex containing RIPK1 and RIPK3, as well as FADD and caspase-8, following treatment of OVCAR3 cells expressing mCherry-tagged RIPK3 with either I (1\u2009*\u03bc*M), Z (20\u2009*\u03bc*M) or both (I+Z) for 24\u2009h. (**d**) Formation of a RIPK3 complex containing MLKL was evaluated by immunoblotting the mCherry immunoprecipitates described. The MLKL-reactive species are shown as open arrowheads. (**e**) Evaluation of MLKL phosphorylation in OVCAR3 cells or OVCAR3 cells expressing ectopic RIPK3 as a positive control. Cells were again treated with I, Z or I+Z as detailed above, and expression of pMLKL or total MLKL was evaluated by immunoblot analysis. (**f**) Effect of I, Z, necrostatin (Nec, 10\u2009*\u03bc*M) or necrosulfonamide (2\u2009*\u03bc*M) treatment, alone or added in combinations as shown, for 48\u2009h, on ovarian cancer cell death. Mean\u00b1S.D. of triplicates from representative experiments are shown. \\**P*\\<0.05 *versus* control](cddis2014448f2){#fig2}\n\n![RIPK1 suppression compromises I+Z-mediated death in ovarian cancer cells. (**a**) The effect of suppression of RIPK1 expression by shRNA (see inset) on OVCAR3 cell death was examined following treatment with I (1\u2009*\u03bc*M), Z (20\u2009*\u03bc*M), necrostatin (Nec), diluent controls (Con) or combinations of these agents as shown for 48\u2009h. (**b**) The effect of suppression of the adaptor protein FADD (inset) was similarly evaluated. Mean\u00b1S.D. of triplicates from representative experiments are shown. \\**P*\\<0.05 *versus* control](cddis2014448f3){#fig3}\n\n![RIPK3 expression is required to trigger cell death after I+Z treatment. (**a**) The effect of reconstitution of expression of RIPK3, or a catalytically dead mutant (K50A), in the death-resistant HEY cells was assessed for impact on cell death 48\u2009h following treatment with I (1\u2009*\u03bc*M), Z (10\u2009*\u03bc*M), necrostatin (Nec, 10\u2009*\u03bc*M), diluent controls (Con) or combinations of these agents as shown. (**b**) Expression of low levels of RIPK3 was also evaluated for impact on cell death, as described above. (**c**) The effect of suppression of RIPK3 by shRNA (inset) on OVCAR3 cell death was examined following treatment with I (1\u2009*\u03bc*M), Z (10\u2009*\u03bc*M), diluent control or combinations of these agents as shown. Mean\u00b1S.D. of triplicates from representative experiments are shown. \\**P*\\<0.05 *versus* control](cddis2014448f4){#fig4}\n\n![TNF*\u03b1* transcription is induced by I+Z treatment. (**a**) The relative level of TNF*\u03b1* transcript was evaluated by qPCR in diluent-treated OVCAR3 cells or following treatment with I (1\u2009*\u03bc*M), Z (20\u2009*\u03bc*M) or both. (**b**) Similar evaluation in OVCAR3 cells expressing shRNA to RIPK1 or a control shRNA, or in (**c**) HEY cells reconstituted with RIPK3 or a catalytically dead mutant (K50A) RIPK3. The influence of treatment anti-TNF*\u03b1* (5\u2009*\u03bc*g/ml) on I+Z-induced cell death was evaluated in the OVCAR3 cells (**d**) or HEY cells reconstituted with RIPK3 (**e**). Mean\u00b1S.D. of triplicates from representative experiments are shown. **a**--**c**: \\**P*\\<0.05 *versus* control; **d** and **e**: \\**P*\\<0.05 for I+Z control *versus* either untreated control or anti-TNF*\u03b1*](cddis2014448f5){#fig5}\n\n![Expression of ripoptotic effectors in ovarian cancer. (**a**) The distribution of RIPK3 transcripts across all aggressive serous ovarian cancers curated in the TCGA web portal. (**b**) Immunoblot analysis for key effectors of apoptosis/necroptosis in xenograft tumor cells lysates derived from ovarian cancer patients. (**c**) TNF*\u03b1* transcript was evaluated by qPCR in patient-derived serous ovarian cancer cells following treatment with diluent (Con), I (1\u2009*\u03bc*M), Z (10\u2009*\u03bc*M) or both agents. (**d**) Patient-derived ovarian cancer cells were treated for 48\u2009h with diluent (Con), I (1\u2009*\u03bc*M), Z (20\u2009*\u03bc*M) or both agents (I+Z) and evaluated cell death. (**e**) The impact of treatment with nonspecific Ig or anti-TNF*\u03b1* (5\u2009*\u03bc*g/ml) on IZ-induced cell death of patient-derived ovarian cancer cells. Mean\u00b1S.D. from one of two similar studies are shown. **c** and **d**: \\**P*\\<0.05 *versus* control; **e** and **f**: \\**P*\\<0.05 I+Z for control *versus* either untreated control or anti-TNF*\u03b1*](cddis2014448f6){#fig6}\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nDry eye disease (DED) affects millions of people around the world \\[[@CR1]\\]. Whether due to aqueous deficiency or evaporative dry eye (or most commonly a combination of factors), it is a potentially debilitating condition. Functional limitation is variable, but moderate to severe DED may make it difficult for patients to do computer work, reading, and other activities that limit their daily lives and productivity. DED was traditionally thought to be a disease of age and more common in females. In recent years, a younger population is the most rapidly growing segment of dry eye sufferers, likely in part to shifts in our lifestyles toward frequent computer and visual display tasking. Post-refractive surgery (e.g., post-LASIK) dry eye is also a significant clinical obstacle \\[[@CR2]\\]. The current mainstay treatment for mild dry eye includes various types of lubricating drops and ointments. These must be frequently re-applied and do not resolve the underlying disease process, often only temporarily relieving the symptoms.\n\nThis article is based on previously conducted studies and does not involve any new studies of human or animal subjects performed by any of the authors.\n\nDry Eye as an Inflammatory State---Current Treatments and Implications for Pharmaceutical Development {#Sec2}\n-----------------------------------------------------------------------------------------------------\n\nDry eye has been recognized and treated as an inflammatory process. The first pharmaceutical to target an inflammatory pathway in order to treat dry eye is cyclosporine 0.05% (Restasis, Allergan; Irvine, CA, USA), which has become a mainstay of treatment for moderate to severe dry eye disease. While many patients benefit from this drug, some also have mixed reactions and may discontinue it over time due to intolerance of associated side effects (temporary redness and/or burning upon instillation) \\[[@CR3]\\]. The need for additional treatments for dry eye is clear to patients and physicians alike, and the recognition of dry eye as an inflammatory process has opened the door to pharmaceuticals targeting the inflammatory cascade. Fortunately for dry eye patients, there are numerous promising pharmaceuticals in the pipeline, some borrowing insight from treatment of inflammatory conditions in other parts of the body, and many of them having completed or currently undergoing phase 3 clinical testing. Additionally, dry eye has become a target for device development, including for example the Oculeve Neurostimulator device and the EyeGate Ocular Iontophoresis device. These devices address the treatment of dry eye with very different and innovative approaches, which are reviewed below. Clinical trials for medications and devices to treat DED need to be well-designed in order to be successful, as dry eye is a disease in which symptoms may not correlate well with objective testing \\[[@CR1]\\]. Even prior large-scale studies of DED have varied significantly in their method of assessing disease severity \\[[@CR4]--[@CR6]\\]. Despite this challenge, using a combination of subjective patient symptom questionnaires and objective outcome measures, several promising developments are on the horizon. This review summarizes the major developments for dry eye treatment currently in the therapeutic pipeline.\n\nPharmaceuticals in Phase 3 Clinical Trials {#Sec3}\n==========================================\n\nLifitegrast 5%: Small-Molecule Integrin Antagonist {#Sec4}\n--------------------------------------------------\n\nLifitegrast (Shire Pharmaceuticals; Lexington, MA, USA) is a small-molecule antagonist of the T cell integrin, lymphocyte function antigen-1 (LFA-1) \\[[@CR7]--[@CR9]\\]. LFA-1 plays a role in the inflammatory cascade of dry eye, binding intercellular adhesion molecule-1 (ICAM-1) and promoting T-cell activation and eventually cytokine release. Therefore, inhibiting LFA-1 effectively limits T-cell mediated inflammation. OPUS-1, a phase 3, randomized, prospective, double-masked placebo-controlled study compared lifitegrast twice daily with placebo in 565 subjects with a history of bilateral dry eye disease over the course of 12\u00a0weeks \\[[@CR7]\\]. Measured outcomes included corneal fluorescein staining and conjunctival lissamine staining, Schirmer tear test, visual analog scale, and ocular discomfort score. Patients receiving lifitegrast showed statistically significant superior improvement over placebo in total corneal staining, total lissamine staining, mean eye dryness score, and ocular discomfort score over the course of the study, while other measured outcomes were not statistically different between the treatment and placebo groups at 12\u00a0weeks. Decreased lissamine staining in the treatment group was demonstrated as early as day 14, which may indicate a relatively rapid decrease in ocular surface inflammation with lifitegrast. Some limitations of this phase 3 study include the relatively short follow-up period. OPUS-2, completed in 2013, again compared lifitegrast twice daily with placebo for 12\u00a0weeks in dry eye patients with a history of active artificial tear use \\[[@CR10]\\]. The primary endpoint of patient-reported symptoms was significantly better versus placebo, whereas there was no significant difference among the treatment and placebo groups in inferior corneal fluorescein staining score. OPUS-3, a third phase 3 trial, is in progress to further evaluate the efficacy and safety of lifitegrast. SONATA was a 1-year prospective study of lifitegrast versus placebo in 331 patients, with the goal to evaluate safety of long-term treatment \\[[@CR11]\\]. The protocol for this study, completed in 2014, also allowed for use of concomitant artificial tears and other adjunctive topical treatments. Overall the drug was found to be well tolerated, but of note, almost 50% of patients receiving lifitegrast experienced some mild treatment-related side effect, most commonly dysgeusia and instillation site irritation. There were no vision-threatening side effects. Future studies may investigate the potential role for lifitegrast in patients with blepharitis and other diseases of ocular surface inflammation. Overall, lifitegrast shows promise in improving dry eye and is pending review by the FDA for approval in late 2015.\n\nRebamipide: Quinolinone Derivative Mucin Secretogogue {#Sec5}\n-----------------------------------------------------\n\nRebamipide (Otsuka Pharmaceutical Co.; Tokyo, Japan) is a mucin secretogogue initially marketed in Japan for the treatment of gastric mucosal disorders and gastritis \\[[@CR12]--[@CR16]\\]. As decreased numbers of goblet cells have been observed in the conjunctiva and decreased mucin at the corneal surface has been observed in patients with dry eye, rebamipide was thus a potential target for dry eye treatment by stabilizing the mucin component of the tear film. Early studies demonstrated an increased production of periodic acid-Schiff-positive cells (goblet cells) when rebamipide was instilled into rabbit eyes and increased mucin-like glycoprotein when rebamipide was incubated with human corneal epithelial cells \\[[@CR17], [@CR18]\\]. Recent phase 2 and phase 3 studies have yielded promising results for clinical use of rebamipide in the treatment of dry eye \\[[@CR13], [@CR16]\\]. In the phase 3 trial, 188 patients with dry eye syndrome were randomized to 2% rebamipide four times daily or 0.1% sodium hyaluronate six times daily for 4\u00a0weeks. The main outcome measures included two primary end points of corneal fluorescein staining score and lissamine green conjunctival staining score, in addition to Schirmer's test, tear film breakup time, subjective report of foreign body sensation, dryness, pain and blurred vision, and patient's treatment impression scores. At 4\u00a0weeks, mean change from baseline in corneal fluorescein staining scores verified non-inferiority to sodium hyaluronate. Lissamine conjunctival staining scores demonstrated superiority of rebamipide. While there was not a significant difference between the two groups in Schirmer's test and tear breakup time, foreign body sensation and eye pain scores were significantly superior in the rebamipide group, with improved efficacy noted at the 2\u00a0week time point. Of note, patients' overall impression of the improvement in symptoms was significantly better in the rebamipide group as well. Adverse events were rare, and included bitter taste, eye pruritis, nasopharyngitis, headache, and decrease in white cell count (which was noted in three patients in the rebamipide group and none in the sodium hyaluronate group). The decrease in white blood cell count was not further characterized in this study, and the implications of this finding may require further investigation. Additionally, the short duration of this phase 3 study, only 4\u00a0weeks, means that further work will be needed to demonstrate the potential role of rebamipide in the treatment of chronic dry eye syndrome. This medication was launched in Japan for the treatment of dry eye in a 2% ophthalmic suspension in January 2012 and has yet to be approved for use in the USA.\n\nMiM-D3: Nerve Growth Factor Peptidomimetic, Mucin Secretogogue {#Sec6}\n--------------------------------------------------------------\n\nMIM-D3 (Mimetogen Pharmaceuticals; Gloucester, MA, USA) is a first in its class, small-molecule nerve growth factor (NGF) peptidomimetic that completed a phase 3 clinical trial for the treatment of dry eye in 2014 \\[[@CR19]--[@CR22]\\]. NGF plays a role in corneal wound healing and has previously been shown to have mucin secretogogue activity in conjunctival cells. Early studies in a rat model with scopolamine-induced dry eye demonstrated a favorable effect on glycoconjugate secretion with topical MIM-D3 1% \\[[@CR20]\\]. A subsequent phase 2 clinical trial enrolled 150 patients with dry eye who were randomized to 1% MIM-D3, 5% MIM-D3 or placebo, dosed twice daily for 28\u00a0days \\[[@CR19]\\]. This study, as well as the phase 3 trial, used the Controlled Adverse Environment (CAE) challenge to measure dry eye severity, where patients are subjected to an environment that exacerbates dry eye symptoms. Compared to placebo, fluorescein corneal staining post-CAE after 28\u00a0days of treatment with MIM-D3 was significantly improved in the 1% MIM-D3 group. Patients in the 5% MIM-D3 group showed better daily ocular dryness scores, and patients with higher symptom scores reported improvement of symptoms for both MIM-D3 doses. Adverse events were rare, including eye irritation and dellen formation, and systemic adverse events were not thought to be attributable to treatment. Limitations of the phase 2 study include the multiple post hoc analyses and short follow-up period. The subsequent phase 3 trial enrolled 403 patients with dry eye who received MIM-D3 1% or placebo twice daily for 28\u00a0days. Primary outcome measures of corneal fluorescein staining and ocular dryness as well as several secondary outcomes were examined, again in response to the CAE \\[[@CR23]\\]. The phase 3 study has not been published as yet; however, initial top-line results from Mimetogen demonstrated superiority over placebo in total corneal fluorescein staining. Notably, the mean blurred vision, reading, and TV-watching scores were improved with MIM-D3 compared with placebo \\[[@CR23]\\]. Adverse events were described as rare and transient.\n\nOTX-DP: Sustained Release Dexamethasone Loaded Punctal Plug 0.4\u00a0mg {#Sec7}\n------------------------------------------------------------------\n\nIn 2015, Ocular Therapeutix (Bedford, MA, USA) announced the outcomes of their second phase 3 clinical trial for OTX-DP, the sustained release dexamethasone-loaded punctal plug \\[[@CR24]\\]. The punctal plug dispenses a tapered release of dexamethasone onto the ocular surface over the course of (up to) 30\u00a0days after insertion. OTX-DP has primarily been investigated for treatment of inflammation and pain in the postoperative period after cataract surgery. An initial phase 2 trial demonstrated superiority over placebo for absence of anterior chamber cells at days 14 and 30 and for absence of pain at all measured time points through 30\u00a0days after cataract surgery \\[[@CR25]\\]. The plug was retained in all patients through day 14 and in about 97% of patients at 30\u00a0days. Importantly, the OTX-DP was also reported to be well tolerated by patients, and there were no significant elevations in intraocular pressure. The first phase 3 trial enrolled 247 patients undergoing cataract surgery in the US who were randomized to placement of the OTX-DP or standard, non-medicated vehicle punctal plug at the end of the surgical procedure. Outcome measures included anterior chamber inflammation and pain in the postoperative period. Patient-reported pain at day 8 and anterior chamber inflammation at day 14 after surgery were both statistically significantly better in the OTX-DP group versus placebo. In a second phase 3 trial enrolling 240 patients, however, a statistically significant decrease in patient-reported pain at day 8 after surgery was found in the OTX-DP group, but no significant difference in anterior chamber inflammation was detected at day 14. Further post hoc analysis is underway to investigate the differences in outcomes of these two trials. Potential advantages of OTX-DP include more reliable dosing of steroid and fewer self-administered medications for patients in the postoperative period. Important points for clinicians to consider in evaluating the upcoming results will include the relative risk of steroid-induced glaucoma with use of OTX-DP versus conventional topical steroid drops as well as reimbursement practices for insertion of the plug in the operating room following surgery.\n\nMore recently, OTX-DP has been investigated for use in the treatment of dry eye \\[[@CR26]\\]. In 2015, a phase 2 clinical trial enrolled 40 patients with dry eye disease. Patients will initially receive a vehicle placebo plug for 30\u00a0days, with patients continuing to exhibit symptoms then being randomized to OTX-DP or vehicle placebo plug. Primary outcome measures will include corneal and conjunctival staining, tear breakup time, and resorption of the plug following therapy. Of note, no comparison to conventionally used topical steroid has been done---therefore, it remains unclear whether the low, steady dosing of the OTX-DP is superior to a tapered drop regimen. Further evaluation will be needed to determine whether patients who demonstrate earlier plug resorption will show any decreased efficacy in treatment in either the postoperative or dry eye treatment scenarios and likewise whether the few patients who showed persistence of the plug at 2\u00a0months time are at higher risk for adverse events including steroid-induced glaucoma.\n\nEBI 005 (Eleven Biotherapeutics): Protein-Based IL-1 Inhibitor {#Sec8}\n--------------------------------------------------------------\n\nEBI 005, designed by Eleven Biotherapeutics (Cambridge, MA, USA), is the first protein-based IL-1 inhibitor designed for topical ophthalmic use and has completed phase 2 clinical testing \\[[@CR27], [@CR28]\\]. IL-1 is an inflammatory mediator of dry eye in addition to multiple systemic diseases. Notably it has been successfully targeted for the treatment of inflammatory conditions such as rheumatoid arthritis. A topically administered ocular therapeutic formulation IL-1 inhibitor was investigated in 2012, and the result, EBI-005, was found to bind its target, IL-1R1, with high specificity, which conferred a high in vivo potency in a therapeutic rat model \\[[@CR27]\\]. A phase 2 study was completed in 2015, investigating the use of EBI-005 5\u00a0mg/ml three times daily for the treatment of allergic conjunctivitis \\[[@CR28]\\]. One hundred fifty-six patients with moderate to severe allergic conjunctivitis were randomized to EBI-005 or vehicle groups and were subjected to aerosolized allergen challenge or direct conjunctival allergen challenge. In the conjunctival allergen challenge test, EBI-005-treated subjects showed statistically significant improvement in ocular itching compared to vehicle. Treated patients also showed greater improvement compared to vehicle in tearing and nasal symptoms. Of note, the primary pre-specified endpoint of ocular itching was not met in the aerosolized challenge test. Further phase 3 clinical testing is under way and will elucidate the role for EBI-005 in both allergic conjunctivitis and dry eye in the clinic.\n\nDiquafosol: P2Y~2~ Receptor Agonist {#Sec9}\n-----------------------------------\n\nDiquafosol is a purinergic agonist of the ocular surface P2Y~2~ receptor, which promotes fluid transfer and mucin secretion via a pathway involving the activation of phospholipase proteins \\[[@CR29]\\]. Diquafosol was approved in 2010 in Japan for use in treating dry eye, and it recently concluded a phase 3 study in the US \\[[@CR30]\\]. In earlier randomized clinical trials of this medication, mucin production and ocular surface damage appear to be improved, while effects on aqueous production are less certain \\[[@CR29], [@CR31], [@CR32]\\]. There have been no serious ocular side effects reported. Similar to other compounds for treating dry eye, heterogeneity in study design in evaluating clinical endpoints may have contributed to inconsistent findings among the several randomized trials that have been completed for this drug. Further larger trials with longer follow-up periods may better demonstrate its efficacy.\n\nPharmaceuticals in Phase 2 Clinical Trials {#Sec10}\n==========================================\n\nRU-101: Recombinant Human Serum Albumin {#Sec11}\n---------------------------------------\n\nPhase 1 and 2 clinical trials have been completed for RU-101 (R-Tech Ueno, Tokyo, Japan), a topically administered recombinant human serum albumin applied six times daily for the treatment of severe dry eye disease \\[[@CR33]\\]. The phase 1 trial investigated the safety and tolerability of escalating doses of RU-101 over the course of 4\u00a0weeks in patients with severe dry eye, and the phase 2 trial investigated the safety and efficacy of the RU-101 dosage established in phase 1 over the course of 12\u00a0weeks. The primary outcome measure was safety and secondary outcome measure change from baseline in dry eye symptoms, ocular surface disease index, visual acuity, tear breakup time, corneal fluorescein staining, corneal sensitivity and Schirmer testing \\[[@CR34]\\]. The corneal staining score at 12\u00a0weeks was chosen as the primary endpoint for effectiveness. Although there was significant improvement with treatment from baseline to 12\u00a0weeks, this was not significantly different from placebo. According to a press release from R-Tech Ueno, plans are underway to further investigate RU-101 in patients with severe dry eye, including dosing optimization. RU-101 or products like it, if effective in clinical trials, could replace human serum tears currently used to treat dry eye, eliminating the challenging process of obtaining these medications and avoiding the associated infection risk.\n\nKPI-121/LE-MMP 0.25%: Loteprednol Etabonate Mucus-Penetrating Particle, Glucocorticoid Receptor Agonist {#Sec12}\n-------------------------------------------------------------------------------------------------------\n\nKPI-121 (Kala Pharmaceuticals, Waltham, MA, USA) is a mucus-penetrating nanoparticle loteprednol etabonate 0.25% topical ophthalmic medication, which has recently completed phase 2 clinical testing in 2015 \\[[@CR35]\\]. The mucus-penetrating particle was designed to enhance delivery into ocular tissues by facilitating tear film mucus penetration \\[[@CR36]\\]. Similar strategies have been previously investigated to deliver medications across the cervical mucosa, gastrointestinal mucosa, and others \\[[@CR37]\\]. The 0.25% formulation is designed for use in dry eye and meibomian gland disease, whereas a 1% formulation is targeted for the treatment of postoperative inflammation. The phase 2 trial included 150 patients with dry eye who were randomized to 0.25% KPI-121 or vehicle placebo, dosed four times daily for 28\u00a0days. The primary clinical endpoints were bulbar conjunctival hyperemia and ocular discomfort. There was a statistical difference in conjunctival hyperemia favoring KPI-121, while ocular discomfort showed a favorable trend versus placebo but without a statistical difference. The authors noted that patients with worse baseline ocular discomfort showed a greater trend favoring outcomes with KPI-121 versus placebo. Instillation site pain was reported in 6.9% of patients receiving the medication versus 3.8% receiving placebo.\n\nDevices in Development for Dry Eye Therapy {#Sec13}\n==========================================\n\nOculeve Neurostimulator Device: Intranasal Lacrimal Stimulator for Dry Eye {#Sec14}\n--------------------------------------------------------------------------\n\nA pilot study has been completed for the Oculeve Neurostim Device (Oculeve, Inc, San Francisco, CA, USA), a lacrimal neural stimulator for the treatment of dry eye \\[[@CR38]\\]. The device, initially conceived by a fellow at the Stanford Biodesign program, is inserted into the mucous membrane in the nasal cavity. Tear production stimulation is then modulated by a wireless controller. The 2015 pilot study, for which results have not yet been published, evaluated the reduction in exacerbation of dry eye symptoms in 40 patients upon exposure to a controlled adverse environment (CAE) in patients with the neurostimulator in place.\n\nOcular Iontophoresis with EG-437 (40\u00a0mg/ml Dexamethasone Phosphate Solution) {#Sec15}\n----------------------------------------------------------------------------\n\nThe EyeGate^\u00ae^ II system (Eyegate Pharmaceuticals, Inc, Waltham, MA, USA) is an ocular iontophoriesis system designed to deliver drugs to the conjunctiva and sclera \\[[@CR39], [@CR40]\\]. In ocular iontophoresis, a small current is applied to the ocular surface creating an electrical field, which enhances the mobility of charged particles across the anterior and posterior segments. A drug delivered by ocular iontophoresis may achieve higher concentrations than it would via topical drop form. A recent study of patients with dry eye receiving delivery of EG-437 (40\u00a0mg/ml dexamethasone phosphate solution) via ocular iontophoresis found statistically significant improvement in signs and symptoms of dry eye in response to a controlled adverse environment challenge. However, the primary endpoints of corneal staining and ocular discomfort were not met in the phase II trial. It should be noted also that 87% of patients in the safety population had at least one adverse event, most commonly conjunctival hyperemia and keratitis, which required treatment \\[[@CR40]\\]. Some of the patients experiencing adverse events were in the placebo group who received ocular iontophoresis without medication delivery. No severe adverse events were reported. While the clinical results in terms of dry eye endpoints are interesting, future studies and possibly refinements to mitigate the percentage of patients who experience non-serious adverse events with this device may be needed to make it a success in the clinical setting.\n\nConclusion {#Sec16}\n==========\n\nWe are at an exciting crossroads with respect to dry eye treatment. Multiple pharmaceutical agents are moving toward the market for the treatment of dry eye syndrome. Topical agents range from twice daily to four times daily administration, and phase 3 trials have shown varied degrees of improvement in subjective and objective signs of dry eye syndrome. One important remaining question is, given the chronic nature of the disease, how well will these medications be tolerated when taken over a long time period and will their effectiveness remain the same when taken chronically? In addition to studies of potential long-term local side effects, whether there are any systemic side effects of the immunomodulatory medications remains an important question, despite their low levels of systemic absorption. This remains to be shown through longer-term studies following phase 3 clinical trials.\n\nInflammation is a common feature in multiple ocular surface diseases in addition to dry eye, including blepharitis, allergic conjunctivitis, and others. Thus, agents designed to treat dry eye may have cross-over benefit to those suffering from other ocular surface conditions, and vice versa. Further study of some of these medications for patients with other inflammatory diseases is merited and in some cases is already underway. Patients with more than one inflammatory ocular surface disease (as is often the case) represent an important cohort to study as well. The future is bright as medications and devices continue to flow into the pipeline. Dry eye sufferers and physicians alike can only hope that in the decade to come we will have many more approved therapeutics available to treat dry eye.\n\nNo funding or sponsorship was received for this study or publication of this article.\n\nAll named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this manuscript, take responsibility for the integrity of the work as a whole, and have given final approval for the version to be published.\n\nConflict of interest {#d30e420}\n====================\n\nP. K. Gupta is a consultant for AMO, Shire, Allergan, Tear Science, Novabay, and BioTissue, all of which have products related to the treatment of dry eye. L. A. Vickers declares no conflict of interest.\n\nCompliance with ethics guidelines {#d30e425}\n=================================\n\nThis article is based on previously conducted studies and does not involve any new studies of human or animal subjects performed by any of the authors.\n\nOpen Access {#d30e430}\n===========\n\nThis article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (<http://creativecommons.org/licenses/by-nc/4.0/>), which permits any noncommercial use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.\n"}
+{"text": "Introduction {#sec1}\n============\n\nMicroneedle systems are increasingly being heralded as a technological step change in drug delivery applications and can offer a multitude of advantages over conventional syringe-based approaches.^[@ref1]\u2212[@ref9]^ These microneedle (MN) systems typically comprise an array of sub-millimeter sized projections (50--900 \u03bcm) and, in contrast to conventional hypodermic injections, are sufficiently small that the shallow penetration depth typically fails to trigger the dermal nerve network.^[@ref9]\u2212[@ref11]^ The near-painless puncture of the skin barrier allows the transport of a large variety of drugs and vaccines to the underlying microcirculation, and the success of the general strategy has seen a near exponential rise in publications.^[@ref3]^ There are generally five classes of MN system: solid, coated, hollow, dissolvable, and swellable. Each has its own merits and limitations, and they have been extensively reviewed.^[@ref1]\u2212[@ref9]^ As the development of MN patches has continued apace, there has been a trickle-down availability of the technologies required to produce them with silicone molds allowing rapid, low cost, soft lithographic production within conventional laboratory environments.^[@ref1],[@ref2],[@ref12],[@ref13]^ The processes can be adapted for preparing all bar the hollow designs. The aim of the present communication has been to investigate the development of composite solid microneedles consisting of polymer encapsulated nanocarbon. The intention was to create electrochemically conductive MN arrays that could, in principle, offer the potential for electrochemical techniques to be applied such that the MN patches could possess sensing capabilities but, more importantly, could also be used to control drug release. The latter sits at the core of the approach, as the electronic control over the release mechanism could offer a new approach in MN design.\n\nThe rationale adopted here revolves around the use of cellulose acetate phthalate (CAP) as the binding polymer used to maintain the integrity of the needle structure. In contrast to more benign polymer systems used in MN manufacture (i.e., polycarbonate or polystyrene^[@ref12],[@ref13]^), CAP is pH sensitive and is already used in a multitude of oral drug formulations.^[@ref14]\u2212[@ref18]^ The CAP outer layer is stable in the acid of the stomach but dissolves upon reaching the alkaline environments of the colon.^[@ref16],[@ref17]^ Exploitation of the pH sensitivity of the polymer for use in electrochemically controlled drug release has been described previously but in the context of its use as a film encapsulating a drug loaded reservoir.^[@ref19]^ The intention here was to use the CAP matrix as the binding medium within which nanocarbon particles would be mixed to yield a composite microneedle but which could swell or dissolve upon changes in local pH brought about by the imposition of a suitably reducing potential. The latter relates to the hydrogen evolution reaction (HER), whereby there is an increase in local pH at the electrode as a consequence of the reduction process.^[@ref19]^ It could be envisaged therefore that, were the HER reaction to be imposed at the MN array, the increase in local pH would lead to the swelling (and possible dissolution) of the CAP polymer that constitutes the needle structure. Thus, were drugs to be entrapped within the network at the time of formulation, then the electrochemically induced swelling could allow control over their release. The ultimate embodiment of the proposed electrochemically driven strategy is highlighted in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}. There is a burgeoning interest in the use of dissolvable and swellable MN systems for drug release,^[@ref1]\u2212[@ref9]^ and although the formulation process is similar to that proposed here, the release process for those is inherently passive. It was therefore of interest to determine whether the introduction of the nanocarbon particles could facilitate the development of a controlled release mechanism and to critically evaluate the merits and limitations that such systems would present for transdermal drug delivery.\n\n![Overview of the proposed, electrochemically initiated release mechanism](am9b09674_0001){#fig1}\n\nMaterials and Methods {#sec2}\n=====================\n\nElectrochemical analysis was performed utilizing a \u03bcAutolab Type III computer operated potentiostat (Eco-Chemie, Utrecht, The Netherlands). All measurements were conducted at 22 \u00b1 2 \u00b0C. Nanocarbon powder was supplied by Sigma-Aldrich with a mean particle size of 100 nm. Silicone MPatch microneedle templates were purchased from Micropoint Technologies Pte Ltd. (CleanTech Loop, Singapore). The initial measurements involved a three-electrode configuration consisting of a carbon--polymer composite MN working electrode, a counter electrode in the form of a platinum wire, and a standard silver/silver chloride (3 M NaCl, BAS Technicol UK) reference electrode.\n\nMicroneedle Fabrication {#sec2.1}\n-----------------------\n\nProduction of the MNs was generally achieved through the simple mixing and dispersion of the nanocarbon (Sigma-Aldrich) within the polymer (50/50 wt %) that had previously been dissolved in a suitable solvent such as cyclohexanone. The silicone molds were obtained from Micropoint Technologies and typically consisted of a 10 \u00d7 10 MN array. The needle dimensions used in the studies were either 200 \u00d7 200 \u00d7 350 \u03bcm or 200 \u00d7 200 \u00d7 700 \u03bcm. The molds were cleaned by sonication, the polymer--carbon mixture was added, and the solvent was allowed to evaporate. The use of a vacuum oven (under ambient temperature) was found to greatly improve the speed of the production process and needle quality through accelerating the removal of trapped solvent. In the case of microneedle arrays loaded with Toluidine Blue (TBO), the silicone molds were initially coated with carbon--CAP solution to yield a coherent but conductive interface. A second mixture of the carbon--CAP containing 1% TBO was then introduced, and the needles were cured as specified previously.\n\nMicroneedle Modification {#sec2.2}\n------------------------\n\nThe nanocarbon--CAP microneedles were sputtered with a thin layer of palladium using a 80:20 Pd/Au target at 30 mA for 3 min (Emitech K500X Sputter Coater, Quorum Technologies Ltd., England). X-ray photoelectron spectroscopy (XPS) of the palladium samples before and after modification with cysteine was performed using an Axis Ultra DLD spectrometer (Kratos Analytical, Japan) using monochromated Al K\u03b1 X-rays (15 kV and 10 mA) with an operating pressure lower than 6 \u00d7 10^--8^ Pa. A hybrid lens mode was used during analysis, and charge neutralization was achieved using an immersion lens with a filament current of between 1.7 and 2.1 mA at a charge balance voltage of between 3.0 and 3.6 V. Three spots were analyzed per sample with wide energy survey scans (0--1300 eV binding energy) as well as high resolution spectra for Pd 3d, C 1s, N 1s, O 1s, and S 2p. The pass energy was 160 eV for the wide energy survey scans and 20 eV for the high resolution spectra.\n\nBiocompatibility Studies {#sec2.3}\n------------------------\n\nThe biocompatibility of the nanocarbon--cellulose acetate composite was assessed in relation to skin irritation (DIN EN ISO 10993-10:2014) and cytotoxicity (DIN EN ISO 10993-5:2009) and conducted by Bioserv Analytik Un Medizinprodukte GMBH (Rostock Germany) under GLP conditions.\n\nSkin irritation assessments were conducted using healthy young female albino rabbits with a weight not less than 2 kg (as per ISO recommendations). The rabbits were kept caged for at least 5 days prior to the test to enable acclimatization and were immunized against myxomatosis and RHD. The fur on the back of the rabbits was closely clipped on both sides of the spinal column (10 \u00d7 15 cm) 4 h before the test procedure was initiated. The test material (C--CAP) was applied directly to the clipped skin along with gauze patches that served as a control (25 \u00d7 25 mm). The sites were then covered with a nonocclusive gauze patch and then wrapped with an occlusive bandage for a period of 4 h. At the end of the test, the dressings were removed, and any residual substances were washed with warm water and the skin blown dry. The application sites were typically monitored at 1, 24, 48, and 72 h after removal of the material and scored in terms of extent of erythema, eschar, and edema formation. The results indicated that there was no skin irritation at any point in the course of the 72 h observation.\n\nCytotoxicity studies were conducted using 6 cm^2^ samples of the C--CAP material. Polypropylene and DMSO were used as the negative and positive control samples, respectively. Extracts were prepared in accordance with ISO 10993-12:2012 using Dulbecco's Modified Eagle Medium with 10% fetal calf serum (DMEM--FCS). The extraction process was run with gentle shaking for 24 h at 37 \u00b0C. Cell cultures were prepared using L929 cells (ATCC CCL 1, NCTC clone 929, connective tissue mouse) as per recommendations in 10993-5:2009. Cells were grown in DMEM--FCS at 37 \u00b0C and 5% CO~2~ in a humidified incubator. Cells were harvested 24 h before determination of cytotoxicity using a trypsin/EDTA solution and resuspended in fresh DMEM--FCS. The cell density was adjusted to 1.75 \u00d7 10^5^ cells/mL. The wells of a tissue culture plate were then inoculated with 1 mL of the cell suspension and incubated for 24 h, during which the cells formed a subconfluent monolayer. Serial dilutions of the C--CAP extract were prepared to give concentrations of 100, 66, 44, 30, and 20% using DMEM--FCS as diluent. Each dilution extract was then tested through triplicate pipetting of 1 mL aliquots into the respective cell culture wells (after removal of the culture medium). The well plate was then incubated for 24 h prior to assessment. The cell culture plates were examined microscopically and graded according to their reactivity (0: no growth inhibition/no cell lysis through to 4: almost complete destruction of the cell layers). The cell layers were also quantitatively assessed through staining with 0.25% crystal violet, washed, and dried, and the cell bound stain was extracted with 33% glacial acetic acid. The dissolved stain samples were read by a microplate reader at 550 nm. The absorbance for each sample was determined in triplicate with the mean value of the negative polypropylene estimated at 100% cell growth. The relative inhibition of cell growth (ICG) was calculated as %ICG = 100 -- (100 \u00b7 *A*~550\u00a0test~/*A*~550\u00a0negative\u00a0control~), whereby, in accordance with ISO 10993-5, an ICG of more than 30% is regarded as a cytotoxic effect. The extract solutions all fell below this threshold and along with the overall assessment indicated that the material did not cause any relevant toxicological or biological damage to the subconfluent monolayer of L929 cells under the test conditions of DIN EN ISO 10993-5:2009.\n\nResults and Discussion {#sec3}\n======================\n\nScanning electron micrographs detailing the structure of an MN array (height: 350 \u03bcm) cast from a solution of CAP without any additional components are highlighted in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}A. Well-defined peaks were obtained and were found to be stable in solutions where the pH is acidic/neutral. Upon exposing the CAP MN array to pH 8 buffer, the needle definition deteriorates and gradually dissolves as indicated in the electron micrographs ([Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}B--D) recorded after 1, 3, and 5 min intervals, respectively.\n\n![Scanning electron micrographs of the cellulose acetate phthalate microneedles upon exposure to pH 8 Britton--Robinson buffer. Recorded at (A) 0, (B) 1, (C) 3, and (D) 5 min. Microneedles: 200 \u00d7 200 \u00d7 350 \u03bcm.](am9b09674_0002){#fig2}\n\nThe process was repeated using the nanocarbon--CAP (C--CAP) formulation, and representative electron micrographs recorded after 1 min in pH 8 BR buffer are detailed in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}A--C. The initial morphology of the composite needle ([Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}A) is relatively smooth with no pronounced defects or granularity. The latter could have been expected given the particulate nature of the carbon, and such features have been previously observed with palladium systems.^[@ref13]^ Upon being immersed in the pH 8 buffer, the CAP layer begins to dissolve ([Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}B), and while the needle framework is retained, the outer surface begins to transform. The dissolution of the CAP leaves the residual carbon substructure exposing the platelet-like carbon formation ([Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}C). As the needles are left in contact with the buffer, dissolution of the cap continues and compromises the integrity of the structure, leading to sustained erosion and disappearance of the needles---akin to the behavior observed with the pure cap systems ([Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}B,C)\n\n![Scanning electron micrographs of the C--CAP microneedles before (A) and after (B) 1 min of exposure to pH 8 Britton--Robinson buffer. Dissolution of the CAP polymer leaving carbon platelets is highlighted in (C). Scanning electron micrographs (D,E) and computerized tomography (F) of microneedles' piercing of needles through defleshed tomato skin. Microneedle array: 200 \u00d7 200 \u00d7 700 \u03bcm.](am9b09674_0003){#fig3}\n\nThe ability of the needles to pierce and retain integrity was assessed through electron microscopy and computerized tomography (CT) of MN arrays puncturing tomato skin. Representative images are shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}D--F. The tomato flesh was removed from the cuticle in order to enable inspection of the needles post pierce with the web-like indentations on the skin's inner surface ([Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}D,E), characteristic of cell wall structures that remain strongly adherent to the cuticle. The ability to section through the microneedle array using the CT scans also confirmed that the needle structures were solid and free of fissures or voids.\n\nCyclic voltammograms detailing the response of the carbon--CAP MN array (200 \u00d7 200 \u00d7 350 \u03bcm) to ferrocyanide and ruthenium hexamine (individual solutions containing 2 mM redox probe, 0.1 M KCl, 50 mV/s) are detailed in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}A,B, respectively. In both cases, well-defined peaks are obtained, although there is marked deviation from the ideal peak separations of 59 mV. This can be attributed to the composite nature of the MN array; similar behavior has been observed with the composite Pd--polystyrene microneedle.^[@ref12]^ Nevertheless, the ability of the needles to function as viable electrochemical sensors is apparent.\n\n![Cyclic voltammograms detailing the response of the C--CAP microneedles to ferrocyanide (A) and ruthenium hexamine (B). Each redox probe present at 2 mM in 0.1 M KCl. Scan rate: 50 mV/s.](am9b09674_0004){#fig4}\n\nElectrochemically Induced Dissolution of CAP {#sec4}\n--------------------------------------------\n\nPalladium is widely used to enhance the HER process^[@ref20],[@ref21]^ and has been employed in the form of clusters and coatings with graphite,^[@ref22]\u2212[@ref24]^ carbon nanotubes,^[@ref25]^ graphene,^[@ref26],[@ref27]^ molybdenum nanosystems,^[@ref22],[@ref28],[@ref29]^ and various metal nanoparticle systems^[@ref30]\u2212[@ref32]^ for use in fuel cell applications. In contrast to most metals and alloys, the dissociation of H~2~ molecules within Pd structures is facile, occurs with almost no activation barrier, and serves as an ideal catalyst for hydrogen sorption and desorption.^[@ref33],[@ref34]^ The core approach here, however, is not to utilize the hydrogen being produced but rather to exploit the change in local pH that arises as a consequence of the electrolysis. Given that carbon is a relatively poor substrate for the HER process, it was assumed that the presence of the metallic Pd on the surface would enable a much more effective response. A thin layer of palladium was sputtered on the C--CAP microneedles to assess whether or not the presence of the metal would improve the overall performance---both in terms of sensing application but, more importantly, as a means of enhancing the HER process, which would be required to promote swelling of the needles as indicated in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}. The influence of the Pd layer on the electrode response to ferrocyanide is highlighted in [Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}A. The addition of the cysteine onto Pd layers was confirmed through XPS analysis and comparison of the S 2p peak before and after modification as indicated in [Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}B.\n\n![(A) Cyclic voltammograms comparing the responses of the C--CAP and C--CAP--Pd microneedles to 2 mM ferrocyanide before and after modification with palladium and cysteine. Scan rate: 50 mV/s. (B) XPS spectra highlighting the modification of C--CAP--Pd microneedles with cysteine.](am9b09674_0005){#fig5}\n\nThe Pd layer was found to improve the electrochemical performance through increasing the electron transfer kinetics as indicated by the improvement in the peak definition and separation; however, the greatest gain in performance was achieved through the further modification of the surface with an adsorbed cysteine layer as described by McFie and Feliciano-Ramos and associated colleagues.^[@ref35],[@ref36]^ The latter has been attributed to the protonated amino group of the cysteine improving electron transfer from the negatively charged ferrocyanide,^[@ref36]^ but it must be noted that modification with 2-mercaptoethansulfonate (presenting a net negative charge at the interface) was equally capable of facilitating ferrocyanide electrochemistry.^[@ref35]^ It is likely that the cysteine promoted electrode enhancement is through the removal of the surface oxides at the Pd interface.^[@ref35]^\n\nThe responses detailed in [Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}A highlight the benefits of employing Pd--cysteine layers for sensing purposes---where the microneedle array is to be used as both sensor and actuator. It must be noted however that the imposition of the large reducing potential necessary to initiate HER (and hence drug release) would result in the reduction the Pd--thiol bond, leading to the desorption of the cysteine^[@ref37]^ such that the latter has little influence on the kinetics of drug release. As such, the cysteine modification step was omitted from the subsequent investigations of microneedle dissolution. Linear sweep voltammograms detailing the response of the MN arrays (C--CAP and C--CAP--Pd) are compared in [Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"} under two different pH regimes. It is clear that, in both cases, the presence of the Pd has a dramatic impact on the relative responses.\n\n![Linear sweep voltammograms of the unmodified and Pd coated microneedles in pH 3 and 7 Britton--Robinson buffer. Scan rate: 50 mV/s.](am9b09674_0006){#fig6}\n\nThe next phase was to assess the ability to control the swelling or dissolution of the microneedles. A skin mimic assembly was constructed, in which a calcium alginate gel containing 2 mM ferrocyanide was prepared as previously described.^[@ref12]^ The pH was controlled to ensure that the matrix itself did not lead to the dissolution of the needles. A thin layer of parafilm was then stretched over the gel to act as the skin's stratum corneum. The modified microneedles were then pressed onto the parafilm layer (100 \u03bcm) through simple application of thumb pressure with the needle tips piercing through the polymer into the underlying gel. A separate reference electrode (3 M NaCl, Ag\\|AgCl) and Pt counter were then inserted through the parafilm to complete the cell. Ferrocyanide was included within the gel to acts as an in situ probe that could be used to assess the structural integrity of the microneedle. Voltammograms were recorded before the application of any cathodic potential to serve as a control. It was envisaged that as the imposition of a cathodic potential would increase the pH and lead to the swelling or dissolution of the CAP binder within the core of the needle framework, and as such, the surface area of the needles would change. The latter could therefore be assessed through changes in the peak magnitude of the ferrocyanide within the gel. Thus, the ferrocyanide voltammograms were recorded before and after each cathodic step. Cyclic voltammograms detailing the response of the MN array within the alginate layer before and after the imposition of a potential of \u22122 V are shown in [Figure [7](#fig7){ref-type=\"fig\"}](#fig7){ref-type=\"fig\"}. Before the cathodic potential is applied, the voltammetric response of the MN array to ferrocyanide is consistent with that observed in [Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}A.\n\n![Cyclic voltammograms detailing the response of the C--CAP--Pd microneedles to 2 mM ferrocyanide before and after holding the electrode at \u22122 V for given time periods.](am9b09674_0007){#fig7}\n\nThe voltammetric profile changes dramatically upon stimulating hydrogen evolution with a loss of definition. The latter can be attributed to increased resistance within the bulk of the needle structure as a consequence of the binder swelling and increasing the spatial separation between the carbon particles. The swelling and dissolution of the MN were confirmed through placing the array within pH 6 Britton--Robinson buffer (without a barrier layer) and examining the needle morphology after various periods of imposing the cathodic potential (\u22121 and \u22122 V). Electron micrographs comparing needle integrity before and after the onset of hydrogen evolution are detailed in [Figure [8](#fig8){ref-type=\"fig\"}](#fig8){ref-type=\"fig\"}A--C. The electron micrographs highlight the gradual dissolution of the microneedle structure as the electrode is held at varying potentials (\u22121 and \u22122 V) for a period of 30 s. The dissolution of the CAP polymer is noticeable at \u22121 V with the structure of the needle being analogous to that observed in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}B. The application of \u22122 V however leads to the near complete removal of the needle after 30 s. However, it is clear from inspecting the cyclic voltammograms ([Figure [7](#fig7){ref-type=\"fig\"}](#fig7){ref-type=\"fig\"}) that the resistance increases dramatically upon imposing the reducing potential for very short periods, indicating that swelling occurs relatively readily, but complete dissolution (and degradation of the needle bulk) requires much more forceful conditions. The continued response observed in [Figure [7](#fig7){ref-type=\"fig\"}](#fig7){ref-type=\"fig\"} after the apparent destruction of the needle (as indicated in [Figure [8](#fig8){ref-type=\"fig\"}](#fig8){ref-type=\"fig\"}C) is attributed to the continuing electroactivity of the base plate.\n\n![Scanning electron micrographs of the C--CAP--Pd microneedles after the electrode is poised at a reducing potential. (A) Open circuit. (B) \u22121 V for 30 s. (C) \u22122 V for 30 s. Microneedle: 200 \u00d7 200 \u00d7 700 \u03bcm.](am9b09674_0008){#fig8}\n\nElectrochemical Release of Model Drug {#sec5}\n-------------------------------------\n\nThe efficacy of the electrochemical release method was assessed with a model drug using a gelatin matrix to mimic transport into the skin. The pH of the gelatin was adjusted to pH 4.02, which would prevent the dissolution of the CAP and retain the integrity of the composite microneedle assembly. TBO was chosen as the model drug agent, as it could be easily monitored through visual inspection through a blue coloration (\u03bb~max~ = 630 nm) and could be readily incorporated into the C--CAP formulation. The C--CAP was added to a solution of TBO (5%) in cyclohexanone and stirred until a viscous solution was achieved. It was envisaged that the imposition of the reducing potential at the microneedle array would initiate the HER process and lead to an increase in the local pH. This would induce the dissolution of the CAP, resulting in the release of TBO into the gelatin. To ensure the TBO was released exclusively from the needles, the MN array was first pierced through a layer of parafilm, which acted as a skin mimic, before being placed in the gel in order to separate the baseplate from the gelatin. The TBO containing MN was washed in acid to remove extraneous TBO and then pierced into the gelatin layer and left to sit. No blue coloration was observed, which confirmed that the TBO was entrapped within the MN structure. The effect of imposing the reducing potential is highlighted in [Figure [9](#fig9){ref-type=\"fig\"}](#fig9){ref-type=\"fig\"} where the blue coloration associated with TBO can be seen to develop with duration of the HER electrolysis.\n\n![A C--CAP--Pd MN array loaded with TBO and pierced into gelatin (pH 4.02). The imposition of a potential of \u22122 V over a period of 100 s is shown to affect the release of TBO (blue coloration). Microneedle: 200 \u00d7 200 \u00d7 700 \u03bcm.](am9b09674_0009){#fig9}\n\nThe release of TBO within the gelatin was induced through the application of a large reducing potential (\u22122 V), which has the effect of rapidly changing the pH at the electrode surface and speeds the dissolution of the CAP component. The prime disadvantage of this approach relates to the fact that the integrity of the needles is rapidly compromised ([Figure [8](#fig8){ref-type=\"fig\"}](#fig8){ref-type=\"fig\"}C). A more subtle approach, involving less negative potentials, can also be used (as previously highlighted in [Figure [8](#fig8){ref-type=\"fig\"}](#fig8){ref-type=\"fig\"}B). The release of the TBO model drug was further investigated through employing a \u22121 V release step applied repetitively over a period of 40 min in pH 5 buffer. The microneedle array was held for 3 min at the reducing potential, and thereafter, the solution was sampled and analyzed using conventional colorimetry (TBO \u03bb~max~ = 632 nm; \u03b5 = 30\u202f000 M^--1^ cm^--1^).^[@ref38]^ A control microneedle array containing TBO was left in contact with the buffer without any applied potential, and the results are detailed in [Figure [10](#fig10){ref-type=\"fig\"}](#fig10){ref-type=\"fig\"}A. In the absence of an applied potential, there is effectively no release of TBO, whereas the repeated application of \u22121 V results in sustained release of the model drug. The typical yield per cycle is highlighted in [Figure [10](#fig10){ref-type=\"fig\"}](#fig10){ref-type=\"fig\"}B and increases with each cycle before falling dramatically to zero---at which point the needle has effectively dissolved. The yield of drug fall within the nanomole region is as expected given the relatively low capacity of the needles.\n\n![Effect of repetitive application of a potential (\u22121 V, 300 s) on the release of TBO into pH 5 buffer (A) and the corresponding yield per cycle (B). Microneedle: 200 \u00d7 200 \u00d7 700 \u03bcm.](am9b09674_0010){#fig10}\n\nThe micromolding technique clearly provides a facile means of producing high quality microneedles, which, through judicious selection of the casting components, could be applied to a wealth of sensing and drug delivery applications.^[@ref1]\u2212[@ref9]^ The nanocarbon systems are conductive and exhibit reasonable electrochemical properties, which could be harnessed in analytical scenarios. The modification with Pd is easily achieved through employing standard techniques used for coating SEM samples, yet the presence of the metal can provide significant enhancement to the voltammetric performance. The main impact of the work described however relates to the ability to control the integrity of the microneedle through controlling the potential applied at the needles themselves. It is easy to envisage the incorporation of a suitable drug candidate within the carbon--CAP matrix at the time of casting. Drug yield will however be an issue, as it is for many microneedle systems, and it is inevitable that the system proposed here would only be viable for low yield, high potency agents.^[@ref1]\u2212[@ref9]^ The imposition of the reducing potential, although necessary to induce swelling, could also, in principle, lead to the reduction of functional groups within the drug, such as nitro groups, and care would be required to ensure there were no inadvertent modifications to the therapeutic agent as a consequence of the release mechanism. The possibility of fragmentation and the loss of carbon/Pd particulates must also be considered. It must be noted however that, providing the needles are sufficiently shallow, residual needle fragments are liable to be expunged from the skin through normal skin turnover in the outer layers, which would normally occur over a period of weeks.^[@ref39]^ The biocompatibility studies of the C--CAP material provide some positive insights into the potential use of the material with no apparent skin irritation nor cytotoxicity; however, these need to be viewed with some caution---particularly where the devices may be used for long-term applications where sensitization may occur.\n\nConclusions {#sec6}\n===========\n\nThe electrode potentials used in this work to effect the swelling and dissolution are significantly large and have been selected on the basis of enabling gross characteristics to be observed with relative ease for the purpose of confirming proof of concept. The use of less negative potentials would necessarily reduce the degree of pH modification and, thus, rather than having fairly rapid release, could enable a much more metered dosing without the dramatic deterioration of the needle structure. This would mitigate against issues associated with the loss of nanocarbon/Pd particles into the skin. The approach presented highlights a new route through which microneedle structure could be controlled in situ and thereby offer alternative means of controlling dosage.\n\nThe manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. The authors contributed equally.\n\nThe authors declare no competing financial interest.\n\nThe authors are pleased to acknowledge financial support from the European Union's INTERREG VA Programme, managed by the Special EU Programmes Body (SEUPB) and the Department for the Economy (DfE) Northern Ireland.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nVisuo-motor adaptation to rightward prismatic shifts reduces signs of left spatial neglect on a wide range of measures ([@bib11; @bib2; @bib6; @bib9; @bib13]). As there are hemispheric asymmetries in spatial attention mechanisms, it may be useful to examine whether prism adaptation can produce similar improvements in neglect of the *right* hemispace following left hemisphere damage. We report improvement in a patient with mild right spatial neglect following adaptation to leftward-shifting prisms.\n\n2. Participants and methods {#sec2}\n===========================\n\n2.1. Participants {#sec2.1}\n-----------------\n\nParticipants were one patient with right spatial neglect (female, age\u00a0=\u00a075 years, left-handed) and eight right-handed, neurologically healthy age- and gender-matched control participants (Mean age\u00a0=\u00a073.4 years, Standard Error of the Mean -- SEM\u00a0=\u00a0.82).\n\nPatient DS was hospitalised with unintelligible speech, left gaze deviation, right neglect, right facial weakness and hemiplegia with brisk reflexes on the right and bilateral Babinski signs. A Computerised Tomography (CT) scan revealed a large left fronto-parietal haematoma due to haemorrhagic infarction.\n\nThree months later, at the time of the current investigation, DS was referred to us after her speech therapist noticed a tendency to leave the rightmost part of her workbooks uncompleted. At this time she had anomic aphasia with impaired repetition; however, her comprehension was relatively preserved and judged sufficient to enable informed consent. She had dense right hemiplegia and completed pen and paper tests for neglect with her left hand. She failed to copy the rightmost detail of a simple scene ([@bib8]); showed a mean 8.9% leftward line bisection bias on three lines ([@bib14]); and failed to cancel 2--3 rightmost targets on three cancellation tests ([@bib4; @bib5]). A Magnetic Resonance Imaging (MRI) scan revealed a large left hemisphere lesion involving the frontal eye field, motor and premotor cortices, cingulate gyrus, posterior superior temporal gyrus and the parietal lobe (including the superior parietal lobule, precuneus, angular and supramarginal gyri).\n\n2.2. Design and procedure {#sec2.2}\n-------------------------\n\nThe effects of adaptation to leftward-shifting prisms on the neglect symptoms of DS were examined using a multiple baselines design, with eight testing sessions spanning eighteen days. Effects of both sham treatment (day 2) and prism treatment (day 7) were examined. In these sessions neglect was assessed using the Ogden copying task ([@bib8]), the Bells cancellation task ([@bib5]), and a modified version of the line bisection subtest of the Behavioral Inattention Test ([@bib14]). Performance on the copying and cancellation tasks were at ceiling, therefore analyses focused on the results of the line bisection test. In each session DS bisected twelve 203\u00a0mm horizontal lines positioned to the left, middle or right of 4 sheets of A4. Bisection deviations from veridical were expressed as a percentage of line length, with negative values indicating a leftward bisection bias (right-sided neglect).\n\nComparison bisection data were collected from the healthy participants in three sessions completed on one day: at baseline, after sham adaptation, and after prism adaptation. Like DS, control participants used their left hands for all tasks.\n\nDuring prism adaptation participants wore goggles containing adjustable wedge prism lenses that were set to induce no shift ('sham' treatment) or a 15\u00b0 leftward visual shift ('prism' treatment). They made fifty pointing movements with their left hand, alternating between two targets positioned at eye level and arm\\'s length 10\u00b0 to the left and right of straight ahead. The goggles restricted the visual field such that participants received visual feedback of the second half of their pointing movement only. After touching each target, they returned their hand to the surface of the table in front of them.\n\nAdaptation was confirmed by measuring an after-effect: Participants pointed to three targets located straight ahead and 10\u00b0 to the left and right of body midline in a pre-determined pseudorandom order. A panel positioned under the chin occluded vision of the pointing arm. Pointing error was measured in degrees with the aid of markings on the underside of the panel, with negative numbers indicating leftward deviation. Twelve pointing measurements were taken in four sessions for DS (pre-sham, post-sham, pre-prism and post-prism), and in three sessions for the control participants (at baseline, post-sham and post-prisms).\n\n3. Results {#sec3}\n==========\n\n3.1. Prism adaptation {#sec3.1}\n---------------------\n\n### 3.1.1. Healthy controls {#sec3.1.1}\n\nA repeated-measures Analysis of Variance (ANOVA) of pointing error with one factor, session (baseline, post-sham, and post-prism), revealed a significant main effect \\[*F*(2,14)\u00a0=\u00a014.45, *p*\u00a0\\<\u00a0.001\\] where mean pointing error was unchanged between the baseline (*M*\u00a0=\u00a0.9\u00b0, SEM\u00a0=\u00a0.67) and post-sham session \\[*M*\u00a0=\u00a01.9\u00b0, SEM\u00a0=\u00a0.72; *t*(7)\u00a0=\u00a01.12, *p*\u00a0=\u00a0.26\\] but shifted significantly rightward after prism adaptation (*M*\u00a0=\u00a04.9\u00b0, SEM\u00a0=\u00a0.55) compared to both baseline \\[*t*(7)\u00a0=\u00a05.37, *p*\u00a0\\<\u00a0.005\\] and post-sham \\[*t*(7)\u00a0=\u00a03.90, *p*\u00a0\\<\u00a0.01\\] pointing.\n\n### 3.1.2. DS {#sec3.1.2}\n\nA repeated-measures ANOVA of pointing error with two factors, treatment (sham and prism)\u00a0\u00d7\u00a0session (pre and post), revealed a significant two-way interaction \\[*F*(1,11)\u00a0=\u00a021.9, *p*\u00a0\\<\u00a0.001\\] where pointing error was unchanged following sham treatment \\[Pre: *M*\u00a0=\u00a03.3\u00b0, SEM\u00a0=\u00a0.28; Post: *M*\u00a0=\u00a04.8\u00b0, SEM\u00a0=\u00a0.70; *t*(11)\u00a0=\u00a02.02, *p*\u00a0=\u00a0.07\\] but shifted 4.9\u00b0 rightward following prism treatment \\[Pre: *M*\u00a0=\u00a05.6\u00b0, SEM\u00a0=\u00a0.55; Post: *M*\u00a0=\u00a010.5\u00b0, SEM\u00a0=\u00a0.35; *t*(11)\u00a0=\u00a08.48, *p*\u00a0\\<\u00a0.001\\].\n\n3.2. Line bisection {#sec3.2}\n-------------------\n\n### 3.2.1. Healthy controls {#sec3.2.1}\n\nThere was a decrease in bisection errors across the baseline (*M*\u00a0=\u00a0\u22124.68, SEM\u00a0=\u00a01.53), post-sham (*M*\u00a0=\u00a0\u22123.66, SEM\u00a0=\u00a01.49) and\u00a0post-prism sessions (*M*\u00a0=\u00a0\u22123.11, SEM\u00a0=\u00a01.44), although a one-way ANOVA revealed no main effect \\[*F*(2,14)\u00a0=\u00a01.63, *p*\u00a0=\u00a0.23\\]. Bisection errors were pooled across sessions and a 95% confidence interval around the mean confirmed a significant leftward bias \\['pseudoneglect'; CI~.95~\u00a0=\u00a0(\u22125.54, \u22122.10)\\].\n\n### 3.2.2. DS {#sec3.2.2}\n\n[Fig.\u00a01](#fig1){ref-type=\"fig\"} shows DS\\'s bisection performance across sessions compared to controls'. The sessions were grouped into three stages: baseline (day 1 and day 2 pre), post-sham (day 2 post, day 6 and day 7 pre), and post-prism (day 7 post, day 8, day 18). Bisection errors in all baseline and post-sham sessions were outside the 95% confidence interval for controls, but were within normal bounds in the first two post-prism sessions. A one-way ANOVA revealed a main effect of stage \\[*F*(2,93)\u00a0=\u00a07.49, *p*\u00a0\\<\u00a0.005\\]. *T*-tests revealed no difference between baseline (*M*\u00a0=\u00a0\u22129.52, SEM\u00a0=\u00a01.52) and post-sham performance \\[*M*\u00a0=\u00a0\u221210.44, SEM\u00a0=\u00a0.97; *t*(58)\u00a0=\u00a0.537, *p*\u00a0=\u00a0.593\\]. Bisection deviation in the post-prism stage \\[*M*\u00a0=\u00a0\u22125.46, SEM\u00a0=\u00a0.68\\] was smaller than both baseline \\[*t*(58)\u00a0=\u00a02.72, *p*\u00a0\\<\u00a0.01\\] and post-sham \\[*t*(70)\u00a0=\u00a04.20, *p*\u00a0\\<\u00a0.001\\].\n\nStability of improvement was evaluated with a one-way ANOVA, which showed no difference between bisection errors in the three post-prism sessions \\[*F*(2,33)\u00a0=\u00a0.45, *p*\u00a0=\u00a0.641\\], although there appeared to be a trend for a return to baseline. Mean error on day 18 was within one SEM of the 95% confidence interval around the mean for the control group, but was also not different from baseline \\[*t*(34)\u00a0=\u00a01.34, *p*\u00a0=\u00a0.19\\]. Five additional daily adaptation sessions administered after day 18 resulted in no further reduction in bisection error (*M*\u00a0=\u00a0\u22125.58, SEM\u00a0=\u00a0.80; not shown in Figure).\n\n4. Discussion {#sec4}\n=============\n\nTo our knowledge this is the first report of a patient with right spatial neglect to be treated with prism adaptation. In previous work patients with left spatial neglect adapted to rightward- but not leftward-shifting prisms ([@bib11]). We therefore used leftward-shifting prisms to treat this patient with right spatial neglect to induce a rightward orienting after-effect. Prism adaptation was effective both in inducing a rightward after-effect and in improving neglect as measured on a bisection task.\n\nHealthy participants make small but systematic leftward errors in line midpoint estimations ('pseudoneglect'; [@bib3]). These shift rightward by 1--2% following adaptation to leftward-shifting prisms ([@bib1; @bib7]) but not always significantly so for manual bisection ([@bib15]). The baseline bisection errors of DS were larger than the controls\\', and her post-prism error reduction of approximately 5% was greater than the shifts previously reported for healthy participants. These results indicate both the presence of mild neglect, and a reduction of bisection bias after prism adaptation beyond that which would be expected for controls.\n\nAlthough further studies with greater numbers of patients are needed, our results suggest that the neurological process by which adaptation to rightward-shifting prisms ameliorates left neglect can occur in a similar fashion with leftward-shifting prisms for patients with right neglect. A proposed mechanism for the prism-induced improvements in left spatial neglect involves signals from right cerebellum that influence activity in the left parietal lobe via a network of left and right hemisphere areas ([@bib10]). These left parietal areas could be recruited for previously right parietal functions ([@bib10]), or may further influence the right superior parietal lobe via colossal communication ([@bib12]). Adaptation to leftward-shifting prisms using the left hand may result in a symmetrical process in patients with right spatial neglect, recruiting right parietal areas or influencing spared left hemisphere areas to restore rightward attention.\n\nThe authors would like to thank DS and eight healthy controls for their participation. Data collection was supported by the Wellcome and North-West Wales NHS Trusts, and JB was supported by the British Federation for Women Graduates.\n\nThis study was approved by the ethics committee of the School of Psychology, Bangor University in accordance with the 1964 Declaration of Helsinki.\n\n![Average line bisection errors (\u00b11 SEM) of patient DS for each session compared to the 95% confidence interval around the mean for healthy control participants. Sessions were grouped into three stages: baseline, post-sham, and post-prism as indicated by the shaded areas. Negative numbers indicate leftward bisection errors (right-sided neglect).](gr1){#fig1}\n"}
+{"text": "Introduction {#Sec1}\n============\n\nParkinson's disease (PD), the most common neurodegenerative disorder, is characterized by the loss of nigrostriatal dopamine neurons and the formation of intracellular Lewy bodies (LBs), which consist primarily of \u03b1-synuclein (hereafter referred to as synuclein) and ubiquitin^[@CR1]^. Interactions between genetic predisposition and environmental factors are likely the primary events inducing mitochondrial dysfunction and oxidative damage, resulting in oligomerization and aggregation of synuclein, but the underlying molecular mechanisms remain poorly understood. The vast majority of PD occurs sporadically, with inherited familial forms of the disease accounting for roughly 5% of all cases^[@CR2]^. The identification of PD-related genes and risk factors has implicated several pathways in PD etiology, with growing evidence suggesting a link between dysfunctional intracellular protein catabolism and PD pathogenesis. In *PARK1*-linked PD, intrinsically disordered mutant synuclein initiates the disease process. Given that highly aggregated proteins are deposited in nigral neurons in PD, dysfunctions of proteolytic systems, i.e., the ubiquitin--proteasome system and autophagy--lysosomal pathway, seem to contribute to the final neurodegenerative process.\n\nMacroautophagy (hereafter, referred to as autophagy) is a highly conserved bulk protein degradation pathway in eukaryotes. Cytoplasmic proteins and organelles are engulfed within autophagosomes, which fuse with the lysosome, where they are degraded along with their cargo^[@CR3]^. Several lines of evidence indicate that synuclein is predominantly degraded by autophagy, but also by the proteasome. However, mutant forms of synuclein and oligomers are dependent on the autophagy--lysosome pathway for their clearance^[@CR4]--[@CR6]^. Although the phenotypes of mice harboring brain-specific deletion of *Atg5* or *Atg7* reveal the critical role of autophagy in the removal of aggregated proteins^[@CR7],[@CR8]^, Friedman *et al*.^[@CR9]^ demonstrated that DA neuron-specific autophagy deficiency leads to the restrictive presynaptic accumulation of synuclein in the dorsal striatum, suggesting that impaired autophagy plays a role in PD pathogenesis. In DA neurons, the primary site of endogenous pathology, no detailed reports to date have examined the endogenous synuclein accumulation in dopaminergic cell bodies and neurites, which is associated with Lewy pathology. In this context, the stress-inducible protein p62 (encoded by the gene Sqstm1) is of importance because it is the ubiquitin- and LC3-binding protein that functions as a selective autophagy adaptor/receptor for degradation of ubiquitinated substrates^[@CR10]^. Indeed, p62 is present in neuronal inclusions of individuals with Alzheimer's disease and other neurodegenerative diseases^[@CR11]^, although its actual roles in pathogenesis remains largely unknown.\n\nIn order to understand the effects of autophagy impairment on DA neurons, we characterized conditional knock-out mice harboring a tyrosine hydroxylase (TH) neuron--specific deletion of *Atg7* and observed their age-related pathological and motor phenotypes.\n\nResults {#Sec2}\n=======\n\nGeneration of DA neuron--specific *Atg7* conditional knockout mice, and characterization of their locomotor impairments {#Sec3}\n-----------------------------------------------------------------------------------------------------------------------\n\nWe generated tyrosine hydroxylase (TH) cell--specific *Atg7* conditional knockout mice (*Atg7*^*flox*/*flox*^:TH-Cre) by crossing the previously characterized *Atg7* floxed mice (*Atg7*^*flox*/*flox*^)^[@CR8]^ with TH-Cre mice (TH-Cre) harboring the Cre recombinase coding sequence downstream of a characterized fragment of the TH promoter^[@CR12]^ instead of the IRES version (Fig.\u00a0[1A](#Fig1){ref-type=\"fig\"}). In *Atg7*^*flox*/*flox*^:TH-Cre midbrains, the quantification of Atg7 after normalization by actin (Supple. Figure\u00a0[1B](#MOESM1){ref-type=\"media\"}) showed 13.54\u2009\u00b1\u20092.27% residual amount, compared to controls. Considering the percentage of dopamine neurons is not so high, it is intriguing that the decrease in Atg7 is remarkable.Figure 1Behavioral and neuropathological examinations of *Atg7*^*flox*/*flox*^:TH-Cre mice. (**A**) Schematic representation of the genetic cross between mice carrying the floxed *Atg7* allele (*Atg7* F/F) and knock-in mice carrying Cre inserted at the 3'UTR of the TH gene (TH-Cre) to generate *Atg7* F/F:TH-Cre (or *Atg7*^*flox*/*flox*^:TH-Cre) mice. (**B**) Runway test of *Atg7*^*flox*/*flox*^ mice (a) and *Atg7*^*flox*/*flox*^:TH-Cre mice. (b) In contrast to the *Atg7*^*flox*/*flox*^ mice, which exhibited well-coordinated movement and almost no slips of the forepaw or hindpaw from the beam, *Atg7*^*flox*/*flox*^:TH-Cre mice could hardly move on the beam and the hindpaw slipped frequently (over 110 weeks of age). In this figure, the hindpaw of the *Atg7*^*flox*/*flox*^:TH-Cre mice has slipped off the beam. (c) The number of hindlimb slips was recorded as 85-week-old and 120-week-old mice cross the challenging beam. Data show means\u2009\u00b1\u2009SE (error bars), and statistical significance was evaluated using Student's t-test. \\*\\*P\u2009\\<\u20090.01. (**C**) In the accelerating rotarod assay, rotation was accelerated from 3\u2009rpm to 35\u2009rpm over 5\u2009min, and fall latency was recorded, in mice from 95 to 120 weeks of age. Red line, *Atg7*^*flox*/*flox*^:TH-Cre mice; blue line, *Atg7*^*flox*/*flox*^ mice. Data show means\u2009\u00b1\u2009SE (error bars), and statistical significance was evaluated using Student's t-test. \\*\\*P\u2009\\<\u20090.01. (**D**) Histological analyses of locus coeruleus (LC) in a 9-month-old *Atg7*^*flox*/*flox*^:TH-Cre mouse. Paraffin sections were stained with hematoxylin--eosin (HE) (a) and immunostained for TH (b), ubiquitin (c), and p62 (d). In the each staining (a-d), many inclusions are apparent (arrow). Scale bars: 10\u2009\u00b5m. (**E**) Immunofluorescence labeling of p62 (green) or ubiquitin (red) in the LC of an *Atg7*^*flox*/*flox*^:TH-Cre mouse (a--c) and an *Atg7*^*flox*/*flox*^ mouse (d--f)., Scale bars: 20\u2009\u00b5m.\n\n*Atg7*^*flox*/*flox*^:TH-Cre mice were viable at birth and indistinguishable in appearance from their littermates, and their survival rate was not markedly diminished. Although *Atg7*^*flox*/*flox*^:TH-Cre mice have not yet been observed over the entire lifespan, they began to show impairment in motor coordination tasks around 100 weeks and apparent motor behavioral deficits around 110 weeks. These clinical abnormalities could be demonstrated by the runway test (Fig.\u00a0[1Ba,b and c](#Fig1){ref-type=\"fig\"}) and rotarod test (Fig.\u00a0[1C](#Fig1){ref-type=\"fig\"}). In contrast to *Atg7*^*flox*/*flox*^ mice, which exhibited well-coordinated movement and almost no slips of the forepaw or hindpaw from the beam, the, *Atg7*^*flox*/*flox*^:TH-Cre mice could hardly move on the beam and slipped frequently. In particular, the hindpaws of *Atg7*^*flox*/*flox*^:TH-Cre mice often slipped off the beam (Fig.\u00a0[1Bb and c](#Fig1){ref-type=\"fig\"}). Furthermore, in the accelerating rotarod test, the fall latency was reduced in *Atg7*^*flox*/*flox*^:TH-Cre mice (Fig.\u00a0[1C](#Fig1){ref-type=\"fig\"}). Gait disturbance progressed, and by the terminal stage, the majority of affected mice could hardly move.\n\nLoss of *Atg7* leads to age-related development of p62 inclusions in the DA neuron {#Sec4}\n----------------------------------------------------------------------------------\n\nPrevious studies showed that deletion of *Atg5* or *Atg7* in the central nervous system (CNS) leads to formation of inclusions positive for the autophagy adaptor/receptor protein p62 in various neuronal populations. However, the mechanisms underlying age-related progression and intracellular localization of these inclusions in dopaminergic neurons remain elusive. Initially, we confirmed aggregate formation in the locus coeruleus (LC), because the somata of neurons in the LC are wide enough to easily visualize the inclusions. In *Atg7*^*flox*/*flox*^:TH-Cre mice, these neurons contained 7 eosinophilic aggregates, which are characteristic of LBs (Fig.\u00a0[1Da](#Fig1){ref-type=\"fig\"}). In addition, these inclusions were observed in TH-positive neurons (Fig.\u00a0[1Db](#Fig1){ref-type=\"fig\"}), including ubiquitin (Fig.\u00a0[1Dc](#Fig1){ref-type=\"fig\"}) and p62 (Fig.\u00a0[1Dd](#Fig1){ref-type=\"fig\"}). In fact, p62 and ubiquitin colocalization can make a strong prediction of LB (Fig.\u00a0[1E](#Fig1){ref-type=\"fig\"}). Next, we observed the staining of dopaminergic neurons in the substantia nigra (SN). In contrast to *Atg7*^*flox*/*flox*^ mice, in which no aggregates were seen even at the age of 18 months (Fig.\u00a0[2Ab](#Fig2){ref-type=\"fig\"}), p62-positive inclusions were present in 2-month-old *Atg7*^*flox*/*flox*^:TH-Cre mice (Fig.\u00a0[2Aa](#Fig2){ref-type=\"fig\"}). Notably, Western blotting revealed p62 and ubiquitin accumulations in the midbrain of 2 month-old *Atg7*^*flox*/*flox*^:TH-Cre mice, whereas synuclein did not accumulate in the midbrain, at least at this stage (data not shown). In immunohistochemical staining, dopaminergic neurons exhibited no synuclein accumulation (data not shown). In 18-month-old mice, p62 inclusions were larger (Fig.\u00a0[2Ae and B](#Fig2){ref-type=\"fig\"}) and mainly localized outside the soma (Fig.\u00a0[2Ac](#Fig2){ref-type=\"fig\"}). To determine the location of these inclusions, we conducted double DAB staining with p62 (Fig.\u00a0[2C](#Fig2){ref-type=\"fig\"} black) and TH antibodies (Fig.\u00a0[2C](#Fig2){ref-type=\"fig\"} brown). P62 inclusions were present along the TH fibers, and were mainly located in the branch of dopaminergic neurons (Fig.\u00a0[2Cb](#Fig2){ref-type=\"fig\"}). To investigate the localizations of aggregates outside the soma, we conducted double labeling (Fig.\u00a0[2D](#Fig2){ref-type=\"fig\"}). Specifically, we labeled dopaminergic neurons with TH antibody (Fig.\u00a0[2D](#Fig2){ref-type=\"fig\"} red) and followed individual fibers, revealing that p62-positive aggregates (Fig.\u00a0[2D](#Fig2){ref-type=\"fig\"} green) were present along neurites. To further characterize these aggregates, we performed ultrastructural analysis in SN (Fig.\u00a0[2E](#Fig2){ref-type=\"fig\"}). Fibrous inclusions localized in the soma (Fig.\u00a0[2Ea,b](#Fig2){ref-type=\"fig\"}) and neurites (Fig.\u00a0[2Ec,d](#Fig2){ref-type=\"fig\"}) in DA neurons. In these inclusions, autophagosome-like structures were observed (Fig.\u00a0[2Ec](#Fig2){ref-type=\"fig\"}). But, as to autophagosome formation, ATG conjugation systems including Atg7 are not absolutely essential^[@CR13]^. In fact, these autophagosome-like structures were previously reported in Atg7-deficient brains^[@CR14]^.Figure 2Development of p62-positive aggregates in dopaminergic neurons. (**A**) Histological analyses of 2-month-old mice (a), 18-month-old *Atg7*^*flox*/*flox*^ control (b), and 18-month old *Atg7*^*flox*/*flox*^:TH-Cre mice (c). Cryosections were immunostained for p62. In contrast to *Atg7*^*flox*/*flox*^ mice, in which no aggregates were seen in dopaminergic neurons, p62 positive aggregates increased in number (d) and size (e) in the dopaminergic neurons of *Atg7*^*flox*/*flox*^:TH-Cre mice as a function of age. Scale bars: 20\u2009\u00b5m. High magnification of p62 inclusions is shown in the inset. Scale bar: 50\u2009\u00b5m (inset) Data show means\u2009\u00b1\u2009SE (error bars), and statistical significance was evaluated using Student's t-test. \\*\\*P\u2009\\<\u20090.01. (**B**) Immunoblot for ubiquitin, synuclein, p62, and actin. Lanes 1--3: 18-month-old *Atg7*^*flox*/*flox*^ mice; lanes 4--6: 18-month-old *Atg7*^*flox*/*flox*^:TH-Cre mice. (**C**) (a) Histological analyses of dopaminergic neurons in an 18-month-old *Atg7*^*flox*/*flox*^:TH-Cre mouse. Cryosections were immunostained for p62 (black) and TH (brown) Scale bars: 50\u2009\u00b5m. Square areas in (a) are enlarged in (b). P62 inclusion (arrow) is present along TH fibers. Scale bars: 10\u2009\u00b5m. (**D**) The number of p62-positive aggregates within TH fibers was identified by high-resolution image through dopaminergic neurons revealed immunofluorescence labeling of TH (red) and p62 (green). Scale bars: 20\u2009\u00b5m. High magnification of inclusion is shown in the inset. Scale bars: 20\u2009\u00b5m (inset). Data show means\u2009\u00b1\u2009SE (error bars), and statistical significance was evaluated using Student's t-test. \\*\\*P\u2009\\<\u20090.01. (**E**) Electron micrographs of neurons in the substantia nigra (SN). Somata (a,b) and neurites (c,d) of 18-month-old *Atg7*^*flox*/*flox*^:TH-Cre mice. Square areas in (a) are enlarged in (b). Electron-dense filamentous inclusions were observed in neurons of the SN. (c) Inclusion in a neurite, containing mitochondria and autophagosomes (white arrows). Inclusion is labeled with a dashed line. (d) Electron micrograph of an inclusion immunostained for p62. Electron-dense area (white asterisk) indicates a p62-positive aggregate, which contains many mitochondria (white arrowheads). Inclusion is labeled with a dashed line. Scale bars: (a) 5\u2009\u00b5m, (b) 2\u2009\u00b5m, (c) 1\u2009\u00b5m, (d) 500\u2009nm.\n\nTo distinguish the p62 positive inclusions, we immunostained SN slices with p62 antibody and identified lesions based on the DAB-labeled products. Ultrathin sections containing these target lesions were cut and examined under an electron microscope. In Fig.\u00a0[2Ed](#Fig2){ref-type=\"fig\"}, the black region contains p62 positive aggregates. Because such aggregates involve many mitochondria, we speculate that mitochondrial transport might be disturbed in DA neurons if these aggregates were formed in neurites.\n\nP62-positive inclusions contain endogenous synuclein {#Sec5}\n----------------------------------------------------\n\nPrevious *in vivo* analysis of DA neurons suggested that synuclein regulation is linked to autophagy^[@CR9],[@CR15]^, but it remained unclear how synuclein accumulation is associated with PD pathology. To address this issue, we asked whether endogenous synuclein colocalizes with p62-positive inclusions in *Atg7*^*flox*/*flox*^:TH-Cre mice. To this end, we carefully observed thickly sliced midbrain blocks, which includes the SN, using a VS120 microscopic 3D measurement system (Olympus, Tokyo, Japan). Surprisingly, around 9 months, we pathologically confirmed endogenous synuclein accumulation in somata and neurites (Fig.\u00a0[3A](#Fig3){ref-type=\"fig\"}). High resolution confocal images through SN show immunofluorescence labeling of TH and synuclein, where 89.30\u2009\u00b1\u20091.65% (n\u2009=\u20095, each 2 slices) synuclein inclusions located in TH fibers (Supple. Figure\u00a0[2C](#MOESM1){ref-type=\"media\"}). Furthermore, the level of synuclein increased in the midbrain of *Atg7*^*flox*/*flox*^:TH-Cre mice (Fig.\u00a0[3B and C](#Fig3){ref-type=\"fig\"}). Although p62 is a representative substrate of autophagy that is rapidly influenced by its dysfunction, it seemed that not only p62, but also endogenous synuclein, was regulated to some extent by autophagy. Next, we investigated whether endogenous synuclein colocalized with p62 aggregates. In confocal microscopy, *Atg7*^*flox*/*flox*^ mice exhibited no p62-positive aggregates or endogenous synuclein accumulation (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"} upper panel). On the other hand, *Atg7*^*flox*/*flox*^:TH-Cre mice exhibited synuclein deposition and inclusions that colocalized with p62 in cell bodies (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"} middle panel). Interestingly, we observed these inclusions not only within the soma, but also outside the soma (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"} lower panel). To determine the localization of these inclusions, we analyzed many samples by confocal microscopy, and found that most inclusions localized in neurites, as demonstrated in Fig.\u00a0[2D](#Fig2){ref-type=\"fig\"}. When we observed these inclusions at higher magnification, they have multiple cavities (Fig.\u00a0[3E](#Fig3){ref-type=\"fig\"}), consistent with the results of ultrastructural analysis (Fig.\u00a0[2Ed](#Fig2){ref-type=\"fig\"}). Thus, loss of autophagy induces synuclein accumulation and formation of LBs, similar to the pathology observed in DA neurons.Figure 3Endogenous synuclein accumulates in dopaminergic neurons of *Atg7*^*flox*/*flox*^:TH-Cre mice. (**A**) (a) Histological analyses of substantia nigra (SN) in a 9-month-old *Atg7*^*flox*/*flox*^:TH-Cre mouse. Thickly sliced (40-\u00b5m) cryosections were immunostained for synuclein, and the SN was broadly observed on a VS120 microscopic 3D measurement system (Olympus). Scale bars: 100\u2009\u00b5m. Square areas in (a) are enlarged in (b) (Scale Bar: 20\u2009\u00b5m) and square areas in (b) are enlarged in (c) (Scale bars: 10\u2009\u00b5m). Arrows indicate synuclein inclusions within soma and neurites. (**B**) Immunoblot for synuclein and actin. Lanes 1--3: 9-month-old *Atg7*^*flox*/*flox*^ mouse; lanes 4--6: 9-month-old *Atg7*^*flox*/*flox*^:TH-Cre mouse. (**C**) Quantitation of the immunoblot for synuclein; ratio of synuclein to actin is shown. Synuclein accumulated in the midbrains of *Atg7*^*flox*/*flox*^:TH-Cre mice. Data show means\u2009\u00b1\u2009SE (n\u2009=\u20093). \\*\\*P\u2009\\<\u20090.01 (Student's *t-*test). (**D**) Immunofluorescence labeling of synuclein (green) or p62 (red) in the SN of an *Atg7*^*flox*/*flox*^ mouse (Upper panels) and an *Atg7*^*flox*/*flox*^:TH-Cre mouse (middle and lower panels). Synuclein accumulates in somata (middle panels) and/or neurites (lower panels), where synuclein colocalizes with p62-positive inclusions. DAPI nuclear staining is shown in blue, Scale bars: 20\u2009\u00b5m. (**E**) High-resolution confocal three-dimensional (3D) image of p62 and synuclein positive inclusions, which localize around soma in the midbrain. DAPI nuclear staining is shown in blue. Scale bars: 10\u2009\u00b5m.\n\nThe number of TH neurons decreases in aged *Atg7*^*flox*/*flox*^:TH-Cre mice, contributing to locomotor dysfunction {#Sec6}\n-------------------------------------------------------------------------------------------------------------------\n\nTo date, many genetically modified mice have been developed as PD models, but most of them do not exhibit neuronal loss. To assess the influence of aggregate formation, we compared the number of TH neurons between aged *Atg7*^*flox*/*flox*^:TH-Cre and *Atg7*^*flox*/*flox*^ mice. As demonstrated in the rotarod test (Fig.\u00a0[1C](#Fig1){ref-type=\"fig\"}), *Atg7*^*flox*/*flox*^:TH-Cre mice exhibited locomotor dysfunction at ages above 110 weeks. We sacrificed these mice (*Atg7*^*flox*/*flox*^:TH-Cre, n\u2009=\u20094; *Atg7*^*flox*/*flox*^ mice, n\u2009=\u20095) and counted TH neurons in three sections (VTA: Ventral Tegmental Area, SNcc: center area of substantia nigra pars compacta, SNcl: lateral area of substantia nigra pars compacta) (Fig.\u00a0[4A](#Fig4){ref-type=\"fig\"}). *Atg7*^*flox*/*flox*^:TH-Cre mice had fewer TH neurons than *Atg7*^*flox*/*flox*^ mice. The reduction in TH cell number was most prominent in the center area of substantia nigra pars compacta (SNcc) (Fig.\u00a0[4B](#Fig4){ref-type=\"fig\"}). No neuronal loss was observed in young mice, which did not exhibit motor dysfunction (data not shown). In PD, the reduction in the abundance of TH neurons also occurs primarily in the SNc, consistent with our results. Dopaminergic neuronal loss may contribute to motor impairment observed at the late stages of disease. Furthermore, we tested DA physiology in these 120-week-old *Atg7*^*flox*/*flox*^:TH-Cre mice by neurochemical analysis of the dorsal striata. High-performance liquid chromatography (HPLC) revealed a reduction in striatal DA levels and metabolites in *Atg7*^*flox*/*flox*^:TH-Cre versus control mice (Fig.\u00a0[4C](#Fig4){ref-type=\"fig\"}). Thus, DA content is affected by dopaminergic neuronal loss.Figure 4Delayed dopaminergic neuronal degeneration in *Atg7*^*flox*/*flox*^:TH-Cre mice. (**A**) Histological analyses of SN in 120-week-old *Atg7*^*flox*/*flox*^:TH-Cre and *Atg7*^*flox*/*flox*^ mice. Paraffin sections were immunostained for TH and demonstrated as shown the Table. SNcl: lateral area of substantia nigra pars compacta; SNcc: center area of substantia nigra pars compacta; VTA: Ventral Tegmental Area. Scale bars: 20\u2009\u00b5m. (**B**) The number of dopaminergic neurons in the SN was identified by TH immunoreactivity. Data are means\u2009\u00b1\u2009SE (*Atg7*^*flox*/*flox*^:TH-Cre, n\u2009=\u20094; *Atg7*^*flox*/*flox*^ mice, n\u2009=\u20095) \\*\\*P\u2009\\<\u20090.01 (Student's *t-*test). (**C**) HPLC analysis of DA, DOPAC, and HVA levels in the dorsal striatum at 120 weeks. Data are means\u2009\u00b1\u2009SE (*Atg7*^*flox*/*flox*^:TH-Cre, n\u2009=\u20093; *Atg7*^*flox*/*flox*^ mice, n\u2009=\u20093) \\*P\u2009=\u20090.05 (Student's *t-*test).\n\nDiscussion {#Sec7}\n==========\n\nGreat progress has been made toward understanding the pathogenesis of familial PD, mainly due to the discovery of specific mutations, whereas little is understood about the mechanisms underlying idiopathic PD pathogenesis. In this study, we carried out detailed observations of *Atg7*-deficient mice, and observed a loss of autophagy in dopaminergic neurons, resulting in Lewy pathology and motor dysfunctions associated with aging. Autophagy maintains the level of endogenous synuclein in DA neurons. In aged *Atg7*^*flox*/*flox*^:TH-Cre mice, synuclein accumulated in the neurites and/or cell bodies involved in inclusions. In addition, these inclusions had characteristics of LBs.\n\nOur results indicate that long-term disruption of autophagy results in widespread synuclein accumulation in DA neurons. In previous report, Friedman *et al*.^[@CR9]^ reported that dopaminergic axons developed synuclein aggregates at the age of 20 months, and suggested that autophagy is involved in axonal synuclein homeostasis in mice. In our careful observation of *Atg7*^*flox*/*flox*^:TH-Cre mice, we found that the level of endogenous synuclein protein increased at around 9 months in cell bodies and neurites (Fig.\u00a0[3A,B and C](#Fig3){ref-type=\"fig\"}), and these accumulated synuclein deposits contained p62 positive aggregates (Fig.\u00a0[3D](#Fig3){ref-type=\"fig\"}). Our synuclein antibody efficiently detects endogenous mouse synuclein, and a virtual slide system could be used to observe neurons extensively.\n\nIn any case, it is clear that autophagy plays a significant role in synuclein degradation in dopaminergic neurons. The protein catabolic pathways that regulate synuclein degradation have been discussed by multiple authors^[@CR4],[@CR16],[@CR17]^, and the extent to which autophagy contributes to synuclein degradation depends on the presence of mutations^[@CR18]^. Several lines of evidence indicate a role for autophagy in the regulation of synuclein homeostasis^[@CR5],[@CR6],[@CR15]^, suggesting that dysfunctional autophagic clearance may contribute to the development of synuclein inclusions. Conversely, overexpression of wild-type synuclein impairs autophagic activity, implying a functional relationship between synuclein and autophagic degradation^[@CR19]^. Furthermore, LB-like synuclein aggregates resist degradation and impair autophagy^[@CR20]^. In this way, synuclein accumulation may be influenced by the interaction between autophagic impairment and inclusion formation. Intriguingly, in this study we observed for the first time that synuclein deposition is preceded by the formation of p62 inclusion in DA neurons. It is well known that homeostasis of p62 is strictly regulated by autophagy in the brain^[@CR10]^, and that dysregulation of autophagy causes formation of p62 aggregates. Our results suggest that autophagy regulates synuclein levels to some extent, and we speculate that p62 aggregates caused by autophagic defects serve as 'seeds' for synuclein inclusions.\n\nMany questions remain unanswered with respect to LB formation. One hypothesis is that Lewy pathology spreads from the axon to the soma^[@CR21]^. DA neurons have a unique structural characteristic, long hyper-branched projection axons, that innervate wide areas in the brain^[@CR22]^. It is likely that this feature increases the chance of developing deposits in peripheral axons and dendrites. In the human brain, synuclein deposition and neuronal degeneration are accentuated in distal regions^[@CR23]^. In our study, *Atg7*^*flox*/*flox*^:TH-Cre mice exhibited age-related growth and spreading of p62 aggregates in the peripheral region (Fig.\u00a0[2A](#Fig2){ref-type=\"fig\"}). It is likely that aspects of the local environment related to branching may predispose branch regions to synuclein deposition^[@CR21]^. Our observations that p62- and synuclein-positive inclusions are prone to colocalize in peripheral regions (Fig.\u00a0[3D and E](#Fig3){ref-type=\"fig\"}) support this hypothesis.\n\nOur ultrastructural study revealed that many mitochondria were present in aggregates (Fig.\u00a0[2Ec,d](#Fig2){ref-type=\"fig\"}). During mitochondrial degradation in cultured cells, depolarized mitochondria cluster with p62^[@CR24]^. Depolarized mitochondria are ubiquitylated by Parkin, recognized by p62, transported via microtubules to aggregates, and degraded by autophagy. Inhibition of physiological microtubule transport and clustering of depolarized mitochondria may contribute to inclusion formation. Electron microscopic analysis by Gai *et al*. revealed that many mitochondria are concentrated in the early forms of LBs^[@CR25]^, and Bedford *et al*. also reported deposits of mitochondria in early LBs (pale bodies) in PD patients^[@CR26]^. We speculate that p62 aggregates containing mitochondria (Fig.\u00a0[2Ec,d](#Fig2){ref-type=\"fig\"}) represent inclusions at the early phase of biogenesis of LBs. Thus, Parkin-mediated p62 clustering may be a key process of LB formation, consistent with the fact that LBs are usually absent in autosomal recessive early-onset Parkinsonism caused by Parkin mutations^[@CR27]^.\n\nNone of the current genetic PD models in mouse recapitulates all features of PD. Additionally, only a few of these models develop mild DA neurodegeneration. The most parsimonious explanation for the lack of DA neurodegeneration in genetic PD models is a compensatory mechanism that results from adaptive changes during development, making it hard to observe the degenerative phenotype over the lifespan of mice.\n\nIn this study, the age-related motor dysfunction and pathology in mice with *Atg7* deficiency in DA neurons suggests that impairment of autophagy is a potential mechanism underlying the pathology of PD. Our *Atg7*-deficient mice demonstrate typical Lewy pathology, including endogenous synuclein and neuronal loss, which resembles PD. Furthermore DA levels are affected by dopaminergic neuronal loss. Intriguingly, the significant decrease in dopamine is not accompanied by concomitant reductions in the dopamine catabolites DOPAC and HVA. Until now, we did not clarify the reason, but one of the explanation may be presynaptic dysfunction in Atg7-deficient mice^[@CR9]^. These PD models will provide insight into the process of autophagy in PD pathology, and will be crucial for the development of novel therapeutic targets.\n\nMaterials and Methods {#Sec8}\n=====================\n\nAnimals {#Sec9}\n-------\n\nAll animals were kept in a pathogen- and odor-free environment, which was maintained under a 12\u2009h light/dark cycle at ambient temperature. Procedures were approved by the Animal Experimental Committee of the Juntendo University Graduate School of Medicine, and were performed in accordance with the guidelines of NIH and the Juntendo University Graduate School of Medicine. Floxed *Atg7* mice were characterized previously^[@CR8]^ and were crossed with TH-Cre mice carrying the knock-in construction containing TH fused to Cre in the 3\u2032 end (gift from Dr. Ted M. Dawson, Johns Hopkins University, USA) to generate *Atg7*^*flox*/*flox*^:TH-Cre mice.\n\nBehavioral tests {#Sec10}\n----------------\n\nLocomotor behavior was assessed in mice from 90 to 120 weeks of age. Accelerating rotarod tests were performed on a rotarod machine with automatic timers and falling sensors (MK-660D, Muromachi Kikai). Male *Atg7*^*flox*/*flox*^ mice and *Atg7*^*flox*/*flox*^:TH-Cre mice (n\u2009=\u200910 for each genotype) were placed on a 3-cm diameter rotating rod covered with rubber, and rotation was accelerated from 3 to 35\u2009rpm over 5\u2009min. Fall latency was recorded, and the first fall latency of the third trial was used for analysis. The runway test was performed using a narrow, horizontally fixed beam (1\u2009cm wide, 80\u2009cm long, held at a height of 40\u2009cm from the table). The animal was placed at one end of the beam and urged to move toward the opposite end, where an escape platform was located.\n\nImmunoblot analyses {#Sec11}\n-------------------\n\nAnesthetized mice were dissected by decapitation, and lysates for SDS-PAGE were prepared from midbrain tissue as described previously^[@CR28]^. After electrophoresis, separated proteins were transferred to PVDF membranes (Hybond-P; Amersham Biosciences), which were incubated overnight at 4\u2009\u00b0C with the indicated primary antibodies: anti-ubiquitin (Dako), anti-synuclein-1 (BD bioscience), and anti-actin (Millipore: MAB1501). Antibody binding was visualized using the Enhanced Chemiluminescence Kit (GE Healthcare), and signals were detected on an ImageQuant LAS4000 (GE Healthcare).\n\nHistological analyses {#Sec12}\n---------------------\n\nMice were perfused with 4% paraformaldehyde (PFA), and their brains were immersion-fixed at 4\u2009\u00b0C for 36 hrs. The fixed samples were cryoprotected with 20% sucrose and sliced on a freezing microtome to obtain 40-\u00b5m-thick floating sections. For double immunohistochemistry of p62 and TH, sections were initially incubated with the anti-p62 antibody (PROGEN Biotechnik, GmbH) and visualized with diaminobenzidine (DAB)-containing nickel ammonium sulfate (DABNi), which generates dark purple precipitates. The same section was then incubated with the anti-TH antibody (657012, Calbiochem, Germany), and visualized with DAB, which generates brown precipitates. The sections were observed on a VS120 (Olympus, Tokyo, Japan). For double immunofluorescence of p62 (green) and TH (red), floating sections were incubated with rabbit anti-TH antibody (657012, Calbiochem, Germany) and guinea pig anti-p62 antibody (PROGEN Biotechnik, GmbH). They were then incubated with anti--guinea pig IgG conjugated with Alexa Fluor 488 and anti-rabbit IgG conjugated with Alexa Fluor 546. For double immunofluorescence study of synuclein (green) and p62 (red), floating sections were incubated with the rabbit anti-synuclein antibody (AB5038P, Millipore, USA) and guinea pig anti-p62 antibody (PROGEN Biotechnik, GmbH). They were then incubated with anti-rabbit IgG conjugated with Alexa Fluor 488 and anti--guinea pig IgG conjugated with Alexa Fluor 546. Fluorescent signals were captured by LSM 780 confocal microscopy (Zeiss).\n\nElectron microscopy and immunoelectron microscopy with DAB {#Sec13}\n----------------------------------------------------------\n\nFor conventional electron microscopy, mice were fixed by cardiac perfusion with 2% PFA and 2% glutaraldehyde in 0.1\u2009mol/L PB (pH 7.2). Brain slices were embedded in epoxy resin, and ultrathin sections (70\u2009nm thick) were cut and observed on a Hitachi HT7700 electron microscope (Hitachi, Tokyo, Japan). For immunoelectron microscopy with DAB, mice were perfused and fixed with 4% paraformaldehyde, 0.5% glutaraldehyde (GLA), and 15% saturated picric acid in 0.1\u2009M phosphate buffer (PB) for 1 hr at room temperature (RT). Coronal 50 \u00b5m-thick slices of substantia nigra (SN) were prepared with a microslicer (VT1200S, Leica Microsystems, GmbH), and SN-containing slices were subjected to the indicated immunohistochemical analysis. Slices were treated with 0.6% hydrogen peroxide in saline for 15\u2009mins. After washing, the slices were incubated for 6 weeks at 4\u2009\u00b0C with guinea pig anti-p62 antibody (PROGEN Biotechnik, GmbH) in 1% bovine serum albumin in 0.1\u2009M phosphate-buffered saline (PBS). Subsequently, the slices were incubated for 2 hrs at RT with horseradish peroxidase--conjugated anti--guinea pig IgG antibody (Jackson ImmunoResearch, USA), and then developed with DABNi. The immunoreacted slices were re-fixed with 2.5% GLA in 0.1\u2009M cacodylate buffer (CB) for 15\u2009min at 4\u2009\u00b0C. After several rinses in 0.1\u2009M CB with 4.5% sucrose, the slices were post-fixed with 2% osmium tetroxide in 0.1\u2009M CB for 2 hrs at 4\u2009\u00b0C, dehydrated in a graded series of ethanol followed by propylene oxide, and then horizontally embedded in epoxy resin (EPON 812, TAAB, UK). Embedded samples were sectioned into 4 \u00b5m-thick serial semi-thin sections, and identified based on DABNi color development under a light microscope. Subsequently, sections that contained target lesions were adhered onto new blocks of epoxy resin. Ultrathin sections for silver and/or gray interference color were cut from the region, placed in formvar-coated one-slot grids, stained with uranyl acetate and lead citrate, and examined on a transmission electron microscope (JEM-1400, JEOL, Tokyo).\n\nHPLC Analysis {#Sec14}\n-------------\n\nDorsal striata from 120-week-old mice were dissected, quickly frozen on dry ice, and then homogenized with 0.5\u2009mL of 0.2\u2009M perchloric acid containing 100\u2009\u00b5M EDTA-2Na per 100\u2009mg wet tissue. Samples were centrifuged at 20,000\u2009\u00d7\u2009g for 15\u2009min at 4\u2009\u00b0C. The supernatant was collected and analyzed by HPLC.\n\nStatistical analysis {#Sec15}\n--------------------\n\nStatistical significance was determined by Student's t-test (STATVIEW; SAS Institute). Data are expressed as means\u2009\u00b1\u2009S.E. P\u2009\\<\u20090.05 was considered significant.\n\nElectronic supplementary material\n=================================\n\n {#Sec16}\n\nSupplemental Info\n\n**Electronic supplementary material**\n\n**Supplementary information** accompanies this paper at 10.1038/s41598-018-21325-w.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nWe thank Dr. Ted M. Dawson for providing us with the TH-Cre mice, and Kentaro Endo (Tokyo Metropolitan Institute), Kyohei Mikami (Tokyo Metropolitan Institute), and Souichiro Kakuta (Juntendo University) for excellent technical assistance with electron microscopy. This work was supported by a KAKENHI Grant-in-Aid for Scientific Research on Innovative Areas, \"Multidisciplinary research on autophagy: from molecular mechanisms do disease states\" from JSPS.\n\nS.S., T.U., M.K., and N.H. designed the research; S.S., T.F., S.N., and H.K. performed the research; T.U., M.K., and N.H. analyzed data; and S.S. and N.H. wrote the manuscript with help and supervision from Y.U. and K.T.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare no competing interests.\n"}
+{"text": "Background\n==========\n\nIntimate partner violence (IPV) against women is a serious public health issue. Estimates of rates of IPV vary in the literature, depending on the definition used, data collection procedures and sampling strategies. Annual Canadian population estimates have varied over time from 10% in 1993 \\[[@B1]\\] to 1.9% in 2009 \\[[@B2]\\]. However, other surveys have found higher rates of abuse experienced by women, depending in part on the sample population. For example, in 2002--2003, Thurston and colleagues found a disclosure rate of 19.0% when women in an urgent care clinic were directly asked by nurses during the first year of implementation of a universal domestic violence screening protocol \\[[@B3]\\]. Depending on the tool used, recent Ontario, Canada trials have identified between 4.1--22% of adult women presenting to emergency departments (EDs), family medicine practices and women's health clinics reporting IPV over the past year \\[[@B4]-[@B6]\\].\n\nWomen are not likely to disclose abuse unless directly asked \\[[@B7]\\]. Using focus groups with women who had experienced IPV and who were currently using support programs, Chang and colleagues noted that these women offered specific advice to health care providers when asking about IPV (i.e., provide a rationale for the inquiry to lessen feelings of shame and apprehension; ask when alone in a safe and supportive environment; and offer information, support and access to resources even if women do not disclose \\[[@B8]\\]). Yet, few women are asked despite presenting with signs and symptoms suggestive of exposure to IPV. A study by Glass and colleagues demonstrated that less than 25% of women presenting for any reason were asked about physical, sexual and emotional IPV by ED staff, including only 39% of women presenting with acute trauma, and 13% with past-year IPV \\[[@B9]\\]. The issue regarding whether to ask by routine universal screening, versus clinically-indicated case-finding, has been addressed elsewhere \\[[@B6],[@B10]\\]. Recent randomized trial level evidence does not support universal screening \\[[@B4],[@B11]\\].\n\nBarriers and facilitators to asking about IPV vary according to a number of factors. For example, Rodriguez and colleagues surveyed 375 culturally diverse women who attended public clinics to examine factors associated with abuse disclosure to physicians \\[[@B12]\\]. Forty-two percent of these women stated they had talked to a physician about their abuse. These respondents perceived that physicians did not ask directly about abuse and that they had insufficient time and interest in discussing abuse. Furthermore, respondents described fear of legal involvement and concerns about confidentiality as barriers to disclosing abuse. However, of all the factors measured in this study, clinician inquiry was the strongest determinant of abuse disclosure.\n\nOther studies have focused on the experiences of health care professionals. In an ethnographic study of 38 physicians who were primarily family practitioners working in an urban health maintenance organization, barriers to addressing IPV included lack of comfort, fear of offending, powerlessness, frustration, loss of control and time constraints \\[[@B13]\\]. Similarly, a review article indicated that lack of provider education regarding woman abuse, fear of offending women, lack of time, clients not disclosing and lack of effective interventions were important barriers to asking about IPV \\[[@B14]\\]. Rodriguez et al. also noted that physicians were more likely to identify patient-related barriers to identification and intervention than physician-related barriers \\[[@B15]\\].\n\nSimilar issues were identified by nurses, including lack of time, lack of training in both assessment and how to respond as well as unique challenges related to their role including pressure from physicians to see patients quickly, presence of family members, language and cultural differences and challenges of screening older clients and clients with mental health issues \\[[@B16]-[@B18]\\]. Other studies of health care practitioners have found that older, more experienced clinicians, and those with histories of exposure to abuse, were more likely to ask about IPV \\[[@B19]\\].\n\nStudy objective\n---------------\n\nThe overall objective was to identify barriers and facilitators to asking about IPV among a large, randomly selected sample of nurses and physicians in specified areas of practice where abused women are likely to present. Specific goals were to: 1) explore physician's and nurses' experiences, both professional and personal, when asking about IPV; 2) determine variation by discipline i.e., nurse or physician; and 3) identify implications for practice, workplace policy and curriculum development. This study was part of the McMaster Violence Against Women Research Program, a multi-study program of research developed to investigate the health care response to IPV experienced by women, including how best to identify women presenting to health care settings.\n\nMethods\n=======\n\nSample and data collection\n--------------------------\n\nThis paper summarizes responses to two open-ended questions about barriers and facilitators to asking about IPV: 1) What do you experience as barriers to screening for woman abuse?; and 2) What has helped or would help make screening for woman abuse easier for you? These were the final questions on a mailed self-administered 43-item survey that addressed barriers and facilitators to asking about IPV and also included questions related to respondent demographics, education, training, and professional and personal experience with IPV. The survey was mailed to a random sample of 1,000 physicians, weighted by specialty (family practice, emergency medicine, public health, obstetrics and gynaecology) and an unweighted random sample of 1,000 nurses working in family physician offices, emergency care, maternal/newborn and public health. The data were collected from March to June 2004 and analysis of the qualitative data occurred from 2004 until 2009.\n\nA modified Tailored Design Method was used to enhance the response rate \\[[@B20]\\]. Potential respondents received an advance notice advising them of the survey; one week later, the 1,000 nurses and 1,000 physicians were sent a personalized letter of information, the survey and a token gift certificate for an Ontario-wide coffee shop. A reminder letter and a replacement survey were mailed to all potential participants three weeks later. A complete description of the survey development, sampling and methods is available in our previous publication of the quantitative results \\[[@B21]\\].\n\nData management and analysis\n----------------------------\n\nResponses to the two open-ended questions were entered into a word processing document. To ensure data quality, one of the study investigators reviewed all transcribed responses; any questionable responses were identified and another investigator checked these responses with the original submission. Inductive content analysis was used as the method of analysis; that is patterns and categories emerged from the data and were not decided upon before the analysis \\[[@B22]\\]. The investigators independently examined the transcribed data for common categories regarding barriers and facilitators by identifying key words, phrases or concepts used by respondents. The project team vetted the resulting categories. The first author then coded the written responses by discipline i.e., nurse or physician. As each category was analyzed, attention was given to similar and contrasting perspectives, as well as unique viewpoints. Random checks were performed by the second author to ensure that the coding was consistent and that the frequencies identified by categories accurately reflected the responses. Those few coding discrepancies that were identified were resolved through dialogue.\n\nFrequencies were calculated for the commonly identified categories and because of the relatively low number of observations in some of the two-by-two tables, the Fisher's Exact Test was performed to determine statistical significance when examining differences between nurses and physicians \\[[@B23]\\]. Although such an approach to data analysis is infrequent when conducting qualitative analyses, it does reflect the positivist epistemology that underpinned development of the self-administered survey tool \\[[@B24],[@B25]\\].\n\nThe intent of the analysis for this paper was to focus on differences between nurses and physicians and did not include an analysis by practice setting. Some exceptions where practice settings are identified are noted. Ethics approval was obtained from the Research Ethics Board of Western University, London, Canada.\n\nResults\n=======\n\nIn total, 931 individuals returned completed questionnaires; 597 identified themselves as nurses (59.7% response rate) and 328 indicated that they were physicians (32.8% response rate). Six respondents did not specify their discipline. Of the total 931 responses, 769 (82.6%) provided written comments to the two open-ended questions, i.e., 527 nurses (88.3% of nurse respondents), 238 physicians (72.6% of physician respondents) and four who did not identify their discipline. Table\u2009[1](#T1){ref-type=\"table\"} shows the demographic characteristics, training and professional/personal experience of those who provided written comments. This group is very similar to the larger sample of 931 in terms of gender, age, area of practice, formal training and professional and personal experience with woman abuse. Consistent with the full study sample, the majority of those providing comments were female (81.1%, n\u2009=\u2009624), which was primarily driven by the nursing data. Of those physicians who provided comments, the data were fairly evenly split by gender, with slightly more male (53.4%, n\u2009=\u2009127) than female (46.6%, n\u2009=\u2009111) physicians offering written comments. Furthermore, of those who provided written comments, the majority of nurses (61.5%, n =324) and physicians (58.0%, n\u2009=\u2009138) had not received formal IPV training and the majority of both nurses and physicians experienced less than 20 disclosures in the past year.\n\n###### \n\nDemographic characteristics of study sample\n\n  \u00a0                                                     **Full study sample (931)**   **Sample providing comments (769)**   **Nurses with comments (527)\\***   **Physicians with comments (238)\\***\n  ----------------------------------------------------- ----------------------------- ------------------------------------- ---------------------------------- --------------------------------------\n  **Sex**                                                                                                                                                      \n  Female                                                77.6% (722)                   81.1% (624)                           97.2% (512)                        46.6% (111)\n  Male                                                  21.8% (203)                   18.1 (139)                            2.3% (12)                          53.4% (127)\n  Missing                                               0.6% (6)                      0.8% (6)                              \\-\\--                              0\n  **Age (years)**                                                                                                                                              \n  20--29                                                6.9% (64)                     7.4% (57)                             9.7% (51)                          2.1% (5)\n  30--39                                                24.9% (232)                   25.0% (192)                           24.1% (127)                        27.3% (65)\n  40--49                                                33.1% (308)                   32.9% (253)                           32.1% (169)                        35.3% (84)\n  50--59                                                28.8% (268)                   27.7% (213)                           27.7% (146)                        28.2% (67)\n  60+                                                   5.6% (52)                     6.1% (47)                             5.7% (30)                          7.1% (17)\n  Missing                                               0.8% (7)                      0.9% (7)                              \\-\\--                              0\n  **Current area of practice**                                                                                                                                 \n  Family medicine                                       32.2% (300)                   29.3% (225)                           8.7% (46)                          74.8% (178)\n  Emergency medicine                                    21.2% (197)                   21.7% (167)                           27.1% (143)                        9.7% (23)\n  Public health                                         17.8% (166)                   19.0% (146)                           26.9% (142)                        *\\-\\--*\n  OB/gyn/newborn                                        22.6% (210)                   23.8% (181)                           30.7% (162)                        8.8% (21)\n  Retired\u2009+\u2009other                                       4.0% (37)                     4.0% (31)                             5.3% (28)                          \\-\\--\n  Missing                                               2.3% (21)                     2.2% (17)                             1.1% (6)                           4.6% (11)\n  **Any disclosure**                                                                                                                                           \n  Never                                                 23.1% (215)                   22.6% (174)                           30.4% (160)                        5.0% (12)\n  None this year                                        10.3% (96)                    10.4% (80)                            12.1% (64)                         6.7% (16)\n  Less than 20 this year                                63.1% (587)                   63.1% (485)                           54.1% (285)                        83.2% (198)\n  20 or more this year                                  2.5% (23)                     2.9% (22)                             2.1% (11)                          4.6% (11)\n  Missing                                               1.1% (10)                     1.0% (8)                              1.3% (7)                           \\-\\--\n  **Formal intimate partner violence (IPV) training**                                                                                                          \n  No                                                    61.5% (573)                   60.3% (464)                           61.5% (324)                        58.0% (138)\n  Yes                                                   36.7% (342)                   38.2% (294)                           37.0% (195)                        40.8% (97)\n  Missing                                               1.7% (16)                     1.4% (11)                             1.5% (8)                           \\-\\--\n  **Respondent, friend of relative experience**                                                                                                                \n  No                                                    49.7% (463)                   46.4% (367)                           42.7% (225)                        54.2% (129)\n  Yes                                                   48.4% (451)                   52.1% (401)                           55.6% (293)                        45.0% (107)\n  Missing                                               1.8% (17)                     1.4% (11)                             1.7% (9)                           \\-\\--\n\n**Note**: \\* 4 did not indicate if they were a physician or a nurse; the number in the parenthesis is the sample size; \\-\\--: proportion suppressed, based on fewer than 5 observations; OB/gyn/newborn: Obstetrics/gynaecology/care of newborns.\n\nWhile some responses were a single word or a brief phrase, other respondents provided several sentences or a paragraph. Respondents offered both professional and personal experiences, shared their opinions and asked questions. Quotations provided in the text have been purposefully selected to reflect common themes and experiences of both physicians and nurses across practice settings.\n\nBarriers to asking about IPV\n----------------------------\n\nAnalysis of the data from the open-ended questions identified nine categories of barriers (Table\u2009[2](#T2){ref-type=\"table\"}). Overall, the barriers described most often by nurses and physicians were lack of time, behaviours attributed to women living with abuse and lack of training. The top five barriers described by nurses were lack of time (27.3%), lack of training (20.9%), behaviours attributed to women living with abuse (19.9%), partner presence (19.5%) and language/cultural practices (18.4%) while the top five barriers described by physicians were lack of time (46.2%), behaviours attributed to women living with abuse (25.2%), lack of resources (18.9%), language/cultural practices (8.8%) and lack of training (5.5%). Table\u2009[3](#T3){ref-type=\"table\"} summarizes the top five barriers by discipline.\n\n###### \n\nBarriers to asking about IPV described by nurses and physicians (n\u2009=\u2009769)\n\n  **Category**                                         **All Respondents (n\u2009=\u2009769)**   **Nurses (n\u2009=\u2009527)**   **Physicians (n\u2009=\u2009238)**\n  ---------------------------------------------------- ------------------------------- ---------------------- --------------------------\n  1 Lack of time^a^                                    33.0% (254)                     27.3% (144)            46.2% (110)\n  2 Behaviours attributed to women living with abuse   21.5% (165)                     19.9% (105)            25.2% (60)\n  3 Lack of training^a^                                16.4% (126)^b^                  20.9% (110)            5.5% (13)\n  4 Language/cultural practices^a^                     15.5% (119)^c^                  18.4% (97)             8.8% (21)\n  5 Partner presence^a^                                14.0% (108)                     19.5% (103)            2.1% (5)\n  6 Lack of resources^a^                               13.0% (100)                     10.4% (55)             18.9% (45)\n  7 Lack of space/privacy^a^                           8.2% (63)                       11.2% (59)             1.7% (4)\n  8 Discomfort with topic                              4.6% (35)                       4.7% (25)              4.2% (10)\n  9 Lack of practitioner knowledge of resources        2.6% (20)                       2.7% (14)              2.5% (6)\n\n**Note**:\n\n*IPV* Intimate Partner Violence; ^a^ significant differences between nurses and physicians, Fisher Exact Test, p\u2264 0.005; ^b^ three respondents did not identify discipline; ^c^ one respondent did not identify discipline.\n\n###### \n\nTop five barriers by discipline\n\n  **Nurses (n\u2009=\u2009527)**                                         **Physicians (n\u2009=\u2009238)**\n  ------------------------------------------------------------ ------------------------------------------------------------\n  1 Lack of time (27.3%)                                       1 Lack of time (46.2%)\n  2 Lack of training (20.9%)                                   2 Behaviours attributed to women living with abuse (25.2%)\n  3 Behaviours attributed to women living with abuse (19.9%)   3 Lack of resources (18.9%)\n  4 Partner presence (19.5%)                                   4 Language/cultural practices (8.8%)\n  5 Language/cultural practices (18.4%)                        5 Lack of training (5.5%)\n\nTwo physicians used the metaphor of *opening up a can of worms* when describing barriers. This analogy likely highlights the multi-faceted complexities faced by practitioners and the potential challenges that the respondents may have experienced when asked how much they agreed with the statements on the 43-item mailed survey regarding possible barriers to asking about IPV. Further details regarding barriers are profiled below.\n\n### Lack of time\n\nAs shown in Table\u2009[2](#T2){ref-type=\"table\"}, lack of time was cited as the most frequent barrier by both nurses (N) and physicians (P), yet physicians described this barrier significantly more often than nurses (physicians: 46.2%; nurses 27.3%).\n\nOne physician noted:\n\n\\\"*Tend to be lengthy issues...therefore very difficult to address these with the time constraints of a busy ER* \\[Emergency Room\\]*. Also I will never see the patient again so difficult to develop appropriate patient/doctor relationships with such a major issue in a brief time period.* (P)\\\"\n\nOften, lack of time was associated with other factors such as heavy workloads, the time required to adequately deal with the issue and the demands of a busy workplace as this nurse shared:\n\n\\\"*Time issues. If you are going to ask, you have to have the time to listen to the response and deal with the issue.* (N)\\\"\n\nAnother nurse also wrote about the personal impact:\n\n\\\"*We have very little time for patient care. Emotional support is not always there, heavy workload and fast paced environment. I walk away from a shift feeling I really have not done the best job for some of the women.* (N)\\\"\n\n### Behaviours attributed to women living with abuse\n\nSimilar proportions of physicians (25.2%) and nurses (19.9%) described the behaviours attributed to women living with abuse as a barrier. Some respondents expressed a sense of frustration if a woman had disclosed abuse and then stayed or returned to the abusive partner. This was especially challenging when the practitioner had invested effort in developing a relationship as this physician wrote:\n\n\\\"*I have been repeatedly frustrated by women who, after I have taken the trouble to provide alternatives for them, have 'backed down' and returned to their abusive partners.* (P)\\\"\n\nOthers seemed to be more confused than frustrated by such behaviour as this nurse shared:\n\n\\\"*I have heard women say 'but I love him'. I don't understand that thinking especially when children are involved.* (N)\\\"\n\nHowever, others appeared more judgmental and critical as evident by the following responses:\n\n\\\"*I find they defend their partner or don't want what I am offering.* (N)\\\"\n\n\\\"*They don't disclose and are reluctant or unable to help themselves.* (P)\\\"\n\n\\\"*Certain clients don't talk and refuse to do anything about it.* (P)\\\"\n\nIn contrast, a physician wrote the following comment capturing the secrecy and the layers of complexity associated with IPV:\n\n\\\"*Some women have great 'masks'- trying to peel off these 'masks' so as to see the true problem can be difficult.* (P)\\\"\n\nSome respondents indicated that such a lack of action by abused women could be linked to feelings of social stigma, fear, a personal sense of embarrassment or failure and a perception that the abuse is deserved. Seven nurses from different practice settings specifically commented that women might deny abuse since they were concerned that a disclosure could result in the apprehension of their children by child protective services. One nurse provided the following comment about the impact on her practice of her interpretation of the legislation regarding her duty to report situations where children are perceived to be at risk of abuse to an Ontario child protective agency (generally referred to as the Children's Aid Society or CAS):\n\n\\\"*I feel****obliged****to tell women that if they****now****disclose violence****we****must call CAS. I think that this is a****huge****barrier to what we are trying to achieve- safety of the woman and in this way safety of the child. I believe many health workers have changed practice due to this legal requirement... I believe this law has set this issue back 10\u2009years.* (N) - emphasis in original\\\"\n\nThis nurse's description of her experiences captures the complexities of IPV inquiry and reflects that IPV is not owned solely by the health care sector but rather requires links with multiple sectors including child welfare and law enforcement. This written response also profiles the internal conflicts encountered by practitioners when balancing professional expectations and the personal face of IPV with a desire to avoid potentially undesirable consequences and making the situation worse for the woman. There is also a sense of a \"conveyor belt\" syndrome that once events are set in motion there is no turning back.\n\n### Lack of training\n\nA significantly lower proportion of physicians (5.5%) than nurses (20.9%) described a lack of training as a barrier. Nevertheless, as evident by the following written responses, some nurses and physicians were very open about their lack of training and knowledge often linking the need for additional training to fears of offending women, wanting to know how and when to initiate the topic and what to do following disclosure:\n\n\\\"*Knowing when and where to ask, without offending the woman.* (N)\\\"\n\n\\\"*Not being afraid to ask. Being candid - ask questions simply and do not tread on eggshells.* (P)\\\"\n\n### Partner presence\n\nThe abusive partner's tendency to stay by the woman's side was described as a barrier significantly more by nurses (19.5%) than by physicians (2.1%). Nurses wrote that partner presence made it especially challenging to interview the woman alone. This observation was not limited to the hospital sector and was noted for home visits, prenatal and ED visits as well as for labour and delivery.\n\n### Lack of space and privacy\n\nSignificantly more nurses (11.2%) than physicians (1.7%) described lack of space and privacy as barriers. This included the presence of other family members as well as other patients and work environments that were not conducive to confidential one-on-one interviews. Considering both partner presence and lack of space and privacy suggest that these are substantial barriers for nurses.\n\n### Language and cultural practices\n\nA significantly greater percentage of nurses (18.4%) than physicians (8.8%) cited language and cultural practices as barriers. Frequently, the written responses from both nurses and physicians for this barrier were brief (e.g., *cultural differences/language)*. Others provided more descriptive comments as evident by the following:\n\n\\\"*Many women in the practice where I work would not admit to physical or mental abuse as it appears to be accepted in their culture and no amount of questioning will get them to admit it.* (N)\\\"\n\n\\\"*Cultural difficulties- fear of inadequate protection from the police. Therefore many instances, women will decline anything more than just reporting to me.* (P)\\\"\n\n### Lack of quality resources\n\nA significantly greater percentage of physicians (18.9%) than nurses (10.4%) described the lack of quality resources as a barrier as evident by the following responses:\n\n\\\"*Ability to get help and treatment quickly and safely.* (P)\\\"\n\n\\\"*No real effective support both in psychosocial and legal.* (P)\\\"\n\n\\\"*Once we have identified the problem, there are serious dangers in women leaving the situation and not enough community resources to help them if family and friends cannot help them.* (P)\\\"\n\n\\\"*In the North there are no resources or ways out for these women and I spent a lot of time listening to them. That was all we could provide and it was quite sad.* (N)\\\"\n\nConfidential and culturally sensitive interpreters and other specialized services for immigrants were described as resources that were often missing. Although lack of resources was described as a barrier, very few physicians and nurses described lack of knowledge of resources as a barrier (physicians 2.5%; nurses 2.7%).\n\n### Rural settings\n\nIssues related to providing care in rural settings were described by a few respondents as a barrier to asking about IPV. One physician referred to *lack of resources in small rural setting. \\[sic\\] \\[s\\]mall town attitude*, while another physician described the small community as a facilitator: *I often know of people's problems before they come to me.* Another respondent seemed to question the perception that urban centres have better resources.\n\n\\\"*I work in an urban ER and it's great to be able to identify and approach women at risk but it seems that (sometimes) the support system is not available to help her when she needs it.* (N)\\\"\n\n### Personal discomfort\n\nVery few respondents described their own personal discomfort with IPV as a barrier (nurses 4.7%; physicians 4.2%).\n\n### Inadequate management support\n\nIn addition to the barriers noted in Table\u2009[2](#T2){ref-type=\"table\"}, several nurses referenced inadequate management support as a barrier.\n\n\\\"*Administrators who do not see the importance of this as a social broad determinant of health issue - concerned about backed up clinic schedules and the cost of time involved*. (N)\\\"\n\n\\\"*I feel very comfortable screening and do screen in my scope of practice, however this is not supported by my manager in the agency I am employed by. Therefore, we do not have a consistent tool and do not receive training. (N)*\\\"\n\n\\\"*Many abusers complain about the nurses and care and we have little support from management. We end up having to write up the situations. The fear of being reported as more complaints are being sent to the CNO (*College of Nurses of Ontario) *is now a factor.*(N)\\\"\n\nRecognizing the differences in employment models, organizational structures and reporting relationships experienced by nurses and physicians, it is not surprising that this barrier was described by nurses and not physicians. These experiences suggest that management plays a critical role in supporting or hindering IPV inquiry by nurses.\n\nIn summary, as noted in Table\u2009[2](#T2){ref-type=\"table\"}, there were statistically significant differences between nurses and physicians regarding the following barriers: lack of time; lack of training; language/cultural practices; partner presence; lack of resources; and lack of space/privacy.\n\nFacilitators to asking about IPV\n--------------------------------\n\nAnalysis of the data from the open-ended questions identified eight categories of facilitators, some of which differed by discipline (Table\u2009[4](#T4){ref-type=\"table\"}). The top facilitators for nurses were training (47.8%), community resources/professional supports (25.2%) and professional tools/protocols/policies (20.3%), while the most frequent facilitators reported by physicians were community resources/professional supports (21.8%), training (17.6%) and client educational materials (13.0%). Table\u2009[5](#T5){ref-type=\"table\"} summarizes the top five facilitators by discipline. Further details regarding facilitators are outlined below.\n\n###### \n\nFacilitators to asking about IPV described by nurses and physicians (n\u2009=\u2009769)\n\n  **Category**                                  **All respondents (n\u2009=\u2009769)**   **Nurses (n\u2009=\u2009527)**   **Physicians (n\u2009=\u2009238)**\n  --------------------------------------------- ------------------------------- ---------------------- --------------------------\n  1 Training^a^                                 38.5% (296)^b^                  47.8% (252)            17.6% (42)\n  2 Community resources/professional supports   24.1% (185)                     25.2% (133)            21.8% (52)\n  3 Professional tools/protocols/policies^a^    17.3% (133)                     20.3% (107)            10.9% (26)\n  4 Client educational materials^a^             8.6% (66)                       6.6% (35)              13.0% (31)\n  5 Routine screening                           8.6% (66)                       9.5% (50)              6.7% (16)\n  6 Having time                                 6.2% (48)                       6.5% (34)              5.9% (14)\n  7 IPV experience: professional & personal     6.1% (47)                       7.2% (38)              3.8% (9)\n  8 Societal awareness                          6.0% (46)                       6.1% (32)              5.9% (14)\n\n**Note**:\n\n*IPV* Intimate Partner Violence; ^a^ significant differences between nurses and physicians, Fisher Exact Test, p\u2264 0.005; ^b^ two respondents did not identify discipline.\n\n###### \n\nTop five facilitators by discipline\n\n  **Nurses (n\u2009=\u2009527)**   **Physicians (n\u2009=\u2009238)**                                \n  ---------------------- --------------------------------------------------- --- ---------------------------------------------------\n  1                      Training (47.8%)                                    1   Community resources/professional supports (21.8%)\n  2                      Community resources/professional supports (25.2%)   2   Training (17.6%)\n  3                      Professional tools, protocols/policies (20.3%)      3   Client educational materials (13.0%)\n  4                      Routine screening (9.5%)                            4   Professional tools/protocols/policies (10.9%)\n  5                      IPV experience: professional & personal (7.2%)      5   Routine screening (6.7%)\n\n### Training\n\nA significantly greater percentage of nurses (47.8%) described training as a facilitator in comparison to physicians (17.6%). Overall, the respondents clearly described how they wanted to learn about IPV. The approaches described included: receiving literature and written material; developing and disseminating \"best practices\"; using \"real-life\" scenarios; role playing; talking with colleagues; having opportunities to practice newly learned skills with others who have greater comfort and experience with the topic; having discussions with women who are survivors; and touring women's shelters. Two nurses commented on the responsibility of the employer to offer training at the worksite, while another acknowledged the challenges of shift work and being able to participate in educational sessions.\n\n### Community resources and professional supports\n\nBoth physicians (21.8%) and nurses (25.2%) cited the availability of community resources and professional supports as facilitators for them as practitioners. Examples described as facilitators included: inventories of local resources; access to colleagues for consultation; being part of a multidisciplinary team; and the availability of staff with specialized expertise for consultation. References to other disciplines or services focused primarily on social workers and Children's Aid Society staff. Social workers clearly emerged as playing a critical role, one respected by both physicians and nurses. Despite the favourable comments regarding the role played by social workers, not all practitioners felt supported 24/7*.*\n\n### Professional tools, protocols and policies\n\nSignificantly more nurses (20.3%) described tools, protocols or policies as facilitators to support screening as a facilitator, whereas this was described by only 10.9% of physicians. Routine screening was mentioned by 9.5% of nurses and 6.7% of physicians and was believed to make asking about IPV easier. Some respondents were working in settings where this was an expectation and others recommended that routine screening be instituted. These results included 15 nurses who specifically recommended that screening be *mandatory/required/obligatory* and one physician, who advocated that screening be *recommended by law*. In contrast, one nurse recommended strongly against universal screening.\n\n\\\"*It is an unrealistic expectation to be screening every woman/female over the age of 12 for abuse on each encounter with health care personnel as the guidelines*\\[[@B26]\\]*have suggested. This expectation turned many emergency room nurses against screening when it was first introduced and compliance remains very low.* (N)\\\"\n\n### IPV experience\n\nBoth professional and personal experiences were described by 7.2% of nurses and 3.8% of physicians as being facilitators to asking about IPV. Thirteen nurses and one physician shared that they had personally experienced abuse in their own lives and that this experience enhanced their capacity to work with abused women as evident by the following written comments:\n\n\\\"*My personal experience with abuse provides me with a comfort level, knowledge of the system and a desire to support and empower women.* (N)\\\"\n\n\\\"*The fact that I have been a victim of domestic violence and abuse makes it easier for me to identify women who are experiencing a similar situation.* (N)\\\"\n\n\\\"*My own experience has helped. Having the time in my own life to deal with my own issues has helped the most. I cannot imagine that you can teach this to someone - it is so intricate and complicated.* (P)\\\"\n\nOne nurse who acknowledged her inexperience with abuse described how she relied on a colleague who had personally experienced abuse.\n\n### Educational materials\n\nA significantly greater percentage of physicians (13.0%) described educational materials for women as a facilitator, while this was the case for only 6.6% of nurses. This difference may be accounted for by the difference in roles, where the physician was more likely to experience disclosures and depending on the practice setting assume responsibility for initiating care.\n\nSome respondents acknowledged an increased awareness of IPV in the larger community, yet were concerned that more awareness is needed. Only 6.1% of nurses and 5.9% of physicians described societal awareness as a facilitator. As an example, one nurse commented: *It will always remain a serious problem until more information e.g. ads, commercials etc. show women it is not their fault.*\n\nPublic awareness campaigns, as well as community forums and meetings, were recognized for the key role they play in promoting awareness. Several respondents referred to the role that the media, such as television, radio and print play. The media was recognized for its critical role in promoting awareness of both IPV as well as available community services. Furthermore, the media was seen as a mechanism to normalize routine questioning by practitioners regarding abuse, so women are not surprised or offended by such questions. Mandatory education in schools was also recommended as a strategy to enhance greater societal awareness.\n\n### Time\n\nConsistent with the finding that lack of time was the most frequently described barrier by both physicians (46.2%) and nurses (27.3%), only 5.9% of physicians and 6.5% of nurses cited having time as a facilitator.\n\nIn summary, Table\u2009[4](#T4){ref-type=\"table\"} highlights several statistically significant differences between nurses and physicians regarding the following facilitators: training; professional tools/protocols/policies; and client educational materials.\n\nDiscussion\n==========\n\nThe majority of 931 physicians and nurses completing a survey regarding their experience with IPV inquiry provided written responses to open-ended questions. These questions may have encouraged respondents to reflect and share their personal experiences. Perhaps respondents were also indicating that the short statements regarding barriers and facilitators included in the 43-item survey were too simplistic and did not fully capture the complexities of the issues they face related to IPV.\n\nOverall, the top barriers to asking about IPV were lack of time, behaviours attributed to women living with abuse, lack of training, language/cultural practices and partner presence, while the facilitators cited most often included training, community resources/professional supports, and professional tools/protocols/policies. The statistically significant differences between nurses and physicians regarding both barriers and facilitators are most likely related to differences in roles and work environments.\n\nAfter reviewing the comments, the study investigators were left with the impression that this is a very emotionally charged and complex practice issue for both nurses and physicians. Many nurses and physicians in this study continued to struggle with IPV inquiry. The sharing of personal stories -- even in this self-administered, written format, was especially poignant and a reminder of additional burdens that some practitioners face related to IPV.\n\nThe two images of *peeling off a mask* and *opening a can of worms* further suggest the perceived complexities of this issue. The analogy of opening up a can of worms (or a \"Pandora's Box\"), also noted by Sugg and Inui \\[[@B13]\\] and McCauley and colleagues \\[[@B27]\\] implies a sense of unpredictability and concern about having the necessary time and skills to deal with the many issues associated with abuse inquiry and disclosure. It is possible that for those who operate from a results-driven model of care, where actions are expected to solve problems, the inability to control a situation or the outcome, can be personally and professionally intimidating or frustrating. Understanding and accepting the lived experience of abused women, which may challenge practitioner logic, can be emotionally draining, while customizing care to unique circumstances and searching out resources may be time consuming and challenging. The image of peeling off a mask illustrates the challenges faced by some practitioners. It highlights the energy and time required of practitioners, the intensity of the experience for both abused women and practitioners and the secretiveness of IPV as a societal and practice issue. Such metaphors challenge us to explore approaches at multiple levels (i.e., practitioner, practice setting, workplace and community) and to ensure that practitioners have the necessary skill sets as well as on-going education and supports in their work environment.\n\nGiven that dealing with complex practice issues, is by definition, typically time consuming, the authors were not surprised that lack of time was the most frequently described barrier for nurses and physicians. Insufficient time is often cited as a barrier for a variety of practice issues. Lack of time is an important factor in health care environments, yet it can be a quick and almost automatic, impersonal response when asked to identify possible barriers. Even though very few nurses and physicians described personal discomfort as a barrier, focusing on lack of time may mask other barriers that may be more challenging for practitioners to address such as feelings of frustration, a sense of futility or helplessness about how best to respond.\n\nIt is noteworthy that of the barriers described by respondents in this study, 51.0% were attributed to the abused women themselves (57.8% nurses; 36.1% physicians) suggesting that practitioners described fewer barriers related to their own behaviours. Their frustration with women choosing not to accept their advice or returning to partners after leaving, in and of itself, suggests the need for more education with respect to the complex dynamics of IPV.\n\nAs noted in Table\u2009[1](#T1){ref-type=\"table\"}, formal IPV training is not common and others have found that this kind of basic education is not the norm \\[[@B28],[@B29]\\]. This reported lack of formal training points to potential curriculum gaps and the need for continuous learning opportunities in the work place. Both nurses and physicians indicated that they wanted to know how to introduce the topic and what to do if the woman discloses. Challenges in engaging women in such sensitive discussions have been noted by other investigators \\[[@B13],[@B14],[@B30]\\]. It is recognized that training can heighten sensitivity and enhance awareness. Hence, feelings of frustration expressed by both nurses and physicians accompanied by attitudes that appear to blame women may be consistent with low rates of formal training. Furthermore, it is not surprising that with the low rates of formal training, 30.4% of nurses had never heard a disclosure compared to only 5.0% of physicians (Table\u2009[1](#T1){ref-type=\"table\"}). However, recognizing that reports of formal IPV training were low for both nurses (37.0%) and physicians (40.8%), the stark differences in disclosures between nurses and physicians may be explained by role differences, patient behaviour, the nature of the patient interaction and the practice setting, especially in terms of safe and confidential spaces for these discussions. Nonetheless, this finding is alarming given the prevalence of woman abuse and that respondents were recruited from practice areas where IPV is likely to be encountered.\n\nThe seemingly greater emphasis on training by nurses requires further inquiry. This finding is interesting considering that the proportion of physicians (58.0%) and nurses (61.5%) who had no formal IPV training is similar. Many respondents, especially nurses, recommended training during their formal educational programs, continuing professional education, the opportunity to practice such skills and the opportunity to learn from the experiences of others. Greater emphasis on training by nurses may be related to role differences in client interactions depending on the discipline, and/or lack of experience with disclosure. Although training can result in greater knowledge, confidence and skill development, it is likely overly simplistic to suggest that training without periodic refreshers, other structural supports and organizational policies will result in significant practice changes \\[[@B15]\\]. This may be a practice issue that needs to be repeatedly addressed in a supportive practice setting to be understood and one that is very challenging to learn through simulated or theoretical experiences.\n\nThese findings suggest that in order not to blame women and to better address barriers to IPV inquiry, nurses and physicians need greater understanding of the complex dynamics and contextual factors that result in women continuing to deny abuse, not following through on intended actions or returning to abusive partners. Roberts et al. cite multiple stressors that abused women frequently encounter, including financial challenges, child custody battles, a sense of fear, altered social supports and feelings associated with the loss of emotional attachments with their abusers \\[[@B31]\\]. Furthermore, these authors suggest that abused women need to balance the rewards of leaving with the costs of this decision. Bennett and colleagues refer to the challenges that abused women often encounter when entering the justice system, such as a sense of confusion with the process, time delays, minimal information and the need for multiple appointments at a time when access to resources including child care, transportation and finances, are reduced \\[[@B32]\\]. Moreover, Bonomi et al. refer to both subtle and blatant pressures that abusers may direct towards abused women \\[[@B33]\\]. The impact of these cumulative stressors may be initially downplayed by abused women until experienced first-hand, resulting in the increased likelihood that women return to abusive relationships \\[[@B31]\\].\n\nThe impact on abuse disclosure and practitioner behaviour of legislation, that requires practitioners to contact child protective agencies when child abuse is suspected, is an area requiring further study. While further research about how best to address barriers to IPV inquiry faced by nurses and physicians is warranted, some actions in the workplace can be instituted. The potential challenges of collaborating across sectors can be minimized by developing strong relationships and an understanding of other sector roles and responsibilities. Providing training and regular updates, facilitating quick and easy access for consultations and the opportunity to discuss individual situations are important supportive strategies rather than leaving practitioners to face such decision points alone.\n\nOverall, the nurses and physicians cited similar barriers and facilitators; however there were important differences in responses between the two disciplines. Some differences, such as language barriers, partner presence and lack of privacy, which were described more often by nurses, may be related to differences in role expectations and work environments between nurses and physicians. Similarly, because nurses are more likely to be employees, management support and agency policy are important factors for nurses when dealing with IPV. Knowing the types of barriers experienced by practitioners can influence the need for workplace policies and the development of pre-service and in-service training materials. Such resources can be customized by discipline, while customization by practice setting is an area for future study.\n\nOur findings are consistent with previous studies of barriers and facilitators to asking about IPV, in particular barriers such as lack of provider education, frustration, lack of time and fear of offending a patient exposed to IPV \\[[@B13],[@B14],[@B18]\\]. It is disappointing that many of the barriers to abuse inquiry have not changed in the last 15\u2009years. However, this finding suggests that approaches designed to assist practitioners in handling these commonly cited barriers may rest in the synergistic nature of the barriers and such approaches should not attempt to deal with barriers one at a time. A systems perspective that recognizes the complexity and inter-connectedness across barriers may yield more promising results.\n\nIn addition to corroborating the results of past studies, this study provides further insight into why some barriers have not changed. Lack of clinician confidence is not surprising, given the paucity of evidence-based interventions to which abused women can be referred \\[[@B34]\\]. IPV is a complex and multidimensional issue that is not solely the domain of the health care sector. Interprofessional and cross-sectoral collaboration is required. Expressions of professional challenges in being able to adequately meet the needs of abused women, and feelings of frustration and concerns about not knowing what to do if women disclose, suggest that more is needed to support practitioners. In order to navigate the complexities that surround asking about IPV, practitioners require specific personal skills and knowledge, access to community resources and work environments that encourage the development and sustainability of these skills. Mentoring, coaching and opportunities for debriefing and reflective practice are important supports for sustaining practitioner capacity for IPV inquiry.\n\nAlthough barriers and facilitators were categorized, it is evident from the analysis that there is some overlap across categories. This suggests the need for comprehensive frameworks to explore further the many issues associated with IPV inquiry and the interplay across these issues. Multifaceted i.e., practitioner, workplace, community and patient-centred, intersectoral approaches are warranted. The complexities associated with IPV inquiry require action at many levels and failing to systematically address all levels may yield less than desirable results.\n\nThese findings highlight the need for on-going rigorous research to assist in identifying best practices from the perspective of both abused women and practitioners. In addition, the needs of all women should be considered when developing patient-centred approaches that are sensitive to race, ethnicity, socioeconomic status, religious/spiritual beliefs, age, ability and sexual orientation.\n\nStrengths and limitations of the study\n--------------------------------------\n\nThis study has several strengths. Existing practice directories were used to randomly identify possible participants; responses from these participants provided rich data to explore the themes outlined above. The willingness of these respondents to provide detailed comments suggest that these practitioners were committed to IPV related issues. However, despite the high proportion of physicians (72.6%) who provided written comments among those who responded, only 32.8% (328) of physicians invited to participate in the study returned the survey. This may have limited the identification by physicians of perceived barriers and facilitators. However, while there is, in theory, a greater chance for bias when response rates are low, such response rates do not always mean biased responses \\[[@B35]\\]. Others have noted the challenges to obtaining high response rates to mailed surveys targeting physicians \\[[@B36]\\].\n\nFurthermore, respondents may have been inclined to provide more socially acceptable and brief answers considering the sensitivity of the topic and the method of data collection, yet the volume and the richness of the written responses mitigate these limitations. Although the mailed survey offered respondents anonymity and a sense of privacy in which to share their perspectives and experiences without sanction, there was no opportunity for the researchers to seek clarification or further detail. The participants completed the open-ended questions without the benefit of interviewer prompts and without the stimulus of other participants. Future research would benefit from follow-up interviews which could provide further insights into the differences between nurses and physicians.\n\nConclusions\n===========\n\nIPV is a significant public health concern. This study identified barriers and facilitators to asking about IPV among Ontario, Canada nurses and physicians working in family practice, emergency medicine, public health, obstetrics/gynaecology and newborn care. Exploring why barriers to IPV inquiry continue to persist is important given the stigma associated with IPV, low disclosure rates and the impact of IPV on individuals, families and the larger community. A number of statistically significant differences between nurses and physicians were noted. The most frequently described barriers included lack of time and behaviours attributed to women living with abuse, such as denial and lack of action. The study respondents described professional supports such as the availability of social workers and access to community resources as facilitators. Although training was noted as a facilitator by both physicians and nurses, it was noted more often by nurses.\n\nWhile these findings echo those of previous studies, this research provides new insights into the complex relationships among barriers and facilitators faced by physicians and nurses and why barriers to IPV inquiry have not changed in the past 15\u2009years. This study also highlights the need for multifaceted strategies such as pre-service, service and continuous learning, practitioner supports, supportive work environments and workplace policies that collectively address individual, interpersonal, workplace and systemic issues faced by practitioners in addressing the challenges and complexities associated with IPV inquiry.\n\nAbbreviations\n=============\n\nCAS, Children's Aid Society; ED, Emergency department; IPV, Intimate partner violence; N, Nurse; P, Physician.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nCEB assumed responsibility for the qualitative analysis and prepared the initial draft. IAG conducted the statistical analysis and vetted the qualitative analysis. HLM and CNW obtained study funding. All authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2458/12/473/prepub>\n\nAcknowledgements\n================\n\nWe wish to acknowledge: Dr. Bonnie Lynn Wright, former Nurse Researcher/Educator, Public Health Research, Education & Development (PHRED) Program, Middlesex-London Health Unit (MLHU), who played a lead role in designing the study and developing the survey, and Anita Evans, former Research Assistant, PHRED Program, MLHU, who provided research support in checking the accuracy of the transcribed data and assisted in the initial categorization of the open-ended responses.\n\nThis study was part of the McMaster Violence Against Women Research Program, funded by Echo: Improving Women's Health in Ontario, an agency of the Ontario Ministry of Health and Long-Term Care (formerly the Ontario Women's Health Council) (H. MacMillan, PI). The authors are solely responsible for the design, implementation, interpretation, and reporting of this study.\n\nNadine Wathen is supported by a Canadian Institutes of Health Research (CIHR)-Ontario Women's Health Council New Investigator Award. Harriet MacMillan holds the David R. (Dan) Offord Chair in Child Studies at McMaster University.\n"}
+{"text": "Introduction {#S0001}\n============\n\nThe human genome encodes at least 400 ion channel family members (\\~1.5%^[1](#CIT0001)^), representing the second largest class of membrane proteins for drug discovery after G protein-coupled receptors (GPCRs) ([Figure1(a](#F0001))).^[2](#CIT0002)-[5](#CIT0005)^ Roughly 18% of small molecule drugs listed in the ChEMBL database are targeted towards ion channels,^[5](#CIT0005)^ with global sales estimated to be \\$12 billion.^[6](#CIT0006)^ Although it is well validated that ion channels are at the core of many diseases, approved drugs are available for only a small percentage of this protein class (approx. 8%) despite focused drug discovery efforts over the past 30\u00a0years.^[7](#CIT0007)^ Ion channels function by transporting ions across cell membranes and play important roles in a broad range of physiological and pathophysiological processes. Mutations of single ion channel proteins have been demonstrated to be the cause of genetic diseases, collectively known as channelopathies.^[8](#CIT0008)^ For example, mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) lead to cystic fibrosis, whereas various pain syndromes, including congenital indifference to pain and paroxysmal extreme pain disorder, are associated with either loss or gain of function mutations, respectively, in the *SCN9A* gene encoding the voltage-gated sodium channel Na~v~1.7. Along with direct effects on the functionality of ion channel subunits or the proteins that regulate them, channelopathies can also result from autoimmune responses to channel proteins.^[9](#CIT0009)^10.1080/19420862.2018.1548232-F0001Figure 1.Market opportunities and global clinical pipeline for ion channel drug targets. (a) Market opportunities for targeting ion channels which represent the second largest membrane protein target class after GPCRs, adapted from Santos *et al* 2017.[^5^](#CIT0005) (b) Ion channel drugs in development and the clinical pipeline (sourced from Pharmaprojects as of March/April 2016).\n\nTo date, most ion channel drug development has focused on identifying and developing small molecule and peptide modulators, mainly through serendipitous discovery due to a lack of information on structure and function. Many ion channel modulators have been discovered from studies of naturally occurring substances, such as toxins from plants and venomous animals.^[10](#CIT0010)^ The conotoxin family is the most well-known of the animal-derived toxins,^[11](#CIT0011)^ with ziconotide, a selective Ca~v~2.2 antagonist, a frequently cited example of a synthetic peptide analogue of cone snail \u03c9-conotoxin used for the treatment of severe chronic pain.^[12](#CIT0012)^ Despite the initial successes in identifying ion channel modulators, only two novel ion channel drugs have been approved by the US Food and Drug Administration (FDA) since the 1990s, despite vastly improved screening tools for small molecule/compound libraries.^[13](#CIT0013)^ The most recently approved drugs are ivacaftor (Kalydeco), which potentiates the cystic fibrosis CFTR chloride channel^[14](#CIT0014)^ and crofelemer (Mytesi), a proanthocyanidin oligomer, which inhibits both CFTR and the calcium-activated chloride channel TMEM16A.^[15](#CIT0015)^ As with the vast majority of other drugs targeting ion channels, ivacaftor and crofelemer are both small molecule chemical entities.^[16](#CIT0016)^\n\nAlternative modalities for targeting ion channels have recently included monoclonal antibodies (mAbs), but their therapeutic potential has been vastly underexploited.^[17](#CIT0017)^ An in-house analysis using information gleaned from the public domain revealed that only one antibody drug (a polyclonal or pAb) is in early clinical study among the \\>\u00a0650 ion channel targeting drugs under active development in the global pipeline ([Figure 1(b](#F0001))).\n\nAdvantages of targeting ion channels with antibodies {#S0002}\n====================================================\n\nAlthough therapeutic antibodies are typically more expensive to develop, they generally attain higher approval success rates compared with their small molecule counterparts.^[18](#CIT0018)^ As with antibodies targeting GPCRs,^[19](#CIT0019),[20](#CIT0020)^ antibodies directed towards ion channels have the potential to offer many additional advantages relative to selectivity, bioavailability and effector function as summarized below.\n\nSelectivity {#S0002-S2001}\n-----------\n\nObtaining target selectivity in small molecule drug discovery is one of the foremost technical hurdles for drug development, regardless of the route from which the molecule is derived, i.e., rational design or random screening of large compound libraries. With respect to ion channels, this has been particularly challenging as ion channels within a given family often share high levels of homology, notably within the pore-forming domains where many channel blockers exert their effect, but have vastly different physiological roles. For example, the sodium channel isoforms Na~v~1.7, Na~v~1.8 and Na~v~1.9 have been identified as targets in nociceptor neurons where modulation ameliorates different pain states. However, stringent counter-screens are required to characterize potential modulators of these channels for effects on other Na~v~ family members, such as Na~v~1.5, which initiates the cardiac action potential. Superior specificity and selectivity compared to small molecules are particularly relevant when the desire is to target specific ion channel isoforms, for example, the non-functional variant of P2X7 (nfP2X7),^[21](#CIT0021)^ the neo-natal splice variant of Na~v~1.5 (nNa~v~1.5),^[22](#CIT0022)^ or isoforms of K~v~11.1B that are up-regulated in certain tumors.^[23](#CIT0023),[24](#CIT0024)^ An obvious alternative to small molecule promiscuity is the development of mAbs, where high levels of specificity would be expected to mitigate off-target effects, and therefore generate safer classes of drugs.\n\nBiodistribution, half-life and effector functions {#S0002-S2002}\n-------------------------------------------------\n\nMAbs offer a number of potential benefits beyond selectivity, including 1) limiting central nervous system (CNS) penetration (when targeting a therapeutic to the periphery); 2) low variability in patient pharmacokinetics; and 3) longer duration of action leading to reduced dosing. The half-lives of native antibodies can be further extended through alterations to the variable domain that enhance FcRn-mediated recycling^[25](#CIT0025)^ and for antibody fragments via modification with polyethelene glycol (PEG) (i.e., pegylation)^[26](#CIT0026)^ or binding to human serum albumin.^[27](#CIT0027)^ Other types of protein engineering apply approaches directed to the Fc domain that can be used to ablate or increase antibody-mediated effector functions, such as antibody-dependent cell-mediated cytotoxicity (ADCC), complement mediated cytotoxicity, or antibody-dependent cell-mediated phagocytosis,^[28](#CIT0028)^ which are relevant in the case of autoimmune diseases and cancer. MAbs can also be conjugated to radioisotopes or toxic compounds, or linked to the T-cell receptor (so-called CAR-T technology) to directly kill tumors or elicit T-cell mediated tumor cell destruction, respectively. Given their exquisite specificity, it may also be possible to generate mAbs that recognize different conformational states of an ion channel, such as a depolarization-induced conformational change that may render an epitope more accessible to antibody binding.^[29](#CIT0029)^ In addition, affinity and potency against a target can be further enhanced by well-established protein engineering methodologies for lead mAb optimization.\n\nThe challenges for ion channel antibody drug discovery {#S0003}\n======================================================\n\nDespite considerable interest, only one polyclonal antibody, BIL010t (Biosceptre, Cambridge, UK), which recognizes a non-functional form of P2X7 and is formulated as a topical ointment for the treatment of basal cell carcinoma, has completed Phase 1 clinical trials.^[30](#CIT0030)^ The lack of success in generating such antibodies, particularly mAbs, is attributable to a number of important challenges. For example, for many of the voltage-gated ion channels (VGICs), extracellular loops (where mAbs are most likely to elicit a modulating effect) are short and contain few potential epitopes ([Figure 2](#F0002)). Additionally, these loops tend to be highly conserved at the primary amino acid sequence level, and thus lack sufficient immunogenicity to generate robust antibody responses in mammalian hosts. Even in cases where the extracellular domains are large, the proteins themselves are either poorly expressed or difficult to purify from conventional platforms used for recombinant protein production. This, in turn, can limit the starting material available for large-scale immunization and screening campaigns.10.1080/19420862.2018.1548232-F0002Figure 2.Ion channel extracellular domains can influence the difficulty in generating functional antibodies. A comparison of the structural topology of P2X, acid sensing (ASIC), voltage-gated (VGIC) and transient-receptor potential (TRP) ion channel families is shown with the relative mass of the extracellular domains (ECDs) highlighted by dashed red lines. Structural information was adapted from the Protein Data Bank (PDB) figures for P2X4 (3I5D)[^31^](#CIT0031), VGCC (4MTO)[^32^](#CIT0032), TRPA1 (3J9P)[^33^](#CIT0033) and ASIC1 (6AVE)[^34^](#CIT0034). The plasma membrane is represented by blue horizontal lines. Channels with large ECDs (e.g. P2X and ASIC) are expected to display a proportionally larger epitope target area than channels with much smaller ECDs (e.g. VGIC and TRP) and would therefore present less challenging targets in antibody discovery campaigns. Conversely, VGICs and TRP channels that display much smaller epitope target areas represent more challenging targets.\n\nDespite these challenges, autoantibodies that bind ion channels (and presumably alter their activity) have been identified in patients with a number of diseases, including myasthenia gravis (nAChR),^[35](#CIT0035)^ multiple sclerosis (Kir4.1),^[36](#CIT0036),[37](#CIT0037)^ Lambert-Eaton Myasthenic Syndrome or LEMS (VGCC),^[38](#CIT0038)^ neuromyotonia (K~v~1 family) (voltage-gated potassium channels),^[39](#CIT0039)^ melanoma retinopathy (TRPM1),^[40](#CIT0040)^ autoimmune encephalitis (NMDA),^[41](#CIT0041)^ progressive encephalo-myelitis with rigidity and myoclonus, also known as PERM (glycine receptor)^[42](#CIT0042)^ and Morvan's syndrome (the K~v~1 family of voltage-gated potassium channels).^[43](#CIT0043)^ Some autoantibodies^[44](#CIT0044)^ and at least one mAb^[45](#CIT0045)^ induce ion channel internalization, suggesting that antibody drug conjugates could also be a feasible therapeutic modality for targeting this drug class. Moreover, therapeutic antibodies generated in response to DNA immunization using an expression vector encoding the K~v~1.3 potassium efflux channel have been shown to be effective in ameliorating autoimmune encephalomyelitis in rats, underscoring the validity of antibodies as ion channel drug candidates.^[46](#CIT0046)^\n\nIon channel structural topology {#S0004}\n===============================\n\nThe classification of ion channels can be based upon their ion selectivity, gating mechanism and/or sequence similarity. The ion channel gating mechanism system identifies three main groups, namely the voltage-gated channels, the extracellular ligand-gated channels and channel proteins utilizing other gating mechanisms, such as mechano-sensitive channels. The structural architecture associated with each family of ion channel has been described extensively elsewhere,^[47](#CIT0047)-[52](#CIT0052)^ and is not reviewed here.\n\nAmong the key factors governing successful discovery of antibodies that can modulate ion channel activity are the size, complexity, immunogenicity and mechanistic properties of the extracellular domains where antibodies are expected to bind. The topology of select ion channel family members, as shown in [Figure 2](#F0002), demonstrates stark differences in the size of the extracellular domains and loops relative to the whole ion channel, such as those observed between acid sensing and P2X channels (large extracellular domains) and voltage-gated and TRP channels (small extracellular loops). Owing to the paucity of potential immunogenic extracellular epitopes of the latter group, perhaps it is not surprising that the single antibody drug in clinical development targets an ion channel belonging to the former group (nfP2X7), and is actually polyclonal.\n\nThe challenges noted above have led to targeting specific ion channel extracellular domains with varying levels of success. For example, the E3 re-entrant loop of ion channels comprising six transmembrane (TM) motifs has held a particular interest since this region is thought to maintain positioning of the ion selectivity filter and, at least in some cases, appears to interact with toxins and physiological modulators.^[53](#CIT0053)^ The length and accessibility of the E3 region between the fifth (S5) and sixth (S6) transmembrane domains (TMDs) presents a suitable targeting region for antibodies, and it is rarely post-translationally modified.^[50](#CIT0050)^ The amino acid sequence of channel subtypes can be varied in this region, which also offers the opportunity for isoform-specific interactions to disrupt channel function. Many antibodies reported to have been generated to this region tend to be polyclonal, namely K~v~1.2, K~v~3.1, K~v~10.1, TRPC1, TRPC5, TRPM3, TRPV1, Na~v~1.5, Ca~v~2.1/Ca~v~2.2,^[50](#CIT0050)^ and exhibit functional activity, such as modulation of store-operated or agonist-evoked Ca^2+^ entry,^[54](#CIT0054)-[59](#CIT0059)^ promotion of oligodendrocyte proliferation and migration^[60](#CIT0060)^ and inhibition of tumor growth.^[61](#CIT0061),[62](#CIT0062)^ Additionally, NESOpAb is a polyclonal antibody that specifically recognizes a neonatal epitope presented on the second extracellular loop in Na~v~1.5 domain I and inhibits sodium currents up to 60% with an IC~50~ value of less than 25\u00a0nmol/L. Furthermore, it demonstrates selectivity, being able to distinguish between the neonatal and adult splice variant forms, which differ by seven amino acids.^[63](#CIT0063)^\n\nSources of ion channels for generating and screening ion channel antibodies {#S0005}\n===========================================================================\n\nIon channels are typically low abundance proteins in the cells and tissues in which they are produced. Furthermore, when expressed as recombinant proteins in heterologous systems (e.g., mammalian, insect, yeast and bacterial cells) yields of purified functional protein are often low. Therefore, production of antibodies that can recognize and/or block channel activity has relied primarily on immunization of animals with either: 1) whole cells; 2) crude membrane fractions; 3) plasmid DNA expression vectors encoding channel protein subunits; or 4) peptide-based antigens that preferably mimic a targeted extracellular loop structure. As an alternative to immunization, antibody discovery can also be achieved by screening pre-existing libraries of antibody single-chain variable fragments (scFvs) from na\u00efve or immunized animals via phage or yeast display.^[64](#CIT0064)^ While this latter approach can preclude immunization altogether, the need for purified, correctly folded protein is generally required for the panning and screening phases of the process. Some of the sources of protein used as antigen and to screen for antibody identification are described briefy as follows.\n\nPurified ion channels from native sources {#S0005-S2001}\n-----------------------------------------\n\nPrevious studies have shown that ion channels can be isolated from their native source in a way that maintains functionality of the purified protein following reconstitution into proteoliposomes. For example, a number of laboratories in the 1980s and 1990s purified voltage-gated sodium ion channels from rat brains,^[65](#CIT0065)^ as well as from the electroplax of the electric eel, *Electrophorus electricus*,^[66](#CIT0066)^ and were able to reconstitute functional activity from purified components.^[67](#CIT0067)-[70](#CIT0070)^ Given the high degree of conservation amongst ion channel orthologs, channel proteins from animal sources might therefore serve as antigens and screening reagents to identify antibodies that recognize and modulate their human counterparts. The obvious drawbacks here are the need to obtain sufficient amounts of material from sources whose channel proteins closely match their human orthologs, the typically complicated purifications required to generate that material, and the need to break tolerance in animal hosts being used for immunization.\n\nRecombinant proteins expressed in mammalian cells {#S0005-S2002}\n-------------------------------------------------\n\nMammalian cell lines (e.g., HEK293, CHO, U2OS) are arguably the gold standard for generating recombinant ion channels closest to their 'native' configurations and functionalities. However, as noted above, mammalian cells typically produce low levels of surface localized, functional recombinant ion channels, making them difficult to purify. The use of whole cells or crude cell fractions diminishes the antigenic load of the target protein and introduces additional contaminating proteins that are likely to be more abundant and more immunogenic than the recombinant ion channel of interest, making it difficult to generate an immune response or *in vitro* display output that is sufficiently enriched to effectively screen.\n\nChimeric channels expressed in E. coli {#S0005-S2003}\n--------------------------------------\n\nDespite the phylogenetic divergence between prokaryotes and eukaryotes, it has been possible to generate chimeric bacterial-human ion channels that facilitate protein expression and purification in bacterial host expression systems. For example, a functional chimera in which the extracellular domain (ECD) of the bacterial protein GLIC was fused to the transmembrane domain of the human \u03b11 glycine receptor (\u03b11GlyR) has been reported,^[71](#CIT0071)^ as have functional pentameric ligand-gated ion channel chimeras containing large eukaryotic intracellular domains from nAchR-\u03b17, GABAp1 and Gly\u03b11 fused to the *Gloeobacter violaceus* GLIC channel.^[72](#CIT0072)^ Similarly, the use of bacterial Nav channels has been elegantly exemplified in structural studies of VGICs^[73](#CIT0073)^ and enabled crystallization of a chimeric voltage-gated sodium channel from *Arcobacter butzleri* fused to portions of the human Na~v~1.7 voltage-sensor 4 domain bound to aryl sulphonamide antagonists.^[74](#CIT0074)^ Although chimeric channels could offer a possible solution to the generation of sufficient amounts of antigen and screening reagents to implement antibody discovery processes, they would nevertheless require extensive counter-screening. Moreover, the bacterial elements of the channel may present immunodominant epitopes that could overwhelm the response to the human components of the chimera.\n\nAlternative platforms for recombinant protein expression {#S0005-S2004}\n--------------------------------------------------------\n\nCilated protozoa devote a large part of their metabolism towards membrane protein production and have expanded gene families for all four of the major classes of membrane transporters, including P-type ATPases, major facilitator superfamily members, ABC transporters and voltage-gated ion channels.^[75](#CIT0075)^ *Tetrahymena thermophila*, in particular, has been identified as an attractive platform for over-expression of recombinant human ion channels based on the fact that its macronuclear genome encodes approximately three times as many voltage-dependent K^+^ channels as do human cells.^[75](#CIT0075)^ Although a complex eukaryote, *Tetrahymena* shares many of the features of microbial expression hosts, including ease of growth in peptone-based media at scale with relatively short doubling times of 2 to 3\u00a0hours.^[76](#CIT0076)^\n\nTetraGenetics Inc., an early-stage biotechnology company in Arlington, MA, has demonstrated expression of approximately 20 recombinant human voltage-gated, ligand-gated and mechano-sensitive ion channels in *Tetrahymena* (unpublished data). Efficient recovery of purified channel proteins (in many cases, in the order of \\>\u00a01mg/L culture) has enabled the development of antigen preparations and screening tools that have recently been used to generate a panel of blocking anti-K~v~1.3 antibodies.^[77](#CIT0077)^ Besides the propensity for this organism to encode hundreds of native ion channels, it remains unclear why it bypasses the limiting factors associated with low expression yields in mammalian cells where there appears to be an upper limit on how many functional recombinant channels can reach the plasma membrane before creating a toxic metabolic environment. No such toxicity is observed in *Tetrahymena*, allowing many more recombinant channels to reach the cell surface. In mammalian cells, plasma membrane channel number is likely to be regulated by a variety of mechanisms, including manipulation of various retention signals by auxiliary subunits. *Tetrahymena* does not encode any obvious orthologs of mammalian auxiliary subunits and, in the case of Na^+^-selective voltage gated channels, co-expression of the \u03b2 subunits is not required for cell surface localization in the *Tetrahymena* system (unpublished data). While it is clear that *Tetrahymena* is well suited to the production of recombinant ion channel proteins, a number of other systems are being developed for this purpose, including virus-like particles,^[78](#CIT0078)^ cell-free lysates,^[79](#CIT0079)-[82](#CIT0082)^ synthetic and semi-synthetic chemistries.^[83](#CIT0083),[84](#CIT0084)^\n\nMaintaining the native protein fold -- SMALPs and nanodiscs {#S0005-S2005}\n-----------------------------------------------------------\n\nCharacterization of modulating antibodies following immunization or *in vitro* display methods requires that a purified ion channel protein be maintained in its proper three-dimensional conformation regardless of its source. One approach towards stabiliizing the structure involves incorporation of detergent-solubilzed and purified membrane proteins into nanodiscs that utilize a supporting protein scaffold and lipids to generate an artificial bilayer into which the membrane protein of interest is embedded.^[85](#CIT0085)-[87](#CIT0087)^ Potential limitations to this approach are that transfer of proteins into nanodiscs requires initial solubilization by detergent, as well as reconstitution into a non-native lipid environment. In addition, solubilization of membrane proteins in detergents faces the technical challenge of maintaining a physiologically relevant conformation in addition to stability. An alternative strategy that provides a detergent-free route to membrane protein isolation with retention of the native lipid environment (as much as is possible) is to replace the detergent with amphipathic styrene-maleic acid (SMAs), where the polymer self-inserts into biological membranes and is capable of extracting small discs of the native lipid bilayer containing the membrane protein of interest, generating SMA-encapsulated lipid particles (SMALPs).^[88](#CIT0088)^ More recently, the tetrameric potassium channel KcsA was isolated directly from the membranes of *Escherichia coli* without the need for detergent by using SMALPs.^[89](#CIT0089)^ SMALPs have also been reconstituted into planar lipid bilayers directly from native nanodiscs, which enabled functional characterization of the TRPV1 channel by electrophysiology.^[90](#CIT0090)^ The nanodisc approach was implemented in the reconstitution of tetrameric KirBac1.1 potassium channels into lipid nanodiscs, enabling single-molecule fluorescence resonance energy transfer confocal microscopy, which permitted the elucidation of structural changes that occur upon channel activation and inhibition.^[91](#CIT0091)^\n\nDNA immunization {#S0005-S2006}\n----------------\n\nPlasmid expression vectors encoding ion channel proteins are likely to produce correctly folded and functional antigens following immunization of mammals. Nevertheless, the yield of protein presented at the cell surface may still be low and the resulting immune response may not be sufficiently robust to generate an antibody titer high enough to identify potential modulators. The inclusion of adjuvants and the use of tailored expression vectors with strong promoters, such as the CMV promoter, are often applied in this instance.^[92](#CIT0092)^\u00a0The use of T cell helper epitopes, such as PADRE, are also proving successful.^[46](#CIT0046)^ The DNA immunization approach has been used to generate K~v~1.3 nanobodies using the Ablynx platform,^[93](#CIT0093)^ as well as in combination with purified antigen to generate conventional K~v~1.3 antibodies.^[77](#CIT0077)^ These are described in a later section.\n\nPeptides {#S0005-S2007}\n--------\n\nPeptide antigens have been used to generate functional polyclonal antibodies against multiple ion channels,^[50](#CIT0050)^ as well as mAbs targeting select ion channels.^[61](#CIT0061)^ Peptides usually do not suffer from issues of quantity or purity because they can typically be produced via chemical synthesis or robust cell expression systems, such as *E. coli*, to serve both antigen and screening requirements. However, the physiological relevance of peptide-based antigens, even those that are three-dimensionally accurate representations of surface loop structures, will always be limited as they lack the context of other molecular determinants associated with the ion channel surface 'epitome'.\n\nIon channel antibody generation and screening -- additional considerations {#S0006}\n==========================================================================\n\nWhile the source of ion channel antigen is a critical consideration for any antibody discovery program, the approach to the generation and screening of ion channel antibodies should be scrutinized in the context of the challenges described above. The relative lack of success in the identification and clinical progression of ion channel antibodies suggests that therapeutically valuable antibodies are typically rare and difficult to identify in any given discovery program. Therefore, it would seem prudent to structure a discovery program that can either increase a specific immune response against the target ion channel and/or deeply mine an immune repertoire in an effort to capture as many potential hits as possible. In the case of the former, amino acid conservation and, depending on the ion channel, the topology of the ion channel can affect the generation of a robust immune response. Options that may mitigate this challenge include the use of phylogenetically diverse immune hosts; immunization strategies that break immune tolerance, such as DNA immunization; inclusion of T-helper cell epitopes; transgenic animals overexpressing the neonatal Fc receptor;^[94](#CIT0094)-[96](#CIT0096)^ and, related to this, transgenic animals where the target gene has been deleted. Of course, the latter approach will be dependent on the viability of the knock-out animal. With regards to mining the immune repertoire, a number of platform technologies, e.g., direct B-cell cloning and/or deep sequencing,^[97](#CIT0097)-[99](#CIT0099)^ have been developed recently that increase the probability of identifying rare antibodies and avoid standard hybridoma-based technologies where valuable antibodies may be missed owing to inefficient fusion events or the loss of rare B cell clones.^[100](#CIT0100)^\n\nFollowing the identification of antibodies that specifically recognize the ion channel target, it is important to select extracellular binders by some means to advance these clones into functional characterization. Typically, the method most commonly used to identify extracellular binders is flow cytometry using native or transfected mammalian cell lines expressing the ion channel of interest. This, however, is not necessarily straight-forward because many cell lines with confirmed electrophysiological activity may nevertheless express low surface channel numbers making definitive identification difficult. Alternatively, ELISA assays using peptide and protein fragments comprising various extracellular loops is relatively quick and simple. However, peptides applied in this manner are generally not conformational and false-negative results are likely for antibodies that recognize discontinuous or conformational epitopes.\n\nDepending on the number of hits that are recovered from a given discovery program, it may be feasible to forego the antibody sorting described above and move directly to functional characterization. The accepted gold standard for ion channel functional characterization is patch-clamp electrophysiology, which allows real-time kinetic and pharmacological analysis of the effects of drug molecule candidates. Whilst electrophysiology is the most detailed analytical tool available for ion channel functional modulation and is key in making hit-to-lead candidate determinations, it is resource intensive and has suffered historically from low throughput. Progress has been attained with increasing screening throughput and maintaining accuracy with platforms that utilize robotic multi-patch clamp configurations (Patchliner\u00ae, Nanion Technologies; Qube384, Sophion; PatchXpress, Molecular Devices; IonFlux, Fluxion). Interestingly, some investigators and platform manufacturers have reported instances of compounds (including antibodies) that demonstrate functionality when analyzed by manual patch-clamp, but are inactive when analyzed by automated platforms (Colussi, personal communication). The reasons for this remain unclear, but the possibility of false-negative results may lead to the abandonment of potentially interesting antibodies. Moreover, antibodies may reasonably be assumed to exhibit slower binding and efficacy kinetics compared to small molecules, which should be considered when analyzing electrophysiological currents over the course of 5 to 10\u00a0minutes in which measurements are usually made.\n\nIn addition to electrophysiology, a number of other technologies are available that can offer effective screening of modulating compounds that each have their own advantages and disadvantages. These include flux-based assays that measure the cellular influx or efflux of radioactive Na^+^, Ca^2+^ and Rb^+^ for studying sodium, calcium and potassium channels, respectively, and fluorescence-based assays that utilize either voltage-sensitive dyes that measure cell membrane voltage changes or ion-specific fluorescent probes that measure changes in intracellular ion concentrations.^[101](#CIT0101)^ A recent and more detailed review of ion channel antibody screening strategies can be found in Colley et al.^[102](#CIT0102)^\n\nIon channel therapeutic opportunities {#S0007}\n=====================================\n\nThe wide range of physiological processes involving ion channels can be broadly summarized as follows: maintenance of cell resting potential, conductance of electrical signals, synaptic transmission at nerve terminals, intracellular transfer of ions and metabolites, cell volume regulation, excitation-contraction coupling and stimulation-secretion coupling, such as that involved in the release of insulin from the pancreas.\n\nUsing information in the public domain, we analyzed the ion channel target landscape against which antibody therapeutics could have potential in the treatment of disease, excluding those that would require CNS penetration or intracellular distribution. This analysis identified over 150 potential antibody targets, with over 35 of those possessing clinical or preclinical levels of validation from small molecule and peptide studies. Many opportunities fall within the oncology, autoimmune and inflammatory/neuropathic pain (including migraine) therapeutic areas ([Figure 3](#F0003)), but there is also significant potential in the respiratory, metabolic and rare or neglected disease areas. Several ion channel targets fall into more than one therapeutic indication.^[103](#CIT0103)^ Our findings are summarized in [Table 1](#T0001). The main subclasses of targets are voltage-gated and calcium-activated potassium ion channels, voltage-gated calcium and sodium channels, acid-sensing ion channels, transient receptor potential (TRP) channels, purinergic P2X channels, calcium-release activated channels and chloride channels, which are discussed below.10.1080/19420862.2018.1548232-T0001Table 1.Examples of ion channel therapeutic opportunities with level of validation attained by different drug entities, or associated biology, including genetic evidence, knock-out models, etc.Ion channelTherapeutic Area/IndicationModality &/or EntityIn vitro validationIn vivo validation/preclinicalClinical validationReferenceK~v~1.3MS, RA, T1D, atopic dermatitis, uveitis, DED, psoriasis myositis, cutaneous lupus, psoriatic arthritis, IBD, allergic asthmaAntagonist -- peptide analogs of ShK toxin, e.g., dalazatideInhibition of T~EM~ cell proliferation and migration, IL-2 secretion, Ca^2+^ signalling, inhibition of K~v~1.3 currents, inhibition of CD3-antibody- and alloantigen-induced proliferationInhibition of T~EM~ cell proliferation, blocking K~v~1.3 in psoriasiform SCID mouse model, efficacy in DTH and EAE rat models.\\\nClears viral and bacterial infectionsValidation in DED from T cells isolated from patient tissue; suppression of chemokine-induced migration of peripheral blood T cells isolated from healthy donors Dalazatide Ph1 & Ph2[77](#CIT0077),[93](#CIT0093),[104](#CIT0104)-[112](#CIT0112)Atopic dermatitis, psoriasisAntagonist -- small molecule, e.g., PAP1Blocking of K~v~1.3 currents.\\\nSignificant dose-dependent inhibition of proliferation and suppression of IL-2 and IFN-\u03b3 productionPotent suppression of oxazolone-induced inflammation by inhibiting the infiltration of CD8\u00a0+\u00a0T cells in rat allergic contact dermatitis model; significant clinical and histological improvement of plaques in SCID mouse-psoriasis skin xenograft model with reduction in T~EM~ cellsPatient psoriatic plaques enriched in T~EM~ cells\\\nPh1 (inactive)[108](#CIT0108)-[110](#CIT0110),[113](#CIT0113)K~v~10.2Brain cancer, lung and cervicalAntagonist -- small molecule, e.g., TDZInduction of caspase-dependent apoptosis and cell cycle arrestReduction in xenografted MB growth and metastasis, inhibition of balbc/c nude mouse xenografts established using A549 sphere cellsCase report of MB patient demonstrated therapeutic efficacy although not without side effects[114](#CIT0114)-[116](#CIT0116)K~v~11.1BSome cancers (leukemias, gastric, colon)Antagonist -- small molecule, e.g., CD-60,130Reduction in cell proliferation of tumor cells and tumor cell invasiveness, reduction in VEGF secretionReduced leukemic cell infiltration in NOD/SCID and higher survival ratesEpigenetically silenced in ovarian cancer[24](#CIT0024),[117](#CIT0117)-[122](#CIT0122)K~Ca~3.1Autoimmune, e.g., IBD, MS, RA, asthma, fibrosis, sickle cell anemiaAntagonist -- small molecule, e.g., TRAM-34, NS6180, SenicapocGenetic knockdown of K~Ca~3.1 suppresses T cell activationNo toxicities observed. K~Ca~3.1 blockers validated in a number of animal models, e.g., rodent EAE and experimental colitis modelsRestores corticosteroid sensitivity in cytokine-treated ASM cells from COPD and asthmatic patients\\\nPh2 (inactive)[123](#CIT0123)-[133](#CIT0133)Breast, prostate, pancreatic, endometrial, GBM, HNSCC, leukemia, ICC, melanomaCombined activation of K~Ca~3.1 and inhibition of K~v~11.1 -- small molecule, e.g., RiluzoleCisplatin-resistant CRC cells express higher levels of K~Ca~3.1 and K~v~11.1 channels; K~Ca~3.1 activators and\\\nK~v~11.1 inhibitors have a synergistic action with cisplatin in triggering apoptosis and inhibiting\\\nproliferation; TRAM-34 also potentiates response of TMZIn nude mice xenografted with human NSCLC, Senicapoc reduced tumor growth.\\\nIn SCID mice xenografted with human GL-15 glioma cells, TRAM-34 reduced tumor infiltration and astrogliosis surrounding the tumorPh1 NCT01303341\\\nPh2 NCT00866840[134](#CIT0134)-[144](#CIT0144)Ca~v~3.1Breast and prostate cancerAgonistTumor suppressor functionExpression inhibits proliferation and apoptosis of MCF7 cells.\\\nInhibition of prostate cancer cell proliferationMutations in Ca~v~3.1 confer gain-of-function in adenomas[145](#CIT0145)-[147](#CIT0147)Na~v~1.7PainAntagonist -- peptideBlocking of Na~v~1.7 currents but also acts at Ca~v~2.2Synthetic peptides based on spider-derived venom have reversed pain behaviours in mouse models of peripheral spontaneous painGenetic evidence provided by loss-of-function and gain-of-function channelopathies.[9](#CIT0009),[148](#CIT0148)\u00a0Antagonist -- small molecule, e.g., PF-04856264, PF-05089771, CNV1014802/BIIB074Bind preferentially to slow inactivated state of Na~v~1.7, blocks TTX-induced current in DRG neuronsLess selective but more potent with respect to analgesia\\\nOD1 mouse model of Na~v~1.7 mediated painPh2 NCT01529346\\\nNCT0156102, NCT02935608[52](#CIT0052),[149](#CIT0149)-[151](#CIT0151)ASIC1PainAntagonist -- small molecule, e.g, PPC-5650Inhibition of ASIC1 mediated currentsPreclinical cancer models demonstrate nociceptive neuronal expression of ASIC receptors, that respond to a significant increase in an acidic cancer-induced environment within the bonePh1 (inactive)\\\nNCT01818570\\\nNCT01449487\\\nReduction of IBS-related pain[152](#CIT0152)-[154](#CIT0154)TRPA1Pain and inflammationAntagonistSmall molecule in vitro inhibition of AITC-induced\u00a0Ca2+ uptakeFunctional upregulation in OVA-sensitized mice challenged with fine particulate matter\\\nInhibition of TRPA1 eliminates the mechanical and cold allodynia induced by cisplatin and oxaplatin\\\nEndogenous agonists known to cause pain sensation\\\nTRPA1 knockout mice showed higher tolerance to heatPh2 NCT01726413\\\nGenetic evidence demonstrated by TRPA1 gain-of-function mutation linked to FEPS[155](#CIT0155)-[158](#CIT0158)TRPC3AP, SS, hypertension, atherosclerosis, COPDAntagonist -- small molecule, e.g., Pyr3, SalBMutated TRPC3 channels on Jurkat cells show decreased Ca2+ influx after TCR stimulation, which can be rescued by overexpression of wild-type TRPC3Rat model of sepsis demonstrated upregulation of TRPC3 in T cells enhancing T cell apoptosis\\\nTRPC3 inhibition protects salivary glands and pancreas cells from Ca2+ mediated toxicity by inhibiting the TRPC3-mediated component of Ca2+ influxGenetic evidence provided by TRPC6 mutations in FSGS resulting in excessive Ca2+ influx and subsequent injury or loss of podocytes[159](#CIT0159)-[162](#CIT0162)TRPV3Skin health, including inflammation and painAntagonist -- small molecule, e.g., FTP-THQ, GRC15300Inhibits agonist-induced release of ATP and GM-CSF in m308 keratinocytesDose-dependently blocks histamine-induced itch in mouse models\\\nReduction of CFA-induced thermal hypersensitivityGain-of-function TRPV3 mutations identified in rodent and man that are associated with pain and Olmsted syndrome\\\nPh2 (inactive)[163](#CIT0163),[164](#CIT0164)P2X3 and P2X2/3Pain, fibrosis, chronic coughAntagonist -small molecule, e.g., gefapixant, AF-219Block homo- and hetero-trimer formsBlocks peripheral action in afferent neurons when ATP is released causing sensitization to pain signalsPh2 and Ph3 NCT02477709[165](#CIT0165)P2X4PainAntagonist -- small molecule, e.g., NC-2600, ivermectinInhibition of ATP-evoked intracellular Ca2+ influxEfficacy demonstrated in CCI nerve neuropathic pain model and EAN. Dose-dependent inhibition of ATP-induced BDNF releasePh1\\\nhttps://adisinsight.springer.com/drugs/800038339[166](#CIT0166)-[168](#CIT0168)\\\nWO2010/093061P2X7\u00a0Antagonist- -- small molecule, e.g., EVT-401Blocks ATP-induced IL-1\u03b2 release from monocytes.\\\nBlocks P2X7 mediated currentsEfficacy demonstrated in CGN-induced model of inflammation, DNBS-induced model of distal colitis, EAE rodent model and CIA modelPh1\\\nhttps://adisinsight.springer.com/drugs/800025672US20110118287nfP2X7Basal cell carcinomaAntagonist -- antibody, e.g., pAb BIL010tExpression of nfP2X7 in basal cell carcinoma confirmed by IHCLesion size reduced in a mouse model of melanomaPh1 NCT02587819[30](#CIT0030)Orai1 (CRAC)Autoimmune/inflammatory disease, e.g., psoriasis, APAntagonist -- small molecule, e.g., CM-4620Inhibits increase in intracellular Ca2+ in pancreatic acinar cells that leads to enzyme activation, mitochondrial dysfunction, ER stress and necrosisInhibits CRAC pathway in T cells, blocking the release of IL-2 and TNF\u03b1 and reduces neutrophil activationPh1\\\nNCT03401190[169](#CIT0169)Antagonist -- mAbInhibition of T-cell effector function, T cell proliferation and cytokine release. Triggers internalization of Orai1Demonstrates efficacy in rodent T-cell mediated GvHD modelLoss-of-function mutations cause severe immunodeficiency with recurrent infections due to impaired T cell function[170](#CIT0170),[171](#CIT0171)[^1][^2] 10.1080/19420862.2018.1548232-F0003Figure 3.Therapeutic opportunities in the ion channel antibody target landscape shown by therapeutic area. The percentage values in the outer ring represent the number of ion channels implicated for that therapeutic area from the \\>150 potential antibody targets identified. The inner ring depicts each therapeutic area with the number of clinically (in Phase 2 or further development) validated targets in bold font and the number of preclinically validated targets indicated in bold italicized font and bracketed for distinction. In a few instances, an ion channel has presented targeting opportunities in multiple indications within a therapeutic area and therefore different levels of validation have been presented. Therefore, the highest level of validation is taken to avoid duplication, for example, P2X3 in different respiratory conditions. However, where there are ion channels representing a targeting opportunity in multiple therapeutic areas these have been treated separately and accordingly can demonstrate different levels of validation, for example, K~v~1.3 (implicated in autoimmune conditions, such as type 1 diabetes, psoriasis, cutaneuous lupus; respiratory indications (asthma); inflammatory conditions (uveitis and dry eye disease), K~Ca~3.1 (implicated in autoimmune condtions, such as IBD, multiple sclerosis, rheumatoid arthritis; oncology (glioma, renal cancer, NSCLC), respiratory indications (asthma), sickle cell anemia) and TRPC6 (pain; respiratory; metabolic; autoimmune/inflammation; oncology). For further details of the role of each of these ion channels in disease, refer to the main text. There are at least 35 ion channels with clinical or a preclinical level of validation provided by small molecule or peptidic approaches that are suitable for targeting with therapeutic antibodies. Abbreviations: DED dry eye disease; RP retinitis pigmentosa.\n\nVoltage-gated potassium ion channels {#S0007-S2001}\n------------------------------------\n\nThere are at least 40 voltage-gated potassium channels (K~v~) in the human genome with physiological roles.^[172](#CIT0172)^ K~v~ channels represent an ion channel subclass that offers substantial potential for drug development in a range of diseases, such as cancer, autoimmune diseases and metabolic, neurological and cardiovascular conditions. Their roles range from regulating calcium signalling, cell cycle progression, apoptosis and cell volume to driving cellular proliferation and migration, as well as repolarization of neuronal or cardiac action potentials,^[173](#CIT0173)-[178](#CIT0178)^ and thus present the potential for various pharmacological strategies to target K~V~ channels with specific antibodies. The therapeutic potential of selected potassium channels is highlighted below.\n\n### K~v~1.3 {#S0007-S2001-S3001}\n\nK~v~1.3 is encoded by the gene *KCNA3*, expressed on human T cells and was initially recognized as a target for immunosuppression based on the observation that non-selective K^+^ channel blockers can inhibit T cell proliferation and interleukin (IL)-2 secretion.^[104](#CIT0104),[105](#CIT0105)^ It has since been validated as a therapeutic target in numerous preclinical animal models using a variety of small molecule and peptide toxin blockers^[104](#CIT0104),[106](#CIT0106)-[111](#CIT0111)^ and, importantly, in the clinic through the development of the potent venom peptide analog Shk-186 (dalazatide).^[113](#CIT0113)^\n\nK~v~1.3 is the predominant potassium channel expressed on effector memory T cells (T~EM~), which are implicated in a range of T-cell mediated diseases, such as, multiple sclerosis,^[106](#CIT0106)^ rheumatoid arthritis,^[104](#CIT0104)^\u00a0Type-1 diabetes mellitus,^[104](#CIT0104)^ allergic contact dermatitis^[109](#CIT0109)^ and psoriasis.^[108](#CIT0108)^ K~v~1.3 blockers selectively inhibit Ca^2+^ signalling, proliferation and *in vivo* migration of CCR7 negative and positive T~EM~ cells; however, stronger effects are observed on CCR7^\u2212^ T~EM~ cells compared to CCR7^+^ central memory CD4 T cells.^[179](#CIT0179)^ More recently, K~v~1.3 knock-down in memory T cells was shown to suppress CD40L expression and memory phenotype.^[180](#CIT0180)^ CD40L is also a target for autoimmune disorders, and this finding provides further validation of the therapeutic potential of K~v~1.3 blockade in immunomodulation.^[123](#CIT0123)^\n\nThere are no significant intracellular calcium stores in T cells due to their small size, therefore Ca^2+^ influx through CRAC is paramount for NFAT translocation to the nucleus to elicit cytokine secretion and T cell proliferation. The T cell needs to retain a negative membrane potential through a counterbalancing K^+^ efflux via K~v~1.3 and/or the other T cell K^+^ channel, Ca^2+^-activated channel K~Ca~3.1, in order to be fully activated.^[172](#CIT0172)^ Thus, pharmacological inhibition of K~v~1.3 activity blocks activated T-cell proliferation and cytokine production by disrupting the driving force of sustained Ca^2+^ influx via CRAC.\n\nK~v~1.3 is also expressed in breast, prostate and colon cancer and is linked to resistance to apoptosis as observed by the upregulation of K~v~1.3 expression in diffuse large B-cell lymphoma and glioma.^[181](#CIT0181)^ Nevertheless, the role K~v~1.3 plays in proliferation and apoptosis appears to be complex and context dependent, and further studies are required to validate K~v~1.3 as a potential cancer target and biomarker.^[182](#CIT0182)^\n\n### K~v~10.1 {#S0007-S2001-S3002}\n\nA comprehensive overview of the biophysical and pharmacological roles of K~v~10.1 and its potential mechanisms in disease has been described elsewhere.^[183](#CIT0183),[184](#CIT0184)^ K~v~10.1, also known as ether-a-go-go-related gene 1 (EAG1), is encoded by the gene *KCNH1* and expression is predominantly restricted to the CNS in health. Most of the interest in K~v~10.1 as a therapeutic target originates from the observation that it is aberrantly expressed in up to 70% of tumor cell lines and human cancers,^[185](#CIT0185)^ including colon cancer,^[186](#CIT0186)^ gastric cancer,^[187](#CIT0187)^ breast cancer,^[188](#CIT0188),[189](#CIT0189)^ soft tissue sarcoma,^[190](#CIT0190)^ acute myeloid leukaemia (AML),^[62](#CIT0062)^ adenoma,^[191](#CIT0191)^ hepatocarcinoma,^[192](#CIT0192)^ head and neck cancer,^[193](#CIT0193)^ brain metastases and glioblastoma multiforme.^[194](#CIT0194)^ Hence, K~v~10.1 presents a good opportunity as an antibody target, in the context of disease association where targeting would be restricted to the periphery due to the inability of antibodies to cross the blood-brain-barrier. K~v~10.1 expression also has potential as a biomarker in several of these tumor types and can correlate with a poor prognosis.^[187](#CIT0187),[195](#CIT0195),[196](#CIT0196)^\n\nAs such, K~v~10.1 has been extensively studied in terms of its role in aberrant cell proliferation and tumor growth, where expression has been reported to be activated by epidermal growth factor receptor (EGFR) tyrosine kinase^[197](#CIT0197)^ and regulated by p53 and E2F1 that are also often altered in cancer.^[198](#CIT0198)^ As well as presenting a potential therapeutic target, K~v~10.1 is being explored as a diagnostic marker through tumor xenograft imaging *in vivo* studies.^[199](#CIT0199)^\u00a0K~v~10.1 also plays a key role in cytoskeletal organisation, which in turn affects cell viability, angiogenesis, migration and invasion,^[200](#CIT0200)^ thereby conferring an advantage to tumor growth through increased vascularization and resistance to hypoxia. It has also been shown that K~v~10.1 is constitutively and rapidly internalized by endocytosis and lysosomal sorting,^[201](#CIT0201)^ and that recycling contributes to maintaining K~v~10.1 surface level expression. This property is an important consideration for the potential of a drug's mechanism of action, such as an antibody-drug conjugate or prodrug format, for the targeting of tumor cells. K~v~10.1 knock-out mice show no apparent deleterious phenotype during embryogenesis and develop normally to adulthood with no behavioural abnormalities,^[202](#CIT0202)^ suggesting that blockade or antagonism of aberrant ion channel expression is a feasible targeting strategy. Moreover, further validation is provided by experimental evidence generated from the specific inhibition of K~v~10.1 by antisense technology^[203](#CIT0203)^ and siRNA^[204](#CIT0204)^ resulting in the reduction of tumor cell line proliferation *in vitro.*\n\nSubsequently, a closely related family member, K~v~10.2, has been identified as a potential target for brain, lung and cervical cancer where clinical proof-of-concept has been attained for the treatment of glioma using the re-purposed FDA-approved antipsychotic drug thioridazine as a K~v~10.2 blocker achieving a reduction in tumor volume.^[114](#CIT0114)^\n\n### K~v~2.1 {#S0007-S2001-S3003}\n\nK~v~2.1, encoded by the *KCNB1* gene, mediates a classical delayed rectifier current that is involved in neuronal repolarization. In addition to their role in the CNS, K~v~2.1 ion channels are also involved in cell differentiation and growth of non-excitable cells, and inhibition of K~v~2.1 in pancreatic \u03b2--cells enhances insulin secretion, which offers a potential therapeutic strategy for the treatment of Type-2 diabetes.^[205](#CIT0205)^ Over-expression and aberrant behaviour of K~v~2.1 has also been reported in several tumor types including uterine cancers,^[206](#CIT0206)^ gastric cancers^[207](#CIT0207)^ and medullablastoma.^[208](#CIT0208)^ Further evidence of the potential therapeutic benefit of targeting K~v~2.1 is provided by studies with perifosine, which is a third generation alkylphospholipid analog with anti-tumor properties. The principal mechanism of action of perifosine is the inhibition of Akt signalling by disrupting lipid rafts to which K~v~2.1 ion channels preferentially localize. A recent study demonstrated that perifosine induces a hyperpolarizing shift in the voltage dependence of K~v~2.1 inactivation, accelerating the kinetics of closed-state inactivation but without altering the voltage dependence of activation.^[209](#CIT0209)^\n\n### K~v~11.1B {#S0007-S2001-S3004}\n\nK~v~11.1 (or hERG) is encoded by the *KCNH2* gene and is the pore-forming \u03b1 subunit of the voltage-gated inwardly rectifying potassium channel associated with cardiac arrhythmias and rhythmic excitability of the pituitary. K~v~11.1 mediates the rapidly activating component of the delayed rectifying potassium current in heart (IKr)^[210](#CIT0210)^ and its properties are modulated by cAMP and auxiliary subunit assembly. It was one of the first voltage-gated channels linked to cancer, and has been shown to be abundantly expressed in several leukemias, gastric and colon cancer, whereas it is epigenetically silenced in ovarian cancer, and thus does not seem required for tumorigenesis in all tumor types.^[24](#CIT0024)^\n\nAs with some other ion channels, K~v~11.1 is associated with the sigma-1 receptor (SigR1), a stress-activated transmembrane chaperone.^[211](#CIT0211)^ SigR1 promotes the formation of K~v~11.1/\u03b21-integrin signalling complexes that trigger the activation of the PI3K/AKT pathways. The presence of Sig1R in tumor cells increases cell motility and vascular endothelial growth factor (VEGF) secretion. *In vitro* therapeutic validation has been illustrated by several experimental observations following blockade of K~v~11.1, such as the reduction of cell proliferation in cultured tumor cells,^[117](#CIT0117)^ ablation of the invasiveness of colorectal cancer cells^[118](#CIT0118)^ and gastric cancer cells,^[119](#CIT0119)^ as well as secretion of VEGF, a well-known driver of tumorigenesis and angiogenesis from glioma^[120](#CIT0120)^ and myeloid leukaemia cells.^[117](#CIT0117)^ NOD/SCID mice engrafted with acute lymphoid leukemia cells and treated with K~v~11.1 channel blockers showed reduced leukemic infiltration and had higher survival rates, suggesting that potential therapeutic effects are relevant in an *in vivo* setting.^[121](#CIT0121),[122](#CIT0122)^\n\nHowever, the use of general K~v~11.1 blockers in cancer therapy presents a risk for causing cardiotoxicity (by lengthening of the EGC QT interval), which would expose the patient to ventricular arrhythmias. The potential to circumvent this relies on the existence of at least 2 isoforms of K~v~11.1, namely K~v~11.1A and K~v~11.1B.^[212](#CIT0212)^ Tumor and cardiac cells express different ratios of the A and B K~v~11.1 isoforms, thus side effects could be potentially avoided by specifically inhibiting the channel splice variant (K~v~11.1B) that is highly expressed in certain tumors,^[121](#CIT0121)^ such as, AML, neuroblastoma^[24](#CIT0024)^ and acute lymphoblastic leukemia.^[122](#CIT0122),[213](#CIT0213)^ Some progress towards achieving this goal has been made using the small molecule CD-160,130,^[121](#CIT0121)^ which blocks the K~v~11.1B isoform with higher specificity than the K~v~11.1A isoform. MAbs targeting K~v~11.1B could provide the wherewithal for superior selectivity, and thus provide a safer therapeutic strategy. This is discussed further in a later section.\n\nCalcium-activated potassium ion channels {#S0007-S2002}\n----------------------------------------\n\nCalcium-activated potassium channels are potassium channels gated by calcium or that are structurally or phylogenetically related to calcium-gated channels.^[214](#CIT0214)^ Intracellular calcium can also trigger potassium currents.^[215](#CIT0215)^ These channels can be broadly categorized into three types based on their unitary conductance: large-, intermediate-, and small-conductance.^[216](#CIT0216)^ Large-conductance channels are activated by both voltage and increases in cytosolic Ca^2+^ and represent a distinct gene family. Intermediate and small conductance K^+^ channels are activated exclusively by increases in intracellular Ca^2+^ and represent a distinct gene sub-family.^[217](#CIT0217),[218](#CIT0218)^ Calcium-activated 6 or 7 TM K^+^ channels (K~Ca~), represent another structural sub-type in the potassium channel family where K~Ca~3.1 is the most well-characterized member.\n\n### K~ca~3.1 {#S0007-S2002-S3001}\n\nFirst described in the 1950s, K~Ca~3.1, encoded for by the *KCNN4* gene, is a voltage-independent potassium channel that is activated by intracellular calcium mediated by calmodulin.^[214](#CIT0214)^ Activation triggers membrane hyperpolarization, which in turn promotes calcium influx. K~Ca~3.1 is expressed on activated T and B cells, macrophages, microglia, vascular endothelium, epithelia, proliferating vascular smooth muscle cells and fibroblasts, and therefore presents a potential therapeutic target for inflammatory and autoimmune diseases, such as inflammatory bowel disease and multiple sclerosis.^[123](#CIT0123)^ This ion channel is also expressed on erythrocytes, hence also has potential as a therapeutic target for sickle cell anemia.^[124](#CIT0124)^ Elevated intracellular calcium activates K~Ca~3.1, thereby maintaining a negative membrane potential, which is required for production of inflammatory chemokines and cytokines by T cells, macrophages and mast cells. Potassium efflux through K~Ca~3.1 can be significant, resulting in efflux of \\>\u00a050% of intracellular potassium content with the associated cell shrinkage being linked to apoptosis in certain circumstances.^[219](#CIT0219),[220](#CIT0220)^ Functional cooperation between TRPC1 and K~Ca~3.1 in the regulation of Ca^2+^ entry has been suggested as both these ion channels co-localize into lipid rafts, and knockdown of TRPC1 suppresses the Ca^2+^ entry induced by K~Ca~3.1 activation.^[221](#CIT0221)^\n\nPreclinical proof-of concept studies in animal models have validated the therapeutic potential of K~Ca~3.1 blockers, with no toxicities observed in K~Ca~3.1 knock-out mice and inhibition from developing severe colitis in two mouse models of inflammatory bowel disease,^[125](#CIT0125)^ a mouse model of experimental autoimmune encephalomyelitis,^[126](#CIT0126)^ several models of cardiovascular diseases^[127](#CIT0127)^ and unilateral ureteral obstruction-induced renal fibrosis in wild-type mice and rat.^[128](#CIT0128)^ K~Ca~3.1 blockers, such as senicapoc, have been evaluated in clinical trials for sickle cell anemia and exercise-induced asthma, but have so far not shown efficacy, although the results have confirmed that targeting K~Ca~3.1 is safe.^[129](#CIT0129),[130](#CIT0130)^ Although senicapoc did not reduce the number of painful sickling crises, which was the clinical endpoint that the sponsoring company, Icagen Inc., had selected for their trial,^[130](#CIT0130)-[133](#CIT0133)^ the compound did demonstrate a reduction in hemolysis with increasing hemoglobin and hematocrit levels, a non-significant reduction in late asthmatic response, and it was well tolerated. In addition, significant inter-patient variation was observed in senicapoc's half-life, making dosing difficult. In a similar manner to K~v~1.3 blockers, K~Ca~3.1 blockade acts on specific T or B cell subsets, providing the wherewithal for targeted immunomodulation rather than whole-sale immunosuppression, and an antibody would provide a longer half-life, yielding a better pharmacokinetics/pharmacodynamics profile.\n\nK~Ca~3.1 has also been implicated in several cancers, presumably mediated by its role in the proliferative response of many cell types. Several reports describe successful inhibition of tumor cell proliferation and pro-invasive behaviour following K~Ca~3.1 blockade in both *in vitro* and *in vivo* studies, including prostate,^[134](#CIT0134)^ breast,^[135](#CIT0135)^ pancreatic,^[136](#CIT0136)^ endometrial,^[137](#CIT0137)^ glioblastoma,^[138](#CIT0138)-[140](#CIT0140)^ head and neck squamous cell carcinoma (HNSCC)^[141](#CIT0141)^ and leukemia.^[142](#CIT0142)^ It has also been proposed that targeting K~Ca~3.1 could provide a potential adjuvant therapy for the inhibition of tumor angiogenesis and tumor progression.^[139](#CIT0139),[143](#CIT0143)^\n\nRecently, the combined activation of K~Ca~3.1 and inhibition of K~v~11.1 has been identified as a potential alternative strategy to overcome cisplatin resistance in colorectal cancer (CRC) from studies using molecular and electrophysical approaches with the cisplatin-resistant CRC cell lines HCT-116 and HCT-8.^[144](#CIT0144)^ Several previously characterized K^+^ channel modulators were tested *in vitro* individually and in combination for their action on K^+^ currents, cell viability, apoptosis, cell cycle, proliferation, intracellular signalling and platinum uptake. These effects were also analyzed in a mouse xenograft model that mimics chemoresistance. Cisplatin-resistant CRC cells express higher levels of K~Ca~3.1 and K~v~11.1 channels compared with cisplatin-sensitive CRC cells. In resistant cells, the K~Ca~3.1 activator, SKA-31, and K~v~11.1 inhibitor, E4031, revealed a synergistic action with cisplatin resulting in apoptosis and inhibition of cell proliferation. Similarly, riluzole is able to both activate K~Ca~3.1 and inhibit K~v~11.1, which suggests a combined approach or potential use of a bispecific antibody as a targeting strategy, for example, in patients with ovarian cancer where cisplatin resistance also presents a challenge in adjuvant therapy.\n\nVoltage-gated calcium ion channels {#S0007-S2003}\n----------------------------------\n\nVoltage-gated calcium channels (VGCC) are a group of voltage-gated ion channels found in the membrane of excitable cells (e.g., muscle, glial cells, neurons) with selectivity for Ca^2+^. At resting membrane potential, VGCCs are normally closed. They are activated at depolarized membrane potentials and are key transducers of cell surface membrane potential changes into intracellular calcium influx that regulates intracellular processes such as contraction, secretion, neurotransmission and gene expression in many different cell types. There are 10 members of the voltage-gated calcium channel family that have been characterized in mammals, and these serve distinct roles in cellular signal transduction.^[222](#CIT0222)^\n\n### Ca~v~3.1 and Ca~v~3.2 {#S0007-S2003-S3001}\n\nCa~v~3.1 and Ca~v~3.2 are encoded by the *CACNA1G* and *CACNA1H* genes, respectively, and both belong to the T-type calcium channel subfamily. Although very closely related with similar biophysical properties, their functional effects are very different and emphasize the importance of being able to develop ion channel modulators with high selectivity. Ca~v~3.1, but not Ca~v~3.2, is thought to act as a tumor suppressor because it is involved in the inhibition of proliferation and promotes apoptosis in MCF-7 human breast cancer cells.^[145](#CIT0145),[146](#CIT0146)^ Whereas overexpression of Ca~v~3.1 suppresses cell proliferation and siRNA knockdown or treatment with ProTx-I, a selective inhibitor for Ca~v~3.1, promotes cell proliferation of MCF-7 cells, gene knockdown or overexpression of Ca~v~3.2 exhibits no effect on cell proliferation in this cancer cell line. Moreover, Ca~v~3.1 expression has been shown to correlate with sensitivity to apoptosis^[145](#CIT0145)^ and inhibition of prostate cancer cell proliferation.^[147](#CIT0147)^\n\nCa~v~3.2 has been suggested to promote a constitutive calcium entry influx due to the influence of K~Ca~3.1.^[223](#CIT0223)^ It is thought that Ca~v~3.2 is responsible for the neuroendocrine differentiation associated with the increase in calcium-dependent secretion of mitogenic factors in prostate cancer,^[224](#CIT0224)^ and has been nominated as a biomarker for breast cancer progression and treatment.^[225](#CIT0225)^ Thus, a different mode of action would be required for modulating cancer drugs targeting each of these calcium channels,^[226](#CIT0226)^ namely, one as a channel agonist (Ca~v~3.1) and one as an antagonist (Ca~v~3.2).\n\nVoltage-gated sodium ion channels {#S0007-S2004}\n---------------------------------\n\nThe voltage-gated sodium channel family has 9 members (Na~v~1.1 to Na~v~1.9) that are encoded by the genes *SCN1A* to *SCN11A*. Na~v~ sodium channels have key roles in the initiation and propagation of action potentials in excitable neuronal cells, muscles and heart tissues, and as such have historically been regarded as therapeutic targets for pain, arrhythmia and epilepsy. A range of inherited disorders affecting skeletal muscle, heart rhythm and the central and peripheral nervous systems have been linked to mutations in the Na~v~ genes^[227](#CIT0227)^ that confer loss-of-function or gain-of-function properties.^[228](#CIT0228),[229](#CIT0229)^\n\n### Nav1.7, Nav1.8 and Nav1.9 {#S0007-S2004-S3001}\n\nNa~v~1.7, Na~v~1.8 and Na~v~1.9 are expressed in peripheral sensory neurons and hereditary gain of function mutations have been identified as the cause of pain disorders, including hereditary erythromelagia and paroxysmal extreme pain disorders (Na~v~1.7), as well as other painful peripheral neuropathies (Na~v~1.8, Na~v~1.9). Conversely, loss of function mutations in Na~v~1.7 lead to a congenital insensitivity to pain.^[230](#CIT0230)^ It is not surprising then that these Na~v~ isoforms, particularly Na~v~1.7, have generated substantial interest as targets for the development of non-opioid pain therapeutics. Additionally, recent evidence is growing that implicates Na~v~1.7 in cancer. Functional Na~v~1.7 expression has been found to be involved in EGF-mediated tumor cell invasion in non-small lung cancer cells^[231](#CIT0231)^ and has also been reported to be abundantly expressed in prostate cancer^[146](#CIT0146)^ and breast cancer.^[232](#CIT0232)^\n\n### Nav1.5 and nNa~v~1.5 {#S0007-S2004-S3002}\n\nIn metastatic breast cancer cells, Na~v~1.5 is upregulated and found to potentiate tumor cell migration and invasion in both *in vitro* and *in vivo* experimental models,^[233](#CIT0233)^ whereas stable down-regulation of Na~v~1.5 expression significantly reduces tumor growth, local invasion into surrounding tissue and metastasis to liver, lungs and spleen, thus providing a further body of evidence for its role in metastasis. Furthermore, a neonatal splice variant (nNa~v~1.5) has been shown to be upregulated and associated with metastasis and breast cancer progression *in vitro, in vivo* and in clinical samples of patient lymph node tissue.^[234](#CIT0234)^ In developmentally regulated D1:S3 splicing of the Na~v~1.5 gene, *SCN5A*, there are 31 nucleotide differences between the 5\u02b9-exon ('neonatal') and the 3\u02b9-exon ('adult') forms, resulting in seven amino-acid substitutions in the S3/S4 extracellular region of Domain 1 (D1:S3-S3/S4 linker). Functional activity of nNa~v~1.5 can be suppressed by both siRNA and a specific polyclonal antibody, NESO-pAb.^[22](#CIT0022)^ The siRNA rapidly reduced the level of nNa~v~1.5 mRNA by \\~\u00a090%, but not adult Na~v~1.5 mRNA; however, the effects on protein reduction were considerably less. NESO-pAb reduced metastatic activity of the breast cancer cell line, MDA-MB-231, in a dose-dependent manner. Other studies from the same group demonstrated that blockade of the Na~v~1.5 channel with small molecules or siRNA also inhibited the invasiveness of endocrine-resistant breast cancer cells.^[235](#CIT0235)^ Recent work has shown that increasing the level of nNa~v~1.5 cell surface expression increased the metastatic behavior of breast cancer cells^[236](#CIT0236)^ due to a reduction in cell adhesion, and suggested a possible interaction with the Sig1R transmembrane chaperone. Further work is necessary to elucidate the nature of this interaction and provide further understanding of the role of nNa~v~1.5 function in an oncology setting. The adult form of Na~v~1.5 is responsible for propagating the action potential in cardiac muscle, and therefore, like K~v~11.1, targeting isoforms of these channels as a therapeutic strategy necessarily stresses the importance of selectivity.\n\nASIC channels {#S0007-S2005}\n-------------\n\nThe acid-sensing ion channel (ASIC) family, encoded by *ASCI1-5* genes, are part of the epithelial sodium channel superfamily and are voltage independent. Instead, ASICs are activated in response to reduced extracellular pH,^[237](#CIT0237)^ particularly tissue acidosis and are Na^+^ permeable with an isoform, ASIC1a, showing low Ca^2+^ permeability. ASICs are expressed in the peripheral and central nervous system and are potential therapeutic targets for neurological conditions.^[238](#CIT0238)^\n\n### ASIC1a {#S0007-S2005-S3001}\n\nASIC1a mediates Ca^2+^ overload and has been reported to contribute to ischemic nerve cell death and inflammation in multiple sclerosis.^[239](#CIT0239)^ It has also been implicated in pain,^[240](#CIT0240)^ migraine,^[241](#CIT0241)^ pain associated with bone cancer,^[152](#CIT0152)^ glioblastoma^[232](#CIT0232),[242](#CIT0242)^ and breast cancer.^[243](#CIT0243)^ Additionally, ASIC1 inhibitors have been shown to cause a significant reduction of tumor growth and load.^[243](#CIT0243)^\n\nTRP channels {#S0007-S2006}\n------------\n\nThe mammalian TRP channel family consists of six gene families comprising 28 members of cation channels activated by different stimuli and ligands with diverse physiological functions that range from pain and thermal perception to Ca^2+^-mediated cell processes and homeostatic reabsorption of calcium and magnesium. The mammalian TRP channel family can be broadly sub-divided into 6 sub-families that consist of TRPA (ankyrin), TRPV (vanilloid), TRPM (melastatin), TRPC (canonical), TRPP (polycystin) and TRPML (mucolipin). Known naturally occurring compounds that act as ligands include capsaicin (TRPV1) and menthol (TRPM8). Trp channels are thought to play a key role in several diseases encompassing inflammation, allergy, autoimmune, fibrosis, oncology and pain indications.^[244](#CIT0244)^ Calcium entry through Trp channels may inhibit apoptosis, and is an effect that has been partly attributed to the stimulation of NF-kB.^[245](#CIT0245)^ Examples of Trp channels in various disease settings are described below and have been reviewed in depth elsewhere with regards to targeting pharmacology strategies.^[155](#CIT0155),[246](#CIT0246)^\n\n### TRPM7 {#S0007-S2006-S3001}\n\nTRPM7 is associated with proliferation, motility and metastasis of cancer cells. Inhibition of TRPM7-regulated PI3K/AKT and MEK/ERK signalling pathways has been demonstrated to suppress glioblastoma cell proliferation and migration *in vitro*.^[247](#CIT0247),[248](#CIT0248)^ This inhibition is thought to enhance apoptosis induced by TRAIL,^[249](#CIT0249)^ and induce replicative senescence and enhanced cytotoxicity with gemcitabine in pancreatic adenocarcinoma.^[250](#CIT0250)^ Aldehyde dehydrogenase (ALDH1), which is a cytoplasmic stem cell marker in many malignancies, has recently been suggested to be a tumor stem cell marker in glioblastoma. The Notch signalling pathway plays a key role in cancer stem cell (CSC) survival, proliferation and maintenance of the CSC population. TRPM7 gene silencing down-regulates both the Notch and STAT3 pathways in glioma stem cells, whereas increased ALDH1 expression and activity is induced by TRPM7.^[251](#CIT0251)^ Moreover, phosphorylated STAT3 binds and activates ALDH1 promoters in glioma cells. Thus, the authors concluded that these findings demonstrate that TRPM7 activates the Notch and STAT3 pathways leading to activation of ALDH1 and subsequent increases in cell proliferation and migration.^[251](#CIT0251)^\n\n### TRPM8 {#S0007-S2006-S3002}\n\nTRPM8 is a receptor-activated non-cationic ion channel. In prostate epithelial cells, expression of TRPM8 is regulated by androgen, is elevated in androgen-sensitive cancerous cells compared with normal cells and has been confirmed as an ionotropic receptor for testosterone.^[252](#CIT0252)^ As such, TRPM8 has been identified as a novel target for androgen-regulated prostate cancer,^[253](#CIT0253)^ where overexpression is androgen-dependent and required for tumor cell survival. Although the precise mechanism involved is unknown,^[254](#CIT0254)^ the influx of Ca^2+^ and Na^+^ in prostate cancer cells has been shown as necessary for survival and function,^[255](#CIT0255)^ and it is well established that Ca^2+^ signaling regulates proliferation and apoptosis in cancer cells. In androgen-sensitive cell lines, such as LNCaP, testosterone activation of TRPM8 elevates basal Ca^2+^ levels^[252](#CIT0252)^ whilst TRPM8 inhibition with a small molecule antagonist or siRNA results in cell death.^[255](#CIT0255)^ Anti-androgen therapy also significantly reduces the expression of TRPM8.^[256](#CIT0256)^\n\nThe tissue-specific function of TRPM8 in prostate physiology and carcinogenesis remain unknown. This is complicated by the different cellular locations for TRPM8, namely the cytoplasm (in the endoplasmic reticulum) and plasma membrane. Testosterone-induced plasma membrane TRPM8 activity elicits calcium uptake causing apoptotic cell death. In addition, the promoter region of *trpm8* possesses a consensus p53 binding site, suggesting that TRPM8 may serve as a downstream target of tumor-suppressor genes potentially providing a protective role.\n\nConversely, TRPM8 expression is significantly lower in androgen-independent and metastatic prostate cancer.^[257](#CIT0257)^ The androgen independent pathways do not require androgens, but can be activated by growth factors acting through kinase pathways. These initial observations suggested that TRPM8, as a therapeutic target, may only be suitable for androgen-sensitive prostate cancer. However, application of an adjuvant therapy that rescues TRPM8 expression or enhances its activity or acts as an agonist could pose a potential strategy for the treatment of androgen-independent prostate cancer.^[252](#CIT0252),[258](#CIT0258),[259](#CIT0259)^ Such a hypothesis is substantiated by the observations that, whilst TRPM8 may not be essential for the survival of the androgen-independent prostate cancer cell line PC3, overexpression of this ion channel mediates a reduction in proliferation and migration, as well as facilitating apoptosis.^[258](#CIT0258)^\n\nThus, androgen sensitivity would need to be taken into consideration when selecting the desired mechanism of action of an antibody targeting TRPM8 in prostate cancer. That is, an antibody with ADCC effector function targeting plasma membrane-associated TRPM8 might be preferred in the case of androgen-dependent prostate cancer, while an antibody with agonist activity combined with an adjuvant therapy to enhance expression might be more effective for androgen-independent prostate cancer.\n\n### TRPV1 {#S0007-S2006-S3003}\n\nTRPV1 is overexpressed in many tumor types, including endometrial,^[260](#CIT0260)^ thyroid,^[261](#CIT0261)^ breast,^[262](#CIT0262)^ astrocytoma,^[263](#CIT0263)^ prostate,^[264](#CIT0264)^ pancreas,^[265](#CIT0265)^ colon,^[266](#CIT0266)^ melanoma^[267](#CIT0267)^ and bladder.^[267](#CIT0267)^ TRPV1 is activated by capsaicin and is probably the most well-known TRP channel targeted for pain with marketed products, such as Qutenza\u00ae (Acorda Therapeutics) and Zuacta^TM^ (Sanofi), both of which are capsaicin-based. As an agonist, administration of capsaicin causes an initial enhanced stimulation of TRPV1-expressing nociceptors that may be associated with painful sensations, but this is followed by pain relief thought to be mediated by a reduction in TRPV1-expressing nociceptive nerve endings. However, there may be a gradual re-emergence of painful neuropathy over time, and this is thought to be due to TRPV1 nerve fibre reinnervation. This potentially could be circumvented by the use of an antagonist antibody. TRPV1 also plays a key role in deep tissue pain,^[268](#CIT0268)^ joint pain in arthritis^[269](#CIT0269)^ and bone cancer pain.^[152](#CIT0152)^\n\n### TRPV3 {#S0007-S2006-S3004}\n\nThe clinical significance of TRPV3 in non-small cell lung cancer (NSCLC) was recently reported^[270](#CIT0270)^ where it was observed to be overexpressed in \\~\u00a068% of lung cancer cases, correlating with the differentiation and tumor node metastasis stage of the tumor. Significantly, TRPV3 expression was associated with short overall survival. Blocking or knockdown of TRPV3 has been shown to inhibit lung cancer cell proliferation and arrest the cell cycle at the G1/S transition stage.^[270](#CIT0270)^ The rate of proliferation of epithelial cells in TRPV3 knockout mice is less than that in wild-type mice and TRPV3 up-regulation has been shown to be associated with a high risk for development of CRC.^[271](#CIT0271)^ TRPV3 has been proposed as a potential companion drug target for NSCLC.^[270](#CIT0270)^\n\n### TRPV6 {#S0007-S2006-S3005}\n\nTRPV6 demonstrates higher calcium selectivity over other TRP channels and plays an important role in regulation of calcium homeostasis in the body. In cancer, evidence points to its upregulation and correlation with the advanced stages in prostate, colon, breast, thyroid and ovarian carcinomas where it translocates to the plasma membrane via an Orai1-mediated mechanism and promotes tumor cell survival by enhancing proliferation and conferring apoptosis resistance.^[272](#CIT0272)^\n\n### TRPA1 {#S0007-S2006-S3006}\n\nTRPA1 is implicated in inflammatory pain and naturally derived compounds from plants, such as mustard oil, act as agonists on TRPA1 causing pain by excitation of nerve fibres.^[273](#CIT0273)^ The closed-state structure of TRPA1 was recently solved,^[33](#CIT0033)^ and further study using molecular dynamics simulations, in parallel with mutagenesis and functional evaluation by electrophysiology, explored conformational changes on the proposed open state for an informed understanding on the structure and function of this ion channel.^[274](#CIT0274)^ However, selection of appropriate animal disease models requires careful consideration because cross-species variations in metabolic mechanisms and signal transduction pathways can lead to species-specific differences in TRPA1 function. Paclitaxel-induced neuropathy is thought to trigger the release of mast cell tryptase, which activates the protease-activated receptor 2 that in turn sensitizes TRPA1 (as well as TRPV1 and TRPV4) through the PLC, PKC and PKA signalling pathways.^[275](#CIT0275)^ TRPA1 expression can be modulated by other GPCRs, including the bradykinin receptor, the bile acid receptor TGR5 and the MAS-related GPCR.^[156](#CIT0156)^ Inhibition of TRPA1 eliminates the mechanical and cold allodynia induced by cisplatin and oxaplatin, which are commonly used chemotherapies.^[155](#CIT0155)^ In addition, selective blockade of TRPA1 attenuates pain without altering body temperature regulation or the ability to feel cold.^[275](#CIT0275)^ TRPA1 also presents a therapeutic opportunity in the treatment of migraine.^[276](#CIT0276)^ Similar observations have been made for TRPM8 where it plays a major role in cold hypersensitivity^[277](#CIT0277),[278](#CIT0278)^ and presents a therapeutic opportunity both in the treatment of pain and migraine.^[279](#CIT0279)^\n\nIn addition to its role in pain signalling, TRPA1 is found in nerve fibers that innervate the respiratory tract, in the peripheral nervous system, as well as on non-neuronal cells, such as fibroblasts and epithelial cells,^[280](#CIT0280)^ and it is an emerging target for respiratory conditions such as cystic fibrosis, asthma, allergic rhinitis, chronic cough and itch. A large body of evidence accumulated from *in vitro* experiments, animal disease models and patient data indicates that TRPA1 functions as a chemosensor for exogenous irritants and endogenous mediators of inflammation. Additionally, the presence of fine particulate matter in OVA-sensitized mice has been demonstrated to upregulate TRPA1 expression.^[157](#CIT0157)^ Based on these observations, a therapeutic strategy targeting TRPA1 in respiratory disease would require blockade of this important ion channel.^[158](#CIT0158)^\n\n### TRPC3 {#S0007-S2006-S3007}\n\nExcessive Ca^2+^ influx regulates cytotoxic processes associated with immune-mediated diseases, such as acute pancreatitis and Sj\u00f6gren's syndrome causing dry mouth and/or dry eye. Inhibition of TRPC3, and therefore Ca^2+^ influx, has been shown to protect pancreatic and salivary gland secretory cells from damage caused by Ca^2+^ cytotoxicity.^[159](#CIT0159)^ TRPC3 also plays a role in airway smooth muscle proliferation associated with airway remodelling, a histological characteristic of chronic respiratory diseases such as asthma and chronic obstructive pulmonary disease (COPD).^[160](#CIT0160)^ Inhibition or knockdown of TRPC3 blocks increased activity of TRPC3 and membrane depolarization in OVA-sensitized/-challenged cells^[281](#CIT0281)^ and suppresses airway smooth muscle cell proliferation and airway remodelling in mouse models of disease.^[282](#CIT0282)^\n\n### TRPC6 {#S0007-S2006-S3008}\n\nThe classical or canonical TrpC family (formerly short-TRPs, STRPs) encompasses channels presenting a large number of different activation modes. Some are store-operated, whereas others are receptor-operated channels activated by the production of diacylglycerol or redox processes. TrpC6 is amongst the latter subgroup of TrpC channels.^[283](#CIT0283)^\n\nOf clinical relevance, TrpC6 channels are upregulated in a wide range of cell types across a broad spectrum of disease indications, such as focal segmental glomerulosclerosis (FSGS),^[161](#CIT0161)^ pulmonary fibrosis,^[284](#CIT0284)^ cancer,^[285](#CIT0285)-[290](#CIT0290)^ hypertension^[291](#CIT0291),[292](#CIT0292)^ and allergic asthma.^[293](#CIT0293)^ Expression or over-expression of TrpC6 has been shown to have a pro-proliferative effect. For example, the presence of TrpC6 has been determined to be essential for the progression of gastric cancer in studies comparing normal and cancerous epithelial cells of humans, and TRPC6 inhibitors, and a dominant negative TRPC6 channel mutant, have been shown to promote cell cycle arrest in gastric cancer cell lines.^[294](#CIT0294)^ There are no specific TrpC6 inhibitors in development that effectively suppress these processes, and therefore this ion channel presents a potential therapeutic opportunity. In the kidney, TrpC6 resides in podocyte membranes where it plays a role in maintaining glomerular function by acting as a non-selective cation channel that primarily transports Ca^2+^. TRPC6 has also been implicated in renal disease, such as primary forms of FSGS where circulating factors cause dysfunction or loss of podocytes via TrpC6 channel activation. TRPC6 mutations from both familial and sporadic cases of FSGS map to the N- and C-termini, often resulting in excessive calcium influx and subsequent injury or loss of podocytes.^[161](#CIT0161)^ This in turn promotes glomerular mesangial cell apoptosis via calcineurin/NFAT and FasL/Fas signaling pathways.^[295](#CIT0295)^ Additionally, mice with podocyte-specific overexpression of TrpC6 recapitulate many of the pathological features of FSGS.^[296](#CIT0296)^ Despite the clear association of TrpC6 with FSGS, there are currently no clinical trials of therapeutics targeting TrpC6 for any condition.\n\nLigand-gated purinergic P2X channels {#S0007-S2007}\n------------------------------------\n\nThere are seven members of the P2X purinergic ligand-gated ion channel family (P2X1-P2X7) reported to date, which are encoded by the genes *P2RX1-P2RX7* that have dispersed chromosomal locations.^[297](#CIT0297)^ Of these, the homotrimers of P2X3, P2X4, P2X7 and heterotrimers of P2X2/3 offer the greatest potential for therapeutic targeting.^[298](#CIT0298)^ On activation by ATP released by damaged cells, these ion channels open to allow the influx of calcium, leading to cell membrane depolarization.^[299](#CIT0299)^ Whilst P2X channels have a wide tissue distribution, they are activated by ATP with differing affinities that trigger distinct physiological functions, such as central synapse transmission smooth muscle cell contraction, platelet aggregation, macrophage activation, cell proliferation and cell death.^[300](#CIT0300)^\n\n### P2X3 {#S0007-S2007-S3001}\n\nActivation of P2X3 homomers contributes to acute nociception and activation of P2X2/3 heteromers that modulate nociceptive sensitivity associated with nerve injury or chronic inflammation. P2X3 receptors undergo rapid turnover that is dependent on their activity status (i.e., agonist-stimulated internalization), where expression and function are modulated by phosphorylation of intracellular domains in response to extracellular pain-causing molecules.^[301](#CIT0301)^ Calcium/calmodulin-dependent serine protein kinase has been demonstrated to regulate purinergic nociceptive signalling of P2X3.^[302](#CIT0302)^\n\n### P2X4 {#S0007-S2007-S3002}\n\nP2X4 has also been shown to play a role in neuropathic pain where knockdown, knockout and inhibition of P2X4 in rodent *in vivo* models reverses neuropathic pain.^[303](#CIT0303)^ Studies show that the P2X4 channel expressed in the satellite glial cells of dorsal root ganglia is associated with neuropathic pain.^[304](#CIT0304)^ Additionally, activation of P2X4 (and P2X7) on microglia is thought to maintain nociceptive sensitivity through neural-glial cell interactions as antagonists to these receptors have been observed to reduce neuropathic pain. These mechanisms are reviewed in depth elsewhere.^[305](#CIT0305),[306](#CIT0306)^\n\n### P2X7 {#S0007-S2007-S3003}\n\nP2X7 is expressed on T cells and is implicated in the immune response where it is upregulated on macrophages and microglial cells following stimulation by IFN\u0264 or LPS driving T cell activation and proliferation with the release of IL-1\u03b2.^[307](#CIT0307)-[310](#CIT0310)^ P2X7 has also been directly implicated in systemic lupus erythematosus,^[311](#CIT0311)^ arthritis^[312](#CIT0312)^ and diabetic neuropathic pain.^[313](#CIT0313)^\n\nIn cancer, P2X7 is thought to act as a danger \"sensor\" for high local ATP in inflammation^[314](#CIT0314)^ and the tumor microenvironment,^[315](#CIT0315)^ with overexpression of P2X7 linked to tumor growth and metastasis.^[316](#CIT0316),[317](#CIT0317)^ It was recently demonstrated that the anti-parasitic agent ivermectin can modulate sensitivity of both P2X4 and P2X7 receptors to extracellular ATP in association with the pannexin-1 channel, resulting in the induction of a non-apoptotic and inflammatory form of cancer cell death.^[318](#CIT0318)^ A non-functional conformational form of P2X7 (nfP2X7) has been described that is unable to form the large pore conformation associated with driving cell death and is specifically expressed on the surface of cancer cells sourced from a wide range of tumors.^[21](#CIT0021)^ It is thought that this variant of P2X7 could result from rare splice isoforms or single nucleotide polymorphisms,^[319](#CIT0319)^ and the resulting molecule only has one functional ATP binding site of the three available, thus providing a unique targeting epitope on nfP2X7 that is not present on functional P2X7.^[21](#CIT0021)^ To date, this is the only ion channel target with an antibody (BIL010t) in clinical study, as described below in the \"Antibodies in research and development\" section.\n\nAdditionally, several P2X7 splice variants have been described, with the full-length and a truncated version (P2X7RB) specifically studied in osteosarcoma.^[320](#CIT0320)^ In the case of glioma, Fang and co-workers have reported that suppression of P2X7 promotes tumor growth via the EGFR signalling pathway.^[321](#CIT0321)^ The same group evaluated the potential of transplanted endothelial progenitor cells (EPCs) as a therapeutic strategy and imaging probe to overcome anti-angiogenic resistance of gliomas in the brain. This approach takes advantage of P2X7 receptor activation on EPCs prior to transplantation, which promotes their proliferation and homing to gliomas *in vivo*.^[322](#CIT0322)^ P2X7 has also been shown to act as an upstream modulator of the VEGF signalling pathway with proof-of-concept demonstrated in experimental neuroblastoma^[323](#CIT0323)^ using ACN human NB cells. *In vivo* studies also demonstrated ACN-derived tumor growth could be reduced in nude/nude mice, and an even stronger effect of P2X7 blockade was obtained in a syngeneic immune-competent neuroblastoma model where Neuro2A cells were injected in albino mice. Lastly, it has been reported that CD40 upregulates P2X7 in retinal endothelial cells, rendering them susceptible to ATP/P2X7--mediated programmed cell death.^[324](#CIT0324)^\n\nCalcium-release activated channels {#S0007-S2008}\n----------------------------------\n\nCalcium-induced calcium release is a process that occurs in many cells and tissues whereby an increase in concentration of Ca^2+^ in the cytosol causes a further increase as Ca^2+^ is released from intracellular stores. In this case, Ca^2+^ release from the sarcoplasmic/endoplasmic reticulum occurs following activation of Ca^2+^ release channels in response to elevated levels of cytosolic Ca^2+^, inositol triphosphate (IP3), or changes in membrane potential.\n\nOn the other hand, Ca^2+^ release-activated Ca^2+^(CRAC) channels represent one of the main Ca^2+^ entry pathways into the cell. They are fully reconstituted via two proteins, the stromal interaction molecule 1 (STIM1), a Ca^2+^ sensor in the endoplasmic reticulum, and the Ca^2+^ ion channel Orai in the plasma membrane. After Ca^2+^ store depletion, STIM1 and Orai couple to each other, allowing Ca^2+^ influx.^[325](#CIT0325)^\n\n### CRAC (Orai, STIM) {#S0007-S2008-S3001}\n\nCalcium release-activated calcium (CRAC) channel protein 1 is a calcium selective ion channel that is encoded by the *ORAI1* gene in humans. Four Orai1 subunits comprise the CRAC channel, which is indirectly activated by low intracellular Ca^2+^, whereby decreased calcium concentration in the endoplasmic reticulum is sensed by STIM1, which in turn aggregates and relocates near the plasma membrane, where it activates Orai1 via direct interaction with the ion channel.^[326](#CIT0326),[327](#CIT0327)^ Sig1R inhibits store-operated Ca^2+^ entry (SOCE) by attenuating coupling of STIM1 to Orai1.^[328](#CIT0328)^\n\nCRAC channels are critical to the activation of T lymphocytes, mast cells and other hematopoietic cells, as they provide the primary route for the influx of calcium into these cells.^[329](#CIT0329),[330](#CIT0330)^ CRAC channel inhibition could therefore provide a direct way of modulating the immune response for the treatment of multiple diseases and disorders. In addition, growth factors can stimulate CRAC channels (CRACs) to mediate Ca^2+^ entry and subsequent Ca^2+^ oscillations in proliferating cells. For example, tumor cell migration and metastases require Ca^2+^ influx and Orai and STIM have been shown to have a role in apoptosis resistance associated with proliferating and migrating tumor cells.^[331](#CIT0331)-[334](#CIT0334)^ In the context of metastatic CRC, where only 10--20% of patients receive a clinical benefit from the use of anti-EGFR mAbs, studies suggest that efficacy of these mAbs could depend on their ability to reduce SOCE that is known to promote cancer cell growth.^[335](#CIT0335)^ Subsequent studies revealed that a lipid raft-ion channel complex, consisting of K~Ca~2.3, TRPC1 and Orai1, regulates SOCE-dependent colon cancer cell migration,^[335](#CIT0335)^ in agreement with previous observations.^[336](#CIT0336)^ The formation of this lipid raft is triggered by STIM1 phosphorylation by EGF and activation of the AKT pathway. Additionally, evidence of an association of Orai and STIM with TRPC1 has been implicated in tumorigenesis.^[337](#CIT0337)^\n\nChloride channels {#S0007-S2009}\n-----------------\n\nChloride channels are a superfamily of ion channels with diverse structures (e.g., voltage-gated and ligand-gated channels),^[338](#CIT0338)^ but their biology is still poorly understood and as such remain a relatively under-exploited target class for drug discovery.^[339](#CIT0339)^ These channels may conduct many different ions, but they are named for chloride (as this is the most abundant). They are involved in a wide range of biological functions, including epithelial fluid secretion, cell-volume regulation, neuroexcitation, smooth-muscle contraction and acidification of intracellular organelles. Mutations in several chloride channels cause human diseases, including cystic fibrosis, macular degeneration, myotonia, kidney stones, renal salt wasting and hyperekplexia. Chloride-channel modulators have potential applications in the treatment of some of these disorders, as well as in secretory diarrhoeas, polycystic kidney disease, osteoporosis and hypertension. The chloride ion channel family comprises CLCs (that contain 10 or 12 transmembrane helices), CLICs (comprising two TMDs linked by a large central pore loop that may be glycosylated and can switch from a soluble state to a membrane bound state), CFTR (an ABC transporter) and CACC (calcium activated chloride channels, e.g., TMEM16A). A number of chloride channels are intracellular in distribution, e.g., CLICs, and would therefore not be suitable for targeting with antibodies unless an \"intrabody\" approach could be successfully applied.\n\n### CLCs {#S0007-S2009-S3001}\n\nCLCs have been implicated in osteopetrosis (CLC-7)^[340](#CIT0340)^ and glioblastoma (CLC-2, CLC-3 and CLC-5),^[341](#CIT0341)^ where in normal brain tissue only CLC-2 is expressed on the plasma membrane.\n\n### CACCs {#S0007-S2009-S3002}\n\nCACCs, such as CLCA1 and CLCA2, are thought to have an apoptotic role, but are downregulated in cells resistant to detachment-induced apoptosis (knowns as anoikis), as shown in mammary gland cells,^[342](#CIT0342)^ suggesting a role in cancer cells where reduction in expression is thought to contribute towards tumor cell survival.^[343](#CIT0343),[344](#CIT0344)^ Moreover, the reduction in expression of CLCA2 by lentiviral shRNA causes cell overgrowth and focus formation, enhanced migration and invasion with an increased risk of metastasis.^[344](#CIT0344)^ Overexpression of CLCA2, on the other hand, inhibits cell proliferation with increases in chloride current at the plasma membrane and accompanying reduced intracellular pH.^[345](#CIT0345)^ CACCs are also thought to have a role in asthma, COPD, cystic fibrosis and neuropathic pain.^[346](#CIT0346),[347](#CIT0347)^ Currently, there are no antibodies in discovery or development, and only 2 peptides targeting chloride channels, including lancovutide (duramycin) for the treatment of COPD and cystic fibrosis, were reported to be in development, but both have been discontinued (LanthiBio, TransMolecular).\n\nIon channel targeting antibodies in development {#S0008}\n===============================================\n\nGiven the advances made in targeting GPCRs over the past decade,^[20](#CIT0020)^ there is now substantial interest from the biotechnology and pharmaceutical industries to extend these capabilities to therapeutic mAbs that target ion channels.^[348](#CIT0348)-[351](#CIT0351)^ Several efforts are ongoing to succeed with this target class, and, whilst the pipeline is in its infancy, the ion channel-antibody pipeline in 2016 was reminiscent of the early stage GPCR-antibody pipeline 10\u00a0years ago ([Figure 4(a](#F0004))). A review of information available in the public domain that includes company websites, publications, conferences and searches on commercial databases, such as Pharmaprojects, has identified over 20 ion channels that are the focus of research and development activities. However, the majority of these are in early discovery or preclinical development, with only one antibody that has recently completed Phase 1 studies for basal cell carcinoma. Nevertheless, it is evident that there is a continued and burgeoning interest in ion channels as therapeutic antibody targets ([Figure 4(b](#F0004))), where substantial progress has been made in this field recently, as reflected by the increase in the number of preclinical stage programs in 2018 ([Figure 4(c](#F0004))). The range of ion channel targets under investigation has broadened with 37 programs listed in 2016 directed to at least 17 ion channel targets compared to 56 programs directed to at least 23 ion channel targets in the research and development pipeline for 2018. This and several other ion channel-targeting antibodies are discussed below as case studies of antibodies with commercial interest.10.1080/19420862.2018.1548232-F0004Figure 4.Ion channel targeting antibody programs in the R&D pipeline. Shown is a comparison between 2016 (a) and 2018 (b). Several ion channel targeting programs are undisclosed, such as Integral Molecular, Merck, Amgen, MedImmune, Theranyx, Ablynx and Innovative Targeting Solutions. The range of ion channel targets under investigation has broadened with 37 programs listed in 2016 compared to 56 programs in 2018. These antibody programs are directed to at least 17 targets in 2016 compared to at least 23 targets of interest in 2018, as can be observed by the increase in number of pie sectors. Selected targets of interest are denoted within the piechart layout with the number of programs indicated in brackets as the color coding of the pie sectors shifts due to the delisting of TRPM8 and the emergence of P2X2/P2X3. Since 2016, the number of programs underway for targeting Orai1 has decreased; there is a noticeable Increase in activity for K~v~1.3 and P2X3; whereas Na~v~1.7 and Na~v~1.8 activity remains at a similar level. The P2X family is indicated by the black bracket line. Each target is color coded as depicted in the key to the right-hand side of each piechart. Information sourced from the public domain, such as scientific literature, company websites, etc. c. Shown are the ion channel antibodies in the R&D pipeline by stage depicting progress since 2016 to date. There is only 1 antibody program in clinical development (Ph1): nfP2X7 for basal cell carcinoma (Biosceptre). Some ion channel targets have more than one program for different therapeutic indications (for example, K~v~1.3, P2X7 and CACNA2D1). Inactive programs include TRPA1 (Juno Therapeutics) and nAchR (NIH) and are listed as inactive (but not as terminated, unlike TRPM8 which is not currently listed).\n\nAntibodies in research and development {#S0009}\n======================================\n\nClinical development {#S0009-S2001}\n--------------------\n\n### nfP2X7 {#S0009-S2001-S3001}\n\nBiosceptre recently published Phase 1 study results for BIL010t, a polyclonal antibody that targets a non-functional form of P2X7 (nfP2X7) for the treatment of basal cell carcinoma.^[30](#CIT0030)^ BIL010t is the first ion channel-targeting antibody to enter the clinic with the potential to become a first-in-class therapy. The company has built a pipeline focused on targeting nfP2X7 with various modalities in development (<http://www.biosceptre.com/pipeline/>) as outlined in further detail below.\n\nP2X7 is an ATP-sensing pore-forming channel that can drive apoptosis by allowing rapid Ca^2+^ influx and downstream caspase activation.^[352](#CIT0352)^ Non-functional variants of this channel exist, in particular, nfP2X7, which allows residual calcium flux but fails to form an apoptotic pore. This variant is expressed at high levels in many cancers, including melanomas, and presents a unique epitope, E200, at the cell surface which is not present on normal healthy cells.^[353](#CIT0353)^ Selective exposure of this epitope makes it an ideal target for the development of therapeutic antibodies against a variety of different cancers.\n\nBIL010t was generated by immunization of sheep using the E200 peptide sequence conjugated to keyhole limpet hemocyanin and then immunoglobulin G (IgG) was purified from the resulting sera. The disadvantage to this approach is that immune responses can differ from one individual host animal to another, leading to at least some batch-to-batch variation. Nevertheless, *in vivo* studies of BIL010t in a mouse melanoma preclinical model have demonstrated significant inhibition of tumor growth. BIL010t is being investigated as a topical therapy and has demonstrated safety and tolerability in this first clinical study,^[354](#CIT0354)^ as well as providing an indication of efficacy that was confirmed by histopathological analysis of post-treatment biopsies.^[30](#CIT0030)^ Biosceptre is building a pipeline of nfP2X7-targeting modalities: BIL03s is a human single domain antibody that has been developed for systemic administration for a number of solid tumors and will enter a Phase 1 trial imminently in Australia; a vaccine is in Phase 1 study for solid tumors (BIL06v), and an antibody-drug conjugate has progressed to preclinical development. The company's pipeline presents the opportunity to target a broad range of other cancers, including breast, lung and prostate cancer.\n\nPreclinical development {#S0009-S2002}\n-----------------------\n\n### P2X7 {#S0009-S2002-S3001}\n\nP2X7 has been closely studied in cells of the hematopoietic lineage, particularly innate immune cells and lymphocytes. A collaboration between the University Medical Center Hamburg-Eppendorf and Ablynx has reported the generation of antagonist nanobodies (another type of single domain antibody) that can either block or potentiate P2X7 on T cells.^[355](#CIT0355)^ ATP-induced gating leads to shedding of CD62L on T cells and IL-1\u03b2 release, and antagonizing P2X7 may provide an alternative or complementary strategy to blocking IL-1\u03b2. With excellent specificity for P2X7, antagonist nanobodies in modular bivalent format significantly increased the potency of the nanobody (pM) to block the IL-1\u03b2 inflammatory response in whole human blood, preventing pore formation, which in turn led to cell death in P2X7 transfected cells. In the presence of 100\u00a0\u00b5M ATP, these nanobodies demonstrate an IC~50~ of 0.1\u00a0nM with 1000-fold superior potency and efficacy over existing benchmark compounds (KN-62 and A438079). *In vivo* function was demonstrated using surrogate mouse nanobodies in an antibody-induced nephritis model, where administration of trivalent formated, half-life extended anti-P2X7 nanobodies could modulate P2X7 dependent pathology, as well as in a DNFB-sensitized mouse model for allergic dermatitis. The lead candidate, Dano1, is being progressed through preclinical development for the treatment of inflammation and neurological disease, such as rheumatoid arthritis, inflammatory bowel disease, COPD, multiple sclerosis, renal injury and graft-vs-host disease (GvHD), however no development has been reported since the acquisition of Ablynx by Sanofi in early 2018. It is noteworthy that agonist nanobodies were also identified, e.g., Dano5, which induced shedding of CD27 and cell death of P2X7-postive T cells, that could have utility for immuno-oncology applications.\n\n### P2X3 and P2X2/P2X3 {#S0009-S2002-S3002}\n\nRinat (a subsidiary of Pfizer) recently described the modulation of P2X3 and P2X2/3 ion channels by mAbs.^[356](#CIT0356)^ These ligand-gated ion channels have clinical relevance in pain sensation, such as inflammatory and visceral pain, cancer pain (particularly in bone and HNSCC), as well as chronic cough. Mouse mAbs were generated using standard hybridoma technology following immunization of balb/c mice with purified recombinant full-length P2X3 expressed in mammalian cells that was solubilized in dodecyl maltoside detergent and combined with Gerbu adjuvant to enhance the immune response.^[356](#CIT0356)^ The resulting panel of hybridoma antibodies exhibited different functional effects depending on homomeric or heteromeric composition of the ion channel, as well as the kinetic state and the duration of antibody exposure. Binding to the native channel was confirmed by fluorescence-activated cell sorting (FACS) and immunocytochemistry with functional activity assessed by Ca^2+^ flux and electrophysiology using the whole cell voltage clamp technique. Short-term exposure with one mAb, 12D4, that bound the desensitized conformation was shown to block 80% of homomeric P2X3 \u03b1\u03b2-meATP-activated inward currents at 0.3\u00a0\u00b5M with an IC~50~ of 16\u00a0nM, whereas the same antibody potentiated \u03b1\u03b2-methylene ATP-evoked currents mediated by heteromeric P2X2/3. Interestingly, long-term exposure (24h) reversed the potentiating effect of 12D4 on P2X2/3 mediated current and led to potent inhibition. It was suggested that these differences are based on the composition of the channel (i.e., homomeric vs heteromeric) and the length of time for exposure, as well as differences in binding affinities and epitopes. In addition, mAb 12D4 rapidly internalizes and disappears from plasma membranes of P2X3 expressing cells. Efficacious *in vivo* activity was confirmed for reversing visceral pain in a 2,4,6-trinitrobenzene sulphonic acid-induced colitis rodent model, where the effects were reversible. However, no effect was seen in a complete Freunds adjuvant rodent model for inflammatory pain, nor in a 0.5% formalin test rodent model. Whilst showing promise for the therapeutic treatment of chronic pain, further work will be required to humanize mAb 12D4 and evaluate its effect in other models, including non-human primate.\n\nShark-derived VNAR (another type of single domain antibody) antagonists to P2X3 (OSX300) are being developed by Ossianix in collaboration with Lundbeck A/S for the treatment of chronic pain. The Ossianix discovery platform encompasses both semi-synthetic phage display libraries based on specific VNAR isoforms and immunized repertoires derived from nurse shark lymphocytes.^[357](#CIT0357)^ Integral Molecular in collaboration with Crystal Bioscience (acquired in late 2017 by Ligand Pharmaceuticals for their transgenic chicken platform) presented data in 2016 indicating that they had also successfully raised chicken antibodies to P2X3, achieving sub-nanomolar affinities, potent inhibition of Ca^2+^ flux and in an *ex vivo* model for pain inhibition (90% dorsal root ganglion inhibition).^[358](#CIT0358)^\n\n### K~v~1.3 {#S0009-S2002-S3003}\n\nK~v~1.3 is a well-validated therapeutic target for immune-modulation, and a number of companies have described preclinical data of Ig-based drug candidates in their pipelines. Ablynx has reported generating highly potent and selective nanobodies with *in vivo* proof-of-concept^[93](#CIT0093)^ where the first extracellular loop (ECL1) has been shown to be essential for binding. Binding affinities are sub-nanomolar in bivalent/biparatopic format with greater than 10,000-fold selectivity for K~v~1.3 than other related ion channels. Electrophysiology studies confirmed the lack of measurable off-target current blocking over the closest-related K~v~1 family members and hERG, as well as functional activity that is comparable to the benchmark ShK toxin. Bivalent nanobody constructs demonstrated a fast onset of blocking activity on T~EM~ cells, thereby halting T-cell activation, as well as an increased duration of effects. Both bivalent and trivalent constructs (anti-human serum albumin improved serum half-life) were assessed *in vivo* using a rodent delayed type hypersensitivity model and found to be comparable to ShK in efficacy. Whilst these nanobodies show promise as potential therapeutics, no further development has been reported recently.\n\nThe rational design of a bovine antibody has been used to generate a selective immunosuppressive mAb targeting K~v~1.3.^[359](#CIT0359)^ This was achieved by grafting the toxin peptide sequences for Moka-1 toxin and Vm24-toxin into the ultra-long bovine heavy chain complementarity-determining region 3 (CDR3). The resulting mAb, SVN-001, demonstrated good selectivity and potency against effector human T~EM~ cells, a significantly improved plasma half-life and serum stability compared with the parent peptide, as well as potent *in vivo* efficacy.^[359](#CIT0359)^ By targeting a unique subset of immune cells, SVN-001 is not broadly immunosuppressive, which improves the safety profile compared to typical immunosuppressants.\n\nTetraGenetics, in collaboration with Crystal Biosciences (acquired by Ligand Pharmaceuticals) and argenx, has generated the first conventional light and heavy chain antagonist anti-K~v~1.3 mAbs to be advanced into preclinical studies.^[77](#CIT0077)^ Recombinant K~v~1.3 was used for the immunization of chickens and llamas, as well as for antibody screening. The resulting panel of lead candidate antibodies show a high degree of potency (IC~50~\u00a0\\<\u00a010\u00a0nM) in blocking K~v~1.3 currents and the desired selectivity over related K~v~ family members that would be expected for mAbs. These antibodies are currently being developed for the treatment of Type 1 diabetes, although, as is common for K~v~1.3 antagonists, the potential treatment of other autoimmune diseases by targeting K~v~1.3 overexpression in T~EM~ cells is being explored.\n\n### K~v~10.1 {#S0009-S2002-S3004}\n\nThe aggressive behavior of pituitary tumor cells has been shown to correlate with high expression levels of HER2 and the K~v~10.1 channel.^[191](#CIT0191)^ Therefore, it may be possible to target these cells via combination antibody therapy or possibly even a bispecific antibody format. Furthermore, K~v~10.1 shares some homology with K~v~11.1 in the inner vestibule area of the ion channel,^[360](#CIT0360)^ and so antibodies with superior selectivity compared to small molecules (where cardiac safety is a major concern) would presumably present a significant advantage.\n\nUntil very recently, no specific K~v~10.1 peptide toxin has been reported; however, a novel specific K~v~10.1 inhibitor from the sea anemone *Anthopleura elegantissima* has now been identified.^[361](#CIT0361)^ Prior to this, there was only one group (at the Max-Planck Institute of Experimental Medicine) who had described an antibody (mAb56) that demonstrates exquisite specificity for K~v~10.1 mediated by binding to the E3 region^[61](#CIT0061)^ and does not block K~v~11.1 or the sub-family member K~v~10.2. mAb56 was generated by immunization of mice using a fusion protein that incorporated the E3 region and tetramerizing domain of K~v~10.1, followed by standard hybridoma methodology.^[185](#CIT0185)^ Current inhibition in K~v~10.1-expressing HEK293 and neuroblastoma cells has been demonstrated with mAb56 showing dose-dependent effects and yielding an IC~50~ value of 73\u00a0\u00b1\u00a047\u00a0nmol/L in HEK293 cells. Additionally, the antibody's ability to inhibit tumor cell growth was confirmed both *in vitro* using anchorage-independent cancer cell growth assays and *in vivo* in both MDA-MB-435S human breast cancer and PAXF1657 human pancreas carcinoma xenograft models.^[61](#CIT0061)^ It is not known, however, if this mAb will progress into clinical development.\n\nThe same group has also demonstrated the induction of tumor cell-selective apoptosis by targeting K~v~10.1 via a bifunctional antibody that is a fusion protein comprising an anti-K~v~10.1 scFv antibody fragment and the tumor necrosis factor-related apoptosis-inducing ligand (TRAIL).^[362](#CIT0362)^ This antibody entity, scFv62-TRAIL, has also been reported to sensitize prostate cancer cells and other cancer cells to chemotherapy drugs, such as doxorubicin,^[362](#CIT0362),[363](#CIT0363)^ which could provide a potential means to overcome drug resistance. The full-length IgG from which scFv62 was derived has been used for *in vivo* imaging of tumor xenografts, providing further evidence of the utility of these antibodies in targeted cancer therapy.^[199](#CIT0199)^\n\n### K~v~11.1B {#S0009-S2002-S3005}\n\nK~V~11.1 (also known as hERG1) channels are often overexpressed in human cancers, but targeting them risks cardiotoxicity. CD-160,130 is a small molecule compound that blocks K~v~11.1 channels with a higher efficacy for the K~v~11.1 isoform B,^[121](#CIT0121)^ and shows anti-tumor effects without the cardiovascular toxicity induced by K~v~11.1 blockade. This validates the strategy to explore selective targeting of this isoform with antibodies for therapeutic benefit without the associated risk of cardiac arrhythmia, for example, by using a tumor-specific bifunctional antibody (similar to the scFv62-TRAIL strategy) or tumor-targeted nanoparticles.^[24](#CIT0024),[364](#CIT0364)^ In the latter case, by linking antibodies that recognize tumor-specific cell surface receptors to nanoparticles, cellular uptake (and drug delivery) is dramatically enhanced,^[365](#CIT0365)^ which could further improve the safety and efficacy of targeting K~v~11.1 on cancer cells.^[24](#CIT0024),[366](#CIT0366)^ Indeed, a proof-of-concept study has been described for the conjugation of a K~v~11.1 mAb (specific against the S5 pore of K~v~11.1)^[367](#CIT0367)^ to dicarboxylic acid-terminated pegylated titanium oxide (PEG TiO2) nanoparticles for the targeting of pancreatic adenocarcinoma cells.^[364](#CIT0364)^ Recent preclinical studies have demonstrated that antibody-targeted nanoparticles have better anti-tumor activity compared to non-targeted nanoparticles, due to more efficient localization and penetration into the tumor.^[365](#CIT0365)^ It is also worth noting that neoplastic cells are often depolarized and their changes in membrane potential slow, even when these changes oscillate in phase with the cell cycle stages. Therefore, the proportion of time spent by a voltage-gated channel in a given state can be very different in tumors than in excitable cells.^[368](#CIT0368)^ In another study, *in vitro* proof-of-concept was attained using doxorubicin-loaded, PEGylated gold nanoparticles conjugated to a commercially available anti-K~v~11.1 polyclonal antibody preparation for the targeting of PANC1 cells.^[369](#CIT0369)^ Preliminary results suggest that this approach has the potential to significantly enhance the therapeutic index of doxorubicin. The next logical step would be *in vivo* evaluation of these strategies using a mAb specific to the K~v~11.1B isoform in order to assess targeting efficacy and safety, functional potency, biodistribution and bioavailability.\n\nDiscovery {#S0009-S2003}\n---------\n\nSodium channels, particularly members of the Na~v~ family, have been the focus of intense drug discovery efforts. Notably, the key role that Na~v~1.7 plays in pain pathways and obtaining a selective subtype targeting molecule that avoids off-target related functions, such as cardiovascular side effects,^[52](#CIT0052),[370](#CIT0370)^ has made this a priority in many neuroscience departments. Progress has been impeded due to the high sequence similarity with other Na~v~ channels, especially in the pore region where the binding sites of small molecules are located.^[149](#CIT0149)^ A selective neutralizing mAb to Na~v~1.7, SVmab1, was reported to have been generated by immunization with a peptide sequence corresponding to the loop between S3 and S4 using Abmart's SEAL\u2122 technology with standard hybridoma methodology.^[371](#CIT0371)^ Subsequent characterization determined that SVmab1 bound to the voltage-sensor paddle of Na~v~1.7 and inhibited Na~v~1.7 in transiently expressing HEK cells as determined by whole cell voltage clamping.^[371](#CIT0371)^ SVmab1 was also found to suppress acute and chronic itch in mouse models and inhibit chronic itch-potentiated synaptic transmission in spinal cord slices.^[371](#CIT0371)^ However, following the excitement surrounding the initial discovery, subsequent work was unable to replicate the previously reported effects using a recombinant version of the mAb generated using published sequences.^[372](#CIT0372)^ The lack of effectiveness of recombinant SVmab1 was recently confirmed by the group who initially discovered the antibody, but who also reaffirmed the activity of SVmab1 derived from two different batches of hybridoma-derived SVmab1, albeit with batch differences observed.^[373](#CIT0373)^ The reasons offered for the apparent discrepancy in functionality between the different mAb preparations were potential differences in post-translational modifications between hybridoma- and HEK293- derived antibodies, batch-to-batch variability in hybridoma-generated antibody material and possible errors in antibody sequencing, although the authors indicate the latter possibility is remote.^[373](#CIT0373)^\n\nWhilst many small molecules have been investigated for their potential use for pain control, an interesting observation has been reported that the more specific the molecule, the less analgesic the effect.^[156](#CIT0156)^ It remains to be seen if antibodies that target specific members of the Na~v~ family can overcome this conundrum and provide an advantage, for example, by modulating channel activity in a therapeutically beneficial way. It appears the biopharmaceutical industry is not yet convinced, as many antibody programs highlighted only a few years ago have since been discontinued.^[374](#CIT0374)^\n\nAt the same time, this challenging subclass of ion channel is still the target focus of many efforts with various modalities under investigation, including the insertion of toxin peptides or knottins (Cysteine knot mini-proteins) into the peripheral CDR loops of an IgG structure to gain specificity and potency, as shown by IONTAS in 2017.^[375](#CIT0375)^ Similar proof-of-concept experiments that demonstrate the viability of this approach have been achieved with K~v~1.3 (ShK toxin), K~Ca~3.1 (kaliotoxin) and ASIC1a (psalmatoxin). A similar warhead approach using toxin analogs has been reported whereby the Na~v~1.7 GpTx-1 peptide toxin was tethered to antibodies to generate bifunctional molecules and utilizes FcRn-based antibody recycling to target Na~v~1.7 function.^[376](#CIT0376)^\n\nIn the absence of purified ion channel protein that maintains a biologically relevant conformation, another approach that is under assessment by Visterra Inc. for the generation of mAbs directed to the voltage-sensing domain of Na~v~1.7 harnesses several strategies, namely, the use of yeast surface display and immunization using a chimeric format of the voltage-sensing domain fused to the prokaryotic NavAb channel, as described in 2017.^[377](#CIT0377)^ The chimeric ion channel was purified and reconstituted into nanodiscs for use as antigen. Interestingly, in this study, single domain antibodies were found to engage the ion channel more efficiently. Immunizations of multiple host species that implement both DNA and nanodiscs with and without adjuvant are underway. The resulting immune repertoire will be deep-mined using microfluidics to identify specific mAbs, so further progress remains to be reported.\n\nMore encouragingly for antibody discovery, Amgen has reported the successful targeting of TRPA1 with antagonist murine mAbs that can block multiple modes of TRPA1 activation.^[378](#CIT0378)^ Rather than the use of peptides as antigen, various other formats were utilized for immunization, including DNA, whole cells stably expressing TRPA1 and recombinant adenoviral vector expressing TRPA1 (which incorporated immune-modulating modifications to enhance the immune response). Standard hybridoma methodology was employed, and resulting stable clones were screened for binding to TRPA1 by FACS, followed by evaluation of the purified IgG molecules for functional activity in calcium uptake assays and blocking of TRPA1 activation. A panel of mAbs was identified that demonstrated dose-dependent inhibition of TRPA1 with the most potent mAb exhibiting IC~50~ values of 260\u00a0nM in the agonist (allyl isothiocyanate) blocking assay and 90\u00a0nM in the cold activation assay. Although not confirmed, it has been suggested that these mAbs bind to the pore region of TRPA1, which would be in keeping with other observations that antagonist ion channel targeting antibodies often bind to the third extracellular loop that forms the pore in many ion channels.^[50](#CIT0050),[348](#CIT0348)^ The study also notes that only partial inhibition was achieved. Perhaps other antibody formats, such as nanobodies or Cowbodies that have an ultra-long heavy chain CDR3 loop, would bind epitopes otherwise inaccessible to full-length IgG antibodies and thereby achieve better inhibition.\n\nNovo Nordisk has described the generation of antagonist antibodies to the Orai1 pore-forming subunits of CRAC in a study to assess antibody-mediated effects for SOCE in T cells.^[170](#CIT0170)^ They have successfully identified a specific anti-human Orai1 mAb directed to the second extracellular loop that was generated by using a peptide sequence from the extracellular loop conjugated to bovine serum albumin for immunization and standard hybridoma generation. An ELISA- and FACS-positive antibody, 10F8, was able to inhibit T cell effector function *in vitro*, possibly through the contribution of antibody-mediated internalization of Orai1. Inhibition of effector function was also demonstrated *in vivo* by using a humanized GvHD mouse model, which confirmed a reduction in T cell proliferation and pro-inflammatory cytokine production. In addition, 10F8 was used to characterize Orai1 expression on immune cell subsets from blood and rheumatoid arthritis synovial fluid providing further validation of Orai1 as a target for autoimmune disease. Whilst the efficacy of this mAb has been demonstrated both *in vitro* and *in vivo*, no further information is currently available as to the progress of this molecule.\n\nAmgen has described the targeting of Glycine Receptor \u03b13 (GlyR\u03b13), a Cys-loop receptor class ion channel implicated in pain. Glycine binding to the ECD triggers a conformational change, opening the ion channel to chloride ions. A specific antibody, AM-3607, directed to GlyR\u03b13 was shown to enhance the response to exogenous glycine and was used in structural studies in complex with the pentameric GlyR\u03b13 to further understand potentiator mechanism. The resulting crystal structure was resolved to 2.6\u212b, revealing that the mAb binding site was 10\u212b above the agonist binding site, suggesting a novel positive allosteric binding site.^[379](#CIT0379)^ Whilst this provides valuable insights for structure-based drug discovery, no further therapeutic development is planned for this mAb.\n\nFinally, Regeneron has generated mAbs to the ASIC1 ion channel using a DNA immunization approach and the VelocImmune transgenic mouse platform, as presented in 2016.^[380](#CIT0380)^ Binders to cells expressing ASIC1 were identified from hybridoma clones and antibody diversity was evaluated by differential antigen disruption,^[381](#CIT0381)^ where the effect of chemical disruption of the cell surface antigen is assessed by FACS analysis to produce a heat map that indicates the diversity of epitope coverage. Twelve individual mAbs were identified from 106 binders for further profiling.^[382](#CIT0382)^ Select mAbs were shown to inhibit the pain response *in vivo* in a model of carrageen-induced muscle hyperalgesia, using a dose range of 10--40 mg/kg. However, no further development has recently been reported. More recently, the Shanghai Institute for Advanced Immunochemical Studies in collaboration with the Scripps Research Institute reported the successful isolation of ASIC1a antagonist mAbs by using the nanodisc antigen format. These mAbs were selective and potent in both *in vitro* testing and an *in vivo* middle cerebral artery occlusion (MCAO)-induced ischemia stroke model.[^383^](#CIT0383)\n\nFuture directions {#S0010}\n=================\n\nCurrently, there are no marketed mAb-based therapeutics that target an ion channel, with only one polyclonal antibody targeting a non-functional form of P2X7 (nfP2X7) in Phase 1 clinical development for the treatment of basal cell carcinoma.^[30](#CIT0030)^ Ion channels still remain significantly underexploited as antibody drug targets^[17](#CIT0017)^ due mainly to the challenges of expressing sufficient protein in a biologically relevant conformation for antibody discovery purposes, as well as epitope accessibility and screening approaches. However, advances made in generating crystal structures and, more recently, cryo-electron microscopy structures, coupled with the deepening knowledge of ion channel gating and target biology validation now provide an informed base on which to progress and streamline antibody-based approaches in targeting ion channels to treat a variety of diseases.\n\nConclusions {#S0011}\n===========\n\nIon channels are widely recognized as important therapeutic targets in a range of diseases, but remain a challenge for drug discovery. There is still a paucity of functional ion channel mAbs described in the literature, and several of the examples cited here are drawn from recent conferences reflecting the most current advances and innovations enabling the discovery and development of therapeutic mAbs directed towards this important drug target subclass. Despite the significant focus on Na~v~ channels, in particular Na~v~1.7, very little success has been reported. However, a review of the therapeutic pipeline suggests that the ligand-gated P2X family and K~v~1.3 are likely to yield the initial successes for ion channel-targeting mAbs. The P2X family possesses a larger extracellular region than many other ion channels, and therefore is easier to target. The increasing number of ion channel structures being published will assist our understanding of this drug class and enable a deeper knowledge of the biology involved. Progress has been made in overcoming some of the technical challenges associated with ion channel expression and antibody screening, which are likely to facilitate the identification of new functional antibodies moving forward.\n\nAbbreviations {#S0012}\n=============\n\nADCCantibody-dependent cell-mediated cytotoxicityALDH1aldehyde dehydrogenaseAMLacute myeloid leukaemiaASICacid-sensing ion channelCACCcalcium activated chloride channelCa~v~voltage gated calcium channel family memberCDRComplementarity-determining regionCNScentral nervous systemCFTRcystic fibrosis transmembrane conductance regulatorCRACcalcium release-activated calcium channelCRCcolorectal cancerCSCcancer stem cellECDextracellular domainE. coliEscherichia coliEPCsendothelial progenitor cellsFSGSFocal segmental glomerulosclerosisGPCRG protein-coupled receptorHNSCChead and neck squamous cell carcinomaK~v~voltage-gated potassium channel family membermAbmonoclonal antibodyNa~v~voltage-gated sodium channel family memberNSCLCnon-small cell lung cancerpAbpolyclonal antibodyscFvsingle-chain variable fragmentShKStichodactyla toxinSigR1sigma-1 receptorSMALPsstyrene-maleic acid lipid particlesSOCEstore-operated calcium entrySTIM1stromal interaction molecule 1TMtransmembraneTMDtransmembrane domainTRAILtumor necrosis factor-related apoptosis-inducing ligandVGICvoltage gated ion channelVGCCvoltage gated calcium channels\n\nDisclosure statement {#S0013}\n====================\n\nAuthors T.G. Clark and P. Colussi are employees of TetraGenetics Inc. and C.J. Hutchings provides consulting services to TetraGenetics Inc.\n\n[^1]: Abbreviations and acronyms used in table:\n\n[^2]: AITC, allyl isothiocyanate; AP, acute pancreatitis; ASM, airway smooth muscle; ATP, adenosine triphosphate; BDNF, brain-derived neurotrophic factor; CCI, chronic constriction injury; CFA, complete Freund's adjuvant; CGN, carrageenan; CIA, collagen-induced arthritis; COPD, chronic obstructive pulmonary disease; CRC, colorectal cancer; DED, dry eye disease; DNBS, 2,4-dinitrobenzene sulfonic acid; DTH, delayed type hypersensitivity; DRG, dorsal root ganglion; EAE, experimental autoimmune encephalitis; EAN, experimental neuritis; ER, endoplasmic reticulum; FEPS, familial episodic pain syndrome; FSGS, focal segmental glomerulosclerosis; GBM, glioblastoma; GM-CSF, granulocyte-macrophage colony-stimulating factor; GvHD, graft versus host disease; HNSCC, head and neck squamous cell carcinoma; IBD, inflammatory bowel disorder; IBS, irritable bowel syndrome; ICC, intrahepatic cholangiocarcinoma; IFN-\u03b3, interferon gamma; IHC, immuno-histochemistry; IL-2, interleukin-2; MB, medullablastoma; MCF-7, Michigan Cancer Foundation-7 breast cancer cell line; MS, multiple sclerosis; NOD, non-obese diabetic; NSCLC, non-small-cell lung cancer; OD1, mouse model of Na~v~1.7-mediated pain based on intraplantar injection of the scorpion toxin OD1; OVA, ovalbumin; Ph1, Phase 1 clinical trial; Ph2, Phase 2 clinical trial; Ph3, Phase 3 clinical trial; RA, rheumatoid arthritis; SalB, salvianolic acid B; SCID, severe combined immunodeficiency; ShK, Stichodactyla toxin; SS, Sj\u00f6gren's syndrome; TCR, T cell receptor; T1D, type 1 diabetes; T~EM~, effector memory T lymphocytes; TNF\u03b1, tumor necrosis factor alpha; TDZ, Thioridazine; TTX, tetrodotoxin; VEGF, vascular endothelial growth factor.\n"}
+{"text": "![](brjcancer00155-0012.tif \"scanned-page\"){.375}\n\n![](brjcancer00155-0013.tif \"scanned-page\"){.376}\n\n![](brjcancer00155-0014.tif \"scanned-page\"){.377}\n\n![](brjcancer00155-0015.tif \"scanned-page\"){.378}\n\n![](brjcancer00155-0016.tif \"scanned-page\"){.379}\n\n![](brjcancer00155-0017.tif \"scanned-page\"){.380}\n\n![](brjcancer00155-0018.tif \"scanned-page\"){.381}\n"}
+{"text": "\"Diversity and Quality\". With that in mind, we worked for the 24^th^ International Conference of Indian Association of Palliative Care organized at Coimbatore, India from 10^th^ to 12^th^ February 2017. The conference was jointly organised by the G. Kuppuswamy Naidu Memorial Hospital and Coimbatore Cancer Foundation.\n\nThe theme of our conference was Adding Value to\n\nAwarenessLearningDeciding RightCare until the End.\n\nAs with the previous conferences, we focused to improve the credibility of information and to help advance the effectiveness, and equitability of scientific information disseminated.\n\nOur emphasis on giving members of the audience a range of topics from recent evidence, interactive hands on sessions, topics related to advocacy, team work, organization, and philosophy was well received.\n\nWe are glad we were given a good feedback from the multidisciplinary audience, which numbered close to 1000, which was probably one of the highest numbers participating in the history of this conference. The delegates represented 16 countries and from19 states and 4 Union Territories of India.\n\nWe received a significant number of abstracts from a whole lot of students, doctors, nurses, social workers, psychologists, and members of several palliative care organisations. We had a committee to review and select those for oral and poster presentations. The top three abstracts selected by the reviewers were presented as part of the plenary sessions.\n\nWe also held five excellent Pre conference workshops, three of those were organized by the conference committee; focused on advocacy and training to help awareness locally and in Tamil Nadu.\n\nThe present issue of IJPC is the collection of abstracts presented as papers, posters, and talks. On behalf of the scientific committee, I would like to express my deep gratitude to everyone who has showed immense interest and passion to submit such a variety of abstracts, share their presentations, and articles for publication.\n\nWe would like to thank the Scientific Committee, and the Advisory Board for their expertise and support.\n\nI would like to thank Dr. Odette Spruyt, Dr. Charu Singh, Dr. Anand Narayan, Dr. Latha Balasubramani, Dr. Vidya Kumarasamy, Dr. Rajith and of course Dr. Balaji, the Organising Secretary, who showed us an exemplary example of team work and excellent planning.\n"}
+{"text": "\n"}
+{"text": "INTRODUCTION\n============\n\nEndometriosis is estimated to occur in 6% to 10% of women of reproductive age,^[@B1]^ with a prevalence of 38% (range, 20%--50%) in infertile women,^[@B2]^ and in 71% to 87% of women with chronic pelvic pain.^[@B3]^ In the past, symptomatic moderate to severe endometriosis was most commonly treated by laparotomy with the removal of affected tissue, with or without hysterectomy and bilateral salpingo-oophorectomy. More recently, many women with advanced endometriosis have been treated with a laparoscopic approach because it results in a shorter hospital stay and recovery period compared with laparotomy. However, conventional laparoscopy has inherent limitations for the treatment of advanced endometriosis because of the adhesive nature of the disease, obliteration of the surgical plans, variability of surgical skill levels, and normal mechanics of the human hand.\n\nThe use of a robotic system to assist with laparoscopy has been shown to overcome some of the technical limitations of laparoscopy while maintaining its minimally invasive nature. The most widely used system for robotically assisted laparoscopy is the da Vinci Surgical System (Intuitive Surgical, Sunnyvale, CA, USA). Advantages of this system include a steady 3-dimensional (3D) image and articulating instrumentation that allow for 7 degrees of movement, mimicking the human wrist. It also filters the surgeon\\'s hand tremors at the console and eliminates the fulcrum effect encountered in conventional laparoscopy. Moreover, it allows movement downscaling, increasing accuracy and precision. To date, only a small study has been published that examines the feasibility of robotically assisted laparoscopy for the management of advanced endometriosis.^[@B4]^ The primary objective of this study is to report our experience with the safety and feasibility of robotic surgical treatment of advanced pelvic endometriosis. The secondary objective was to explore whether the stage of endometriosis affected surgical outcome.\n\nMATERIALS AND METHODS\n=====================\n\nThe institutional review board of the University Hospitals Case Medical Center, Cleveland, Ohio, approved this retrospective chart review study. Our series consists of 43 patients who underwent robotically assisted laparoscopic surgery for the treatment of advanced endometriosis between April 2008 and March 2010. All data were collected directly from the patients\\' charts by use of a standardized data collection sheet. The electronic records of 422 patients who underwent robotic surgery during the study period for gynecologic indications were evaluated. Patients with histologically confirmed endometriosis were identified. Operative reports of this cohort were reviewed. The extent of disease as dictated by the primary surgeon was matched against the American Society for Reproductive Medicine (ASRM) classification of endometriosis for staging.^[@B5]^ Patients with stage I and II endometriosis were excluded (n = 11). Patients who underwent conservative surgical management with uterine preservation were also excluded (n = 7). There were no selection criteria to use the robotic platform. This was based solely on surgeon preference, scheduling, and availability.\n\nDemographic and clinical characteristics were recorded, and the patients were divided into 2 groups according to the ASRM classification by endometriosis stage III or IV.^[@B5]^ Outcome variables included total operating room, actual operative time, and postanesthesia care unit (PACU) time (in minutes); estimated blood loss (in milliliters); and uterine weight (in grams). Complications were determined by examination of the hospital chart and the postoperative visit notes within 6 mo after surgery.\n\nOperative Technique\n-------------------\n\nAfter induction of general anesthesia, the patient was positioned in the dorsal lithotomy position with both arms tucked by the patient\\'s side. A beanbag was adjusted to keep the arms and the shoulders in place. Pneumoperitoneum was usually induced with a Veress needle. A 12-mm trocar was placed 2 to 5 cm supraumbilically. Two 8-mm robotic trocars were placed bilaterally, 10 cm lateral to and at the level of the umbilicus. An accessory 10-mm trocar was placed in the left lower quadrant. Monopolar scissors were inserted through the right robotic trocar, and a plasma kinetic dissecting forceps was inserted through the left robotic trocar.\n\nOophorectomy was started by entering the retroperitoneal space of the lateral pelvic sidewall with unipolar electrosurgery. The infundibulopelvic (IP) ligament was then skeletonized. The ureter was identified. Subsequently, unipolar electrosurgery was used to create a window below the IP ligament and above the ureter. The IP was then coagulated 3 times and divided. Alternatively, the IP was coagulated and cut if the ureter could be visualized easily. This process was repeated on the contralateral side when both ovaries were removed. Lysis of adhesions was then performed to restore normal anatomy in patients with extensive adhesions.\n\nUnilateral lateral pelvic sidewall dissection and ureterolysis were performed in a total of 8 patients (3 with stage III and 5 with stage IV disease) to resect deep infiltrating lesions. Partial obliteration of the cul-de-sac was treated with sharp dissection in only 3 patients with stage IV disease. Careful dissection with and without the use of electrocautery was used to preserve the integrity of the ureter, pelvic vessels, and rectum. The hysterectomy portion of the procedure was then completed by a standard approach according to the surgeon\\'s preference.\n\nThe Foley catheter was removed at the end of the procedure. A regular diet was started on the same day of the operation, and patients were allowed to ambulate as soon as possible. The patient was routinely discharged the morning after surgery, unless the patient\\'s clinical condition required continued hospitalization. A routine postoperative visit was made 4 to 6 wk after the surgery and as needed thereafter. Total operative time was defined as the time elapsed from intubation to extubation in minutes. Actual operative time was defined as the time elapsed from skin incision to skin closure. The PACU time was defined as the time from arriving to the PACU until discharge of the patient to the floor.\n\nStatistical Analysis\n--------------------\n\nPatient characteristics and surgical parameters for the entire cohort of patients were described. A secondary goal was to compare these parameters among patients with stage III and stage IV endometriosis. Patient demographic and clinical characteristics were compared between stage III and IV endometriosis patients by use of the Mann-Whitney *U* test for continuous variables and \u03c7^2^ tests for categorical variables. The median total operating room, actual operative, and PACU times were compared for women with stage III and IV endometriosis by use of the Mann-Whitney *U* test. *P* \\< .05 was considered statistically significant. Data were analyzed with SPSS software, version 17.0 (SPSS, Chicago, IL, USA).\n\nRESULTS\n=======\n\nPatient Characteristics\n-----------------------\n\nForty-three women underwent a robotic procedure for advanced endometriosis **([Table 1](#T1){ref-type=\"table\"})**. Of these, 19 (44.2%) had stage III and 24 (55.8%) had stage IV endometriosis. The median patient age was 46 y (range, 32--68 y), and the mean body mass index was 28 kg/m^2^ (range, 20.2--50.2 kg/m^2^). Baseline characteristics were similar in both groups.\n\n###### \n\nBaseline Characteristics and Intraoperative Outcome Measures of Study Population\n\n  Patient/Clinical Demographics                                              Overall (N = 43)   Stage III Endometriosis (n = 19)   Stage IV Endometriosis (n = 24)   *P Value*^[a](#TF1-1){ref-type=\"table-fn\"}^\n  -------------------------------------------------------------------------- ------------------ ---------------------------------- --------------------------------- ---------------------------------------------\n  Age \\[median (range)\\] (y)                                                 46 (32--68)        49 (34--68)                        43.5 (32--61)                     .339\n  BMI^[b](#TF1-2){ref-type=\"table-fn\"}^ \\[mean (SD)\\] (kg/m^2^)              28 (20.2--50.2)    26.8 (21.3--45.4)                  30.4 (20.2--50.2)                 .695\n  Previous surgery                                                           9 (20.9%)          3 (15.8%)                          6 (25%)                           .708\n  Total OR^[b](#TF1-2){ref-type=\"table-fn\"}^ time \\[median (range)\\] (min)   190 (97--368)      190 (97--290)                      183 (138--368)                    .934\n  Actual operative time \\[median (range)\\] (min)                             145 (67--325)      145 (67--234)                      146.5 (102--325)                  .882\n  Uterine weight \\[median (range)\\] (g)                                      121.3 (48--570)    105 (52--376)                      125.1 (48--570)                   .839\n  EBL^[b](#TF1-2){ref-type=\"table-fn\"}^ \\[median (range)\\] (mL)              100 (20--400)      100 (25--325)                      100 (20--400)                     .503\n  Complications intraoperatively                                             2 (4%)             1 (5.2%)                           1 (4.2%)                          .999\n  PACU time \\[median (range)\\] (min)                                         80 (32--165)       87.5 (32--135)                     80.0 (32--165)                    .752\n  Hospital stay                                                                                                                                                      \n  \u2003\u2003\u2003\u20031 d                                                                    41 (95.3%)         19 (100%)                          22 (91.7%)                        .495\n  \u2003\u2003\u2003\u2003\\>1 d                                                                  2 (4.7%)           0                                  2 (8.3%)                          \n  Complications postoperatively                                              1 (2%)             0                                  1 (4.2%)                          .999\n\nComparison of stage III and stage IV endometriosis with Mann-Whitney *U* or \u03c7^2^ test.\n\nBMI = body mass index; EBL = estimated blood loss; OR = operating room.\n\nNine patients (20.9%) had previously undergone laparotomy for various indications. Along with the hysterectomy, both ovaries were removed in 29 of 43 patients (67.5%). In 14 patients (32.5%), only 1 ovary was removed and the contralateral ovary was preserved.\n\nIntraoperative Outcomes\n-----------------------\n\nThe median total operative time was 190 min (range, 97--368 min), including patient positioning, robot docking, performing surgery, and performing closure of the port sites. The median actual operative time was 145 min (range, 67--325 min), and both total operating room time and actual operative time were comparable between the 2 groups **([Table 1](#T1){ref-type=\"table\"})**. There was no difference between the 2 groups regarding estimated blood loss and uterine weight. Pathologic evaluation confirmed the endometriosis diagnosis in all patients. There was 1 conversion to laparotomy because of the size and location of multiple fibroids in a woman with a 14 wk--sized uterus. One patient also required vaginal assistance to expedite completion of her hysterectomy.\n\nPostoperative Outcomes\n----------------------\n\nThere was no significant difference between groups in PACU times, with a median of 80 min (range, 32--165 min) for all patients. Narcotics were given postoperatively as needed. All patients were sent home with oral oxycodone/acetaminophene, 5/325 mg. Histopathologic confirmation of the disease was established postoperatively. Concomitant fibroids were confirmed in 12 patients (27.9%). Almost all patients were discharged the day after surgery (41 of 43 \\[95%\\]); 1 patient who was converted to laparotomy was discharged after 5 d, and another patient stayed for 3 d because of postoperative ileus that resolved spontaneously. One patient was readmitted on postoperative day 11 with fever, chills, and lower abdominal pain. Computed tomography showed a vaginal cuff abscess. She was treated with antibiotics, the abscess was drained vaginally under general anesthesia, and a Foley catheter was placed into the abscess cavity. She remained afebrile and was discharged after 4 d. The Foley catheter was removed after 7 d.\n\nDISCUSSION\n==========\n\nOur study represents the largest cohort of patients with advanced endometriosis managed with definitive surgery by robotic laparoscopy reported in the literature. In this study of 43 patients with stage III and IV endometriosis, intraoperative complication rates were low and only 1 patient required conversion to laparotomy. Moreover, only 1 minor and 1 major postoperative complication occurred, manifesting as a self-limited ileus and a vaginal cuff abscess, respectively. Together these data suggest both the feasibility and safety of this surgical modality for the definitive treatment of severe endometriosis.\n\nThere are very few randomized trials evaluating the different surgical approaches to pelvic endometriosis. Whereas Nezhat et al.^[@B4]^ observed no benefit of robotic over conventional laparoscopy for the surgical treatment of stage I or II endometriosis, we speculate that such comparisons are unlikely to be equivocal for more severe manifestations of endometriosis. A 10% rate of conversion to laparotomy has been reported in patients with severe endometriosis managed with conventional laparoscopy when performed by high-volume, experienced laparoscopic surgeons.^[@B6]^ In our series there were no conversions to laparotomy because of an intraoperative complication; moreover, a 2% overall conversion rate (1 of 50) compares favorably with that reported with conventional laparoscopy. Furthermore, as the complexity of pelvic dissection necessary for the surgical management of severe pelvic endometriosis increases, the advantages provided by the robotic platform become more indispensible. Isolated reports have documented its use in patients with severe endometriosis involving the urinary and gastrointestinal systems.^[@B7],[@B8]^ These reports encourage further exploration of robotic surgery in this population of patients.\n\nAmong its inherent characteristics, the 3D technology of robotic surgery is of particular importance in the surgical management of severe endometriosis. The robotic platform improves the depth of perception and facilitates the resection of deep infiltrating lesions. In addition, the robotic system improves dexterity, filters the surgeon\\'s tremor, and improves intuitive movements.^[@B9],[@B10]^ These features enable the surgeon to execute complicated surgical steps such as re-creation of an obliterated cul-de-sac, lateral pelvic wall dissection and resection of densely adherent endometriomas, ureterolysis, and enterolysis. Moreover, the 7 *df* and 3D visual image permit easier handling of the tissue.^[@B10],[@B11]^\n\nRobotic surgery has several disadvantages compared with traditional laparotomy. These include increased cost; the lack of tactile feedback to the surgeon; the presence of bulky robotic arms, as well as long and thick cords; the inability to move the surgical table once the robot arms are attached; and a limited range of motion with respect to operating in different quadrants in the same case.^[@B4]^ A major potential limitation of robotic surgery is the absence of tactile (also called \"haptic\") feedback. This is of particular importance when the dissection is close to delicate structures, such as the ureter, blood vessels, and rectum. This could lead to an inability to determine the strength needed for suturing without breaking the suture or possible injury to adjacent structures. It is especially true in early cases performed by inexperienced surgeons, and its effect will decrease but not disappear with experience. In addition, visual cues from the 3D image may dampen this limitation.^[@B12]^\n\nThis study has several limitations. First, it is a retrospective study and, as such, has inherent weaknesses. Ideally, a prospective randomized trial with laparotomy or laparoscopy without robotic assistance as a control group will be of interest when analyzing the perioperative outcomes. The study is also limited by the lack of a control group. In addition, it is limited by the lack of long-term outcomes including recurrence of endometriosis. Given the fact that the long-term objective of most patients with pelvic pain due to endometriosis is pain relief, a study with the long-term goal of pain relief is needed. However, most our patients were followed up for up to 1 y after surgery without recurrence of symptoms.\n\nThe ASRM scoring system did not correlate well with either perioperative or postoperative outcome measures in our cohort of patients. Although the ASRM system\\'s limitations in predicting clinically relevant outcomes in patients with endometriosis, such as pelvic pain and fertility, are well described, the scoring system was developed to describe the extent and location of anatomic distortion in patients with endometriosis. Anatomic findings, such as cul-de-sac obliteration, adhesions to the broad ligament and pelvic sidewall, and deep peritoneal implants, are anticipated to add to the complexity of definitive surgery for endometriosis. We regard our finding that the operative time was not significantly correlated with ASRM stage as a significant negative finding. Although these data may suggest limitations in either the functionality or the reproducibility of the ASRM system (or both), one may also surmise that these findings further attest to the ability of the robotic technique to manage distorted pelvic anatomy with greater ease. Because of these questions, long-term prospective studies on the definitive surgical management of severe endometriosis are of great interest. However, investigators will also be challenged to prospectively define meaningful and reproducible inclusion criteria for these studies.\n\nCONCLUSION\n==========\n\nIn this retrospective cohort of 43 patients with stage III and IV endometriosis, we found robotic-assisted laparoscopy to be both safe and feasible. Moreover, we speculate that the unique features of the robotic platform may offer advantages over conventional laparoscopy in patients with more severe forms of endometriosis and allow more women with this condition the opportunity for surgical management by a minimally invasive approach. A prospective study is essential to more fully evaluate the relative merits of the robotic platform in the surgical treatment of severe endometriosis.\n\n[^1]: Presented in part at the annual meeting of the Society of Gynecologic Surgeons, San Antonio, TX, USA, April 10--13, 2011.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nObesity-linked genetic variations in the presence of other routine habits such as smoking, physical inactivity, and unhealthy food intake may greatly raise the risk of a person developing heart diseases (cardiovascular diseases, CVD).\n\nExcess body fat, obesity, is one of the most common disorders in clinical practice. The location of the body fat is a major determinant of the degree of excess morbidity and mortality due to obesity \\[[@B1]\\]. At least two components of body fat are associated with obesity-related adverse health outcomes. These are the amount of subcutaneous truncal or abdominal fat, and the amount of visceral fat located in the abdominal cavity. Each of these components of body fat is associated with varying degrees of metabolic abnormalities and independently predicts adverse health outcomes. Many complex traits are thought to be inherited since they often run in families. However, these complex traits do not show typical mendelian pedigree patterns. These nonmendelian diseases may depend on several susceptibility loci, with a variable contribution from environmental factors. Discovering the major susceptibility locus may be the key to advances in understanding the pathophysiology of a disease.\n\nThere have been several studies using association approaches in order to undertake systematic searches for candidate genes in obesity defined as elevated body mass index (BMI, kg/m^2^) \\[[@B2]\\].\n\nApolipoprotein A-II (APOA-II) is the second most common protein in high-density lipoproteins. APOA-II appears to impair the reverse cholesterol transport and antioxidant function of high-density lipoprotein, which is consistent with the observation that increased APOA-II levels promote the development of atherosclerosis \\[[@B3]\\]. *APOA2* polymorphism (-265T\\>C) has been renamed to c.-492T\\>C, according to the Human Genomic Variation Society version 2012. A functional polymorphism representing a T-to-C substitution at the -492 position of this gene has been associated with waist circumference and lower levels of plasma APOA-II in European men \\[[@B4]\\], suggesting that genetic variation at the *APOA2* may be associated with body fat distribution phenotypes. Lower levels of visceral adipose tissue (VAT), both absolute and relative to their total body fat, have been reported in African-American compared with white women \\[[@B5], [@B6]\\], which may be related to differences in genetic make-up between women of different ethnic backgrounds.\n\nNew obesity loci continue to be identified through genome-wide association studies in populations of increasing size and ethnic diversity \\[[@B7], [@B8]\\] but understanding of the mechanisms by which known genetic variants contribute to obesity remains limited. Several well established obesity candidates encode proteins that appear to modulate obesity risk via energy intake, a key determinant of obesity risk \\[[@B9]\\].\n\nIn a previous investigation carried out on White Americans participating in the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study, the recessive effects for the c.-492T\\>C polymorphism was observed \\[[@B10]\\]. Homozygous individuals for the C allele had higher body mass index (BMI) and obesity risk than the carriers of the T allele, but relationships between *APOA2* c.-492T\\>C genotype and obesity among Egyptian adolescents are unexplored till now.\n\nTherefore, our objectives were to analyzing the association between the *APOA2*c.-492T\\>C polymorphism SNP and the risk of obesity and study its association with anthropometric measurements, body fat distribution, food consumption, and lipid metabolism in a sample of Egyptian adolescents.\n\n2. Materials and Methods {#sec2}\n========================\n\nA descriptive cross-sectional study was conducted on randomly selected 303 Egyptian adolescents. They were 196 obese and 107 nonobese. Their age ranged 16--19 years old and the mean age was 17.45 \u00b1 2.54 years. Obese cases had BMI greater than 95th percentile for age and gender according to the National Egyptian Growth Curves of Children and Adolescents \\[[@B11]\\].\n\nThe data were collected from June 2011 to July 2012 and were extracted from a project entitled \"Obesity among Youth: Lifestyle and Genetic Factors\" funded by the Science and Technology Development Fund (STDF), Egypt. This study protocol was approved by the ethical committee board of the National Research Centre of Egypt (no. 10/223). An informed written consent was obtained from all participants. All individuals were clinically evaluated and anthropometric data were collected.\n\n2.1. Anthropometric Measurements {#sec2.1}\n--------------------------------\n\nAnthropometric variables including height, weight, waist, and hip were measured. Body weight was measured with the patients in light clothing and without shoes. Height was measured with the patients standing with their backs leaning against the stadiometer of the same scale. BMI was calculated as weight in kilograms divided by height in meters squared (kg/m^2^). WC and hip circumference (HC) were measured in cm using a plastic, nonstretchable tailor\\'s tape. WC was measured with light clothing at a level midway between the lower rib margin and the iliac crest standing and breathing normally. HC was measured at the level at the widest circumference over the buttocks (at the greater trochanter). Subsequently, the waist hip ratio (WHR) was calculated as WC divided by HC. Anthropometric measurements were obtained according to standardized equipment and following the recommendations of the International Biological Program \\[[@B12]\\].\n\nSystolic and diastolic blood pressure (BP) were measured with the patients sitting with their left arm at heart level using a professional Riester sphygmomanometer manufactured in Japan. Several measurements were made, from which an average BP measurement was obtained. BF% was measured by Tanita Body Composition Analyzer (SC-330).\n\n2.2. Genotyping {#sec2.2}\n---------------\n\nGenomic DNA was extracted from peripheral blood leukocytes using GeneJET Genomic DNA Purification Kit---(Fermentas, German) according to the manufacturer instructions. Genotyping of c.-492T\\>C polymorphism in the *APOA2* gene was carried out using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis \\[[@B4]\\]. Two pair of primers were used to amplify the promoter region of the *APOA2* gene containing the polymorphism; upstream primer 5\u2032CAT GGG TTG ATA TGT CAG AGC-3\u2032 and downstream primer 5\u2032 TCA GGT GAC AGG GAC TAT GG 3\u2032.\n\nPCR was carried out in a 25\u2009*\u03bc*L total final volume containing 200\u2009*\u03bc*M dNTPs (Finzyme, Finland), 10 pmole of each primer, 2\u2009U of Taq polymerase (Finzyme, Finland), and 500\u2009ng DNA. Thermal cycling conditions were as follows: denaturation at 95\u00b0C for 10\u2009min, followed by 30 cycles of denaturation at 95\u00b0C for 30\u2009sec, annealing at 59.5\u00b0C for 30\u2009sec, and elongation at 72\u00b0C for 30\u2009sec followed by a final elongation of 5\u2009min.\n\nTen *\u03bc*L of successfully amplified PCR products were digested with Fast Digest BsmI enzyme (Fermentas, Germany), incubated at 37\u00b0C for 5\u2009min and the fragments were run in 3% agarose gel stained with ethidium bromide, and analyzed under ultraviolet light. The BsmI enzyme cuts the PCR product (273\u2009bp) in two fragments 215 and 58\u2009bp in presence of the T allele.\n\n2.3. Abdominal Ultrasonographic Examination {#sec2.3}\n-------------------------------------------\n\nUltrasonography was carried out by using GE logic *\u03b1* 200-ultrasound machine. Visceral fat layer was measured from the region just above the umbilicus \\[[@B13]\\]. The convex-array probe (3.5\u2009MHz) was used for measuring visceral abdominal fat and anterior wall of the aorta.\n\n2.4. Dietary Intake {#sec2.4}\n-------------------\n\nFood intake carried out using 24 hours dietary recall. Cases were asked to recall their dietary intakes of the previous 24 hours. In particular, we asked about intake of carbonated beverages, juices, and other casual intake. Food frequency method assessed food consumption frequencies per day and week and month basis by using a questionnaire. It was focused on different kinds of food consumption frequencies rather than consumption of specific nutrients. The energy and nutrient contents were computed.\n\n2.5. Biochemical Analyses {#sec2.5}\n-------------------------\n\nHDL cholesterol was measured after precipitation of non-HDL cholesterol with magnesium/dextran. We measured LDL cholesterol by use of a homogeneous direct method (LDL Direct Liquid Select Cholesterol Reagent; Equal Diagnostics) on the Hitachi autoanalyzer 704 (Roche Diagnostics Switzerland).\n\n2.6. Statistical Analysis {#sec2.6}\n-------------------------\n\nQuantitative variables were expressed as mean \u00b1 S.D., and qualitative variables were expressed as percentages. Differences between groups were tested using an independent two-sample *t*-test and chi-square test was used to test for differences in the distribution of categorical variables. *P* values \\< 0.05 were considered statistically significant.\n\n3. Results {#sec3}\n==========\n\nThe characteristics of the obese cases and nonobese are given in [Table 1](#tab1){ref-type=\"table\"}, where we compare the anthropometric and clinical variables for the obese and nonobese individuals. There were significant differences in BMI, WC, WHR, BF%, and visceral fat thickness between the two groups. Obese adolescents had higher values than nonobese in both genders. *APOA2*genotype frequencies did not deviate from Hardy-Weinberg equilibrium expectations and did not differ between males and females. Therefore, males and females were analyzed together. The genotype and allele distribution are presented in [Table 2](#tab2){ref-type=\"table\"}. The data indicate that genotype tends to differ significantly between obese and nonobese adolescents in the CC and TT genotypes, remaining significant when the genotypes in CC and CT+TT were grouped; the CC genotype was more common in obese cases than in nonobese (CC frequency 32.1% in obese cases and 9.3% in lean controls, *P* \\< 0.001). The allele frequency of the *APOA2* c.-492T\\>C polymorphism was also significantly different between the two groups ([Table 2](#tab2){ref-type=\"table\"}).\n\n*APOA2* genotype was evaluated by comparing homozygous minor allele carriers (CC) with combined homozygous major (TT) and heterozygous (TC) subjects ([Table 3](#tab3){ref-type=\"table\"}). After adjusting for age, sex, and BMI, anthropometric measures showed significant differences between homozygous and heterozygous carriers. BMI, WC, BF%, and visceral fat were significantly higher in CC subjects compared with combined heterozygotes (TC) and homozygous major (TT) carriers (*P* \\< 0.001). Significant elevated HDL-C was observed in CC subjects compared with the carriers of T allele.\n\nHomozygous individuals for the CC allele had a statistically higher mean of energy intake, total fat intake (g/day), and saturated fat (SATFAT) than carriers of the T allele.\n\n4. Discussion {#sec4}\n=============\n\nThe present study found strong association between the *APOA2* c.-492T\\>C SNP polymorphism and obesity risk and anthropometric measures. The study observed that CC homozygotes had higher BMI, WC, BF%, visceral fat, food consumption, and HDL-C than carriers of the T allele. These results were consistent with the findings of the previous study of overweight individuals in other populations. Relatively few studies reported the association between *APOA2* polymorphisms and phenotypic traits \\[[@B14]--[@B17]\\]. Few genetic variants have been identified in the *APOA2* gene \\[[@B18]\\]. Interestingly, a T-C transition at position -492 affecting element of the *APOA2* promoter has been reported to be functional in 2 independent studies, both demonstrating an \u221230% drop in basal transcription activity \\[[@B4], [@B19]\\]. In one of these studies, the *APOA2* polymorphism was associated with waist circumference in men \\[[@B4]\\]. Another study \\[[@B19]\\] reported an association between this polymorphism and abdominal fat depots in women.\n\nAssociation between *APOA2* c.-492T\\>C SNP and BMI or obesity only in the presence of high-saturated fat intake in three American populations has been observed \\[[@B20]\\]. Moreover, with this gene-diet interaction other studies extend the findings to other geographical areas (Europe and Asia), reporting that when saturated fat intake is low (\\<22\u2009g/d), this SNP does not have any effect on BMI or obesity. However, when saturated fat intake is high (\u226522\u2009g/d), significant differences in anthropometric variables were detected between CC individuals and T allele carriers. Further adjustment for other macronutrients did not change the significance of these findings, supporting the specificity of saturated fat as a driver of this interaction \\[[@B21]\\]. Moreover, other study reported genotype-associated differences in specific intake-related behaviours, which may contribute to obesity risk, identifying the possible role of ghrelin in modulating *APOA2*-nutrient interactions \\[[@B22]\\]. Eating behaviours have been identified as related to obesity risk \\[[@B23], [@B24]\\] and appear to be associated with *APOA2* genotype in a manner consistent with obesity risk. Relationship between *APOA2*, saturated fat, and hormonal regulation of food intake has also been identified, which may be relevant to weight control. The interactions between *APOA2* and saturated fat for obesity may be mediated via modulation of plasma ghrelin and expansion of knowledge of APOA2 and obesity to include modulation of specific behaviours and hormonal mediators not only broadens understanding of gene-diet interactions, but also facilitates the pragmatic, future goal of developing dietary guidelines based on genotype \\[[@B22]\\]. Lower saturated fat was associated with lower ghrelin in CC carriers, which may theoretically be expected to accompany lower energy intake and smaller body size.\n\nDespite the scarcity of previous data supporting a role of APOA2 in regulating food intake, numerous experimental evidence demonstrates a pivotal role of another apolipoprotein, APOA4, as a satiety signal \\[[@B25], [@B26]\\]. Fujimoto et al. \\[[@B25]\\] were the first to report that APOA4 is a satiety factor secreted by the intestine after fat absorption and that this function of APOA4 is not shared by gut APOA1. APOA2 is a member of the apolipoprotein multigene super family, which includes genes encoding soluble apolipoproteins (e.g., APOA1 and APOA4) that share genomic structure and several functions. Although all these apolipoprotein genes have been found to be related to obesity in at least one epidemiological study \\[[@B27]\\], only APOA4 has been subscribed in regulation of food intake, acting as a satiety signal. The present study shows association between the *APOA2* polymorphism and food consumption, suggesting a potential new role of APOA2 in the regulation of human appetite. Moreover, the present study shows that the c.-492T\\>C locus is an important genetic determinant of HDL cholesterol concentration. The mechanisms of this proatherogenic capability of increased human apoA-II could be due to increased concentration of apoB-containing lipoprotein and decreased HDL cholesterol, impairment of reverse cholesterol transport due to decreased cholesterol efflux and esterification \\[[@B28], [@B29]\\].\n\nIn summary, the present study emphasized that the homozygous individuals for the CC allele had higher obesity risk than carriers of the T allele. The functional polymorphism representing a T-to-C substitution at the -492 position of this gene is associated with visceral adipose tissue and food consumption. Moreover, this polymorphism had a significant role on HDL cholesterol concentration and could be a modifier gene for familial combined hyperlipidemia.\n\nThe authors are greatly thankful to the Science and Technology Development Fund (STDF) for funding the project entitled \"Obesity among Youth: Lifestyle and Genetic Factors\" (1225) that enabled us to use the data to establish this work.\n\n###### \n\nGeneral characteristics of obese and non-obese adolescents.\n\n  -----------------------------------------------------------------------------------------------\n  Parameters\\           Females             Males                                \n  --------------------- ------------------- ---------------- ------------------- ----------------\n  BMI (kg/m^2^)         34.40 \u00b1 5.96\\*\\*    21.99 \u00b1 3.45     32.75 \u00b1 6.075\\*\\*   19.86 \u00b1 3.33\n\n  Waist (cm)            95.90 \u00b1 12.23\\*\\*   72.57 \u00b1 8.94     98.05 \u00b1 18.90\\*\\*   71.79 \u00b1 8.54\n\n  WHR                   0.82 \u00b1 0.06\\*\\*     0.77 \u00b1 0.16      0.88 \u00b1 0.129\\*\\*    0.81 \u00b1 0.06\n\n  BF%\\                  33.74 \u00b1 13.05\\*\\*   30.96 \u00b1 12.93    33.88 \u00b1 13.78\\*\\*   26.63 \u00b1 11.61\n\n  Systolic BP (mmHg)    110.77 \u00b1 14.81      109.60 \u00b1 13.09   110.67 \u00b1 16.37      108.80 \u00b1 13.23\n\n  Diastolic BP (mmHg)   71.79 \u00b1 10.31       71.80 \u00b1 7.93     71.42 \u00b1 9.746       73.00 \u00b1 10.00\n  -----------------------------------------------------------------------------------------------\n\n\\*\\*Statistically significant differences (*P* \\< 0.01) between obese and non-obese subjects.\n\n###### \n\nGenotype and allelic distribution of *APOA2* c.-492T\\>C polymorphisms.\n\n  -------------------------------------------------------------------------------------------\n  Polymorphisms   Genotypes   Alleles   Grouped genotypes                             \n  --------------- ----------- --------- ------------------- -------- ------- -------- -------\n  Controls\\       10          64        33                  84       130     10       97\n  (*n* = 107)                                                                         \n\n  9.34%           59.81%      30.84%    39.25%              60.74%   9.34%   90.65%   \n\n  Obese\\          63          62        71                  188      204     63       133\n  (*n* = 196)                                                                         \n\n  32.1%           31.6%       36.2%     47.96%              52.04%   32.1%   67.9%    \n\n  *P* value       0.001       0.001     0.022               0.01     0.01    0.001    0.001\n  -------------------------------------------------------------------------------------------\n\n###### \n\nAssociations of *APOA2* c.-492T\\>C polymorphism genotype with anthropometric parameters^a^.\n\n  ---------------------------------------------------------------\n  Parameters\\               CC (*n* = 73)     TT+TC (*n* = 230)\n  ------------------------- ----------------- -------------------\n  BMI (kg/m^2^)             32.29 \u00b1 7.48\\*    23.77 \u00b1 8.25\n\n  Waist (cm)\\               87.91 \u00b1 15.81\\*   82.84 \u00b1 17.04\n\n  WHR\\                      0.89 \u00b1 0.11       0.82 \u00b1 0.19\n\n  BF%\\                      35.69 \u00b1 11.37\\*   30.87 \u00b1 12.43\n\n  Systolic BP (mmHg)\\       108.70 \u00b1 16.24    110.66 \u00b1 13.89\n\n  Diastolic BP (mmHg)\\      71.30 \u00b1 10.41     72.11 \u00b1 9.08\n\n  Energy intake (Kcal/d)\\   1957.1 \u00b1 87.9\\*   1499.7 \u00b1 87.1\n\n  Total fat (g/d)\\          93.6 \u00b1 31.6\\*     51.5 \u00b1 29.5\n\n  SATFAT (g/d)\\             33.9 \u00b1 8.95\\*     22.2 \u00b1 7.91\n\n  HDL-C (mg/dL)\\            56.60 \u00b1 7.88\\*    38.41 \u00b1 7.81\n\n  LDL-C (mg/dL)\\            115.68 \u00b1 38.87    110.64 \u00b1 36.10\n  ---------------------------------------------------------------\n\n\\*Statistically significant differences (*P* \\< 0.05) between CC homozygous subjects and carriers of the T allele for the corresponding variable.\n\n^a^Data are adjusted for age, gender, and BMI.\n\n[^1]: Academic Editor: Francisco Blanco-Vaca\n"}
+{"text": "1. Introduction {#sec1-polymers-10-00758}\n===============\n\nBiodegradable polymers refer to a category of polymers that can be cleaved into small polymer fragments in vivo. The biodegradability endows these polymers with many special applications particularly in drug delivery, tissue regeneration and biotherapeutics \\[[@B1-polymers-10-00758],[@B2-polymers-10-00758],[@B3-polymers-10-00758]\\]. Methods for the preparation of biodegradable polymers can be versatile. Voit and Lederer reviewed the synthesis and major characterizations of hyperbranched and highly branched polymer architectures using polycondensation, addition step-growth reaction and cycloaddition reactions, self-condensing vinyl polymerization and ring-opening multi-branching techniques \\[[@B4-polymers-10-00758]\\]. The exploitation of \"green\" atom transfer radical polymerization (ATRP) and ring-opening polymerization (ROP) to design well-defined and eco-friendly polymeric materials such as biodegradable polymers, polymer brushes, nonionic polymeric surfactants, etc. was reviewed by Tsarevsky and Matyjaszewski \\[[@B5-polymers-10-00758]\\]. Utilizing various polymers for fabricating the more complicated polymeric particles, e.g., micelles, vesicles and capsules, has also been well-documented \\[[@B6-polymers-10-00758],[@B7-polymers-10-00758],[@B8-polymers-10-00758]\\]. Reversible deactivation radical polymerizations (RDRPs) is a relatively new polymerization technique but has already been well-explored. Due to its advantages over other techniques on the preparation of well-defined polymers with low molecular weight distributions, particularly in the preparation of versatile hyperbranched and multi-functional polymeric architectures, in this review, we mainly focus on discussing the preparation of versatile polymeric architectures via RDRPs.\n\n1.1. Varied Polymeric Architectures {#sec1dot1-polymers-10-00758}\n-----------------------------------\n\nPolymeric architectures are very versatile. Based on the composition, they can be homopolymers, or block, statistical, gradient and graft copolymers. Based on the structure, they can be designed as linear, multi-armed, comb-like, networks, and hyperbranched polymers. They can also be tailored with single, multi-, homo-, hetero- or multi-functionalities. These broad polymeric architectures can be fabricated into various complicated particles via either self-assembly or designed interactions, such as micelles, vesicles, capsules, hydrogels and nanogels ([Scheme 1](#polymers-10-00758-sch001){ref-type=\"scheme\"}). Because RDRPs have controlled and living polymerization properties and the chain transfer agent (CTA) employed for the RDRPs can be flexibly designed, for instance, as linear, multi-armed or functional, they are convenient tools for the synthesis of the more complicated architectures. The combination of different RDRPs methods is usually the solution for generation of the more complicated polymeric architectures \\[[@B9-polymers-10-00758],[@B10-polymers-10-00758]\\].\n\n1.2. Reversible Deactivation Radical Polymerizations {#sec1dot2-polymers-10-00758}\n----------------------------------------------------\n\nIn addition to ionic and coordination ring-opening polymerization \\[[@B11-polymers-10-00758],[@B12-polymers-10-00758]\\], free radical polymerization RDRPs have been exploited extensively to generate multi-armed structures with predetermined molecular weights and narrow molecular weight distributions. ATRP \\[[@B13-polymers-10-00758],[@B14-polymers-10-00758]\\], nitroxide mediated radical polymerization (NMRP) \\[[@B15-polymers-10-00758]\\] and reversible addition fragmentation chain transfer (RAFT) polymerization \\[[@B16-polymers-10-00758],[@B17-polymers-10-00758],[@B18-polymers-10-00758],[@B19-polymers-10-00758],[@B20-polymers-10-00758]\\] are the most explored RDRPs ([Figure 1](#polymers-10-00758-f001){ref-type=\"fig\"}). ATRP is one of the most studied RDRPs and many articles have been published about this topic since its development in 1995 by Matyjaszewski \\[[@B21-polymers-10-00758],[@B22-polymers-10-00758]\\]. ATRP is usually initiated by a halogenated organic species in the presence of a metal halide. The metal has a number of different oxidation states that allows it to attract a halide from the organohalide, creating a radical that then starts free radical polymerization. ATRP is an excellent tool for the synthesis of well-defined polymers, however the low solubility of metal halides may limit the catalyst availability and the residual catalyst among the as-prepared polymers may limit the applications in biological field and electronic devices \\[[@B23-polymers-10-00758]\\]. RAFT polymerization was discovered by Rizzardo et al. only two decade ago, but has also been well-explored and -employed to synthesize polymers with predetermined molecular weight and narrow molecular weight distributions over a wide range of monomers. RAFT technique is suitable for polymerizing versatile monomers in different media, where solution (either in organic or aqueous media), emulsion and suspension polymerizations can be carried out for purposely generating functionalized polymers. These functional groups can also be exploited for further polymerization or further reaction to form complicated architectures. RAFT polymerization, in comparison with ATRP, can be undertaken without the introduction of metal ion catalysts, therefore, it will be a secure tool particularly in biological and electrical applications \\[[@B17-polymers-10-00758],[@B24-polymers-10-00758],[@B25-polymers-10-00758]\\].\n\n1.3. Necessity for Making Biodegradable Polymeric Architectures {#sec1dot3-polymers-10-00758}\n---------------------------------------------------------------\n\nBiodegradable polymeric architectures have many advantages that could be envisioned \\[[@B26-polymers-10-00758]\\]. First, previous research revealed that polymers with high molecular weight over 50,000 g\u00b7mol^\u22121^ will exhibit significantly increased circulation time in the body since the glomerular filtration in the kidney has a molecular weight cut-off of about 50,000 g\u00b7mol^\u22121^ \\[[@B27-polymers-10-00758]\\]. Biodegradable polymeric architectures tend to be cleaved into smaller fragments in vivo and subsequently excreted out of the body, which will greatly help clean the polymer fragments within the body. Second, biodegradable polymeric architectures will offer important applications in bio-therapeutics. For example, protein and peptide drugs hold great promise as therapeutic agents. However, most of these drugs can be degraded by proteolytic enzymes and rapidly cleared by the kidneys, resulting in a short circulating half-life. Fortunately, when polyethylene glycol chains are attached to protein and peptide drugs, their circulation time and pharmacokinetics can be significantly improved \\[[@B28-polymers-10-00758]\\]. Third, another advantage is that when the biodegradable polymers are employed to fabricate nanoparticles as drug carriers, the drug release can be realized via the disintegration of polymeric nanoparticles upon biodegradation in vivo.\n\n1.4. How to Confer Biodegradability to Polymeric Architectures {#sec1dot4-polymers-10-00758}\n--------------------------------------------------------------\n\nTo confer biodegradability to polymers, they have to be designed with intra-linkers that can be cleaved by either physiological substances (e.g., glutathione) or enzymatic catalysis \\[[@B29-polymers-10-00758],[@B30-polymers-10-00758]\\]. The biodegradable linkages can be tailored on the polymer backbones, on the side chains, on the cross-linking agents, etc. Several covalent linkages are biodegradable, e.g., the acetal linkage is acid labile \\[[@B31-polymers-10-00758]\\]; the ester linkage is degradable upon hydrolysis \\[[@B32-polymers-10-00758],[@B33-polymers-10-00758]\\]; disulfide bond is cleavable in the presence of glutathione (GSH), the most abundant intracellular thiol (0.2--10 mM) in most mammalian and many prokaryotic cells \\[[@B34-polymers-10-00758],[@B35-polymers-10-00758],[@B36-polymers-10-00758]\\]; and polymers such as polycaprolactone (PCL) \\[[@B37-polymers-10-00758]\\] and poly(amino acid)s \\[[@B38-polymers-10-00758]\\] with polypeptide backbone can be degraded in biological environments by enzymes such as proteinases and peptidases.\n\n1.5. Scope of the Review {#sec1dot5-polymers-10-00758}\n------------------------\n\nThis review discusses the synthesis of versatile biodegradable polymeric architectures that undergo biodegradation using the technique of RDRPs, and their biomedical applications, such as gene/drug delivery, controlled release, targeting biotherapeutics, nanomedicine and so on are also highlighted.\n\n2. Biodegradable Polymeric Architectures {#sec2-polymers-10-00758}\n========================================\n\n2.1. Well-Defined Star-Shaped Structures {#sec2dot1-polymers-10-00758}\n----------------------------------------\n\nWell-defined polymeric structures, e.g., star polymers, are of particular significance in biological applications such as drug delivery and bio-therapeutics \\[[@B39-polymers-10-00758]\\]. Generally, star polymeric structures can be synthesized via \"arm-first\" or \"core first\" methodologies. The \"arm-first\" methodology can be used to generate multi-armed structures by either cross-linking the linear polymeric chains or post-polymerization conjugation of linear functionalized polymeric chains to a multi-functional core via chemo-selectively covalent coupling or non-covalent interactions, e.g., metal ion mediated coordination \\[[@B9-polymers-10-00758],[@B40-polymers-10-00758],[@B41-polymers-10-00758],[@B42-polymers-10-00758],[@B43-polymers-10-00758],[@B44-polymers-10-00758],[@B45-polymers-10-00758],[@B46-polymers-10-00758],[@B47-polymers-10-00758],[@B48-polymers-10-00758]\\]. The \"core first\" strategy is more straightforward, and therefore has attracted an increasing interest for generating multi-armed polymeric architectures in a more controllable mode using multi-functional chain transfer agent \\[[@B49-polymers-10-00758],[@B50-polymers-10-00758],[@B51-polymers-10-00758],[@B52-polymers-10-00758],[@B53-polymers-10-00758],[@B54-polymers-10-00758]\\]. Star polymers consisting of miktoarms have also been tailored to achieve different properties \\[[@B10-polymers-10-00758],[@B42-polymers-10-00758],[@B55-polymers-10-00758],[@B56-polymers-10-00758]\\].\n\nMulti-armed star polymeric architectures have attracted increasing interest due to their potential applications in a number of areas, e.g., encapsulation, sensing, catalysis, electronics, optics, biological engineering, coatings, additives, and drug and gene delivery \\[[@B57-polymers-10-00758],[@B58-polymers-10-00758]\\]. In recent studies, Davis and coworkers successfully demonstrated the synthesis of three-armed star polymeric architectures using both \"core first\" and \"arm first\" methodologies to generate three-armed architecture containing biodegradable disulfide linkages. When \"arm first\" method was adopted, the linear polymer chain was tailored with thiol-reactive pyridyl disulfide groups, through which the linear chains were attached onto a tri-thiol functional core to afford three-armed star polymeric structure. At the same time, the \"core first\" technique was also utilized to generate the same three-armed star polymers from RAFT controlled polymerization using a trifunctional RAFT agent ([Figure 2](#polymers-10-00758-f002){ref-type=\"fig\"}a). Gel permeation chromatography (GPC) and electrospray ionization (ESI) mass spectroscopy analysis evidenced the well-controlled RAFT polymerization which yielded well-defined three-armed star structures with polydispersity index (PDI) less than 1.28. The R group was designed at the end of the RAFT agent, through which the as-prepared polymer chains would sit outside of the RAFT active centers, that is, at the end of each arm. Further modification of the RAFT cores would risk polymeric chain loss. This design would compromise the application when modification of the trithiocarbonate or dithioester RAFT cores is required. To overcome this drawback, Davis and coworkers designed a three-armed RAFT agent via a condensation reaction between the *R*-group of the RAFT agent and a trifunctional core to afford a trifunctional RAFT agent with *Z*-groups at the end of each arm. The subsequent polymerizations of styrene and PEG-A using this RAFT agent generated three-armed polymeric structures with trithiocarbonate cores at the end of each arm, endowing the potential for further modifications through the RAFT cores ([Figure 2](#polymers-10-00758-f002){ref-type=\"fig\"}b) \\[[@B59-polymers-10-00758]\\]. Aminolysis of the trithiocarbonate cores and further reaction with dithiodipyridine (DTDP) yielded sulfhydryl groups and subsequently pyridyldisulfide (PDS) terminal groups, available for further reactions with any free thiol-tethered precursors. When the ends of the star polymers were modified with cholesterol groups, \u03b1-cyclodextrin (\u03b1-CD) groups were attached successfully via inclusion complexation. The generated architecture can be easily degraded in the presence of DTT due to the introduction of disulfide linkages. The methodology presented here can be a prototype research for post-polymerization modifications of various polymeric architectures prepared by RAFT mechanism.\n\nAs an extension, a six-armed star architecture with disulfide intra-linkages on each arm was also synthesized using \"core-first\" methodologies, where a six-armed RAFT agent was synthesized first by attaching the RAFT agent via its *Z*-group to a core that has six RAFT active sites, followed by the RAFT mediated polymerization \\[[@B49-polymers-10-00758],[@B60-polymers-10-00758]\\]. The PDIs of the six-armed star polymers with amphiphilic copolymer arms of poly(St-*b*-PEG-A) were less than 1.31 for the copolymers up to 80% conversion, indicating a well-controlled mechanism by RAFT. After cleavage in the presence of [dl]{.smallcaps}-Dithiothereitol (DTT), the PDI of the single-armed chains was measured to be 1.20 by GPC, in accordance with the successful living polymerization. It should be emphasized that a lower PDI is not necessarily indicative of instantaneous arm growth from all thiocarbonate sites \\[[@B61-polymers-10-00758]\\], as the fragmentation of the initial RAFT functionality may not favor the initiating group (R-group). This may be a noticeable problem at very low conversions, but as conversion proceeds, and the main RAFT equilibrium is attained, this is unlikely to become a significant influence on the kinetics and/or architectures.\n\nIn Li's study, biodegradable star-shaped poly(*\u025b*-caprolactone) and poly(*\u025b*-caprolactone-*b*-[l]{.smallcaps}-lactide) (5sPCL-*b*-PLLA) with five arms were synthesized by ring-opening polymerization (ROP) from an asymmetric core. Subsequently, a series of amphiphilic and double responsive star-block copolymers were synthesized by RAFT star polymerization of *N*,*N*-dimethylamino-2-ethyl methacrylate (DMAEMA) from the star-shaped macro-RAFT agent, which was prepared by attaching 3-benzylsulfanylthiocarbonylthiocarbonylsufanylpropionic acid (BSPA) to 5sPCL-*b*-PLLA using a simple two-step reaction sequence. GPC and ^1^H-NMR measurements demonstrated the polymerization courses are under control. The molecular weight of 5sPCL-*b*-PLLA-*b*-DMAEMA increased with increasing monomer conversion and the molecular weight distribution ranged 1.19--1.37. Spherical micelles with degradable core and pH and thermo-double sensitive shell were prepared from the aqueous medium of the amphiphilic star-shaped copolymers through a dialysis method. Both pH and thermal-responsive behaviors of the copolymer micelles in this study were investigated ([Figure 3](#polymers-10-00758-f003){ref-type=\"fig\"}) \\[[@B62-polymers-10-00758]\\]. In addition to the well-defined symmetrical multi-armed polymeric structures, biodegradable, penta-armed star-block copolymers were also synthesized via an asymmetric core by combination of ROP and RAFT polymerizations, where the five-armed macro-RAFT agent was prepared by ROP on each arm.\n\nTo generate more complicated polymeric architectures, combined methods should be more effective \\[[@B63-polymers-10-00758]\\]. Qiao and Wiltshire \\[[@B64-polymers-10-00758]\\] synthesized the degradable polyester-based star polymers with a high level of functionality in the arms via the \"arms first\" approach using an acetylene-functional block copolymer macroinitiator. This was achieved by using 2-hydroxyethyl 2\u2032-methyl-2\u2032-bromopropionate to initiate the ROP of caprolactone monomer, followed by ATRP of a protected acetylene monomer, (trimethylsilyl) propargyl methacrylate. The hydroxyl end-group of the resulting block copolymer macroinitiator was subsequently cross-linked under ROP conditions using a bislactone monomer, 4,4\u2032-bioxepanyl-7,7\u2032-dione, to generate a degradable core cross-linked star (CCS) polymer with protected acetylene groups in the corona. After removal of trimethylsilyl-protecting groups the resulting pendent acetylene groups were then reacted with azide-functionalized linear polystyrene via a copper-catalyzed cycloaddition reaction between azide and acetylene functionalities. The \"brush-like\" arms could be cleaved via the hydrolysis of polyester star structure to generate molecular brushes. Combining RAFT polymerization with ATRP and hetero-Diels-Alder chemistry, Sinnwell et al. successfully prepared 12-armed star block copolymers. The biodegradable ester linkages between the arm and core confer the biodegradability to the generated polymeric architectures \\[[@B65-polymers-10-00758]\\]. Star-shaped block copolymers with a biodegradable poly(lactide) core were also synthesized using RAFT polymerization combining copper-catalyzed Huisgen 1, 3-dipolar cycloaddition and thiol-ene Michael additions \\[[@B66-polymers-10-00758]\\].\n\nFew star-shaped thermoresponsive polymers with six arms were prepared via RAFT polymerization by Cortez-Lemus's group. Star polymers with homopolymeric arms of poly(*N*-vinylcaprolactam) (PNVCL), copolymeric arms of poly(*N*-vinylcaprolactam-*co*-*N*-vinylpyrrolidone) (PNVCL-*co*-PNVP) and arms of block copolymers of poly(*N*-vinylcaprolactam-*b*-Vinyl acetate) (PNVCL-*b*-PVAc) and (PNVCL-*co*-PNVP)-*b*-PVAc were achieved by exploiting the R-RAFT synthetic methodology (or R-group approach), where the thiocarbonyl group is transferred to the polymeric chain end. Removing the xanthate group of the star polymers allowed for the introduction of specific functional groups at the ends of the star arms and resulted in an increase of the lower critical solution temperature (LCST) values. These star block copolymers could self-assemble into single flowerlike micelles, showing great stability in aqueous solution. Micellar aggregates of selected star polymers were used to encapsulate methotrexate showing their potential in the temperature controlled release of this antineoplasic drug ([Figure 4](#polymers-10-00758-f004){ref-type=\"fig\"}) \\[[@B67-polymers-10-00758]\\].\n\nIn another study reported by Qiao and Wiltshire, the synthesis of selectively degradable core cross-linked star polymers using ATRP and ROP was presented \\[[@B68-polymers-10-00758]\\]. In their study, both the arms and the core can be designed to be biodegradable and selectively degraded. The arms were also designed to be the same or different. The multifunctional initiator, 2-hydroxyethyl 2\u2032-methyl-2\u2032-bromopropionate was used to synthesize degradable poly(*\u03b5*-caprolactone) (PCL) and nondegradable polystyrene (PSt) and poly(methyl methacrylate) (PMMA) macro-initiators, which were subsequently cross-linked to generate core cross-linked star (CCS) polymers. By using the non-degradable divinylbenzene (DVB) and ethylene glycol dimethacrylate (EGDMA) as well as the degradable (4,4\u2032-bioxepanyl-7,7\u2032-dione (BOD) and 2,2-bis(*\u03b5*-caprolactone-4-yl)propane (BCP) monomers to cross-link the different macro-initiators, a range of CCS polymers were synthesized where either the arm or the core domain can be selectively degraded. Hydrolysis of PCL/PMMA/EGDMA miktoarm CCS polymer resulted in CCS polymer with a reduced number of arms, whereas PSt/BOD core-degradable CCS polymer yielded the original linear PSt arms upon hydrolysis.\n\nSimilarly, Schramm et al. \\[[@B69-polymers-10-00758]\\] also reported the synthesis of well-defined 4-, 6-, 8- and 12-armed star polymers with biodegradable PCL biodegradable cores, poly(*\u03b5*-caprolactone)-*b*-poly(ethylene glycol) methacrylates (PEGMAs) using ATRP and ROP. These multi-armed star architectures exhibited unimolecular behavior and the capability of encapsulation of hydrophobic molecules, therefore they are potential candidates as hydrophobic anticancer drug carriers. Likewise, thermosensitive four armed triblock copolymers comprised of poly(*\u03b5*-caprolactone), poly(olego(ethylene oxide) methacrylate) and poly(di(ethylene oxide)methyl ether methacrylate) segments were also synthesized by ATRP and ROP joint methods using a four armed initiator. These four armed polymeric structures were found to be able to self-assemble into spherical micelles which undergo reversible sol-gel transitions between room temperature (22 \u00b0C) and human body temperature (37 \u00b0C) \\[[@B70-polymers-10-00758]\\]. Well-defined dendrimer-like star block copolymers up to 24 arms have also been successfully achieved by combination of ROP and ATRP using \"core-first\" methodology \\[[@B71-polymers-10-00758]\\].\n\n2.2. Cross-Linked (Highly Branched) Structures {#sec2dot2-polymers-10-00758}\n----------------------------------------------\n\nRAFT polymerization can be a convenient tool for generating functionalized and biodegradable macro-monomers via wisely tailored RAFT agent. Davis and coworkers synthesized a novel AB~2~ macro-monomers bearing \u03b1-dithiobenzoate and \u03c9-double pyridyl disulfide end-groups through a straightforward synthetic approach \\[[@B72-polymers-10-00758]\\]. These monomers were prepared by RAFT polymerization, after which the \u03b1-dithobenzoate functionality was aminolyzed to yield thiols that were simultaneously subjected to an exchange reaction with pyridyl disulfide at the chain ends, resulting in the formation of hyperbranched structures, which could proceed disulfide mediated degradation in the presence of reducing agent such as [dl]{.smallcaps}-Dithiothreitol (DTT), Tris(2-carboxyethyl)phosphine (TCEP) or glutathione. Biodegradable hyperbranched cationic polymers, poly(2-(dimethylamino)ethyl methacrylate) (PDMAEMA), have also been synthesized via RAFT mechanism for DNA binding and delivery \\[[@B73-polymers-10-00758]\\].\n\nIn addition to the biodegradable linkages, when the biodegradable polymers such as poly(lactide) (PLA) or polycaprolactone (PCL) are incorporated into cross-linked or self-assembled polymeric architectures, biodegradability can also be achieved. Schubert, Hoogenboom and coworkers synthesized a well-defined biodegradable macro-monomer, oligo(2-ethyl-2-oxazoline) methacrylate by direct end-capping of living oligo(2-ethyl-2-oxazoline) chains with in situ formed triethylammonium methacrylate, followed by homopolymerization via RAFT mechanism and then copolymerization using the homopolymer as macro-RAFT agent to achieve comb-like biodegradable architectures \\[[@B74-polymers-10-00758]\\]. Despite the same combined polymerization techniques being used, different polymerization sequence may afford completely different polymeric structures. In research by Thurecht and coworkers, RAFT polymerization and ROP were used to synthesize both hyperbranched and microgel particles \\[[@B75-polymers-10-00758]\\]. The core-first method afforded the hyperbranched core--shell structure, whereas the arm-first method gave core-cross-linked shell particles ([Figure 5](#polymers-10-00758-f005){ref-type=\"fig\"}).\n\nIn addition to RAFT polymerization, ATRP was also used as a convenient tool for the synthesis of highly branched biocompatible poly(2-hydroxyethyl methacrylate), which was then used to prepare biocompatible fibers. The incorporation of partial disulfide-based dimethacrylate monomer in the polymerization conferred the biodegradability to the highly branched polymers \\[[@B76-polymers-10-00758]\\]. Hedrick and coworkers described a new functional lactone containing a pendant acrylate group that can be of great interest for the design of new cross-linked biodegradable materials using combined ATRP and ROP techniques \\[[@B77-polymers-10-00758]\\]. Similarly, Xu et al. reported the generation of comb-shaped copolymers composed of biocompatible hydroxypropyl cellulose backbones and cationic poly(2-dimethyl amino)ethyl methacrylate) side chains for gene delivery. The generated complex exhibited a stronger ability to bind with DNA, due to the increased surface cationic charges \\[[@B78-polymers-10-00758]\\]. Comb-like and biodegradable supramolecular architectures can also be prepared using amphiphilic copolymer, poly(lactide)-*b*-poly(2-hydroxyethyl methacrylate) (PLA-*b*-PHEMA) with partially biodegradable PLA block and PHEMA biocompatible one using an orthogonal polymerization strategy via ROP and ATRP \\[[@B79-polymers-10-00758]\\]. Likewise, the same methodology was also adopted to prepare ABA and star amphiphilic block copolymers composed of polymethacrylate bearing a galactose fragment and biodegradable poly(epsilon-caprolactone) \\[[@B80-polymers-10-00758]\\]. Series of degradable branched poly(dimethylaminoethyl methacrylate) (PDMAEMA) copolymers were investigated by Zhao's group. The branched PDMAEMA copolymers were synthesized by controlled radical cross-linking copolymerization. Efficient degradation processes were experimented for all of the copolymers. The degree of branching exhibited a big impact on the performance of transfection when tested on different cell types. The product with the highest degree of branching and highest degree of functionality had a superior transfection profile in terms of both transfection capability and the preservation of cell viability. The branched PDMAEMA copolymers show high potential for gene-delivery applications through a combination of the simplicity of their synthesis, their low toxicity and their high performance ([Figure 6](#polymers-10-00758-f006){ref-type=\"fig\"}) \\[[@B81-polymers-10-00758]\\].\n\n2.3. Hydrogels and Nanogels {#sec2dot3-polymers-10-00758}\n---------------------------\n\n### 2.3.1. Hydrogels {#sec2dot3dot1-polymers-10-00758}\n\nHydrogels are optimal materials for tissue engineering scaffolds due to their tissue-like mechanical compliance and mass transfer properties. However, many hydrogels that have been widely used in medical science are not biodegradable, thus cannot be easily and quickly cleared out of the body. Therefore, using biocompatible and biodegradable co-polymers for fabricating hydrogels is much desired. Ratner and coworkers successfully prepared cross-linked nanogels of biodegradable poly(2-hydroxyethyl methacrylate) (PHEMA) as engineered tissue constructs using ATRP technique and an enzyme degradable cross-linking agent, polycaprolactone (PCL) and a degradable macro-initiator that also contained oligomeric PCL \\[[@B37-polymers-10-00758]\\].\n\nThe hydrogel of nanostructured hyaluronic acid has also be generated in situ by Matyjaszewski's group under physiological conditions (pH 7.4, 37 \u00b0C) by a combination of ATRP and Michael-type addition reaction using biodegradable nanogel precursors, 2-hydroxyethyl p(OEO300MA-*co*-methacrylate) (POEO300MA-*co*-PHEMA) \\[[@B82-polymers-10-00758]\\]. RAFT agent in the form of \"RAFT gel\" has also been prepared by Takasu's group via chemoselective polycondensations of a dicarboxylic acid containing a mercapto group and further used for the polymerization of methyl methacrylate to afford polyester containing biodegradable hydrogels \\[[@B83-polymers-10-00758]\\]. It is well known that the synthetic poly(amino acid)s that have polypeptide backbone can be degraded in biological environments by enzymes such as proteinases and peptidases. Kubies et al. successfully prepared such cross-linked biodegradable hydrogels of a series of polymer architectures with the same polypeptide backbone via ring opening polymerization. They also found the enzyme-catalyzed hydrolysis can be controlled through copolymerization and/or side-chain modifications \\[[@B38-polymers-10-00758]\\]. A combination of anionic and RAFT polymerization was used to synthesize an triblock polymer poly-\\[(propylenesulfide)-*b*-(*N*,*N*-dimethylacrylamide)-*b*-(*N*-isopropylacrylamide)\\] (PPS-*b*-PDMA-*b*-PNIPAAM) that forms physically cross-linked hydrogels when transitioned from mechanisms for reactive oxygen species (ROS) triggered degradation and drug release. At ambient temperature, PPS-*b*-PDMA-*b*-PNIPAAM assembled into 66 \u00b1 32 nm micelles comprising a hydrophobic PPS core and PNIPAAM on the outer corona. The PPS-*b*-PDMA-*b*-PNIPAAM micelles were preloaded with the model drug Nile red and the resulting hydrogels demonstrated ROS-dependent drug release. The hydrogels were cyto-compatible in vitro and demonstrated to have utility for cell encapsulation and delivery. These hydrogels also possessed inherent cell-protective properties and reduced ROS-mediated cellular death in vitro. Subcutaneously injected PPS-*b*-PDMA-*b*-PNIPAAM polymer solutions formed stable hydrogels that sustained local release of the model drug Nile red for 14 days in vivo. These collective data demonstrate the potential use of PPS-*b*-PDMA-*b*-PNIPAAM as an injectable, cyto-protective hydrogel that overcomes conventional PNIPAAM hydrogel limitations such as syneresis, lack of degradability, lack of inherent drug loading and environmentally responsive release mechanisms ([Figure 7](#polymers-10-00758-f007){ref-type=\"fig\"}) \\[[@B84-polymers-10-00758]\\].\n\n### 2.3.2. Nanogels {#sec2dot3dot2-polymers-10-00758}\n\nNanogels have drawn enormous attention due to their applications as targeted drug delivery scaffolds in biomedical science. Matyjaszewski's group is pioneering the fabrication of hyperbranched polymeric architectures, particles, hydrogels and nanogels using ATRP strategies \\[[@B85-polymers-10-00758]\\]. They reported the synthesis of stable biodegradable nanogels cross-linked with disulfide linkages using inverse miniemulsion ATRP methods. The biodegradation in the presence of glutathione tripeptide can trigger the release of encapsulated molecules including rhodamine 6 G, a fluorescent dye and doxorubicin (Dox), an anticancer drug, as well as facilitate the removal of empty vehicles \\[[@B86-polymers-10-00758]\\]. They also prepared biodegradable nanogels as delivery carriers for carbohydrate drugs using ATRP in a cyclohexane inverse miniemulsion in the presence of a disulfide functionalized dimethacrylate cross-linker. These nanogels exhibited the high loading efficiency of rhodamine B isothiocyanate-dextran (RITC-Dx) exceeding 80% \\[[@B87-polymers-10-00758]\\]. The same inverse miniemulsion ATRP strategy was also utilized to make biodegradable nanogels. Likewise, nanogels that can be degraded under various pH conditions were also prepared from biodegradable amphiphilic polymers synthesized by ATRP combined with ROP synthetic methodologies \\[[@B88-polymers-10-00758]\\].\n\nRecent advances in drug carrier design in the field of photodynamic therapy (PDT) have stimulated the development of numerous sophisticated drug delivery carriers. Kim and coworkers designed a novel biodegradable and biocompatible nanogels used as PDT carriers. The nanogels were synthesized through ATRP method using inverse miniemulsion and their biodegradability was determined in the presence of glutathione. The model photosensitizer (PS) was encapsulated in the biodegradable nanogels by simple mixing and sonication. The cellular uptake and the cytotoxicity of the nanogels before and after laser irradiation were determined. The results showed that the Ce6-loaded nanogels did not influence the cellular viability of the cells before light irradiation. Under light exposure, the Ce6-nanogel complex revealed strong photoactivity. These nanogels may enhance therapeutic efficacy of PSs without any complex chemical modifications with PSs ([Figure 8](#polymers-10-00758-f008){ref-type=\"fig\"}) \\[[@B89-polymers-10-00758]\\].\n\nThe stability of encapsulation in self-assembled system is usually limited by the requisite concentration for self-assembly formation. Once the encapsulation is achieved, the lack of targeting molecules on the drug carriers will compromise the efficiency for targeted delivery. To tackle this issue, Thayumanavan and coworkers successfully fabricated surface-functionalized polymer nanogels with facile hydrophobic guest encapsulation capabilities \\[[@B90-polymers-10-00758]\\]. These biodegradable nanogels were first prepared from pyridyl disulfide pedant random copolymers that were prepared through RAFT mechanism via cross-linking through disulfide bonding, followed by the surface modification with a thiol-modified cell-penetrating peptide, Tat-SH. The internalization of Tat-SH modified nanogels occurred much more readily than that observed with the control gels, confirming the effectiveness of the modification of the nanogel surface. This presents a clear method for incorporating ligands onto the polymer nanoparticles and thus achieves specificity to pathogenic cells. Biodegradable nanogels/microgels have also been successfully prepared by RAFT polymerization using cross-linking agents that contain acid sensitive or disulfide intra-linkages. The surface tethered RAFT active centers allow further modifications and functionalizations via thiol-pyridyl disulfide exchange or thiol-ene reactions \\[[@B91-polymers-10-00758]\\].\n\n2.4. Micelles, Vesicles and Capsules {#sec2dot4-polymers-10-00758}\n------------------------------------\n\nPolymers have been widely explored for the preparation of varied particles, e.g., micelles, vesicles and capsules, based on the expectation that these particles can be the appropriate reservoirs for controlled drug delivery. The advantage of using these polymer particles as drug carriers over traditional administration of free drugs lies in the increased circulation time in the body as these particles are usually big enough to prevent fast clearance through kidney filtration which has a cut-off molecular weight of 50,000 g\u00b7mol^\u22121^ \\[[@B27-polymers-10-00758]\\]. Another advantage of polymer particles is called \"stealth-like\" effect which can be observed with the particles smaller than 200 nm or those surface decorated with specific polymers, e.g., poly(ethylene glycol) \\[[@B92-polymers-10-00758]\\]. This \"stealth-like\" effect will greatly increase the circulation time. Polymer particles are usually prepared by amphiphilic block copolymers, where the hydrophobic block is used to form the core and the hydrophilic block from the corona in polar media. Research has revealed that the morphology of the polymeric particles might be mainly determined by the ratio of hydrophilic segment to the hydrophobic one \\[[@B93-polymers-10-00758]\\]. Using block copolymers to prepare micelles has been extensively conducted and well-reviewed \\[[@B8-polymers-10-00758]\\]. The design of polymeric particles with hydrophobic cores is based on the fact that anti-cancer drugs are usually hydrophobic and can be impregnated within the particle cores. Of course, the polymer particles can also be tailored with hydrophilic core and hydrophobic corona when required, mostly by manipulation of the polarity of preparation solvent. Once the drug is impregnated within particles, another issue arises with how to control the drug release. The traditional drug release from the particles is usually controlled by the self-degradation. However, if the polymers are designed as biodegradable, better control or more controlling means can be realized. The preparation of polymeric micelles for drug delivery using RAFT polymerization was reviewed by Stenzel \\[[@B7-polymers-10-00758]\\]. In this section, we mainly discuss the preparation of polymer particles that can undergo biodegradation and their potential applications.\n\n### 2.4.1. Micelles {#sec2dot4dot1-polymers-10-00758}\n\nMicelles generated from well-defined diblock copolymers of thermoresponsive poly(*N*-isopropylacrylamide-*co*-*N*,*N*-dimethylacrylamide) blocks and biodegradable poly([d]{.smallcaps},[l]{.smallcaps}-lactide) blocks by the combination of RAFT polymerization and ROP were also prepared by Akimoto et al. The biodegradable polylactide (PLA) cores conferred the degradability to the micelles at acidic condition (pH 5.0). A much similar work was carried out by Zhu et al. who fabricated micelles using a thermal responsive poly(*N*-isopropylacrylamide) block, thus drug release could be thermally controlled. The presence of polycaprolactone (PCL) block makes the micelle biodegradable in biological environments \\[[@B94-polymers-10-00758]\\].\n\nIn Ning's study, well-defined, novel, linear, biodegradable and amphiphilic thermo-responsive ABA-type triblock copolymers, poly(2-(2-methoxyethoxy) ethyl methacrylate-*co*-oligo(ethylene glycol) methacrylate)-*b*-poly(*\u03b5*-caprolactone)-*b*-poly(2-(2-methoxyethoxy) ethyl methacrylate-*co*-oligo (ethylene glycol) methacrylate) (P(MEO~2~MA-*co*-OEGMA)-*b*-PCL-*b*-P(MEO~2~MA-*co*-OEGMA)) (tBPs), were synthesized via a combination of ring-opening polymerization (ROP) of *\u03b5*-caprolactone (*\u03b5*CL) and RAFT polymerization of MEO~2~MA and OEGMA monomers. Thermo-responsive micelles were obtained through a self-assembly process of copolymers in aqueous medium. The hydrophobic drug of anethole was encapsulated in micelles through the dialysis method. The average particle sizes of drug-loaded micelles were determined by dynamic light scattering measurement. In vitro, the sustained release of the anethole was performed in pH 7.4 phosphate buffered saline at different temperatures. Results showed that the triblock copolymer micelles were quite effective in the encapsulation and controlled release of anethole. The vial inversion test demonstrated that the triblock copolymers could trigger the sol-gel transition which also depended on the temperature, and its sol-gel transition temperature gradually decreased with the concentration increasing ([Figure 9](#polymers-10-00758-f009){ref-type=\"fig\"}a,b) \\[[@B95-polymers-10-00758]\\].\n\nAnother issue arising with the particle delivered drug delivery is how to enhance the delivering efficiency. The unmodified particles are usually evenly distributed in the body, if this is the case side effect might happen. Therefore, achieving targeted drug delivery has attracted enormous interest. Davis and coworkers prepared surface functionalized micelles using amphiphilic triblock copolymers of oligo(ethyleneglycol) acrylate (PEG-A) and styrene (St), poly(PEG-A)-*b*-poly(St)-*b*-poly(PEG-A) by RAFT polymerization using a new bifunctional RAFT agent, *S*,*S*-bis\\[\u03b1, \u03b1\u2032-dimethyl-\u03b1\u2033-(2-pyridyl disulfide) ethyl acetate\\] trithiocarbonate (BDPET) \\[[@B96-polymers-10-00758]\\]. These micelles were tailored with surface bound pyridyldisulfide (PDS) groups that are active to a free thiol group bearing model peptide, reduced glutathione, and a thiol modified fluorophore, rhodamine B, under mild reaction conditions ([Figure 9](#polymers-10-00758-f009){ref-type=\"fig\"}c). It can be envisioned that, when these micelles are tailored with specific targeting molecules, the delivery efficiency should be greatly enhanced and the unwanted side effect can then be avoided.\n\nIt is usually difficult to obtain the complicated polymer architectures using a single polymerization technique. Combining with organo-base catalyzed polymerization of [l]{.smallcaps}- or [d]{.smallcaps}-lactide Frey and coworkers, using ATRP technique, prepared biodegradable poly(isoglycerol methacrylate)-*b*-poly([l]{.smallcaps}- or [d]{.smallcaps}-lactide) copolymer as building block for fabrication of spherical and large superamolecular vesicles via self-assembly in aqueous medium \\[[@B97-polymers-10-00758]\\]. In most cases, the micelle cores are employed as drug reservoirs, however some novel micelles based on biodegradable poly ([l]{.smallcaps}-glutamic acid)-*b*-polylactide with paramagnetic Gd ions chelated to the shell layer were also prepared as a potential nanoscale magnetic resonance imaging (MRI)-visible delivery system \\[[@B98-polymers-10-00758]\\].\n\nIn addition to RAFT polymerization, ATRP incorporating with ROP were also employed to synthesize a new class of supramolecular and biomimetic glycopolymer/poly(*\u03b5*-caprolactone)-based polypseudorotaxane/glycopolymer triblock-copolymers. The polypseudorotaxane block was prepared by an inclusion reaction between biodegradable poly(*\u03b5*-caprolactone) and \u03b1-cyclodextrin. These triblock biohybrids were then utilized to fabricate micelles or vesicles that possess hydrophilic glycopolymer shell and oligosaccharide threaded polypseudorotaxane core \\[[@B99-polymers-10-00758]\\]. Likewise, quite similar biodegradable amphiphilic block copolymers with poly(\u03b3-methyl-*\u03b5*-caprolactone) (PmCL), o-nitrobenzyl (ONB) and polyacylic acid (PAA) blocks, and the same synthetic methodologies have also been prepared by Cabane et al. for fabrication of micelles as well. Furthermore, the as-fabricated micelles and vesicles are also photoresponsive due to the presence of a photodegradable OCN linker as a junction point between hydrophilic and hydrophobic chains \\[[@B100-polymers-10-00758]\\].\n\nAcetal is a pH sensitive group that is stable at pH 7 and prone to go hydrolysis at mild acidic pH of 4.0--5.0, with a half-life of 6.5 h, respectively. Zhong and coworkers \\[[@B101-polymers-10-00758]\\] incorporated acetal groups into block copolymers comprising of a novel acid-labile polycarbonate and poly(ethylene glycol) (PEG) to generate pH-responsive biodegradable micelles as potential smart nano-vehicles for targeted delivery of anticancer drugs. Biodegradable cross-linked micelles were also prepared with a stimulus-responsive triblock copolymer synthesized via a bifunctional ATRP initiator containing intra-disulfide linkage \\[[@B102-polymers-10-00758]\\]. When the micelles were prepared by stimulus-responsive copolymer and self-assembled on mica surface, pH manipulated switchable surface was achieved \\[[@B103-polymers-10-00758]\\]. NMRP in combination with ROP were also utilized to prepare poly(*\u03b5*-caprolactone-*b*-4-vinylpyridine) for preparation of micelles. As the so-prepared micelle impregnated a cationic core it can mediate the transportation of AuCl~4~^\u2212^ anions from aqueous phase to the micelle core to afford micelle protected Au nanoparticles after reduction with NaBH~4~ \\[[@B104-polymers-10-00758]\\].\n\n### 2.4.2. Vesicles {#sec2dot4dot2-polymers-10-00758}\n\nPolymeric vesicles or polymersomes are nano- or micrometer sized polymeric capsules with a bilayered membrane. Extensive applications can be envisioned in nanomedicine, in vivo diagnostics and drug delivery \\[[@B6-polymers-10-00758]\\]. Du and Armes reported the facile preparation of vesicles in pure water medium using diblock copolymer, poly(*\u03b5*-caprolactone)-*b*-poly\\[2-(methacryloyloxy) ethyl phosphorylcholine\\] (PCL-*b*-PMPC), which was synthesized using the combined methods of ROP, end-group modification and ATRP. These vesicles can be stabilized by sol-gel chemistry within the vesicle membrane \\[[@B105-polymers-10-00758]\\]. In addition to the well-defined routine chemical polymerization methods, lipase-catalyzed condensation polymerization method was also used to synthesize biodegradable poly(10-hydroxydecanoic acid) (PHDA) and further modify it with ATRP initiator for grafting another hydrophobic polystyrene block for fabrication of polymeric nanoparticles in aqueous medium \\[[@B106-polymers-10-00758]\\]. Vesicles can also be designed as pH sensitive for efficient DNA encapsulation and delivery, where the particles were prepared by poly(2-(methacryloyloxy)ethyl-phosphorylcholine)-*co*-poly(2-(diisopropylamino)ethyl methacrylate) (PMPC-*b*-PDPA) diblock copolymers. The PMPC block is highly biocompatible and nonfouling, while the PDPA block is pH-sensitive (pKa \\~5.8--6.6, depending on the ionic strength) \\[[@B107-polymers-10-00758]\\]. Wang and coworkers reported a novel method for the preparation of biodegradable large compound vesicles with controlled size and narrow size distribution by using aqueous nanodroplets as templates. PEG-based large compound vesicles (LCVs) were prepared through a self-assembly process of the temperature-responsive 2-(2-methoxyethoxy) ethyl methacrylate-oligo(ethylene glycol) methacrylate-*N*,*N*\u2032-cystamine bisacrylamide (MEO~2~MA-OEGMA-CBA) branched copolymer. The sizes of the LCVs can be easily tuned by the amount of surfactants and the cross-linked reaction in LCVs occurred during the fusion process of small vesicles without any additional cross-linking agent. The formed LCVs are uniform, low toxic and resistant to nonspecific protein adsorption. The biodegradable and biocompatible LCVs can act as a vector for proteins ([Figure 10](#polymers-10-00758-f010){ref-type=\"fig\"}) \\[[@B108-polymers-10-00758]\\].\n\n### 2.4.3. Capsules {#sec2dot4dot3-polymers-10-00758}\n\nMultilayered polymer capsules assembled via layer-by-layer (LbL) technology have generated significant scientific and technological interest over the past decade because of their potential as advanced delivery and microreactor systems \\[[@B109-polymers-10-00758],[@B110-polymers-10-00758]\\]. Caruso and coworkers are pioneering the preparation of versatile capsules via self-assembly for drug and gene delivery and controlled release, among which some of them are biodegradable \\[[@B111-polymers-10-00758],[@B112-polymers-10-00758],[@B113-polymers-10-00758],[@B114-polymers-10-00758]\\]. For example, they fabricated low-fouling poly(*N*-vinyl pyrrolidone) (PVPON) capsules with engineered biodegradable properties via LbL process mediated by hydrogen bonding interaction. Due to the introduction of intra-disulfide linkages among the capsules they underwent destruction within 4 h in the presence of 5 mM glutathione. The cross-linked multilayers endowed the capsule with low-fouling properties to a range of proteins, including fibrinogen, lysozyme, immunoglobulin G, and bovine serum albumin \\[[@B115-polymers-10-00758]\\]. Disulfide-stabilized poly(methacrylic acid) capsules that undergo reversible swelling in response to changes of external pH and degrade in the presence of a physiological concentration of glutathione were also prepared and investigated.\n\nIn Cui's study, the preparation of pH responsive, biodegradable, biocompatible and cross-linked polymer capsules for controlled drug release was presented. The capsules were prepared using silica particles as templates for surface grafting of poly (acrylic acid) (PAA) and PAA-*co*-poly(polyethylene glycol) acrylate) (PAA-*co*-PPEGA) block copolymer via RAFT polymerization directly from silica particles, followed by cross-linking with cystamine dihydrochloride and removal of the silica template in the presence of hydrofluoric acid, respectively. The resultant polymer capsules were water soluble and biocompatible with a mean diameter of approximately 260 \u00b1 10 nm. These polymer capsules were non-toxic to human cells at a low concentration, which are favorable to be utilized as drug carriers for pH responsive and biodegradation controlled drug release. Doxorubicin hydrochloride (DOX) was used as a model drug to test the drug loading and releasing properties of the polymer capsules. It was found that the DOX could be effectively loaded into the PAA and PAA-*co*-PPEGA capsules with a loading capacity up to 52.24% and 36.74%, respectively. The pH and biodegradation controlled release behaviors of DOX loaded PAA-PPEGA capsules were also explored. The results implied that both PAA and PAA-*co*-PPEGA capsules are promising platforms for pH and biodegradation controlled drug delivery systems, while the PAA-*co*-PPEGA capsules exhibit less cytotoxicity ([Figure 11](#polymers-10-00758-f011){ref-type=\"fig\"}) \\[[@B116-polymers-10-00758]\\].\n\n2.5. Polymeric Architectures Based on Biodegradable Synthetic or Natural Precursors {#sec2dot5-polymers-10-00758}\n-----------------------------------------------------------------------------------\n\nBiodegradable architectures can also be achieved by grafting polymer chains onto the biodegradable precursors. The degradation of biodegradable precursors will consequently disintegrate the as-prepared architectures. These precursors can be synthetic biocompatible and biodegradable films, such as poly(3-hydroxybutyrate-*co*-3-hydroxyhexanoate) (P(HB-*co*-HHx)) ([Figure 12](#polymers-10-00758-f012){ref-type=\"fig\"}a) \\[[@B117-polymers-10-00758]\\]. Cellulose is a natural polysaccharide consisting of a linear chain of several hundreds to over ten thousand linked [d]{.smallcaps}-glucose units. It is the major constituent of paper, paperboard, and card stock and of textiles made from cotton, linen, and other plant fibers. Its high hydrophilicity is right due to the multi-hydroxy groups from the glucose units. These hydroxy groups not only make the cellulose chains holding firmly together side-by-side and forming microfibrils with high tensile strength, they can also be used to make soluble and functionalized cellulose. The biodegradable cellulose would be a good precursor for generation of biodegradable architectures. A few groups have explored the possibility of modifying cellulose. Carlmark et al. successfully modified the cellulose using ATRP via \"graft from\" methodology. They first attached 2-bromoisobutyryl bromide on the cellulose surface through the condensation reaction with the surface hydroxyl group, followed by the ATRP reaction to create an amphiphilic block copolymer layer on it \\[[@B118-polymers-10-00758]\\].\n\nUsing the same methodology, Perrier and coworkers successfully attached different RAFT agents through the surface hydroxyl groups for direct grafting an amphiphilic copolymer, poly(ethylene glycol)-*b*-poly(*[l]{.smallcaps}*-lactic acid), from the cellulose surface. The biodegradable poly([l]{.smallcaps}-lactic acid) block further facilitates the biodegradability of the so-prepared architecture \\[[@B119-polymers-10-00758]\\]. One advantage of RAFT polymerization is the versatile initiation methods. In addition to the commonly used thermal initiation, other ionizing radiation sources, such as \u03b3-ray, ultraviolet, microwave and X-ray radiation, have also been used to initiate RAFT controlled polymerizations \\[[@B20-polymers-10-00758],[@B120-polymers-10-00758],[@B121-polymers-10-00758],[@B122-polymers-10-00758]\\]. Barsbay and coworkers used both thermal and \u03b3-ray initiations and RAFT polymerization to modify cellulose with styrene and sodium 4-styrenesulfonate polymeric brushes using \"graft from\" methodology \\[[@B123-polymers-10-00758],[@B124-polymers-10-00758]\\]. Cellulose fiber was also modified with biodegradable polyesters by the aid of host-guest inclusion complexation between \u03b2-cyclodextrin and adamantine motifs \\[[@B125-polymers-10-00758]\\].\n\n\"Glycopolymers\", particularly the multivalent ones have attracted tremendous attention due to the potential applications in biomedicine and biomaterials. Dong and coworkers synthesized a four-armed star glycopolymer composed of block copolymer arms bearing lactone end groups. These star polymers could self-assemble onto nanoparticles that carry the lactose groups on their surface, allowing for the further complexing with lectins to achieve biodegradable biohybrids \\[[@B126-polymers-10-00758]\\]. Glycopolymers were synthesized by Stenzel and coworkers using RAFT polymerization and thio-ene click chemistry to fabricate glucose surface tethered glycomicelles for further complexation with concanavalin A, a mannose and glucose specific lectin. These biodegradable and biocompatible glycomicelles could be utilized as potential drug carriers \\[[@B127-polymers-10-00758]\\]. Qiu et al. also prepared large spherical micelles in aqueous solution, using star-shaped polypeptide/glycopolymer biohybrids composed of poly(\u03b3-benzyl [l]{.smallcaps}-glutamate) and poly([d]{.smallcaps}-gluconamidoethyl methacrylate) prepared via ROP and ATRP. The generated micelles had a helical polypeptide core surrounded by a multivalent glycopolymer shell, which potentially provides a platform for fabricating targeted anticancer drug delivery system and for studying the glycoprotein functions in vitro \\[[@B128-polymers-10-00758]\\].\n\nIn contrast with the polymer--drug conjugates prepared thus far, in which the drug is typically attached via an enzymatically or hydrolytically cleavable linker, Apostolovic's group reported the noncovalent polymer therapeutics based on a conceptually novel class of polymers prepared using RAFT mechanism. The polymer backbone was used to attach the cargo via a noncovalent, biologically inspired coiled coil linker, which was formed by heterodimerization of two complementary peptide sequences that are linked to the polymer carrier and the cargo, respectively \\[[@B129-polymers-10-00758]\\].\n\n2.6. Biodegradable Biomolecule-Polymer Conjugates {#sec2dot6-polymers-10-00758}\n-------------------------------------------------\n\nBioconjugates refer to a category of polymer conjugates with widespread biomolecules, which have attracted increasing interest as they have numerous potential applications in biotherapeutics, bioseparation and functional materials field. The importance of bioconjugates lies in the fact that the biomolecules will exhibit prolonged circulation time in biofluids \\[[@B130-polymers-10-00758],[@B131-polymers-10-00758]\\] and their immunogenicity and antigenicity can also be reduced by the incorporation of biocompatible polymer fragments \\[[@B2-polymers-10-00758],[@B132-polymers-10-00758],[@B133-polymers-10-00758]\\]. When the bioconjugates are designed with biodegradable linker between the biomolecules and the polymer fragments these biomolecules can be released in vivo, therefore, their bioactivities can be reversed \\[[@B20-polymers-10-00758],[@B134-polymers-10-00758]\\]. On the other hand, most biomolecules, e.g., proteins and enzymes, consist of peptides that are linked by biodegradable disulfide bonding. In this case, these biomolecules are also biodegradable, making the whole bioconjugates biodegradable.\n\nDavis and coworkers delivered elegant research on the preparation of biodegradable conjugates. Free thiol tethered biomolecule, e.g., bovine serum albumin (BSA), has been successfully modified with several polymers to afford biodegradable homo- or hetero-bioconjugates under ambient condition using room temperature initiation via RAFT polymerization ([Figure 12](#polymers-10-00758-f012){ref-type=\"fig\"}d) \\[[@B20-polymers-10-00758],[@B135-polymers-10-00758],[@B136-polymers-10-00758],[@B137-polymers-10-00758],[@B138-polymers-10-00758]\\]. By tailoring the bioconjugates with disulfide linkage between lysozyme and the polymer chains, the bioactivity of lysozyme can be reversed during the biodegradation process \\[[@B134-polymers-10-00758]\\]. They have also successfully modified lysozyme with well-defined poly-*N*-(2-hydroxypropyl) methacrylamide via surface modifications through amide bonding to tailor the enzyme's bioactivity \\[[@B139-polymers-10-00758]\\]. A latest study reported the modification of fragile glucose oxidase (GOx) with biocompatible polymer, poly(ethyleneglycol) acrylate (polyPEG-A) and thermoresponsive copolymer of poly(ethyleneglycol) acrylate and di(ethyleneglycol) ethyle ether acrylate \\[poly(PEG-A-*co*-DEG-A)\\] to afford biodegradable enzyme--polymer conjugates. Bio-cleavage of the polymer chains from the GOx surface obviously recovered the enzymatic activity \\[[@B62-polymers-10-00758]\\]. These smart enzyme--polymer conjugates would envision promising applications in biotechnology and biomedicine. Maynard and coworkers also achieved significant advances in the preparation of bioconjugates using either ATRP or RAFT polymerization. They successfully modified si-RNA with a biodegradable polymer fragment. Since *si*-RNA is considered an effective targeting molecule, its biodegradable polymer conjugates could be good candidates for potential bio-therapeutics \\[[@B140-polymers-10-00758]\\]. In addition to RAFT polymerization, ATRP has also been successfully applied to prepare biodegradable polymer conjugates with BSA \\[[@B141-polymers-10-00758],[@B142-polymers-10-00758]\\] and engineered lysozyme \\[[@B143-polymers-10-00758]\\].\n\n3. Conclusions and Perspectives {#sec3-polymers-10-00758}\n===============================\n\nThis review has discussed the synthesis and applications of biodegradable polymeric architectures using different RDRPs. These biodegradable polymeric structures can be designed as well-defined star-shaped, cross-linked or hyperbranched, through which more complicated nanoparticles such as micelles, vesicles and capsules can be fabricated via either self-assembly or cross-linking methodologies. Nanogels and hydrogels can also be prepared via RDRPs. Their applications in biomedical science are also discussed. Biodegradable polymeric architectures can be prepared with both synthetic and natural precursors.\n\nAs discussed in this review, RDRPs have proven to be convenient tools for the synthesis of the versatile biodegradable polymeric architectures to meet varied applications. Driven by the practical application and commercialization, the design of more complicated polymeric architectures with controllable biodegradability will be expected. However, it is worth noting that a fast biodegradable process in vivo is not desired in some situations. Therefore, designing and fabricating the polymeric architectures with controllable and slow biodegradability would be a critical issue in this field. To achieve this, many other different polymerization techniques are required besides RDRPs.\n\nY.X., J.L. and L.C. conceived and designed the structure of this article; F.C. performed the literature search; A.Z. and F.Q. wrote the paper and organized the figures; and A.Z. and F.Q. contributed equally to this work.\n\nThis work was funded by Qingdao Innovation Leading Talent Program; Natural Science Foundation of China (51173087) and Qingdao (12-1-4-2-2-jch); Taishan Scholars Program and Shandong Provincial Natural Science Foundation, China (ZR2018BEM020).\n\nThe authors declare no conflict of interest.\n\nFigures and Scheme\n==================\n\n![Schematic illustration for biodegradable polymeric architectures via reversible deactivation radical polymerizations.](polymers-10-00758-sch001){#polymers-10-00758-sch001}\n\n![(**a**) Generally accepted mechanism for a RAFT polymerization. Copyright 2009, American Chemical Society; (**b**) The mechanism for NMRP polymerization. Copyright 2001, American Chemical Society; (**c**) Transition-metal-catalyzed ATRP. Copyright 2001, American Chemical Society.](polymers-10-00758-g001){#polymers-10-00758-f001}\n\n![(**a**) Multi-armed RAFT agents with biodegradable disulfide intra-linkers: three-armed RAFT agent with *R*-group at the end of each arm; three-armed RAFT agent with *Z*-group at the end of each arm and six-armed RAFT agent with *R*-group at the end of each arm; (**b**) Post polymerization modification of the three-armed star polymers with RAFT active center at the end of each arm. Copyright 2009, American Chemical Society.](polymers-10-00758-g002){#polymers-10-00758-f002}\n\n![Synthesis of the star-block amphiphilic copolymer via ring-opening polymerization (**a**) and RAFT polymerization (**b**). Copyright 2010, Elsevier.](polymers-10-00758-g003){#polymers-10-00758-f003}\n\n![(**a**) Synthesis of hexafunctional star polymers and block copolymers based on PNVCL using amultifunctional xanthate as a RAFT agent; (**b**) Synthesis of star (PNVCL-*b*-PVAc)~6~ copolymers; (**c**) Self-assembly in aqueous solution of star (PNVCL-*b*-PVAc)~6~ block copolymers; (**d**) Synthesis of the star \\[PVAc-*b*-(PNVCL-*co*-PNVP)\\]~6~ block copolymers. The blue and red colors represent PNVCL and PVAc, respectively. Copyright 2017, Elsevier.](polymers-10-00758-g004){#polymers-10-00758-f004}\n\n![Scheme of degradable core--shell polymer via RAFT and ROP route. Routes (**A**,**B**) demonstrate the core-first method (for hyperbranched particles) and arm-first method (for CCS particles), respectively. Copyright 2011, American Chemical Society.](polymers-10-00758-g005){#polymers-10-00758-f005}\n\n![(**a**) (Left) Controlled radical cross-linking copolymerization through in situ DE-ATRP, followed by a post-functionalization process; and (Right) graphical representation of structures with different degrees of branching. The efficacy of functionalization depends on the content of the pendent vinyl groups; (**b**) Graphical representation of the degradation of structures with different degrees of branching. Copyright 2014, Wiley-VCH.](polymers-10-00758-g006){#polymers-10-00758-f006}\n\n![(**a**) Schematic representation of micelle gelation at 37 \u00b0C and polymer architecture coordinating with STEM-EDS element maps; (**b**) TEM images of PPS60-*b*-PDMA150-*b*-PNIPAAM150 micelles at 25 and 37 \u00b0C; (**c**) STEM-EDS element maps for sulfur (red) and oxygen (green) of PPS60-*b*-PDMA150-*b*-PNIPAAM150 core--shell compartments at 37 \u00b0C with image thresholding and background subtraction. Core-forming PPS produces the red signal for sulfur, while oxygen (appearing green) is present in the PDMA and PNIPAAM corona-forming blocks. Copyright 2014, American Chemical Society.](polymers-10-00758-g007){#polymers-10-00758-f007}\n\n![(**a**) Scheme for biodegradable nanogels synthesized by inverse miniemulsion ATRP; (**b**) Fluorescent images of cells with Ce6-loaded nanogels as a function of incubation time (scale bar 300 \u00b5m). Copyright 2016, Springer.](polymers-10-00758-g008){#polymers-10-00758-f008}\n\n![(**a**) Synthesis of triblock copolymer P(MEO~2~MA-*co*-OEGMA)-*b*-PCL-*b*-P(MEO~2~MA-*co*-OEGMA); (**b**) Schematic representation of the self-assembled thermo-sensitive core--shell micelles for temperature-stimulated drug release and gelation of tPBs. Copyright 2018, Elsevier; (**c**) Functionalized micelles and further attachment of fluorophores and targeting tripeptides.](polymers-10-00758-g009){#polymers-10-00758-f009}\n\n![(**a**) Schematic outline of preparation of inverse emulsion nanodroplet templates containing the branched copolymer in the water phase and the possible mechanism for the formation of cross-linked large compound vesicles in nanodroplet templates; (**b**) TEM images of the self-assembled nanostructure of the branched copolymer in nanodroplet templates at different stages. Copyright 2014, Royal Society of Chemistry.](polymers-10-00758-g010){#polymers-10-00758-f010}\n\n![The preparation of pH responsive, biodegradable, biocompatible and cross-linked polymer capsules for controlled drug DOX (the red dots) release. Copyright 2014, Elsevier.](polymers-10-00758-g011){#polymers-10-00758-f011}\n\n![(**a**) Synthesis of the gradient copolymers; (**b**) Synthesis of the block copolymers; (**c**) Synthesis of thermo-responsive surfaces. Copyright 2010, Wiley-VCH; (**d**) Synthesis of branched poly (*N*-(2-hydroxypropyl) methacrylamide) (PHPMA) and the subsequent conjugation with protein. Copyright 2009, American Chemical Society.](polymers-10-00758-g012){#polymers-10-00758-f012}\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Background & Summary\n====================\n\nComputed tomography (CT) scanning has opened new ways for studying various parts of an organism's biology. This technique has been used in sharks to elucidate the development^[@b1]^, function^[@b2]^ and morphology of their feeding mechanics^[@b6]^. The majority of these studies have focused on small taxonomic subsets to gain detailed anatomical knowledge. However, phylogenetically broad statistical analyses of shark cranial mechanics are still lacking. The lead author of this work is currently undertaking such studies. This contribution provides a descriptor of a large CT dataset of shark anatomy. We used whole specimens from museum collections (a few specimens only comprise the head due to their conservation) to create a dataset of medical CT scans which covers approximately 75% of all extant shark families^[@b9]^.\n\nThe x-ray computed tomography library presented here was created for investigating the ecomorphological diversity of shark feeding systems. The data from the CT scans were used to create a second dataset comprising three-dimensional (3D) models of the lower jaw. These models were used to examine lower jaw disparity in extant sharks species in a separate study by quantifying jaw shape using landmark based geometric morphometrics. The lower jaw was selected because it displays a diversity of jaw morphologies between shark species^[@b10]^. Lower jaws present biomechanically important features related to jaw closing mechanics and are therefore expected to function as a predictor of ecological specialisation^[@b16]^.\n\nThis tomography and virtual 3D dataset can be applied to or used to supplement further comparative and functional analyses of shark morphology. These may include investigations of (a)symmetry, development, integration and modularity or shape change through evolutionary time^[@b22]^. Furthermore, 3D anatomical models provide an excellent visual resource for outreach and education. For example, 3D models can be integrated into 3D PDF documents as interactive figures or can be physically reproduced using rapid prototyping (also referred to as stereolithography or 3D printing)^[@b23]^.\n\nMethods\n=======\n\nSpecimens\n---------\n\n122 individuals representing 73 extant species of 25 out of 34 families from all 9 orders of extant sharks^[@b9]^ (Selachimorpha, Chondrichthyes) are used in the CT scan dataset (Data Citation 1). The specimens are, at the time of CT scanning, formalin preserved and stored in 70% alcohol except for 4 specimens which were stored frozen (RMNH.PISC.36345, RMNH.PISC.verznr.2, 3 and 7). All CT scanned specimens are curated in the spirit collections of the British Museum of Natural History (BMNH) and Naturalis Biodiversity Center (NBC) (see the accompanying metadata for a complete list of specimens). The NBC collections comprise material from the Rijksmuseum van Natuurlijke Historie (RMNH) and the Zo\u00f6logisch Museum Amsterdam (ZMA); those institutional identifiers continue to be used here.\n\nCT scanning\n-----------\n\nCT scans of the specimens housed in the NBC collections were made at the Leiden University Medical Center, the Netherlands (LUMC) with a Toshiba Aquilion 64 medical scanner (Toshiba Medical Systems, Otawara, Japan) using a for the sharks customized scanning and reconstruction protocol (100\u2009kV tube voltage, 150 mAs tube charge per rotation, 64 active channels, acquisition and reconstructed slice thickness 0.5\u2009mm, reconstructed slice increment 0.5\u2009mm, FC03 reconstruction filter, pitch factor 0.83). Specimens from the BMNH collection were scanned at the CT scanning facility of the Royal Brompton and Harefield NHS Trust (RBH), London, United Kingdom, with a Siemens Somatom Sensation 64 medical scanner using a customized scanning protocol with a variable reconstructed slice increment (100\u2009kV tube voltage, 210 mAs tube charge per rotation, acquisition and reconstructed slice thickness 1.0\u2009mm, B30f reconstruction filter).\n\n3D segmentations\n----------------\n\nThe image series from the CT scans were imported into Mimics, v 15.01, (Materialise Software) for segmentation and 3D modelling. We used manual segmentation with a threshold edit of Hounsfield units (i.e., grey values) that are associated to the calcified cartilage of the lower jaw ([Fig. 1](#f1){ref-type=\"fig\"}). The Hounsfield units produced by the medical CT scanners are calibrated according to standard procedures of the manufacturer. The structures of calcified cartilage are not very dense in the scans (meaning relatively low hounsfield units, ranging from 400 to 1,500), and at the border between calcified cartilage and soft tissue, there will always be a partial volume effect that give a \\'smoothed\\' transition from calcified cartilage to soft tissue. Structures with high density, such as the lower jaws, are depicted bright, while structures with low or intermediate density are depicted as dark using optimized greyscales ([Fig. 1a](#f1){ref-type=\"fig\"}). A mask based on measured threshold values of the Hounsfield units, which aligned very closely with the high relative density of the calcified cartilages of the lower jaw is set to produce an exact overlay on the slice images ([Fig. 1b](#f1){ref-type=\"fig\"}). We used these masks to create 3D models of the left lower jaws ([Fig. 1c](#f1){ref-type=\"fig\"}). Threshold values couldn't be standardised across different scans due to variable tissue densities across specimens. Most 3D models could be modeled without ambiguity. Where manual thresholds were required, ambiguity was checked against left-right symmetry of the skeleton.\n\nSubsequently, each 3D model of the left lower jaw was exported as a \\*.PLY file from MIMICS (Data Citation 1).\n\nThirty-one additional 3D models of the left lower jaw were segmented using CT scans provided by other researchers ([Table 1](#t1){ref-type=\"table\"}), resulting in a total of 153 3D models of 94 extant shark species.\n\nData Records\n============\n\nData record 1---The data for this manuscript have been deposited in a Figshare repository (Data Citation 1). It comprises a dataset of 121 of volumes that consist of tomography images ('slices') in DICOM format reconstructed from medical computed tomography (CT) scans of sharks. The matrices of voxels in the dataset represent the Hounsfield units of the corresponding materials and tissues. The Hounsfield unit is associated with a well-defined physics quantity, being the linear attenuation coefficient of these materials and tissues. The majority of the data are whole body CT scans, while a few are CT scans of the head. From the CT scans we generated 153 3D models of the left lower jaws. The accompanying metadata describes the list of CT scanned specimens, their scan parameters and the derived 3D models included in Data record 1. [Table 1](#t1){ref-type=\"table\"} describes the specimen list and scan parameters of the additional 3D models of the left lower jaws that are created from CT scans provided by other researchers.\n\nTechnical Validation\n====================\n\nCT scanner\n----------\n\nMedical CT scanners are subject to a regular program for quality control and maintenance under the responsibility of a qualified medical physicist. However, discrepancies between the reconstructed values in an image and the true attenuation coefficients of the scanned object (i.e., image artifacts) can occur. Three common categories of CT image artefact appearances can be distinguished: streaking, shading, and rings and bands. Streaking artifacts appear as straight lines (bright and/or dark) across the image and are the result of the nature of the filtered backprojection reconstruction process. Shading artifacts can occur near objects of high contrast and usually appear in the soft tissue region near bony structures or near air pockets. This type of artifact is hard to identify since it shows a similar shape as the structure creating the shading. Ring and band artifacts can be visible as rings or bands overlaying the original image structure. The occurrence of artifacts can originate from the system design, x-ray tubes, detector, the specimen or operator^[@b24]^. In our dataset only two CT scans show significant artifacts. BMNH 1978.6.22.1 *Cetorhinus maximus* show ring artifacts in the dense vertebrae and BMNH 1978.6.22.1 *Cetorhinus maximus* shows ring and streaking artifacts in the dense vertebrae and posterior region of the head. Despite these artifacts skeletal structures were identifiable and useable for 3D segmentation. All other scans are free of artifacts or show only negligible artifacts.\n\nThe X-rays, generated by the CT scanner, are used to measure the transmission of X-ray through the specimens under hundreds of different angles. All these measurements are referred to as the raw data. This data is processed with a filtered backprojection, which generates a series of cross-sectional images^[@b25]^. Internal structures are visualized by their ability to attenuate the X-ray beams based on the linear attenuation coefficient. The parameters of the CT scanner were set to optimally visualize the jaws of the sharks. The jaws are well calcified compared to other skeletal structures in the head, therefore structures such as the neurocranium, basihyal and branchial chamber appear less clear in the scans.\n\nUsage Notes\n===========\n\nThe CT scan data in DICOM format can be loaded into 3D analysis software such as the free software package SPIERS^[@b26]^ or in license based software packages such as MIMICS (<http://biomedical.materialise.com/mimics>), AVIZO (<http://www.fei.com/software/avizo3d/>), or VG StudioMax (<http://www.volumegraphics.com/en/products/vgstudio-max/basic-functionality/>). 3D models of the structures of interest can be produced and exported in various formats using one of these software packages. The exported models can be used in landmark-based geometric morphometric methods^[@b27],[@b28]^ to quantitatively test hypotheses of morphological diversity. The IDAV Landmark Editor^[@b29]^ is a free software package well suited for placing landmarks on the 3D models and exporting the landmark coordinates. Note that Landmark Editor only accepts \\*.PLY files without binary encoding. The software package Meshab (<http://meshlab.sourceforge.net/>) can be used to analyse, view and convert the \\*.PLY files to a usable format for Landmark Editor.\n\nThe landmark coordinates generated in Landmark Editor can be used in statistical software designed for geometric morphometric approaches, such as MorphoJ^[@b30]^, Morphologika^[@b31]^, PAST^[@b32]^ or the geomorph package^[@b33]^ in R^[@b34]^.\n\nAdditional Information\n======================\n\n**How to cite this article:** Kamminga, P. *et al.* X-ray computed tomography library of shark anatomy and lower jaw surface models. *Sci. Data* 4:170047 doi: 10.1038/sdata.2017.47 (2017).\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\nWe thank R. de Ruiter and J. Maclaine for granting access to the museum collections of Naturalis Biodiversity Center and the British Museum of Natural History respectively. For the reconstruction additional 3D models of the lower jaws of species that were not present in the museum collection we received scan from several scientists. Therefore, we thank F. Mollen, M. Dean, K. Mara, P. Motta, D. Huber, R. Robins, C. Crawford, D. Walpole, C. Perry, A. Frew and R. Berquist for sharing their data with us. Finally, we thank T. Patel of the RBH for granting access to the CT scan facility and her technical assistance.\n\nThe authors declare no competing financial interests.\n\n![3D segmentation workflow in RMNH.PISC.24047 (*Squatina squatina*).\\\n(**a**) Each CT scan is build up of multiple slices (tomograms) showing structures with high attenuation of X-rays bright and structures with low attenuation dark, using a grey scale. Each pixel in the tomogram is associated with a Hounsfield unit. (**b**) A mask (green) is superimposed on the left lower jaw. This mask is based on predetermined threshold values of the Hounsfield unit of to produce an exact overlay on each tomogram. (**c**) When in each slice a mask is superimposed on the left lower jaw a 3D model is calculated based on the mask. The black line through the 3D model indicates the location of the tomogram in (**a**,**b**), mc l=left lower jaw.](sdata201747-f1){#f1}\n\n###### Additional MC L.\n\n  **Registration number**           **Species**               **Family**       **Sex**   **TL (cm)**                 **Scanned by**                            **Institute**                         **Scanner**             **Scan dose (kV)**   **Scan exposure time (mAs)**   **Algorithm**   **Slice increment (mm)**   **Slice thickness (mm)**\n  ------------------------- ---------------------------- -------------------- --------- ------------- -------------------------------------------- -------------------------------------- --------------------------------- -------------------- ------------------------------ --------------- -------------------------- --------------------------\n  ERB 0854                         Lamna ditropis              Lamnidae           f          234                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0929                          Lamna nasus                Lamnidae           m          174                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0930                          Lamna nasus                Lamnidae           m          166                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0932                     Carcharodon carcharias          Lamnidae           f          212                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0933                       Isurus oxyrinchus             Lamnidae           f          194                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0934                       Isurus oxyrinchus             Lamnidae          NA          230                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0935                         Isurus paucus               Lamnidae           f          254                       F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  ERB 0937                         Lamna ditropis              Lamnidae           f          90                        F. Mollen                            ZNA hospital Antwerp                Philips/Brilliance 40               120                        \u00a0                       B                   0.5                         1\n  FLMNH 44978                 Centroscymnus coelolepis       Somniosidae          f          NA                         D. Huber                       USF Center for Advanced Health      GE_Medical Systems/Lightspeed V          120                       100                    LUNG                 0.625                      0.625\n  LACMNH 3211                Cephaloscyllium ventriosum     Scyliorhinidae       NA          NA                         M. Dean                           UC Irvine Medical Center              Siemens/Sensation 16                120                       196                    H40f                 0.699                       0.75\n  LACMNH 38139                  Stegostoma fasciatum       Stegostomatidae       NA          NA                         M. Dean                           UC Irvine Medical Center              Siemens/Sensation 16                120                       196                    H40f                 0.699                       0.75\n  LACMNH 38291                 Chiloscyllium griseum        Hemiscylliidae       NA          NA                         M. Dean                           UC Irvine Medical Center              Siemens/Sensation 16                120                       196                    H40f                 0.699                       0.75\n  LACMNH 43856                   Galeocerdo cuvier          Carcharhinidae       NA          NA                         M. Dean                           UC Irvine Medical Center              Siemens/Sensation 16                120                       196                    H40f                 0.699                       0.75\n  LACMNH 45857-1             Pseudocarcharias kamoharai   Pseudocarchariidae      f          NA                         M. Dean                       Toshiba America Medical Systems             Toshiba/Aquilion                  120                       125                    FC04                variable                      2\n  LACMNH 47362-1                Mitsukurina owstoni         Mitsukurinidae       NA          NA                         M. Dean                       Toshiba America Medical Systems             Toshiba/Aquilion                  120                       125                    FC04                   1                          2\n  no id                          Carcharias taurus          Odontaspididae       NA          NA                         D. Huber                           Tampa General Hospital               Philips/Brilliance 64               120                        \u00a0                      UA                 variable                     0.9\n  no id                        Carcharhinus acronotus       Carcharhinidae       NA          NA        K. Mara; P. Motta, download from Digimorph     University Diagnostic Institute             Toshiba/Aquilion                  120                       100                    FC30                  0.5                         1\n  no id                           Eusphyra blochii            Sphyrnidae         NA          NA        K. Mara; P. Motta, download from Digimorph     University Diagnostic Institute             Toshiba/Aquilion                  120                        20                    FC13                   2                          2\n  no id                           Sphyrna mokarran            Sphyrnidae         NA          NA        K. Mara; P. Motta, download from Digimorph     University Diagnostic Institute             Toshiba/Aquilion                  120                        20                    FC13                variable                      1\n  no id                            Sphyrna tudes              Sphyrnidae         NA          NA        K. Mara; P. Motta, download from Digimorph     University Diagnostic Institute             Toshiba/Aquilion                  120                        75                    FC30                  0.5                        0.5\n  no id                      Rhizoprionodon terraenovae     Carcharhinidae       NA          NA        K. Mara; P. Motta, download from Digimorph     University Diagnostic Institute            Toshiba/Aquilion 64                120                        75                    FC30                  0.5                         1\n  no id                          Hexanchus griseus           Hexanchidae         NA          NA                         D. Huber                       USF Center for Advanced Health      GE_Medical Systems/Lightspeed V          120                       100                    LUNG                 0.625                      0.625\n  SIO CCS-79-4-5                Negaprion acutidens         Carcharhinidae       NA          NA                         M. Dean                           UC Irvine Medical Center                        \u00a0                          \u00a0                         \u00a0                       \u00a0                  0.499                        \u00a0\n  SIO CCS-79-4-6                Negaprion acutidens         Carcharhinidae       NA          NA                         M. Dean                           UC Irvine Medical Center                        \u00a0                          \u00a0                         \u00a0                       \u00a0                   0.5                         \u00a0\n  UF 160188                    Alopias superciliosus          Alopiidae          NA          NA           R. Robins (provided by C. Crawford)       Medical University of South Carolina        Siemens/Sensation 64                 80                       287                    B10s                  0.6                        0.6\n  USNM 3999                    Orectolobus maculatus        Orectolobidae         f          NA                       C. Crawford                   Medical University of South Carolina        Siemens/Sensation 64                 80                        48                    B10s                  0.6                        0.6\n  SIO 50-200                  Carcharhinus brachyurus       Carcharhinidae       NA          NA                        D. Walpole                    Marine Vertebrate Collection, SIO          3T GE Signa Exite HDx                \u00a0                         \u00a0                       \u00a0                 variable                     0.7\n  SIO 62-1                   Carcharhinus galapagensis      Carcharhinidae       NA          NA                         C. Perry                     Marine Vertebrate Collection, SIO          3T GE Signa Exite HDx                \u00a0                         \u00a0                       \u00a0                    1                          1\n  no id                           Mustelus henlei             Triakidae           f          NA                       R. Berquist                    California Institute of Technology      7T Bruker Biospec Avance II             \u00a0                         \u00a0                       \u00a0                   0.2                        0.2\n  SIO 91-85                     Orectolobus ornatus         Orectolobidae        NA          NA                         A. Frew                      Marine Vertebrate Collection, SIO            7T Bruker---UCLA                   \u00a0                         \u00a0                       \u00a0                   0.1                        0.1\n  PSRC Uncat                    Proscyllium habereri        Proscylliidae        NA          NA                       R. Berquist                      Pacific Shark Research Center         7T Bruker Biospec Avance II             \u00a0                         \u00a0                       \u00a0                   0.1                        0.1\n\n[^1]: P.K. and M.D.B. conceived and designed the project, P.K. reconstructed the 3D models and wrote the paper. P.K., P.W.D.B. and J.G. collected and reconstructed the CT scan data.\n"}
+{"text": "Introduction\n============\n\nMyxopapillary ependymoma was first described as a subtype of ependymoma by Kernohan \\[[@B1]\\] in 1932. Myxopapillary ependymoma is a tumor that typically arises in the filum terminale and conus medullaris \\[[@B1],[@B2]\\]. This tumor has a significantly better prognosis than other ependymomas, because it is slow-growing, unlikely to metastasize and does not tend to infiltrate neural tissues \\[[@B3]\\]. Intracranial ependymomas represent 6%-9% of primary central nervous system neoplasms and generally present in young children, with a mean age of four-year. Spinal myxopapillary ependymomas are rarer than intracranial types. Myxopapillary ependymomas generally present within the spinal canals of adults. Spinal myxopapillary ependymomas in children \\<5 years requiring surgery are very rare.\n\nHere we describe a rare case of a 5-year-old boy who presented with cauda equina compression due to myxopapillary ependymoma, which was treated with subtotal resection and postoperative radiation therapy.\n\nCase Report\n===========\n\nA 5-year-old boy presented with a history of right popliteal pain and abnormal gait that began after March 2000, four months prior to admission. Three months before admission, the patient visited a neighborhood hospital complaining of right popliteal pain, pain in the right hip and an abnormal gait. A magnetic resonance imaging (MRI) scan of the lumbar spine showed a tumor in the cauda equina. He was referred to our hospital and was admitted for examination. On admission he had no motor, sensory or urinary disturbances. Decreased tone and hyporeflexia in the lower extremities were observed. His medical history indicated no abnormalities. Radiographs revealed wide interpedicular distances in the lower lumbar vertebrae ([Fig. 1](#F1){ref-type=\"fig\"}). T2-weighted sagittal MRI showed a high-signal mass of cauda equina extending from L1 to L4 ([Fig. 2](#F2){ref-type=\"fig\"}). The mass was enlarged compared to that on the MRI performed at the neighborhood hospital. As such we performed subtotal resection with laminoplasty from T12 to L4. Histological findings from the tissue removed during surgery showed myxopapillary ependymoma with a positive margin ([Fig. 3](#F3){ref-type=\"fig\"}). Our attempt to perform radiation therapy could not proceed due to his parents\\' refusal of the therapy. A vesicorectal disorder occurred immediately after surgery, but it improved naturally over time.\n\nTwo years after admission, follow-up MRI revealed a recurrent enhanced tumor from L2 to L4, with a small seeding lesion at the sacral area ([Fig. 4](#F4){ref-type=\"fig\"}). There was a sensory deficit of analgesia from the L2 to L5 area. Leg pain and motor neuron deficits were not observed. We performed subtotal resection with laminectomy from L1 to L4 ([Fig. 5](#F5){ref-type=\"fig\"}). The tumor was tightly adhered to many bundles of the cauda equina, and several bundles were coursing through the tumor. Thus the tumor was only partially removed. Postoperative radiation therapy (45 Gy for the whole brain and spinal cord) was performed. After the second surgery, a urinary disorder occurred, which persists today.\n\nMRI of the spine nine years after the second surgery showed no evidence of regrowth of the residual tumor ([Fig. 6](#F6){ref-type=\"fig\"}). Neutral standing radiographs ten years after the second surgery showed no spinal deformity ([Fig. 7](#F7){ref-type=\"fig\"}). No further evidence of the disease has been noted elsewhere in the patient in the ten-year of follow-up.\n\nThe patient and his family provided consent for data in this case to be published.\n\nDiscussion\n==========\n\nMyxopapillary ependymomas arise almost exclusively from the conus medullaris, cauda equina and filum terminale \\[[@B1],[@B2]\\]. The World Health Organization (WHO) classifies ependymomas into the following four categories: subependymomas (WHO grade 1), myxopapillary ependymomas (WHO grade 1), ependymomas (WHO grade 2) and anaplastic ependymomas (WHO grade 3) \\[[@B4]\\]. Myxopapillary ependymomas are slow-growing tumors that may become very large before clinical detection, occasionally filling and expanding throughout the lumbosacral canal \\[[@B5]\\].\n\nMyxopapillary ependymomas in children typically occur in the central nervous system. Therefore there are few cases of spinal myxopapillary ependymoma in children. There are some reports on surgical procedures used to treat spinal myxopapillary ependymoma. Nakamura et al. \\[[@B6]\\] and Quraishi et al. \\[[@B7]\\] reported the surgical outcomes of this tumor in patients of 14-58 years and patients of 20-50 years, respectively. The 5-year-old boy in the current report was the youngest of patients in the earlier reports.\n\nComplete resection of myxopapillary ependymomas is associated with excellent long-term prognosis. However it is sometimes difficult to completely resect these tumors because of their tight adherence to the spinal cord and cauda equina. Al-Habib et al. \\[[@B8]\\] reviewed fifteen myxopapillary ependymoma patients (age range, 18-71 years). They reported that 17% of patients developed tumor regrowth requiring reoperations, while none of the patients with total resection developed recurrence.\n\nThe role of radiation therapy has not been adequately studied. Radiation therapy is likely warranted in cases of subtotal resection and recurrence of the intradural tumors, but is controversial in cases of total resection and extradural ependymomas \\[[@B9]\\]. In our case, the tumor recurred after the first surgery in the absence of postoperative radiation therapy. In contrast when postoperative radiation therapy was used following the second surgery, a relapse did not occur. As such radiation therapy is considered effective against myxopapillary ependymomas.\n\nThis patient represents a young case of spinal myxopapillary ependymoma of the cauda equina with compression. We successfully treated the patient with subtotal tumor resection and postoperative radiation therapy of the entire brain and spinal cord. No further evidence of the disease has been noted elsewhere in the patient during ten years of follow-up. Spinal deformity is thought to occur when young children undergo laminectomy. However in our case, no spinal deformity occurred in the 5-year-old boy over a ten-year follow-up period.\n\nWe reported a rare case of spinal myxopapillary ependymoma with cauda equina compression. Subtotal tumor resection and postoperative radiation therapy of the entire brain and spinal cord were effective, and the patient had a good prognosis.\n\nNo potential conflict of interest relevant to this article was reported.\n\n![Plain radiographs revealed wider interpedicular distances in the lower lumbar vertebrae. Lumbar spine alignment was normal. There were no sclerotic lesions or lytic destruction noted.](asj-8-846-g001){#F1}\n\n![Magnetic resonance imaging showed a mass within the cauda equina extending from L1 to L4. The tumor was hypointense on T1-weighted images and hyperintense on T2-weighted images with gadolinium-enhancement. Gd-Enhanced, gadolinium-enhanced.](asj-8-846-g002){#F2}\n\n![Histological examination of the tissue removed during the surgery showed that the myxoid area was positioned between small vessels, around which small tumor cells aggregated. The tumor cells were positive for glial fibrillary acidic protein. (**A**) H&E, \u00d7100. (**B**) Glial fibrillary acidic protein, \u00d7100.](asj-8-846-g003){#F3}\n\n![Two years after admission, follow-up magnetic resonance imaging revealed a recurrent enhanced tumor from L2 to L4 with a small seeding lesion within the sacral area. Gd-Enhanced, gadolinium-enhanced.](asj-8-846-g004){#F4}\n\n![Magnetic resonance imaging after the second surgery showed that laminectomy was performed from L1 to L4 to resect the tumor. However the tumor was only partially removed, because it was tightly adhered to many bundles of the cauda equina. Gd-Enhanced, gadolinium-enhanced.](asj-8-846-g005){#F5}\n\n![Magnetic resonance imaging of the entire spine nine years after the second surgery showed no evidence of residual tumor regrowth.](asj-8-846-g006){#F6}\n\n![Neutral standing radiographs ten years after the second surgery showed no spinal deformities.](asj-8-846-g007){#F7}\n"}
+{"text": "INTRODUCTION {#S1}\n============\n\nGenomic integrity is constantly threatened by various endogenous and environmental insults. Fortunately, cells are equipped with specialized molecular machineries to correct specific types of DNA damage. In particular, nucleotide excision repair (NER) is a versatile DNA repair pathway that eliminates a wide variety of helix-distorting base lesions induced by environmental carcinogenic sources**,** including two products induced by solar ultraviolet B (UVB) radiation, cyclobutane pyrimidine dimers (CPD) and ([@R6]-[@R4]) photoproducts (6-4PP), as well as other bulky adducts induced by air pollutants ([@R1]-[@R7]).\n\nMammalian NER consists of two distinct subpathways: global genome NER (GG-NER), which operates throughout the entire genome, and transcription-coupled NER (TC-NER), which specifically removes lesions on the transcribed DNA strand of active genes ([@R4], [@R5], [@R8]-[@R10]). A major difference between these two pathways appears to lie in the strategies for detecting damaged bases and the phenotypes in human patients. Defects in GG-NER cause xeroderma pigmentosum (XP), an autosomal recessive disorder predisposing affected individuals to cancer development not only in the skin, but also in the brain and lungs ([@R4], [@R5], [@R9], [@R11]). Seven XP factors (XPA-XPG) have been identified as indispensable NER factors ([@R4], [@R5], [@R8]-[@R10]). Accumulating evidence indicates that XPC plays an essential role in GG-NER-specific damage recognition ([@R12]-[@R14]). XPC activity in NER requires polyubiquitination by the DDB E3-ligase complex that comprises the DDB1 (DNA damage-binding protein 1), DDB2, CUL4A (Cullin-family E3-ligase adaptor protein) and ROC1 (E3-ligase RING domain) subunits following UV radiation ([@R15]-[@R17]).\n\nRecent studies from our group and others have demonstrated that the availability and the activity of these NER factors are regulated by endogenous and environmental factors to control the NER capacity ([@R18], [@R19]). Since these XP factors are difficult to target directly, identification of their upstream regulators will provide valuable tools for enhancing NER and therefore tumor suppression.\n\nAs a calcium-dependent cell adhesion molecule, E-cadherin is a key component of adherens junctions, structures that are crucial for the maintenance of epithelial integrity ([@R20], [@R21]). It is a calcium-dependent adhesion molecule that mediates homophilic interactions and the formation of catenin-containing complexes ([@R20]-[@R22]). In cancer, loss of E-cadherin is the best-characterized alteration in the activation of cancer cell invasion and metastasis in multiple cancer types ([@R21], [@R23], [@R24]), including skin cancer ([@R25]-[@R29]). However, the role of E-cadherin in DNA repair and tumor initiation remains unknown.\n\nUsing mouse models with tissue-specific E-cadherin inactivation, recent studies have shown that loss of E-cadherin induces spontaneous tumorigenesis in the liver ([@R30]), skin and mammary gland in the absence of p53 ([@R31]), and increases chemical or oncogene-induced liver tumorigenesis ([@R30]). Furthermore, ultraviolet radiation (UV) down-regulates E-cadherin in keratinocytes and melanocytes ([@R28], [@R32]). These findings suggest that loss of E-cadherin is an early event and may play an active role in early stage of tumorigenesis. It is possible that the early loss of E-cadherin actively contributes to tumor initiation by impairing UV-induced DNA repair, leading to increased genetic mutations and thus UV-induced tumorigenesis. Here we show that E-cadherin positively regulates UVB-induced DNA repair by promoting the transcription of XPC and DDB1 to facilitate damage recognition.\n\nRESULTS {#S2}\n=======\n\nE-cadherin loss inhibits nucleotide excision repair and down-regulates XPC and DDB1 {#S3}\n-----------------------------------------------------------------------------------\n\nTo determine whether E-cadherin has a role in the regulation of NER, we compared NER capability and NER protein levels in control and E-cadherin-inhibited cells. E-cadherin inhibition significantly reduced both CPD repair ([Fig. 1A-B](#F1){ref-type=\"fig\"}) and 6-4PP repair ([Fig. 1C-D](#F1){ref-type=\"fig\"}). As compared with the control group, E-cadherin knockdown by either siRNA or shRNA decreased the levels of XPC and DDB1, while it did not affect other NER factors, including XPA, XPB, XPD, XPF, XPG, DDB2 and Cul4A ([Fig. 1E](#F1){ref-type=\"fig\"}). In addition, as compared with sham-irradiated mouse epidermis, E-cadherin was down-regulated in UVB-irradiated mouse epidermis and tumors ([Fig. S1A](#SD2){ref-type=\"supplementary-material\"}). Immunohistochemical analysis showed that, as compared with normal human skin, in actinic keratosis (AK) and human squamous cell carcinoma (SCC) the protein levels of E-cadherin, DDB1 and XPC were decreased (score 0 or 1) ([Fig. 1F-G](#F1){ref-type=\"fig\"}). This reduction was statistically significant as analyzed by the Mann-Whitney U test (*P* \\< 0.0005 for AK or SCC *vs* normal skin). The protein levels of XPC and DDB1 were positively correlated with the E-cadherin protein level in human skin and tumors ([Fig. S1B](#SD2){ref-type=\"supplementary-material\"}) (*P* \\< 0.0001). These results indicate that E-cadherin loss inhibits NER and down-regulates XPC and DDB1.\n\nE-cadherin regulates 6-4PP repair but not CPD repair *via* increasing XPC availability {#S4}\n--------------------------------------------------------------------------------------\n\nTo determine the mechanism by which E-cadherin regulates the repair of CPD and 6-4PP, we assessed the functional significance of XPC and DDB1 regulation by E-cadherin. UVB down-regulated E-cadherin protein levels ([Fig. 2A](#F2){ref-type=\"fig\"}), consistent with previous studies ([@R28]). E-cadherin knockdown reduced UVB-induced XPC polyubiquitination ([Fig. 2A](#F2){ref-type=\"fig\"}), a biochemical process critical for DNA damage recognition mediated by the UV-DDB complex ([@R15]-[@R17]). E-cadherin knockdown inhibited the recruitment of XPC to subnuclear UV-induced CPD sites, while it had no effect on the recruitment of DDB2 to CPD sites ([Fig. 2B-D](#F2){ref-type=\"fig\"}, [S2A](#SD3){ref-type=\"supplementary-material\"}). It is possible that the recruitment of XPC protein was due to the reduced XPC protein level or the reduced XPC ubiquitination. Our findings are in line with previous reports supporting the recruitment of DDB2 independent of XPC as well as DDB1 ([@R33], [@R34]). In E-cadherin knockdown cells, XPC addition by transient transfection with the pCMV-XPC construct did not affect DDB1 expression or XPC polyubiquitination ([Fig. 2E](#F2){ref-type=\"fig\"}). However, XPC addition increased 6-4PP repair to an extent similar to control cells but did not affect CPD repair ([Fig. 2F-I](#F2){ref-type=\"fig\"}). At the UVB dose used (20 mJ/cm^2^), UVB did not induce apoptosis, nor did the addition of XPC and/or DDB1 affect post-UVB survival ([Fig. S2B](#SD3){ref-type=\"supplementary-material\"}). In contrast, in XPC null cells with the same DDB1 protein levels, XPC addition by lentiviral infection significantly increased the repair of both CPD and 6-4PP ([Fig. S2C-G](#SD3){ref-type=\"supplementary-material\"}). These results demonstrate that E-cadherin regulates 6-4PP repair via increasing basal XPC levels while it may regulate CPD repair via increasing basal XPC levels and activity.\n\nE-cadherin regulates CPD repair through increasing the levels of both XPC and DDB1 {#S5}\n----------------------------------------------------------------------------------\n\nTo determine how E-cadherin regulates CPD repair, we assessed the role of both DDB1 and XPC. In E-cadherin-inhibited cells, adding both XPC and DDB1 by transient transfection with the combination of the pCMV-XPC and pcDNA-DDB1 constructs increased XPC polyubiquitination ([Fig. 3A](#F3){ref-type=\"fig\"}) and CPD repair as well as 6-4PP repair ([Fig. 3B-E](#F3){ref-type=\"fig\"}). Immunofluorescence analysis demonstrated that addition of XPC alone did not affect recruitment of XPC to the subnuclear CPD sites, while addition of both XPC and DDB1 restored the recruitment of XPC to CPD sites to a level similar to control cells ([Fig. 3F-G](#F3){ref-type=\"fig\"}). These data demonstrate that both XPC and DDB1 are required for E-cadherin regulation of CPD repair.\n\nThe TGF-\u03b2 pathway but not \u03b2-catenin is required for E-cadherin loss to induce inhibition of Nucleotide Excision Repair {#S6}\n----------------------------------------------------------------------------------------------------------------------\n\nTo determine the downstream effector pathway in E-cadherin regulation of NER, we first assessed the role of \u03b2-catenin, which has been shown to translocate from the plasma membrane to the nucleus and become activated ([@R35]). E-cadherin knockdown induced nuclear translocation of \u03b2-catenin and the activation of TCF/LEF complex, indicating the activation of the Wnt/\u03b2-catenin pathway ([Fig. S3A-B](#SD4){ref-type=\"supplementary-material\"}). However, in E-cadherin-inhibited cells, \u03b2-catenin knockdown had no effect on XPC and DDB1 expression, XPC polyubiquitination ([Fig. S3C](#SD4){ref-type=\"supplementary-material\"}), or the repair of CPD and 6-4PP ([Fig. S3D-G](#SD4){ref-type=\"supplementary-material\"}), indicating that E-cadherin regulation of nucleotide excision repair does not require the Wnt/\u03b2-catenin pathway.\n\nNext we determined the role of the TGF-\u03b2 pathway, which has been found to be suppressed by E-cadherin ([@R11]). As compared with sun-protected normal human skin, in human SCC E-cadherin down-regulation was associated with the nuclear translocation of Smad3 ([Fig. S4A](#SD5){ref-type=\"supplementary-material\"}), suggesting activation of the TGF-\u03b2 pathway ([@R36]-[@R38]). A luciferase reporter assay of the SBE4 promoter showed that E-cadherin knockdown activated the TGF-\u03b2 pathway, which was sensitive to the pan TGF-\u03b2 neutralization antibody ([Fig. 4A](#F4){ref-type=\"fig\"}) and the TGF-\u03b2 pathway inhibitor LY364947 ([Fig. 4B](#F4){ref-type=\"fig\"}). E-cadherin inhibition increased the TGF-\u03b21 mRNA level ([Fig. 4C](#F4){ref-type=\"fig\"}). These data indicate that increased TGF-\u03b21 expression has a role in TGF-\u03b2 pathway activation by E-cadherin knockdown.\n\nIn E-cadherin-inhibited cells, the pan TGF-\u03b2 neutralization antibody increased the levels of both XPC and DDB1 ([Fig. S4B](#SD5){ref-type=\"supplementary-material\"}); LY364947 increased the levels of both XPC and DDB1 and polyubiquitination of XPC ([Fig. 4D](#F4){ref-type=\"fig\"}). LY364947 increased the recruitment of XPC to subnuclear CPD sites ([Fig. S4C-D](#SD5){ref-type=\"supplementary-material\"}) and the repair of both CPD and 6-4PP ([Fig. 4E-F](#F4){ref-type=\"fig\"}). At the UVB irradiation dose used for NER analysis (20 mJ/cm^2^), LY364947 did not affect UVB-induced apoptosis ([Fig. S4E](#SD5){ref-type=\"supplementary-material\"}). In HaCaT cells TGF-\u03b21 treatment down-regulated both XPC and DDB1 ([Fig. 4G](#F4){ref-type=\"fig\"}). E-cadherin was also decreased by TGF-\u03b21 treatment ([Fig. 4G](#F4){ref-type=\"fig\"}), consistent with previous studies in HaCaT and other cell lines detecting TGF-\u03b21-induced E-cadherin suppression ([@R29], [@R39]), ([@R40]). Similarly, in normal human epidermal keratinocytes (NHEK), TGF-\u03b21 down-regulated XPC and DDB1 ([Fig. 4H](#F4){ref-type=\"fig\"}) and suppressed the repair of both CPD and 6-4PP ([Fig. 4I-J](#F4){ref-type=\"fig\"}). However, in these cells TGF-\u03b21 did not affect the E-cadherin level ([Fig. 4H](#F4){ref-type=\"fig\"}). These results demonstrate that the TGF-\u03b2 pathway produces phenocopies of E-cadherin loss and is the downstream effector of E-cadherin loss in suppressing NER.\n\nTGF-\u03b2 pathway inhibits XPC and DDB1 transcription {#S7}\n-------------------------------------------------\n\nThe TGF-\u03b2 pathway regulates various physiological and pathological responses largely through controlling gene transcription *via* the Smad proteins ([@R36]-[@R38]). To determine the mechanism in the regulation of XPC and DDB1 by the E-cadherin/TGF-\u03b2 pathway, we tested the hypothesis that the TGF-\u03b2 pathway suppresses the transcription of XPC and DDB1 in E-cadherin-inhibited cells. Indeed, E-cadherin knockdown reduced the mRNA levels of both XPC and DDB1, which was prevented by the TGF-\u03b2 pathway inhibitor ([Fig. 5A](#F5){ref-type=\"fig\"}).\n\nNext we generated the XPC promoter luciferase construct between the \u22121187 and 27 positions relative to the XPC TSS (transcription start site) and the DDB1 promoter luciferase construct between the \u22121138 and 124 positions relative to the DDB1 TSS. E-cadherin knockdown inhibited transcription of both XPC and DDB1, which was prevented by the TGF-\u03b2 pathway inhibitor ([Fig. 5B](#F5){ref-type=\"fig\"}). Similar results were obtained from the shorter promoters of both XPC (\u2212313-27) and DDB1 (\u2212467-124) ([Fig. 5C](#F5){ref-type=\"fig\"}). TGF-\u03b21 inhibited the transcription of the promoters of both XPC (\u2212313-27) and DDB1 (\u2212467-124) ([Fig. 5D](#F5){ref-type=\"fig\"}). Using the TRANSSERCH and TRANSFAC programs, we predicted several candidate transcription factor-binding elements in the shorter promoters of both XPC (\u2212313-27) and DDB1 (\u2212467-124) ([Fig. S5A-B](#SD6){ref-type=\"supplementary-material\"}).\n\nNext we generated promoters with deletions of the five candidate transcription factor-binding elements of the XPC promoter (CREB2, SP1-1/SP1-2, CREB1, E2F and NKX2.5/SBE) and the DDB1 promoter (N-Myc, NF1, SP1-2, E2F/SP1-1(TIE)/SBE-3 and SBE-1/SBE-2) identified. As compared with the WT XPC promoter, SP1-1/SP1-2 deletion significantly inhibited basal XPC transcription ([Fig. 5E-F](#F5){ref-type=\"fig\"}). However, as compared with the same WT XPC promoter, deletion of the SP1-1/SP1-2, CREB2, CREB1, or NKX2.5/SBE elements had no effect on TGF-\u03b21-induced suppression of XPC transcription. In contrast, deletion of the E2F site completely blocked TGF-\u03b21-induced suppression of XPC transcription ([Fig. 5E-F](#F5){ref-type=\"fig\"}), indicating that the E2F site in the XPC promoter is required for the regulation of XPC transcription by TGF-\u03b21.\n\nAs compared with the WT DDB1 promoter, deletion of the SP1-2 or N-Myc element significantly reduced basal DDB1 transcription ([Fig. 5E and G](#F5){ref-type=\"fig\"}). However, as compared with the WT DDB1 promoter, deletion of the N-Myc, NF1, SP1-2 or SBE-1/SBE-2 elements had no effect on TGF-\u03b21-induced suppression of DDB1 transcription. In contrast, deletion of the E2F/SP1-1(TIE)/SBE-3 site completely blocked TGF-\u03b21-induced suppression of DDB1 transcription ([Fig. 5E and G](#F5){ref-type=\"fig\"}). Site-specific mutation of the XPC E2F element completely blocked TGF-\u03b21-induced suppression of XPC transcription ([Fig. 5E and H](#F5){ref-type=\"fig\"}). In comparison, site-specific mutation of the DDB1 E2F or SBE-3 elements, but not the SP1-1(TIE) element, completely blocked TGF-\u03b21-induced suppression of DDB1 transcription ([Fig. 5E and I](#F5){ref-type=\"fig\"}, [S5C](#SD6){ref-type=\"supplementary-material\"}). Taken together, these data demonstrate that E-cadherin loss suppresses XPC and DDB1 transcription through activating the TGF-\u03b2 pathway. TGF-\u03b2 activation inhibits XPC and DDB1 transcription through the E2F element of the XPC promoter and the E2F and SBE-3 elements of the DDB1 promoter.\n\nThe TGF-\u03b2 pathway promotes binding of the E2F4/P130 complex to the XPC promoter and the E2F4/Smad3/p107 complex to the DDB1 promoter {#S8}\n------------------------------------------------------------------------------------------------------------------------------------\n\nThe TGF-\u03b2 pathway has been demonstrated to induce the transcription-repressing activity of E2F4/5/p107/Smad via the E2F and Smad-binding sites ([@R41]). We found that E2F4 and p130 bind to the XPC promoter with the E2F element, and E2F4, p107 and Smad3 bind to the DDB1 promoter with the E2F/SP1-1(TIE)/SBE-3 element ([Fig. 6A-B](#F6){ref-type=\"fig\"}). TGF-\u03b21 significantly increased these bindings ([Fig. 6A-B](#F6){ref-type=\"fig\"}). Immunoprecipitation analysis demonstrated that Smad3 binds with E2F4 and p107 but not p130; TGF-\u03b21 does not affect this binding ([Fig. 6C](#F6){ref-type=\"fig\"}). p130 bound with E2F4 but not Smad3; TGF-\u03b21 increased the E2F-p130 interaction ([Fig. 6C](#F6){ref-type=\"fig\"}).\n\np130 knockdown blocked TGF-\u03b21-induced suppression of XPC transcription, while it had no effect on DDB1 transcription ([Fig. 6D-E](#F6){ref-type=\"fig\"}). In contrast, p107 knockdown prevented TGF-\u03b21-induced suppression of DDB1 transcription, while it had no effect on XPC transcription ([Fig. 6F-G](#F6){ref-type=\"fig\"}). p130 knockdown also prevented E-cadherin-loss-induced down-regulation of XPC but not DDB1, whereas p107 knockdown blocked E-cadherin-loss-induced down-regulation of DDB1 but not XPC ([Fig. 6H](#F6){ref-type=\"fig\"}). Taken together, these data demonstrate that the TGF-\u03b2 pathway promotes the binding of E2F4 and p130 to the E2F element in the XPC promoter and thereby suppresses XPC transcription, and enhances the binding of the E2F4/Smad3/p107 complex to the E2F/SP1-2(TIE)/SBE-3 element in the DDB1 promoter and thereby suppresses DDB1 transcription.\n\nDISCUSSION {#S9}\n==========\n\nLoss of E-cadherin is the best-studied hallmark in cancer progression and metastasis ([@R21], [@R23]-[@R29]). Here we show that E-cadherin is critical for the efficient repair of UVB-induced DNA damage. Loss of E-cadherin impairs UV-induced DNA repair by suppressing the transcription of XPC and DDB1. In human AK and SCC specimens, E-cadherin down-regulation is associated with reduced levels of XPC and DDB1, as compared with normal human skin samples. The activation of the TGF-\u03b2 pathway mimics E-cadherin loss and is required for the suppression of XPC and DDB1 expression and UV-induced DNA damage repair in E-cadherin-inhibited cells. Our findings elucidate an unexpected role of E-cadherin in DNA damage repair and suggest a molecular mechanism for E-cadherin's suppressive role in tumor initiation.\n\nNucleotide excision repair is regulated at multiple levels by extracellular cues and intracellular signaling pathways through regulating the expression or activity of DNA repair factors ([@R18], [@R19]). In this study we found that two NER factors, XPC and DDB1, are transcriptionally regulated by E-cadherin via regulation of different E2F4 transcription repressor complexes. E-cadherin loss increased the enrichment of the E2F4/p130 complex at the XPC promoter. In contrast, at the DDB1 promoter, E-cadherin loss elevated the E2F4/p107/Smad3 complex, a signaling effector complex found in TGF-\u03b2-induced suppression of c-Myc ([@R41]). The recruitment of Smad3 to the DDB1 promoter is TGF-\u03b21-dependent ([Fig. 6B](#F6){ref-type=\"fig\"}), but its interaction with E2F4/p107 complex is TGF-\u03b21-independent. It is possible that TGF-\u03b21 induces the nuclear translocation of the existing Smad3/E2F4/p107 and thus the binding of this complex to the DDB1 promoter. Indeed, previous reports indicated that the XPC gene is repressed by the E2F4-p130 repressor complex ([@R42]), and that disruption of this complex by the tumor suppressor ARF (alternative reading frame) or SIRT1-mediated PTEN deacetylation stimulates XPC transcription ([@R43], [@R44]).\n\nUsing a promoter luciferase reporter assay, a recent study showed that the basal DDB1 transcription regulation involves the E2F element, as well as the Sp1, N-Myc, and NF1 elements ([@R45]). The E2F sites in the promoters of XPC and DDB1 are required for TGF-\u03b2-induced suppression of XPC and DDB1 transcription, whereas the SBE site is only required for DDB1 promoter transcription. Together with previous studies on TGF-\u03b2-induced c-Myc suppression, which requires the SBE element ([@R46]), our findings suggest a gene-specific regulatory mechanism for TGF-\u03b2 signaling. In addition to E2F4, E2F5 may also play an important role, as detected in TGF-\u03b2-induced c-Myc repression ([@R41]). Further investigation can elucidate the specific E2F family members required for XPC and DDB1 regulation by TGF-\u03b2. Together with the recent report showing that XPC also positively regulates E-cadherin ([@R47]), XPC and E-cadherin may form a positive feedback loop to suppress tumorigenesis and tumor progression.\n\nAlthough the DDB2 promoter contains a functional E2F element as does the DDB1 promoter ([@R45]), DDB2 expression was not affected by E-cadherin loss, suggesting that E-cadherin specifically regulates DDB1. Furthermore, it is noteworthy that E-cadherin promotes 6-4PP repair through regulating XPC, while it promotes CPD repair via regulating the expression of both XPC and DDB1. These distinct mechanisms are supported by the specific requirement of NER factors for 6-4PP and CPD: XPC is required for the repair of both 6-4PP and CPD, whereas the UV-DDB complex including DDB1 is required for efficient repair of CPD, but not 6-4PP ([@R33], [@R48]). Our findings demonstrate a critical molecular link of E-cadherin to nucleotide excision repair.\n\nE-cadherin interacts with multiple proteins and regulates related downstream pathways. E-cadherin is known to recruit \u03b2-catenin to the cell membrane and thus prevent \u03b2-catenin's nuclear localization and transactivation ([@R35]). However, we found that regulation of nucleotide excision repair by E-cadherin loss does not require \u03b2-catenin nuclear translocation and its transactivation ([Fig. S3A-G](#SD4){ref-type=\"supplementary-material\"}). Instead, we found that E-cadherin inhibition suppresses nucleotide excision repair through activating the TGF-\u03b2 pathway. E-cadherin knockdown increased the TGF-\u03b21 mRNA level in keratinocytes ([Fig. 4C](#F4){ref-type=\"fig\"}), consistent with the findings in liver cells ([@R49]).\n\nTGF-\u03b2 signaling has a pleotropic and context-dependent role in tumorigenesis and tumor progression ([@R36]-[@R38]), including in the skin ([@R50]). Recent work has shown that deletion of Smad2 or Smad4 increases skin tumorigenesis ([@R51], [@R52]), and Smad4 loss reduces UV-induced DNA repair ([@R52]). In contrast, Smad3 deletion inhibits skin tumorigenesis ([@R53]). It is possible that suppression of CPD repair by TGF-\u03b21/Smad3 may play an active role in tumor initiation, since CPD but not 6-4PP drives skin tumorigenesis ([@R54]). Although TGF-\u03b21 promotes non-homologous end-joining repair and protects cells from ionizing radiation ([@R55]), TGF-\u03b21/Smad3 has been shown to inhibit the BRCA1-dependent repair of double-strand breaks induced by mitomycin C ([@R56]). In the chemical carcinogenesis model, TGF-\u03b21 enhances early skin tumorigenesis in mice ([@R57]); UVB radiation induces TGF-\u03b2 biosynthesis and activation ([@R58]); and TGF-\u03b21 signaling promotes UVB-induced skin tumorigenesis ([@R59]), suggesting that TGF-\u03b2 signaling is oncogenic in the UVB response. Further studies can assess the precise role of the TGF-\u03b2 signaling and its individual molecular components in the UVB-induced skin tumorigenesis model. Nevertheless, our findings provide a molecular mechanism for TGF-\u03b2 signaling linking E-cadherin loss to suppression of nucleotide excision repair and suggest a crucial role of the E-cadherin/TGF-\u03b2 axis in tumor initiation following UVB damage.\n\nWe primarily focused on UVB irradiation for the majority of these experiments and only used UVC for the localized UV irradiation experiments, since ([@R1]) both UVC and UVB cause the formation of CPD and 6-4PP as the primary DNA damage type, and ([@R2]) UVB is clinically relevant to UV-associated human skin damage. However, we were unable to detect subnuclear formation of CPD following local UVB irradiation. It is possible that the Millipore filters block UVC while they may not sufficiently block UVB. Therefore all the localized UV irradiation experiments were performed using UVC but not UVB, as in previous reports ([@R12], [@R60]-[@R64]). Further investigation in greater molecular detail will clarify whether the NER process is regulated differently in response to UVC and UVB irradiation.\n\nIn summary, our findings demonstrate that E-cadherin positively regulates nucleotide excision repair by promoting the transcription of XPC and DDB1. E-cadherin loss causes the activation of TGF-\u03b2 signaling, which induces transcription repression of both XPC and DDB1. Our findings may shed light on the previously unrecognized positive role of E-cadherin in DNA repair and tumor initiation.\n\nMATERIALS AND METHODS {#S10}\n=====================\n\nAll human specimens were studied after approval by the University of Chicago Institutional Review Board. All animal procedures have been approved by the University of Chicago Institutional Animal Care and Use Committee. HaCaT cells (human keratinocytes and epithelial cells, kindly provided by Dr. Fusenig) were used. Detailed descriptions of all methods are provided in *SI* Methods.\n\nSupplementary Material {#SM}\n======================\n\nWe are grateful to Dr. Altaf Wani for his helpful suggestions. We thank Terri Li for immunohistochemistry and Dr. Ann Motten for a critical reading of the manuscript. This work was supported by the NIH/NIEHS grant ES016936 and ES024373 (YYH), the American Cancer Society (ACS) grant RSG-13-078-01 (YYH), the University of Chicago Cancer Research Center (P30 CA014599), the CTSA (UL1 TR000430), and the University of Chicago Friends of Dermatology Endowment Fund.\n\n**CONFLICT OF INTEREST**\n\nThe authors have no conflict of interest.\n\n![E-cadherin loss inhibits nucleotide excision repair and down-regulates XPC and DDB1\\\n(A and C) Slot blot analysis of the levels of CPD (A) and 6-4PP (C) in HaCaT cells transfected with vector (shCon) or shRNA targeting E-cadherin (shE-cadherin, or shE-cad) at 0, 6 and 24 h post-UVB (20 mJ/cm^2^) for CPD and 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^) for 6-4PP. (B and D) Quantification of percentage (%) of CPD repair (B) from A and 6-4PP repair (D) from C. \\*, *P* \\< 0.05, compared with shCon group, Student's *t*-test and two-way ANOVA. (E) Immunoblot analysis of E-cadherin, XPC, DDB1, XPA, XPB, XPD, XPD, XPF, XPG, DDB2, Cul4a and GAPDH in HaCaT cells transfected with shCon or shE-cadherin, or siRNA targeting E-cadherin (siE-cadherin) or control siRNA (siCon). The results were obtained from three independent experiments. (F) Representative immunohistochemical analysis of E-cadherin, XPC, and DDB1 protein levels (brown) in normal skin (n=14), actinic keratosis (AK) (n=25), and squamous cell carcinoma (SCC) (n=25). Scale bar, 50 \u03bcm. (G) Comparison of E-cadherin staining (score) in normal epidermis compared with the panel of AK and SCCs. Statistical analysis was performed using a Mann--Whitney U test.](nihms-719076-f0001){#F1}\n\n![E-cadherin regulates 6-4PP repair but not CPD repair via increasing XPC availability\\\n(A) Immunoblot analysis of E-cadherin, XPC, Ub-XPC (polyubiquitinated XPC), DDB1 and GAPDH in HaCaT cells transfected with shCon or shE-cadherin at 0, 0.5, 1.5, 6 and 24 h post-UVB (20 mJ/cm^2^). The results were obtained from three independent experiments. (B-C) Immunofluorescence assay of the subnuclear colocalization of CPD and XPC (B) and CPD and DDB2 (C) in HaCaT cells transfected with shCon or shE-cadherin at 0.5 h post-UVC (10 mJ/cm^2^) through a 5 \u03bcm micropore filter. Scale bar, 10 \u03bcm. (D) The relative intensity of XPC and DDB2 focus was calculated by analyzing 100 foci and normalized to that of CPD (n= 100, bar: SD). (E) Immunoblot analysis of E-cadherin, XPC, DDB1 and GAPDH in HaCaT cells transfected with shCon, shE-cadherin, or the combination of shE-cadherin with XPC plasmids at 0, 1.5, 6 and 24 h post-UVB (20 mJ/cm^2^). The results were obtained from three independent experiments. XPC protein levels in A and E were quantified using ImageJ software (below each band in arbitrary units). (F-G) Slot blot analysis of 6-4PP (F) and CPD (G) in HaCaT cells transfected with shCon, shE-cadherin, or the combination of shE-cadherin with XPC plasmids at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^) for 6-4PP and 0, 6 and 24 h post-UVB (20 mJ/cm^2^) for CPD. (H and I) Quantification of percentage (%) of 6-4PP repair from F and CPD repair from G. \\*, *P* \\< 0.05, compared with shCon group; \\#, *P* \\< 0.05, compared with shE-cadherin together with shCon groups, Student's *t*-test and two-way ANOVA.](nihms-719076-f0002){#F2}\n\n![E-cadherin regulates CPD repair through increasing the levels of both XPC and DDB1\\\n(A) Immunoblot analysis of E-cadherin, XPC, Ub-XPC, DDB1 and GAPDH in HaCaT cells transfected with shCon, shE-cadherin, or the combination of shE-cadherin with vector (Con), XPC, DDB1, or the combination of XPC and DDB1 plasmids at 0, 1.5, 6 and 24 h post-UVB (20 mJ/cm^2^). The results were obtained from three independent experiments. XPC protein levels in A were quantified using ImageJ software (below each band in arbitrary units). (B) Slot blot analysis of 6-4PP in HaCaT cells transfected with shCon, shE-cadherin, or the combination of shE-cadherin with vector (Con), XPC, or DDB1 plasmids at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^). (C) Quantification of percentage (%) of 6-4PP repair from B. \\*, *P* \\< 0.05, compared with shCon group; \\#, *P* \\< 0.05, compared with shE-cadherin/Con groups. (D) Slot blot analysis of CPD in HaCaT cells treated as in B at 0, 6, or 24 h post-UVB. (E) Quantification of percentage (%) of CPD repair from D. \\*, *P* \\< 0.05, compared with shCon group; \\#, *P* \\< 0.05, compared with shE-cadherin/Con groups. (F) Immunofluorescence assay of the colocalization of XPC and subnuclear CPD in HaCaT cells transfected as in B at 0.5 h post-UVC (10 mJ/cm^2^) through a 5 \u03bcm micropore filter. Scale bar, 10 \u03bcm. (G) The relative intensity of XPC focus was calculated by analyzing 100 foci and normalized to that of CPD (n= 100, error bar: SD).](nihms-719076-f0003){#F3}\n\n![The TGF-\u03b2 pathway is required for suppression of nucleotide excision repair by E-cadherin loss\\\n(A) Luciferase reporter assay of the SBE4 luciferase reporter in HaCaT cells transfected with shCon or shE-cadherin and then treated with normal IgG or anti-TGF-\u03b2 antibody for 24 h. (mean\u00b1S.D. (error bars), n=3; \\*, p\\<0.05, compared with the shCon group; \\#, p\\<0.05, compared with the shE-cadherin group). (B) Luciferase reporter assay of the SBE4 promoter in HaCaT cells transfected with shCon or shE-cadherin and then treated with vehicle or the TGF-\u03b2 pathway inhibitor LY364947 (2 \u03bcM) for 24 h. (C) Real time RT-PCR analysis of TGF-\u03b21 mRNA levels in HaCaT cells transfected with shCon or shE-cadherin. (D) Immunoblot analysis of E-cadherin, XPC, Ub-XPC, DDB1, and GAPDH in HaCaT cells transfected with shCon or shE-cadherin and then treated with vehicle or LY364947 (2 \u03bcM) for 24 h and then collected at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^). The results were obtained from three independent experiments. (E-F) Quantification of repair percentage (%) of CPD (E) and 6-4PP (F) in HaCaT cells transfected with shCon or shE-cadherin and then treated with vehicle or TGF-\u03b2 inhibitor LY364947 (2 \u03bcM) for 24 h and collected at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^) for 6-4PP and 0, 6 and 24 h post-UVB (20 mJ/cm^2^) for CPD. \\*, *P* \\< 0.05, compared with shCon group; \\#, *P* \\< 0.05, compared with shE-cadherin/Veh groups. (G) Immunoblot analysis of E-cadherin, XPC, Ub-XPC, DDB1, and GAPDH in HaCaT cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 48 h and then collected at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^). The results were obtained from three independent experiments. (H) Immunoblot analysis of E-cadherin, XPC, Ub-XPC, DDB1, and GAPDH in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 48 h and then collected at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^). Protein levels in D, G and H were quantified using ImageJ software (below each band in arbitrary units). (I-J) Quantification of repair percentage (%) of CPD and 6-4PP in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 48 h and then collected at 0, 1.5 and 6 h post-UVB (20 mJ/cm^2^) for 6-4PP and 0, 6 and 24 h post-UVB (20 mJ/cm^2^) for CPD. \\*, *P* \\< 0.05, compared with Con groups.](nihms-719076-f0004){#F4}\n\n![TGF-\u03b2 pathway activation inhibits both XPC and DDB1 transcription\\\n(A) Real time RT-PCR analysis of XPC and DDB1 mRNA levels in HaCaT cells transfected with shCon or shE-cadherin and then treated with vehicle or LY364947 (2 \u03bcM) for 24 h. (B) Luciferase reporter assay of the XPC promoter (\u22121187-27) and DDB1 promoter (\u22121138-124) in HaCaT cells transfected with shCon or shE-cadherin and then treated with vehicle or LY364947 (2 \u03bcM) for 24 h. (C) Luciferase reporter assay of the XPC promoter (\u2212313-27) and DDB1 promoter (\u2212467-124) in HaCaT cells transfected with shCon or shE-cadherin and treated with vehicle or LY364947 (2 \u03bcM) for 24 h. (D) Luciferase reporter assay of the XPC (\u2212313-27) and DDB1 promoter (\u2212467-124) in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (E) Schematic representation of the deletion sites of both XPC and DDB1 promoters and the E2F site of human XPC promoter and E2F/SP1-1/SBE-3 site of human DDB1 promoter. Red nucleotides indicate mutations made in human XPC and DDB1 promoters. (F) Luciferase reporter assay of the XPC promoter with wild-type sequence (\u22121187-27), or deletion of the CREB2, SP1-1/SP1-2, CREB1, E2F or NKX2.5/SBE element in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (G) Luciferase reporter assay of the DDB1 promoter with wild-type sequence (\u22121138-124), or deletion of the N-Myc, NF1, SP1-2, E2F/SP1-1(TIE)/SBE-3 and SBE-1/SBE-2 site in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (H) Luciferase reporter assay of the XPC promoter with an intact (\u22121187-27) XPC or E2F site-mutated XPC in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (I) Luciferase reporter assay of the DDB1 promoter with an intact (\u22121138-124) DDB1 or mutation of E2F site, SP1-1(TIE) site or SBE-3 site DDB1 in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h.](nihms-719076-f0005){#F5}\n\n![The TGF-\u03b2 pathway promotes binding of the E2F4/P130 complex to the XPC promoter and the E2F4/Smad3/p107 complex to the DDB1 promoter\\\n(A) Chromatin immunoprecipitation (ChIP) was performed in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 6 h using the indicated antibodies. qPCR was performed with primers specific for the XPC promoter region harboring the E2F site (upper panel) and for the human GAPDH promoter region as negative and positive control (Lower panel; AcH3, acetylated histone H3). (B) Chromatin immunoprecipitations were performed in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 6 h using the indicated antibodies. qPCR was performed with primers specific for the DDB1 promoter region upstream of the E2F/SP1-1/SBE-3 binding site. (C) Immunoprecipitation was performed in NHEK cells treated with vehicle or TGF-\u03b21 (10 ng/ml) for 6 h with the indicated antibodies, followed by immunoblot analysis of p107, p130, Smad3, and E2F4. The results were obtained from three independent experiments. (D and E) Luciferase reporter assay of the XPC (\u22121187-27) and DDB1 promoter (\u22121138-124) in NHEK cells transfected with siCon or siRNA target p130 (sip130) and then treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (F and G) Luciferase reporter assay of the DDB1 (\u22121138-124) and XPC promoter (\u22121187-27) in NHEK cells transfected with siCon or siRNA target p107 (sip107) and then treated with vehicle or TGF-\u03b21 (10 ng/ml) for 24 h. (H) Immunoblot analysis of p107, p130, XPC, DDB1, and GAPDH in HaCaT cells transfected with shCon or shE-cadherin and then transfected with siCon, sip107 or sip130. Protein levels in C and H were quantified using ImageJ software (below each band in arbitrary units). The results were obtained from three independent experiments.](nihms-719076-f0006){#F6}\n\n![Schematic diagram of E-cadherin regulation of nucleotide excision repair (NER)](nihms-719076-f0007){#F7}\n"}
+{"text": "Introduction {#s1}\n============\n\nIschemic stroke, caused by the formation of a clot in a main cerebral blood vessel, results in a sharp drop in perfusion of the ischemic core and is a leading cause of death and disability worldwide. An ischemic stroke deprives the brain of oxygen and nutrients, leading to permanent necrotic neuronal death in the region of brain tissue supplied by the affected cerebral artery (Singh et al., [@B76]). The pathophysiology of the complex brain injury that occurs following an ischemic stroke has not yet been fully elucidated. Cerebral ischemia--reperfusion causes a loss of cellular ion homeostasis, activation of caspases, generation of reactive oxygen species (ROS), bioenergetic failure, impaired mitochondrial function, and excitotoxicity in brain tissue, but these complex reactions are not yet fully understood. Activation of complementary pathways promotes the generation of arachidonic acid products and cytokines, infiltration of immune cells, and disruption of the blood--brain barrier (BBB), which initiates inflammatory cascades (Fann et al., [@B18]; Ballarin and Tymianski, [@B6]; Bao et al., [@B8]).\n\nStroke-induced inflammation and the activation of proinflammatory mediators have been the focus of recent research into the mechanisms of stroke-induced brain damage (Chen et al., [@B12]; Stonesifer et al., [@B78]), and high-mobility group box 1 protein (HMGB1), a typical damage-associated protein, has gained particular interest (Choi et al., [@B13]; Shichita et al., [@B74]). The HMGB proteins were first identified by Goodwin et al. ([@B22]). Since then, a number of studies have confirmed that HMGB1 plays a central role in the pathogenesis of many diseases, including systemic lupus erythematosus (SLE), acute liver failure (ALF), tumors, and cerebrovascular diseases (Cully, [@B14]; Majumdar et al., [@B54]; Jian et al., [@B32]; Xiong et al., [@B98]; Seidu et al., [@B73]; Hossain et al., [@B29]). The HMGB1 protein was originally reported to be a ubiquitous, non-histone chromosomal protein that plays a role in DNA replication and repair in eukaryotic cells (Liu et al., [@B48]; Thomas and Stott, [@B85]). When released or secreted, HMGB1 acts as a sentinel for the immune system and triggers cell survival or death pathways in response to stress or damage. It is a multifunctional protein, with the functions depending on its location in the cell. In normal brain tissue, HMGB1 is usually located in the nuclei. However, following a stroke, HMGB1 is translocated to the cytosol and secreted into the extracellular space. Studies have indicated that intracellular HMGB1 plays an important role in the regulation of energy homeostasis and transcription (Tang et al., [@B83]). In contrast, it has been reported that extracellular HMGB1 directs BBB breakdown, neuroimmune activities, and neuronal death (Gardella et al., [@B20]; Faraco et al., [@B19]).\n\nIn this review article, we aim to illustrate the biological functions of HMGB1 and the role of this protein in ischemic stroke, as well as fully clarify the mechanisms of HMGB1's role in stroke pathology, to highlight HMGB1 and pathways which may be potential drug targets in an attempt to provide new prospects and directions for the treatment of ischemic stroke.\n\nStructure and Characteristics of HMGB1 {#s2}\n======================================\n\nThe HMGB protein family comprises four proteins: HMGB1, HMGB2, HMGB3, and HMGB4. All have HMG box domains, which are DNA-binding motifs (Stros, [@B79]). HMGB1 protein is highly conserved in evolution as a chromatin-binding molecule, and since its discovery in 1973, it has attracted the attention of researchers (Goodwin et al., [@B22]).\n\nHMGB1 has a molecular weight of 30 kD and consists of 214 amino acid residues. Its amino acid sequence is highly conserved, with over 98% homology between humans and rodents. HMGB1 consists of three distinct structural domains: two positively charged DNA-binding motifs (boxes A and B) and a highly negatively charged C-terminal acidic tail ([Figure 1](#F1){ref-type=\"fig\"}). The A box is located at the N-terminal, and the B box is between the A box and the C tail. The A and B boxes are evolutionarily conserved and are each composed of three \u03b1-helical structures, which can nonspecifically bind to DNA. The A box is important for anti-inflammatory action (Lotze and Tracey, [@B53]), while the B box is critical for proinflammatory activity and cell differentiation (Sparatore et al., [@B77]). However, the most important part of HMGB1 is the C-terminal domain, which is associated with regulating HMGB1's DNA binding affinity.\n\n![Structure of high-mobility group box 1 protein (HMGB1; 30 kD, 214 amino acids). HMGB1 is divided into three distinct structural domains: A box, B box, and C tail. The three regions have their respective positions. The three cysteines (C23, C45, and C106) in the molecular structure of HMGB1 contribute to its redox state.](fncel-13-00127-g0001){#F1}\n\nThere are three cysteines (C23, C45, and C106) in the molecular structure of HMGB1 (Yang et al., [@B104]), and the redox state of these three cysteines regulates biological activity and receptor binding. HMGB1 with all-thiol state of the three cysteine residues has been reported to cooperate with C-X-C motif chemokine 12 (CXCL12) to form a heterocomplex with chemotactic activity, which binds to the CXCL12 reciprocal receptor C-X-C chemokine receptor type 4 (CXCR4) in a synergistic fashion, contributing to the cytokine-inducing and chemoattractant activities of HMGB1 (Schiraldi et al., [@B70]; Venereau et al., [@B91]). However, it is reported that all-thiol HMGB1 is oxidized, leading to the formation of disulfide state at the three cysteine residues, when it is released into the circulation after cerebral ischemia. Oxidized HMGB1 possesses a cytokine-stimulating function, inducing the translocation nuclear factor \"kappa-light-chain-enhancer\" of activated B-cells (NF-\u03baB) to the nucleus and synthesis of tumor necrosis factor alpha (TNF-\u03b1) in activated macrophages (Yang et al., [@B106]; Singh et al., [@B76]).\n\nThe Biological Functions of HMGB1 {#s3}\n=================================\n\nUnder physiological conditions, as a non-histone chromosome binding protein, HMGB1 remains in the nucleus and nonspecifically binds to DNA, stabilizes nucleosomes, and assists in DNA replication and transcription (Liu et al., [@B48]; Thomas and Stott, [@B85]; Tang et al., [@B83]). In addition, HMGB1 may also interact with nucleotides to repair related proteins, play a role in DNA repair, and identify damaged DNA fragments and remove them. During the maturation of T and B lymphocytes, HMGB1 can affect the recombination, differentiation, and development of V(D)J gene fragments. Cell activation, injury, or death caused by some pathological conditions, such as hypoxia, result in HMGB1 translocation from nucleus to cytoplasm or extracellular space, due to the separation of HMGB1 with damaged DNA. The translocation of HMGB1 between the nucleus and the cytoplasm is associated with the acetylation of lysine in nuclear localization sites (NLSs; Andersson et al., [@B5]).\n\nThere are two ways for HMGB1 to be released: passive release or active secretion. During disease development, these two mechanisms are not completely independent but are mutually causal. It has been confirmed that the secretion of HMGB1 from necrotic cells functions as danger-associated molecular patterns (DAMPs), and contributes to the inflammatory cascade (Scaffidi et al., [@B69]; Singh et al., [@B76]). Okuma et al. ([@B59]) reported that HMBG1 is secreted within a few hours of stroke onset and is a hyperacute DAMP that devastates the BBB. When HMGB1 is released into the extracellular space, it is recognized by receptors such as toll-like receptors 2 and 4 (TLR2, TLR4; Park et al., [@B61]); and receptor for advanced glycation end product (RAGE); (Rauvala and Rouhiainen, [@B67]), thereby activating the NF-\u03baB signaling pathway and contributing to the inflammatory response (Lok et al., [@B50]) ([Figure 2](#F2){ref-type=\"fig\"}). Inhibiting the expression and translocation of HMGB1 and its receptors has been demonstrated to have anti-inflammatory and neuroprotective effects on ischemic injury (Tao et al., [@B84]).\n\n![Pathways of HMGB1 secretion. There are two mechanisms used by cells to liberate HMGB1 into the extracellular milieu. Somatic cells contain large amounts of HMGB1 that is passively released into the extracellular space during cell apoptosis or necrosis. A second mechanism is the active secretion of HMGB1 from activated immune or nerve cells.](fncel-13-00127-g0002){#F2}\n\nReceptors and Transduction Pathways of HMGB1 {#s4}\n============================================\n\nAlthough multiple receptors have been reported for HMGB1, only four receptors have been identified: TLR2, TLR4, RAGE, and CXCR4, as well as matrix metalloproteinases (MMPs; Hori et al., [@B28]; Carty and Bowie, [@B10]). Other receptors are likely to present molecules to bind HMGB1. The binding of HMGB1 to these receptors mediates inflammatory factors production and eventually results in systemic inflammation.\n\nThe transmembrane protein RAGE is a member of the immunoglobulin superfamily and is widely expressed in the cell surface of many cells including mononuclear macrophages, epithelial cells, endothelial cells, and nerve cells (Kokkola et al., [@B40]). It binds to a variety of active proteins, including glycated proteins, cytoplasmic protein S100, amyloid \u03b2-peptide, and HMGB1. Under physiological conditions, cells have low RAGE expression, but when its ligand molecules increase, its expression also increases. RAGE has a high affinity for HMGB1. When HMGB1 binds to the upregulated RAGE, some protein kinases in mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3 kinase/protein kinase B (PI3K/Akt) pathways, such as p38 kinase, SAPK/JNK, extracellular regulated protein kinases1/2 (ERK 1/2) and Akt are activated by this mitogen (Qin et al., [@B64]; Tong et al., [@B87]); in addition, cell division cycle 42 (Cdc42), Ras-related C3 botulinum toxin substrate (Rac), and just another kinase/signal transducer and activator of transcription 1 (JAK/STAT1)-mediated signal transduction pathways are also activated (Tsoyi et al., [@B88]). These events finally lead to the translocation of NF-\u03baB, inducing the expression of inflammatory cytokines and chemokines that participate in the maturation and migration of immune cells, the expression of surface receptors and growth of neurites, as well as tumor proliferation (Muhammad et al., [@B56]; Akirav et al., [@B4]; Zhang et al., [@B110]).\n\nTLR2 and TLR4 are also HMGB1 receptors that induce a proinflammatory response (Yang et al., [@B105], [@B107]). TLR-mediated signaling is important for cytokine release and activation of innate immunity. Kang and Lee demonstrated that HMGB1 binding to TLRs can mediate activation of certain pathways, including myeloid differentiation factor 88 (MyD88)-dependent and MyD88-independent pathways, which result in the expression of inflammatory genes, leading to the production of cytokines and chemokines (Kang and Lee, [@B34]). Recently, Kim E. J. et al. ([@B35]) reported that binding of HMGB1 to TLR-4 increases interleukin-1\u03b2 (IL-1\u03b2) production in vascular smooth muscle cells *via* the activation of Nod-like receptor protein 3 (NLRP3) inflammasome.\n\nCXCR4 (a CXCL12 reciprocal receptor) is another receptor that is reported to bind HMGB1. A CXCL12/HMGB1 heterocomplex can interact with CXCR4 receptors and induce migration of inflammatory cells. Recently, Tirone et al. ([@B86]) demonstrated that fully reduced HMGB1 (fr-HMGB1) orchestrates tissue regeneration *via* CXCR4.\n\nHMGB1 and Cerebral Ischemic Stroke {#s5}\n==================================\n\nAs the global population ages, cerebral ischemic stroke and its complications have become the main cause of disability and death worldwide (Neumann et al., [@B58]). During ischemic stroke, cerebral artery occlusion leads to oxygen and nutrient depletion in neural tissue (Lo, [@B49]). In response to ischemic injury, astrocytes and microglia in the brain activate and release reactive nitrogen species (RNS), ROS, and proinflammatory cytokines that cause secondary damage to the infarct area. Studies have shown that HMGB1 is involved in the pathogenesis of ischemic stroke and reperfusion injury (Kim et al., [@B38], [@B37]; Liu et al., [@B47]; Muhammad et al., [@B56]; Qiu et al., [@B65]; Shichita et al., [@B75]). Furthermore, the application of anti-HMGB1 neutralizing antibodies has been shown to reduce infarct volume and ameliorate infarction after middle cerebral artery occlusion (MCAO) in rats (Kim et al., [@B37]). Recently, it has been reported that HMGB1 works as an immune system signal (or DAMP), and that HMGB1 inhibition has a protective effect against damage following an ischemic stroke. There are two phases of the post-ischemic stroke inflammatory response: the early phase, which is involved in neural tissue destruction, and the late phase, which consists of tissue remodeling (Agnello et al., [@B2]). In the early phase, HMGB1 is released from nerve cells to accelerate the inflammatory response. However, in the late phase, HMGB1 release from reactive astrocytes may affect the regeneration of nerves and blood vessels, and promote tissue remodeling (Bianchi et al., [@B9]).\n\nProcess of Activation and Secretion of HMGB1 {#s5-1}\n--------------------------------------------\n\nIn the early stages of stroke, a large number of neurons undergo sustained hypoxia and oxidative toxicity. The cell membranes of neurons are destroyed, \"holes\" appear, and HMGB1, loosely bound to chromosomes, is passively released into the extracellular space (Tsung et al., [@B89]). Meanwhile, microglia and astrocytes are activated and intracellular HMGB1 is modified by a series of acetylation and phosphorylation reactions, thus decreasing its affinity for DNA, finally leading to active secretion of HMGB1 into extracellular space. This is accompanied by increases in RNS and ROS levels. These factors all contribute to the formation of highly oxidative conditions. The activity and function of many proinflammatory cytokines are adjusted by the oxidation of methionine, cysteine, and tyrosine residues (Singh et al., [@B76]). HMGB1 is also modified to the restored or oxidized forms, which have specific cellular functions in the ischemic brain (Zhang J. et al., [@B111]; Lorenzen et al., [@B52]). Once HMGB1 is released, an inflammation signal is emitted, and the immune system responds with a proportional positive feedback amplification.\n\nIn a model of MCAO in mice, Kim et al. ([@B37]) reported that, through phosphorylation and acetylation, HMGB1 is translocated into the cytoplasm and then secreted into the extracellular area. Moreover, using *in vitro* experiments, Hua et al. ([@B30]) demonstrated that when primary cultures of neurons undergo oxygen glucose deprivation (OGD), HMGB1 is detected in the liquid supernatant. Fully reduced HMGB1 has been shown to be prevalent in the serum and brain samples of mice after 2 h of stroke (Liesz et al., [@B46]). Furthermore, cytokine-inducing disulfide HMGB1 was discovered in mouse serum after 24 h of cerebral ischemia (Laird et al., [@B41]). These studies demonstrate that early necrotic brain tissue releases fully reduced HMGB1 to the blood stream, where it is oxidized to its cytokine-inducing form.\n\nDynamics of HMGB1 in Ischemic Stroke {#s5-2}\n------------------------------------\n\nWhen an ischemic stroke occurs, HMGB1 is released into peripheral blood from the central nervous system (CNS). This extracellular HMGB1 has been demonstrated to provoke an inflammatory response in many experimental animal models (Yang et al., [@B108]). The HMGB1 levels in serum and plasma reflect the expression level of HMGB1 in the CNS and the extent of brain injury. In experimental animal models of cerebral ischemia--reperfusion injury (IRI) and in ischemic stroke patients, HMGB1 levels in the cerebral spinal fluid (CSF) and serum are significantly increased. Goldstein et al. ([@B21]) reported that HMGB1 levels in ischemic stroke patients rapidly increase, and are up to 13 times higher than those in the control group within 24 h. Kim et al. ([@B37]) reported that CSF and serum HMGB1 levels increase rapidly after 3 h of ischemic stroke, and generate 2 peaks: one on the 1st day and the other on the 6th and 7th days after stroke. The authors suggested that the HMGB1 peak on the 1st day may be caused by acute necrosis of nerve cells induced by excitotoxicity and that HMGB1 activates microglia in the early stages of inflammation. The second peak of HMGB1 after ischemic stroke, however, is thought to be secreted by various immune cells such as microglia, macrophages, astrocytes, and vascular endothelial cells. More recently, Umahara et al. ([@B90]) also observed that, at the acute stage of cerebral infarction in patients, HMGB1 is located in the neuronal cytoplasm, while during the late stage of cerebral infarction, HMGB1 is mainly secreted by macrophages located in the basal ganglia and in some ischemic regions. Xiong et al. ([@B99], [@B98]) confirmed that HMGB1 transposes from the neuronal cell nucleus to the cytoplasm, and finally to the extracellular environment after ischemic stroke in the MCAO rats and mice. They found that HMGB1 is comprehensively expressed in the nuclei of neurons in the control group and significantly reduced after MCAO, and its subcellular translocation is observed at an early stage (12 h) of ischemic stroke (Xiong et al., [@B99]). At the same time, increased numbers of HMGB1 positive microglia/macrophages are observed infiltrating the stroke area and exacerbating inflammation (Xiong et al., [@B99]). It was suggested that at the early stage of stroke, HMGB1 is first passively released from the dying neurons, accompanied by active secretion by the actively infiltrated microglia/macrophages (Xiong et al., [@B99]). Meanwhile, they found that HMGB1 levels are increased in the CSF and circulation at 5, 24, and 48 h of reperfusion (Gu et al., [@B24]; Xiong et al., [@B99], [@B98]). This phenomenon may be related to the disruption of the BBB and an increase of vascular permeability, or to other unknown causes. HMGB1 is reported to be highly expressed in the blood not only during the acute phase but also for at least 2 weeks following ischemia in rats (Kim et al., [@B38]), even longer to around 1 month in ischemic stroke patients (Schulze et al., [@B72]).\n\nContribution of HMGB1 in Early Cerebral Ischemic Stroke {#s5-3}\n-------------------------------------------------------\n\nEarly restoration of reperfusion is a fundamental step in the prevention of decreased brain perfusion in cerebral ischemia patients. However, while ischemia as a stimulus may activate macrophages to release the inflammatory mediator HMGB1, reperfusion may exacerbate the inflammatory response by stimulating the release of more HMGB1 into the extracellular environment and aggravating brain tissue damage. Kim et al. ([@B37]) constructed an HMGB1 short hairpin RNA (shRNA) plasmid and injected it into MCAO mice to interfere with HMGB1 expression. This treatment reduces the size of cerebral infarcts in mice (Kim et al., [@B37]). Meanwhile, they also found that the activation and infiltration of microglia in the ischemic region is reduced with this treatment, and the expression of TNF-\u03b1, IL-1\u03b2, cyclo-oxygenase-2 (COX-2), and inducible nitric oxide synthase (iNOS) are decreased (Kim et al., [@B37]). The anti-HMGB1 antibody also significantly reduces the size of the cerebral infarct, improves the permeability of the BBB, inhibits the activation of microglia, and reduces the expression of TNF-\u03b1 and iNOS in the MCAO mouse model (Liu et al., [@B47]; Muhammad et al., [@B56]). Conversely, the infarct volume is increased and the extent of inflammation response is aggravated when recombinant HMGB1 is injected into mice (Goldstein et al., [@B21]). After treatment with glycyrrhizin, a direct inhibitor of HMGB1, the level of HMGB1 in neuronal cells is significantly increased, while translocation and release of HMGB1 are inhibited, neuronal death in the infarct areas is significantly reduced, as is infarct volume. This was accompanied by a reduction in the activation and infiltration of inflammatory cells including microglia/macrophages, neutrophils and T lymphocytes, as well as the production of proinflammatory cytokine TNF-\u03b1, IL-1\u03b2 and interferon-\u03b3 (IFN-\u03b3; Kim et al., [@B39]; Gu et al., [@B24]; Xiong et al., [@B98]). HMGB1 is also expressed in microglia, and the extent of cerebral infarct can be significantly reduced by the inhibition of microglial HMGB1 expression (Hayakawa et al., [@B26]). Recently, Balosso et al. ([@B7]) reported that different redox forms of HMGB1 can induce differential activation patterns of microglia and that disulfide HMGB1 may promote neuronal cell death induced by N-methyl-D-aspartic (NMDA) acid receptor through TLR-4 receptors.\n\nExtracellular HMGB1 acts as a proinflammatory factor and activates microglia and macrophages to amplify inflammatory responses by recognizing TLRs or RAGE receptors. The RAGE receptor is involved in cerebral ischemic injury caused by HMGB1, and the level of soluble RAGE (sRAGE) in the serum of patients with acute stroke is significantly increased. The expression of RAGE is also increased in the brain tissue of patients with unilateral cerebral infarction. Liesz et al. ([@B46]) reported that HMGB1 is released at the acute stage of ischemia from the injured brain in both the mice model and patients, and HMGB1-RAGE signaling participates in the ischemic stroke, which was believed to be critical for clarifying the mechanism of the brain--immune interaction after ischemia. The TLRs are also implicated in the process of ischemic injury. In experiments using mice, Zhang et al. ([@B112]) observed that inhibiting the binding of HMGB1 to TLR4 downregulates IL-17A levels, thereby inhibiting neuronal apoptosis, improving nerve repair and reducing infarct volume. By blocking the TLR4 receptor, the extent of brain tissue edema, infarct size, and increased neurological damage scores after stroke are reduced *in vivo*, which possibly occurs *via* the TLR4/MyD88 signal pathway. Therefore, HMGB1 participates in the destruction of the BBB after stroke and leads to inflammation in brain tissue.\n\nThe Role of HMGB1 in the Advanced Stages of Ischemic Stroke {#s5-4}\n-----------------------------------------------------------\n\nDuring the advanced stages of ischemic stroke, the function of HMGB1 is not fully understood. At present, it is generally thought that HMGB1 can promote the regeneration of nerve cells, remodeling of blood vessels and recovery of neurological function in the late infarct stage (Le et al., [@B42]). Stroke-activated astrocytes increase the viability and migration of endogenous endothelial progenitor cells (EPCs) by releasing HMGB1, and promote neurovascular repair after stroke, while inhibition of HMGB1 by an siRNA restrains the EPC proliferation, blocks the peri-infarct angiogenesis, and increases neurological scores. Similarly, exogenous EPC transplantation promotes the regeneration of blood vessels in the ischemic region during the chronic phase of stroke, reduces the volume of brain atrophy, and improves neurological function, mainly through the involvement of HMGB1-RAGE initiated MAPK kinase/extracellular signal-regulated kinase (MEK/ERK) pathway (Hayakawa et al., [@B27]).\n\nMoreover, another important role of HMGB1 after stroke is to promote endothelial cell sprouting. When different concentrations of HMGB1 are used to treat human umbilical vein endothelial cells, the degree of endothelial cell sprouting is directly proportional to the HMGB1 concentration. Schlueter et al. ([@B71]) reported that high concentrations of HMGB1 have proinflammatory effects that cause endothelial damage, while low concentrations of HMGB1 improves EPCs activity, then promoting neurovascular growth. Thus, HMGB1 released from ischemic brain mediates post-stroke angiogenesis at the advanced stage, subsequently promoting brain repair and disease recovery.\n\nThe Mechanisms of HMGB1 Participating in Ischemic Stroke {#s5-5}\n--------------------------------------------------------\n\nAt present, although a pivotal role of HMGB1 in cerebral ischemia is widely accepted, the specific mechanisms are not yet fully understood. However, possible mechanisms will be discussed in the following paragraphs.\n\n### Regulation of Inflammation by HMGB1 {#s5-5-1}\n\nAs mentioned previously, post-ischemic inflammation appears to be a critical component of the progression of pathogenic stroke. Recent studies agree that HMGB1 is a recognized proinflammatory factor in ischemic stroke and positively correlates with stroke severity in animal models and patients (Harris et al., [@B25]; Le et al., [@B42]). HMGB1, as an endogenous inflammatory mediator, is passively released by necrotic cells or actively secreted by macrophages/monocytes into the ischemic core, triggering and amplifying inflammatory processes. In turn, the released HMGB1 also induces the activation of microglia, macrophages, and endothelial cells among others (Wang C. et al., [@B92]), which results in the production of proinflammatory mediators such as TNF-\u03b1, iNOS, IFN-\u03b3, NO, chemokines, and cell adhesion molecules. On one hand, these proinflammatory mediators recruit more immune cells from the circulatory system into the CNS, thus aggravating the development of inflammation in the brain (Young et al., [@B109]). On the other hand, these proinflammatory mediators also stimulate microglia and macrophages to actively secrete HMGB1 in a positive feedback loop, which exerts an effect as late inflammatory mediators. Xiong et al. ([@B98]) demonstrated that microglia/macrophages express HMGB1 within the ischemic core. However, when HMGB1 levels are reduced, infiltration of microglia/macrophages and leukocytes in ischemic brain tissue is significantly inhibited (Gu et al., [@B24]). Several studies have demonstrated 2 peaks of HMGB1 levels in CSF and serum post-MCAO (Abraham et al., [@B1]; Kim et al., [@B38]; Kim I. D. et al., [@B36]; Umahara et al., [@B90]), which is likely to be related to two separate sources of circulating HMGB1. The first peak of HMGB1 may originate from necrotic neurons and activated microglia/macrophages of the CNS, and the other may be derived from active secretion by delayed activated inflammatory cells in the postischemic hemisphere or peripheral immune cells (Le et al., [@B42]). This suggests that accumulated extracellular HMGB1 may not only mediate acute damaging processes in the brain but also aggravate inflammation in the brain and increase vulnerability to post-stroke infection (PSI) Recently, a study reported that HMGB1 can exacerbate inflammatory damage to the BBB during the process of brain ischemia--reperfusion, which may be the cause of the release of HMGB1 from the brain to the CSF and circulation (Li M. et al., [@B44]).\n\nHMGB1 is \"sticky\" and binds to a variety of different molecules including RACE, TLRs and CXCR4 on the cell surface. This \"sticky\" may partially explain the limitation of the diffusion of extracellular HMGB1, then localizing the damaging effects of HMGB1. After ischemic stroke, extracellular HMGB1 works as a DAMP, through three signaling pathways ([Figure 3](#F3){ref-type=\"fig\"}): (1) by binding to TLR4, HMGB1 induces MyD88 or Toll/IL-1 receptor domain-containing adaptor-inducing IFN-\u03b2 (TRIF) signaling cascades, leading to the activation of transcription factors, such as NF-\u03baB and activator protein-1 (AP-1) *via* TNF receptor-associated factor 6 (TRAF6)-mediated JNK, p38 MAPK, and ERK signaling activation (O'Neill and Bowie, [@B60]); (2) by interacting with RAGE, PI3K/Akt and MAPK pathways are activated, resulting in nuclear NF-\u03baB translocation and production of inflammatory mediators including TNF-\u03b1 and IL-1\u03b2 (Zhang et al., [@B110]). PI3K/Akt pathway is usually considered to be an antiapoptotic pathway, but there is also evidence to suggest that it can facilitate the production of inflammatory cytokines (Xue et al., [@B101]; Li H. et al., [@B43]; Xu et al., [@B100]). Some studies have reported that when HMGB1 binds to RAGE, it also generates direct intracellular signaling resulting in nuclear NF-\u03baB translocation (Li et al., [@B45]). However, other studies have suggested that TLR4 is required for HMGB1-induced NF-\u03baB activation and cytokine formation *via* the RAGE receptor (Yang et al., [@B105]) and (3) when bound to CXCR4, the ERK, cyclooxygenase2/janus kinase/signal transducer and activator of transcription (COX2/JAK/STAT), and PI3K/Akt signaling pathways are activated (Schiraldi et al., [@B70]; Yamamoto and Tajima, [@B102]; Cecchinato et al., [@B11]), contributing to inflammatory cell migration and inflammatory mediator production.\n\n![Potential mechanisms by which HMGB1 contributes to stroke pathogenesis. HMGB1 acts as an early mediator at the initial stages of stroke. An ischemia causes nerve cell injury, which leads to the passive release of the disulfide form of hypoacetylated HMGB1 from damaged cells. Extracellular HMGB1, which is either acetylated or oxidized at residues 23 and 45 to its the disulfide form, stimulates inflammatory signaling by binding to cell-surface receptors TRL4, receptor for advanced glycation end product (RAGE), and CXCR4 on microglia. Acetylated or disulfide HMGB1, together with cytokines and danger-associated molecular patterns (DAMPs), is also actively released from activated immune cells, causing additional nerve cell damage and microglia activation *via* positive feedback, thereby acting as a late inflammatory mediator.](fncel-13-00127-g0003){#F3}\n\n### Excitotoxic Injury Caused by HMGB1 Release {#s5-5-2}\n\nHMGB1 can induce excitatory neurotransmitter release in the brain after stroke. Studies have shown that HMGB1 inhibits mouse neural glial glutamate transporters by glutamate/aspartate transporter (GLAST) neural activation particles and increases extracellular levels of glutamate and its receptor. The activation of glutamate receptors causes Ca^2+^ influx, eventually leading to Ca^2+^ overload and a loss of cell function due to dyshomeostasis (Zhang J. et al., [@B111]). An intravenous injection of anti-HMGB1 monoclonal antibody could, therefore, reduce cerebral infarct volume and improve neurological function by preventing the elevation of glutamate levels and reducing excitotoxicity in neurons.\n\n### BBB Damage by HMGB1 Release {#s5-5-3}\n\nHMGB1, as an inflammatory cytokine, also can contribute to BBB breakdown, and BBB permeability is significantly reduced by using of anti-HMGB1 monoclonal antibody in experimental stroke models (Zhang J. et al., [@B111]). Ischemic stroke results in upregulation of MMPs, then increasing BBB permeability and even BBB damage because MMPs can break down a myriad of extracellular matrix (ECM). MMPs are usually confined to the cytosol in their inactivated state and are activated by plasmin or other MMPs. Sapojnikova et al. ([@B68]) demonstrated that HMGB1 levels are strongly correlated with MMP-9 secretion. Qiu et al. ([@B66]) reported that HMGB1 upregulates MMP-9 *via* the TLR4 signaling pathway; when TLR4 signaling is blocked, HMGB1-induced MMP-9 upregulation is mostly suppressed. In addition, Tissue-type plasminogen activator (tPA) is a putative pharmacotherapy for ischemic stroke, but the neurovasculature complications including edema and hemorrhagic transformation, due to BBB damage, also can occur following the using of this drug, along with the reperfusion of blood flow. When using the HMGB1-binding heptamer peptide (HBHP) to inhibit HMGB1 activity, significantly improves the BBB leakage, rescues the loss of occludin, a tight junction protein, and promotes BBB integrity, thereby reducing the complications resulted from tPA treatment (Li M. et al., [@B44]).\n\n### Regulation of Autophagy by HMGB1 {#s5-5-4}\n\nThe release of HMGB1 is also related to the functional state of autophagy in the early stages of stroke. There is some evidence to suggest that ischemic stroke postprocessing forms a two-way feedback mechanism to protect the brain, restraining autophagy and reducing the secretion of HMGB1. However, the relationship between HMGB1 and autophagy needs to be explored. Wang J. et al. ([@B94]) indicated that the expression and location of HMGB1 are closely related to autophagy; remote ischemic preconditioning (RIPerC) and ischemic postconditioning (IPOC) has been shown to inhibit autophagy, blocking the translocation of HMGB1 from the nucleus to the cytoplasm. It is also clear that the inhibition of autophagy reduces the secretion of HMGB1; the reduction in HMGB1 secretion, in turn, causes the inhibition of autophagy (Hayakawa et al., [@B26]). Tang et al. ([@B81]) demonstrated that the different redox states of HMGB1 lead to different cellular effects, which may be the crucial function of HMGB1: to promote information transfer between autophagy and apoptosis. Thiol dehydrogenation forms a disulfide bond (for example, when it is oxidized) between C23 and C45 of the HMGB1 protein A-box, which is required for induction of autophagy because HMGB1 binds to Beclin-1, causing Beclin-1 to separate from Bcl-2 (Tang et al., [@B83]).\n\n### Relationship Between HMGB1 and Mitochondrial Oxidative Stress/ERS {#s5-5-5}\n\nEndoplasmic reticulum (ER) and mitochondria constituted the centers of metabolic networks. After a stroke, mitochondrial oxidative stress results in a large number of ROS production, which can induce ER stress (ERS), finally causing inflammation and cell apoptosis, *via* C/EBP homologous protein (CHOP), Caspase12, JUN activation (L\u00f3pez-Hern\u00e1ndez et al., [@B51]; Poone et al., [@B62]). As mitochondria are quantifiable sources of ROS, and ER is a factory to folding, processing and trafficking of proteins, immune-metabolic dysregulation can occur in these compartments of neurons during cerebral ischemia--reperfusion (Narne et al., [@B57]; Wang et al., [@B93]). In response to ERS and mitochondrial oxidative stress, macrophages and monocytes actively release HMGB1. Tang et al. ([@B82]) reported that oxidative stress plays a potential role in the regulation of HMGB1 release. They illuminated that a ROS, H~2~O~2,~ stimulates HMGB1 release actively from macrophages and monocytes, possibly *via* MAPK- and CRM1-dependent pathways (Tang et al., [@B82]). In addition, a recent study have shown that oxygenized low density lipoprotein (OxLDL), a risk factor of stroke, stimulates HMGB-1 secretion in macrophages resulted from oxidative stress, then HMGB1 can contribute to macrophage-derived foam cells formation *via* ERS/CHOP pathway (Wu et al., [@B97]).\n\n### Relationship Between the Immune System and HMGB1 {#s5-5-6}\n\nHMGB1 an extracellular signal molecule and inflammatory mediator has been demonstrated to participate in the adaptive immune response not only by indirectly acting on dendritic cells (DCs; Yang et al., [@B103]) but also by directly acting on T cells to regulate their function (Dumitriu et al., [@B15]). The effect of HMGB1 on T cells is bidirectional (Sundberg et al., [@B80]). Low dose, short-term *in vitro* HMGB1 stimulation promotes the activation, maturation, and migration of DCs (Dumitriu et al., [@B16]). This HMGB1 stimulation can also act directly on T cells (Zhang Y. et al., [@B114]), stimulating T cell activation and proliferation, and polarization of (T helper cells 1) Th1, resulting in the secretion of IL-2 and INF-\u03b3, as well as inhibiting the expression of cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and forkhead or winged helix transcription factor proteins 3 (Foxp3) in regulatory T cells and reducing the secretion of IL-10 (Wang et al., [@B95]; Zhu et al., [@B115]). In contrast, long-term, high doses of HMGB1 have the opposite effects, directly or indirectly inhibiting the activation and proliferation of DCs and Th1 polarization of T cells (Huang et al., [@B31]), reducing IL-2 secretion, thus blocking T cell function and inducing immunosuppression (Zhang et al., [@B113]; Zhu et al., [@B116]; Wild et al., [@B96]).\n\n#### The Effect of HMGB1 on the Ischemic Brain *via* T Lymphocytes {#s5-5-6-1}\n\nCurrently, activated T lymphocytes are considered to be a key factor in secondary tissue injury, affecting the prognosis of stroke. Xiong et al. ([@B98]) reported that T cells participate in HMGB1 effects on the ischemic brain in both *in vivo* and *in vitro* experiments. The study showed that the HMGB1 inhibitor glycyrrhizin protects against ischemia partly by inhibiting the infiltration of T cells and their subtypes into the ischemic brain. Injection of glycyrrhizin reduces the infarct size in WT mice but not T or B cell-deficient severe combined immune deficiency (SCID) mice, while restoring T and B cells in SCID mice elicit a reduction in infarct size following glycyrrhizin treatment. The *in vitro* experiments demonstrated that glycyrrhizin inhibits neuronal death in a splenocyte and neuron coculture system with splenocytes derived from WT mice, but not from SCID mice. These results suggest a pivotal role for T cells in the detrimental effects of HMGB1 in the brain after ischemic stroke (Xiong et al., [@B98]).\n\n#### Role of HMGB1 in Stroke-Induced Immunodepression {#s5-5-6-2}\n\nIschemic stroke initiates not only a brain inflammation that causing cerebral injury, but also a stroke-induced immunodepression (SIID), resulting in an incidence of PSI, most typically urinary tract infections and pneumonia (Gu et al., [@B23]). PSI is one of the leading causes of delayed death in stroke patients (Prass et al., [@B63]; Emsley and Hopkins, [@B17]). Prass et al. ([@B63]) first reported that the PSI is closely associated with the reduction of lymphocytes resulted from apoptosis in the peripheral organs in 2003, which is simultaneously accompanied by the shift of cytokine component from proinflammatory to anti-inflammatory profile. After that, SIID is commonly accepted, which is characterized by lymphopenia and dysfunction of lymphocytes. It is believed that HMGB1 may be related to lymphopenia and immune-function failure after ischemic stroke. Then, the correlation between levels of HMGB1 and the number of immune cells in the blood after stroke was investigated (Gu et al., [@B24]). When compared with the sham group, the number of total peripheral blood mononuclear cells (PBMCs) is significantly reduced while plasma HMGB1 levels increases, which suggests that HMGB1 may regulate lymphopenia and immunodepression (Gu et al., [@B24]). These findings are supported by a report that HMGB1 release into plasma is reduced after splenectomy, which has been previously shown to reduce infarct size and improve lymphopenia after stroke (Juneja et al., [@B33]; Mili\u0107evi\u0107 et al., [@B55]; Ajmo et al., [@B3]). Furthermore, treatment with glycyrrhizin to inhibit HMGB1 activity limits the release of HMGB1 into the blood and attenuates the reduction of total PBMCs (Gu et al., [@B24]; Xiong et al., [@B98]). All of these results confirmed the negative correlation between HMGB1 release and lymphopenia, suggesting that HMGB1 participates in the process of SIID. HMGB1 was reported to bind to RAGE or activate the TLR4/MyD88 pathway, both of which promote the reduction of mature monocytes and lymphocytes in the circulation, and lead to subsequent post-stroke immunosuppression (Huang et al., [@B31]; Wild et al., [@B96]). In addition, a recent study has shown that alarmin HMGB1 aggravates brain and systemic inflammation *via* TLR4-dependent pathway in a rat PSI model induced by lipopolysaccharides (LPS), which forms a positive feedback loop between PSI-mediated HMGB1 release and following HMGB1-release-induced enhancement of LPS function and deteriorative PSI (Kim I. D. et al., [@B36]).\n\nConclusion and Perspective {#s6}\n==========================\n\nAt present, the incidence of ischemic stroke is not only increasing year by year but also becoming more severe. Ischemic stroke treatment is limited and the mortality and disability rate is extremely high. The study of ischemic stroke still requires further work. A growing body of evidence supports the idea that HMGB1 is a cytokine that regulates inflammation and immune response. It can mediate and amplify the inflammatory response after ischemia and aggravate brain damage, as well as exacerbate SIID and PSI. HMGB1 may be a potential biomarker for independently predicting the poor stroke prognosis because its upregulated blood lever is positively correlated with the infarct volume, neurological deficiency degree, BBB damage and serum inflammatory cytokine production (Le et al., [@B42]). However, the distribution and functions of HMGB1 are varied, but its biological mechanism is not yet clear, particularly regarding the relationship between HMGB1 and stroke. It is therefore important to study the mechanisms of HMGB1 in the absence of stroke, with the aim of using HMGB1 as a target for stroke treatments and provide new prospects and directions for stroke diagnosis and prognosis.\n\nAuthor Contributions {#s7}\n====================\n\nXX and LG designed this review article. YY wrote the manuscript. ZZ and HZ and TJ polished the article.\n\nConflict of Interest Statement {#s8}\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe thank Bronwen Gardner, PhD, and Hanne Gadeberg, PhD, from Liwen Bianji, Edanz Editing China ([www.liwenbianji.cn/ac](http://www.liwenbianji.cn/ac)), for editing the English text of a draft of this manuscript.\n\n**Funding.** This work was supported by the National Natural Science Foundation of China (no. 81771283 and 81301019 to LG and no. 81870939 and 81571147 to XX).\n\n[^1]: Edited by: Xiaohong Li, Institute for Basic Research in Developmental Disabilities (IBR), United States\n\n[^2]: Reviewed by: Yu-Feng Wang, Harbin Medical University, China; Amit U. Joshi, Stanford University, United States\n\n[^3]: ^\u2020^These authors have contributed equally to this work\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\nThis article narrates a few of my experiences and their effects on my understanding, selected for their novelty. Some of NIMHANS alumni may remember having heard bits and pieces of these experiences. The experiences profoundly influenced my clinical conduct by: (i) affecting the manner of my 'professional' cognition, (ii) prompting me to develop few constructs or concepts for bringing about some semblance of holism to my understanding, (iii) contributing to gradual emergence of a few ideas as principles of therapy that I try to adapt in my therapeutic work. I suspect that most fellow professionals would have encountered similar experiences and arrived at similar or different conclusions. After all, variability is a characteristic of natural phenomena in every dimension of its manifestation. If we can pool such experiences, some principles may eventually emerge, common to most professionals. Consequently, I hope that the following narrative will prompt the reader to reflect on them. I expect that, in the long run, investigative-psychiatry will accord equal importance to subjective experiences and their possible validity, leading to enrichment of psychotherapeutic literature.\n\n'SCIENTIFIC PRECISION,' OR PROBABILITIES AND 'SPECTRUMS'? {#sec1-2}\n=========================================================\n\nI followed-up same patients for over 3\u00bd years while working at Napsbury Hospital near St. Albans (UK) from 1971 to mid 1974. Many of them were readmitted twice or thrice. Nearly one third of them used to require review of diagnosis and its change. For example, a functioning obsessive-compulsive neurotic would become a dysfunctional paranoid schizophrenic a year later. Switchover from a neurotic to an affective disorder and vice-versa was most common. These changes occurred in spite of elaborate and repeated clinical assessments. Similarly, I found that the more and more I interacted with the patient, my understanding about that patient kept on evolving.\n\nFor me, this inconsistency of diagnosis was a rude shock. The effect was, ***I lost absolute faith in diagnosis.** I do not mean diagnosis is invalid or un-necessary. The profession cannot clinically and academically function without the instrument of diagnosis, which represents a set of commonly agreed upon concepts and constructs for professional convenience of understanding and communication*. Diagnosis represents a tentative assumption about the probable clinical state of a patient at a given point in time. Like the latitudes and longitudes, the underlying concepts and constructs serve as a reference-grid to infer a workable meaning from the clinical data. Thus, when two or more professionals with different 'reference-grids' were to read through recorded clinical histories, they are likely to deduce different diagnosis. But, we cannot expect the patient\\'s psychopathology, un-manifest and manifest to be according to our scheme of things.\n\n*Thus, our diagnostic constructs need not and do not always correspond to what is happening in the patient. As mentioned earlier, their best use is only as reference-points and conceptual templates for the convenience of understanding and communication*. Giving them any importance beyond these purposes will jeopardize the efficient management of patient, by distracting and diluting our conscious and un-conscious involvement with *real problems* that require attention or even masking the problems. About a decade later, the concept of 'spectrum of disorders' began to appear in professional literature.\n\nExcessive importance to diagnosis is an obstacle in the field of psychotherapy.\n\n'SCHOOLS' OF PSYCHOTHERAPY, OR JUST DIFFERENT PERSPECTIVES? {#sec1-3}\n===========================================================\n\nWhile at Napsbury, I simultaneously worked for three consultants, looking after some of their patients. One was a senior Jungian training analyst, another was a disciple of Maxwell Jones and an expert in therapeutic community, and the like. The other was an eclectic. We used to conduct therapeutic ward-group meetings twice a week in a 'sub-acute' ward, and once a week in an 'acute ward.' My psychotherapy supervisor was a Kleinian analyst. In that hospital, a clinical psychologist colleague of mine was undergoing training in Freudian analytical group psychotherapy. Upon my request, his supervisor permitted me to participate in the group-sessions as a silent participant. After over a year, this privilege was terminated, and I never came to know what complications, if any, in the group process or in the trainee-therapist led to this decision.\n\nPossibly, as a consequence of this exposure to many 'schools' and modalities, ***I lost absolute faith in 'schools of psychotherapy.'** The 'schools of psychotherapy' are not fiction. On the contrary, they are valid, as I have been able to observe most of the described concepts in my clinical work. All of them are equally valid from respective 'reference points.'* For example, suppose that two psychiatrists believe in two different sets of concepts. If they happen to read through a detailed case history, each will 'see' evidences for one\\'s concepts. Also, in my patients, I have been able to understand different parts of pathology by applying concepts from different schools. As mentioned above in respect of diagnosis, we cannot expect the patient and his/her family system to develop psychopathology according to our concepts and constructs. This is the reason for emergence of so many schools of psychotherapy to make some workable sense of the extreme variability in clinical manifestation.\n\nSimilar to diagnostic system, while concepts and constructs are essential for any clinical or therapeutic work, rigid adherence to them can be counter-productive. Once, I suffered from the concept of 'castrating patient.'\n\nFACE, OR LOOSE OUT {#sec1-4}\n==================\n\nAs mentioned earlier, I was an active participant in therapeutic ward group meetings on the lines of therapeutic milieu. The participants were patients, nurses, doctors, consultants, and at times, even patients' key-relatives. After a year, I started conducting weekly-once ward group meeting on my own in a 'sub-acute' ward, consisting of only patients and myself. I discussed the proceedings of each meeting with the consultant. After a few weeks, when a meeting was about to end, one patient said, \"you are no good,\" and another joined, \"you are useless.\" I was stunned, and momentarily lost my bearings, but, some how managed to say \"we will discuss this issue in our next meeting.\" I continued to be preoccupied with this experience, and was sleepless for two days.\n\nAfter four days, while discussing about this session with the consultant, I expressed my preference that either the consultant attends the meeting with me to 'buffer the crisis' ('bail-me-out'), or gets another registrar to take-over the group. The consultant\\'s advice was stern: \"Go and face the group. Otherwise, you will never be able to face a group again in your life.\" He suggested a few alternative strategies to adapt in the next meeting. I did attend it, invited discussion about my conduct and role in the group. I was astonished to find the same two patients extremely supportive and insightful in the discussions. When that meeting ended, I felt extremely confident of my self, that 'I can manage any group any where.' Then, ***I lost absolute faith in the 'castrating patient.'** I do not mean that such a phenomena does not exist. It had happened to me! I mean that it should eventually be possible to successfully 'wade' through any therapeutic situation, however unpleasant. The blunt message is: therapist\\'s despair is suicidal to further learning. Be hopeful and think positively. There is always a way out!* Rigid adherence to any concept or construct can be dangerous.\n\n*At a personal level, patients are not hostile to therapists; they dare not be so to their benefactor!* Negative-feelings emerging in therapy are part of the therapeutic process. Therapist has to make positive use of them so that both patient and therapist learn from it. Believe in the ancient wisdom, \"whatever happens is for the good.\" There may however be exceptions, as in the case of patients with active and violent paranoid delusions.\n\nIn this context, it is necessary to underscore a point: *Do not be afraid of making mistakes,* while at the same time, being cautions not to commit known mistakes. Fear of making mistakes prevents useful learning, and perpetuates wrong learning. There are too many variables operative, most of them un-known, and it is impossible to negotiate an 'error-free therapeutic conduct.' The only insurance against consequences of 'mistakes' is the triad of sincerity (honesty), genuinity (absence of pretence), and empathic good-will.\n\nDuring that period, I was fortunate to recognize another concept, the professional\\'s self-concept of \"professional-ego\" that could perhaps, 'castrate' further learning if given a free reign.\n\nPATIENTS CAN GET WELL WITHOUT ME ! {#sec1-5}\n==================================\n\nA patient whom I followed-up regularly was an unmarried man in his thirties, living alone and working as a fitter in a factory. His diagnosis was 'paranoid depressive psychosis.' His delusions were centered around a perambulator in his back yard. It was the perambulator in which his grand mother used to take him out daily after he was orphaned early in his childhood. Even though manageable and functioning most of the time, occasionally he used to lapse into acute excarbations. At such times, he requied admission and ECTs. Over a period of time, I began to dread the prospect of seeing him. Every encounter with him was a reminder of my clinical impotence in bringing about any further improvement in his clinical state.\n\nDuring one such follow-up review, I was very surprised to see him happy and smiling. I had never seen him smiling earlier, and for those readers who have read P.G.Wodehouse, 'my jaw fell with a thud!'. I greeted him, \"I am glad to see you well and happy. I am interested to learn how this excellent change came about.\" He reported: (i) a few days ago he happened to cry in the factory, (ii) the personnel-officer took him to his office and enquired, (iii) he cried out his fears and worries relating to the perambulator, (iv) the personnel-officer drove the patient home and took away the perambulator, and (v) his fears and worries ceased, and depression disappeared! This account was a terrible shock to me and an affront to my professional ego. Momentarily, I even felt angry, \"how dare he get well with out my help?\" Of course, a few months later, his symptoms recurred; but, were milder.\n\n***I lost absolute faith in the indispensability of mental health profession(al).** A patient can get well, and survive by means and methods other than offered by mental health professional! Also, the patient is capable of and responsible for doing so. If so, 'why has he come to consult?'* 'Why the patient has chosen not to put these abilities to use?' Psychotherapeutic effort with any patient should attempt to answer this question. *It is this attempt that contributes to favorable therapeutic outcome.* Psychotherapy happens while the therapist is attempting to understand 'why this patient has not been able to manage his or her problems?' But, the outcome can be extremely variable.\n\n*Sincerely believe that patient is capable of getting well and surviving by means and methods other than offered by mental health profession. Actively and positively, give responsibility to patient.* Of course, this principle may not apply in its entirety in such cases as mental retardation, etc.\n\n*Professional or personal pride (as different from 'self-respect' and 'confidence') is a serious handicap to learn and practice psychotherapy.* I believe that *effective and continued learning* requires a state analogous to humility, something like: 'I do not know, I want to know....\" In contrast, professional pride is a characteristic of one who knows!\n\nSWAMI VIVEKANADA\\'S CANNONS {#sec1-6}\n===========================\n\nIn this context, it is interesting to consider what swami Vivekananda has said about teaching and learning. Here, we have to remind ourselves, that in a therapeutic situation bilateral 'teaching-learning' is taking place between the patient and the therapist. The cannons are: \"Real teaching is: teaching how to learn. The student will then do his own learning.\" In psychotherapeutic setting, best therapeutic benefit is when the patient learns how to cope. The next cannon gives clue to how this can best happen.\"Best teaching is when the teacher is not aware he is teaching, and best learning is when the student is not aware of being taught.\" While applying this cannon to psychotherapeutic situation, two questions pose themselves: (i) When is the therapist unaware of 'doing psychotherapy or teaching patient how to cope?' (ii) When is the patient unaware of being 'taught or, helped?' The answers are: The therapist is unaware of doing psychotherapy when he or she is busy understanding the patient. Even interpretations are only meant to lessen resistances and/or 'loosen-up' further abreaction or inflow of information. The patient is unaware of 'learning to cope' when he or she is busy interacting with and responding to the therapist.\n\nConsequent to the above points, a few corollaries emerge: *Psychotherapy is not done, but happens. It happens when an individual\\'s distress is favorably modified by interaction with another individual.* It happens when there is concern for another\\'s distress, empathy, good-will, intent to relieve distress, and hopeful and sincere effort. These are 'desirable therapist qualities' advocated by Rogerians.As said earlier, *patient benefits not by therapist\\'s objective of doing psychotherapy and consequent behavior, but by his/her objective of understanding: why, how, under what circumstances, and for what reasons is the patient in distress?* Consciously, or un-consciously, in a parallel fashion, the patient too will be following this 'understanding' and learns coping strategies in the bargain.*Psychotherapeutic benefit is a consequence of desirable therapist qualities. Without these, mere methods, techniques, and concepts are useless*. Therefore, these qualities should ideally become focus of training in psychotherapy. Even if they belong to the phenomena of personality, I believe that they can be cultivated and/or strengthened.\n\nNow, it is very easy to appreciate how the pride of being a therapist (a product of professional-ego), of being a 'guru,' or of 'doing psychotherapy' can be a serious obstacle to therapeutic benefit.\n\nTHERAPIST AND HIS TOOLS {#sec1-7}\n=======================\n\nI now shift the time and place to eighties and nineties at NIMHANS. From 1983 to 1994, I was involved in the department\\'s psychotherapy training programme for psychiatry residents, and in evaluating their therapeutic performance. During this period, about 250 residents were trained and assessed. The department\\'s effort was to offer a uniform training with the help of: (i) a structured format of training and supervision, both theoretical and supervisory, (ii) weekly meeting of the supervisors to exchange experiences in their weekly-once supervision-sessions, and (iii) structured assessment of the residents\\' skills at the end of each course. There were three important variables in the training programme: the supervisor, the resident, and the patient.\n\nOver the period, my and fellow supervisors\\' cumulative observation was: (i) In spite of fairly uniform training procedure, there were individual variations in supervisory sessions. (ii) The residents, varied widely in how they conducted themselves in therapy sessions with their patients, in respect of concepts, methods and techniques used. (iii) Every resident was successful with some patients, unsuccessful with a few, and with variably moderate results with others. ***I lost absolute faith in methods and techniques.** It is not that they are useless; on the contrary, they are the tools of therapeutic management; but, they are less important than the person who uses them. They are useful only to the extent that their user makes them useful!* In a way, a therapist is to therapy what a director (*'Sutradhari'*) is to a play. The therapist plays a major role in the therapeutic outcome.\n\nApart from the above events, there were also many, less dramatic events and, perhaps many more events that did not register in my conscious awareness that contributed to my learning.\n\nTOO MANY VARIABLES (OR, \"TOO MANY COOKS\") {#sec1-8}\n=========================================\n\nThere are many variables operative in psychotherapy. Therapist\\'s personality and choices of concepts and techniques contribute to a continuum of therapeutic methods at least in two dimensions, as figuratively represented in [Figure 1](#F0001){ref-type=\"fig\"}.\n\n![Continua in psychotherapeutic methods](IJP-50-213-g001){#F0001}\n\nThen, there are variables that influence the psycho-therapeutic outcome. The important ones are: (i) Patient. (ii) Diagnosis. (iii) Socio-environmental context. (iv) Therapist. Consequently, the outcome necessarily follows Guassian distribution on each of the four dimensions. Therefore, absolute therapeutic effectiveness is a myth. The logical consequences of this fact are shown in [Table 1](#T0001){ref-type=\"table\"}.\n\n###### \n\nTherapist-effectiveness and patient-potential to benefit\n\n  Therapist-effectiveness                           Patient-potential to benefit\n  ------------------------------------------------- --------------------------------------------------------\n  Can benefit any patient (Very rare)               Can benefit from any therapist (Very rare)\n  Can benefit most patients                         Can benefit from most therapists\n  Can benefit particular types of patients          Can benefit from particular types of therapists\n  Cannot benefit any patient (Probably, not true)   Cannot benefit from any therapist (Probably, not true)\n\nBut, it is essential to remember our ancient tenet that no sincere effort ever gets wasted. Even if the patient and therapist are unaware of it, even if there is no manifest therapeutic benefit, the patient carries forward 'credits' from previous therapy or therapies to reap the benefit with some later therapist!\n\nLEARNING FROM INDIAN MYTHOLOGY {#sec1-9}\n==============================\n\nIn 1980s, when I became interested in collecting concepts relating to mental health in Indian mythology and related literature, I noticed a few shifts in my therapeutic orientation, which may appear radical in some respects.\n\n1.  A chapter (no. 4) titled 'Psychotherapeutic Paradigms from Indian Mythology' that I wrote for *Therapeutic Use of Stories* (Ed. By Kedar Nath Dwivedi, 1997, Routledge, London) describes the first of the effects. I was convinced that components of ideal human behavior are almost same as those of desirable therapist qualities, as well as those that correlate with state of well-being. That is, an individual who has components of ideal behavior as a part of personality will not only be inherently mentally healthy, but will also have a potential to be a 'natural' psychotherapist.\n\n2.  The second was a conviction that: as illnesses and difficulties (miseries) in human life are inevitable, a state of health is not defined by their absence, but by the individual\\'s ability to effectively manage these inevitable challenges.\n\n3.  Consequent to the second, the third was a conviction that mental health is not a passive state, but has to be actively earned by facing the challenges, bearing the difficulties, managing the adversities, gracefully accepting un-avoidable limitations and failures, and re-attempting as appropriate, etc.\n\n4.  The fourth was a conviction that a **set of dynamic factors** that are opposites of components of ideal behavior can explain why and how individuals suffer psychologically. The type of suffering (viz. diagnosis) is dependent on pre-morbid-personality, nature of stress, and related factors. *These factors are also dynamic forces similar and parallel to the traditional 'psycho-dynamics' that we are usually familiar with. By keenly studying any detailed clinical-psycho-social case-history, it is possible to explain the patient\\'s suffering as a consequence of aberrations in one or more of these factors.*\n\n*These factors can be called **'Humanistic Factors,'*** and can be used to explain any individual\\'s psychopathology, in the same way as one uses classical psychodynamic theories. Moreover, these factors can complement other existing theories. The Humanistic Factors are: *Excessive or abnormal selfishness* that encroaches upon and obstructs one\\'s civic, social, and other roles and responsibilities. It can contribute to such motivating forces as greed, competitiveness, intolerance, un-attainable ambitions and expectations, aggression, and inter-personal strife. This factor can reach insatiable proportions.*Excessive or abnormal pride,* which can also lead to consequences similar to that of selfishness. This can also reach insatiable proportions.*Not wanting to 'pay the price,' not wanting to 'sacrifice,' not wanting to put in necessary hard work or bear the necessary burden.* For example, let us assume the known fact that mental health is intimately related to 'adjustment' to one\\'s needs, abilities, stresses, strains, and demands of living. What else is adjustment if it is not: (i) giving up some advantage for some other advantage, (ii) accepting and bearing some disadvantage in order to avoid some other disadvantage, (iii) giving up some goal/objective for some other goal, and (iv) accepting and tolerating some degree of unpleasantness, defects, burdens, etc. People suffer and make others also miserable because they do not want to sacrifice; they want every thing, with little or no effort.*Pretence (lack of genuinity, or hypocrisy)*. For example, pretending to be: (i) what one is not, or (ii) in a psychological state different from what one is in, or (iii) having an ability one does not have. Pretence, like acting, is extremely stressful. This factor is most probably, a consequence of the above three factors.*Dis-honesty*. This includes unfairness, exploitation, and not applying to oneself the same principles that one expects from others, etc. This factor is most probably, a consequence of the above four factors. The behavior of *lying* is related to this factor. Technically, any distortion, exaggeration, or denial, etc. of a known fact constitutes a lie. This is so even if such a process take place intra-psychically.*Desire or addiction to either sensation-seeking (thrill, or excitement seeking), or dulling or altering the conscious state for the purpose of avoiding the responsibility of experiencing the current state of being*. It is interesting that our pharmacological mode of management does some thing similar under the control of the therapist!\n\n*It is frightening that in the present, modern culture, the entire media are brain-washing the population to cherish and imbibe the first four factors mentioned above. As a consequence, inevitably, they are forced to develop the last two factors also in the bargain.* Every one of us is a witness to this phenomenon.\n\n\\[Ed: The concluding part of this series will appear in the next issue\\]\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared\n"}
+{"text": "1 INTRODUCTION {#SEC1}\n==============\n\nGenome-wide association studies (GWAS) are an increasingly popular approach for identifying associations between genotype and phenotype. A large number of such studies have been performed recently to try to identify the genetic basis of a wide variety of diseases, and explore how this genetic basis differs depending on the geographic origin of the studied population. High-throughput genotyping chips are used to obtain the genotype of an individual at several hundreds of thousands of single nucleotide polymorphisms (SNPs). These sets of SNPs are able to represent most of the variability at the single locus level that was identified by the HapMap project (Frazer *et al.*, [@B3]). In a GWAS study, several thousands of disease individuals, and several thousands of healthy controls are genotyped. Statistical tests are used to identify SNPs that show a strong association with the disease. Strong association between a SNP and a disease can be evidence that the SNP is related to the disease, or that it is in linkage disequilibrium with SNPs that are related to the disease. In both cases significant associations provide promising leads for further experimental investigation into the genetic etiology of diseases. These studies have led to the identification of more than 150 risk loci in more than 60 diseases (Manolio and Collins, [@B9]). The Wellcome Trust Case-Control Consortium (WTCCC) genotype 500 000 SNPs in seven common diseases: type 1 diabetes (T1D), type 2 diabetes (T2D), coronary artery disease (CAD), Crohn\\'s disease (CD), bipolar disease (BD), hypertension (HT) and rheumatoid arthritis (RA) (WTCCC, [@B17]). In this article we use the individual genotype data from this study.\n\nComputational methods have been used to identify disease similarities using a variety of data sources, including gene expression in cancer (Rhodes *et al.*, [@B14]) and known relationships between mutations and phenotypes (Goh *et al.*, [@B5]). However, while a large number of GWAS focusing on individual diseases have been recently published, the attempts to integrate the results of multiple studies have been limited. Most of these integration approaches focus on combining multiple studies of the same disease in order to increase the statistical power (Zeggini *et al.*, [@B18]), or use data from other high-throughput measurement modalities to improve the results of GWAS studies (Chen *et al.*, [@B2]). Comparison between the genetic components of diseases have been done using four different approaches. The first approach is based on the identification of the association between one SNP in two different diseases in two independent studies. The second approach selects a group of SNPs that have been previously associated with some disease and tests if they are also associated with a different disease. An example of this approach is the genotyping of a large number of individuals with T1D at 17 SNPs that have been associated with other autoimmune diseases, which leads to the identification of a locus previously associated with only RA as being significantly associated with T1D as well (Fung *et al.*, [@B4]). The third approach pools data from individuals with several diseases prior to the statistical analysis, and has been used in the original WTCCC study. Several similar diseases (autoimmune diseases, metabolic and cardiovascular diseases) are grouped in order to increase the statistical power for identifying SNPs that are significantly associated with all the diseases in the pool. The fourth approach compares the results of multiple GWAS, and has been previously applied to the WTCCC dataset (Torkamani *et al.*, [@B16]). They use the *P*-values indicating the significance of the association between a SNP and a single disease, and compute the correlations between these *P*-values in pairs of diseases, as well as the size of the intersection of the 1000 most significant SNPs in pairs of diseases. They identify strong similarities between T1D and RA, between CD and HT, and between BD and T2D.\n\nIn this work, we introduce a novel approach to identify similarities in the genetic architecture of diseases. We train a classifier that distinguishes between a *reference disease* and the control set. We then use this classifier to classify all the individuals that have a *query disease*. If there is a similarity at the genetic level between the query disease and the reference disease, we expect more individuals with the query disease to be classified as belonging to the disease class than if there is no similarity. We generalize our procedure to multiple disease comparison: given a set of multiple diseases, we use each in turn as the reference disease while treating all others as query diseases.\n\nThere are two main differences between our new approach and existing analyses. First, previous approaches \\[such as Torkamani *et al.* ([@B16])\\] compute a significance score for each SNP, and then use these scores for comparing diseases. In our approach, we first compute a classification for each individual, and then compare diseases using these classifications. Second, we train the classifier using information from all SNPs, and during this learning process select the SNPs that contribute to the classification based on the genotype data only. This genome-wide approach makes it possible to see the classifier as a statistical representation of the differences between the disease set and the control set.\n\nThe use of classifiers in the context of GWAS has been limited so far. In particular, attempts at using them for predicting outcome based on genotype have been unsuccessful. For example, a recent prospective study in T2D (Meigs *et al.*, [@B11]) found that using 18 loci known to be associated with T2D in a logistic regression classifier together with known phenotypic risk factors does not significantly improve the risk classification, and leads to a reclassification in only 4% of the patients. A particular challenge in the context of outcome prediction is that the prevalence of most diseases is relatively low and that it is therefore necessary to achieve high precision in order for the classifier to be usable. Our goal is not predicting individual outcomes, and we only compare predictions made by a single classifier. We can therefore ignore disease prevalence.\n\nA second challenge in the use of a classification approach for finding disease similarities is that the classifier does not explicitly identify genetic features of the disease, but rather learns to distinguish the disease set from the control set. Differences between the two sets that are due to other factors might therefore lead to incorrect results. In most GWAS, a careful choice of matched controls limits this risk. However, when using a classifier trained on one GWAS to classify individuals from a different study, there is a risk that the background distribution of SNPs is very different between the populations in which the datasets have been collected, which could lead to errors, particularly when comparing diseases using datasets from different geographic origins. This risk can be limited by using disease data from a single source. In this work, we use genotype data provided by the WTCCC study, in which all individuals were living in Great Britain and individuals with non-Caucasian ancestry were excluded.\n\nIn this article, we first provide a detailed description of the analysis approach. We then show that we are able to train classifiers that achieve a classification error that is clearly below the baseline error for T1D, T2D, BD, HT and CAD. We use these classifiers to identify strong similarities between T1D and RA, as well as between HT and BD, and weak similarities between T1D and both BD and HT. We also show that we are able to train a classifier that distinguishes between the two control sets in the WTCCC data. We use this classifier to identify similarities between some diseases and individual control sets. This finding matches observations made during the quality check phase of the original study. The implications of this finding on our approach are addressed in the [Section 5](#SEC5){ref-type=\"sec\"}. Finally, we discuss the implications of the similarities we find, and propose extensions of this approach. A detailed description of the dataset used in this work, the data pre-processing, the decision tree classifier and the comparison procedure are provided in [Sections 3](#SEC3){ref-type=\"sec\"} and [4](#SEC4){ref-type=\"sec\"}, respectively, at the end of the article.\n\n2 APPROACH {#SEC2}\n==========\n\nIn this section, we define the general classifier-based approach to identify genetic similarities between diseases. The approach can be separated into four steps: data collection, preprocessing, classifier training and disease comparison. [Figure 1](#F1){ref-type=\"fig\"} provides an overview of the training and comparison steps. Fig. 1.Overview of the approach. This figure presents the *classification* and *comparison* steps of our analysis pipeline. These steps are repeated using a different *reference disease* each time. The classifier returns a real value between 0.0 and 1.0 which we call *disease-class probability*. The histograms represent the distribution of the disease-class probability of the individuals with the reference disease (left) and of the controls (right). In the situation depicted in this figure, there is evidence that query disease C is more similar to the reference disease than the other query diseases.\n\nThe data collection step consists of collecting samples from individuals with several diseases, as well as matched controls, and genotyping them. Alternatively, existing data can be reanalysed. In both cases, it is important to limit the differences between the disease sets and the control sets that are not related to the disease phenotype. Similarly, differences between the different disease sets should also be limited. In particular, it is recommended to use individuals with the same geographic origin, the same ancestry, and a single genotyping technology for the whole study. In this work, we use existing data from the WTCCC which satisfies these criteria.\n\nIn the preprocessing step, the data are filtered and uncertain genotype measurements, as well as individuals and SNPs that do not fit quality requirements are discarded. It is important to develop preprocessing steps that ensure good data quality. Approaches that analyze each SNP individually can afford to have a more stringent, often manual post-processing step on the relatively few SNPs that show strong association. The SNPs that do not pass this quality inspection can be discarded without affecting the results obtained on other SNPs. In our approach however, classifier training is done using genome-wide information, and removing even a single SNP used by the classifier could potentially require retraining the entire classifier. It is therefore impractical to perform any kind of post-processing at the SNP level. The [Section 3](#SEC3){ref-type=\"sec\"} of this article describes the data used in this work, as well as the quality control measures we take.\n\nThe classifier training and comparison steps are interleaved. We start with a list of diseases and a set of individual genotypes for each disease, as well as at least one set of control genotypes. We pick one disease as *reference disease*, and refer to the remaining diseases as *query diseases*. We train a classifier distinguishing the corresponding disease set from the control set. For any individual, this classifier could either return a binary classification (with values 0 and 1 indicating that the classifier believes the individual is part of, respectively, the controls class or the disease class) or a continuous value between 0 and 1. This continuous value can be seen as the probability of the individual to be part of the disease class, as predicted by the classifier. We refer to this value as *disease-class probability*. For simplicity, we will only use the disease-class probability values for the rest of this section, but the comparison step can be performed similarly using binary classifications. During the comparison step, we classify individuals from the query disease sets using the classifier obtained in the training step, and for each query disease, compute the average disease-class probability. The training and comparison steps are then repeated so that each disease is used once as reference disease.\n\nWe can compare the average disease-class probability of the different query diseases to identify similarities between them. Diseases that have a higher average disease-class probability are more likely to be similar to the reference disease than diseases with a lower average disease-class probability. Using cross-validation, we can obtain the average disease-class probability of the reference disease set and the control set used for training the classifier, and compare them with the values of the other diseases. One particular caveat that needs to be considered in this analysis is that while the classifier does distinguish the control set from the disease set, there is no guarantee that it will only identify genetic features of the disease set. It is also possible that it will identify and use characteristics of the training set, especially if there are data quality issues. This case can be identified during the comparison step if the average disease-class probability of most query diseases is close to the average disease-class probability of the reference disease, but very different from the average disease-class probability of the control set. It is therefore important to look at the distribution of the average disease-class probabilities of all query diseases before concluding that an individual disease is similar to the reference disease.\n\nIt is important to note that the disease-class probability of a given individual does not correspond to the probability of this individual actually having the disease. The disease frequency is significantly higher in the datasets we use for training the classifier than in the real population. In a machine learning problem in which the test data are class-imbalanced, training is commonly done on class-balanced data, and class priors are then used to correct for the imbalance. Such priors would, however, scale all probabilities linearly, and would not affect the relationships we identify, nor their significance. Estimating the probability of an individual having the disease is not the goal of this project and we can therefore ignore class priors.\n\nA large variety of classifiers can be integrated into the analysis pipeline used in our approach. The [Section 4](#SEC4){ref-type=\"sec\"} provides a more formal description of the classification task. In this article, we use a common classifier, decision trees, to show that this approach allows us to identify similarities. The specific details about the decision tree classifier, and how its outputs are used in the analysis step are described in the [Section 4](#SEC4){ref-type=\"sec\"}.\n\n3 RESULTS {#SEC3}\n=========\n\nWe evaluate the ability of our analysis approach to identify similarities between diseases using the set of seven diseases provided by the WTCCC. In this section, we first evaluate the performance of individual classifiers that distinguish one disease from the joint control set. We then show that these classifiers can identify similarities between diseases. Finally, we use our classifier to identify differences between the two control sets, and provide evidence indicating that these differences do not affect the disease similarities we identify.\n\n3.1 Classifier performance {#SEC3.1}\n--------------------------\n\nWe first train one classifier for each disease using both the *58C* and the *UKBS* sets as controls. The performance of each classifier is evaluated using cross-validation, and reported in [Table 1](#T1){ref-type=\"table\"}. We compare our classifier to a baseline classifier that classifies all individuals into one class without using the SNP data at all. The best error such a classifier can achieve during cross-validation is the frequency of the smaller class in the training set. We refer to this value as the *baseline error*. Table 1.Classifier performance (cross-validation)DiseaseBaseline (%)Error (%)Precision (%)Recall (%)\u0394~*p*~LeavesT1D40.0522.9371.6570.710.3839RA38.4333.4559.1242.090.13012BD38.2433.5962.6030.180.08711HT39.9236.7757.9828.640.08012CAD39.0536.6255.2532.730.07512T2D39.538.054.1225.050.05214CD36.6336.2829.8318.430.04611[^1]\n\nThe disease for which the classifier performs best is T1D, with a classification error of 22.93%, compared with a baseline error of 40.05%. The classification error obtained by the decision tree classifier is also below the baseline error for several other diseases, although by a substantially smaller margin. This is the case for RA (with an error of 33.45% versus 38.43%), BD (33.59% versus 38.24%), HT (36.77% versus 39.92%) and CAD (36.62% versus 39.05%). For two diseases, T2D and CD, the improvement compared with the baseline error is only minimal, and we choose not to use these classifiers in our analysis. While the classifiers that we keep only provide small improvements in terms of classification error (with the exception of T1D), they have a significantly better trade-off between precision (at least 55%) and recall (at least 28%) than the baseline classifier (which would classify all individuals as controls).\n\nWe do not use these classifiers in a binary way, but rather use the disease-class probability, which is the conditional probability of an individual to be part of the disease-class given its genotype, under the model of the reference disease learned by the classifier (see [Section 7](#SEC7){ref-type=\"sec\"} for a precise definition for decision trees). It is therefore interesting to consider the distributions of the disease-class probability, as obtained during cross-validation. [Figure 2](#F2){ref-type=\"fig\"} illustrates that these distributions differ significantly for T1D. It can also be seen that there are individuals for which the disease-class probability is close to 50%, meaning that there are leaf nodes in the classifier that represent subsets of the data that cannot be distinguished well. Our approach takes this into account by using disease-class probabilities rather than binary classifications. In order to evaluate the ability of our classifiers to distinguish between the disease set and the control set using the disease-class probability metric, we use the difference \u0394~*p*~ of the average disease-class probability between the two sets. The classifiers that we keep all have values of \u0394~*p*~ above 0.075. This illustrates that while there are only small improvements in binary classification performance, the classifiers are able to distinguish between the disease set and the control set in the way we intend to use them. Fig. 2.Distribution of the disease-class probabilities for the T1D classifier. The two histograms show the distribution of the disease-class probability of the individuals, respectively, in the joint control set (top) and in the T1D set (bottom), as computed during cross-validation. The red lines represent the average disease-class probabilities, and the black line indicates the 0.5 probability cut-off used for binary classification. The plot in between the histograms shows the average disease-class probabilities of the six other diseases on the interval between the average disease-class probabilities of the control set and of the disease set.\n\n3.2 Disease similarities {#SEC3.2}\n------------------------\n\nFor each of the five classifiers with sufficiently good performance, we compute the average disease-class probability of each of the six query diseases. In summary, we identify strong symmetrical similarities between T1D and RA, as well as between BD and HT. Furthermore, we find that T1D is closer to both BD and HT than other diseases, even though we did not find the symmetrical relation using the T1D classifier. This section provides a detailed presentation of these results.\n\nFor T1D, the average disease-class probability for the control set and the disease set, as computed using cross-validation, are 0.259 and 0.642, respectively. [Figure 2](#F2){ref-type=\"fig\"} shows the distribution of the average disease-class probabilities for the query diseases. RA, another autoimmune disease, is clearly the closest to T1D (average disease-class probability of 0.337). This result is significant, with *P*-value \\<10^\u22125^ (see the [Section 4](#SEC4){ref-type=\"sec\"} for details on how *P*-values are obtained). All other diseases have an average disease-class probability that is close to that of the control set, which means that there is no evidence of similarity with T1D.\n\nFor RA, the average disease-class probabilities are 0.303 for the control set and 0.433 for the disease set. The distribution of the average disease-class probabilities for the other diseases are shown on [Figure 3](#F3){ref-type=\"fig\"}a. We can observe that T1D (average disease-class probability of 0.397) is closest to RA (*P*\\< 10^\u22125^), meaning that we find a symmetrical similarity between the two diseases. All other diseases have an average disease-class probability close to the one of the control set. Fig. 3.Disease-class probabilities comparisons. The plots represent the interval between the average disease-class probabilities of the control set and of the disease set for RA (**a**), BD (**b**), HT (**c**) and CAD (**d**), respectively. The average disease-class probabilities for all the query diseases are shown in blue on every plot. Note that while all plots on this figure use the same scale, different scales are used for the central plots of [figures 2](#F2){ref-type=\"fig\"} and [4](#F4){ref-type=\"fig\"}.\n\nFor BD, the average disease-class probabilities are 0.297 for the control set and 0.384 for the disease set. The distribution of the average disease-class probabilities for the query diseases are shown in [Figure 3](#F3){ref-type=\"fig\"}b. We can observe that there is a wider spread in the average disease-class probabilities, and that there is no cluster of diseases close to the control set. We can also observe that HT (average disease-class probability of 0.359, *P* \\< 10^\u22125^) is closest to BD, followed by T1D (average disease-class probability of 0.354, *P*-value of 0.001).\n\nFor HT, the average disease-class probabilities are 0.315 for the control set and 0.395 for the disease set. The distribution of the average disease-class probabilities for the other diseases are shown in [Figure 3](#F3){ref-type=\"fig\"}c. We can observe that BD (average disease-class probability of 0.381, *P*-value \\< 10^\u22125^) is clearly closest to HT. T1D (average disease-class probability of 0.368, *P* \\< 10^\u22125^) is also closer to HT than the remaining diseases.\n\nFor CAD the average differences between the query diseases are smaller than for all the other classifiers ([Fig. 3](#F3){ref-type=\"fig\"}d). Furthermore, the classifier for CAD is the one with the worst performance amongst the ones we use in the comparison phase. Therefore, we believe that the results are not strong enough to report putative similarities identified using this classifier, even though some differences between diseases have significant *P*-values.\n\n3.3 Differences between control sets {#SEC3.3}\n------------------------------------\n\nThe original WTCCC study found several SNPs that are significantly associated with one of the two control sets. These SNPs are filtered out during preprocessing, both in the WTCCC study and in this work. However, the mere existence of differences between two control sets prompted the question whether a classifier could distinguish the two sets, and if so, what the implications of this finding would be on the validity of results obtained with these control sets.\n\nWe perform several experiments using the two control sets separately, and report the results in [Table 2](#T2){ref-type=\"table\"}. First, we train a *control--control classifier* that distinguishes the two control sets from each other. This classifier achieves an error of 41.15% compared with a baseline error of 49.62%, and a \u0394~*p*~ of 0.093. This shows that we are able to distinguish to some extent between the two control sets. [Figure 4](#F4){ref-type=\"fig\"} shows the distribution of the *58C class probability* (which corresponds to the value called *disease-class probability* when the classifier distinguishes between one disease and the controls). In order to verify that this result is due to differences between the two specific control set, and not the ability of our classifier to distinguish between any two sets, we randomly split all control individuals into two sets, *R1* and *R2*. We train a classifier to distinguish between these two sets. We find that this classifier does only minimally improves the classification error (error of 49.45%, baseline error of 50.03%, \u0394~*p*~ of \u22120.003). Fig. 4.Distribution of the class probabilities for the control--control classifier distinguishing the *UKBS* control set from the *58C* control set. The two histograms show the distribution of the *58C* class probability of the individuals, respectively, in the *UKBS* control set (top) and in the *58C* control set (bottom), as computed during cross-validation. The red lines represent the average class probabilities, and the black line indicates the 0.5 probability cut-off used for binary classification. The plot in between the histograms shows the average disease-class probabilities of all seven other diseases on the interval between the average class probabilities of the two control sets. Table 2.Separate training set classifier performanceExperimentBaseline (%)Error (%)Precision (%)Recall (%)\u0394~*p*~LeavesUKBS/58C49.6241.1558.3364.050.09311R1/R250.0349.4550.5946.42\u22120.00311UKBS/T1D42.6223.1579.5380.340.402858C/T1D42.9924.4676.6082.220.3708UKBS/RA44.2936.4266.2170.720.1441058C/RA44.6638.1164.8967.830.1359[^2]\n\nWe apply the comparison step of our pipeline using the control--control classifier in order to identify possible similarities between the disease set and one of the control sets. [Figure 4](#F4){ref-type=\"fig\"} shows the distribution of the average *58C* class probabilities for each disease. The average disease-class probabilities obtained during cross-validation are 0.477 for the *UKBS* set and 0.561 for the *58C* set. Both HT (average *58C* class probability of 0.521, *P* \\< 10^\u22125^) and BD (average *58C* class probability of 0.514, *P*-value of 0.0002) are closer to the *58C* control set, whereas both RA (average *58C* class probability of 0.487, *P* \\< 10^\u22125^) and CAD (average *58C* class probability of 0.489, *P*-value of 0.0003) are closer to the *UKBS* control set.\n\nGiven the differences between the control sets, and the unexpected similarities between control sets and diseases, we are interested in verifying that the performance of the disease classifiers used in the analysis is not an artifact caused by these differences. We therefore train two new classifiers for each disease, one using only *UKBS* as control set, and one using only *58C* as control set. The performance of these classifiers for T1D and RA is shown in [Table 2](#T2){ref-type=\"table\"}, and is similar to the performance of the classifiers that use both control sets together. For the remaining diseases (including HT and BD), the classifiers using only one of the control sets do not achieve a classification error below the baseline error, most likely due to the smaller training set (i.e. overfitting). For each of the classifiers for T1D and RA, we compute the average disease-class probability for the other six diseases as well as the unused control set. The similarities between the two diseases are significant in all four classifiers. Furthermore, the average disease-class probability of the unused control set is similar to the average disease-class probability of the other five diseases, and not significantly closer to T1D or RA. Therefore, we can conclude that the results obtained using the T1D and RA classifiers are not due to differences between the control sets. Furthermore, the results using a single control set provide further evidence indicating that the classifiers do identify relevant features of T1D and RA, respectively, rather than relevant features of the control set.\n\n4 DISCUSSION {#SEC4}\n============\n\nIn this work, we introduce a novel approach for identifying genetic similarities between diseases using classifiers. We identify genetic similarities between several diseases. In this section, we first discuss the implications of these findings. We then consider challenges in the application of classifiers to GWAS data. Finally, we propose possible extensions of this approach.\n\nWe identify a strong similarity between T1D and RA. Genetic factors that are common to these two autoimmune diseases were identified well before the advent of GWAS, and linked to the HLA genes (Torfs *et al.*, [@B15], Lin *et al.*, [@B7]). The original WTCCC study (WTCCC, [@B17]) identifies several genes that appear to be associated with both diseases. We look at the classifiers corresponding to these two diseases. The SNP with the highest information gain in T1D is rs9273363, which is located on chromosome 6, near MHC class II gene HLA-DQB1, and is also the SNP that is most strongly associated with T1D in the initial analysis of the WTCCC data, with a *P*-value of 4.29 \u00d7 10^\u2212298^ (Nejentsev *et al.*, [@B12]). This is the strongest association reported for any disease in the WTCCC study, which explains to a large extent why the T1D classifier so clearly outperforms the classifiers for the other diseases. This SNP is also significantly associated with RA (*P*-value of 6.74 \u00d7 10^\u221211^). The SNP with the highest information gain in RA is rs9275418, which is also part of the MHC region, and is strongly associated with both RA (*P*-value of 1.00 \u00d7 10^\u221248^) and T1D (*P*-value of 7.36 \u00d7 10^\u2212126^). This shows that our approach is able to recover a known result, and uses SNPs that have been found to be significantly associated with both diseases in an independent analysis of the same data.\n\nThe similarity we identify between HT and BD is interesting, since there does not appear to be previous evidence of a link between the two diseases at the genetic level. However, a recent study identified an increased risk of HT in patients with BD compared with general population, as well as compared to patients with schizophrenia in the Dannish population (Johannessen *et al.*, [@B6]). The WTCCC study only identified SNPs with moderate association to HT (lowest *P*-value of 7.85 \u00d7 10^\u22126^) and a single SNP with strong association with BD (*P*-value of 6.29 \u00d7 10^\u22128^). The decision trees for both diseases use a large number of SNPs that have a very weak association with the respective disease. Both classifiers have a classification error that is clearly below the baseline error, and provide evidence of similarity between the two diseases. This indicates that our classifier-based approach is able to use the weak signals of a large number of SNPs to identify evidence for similarities that would be missed by comparing only SNPs that show moderate or strong association with the diseases. Further analyzes are necessary to identify the nature and implications of the similarity we find between HT and BD, as well as the weaker similarity we identified between these two diseases and T1D.\n\nWe also show that we can train a classifier that can distinguish the two control sets, and we use it to identify diseases that are more similar to one of the control set than the other. This is not an unexpected finding, since SNPs that were strongly associated with a control set were identified and discarded in the WTCCC study. These SNPs were also removed in the preprocessing step of our study, and the results we obtain when trying to distinguish the two control sets therefore show that the decision tree classifier is able to achieve a classification error below the baseline error even though the SNPs with the strongest association could not be used by the classifier. The similarities between some diseases and one of the control sets can most likely be explained by some subtle data quality issue. During quality control, the authors of the WTCCC study found several hundreds of SNPs in which some datasets exhibited a particular probe intensity clustering \\[see the [Supplementary Material](http://bioinformatics.oxfordjournals.org/cgi/content/full/btp226/DC1) of the original WTCCC study (WTCCC, [@B17]) for details\\]. This particular pattern was always observed in *58C*, *BD*, *CD*, *HT*, *T1D*, *T2D*, but not in *UKBS*, *RA* and *CAD*. This matches the result obtained using our classifier-based approach, in which *RA* and *CAD* were predicted to be most similar to *UKBS*, and could therefore be a possible explanation of the similarities we find.\n\nWhile we do find several interesting similarities between diseases, we also observe that training a classifier that distinguishes between individuals with a disease and controls using SNP data poses numerous challenges. The first is that whether someone will develop a disease is strongly influenced by environmental factors. The genetic associations that can be identified using GWAS are only predispositions, and it is therefore likely that some fraction of the control set will have the predispositions, but will not develop the disease. Furthermore, depending on the level of screening, the disease might be undiagnosed in some control individuals, and individuals that are part of a disease set might have other diseases as well. This is especially true for high-prevalence diseases like HT.\n\nObtaining good classifier performance by itself is not, however, the main goal of our approach. We show that we can find similarities even when the classifier performance only shows small improvements compared with the baseline error. In this work, we focus on the comparison approach, not on developing a classifier specially suited for the particular task of GWAS classification. We use decision trees because they are a simple, commonly used classification algorithm.\n\nThis work shows that classifiers can be used to identify similarities between diseases. This novel approach can be expanded into several directions. First, classification performance can be potentially improved by using a different generic classifier, or by developing classifiers that do take into account the specific characteristics of SNP data. Second, further analysis methods need to be developed in order to analyze the trained classifiers, and identify precisely the SNPs that do lead to the similarities this approach detects. Such a methodology would be useful, for example, to further analyze the putative similarity between HT and BD. Third, building on the fact that our approach considers the whole genotype of an individual, it could be possible to identify subtypes of diseases, and cluster individuals according to their subtype. Finally, modifying the approach to allow the integration of studies performed in populations of different origins or using different genotyping platforms would allow the comparison of a larger number of diseases.\n\nOur approach identifies similarities between the genetic architecture of diseases. This is, however, only one of the many axes along which disease similarities could be described. In particular, both genetic and environmental factors interact in diseases, and the genetic architecture for two diseases could be similar, but the environmental triggers could be different, leading to low co-occurrence. There is therefore a need for methods that integrate similarities of different kinds that were identified using different measurement and analysis modalities. An example of such an approach is the computation of disease profiles that integrate both environmental ethiological factors and genetic factors (Liu *et al.*, [@B8]).\n\n5 CONCLUSION {#SEC5}\n============\n\nGWAS have been used to identify candidate loci likely to be linked to a wide variety of diseases. In this article, we introduce a novel approach that allows identifying similarities between diseases using GWAS data. Our approach is based on training a classifier that distinguishes between a reference disease and a control set, and then using this classifier for comparing several query diseases to the reference disease. This approach is based on the classification of individuals using their full genotype, and is thus different from previous work in which the independent statistical significance of each SNP is used for comparing diseases.\n\nWe apply this approach to the genotype data of seven common diseases provided by the WTCCC, and show that we are able to identify similarities between diseases. We replicate the known finding that there is a common genetic basis for T1D and RA, find strong evidence for genetic similarities between BD and HT, as well as evidence for genetic similarities between T1D and both BD and HT. We also find similarities between one of the control sets used in the WTCCC (*UKBS*) and two disease sets, RA and CAD. This similarity can possibly be a consequence of the subtle differences in genotyping quality that were observed during the initial quality control performed by the WTCCC.\n\nOur results demonstrate that it is possible to use a classifier-based approach to identify genetic similarities between diseases, and more generally between multiple phenotypes. We expect that this approach can be improved by using classifiers that are more specifically tailored for the analysis of GWAS data, and by the integration of a larger number of disease phenotypes. The ability to compare similarities between diseases at the whole-genome level will likely identify many more currently unknown similarities. Genetic similarities between diseases provide new hypotheses to pursue in the investigation of the underlying biology of the diseases, and have the potential to lead to improvements in how these diseases are treated in the clinical setting.\n\n6 DATA {#SEC6}\n======\n\nWe use the individual genotypes provided by the WTCCC. These genotypes come from a GWAS (WTCCC, [@B17]) of seven common diseases: T1D, T2D, CAD, CD, BD, HT, and RA. The data consist of a total of 2000 individuals per disease and 3000 shared controls, with 1500 control individuals from the 1958 British Birth Cohort (58C control set) and 1500 individuals from blood donors recruited specifically for the project (UKBS control set). The genotyping of 500 568 SNPs per individual was performed using the Affymetrix GeneChip 500 K Mapping Array Set. In the original analysis of this dataset by the WTCCC, a total of 809 individuals and 31 011 SNPs that did not pass quality control checks are excluded. In addition, SNPs that appear to have a strong association in the original study have been manually inspected for quality issues, and 578 additional SNPs were removed. In this work, we exclude all individuals and SNPs that were excluded in the WTCCC study, as well as an additional 9881 SNPs that do not appear in the WTCCC summary results.\n\nOne concern with these quality control steps is the identification of SNPs for which the genotype calling is of poor quality. In the WTCCC study, this is done after the analysis, which makes it possible to visually inspect the small subset of SNPs that are potentially significant. In a classifier-based approach, it is impractical to perform any kind of visual inspection, and we must try to minimize the errors due to genotype calling prior to the analysis. The WTCCC study only uses genotype calls made by a custom algorithm, Chiamo (Marchini *et al.*, in preparation), but the genotype calls made using the standard Affymetrix algorithm BRLMM are also available. While the study does show that Chiamo has, on average, a lower error rate than BRLMM, there are SNPs that are discarded during the quality control process that show errors in the genotype calls made by Chiamo. We use the two genotype sets to create a consensus dataset in which the genotype of a given individual at a given SNP is used only if there is agreement between the call made by Chiamo and the call made by BRLMM, and is considered to be unknown if the calls are different. This approach individually considers the call made for every individual at every SNP, and does not discard entire SNPs. The handling of SNPs that have a high proportion of unknown genotypes is left to the classification algorithm, and will be discussed in the corresponding section. While this approach does reduce the errors in genotype calling, this comes at the cost of discarding cases in which Chiamo is right but BRLMM is not. Overall, the frequency of unknown genotypes is 2% using the consensus approach, compared with 0.65% using Chiamo and 0.74% using BRLMM. Furthermore, BRLMM genotype calls are entirely missing for a total of 184 individuals, which are thus excluded from our study.\n\nAfter performing these preprocessing steps, the data set used in this study consists of 459 075 SNPs measured in 2938 control individuals (58C: 1480, UKBS: 1458), 1963 with T1D, 1916 individuals with T2D, 1882 individuals with CAD, 1698 individuals with CD, 1819 individuals with BD, 1952 individuals with HT and 1834 individuals with RA.\n\n7 METHODS {#SEC7}\n=========\n\nIn this section, we first formally define the classification task that is central to our approach, then describe the specific classifier we use in this work and how we evaluate its performance, and finally describe how we use the classification results to infer relationships between diseases.\n\n7.1 Classification Task {#SEC7.1}\n-----------------------\n\nThe data consist of a list of individuals *i*, a list of SNPs *s* \u2208 *S*, and the measurement of the genotype *g*(*s*, *i*) of individual *i* at SNP *s*. We use *G*~*i*~={*g*(1, *i*),...,*g*(\\|*S*\\|, *i*)} to denote the genotype of individual *i* at all the SNPs in the study. The genotype measurement is a discrete variable which can take four values: homozygote for the major allele, homozygote for the minor allele, heterozygote and unknown: *g*(*s*, *i*)\u2208{*maj*, *min*, *het*, *unk*}. Each individual belongs to one of several disease sets, or to the control set. For the WTCCC data used in this work, we have seven disease sets: *T1D*, *T2D*, *CAD*, *CD*, *BD*, *RA*, *HT*, and we use the union of the *58C* and *UKBS* sets as control set.\n\nFor each disease *d*, we train a classifier that distinguishes between that disease set and the controls. The individuals that are not part of these sets are ignored during the training of this classifier. For each individual *i* used during training, a binary class variable *c*~*i*~ indicates whether the individual belongs to the disease set (*c*~*i*~==*disease*) or to the control set (*c*~*i*~==*control*). The supervised classification task consists of predicting the class *c*~*i*~ of an individual *i* given its genotype *G*~*i*~. In this work, we use a decision tree classifier, but any algorithm able to solve this classification task can be easily integrated into our analysis pipeline.\n\n7.2 Decision trees {#SEC7.2}\n------------------\n\nIn this section, we describe the decision tree classifier (Breiman *et al.*, [@B1]). We use cross-validation in order to train the classifier, prune the trained decision tree and evaluate its performance on distinct sets of individuals.\n\nWe train a decision tree *T* by recursively splitting the individuals in each node using maximum information gain for feature selection. We use binary categorical splits, meaning that we find the best rule of the form *g*(*s*, *i*)==\u03b3, where \u03b3\u2208{*maj*, *min*, *het*}. Binary splits make it possible to handle cases in which only one of the three possible genotypes is associated with the disease without unnecessarily splitting individuals that have the two other genotypes. Unknown values are ignored when computing information gain. This is necessary since there is a correlation between the frequency of unknown values and the quality of the genotyping, which in turn is variable between the different datasets. Counting unknown values during training could therefore lead to classifiers separating the two sets of individuals based on data quality differences, rather than based on genetic differences. However, if a large number of measurements are unknown for a given SNP, the information gain for that SNP will be biased. This is particularly true if the fraction of unknowns is very different between the cases and the controls. In order to avoid this situation, we discard all SNPs that do have \\>5% of unknown genotypes amongst the training individuals in the node we are splitting. In each leaf node *L*, we compute the fraction *f*~*L*~ of training individuals in that node of that are part of the disease class: .\n\nIn order to choose a pruning algorithm, we compare the cross-validation performance obtained using Cost-Complexity Pruning (Breiman *et al.*, [@B1]), Reduced Error Pruning (Quinlan, [@B13]), as well as a simple approach consisting of limiting the tree depth. We find that Reduced Error Pruning outperforms Cost-Complexity Pruning, and performs similarly well than limiting the tree depth, but results in smaller decision trees. We therefore use Reduced Error Pruning, which consists of recursively eliminating subtrees that do not improve the classification error on the pruning set (which only contains individuals that were not used during training).\n\nThe classification of an individual *i* using a decision tree *T* is done by traversing the tree from the root towards a leaf node *L*(*i*) according to the genotype of the individual which is classified. If *f*~*L*(*i*)~ \\>0.5, then the individual is classified as *disease*, else the individual is classified as *control*. We can consider the decision tree *T* as a high-level statistical model of the difference between the disease and the control sets. Under this model, the fraction *f*~*L*(*i*)~ represents the conditional probability of individual *i* to be part of the disease class given its genotype: *P*~*T*~(*c*~*i*~==*disease* \\| *G*~*i*~)=*f*~*L*(*i*)~. This value is the *disease-class probability* of individual *i*. In order to compute the fractions *f*~*L*~ over sufficiently large numbers of individuals, we further prune our tree to only have leaf nodes containing at least 100 training individuals. The benefit of using this probability rather than the binary classification is that it allows to distinguish leaf nodes in which there are mainly training individuals from one class from those in which both classes are almost equally represented.\n\nIn order to assess the performance of our classifier, we perform 5-fold cross-validation. We start by separating the data into five random sets containing 20% of the individuals each. A decision tree *T* is trained using four of these sets, while one set is reserved for pruning and testing. The unused set is split randomly into two equal sets. The first of these sets is used to obtain pruned tree *T*\u2032 from tree *T*, and the individuals in the second set are used to evaluate the performance of tree *T*\u2032. The last step is then repeated using the second set for pruning, and the first for testing. Finally, we repeat the training and evaluation four more times, each time leaving out a different set for pruning and testing. This ensures that for every individual in our dataset, there is one pruned decision tree for which the individual was used neither for training nor for pruning. We can therefore evaluate the performance of the classifier on unseen data. We can also compute the average disease-class probability *p*(*C*) of the control individuals, and the average disease-class probability *p*(*d*) of the individuals with disease *d*. The difference \u0394~*p*~ between those two probabilities indicates how well the classifier is able to distinguish controls from diseases. We use the cross-validation results to compare the performance of the classifier against a baseline classifier which simply assigns the most frequent label amongst the training set to all individuals. Classifiers that do not outperform this baseline classifier, or for which the difference \u0394~*p*~ is small, are not used to identify similarities between diseases.\n\nGiven the cross-validation scheme used, we end up training not one, but several possibly distinct decision trees. Rather than arbitrarily choosing one, we use the set *T*~*d*~ of all decision trees trained during cross-validation for a given disease *d*. In order to classify a new individual *i*, we first classify *i* using each classifier independently, and then return the average classification. Similarly, we average the results of individual classifiers to obtain the average disease-class probability: .\n\n7.3 Identifying similarities {#SEC7.3}\n----------------------------\n\nOnce a classifier has been trained to distinguish the set of individuals with reference disease *d* from the control set, we can use it to identify diseases that are similar to disease *d*. Using the classifier, we can compute the disease-class probability of an individual with a query disease *d*\u2032. In order to be able to compare diseases, we are interested in computing the average disease-class probability of all individuals in *d*\u2032: . We expect this average probability to be in, or close to the interval between *p*(*C*) and *p*(*d*), which were the averages computed on, respectively, the control set and the disease set *d* during cross-validation. If *p*(*d*\u2032) is close to *p*(*C*), then *d*\u2032 is not very different from the control set, whereas a value *p*(*d*\u2032) that is close to *p*(*d*) indicates similarity between the two diseases. Using this method, we can compare all query diseases to the reference disease *d*, and identify if there are diseases that are more similar to *d* than others.\n\nIf we find that a query disease *d*\u2032 is closer to reference disease *d* than the other query diseases, then we need to assess the significance of this finding. In order to do so, we randomly sample a set *r* of individuals from all the disease sets except *d*, such that *r* is of the same size as *d*\u2032, and compute *p*(*r*). We repeat this procedure 10 000 times. The fraction of random samples *r* for which *p*(*r*) \u2265*p*(*d*\u2032) indicates how often a random set of individuals would obtain a probability of being part of the disease-class at least as high as the set *d*\u2032, and is therefore a *P*-value indicating how significant the similarity between *d*\u2032 and *d* is.\n\nThis study makes use of data generated by the Wellcome Trust Case-Control Consortium. A full list of the investigators who contributed to the generation of the data is available from [www.wtccc.org.uk](www.wtccc.org.uk).\n\nFunding for the project was provided by the Wellcome Trust under award 076113; Richard and Naomi Horowitz Stanford Graduate Fellowship (to M.A.S.).\n\n*Conflict of Interest*: none declared.\n\n[^1]: *Baseline* corresponds to the baseline error; *Error*, *Precision* and *Recall* to the cross-validation performance of the decision tree classifier; \u0394~*p*~ to the difference between the average disease-class probability of the control set and the average disease-class probability of the disease set; and *Leaves* to the maximum number of leaves in the pruned classifiers for this disease.\n\n[^2]: *Baseline* corresponds to the baseline error; *Error*, *Precision* and *Recall* to the cross-validation performance of the decision tree classifier; \u0394~*p*~ to the difference between the average disease-class probability of the control set, and the average disease-class probability of the disease set; and *Leaves* to the maximum number of leaves in the pruned classifiers for this experiment. *R1* and *R2* represent two random splits of the joint control set.\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nLymphoma is a hematological malignancy treated with multiagent chemo-immunotherapy, radiation therapy, and other biological agents. Lymphoma patients receiving these therapies often report fatigue as one of the most distressing symptoms, even worse than other symptoms such as pain, nausea, and vomiting, which are usually treatable with medications. They suffer 2.5--3 times more fatigue than the general population.\\[[@ref1]\\] Fatigue leads to decline in physical activity during the period of chemotherapy treatment. The chemotherapeutic medications are likely to negatively affect the health-related quality of life (HRQOL).\\[[@ref2]\\]\n\nOver the past few years, yoga has emerged as an effective intervention in improving overall well-being. It has been reported to improve sleep and reduce anxiety and depression, thus improving overall quality of life among breast cancer patients.\\[[@ref3]\\]\n\nOne limitation for integration of yoga in mainstream oncology practice is lack of availability of a trained yoga instructor to every patient. Often, cancer patients are advised isolation in post-chemo phase. Therefore, they cannot attend community yoga program during chemotherapy treatment. In the current era, access to audio-visual (AV) devices is very easy even in rural parts of our country. Hence, this study was designed to assess feasibility and safety of home-based yoga intervention administered through AV devices and printed guidebooks after a direct supervised yoga session.\n\nSubjects and Methods {#sec1-2}\n====================\n\nStudy design and settings {#sec2-1}\n-------------------------\n\nThis was a single-arm pre- and postdesign interventional study conducted at Postgraduate Institute of Medical Education and Research in Chandigarh, India.\n\nAs a feasibility pilot study, a total of 14 patients diagnosed with malignant lymphoma, planned to receive chemotherapy were recruited and study intervention was carried out during first 2 months of chemotherapy (September to November 2016). The inclusion criteria were: (a) aggressive lymphoma patients (Stage II--IV) planned to receive any of the three chemotherapy schedules, i.e., cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP); CHOP + rituximab; adriamycin, bleomycin, vinblastine, and Dacarbazine(ABVD); (b) age group of 18--65 years; (c) Eastern Cooperative Oncology Group (ECOG) performance status score 0--2 (If the ECOG score was \\>2 at the time of starting chemotherapy, then these patients were reassessed at the time of beginning of 2^nd^ chemotherapy cycle). In case, the score improved to ECOG \\<2, then these patients were also offered to participate in the study; (d) participants having access to AV display device (CD player, multimedia/smartphone, and laptop); (e) able to complete questionnaire (by self or by interview) in Hindi, Punjabi, or English. Exclusion criteria were Stage I lymphoma patients, ECOG performance score of 3--4 even at 2^nd^ chemotherapy cycle.\n\nFeasibility of yoga-based intervention in lymphoma patients {#sec2-2}\n-----------------------------------------------------------\n\nThe feasibility criteria were taken from a yoga feasibility study done in Japan among breast cancer patients.\\[[@ref4]\\] It included recruitment, retention, and acceptability rates which were defined as:\n\nRecruitment rate: The rate of consenting patients divided by the number of eligible patients approached to participate in the studyRetention rate: Number of recruited participants who completed pre and postintervention questionnairesAdherence was assessed using the self-reported number of completed yoga sessions over a period of 2 months, if participants practiced a complete session in a day; it was counted as one yoga session. 20 sessions/month were taken as good adherenceAcceptability of the intervention was assessed after completion of the program using a 3-point Likert scaleSafety: Participants were asked to report untoward problem that occurred due to yoga practice in their yoga log books.\n\nStudy tools {#sec2-3}\n-----------\n\nWell-validated and reliable tools were used in the study:\n\nSociodemographic profile was assessed on the day of registration in the clinicPhysical performance status was assessed using ECOG performance scale\\[[@ref5]\\]HRQOL and fatigue were assessed using lymphoma-specific instruments, i.e., functional assessment of cancer therapy-lymphoma scale\\[[@ref6]\\]Functional assessment of chronic illness therapy-Fatigue scale\\[[@ref7]\\]Quality of sleep was measured with Pittsburgh sleep quality index\\[[@ref8]\\]Depression was assessed using Patient Health Questionnaire-9\\[[@ref9]\\]Generalized anxiety disorder-7 tool (GAD-7) for measuring anxiety\\[[@ref10]\\]Numerical pain rating scale to determine intensity of pain.\\[[@ref11]\\]\n\nThe participants were asked to report the adverse events occurring due to yoga practice on their yoga log books.\n\nIntervention {#sec2-4}\n------------\n\nYoga protocol was taken from the common yoga protocol followed by Government of India on the occasion of \"International Yoga Day\" 2016.\\[[@ref12]\\] It included warm-up exercises, Asanas, and Pranayam as detailed in [Table 1](#T1){ref-type=\"table\"}. At the time of recruitment, participants were taught and then supervised yogic exercises on one-to-one basis by the certified yoga instructor (first author). Once this author was convinced with their practice, they were asked to do same yogic exercises daily at home for 30 min. Yoga video and booklets in different languages guided them in the same. Adherence was measured through yoga logbooks \\[[Table 2](#T2){ref-type=\"table\"}\\] maintained by the participants at their home. Biweekly telephonic reinforcement was done to countercheck and enhance adherence.\n\n###### \n\nVarious yogic exercises included in the Yoga protocol\n\n![](IJY-11-249-g001)\n\n###### \n\nPerforma on adherence (please put \u221a for particular activity done and \u00d7 for not done)\n\n![](IJY-11-249-g002)\n\nStudy procedure {#sec2-5}\n---------------\n\nEthical approval was obtained from the Institute\\'s Ethical Committee. Informed consent was obtained from all participants. In orientation phase, every participant was informed about the yogic exercises to be done at home during the chemotherapy period. Subsequently, they were taught yogic exercises and maintenance of daily yoga logbook. A yoga booklet and video were also provided to them. Biweekly telephonic reinforcement emphasized about protocol adherence. Follow-up was done at 2^nd^ month of commencement of the program.\n\nStatistical analysis {#sec2-6}\n--------------------\n\nStatistical Package for the Social Sciences (IBM Corp.SPSS version 22. Armonk, New York) was used for data analysis. Descriptive statistics was done for socioeconomic characteristics of the participants. Paired *t*-test was applied to check the difference between baseline and postintervention mean scores.\n\nResults {#sec1-3}\n=======\n\nRecruitment and participant\\'s characteristics {#sec2-7}\n----------------------------------------------\n\nAfter assessing 25 patients for eligibility, 10 patients were found ineligible to participate in the study (ECOG performance score was 3--4 even at 2^nd^ chemotherapy cycle). One patient refused to participate due to personal reasons; finally, 14 patients were included in the study.\n\nTables [3](#T3){ref-type=\"table\"} and [4](#T4){ref-type=\"table\"} show the summary of a sociodemographic and clinical profile of the participants, respectively.\n\n###### \n\nBaseline characteristics of the participants (*n*=14)\n\n![](IJY-11-249-g003)\n\n###### \n\nParticipant's clinical details and adherence in terms of number of days of performance of yogic exercises during 1st and 2nd month of chemotherapy\n\n![](IJY-11-249-g004)\n\nReported access of audio-visual gadgets and booklet by the patients {#sec2-8}\n-------------------------------------------------------------------\n\nExcept one, all participants reportedly used yoga video as a guide to yoga practice at home (11 watched it on smartphone, 1 on CD player, and 2 on laptop). Thirteen literate participants reportedly read the yoga booklet at home.\n\nFeasibility of the yoga program {#sec2-9}\n-------------------------------\n\nRecruitment rate was 93% and retention rate was reported to be 100%. All the participants maintained the Yoga Log book daily at home. [Table 4](#T4){ref-type=\"table\"} shows reported adherence to yoga by the participants during 1^st^ and 2^nd^ month of chemotherapy. [Table 5](#T5){ref-type=\"table\"} shows reported adherence to the various components of the yogic exercises.\n\n###### \n\nGood adherence\\* in terms of number of participants practicing various components and complete yoga session (*n*=14)\n\n![](IJY-11-249-g005)\n\nSide effects of chemotherapy such as fatigue, pain in upper extremity where chemotherapy was administered, nausea and vomiting, fever, and diarrhea along with the absence of companion, and low mood were cited as the most frequent reason for skipping the yoga practice at home. The improvement was observed in all secondary outcome variables except pain \\[[Table 6](#T6){ref-type=\"table\"}\\]. At the end of 2^nd^ month of intervention, all participants completed intervention acceptability questionnaire and acceptability rate was found to be 97% \\[[Table 7](#T7){ref-type=\"table\"}\\].\n\n###### \n\nEffects of Yoga on patient.rated outcome at baseline (before start) and 2^nd^ month of chemotherapy\n\n![](IJY-11-249-g006)\n\n###### \n\nAcceptability of the yoga program (*n*=14)\n\n![](IJY-11-249-g007)\n\nAdverse events {#sec2-10}\n--------------\n\nNo yoga-related serious adverse event was reported by the patients during the study period.\n\nDiscussion {#sec1-4}\n==========\n\nConsidering the good 100% retention rate, 93% recruitment rate, 97% acceptability rate, 78.6% adherence rate, and no \"yoga\" related serious adverse events during chemotherapy, it can be suggested that home-based yoga intervention among lymphoma patients undergoing chemotherapy is feasible and safe.\n\nAdherence to the warm-up exercises was good, i.e., 100% in both 1^st^ and 2^nd^ month, and it improved for Asanas from 57% to 78.6% and for Pranayam from 86% to 100% in the 2^nd^ month as compared to the 1^st^ month of chemotherapy. It is understandable that Yogic Asanas are more complex than warm-up exercises. They need slightly more physical fitness than the warm-up activity. In our study, we believe that, after one/two chemotherapy, cycle reduction in tumor burden might have led to improvement in physical condition of the patients which would have enabled them to perform Yogic Asanas in an easier manner than at the time of recruitment. Another possible explanation for better performance is the knowledge that chemotherapy-related adverse events are more likely in the first few weeks of starting of chemotherapy and they may hinder any kind of physical activity.\\[[@ref13]\\] We feel that good adherence in our study was may be due to cultural aspects of our study population, as acceptance for Yoga is higher in Indian culture.\n\nOur intervention program was home based, so it was more convenient for the participants to comply with it. Further, patients undergoing chemotherapy are neutropenic so are extremely susceptible to get infection which may be of serious life-threatening infections in nature.\\[[@ref13]\\] Hence, practicing yoga at home is preferable for this highly immunosuppressed cohort.\n\nOne of the hindrances in use of yoga with the modern medicine is lack of availability of trained yoga instructor at every place, especially semi-urban/rural areas do not have access to trained yoga experts. On the contrary, these days, electronic sources such as smartphone/laptops are readily available even in the rural settings. The results of this study show that the use of these gadgets can help patients to practice yoga at home.\n\nThe secondary outcomes, for example, quality of life, fatigue, anxiety, depression and sleep also improved in our study and these results are consistent with the result of other similar studies done in cancer survivors.\\[[@ref14][@ref15]\\] Hence, yoga-based programs should be considered in the management of lymphoma patients.\n\nAs this was a feasibility study for home-based yoga intervention, however, it needs further validation about improvement in HRQOL, fatigue, sleep, anxiety, and depression by taking up future randomized study on a larger study sample.\n\nConclusion {#sec1-5}\n==========\n\nThe results of the present study support the safety and feasibility of yogic interventions in patients diagnosed with lymphoma undergoing multiagent chemotherapy.\n\nFinancial support and sponsorship {#sec2-11}\n---------------------------------\n\nNil.\n\nConflicts of interest {#sec2-12}\n---------------------\n\nThere are no conflicts of interest.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nIntegrated surveillance is considered a promising working strategy \\[[@B1]--[@B4]\\] to enhance early warning of emerging infections such as arboviral diseases. In addition to providing early signals, integrated surveillance by systematically integrating multiple sources of surveillance data in a timely manner (indicator- and event-based surveillance, case-based surveillance, vector surveillance, and virus and environmental data and information) could contribute more effectively to accurate risk assessments (RA) \\[[@B5]\\]. Unfortunately, very few countries worldwide \\[[@B6]--[@B8]\\], and in the Mediterranean Region \\[[@B9][@B10]\\], have managed to collect and analyse surveillance data across sectors related to arbovirus transmission, and even fewer have interoperable databases. Ultimately, this limits early warning and risk assessment capacity with impact on the prevention and control of arbovirus infections. This is in line with the recognised challenges of sharing data and information, although the evidence for the public health benefits of sharing is growing with well-documented instances of an improved outcome as a result of sharing surveillance data \\[[@B11]--[@B16]\\]. Efficient data sharing also prompted an early response to the emergence of the H7N9 influenza virus in humans in China \\[[@B14]\\]; conversely, reluctance to share can hinder or slow down the response and global outbreaks have shown that inadequate surveillance and response capacity in a single country can endanger national populations and the public health security of the entire world \\[[@B13]\\]. One relevant issue is how to enhance trust within and between countries, considering that trust facilitates the sharing of data and information. Trust-building measures can take the form of face-to-face meetings, regular regional workshops, desktop exercises, joint outbreak investigations, and networking activities. These promote the sense of working towards a common goal \\[[@B11]\\].\n\nTo this aim, working with a multisectoral and transdisciplinary approach, often mentioned as a One Health approach, can help to mediate different assumptions and views and to fill knowledge gaps \\[[@B17]--[@B20]\\].\n\nFocusing specifically on those threats which occur at the animal-human-ecosystem interfaces, several international organisations, including the World Health Organization (WHO), Food and Agriculture Organization of the United Nations (FAO), World Organisation for Animal Health (OIE), and World Bank, have recognised the critical role of multisectoral risk assessment (MRA) (multisectoral risk assessment (MRA): assessment with the concomitant participation of all the relevant sectors involved in the surveillance of a given arbovirus infection) to enhancing cross-sectoral collaboration and improving data collection and data-sharing from different sectors \\[[@B21]--[@B24]\\].\n\nIn fact, for health threats that are either emerging or existing at the interface, including food safety issues, neither the technical data nor other information important to conduct a comprehensive assessment nor the appropriate breadth of technical expertise and experience are routinely available within a single agency or sector \\[[@B22]\\].\n\nIn the dimension of capacity building and training, risk assessment exercises implemented with a multisectorial approach can foster data and information sharing across sectors reducing information gaps, highlight experiences and contributions across countries, develop the concept of a national/regional \"cross-sectoral\" risk assessment outcome, and guide prioritisation of actions and allocation of funds also taking into account the cross-border dimension.\n\nIn fact, regional public health threats are often presenting common characteristics such as the need of joint prevention and response activities, common coordination, and comprehensive lessons learned analysis across the actors involved, especially at borders (see the cases of Crimean--Congo haemorrhagic fever cluster in 2008 at the borders between Greece and Bulgaria \\[[@B25]\\] and in 2009 between Georgia and Turkey \\[[@B26]\\]).\n\nThese characteristics can be easily integrated in the framework of MRA.\n\nTowards this aim, we organized three MRA exercises: one on West Nile virus (WNV) infection, one on Crimean--Congo haemorrhagic fever (CCHF), and one on Rift Valley fever (RVF) in the framework of the MediLabSecure (MLS) project \\[[@B27]\\].\n\nThe aim of these exercises was not only to formulate more reliable risk assessments but also to promote a process leading to a homogenous understanding of risk across different sectors in a given country, and across neighbouring countries, using a structured strategy of assessment. This article describes their implementation and discusses the added value of the adopted multisectoral approach.\n\n2. Materials and Methods {#sec2}\n========================\n\nThe MLS project started in 2014 and aims at consolidating a regional network of public health institutions and laboratories, belonging to 19 non-European Union (EU) countries (Albania, Algeria, Armenia, Bosnia and Herzegovina, Egypt, Georgia, Jordan, Kosovo, Lebanon, Libya, Moldova, Montenegro, Morocco, Palestine, Former Yugoslav Republic of Macedonia, Serbia, Tunisia, Turkey, and Ukraine), for the control of zoonotic emerging viruses. It represents a cluster for awareness, risk assessment, surveillance, monitoring, and control of relevant emerging diseases, with special focus on arbovirus infections.\n\nIn this context, we designed three MRA exercises in coordination with the MLS working group and the subject-matter experts of the European Centre for Disease Prevention and Control (ECDC) and of the Italian Animal Health Institute \"Istituto Zooprofilattico Sperimentale dell\\'Abruzzo e del Molise (IZSAM).\"\n\nFor the development of the three MRA exercises, we relied on the following existing RA methodology and guidance documents: the ECDC \"WNV risk assessment tool\" \\[[@B28]\\], the ECDC \"operational guidance on rapid risk assessment (RRA) methodology\" \\[[@B29]\\], and the Food and Agriculture Organization of the United Nations (FAO) methodology of \"The RVF in Niger: Risk Assessment\" \\[[@B30]\\]. All mentioned tools and guidance documents were developed by subject-matter experts, had been piloted in other contexts, and were in line with the pathogens and methodological priorities identified by the MLS countries.\n\nWe invited country representatives of the human virology, animal virology, medical entomology, and public health sectors (hereby called \"*sectors*\") involved in the surveillance of vector-borne diseases to participate in the three MRA exercises. Background documentation (including selected references) was sent by e-mail to participants one week before each exercise. An exercise implementation guide was also developed and sent to the facilitators together with the background documentation. The participants were asked to send national epidemiological data on the concerned pathogens that were then shared with all participants. At the start of each exercise session, participants were provided with a participant\\'s guide.\n\nAll three exercises developed in three phases, the first always consisted in technical and methodological presentations by subject-matter experts. The second and third phases differed as shown in [Table 1](#tab1){ref-type=\"table\"}.\n\nAdditional details on the developed exercises, background documents, and guidance for facilitators and participants are available in the MRA exercise reports \\[[@B32]--[@B34]\\].\n\nThe added value of the multisectoral approach during the CCHF and the RVF assessments was collected by asking each participant to rank (high, medium, or low) the level of perceived benefit of the multisectoral collaboration when answering each \"risk question\" included in the exercises.\n\nPre- and posttest questionnaires, designed to assess if the MRA had increased the participant\\'s knowledge, were prepared and submitted for the CCHF (Annex 5 of \\[[@B33]\\]) and RVF (Annex 6 of \\[[@B34]\\]) exercises. We deemed that, considering the aim of the exercises, it would have been particularly important to assess knowledge of participants on key parameters on which to rely on for the assessment, notably, surveillance data, source, and type of information and disease/infection risk factors.\n\nParticipants were also asked to compile an exercise evaluation form (Annex 6 of \\[[@B34]\\]) at the end of each exercise to provide MediLabSecure project with feedback on the quality and the pertinence of the training sessions.\n\n2.1. The WNV Exercise {#sec2.1}\n---------------------\n\nAll the 19 countries involved in the MLS network took part in the exercise ([Table 1](#tab1){ref-type=\"table\"}). The participants were divided in smaller groups by country according to regional proximity.\n\nEach participant was asked to identify the risk area typology that was mostly representative of his/her country on the basis of the six risk area types defined by ECDC for WNV transmission ([Figure 1](#fig1){ref-type=\"fig\"}).\n\nSubsequently, the participants discussed the reasons for their identified risk area in groups, considering both national and cross-border factors. They were allowed to modify their risk area after the discussion. Then, the participants discussed in country groups the level of risk with regard to national surveillance system characteristics using the SWOT \\[[@B35]\\] analysis framework (strengths, weaknesses, opportunities, and threats analysis) guided by the ECDC tool \\[[@B28]\\]. The final risk area typology and the main aspects that had emerged from all the national SWOT analyses were presented and discussed in plenary with all other groups \\[[@B31]\\].\n\n2.2. The CCHF Exercise {#sec2.2}\n----------------------\n\nCCHF MRA was implemented with the countries of the Balkans and Black Sea Region of MLS ([Table 1](#tab1){ref-type=\"table\"}) that considered this disease as a priority for the area. The exercise was developed by adapting the information table for rapid risk assessment and the risk-ranking algorithm of the ECDC operational guidance on rapid risk assessment methodology (Annex 2 and 3 in \\[[@B33]\\]) to rate the potential of CCHF virus transmission in each participating country integrating the views of the different sectors. The assessment was done in two steps: first, the participants assessed the risk in small groups of neighbouring countries on the basis of the information delivered with the technical presentations, available national data, and the background document sent in advance; second, an assessment was made by each country over the different sectors. Each country provided the multisectoral added value to the rapporteur for plenary audience restitution.\n\n2.3. The RVF Exercise {#sec2.3}\n---------------------\n\nThe RVF exercise was implemented with the countries of North Africa and the Middle East Region of MLS ([Table 1](#tab1){ref-type=\"table\"}) which considered RVF a priority for the area. The RVF exercise was developed by adapting the risk questions of the FAO RVF in Niger Risk assessment (Annex 3 in \\[[@B34]\\]) to identify the risk of RVF virus infection introduction, spread and/or persistence in each participating country. As for the CCHF exercise, the participants were divided in small groups of neighbouring countries to discuss the regional situation with the colleagues of the other sectors in the group.\n\nFor the last phase, the group was divided by country with all sectors represented because the expected outcome was the level of risk by country. Each country provided the multisectoral added value to the rapporteur for plenary audience restitution.\n\n3. Results {#sec3}\n==========\n\nA total of 159 participants from the 19 non-EU countries of the MLS network took part in the three exercises: 73 participants in the WNV, 42 in the CCHF, and 44 in the RVF exercise.\n\n3.1. The WNV Exercise {#sec3.1}\n---------------------\n\nThe WNV exercise highlighted a high heterogeneity in assessing the level of risk across the involved sectors. The sharing of information and discussion between sectors and neighbouring countries reduced intersectoral variability towards a single level of risk in each country.\n\nEach participant was provided with dots coloured as per his/her sector (i.e., yellow for human virology, blue for animal virology, green for medical entomology, and red for public health), and these dots were used to mark the identified risk area on a wall poster.\n\nAs an example, we report here the outcomes of two groups. In [Figure 2](#fig2){ref-type=\"fig\"}, country 1 assessed risk level 5 (affected risk area), country 2, risk level 2 (imperilled risk area), and country 3, risk level 1 (predisposed risk area) with final good agreement between different sectors within country. In [Figure 3](#fig3){ref-type=\"fig\"}, countries 1 and 2 assessed risk level as 1 and 2, respectively, without final agreement between different sectors in one country.\n\nThe SWOT analysis underlined the critical role of integrated surveillance systems, laboratory capacity, and intersectoral collaboration for reliable risk assessments of arbovirus infections. The implementation of the first MRA exercise on WNV highlighted the need for enhancing the collaboration between sectors to reduce heterogeneity in risk assessment and for analysing the added value of a multisectoral approach.\n\n3.2. The CCHF Exercise {#sec3.2}\n----------------------\n\n### 3.2.1. Knowledge and Capacity {#sec3.2.1}\n\nThe results of the pre- and posttests completed by thirty-five (83%) participants of CCHF exercise showed that the exercise led to improvements in the capacity to determine *risk factors* and to identify *sources of reliable information* to assess the risk. For example, with reference to the question of the test \"*Would CCHF be an unusual or unexpected threat in your country?*\"10 out of 35 (29%) of the respondents replied \"yes\" in the pretest, while in the posttest, all the respondents (35) replied \"no\" to this question. This suggests that the discussion between countries and the assessment exercise helped to identify possible risk factors also at cross-border or regional level (i.e., knowledge that neighbouring countries host the pathogen).\n\nRegarding documentation for risk assessment, we reported in [Table 2](#tab2){ref-type=\"table\"} the documents mentioned by the participants to assess the level of risk for CCHF in their country.\n\n### 3.2.2. The Added Value of the Multisectoral Approach {#sec3.2.2}\n\nThe added value of the concomitant participation of several sectors to the RA for each risk question of the exercise is reported in [Figure 4](#fig4){ref-type=\"fig\"}. These specific aspects related to the added value of the exercise were considered particularly relevant by the project\\'s stakeholders and therefore reported in the MediLabSecure Strategic Document \\[[@B31]\\] for further developments.\n\nThe multisectoral approach was assessed as particularly valuable in \"setting the scene\" and in analysing comprehensively the situation having access to information and knowledge provided by each of the sectors involved in the exercise (see the added value for risk questions 1 and 5 in [Figure 4](#fig4){ref-type=\"fig\"} and data analysis in additional file 1).\n\n3.3. The RVF Exercise {#sec3.3}\n---------------------\n\n### 3.3.1. Knowledge and Capacity {#sec3.3.1}\n\nThe results of the pre- and posttests, completed by twenty-one (48%) participants of the RVF exercise, showed that the exercises led to improvements in the capacity to determine risk factors. Although the participants were all able to identify several relevant risk factors, some specific risks were only identified in the posttest. Among them \"*animal movements*\" included by 11 (52%) and 10 (48%), as relevant risk of spread of the virus in endemic and new areas, respectively, \"*social and economic instability*\" included by six (29%) both as relevant risk of endemic and new areas, and \"*climate changes*\" included by eight (38%) and seven (33%) as relevant risk of endemic and new areas, respectively. In relation to \"list kind of documents to rely on to assess the level of risk for RVF in your country,\" in total, 18 (88%) and 19 (90%) of participants were able to mention kind of documents useful for RA of RVF in their countries in pre- and posttest, respectively.\n\n3.4. The Added Value of the Multisectoral Approach {#sec3.4}\n--------------------------------------------------\n\nThe country perception of the added value of the multisectoral approach is reported in [Figure 5](#fig5){ref-type=\"fig\"}. Also, for this exercise, the multisectoral approach was particularly valuable in \"setting the scene\" and in analysing comprehensively the situation having access to wide range of information and knowledge provided by each of the sectors involved in the exercise (see the added value for risk questions 3, 4, and 6 in [Figure 5](#fig5){ref-type=\"fig\"} and data analysis in additional file 1). As for CCHF and also for RVF, the aspects related to the added value of the exercise were considered particularly relevant by the project\\'s stakeholders and therefore reported in the MediLabSecure Strategic Document \\[[@B31]\\].\n\n3.5. Results of the Evaluation of Three Exercises {#sec3.5}\n-------------------------------------------------\n\nResponse rates to the evaluation questionnaire were 90% (66/73), 88% (37/42), and 68% (30/44) for WNV, CCHF, and RVF exercises, respectively.\n\nOverall, 92% (WNV), 94% (CCHF), and 83% (RVF) of respondents found the exercise satisfactory. Ninety percent or more of respondents for each exercise found the discussion between sectors useful to identify the level of risk.\n\nAlmost all respondents reported that the objectives of the exercises were clearly communicated (99% for the WNV MRA exercise, 89% for the CCHF MRA, and 83% for the RFV MRS), while agreement on the appropriateness of the time allotted for the exercises was 92% for both WNV and CCHF and just 54% for the RVF exercise (see data analysis in additional file 1).\n\n4. Discussion {#sec4}\n=============\n\nAs reported, the main aims of these exercises were to increase knowledge on MRA and raise awareness of multisectoral collaboration for conducting risk assessment of arbovirus infection with a One Health approach in the Mediterranean region. Using available tools and guidance documents allowed to avoid duplications and to refer to existing recognized published guidance.\n\nAlso, using different guidance documents helped to identify methods needed to facilitate risk assessments. For example, the WNV and CCHF exercises seem to have been facilitated by the concomitant presence of \"risk questions\" and algorithms in the method that guide in a stepwise manner the participants towards the final assessment. The RVF exercise instead relied only on \"risk questions\" to guide the participants. Replying to those questions might be difficult for people not familiar with RA methodologies and/or without access to relevant information and data. This has probably generated the perception of lack of adequate time allotted for the RVF exercise, and it is also in line with the best practice identified for a joint risk assessment by WHO-OIE-FAO \\[[@B21]\\]: \"at least one member of the Joint Risk Assessment (JRA) Technical Team should have experience in risk assessment to guide the process and advise on the JRA methodology.\"\n\nConsidering that different sectors may rightfully assess the risk differently, this approach has the advantage of enabling actors in each sector to recognize this variability and the reasons behind it. This awareness is a first step towards the identification of national intersectoral priorities in terms of surveillance and response that, in turn, can guide a OneHealth approach to resource allocation. In fact, MRA can facilitate prioritization of zoonosis in line with other proposed integrated approaches \\[[@B19], [@B21], [@B36]\\] and, in addition, allow joint evaluation of the risk of a specific zoonosis and prepare for a coordinated integrated response.\n\nThe pre- and posttests implemented during the exercises have highlighted that many participants did not perceive the relevance and need of recent published and unpublished documents (including those from neighbourhood countries) to support risk assessments. The exercises helped in understanding the relevance of different sources of information and data for RAs. However, it has to be noted that, in order to save time during the implementation, the ISS team searched and analysed in advance the available relevant documentation and synthetized the outcomes of the research in background documents distributed to participants. Data review was therefore not fully simulated. The identification of relevant sources of information by each of the sectors involved in the assessment and their sharing is the first step of the RA, and it should be considered among the relevant outcomes of intersectoral collaboration.\n\nAs highlighted by the WNV exercise, the multisectoral collaboration helped in the identification of the level of risk, and with the CCHF and RVF exercises, we explored at what stage of the RA this collaboration was more beneficial. Our findings suggest that the strategic added value of the multisectoral approach lies in its ability to create a common base of comprehensive and critical information, filing knowledge gaps, and to reduce uncertainty in risk assessment. This, in turn, facilitates the achievement of consensus on the comprehensive level of risk for the country taking into account the perspective of all sectors involved. The concomitant participation to the assessment of other countries of the region has also contributed to the identification of possible cross-border risk factors and to the assessment of a \"regional\" risk level. Similar outcomes were reported following the 2003 International Workshop \\[[@B37]\\] on the possibility, benefits, and obstacles of integration of ecological and health risk assessments based on the WHO \"framework for integrated assessment of human health and ecological risks\" \\[[@B38]\\]. Improved assessment quality, efficiency, and predictive capability were considered to be principal benefits of integration of risk assessments. Unfortunately, some of the obstacles to the acceptance and implementation of this approach, identified at the time, such as disciplinary and organizational barriers between disciplines, are still present. The workshop\\'s recommendations, such as harmonization of exposure characterization, surveillance methods and models, and development of methods to facilitate comparison of risks, are still being addressed \\[[@B21], [@B31], [@B39]\\] underlying both the relevance and the complexity of the issue.\n\n5. Conclusions {#sec5}\n==============\n\nIncreasing awareness of intersectoral priorities, including cross-border ones, through MRA is a new frontier which can support early warning capacities. This approach is relevant to reduce gaps due to unavailability of shared data and information, and it can also promote the use of multiple sources of information across sectors and facilitate consensus on operational arrangements for the RA, e.g., as recommended by the World Health Organisation (WHO) in the Western Pacific Regional Action Plan for Dengue Prevention and Control \\[[@B5]\\]. Given that six out of the ten threats to global health listed by WHO \\[[@B40]\\] are issues occurring at the human, animal, and environmental interface, the implementation of comprehensive regional assessments with a One Health approach made by national authorities using similar frameworks is promising in terms of the potential added value for the global health security agenda. This justifies further efforts in fine-tuning methodological approaches and addressing implementation challenges.\n\nThe authors acknowledge the active participation of the members of MediLabSecure network to the MRA exercises. The MediLabSecure project was supported by the European Commission (DEVCO: IFS/21010/23/\\_194 & IFS/2018/402-247). The funding body was not involved in the design of the study, in collection, analysis, and interpretation of data, and in writing the manuscript.\n\nCCHF:\n\n:   Crimean--Congo haemorrhagic fever\n\nECDC:\n\n:   European Centre for Disease Prevention and Control\n\nEU:\n\n:   European Union\n\nFAO:\n\n:   Food and Agriculture Organization of the United Nations\n\nIZSAM:\n\n:   Istituto Zooprofilattico Sperimentale dell\\'Abruzzo e del Molise\n\nMLS:\n\n:   MediLabSecure project\n\nMRA:\n\n:   Multisectoral risk assessment\n\nRA:\n\n:   Risk assessment\n\nRRA:\n\n:   Rapid risk assessment\n\nRVF:\n\n:   Rift Valley fever\n\nSWOT analysis:\n\n:   Strengths, weaknesses, opportunities, and threats analysis\n\nWHO:\n\n:   World Health Organisation\n\nWNV:\n\n:   West Nile virus.\n\nData Availability\n=================\n\nThe data generated or analysed during this study, including documentation and tools prepared for the exercises, are available in the references reported in this published article (refer to \\[[@B32]--[@B34]\\]) and its supplementary information files. Pre-test and post-test forms filled in by participants are in hard copies available from the corresponding author and can be provided on reasonable request making the copies anonymous.\n\nEthical Approval\n================\n\nThe implementation of the exercises reported in this manuscript did not need formal ethical approval and informed consent and complies with national guidelines as per the code of ethics of Istituto Superiore di Sanit\u00e0 <https://www.iss.it/wp-content/uploads/2017/12/CE_Codice_di_etica_2015_01_13.pdf>\n\nConflicts of Interest\n=====================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' Contributions\n=======================\n\nMGD developed and implemented the exercises and wrote the manuscript; FR supported the development of the exercises, the implementation of WNV exercise, and the drafting of the manuscript; WVB and LM supported the development and the implementation of the WNV exercises and critically revised the manuscript; TM and TD supported the development and the implementation of the CCHF exercise and critically revised the manuscript; BS and PC supported the development and the implementation of the RVF exercise and critically revised the manuscript; SD supported the development and the implementation of all the three exercises and critically revised the manuscript; the persons of the MediLabSecure Working Group collaborated in the exercise development and implementation. All authors reviewed and approved the final manuscript.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\nAdditional file 1: file format: excel; title of data: data analysis of the evaluation of the exercises and the added value of multisectoral approach. Description of data: replies of the participants to the questions related to the evaluation of the exercises and perceived participants\\' added value of the multisectoral approach.\n\n###### \n\nClick here for additional data file.\n\n![Seasonal risk levels of WNV transmission to humans with the corresponding risk area and the indicators used to define the level (source ECDC), Source: \\[[@B28]\\].](BMRI2020-4832360.001){#fig1}\n\n![Perceived risk of West Nile virus using the ECDC risk assessment tool. Risk areas identified by three countries with consensus between sectors.](BMRI2020-4832360.002){#fig2}\n\n![Perceived risk of West Nile virus using the ECDC risk assessment tool. Risk areas identified by two countries with less consensus between sectors.](BMRI2020-4832360.003){#fig3}\n\n![Added value of the multisectoral approach as assessed by participants to the CCHF exercise (11 countries).](BMRI2020-4832360.004){#fig4}\n\n![Added value of the multisectoral approach as assessed by participants to the RVF exercise (8 countries).](BMRI2020-4832360.005){#fig5}\n\n###### \n\nOverview of the three multisectoral risk assessment exercises conducted, Source: \\[[@B31]\\].\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Exercise (place and date)                                              Participant countries from MediLabSecure network                                                                                                                                                                                  Objectives                                                                                                                       Methodology                                                                                                                                                                                   Guidance documents\n  ---------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------\n  West Nile virus exercise (Paris, December 2015)                        Albania, Algeria, Armenia, Bosnia and Herzegovina, Egypt, Georgia, Jordan, Kosovo, Lebanon, Libya, Moldova, Montenegro, Morocco, Palestine, former Yugoslav Republic of Macedonia (FYROM), Serbia, Tunisia, Turkey, and Ukraine   \\(i\\) Describe risk level assessment between sectors and countries\\                                                              \\(1\\) Map the assessment of WNV risk across four sectors (human and animal virology, medical entomology, and public health) by country and by regions\\                                        ECDC \"West Nile virus risk assessment tool\" \\[[@B28]\\]\n                                                                                                                                                                                                                                                                                                           (ii) Assess the cross-sectoral collaboration during the initial phase of the MediLabSecure project\\                              (2) Conduct a SWOT analysis to assess strengths, weaknesses, opportunities, and threats in relation to the surveillance systems in place at national level, to support the risk assessment\\   \n                                                                                                                                                                                                                                                                                                           (iii) Make participants aware of the ECDC tool\\                                                                                  (3) Compile an evaluation questionnaire on exercise satisfaction                                                                                                                              \n                                                                                                                                                                                                                                                                                                           (iv) Provide indications for the next MRA exercises                                                                                                                                                                                                                                                                            \n\n                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n\n  Crimean--Congo haemorrhagic fever exercise (Belgrade, November 2016)   Albania, Armenia, Bosnia and Herzegovina, Former Yugoslav Republic of Macedonia (FYROM), Georgia, Kosovo, Moldova Montenegro, Serbia, Turkey, and Ukraine                                                                         \\(i\\) Enhance knowledge and capacity on MRA\\                                                                                     \\(1\\) Table0top exercise on multisector risk assessment with four sectors (human and animal virology, medical entomology, and public health) by country and by regions\\                       ECDC \"operational guidance on rapid risk assessment methodology\" \\[[@B29]\\]\\\n                                                                                                                                                                                                                                                                                                           (ii) Encourage multisectoral collaboration and exchange, also among neighbouring countries and assess the related added value\\   (2) Questionnaire on the value of multisector approach\\                                                                                                                                       FAO \"RVF in Niger risk assessment\" \\[[@B30]\\]\n                                                                                                                                                                                                                                                                                                           (iii)Provide consensus on a single national level of risk across all the sectors\\                                                (3) Evaluation questionnaire on exercise satisfaction                                                                                                                                         \n                                                                                                                                                                                                                                                                                                           (iv) Make participants aware of ECDC RRA guidance and FAO RA methodology\\                                                                                                                                                                                                                                                      \n                                                                                                                                                                                                                                                                                                           (v) Make participants aware of ECDC RRA guidance and FAO RA methodology                                                                                                                                                                                                                                                        \n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n###### \n\nNumber of participants of the CCHF exercise who identified useful documents for RA, by type of document.\n\n  Type of document                                                                    Pretest   Posttest        \n  ----------------------------------------------------------------------------------- --------- ---------- ---- -----\n  No documents mentioned                                                              9         25         2    6\n  Guidance, law decrees, plans                                                        22        63         17   49\n  Guidance, law decrees, plans, scientific articles, unpublished documents, studies   2         6          14   39\n  Scientific articles                                                                 2         6          2    6\n  Total responders                                                                    35        100        35   100\n\n[^1]: Academic Editor: Surender Khurana\n"}
+{"text": "Introduction {#section5-2040622320922020}\n============\n\nThe rupture and erosion of unstable intravascular coronary plaques and the formation of thrombosis are direct causes of acute coronary syndrome (ACS).^[@bibr1-2040622320922020],[@bibr2-2040622320922020]^ Usually, plaque rupture and secondary thrombosis are considered to be complex pathophysiological events, and typical coronary heart disease risk factors play an important role in the development of ACS.^[@bibr3-2040622320922020][@bibr4-2040622320922020][@bibr5-2040622320922020]--[@bibr6-2040622320922020]^ We know that the occurrence of ACS can be caused by thrombosis or luminal obstruction after the rupture of a coronary plaque in patients with unstable angina pectoris; therefore, early identification and prediction of coronary plaque in unstable angina pectoris (UAP) patients is conducive to the clinical treatment and clinical prognosis of patients.\n\nThe immune response theory suggests that T lymphocytes including CD8+ and CD4+ T cells can be observed at various stages of atherosclerosis in both human and non-human primates. This suggests that the immune response plays a part in the progression of the disease.^[@bibr7-2040622320922020],[@bibr8-2040622320922020]^ Previous studies have confirmed that fatty macrophages, T lymphocytes (including CD8+ and CD4+), and smooth muscle cells are present mainly in the lipid stripes of atherosclerosis.^[@bibr9-2040622320922020],[@bibr10-2040622320922020]^ Therefore, the pathological mechanism suggests that CD4+ and CD8+ T cells are involved in the process of atherosclerotic plaque stability. A recent study showed that the CD4+/CD8+ ratio was correlated positively with the occurrence of ACS.^[@bibr11-2040622320922020]^ However, the association between the CD4+/CD8+ ratio and the risk of acute coronary events is unclear.\n\nAt present, coronary angiography is commonly used to evaluate coronary vascular status, and more accurate data can be obtained. Optical coherence tomography (OCT) can not only detect the vascular morphology of the coronary artery, but also analyze the details of lesions; thus OCT technology can evaluate the components and characteristics of coronary plaques. The use of imaging techniques can therefore better identify unstable plaques.^[@bibr2-2040622320922020],[@bibr7-2040622320922020]^\n\nWe investigated the potential association between the CD4+/CD8+ ratio and OCT-confirmed coronary artery instability in patients with UAP. Our results may help predict the risk of plaque rupture and provide an updated clinical basis for the prevention and treatment of coronary atherosclerotic unstable plaques in UAP patients.\n\nMaterials and methods {#section6-2040622320922020}\n=====================\n\nStudy population and design {#section7-2040622320922020}\n---------------------------\n\nContinuous screening included a total of 266 patients. All participants were selected from UAP patients who were admitted to our center from January 2016 to January 2018 for percutaneous coronary interventional therapy (PCI) and OCT examination ([Figure 1](#fig1-2040622320922020){ref-type=\"fig\"}). All patients underwent and completed PCI successfully without complications. The diagnosis of UAP was based on pre-established guidelines.^[@bibr12-2040622320922020]^ Meanwhile, demographic characteristics, clinical features, risk factors for coronary heart disease, blood biochemical data, electrocardiography, echocardiography, coronary angiography (CAG), and OCT results were collected. CAG diagnosed the culprit vessels, while OCT confirmed the vulnerability of plaques at these sites.\n\n![Flowchart depicting patient selection. Among the 266 patients enrolled, OCT examination confirmed plaque rupture (*n*\u2009=\u20099), TCFA (*n*\u2009=\u200966), non-rupture and non-TCFA (*n*\u2009=\u2009191). Plaque rupture and TCFA were classified as the instability group, non-rupture and non-TCFA comprised the stability group.\\\nOCT, optical coherence tomography; TCFA, thin-cap fibroatheroma.](10.1177_2040622320922020-fig1){#fig1-2040622320922020}\n\nFollow up {#section8-2040622320922020}\n---------\n\nAll patients were subsequently followed for 1\u2009year. Primary endpoint was occurrence of a major adverse cardiovascular event (MACE), including all-cause death, target vessel myocardial infarction (MI), and clinically driven target vessel revascularization.\n\nDefinitions of coronary heart disease risk factors {#section9-2040622320922020}\n--------------------------------------------------\n\nThe diagnostic criteria of hyperlipidemia were based on the guidelines for prevention and treatment of dyslipidemia in Chinese adults (2016).^[@bibr13-2040622320922020]^ Hypertension was diagnosed if the patient had blood pressure of \u2a7e140/90\u2009mmHg during at least three random examinations, or was actively using antihypertensive drugs.^[@bibr14-2040622320922020]^ Diabetes was diagnosed if the patient's fasting blood glucose was \u2a7e7.1\u2009mmol/l or \u2a7e11.1\u2009mmol/l 2\u2009h after a meal, or there was a clear history of diabetes and the patient was receiving hypoglycemic drugs.^[@bibr15-2040622320922020]^ Body mass index (BMI) was calculated by dividing a patient's weight in kilograms by the square of their height in meters. Smoking was defined as current or previous smoking with cessation less than 3\u2009months prior.\n\nLaboratory data {#section10-2040622320922020}\n---------------\n\nBlood samples were taken from all patients before procedures and immediately sent to the laboratory for analysis. Routine examination included standard blood work-up, liver function, renal function, blood lipid levels, fasting blood glucose, etc. T lymphocyte counts were expressed as absolute and proportional CD3+ cell numbers and CD4+ and CD8+ cells were counted in all 266 test subjects, from which the CD4+/CD8+ ratio was calculated. Blood tests were conducted using standard methods at the central laboratory.\n\nPCI procedure and OCT image acquisition {#section11-2040622320922020}\n---------------------------------------\n\nAll patients underwent successful CAG and PCI after admission. After intracoronary nitroglycerine administration, the OCT catheter was advanced distal to the culprit lesion. Imaging of the culprit lesion was then acquired using the frequency-domain OCT C7XR system and the Dragonfly catheter (Lightlab Imaging, St. Jude Medical, Westford, MA).^[@bibr2-2040622320922020],[@bibr16-2040622320922020]^ All treatment strategies were determined by experienced interventional cardiologists. Angiographic images measured reference diameter, minimum lumen diameter, lesion length, and percent diameter stenosis. At the end of the procedure, patients undergoing stent implantation were subjected to final OCT imaging. The images were stored digitally and identified for offline analysis.\n\nOCT image analysis {#section12-2040622320922020}\n------------------\n\nAfter determining the culprit lesions, two experienced cardiac interventional technicians analyzed the morphological characteristics of plaques according to the standards.^[@bibr2-2040622320922020],[@bibr16-2040622320922020][@bibr17-2040622320922020]--[@bibr18-2040622320922020]^ Plaques were divided into fibrous (uniform, high-backscatter area) and lipid (diffuse boundary of low signal region) plaques.^[@bibr16-2040622320922020],[@bibr19-2040622320922020]^ In lipid plaques, the thickness of the fiber cap at the thinnest point was measured three times and the mean value was calculated. Thin-cap fibroatheroma (TCFA) was defined as a plaque with lipid content in at least two quadrants, a plaque with a maximal lipid arc \\>90\u00b0 with the thinnest part of the fibrous cap measuring \\<65\u2009\u03bcm.^[@bibr20-2040622320922020]^ Plaque rupture (PR) was identified by the presence of fibrous cap discontinuity with a clear cavity formed inside the plaque ([Figure 2](#fig2-2040622320922020){ref-type=\"fig\"}).^[@bibr2-2040622320922020],[@bibr21-2040622320922020],[@bibr22-2040622320922020]^\n\n![Representative OCT images of normal, TCFA, and plaque rupture (left to right, respectively).\\\nOCT, optical coherence tomography; TCFA, thin-cap fibroatheroma.](10.1177_2040622320922020-fig2){#fig2-2040622320922020}\n\nStatistical analysis {#section13-2040622320922020}\n--------------------\n\nAll analyses were performed using SPSS 23.0 for Windows statistical software (SPSS Inc., Chicago, IL, USA). The CD4+/CD8+ ratio was not normally distributed, so non-parametrical testing was used to determine differences. The Spearman correlation coefficient was calculated to assess associations between CD4+/CD8+ ratio and all continuous variables in this study. The Mann--Whitney *U* test was used to study CD4+/CD8+ ratio as a continuous variable for all risk factors. Normally distributed continuous variables were expressed as means and standard deviations (SDs), whereas categorical variables were presented as percentages. The Chi-square (\u03c7^2^) test was used for comparing categorical variables. Significant variables in univariate analysis were subsequently included in the multivariate logistic analysis. The cumulative event rate was estimated from Kaplan--Meier curves and compared using the log-rank test. A *p* value of \\<0.05 was considered statistically significant.\n\nResults {#section14-2040622320922020}\n=======\n\nCD4+/CD8+ ratio and clinically relevant characteristics {#section15-2040622320922020}\n-------------------------------------------------------\n\nBaseline clinical characteristics for all patients are described in [Table 1](#table1-2040622320922020){ref-type=\"table\"}. A higher CD4+/CD8+ ratio was observed in patients who smoked (*p\u2009=\u2009*0.021), whereas laboratory data showed a positive correlation between CD4+/CD8+ ratio and low density lipoprotein (LDL) levels (*r*\u2009=\u20090.22, *p*\u2009\\<\u20090.001), and total cholesterol (*r*\u2009=\u20090.191, *p*\u2009=\u20090.002). CD4+/CD8+ ratio was shown to have a negative correlation with high density lipid levels (*r*\u2009=\u2009--0.167, *p*\u2009=\u20090.006).\n\n###### \n\nBaseline characteristics of the patients in relation to ratio of CD4+/CD8+ plasma levels.\n\n![](10.1177_2040622320922020-table1)\n\n                                             (*n*\u2009=\u2009266)          CD4/8               *p* value\n  ------------------------------------------ -------------------- ------------------- -----------------------------------------------------------------\n  Age, year                                  65 (24--88)          *r*\u2004=\u2004\u22120.42         0.490\n  \u2003Male                                      194 (72.9%)          1.74 (0.77--2.79)   0.360\n  \u2003Female                                    72 (27.1%)           1.75 (0.66--2.80)   \n  BMI, kg/m^2^                               24.8 (17.6--37.8)    *r*\u2004=\u2004\u22120.05         0.450\n  Hypertension, *n* (%)                                                               \n  \u2003Yes                                       183 (68.8)           1.74 (0.66--2.80)   0.310\n  \u2003No                                        83 (31.2)            1.77 (0.74--2.39)   \n  Hyperlipidemia, *n* (%)                                                             \n  \u2003Yes                                       185 (69.5)           1.77 (0.66--2.80)   0.319\n  \u2003No                                        81 (30.5)            1.71 (0.74--2.79)   \n  Diabetes, *n* (%)                                                                   \n  \u2003Yes                                       81 (30.5)            1.75 (0.84--2.63)   0.980\n  \u2003No                                        185 (69.5)           1.74 (0.66--2.80)   \n  Prior or current smoker, *n* (%)                                                    \n  \u2003Yes                                       126 (47.4)           1.80 (0.68--2.80)   0.021\n  \u2003No                                        140 (52.6)           1.73 (0.66--2.77)   \n  Prior stroke, *n* (%)                                                               \n  \u2003Yes                                       33 (12.4)            1.75 (0.68--2.36)   0.890\n  \u2003No                                        233 (87.6)           1.74 (0.66--2.80)   \n  Prior MI, *n* (%)                                                                   \n  \u2003Yes                                       27 (10.2)            1.71 (0.84--2.80)   0.606\n  \u2003No                                        239 (89.8)           1.75 (0.66--2.79)   \n  Prior peripheral vessel disease, *n* (%)                                            \n  \u2003Yes                                       14 (5.3)             1.77 (1.29--2.36)   0.670\n  \u2003No                                        252 (94.7)           1.75 (0.66--2.80)   \n  Prior kidney failure, *n* (%)                                                       \n  \u2003Yes                                       5 (1.9)              1.59 (1.18--1.99)   0.681\n  \u2003No                                        261 (98.1)           1.75 (0.66--2.80)   \n  Prior COPD, *n* (%)                                                                 \n  \u2003Yes                                       12 (4.5)             1.60 (1.22--1.99)   0.293\n  \u2003No                                        254 (95.5)           1.75 (0.66--2.80)   \n  Prior HF, *n* (%)                                                                   \n  \u2003Yes                                       31 (11.7)            1.73 (1.00--2.37)   0.518\n  \u2003No                                        235 (88.3)           1.75 (0.66--2.80)   \n  **Medication:**                                                                     \n  Prior use of statin, *n* (%)                                                        \n  \u2003Yes                                       113 (42.5)           1.74 (0.66--2.80)   0.780\n  \u2003No                                        153 (57.5)           1.75 (0.69--2.79)   \n  Prior use of \u03b2-blockers, *n* (%)                                                    \n  \u2003Yes                                       65 (24.4)            1.77 (0.66--2.80)   0.954\n  \u2003No                                        201 (75.6)           1.74 (0.68--2.79)   \n  Prior use of CCB, *n* (%)                                                           \n  \u2003Yes                                       106 (39.8)           1.74 (0.68--2.67)   0.572\n  \u2003No                                        160 (60.2)           1.75 (0.66--2.80)   \n  Prior use of ACEI/ARB, *n* (%)                                                      \n  \u2003Yes                                       90 (33.8)            1.75 (0.68--2.80)   0.718\n  \u2003No                                        176 (66.2)           1.74 (0.66--2.63)   \n  **Laboratory**                                                                      \n  WBC                                        6.54 (3.48--15.47)   0.029               0.637\n  Hemoglobin                                 136 (72--179)        0.022               0.725\n  PLT                                        197 (62--539)        \u22120.03               0.623\n  HDL                                        1.10 (0.6--2.21)     \u22120.167              0.006[\\*\\*](#table-fn1-2040622320922020){ref-type=\"table-fn\"}\n  LDL-c                                      1.87 (0.70--4.72)    0.22                \\<0.001[\\*\\*](#table-fn1-2040622320922020){ref-type=\"table-fn\"}\n  TC                                         3.49 (1.50--6.19)    0.191               0.002[\\*\\*](#table-fn1-2040622320922020){ref-type=\"table-fn\"}\n  TG                                         1.58 (0.43--6.34)    0.092               0.133\n  Apo A1                                     1.22 (0.76--1.86)    \u22120.074              0.348\n  Apo B                                      0.73 (0.29--1.43)    0.128               0.105\n  Lipoprotein a                              169 (7--1694)        \u22120.015              0.848\n  Creatinine                                 72 (38--723)         \u22120.005              0.940\n  Uric acid                                  319 (175--731)       \u22120.054              0.494\n  AST                                        21 (9--475)          0.006               0.923\n  ALT                                        21.5 (5--404)        \u22120.009              0.881\n  GLU                                        5.21 (3.14--19.83)   \u22120.055              0.369\n\nCorrelation is significant at the 0.01 level (2-tailed).\n\nData are presented as *n* (%) and median\u2004\u00b1\u2004\\[IQR\\] or *r*\u2004=\u2004Spearman's rank correlation coefficient.\n\nACEI, angiotensin-converting enzyme inhibitors; ALT, alanine aminotransferase; apo-AI, apolipoprotein A1; apo-B, apolipoprotein B; ARB, angiotensin receptor blocker; AST, aspartate transaminase; BMI, body mass index; CCB, calcium channel blockers; COPD, chronic obstructive pulmonary disease; HDL, high density lipoprotein; HF, heart failure; GLU, glucose; IQR, interquartile range; LDL-c, low density lipoprotein-cholesterol; MI, myocardial infarction; PLT, platelet; TC, total cholesterol; TG, total triglycerides; WBC, white blood cell.\n\nCAG and OCT findings in relation to CD4+/CD8+ ratio {#section16-2040622320922020}\n---------------------------------------------------\n\nCAG and OCT findings in relation to CD4+/CD8+ ratio are presented in [Table 2](#table2-2040622320922020){ref-type=\"table\"}. A higher CD4+/CD8+ ratio was observed in patients with plaque rupture (*p*\u2009=\u20090.019) and TCFA (*p*\u2009=\u20090.027), whereas a lower CD4+/CD8+ ratio was found in patients with calcified plaques (*p*\u2009=\u20090.045). There were no significant differences in other CAG or OCT findings between those with high or low CD4+/CD8+ ratio levels.\n\n###### \n\nCD4+/CD8+ ratio according to OCT and angiography findings.\n\n![](10.1177_2040622320922020-table2)\n\n                               (*n*\u2009=\u2009266)   CD4+/CD8+           *p* value\n  ---------------------------- ------------- ------------------- -----------\n  Target vessel                                                  \n  LM                           5 (1.9)       1.93 (1.64--2.08)   0.141\n  LAD                          207 (77.8)    1.73 (0.66--2.80)   0.611\n  LCX                          25 (9.4)      1.84 (0.89--2.36)   0.517\n  RCA                          39 (14.7)     1.78 (0.68--2.63)   0.764\n  Characteristic of plaque                                       \n  Lipid                        92 (34.6)     1.81 (0.68--2.80)   0.066\n  Calcified                    68 (25.6)     1.68 (0.66--2.40)   0.045\n  Fibrotic                     106 (39.8)    1.75 (0.77--2.58)   0.998\n  Rupture                      9 (3.4)       1.92 (1.74--2.63)   0.019\n  TCFA, *n*(%)                 66 (24.8)     1.84 (0.91--2.8)    0.027\n  Number of vascular lesions                                     \n  1                            123 (46.2)    1.74 (0.68--2.80)   0.610\n  2                            79 (29.7)     1.77 (0.77--2.35)   0.976\n  3                            64 (24.1)     1.74 (0.66--2.63)   0.530\n  Bifurcation lesion           72 (27.1)     1.82 (0.89--2.58)   0.151\n  CTO                          12 (4.5)      1.57 (0.91--2.80)   0.208\n\nData are presented as *n* (%) and median\u2004\u00b1\u2004\\[IQR\\] or *p* value for Mann--Whitney *U* test.\n\nCTO, chronic total occlusion; IQR, interquartile range; LAD, left anterior descending artery; LCX, left circumflex; LM, left main; OCT, optical coherence tomography; RCA, right coronary artery; TCFA, thin-cap fibroatheroma.\n\nCoronary risk factors and laboratory data in relation to plaque instability {#section17-2040622320922020}\n---------------------------------------------------------------------------\n\nCoronary risk factors and laboratory data based on plaque instability are shown in [Table 3](#table3-2040622320922020){ref-type=\"table\"}, including the instability (plaque rupture and TCFA patients, *n*\u2009=\u200975), and stable (non-rupture and non-TCFA patients, *n*\u2009=\u2009191) groups. CD4+/CD8+ ratio \\>1.725, prior stoke, hemoglobin, and triglyceride (all *p*\u2009\\<\u20090.05) were statistically different between the two groups.\n\n###### \n\nCharacteristics of coronary risk factors and laboratory data according to OCT indicated plaque vulnerability.\n\n![](10.1177_2040622320922020-table3)\n\n                           Ruptured plaque and TCFA   Nonrupture and non-TCFA   t/\u03c7^2^   *p* value\n  ------------------------ -------------------------- ------------------------- -------- -----------\n  Male                     52 (69.3)                  142 (74.3)                0.147    0.701\n  Age (year)               65.38\u2009\u00b1\u200911.09              63.83\u2009\u00b1\u200910.37             1.067    0.287\n  Hypertension             54 (72.0)                  129 (67.5)                1.255    0.263\n  Diabetes mellitus        22 (29.3)                  59 (30.9)                 0.005    0.945\n  Prior or current smoke   35 (46.7)                  91 (47.6)                 0.013    0.908\n  Hyperlipidemia           55 (73.3)                  130 (68.1)                1.595    0.207\n  CHF                      10 (13.3)                  21 (11.0)                 0.408    0.523\n  Family history of CAD    9 (12.0)                   27 (14.1)                 0.125    0.724\n  Prior MI                 6 (8.0)                    21 (11.0)                 0.411    0.521\n  Prior stroke             35 (46.7)                  18 (9.4)                  5.543    0.019\n  Prior COPD               1 (1.3)                    11 (5.8)                  1.409    0.235\n  BMI                      24.92\u2009\u00b1\u20092.77               25.05\u2009\u00b1\u20093.08              0.312    0.775\n  WBC (mmol/l)             6.66\u2009\u00b1\u20091.75                6.93\u2009\u00b1\u20092.05               0.982    0.326\n  Hemoglobin (mmol/l)      130.34\u2009\u00b1\u200919.30             136.66\u2009\u00b1\u200915.46            6.273    0.014\n  PLT (mmol/l)             199.83\u2009\u00b1\u200956.00             203.22\u2009\u00b1\u200960.75            0.41     0.682\n  HDL-c (mmol/l)           1.16\u2009\u00b1\u20090.33                1.15\u2009\u00b1\u20090.27               0.315    0.753\n  LDL-c (mmol/l)           1.94\u2009\u00b1\u20090.67                1.87\u2009\u00b1\u20090.62               0.747    0.455\n  TC (mmol/l)              3.22\u2009\u00b1\u20090.96                3.63\u2009\u00b1\u20090.96               0.835    0.393\n  TG (mmol/l)              2.10\u2009\u00b1\u20091.09                1.72\u2009\u00b1\u20090.88               7.305    0.008\n  ApoA1 (g/L)              1.21\u2009\u00b1\u20090.16                1.25\u2009\u00b1\u20090.21               1.397    0.165\n  ApoB (g/L)               0.74\u2009\u00b1\u20090.22                0.75\u2009\u00b1\u20090.23               2.69     0.789\n  Lipoprotein a (g/L)      139 (100,292)              202 (98,363)              1.319    0.189\n  Creatinine (\u03bcmol/L)      73 (39,723)                72 (39,162)               1.225    0.225\n  Uric acid (\u03bcmol/L)       351.64\u2009\u00b1\u200998.49             326.41\u2009\u00b1\u200996.70            1.501    0.135\n  GLU (mmol/l)             5.94\u2009\u00b1\u20091.92                5.92\u2009\u00b1\u20092.14               0.046    0.964\n  CD4+/CD8+ \\>1.725        53 (70.7)                  91 (47.6)                 11.497   0.001\n\nData are presented as *n* (%) and mean\u2004\u00b1\u2004(SD).\n\napo-AI, apolipoprotein A1; apo-B, apolipoprotein B; BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; GLU, glucose; HDL-c, high density lipoprotein-cholesterol; LDL-c, low density lipoprotein-cholesterol; OCT, optical coherence tomography; PLT, platelet; SD, standard deviation; TC, total cholesterol; TCFA, thin-cap fibroatheroma; TG, total triglycerides; WBC, white blood cell.\n\nPredictors of plaque vulnerability {#section18-2040622320922020}\n----------------------------------\n\nReceiver operating curve (ROC) analysis showed that CD4+/CD8+ ratio was predictive of plaque instability ([Figure 3](#fig3-2040622320922020){ref-type=\"fig\"}), and the threshold for CD4+/CD8+ ratio was 1.725 with a sensitivity and specificity of 70.7% and 52.4%, respectively (Youden's index\u2004=\u20040.23). Based on the threshold for the CD4+/CD8+ ratio, the cohort was divided into two groups: a higher CD4+/CD8+ ratio group (CD4+/CD8+ ratio \\>1.725, *n*\u2009=\u2009144) and a lower CD4+/CD8+ ratio group (CD4+/CD8+ ratio \u2a7d1.725, *n*\u2009=\u2009122). Significant variables in the univariate analysis were subsequently included in a binary logistic regression analysis, which showed that CD4+/CD8+ ratio \\>1.725 and prior stroke were predictors of plaque instability (all *p*\u2009\\<\u20090.05, [Table 4](#table4-2040622320922020){ref-type=\"table\"}). In this study, 53 (70.7%) of the 75 patients with unstable plaques had a CD4+/CD8+ ratio \\>1.725 (\u03c7^2^\u2009=\u200911.497, *p*\u2009\\<\u20090.01). Patients with a CD4+/CD8+ ratio \\>1.725 had an increased risk of plaque instability compared with patients with a CD4+/CD8+ ratio \\<1.725 \\[odds ratio (OR)\u2004=\u20042.679, 95% confidence interval (CI): 1.497--4.794\\].\n\n![ROC analyses for the predictive efficacy of CD4+/CD8+ ratio for plaque instability.\\\nROC, receiver operating curve.](10.1177_2040622320922020-fig3){#fig3-2040622320922020}\n\n###### \n\nAssociation between patient characteristics and the prevalence of plaque vulnerability: results of binary logistic regression analysis.\n\n![](10.1177_2040622320922020-table4)\n\n  independent variables           OR      95% CI           *p* value\n  ------------------------- ----- ------- ---------------- -----------\n  CD4+/CD8+ ratio \\>1.725   No    1                        \n                            Yes   2.651   (1.474--4.768)   0.001\n  Prior stroke              No    1                        \n                            Yes   2.960   (1.380--6.348)   0.005\n\nOutcome {#section19-2040622320922020}\n-------\n\nOf 266 patients, 23 suffered from MACE and 7 died during a 1-year follow-up. The Kaplan--Meier estimate for all-cause mortality and MACE showed no significant differences in the lower CD4+/CD8+ ratio group compared with the higher CD4+/CD8+ ratio group (all-cause mortality, 2.5% *versus* 2.8%, *p*\u2009=\u20090.858; MACE, 7.4% *versus* 9.7%, *p*\u2009=\u20090.488; [Figures 4](#fig4-2040622320922020){ref-type=\"fig\"} and [5](#fig5-2040622320922020){ref-type=\"fig\"}). However, after correcting for smoking in a Cox regression analysis, prior stroke, ruptured plaque, CD4+/CD8+ ratio, and hemoglobin were not predictive (all *p*\u2009\\>\u20090.05), but levels of LDL were statistically significant for MACE \\[hazard ratio (HR)\u2009=\u20092.26; 95% CI, 1.039--4.915, *p*\u2009=\u20090.04\\] and all-cause mortality (HR\u2009=\u20092.66; 95% CI, 1.068--6.623, *p*\u2009=\u20090.036).\n\n![The Kaplan--Meier analysis from MACE.\\\nMACE, major adverse cardiovascular event.](10.1177_2040622320922020-fig4){#fig4-2040622320922020}\n\n![The Kaplan--Meier analysis from all-cause mortality.](10.1177_2040622320922020-fig5){#fig5-2040622320922020}\n\nDiscussion {#section20-2040622320922020}\n==========\n\nIn this retrospective study, we found that CD4+/CD8+ ratio correlated with coronary plaque instability in patients with UAP and could predict plaque rupture and TCFA. These findings suggest that CD4+/CD8+ ratio may represent a cost-effective marker for risk stratification in patients with UAP.\n\nThe current immune response theory of atherosclerotic lesions suggests that T cells exist in atherosclerotic plaques in experimental animals and humans,^[@bibr23-2040622320922020]^ and there is evidence of different T cell subsets in coronary plaques.^[@bibr24-2040622320922020][@bibr25-2040622320922020][@bibr26-2040622320922020][@bibr27-2040622320922020]--[@bibr28-2040622320922020]^ Multiple studies have shown that CD4+T cells are involved in the induction and regulation of atherosclerosis,^[@bibr29-2040622320922020][@bibr30-2040622320922020][@bibr31-2040622320922020][@bibr32-2040622320922020]--[@bibr33-2040622320922020]^ and these studies also confirm that CD4+/CD8+ ratio was significantly correlated with LDL levels. Furthermore, LDL is also an important factor in the progression of coronary atherosclerosis, and the positive correlation between CD4+/CD8+ ratio and LDL was also confirmed in our study. It is undeniable that LDL is related closely to the progression of coronary plaque, but the vulnerability of coronary plaque is not completely dependent on LDL, such as unstable calcification nodules, and there is a high risk of plaque rupture. Many patients with coronary plaque instability are detected by OCT, even if they have no history of smoking, hypertension, diabetes, dyslipidemia or any other traditional coronary heart disease risk, so we sought a new indicator to predict coronary plaque instability.\n\nRecent reports suggest that CD8+ T cells are also important regulatory factors involved in atherosclerosis.^[@bibr34-2040622320922020][@bibr35-2040622320922020][@bibr36-2040622320922020]--[@bibr37-2040622320922020]^ Immunohistochemical staining of CD3, CD4, and CD8 showed progressive T cell accumulation during atherosclerosis, the early stages of which were mainly diffuse cytotoxic T cell infiltration, but increasing numbers of T helper cells were found during disease progression, leading to an elevated CD4+/CD8+ ratio.^[@bibr38-2040622320922020],[@bibr39-2040622320922020]^ Previous studies have found that the number of T lymphocytes in the blood and lymph nodes of patients with atherosclerosis and the ratio of CD4+/CD8+ were different. In this study, the ratio of CD4+/CD8+ T cells and coronary plaque instability was detected in the peripheral blood of patients in a predictive manner.^[@bibr34-2040622320922020],[@bibr38-2040622320922020]^\n\nCoronary atherosclerosis is a chronic process of development; we know that UAP is caused mainly by the development of atherosclerotic plaques in coronary arteries, such as ulcers, erosion, rupture, etc., which leads to an inflammatory response in the vascular intima or lumen blockage. Patients with UAP are more likely to have acute MI, and the probability of cardiovascular events is increased. We therefore screened UAP patients with coronary plaque progression for plaque instability or even plaque rupture. In this study, we found that the ratio of CD4+/CD8+ increased with instability of plaques, further validating the above studies.^[@bibr34-2040622320922020],[@bibr38-2040622320922020],[@bibr39-2040622320922020]^\n\nRecently, in one study, Gao and his team confirmed that the Chinese elderly patients with ACS in the peripheral blood CD4+/CD8+ ratio is increased significantly compare to stable angina pectoris patients. ACS patients showed a significantly increased CD4+/CD8+ ratio, which was involved in the progress of coronary artery plaque,^[@bibr11-2040622320922020]^ but only based on the clinical diagnosis of ACS in the study. We know that in the real world, for patients with angina pectoris the situation can be quite different, with many interfering factors making the diagnosis of stable or unstable angina pectoris difficult. We therefore used the most direct and ideal intravascular imaging technique, OCT, to diagnose plaque instability. In our study, the correlation between plaque characteristics confirmed by OCT and the ratio of CD4+/CD8+ cells was then compared. CD4+/CD8+ ratio was found to be an predictor of coronary plaque instability in patients with UAP.\n\nWe also followed the all patients for 1\u2009year, with telephone and outpatient follow up being conducted at the 1st, 3rd, 6th and 12th months after the operation. Postoperative anti-platelet aggregation drug therapy was standardized, and CAG review was conducted at the 13th month. It was found that CD4+/CD8+ ratio had no significant statistical significance on patient MACE or death. However, the survival time of the group with a high CD4+/CD8+ ratio was lower than that of the group with a low CD4+/CD8+ ratio, although patients' prognosis was affected by various other factors. We will await the 3-year follow-up information to further elucidate any role of CD4+/CD8+ ratio in patient survival.\n\nIn our study, most MACE incidences in our higher CD4+/CD8+ ratio group occurred in the first 6\u2009months, and the probability of MACE in months 6--12 was almost identical in both groups. We can assume that PCI can be optimized under precise guidance by OCT, while the early detection of plaque instability may lead to personalized post-operative treatment, such as intensive statin therapy and antithrombotic therapy. After individualized treatment, the prognosis of patients with unstable plaques is not significantly different from that of patients with stable plaques. Therefore, early prediction of coronary plaque instability in patients with UAP is of utmost importance for PCI and post-operative treatment regimens. Our findings further highlight the importance of correctly predicting diagnosis and optimizing treatment. We emphasize the need for early detection and individualized treatment to achieve survival benefits. We know that, in China, some basic hospitals lack OCT devices, even in large medical or cardiology centers because OCT examination is expensive. This examination method has further drawbacks, such as the need for enhanced contrast, specially trained professional technical personnel, limitations of blood vessel lumen diameter, and low penetration. Despite these, OCT examination remains vastly superior to other intravascular examinations. Therefore, when OCT cannot be used because of the limitations detailed above, the CD4+/CD8+ ratio in the peripheral blood can be detected to predict coronary plaque instability, so as to optimize personalized treatment and reduce the MACE rate.\n\nSeveral limitations should be recognized in this study. First, this study was a single-center retrospective observational study. Secondly, we observed the coronary artery plaques of patients with targeted lesions and did not analyze predictors of the composition of non-target lesions. Furthermore, this study only included UAP patients with successful PCI, so the results could not be extended to all ACS patients. Finally, prognostic survival analysis was based on multi-factor interventions, such as smoking, lifestyle, nutrition, and exercise habits, that might affect mortality, but this information was not available in the data, and the small sample size limited the predicted results.\n\nDespite these limitations, our research also has a number of strengths. Most reports on the prediction of atherosclerotic plaque characteristics by lymphocytes comprise studies of the aorta, carotid and renal arteries, and animal experiments, but there are few studies regarding the direct prediction of coronary plaque. Our study directly predicted the characteristics of coronary plaque in UAP patients by calculating the CD4+/CD8+ T lymphocyte ratio. Extravascular ultrasound, angiography, intravascular ultrasound, pathological biopsy, and other techniques are most commonly used to define the characteristics of plaques; however, studies on the relationship between the determination of atherosclerotic plaque and CD4+/CD8+T lymphocyte ratio by OCT have not been previously reported. It is of great scientific and clinical significance to predict the instability of coronary atherosclerotic plaques for efficient prevention and treatment. This study is thus both innovative and practical.\n\nConclusion {#section21-2040622320922020}\n==========\n\nHigher CD4+/CD8+ ratio is associated with higher risk of plaque instability, as confirmed by OCT in our cohort of UAP patients. CD4+/CD8+ ratio was not an independent predictor of 1-year MACE and all-cause mortality.\n\nWe deeply appreciate Ling Lin for remote monitoring and data collection throughout the study. We thank Xiao-Fei Gao for analysing data. We also thank Gillian Campbell, from Liwen Bianji, Edanz Group China, for editing the English text of a draft of this manuscript.\n\n**Author contributions:** ZJJ and LS designed the statistical analysis, led the interpretation of research findings, and revised the manuscript. CDY and KXQ participated in design, data collection, dataset generation, statistical analysis, and drafting of the manuscript. All of the authors read and approved the final manuscript.\n\n**Funding:** The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a project grant from Science and Technology Program of Taicang City, China (No.TC2018JCYL25).\n\n**Conflict of interest statement:** The authors declare that there is no conflict of interest.\n\n**Ethics approval and consent to participate:** The study was approved by the institutional ethics committee of the First People's Hospital of Taicang (No.TCYY2019-KY0013) and was carried out in accordance with the guidelines of the Declaration of Helsinki. Written informed consent was formally obtained from all participants.\n\n**ORCID iD:** Dayang Chai ![](10.1177_2040622320922020-img1.jpg) <https://orcid.org/0000-0002-3389-5555>\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction {#acel12654-sec-0001}\n============\n\nHealthy mitochondria are critical for maintaining normal bioenergetically demanding activities of neurons. Such energy demand in neuronal axons is likely to be especially extreme due to the activities such as synaptic transmission, generating and propagating action potentials, and transporting biomaterials over a long distance. Deleterious mitochondrial changes such as a decrease in mitochondrial integrity and function are associated with aging and neurodegenerative diseases (Bratic & Larsson, [2013](#acel12654-bib-0004){ref-type=\"ref\"}; L\u00f3pez\u2010Ot\u00edn *et\u00a0al*., [2013](#acel12654-bib-0021){ref-type=\"ref\"}). Mitophagy is an important mitochondrial quality control mechanism that selectively eliminates damaged mitochondria by autophagy (Wang and Klionsky, [2014](#acel12654-bib-0502){ref-type=\"ref\"}; Wei *et\u00a0al*., [2015](#acel12654-bib-0044){ref-type=\"ref\"}). The process is regulated by the PTEN\u2010induced putative kinase 1 (PINK1) and the E3 ubiquitin ligase Parkin (Pickrell & Youle, [2015](#acel12654-bib-0032){ref-type=\"ref\"}), whose mutations can cause familial forms of Parkinson\\'s disease (PD).\n\nIt is widely assumed that damaged mitochondria accumulate in neurons if mitophagy fails, which causes PD in PINK1\u2010Parkin deficiency. For axons, the studies based on *in\u00a0vitro* cultured neurons have been controversial, especially regarding whether mitophagy occurs locally in axons. PINK1\u2010Parkin\u2010mediated mitophagy was reported to occur in neurons but restricted to the somatodendritic regions (Seibler *et\u00a0al*., [2011](#acel12654-bib-0038){ref-type=\"ref\"}; Cai *et\u00a0al*., [2012](#acel12654-bib-0006){ref-type=\"ref\"}). However, Maday *et\u00a0al*. ([2012](#acel12654-bib-0023){ref-type=\"ref\"}); Maday & Holzbaur ([2014](#acel12654-bib-0022){ref-type=\"ref\"}) showed that autophagosomes were generated at neurite tips, which fused with axonal mitochondria and together were transported to the neuronal soma for lysosome\u2010mediated degradation. By contrast, another study reported that autophagosomes were readily available in axons and were recruited to damaged mitochondria in a PINK1\u2010Parkin\u2010dependent manner followed by local turnover in axons (Ashrafi *et\u00a0al*., [2014](#acel12654-bib-0001){ref-type=\"ref\"}). Other than by inference, evidence *in\u00a0vivo* supporting an essential role of PINK1\u2010Parkin in regulating mitochondrial quality control in axons during aging is lacking. As axonal mitochondria may behave differently in neurons cultured *in\u00a0vitro* (Sung *et\u00a0al*., [2016](#acel12654-bib-0040){ref-type=\"ref\"}), it is important to elucidate how mitochondria are maintained in the lengthy axons under physiological conditions *in\u00a0vivo*.\n\nMitochondria are dynamic organelles that undergo constant fission and fusion. The fission--fusion balance is critical for determining mitochondrial morphology, distribution, abundance, and function (Chan, [2012](#acel12654-bib-0008){ref-type=\"ref\"}), and allows mitochondria to quickly adapt to different metabolic needs in different subcellular compartments or upon different energetic states (Wai & Langer, [2016](#acel12654-bib-0043){ref-type=\"ref\"}). Moreover, fission--fusion is also an important means to control mitochondrial quality, via removing the damaged portion of a mitochondrion by fission or via refreshing a dysfunctional mitochondrion by fusion to a healthy one (Schrepfer & Scorrano, [2016](#acel12654-bib-0037){ref-type=\"ref\"}). The key regulators of mitochondrial dynamics are an evolutionarily conserved family of dynamin\u2010related GTPases (Chan, [2012](#acel12654-bib-0008){ref-type=\"ref\"}). Specifically, mitofusin\u20101 (Mfn1) and \u20102 (Mfn2) mediate fusion of the mitochondrial outer membrane and optic atrophy 1 (Opa1) mediates that of the inner membrane; fission is mainly controlled by dynamin\u2010related protein 1 (Drp1) (Chan, [2012](#acel12654-bib-0008){ref-type=\"ref\"}; Bertholet *et\u00a0al*., [2016](#acel12654-bib-0003){ref-type=\"ref\"}). However, the role of mitochondrial fission--fusion, or its relative contribution compared to that of mitophagy, in axonal mitochondrial quality control *in\u00a0vivo* is unknown.\n\nDissecting the mechanism maintaining axonal mitochondria in animals during their lifetime has been challenging, which is at least in part due to the limitation of classical sample fixation and staining processes that grossly alter mitochondrial morphology. In this regard, the *Drosophila* wing model that we developed provides a unique system (Fang *et\u00a0al*., [2012](#acel12654-bib-0014){ref-type=\"ref\"}, [2013](#acel12654-bib-0015){ref-type=\"ref\"}; Fang & Bonini, [2015](#acel12654-bib-0013){ref-type=\"ref\"}) that allows direct visualization of the morphology of axonal mitochondria under near physiological conditions *in\u00a0vivo*. Using this system, we critically evaluate the role of PINK1\u2010Parkin\u2010dependent and PINK1\u2010Parkin\u2010independent mitophagy as well as mitochondrial fission--fusion in maintaining axonal integrity during aging. We find that fragmented mitochondria accumulate in aged axons, which is consistent with the general idea that mitochondrial quality and function decline with age (Bratic & Larsson, [2013](#acel12654-bib-0004){ref-type=\"ref\"}; L\u00f3pez\u2010Ot\u00edn *et\u00a0al*., [2013](#acel12654-bib-0021){ref-type=\"ref\"}). However, our study reveals a striking scarcity and dispensability of axonal mitophagy in neuronal aging, whether PINK1\u2010Parkin dependent or independent. Instead, our findings raise the possibility that mitophagy\u2010independent mechanisms such as fission--fusion play a central role in the maintenance of axonal mitochondria during normal aging.\n\nResults {#acel12654-sec-0002}\n=======\n\n*In vivo* neuroimaging reveals morphological alterations of axonal mitochondria during aging {#acel12654-sec-0003}\n--------------------------------------------------------------------------------------------\n\nTo characterize the dynamic morphological changes in neuronal mitochondria during aging, we used the GAL4/UAS system to label neuronal mitochondria with mitochondria\u2010localized GFP (mitoGFP) in the *Drosophila* wing nerve. The flies were aged to different time points and the axons were imaged in *live* fly wings without fixation by confocal microscopy. Mitochondria in the neuronal soma and axons of the costal wing nerve (the boxed area in Fig.\u00a0[1](#acel12654-fig-0001){ref-type=\"fig\"}A) were directly visualized *in\u00a0vivo* (Fig.\u00a0[1](#acel12654-fig-0001){ref-type=\"fig\"}B). The axonal mitochondria displayed a mixed population of morphology, from round or almost round (1.0\u00a0\u2264\u00a0length/width\u00a0\\<\u00a01.5), intermediate (1.5\u00a0\u2264\u00a0length/width\u00a0\\<\u00a02.0), tubular (2.0\u00a0\u2264\u00a0length/width\u00a0\\<\u00a05.0), to hyperfused (length/width\u00a0\u2265\u00a05) (Fig.\u00a0[1](#acel12654-fig-0001){ref-type=\"fig\"}C,D). We quantified the length and width of each mitochondrion in the costal wing nerve, and calculated the average length/width ratio at different ages in Fig.\u00a0[1](#acel12654-fig-0001){ref-type=\"fig\"}E. It is evident that both the average length/width ratio and the proportion of long mitochondria (tubular and hyperfused) were decreased, whereas those of short mitochondria (intermediate and round) were increased in aged flies (D30 and D50). Concurrently, the number of axonal mitochondria was increased with age (Fig.\u00a0[1](#acel12654-fig-0001){ref-type=\"fig\"}F). These data suggest that mitochondria became fragmented and accumulated in aged axons.\n\n![*In vivo* neuroimaging reveals the accumulation of fragmented mitochondria in aged axons. (A) A cartoon illustration of the *Drosophila* wing. The green line highlights the wing nerve in the costal, L1, and L3 wing veins, and the red dots denote the neuronal soma. (B, C) Representative confocal images of mitochondria in the axons of the *Drosophila* costal wing nerve (the blue box in A) at days D3, D15, D30, and D50. Mitochondria are labeled by mitoGFP using a *dpr*\u2010Gal4 driver. (C) A higher magnification of the area in the white box in B. Arrowheads: short; chevron: intermediate; arrow: tubular; and asterisks: hyperfused axonal mitochondria. (D) The axonal mitochondria imaged in (B) are classified into four groups based on the length/width ratio as specified. (E, F) The average length/width ratio (E) and number (F) of axonal mitochondria at indicated ages. Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a09--16 wings per group. \\**P\u00a0*\\<*\u00a0*0.05, \\*\\*\\**P\u00a0*\\<*\u00a0*0.001. Scale bars, 10\u00a0\u03bcm in (B) and 2\u00a0\u03bcm in (C).](ACEL-16-1180-g001){#acel12654-fig-0001}\n\nThe PINK1\u2010Parkin pathway is dispensable for axonal maintenance in both sensory and motor neurons in adult *Drosophila* {#acel12654-sec-0004}\n----------------------------------------------------------------------------------------------------------------------\n\nThe accumulation of fragmented mitochondria in axons might be a result of declined mitochondrial turnover during aging. Damaged axonal mitochondria in cultured neurons could be cleared by local mitophagy, which requires the function of PINK1 and Parkin (Ashrafi *et\u00a0al*., [2014](#acel12654-bib-0001){ref-type=\"ref\"}). Hence, we first investigated the contribution of the PINK1\u2010Parkin pathway to the age\u2010associated accumulation of axonal mitochondria *in\u00a0vivo*. The *Pink1* loss\u2010of\u2010function (LOF)\u2010mutant flies, *Pink1* ^*B9*^ and *Pink1* ^*5*^, exhibited collapsed thorax and abnormal wing posture (Fig.\u00a0[S1](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A,B, Supporting information), which was consistent with the mitochondrial defects in muscles as previously reported (Clark *et\u00a0al*., [2006](#acel12654-bib-0009){ref-type=\"ref\"}; Park *et\u00a0al*., [2006](#acel12654-bib-0030){ref-type=\"ref\"}). Unlike in muscles, the morphology of axonal mitochondria in the wing axons of *Pink1* ^*B9*^ and *Pink1* ^*5*^ flies looked mostly normal (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}A). Neither the average mitochondrial length/width ratio nor the number of mitochondria showed a significant difference between *Pink1* mutants and *w* ^*1118*^ control flies (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}B,C). Moreover, the axonal integrity of the *Pink1* mutants was well maintained and no axonal degeneration was observed during aging (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}D).\n\n![Flies with Pink1\u2010Parkin deficiency exhibit normal mitochondrial morphology and intact axonal integrity. (A) Axonal mitochondria labeled by mitoGFP in the costal wing nerve of control (*w* ^*1118*^), *Pink1* ^*B9*^ *,* and *Pink1* ^*5*^ flies at D3 or D30. (B\u2010C) Quantification of length/width ratio (B) and the number (C) of the axonal mitochondria. (D) Representative *in\u00a0vivo* images of distal axons of the L1 wing nerve of control (*w* ^*1118*^), *Pink1* ^*B9*^ *,* and *Pink1* ^*5*^ flies at D3 or D30. (E) Representative images of axonal mitochondria of RNAi\u2010*Luc* (control) and RNAi\u2010*Parkin* flies at D3 or D30, the average length/width ratio, and the number of axonal mitochondria are quantified in (F, G). Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a08--10 wings per group. Two\u2010way ANOVA shows no statistically significant difference between any of the mutant or RNAi groups and their respective controls, whereas the mitochondrial changes with age are significant (*P\u00a0*\\<*\u00a0*0.05) except for (F). (H) Representative *in\u00a0vivo* images of neuronal soma and proximal axons of the L1 nerve of indicated genotypes at D3 and D30. No axonal degeneration is detected. Scale bars, 5\u00a0\u03bcm in (A) and (E), and 10\u00a0\u03bcm in (D) and (H).](ACEL-16-1180-g002){#acel12654-fig-0002}\n\nSimilarly, despite an age\u2010associated decrease in *Parkin* expression (Fig.\u00a0[S1](#acel12654-sup-0001){ref-type=\"supplementary-material\"}C, Supporting information), downregulation of *Parkin* in the wing nerve did not result in the accumulation of fragmented mitochondria (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}E--G) or axonal degeneration (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}H). The knockdown (KD) efficiency was examined by quantitative RT--PCR (qPCR) (Fig.\u00a0[S1](#acel12654-sup-0001){ref-type=\"supplementary-material\"}D, Supporting information). In contrast, downregulation of *Parkin* or *Pink1* in muscles using the same transgenic RNAi strains caused muscle degeneration and collapsed thoraces (Fig.\u00a0[S1](#acel12654-sup-0001){ref-type=\"supplementary-material\"}E, Supporting information), confirming their essential role in the muscle cells. As the fly wing nerve consists of only sensory neurons (Nakamura *et\u00a0al*., [2002](#acel12654-bib-0027){ref-type=\"ref\"}), we asked whether the lack of axonal degeneration in PINK1\u2010Parkin deficiency was sensory neuron\u2010specific. We then examined the axons of the motor neurons in the adult *Drosophila* leg (Fig.\u00a0[S2](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A, Supporting information). Similar to the sensory wing nerve, the motor neurons of the *Pink1* LOF mutants showed no axonal degeneration, even in aged flies (Fig.\u00a0[S2](#acel12654-sup-0001){ref-type=\"supplementary-material\"}B, Supporting information).\n\nAs overexpression of *Parkin* in adult neurons was shown to enhance mitochondrial activity and extend longevity in *Drosophila* (Rana *et\u00a0al*., [2013](#acel12654-bib-0035){ref-type=\"ref\"}), we were keen to test whether upregulation of PINK\u2010Parkin could promote turnover of fragmented mitochondria and improve axonal maintenance during aging. Unexpectedly, overexpression of *Pink1* or *Parkin* in the wing neurons led to extremely fragmented mitochondria (Fig.\u00a0[S3](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A,B, Supporting information). Expression of the enzymatically inactive mutants, *Pink1* ^*L464P*^ (Song *et\u00a0al*., [2013](#acel12654-bib-0039){ref-type=\"ref\"}), *Parkin* ^*T240R*^ *,* or *Parkin* ^*R275W*^ (Lee *et\u00a0al*., [2010](#acel12654-bib-0020){ref-type=\"ref\"}; Kim *et\u00a0al*., [2013](#acel12654-bib-0017){ref-type=\"ref\"}) did not show such effect, indicating that the impact of upregulating PINK1\u2010Parkin on axonal mitochondrial morphology requires their enzymatic activity. Furthermore, in striking contrast to LOF and KD of *Pink1* and *Parkin* (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}D,H), the wing nerve of *Pink1\u2010* or *Parkin*\u2010overexpressing flies degenerated rapidly with age (Fig.\u00a0[S3](#acel12654-sup-0001){ref-type=\"supplementary-material\"}C,E, Supporting information). This is likely due to a mitophagy\u2010independent function of Parkin in regulating Marf, the *Drosophila* homologue of mammalian mitofusins that promote mitochondria fusion (Chan, [2012](#acel12654-bib-0008){ref-type=\"ref\"}; Schrepfer & Scorrano, [2016](#acel12654-bib-0037){ref-type=\"ref\"}; Fig.\u00a0[S3](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A, Supporting information). Parkin was known to ubiquitinate Marf for proteasome\u2010mediated degradation and resulted in shortened mitochondria in insect cells (Poole *et\u00a0al*., [2010](#acel12654-bib-0033){ref-type=\"ref\"}; Ziviani *et\u00a0al*., [2010](#acel12654-bib-0045){ref-type=\"ref\"}). Consistently, we found that both mitochondrial morphology (Fig.\u00a0[S3](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A,B, Supporting information) and axonal integrity (Fig.\u00a0[S3](#acel12654-sup-0001){ref-type=\"supplementary-material\"}D,E, Supporting information) of *Parkin*\u2010overexpressing neurons were significantly improved by co\u2010expression of *Marf*. This result is in line with the idea that mitophagy\u2010independent PINK1\u2010Parkin functions may cause or contribute to the pathogenesis of PD (Pryde *et\u00a0al*., [2016](#acel12654-bib-0034){ref-type=\"ref\"}).\n\nMitophagy rarely occurs in intact axons *in\u00a0vivo* during normal aging {#acel12654-sec-0005}\n---------------------------------------------------------------------\n\nIn addition to the PINK1\u2010Parkin\u2010dependent mitophagy, mitophagy can take place independently of PINK1\u2010Parkin (Wei *et\u00a0al*., [2015](#acel12654-bib-0044){ref-type=\"ref\"}). Hence, the accumulation of axonal mitochondria might also be due to an age\u2010associated decline of autophagic clearance of damaged mitochondria that is independent of PINK\u2010Parkin. To test this possibility, we generated the fly strains to express mCherry\u2010labeled Atg8a (an homologue of mammalian LC3) in the wing nerve to visualize and investigate autophagy by *in\u00a0vivo* imaging. Induction of autophagy leads to the cleavage and translocation of Atg8a to the autophagosome membrane, which transforms mCherry\u2010Atg8a signal from diffuse to punctate. Formation of Atg8a/LC3 puncta is widely used as an autophagosome marker (Klionsky *et\u00a0al*., [2016](#acel12654-bib-0018){ref-type=\"ref\"}). As illustrated in Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}A, in the neuronal soma, autophagosomes (mCherry\u2010Atg8a puncta) were frequently observed in both young (D3) and aged (D30) flies. The neuronal soma maintained relatively constant basal levels of autophagy during aging, as neither the number nor the size of mCherry\u2010Atg8a puncta was significantly changed in the soma during aging (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}B,C). In contrast, autophagosome puncta were sparse in young axons. With age, an increasing number of mCherry\u2010Atg8a puncta were detected in axons (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}A,F); the size of axonal autophagosomes showed a trend to increase although this was not statistically significant (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}G). These results suggested an increase in autophagy induction and/or a decrease in autophagic flux in axons during aging.\n\n![Mitophagy is both rare and dispensable for intact axons *in\u00a0vivo* during aging. (A--I) Axonal mitophagy rarely occurs *in\u00a0vivo* under physiological conditions. (A) Representative images showing neuronal soma (left) and axons (right) with autophagosomes (AP) labeled by mCherry\u2010Atg8a and mitochondria labeled by mitoGFP. mCherry\u2010Atg8a puncta are commonly seen in the soma of both young and aged flies; axonal AP are rarely observed in young flies but significantly increase with age. Arrows, colocalization of mitoGFP and mCherry\u2010Atg8a puncta in the soma; arrowheads, nearby but not colocalized mitochondria and autophagosome in axons. The number and size of somatic AP (B\u2010C) and axonal AP (F\u2010G) as well as colocalization of mitoGFP and mCherry\u2010Atg8a in the soma (D, E) are quantified. Colocalization of mitochondria and autophagosomes is almost never detected in axons *in\u00a0vivo* even in aged animals; hence, the colocalization occurrence is essentially zero in (H, I). Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a07--12 wings per group. \\**P\u00a0*\\<*\u00a0*0.05; ns, not significant; ud, undetected. (J--N) Blocking axonal autophagy has little effect on the maintenance of axonal mitochondria or axonal integrity *in\u00a0vivo*. (J) Representative images showing AP (mCherry\u2010Atg8a, arrow) and mitochondria (mitoGFP) in the costal wing axons of RNAi\u2010*Luc*,RNAi\u2010*Atg12*, and RNAi\u2010*Atg17* flies. (K) The percentage of flies containing axonal AP in the entire wing nerve. (L, M) Quantification of the length/width ratio (L) and the number (M) of the axonal mitochondria. Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a011--14 wings per group. \\**P\u00a0*\\<*\u00a0*0.05. (N) Representative images of the distal wing nerve of the indicated genotypes at D3 and D30. No sign of axonal degeneration is observed in the RNAi\u2010*Atg12* or RNAi\u2010*Atg17* flies. Scale bars, 5\u00a0\u03bcm in (A) and (J), and 10\u00a0\u03bcm in (N).](ACEL-16-1180-g003){#acel12654-fig-0003}\n\nAs the numbers of fragmented mitochondria and autophagosomes were both increased in aged axons, we asked whether this concurrence represented an increase in mitophagy in axons. A previous study showed that depolarized axonal mitochondria became colocalized with autophagosomes, leading to local mitophagy in axons (Ashrafi *et\u00a0al*., [2014](#acel12654-bib-0001){ref-type=\"ref\"}). Hence, we carefully examined colocalization of mitoGFP and mCherry\u2010Atg8a puncta in the wing nerve. Unexpectedly, however, we found mitoGFP and mCherry\u2010Atg8a puncta were essentially not colocalized *in\u00a0vivo*, even in aged flies (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}A, arrowheads). Quantification showed that neither the proportion of mCherry\u2010Atg8a puncta containing mitochondria nor the percentage of mitochondria colocalized with mCherry\u2010Atg8a changed significantly during aging (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}H--I). Of note, among over 110 fly wings examined in this study, we only detected two individual cases of a mitochondrion colocalized with mCherry\u2010Atg8a puncta in the entire wing nerve, which is composed of more than 200 axons (Fang *et\u00a0al*., [2013](#acel12654-bib-0015){ref-type=\"ref\"}). As such, the occurrence of axonal mitophagy under physiological conditions *in\u00a0vivo* is \\<1 of 10\u00a0000 axons, suggesting that in axons virtually no mitophagy occurs, regardless of age. Similarly, McWilliams *et\u00a0al*. ([2016](#acel12654-bib-0024){ref-type=\"ref\"}) reported that the majority of mitochondrial turnover occurred in the soma of Purkinje neurons with minimal mitophagy observed in mouse axons *in\u00a0vivo*.\n\nBlocking autophagy in the *Drosophila* wing nerve has a minimal effect on mitochondrial turnover or axonal integrity during aging {#acel12654-sec-0006}\n---------------------------------------------------------------------------------------------------------------------------------\n\nAlthough axonal mitophagy rarely occurred *in\u00a0vivo*, it did not exclude the possibility that occasional axonal mitophagy played an essential role. As both PINK1\u2010Parkin\u2010dependent and PINK1\u2010Parkin\u2010independent mitophagy pathways rely on the autophagy machinery for the final clearance of mitochondria, we sought to examine the effect of blocking autophagy on mitochondrial turnover and axonal integrity during aging.\n\nAutophagy is controlled by a family of evolutionarily conserved ATG genes. One of the core ATG genes is *Atg12*, required for the formation of autophagosomes in promoting the conjugation of ATG8 protein to the lipid phosphatidylethanolamine (Ktistakis & Tooze, [2016](#acel12654-bib-0019){ref-type=\"ref\"}). As the aging context limited the use of the null mutants in this study, we could only assess the impact of these genes in a partial compromise situation. Nevertheless, in the RNAi\u2010*Atg12* flies, both the number and size of mCherry\u2010Atg8a puncta in the soma were significantly decreased; meanwhile, diffused mCherry\u2010Atg8a signal was increased (Fig.\u00a0[S4](#acel12654-sup-0001){ref-type=\"supplementary-material\"}A--C, Supporting information). The KD efficiency of RNAi\u2010*Atg12* and RNAi\u2010*Atg17* was examined in Fig.\u00a0[S4](#acel12654-sup-0001){ref-type=\"supplementary-material\"}D (Supporting information). Importantly, despite the low basal level of autophagy in axons (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}J, control flies, RNAi\u2010*luc*), *Atg12* KD caused a further decrease in axonal autophagy (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}J, RNAi\u2010*Atg12*), which was seen as a drastic reduction in flies with detectable mCherry\u2010Atg8a puncta in axons (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}K, from \\~80% in the control to only \\~10% in the RNAi\u2010*Atg12*). As such, axonal autophagy was inhibited in the wing nerve in the majority of RNAi\u2010*Atg12* flies. Nonetheless, the maintenance of axonal mitochondria was not impaired, as RNAi\u2010*Atg12* flies showed no difference in either the average length/width ratio or the density of axonal mitochondria (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}J,L,M).\n\nThe lack of a deleterious effect on axonal mitochondrial turnover was further confirmed by manipulation of another core ATG gene *Atg17* (required for the autophagosome assembly, Ktistakis & Tooze, [2016](#acel12654-bib-0019){ref-type=\"ref\"}) in the wing nerve (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}J--M). In nearly 100% of RNAi\u2010*Atg17* flies, an axonal autophagosome was never found in over two hundred axons examined in the entire wing nerve (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}K). Nevertheless, no degeneration of a single axon was observed, even in aged RNAi\u2010*Atg17* flies (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}N, D30). These results further support the conclusion that axonal mitochondria do not rely on mitophagy for their maintenance during normal aging.\n\nMitophagy\u2010independent mechanisms may regulate mitochondrial maintenance and axonal integrity during aging {#acel12654-sec-0007}\n---------------------------------------------------------------------------------------------------------\n\nNeuronal integrity and functions rely on mitochondria for ATP production. Because we revealed that mitophagy was dispensable for axons under physiological conditions *in\u00a0vivo*, neurons must use other means to maintain axonal mitochondria and axonal integrity. Mitochondrial quality and abundance is also controlled by mitochondrial transport and fission--fusion (Saxton & Hollenbeck, [2012](#acel12654-bib-0036){ref-type=\"ref\"}; Schrepfer & Scorrano, [2016](#acel12654-bib-0037){ref-type=\"ref\"}). Indeed, we and others previously showed that proper mitochondrial transport and distribution are essential for the integrity of *Drosophila* axons (Avery *et\u00a0al*., [2012](#acel12654-bib-0002){ref-type=\"ref\"}; Fang *et\u00a0al*., [2012](#acel12654-bib-0014){ref-type=\"ref\"}). As fission--fusion also regulates the morphology and turnover of mitochondria (Chan, [2012](#acel12654-bib-0008){ref-type=\"ref\"}; Schrepfer & Scorrano, [2016](#acel12654-bib-0037){ref-type=\"ref\"}), we next investigated the possibility that the accumulation of fragmented mitochondria in aged axons resulted from an age\u2010associated change in the fission--fusion dynamics. Interestingly, we found that *Opa1* expression in the fly head became reduced with age (Fig.\u00a0[4](#acel12654-fig-0004){ref-type=\"fig\"}A, D30). However, as a similar expression decrease was detected with *Parkin* (Fig.\u00a0[S1](#acel12654-sup-0001){ref-type=\"supplementary-material\"}C, Supporting information), but KD of *Parkin* did not impair mitochondrial turnover or axonal integrity (Fig.\u00a0[2](#acel12654-fig-0002){ref-type=\"fig\"}E--H), it was yet to be proven that fission--fusion was indeed essential for axonal maintenance during aging.\n\n![The fission--fusion balance is critically required for the maintenance of axonal mitochondria and axonal integrity during aging. (A) The relative mRNA levels of mitochondrial fission--fusion genes in the fly head at the indicated time points. *Opa1* expression is significantly decreased at D30. (B--D) Representative images of axonal mitochondria of indicated genotypes, and the axonal mitochondrial length and number are quantified in (C) and (D), respectively. Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a08--14 wings per group except for RNAi\u2010*Marf*, of which only three adult flies were available for imaging due to early lethality. (E) The knockdown efficiency of the above RNAi lines was examined by qPCR. (F, G) Age\u2010dependent axonal degeneration by downregulation of the fission--fusion genes in the adult fly wing nerve. The representative images of the L1 wing nerve of indicated genotypes at D3, D5, and D30 are shown in (F) and the degeneration scores are quantified in (G) as described in the Methods. Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a06--8 wings per genotype per time point except for RNAi\u2010*Marf* (*n*\u00a0=\u00a03). *n.a*., not available; because KD of *Marf* or *Opa1* by the *dpr*\u2010Gal4 driver (which is highly expressed in the adult wing neurons that facilitates the *in\u00a0vivo* imaging but is also expressed in neurons elsewhere) causes early lethality in young adults, the imaging data at later time points for RNAi\u2010*Marf* and RNAi\u2010*Opa1* flies are unavailable. \\**P\u00a0*\\<*\u00a0*0.05, \\*\\**P\u00a0*\\<*\u00a0*0.01, \\*\\*\\**P\u00a0*\\<*\u00a0*0.001; ns, not significant. Scale bars, 2\u00a0\u03bcm in (B) and 10\u00a0\u03bcm in (F).](ACEL-16-1180-g004){#acel12654-fig-0004}\n\nAs shown in Fig.\u00a0[4](#acel12654-fig-0004){ref-type=\"fig\"}B--D, downregulation of *Marf* or *Opa1* significantly decreased, whereas KD of *Drp1* increased, the length of axonal mitochondria in the wing nerve. We then examined the impact on axonal integrity. RNAi\u2010*Marf* flies showed remarkable axonal degeneration as early as D3, RNAi\u2010*Opa1* flies at D15, and RNAi\u2010*Drp1* flies at D30 (Fig.\u00a0[4](#acel12654-fig-0004){ref-type=\"fig\"}E). Of note, because KD of *Marf* or *Opa1* using the *dpr*\u2010Gal4 driver (which is highly expressed in the adult wing neurons to facilitate *in\u00a0vivo* imaging, but is also expressed in neurons elsewhere) caused early lethality of young adults, the axonal imaging data at later time points for RNAi\u2010*Marf* (D15 and D30) and RNAi\u2010*Opa1* flies (D30) were unavailable. Nevertheless, downregulation of the fission--fusion genes by genetic manipulations caused age\u2010dependent, progressive axonal degeneration. This result suggests that the change in their expression during normal aging should not be overtly large; otherwise, aging itself would cause spontaneous axonal degeneration. Indeed, the small (\\~30%) but significant reduction in *Opa1* expression in aged neurons (Fig.\u00a0[4](#acel12654-fig-0004){ref-type=\"fig\"}A) is consistent with the moderate alteration of mitochondrial morphology in axons during normal aging.\n\nInhibition of fission--fusion but not mitophagy in adult neurons significantly accelerates aging in *Drosophila* {#acel12654-sec-0008}\n----------------------------------------------------------------------------------------------------------------\n\nIt was possible that disruption of the PINK1\u2010Parkin pathway or the autophagy machinery led to functional deficits before morphological changes were detectable. The climbing assay is widely used to evaluate the function of neurons and muscles in *Drosophila* models of neurodegenerative diseases including PD. As the function of mitophagy/autophagy in neural development might compromise the analysis, we used an inducible, pan\u2010neuronal driver *elav*\u2010GeneSwitch (*elav*\u2010GS, Osterwalder *et\u00a0al*., [2001](#acel12654-bib-0029){ref-type=\"ref\"}) and induced the expression of the RNAi transgenes at adulthood only. As shown in Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}A, no significant difference in the climbing ability was found in RNAi\u2010*Pink1* or RNAi\u2010*Parkin* flies. In contrast, KD of *Pink1* or *Parkin* in muscles (*Mef2*\u2010Gal4) caused age\u2010dependent climbing defects as early as D24 in RNAi\u2010*Parkin* and D36 in RNAi\u2010*Pink1* flies (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}B). In addition, KD of the core ATG gene *Atg12* or *Atg17* in adult neurons also led to accelerated decline of climbing ability (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}C). This was even more dramatic when the flies were raised at an elevated environmental temperature of 29\u00a0\u00b0C (Fig.\u00a0[S4](#acel12654-sup-0001){ref-type=\"supplementary-material\"}E, Supporting information). Heat stress is known to disturb the proteostasis and require the function of autophagy to maintain the homeostasis (Dokladny *et\u00a0al*., [2015](#acel12654-bib-0012){ref-type=\"ref\"}). Thus, although the autophagy machinery is dispensable for the turnover of axonal mitochondria (Fig.\u00a0[3](#acel12654-fig-0003){ref-type=\"fig\"}J), these results confirmed the essential role of autophagy in neuronal aging likely due to the important functions in regulating the proteostasis in neurons.\n\n![Inhibition of fission--fusion but not mitophagy in adult neurons significantly accelerates aging. (A\u2010C) Climbing assays of flies with (A) *Pink1* or *Parkin* downregulated in adult neurons by *elav\u2010* GS, (B) in muscles by *Mef2*\u2010Gal4 (B), or (C) *Atg12* or *Atg17* downregulated in adult neurons. The locomotive ability was assessed as the average percentage of flies climbing over 5\u00a0cm within 15\u00a0s. Data are shown as mean\u00a0\u00b1\u00a0SEM,*n*\u00a0=\u00a020 flies per vial and 6--8 vials each group. \\**P\u00a0*\\<*\u00a0*0.05, \\*\\**P\u00a0*\\<*\u00a0*0.01; ns, not significant. (D) Knockdown of *Pink1* or *Parkin* in all cells by the *Da*\u2010Gal4 driver dramatically reduces the longevity. (E\u2010G) Lifespan assays of the flies with adult\u2010onset, neuronal downregulation (*elav*\u2010GS) of *Pink1* or *Parkin* (E), *Atg12* or *Atg17* (F), or *Marf*,*Opa1,* or *Drp1* (G). A copy of UAS\u2010*Dcr2* was co\u2010expressed to boost the RNAi knockdown efficiency in neurons (Ni *et\u00a0al*., [2008](#acel12654-bib-0028){ref-type=\"ref\"}), but was not needed with the *Mef2*\u2010Gal4 or *Da*\u2010Gal4 drivers, in which cases the RNAi\u2010*mCherry* was used as a control. *n*, the number of flies tested for each genotype is indicated. (H) Summary of the median lifespans, shown as mean\u00a0\u00b1\u00a0SEM, \\**P\u00a0*\\<*\u00a0*0.01, \\*\\**P\u00a0*\\<*\u00a0*0.001, \\*\\*\\**P\u00a0*\\<*\u00a0*0.0001.](ACEL-16-1180-g005){#acel12654-fig-0005}\n\nFinally, as an attempt to compare the overall involvement of mitophagy and fission--fusion in neuronal aging, we examined the effect of their downregulation on longevity. Similar to their LOF mutants (Greene *et\u00a0al*., [2003](#acel12654-bib-0016){ref-type=\"ref\"}; Cha *et\u00a0al*., [2005](#acel12654-bib-0007){ref-type=\"ref\"}; Clark *et\u00a0al*., [2006](#acel12654-bib-0009){ref-type=\"ref\"}; Park *et\u00a0al*., [2006](#acel12654-bib-0030){ref-type=\"ref\"}), ubiquitous KD of *Pink1* or *Parkin* in all cells using a *Daughterless*\u2010Gal4 (*Da*\u2010Gal4) driver dramatically shortened lifespan (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}D,H). To our great surprise, however, the adult\u2010onset, neuronal downregulation of *Pink1* or *Parkin* (*elav\u2010*GS) did not decrease the longevity; instead, the RNAi\u2010*Parkin* and one of the RNAi\u2010*Pink1* lines showed a puzzling extension of the lifespan (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}E,H). In addition, KD of the core autophagy genes *Atg12* or *Atg17* in adult neurons led to a modest reduction in the lifespan (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}F,H). In contrast, downregulation of the fission--fusion genes in adult neurons all dramatically reduced the longevity (Fig.\u00a0[5](#acel12654-fig-0005){ref-type=\"fig\"}G,H). Of note, disruption of fission--fusion consequently leads to impaired mitochondrial functions (Pich *et\u00a0al*., [2005](#acel12654-bib-0031){ref-type=\"ref\"}), which may add to the deleterious effects on axonal integrity and longevity. Nevertheless, our data demonstrate that mitophagy is dispensable for axonal integrity and physiological aging in neurons, whereas mitophagy\u2010independent mechanisms such as fission--fusion play an essential role in the maintenance of axonal mitochondria and neural integrity (Fig.\u00a0[6](#acel12654-fig-0006){ref-type=\"fig\"}).\n\n![A schematic model of differential requirements of mitophagy and fission--fusion in the maintenance of axonal mitochondria and axonal integrity during aging. During aging, fragmented mitochondria accumulate in axons and the basal levels of axonal autophagy increase. However, mitophagy (evident by the colocalization of mitochondria and autophagosomes) is predominantly observed in the soma but rarely in axons *in\u00a0vivo*, even in aged neurons. Further, disrupting the PINK1\u2010Parkin pathway (A) or blocking axonal autophagy by knockdown of the core autophagy genes *Atg12* or *Atg17* (B) does not impair mitochondrial turnover or axonal integrity. Instead, knockdown of the fission--fusion genes *Opa1 Marf* or *Drp1* (C) causes profound changes to the morphology and abundance of axonal mitochondria, which leads to age\u2010dependent, progressive axonal degeneration and significantly shortens lifespan. Together, our study indicates that mitophagy is dispensable for axonal maintenance *in\u00a0vivo*, whereas mitophagy\u2010independent mechanisms such as mitochondrial fission--fusion may play an essential and central role in the maintenance of axonal mitochondria and axonal integrity during normal aging.](ACEL-16-1180-g006){#acel12654-fig-0006}\n\nDiscussion {#acel12654-sec-0009}\n==========\n\nNeuronal aging is known to be associated with deleterious changes in mitochondria including a decrease in mitochondrial biogenesis, the respiratory chain efficacy and ATP generation, an increase in the production of reactive oxygen species, accumulation of mitochondrial DNA mutations, and reduction in mitochondrial transport and turnover (Green *et\u00a0al*., [2011](#acel12654-bib-0501){ref-type=\"ref\"}; L\u00f3pez\u2010Ot\u00edn *et\u00a0al*., [2013](#acel12654-bib-0021){ref-type=\"ref\"}; Bratic & Larsson, [2013](#acel12654-bib-0004){ref-type=\"ref\"}). The development of the *in\u00a0vivo* imaging paradigm of the *Drosophila* wing nerve has enabled the systematic characterization of mitochondrial morphology in axons during aging in this study. Moreover, taking the advantage of vast genetic tools in *Drosophila*, we have manipulated mitochondrial quality control genes in a spatially and temporally controlled manner, and directly visualized the consequence on mitochondria and axonal integrity *in\u00a0vivo*. The imaging data show a clear change in mitochondrial heterogeneity from long, tubular mitochondria in young axons toward short, round mitochondria in aged axons. This change is concurrent with an increased number of axonal mitochondria. Thus, aging is associated with an accumulation of fragmented mitochondria in axons. Our finding is consistent with a recent study showing that the axonal mitochondria size decreased in aged *C.\u00a0elegans* (Morsci *et\u00a0al*., [2016](#acel12654-bib-0026){ref-type=\"ref\"}). This is in addition to the reported decline of axonal mitochondrial transport during aging in nematode neurons (Morsci *et\u00a0al*., [2016](#acel12654-bib-0026){ref-type=\"ref\"}) and mouse retinal ganglion cells (Takihara *et\u00a0al*., [2015](#acel12654-bib-0041){ref-type=\"ref\"}).\n\nUnlike in the neuronal soma where lysosomes are abundant and mitochondria can be readily cleared by mitophagy, mitochondria in distal axons face unique challenges. It remains debated whether damaged axonal mitochondria are transported back to the soma for turnover (Maday *et\u00a0al*., [2012](#acel12654-bib-0023){ref-type=\"ref\"}; Maday & Holzbaur, [2014](#acel12654-bib-0022){ref-type=\"ref\"}) or are degraded locally in axons by PINK1\u2010Parkin\u2010dependent mitophagy (Ashrafi *et\u00a0al*., [2014](#acel12654-bib-0001){ref-type=\"ref\"}). In a recent study, Sung *et\u00a0al*. ([2016](#acel12654-bib-0040){ref-type=\"ref\"}) elegantly demonstrated that the appearance and the regulation of axonal mitochondria and autophagy differ between *in\u00a0vivo* conditions and *in\u00a0vitro* setups -- autophagosomes and mitochondria did not colocalize in axons of *Drosophila* motor neurons *in\u00a0vivo*; however, when these neurons were cultured *in\u00a0vitro*, numerous autophagosomes were colocalized with mitochondria in axons, even in the absence of any drug\u2010induced mitochondrial damage. In this study, we provide *in\u00a0vivo* evidence that the PINK1\u2010Parkin pathway is not required for mitochondrial turnover or axonal integrity in either sensory (wing) or motor (leg) neurons of the adult *Drosophila* during aging. Certainly, we cannot rule out the possibility that PINK1\u2010Parkin deficiency may still compromise the ultrastructure or function of axonal mitochondria. Interestingly, Devireddy *et\u00a0al*. ([2015](#acel12654-bib-0011){ref-type=\"ref\"}) showed that *Pink1* deletion led to a slight decrease in mitochondrial membrane potential and abnormal mitochondrial morphology in the soma of *Drosophila* larval neurons but did not change mitochondrial density or length in axons. Furthermore, the results of climbing and lifespan assays strongly suggest that the PINK1\u2010Parkin pathway is dispensable for adult neurons at the animal level. And the reduced longevity of the LOF mutants of *Pink1* and *Parkin* (Greene *et\u00a0al*., [2003](#acel12654-bib-0016){ref-type=\"ref\"}; Cha *et\u00a0al*., [2005](#acel12654-bib-0007){ref-type=\"ref\"}; Clark *et\u00a0al*., [2006](#acel12654-bib-0009){ref-type=\"ref\"}; Park *et\u00a0al*., [2006](#acel12654-bib-0030){ref-type=\"ref\"}) is likely due to their essential role in non\u2010neuronal cells such as muscles. As the difference 'between neurons and muscles' may represent a cell type\u2010specific expression and/or requirement of PINK1\u2010Parkin, our data do not exclude the possibility that the PINK1\u2010Parkin pathway plays an essential role in select subtypes of neurons, such as dopaminergic neurons, and thus contributes to the pathogenesis of PD.\n\nThe autophagy machinery is required for the final turnover of mitochondria in all mitophagy pathways (Pickrell & Youle, [2015](#acel12654-bib-0032){ref-type=\"ref\"}; Wei *et\u00a0al*., [2015](#acel12654-bib-0044){ref-type=\"ref\"}). Our *in\u00a0vivo* imaging data reveal that, accompanying the accumulation of axonal mitochondria, the number of axonal autophagosomes increases with age. However, mitochondria and autophagosomes are essentially not colocalized in axons *in\u00a0vivo*, even in aged neurons. This is in sharp contrast to the neuronal cell bodies where colocalization of mitochondria and autophagosomes is frequently observed, lending support to an important role of mitophagy in the soma (Cai *et\u00a0al*., [2012](#acel12654-bib-0006){ref-type=\"ref\"}; Sung *et\u00a0al*., [2016](#acel12654-bib-0040){ref-type=\"ref\"}). Moreover, knockdown of *Atg12* or *Atg17* inhibits axonal autophagy but does not affect mitochondrial turnover or axonal integrity. As all mitophagy pathways rely on the autophagy machinery, our data demonstrate that neither PINK1\u2010Parkin\u2010dependent nor PINK1\u2010Parkin\u2010independent mitophagy is required for maintaining axonal mitochondria *in\u00a0vivo*. Although the function of PINK1\u2010Parkin in axons has been debated (Ashrafi *et\u00a0al*., [2014](#acel12654-bib-0001){ref-type=\"ref\"}; Devireddy *et\u00a0al*., [2015](#acel12654-bib-0011){ref-type=\"ref\"}; Sung *et\u00a0al*., [2016](#acel12654-bib-0040){ref-type=\"ref\"}), our study is the first to show that overall mitophagy is dispensable for axonal maintenance during normal aging (to the best of our knowledge). Nevertheless, blocking autophagy in adult *Drosophila* neurons results in accelerated aging and slightly shortened lifespan, signifying the importance of autophagy in neuronal aging.\n\nMitochondrial fission--fusion is required for proper axonal projections and synaptic development and is involved in human neurological disorders (Burt\u00e9 *et\u00a0al*., [2015](#acel12654-bib-0005){ref-type=\"ref\"}; Bertholet *et\u00a0al*., [2016](#acel12654-bib-0003){ref-type=\"ref\"}). However, whether fission--fusion constitutes an essential quality control mechanism for mitochondria in axons and how it impacts on neuronal aging compared to mitophagy are unclear. We show that the proper levels of the fission--fusion genes are critically required in axons during aging and that disturbance of the fission--fusion balance in adult neurons dramatically reduces the longevity. When comparing mitophagy and fission--fusion, it is important to bear in mind that axons can be hundreds or even thousands of times longer than the size of the soma. Should neurons routinely turnover axonal mitochondria by mitophagy, the amount of energy it would cost to constantly transport mitochondria from the soma to distal axons and then back to the soma for lysosome\u2010mediated degradation would be unmanageably enormous and wasteful. Fission--fusion provides an important alternative mechanism to control mitochondrial quality by 'repairing and reusing' mitochondria (Schrepfer & Scorrano, [2016](#acel12654-bib-0037){ref-type=\"ref\"}). As such, maintaining axonal mitochondria by fission--fusion can be much more 'cost\u2010effective' than mitophagy. This alternative model is especially attractive as damage to axonal mitochondria under physiological conditions during aging is likely mild and gradual.\n\nTogether, our *in\u00a0vivo* data strongly suggest that maintenance of axonal mitochondria does not require mitophagy, whereas mitophagy\u2010independent mechanisms such as fission--fusion may play a pivotal role in maintaining axonal mitochondria and neural integrity during normal aging. Finally, it is important to point out that there are multiple levels of mitochondrial quality control, including mitochondrial proteases, ubiquitin\u2010dependent turnover, vesicle transport to lysosomes, and the new emerging mechanisms of mitochondrial 'expulsion' that have been recently reported as an alternative means of autophagy and PINK1\u2010Parkin\u2010independent turnover (Melentijevic *et\u00a0al*., [2017](#acel12654-bib-0025){ref-type=\"ref\"}). Each of them may be active or inactive under certain conditions, in select types of cells, or at restricted subcellular compartments. Thus, it will be of great interest to investigate the role of mitophagy\u2010independent mechanisms of mitochondrial quality control in aging in the future.\n\nExperimental procedures {#acel12654-sec-0010}\n=======================\n\nFly stocks {#acel12654-sec-0011}\n----------\n\nFlies tested in this study were raised on standard cornmeal media and maintained at 25\u00a0\u00b0C and 60% relative humidity. The following strains were obtained from the Bloomington *Drosophila* Stock Center (BDSC): *w* ^*1118*^ (5905), *Pink1* ^B9^ (34749), *Pink1* ^5^ (51649), *Da*\u2010Gal4 (5460), *elav\u2010*GS (43642), UAS\u2010mCherry\u2010*Atg8a* (37750), UAS\u2010*LacZ* (8529), UAS\u2010*Pink1* (51648), UAS\u2010*Parkin* (51651), UAS\u2010RNAi\u2010*Parkin* (38333), UAS\u2010RNAi\u2010*Atg12* (34675), UAS\u2010RNAi\u2010*Atg17* (36918), and UAS\u2010RNAi\u2010*Marf* (55189). The following strains were obtained from the Tsinghua Fly Center (TFC): UAS\u2010RNAi\u2010*Opa1* (THU0811), UAS\u2010RNAi\u2010*Drp1* (TH02258.N), *Mef2*\u2010Gal4 (THJ0244). For neuronal expression (*Dpr*\u2010Gal4, *D42*\u2010Gal4, and *elav\u2010*GS) of long hairpin RNAi lines used in this study, the RNAi\u2010*luciferase* (31603) was used as a control and a copy of UAS\u2010*Dcr2* was co\u2010expressed to boost the knockdown efficiency (Ni *et\u00a0al*., [2008](#acel12654-bib-0028){ref-type=\"ref\"}); for ubiquitous expression (*Da*\u2010Gal4) or expression in muscles (*Mef2*\u2010Gal4), no additional UAS\u2010Dcr2 was needed and the RNAi\u2010*mCherry* (35785) was used as a control. The UAS\u2010*Pink1* ^*L464P*^, UAS\u2010*Parkin* ^*T240R*^ *,* and UAS\u2010*Parkin* ^*R275W*^ flies are kind gifts from J. Chung, K.L. Lin, and C.H. Chen.\n\nFor simultaneous *in\u00a0vivo* imaging of axonal mitochondria and axon membrane integrity, the following stable fly strain containing multiple transgenes was generated by chromosomal recombination: *yw;dpr\u2010Gal4,UAS\u2010mitoGFP/CyO;UAS\u2010mCD8\u2010mCherry/TM6B.Tb*.\n\nFor simultaneous *in\u00a0vivo* imaging of axonal mitochondrial and axonal autophagosomes, the following stable fly line containing multiple transgenes was generated by chromosomal recombination: *yw; dpr\u2010Gal4, UAS\u2010mCherry\u2010Atg8a, UAS\u2010mitoGFP/Cyo*.\n\nStatistical analysis {#acel12654-sec-0012}\n--------------------\n\nUnless otherwise noted, statistical significance in this study is determined by unpaired, two\u2010tailed Student\\'s *t*\u2010test with unequal variance at \\**P\u00a0*\\<*\u00a0*0.05, \\*\\**P\u00a0*\\<*\u00a0*0.01, and \\*\\*\\**P\u00a0*\\<*\u00a0*0.001. Error bars represent standard error of mean (SEM).\n\nAdditional experimental procedures are available in Supporting Information.\n\nFunding {#acel12654-sec-0014}\n=======\n\nThis study is supported by the National Key R&D Program of China (No. 2016YFA0501902), the National Natural Science Foundation of China (No. 31471017 and No. 81671254), and the State High\u2010Tech Development Plan of China (the '863 Program', No. 2014A020526) to Y.F.\n\nAuthor contributions {#acel12654-sec-0015}\n====================\n\nX.C. planned and performed experiments, analyzed data, and wrote the manuscript. H.W. and Q.W. performed experiments. X.C., H.W., and Z.W. contributed critical reagents. S.Z. and Y.D. provided technical assistance. Y.F. designed and supervised research and wrote the manuscript.\n\nConflict of interest {#acel12654-sec-0016}\n====================\n\nThe authors declare no conflict of interest.\n\nSupporting information\n======================\n\n###### \n\n**Fig.\u00a0S1** Disruption of the PINK1\u2010Parkin pathway leads to overt abnormal muscle phenotypes.\n\n**Fig.\u00a0S2** The PINK1\u2010Parkin pathway is also dispensable in *Drosophila* motor neurons.\n\n**Fig.\u00a0S3** Upregulation of the PINK1\u2010Parkin pathway is detrimental to axons *in\u00a0vivo*.\n\n**Fig.\u00a0S4** The core ATG gene Atg12 and Atg17 are required for neuronal autophagy.\n\n###### \n\nClick here for additional data file.\n\nWe thank J. Chung, K.L. Lin, and C.H. Chen for providing the *Pink1* ^*L464P*^, *Parkin* ^*T240R*^ *,* and *Parkin* ^*R275W*^ flies, the BDSC and the TFC for providing fly strains, and A. Li, C. Liu, and J. Yuan for helpful discussion and critical reading of the manuscript.\n"}
+{"text": "In December 2013, my colleagues and I published the null results of two randomized trials investigating effects of brief parent-child music enrichment on preschoolers\\' cognitive skills (Mehr et al., [@B16]). Fully aware of the limitations of our studies, including, of course, that a failure to reject the null hypothesis does not imply evidence in support of the null hypothesis, we conservatively titled the paper \"Two randomized trials reveal no consistent evidence for nonmusical cognitive benefits of brief preschool music enrichment.\" In the discussion we wrote over 1000 words on why we might not have detected a positive effect, should one exist (pp. 9--10). Nonetheless, a media firestorm ensued, in which press reports claimed not only that our findings affirmed the null hypothesis, but also that they implied a broader conclusion: music lessons confer no cognitive benefits whatsoever (e.g., regardless of child age or training content, duration, or intensity). For instance, the *Times of London* reported, \"Academic benefits of music 'a myth'\" (Devlin, [@B5]); a *Boston Globe* headline read, \"Music doesn\\'t make you smarter, Harvard study finds\" (Johnson, [@B13]); and *TIME* reported \"Music can soothe the soul and speed along creativity, but it won\\'t, according to researchers from Harvard, boost intelligence\" (Sifferlin, [@B23]). These headlines appeared alongside 100 other reports from over 40 countries (partial listing: <https://goo.gl/pCwvqG>), despite efforts to clarify our findings via numerous interviews, a live Q&A on Reddit, and a New York Times op-ed (Mehr, [@B15a]).\n\nPerhaps we should not have been surprised. The idea that \"music makes you smarter\" is widely accepted by the general public (e.g., Mehr, [@B15]) and traces back to a sensationalist media interpretation of a *Nature* paper describing improved spatial task performance after listening to a Mozart sonata (Rauscher et al., [@B21]). This \"Mozart effect\" has been called a \"scientific legend\" (Bangerter and Heath, [@B1]) and was conclusively debunked (e.g., Chabris, [@B4]; Thompson et al., [@B25]), but not before it elicited a media frenzy that affected political policy: it was cited in a US House of Representatives debate on public funding of arts programs (Trescott, [@B26]) and by President Bill Clinton in a speech on arts education (Hershenson, [@B9]), and even prompted the governor of Georgia to sign a bill funding the distribution of classical music CDs in maternity wards (Bangerter and Heath, [@B1]).\n\nThis sort of sensationalism also occurs in coverage of other scientific fields. A year before our *PLoS* findings were published, Barron and Brown ([@B2]) described in *Nature* misleading media coverage of research on animal sexual behavior. For instance, a study of the evolution of cooperative breeding (Young et al., [@B27]) was described in the press using language of human sexuality (\"The love that daren\\'t squawk its name: when animals come out of the closet\"; Mooallem, [@B17]) despite the study\\'s actual topic of mate choice in the Laysan albatross. Taking Barron and Brown\\'s lead, I reviewed media coverage of recent scientific research in music cognition. I searched through roughly 15 years\\' worth of articles from the world\\'s fifty highest-circulation print and online publications (Pew Research Center, [@B22]). Here, I present some descriptive results of my review. A full empirical study of trends in science journalism is outside the scope of a brief commentary, but I hope that this article might prompt such research and also help to stop exaggerated media claims about music cognition research.\n\nI found reporting errors in two categories. First, in many papers of record, journalists fell prey to the *post hoc ergo propter hoc* fallacy, misinterpreting the findings of correlational studies as causal effects. For instance, Hanna-Pladdy and MacKay ([@B8]) compared elderly musicians to elderly non-musicians and found the musicians had stronger cognitive skills. The *Philadelphia Inquirer* reported that musical training was the cause: \"Those clarinet lessons helped you tune up for your later years: Getting to Carnegie hall isn\\'t the only reason to practice, practice, practice\" (Bauers, [@B3]). However, the study\\'s retrospective design precluded any causal inference, which the authors stated clearly: \"...the correlational design of our studies does not allow us to comment on whether musical participation causally enhanced cognition or whether other variables were responsible for the findings\" (p. 384). This measured interpretation fell on deaf ears, as other widely-read publications joined the *Inquirer* in a failure to distinguish correlation from causation: *FOX News, The Huffington Post*, and *The New York Times* reported the result in causal language (FOX News, [@B6]; Huffington Post, [@B12]; Klass, [@B14]).\n\nSecond, press reports made critical errors in interpreting psychometric outcome measures. For instance, improvement in word recall after music lessons (Ho et al., [@B10]) was reported in *The New York Times* under the headline \"More music yields more words,\" stating, \"If you want to improve the vocabulary of your children, sign them up for the orchestra...\" (O\\'Neil, [@B19]). This is dramatically inaccurate: a word recall test assesses memory, not vocabulary. Similarly, when Fujioka and colleagues reported (Fujioka et al., [@B7]) that musical training affected children\\'s neural response to sound, the *Daily Mail* reported \"Music lessons are an IQ booster for young minds\" (Hope, [@B11]). That study made no mention of IQ. And our *PLoS* findings were reported with \"...scientists are not so sure that \\[music\\] boosts IQ\" on *National Public Radio* (Neuman, [@B18]), despite our explicit decision not to use IQ as an outcome measure (a decision to which we devoted 300 words of explanation; Mehr et al., [@B16], p. 3).\n\nThat dry scientific titles are translated into catchy headlines is not necessarily worrisome; after all, science journalism can only thrive if the general public actually reads its journalistic product. However, these catchy headlines often include both error types I have described. For instance, \"Voxel-based morphometry reveals increased gray matter density in Broca\\'s area in male symphony orchestra musicians\" (Sluming et al., [@B24]) became \"Music improves brain power in some performers,\" with the subhead \"Mozart increases mental mass\" (Radford, [@B20]). The causal direction was unknown, and gray matter density is not \"brain power.\" Our *PLoS* paper\\'s title became \"Do, Re, Mi, Fa-get the piano lessons: Music may not make you smarter\" (Sifferlin, [@B23]). We studied neither piano lessons nor general intelligence.\n\nI have not attempted to estimate the overall frequency of media misrepresentation of music cognition research, nor have I attempted a systematic comparison of that frequency to that of other fields, and so I cannot comment on the prevalence or relative severity of the problems I describe above. But granting that misrepresentation occurs with nonzero frequency, and that it can occur in such prominent news outlets as those cited above, these descriptive findings raise a sticky question: Whose fault is all this? Are journalists sensationalizing research findings to garner page-views and sell papers, or are scientists exaggerating the importance of their own work?\n\nI speculate that the answer is \"both,\" and hope that this commentary might encourage music cognition scholars to do what we can to anticipate and avoid such issues when publishing new work. We should not take for granted public understanding of basic principles of scientific inference (e.g., correlation is not causation; effects reported are specific to the intervention that was tested) and in interviews and/or public discussion of our research, we should distinguish clearly between data-driven conclusions and idea-driven speculation. If and when our work is misrepresented, we must engage directly with journalists and with the public to correct the record, rather than throwing up our hands in frustration and keeping quiet. Music cognition research is in a particularly sensitive position when it comes to the press: our studies often involve three topics subject to intense public interest---children, brains, and music---and so we must make every effort to promote accurate and responsible public dissemination of our work.\n\nConflict of interest statement\n==============================\n\nThe author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nMy thanks to Sandra Trehub for her review; to Howard Gardner, Alia Martin, Adena Schachner, and Ellen Winner for their comments on the manuscript; and to Elizabeth Spelke for her support.\n\n[^1]: Edited by: Eduardo Coutinho, Imperial College London and University of Liverpool, UK\n\n[^2]: Reviewed by: Sandra E. Trehub, University of Toronto, Canada\n\n[^3]: This article was submitted to Cognition, a section of the journal Frontiers in Psychology\n"}
+{"text": "One well-documented method to control immunity and tolerance is through the regulation of nutrients in their immune microenvironment. Best described is the tryptophan deficiency mediated by the catabolic enzyme indoleamine 2,3-dioxygenase (IDO), which locally depletes tryptophan and liberates the immunoregulatory metabolites known as kynurenines. T cell activation is exquisitely sensitive to local tryptophan catabolism, and thus this enzyme exerts profound protective effects in allo-fetal rejection, autoimmunity, and inflammation. IDO can also be detected in tumors and draining LNs, and DC expression of IDO limits T cell responsiveness to antigen ([@bib17]; [@bib10]). In addition to IDO, some human cancers express tryptophan 2,3-dioxygenase (TDO), which also utilizes tryptophan as a substrate to produce kynurenines ([@bib22]). Furthermore, cysteine and arginine deficiency in tumors can inhibit T cell activation ([@bib25]; [@bib27]), suggesting that loss of any number of amino acids may serve as a common tumor escape mechanism.\n\nIn accordance with these findings, in a model of skin allograft tolerance, it was observed that multiple catabolic enzymes were up-regulated that consume a litany of essential amino acids. It was shown that these enzymatic activities could dampen T cell proliferation through nutrient deprivation ([@bib2]). One of the many enzymes that were up-regulated was Tph-1 (tryptophan hydroxylase-1), a synthase which utilizes tryptophan as a substrate to produce serotonin and melatonin ([@bib31]), and it was speculated that Tph-1 may precipitate the loss of tryptophan in the local microenvironment ([@bib32]). In the present study, we sought to determine whether Tph-1, the isoform expressed in the periphery ([@bib30]), could function as a regulator of immunity through the control of tryptophan metabolism and uncovered a novel and profound immunoregulatory function for this enzyme.\n\nRESULTS\n=======\n\nMast cells (MCs) express Tph-1\n------------------------------\n\nA study from our laboratory previously reported that MCs are critical in maintaining regulatory T cell--dependent skin allograft tolerance ([@bib13]). Concordant with these observations, gene array analyses performed by two different groups established that tolerant skin allografts display an MC signature with heightened signal for Tph-1 and other MC gene products ([@bib32]; [@bib13]). Pilot experiments showed that purified MCs from tolerant allografts expressed heightened levels of Tph-1 compared with MCs from syngeneic grafts (not depicted). It was also observed that peritoneal MCs from naive mice express \u223c1,000-fold higher expression for Tph-1 message than any other hematopoietic cell type examined by quantitative RT-PCR (qRT-PCR) and that Tph protein expression appears to be restricted to MCs ([Fig. 1](#fig1){ref-type=\"fig\"}). Heightened Tph-1 expression in MCs appeared unique, as the expression for other catabolic enzymes was absent or minimal ([Fig. 1 B](#fig1){ref-type=\"fig\"}).\n\n![**MCs have specific expression of Tph-1.** (A) Tph-1 expression (as determined by qRT-PCR) of the indicated cell types standardized to \u03b2-actin expression (mean \u00b1 SEM) is shown. For each cell type, two independent FACS-sorted samples were used. Each MC and macrophage sample was pooled from the peritoneal lavage of 20 mice. All other samples were pooled from the LN and spleen of two to four mice. (B) Peritoneal MC expression of the following genes was determined by qRT-PCR: arginase (Arg), dopa decarboxylase (Ddc), histidine decarboxylase (Hdc), IDO, inducible nitric oxide synthase (iNos), [l]{.smallcaps}-threonine dehydrogenase (Tdh), TDO, Tph-1, and Tph-2. Two independent samples were used to determine mean \u00b1 SEM shown. (A and B) ND indicates that a signal was not detected. (C) Intracellular staining for Tph-1 was performed for MCs, CD4^+^ T cells, CD8^+^ T cells, B cells, CD11c^hi^ DCs, and F4/80^+^ CD11b^+^ macrophages in WT (bold lines) and Tph-1^\u2212/\u2212^ (shaded areas) mice (*n* = 3--4 mice/group, and each cell type was tested in at least two independent experiments). (D) Naive LN sections were sectioned and stained for Tph-1 expression on MCs. CD117 staining is blue, Tph-1 staining is green, and overlap is white. Results are representative of three different mice/group. Bar, 10 \u00b5m.](JEM_20120408_Fig1){#fig1}\n\nBased on these findings and the recognition that Tph-1 can locally deplete tryptophan, the potential functional involvement of Tph-1 in mediating immune tolerance within the skin microenvironment was studied in Tph-1^\u2212/\u2212^ mice. Tph-1^\u2212/\u2212^ mice have been previously described as harboring no gross physical abnormalities except diminished cardiac function ([@bib3]) and enhanced liver microcirculation ([@bib11]). However, the immune compartments of Tph-1^\u2212/\u2212^ mice have not been well characterized. The genetic absence of Tph-1 appears to impact the steady-state levels of tryptophan as serum from Tph-1^\u2212/\u2212^ mice contains \u223c5 \u00b5g/ml more tryptophan in serum than that of WT controls (P = 0.0152; [Fig. 2 A](#fig2){ref-type=\"fig\"}). Although systemic levels of tryptophan are generally attributed to the activity of TDO, it appears that IDO can modulate the relative amounts of tryptophan under inflammatory conditions ([@bib26]). Therefore, it appears that constitutive loss of Tph-1 can also alter tryptophan levels in the periphery. The analysis of immune cell phenotypes on WT and Tph-1^\u2212/\u2212^ mice revealed that these mice were indistinguishable across all tissues examined (not depicted). Furthermore, the lack of Tph-1 does not appear to impact the frequency ([Fig. 2 B](#fig2){ref-type=\"fig\"}) or numbers (not depicted) of MCs, which have been described as the cells most abundantly expressing Tph-1 ([@bib28]; [@bib16]; [@bib4]).\n\n![**Tph-1^\u2212/\u2212^ mice have elevated serum tryptophan levels and a normal MC compartment.** (A) Serum tryptophan levels from in-house-bred naive WT and Tph-1^\u2212/\u2212^ mice (*n* = 13--14 mice/group from three independent experiments) were determined by ELISA. Mean \u00b1 SEM is shown. The p-value was determined by Student's *t* test. (B) Peritoneal lavage, lungs, intestines, spleen, and skin of Tph-1^\u2212/\u2212^ mice were analyzed by flow cytometry for the relative number of MCs and compared with WT mice. Data are pooled from two independent experiments (*n* = 7--8 mice/group), and mean is indicated.](JEM_20120408_Fig2){#fig2}\n\nDonor-specific transfusion (DST)/\u03b1-CD40L--treated mice reject allogeneic skin grafts with Tph-1 deficiency\n----------------------------------------------------------------------------------------------------------\n\nThe potential role of Tph-1 deficiency in immune tolerance was studied in a model of skin allograft tolerance. Here mice were treated with a tolerance-inducing regimen of DST in combination with \u03b1-CD40L on days \u22127, \u22125, and \u22123 and then received an allogeneic CB6F~1~ (C57BL/6 \u00d7 BALB/c) skin graft on day 0 ([@bib23]). In the first experiments, the specific and irreversible inhibitor parachlorophenylalanine (PCPA), which has been reported to effectively block the effects of Tph with multiple treatments ([@bib3]), was used. Approximately 90% of the PCPA-treated mice receiving DST/\u03b1-CD40L rejected their skin allografts within 30 d. In contrast, 100% of DST/\u03b1-CD40L--treated WT mice retained their skin allografts for \\>50 d ([Fig. 3 A](#fig3){ref-type=\"fig\"}). The role of Tph-1 in allograft tolerance was then studied in Tph-1^\u2212/\u2212^ mice ([Fig. 3, A and B](#fig3){ref-type=\"fig\"}). It was shown that 80% of the DST/\u03b1-CD40L--treated Tph-1^\u2212/\u2212^ mice rejected their skin grafts within 30 d as compared with DST/\u03b1-CD40L--treated WT mice (P \\< 0.0001 between these two groups by log-rank test). Tph-1^\u2212/\u2212^ mice grafted with syngeneic grafts retained those grafts as did WT controls, establishing that Tph-1 is not involved in wound healing. Furthermore, allogeneic grafts placed on unmanipulated WT or Tph-1^\u2212/\u2212^ mice mounted similar rejection responses with regard to the frequency and rates of graft rejection ([Fig. 3 B](#fig3){ref-type=\"fig\"}). It was also noted that the majority of MCs in Tph-1^\u2212/\u2212^ mice in donor skin allografts were negative for Tph-1, indicating that they were host derived ([Fig. 3 C](#fig3){ref-type=\"fig\"}). The loss of tolerance in this model was not caused by the absence of peripheral serotonin synthesis in the Tph-1^\u2212/\u2212^ mice as adding back 5-hydroxytryptophan (5-HTP), which bypasses the Tph-1 deficiency to restore serotonin levels, did not allow for long-lived tolerance in Tph-1^\u2212/\u2212^ mice ([Fig. 3 B](#fig3){ref-type=\"fig\"}). Because MCs store serotonin in their granules, reconstitution of serotonin biosynthesis by 5-HTP administration in Tph-1^\u2212/\u2212^ mice was confirmed by histological analysis of cytospins of peritoneal lavage MCs ([Fig. 3 D](#fig3){ref-type=\"fig\"}). It is therefore concluded that Tph-1 is required for the maintenance of peripheral tolerance independently of serotonin.\n\n![**Tph-1 is essential for the maintenance of tolerance.** (A and B) WT and Tph-1^\u2212/\u2212^ mice were grafted with syngeneic C57BL/6 or allogeneic CB6F~1~ skin. DST- and \u03b1-CD40L--treated mice received an allogeneic skin graft. In the indicated groups, PCPA and 5-HTP were given every other day for the duration of the experiment starting at day \u22128 for DST/\u03b1-CD40L--treated mice and day \u22121 for syngeneic and allogeneic mice. The number of mice/group is indicated, and data were pooled from two or more independent experiments. (C) Histology staining on skin allografts in DST/\u03b1-CD40L--treated mice of Tph-1 or WT mice at day 10 after grafting. Tph-1 is shown in red, and cKit (CD117) is shown in blue. In the overlay, white arrows indicated double-positive cells, and blue arrows indicate cells positive for cKit but not for Tph-1. (D) Reconstitution of serotonin in the granules of MCs was confirmed by immunofluorescent staining on purified peritoneal lavage MCs. Serotonin staining is in red, cKit staining in blue, and nuclear staining in yellow. The white arrows indicate MCs positive for serotonin, whereas the red arrows indicate MCs negative for serotonin. Bars, 20 \u00b5m. (E) Graft draining LNs from DST/\u03b1-CD40L--treated mice were depleted of antigen-presenting cells and restimulated with irradiated WT of F1 antigen-presenting cells for 18 h for IFN-\u03b3 and for 36 h for IL-17A before ELISPOT development. Data show *n* = 7--8 mice/group, and mean \u00b1 SEM is representative of results from two independent experiments. P-values were determined by Student's *t* test. (F) Mast-deficient (W^sh^) mice were locally (skin) reconstituted with 6--8-wk-old MCs from WT or Tph-1^\u2212/\u2212^ mice 8 wk before DST/\u03b1-CD40L treatment. After receiving a skin graft, mice were monitored for rejection for another 50 d. Data shown were combined from two independent experiments.](JEM_20120408_Fig3){#fig3}\n\nTo determine whether rejection in the DST/\u03b1-CD40L--treated Tph-1^\u2212/\u2212^ mice was caused by the emergence of an allogeneic T cell response, T cells from draining LNs 14 d after the initial grafting were recalled with F1 antigen-presenting cells to assess their cytokine production ([Fig. 3 E](#fig3){ref-type=\"fig\"}). An increase in the allogeneic-specific IFN-\u03b3 response (P = 0.0334) as well as IL-17A response (P = 0.0084) was observed in the DST/\u03b1-CD40L--treated Tph-1^\u2212/\u2212^ mice compared with similarly prepared WT mice.\n\nBased on the finding that the highest signal of Tph-1 message and protein expression occurs in MCs, it was envisioned that MC-derived Tph-1 is responsible for the phenotype seen in the deficient mice. W^sh^ mice harbor a MC deficiency caused by a c-Kit mutation and can be readily reconstituted with genetically deficient MCs to test the role of MC-derived genes in immune responses ([@bib7]). As such, W^sh^ mice were reconstituted with WT or Tph-1^\u2212/\u2212^ BM-derived MCs (BMMCs). As previously reported, DST/\u03b1-CD40L--treated W^sh^ mice reconstituted with WT BMMCs show a gradual erosion of tolerance leading to rejection of \u223c50% of the allografts between days 30 and 60 after transplantation ([@bib13]). However, when DST/\u03b1-CD40L--treated W^sh^ mice were reconstituted with Tph-1^\u2212/\u2212^ BMMCs, it was observed that all mice rejected their allograft within 30 d with P = 0.0002 in comparison with WT BMMC-reconstituted mice by log-rank test ([Fig. 3 F](#fig3){ref-type=\"fig\"}). Therefore, it appears that MC-derived Tph-1, and not Tph-1 produced by other cells, is necessary for long-term graft tolerance.\n\nTph-1 and immune suppression to tumor\n-------------------------------------\n\nAs MC expression of Tph-1 is critical for allograft tolerance and MCs have been implicated in suppressing antitumor immunity ([@bib14]), experiments were designed to determine whether MC-derived Tph-1 mediated immune tolerance to an intradermal (i.d.) skin tumor model. MB49 is a bladder cell carcinoma that expresses male minor histocompatibility antigen and readily grows and kills male mice because they are centrally tolerant to H-Y ([@bib29]; [@bib8]). In contrast, female WT mice have delayed tumor growth kinetics in comparison with WT males yet nonetheless ultimately succumb to tumor. Strikingly, female Tph-1^\u2212/\u2212^ mice have reduced tumor growth kinetics, and \u223c50% completely reject MB49. It was also confirmed that the effect of Tph-1 deficiency was independent of serotonin biosynthesis by the lack of effect of administration of 5-HTP ([Fig. 4 A](#fig4){ref-type=\"fig\"}). 11 d after tumor inoculation, tumor-specific recall responses from T cells from the draining LNs were measured ([Fig. 4 B](#fig4){ref-type=\"fig\"}). Female Tph-1^\u2212/\u2212^ mice had significantly higher numbers of IL-17A--specific spots (P \\< 0.01) in comparison with all other groups, and IFN-\u03b3--specific spots were elevated in female Tph-1^\u2212/\u2212^ mice although it did not reach statistical significance. Through reconstitution of W^sh^ mice with either BMMCs from either WT or Tph-1^\u2212/\u2212^ mice, it was confirmed that Tph-1^\u2212/\u2212^ BMMCs conferred protection to the same extent as female Tph-1^\u2212/\u2212^ mice with P = 0.0052 between WT and Tph-1^\u2212/\u2212^ BMMC-reconstituted mice ([Fig. 4 C](#fig4){ref-type=\"fig\"}). The data establish that MCs maintain a suppressive antitumor microenvironment and that Tph-1 is a major mediator within this context.\n\n![**Tph-1 is essential for the generation of protective T cell--mediated antitumor immunity.** (A) MB49 was inoculated, and tumor growth was monitored over time in all the groups. One group received 5-HTP every other day from the day before the start of experiment. Data show mean \u00b1 SEM of *n* = 6--8 mice/group and are representative of results from three independent experiments. P-values listed are for the group they are shown next to in comparison with WT female mice and were determined by two-way ANOVA. (B) On day 11 after tumor inoculation, whole tumor draining LNs were isolated, plated, and restimulated with irradiated tumor to determine IFN-\u03b3 ELISPOTS after 18 h and IL-17A ELISPOTS after 36 h, restimulated for 18 and 36 h, respectively. Data show mean \u00b1 SEM for *n* = 10--16 mice/group and were pooled from three independent experiments. The p-value shown was determined for the indicated groups by Tukey posttest after one-way ANOVA analysis. (C) MB49 was inoculated systemically through the tail vein in age-matched female WT and Tph-1^\u2212/\u2212^ mice as well as female W^sh^ mice systemically reconstituted with BMMCs from WT and Tph-1^\u2212/\u2212^ mice. The graph is compiled from three independent experiments with total numbers of mice indicated in the figure. The p-value was determined by log-rank comparison between the two BMMC-reconstituted groups.](JEM_20120408_Fig4){#fig4}\n\nTph-1 deficiency exacerbates experimental autoimmune encephalomyelitis (EAE)\n----------------------------------------------------------------------------\n\nThe question arose as to whether Tph-1 only enhances immunity under immune-tolerant or -suppressive conditions, or can it function in tempering inflammation in general? To address this question, the impact of Tph-1 deficiency on a central nervous system (CNS) inflammatory disease was studied. EAE is a model of CNS inflammation in which encephalitogenic Th17 lineage T cells infiltrate the CNS, mediate damage, and cause ascending paralysis. Using suboptimal disease conditions, it was found that Tph-1^\u2212/\u2212^ mice develop earlier and more severe disease than controls (P = 0.0064; [Fig. 5 A](#fig5){ref-type=\"fig\"}). As seen in all other models tested, restoration of serotonin levels by administration of 5-HTP had no effect on the immunological impact of Tph-1 deficiency. In this specific case, it was shown that serotonin levels in serum were restored to the same levels of controls after just one injection with 5-HTP ([Fig. 5 B](#fig5){ref-type=\"fig\"}). Furthermore, long-term treatment of 5-HTP restored serotonin to a level greater than WT controls (P \\< 0.001; [Fig. 5 C](#fig5){ref-type=\"fig\"}). The finding that serotonin is not involved is also supported by the observation that serotonin transporter--deficient mice have attenuated EAE ([@bib9]).\n\n![**Tph-1^\u2212/\u2212^ mice have exacerbated EAE independent of their serotonin level.** (A) WT, Tph-1^\u2212/\u2212^, and Tph-1^\u2212/\u2212^ + 5-HTP--treated (starting on day \u22121 and continuing every other day for the duration of the experiment) mice were monitored and scored. Data are pooled from three independent experiments and show mean \u00b1 SEM for *n* = 13 mice/group. P-values indicate significance in comparison with WT mice and were calculated by Mann--Whitney *U* test. (B) Serum serotonin levels of WT, Tph-1^\u2212/\u2212^, and Tph-1^\u2212/\u2212^ + 5-HTP--treated (one shot 2 d prior) mice on day 1 after EAE induction were determined by ELISA (*n* = 11--12 mice/group pooled from three independent experiments). (C) Serum serotonin levels of WT, Tph-1^\u2212/\u2212^, and Tph-1^\u2212/\u2212^ + 5-HTP--treated (last shot given 2 d before sample collection) mice on day 21 after EAE induction were determined by ELISA (*n* = 13 mice/group pooled from three independent experiments). The p-value was determined by one-way ANOVA analysis and indicates difference from WT mice. (B and C) Mean \u00b1 SEM is shown.](JEM_20120408_Fig5){#fig5}\n\nFurther characterization of immune responses during EAE found that Tph-1^\u2212/\u2212^ mice have an enhanced frequency of encephalitogenic T cells. Analysis of T cell responses during EAE revealed that on day 7 after immunization (when neither group showed disease), Tph-1^\u2212/\u2212^ mice had greater numbers of CD4^+^ T cells expressing CCR6 (P = 0.0484), one of the adhesion molecules necessary to be on the first wave of T cells that infiltrate the CNS ([Fig. 6 A](#fig6){ref-type=\"fig\"}; [@bib12]; [@bib24]). Tph-1^\u2212/\u2212^ mice also had elevated recall IL-17A production by CD4^+^ T cells (P = 0.0002). At late stages of disease, Tph-1^\u2212/\u2212^ mice also had elevated numbers of CD4^+^ and CD8^+^ T cells that were infiltrating the CNS, which have high IL-17A and IFN-\u03b3 recall responses ([Fig. 6 B](#fig6){ref-type=\"fig\"}). To determine whether this effect is just caused by differences in T cell priming, myelin-specific Th17 cells were adoptively transferred into WT and Tph-1^\u2212/\u2212^ mice. Data show that these T cells could effectively initiate disease in Tph-1^\u2212/\u2212^ mice but not controls (P \\< 0.0001), showing that Tph-1^\u2212/\u2212^ mice are better able to sustain and facilitate encephalitogenic T cell--mediated pathogenesis ([Fig. 6 C](#fig6){ref-type=\"fig\"}). Although hampered by differences in EAE severity caused by the increased age of the mice, the phenotype of W^sh^ mice reconstituted with Tph-1^\u2212/\u2212^ BMMCs looked consistent with that of Tph-1^\u2212/\u2212^ mice, suggesting MC-derived Tph-1 is responsible for this effect ([Fig. 6 D](#fig6){ref-type=\"fig\"}).\n\n![**Tph-1^\u2212/\u2212^ mice have increased encephalitogenic T cell responses.** (A) Day 7 after inoculation for EAE, mice were taken down and draining LNs were stained for the indicated parameters (after restimulation for cytokine stains). Data are pooled from three independent experiments (*n* = 11 mice/group), and mean is indicated. (B) Day 21 after inoculation, lymphocytes from the CNS were isolated and stained for the indicated parameters (after restimulation for cytokine stains). Data are pooled from three independent experiments (*n* = 15--18 mice/group), and mean is indicated. (A and B) P-values were determined by Student's *t* test. (C) Mice were injected with activated MOG~35--55~-specific T cells, and disease severity was monitored over time. Data are pooled from two independent experiments and show mean \u00b1 SEM for *n* = 14--16 mice/group. The p-value was determined by Mann--Whitney *U* test. (D) W^sh^ mice were given WT or Tph-1^\u2212/\u2212^ BMMCs systemically and given 8 wk to reconstitute. These mice and age-matched controls were immunized for active EAE and monitored for disease severity over time. Data show mean \u00b1 SEM for *n* = 10--15 mice/group pooled from two independent experiments.](JEM_20120408_Fig6){#fig6}\n\nTph-1^\u2212/\u2212^ mice have enhanced signaling of mammalian target of rapamycin (mTOR)\n-------------------------------------------------------------------------------\n\nThe mTOR pathway can promote CD4 T cell differentiation to Th1 and Th17 cells ([@bib6]), and its activity is inhibited in vitro with loss of essential amino acids ([@bib2]). Therefore, it presented itself as a potential target for mediating the differences observed in WT and Tph-1^\u2212/\u2212^ mice. The phosphorylation of S6 riboprotein, one of the targets of mTOR, in CD4^+^ T cells was measured in the draining LNs 4 d after EAE immunization, and we found that it was elevated in Tph-1^\u2212/\u2212^ mice (P = 0.04; [Fig. 7](#fig7){ref-type=\"fig\"}). This suggests that comparatively there is a reduction of mTOR activity in WT mice that is consistent with there potentially being a tryptophan-deficient microenvironment imposed by Tph-1.\n\n![**Tph-1^\u2212/\u2212^ mice have enhanced mTOR activity.** (A) Example staining of P-S6 and its isotype control antibody is shown. (B) Day 4 after inoculation for EAE, mice were taken down, draining LNs were harvested, and CD4^+^ T cells were stained for P-S6 expression (and isotype control). Data are pooled from three independent experiments with *n* = 15 mice/group. Mean is indicated, and the p-value was determined by Student's *t* test.](JEM_20120408_Fig7){#fig7}\n\nDISCUSSION\n==========\n\nThe findings presented herein describe Tph-1 as an important and novel MC-derived regulator of immunological tolerance. This single molecule exerts striking alterations in immunological outcomes in models of transplantation tolerance, tumor growth, and autoimmunity. Tph-1 metabolizes tryptophan for the purpose of producing serotonin. However, the experiments presented here clearly show that the major immunological impact of this pathway is not through the regulation of serotonin levels and therefore must be caused by its ability to exhaust tryptophan as suggested by the mTOR experiments.\n\nThere are several catabolic enzymes that have been suggested to contribute to the establishment of immune tolerance. For example, during skin allograft tolerance, DCs activated by regulatory T cells can express enzymes to consume 9 of the 10 essential amino acids and cause a reduction in mTOR signaling ([@bib2]). Extensive work on IDO, one of the tryptophan-catabolizing enzymes, also shows that it can limit immune responses through the induction of the GCN2 stress response, which promotes anergy, as well as the production of tryptophan metabolites that suppress inflammation ([@bib17]). There are also some indications that different catabolic enzymes may counter-regulate the activity of each other. For example, the production of nitric oxide by inducible nitric oxide synthase can prevent IDO activity ([@bib10]). However, in the case of Tph-1 expression by MCs, this does not appear to be occurring because MCs have little to no messenger RNA expression for other catabolic enzymes, so competition for substrate is not likely contributing to the immunosuppressive activity of Tph-1.\n\nIn the case of Tph-1, the question arises as to which cell type is the target of the tryptophan deficiency in the tolerant microenvironment. Although we observe enhancement of mTOR activity in CD4 T cells with Tph-1 deficiency during EAE, it is not yet clear whether this is the case in our transplant and tumor models. In addition, there could be additional targets of Tph-1 activity. For example, we have recently reported that within the tolerant allograft, graft-derived DCs mediate regional allo-specific unresponsiveness. Upon analysis of gene expression of these DCs, it appears that they have experienced a nutrient-stressed environment ([@bib5]). We would contend that this is caused by Tph-1 activities on DCs from the increased MCs that infiltrate the tolerant allograft. It would also be of interest to determine the factors that up-regulate Tph-1 expression in MCs as well as additional cell types to potentially mediate immune tolerance.\n\nThe experiments presented provide a new prospective on this immunologically important enzyme. Its end product serotonin can impact inflammation, particularly in the gut ([@bib21]; [@bib18]). It has also been observed in Tph-1^\u2212/\u2212^ mice that there is an enhanced ability to clear lymphocytic choriomeningitis virus ([@bib11]) and reject the MC-38 colon cancer line ([@bib19]). However, in these cases, this is dependent on the loss of serotonin in these mice. Therefore, Tph-1 likely regulates immunity by regulating serotonin levels or by exhausting tryptophan depending on the nature of the immune response invoked. Certainly, the fact that Tph-1 can so profoundly impact tolerance and inflammation provides compelling incentive to consider Tph-1 as a novel target in immune intervention.\n\nMATERIALS AND METHODS\n=====================\n\n### Mice.\n\nTph-1^\u2212/\u2212^ mice fully backcrossed to a C57BL/6 background were provided by N. Horseman (University of Cincinnati, Cincinnati, OH) and maintained at the Dartmouth College animal facility. Male and female 6--8-wk-old C57BL/6 mice were bred in-house for measurement of tryptophan in serum, all EAE experiments, and all ELISPOT experiments. In all other experiments, C57BL/6 mice were purchased from the National Cancer Institute. 2D2 TCR transgenic mice were provided by V. Kuchroo (Harvard University, Cambridge, MA) and maintained in-house. C57BL/6 Kit^W-sh^ (W^sh^) mice were purchased from the Jackson Laboratory. CB6F~1~ (C57BL/6 \u00d7 BALB/c hybrid) were purchased from the National Cancer Institute. Experiments were performed under protocols approval by the Institutional Animal Care and Use Committee of Dartmouth College, and mice were maintained in a specific pathogen--free facility at Dartmouth Medical School.\n\n### MC reconstitution.\n\nBMMCs for MC reconstitution were generated by culturing BM cells with 20 ng/ml IL-3 (PeproTech) and 50 ng/ml stem cell factor (SCF; PeproTech) for 5--8 wk as shown previously ([@bib13]). Purity was assessed by anti-CD117 and anti-FC\u03b5RI staining on BMMC cultures ([@bib13]). A total of 4--5 \u00d7 10^6^ BMMCs were then injected i.d., i.v., and i.p. into W^sh^ recipients. Mice were allowed to rest for 8--12 wk before use in experiments. At the end of the experiment, reconstitution was confirmed by flow cytometry or histology.\n\n### Skin grafting.\n\nSkin grafting was performed as previously described ([@bib15]). In brief, 1-cm^2^ full-thickness tail skins were collected from CB6F~1~ allogeneic donor or C57BL/6 syngenic donor mice. Skins were then stored on PBS-soaked gauze and were on the following day applied to the dorsal surface of age-matched WT or Tph-1^\u2212/\u2212^ host mice. Indicated groups were treated 7 d before grafting, via the i.v. injection of 3 \u00d7 10^7^ DST in conjunction with an injection of 250 \u00b5g anti-CD40L (clone MR1), followed by further injections on days \u22125 and \u22123 before graft. This regimen results in long-term tolerization of the mice to alloantigen.\n\n### Cell culture, tumor challenge, and vaccination.\n\nMurine bladder carcinoma cell line MB49 was maintained in complete medium (RPMI 1640 containing 10% fetal calf serum, 100 U/ml penicillin, 100 \u00b5g/ml streptomycin, 2 mM glutamine, and 50 \u00b5M 2-mercaptoethanol). Mice were injected with 2.5 \u00d7 10^5^ MB49 tumor cells i.d. on the right flank, and tumor diameters were measured with a caliper thrice weekly. Alternatively, mice were challenged with 2.5 \u00d7 10^5^ MB49 tumor cells i.v. in the tail vein and were monitored for survival.\n\n### EAE immunization and clinical evaluation.\n\nAge-matched WT and Tph-1^\u2212/\u2212^ mice were immunized subcutaneously with 125 \u00b5g MOG~35--55~ peptide (Peptides International) emulsified in IFA (Sigma-Aldrich) supplemented with 0.5 mg/mouse *Mycobacterium tuberculosis H37* (Difco Laboratories) on day 0 and an i.p. injection of 200 ng pertussis toxin (List Biologicals) on days 0 and 2. For adoptive transfer EAE, donor 2D2 transgenic T cells received a standard immunization as previously described ([@bib20]), and lymphocytes were isolated out of mice 10 d later. Cells were stimulated with 20 \u00b5g/ml MOG~35--55~, 10 \u00b5g/ml anti--IFN-\u03b3 (BioXCell), 20 ng/ml IL-23 (BD), 10 ng/ml IL-6 (PeproTech), and 10 ng/ml IL-1\u03b2 (PeproTech) for 4 d before reisolation of live T cells (\u223c80% positive for IL-17A). Recipient mice were injected with 1.0 \u00d7 10^6^ T cells i.v. Mice were scored as previously described ([@bib1]).\n\n### Antibodies and reagents.\n\nMouse monoclonal antibodies to CD8 (53-6.7) and CD4 (GK1.5) were purchased from eBioscience. Mouse monoclonal antibodies to CD4 (RMA4.5), CD45 (30-F11), IL-17A (TC11-18H10.1), IFN-\u03b3 (XMG1.2), CCR6 (29-2L17), CD117 (ACK2), FC\u03b5R1 (MAR-1), CD11b (M1/70), CD11c (N418), F4/80 (BM8), and CD19 (6D5) were purchased from BioLegend. Tph-1 antibody was purchased from Santa Cruz Biotechnology, Inc. Secondary F(ab\u2032)2 anti--rabbit IgG antibody was purchased from eBioscience. P-S6 antibody and its isotype control antibody were purchased from Cell Signaling Technology.\n\nELISPOT antibodies for IFN-\u03b3 were purchased from Mabtech and those for IL-17A were purchased from BioLegend. Both were developed using the AEC substrate kit from BD. ELISA kits for serotonin (Enzo Life Sciences) and tryptophan (Rocky Mountain Diagnostics) were used according to the manufacturers' directions. 5-HTP (Sigma-Aldrich) and PCPA (Sigma-Aldrich) were given i.p. at 5.5 mg/mouse every other day in all indicated experiments starting the day before any other treatment.\n\n### Histology.\n\nSections were cut and stained as previously described ([@bib13]). In brief, tissues were fixed in OCT, sectioned onto slides, fixed with methanol and acetone, stained with antibody in 10% serum, washed extensively, and mounted with Prolong Gold (Molecular Probes) according to the manufacturer's directions. Unstained and single-stain sections were performed for each tissue. In addition, the same staining cocktails were used on tissue from Tph-1^\u2212/\u2212^ mice to serve as a negative control for Tph-1 and serotonin stains. Images were taken on an LSM 510 confocal microscope (Carl Zeiss) and analyzed using LSM 5 Image Browser (Carl Zeiss).\n\n### Flow cytometry.\n\nSingle cell suspensions were incubated with antibodies conjugated with FITC, PE, PerCP, APC, Alexa Fluor 647, and/or Alexa Fluor 700. Intracellular staining and restimulation for cytokine staining was performed as previously described ([@bib20]). To stain for Tph-1, cells were surface stained, washed, fixed with Fixation/Permeabilization Buffer (BD), washed, permeabilized with Perm/Wash (BD), stained with Tph-1 antibody, washed, stained with secondary antibody, washed, and resuspended for analysis. For all staining steps, 10% normal rat serum was included. P-S6 staining was performed as follows: cells were blocked in 10% serum, surface stained, washed, fixed with 4% paraformaldehyde, washed, fixed and permeabilized with methanol, washed, blocked with 10% serum, intracellularly stained, washed, and resuspended for analysis. Five-color analyses were performed on a modified FACScan (BD) running CellQuest software (BD) and Rainbow software (Cytek).\n\n### RNA preparation and qRT-PCR.\n\nAll cell populations were sorted on a FACS Aria (BD). For MC samples, the cell number acquired was insufficient for postsort analysis, so the samples were assessed for expression of several MC markers and products to confirm specificity of cells. RNA was prepared according to the manufacturer's directions using an RNeasy Mini kit (QIAGEN). MC samples were further amplified and converted to cDNA using QuantiTect Whole Transcriptome kit (QIAGEN). All other samples were transcribed to cDNA using the iScript cDNA synthesis kit (Bio-Rad Laboratories). qRT-PCR was performed as previously described ([@bib1]). All samples were standardized to expression of \u03b2-actin. Primers used are as follows: \u03b2-actin (5\u2032-CCACACCCGCCAGTTCG-3\u2032 and 5\u2032-TCTGGGCCTCGTCACCCACAT-3\u2032), Tph-1 (5\u2032-GAAGACAACATCCCGCAACT-3\u2032 and 5\u2032-GTTCAGCCAAGAGAGGAACG-3\u2032), arginase (5\u2032-CAGAAGAATGGAAGAGTCAG-3\u2032 and 5\u2032-CAGATATGCAGGGAGTCACC-3\u2032), dopa decarboxylase (5\u2032-AGGGCAGAGAAAGAATGAAAGCA-3\u2032 and 5\u2032-GGAGTGGTAGTTATTTTTCTCTTTCCA-3\u2032), histidine decarboxylase (5\u2032-GATCAGATTTCTACCTGTGG-3\u2032 and 5\u2032-GTGTACCATCATCCACTTGG-3\u2032), IDO (5\u2032-TGGCAAACTGGAAGAAAAAG-3\u2032 and 5\u2032-ATTGCTTTCAGGTCTTGACG-3\u2032), inducible nitric oxide synthase (5\u2032-ACCCCTGTGTTCCACCAGGAGATGTT-3\u2032 and 5\u2032-TGAAGCCATGACCTTTCGCATTAGCA-3\u2032), [l]{.smallcaps}-threonine dehydrogenase (5\u2032-AAGCACGCGCCTGACTTC-3\u2032 and 5\u2032-CCGAGCATTGCTGTCATCTAGA-3\u2032), TDO (5\u2032-TGGGAACTAGATTCTGTTCG-3\u2032 and 5\u2032-TCGCTGCTGAAGTAAGAGCT-3\u2032), and Tph-2 (5\u2032-CAGGAGAGGGTTGTCCTTGG-3\u2032 and 5\u2032-TTTGCCGCTTTTCTTGTCCT-3\u2032).\n\n### Statistical analysis.\n\nData graphs were made using Prism software (GraphPad Software) and expressed as the mean \u00b1 SEM. Differences for graphs with one grouping variables were analyzed by Student's *t* test (two groups) or one-way analysis of variance (ANOVA) and Tukey analysis (three or more groups). Log-rank tests were used to compare skin graft survival and survival of mice inoculated with MB49. For the study of MB49 growth kinetics, two-way ANOVA was used to assess significance. In EAE time course experiments, statistical relevance was determined using Mann--Whitney *U* Test.\n\nThis work was supported by grants from the National Institutes of Health (A1084089, CA123079, and A1048667) and by the Medical Research Council Centre for Transplantation and Biomedical Research Center at King's College London. E.C. Nowak is supported by a postdoctoral fellowship from the National Multiple Sclerosis Society.\n\nThe authors have no conflicting financial interests.\n\nAbbreviations used:5-HTP5-hydroxytryptophanANOVAanalysis of varianceBMMCBM-derived MCCNScentral nervous systemDSTdonor-specific transfusionEAEexperimental autoimmune encephalomyelitisi.d.intradermal(ly)IDOindoleamine 2,3-dioxygenaseMCmast cellmTORmammalian target of rapamycinPCPAparachlorophenylalanineqRT-PCRquantitative RT-PCRTDOtryptophan 2,3-dioxygenase\n\n[^1]: E.C. Nowak, V.C. de Vries, and A. Wasiuk contributed equally to this paper.\n"}
+{"text": "The growth of human gastric cancer cells involves a variety of growth factors, gut hormones, and cytokines ([Tahara *et al*, 1993](#bib38){ref-type=\"other\"}). In particular, the epidermal growth factor receptor (EGF-R) pathway appears to play a crucial role in gastric cancer progression. A large percentage of gastric cancer cell lines express EGF-R ([Yokozaki, 2000](#bib50){ref-type=\"other\"}), and gastric cancer cells grow in response to EGF/transforming growth factor-*\u03b1* (TGF-*\u03b1*) activation of EGF-R in an autocrine loop ([Yoshida *et al*, 1990](#bib51){ref-type=\"other\"}; [Piontek *et al*, 1993](#bib35){ref-type=\"other\"}). Expression of EGF and its receptor has been found to correlate with prognosis in patients with gastric cancer ([Yasui *et al*, 1988](#bib49){ref-type=\"other\"}; [Jonjic *et al*, 1997](#bib14){ref-type=\"other\"}).\n\nTumour angiogenesis is essential for the growth and metastasis of solid tumours, and the process of angiogenesis is mediated by numerous stimulatory and inhibitory factors ([Folkman, 1995](#bib7){ref-type=\"other\"}). One such factor is vascular endothelial growth factor (VEGF), a potent mitogenic and chemotactic factor for endothelial cells (ECs) *in vitro* and an angiogenic factor *in vivo* ([Leung *et al*, 1989](#bib28){ref-type=\"other\"}; [Kondo *et al*, 2000](#bib25){ref-type=\"other\"}). The expression of VEGF has been correlated with tumour progression and poor clinical outcome in various cancer systems including gastric cancer ([Maeda *et al*, 1996](#bib29){ref-type=\"other\"}; [Takahashi *et al*, 1996](#bib39){ref-type=\"other\"}; [Kido *et al*, 2001](#bib21){ref-type=\"other\"}). VEGF is expressed as four isoforms derived from alternate splicing of the mRNA ([Tischer *et al*, 1991](#bib43){ref-type=\"other\"}). The smaller isoforms, VEGF-121 and VEGF-165, are secreted with VEGF-165 being the predominant isoform in most tumours. The classic receptors for VEGF possess tyrosine kinase activity and are expressed primarily on ECs. The current nomenclature for the VEGF receptors lists three receptors: VEGFR-1 (flt-1), VEGFR-2 (kdr/flk-1), and VEGFR-3 (flt-4) ([Fournier *et al*, 1997](#bib8){ref-type=\"other\"}). VEGF-induced mitogenesis and angiogenesis are mediated largely by VEGFR-2 ([Veikkola *et al*, 2000](#bib44){ref-type=\"other\"}).\n\nNeuropilin-1 (NRP-1) was first described as a semaphorin receptor important for the guidance of developing neurons ([He and Tessier-Lavigne, 1997](#bib11){ref-type=\"other\"}; [Kolodkin *et al*, 1997](#bib24){ref-type=\"other\"}). Transgenic overexpression or knockout of the *NRP-1* gene results in lethal abnormalities in the cardiovascular system, suggesting that NRP-1 plays a role in vasculogenesis and possibly angiogenesis ([Kitsukawa *et al*, 1995](#bib23){ref-type=\"other\"}; [Kawasaki *et al*, 1999](#bib19){ref-type=\"other\"}). More recently, NRP-1 has been found to be expressed on ECs, and coexpression of NRP-1 and VEGFR-2 on ECs enhances the biological activity of VEGFR-2 in response to the VEGF-165 isoform ([Soker *et al*, 1998](#bib37){ref-type=\"other\"}; [Whitaker *et al*, 2001](#bib45){ref-type=\"other\"}). These findings suggest that NRP-1 acts as a coreceptor for VEGFR-2 in ECs and functions in VEGF-mediated angiogenesis and vasculogenesis. NRP-1 is also expressed by several types of tumour cells, such as breast and prostate cancers ([Soker *et al*, 1998](#bib37){ref-type=\"other\"}), and overexpression of NRP-1 in prostate carcinoma cells has been shown to enhance tumour angiogenesis and growth ([Miao *et al*, 2000](#bib31){ref-type=\"other\"}).\n\nExperimental evidence suggests a role for the EGF ligand--receptor system in the induction of angiogenesis. In fact, it has been shown that EGF and TGF-*\u03b1* can upregulate the production of VEGF, currently regarded as the major proangiogenic factor for most types of human cancer ([Goldman *et al*, 1993](#bib10){ref-type=\"other\"}; [Kerbel *et al*, 1998](#bib20){ref-type=\"other\"}). A recent study has shown that EGF induces NRP-1 mRNA expression in astrocytoma ([Ding *et al*, 2000](#bib6){ref-type=\"other\"}), but the expression and regulation of NRP-1 in gastric cancer has not previously been investigated. We hypothesized that EGF, as a regulator of other angiogenic factors such as VEGF in other tumour types, regulates both VEGF and NRP-1 in human gastric cancer cell lines. In this study, we examined NRP-1 and EGF-R expression in human gastric cancer cell lines and evaluated the effect of EGF on NRP-1 and VEGF expression. We also investigated major signalling pathways involved in NRP-1 and VEGF induction by EGF.\n\nMATERIALS AND METHODS\n=====================\n\nMaterials\n---------\n\nRecombinant human EGF was purchased from R&D Systems, Inc. (Minneapolis, MN, USA). The anti-human EGF-R monoclonal antibody C225 was provided by ImClone Systems (New York, NY, USA). The mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (MEK) 1/2 inhibitor U0126 and MEK1 inhibitor PD98059 were obtained from New England Biolabs Inc. (Beverly, MA, USA). The phosphatidylinositol-3 (PI-3) kinase/Akt inhibitor wortmannin was purchased from Sigma Chemical Company (St Louis, MO, USA). The P38 MAPK inhibitor SB203580 was purchased from Calbiochem (San Diego, CA, USA). Diaminobenzidine substrate (DAB) and Universal Mount was purchased from Research Genetics (Huntsville, AL, USA). Superfrost slides were purchased from Fisher Scientific Co. (Houston, TX, USA). Streptavidin HRP was purchased from DAKO (Carpinteria, CA, USA), TSA Biotin System was purchased from Perkin Elmer Life Science Inc. (Boston, MA, USA). Antibodies for the immunohistochemical analysis were obtained as follows: rabbit anti-NRP-1 antibody from Santa Cruz (Santa Cruz, CA, USA); mouse anti-EGF-R monoclonal antibody from Zymed (South San Francisco, CA, USA); biotinated goat anti-mouse IgG from Biocare Medical (Walnut Creek, CA, USA); peroxidase-conjugated goat anti-rabbit IgG from Jackson Immuno-Research Laboratories (West Grove, PA, USA).\n\nCell lines and culture conditions\n---------------------------------\n\nThe human gastric carcinoma cell lines AGS and NCI-N87 were obtained from American Type Culture Collection (Manassas, VA, USA). The ST-2, ST-4, and ST-8 human gastric cancer cells were established from patients at The University of Texas MD Anderson Cancer Center (a gift from Dr Bradley McIntyre) ([Yadav *et al*, 1996](#bib46){ref-type=\"other\"}). TMK-1 cells were a gift from Dr Eiichi Tahara (Hiroshima University School of Medicine, Hiroshima, Japan) ([Ito *et al*, 1989](#bib13){ref-type=\"other\"}). KKLS cells were kindly provided by Dr Yutaka Takahashi (Cancer Research Institute, Kanazawa University, Kanazawa, Japan).\n\nThe cell lines were grown in minimum essential medium or RPMI 1640 (Life Technologies, Grand Island, NY, USA) supplemented with 10% foetal bovine serum (FBS) and 2\u2009U\u2009ml^\u22121^ penicillin -- streptomycin. All experiments were performed at subconfluence (70--80%) to avoid variations in VEGF expression due to confluence ([Koura, 1996](#bib26){ref-type=\"other\"}). Before treatment, cells were incubated in 5% FBS-containing medium overnight to minimise the effect of serum starvation on VEGF induction ([Jung *et al*, 1999](#bib17){ref-type=\"other\"}).\n\nReverse transcriptase--polymerase chain reaction\n------------------------------------------------\n\nTotal RNA was extracted from human gastric carcinoma cell lines using Tri Reagent (Molecular Research Center, Inc., Cincinnati, OH, USA) following the manufacturer\\'s instructions. For reverse transcriptase--polymerase chain reaction (RT--PCR), 3\u2009*\u03bc*g of total RNA was used for cDNA synthesis with avian myeloblastosis virus reverse transcriptase (Life Technologies) in a final volume of 20\u2009*\u03bc*l. The reaction mixture included 0.5\u2009[M]{.smallcaps} Tris-HCl (pH 8.0), 0.5\u2009[M]{.smallcaps} KCl, 0.05\u2009[M]{.smallcaps} MgCl~2~, 2.5\u2009m[M]{.smallcaps} dNTP, 40\u2009U of RNase inhibitor (Boehringer Mannheim, Indianapolis, IN, USA), 50\u2009U of reverse transcriptase, and 0.5\u2009*\u03bc*g of random primers. The cDNA synthesis reaction was performed for 1\u2009h at 42\u00b0C. A portion of the reaction mixture (5\u2009*\u03bc*l) was subjected to PCR amplification in a reaction mixture (50\u2009*\u03bc*l) that contained 1\u2009*\u03bc*mol\u2009l^\u22121^ of each of two primers (sense and antisense), 1.5\u2009mmol\u2009l^\u22121^ of MgCl~2~, 0.2\u2009mmol\u2009l^\u22121^ of each of four deoxynucleotides, and 2.5\u2009U of *Taq* polymerase (Promega, Madison, WI, USA). PCR amplification was performed under the following conditions: 94\u00b0C for 5\u2009min; 35 cycles of 1\u2009min denaturing at 94\u00b0C, 30\u2009s of annealing at 57\u00b0C, and 1\u2009min of extension at 72\u00b0C. PCR products were analysed by electrophoresis of 20\u2009*\u03bc*l of each PCR reaction mixture in a 3% agarose gel, and bands were visualised by ethidium bromide staining. The primers chosen were as follows: NRP-1 sense, 3\u2032-ACGATGAATGTGGCGATACT-5\u2032; antisense, 5\u2032-AGTGCATTCAAGGCTGTTGG-3\u2032. Human umbilical vein endothelial cell RNA was used as a positive control.\n\nDetermination of EGF\\'s effects on NRP-1 and VEGF mRNA expression in NCI-N87 and ST-2 cells\n-------------------------------------------------------------------------------------------\n\nTo determine the effects of EGF on NRP-1 and VEGF mRNA expression, NCI-N87 and ST-2 cells were grown to subconfluence in standard medium as described above, and the medium was changed to 5% FBS-containing medium overnight. Cells were then incubated with EGF (50\u2009ng\u2009ml^\u22121^) for 4 or 24\u2009h in 1% FBS-containing medium. Total RNA was extracted, and VEGF expression and NRP-1 expression were determined by Northern blot analysis.\n\nDetermination of C225\\'s effects on NRP-1 and VEGF mRNA induction by EGF in NCI-N87 cells\n-----------------------------------------------------------------------------------------\n\nWe evaluated the ability of C225 to block VEGF and NRP-1 mRNA induction by EGF. NCI-N87 cells grown under the conditions described above were pretreated with or without C225 (10 or 50\u2009*\u03bc*g\u2009ml^\u22121^) in 1% FBS-containing medium for 24\u2009h, and EGF (50\u2009ng\u2009ml^\u22121^) was then added. Cells were harvested after 4 or 24\u2009h, and the relative expression levels of VEGF and NRP-1 mRNA were determined by Northern blot analysis.\n\nDetermination of EGF\\'s effects on Erk, Akt, and P38 phosphorylation in NCI-N87 cells\n-------------------------------------------------------------------------------------\n\nTo determine the effect of EGF on the protein levels and phosphorylation of the signalling intermediates Erk, Akt, and P38 MAPK cells grown under the conditions described above were incubated with EGF (50\u2009ng\u2009ml^\u22121^) for 0, 5, 10, 15, 30, or 60\u2009min in 1% FBS-containing medium, and cell lysates were obtained. Phosphorylated and total protein levels were determined by Western blot analyses as described below.\n\nDetermination of effects of Erk1/2, Akt, and P38 MAPK inhibition on VEGF and NRP-1 induction by EGF\n---------------------------------------------------------------------------------------------------\n\nTo determine the effects of Erk1/2, Akt, and P38 MAPK inhibition on VEGF and NRP-1 induction, NCI-N87 cells grown under the conditions described above were pretreated with 50\u2009*\u03bc*[M]{.smallcaps} PD98059, 10\u2009*\u03bc*[M]{.smallcaps} U0126, 200\u2009n[M]{.smallcaps} wortmannin, or 25\u2009*\u03bc*[M]{.smallcaps} SB203580 for 1\u2009h in 1% FBS-containing medium, and then EGF (50\u2009ng\u2009ml^\u22121^) was added for 24\u2009h. Total RNA was extracted, and Northern blot analysis was performed. Preliminary experiments had shown that inhibitors at these concentrations blocked activation of their targets without inducing cell death (data not shown).\n\nRNA extraction and Northern blot analysis\n-----------------------------------------\n\nTotal RNA was extracted from cells using Tri Reagent following the manufacturer\\'s instructions. Northern blot analysis was performed as previously described ([Jung *et al*, 1999](#bib17){ref-type=\"other\"}). A human VEGF-specific 204-bp cDNA probe was a gift from Dr Brygida Berse (Harvard Medical School, Boston, MA, USA), a human NRP-1-specific 639-bp cDNA probe was a gift from Dr Michael Klagsbrun (Harvard Medical School), and a glyceraldehyde-phosphate dehydrogenase (GAPDH) cDNA probe was purchased from American Type Culture Collection. The VEGF-specific probe identifies all alternatively spliced forms of its mRNA transcripts. Probes were purified by agarose gel electrophoresis using a QIAEX gel extraction kit (QIAGEN, Inc., Chatworth, CA, USA). Each cDNA probe was radiolabelled with \\[*\u03b1*-^32^P\\] deoxyribonucleotide triphosphate according to the random-primer technique using the Rediprime labelling system (Amersham Corp., Arlington Heights, IL, USA). Aliquots (25\u2009g) of total RNA were subjected to electrophoresis in 1% denaturing formaldehyde--agarose gels. The RNA was transferred to a Hybond-N+ positively charged nylon membrane (Amersham Corp.) overnight by capillary elution and subjected to ultraviolet crosslinking at 120\u2009000\u2009*\u03bc*J\u2009cm^\u22122^ using an ultraviolet Stratalinker 1800 (Stratagene, La Jolla, CA, USA). After blots were incubated for 3--4\u2009h at 65\u00b0C in rapid hybridisation buffer (Amersham), the membranes were hybridised overnight at 65\u00b0C with the cDNA probe for VEGF, NRP-1, or GAPDH. The probed nylon membranes were washed and exposed to radiographic film (Life Technologies).\n\nWestern blot hybridisation\n--------------------------\n\nFor Western blot hybridisation, cells were rinsed twice with ice-cold phosphate-buffered saline and then lysed with protein lysis buffer (20\u2009m[M]{.smallcaps} sodium phosphate (pH 7.4), 150\u2009m[M]{.smallcaps} sodium chloride, 1% Triton X-100, 5\u2009m[M]{.smallcaps} EDTA, 5\u2009m[M]{.smallcaps} phenylmethylsulphonyl fluoride, 1% aprotinin, 1\u2009*\u03bc*g\u2009ml^\u22121^ leupeptin, and 500\u2009*\u03bc*[M]{.smallcaps} Na^3^VO^4^). The protein was quantitated spectrophotometrically using a BCA assay (Pierce, Rockford, IL, USA). Aliquots (50\u2009*\u03bc*g) of the protein were subjected to electrophoresis on 8--10% polyacrylamide gels. The protein was then transferred onto a nitrocellulose membrane (Schleicher and Schuell, Keene, NH, USA) by electrotransfer. Following blocking with 5% milk in 0.5% Tween 20 in phosphate-buffered saline, the membrane was probed with the primary antibody (1:1000 dilution of rabbit polyclonal anti-EGF-R antibody (Upstate Biotechnology, Waltham, MA, USA), mouse monoclonal antiphospho-specific p44/42 MAPK (Erk1/2) antibody, rabbit polyclonal antiphospho-specific Akt (site S473) antibody, or rabbit polyclonal antiphospho-specific P38 MAPK antibody (all from Cell Signalling Technology, Beverly, MA, USA)). The membranes were then washed and treated with the secondary antibody labelled with horseradish peroxidase (goat anti-rabbit or anti-mouse immunoglobulin (Amersham) at a 1:3000 dilution). Protein bands were visualised using a commercially available chemiluminescence kit (Amersham). For assaying total protein levels, the membrane was washed with stripping solution (100\u2009m[M]{.smallcaps} 2-mercaptoethanol, 2% sodium dodecyl sulphate, and 62.5\u2009m[M]{.smallcaps} Tris-HCl (pH 6.7)) for 30\u2009min at 65\u00b0C and reprobed with rabbit polyclonal anti-p44/42 (Erk1/2), anti-Akt, or anti-p38 antibody (all at a 1:1000 dilution).\n\nDensitometric quantification\n----------------------------\n\nDensitometric analysis was performed using Image Quant software (Molecular Dynamics, Sunnyvale, CA, USA) to quantify the results of Northern blot analysis (VEGF, NRP-1, and GAPDH mRNA expression) in the linear range of the film. GAPDH mRNA was used as an internal control for loading Northern blots.\n\nImmunohistochemical analysis of human gastric cancer specimens\n--------------------------------------------------------------\n\nExpression of EGF-R and NRP-1 in human gastric cancer specimen was analysed immunohistochemically as previously described ([Jung *et al*, 2002](#bib16){ref-type=\"other\"}) with some modifications. For antigen retrieval, slides for EGF-R were treated with pepsin and incubated at 37\u00b0C for 20\u2009min, slides for NRP-1 were placed in 0.1\u2009[M]{.smallcaps} citrate buffer and heated in a microwave oven for 5\u2009min. To intensify the signals, we used the Tyramide amplification method using the TSA kit (Perkin Elmer). Ten paraffin-embedded human gastric cancer specimens were stained for EGF-R and NRP-1 and the expression pattern was analysed at \u00d7 100 magnification for determination of colocalisation of proteins in serial sections.\n\nRESULTS\n=======\n\nExpression of NRP-1 and EGF-R by human gastric cancer cells\n-----------------------------------------------------------\n\nThe expression of NRP-1 mRNA and EGF-R protein by human gastric cancer cell lines was examined by RT--PCR and Western blot analysis, respectively. RT--PCR revealed that five (TMK-1, AGS, NCI-N87, ST-2, ST-7) of seven gastric cancer cell lines constitutively expressed NRP-1 mRNA. NRP-1 expression was not detected in KKLS and ST-8 cells ([Figure 1](#fig1){ref-type=\"fig\"}Figure 1Expression of NRP-1 and EGF-R by seven human gastric cancer cell lines. Cells were grown to 80% confluence in 10% serum-containing medium. RT--PCR with NRP-1-specific primers was performed to detect NRP-1 RNA expression. Western blots were performed to detect EGF-R protein expression. NRP-1 expression closely mirrored EGF-R expression.). Northern blot hybridisation produced similar results (data not shown). The five NRP-1-positive cell lines also expressed EGF-R protein to various degrees, but EGF-R protein was not detected in the two NRP-1-negative cell lines ([Figure 1](#fig1){ref-type=\"fig\"}).\n\nEffect of EGF on NRP-1 and VEGF expression in NCI-N87 and ST-2 cells\n--------------------------------------------------------------------\n\nTo analyse the regulation of NRP-1 expression by EGF, we studied NCI-N87 and ST-2 human gastric cancer cells, which expressed relatively high levels of EGF-R. Furthermore, others have shown that high levels of EGF-R protein are present in the membrane of NCI-N87 cells ([Basque *et al*, 2001](#bib3){ref-type=\"other\"}). ST-2 cells expressed relatively high levels of EGF-R protein ([Figure 1](#fig1){ref-type=\"fig\"}) when treated with EGF for 4 or 24\u2009h. Control cells were harvested at each time point to exclude any effects on gene expression from increasing cell confluence ([Koura, 1996](#bib26){ref-type=\"other\"}). NRP-1 mRNA expression was slightly induced at 4\u2009h and strongly induced at 24\u2009h by incubation with EGF (3.0-fold over control for NCI-N87; 3.3-fold over control for ST-2) ([Figure 2A](#fig2){ref-type=\"fig\"}Figure 2Induction of NRP-1 (**A**) and VEGF (**B**) in NCI-N87 and ST-2 cells after treatment with EGF. The cells were incubated in 5% serum-containing medium overnight and then were incubated with or without 50\u2009ng\u2009ml^\u22121^ EGF for 4 or 24\u2009h in 1% serum-containing medium. Total RNA was extracted for Northern blot analysis. EGF induced NRP-1 and VEGF mRNA expression at 4 and 24\u2009h.). Similar NRP-1 induction by EGF was also observed in AGS cells (data not shown).\n\nIn preliminary experiments, we found that all gastric cell lines studied expressed VEGF mRNA at various levels (data not shown). We therefore examined the effect of EGF on VEGF mRNA induction using NCI-N87 and ST-2 cells. EGF induced VEGF mRNA expression at 4\u2009h (2.3-fold over control for NCI-N87 cells; 3.8-fold over control for ST-2 cells), and the induction continued at 24\u2009h ([Figure 2B](#fig2){ref-type=\"fig\"}).\n\nEffect of C225 on NRP-1 and VEGF induction by EGF in NCI-N87 cells\n------------------------------------------------------------------\n\nWe investigated the effect of blockade of EGF-R with C225 on the induction of NRP-1 and VEGF following treatment with EGF. Pretreatment of the cells with C225 (10 or 50\u2009*\u03bc*g\u2009ml^\u22121^) for 24\u2009h followed by treatment with EGF inhibited induction of NRP-1 and VEGF in a dose-dependent manner ([Figure 3](#fig3){ref-type=\"fig\"}Figure 3Inhibitory effect of C225 on NRP-1 (**A**) and VEGF (**B**) mRNA induction by EGF in NCI-N87 cells. The cells were incubated in 5% serum-containing medium overnight and then were pretreated with or without C225 (10 or 50\u2009*\u03bc*g\u2009ml^\u22121^) in 1% FBS-containing medium for 24\u2009h. EGF (50\u2009ng\u2009ml^\u22121^) then was or was not added, and cells were harvested after 4 or 24\u2009h. Relative expression levels of VEGF and NRP-1\u2009mRNA were determined by Northern blot analysis. Pretreatment of the cells with C225 inhibited EGF\\'s effect in a dose-dependent manner.). At a C225 dose of 50\u2009*\u03bc*g\u2009ml^\u22121^, VEGF expression induced by EGF was nearly completely blocked ([Figure 3B](#fig3){ref-type=\"fig\"}), whereas NRP-1 expression was only partially blocked ([Figure 3A](#fig3){ref-type=\"fig\"}).\n\nEffect of EGF on signalling pathways involved in NRP-1 and VEGF induction in NCI-N87 cells\n------------------------------------------------------------------------------------------\n\nTo determine the signalling pathways induced by EGF in NCI-N87 cells, Western blot analysis was performed after incubation of cells with EGF for various durations. As shown in [Figure 4](#fig4){ref-type=\"fig\"}Figure 4Effect of EGF on Erk1/2, Akt, and P38 phosphorylation in NCI-N87 cells. The cells were incubated in 5% serum-containing medium overnight and then were incubated with 50\u2009ng\u2009ml^\u22121^ EGF for the indicated duration in 1% serum-containing medium. Phosphorylated and total protein levels were determined by Western blot analyses. EGF led to induction of phosphorylated Erk1/2, Akt, and P38., a moderate increase in the phosphorylation of Erk1/2 began within 5\u2009min of EGF treatment and returned to the basal level at 60\u2009min. EGF also moderately increased Akt phosphorylation as early as 5\u2009min of incubation, and the effect continued for at least 60\u2009min. EGF only minimally increased phosphorylation of P38 MAPK 15\u2009min after treatment. EGF also minimally increased phosphorylation of c-jun amino-terminal kinase (JNK) (data not shown). The relative expression levels of total Erk1/2, Akt, and P38 were not significantly altered after EGF treatment ([Figure 4](#fig4){ref-type=\"fig\"}).\n\nWe next selectively blocked the Erk1/2, Akt, or P38 MAPK pathways to determine which pathway was essential for EGF induction of NRP-1 and VEGF mRNA expression in NCI-N87 cells. The results of Northern blot analysis are shown in [Figure 5](#fig5){ref-type=\"fig\"}Figure 5Effect of Erk1/2, Akt, and P38 MAPK inhibition on NRP-1 and VEGF induction by EGF in NCI-N87 cells. The cells were incubated in 5% serum-containing medium overnight and then were pretreated with 50\u2009*\u03bc*[M]{.smallcaps} PD98059, 10\u2009*\u03bc*[M]{.smallcaps} U0126, 200\u2009n[M]{.smallcaps} wortmannin, or 25\u2009*\u03bc*[M]{.smallcaps} SB203580 for 1\u2009h in 1% FBS-containing medium. EGF (50\u2009ng\u2009ml^\u22121^) was then added for 24\u2009h. Control cells were not treated with EGF (lane 1) and cells treated with EGF without addition of signalling inhibitors served as another internal control (lane 2). Total RNA was extracted, and Northern blot analysis was performed. Blockade of the Erk1/2, Akt or P38 pathways all led variable decreases in NRP-1 and VEGF mRNA expression.. PD98059, U0126, and SB203580 effectively inhibited NRP-1 induction by EGF, whereas wortmannin did not. In addition, SB203580 nearly completely abrogated VEGF induction by EGF, and wortmannin and U0126 partially blocked VEGF induction, to a similar degree. Thus, blockade of the Erk1/2 or P38 MAPK pathway effectively suppressed NRP-1 mRNA induction by EGF. VEGF mRNA induction was completely inhibited by blockade of the P38 pathway.\n\nImmunohistochemical analysis of human gastric cancer specimens for NRP-1 and EGF-R\n----------------------------------------------------------------------------------\n\nWe stained 10 paraffin-embedded human gastric cancer specimen for EGF-R and NRP-1 ([Figure 6](#fig6){ref-type=\"fig\"}Figure 6Immunohistochemical staining of intestinal-type human gastric cancer for EGF-R and NRP-1. EGF-R and NRP-1 were colocalised in a glandular pattern in a moderately differentiated gastric cancer specimen. Original magnification \u00d7 100.). Six samples were differentiated (intestinal-type) and four were poorly differentiated (diffuse-type). Overall, eight of 10 specimens expressed nrp-1 in the tumour epithelial cells and nine of 10 expressed EGF-R. Among the differentiated tumours, half of them expressed nrp-1 and two-thirds expressed egf-r. Two of the six intestinal-type tumours demonstrated colocalisation of NRP-1 and EGF-R. Among the poorly differentiated tumours, half of them demonstrated colocalisation of NRP-1 and EGF-R.\n\nDISCUSSION\n==========\n\nIn the present study, we examined the expression of NRP-1 and EGF-R and the effect of EGF on NRP-1 and VEGF expression in human gastric cancer cells. Five of seven gastric cancer cell lines constitutively expressed NRP-1\u2009mRNA. The expression pattern of NRP-1 was directly related to EGF-R protein expression, and EGF increased not only NRP-1 but also VEGF mRNA expression in two different cell lines. We also confirmed the expression pattern of NRP-1 and EGF-R in human gastric cancer specimens and found one-third of the differentiated cancers and half of the undifferentiated cancers demonstrated colocalisation of these two proteins (this implies an association, but causality cannot be confirmed by immunohistochemical studies of human specimens). These data suggest that regulation of NRP-1 expression is associated with EGF-R activation and that ligands for EGF-R could theoretically contribute to tumour angiogenesis by a mechanism that involves upregulation of VEGF and NRP-1 expression in this tumour system. Furthermore, EGF induction of NRP-1 and VEGF might involve multiple signalling pathways, thus providing multiple potential targets for inhibiting the induction of NRP-1 and VEGF. Further study is needed to validate the importance of EGF-R activation and NRP-1 induction.\n\nNRP-1 is expressed on ECs and tumour cells. Various tumour cell types, such as breast and prostate cancers, express abundant levels of NRP-1 mRNA, whereas low levels of NRP-1 are found in some normal adult tissues ([Soker *et al*, 1998](#bib37){ref-type=\"other\"}). Moreover, NRP-1 expression appears to correlate with advanced stage or grade in prostate cancer and astrocytoma ([Ding *et al*, 2000](#bib6){ref-type=\"other\"}; [Latil *et al*, 2000](#bib27){ref-type=\"other\"}). In our study, five of seven gastric cancer cell lines expressed NRP-1 mRNA. The exact role of NRP-1 in tumour cells remains to be elucidated; however, it is possible that NRP-1 may augment tumour angiogenesis and/or tumour cell survival ([Miao *et al*, 2000](#bib31){ref-type=\"other\"}; [Bachelder *et al*, 2001](#bib2){ref-type=\"other\"}).\n\nLittle is known about the role of NRP-1 in tumour cells. NRP-1 is a nontyrosine kinase transmembrane protein ([Soker *et al*, 1998](#bib37){ref-type=\"other\"}). Unlike ECs, tumour cells may express NRP-1 as the only VEGF receptor. Overexpression of NRP-1 in prostate carcinoma cells enhances angiogenesis and increases proliferation of ECs, suggesting that the expression of NRP-1 by tumour cells themselves can influence tumour growth and angiogenesis ([Miao *et al*, 2000](#bib31){ref-type=\"other\"}). Furthermore, expression of NRP-1 in tumour cells enhances binding of VEGF-165 to these cells ([Miao *et al*, 2000](#bib31){ref-type=\"other\"}). It is possible that NRP-1 on tumour cells binds to VEGF-165 and acts as a carrier of VEGF to enhance VEGF binding to VEGF tyrosine kinase receptors on ECs. Alternatively, VEGF-165 might stimulate tumour cells directly via NRP-1 through an unknown mechanism. A recent study has shown that VEGF-165 acts as an autocrine survival factor in NRP-1-positive breast carcinoma cells lacking VEGFR-2 and that this likely occurs through activation of the PI-3 kinase pathway ([Bachelder *et al*, 2001](#bib2){ref-type=\"other\"}). It has also been reported that VEGF-165 may act as an autocrine growth factor in a VEGFR-positive human gastric cancer cell line ([Tian *et al*, 2001](#bib42){ref-type=\"other\"}). NRP-1 could act as a coreceptor that enhances VEGF-165 function in both VEGFR-2- and NRP-1-positive tumour cells. We investigated the effect of VEGF-165 on proliferation of tumour cells in NRP-1-positive NCI-N87, ST-2, and TMK-1 cells, but VEGF-165 had no effect on cell growth in these cells (data not shown).\n\nVEGF expression is known to be regulated by numerous cytokines and growth factors ([Akagi *et al*, 1998](#bib1){ref-type=\"other\"}). In contrast, factors that regulate NRP-1 expression in ECs and tumour cells are not fully elucidated. Recent studies have shown that tumour necrosis factor-*\u03b1* and VEGF increase NRP-1 expression in human and bovine retinal ECs, respectively ([Giraudo *et al*, 1998](#bib9){ref-type=\"other\"}; [Oh *et al*, 2002](#bib32){ref-type=\"other\"}). Other investigators have shown that EGF increases NRP-1 expression in human malignant astrocytoma cell lines and that the expression of NRP-1 mRNA peaks at 4\u2009h and returns to basal levels 8\u2009h after EGF treatment ([Ding *et al*, 2000](#bib6){ref-type=\"other\"}). In our study, NRP-1 mRNA expression was induced by EGF treatment at 24\u2009h in three different gastric cancer cell lines (AGS, NCI-N87, and ST-2). Further experiments revealed that EGF also increased VEGF mRNA expression in NCI-N87 and ST-2 cells but not in AGS cells (data not shown for AGS cells). The factors involved in the regulation of VEGF may thus be dependent upon the tumour system under study ([Akagi *et al*, 1998](#bib1){ref-type=\"other\"}). In TMK-1 cells (cells that express relatively low levels of EGF-R protein), EGF did not increase either NRP-1 or VEGF (data not shown).\n\nC225, an anti-EGF-R monoclonal antibody, binds EGF-R with affinity similar to that of EGF and is able to block EGF binding and EGF activation of the receptor. EGF-R signalling is critical for cell proliferation. Furthermore, EGF-R-mediated signals contribute to other processes that are crucial to cancer progression, including angiogenesis, metastatic spread, and the inhibition of apoptosis ([Kim *et al*, 2001](#bib22){ref-type=\"other\"}). Recently, several studies have demonstrated that *in vitro* treatment of cancer cells with C225 can downregulate the production of angiogenic factors such as VEGF, interleukin-8, or basic fibroblast growth factor and that *in vivo* inhibition of EGF-R results in growth inhibition and reduction in microvessel density accompanied by decreases in angiogenic factor expression ([Petit *et al*, 1997](#bib34){ref-type=\"other\"}; [Perrotte *et al*, 1999](#bib33){ref-type=\"other\"}; [Bruns *et al*, 2000](#bib4){ref-type=\"other\"}; [Ciardiello *et al*, 2000](#bib5){ref-type=\"other\"}). These results strongly support the involvement of the EGF-R signalling pathways in the regulation of angiogenesis and suggest that C225 could have therapeutic utility, not only through its ability to inhibit the growth of tumour cells but also through its ability to suppress tumour angiogenesis. In our study, C225 blocked both VEGF and NRP-1\u2009mRNA induction by EGF in a dose-dependent manner in NCI-N87 cells.\n\nEGF has been shown to influence various signalling pathways, including ras--raf--MEK--MAPKs, PI-3 kinase/Akt, and the stress-activated protein kinases (SAPKs or JNKs) ([Yarden and Sliwkowski, 2001](#bib48){ref-type=\"other\"}). We demonstrated that, in NCI-N87 cells, the Erk1/2, Akt, and P38 MAPK pathways are activated by EGF treatment. The modest level of activation may be secondary to the relative moderate levels of EGF-R compared to other cell types ([Figure 1](#fig1){ref-type=\"fig\"}). Blockade of the Erk1/2 or P38 MAPK pathway inhibited EGF-induced NRP-1 expression more effectively than did inhibition of the Akt pathway. In addition, blockade of P38 activity nearly completely inhibited EGF-induced VEGF expression, and blockade of the Erk1/2 or Akt pathway partially inhibited VEGF induction. Taken together, the data suggest that the signalling mechanisms leading to NRP-1 induction by EGF are distinct from those that induce VEGF. However, blockade of P38 MAPK inhibited EGF induction of both NRP-1 and VEGF and might be a therapeutic target in this tumour system. The P38 pathway has not been extensively investigated in angiogenic systems; however, we previously found that treatment of human vascular smooth muscle cells with interleukin-1*\u03b2* leads to VEGF induction via P38 MAPK activation ([Jung *et al*, 2001](#bib15){ref-type=\"other\"}). Others have also shown that P38 can be phosphorylated by EGF-R activation ([Kanda *et al*, 2001](#bib18){ref-type=\"other\"}; [Yamanaka *et al*, 2001](#bib47){ref-type=\"other\"}). Taken together, our studies along with others support the role of EGF-R activation of angiogenic pathways through P38. Thus, P38 may be a common angiogenic signalling pathway in multiple cell types.\n\nThe mechanisms by which the EGF-R signalling pathways regulate VEGF and NRP-1 are unclear. Stimulation of the EGF-R signalling pathways is known to activate ras and raf, resulting in phosphorylation of c-fos and c-jun and leading to increased AP-1 transcriptional activity. The VEGF promoter has several AP-1 binding sites and increased AP-1 activity leads to transcription of VEGF ([Rozakis-Adcock, 1993](#bib36){ref-type=\"other\"}; [Kerbel *et al*, 1998](#bib20){ref-type=\"other\"}). The PI-3 kinase pathway also plays a role in VEGF induction by EGF-R signalling ([Maity *et al*, 2000](#bib30){ref-type=\"other\"}).\n\nStudies in an astrocytoma cell line showed that activation of p21-Ras induces not only VEGF but also NRP-1 expression ([Ding *et al*, 2000](#bib6){ref-type=\"other\"}). A recent study has shown that NRP-1 is the downstream target of transcription factor Ets-1 in ECs ([Teruyama *et al*, 2001](#bib41){ref-type=\"other\"}). VEGF is a potent inducer of Ets-1 in ECs, and this induction of Ets-1 is mediated by the activation of Erk1/2 ([Tanaka *et al*, 1999](#bib40){ref-type=\"other\"}).\n\nIn summary, we have shown that EGF and EGF-R play a role in the regulation of NRP-1 and VEGF expression via multiple signalling pathways in human gastric cancer cells. Further studies are required to determine the clinical importance of activation of the EGF-R signalling pathways and the downstream effect on VEGF and NRP-1 expression.\n\nThis work was supported, in part, by the Carlos Cantu Foundation (PFM), the Gillsohn Longenbaugh Foundation (LME), and National Institutes of Health Cancer Center support Grant CA16672. We thank Melissa G Burkett of the Department of Scientific Publications and Rita Hernandez of the Department of Surgical Oncology, MD Anderson Cancer Center, for their editorial assistance.\n"}
+{"text": "INTRODUCTION\n============\n\nAcquired hemophilia A (AHA) is a rare bleeding disorder caused by autoantibodies against factor VIII (FVIII) \\[[@B1]\\]. It is diagnosed in patients without previous or familial histories of bleeding who have isolated prolongation of the activated partial thromboplastin time (aPTT) that cannot be corrected by mixing study and who have reduced FVIII levels as well as evidence of FVIII inhibitor activity \\[[@B2]\\]. FVIII autoantibodies occur in patients in the postpartum period and those with concomitant diseases or conditions (i.e., autoimmune disorders, malignancies, and drug abuse). However, up to 50% of cases develop without any relevant medical illness \\[[@B3]\\]. The mortality rate of AHA is as high as 16% \\[[@B4]\\].\n\nClopidogrel, an antiplatelet agent commonly used for coronary disease or cerebral vascular disease, is a cause of AHA \\[[@B5]\\]. However, only 5 cases of clopidogrel-related AHA have been described to date \\[[@B5], [@B6]\\]. Because the exact natural history of clopidogrel-associated AHA is unknown, knowledge of its clinical features is important. The present report describes a patient with AHA associated with clopidogrel use who showed complete remission after steroid administration and discontinuation of clopidogrel.\n\nCASE REPORT\n===========\n\nA 65-year-old man was hospitalized for cerebellar infarction. He was referred to the hematology department due to prolonged aPTT. He had been administered aspirin, clopidogrel, and valproate for cerebellar infarction, a statin for hyperlipidemia, and cefoperazone, clindamycin, and teicoplanin for aspiration pneumonia. He had no previous or familial medical history of bleeding or coagulopathy and no clinical symptoms or signs of malignancy, antiphospholipid syndrome, or collagen vascular disease.\n\nAt admission, he was drowsy without clinical evidence of bleeding. He was intubated, and a tracheostomy was performed. Initial laboratory results showed normal complete blood cell counts, kidney and liver function, and coagulation parameters. However, 3 weeks after initiation of clopidogrel, blood oozed from his tracheostomy site, and his aPTT was 40.3 seconds (normal range, 27.9-37.8 seconds) ([Fig. 1](#F1){ref-type=\"fig\"}). Clopidogrel was considered the cause of bleeding, and its administration was discontinued. However, blood continued to ooze from his tracheostomy stoma, and his aPTT gradually rose up to 98.8 seconds ([Fig. 1](#F1){ref-type=\"fig\"}) without platelet or PT changes despite discontinuation of clopidogrel and a transfusion with fresh frozen plasma. A mixing test did not correct the aPTT. Tests for lupus anticoagulant, anticardiolipin antibody, and antinuclear antibody were negative. His FVIII was below 1% (reference range, 60-140%), and his FVIII inhibitor titer was 5.4 Bethesda units/mL (BU/mL) ([Fig. 1](#F1){ref-type=\"fig\"}).\n\nWe suspected clopidogrel-associated AHA and began steroid administration. Within a few days, the bleeding was controlled, and the patient\\'s aPTT decreased to normal levels. However, 3 weeks after steroid use, active bleeding developed from his stoma. His aPTT was normal, and FVIII inhibitor was undetectable ([Fig. 1](#F1){ref-type=\"fig\"}). However, his FVIII was only 8% ([Fig. 1](#F1){ref-type=\"fig\"}). We performed a left inferior thyroid arterial branch ligation and administered human FVIII concentrates (GreenMono, Greencross, Gyeonggi-do, Korea). Afterward, there were no further bleeding episodes, and his FVIII was normalized 2 months after steroid treatment. The patient was gradually tapered off steroids in 3 months. During the 2-year follow-up period, his aPTT remained within normal levels.\n\nDISCUSSION\n==========\n\nClopidogrel is increasingly used as a first-line antiplatelet agent in patients with coronary artery disease, cerebral vascular disease, and peripheral vascular disease \\[[@B7]\\]. Clopidogrel is known to have adverse hematologic effects, such as neutropenia \\[[@B8], [@B9]\\], idiopathic immune thrombocytopenia \\[[@B10]\\], thrombotic thrombocytopenic purpura \\[[@B11]\\], and hemolytic uremic syndrome \\[[@B12]\\]. Recently, 5 cases of clopidogrel-associated AHA have been reported \\[[@B5], [@B6]\\].\n\nIn general, there is an equal distribution of AHA between the sexes. However, there is a female predominance in the younger age group because of an association with pregnancy, and males constitute the majority of patients over the age of 60 who develop AHA \\[[@B13], [@B14]\\]. However, in 5 cases of clopidogrel-associated AHA, 4 were females, and all were over the age of 60 years (67-78 years) \\[[@B6]\\]. Nevertheless, a predominance of clopidogrel-associated AHA in females is difficult to conclude due to the rarity of the condition. It appears more likely to occur in older patients, who tend to have vascular diseases and, as a result, more commonly use clopidogrel. In the present case, soft-tissue bleeding and aPTT prolongation developed within 3 weeks after clopidogrel treatment. Because the patient was hospitalized during that episode, we were able to detect the abnormal aPTT earlier than in 2 previous cases, where AHA developed in 2-3 months after clopidogrel use \\[[@B5]\\].\n\nIn the literature, there are 34 cases of drug-induced AHA with FVIII inhibitor titers ranging from 1.6 to 250 BU/mL (average, 67.7 BU/mL) \\[[@B6]\\]. Those with clopidogrel-associated AHA had FVIII titers from 2.2 to 17 \\[[@B6]\\]. The present patient had a FVIII titer of 5.4 BU/mL. However, according to that analysis \\[[@B6]\\], FVIII inhibitor titers were not well correlated with FVIII levels or clinical manifestation. In the present case, FVIII levels were normalized in about 7 weeks after the FVIII inhibitor titer was normalized. This discordance may arise from the different pattern of FVIII inactivation as a result of autoantibodies in AHA versus alloantibodies in congenital hemophilia \\[[@B15]\\]. In fact, whereas alloantibodies usually completely inactivate FVIII activity (type I kinetics), autoantibodies often incompletely inactivate FVIII activities (type II kinetics), and some residual FVIII activity remains in the patient\\'s plasma \\[[@B15]\\]. Therefore, the Bethesda assay, which quantifies the in vitro inhibitor titer, may underestimate in vivo inhibitor potency in AHA \\[[@B6]\\].\n\nThere are no guidelines with respect to treatment for AHA, although its mortality rate is as high as 16% \\[[@B4]\\]. In over 80% of the cases, FVIII inhibitor disappeared after discontinuation of clopidogrel or after immunosuppressive therapy \\[[@B6]\\]. The present patient\\'s FVIII was normalized after 7 weeks of steroid use. This is similar to 2 cases reported by Haj et al., in which FVIII returned to normal levels, and FVIII inhibitor was undetectable within 8 weeks of steroid treatment \\[[@B5]\\] and remained normal for 2 years. Therefore, bleeding tendencies in patients using clopidogrel should not be regarded as an antiplatelet effect of clopidogrel unless a platelet count and coagulation screen are normal. Immunosuppressive therapy should be initiated as soon as AHA is diagnosed.\n\n![Changes in aPTT, FVIII, and FVIII inhibitor levels as well as the use of blood products or medication after clopidogrel treatment. Abbreviations: aPTT, activated partial thromboplastin time; FVIII, factor VIII; FFP, fresh frozen plasma.](kjh-47-80-g001){#F1}\n"}
+{"text": "1. Introduction {#sec1-ijms-20-01974}\n===============\n\nDiapause is an adaptation to seasonality that is widespread across invertebrate taxa which allows them to respond to periodic environmental changes in different developmental stages. Diapause can occur during any stage of development in insects including the egg, larva, pupa and adult stages \\[[@B1-ijms-20-01974],[@B2-ijms-20-01974],[@B3-ijms-20-01974],[@B4-ijms-20-01974]\\]. However, this diapause induction phase occurs at a genetically predetermined sensitive stage of life, which can be in diapausing individuals or in preceding generations of insects such as silkworm \\[[@B5-ijms-20-01974],[@B6-ijms-20-01974]\\]. The mosquito *Culex pipiens* can enter a reproductive diapause characterized by an arrest in ovarian development \\[[@B5-ijms-20-01974]\\]. Similarly, the embryonic diapause termination of *Bombyx mori* is associated with the activation of sorbitol dehydrogenase gene (*SDH*) \\[[@B7-ijms-20-01974]\\]. A photoperiod signal can regulate the diapause induced by FOXO through an insulin signaling pathway with circadian genes as the input module, meanwhile at the same time, insulin can also regulate the synthesis of juvenile hormone to achieve the diapause process \\[[@B5-ijms-20-01974],[@B8-ijms-20-01974]\\]. Unlike most insects, diapause induction of locusts is a trans-generational process. Changes due to the environment in the maternal parent could lead to transference of the diapause factor to the offspring eggs \\[[@B9-ijms-20-01974],[@B10-ijms-20-01974],[@B11-ijms-20-01974]\\]. However, the physiological changes during diapause are largely conserved across species and are hypothesized to be regulated by a conserved suite of genes \\[[@B12-ijms-20-01974]\\]. The offspring eggs of locusts need to be induced under a low temperature until they cease development in the late anatrepsis stage before the embryo enters diapause \\[[@B13-ijms-20-01974]\\]. Diapause eggs were found to have relatively strong resistance to cold. The success of egg diapause in winter has a direct effect on the size of locust populations during spring \\[[@B14-ijms-20-01974]\\]. Maternal effects on transgenerational diapause, especially for the photoperiod, are critical for understanding the locust outbreak dynamics after winter. Hence, *Locusta migratoria* has been used as a model insect to understand the mechanism of insect diapause induction from maternal parents to their offspring.\n\nOur previous studies on diapause induction in migratory locust eggs by transcriptome and proteomic analysis have shown that cellular metabolism in diapause eggs is more active compared to non-diapause eggs where specific enzymes played a role in cryoprotection and provided stored energy for up-regulation in the diapause induction stage \\[[@B15-ijms-20-01974]\\]. But how the maternal parent is being induced by photoperiod to produce diapause eggs is still unknown. Therefore, it is particularly important to know the molecular mechanism of the maternal effect induced by either a short or long photoperiod. Fat body, a loosely organized tissue in arthropods, has major functions of nutrient storage, hormone synthesis and vitellogenesis besides other vital activities. Vitellogenesis is the process of yolk formation (vitellogenin or egg yolk protein) via nutrients being deposited in the oocyte or female germ cell involved in reproduction of lecithotrophic organisms. In insects, vitellogenesis starts when the fat body stimulates the release of juvenile hormones and produces proteins. Entry into vitellogenesis is an important stage of oogenesis and by forcing females into reproductive diapause, oogenesis can easily be arrested at the pre-vitellogenic stages \\[[@B16-ijms-20-01974]\\]. Photoperiodic signals are possibly transmitted to eggs by proteins synthesized from fat bodies. Hence, we analyzed the transcriptomes of fat bodies and ovaries of adult migratory locusts induced by long and short photoperiod. The key genes related to diapause induction were obtained and verified for their specific functions by RNAi, and we determined their regulatory relationship with the FOXO signaling pathway. This could be helpful to provide a reference for studying the mechanism of diapause induction in many other insects as well. It also generates insight into the monitoring and managing of pest outbreaks in a specific environment.\n\n2. Results {#sec2-ijms-20-01974}\n==========\n\n2.1. Transcriptomic Analyses {#sec2dot1-ijms-20-01974}\n----------------------------\n\nThe fat body (FAT) and ovary (OVA) transcriptomes of both long (L) and short (S) photoperiods were sequenced independently. Twelve mRNA libraries were generated from fat body under long (L_FAT) and short photoperiods (S_FAT), and ovary under long (L_OVA) and short photoperiods (S_OVA). Three biological repeats were detected for each group. 75.6--90.7 million clean reads with Q20 \\> 95% were obtained along with 10.6--13.6 clean bases ([Table 1](#ijms-20-01974-t001){ref-type=\"table\"}). FAT and OVA transcriptomes were then assembled into 260,779 and 323,527 transcripts individually. Similarly, 102,273 and 132,147 unigenes obtained from FAT and OVA transcriptomes were annotated. To uncover the molecular mechanism underlying these transcriptomic profiles, gene function was annotated based on seven databases including NR, GO, Pfam, SwissProt, KEGG, COG and NT ([Table 1](#ijms-20-01974-t001){ref-type=\"table\"}) by BLAST (*e*-value \\< 0.00001). A total of 517 up-regulated and 236 down-regulated transcripts were found in S_FAT versus the L_FAT group. Similarly, for the ovary samples, a total of 3582 up-regulated and 1371 down-regulated transcripts were found in S_OVA versus the L_OVA group ([Figure 1](#ijms-20-01974-f001){ref-type=\"fig\"}). Up-regulated transcripts were \\~2.5 times greater than the down-regulated transcripts in both groups, which suggested most genes were induced by a short photoperiod. In addition, the number of DEGs found in the OVA group was \\~6 times greater than in the FAT group, which indicated that genes related to photoperiod in ovaries were greater than in fat bodies. Correlation analysis showed that a total of 137 transcripts, including 90 positive and 47 negative correlation DEGs, were differentially expressed in both OVA and FAT groups ([Figure S1](#app1-ijms-20-01974){ref-type=\"app\"}). To identify the photoperiod induced expression profile in both OVA and FAT groups, the 137 correlated DEGs were subsequently clustered by CLUSTER 3.0 software ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}).\n\n**Cluster I (eight genes)**: photoperiod related genes unique to fat bodies. Total of eight genes were identified in this cluster, including glypican 6, phosphoenolpyrucate carboxykinase, HIRA-interacting protein 3, etc. For this cluster, only DEGs of S_FAT down-regulated while DEG associated with L_FAT and S/L_OVA up-regulated ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}, [Table S1](#app1-ijms-20-01974){ref-type=\"app\"}). **Cluster II (74 genes)**: negative regulation of photoperiod related genes. We obtained a total of 74 genes where the key negative regulating photoperiod related genes include NADH dehydrogenase, retinoic acid-induced protein 1, insulin receptor substrate 1, etc. For this cluster, expressions of most genes were lower under a short photoperiod than under a long photoperiod in both FAT and OVA ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}, [Table S1](#app1-ijms-20-01974){ref-type=\"app\"}). **Cluster III (25 genes)**: Cluster III consists of 25 photoperiod related genes that are also unique to fat bodies including actin, longitudinals lacking protein, glyceraldehyde-3-phosphate dehydrogenase, etc. For this cluster, expressions of genes were significantly different in the FAT group as compared to OVA, which was non-significant ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}, [Table S1](#app1-ijms-20-01974){ref-type=\"app\"}). **Cluster IV (five genes):** This cluster has the minimum number of genes. Photoperiod related genes were inversely expressed in fat bodies and ovaries including kazal-type serine protease inhibitor, myelin regulatory factor, quinohemoprotein amine dehydrogenase, etc. For this cluster, expressions of genes were higher in S_FAT and L_OVA as compared to L_FAT and S_OVA where the expression was lower ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}, [Table S1](#app1-ijms-20-01974){ref-type=\"app\"}). **Cluster V (25 genes)**: Cluster V has a total of 25 genes where the positive regulation of photoperiod related genes includes ribosomal protein, succinate dehydrogenase, filamin, etc. For this cluster, expression of most genes was higher under a short photoperiod than under a long photoperiod in both FAT and OVA ([Figure 2](#ijms-20-01974-f002){ref-type=\"fig\"}, [Table S1](#app1-ijms-20-01974){ref-type=\"app\"}).\n\n2.2. KEGG Pathways Analysis {#sec2dot2-ijms-20-01974}\n---------------------------\n\nTo investigate the biological functions, all of the selected DEGs of FAT group were mapped to 193 pathways, while DEGs of OVA group were mapped to 288 pathways by aligning to Kyoto Encyclopedia of Genes and Genomes (KEGG) database. To identify the DEGs related to diapause in FAT and OVA group under the two photoperiods, enrichment analysis was performed for all KEGG pathways. Results showed that ten biological pathways were enriched (*p* \\< 0.05) in S_FAT vs. L_FAT up-regulated DEGs, including ribosome (ko03010), TCA cycle (ko00020), glycolysis (ko00010), etc., while ten biological pathways were enriched (*p* \\< 0.05) in S_FAT vs. L_FAT down-regulated DEGs, including oxidative phosphorylation (ko00190), glyoxylate and dicarboxylate metabolism (ko00630), peroxisome (ko04146), etc. Similarly, 19 biological pathways were enriched (*p* \\< 0.05) in S_OVA vs. L_OVA up-regulated DEGs, including ribosome (ko03010), protein processing in endoplasmic reticulum (ko04141), oocyte meiosis (ko04114), etc., while ten biological pathways were enriched (*p* \\< 0.05) in S_OVA vs. L_OVA down-regulated DEGs, including the thyroid hormone signaling pathway (ko04919), focal adhesion (ko04510), the longevity regulating pathway (ko04213), etc. ([Table 2](#ijms-20-01974-t002){ref-type=\"table\"}). To find out some common pathways in both FAT and OVA groups, the correlated DEGs were also annotated to the KEGG database. Oxidative phosphorylation (ko00190), ribosome (ko03010), biosynthesis of secondary metabolites (ko01110), TCA cycle (ko00020) and glycolysis (ko00010) were the top 5 annotated pathways ([Table S2](#app1-ijms-20-01974){ref-type=\"app\"}). DEGs related to oxidative phosphorylation, including COX1, COX2, COX3, ATPeF0A, ND1, ND2, ND3, ND4, ND5 were down-regulated under a short photoperiod than a long photoperiod in both FAT and OVA groups, while TCA cycle and glycolysis, including MDH1, SDHA, GAPDH, PGM2, were up-regulated. These pathways and genes mentioned above were closely related to the FOXO signaling pathway. In addition, ribosome genes including RP-L10e, RP-L18Ae and RP-L26e were up-regulated under a short photoperiod rather than under a long photoperiod in both FAT and OVA groups ([Table S2](#app1-ijms-20-01974){ref-type=\"app\"}).\n\n2.3. Validations of DEGs Quantitative Real-Time PCR (qRT-PCR) {#sec2dot3-ijms-20-01974}\n-------------------------------------------------------------\n\nTo confirm the reliability of RNA sequencing data, the mRNA level of 24 DEGs in both FAT and OVA groups including *RP-L10e*, *cox2*, *dual specificity protein kinase* (*shkC*), *retinoic acid-induced protein 1* (*rai1*), etc., were selected randomly for quantitative real-time reverse transcription PCR (qRT-PCR) analysis. The result showed that 43 out of 48 DEGs expression patterns were consistent with RNA sequencing data, though the fold change findings of qRT-PCR and RNA-Seq were not exactly matched ([Figure 3](#ijms-20-01974-f003){ref-type=\"fig\"}). Scatter plots of DEGs mRNA level versus FPKM from RNA-Seq ([Figure S2](#app1-ijms-20-01974){ref-type=\"app\"}) revealed that our RNA sequencing data are valid as they were significantly correlated (Pearson *R* = 0.7141, *R*^2^ = 0.5099, *F* = 47.860, *p* \\< 0.0001).\n\n2.4. Rai1 and foxo Functions Identified by RNAi {#sec2dot4-ijms-20-01974}\n-----------------------------------------------\n\nPhotoperiod's closely related gene '*rai1*' might play an important part in insect diapause induction. The *rai1* gene was then selected for further identification of the diapause related functions. The *rai1* sequence was initially analyzed by NCBI BLAST. The length of the *rai1* coding sequence was 3243 bp, which could have encoded a protein with 1080 amino acid. However, the percent identity and coverage with *rai1* of *Zootermopsis nevadensis* was 49 and 67, respectively. Similarly, an extended PHD finger found in *rai1* (Ephd_rai1_like domain), in interval of 888--1067 aa of *rai1*. To verify the function of *rai1* on regulating locust diapause, dsRNA of *rai1* was synthesized and subsequently injected into a female adult of *L. migratoria* to RNAi *rai1* under long and short photoperiods, followed by confirming RNAi efficiency via qRT-PCR. It was observed that the mRNA level of *rai1* was significantly (*p* \\< 0.05) lower in treatment than in control (inject ddH~2~O) under long and short photoperiods, which indicated acceptability of RNAi efficiency of *rai1* ([Figure 4](#ijms-20-01974-f004){ref-type=\"fig\"}A). Under a short photoperiod, average egg diapause rate (DR) was 79.4% in RNAi *rai1* treatment, significantly (*p* = 0.0214) higher than control (62.1%). However, diapause rate had no significant difference between RNAi treatment and control under a long photoperiod. We therefore concluded that RNAi maternal *rai1* could promote offspring diapause under short photoperiod conditions ([Figure 4](#ijms-20-01974-f004){ref-type=\"fig\"}B).\n\nPrevious study has suggested that *foxo* was a key gene for diapause induction of the mosquito *Culex pipiens* \\[[@B8-ijms-20-01974]\\]. To verify the function of *foxo* on inducing locust diapause, ds*foxo* was injected into female adults of *L. migratoria* to knock down the *foxo* gene under long and short photoperiods, followed by confirming RNAi efficiency via qRT-PCR. It was observed that the mRNA level of *foxo* was significantly (*p* \\< 0.05) lower in treatment than in the control (inject ddH~2~O) under long and short photoperiods, which indicated acceptability of RNAi efficiency of *foxo* ([Figure 4](#ijms-20-01974-f004){ref-type=\"fig\"}C). However, under short photoperiod conditions, the average egg diapause rate (DR) was 97.4% in the negative control treatments, while it was significantly (*p* = 0.0173) lower in the ds*foxo*-treated insects (70.9%). In contrast, the diapause rate was not affected by ds*foxo* treatment under long photoperiod conditions. Thus, we suggest that RNAi maternal *foxo* could inhibit the offspring diapause under short photoperiod conditions ([Figure 4](#ijms-20-01974-f004){ref-type=\"fig\"}D).\n\n2.5. SOD and ROS Activity Changes after RNAi rai1 {#sec2dot5-ijms-20-01974}\n-------------------------------------------------\n\nPrevious study has shown that high ROS level could promote FOXO phosphorylation and subsequently induced diapause. SOD being a typical antioxidant enzyme shows resistance to oxidative damage, diapause and lifespan extension. The effect of ds*rai1* injection on SOD and ROS, and subsequent detection and measurement of SOD and ROS activities in fat body of ds*rai1*-treated and control locusts under both long and short photoperiods were verified. Results showed that SOD activity in RNAi *rai1* treatment was significantly (*p* = 0.0013) higher than control under a long photoperiod. Similarly, SOD activity in RNAi *rai1* treatment was also significantly (*p* = 0.0387) higher than control under a short photoperiod ([Figure 5](#ijms-20-01974-f005){ref-type=\"fig\"}A). Furthermore, SOD activity in a short photoperiod treated sample was significantly higher than that in a long photoperiod treated sample in both ds*rai1* treatment and control ([Figure 5](#ijms-20-01974-f005){ref-type=\"fig\"}B). ROS activities were 557.3 IU/g in RNAi *rai1* treatment, significantly (*p* = 0.0002) higher than control (349.9 IU/g) under a long photoperiod. Meanwhile, ROS activities were 553.8 U/g in RNAi *rai1* treatment, significantly (*p* = 0.0370) higher than control (492.7 U/g) under a short photoperiod ([Figure 5](#ijms-20-01974-f005){ref-type=\"fig\"}C). ROS activities in a short photoperiod treated sample were significantly higher than that in a long photoperiod treated sample in control. However, no significant difference for ROS activities between samples under long and short photoperiods in ds*rai1* treatments were observed ([Figure 5](#ijms-20-01974-f005){ref-type=\"fig\"}D). Results indicated that *rai1* negatively regulated SOD and ROS. SOD and ROS levels were higher under a short photoperiod than a long photoperiod.\n\n3. Discussion {#sec3-ijms-20-01974}\n=============\n\nPhotoperiod is considered as one of the most important environmental factors affecting insect diapause \\[[@B17-ijms-20-01974],[@B18-ijms-20-01974],[@B19-ijms-20-01974]\\]. The embryonic diapause is transgenerationally induced and controlled by the maternal parent in *Bombyx mori* \\[[@B20-ijms-20-01974]\\]. Our previous studies showed that female adults of *L. migratoria* could also be induced by a short photoperiod and subsequently produce diapause eggs \\[[@B13-ijms-20-01974],[@B21-ijms-20-01974]\\]. Here, fat bodies and ovaries of locust under long and short photoperiods were used for transcriptomic analysis. The correlated DEGs between OVA and FAT group were validated by qRT-PCR. In this study, we focused on finding the important maternal DEGs, which respond to diapause induction signals (photoperiod), and contribute to a better understanding of the molecular mechanisms underlying insect diapause induction.\n\n3.1. DEGs and Pathways Related to FOXO Signaling Pathway {#sec3dot1-ijms-20-01974}\n--------------------------------------------------------\n\nPhotoperiods are important cues for seasonal adaptations in insects. Although egg diapause is one of the seasonal adaptations, a photoperiod change is likely to affect many aspects of physiological status. FOXO signaling is involved in a wide variety of insect phenomena. Thus, FOXO signaling genes might be stimulated by photoperiods. FOXO's role in diapause induction has been noted in some other invertebrates, including nematodes \\[[@B22-ijms-20-01974]\\] and mosquitoes \\[[@B8-ijms-20-01974]\\]. FOXO activity is inhibited by growth factors that ultimately affect insulin signaling pathways, stimulated by nutrient depletion and a plethora of ROS-induced post-translational modifications \\[[@B23-ijms-20-01974]\\]. It has been observed that the insulin signaling and FOXO, a downstream molecule in the insulin signaling pathway, mediate the diapause response in *Culex pipiens* \\[[@B8-ijms-20-01974]\\]. In this study, DEGs were grouped into five clusters to find out their functions related to diapause induction under photoperiod regulation. These DEGs were also mapped to KEGG pathways to understand their role in up or down regulation. Results indicated that many of these DEGs (INSR, IRS, MnSOD, EGFR, etc.) and pathways (ko00010, ko00020, ko00190) were involved in the FOXO signaling pathway ([Table S3](#app1-ijms-20-01974){ref-type=\"app\"}).\n\nqRT-PCR results showed that relative expression of *INSR* and *IRS* was significantly lower in samples under a short photoperiod than in samples under a long photoperiod. Under a short photoperiod, down-regulation of *INSR* and *IRS* expression suggest that the insulin signaling pathway was possibly inhibited and as a result of that the FOXO was more active in locusts under a short photoperiod. In addition, *MnSOD* has already been identified as an important target downstream gene of FOXO in mice and nematodes \\[[@B24-ijms-20-01974],[@B25-ijms-20-01974]\\]. Expression of *MnSOD* (*sod2*) and SOD activity was significantly higher in samples under a short photoperiod than in samples under a long photoperiod. The up-regulation of *MnSOD* expression and SOD activity also implied that FOXO was activated during the diapause induction stage of *L. migratoria*. RNAi results demonstrated that knocking down the *foxo* gene of maternal locusts decreased the diapause incidence of offspring under a short photoperiod. This result is consistent with the previous findings that FOXO is active under short photoperiod conditions in mosquitoes \\[[@B8-ijms-20-01974]\\], and indicates that FOXO could likewise be important in the photoperiod-induced diapause process of *L. migratoria*.\n\nThe insect hormone biosynthesis pathway (ko00981) was also very important in diapause induction according to the transcriptome. Transcriptome analysis suggested up-regulation of ecdysteroid biosynthesis in a diapause-inducing group of *Bombyx mori* \\[[@B20-ijms-20-01974]\\]. Furthermore, Juvenile hormone and ecdysteroids were involved in egg diapause of *L. migratoria* \\[[@B26-ijms-20-01974],[@B27-ijms-20-01974]\\]. Gene such as *fpps* has a key role in synthesis of terpenoid backbone \\[[@B28-ijms-20-01974],[@B29-ijms-20-01974]\\]. Similarly, *sad* has been identified to be vital for ecdysone synthesis in *Drosophila melanogaster* \\[[@B30-ijms-20-01974]\\]. In our study the transcriptome analysis revealed the up-regulation of *fpps* and *sad* in a diapausing-induced group of *L. migratoria* as compared to a non-diapausing group. Meanwhile, *JHamt* is known for its critical and closely related role in JH biosynthesis, which is associated with insulin signaling pathways \\[[@B31-ijms-20-01974],[@B32-ijms-20-01974]\\]. *JHamt* is down-regulated in locusts under a short photoperiod, which means JH biosynthesis must have decreased. Studies have also revealed that the insulin promotes JH biosynthesis \\[[@B33-ijms-20-01974],[@B34-ijms-20-01974],[@B35-ijms-20-01974]\\]. The decrease in JH synthesis also indicated that the insulin signaling pathway has been inhibited and subsequently indicated the activation of FOXO ([Figure 6](#ijms-20-01974-f006){ref-type=\"fig\"}).\n\nAnother important feature of the FOXO signaling pathway (ko04068) was ROS. DEGs involved in glycolysis and TCA cycle, including *pgm*, *gapdh*, *adh*, *sdh1*, *mdh1*, were up-regulated in samples under a short photoperiod. In contrast, DEGs involved in oxidative phosphorylation and peroxisome, including *prx6*, *nd4*, *cox2*, were down-regulated in fat body samples under a short photoperiod. However, both qRT-PCR and transcriptomic results showed that *nd4* was up-regulated in ovary samples under a short photoperiod. This has clearly indicated almost all of the DEGs involved in glycolysis and TCA cycle induced ROS. ROS was reported as central regulator in inducing the diapause in *Helicoverpa armigera* \\[[@B36-ijms-20-01974]\\], while also keeping the FOXO active. ROS activities detection also showed that ROS level is higher in samples under a short photoperiod than a long photoperiod ([Figure 5](#ijms-20-01974-f005){ref-type=\"fig\"}D). This suggested that the ROS must keep a relatively higher level in maternal locust for diapause induction ([Figure 6](#ijms-20-01974-f006){ref-type=\"fig\"}).\n\nInterestingly, DEGs of ribosome were all up-regulated in samples under a short photoperiod. This indicates that ample supplies of ribosomal bodies are needed during diapause induction. These ribosomal reserves may even be transmitted directly to the eggs. In addition, there are other pathways associated with diapause induction, such as focal adhesion including *\u03b2-actin*, *flamin*, *tilin*, etc.; neurogenesis, including *pecanex*, *mrf* and *mbp*; synaptic vesicle cycle, including *SNAP25* and *syntaxin* ([Table S3](#app1-ijms-20-01974){ref-type=\"app\"}).\n\n3.2. Maternal rai1 Regulates Locust Diapause {#sec3dot2-ijms-20-01974}\n--------------------------------------------\n\nBoth transcriptome analysis and RT-PCR results showed that mRNA level of *rai1* was lower in samples under a short photoperiod than that of under a long photoperiod. This indicated that *rai1* expression was repressed by short photoperiod. In addition, egg diapause could have induced under short photoperiod. This suggested that the diapause induction in maternal locust was possibly negatively regulated by *rai1*. To confirm our hypothesis, RNAi was performed to identify the diapause regulation function of *rai1*. The result showed that the diapause incidence in *L. migratoria* increased after knocking down of the *rai1* gene ([Figure 4](#ijms-20-01974-f004){ref-type=\"fig\"}B). The RNAi result was consistent with the *rai1* expression profiles obtained through transcriptome analysis and RT-PCR. Hence, we concluded that the offspring diapause could be inhibited by a short photoperiod through repressing maternal *rai1* expression. For the first time in 1997, *rai1* was reported as a retinoic acid induced gene of acute promyelocytic leukemia cell \\[[@B37-ijms-20-01974]\\]. Vitamin A is a group of unsaturated nutritional organic compounds that includes retinol, retinal, retinoic acid, and several provitamin A carotenoids. These retinoid compositions were found in insect eyes \\[[@B38-ijms-20-01974]\\]. Photoperiodic induction of diapause in *Apanteles glomeratus* and *Amblyseius potentillae* has also been known to be Vitamin A dependent \\[[@B39-ijms-20-01974],[@B40-ijms-20-01974]\\]. Previous investigation into the role of *rai1* revealed that it has regulated circadian genes including *per*, *cry1*, etc. in mice. \\[[@B41-ijms-20-01974]\\]. In addition, circadian genes had an input module to transmit a light signal to the insulin signaling pathway and finally lead *C. pipiens* diapause \\[[@B33-ijms-20-01974]\\]. Circadian clock genes along with carbohydrate, lipid and energy metabolism are the known vital drivers for activating the environmental signals in summer diapause of *Delai antiqua* \\[[@B42-ijms-20-01974]\\]. Lipid metabolism can be seen distinctly in later phase of diapause however, understanding the difference in dormancy and diapausing, the lipid metabolism along with hormonal control clearly explained *Aedes albopictus* (Asian tiger mosquito) development through transcriptome analysis \\[[@B43-ijms-20-01974]\\]. Due to the complex nature of diapause, a number of genes and processes regulate the diapausing in *Drosophila montana* including metabolism and fatty acid biosynthesis, which assist diapausing females to survive overwintering \\[[@B44-ijms-20-01974]\\]. So, we speculate that photoperiodic signals are transmitted to maternal *rai1* via retinoic acid and eventually to the circadian genes and downstream insulin signaling pathway, which leads to offspring's egg diapause ([Figure 6](#ijms-20-01974-f006){ref-type=\"fig\"}) which supports the previously reported findings in *Delia antiqua* \\[[@B42-ijms-20-01974]\\]. Similarly, a high ROS level induced pupal diapause of *Helicoverpa armigera* and extended the lifespan of insects \\[[@B45-ijms-20-01974],[@B46-ijms-20-01974]\\]. *Sod-2*, a downstream gene of FOXO, was dramatically higher in diapausing females than in non-diapausing females of mosquito *C. pipiens* \\[[@B47-ijms-20-01974]\\]. Results showed that ROS and SOD activities were higher in ds*rai1* treatment than in control. The up-regulation of SOD activity implied that FOXO is being activated after RNAi *rai1*. These results suggested that RAI1 might inhibit diapause induction by repressing ROS and FOXO activities. On the other hand, retinoic acid has also been found to be critical for neurogenesis of *L. migratoria* \\[[@B48-ijms-20-01974]\\]. Results in the present study suggested that other factors might inhibit induction of *L. migratoria* diapause under a long photoperiod. Only knocking down ds*rai1* might not enough to cut-off the insulin signaling and subsequently regulates the FOXO activity under a long photoperiod. This could be the main reason why egg diapause enhancement by knocking down *rai1* was confirmed only under a short photoperiod, whereas it was not observed under a long photoperiod.\n\nEgg diapause enhancement through ds*rai1* treatments was confirmed under short photoperiod. Due to the complex nature of diapause regulation, the *rai1* can regulate diapause; however, this may not be the final decisive role of *rai1*. Results showed that knocking down *rai1* could only increase the diapause rate under short photoperiod; although, a 100% increase in diapause is not determined by the study undertaken. In mosquitoes, the phosphorylated FOXO was blocked to translocate into the nucleus by insulin signaling pathway under long-day photoperiod \\[[@B8-ijms-20-01974]\\]. We presumed that insulin signaling pathway strongly arrested FOXO activity. Only knocking down ds*rai1* might not enough to cut-off the insulin signaling and subsequently regulates the FOXO activity under a long photoperiod. This could be the main reason why knocking down *rai1* cannot promote egg diapause of *L. migratoria* under a long photoperiod.\n\nIn the present study, dozens of DEGs were obtained in different tissues of maternal *L. migratoria* induced by different photoperiods. Most of these DEGs, including *INSR*, *IRS*, *MnSOD*, *EGFR*, were closely related to the FOXO signaling pathway. Among these DEGs, the *rai1* was selected and confirmed to be involved in diapause induction of maternal locust under a short photoperiod. Because of the special relationship between retinoic acid and light, we are of the opinion that the photoperiod was likely to affect the FOXO signaling pathway through the maternal *rai1*, and ultimately established environmental signals, transmitted to the next generation to induce locust egg diapause. Future researchers can work in the area to explore other factors and genes that can promote diapause induction under a long photoperiod, and can highlight the specific mechanism of related DEGs regulating diapause through the FOXO signaling pathway.\n\n4. Materials and Methods {#sec4-ijms-20-01974}\n========================\n\n4.1. Insect Rearing and Tissue Collection {#sec4dot1-ijms-20-01974}\n-----------------------------------------\n\nThe *L. migratoria* L. colony used in this study was originally collected from the field at Tianjin, China (38\u00b049\u2032 N, 117\u00b018\u2032 E) in November 2007 and was maintained by the State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chines Academy of Agricultural Sciences. Third instar locusts were collected from the rearing cages and transferred to 20 cm \u00d7 20 cm \u00d7 28 cm mesh cages in artificial climate chambers (PRX-250B-30, Haishu Saifu Experimental Instrument Factory, Ningbo, China) where the insects were raised to adults at either 27 \u00b0C and 60% RH under a long photoperiod of 16:8 L:D to produce non-diapause eggs, and at 27 \u00b0C, 60% RH under a short photoperiod of 10:14 L:D to produce diapause eggs. When *L. migratoria* matured after 72 h under both long and short photoperiods, ovaries and fat bodies were separately collected from female individuals. Ovaries were dissected at the stage of developmental phase-II. In phase II, during the process of vitellogenesis, the color of the ovaries turned to light yellow from white, had an ovary size of around 10 mm \u00d7 5 mm (length \u00d7 width), whereas the fallopian tube shifted from thin to round. Tissues were dissected into the ice-cold RNase-free saline solution. There were four tissues and three replicates in each tissue. All of the samples were immediately frozen by liquid nitrogen and then kept at \u221280 \u00b0C until RNA extraction. The *L. migratoria* L. colony used in this study has been maintained by the State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chines Academy of Agricultural Sciences. The locusts are common agricultural pests and are not included in the \"List of Endangered and Protected Animals in China\".\n\n4.2. RNA Extraction and RNA-Seq {#sec4dot2-ijms-20-01974}\n-------------------------------\n\nA total of twelve *L. migratoria* tissues were used whereas RNA extraction, library construction and sequencing were carried out as described in previously reported study \\[[@B21-ijms-20-01974]\\]. Total RNA was isolated from frozen medullar tissue by using the RNA plant mini kit with column DNase digestion (Qiagen, Hilden, Germany) following the manufacturer's instructions. RNA concentration was then measured using Qubit RNA Assay Kit in Qubit 2.0 Flurometer (Life Technologies, Carlsbad, CA, USA). Additionally, RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, USA). All of the 12 samples with RNA integrity numbers (RIN) above 8 were used for construction of the libraries. Sequencing libraries were generated using NEBNext Ultra\u2122 RNA Library Prep Kit for Illumina (NEB, Illumina, San Diego, CA, USA) following manufacturer's recommendations, whereas index codes were added to attribute sequences to each sample. Finally, PCR products were purified (AMPure XP system, Beckman Coulter Life Sciences, Indianapolis, IA, USA) and library quality was assessed on the Agilent Bioanalyzer 2100 system. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq 2500 platform and 125 bp paired-end reads were generated.\n\n4.3. Sequence Assembly, Annotation and DEGs Analysis {#sec4dot3-ijms-20-01974}\n----------------------------------------------------\n\nSequencing data of ovary and fat body samples were analyzed independently. Raw data were processed into clean data by eliminating adapters, poly-N and inferior quality reads. Transcriptome assembly was achieved based on the clean data, using Trinity with min_kmer_cov set to default value of 2, whereas all other parameters were set-up to default \\[[@B49-ijms-20-01974]\\]. Gene function was annotated based on the following seven databases: Nr (NCBI non-redundant protein sequences), Nt (NCBI non-redundant nucleotide sequences), Pfam (Protein family), KOG/COG (Clusters of Orthologous Groups of proteins), Swiss-Prot (a manually annotated and reviewed protein sequence database), KO (KEGG Ortholog database) and GO (Gene Ontology). Data for each sequenced library was analyzed using BLAST with a cut-off *e*-value of 10^\u22125^. Prior to differential gene expression analysis, for each sequenced library, the read counts were then adjusted by the edge R program package through one scaling normalized factor. Differential expression analysis of two samples were performed using the DEGseq (2010) R package (1.10.1) \\[[@B50-ijms-20-01974]\\]. The *p*-value was adjusted using *q*-value \\[[@B51-ijms-20-01974]\\] whereas the *q*-value \\< 0.05 and \\|log2 (fold change)\\| \\> 1 was specified as the threshold for significantly differential expression.\n\n4.4. cDNA Synthesis and qRT-PCR {#sec4dot4-ijms-20-01974}\n-------------------------------\n\nTotal RNAs of L_FAT (L_FAT1, L_FAT2 and L_FAT3 mixed together), S_FAT (S_FAT1, S_FAT2 and S_FAT3 mixed together), L_OVA (L_OVA1, L_OVA2 and L_OVA3 mixed together) and S_OVA (S_OVA1, S_OVA2 and S_OVA3 mixed together) were respectively used to move along with further research. cDNA was synthesized from the four mixed RNA samples using M-MLV reverse transcriptase and recombinant RNase inhibitor (Takara, Beijing, China). The expression levels of 24 DEGs in all four tissues were determined by qRT-PCR using SYBR Premix Ex Taq kit (Takara) as per the manufacturer's instructions in an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). qRT-PCR was performed as per following conditions: 95 \u00b0C for 10 min; 40 cycles of 95 \u00b0C for 15 s, 60 \u00b0C for 45 s. Gene expression was quantified using 2^\u2212\u0394\u0394*C*t^ method \\[[@B52-ijms-20-01974]\\], with elongation factor 1 (*ef-1*) as the internal control for normalization of data. The specific primers used for qRT-PCR are listed in [Table S4](#app1-ijms-20-01974){ref-type=\"app\"}.\n\n4.5. RNA Interference {#sec4dot5-ijms-20-01974}\n---------------------\n\nThe dsRNA was generated by in vitro transcription using the T7 RiboMAX system (Promega, Fitchburg, WI, USA) as per prescribed manufacturer's protocol. Templates for in vitro transcription reactions were prepared by PCR amplification from plasmid DNA of the cDNA clone of *rai1* and *foxo* using the primer pairs with T7 polymerase promoter sequence at 5\u2032-end ([Table S4](#app1-ijms-20-01974){ref-type=\"app\"}). The length of ds*rai1* was 583 bp, whereas the length of ds*foxo* was 194 bp. A Total of 5 \u03bcL of dsRNAs (2 \u03bcg/\u03bcL) for the target gene (*rai1* or *foxo*), water as controls were used to inject into the ventral part between 2nd and 3rd abdominal segments of the female adults within 72 h after molting. For each gene, 75 female adults were injected and divided into three groups. The effects of RNAi on the mRNA levels were investigated by qRT-PCR at 48 h after injection. To keep track of the transcript levels of *rai1* and *foxo*, total RNA was extracted from entire bodies of locusts. For each target gene, three individuals from each group were used for RNA extraction.\n\n4.6. Diapause Rate Detection {#sec4dot6-ijms-20-01974}\n----------------------------\n\nLocusts were placed in new mesh cages (20 cm \u00d7 20 cm \u00d7 28 cm) and provided with bouquets of greenhouse grown wheat after applying the above-mentioned treatments. Meanwhile, 25 adult males were introduced to each replicate. The floor of the cages was covered in a 2 cm layer of sieved sterile sand with the cages maintained until eggs laying by the locusts. Once oviposition was observed, eggs were collected at an interval of 48 h for 10 days using a camel paint brush and transferred into plastic Petri dishes (90 mm \u00d7 50 mm), where the eggs were incubated on vermiculite, before shifting to 27 \u00b0C and 60% RH to slow down the development. Around 150 eggs were obtained from 2--3 pods, which were then used in each experimental replication. Eggs were kept under 27 \u00b0C for 30 days until eclosion of 1st instar nymphs ceased (H1). To account for non-viable eggs, all remaining un-eclosed eggs were kept at 4 \u00b0C for 60 days to receive ample time to break the diapause, afterwards they were incubated at 27 \u00b0C for 30 days and for any further 1st instar emergence recording (H2). The diapause rate (DR) was calculated as: DR (%) = H2/(H1 + H2).\n\n4.7. SOD and ROS Activity Detection {#sec4dot7-ijms-20-01974}\n-----------------------------------\n\nRapid ELISA-based quantification was used to monitor and measure the SOD and ROS activities in the fat bodies of *L. migtatoria* using the specified manufacturer instructions for catalogue SU-B97128; SU-B97124; SU-B97136; SU-B97141 (Collodi Biotechnology Co., Ltd., Quanzhou, China). The fat body samples were homogenized in a 1 mL PBS, and the resulting suspension was subjected to ultra-sonication to further disrupt the cell membranes. After homogenates were centrifuged for 15 min at 2348\u00d7 *g* (5000 rpm), the supernatants were collected and were stored at \u221220 \u00b0C until used for further analysis. All of the required reagents and samples were prepared and were properly maintained at room temperature (18--25 \u00b0C) for 30 min prior initiating further assay procedure ([Table S5](#app1-ijms-20-01974){ref-type=\"app\"}). We set-up standard wells, sample wells and blank (control) wells. We then added 50 \u03bcL standard to each standard well, 50 \u03bcL of sample to each sample well and 50 \u03bcL of sample diluent to each blank/control well. A 100 \u03bcL amount of HRP-conjugate reagent was added to each well and covered with an adhesive strip and incubated for 60 min at 37 \u00b0C. The Micro titer plates were rinsed using Wash Buffer (1\u00d7) 4-times followed by adding gently mixed Chromogen Solution A (50 \u03bcL) and Chromogen Solution B (50 \u03bcL) to each well in succession, protected from light and incubated for 15 min at 37\u00b0C. Finally, 50 \u03bcL Stop Solution was added to each well. During the process, well color changed from blue to yellow, showed a proper sign and confirmation of uniformity. Colorless or green color is usually a sign of no uniformity. In such case, the plate was then gently tapped to ensure thorough mixing. The Optical Density (OD) at 450 nm was read using a Micro Elisa Strip plate reader (Multiskan\u2122 FC 51119000, Thermo Fisher Scientific Inc., Waltham, MA, USA) within 15 min of adding the Stop Solution. Standard curves of SOD and ROS were constructed respectively and calculated accordingly to quantify the amount of SOD and ROS of each sample.\n\n4.8. Statistical Analysis {#sec4dot8-ijms-20-01974}\n-------------------------\n\nThe differences between treatments were compared by Student's *t*-test. Differences were considered significant at *p* \\< 0.05. Values were reported as mean \u00b1 SE. Data were analyzed using the SPSS software (version 15.0; SPSS Inc., Chicago, IL, USA) and GraphPad Prism software (version 6.01; GraphPad Software Inc., San Diego, CA, USA).\n\n4.9. Availability of Data and Materials {#sec4dot9-ijms-20-01974}\n---------------------------------------\n\nAll RNA-seq.fastq files are available from the NCBI Sequence Read Archive database (accession number SRR5762723, SRR5762724, SRR5762725, SRR5762726, SRR5762727, SRR5762728, SRR5754272, SRR5754273, SRR5754274, SRR5754275, SRR5754276, SRR5754277, SRR5754278). All other data is contained in the manuscript and in the [Supplementary Materials](#app1-ijms-20-01974){ref-type=\"app\"}.\n\nWe thank Guangjun Wang at CAAS for technical assistance. We also thank Shan Xue for the figure editing. We appreciate the efforts of Mark R. McNeill, AgResearch, Canterbury Agriculture and Science Centre, New Zealand for English editing and proof reading.\n\nSupplementary materials can be found at <https://www.mdpi.com/1422-0067/20/8/1974/s1>.\n\n###### \n\nClick here for additional data file.\n\nConceptualization, K.H. and Z.Z.; Formal analysis, K.H.; Investigation, K.H. and A.R.J.; Methodology, K.H., A.R.J. and Z.Z.; Project administration, Z.Z.; Resources, X.T. and X.N.; Writing---original draft, K.H. and A.R.J.; Writing---review & editing, K.H. and H.U.\n\nThe China Agriculture Research System (CARS-34-07B), National Natural Science Foundation of China under Grant (No. 61661136004), of P.R. China, and the STFC Newton Agritech Programme under Grant (No. ST/N006712/1) of UK funded this research.\n\nThe authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.\n\nFAT\n\nfat body\n\nOVA\n\novary\n\nL\n\nlong photoperiod\n\nS\n\nshort photperiod\n\nDEGs\n\ndifferentially expressed genes\n\nvs\n\nversus\n\nFDR\n\nfalse discovery rate\n\nBLAST\n\nbasic local alignment search tool\n\nGO\n\nGene Ontology\n\nKEGG\n\nKyoto Encyclopedia of Genes and Genomes\n\nKO\n\nKEGG Ortholog database\n\nKOG/COG\n\nClusters of Orthologous Groups of proteins\n\nNR\n\nNCBI non-redundant protein sequences\n\nNT\n\nNCBI nucleotide sequences\n\nPfam\n\nprotein families\n\nSwiss-Prot\n\na manually annotated and reviewed protein sequence database\n\nTCA cycle\n\ntricarboxylic acid cycle\n\nfpps\n\nfarnesyl diphosphate synthase\n\nJHamt\n\njuvenile hormone acid\n\nO\n\n-methyltransferase\n\nSOD\n\nsuperoxide dismutase\n\nPGM\n\nphosphoglucomutase\n\nGAPDH\n\nglyceraldehyde 3-phosphate dehydrogenase\n\nADH\n\nalcohol dehydrogenase\n\nMDH\n\nmalate dehydrogenase\n\nINSR\n\ninsulin receptor\n\nIRS\n\ninsulin receptor substrate\n\nEGFR\n\nendothelial growth factor receptor\n\nCOX\n\ncytochrome c oxidase\n\nND\n\nNADH dehydrogenase\n\nDR\n\ndiapause rate\n\nfoxo\n\nfoxo\n\ngene\n\nFOXO\n\nFOXO protein\n\n![DEGs between long and short photoperiods treatments in *L. migratoria* fat body (left) and ovary (right) samples. DEGSeq (2010) R package (1.10.1) was used to carry out the differential expression analysis in digital gene expression and determining the expression via model based negative binomial distribution. Resulting *p* values were adjusted using the Benjamini and Hochberg's approach for controlling the false discovery rate. Genes with an adjusted *p*-value of \\<0.05 explained by DEGSeq were assigned as differentially expressed. The *x*-axis represents the change of gene expression in different groups; whereas the *y*-axis represents the statistical significance of gene expression change. \u2212log10(padj) means \u2212log10 (adjusted *p*-value). The smaller the adjusted *p*-value in \u2212log10(padj), the greater the difference will be (significant). Blue dots in the figure represent the genes with no significant difference; red dots represent the up-regulated genes with significant difference, whereas green dots represent the down-regulated genes with significant difference.](ijms-20-01974-g001){#ijms-20-01974-f001}\n\n![Heatmap of DEGs across the four treatments. Each line in the figure represents a gene, with the columns representing S_FAT, L_FAT, S_OVA and L_OVA. Red indicates relatively high expression and blue indicates relatively low expression. All of the DEGs fall in the range of \u00b15.0. The spectrum of color from red to blue, indicating that log10 (FPKM + 1) has flow from large to small.](ijms-20-01974-g002){#ijms-20-01974-f002}\n\n![qRT-PCR validations of 24 DEGs in FAT and OVA groups. (**A**) Relative mRNA level of 24 DEGs in FAT group. (**B**) Relative mRNA level of 24 DEGs in OVA group. (**C**) FPKM value of 24 DEGs in FAT group. (**D**) FPKM value of 24 DEGs in OVA group. For each DEG, three technical replications were performed; *ef-1* gene was used as internal control. For each treatment, three replications were performed. Values were reported as mean \u00b1 SE. \\* Indicates an error probability of *p* \\< 0.05 using Student's *t*-test.](ijms-20-01974-g003){#ijms-20-01974-f003}\n\n![(**A**) RNAi efficiency verified in the whole body of adult females after injecting dsRNA of *rai1* under both long and short photoperiods. (**B**) Diapause rate detected after injecting dsRNA of *rai1* under both long and short photoperiods. (**C**) RNAi efficiency verified in the whole body of adult females after injecting dsRNA of *foxo* under both long and short photoperiods. (**D**) Diapause rate detected after injecting dsRNA of *foxo* under both long and short photoperiods. Numbers of replications were set as three for each treatment. Values were reported as mean \u00b1 SE. \\* Indicates an error probability of *p* \\< 0.05 using Student's *t*-test.](ijms-20-01974-g004){#ijms-20-01974-f004}\n\n![(**A**,**B**) SOD activity detected after injecting dsRNA of *rai1* to locusts under both long and short photoperiods. (**C**,**D**) ROS activity detected after injecting dsRNA of *rai1* to locusts under both long and short photoperiods. For each treatment, three replications were performed. Values were reported as mean \u00b1 SE. \\* Indicates an error probability of *p* \\< 0.05 using Student's *t*-test.](ijms-20-01974-g005){#ijms-20-01974-f005}\n\n![Potential pathway of maternal effect on egg diapause induction of *L. migratoria*. The 'blue' mean 'down-regulated', 'red' means 'up-regulated' whereas 'black' means 'not confirmed in our experiment'.](ijms-20-01974-g006){#ijms-20-01974-f006}\n\nijms-20-01974-t001_Table 1\n\n###### \n\nSummary of RNA-seq metrics from *L. migratoria* transcriptomes for ovaries and fat bodies under both long and short photoperiods.\n\n  Sample   Clean Reads   Clean Bases (G)   Q20 (%)   Number of Transcripts   Number of Unigenes\n  -------- ------------- ----------------- --------- ----------------------- --------------------\n  L_FAT1   78872568      11.83             95.61     260,779                 102,273\n  L_FAT2   79227948      11.88             95.62                             \n  L_FAT3   88069778      13.21             95.86                             \n  S_FAT1   77266106      11.59             96.06                             \n  S_FAT2   77647066      11.65             96.78                             \n  S_FAT3   76502530      11.48             95.71                             \n  L_OVA1   80389650      12.06             95.58     323,527                 132,147\n  L_OVA2   90730566      13.61             95.84                             \n  L_OVA3   71022586      10.65             95.30                             \n  S_OVA1   75628016      11.34             95.92                             \n  S_OVA2   88328188      13.25             96.35                             \n  S_OVA3   76316318      11.45             94.97                             \n\nijms-20-01974-t002_Table 2\n\n###### \n\nDEGs of different groups enriched KEGG pathways.\n\n  Group                       Num.                                                  Term                                          ID              *p*-Value\n  --------------------------- ----------------------------------------------------- --------------------------------------------- --------------- ----------------\n  S_FAT vs. L_FAT up DEGs     1                                                     Ribosome                                      ko03010         1.62 \u00d7 10^\u22129^\n  2                           Citrate cycle (TCA cycle)                             ko00020                                       0.003078        \n  3                           Regulation of actin cytoskeleton                      ko04810                                       0.012542        \n  4                           Glyoxylate and dicarboxylate metabolism               ko00630                                       0.015929        \n  5                           Glycosphingolipid biosynthesis                        ko00603                                       0.019268        \n  6                           Protein processing in endoplasmic reticulum           ko04141                                       0.02064         \n  7                           Endocytosis                                           ko04144                                       0.021813        \n  8                           Glycolysis / Gluconeogenesis                          ko00010                                       0.022633        \n  9                           Cysteine and methionine metabolism                    ko00270                                       0.029743        \n  10                          Fc \u03b3 R-mediated phagocytosis                          ko04666                                       0.029743        \n  S_FAT vs. L_FAT down DEGs   1                                                     Glycine, serine and threonine metabolism      ko00260         2.21 \u00d7 10^\u22125^\n  2                           Oxidative phosphorylation                             ko00190                                       0.000119        \n  3                           Glyoxylate and dicarboxylate metabolism               ko00630                                       0.000133        \n  4                           Pentose phosphate pathway                             ko00030                                       0.000141        \n  5                           Phenylalanine, tyrosine and tryptophan biosynthesis   ko00400                                       0.002921        \n  6                           Fructose and mannose metabolism                       ko00051                                       0.006848        \n  7                           Phenylalanine metabolism                              ko00360                                       0.013337        \n  8                           Alanine, aspartate and glutamate metabolism           ko00250                                       0.019824        \n  9                           Peroxisome                                            ko04146                                       0.036338        \n  10                          Arginine biosynthesis                                 ko00220                                       0.04393         \n  S_OVA vs. L_OVA up DEGs     1                                                     Protein processing in endoplasmic reticulum   ko04141         3.86 \u00d7 10^\u221211^\n  2                           Proteasome                                            ko03050                                       7.15 \u00d7 10^\u22127^   \n  3                           Spliceosome                                           ko03040                                       8.36 \u00d7 10^\u22125^   \n  4                           Cell cycle                                            ko04110                                       0.000107        \n  5                           Protein export                                        ko03060                                       0.000334        \n  6                           RNA transport                                         ko03013                                       0.000422        \n  7                           Antigen processing and presentation                   ko04612                                       0.000672        \n  8                           Ubiquitin mediated proteolysis                        ko04120                                       0.001973        \n  9                           Pyrimidine metabolism                                 ko00240                                       0.009155        \n  10                          NOD-like receptor signaling pathway                   ko04621                                       0.010726        \n  11                          DNA replication                                       ko03030                                       0.013664        \n  12                          Toll-like receptor signaling pathway                  ko04620                                       0.019418        \n  13                          RNA degradation                                       ko03018                                       0.021142        \n  14                          Oocyte meiosis                                        ko04114                                       0.021603        \n  15                          Ribosome                                              ko03010                                       0.027409        \n  16                          Cytosolic DNA-sensing pathway                         ko04623                                       0.033442        \n  17                          Progesterone-mediated oocyte maturation               ko04914                                       0.034639        \n  18                          NF-kappa B signaling pathway                          ko04064                                       0.042894        \n  19                          N-Glycan biosynthesis                                 ko00510                                       0.048725        \n  S_OVA vs. L_OVA down DEGs   1                                                     Thyroid hormone signaling pathway             ko04919         0.000812\n  2                           Focal adhesion                                        ko04510                                       0.000986        \n  3                           Hippo signaling pathway - fly                         ko04391                                       0.00245         \n  4                           Notch signaling pathway                               ko04330                                       0.009375        \n  5                           Tight junction                                        ko04530                                       0.01335         \n  6                           Dorso-ventral axis formation                          ko04320                                       0.018136        \n  7                           Cardiac muscle contraction                            ko04260                                       0.021591        \n  8                           ECM-receptor interaction                              ko04512                                       0.027292        \n  9                           Longevity regulating pathway                          ko04213                                       0.034335        \n  10                          cGMP-PKG signaling pathway                            ko04022                                       0.042573        \n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction {#s1}\n============\n\nThirty years after the discovery of HIV/AIDS, the search for a safe and effective vaccine has intensified, as a number of promising candidate vaccines progressing to phase IIb/III clinical trials have failed to show efficacy. One of the greatest barriers to HIV vaccine development is the enormous virion diversity (depicted by the existence of numerous clades and subtypes in distinct geographic demarcations) and the continuous evolution which generates numerous quasi-species within an infected individual (Hemelaar et al., [@B125]). This not only makes it challenging to create immunogens which are effectively matched to the circulating target viruses, but also provides room for immune escape of HIV from potent vaccine-induced immune responses. Therefore, it has emerged that immunogens derived from the most conserved regions of HIV and covering multiple variants (conserved mosaics) stand out as the most suitable candidates for T-cell based vaccines, while immunogens covering the most potent and broadly neutralizing and non-neutralizing antibody epitopes are better for antibody-based vaccines (Emini and Koff, [@B75]; Robinson and Amara, [@B222]; McMichael, [@B183]; Letourneau et al., [@B160]; Thorner and Barouch, [@B263]; Sekaly, [@B241]; Korber et al., [@B155]; Barouch et al., [@B22]; Santra et al., [@B231]; Borthwick et al., [@B38]). However, the development of a vaccine based on conserved antibody epitopes to provide protective global coverage and to minimize immune escape is hampered by inaccessibility of the highly shielded conserved envelope domains. Furthermore, the observation that development of broadly neutralizing antibodies requires prolonged stimulation with higher antigenic loads from divergent virus species (van Gils and Sanders, [@B267]) implies that HIV vaccine strategies must provide a continuous high level expression of a cocktail of immunogens. Although the use of polyvalent T-cell and B-cell mosaic constructs or the conserved consensus sequences may effectively overcome the challenges of HIV diversity and significantly improve vaccine efficacy (Santra et al., [@B231], [@B232]), the lack of clearly defined correlates of efficacy means that it remains unclear what immune responses an HIV vaccine should aim to induce. Recently, a non-human primate (NHP) study based on the RhCMV vector induced exceptionally broad and persistent atypical CD8+ T cells which effectively cleared SIV and maintained durable suppression of virus replication (Hansen et al., [@B118], [@B116], [@B117]), suggesting that HIV vaccine development research may have to adapt immunogen design and delivery strategies that stimulate similar responses.\n\nDelivery vectors are vital and integral components of a successful vaccine as they play an important role in modulating both innate and adaptive immunity. Therefore, vaccine vectors can significantly influence the magnitude and breadth, as well as the phenotypic and functional qualities of vaccine-induced immune responses. Moreover, as the type of delivery vector, in conjunction with the route of vaccine administration often determine whether or not vaccine-specific immune responses persist within the systemic and/or mucosal compartments (Masopust et al., [@B174]; Kiyono and Fukuyama, [@B152]; Ranasinghe et al., [@B212]; Czerkinsky and Holmgren, [@B62]), vector choice remains a critical determinant of the overall efficacy of any given vaccine. A part from the immunostimulatory potential to induce strong and persistent immunity, several other factors such as stability and ease of large scale manufacturing, safety, capacity for transgene insertion and pre-existing immunity also influence vector choice. It is now well-documented that pre-existing anti-vector immunity (especially neutralizing antibodies) can prevent transduction and/or expression of vaccine transgenes thus reducing vaccine-specific immune induction (Xiang et al., [@B284]; Fitzgerald et al., [@B82]; Lasaro and Ertl, [@B158]). This is a common phenomenon, clearly demonstrated with certain vectors which show superior immunogenicity in animal models yet induce only modest immune responses due to neutralization by pre-existing antibodies in humans (McCoy et al., [@B180]). Additionally, pre-existing immunity can alter the natural course of infection leading to catastrophic consequences such as enhanced HIV acquisition and possibly accelerated disease progression (Buchbinder et al., [@B45]; McElrath et al., [@B181]). Thus strategies that concurrently maximize vaccine immunogenicity while minimizing safety concerns remain an urgent priority in the development of a safe and efficacious vaccine for HIV/AIDS.\n\nA good number of HIV vaccine candidates (both prophylactic and therapeutic) employing a broad range of vaccine delivery vectors have been tested and some have progressed to evaluation of potential efficacy in phase IIb/III trials. Of significant relevance as far as safety is the STEP trial that used human adenovirus serotype 5 (Ad5) to deliver a well-designed HIV immunogen expressing Gag/Pol/Nef, which was associated with increased risk of HIV acquisition in uncircumcised male vaccinees with pre-existing immunity to Ad5 (Buchbinder et al., [@B45]; McElrath et al., [@B181]). This unexpected and rather worrisome finding prompted the premature halting of two related efficacy trials due to futility (Gray et al., [@B110]; Hammer et al., [@B115]). As disappointing as this might have been at the time, invaluable lessons have been learned and there is still great optimism as these lessons are now taken on board. Focussing on some of the promising HIV vaccine candidates in preclinical and clinical development, this review discusses pertinent issues relating to safety and immunogenicity of replicating and non-replicating viral vectors, pre-existing anti-vector immunity and how these can potentially influence the natural history of HIV infection and progression. In particular, this article highlights the safety profiles, immuno-stimulatory potential and possible limitations of plasmid DNA, MVA (modified vaccinia virus Ankara), ALVAC (canarypox virus), NYVAC (New York attenuated vaccinia virus), influenza virus and adenovirus vectored vaccines in preclinical and clinical studies for HIV vaccines. Some of the delivery vectors evaluated in clinical studies are summarized in Table [1](#T1){ref-type=\"table\"}, while those in preclinical development are summarized in Table [2](#T2){ref-type=\"table\"}.\n\n###### \n\n**Representative clinical studies**.\n\n  **Study name and phase**                                 **Immunogen**                                                                         **Vectors, regimen and route of immunization**                                            **Immune responses generated**                                    **References**\n  -------------------------------------------------------- ------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------- ----------------------------------------------------------------- -----------------------------------------------------------\n  **(I) HETEROLOGOUS PRIME-BOOST STUDIES**                                                                                                                                                                                                                                                                   \n  HIVCORE002 (Phase I study)                               HIVconsv (T cell immunogen based on conserved regions)                                ChAdV63/MVA (i.m.)                                                                        -CD4+ and CD8+ T cells -*In vitro* virus inhibition               Borthwick et al., [@B38]\n  DNA/ChAdV63/MVA (i.m.)                                                                                                                                                                                                                                                                                     \n  DNA/MVA/ChAdV63 (i.m.)                                                                                                                                                                                                                                                                                     \n  HVTN 505 (Phase IIb study)                               VRC-HIVDNA016-00-VP/VRC-HIVADV014-00-VP                                               DNA-prime (i.m. biojector device)/rAd5 boost (i.m. needle and syringe)                    -T cells and gp140 binding IgG antibodies                         Hammer et al., [@B115]\n  HVTN 503/Phambili (Phase IIb study)                      MRKAd5 HIV-1 Gag/Pol/Nef                                                              DNA-prime (i.m.)/Ad5 boost (i.m.)                                                         -CD8+ and CD4+ T cells                                            Gray et al., [@B110]\n  Phase 1 study                                            Gag and Env DNA and recombinant trimeric Env glycoprotein                             DNA-prime (i.m.)/Protein boost with MF59 adjuvant                                         -Robust B and T cells                                             Spearman et al., [@B251]\n  -Strong NAbs to SF162                                                                                                                                                                                                                                                                                      \n  -ADCC and neutralization of tier 2 strains                                                                                                                                                                                                                                                                 \n  Phase I/II study                                         Multi-clade, multigene: DNA/HIV-1 gp160, p17/p24 Gag and MVA/HIV-1 Gag/Pol            Low dose (i.d.) DNA-prime (x3)/MVA-boost (i.m. x2)                                        -High magnitude and broad CD4+ and CD8+ T cell responses          Bakari et al., [@B13]\n                                                                                                                                                 (DDDMM)                                                                                   -Env antibodies                                                   \n  Phase I study DP6-001                                    Multigene polyvalent gp120 and Gag DNA and polyvalent gp120 protein                   i.m. or i.d. Polyvalent DNA-prime/i.m. protein-boost (with QS21 adjuvant)                 -High titer binding and BNAbs, ADCC and multifunctional T cells   Bansal et al., [@B17]; Vaine et al., [@B265]\n  RV144 (Phase III study)                                  ALVAC-HIV vCP1521/AIDSVAX gp120 B/E                                                   ALVAC-prime (i.m.)/gp120 protein-boost                                                    -T cells and non-neutralizing antibodies to V1/V2 loop            Rerks-Ngarm et al., [@B215]\n  Phase I study                                            Multigenic HIV DNA (gp160- A/B/C; Rev B, Gag A/B and RT- B and HIV-MVA Env/Gag/Pol)   DNA- prime (i.d. with Biojector)/MVA-boost (i.d./i.m.); with or without GM-CSF adjuvant   -Broad and potent cellular immune responses                       Sandstrom et al., [@B230]; Gudmundsdotter et al., [@B112]\n  HVTN 502/STEP Study (Phase IIb)                          MRKAd5 HIV-1 Gag/Pol/Nef                                                              DNA-prime (i.m.)/Ad5 boost (i.m.)                                                         -Strong CD8+ T cell responses                                     Buchbinder et al., [@B45]; McElrath et al., [@B181]\n  Phase 1 study                                            HIVA (HIV-1 clade A and a CTL epitope string)                                         DNA-prime (i.m.)/MVA-boost (i.m.)                                                         -Multifunctional CD4+ and CD8+ T cells                            Mwau et al., [@B193]; Goonetilleke et al., [@B107]\n  Phase I study (EuroVacc: EV02)                           HIV-1 clade C-Env/Gag/Pol/Nef (DNA-C and NYVAC-C)                                     DNA-prime (i.m.)/NYVAC- boost (i.m.)                                                      -Durable, broad and poly-functional CD4+ and CD8+ T cells         Harari et al., [@B119]; McCormack et al., [@B179]\n  Phase I study                                            ALVAC-HIV(vCP300)                                                                     ALVAC-prime (i.m.)/i.m. Protein-boost (with MF59 adjuvant)                                -Durable CTLs                                                     Evans et al., [@B78]\n                                                           gp120/gp41, Gag, Pro, Nef, Pol and SF-2 rgp120                                                                                                                                  -Antibody responses                                               \n  Phase I study                                            ALVAC-HIV(vCP205)                                                                     ALVAC-prime (i.m.)/i.m. Protein-boost (with MF59 adjuvant)                                -Strong CD8+ T cell responses and NAbs                            Belshe et al., [@B26]\n                                                           gp120/gp41, Gag, Pol and SF-2 rgp120                                                                                                                                                                                                              \n  **(II) HOMOLOGOUS PRIME-BOOST OR SINGLE DOSE STUDIES**                                                                                                                                                                                                                                                     \n  HVTN-070 and -080 Phase I studies                        PV (PENNVAX(R)-B DNA expressing Gag, Pol, Env and DNA/IL-12                           DNA+IL-12 (i.m. or by electroporation)                                                    -CD4+ and CD8+ T cell responses                                   Kalams et al., [@B142]\n  IPCAVD-001                                               Ad26.ENVA.01                                                                          Intramuscular delivery of rAd26                                                           -Binding antibodies                                               Baden et al., [@B12]; Barouch et al., [@B21]\n  -Multiple CD8+ and CD4+ T cell responses                                                                                                                                                                                                                                                                   \n  -ADCC and virus inhibition                                                                                                                                                                                                                                                                                 \n  HVTN 090 Phase Ia study                                  VSV~IN~N4CT1~Gag~1 (recombinant VSV expressing HIV-1 Gag)                             Dose-escalating i.m. delivery                                                             Low level T cell responses detected following initial dosing      Fuchs et al., [@B85], [@B86]\n  Phase I study                                            Ad35-GRIN (Gag, RT, Integrase, Nef) and Ad35-GRIN/ENV                                 Intramuscular delivery of Ad35-GRIN/Env or Ad35-GRIN                                      -Robust, broad and polyfunctional CD4 and CD8+ T cells            Keefer et al., [@B146]\n  Phase I/II study (RISVAC02)                              MVA-B (monomeric gp120 and clade B Gag/Pol/Nef poly-protein)                          Three doses of MVA (i.m.)                                                                 -Durable antibody and cellular immune responses                   Garcia et al., [@B89]; Gomez et al., [@B105]\n  Phase I study                                            ADVAX (multigenic HIV-1 DNA vaccine)                                                  DNA by i.m. electroporation                                                               -CD4 and CD8+ T cells with multiple cytokines                     Vasan et al., [@B270]\n  VAX 003 (Phase III study)                                Bivalent recombinant gp120 vaccine: AIDSVAX B/E                                       Seven i.m. injections; with Alum adjuvant                                                 -Binding and neutralizing antibodies to gp120                     Pitisuttithum et al., [@B207]\n  VAX 004 (Multicentre Phase III study)                    Bivalent recombinant gp120 vaccine: AIDSVAX B/B                                       Seven i.m. injections; with Alum adjuvant                                                 -Binding and neutralizing antibodies to gp120                     Flynn et al., [@B83]; Gilbert et al., [@B100]\n\n*i.m., intramuscular; i.n., intranasal; i.d., intradermal; s.c., subcutaneous; i.p., intraperitoneal; ADCC, antibody dependent cytotoxicity; NAbs, neutralizing antibodies; BNAbs, broadly neutralizing antibodies*.\n\n###### \n\n**Representative preclinical studies**.\n\n  **Animals**                                              **Immunogen**                                                                                                   **Vectors, regimen and route of immunization**                                                  **Immune responses generated**                                               **Outcomes**                                                                               **References**\n  -------------------------------------------------------- --------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------- ------------------------------------------------------------------------------------------ ------------------------------------------------------------------------\n  **(I) HETEROLOGOUS PRIME-BOOST STUDIES**                                                                                                                                                                                                                                                                                                                                                                                                         \n  Mice and rabbits                                         HIV Env/Gag-Pol-Nef DNA, MVA-C (HIV Env/Gag-Pol-Nef and CN54gp140 protein)                                      Intramuscular delivery of DNA/MVA/Protein with TLR4 (GLA-AF adjuvant) for protein boost         Antibody and T cell responses                                                --                                                                                         McKay et al., [@B182]\n  Rhesus macaques                                          SIVmac239 Env/Gag DNA, rmIL-12 DNA and SIVmac239 protein vaccines                                               DNA-prime (by electroporation)/i.m. or i.d. Protein-boost, or DNA and protein co-immunization   Persistent mucosal Envelope-specific antibody responses                      Enhanced immunity by the co-immunization modality                                          Jalah et al., [@B139]\n  Rhesus macaques                                          SIV-Gag mosaic                                                                                                  DNA-prime (x3, i.m.) Ad5-boost (i.m.)                                                           -NAbs                                                                        Protection against SIVsmE660 challenge                                                     Roederer et al., [@B224]\n                                                           SIV-Env mosaic                                                                                                                                                                                                  -ADCC                                                                                                                                                                   \n                                                           SIV~mac239~ Env                                                                                                                                                                                                 -Cellular responses                                                                                                                                                     \n  Rhesus monkeys                                           DNA expressing SIVmac239 antigens + rmIL-12 and inactivated SIVmac239 virus particles as protein                DNA prime (i.m. followed by *in vivo* electroporation) /protein-boost                           -SIV-specific CTLs                                                           -Protection from SIV~SM~E660 acquisition                                                   Patel et al., [@B202]\n                                                                                                                                                                                                                                                                           -CD4+ and CD8+ memory T cells                                                                                                                                           \n                                                                                                                                                                           -Binding antibodies                                                                             -Reduced peak and chronic phase viremia                                                                                                                                 \n  Mice                                                     pCCMp24                                                                                                         DNA prime/Tiantan boost (i.m.)                                                                  Antibody and T cells                                                         --                                                                                         Excler et al., [@B79]; Liu et al., [@B166]\n  rddVTT-CCMp24                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  Rhesus macaques                                          SIV~SME543~-Gag/Pol/Env                                                                                         Prime-boost (i.m.) with: Ad26/MVA, Ad35/Ad26, DNA/MVA, MVA/Ad26                                 -NAbs                                                                        Protection from SIVmac251 acquisition or disease progression                               Barouch et al., [@B20]\n  -Binding antibodies                                                                                                                                                                                                                                                                                                                                                                                                                              \n  -Cellular responses                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Mice                                                     Ad35-GRIN/ENV and MVA-C (Gag/Env/Pol)                                                                           Ad35-GRIN/ENV-prime (i.m.)/MVA-boost (i.m.)                                                     Polyfunctional CD8+ T cells                                                  --                                                                                         Ratto-Kim et al., [@B213]\n  Macaques                                                 SIV DNA/GM-CSF (SIV239 Gag/PR/RT/Env/Tat/Rev) and MVA-SIVgpe                                                    DNA/GM-CSF- prime (i.m.)/MVA-boost (i.m.)                                                       -Neutralizing antibody responses                                             Sterile protection after SIVsmE660 challenge                                               Lai et al., [@B157]\n  -ADCC                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  Murine                                                   DNA-Env and gp120 protein vaccines                                                                              DNA Env-prime/gp120 protein-boost (i.m. and i.n.)                                               -Persistent mucosal and systemic Abs                                         --                                                                                         Cristillo et al., [@B60]\n                                                                                                                                                                           (Advax-M and Advax-P adjuvants)                                                                 -T cell responses                                                                                                                                                       \n  New-born and adult mice                                  BCG-HIVA, MVA-HIVA and HAdV5.HIVA                                                                               BCG-prime (i.p./i.d./s.c.) followed with i.m. MVA- or HAdV5- boost                              -Strong, cytotoxic CD8+ T cell responses                                     --                                                                                         Hopkins et al., [@B132]; Saubi et al., [@B233]\n  Rhesus macaques                                          VSV and SFV replicon expressing SIV-Gag/Env                                                                     VSV-prime (i.m. and i.n.)/SFVG-boost (i.m.)                                                     -High titer NAbs to Env proteins and weak cellular responses                 -Sterilizing immunity Control of SIVsmE660 breakthrough infections                         Schell et al., [@B236]\n  New-born macaques                                        VSV-SIVgpe (rVSV- Gag/Pol/Env) and MVA-SIVgpe                                                                   VSV-prime (oral)/MVA-boost (i.m.)                                                               -Systemic Abs, both systemic and local cellular responses                    --                                                                                         Van Rompay et al., [@B268]\n  Mice, rabbits and macaques                               Consensus or Polyvalent mosaic DNA and protein (gp120) vaccines                                                 DNA-prime (i.m.)/i.m. and i.d. rVaccinia-boost.                                                 -Broadly neutralizing antibodies and CD8+ T cell responses                   Enhanced immunogenicity                                                                    Wang et al., [@B275]; Santra et al., [@B231]\n  DNA-prime (gene gun)/ Protein-boost (i.d.) + IFA                                                                                                                                                                                                                                                                                                                                                                                                 \n  Rhesus macaques                                          VSV-SHIVGag/Pol/Env                                                                                             VSV-prime (i.m.)/MVA-boost (i.m.)                                                               -Persistent multi-functional                                                 Durable (over 5 years) control of SHIV89.6P replication                                    Rose et al., [@B227] Schell et al., [@B235]\n                                                           MVA-SHIVGag/Pol/Env                                                                                                                                                                                             CD8+ T cells and NAbs                                                                                                                                                   \n  Rabbits macaques                                         HIV-1 Env gp120                                                                                                 DNA (electroporation)/gp120 protein boost                                                       -Persistent Th1, CTL and Env responses                                       Neutralization of sensitive SHIV isolates                                                  Cristillo et al., [@B61]\n  Rhesus macaques                                          CMV-SHIVdEN and SeV-Gag                                                                                         DNA prime (i.m.)/Sendai Virus boost (i.n.)                                                      -CD8+ T cells                                                                Durable control of SIVmac239 and SHIV89.6PD                                                Matano et al., [@B176]; Takeda et al., [@B254]; Kawada et al., [@B144]\n  Rhesus Macaques                                          replication-defective SHIV particles and MVA-SHIV (SIV Gag, SIV Pol and HIV Env)                                Intrarectal DNA prime/MVA boost                                                                 -Antibodies in plasma                                                        -Preserved CD4 T cells -Reduced disease progression after SHIV 89.6P challenge             Wang et al., [@B277]\n  -Cellular responses                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Rhesus macaques                                          SHIV-DNA plus IL-2 and rMVA                                                                                     DNA + IL-12-prime (i.n.)/MVA-boost (i.n.)                                                       -Mucosal and systemic antibody and cellular responses                        Protection from SHIV 89.6P challenge                                                       Bertley et al., [@B33]\n  Mice and monkeys                                         E1/E3-deleted AdHu5 and E1-deleted AdC7 or AdC6, expressing Gag37                                               i.m. prime-boost with: AdC7/AdC6/AdHu5 or AdHu5/AdC6/AdC7                                       -Robust CD8+ CD4+ T cells                                                    --                                                                                         Reyes-Sandoval et al., [@B217]\n  -Antibody responses                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Cynomolgus macaques                                      DNA- HIV-1 IIIB Env/Gag/RT/Rev/Tat/Nef, MVA- HIV-1 IIIB Nef-Tat- Rev, SIVmacJ5 Gag/Pol and Vaccinia HIV-1 Env   DNA prime/MVA boost (i.m. or mucosally)                                                         -Antibody and cellular responses                                             Protection from infection                                                                  Makitalo et al., [@B171]\n  Mice                                                     HIV-1 Env IIIB Ag (DNA-Env and MVA-Env)                                                                         DNA-Env-prime/MVA-Env-boost (i.n. with Cholera toxin adjuvant)                                  -Mucosal CD8+ T cells, mucosal and systemic antibodies                       --                                                                                         Gherardi et al., [@B97]\n  -Beta-chemokines                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  Rhesus monkeys                                           DNA, MVA and Ad5 vectors expressing SIVmac239 Gag                                                               DNA Prime (i.m.)/MVA- or Ad5- boost (i.m.)                                                      -Robust CD8+ T cells with cytotoxic activity                                 Pronounced attenuation of SHIV infection and mitigated disease progression                 Shiver et al., [@B248]\n  Macaques                                                 DNA and NYVAC SIV-gpe (Gag/Pol/Env)                                                                             DNA-prime (i.m.)/NYVAC-boost (i.m.)                                                             -Durable CD8+ T cell responses                                               --                                                                                         Hel et al., [@B124]\n  **(II) HOMOLOGOUS PRIME-BOOST OR SINGLE DOSE STUDIES**                                                                                                                                                                                                                                                                                                                                                                                           \n  Mice and rabbits                                         Ad4Env160                                                                                                       i.m., i.n., or s.c. delivery of rAd4                                                            -T cell and antibody responses                                               Neutralization of tier-1 and tier-2 pseudoviruses                                          Alexander et al., [@B6]\n  Ad4Env140                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Ad4Env120                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Mice                                                     Ad35-GRIN/ENV and MVA- Gag/Env/Pol                                                                              Ad35-prime (i.m.)/Ad35-boost i.m.): MVA-prime (i.m.)/MVA-boost (i.m.)                           -Polyfunctional CD8+ T cells                                                 --                                                                                         Ratto-Kim et al., [@B213]\n  Rhesus macaques                                          SIV~SME543~-Gag/Pol/Env                                                                                         MVA-prime (i.m.)/MVA-boost (i.m.)                                                               -Neutralizing Abs, binding antibodies and cellular responses                 Protection from SIVmac251 acquisition or disease progression                               Barouch et al., [@B20]\n  Rhesus macaques                                          RhCMV-SIV/Gag, Rev/Nef/Tat, Pol, Env                                                                            RhCMV vectors delivered by s.c. injection                                                       -Strong and persisting, polyfunctional effector memory CD8+ and CD4+ cells   Viral clearance and durable protection from SIVmac239 disease progression                  Hansen et al., [@B118], [@B116]\n  Rhesus monkeys                                           SIV-Gag, SIV-Env and SIV Rev-Tat-Nef fusion protein                                                             Intravenous delivery of recombinant Rhadinovirus                                                -Persistent effector memory CD8+ T cells                                     Control of SIVmac239 replication                                                           Bilello et al., [@B34]\n  Rhesus macaques                                          Rabies virus (RV) expressing SIVmac239 Gag/Pol or Env                                                           Intramuscular delivery of rRV constructs                                                        -Polyfunctional CD8+ T cells in the mucosa                                   Control of SIVmac251-CX challenge                                                          Faul et al., [@B80]\n  -NAbs                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  Rhesus and Cynomolgus macaques                           SIV-Gag DNA + rIL-12 DNA vaccines                                                                               Intramuscular DNA delivery                                                                      T cell and Antibody responses                                                Improved clinical outcome after SHIV\\[89.6P\\] challenge                                    Boyer et al., [@B39]; Chong et al., [@B56]\n  Juvenile and Infant Rhesus macaques                      ALVAC-SIV and MVA-SIV both expressing SIV-Gag/Pol/Env                                                           Multiple immunizations with ALVAC-SIV (i.m.) or MVA-SIV (i.m.)                                  -High titres of binding antibodies, low-level T cell responses               Protection from oral SIVmac251 challenge, and reduced viremia in breakthrough infections   Van Rompay et al., [@B269]\n  Mice                                                     HIV-1 Env IIIB Ag (DNA-Env and MVA-Env)                                                                         MVA-Env/MVA-Env                                                                                 -Mucosal CD8+ T cells, mucosal and systemic antibodies                       --                                                                                         Gherardi et al., [@B97]\n                                                                                                                                                                           DNA-Env/DNA-Env (i.n. with Cholera toxin adjuvant)                                              -Beta-chemokines                                                                                                                                                        \n  Mice                                                     Influenza virus expressing HIV-1 ELDKWA epitope                                                                 i.n. prime/boost with chimeric influenza virus, followed with i.p. boost with live virus        -Neutralizing antibodies                                                     Neutralization of distantly related HIV-1 isolates                                         Muster et al., [@B191]\n\n*i.m., intramuscular; i.n., intranasal; i.d., intradermal; s.c., subcutaneous; i.p., intraperitoneal; ADCC, antibody dependent cytotoxicity; NAbs, neutralizing antibodies; BNAbs, broadly neutralizing antibodies*.\n\nRecombinant DNA vaccine vectors\n===============================\n\nDNA plasmid vaccines can induce both T and B cell immune responses, and are popular for their safety, stability, versatility and ease of large scale production. Most importantly is the fact that they can be used repetitively to boost immunity (Valentin et al., [@B266]) without the risk of immune interference as is the case with viral vectors with high prevalence of pre-existing immunity. However, on their own DNA plasmid vaccines have exhibited very limited immunostimulatory capacity and often induced sub-optimal immune responses. Recent advances in DNA delivery such as intramuscular, skin or intradermal electroporation (Selby et al., [@B242]; Widera et al., [@B282]; Brave et al., [@B41]; Vasan et al., [@B270]; Kopycinski et al., [@B154]) or use of other physical delivery methods such as gene gun and biojector devices (Drape et al., [@B67]; Wang et al., [@B276]; Graham et al., [@B108]), together with concurrent use of cytokine adjuvants including IL-2, IL-12, and IL-15 (Winstone et al., [@B283]; Kalams et al., [@B141], [@B142]) have greatly improved the immunogenic potential of DNA vaccines. In particular, IL-12 was shown to significantly augment the frequency, magnitude and breadth of Gag-specific immune responses in healthy volunteers immunized with a recombinant DNA vaccine expressing HIV-1 Gag (Kalams et al., [@B141], [@B142]). Similarly, when macaques were co-immunized with a plasmid encoding IL-12 and a DNA plasmid expressing SIV-Gag, strong antibody and cellular responses which correlated with a better clinical outcome were induced (Boyer et al., [@B39]; Chong et al., [@B56]). More impressively, co-delivery of a plasmid encoding GM-CSF with a DNA vaccine expressing SIV genes induced strong neutralizing antibody responses and ADCC, which protected against infection with SIVsmE660 (Lai et al., [@B157]). The use of strong adjuvants such as glucopyranosyl lipid A (a TLR4 agonist) in a DNA/MVA/protein immunization regimen was shown to enhance both antibody and T cell responses (McKay et al., [@B182]), while plasmids encoding the TLR5 agonist, flagellin, enhanced both antibody and T cell immunity to influenza virus (Applequist et al., [@B10])\n\nOther significant improvements in DNA vaccine technology include codon optimization, use of stronger promoters/enhancers and signal peptides such as the tissue plasminogen activator (tPA) and lysosome associated membrane protein (LAMP1), all of which significantly enhance transgene expression and trafficking, thus leading to increased vaccine immunogenicity (Wang et al., [@B273]; Yan et al., [@B290]; Wallace et al., [@B272]). Furthermore, ease of DNA manipulation provides a platform to deliver polyvalent or multi-gene vaccine components which can increase the breadth and depth of vaccine-induced immunity to reduce immune escape. This strategy showed remarkable success in rabbit experiments where a polyvalent gp120 vaccine induced broadly neutralizing antibody responses as opposed to the monovalent vaccine (Wang et al., [@B275]). Similarly, polyvalent mosaic plasmid DNA vaccines have demonstrated enhanced immunogenicity in mice (Kong et al., [@B153]) and rhesus monkeys (Santra et al., [@B231]).\n\nSeveral studies indicate that delivery of DNA vaccines by electroporation induces both cellular and humoral immune responses which are long-lived and can persist for several years with or without subsequent heterologous boosting (Cristillo et al., [@B61]; Patel et al., [@B203]; Jalah et al., [@B139]). In particular, the level of HIV-specific immune responses to the multigenic ADVAX vaccine was increased by up to 70-fold when electroporation was used for delivery (Vasan et al., [@B270]). Nonetheless, DNA vaccines consistently show much better immunogenicity when used as priming components in conjunction with viral vectors such as adenoviruses (Shiver et al., [@B248]; Hammer et al., [@B115]; Borthwick et al., [@B38]), MVA (Sandstrom et al., [@B230]; Gudmundsdotter et al., [@B112]; Bakari et al., [@B13]; Borthwick et al., [@B38]), fowlpox (Kent et al., [@B148]), and NYVAC (Hel et al., [@B124]) in heterologous prime boost regimens delivering the same vaccine inserts, or in co-immunization strategies that combine DNA-prime with protein boosting (Kennedy et al., [@B147]; Wang et al., [@B274]). As a matter of fact, prime-boost regimens still remain the most successful strategies that emphasize the potential of DNA vaccines. It was recently shown that a DNA-prime/protein-boost regimen was significantly better than either DNA/DNA or protein/protein alone regimens for generating long-term protection of mice against *Leishmania donovani* (Mazumder et al., [@B177]). The DNA and protein co-immunization modalities are particularly desirable as they maximize induction of long-lived humoral and cellular immune responses which can disseminate to mucosal sites, including the genito-rectal mucosae (Patel et al., [@B202]; Jalah et al., [@B139]). A recent study has demonstrated in small animal models that concurrent, multiple-route DNA vaccinations comprising DNA prime by electroporation, followed with intranasal, intramuscular, subcutaneous or transcutaneous homologous protein boost induced strong HIV-specific B and T cell responses (Mann et al., [@B172]). Independently, another study showed enhancement of HIV gp120-specific IgA responses in serum and mucosal secretions following a DNA env-prime and gp120 protein-boost delivered with novel carbohydrate-based adjuvants (Advax-M and Advax-P) which were specifically designed for mucosal and systemic immune enhancement (Cristillo et al., [@B60]). The tremendous effect of a DNA prime in enhancing antibody responses to protein vaccines was also documented in a Phase 1 clinical study, where intramuscular delivery of a DNA priming vaccine followed with recombinant protein boost stimulated higher frequencies of B and T cells, as well as higher neutralizing antibody titres and ADCC in contrast to immunization with protein alone (Spearman et al., [@B251]). Perhaps the most exciting of the DNA-prime/protein-boost studies is the 6-plasmid polyvalent DNA vaccine expressing gp120 and Gag, followed by QS21-adjuvanted polyvalent gp120 protein boost (DP6-001 study) in which multifunctional T cells and high-titre gp120-specific binding and broadly-neutralizing antibodies as well as ADCC were induced (Graham et al., [@B109]; Bansal et al., [@B17]; Wang et al., [@B274]; Vaine et al., [@B265]).\n\nApart from effective delivery strategies and routes of immunization, there is evidence showing that expression of DNA vaccines and subsequent immunogenicity in humans and other primates can be limited by serum amyloid P component (SAP), a protein found in blood and known to bind strongly to DNA (Wang et al., [@B278], [@B279]). In small animals this protein either binds weakly or does not exist at all. Thus, depletion of SAP protein prior to administration of DNA vaccines is another new strategy being tested to improve DNA vaccine immunogenicity. This concept has been proven in mice, where depletion of SAP using the bis-d-proline compound CPHPC (Bodin et al., [@B35]; Gillmore et al., [@B103]) was shown to augment antibody and cellular immune responses to a DNA vaccine expressing Hepatitis B surface antigens (Wang et al., [@B279]). The concept is currently being tested in a Phase 1 clinical trial (HIVCORE003) of healthy adults using the T-cell based HIV candidate vaccine, HIVconsv.\n\nAlthough the efficacy of an HIV DNA vaccine is yet to be demonstrated in humans, various studies (prophylactic and therapeutic) in the macaque model have reported protective immune responses which controlled SIV/SHIV replication or protected from infection (Rosati et al., [@B226], [@B225]; von Gegerfelt et al., [@B271]; Valentin et al., [@B266]; Patel et al., [@B202]). In particular, a study combining a DNA/MVA mucosal delivery of a DNA construct expressing replication-defective SHIV particles and MVA expressing SIV-Gag/Pol and HIV Env (MVA-SHIV) demonstrated significant protection from disease progression after a SHIV89.6P challenge (Wang et al., [@B277]). Furthermore, mucosal co-delivery of a DNA priming vaccine together with an IL-2 encoding vector, followed by MVA boost also induced protective immunity against SHIV89.6P challenge (Bertley et al., [@B33]). The results in these macaque models, together with the documented efficacy of DNA vaccines against animal diseases \\[e.g., equine West Nile Virus (WNV) (Davis et al., [@B64]), melanoma in dogs (Bergman et al., [@B31]) and infectious hematopoietic necrosis virus (IHNV) in salmon (Garver et al., [@B93]; Kurath et al., [@B156])\\] raise hopes that with the right immunogen and effective delivery strategies (including adjuvants), plasmid DNA vaccines for HIV/AIDS could achieve efficacy in clinical trials, when used alone, but more realistically in prime-boost combinations with live viral-vectored or protein vaccines.\n\nNon-replicating recombinant viral vectors\n=========================================\n\nAdenovirus vaccine vectors\n--------------------------\n\nAdenoviruses are the most powerful vectors for inducing both antibody and cell-mediated immunity to inserted transgenes and are known to elicit between 5- and 10-fold stronger T cell responses compared to conventional naked DNA or MVA/pox-like virus vectors (Xiang et al., [@B289]; He et al., [@B123]; Fitzgerald et al., [@B82]; Casimiro et al., [@B50], [@B49]; Tatsis and Ertl, [@B258]; Catanzaro et al., [@B52]). The Adenovirus vectors use either the Coxsackie and Adenovirus Receptor (CAR) or CD46 receptors (Bergelson et al., [@B30]; Gaggar et al., [@B88]) and can infect a wide variety of cells, including dendritic cells. In particular, group B adenoviruses such as Ad35 recognize CD46 surface protein and infect DCs more efficiently than group C isolates. These vectors achieve higher levels of transgene expression which in turn results in stronger and persistent immune effector functions (Zhang et al., [@B301]; Hutnick et al., [@B138]; Suleman et al., [@B253]). Several studies indicate that adenoviruses predominantly stimulate persistent effector memory CD8+ T cell responses (Yang et al., [@B292], [@B293]; Tatsis et al., [@B259]) which are more suitable for immediate control of invading pathogens at peripheral entry sites such as the genital mucosa (Cerwenka et al., [@B53]; Sallusto et al., [@B229]; Huster et al., [@B137]), and have shown tremendous success in animal studies (Liu et al., [@B167]). In addition to the effector memory T cells, stable central memory CD8+ T cell populations are also generated, thus providing surveillance in both peripheral and lymphoid sites. Although persisting adenovirus-driven immune responses could also be due to the long-term presentation of antigens by non-haematopoietic cells serving as unlimited antigen depot (Finn et al., [@B81]; Kim et al., [@B150]; Bassett et al., [@B23]), long-lived immunity is largely attributed to persisting low-level expression of inserted immunogens. Adenovirus genomes are known to persist for prolonged periods in various cell types (including those at inoculation sites) where they remain transcriptionally active and continuously produce low-levels of antigen to prime na\u00efve T cells while also maintaining the effector memory T cells (Yang et al., [@B294], [@B295]; Tatsis et al., [@B259]). Furthermore, the arising effector memory T cells express the IL-7 receptor (CD127) which allows their prolonged survival in the absence of antigen. Besides induction of potent adaptive immune responses, adenoviruses also stimulate innate immunity via highly inflammatory responses which involve TLR2, TLR9, NOD-like receptors and the type 1 interferon pathways that result in abundant cytokine and chemokine secretion (Hensley et al., [@B126]; Nazir and Metcalf, [@B194]; Appledorn et al., [@B9]; Muruve et al., [@B189]). Another attractive feature of adenovirus vectors is their ability to induce both systemic and mucosal immune responses following parenteral delivery, as well as their suitability for mucosal immunization (Sharpe et al., [@B244]; Xiang et al., [@B286]; Bangari and Mittal, [@B16]; Haut et al., [@B122]).\n\nThe most well-characterized of the adenovirus vectors is human Ad5, successfully used as a delivery vector for a rabies vaccine and found to be very good at inducing protective virus neutralizing antibodies concurrently with CD8+ and CD4+ T cells (Xiang et al., [@B287], [@B289]). In the HIV field, Ad5 was used as a booster immunization following DNA priming and induced strong CD8+ T cell responses in a large proportion of the STEP study vaccinees (Buchbinder et al., [@B45]; McElrath et al., [@B181]). However, clinical efficacy may have been significantly compromised by pre-existing neutralizing antibodies (ranging from 40 to 70% in developed countries and greater than 90% in developing countries) and cellular immunity (Fitzgerald et al., [@B82]; Holterman et al., [@B131]; Bangari and Mittal, [@B16]; Xiang et al., [@B285]; Lasaro and Ertl, [@B158]; Ersching et al., [@B76]; Mast et al., [@B175]; Barouch et al., [@B19]). These results were recapitulated in a non-human primate study using low-dose penile exposure to SIVmac251 in Ad5 seropositive animals immunized with SIV-Gag/Pol/Nef (Qureshi et al., [@B210]). Possibly, adenovirus vaccination boosted the numbers of activated CD4+ T cells which are targets for HIV-1 (Benlahrech et al., [@B29]). While this might seem a plausible explanation, especially when considering the potential of such activated targets to traffic to the genito-rectal mucosae (Tatsis et al., [@B259]; Benlahrech et al., [@B29]), this argument is strongly contested by observations that other vaccine carriers such as DNA and MVA do stimulate CD4+ T cell activation but have not been associated with increased HIV acquisition. However, it is worth noting that DNA/MVA vaccines are yet to be tested for efficacy in large clinical trials and as such their potential to enhance HIV acquisition has never assessed. Furthermore, DNA/MVA vaccines combinations have not been associated with long-term persistence of activated T cells or mucosal homing. Another postulated theory is the formation of adenovirus-specific antibody immune complexes that activate both dendritic and CD4+ T cells hence fuelling infection (Perreau et al., [@B205]). In this study, Ad5 immune complexes were strongly correlated with higher HIV infection in the *in vitro* cultures, thus supporting a stronger likelihood of enhanced HIV acquisition. Should either or both of these theories be true, this would have dire consequences for other clinical trials using Ad5 to deliver non-HIV immunogens such as malaria (Sedegah et al., [@B240]; Tamminga et al., [@B257]; Chuang et al., [@B57]) and TB (Smaill et al., [@B249]) vaccines which will induce similar phenotypes and pre-dispose the vaccinees to increased HIV acquisition risk, although this may not be apparently detectable as these studies may not monitor HIV acquisition.\n\nApart from the issue of pre-existing immunity, immunization with Ad5 can induce neutralizing antibodies in na\u00efve individuals which can be a hindrance for successive immunizations with the same or cross-reactive adenoviral vectors (Casimiro et al., [@B51]; Bangari and Mittal, [@B16]). Thus new rare adenovirus vectors with lower pre-existing immunity such as Ad26 and Ad35 are becoming more attractive (Holterman et al., [@B131]; Abbink et al., [@B1]; Barouch et al., [@B20]; Zhang et al., [@B299]), although these are relatively less immunogenic compared to Ad5 (Colloca et al., [@B58]). Besides the lower sero-prevalence, Ad26 neutralizing antibody titres are usually very low compared to Ad5 (Abbink et al., [@B1]; Chen et al., [@B55]; Mast et al., [@B175]). As an HIV vaccine delivery vector, Ad26 was shown to induce broadly functional cellular and antibody responses with viral inhibitory capacity in a first-in-human (IPCAVD-001) clinical trial of an HIV envelope immunogen (Ad26.ENVA.01) (Baden et al., [@B12]; Barouch et al., [@B21]). In this study, a dose-dependent expansion of the magnitude, breadth, and epitopic diversity of Env-specific binding antibody responses were observed. The responses comprised multiple CD8+ and CD4+ T cell memory subpopulations and cytokine secretion phenotypes. Antibody-dependent cell-mediated phagocytosis and degranulation functional activity were also observed. Ad35 has also shown high immunogenicity in healthy volunteers, eliciting robust and polyfunctional CD8+ and CD4+ T cells in a majority of volunteers immunized with Ad35-GRIN (an immunogen based on Gag, RT, integrase and nef) or Ad35-GRIN/ENV (premixed Ad35-GRIN and Ad35-ENV vaccines) (Keefer et al., [@B146]). Similarly, in BALB/c mice, an Ad35-GRIN/ENV-prime followed by a boost with rMVA containing Gag/Env/Pol genes from various HIV-1 clades induced polyfunctional CD8+ Gag-specific central and effector memory T cells which were superior to those elicited in homologous Ad35/Ad35 or MVA/MVA prime boosts (Ratto-Kim et al., [@B213]).\n\nOther rare adenovirus vectors include human Ad6, chimpanzee Ad3, Ad63, and Ad68 (Barnes et al., [@B18]; Colloca et al., [@B58]; Dicks et al., [@B65]; O\\'Hara et al., [@B197]; Roshorm et al., [@B228]). The chimpanzee adenoviruses remain attractive in particular due to their high immunological potency and low sero-prevalence, as well as extremely low or virtually absent cross-reactivity with human adenoviruses (Xiang et al., [@B285]; Chen et al., [@B55]; Colloca et al., [@B58]). Furthermore, chimpanzee adenoviruses induce stronger T and B cell responses in heterologous prime-boost regimens even in the presence of pre-existing immunity to Ad5 (Tatsis et al., [@B260]). Apart from using these naturally occurring human and chimpanzee adenoviruses, new derivatives of adenovirus vectors that have equivalent immunogenicity but with significantly lower pre-existing antibodies are currently being developed (Dicks et al., [@B65]; Lopez-Gordo et al., [@B168]). However, it is worth noting that pre-existing cellular immunity (CD8+ and CD4+ T cells) may be a major deterrent as unlike antibodies, these cells are highly cross-reactive across adenovirus serotypes because they are directed to conserved sequences of adenovirus (Olive et al., [@B199]; Fitzgerald et al., [@B82]; Frahm et al., [@B84]). Nevertheless, some studies indicate that Ad5 and Ad26 vectors can still elicit significant systemic and mucosal responses even in people with pre-existing immunity (Barouch et al., [@B21]; Smaill et al., [@B249]). Immunogenic adenoviruses faced with significant pre-existing immunity problems can be improved by modification of the antibody-binding sites, especially within the variable hexon loops in order to reduce NAb binding whilst maintaining immunogenicity (Bruder et al., [@B44]). This can be achieved via point mutations or complete replacement (Roberts et al., [@B221]; Abe et al., [@B2]; Pichla-Gollon et al., [@B206]; Bruder et al., [@B43]).\n\nBesides their immunogenicity when used alone, adenovirus vaccines are also very immunogenic when used to prime responses which are then boosted by other vaccine vectors (Tatsis et al., [@B261]; Ratto-Kim et al., [@B213]). In particular, adenovirus-prime followed with MVA-boost can induce high frequencies of much more long-lived, potent T cells (Reyes-Sandoval et al., [@B218], [@B216]; Capone et al., [@B47]; Hill et al., [@B127]). A Phase I clinical trial of a T-cell HIV vaccine based on the conserved regions was recently shown to elicit exceptionally high magnitude and polyfunctional T cell responses (circa 5000 IFN-\u03b3 ELISPOT SFU/million cells) in HIV-negative healthy volunteers when primed with chimpanzee Ad63 (ChAdV63-HIVconsv) followed with MVA-HIVconsv boost (Borthwick et al., [@B38]). The vaccine-induced CD8+ T cells exhibited potent *in vitro* antiviral activity. This study also demonstrated that the magnitude and functional capacity of T cells induced in a regimen comprising three priming doses of DNA followed with ChAdV63 and MVA (DDDCM) did not differ significantly from those in a simplified ChAdV63-prime and MVA-boost (CM) regimen. The superior immunogenicity of this regimen is not unique to HIV immunogens, as it has also been demonstrated in preclinical and clinical studies of experimental malaria vaccines (Dunachie et al., [@B72]; Draper et al., [@B68]). Such repeated heterologous immunizations with the same transgene are known to increase both the magnitude and functional quality of vaccine-specific T cells and to allow more efficient migration to mucosal-associated tissues (Tatsis et al., [@B261]). This is important in HIV infection, as effector immune cells in mucosal sites could block HIV transmission. It has also been shown that DNA priming followed with adenovirus boosting can reduce the level of anti-vector antibodies and increase transgene-specific immune responses (Xiang et al., [@B288]; Yang et al., [@B296]), although this is questionable when considering the STEP study which employed a DNA-prime/Ad5-boost regimen. However, it is possible that this regimen effectively reduced the anti-vector antibody effect, thus curtailing a potentially worse outcome in the absence of DNA priming. Furthermore, prime-boost regimens with various combinations of adenovirus vectors were shown to induce robust frequencies of HIV-1 Gag-specific CD8+ T cells in nonhuman primates (Reyes-Sandoval et al., [@B217]), although it has to be appreciated that the level of pre-existing Ad5 immunity in NHPs would be lower or absent.\n\nAdenoviruses are only associated with benign human pathologies, but their greatest limitation is pre-existing immunity which dampens vaccine-specific immunity by limiting transgene expression, while potentially exacerbating HIV acquisition. However, all else considered, Adenoviruses remain by far the most promising vaccine carriers for HIV-1, because unlike other vectors, they induce exceptionally high and persistent frequencies of vaccine specific T cells, which is a requirement for sustained HIV control. Although their efficacy has probably been hampered by high sero-prevalence, this no longer seems an insurmountable hurdle in light of the enormous amount of research efforts directed at finding strategies to circumvent the problems of pre-existing immunity (Gabitzsch et al., [@B87]). Additionally, replicating adenoviruses such as AdH4 and AdHu7 which can be delivered orally in the form of edible capsules might help to overcome pre-existing immunity (Xiang et al., [@B286]). Moreover, intranasal or oral delivery of adenoviruses has been shown to provide superior protection in animal models, and might trigger mucosal immune responses well-situated for preventing HIV acquisition. Perhaps adenovirus vectors engineered not to induce CD4+ T cells could be an alternative to overcome increased HIV-1 acquisition risk, although lacking CD4+ T cell help for the CD8+ T cells might compromise the differentiation and stability and thus efficacy of both CD8+ T cells and antibody responses (Yang et al., [@B295]).\n\nRecombinant MVA (rMVA) vectors\n------------------------------\n\nApart from their excellent safety profile, inherent adjuvant properties and ease of large scale production, recombinant vaccinia virus vectors are also popular for their large genomes which facilitate insertion of larger immunogens (Smith and Moss, [@B250]). MVA does not replicate in humans (Carroll and Moss, [@B48]) due to serial passaging in chick embryo fibroblasts which resulted in loss of more than 10% of its genome (Meyer et al., [@B186]), and its safety was well-documented during the smallpox eradication campaign (Mahnel and Mayr, [@B170]). MVA\\'s potent immunostimulatory properties are achieved in a cascade of events involving induction of type 1 interferons, various chemokines for cell migration and activation of several cellular signaling pathways (Price et al., [@B208]). The immunostimulatory potency of MVA is largely attributed to the absence of genes involved in immune evasion (such as those that interfere with IFN-\u03b1, IFN-\u03b2, and TNF-\u03b1), thus allowing for stronger innate immunity to be generated (Antoine et al., [@B8]). MVA vectors are particularly important for generating strong T cell immunity against intracellular pathogens and cancers, but have also been shown to induce potent, high titre antibodies in a variety of disease models including SIV and malaria (Gherardi et al., [@B96]; Draper et al., [@B70], [@B69]; Barouch et al., [@B20]). However, it is now well established that MVA vectors are more suited for boosting rather than priming, and depending on the priming vector (e.g., DNA or live vectors such as fowlpox and adenoviruses), MVA can induce various phenotypes of T cells, either predominated by CD4+ or CD8+ subsets or a combination of both.\n\nIn pre-clinical and clinical studies of malaria, recombinant MVA was shown to be highly immunogenic as it induced strong (and protective) cellular and antibody responses to malaria antigens, either on its own or when used to boost responses primed by vectors such as DNA, fowlpox or AdHu5 (Schneider et al., [@B237], [@B238]; Gilbert et al., [@B102], [@B101]; McConkey et al., [@B178]; Anderson et al., [@B7]; Webster et al., [@B280]; Bejon et al., [@B25]; Sheehy et al., [@B246]). Recombinant MVA85A (expressing the mycobacterial antigen Ag85A) was also shown to induce strong and durable T cell responses in various clinical studies (Scriba et al., [@B239]; Tameris et al., [@B256], [@B255]). Furthermore, it was demonstrated that MVA expressing influenza A virus antigens (MVA-NP+M1) efficiently boosted CD8+ T cell responses to achieve clinical efficacy in humans (Berthoud et al., [@B32]; Lillie et al., [@B165]). As a therapeutic vaccine for cancer, recombinant MVA expressing the human papilloma virus antigens E2, E6, or E7, with or without IL-12 was shown to induce T and B cell immunity resulting in controlled HPV load and subsequent regression or complete elimination of precancerous lesions in a majority of vaccinees (Corona Gutierrez et al., [@B59]; Garcia-Hernandez et al., [@B91]; Albarran et al., [@B4]). Additionally, MVA expressing 5T4 antigen (TroVax) induced 5T4-specific antibody and cellular responses which correlated with tumor regression in a clinical trial of patients with advanced colorectal cancer (Harrop et al., [@B121]).\n\nAlthough there is clear demonstration of the clinical efficacy of prophylactic and therapeutic MVA-vectored vaccines for malaria, TB, influenza virus and cancer, MVA vaccines for HIV are yet to be evaluated for clinical efficacy. However, Phase I and II studies of MVA expressing HIV antigens, either alone or in various prime-boost combinations indicate modest to strong immunogenicity (Guimaraes-Walker et al., [@B113]; Howles et al., [@B135]; Bakari et al., [@B13]; Garcia et al., [@B89]; Goepfert et al., [@B104]; Gomez et al., [@B105]). In particular, the MVA-B candidate HIV vaccine expressing monomeric gp120 and Gag-Pol-Nef poly-protein of clade B where MVA was administered without prior priming, induced long-lasting robust and polyfunctional effector memory T cell and antibody responses in Phase I/II studies (Garcia et al., [@B89]; Gomez et al., [@B105]). Furthermore, MVA has shown much higher immunogenicity when combined in prime-boost regimens with other priming vectors such as DNA, fowlpox or adenovirus (Goepfert et al., [@B104]; Keefer et al., [@B145]; Borthwick et al., [@B38]). In Phase 1 studies of the HIVA immunogen (based on HIV clade A and a string of CTL epitopes), priming with DNA (pTHr.HIVA) followed with MVA boosting (MVA.HIVA) was found to be immunogenic, inducing multifunctional and proliferative CD8+ and CD4+ T cell responses in greater than 70% of the vaccinees (Mwau et al., [@B193]; Goonetilleke et al., [@B107]).\n\nAs discussed earlier, a Phase I study combining DNA- and/or ChAdV63-prime followed with MVA boost to deliver an HIV-1 T cell immunogen induced high magnitude T cell responses with potent antiviral capacity (Borthwick et al., [@B38]). This study and similar studies of malaria vaccines (Sheehy et al., [@B246], [@B245]; O\\'Hara et al., [@B197]) showed that the magnitude of T cell responses induced by ChAdV63 alone were modest, but significant boosting was achieved following MVA administration, thus highlighting the superior immunogenic potential of MVA when combined with appropriate priming vectors such as BCG (Whelan et al., [@B281]; Scriba et al., [@B239]), natural influenza A virus (Berthoud et al., [@B32]) or ChAdV63 (Colloca et al., [@B58]). Remarkably, a DNA/MVA prime boost of a vaccine expressing multiple HIV antigens induced responses in about 90% of volunteers and demonstrated strong immunogenicity despite pre-existing immunity to vaccinia virus (Sandstrom et al., [@B230]). As a therapeutic HIV vaccine vector, rMVA was found to be safe and to significantly augment HIV-specific CD4+ and CD8+ T cell responses in HAART-treated HIV-infected volunteers immunized with the MVA.HIVA candidate vaccine (Dorrell et al., [@B66]; Ondondo et al., [@B200]; Yang et al., [@B291]). Furthermore MVA was found to be safe in neonates in a Phase 1 trial where MVA.HIVA was administered to infants born to HIV-infected or uninfected mothers (Afolabi et al., [@B3]). Therapeutic administration of MVA prime followed with fowlpox boost expressing Env, Gag, Tat, Rev, and Nef-RT fusion antigens increased the frequencies and breadth of T cell responses in young adults (Greenough et al., [@B111]).\n\nOne very attractive feature of rMVA (and other poxvirus vectors) is their ability to induce mucosal immune responses when administered via mucosal routes (Gherardi and Esteban, [@B94], [@B95]). In particular, murine and macaques studies using rMVA vectors demonstrated induction of protective HIV-specific immune responses within the genito-rectal mucosae, which in some cases correlated with reduced disease progression (Belyakov et al., [@B28]; Makitalo et al., [@B171]; Wang et al., [@B277]). Enhanced immunogenicity of rMVA in combination with DNA priming was also achieved by using the non-toxic B subunit of cholera toxin (CTB) as mucosal adjuvant (Gherardi et al., [@B97]). Thus, even though MVA may be inadequate as a stand-alone delivery platform, it definitely shows greater potential as a boosting vector (especially for the chimpanzee adenoviruses) and should be evaluated for efficacy in advanced HIV vaccine trials.\n\nRecombinant NYVAC vaccine vectors\n---------------------------------\n\nNYVAC vector is also a vaccinia-based vector which was highly attenuated by deletion of 18 genes involved in host range virulence. It has been shown to induce mainly CD4+ T cell responses, in contrast to MVA which has a stronger immunostimulatory potential and is known to induce both CD8+ and CD4+ responses (Mooij et al., [@B187]). However, in a trial of chronically infected patients on HAART, a NYVAC-based vaccine expressing Gag/Pol/Nef/Env from an HIV-1 clade B isolate (NYVAC-B) was found to be highly immunogenic and induced high magnitude, broad and polyfunctional CD4+ and CD8+ T cells (Harari et al., [@B120]). Similar to MVA, NYVAC elicits greater immune responses when used in prime-boost combinations rather than on its own (Harari et al., [@B119]; McCormack et al., [@B179]). In these EuroVacc studies, priming with DNA-C followed with NYVAC-C boost elicited broad, polyfunctional and durable CD4+ T cell responses in greater than 90% of volunteers, compared to only 40% when NYVAC was used alone (Harari et al., [@B119]). Moreover, in a preclinical study with a DNA prime followed with NYVAC boost, responses to a vaccine expressing SIV-Gag/Pol/Env were boosted 10-fold with improved quality and quantity of T cell responses (Hel et al., [@B124]). A NYVAC/SIV-gpe vaccine (expressing SIV Gag/Pol/Env) also elicited mucosal immune responses in macaques following both mucosal and systemic delivery (Stevceva et al., [@B252]). Despite the skewing toward CD4+ T cell responses, NYVAC has potential to stimulate and boost more balanced immune responses when combined with other vectors, and its potential should be fully explored, especially for therapeutic HIV vaccines which require re-invigoration of CD4+ T cell functions (and frequencies).\n\nCanarypox (ALVAC) vaccine vectors\n---------------------------------\n\nALVAC is an attenuated derivative of the canarypox virus that was repeatedly passaged in chick embryo fibroblasts and thus has restricted tropism with very minimal pathogenicity in humans (Yu et al., [@B297]). Despite the comparatively lower immunogenicity with respect to other poxvirus vectors such as MVA (Zhang et al., [@B300]) and NYVAC, the fact that ALVAC has no potential pre-existing immunity in humans makes it a more attractive HIV vaccine delivery vector. The ALVAC vector (vCP205) was shown to be safe and to induce strong CD8+ CTL and antibody responses to an HIV vaccine expressing gp120/41 and Gag/Pol sequences \\[ALVAC-HIV(vCP205)\\] in a Phase 1 clinical trial in the USA in the 1990 s (Belshe et al., [@B26]). A related ALVAC-based vaccine expressing multiple HIV antigens comprising Gag, Env, Nef, Pol and Pro \\[ALVAC-HIV(vCP300)\\] also induced durable CTL responses in healthy volunteers (Evans et al., [@B78]). In preclinical studies, ALVAC expressing SIV Gag/Pol/Env protected against low-dose oral SIVmac251 challenge of neonate rhesus macaques in a study design aiming to mimic HIV transmission through breast milk (Van Rompay et al., [@B269]). More recently ALVAC-based HIV vaccines have been tested in both adults and infants, where they have shown modest immunogenicity (Kintu et al., [@B151]; Kaleebu et al., [@B143]) and in the RV144 trial of ALVAC prime \\[ALVAC-HIV(vCP1521)\\] and protein boost (AIDSVAX B/E rgp120), the only HIV vaccine candidate to show efficacy (Rerks-Ngarm et al., [@B215], [@B214]).\n\nWhile it is unclear whether the modest success of RV144 was due to the immunostimulatory potential of canarypox virus vector or immunogenicity of the vaccine inserts, the fact that the immunogens in the RV144 trial vaccines are not significantly distinct from those used in other HIV vaccines in the field eliminates the \"immunogen effect,\" thus leaving the vectors and delivery methods as possible explanations. But, as the AIDSVAX vaccine (recombinant gp120) showed no efficacy in earlier trials (VAX003 and VAX004), the success of RV144 points to the delivery vector (ALVAC) and possibly the benefits of a combined viral vector and protein immunization regimen as opposed to homologous boosts. This might suggest that combined live vector-priming and protein-boost immunization modalities could be further refined to achieve greater potential for increased efficacy. Alternatively, protection by the combined vaccines could be attributed to T cell help for the antibody responses. It must however be noted that unlike the RV144 study, VAX003, and VAX004 were conducted in high-risk populations, which might be a strong confounding factor, although this might as well be reflective of the very limited efficacy of stand-alone protein subunit vaccines for HIV. Despite the modest efficacy of RV144, the immune responses waned within a short time indicating that ALVAC may not be a particularly suitable vector to induce long-lived anti-HIV immunity, unless it is combined with other powerful vectors. In direct comparison of immunogenicity, ALVAC was found to be less immunogenic than MVA, possibly due to MVA\\'s enhanced antigen expression within dendritic cells (Zhang et al., [@B300]). Nonetheless, ALVAC is still quite promising for HIV vaccine delivery, as it is also already licensed for delivery of several veterinary vaccines including the feline leukemia virus (FeLV) and feline rabies vaccine (PUREVAX) and RECOMBITEK vaccine which protects against canine distemper, equine influenza and West Nile Virus.\n\nMycobacterium bovis bacillus calmette-guerin (BCG) vaccine vectors\n==================================================================\n\nPrevention of breast milk transmission of HIV-1 remains an important goal for HIV vaccine researchers. BCG is an attenuated vaccine proven to be safe and has for many years been administered to new-born babies to immunize against Mycobacterium tuberculosis (Mtb). As such, BCG provides a platform to co-deliver HIV immunogens in neonates to potentially protect against mother-to-child transmission of HIV-1. The potential use of BCG as an HIV vaccine vector was explored in preclinical studies of adult and new-born BALB/c mice using the HIV-1 clade A Gag immunogen (HIVA) (Mwau et al., [@B193]). Priming with recombinant BCG expressing HIVA (BCG.HIVA) induced HIV-specific T cell responses which were efficiently boosted with rMVA (MVA.HIVA) (Hopkins et al., [@B132],[@B133]; Saubi et al., [@B233], [@B234]). In further related studies, priming with BCG.HIVA and boosting with a combination vaccine expressing HIVA and the Mtb antigen 85A (mMVA.HIVA.85A) induced robust IFN-\u03b3-producing T cells to both HIV-1 and Mtb antigens. Moreover, in adult mice, BCG.HIVA primed weak HIV-1-specific CD8+ T cell responses, which were strongly boosted with either Ad5 (HAdV5.HIVA) or rMVA (MVA.HIVA). Thus, immunization of neonates with recombinant BCG expressing HIV-1 immunogens, followed with an MVA boost expressing the same HIV immunogen might concurrently protect against Mtb and HIV-1. It remains to be seen how these rBCG-vectored HIV-1 vaccines will perform in clinical studies.\n\nReplication-competent viral vectors\n===================================\n\nThe unprecedented success of the SIVmac239 \u0394nef experimental vaccine in rhesus macaques (Reynolds et al., [@B220], [@B219]) gives a hint that possibly, a successful HIV vaccine will require a live delivery vector, as these are known to induce high magnitude, durable and broadly effective immunity. But as exciting as this may sound, there are significant challenges in terms of balancing the safety and immunogenicity vs. replicative capacity. Of the adenoviruses, Ad4 and Ad7 have been tested in clinical studies (by oral delivery) and were successfully used for the prevention of respiratory and enteric illnesses (Hoke and Snyder, [@B128]). These replication competent adenoviruses naturally infect and replicate in mucosal tissues (Patterson and Robert-Guroff, [@B204]) and could thus be quite relevant for HIV vaccines. Preclinical studies of recombinant Ad4 expressing HIV-1 clade C envelope gp160 (Ad4Env160), gp140 (Ad4Env140), and gp120 (Ad4Env120) demonstrated induction of envelope-specific T cells in mice and antibody responses in rabbits (Alexander et al., [@B6]). Serum from the rabbits was able to neutralize a tier 1 clade C pseudovirus and to a lesser extent, homologous and heterologous tier 2 pseudoviruses.\n\nA replicating CMV vectored SIV vaccine (RhCMV-SIV/Gag, Rev/Nef/Tat, Pol, Env) was shown to persist in vaccinated rhesus macaques and conferred durable protection from disease progression owing to induction of high magnitude effector memory CD8+ T cells, despite pre-existing CMV immunity (Hansen et al., [@B118], [@B116], [@B117]). Other replication-competent viruses in clinical development include the TianTan vaccinia virus (TT), Vesicular stomatitis virus (VSV), a derivative of NYVAC (NYVAC-C-KC) and Sendai virus (SeV). The TianTan vaccinia virus was used in a DNA-prime (pCCMp24)/Tiantan boost (rddVTT-CCMp24) regimen where it was shown to induce antibody and HIV-specific T cell responses (including memory phenotypes) following intramuscular delivery and has now been advanced to Phase II clinical study in China (Excler et al., [@B79]; Liu et al., [@B166]). The NYVAC-C-KC vectors have shown superior cellular and humoral immunity compared to the non-replicating NYVAC, at least in mice (Kibler et al., [@B149]; Gomez et al., [@B106]).\n\nA Sendai virus vector expressing SIV Gag (SeV-Gag) administered intranasally as a boost following intramuscular priming with an envelope-independent DNA vaccine (CMV-SHIVdEN) demonstrated very strong suppression of intravenous SIVmac239 challenge in rhesus macaques, which was extended over a 3-year period (Matano et al., [@B176]; Takeda et al., [@B254]; Kawada et al., [@B144]). Clinical investigations of a SeV-based candidate HIV vaccine expressing Gag \\[SeV-G (NP)\\] are ongoing in Rwanda, Kenya and the UK, and it is expected that results of these trials will provide a feel of the potential of Sendai virus as an HIV vaccine vector. Attenuated VSV is a non-pathogenic, low sero-prevalence vector that was also found to be quite promising as it achieved virus control during SHIV89.6P challenge experiments in rhesus macaques immunized with rVSV expressing Gag and Env (Rose et al., [@B227]). Recombinant VSV vector was shown to induce strong memory CTL responses to HIV-1 Gag and Env in mice, which were significantly amplified by boosting with heterologous recombinant vaccinia virus vectors (Haglund et al., [@B114]). It is postulated that intranasal delivery of rVSV vaccines in combination with IL-12 administered during DNA priming may elicit mucosal immunity for HIV (Egan et al., [@B74], [@B73]). Priming with rVSV-Gag/Pol/Env (VSV-SIVgpe) followed with MVA-Gag/Pol/Env (MVA-SIVgpe) boost was shown to induce strong and long-lived antibody and cellular responses that achieved long-term control of SHIV replication (Schell et al., [@B235]; Van Rompay et al., [@B268]). An ongoing phase 1 trial of rVSV-HIV-1 Gag vaccine (HVTN090) has demonstrated clinical safety and T cell immunogenicity following intramuscular delivery (Fuchs et al., [@B85], [@B86]), although the magnitude of responses was limited and will most likely require priming (or boosting) with suitable vectors.\n\nOther vectors being explored include rhadinovirus (Bilello et al., [@B34]), yellow fever virus (Bonaldo et al., [@B37]), rabies virus (Faul et al., [@B80]), Venezuelan equine encephalitis virus (VEEV) (Caley et al., [@B46]) and Semliki Forrest virus (Schell et al., [@B236]), all of which have shown strong immunogenicity, with some achieving efficacy in NHP challenge protection models. Influenza virus vaccine vectors have also been studied extensively and have been successfully used as delivery vehicles for several experimental HIV vaccines (Li et al., [@B162], [@B161]; Muster et al., [@B191], [@B190]; Garcia-Sastre and Palese, [@B92]; Palese et al., [@B201]; Sexton et al., [@B243]). As natural mucosal pathogens, influenza virus vectors are well-adapted for stimulating robust mucosal and systemic immunity comprising both antibody and cellular immune responses (Garcia-Sastre and Palese, [@B92]; Palese et al., [@B201]; Li et al., [@B161]). Mucosal immunization of mice with chimeric influenza virus vectors expressing the HIV-1 gp120 V3 loop peptide (IHIGPGRAFTYTT) (Li et al., [@B162]) or the gp41 epitope (ELDKWA) (Muster et al., [@B192], [@B191], [@B190]) was shown to induce persistent antibody and CTL responses. Influenza virus vectors might be successfully combined in prime-boost regimens as demonstrated in influenza virus-prime and MVA-boost studies in mice (Gherardi et al., [@B96]), although they have a limited capacity for immunogen insertion.\n\nHeterologous prime-boost strategies for enhanced HIV vaccine efficacy\n=====================================================================\n\nRepeated vaccination in heterologous prime boost approaches employing different vector combinations in a specific order is widely accepted as the most efficient means to induce superior quality and quantity of vaccine-specific immune responses (Li et al., [@B163]; Ramshaw and Ramsay, [@B211]; Estcourt et al., [@B77]; McShane, [@B184]; Newman, [@B195]). Heterologous prime boost regimes allow immune boosting without creating problems of anti-vector immunity. Furthermore, heterologous prime-boosts result in increased frequencies of memory T cells, and it has been shown that the number of immunizations can significantly influence the phenotype of vaccine-specific memory T cells, with secondary and tertiary immunizations generating effector-like memory T cells which preferentially accumulate in non-lymphoid organs (Masopust et al., [@B173]; Nolz and Harty, [@B196]). These findings have huge implications on the quality and potential of mucosal surveillance of cells induced in prime-boost vaccination protocols.\n\nDistinct live viral vectors can be combined in prime-boost regimes to maximize immune responses. In most studies DNA has been used for priming, but recently a number of virus vectors including Adenoviruses, influenza viruses as well as fowlpox and canarypox have been tested in prime-boost regimens. Prime-boost regimens comprising Adenovirus and MVA or heterologous Adenovirus strains have recently been shown to induce both cellular and humoral immune responses to SIV and malaria antigens (Draper et al., [@B70]; Liu et al., [@B167]; Tatsis et al., [@B260]; Barouch et al., [@B20]). In particular, impressive protection against SIV acquisition in rhesus monkeys was achieved following immunization with a SIV~SME543~-Gag/Pol/Env vaccine delivered by Ad26/MVA and Ad35/Ad26 prime-boost regimens which induced a mixture of neutralizing and binding antibody as well as cellular immune responses (Barouch et al., [@B20]). This study further demonstrated induction of both systemic and mucosal immune responses and achieved protection from both acquisition and disease progression, thus providing proof of concept that HIV-1 acquisition and post-infection control might be achieved by improved immunogen design and delivery strategies. Heterologous or homologous regimens comprising DNA/MVA, MVA/Ad26, and MVA/MVA were comparatively less efficacious than Ad26/MVA or Ad35/Ad26, which reduced viral load set-points by greater than 100-fold. A Phase 1 clinical trial (B003/IPCAVD-004) assessing the immunogenicity of various prime-boost combinations of Ad26 and Ad35 is ongoing, and will inform the field on the clinical utility of these two promising human adenovirus vector combinations. Another NHP study employing three doses of plasmid DNA followed with Ad5 to deliver various immunogens comprising SIV-Gag, SIV-Env mosaic immunogens or SIV~mac239~ Env also induced cellular and antibody responses (neutralizing antibodies and ADCC) and achieved significant protection against intra-rectal challenge of rhesus macaques with SIV~smE660~ that was a mismatch of the vaccine strain (Roederer et al., [@B224]). Moreover, superior immunogenicity of prime-boost combinations using DNA/ChAdV63/MVA or ChAdV63/MVA has been demonstrated in a Phase I study (Borthwick et al., [@B38]).\n\nThe success of a viral vector for priming has already been demonstrated in the RV144 study which used ALVAC to prime antibody and T cell responses, followed with a protein boost (Rerks-Ngarm et al., [@B215]). Although priming with DNA has always seemed a better strategy as it focuses the immune response to the immunogen transgene, as opposed to viral vectors which carry multitudes of immunogenic antigens within their backbones, the efficacy of viral-vector priming followed by protein boosting in the RV144 study and the superior immunogenicity of virus-prime/virus-boost in the studies discussed above support the use of viral vectors for both priming and boosting. Therefore, heterologous prime-boost regimens combining DNA, Adenovirus and MVA or ALVAC are likely to achieve efficacy against HIV in clinical trials, although this will require that HIV Env or genes encoding NAb epitopes are included in the immunogen formulations (Barouch et al., [@B20], [@B21]). Preclinical studies investigating the potential of combined chimpanzee adenovirus, MVA and protein prime-boost regimens to deliver immunogens which can stimulate broadly neutralizing antibodies such as BG505 are underway. The success of recombinant adenovirus vector priming followed with MVA boost in inducing high-titre antibodies either on their own or in conjunction with molecular adjuvants has already been proven in preclinical studies of malaria (Draper et al., [@B70]). Possibly the persistence of adenovirus ensures continuous antigen supply which is suitable for B cell priming. It is envisaged that optimal delivery modalities which combine HIV immunogens eliciting BNAbs with those that stimulate strong T cell immunity will achieve enhanced vaccine efficacy. Of course a major caveat of combining strong T cell vectors with antibody-producing immunogens is the possible immune interference of antibody production by these vectors. Nevertheless, this can be optimized perhaps by employing several protein boosts with powerful adjuvants in order to deliver the most balanced immune responses.\n\nPotential vaccine-associated risk of HIV acquisition\n====================================================\n\nThe increased risk of HIV-1 acquisition in the STEP and HVTN505 trial vaccinees despite strong immune responses has raised many unanswered questions as to whether the vaccine delivery modalities, suboptimal potency of the HIV immunogens or other unknown external factors are responsible for vaccine failure. As far as immunogen design, the vaccine construct used in the STEP, Phambili and HVTN505 studies represents one of the most comprehensive immunogens with broad coverage, as it comprised a 6-plasmid DNA and rAd5 vectors expressing Gag/Pol/Nef/Env proteins from multiple clades. Other immunogens based on similar or far less comprehensive HIV protein coverage have also been tested and showed varied degrees of immunogenicity. Thus, an understanding on whether the outcomes of the STEP/Phambili/HVTN505 studies (efficacy, immunogenicity or increased risk of acquisition) would have been different if other delivery vectors (such as DNA/MVA, DNA/ALVAC or DNA/Ad35/Ad26 or even a replicating CMV vector) had been used to deliver the same immunogens in these trials is key for further progression in the field. An Alternative way to look at this is to ask whether the results of RV144 trial would have been worse if Ad5 was used instead of ALVAC, assuming that the prevalence of Ad5 neutralizing antibodies in the RV144 population does not differ significantly from the STEP and Phambili study populations.\n\nThe finding that the vaccine was not at all efficacious amongst men who were circumcised or in uncircumcised men who did not have pre-existing Ad5 immunity raises doubts as to whether efficacy was genuinely hindered by Ad5 serostatus. This is further supported by the results of HVTN505 study which tested only circumcised individuals without Ad5 antibodies, yet no protection was observed. Moreover, the absence of Ad5 antibodies in the HVTN505 study participants (which should in theory allow for higher immunogenicity) was not associated with any significant enhancement of the magnitude and quality of immune responses over those seen in the STEP and Phambili studies. Therefore, Ad5 serostatus can be safely removed from the equation, leaving the only plausible explanation for vaccine failure to be the quality and quantity of immune responses. If this can be fully documented beyond doubt then it implies that either the Ad5 delivery vector or the HIV-1 antigens used were not immunogenic enough to afford protection from infection or post-infection virus control. However, considering that Ad5 is one of the most immunogenic vectors currently available, (and that the immunogen used in these studies was comprehensive and well-designed), this would have serious implications for vaccine design, as it sets the bar really high for new candidate vaccines which would be expected to stimulate responses of extremely higher magnitudes and superior qualitative properties in order to achieve even the minimal efficacy. On a brighter side, this would perhaps instigate intense scrutiny of the current methods used for assessing vaccine immunogenicity in order to standardize and synchronize with those for efficacy measurements.\n\nOne other interesting question is whether (and how) Ad5 sero-positivity is intrinsically associated with HIV acquisition. Although studies of uncircumcised men document increased risk of natural HIV acquisition due to a high frequency of CD4+CCR5+ target cells in the foreskin (Prodger et al., [@B209]), how this relates their Ad5 sero-positivity and titre levels with infection risk is not very clear. However, the fact that the risk of HIV-1 acquisition in the STEP study diminished with time after immunization, and eventually leveled up with placebo recipients (Buchbinder et al., [@B45]) might in actual fact support a role for vaccine-induced immune activation in HIV acquisition (Tenbusch et al., [@B262]). Perhaps this could be as a result of generalized immune activation or induction of activated vaccine-specific HIV-1 targets with mucosal-homing properties. Should this be the case, then this would not be unique to Ad5 vectors alone and it would therefore be expected to equally affect other delivery vectors capable of inducing activated mucosal-homing target cells. However, as there were no notable differences in activated circulating T cells between vaccinees and placebos, it is unlikely that generalized vaccine-induced immune activation played a role, although it remains possible that there could have been significant differences in activated targets at mucosal sites which were not measured.\n\nThis then raises another interesting question as to whether the outcome of the STEP/Phambili/HVTN505 studies would have been significantly worse (or better) had the vaccines been administered mucosally. This question might have two sides to it, in the sense that mucosal delivery would probably have generated higher frequencies of activated HIV targets at the genital mucosae, hence increasing the potential of fuelling infection. On the other hand, induction of robust and polyfunctional effector immune responses at mucosal portals of HIV entry would probably have cleared the incoming HIV before infection became established. Although these questions have no clear cut answers and cannot be addressed retrospectively in the context of the clinical trials they relate to, they however highlight the extreme challenges in HIV vaccine delivery, and new studies designed to directly tackle these issues will be quite informative for future vaccine development research. Studies looking at whether the most promising delivery vectors (and the respective immunogens) can concurrently induce activated HIV-1 target cells that preferentially home to and persist in the genito-rectal and GALT mucosae, and whether or not such vaccine-induced cells become highly permissive to HIV infection will be of particular interest in efforts aimed at limiting the risk of vaccine-induced HIV-1 acquisition and accelerated disease progression.\n\nPerspectives and conclusion\n===========================\n\nIdeally, vectors for HIV-1 vaccines should directly target antigen presenting cells (APCs) or other immune cells to induce long-lived, strong antibody and cellular responses that can broadly disseminate to systemic and mucosal compartments. The vaccine-specific T cells in particular should be broad and contain activated effector, effector memory and central memory phenotypes in various proportions in order to achieve a proper balance between immediate virus clearance and sustained immune-surveillance for long-term protection, as demonstrated by the RhCMV-SIV vaccine which controlled and cleared pathogenic SIV infection (Hansen et al., [@B118], [@B116], [@B117]). Furthermore, vectors which can stimulate polyfunctional CD4+ and CD8+ T cells that act in concert with B cells to inhibit HIV replication through a variety of mechanisms would be more successful than those inducing only mono-functional T cells of either subset alone.\n\nOf particular relevance to protection from infection would be vaccine vectors associated with homing and long-term persistence of vaccine-induced immune responsive cells at the genito-rectal mucosae (Chanzu and Ondondo, [@B54]) as well as other mucosal sites serving as HIV reservoirs. This remains a very important priority in consideration of the significant rapid CD4+ T cell depletion in the intestinal mucosa despite successful HAART (Brenchley et al., [@B42]; Mehandru et al., [@B185]). Thus, vaccine vectors which naturally infect cells within mucosal inductive sites, especially the replication-competent viruses such as adenovirus and influenza virus vectors (Gherardi et al., [@B96]; Sexton et al., [@B243]) which can be administered mucosally to trigger mucosal immunity, would be more suited for HIV vaccine delivery. Alternatively, delivery of vaccines via routes which enhance mucosal immunity (Holmgren et al., [@B130]; Holmgren and Czerkinsky, [@B129]; Czerkinsky and Holmgren, [@B62]) or vectors possessing an inherent ability to induce mucosal immunity in addition to systemic immune responses following parenteral or mucosal vaccine delivery (Moser et al., [@B188]) may be employed. Virosome vectors for instance, possess intrinsic adjuvant properties and a unique ability to target antigen presenting cells, hence have been very successful at inducing protective mucosal immunity in SHIV challenge models (Moser et al., [@B188]; Bomsel et al., [@B36]; Leroux-Roels et al., [@B159]). Other vectors suitable for mucosal vaccine delivery include VEEV (Caley et al., [@B46]). In the absence of mucosal delivery vectors, new delivery technologies such as the \"prime and pull\" approach may be utilized in conjunction with systemic delivery methods to enhance mucosal homing and subsequent immunity (Azizi et al., [@B11]; Shin and Iwasaki, [@B247]; Tregoning et al., [@B264]). In this approach, specialized chemokines are administered in mucosal compartments following parenteral immunization in order to chemo-attract the activated vaccine-specific immune cells from the systemic compartments. Furthermore, use of mucosal adjuvants such as CTB and LT-B (Albu et al., [@B5]; Yuki and Kiyono, [@B298]), pro-inflammatory cytokines (IL-1\u03b1, IL-12, and IL-18) (Belyakov et al., [@B27]; Bradney et al., [@B40]; Albu et al., [@B5]) or immunostimulatory CpG motifs (Horner et al., [@B134]; Dumais et al., [@B71]; Daftarian et al., [@B63]; Jiang et al., [@B140]) which target recruitment of immune cells to the mucosal sites would be useful. Co-delivery of vaccines with genes encoding CCL19 and CCL28 was also shown to enhance HIV-1-specific T and B cell responses in the systemic as well as mucosal compartments (Hu et al., [@B136]).\n\nIn consideration of both safety and immunogenicity goals as already discussed, and with particular emphasis on the pivotal role of CTL responses in controlling HIV replication, it seems that non-replicating viral vectors with lower sero-prevalence would be highly desirable, mainly due to excellent safety profiles and potent adjuvant effect allowing for induction of very strong, high quality and long-lived cellular and humoral immunity. However, although safety and reduced immune interference would be guaranteed, a major caveat would be that these lower sero-prevalence vectors may not be adequately immunogenic. Perhaps these vectors can be re-engineered to improve their immunogenic potential. For instance, the immunogenicity of vectors such as MVA and NYVAC can be improved by removal of genes associated with immune evasion which counteract immune responses to the vaccine (Kibler et al., [@B149]; Gomez et al., [@B106]; Garcia-Arriaza et al., [@B90]). In other cases, addition of cytokine-encoding genes such as type 1 interferons, IL-12 or GM-CSF can enhance vaccine efficacy (Gherardi et al., [@B99], [@B98]; Rodriguez et al., [@B223]; Ramshaw and Ramsay, [@B211]; Bayer et al., [@B24]). Furthermore, chemokines such as CCL3 which recruits professional APCs can be co-delivered with HIV antigens to enhance vaccine immunogenicity (Lietz et al., [@B164]).\n\nAlternatively, vectors capable of inducing substantial immunogenicity in the presence of pre-existing natural or vaccine-induced anti-vector immunity may be worth considering, although it is expected that finding highly attenuated vectors which are safe and remain immunologically potent will be equally challenging. As discussed earlier, combining some of the most promising vectors in heterologous prime-boost regimens will significantly enhance the quantity, quality and protective efficacy of immune responses. However, in consideration of the possible catastrophic effects of elevated immune activation likely to arise from various vector combinations, it would be expected that suitable HIV vaccine vectors maintain lower levels of immune activation to limit the numbers of activated HIV-1 targets (Perreau et al., [@B205]; Benlahrech et al., [@B29]) likely to fuel infection in the event of exposure. Furthermore, it is documented that in the absence of a very strong protective immune responses to counteract the incoming virus, the presence of vaccine-specific T cells which are activated and hence more susceptible to infection may increase the risk of acquisition (Tenbusch et al., [@B262]). Whether it is possible to achieve potent immunostimulatory capacity but with minimal immune activation still remains a subject of intense investigation.\n\nWhen safety and versatility are considered, and in full view of the enormous technology advancements in DNA plasmid formulations and delivery, in conjunction with other immunomodulatory interventions such as SAP depletion and use of molecular adjuvants, recombinant DNA vaccines remain very attractive, although efforts to improve stimulation of long-lived effector/memory CD8+ T cell phenotypes are still needed to achieve long-term efficacy. Undoubtedly, repeated immunizations or combining DNA vaccines with persistent (replicating) vectors or vectors with slow immunogen release features would induce durable immunity. Nonetheless, replicating vectors with lower sero-prevalence and minimal pathogenicity (Rose et al., [@B227]; Kawada et al., [@B144]; Fuchs et al., [@B86]; Liu et al., [@B166]) are being considered as they would provide a persistent pool of HIV vaccine-specific effector memory phenotype cytotoxic T cells which are critical for long-term protection from disease progression (Hansen et al., [@B118], [@B116], [@B117]). Such effector memory responses would otherwise be expected to wane with time, in the absence of antigen. Replicating vectors may also be better-suited for induction of broadly neutralizing antibodies since persisting expression of the Env antigens is likely to drive high levels of somatic mutations required for affinity maturation of these antibodies (van Gils and Sanders, [@B267]). A new strategy that has been proven to induce durable and protective antibody responses in humanized mice challenged with high doses of diverse HIV strains is vectored immunoprophylaxis, which involves insertion of immunoglobulin genes into viral vectors such as the adeno-associated virus (AAV) to provide long-term expression of neutralizing antibodies (Balazs et al., [@B14], [@B15]). Moreover, inclusion of Th2 cytokines such as IL-4, IL-5, and IL-6 which enhance B cell maturation into long-lived antibody secreting cells is yet another strategy already shown to induce high titres of neutralizing antibodies which protected mice from Friend Virus (Ohs et al., [@B198]). Other possible strategies include use of lentiviral vectors expressing B cell receptor genes encoding neutralizing antibodies to HIV-1 to transduce haematopoietic stem cells (Luo et al., [@B169]).\n\nSince optimum induction of immune responses to vaccines strongly depends on innate immune triggering as well as the levels of transgene expression, vectors with natural adjuvant properties and therefore capable of strongly inducing innate immunity are particularly immunogenic and thus highly desirable. However, care must be taken to balance between strong innate function stimulation and the potential risk of inducing potent stimulation of immuno-pathological effects, including immune hyper-activation.\n\nIn conclusion, a successful vaccine for HIV will have to stimulate potent antibody and CTL responses broad enough to cover multiple HIV variants and with potential to neutralize, bind or suppress HIV-1 replication for sustained (possibly infinite) lengths of time. Of utmost importance, however is generation of vaccine-specific immune responses in the genito-rectal mucosae, the major portals of HIV entry. Emerging evidence strongly suggests that non-pathogenic, low-level replicating viral vectors which can mimic live attenuated vaccines, but with low sero-prevalence might be the best way to achieve HIV vaccine efficacy. As these vectors persist long after immunization, they are capable of inducing and maintaining effector/memory CTLs for continued immune surveillance that is necessary to protect from infection, disease progression and to clear or prevent establishment of latent reservoirs. Thus, to achieve protective efficacy HIV vaccine development will need ingenious state of the art technologies to create the very best of T cell and antibody immunogens, delivered by the most potent but safe vectors possessing remarkably high capacity to induce both systemic and mucosal immunity, but without significant immune activation likely to fuel HIV acquisition. Recent significant advances in vaccine delivery technologies and HIV immunogen design provide hope that this is not far from reality.\n\nConflict of interest statement\n------------------------------\n\nThe author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nBeatrice O. Ondondo is a senior HIV vaccine development immunologist employed on a research grant supported by MRC UK.\n\n[^1]: Edited by: Shuo Li, The University of Melbourne, Australia\n\n[^2]: Reviewed by: Andrew Lucas, Murdoch University, Australia; Cody Charles Allison, Walter and Eliza Hall Institute, Australia\n\n[^3]: This article was submitted to Microbial Immunology, a section of the journal Frontiers in Microbiology.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nIn the Western world, the main risk factor for developing oropharyngeal squamous cell carcinoma (OPSCC) is infection with high-risk human papillomavirus (HPV); while, a smaller proportion is related to the high consumption of alcohol and smoking tobacco. Patients with HPV-associated OPSCCs have improved survival, which may be related to a different mutational profile \\[[@R1], [@R2]\\], histopathology \\[[@R3]\\], and clinical features \\[[@R4]\\]. The first reports of improved survival for HPV-positive (HPV+) OPSCCs were published in the early 2000\\'s \\[[@R5], [@R6]\\]. Shortly after, p16 (a surrogate-marker for HPV- infection) was demonstrated to have prognostic value \\[[@R7], [@R8]\\]. The importance of HPV and p16 has since been confirmed, and the combination (e.g. HPV/p16-status) has shown better prognostication \\[[@R9]\\]. However, long-term results are missing for overall survival (OS), time to progression (TTP), and survival after progression (SAP), especially their relationship to clinico-pathological characteristics from a large, population-based cohort.\n\nProgression is a strong predictor of survival in patients with OPSCCs. The subgroup of patients who are HPV and p16-positive have a lower risk of recurrence and improved SAP compared with patients without HPV/p16 \\[[@R10]--[@R12]\\]. These findings are based on selected patients from small cohort-studies examining only p16 or HPV without a detailed account of progression. Therefore, HPV+ OPSCC may have an increased OS due to a better response to salvage therapy (i.e. SAP). A nomogram is a graphic demonstration of a statistical model for calculating the cumulative effect of weighted variables on the probability of a particular outcome, and enables continuous estimation of the probability of specific outcomes (i.e. death or progression). Furthermore, they can combine multiple independent variables while considering the prognostic weight for each variable when calculating the probability of an outcome. Although nomograms have been developed to predict clinical end-points for patients with several types of malignancies, to the best of our knowledge, this is the first study addressing OS, TTP, and SAP in OPSCCs \\[[@R13]--[@R17]\\].\n\nTo avoid selection-bias, we aimed to include all patients diagnosed with OPSCC in the Eastern part of Denmark, which has an approximate population size of 2.4 million inhabitants. The cohort is well described and includes detailed information on both tumour (HPV, p16, stage, site) as well as clinical characteristics (follow-up, treatment, progression) with virtually no patients lost to follow-up. The primary aim was to identify tumour- and patient-factors associated with OS, TTP, and SAP. Due to our unique, complete population-based cohort spanning 15-years, we were able to construct and validate reliable, predictive nomograms.\n\nRESULTS {#s2}\n=======\n\nPopulation demographics {#s2_1}\n-----------------------\n\nA total of 1,542 patients diagnosed with OPSCC between 2000--2014 were included, and their characteristics are summarized in Table [1](#T1){ref-type=\"table\"} with univariate hazard ratios for OS. The majority of patients were males (72%), with a median age of 60 years at diagnosis, and typically presented with tumours in an advanced nodal stage, but with small primary tumours (Table [1](#T1){ref-type=\"table\"}). Of the total cohort, 54% were stage III or IV, 61% had T1 or T2 primary tumours, and 60% had N2 or N3 nodal classification (Table [1](#T1){ref-type=\"table\"}). The median follow-up for patients alive at the last date of follow-up was 4.0 years (range, 0.8 to 15.8 years). Tobacco use was high with a median number of pack-years of 27 (mean 30). Subsequently, we examined which of the variables were correlated (Figure [1D](#F1){ref-type=\"fig\"}). It was observed that the subgroup of patients with an HPV+/p16+ tumour correlated with: a shorter smoking history, higher N-stage, better performance status, lower T-stage, and younger age (Figure [1D](#F1){ref-type=\"fig\"}).\n\n###### Patient characteristics and their relationship to overall survival\n\n![](oncotarget-07-71761-t001)\n\nFootnote: The hazard ratios for age, year of diagnosis, and pack years represents the hazard ratio per year increase. RT: radiotherapy. BSCC: Base of tongue squamous cell carcinoma. STSCC and NSTSCC: Specified- and non-specified tonsillar squamous cell carcinomas.\n\n![Kaplan-Meier curves for (**A**) overall survival, (**B**) time to progression, and (**C**) survival after progression. Correlation heatmap between variables (**D**). Comparison between survival in an age- and gender-matched background population and the HPV+/p16+ subgroup (**E**) and the HPV\u2212/p16\u2212 subgroup (**F**) based on T classification of tumours.](oncotarget-07-71761-g001){#F1}\n\nUnivariate and multivariate analysis of factors influencing overall survival {#s2_2}\n----------------------------------------------------------------------------\n\nIn the OS analysis, 1,524 patients (18 were lost to follow-up) were included, with a total of 690 deaths. In the total cohort, 1,336 patients were treated with curative regimes; of these patients, 543 died during follow-up. The subgroup of HPV+/p16+ had a markedly better prognosis compared with the other combinations of HPV/p16 (Table [1](#T1){ref-type=\"table\"} and Figure [1A](#F1){ref-type=\"fig\"}) (*p* \\< 0.05). Other factors significantly associated with a poorer OS in the univariate analysis were: pack years of smoking, treatment, tumour location, and T, N, and M classification (Table [1](#T1){ref-type=\"table\"}). We subsequently constructed a multivariate Cox regression model to examine which factors independently predicted OS. The HPV/p16 status was an independent predictor for OS, even when adjusted for T-stage, N-stage, treatment, smoking history, age, and performance status ([Supplementary Table S1](#SD1){ref-type=\"supplementary-material\"}). To better illustrate any excess mortality in the HPV+/p16+ and HPV\u2212/p16\u2212 subgroups, Kaplan-Meier survival curves were compared with the mortality rates of the background population matched for age, gender, and calendar period (Figure [1E and 1F](#F1){ref-type=\"fig\"}). Interestingly, patient with stage T1-T2 tumours in the HPV+/p16+ subgroup had almost no excess mortality compared with the background population (Figure [1E](#F1){ref-type=\"fig\"}), which was in stark contrast to the HPV\u2212/p16\u2212 subgroup (Figure [1F](#F1){ref-type=\"fig\"}).\n\nUnivariate and multivariate analysis of factors influencing time to progression {#s2_3}\n-------------------------------------------------------------------------------\n\nIn total, 376 patients experienced a progression, with 153 (19%) in the HPV+/p16+ subgroup vs. 182 (36%) in the HPV\u2212/p16\u2212 subgroup (*P* \\< 0.01). The subgroup of HPV+/p16+ had a significantly longer TTP compared with the other combinations of HPV/p16 (Figure [1B](#F1){ref-type=\"fig\"}) (*P* \\< 0.05). The median TTP for the total cohort was 9.7 months. The median TTP was 13 months for the HPV+/p16+ subgroup and 8.5 months (*P* \\< 0.05) for the HPV\u2212/p16\u2212 subgroup. Median time to loco-regional progression was 7.6 months for the HPV+/p16 subgroup and 7.4 months for the HPV\u2212/p16\u2212 group. The corresponding figures for distant progression were 18 and 11 months, respectively, (*P* \\< 0.05). The univariate analysis of factors influencing TTP included: pack years of smoking, HPV/16 status, and T, N, and M classification of tumours ([Supplementary Table S2](#SD1){ref-type=\"supplementary-material\"}). A multivariate Cox regression model identified HPV/p16 status as an independent predictor of TTP with a hazard ratio of 3.0 (95% CI: 2.3--3.8) in the HPV\u2212/p16\u2212 subgroup compared with the HPV+/p16+ subgroup, after adjustment ([Supplementary Table S3](#SD1){ref-type=\"supplementary-material\"}).\n\nUnivariate and multivariate analysis of factors influencing survival after progression {#s2_4}\n--------------------------------------------------------------------------------------\n\nThe median SAP was 13 months in the HPV+/p16+ subgroup vs. 6 months in the HPV\u2212/p16\u2212 subgroup (*P* \\< 0.05). The HPV+/p16+ status also independently influenced SAP positively with a hazard ratio of 2.2 (95% CI: 1.6--2.9) compared with the HPV--/p16-- subgroup ([Supplementary Table S4](#SD1){ref-type=\"supplementary-material\"}), when adjusted for other significant covariates. Performance status and progression location also independently influenced SAP ([Supplementary Table S4](#SD1){ref-type=\"supplementary-material\"}).\n\nPredictive nomograms {#s2_5}\n--------------------\n\nNomograms (Figure [2A and 2B](#F2){ref-type=\"fig\"}) were constructed to predict survival using the independent covariates identified in the multivariate Cox regression models. The median total points for the 1,523 patients used to fit the multivariate Cox model for OS was 101 (range, 0--421). The nomogram is used by totalling the points identified on the top scale for each independent covariate which is identified on the total points scale to identify the estimated median survival time (years) and the probability of 1-, 5- and 10-year survival.\n\n![Nomogram for overall survival (A) and time to progression (B)\\\nThe nomogram is used by totalling the points identified on the top scale for each independent covariate. The total points scale is used to identify the estimated median survival time (years) and the probability of 1-, 5-, and 10-year survival.](oncotarget-07-71761-g002){#F2}\n\nThe bias-corrected concordance index for this nomogram was 0.79 based on the bootstrap validated Cox model. The calibration curve ([Supplementary Figure S1](#SD1){ref-type=\"supplementary-material\"}) illustrates how the predictions from the nomogram compare with actual outcomes. The corresponding information for the nomogram for TTP can be found in [Supplementary Figure S2](#SD1){ref-type=\"supplementary-material\"}.\n\nPrognostic index {#s2_6}\n----------------\n\nA simplified prognostic index was constructed for the 5-year OS using seven independent prognostic factors. The index score is based on the sum total of factors, with points given for each of the following: treatment, HPV/p16-status, age, pack years, T-status, N-status, and performance status. Risk is assigned as follows: index score 0 to 4, low (\u2265 80% probability of surviving 5 years); index score 5 to 7, intermediate (\\< 80% and \\> 20% probability of surviving 5 years); and index score 8 or greater, high risk (\u2264 20% probability of surviving 5 years) (Table [2](#T2){ref-type=\"table\"}).\n\n###### Prognostic index for overall survival\n\n  Points               0                     1            2                           3              \n  -------------------- --------------------- ------------ --------------------------- -------------- --------------\n  Treatment            RT and chemotherapy   RT           Palliative                  NA             No treatment\n  Age                  0--44                 45--64       65-89                       90 and above   NA\n  HPV/p16              HPV+/p16+             HPV--/p16+   HPV--/p16-- or HPV+/p16--   NA             NA\n  T classification     T1                    T2-T3        T4                          NA             NA\n  N classification     N0                    N1-N2        N3                          NA             NA\n  Pack years           0--59                 59--179      180 and above                              \n  Performance status   0--1                  2--3         NA                          NA             4\n\nFootnote: It is possible to obtain between 0 and 19 points. Low-risk is from 0--4 points (\u2265 80 % risk of being alive after 5 years), medium risk 5--7 points (between \\< 80 and \\> 30% chance of being alive after 5 years), and high-risk \\> 7 points (\u2264 30% change of being alive after 5 years).\n\nDISCUSSION {#s3}\n==========\n\nThis study included the largest-to-date collection of 1,542 OPSCCs from a consecutive, population-based, non-selected cohort and assessed OS, TTP, and SAP as well as their relation to multiple clinico-pathological characteristics. From these results, we constructed and internally validated nomograms to predict OS and TTP. Furthermore, we established a simplified prognostic index based on the independent covariates for OS for use in categorizing patients into low-, intermediate-, and high-risk groups. Besides being useful for interpretation of the underlying Cox model, the nomogram combines independent prognostic factors and considers the importance of each on the probability of survival and progression. The prognostic models we present may facilitate discussions in clinical settings and aid in identifying low-risk patients that could be candidates for de-escalation therapy, as well as high-risk patients eligible for new clinical trials. In particular, HPV+/p16+ patients with T1-T2 tumours should be considered candidates for de-escalation therapy, as we demonstrated their survival is similar to the background population, and this would avoid the morbidity associated with therapy. Notably, at least nine de-escalation treatment trials are on-going or finishing \\[[@R18]\\]. Our nomograms are likely to be applicable to these and future trials, as we reported similar 5-year survival or progression rates as studies in North America;\\[[@R19]--[@R21]\\] Western \\[[@R22]--[@R24]\\], Southern \\[[@R25]\\], and Northern Europe \\[[@R6], [@R26]\\]; Australia \\[[@R27]\\]; and China \\[[@R28]\\].\n\nAn estimated 10% of all head and neck SCCs are p16+, but HPV-- caused by alternative cellular misconfigurations leading to p16-overexpression.\\[[@R29], [@R30]\\] Therefore, it is suboptimal to include patients in de-escalation trials based on evaluation of a single biomarker (i.e. HPV or p16 alone) due to the risk of misclassification of tumours and thereby misallocation of patients with an undesired prognosis. \\[[@R31], [@R32]\\] One of the main findings in this study includes the identification of HPV+/p16+ as an important and independent predictor for improved OS, TTP, and SAP. Importantly, even though the subgroup of HPV+/p16+ patients was more likely to have a shorter smoking history and a better performance status, even when adjusted for these covariates, HPV+/p16+ was a significant and strong predictor for OS. The calibration for this model was robust and could explain 79% of the observed variability in OS in the cohort. Our findings are in accordance with similar studies addressing survival and progression in HPV+ or p16+; although, we present the first results of a population-based cohort with long-term follow-up. While another smaller study in a region with low HPV-prevalence (below 20%) constructed nomograms for OS and TTP in OPSCC, this study was not population-based and did not address SAP.\\[[@R33]\\] Furthermore, this study did not convincingly demonstrate which patients were lost to inclusion and follow-up, resulting in less robust external validity.\n\nOur nomograms support the latest AHR-stage classification proposed by O\\'sullivan *et al.* \\[[@R34]\\] and comply with the notion that T1--T2 and N0--N2 show significantly better OS, TTP, and SAP compared with \u2265 T3 and \u2265 N3 tumours, indicating that down-staging of N-classified OPSCCs is reasonable. Noticeably, our findings, in contrast to O\\'sullivan *et al*, include both HPV and p16 status and patients with distant disease. Our models encourage further studies to better understand whether all N-classified tumours are eligible for down-staging. Although our findings rely on a population-based cohort and selection-bias is minimized, the nomograms have certain limitations. It would be most appropriate to apply the model as the last step in a clinical setting, and the models are not necessarily applicable for changing patients\\' decisions not to accept further treatment since we do not know her/his response to therapy. With respect to accuracy, the CIs at the various predicted probabilities of recurrence should be considered if using these nomograms in clinical settings. Although this model is internally validated, it could be strengthened by external validation, e.g. in similar population-based cohorts. Future work might focus on validating our results and incorporating additional prognostic factors including robotic, curative surgery, and specific salvage treatment for relapsed disease as well as further outcome measures such as time to treatment or to pathological evaluation.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nWe included patients diagnosed with OPSCC in Eastern Denmark between 2000--2014. The reasons for exclusion are illustrated in Figure [3](#F3){ref-type=\"fig\"}. We targeted all patients diagnosed with OPSCC in the Eastern part of Denmark cohort with detailed information on tumour (HPV, p16, stage, site) and clinical characteristics (follow-up, treatment, progression). Using the unique resident-code from the Danish Civil Registration System, we linked two national registries, the prospectively maintained Danish Head and Neck Cancer Group (DAHANCA) \\[[@R35]\\] database and the Danish Pathology Data Registry (DPDR) \\[[@R36]\\], to identify patients. Patient-data were retrieved from these databases as well as from medical records. The patient cohort from 2000 to 2010 has previously been described \\[[@R37], [@R38]\\].\n\n![CONSORT flow-diagram](oncotarget-07-71761-g003){#F3}\n\nTumour-site, histology, tumour-grade, HPV- PCR, and p16 immunohistochemistry {#s4_1}\n----------------------------------------------------------------------------\n\nAn H&E-stained section of each tumour was reviewed by an expert head and neck pathologist. Based on the finding of tissue-specific structures and clinical information, tumours were divided into base of tongue squamous cell carcinoma (BSCC) and specified- and non-specified tonsillar squamous cell carcinomas (STSCC and NSTSCC). The latter was according to the certainty of origin. The pathologist validated the diagnosis of OPSCC. The p16 staining was considered positive if there was a strong and diffuse nuclear and cytoplasmic reaction in more than 75% of the tumour cells \\[[@R39]\\]. The HPV DNA-PCR was performed with DNA isolated from FFPE tumour sections. Methods for immunohistochemistry and HPV-PCR have previously been described \\[[@R37], [@R38]\\].\n\nOverall survival, time to progression, and statistics {#s4_2}\n-----------------------------------------------------\n\nThe date of the last follow-up was the 1st of November 2015. The OS was defined as the time from diagnosis of OPSCC to death due to any cause. Data on vital status (death) stems from the real-time, national, patient-administrative *Green System (GSOpen).* Progression was only considered if it was biopsy-verified. Data on progression stem from the DPDR. The TTP was defined as the time from diagnosis of OPSCC to time of progression at any site. In the analysis of TTP, patients were censored either at the last date of follow-up or at time of death. Patients alive at the last date of follow-up were censored in survival analyses for OS and SAP. The date of diagnosis was used to reflect the date of treatment, since the vast majority of Danish patients initiate treatment within 1 month \\[[@R40], [@R41]\\]. The SAP was defined as the time from progression to death due to any cause. Curative radiotherapy regimens consisted of 66--68 GY, divided into 33--34 fractions given 6 days a week. From 2007, stage III-IV patients were offered concomitant cisplatin, if tolerated. Covariates available for adjustment are described in Table [1](#T1){ref-type=\"table\"}. The following variables were coded as continuous variables in the survival analyses: age and year of diagnosis. The remaining variables were coded as categorical variables unless otherwise stated. Data on smoking (20 cigarettes per day for 1 year = 1 pack year) were retrieved from the DAHANCA database or medical files. Kaplan-Meier curves were used to illustrate survival differences and significant differences were assessed with log-rank tests. Hazard ratios for death, OS, and TTP were calculated by univariate Cox-regression with log-rank tests for each parameter (Table [1](#T1){ref-type=\"table\"}). To evaluate which covariates independently influenced survival, we performed multivariate cox regression analyses with the same factors used in the univariate analyses (Table [3](#T3){ref-type=\"table\"} and [Supplementary Tables S2 and S3](#SD1){ref-type=\"supplementary-material\"}) with backward elimination, until only significant factors remained. The stopping rule in the backward elimination of factors was based on Akaike\\'s Information Criteria, with the statistics based on the pooled residual chi-square of the model with the R package rms and the function fastbw \\[[@R42]\\]. The models were internally validated with 200 bootstrappings. The difference in Somers\\' D between the training and test set was used to evaluate the optimism in the predictive accuracy (i.e. a measure of the difference between observed and predicted values). Somers\\' D can range from \u22121 (all pairs disagree) to 1 (all pairs agree) and can be converted to the more well-known AUC via the following formula, AUC = Dyx/2 + 0.5. The calibration of the models was subsequently tested using the rms package in R. To test whether the assumption of proportional hazards was violated, we plotted Schoenfeld residuals for the univariate analyses \\[[@R43]\\]. Additionally, the multivariate models for survival were examined for violations of the proportional hazards assumption with the function cox.zph. None of the final models violated the proportional hazards assumption. To test for correlations between the covariates, we used Spearman\\'s rank correlation (Figure [1D](#F1){ref-type=\"fig\"}) \\[[@R44]\\]. The analysis comparing survival in the cohort with the background population was performed for each year of observation according to the Kaplan-Meier survival rates in the cohort. The survival analysis for the background population was performed for a gender and age-matched population in the same calendar-years as the cohort with mortality figures from the Danish National Department of Statistics (<http://www.statistikbanken.dk/statbank5a/default.asp?w=1344>). Missing data were left out of the analysis, and variables with missing values are illustrated in Table [1](#T1){ref-type=\"table\"}. A *p* value \\< 0.05 was considered significant. Data were analysed with SPSS version 23 (SPSS Inc., Chicago, IL, USA) and *R* statistics version 3.03.\n\n###### Independent covariates for overall survival in the multivariate Cox regression model\n\n                          Hazard ratio for death   Lower CI   Upper CI   *P*\n  ----------------------- ------------------------ ---------- ---------- -----------\n  **T classification**                                                   \n  T1                      Ref                                            \n  T2                      1.55                     1.13       2.11       0.0060\n  T3                      2.32                     1.69       3.19       \\< 0.0001\n  T3                      3.59                     2.52       5.11       \\< 0.0001\n  **N classification**                                                   \n  N0                      Ref                                            \n  N1                      1.59                     1.17       2.17       0.0034\n  N2                      2.00                     1.54       2.60       \\< 0.0001\n  N3                      3.59                     2.48       5.19       \\< 0.0001\n  **Treatment**                                                          \n  Radiotherapy            Ref                                            \n  Chemoradiotherapy       0.7                      0.6        0.9        0.0042\n  Palliative              2.4                      1.7        3.5        \\< 0.0001\n  No treatment            7.2                      4.2        12.1       \\< 0.0001\n  **HPV/p16**                                                            \n  HPV+/p16+               Ref                                            \n  HPV+/p16--              2.5                      1.8        3.7        \\< 0.0001\n  HPV--/p16+              2.1                      1.4        3.0        0.0001\n  HPV--/p16--             3.4                      2.7        4.3        \\< 0.0001\n  **Age**                 1.03                     1.01       1.04       \\< 0.0001\n  **Pack years**          1.005                    1.002      1.008      0.0015\n  **Performance score**                                                  \n  0                       Ref                                            \n  1                       1.3                      1.1        1.7        0.0105\n  2                       2.1                      1.5        2.9        \\< 0.0001\n  3                       1.6                      0.9        2.8        0.0905\n  4                       7.9                      2.1        29.9       0.0024\n\nFootnote: CI, 95% confidence interval. The hazard ratio for age and pack years represents the increase in hazard ratio per 1 year increase.\n\nValidation and calibration of multivariate cox regression models {#s4_3}\n----------------------------------------------------------------\n\nWhen performing multivariate analysis of survival, it is of interest to examine how well the models can explain the variation in survival observed in the cohort. The Somers\\' D for the final model for OS was 0.60, which corresponds to an AUC of 0.80. This was validated with bootstrapping to prevent possible over fitting using the Somers\\' D rank correlation for the predicted log hazard and observed survival time. The optimism in the model was 0.0064, and the expected bias-corrected AUC was 0.79, suggesting that the model was not over fitted. Finally, the models were evaluated for calibration accuracy in predicting the probability of surviving 5 years. This was performed with bootstrapping to estimate the optimism from the models for how the predicted 5-year survival estimates compared with the observed survival estimates ([Supplementary Figure S1](#SD1){ref-type=\"supplementary-material\"}). The corresponding AUC and optimism in the multivariate Cox regression model for TTP were 0.68 and 0.010, respectively, and the calibration curve is shown in [Supplementary Figure S2](#SD1){ref-type=\"supplementary-material\"}.\n\nSUPPLEMENTARY MATERIALS FIGURES AND TABLES {#s5}\n==========================================\n\n**CONFLICTS OF INTEREST**\n\nThe authors declare no conflicts of interest.\n\n**FUNDING**\n\nCGL is funded by the Candys Foundation and Kr\u00e6ftfonden (The Cancer Foundation).\n\n**Authors\\' contributions**\n\nCGL, DHJ, and CVB designed the study. CGL, EG, and ALC extracted data from files and registries. EG and ALC performed the laboratory analysis. KK, CHO, and LA performed the pathology review. DHJ performed the statistical analysis. CGL, DHJ, and CVB led the writing. All authors provided conceptual input, interpreted the findings, and contributed in significant ways to the final article.\n"}
+{"text": "1. INTRODUCTION\n===============\n\nThe 4th annual Ontario Thoracic Cancer Conference was held at Niagara-on-the-Lake, April 17--19, 2009, bringing together health care professionals interested in thoracic oncology in the province of Ontario. Attendees at this conference spanned the disciplines of surgical, radiation, and medical oncology, respirology, pathology, nursing, support services, and radiation therapy. Advocates for lung cancer patients were also in attendance. For the first time this year, a session on the management of esophageal cancer was presented, as were a session on new innovations in radiation therapy and an update on molecular targeted therapy. A poster session highlighted research work being done by trainees, whose abstracts are published in the [Appendix](#app1-co16-5-111){ref-type=\"app\"} to this report.\n\n2. HIGHLIGHTS\n=============\n\nThis year's meeting highlighted three themes:\n\n-   Innovations in the management of lung cancer\n\n-   Controversies in the management of esophageal cancer\n\n-   Molecular targeted therapies for lung cancer\n\n2.1 Innovations in Lung Cancer Management\n-----------------------------------------\n\n### 2.1.1 Radiation Therapy\n\nProfessor Jake Van Dyk, from the University of Western Ontario and the London Health Sciences Centre, delivered the first keynote address, \"New Advances in Radiation Therapy for Non-Small Cell Lung Cancer\" ([nsclc]{.smallcaps}).\n\nThree major problems are encountered in the treatment of [nsclc]{.smallcaps}:\n\n-   Accurate delineation of the target\n\n-   Precision delivery of high-dose radiation to the target\n\n-   Minimization of radiation exposure to surrounding normal critical tissues---for example, normal lung and esophagus\n\nThe potential solutions for these problems involve better imaging by incorporating positron-emission tomography ([pet]{.smallcaps}) for accurate localization and by avoiding geometric misses [@b1-co16-5-111]--[@b3-co16-5-111], by using image-guided radiation therapy [@b4-co16-5-111] for dose escalation and tumour adaptive changes to improve local control [@b5-co16-5-111],[@b6-co16-5-111], and by minimizing collateral damage to critical tissues.\n\nInnovative new radiation delivery treatment systems that include the use of tomotherapy [@b7-co16-5-111], robotic radiosurgery, stereotactic body radiation treatment, and magnetic resonance--guided radiation have been evaluated. The use of respiratory gating methods to minimize exposure of normal lung tissue is important in the development of these new, highly conformal radiation techniques, so as to reduce the risk of radiation pneumonitis as the radiation dose is escalated beyond traditional levels.\n\nDr. Stewart Gaede, from the London Health Sciences Centre, spoke on \"Respiratory Gating in Lung Cancer Applications, Including 4D CT--Based Treatment Planning.\" Respiratory management techniques that include tumour tracking methods, tumour immobilization, breath-hold methods, and respiratory gating were reviewed. At the London Regional Cancer Centre, 96 patients have been treated using respiratory-gated radiotherapy with either liver metastases or lung tumours. Based on evaluation of dose--volume histogram parameters, respiratory gating reduces the amounts of normal lung and liver that receive a significant dose. However, optimal techniques are still being investigated to correlate the use of external marker motion with internal tumour or organ motion [@b8-co16-5-111]. Despite the uncertainties of dose distribution and organ tracking, the use of respiratory gating is a promising strategy to aid in dose escalation, in the avoidance of critical structures influenced by respiration, and in the delivery of intensity-modulated radiation treatments.\n\n### 2.1.2 Reducing Wait Times\n\nDr. Carol Sawka, Vice President, Clinical Programs and Quality Initiatives, Cancer Care Ontario ([cco]{.smallcaps}), gave the second keynote address on \"Access to Cancer Services in Ontario: A Progress Report.\"\n\nAccess to care can be defined as \"equitable and timely access to appropriate care when needed.\" The patient journey from onset of symptoms to treatment is influenced by many factors, which include the type and stage of cancer, the treatment and services selected by the physician, and patient choices for therapy. Cancer Care Ontario is dedicated to improving the patient experience, and the organization uses a number of strategies to improve the performance of cancer services. It has created the Cancer System Quality Index ([csqi]{.smallcaps}) [@b9-co16-5-111], a Web-based public reporting tool that serves as a system-wide monitor tracking the quality and consistency of key cancer services that span the spectrum from prevention to end-of-life care. The [csqi]{.smallcaps} has about 30 key indicators, and each indicator is a specific measure of progress against one of six goals that help focus efforts to improve the cancer system in Ontario. One of the key indicators is wait times, whose principle is that wait time targets should be based on the biologic behaviour of the cancer. Wait times for surgical and radiation therapy are improving, but access to systemic therapy still needs improvement. The Lung Cancer Disease Pathway Management initiative at [cco]{.smallcaps} will focus on the patient journey to make improvements in the cancer system. It will bring together experts focused on lung cancer to evaluate the continuum of care, to map the patient journey, to evaluate the system's performance, and to develop an integrated improvement program.\n\nMethods to improve the performance of the system have included the development of thoracic diagnostic assessments units ([dau]{.smallcaps}s) as presented by Dr. Matthew Kilmurry and Ms. Jennifer Parkins from the Grand River Regional Cancer Centre.\n\nThe [dau]{.smallcaps} was a joint venture between the regional cancer centre and its two host hospitals in the Waterloo--Wellington Local Health Integration Network ([lhin]{.smallcaps}). The [lhin]{.smallcaps}'s lung cancer patients have a long wait time and uncoordinated care pathways for referrals and diagnostic imaging. The [dau]{.smallcaps} provides timely access to diagnosis and treatment, interdisciplinary focused care, multidisciplinary case conferencing, and implementation of evidence-based care. The most common diagnostic test ordered is a computed tomography ([ct]{.smallcaps})--guided biopsy, followed by bronchoscopy, mediastinoscopy, and [pet]{.smallcaps} imaging. To reduce the wait time from [ct]{.smallcaps} to diagnosis, pre-booked slots were made available in medical imaging. As a result, wait time was reduced to 27 days from 74 days, which parallels the initial experience of the Time to Treat initiative at the Toronto East General Hospital [@b10-co16-5-111]. Future plans include integrating the [dau]{.smallcaps} into the surgical oncology program, expanding the nursing role, and further engaging regional physicians. Cancer Care Ontario has set up a guideline on the establishment of diagnostic programs [@b11-co16-5-111].\n\nA diagnostic test commonly requested through the [dau]{.smallcaps} is [pet]{.smallcaps} imaging. An update on the role of [pet]{.smallcaps} in staging and managing lung cancer was presented by Dr. Yee Ung, from the Odette Cancer Centre. In lung cancer, [pet]{.smallcaps} imaging shows high sensitivity and specificity over conventional imaging, a finding that has been systematically reviewed[@b12-co16-5-111]. More recently, two clinical trials in lung cancer by the Ontario Clinical Oncology Group have shown the utility of [pet]{.smallcaps} for staging the mediastinum in early-stage resectable lung cancer [@b13-co16-5-111] and in selecting appropriate locally advanced lung cancer patients for aggressive combined-modality therapy [@b14-co16-5-111]. As a result of these clinical trials, lung cancer patients in Ontario now have access to [pet]{.smallcaps} imaging as part of their care when they fit the foregoing criteria.\n\n### 2.1.3 Interdisciplinary Care\n\nThe patient journey for lung cancer involves interaction with many disciplines, and there is an expanding role for the advanced practice nurse ([apn]{.smallcaps}), as presented by Ms. Lorraine Martelli--Reid, an [apn]{.smallcaps} from the Juravinski Cancer Centre.\n\nThe roles of the [apn]{.smallcaps} span the spectrum from clinical care to education, research, and organizational leadership. A retrospective review at the Juravinski Cancer Centre ([jcc]{.smallcaps}) of patients undergoing postoperative adjuvant chemotherapy with cisplatin and vinorelbine in the National Cancer Institute of Canada ([ncic]{.smallcaps}) Clinical Trials Group ([ctg]{.smallcaps}) [br]{.smallcaps}.10 [@b15-co16-5-111] trial showed an absolute survival difference of 15%. However, the chemotherapy regimen is difficult to complete: only 50% of patients on the [ncic br]{.smallcaps}.10 clinical trial were able to complete all 4 cycles of chemotherapy. At the [jcc]{.smallcaps}, however, 84% of patients were able to complete the 4 cycles. Support from an [apn]{.smallcaps} is vital in helping to manage symptoms during chemotherapy and in providing education and counselling for the patients. A novel [apn]{.smallcaps}-led \"Take a Breather Clinic\" was established to help lung cancer patients with symptoms of dyspnea. Dyspnea had been identified as a significant symptom using the Edmonton Symptom Assessment System ([esas]{.smallcaps}).\n\nUse of the [esas]{.smallcaps} as a common tool for assessment was initiated through the Provincial Palliative Care Integration Project ([ppcip]{.smallcaps}) funded by the Ministry of Health and Long-Term Care and [cco]{.smallcaps}. Dr. Jeff Myers, palliative physician and Toronto Central [lhin]{.smallcaps} lead for the project, indicated that the aim was to target all lung cancer and palliative care patients in the regional cancer centres and all palliative patients in the home setting. By using common tools that incorporate a symptom measurement scale (that is, the [esas]{.smallcaps} [@b16-co16-5-111]), symptom management guidelines for intervention, and a palliative performance scale, it would be possible to evaluate patient symptoms and to monitor progress through the course of a patient's care. The success of the [ppcip]{.smallcaps} led to the next stage (that is, the Ontario Cancer Symptom Management Collaborative), which includes all cancer patients with participation of all regional cancer centres and community care access centres. This project has given a \"voice\" to the patient's symptoms, which may or may not be usually discussed, and has provided a common language for communication between care providers.\n\n2.2 Controversies in the Management of Esophageal Cancer\n--------------------------------------------------------\n\nThis year's meeting explored selected issues in esophageal cancer. Adenocarcinomas of the esophagus and gastroesophageal junction are increasing in incidence, and squamous cell cancers are decreasing. The 5-year survival rates are poor in surgically resected patients, emphasizing the need for more effective adjunctive therapies. Dr. Jennifer Knox from the Princess Margaret Hospital reviewed the major clinical trials involving preoperative chemotherapy [@b17-co16-5-111]--[@b19-co16-5-111] or preoperative chemoradiation [@b20-co16-5-111].\n\nPreoperative chemoradiation for adenocarcinomas improves overall survival and achieves higher pathologic complete response rates than are seen with chemotherapy alone, but the regimen is more toxic. Improvement in overall survival is less certain with preoperative chemoradiation in squamous cell carcinomas, but local control is improved over that with surgery alone. The ability to evaluate response to neo-adjuvant therapy would be useful for prognostication. The [municon]{.smallcaps} trial [@b21-co16-5-111] evaluated the strategy of using [pet]{.smallcaps} to determine the length of preoperative chemotherapy before surgery in locally advanced gastroesophageal junction cancers. In that trial, patients underwent a [pet]{.smallcaps} scan at baseline and then proceeded to neoadjuvant chemotherapy with cisplatin and 5-fluorouracil. After 2 weeks of therapy, a re-evaluation [pet]{.smallcaps} scan separated the metabolic non-responders (less than 35% decrease in standardized uptake value) from the responders. The non-responders by [pet]{.smallcaps} proceeded to surgery immediately; the responders continued the rest of their chemotherapy to 12 weeks before receiving surgery. Responders by [pet]{.smallcaps} had a 96% R0 resection rate and a 58% major pathologic response (defined as less than 10% residual tumour).\n\nThe use of [pet]{.smallcaps} in radiation treatment planning was evaluated by Dr. Danny Vesprini and colleagues from the Odette Cancer Centre. Their study evaluated the effect of the addition of fused [pet]{.smallcaps}--[ct]{.smallcaps} imaging over [ct]{.smallcaps} alone in the identification of the gross tumour volume ([gtv]{.smallcaps}) in patients with esophageal cancer [@b22-co16-5-111]. Ten patients with esophageal cancer underwent [pet]{.smallcaps} and [ct]{.smallcaps} imaging in radiation treatment position, and the resulting image sets were fused. Six radiation oncologists independently contoured the [gtv]{.smallcaps} using [ct]{.smallcaps} data alone, supplemented with standardized clinical and diagnostic imaging information. The same radiation oncologists then contoured the [gtv]{.smallcaps} using the co-registered [pet]{.smallcaps}--[ct]{.smallcaps} images. The standard deviation of the [gtv]{.smallcaps} length and volume were used a measure of inter-observer and intra-observer variation. The average observer agreement index using [pet]{.smallcaps}--[ct]{.smallcaps} was 72.7% as compared with 69.1% using [ct]{.smallcaps} alone. The [pet]{.smallcaps}--[ct]{.smallcaps} significantly improved both inter-observer and intra-observer variability in the identification of the primary [gtv]{.smallcaps}.\n\n2.3 Molecular Targeted Therapies for Lung Cancer\n------------------------------------------------\n\nThe third keynote address on \"Current Perspectives in the Treatment of Advanced Non-Small Cell Lung Cancer\" was given by Dr. Natasha Leighl, from the Princess Margaret Hospital.\n\nAs the biology of tumour progression becomes better understood, newer targets for biologic therapies will become available for clinical trials. Currently, a wealth of molecular targeted therapies are under investigation in [nsclc]{.smallcaps}. The most promising new therapies target either the vascular endothelial growth factor ([vegf]{.smallcaps}) [@b23-co16-5-111] or the epidermal growth factor receptor ([egfr]{.smallcaps}) [@b24-co16-5-111]. Two important randomized phase [iii]{.smallcaps} clinical trials evaluated the addition of bevacizumab to standard chemotherapy as compared with standard chemotherapy alone in advanced-stage ([iiib]{.smallcaps}/[iv]{.smallcaps}) and recurrent [nsclc]{.smallcaps}. The Eastern Cooperative Oncology Group E4599 trial [@b25-co16-5-111] and the [ava]{.smallcaps}i[l]{.smallcaps} trial [@b26-co16-5-111] showed improved progression-free survival for their bevacizumab arms. In the E4599 trial, overall median survival also improved to 12.3 months from 10.3 months \\[hazard ratio ([hr]{.smallcaps}): 0.79; *p* = 0.003\\], and the adenocarcinoma subgroup had a more significant improvement in overall median survival to 14.2 months ([hr]{.smallcaps}: 0.69). The [ava]{.smallcaps}i[l]{.smallcaps} trial did not demonstrate a survival benefit. The incidence of grade 3 or more serious adverse events on these trials was low, ranging from 0.3% to 9% for bleeding, hypertension, proteinuria, febrile neutropenia, and arterial thrombosis.\n\nA promising [vegf]{.smallcaps} small-molecule inhibitor, cediranib, was evaluated by the [ncic ctg]{.smallcaps} in a phase [ii]{.smallcaps}/[iii]{.smallcaps} study design---the [ncic br]{.smallcaps}.24 trial. Patients were randomized to carboplatin and paclitaxel with cediranib or a placebo; the cediranib arm had an increased response rate of 38% as compared with 16% ([hr]{.smallcaps}: 0.77) [@b27-co16-5-111], but some toxic deaths from dehydration and diarrhea occurred. The follow-up trial, [ncic br]{.smallcaps}.29 will use a lower dose of cediranib to reduce the occurrence of side effects.\n\nThe [flex]{.smallcaps} trial evaluated the use of cisplatin and vinorelbine with or without cetuximab (a monoclonal antibody against [egfr]{.smallcaps}) in first-line treatment of stage [iiib]{.smallcaps}/[iv nsclc]{.smallcaps}. The cetuximab arm had a median overall survival of 11.3 months as compared with 10.1 months, and a 1-year survival of 47% as compared with 42% ([hr]{.smallcaps}: 0.871; *p* = 0.044) [@b28-co16-5-111]. Patients who developed an early acne-like rash experienced a median overall survival of 15 months.\n\nA current controversy is the role for maintenance therapy after completion of first-line treatment with a platinum doublet. In a trial by Ciuleanu *et al.* [@b29-co16-5-111] of maintenance after completion of platinum chemotherapy, patients were randomized to either pemetrexed or placebo (2:1 randomization), resulting in a median progression-free survival of 4.3 months as compared with 2.6 months ([hr]{.smallcaps}: 0.502; *p* \\< 0.00001), and an improvement in overall survival to 13.4 months as compared with 10.6 months ([hr]{.smallcaps}: 0.79; *p* = 0.012), with a significant difference of 14.4 months as compared with 9.4 months (*p* = 0.0025) in non-squamous histology. Similar results were seen for the use of targeted therapies for maintenance on the [saturn]{.smallcaps} [@b30-co16-5-111] and [atlas]{.smallcaps} [@b31-co16-5-111] trials that used erlotinib, or bevacizumab with or without erlotinib, although survival data are pending.\n\n3. THE GREAT DEBATES\n====================\n\nThree issues were debated at this year's meeting:\n\n-   Stereotactic body radiation therapy ([sbrt]{.smallcaps}) compared with surgery for T1N0 lung cancer\n\n-   Preoperative compared with postoperative chemoradiation for esophageal cancer\n\n-   Endoscopic mucosal resection compared with surgery for esophageal cancer\n\n3.1 SBRT Versus Surgery for T1N0 NSCLC\n--------------------------------------\n\nDr. Patrick Cheung, Odette Cancer Centre, debated Dr. Richard Inculet, London Health Sciences Centre in the first debate of radiation versus surgery for early-stage [nsclc]{.smallcaps}.\n\nA comparison of current outcomes in patients treated using [sbrt]{.smallcaps} with those in patients undergoing surgery for stage [i nsclc]{.smallcaps} are limited by the accuracy of staging. Patients referred for radiation often have significant medical comorbidities that preclude surgical resection, and they are often clinically staged where surgical candidates are pathologically staged. In addition, the radiation dose used to control early-stage [nsclc]{.smallcaps} is very important: The dose given must be effective enough to eradicate small lung cancers.\n\nIn a large retrospective multi-institutional study of [sbrt]{.smallcaps} using biologically effective doses of radiation for stage [i nsclc]{.smallcaps}, a 5-year survival rate of 53.9% was achieved, and in the subset of operable lung patients, the survival increased to 70.8% [@b32-co16-5-111]. The toxicities associated with [sbrt]{.smallcaps} for peripheral locations are minimal; they include radiation pneumonitis (5.4%), mild dermatitis (1.2%), and rib fracture (1.6%). Centrally located lesions may carry a higher risk of bronchial stenosis with lung collapse, and current clinical trials are evaluating the safety of treating central lesions with [sbrt]{.smallcaps}. The major concern with [sbrt]{.smallcaps} is the effect of radiation on patients with poor pulmonary function. However, an analysis of 70 medically inoperable stage [i nsclc]{.smallcaps} patients with poor baseline pulmonary function did not predict for decreased survival or decreased pulmonary function after treatment [@b33-co16-5-111].\n\nSurgical resection is still the standard of care for resectable early-stage [nsclc]{.smallcaps}. Innovations with minimally invasive surgical techniques---that is, video-assisted thoracoscopic surgery ([vats]{.smallcaps})---have reduced surgical morbidity. In selected cases, outcomes may be better than those with standard lobectomy [@b34-co16-5-111]. The safety and efficacy of [vats]{.smallcaps} lobectomy compared with open lobectomy have been systematically reviewed, and no statistically significant differences were observed in terms of postoperative prolonged air leak, arrhythmia, pneumonia, mortality, or risk of locoregional recurrence [@b35-co16-5-111]. Today, patients that might not have been considered for open lobectomy may therefore, with [vats]{.smallcaps}, still be surgical candidates.\n\nThe choice of [sbrt]{.smallcaps} or surgery for stage [i nsclc]{.smallcaps} will be multifactorial, but appropriately selected patients will do well with either option.\n\n3.2 Preoperative Versus Postoperative Chemoradiation for Esophageal Cancer\n--------------------------------------------------------------------------\n\nDr. Rebecca Wong, Princess Margaret Hospital, debated Dr. Richard Malthaner, London Health Sciences Centre, in the second debate on esophageal cancer.\n\nSurgery alone is insufficient treatment for resectable, but locally advanced, cancers of the esophagus because locoregional and distant recurrence rates are significant. Therefore using either preoperative or postoperative therapy may be useful in improving outcomes. The advantage of using a preoperative approach are these:\n\n-   Tumour downstaging can occur before surgical resection.\n\n-   Radiation target volumes are smaller.\n\n-   Perioperative morbidity is less.\n\n-   Radiation dose is more effective in an undisturbed tumour.\n\nIn esophageal cancer, 10 randomized controlled clinical trials have involved 1209 patients. The [hr]{.smallcaps} for all-cause mortality was 0.81 for neoadjuvant chemoradiotherapy as compared with surgery alone, corresponding to a 13% absolute difference in survival at 2 years favouring neoadjuvant chemoradiation [@b20-co16-5-111].\n\nPostoperative chemoradiotherapy holds these advantages:\n\n-   Postoperative adjuvant therapy can be tailored using the accurate stage.\n\n-   Unnecessary treatment of early-stage esophageal cancer is avoided.\n\n-   Surgery is better tolerated.\n\n-   Immediate improvement is achieved in the major presenting symptom, dysphagia.\n\nNo randomized controlled trials have compared postoperative chemoradiation with surgery alone, and none have compared preoperative chemoradiation with postoperative chemoradiation. At the London Regional Cancer Centre, a retrospective review of patients with lymph-node-positive disease who, after surgical resection, were given postoperative chemoradiation showed that the postoperative treatment, as compared with no treatment, was associated with significantly longer survival [@b36-co16-5-111]. A definitive clinical trial comparing preoperative with postoperative chemoradiation would be useful in determining the precise benefit in terms of survival and quality of life.\n\n3.3 Endoscopic Mucosal Resection Versus Surgery for Esophageal Cancer\n---------------------------------------------------------------------\n\nDr. Norman Marcon, St. Michael's Hospital, debated Dr. Richard Inculet, London Health Sciences Centre, in the third debate on the treatment of early esophageal cancer.\n\nThe dilemmas faced in the treatment of high-grade dysplasia and intramucosal adenocarcinoma [@b37-co16-5-111] include\n\n-   the confidence of the pathologic diagnosis,\n\n-   the malignant risk of the lesion,\n\n-   the completeness of the resection,\n\n-   the morbidity and mortality of the treatment, and\n\n-   the eradication of the disease.\n\nEndoscopic therapy can be either endoscopic mucosal resection or endoscopic submucosal dissection (basically removing the mucosal tissue down to and including the submucosa). For endoscopic resection to be successful, there must be accurate staging of the disease, a low failure rate, an accurate method of surveillance, good functional results post treatment, and an effective way to deal with the underlying cause---that is, gastroesophageal reflux disease [@b38-co16-5-111]. For surgical resection to be successful, there must be a low complication rate, a reasonable functional result, and a high curative potential [@b39-co16-5-111]. Both treatment options are suitable, depending on patient compliance, disease characteristics, extent of disease, and expertise of the treating physician.\n\n4. POSTER PRESENTATIONS\n=======================\n\nResearch by the medical trainees was highlighted in the poster presentations. The abstract review committee selected two posters for oral presentation. The first, by Dr. Meredith Giuliani, was titled \"Prophylactic Cranial Irradiation Utilization Rates in Limited-Stage Small-Cell Lung Cancer.\" The second, by Dr. Jeffrey Cao, was a \"Systematic Review of the Cost-Effectiveness of PET in Staging of Non- Small-Cell Lung Cancer and Management of Solitary Pulmonary Nodules.\" The abstracts are published in the [Appendix](#app1-co16-5-111){ref-type=\"app\"} to this report.\n\n5. SUMMARY\n==========\n\nThe Ontario Thoracic Cancer Conference continues to bring together people interested in the management of patients with thoracic malignancies. It remains an excellent forum to foster research and wide multidisciplinary interaction. We extend our thanks to all who made this meeting such a success, including our sponsors, Astra Zeneca and Lilly (platinum level), Olympus (gold level), and Boehringer Ingelheim (silver level).\n\nWaters E, Rodrigues G, Vincent M, Dingle B. London Regional Cancer Centre, University of Western Ontario, London, Ontario.\n\nBackground and Objectives\n=========================\n\nThe London Regional Cancer Program ([lrcp]{.smallcaps}) employs a unique schedule of concurrent chemoradiation, termed [vcrt]{.smallcaps} (vinblastine, cisplatin, radiation therapy), for the treatment of unresectable stage [iiia]{.smallcaps} and [iiib]{.smallcaps} non-small-cell lung cancer ([nsclc]{.smallcaps}). The protocol consists of 100 mg/m^2^ cisplatin and 6 mg/m^2^ vinblastine (reduced to 4.2 mg/m^2^ during cycles 3 and 4), split over days 1--3, given once every 3 weeks for 4 cycles, with 60 Gy of concurrent radiation over 6 weeks during cycles 3 and 4.\n\nThe objective of the present study was to determine overall survival and to characterize outcomes of patients treated with [vcrt]{.smallcaps}.\n\nMethods\n=======\n\nThis retrospective analysis, with a focus on overall survival and toxicities, reports a cohort of 294 patients who underwent [vcrt]{.smallcaps} at the [lrcp]{.smallcaps} between 1996 and 2006.\n\nResults\n=======\n\nThe overall 5-year survival, determined using Kaplan--Meier methodology, was 19.8%, and the median survival duration was 18.2 months. Reported grades 3 and 4 toxicities included neutropenia (39%), anemia (10%), pneumonitis (1%), and esophagitis (3%). Log-rank tests demonstrated significant differences in survival between groups of patients for completion surgery, use of radiation therapy, and cisplatin dose. Similarly, univariate Cox regression showed that completion surgery, use of radiation therapy, cisplatin dose, and vinblastine dose were significant factors in the survival of all stage [iii nsclc]{.smallcaps} patients treated with [vcrt]{.smallcaps}.\n\nDiscussion and Conclusions\n==========================\n\nThis retrospective analysis reveals an overall survival comparable to that of other current combined chemoradiation protocols. The success of the [vcrt]{.smallcaps} protocol seems to be dose-dependent.\n\nAlam Y, Zhao Y. Windsor Regional Cancer Centre, Windsor, Ontario; Schulich School of Medicine, University of Western Ontario, London, Ontario.\n\nBackground and Objectives\n=========================\n\nLung cancer is strongly correlated with cigarette smoking, although other environmental and genetic components may also play a role. There have been public concerns in Southwestern Ontario that environmental and occupational exposures are resulting in a higher lung cancer rate.\n\nThe present study aimed to further characterize the smoking habits, histology subtypes, stage at diagnosis, and 5-year survival rates in newly diagnosed lung cancer patients in the Southwestern Ontario region in 2003.\n\nMethods\n=======\n\nWe retrospectively reviewed all incident lung cancer patients seen at the Windsor Regional Cancer Centre ([wrcc]{.smallcaps}) in 2003. Data were collected from the e-chart database used at the centre. Age, presenting symptoms, histology, initial clinical staging, and 5-year survival were recorded.\n\nResults\n=======\n\nIn 2003, 189 new lung cancer cases were seen at the [wrcc]{.smallcaps}. In 93.65% of the cases, the patient was 50 years of age or older. In 93.12% of the cases, the patient was a smoker or ex-smoker. Weight loss, cough, and dyspnea were the most common presenting symptoms. Non-small-cell lung cancer ([nsclc]{.smallcaps}) constituted 74.07% of the cases, of which 64% were stage [iii]{.smallcaps} or [iv]{.smallcaps} at initial diagnosis. The small-cell subtype accounted for 14.99% of the cases, and 61.76% of those patients had extensive disease at diagnosis. Survival at 5 years was 5.82%.\n\nDiscussion and Conclusions\n==========================\n\nMore than 90% of newly diagnosed lung cancer patients from the Southwestern Ontario region in 2003 were smokers and 50 years of age or older. The histology subtypes and survival trends for these patients were similar to trends published elsewhere in the literature. Another review for 2008, to see if the characteristics have changed, would be interesting.\n\nKuruvilla MS, Martelli-Reid L, Goffin JR, Arnold A, Ellis PM. Juravinski Cancer Centre, McMaster University, Hamilton, Ontario.\n\nBackground and Objectives\n=========================\n\nPublished data support the use of cisplatin--vinorelbine doublet as adjuvant chemotherapy in completely resected non-small-cell lung cancer ([nsclc]{.smallcaps}). However, data also demonstrate a need for frequent dose reductions. Consequently, many centres across Canada have been reluctant to adopt this practice.\n\nThe primary objective of the present study was to assess the deliverability of this adjuvant regimen in patients with stage [ib]{.smallcaps}, [ii]{.smallcaps}, and [iii nsclc]{.smallcaps} within our center. Secondary objectives were to determine the tolerability and toxicity of this regimen.\n\nMethods\n=======\n\nWe retrospectively reviewed patients with [nsclc]{.smallcaps} receiving adjuvant cisplatin--vinorelbine at the Juravinski Cancer Centre between January 2005 and September 2007. Demographics, total chemotherapy dose, treatment duration, and toxicity profiles were abstracted. Relative dose intensity ([rdi]{.smallcaps}) was calculated as a marker of deliverability.\n\nResults\n=======\n\nAdjuvant cisplatin--vinorelbine was administered to 41 patients. The median weekly dose intensity was 23.5 mg/m^2^ (range: 13.1--27.9 mg/m^2^) for cisplatin and 18.8 mg/m^2^ (range: 8.4--25.1 mg/m^2^) for vinorelbine. The median [rdi]{.smallcaps}s for cisplatin and vinorelbine were 94% and 63%. Of the treated patients, 71% underwent all 4 cycles; 61% received all treatments of cisplatin, and 16%, all treatments of vinorelbine. Toxicities at grade 3 or higher included anemia (12%), neutropenia (63%), febrile neutropenia (7%), constipation (2%), and fatigue (18%). Blood transfusions were given to 24% of the patients, and 18% received erythropoiesis-stimulating agents. Less severe toxicities included peripheral neuropathy, ototoxicity, mucositis, vomiting, alopecia, and flare of surgical site pain.\n\nDiscussion and Conclusions\n==========================\n\nThe deliverability of adjuvant cisplatin--vinorelbine, administered weekly for 16 weeks, was as good as or better than that in the [anita]{.smallcaps} trial. This finding helps to attenuate concerns about the tolerability of the regimen.\n\nChan E, Kiss A, Balogh J, Barbera L, Cheung P, Poon I, Spayne J, Ung YC. Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario.\n\nBackground\n==========\n\nImaging by positron-emission tomography ([pet]{.smallcaps}) is increasingly used to stage and plan radiotherapy ([rt]{.smallcaps}) in patients with non-small-cell lung cancer ([nsclc]{.smallcaps}). The effect of this approach on [rt]{.smallcaps} volumes is not yet fully known. The present study evaluated differences in *V*~20~ (a predictor for radiation pneumonitis) and contoured [rt]{.smallcaps} volumes in [pet]{.smallcaps}-versus computed tomography ([ct]{.smallcaps})--based plans for stage [iii nsclc]{.smallcaps}.\n\nMethods\n=======\n\nAs part of their work-up, 18 patients underwent [pet]{.smallcaps} and [ct]{.smallcaps}. The [rt]{.smallcaps} volumes were initially contoured using the [ct]{.smallcaps} data alone. Contours were then modified by information from the [pet]{.smallcaps} imaging. Differences between the paired contours and *V*~20~ determinations for each patient were calculated.\n\nResults\n=======\n\nThe average difference between the total contoured gross tumour volume for the [pet]{.smallcaps} and [ct]{.smallcaps} approaches was similar (1.13 cm^3^). In 5 of 18 patients, volume differences of more than 25 cm^3^ were observed. The average difference between the contoured planning target volumes for the two approaches was more varied ([pet]{.smallcaps} volumes were larger on average by 12.99 cm^3^). In 6 of 15 patients, volume differences of more than 50 cm^3^ were observed. The calculated *V*~20~ was similar between the two approaches ([ct]{.smallcaps} plans were larger on average by 0.19%) with a range of \u22122.8% to 4.5%. The range of *V*~20~ based on [ct]{.smallcaps} was 15.2% to 37.7%; on [pet]{.smallcaps}, it was 14.7% to 36.7%.\n\nConclusions\n===========\n\nOverall, there appears to be only a small change between the contoured [gtv]{.smallcaps} and the planning target volume when using [pet]{.smallcaps} or [ct]{.smallcaps} imaging. However, in a proportion of patients, the contoured volumes are quite different. The ultimate effect of these differences will need to be validated by clinical outcomes.\n\nLouie A, Rodrigues G, Gaede S. London Regional Cancer Centre, University of Western Ontario, London, Ontario.\n\nObjectives\n==========\n\nThe purpose of the present study was to investigate inter- and intra-observer target volume delineation ([tvd]{.smallcaps}) error in the setting of four-dimensional (4D) computed tomography ([ct]{.smallcaps}) image data acquisition in thoracic tumours.\n\nMethods\n=======\n\nSix radiation oncologists contoured the primary and nodal gross tumour volume ([gtv]{.smallcaps}) of 10 lung tumours on the 10 respiratory phases of a 4D [ct]{.smallcaps} scan. The coefficient of variation ([cov]{.smallcaps}) and the percentage shared internal target volume (s[itv]{.smallcaps}) of the 6 physicians for each patient was used to assess inter- and intra-observer variability.\n\nAnalysis of variance was performed to assess differences in inter- and intra-physician variability based on patient case difficulty, respiratory phase, physician seniority, and physician observer.\n\nResults\n=======\n\nInter-physician percentage s[itv]{.smallcaps} for primary tumour ranged from 31.1% to 83.3% \\[standard deviation ([sd]{.smallcaps}): 4.4%--15.8%\\] and from 16.4% to 66.8% ([sd]{.smallcaps}: 5.8%--21.1%) for nodes. Intra-physician s[itv]{.smallcaps} for primary tumour ranged from 59.6% to 72.7% ([sd]{.smallcaps}: 13.0%--23.9%), and from 28.3% to 57.0% ([sd]{.smallcaps}: 18.6%--34.2%) for nodes. Analysis of variance for [cov]{.smallcaps}s found case difficulty (easy vs. difficult) to be significant for inter-physician primary tumour and intra-physician nodal disease delineation. Physician seniority, respiratory phase, and individual physician were not found to be significant for [tvd]{.smallcaps} error.\n\nConclusions\n===========\n\nHigh observer variability in [tvd]{.smallcaps} continues to be a major source of error in the 4D [ct]{.smallcaps} era for lung cancer. Inter-physician variability appears to be the more significant source of this error than intra-physician variability. Development of measures to reduce inter- and intra-observer [tvd]{.smallcaps} variability are necessary to the delivery of high-quality radiotherapy.\n\nTaremi M, Dahele M, Bezjak A. Princess Margaret Hospital, University Health Network, University of Toronto, Toronto, Ontario.\n\nBackground and Objectives\n=========================\n\nStereotactic body radiotherapy ([sbrt]{.smallcaps}), a technique to deliver high-dose radiation in each fraction, is expected to provide high rates of local control.\n\nMethods\n=======\n\nPatients eligible for our lung [sbrt]{.smallcaps} Research Ethics Board--approved protocol included those with inoperable early-stage [nsclc]{.smallcaps} (T1/T2, N0, M0), and patients with a limited number of pulmonary metastases.\n\nTwo dose/fractionation (fr) schedules for peripheral tumours are 48 Gy/4 fr for T1 and 54--60 Gy/3 fr for T2 tumours. If the tumour is in proximity to midline structures, 60 Gy/8 fr or 50 Gy/10 fr is used. Toxicity and tumour response are assessed using Common Terminology Criteria for Adverse Events v.3 and the Response Evaluation Criteria in Solid Tumors criteria respectively.\n\nResults\n=======\n\nBetween December 2004 and July 2008, 111 patients (median age: 72 years) were treated. The data for 96 patients with early-stage non-small-cell lung cancer (100 lesions) and a median follow-up of 16.4 months were analyzed for this report. In patients with at least 6 months of follow-up, we observed 42 partial responses, 35 complete responses, 10 stable disease, and 11 disease progression. Local failure occurred for 9 lesions, 5 of which were treated with 50 Gy/10 fr.\n\nThe estimated 3-year overall survival was 48% \\[95% confidence interval ([ci]{.smallcaps}): 32% to 62%\\]; cause-specific survival was 83% (95% [ci]{.smallcaps}: 72% to 94%). The most common acute toxicity was fatigue (42 patients). No patient had grade 4 or 5 toxicity.\n\nDiscussion and Conclusions\n==========================\n\nIn early-stage [nsclc]{.smallcaps}, [sbrt]{.smallcaps} is an effective and well-tolerated treatment. However, careful patient selection, attention to planning and treatment delivery, and ongoing follow-up is needed to fully define the therapeutic ratio for this technique.\n\nGiuliani M, Hope A, Sun A, Ma C, Brade A, Cho JBC, Bezjak A. Princess Margaret Hospital, University Health Network, University of Toronto, Toronto, Ontario.\n\nBackground and Objectives\n=========================\n\nProphylactic cranial irradiation ([pci]{.smallcaps}) improves survival in patients with limited-stage small-cell lung cancer ([lssclc]{.smallcaps}). The objective of the present audit was to assess the adoption of [pci]{.smallcaps} at our institution.\n\nMethods\n=======\n\nFrom 1997 to 2007, 796 patients were treated at Princess Margaret Hospital for [sclc]{.smallcaps}. Of these 796 patients, the 226 (28.4%) who received radical [ls]{.smallcaps}-[sclc]{.smallcaps} treatment formed the basis of this project. Brain failure-free survival ([ffs]{.smallcaps}) was estimated by the Kaplan--Meier method, comparing patients treated with and without [pci]{.smallcaps}.\n\nThe [pci]{.smallcaps} uptake was determined, and for patients not receiving [pci]{.smallcaps}, the reason was recorded if it was available. The Fisher exact test was used to compare rates of [pci]{.smallcaps} use during 1997--2001 and 2002--2007.\n\nResults\n=======\n\nMedian follow-up was 15.5 months (range: 2--130 months). From 1997 to 2007, 55.3% (*n* = 125) of radically treated [ls]{.smallcaps}-[sclc]{.smallcaps} patients received [pci]{.smallcaps}. Brain [ffs]{.smallcaps} at 6, 12, and 24 months was 94.3%, 69.3%, and 32.0% respectively for patients who did not receive [pci]{.smallcaps} and 100.0%, 94.6%, and 76.8% respectively for patients who did receive that treatment (*p* \\< 0.0001). A nonsignificant increase in [pci]{.smallcaps} uptake, 51.7% to 59.1% (*p* = 0.29), occurred from 1997--2001 to 2002--2007. The most common reasons for not receiving [pci]{.smallcaps} were patient refusal and disease progression.\n\nDiscussion and Conclusions\n==========================\n\nProphylactic cranial irradiation significantly improves brain [ffs]{.smallcaps}; only half of [ls]{.smallcaps}-[sclc]{.smallcaps} patients at our institution received [pci]{.smallcaps}. Uptake of [pci]{.smallcaps} has not significantly increased in recent years. A thorough exploration of patient concerns regarding [pci]{.smallcaps} may increase the uptake rate.\n\nCao JQ, Rodrigues GB. London Health Sciences Centre, University of Western Ontario, London, Ontario.\n\nBackground and Objective\n========================\n\nThis systematic review describes the cost-effectiveness of positron-emission tomography ([pet]{.smallcaps}) imaging in the staging of [nsclc]{.smallcaps} and the management of solitary pulmonary nodules ([spn]{.smallcaps}s).\n\nMethods\n=======\n\nWe conducted systematic literature searches in the [medline]{.smallcaps}/Pre[medline]{.smallcaps}, [embase]{.smallcaps}, and U.K. National Health Service databases. Measurement of study quality was assessed by the validated Quality of Health Economic Studies ([qhes]{.smallcaps}) instrument. Studies with a [qhes]{.smallcaps} score below 75 were excluded. Characteristics including study methodology, assumptions, and cost effectiveness metrics \\[incremental cost-effectiveness ratio---[icer]{.smallcaps}---based on life-years saved and average cost savings per patient ([acsp]{.smallcaps})\\] were abstracted. Descriptive statistics were generated with cost amounts converted to a common inflation-adjusted 2007 U.S. dollar.\n\nResults\n=======\n\nThe 20 studies that met all inclusion criteria, including acceptable [qhes]{.smallcaps} scores as determined by two reviewers (mean: 87.8), were based on the national health insurance payer perspective of 8 different countries. Investigations assessed the [spn]{.smallcaps} scenario (*n* = 8), the staging scenario (*n* = 11), and the [spn]{.smallcaps} and staging scenarios (*n* = 1) together. Mean assumed cost of [pet]{.smallcaps} scanning was \\$1267 (range: \\$769--\\$2580) in these studies. Median [icer]{.smallcaps}s for [spn]{.smallcaps} and staging were \\$2039 (range: \\$181--\\$3927) and \\$4037 (range: \\$527--\\$32618) respectively. Median [acsp]{.smallcaps}s for [spn]{.smallcaps} and staging were \\$518 (range: \\$66--\\$1480) and \\$1390 (range: \\$143--1633) respectively.\n\nConclusions\n===========\n\nReported cost-effectiveness metrics are highly variable and depend on input variables and assumptions including: cost, disease prevalence, diagnostic operating characteristics, the diagnostic strategies assessed, and the methodologies used. Despite this variation, these studies have consistently concluded that [pet]{.smallcaps} has favourable cost-effectiveness characteristics as compared with non-[pet]{.smallcaps} strategies.\n\nYu E, Tai P, Malthaner R, Stitt L, Rodrigues G, Dar R, Yaremko B, Younus J, Sanatani M, Vincent M, Dingle B, Fortin D, Inculet R. London Health Science Centre, University of Western Ontario, London, Ontario, and Allan Blair Cancer Center, Regina, Saskatchewan.\n\nObjective\n=========\n\nThis study investigated the effect of resection margin status and time interval to relapse on outcomes in high-risk esophageal cancer patients.\n\nPatients and Methods\n====================\n\nDuring 1989--1999, we followed high-risk resected esophageal cancer patients who completed postoperative chemoradiation therapy. Adjuvant chemotherapy consisted of 4 cycles of epirubicin, cisplatin, and 5-flurouracil ([ecf]{.smallcaps}), with epirubicin omitted during the radiation therapy ([rt]{.smallcaps}) phase. Total [rt]{.smallcaps} dose was 45--60 Gy at 1.8--2.0 Gy per fraction. Patients who relapsed after a disease-free interval of more than 3 months were treated with palliative chemoradiation when appropriate. Patients who relapsed after a disease-free interval of 3 months or less were treated with best supportive care. Logistic regression and log-rank tests were used for post-recurrence survival analysis.\n\nResults\n=======\n\nOf the 69 patients treated with adjuvant chemoradiation post-esophagectomy, 46 patients experienced recurrence. Median time to relapse was 28 months (range: 0.1--40 months). Median age of relapse patients was 61 years (range: 37--82 years). There were 42 male, 44 node-positive, 31 adenocarcinoma, and 33 clear resection margin post-esophagectomy patients. Median follow-up after recurrence was 30.5 months (range: 1.3--100 months). The median post-recurrence overall survival duration was 5.8 months, with overall survival rates of 20%, 10%, and 5% at 12 months, 24 months, and 36 months respectively. Of the prognostic factors analyzed, only resection margin status and time interval to recurrence were statistically significant for patient outcome in univariate and multivariate analysis.\n\nConclusions\n===========\n\nSurgical resection margin status and time interval to disease relapse were independent prognostic factors for patient outcome.\n\nThe opinions expressed in the abstracts are those of the authors and are not to be construed as the opinion of the publisher (Multimed Inc.), or the organizers of the 4th Annual Ontario Thoracic Cancer Conference.\n\nAlthough the publisher (Multimed Inc.) has made every effort to accurately reproduce the abstracts, Multimed Inc. and the organizers of the 4th Annual Ontario Thoracic Cancer Conference assume no responsibility and/or liability for any errors and/or omissions in any abstract published.\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nChanges in diet and lifestyle are two important factors causing the worldwide dramatic increase in the incidence of diabetes \\[[@pone.0206169.ref001]\\]. Currently, about 0.7% of the world population suffer from insulin-dependent diabetes mellitus \\[[@pone.0206169.ref002]\\]. Both type I and type II diabetic patients require insulin, but because of developing insulin resistance, the late stage type II diabetes patients use larger amounts of this hormone \\[[@pone.0206169.ref003], [@pone.0206169.ref004]\\]. Due to the cumulative incidence of this metabolic disease among children and adults, insulin is now at the top of the list of therapeutic peptides in high demand \\[[@pone.0206169.ref005]\\]. Insulin controls the storage and use of sugars and if by its dysfunction, the blood sugar homeostasis is not well preserved for long periods, serious complications will appear in different tissues \\[[@pone.0206169.ref006]\\].\n\nMethods for producing insulin initially used extraction of the native hormone from the animal pancreas \\[[@pone.0206169.ref007]\\]. However, the use of bovine or pig insulin is associated with the appearance of significant adverse effects (e.g. allergy) \\[[@pone.0206169.ref008]\\]. Due to its reputable production efficiency, recombinant DNA technology is now used for human insulin production. *Escherichia coli* and *Saccharomyces cerevisiae* are the most important hosts for production of human insulin \\[[@pone.0206169.ref009]\\]. Because of limitation in production capacity and high production cost of insulin, current manufacturing technologies are not able to meet the human demand \\[[@pone.0206169.ref010]\\]. In spite of the several attempts which have been recently made to express proinsulin and insulin in bacteria, the direct production of (pro) insulin in the bacterial host is generally difficult. Insulin is composed of two chains, A and B, that are joining by two disulfide bonds with a third disulfide bond within the A-chain. Due to their small size, insulin peptide chains are very prone to be degraded by the host's proteases \\[[@pone.0206169.ref011]\\]. To overcome this challenge, an approach for a successful expression and purification system is to design recombinant insulin chains fused to an appropriate partner protein. Subsequently, the target insulin peptides are released from the fusion protein by either chemical or enzymatic cleavage at the corresponding site of carrier-peptide junction. Choosing an appropriate fusion protein partner is very important because according to previous studies, certain fusion protein partners could greatly improve the stability and the level of expression of the target proteins \\[[@pone.0206169.ref012]--[@pone.0206169.ref014]\\]. In the pancreatic tissue, a single chain proinsulin molecule is produced with the A-chain (21 amino acids) and B-chain (30 amino acids) linked together by the C-peptide (31 amino acids). The formation of native insulin from proinsulin requires a two-step process i.e. formation of the characteristic pattern of disulfide bridges which is followed by proteolytic cleavage and subsequent release of the C-peptide \\[[@pone.0206169.ref015]\\].\n\nRecently \u03b2-galactosidase (\u03b2-gal) has been used as a fusion protein partner, with plasmids containing the DNA sequence encoding the A- or B-chain of human insulin. The production of the fusion protein is intracellular and it appears as cytoplasmic inclusion bodies in the bacterial host (*Escherichia coli*) \\[[@pone.0206169.ref016]\\]. However, the large size of \u03b2-gal fusion protein partner (1000 amino acids) limits the yield of insulin to a significant level. Moreover, during translation, the fusion proteins (\u03b2-gal and insulin A- or B-chain) became detached from the ribosome resulting in premature chain termination \\[[@pone.0206169.ref017], [@pone.0206169.ref018]\\]. Also, relatively high concentrations of the A- and B-chain are required for the efficient folding during insulin chain combination to produce functional insulin. The large amounts of insulin chains at this step usually result in peptide chain aggregation which significantly reduces the yield of natively folded insulin \\[[@pone.0206169.ref019], [@pone.0206169.ref020]\\]. The big demand for affordable human insulin therefore requires the development of a more efficient production system and also introduction of suitable folding conditions. To overcome the above mentioned complications, in our system for expression, purification and chain combination, we have used human \u03b1B-crystallin (\u03b1B-Cry) as both fusion protein partner and a molecular chaperone mediating insulin folding. The intrinsic ability of this chaperone protein to form large oligomers \\[[@pone.0206169.ref021]\\] may also facilitate the purification of the target peptides. Overall, we have designed a streamlined expression and purification approach for the insulin A- and B-chain which each step is straightforward with the high yields. Our method might be also applicable to the production of other therapeutic peptides.\n\nMaterials and methods {#sec002}\n=====================\n\nBis-1-anilino-8-naphthalene sulfonate (bis-ANS), thioflavin T (ThT), sodium tetrathionate, sodium sulfide, cyanogen bromide, and other chemicals were purchased from Sigma. Standard insulin and goat anti-rabbit IgG peroxidase were purchased from Sigma-Aldrich Company. The dialysis tube (2 kDa cut-off) was purchased from Spectrum Scientific Company. Anti \u03b1B-Cry antibody was a generous gift of Professor Samuel Zigler (Johns Hopkins School of Medicine). Gel filtration media and Ni-NTA matrix were from GE Healthcare and Qiagen, respectively.\n\nGene and plasmid construction {#sec003}\n-----------------------------\n\nA schematic representation of the expression vector pET28b (+) (Novagene) and additional sequence information of the construct are given in **[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**. As shown in [**Fig 1A** and **1B**](#pone.0206169.g001){ref-type=\"fig\"}, the expression constructs consist of human \u03b1B-Cry (*CRYAB*) ligated to the B-chain (\u03b1B-BC) or the A-chain (\u03b1B-AC) of the human insulin genes. The fusion gene is flanked by *Nco*I and *Not*I restriction sites. Also, the *Nde*I site was designed to encode a methionine residue which provides a CNBr cleavage site at the end of the fusion partner protein (\u03b1B-Cry). Additionally, in the sequence of \u03b1B-Cry, methionine 68 and proline 130 were replaced, respectively, with isoleucine and valine to provide resistance against cleavage by CNBr and formic acid. These substitutions were made using QuikChange II XL Site-Directed Mutagenesis Kit (Stratagene) following the manufacturer\\'s instruction. To produce the constructs, the fusion gene sequences with designated restriction sites were chemically synthesized and then cloned into pET28b (+). The amino acid sequence of the desired fusion genes is presented in **[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**. The pET28b (+) containing the A-chain was engineered by designing a 6 His-tag at the N-terminus of the *CRYAB* gene. This affinity tag was incorporated to facilitate the purification of the A-chain.\n\n![Schematic illustration of sequence of the designated vectors.\\\npET28b (+) vectors harboring fusion genes are depicted here. The positions of *Nco*I, *Nde*I and *Not*I restriction sites are also indicated. (**A**) and (**B**) respectively stand for the vectors including the fusion gene of \u03b1B-BC and \u03b1B-AC. The translation frame under the vectors correspond to the amino acid sequence of the fusion proteins. The sequence of \u03b1B-Cry fusion partner is indicated in blue. (**M**) stands for the single letter abbreviation of methionine serving as the CNBr cleavage site. The amino acid sequence of B- and A-chain of human insulin is indicated in red. Isoleucine and valine residues (in pink) generated by site direct mutagenesis in \u03b1B-Cry gene instead of methionine and proline, respectively.](pone.0206169.g001){#pone.0206169.g001}\n\nExpression of the fusion proteins and isolation of inclusion bodies {#sec004}\n-------------------------------------------------------------------\n\nThe constructs were transformed into *E*. *coli* XL1-blue \\[[@pone.0206169.ref022]\\]. After extraction of the vectors (Miniprep Kit), the correct DNA sequence was confirmed by sequencing. Then, the constructs (\u03b1B-AC and \u03b1B-BC) with correct sequences were transformed into *E*. *coli* BL21 (DE3) (Stratagene). Typically, a single colony was used to inoculate a 15 mL overnight culture grown at 37\u00b0C in LB medium containing antibiotic (50 \u03bcg/mL kanamycin) with constant shaking. The overnight cultures were then used to inoculate 1000 ml Luria-Bertani (LB) medium supplemented with antibiotics. The cells were grown at 37\u00b0C to an optical density of 1 at 600 nm, prior to induction of protein expression by the addition of isopropyl-\u03b2-D-thiogalactopyranoside (IPTG) at a final concentration of 0.25 mM. The cells were grown overnight, then centrifuged at 7000 \u00d7 g for 7 min and stored at -20\u00b0C, prior to further processing. To constitute an effective expression system for the peptides, we took advantage of the fusion partner \u03b1B-Cry, which promotes the formation of inclusion bodies in *E*. *coli*. These misfolded and insoluble aggregates offer multiple benefits such as enhanced protein expression (\\>30%), high purity, resistance to proteolytic digestion, and convenient isolation of the inclusion bodies \\[[@pone.0206169.ref018], [@pone.0206169.ref023], [@pone.0206169.ref024]\\]. For partial purification and complete solubilization of the inclusion bodies, the cell pellets (1 gr) containing \u03b1B-AC and \u03b1B-BC fusion proteins, were resuspended separately in 10 mL of lysis buffer (50 mM Tris-HCl, pH 8.0, 5 mM EDTA) and then homogenized. The bacterial cells were lysed by sonication, and the cell debris was centrifuged by a low-speed centrifugation (6000 \u00d7 g for 7 min). The pellets were further washed as a suspension in a wash buffer A (50 mM Tris-HCl, pH 8.0, 5 mM EDTA) with 0.1% (v/v) Triton X-100 and 1 M urea, sonicated for 30 seconds in an ice bath, and centrifuged (8000 \u00d7 g, 4\u00b0C, 20 min). To collect the inclusion bodies (\u03b1B-AC and \u03b1B-BC), the debris was resuspended in 0.1gr/mL in the binding buffer B (20 mM Tris-HCl, pH 6.7, 5 mM 2-Mercaptoethanol) for \u03b1B-AC and binding buffer C (50 mMTris-HCl, pH 8.4, 5 mM 2-Mercaptoethanol) for \u03b1B-BC, both containing 8 M urea. The inclusion bodies containing the fusion proteins were also recovered from the debris by centrifugation at 10000 \u00d7 g for 30 min at 4\u00b0C and the supernatants were passed through a filter (0.2 \u03bcM pore diameter) prior to the purification. In each step, purity of the protein samples was analyzed by 12% SDS--PAGE gel. Also, a 10 L of the culture medium was used for each fusion protein.\n\nPurification of the fusion proteins \u03b1B-AC and \u03b1B-BC {#sec005}\n---------------------------------------------------\n\nPurification of \u03b1B-AC and \u03b1B-BC fusion proteins were done using Ni-NTA (13 mL) (Qiagen, Germany) and DEAE Sepharose (Pharmacia) columns, respectively. The solubilized inclusion body containing \u03b1B-AC under reducing conditions (5 mM \u03b2-mercaptoethanol) was applied directly to a Ni-NTA (13 mL) column pre-equilibrated with binding buffer B (20 mM Tris-HCl, pH 6.7, 5 mM \u03b2-mercaptoethanol) containing 8 M urea. The column was washed with the respective binding buffer and the bound protein was eluted with elution buffer (20 mM Tris-HCl, pH 6.7, 5 mM \u03b2-mercaptoethanol) containing 8 M urea and 100 mM imidazole at a 1 mL/min flow rate. The eluted fractions were collected and analyzed on reducing SDS--PAGE and western blot. For purification of \u03b1B-BC, the solubilized inclusion body containing \u03b1B-BC was loaded on DEAE Sepharose column, pre-equilibrated in binding buffer C containing 8 M urea, at about 30 mg of total protein per mL of resin. The protein fraction of \u03b1B-BC was washed with equilibration buffer at pH 8.0 and the bound material was eluted with 100 mM NaCl in the same buffer.\n\nGel electrophoresis and western blot analysis {#sec006}\n---------------------------------------------\n\nExpression, isolation of the inclusion body and quality of purification were determined by visualization of total proteins, following separation by a 12% polyacrylamide gel electrophoresis. The protein bands were visualized using Coomassie brilliant blue (CBB) staining \\[[@pone.0206169.ref025]\\]. Also, cleavage of the fusion proteins and isolation of the peptides were analyzed on an 18% SDS-PAGE gel. \u03b1B-AC and \u03b1B-BC were identified by western blot analysis using a primary antibody which specifically recognizes the \u03b1B-Cry, and addition of the horseradish peroxidase second antibody (goat anti-rabbit IgG peroxidase). The visualization was achieved with diaminobenzidine as the enzyme substrate \\[[@pone.0206169.ref026]\\].\n\nCleavage of the fusion proteins and purification of insulin A- and B-chain {#sec007}\n--------------------------------------------------------------------------\n\nThe peptides (A- and B-chain of insulin) were cleaved from the fusion protein partner (\u03b1B-Cry) using CNBr. There are various chemical cleavage methods that offer an inexpensive procedure for removal of the fusion tags or partners. In this case, the purified and lyophilized \u03b1B-AC and \u03b1B-BC proteins which obtained from a 10 L of Luria broth (LB) culture medium were resuspended individually in 70% (v/v) formic acid at a concentration of 50 mg/mL. Also, CNBr (350 mg/mL) prepared in 70% (v/v) formic acid was added at a 100-fold molar excess over protein and the mixture was allowed to react for 24 h, at room temperature. The CNBr and formic acid were removed by overnight dialyzing against 5 L of double distilled water using an appropriate dialysis tube (2 kDa molecular weight cut-off). The precipitates containing \u03b1B-AC (\\~ 1300 mg) and \u03b1B-BC (\\~ 1700 mg) were stored at -20\u00b0C.\n\nThe sulfitolysis experiment {#sec008}\n---------------------------\n\nThe digested and lyophilized \u03b1B-AC and \u03b1B-BC were individually dissolved (100 mg/mL) in 300 mM Tris-HCl, pH 8.6 containing 8 M urea. S-sulfonated derivatives of the peptide mixtures were prepared by adding sodium sulfite and sodium tetrathionate to the final concentration of 250 and 80 mM, respectively. After incubation at room temperature for 3 h, the pH was adjusted to 5.0 with acetic acid \\[[@pone.0206169.ref027]\\] and the mixtures were dialyzed twice against 3 L of double distilled water. The resulting white precipitates containing \\~1210 mg \u03b1B-AC and \\~1630 mg \u03b1B-BC were centrifuged at 9000 \u00d7 g for 45 min at 4\u00b0C and then lyophilized.\n\nPurification of the insulin peptide chains {#sec009}\n------------------------------------------\n\nThe B-chain was separated from its fusion partner protein using gel filtration chromatography on a Sephadex G-50 (90 \u00d7 1, GE Healthcare) column which had been equilibrated in 1M glacial acetic acid and eluted with the same solvent \\[[@pone.0206169.ref027], [@pone.0206169.ref028]\\]. Due to the insolubility of the A-chain in 1M acetic acid, 20 mM Tris-HCl (pH 8.0) containing 8 M urea was used for purification of this peptide \\[[@pone.0206169.ref029]\\]. For purification of B-chain, the lyophilized S-Sulfonated derivatives (100 mg) were dissolved in 2 mL acetic acid (1M) and then loaded onto a Sephadex G-50 column. The column was equilibrated and the B-chain was eluted with 1M acetic acid. The flow rate and fraction size were 0.2 mL/min and 2 mL, respectively. The optical density of the fractions was determined at 276 nm. The A-chain was purified in the same way in 20 mM Tris-HCl pH 8.0 containing 8 M urea as the elution solvent. For gel filtration chromatography, all the buffers and solvents were degassed.\n\nApplication of \u03b1B-Cry to increase the yield of insulin A-B chain combination {#sec010}\n----------------------------------------------------------------------------\n\nThe combination of A- and B-chain was performed according to a standard protocol \\[[@pone.0206169.ref020], [@pone.0206169.ref027]\\]. The A- and B-chain (weight ratio of 2:1) were dissolved in degassed buffer (0.1 MGly/NaOH pH 10.5). Dithiothreitol was rapidly added to the peptide solution to a molar ratio 1:1 of SH to SSO~3~^2^. At the indicated time points (0, 12, 24, 36 and 48 h), the aliquots were withdrawn and precipitated proteins were removed by centrifugation (9000 \u00d7 g for 30 min at 4\u00b0C). Refolding was performed under similar conditions in the presence of \u03b1B-Cry in a molar ratio of 0.04 for \u03b1B-Cry and insulin B-chain. The retention time of \u03b1B-Cry was determined on a C18 RP-HPLC column ([S1 Fig](#pone.0206169.s001){ref-type=\"supplementary-material\"}). The yield of insulin A-B chain combination was analyzed by reverse phase chromatography, measuring the optical density at 214 nm. The positions of the elution peaks corresponding to insulin A- and B-chain as well as native insulin were indicated by subjecting their standard counterparts. The experiments were repeated 3 times. The standard A- and B-chain were also prepared from insulin of Sigma, following previous publication \\[[@pone.0206169.ref030]\\].\n\nPurification of natively folded human insulin {#sec011}\n---------------------------------------------\n\nAmmonium sulfate was added to the supernatant containing the active insulin to a final concentration of 500 mM. Then, the solution was passed through a 0.22 \u03bcm filter (Biofil), and subjected to phenyl sepharose (GE Healthcare) column (1 \u00d7 20 cm, Pharmacia) at about 20 mg per mL resin. The column was connected to a Pharmacia liquid chromatography system, pre-equilibrated with 20 mM Tris-HCl, 500 mM ammonium sulfate at pH 8.0. The equilibration buffer was also used to wash the bound materials. Then, insulin was eluted with a descending linear gradient of ammonium sulfate (500--0 mM) in 20 mM Tris-HCl, pH 8.0. The experiment was done at flowrate of 1 mL/min with a fraction size of 2 mL and the eluates were analyzed by measuring the optical density at 276 nm. In the elution profile, the position of native insulin was identified by the addition of an authentic sample (standard human insulin). The elution with the same condition was considered for the natively folded insulin. The fractions corresponding to the native insulin were collected and dialyzed extensively against acetic acid (1 M) at 4\u00b0C \\[[@pone.0206169.ref031], [@pone.0206169.ref032]\\].\n\nComparing the natively folded human insulin with a standard insulin {#sec012}\n-------------------------------------------------------------------\n\n### RP-HPLC analysis {#sec013}\n\nFor comparing the surface features of the recombinant and standard insulins, these two proteins were separately subjected to RP-HPLC column at 25\u00b0C (ProntoSIL 200-5-C18, 250 \u00d7 4.6 mm; Apex Scientific) equipped with UV detector (Smartline 2500; KNAUER). The insulin samples were chromatographed at a flow rate of 1 mL/min with a linear gradient of acetonitrile (24--60%) for 10 min at a constant temperature 25\u00b0C.\n\n### The fluorescence assessments {#sec014}\n\nThe intrinsic fluorescence was recorded in 50 mM sodium phosphate buffer pH 7.4, using an Agilent fluorescence spectrophotometer (Varian Cary Eclipse, USA). In order to measure the fluorescence of tyrosine (Tyr) residues, the insulin samples (1 mg/mL) were excited at 276 nm and the emission spectra were collected at the wavelength range of 280--400 nm. The excitation/emission band passes for Tyr fluorescence were set at 5/5 nm\\[[@pone.0206169.ref033]\\]. Also, bis-ANS fluorescence for the surface hydrophobicity assessment of the protein samples was performed \\[[@pone.0206169.ref034]\\]. The scans were average of three replicates.\n\n### Circular dichroism measurement {#sec015}\n\nThe CD measurements were performed in 50 mM sodium phosphate buffer pH 7.4 on a Jasco J-720 spectropolarimeter. The spectra were recorded in a 0.1 cm path length cell from 250 to 190 nm. The background CD spectrum of the buffer was subtracted from the CD spectra of the insulin chains. The reported spectra are the average of three scans \\[[@pone.0206169.ref035]\\]. The protein concentration was adjusted to 2 mg/mL by UV absorption at 276 nm using an extinction coefficient of 1.08 for 1.0 mg/mL. The CD experiments were done at room temperature.\n\n### Near infrared (NIR) spectroscopy assessment {#sec016}\n\nRecombinant and standard insulin were analyzed by near infrared (NIR) Spectroscopy. The NIR spectra were collected on a NIRS XDS series Vis-NIR spectrometer (Metrohm, Switzerland) with the reflectance mode, in the range of 8,000 to 4,000 cm^-1^ and a resolution of 8 cm^-1^ at room temperature. For each dried sample, the NIR spectra were recorded through the bottom of the glass plate. To decrease the possible effect of uneven distribution of dried insulin, the protein samples were rotated and measured again \\[[@pone.0206169.ref036]\\]. Each spectrum consisted of the average of 3 replicates.\n\n### Insulin aggregation analysis {#sec017}\n\nThe aggregation of insulin (0.5 mg/mL) was induced in the presence of 20 mM dithiothreitol (DTT) in phosphate buffer (50 mM, pH 7.2) \\[[@pone.0206169.ref037]\\]. The kinetic of aggregation was assessed at 40\u00b0C for 20 min by recording the optical density at 360 nm, using a T90^+^ UV-Vis spectrophotometer (PG Instruments) which equipped with Peltier temperature controller. Also, formation of amyloid fibril was monitored by ThT fluorescence analysis \\[[@pone.0206169.ref038]\\]. Fibril formation was initiated by incubation of insulin (2 mg/mL) in 20% acetic acid pH 2.0, containing 150 mM NaCl for 5 h, at 60\u00b0C \\[[@pone.0206169.ref039], [@pone.0206169.ref040]\\]. At the end of incubation, 10 \u03bcM ThT was added to the protein solution and fluorescence measurement was done using an excitation wavelength at 450 nm \\[[@pone.0206169.ref030]\\]. The excitation and emission slits were 5 and 10, respectively.\n\n### The size exclusion chromatography assessment {#sec018}\n\nThe standard and recombinant insulin samples (2 mg/mL) were individually dissolved in Tris HCl (10 mM, pH 7.4), containing 0.11 mM ZnCl~2~ \\[[@pone.0206169.ref041]\\]. Under this condition, the molar ratio of insulin to zinc was about 3, favoring formation of the protein hexameric form \\[[@pone.0206169.ref042], [@pone.0206169.ref043]\\]. In this study, a KNAUER HPLC system equipped with the analytical size exclusion column (300 \u00d7 8 mm SEC column PSS SUPREMA) and DAD 2.1 UV detector was used. During the experiments, the temperature of HPLC column was also kept at 37\u00b0C. The HPLC results were average of three independent repetitions.\n\n### The insulin tolerance test (ITT) {#sec019}\n\nFor activity assessment of the natively folded fraction of final product, the insulin tolerance test (ITT) was performed in fasted mice \\[[@pone.0206169.ref027]\\]. Inbred *BALB/c* male mice between twelve to fifteen week old mice (weighing 22--27 g) were randomly divided into three groups of 7 mice each. The food was taken away 6 h before the start of the test. The control group received a subcutaneous saline injection (0.1 mL per 100 g body weight) while the experimental groups received a saline injection containing the standard and recombinant insulin. The samples were injected subcutaneously at a dose of 0.75 U/kg based on the unit of standard insulin \\[[@pone.0206169.ref027]\\]. One unit (U) of insulin is defined as 36.3 \u03bcg of the polypeptide. Blood glucose was measured with a glucometer (Accu-Chek Performa, Germany). To determine a basal glucose level, blood glucose was measured in a drop of blood obtained from the tail vein before the application of the samples. After injection of insulin product or saline, blood glucose measurements were done at time intervals of 10, 20, 30, 40, 60, 90, 120, 150 and 180 min. Before injection, blood glucose measurements were done for 60 min. The mice were given free access to food immediately after the experiment. All experiments followed the ethical guidelines for animal experiments were described and approved by the committee for experiments with laboratory animals of the National Research Ethics Committee. In this study, the mice were purchased from Shiraz University of Medical Sciences (Shiraz, Iran). The mice were fed normal chow which has been supplied in wire lid food hoppers, making the food and water easily accessible to them. Also the mice were housed at 23--24\u00b0C under a daily cycle of 12 h of light and dark.\n\nStatistical analysis {#sec020}\n--------------------\n\nThe data were analyzed in GraphPad Prism and presented as mean \u00b1S.E.M. The significance of data was calculated using a two-tailed t-test for independent samples \\[[@pone.0206169.ref027]\\].\n\nResults {#sec021}\n=======\n\nDesigning an appropriate fusion partner for individual expression of insulin A- and B-chain in *Escherichia coli* {#sec022}\n-----------------------------------------------------------------------------------------------------------------\n\nIn order to increase the expression level in *Escherichia coli*, we fused the A- and B-chain of human insulin to the \u03b1B-Cry gene. The internal methionine (residue 68) in wild-type \u03b1B-Cry was mutated to isoleucine in order to prevent the formation of additional unwanted cleavage products in the presence of CNBr that would otherwise need to be removed from the final product (**[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**). The Asp-Pro sequence (residues 129 and 130) in \u03b1B-Cry is sensitive to the cleavage by formic acid and this reagent is needed for releasing of the insulin chains from the fusion proteins in the cleavage step. Therefore, Pro 130 was mutated to valine (**[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**).\n\nAlso, design of the restriction site of *Nde*I (5\\'CATATG3\\') incorporated a methionine residue at the carboxy terminal end of human \u03b1B-Cry, at the boundary between the carrier protein and the desired peptide. Therefore, the expressed fusion protein can be cut at the boundary by cyanogen bromide treatment. Finally, the engineered expression plasmids were constructed and their sequence precision was confirmed (**[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**). Hence, the final constructs contained the N-terminal \u03b1B-Cry partner, a CNBr cleavage site and the C-terminal A- or B-chain of human insulin. In addition, the construct for A-chain (\u03b1B-AC) contained an N-terminal hexa-histidine tag to enable rapid and straightforward purification. The details of these constructs are given in **[Fig 1](#pone.0206169.g001){ref-type=\"fig\"}**.\n\nExpression and purification of the fusion proteins {#sec023}\n--------------------------------------------------\n\nThe desired constructs were transformed to XL1blue cells. Following extraction and sequencing, the vectors were then transformed to BL21 cells. BL21 cells harboring recombinant plasmids pET28b (+) were induced separately to express the \u03b1B-AC and \u03b1B-BC fusion proteins, using IPTG inducer. The SDS--PAGE analysis suggested that the expressed fusion proteins were completely insoluble and supernatant of the cell lysates after sonication had no obvious desired proteins (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**, lanes 2 and 5). Based on SDS-PAGE results the expressed fusion proteins (\u03b1B-AC and \u03b1B-BC) were constituted approximately 31% of the whole cell extract with the complete partitioning into the insoluble fraction (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**). The inclusion bodies containing the \u03b1B-AC and \u03b1B-BC were isolated by a low speed centrifugation. As a pre-step purification, the partially soluble contaminating proteins were removed by Buffer A (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**, lanes 3 and 6). Also, a solution of 8 M urea was used for solubilization of the fusion proteins (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**, lanes 4 and 7). The solubilized inclusion bodies contain approximately 48.1% and 54.4% of the \u03b1B-AC and \u03b1B-BC, respectively.\n\n![The SDS--PAGE analysis of the fusion proteins expression and purification.\\\nThe protein samples in different steps of the purification of fusion proteins (\u03b1B-AC and \u03b1B-BC) were subjected to gel electrophoresis analysis. Lane 1 indicates molecular mass marker. Lanes 2, 3 and 4, respectively, stand for the supernatant, the pellet subjected to the subsequent washing with Triton X-100 and urea, and \u03b1B-AC fusion protein solubilized in 8 M urea. Lanes 5, 6 and 7 indicate SDS-PAGE analyses of protein in the similar steps during purification of \u03b1B-BC fusion protein.](pone.0206169.g002){#pone.0206169.g002}\n\nThe solubilized inclusion bodies were further purified by a one-step anion exchange chromatography, using DEAE matrix (for \u03b1B-BC) and immobilized metal affinity chromatography by Ni-NTA column (for \u03b1B-AC) (**[Fig 3A](#pone.0206169.g003){ref-type=\"fig\"}**, lanes 3 and 4). Eventually, using these methods, about 130 and 170 mg of the \u03b1B-AC and \u03b1B-BC were obtained from 1 L of the LB medium. The purified fusion proteins have an apparent molecular weight of 23--25 kDa (**[Fig 3A](#pone.0206169.g003){ref-type=\"fig\"}**) which is corresponding to the sum of molecular mass of the \u03b1B-Cry fusion partner (20 kDa) and insulin A-chain (2.4 kDa) or B-chain (3.4 kDa).\n\n![SDS--PAGE and western blot analyses of \u03b1B-Cry and the fusion proteins.\\\nThe fusion proteins were analyzed by reducing SDS-PAGE (gel 12%) (**A**), and western blot (**B**). Lane 1 shows molecular mass marker. Also, lanes 2, 3 and 4, respectively, indicate \u03b1B-Cry, \u03b1B-AC and \u03b1B-BC. In the western blot analyses, the antibody against \u03b1B-Cry was used.](pone.0206169.g003){#pone.0206169.g003}\n\nAs shown in the **[Fig 3A](#pone.0206169.g003){ref-type=\"fig\"}**, the expressed fusion proteins appeared highly pure. In addition, to confirm the identity of the purified bands, the fusion proteins were detected by western blot analysis which has been developed with an anti \u03b1B-Cry antibody recognizing partner protein in the fusion molecules (**[Fig 3B](#pone.0206169.g003){ref-type=\"fig\"}**, lanes 3 and 4).\n\nCleavage of the fusion protein, sulfitolysis and isolation of S-sulfonated insulin chains {#sec024}\n-----------------------------------------------------------------------------------------\n\nAfter purification of the fusion proteins, the \u03b1B-Cry carrier was efficiently removed by the action of a site-specific cleaving reagent. As shown in **[Fig 4](#pone.0206169.g004){ref-type=\"fig\"}** after 24 h incubation with CNBr, the fusion proteins were cleaved to the significant level. At this stage three fragments including \u03b1B-Cry partner protein (20 kDa), insulin A-chain (2.4 kDa) or B-chain (3.4 kDa) and trivial quantity of uncleaved fusion protein (\\~25 kDa) were detectable on the SDS-PAGE gel (**[Fig 4](#pone.0206169.g004){ref-type=\"fig\"}**). Quantitative comparison of the digest efficiency, using Image J software, revealed the significant CNBr cleavage of the fusion protein. The cleavage of \u03b1B-BC and \u03b1B-AC was estimated to be approximately 68.8 and 75.2%, respectively **([Fig 4](#pone.0206169.g004){ref-type=\"fig\"}**, lanes 5 and 6). The efficient cleavage occurred within 24 h with a CNBr to methionine molar ratio of 100:1 at room temperature. Increasing the incubation time or amount of CNBr did not improve the cleavage efficiency. Chemical cleavage experiment demonstrated that the full length fusion protein bands appreciably disappeared giving rise to the new bands of the smaller sizes corresponding to fragments \u03b1B-Cry (20 kDa) and insulin A- or B-chain. Prior to separation of the insulin peptides from their corresponding fusion proteins oxidative sulfitolysis which prevents improper disulfide binding, was performed. This experiment was done by adding sodium sulfite and sodium tetrathionate to the solutions containing insulin chains as described in the experimental section.\n\n![SDS--PAGE analyses of the fusion proteins cleavage.\\\nThe fusion proteins were subjected to CNBr cleavage and the protein samples were analyzed by reducing SDS-PAGE (gel 18%). Lane 1 stands for the molecular mass marker. Lanes 2, 3 and 4 indicate \u03b1B-Cry, \u03b1B-AC and \u03b1B-BC, respectively. Lanes 5 and 6 show \u03b1B-BC and \u03b1B-AC fusion proteins after cleavage by the CNBr.](pone.0206169.g004){#pone.0206169.g004}\n\nAfter that, using a single-step gel filtration chromatography, the digested mixtures were separated into two peaks (**[Fig 5](#pone.0206169.g005){ref-type=\"fig\"}**). The eluted fractions were then run on SDS--PAGE (insets of **[Fig 5](#pone.0206169.g005){ref-type=\"fig\"}**). In [**Fig 5A** and **5B**](#pone.0206169.g005){ref-type=\"fig\"}, the fractions eluted as the first peak contain undigested fusion protein and \u03b1B-Cry. Also, the fractions eluted as the second elution peak which belong to the insulin chains were collected and concentrated. After confirming suitable purification of the insulin peptides from residual fragments, the peptide chains were dialyzed twice against double distilled water (pH 5.5) and then lyophilized.\n\n![Gel filtration and SDS-PAGE analyses of the digested fusion proteins.\\\nThe digest products of two fusion proteins with CNBr were individually applied to a Sephadex G50 gel filtration column (1.6 x 94 cm). The experiments were done by dissolving \u03b1B-BC in 1 M acetic acid and \u03b1B-AC in 20 mM Tris containing 8 M urea at pH 8.0. The elution profile and reducing SDS-PAGE profile (gel 18%) (the insets) corresponding to \u03b1B-AC (**A**) and \u03b1B-BC (**B**) are indicated.](pone.0206169.g005){#pone.0206169.g005}\n\nFinally, to verify the efficiency of our purification methods and to validate the identity of the recombinant A- and B-chain, purified and lyophilized peptides were analyzed by SDS--PAGE (gel 18%) and analytical RP-HPLC (**[Fig 6](#pone.0206169.g006){ref-type=\"fig\"}**). Then, the dialysis against pure water at pH 5.5 results in precipitation of insulin A- and B-chain. Under such condition, the insulin chains are known to be highly insoluble whereas other possible contaminations were appeared in the soluble phase. The migration pattern of insulin A- and B-chain in the SDS--PAGE gel is also indicated in the inset of **[Fig 6](#pone.0206169.g006){ref-type=\"fig\"}**.\n\n![**RP-HPLC and SDS-PAGE analyses of A- and B-chain of human insulin.** The purified samples of A- and B-chain of human insulin were subjected to RP-HPLC and SDS-PAGE analyses. RP-HPLC analysis was conducted on C18 analytical column for the A- and B-chain of both standard insulin (STD) and human recombinant insulin product (R). The inset also shows the SDS--PAGE profile (gel 18%) of the purified chains of human insulin.](pone.0206169.g006){#pone.0206169.g006}\n\nSince the insulin peptides have small molecular masses than the fusion partner, the highly pure peptides could be obtained by gel filtration chromatography without needing a help of other kinds of affinity tags or high-resolution chromatographic methods, such as HPLC. Approximately, 27.6 mg and 43.2 mg of the purified S-sulfonated A- and B chain were generally obtained from 500 mg of the digested mixture.\n\nAnalytical RP-HPLC indicated the presence of a single major component at appreciate purity (**[Fig 6](#pone.0206169.g006){ref-type=\"fig\"}**). Also, using standard A- and B-chain and measuring their retention time of elution, the identity of the recombinant peptides was analyzed. Our insulin peptides were highly pure and eluted synchronously with the corresponding peptide chains of the standard insulin (**[Fig 6](#pone.0206169.g006){ref-type=\"fig\"}**).\n\nApplying human \u03b1B-Cry to increase the efficacy of insulin chain combination and folding {#sec025}\n---------------------------------------------------------------------------------------\n\nAs shown in **[Fig 7](#pone.0206169.g007){ref-type=\"fig\"}**, the chain combination experiment was conducted by the incubation of A- and B-chain of human insulin with a weight ratio of 2:1, respectively, for 48 h. As revealed by the analytical RP-HPLC analysis, in the absence of appropriate chaperone, approximately after 12 h of incubation, the formation of natively folded human insulin molecules was reached to its steady state (**[Fig 7](#pone.0206169.g007){ref-type=\"fig\"}**). As shown in **[Fig 7](#pone.0206169.g007){ref-type=\"fig\"}**, the usable chains available for the further combination and pairing were also aggregated at the initial steps of the folding process, giving rise the amount of insulin aggregates species. Spending more times had no significant effect on the formation yield of insulin molecules with native folding.\n\n![The chain combination analysis with RP-HPLC in the presence of \u03b1B-Cry.\\\n\u03b1B-Cry chaperone was used to increase efficacy of the chain combination step. (**A**) Peaks 1--4, respectively, indicate the position of oxidized A-chain, native insulin, oxidized B-chain and intermediate products presumably with the incorrect disulfide linkage. (**B**) Peaks 1--5 stand for the oxidized A-chain, native insulin, oxidized B-chain, \u03b1B-Cry and presumably the intermediate insulin species with the incorrect disulfide linkage. In both (**A**) and (**B**), the peaks correspond to the native insulin are distinguished by the white solid diamond as indicated by authentic insulin in the (**N**) line. The experiments were repeated three times.](pone.0206169.g007){#pone.0206169.g007}\n\nAs shown in **[Fig 7A](#pone.0206169.g007){ref-type=\"fig\"}**, at 12 h interval, some undesirable intermediates were developed. Due to their longer retention time, the incorrect species may present at the high hydrophobic state. Formation of these intermediates is possibly due to the wrong disulfide pairing or hydrophobic collapse of insulin A- and/or B-chain. At this condition, the amount of A- and B-chain of human insulin can be gradually decreased and this reduction was accompanied with the increment in the quantity of insulin misfolded intermediates. To solve this problem, we applied human \u03b1B-Cry to increase the yield of insulin chain combination and folding (**[Fig 7B](#pone.0206169.g007){ref-type=\"fig\"}**). As shown by the analytical RP-HPLC analysis the joining of the S-sulfonated A- and B-chain to form natively folded and active hormone, as assisted by the chaperone (\u03b1B-Cry), gives a significantly improved yield. Moreover, the incorrect protein species, indicating longer retention time, were fully disappeared as \u03b1B-Cry was used to assist the chain combination and folding of the human recombinant insulin.\n\nBased on the results of respective RP-HPLC profiles, during 48 h of the incubation, the A- and B-chain in the presence of \u03b1B-Cry remain largely soluble to have further chance of the correct pairing which subsequently increase the quantity of natively folded insulin molecules. However, in the absence of \u03b1B-Cry, the amount of insulin chains was gradually reduced without generating native insulin molecules (peaks that marked by diamond correspond to native insulin). The efficiency of insulin chain combination and folding during the mentioned times were calculated from the total and corresponding peak areas. The efficiency in each incubation time is indicated in **[Table 1](#pone.0206169.t001){ref-type=\"table\"}**. In particular, by developing the incubation time in the presence of \u03b1B-Cry, the area corresponding to the insulin molecules with correct folding was remarkably increased (**[Fig 7B](#pone.0206169.g007){ref-type=\"fig\"}**, diamond marked peaks). It should be noted that with the increase in the intensity of the peak for native insulin, the intensity of the peaks corresponding to the A- and B-chain was decreased. However, the samples which were exposed to folding conditions in the absence of \u03b1B-Cry, did not show significant differences in the peak corresponding to native insulin with the respect to the increase in incubation time. It should be noted that the yield was calculated on the basis of the amounts of all products for each incubation time.\n\n10.1371/journal.pone.0206169.t001\n\n###### The refolding efficiency (%) of recombinant human insulin.\n\n![](pone.0206169.t001){#pone.0206169.t001g}\n\n  Time of incubation (h)   \\- \u03b1B-Cry     \\+ \u03b1B-Cry\n  ------------------------ ------------- -------------\n  12                       5.0 \u00b1 1.32    6.3 \u00b1 0.86\n  24                       7.8 \u00b1 1.03    11.7 \u00b1 1.05\n  36                       8.0 \u00b1 0.91    19.3 \u00b1 1.79\n  48                       10.2 \u00b1 1.11   26.7 \u00b1 2.02\n\nHydrophobic interaction chromatography (HIC) is commonly used to separate natively folded and unfolded species during the purification of therapeutic proteins. In general, the separation of proteins by HIC largely depends on the hydrophobicity of protein, the temperature and the properties of the stationary and mobile phases. In this system, the elution times are related to hydrophobic contact area, therefore an increase in what affecting surface hydrophobic area of the sample results in increased binding to HIC media \\[[@pone.0206169.ref044]\\]. For purification of natively folded recombinant insulin, HIC with a phenyl sepharose hydrophobic column was used. To perform this experiment, a mixture of BSSO3^-^ (40 mg) and ASSO3^-^ (80 mg) which has been incubated with \u03b1B-Cry (\u03b1B-Cry to B-chain molar ratio of 0.04) for 48 h was used (**[Fig 8](#pone.0206169.g008){ref-type=\"fig\"}**). By reducing the concentration of ammonium sulfate, a mixture of misfolded intermediates was eluted at the initial step of the purification process (**[Fig 8](#pone.0206169.g008){ref-type=\"fig\"}**). In this purification step, 27.08 mg of the human recombinant insulin, similar to the standard insulin, was obtained which corresponds to a yield of 22.57%. As authentic insulin was subjected to the same column, it displayed a similar elution time to that of natively folded human recombinant insulin. It should be noted that, as far as the elution time is concerned, our recombinant human insulin product and standard insulin of Sigma was eluted closely at the similar concentration of ammonium sulfate (250 mM).\n\n![Purification of the insulin peptide chains on phenyl sepharose column.\\\nThe natively folded human insulin was purified using hydrophobic interaction chromatography (HIC) on a phenyl sepharose column. The elution was achieved with a reverse linear gradient of ammonium sulfate (500--0 mM) in 20 mM Tris HCl, pH 8.0. The standard insulin in the similar condition was also subjected to the same column for comparing the elution profile of the natively folded human recombinant insulin.](pone.0206169.g008){#pone.0206169.g008}\n\nStructural and functional characterization of the human recombinant insulin product {#sec026}\n-----------------------------------------------------------------------------------\n\n### Analytical RP-HPLC for validation of native folding of the recombinant insulin {#sec027}\n\nThe lyophilized powder of purified insulin product obtained from the phenyl sepharose column was dissolved in 0.5 mL acetic acid (20%), filtered through 0.22 \u03bcm filter and subjected to a hydrophobic C18 analytic column. As shown in **[Fig 9](#pone.0206169.g009){ref-type=\"fig\"}**, to confirm the surface identity of our recombinant insulin product, the standard insulin was also individually subjected to the HPLC column, under similar conditions. As indicated in **[Fig 9](#pone.0206169.g009){ref-type=\"fig\"}**, human insulin product and standard insulin displayed similar retention time on this column.\n\n![Comparing the RP-HPLC elution profile of the two insulin.\\\nThe human recombinant insulin was compared with standard insulin based on their elution profiles on the analytical RP-HPLC using C18 column.](pone.0206169.g009){#pone.0206169.g009}\n\nIt is essential to note that the standard insulin and our insulin product might be in their monomeric state as they dissolved in a solvent containing 20% acetic acid \\[[@pone.0206169.ref040]\\]. When recombinant and standard insulin were mixed together and subjected to the same chromatography procedure, they were eluted as a single peak. Their similar elution profile on C18 column suggests their comparable structural folding and exposed surface characteristics.\n\n### Structural analyses of the human recombinant insulin product {#sec028}\n\nThe structure of our insulin product was compared with that of standard insulin using fluorescence, circular dichroism (CD) and near infrared (NIR) spectroscopic analyses. The highly sensitive technique of intrinsic protein fluorescence has been widely used to compare the structure of proteins. It has been previously indicated that the structural alteration of insulin can be directly detected by Tyr fluorescence measurement. Human insulin contains four Tyr residues, upon exposing and/or unfolding its molecular structure, the quantum yield of these residues is altered \\[[@pone.0206169.ref033]\\]. Tyr fluorescence analyses (**[Fig 10A](#pone.0206169.g010){ref-type=\"fig\"}**) showed that the two insulin samples hold similar intensity. Also, the ability of bis-ANS to bind to the exposed hydrophobic patches of the two insulin samples indicated no significant difference (**[Fig 10B](#pone.0206169.g010){ref-type=\"fig\"}**). The results of both Tyr and bis-ANS florescence study suggested that two insulin samples have appreciable structural similarity. Also, the secondary structure content of the two insulin samples was analyzed and compared by CD and NIR spectrometry. The far UV-CD spectra of both insulin samples (**[Fig 10C](#pone.0206169.g010){ref-type=\"fig\"}**) are very similar, suggesting that they both adopted a similar \u03b1-helical secondary structure. Due to the high ratio of disulfide bonds existing in insulin, the CD signal at 251 nm has been previously used to monitor the conformation and overall disulfide bond pattern in this hormone \\[[@pone.0206169.ref045], [@pone.0206169.ref046]\\]. Judging from the CD signal at this wavelength, it is concluded that disulfide bond arrangement is similar in these two insulin samples.\n\n![Structural characterization of natively folded human recombinant insulin.\\\nTyrosine fluorescence emission spectra of the natively folded and standard insulin. The insulin samples (1.0 mg/mL) were excited at 276 nm, with emission scans from 280 to 400 nm. (**B**) The solvent exposed hydrophobic surface was detected by bis-ANS fluorescence. The protein samples (1.0 mg/mL) containing 10 \u03bcM bis-ANS were incubated at room temperature for 20 min. Upon excitation at 350 nm, the bis-ANS fluorescence spectra were measured from 400 to 600 nm. (**C**) The far UV-CD spectra of the insulin samples. The protein samples (2 mg/mL) were dissolved in phosphate-buffered (50 mM phosphate pH 7.4) and the CD spectra were scanned from 260 to 195 nm (**D**) The NIR spectra of powder of the two insulin are indicated. Band assignments are depicted in the figure.](pone.0206169.g010){#pone.0206169.g010}\n\nThe NIR spectra of these insulin samples were also collected (**[Fig 10D](#pone.0206169.g010){ref-type=\"fig\"}**). Likewise, two other bands in the regions 5900--5700 cm^-1^ correspond to the first overtones of the C--H stretching. The band at 5,150 and 6,900 originate from the vibrations in the water molecules and the frequency depends on the environment and hydrogen bonding of the water molecules. These two peaks show that the insulin samples contain water molecule in their structures. The bands at the 5200--4000 cm^-1^ have been attributed to the combinations of amide I, amide II, amide III, amide A, amide B and C--H stretching bands \\[[@pone.0206169.ref036], [@pone.0206169.ref047]\\]. Overall, these bands correlated to the secondary structure of the insulin samples. In agreement with the results of CD analyses, the data of NIR also suggest that the secondary structures in two insulin samples are in the appreciable similarity. Overall, the results of spectroscopic analyses suggest that our insulin product revealed a substantial structural similarity with that of the standard insulin.\n\nThe reduction of S--S bridges between A- and B-chain of insulin results in formation of the amorphous aggregates \\[[@pone.0206169.ref037]\\]. In the current study, the aggregation propensity of two insulin samples (standard and recombinant) were analyzed in the presence of DTT (**[Fig 11A](#pone.0206169.g011){ref-type=\"fig\"}**). Our result suggests a similar aggregation pattern for these two insulin samples. The three dimensional structure of insulin monomer also displayed in **[Fig 11B](#pone.0206169.g011){ref-type=\"fig\"}**. As reported previously, the disulfide bonds in insulin structure have specific accessibility to the environment which highly depend on the structural feature of this hormone \\[[@pone.0206169.ref048]\\]. Therefore, the similar aggregation pattern of two insulin samples is likely due to their comparable structures (**[Fig 11A](#pone.0206169.g011){ref-type=\"fig\"})**. The ThT has been previously used to monitor development of amyloid fibrils in proteins \\[[@pone.0206169.ref030], [@pone.0206169.ref040]\\]. After incubation of the insulin samples under fibrillogenic conditions, the ThT fluorescence measurement was done. The inset of **[Fig 11A](#pone.0206169.g011){ref-type=\"fig\"}** shows a significant increment in the ThT fluorescence emissions, suggesting formation of the amyloid fibrils by the both insulin samples. Moreover, the two insulin samples indicated a very similar ThT emission profiles which may suggest a comparable folding in their native structures.\n\n![The aggregation and size exclusion analyses of the recombinant and standard insulin samples.\\\n**(A)** The tendencies of two insulin samples (0.5 mg/mL) were analyzed in the presence of DTT at 40\u00b0C. The **Inset** displays ThT fluorescence emissions of two insulin samples which incubated for 5 h under fibrillogenic conditions. **(B)** Schematic representation of insulin monomer extracted from 3AIY pdb. The disulfide bonds indicated in yellow. **(C)** The size exclusion chromatography was performed in a condition favoring formation of the hexameric form. The **STD** and **R** stand for standard and the recombinant insulin samples, respectively. **(D)** The hexameric form of human insulin extracted from 3AIY pdb and red residues show His^B10^.](pone.0206169.g011){#pone.0206169.g011}\n\nIn this study, the oligomerization patterns of the two insulin samples were also compared by size exclusion chromatography (**[Fig 11C](#pone.0206169.g011){ref-type=\"fig\"}**). As reported before, when the molar ratio of insulin to zinc is about 3, this peptide hormone at physiological pH and appropriate concentration is capable to appear in its hexameric state \\[[@pone.0206169.ref042], [@pone.0206169.ref043], [@pone.0206169.ref049]\\].\n\nAs shown in **[Fig 11C](#pone.0206169.g011){ref-type=\"fig\"}**, and in agreement with previous study \\[[@pone.0206169.ref050]\\], the retention time of the standard and recombinant insulin samples were almost similar to each other. Therefore, these insulin samples are likely to be in their hexameric state (**[Fig 11D](#pone.0206169.g011){ref-type=\"fig\"}**).\n\n### *In vivo* activity assessment of the insulin product {#sec029}\n\nThough, for the *in vivo* activity assessment, insulin tolerance test (ITT) in mice was studied. An ITT is a procedure in which insulin is injected into the mice, then concentration of the blood glucose is measuring at the regular intervals \\[[@pone.0206169.ref027], [@pone.0206169.ref051]\\]. After subcutaneous injection of the insulin samples, a significant reduction in the blood glucose level was observed. Also, both standard insulin and our insulin product displayed a similar time-dependent effect on the level of blood glucose in mice (**[Fig 12](#pone.0206169.g012){ref-type=\"fig\"}**). A negative control was also prepared by the injection of saline alone into mice.\n\n![The insulin tolerance test in mice.\\\nInsulin tolerance test (ITT) was performed in mice with the age of 12 and 15 weeks, and average 25 g of their body weight. Values are mean \u00b1 S.E., n = 7/group. \\*\\*\\*p\\<0.01 compared to human insulin treated group.](pone.0206169.g012){#pone.0206169.g012}\n\nOverall, our results suggest a similar pattern of glucose lowering activity for these two insulin samples.\n\nDiscussion {#sec030}\n==========\n\nAlthough insulin does not require complex post-translational modifications its expression in *Escherichia coli* is challenging because its two peptide chains are highly susceptible to degradation by the host's enzymes \\[[@pone.0206169.ref052]\\]. To overcome the difficulties associated with expressing insulin genes in *Escherichia coli*, a suitable fusion partner (tag) protein could ensure its stability in microorganisms and facilitate its purification. The role of a fusion partner is to sequester the fused peptide in the inclusion bodies, conferring resistance to its degradation, increasing the expression rate and simplifying the recovery of the small peptide \\[[@pone.0206169.ref011], [@pone.0206169.ref053]\\]. In the current study, when human \u03b1B-Cry was selected and designed as a fusion partner for high expression level of the insulin chains, the following beneficial factors were considered. First, the size of \u03b1B-Cry (175 amino acids) is \\~6 times smaller than the \u03b2-gal fusion protein partner (1000 amino acids) \\[[@pone.0206169.ref016]\\], which subsequently increases the insulin production yield to a significant level. Second, \u03b1B-Cry has an intrinsic ability to form large oligomers in solution \\[[@pone.0206169.ref021], [@pone.0206169.ref054], [@pone.0206169.ref055]\\], and therefore, it may provide an easier condition to separate the A- and B-chain from their fusion partners by a suitable size exclusion protocol. Third, the \u03b1B-Cry fusion partner lacks cysteine, therefore, it will not engage in disulfide bonding with the six cysteine residues in two peptide chains of human insulin. Fourth, the bacterial expression host may take the advantage of the chaperone activity of the carrier protein (\u03b1B-Cry) which may provide an intrinsic tolerance to overcome harmful environmental stresses \\[[@pone.0206169.ref056]\\]. Moreover, according to the results of previous studies, \u03b1-Cry promotes the dissociation of insulin oligomers to the lower associated species (dimers and monomers) \\[[@pone.0206169.ref042]\\] and displays a superior stabilizing effect on insulin. A chemical cleavage site (CNBr cleavage site) at the end of \u03b1-Cry carrier protein provides a convenient condition for recovering the target peptides from the digested fusion proteins (\u03b1B-AC and \u03b1B-BC). We also took advantage of the chaperone activity of this carrier protein to increase the yield of *in vitro* chain combination and folding of human recombinant insulin. A summary of our expression, purification and folding strategy is shown as **[Fig 13](#pone.0206169.g013){ref-type=\"fig\"}**.\n\n![Schematic diagram of the experimental procedure.\\\nThe constructs contain ATG codon for coding methionine (**M**) as the specific CNBr cleavage site. The fusion genes or proteins \u03b1B-Cry/A-chain and \u03b1B-Cry/B-chain indicated as \u03b1B-AC and \u03b1B-BC, respectively.](pone.0206169.g013){#pone.0206169.g013}\n\nBased upon the above information, \u03b1B-Cry fully meets the requirements as a suitable carrier protein and chaperone molecule for expression of the insulin chains. Moreover, this chaperone has been already suggested as a suitable excipient candidate in the formulation of therapeutic insulin \\[[@pone.0206169.ref042]\\].\n\nThe SDS-PAGE results suggested that the fusion proteins (\u03b1B-AC and \u03b1B-BC) were fully expressed as inclusion bodies in *Escherichia coli* (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**). The intrinsic chaperone ability of \u03b1B-Cry through the intermolecular association of hydrophobic domains may facilitate the sequestering of the fusion proteins in the inclusion bodies (**[Fig 2](#pone.0206169.g002){ref-type=\"fig\"}**) \\[[@pone.0206169.ref057], [@pone.0206169.ref058]\\]. Cell lysis followed by washing of the inclusion bodies led to the fusion proteins with a purity more than 48%. The fusion proteins were then purified using a single-step purification approach which produced these proteins in high yield and reasonable purity. The CNBr release of insulin peptide chains from the fusion proteins was also efficient (68.8% and 75.2% for \u03b1B-BC and \u03b1B-AC, respectively). The fusion proteins indicated high solubility in the CNBr digesting solution and this chemical cleaving agent was also simply removed from the digested mixture by dialysis. Using a single-step gel filtration, we purified the A-and B-chain of human insulin (**Figs [5](#pone.0206169.g005){ref-type=\"fig\"}** and **[6](#pone.0206169.g006){ref-type=\"fig\"}**).\n\nSubsequently, combination of these chains to form active insulin was our main goal. Six cysteine residues are present in the A- and B-chain, giving rise to three disulfide bonds \\[[@pone.0206169.ref059]\\]. Therefore, statistically, there are 15 possibilities of forming disulfide bridges within one human insulin molecule and only one giving rise to native functional human insulin with the following disulfide bridges: A6-A11; A7-B7 and A20-B19 \\[[@pone.0206169.ref060]\\]. Fortunately, previous studies achieved correct folding of insulin molecules in the presence of a mixture of S-sulfonated A- and B-chain under a reducing environment \\[[@pone.0206169.ref031]\\]. However, Dixon and Wardlaw indicated an approximate yield of only 2% for the chain combination of the oxidized insulin peptide chains \\[[@pone.0206169.ref020], [@pone.0206169.ref061]\\]. Later, another study suggested a combination of the B-chain with a large excess of the A-chain led to greater yield of natively folded insulin \\[[@pone.0206169.ref020]\\]. During the air oxidation step, where the reduced chains are combined to form folded insulin, only a fraction of the total chains is available. The loss of insulin chains in this step is mainly due to their partial insolubility and aggregation, and incorrect disulfide bond arrangement during the initial steps of the protein folding \\[[@pone.0206169.ref062], [@pone.0206169.ref063]\\]. Moreover, it has been previously indicated that in solution, B-chain has a strong tendency to form soluble aggregates of approximately 40 kDa and its aggregation is also pH dependent \\[[@pone.0206169.ref037]\\]. All these facts could explain the very low yield of insulin chain combination in the absence of an appropriate chaperone (**[Fig 7A](#pone.0206169.g007){ref-type=\"fig\"}**). The insulin chain combination has been typically applied at pH values between 9.5 and 11 and in the present study it was undertaken in 0.1 M glycine at pH 10.6. The basic pH results in deprotonation of sulfhydryl groups on cysteines (thiolate moieties) and enhances the yield by limiting aggregation of the reduced B-chains. However, due to hydrophobic interactions, partial aggregation of the B-chains occurs at basic pH. All of these challenges reduce the efficiency of insulin folding during the chain combination step \\[[@pone.0206169.ref020], [@pone.0206169.ref039]\\]. To overcome these problems, Tang et al. used protein disulfide isomerase (PDI) and obtained significantly higher yield compared with the cross-linking of the chains in the absence of this enzyme \\[[@pone.0206169.ref062]\\]. PDI has both chaperone and catalytic activities and its chaperone domain prevents aggregation of the target protein, while its catalytic domain catalyzes the correct disulfide formation. However, Winter et al. showed that upon suppressing the catalytic activity of PDI, the yield of insulin combination was maintained at an appreciable level while inhibition of the chaperone function drastically reduced PDI-assisted folding. These findings suggest that the chaperone function of PDI plays the main role during the chain combination and folding of insulin molecules \\[[@pone.0206169.ref064]\\]. Hence, the crucial aspect for high efficiency of the chain combination is to select an appropriate chaperone that prevents aggregation of the insulin chains. There are many reports on \u03b1B-Cry and lens \u03b1-Cry preventing insulin aggregation under reducing conditions \\[[@pone.0206169.ref048], [@pone.0206169.ref065]\\].\n\n\u03b1B-Cry is an ATP-independent molecular chaperone that has little ability to fold target proteins but it stabilizes unfolding target proteins to prevent their aggregation. Therefore, we used \u03b1B-Cry as both fusion partner to prepare a suitable construct for the individual expression of human insulin A- and B-chain and later a chaperone to ensure correct folding during the chain combination step (**[Fig 13](#pone.0206169.g013){ref-type=\"fig\"}**). Fortunately, the chaperone activity of \u03b1B-Cry is still efficient at high pH values \\[[@pone.0206169.ref066]\\]. Indeed, \u03b1B-Cry significantly increased the yield of insulin chain combination and folding at basic pH (**[Fig 7B](#pone.0206169.g007){ref-type=\"fig\"}**). The chaperone also stabilized the native insulin molecules against chemical stress during the subsequent purification step (**Figs [10](#pone.0206169.g010){ref-type=\"fig\"} and [11](#pone.0206169.g011){ref-type=\"fig\"}**) \\[[@pone.0206169.ref042]\\]. As indicated in **[Fig 7B](#pone.0206169.g007){ref-type=\"fig\"}**, after 48 h, the yield of natively folded insulin molecules increased from 10.2% in the absence of \u03b1B-Cry to 26.7% in its presence. The residual fractions belonged to the isolated A- and B-chain and incorrectly folding intermediates. A highly pure sample of natively folded human recombinant insulin was achieved with a suitable phenyl sepharose column and elution at 250 mM ammonium sulfate. After purification, we compared our insulin with the standard insulin sample according to their structure and biological activity. The results of RP-HPLC analysis, structural assessment by fluorescence, CD and NIR, aggregation and oligomerization studies, as well as *in vivo* activity assessments, all showed that our recombinant insulin exhibited substantial similarity to standard insulin.\n\nOverall, we designed a streamlined insulin chain expression and purification protocol in which all individual steps are straightforward with high yield. Additionally, this study provided a simple and efficient route to express and purify insulin peptide chains that are difficult or expensive to produce by chemical synthesis or by ordinary recombinant methods. Our results also suggest \u03b1B-Cry as a remarkable candidate not only to increase the efficacy of insulin folding during the chain combination step but also as a possible interesting excipient in the formulation of this therapeutic hormone.\n\nSupporting information {#sec031}\n======================\n\n###### RP-HPLC analysis of human \u03b1B-Cry.\n\nThe retention time of \u03b1B-Cry used for the chain combination experiment was determined using C18 reverse phase column. A 20 \u03bcL of \u03b1B-Cry (2 mg/m) in the refolding buffer was subjected to the column and analyzed by Knauer HPLC system. The absorbance signals were recorded at 214 nm using DAD 2.1 UV-Visible detector (Knauer, Germany).\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\nWe would like to sincerely appreciate Professor John A. Carver (College of Physical and Mathematical Sciences, The Australian National University) for, his valuable comments, critical reading and language refinement of this manuscript. We also sincerely appreciate Professor Bernhard Erni (Emeritus Professor of Biochemistry from Department of Chemistry and Biochemistry of Bern University, Bern, Switzerland) for his valuable comments on this manuscript.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "![](jpmedmentpath268705-0126){#sp1 .122}\n\n![](jpmedmentpath268705-0127){#sp2 .123}\n\n![](jpmedmentpath268705-0128){#sp3 .124}\n\n![](jpmedmentpath268705-0129){#sp4 .125}\n\n![](jpmedmentpath268705-0130){#sp5 .126}\n\n![](jpmedmentpath268705-0131){#sp6 .127}\n"}
+{"text": "1. Metformin: Historical re-purposing and pleiotropy {#s0005}\n====================================================\n\nIt is a lesser-known fact that metformin was originally introduced as an anti-influenza drug and that glucose-lowering was only one of its side effects [@b0005]. The many pleiotropic effects of metformin and its widespread utility in medicine today have led scientists to call it the aspirin of the 21st century [@b0010].\n\nIn the current scenario, when there is no specific agent against COVID-19, and when re-purposing of drugs is the primary weapon, we suggest that metformin be used as one of the drugs to combat the virus.\n\n2. Metformin: Mechanism of action on molecular level {#s0010}\n====================================================\n\nMetformin activates AMP-activated protein kinase (AMPK) in hepatocytes by causing its phosphorylation. This is the main mechanism by which metformin brings about favourable effects on glucose and lipid metabolism [@b0015].\n\n2.1. Metformin-AMPK-ACE2-SARS-CoV-2 {#s0015}\n-----------------------------------\n\nThe juggernaut virus, SARS-CoV-2, that has led to the deaths of over 1.7 lakh people across the world uses angiotensin-converting enzyme 2 (ACE2) as its receptor. It enters the human body through interaction between its spike proteins (S1) and the N-terminal region of ACE2 [@b0020], [@b0025]. The receptor binding domain (RBD) of the virus binds with the protease domain (PD) of the ACE2 receptor and forms an RBD-PD complex [@b0020].\n\nAcute Respiratory Distress Syndrome (ARDS) is one of the commonest complications developing in patients with COVID-19 [@b0030]. There have been animal studies that have implicated ACE2 in the acute lung injury (ALI) caused due to SARS-CoV [@b0020]. It has been hypothesized that ACE2 causes ALI by bringing about autophagy through the AMPK/mTOR pathway [@b0035]. AMPK has been shown to increase the expression of ACE2 as well as to increase its stability by phosphorylating ACE2 Ser^680^ in human umbilical vein endothelial cells (HUVECs) and human embryonic kidney 293 (HEK293T) cells [@b0040].\n\nSince metformin works through AMPK activation, which leads to phosphorylation of ACE2 [@b0040], we can consider that theoretically this addition of a phosphate group (PO~4~ ^\u22123^) would bring about conformational and functional changes in the ACE2 receptor [@b0045]. This could lead to decreased binding with SARS-CoV-2 RBD due to steric hindrance by the addition of a large sized PO~4~ ^\u22123^ molecule.\n\nNonetheless, once the virus is inside, there is a downregulation of ACE2 receptors. This in turn leads to an imbalance in the renin-angiotensin-aldosterone system (RAS) promoting the deleterious effects of its pro-inflammatory and pro-fibrotic arm, further giving rise to the lethal cardio-pulmonary complications [@b0050]. By upregulating ACE2, the imbalance in RAS could be averted. Hence, metformin would not only prevent the entry of SARS-CoV-2 as described above, but also prevent the detrimental sequelae by causing activation of ACE2 through AMPK-signalling.\n\n2.2. Metformin-mTOR-Coronavirus {#s0020}\n-------------------------------\n\nThe mammalian target of rapamycin (mTOR) signalling plays an important role in the pathogenesis of influenza, besides modulating antibody response for cross-protective immunity against infective influenza viruses. Metformin activates AMPK via liver kinase B1 (LKB1), inhibiting the mTOR pathway. It also indirectly attenuates AKT activation through phosphorylation of insulin receptor substrate 1 (IRS-1) resulting in inhibition of the mTOR signalling cascade [@b0055]. Other biguanide molecules, buformin and phenformin have been associated with better survival outcomes in animal models of influenza [@b0060], [@b0065]. Further, the PI3K/AKT/mTOR pathway plays major roles in MERS\u2011CoV infection [@b0070]. Since metformin inhibits the same pathway, it would be interesting to decipher its role against SARS-CoV-2.\n\n2.3. Protein-protein interaction map and network-based drug repurposing {#s0025}\n-----------------------------------------------------------------------\n\nA study was attempted to narrow the existing molecular-level knowledge gap of SARS-CoV-2 by mapping the interactions between SARS-CoV-2 and human proteins [@b0075]. With the help of affinity purification mass spectrometry (AP-MS), 332 protein--protein interactions (PPIs) could be identified. Further, 66 druggable human proteins/factors targeted by 69 medicines which were either FDA-approved or in clinical trials or pre-clinical molecules were recognized. To our interest, it was found that human proteins regulated by the mTORC1 signalling pathway, specifically LARP1 and FKBP7, interact with important viral proteins, N and Orf8 [@b0075]. Since metformin inhibits mTOR signalling, it could act as an indirect modulator of the protein--protein complex, thus preventing the viral replication and pathogenesis.\n\n2.4. Viral replication: Lessons learnt from Zika virus {#s0030}\n------------------------------------------------------\n\nZika virus (ZIKV), a single-stranded RNA virus, is a mosquito transmitted flavivirus. A study using HUVECs and human retinal vascular endothelial cells (HRvECs) showed that AMPK restricts the replication of ZIKV in the endothelial cells [@b0080]. Activation of AMPK with the help of two well-known AMPK activators, metformin and 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) led to attenuation of ZIKV replication. Activated AMPK further potentiated the expression of certain genes with known antiviral properties such as IFNs, OAS2, ISG15, and MX1 while inhibiting inflammatory mediators like TNF-\u03b1 and CCL5. It would be useful to explore whether the same is observed for SARS-CoV-2. A recent study has reported that inhibition of glycolysis by non-toxic concentration of 2-DG completely attenuated SARS-CoV-2 replication in Caco-2 cells [@b0085]. All of these indicate towards a possible frontline role of metformin against COVID-19.\n\n2.5. Insulin resistance and SARS-CoV-2 {#s0035}\n--------------------------------------\n\nA few case reports from China and Italy, along with a Chinese meta-analysis, have shown diabetes to be an important risk factor for severe disease requiring ventilation [@b0090]. Further, a study had shown a direct metabolic link between SARS-CoV and diabetes, postulating that the virus enters the pancreatic islets which express ACE2, leading to acute \u03b2-cell damage and transient Type 2 diabetes mellitus (T2DM) [@b0095]. Evidence from an animal study points toward increased ACE2 activity in pancreas of persons with diabetes besides its elevated expression in other tissues such as lung, liver and heart [@b0100]. Hence, optimal control of T2DM, for both chronic and transient cases, might help in the treatment of COVID-19. Although recent discussions point out that oral hypoglycaemic agents such as Sodium-Glucose-Transporter-2 inhibitors (SGLT-2i), Glucagon-Like-Peptide-1 Receptor Agonists (GLP-1RAs), Pioglitazone and even Insulin might actually be harmful for COVID-19 individuals with diabetes [@b0105], [@b0110], limited evidence is available on metformin for the same. Considering its pleiotropic effects and a possible role in combating hepatitis C virus (HCV), hepatitis B virus (HBV) and human immunodeficiency virus (HIV) through increasing insulin sensitivity [@b0115], metformin can be a real game-changer for treating this pandemic.\n\nFunding {#s0040}\n=======\n\nNo funding has been received for the preparation of this manuscript.\n\nAuthorship {#s0045}\n==========\n\nAll authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship of this article, take responsibility for the integrity of the work as a whole and that it will not be published elsewhere in the same form, in English or in any other language, including electronically, and have given their approval for this version to be published.\n\nAuthorship contributions {#s0050}\n========================\n\nSS, AR and BS co-wrote and revised the manuscript.\n\nDeclaration of Competing Interest\n=================================\n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nPancreatic cancer (PC) is one of the most aggressive and lethal malignancies, causing the deaths of an estimated 330,400 men and women worldwide in 2012 \\[[@R1]\\]. Total deaths due to PC are projected to increase dramatically, making it second leading cause of cancer-related deaths in the United States by 2030 \\[[@R2]\\]. Gemcitabine, the current standard first-line treatment, offers marginal symptom control and prolongation of life. Clinical trials aiming to improve the efficacy of gemcitabine have provided little improvement in survival outcomes \\[[@R3]\\]. New therapeutic strategies, including therapeutic antibodies or/and small molecule inhibitors, have been successful for a number of malignancies, but results obtained on PC treatments have so far been extremely frustrating \\[[@R4]\\]. A number of molecular mechanisms responsible for transformation and progression of PC have been identified, providing a set of potential pharmacological targets \\[[@R5]\\]. Among these is loss of adhesion between tumor cells caused by downregulation of CDH1 (also called E-cadherin) in response to genetic or epigenetic changes \\[[@R6]--[@R8]\\].\n\nHistone acetylation is a dynamic epigenetic mechanism regulated by the histone acetyltransferases (HAT) and histone deacetylases (HDACs). HDAC3 (histone deacetylases 3), a member of class I HDACs, is overexpressed in the majority of carcinomas \\[[@R9], [@R10]\\], and is one of the most frequently upregulated genes in cancer \\[[@R11]\\]. Our previous study shows increased HDAC3 expression in PC \\[[@R12]\\]. HDAC3 could function as an oncogenic protein, promoting PC cell proliferation, migration, and invasion, as well as increasing drug resistance \\[[@R12]\\]. HDAC3 inversely correlates with CDH1 expression in ovarian carcinoma, and HDAC3 siRNA knock down in ovarian carcinoma cells reduced cell migration and increased CDH1 expression \\[[@R13]\\]. HDAC3 represses CDH1 through interactions with epithelial-mesenchymal transition (EMT) regulators including Snail and Twist1 \\[[@R14]\\].\n\nThis study uses high-throughput tissue microarray (TMA) and immunohistochemistry to investigate the expression and subcellular localization of CDH1 and HDAC3 in PC tissues. We analyze their association with clinicopathological factors, and address their possible value as prognostic indicators.\n\nRESULTS {#s2}\n=======\n\nExpression of CDH1 and HDAC3 in PC tissues and adjacent normal tissues {#s2_1}\n----------------------------------------------------------------------\n\nImmunohistochemistry results are summarized in Tables [1](#T1){ref-type=\"table\"} and [2](#T2){ref-type=\"table\"}. Strong membrane localization of CDH1 was observed in 85.7% (72/84) of normal tissues adjacent to PC (Figure [1A](#F1){ref-type=\"fig\"}). In contrast, cell membrane expression of CDH1 was greatly reduced in PC tissues (Figure [1B](#F1){ref-type=\"fig\"}), with high expression in 63.1% (53/84) of cases. Interestingly, higher cytoplasmic CDH1 expression was observed in PC samples (Figure [1C](#F1){ref-type=\"fig\"}); 33.3% of tumor samples (28/84) but only 11.9% (10/84) of adjacent tissue samples displayed high cytoplasmic CDH1.\n\n###### Comparisons with CDH1 expression between PC and paired adjacent normal tissues (*n* = 84)\n\n  Tissue sample     No.of patients   Membrane CDH1 (*n*, %)   *P*-value   Cytoplasmic CDH1 (*n*, %)   *P*-value               \n  ----------------- ---------------- ------------------------ ----------- --------------------------- ----------- ----------- ---------\n  Tumor             84               31 (36.9)                53 (63.1)   0.001\\*                     56 (66.7)   28 (33.3)   0.001\\*\n  Adjacent normal   84               12 (14.3)                72 (85.7)   74 (88.1)                   10 (11.9)               \n\n###### Comparisons with HDAC3 expression between PC and paired adjacent normal tissues (*n* = 84)\n\n  Tissue sample     No.of patients   Nuclear HDAC3 (*n*, %)   *P*-value   Cytoplasmic HDAC3 (*n*, %)   *P*-value               \n  ----------------- ---------------- ------------------------ ----------- ---------------------------- ----------- ----------- -------\n  Tumor             84               38 (45.2)                46 (54.8)   \\< 0.001\\*                   38 (45.2)   46 (54.8)   0.641\n  Adjacent normal   84               68 (81.0)                16 (19.0)   35 (41.7)                    49 (58.3)               \n\n![Immunohistochemical expression levels and localization of CDH1 and HDAC3 in PC tissues\\\nStrong membrane-associated CDH1 was observed in adjacent normal tissues (**A**). Low membrane CDH1 (**B**) and high cytoplasmic CDH1 (**C**) was found in tumor cells. Higher level of nuclear HDAC3 was observed in PC tissues (**D**), than in adjacent normal tissues (**E**). There was no difference in cytoplasmic HDAC3 expression between PC tissues and noncancerous samples (E, normal tissue; (**F**), tumor tissue). Scale bar, 50 \u03bcm.](oncotarget-07-16505-g001){#F1}\n\nHDAC3 was distributed in the cytoplasm and nucleus. As shown in Figure [1D](#F1){ref-type=\"fig\"}, nuclear HDAC3 was highly expressed in 54.8% (46/84) of PC tissues. In contrast, HDAC3 was only seen in the nucleus of 19% (16/84) of noncancerous tissues (Figure [1E](#F1){ref-type=\"fig\"}). There was no difference in cytoplasmic HDAC3 expression between PC tissues and noncancerous samples (54.8%, 46/84 vs. 58.3%, 49/84; Figure [1E, 1F](#F1){ref-type=\"fig\"}).\n\nCorrelations of CDH1 and HDAC3 expression in PC tissues {#s2_2}\n-------------------------------------------------------\n\nAn inverse correlation was identified between low membrane expression of CDH1 and high nuclear HDAC3 expression (Spearman correlation coefficient *r* = \u22120.348, *P* = 0.001, [Supplementary Table S1](#SD1){ref-type=\"supplementary-material\"}). High cytoplasmic CDH1 expression positively correlated with high nuclear HDAC3 expression (Spearman correlation coefficient *r* = 0.440, *P* \\< 0.001, Table [3](#T3){ref-type=\"table\"}). No correlations were found between cytoplasmic HDAC3 expression and CDH1 expression location (*P* \\> 0.05, [Supplementary Tables S2, S3](#SD1){ref-type=\"supplementary-material\"}).\n\n###### Association between nuclear HDAC3 and cytoplasmic CDH1 expression\n\n  Tumor tissue sample     Nuclear HDAC3   Correlation coefficient   *P*-value   \n  ----------------------- --------------- ------------------------- ----------- ------------\n  Cytoplasmic CDH1 Low    34              22                        0.440       \\< 0.001\\*\n  Cytoplasmic CDH1 High   4               24                                    \n\nRelationship of clinicopathological features with CDH1 and HDAC3 expression in PC patients {#s2_3}\n------------------------------------------------------------------------------------------\n\nThe relationships of CDH1 and HDAC3 expression levels with clinicopathological features of PC were evaluated by immunohistochemistry. As summarized in Table [4](#T4){ref-type=\"table\"}, CDH1 cell membrane expression correlated with lymph node metastasis (*P* = 0.026) and clinical stage (*P* = 0.020). High cytoplasmic CDH1 strongly correlated with lymph node metastasis (N classification, *P* \\< 0.001) and advanced clinical stage (*P* \\< 0.001). Neither cytoplasmic nor membrane CDH1 were associated with patients' gender, age, tumor location, tumor size, tumor differentiation, invasion depth, distant metastasis, abdominal pain, jaundice or nervous invasion (*P* \\> 0.05).\n\n###### Correlation between the clinicopathologic characteristics and CDH1 expression (*n* = 84)\n\n  Clinicopathological parameters   No.of patients   Membrane CDH1 (*n*, %)   Cytoplasmic CDH1 (*n*, %)                                                                                                          \n  -------------------------------- ---------------- ------------------------ --------------------------- ------------------------------------------------------------------------------ ----------- ----------- ---------------------------------------------------------------------------------\n  Cases                            84               31 (36.9)                53 (63.1)                                                                                                  56 (66.7)   28 (33.3)   \n  **Age (years)**                                                                                                                                                                                               \n  \u2003\u2264 60                            39               17 (43.6)                22 (56.4)                   0.237[^a^](#tfn_001){ref-type=\"table-fn\"}                                      26 (66.7)   13 (33.3)   1.000[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003\\> 60                           45               14 (31.1)                31 (68.9)                   30 (66.7)                                                                      15 (33.3)               \n  **Gender**                                                                                                                                                                                                    \n  \u2003Male                            51               21 (41.2)                30 (58.8)                   0.313[^a^](#tfn_001){ref-type=\"table-fn\"}                                      34 (66.7)   17 (33.3)   1.000[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Female                          33               10 (30.3)                23 (69.7)                   22 (33.3)                                                                      11 (33.3)               \n  **Tumor location**                                                                                                                                                                                            \n  \u2003Head, neck                      56               24 (42.9)                32 (57.1)                   0.110[^a^](#tfn_001){ref-type=\"table-fn\"}                                      35 (62.5)   21 (37.5)   0.252[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Body, tail                      28               7 (25.0)                 21 (75.0)                   21 (75.0)                                                                      7 (25.0)                \n  **Tumor size (cm)**                                                                                                                                                                                           \n  \u2003\u2264 3                             25               9 (36.0)                 16 (64.0)                   0.911[^a^](#tfn_001){ref-type=\"table-fn\"}                                      18 (72.0)   7 (28.0)    0.500[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003\\> 3                            59               22 (37.3)                37 (62.7)                   38 (64.4)                                                                      21 (35.6)               \n  **Tumor differentiation**                                                                                                                                                                                     \n  \u2003Well, moderate                  57               21 (36.8)                36 (63.2)                   0.986[^a^](#tfn_001){ref-type=\"table-fn\"}                                      40 (70.2)   17 (29.8)   0.322[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Poor                            27               10 (37.0)                17 (63.0)                   16 (59.3)                                                                      11 (40.7)               \n  **Invasion depth**                                                                                                                                                                                            \n  \u2003T1 + T2                         71               27(38.0)                 44(62.0)                    0.618[^a^](#tfn_001){ref-type=\"table-fn\"}                                      49 (69.0)   22 (31.0)   0.286[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003T3 + T4                         13               4(30.8)                  9(69.2)                     7 (53.8)                                                                       6 (46.2)                \n  **Lymph nodes metastasis**                                                                                                                                                                                    \n  \u2003N0 (negative)                   51               14 (27.5)                37 (72.5)                   0.026[^a^](#tfn_001){ref-type=\"table-fn\"}[\\*](#tfn_003){ref-type=\"table-fn\"}   43 (84.3)   8 (15.7)    \\< 0.001[^a^](#tfn_001){ref-type=\"table-fn\"}[\\*](#tfn_003){ref-type=\"table-fn\"}\n  \u2003N1 (positive)                   33               17 (51.5)                16 (48.5)                   13 (39.4)                                                                      20 (60.6)               \n  **Distant metastasis**                                                                                                                                                                                        \n  \u2003Absent                          82               29 (35.4)                53 (64.6)                   0.133[^b^](#tfn_002){ref-type=\"table-fn\"}                                      56 (68.3)   26 (31.7)   0.108[^b^](#tfn_002){ref-type=\"table-fn\"}\n  \u2003Present                         2                2 (100)                  0 (0)                       0 (0)                                                                          2 (100)                 \n  **Clinical stage**                                                                                                                                                                                            \n  \u2003Early stages (\u2264 IIa)            49               13(26.5)                 36(73.5)                    0.020[^a^](#tfn_001){ref-type=\"table-fn\"}[\\*](#tfn_003){ref-type=\"table-fn\"}   43(87.8)    6(12.2)     \\< 0.001[^a^](#tfn_001){ref-type=\"table-fn\"}[\\*](#tfn_003){ref-type=\"table-fn\"}\n  \u2003Advanced stages (\\> IIa)        35               18(51.4)                 17(48.6)                    13(37.1)                                                                       22(62.9)                \n  **Abdominal pain**                                                                                                                                                                                            \n  \u2003Absent                          38               13 (34.2)                25 (65.8)                   0.642[^a^](#tfn_001){ref-type=\"table-fn\"}                                      22 (57.9)   16 (42.1)   0.121[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Present                         46               18 (39.1)                28 (60.9)                   34 (73.9)                                                                      12 (26.1)               \n  **Jaundice**                                                                                                                                                                                                  \n  \u2003Absent                          69               23 (33.3)                46 (66.7)                   0.146[^a^](#tfn_001){ref-type=\"table-fn\"}                                      47 (68.1)   22 (31.9)   0.546[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Present                         15               8 (53.3)                 7 (46.7)                    9 (60.0)                                                                       6 (40.0)                \n  **Nervous invasion**                                                                                                                                                                                          \n  \u2003Negative                        51               20 (39.2)                31 (60.8)                   0.585[^a^](#tfn_001){ref-type=\"table-fn\"}                                      33 (64.7)   18 (35.3)   0.636[^a^](#tfn_001){ref-type=\"table-fn\"}\n  \u2003Positive                        33               11 (33.3)                22 (66.7)                   23 (69.7)                                                                      10 (30.3)               \n\nChi-square test.\n\nFisher\\'s exact test.\n\n*P* \\< 0.05 indicates a significant association among the variables.\n\nAs summarized in Table [5](#T5){ref-type=\"table\"}, no correlations were observed between cytoplasmic levels of HDAC3 and patients' clinicopathologic features. Nuclear HDAC3 staining correlated with lymph node metastasis (*P* \\< 0.001) and clinical stage (*P* \\< 0.001), but did not correlate with patient\\'s gender, age, tumor location, tumor size, tumor differentiation, invasion depth, distant metastasis, abdominal pain, jaundice, or nervous invasion (*P* \\> 0.05).\n\n###### Correlation between the clinicopathologic characteristics and HDAC3 expression (*n* = 84)\n\n  Clinicopathological parameters   No.of patients   Nuclear HDAC3 (*n*, %)   Cytoplasmic HDAC3 (*n*, %)                                                                                                               \n  -------------------------------- ---------------- ------------------------ ---------------------------- ----------------------------------------------------------------------------------- ----------- ----------- -------------------------------------------\n  Cases                            84               38 (45.2)                46 (54.8)                                                                                                        38 (45.2)   46 (54.8)   \n  **Age (years)**                                                                                                                                                                                                     \n  \u2003\u2264 60                            39               19 (48.7)                20 (51.3)                    0.551[^a^](#tfn_004){ref-type=\"table-fn\"}                                           18 (46.2)   21 (53.8)   0.875[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003\\> 60                           45               19 (42.2)                26 (57.8)                                                                                                        20 (44.4)   25 (55.6)   \n  **Gender**                                                                                                                                                                                                          \n  \u2003Male                            51               20 (39.2)                31 (60.8)                    0.168[^a^](#tfn_004){ref-type=\"table-fn\"}                                           25 (49.0)   26 (51.0)   0.387[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Female                          33               18 (54.5)                15 (45.5)                                                                                                        13 (39.4)   20 (60.6)   \n  **Tumor location**                                                                                                                                                                                                  \n  \u2003Head, neck                      56               23 (41.1)                33 (58.9)                    0.278[^a^](#tfn_004){ref-type=\"table-fn\"}                                           26 (46.4)   30 (53.6)   0.757[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Body, tail                      28               15 (53.6)                13 (46.4)                                                                                                        12 (42.9)   16 (57.1)   \n  **Tumor size (cm)**                                                                                                                                                                                                 \n  \u2003\u2264 3                             25               10 (40.0)                15 (60.0)                    0.530[^a^](#tfn_004){ref-type=\"table-fn\"}                                           12 (48.0)   13 (52.0)   0.741[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003\\> 3                            59               28 (47.5)                31 (52.5)                                                                                                        26 (44.1)   33 (55.9)   \n  **Tumor differentiation**                                                                                                                                                                                           \n  \u2003Well, moderate                  57               27 (47.4)                30 (52.6)                    0.569[^a^](#tfn_004){ref-type=\"table-fn\"}                                           26 (45.6)   31 (54.4)   0.920[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Poor                            27               11 (40.7)                16 (59.3)                                                                                                        12 (44.4)   15 (55.6)   \n  **Invasion depth**                                                                                                                                                                                                  \n  \u2003T1 + T2                         71               35 (49.3)                36 (50.7)                    0.081[^a^](#tfn_004){ref-type=\"table-fn\"}                                           29 (40.8)   42 (59.2)   0.059[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003T3 + T4                         13               3 (23.1)                 10 (76.9)                                                                                                        9 (69.2)    4 (30.8)    \n  **Lymph nodes metastasis**                                                                                                                                                                                          \n  \u2003N0 (negative)                   51               32 (62.7)                19 (37.3)                    \\< 0.001[^a^](#tfn_004){ref-type=\"table-fn\"}[^\\*^](#tfn_006){ref-type=\"table-fn\"}   23 (45.1)   28 (54.9)   0.974[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003N1 (positive)                   33               6 (18.2)                 27 (81.8)                                                                                                        15 (45.5)   18 (54.5)   \n  **Distant metastasis**                                                                                                                                                                                              \n  \u2003Absent                          82               38 (46.3)                44 (53.7)                    0.499[^b^](#tfn_005){ref-type=\"table-fn\"}                                           36 (43.9)   46 (56.1)   0.202[^b^](#tfn_005){ref-type=\"table-fn\"}\n  \u2003Present                         2                0 (0)                    2 (100)                                                                                                          2 (100)     0 (0)       \n  **Clinical stage**                                                                                                                                                                                                  \n  \u2003Early stages (\u2264 IIa)            49               31 (63.3)                18 (36.7)                    \\< 0.001[^a^](#tfn_004){ref-type=\"table-fn\"}[\\*](#tfn_006){ref-type=\"table-fn\"}     22 (44.9)   27 (55.1)   0.941[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Advanced stages (\\> IIa)        35               7 (20.0)                 28 (80.0)                                                                                                        16 (45.7)   19 (54.3)   \n  **Abdominal pain**                                                                                                                                                                                                  \n  \u2003Absent                          38               16 (42.1)                22 (57.9)                    0.600[^a^](#tfn_004){ref-type=\"table-fn\"}                                           19 (50.0)   19 (50.0)   0.425[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Present                         46               22 (47.8)                24 (52.2)                                                                                                        19 (41.3)   27 (58.7)   \n  **Jaundice**                                                                                                                                                                                                        \n  \u2003Absent                          69               33 (47.8)                36 (52.2)                    0.307[^a^](#tfn_004){ref-type=\"table-fn\"}                                           31 (44.9)   38 (55.1)   0.902[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Present                         15               5 (33.3)                 10 (66.7)                                                                                                        7 (46.7)    8 (53.3)    \n  **Nervous invasion**                                                                                                                                                                                                \n  \u2003Negative                        51               19 (37.3)                32 (62.7)                    0.068[^a^](#tfn_004){ref-type=\"table-fn\"}                                           26 (51.0)   25 (49.0)   0.189[^a^](#tfn_004){ref-type=\"table-fn\"}\n  \u2003Positive                        33               19 (57.6)                14 (42.4)                                                                                                        12 (36.4)   21 (63.6)   \n\nChi-square test.\n\nFisher\\'s exact test.\n\n*P* \\< 0.05 indicates a significant association among the variables.\n\nAssociations between CDH1 and HDAC3 expression and survival {#s2_4}\n-----------------------------------------------------------\n\nKaplan-Meier analysis and log-rank test were used to investigate the prognostic value of CDH1 and HDAC3 expression and classic clinicopathologic characteristics on patient survival. In univariate analysis, both membrane and cytoplasmic CDH1 expression, as well as nuclear HDAC3, were closely associated with overall survival (OS) of PC patients (*P* = 0.012, *P* \\< 0.001, and *P* \\< 0.001, respectively; Table [6](#T6){ref-type=\"table\"}), with Spearman correlation coefficients of 0.240, \u22120.435, and \u22120.530 ([Supplementary Table S4](#SD1){ref-type=\"supplementary-material\"}), respectively. The log-rank test results showed that the aberrant expression levels of these proteins correlated strongly with poorer survival in PC patients (*P* \\< 0.001; Figure [2](#F2){ref-type=\"fig\"}). As shown in Table [7](#T7){ref-type=\"table\"}, the cumulative 1-year survival rate was 58% in the high membrane CDH1 group (95% confidence interval \\[CI\\], 0.443--0.717), whereas it was only 32% (95% CI, 0.163--0.477) in the low expression group (Figure [2A](#F2){ref-type=\"fig\"}). The cumulative 1-year survival rate was 63% (95% CI, 0.512--0.748) in the low cytoplasmic CDH1 group, whereas it was only 21% (95% CI, 0.053--0.367) in the high-expression group (Figure [2B](#F2){ref-type=\"fig\"}). The 1-year survival rate was 79% in the low nuclear HDAC3 group (95% CI, 0.653--0.927), whereas it was only 24% (95% CI, 0.122--0.358) in the high staining group (Figure [2C](#F2){ref-type=\"fig\"}). There was no difference in survival time associated with cytoplasmic HDAC3 expression (low vs. high, 47% (95% CI, 0.313--0.627) vs. 50% (95% CI, 0.363--0.637); Figure [2D](#F2){ref-type=\"fig\"}).\n\n###### Summary of univariate and multivariate Cox regression analysis of overall survival duration in all PCs\n\n  Clinicopathological parameters   Univariate analysis   Multivariate analysis                                       \n  -------------------------------- --------------------- ----------------------- ------------ ------- -------------- ---------\n  **Membrane CDH1**                                                                                                  \n  \u2003Low                             1                                                                                 \n  \u2003High                            0.500                 0.290--0.861            0.012\\*                             \n  **Cytoplasmic CDH1**                                                                                               \n  \u2003Low                             1                                                          1                      \n  \u2003High                            2.996                 1.725--5.204            \\< 0.001\\*   2.204   1.210--4.012   0.010\\*\n  **Nuclear HDAC3**                                                                                                  \n  \u2003Low                             1                                                          1                      \n  \u2003High                            4.020                 2.182--7.405            \\< 0.001\\*   3.033   1.572--5.852   0.001\\*\n  **Cytoplasmic HDAC3**                                                                                              \n  \u2003Low                             1                                                                                 \n  \u2003High                            0.716                 0.418--1.227            0.224                               \n  **Age (years)**                                                                                                    \n  \u2003\u2264 60                            1                                                                                 \n  \u2003\\> 60                           0.956                 0.558--1.639            0.870                               \n  **Gender**                                                                                                         \n  \u2003Male                            1                                                                                 \n  \u2003Female                          0.531                 0.295--0.957            0.035\\*                             \n  **Tumor location**                                                                                                 \n  \u2003Head, neck                      1                                                                                 \n  \u2003Body, tail                      1.189                 0.678--2.085            0.546                               \n  **Tumor size(cm)**                                                                                                 \n  \u2003\u2264 3                             1                                                                                 \n  \u2003\\> 3                            0.797                 0.451--1.409            0.436                               \n  **Tumor differentiation**                                                                                          \n  \u2003Well, moderate                  1                                                          1                      \n  \u2003Poor                            2.077                 1.192--3.620            0.010\\*      2.119   1.210--3.711   0.009\\*\n  **Invasion depth**                                                                                                 \n  \u2003T1 + T2                         1                                                                                 \n  \u2003T3 + T4                         0.983                 0.463--2.088            0.965                               \n  **Lymph nodes metastasis**                                                                                         \n  \u2003N0(negative)                    1                                                                                 \n  \u2003N1(positive)                    2.060                 1.196--3.546            0.009\\*                             \n  **Distant metastasis**                                                                                             \n  \u2003Absent                          1                                                                                 \n  \u2003Present                         2.372                 0.574--9.798            0.233                               \n  **Clinical stage**                                                                                                 \n  \u2003Early stages (\u2264 IIa)            1                                                                                 \n  \u2003Advanced stages (\\> IIa)        2.230                 1.294--3.845            0.004\\*                             \n  **Abdominal pain**                                                                                                 \n  \u2003Absent                          1                                                                                 \n  \u2003Present                         0.913                 0.531--1.569            0.742                               \n  **Jaundice**                                                                                                       \n  \u2003Absent                          1                                                                                 \n  \u2003Present                         0.976                 0.476--2.000            0.947                               \n  **Nervous invasion**                                                                                               \n  \u2003Negative                        1                                                                                 \n  \u2003Positive                        1.168                 0.678--2.012            0.576                               \n\nHR hazard ratio, 95% CI 95% confidence interval.\n\n![Cumulative kaplan-meier overall survival curves of 84 PC patients segmented by CDH1 (A), membrane CDH1; (B), cytoplasmic CDH1), HDAC3 (C), nuclear HDAC3; (D), cytoplasmic HDAC3), and high-risk combination group (cytoplasmic CDH1 and nuclear HDAC3 combinations) (E)\\\n*P*-values were calculated by the log-rank test.](oncotarget-07-16505-g002){#F2}\n\n###### Comparisons with cumulative 1-year survival rate between different groups\n\n  Variables                    Cumulative 1-year survival rate   95% CI\n  ---------------------------- --------------------------------- --------------\n  **Membrane CDH1**                                              \n  \u2003Low                         32%                               0.163--0.477\n  \u2003High                        58%                               0.443--0.717\n  **Cytoplasmic CDH1**                                           \n  \u2003Low                         63%                               0.512--0.748\n  \u2003High                        21%                               0.053--0.367\n  **Nuclear HDAC3**                                              \n  \u2003Low                         79%                               0.653--0.927\n  \u2003High                        24%                               0.122--0.358\n  **Cytoplasmic HDAC3**                                          \n  \u2003Low                         47%                               0.313--0.627\n  \u2003High                        50%                               0.363--0.637\n  **High risk combinations**                                     \n  \u2003Both high expression        12%                               0.002--0.238\n  \u2003Other expression group      64%                               0.522--0.758\n\n95% CI, 95% confidence interval.\n\nUnivariate analysis also indicated that gender, tumor differentiation, lymph node metastasis, and clinical stage correlated with patient survival (*P* = 0.035, *P* = 0.010, *P* = 0.009, and *P* = 0.004, respectively). Multivariate analysis shows that cytoplasmic CDH1 expression, nuclear HDAC3 expression, and tumor differentiation were independent prognostic factors for PC patients (Table [6](#T6){ref-type=\"table\"}). Membrane CDH1 expression, gender, lymph node metastasis, and clinical stage were not associated with survival (Table [6](#T6){ref-type=\"table\"}). To further investigate the association of survival time with cytoplasmic CDH1 and nuclear HDAC3 expression, a final concomitant model was constructed. As shown in Figure [2E](#F2){ref-type=\"fig\"}, the log-rank test showed that high co-expression of these two proteins correlated with shorter survival time of PC patients (*P* \\< 0.001). The cumulative proportion of 1-year survival was only 12% (95% CI, 0.002--0.238) in the high co-expression group and 64% (95% CI, 0.522--0.758) in other combination groups (Table [7](#T7){ref-type=\"table\"}). Moreover, Spearman correlation analysis revealed a positive correlation between the high co-expression group and lymph nodes metastasis, clinical stage (*r* = 0.436 and *r* = 0.506, respectively, [Supplementary Table S5](#SD1){ref-type=\"supplementary-material\"}).\n\nDISCUSSION {#s3}\n==========\n\nCellular functions are dictated by protein activity and content. There are numerous strategies to regulate proteins varying from modulating gene expression to post-translational modifications to control of protein localization \\[[@R15]\\]. Numerous studies demonstrate functionally relevant subcellular translocation of specific individual proteins \\[[@R16]\\]. For example, \u03b2-catenin is found at multiple subcellular localizations, including at cell junctions, where it stabilizes cell-cell contacts; in the cytoplasm, where \u03b2-catenin levels are controlled by protein stability regulating processes; and in the nucleus, where \u03b2-catenin is involved in transcriptional regulation and chromatin interactions \\[[@R17], [@R18]\\]. Moreover, \u03b2-catenin nuclear import and accumulation drives tumor formation and correlates with clinical tumor grade \\[[@R19]\\]. Another example is BRCA1, whose prognostic significance varies with its subcellular distribution. Nuclear detection of the protein is associated with a worse prognosis, while cytoplasmic localization predicts lower probability of recurrence due to fewer lymph node metastases \\[[@R20]\\].\n\nDysfunction of the CDH1-mediated cell adhesion system plays an important role in pancreatic tumor progression to invasive, metastatic carcinoma \\[[@R21], [@R22]\\]. Epigenetic modifications contribute to loss of CDH1 expression \\[[@R23], [@R24]\\]. Yao R *et al.* \\[[@R25]\\] found that HDAC3 binds the CDH1 promoter, resulting in reduced local histone acetylation and CDH1 transcriptional repression \\[[@R25]\\]. We previously revealed that HDAC3 is overexpressed in PC tissue, and increased HDAC3 can promote malignant tumor phenotypes \\[[@R12]\\]. Moreover, Hayashi A *et al.* \\[[@R13]\\] found that HDAC3 was inversely correlated with CDH1 expression in ovarian carcinoma. In this study, we determined the expression pattern of CDH1 and HDAC3 proteins in PC tissues, and the clinicopathological and prognostic value of those subcellular localizations.\n\nHigh-throughput TMA was employed to perform our research. First, we found that CDH1 was predominantly found on the cell membrane and in the cytoplasm, while HDAC3 localized to cell nucleus and cytoplasm. Further analysis revealed that the cell membrane CDH1 was greatly reduced in PC tissues compared to noncancerous epithelia, whereas nuclear HDAC3 was abnormally upregulated. Furthermore, there was an inverse association between these two proteins in PC tissues, consistent with recent reports on ovarian carcinoma \\[[@R13]\\].\n\nIt is worth noting that abnormal cytoplasmic CDH1 in PC tissues, and higher cytoplasmic CDH1 expression were associated with more aggressive tumor-associated variables, including lymph node metastasis and advanced clinical stage. Moreover, PC patients with high cytoplasmic CDH1 expression had shorter OS than the low-expression group. In contrast, reduced membrane CDH1 correlated with lymph node metastasis, advanced clinical stage, and shorter survival time. Multivariate analyses demonstrate that cytoplasmic but not membrane CDH1 expression was an independent prognostic factor for PC. Previously, Deeb G *et al.* \\[[@R26]\\] found that cytoplasmic staining of CDH1 in lung cancer tissues correlates with shorter patient survival. Ito K *et al.* \\[[@R27]\\] revealed that CDH1 cytoplasmic staining may be due to CDH1 proteolytic cleavage by a membrane-bound metalloprotease, yielding a soluble form. Although nuclear staining of CDH1 protein has been associated with skin Merkel cell carcinomas \\[[@R28]\\], we did not observe nuclear CDH1 in our PC patient cohort. Taken together, cytoplasmic CDH1 expression appears to represent altered protein localization related to PC tumorigenicity.\n\nHDAC3 is the only class I HDAC found in the nucleus, cytoplasm, and plasma membrane \\[[@R29], [@R30]\\]. Previous studies focused on its function as an epigenetic modifier, repressing transcription through histone deacetylation \\[[@R10], [@R31], [@R32]\\]. Few studies have investigated the prognostic role of altered HDAC3 localization in PC. In this study, we found HDAC3 in the cytoplasm and nucleus of tumor cells, but not on the plasma membrane. Higher nuclear HDAC3 expression was observed in PC relative to adjacent normal tissues, while cytoplasmic expression of HDAC3 was indistinguishable. Cytoplasmic staining of HDAC3 was not associated with any clinicopathologic features or survival in PC patients. In contrast, increased nuclear HDAC3 expression was strongly associated with N classification and advanced clinical stage. For example, nuclear HDAC3 expression was detected in 80.0% of patients with high tumor grade (\\> IIa), but only 36.7% in the low tumor grade group (\u2264 IIa), suggesting that nuclear HDAC3 plays an important role in tumor progression in PC patients. Univariate analysis showed that nuclear HDAC3 in PC was associated with patients' OS. Higher nuclear HDAC3 correlates with worse prognosis. Furthermore, according to multivariate analysis, overexpression of nuclear HDAC3 has independent prognostic significance for PC. It is of particular note that high nuclear HDAC3 expression was positively associated with increased cytoplasmic CDH1. High co-expression of these two proteins correlated with shorter patient survival, with a cumulative 1-year survival of 12% (95% CI, 0.002--0.238) compared to that of 64% (95% CI, 0.522--0.758) in other expression levels group. Escaffit F *et al.* \\[[@R33]\\] reported that nuclear localization of HDAC3 decreases the efficiency of apoptosis induction, and HDAC3 cytoplasmic relocalization is important for the apoptotic process.\n\nWe speculate that first, pancreatic tumor cells may have escaped apoptosis, at least in part, through HDAC3 overexpression in cell nucleus. Secondly, high concentrations of nuclear HDAC3 may directly inhibit CDH1 promoters, leading to reduced CDH1 cell membrane expression. Additionally, nuclear HDAC3 expression may upregulate membrane-bound metalloprotease expression through epigenetic modification of the associated target gene, leading to increased cytoplasmic CDH1. Together, our findings strongly indicate that nuclear HDAC3 upregulation is crucial for the aggressive behaviors and worse prognosis of PC patients, which suggest that HDAC3 may be an effective therapeutic target. Unfortunately, clinical data for HDAC inhibitors (HDACIs) are inadequate, because few studies have included patients with PC and few PC patients entered the HDACIs phase II/III trials that did \\[[@R34]\\]. More high quality clinical trials recruiting candidates with PC are required to determine the efficacy of these therapies. Selective HDACIs, potentially targeting HDAC3, may yield more potent efficacy and fewer side effects than pan-HDACIs.\n\nIn summary, these data strongly suggest the importance of nuclear HDAC3 and cytoplasmic CDH1 in the progression and clinical outcome of human PC. These markers provide strong candidates for targeted therapy of PC patients. Larger prospective studies could further validate these findings.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nPatients and tissue samples {#s4_1}\n---------------------------\n\nThis study was approved by the Ethics and Research Committees of Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, and was conducted in accordance with the Declaration of Helsinki Principles. TMAs containing 90 PC tissues and corresponding non-tumor tissues were purchased from ShGnghGi Outdo Biotech Company (China). The TMAs contained well-documented clinicopathological information, including patients' age, sex, tumor size and location, tumor differentiation, invasion depth, lymph node metastasis, distant metastasis, clinical stage, abdominal pain, jaundice, nervous invasion, and follow-up data (ended in December, 2011). Six patients were excluded due to lack of completed clinical and follow-up data. In total, 84 patients were included, 51 males and 33 females, with a median age of 62 years old (ranging from 38 to 85 years old). The overall survival time ranged from 0 to 87 months, with a median of 15 months. Detailed information can be found in Table [8](#T8){ref-type=\"table\"}.\n\n###### Detailed clinical information of patients with PC\n\n  Characteristics                           Categories             Number\n  ----------------------------------------- ---------------------- --------\n  Overall survival median (range, months)   15 (0--87)             \n  Age median (range, years)                 62 (38--85)            \n  Tumor location                            Head, neck             56\n  Body, tail                                28                     \n  Tumor size (cm)                           \u2264 3                    25\n  \\> 3                                      59                     \n  Tumor differentiation                     Well, moderate         57\n  Poor                                      27                     \n  Invasion depth                            T1 + T2                71\n  T3 + T4                                   13                     \n  Lymph nodes metastasis                    N0 (negative)          51\n  N1 (positive)                             33                     \n  Distant metastasis                        Absent                 82\n  Present                                   2                      \n  Clinical stage                            Early stages (\u2264 IIa)   49\n  Advanced stages (\\> IIa)                  35                     \n  Abdominal pain                            Absent                 38\n  Present                                   46                     \n  Jaundice                                  Absent                 69\n  Present                                   15                     \n  Nervous invasion                          Negative               51\n  Positive                                  33                     \n\nImmunohistochemistry {#s4_2}\n--------------------\n\nImmunohistochemistry was performed based on the standard streptavidin-peroxidase (S-P) method (Zymed, San Francisco, CA). After deparaffinization and rehydration, TMA sections were subjected to high pressure for antigen retrieval for 5 minutes. Endogenous peroxidase activity was blocked using 100 \u03bcL of peroxidase block for 10 min. The slides were subsequently incubated overnight at 4\u00b0C with primary antibodies as follows: CDH1 (dilution 1:300, BD Biosciences), HDAC3 (dilution 1:500, Abcam). After washing in 1\u00d7 phosphate buffered saline (PBS), the sections were incubated with biotinylated secondary antibodies (Zymed, San Francisco, CA) for 30 min at room temperature, followed by incubation with streptavidin horseradish peroxidase complex. Finally, sections were incubated with DAB for 2 min. Positive controls were used in each experiment following supplier\\'s instructions. Negative controls applying appropriate IgG to replace primary antibody were also run in each experiment ([Supplementary Figure 1A, 1B](#SD1){ref-type=\"supplementary-material\"}).\n\nScoring of immunohistochemistry {#s4_3}\n-------------------------------\n\nA double-blind method, carried out independently by two investigators without access to the patients' clinical and pathological features, was used to analyze immunohistochemistry results. Five visual fields from different areas of each specimen were chosen at random for the immunohistochemistry evaluation. HDAC3 and CDH1 expression was scored according to staining intensity and the percentage of positive cells as previously described \\[[@R35]\\]. The percentage of positive cells was scored as follows: 0% (0), 1%--10% (1), 11%--50% (2) and 51%--100% (3). Staining intensity was scored as follows: no staining (0), week (1), moderate (2), and strong (3). Comprehensive score = staining percentage \u00d7 intensity. CDH1 or HDAC3 expression was classified as follows: \\< 6 low expression, \u2265 6 high expression.\n\nStatistical analysis {#s4_4}\n--------------------\n\nAll statistical analyses were carried out using the SPSS 13.0 software. The \u03c7^2^ test and Fisher\\'s exact test were used to analyze the correlation between the clinicopathologic characteristics and CDH1 and HDAC3 expression as appropriate. Overall survival (OS) was defined as the interval from date of diagnosis until death from any cause. Data were censored for living patients and patients lost between follow-ups. The OS was estimated using the Kaplan-Meier method and compared using the log-rank test. Significant variables were further analyzed by multivariate analysis to test for independent prognosis. Bivariate correlations between variable factors were calculated by Spearman rank correlation coefficients. *P*-values \\< 0.05 were considered statistically significant.\n\nSUPPLEMENTARY MATERIALS FIGURE AND TABLES {#s5}\n=========================================\n\nThis study was supported in part by the National Natural Science Foundation of China (grant NO. 81502017, 81502018, 81572315, 81101846, 81171887 and 91229117), by Research Grant from Shanghai Hospital Development Center (SHDC12014128), by National Key Clinical Discipline-Oncology, and by Songjiang Liandong Program(0702N14002).\n\n**CONFLICTS OF INTEREST**\n\nThe authors declare no conflicts of interests.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nThe global burden of diabetes among older people is growing. The age-standardised prevalence of diabetes has doubled since 1980 \\[[@CR1]\\] and in the year 2017 globally over 120\u00a0million people aged 65\u00a0years or older had diabetes \\[[@CR2]\\]. In Finland in 2005, approximately 20% of adults over 65\u00a0years were diagnosed with diabetes, whereas 30% of the older population had prediabetes, consisting of impaired glucose tolerance (IGT) or impaired fasting glucose (IFG) or a combination of the two \\[[@CR3]\\]. The prevalence of diabetes and prediabetes in Finland is similar to the prevalence in other European countries \\[[@CR2]\\]. Prediabetes has an annual risk of 5--10% of progressing to diabetes \\[[@CR4]\\] and it also increases the risk for microvascular complications traditionally associated with diabetes \\[[@CR5], [@CR6]\\]. As the population ages due to increased life expectancy and a fall in fertility, non-communicable diseases such as heart disease and diabetes will become even more important causes of morbidity and mortality, regardless of income level \\[[@CR7]\\]. This calls for effective primary and secondary interventions in older people to reduce the disease burden and health care costs \\[[@CR8]\\].\n\nDecreased physical performance, an important predictor of disability and functional decline, has been shown to have negative consequences on the daily life of older people \\[[@CR9]--[@CR11]\\]. Diabetes is one of the major causes of physical limitation and individuals with diabetes have approximately 50--80% greater risk of disability compared to those without diabetes \\[[@CR12]\\]. Age-related sarcopenia, characterised by loss of muscle strength and muscle quality, is accelerated by diabetes especially in the lower extremities \\[[@CR13], [@CR14]\\]. This increases the prevalence of mobility limitation and frailty at an earlier age, which reduces quality of life and results in loss of independence and institutionalisation \\[[@CR15], [@CR16]\\]. Disability and poor physical performance also lead to a rise in health care costs and an increased mortality rate \\[[@CR17], [@CR18]\\].\n\nAlthough there is an apparent correlation between diabetes and physical disability, most studies focusing on this subject have been cross-sectional \\[[@CR12], [@CR19], [@CR20]\\]. On the other hand, longitudinal studies have often relied on self-reported diagnosis of diabetes or self-reported physical performance, mainly focusing on activities of daily living (ADL), instrumental activities of daily living (IADL) and motility \\[[@CR21]--[@CR23]\\]. Furthermore, there are no longitudinal studies assessing the relationship between prediabetes and physical performance.\n\nTo address this question, we studied the association between glucose regulation at an average age of 61\u00a0years and the objectively measured physical performance evaluated with the Senior Fitness Test (SFT) approximately 10\u00a0years later using data from the Helsinki Birth Cohort Study.\n\nMaterials and methods {#Sec2}\n=====================\n\nStudy population {#Sec3}\n----------------\n\nThe Helsinki Birth Cohort Study (HBCS) includes 13,345 individuals born between 1934 and 1944\u00a0at the Helsinki University Central Hospital or the Helsinki City Maternity Hospital \\[[@CR24]\\]. All subjects included in this study cohort attended child welfare clinics in Helsinki and were still living in Finland in 1971, when all Finnish residents received a personal identification number \\[[@CR25]\\]. A random sample of 2902 from those 8760 individuals who were born at the Helsinki University Central Hospital was invited to a baseline clinical examination in the year 2000. A total of 2003 cohort members participated in an examination conducted between 2001 and 2004. In 2011, members of the clinical cohort still alive and living within a 100-km radius of Helsinki were invited to participate in a clinical follow-up examination. A total of 1094 subjects participated in this clinical examination between 2011 and 2013. The cohort members that participated only in the baseline examination (*n*\u2009=\u2009925) were older, more frequently men and smokers and they had higher BMI (all *p*\u2009\\<\u20090.028) compared to those who were included in our study. The main reasons for declining invitation to the follow-up examination were related to personal or a family member's health conditions. This study includes the 1078 subjects who had adequate information on physical performance and glucose regulation \\[[@CR26]\\]. The study was approved by the Ethics Committee of Epidemiology and Public Health of the Hospital District of Helsinki and Uusimaa and that of the National Public Health Institute, Helsinki and follows the guidelines of the Declaration of Helsinki. All participants gave a written, informed consent.\n\nGlucose regulation {#Sec4}\n------------------\n\nFasting plasma glucose was measured in all subjects at the baseline clinical examination in 2001 to 2004. A standard 2-h 75-g oral glucose tolerance test (OGTT) was conducted in all individuals, except for those with previously known type 1 or type 2 diabetes (*n*\u2009=\u200950), defined by self-report, medical records or use of medication for diabetes. The World Health Organization 1999 criteria \\[[@CR27]\\] were used for diagnosing diabetes, IGT, and IFG. Subjects reporting a diagnosis of diabetes or taking medications for diabetes before the clinical examination were classified as having previously known diabetes. Those diagnosed with type 2 diabetes for the first time at the OGTT were classified as having newly diagnosed diabetes. Subjects meeting the criteria for both IGT and IFG were categorised as having IGT.\n\nPhysical performance {#Sec5}\n--------------------\n\nSubjects in the clinical follow-up were assessed for physical performance with the validated Senior Fitness Test (SFT) \\[[@CR28], [@CR29]\\]. A modified test battery consisting of five tests was carried out: (1) 30-s chair stand: number of full stands completed in 30\u00a0s with arms folded across chest to assess lower body strength, (2) Arm curl: number of bicep curls recorded in 30\u00a0s holding hand weight (3\u00a0kg for men and 2\u00a0kg for women) to assess upper body strength, (3) Back scratch: with one hand reaching over shoulder and the other one up middle back, number of centimetres between extended middle fingers to assess upper body (shoulder) flexibility, (4) Chair sit-and-reach: sitting at the front of chair with leg extended, number of centimetres between extended fingers and tip of toe to assess lower body (hamstring) flexibility, (5) 6\u00a0min walk: number of meters walked in 6\u00a0min to assess aerobic endurance. Measurements were performed by trained research assistants \\[[@CR26]\\]. Subjects were rated for each test using percentile tables of normative data for 5-year age groups \\[[@CR29]\\]. The rating varied between 1 and 20, based on 5-\u2030 ranges, with 1 indicating a test result below the fifth percentile and 20 indicating a test result in the top 5\u00a0\u2030. Finally, the overall score was calculated as the sum of the normalised ratings for all five SFT items. The overall score varied between 5 and 100.\n\nCovariates {#Sec6}\n----------\n\nQuestionnaires were used at the baseline clinical examination to assess current health situation, use of medication, educational attainment and lifestyle characteristics. Anthropometric measurements including height, weight and waist circumference were measured. Body mass index (kg/m^2^) was calculated, and lean body mass and fat percentage were estimated with bioelectrical impedance using the InBody 3.0 eight polar tactile electrode system (Biospace Co., Ltd., Seoul, Korea). Physical activity was assessed using a validated exercise questionnaire; the Kuopio Ischaemic Heart Disease Risk Factor Study (KIHD) 12-month leisure time physical activity (LTPA) history \\[[@CR30]\\]. The questionnaire assigned a metabolic equivalent of task (MET) value for each specific activity and intensity. Physical activity is presented as MET-hours per day based on a 12-month history.\n\nStatistical analysis {#Sec7}\n--------------------\n\nMeans and standard deviations or medians and interquartile ranges were calculated for continuous variables, whereas categorical variables are presented as frequencies and proportions. The associations between the characteristics of the cohort members and physical performance were assessed using linear regression analyses. Multiple linear regression models were used to assess the association between glucose regulation and physical performance. The results are presented pooled by sex, as none of the interactions for sex and the glucose regulation variables on the Senior Fitness Test were statistically significant (all *p* values\u2009\\>\u20090.52). In model 1, we adjusted for sex and age. Next, we added adult socioeconomic status, smoking status, alcohol consumption and physical activity to model 2. In model 3, we further adjusted for body fat percentage. All tests were performed two-tailed and the level of significance was set at *p*\u2009\\<\u20090.05. Statistical analyses were carried out using IBM SPSS Statistics Version 24 (IBM Corp., Armonk, NY, USA).\n\nResults {#Sec8}\n=======\n\nA total of 475 men and 603 women were included in the analysis (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The mean age at the baseline clinical examination was 61.2 (SD 2.6) years for men and 61.3 (SD 2.9) years for women. The prevalence of previously known diabetes was 6.7% for men and 3.0% for women. At the OGTT, 9.7% of the men and 5.1% of the women were diagnosed with new diabetes. Prediabetes had a higher prevalence, with 30.1% of the men and 29.2% of the women diagnosed with either IGT or IFG. The mean follow-up time between the OGTT and the SFT was 9.7 (SD 0.9) years for the entire cohort. Men had a mean age of 70.8 (SD 2.6) years and women a mean age of 71.0 (SD 2.8) years when physical performance was assessed.\n\nTable 1Baseline characteristics for men and women in the Helsinki birth cohort studyVariableMen (*n*\u2009=\u2009475)Women (*n*\u2009=\u2009603)MeanSDMeanSDAge at baseline (years)61.22.661.32.9Weight (kg)85.512.272.412.7Height (cm)177.25.9163.45.7BMI (kg/m^2^)27.23.527.14.6Lean body mass (kg)65.57.347.65.3Body fat (percentage)22.95.533.26.6Fat mass (kg)19.97.124.68.9Waist circumference (cm)99.610.389.311.7Smoking status (*r*/*n*, %)\u00a0Never smoked134/47428.3344/60357.0\u00a0Ex-smoker228/47448.1154/60325.5\u00a0Smoker112/47423.6105/60317.4Alcohol consumption (*r*/*n*, %)\u00a0Never used/ex-user29/4716.228/6004.7\u00a0Less than 1 day a week126/47126.8311/60051.8\u00a0Weekly316/47167.1261/60043.5Adult socioeconomic status (*r*/*n*, %)\u00a0Labourer186/47539.2123/60320.4\u00a0Lower middle133/47528.0363/60360.2\u00a0Upper middle112/47523.669/60311.4\u00a0Self-employed44/4759.348/6038.0Self-reported physical activity (*r*/*n*, %)\u00a0Sedentary35/4737.481/60213.5\u00a01--2 times/week227/47348.0264/60243.9\u00a0At least 3 times/week211/47344.6257/60242.7LTPA (MET-h/Day)^a^5.23.0 to 9.04.92.8 to 8.2Education level (*r*/*n*, %)\u00a0Basic116/47524.4230/60338.1\u00a0Upper secondary115/47524.2157/60326.0\u00a0Lower tertiary143/47530.1159/60326.4\u00a0Upper tertiary101/47521.357/6039.5Glucose regulation (*r*/*n*, %)\u00a0Normoglycaemic254/47553.5378/60362.7\u00a0IFG45/4759.529/6034.8\u00a0IGT98/47520.6147/60324.4\u00a0Newly diagnosed diabetes46/4759.731/6035.1\u00a0Previously known diabetes32/4756.718/6033.0^a^ *LTPA* leisure time physical activity, values are given as median and interquartile ranges. *MET* metabolic equivalent of task, *IFG* impaired fasting glucose, *IGT* impaired glucose tolerance, Newly Diagnosed Diabetes: Diabetes diagnosed with oral glucose tolerance test at clinical examination during 2001--2004, Previously Known Diabetes: Diabetes diagnosed before clinical examination in 2001--2004\n\nThe mean overall SFT score was 45.0 (SD 17.5) points for the entire cohort. Women had a 4.40 point (*p*\u2009\\<\u20090.001) higher SFT score compared to men (Table\u00a0[2](#Tab2){ref-type=\"table\"}). Baseline lean body mass and body fat percentage showed a significant inverse association with the SFT score. Subjects with previously known diabetes at baseline had a 17.77 lower mean SFT score (*p*\u2009\\<\u20090.001) compared to subjects with normoglycaemia. Those with newly diagnosed diabetes (*p*\u2009\\<\u20090.001) and IGT (*p*\u2009\\<\u20090.001) also had a significantly lower SFT score compared to their normoglycaemic counterparts. There was no significant difference in SFT score between those with IFG and those with normoglycaemia (*p*\u2009=\u20090.26).\n\nTable 2Associations (univariate) between subject characteristics and SFT scoreVariable*N*\u2009=\u20091078*b*95% CI*p*Age at senior fitness test (years)\u2212\u20090.09\u2212\u20090.48 to 0.290.64Sex\u00a0MenRef----\u00a0Women4.402.23 to 6.49\\<\u20090.001Height (cm)\u2212\u20090.10\u2212\u20090.21 to 0.020.11Lean body mass (kg)\u2212\u20090.27\u2212\u20090.36 to \u2212\u20090.17\\<\u20090.001Body fat (%)\u2212\u20090.51\u2212\u20090.64 to \u2212\u20090.38\\<\u20090.001Smoking status\u00a0Never smokedRef----\u00a0Ex-smoker\u2212\u20091.00\u2212\u20093.31 to 1.320.40\u00a0Smoker\u2212\u20099.18\u2212\u200911.93 to \u2212\u20096.42\\<\u20090.001Alcohol consumption\u00a0Never used/ex-userRef----\u00a0Less than 1 day a week4.68\u2212\u20090.13 to 9.500.06\u00a0Weekly5.921.17 to 10.660.02Adult socioeconomic status\u00a0LabourerRef----\u00a0Lower middle5.683.22 to 8.14\\<\u20090.001\u00a0Upper middle7.524.34 to 10.70\\<\u20090.001\u00a0Self-employed5.901.88 to 9.930.004LTPA (MET-h/day)0.15\u2212\u20090.05 to 0.350.14Glucose regulation\u00a0NormoglycaemicRef----\u00a0IFG\u2212\u20092.37\u2212\u20096.47 to 1.740.26\u00a0IGT\u2212\u20095.15\u2212\u20097.65 to \u2212\u20092.65\\<\u20090.001\u00a0Newly diagnosed diabetes\u2212\u200910.55\u2212\u200914.56 to \u2212\u20096.54\\<\u20090.001\u00a0Previously known diabetes\u2212\u200917.77\u2212\u200922.65 to \u2212\u200912.90\\<\u20090.001*SFT* senior fitness test, *b* regression coefficient (unstandardised), *LTPA* leisure time physical activity, *MET* metabolic equivalent of task, *IFG* impaired fasting glucose, *IGT* impaired glucose tolerance\n\nTable\u00a0[3](#Tab3){ref-type=\"table\"} shows the association between glucose regulation and physical performance 10\u00a0years later. There was no significant difference in physical performance between subjects with IFG and subjects with normoglycaemia. Those with previously known diabetes, newly diagnosed diabetes and IGT all had a lower SFT score compared to those with normoglycaemia. There was a graded increase in the inverse association between more severe impairment in glucose regulation and physical performance, as subjects with IGT showed the weakest association and subjects with previously known diabetes showed the strongest inverse association with the overall SFT score. After adjusting for sex, age, adult socioeconomic status, and lifestyle factors, the regression coefficients were only slightly attenuated. However, further adjustment for body fat percentage had an impact on these associations. In the fully adjusted model, the mean difference in SFT score between subjects with previously known diabetes and normoglycaemic subjects was \u2212\u200911.56 points (95% CI \u2212\u200916.15 to \u2212\u20096.98, *p*\u2009\\<\u20090.001).\n\nTable 3Regression coefficients for the association between glucose regulation and SFT scoreGlucose regulation^a^Model 1Model 2Model 3*b*95% CI*pb*95% CI*pb*95% CI*p*IFG\u2212\u20091.69\u2212\u20095.80 to 2.420.42\u2212\u20092.03\u2212\u20096.04 to 1.980.320.47\u2212\u20093.47 to 4.400.82IGT\u2212\u20095.13\u2212\u20097.63 to \u2212\u20092.63\\<\u20090.001\u2212\u20095.31\u2212\u20097.77 to \u2212\u20092.86\\<\u20090.001\u2212\u20092.56\u2212\u20094.96 to \u2212\u20090.160.04Newly diagnosed diabetes\u2212\u20099.86\u2212\u200913.87 to \u2212\u20095.84\\<\u20090.001\u2212\u20098.97\u2212\u200912.89 to \u2212\u20095.05\\<\u20090.001\u2212\u20095.49\u2212\u20099.26 to \u2212\u20091.720.004Previously known diabetes\u2212\u200916.95\u2212\u200921.84 to \u2212\u200912.07\\<\u20090.001\u2212\u200915.72\u2212\u200920.54 to \u2212\u200910.91\\<\u20090.001\u2212\u200911.56\u2212\u200916.15 to \u2212\u20096.98\\<\u20090.001^a^Normoglycaemia serves as the reference. Model 1: Adjusted for sex and age. Model 2: Further adjusted for adult socioeconomic status, smoking status, alcohol consumption and physical activity. Model 3: Further adjusted for body fat percentage. *SFT* senior fitness test, *b* regression coefficient (unstandardised), *IFG* impaired fasting glucose, *IGT* impaired glucose tolerance\n\nThe association between glucose regulation and the age standardised scores for the individual SFT tasks are presented in Table\u00a0[4](#Tab4){ref-type=\"table\"}. In the fully adjusted model, subjects with previously known diabetes performed poorer in all tasks except the chair sit-and-reach test compared to those with normoglycaemia. The greatest difference between subjects with previously known diabetes and subjects with normoglycaemia was in the 6-min walk test (mean difference \u2212\u20093.38, 95% CI \u2212\u20094.76 to \u2212\u20092.00, *p*\u2009\\<\u20090.001). Subjects with newly diagnosed diabetes performed significantly poorer in both the arm curl test and the back scratch test. Subjects with IGT performed poorer in the arm curl test, but there was no difference in the rest of the individual SFT tasks compared to subjects with normoglycaemia.\n\nTable 4Regression coefficients for the association between glucose regulation and individual SFT test components^a^Glucose regulation^b^Chair standArm curlBack scratchChair sit-and-reach6-min walk*b*95% CI*pb*95% CI*pb*95% CI*pb*95% CI*pb*95% CI*p*IFG0.32\u2212\u20090.57 to 1.210.48\u2212\u20090.37\u2212\u20091.52 to 0.790.53\u2212\u20090.33\u2212\u20091.75 to 1.080.65\u2212\u20090.14\u2212\u20091.60 to 1.320.851.08\u2212\u20090.10 to 2.250.07IGT\u2212\u20090.29\u2212\u20090.84 to 0.250.29\u2212\u20090.71\u2212\u20091.43 to \u2212\u20090.000.049\u2212\u20090.36\u2212\u20091.23 to 0.520.42\u2212\u20090.72\u2212\u20091.61 to 0.180.12\u2212\u20090.50\u2212\u20091.22 to 0.220.17Newly diagnosed diabetes\u2212\u20090.73\u2212\u20091.58 to 0.120.09\u2212\u20091.48\u2212\u20092.60 to \u2212\u20090.370.01\u2212\u20091.72\u2212\u20093.10 to \u2212\u20090.350.01\u2212\u20090.51\u2212\u20091.92 to 0.900.48\u2212\u20091.00\u2212\u20092.14 to 0.140.09Previously known diabetes\u2212\u20091.95\u2212\u20092.98 to \u2212\u20090.92\\<\u20090.001\u2212\u20092.25\u2212\u20093.61 to \u2212\u20090.900.001\u2212\u20092.85\u2212\u20094.58 to \u2212\u20091.120.001\u2212\u20091.14\u2212\u20092.84 to 0.570.19\u2212\u20093.38\u2212\u20094.76 to \u2212\u20092.00\\<\u20090.001^a^Adjusted for sex, age, adult socioeconomic status, smoking status, alcohol consumption, physical activity and body fat percentage^b^Normoglycaemia serves as the reference. *SFT* senior fitness test, *b* regression coefficient (unstandardised), *IFG* impaired fasting glucose, *IGT* impaired glucose tolerance\n\nDiscussion {#Sec9}\n==========\n\nIn this study of community dwelling older people, we found that impaired glucose regulation was strongly associated with decreased physical performance nearly 10\u00a0years later. We showed that physical performance was poorer not only among individuals with diabetes, but also among those with IGT, even after controlling for confounding factors. There was a gradual decline in physical performance when transitioning to more severe disturbances in glucose regulation, with subjects previously diagnosed with diabetes having the poorest physical performance. To the best of our knowledge, we are the first to report this kind of relationship between glucose regulation and physical performance in a longitudinal study setting.\n\nWe confirmed the results from previous studies, suggesting an association between diabetes and decreased physical performance. We found that after controlling for confounders, subjects with previously diagnosed diabetes had approximately 0.7 SD lower SFT score compared to subjects with normoglycaemia. In a meta-analysis by Wong and colleagues \\[[@CR12]\\] including 26 studies, diabetes was shown to increase the risk of mobility disability, IADL disability and ADL disability. In a 7-year follow-up study of Mexican-Americans, diabetes was associated with roughly a 1.5--2 times greater risk of lower body disability compared to normoglycaemia \\[[@CR31]\\]. Some data, although limited, are also available on the relationship between IGT and physical performance. In a cross-sectional study of 1391 individuals, both IGT and diabetes increased the risk of poor physical performance \\[[@CR32]\\]. These results together with our findings suggest that physical performance is affected already at the stage of IGT.\n\nSeveral mechanisms may explain the association between impaired glucose regulation and decreased physical performance. Insulin resistance has been shown to cause increased protein degradation and decreased protein synthesis in skeletal muscle. In the long term, this may cause loss of muscle mass, which further aggravates glucose regulation, as skeletal muscle is important for the uptake of glucose \\[[@CR33]\\]. Inflammatory markers associated with insulin resistance and obesity, such as interleukin-6 and C-reactive protein (CRP), may also contribute to functional decline by causing negative changes in the regulation of skeletal muscle homeostasis \\[[@CR16]\\]. In addition, physical inactivity, mitochondrial dysfunction, decreased skeletal muscle blood flow and hyperglycaemia itself may explain the excess disability among adults with abnormal glucose regulation \\[[@CR34]\\].\n\nIn our analysis, adjusting for body fat percentage weakened the association between glucose regulation and physical performance. Body fat percentage and obesity are strongly associated with disturbances in glucose regulation and have been shown to be important predictors of poor physical performance \\[[@CR26], [@CR35]\\]. Excess body fat triggers low-grade systemic inflammation and may cause accumulation of both intermuscular and intramuscular fat deposits \\[[@CR36]\\]. This has been shown to have a negative impact on muscle strength and increase the risk of mobility disability \\[[@CR34]\\]. Additionally, ectopic fat also affects glucose metabolism and it is suggested that this could be one of the triggers of insulin resistance \\[[@CR33]\\]. Functional decline and physical inactivity, on the other hand, further increases the risk of obesity, metabolic syndrome and other chronic diseases \\[[@CR37]\\].\n\nCompared to subjects with normoglycaemia, we found that people with previously known diabetes had, besides the lowest overall SFT score, also a poorer result on most SFT test components. This suggests that over time, many different elements of physical performance important in daily life are affected among those with diabetes, including strength, endurance and flexibility. Previously known diabetes had the strongest inverse association with the 6-min walk test, indicating that lower body function in particular is decreased among individuals with diabetes. Newly diagnosed diabetes, on the other hand, was inversely associated with only two out of five performance tests. This indicates that longer duration of diabetes is an important risk factor of decreased physical function \\[[@CR9], [@CR13]\\].\n\nFrom a public health point of view, our results show optimistic possibilities for potentially slowing down functional decline among older people. By diagnosing impairments in glucose regulation at an early stage, lifestyle interventions and patient education might prevent further decline in glucose regulation \\[[@CR5]\\], thereby promoting the maintenance of physical performance. This could ultimately lead to a reduction of health care costs and an increase in functional independence of older people \\[[@CR17]\\]. Pharmacotherapy, such as metformin, reduces the risk of progression to diabetes among subjects with prediabetes, however, this treatment has been shown to be less effective compared to lifestyle interventions \\[[@CR38]\\]. A topic for future research would be to investigate if metformin could slow down the decline in physical performance among individuals with IGT.\n\nThis study has several strengths. We studied a large population which comprised both men and women. The long follow-up time of approximately 10\u00a0years enabled us to evaluate the long-term association between glucose regulation and physical performance. We used the OGTT to assess glucose regulation and were thereby able to correctly classify subjects with a high sensitivity. Relying solely on fasting plasma glucose or HbA~1c~ when diagnosing diabetes has been shown to be less sensitive compared to using an OGTT, with nearly 50% of undiagnosed diabetes detected only with an OGTT \\[[@CR39]\\]. Another strength is that we used the validated SFT to objectively measure physical performance among our subjects. Compared to testing single muscle groups, such as grip strength or gait speed, the SFT can be used to evaluate full-body physical performance, including strength, flexibility and endurance. Another popular test of physical performance, the Short Physical Performance Battery (SPPB), although effective in predicting disability, nursing home admittance and survival rates among older people, includes test components that have been shown to be either too easy or too difficult to perform \\[[@CR40]\\].\n\nThis study also has some limitations. First, although we were able to adjust for several covariates, we did not adjust for complications associated with diabetes, including neuropathy, retinopathy and peripheral vascular disease. These factors have been shown to be associated with disability and functional impairment among older people \\[[@CR35], [@CR41]\\]. In addition, there may be other unmeasured variables affecting our results. Second, physical performance was not assessed at the baseline clinical examination, thus, we were not able to report potential changes in physical performance during the follow-up time. This prevented us from addressing causality. The fact that individuals with the longest duration of diabetes had the poorest physical performance suggests, however, that impaired glucose regulation causes a decline in physical performance over time. Third, we acknowledge the possibility of a selection bias, as only half of the subjects from the baseline clinical examination participated in the examination of physical performance. Fourth, we studied a homogenous population of Caucasians only living in a restricted area of Finland, and this should be taken into account when implementing our findings in other populations.\n\nIn summary, there is a strong association between disturbances in glucose regulation and poor physical performance among older people. Decline in physical function is known to begin already at an early stage of impaired glucose regulation, however, diabetes, and in particular, a long duration of diabetes further exacerbates this decline and affects several different aspects of physical performance among individuals with advanced age.\n\nHBCS was supported by Emil Aaltonen Foundation, Finnish Foundation for Cardiovascular Research, Finnish Foundation for Diabetes Research, Finnish Foundation for Pediatric Research, Juho Vainio Foundation, Novo Nordisk Foundation, Signe and Ane Gyllenberg Foundation, Samfundet Folkh\u00e4lsan, Finska L\u00e4kares\u00e4llskapet, Liv och H\u00e4lsa, European Commission FP7 (DORIAN) Grant Agreement No. 278603 and EU H2020-PHC-2014-DynaHealth Grant No. 633595. The Academy of Finland supported MBvB (Grant No. 257239); EK (Grant Nos. 127437, 129306, 130326, 134791, 263924 and 274794); JGE (Grant Nos. 129369, 129907, 135072, 129255 and 126775).\n\nConflict of interest {#FPar1}\n====================\n\nThe authors declare that they have no conflict of interest.\n\nHuman and animal rights statement {#FPar2}\n=================================\n\nAll procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.\n\nInformed consent {#FPar3}\n================\n\nInformed consent was obtained from all individual participants included in the study.\n\n[^1]: Managed by Massimo Federici.\n"}
+{"text": "Introduction\n============\n\n*Mycobacterium tuberculosis* H37Rv (Mtb), a causative agent of tuberculosis (TB), has remained a major health concern globally. Based on the World Health Organization (WHO) latest reports, it is estimated that there have been 8.6 million new cases of TB reported in 2012 and a total of 1.3 million TB deaths ([@bav060-B1]). Most patients are treated for TB using first-line drugs, rifampicin and isoniazid. Together with the other first-line drugs, ethambutol and pyrazinamide, these two drugs form the basic ingredients of combination chemotherapy followed by the WHO directly observed treatment short course strategy ([@bav060-B2]). Second-line drugs, such as fluoroquinolones, and injectables like kanamycin, capreomycin and amikacin, are relied upon when the first-line drugs fail to control the disease. However in recent times, many antibiotic-resistant strains of *Mtb* have been reported. Multi-drug resistant TB is an Mtb strain resistant to rifampicin and isoniazid. Furthermore, acquisition of resistance towards fluoroquinolone, along with at least one of the injectable drugs, causes extensively drug resistant TB ([@bav060-B3]). The emergence of resistant strains to the first- and second-line drugs currently used poses a mammoth challenge for control of TB and cure of the infected.\n\nOff-target effects, which are often discovered at the later stages of drug discovery research, have led to failure of many new medicines. Thus, there is an urgent need to discover ways of identifying off-target sites for drugs at an early stage in research. Detailed structural knowledge of the interactions between molecules in the cell provides one way of approaching this problem. The objective would be to define the structural interactome, an inventory of the various interactions between macromolecules and both natural and synthetic small molecules. The structural interactome can augment molecular-level interaction networks and provide a rich source of information on interactions between biological molecules and natural or synthetic ligands. Information on interactions between host and pathogen proteins will be helpful in identifying targets among pathogen proteins.\n\nIntegrated databases defining the structural interactome, bringing together information on protein sequences and structures, binding site properties, small molecules and their interactions, provide a valuable resource. Existing integrated databases on TB, TB Database ([@bav060-B4], [@bav060-B5]) and Tuberculist ([@bav060-B6]), provide information on genome, proteome, expression as well as corresponding references in the scientific literature but provide no information on structural interactomics comprising of binding sites, small molecules, druggability analysis of targets and functional domain assignments. This work from the CamBan (Cambridge---Bangalore) collaboration, involving four independent research groups from Cambridge and Bangalore, brings together various resources developed by these research groups and elsewhere to provide an extended Mtb structural-interactome resource. Each group has contributed towards the data specific to TB using in-house algorithms and databases developed and established individually over the years. The algorithms used to generate the data are designed to address and enrich sequence and structural data along with various small-molecule interactions. The database also incorporates systems-based analysis and provides list of high-confidence targets.\n\nSensitive profile-based techniques such as hmmscan of HMMER3.0 ([@bav060-B7]), Reverse PSI-BLAST (Reverse Position-Specific Iterative Basic Local Alignment Search Tool) ([@bav060-B8]) and HHblits ([@bav060-B9]) were used to achieve enhanced domain annotation for the proteins. Structural annotation of *M.* *tuberculosis* proteome ([@bav060-B10]) and CHOPIN ([@bav060-B11]) database provided structural data for many proteins. PocketDepth ([@bav060-B12]), PocketMatch ([@bav060-B13]) and PocketAlign ([@bav060-B14]) algorithms are used for binding site prediction and comparison. Protein domain analysis of unannotated genes was pursued using a computationally intensive bioinformatics pipeline called PURE (Prediction of Unassigned Regions) ([@bav060-B15]). The dataset from CREDO ([@bav060-B16]), a protein--ligand interaction database for drug discovery, TIMBAL ([@bav060-B17]), a database of small molecules disrupting protein--protein interactions and TIBLE (<http://mordred.bioc.cam.ac.uk/tible/>), a database of small molecules against *Mtb* and ligand-based off-target predictions, are connected in structural interactome computational resource (SInCRe). High-confidence drug targets derived from targetTB ([@bav060-B18]) have been included in the database. Drug targets have also been identified by a sequence-based approach with the help of FDA-approved drugs and are incorporated into the database. An interface has also been provided to integrate the data available from other external resources like STRING ([@bav060-B19]), STITCH ([@bav060-B20]) and Tuberculist ([@bav060-B6]). Future works from these groups will be mapped on to Mtb-specific dataset, and SInCRe will be updated on a regular basis.\n\nDatabase\n========\n\nThe integrated suite of databases was developed to provide detailed sequence and structure-based dimensionality to aid in drug-discovery pipelines. The data integrated here were obtained from the databases and webservers developed individually by the four research groups in CamBan. In-house algorithms and databases are the primary resources for the database. [Figure 1](#bav060-F1){ref-type=\"fig\"} details the various data types. Figure 1.Various data resources contributing to SInCRe.\n\nSequence-based analysis\n-----------------------\n\n### Algorithms and datasets used.\n\nThe development of similarity-search procedures with the use of profiles such as Position Specific Scoring Matrices (PSSMs) ([@bav060-B8], [@bav060-B21]), Environment-Specific Substitution Tables ([@bav060-B22]) and Hidden Markov Models (HMMs) ([@bav060-B7]) has proven to be sensitive in detecting remote homologues reliably. Combination of such sensitive profile-based techniques resulted in the structural and functional annotations for \u223c95% of the Mtb proteome. Sensitive approaches such as hmmscan available through HMMER3.0 package, RPS-BLAST and HHblits were employed against sequence and structural profiles of various domain families obtained from Pfam ([@bav060-B23]), SUPERFAMILY ([@bav060-B24]), MulPSSM ([@bav060-B25], [@bav060-B26]) and the HH-suite database (<ftp://toolkit.genzentrum.lmu.de/pub/HHsuite/databases/hhsuite_dbs/>).\n\nMulPSSM, developed by one of our research groups, is a searchable database of multiple PSSM profiles. The multiple profiles for a given domain family correspond to an alignment, wherein multiple sequences from that family are used as reference. The current version comprises 403 107 profiles for 14 831 Pfam domain families (Pfam v.27) and 14 235 profiles corresponding to 3856 structural families in PALI (Phylogeny and ALIgnment of homologous protein structures) ([@bav060-B27]) database based on SCOP (Structural Classification Of Proteins) (version 1.75).\n\n### Assessment of structure and function predictions.\n\nEach of the associations made was assessed based on e-value and alignment length. For associations made using RPS-BLAST against MulPSSM database, an e-value cut-off of 0.01 was used in addition to a profile-coverage threshold of 70%. For domain assignments made using hmmscan against the Pfam database, profile-specific gathering threshold cut-offs were used to extract reliable hits. For hits identified by HHblits, an e-value threshold of 0.001 was used to associate domain families. In the searches against SUPERFAMILY HMM database using hmmscan software, the hits with e-values better than 0.0001 were considered to be reliable and are included in the database. Domain assignments for all the proteins were manually curated to maximize residue (64%) and sequence (89%) coverage.\n\nStructural/functional domains for un-annotated proteins as well as those with unassigned regions were determined with help of the computationally intensive pipeline, PURE ([@bav060-B15]), developed by one of our groups. Cases where all the earlier approaches were unsuccessful in recognition of structural or functional domains, a fold-recognition algorithm PHYRE2 ([@bav060-B28]) was employed. A confidence cut-off of 90% was considered to retrieve folds reliably. This exercise was essentially an attempt to assess the foldability of the protein in question. Also, transfer of function based on homology was pursued using HHblits against non-redundant sequence database at an e-value cut-off of 0.001 and query coverage threshold of 60%.\n\nAll the hits of the earlier approaches were coupled with manual intervention to ensure maximum residue and sequence coverage of Mtb proteome ([@bav060-B29]).\n\n### Drug target identification based on sequence information.\n\nRepurposing drugs has been regarded as a promising strategy mainly due to the reduced cost and time involved. A target identification methodology which essentially integrates homology and pharmacological information (G. Ramakrishnan, N. Chandra and N. Srinivasan, in preparation) facilitated recognition of 132 FDA-approved drugs which could be repurposed for 56 potential targets in *Mtb.* This methodology comprises three steps: exploration of evolutionary relationship between targets of known FDA-approved drugs and *Mtb* proteins, structural elucidation of binding sites of *Mtb* proteins homologous to known targets and evaluation of predicted binding sites with the help of protein--ligand docking. Evolutionary relationships were explored with the help of a sensitive profile-based iterative search tool, jackhmmer([@bav060-B30]), at an e-value threshold of 0.0001. An initial filtering step to eliminate drugs known to act on human proteins ensured that the 'anti-targets' in host are not picked up. The reliably identified relationships picked were further probed for the conservation of ligand-binding site residues across known targets and the *Mtb* proteins homologous to these targets. Structural information was taken from Protein Data Bank (PDB) ([@bav060-B31]) for *Mtb* and high-confidence structural models obtained from ModBase ([@bav060-B32]) for proteins with no known structure were used to assess the binding pockets. A structural alignment algorithm, TM-align ([@bav060-B33]), was effective in identification of highly similar local structural matches (TM-score\u2009\\>\u20090.50) between targets and their corresponding homologues in *Mtb.* Finally, the shortlisted proteins in *Mtb* predicted to serve as potential targets were evaluated using FDA-approved drugs with the help of an efficient protein--ligand docking tool, Glide ([@bav060-B34]) (<http://www.schrodinger.com/Glide>). A total of 132 FDA-approved drugs ([Supplementary Table S1](http://database.oxfordjournals.org/lookup/suppl/doi:10.1093/database/bav060/-/DC1)) were thus identified, which could be repurposed for 56 potential target proteins in *Mtb.*\n\nStructure-based analysis\n------------------------\n\n### Structural proteome of Mtb.\n\nStructural annotation of the *M.* *tuberculosis* proteome was carried out by one of the groups ([@bav060-B10]). PDB holds a total of 324 crystal structures of Mtb proteins and comparative models were generated for 2737 proteins, thus giving structure availability for 70% of the Mtb proteome. Structural models were generated using Modpipe, a software suite along with ModBase ([@bav060-B32]), a database of models generated using comparative modelling. The structural models need to be of high confidence and reliability as they play a central role to all the further analysis carried out. To assess the reliability of the protein structural models, various structure verification methods including statistical scoring potential ([@bav060-B37], [@bav060-B38]), secondary structure compatibility ([@bav060-B39]) and stereochemical quality check ([@bav060-B40]) were used. In the case of multi-domain proteins 3D models of individual domains are presented. Only those binding sites that were detected within the domains are analysed.\n\nThe CHOPIN ([@bav060-B11]) database (<http://structure.bioc.cam.ac.uk/chopin>) assigns structural domains and generates homology models for 2911 sequences, corresponding to \u223c73% of the proteome. Conformational states, characteristic of different oligomeric states and ligand binding, reflect various functional states of the proteins. Additionally, CHOPIN includes structural analyses of mutations potentially associated with drug resistance. The model number, sequence coverage and zscore are displayed on the SInCRe result page with links provided to CHOPIN webpage (<http://mordred.bioc.cam.ac.uk/chopin/about>) that provides model details and an option to download the models.\n\n### Detection of binding sites.\n\nComputational methods for binding site detection can be classified into three broad categories based on their approaches: (i) evolutionary methods based on structure--sequence alignment (ii) energy-based methods using chemical probes and (iii) geometric approaches that scan the 3D structure of the protein to detect pockets. Each of these methods has its own strengths and limitations with respect to different aspects such as accuracy in detection and prediction, computational time, complexity and features captured. All the three methods were used in this study to minimize the prediction error and increase the confidence. The methods used are, a grid-based geometric method, PocketDepth ([@bav060-B12]), evolutionary method, Ligsite ([@bav060-B41]) and energy-based method, SiteHound. PocketDepth is an in-house method that uses depth-based clustering algorithm for detecting putative binding sites in the given protein structures. The idea that depth is defined by the centrality of empty subspaces in a protein structure is used to identify the pockets from all the protein structures. The PocketDepth algorithm was later combined with LIGSITEcsc, which uses Connolly's surface ([@bav060-B42]) to identify surface--solvent--surface events that involves grooves and then detects binding sites in a given protein by mapping the degree of conservation of the residues in the selected surface. All the pockets detected by PocketDepth that are within 5 \u00c5 radius of the predicted LIGSITEcsc pockets were selected. SiteHound ([@bav060-B43]), an energy method that searches for interaction zones favourable for a methyl probe within the protein, was used on all the pockets identified as a filter to fetch out final set of consensus ligand binding sites.\n\nOther than the binding sites identified by these methods, pockets were also selected based on the experimentally characterized binding site residues in each protein in the proteome or in their homologues. This was done by fetching entries from the database using respective general feature format files obtained from UniProt database ([@bav060-B44]). Possible binding sites were identified by scanning each protein sequence in the proteome with known binding motifs from the Prosite ([@bav060-B45]) database to make sure they were not missed out by other methods in the workflow ([@bav060-B46]). The binding sites detected can be viewed using Jmol plugin and also co-ordinates of these binding pockets can be downloaded in pdb format.\n\n### Drug binding site database and comparison.\n\nDrugBank ([@bav060-B47]) and DrugPort were used to prepare a combined list of drugs or drug-like compounds; these included approved and experimental drugs and nutraceuticals. XML data files were obtained from these two databases and later parsed to extract information on proteins complexed with any of these drugs present in PDB. The binding sites were then extracted from these complexes. Residues of all atoms that lie within 4.5\u00c5 of any atom in the drug molecule were extracted as part of the binding site. Ten thousand six hundred and fifty-eight (from Drugbank) ([Supplementary Table S2](http://database.oxfordjournals.org/lookup/suppl/doi:10.1093/database/bav060/-/DC1))\u2009+\u20092516 (from Drugport) ([Supplementary Table S3](http://database.oxfordjournals.org/lookup/suppl/doi:10.1093/database/bav060/-/DC1)) drug-binding sites were obtained from PDB through this process. High-confidence targets from Mtb were scanned using these known drug-binding sites, and also drug-binding sites were scanned for similarities against different binding site clusters.\n\n### Structural interactome.\n\nThe structural interactomics database CREDO ([@bav060-B16]) provides details of pairwise atomic interactions of intermolecular and intramolecular contacts between ligands and macromolecule for the structures in PDB. The PDB codes in the database are linked to the results of CREDO. This database stores interaction between atoms as structural interaction fingerprints as implemented by Deng *et\u00a0al*. ([@bav060-B48]). Thirteen different interaction types such as hydrogen bonds, halogen bonds, carbonyl interactions and more are currently implemented in CREDO. Polypeptide-residue mapping is done onto UniProt. This allows identification of modified, non-standard or mutated proteins in the PDB compared with sequence in UniProt. Further, small-molecule and protein interaction details are provided in the database. Physico-chemical properties are calculated for all the small molecules in PDB and these properties are important for evaluating its drug-likeness. Topological similarities of the small molecules based on 2D and 3D descriptors are also retrieved from the database. With these data, CREDO provides major structural interaction details to study small-molecule binding properties. The PDB structures used as templates for building models in SInCRe are linked to the CREDO database.\n\n### Structure binding molecules.\n\nTIMBAL ([@bav060-B17]), a database of small molecules disrupting protein--protein interactions, provides us with a list of small molecules relevant to the proteins of *Mtb*. Previously constructed by manual curation, now TIMBAL is automated to identify a list of protein--protein interaction modulators. The PPI targets and their orthologs are identified by UniProt identifiers. Small molecules related to these proteins are searched using UniProt identifiers in ChEMBL database. The homologues of known protein--protein interactions to the proteins in Mtb are identified and corresponding small molecules are listed. Totally 21 Mtb proteins are homologous to proteins in TIMBAL database corresponding to 11 targets.\n\n### Ligand-based off-target prediction and small-molecule data.\n\nThere are two main approaches to predict off-target activity. The structure-based approach relies on the similarity of the targets binding pockets, whereas the ligand-based approach connects targets based on the similarity of their ligands. The two methodologies complement each other ([@bav060-B49]). TIBLE (<http://mordred.bioc.cam.ac.uk/tible/>) collects small-molecule data (Minimal Inhibitory Concentration (MIC) for mycobacterium and binding to isolated Mtb targets) from the ChEMBL database ([@bav060-B50]) and the CDD ([@bav060-B51]). There are 75 Mtb targets with small-molecule binding data. For each of these targets, three independent algorithms---SEA ([@bav060-B52]), PharmMapper ([@bav060-B53]) and PASS ([@bav060-B54]) are used to derive off-target ligand-based predictions. Link from TIBLE to PharmMapper offers pharmacophore-matching platform for potential target identification. The details of small molecules and ligand-based off-target are integrated into the SInCRe database and also linked to the TIBLE page for detailed information.\n\nSystems-based target identification\n-----------------------------------\n\nIdentification of high confidence drug targets is a primary factor for efficient drug treatment. TargetTB ([@bav060-B18]), a comprehensive *in silico* target identification pipeline, was developed by one of the groups. The pipeline is built by incorporating network-based analysis of the protein--protein interactions, a flux-balance analysis of the reactome, phenotype-essentiality data derived from experiments, targetability assessment based on sequence and structure analysis using in-house novel algorithms. Initially proteins that are important for the survival of *Mtb* were identified using flux balance and network analyses. Subsequently comparative genomics with the host was carried out. Finally the viability of a protein to be a potential drug target was assessed using novel methods for structural analysis of binding sites. Further, expression-data analysis, providing correlation and non-similarity measures of target proteins to gut flora proteins and also to 'anti-target' proteins in the host, was analysed extensively. Four hundred and fifty-one high-confidence entries were identified by this analysis pipeline. These short-listed targets have been further analysed through phylogenetic profiling against 228 pathogen genomes to identify antibiotic targets of broad spectrum especially those specific to TB. Target proteins significant to mycobacterial persistence and drug resistance mechanisms have also been analysed and reported. The details of the targets identified through TargetTB pipeline has been integrated into this database.\n\nOther resources\n---------------\n\nExternal data from STRING ([@bav060-B19]), a database of known and predicted protein--protein interactions, STITCH ([@bav060-B20]), a database of protein--small molecule interactions and Tuberculist ([@bav060-B6]) for primary details about each Mtb protein are integrated into the SInCRe database.\n\nCoverage of the *M. tuberculosis* proteome in the database\n==========================================================\n\nOur analysis of the repertoire of *M. tuberculosis* proteins, using a multitude of sensitive techniques, has generated a resource of information including structural and functional domain assignments, potential drug-targets and small-molecule binders including FDA-approved drugs. [Figure 2](#bav060-F2){ref-type=\"fig\"} summarizes the percentage coverage achieved for *M. tuberculosis* proteins and indicates that 3495 of 4018 proteins could be associated with at least one functional domain (Pfam domain) assignment while 3131 proteins could either be associated with structural domains (SCOP domains) or with proteins of known structure. In terms of domain assignment alone, a total of 3566 proteins (89%) could be associated with at least one structural or functional domain. Due to the combined use of sensitive profile-based techniques, the percentage of *M. tuberculosis* proteins associated with functional domains is 3% higher than the annotations available in databases such as Pfam; and the percentage coverage achieved in terms of structural domains is 8% higher than the structural annotations available in databases such as SUPERFAMILY. Figure 2.Percentage coverage of *M. tuberculosis* proteins in the database. Numbers in brackets denote absolute values.\n\nSystematic means to identify potential drug targets in *M. tuberculosis* has resulted in recognition of 498 high-confidence targets, constituting 12% of the proteome. The SInCRe database also includes information on protein--protein interactions within *M. tuberculosis* as documented in resources such as STRING. Approximately 23 000 known or predicted protein--protein interactions in *M. tuberculosis* are mediated by 2386 (59%) proteins.\n\nOur attempt to integrate information from diverse resources provides a unified platform to explore and investigate the usefulness of a predicted target or a small molecule in the context of drug development and drug discovery for TB.\n\nDatabase and web interface\n==========================\n\nThe SInCRe database is created by integrating resources from various other databases for 4018 Mtb proteins. This database has been developed on the Linux-Apache-MySQL-PHP platform. Sequence- and structure-level datasets have been stored in efficiently designed relational database schema. The web interface is developed using BootStrap (<http://twitter.github.com/bootstrap>). This provides cascading style sheets framework and javascript functionality. CytoscapeWeb ([@bav060-B55]), a java plugin, is used for interactive display of protein--protein and protein--small molecules interaction networks. Protein structures are represented in 3D using JSmol, a JavaScript-based molecular viewer from Jmol, an open-source Java viewer for chemical structures in 3D (<http://www.jmol.org/>). The modelled structures and sequences can be downloaded in PDB and FASTA formats, respectively. The tables in webpages are sortable and searchable, giving the user ease of acquiring data of interest.\n\nThe database can be queried using Rv IDs, gene name, UniProt ID, Pfam ID and Tuberculist functional classification. The dataset can be browsed for information available based on a few methods for limited list of Rv IDs.\n\nPrediction of drug interactions using SInCRe\n============================================\n\nProtein kinases and phosphatases constitute important classes of drug targets due to the critical roles played by them in reversible protein phosphorylation that regulates many biological processes. There are many studies that report the development of potent inhibitors for these enzymes involved in protein phosphorylation to treat different types of cancer and autoimmune diseases ([@bav060-B56]). Serine/threonine protein kinases (STPKs) are one such class of kinases that specifically phosphorylate the hydroxyl group of one or more serine and threonine residues in the substrate protein. *Mycobacteriumtuberculosis* (Mtb) genome houses 11 of such STPK genes and all of these are known to regulate crucial signalling processes, playing an important role in regulating physiology and virulence of the pathogen ([@bav060-B57]).\n\nOf the 11 STPKs in Mtb, nine (PknA, PknB, PknD, PknE, PknF, PknH, PknI, PknJ and PknL) are receptors containing a transmembrane helix with extracellular sensory domain and intracellular kinase domain, thus acting as signal transducers. The other two kinases (PknG and PknK) are cytoplasmic containing a regulatory domain and could hence play a role in intracellular responses. Here, we explore the role of one such STPK -- PknD (Rv0931c), as a putative drug target through the information present in SInCRe database. PknD acts a receptor kinase with extracellular sensory domain adopting a six-bladed \u03b2 propeller structure (PDB ID: 1RWL, 1RWI) ([@bav060-B58]), and an intracellular kinase domain. The 3D structure of intracellular kinase domain could be derived using homology modelling using the crystal structure of PknE (PDB ID: 2H34) kinase domain as the template which share 59.7% sequence identity with the target. Although the substrate and the ligand for the PknD is yet to be discovered, the gene neighbourhood analysis reveals that it could play an essential role in phosphate transport. This is complemented by the fact that the growth of \u0394*pknD* strain is compromised in a phosphate deficient medium ([@bav060-B59]). Recently, PknD has been observed to phosphorylate the N-terminal domain of Rv0516c, a putative regulator of sigma factor SigF ([@bav060-B60]). These three genes---PknD, Rv0516c and SigF play an important role in osmosensory signalling pathway ([@bav060-B61]). Moreover, a screen for identifying important genes for central nervous system infection by Mtb also identified PknD to be essential as \u0394*pknD* strain was observed to be defective for invasion of central nervous system ([@bav060-B62]).\n\nThe binding site prediction exercise carried out on a proteome-scale involving a consensus of different types of algorithm ([@bav060-B46]) identified a putative binding site present at the interface of N-terminal and C-terminal lobe of kinase domain in PknD ([Figure 3](#bav060-F3){ref-type=\"fig\"}A). A systematic binding site comparison of this predicted pocket against a database of approved drug-binding sites yielded nilotinib (NIL) binding site from human mitogen activated protein kinase 11 protein (PDB ID: 3GP0) as the topmost hit with binding site similarity score (PMAX) ([@bav060-B13]) of 0.703. A binding site alignment of the predicted pocket with this known NIL binding site using PocketAlign algorithm ([@bav060-B14]) reveals the observed similarity and the differences in the binding sites ([Figure 3](#bav060-F3){ref-type=\"fig\"}B). Although the similarity of these protein kinases with the human counterparts can increase the risk of toxicity, there are supporting evidences in the literature that have successfully exploited the ATP-binding sites to achieve the selectivity. There are FDA-approved drugs that selectively bind to active and inactive conformations of the protein kinases to achieve the selectivity ([@bav060-B56]). The differences in kinase inhibitor binding sites (depicted as wireframe in [Figure 3](#bav060-F3){ref-type=\"fig\"}B) could be used as anchor points in fragment-based drug discovery to achieve the selectivity towards Mtb protein kinases. Interestingly, the binding sites of many of the anti-retroviral protease inhibitors like nelfinavir and lopinavir were also observed to have similarity to the predicted binding site in PknD. These observations are supported by the fact that nelfinavir is found to have anti-cancerous property attributed to its ability to weakly inhibit multiple protein kinases ([@bav060-B63]). One such anti-retroviral protease inhibitor---saquinavir (Ligand code: ROC), having high binding site similarity with the predicted binding site in PknD was explored further through computational docking using AutoDock Vina ([Figure 3](#bav060-F3){ref-type=\"fig\"}C) ([@bav060-B64]). The computationally predicted binding affinity (\u22128.1 kcal/mol) was found to be comparable to the native saquinavir complexed with HIV-protease (\u22129.4 kcal/mol). The best pose obtained through computational docking predicted the residues---ARG101, GLU142, ARG93 and GLU31 present in the predicted binding site to have crucial interaction with the saquinavir. These interesting drug associations can be readily obtained from the 'protein--small molecule associations' tab presented in the SInCRe database. The SInCRe database can thus, be used to generate readily testable hypothesis for anti-tubercular drug discovery. Figure 3.Example for predicted drug interactions using SInCRe. (**A**) A predicted binding site for PknD, a STPK, is depicted in the form of spacefill. (**B**) The alignment of predicted binding site from PknD (Rv0931c, in red) with the NIL binding site from Human Mitogen Activated Protein Kinase (PDB ID: 3GP0). The corresponding residues are highlighted in sticks, whereas unique residues with no correspondences are represented as wireframe. These distinguishing residues can be targeted to achieve the selectivity. (**C**) The best pose derived from computational docking depicting the interaction of saquinavir (ROC, shown as green ball and stick model) with the residues (represented as sticks) of the predicted binding site in PknD.\n\nConclusion\n==========\n\nSInCRe is an integrated suite of databases that provides the outcome of extensive sequence and structural studies of Mtb proteins. Sequence-based domain assignment and structural analysis of binding sites act as a resource to help in the identification off-target interactions of drug molecules, knowledge of which is useful in the design of novel drugs for *M.tuberculosis.* Future updates will include incorporation of other resources from Cambridge and Bangalore.\n\nSupplementary Data\n==================\n\n[Supplementary data](http://database.oxfordjournals.org/lookup/suppl/doi:10.1093/database/bav060/-/DC1) are available at *Database* Online.\n\n###### Supplementary Data\n\nAuthors thank the two reviewers of this article, Samir Brahmachari, Anshu Bharadwaj and other colleagues from Council for Scientific and Industrial Research (CSIR), India for valuable comments and suggestions. They also thank Sumanta Mukherjee for his inputs on developing the database. Authors also thank Harry Jubb, John Overington, Yvonne Light, Sean Ekins, Xiaofeng Liu and John Irwin with their help in the development of the TIBLE resource.\n\nFunding\n=======\n\nThis research is supported by Open Source Drug Discovery (OSDD) program of CSIR, India as well as by the Department of Biotechnology and Mathematical Biology Program sponsored. by Department of Science and Technology. N.S. is a J.C. Bose National Fellow. Funding for open access charge: National Centre for Biological Sciences (NCBS) to R. Sowdhamini.\n\n*Conflict of interest*. None declared.\n\n[^1]: Citation details: Metri,R., Hariharaputran,S., Ramakrishnan,G., *et\u00a0al*. SInCRe---structural interactome computational resource for *Mycobacterium tuberculosis*. *Database* (2015) Vol. 2015: article ID bav060; doi:10.1093/database/bav060\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nVasovagal syncope (VVS) is characterized by a transient, self-limited loss of consciousness: loss of postural tone, collapse, and spontaneous recovery \\[[@pone.0163280.ref001]\\]. Between 20--50% of adults will experience at least one syncopal episode during their lifetime, and VVS is most common in the elderly with up to 23% of individuals over 70 years old affected and a 30% chance of reoccurrence \\[[@pone.0163280.ref002]\\]. Each year over 400,000 new syncope patients are diagnosed with VVS, of which 2--5% require emergency room visits, placing an annual burden of \\$2.4 billion on U.S. healthcare \\[[@pone.0163280.ref003], [@pone.0163280.ref004]\\].\n\nThe mechanism underlying VVS is not fully understood, yet the current thinking is decreased venous return to the heart causes vigorous contraction of the myocardium against inadequately filled atria, triggering the Bezold-Jarisch reflex resulting in paradoxical hypotension, bradycardia \\[[@pone.0163280.ref001], [@pone.0163280.ref005]\\], and consequent decreased cerebral perfusion, leading to a loss of consciousness \\[[@pone.0163280.ref005]\\]. The head-up tilt test is valuable in diagnosing VVS and understanding its mechanism. Head-up tilt tests on patients (mean age: 42.5 years) have shown that sympathetic nerve activity and myocardial contractility are reduced preceding the onset of syncope, followed by a lowering of blood pressure \\[[@pone.0163280.ref006], [@pone.0163280.ref007]\\].\n\nIn a series of compelling experiments, Cohen and colleagues proposed a rat model of VVS using sinusoidal galvanic vestibular stimulation (sGVS) \\[[@pone.0163280.ref008]--[@pone.0163280.ref010]\\]. Their groups found that sGVS could produce temporary hypotension and bradycardia. These studies imply that sGVS in rats mimics human vasovagal responses through activation of the otolith system \\[[@pone.0163280.ref009]\\]. While Cohen and colleagues have provided considerable research in understanding the role of sGVS on precipitating the vasovagal response, a critical aspect, and the final sequela, of VVS has not been explored: cerebral perfusion.\n\nDespite VVS pathophysiology including the reductions in blood pressure and heart rate, VVS culminates in loss of conscious, typically leading to a fall \\[[@pone.0163280.ref011]\\]. Although the consensus is that cerebral hypoperfusion ultimately causes the fainting episode, currently there is limited research and understanding of the role cerebral blood flow (CBF) plays in VVS pathophysiology. To date, few clinical studies have been published measuring CBF before and during syncope. In one study by Grubb *et al*., thirty syncope patients were subjected to the head-up tilt test. During syncope, the mean CBF was observed to be 46 \u00b1 17% lower than Baseline values \\[[@pone.0163280.ref012]\\]. A second study evaluated CBF changes in syncope patients with no bradycardia or hypotension during the episode. In these patients, the mean CBF was still markedly reduced during syncope (26 \u00b1 13% lower than Baseline values) \\[[@pone.0163280.ref013]\\]. Sung *et al*. also investigated CBF changes during syncope in patients and found that standing or head-up tilt test induced syncopal-related lowering of the mean CBF by 32 \u00b1 11.6% (standing) and 30 \u00b1 16.7% (head-up tilt test) \\[[@pone.0163280.ref014]\\]. These studies, as well as the VVS paradigm, consider cerebral hypoperfusion as the primary sequelae of VVS. Therefore, an animal model of VVS must show reduction in cerebral blood flow during syncope. Additionally, VVS patients are plagued by spontaneous loss of consciousness. Therefore, a second key to an animal model of VVS is that behavioral changes mimicking VVS symptoms are experienced by awake animals.\n\nHerein we utilize the sGVS model in rats to examine its impact on CBF and, in awake animals, behavior changes caused by sGVS. Our first hypothesis is sGVS in rats results in decreased blood pressure and heart rate (HR), followed by reduced CBF, mimicking vasovagal response in humans. Our second hypothesis is sGVS in awake animals will induce VVS-like symptoms, namely fatigue followed by spontaneous fainting and recovery. Our third hypothesis is isoflurane preconditioning will induce cardio- and neuro-protective mechanisms, thereby attenuating the effects of sGVS on blood pressure, HR, and CBF.\n\nMaterials and Methods {#sec002}\n=====================\n\nAll experiments were carried out in accordance to the methods and procedures approved by the Loma Linda University IACUC and Research Protection Programs, and conducted in compliance with the *NIH Guidelines for the Use of Animals in Neuroscience Research*. A protocol was in place for early termination of experiments if animals became severely ill (determined by the Loma Linda University veterinary staff).\n\nAdult male Sprague-Dawley rats (280--320 g) were obtained from Harlan Laboratories (Harlan Laboratories, Indianapolis, IN, USA). Animals were housed in a humidity and temperature controlled room on a 12 hour light-dark cycle, and given food and water *ad libitum*. During all surgical procedures and methods during which rats were anesthetized, the body temperature was maintained at 37 \u00b1 0.5\u00b0C using a heating pad controlled using a rectal probe. Animals were continuously monitored during all surgical procedures and methods. Animals were randomly assigned to the groups in each experiment. After surgical procedures, betadine was applied to the wound(s), and buprenorphine was administered subcutaneously (0.01 mg/kg). At the end of the experiment, rats were euthanized via decapitation after deeply anesthetized with 5% isoflurane. No animals died before the experimental endpoints were achieved.\n\nSixty-six rats were used in this study. Four animals were used in a pre-experiment to determine the optimal stimulation parameters for lowering mean arterial pressure (MAP) and HR, while depressing CBF. Experiment 1 investigated the effect of sGVS on CBF (groups: Sham and sGVS, n = 10/group). Experiment 2 examined the effect of sGVS in awake, freely moving rats (groups: Sham and sGVS, n = 6/group). Experiment 3 evaluated isoflurane preconditioning as a potential therapy to prevent sGVS sequelae (groups: Sham, sGVS, and sGVS + isoflurane preconditioning, n = 10/group). Sham animals were normal rats which underwent all surgeries and electrode placement, had the MAP, HR, and CBF measured, but without ever undergoing sGVS stimulation. Additional methodological details are available in [S1 File](#pone.0163280.s002){ref-type=\"supplementary-material\"}.\n\nsGVS {#sec003}\n----\n\nsGVS was induced using a computer-controlled stimulator (Grass Technologies, West Warwick, RI, USA) which generates sinusoidal currents and applied via electrodes \\[[@pone.0163280.ref008]--[@pone.0163280.ref010]\\]. After laser Doppler probe placement, two Ag/AgCl needle electrodes were inserted into the skin over the mastoids. The computer-controlled stimulator generated sinusoidal currents binaurally (2--4 mA, 0.025--0.5 Hz). Rats were kept under anesthesia for a total of 67 minutes: 30 minutes for surgeries (femoral artery catheterization, laser Doppler probe placement, and electrode placement), during measurement of the baseline (for MAP, HR, and CBF, 4 minutes), during sGVS stimulation (3 minutes), and for 30 minutes post-stimulation.\n\nPre-Experiment: Determine the Optimal Parameters for Inducing sGVS {#sec004}\n------------------------------------------------------------------\n\nIn four rats, to determine the optimal stimulation parameters for inducing MAP and HR reductions, various combinations of amplitude and frequencies were tested: 2 or 4 mA at 0.025--0.5 Hz. The optimal stimulation parameters were determined to be the frequency and amplitude which induced the greatest, most reproducible drop in MAP and HR among each of the four animals. CBF was monitored in each rat to identify any CBF changes during stimulation using the various sets of stimulation parameters. These four animals were not used in any other aspect of this study. Rats were deeply anesthetized with 5% isoflurane and then decapitated after completing the experiment.\n\nExperiment 1: Effect of sGVS on Cerebral Blood Flow {#sec005}\n---------------------------------------------------\n\nTwenty animals were randomly assigned to either the Sham or sGVS group (n = 10/group). Sham animals were normal rats which underwent all surgeries (burr hole, femoral artery catheterization) and electrode placement, had the MAP, HR, and CBF measured, but without ever undergoing sGVS stimulation. These animals were anesthetized with isoflurane for the same duration as animals subjected to sGVS and were monitored for MAP, HR, and CBF changes for the same duration as sGVS animals (4 minutes of \"Baseline\", 3 minutes of \"Stimulation\" without electrical stimulation, and 30 minutes \"Post-Stimulation\"). Sham animals were used as the control group.\n\nOne day before sGVS, rats were anesthetized and a burr hole was drilled in the skull (center: 5 mm proximal to the coronal suture, 4 mm right lateral of the sagittal suture) using a microdrill. Then bone wax was applied and the skin sutured. Betadine was applied to the wound and buprenorphine was administered subcutaneously (0.01 mg/kg). The animal was allowed to recover before being returned to its home cage.\n\nOn the day of sGVS, a PE50 catheter was inserted and advanced 10--15 mm into the femoral artery of an anesthetized rat. The catheter was connected to a transducer for continuous measurement of MAP and HR using a blood pressure analyzer (Digi-Med BPA 400a, Micro-Med, INC., Louisville, KY, USA) and the DMSI-400 software (Micro-Med, INC., Louisville, KY, USA). After femoral artery catheterization, the burr hole was reopened, exposing the brain tissue. A laser Doppler probe (OxyFlo probe, MNP100XP, AdInstruments Inc., Colorado Springs, CO, USA) was used for continuous measurement of CBF using a blood flow monitor (INI191, AdInstruments Inc., Colorado Springs, CO, USA) and PowerLab (PL3504, AdInstruments Inc., Colorado Springs, CO, USA) with the LabChart software (LabChart Pro, AdInstruments Inc., Colorado Springs, CO, USA).\n\nsGVS was induced for three minutes with a 4 mA amplitude and 0.025 Hz frequency. MAP, HR, and CBF were continuously monitored for four minutes before sGVS, during stimulation (3 minutes), and for 30 minutes post-sGVS. Thirty minutes after stimulation, the laser Doppler probe was removed and the rats were deeply anesthetized with 5% isoflurane and then decapitated.\n\nExperiment 2: Effect of sGVS on Awake, Freely Moving Animals {#sec006}\n------------------------------------------------------------\n\nTwelve animals were used to examine the effects of sGVS on awake animals. Animals were randomly assigned to either the Sham or sGVS group (n = 6/group). Sham animals had electrodes placed, but received no stimulation. MAP, HR, and CBF were not monitored in any animal for experiment 2. The only surgical procedure performed was electrode placement.\n\nElectrodes were placed while the animal was under anesthesia. Total duration of isoflurane exposure during electrode placement was not longer than 10 minutes (range: 6--10 minutes). Animals were allowed to recover from the effects of isoflurane for a minimum of 50 minutes (range: 50--70 minutes) in a cage identical to its home cage without bedding, food, and water. After the animal fully recovered, sGVS was induced for 3 minutes. The animal's behavior was recorded for 5 minutes before stimulation, during stimulation (3 minutes) and for 60 minutes post-stimulation. Immediately after recording stopped (60 minutes post-stimulation), animals were deeply anesthetized and decapitated.\n\nExperiment 3: Effect of Isoflurane Preconditioning on sGVS {#sec007}\n----------------------------------------------------------\n\nThirty animals were randomly assigned to either the Sham, sGVS, or isoflurane preconditioning followed by sGVS group (n = 10/group). Sham animals served as the control group and underwent all procedures as described in Experiment 1. All animals receiving sGVS (sGVS only group and isoflurane preconditioning followed by sGVS group) underwent all surgical procedures as described in Experiment 1.\n\nTen animals were subjected to five days of isoflurane preconditioning. Daily, isoflurane (2.5%) was administered to animals for 90 minutes. The isoflurane preconditioning regimen was completed five days before subjecting rats to sGVS. Four days after completing the isoflurane preconditioning regimen, rats underwent the surgical procedure for creating the burr hole (see [Experiment 1](#sec005){ref-type=\"sec\"} and [S1 File](#pone.0163280.s002){ref-type=\"supplementary-material\"} for details). One day later (five days after completing the isoflurane preconditioning regimen), rats were subjected to femoral artery catheterization, laser Doppler probe placement, and sGVS as described in Experiment 1. MAP, HR, and CBF were monitored before, during, and for 30 minutes post-stimulation as described in Experiment 1. Animals were anesthetized before, during, and after stimulation. All animals were deeply anesthetized and decapitated 30 minutes after stimulation (immediately after recording was completed).\n\nData Processing and Statistical Analysis {#sec008}\n----------------------------------------\n\nAll raw data was processed and analyzed by a blinded investigator. The data was divided into three sections: Baseline (minutes 0--4), Stimulation (minutes 3--7), and Post-Stimulation (minutes 7--37). Within each section the data was averaged and the standard deviation (SD) calculated. Upon processing, the Post-Stimulation data had very large SD due to the variation in the return to baseline values, thus the Post-Stimulation section was subdivided into 0--5 Min, 5--10 Min, 10--20 Min, and 20--30 Min Post-Stimulation. Since there was significant variation between animals (within the same group) in the raw Baseline values for the physiological parameters measured, the change from Baseline values were computed for each section/subsection. The data is presented as the mean (% change from Baseline) \u00b1 SD. In Experiment 2, the behavioral changes for animals during and post-sGVS are anecdotally described since functional tests specifically related to VVS symptoms were lacking. Finally, the data was processed to determine the amount of time after beginning stimulation that was required to observe a rapid, sustained drop in CBF (CBF drop delay). The local minimums for the data obtained during stimulation were identified. Each local minimum was examined for 1) being greater than 3 SD change from Baseline, and 2) sustained depression compared to Baseline (more than 15 seconds). The delay in CBF drop was the amount of time after beginning stimulation that the minimum occurred. Data was analyzed using two-way ANOVA with Sidak's *post-hoc* test (GraphPad Prism 6, La Jolla, CA). p\\<0.05 was considered statistically significant; p\\<0.1 was considered as a tendency.\n\nResults {#sec009}\n=======\n\nIn the pre-experiment, four rats were used to determine the optimal stimulation parameters required to induce reproducible vasovagal-like response: reduction in blood pressure and heart rate. Each animal was subjected to sGVS using the following stimulation parameters: current amplitude was either 2 or 4 mA, frequencies were 0.025, 0.05, 0.1, or 0.5 Hz. The optimal stimulation parameters for inducing cardiovascular depression in rats was found to be a 4 mA current at 0.025 Hz ([Fig 1II](#pone.0163280.g001){ref-type=\"fig\"}). This set of stimulation parameters induced depression of MAP and HR in each animal. Monitoring of the cerebral blood flow indicated that this set of stimulation parameters was also capable for causing a reduction in cerebral blood flow, as shown by decreased cerebral perfusion units measured using laser Doppler flow. Other sets of parameters were observed to produce MAP and HR depression in some animals but not others. For example, the representative plot ([Fig 1](#pone.0163280.g001){ref-type=\"fig\"}) shows that the 4 mA at 0.05 Hz, 4 mA at 0.1 Hz and 4 mA at 0.5 Hz currents also induced marked decreases in MAP and HR (IV, VI, and VIII, respectively). However, these sets of stimulation parameters did not cause reductions in MAP and HR in some animals. Thus, these sets of stimulation parameters were unable to cause reproducible lowering of MAP and HR. Additionally, the 4 mA at 0.1 Hz and 4 mA at 0.5 Hz currents did not have any consequent CBF reductions (VI and VIII, respectively) despite the marked lowering of MAP and HR. Based on the preliminary experiment, a 4 mA current at a frequency of 0.025 Hz was used in Experiments 1--3.\n\n![Effects of Various Sets of Stimulation Parameters on HR, MAP, and CBF.\\\nA representative plot of the stimulator current (mA), heart rate (BPM), mean arterial pressure (mmHg), and cerebral blood flow (perfusion units) is shown. Baseline data was collected for 4 minutes before beginning stimulation (minutes 0--4). Five minutes of recovery was allowed after each stimulation event. Stimulation occurred for 3 minutes using eight different sets of stimulation parameters: 2 mA at 0.025 Hz (I), 4 mA at 0.025 Hz (II), 2 mA at 0.05 Hz (III), 4 mA at 0.05 Hz (IV), 2 mA at 0.1 Hz (V), 4 mA at 0.1 Hz (VI), 2 mA at 0.5 Hz (VII), and 4 mA at 0.5 Hz (VIII). Red lines highlight the start and stop of each stimulation event. The greatest drop in CBF (13.9%) occurred for the 4 mA, 0.025 Hz current (II).](pone.0163280.g001){#pone.0163280.g001}\n\nNo mortality was observed in this study. A total of seven animals were removed from the study since no vasovagal-like response was observed (*i*.*e*. lack of a decrease in MAP or HR combined by a lack of oscillatory response of MAP or HR): two from the sGVS group in Experiment 1, two from the sGVS group in Experiment 3, and three from the sGVS + Isoflurane group in Experiment 3. In total, data was collected and analyzed for 18 animals in Experiment 1 (n = 10 for Sham, n = 8 for sGVS), 12 animals in Experiment 2 (n = 6 for Sham, n = 6 for sGVS), and 25 animals in Experiment 3 (n = 10 for Sham, n = 8 for sGVS only, and n = 7 for sGVS + Isoflurane Preconditioning).\n\nExperiment 1 --sGVS Causes Decreased Cerebral Blood Flow {#sec010}\n--------------------------------------------------------\n\nsGVS in rats causes a delayed but significant drop in CBF ([Fig 2](#pone.0163280.g002){ref-type=\"fig\"}). Compared to sham animals, sGVS causes a decrease in HR (p\\<0.05) and CBF (p\\<0.05), and a tendency to decrease mean arterial pressure (MAP) (p\\<0.1) ([Fig 3](#pone.0163280.g003){ref-type=\"fig\"}). Upon halting sGVS, MAP rapidly returned to a level indistinguishable from that of Sham (p\\>0.1). However, HR remained depressed compared to that of Sham for 10 minutes post-sGVS (p\\<0.05 for 0--5 and 5--10 Min Post-Stimulation). CBF was significantly lowered compared to that of Sham up to 30 minutes post-sGVS (p\\<0.05 for each time range).\n\n![sGVS Causes CBF Reduction.\\\nRepresentative plot of CBF (bottom, blue plot) changes during sGVS (top, red plot). Baseline CBF (\\~1780 perfusion units) was collected for 4 minutes before beginning sGVS (at minute 4). Approximately 1 minute after starting sGVS a significant drop in CBF was observed, which was maintained throughout stimulation. After sGVS, CBF recovers, but does not return to baseline values.](pone.0163280.g002){#pone.0163280.g002}\n\n![MAP, HR, and CBF Changes Induced by sGVS.\\\n**(A)** MAP tends to decrease during sGVS (p\\<0.1). After sGVS, MAP recovers. Prolonged anesthesia causes a steady MAP drop in all rats (Sham: p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation, p\\<0.1 for 0--5 Min Post-Stimulation vs 20--30 Min Post-Stimulation; sGVS: p\\<0.1 between Baseline and 20--30 Min Post-Stimulation). **(B)** HR is significantly reduced for sGVS rats compared to Sham during sGVS (p\\<0.05), and for up to 10 Min Post-Stimulation (p\\<0.05). Prolonged anesthesia causes steady decline of HR in all rats (Sham: p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation, p\\<0.05 for 0--5 and 5--10 Min Post-Stimulation vs 10--20 and 20--30 Min Post-Stimulation; sGVS: p\\<0.05 for Baseline vs each time range post-sGVS, p\\<0.05 for 0--5 and 5--10 Min Post-Stimulation vs 20--30 Min Post-Stimulation). **(C)** CBF is significantly reduced during sGVS compared to Sham (p\\<0.05) and remains depressed for up to 30 minutes post-sGVS (p\\<0.05). Prolonged anesthesia does not significantly lower CBF in either sham or sGVS rats. \\* p\\<0.05 for Sham vs sGVS for the given time variable.](pone.0163280.g003){#pone.0163280.g003}\n\nA tendency to decrease over time was observed in Sham and sGVS rats for MAP (Sham: p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation; sGVS: p\\<0.1 for Baseline vs 10--20 and 20--30 Min Post-Stimulation) and HR (Sham: p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation; sGVS: p\\<0.05 for Baseline vs 0--5, 10--20, and 20--30 Min Post-Stimulation). No time-dependent changes in CBF were observed for Sham. Other than the marked reduction in CBF caused by sGVS, the CBF in rats subjected to sGVS was also unaffected over time.\n\nExperiment 2 --sGVS in Awake Animals Mimics Vasovagal Syncope Symptoms {#sec011}\n----------------------------------------------------------------------\n\nAwake, freely moving animals were subjected to sGVS to evaluate the effect of sGVS on behavior. Before beginning sGVS, all animals exhibited normal behavior (balance, movements, walking, body tone). Sham animals continued to have normal behavior throughout the duration of electrode implantation (1.5--2 hrs). At the onset of sGVS, rats' behaviors began to change ([Fig 4](#pone.0163280.g004){ref-type=\"fig\"}). First, the breathing rate began to fluctuate (slow and deep or shallow). Within 30 seconds, animals exhibited fatigue-like behavior, such as labored breathing and reduced responsiveness to external stimuli (sound and light). Between 0.5 and 2 minutes (and for the remainder duration of stimulation), rats began to show signs of abnormal movement; use of limbs was slower, head swaying and/or head rotation, body swaying and/or leaning towards one side. During this time, when animals walked, it was unbalanced/uncoordinated. During sGVS, 2 animals had spastic movements and 2 animals had fainting-like behavior. The latter two rats fell on their side, then had 1--2 seconds of either no movement or one limb spasm, then recovered spontaneously. After stopping stimulation, the animals subjected to sGVS exhibited fatigue-like behavior for up to 45 minutes after sGVS. Animals subjected to sGVS recovered normal behavior 28 \u00b1 9.5 minutes post-stimulation (range: 15--45 minutes).\n\n![Behavioral Changes during sGVS in Awake Rats.\\\nRepresentative images of the changes in behavior observed during sGVS in awake rats. Sinusoidal galvanic vestibular stimulation in the awake animal induces similar symptoms as that experienced by VVS patients. Awake animals during stimulation exhibit signs of fatigue-like behavior (reduced responsiveness and lethargy), labored breathing, altered coordination, and even faint-like behavior (*i*.*e*. falling). Representative images from fainting-like behavior in two animals are shown (**A**, **B**) with pre-faint-like stance (**i**), stance during the faint-like behavior (**ii**), and spontaneous recovery within 1--2 seconds after the onset of the faint-like behavior (**iii**). **(C)** Representative images of altered coordination and head swaying.](pone.0163280.g004){#pone.0163280.g004}\n\nExperiment 3 --Isoflurane Preconditioning Reduces the Effect of sGVS Cerebral Blood Flow Depression {#sec012}\n---------------------------------------------------------------------------------------------------\n\n### Mean Arterial Pressure {#sec013}\n\nCompared to sham animals, sGVS decreased MAP in rats preconditioned with isoflurane (sGVS + Isoflurane Preconditioning (PC)) (p\\<0.05) and tended to decrease MAP in unconditioned animals (p\\<0.1) ([Fig 5A](#pone.0163280.g005){ref-type=\"fig\"}). During sGVS, no difference was observed for MAP between unconditioned and isoflurane preconditioned rats. After stopping sGVS, the MAP of animals subjected to sGVS only and those that were isoflurane preconditioned returned to levels indistinguishable from that of Sham. MAP for Sham and sGVS only animals tended to decrease over time (p\\<0.1 Baseline vs 20--30 Min Post-Stimulation), which was not observed Post-sGVS in isoflurane preconditioned rats.\n\n![Effects of Isoflurane Preconditioning on MAP, HR, and CBF Changes During and Post-sGVS.\\\n**(A)** MAP is significantly reduced by sGVS in unconditioned and isoflurane preconditioned rats subjected to sGVS compared to Sham (p\\<0.1 Sham vs sGVS, p\\<0.05 Sham vs sGVS + Isoflurane PC). MAP recovers immediately upon stopping sGVS, but prolonged anesthesia has a tendency to steadily decrease MAP (Sham: p\\<0.1 Baseline vs 20--30 Min Post-S-Stimulation; sGVS: p\\<0.1 Baseline vs 20--30 Min Post-Stimulation). **(B)** During sGVS, HR is significantly lowered by sGVS in unconditioned and isoflurane preconditioned rats compared to Sham (p\\<0.05 for Sham vs sGVS and sGVS + Isoflurane PC). HR in unconditioned animals remains decreased for up to 10 minutes post-sGVS, while the HR of isoflurane preconditioned rats recovers after sGVS stops. Prolonged anesthesia causes continued decreased in HR for sham and unconditioned sGVS animals (Sham: p\\<0.1 between Baseline and 20--30 Min Post-Stimulation; sGVS: p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation). **(C)** CBF is reduced by sGVS in unconditioned and isoflurane preconditioned animals compared to Sham (p\\<0.05 for Sham vs sGVS and sGVS + Isoflurane PC), yet isoflurane preconditioning attenuates the CBF reduction caused by sGVS (p\\<0.05 sGVS vs sGVS + Isoflurane PC). The CBF reduction caused by sGVS remains for up to 30 minutes post-stimulation (p\\<0.05 for Sham vs sGVS and sGVS + Isoflurane PC for each time frame Post-Stimulation). \\* p\\<0.05 for Sham vs sGVS for the given time variable, **\\#** p\\<0.05 for Sham vs sGVS + Isoflurane PC, **&** p\\<0.05 for sGVS vs sGVS + Isoflurane PC.](pone.0163280.g005){#pone.0163280.g005}\n\n### Heart Rate {#sec014}\n\nRats subjected to sGVS had reduced HR during stimulation compared to that of Sham for unconditioned and isoflurane preconditioned rats (p\\<0.05 for Sham vs sGVS and sGVS + Isoflurane PC) ([Fig 5B](#pone.0163280.g005){ref-type=\"fig\"}). Isoflurane preconditioning did not attenuate the impact of sGVS on HR. The HR of animals subjected to only sGVS remained depressed for up to 10 minutes post-sGVS whereas HR of sGVS animals preconditioned with isoflurane returned to values that were not statistically different from that of Sham immediately after stopping stimulation (p\\<0.05 Sham vs sGVS at 0--5 and 5--10 Min Post-Stimulation). Over time, the HR in Sham animals and unconditioned animals subjected to sGVS decreased (Sham: p\\<0.05 Baseline vs 20--30 Min Post-Stimulation; sGVS: p\\<0.1 for Baseline vs 0--5 and 5--10 Min Post-Stimulation, p\\<0.05 for Baseline vs 10--20 and 20--30 Min Post-Stimulation), whereas isoflurane preconditioned sGVS rats' HR did not have any tendency to decrease over time.\n\n### Cerebral Blood Flow {#sec015}\n\nDuring stimulation, CBF is significantly reduced in animals subjected to sGVS compared to that of sham animals (p\\<0.05 Sham vs sGVS), and this reduction is sustained for up to 30 minutes post-stimulation (p\\<0.05 Sham vs sGVS for each time range post-sGVS) ([Fig 5C](#pone.0163280.g005){ref-type=\"fig\"}). CBF of isoflurane preconditioned animals was also statistically lower than that of sham animals during stimulation (p\\<0.05 Sham vs sGVS + Isoflurane PC). In isoflurane preconditioned animals, the reduced CBF continued for up to 30 minutes after sGVS (p\\<0.05 Sham vs sGVS + Isoflurane PC for 0--5, 5--10, 10--20, and 20--30 Min Post-Stimulation). However, isoflurane preconditioned rats had a significantly attenuated CBF drop compared to unconditioned rats both during stimulation (p\\<0.05 sGVS vs sGVS + Isoflurane PC) and at each time range post-stimulation (p\\<0.05 sGVS vs sGVS + Isoflurane PC at 0--5, 5--10, 10--20, and 20--30 Min Post-Stimulation).\n\nDiscussion {#sec016}\n==========\n\nSinusoidal galvanic vestibular stimulation in rats provides an experimental mimetic of human vasovagal-like response \\[[@pone.0163280.ref008]--[@pone.0163280.ref010]\\]. Herein, sGVS induced cardiovascular and cerebrovascular depression in rats, as well as resulting in VVS-like symptoms in awake animals. The reduction in CBF during sGVS was partially attenuated with isoflurane preconditioning despite isoflurane preconditioning having no effect on sGVS-induced lowering of MAP or HR.\n\nThe sGVS model has been previously shown to cause lower MAP and HR in rats through activation of the vestibular neurons in the otolith organs \\[[@pone.0163280.ref009], [@pone.0163280.ref015]\\]. In the landmark experiments by Cohen and colleagues, sGVS in rats was established as an experimental model mimicking human vasovagal response and VVS \\[[@pone.0163280.ref008]--[@pone.0163280.ref010]\\]. In that series of experiments, sGVS was found to cause rapid lowering of MAP and HR at stimulus amplitudes of 2--4 mA and frequencies of 0.008--0.4 Hz \\[[@pone.0163280.ref008], [@pone.0163280.ref009]\\] with the most effective stimulation frequency being 0.025 Hz \\[[@pone.0163280.ref010]\\]. Herein, our findings support the findings that the optimal stimulation parameters for inducing vasovagal-like responses in rats is a 4 mA current at 0.025 Hz. Additionally, we found that not only is sGVS capable of producing the cardio-vascular depression that mimics human VVS, but sGVS also results in marked reduction of CBF. While HR and MAP may contribute to the onset of VVS-induced fainting, VVS pathophysiology culminates in decreased cerebral perfusion (*i*.*e*. results in loss of consciousness) \\[[@pone.0163280.ref005], [@pone.0163280.ref016]\\].\n\nSimilarities between the sGVS Rodent Model and Vasovagal Syncope in Patients {#sec017}\n----------------------------------------------------------------------------\n\n### VVS Phase 1: Progressive Early Hypotension {#sec018}\n\nIn VVS patients, there are four distinct phases of vasovagal syncope: progressive early hypotension, terminal vasodilatation, the syncopal episode, and the postfaint phase. The first phase, termed \"progressive early hypotension,\" occurs for patients, during head-up tilt test, who are destined to develop syncope \\[[@pone.0163280.ref005]\\]. This phase is characterized by a slow but continuous lowering of blood pressure and cardiac output, as well as an increase in HR \\[[@pone.0163280.ref005]\\]. This phase may last up to 5 minutes. In rats, 7 of 16 animals subjected to sGVS showed a brief (less than 20 seconds) rise in HR, with only 3 of those 7 animals showing a slow decline in BP during that period. Four of the 7 animals which showed a rise in HR immediately following the start of stimulation had rapid lowering of blood pressure. Further studies need to be performed to determine if these observations in fact correlate with human VVS by measuring cardiac output \\[[@pone.0163280.ref005]\\].\n\n### VVS Phase 2: Terminal Vasodilatation {#sec019}\n\nThe second phase of VVS, termed \"terminal vasodilatation,\" occurs during the last 2 minutes before syncope \\[[@pone.0163280.ref005]\\]. The characteristics of this phase include a rapid reduction in blood pressure and a slowing of HR, with the decrease in HR occurring after the rapid drop in blood pressure \\[[@pone.0163280.ref005]\\]. While in humans the sequence of physiological changes leading to syncope may be induced by a number of factors \\[[@pone.0163280.ref005], [@pone.0163280.ref017]\\], sGVS in rodents more closely mimics this phase (typically skipping the progressive early hypotension phase). In the majority of animals, MAP rapidly decreases after stimulation is started \\[[@pone.0163280.ref008]--[@pone.0163280.ref010]\\], followed by slow decline in HR. A few animals (4 of 16) had a sharp decline in HR immediately following the blood pressure drop.\n\nAlso observed in humans during the second phase of VVS is a slow decline in CBF which precedes a rapid drop immediately before the syncopal episode \\[[@pone.0163280.ref005]\\]. In 11 of 16 rats, CBF did not significantly change from Baseline before the sharp drop occurred. Only 5 of 16 animals experienced a slow gradual decline in CBF preceding a rapid drop. In these five animals, the decline in CBF (before the drop-off) occurred for 21 \u00b1 5.6 seconds, with one exception in which the gradual decrease in CBF happened over 120 seconds before the rapid drop.\n\n### VVS Phase 3: Syncope {#sec020}\n\nThe third phase of VVS is the transient loss of consciousness. In this phase, blood pressure and HR are at their minimum values and have experienced sharp falls just prior to the faint. However, this phase is primarily characterized by the rapid drop in CBF. In VVS patients, the rapid decrease in blood pressure and HR precedes the syncopal episode by 30--120 seconds, but can be variable \\[[@pone.0163280.ref005], [@pone.0163280.ref018]--[@pone.0163280.ref023]\\]. In the rat model of sGVS, a delay from the start of stimulation (and thus MAP and HR reductions) until the rapid drop in CBF was observed. Rats subjected to sGVS experienced a sharp reduction in CBF 51 \u00b1 21.7 seconds after starting stimulation. When rats were preconditioned with isoflurane before sGVS, the delay in the CBF drop was 72 \u00b1 36.4 seconds. If one assumes that the rapid drop in CBF in rats corresponds to the syncopal episode (*i*.*e*. loss of consciousness), then the delay in this drop is similar to the amount of time observed in human VVS \\[[@pone.0163280.ref005], [@pone.0163280.ref018]--[@pone.0163280.ref023]\\]. Interestingly, this delay in CBF drop observed in the sGVS rat model mirrors the \"pre-syncope\" window reported in humans. Therefore, this model may be utilized for examining therapies, administered during the pre-syncopal event, which are aimed at preventing the loss of consciousness.\n\nTo date, there has been a limited number of studies in which CBF was monitored during syncope. Grubb *et al*. performed two independent studies measuring the change in CBF in humans with VVS. In the authors' first study, the mean CBF during syncope in twenty VVS patients dropped by 46 \u00b1 17% \\[[@pone.0163280.ref012]\\]. The second study by Grubb *et al*. evaluated the changes in CBF during syncope in five patients in which hypotension and bradycardia was absent. Despite these patients not having reduced blood pressure or HR, they experienced a syncopal episode which correlated with a decrease in the mean CBF of 26 \u00b1 13% during syncope \\[[@pone.0163280.ref013]\\]. Sung *et al*. observed reductions of the mean CBF in syncope patients during standing or head-up tilt test to be 32 \u00b1 11.6% and 30 \u00b1 16.7%, respectively \\[[@pone.0163280.ref014]\\]. In rats, sGVS induces changes in CBF of 22 \u00b1 11.7% ([S1 Fig](#pone.0163280.s001){ref-type=\"supplementary-material\"}). The minor differences in the percent CBF reduction during syncope between the rat sGVS model and human VVS may be attributed to the use of anesthesia in rats (*i*.*e*. isoflurane blocks some of the CBF depressive effects of sGVS).\n\nFinally, sGVS in awake animals produces behavior similar to that observed in humans during VVS. During sGVS, awake animals were fatigued, had variable breathing, had uncoordinated movements, and had head swaying \\[[@pone.0163280.ref019]\\]. Two animals even had fainting-like behavior: spontaneous fall followed by 1--2 seconds of either no movement or jerky movements of one limb, then spontaneous recovery. Humans have been reported to experience a number of behavioral changes preceding the syncopal episode, including light-headedness, fatigue, blurred vision, sweating, shortness of breath, palpitations, and nausea \\[[@pone.0163280.ref011], [@pone.0163280.ref019]\\]. These symptoms are reported in the majority of VVS patients, with fatigue being the most common (in 92% of patients) \\[[@pone.0163280.ref011]\\].\n\n### VVS Phase 4: Postfaint Phase {#sec021}\n\nThe final phase of VVS is the postfaint phase which is characterized by recovery from the loss of consciousness. During this phase, blood pressure and HR recover immediately in the majority of VVS patients, however some patients remain hypotensive for up to 5 minutes \\[[@pone.0163280.ref005]\\]. In the sGVS rat model, MAP tends to return to levels indistinguishable from sham animals. However, due to the hypotensive effects of isoflurane, the MAP of sham animals begins to reduce Post-Stimulation. Yet the HR in sGVS rats remains depressed for up to 10 minutes after stimulation is stopped.\n\nAnother postfaint observation, is that most VVS patients (up to 75%) experience post-syncope behavioral changes, such as fatigue and lethargy, light-headedness, disorientation, nausea, confusion, palpitations, and altered mental status \\[[@pone.0163280.ref011], [@pone.0163280.ref019]\\]. Although the exact cause of these symptoms after VVS is unknown, the CBF in the rodent model of sGVS suggests that it may be a prolonged depression of CBF. Despite MAP and HR recovering within 15 minutes after stimulation in the rat model, CBF was reduced for up to 30 minutes after sGVS. This observation may provide a physiological explanation for the postfaint symptoms observed in the awake rat model (such as extreme fatigue) as well as the postfaint symptoms reported by VVS patients \\[[@pone.0163280.ref011], [@pone.0163280.ref019]\\].\n\nLimitations and Future Studies {#sec022}\n------------------------------\n\nTo date, there still remains some aspects of VVS which need to be investigated in the sGVS rat model. First, the mechanism of sGVS is still unclear. VVS in humans is related to vagal activation and sympathetic inhibition. Since the mechanism of sGVS is not known, it cannot yet be labeled as a true model of VVS. Second, the mechanism of human VVS is characterized by decreased cardiac output (lowered HR and/or stroke volume) resulting in hypotension, culminating in cerebral hypoperfusion \\[[@pone.0163280.ref016]\\]. The current understanding of the sGVS rat model does not focus on cardiac output, but rather HR. Future studies should investigate cardiac output during sGVS in rats. Third, the decreased blood pressure is within the range of cerebral autoregulation (50--150 mmHg), thus healthy humans are unlikely to experience reduced CBF. However, studies on the blood pressure and CBF changes in VVS patients have also found that syncope can be induced despite the reduced blood pressure being within the cerebral autoregulation range \\[[@pone.0163280.ref007], [@pone.0163280.ref012], [@pone.0163280.ref014]\\]. These reports are also supported by Grubb *et al*. and Sung *et al*. in studies which identified significant changes in CBF during syncope in patients with reduced blood pressure that is within the cerebral autoregulation range \\[[@pone.0163280.ref012], [@pone.0163280.ref014]\\]. Furthermore, in syncope patients, CBF lowering and syncope can occur even in the absence of hypotension and bradycardia \\[[@pone.0163280.ref013]\\]. Thus syncope may rely less on blood pressures outside of the cerebral autoregulation range than previously thought. However, this phenomenon, as well as the effects of sGVS on CBF need to be examined in greater detail to understand the cause of VVS despite blood pressure being within the cerebral autoregulatory range. Fourth, the mean drop in blood pressure induced by sGVS in rats is 8% which is less than observed in syncopal patients subjected to head-up tilt testing (approximately 20--30% drop in blood pressure) \\[[@pone.0163280.ref005], [@pone.0163280.ref012]\\]. The mechanism causing the CBF reduction in rats with only an 8% drop in blood pressure remains to be investigated, but may in part be due to the animals being anesthetized, thereby having a lower baseline MAP (compared to the MAP in awake rats), whereas VVS patients are fully awake when syncope occurs. Fifth, the results of Experiment 2 (observing sGVS in freely moving rats) are qualitative. Currently no behavioral tests are available for quantifying the altered behavior of rats during sGVS. In future studies we will develop a behavioral test(s) for identifying behavioral changes in rodents during sGVS.\n\nConclusion {#sec023}\n==========\n\nHerein, we advance the understanding of the sGVS rat model and its use as a pre-clinical mimetic of VVS, as well as investigate a therapy for vasovagal syncope. To our knowledge, we are the first to show that sGVS causes a CBF decrease, in addition to MAP and HR, with a rapid decrease in MAP and HR preceding CBF reduction. After stimulation, while MAP and HR recover, CBF remains depressed for up to 30 minutes post-sGVS. We are also the first to show that sGVS in awake rats produces VVS-like symptoms, including fatigue and fainting-like behavior. Finally, we are the first to begin investigating therapies for VVS using the sGVS rat model with the focus on attenuating cerebral hypoperfusion. We identified that isoflurane preconditioning reduces CBF depression caused by sGVS, which may point to potential therapies that can decrease syncopal episodes. The current study not only provides the basis for using the sGVS rat model to examine the vasovagal response and develop novel therapies for VVS, but also further establishes the sGVS rodent model as a close experimental mimetic of VVS in humans, in not only cardio- and cerebro-vascular effects, but also in behavioral changes.\n\nSupporting Information {#sec024}\n======================\n\n###### Maximum Changes from Baseline During sGVS for MAP, HR, and CBF.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nThis work was supported by a Loma Linda University Seed grant known as GCAT (Grants to promote Collaborative and Translational research) to RA and JT. Dr. McBride was supported by an NIH R01 grant NS081740 to JHZ. The authors thank Sameer Soliman and Andrew Barnett for their help with preliminary experiments; written consent was provided to the authors by Drs. Soliman and Barnett.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: **Conceptualization:** DWM JHZ RA JT.**Data curation:** DWM CR.**Formal analysis:** DWM EF.**Funding acquisition:** RA JT.**Investigation:** DWM CR RA JT.**Methodology:** DWM RA JT.**Project administration:** JT.**Resources:** RA JT.**Software:** DWM.**Supervision:** JHZ RA JT.**Validation:** DWM DK.**Visualization:** DWM.**Writing -- original draft:** DWM.**Writing -- review & editing:** DK JHZ RA JT.\n"}
+{"text": "\n"}
+{"text": "INTRODUCTION\n============\n\nThe adverse health effects of air pollution on respiratory and cardiovascular diseases are well known to the public. Regulation and monitoring of air pollution are performed at both the national and international levels. Particulate matter (PM) is one type of air pollutant. It is not a specific chemical entity, unlike other commonly known pollutants such as ozone, sulphur dioxide, and nitrogen dioxide. It is a physical category of dust with different components mixed together \\[[@b1-jpmph-49-4-205]\\]. The particle size determines the different categorizations: PM~10~ (less than 10 \u03bcm aerodynamic diameter) and PM~2.5~ (less than 2.5 \u03bcm aerodynamic diameter). PM~2.5~ is also known as fine PM.\n\nPM~2.5~ has been reported to play a major role in increasing the chance of mortality due to cardiovascular diseases because it can penetrate the capillary vessel of the lungs and reach the alveoli \\[[@b2-jpmph-49-4-205],[@b3-jpmph-49-4-205]\\]. Extensive research has been conducted on the association between PM~2.5~ and respiratory diseases including asthma. Asthma is a syndrome in which reversible respiratory obstruction occurs and is characterized by hypersensitiveness to allergens. When stimulated, a person experiences wheezing and dyspnea. In most cases, asthma is caused by a genetic predisposition and is triggered by environmental allergens.\n\nThe prevalence rate of asthma is high in children. In the case of South Korea (hererafter Korea), the prevalence rate of asthma in children steadily increased due to urbanization and westernization. In 2010, a national study based on the International Study of Asthma and Allergies in Childhood questionnaire found that 10.1% of elementary school students and 8.5% of middle school students had experienced symptoms of asthma in the past 12 months \\[[@b4-jpmph-49-4-205]\\]. These numbers should not be ignored.\n\nRecently published systematic reviews and meta-analyses reported the pooled relative risk (RR) of the number of hospital admissions and emergency department (ED) visits due to asthma as 1.023 (95% confidence interval \\[CI\\], 1.015 to 1.031, per 10 \u03bcg/m^3^ increase) when examining the effects of PM~2.5~ on the total population, and 1.025 (95% CI, 1.013 to 1.037, per 10 \u03bcg/m^3^ increase) when the subject was confined to children only \\[[@b5-jpmph-49-4-205]\\]. Another review that examined the effects of PM~2.5~ on ED visits due to asthma reported a pooled RR of 1.036 (95% CI, 1.018 to 1.053, per 10 \u03bcg/m^3^ increase) \\[[@b6-jpmph-49-4-205]\\]. However, existing studies contain several limitations. These studies were not focused on childhood asthma and only presented pooled effect estimates in children as subgroup analysis. Moreover, most of the relevant studies were conducted in North America and Europe \\[[@b7-jpmph-49-4-205]-[@b28-jpmph-49-4-205]\\], and although studies conducted in other regions exist \\[[@b29-jpmph-49-4-205]-[@b32-jpmph-49-4-205]\\], they did not consider the varying effects of PM~2.5~ according to different regions. The design of the study, the background PM~2.5~ mean concentration and variation of the region where the study was conducted, and the time of study may change the effects as well, but these factors were not adequately considered in existing studies.\n\nIn addition to the two reviews mentioned above, seven new relevant papers have recently been published \\[[@b22-jpmph-49-4-205]-[@b28-jpmph-49-4-205]\\]. Of these, the time-series studies assessed the exposure to air pollution by using the exposure value of the population-weighted average in between the measuring points of air pollution \\[[@b22-jpmph-49-4-205],[@b24-jpmph-49-4-205],[@b27-jpmph-49-4-205]\\], and the case-crossover design studies used the method of matching individual addresses with the PM~2.5~ measures \\[[@b25-jpmph-49-4-205],[@b26-jpmph-49-4-205]\\], which yielded more accurate results. Therefore, by including these recent developments, we tried to calculate more accurate pooled effect estimates of the effects of PM~2.5~ on childhood asthma and assess the variations of effects induced by differences in some factors such as region or date of research, which have been not adequately examined yet.\n\nMETHODS\n=======\n\nSelection Criteria\n------------------\n\nWe first determined some criteria for selecting relevant studies. They are as follows:\n\n1.  1\\) The subject of study was limited to children and adolescents under the age of 20.\n\n2.  2\\) Study results were limited to computerized records of hospital admissions and ED visits. Outpatient visits were excluded. Hospital admissions confirmed through interviews were not eligible. Subjective symptoms, decrease in pulmonary function, and use of emergency inhalers were not considered endpoints.\n\n3.  3\\) Effect estimates had to be presented as an odds ratio (OR) or RR.\n\nSearch Terms and Study Selection\n--------------------------------\n\nWhen deciding on search terms, we minimized keywords in order to increase the sensitivity of our searches. Some of the search terms we used were child^\\*^, pediatric^\\*^, fine particulate matter^\\*^, fine particle^\\*^, PM~2.5~, asthma^\\*^, hospitalization, hospitalisation, admission^\\*^, ed, er, and emergency. We searched studies to include in our meta-analysis using PubMed and EMBASE in March of 2016. Moreover, we selected the final eligible studies after having two authors each independently select references according to the criteria above and the same search terms and then comparing the two lists.\n\nStatistical Methods\n-------------------\n\nThe effect size was expressed as RR. We considered the OR as a proxy to the RR. In order to have all the effect estimates chosen from the selected studies to reflect the same 10 \u03bcg/m^3^ increment of PM~2.5~ concentration, we implemented meta-analyses after recalculating the \u03b2 coefficient and 95% CI presented in each study. Because the purpose of this study is to combine and identify the effects from regions all over the world, generalization of heterogeneous parts of the research group was its goal. Therefore, the random effects model using the DerSimonian and Laird \\[[@b33-jpmph-49-4-205]\\] estimation method was mainly considered, rather than the fixed effects model \\[[@b33-jpmph-49-4-205],[@b34-jpmph-49-4-205]\\]. When estimating the pooled effect, the model takes into account both the between-study variation and the within-study variation and provides a greater confidence level than the fixed effects model. The I-squared value (%) was calculated in order to identify heterogeneity.\n\nIn the primary meta-analysis of this study, an effect estimate that could represent the selected studies was used. We used the same lag value that was presented in the original paper \\[[@b35-jpmph-49-4-205]\\], but if a study presented multiple estimates from different lags, we selected the one with the largest effect size. This is because, generally, these works report the greatest effect size \\[[@b36-jpmph-49-4-205]\\]. If a study did not have one effect estimate that could represent the research, we selected two or more values that were obtained from subjects that were mutually exclusive (that is, if a study did not present an effect estimate in whole participants but presented two or more separate values from stratified groups, we included those in the meta-analysis). In order to identify publication bias, we conducted the Egger's test and identified the degree of asymmetry through a funnel plot \\[[@b37-jpmph-49-4-205]\\].\n\nMoreover, we conducted category-specific meta-analyses in order to determine what factors influenced the effect of PM~2.5~, if those influences were robust, and what factors contributed to the heterogeneity of effect estimates. We conducted the analyses by sorting the effect estimates into categories of age, results (records of hospital admissions or ED visits), season, design of the study, region, and the lag of exposure. We also conducted a separate analysis according to whether or not different pollutants were adjusted in the statistical model.\n\nWe hypothesized that the components of PM~2.5~ would change according to the time of the study and that the size of the effect could change according to the components. In addition, we thought that the variation and the mean concentration of PM~2.5~ in the region where the study was conducted might change the size of the effect. Therefore, through mixed-effects meta-regression, we derived an effect estimate of the time of the study, and the mean and standard deviation of the concentration in the study region on RR for childhood asthma.\n\nAll statistical analyses performed using R version 3.1.3 (Comprehensive R Archive Network: <http://cran.r-project.org>) and we carried out a series of statistical analyses described above through the meta package. All statistical analyses set a 5% significance level for the two-tailed test.\n\nRESULTS\n=======\n\nSelection of Relevant Studies and Extracting Effect Estimates and Their Confidence Intervals\n--------------------------------------------------------------------------------------------\n\nA total of 661 references were searched using the search terms mentioned above, and of those, we first selected 56 to examine in whole by excluding overlapping studies (n=171) and reading the titles and abstracts (n=490). Then we ultimately selected 26 studies according to the selection criteria and extracted effect estimates ([Figure 1](#f1-jpmph-49-4-205){ref-type=\"fig\"}). The 26 studies were published between 1999 and 2016, and we summarized each of the research outlines and the main research results in [Table 1](#t1-jpmph-49-4-205){ref-type=\"table\"}. Most of the research was conducted in North America and Europe and both time-series and case-crossover designs were almost equally represented.\n\nAfter extracting all effect estimates and CIs from the main body of each research paper and its supplementary materials, we broke it down to a total of 244 effect estimates. Of those, we selected 33 representative effect estimates from each study to use in our primary meta-analysis.\n\nPrimary Meta-analysis\n---------------------\n\nIn the random effects model, we were able to find that when the concentration of PM~2.5~ increased by 10 \u03bcg/m^3^, the risk of a child's hospital admission or ED visit increased by 4.8% (RR, 1.048; 95% CI, 1.028 to 1.067). The I-squared value, which shows the heterogeneity of the included studies, was 95.6%, a high figure. We presented a forest plot for the included effect estimates and pooled estimates ([Figure 2](#f2-jpmph-49-4-205){ref-type=\"fig\"}).\n\nPublication Bias\n----------------\n\nTo schematically examine the tendency toward publication bias, we found a relatively symmetrical shape in the funnel plot and confirmed that there was not much of a bias because there was not statistically significant (*p*=0.42) in the Egger 's test ([Figure 3](#f3-jpmph-49-4-205){ref-type=\"fig\"}).\n\nCategory-specific Meta-analyses\n-------------------------------\n\nWe found that the effects are greater on children below the age of five than on children ages 5 to 19, in warmer seasons, and in North America and Europe than in Asia. The pooled effect estimates extracted through the multi-pollutant model was also statistically significant (RR, 1.040; 95% CI, 1.022 to 1.057). According to the lags, the effect changed greatly from 0.2% to 6.5%, and the effect was large for 3-day lag and 3-day average lag ([Table 2](#t2-jpmph-49-4-205){ref-type=\"table\"}).\n\nMeta-regression Analyses\n------------------------\n\nWe did not find a tendency toward change in the statistically significant RR according to the time of study and the standard deviation of the background concentration of the region of study. We found a negative tendency in the mean PM~2.5~ concentration by the region of study, but it was not statistically significant (\u03b2=-0.0008, *p*=0.14) ([Figure 4](#f4-jpmph-49-4-205){ref-type=\"fig\"}).\n\nDISCUSSION\n==========\n\nIn the primary meta-analysis of the effect estimates obtained from the 26 studies, we found that in the short-term, when the concentration of PM~2.5~ increased 10 \u03bcg/m^3^, the risk of a child's hospital admission or ED visit increased 4.8%, which is statistically significant. The effect of PM~2.5~ could be considered quite robust, since the effect was maintained to 4.0% even when we pooled the estimates extracted by the multi-pollutant model in this study. This number is greater than the 2.3% found among the total population presented in the aforementioned study of Zheng et al. \\[[@b5-jpmph-49-4-205]\\]. These results show that children are more vulnerable to air pollution because their alveoli and airways are still growing, their immune systems are underdeveloped, and they spend more time outdoors, which increases ventilation \\[[@b38-jpmph-49-4-205]\\].\n\nBased on known biological mechanisms, the generation of reactive oxygen species (ROS) is accelerated because of the transition metal included in PM~2.5~. Oxidative stress from ROS may be related to epithelial cell destruction and allergic inflammation, and this process is known to be related to exacerbation of asthma \\[[@b39-jpmph-49-4-205]\\]. Meanwhile, previous studies reported that arginase may participate in a process that fine particles exacerbate childhood asthma \\[[@b40-jpmph-49-4-205]\\]. *In vivo* studies report that the overexpression of arginase influences the hyperresponsiveness of airways \\[[@b41-jpmph-49-4-205]\\] and that fine particles exacerbate the airway's responsiveness in asthma in murine models \\[[@b42-jpmph-49-4-205]\\]. Human epidemiological studies have shown that the variation of the ARG1 and ARG2 genes---which are related to the manifestation of arginase in childhood asthma patients---is statistically significant \\[[@b40-jpmph-49-4-205],[@b43-jpmph-49-4-205]\\].\n\nIn the preceding meta-analyses by Zheng et al. \\[[@b5-jpmph-49-4-205]\\], they suggested 20 relevant studies on children's asthma. We found a discrepancy between the selected studies of Zheng et al. \\[[@b5-jpmph-49-4-205]\\] and ours even aside from seven papers published more recently. They cited several studies that we excluded in the process of extracting eligible studies. On the other hand, the six studies included in this study were not cited by the preceding study. We selected studies and extracted results carefully focusing on children. Therefore, we believe that the 26 references selected for this study comprise the best selection.\n\nWe found that when the concentration of PM~2.5~ increased by 10 \u03bcg/m^3^, the risk of a child's hospital admission or ED visit increased by 4.8%. This value is greater than the 2.5% increase in children found in the preceding meta-analysis by Zheng, et al. \\[[@b5-jpmph-49-4-205]\\]. The following are some reasons to explain this difference. First, the newly added original studies included several studies in which the RR exceeded 1.10 when the measure of effect estimates was converted to 10 \u03bcg/m^3^ per increase \\[[@b7-jpmph-49-4-205],[@b9-jpmph-49-4-205],[@b13-jpmph-49-4-205],[@b20-jpmph-49-4-205],[@b23-jpmph-49-4-205],[@b28-jpmph-49-4-205]\\]. Second, while the previous study pooled the effect estimates from the 0-day, 1-day, or 2-day average lags, we used the model with the greatest effect size out of the lags reported in the original studies.\n\nIn this study, we found a difference in RR according to the season, and during the warmer seasons, the RR was 1.085 (95% CI, 1.051 to 1.119). The studies included in our meta-analysis showed quite consistent results \\[[@b9-jpmph-49-4-205],[@b20-jpmph-49-4-205],[@b22-jpmph-49-4-205],[@b23-jpmph-49-4-205],[@b31-jpmph-49-4-205]\\]. We thought the reason for this was that during warmer seasons, children spend more time outdoors and therefore spend more time exposed to PM~2.5~. In addition, greater ventilation of buildings during these seasons makes it easier for air pollutants to penetrate inside the buildings. It was reported that the individual exposure concentration of PM~2.5~ that people living in well-ventilated environments showed high correlation to the concentration of the atmosphere \\[[@b44-jpmph-49-4-205]\\]. The difference in components of PM~2.5~ according to the season may also be related, but because the extent of heterogeneity by region is too great, the evidence is not yet definitive \\[[@b45-jpmph-49-4-205]-[@b47-jpmph-49-4-205]\\].\n\nIn terms of the design of the studies, the pooled RRs for the time-series and the case-crossover design studies were 1.028 and 1.051, respectively. For the case-crossover design, the OR was calculated using the conditional logistic regression model. Compared to the RR, the OR has a tendency to overestimate the actual risk. However, it may be thought as a closer representation of reality than the exposure assessment of the time-series because a recently published case-crossover study more accurately matched air pollutants using the addresses of individuals \\[[@b20-jpmph-49-4-205],[@b21-jpmph-49-4-205],[@b25-jpmph-49-4-205],[@b26-jpmph-49-4-205]\\]. Residential information of patients entering hospitals or visiting the ED cannot be reflected in time-series. If we suppose that PM~2.5~ having an influence on exacerbating asthma as true, even in one study region, there is a possibility that the large effect in certain area with a high concentration could be diluted because of smaller effects in other area with a low concentration. We think that the actual effect is somewhere between the RRs of the time-series studies and the ORs of the case-crossover studies.\n\nWhen we examine the pooled RR of each lag, we can see that there is up to a 6% difference in value depending on the type of model. The effects of both the concentration three days before (3-day lag) and the average concentration over three days (3-day average lag) were considerable. This result is somewhat difficult to interpret. We need to consider the following factors when dealing with lags: that the ethnicities of the subject of study differ by regions and an accessibility to health services could change depending on the time of study. Through meta-regression analysis, we found a negative tendency among effect sizes depending on the mean concentration of PM~2.5~, but it was not statistically significant. Aside from the three studies in China which the mean concentrastion exceeded 30 \u03bcg/m^3^ ([Figure 4](#f4-jpmph-49-4-205){ref-type=\"fig\"}), we did not find a negative tendency in the meta-regression analysis (\u03b2=-0.0004, *p*=0.90). Therefore, we could not draw conclusions in this study regarding such a limited tendency. A negative tendency means that the effect on asthma is smaller for regions where the mean concentrations of PM~2.5~ are higher. This means that the relationship between the mean concentration and the childhood asthma could be non-linear, or more specifically, supra-linear.\n\nWhen we examine the results according to region in the category-specific analyses, the pooled RR of the three studies conducted in Shanghai and Hong Kong was 1.019, which is a smaller value than those in North America (1.047) and Europe (1.075). This is similar to the results of the previous meta-analysis that examined the short-term effects of PM~2.5~ on total mortality and cardiorespiratory mortality, and found that the pooled estimate in China was lower than in the US, Europe, Japan, and Australia \\[[@b48-jpmph-49-4-205]\\]. A hypothesis that the components of PM~2.5~ in China are different from those of developed countries was raised regarding this finding. In other words, in China, the contributions of coal combustion and desert dust---rather than exhaust from automobiles---were greater than in other regions.\n\nHowever, in a preceding meta-regression analysis including studies on PM~10~ and cardiorespiratory mortality conducted in China only, a statistically significant negative tendency was reported regarding the association between the mean concentrations of study regions and the effect sizes \\[[@b49-jpmph-49-4-205]\\]. A study conducted across 27 US regions also reported that the effect of PM~2.5~ was greater in regions with lower background mean concentration, even though the result was not statistically significant \\[[@b50-jpmph-49-4-205]\\]. In a cohort study on the effect of PM~2.5~ on cardiorespiratory mortality, the risks with the concentration level formed a supra-linear shape \\[[@b2-jpmph-49-4-205]\\]. Therefore, for the regional effect variation in this study, the hypothesis that the effect was lowered in high concentrations seems more plausible, since only groups with resistance remain and detrimental effects on individuals vulnerable to PM~2.5~ occur in lower concentrations.\n\nThere are many other genetic and environmental factors reported to cause childhood asthma besides PM~2.5~. Another hypothesis, following hygiene theory, states that allergic reactions decrease when children are exposed to micro-organisms because immune reactions are suppressed. Since westernization is still in progress in China, the effects of PM~2.5~ on asthma may be small \\[[@b51-jpmph-49-4-205],[@b52-jpmph-49-4-205]\\]. There may be an objection to this statement since the three Chinese studies included in this study were conducted in Shanghai and Hong Kong, two very westernized large cities, but the infrastructure of the residences and the lifestyles of children growing up in such regions are different from those of North America and Europe.\n\nThere are some limitations of this study. First, outpatient visits, use of inhalers, and other symptom outbreaks could all be considered health effects and consequences, but we confined the results to hospital admissions and ED visits which were mainly reported in previous studies. Therefore, the pooled effect estimate reported in our study might be underestimated. But in a study that uses surveys on symptoms and use of inhalers, the period between the exposure and outbreak could be imprecise. Moreover, results from a survey could be subjective. In cases of outpatient visits, we cannot exclude periodic follow-up cases. Second, we combined the RRs with the ORs because we deemed the OR to be proxy to the RR. Because of this, we may have calculated an overestimated value rather than the actual risk. However, in the case of Korea, hospitalization due to asthma among children between the ages of zero to 19 was 0.14% in 2014 \\[[@b53-jpmph-49-4-205]\\]. The frequency of hospital admissions or ED visits due to asthma is rare so a possible bias will be negligible. Third, we could not control the innate heterogeneity of the selected studies. Components of PM~2.5~, ethnicities of the study population, and accessibility to health service as well as different age range, season, and adjusting variables or parameters in statistical models all probably affected the heterogeneity of the studies. However, we did not find a significant decrease of heterogeneity ([Table 2](#t2-jpmph-49-4-205){ref-type=\"table\"}). In order to obtain a more accurate pooled effect estimate, a meta-analysis should be conducted after an in-depth examination of the methods and quality of research.\n\nThe strength of this study is that we newly included seven recent studies in our meta-analysis. In addition, with a focus on children, we examined variations in effect of different possible factors, and presented the direction for future studies. In particular, we raised the need for an epidemiological study on regions besides China with high concentrations of PM~2.5~.\n\nCONCLUSION\n==========\n\nWe found that in the short-term, when the concentration increased by 10 \u03bcg/m^3^, the risk of a child's hospital admission or ED visit increased by 4.8%. If we consider the fact that air pollution affects a vast range of regions and many populations, this is not a negligible figure. A more fundamental solution is the reduction of the matter from emission sources, so we need to conduct studies on sources that emit PM~2.5~ and draft feasible environment-friendly policies for such emission sources.\n\nThis study was supported by the R&D Program for Society of the National Research Foundation funded by Ministry of Science, ICT and Future Planning, Republic of Korea (grant no. 2014M3C8A5030619).\n\n**CONFLICT OF INTEREST**\n\nThe authors have no conflicts of interest associated with the material presented in this paper.\n\n![Selection process for systematic review and metaanalysis. PM~2.5~, fine particulate matter; ER, emergency room; CI, confidence interval.](jpmph-49-4-205f1){#f1-jpmph-49-4-205}\n\n![Forest plot for selected effect estimates in primary meta-analysis. RR, relative risk; CI, confidence interval; Wt, weight.](jpmph-49-4-205f2){#f2-jpmph-49-4-205}\n\n![Funnel plot for a possible selection bias in the primary meta-analysis (A). After removing three estimates (Anderson et al. \\[[@b8-jpmph-49-4-205]\\], Tecer et al. \\[[@b12-jpmph-49-4-205]\\], and Halonen et al. \\[[@b13-jpmph-49-4-205]\\]) from the right-lower area in A, still symmetrical shape is shown (B). Each black circle denotes each effect estimate of the selected studies, and the vertical red dotted line denotes the pooled random effect risk ratio in the primary meta-analysis. The p-value is derived from Egger's test.](jpmph-49-4-205f3){#f3-jpmph-49-4-205}\n\n![Bubble plot and regression line for mixed-effect meta-regression of study mean fine particulate matter (PM~2.5~) concentration and effect estimate. The black circles denote each effect estimate and their sizes represent each weight. The bold red line indicates a linear relationship between study mean PM~2.5~ concentration and relative risk and the black dotted lines indicate a 95% confidence interval.](jpmph-49-4-205f4){#f4-jpmph-49-4-205}\n\n###### \n\nSummary of selected studies on the association of short-term fine particulate matter (PM~2.5~) exposure with pediatric HA and ED visits for asthma\n\n  Author (publication year) \\[Ref\\]                                   Study period                                                        Location              Sample                                                                                Exposure assessment                                                                                                                                                                             Outcome                                                          Study design   Statistical model                         PM~25~ arithmetic mean concentration (\u03bcg/m^3^)(SD)                                                                                     Major effect estimates (risk ratio) (95% CIs)\n  ------------------------------------------------------------------- ------------------------------------------------------------------- --------------------- ------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------- -------------- ----------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------- -----------------------------------------------\n  Norris et al. (1999) \\[[@b7-jpmph-49-4-205]\\]                       Sep 1, 1995-Dec 31, 1996                                            Seattle, USA          \\<18y, 900 patients                                                                   3 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  ED visits                                                        TS             GAM with Poisson distribution             12.0 (9.5)                                                                                                                             Single-pollutant model\n  \u20031.15 (1.08, 1.23) for 1-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  Multi-pollutant model with SO~2~ and NO~2~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  \u20031.17 (1.08, 1.26) for 1-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  Lin et al. (2002) \\[[@b8-jpmph-49-4-205]\\]                          Jan 1, 1981-Dec 31, 1993                                            Toronto, Canada       6-12 y, 7319 (boys: 4629, girls: 2690) patients                                       1 Fixed site; the authors obtained data on every 6-d period from 1984 to 1990 and instructed a daily predicted value via modeling                                                               HA                                                               TS and CCD     GAM and conditional logistic regression   18.0 (8.5)                                                                                                                             Single-pollutant model\n  \u2003(a) Boys,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  \u2003\u20031.00 (0.97, 1.04) for the same day IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  \u2003\u20031.01 (0.97, 1.06) for the same day IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             \n  \u2003(b) Girls,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.06 (0.99, 1.13) for 5-d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u20031.04 (0.95, 1.15) for 5-d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Multi-pollutant model with CO, SO~2~, NO~2~ and O~3~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003(a) Boys,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  \u2003\u20030.96 (0.90, 1.02) for 5-d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u20030.94 (0.85, 1.03) for 5-d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  \u2003(b) Girls,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.01 (0.93, 1.10) for 5-d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u20030.96 (0.85, 1.09) for 5-d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Lee et al. (2006) \\[[@b29-jpmph-49-4-205]\\]                         Jan 1, 1997-Dec 31, 2002                                            Hong Kong, China      \u226418 y, 26 663 patients                                                                13 Fixed sites (before 2000, 11 sites); a daily arithmetic mean was calculated and used                                                                                                         HA                                                               TS             GAM with Poisson distribution             45.3 (16.2)                                                                                                                            Single-pollutant model\n  \u20031.066 (1.045, 1.087) for 4-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       \n  Multi-pollutant model with CO, SO~2~, NO~2~ and O~3~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u20031.032 (1.009, 1.056) for 1-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       \n  Ko et al. (2007) \\[[@b30-jpmph-49-4-205]\\]                          Jan 1, 2000-Dec 31, 2005                                            Hong Kong, China      \u226414 y, 23 596 patients                                                                3 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  HA                                                               TS             GAM with Poisson distribution             65.4 (21.1)                                                                                                                            Single-pollutant model\n  \u20031.024 (1.013, 1.034) for 5-d average 10 \u03bcg/m^3^ increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  Villneneuve et al. (2007) \\[[@b9-jpmph-49-4-205]\\]                  Jan 1, 1998-Mar 31, 2002                                            Edmonton, Canada      2-4 y, 7247 patients;                                                                 3 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  ED visits                                                        CCD            Conditional logistic regression           7.0^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^ in Apr to Sep; 7.3^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^ in Oct to Mar   Single-pollutant model:\n  5-14 y, 13 145 patients                                             \\(a\\) 2-4 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \u20031.06 (0.97, 1.15) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \u2003\u2003- Oct to Mar: 0.95 (0.84, 1.07)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \u2003\u2003- Apr to Sep: 1.16 (1.04, 1.28)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \\(b\\) 5-14 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u20031.06 (1.00, 1.12) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \u2003\u2003- Oct to Mar: 0.99 (0.91, 1.09)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \u2003\u2003- Apr to Sep: 1.10 (1.02, 1.17)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  Andersen et al. (2008) \\[[@b10-jpmph-49-4-205]\\]                    Oct 3, 2003-Dec 31, 2004                                            Copenhagen, Denmark   5-18 y, 559 patients in single pollutant model; 318 patients in two-pollutant model   1 Fixed site; a daily arithmetic mean was calculated and used                                                                                                                                   HA                                                               TS             GLM with Poisson regression               10.0 (5.0)                                                                                                                             Single-pollutant model\n  \u20031.15 (1.00, 1.32) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  Two-pollutant model with total number concentration of particles                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \u20031.13 (0.98, 1.32) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  Halonen et al. (2008) \\[[@b11-jpmph-49-4-205]\\]                     Jan 1, 1998-Dec 31, 2004                                            Helsinki, Finland     \\<15y, 4807 patients                                                                  Fixed monitoring site, no specific information available                                                                                                                                        ED visits                                                        TS             GLM with Poisson regression               9.5^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^                                                                                    Single-pollutant model\n  1.026 (0.083, 1.054) for 4-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Jalaludin et al. (2008) \\[[@b31-jpmph-49-4-205]\\]                   Jan 1, 1997-Dec 31, 2001                                            Sydney, Australia     1-14y, 317 724 patients                                                               14 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                 ED visits                                                        CCD            Conditional logistic regression           9.4 (5.1)                                                                                                                              Single-pollutant model\n  \u2003(a) 1-4 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.014 (1.007, 1.021) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003\u2003\u2003- Warm months: 1.009 (1.002, 1.017)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u2003\u2003- Cool months: 1.010 (0.999, 1.024)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003(b) 5-9 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.016 (1.005, 1.027) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003\u2003\u2003- Warm months: 1.013 (1.003, 1.024)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u2003\u2003- Cool months: 0.995 (0.976, 1.015)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003(c) 10-14 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u2003\u20031.012 (.0998, 1.027) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003\u2003\u2003- Warm months: 1.001 (0.987, 1.024)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  \u2003\u2003\u2003- Cool months: 1.017 (0.991, 1.044)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  Two-pollutant model with NO~2~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       \n  \u2003(a) 1-4 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.008 (1.001, 1.015) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003(b) 5-9 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u20031.016 (1.006, 1.026) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  \u2003(c) 10-14 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u2003\u20031.011 (0.999, 1.024) for the same-day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  Tecer et al. (2008) \\[[@b12-jpmph-49-4-205]\\]                       Dec 31, 2004-Oct 31, 2005                                           ZiDnguldak, Turkey    \\<15y, 187 patients                                                                   1 Fixed site; a daily arithmetic mean was calculated and used                                                                                                                                   HA                                                               CCD            Conditional logistic regression           29.1 (NA)                                                                                                                              Single-pollutant model\n  \u20031.25 (1.05, 1.50) for 4-d lag 10 \u03bcg/m^3^ increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   \n  \u20031.37 (1.06, 1.76) for 4-d lag IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  Halonen et al. (2010) \\[[@b13-jpmph-49-4-205]\\]                     Jan 1, 1998-Dec 31, 2004                                            Helsinki, Finland     Restricted to the warm season (May to Sep)                                            2 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  ED visits                                                        TS             GAM with Poisson distribution             8.8^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^                                                                                    Two-pollutant model with O~3~\n  \\<15 y, 1972 patients                                               \u20031.148 (1.038, 1.270) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  Silverman et al. (2010) \\[[@b14-jpmph-49-4-205]\\]                   Jan 1, 1999-Dec 31, 2006                                            New York City, USA    Restricted to the warm season (Apr to Aug)                                            24 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                 HA                                                               TS             GLM with Poisson regression               13^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^                                                                                     Single-pollutant model\n  \\<6 y                                                               \u2003(a) \\<6 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \\- Non-ICU admission: 15 185,                                       \u2003\u2003- Non-ICU: 1.14 (1.10, 1.19) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \\- ICU admission: 1141 patients                                     \u2003\u2003- ICU: 1.03 (0.91, 1.17) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  6-18y                                                               \u2003(b) 6-18 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \\- Non-ICU admission: 10 332,                                       \u2003\u2003- Non-ICU: 1.19 (1.11, 1.27) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \\- ICU admission: 994 patients                                      \u2003\u2003- ICU: 1.26 (1.10, 1.44) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  Two-pollutant model with O~3~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u2003(a) \\<6 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u2003- Non-ICU: 1.13 (1.08, 1.18) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u2003- ICU: 1.04 (0.91, 1.19) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  \u2003(b) 6-18 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         \n  \u2003\u2003- Non-ICU: 1.16 (1.08, 1.23) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n  \u2003\u2003- ICU: 1.23 (1.07-1.41) for 2-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  Strickland et al. (2010) \\[[@b15-jpmph-49-4-205]\\]                  Aug 1, 1998-Dec 31, 2004                                            Atlanta, USA          5-17 y, 91 386 patients                                                               11 Fixed sites; a population-weighting average across monitors was calculated and used                                                                                                          ED visits                                                        TS             GLM with Poisson regression               16.4 (7.4)                                                                                                                             Single-pollutant model\n  \u2003- Whole period: 1.020 (1.002,1.039) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \u2003- Warm season: 1.043 (1.016, 1.070) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \u2003- Cold season: 1.005 (0.978, 1.031) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  Li et al. (2011) \\[[@b16-jpmph-49-4-205]\\]                          Jan 1, 2004-Dec 31, 2006                                            Detroit, USA          2-18 y, 7063 patients                                                                 4 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  ED visits + HA^[2](#tfn2-jpmph-49-4-205){ref-type=\"table-fn\"}^   TS and CCD     GAM and conditional logistic regression   15.0 (7.9)                                                                                                                             Single-pollutant model\n  \u20031.030 (1.001, 1.061) for 5 d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             \n  \u20031.039 (1.013, 1.066) for 5 d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  Glad et al. (2012) \\[[@b17-jpmph-49-4-205]\\]                        Jan 1, 2002-Dec 31, 2005                                            Pittsburgh, USA       0-17 y, 978 patients                                                                  2 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  ED visits                                                        CCD            Conditional logistic regression           NA                                                                                                                                     Single-pollutant model\n  \u20031.012 (0.916, 1.118) for the same-day 10 \u03bcg/m^3^ increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Iskandar et al. (2012) \\[[@b18-jpmph-49-4-205]\\]                    May 15, 2001-Dec 31, 2008                                           Copenhagen, Denmark   0-18 y, 6329 patients                                                                 1 Fixed site; a daily arithmetic mean was calculated and used                                                                                                                                   HA                                                               CCD            Conditional logistic regression           10.3 (5.4)                                                                                                                             Single-pollutant model\n  \u20031.09 (1.04, 1.13) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  Two-pollutant model with NO~2~:                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  \u20031.06 (1.02, 1.11) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  Winquist et al. (2012) \\[[@b19-jpmph-49-4-205]\\]                    Jan 1, 2001-Jun 27, 2007                                            St. Louis, USA        0-1 y.                                                                                1 Fixed site; a daily arithmetic mean was calculated and used                                                                                                                                   ED visits & HA                                                   TS             GLM with Poisson regression               14.4 (7.5)                                                                                                                             Single-pollutant model\n  \\- ED: 12 236 patients                                              \u2003(a) 0-1 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      \n  2-18 y.                                                             \u2003\u2003- ED: 1.047 (0.999, 1.097) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \\- ED: 49 978 patients                                              \u2003(b) 2-18 y,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \\- All HA: 7095 patients                                            \u2003\u2003- ED: 1.050 (1.021,1.080) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         \n  \u2003\u2003- HA: 1.052 (0.985, 1.123) for 5-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  Delfino et al. (2014) \\[[@b20-jpmph-49-4-205]\\]                     Jan 1, 2000-Dec 31, 2008                                            California, USA       0-18 y, 11 390 patients                                                               Subject addresses were geocoded; using a modified, California LINE Source Dispersion Model, version. 4 to estimate pollutants at each residence                                                 ED visits + HA^[2](#tfn2-jpmph-49-4-205){ref-type=\"table-fn\"}^   CCD            Conditional logistic regression           \\- Warm season: 16.0 (9.5)                                                                                                             Single-pollutant model\n  \\- Cool season: 19.0 (13.8)                                         \u2003- Warm season: 1.079 (1.008, 1.154) for 7-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n  \u2003- Cool season: 1.162 (1.076, 1.254) for 7-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  Gleason et al. (2014) \\[[@b21-jpmph-49-4-205]\\]                     Jan 1, 2004-Dec 31, 2007                                            New Jersey, USA       3-17 y, 21 854 patients                                                               Subject addresses were geocoded; using 12\u00d712-km grid from the Multi-Scale Air Quality Model to estimate pollutants at each residence                                                            ED visits                                                        CCD            Conditional logistic regression           NA                                                                                                                                     Single-pollutant model\n  \u20031.03 (1.02, 1.04) for the same day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  Multipollutant model with O~3~ and other pollens                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \u20030.99 (0.98, 1.01) for the same day IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  Hua et al. (2014) \\[[@b32-jpmph-49-4-205]\\]                         Jan 1, 2007-Jul 31, 2012                                            Shanghai, China       0-14 y, 114 673 patients                                                              1 Fixed site; a daily arithmetic mean was calculated and used                                                                                                                                   HA                                                               TS             Polynomial distributed lag model          40.9 (27.7)                                                                                                                            Single-pollutant model\n  \u20031.04 (1.02, 1.05) for IQR increase with a maximum lag of 3 d                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u20031.06 (1.05, 1.08) for IQR increase with a maximum lag of 5 d                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Multipollutant model with NO~2~ and SO~2~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  \u20031.03 (1.02, 1.05) for IQR increase with a maximum lag of 3 d                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u20031.06 (1.04, 1.08) for IQR increase with a maximum lag of 5 d                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Strickland et al. (2014) \\[[@b22-jpmph-49-4-205]\\]                  Jan 1, 2002-Jun 30, 2010                                            Atlanta, USA          2-16 y, 109 758 patients                                                              6 Fixed sites; a population-weighting average across monitors calculated and used                                                                                                               ED visits                                                        TS             GLM with Poisson regression               13.3 (5.4)                                                                                                                             Single-pollutant model\n  \u20031.032 (1.019, 1.044) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   \n  Two-pollutant model with O~3~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u20031.022 (1.009, 1.035) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   \n  Wendt et al. (2014) \\[[@b23-jpmph-49-4-205]\\]                       Jan 1, 2005-Dec 31, 2007                                            Boston, USA           0-17 y                                                                                3 Fixed sites; a daily arithmetic mean was calculated and used                                                                                                                                  HA                                                               CCD            Conditional logistic regression           15.0 (6.0)                                                                                                                             Single-pollutant model\n  \\- May to Oct: 6061 patients                                        \u2003- May to Oct: 1.10 (1.03, 1.17) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \\- Nov to Apr: 7894 patients                                        \u2003- Nov to April: 1.06 (1.00, 1.14) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  \n  Two-pollutant model with NO~2~                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       \n  \u2003- May to Oct: 1.13 (1.04, 1.24) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  \u2003- Nov to Apr: 1.00 (0.93, 1.07) for 6-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n  Byers et al. (2016) \\[[@b24-jpmph-49-4-205]\\]                       Jan 1, 2007-Dec 31, 2011                                            Indianapolis, USA     5-17 y, 33 981 patients                                                               3 Fixed sites; a population-weighting average across monitors calculated and used                                                                                                               ED visits                                                        TS             GLM with Poisson regression               13.6 (7.1)                                                                                                                             Single-pollutant model\n  \u2003- All seasons: 1.007 (0.986, 1.029) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  \u2003- Apr to Sep: 0.985 (0.934, 1.040) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \u2003- Oct to Mar: 0.976 (0.930, 1.025) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  Gleason et al. (2015) \\[[@b25-jpmph-49-4-205]\\]                     Jan 1, 2004-Dec 31, 2007                                            Newark, USA           3-17 y, 3675 patients                                                                 Subject addresses were geocoded; using grid from the Multi-Scale Air Quality Model to estimate pollutants at each residence                                                                     ED visits                                                        TS and CCD     GLM and conditional logistic regression   NA                                                                                                                                     Single-pollutant model\n  \u20031.00 (0.96, 1.05) for 3-d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n  \u20031.00 (0.96, 1.04) for 3-d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  Multipollutant model with O~3~ and other pollens                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  \u20030.93 (0.89, 0.98) for 3-d average IQR increase in TS                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n  \u20030.95 (0.91, 1.00) for 3-d average IQR increase in CCD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  Strickland et al. (2015) \\[[@b26-jpmph-49-4-205]\\]                  Jan 1, 2002-Jun 30, 2010                                            Georgia, USA          2-18 y, 189 816 patients                                                              Subject addresses were geocoded; using a two-stage model that includes land use parameters and satellite aerosol optical depth measurements at 1-km resolution to estimate pollutants           ED visits                                                        CCD            Conditional logistic regression           12.9^[1](#tfn1-jpmph-49-4-205){ref-type=\"table-fn\"}^                                                                                   Single-pollutant model\n  \u20031.013 (1.003, 1.023) for the same day 10 \u03bcg/m^3^ increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Alhanti et al. (2016) \\[[@b27-jpmph-49-4-205]\\]                     Jan 1, 2006-Dec 31, 2009                                            Dallas, USA           0-4 y, mean daily counts: 16.91 patients                                              All available monitors; the monitoring data were first spatially interpolated across the study's geographic domain and then a population-weighted average across monitors calculated and used   ED visits                                                        TS             GLM with Poisson regression               11.1 (4.7)                                                                                                                             Single-pollutant model\n  5-18 y, mean daily counts: 25.75 patients                           \u20030-4 y, 0.98 (0.94, 1.02) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           \n  \u20035-18 y, 0.99 (0.95, 1.03) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \n  Weichenthal et al. (2016) \\[[@b28-jpmph-49-4-205]\\]                 Jan 1, 2004-Dec 31, 2011                                            Ontario, Canada       Total; 127 836 patients,                                                              Fixed site in Ontario which is part of Canada's National Air Pollution Surveillance network; a daily arithmetic mean was calculated and used                                                    ED visits                                                        CCD            Conditional logistic regression           7.1 (6.3)                                                                                                                              Single-pollutant model\n  \\<9y, NA                                                            \u20031.072 (1.042, 1.100) for 3-d average IQR increase                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n\nRef, reference number; HA, hospital admission; ED, emergency department; GLM, generalized linear model; GAM, generalized additive model; NA, not available; IQR, interquartile range; TS, time series; CCD, case-crossover design; PM, particulate matter; SD, standard devaition; CI, confidence interval; ICU, intensive care unit; CO, carbon monoxide; SO~2~, sulfur dioxide; NO~2~, nitrogen dioxide; O~3~, ozone.\n\nMedian value of the daily PM~2.5~ distribution during the entire study period. This study doesn't present the arithmetic mean of PM~2.5~.\n\nThe authors regarded asthma morbidity as hospital encounters which counted both HA and ED visits.\n\n###### \n\nResults of category-specific meta-analyses\n\n                                                        No. of study (no. of estimate)   RR (95% CIs)^[1](#tfn3-jpmph-49-4-205){ref-type=\"table-fn\"}^   I^[2](#tfn4-jpmph-49-4-205){ref-type=\"table-fn\"}^ (%)\n  ----------------------------------------------------- -------------------------------- -------------------------------------------------------------- -------------------------------------------------------\n  Age^[2](#tfn4-jpmph-49-4-205){ref-type=\"table-fn\"}^                                                                                                   \n  \u2003\\< 5                                                 7 (9)                            1.044 (1.017, 1.071)                                           81.9\n  \u20035-18                                                 12 (15)                          1.027 (1.011, 1.043)                                           76.8\n  Outcome                                                                                                                                               \n  \u2003HA                                                   10 (15)                          1.048 (1.029, 1.067)                                           77.7\n  \u2003ED visits                                            15 (17)                          1.027 (1.011, 1.044)                                           79.5\n  Season                                                                                                                                                \n  \u2003Cold                                                 7 (8)                            1.015 (0.994, 1.037)                                           57.1\n  \u2003Warm                                                 9 (11)                           1.085 (1.051, 1.119)                                           94.8\n  Study design                                                                                                                                          \n  \u2003TS                                                   15 (19)                          1.028 (1.015, 1.041)                                           76.9\n  \u2003CCD                                                  13 (17)                          1.051 (1.020, 1.084)                                           96.6\n  Area                                                                                                                                                  \n  \u2003North America                                        14 (19)                          1.047 (1.019, 1.076)                                           96.1\n  \u2003Europe                                               8 (11)                           1.075 (1.030, 1.123)                                           65.9\n  \u2003China                                                3 (3)                            1.019 (1.013, 1.025)                                           0.0\n  Multipollutant model                                                                                                                                  \n  \u2003No                                                   25 (33)                          1.054 (1.037, 1.071)                                           96.0\n  \u2003Yes                                                  13 (18)                          1.040 (1.022, 1.057)                                           83.1\n  Time lag (d)                                                                                                                                          \n  \u20030 (same day)                                         12 (14)                          1.018 (1.005, 1.028)                                           60.9\n  \u20031                                                    11 (13)                          1.018 (1.005, 1.030)                                           59.6\n  \u20032                                                    8 (8)                            1.002 (0.984, 1.021)                                           84.6\n  \u20033                                                    10 (11)                          1.030 (1.015, 1.045)                                           66.6\n  \u20034                                                    4 (4)                            1.016 (0.969, 1.065)                                           83.1\n  \u20035                                                    5 (6)                            1.019 (0.975, 1.065)                                           93.5\n  Average                                                                                                                                               \n  \u20032                                                    3 (7)                            1.065 (1.020, 1.113)                                           81.7\n  \u20033                                                    11 (15)                          1.019 (1.006, 1.033)                                           82.2\n  \u20035                                                    10 (14)                          1.025 (1.007, 1.043)                                           77.4\n  \u20036                                                    3 (5)                            1.029 (0.938, 1.129)                                           69.9\n\nRR, relative risk; CI, confidence interval; HA, hospital admission; ED, emergency department; TS, time-series; CCD, case-crossover design.\n\nCalculated by DerSimonian and Laird random effects model \\[[@b33-jpmph-49-4-205]\\].\n\nThere are two exceptions: Silverman et al. \\[[@b14-jpmph-49-4-205]\\] and Iskandar et al. \\[[@b18-jpmph-49-4-205]\\]: cut-off age is six.\n"}
+{"text": "Transition elements, such as copper, iron, or zinc, are essential nutrients for plants. They participate in every biological process. However, intracellular metal levels must be maintained within a narrow physiological concentration. Too little, and not enough cofactors are available to the cell; too much, and Fenton-type reactions and mismetallation events will disrupt many cellular processes. As a result, plants have developed complex systems to control metal uptake and to deliver them to all tissues and cells. In this Research Topic, we have collected some of the most recent work furthering our understanding of Metallic Micronutrient Homeostasis in Plants.\n\nPlant dependence on metals contrasts to the common low bioavailability of these nutrients in many soil types. Growing evidence indicates that plants have optimized the use of metals facilitating their relocation from one cellular compartment to another, depending on their need in different physiological processes. In this Topic, [Zhang and Kr\u00e4mer](https://doi.org/10.3389/fpls.2018.01641) have reanalyzed available bioinformatic data to show that there is a diurnal regulation of the translation of proteins involved in Fe-S metabolism in *Arabidopsis thaliana*, including that of frataxin, a protein involved in iron transfer in Fe-S protein biogenesis, and also in copper control in mitochondria and chloroplasts, as reviewed in this issue by [Gomez-Casati et al.](https://doi.org/10.3389/fpls.2018.01706) Iron is not the only metal micronutrient that is regulated by day-night cycles. Copper homeostasis is also controlled in this way, the likely consequence of plastocyanin synthesis for photosynthesis, and the toxic effect of copper in Fe-S synthesis. [Andr\u00e9s-Col\u00e1s et al.](https://doi.org/10.3389/fpls.2018.00910) showed that copper transporter COPT3 transcription is controlled in this way, in a process that is dependent of copper availability to the plant and on transcription factor TCP16.\n\nThe study of metal (re)distribution in plants can be greatly assisted by metal-imaging methods. Using X-ray fluorescence approaches, [Vigani et al.](https://doi.org/10.3389/fpls.2018.01112) showed how iron deficiency affects zinc distribution in leaves. With a similar approach, [Ibeas et al.](https://doi.org/10.3389/fpls.2018.01985) have determined the iron distribution in *Chenopodium quinoa* seeds and used it to validate the Perls-DAB histochemical method to visualize iron. This approach not only led to illustrate how iron distribution in seeds changed during dicot evolution but showed how subcellular iron distribution changes during embryo development ([Ibeas et al.](https://doi.org/10.3389/fpls.2017.02186)). Metal imaging approaches are also important to evaluate how metal nanoparticles could enter into trophic chains through plants, and the physiological effects that might cause there ([Mosa et al.](https://doi.org/10.3389/fpls.2018.00872)).\n\nThe evolving metal distribution in a cell and an organism could also reflect how some metalloenzymes are expressed with different demand for metal cofactors. For this, we would need high throughput metalloproteomics methods to identify the precise nature of the accepting metalloprotein to put into context many of the observed phenotypes. For instance, symbiotic nitrogen fixation requires one or several zinc-proteins that although dependent on zinc transfer into the endoplasmic reticulum of nodule cells, exert their effect in nitrogen-fixing symbiosomes and in nodule development ([Le\u00f3n-Mediavilla et al.](https://doi.org/10.3389/fpls.2018.00990)).\n\nHormones have been known to control metal homeostasis, but the precise mechanisms of how hormone levels are controlled by metals still remains elusive. In this topic, [Garc\u00eda et al.](https://doi.org/10.3389/fpls.2018.01325) reported how long-distance iron trafficking mediated by OPT3 could control ethylene metabolism as well as GSNO levels. In addition, micronutrients such as silicon, modulate hormone levels to overcome abiotic stress such as chilling and also improve the nutrition by metals (particularly zinc and manganese) ([Moradtalab et al.](https://doi.org/10.3389/fpls.2018.00420)).\n\nIn summary, the work shown in this Research Topic illustrates our advances in defining how cells and organisms control and redistribute essential metal nutrients. This distribution is affected by biotic and abiotic interactions, hormones and responds to different physiological states of plant development. The work also illustrates the need to have new methods to determine and visualize metals, as well-new developments in metalloproteomics and in metal-speciation analyses.\n\nAuthor Contributions {#s1}\n====================\n\nAll authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.\n\nConflict of Interest Statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n[^1]: Edited and reviewed by: Felipe Klein Ricachenevsky, Universidade Federal de Santa Maria, Brazil\n\n[^2]: This article was submitted to Plant Nutrition, a section of the journal Frontiers in Plant Science\n"}
+{"text": "**Funding:** This study was supported by Emergency research project of novel coronavirus infection of Anhui Medical University (YJGG202003).\n\nThis article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/jmv.25934\n\n[^1]: co\u2010first authors of this paper\n"}
+{"text": "Introduction {#section1-1536012118821034}\n============\n\nMET (also known as protein product of the c-*MET* proto-oncogene) is a receptor tyrosine kinase, which binds the ligand hepatocyte growth factor (HGF).^[@bibr1-1536012118821034]^ Decreased *MET* function is associated with an increased risk of autism spectrum disorder (ASD).^[@bibr2-1536012118821034][@bibr3-1536012118821034][@bibr4-1536012118821034]--[@bibr5-1536012118821034]^ *MET* transcript expression is prominent in the cerebral cortex, hippocampus, and amygdala.^[@bibr6-1536012118821034]^ The changes in *Met* signaling during development could affect neuronal number as well as the complexity of the neuropil,^[@bibr7-1536012118821034][@bibr8-1536012118821034]--[@bibr9-1536012118821034]^ consequently altering the structure or connectivity in the brain and ultimately impacting its function. In a recent neuroimaging study,^[@bibr10-1536012118821034]^ when children and adolescents with autism-associated promoter variant in *MET* were clustered into homozygous and heterozygous groups, the *MET* risk allele had strong impact across individuals within the heterozygous group. The results suggested that *MET* risk genotype should be taken into account to stratify individuals with ASD. Human neuroimaging studies likely reflect the effects of multiple different changes at the genetic level, as reflected in the considerable heterogeneity observed between studies.\n\nImpaired somatosensory processing and subsequent associated deficits in motor skills, interpersonal relations, social skills, and communication are well-documented in neurodevelopmental disorders, including ASD.^[@bibr11-1536012118821034]^ The term \"tactile defensiveness\" has emerged to depict the tactile hypo- or hypersensitivity of children with ASD. Mouse models offer a particularly useful paradigm to study the tactile sensitivity deficits. Mice rely heavily on tactile discrimination with their whiskers. A large portion of the primary somatosensory cortex is devoted to representation of whiskers by neural modules in a one-to-one and patterned fashion.^[@bibr12-1536012118821034]^ These neural modules known as \"barrels\" in the cortex also exist in subcortical structures, such as the sensory thalamus and the trigeminal brainstem, which carry the whisker inputs to the cortex. Thus, the mouse whisker-barrel system is uniquely advantageous to study the development, plasticity, adult organization, and function of the somatosensory system.\n\nPreviously, we have used a mouse line in which Met signaling is inactivated specifically in the cerebral cortex and hippocampus of *Met--Emx1* mice. This mouse line was generated using a *Cre-loxP* recombination strategy. Unlike global *Met* null mutants,^[@bibr13-1536012118821034]^ *Met--Emx1* mice live to adulthood,^[@bibr14-1536012118821034]^ allowing examination of brain structure during the postnatal and adult periods. As reported in an earlier anatomical magnetic resonance imaging (MRI) study,^[@bibr15-1536012118821034]^ the rostral cortex, caudal hippocampus, dorsal striatum, thalamus, and corpus callosum were all larger in adult, but not in juvenile, *Met--Emx1* mice relative to wild-type mice. The specificity of the changes suggests that aberrant expansion of the forebrain is consistent with continued axonal and dendritic growth, potentially leading to improper circuit formation and maintenance. In a separate electrophysiological study, we observed an excitatory/inhibitory (E/I) imbalance in somatosensory thalamocortical transmission in an in vitro slice preparation in the same mouse model,^[@bibr16-1536012118821034]^ consistent with the E/I imbalance theory in autism.^[@bibr17-1536012118821034]^ Behavioral deficits in *Met--Emx1* mice have been reported as well.^[@bibr18-1536012118821034],[@bibr19-1536012118821034]^ Given the tactile impairments associated with neurodevelopmental disorders, genetic associations of *MET* with ASD, and our previous data of anatomical differences in structures and electrophysiology, the *Met--Emx1* mouse line was chosen to study altered somatosensory responses at the circuit level in the intact live animal.\n\nWe employed in vivo high-resolution proton magnetic resonance spectroscopy (^1^H-MRS) and resting-state functional magnetic resonance imaging (rsfMRI) to investigate the neurotransmitters (glutamate and gamma-aminobutyric acid \\[GABA\\]) and functional connectivity alterations, respectively, within the somatosensory thalamocortical circuitry. The neuroimaging method rsfMRI is used to investigate regional brain interactions when a subject is not performing an explicit task. Resting-state functional MRI signals are thought to arise from spontaneous low-frequency fluctuations in blood oxygen level-dependent (BOLD) signal. The coherence of BOLD signal---also named \"functional connectivity\"---can be examined using hypothesis-driven seed-based analysis, which is based on the correlation of the resting-state time series of a selected brain region (seed) with the time series of other regions.^[@bibr20-1536012118821034][@bibr21-1536012118821034]--[@bibr22-1536012118821034]^ Functional connectivity detects tightly coupled regions in brain baseline functional systems.^[@bibr23-1536012118821034][@bibr24-1536012118821034][@bibr25-1536012118821034]--[@bibr26-1536012118821034]^ Therefore, rsfMRI can display basic brain functional organization, which reflects the formation and maintenance of neural networks.^[@bibr27-1536012118821034]^\n\nIn vivo ^1^H-MRS is a noninvasive spectroscopic technique capable of measuring neurochemical concentrations in specific regions of the central nervous system.^[@bibr28-1536012118821034][@bibr29-1536012118821034]--[@bibr30-1536012118821034]^ At high field strength (\u22657 T) with a short echo time (\u226425 ms), the resolution of coupled peaks, such as the CH~2~ from glutamate and glutamine and GABA, can be clearly resolved, suggesting the potential of measuring GABA at 2.30 ppm without the need of spectral *J*-editing.^[@bibr29-1536012118821034],[@bibr30-1536012118821034]^ The application of in vivo ^1^H-MRS provides biochemical information on glutamatergic and GABAergic functions without bias introduced from sample preparation. Alterations of correlation between GABA or glutamate and functional connectivity have been reported in mental disorders or following experimental modulation^[@bibr31-1536012118821034][@bibr32-1536012118821034][@bibr33-1536012118821034]--[@bibr34-1536012118821034]^ but not yet in ASD. Here, we provide correlations between neurotransmitter levels and functional connectivity within the somatosensory thalamocortical circuitry in *Met--Emx1* mice.\n\nSets of discrete neural elements linked by connections can be considered a network. An analytical technique---graph theory analysis---is helpful in assessing the brain network properties. A network can be defined in graph theory as a set of nodes or vertices and the edges or lines between them can be quantitatively described by various measures (eg, clustering coefficient, local efficiency, small-worldness \\[SW\\]) to assess the network features such as integration, segregation, and resilience.^[@bibr35-1536012118821034]^ This method is used to characterize organization of networks in neuroimaging studies in humans, nonhuman primates, and rats,^[@bibr36-1536012118821034][@bibr37-1536012118821034][@bibr38-1536012118821034]--[@bibr39-1536012118821034]^ and it has paved the way for increasingly sophisticated investigations of brain connectivity in the study of ASD.^[@bibr40-1536012118821034],[@bibr41-1536012118821034]^ Recently, modular structure and hub community were detected in the mouse brain using rsfMRI,^[@bibr42-1536012118821034],[@bibr43-1536012118821034]^ indicating the presence of a well-organized brain network that shares similar features as primates and rats. In addition to seed-based functional connectivity analysis, we performed a graph theory analysis to compare the transgenic mice with homozygous and heterozygous genotype with wild-type mice in topological organization across brain regions.\n\nMethods {#section2-1536012118821034}\n=======\n\nAnimals {#section3-1536012118821034}\n-------\n\nFounder mice, *Met-fx* (\\#016974; Jackson Laboratory, Bar Harbor, Maine^[@bibr44-1536012118821034]^) and *Emx1-Cre* (\\#005628; Jackson Laboratory), were crossed and maintained on a C57BL/6J background for \\>30 generations. Mice with a single inactive *Met* allele, *Met-fx/Emx1-Cre*, are denoted as *Met--Emx1* or heterozygous *Met--Emx1*, and those with 2 inactive alleles are denoted as *Met-Met-Emx1* or homozygous *Met--Emx1*. Mice used in these experiments were adult male and female littermates from mating nonsibling heterozygotes. Wild-type mice included the *Emx1-Cre* allele alone, the *Met-fx* allele alone, and mice lacking any transgenes. In our previous studies, we observed no differences between these groups,^[@bibr15-1536012118821034],[@bibr16-1536012118821034]^ and therefore, we combined the data from these groups. We imaged 6 groups: wild-type male (n = 12), wild-type female (n = 12), heterozygous *Met--Emx1* male (n = 12), heterozygous *Met--Emx1* female (n = 12), homozygous *Met-Met-Emx1* male (n = 12), and homozygous *Met-Met-Emx1* female (n = 12) mice. All mice were adults, more than 90 days of age. All procedures conformed to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee, University of Maryland, Baltimore.\n\nImage Acquisition {#section4-1536012118821034}\n-----------------\n\nWe performed all experiments on a Bruker Biospec 7 T 30-cm horizontal bore scanner (Bruker Biospin MRI GmbH, Ettlingen, Germany) equipped with a BGA12S gradient system and interfaced to a Bruker Paravision 5.1 console. A Bruker 72-mm linear-volume coil served as the transmitter and a Bruker ^1^H 4-element surface coil array served as the receiver. Anesthesia was induced using isoflurane (\u223c1%) with enriched O~2~, which is considered appropriate based on previous evidence in rodent models.^[@bibr45-1536012118821034][@bibr46-1536012118821034]--[@bibr47-1536012118821034]^ An MR compatible small-animal monitoring and gating system (SA Instruments, Inc., Stony Brook, New York) monitored the respiration rate and body temperature, which was maintained at 36\u00b0C to 37.5\u00b0C using a warm water bath circulation.\n\nA 3-slice (axial, mid-sagittal, and coronal) scout image using rapid acquisition with fast low-angle shot localized the mouse brain. A fast shimming procedure (FASTMAP) improved the B~0~ homogeneity covering the brain. Anatomic images (repetition time/effective echo time \\[TR/TE~eff~ = 5000/18.42 ms\\]) were obtained using a 2-dimensional rapid acquisition with relaxation enhancement sequence covering the entire brain. Imaging was performed over a 1.75-cm field of view (FOV) in the coronal plane with an in-plane resolution of 146 \u03bcm using 18 slices at 1-mm thickness.\n\nResting-state functional MRI was acquired matching the anatomic images using a single shot, spin echo planar imaging sequence (TR/TE = 1000/27.6 ms) with a 1.75-cm FOV and an in-plane resolution of 273 \u03bcm^2^ using 18 slices at 1-mm thickness. Despite a weaker BOLD signal, spin echo provides much less imaging distortion than gradient echo when imaging mouse brains. Therefore, spin echo imaging was used in this study. [Figure 1A](#fig1-1536012118821034){ref-type=\"fig\"} shows the raw acquisition EPI images from a female mouse. Six hundred repetitions were taken, resulting in a total scanning time of around 10 minutes for each data set. Two rsfMRI sessions were acquired for each animal. During the rsfMRI experiment, we adjusted the isoflurane anesthesia to 0.75% to 1% and the respiratory rate of the mouse was \u223c50 breaths/minute without any detectable motion.\n\n![Brain imaging and spectroscopy. Demonstration of the raw EPI images of a female control mouse (A), designation of ROIs for rsfMRI analysis (B), connectivity map of control animals with right-side somatosensory cortex as the seed ROI (C), and representative ^1^HMRS spectra from somatosensory cortex (top) and thalamus (bottom) (D). Amy indicates amygdala; Au, auditory cortex; Cg, cingulate cortex; CIC, inferior colliculus; GABA, gamma-aminobutyric acid; Gln, glutamine; Glu, glutamate; Glx, glutamate and glutamine; GSH, glutathione; Hip, hippocampus; hypotha, hypothalamus; Ins, *myo*-inositol; lCPu/mCPu, lateral/medial caudate putamen; M1/M2, primary/secondary motor cortex; MM, macromolecules; NAA, *N*-acetyl-aspartate; NAc, nucleus accumbens; Pir, piriform cortex; Pn, pontine nuclei; PnO, rostral pontine reticular nucleus; PrL/IL, prelimbic/infralimbic cortex; RSC, retrosplenial cortex; rsfMRI, resting-state functional magnetic resonance imaging; S1/S2, primary/secondary somatosensory cortex; Tau, taurine; tCho, glycerophosphocholine and phosphocholine; tCr, creatine and phosphocreatine; Tha, thalamus; V, visual cortex; ^1^HMRS, high-resolution proton magnetic resonance spectroscopy.](10.1177_1536012118821034-fig1){#fig1-1536012118821034}\n\nHigh-resolution proton magnetic resonance spectroscopy data were obtained from the right somatosensory cortex (2.5 \u00d7 1 \u00d7 1.5 mm^3^, number of averages = 1000) and the right thalamus (2.0 \u00d7 2.5 \u00d7 1.5 mm^3^, number of averages = 600). Prior to acquiring the spectra, the FASTMAP procedure aided the adjustments of all first- and second-order shims over the voxel of interest. After shimming, a typical water line-width was 10 to 11 Hz (0.033--0.037 ppm). This allowed for good separation of the glutamate (\u03b3CH~3~ 2.35 ppm, \u03b1CH 3.75 ppm) and GABA (\u03b1CH~2~ 2.28 ppm, \u03b2CH~2~ 1.89 ppm). A short echo time Point-RESolved Spectroscopy (PRESS) pulse sequence (TR/TE = 2500/10 ms) was used for MRS data acquisition.^[@bibr48-1536012118821034]^ The unsuppressed water signal from the prescribed voxel was collected as a biochemical concentration reference for each scan. The total in vivo imaging acquisition time was 2.5 hours.\n\nResting-State Functional MRI Processing {#section5-1536012118821034}\n---------------------------------------\n\nWe conducted all rsfMRI image preprocessing and processing using SPM12 (<http://www.fil.ion.ucl.ac.uk/spm/>) and Analysis of Functional NeuroImages (AFNI) (<http://afni.nimh.nih.gov/afni>). The processing pipeline included slice-timing correction, motion correction, alignment to a study-specific mouse anatomic template, band-pass filtering, orthogonalization of motion-derived parameters, and smoothing. Specifically, the first 10 and last 50 volumes were excluded from each data set according to the stability of animals and the equipment. Slice-timing correction and motion correction were performed in SPM12 with the ninth slice as the reference slice and second-degree B-spline interpolation as the estimation method. The study-specific mouse anatomical template was created in SPM12 with 2 steps. First, the anatomical image of each mouse was aligned to the anatomical image of a representative mouse with fourth-degree B-spline interpolation. Then, the aligned anatomical images were averaged, yielding the study-specific template. Resting-state functional MRI of each mouse was aligned to the study-specific template with the original anatomic image as the source image in SPM12. Next, the motion-derived parameters were regressed and the data were band-pass filtered (0.01--0.1 Hz) and smoothed (Full width at half maximum \\[FWHM\\] = 0.6 mm) with 3dBandpass function in AFNI.\n\nThirty-seven regions of interest (ROIs) were manually defined based on the standard Paxinos and Watson mouse brain atlas^[@bibr49-1536012118821034]^ ([Figure 1B](#fig1-1536012118821034){ref-type=\"fig\"}). For each animal, the regionally averaged BOLD time series was extracted from each ROI, and then these extracted time courses of ROIs were correlated using the Pearson correlation coefficient, producing a 37 \u00d7 37 correlation matrix. A separate correlation matrix was generated each session for each mouse. A Fisher transformation yielded the *z*-score association matrices. Two sessions of *z*-score matrices were averaged for graph theory analysis.\n\nGraph Theory Analysis {#section6-1536012118821034}\n---------------------\n\nGraph theory analysis was performed in MATLAB (Ver. R2014a; MathWorks, Inc. Natick, Massachusetts, USA)-based program graph analysis toolbox (<https://www.nitrc.org/projects/gat/>). Binary adjacency matrices were derived by thresholding the 37 \u00d7 37 association matrices at a range of densities (0.01--0.6). Six global network measures (global efficiency \\[GEff\\], local efficiency, SW, Louvian modularity, normalized clustering coefficient, and normalized characteristic path lengths) were calculated at the selected range of densities. To avoid the bias of density selections, a range of densities (0.33--0.4) were used as the target range. The lower boundary was selected at the density when all nodes in the network had at least 1 connection based on the data from the wild-type animals in this study. The upper boundary was set to 0.4 when the averaged SW index was larger than 1.15, based on the range of the SW (eg, 1.1--1.2) used in previous publications.^[@bibr50-1536012118821034][@bibr51-1536012118821034]--[@bibr52-1536012118821034]^ Values of the 6 global network measures were calculated for densities from 0.33 to 0.4 with the step of 0.01. The averages were calculated across the target density range and were used for estimating the effect of genotype, sex, and the genotype \u00d7 sex interaction with 2-way analysis of variance (ANOVA). False discovery rate correction was performed to maintain a 5% type 1 error rate.^[@bibr53-1536012118821034]^ Corrected *P* values were demonstrated as *Q* values. One-way ANOVA was further performed in males and females on measurements that showed significant genotype \u00d7 sex interactions. Post hoc analysis was performed with Tukey multiple comparison test. All of the statistical analyses in this study were performed in SPSS Statistics (Version 23.0; SPSS Inc., Chicago, Illinois).\n\nSeed-Based Analysis {#section7-1536012118821034}\n-------------------\n\nFor each rsfMRI session, the regionally averaged time course of somatosensory cortex (S1 and S2 combined) was extracted and correlated with the time courses of other voxels in the whole brain to create a correlation map with somatosensory cortex as the seed ROI. For each animal, the correlation map was then transformed to a *z*-score connectivity map with the Fisher transformation, averaged across sessions, and subjected to a 1-way ANOVA (3dANOVA) analysis with males and females separated. 3dFWHMx determined the spatial smoothness of error variance, from which 3dClustSim estimated the required minimum cluster size, maintaining a 5% type 1 error rate (indicated as \u03b1 in the [Figure 3](#fig3-1536012118821034){ref-type=\"fig\"}).^[@bibr54-1536012118821034]^ The averaged connectivity *z* scores (from the significant clusters in thalamus) were subjected to 2-sample *t* test for quantitative analysis with respect to genotype. Values are reported as mean \u00b1 standard error of the mean (SEM). Averaged *z* scores were used in the correlation analysis with neurotransmitter level. The connectivity profile of wild-type animals with right somatosensory cortex as the seed ROI is demonstrated in [Figure 1C](#fig1-1536012118821034){ref-type=\"fig\"}.\n\nHigh-resolution proton Magnetic Resonance Spectroscopy Data Analysis {#section8-1536012118821034}\n--------------------------------------------------------------------\n\nHigh-resolution proton magnetic resonance spectroscopy data were fitted using the LCModel package.^[@bibr55-1536012118821034]^ A simulated basis set of model metabolites appropriate for our acquisition parameters was obtained from Stephen Provencher (PhD; LCMODEL Inc., Oakville, ON, Canada; personal communication, November 2012). We based the criteria for selection of the reliable metabolite concentrations on the Cramer-Rao lower bounds (CRLB) with CRLB \u226420% for the rodent brain.^[@bibr56-1536012118821034],[@bibr57-1536012118821034]^ All concentrations were expressed as mean \u00b1 SEM. In the current study, the CRLB values for both GABA (14.23% \u00b1 0.56%) and glutamate (4.89% \u00b1 0.11%) fulfilled the criteria. [Figure 1D](#fig1-1536012118821034){ref-type=\"fig\"} illustrates representative in vivo high-resolution ^1^HMRS spectrum from the right somatosensory cortex and thalamus of 1 male wild-type mouse. We analyzed the levels of GABA, glutamate, and the ratio of glutamate to GABA (glutamate/GABA) with 2-way ANOVA to assess the effects of genotype, sex, and genotype \u00d7 sex interaction. One-way ANOVA was further performed in males and females when significant genotype \u00d7 sex interaction was observed. Post hoc analysis was performed with Tukey test. The significance level was set at *P* \\< .05. Since only females demonstrated significant genotype effect in the seed-based analysis, we performed the correlation between functional connectivity (averaged connectivity *z* scores from the significant cluster in thalamus) and neurotransmitter levels only in females with Pearson correlation test within each group.\n\nData Availability {#section9-1536012118821034}\n-----------------\n\nThe data sets generated and analyzed during the current study are available from the corresponding author upon request.\n\nResults {#section10-1536012118821034}\n=======\n\nNetwork Topology Alteration {#section11-1536012118821034}\n---------------------------\n\nThe average of global network measures in the target density range were significantly different with respect to the genotype effect in 2-way ANOVA ([Figure 2A](#fig2-1536012118821034){ref-type=\"fig\"}), that is, SW (genotype: *F*\\[2, 66\\] = 5.046, *P* = .009, *Q* = 0.024), normalized clustering coefficient (C/C~rand~; genotype: *F*\\[2, 66\\] = 4.745, *P* = .012, *Q* = 0.024; sex: *F*\\[1, 66\\] = 4.143, *P* = .046, *Q* = 0.069), and modularity (ModL; genotype: F\\[2, 66\\] = 5.004, *P* = .009, *Q* = 0.024; sex: *F*\\[1, 66\\] = 5.353, *P* = .024, *Q* = 0.048). Post hoc analysis of the genotype effect showed significantly reduced values in heterozygous *Met--Emx1* mice when compared with wild-type animals (SW: *P* = .006; C/C~rand~: *P* = .008; ModL: *P* = .007) but not in homozygous *Met--Emx1* mice. Significant effect of genotype, sex, and genotype \u00d7 sex interaction was observed in local efficiency (MLocEff; genotype: *F*\\[2, 66\\] = 3.828, *P* = .027, *Q* = 0.04; sex: *F*\\[1, 66\\] = 6.523, *P* = .013, *Q* = 0.039; genotype \u00d7 sex: *F*\\[2, 66\\] = 7.137, *P* = .002, *Q* = 0.012). Therefore, 1-way ANOVA was performed to further investigate the effect of genotype on MLocEff in males and females separately. A main effect of genotype was observed in females with *F*\\[2, 33\\] = 14.678, *P* \\< .001 ([Figure 2B](#fig2-1536012118821034){ref-type=\"fig\"}). MLocEff was significantly increased in female homozygous *Met--Emx1* mice (*P* = .016) and decreased in heterozygous *Met--Emx1* mice (*P* = .047) when compared with wild-type mice. A significant difference between heterozygous and homozygous *Met--Emx1* mice was also observed (*P* \\< .001) that MLocEff in homozygous *Met--Emx1* mice was significantly higher than in heterozygous *Met--Emx1* mice. Only a main effect of sex was observed in the area under the curve of normalized characteristic path lengths (L/L~rand~; *F*\\[1, 65\\] = 8.276, *P* = .005, *Q* = 0.03). No difference in the effect of genotype, sex, or genotype \u00d7 sex interaction was observed in the area under the curve of GEff.\n\n![Loss of *Met* allele changed global measurements. Alterations were observed within the densities between 0.33 and 0.4. Two-way ANOVA showed significant effects of genotype or sex and a genotype \u00d7 sex interaction (A). One-way ANOVA was further performed on MLocEff in both males and females. Only females showed significant genotype effect (B). Values are means \u00b1 SEM. ^\\#^ *P* \\< .05, ^\\#\\#^ *P* \\< .01, ^\\#\\#\\#^ *P* \\< 0.001 main genotype effect; ^\u2020^ *P* \\< 0.05, ^\u2020\u2020^ *P* \\< 0.01 main sex effect; ^\\$\\$^ *P* \\< .01 genotype \u00d7 sex interaction; \\**P* \\< .05, \\*\\**P* \\< .01, \\*\\*\\**P* \\< .001 between group difference. ANOVA indicates analysis of variance; C/C~rand~, normalized clustering coefficient; GEff, global efficiency; HET, heterozygous; HOM, homozygous.; L/L~rand~, normalized characteristic path lengths; MLocEff, local efficiency; ModL, modularity; SEM, standard error of the mean; SW, small-worldness; WT, wild-type.](10.1177_1536012118821034-fig2){#fig2-1536012118821034}\n\nAlteration of Somatosensory Thalamocortical Functional Connectivity {#section12-1536012118821034}\n-------------------------------------------------------------------\n\nThe functional connectivity maps of somatosensory cortex (S1 + S2) were subjected to 1-way ANOVA with males and females analyzed separately. Clusters in thalamus with main effect of genotype were only observed in females, *P* \\< .02 (type 1 error rate, \u03b1 \\< 0.05), as shown in [Figure 3A](#fig3-1536012118821034){ref-type=\"fig\"}. In addition, between-group comparison showed increased somatosensory thalamocortical connectivity in female homozygous *Met--Emx1* mice when compared with wild-type female mice, *P* \\< .01 (type 1 error rate, \u03b1 \\< 0.05) ([Figure 3B](#fig3-1536012118821034){ref-type=\"fig\"}). The averaged connectivity *z* score of clusters with a main effect of genotype significantly increased 189% in female homozygous *Met--Emx1* mice (0.17 \u00b1 0.02) when compared with wild-type females (0.06 \u00b1 0.01), *P* \\< .001. The *z* scores were subsequently used for investigation on the correlation between somatosensory thalamocortical functional connectivity and neurotransmitter levels. Heterozygous mice did not differ significantly from the wild-type mice.\n\n![Impaired Met signaling leads to sex-specific differences in functional connectivity. One-way ANOVA shows significant functional connectivity alterations in thalamus with somatosensory cortex (S1 + S2) as the seed ROI in females but not in males (A). Additional post hoc testing demonstrated clusters of voxels in homozygous female mice that were significantly different to control animals (B). \u03b1 indicates type 1 error rate; ANOVA, analysis of variance; HET: heterozygous, HOM, homozygous; ROI, region of interest; WT, wild-type.](10.1177_1536012118821034-fig3){#fig3-1536012118821034}\n\nExcitatory/Inhibitory Alteration in Met--Emx1 mice {#section13-1536012118821034}\n--------------------------------------------------\n\nOur previous in vitro electrophysiology study implicated impaired *Met* signaling in decreased GABAergic inhibition and increased E/I ratio.^[@bibr16-1536012118821034]^ Magnetic resonance spectroscopy tested the hypothesis that *Met--Emx1* mice have an altered E/I balance, as measured by in vivo glutamate and GABA levels. In the right somatosensory cortex (S1 + S2), no significant alteration in glutamate, GABA, or glutamate/GABA ratio was detected ([Figure 4A](#fig4-1536012118821034){ref-type=\"fig\"}). In the right thalamus, significant genotype \u00d7 sex interaction was observed in glutamate/GABA ratio (sex: *F*\\[1, 66\\] = 5.468, *P* = .022; genotype \u00d7 sex: *F*\\[2, 66\\] = 3.990, *P* = .023; [Figure 4B](#fig4-1536012118821034){ref-type=\"fig\"}). A 1-way ANOVA was performed with males and females separately to assess the differences among genotypes. Only female mice showed a main effect of genotype, with *F*\\[2, 33\\] = 6.483, *P* = .004. Post hoc analysis revealed higher glutamate/GABA ratio in female heterozygous *Met--Emx1* mice when compared with female homozygous *Met--Emx1* mice (*P* = .003). There was a trend toward increased glutamate/GABA ratio in female heterozygous *Met--Emx1* mice when compared with female wild-type mice (*P* = .093; [Figure 4B](#fig4-1536012118821034){ref-type=\"fig\"}).\n\n![Comparison of glutamate and GABA levels. The concentrations of glutamate and GABA and the glutamate/GABA ratio were measured in somatosensory cortex (A) and thalamus (B). Values are means \u00b1 SEM. Two-way ANOVA showed significant effect of sex and genotype \u00d7 sex interaction of Glu/GABA ratio in thalamus. One-way ANOVA was further performed on measurements with significant genotype \u00d7 sex interaction in both males and females. Only females showed significant genotype effect on Glu/GABA ratio in thalamus. ^\\#\\#^ *P* \\< .01 main genotype effect; ^\u2020^ *P* \\< .05 main sex effect; ^\\$^ *P* \\< .05 genotype \u00d7 sex interaction; \\*\\**P* \\< .01 between group difference. \u2041.05 \\< *P* \\< .1 trend of between-group difference. ANOVA indicates analysis of variance; GABA, gamma-aminobutyric acid; Glu, glutamate; HET, heterozygous; HOM, homozygous; SEM, standard error of the mean; WT, wild-type.](10.1177_1536012118821034-fig4){#fig4-1536012118821034}\n\nCorrelation Between Glutamate/GABA Ratio and Functional Connectivity {#section14-1536012118821034}\n--------------------------------------------------------------------\n\nThe averaged connectivity *z* scores from the cluster in thalamus showing significant effect of genotype were used in the correlation analysis with neurotransmitter levels.\n\nCorrelations between thalamic glutamate/GABA ratio and somatosensory thalamocortical functional connectivity were estimated in females, with the 3 genotypic groups analyzed separately ([Figure 5](#fig5-1536012118821034){ref-type=\"fig\"}). Only a significant negative correlation between somatosensory thalamocortical functional connectivity and thalamic glutamate/GABA ratio was observed in heterozygous *Met--Emx1* mice (*r* = \u22120.664, *P* = .019), while this correlation was absent in wild-type and homozygous *Met--Emx1* mice.\n\n![Correlation between thalamic glutamate/GABA ratio and somatosensory thalamocortical functional connectivity. Pearson correlation test revealed significant negative correlation between thalamic glutamate/GABA ratio and functional connectivity between somatosensory cortex and thalamus. GABA indicates gamma-aminobutyric acid.](10.1177_1536012118821034-fig5){#fig5-1536012118821034}\n\nDiscussion {#section15-1536012118821034}\n==========\n\nOur findings reveal several features of altered neural network communication in mice with targeted *Met* receptor inactivation. Decreased SW, normalized clustering coefficient, and modularity in heterozygous *Met--Emx1* mice were observed when compared with wild-type ones. In addition, local efficiency in female mice was significantly increased in homozygous *Met--Emx1* mice and decreased in heterozygous *Met--Emx1* mice when compared with wild-type mice. In female *Met--Emx1* mice, we found (1) a significantly increased functional connectivity between somatosensory cortex and thalamus in homozygous *Met--Emx1* mice when compared with wild-type mice, (2) glutamate/GABA ratio in thalamus is negatively correlated with the somatosensory thalamocortical connectivity in heterozygous *Met--Emx1* mice, and (3) a marginally increased glutamate/GABA ratio in thalamus of heterozygous *Met--Emx1* mice and significantly higher glutamate/GABA ratio in female heterozygous *Met--Emx1* mice when compared with female homozygous *Met--Emx1* mice.\n\nThe reduced network modularity, SW, and normalized clustering coefficient in the current study agree with previous studies involving young autism patients.^[@bibr58-1536012118821034],[@bibr59-1536012118821034]^ Altered clustering coefficient and local efficiency indicated altered brain functional segregation.^[@bibr35-1536012118821034]^ The features of SW of brain functional networks were formed during typical brain maturational processes when functional connectivity tends to weaken locally and strengthen over longer distances.^[@bibr60-1536012118821034][@bibr61-1536012118821034]--[@bibr62-1536012118821034]^ The present results demonstrated that cortical and hippocampal disruptions of HGF-Met signaling during the brain maturational process may impair the maturation of large-scale network community. Hyperconnectivity has been observed in thalamocortical system of individuals with ASD.^[@bibr63-1536012118821034],[@bibr64-1536012118821034]^ Enhanced somatosensory thalamocortical connectivity in the present study provides evidence that the disruption of HGF-Met signaling in the neocortex during development can lead to functional deficit in the somatosensory inhibitory system and contribute to the autistic phenotype.\n\nPrevious studies noted altered E/I balance in ASD, involving both glutamate and GABA (reviewed by Nelson and Valakh).^[@bibr17-1536012118821034]^ Our previous in vitro electrophysiology study^[@bibr16-1536012118821034]^ indicated that the E/I ratio increases due to a dramatic reduction in postsynaptic inhibition, as reflected in decreased amplitude of the spontaneous inhibitory postsynaptic currents in the thalamocortical pathway in *Met--Emx1* mice.^[@bibr16-1536012118821034]^ Our current study showed that although the absolute concentrations of glutamate and GABA were not altered in *Met--Emx1* mice, the glutamate/GABA ratio showed an increase which provides in vivo evidence to support the E/I imbalance theory in kinase-inactive *Met* mice. Loss of *Met* activity, specifically in interneurons, has been reported to impair cognitive and procedural behaviors.^[@bibr8-1536012118821034]^ The negative correlation between the glutamate/GABA ratio in the thalamus with the somatosensory thalamocortical connectivity in *Met--Emx1,* but not in wild-type mice, provides a link between the neurotransmission imbalance and thalamocortical system overconnectivity in this model. Future studies will design appropriate behavior tests to investigate whether *Met--Emx1* mice show impaired thalamocortical system behavior, correlated to neurotransmission imbalance and brain functional overconnectivity.\n\nOther novel findings of the current study are the difference between heterozygous and homozygous *Met--Emx1* mice and between males and females. Sex differences, specifically a \"male bias\" or \"female protective effect,\" have been noticed in previous studies that males are affected more frequently than females.^[@bibr65-1536012118821034],[@bibr66-1536012118821034]^ Recently, there is an emerging awareness that the sex differences may be due to different phenotypic presentation in females and the failure of detecting autistic females by current diagnostic procedures.^[@bibr65-1536012118821034],[@bibr67-1536012118821034],[@bibr68-1536012118821034]^ Autism may manifest itself differently, and in some ways more subtly, in females.^[@bibr69-1536012118821034]^ Recent studies have identified sex-related differences in behavioral characteristics.^[@bibr70-1536012118821034],[@bibr71-1536012118821034]^ But research to date is yet to define or provide a systematic understanding of the female presentation.^[@bibr69-1536012118821034],[@bibr72-1536012118821034]^ In this study, male heterozygous *Met--Emx1* mice presented more alterations than male homozygous *Met--Emx1* mice when compared with wild-type mice in network topology, while both female heterozygous and homozygous *Met--Emx1* mice showed network topological alterations and thalamic E/I imbalance. These findings indicate that fully inactive *Met* in the cerebral cortex and hippocampus does not necessarily result in more alterations than single allele inactivation of *Met.* The lack of differences in homozygous *Met--Emx1* mice may be due to early compensation for the complete lack of *Met* Signaling, due to the inactivated allele. Also, the sex bias that males are more prone to autism-associated impairment may not hold true when focusing on a specific genetic basis.\n\nOur results are consistent with some of the observations from individuals with the *Met* susceptibility alleles^[@bibr10-1536012118821034]^ but not all. The genetic changes are different in each case. In the *Met--Emx1* mice, the Met receptor is inactive, whereas in individuals, the rs1858830 variant is located in a noncoding region and appears to alter *MET* expression. Although the *MET* rs1858830 variant has a strong impact in both heterozygous and homozygous cases, the *Met--Emx1* mice did not show a strong phenotype in the heterozygous group. A possible explanation is that in the *Met--Emx1* mice, *Met* was altered only in the cerebral cortex and hippocampal regions, and the full phenotype may involve subcortical areas and inactivation of *Met* in broader brain areas may generate phenotypes more similar to the human results.\n\nThe present study should be viewed with the caveat that anesthetic isoflurane, also a vasodilator, can alter cerebrovascular activity and can have dose-dependent effects on both task-based and resting-state BOLD responses.^[@bibr46-1536012118821034],[@bibr73-1536012118821034]^ We used the minimum dose of isoflurane (1%) to maintain the immobility of the animals. Imaging with this dose or higher doses has shown preserved brain network activity and network organization in rodents.^[@bibr42-1536012118821034],[@bibr73-1536012118821034],[@bibr74-1536012118821034]^ Therefore, the functional connectivity observed in this study should represent the intrinsic property of functional brain organization. However, the possibility exists that mice with different genotypes respond differently to isoflurane. To resolve this question, future studies testing cerebral blood flow will be necessary.\n\nIn conclusion, our study revealed altered thalamocortical functional connectivity and correlations between neural circuits and neurotransmission in mice with the inactive autism-associated Met tyrosine kinase. The present results provide valuable in vivo evidence to support our hypothesis that aberrant functioning of the somatosensory thalamocortical system is at the core of the conspicuous somatosensory behavioral phenotypes observed in *Met--Emx1* mice and in individuals with ASD. The relative level of glutamate and GABA can play an important role in the functioning of the somatosensory thalamocortical system.\n\n**Authors' Note:** S.T. acquired, analyzed, and interpreted the data; wrote the manuscript; and assisted with final editorial changes. W.Z. assisted with MR data acquisition. F.L. conducted the animal preparations. R.S.E. and E.M.P. designed the MRI experiment and contributed to the writing of the manuscript and interpretation of the data. S.X. assisted significantly with experimental design, MR data acquisition, data analysis, data interpretation, writing of the manuscript, and assisted with final editorial changes.\n\nThe authors are grateful to Dr. Rao Gullapalli for contributing to data interpretation. The authors thank Michelle Monroe and Shuxin Zhao for genotyping and breeding the mice.\n\n**Declaration of Conflicting Interests:** The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.\n\n**Funding:** The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institutes of Health/ National Institutes of Neurological Disorders and Stroke (NS092216).\n"}
+{"text": "\n"}
+{"text": "###### Strengths and limitations of this study\n\n-   This Delphi protocol is informed by the Guidance on Conducting and REporting DElphi Studies recommendations.\n\n-   Expert eligibility is predefined and includes international and multiprofessional representation.\n\n-   This protocol states definitions and a priori criteria for consensus, agreement and stability.\n\n-   The study will use a systematic consensus process to provide expert recommendations on the exercise and dosage variables of an exercise training programme for chronic non-specific neck pain that can be used in clinical practice and future clinical trials.\n\n-   The study results will be specific to chronic non-specific neck pain rehabilitation, limiting the external validity to other musculoskeletal conditions.\n\nIntroduction {#s1}\n============\n\nChronic non-specific neck pain (CNSNP) affects 289\u2009million people worldwide with increasing prevalence.[@R1] The subsequent disability is significant resulting in CNSNP being considered one of the leading causes of years lived with disability.[@R2] Despite multiple guidelines and systematic reviews informing clinical practice,[@R3] patient outcomes are suboptimal, reflected by the increasing rank of CNSNP's cause for global disability-adjusted life years.[@R10]\n\nOne explanation for poor patient outcomes is the vague recommendations of 'exercise' or 'strengthening exercise' from clinical guidelines and systematic reviews that inform clinical practice. Furthermore, exercise dosage recommendations (sets, repetitions, load, frequency and so on) are lacking and considered a research priority.[@R13] Several trials have demonstrated small to very large short-term effects on pain and disability when using exercise training programmes that combine submaximal effort exercises for the deep cervical muscles to improve co-ordination and sequential spinal control (motor control exercises) and exercises for the superficial cervical muscles to improve the ability of the neck to produce, transfer and absorb force (segmental exercises).[@R14] Although packages using a combination of motor control and segmental exercise appear promising, the optimal dosage and long-term effectiveness of this exercise training programme are unknown and require evaluation through an adequately powered low risk of bias clinical trials.\n\nThe Medical Research Council recommends complex interventions such as exercise training programmes to be defined prior to clinical trials.[@R21] The Consensus on Exercise Reporting Template (CERT)[@R22] and the Template for Intervention Description and Replication (TIDieR)[@R23] are reporting checklists that aid in defining and reporting the design of exercise interventions. A key component of each reporting checklist is exercise training variables such as progressive overload, specificity, exercise selection order and dosage, as manipulation of these variables results in different performance and physical outcomes.[@R24] To date, the exercise and dosage variables of an exercise training programme of motor control and segmental exercise for CNSNP cannot be reported in accordance with CERT or TIDieR due to significant heterogeneity or poor reporting of previous trials.[@R14] Consensus on these variables will enable a clearly defined exercise intervention for use in clinical trials and clinical practice.\n\nAims {#s1-1}\n----\n\nThe aim of this e-Delphi study is twofold. Initially, we will obtain exercise and academic professional expert opinion on the exercise and dosage variables of an exercise training programme consisting of motor control and segmental exercise for the management of CNSNP. We will then conduct a systematic process to gain consensus on the exercise and dosage variables reported to inform future research and clinical practice.\n\nMethodology {#s2}\n===========\n\nJustification of Delphi methodology {#s2-1}\n-----------------------------------\n\nExploring exercise variables through qualitative research methodologies such as interviews and focus groups would embrace the diversity of opinion between experts providing rich detail and a deeper understanding, but this diversity maybe problematic when defining an intervention.[@R26] As heterogeneity of motor control and segmental exercise variables already exists across trials, a convergence of information is required. Research methodologies such as Delphi techniques, consensus development conferences and nominal group techniques can all achieve consensus.[@R27] The Delphi method is a systematic approach to achieving consensus among experts through the independent completion of sequential questionnaires that are refined on feedback resulting in a convergence of opinion and eventual consensus.[@R29] A Delphi method is advantageous over other forms of consensus techniques owing to (1) independent and anonymous participation reducing peer pressure and other extrinsic factors present in group techniques that cause subject bias[@R30]; (2) controlled feedback between rounds encouraging consensus by providing participants with the opportunity to refine their opinions[@R29]; (3) using experts with a range of knowledge and experience to improve content validity and response rates[@R32]; and (4) electronic questionnaires removing geographical limitations.[@R29]\n\nDesign {#s2-2}\n------\n\nThis protocol is informed by Guidance on Conducting and REporting DElphi Studies (CREDES)[@R35] ([online supplementary file 1](#SP1){ref-type=\"supplementary-material\"}) and other recommended criteria.[@R36] As no register exists for Delphi research, the protocol has been published to ensure quality, rigour and transparency. The three-round e-Delphi is summarised in [figure 1](#F1){ref-type=\"fig\"}.[@R27] Data collection is planned between March and August 2020. All rounds will be completed electronically and anonymously using REDCap, a secure web application for building and managing online surveys.[@R37] Round 1 will be used to generate statements on exercise and dosage variables for both motor control and segmental exercises. Experts will rate their agreement with statements in rounds 2 and 3 using a 5-point Likert scale. Three rounds are commonly cited to be sufficient to achieve consensus.[@R27] Statements that achieve consensus in round 3 will be used to describe the key exercise and dosage variables for a CNSNP exercise training programme which includes motor control and segmental exercises.\n\n10.1136/bmjopen-2020-037656.supp1\n\n![Delphi study procedures. CERT, Consensus on Exercise Reporting Template; TIDieR, Template for Intervention Description and Replication.](bmjopen-2020-037656f01){#F1}\n\nExpert eligibility and sample {#s2-3}\n-----------------------------\n\nA purposive sampling method will be used to recruit a range of experts in CNSNP exercise prescription. It is recommended that a heterogeneous sample is used representing a spectrum of opinions[@R29] and therefore experts will be recruited from two distinct groups:\n\n-   Exercise professionals: any professional who uses exercise to manage neck pain will be considered (eg, physiotherapists, strength and conditioning coaches, osteopaths and chiropractors). Eligible participants will have a relevant postgraduate qualification or \\>5 years of sports or musculoskeletal experience. Experts will treat \u22655 patients with CNSNP per month using exercise. Experts will be identified through existing professional networks and social media/internet-based searching.\n\n-   Academics: eligible academics will have \u22652 peer-reviewed publications focused on the use of exercise in the management of CNSNP in the past 10 years.[@R39] Academics will be identified through CNSNP systematic reviews/randomised clinical trials published indexed in PubMed and Expertscape searches.[@R13]\n\nExperts will be recruited worldwide, aged 18 or above, able to read and write English and willing to participate. They will be invited to participate by the lead author (JP) through email. Recruitment will be maximised by encouraging identified experts to snowball the invitation with other suitable participants and calls for expressions of interest on social media.[@R40] Upon experts confirming their interest and eligibility, they will be provided with a link to the Research Electronic Data Capture (REDCap) system at the University of Birmingham, where a Participant Information Sheet ([online supplementary file 2](#SP2){ref-type=\"supplementary-material\"}) and Consent Form will be hosted.[@R37] The Participant Information Sheet will clarify study procedures, eligibility criteria, assure anonymity and explain the withdrawal process. Participants may withdraw at any time up until the data analysis of the Round 3 Questionnaire. Due to the nature of the Delphi process, responses will be used up to the point of withdrawal. Participants will be able to withdraw from the study by contacting the Principal Investigator or Primary Academic Supervisor. Consent will be obtained electronically through REDCap.[@R37] Recruitment will continue for 4 weeks with a reminder sent at week 2. Should there be no contact within the 4 weeks then no further communication will be sent.[@R42]\n\n10.1136/bmjopen-2020-037656.supp2\n\nThere is no universal guide to sample size in Delphi studies, and expert panels have ranged from 4 to 3000 participants.[@R43] Previous Delphi studies with an aim of intervention development typically achieved consensus with responses from 10 to 27 experts in the final round[@R39] and therefore, a conservative estimate of 27 final responses is required. Assuming a response rate of 70%, a minimum of 40 experts are required to complete the consent form to ensure at least 27 responses in round 3.[@R29] To prevent over-representation from one expert group, recruitment will be monitored to achieve an approximate 50/50 split between exercise and academic professionals.\n\nProcedure {#s2-4}\n---------\n\n### Round 1 {#s2-4-1}\n\nThe objectives of round 1 will be to obtain participant demographic data and generate statements on exercise and dosage variables based on expert opinion. Participants will complete the Participant Details Form collecting information on professional background, highest qualification, primary country of work, work setting, H-index, publication count, years qualified and grade of clinical work. The Round 1 Questionnaire ([online supplementary file 3](#SP3){ref-type=\"supplementary-material\"}) will consist of open-ended questions informed by CERT and TIDieR.[@R22] Open-ended questions improve the content validity as statements are generated by expert opinion.[@R29] Statements generated from previous clinical trials[@R14] will be included in round 2, rather than round 1 to allow participants to provide their expert opinion without bias from the literature, thereby reducing experimenter bias.[@R34] The questions will ask participants to identify the exercise and dosage variables they consider important when prescribing exercise for CNSNP. They will then be asked to list and explain what patients or other factors may affect or inform their reasoning when prescribing the exercise and dosage variables that they identified. Participants will be asked to answer open-ended questions for both motor control and segmental exercises independently with definitions of both subgroups of exercise provided for clarity ([online supplementary file 3](#SP3){ref-type=\"supplementary-material\"}).[@R48] Participants will have the opportunity to provide general comments at the end. The Round 1 Questionnaire ([online supplementary file 3](#SP3){ref-type=\"supplementary-material\"}) was piloted for feedback on readability, relevance and appropriateness through the Study Steering Group and edited accordingly. Round 1 will be open for 1\u2009month with email reminders, including the withdrawal process, being provided at weeks 1 and 3.[@R49]\n\n10.1136/bmjopen-2020-037656.supp3\n\n### Round 2 {#s2-4-2}\n\nThe objectives of round 2 are to evaluate consensus on statements regarding exercise and dosage variables and to identify any further statements. Participants will be provided with feedback explaining how statements were generated from round 1 and then asked to rate their agreement with the statements using a 5-point Likert scale where 1=strongly disagree and 5=strongly agree.[@R49] A 5-point scale is preferred as it possesses acceptability psychometric properties while being quick and easy for participants to reduce frustration and demotivation.[@R50] An open text box will be included for each statement for any additional comments or further statement generation. All comments will be analysed by the study team and reviewed by the Study Steering Group. All participants will be invited to round 2, including those who did not complete round 1, provided they have not withdrawn from the study. This provides the opportunity for participants to continue their involvement who were unable to complete previous rounds due to time or other commitments.[@R29] As per round 1, the Round 2 Questionnaire will remain active for 4 weeks with email reminders sent at weeks 1 and 3.\n\n### Round 3 {#s2-4-3}\n\nThe objective of round 3 is to strengthen consensus on statements regarding exercise and dosage variables. The Round 3 Questionnaire will include feedback from round 2 using descriptive statistics, promoting reflection before completing the final questionnaire. In round 3, participants will be asked to rate their agreement with the statements achieving consensus from round 2 using the same 5-point Likert scale.[@R39] Statements that do not achieve consensus in round 2 will be discarded. A free-text box will be provided for participants to clarify responses but the generation of new statements will not be encouraged. All participants will be invited to participate in round 3, which will again remain active for 4 weeks with email reminders sent at weeks 1 and 3.\n\nData analysis {#s2-5}\n-------------\n\nQuantitative and qualitative data will be inputted into IBM SPSS Statistics V.25 and QSR International's NVivo V.12 Plus software, respectively, for analyses.[@R51] Data will be analysed independently by two researchers (JP and VT) at each round. The complete agreement between researchers is required for statements to be included, with disagreements resolved by discussion.[@R53] The Study Steering Group will review the data at each stage for feedback and editing before dissemination.\n\n### Round 1 {#s2-5-1}\n\nQualitative data will be examined using a theoretical thematic analysis to generate statements under themes preidentified from CERT/TIDieR and then examined inductively for any new themes.[@R26] Original wording from one expert that best represents the wording across participants with similar statements will be used where possible and all other statements will be discarded.[@R34] Statements generated from previous clinical trial findings not identified from participant responses will also be included[@R14] For any statement to be included, it must be described at least once by any participant or via previous clinical trials; therefore, any stand-alone statements will be kept and included. The Round 2 Questionnaire will be constructed using the statements generated.\n\n### Round 2 {#s2-5-2}\n\nQualitative data will be analysed using thematic analysis for the emergence of any new statements. Descriptive and inferential statistics will be used to evaluate agreement and consensus ([table 1](#T1){ref-type=\"table\"}). Any statements not achieving the a priori criteria for consensus will be discarded (median \u22653; IQR \u22641.5; percentage agreement \u226560%).\n\n###### \n\nDefinitions and statistical measures of consensus, agreement and stability\n\n                         Definition                                                                                                    Statistics                                   Round 2                           Round 3\n  ---------------------- ------------------------------------------------------------------------------------------------------------- -------------------------------------------- --------------------------------- ---------------------------------\n  Consensus              The extent to which the group of experts share the same opinion                                               Median                                       \u22653                                \u22653.5\n  IQR                    \u22641.5                                                                                                          \u22641                                                                             \n  Percentage agreement   \u226560%                                                                                                          \u226570%                                                                           \n  Agreement              A measure of inter-rater agreement where the rating of one expert can be predicted by the rating of another   Kendall's coefficient of concordance (*W*)   Significant agreement (p\\<0.05)   Significant agreement (p\\<0.05)\n  Stability              The consistency of responses between successive rounds                                                        Wilcoxon rank-sum test                       NA                                Significance level p\\<0.05\n\n### Round 3 {#s2-5-3}\n\nDescriptive and inferential statistics will evaluate consensus against a priori criteria (median \u22653.5; IQR \u22641; percentage agreement \u226570%) ([table 1](#T1){ref-type=\"table\"}). Statements achieving consensus after round 3 will be used to describe the key exercise and dosage variables of motor control and segmental exercise training programme. Statements that fail to achieve consensus in round 3 will be discarded.\n\nConsensus, agreement and stability {#s2-6}\n----------------------------------\n\nA discrepancy exists as to the definitions and statistical measures of consensus and agreement within the literature.[@R36] Some argue that consensus and agreement are interchangeable,[@R55] whereas others recommend separate definitions.[@R57] To ensure clarity, the following definitions will be used in this study:\n\n-   Consensus: the extent to which the group of experts share the same opinion.[@R55]\n\n-   Agreement: a measure of inter-rater agreement where the rating of one expert can be predicted by the rating of another.[@R59]\n\n-   Stability: the consistency of responses between successive rounds.[@R55]\n\nConsensus, agreement and stability will be assessed in each round using a combination of descriptive and inferential statistics ([table 1](#T1){ref-type=\"table\"}).[@R36] Consensus will be evaluated using descriptive statistics of central tendency and dispersion. As the Likert scale is considered an ordinal scale,[@R60] median and IQR will be used.[@R55] Percentage agreement, defined as the percentage of responses rated agree/strongly agree, will also be used to evaluate consensus among experts for each statement.[@R39] Progressively increased criteria will be used between rounds 2 and 3 to encourage convergence and strengthen overall consensus.[@R39] Agreement between experts across all items and within categories identified after round 1 will be evaluated using Kendall's coefficient of concordance (*W*) where 0 is no agreement and 1 is perfect agreement.[@R59] The stability of the responses between rounds 2 and 3 will be evaluated using the Wilcoxon rank-sum test.[@R55] Statistical significance will be set at p\\<0.05.\n\nData management {#s2-7}\n---------------\n\nAll personal information and data will be kept secure from any third party using a password-protected computer during the study. Only members of the study team will have access to the study data. On completion of the study, the data will be kept securely for 10 years in the School of Sport, Exercise and Rehabilitation Sciences, University of Birmingham, UK before being securely destroyed in accordance with University guidelines.\n\nStudy steering group {#s2-8}\n--------------------\n\nThe Study Steering Group will provide study oversight with members consisting of coauthors and patient/public, methodological and clinical expertise ([table 2](#T2){ref-type=\"table\"}). The Study Steering Group will meet at key stages throughout the study to provide feedback on questionnaire development, structure and clarity; aid in expert identification; review study results at each round and agree statement inclusion; review study conduct and aid in the dissemination of findings. Members of the Study Steering Group who are not coauthors will not have access to raw data or be able to influence the study process. Feedback and changes suggested by the Study Steering Group must be agreed between the study coauthors before implementation.\n\n###### \n\nStudy steering group members, backgrounds and roles\n\n  Background                                  Professional title                      Role\n  ------------------------------------------- --------------------------------------- ----------------------------------------\n  Patient                                     NA                                      Co-chair/patient representative\n  Academic nurse                              Lecturer                                Co-chair/methodological representative\n  Clinical physiotherapist                    Consultant physiotherapist              Clinical representative\n  Patient                                     NA                                      Patient representative\n  Academic physiotherapist                    Senior lecturer                         Primary supervisor\n  Academic physiotherapist                    Reader in musculoskeletal sciences      Secondary supervisor\n  Clinical physiotherapist                    Physiotherapist                         Co-investigator\n  Clinical academic trainee physiotherapist   Pre-doctoral clinical academic fellow   Principle investigator\n\nNA, not applicable.\n\nEthics {#s2-9}\n------\n\nAppropriate ethical approval has been granted from the University of Birmingham Ethics Committee (Ref: ERN_19--1857). Informed consent will be received from all participants before completing any questionnaires. They will be informed of the withdrawal process and that any feedback will be anonymised for privacy.\n\nPatient and public involvement {#s2-10}\n------------------------------\n\nThe study was conceived from our clinical working with patients with spinal complaints over many years and their views used to highlight the relevance of this research. Two patients were involved in reviewing the findings of the original systematic review that underpinned this study, suggesting alternatively terminology that better reflects patient views. Our Study Steering Group patient representatives helped refine the research aim of this study as well as contributing to the design of Participant Information Sheets, expression of interest emails/social media posts and developing the Round 1 Questionnaire. It is anticipated that our patient representatives will continue to co-chair the Study Steering Group, review study results at each round and study conduct. Our patient representatives will be instrumental in future dissemination of findings to patient cohorts, as well as informing future fellowship applications for the lead author (JP). Patient and public involvement in the full study will be reported using the Guidance for Reporting Involvement of Patients and the Public2-short form (GRIPP2-SF) when disseminating the study results.[@R62]\n\nDiscussion {#s3}\n==========\n\nThis Delphi study will provide expert consensus on the exercise and dosage variables of an exercise training programme consisting of motor control and segmental exercises for CNSNP that could not be determined from the current literature. Conducting an e-Delphi allows the development of expert informed recommendations from a range of worldwide experts who can participate anonymously, which should be considered a strength. The expert eligibility criteria could be considered a limitation of the study as it may exclude experts in exercise prescription who see a small volume of patients experiencing neck pain. However, there is currently no clear guidelines as to how best define an expert.[@R27] The strict eligibility criteria used are important to ensure that findings are appropriate to CNSNP as it is currently unknown whether approaches to exercise are transferable between different musculoskeletal conditions. Future research will require the acceptability and feasibility of exercise and dosage variables to be evaluated by patient and physiotherapists focus groups. Results will inform the development of an intervention that will be defined using CERT/TIDieR and evaluated in a low risk of bias, adequately powered clinical trial investigating long-term outcomes and optimal dosage.\n\nSupplementary Material\n======================\n\n###### Reviewer comments\n\n###### Author\\'s manuscript\n\nThis work was completed as part of an NIHR/HEE Pre-doctoral Clinical Academic Fellowship. Academic support provided from the University of Birmingham and the Centre of Precision Rehabilitation for Spinal Pain. The authors would like to thank the clinical, methodological and patient experts who have agreed to be members of the Study Steering Group.\n\n**Twitter:** \\@PhysioPrice, \\@abrushton, \\@HeneghanNicola\n\n**Contributors:** JP: is the CI leading the protocol development, analyses and dissemination. NRH and AR: supervisors for the study, ensuring rigour and quality. JP and VT: will independently complete data analysis. All authors have contributed to the design and development of the protocol and have contributed to manuscript draft; read, provided feedback and approved the final manuscript.\n\n**Funding:** This Delphi study did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sectors. It will be conducted as part of JPs MRes research project through the University of Birmingham. The MRes is a component of JP's academic training funded through the Health Education England/National Institute of Health Research's Pre-doctoral Clinical Academic Fellowship.\n\n**Competing interests:** None declared.\n\n**Patient and public involvement:** Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.\n\n**Patient consent for publication:** Not required.\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n"}
+{"text": "Introduction {#s01}\n============\n\nPrimary cilia are homeostatic sensory organelles driven by intraflagellar transport (IFT) that integrate multiple and diverse cellular inputs critical to normal development and organ homeostasis. Diverse environmental cues affect their assembly and disassembly in a manner that is tightly coupled to the actin cytoskeleton and changes in gene expression. How cells coordinate these processes to effect a coherent cellular response remains unclear.\n\nActin network architecture is increasingly recognized as a major driver of ciliogenesis in cycling cells ([@bib43]). Chemical and genetic perturbations of multiple actin remodeling factors have been found to promote ciliogenesis or affect ciliary length ([@bib25], [@bib26]; [@bib7]; [@bib18]; [@bib21]) and intriguingly rescue ciliogenesis defects in *Ift88* ([@bib25]) and *IFT121* ([@bib16]) mutant cells. Actin-myosin--mediated contraction participates in ciliary abscission as the neuronal differentiation program is triggered ([@bib13]). Leptin elongates hypothalamic cilia via transcriptional regulation and actin destabilization ([@bib21]), whereas the microRNA *miR-129-3p* regulates cilia assembly through the concomitant transcriptional silencing of centrosomal proteins and repression of actin filament formation ([@bib7]). Altogether, these findings highlight the existence of concurrent regulatory mechanisms coordinating ciliogenesis, actin dynamics, and transcription.\n\nInterestingly, mouse mutants for the histone *N-*lysine demethylase *Kdm3a* (*Jmjd1a*, *Tsga*, or *Jhdm2a*) are obese, predisposed to diabetes ([@bib20]; [@bib52]), and have male infertility phenotypes ([@bib36]) overlapping with mouse models of dysfunctional cilia. Structural abnormalities of the acrosome and manchette were observed in mutant *Kdm3a* mouse spermatids ([@bib22]) similar to those reported for conditional inactivation of *Ift88* ([@bib24]; [@bib49]). KDM3A localizes to the actin-rich acrosome--acroplaxome region, which is altered in *Kdm3a* mutant spermatids ([@bib22]). In spite of these phenotypic and functional overlaps, a link between KDM3A and ciliogenesis has not been investigated.\n\nKDM3A is part of the large family of Jumonji C (JmjC) domain--containing proteins ([@bib11]; [@bib27]; [@bib33]). Deregulated expression of KDM3A is associated with colorectal cancer ([@bib55]), breast cancer ([@bib56]), and hepatocellular carcinoma ([@bib59]). Hypoxia ([@bib4]; [@bib45]; [@bib57]) and the HSP90 chaperone machinery ([@bib22]) are reported to modulate KDM3A levels and activity, whereas its interaction with chromatin is modulated by cold-regulated phosphatases ([@bib1]). Overall, these findings indicate that KDM3A plays fundamental roles in organismal homeostasis, integrating multiple environmental cues with physiological states.\n\nKDM3A presents different regulatory modalities and subcellular distributions. Through the catalytic removal of mono- and *N-*dimethyl groups from lysine 9 of histone H3, KDM3A acts as a transcriptional activator ([@bib60]) of multiple genes involved in spermatogenesis ([@bib36]), fat storage ([@bib37]), and energy expenditure ([@bib20]) as well as of diverse types of cancers ([@bib35]; [@bib62]). More recently, scaffolding and noncatalytic modes of transcriptional regulation by KDM3A have emerged wherein KDM3A binds to the SWI--SNF (Switch-sucrose nonfermentable) chromatin-remodeling complex ([@bib1]) or Hedgehog-responsive transcription factor GLI1 ([@bib50]), regulating expression of target genes. Although these functions relate to the role of KDM3A in the nucleus, KDM3A exits the nucleus in response to mechanosensitive stimuli ([@bib23]) and is found at various cytoplasmic sites of somatic cells and during germ cell development ([@bib36]; [@bib61]; [@bib59]; [@bib22]). Altogether, these findings indicate that KDM3A is a multifunctional protein with highly regulated subcellular distributions and nontranscriptional roles that remain to be explored.\n\nHere, we present evidence on the regulation of actin dynamics by KDM3A and the subsequent impact on IFT and ciliogenesis. KDM3A is required for the stabilization of primary cilia and regulates cilial responsiveness to environmental cues. RNA-sequencing, proteomic, and two-hybrid studies reveal a dual role for KDM3A in this integrated response modulating ciliogenesis. On a global level, it senses extracellular signaling and transcriptionally regulates the free pool of actin available. On a very local level, KDM3A directly binds to the actin cytoskeleton, creating a responsive \"actin gate\" and regulating IFT within cilia and periciliary compartments. Importantly, deregulated expression of an IFT protein, IFT81, coupled with actin depolymerization phenocopies cilial defects of *KDM3A* mutants, demonstrating that KDM3A-mediated cytoskeletal changes tightly regulate IFT during ciliary growth.\n\nResults {#s02}\n=======\n\nKDM3A plays a conserved role in the regulation of ciliary dynamics {#s03}\n------------------------------------------------------------------\n\nDuring the analysis of a deletion mutant mouse model of *Kdm3a* lacking the catalytic domain JmjC (\u0394JC; [@bib22]), we observed that flagella of mutant spermatids were consistently shorter than those of control littermates ([Fig. 1, A--C](#fig1){ref-type=\"fig\"}). We then asked whether KDM3A played a broader role in ciliogenesis, examining primary cilia in somatic *Kdm3a^\u0394JC^* cells. Mouse embryonic fibroblasts (MEFs) derived from *Kdm3a^\u0394JC^* mutants show abnormal ranges of ciliary lengths ([Fig. 1 D](#fig1){ref-type=\"fig\"}), which can be partially rescued by transfecting a short cytoplasmic isoform of KDM3A ([Fig. 1 E](#fig1){ref-type=\"fig\"}, KDM3A~i2~-EGFP; [@bib22]). *Kdm3a^\u0394JC/\u0394JC^* MEFs also displayed accumulations of IFT88 at the ciliary base and tip compared with the uniform distribution of IFT88 in wild-type cilia ([Fig. 1, F and G](#fig1){ref-type=\"fig\"}). These observations suggested that KDM3A could play a role in axonemal dynamics and IFT in primary cilia. Because *Kdm3a^\u0394JC^* mutant cells encode a truncated KDM3A protein, albeit at very low levels ([@bib22]), we generated *KDM3A^null^* cells in the immortalized human retinal epithelium hTERT RPE1 cell line. RPE1 cells have been thoroughly characterized to study ciliary growth, resorption, and axoneme length in response to cell density and serum content of the culture media ([@bib46]; [@bib26]).\n\n![**Axonemal lengths are disrupted in *Kdm3a* mutant mice.** (A) Wild-type (WT) and *Kdm3a^\u0394JC/\u0394JC^* mutant spermatozoa from mouse epididymis; the latter have rounded heads and short flagella. (B) Schematic of the gene-targeted deletion of *Kdm3a^\u0394JC^* mutant mouse model. Domains: yellow, C6-type zinc finger (ZF); blue, LXXLL nuclear receptor binding (NR); red, JmjC domain (JmjC). (C) Mean \u00b1 5th--95th percentile of flagellar lengths of epididymal spermatozoa derived from *N* \u2265 3 per genotype; *n* is indicated below each box. (D) Mean \u00b1 5th--95th percentile of MEF ciliary lengths of MEFs (*N* \u2265 3 per genotype) after 24-h serum depletion. (E) Mean \u00b1 5th--95th percentile of MEF ciliary lengths transfected with GFP vector (control \\[CO\\]) or GFP-KDM3A~i2~, showing partial rescue of the mutant short cilia phenotype. (F) Immunostaining illustrates ciliary accumulations of IFT88 in *Kdm3a^\u0394JC^* mutant MEFs. (G) Ratio of averaged fluorescence intensity of IFT88 relative to acetylated \u03b1-tubulin measured from the basal body to the ciliary tip along ciliary axonemes. Error bars represent \u00b1 SEM for \\>10 cilia per genotype. Asterisk denotes significant p-value from *t* test: \\*, P \\< 0.05; \\*\\*\\*, P \\< 0.001; \\*\\*\\*\\*, P \\< 0.0001.](JCB_201607032_Fig1){#fig1}\n\nUsing CRISPR/Cas9 gene editing directed to exon 22 of *KDM3A*, we generated independent RPE1 cell lines null for *KDM3A*. Lines 1, 2, and 8 (*KDM3A^null^*) did not express KDM3A protein ([Fig. 2 A](#fig2){ref-type=\"fig\"}) and were found to bear compound *KDM3A* mutations by genotyping (Fig. S1, A and B). As controls, we used the parental RPE1 line (*RPE1^WT^*) as well as an RPE1 sister line that has undergone CRISPR/Cas9 editing but was wild type for *KDM3A* (*RPE1^CrWT^*). Cells were plated at equal densities and promoted either to (a) ciliate, by serum withdrawal (\u2212FCS), or (b) resorb cilia, by serum replenishment (+FCS), with or without temperature stress. No differences were observed in the cilial length between controls and *KDM3A* null cultures after 3 or 24 h without serum ([Fig. 2 B](#fig2){ref-type=\"fig\"}, 3 h and 24 h \u2212FCS). In contrast, under resorption conditions, *KDM3A^null^* cilia remained significantly longer than those of controls ([Fig. 2 B](#fig2){ref-type=\"fig\"}, resorption +FCS), suggesting that *KDM3A^null^* cells failed to efficiently resorb cilia. Moreover, the proportion of *KDM3A^null^* cells that remained ciliated was unchanged after serum replenishment, in contrast to the efficient resorbtion of ciliated control cells ([Fig. 2 C](#fig2){ref-type=\"fig\"}).\n\n![**KDM3A is a negative regulator of ciliogenesis.** (A) Total cell extracts of wild-type (*RPE1^WT^*) and *KDM3A* gene-edited RPE1 mutants (*KDM3A^Cr22^*, named herein *KDM3A^null^*) immunoblotted with KDM3A, revealing *KDM3A*-null clones (1, 2, and 8). (B) Mean \u00b1 5th--95th percentile of ciliary lengths (*N* = 4) of wild-type lines (*RPE1^WT^*, *RPE1^CrWT^*) and the *KDM3A^null^* lines (1,2,8) after 3 h or 24 h without serum (\u2212FCS) and after 24 h \u2212FCS followed by 15-min temperature stress (42\u00b0C) in the presence of serum or 90-min serum only. Dots represent cilial lengths outside the indicated percentile (*n* \\> 100 cilia per experiment). One-way analysis of variance, \\*\\*\\*\\*, P \\< 0.0001; paired *t* test, \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001; ns, not significant. The F-test illustrates the length variance between *RPE1^WTs^* and *KDM3A^nulls^* for each ciliary phase. (C) Percentages of ciliated cells after 24 h \u2212FCS at the indicated times after serum replenishment. (D) Segmented nuclei and cilia (stained with anti-ARL13B) of *RPE1^CrWT^* and *KDM3A^null^* (line 1) cultures after 3-h serum depletion highlights differences in the frequency of ciliated cells (41%, SD = 7.6; 64%, SD = 7.02, respectively), in spite of identical cell densities. Percentages are from *n* \\> 100 cells per experiment done in triplicate. \u03c7^2^ \\< 0.001. (E) Wide-field images illustrating the extreme range of lengths often observed in exponentially growing *KDM3A^null^* lines (24 h FCS), with frequent abnormal ciliary bulges in mutants (arrow). (F) Percentage of cycling cells (24 h +FCS) that contain a cilium in control (*RPE1^WT^* and *RPE1^CrWT^*) and *KDM3A^null^* (lines 2 and 8) cells (*N* = 3). \\*\\*\\*, P = 0.001 is from \u03c7^2^ test. (G) Scoring of cilia found in late G2 (24 h +FCS or during serum induced ciliary resorption). WTs represent *RPE1^WT^* and *RPE1^CrWT^*, and *KDM3A^null^* lines 1, 2, and 8. P-value is from \u03c7^2^ test. (H) Example of a mutant cilium remaining after centrosomes have separated (\\>2 \u00b5m). (I) Cell cycle profile of cycling (24 h +FCS) control (*RPE1^WT^* and *RPE1^CrWT^*) and *KDM3A^null^* (lines 1, 2, and 8) cultures. Error bars represent SD.](JCB_201607032_Fig2){#fig2}\n\nIt was noted during these studies that *KDM3A^null^* cilia displayed a wider range of lengths than *RPE1^WT^* under all conditions ([Fig. 2 B](#fig2){ref-type=\"fig\"}, analysis of variance F). Also 3 h after serum withdrawal, *KDM3A^null^* cultures had a higher proportion of ciliated cells than controls ([Fig. 2 D](#fig2){ref-type=\"fig\"}). Together, these results suggest that mammalian KDM3A negatively regulates ciliogenesis.\n\nCiliogenesis is usually promoted through serum withdrawal, but cycling RPE1 cells also form cilia in the presence of serum, albeit at lower frequencies ([@bib7]). To confirm the propensity of *KDM3A^null^* cells to ciliate, we examined ciliogenesis in exponentially growing cultures. In the presence of serum, some extremely long cilia with abnormal bulges were observed in *KDM3A^null^* cells ([Fig. 2 E](#fig2){ref-type=\"fig\"}, arrow). Importantly, higher proportions of ciliated cells were consistently observed in mutants ([Fig. 2 F](#fig2){ref-type=\"fig\"}), in spite of comparable cell density and mitotic index to controls (*RPE1^WT+CrWT^* 4.05%, SD = 1.90; *KDM3A^null^* 6%, SD = 1.63; *N* = 3 biological replicates, *n* \\> 300 cells scored/condition). A higher proportion of *KDM3A^null^* cells retained cilia into late G2 ([Fig. 2, G and H](#fig2){ref-type=\"fig\"}), whereas most wild-type cilia were already resorbed ([@bib51]). However, these changes were neither caused by nor associated with alterations in cell cycle progression ([Fig. 2 I](#fig2){ref-type=\"fig\"}). As we did not observe mutant mitotic cells with cilia, we conclude that mutant cilia must be eventually resorbed albeit with altered kinetics.\n\nBoth *Kdm3a^\u0394JC^* deletion and *KDM3A^null^* mutant models indicate that KDM3A plays a role in axoneme dynamics, yet the deletion mutants have on average shorter cilia ([Fig. 1 D](#fig1){ref-type=\"fig\"}) rather than the wider range of lengths observed in *KDM3A^null^* cells ([Fig. 2 B](#fig2){ref-type=\"fig\"}). These contrasting outcomes may be caused by a gain of function of the truncated KDM3A protein encoded by *Kdm3a^\u0394JC^* mutants, as MEFs derived from a second hypomorphic mouse model of *Kdm3a* (*Kdm3a^GT^*; [@bib22]) present, as observed in *KDM3A^null^* RPE1 cells, a wider range of cilial lengths (Fig. S1, C and D). Ciliary resorption in response to serum replenishment and temperature stress was similarly delayed in *Kdm3a^GT/GT^* MEFs (Fig. S1 E).\n\nIn summary, given the increased proportion of ciliated cells, abnormal range of cilial lengths, and delayed cilial resorption kinetics, our loss-of-function mutants suggest that in the absence of KDM3A, ciliogenesis is facilitated and may be less tightly regulated. Conversely, overexpressing KDM3A reduces cilia formation in the parental *RPE1^WT^* line (Fig. S1, F and G).\n\nKDM3A regulates actin expression and binds actin in vivo and in vitro {#s04}\n---------------------------------------------------------------------\n\nGiven KDM3A's localization to actin-rich structures and altered biochemical properties of actin in *Kdm3a* mutant spermatids ([@bib22]), we hypothesized that the effect of KDM3A depletion on ciliogenesis is a consequence of actin dysregulation. As KDM3A had been shown to bind the promoter and regulate the expression of smooth muscle \u03b1-actin (*Acta2*; [@bib32]), these effects could be caused by altered transcription or protein--protein interactions or both. Indeed, phalloidin staining of *KDM3A^null^* cells ([Fig. 3 A](#fig3){ref-type=\"fig\"}) showed consistently decreased abundance of actin filaments compared with *RPE1^WT^* ([Fig. 3 B](#fig3){ref-type=\"fig\"}), and this was also evident in *Kdm3a^\u0394JC/\u0394JC^* MEFs (Fig. S2 A).\n\n![**KDM3A regulates the expression of and interacts with actin components.** (A) Schematic of experiments used to isolate RNA and protein lysates during ciliary growth and resorption. (B) Phalloidin staining and histograms of pixel intensities reveal decreased abundance of stress fibers in *KDM3A^null^* cells. (C) Principal-component analysis of total cell proteomes from *RPE1^WT^* and *KDM3A^null^* cells during ciliary growth and resorption show that variability between growth conditions (43.8%) is larger than that between genotypes (15.7%). (D) Venn diagram illustrates the number of significantly down-regulated genes in *KDM3A^null^* compared with *RPE1^WT^* in the transcriptomic and proteomic datasets during ciliary growth and resorption. Dysregulation of extracellular environment is the most enriched term by GO analysis (Table S4). (E) Label-free quantification (LFQ) intensities for peptides specific to ACTA2 and ACTG2 (nonseperable because of sequence homology) across the indicated growth conditions obtained from total cell proteomes of control (*RPE1^WT^* and *RPE1^CrWT^*) and *KDM3A^null^* (line 2) lines. \\*\\*, P \\< 0.001; ns, not significant. (F) Proteomic-based identification of endogenous components significantly enriched in KDM3A pull-downs during ciliary growth (3 h \u2212FCS). P-values are from *t* tests of label-free quantifications from *RPE1^WT^* relative to *KDM3A^null^* from three experimental replicates. Select isolated actin-binding proteins are highlighted. (G) Schematic of the KDM3A region used as bait (JMJC domain in red) and ACTA2 prey identified in the two-hybrid screen. (H) Independent colonies (one to three) of rescued bait/prey plasmid cotransformed onto two-hybrid Gold yeast and grown in double selection (Leu/Trp) were resuspended and assessed for aureobasidin A resistance and \u03b1-Gal activity. Only yeast cotransformed with JmjC and ACTA2 grew in the stringent conditions indicating a binary interaction. (I) Pure smooth muscle actin resolved in 4--12% acrylamide gel (three serial dilutions) transferred to nitrocellulose is bound by GST-JMJC, but not GST.](JCB_201607032_Fig3){#fig3}\n\nThe turnover of the actin cytoskeleton is both transcriptionally and posttranscriptionally regulated ([@bib38]), requiring a pool of actin monomers (G-actin) and multiple actin-remodeling factors that determine F-actin formation ([@bib5]). We therefore investigated both transcriptional changes and protein--protein interactions that could contribute to KDM3A ciliary traits. To address the global transcriptional changes that occur in the absence of KDM3A during ciliary growth and resorption, we used a combined approach of RNA sequencing and quantitative label-free mass spectrometric analysis (Tables S2 \\[raw datasets\\] and S3 \\[thresholds\\]). Principal-component analysis of *RPE1^WT^* and *KDM3A^null^* proteomes first showed that differences between samples primarily reflected serum content of the culture media as opposed to genotypes of the cell lines ([Fig. 3 C](#fig3){ref-type=\"fig\"}). This supports previous studies reporting that KDM3A does not cause large global transcriptional changes but is instead a fine regulator of gene expression ([@bib36]). Subsequently, considering that KDM3A is a transcriptional activator ([@bib60]), we identified genes for which protein or RNA abundance were reduced in *KDM3A^null^* cells compared with *RPE1^WT^* controls (Table S3). GO(Gene Ontology) analysis suggests that the transcriptional activity of KDM3A is required to sense the extracellular environment ([Fig. 3 D](#fig3){ref-type=\"fig\"} and Table S4). In agreement with previous findings ([@bib32]), direct interrogation of actin expression in these datasets reveals that smooth \u03b1-actin *ACTA2* expression (Table S3) as well as protein levels ([Fig. 3 E](#fig3){ref-type=\"fig\"}) failed to be up-regulated in response to serum replenishment. A reduced abundance of ACTA2 in mutants could potentially compromise actin polymerization by limiting available actin monomers during ciliary resorption (+FCS). However, as actin levels are normal in *KDM3A^null^* cells in the absence of serum, transcriptional control alone cannot explain perturbations of actin dynamics observed in mutant cells during ciliary growth (\u2212FCS).\n\nNext, we investigated a potentially direct effect of KDM3A on the cytoskeleton through protein interactions using KDM3A immunopurification of total cell extracts followed by mass spectrometry. In parallel, we performed a subsaturation two-hybrid screen using the JMJC domain of KDM3A as bait. The KDM3A antibody significantly enriched 166 proteins from *RPE1^WT^* ([Fig. 3 F](#fig3){ref-type=\"fig\"} and Table S3), where GO analysis revealed actin binding and actin cytoskeleton as the most enriched terms (Table S4). Strikingly, these immunopurifications showed that endogenous KDM3A copurified with ACTA2 and the actin-capping protein CAPZB, both of which were also found in the two-hybrid screen. A binary interaction between ACTA2 and the JMJC domain of KDM3A was observed by yeast two-hybrid screening ([Fig. 3 H](#fig3){ref-type=\"fig\"}), but not with CAPZB (Fig. S2 B). Further confirmation of a direct association between KDM3A and ACTA2 was supported through binding of purified GST-JMJC to pure smooth muscle \u03b1-actin ([Figs. 3 I](#fig3){ref-type=\"fig\"} and S3 C). The multiple actin-related proteins that copurified with endogenous KDM3A may thus be indirectly enriched through KDM3A binding to actin filaments.\n\nIn summary, global analysis of the transcriptional and total proteomic profiles of cells lacking KDM3A show enrichment of GO terms that relate to the extracellular environment that are in line with the increasingly recognized role of KDM3A in environmental adaptation and mechanosensation ([@bib8]; [@bib23]; [@bib41]). Immunopurification of endogenous KDM3A, two-hybrid screening, and far--Western blotting revealed that KDM3A interacts with actin in vivo and in vitro. Together with the transcriptional failure of *KDM3A^null^* cells to induce actin expression in response to serum replenishment, these results reveal a novel role for KDM3A in the modulation of actin dynamics that may underpin phenotypic changes of *KDM3A^null^* cells.\n\nKDM3A modulates actin dynamics {#s05}\n------------------------------\n\nTo characterize the functional interaction between KDM3A and actin, we further investigated cellular phenotypes known to depend on actin dynamics.\n\nIt was recently shown that KDM3A expression is up-regulated when cells are in suspension ([@bib41]) and that its cytoplasmic localization depends on mechanosensitive signals ([@bib23]). Indeed, Halo-tagged KDM3A is cytoplasmic and enriched in cortical actin filaments and cellular edges immediately after attachment of suspended cells ([Fig. 4 A](#fig4){ref-type=\"fig\"}). We reasoned that the fast wave of actin polymerization ([Fig. 4 B](#fig4){ref-type=\"fig\"}) that follows early spreading over solid substrates ([@bib58]) would thus provide functional evidence on the impact of cytoplasmic KDM3A on actin dynamics. Wild-type (*RPE1^WT^* and *RPE1^CrWT^*) and *KDM3A^null^* (lines 1, 2, and 8) cultures were trypsinized and maintained in suspension in serum-free media for 2 h before plating ([Fig. 4 C](#fig4){ref-type=\"fig\"}) or plated directly ([Fig. 4 D](#fig4){ref-type=\"fig\"}) onto plastic plates. We found that in both cases, a greater proportion of *KDM3A^null^* cells remained unspread compared with control cultures ([Fig. 4, C and D](#fig4){ref-type=\"fig\"}; P0/P1 ratio \u03c7^2^ = 0.001), indicating that loss of KDM3A interferes with the early stages of cell spreading that require fast actin polymerization.\n\n![**KDM3A is required for actin polymerization.** (A) 60 min after plating on glass coverslips, 3D reconstruction reveals cytoplasmic distribution of Halo-KDM3A with strong presence at cellular edges and apical cap. (B) Schematic of stages requiring fast actin polymerization occurring from initial attachment (Phase 0 \\[P0\\]) to early spreading (P1). P2 represents fully spread cells. (C) Bright-field image and number of *RPE1^WT^* and *KDM3A^null^* cells that remain round (P0) or began spreading (P1) over plastic plates after suspension in the absence of serum for 2 h. Red arrow and green encircling indicate cells at P0, defined by birefringence at edges, and P1, defined by flattened edges, respectively. (D) Proportion of cells that remain in P0 when cultures are plated immediately after trypsinization (*N* = 3, *n* \\> 50 cells scored/genotype). Error bars: mean \u00b1 SD. \\*, P \\< 0.0001, from \u03c7^2^. (E) Wide-field images illustrate differences in cellular edges between *RPE1^crWT^* and *KDM3A^null^*. (F) Schematic of actin filament capping or branching leading to changes in cell morphology. As KDM3A binds actin, the rounder cellular edges of *KDM3A* cells suggest an imbalance between capping (yellow) and ARP2/3 mediated branching (green). (G) Median and range of the distance migrated by attached *RPE1^WT^* cells and *KDM3A^null^* (lines 1, 2, and 8; 4 h after plating). Dots represent measurement of distance moved by a single cell. . (H) Confocal images illustrate the distinct cellular edges of *RPE1^WT^* cells transfected with Halo vector or Halo-KDM3A. (I) Hyperfilopodia-like projections are most prominent at the apical pole of Halo-KDM3A transfected cells (plane 2 in diagram). (J) Median and range of membrane lengths at cellular edge of Halo transfected cells measured as described in Fig. S2 D. *N* = 2 transfections. Statistics from G and J are from unpaired *t* tests.](JCB_201607032_Fig4){#fig4}\n\nIn the presence of serum, *KDM3A^null^* cells consistently have rounder cellular edges compared with *RPE1^WT^* ([Fig. 4 D](#fig4){ref-type=\"fig\"}). Branching or elongation of actin fibers is dynamically driven by the antagonism between ARP2/3 nucleating activity and capping proteins, ultimately determining the length of the actin fiber ([@bib14]). At the cellular edge, this interplay is manifested as the extent of lamellipodia to filopodia formation ([Fig. 4 E](#fig4){ref-type=\"fig\"}), where the underlying actin-remodeling events steer cell migration ([@bib44]; [@bib28]). Consistent with these findings, *KDM3A^null^* cells presented reduced cell motility ([Fig. 4 F](#fig4){ref-type=\"fig\"}) and prominent leading edges (Fig. S2 D). Conversely, overexpressing KDM3A increases the proportion of cellular membrane ending in spike/filopodia-like protrusions ([Fig. 4, H--J](#fig4){ref-type=\"fig\"}; and Fig. S2 E).\n\nTogether, the morphological and behavioral features found in the absence or overexpression of KDM3A demonstrate that KDM3A modulates global actin dynamics and suggest that KDM3A binding to actin may compete with or facilitate the binding of different actin-remodeling proteins.\n\nAltered actin dynamics underlies the ciliary traits of *KDM3A*-null cells {#s06}\n-------------------------------------------------------------------------\n\nTo confirm that a deregulated actin cytoskeleton in the absence of KDM3A underlies the abnormal ciliary phenotypes observed in mutants, we aimed to rescue these phenotypes using chemical modulators of F-actin formation. Depolymerization of F-actin promotes ciliary growth ([@bib25], [@bib26]). Conversely, actin polymerization is thought to be required for ciliary resorption ([@bib31]).\n\nInterfering with actin dynamics through cytochalasin D (CytoD) treatment promotes ciliogenesis in cycling RPE1 cells without the requirement of serum withdrawal ([@bib7]). Indeed, CytoD treatment alone is sufficient to rescue ciliogenesis in genetic mutants affecting IFT machinery ([@bib25]; [@bib16]). Surprisingly, we instead observed that *KDM3A^null^* cells accumulated abnormal phalloidin-stained foci ([Fig. 5 A](#fig5){ref-type=\"fig\"}, 2 \u00b5M CytoD) and have decreased proportion of ciliated cells ([Fig. 5 B](#fig5){ref-type=\"fig\"}). Live-cell imaging with a ciliary marker (5-HT6-GFP; [@bib3]) suggests that the decreased proportion of *KDM3A^null^* ciliated cells in 0.5 \u00b5M CytoD is likely caused by reduced ciliary stability, as cilia break off the cell body ([Fig. 5 C](#fig5){ref-type=\"fig\"}), without the net increase in ciliary length observed in control cultures ([Fig. 5 D](#fig5){ref-type=\"fig\"}). CytoD treatment of cycling *RPE1^WT^* cultures thus phenocopies the increased proportion of ciliated cells found in *KDM3A^null^* cells in the presence of serum ([Fig. 2 E](#fig2){ref-type=\"fig\"}). The low abundance of actin fibers of *KDM3A* mutants likely facilitates ciliogenesis, but further perturbations in actin polymerization are not tolerated by *KDM3A^null^* cells ([Fig. 5 E](#fig5){ref-type=\"fig\"}).\n\n![**Altered actin polymerization underlies *KDM3A* mutant ciliary traits.** (A) Wide-field images showing phalloidin-stained foci in *KDM3A^null^* (line 2) when treated with 2 \u00b5M CytoD for 4 h +FCS. (B) Percentage of ciliated cycling cells in *RPE1^WT^* and *KDM3A^null^* (line 2) cultures in the presence of CytoD. Symbols represent the mean \u00b1 SD. Lines between symbols were added to ease visualization but are not in scale. (C) Single time frames (spanning \u223c2 h) from live videos of *RPE1^WT^* and *KDM3A^null^* expressing 5HT6-EGFP in the presence of CytoD and FCS show instable mutant cilia (arrows). (D) Relative increase in ciliary lengths observed during 4 h of live imaging of *RPE1^WT^* and *KDM3A^null^* in the presence of CytoD and FCS. (E) Graphical summary: Depolymerizing actin increase ciliogenesis in cycling *RPE1^WT^* cells. Conversely due to the already reduced number of actin fibers in *KDM3A^nulls^* cells further depolymerization may induce ciliary instability and synthetic lethality. (F) Percentage of ciliated cells after resorption is induced by serum replenishment in the absence (DMSO) or presence of jasplakinolide for 80 min. Error bars represent SD (*N* = 3, number of cells scored indicated within each bar). \\*, P \\< 0.05; \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001; ns, not significant. (G) Images of cultures scored in F. (H) Graphical summary of jasplakinolide results.](JCB_201607032_Fig5){#fig5}\n\nCritically, the converse experiment, where we promoted actin polymerization using jasplakinolide treatment ([@bib30]) in the presence of serum, rescued the abnormal resorption of *KDM3A^null^* cilia ([Fig. 5, F and G](#fig5){ref-type=\"fig\"}). This suggests that failure of *KDM3A^null^* cells to polymerize actin in response to serum contributes to the abnormal resorption kinetics of mutant cilia ([Fig. 5 H](#fig5){ref-type=\"fig\"}).\n\nThese pharmacogenetic results support a functionally relevant association between KDM3A and the actin cytoskeleton during ciliogenesis. In the absence of KDM3A, the formation of actin filaments is perturbed and underlies the ciliary defects of *KDM3A^null^* cells.\n\nA requirement for KDM3A in cilial stability and IFT {#s07}\n---------------------------------------------------\n\nLive imaging suggested that *KDM3A^null^* cilia were unstable. We investigated whether this instability was caused by the drug treatment or whether it is an inherent property of mutant cilia. Live imaging of cell during ciliary growth (\u2212FCS) revealed that *RPE1^WT^* cilia remained fairly stable, whereas *KDM3A^null^* cilia underwent excessive elongation and fragmentation (Videos 1 and 2 and [Fig. 6 A](#fig6){ref-type=\"fig\"}). Similar ciliary instability was observed in *Kdm3a^\u0394JC^* mutant MEFs (Video 3). These results suggest that instability of mutant axonemes in the absence of KDM3A reflect an inability to regulate axonemal length. These findings are supported by the unresponsiveness of *KDM3A^null^* cilia to resorb (Video 4) and the wider range of lengths observed in fixed *KDM3A* mutant cilia ([Fig. 2](#fig2){ref-type=\"fig\"}).\n\n![**Perturbing actin polymerization phenocopies IFT abnormalities of *KDM3A* mutants.** (A) Single time frames from Videos 1 and 2 of live *RPE1^WT^* and *KDM3A^null^* (line 2) cultures. Arrows point to bulges forming along *KDM3A^null^* cilia. (B) Confocal images after 48 h \u2212FCS indicate low levels of IFT81 at tip of *RPE1^WT^* cilia (green arrow) but conspicuous accumulations in *KDM3A^null^* mutant cilia, frequently accompanied by IFT81 containing \"vesicles\" (yellow arrows). (C) Wide-field images of *RPE1^WT^* cultures in the presence of serum show accumulations of endogenous IFT81 in CytoD treated cells that appear sequentially as ciliary buds (cb; without acetylated \u03b1-tubulin--stained axonemes, most frequent with 0.2 \u00b5M CytoD) and progress toward ciliary tip (ct; 2 \u00b5M CytoD). (D) Representative images of IFT81 distribution in cilia formed after 3 h with solvent control or 100 \u00b5M CK-869 \u2212FCS. (E) Quantification of cilia with or without IFT81 bulges among treated (2 \u00b5M CytoD and 100 \u00b5M CK-869) and solvent controls. Numbers in bars indicate number of cilia scored. Error bar represents SD. P-value is from \u03c7^2^ test. (F) Frequency of length ranges illustrate increased proportion of long cilia in 100 \u00b5M CK-869--treated cultures. (G) Differential sensitivity of *RPE1^WT^* and *KDM3A^null^* to ARP2/3 inhibition as illustrated by the length of cilia in subconfluent cultures (50--70 cells/0.1 mm^2^) treated for 3 h with 2 and 10 \u00b5M CK-869 \u2212FCS. Mean \u00b1 5th--95th percentile (*N* = 3 replicates, *n* \\> 30 cilia measured/condition). (H) Mean \u00b1 5th--95th percentile of ciliary lengths indicate that *RPE1^WT^* cultures begin to show significant cilial elongation in \u226550 \u00b5M CK-869 (*N* = 3 replicates, *n* \\> 30 cilia measured/condition). \\*, P \\< 0.05; \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001; \\*\\*\\*\\*, P \\< 0.0001 (paired *t* test); ns, not significant. Bars, 5 \u00b5m.](JCB_201607032_Fig6){#fig6}\n\nThe presence of ciliary bulges ([Fig. 2 E](#fig2){ref-type=\"fig\"}) and their shedding observed during live imaging of *KDM3A^null^* cells suggest that the instability of *KDM3A* mutant cilia results from perturbation of IFT. Indeed, IFT staining of MEFs ([Fig. 1 F](#fig1){ref-type=\"fig\"}) and RPE1 null for KDM3A ([Fig. 6 B](#fig6){ref-type=\"fig\"}) showed abnormal distributions of IFT88 and IFT81. The anomalies were most evident at the ciliary tip ([Fig. 6 C](#fig6){ref-type=\"fig\"}) but could also be observed at the ciliary base ([Fig. 1 F](#fig1){ref-type=\"fig\"}).\n\nIn *Chlamydomonas reinhardtii,* actin is required for the recruitment of IFT particles to the basal body, where it controls flagellar entry of IFT trains ([@bib2]). Accordingly, the abnormal distribution of IFTs observed in *KDM3A^null^* cells could be caused by the permissive and depolymerized state of the actin cytoskeleton in these mutants. If this was true, then perturbing actin dynamics in *RPE1^WT^* would phenocopy the IFT anomalies observed in *KDM3A* mutants. To test this, we treated cells with CytoD or CK-869, a small molecule that specifically inhibits ARP2/3 complex components ([@bib19]) copurifying with KDM3A ([Fig. 3 F](#fig3){ref-type=\"fig\"}). Perturbations of ARP2/3 activity could contribute to themorphological phenotype of *KDM3A^null^* cells ([Fig. 4, E--G](#fig4){ref-type=\"fig\"}). *RPE1^WT^* cells treated with either drug in the presence or absence of serum showed large accumulations of IFT81 at the ciliary tip of *RPE1^WT^* cells ([Fig. 6, C--E](#fig6){ref-type=\"fig\"}) accompanied by ciliary elongation ([Fig. 6 F](#fig6){ref-type=\"fig\"}). The effect of CK-869 was particularly interesting, as *KDM3A^null^* cells were extremely responsive to ARP2/3 inhibition, showing ciliary elongation from doses as low as 2 \u00b5M ([Fig. 6 G](#fig6){ref-type=\"fig\"}), whereas *RPE1^WT^* cells only respond from 50 \u00b5M ([Fig. 6 H](#fig6){ref-type=\"fig\"}). This functional synergism between *KDM3A* mutants and CK-869 treatment indicates a common underlying mechanism by which KDM3A influences IFT through actin polymerization involving the ARP2/3--actin assembly pathway.\n\nIn summary, live imaging studies provide evidence that in the absence of KDM3A, cilia become unstable. Antibody staining showed abnormal distributions of endogenous IFTs in mutant cilia, and this trait can be phenocopied in wild-type cells by perturbing actin dynamics. Using two different drugs interfering with actin nucleation, we demonstrate that in mammalian primary cilia, actin dynamics also modulates IFT. Altogether, these findings provide evidence that in the absence of KDM3A, actin dynamics are perturbed, disrupting IFT and destabilizing cilia.\n\n*KDM3A^null^* cilia reveal an actin-driven constraint for IFT in wild-type cilia {#s08}\n--------------------------------------------------------------------------------\n\nThe finding that inhibiting actin polymerization in mammalian control cells phenocopies the abnormal IFT loaded bulges found in *KDM3A^null^* cells suggests that wild-type cilia have an actin-dependent constraint that modulates ciliary IFT. To test this further, we increased IFT protein levels by transfecting tagged IFT81 into *RPE1^WT^* and *KDM3A^null^* cells, reasoning that if an actin gate does exist and is dysfunctional in the absence of KDM3A, IFT overexpression would have opposing effects in wild-type and *KDM3A* mutants.\n\nWe generated a functional IFT81-DDK construct capable of rescuing ciliogenesis in an *IFT81*-null RPE1 line (Fig. S3, A--D). Comparable transfection levels of IFT81-DDK onto *RPE1^WT^* and *KDM3A^null^* cells were confirmed by immunoblotting ([Fig. 7 A](#fig7){ref-type=\"fig\"}). Live imaging revealed that overexpressing IFT81 enabled formation of more cilia and with faster kinetics in *KDM3A^null^* mutants compared with *RPE1^WT^* cells ([Fig. 7 B](#fig7){ref-type=\"fig\"}, slopes; and Fig. S3, E and F). Importantly, expression of 5HT6-GFP alone did not recapitulate this increased rate of ciliogenesis in *KDM3A^null^* cells ([Fig. 7 C](#fig7){ref-type=\"fig\"}), demonstrating that this difference is not simply caused by an increased propensity to ciliate but is dependent on ectopic IFT81. In the presence of IFT81-DDK, *KDM3A^null^* cilia are also longer than controls ([Fig. 7 D](#fig7){ref-type=\"fig\"}). These results did not reflect differences in cell density ([Fig. 7 E](#fig7){ref-type=\"fig\"}) or levels of endogenous IFT-B components ([Fig. 7 F](#fig7){ref-type=\"fig\"}). This indicates that increasing the abundance of IFT promotes ciliogenesis in *KDM3A* mutants, but not wild-type controls.\n\n![**KDM3A uncovers an actin-mediated control on IFT that precludes IFT81 overexpression.** (A) Immunoblots with DDK antibody of total cell extracts harvested 24 h after transfection with IFT81-DDK plasmid. *IFT81*^*null*^ cells transfected with 5HT6-EGFP used as negative control for DDK signal (red arrow). Ponceau red is a loading control. (B) Line graph showing the number of 5HT6-EGFP--positive cells with a cilium at the indicated time points. Error bars represent SEM between fields of view. Linear regression and p-values from comparing the individual slopes by paired *t* test. See also Fig. S3 (E and F). (C) Median and range of the percentage of 5HT6-EGFP--positive cells that contain a cilium under each experimental condition. Each dot is the percentage of one independent experiment (*N* = 4 transfections, *n* \\> 100 GFP+ cells scored/condition). \\*\\*, P \\< 0.01 (paired *t* test). (D) Mean \u00b1 10th--90th percentile of lengths of 5HT6-EGFP--positive ciliafrom the two last cotransfections with IFT81-DDK presented in C. \\*\\*\\*, P \\< 0.001 (paired *t test*). (E) Plates fixed after live imaging confirm equal cell densities in *RPE1^WT^* and *KDM3A^null^* cultures. (F) Immunoblots of total cell extracts from the indicated wild-type and mutant lines after 3 h \u2212FCS (ciliary growth) indicate comparable total levels of endogenous IFT components. (G) Mean \u00b1 10th--90th percentile of *RPE1^WT^* cells cotransfected with 5HT6-EGFP and vector control (\u2212) or IFT81-DDK plasmid that contain a cilium measured 24 h after transfection in the presence or absence of 0.5 \u00b5M CytoD. Each dot is the percentage of one field of view taken (*N* = 2 transfections, *n* = 4, 28, 12, and 20 fields scored/condition). Analysis of variance: P \\< 0.004; *t* test for the indicated paired comparisons: \\*, P \\< 0.05; \\*\\*\\*, P \\< 0.0005. (H) Graphical summary illustrates the insensitivity of *RPE1^WT^* to IFT81 overexpression unless actin is depolymerized or KDM3A is absent. ns, not significant.](JCB_201607032_Fig7){#fig7}\n\nBecause CytoD treatment of control cells phenocopied the ciliary traits of *KDM3A^null^* mutants, we simultaneously induced actin depolymerization along with IFT81 overexpression in *RPE1^WT^* cultures. Indeed, this dual treatment allows wild-type cells to become sensitive to IFT-B overexpression when actin is depolymerized ([Fig. 7 G](#fig7){ref-type=\"fig\"}).\n\nTogether, these findings provide further and independent evidence of deregulated IFT in *KDM3A* mutant cilia that result from perturbations in actin dynamics. When actin is unperturbed, increasing IFT pools in control cells makes no difference to ciliary growth, indicating that actin allows wild-type cells to maintain tight regulation of IFT entry into the ciliary compartment ([Fig. 8 A](#fig8){ref-type=\"fig\"}). In contrast, when actin nucleation is perturbed either genetically (*KDM3A* mutants) or chemically (CytoD treatment ARP2/3 inhibition), this constraint is lifted.\n\n![**Proposed mechanism for KDM3A in ciliary regulation.** (A) Collapsed time frames (3 h) of live *RPE1^WT^* and *KDM3A^null^* cultures transiently transfected with ARL13B-mKate2 and IFT81-mGFP showing consistently (arrows) increased accumulations and/recruitment of IFT81-EGFP at the mutant basal body. (B) Enrichment of Halo-KDM3A in centrosomes in \u223c2% of transfected cells. (C) Centrosomal localization of KDM3A confirmed by construct containing its zinc-finger and nuclear receptor--binding domains fused to GFP, colocalizing with IFT81. See also Video 5. These findings support KDM3A regulating both global and local actin dynamics at specific subcellular compartments with a direct impact on ciliogenesis. (D) Summary of *KDM3A* mutant cellular phenotype: In the absence of KDM3A, the actin cytoskeleton contains reduced abundance of actin filaments, wide lamellipodia at all cellular edges and reduced migration. Without an actin gate, ciliogenesis is facilitated, and these cilia are long but unstable. (E) Summary of serum (+FCS or low serum \\[LS\\]) and pharmacogenetic modulation of actin at the pericentriolar periciliary compartment (PPC) illustrates that promoting actin polymerization either phenocopies ciliary growth in wild-type cells or rescues the abnormal resorption of *KDM3A* mutant cilia. CK-869 treatment of cells pinpoints the actin nucleation activity of ARP2/3 to be necessary to maintain normal IFT within cilia. IFT81 accumulations suggest unbalanced anterograde-retrograde transport (represented by different arrow size) when actin is perturbed. Together, these results indicate that the ciliary traits of *KDM3A* mutants stem from perturbations of actin dynamics, which upset IFT.](JCB_201607032_Fig8){#fig8}\n\nDiscussion {#s09}\n==========\n\nOur results show that KDM3A plays a role in mammalian ciliogenesis through the modulation of actin dynamics. In the absence of KDM3A, the assembly and growth of primary cilia is enhanced. However, cilial stability is compromised such that the length of mutant axonemes and presence of IFT proteins within cilia become unregulated. We find that *KDM3A^null^* cells have reduced abundance of actin filaments. This is because KDM3A plays a transcriptional role regulating actin gene expression and binds directly to actin protein, influencing local actin networks. We demonstrate that the ciliary traits uncovered in *KDM3A* mutants can be chemically reversed or phenocopied in wild-type cultures by altering actin dynamics. Through independent approaches involving genetics, small-molecule inhibitors, and IFT overexpression, we uncover an actin-mediated constraint on IFT within mammalian cilia. Our findings uncover a functional interaction between the actin cytoskeleton and IFT that is perturbed in the absence of KDM3A.\n\nCells adapt to new environments, like changes in the serum content of the media, through simultaneous changes in gene expression, cytoskeletal rearrangements, and cilia signaling. This integrated response requires signal transduction relays that communicate the cytoplasmic actin polymerization status with IFT in cilia and histone code in the nucleus. Through the use of *KDM3A* mutants, we have uncovered a requirement for KDM3A, which acts as a nexus in this intracellular signaling network. RNA sequencing and proteomic profiling reveal a transcriptional role for KDM3A in increasing the pool of actin in response to serum replenishment that is concomitant with actin filament formation and ciliary resorption. Failure of *KDM3A* mutants to up-regulate actin expression will thus likely contribute to the poor resorption of mutant cilia. During disassembly, formation of new actin filaments at the ciliary base would likely require a pool of free actin to restrict access into the ciliary compartment. Nevertheless, in the absence of serum *KDM3A^null^* mutants contain comparable levels of actin transcripts and protein to those found in control cells, yet it is under these growth conditions that ciliary extension is facilitated in *KDM3A* mutants. Here, we propose that direct protein interaction of KDM3A with actin locally creates an actin gate at the ciliary base ([Fig. 8, B and C](#fig8){ref-type=\"fig\"}; and Video 5) to regulate entry of IFT. In the absence of KDM3A, access to cilia is unrestricted and balance and/or kinetics of IFT are perturbed.\n\nThe fact that KDM3A is an enzyme that binds actin through its catalytic domain, together with previous observations that actin proteins could be abnormally methylated in *Kdm3a* mutant testes ([@bib22]), prompted us to test the demethylase activity of KDM3A in lysine-methylated actin peptides found in this study. However, as yet we have failed to detect changes in methylation by *KDM3A* in any tested peptide in vitro (unpublished data). Although it remains to be seen whether KDM3A can posttranslationally modify actin in vivo, the binding of KDM3A to actin itself is likely to modulate binding of other actin modifiers, including ARP2/3 and the capping components, to directly tune actin networks.\n\nThe functional interplay between KDM3A-ARP2/3 is interesting; *KDM3A^null^* cells respond to lower CK-869 doses than *RPE^WT^* cells, suggesting that the ciliary traits of *KDM3A* mutants have already compromised ARP2/3 function. A role for ARP2/3 at the cilium/centrosome has been suggested. ARP2/3 mediates actin nucleation at centrosomes ([@bib15]). Silencing of *ACTR3* or *ARP2* in cultured cells promotes ciliogenesis but with unclear kinetics, as cells were phenotyped 2.5--3 d after siRNA depletion was initiated ([@bib25]; [@bib7]). In our study, we used a small-molecule inhibitor, CK-869, which prevents the formation of new actin filaments and unlike CytoD does not destabilize filaments that are already formed ([@bib19]). We demonstrate that acutely blocking ARP2/3 activity (3-h treatment) alters the distribution of endogenous IFT-B within cilia. This suggests that continual actin nucleation by ARP2/3 is required for regulated entry of IFT into cilia and cilial length maintenance.\n\nOur identified roles for KDM3A in actin dynamics have further implications. Understanding how cancer cells can subvert actin bundling during metastasis is an emerging and exciting theme in research. Although the brunt of the current focus of overexpression of KDM3A in various types of cancer has been on its transcriptional targets ([@bib29]; [@bib55]; [@bib10]; [@bib39]; [@bib53]; [@bib40]), our work suggests that cytoplasmic KDM3A may act as an important cytoskeletal rheostat integrating mechanosensation ([@bib23]; [@bib41]) with cell adhesion and signaling through cilia. Aberrant length ranges and resorption of *KDM3A* mutant cilia in response to environmental cues could also contribute to metabolic imbalance in *Kdm3a* mouse models ([@bib37]), as ciliary length is tightly regulated in response to leptin levels ([@bib17]).\n\nOur findings that KDM3A has roles beyond the \"direct\" regulation of transcription (i.e., acting as an H3K9 demethylase; [@bib60]) links cytoplasmic sensing of actin polymerization, ciliogenesis, and transcriptional activity in response to serum. KDM3A is a versatile protein that is tightly regulated in response to environmental inputs ([@bib45]; [@bib29]; [@bib9]; [@bib22]; [@bib1]) with both catalytic and structural roles ([@bib1]; [@bib50]) involving binding to histone ([@bib60]) and nonhistone proteins ([@bib47]) within nuclear and cytoplasmic compartments ([@bib36]; [@bib61]; [@bib59]; [@bib22]; [@bib23]). KDM3A thus appears to be capable of integrating epigenetic and cytoskeletal regulation with ciliary IFT in response to extracellular cues.\n\nWhile our manuscript was in press, actin-mediated scission of the tips of mammalian primary cilia was uncovered as a mechanism to clear activated ciliary receptors ([@bib34]) or to promote ciliary disassembly ([@bib42]), independent of retrograde IFT involvement. The behavior of *KDM3A^null^* cilia is consistent with these findings; through its role modulating actin dynamics, KDM3A would also modulate ciliary ectocytosis at the tip. The relative contribution of unrestricted IFT recruitment to cilia or reduced IFT clearance through ectosytosis in KDM3A mutants is an exciting topic for future research.\n\nMaterials and methods {#s10}\n=====================\n\nPlasmids {#s11}\n--------\n\nPlasmids encoding the serotonin receptor pEGFPN3-5HT6 ([@bib3]), IFT81-DDK, and Halo-KDM3A were purchased from Addgene (35624), OriGene (RC216437), and Kazusa DNA Res Institute (FHC05559), respectively. IFT81-mGFP was made by transferring Sgf1-Mlu1 restriction fragment from clone RC216437 into mTagGFP vector PS100048 (OriGene). The KDM3A deletion mutant GFP-ZFNR was cloned using PCR primers containing Bgl2 and Kpn1 sites for cloning into EGFP-C1 vector (Takara Bio Inc.; listed in Table S1) with Halo-KDM3A as template and PfuUltra II DNA polymerase (Agilent Technologies), subsequently confirmed by sequencing. KDM3A~i2~-EGFP ([@bib22]) contains all the functional domains of full-length KDM3A except for a nuclear localization signal ([@bib6]).\n\n*Kdm3a* mouse model {#s12}\n-------------------\n\nThe deletion mutant (*Kdm3a^\u0394JC^*) and hypomorphic gene-tarp (*Kdm3a^GT^*) mouse models were described previously ([@bib52]; [@bib22]). Genotyping was done with PCR primers as in Table S1.\n\nCell lines and CRISPR/Cas9 mutants {#s13}\n----------------------------------\n\nPrimary MEF cultures were established by mincing embryonic day 11.5--13.5 embryos in DMEM supplemented with antibiotics, FCS, and \u03b2-mercaptoethanol and grown in 3% O~2~. Telomerase immortalized human retinal pigment epithelial cells (hTERT-RPE1, referred to here as *RPE1^WT^*) were from Takara Bio Inc. and maintained in DMEM/F-12 (1:1; Thermo Fisher Scientific). *RPE1^WT^* cells were engineered with a double nicking strategy using CRISPR/Cas9 nickases following reported guidelines ([@bib48]). Oligonucleotides to produce guide RNAs (listed in Table S1) were designed with CRISPR Design Tool (<http://tools.genome-engineering.org>). In brief, guides cloned onto pX461 or pX462 (Addgene) mammalian expression plasmids were cotransfected onto *RPE1^WT^* cells followed by 24 h of 3 \u00b5g/ml puromycin selection and individual GFP-positive cells sorted by FACS onto 96-well plates to establish the RPE1 lines *KDM3A^null^* and *IFT81^null^*. *RPE1^CrWT^* derives from noneventful CRISPR targeting of *IFT81,* subsequent sequencing confirmed wild-type alleles for both IFT81 and KDM3A. Mutations were confirmed by immunoblotting with KDM3A and IFT81 antibodies and sequencing of genomic PCR (primers are listed in Table S1) cloned into pGEM-T Easy (Promega) using T7 and SP6 primers.\n\nTransfections {#s14}\n-------------\n\nParental and engineered RPE1 lines were transfected with Lipofectamine2000 (Thermo Fisher Scientific) following the manufacturer's instructions, grown without antibiotic the day before, and transfected in the absence of serum for 5 h. All cotransfections were done mixing the DNA of the relevant cotransfected plasmids before the addition of transfection reagents.\n\nAntibodies {#s15}\n----------\n\nAntibodies were from the following sources: *N-*KDM3A (12835; Proteintech), *C-*KDM3A (NB100-77282; Novus Biologicals), anti--acetylated \u03b1-tubulin (T7451; Sigma-Aldrich), \u03b3-tubulin (GTU-88; Sigma-Aldrich), IFT88 (13967-1-AP; Proteintech), IFT81 (11744-1-AP; Proteintech), GAPDH (5019A-2; ImGENEX), ARL13B (17711-1-AP; Proteintech), DDK tag (TA50011; OriGene), and Halo (G9281; Promega). Secondary antibodies for immunofluorescence and immunoblotting are Fab\u20322 IgG Alexa Fluor from Molecular Probes and HRP conjugated from EMD Millipore and Sigma-Aldrich.\n\nDrug treatments {#s16}\n---------------\n\nCytoD (25023; EMD Millipore), CK-869 (C9124; Sigma-Aldrich), and jasplakinolide (1689-05; BioVision) were dissolved as recommended by the manufacturers and used at the indicated concentrations. Controls for each drug treatment contained the relevant solvent only.\n\nCiliary phases {#s17}\n--------------\n\nFor the study of cilia dynamics, cells were plated 24 h in advance. In all cases, stocks of all cell lines were maintained at comparable cell densities. For the studies of ciliogenesis in exponentially growing cultures, cells were not allowed to reach confluency during two consecutive passages before assaying ensuring a consistent number of cycling cells. For cilia resorption studies, cells were plated between 40% and 50% confluency followed by a further 24, 48, or 72 h culture without serum (\u2212FCS, as indicated in each panel); for ciliary growth assays (3--24 h), cells were plated at 60--70% confluency. Resorption was induced by replacing serum-free media with media containing 10% FCS (+FCS) for the indicated times. Temperature-induced resorption was done by adding 10% FCS and floating the tissue culture plates on a water bath maintained at 42\u00b0C for the indicated times.\n\nCell spreading assays {#s18}\n---------------------\n\nActively growing cultures were trypsinized and resuspended in complete media, replated immediately, or spun down and placed in suspension without serum for 90 min in a cell incubator. Replated cells were imaged 90--120 min after replating on uncoated plastic plates.\n\nYeast two-hybrid studies {#s19}\n------------------------\n\nThe initial yeast two-hybrid screen was performed using an embryonic eye mouse cDNA library ([@bib22]) with the JMJC domain of KDM3A as bait cloned through partial EcoRI digestion into the DNA-binding domain encoding plasmid pGBKT7. Bait and prey plasmids from blue colonies growing in quadruple selection (Leu/Trp/Ade/His) were rescued by cell lysis and bacterial transformation in ampicillin (pGADT7) or kanamycin (pGBKT7) plates and sequenced. Plasmids encoding CAPZB and ACTA2 were cotransformed back into the yeast two-hybrid Gold strain (Takara Bio Inc.) along with JMJC-pGKBT7, as described in the Matchmaker two-hybrid system protocol. Independent colonies grown in double selection (Leu/Trp) where resuspended, and a 1/100 dilution was replated in 0.2 \u00b5g/ml aureobasidin A double selection plates in the presence of X-\u03b1-Gal (Takara Bio Inc.).\n\nImmunofluorescence {#s20}\n------------------\n\nCells were grown on glass coverslips and fixed (10 min, 4% methanol-free formaldehyde), except for those shown in [Fig. 6 (C and D)](#fig6){ref-type=\"fig\"}, which were fixed with ice-cold methanol for 5 min. Fixed cells were permeabilized (15 min in 0.25% Triton X-100) and blocked in 2% BSA before incubation with antibodies and DAPI and mounted in Prolong Gold (Thermo Fisher Scientific). Rhodamine--phalloidin (R415; Thermo Fisher Scientific) added to the secondary antibody incubation was used to visualize F-actin of fixed cells.\n\nImaging {#s21}\n-------\n\nLive-cell microscopy was done using glass-bottom plates (662892; Greiner Bio-one) in a heated chamber with a source of nitrogen using a 40\u00d7 plan Fluor 0.75 NA dry lens in an A1R confocal microscope (Nikon) in the relevant culture media for each growth condition. Movies shown were acquired as to minimize photobleaching. Postacquisition analyses were done with NIS-Elements (Nikon) on maximum-intensity projections using Fiji (ImageJ) and IMARIS 8.4 (Bitplane). Ciliary length studies on fixed cells were done with wide-field microscopy or confocal imaging as indicated in each panel using a 40\u00d7 dry lens and FIJI software for measurements.\n\nCell lysis and immunopurifications {#s22}\n----------------------------------\n\nTotal extracts for immunoblots were obtained by lysing cells in 150 mM Tris-HCl, pH 7.5, 50 mM NaCl, 0.5% IGEPAL, 1% Triton X-100, 1 mM EDTA, 0.5% deoxycholate, Benzoase DNase (EMD Millipore), and Halt protease inhibitor (1860932; Thermo Fisher Scientific). Total cell extracts for total cell proteome were done in PBS and 1% SDS in the presence of protease inhibitor, sonicated, and cleared by 10-min centrifugation at 14,000 rpm. For KDM3A interactome, *RPE1^WT^* (in duplicate), *RPE1^WTcr^*, and *KDM3A^null^* (line 2, in triplicate) cells were plated the day before and grown in the absence of serum for 3 h, lysed in 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10% glycerol, 5 mM EDTA, 0.5% IGEPAL, and protease inhibitors and incubated with N-KDM3A antibody (12835; Proteintech) for 2 h at 4\u00b0C followed by 15-min incubation with PureProteome magnetic beads (EMD Millipore) to concentrate the immunoglobulin complexes. Three successive washes were performed to reduce the IGEPAL concentration.\n\nFar--Western blot {#s23}\n-----------------\n\nSerial dilutions (50--200 ng/\u00b5l) of pure actin derived from rabbit skeletal muscle (AKL99; Cytoskeleton, Inc.) resolved in 4--12% acrylamide were transferred onto nitrocellulose and stained with Ponceau red to visualize loadings. Bacterially expressed GST or GST-JMJC immobilized in glutathione magnetic beads (8821; Proteintech) were washed and eluted in glutathione buffer. Membrane blocked with 5% BSA was incubated with eluted GST proteins in 10 mM Tris, pH 7.6, 100 mM KCl, 1 mM MgCl~2~, 0.1 mM EDTA, 1 mM DTT, 0.1% (vol/vol) Tween-20, and 2% BSA for 1 h. Subsequent steps followed standard immunoblotting protocol using GST antibody to detect GST proteins retained on membrane.\n\nMass spectrometry {#s24}\n-----------------\n\nOn-bead purified complexes were digested with trypsin, reduced, alkylated, and desalted, and two runs of each sample were analyzed by liquid chromatography tandem mass spectrometry as described previously ([@bib54]). Datasets can be found in *ProteomeXchange* identifier PXD004334. Data were analyzed using the MaxQuant software suite by searching against the UniProt HUMAN database with protein N terminus (acetylated), M (oxidation). Label-free quantification values were determined by MaxQuant as described previously ([@bib12]). In the search for posttranslational modifications, databases were also searched for peptides with mass modifications as follows: mono-, di-, and trimethylation (delta mass of 14.0157, 28.0313, and 42.04695 D, respectively).\n\nRNA sequencing {#s25}\n--------------\n\nRNA was isolated from RPE1*^WT^* and *KDM3A^null^* (line 2) cells in triplicate cultures under ciliary growing or resorption conditions ([Fig. 3 A](#fig3){ref-type=\"fig\"}) using RNeasy Plus (QIAGEN) according to the manufacturer's instructions. RNA integrity score was calculated with RNA 6000 Nano reagent (Agilent Technologies) in a 2100 Bioanalyzer (Agilent Technologies). All samples used for the preparation of the RNA-sequencing library had a RIN (RNA Integrity Value) score above 9. Library construction used poly(A) RNA selection, and all 100-bp reads were obtained using Illumina HiSeq2000. Cufflinks, CummeRbund, and Picard were used for alignments to reference genome, transcript assembly, quantification, and statistical analysis, respectively, done at Oxford Gene Technology (The Molecular Genetics Company). The mean read depth for each sample was of 13 \u00d7 10^6^.\n\nGO term analyses {#s26}\n----------------\n\nGenes or proteins differentially enriched in proteome or interactome datasets were analyzed using DAVID (the Database for Annotation Visualization and Integrated Discovery).\n\nStatistics {#s27}\n----------\n\nStatistics were done using GraphPad Prism 5 software. One-way analysis of variance was used for the initial assessment of multiple comparisons. In those cases where a significant difference was observed (P \\< 0.01), a pairwise two-tailed Students *t* test was applied to compare individual conditions. \u03c7^2^ *t* test was applied to compare percentages as indicated in each panel. Asterisks indicate p-values (\\*, P \\< 0.05; \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001; \\*\\*\\*\\*, P \\< 0.0001) calculated through the statistical analysis indicated in each case. N represents the number of experimental or biological replicates, and n represents the number of events measured in all experimental or biological replicates as indicate.\n\nOnline supplemental material {#s28}\n----------------------------\n\nTable S1 shows sequences of primers used to clone *KDM3A* expression plasmids, CRISPR/Cas9 editing, and all genotyping. Table S2 contains raw values of RNA sequencing reads and proteomic studies. Table S3 shows thresholding criteria and the resulting lists of genes that were subsequently used for GO term analysis. Table S4 shows enriched GO terms. Fig. S1 shows genotypes of *KDM3A^null^* cells, ciliary anomalies of *Kdm3a^GT/GT^* hypomorph MEFs, and effect of overexpressing KDM3A in the percentage of ciliated cells. Fig. S2 shows phalloidin staining of *Kdm3a^\u0394JC/\u0394JC^* MEFs, assessment of KDM3A interaction with actin by two-hybrid screening and far--Western blotting, and detailed measurement criteria used to quantify the extent of filopodia in transfected cells. Fig. S3 shows genotyping of *IFT81^null^* cells, functionality of IFT81-DDK construct, and identical transfection efficiencies of 5HT6-EGFP+IFT81-DDK onto *RPE1^WT^* and *KDM3A^null^* cells. Videos 1 and 2 show instability of *KDM3A^null^* cilia. Video 3 shows instability of *Kdm3a^\u0394JC/\u0394JC^* cilia. Video 4 illustrates the unresponsive behavior of *KDM3A^null^* cilia to serum-induced resorption. Video 5 illustrates that KDM3A~i2~-EGFP can be found in cilia.\n\nSupplementary Material\n======================\n\n###### Supplemental Materials (PDF)\n\n###### Video 1\n\n###### Video 2\n\n###### Video 3\n\n###### Video 4\n\n###### Video 5\n\nWe thank the Structural Genomics Consortium (Oxford) for assistance in protein purification for catalytic studies and the Institute of Genetics and Molecular Medicine technical staff for support with animal husbandry, imaging, and sequencing. We also thank Georgios Kanellos and Brad J. Nolen for expert advice.\n\nC.J. Schofield thanks Cancer Research UK and the Wellcome Trust for funding. This work was supported by core funding from the Medical Research Council (MC_UU_12018/26; P.L. Yeyati, G. Mali, I. Kasioulis, and P. Mill).\n\nThe authors declare no competing financial interests.\n\nAuthor contributions: P.L. Yeyati contributed to the design, execution, and interpretation of the overall research plan. J. Willis, A. von Kriegsheim, and A. Finch performed mass spectrometric analyses, which were interpreted by A. von Kriegsheim and G. Mali. R. Schiller, A. Kawamura, and C.J. Schofield tested KDM3A catalytic activity. I. Kasioulis contributed to KDM3A cloning and subcellular localization experiments. I.R. Adams and C. Playfoot explored KDM3A chromatin binding. N. Gilbert, V. van Heyningen, and I.J. Jackson contributed to experimental design and provided expert advice. J. Sakai shared unpublished data and expert advice. P. Mill oversaw experimental design and interpretation. P.L. Yeyati, C.J. Schofield, and P. Mill wrote the manuscript with contributions from all authors.\n\nAbbreviations used:CytoDcytochalasin DIFTintraflagellar transportJmjCJumonji CMEFmouse embryonic fibroblast\n\n[^1]: V. van Heyningen\\'s present address is Institute of Ophthalmology, University College London, EC1V 9EL London, England, UK.\n"}
+{"text": "Introduction\n============\n\nBackground\n----------\n\nSkin cancer, the most common cancer in the United States, constitutes a serious public health burden \\[[@ref1]-[@ref4]\\]. Skin cancer may be deadly or disfiguring. The most serious skin cancer, melanoma, resulted in an estimated 9000 deaths in 2018 \\[[@ref4]\\]. Skin cancer treatment costs approximated US \\$8.1 billion in 2011 \\[[@ref1]-[@ref3]\\]. Most skin cancers can be prevented by simple behaviors to protect the skin from ultraviolet radiation (UVR), such as staying in shade, wearing sun protective clothing, applying sunscreen, and avoiding indoor tanning \\[[@ref4],[@ref5]\\]. Early detection of skin cancer greatly decreases its potential morbidity, mortality, and cost \\[[@ref6]-[@ref8]\\]. The probability of early skin cancer detection increases with full body visual skin assessment (VSA) \\[[@ref9]\\]. Despite these effective prevention and early detection strategies, over 5 million skin cancer cases are diagnosed or treated annually \\[[@ref10]\\]. Thus, decreasing the burden of skin cancer depends on concerted and innovative public health efforts that extend beyond the conventional biomedical practitioners to complementary and integrative health care practitioners. These efforts also involve other community-wide sectors and could incorporate electronic learning (e-learning) technology to allow widespread and easy dissemination of knowledge.\n\nIn 2014, the Surgeon General issued a *Call to Action to Prevent Skin Cancer*, endorsing comprehensive community-wide efforts to prevent skin cancer by diverse partners and sectors, including business, health care, and education \\[[@ref5]\\]. Massage therapists (MTs) are community members typically practicing outside of conventional health care settings, yet are professionals involved in promoting health and wellness. Despite their interface with health and wellness, MTs have been overlooked as a community-based resource to (1) help promote skin cancer risk reduction and (2) reinforce consumer-targeted public heath skin cancer awareness messages.\n\nMTs are uniquely positioned to promote skin cancer risk reduction through *eyes on the skin* observation and client-centered communication. During a typical full body massage, the client is unclothed under a drape. MTs systematically undrape and view each anatomical area, allowing the opportunity to visualize skin cancer risk factors such as sunburn, tanning lines, high mole counts, or suspicious lesions. Clients typically see their MTs more often and for longer durations than their primary care provider and are more likely to discuss health promotion \\[[@ref11]-[@ref13]\\], thereby providing greater opportunities for successful client-centered communication and encouragement of effective skin cancer risk reduction behaviors such as reducing UVR exposure \\[[@ref14]\\].\n\nIn our prior work with MTs, we conducted in-person and Web-based tobacco cessation brief behavioral intervention (BBI) training for nonbiomedical health care practitioners (including MTs) in private practice contexts \\[[@ref15],[@ref16]\\]. This electronic training (e-training) significantly increased practitioners' use of client-centered communication, BBI, and referral skills in the form of offering clients a *helping conversation*. The helping conversation is a BBI that emphasizes active listening skills and motivational communication strategies to encourage and support clients' healthy behavior change \\[[@ref15]\\].\n\nSkin cancer education and training for MTs has been inconsistent and not rigorously evaluated. Although many MTs receive some skin cancer education, the format, content, duration, source, and depth of this education varies \\[[@ref17]\\]. The few skin cancer--focused in-person workshops and 1 Web-based course available to MTs \\[[@ref18]\\] have not been systematically evaluated and, to our knowledge, do not include training for VSA, client risk assessment, client-centered communication, BBI, and referral skills \\[[@ref17]\\].\n\nObjective\n---------\n\nThere is a need to develop more comprehensive, accessible skin cancer risk reduction training for MTs. Here, we describe the development of the Massage Therapists Skin Health Awareness, Referral, and Education (MTsSHARE) protocol, including the development of e-training technology, simulated client interactions, online data collection instruments, and in-person assessment of MTs' application of their training. We will describe procedures for usability and feasibility testing of the training.\n\nMethods\n=======\n\nPhase 1 (Complete): Adapting Existing Programs and Development of Training and Assessment Technology\n----------------------------------------------------------------------------------------------------\n\n### Conceptual Framework\n\nSocial cognitive theory (SCT) guided the overall study. Individuals learn and maintain new behaviors in a social context through reciprocal interaction of person, environment, and behavior. In total, 4 SCT constructs guided the overall training: (1) reciprocal determinism, or the dynamic and reciprocal interaction of MTs, their external social context, and behavioral responses to the training; (2) behavioral capability to have a helping conversation; (3) observational learning from e-training vignettes; and (4) self-efficacy, affecting behavior choices, efforts to overcome barriers to behaviors, and mastery of the behaviors \\[[@ref19]\\]. According to SCT, observations of a behavior, in this case conversing with massage clients about skin health, influence observers' perceived ability to perform the behavior (self-efficacy) and their perceived expected outcomes of the behavior, including strategies for effective performance.\n\nTo frame the BBI, we used the 4 steps of a helping conversation (awareness, understanding, helping, and relating), client-centered communication skills, client education and referral skills, and strategies for practice system involvement developed in our prior work \\[[@ref15]\\]. The helping conversation framework emphasizes a brief motivational, client-centered approach that allows a range of MT behaviors in response to the situational context (eg, new, returning, or long-term client; massage routine; practice workflow) and the client's readiness to change behavior---an approach more acceptable to MTs than proscriptive approaches to BBI used often in conventional health care contexts \\[[@ref15]\\]. This framework is also easily adaptable for e-learning dissemination.\n\n### Formative Data Collection\n\nTo initiate training development, we conducted 5 key informant telephone interviews with subject matter experts (SMEs) who were licensed MTs in Arizona and had current or previous experience in MT education or online training. The interview responses illuminated strategies to engage MTs, assets to include in the training, and approaches for discussing health issues within MT practice. Specifically, the informants stressed the importance of considering the scope of practice throughout the training development (*don't diagnose*) and the role MTs play in the health of their clients. They also helped establish the desired level of information throughout the training, for example, suggesting the inclusion of more detailed information regarding skin anatomy and the effects of UVR on the skin.\n\nWe then conducted 1 focus group with 5 additional locally practicing licensed MTs. The focus groups reviewed the themes that emerged during key informant interviews and generated data to further support the training. The key results highlighted the importance of the following:\n\n-   discussion of skin cancer risk reduction during appointments and why this activity is within the scope of practice;\n\n-   myriad ways to begin a conversation about skin cancer risk reduction with clients, including personal experience and nonjudgmental comments and questions;\n\n-   major barriers to conversing with clients about skin cancer risk reduction, such as lack of confidence and knowledge about skin cancer, and how to address these barriers;\n\n-   recommendations for how to teach MTs to have conversations with their clients about skin cancer risk reduction.\n\nThe informants focused on MTs' ethical responsibility to share important health-related information with their clients, ask permission to chart any new or changing lesions noticed, and provide a list of local dermatologists for referral purposes.\n\n### E-Learning Module Development\n\nGuided by our conceptual framework and formative data collection, we adapted content from 2 existing Web-based training programs: (1) a multimedia skin cancer risk reduction academic course, currently tailored for university students in the health sciences \\[[@ref20]\\] and (2) MT client-centered communication and referral skills modules used in a BBI training for tobacco cessation \\[[@ref16]\\].\n\nWe adapted skin cancer risk reduction content from the university's academic course and the *Surgeon General's Call to Action to Prevent Skin Cancer* \\[[@ref5]\\] to include skin cancer risk factors, sun safety, VSA, and skin lesion assessment. We endeavored to provide MTs with a refresher of the information some may have received during their professional training, while focusing on content for the expected MT-client interaction and helping conversation. We adapted client-centered communication and referral skills content from previous studies that trained MTs to offer their clients helping conversations and referrals addressing tobacco cessation \\[[@ref15],[@ref16]\\]. The 4 steps of a helping conversation as applied to skin cancer, awareness, understanding, helping, and relating, are described in [Table 1](#table1){ref-type=\"table\"}.\n\nModule development included (1) creating overall competencies and module-specific learning objectives, (2) reviewing existing curricula for structure, (3) adapting existing curricula resources or creating new multimedia content for e-learning, and (4) reviewing draft modules by SME, revising as needed. The e-training was asynchronous, interactive, and less than 2 hours in length, including accessing the modules via a Web-based learning management system, viewing and completing the modules, and completing study assessments. We chose Articulate Storyline for our e-learning course authoring software. Articulate Storyline provides the ability to create responsive modules that integrate audio, video, quiz, and activity components, allowing for a streamlined development and user interface experience.\n\nThe final training is based on 22 core competencies (see [Table 1](#table1){ref-type=\"table\"}, column 2), adapted from the learning objectives of previous helping conversation-oriented Web-based training modules \\[[@ref15]\\], which integrate the skin cancer content across the 4 steps of a helping conversation (see [Table 1](#table1){ref-type=\"table\"}, column 1). The training comprises 6 modules (1) introduction, (2) awareness, (3) understanding, (4) helping, (5) relating, and (6) closing; each module contains photo and video media produced specifically for this project, as well as interactive activities that serve as *knowledge checks* focusing on specific content and skills.\n\nSMEs in MT education, skin cancer, BBI training, online learning, public health, and information technology critiqued the content using an iterative process of review and structured/open-ended feedback proven successful in prior training projects \\[[@ref15]\\]. The massage therapy SMEs were local and national opinion leaders, respected practitioners, and educators. [Multimedia Appendix 1](#app1){ref-type=\"supplementary-material\"} contains screenshots from the Understanding module of the e-training.\n\n###### \n\nMassage Therapists Skin Health Awareness, Referral, and Education electronic training modules and competencies.\n\n  Helping conversation step and module name                                                                                                                                                                                           Competencies                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            Content examples\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------\n  *Awareness* (asking about/awareness of skin cancer risk/risk behaviors and opportunities to help)                                                                                                                                   Describe benefit of MTs as partners in skin cancer risk reduction; begin a helping conversation in a nonconfrontational and supportive way                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              [Multimedia Appendix 1](#app1){ref-type=\"supplementary-material\"}: Opportunity to help\n  *Understanding* (assessing readiness to change behaviors to reduce skin cancer risk and/or seek medical evaluation of a suspicious skin lesion and seeking understanding of the client's motivations for/against behavior change)   Apply active listening skills: open-ended questions, clarifying questions, reflective questions/statements; use positive communication skills: express empathy, avoid problem solving, avoid lecturing, avoid arguing; assess and acknowledge major barriers to skin cancer risk reduction; elicit motivators that inspire risk-reducing behaviors; reinforce motivators that inspire risk-reducing behaviors; assess and acknowledge manageable risks for skin cancer; assess and acknowledge the client\\'s willingness to take action; assess suspicious skin lesions; set realistic goals for the outcome of helping conversations   [Multimedia Appendix 2](#app2){ref-type=\"supplementary-material\"}: Understanding module screenshots\n  *Helping* (offering information about skin cancer risk reduction and referrals for medical evaluation)                                                                                                                              Recognize how to offer support and encouragement based on the client's risk profile and willingness to take action; identify different types of referral resources for professional help with skin health; provide information about professional skin health services; explain how the skills learned in this training can be applied in different situations                                                                                                                                                                                                                                                                          [Multimedia Appendix 3](#app3){ref-type=\"supplementary-material\"}: Returning client with suspicious lesion\n  *Relating* (arranging client follow-up on skin cancer risk reduction behaviors and referrals and offering ongoing encouragement for behavior change)                                                                                Seek permission to follow-up in a respectful manner; facilitate probability of follow-up by finishing the helping conversation on a positive note                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       [Multimedia Appendix 4](#app4){ref-type=\"supplementary-material\"}: Relating\n\n### Media Asset Development\n\nWe developed 5 brief MT-client interaction scenario example videos (average time of 30 seconds), using procedures from our previous consumer studies \\[[@ref21]\\]. Specifically, we asked our SMEs in MT education to (1) review training content on helping conversation goals and skills, (2) reach consensus on the purpose of the video, and (3) review video scripts and storyboards. We recruited a convenience sample of consumers to read scripts and act as massage clients; our MT consultant acted in the MT role during video recording (see [Multimedia Appendices 2](#app2){ref-type=\"supplementary-material\"} and [3](#app3){ref-type=\"supplementary-material\"} for example scenarios). We also recorded 5 testimonial videos (average time of 1 min) wherein the MTs described their experiences with providing skin cancer risk reduction information in their practices (see [Multimedia Appendix 4](#app4){ref-type=\"supplementary-material\"} for example testimonials). We used the scenario and testimonial videos to enhance and reinforce topics discussed throughout the training by embedding the videos into the Web-based modules.\n\n### Development of Simulated Decision-Making Cases\n\nUsing procedures and cases from our previous BBI research \\[[@ref16]\\], we developed electronic simulations of case-based practice of communication skills and application of skin cancer risk reduction knowledge using the Kynectiv DecisionSim \\[[@ref22]\\] platform. Each case comprises skin cancer--focused scenarios that simulate MT decision making during a helping conversation. Furthermore, 3 MT-client interaction decision paths *optimal*, *feedback required*, and *suboptimal* are accompanied by a rubric for scoring each decision (see [Figure 1](#figure1){ref-type=\"fig\"}). The rubric was informed by the competencies for the training modules ([Table 1](#table1){ref-type=\"table\"}). Each response option is associated with a tag (in the back-end database) for each decision path (see [Figure 2](#figure2){ref-type=\"fig\"}). To successfully complete the training, MTs must meet a minimum level of competence---selection of a response path that is within a specified range of the *ideal* interaction path (eg, appropriate MT response during a helping conversation when a client is open to a discussion about skin cancer risk reduction behavior vs a different MT response if the client is resistant to discussion). We developed 5 case simulations for participants to complete following the 6 training modules.\n\n![Decision path template.](resprot_v8i5e13480_fig1){#figure1}\n\n![Screenshot of decision path tags.](resprot_v8i5e13480_fig2){#figure2}\n\n### Development of Learning Management System\n\nThe learning management platform chosen to host the training modules was Desire2Learn (D2L; Desire2Learn Inc), which could seamlessly link to the case simulations. D2L allowed us to design the course components and navigation to facilitate engagement and ease of use. We designed course content to be accessed sequentially, requiring the completion of a module before accessing the next one; this ensures that participants complete the modules in the intended order, but at their own pace. The course home page features a resources section with downloadable PDFs, intended for both MT and client use. These provide skin cancer--related information, as well as tips for offering helping conversations.\n\n### Development and Adaptation of E-Learning Data Collection Methods and Assessments\n\nThe study personnel collected and managed study data using the Research Electronic Data Capture (REDCap) electronic data capture tool hosted at the University of Arizona \\[[@ref23]\\]. REDCap is a secure, Web-based application designed to support data capture for research studies, which provides (1) an intuitive interface for validated data entry, (2) audit trails for tracking data manipulation and export procedures, (3) automated export procedures for seamless data downloads to common statistical packages, and (4) automatic triggering of surveys and email correspondence.\n\nWe selected and modified our training assessments from the literature and those used in previous research. Participant assessments are timepoint-specific versions of 2 surveys: (1) baseline survey to assess participants' sociodemographic and practice-related data, as well as skin cancer risk reduction knowledge and (2) a case-based skin lesion image assessment that allowed participants the opportunity to view images of skin lesions and determine whether they were suspicious and prompted referral to a physician, or nonsuspicious. The baseline survey was modified into a posttest without sociodemographic data to assess knowledge and practice-related behaviors immediately posttraining and at 3 and 6 months.\n\nWe also developed a 5-item client survey to be advertised in participating MT offices and lobbies, inviting all clients of participating MTs to anonymously share whether their MT engaged in skin cancer risk reduction conversations during their massage visit. To further validate the MTs' application of the training, we asked super clients to conduct an immediate postmassage assessment. A super client is a study participant who participates in the study by receiving a massage and assesses MT's use of helping conversations pertaining to skin cancer risk reduction. This in-person observational assessment was adapted from the concept of unannounced standardized patients commonly used in medical and clinical education \\[[@ref24]\\].\n\n### Iterative Usability Testing\n\nThe University's Institutional Review Board approved all human subject procedures for iterative user testing of the MTsSHARE e-training. To assess the usability of the training modules, assessments, and procedures before implementing a larger quasi-experimental longitudinal study with Arizona MTs, we enrolled a convenience sample of 20 licensed MTs from throughout the United States (except Arizona). We used the predetermined feasibility study eligibility criteria to determine MT eligibility (see below), enrolling eligible MTs in 4 waves of 5. Following all training and assessment components, participants completed a 27-item usability survey adapted from the feedback form used by SMEs during module development. The usability survey questions focused on course content (the content of the modules is at the appropriate level for MTs), accessibility (it is easy to access the helping conversation simulations), and relevance (the content in the modules is relevant to my (MT's) practice). We scored all items on a 5-point scale from strongly disagree to strongly agree. After each wave of 5, we analyzed responses and made appropriate changes to study protocol, assessment, and training components as necessary. We coded open-ended responses for major categories. The massage therapy SMEs were local and national opinion leaders, respected practitioners, and educators categories using established methods for qualitative content analysis \\[[@ref25]\\]. The usability assessment was conducted from March 3, 2018 to July 31, 2018.\n\nThe overall mean scores for usability slightly improved with each iteration, increasing from a low score of 3.5 to a 5 (moderately agree to strongly agree), with an overall usability mean score of 4.96. However, the key findings from usability testing were the appropriateness of simple, seamless technology, the progression and relevance of the information presented, suggestions for additional content and general instruction, and the utility of including interactive assessments and client simulation exercises. Making changes after each wave resulted in the progressive improvement of the modules. The final version of the e-training tested well for usability and satisfaction.\n\nPhase 2 (In Progress)\n---------------------\n\n### Feasibility Study Design\n\nThe feasibility study is a single cohort design (see [Figure 3](#figure3){ref-type=\"fig\"}) with participant assessments at 4 time points: (1) immediately upon study enrollment (baseline survey and image assessment 1), (2) posttest 1 occurring immediately after training completion and image assessment, (3) posttest 2 occurring 3 months after training completion and image assessment, and (4) posttest 3 occurring 6 months after training completion and image assessment. After completing the e-training, participants receive an electronic gift card and a certificate for 1 hour of continuing education (CE), approved by the National Certification Board for Therapeutic Massage and Bodywork.\n\nA subset of 20 Tucson-based MTs will receive a visit from a trained super client at least 3 months after completing the e-training.\n\nThe phase 2 of the study, participant enrollment and data collection, is ongoing. All survey invitations are delivered via automated email from the REDCap system, triggered by items completed in an administrative survey by study staff or timepoints based on completion of the e-training.\n\n![Feasibility study protocol.](resprot_v8i5e13480_fig3){#figure3}\n\n### Participant Recruitment, Eligibility, and Enrollment\n\nOur goal for enrollment is 80 MTs practicing in the state of Arizona. We based our sample size estimation on the published literature on skin cancer training for medical students \\[[@ref26]\\] and our own research on tobacco cessation BBI with licensed MTs. Given the potential for a high attrition rate experienced in Web-based trainings \\[[@ref27]\\], we concluded that a sample of 80 participants is sufficiently powered and allows for possible attrition. The sample size analyses were conducted using PASS (V.12). This sample size is large enough to reasonably estimate, in conjunction with sensitivity analysis, relevant variance components, recruitment, and dropout rates for use in a future definitive trial \\[[@ref28]\\].\n\nTo be eligible, MTs must be aged at least 21 years, be a licensed MT in the state of Arizona, have practiced for at least 3 years, provide mainly full body massages, have access to computer with broadband internet connection and audio, and agree to forego continuing education on skin cancer for the duration of study participation. Excluded are MTs who have received continuing education on skin cancer, sun safety or client communication skills training in the past 2 years, and those who perform only partial body massages.\n\nThe recruitment began with Arizona-based MTs who initially responded to the recruitment efforts for usability testing. We then contacted state and national MT organizations, as well as Arizona-based leaders in the massage therapy business, previously known to study staff, to share recruitment materials via social media accounts.\n\nInterested MTs email or call the designated initial contact study staff member, who enters the MTs' information into a shared recruitment database in REDCap. The study staff then schedule and conduct a screening phone call. During this call, if the MT assents verbally to study participation, study staff enroll the MT into the REDCap system, which immediately sends the baseline survey link containing the electronic consent disclosure.\n\n### Participant Training and Follow-Up (In Progress)\n\nWe designed the REDCap database to send an automated notification of completion of image assessment 1 to study staff; this instructs study staff to send the e-training login instructions to the participant via email. The participants have 2 weeks to complete the 6 Web-based training modules and 5 DecisionSim cases. The study personnel check daily for training completion, emailing the certificate for 1 hour of continuing education, and logging the date of completion in the administrative survey. Completion of the training triggers posttest 1 and image assessment 2. When a participant completes the training, study personnel distribute client survey flyers to the MT in person, electronically or via US mail.\n\n### Super Client In-Person Observational Assessment\n\nSuper clients will visit Tucson-based participating MTs for an average 60-min full body massage. Super clients will receive a simple henna tattoo by study personnel, imitating a suspicious lesion, placed on their foot or ankle region before their first massage; this tattoo will be identical on each super client and will serve as a standardizing feature. Following their massage, the super client will complete a brief electronic survey about their massage experience, focusing on whether their MT engaged in a conversation about skin cancer risk reduction and whether they mentioned the suspicious lesion (henna tattoo).\n\nTo date, we have enrolled and trained 5 super clients (4+1 alternate), who will each visit 5 MTs, for a total of 20 MTs visited. We selected a convenience sample of super clients to represent a variety of demographic characteristics (age, gender, phenotype, health history, and sun protective behaviors).\n\n### Data Analysis\n\nFeasibility outcomes, including recruitment and dropout, training completion, overall client feedback, and MT satisfaction, will be described using frequencies and percentages and 95% CIs.\n\nThe longitudinal measures will use appropriate mixed models (linear for continuous outcomes and generalized linear with a logistic link for binary) using time categorically to protect against model misspecification. Comparisons of baseline with 3- and 6-month measures will be carried out using contrasts within these models. The mixed-models are robust to missing outcome data (including dropout) and model misspecification \\[[@ref29],[@ref30]\\].\n\nSensitivity/specificity across 4 timepoints will be compared for image assessment scores. For each timepoint we will assess the following parameters for image assessment: sensitivity, specificity, the likelihood ratio for a positive result, and the likelihood ratio for a negative test result. We will evaluate separate bivariate logistic regression models for each set of image assessments to determine the odds ratio in predicting the correct image. We will evaluate separate models including the scores for each timepoint to determine the areas under the receiver operating characteristic (ROC) curves for image assessments. The area under the ROC curve measures the probability of correctly identifying a true negative (not suspicious) or true positive (suspicious) image.\n\nClient survey data and super client data will be analyzed with descriptive statistics. We will correlate scores from the super client assessment with the DecisionSim scores to further validate MTs' application of helping conversation skills learned in the training. We will conduct an optional debriefing webinar for the 80 MTs in the third year to gain further information about their experience with the curriculum.\n\nMixed-effects linear regression models for longitudinal data will be \ufb01tted to evaluate intervention outcomes adjusted for participant characteristics, for example, age, gender, years in practice, geographical area, and client workload. The mean differences in each of the primary outcomes will be evaluated in separate models, including the covariates as fixed effects and subjects as random effects. We may also consider geographical area as random effect (urban vs rural). In this case, geographical area and subjects will be fitted as random effects to account for the correlation within geographical area and serial intrasubject correlations. Predictor variables with multiple categories will be entered as indicator variables. For dichotomized intervention effects, we will use mixed-effects logistic regression models.\n\nResults\n=======\n\nFor Phase 2, we have screened 77 MTs who have expressed interest in participating. Of those, 14 were either not interested or not eligible (either lacked time to participate or did not see an average of 10 clients per week) and 15 did not follow up after contact attempts were made. We enrolled and consented the remaining 48 MTs. At the time of the paper submission, 11 enrolled MTs had dropped from the study, owing to lack of time to participate. Of the 37 MTs still enrolled, 32 have completed the training, with the remaining 5 having begun but not yet completed the training. We will close recruitment in August 2019.\n\nDiscussion\n==========\n\nPrincipal Findings\n------------------\n\nThe current prevention and early detection strategies have not had a significant impact on reducing the public health burden of skin cancer \\[[@ref5]\\]. We used a rigorous strategy to enlist MTs as partners in skin cancer prevention and detection, developing innovative e-training and assessment protocols. The American Massage Therapy Association estimates that there are 335,000 to 385,000 licensed or certified MTs in the United States, who see about 39.1 million clients annually \\[[@ref31]\\]. Thus, MTs are a largely untapped resource for reducing skin cancer risk. A search of MT training on PubMed and Google Scholar revealed that most articles focus on massage therapy as an intervention and the health outcomes of massage. Few scholarly articles addressed training MTs for a specific skill following their primary professional education \\[[@ref15],[@ref16]\\]. Muramoto et al \\[[@ref16]\\] were the first to successfully develop and implement e-training (BBI training and competency evaluation) of complementary and alternative medicine providers (including MTs) for screening clients for tobacco use and encouraging tobacco cessation. These authors also were the first to develop e-training for these specialized providers.\n\nThe structured and rigorous development approach for this skin cancer risk reduction and BBI e-training for MTs begins to fill a gap in skin cancer risk reduction research. We surveyed 100 MTs in an elecronic, national survey where we asked for the MTs' perceptions of conversations with clients related to skin cancer prevention, as well as detection \\[[@ref32]\\]. The 2 published studies have targeted skin cancer risk reduction in convenience samples of MTs who were attending national MT conferences. One study surveyed 262 MTs to assess their comfort level regarding potential assessment of suspicious skin lesions \\[[@ref17]\\]. The other study reported findings from a face-to-face, 4-hour education session that provided information only to 114 MTs \\[[@ref33]\\]. No previous studies have addressed how MTs could integrate this information into the context of a client visit via client education, or communication skills, such as a BBI (helping conversation) to encourage skin cancer risk reduction. Our e-training approach offers greater learner accessibility, increased convenience, and greater scalability \\[[@ref34]\\]. Thus, the e-training format has the potential to reach many more MTs, nationally.\n\nWe found few other e-learning opportunities pertaining to skin cancer, most of which targeted conventional health care providers with a goal of increasing competencies in diagnostic knowledge and skills competency \\[[@ref35]-[@ref37]\\]. These ranged from several Web-based modules to video training delivered by electronic links \\[[@ref37]\\]. In these studies, providers had a positive impression of the Web-based curriculum, and in one case, increased the likelihood of discussion with patients about skin cancer. The accessibility, effectiveness, and popularity of the curriculum indicated potential for implementation in the primary care setting. Our e-training is designed to be brief, yet engaging, informative, and integrated into the context of a typical client visit to an MT. MTs can access the training when convenient and move from one module to the next at their own pace, both of which are important for learner control and engagement \\[[@ref38]\\].\n\nPrevious training targeting MTs did not appear to be pilot tested or assessed for usability. Our use of formative and summative evaluations along with predesignated *stopping rules* (ie, 4 iterations) represented the ideal conceptualization of usability \\[[@ref39]\\]. The ease of navigation of the training modules and available resources made this training appealing to the participating MTs. The training incorporated highly interactive, scenario-based, simulated helping conversations focused on skin cancer risk reduction, and the simulations provided participants with opportunity to interact with the training, apply knowledge gained, and practice skills learned, reflecting the SCT theory. These features also are important to enhancing e-learning \\[[@ref34]\\].\n\nBarriers and Opportunities\n--------------------------\n\nThe preliminary results reveal anticipated difficulties with recruitment and retention within the MT population. No previously published studies of MTs as participants have addressed recruitment challenges. For the in-progress feasibility study, barriers to enrollment have related to practice-related eligibility conditions, such as number of years in practice and number of clients seen per week. The primary barrier to retention following enrollment has been a self-professed lack of time to participate. It is encouraging that, of the eligible and enrolled MTs, 67% (32/48) have completed the training and progressed to follow-up assessments. Offering incentives in the form of monetary compensation, as well as continuing education credit, has been a useful approach to address both recruitment and retention.\n\nThis paper provides an overview of our systematic approach to developing rigorous e-training for MTs to enable them to be partners in skin cancer risk reduction. The phase 2 results will explicate the feasibility of the e-training approach for further efficacy testing.\n\nThis study was funded in part by the Arizona Biomedical Research Centre as made available through the Arizona Department of Health Services, ABRC/ADHS16-162518. The content and findings are solely the responsibility of the authors and do not necessarily represent the official views of the Arizona Biomedical Research Centre, Arizona Department of Health Services.\n\nThe services and products in support of the research project were generated by the University of Arizona Cancer Center Behavioral Measurement and Interventions Shared Resource, supported, in part, with funding from National Institutes of Health-National Cancer Institute (NIH-NCI) Cancer Center Support Grant P30 CA023074.\n\nConflicts of Interest: None declared.\n\nOpportunity to help.\n\nUnderstanding module screenshots.\n\nReturning client with suspicious lesion.\n\nRelating.\n\nBBI\n\n:   brief behavioral intervention\n\nD2L\n\n:   Desire2Learn\n\ne-learning\n\n:   electronic learning\n\ne-training\n\n:   electronic training\n\nMT\n\n:   massage therapist\n\nMTsSHARE\n\n:   Massage Therapists Skin Health Awareness, Referral, and Education\n\nREDCap\n\n:   Research Electronic Data Capture\n\nROC\n\n:   receiver operating characteristic\n\nSCT\n\n:   social cognitive theory\n\nSME\n\n:   subject matter expert\n\nUVR\n\n:   ultraviolet radiation\n\nVSA\n\n:   visual skin assessment\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\nCerebral vasospasm is a serious complication of subarachnoid hemorrhage (SAH); delayed narrowing of the large capacity arteries and cerebral vasospasm are associated with significant morbidity and mortality following a SAH.\\[[@ref6]\\] The anterior choroid artery (AChA) is the most important branch of the supraclinoid segment of the ICA. It supplies the optic tract, uncus, lateral part of the geniculate body, posterior two-thirds of the posterior limbs of the internal capsule, optic radiation, and choroid plexus of the lateral ventricles. The classical features of AChA occlusion include contralateral hemiplegia, hemianesthesia and hemianopsia. Vasospasm of the choroid arteries may be severe when a SAH causes damage to the choroid plexus. This can affect structures that play an important role in immune, endocrine, detoxifying, thermoregulatory, and secretory functions of the brain; such abnormalities are associated with a poor prognosis.\\[[@ref10][@ref23][@ref25][@ref30][@ref31]\\] The response of AChAs, in a rabbit model of SAH, was evaluated in this study.\n\nThe cerebral arteries are innervated by several systems that contribute to the autonomic control of cerebral blood flow. Parasympathetic fibers affect vasodilation and sympathetic fibers cause vasospasm of the cerebral arteries.\\[[@ref16]\\] Cerebrovascular sensory nerves originate mainly from the first division of ipsilateral trigeminal ganglion (TGG), known as the trigemino-cerebrovascular system.\\[[@ref2][@ref19][@ref32]\\] Trigeminal sensory nerves project to the ipsilateral ICA, MCA, ACA, the rostral part of the basilar artery, the PCA, and the AChA.\\[[@ref1]\\]\n\nNeuro-humoral mechanisms have been suggested as important factors associated with the cerebral artery vasospasm after a subarachnoid hemorrhage.\\[[@ref5][@ref6][@ref22][@ref29]\\] The goal of this study was to determine the relationship between the number of neurons in the ophthalmic root of the TGG and changes in the diameter of the AChA, associated with SAH-induced vasospasm.\n\nMATERIALS AND METHODS {#sec1-2}\n=====================\n\nThirty-two anesthetized, adult male New Zealand rabbits were used for this study. The Ethics Committee of Atat\u00fcrk University, Medical Faculty, approved the animal protocols. Animal care and experimental protocols were conducted according to the guidelines set forth by the same ethics committee. The animals were randomly divided into three groups: SAH (n=17), isotonic saline solution (SHAM; n=7), and control (n=8) groups. A balanced, injectable anesthetic was used to reduce pain and mortality. Anesthesia was induced with isoflurane, given by a facemask, and 0.2 mL/kg of the anesthetic combination of Ketamine HCL, 150 mg/1.5 mL; Xylazine HCL, 30 mg/1.5 mL; and distilled water, 1 mL, was subcutaneously injected prior to the surgery. All animals were monitored for changes in the electrocardiogram, respiration patterns, and blood oxygen concentration during the experiment. All parameters were recorded by a camera and analyzed by physicians that did not know to which experimental group the individual animals belonged. During the procedure, a dose of 0.1 mL/kg of the anesthetic combination was used when required. In 17 (n=17) of the animals, autologous blood (0.5 mL) was taken from the auricular artery. While the head of the animal was held in a hyperflexed position, the posterior notch of the foramen magnum was identified, and the cisterna magna was entered; CSF was then aspirated. When the identification of the cisterna magna was confirmed, blood was injected using a 22-gauge needle, over about 1 minute, in the SAH group; 1 mL of isotonic saline solution was injected in the same way in the seven (n=7) animals of the SHAM group. The remaining eight (n=8) animals not subjected to this procedure were considered the control group. The animals were followed for 10 days without any medical treatment and then killed. The time was selected based on relief of the vasospasm. The number of ganglion neurons is thought to be associated with vasospasm. The effects of SAH on vasospasm were studied by the removal of all AChAs and trigeminal ganglia, bilaterally, for histological examination. Specimens were stored in a 10% formalin solution for 7 days, after which 5 *\u03bc*m tissue sections were cut and stained with hematoxylin and eosin.\n\nTo estimate the neuron density of the ophthalmic division of the trigeminal nerve, all trigeminal roots, together with their ganglions, were extracted bilaterally. The specimens were then horizontally embedded in paraffin blocks to evaluate all roots during the histopathological examination. The physical dissector method\\[[@ref27]\\] was used to evaluate the number of neurons in the TGG. The advantages of this method are that it easily estimates particle number; can be readily performed; is intuitively simple; is free from assumptions about particle shape, size, and orientation; and is unaffected by overprotection and truncation. Two consecutive sections (dissector pairs) obtained from tissue samples with named references were mounted on each slide. The reference and look-up sections were reversed to double the number of dissector pairs; this eliminated the need for obtaining new sections. The mean neuronal density of the ophthalmic root of the TGG (NvGN) per mm^3^ was estimated using the following formula: NvGN = \u2211Q -- N/t \u00d7 A, where \u2211Q -- N is the total number of counted neurons appearing only in the reference sections, t is the section thickness, and A and B are the areas of the counting frames \\[Figure [1a](#F1){ref-type=\"fig\"},[b](#F1){ref-type=\"fig\"}\\]. The Cavalieri volume estimation method was used to obtain the total number of neurons in each specimen.\\[[@ref8][@ref12][@ref27]\\] The total number of neurons was calculated by multiplying the volume (mm^3^) and the numerical density of neurons in each TGG. In addition, the mean number of neurons in the ophthalmic division of the TGG was counted bilaterally in every animal \\[[Table 1](#T1){ref-type=\"table\"}\\].\n\n![(a, b) Stereologic cell counts of the ophthalmic division of the TGG in a rabbit. Use of the physical dissector method where the micrographs of the same fields of view (a, b) were taken from two parallel, adjacent thin sections separated by 5 \u03bcm. The numerical density of the neurons was calculated as NvGN= \u2211Q -- N/t \u00d7 A. In this application, the nucleoli marked as '2,3,4,8' are dissector particles in A. Section B shows them as they disappeared. The nucleoli marked as '1,5,6,7' do not represent dissector particles in A. Section B shows '1 and 8' as they disappeared (H and E, 100, LM)](SNI-2-77-g001){#F1}\n\n###### \n\nVasospasm index of the AChAs and the neuronal density of the ophthalmic root of the trigeminal ganglion\n\n![](SNI-2-77-g002)\n\nThe AChAs were obtained from the coronal brain sections at the level of the AChAs entering into the lateral ventricles. They were also stained with H&E. For the calculation of the vasospasm index of the AChAs, all AChAs were accepted as a cylinder, in view of their morphological characteristics; simple geometric formulas were used to estimate their wall ring values. As a measure of the degree of vasospasm, the use of the AChAs vasospasm index was preferred to use of only measurements of the lumen radius; it can be readily performed, is intuitively simple, more reliable, free from assumptions about vessel diameter of various segments and is unaffected by overestimation errors of the radius of the AChAs. The AChAs of all animals were cut 20 segments away from the point where the internal carotid arteries entered the choroid plexus. Then, 20 histopathology sections, 5 *\u03bc*m apart, were obtained, using a microtome, for each designation, and are represented by the lines 1, 2, 3,... and 20. The average of 10 diameters, of 10 cross-sectional areas, was recorded as the mean diameter. A single line in the figures represents one of them. The mean external and internal (luminal) diameters of each section was measured; the external radius is represented by R~1~ and the internal radius is represented by r~1~. The mean external radius of the anterior choroid arteries was calculated as R~1~ = R~1~ + R~2~ + R~3~ +....R~20~/20; the lumen radius was calculated as r~1~ = r~1~ + r~2~ + r~3~ +....r~20~/20 \\[[Figure 2](#F2){ref-type=\"fig\"}\\]. The wall ring surface values were calculated using the following formula: S~1~ = \u03c0R~1~^2^ -- \u03c0r~1~^2^. The lumen surface area was calculated using the same method. Therefore, the lumen surface value (S~2~) = \u03c0r~1~^2^. The vasospasm index (VI) was calculated as the proportion of S~1~/S~2~ . The vasospasm index (VI) =S~1~/S~2~ = \u03c0R~1~^2^ -- \u03c0r~1~^2^/\u03c0r~1~^2^ = \u03c0(R~1~^2^ -- r~1~^2^)/\u03c0r~1~^2^ = R~1~^2^ -- r~1~^2^/r~1~^2^.\n\n![The histopathological appearance of the right AChA at the entrance level in a normal rabbit (LM, H and E, \u00d7100). On the left side, minimal forms of CP are observed (LM, H and E, \u00d720)](SNI-2-77-g003){#F2}\n\nIn summary, VI= (R~1~^2^ -- r~1~^2^)/r~1~^2^. In addition, the vasospasm index of AChAs in the rabbits with SAH was calculated as the same manner; VI= (R~2~^2^ -- r~2~^2^)/r~2~^2^ . As the VSI increases, the degree of arterial vasospasm also increases. Vasospasm was defined as mild with a reduction of 25% of the normal diameter of any arterial segment of AChA and as severe vasospasm with a decrease of 40%.\\[[@ref13][@ref14]\\]\n\nThe differences between the VSI of AChAs and neuron density of the ophthalmic root of the TGG were compared statistically. For the statistical analysis, SPSS ver. 15.0 was used. The mean \u00b1 standard deviation of the variables is reported. Since the data showed a normal distribution, intergroup differences were assessed using a one-way ANOVA. The presence of homogeneous variance necessitated the use of the Tukey test for comparisons between two groups. A *P*\\<0.05 was accepted as statistically significant.\n\nRESULTS {#sec1-3}\n=======\n\nTwo of the animals (n=2) died within the first week and the remaining animals (n=30) were followed for 10 days and then killed. The following clinical findings were frequently observed during the premortem period of the dead animals and the five living animals: signs of meningeal irritation, consciousness, seizures, fever, apnea, cardiac arrhythmia, and breathing disturbances.\n\nTGGs were identified from the trigeminal impressions located on the upper surface of the petrous bones. They were fusiform shaped, and the volumes were estimated to be 2 \u00d7 1.5 \u00d7 2 mm^3^ . All trigeminal nerves have three main branches: the ophthalmic, mandibular, and maxillary. The neuron density of only the ophthalmic root division of the TGGs was examined \\[Figure [1a](#F1){ref-type=\"fig\"},[b](#F1){ref-type=\"fig\"}\\].\n\nMorphological examinations of the brains showed the AChAs at the cisternal segment, extending from its origin to the choroid fissure, and the plexus segment, extending from the choroid fissure to the area where it enters into the choroid plexus of the lateral ventricles. The mean diameter of the AChA examined was 0.35\u00b10.10 mm. The AChA convolutions were more prominent in the animals with severe vasospasm associated with SAH compared to the SHAM, control, and mild vasospasm associated with SAH groups.\n\nTo estimate the AChA volume, square-lined glass plates were used and photographs were taken under the microscope during histopathological examinations. The inner elastic membrane (IEM) was less convoluted and the luminal surface area was greater in the SHAM, control \\[[Figure 2](#F2){ref-type=\"fig\"}\\], and mild vasospasm associated SAH groups \\[[Figure 3](#F3){ref-type=\"fig\"}\\]. The following features were observed specifically in the severe vasospasm associated SAH group: arachnoiditis, pia-arachnoid adhesions, AChA narrowing, IEM convolutions, intimal edema formation and endothelial cell shrinkage, desquamation, and loss \\[[Figure 4](#F4){ref-type=\"fig\"}\\].\n\n![The histopathological appearance of the right AChA at the postorigin level in a SAH rabbit model (LM, H and E, \u00d7100). On the left side, minimal forms of CP are observed (LM, H and E, \u00d720) (mild vasospasm associated SAH group)](SNI-2-77-g004){#F3}\n\n![The histopathological appearance of the right AChA at the postorigin level in a SAH rabbit model (LM, H and E, \u00d7100). On the left side, minimal forms of CP are observed (LM, H and E, \u00d720) (severe vasospasm associated SAH group)](SNI-2-77-g005){#F4}\n\nThe mean external diameter/internal diameter of the AChAs using the segmental model was estimated as 115\u00b120 \u03bcm/95\u00b120 \u03bcm; the mean vasospasm index of the AChAs was 0.46; the mean neuronal density of the ophthalmic root of the TGGs was 8290\u00b11480 neurons/mm^3^ for all animals (n=32). The mean external diameter/internal diameter of the AChAs was estimated as 140\u00b130 \u03bcm/120\u00b130 \u03bcm, and the mean vasospasm index of the AChAs was 0.36. The mean neuronal density of the ophthalmic root of the TGGs was 8350\u00b1390 neurons/mm ^3^ in the control group (n=8). The mean external diameter/internal diameter of the AChAs was estimated as 130\u00b120 \u03bcm/110\u00b120 \u03bcm, and the mean vasospasm index of the AChAs was 0.39. The mean neuronal density of the ophthalmic root of the TGGs was 8550\u00b1650 neurons/mm ^3^ in the SHAM group (n=7). The mean external diameter/internal diameter of the AChAs was estimated as 100\u00b115 \u03bcm/80\u00b115 \u03bcm, and the mean vasospasm index of the AChAs was 0.56. The mean neuronal density of the ophthalmic root of the TGGs was 8200\u00b1600 neurons/mm^3^ in the SAH group (n=17). Severe vasospasms were observed in seven rabbits with SAH, and mild vasospasms were observed in the remaining 10 rabbits with SAH. The mean external diameter/internal diameter of the AChAs was estimated as 110\u00b120 \u03bcm/90\u00b115 \u03bcm, and the mean vasospasm index of the AChAs was 0.49. The mean neuronal density of the ophthalmic root of the TGGs was 9800\u00b1724 neurons/mm^3^ for the mild vasospasm with the SAH group (n=10). However, the mean external diameter/internal diameter of the AChAs was estimated as 90\u00b115 \u03bcm/60\u00b110 \u03bcm, and the mean vasospasm index of the AChAs was 1.25. The mean neuronal density of the ophthalmic root of the TGGs was 6342\u00b1557 neurons/mm^3^ in the severe vasospasm with the SAH group (n=7),\n\nThe estimated mean AChAs VSI was significantly higher and the mean neuronal density of the ophthalmic root of the TGG was significantly lower in the severe vasospasm associated with the SAH group when compared to the control, SHAM, and mild vasospasm associated with SAH groups (*P*\\< 0.05) \\[[Table 1](#T1){ref-type=\"table\"}\\]. The ophthalmic root of the TGG neuron density in the seven rabbits that developed severe vasospasm was significantly less than observed in the 10 rabbits with mild vasospasm. The results showed an invers relationship between low neuron density in the ophthalmic root of the TGG and severity of the AChA vasospasm \\[[Figure 5](#F5){ref-type=\"fig\"}\\].\n\n![There was an inverse correlation between low neuron density in the TGG and the severity of the AChA vasospasm (r=-0.55, \\<0.05, Pearson correlation analysis)](SNI-2-77-g006){#F5}\n\nDISCUSSION {#sec1-4}\n==========\n\nVasospasm is pathophysiologically characterized by narrowing of the vascular lumen; this may develop following a SAH. Vasospasm after a SAH is associated with significant morbidity and mortality.\\[[@ref7][@ref16][@ref17]\\] It is widely believed that blood, or its degradation products, causes vasospasm by directly affecting the walls of the cerebral arteries.\\[[@ref29]\\] The pathogenesis of the arterial narrowing that occurs following a SAH is not fully understood.\n\nNumerous neuronal, humoral, and chemical factors are involved in cerebrovascular innervations. Parasympathetic fibers have vasodilation effects, and sympathetic fibers have vasospastic effects on cerebral arteries. A variety of autonomic nerve fibers provide neural innervation to cerebral vascular structures. The postganglionic fibers, of the ciliary ganglion, of the third cranial nerve, the sphenopalatine ganglion of the seventh cranial nerve, the otic ganglion of the ninth cranial nerve, and the ganglion of the fifth cranial nerve, are involved in the parasympathetic outflow, which causes vasodilation of the cerebral arteries.\\[[@ref16][@ref20]\\] The sympathetic innervations, from the postganglionic fibers of the superior sympathetic ganglion, which contribute to the carotid plexus, cause vasospasm of the cerebral vessels.\\[[@ref14]\\]\n\nThe extracerebral and intracerebral cranial circulations are innervated by trigeminal nerve fibers. The branches of the ophthalmic division of the trigeminal nerve provide the primary innervations of the large cerebral arteries. The trigeminal nerve fibers release transmitters that cause vasodilation of the cerebral arteries. Direct stimulation of the trigeminal nerve results in ipsilateral cerebral vasodilation after a SAH.\\[[@ref9][@ref11][@ref28]\\] Unilateral postganglionic trigeminal lesions result in ipsilateral constriction of the cerebral arteries and consequent decrease of the CBF; however, pre-ganglionic lesions do not affect the baseline diameter of cerebral arteries.\\[[@ref26][@ref29]\\]\n\nVasoactive neurotransmitters, neuromodulators, and neurochemicals are released from trigeminal nerve terminals; these compounds affect large intracranial and extracranial blood vessels.\\[[@ref20]\\] Aydin *et al*.\\[[@ref3]\\] reported that decreased neuron density in the petrosal ganglion was an etiologic factor associated with high blood pressure. Onder *et al*.\\[[@ref22]\\] reported that decreased neuron density of the TGG was associated with a lower basilar artery volume and that a decreased number of neurons in the TGG resulted in a poorer prognosis of rabbits with a SAH. In their study, the animals that died in the SAH group had the lowest number of neurons in the TGG. In this study, two animals in the group with severe vasospasm associated with a SAH died, and the number of neurons in their TGG was significantly lower than in the controls, SHAM, and mild vasospasm associated SAH groups. In addition, Kanat *et al*.\\[[@ref15]\\] reported that the neuron density of the C3 dorsal root ganglia (C3DRG) might be an important factor in the regulation of the anterior spinal artery (ASA) volume and spinal cord blood flow. The low neuron density of C3DRG was suggested to be an important factor associated with the pathogenesis of severe ASA vasospasm associated with a SAH.\n\nThe goal of this study was to investigate the potential relationship between the neuron density of the ophthalmic root of the TGG and the severity of AChA vasospasm in rabbits, following a SAH. The two major findings of this investigation were (i) a correlation between the neuron density of the ophthalmic root of the TGG and ACA volume, and (ii) the association of a low neuron density, of the TGG, with the pathogenesis of ACA vasospasm after a SAH. In order to evaluate the severity of AChA vasospasm associated with a SAH, the vasospasm index of the AChA was used as the preferred parameter for assessment rather than the radius of the vessel lumen. As the VSI increased, the degree of arterial vasospasm also increased.\n\nThe results of this study showed that the mean AChA VSI was higher in animals with lower neuron density in the trigeminal ganglia. A correlation was found between the neuron density of the ophthalmic division of the trigeminal ganglion and the degree of AChA vasospasm (*P*\\<0.05). In particular, the ophthalmic root of the TGG neuron density of the seven rabbits that developed severe vasospasm was significantly lower than in the 10 rabbits that had mild vasospasm. Therefore, a lower neuron density in the TGG might be associated with reduced vasodilation of the AChAs. Compounds associated with vasodilation are synthesized by TGG neurons and secreted from the nerve terminals that end in the cerebral arteries. A reduced number of neurons may result in deficient quantities of vasodilator molecules in the trigeminal ganglia and/or cerebrovascular sensory nerves, which affects the trigemino-cerebrovascular system and the AChAs, and might subsequently increase the severity of vasospasm in the AChAs after a SAH. Therefore, novel approaches to improve cerebral blood flow, under these conditions, are being investigated.\n\nCONCLUSION {#sec1-5}\n==========\n\nThe number of neurons in the ophthalmic root of the TGG may play an important role in the regulation of cerebral artery vasospasm (AChAs as well) and, therefore, in the regulation of the cerebral circulation. In rabbits, a significant reduction of the number of the ophthalmic root of the TGG neurons, after a SAH, impaired the sensory innervation of cerebral arteries and the AChAs, and might play an important role in the development of cerebral vasospasm. The results of this study showed that the pathological processes involved in vasospasm, associated with a SAH, significantly affected the choroid arteries. In addition, neurological deficits, resulting in ischemic injury and apoptosis of the choroid plexus, were observed.\n\nAvailable FREE in open access from: <http://www.surgicalneurologyint.com/text.asp?2011/2/1/77/82084>\n"}
+{"text": "1. Introduction {#sec1-molecules-24-03605}\n===============\n\nPlants are a source of a wide spectrum of compounds such as polyphenols, carotenoids, glucosinolates, and lignans, among others. These phytochemicals provide potential beneficial properties to each plant matrix \\[[@B1-molecules-24-03605]\\]. *Aloe vera* has been long used thanks to its curative and therapeutic properties. It has been reported that only the pulp has more than 75 bioactive compounds \\[[@B2-molecules-24-03605]\\]. *Aloe vera*, originally from Africa, belongs to the genus *Aloe*, and it is a perennial, succulent xerophyte grown in temperate and sub-tropical regions of the world. *Aloe vera* or *Aloe barbadensis* is part of the *Asphodelaceae* family, of which there are over 360 known species. There are several species under the genus *Aloe*, including *Aloe vera, Aloe barbadensis, Aloe ferox, Aloe chinensis, Aloe indica, Aloe peyrii* etc. Amongst these, *Aloe vera Linn syn. Alo barbadensis Miller* is accepted unanimously as the ideal botanical source of *Aloe* \\[[@B3-molecules-24-03605]\\]. The pulp or gel of *Aloe vera* is the part of the plant that is of great commercial, pharmacological, alimentary, industrial, and cosmetic importance \\[[@B3-molecules-24-03605],[@B4-molecules-24-03605]\\]. *Aloe vera* gel contains more than 98% water. The carbohydrate acemannan accounts for greater than 60% of the solid matter found in the gel \\[[@B5-molecules-24-03605]\\]. Polysaccharides such as acetylglucomannans (acemannans) are glucose-bound mannose units linked by \u03b2--(1\u21924) bonds constituting the backbone of the polysaccharide \\[[@B6-molecules-24-03605]\\]. [Figure 1](#molecules-24-03605-f001){ref-type=\"fig\"} shows a representation of *Aloe vera* leaves, which have three portions: (1) yellow sap, constituted mainly of anthraquinones (1,8-dihydroxyanthraquinone derivatives), (2) pulp or parenchymal tissue (mannans, galactans, arabinans, arabinogalactans, pectic substances, and polysaccharides containing glucuronic acid), and (3) exocarp (cellulose, hemicelluloses, and lignin) \\[[@B4-molecules-24-03605]\\]. Additionally, the gel contains phenolic compounds that can be soluble free or conjugated soluble and insoluble. Conjugated soluble phenolic compounds bind soluble molecules as carbohydrates, proteins, and lipids by esterification in the carboxylic moiety or etherification in the hydroxyl group. Insoluble phenolic compounds generally covalently bind polymers, such as polysaccharides and lignins through an ester bond and are only released from the matrix through acidic, alkaline or enzymatic hydrolysis \\[[@B7-molecules-24-03605]\\]. It has been reported that these phenolic compounds have properties for the treatment of diabetes, tumours or ulcers, among others \\[[@B8-molecules-24-03605],[@B9-molecules-24-03605]\\]. Patel et al., using HPTLC, reported the presence of alkaloids, tannins, steroids, sugars, and triterpenes in an extract of *Aloe vera* \\[[@B10-molecules-24-03605]\\].\n\nSoluble or fermentable fibre components, such as mannans, galactans, arabinans, arabinogalactans, and pectic substances, could have prebiotic effects. Prebiotics are carbohydrate-like compounds that can be used in the food industry to modify the composition of microbiota species to benefit human health \\[[@B11-molecules-24-03605]\\]. Prebiotics mostly target bifidobacteria and lactobacilli, which are two kinds of probiotics \\[[@B12-molecules-24-03605]\\]. Recent research has suggested that combining both prebiotics and probiotics, namely, symbiotic, can fight chronic diseases since colonic fermentation produces short chain fatty acids (SCFAs). SCFAs are generated as a final product of intestinal microbiota metabolism, mostly out of non-digestible dietary fibre. Acetate (C~2~), propionate (C~3~), and butyrate (C~4~) have different carbon chain lengths. There is research on the beneficial effects of several components of *Aloe vera* on health; however, there are few studies on the changes in fibre and there are no reports available on the estimation of free phenolic during the digestive and fermentative process, where components such as phenolic compounds and the production of SCFAs could have beneficial effects on human health. Therefore, the present work studied the changes in fibre and phenolic compounds produced during in vitro digestion of gel and a polysaccharide extract from *Aloe vera* (AV and AP, respectively), after which these fractions were subjected to in vitro colonic fermentation to evaluate the changes in antioxidant capacity and production of SCFAs during fermentation.\n\n2. Results and Discussion {#sec2-molecules-24-03605}\n=========================\n\n2.1. Physicochemical Characterization of Aloe Vera Gel (AV) {#sec2dot1-molecules-24-03605}\n-----------------------------------------------------------\n\nThe pH of AV was 5.4, similar to the value reported by Calder\u00f3n-Oliver et al., \\[[@B13-molecules-24-03605]\\]. The total acidity was 0.007 \u00b1 0.02 g/100 g moist sample. The content of total soluble solids (TSS) of AV gel was 0.42 \u00baBrix. The components present in the TSS can be minerals, vitamins, enzymes, polysaccharides, phenolic compounds, and organic acids, among others \\[[@B13-molecules-24-03605]\\]. [Table 1](#molecules-24-03605-t001){ref-type=\"table\"} shows that the moisture content of the fresh gel extracted from the leaf was the highest value, which agrees with the water activity of the gel and results reported by Hamman \\[[@B6-molecules-24-03605]\\]. The proteins in the gel could be lectins and lectin-like substances \\[[@B13-molecules-24-03605]\\].\n\nThe dietary fibre content was over 55%, mostly soluble fibre (39.53%), which can be fermented in the colon because it contains soluble oligosaccharides and non-digestible oligosaccharides that could be beneficial to individuals who consume soluble fibre \\[[@B6-molecules-24-03605]\\]. The percentage of non-fermentable or insoluble AV gel fibre (cellulose, lignin, and hemicellulose) was 15.58 \u00b1 2.04%. When comparing the results against the amount of total dietary fibre (TDF), we observed that AV gel possesses a considerable amount of soluble fibre, which has been reported to provide benefits to diabetes, obesity, gastritis, and neurodegenerative diseases patients who consume it regularly \\[[@B14-molecules-24-03605]\\]. Interestingly, the polysaccharide content of AV was approximately 71% on a dry basis, which is relevant because polysaccharides have been considered the main component responsible for most of the beneficial properties attributed to the *Aloe vera* plant \\[[@B15-molecules-24-03605]\\]. The chemical composition may vary with respect to what is reported by other authors, stressing that the ground where the plant grows, the weather, and age of the plant play key roles in the chemical composition of *Aloe vera* \\[[@B15-molecules-24-03605]\\].\n\n2.2. Identification of Phenolic Compounds in AV by UPLC-MS {#sec2dot2-molecules-24-03605}\n----------------------------------------------------------\n\nSix compounds were identified by UPLC-MS in AV ([Table 2](#molecules-24-03605-t002){ref-type=\"table\"}). Two signals were presented between 4 and 5 min, both with masses of *m/z* 417.18 \\[M--H\\]-; these signals were identified as the diastereomers aloin A and aloin B \\[[@B16-molecules-24-03605],[@B17-molecules-24-03605]\\]. In addition, a peak of *m/z* 555.25 \\[M -- H\\]- was observed at 4.5 min that could be related to aloenin-2\u2032-*p*-coumaroyl ester. At a retention time of 3.15 min, there was a peak of *m/z* 561.36 \\[M--H\\]- that could be the aloinoside isomers \\[[@B17-molecules-24-03605]\\]. At a retention time of 3 min, a mass of *m/z* 433.19 \\[M--H\\]- was observed, which suggests the compound 10-hydroxyaloin, and at 1.54 min, a peak of *m/z* 393.18 \\[M--H\\]- was observed and corresponded to aloesin, according to the mass. [Table 2](#molecules-24-03605-t002){ref-type=\"table\"} shows the polyphenols identified, aloin A and B, aloesin, 10-hydroxyaloin A or B, and aloenin-2\u2032-*p*- coumaroyl ester, which are anthraquinones reported as powerful stabilizers in oxidative stress \\[[@B17-molecules-24-03605]\\]. Aloesin is a natural polyphenol originated from *Aloe* plants. Aloesin and/or *Aloe* polysaccharides can reduce systemic oxidative stress by acting directly as potent anti-oxidants and also indirectly by regulating the production of adiponectin and gene expression pathways related to insulin sensitivity, glucose transportation, and fatty acid biosynthesis \\[[@B18-molecules-24-03605]\\].\n\nAloin A and B are some of the main components of *Aloe vera* gel, which are not only laxatives but are also known to have antioxidant properties at low concentrations \\[[@B17-molecules-24-03605]\\]. Aloinoside compounds are also anthraquinones, along with aloesin, which exhibits great antioxidant activity and is used to lower oxidative stress indices in diseases such as diabetes \\[[@B17-molecules-24-03605]\\].\n\n2.3. Total Fibre Content of Non-Digestible Fibre Fractions AV-TNDF and AP-TNDF and of Their Soluble (AV-SNDF and AP-SNDF) and Insoluble (AV-INDF and AP-INDF) Fractions {#sec2dot3-molecules-24-03605}\n-----------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAfter the digestion process insoluble non-digestible fraction of AV (AV-INDF), insoluble non-digestible fraction of AP (AP-INDF), soluble fraction of the non-digestible fibre of AV (AV-SNDF), soluble fraction of non-digestible fibre of AP (AP-SNDF), were obtained.\n\nThe fibre contents of the total non-digestible fibre fraction (TNDF) and its soluble (SNDF) and insoluble (INDF) fractions are shown in [Table 3](#molecules-24-03605-t003){ref-type=\"table\"}. The INDF content was approximately 25% higher in AP than in AV. The content obtained for AP was comparable to these of nopal cactus (57.7 g/100 g) \\[[@B19-molecules-24-03605]\\]. The SNDF proportion was 52% higher in AP than in AV, indicating that AP may have a higher amount of partially acetylated mannans as the primary polysaccharide of the gel, although it also contains pectins, some hemicelluloses, mucilages, and gums \\[[@B4-molecules-24-03605]\\]. Acetylated mannan molecules are mainly responsible for the thick, mucilage-like properties of raw *Aloe vera* gel \\[[@B4-molecules-24-03605],[@B6-molecules-24-03605]\\]. These results are interesting because *Aloe vera* polysaccharides have been considered the main component responsible for most of the beneficial properties attributed to the *Aloe vera* plant \\[[@B15-molecules-24-03605]\\]. They are dietary carbohydrates resistant to digestion in the upper GI tract. Due to the human physiological effect of these polysaccharides, they could be labelled as prebiotics, selectively fermented ingredients that allow for specific changes in the composition and/or activity in the gastrointestinal microflora that confer benefits to the hosts wellbeing and health \\[[@B20-molecules-24-03605]\\].\n\nThe high fermentability of SNDF produces SCFAs (acetic, propionic, and butyric acids) that contribute to the proper functioning of the large intestine and the prevention of pathologies through their actions in the lumen, colonic and vascular musculature as well as in the colonocyte metabolism \\[[@B21-molecules-24-03605]\\]. Furthermore, SCFAs reduce intestinal pH and increase water and salt absorption in the large intestine \\[[@B22-molecules-24-03605]\\]. On the other hand, the fibre contents in AV-INDF and AP-INDF were 26.57 g/100 g and 57.7 g/100 g, respectively. The high proportion of INDF, could increase the stool water content provides bulky/soft/easy-to-pass stools \\[[@B23-molecules-24-03605]\\].\n\n2.4. In Vitro Fermentation {#sec2dot4-molecules-24-03605}\n--------------------------\n\n### 2.4.1. Changes in pH {#sec2dot4dot1-molecules-24-03605}\n\nFermentation of TNDF in the large intestine is associated with reduced levels of pH and the consequent proliferation of beneficial microbes such as bifidobacteria \\[[@B22-molecules-24-03605]\\]. Therefore, changes in pH are often used as indicators of the fermentability of non-digestible matter. [Figure 2](#molecules-24-03605-f002){ref-type=\"fig\"}A shows that the pH decreased with reaction time in both samples: from 7.09 (h0) to 6.76 (h24) in AV and from 7.13 (h0) to 6.79 (h24) in AP. However, the change was gradual when compared to that of the lactulose used as a control. Lactulose has high fermentability and promotes a greater reduction in pH. A significant difference (*p* \\< 0.05) was observed in the pH of AV and AP with respect to the pH of the lactulose control during the reaction time and up to 24 h of incubation. This behaviour occurs because the control is a disaccharide of fast fermentation \\[[@B23-molecules-24-03605]\\]. In contrast, AV and AP are matrices mainly composed of non-digestible oligosaccharides or polysaccharides of slow fermentation that produce lower pH \\[[@B24-molecules-24-03605]\\] or buffer or antacid effects, as Al-Madboly et al. have reported \\[[@B25-molecules-24-03605]\\].\n\n### 2.4.2. Volume of gas produced {#sec2dot4dot2-molecules-24-03605}\n\nGases are produced by microbiota as by-products of their metabolic activities. These microorganisms of the gastrointestinal tract satisfy their energy requirements largely through non-digested carbohydrate fermentation and the subsequent production of SCFAs and certain gas species that include carbon dioxide (CO~2~), methane (CH~4~), and hydrogen (H~2~). The latter is the main by-product of hydrogen fermentation and is generated by hydrogen producers of the phyla Firmicutes and Proteobacteria \\[[@B26-molecules-24-03605]\\]. Hydrogen gas can be generated from pyruvate cleavage or reoxidation of the reduced pyridine and flavin nucleotides.\n\n[Figure 2](#molecules-24-03605-f002){ref-type=\"fig\"}B shows the volume of gas produced by the microbiota feeding on different components during in vitro fermentation. In AV, the increase in the volume of gas was slow, reaching over 12 mL at 24 h. The volumes indicate that, in the first four hours, fermentation is very slow since, in addition to fermentable fibres of AV, other compounds hamper the beginning of the process, such as insoluble fibres and polyphenols. On the other hand, the AP fraction immediately begins with active fermentation and progressively increases the volume of gas until reaching 15.8 mL (24 h). In this case, the available compounds are mostly soluble polysaccharides, which are rapidly fermentable when compared against the control. The AP extract clearly produced more gases from the beginning, and at 24 h, there was no significant difference with that produced by lactulose (17.7 mL). These results were in agreement with the fibre proportion since the total non-digestible fibre proportion (TNDF) in AP was approximately 50% more than in AV. These results indicate that glucooligosaccharides from AP could be the best carbon source for different microorganisms since they show specific preferences for defined substrates, as reported by Gullon et al. \\[[@B21-molecules-24-03605]\\]. Gas production was constant throughout the fermentation period, showing a significant difference (*p* \\< 0.05) against the control at 24 h of fermentation.\n\n### 2.4.3. Unfermented fibre (UF) {#sec2dot4dot3-molecules-24-03605}\n\nThe TNDF that passes to the large intestine is composed of soluble fibre that is mostly fermented and insoluble fibre, which is mostly not fermented by microorganisms. The latter are residual compounds or UF in the expelled faeces. [Figure 2](#molecules-24-03605-f002){ref-type=\"fig\"}C shows that the UF residues in AV and AP were scarcely fermented, as evidenced by 29.8% and 39.2% of the TNDF for AV and AP, respectively, results which match what was reported by Mudgil and Barak \\[[@B27-molecules-24-03605]\\].\n\n### 2.4.4. Quantification of SCFAs {#sec2dot4dot4-molecules-24-03605}\n\n[Table 4](#molecules-24-03605-t004){ref-type=\"table\"} shows the changes in the proportion of SCFAs in AV and AP during fermentation until 24 h, using faeces from healthy volunteers. The SCFAs were quantified as acetate, propionate, and butyrate and were compared with the SCFAs produced by lactulose fermentation, which works as a control given its high fermentability and its consideration as a fermentation standard \\[[@B21-molecules-24-03605]\\]. Acetate was the main SCFA quantified in both samples, and no significant difference (*p* \\< 0.05) was found with that produced by lactulose fermentation. This compound is the most common compound in the large intestine \\[[@B26-molecules-24-03605]\\] and enters the peripheral circulation through peripheral tissues to be metabolized. It has been reported that acetic acid is the main product of bifidobacterial fermentation in a human faecal environment \\[[@B28-molecules-24-03605]\\]. Acetate is produced by most anaerobes, including acetogens that are able to perform reductive acetogenesis from formate or hydrogen combined with CO~2~ \\[[@B29-molecules-24-03605]\\]. Acetate plays an important role in controlling inflammation and in combating pathogen invasion \\[[@B30-molecules-24-03605],[@B31-molecules-24-03605]\\].\n\nPropionate is another key SCFA in large intestine fermentation. At 8 h of fermentation, it was produced in similar amounts in the three substrates. However, at 24 h, AV and the control were the substrates that produced the most amount of propionate (3.68 and 4.44 \u03bcmol/mg, respectively), which was significantly higher (*p* \\< 0.05) than that produced by AP (3.12 \u03bcmol/mg) at 24 h of fermentation. Propionate, a gluconeogenerator, has been shown to inhibit cholesterol synthesis. On the other hand, after absorption, acetate has been shown to increase cholesterol synthesis \\[[@B26-molecules-24-03605]\\]; therefore, the higher propionate proportion obtained from AV fermentation could decrease the acetate: propionate ratio and reduce serum lipids and possibly cardiovascular disease risk. The proportion of acetate:propionate in the three samples at 24 h was AV = 2.90, AP = 3.50, and lactulose = 2.57. This indicates that AV and lactulose fermentation produce the highest concentration of propionate. In AV, the proportion was lower than that reported by Gull\u00f3n et al. \\[[@B32-molecules-24-03605]\\]. This difference could be due to seasonal changes and geographic locations that lead to significant variations in gel polysaccharides \\[[@B13-molecules-24-03605]\\]. Producers of propionate largely belong to the phylum Bacteroidetes but also include some Firmicutes.\n\nFinally, AV (1.92 \u03bcmol/mg) and AP (1.79 \u03bcmol/mg) butyrate yields at 24 h were lower than that of lactulose (4.25 \u03bcmol/mg). Butyrate formation occurs in certain Firmicutes bacteria, either via butyrate kinase (in many *Clostridium* and *Coprococcus* species) or via butyric acetate CoA transferase \\[[@B23-molecules-24-03605]\\]. Butyrate has been studied for its role in nourishing the colonic mucosa and in the prevention of cancer of the colon by promoting cell differentiation, cell-cycle arrest and apoptosis of transformed colonocytes, inhibiting the enzyme histone deacetylase and decreasing the transformation of primary bile acids to secondary bile acids as a result of colonic acidification \\[[@B26-molecules-24-03605]\\]. Butyric acid exerts a positive influence on the colonic mucosa and affects cell differentiation. It has anti-inflammatory properties and reduces the incidence of colon cancer \\[[@B29-molecules-24-03605]\\]. Butyrate irrigation (enema) has also been suggested in the treatment of colitis.\n\nTherefore, SCFA production and their potential delivery to the distal colon due to AV and AP fermentation may result in a protective effect. First, the human large intestine is colonized by dense microbial communities that utilize both diet-derived and host-derived energy sources for growth, predominantly through fermentative metabolism. This highly diverse community has the capacity to perform an extraordinary range of biochemical transformations that go well beyond those encoded by the host genome, and these activities exert an important influence upon many aspects of human health \\[[@B25-molecules-24-03605]\\].\n\n2.5. Quantification of Phenolic Compounds in AV and AP and Identification of AV, AV-TNDF, AP, AP-TNDF by HPTLC {#sec2dot5-molecules-24-03605}\n--------------------------------------------------------------------------------------------------------------\n\n### 2.5.1. Quantification of Phenolic Compounds in AV and AP {#sec2dot5dot1-molecules-24-03605}\n\n[Figure 3](#molecules-24-03605-f003){ref-type=\"fig\"}A shows the amount of phenolic compounds from AV and AP before and after undergoing an in vitro enzymatic process of digestion simulating the passing of these samples throughout the digestive tract. AV had a higher proportion of phenolic compounds than the extract (AP). In the same figure, a significant increase (*p* \\< 0.05) of 66% in the amount of phenolic compounds of AV-TNDF with respect to AV was observed, and a significant increase in the amount of phenolic compounds in AP-TNDF compared to AP was also observed, with an increase greater than 100%. However, it is important to note that even when the number of phenolic compounds of AP was lower compared to AV, the sample AP-TNDF had a proportion of phenolic compounds similar to that of AV-TNDF. This may be due to a major release of insoluble phenolic compounds with the acid, alkaline, and enzymatic digestion process, as Anokwuru, et al., have reported \\[[@B7-molecules-24-03605]\\]. This increase in phenolic compounds, also, could be the result of the opening and breakdown of polysaccharide chains and lattices, liberating most of the phenols when the AV and AP samples undergo the fermentation process \\[[@B29-molecules-24-03605]\\].\n\n[Figure 3](#molecules-24-03605-f003){ref-type=\"fig\"}B shows the change in the total phenolic compounds before and during fermentation. Fermented products of AV had a stable amount of phenolic compounds during the first 4 h (243.65 to 239.18 meq aloin/100 g), which was reduced at 24 h until 25.72 meq aloin/100 g. For AP-TNDF the fall was higher, starting with a concentration of 207.48 meq aloin/100 g at 0 h fermentation time to 4.58 meq aloin/100 g at 24 h. In humans, insoluble phenolic compounds are released from the matrix in the colon during the fermentation of the ingested material. The release of these phenolic compounds has been identified as beneficial against colon cancer \\[[@B33-molecules-24-03605]\\]. Free and conjugated phenolic compounds are known as extractable phenolic compounds while bound or insoluble phenolic compounds are known as non-extractable \\[[@B34-molecules-24-03605]\\]. It is observed that they disappear with fermentation and part of them are degraded, although they remain a constant and significant amount due to the release of insoluble phenolic compounds. There are studies of phenolic compounds that show how the microbiota populations and the phenolic compounds metabolized by this microbiota evolve and change. Cueva et al. (2012) \\[[@B35-molecules-24-03605]\\] observed a great degradation of phenolic compounds in grape seed extract during in vitro fermentation. Similar results to those obtained in this work.\n\n### 2.5.2. Identification of phenolic compounds by HPTLC {#sec2dot5dot2-molecules-24-03605}\n\nTaking into account the results obtained for total phenolic compounds before and after digestion of AV and AP, it was assessed the identification of phenolic compounds in AV, AV-TNDF, AP, and AP-TNDF by HPTLC ([Table 5](#molecules-24-03605-t005){ref-type=\"table\"}). The standards used were vanillic acid, *p*-hydroxybenzaldehyde, *p*-hydroxybenzoic acid, *p*-cresol and ferulic acid.\n\nIn the samples AV, AP, AV-TNDF and AP-TNDF, the presence of ferulic acid and an unidentified compound with Rf's of 0.05 and 0.04, respectively, was detected. However, with the fermentation process for AV, in the samples FAV4, FAV8 and FAV24, ferulic acid disappeared, and other components appeared with Rf\u00b4s of 0.64, 0.57, 0.12 and 0.04. For FPA4, FPA8, and FPA24, only two unidentified compounds with Rf\u00b4s of 0.12 and 0.04 appear in small quantities.\n\nUnidentified components can appear when fibre and phenolic compounds are fermented in the intestine. Lee, et al., \\[[@B36-molecules-24-03605]\\] reported that some bacterial phenolic metabolites were modified in colonic fermentation since specific intestinal bacteria metabolize phenolic compounds to different extents and produce different aromatic metabolites. Changes in intestinal phenolic levels may influence microflora composition. Phenolic compounds are likely to benefit the host by inhibiting pathogen growth and regulating commensal bacteria, including probiotics, and could therefore be considered as prebiotics.\n\n2.6. Antioxidant Activity of AV-TNDF and AP-TNDF {#sec2dot6-molecules-24-03605}\n------------------------------------------------\n\n### 2.6.1. Reducing Power {#sec2dot6dot1-molecules-24-03605}\n\nTo establish the appropriate concentration to determine the antioxidant activity of the AV-TNDF and AP-TNDF samples, several concentrations from 0.78 mg/mL to 12.5 mg/mL were evaluated. It was found that the samples had a concentration-dependent behaviour, with 12.5 mg/mL being the one with the highest activity (*p* \\< 0.05) (data not shown); therefore, the tests of reducing power and radical \u00b7OH sequestration correspond only to this concentration. [Figure 4](#molecules-24-03605-f004){ref-type=\"fig\"}A shows the changes in the reducing power during 24 h of the fermentation kinetics at 12.5 mg/mL of the fraction of non-digestible *Aloe vera* (AV-TNDF) and the non-digestible fraction of *Aloe vera* polysaccharide extract (AP-TNDF). The highest activity of reducing power of AV-TNDF was observed at 4 h (AVF4). This indicated that, in general, fractions derived from AV show a greater reducing power than fractions from AP. A possible explanation is that AV-TNDF had higher concentrations of inactive polyphenols in their matrices that can interact with the medium after gastrointestinal digestion. Acids and enzymes can open the fibres containing the polyphenols, which in turn are more bioavailable at the end of digestion. These results agree with the identification of the phenolic compounds shown in [Table 5](#molecules-24-03605-t005){ref-type=\"table\"}. The appearance of the unidentified compounds with Rf\u00b4s of 0.64, 0.57 and 0.12 may have greater reducing power than those identified in AP-TNDF.\n\nThe fermentation samples of AP-TNDF (AP0-24) had antioxidant capacities and reducing powers that experience limited change over time. As shown in [Table 5](#molecules-24-03605-t005){ref-type=\"table\"}, during fermentation, AP-TNDF lost unidentified phenolic compounds, which leads to a decrease in reducing power, and a correlation between the antioxidant capacity and the reducing power was seen, as Alugoju et al. have reported for certain bioactive compounds \\[[@B37-molecules-24-03605]\\]. The reducing power of 0.2 of inhibition was reported with methanolic extracts from non-digested fresh samples \\[[@B28-molecules-24-03605]\\]. Cueva et al. \\[[@B35-molecules-24-03605]\\] demonstrated the appearance of new phenolic compounds that were not present or were present in small amounts before fermentation such as hydroxyphenylacetic acid, phenylpropionic acid or phenylacetic acid with colonic in vitro fermentation of grape seed extract.\n\n### 2.6.2. OH Radical Sequestration {#sec2dot6dot2-molecules-24-03605}\n\nThe statistical analysis of \u00b7OH radical sequestration revealed that all extracts, except AP, were dose dependent. The elimination of hydroxyl radicals from fractionated polysaccharides and AV gel extracts has been reported by Chun-Hui et al. \\[[@B38-molecules-24-03605]\\]. [Figure 4](#molecules-24-03605-f004){ref-type=\"fig\"}B shows the efficacy of 12.5 mg/mL solutions of AV (35.47%), AV-TNDF (93.54%), AP (3.07%) and AP-TNDF (94.78%) to sequester hydroxyl radicals. The results of AV are even lower than assays reported by authors such as Ray et al. \\[[@B39-molecules-24-03605]\\], who report values of up to 48.01% using methanolic extracts of fresh *Aloe vera* gel without a digestion process. This difference may be due to the age of the plant and edaphoclimatic conditions as S\u00e1nchez-Machado et al., have reported \\[[@B40-molecules-24-03605]\\]. The higher efficacy of AV-TNDF and AP-TNDF could be due to the increase in phenolic compounds, because fresh samples have large concentrations of inactive polyphenols in their matrices that can interact with the medium after a digestion process such as gastro intestinal digestion. Additionally, it could be the result of the opening and breakdown of polysaccharide chains and lattices, liberating most of the phenols when the AV and AP samples undergo the fermentation process \\[[@B29-molecules-24-03605]\\].\n\n[Figure 4](#molecules-24-03605-f004){ref-type=\"fig\"}B shows the \u2219OH sequestration percentage of the synthesized or released compounds in solutions of fermented AV (AVF) and the fermented polysaccharide extract (APF) at 12.5 mg/mL. At this concentration, the ferments exhibited a decrease in antioxidant activity during the first 4 h of fermentation in the two samples up to 1.85% (AVF4) and 11.52% (APF4). The AVF antioxidant activity experienced a slight increase until 24 h of fermentation with 7.76% effectiveness observed for AVF24, and the APF antioxidant activity remained low and constant, reaching 5.74% effectiveness at 24 h.\n\nIn contrast to the reducing power, in this case, the percent of \u00b7OH sequestration drops dramatically over time within the first few hours of fermentation. In this case, the phenolic or other compounds that have effective \u00b7OH sequestration in the undigested fractions AV-TNDF and AP-TNDF are rapidly degraded by the enzymes, bacteria or fermentation conditions, drastically lowering the antioxidant activity. Lee et al., have reported a decrease in concentration after incubation with a faecal bacteria homogenate in compounds present in the tea extract, including epicatechin, catechin, 3-O-methyl gallic acid, gallic acid, and caffeic acid \\[[@B36-molecules-24-03605]\\]. The difference in the behavior of reducing power and \u00b7OH radicals are because they have different ways of sequestering Reactive Oxygen Species (ROS). \u00b7OH radicals react with biomolecules such as lipid, protein and DNA \\[[@B41-molecules-24-03605]\\].\n\n3. Materials and Methods {#sec3-molecules-24-03605}\n========================\n\n3.1. Raw Material {#sec3dot1-molecules-24-03605}\n-----------------\n\nLeaves of *Aloe vera* plants (aged 3--4 years) were obtained from a market in Mexico City. The leaves were washed with soap and water, followed by disinfection with ethanol. The gel was extracted by removing the cortex and was frozen and lyophilized until use.\n\n3.2. Physicochemical Characterization of Aloe Vera (AV) {#sec3dot2-molecules-24-03605}\n-------------------------------------------------------\n\nIn the raw gel of *Aloe vera*, the acidity was determined using AOAC method 939.05 (2015) \\[[@B42-molecules-24-03605]\\], while the pH was determined using a potentiometer (pH-120; Conductronic, Puebla, Pue, Mexico). Soluble solids (\u00baBrix) were assessed using a digital refractometer (HSR-500; Atago, Kobe, Hyogo-ken, Japan) at a 0--32 \u00baBrix scale considering that 1 \u00baBrix = 1 g soluble solid in 100 g of solution \\[[@B43-molecules-24-03605]\\]. Moisture was determined according to method 2005.02, raw protein by the Kjeldahl method (2005.06), raw fat by a Soxhlet method (920.39c), and ash by method 923.03 (AOAC, 2015) \\[[@B42-molecules-24-03605]\\]. Neutral detergent fibre was assessed in the lyophilized sample of the gel, according to the 2002:04/ISO technique (16472:2005) from AOAC \\[[@B42-molecules-24-03605]\\]. Total dietary fibre was determined through the enzymatic-gravimetric method (32--05) outlined by AOAC (2015) \\[[@B42-molecules-24-03605]\\].\n\n3.3. Identification of phenolic Compounds by Ultra-Performance Liquid Chromatography-Tandem Mass Spectrometry (UPLC-MS) in Aloe Vera (AV) {#sec3dot3-molecules-24-03605}\n-----------------------------------------------------------------------------------------------------------------------------------------\n\nThe identification of phenolic compounds in the gel of *Aloe vera* (AV) was performed with ultra-performance liquid chromatography (UPLC) by utilizing an Acquity UPLC H-Class (Waters Corp, Milford, MA, USA) coupled with a Waters Xevo G2-X2 Tof system with an electrospray ionization (ESI) interface. The conditions for the chromatographic analysis were as follows: the mobile phase consisted of 0.1% formic acid in water/acetonitrile at a flow rate of 0.3 mL/min and 35 \u00b0C. A BEH C18 Column (2.1 mm X 100 mm) was used and the chromatograms were registered through detection of the Total Ion Current. The MS conditions of analysis in negative-ion mode were as follows: drying gas (nitrogen) flow rate, 8 L/min; gas temperature, 180 \u00b0C; scan range, *m/z* 50--3000; capillary voltage, 4500 V; and nebulizer pressure, 2.5 bar \\[[@B44-molecules-24-03605]\\]. A solution of 8 mg/mL of lyophilized AV was prepared, homogenized, and passed through a 0.22 \u00b5m cellulose filter prior to analysis.\n\n3.4. Extraction of Polysaccharides (AP) from Aloe Vera Gel {#sec3dot4-molecules-24-03605}\n----------------------------------------------------------\n\nThe extraction of *Aloe vera* gel polysaccharides (AP) followed a modification of the technique reported by Ni et al. \\[[@B4-molecules-24-03605]\\]. One gram of lyophilized *Aloe vera* gel was suspended in 80 mL of a 95% water/methanol solution, homogenized with a Polytron, and centrifuged at 936\u00d7 *g* (Allegra 64R; Beckman, Pasadena, CA, USA) for 10 min. The first precipitate was stored, and the supernatant was centrifuged one more time at 10,285\u00d7 *g* for 15 min. The first and second precipitates were combined, re-suspended in deionized water and lyophilized. The sum of these two fractions was the polysaccharide extract (AP). The first precipitate corresponds to the cell wall fibre, and the second precipitate corresponds to micro-particles from cell organelles of *Aloe vera* \\[[@B4-molecules-24-03605]\\].\n\n3.5. Indigestible Fraction and in Vitro Fermentation {#sec3dot5-molecules-24-03605}\n----------------------------------------------------\n\n### 3.5.1. Indigestible Fermentation of Aloe Vera Gel (AV) and Polysaccharides (AP) {#sec3dot5dot1-molecules-24-03605}\n\n[Figure 5](#molecules-24-03605-f005){ref-type=\"fig\"} shows the flowchart of the isolation of the indigestible fibre fractions of AV and AP The indigestible fraction (IF) of (AV) and (AP) ([Figure 5](#molecules-24-03605-f005){ref-type=\"fig\"}) was determined following the method proposed by Saura-Calixto et al. \\[[@B45-molecules-24-03605]\\] with some modifications. Briefly, 300 mg of both fractions were suspended in HCl-KCl buffer (pH 1.5). Then, 0.2 mL of a pepsin solution at a concentration of 300 mg/mL in HCl-KCl buffer (\\>250 units/mg, P7000; Sigma Aldrich, Toluca, Edo. Mex. Mexico) was added and incubated at 40 \u00b0C for 1 h with constant stirring to simulate the digestive process.\n\nTo simulate small intestine digestion, pH was adjusted to 6.9, and 9 mL of tris-maleate buffer (0.1 M) was added, followed by 1 mL of \u03b1-amylase solution (3480 units/mL, A3176; Sigma Aldrich), at a concentration of 120 mg/mL in tris-maleate buffer. The mix was incubated at 37 \u00b0C for 16 h at constant stirring. Lipid digestion enzymes such as lipase or biliary juices were omitted given that the amount of fat (lower than 4%) was not significant. After the reaction time, the samples were centrifuged at 4000\u00d7 *g* for 15 min, and the supernatants were saved. The precipitate was washed twice with 10 mL of distilled water and centrifuged; supernatants were collected in 80 mL beakers. The tubes containing the solid residue were placed in an oven at 60 \u00b0C for 48 h, and the residue was then weighed as the amount of insoluble non-digestible fibre in AV (AV-INDF) and AP (AP-INDF). Both insoluble non-digestible fibre fractions were stored at 4\u00b0C until the fermentation experiments.\n\nThe supernatant was placed on dialysis membranes previously treated with boiling water for 20 min. The supernatant was dialyzed (cellulose membrane 12000--14000 Dalton MWCO-18) at a constant water flow of 7 L/h (30 mL/15 sec) at 25 \u00b0C for 48 h. Finally, the content of the membranes was lyophilized and weighed to obtain the soluble fraction of the non-digestible fibre of AV (AV-SNDF) and non-digestible fibre of AP (AP-SNDF); these fractions were stored at 5 \u00b0C until use in the fermentation process \\[[@B46-molecules-24-03605]\\].\n\n### 3.5.2. In Vitro Fermentation of AV-TNDF and AP-TNDF {#sec3dot5dot2-molecules-24-03605}\n\nThe samples obtained in the digestion, which were the soluble fraction (SNDF) and the insoluble fraction (INDF), were combined as the total non-digestible fraction (TNDF) of AV and AP to be fermented. In vitro fermentation was assessed in the AV-TNDF and AP-TNDF fractions. Fermentation was prepared according to a modified version of Mart\u00edn-Carr\u00f3n and Go\u00f1i \\[[@B46-molecules-24-03605]\\]. The samples were placed in a culture system under strict anaerobic conditions for 24 h. Human faeces were used as inocula, and anaerobic conditions were maintained using oxygen-free carbon dioxide.\n\n#### Inoculum\n\nTo obtain the 10% *w/v* inoculum, the faeces of four volunteers were weighed and placed in beakers containing a sterile and anaerobic fermentation medium. The inocula were mixed in a Stomacher 80 Lab blender (Seward Medical, London, UK for 10 min and then filtered (1-mm mesh) before use.\n\n#### Medium\n\nA micromineral solution was prepared using CaCl~2~-2H~2~O, MnCl~2~-4H~2~O, CoCl~2~-6H~2~O, and FeCl~3~-6H~2~O mixed in a buffer solution ((NH~4~)HCO~3~ and NaHCO~3~) in distilled water. A macromineral solution containing Na~2~HPO~4~, KH~2~PO~4~, and MgSO~4~-7H~2~O was made in distilled water, as was a reducing solution containing cysteine hydrochloride (C-1276, Sigma-Aldrich, Toluca, Edo. Mex., Mexico).\n\n#### Procedure\n\nSubstrate (100 mg) of the total non-digestible fraction (AV-TNDF or AP-TNDF) was weighed in a 60 mL serum vial (Supelco; Bellefonte, PA, USA), and 8 mL of fresh medium and 2 mL of inoculum were added. Vials were sealed with rubber stoppers and placed in a water bath with constant stirring at 37 \u00b0C. The fermentation kinetics was monitored at 0, 4, 8 and 24 h to obtain fermented samples at each time point (AVF4, AVF8 and AVF24 or APF4, APF8 and APF24). A control for each time containing lactulose (L-7877, Sigma-Aldrich, Toluca, Edo. Mex., Mexico) and a target without substrate were included as controls.\n\nGas volume production and pH were measured at 0, 4, 8 and 24 h. Fermentation was stopped by adding 2.5 mL of1 M NaOH. Samples were centrifuged at 2500\u00d7 *g* for 10 min, and 3 mL of supernatant was taken in duplicate for SCFA determination by gas chromatography.\n\n#### Unfermented Fibre\n\nThe precipitate of each tube from the fermentation kinetics experiment was homogenized in 50 mL of 0.9% NaCl for 3 min. All precipitates were filtered using Dacron filters (pore size 50 \u03bcm), which were dried to a constant weight before use. Unfiltered residues were washed twice with 50 mL of 0.9% NaCl and washed twice with 5 mL of acetone to eliminate mineral and inorganic residues in addition to lipids and other hydrophobic compounds. Filter papers were dried at 60 \u00b0C to constant weight, assessing the unfermented fibre (UF) \\[[@B47-molecules-24-03605]\\] that was calculated by the following formula:\n\nThen, UF(AV-TNDF) and UF(AP-TNDF) were obtained.\n\n### 3.5.3. Quantification of SCFAs in AV-TNDF and AP-TNDF {#sec3dot5dot3-molecules-24-03605}\n\nSCFA quantification was carried out in AV-TNDF, AP-TNDF and lactulose fermented according the methodology reported by Saura-Calixto et al., \\[[@B45-molecules-24-03605]\\] and adapted from Zhao et al., \\[[@B48-molecules-24-03605]\\]. Briefly, 500 \u00b5L of supernatant of each fermentation time sample was mixed with 400 mL of 2-methylvaleric acid as an internal standard (10987-8; Sigma, Toluca, Edo. Mex., Mexico) and 100 mL of HClO~4~ to maintain a constant pH in the samples. The mixture was centrifuged at 10,000 at 4 \u00b0C for 15 min, and the supernatant was placed in gas chromatography vials. A Clarus 500/580 GC gas chromatographer (Perkin-Elmer, Inc. Shelton, CO., USA) was used with a TG-WAXMS-A GC column (Thermo Fisher Scientific, CDMX, Mexico) and a flame ionization detector with an injector temperature of 270 \u00b0C and a detector temperature of 300 \u00b0C. Glacial acetic acid, propionic acid, butyric acid, and 4-methylvaleric acid (Sigma, 320099, 94425, 19215 and 10987-8) were used as standards to obtain calibration curves and assess retention times. The oven was heated to 95 \u00b0C at 2 min and increased at 120 \u00b0C/min to 240 \u00b0C; the carrier flow was 1 mL/min, and the carrier gas flow was 20 mL/min.\n\n### 3.5.4. Quantification of Phenolic Compounds {#sec3dot5dot4-molecules-24-03605}\n\nTotal phenolic compounds were quantified using the Folin--Ciocalteu method \\[[@B49-molecules-24-03605]\\] in the AV, AP, AV-TNDF, AP-TNDF, AVF4, AVF8, AVF24, APF4, APF8 and APF24 samples. The samples were homogenized in 0.5 mL methanol, added Folin-Ciocalteu reagent, and saturated sodium carbonate solution and water. After 60 min, the absorbance was measured at 760 nm. A calibration curve was performed with aloin, because it is the most representative phenolic compound in *Aloe vera* \\[[@B50-molecules-24-03605]\\]. The concentration of total phenolic compounds was expressed as milligram equivalents of aloin/100 g sample.\n\n### 3.5.5. Identification of Phenolic Compounds by HPTLC {#sec3dot5dot5-molecules-24-03605}\n\nHPTLC was accomplished according to a modified version of the methodology published by Paillat, et al., \\[[@B51-molecules-24-03605]\\]. At the beginning volumes of standard solutions of ferulic acid, *p*-hydroxybenzaldehyde (PHB), *p*-hydroxybenzoic acid (APHB), *p*-cresol, *p*-creosol, and vanillic acid (722820, 54590, 54630, 61030, 41340, and 68654, Sigma-Aldrich, Qu\u00edmica S.A. de C.V.) were applied at a concentration of 3 mg/mL. After that, were applied 15 \u03bcL of AV, AV-TNDF, AP, AP-TNDF and the kinetic fermentation samples (AVF4, AVF8, AVF24 and APF4, APF8, APF24) at a concentration of 12.5 mg/mL. Samples were placed on TLC silica gel 60 F254 plates (E. Merck, Darmstadt, Germany) using an ATS 4 TLC sampler (CAMAG, Muttenz, Switzerland) at a constant application rate of 120 nL s^\u22121^ and developed in a CAMAG automated developing chamber ADC2 (47% moisture) saturated and preconditioned for 5 min to a 50-mm distance with an *n*-hexane: chloroform:methanol:acetic acid solvent system (5:36:4:0.5 *v/v/v/v*). Plates were scanned at 254 nm and 366 nm in a CAMAG TLC III scanner (slit size 4 mm \u00d7 0.3 mm) at a scanning speed of 10 mm s^\u22121^ and a data step resolution of 50 \u03bcm.\n\n3.6. Antioxidant Activity {#sec3dot6-molecules-24-03605}\n-------------------------\n\n### 3.6.1. Reducing Power and Hydroxyl Radical Scavenging Activity {#sec3dot6dot1-molecules-24-03605}\n\nThe reducing power of AV, AV-TNDF, AP, and AP-TNDF was assessed by the Oyaizu method \\[[@B52-molecules-24-03605]\\]. Various concentrations of methanolic extracts (12.5, 6.25, 3.12, 1.56, and 0.78 mg/mL) were mixed with 2.5 mL 200 mmol/L sodium phosphate buffer (pH 6.6) and 2.5 mL 1% potassium ferricyanide. The mixture was incubated at 50 \u00b0C for 20 min. After 2.5 mL of 10% trichloroacetic acid (*w/v*) were added, the mixture was centrifuged at 650 rpm for 10 min. The upper layer (5 mL) was mixed with 5 mL deionised water and 1 mL 0.1% of ferric chloride; the absorbance was measured at 700 nm, a higher absorbance indicating a higher reducing power. The assays were carried out in triplicate and the results were expressed as mean values \u00b1 standard deviations. Finally, the reducing power in the kinetic fermentation samples (AVF4, AVF8, AVF24, APF4, APF8 and APF24) was determined but only with a concentration of 12.5 mg/mL. This was the concentration that gave the best results in the previous samples and was chosen for this second part of the work.\n\n### 3.6.2. \u2219OH Radical Sequestration {#sec3dot6dot2-molecules-24-03605}\n\nTo determine the scavenging activity of the hydroxyl radical (\u2219OH), the method described by Li et al. \\[[@B53-molecules-24-03605]\\] was applied. The antioxidant activity of AV, AV-TNDF, AP and AP-TNDF was assessed at 12.5, 6.25, 3.12, 1.56 and 0.78 mg/mL. Both 1,10-phenanthroline (0.75 mM) and FeSO4 (0.75 mM) were dissolved in phosphate buffer (pH 7.4) and mixed thoroughly. To start the reaction, H~2~O~2~ (0.01%) and samples were added. The mixture was incubated at 37 \u00b0C for 60 min and the absorbance was measured at 536 nm. Finally, the scavenging activity of the hydroxyl radical in the kinetic fermentation samples (AVF4, AVF8, AVF24, APF4, APF8 and APF24) was determined but only with a concentration of 12.5 mg/mL his was the concentration that gave the best result in the previous samples and was chosen for this second part of the work.\n\n3.7. Statistical Analysis {#sec3dot7-molecules-24-03605}\n-------------------------\n\nDifferences between experimental groups were analysed by one-way ANOVA and Tukey'post hoc test for repeated samples. To compare the maximum and minimum values of the total SCFA concentration, the data were analysed by a paired one-tailed Student's t-test. Data were processed using GraphPad Prism 6.0 (GraphPad Software Inc., San Diego, CA, USA). Values of *p* \u2264 0.05 were considered statistically significant.\n\n4. Conclusions {#sec4-molecules-24-03605}\n==============\n\n*Aloe vera* gel (AV) and polysaccharide extract of *Aloe vera* (AP) are a matrix mainly composed of non-digestible oligosaccharides or polysaccharides of slow fermentation, since that produces a lower pH. The behavior of AV and AP during in vitro colon fermentation was similar to that of lactulose, what indicates the possibility of using *Aloe vera* and polysaccharide extracts as prebiotics. The SCFA production and their potentially delivery to the distal colon due to AV and AP digestion and fermentation process, may result in a protective effect, firstly of human large intestine, since this could be colonised by dense microbial communities that utilized both diet-derived and host-derived energy sources for growth predominantly through fermentative metabolism. In the same way, the antioxidant activity also increases significantly in both the reducing power test and the \u00b7OH radical sequestration when going from AV and AP to its indigestible fraction AV-TNDF and AP-TNDF. Finally, we can conclude that there were no significant differences during the digestion and fermentation of *Aloe vera* and its extract despite the fact that the content of dietary fibre in AP was significantly higher than that of *Aloe vera* gel. During in vitro colon fermentation, the unfermented fibre of AV and AP had a similar response to that of lactulose, as well as the total volume of gas produced, which indicates that *Aloe vera* and polysaccharide extract can possibly be used as prebiotics.\n\n**Sample Availability:** Samples of the compounds are not available from the authors.\n\nConceptualization, A.T.-M., R.M.-E. and M.E.J.-F.; Methodology, A.T.-M.; P.O.-D., S.V.\u00c1.-R.; G.M.A.-J.; Validation, A.T.-M. and R.C.-O.; Formal analysis, A.T.-M.; Investigation, A.T.-M.; Resources, R.M.-E.; writing A.T.-M., R.M.-E., Supervision, R.M.-E., M.E.J.-F. and P.O.-D.; Funding acquisition, R.M.-E. All authors accepted the final version of the manuscript.\n\nThis research was supported by grants from the Instituto Polit\u00e9cnico Nacional (SIP:20181718 project) and CONACyT (PY-SEP-CONACyT 242860, project), The work was also supported by a PhD scholarship to A.T.-M., from CONACyT.\n\nThe authors declare no conflicts of interest.\n\n![*Aloe vera* plant and its main components.](molecules-24-03605-g001){#molecules-24-03605-f001}\n\n![Fermentation kinetics of *Aloe vera* gel (AV) and its polysaccharide extract (AP) during 24 h. (**A**) pH changes during fermentation, (**B**) Volume of gas produced during fermentation, and (**C**) Unfermented residue of samples AV and AP.](molecules-24-03605-g002){#molecules-24-03605-f002}\n\n![**Total phenolic compounds.** (**A**) Total phenolic compounds before and after digestion of *Aloe vera* gel (AV) and its polysaccharide extract (AP). Where AV and AP correspond to samples before digestion; AV-TNDF and AP-TNDF correspond to samples after digestion. (**B**) Changes in the total content of phenolic compounds during AV-TNDF and AP-TNDF fermentation.](molecules-24-03605-g003){#molecules-24-03605-f003}\n\n![Changes in antioxidant activity during colonic fermentation of *Aloe vera* gel (AV) and its polysaccharide extract (AP). (**A**) Changes in reducing power (Abs 700 nm), due to colonic fermentation. (**B**) Changes in % of sequestration of hydroxyl radical \u00b7OH due to colonic fermentation. (**B**) The results are expressed as the ME \u00b1 DS, n = 3.](molecules-24-03605-g004){#molecules-24-03605-f004}\n\n![Indigestible fraction. The obtaining of indigestible fractions of AV and AP. AV= *Aloe vera gel, AP= Aloe* polysaccharides; AV-NDF = Insoluble non-digestible fraction of AV, AP-NDF = Insoluble non digestible fraction of AP, AV-SNDF = Soluble fraction of the non-digestible fibre of AV, AP-SNDF = Soluble fraction of non-digestible fibre of AP.](molecules-24-03605-g005){#molecules-24-03605-f005}\n\nmolecules-24-03605-t001_Table 1\n\n###### \n\nChemical composition of *Aloe vera* ^1^.\n\n  COMPONENT                       AMOUNT (g/100g db)\n  ------------------------------- --------------------\n  Moisture                        97.50 \u00b1 0.04\n  Proteins                        2.15 \u00b1 0.08\n  Lipids                          3.27 \u00b1 0.85\n  Ashes                           12.62 \u00b1 1.1\n  Total Dietary Fibre (TDF)       55.11 \u00b1 0.55\n  Neutral Detergent Fibre (NDF)   15.58 \u00b1 2.04\n  Nitrogen-free extract           26.85 \u00b1 1.1\n\n^1^ Determinations were made in triplicate and the results are expressed in dry basis as the mean \u00b1 SD.\n\nmolecules-24-03605-t002_Table 2\n\n###### \n\nPhenolic compounds identified in the aqueous extract of *Aloe vera* by UPLC-MS.\n\n  PEAK    Rt ^a^ (min)   *m/z* \\[M-H\\]^\u2212^   *m* \\[M\\]   Tentative Identification\n  ------- -------------- ------------------ ----------- --------------------------------\n  **1**   1.52           393.18             394.18      Aloesin\n  **2**   3.15           433.19             434.19      10-Hydroxyaloin\n  **3**   3.15           561.36             562.36      Aloinoside\n  **4**   4.45           417.19             418.19      Aloin A\n  **5**   4.50           555.25             556.25      Aloenin-2\u2032-*p*-coumaroyl ester\n  **6**   4.68           417.19             418.19      Aloin B\n\n^a^ Retention time.\n\nmolecules-24-03605-t003_Table 3\n\n###### \n\nNon-digestible fraction of *Aloe vera* gel (AV) and its polysaccharide extract (AP).\n\n  Sample   SNDF (g/100g)      INDF (g/100g)      TNDF (g/100g)\n  -------- ------------------ ------------------ ---------------\n  AV       20.57 ^a^ \u00b1 0.83   26.57 ^a^ \u00b1 0.02   47.14\n  AP       39.34 ^b^ \u00b1 0.55   53.05 ^b^ \u00b1 0.40   92.39\n\nSNDF = Soluble non digestible fraction; INDF = Insoluble non digestible fraction; TNDF = Total non-digestible fraction. The results are the average of three repetitions \u00b1 SD. Means between rows with different letters (a and b) are significantly different for a significance level of (*p* \u2264 0.05) by t-student test.\n\nmolecules-24-03605-t004_Table 4\n\n###### \n\nConcentration of SCFA (\u03bcmol/mg) produce during the fermentation of *Aloe vera* gel (AV), its polysaccharide extract (AP) and lactulose as a carbon source\n\n  Fermentation Time   Acetic Acid       Propionic Acid     Butyric Acid       Total SCFA                                                                                                                                         \n  ------------------- ----------------- ------------------ ------------------ ------------------ ----------------- ----------------- ----------------- ----------------- ----------------- ------------------ ------------------ ------------------\n  0                   2.16 \u00b1 1.07 ^b^   0.93 \u00b1 0.36 ^a^    2.16 \u00b1 0.08 ^b^    0.09 \u00b1 0.05 ^a^    0.14 \u00b1 0.03 ^b^   0.09 \u00b1 0.03 ^a^   0.09 \u00b1 0.04 ^b^   0.14 \u00b1 0.01 ^c^   0.05 \u00b1 0.01 ^a^   2.33 \u00b1 1.08 ^ab^   1.21 \u00b1 0.39 ^a^    2.34 \u00b1 0.11 ^b^\n  4                   3.15 \u00b1 0.55 ^a^   3.20 \u00b1 0.64 ^a^    4.50 \u00b1 0.73 ^a^    1.32 \u00b1 0.26 ^a^    1.30 \u00b1 0.14 ^a^   1.07 \u00b1 0.06 ^a^   0.31 \u00b1 0.05 ^a^   0.63 \u00b1 0.04 ^c^   0.47 \u00b1 0.02 ^b^   4.77 \u00b1 0.84 ^a^    5.13 \u00b1 0.80 ^a^    5.15 \u00b1 0.93 ^a^\n  8                   6.12 \u00b1 0.66 ^a^   7.00 \u00b1 0.26 ^a^    9.75 \u00b1 1.92 ^b^    2.50 \u00b1 0.27 ^ab^   2.16 \u00b1 0.32 ^a^   3.42 \u00b1 0.90 ^b^   0.79 \u00b1 0.04 ^a^   1.02 \u00b1 0.13 ^b^   2.09 \u00b1 0.03 ^c^   9.41\u00b10.96 ^a^      10.18 \u00b1 0.70 ^a^   12.51 \u00b1 2.79 ^a^\n  24                  10.67 \u00b1 0.3 ^a^   10.93 \u00b1 1.33 ^a^   11.43 \u00b1 1.57 ^a^   3.68 \u00b1 0.24 ^b^    3.12 \u00b1 0.14 ^a^   4.44 \u00b1 0.69 ^b^   1.92 \u00b1 0.10 ^a^   1.79 \u00b1 0.14 ^a^   4.25 \u00b1 1.12 ^b^   16.27 \u00b1 0.60 ^a^   15.83 \u00b1 1.47 ^a^   17.03 \u00b1 1.51 ^a^\n\nStandard deviation for n = 3. Different letters (a and b) indicate significant differences (*p* \\< 0.05) between columns for the same acid.\n\nmolecules-24-03605-t005_Table 5\n\n###### \n\nIdentification of phenolic compounds of AV, AV-TNDF, AP, AP-TNDF and their fermented fractions obtained by HPTLC.\n\n  Phenolic Compounds        Colour       Rf     AV   AV-TNDF   AP   AP-TNDF   FAV4   FAV8   FAV24   FPA4   FPA8   FPA24\n  ------------------------- ------------ ------ ---- --------- ---- --------- ------ ------ ------- ------ ------ -------\n  Vanillinic acid           Dark blue    0.54   \\-   \\-        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  *p*-Hydroxybenzaldehyde   Dark blue    0.65   \\-   \\-        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  *p*-Hydroxybenzoic acid   Dark blue    0.45   \\-   \\-        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  *p*-Cresol                Dark blue    0.72   \\-   \\-        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  *p*-Creosol               Dark blue    0.86   \\-   \\-        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  Ferulic acid              Light blue   0.55   \\+   \\+        \\+   \\+        \\-     \\-     \\-      \\-     \\-     \\-\n  Unidentified              Violet       0.64   \\-   \\-        \\-   \\-        ++     ++     ++      \\-     \\-     \\-\n  Unidentified              Yellow       0.57   \\-   \\-        \\-   \\-        ++     ++     ++      \\-     \\-     \\-\n  Unidentified              Yellow       0.12   \\-   \\-        \\-   \\-        ++     ++     ++      \\+     \\+     \\+\n  Unidentified              Red          0.06   ++   \\+        \\-   \\-        \\-     \\-     \\-      \\-     \\-     \\-\n  Unidentified              Light blue   0.04   \\+   \\+        ++   +++       \\+     \\+     \\+      \\+     \\+     \\+\n\nAV = *Aloe vera*, AV-TNDF = non-digestible fraction of AV, AP= AV polysaccharides, AP-TNDF= non-digestible fraction of AP and their fermented fractions obtained by HPTLC. Samples obtained at 4, 8 and 24 (FAV4, FAV8, FAV24, FPA4, FPA8, FPA24). The symbols indicate: +++ highly concentrated compound, ++ compound concentrate, + little concentrated compound, - no appreciable concentration of the compound. HPTLC, remission at 366 nm. CAMAG.\n"}
+{"text": "1.. RAGE---A Molecule of the Immunoglobulin Superfamily and Links to Oxidative Stress\n=====================================================================================\n\nRAGE was first identified on account of its ability to bind advanced glycation endproducts (AGEs), the products of nonenzymatic glycation and oxidation of proteins and lipids \\[[@b1-ijms-14-19891]\\]. RAGE is expressed in a wide array of cell types, such as vascular cells, inflammatory cells, glomerular epithelial cells (podocytes), neurons of the central and peripheral nervous systems, cardiomyocytes, retinal Muller ganglion cells and alveolar pneumocytes \\[[@b2-ijms-14-19891]\\]. In homeostasis, RAGE is usually expressed at low levels in adult, non-diseased tissues; in settings such as cardiovascular disease, diabetes, inflammation, tumors and neurodegeneration, expression of RAGE is higher than that observed in control, non-diseased animal models or human subjects \\[[@b3-ijms-14-19891]\\]. The key to the biology of RAGE is its ability to bind to and transduce the molecular effects of multiple ligands. In addition to AGEs, RAGE transduces the signals of distinct ligands, such as multiple members of the S100/calgranulin family \\[[@b4-ijms-14-19891]\\], high mobility group box 1 (HMGB1) \\[[@b5-ijms-14-19891]\\], amyloid beta-peptide and beta-sheet fibrils \\[[@b6-ijms-14-19891]\\], lysophosphatidic acid (LPA) \\[[@b7-ijms-14-19891]\\], Mac1 \\[[@b8-ijms-14-19891]\\], and C1q \\[[@b9-ijms-14-19891]\\].\n\nThe biological impact of RAGE is diverse and depends on the cell type, acuteness *vs*. chronicity of stimulation and interactions with surrounding cell types *in vivo*, such those released in an autocrine and/or paracrine manner which may modulate RAGE action. We propose that RAGE ligands may be sequentially induced and released depending on the nature of the cellular stress. For example, our observation that endothelial cells release AGEs within minutes of exposure to hypoxia suggests that RAGE ligands may amplify and perpetuate acute and chronic stresses within the tissues \\[[@b10-ijms-14-19891]\\]. We and others have shown that RAGE activates a diverse array of signaling cascades with significant and wide-ranging impact on gene expression profiles characterized by inflammation, fibrosis and cell fate, as examples.\n\nIn the context of natural aging and diabetes, it is speculated that the development, perpetuation and unremitting nature of reactive oxidant species (ROS) generation directly contributes to the complications of these two key states \\[[@b11-ijms-14-19891],[@b12-ijms-14-19891]\\]. Compelling evidence highlights pathogenic roles for RAGE in both diabetes and aging. In this review, we will focus on the evidence linking ligand-RAGE interaction to oxidative stress and its consequences.\n\n2.. RAGE and Oxidative Stress---First Evidence\n==============================================\n\nWe first described roles for AGE-RAGE interaction in endothelial cells. AGEs, either those prepared *in vitro* or via immunoisolation from diabetic plasma, resulted in generation of thiobarbituric acid reactive substances (TBARS) and activation of oxidant stress-sensitive NF-\u03baB. Roles for RAGE and oxidative stress in the biological impact of AGEs were evident by blockade of these effects by pre-treatment of the endothelial cells with blocking antibodies to RAGE or anti-oxidants \\[[@b13-ijms-14-19891]\\]. *In vivo* studies demonstrated that infusion of AGEs into animals resulted in increased oxidative stress as evidenced by increased levels of malondialdehyde and TBARS in the vasculature and tissues, activation of NF-\u03baB and upregulation of heme oxygenase 1 mRNA. Pre-treatment of the animals with anti-RAGE antibodies or antioxidants blocked these effects of infused AGEs \\[[@b13-ijms-14-19891]\\]. A key question from this work, however, was by what means AGEs generated oxidative stress via RAGE.\n\nSubsequent studies detailed key roles for AGEs in generation of oxidative stress via activation of NADPH oxidase. Specifically, Wautier and colleagues showed that incubation of AGEs on the surface of diabetic red blood cells or the specific AGE, carboxy methyl lysine (CML) AGE with endothelial cells resulted in generation of hydrogen peroxide, upregulation of vascular cell adhesion molecule 1 (VCAM-1) and generation of tissue factor. Treatment of the endothelial cells with diphenyliodonium (DPI), but not by inhibitors of nitric oxide, blocked these effects of AGEs. The key role of NADPH oxidase was shown in these effects of AGEs was shown for the first time by the failure of macrophages retrieved from mice deficient in gp91phox but not wild-type (WT) mice to upregulate tissue factor upon incubation with AGEs \\[[@b14-ijms-14-19891]\\].\n\nFurther experiments suggested that AGE-dependent activation of oxidative stress via RAGE in renal mesangial cells facilitated the generation of mitochondrial superoxide. Infusion of AGEs into non-diabetic rodents resulted in increased cytosolic ROS, followed subsequently by increase mitochondrial permeability transition and deficiency of mitochondrial complex I. These effects were even further pronounced in diabetic rodents \\[[@b15-ijms-14-19891]\\]. Supportive of the roles of cytosolic ROS generated from AGE-RAGE interaction in the induction of mitochondrial oxidative stress were the findings that reduction of cytosolic ROS generation (using apocynin) or lowering AGEs (using alagebrium, an AGE cross-link breaker agent) reduced mitochondrial superoxide generation. Further, in mice devoid of RAGE, diabetes-induced increases in mitochondrial superoxide in the renal cortex were reduced \\[[@b15-ijms-14-19891]\\]. These data suggest that AGE-RAGE interaction may initiate generation of ROS, largely via cytosolic NADPH oxidase-dependent mechanisms, but that in certain conditions and cell types, AGE-RAGE may contribute to downstream amplification of oxidative stress via the induction of mitochondrial superoxide.\n\nThese studies led therefore to the question of the extent to which RAGE ligands activated oxidative stress in cell types other than endothelial cells, macrophages or renal mesangial cells. In the section to follow, we detail the evidence that RAGE ligands activate oxidative stress in a broad array of cell types, thereby implicating this receptor pathway in fundamental mechanisms of oxidative stress and tissue-specific consequences.\n\n3.. RAGE and Oxidative Stress: Evidence in Multiple Cell Types\n==============================================================\n\nRAGE is expressed on diverse cell types. A common mechanism by which the ligands of RAGE may modulate the biological properties of these cells is via oxidative stress. As discussed above, RAGE ligand action in endothelial cells contributes to the generation of oxidative stress. In the sections to follow, we illustrate examples of RAGE-dependent oxidative stress in cardiovascular cells and beyond.\n\n3.1.. Cardiac Endothelial Cells\n-------------------------------\n\nIn primary rat cardiac microvascular endothelial cells, incubation with AGE prior to simulated ischemia and reperfusion resulted in increased release of lactic dehydrogenase (LDH) and caspase activity. The underlying mechanisms were linked to generation of oxidative stress, as evidenced by increased expression of iNOS, total nitric oxide production, formation of superoxide and peroxynitrite and nitrative inactivation of thioredoxin 1 (Trx-1), the latter essential for cellular protection \\[[@b16-ijms-14-19891]\\]. The adverse effects of AGE pretreatment prior to simulated ischemia/reperfusion were prevented by blockade of RAGE signaling. These findings underscore potential mechanisms by which diabetes or aging (in which AGEs accumulate) may predispose the cardiovasculature to increased vulnerability to damage initiated by ischemia and reperfusion.\n\n3.2.. Vascular Smooth Muscle Cells (SMCs)\n-----------------------------------------\n\nIn SMCs, RAGE ligand S100B resulted in increased production of superoxide, migration and expression of MCP-1 and IL6 in a manner dependent on S100B-induced activation of Src kinase and ERK signaling, together with activation of the transcription factors NF-\u03baB and STAT3 \\[[@b17-ijms-14-19891]\\].\n\nIn other studies, Hofmann Bowman and colleagues showed that transgenic expression of RAGE ligand S100A12 in SMCs mediated oxidative stress in pathways linked to vascular calcification. They showed that inhibition of NADPH oxidase blocked the effects of S100A12 on osteogenic gene programming and calcification mechanisms \\[[@b18-ijms-14-19891]\\]. Specifically, they linked these findings to a direct interaction between Nox1 and S100A12.\n\n3.3.. Inflammatory Cells\n------------------------\n\nAs discussed above, RAGE is expressed on monocytes/macrophages. Macrophages retrieved from mice deficient in gp91phox that failed to respond to AGEs solidified roles for NADPH oxidase in the mechanisms linking RAGE to oxidative stress \\[[@b14-ijms-14-19891]\\]. Other inflammatory cell types express RAGE, suggesting that oxidative stress is linked to the mechanisms of RAGE ligand-mediated perturbation.\n\nNeutrophils play prominent roles in infection, inflammation and diabetes. Omori and colleagues demonstrated that neutrophils retrieved from diabetic subjects displayed basally increased translocation of p47phox, a key component of activation of NADPH oxidase, to the cell membrane, in parallel with enhanced preassembly with p22phox. These findings were modeled in cultured HL60 cells in which it was shown that exposure of these cells to high glucose or to RAGE ligand S100B resulted in translocation of p47phox to the cell membrane in a manner dependent on ERK signal transduction and RAGE \\[[@b19-ijms-14-19891]\\].\n\nIn mast cells, which express RAGE, exposure of these cells to galectin-3 results in release of superoxide with consequent opening of the mitochondrial transition pore and caspase-dependent apoptosis. Consistent with key roles for RAGE, these effects of galectin-3 were blocked by pre-treatment of the cells with antibodies to RAGE or bongkrekic acid, the latter an antagonist of the mitochondrial transition pore \\[[@b20-ijms-14-19891]\\].\n\n3.4.. Endothelial Progenitor Cells (EPCs)\n-----------------------------------------\n\nIn diabetes and aging \\[[@b21-ijms-14-19891],[@b22-ijms-14-19891]\\], repair mechanisms are thwarted in concert with increased cellular stress and damage. In bone marrow derived EPCs, exposure to LPS-free *C*-reactive protein (CRP) (at physiologically-relevant concentrations) induced ROS generation, upregulation of RAGE expression, reduced antioxidant defenses and induction of EPC apoptosis. These processes were prevented by pre-treatment with anti-RAGE IgG \\[[@b23-ijms-14-19891]\\]. In other studies, incubation of EPCs with AGEs resulted in induction of ROS production, reduced anti-oxidant defenses, inhibition of EPC proliferation, migration and adhesion, and induction of EPC apoptosis. These effects were prevented by blockade of RAGE \\[[@b24-ijms-14-19891]\\]. In summary, the authors surmised that AGEs increased the vulnerability of EPCs to exogenous stresses and that RAGE is a key principal mediator of oxidative and cell death stresses in these cells.\n\n3.5.. Cells of the Nervous System\n---------------------------------\n\nAmyloid beta-peptide (A\u03b2) and beta sheet fibrils are specific ligands of RAGE. This ligand-RAGE interaction plays key roles in mechanisms linked to neuronal stress in Alzheimer's disease. In cultured neuronal cells, A\u03b2-RAGE interaction resulted in generation of TBARS and activation of NF-\u03baB, processes which were dependent on oxidative stress as demonstrated by blockade of these effects by *N*-acetyl cysteine \\[[@b6-ijms-14-19891]\\]. Furthermore, A\u03b2-RAGE mediated oxidative stress in neuronal cells results in upregulation of macrophage colony stimulating factor (MCS-F) through an NF-\u03baB-dependent pathway \\[[@b25-ijms-14-19891]\\]. Roles for oxidative stress were uncovered by the blocking effects of *N*-acetyl cysteine in this culture model. Notably, the upregulation of MCS-F was shown to stimulate microglial cell migration, proliferation, and survival, factors contributing to neuroinflammation in Alzheimer brain.\n\nIn cultured human NT2 neurons, exposure of these cells to increasing concentrations of the specific AGE, pentosidine, resulted in generation of ROS and apoptotic cell death. In parallel, expression of RAGE rose in a manner dependent of the concentration of AGEs \\[[@b26-ijms-14-19891]\\]. The mechanistic role for PKC-delta in pentosidine-mediated generation of ROS and upregulation was demonstrated by blockade of these effects by pretreatment of the neurons with rottlerin, an inhibitor of PKC-delta. Furthermore, pre-treatment of the neurons with the antioxidant, Vitamin E, also prevented generation of ROS and RAGE upregulation, thus suggesting that a key mechanism by which pentosidine initiated these effects was via generation of ROS.\n\nIn cultured astrocytes, treatment with AGE-bovine serum albumin (BSA) induced increased production of tumor necrosis factor (TNF)-\u03b1 and Interleukin 1\u03b2, in parallel with increased oxidative stress, as measured by reduced astrocyte glutathione (GSH) levels and increased nitric oxide (NO) release \\[[@b27-ijms-14-19891]\\]. Although the direct link to RAGE signaling was not demonstrated, the authors did show that RAGE was expressed by astrocytes.\n\nIn other studies, RAGE was shown to be expressed by primary sensory neurons, thereby suggesting potential links of the RAGE pathway to peripheral nervous system disorders. Vincent and colleagues showed that incubation of primary sensory neurons with the RAGE ligand S100 resulted in generation of oxidative stress and consequent increased activity of phosphatidyl inositol-3-kinase and activation of apoptotic mechanisms \\[[@b28-ijms-14-19891]\\]. These processes were prevented by pretreatment of the neurons with the anti-oxidant \u03b1 lipoic acid.\n\nIntriguingly, in other cells of the nervous system, specifically oligodendrocytes, oxidative stress modulated shedding of RAGE from the cell surface. Qin and colleagues showed that incubation of neonatal rat oligodendrocytes with low doses of hydrogen peroxide (100 \u03bcM) resulted in shedding of RAGE from the cell surface into the culture medium, in parallel with restoration of homeostatic levels of RAGE on the cell surface. In contrast, when the cells were exposed to high doses of hydrogen peroxide (\\>200 \u03bcM), the cells were shown to undergo cell death \\[[@b29-ijms-14-19891]\\]. Of note, both antioxidants and metalloprotease inhibitors blocked these effects in oligodendrocytes. These findings may have implications for diseases in which demyelination predominates, such as in multiple sclerosis and Batten's disease (neuronal ceroid-lipofuscinosis).\n\n3.6.. Osteoblast Cells\n----------------------\n\nAGEs have been shown to impart adverse consequences on osteoblasts; this bears implications for aging-associated bone disorders and diabetic complications. Schurman and colleagues studied two osteoblast cell lines, UMR106 and MC3T3E1 cells, and showed that incubation with AGEs stimulated oxidative stress, activation of caspase 3 activity and inhibition of alkaline phosphatase activity \\[[@b30-ijms-14-19891]\\]. In these cells, AGEs resulted in upregulation of RAGE protein. Of note, the authors showed that incubation of osteoblast cultured cells with metformin prevented the adverse consequences of AGE treatment.\n\n3.7.. Renal Cells\n-----------------\n\nIn addition to the roles for AGE-stimulated mitochondrial oxidative stress post-activation of cytosolic ROS in mesangial cells as discussed above, distinct studies in cultured mesangial cells highlighted roles for pigment epithelium derived factor (PEDF) in protection against AGE-induced oxidative and cellular damage. Ide and colleagues incubated renal mesangial cells with AGEs and identified increased oxidative stress, activation of NF-\u03baB and upregulation of MCP-1, VCAM-1 and PAI-1, all in a manner suppressed by treatment with PEDF or blocking antibodies to RAGE \\[[@b31-ijms-14-19891]\\].\n\nIn renal interstitial fibroblasts (NRK-49F cells), incubation with AGEs increased ROS and via ERK signaling induced cellular mitogenesis in a manner suppressed by antioxidants (*N*-acetyl cysteine) and reduction of RAGE expression using RAGE short hairpin RNA \\[[@b32-ijms-14-19891]\\].\n\n3.8.. Pancreatic \u03b2 Cells\n------------------------\n\nThe effects of RAGE ligands on pancreatic \u03b2 cells were studied in cultured rat INS-1 (insulinoma) cells. RAGE was found to be expressed by these cells and incubation with two distinct RAGE ligands, S100B or HMGB1, induced intracellular oxidative stress and cellular apoptosis. Consistent with key mechanistic roles for oxidative stress in RAGE ligand-mediated cell stress, pre-treatment of the cells with antioxidants or an NADPH oxidase inhibitor blocked these adverse effects on pancreatic \u03b2 cells \\[[@b33-ijms-14-19891]\\]. These findings might suggest that in pre-diabetes syndromes or in established diabetes, the increased production of AGEs (stimulated likely by hyperglycemia) might contribute to further derangement of glucose metabolism via direct attack on pancreatic insulin-producing cells.\n\nIn the section to follow, we discuss examples in which RAGE was directly linked to oxidative stress *in vivo*.\n\n4.. RAGE and Oxidative Stress: Evidence from *in Vivo* Models\n=============================================================\n\n*In vivo* model systems have provided support for roles for RAGE in oxidative stress in various conditions and tissues.\n\n4.1.. Cardiovascular System\n---------------------------\n\nStudies have addressed the role of RAGE and oxidative stress in the context of such disorders as atherosclerosis, aneurysm formation and ischemia-reperfusion injury.\n\nSun and colleagues tested the role of RAGE in atherosclerosis using mice deficient in the low density lipoprotein receptor (LDLR) and fed a western-type diet (0.15% cholesterol). Compared to RAGE expressing LDLR null mice, those mice devoid of RAGE displayed decreased atherosclerosis in parallel with reduced levels of vascular oxidative stress, as measured by detection of superoxide by dihydroethidium \\[[@b34-ijms-14-19891]\\].\n\nIn a distinct model, Hofmann Bowman and colleagues demonstrated that SMC specific overexpression of RAGE ligand S100A12 resulted in the development of aneurysm formation in the aorta. In parallel, evidence for increased oxidative stress, as measured by myeloperoxidase activity, was significantly higher in the transgenic *vs*. wild-type mouse vasculature \\[[@b35-ijms-14-19891]\\]. In primary cultures of SMCs from the transgenic and wild-type mice, increased oxidative stress was evident in these cells by increased staining for MitoSox (marker of mitochondrial oxidative stress), 8-hydroxy-2-deoxyguanosine and mitochondrial oxidative stress \\[[@b35-ijms-14-19891]\\]. As RAGE is the principal receptor for S100A12, it is highly plausible that these effects *in vivo* and in SMCs examined *ex vivo* were mediated via RAGE, but these concepts regarding potential roles for RAGE were not directly tested in this study.\n\nTikellis and colleagues tested the role of RAGE and oxidative stress *in vivo* in a model of high fat feeding. They showed that RAGE-deficient mice fed a Western-type diet, composed of 21% fat, displayed reduced markers of oxidative stress, including NADH- and NADPH-dependent mitochondrial superoxide production, inflammation and hypertrophy in the heart compared to the wild-type animals. Treatment with the AGE inhibitor, alagebrium, further reduced oxidative stress and inflammation in both RAGE null and the wild-type mice \\[[@b36-ijms-14-19891]\\].\n\nOxidative stress contributes critically to the pathogenesis of ischemia-reperfusion injury. Bucciarelli and colleagues showed that cardiac ischemia-reperfusion led to the upregulation of RAGE, inducible nitric oxide synthase expression and increased levels of nitric oxide, cyclic guanosine monophosphate (cGMP) and nitrotyrosine in both rat and mouse hearts. This was prevented by administration of soluble RAGE or by genetic deletion of RAGE in both non-diabetic and diabetic animals \\[[@b37-ijms-14-19891],[@b38-ijms-14-19891]\\].\n\nA consequence of ischemia-reperfusion is mitochondrial oxidative stress and the opening of the mitochondrial transition pore and cytochrome c release, harbingers to the activation of cell death apoptotic pathways \\[[@b39-ijms-14-19891],[@b40-ijms-14-19891]\\]. In a murine model of ligation and reperfusion of the left anterior descending coronary artery, genetic deletion of RAGE or administration of soluble RAGE significantly reduced release of cytochrome c and resulted in significant attenuation of infarct volume and the loss of cardiac function post-infarct. Signal transduction studies underscored key roles for JNK and STAT signaling in these oxidative stress--cell death mechanisms \\[[@b41-ijms-14-19891]\\].\n\nIn other studies, Tsoporis and colleagues induced myocardial infarction in rats and showed that this resulted in upregulation of RAGE expression in cardiomyocytes and release of RAGE ligand S100B, in parallel with myocyte apoptosis. Direct interactions between S100B and RAGE were illustrated by co-immunoprecipitation experiments. In rat neonatal cardiomyocyte cultures, S100 caused increased release of cytochrome c and cardiomyocyte apoptosis \\[[@b42-ijms-14-19891]\\].\n\n4.2.. Renal System\n------------------\n\nIn the kidney, roles for RAGE in oxidative stress in distinct pathophysiologic settings including diabetes-associated nephropathy, adriamycin (doxorubicin) toxicity and obesity associated renal dysfunction have been demonstrated.\n\nInagi and colleagues developed a triple transgenic mouse model of diabetes characterized by transgene-mediated expression of RAGE, iNOS and megsin. In the kidneys of these mice, significant acceleration of renal indices of nephropathy was shown in parallel with evidence of increased oxidative stress in triple transgenic mice *vs*. the controls, including increased accumulation of the AGEs pentosidine and CML AGE, and increased urinary levels of 8-oxo-2\u2032-deoxyguanosine (8-OHdG) \\[[@b43-ijms-14-19891]\\].\n\nRoles for RAGE in induction of renal dysfunction in obesity were demonstrated. When mice were fed a high fat diet and developed obesity, renal levels of mitochondrial superoxide and cytosolic superoxide were increased in the renal cortices of the animals compared to control lean mice. However, in RAGE null mice fed the high fat diet, levels of these two measures of oxidative stress were significantly attenuated. Furthermore, increased AGE CML noted in the kidneys of obese wild-type mice was significantly lower in obese RAGE null mice \\[[@b44-ijms-14-19891]\\]. These data strongly suggested that obesity-associated renal oxidative stress was in mediated at least in part via AGE-RAGE.\n\nFurther compelling evidence for roles for RAGE in renal associated oxidative stress emerged from experiments in which the nephrotoxin adriamycin (doxorubicin) was administered to wild-type or RAGE null mice (in the Balb/c background). RAGE null mice were strikingly protected against the adverse effects of adriamycin damage on podocyte effacement and glomerular sclerosis \\[[@b45-ijms-14-19891]\\]. A key mediator of adriamycin toxicity is postulated to be the induction of oxidative stress. Furthermore, exposure of wild-type mice to adriamycin resulted in generation of AGEs in the renal cortex. Consistent with roles for RAGE in oxidative stress damage in this model, RAGE null mice treated with adriamycin displayed marked reduction in renal levels of NADPH oxidase activity, malondialdehyde, iNOS transcripts, total nitrite and nitrate, and nitrotyrosine levels. Similar results were obtained when wild-type BALB/c mice were treated with adriamycin and soluble RAGE (sRAGE) \\[[@b45-ijms-14-19891]\\]. In cultured podocytes, RAGE ligand (S100B) stimulation resulted in activation of NADPH oxidase in a manner dependent on ERK signaling. These data provided compelling evidence linking RAGE to oxidative stress and renal fibrosis consequent to injection with adriamycin.\n\n4.3.. Central Nervous System\n----------------------------\n\nTo link RAGE and oxidative stress in the central nervous system, rats were treated with high doses of Vitamin A and the cerebral cortices were removed after treatment. At doses of Vitamin A greater than or equal to 2500 IU/kg/day, increased measures of oxidative stress were noted in the brain, as evidenced by higher levels of superoxide anion, carbonylation and 3-nitrotyrosine. In parallel with increased oxidative stress, RAGE protein levels (as assessed by Western blot) were significantly higher in the brain tissue in rats fed greater than or equal to 4500 IU/kg/day Vitamin A \\[[@b46-ijms-14-19891]\\]. Although these data did not establish a causal effect between RAGE expression and generation of oxidative stress in rats fed high doses of Vitamin A nor did they address the pattern of expression of RAGE ligands in Vitamin A feeding, they nevertheless supported that a relationship between the RAGE pathway and oxidative stress was mechanistically plausible *in vivo*.\n\n4.4.. Liver\n-----------\n\nOne consequence of chronic oxidative stress is the generation of AGEs and this is evidenced by the fact that oxidative stress induces hepatic protein glycation. Kuhla and colleagues showed that mice deficient in Uncoupling Protein 2 (UCP2) demonstrated increased oxidative stress (as determined by lower ratio of GSH to GSSG) \\[[@b47-ijms-14-19891]\\]. In parallel, increased oxidative stress was linked to reduced activity of glyoxalase 1 (Glo1); Glo1 is a chief enzyme responsible for the detoxification of major AGE precursors methylglyoxal and glyoxal \\[[@b48-ijms-14-19891]\\]. The authors showed that AGE and RAGE levels in the livers of UCP2 null mice were increased and that levels of the soluble RAGE decoy, soluble RAGE (sRAGE), were greatly reduced in these animals compared to the wild-type control mice \\[[@b47-ijms-14-19891]\\]. When liver injury was induced in these animals, administration of recombinant soluble RAGE exerted protection. The authors deduced that sRAGE's effects were mediated, at least in part, by the binding of RAGE ligands, as AGE levels were indeed reduced in the sRAGE-treated UCP2 null mice.\n\nFurther evidence for RAGE-mediated oxidative stress in the liver was evident in a murine model of acetaminophen toxicity. Administration of soluble RAGE to mice receiving toxic doses of this agent resulted in decreased levels of nitrotyrosine adducts in liver tissue together with higher levels of glutathione \\[[@b49-ijms-14-19891]\\]. In parallel with reduced measures of oxidative stress, soluble RAGE-treated mice displayed improved survival compared with those animals treated with vehicle \\[[@b49-ijms-14-19891]\\].\n\nThese data further buttress the concept that AGE-RAGE may incite a vicious cycle of oxidative stress and the induction of mechanisms that sustain and amplify ROS generation as well as inflammatory signals, mediated by the release of pro-inflammatory RAGE ligands, namely S100/calgranulins and HMGB1, in chronic stress. In the section to follow, we discuss novel insights into mechanisms by which RAGE signal transduction facilitates the generation of oxidative stress.\n\n5.. RAGE Signaling and Novel Roles for RAGE Cytoplasmic Domain Binding Partner, mDia1, in Oxidative Stress\n==========================================================================================================\n\nThe ligands of RAGE transduce their signals through this receptor by activation of a range of signaling pathways; the precise signaling pathways involved in RAGE action appear to be determined based on the nature of the cellular stress and the cell type. *In vivo*, undoubtedly, the signals emitted by surrounding cells likely further dictate the signaling pathways activated upon RAGE ligand stimulation. We discovered that the cytoplasmic domain of RAGE binds to the formin molecule mDia1 and that mDia1 is required for RAGE ligands to activate cell signaling responses in multiple cell types such as transformed cells, smooth muscle cells, and monocytes/macrophages \\[[@b50-ijms-14-19891]--[@b52-ijms-14-19891]\\]. Formins such as mDia1 (diaphanous1) are actin binding molecules that contribute to signal transduction mechanisms in part via effecting Rho GTPase signals \\[[@b53-ijms-14-19891]\\]. Hence, given the important role of Rac1, one of the Rho GTPases, as a component of activated NADPH oxidase, we tested the hypothesis that mDia1 was essential both for RAGE ligand-mediated generation of oxidative stress and for RAGE ligand-mediated signal transduction and cellular migration.\n\nThe role of mDia1 was studied in experiments in SMCs and in a murine model of arterial injury induced by endothelial denudation of the femoral artery \\[[@b52-ijms-14-19891]\\]. In this *in vivo* model, oxidative stress induced by the injury was significantly attenuated in both RAGE null and mDia1 null mice compared to wild-type control animals. The key role of mDia1 in this process was underscored by experiments which revealed the following: first, deletion of mDia1 *in vivo* reduced NADPH oxidase in the injured vessel; second, *in vitro*, exposure of SMCs retrieved from the aorta of mice to RAGE ligand S100B increased ROS production; this was prevented in SMCs retrieved from RAGE null or mDia1 null mice; third, in wild-type SMCs, treatment with S100B resulted in increased membrane translocation of p47phox and Nox1 with no effect on Nox4; this was prevented in RAGE null or mDia1 null SMCs; fourth, activation of Rac1, a key component of NADPH oxidase, was noted in wild-type but not RAGE null or mDia1 null SMCs treated with S100B; fifth, blockade of Rac1 with dominant negative Rac1 or blockade of Nox1 with siRNAs targeting Nox1 reduced S100B-mediated generation of ROS, suppressed activation of serine 9 GSK-3\u03b2; and sixth, treatment of wild-type SMCs with dominant negative Rac1 blocked S100B-mediated migration \\[[@b52-ijms-14-19891]\\]. These data revealed for the first time that mDia1 was a critical regulator of oxidative stress generation and SMC migration in response to RAGE ligands. Further experimentation is underway to determine if and by what mechanisms mDia1 might be linked to activation of NF-\u03baB.\n\nFinally, in the section to follow, we address evidence linking ligand-RAGE actions to oxidative stress mechanisms in human subjects.\n\n6.. AGE-RAGE and Oxidative Stress: Evidence from Human Subjects\n===============================================================\n\nEvidence is emerging from the study of human tissues and plasma to link ligand-RAGE activity to oxidative stress in human subjects bearing a range of pathological conditions---from central nervous disorders to rheumatoid arthritis. Although seemingly correlative, these findings nevertheless place ligand-RAGE and oxidative stress together *in vivo* in human subjects and suggest mechanistic roles for this pathway in oxidative stress damage as well as the possibility that ligand-RAGE might provide novel biomarkers into these pathobiological mechanisms.\n\n6.1.. Cardiometabolic and Renal Disease\n---------------------------------------\n\nSubjects with renal failure undergoing hemodialysis treatment are highly susceptible to acceleration of cardiovascular disease, irrespective of the etiology of loss of renal function. The link between inflammation and oxidative stress in this disorder is therefore of considerable interest. Rodriguez-Ayala and colleagues examined 7 subjects on hemodialysis with high-grade inflammation (*C*-reactive protein levels \\> 10 mg/L) and 11 subjects with low-grade inflammation (*C-*reactive protein levels \\< 10 mg/L) for at least a six month period. The authors found that the subjects with high-grade inflammation displayed higher levels of IL-6, myeloperoxidase activity and advanced oxidation protein products (AOPPs) *vs*. the low-grade inflammation subjects \\[[@b54-ijms-14-19891]\\]. In the subject group with high-grade inflammation, stimulation of their monocytes with CML-human serum albumin resulted in a rapid increase in activity of NF-\u03baB; this was prevented by pre-treatment with anti-RAGE antibodies \\[[@b54-ijms-14-19891]\\]. The authors suggested that inflammation and oxidative stress markers were possibly interrelated in end-stage renal diseases.\n\nIn addition to measurements of AGE products and immunohistochemical detection of RAGE in human diseased and normal tissue sections, other reports examine the relationship between soluble levels of RAGE in human plasma/serum and pathological disorders. Two forms of soluble RAGEs have been detected in human plasma/serum; the first is cell surface-cleaved soluble RAGE \\[[@b55-ijms-14-19891],[@b56-ijms-14-19891]\\] and the second form of soluble RAGE, is termed endogenous secretory RAGE. In contrast to cell surface cleaved RAGE, esRAGE is produced via an alternative splicing mechanism \\[[@b57-ijms-14-19891],[@b58-ijms-14-19891]\\]. In general, esRAGE is believed to represent approximately 20% of the total sRAGE content.\n\nSantilli and colleagues studied total plasma sRAGE levels, urinary 8-iso-prostaglandin F (PGF) 2\u03b1 excretion in urine and plasma levels of asymmetric dimethylarginine (ADMA) in 60 hypercholesterolemic subjects and 20 controls. They reported that plasma sRAGE was significantly lower and ADMA and urinary 8-iso-PGF2\u03b1 were higher in the group with hypercholesterolemia *vs*. controls. Multivariate regression analyses revealed that only 8-iso-PGF2\u03b1 and ADMA predicted the sRAGE levels \\[[@b59-ijms-14-19891]\\]. In subjects with hypercholesterolemia, treatment for 8 weeks with statin therapy (pravastatin or atorvastatin) resulted in a significant reduction in levels of urinary 8-iso-PGF2\u03b1 and treatment with atorvastatin resulted in a rise in sRAGE levels close to those in the normal controls; no changes in ADMA were noted. The authors concluded that sRAGE might be a protective factor against oxidative stress and endothelial dysfunction in atherosclerosis. Of note, however, the study did not distinguish total sRAGE *vs*. esRAGE components in the experiments.\n\nExperimental evidence has strongly linked RAGE and its ligands to oxidative stress in diabetes \\[[@b3-ijms-14-19891],[@b13-ijms-14-19891]\\]. Devangelio and colleagues tested these concepts in human subjects. They examined levels of sRAGE, ADMA and urinary 8-iso-PGF2\u03b1 in 86 diabetic subjects and 43 controls. They reported that levels of plasma sRAGE were lower and levels of ADMA were higher in the diabetic *vs*. control subjects and that a measure of glycemic control, hemoglobin A1c (HbA1c) and urinary 8-iso-PGF2\u03b1 were independently related to sRAGE levels in diabetic subjects. Of the 86 diabetic subjects, 24 with newly diagnosed diabetes and 12 of the subjects with poor metabolic control received therapies including hypoglycemic agents or insulin, respectively \\[[@b60-ijms-14-19891]\\]. The authors reported that improvement in metabolic control by these agents resulted in a significant rise in sRAGE levels in parallel with reduction in ADMA levels. The authors concluded that oxidative stress and endothelial dysfunction characteristic of diabetes might be associated with lower sRAGE levels and that attempts to improve metabolic control might be marked by improvements (increases) in the levels of sRAGE.\n\nRodino-Janeiro and colleagues examined expression of RAGE and oxidative stress markers in human adipose tissue and the relationship to coronary artery disease \\[[@b61-ijms-14-19891]\\]. They examined human subjects undergoing heart surgery and divided them into two groups, those with and without coronary artery disease. Biopsies of epicardial adipose tissue and subcutaneous adipose tissue were performed and revealed that the expression of RAGE mRNA transcripts and protein in the subcutaneous tissue form subjects with coronary artery disease was lower than that in subjects without coronary artery disease. No changes in RAGE expression were noted in the epicardial adipose tissue between these two groups of subjects. However, p22phox and RAGE gene expression were higher in the epicardial *vs*. the subcutaneous adipose tissue, in parallel with reduced levels of catalase mRNA. Interestingly, the subunits of NADPH oxidase were not impacted by the presence or absence of coronary artery disease. The authors concluded that a potential relationship between RAGE expression levels and ROS production in adipose tissue was possible.\n\n6.2.. Pulmonary Disease\n-----------------------\n\nInghilleri and colleagues examined sections from lung tissue retrieved from subjects with usual interstitial pneumonia (UIP) *vs*. cryptogenic organizing pneumonia (COP) for levels of RAGE, oxidative stress, nitrosylation and fibroblast lesions \\[[@b62-ijms-14-19891]\\]. They reported that in both forms of pneumonia, RAGE expression was equivalently high. However, levels of superoxide dismutases and iNOS, which were diffusely present in COP endoalveolar plugs, were not detected in UIP fibroblasts lesions. Tissue nitrosylation was also found to be lower in UIP *vs*. COP. The authors concluded that while RAGE expression was high in both settings, suppression of antioxidant enzyme expression was only observed in UIP. These data suggest that oxidative stress pathways may exert distinct influences in UIP *vs*. COP, both of which are characterized by increased expression of RAGE and fibroblast lesions. It is notable, however, that the authors did not study specific RAGE ligands---such as AGEs, S100/calgranulins or HMGB1 in this work; it is conceivable that insights into the varied findings might underlie their findings.\n\n6.3.. Central Nervous System\n----------------------------\n\nChronic neurodegenerative disorders in human subjects are well-known for their association with oxidative stress. In the setting of diseases characterized by Lewy bodies, Dalfo and colleagues showed that compared to the brains of healthy matched controls, subjects with Lewy bodies displayed the following: increased mass spectrometric and immunological evidence of increased malondialdehyde lysine (MDAL) and 4-hydroxynonenal lysine (HNE), increased AGEs and heterogeneous expression of RAGE in multiple regions in the brain, such as substantial nigra, amygdala and frontal cortex \\[[@b63-ijms-14-19891]\\]. Increased concentration of docosahexaenoic acid in the amygdala and frontal cortex was also demonstrated. These authors were able to further identify two proteins bearing oxidative damage, \u03b1-synuclein and manganese superoxide dismutase in incidental Lewy body disease cortex. Thus, by localizing AGE-RAGE to oxidative stress, the study supports that a mechanistic role for this pathway in the generation of ROS in Lewy body disease is plausible.\n\nIn a distinct study, Freixes and colleagues examined cerebral cortex in Creutzfeldt-Jakob disease. Immunohistochemistry studies were performed to test the concept that expression of oxidative, glycoxidative, lipoxidative and nitrative protein damage were evident in the cerebral cortex in this disorder. The authors found that expression of CML and carboxy ethyl lysine (CEL) AGEs, 4-hydroxynonenal, MDAL, nitrotyrosine, neuronal, endothelial and inducible NOS isoforms, RAGE and superoxide dismutase1 and manganese superoxide dismutase 2 levels were increased in Creutzfeldt-Jakob brain *vs*. healthy age-matched control brain \\[[@b64-ijms-14-19891]\\]. They concluded that oxidative stress mechanisms might contribute to the pathophysiology of prion disease.\n\n6.4.. Immune/Inflammatory Disorders: Rheumatoid Arthritis\n---------------------------------------------------------\n\nRheumatoid arthritis (RA) is a chronic autoimmune disorder characterized by chronic inflammation leading to destruction of bones and joint tissues. Ferrante and colleagues examined markers of oxidative stress and plasma levels of esRAGE in 54 subjects with RA and 20 healthy control subjects \\[[@b65-ijms-14-19891]\\]. The authors reported that levels of urinary markers of oxidative stress were higher in RA patients *vs*. controls and that levels of esRAGE were lower in RA subjects *vs*. controls. Interestingly, a direct correlation between the urinary markers of oxidative stress was observed only in patients not receiving anti-TNF\u03b1 treatment. In contrast, subjects receiving anti-TNF\u03b1 treatment displayed significantly lower levels of urinary 8-iso-PGF2\u03b1 but not urinary 11-dehydro-thromboxane B2 (TXB) 2 *vs*. control subjects. In the latter case, esRAGE was the only independent predictor of the levels of urinary 11-dehydro-TXB2 \\[[@b65-ijms-14-19891]\\]. The authors concluded that the anti-TNF\u03b1 agents might only exert efficacy against isoprostane generation with no impact on thromboxane synthesis. Further, they posited that hyperactivity of the RAGE pathway might escape TNF\u03b1 blockade, thereby contributing to the synthesis of thromboxane in these subjects.\n\n6.5.. Disorders of Reproduction: Preeclampsia\n---------------------------------------------\n\nPreeclampsia, a complication of pregnancy, is associated with oxidative stress. Chekir and colleagues tested the hypothesis that AGE-RAGE and oxidative stress measures might differ in the placentas of human subjects with or free of preeclampsia. Serum levels of AGEs were found to be higher in subjects with preeclampsia compared to those with pregnancy but without preeclampsia and compared to normal healthy subjects \\[[@b66-ijms-14-19891]\\]. Western blotting performed on the extracts of placental tissues revealed that levels of AGEs and RAGE were higher in women with preeclampsia *vs*. normal pregnancy. By immunohistochemical analyses, levels of 4-hydroxynonenal and 8-hydroxy-2\u2032-deoxyguanosine were higher in the preeclamptic *vs*. normal pregnancy placenta \\[[@b66-ijms-14-19891]\\]. These studies suggested possible mechanistic links between AGE-RAGE and the oxidative stress observed in the placental tissue in preeclampsia.\n\n7.. Summary and Perspectives\n============================\n\nTaken together, these considerations support a compelling multi-step role for ligand-RAGE interaction in diverse cell types and systems *in vivo* (both in animal models and in human subjects) in the pathogenesis of oxidative stress ([Figure 1](#f1-ijms-14-19891){ref-type=\"fig\"}). Although oxidative stress has been proposed to contribute to a wide range of disorders, from aging, neurodegeneration, diabetes and its complications, autoimmunity, cardiovascular diseases, pulmonary disorders, and reproductive disorders such as preeclampsia, common threads linking increased generation of ROS and/or reduced antioxidative potential to precise pathogenic mechanisms have yet to be firmly demonstrated. We thus propose a model implicating AGE formation as an initial inciting event in the diverse array of disorders indicated above. We predict that AGEs, once formed, contribute to attraction and activation of inflammatory cells, or, in other settings, activation of endogenous cells---one consequence of which in both situations is the release/expression of non-AGE RAGE ligands, such as S100/calgranulins, HMGB1 or others. Such release and availability of these RAGE ligands may further amplify RAGE-dependent oxidative stress. Furthermore, as illustrated by enlightening experiments in UCP2 null mice discussed above, oxidative stress may beget further AGE production \\[[@b47-ijms-14-19891]\\], thereby fueling the engines of oxidative and cellular stress and tissue damage. Additional evidence supporting this premise is deduced from the work of Anderson and Heinecke in which they showed that phagocyte-derived oxidants from NADPH oxidase are required for the production of CML AGEs during inflammation \\[[@b67-ijms-14-19891]\\]. Together with RAGE-dependent suppression of Glo1 (as identified in diabetic renal cortical tissue) \\[[@b68-ijms-14-19891]\\], these considerations suggest that RAGE might be a key nodal point in generation of ROS and the pathogenesis of chronic disease. Studies are underway to unravel the precise roles of mDia1 in RAGE-dependent oxidative stress to determine if RAGE signal transduction itself is a logical and effective target for therapeutic intervention in the diverse array of RAGE-associated pathologies characterized by accelerated oxidative stress.\n\nThe authors gratefully acknowledge research support from the US Public Health Service and the Juvenile Diabetes Research Foundation. The authors are grateful to Latoya Woods for her expert assistance in the preparation of this manuscript.\n\nThe authors declare no conflict of interest.\n\n![AGE-RAGE, oxidative stress and amplification loops linked to the pathogenesis of chronic disease. We propose that generation and accumulation of AGEs (**red triangles**) may be an important triggering event in a diverse array of stimuli, such as hyperglycemia, natural aging, inflammatory mechanisms and oxidative stress (such as via the myeloperoxidase pathway), and ischemia-reperfusion. Once AGEs achieve degrees of modification/concentrations sufficient to bind to and activate RAGE, RAGE signaling, at least in part through the interaction of its cytoplasmic domain with the formin, mDia1, results in activation of NADPH oxidase and the generation of ROS. Such ROS, via mitochondrial amplification, may generate even further ROS. One consequence of increased NADPH oxidase- and mitochondrial ROS generation is the consumption of antioxidant defenses. Together with RAGE-dependent suppression of Glo1, AGE formation and accumulation is sustained and amplified. Once set in motion, AGEs may stimulate recruitment of RAGE-expressing inflammatory cells which, when activated, may result in the release of non-AGE RAGE ligands, such as S100/calgranulins and HMGB1 (**blue triangles**). Such non-AGE + AGE ligands of RAGE may stimulate yet an additional amplification loop of RAGE activation, that is, cellular stress and activation of gene programs that augur tissue damage and reduced activation of repair mechanisms. We posit that such cellular stimulation mediated by AGE-RAGE and the indicated amplification loops contribute to the pathogenesis of the complications of diabetes, cardiovascular disease, autoimmunity and innate aging. Blocking AGE formation, RAGE signaling and the indicated amplification loops may be essential for the prevention of these chronic disorders.](ijms-14-19891f1){#f1-ijms-14-19891}\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec007}\n============\n\nOverweight and obesity significantly increase the probability of metabolic disorders and chronic diseases like diabetes; cardiovascular diseases, particularly heart failure and coronary heart disease; nonalcoholic fatty liver disease; neoplasms (endometrial, breast post-menopause, prostate, liver, pancreas, colorectum, and kidney); musculoskeletal disorders; and respiratory diseases \\[[@pone.0235121.ref001]--[@pone.0235121.ref005]\\]. Overweight or obesity (body mass index \\[BMI\\] \u226525 kg/m^2^) characterize 82.5%, 76.4%, and 73.6% of people with Type 2 diabetes, hypertension, and dyslipidemia, respectively \\[[@pone.0235121.ref006]\\].\n\nExcess adipose tissue favors the release of free fatty acids (FFAs) from adiposities to the cardiovascular system \\[[@pone.0235121.ref001],[@pone.0235121.ref006]\\]. This leads to an increase in lipid accumulation, including in hepatocytes and skeletal muscle cells. The accumulation of lipids in these cells through the activation of the diacylglycerol-protein kinase C pathway may contribute to the emergence of insulin resistance \\[[@pone.0235121.ref007]\\]. Weight gain accompanied by an increase in body fat leads to insulin resistance and to disorders in the expression of various adipokines. This includes classical hormones, such as leptin; inflammatory mediators like tumor necrosis factor-alfa, interleukins -1, -6, and -8, resistin, and chemerin; enzymes; and metabolites \\[[@pone.0235121.ref001], [@pone.0235121.ref006], [@pone.0235121.ref007]\\].\n\nThe pathophysiology of obesity-induced dyslipidemia involves, amongst other disorders, a reduced lipolysis of triglyceride (TG)-rich lipoproteins, impairment of the peripheral uptake of FFAs, increased FFA flow from adiposities to the liver and other tissues, overproduction of very low density lipoproteins (VLDL) in the liver, and formation of small dense low density lipoproteins (LDL) \\[[@pone.0235121.ref008]\\]. These disorders lead to the development of nonalcoholic fatty liver disease. A decrease in high density lipoprotein (HDL) concentration in obesity involves both an increased uptake of the HDL2 subfraction by adiposities and the increased catabolism of the HDL apolipoprotein A-I \\[[@pone.0235121.ref009]\\]. Insulin resistance also contributes to changes in lipid metabolism and the development of atherogenic dyslipidemia \\[[@pone.0235121.ref006]\\]. The failure to inhibit the microsomal TG transfer protein and lipoprotein lipase activation observed in insulin resistance leads to hypertriglyceridemia \\[[@pone.0235121.ref010]\\].\n\nThe most important mechanisms in the development of obesity-induced hypertension includes the physical compression of the kidneys by accumulating fatty cells, which impedes sodium excretion, followed by the activation of the renin-angiotensin-aldosterone system (RAAS). There is also an increase in the activity of the sympathetic nervous system, which is probably caused by elevated leptin secretion, the activation of the melanocortin system in the brain, and resistance to insulin \\[[@pone.0235121.ref011], [@pone.0235121.ref012]\\]. Obesity is also related to the overproduction of adipokines, which disturb the functions of the endothelium of the blood vessels, vasoconstriction, and vasodilation \\[[@pone.0235121.ref012]\\]. Consequences of these disorders include metabolic syndrome (MetS) and cardiovascular diseases.\n\nThe early identification of metabolic disorders allows for an early intervention and the prevention of serious consequences. Overweight and obesity indices, calculated based on anthropometric measurements, have found a wide application in the identification of metabolic disorders \\[[@pone.0235121.ref013]--[@pone.0235121.ref018]\\]. The most significant advantages of anthropometric indices include the following: non-invasiveness, low cost, standardized techniques and simplicity of measurements, and the possibility to apply them on a large scale. However, it is still difficult to decide which of the indices is the best tool for the early identification of metabolic disorders. In this study, we analyzed both traditional obesity indicators, which have been used for many decades, and those that have been developed relatively recently. We selected indicators that can be calculated based on 2--3 of the simplest anthropometric measurements and that assess not only weight and height proportions (BMI) but also overall body fat (body fat percentage \\[%BF\\], Cl\u00ednica Universidad de Navarra-body adiposity estimator \\[CUN-BAE\\]) and fat distribution (waist circumference \\[WC\\], waist-to-height ratio \\[WHtR\\]). We also selected indicators that combine several measurements defining the geometry of the human body (body roundness index \\[BRI\\], a body shape index \\[ABSI\\]) and which were developed using allometric analysis. The usefulness of the BMI for estimating obesity is limited because the values at the high end of the BMI scale can be attributed to either increased fat mass or lean body mass. The commonly used BMI classification excludes people with an increased metabolic risk resulting from high body fat \\[[@pone.0235121.ref019]\\]. The CUN-BAE is based on the BMI, but it has the advantage of taking into account age and sex. The percentage of fat calculated using the CUN-BAE showed a stronger correlation with the actual amount of adipose tissue than any other anthropometric fat indicator \\[[@pone.0235121.ref019]\\]. Many studies have shown that the results of central obesity measurements have the strongest correlations with metabolic risk factors \\[[@pone.0235121.ref020]\\]. The disadvantage of the WC is that taller people have larger circumferences. Also, WC values differ between ethnic groups \\[[@pone.0235121.ref021]\\]. However, some studies have found that the use of abdominal obesity measurements does not improve the prediction of metabolic risk factors compared with the BMI \\[[@pone.0235121.ref022], [@pone.0235121.ref023]\\]. According to Ashwell and Gibson, the WHtR identifies more people with cardiovascular risk than a combination of the WC and the BMI \\[[@pone.0235121.ref024]\\]. Other studies indicate that the usefulness of the WHtR differs depending on sex \\[[@pone.0235121.ref025]\\]. The ABSI is minimally correlated with height, mass, and the BMI, and can therefore be used to distinguish the independent contributions of WC and BMI to cardiometabolic risk factors \\[[@pone.0235121.ref026]\\]. The ABSI had a positive linear relationship with all-cause and cardiovascular mortality in Europeans, while the corresponding relationship with BMI, WC, and WHtR was J-shaped \\[[@pone.0235121.ref027]\\]. The ABSI was also useful to identify visceral and sarcopenic obesity in patients with diabetes \\[[@pone.0235121.ref028]\\]. The BRI was developed to assess body shape independently of height. It is a better predictor of body fat and the percentage of visceral tissue compared with traditional indicators, such as the BMI and WC \\[[@pone.0235121.ref029]\\]. The discriminatory power of the BRI in predicting MetS was similar to that of the WC and only slightly lower than that of the WHtR \\[[@pone.0235121.ref030]\\]. It was higher than that of the BMI and ABSI in both sexes. In our previous paper we analyzed the usefulness of these indices to identify people with MetS \\[[@pone.0235121.ref018]\\]. In this study, we sought to validate whether the anthropometric indices previously recommended can also be applied to identify single MetS components to avoid a late diagnosis. Our aims were to assess the usefulness of the anthropometric indices and to determine optimal cut-off points for the identification of the MetS components in adults.\n\nMaterials and methods {#sec008}\n=====================\n\nStudy design and participants {#sec009}\n-----------------------------\n\nThe base population of the study was 13,172 participants in the Polish-Norwegian Study (PONS). From this group, 844 individuals (6.4%) were excluded due to incomplete anthropometric measurements and/or biochemical parameters that were necessary to perform this analysis. Consequently, this study was carried out on 12,328 participants aged 55.7\u00b15.4 years. The group included 4094 men. All the participants came from the \u015awi\u0119tokrzyski region in Poland and were Caucasian. The study design, recruitment of participants, and course of the study were described in detail in our previous publications \\[[@pone.0235121.ref018], [@pone.0235121.ref031]\\].\n\nConsent to collect the data was given by the Ethics Committee from the Cancer Centre and Institute of Oncology in Warsaw, Poland (No. 69/2009/1/2011). Consents to carry out analyses of the data were given by the Committee on Bioethics at the Faculty of Health Sciences from Jan Kochanowski University in Kielce (No. 45/2016). Informed written consent was obtained from all participants enrolled in the study.\n\nAnthropometric obesity indices, Blood Pressure and Serum Biochemical Parameters Weight, height, and WC were used to calculate obesity indices. All anthropometric measurements were conducted by trained nurses using standard protocols and techniques \\[[@pone.0235121.ref032]\\]. WC was measured in the horizontal plane midway between the lower rib edge and the upper iliac crest using a non-elastic metric measure. Height was measured using a stadiometer. Weight and %BF were measured with a body composition analyzer (Tanita SC 240MA). Blood pressure (BP) was measured on the right upper limb artery with an Omron pressure monitor (Model M3 Intellisense). The mean of two measurements was used for subsequent analysis. Serum TGs were measured with an enzymatic method using phosphoglycerol oxidase and determination of H~2~O~2~ (with peroxidase). HDL cholesterol was measured with a colorimetric non-precipitation method, using polyethylene glycol-modified enzymes. The glucose concentration was measured with an enzymatic method using hexokinase.\n\nWe considered the following obesity indices: WC, BMI, WHtR, %BF, CUN-BAE \\[[@pone.0235121.ref019]\\], BRI \\[[@pone.0235121.ref029]\\], and ABSI \\[[@pone.0235121.ref026]\\]. The following equations were used:\n\n-   BMI = weight (kg) / height (m)^2^;\n\n-   WHtR = WC (cm) / height (cm);\n\n-   CUN-BAE was calculated using the following equation: %BF = \u2212 44.988 + (0.503 \u00d7 age) + (10.689 \u00d7 sex) + (3.172 \u00d7 BMI) \u2212 (0.026 \u00d7 BMI^2^) + (0.181 \u00d7 BMI \u00d7 sex) \u2212 (0.02 \u00d7 BMI \u00d7 age) \u2212 (0.005 \u00d7 BMI^2^ \u00d7 sex) + (0.00021 \u00d7 BMI^2^ \u00d7 age), where age is measured in years and sex was codified as 0 for men and 1 for women \\[[@pone.0235121.ref019]\\];\n\n-   BRI = 364.2--365.5 x $\\sqrt{1 - \\left\\lbrack \\frac{\\left( {WC/\\left( 2\\pi \\right))}^{2} \\right.}{\\left( 0.5\\ x\\ Height \\right)^{2}} \\right\\rbrack}$ \\[[@pone.0235121.ref029]\\];\n\n-   ABSI = WC (m) / \\[BMI^2/3^(kg/m^2^) height^1/2^ (m)\\] \\[[@pone.0235121.ref026]\\].\n\nThe definition of metabolic risk factors {#sec010}\n----------------------------------------\n\nFour MetS components were included in the analysis: elevated BP (\u2265130 mmHg and/or diastolic blood pressure \u226585 mmHg or drug treatment for hypertension), high glucose concentration (\u2265100 mg/dl or \u22655.5 mmol/L or diabetes treatment), high TG concentration (\u2265150 mg/dL or \u22651.7 mmol/L or drug treatment for elevated triglycerides), and low HDL cholesterol. (\\<40 mg/dL or \\<1.0 mmol/L in men and \\<50 mg/dl or \\<1.3 mmol/L in women or drug treatment for low HDL cholesterol) \\[[@pone.0235121.ref021]\\].\n\nStatistical analysis {#sec011}\n--------------------\n\nAll statistical analyses were performed using the statistical package Statistica 13.3 (TIBCO SOFTWARE INC, Polish version, PL, Cracow). The participants were divided into two groups, according to the presence or lack of a given metabolic disorder. Additionally, the analyses were performed for at least one MetS component. Groups divided according to the MetS components were compared with the non-parametric Mann-Whitney U-test because of the non-normal distribution of all quantitative variables. The analyses were done separately for both sexes.\n\nWe performed receiver-operating characteristic (ROC) curve analyses to determine the discriminatory power of the anthropometric indices as classifiers. Next, we calculated the areas under the curve (AUCs) and the 95% confidence intervals (CI) to compare the discriminatory power of each index. Sensitivity was defined as the percentage of true positive scores according to the criteria of individual MetS components. Specificity was defined as the proportion of scores identified incorrectly. The AUC is a measure of the precision of a given index in the differentiation between individuals with metabolic disorders and without them. It also characterizes the probability of assigning a patient to the correct group. For the AUC, 0.5 was adopted as the bottom border line. The indices with the biggest AUC were considered the best. Optimal cut-off points for all seven obesity indicators were determined with Youden's J statistic using the following equation: J~max.~ = Sensitivity + Specificity-- 1. The index values corresponding to the maximum value of Youden's J statistic were recognized as optimal cut-off points for these indices.\n\nResults {#sec012}\n=======\n\nAll anthropometric indices were significantly higher in men with abnormal metabolic parameters than in men with parameters within the normal range (according to the definition of MetS by the International Diabetes Federation \\[IDF\\]) \\[[@pone.0235121.ref021]\\] (p\\<0.001, [Table 1](#pone.0235121.t001){ref-type=\"table\"}). Men who were not diagnosed with any of the metabolic disorders had the lowest average indices (BMI = 25.76 kg/m^2^, WC = 91.78 cm, and %BF = 23.06%). Men with abnormal HDL cholesterol concentration had the highest index values (BMI = 29.68 kg/m^2^, WC = 102.76 cm, and %BF = 29.05%).\n\n10.1371/journal.pone.0235121.t001\n\n###### Baseline anthropometric and laboratory characteristics of men (N = 4094).\n\n![](pone.0235121.t001){#pone.0235121.t001g}\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  MetS Components               BMI \\[kg/m^2^\\] X\u00b1SD Me\u00b1IQR   WC \\[cm\\] X\u00b1SD Me\u00b1IQR   WHtR X\u00b1SD Me\u00b1IQR   %BF X\u00b1SD Me\u00b1IQR   ABSI \\[m^11/6^ \u00b7 kg^-2/3^\\] X\u00b1SD Me\u00b1IQR   BRI X\u00b1SD Me\u00b1IQR   CUN-BAE \\[%\\] X\u00b1SD Me\u00b1IQR   \n  ----------------------------- ----------------------------- ----------------------- ------------------ ----------------- ----------------------------------------- ----------------- --------------------------- -------------\n  BP                            normal n = 722                26.73\u00b13.46\\             94.56\u00b19.17\\        0.546\u00b10.054\\      24.62\u00b16.16\\                               0.080\u00b10.004\\      4.60\u00b10.67\\                  27.50\u00b14.51\\\n                                                              26.49\u00b14.31              94.00\u00b111.00        0.544\u00b10.070       24.00\u00b17.20                                0.081\u00b10.005       4.58\u00b10.89                   27.36\u00b16.17\n\n  abnormal n = 3372             28.83\u00b13.94\\                   100.17\u00b110.39\\           0.579\u00b10.061\\       27.63\u00b16.41\\       0.081\u00b10.004\\                              4.96\u00b10.75\\        30.31\u00b14.80\\                 \n                                28.53\u00b14.84                    100.00\u00b113.00            0.576\u00b10.078        27.10\u00b17.90        0.081\u00b10.005                               4.91\u00b10.98         30.15\u00b16.03                  \n\n  Glucose                       normal n = 2237               27.61\u00b13.64\\             96.83\u00b19.73\\        0.559\u00b10.057\\      25.84\u00b16.23\\                               0.081\u00b10.004\\      4.75\u00b10.71\\                  28.70\u00b14.61\\\n                                                              27.28\u00b14.48              96.00\u00b113.00        0.557\u00b10.074       25.20\u00b17.50                                0.081\u00b10.005       4.70\u00b10.92                   28.54\u00b15.88\n\n  abnormal n = 1857             29.49\u00b14.05\\                   102.01\u00b110.50\\           0.589\u00b10.062\\       28.63\u00b16.42\\       0.081\u00b10.004\\                              5.07\u00b10.76\\        31.15\u00b14.84\\                 \n                                29.10\u00b14.90                    101.00\u00b113.00            0.585\u00b10.077        28.00\u00b18.00        0.081\u00b10.005                               5.02\u00b10.97         30.97\u00b15.93                  \n\n  TG                            normal n = 2452               27.72\u00b13.80\\             97.15\u00b110.08\\       0.561\u00b10.059\\      25.90\u00b16.32\\                               0.081\u00b10.004\\      4.77\u00b10.73\\                  28.88\u00b14.79\\\n                                                              27.43\u00b14.68              97.00\u00b113.00        0.558\u00b10.076       25.30\u00b17.60                                0.081\u00b10.005       4.73\u00b10.96                   28.75\u00b16.02\n\n  abnormal n = 1642             29.57\u00b13.90\\                   102.21\u00b110.16\\           0.590\u00b10.060\\       28.90\u00b16.27\\       0.081\u00b10.004\\                              5.09\u00b10.74\\        31.20\u00b14.65\\                 \n                                29.18\u00b14.74                    102.00\u00b113.00            0.587\u00b10.075        28.30\u00b17.60        0.081\u00b10.003                               5.02\u00b10.94         30.98\u00b15.97                  \n\n  HDL                           normal n = 3424               28.72\u00b16.39\\             98.48\u00b110.17\\       0.568\u00b10.060\\      26.72\u00b16.39\\                               0.081\u00b10.004\\      4.84\u00b10.73\\                  29.48\u00b14.83\\\n                                                              27.89\u00b14.87              8.00\u00b112.00         0.566\u00b10.077       26.10\u00b17.90                                0.081\u00b10.005       4.80\u00b10.96                   29.36\u00b16.19\n\n  abnormal n = 670              29.68\u00b14.03\\                   102.76\u00b110.86\\           0.596\u00b10.064\\       29.05\u00b16.48\\       0.082\u00b10.005\\                              5.17\u00b10.78\\        31.49\u00b14.72\\                 \n                                29.20\u00b15.11                    102.00\u00b114.00            0.591\u00b10.079        28.50\u00b17.50        0.081\u00b10.005                               5.09\u00b10.96         31.18\u00b16.05                  \n\n  At least one MetS component   No n = 1755                   25.76\u00b13.10\\             91.78\u00b18.78\\        0.532\u00b10.052\\      23.06\u00b15.77\\                               0.080\u00b10.004\\      4.47\u00b10.66\\                  26.22\u00b14.19\\\n                                                              25.57\u00b13.85              91.50\u00b111.50        0.528\u00b10.070       22.55\u00b16.30                                0.080\u00b10.005       4.45\u00b10.83                   26.29\u00b15.53\n\n  Yes n = 2339                  28.71\u00b13.92\\                   99.85\u00b110.28\\            0.577\u00b10.061\\       27.90\u00b16.40\\       0.081\u00b10.004\\                              4.93\u00b10.74\\        30.14\u00b14.80\\                 \n                                28.41\u00b14.87                    99.00\u00b113.00             0.573\u00b10.077        26.90\u00b17.90        0.081\u00b10.005                               4.88\u00b10.95         30.01\u00b16.07                  \n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nX--arithmetic mean, SD--standard deviation, Me--median, IQR--interquartile range, BP--blood pressure, TG--triglycerides, HDL--HDL cholesterol, MetS--metabolic syndrome, WC--waist circumference, BMI--body mass index, WHtR--waist-to-height ratio, %BF--body fat percentage, ABSI--a body shape index, BRI--body roundness index, CUN-BAE--Cl\u00ednica Universidad de Navarra-body adiposity estimator\n\nSimilarly, all obesity indices were significantly higher in women with abnormal metabolic parameters than in those with normal parameters (p\\<0.001; [Table 2](#pone.0235121.t002){ref-type=\"table\"}). Women who were not diagnosed with any of the metabolic disorders had the lowest indices (BMI = 25.39 kg/m^2^, WC = 81.40 cm, %BF = 32.39%). Women with abnormal glucose concentration had the highest index values (BMI = 30.30 kg/m^2^, WC = 94.03 cm, and %BF = 38.63).\n\n10.1371/journal.pone.0235121.t002\n\n###### Baseline anthropometric and laboratory characteristics of women (N = 8234).\n\n![](pone.0235121.t002){#pone.0235121.t002g}\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  MetS Components               BMI \\[kg/m^2^\\] X\u00b1SD Me\u00b1IQR   WC \\[cm\\] X\u00b1SD Me\u00b1IQR   WHtR X\u00b1SD Me\u00b1IQR   %BF X\u00b1SD Me\u00b1IQR   ABSI \\[m^11/6^ \u00b7 kg^-2/3^\\] X\u00b1SD Me\u00b1IQR   BRI X\u00b1SD Me\u00b1IQR   CUN-BAE \\[%\\] X\u00b1SD Me\u00b1IQR   \n  ----------------------------- ----------------------------- ----------------------- ------------------ ----------------- ----------------------------------------- ----------------- --------------------------- -------------\n  BP                            normal n = 2470               25.86\u00b13.83\\             82.90\u00b19.87\\        0.517\u00b10.064\\      33.18\u00b16.43\\                               0.076\u00b10.005\\      4.74\u00b10.84\\                  38.15\u00b14.79\\\n                                                              25.38\u00b14.95              82.00\u00b113.00        0.512\u00b10.085       33.70\u00b18.20                                0.075\u00b10.006       4.68\u00b11.11                   37.92\u00b16.57\n\n  abnormal n = 5764             28.90\u00b15.11\\                   90.36\u00b112.02\\            0.567\u00b10.078\\       37.21\u00b16.47\\       0.075\u00b10.005\\                              5.36\u00b11.02\\        41.83\u00b15.53\\                 \n                                28.28\u00b16.65                    90.00\u00b116.00             0.561\u00b10.104        37.70\u00b18.40        0.075\u00b10.006                               5.28\u00b11.34         41.69\u00b17.67                  \n\n  Glucose                       normal n = 5982               27.12\u00b14.51\\             85.90\u00b110.95\\       0.537\u00b10.071\\      35.01\u00b16.59\\                               0.077\u00b10.005\\      4.99\u00b10.93\\                  39.72\u00b15.29\\\n                                                              26.52\u00b15.79              85.00\u00b115.00        0.531\u00b10.095       35.50\u00b18.60                                0.077\u00b10.006       4.91\u00b11.24                   39.46\u00b17.26\n\n  abnormal n = 2252             30.30\u00b15.35\\                   94.03\u00b112.36\\            0.591\u00b10.080\\       38.63\u00b16.32\\       0.075\u00b10.005\\                              5.66\u00b11.05\\        43.38\u00b15.46\\                 \n                                29.74\u00b16.97                    93.00\u00b117.00             0.587\u00b10.107        39.30\u00b18.00        0.075\u00b10.005                               5.59\u00b11.38         43.36\u00b17.56                  \n\n  TG                            normal n = 5650               27.27\u00b14.74\\             86.11\u00b111.41\\       0.538\u00b10.074\\      35.03\u00b16.78\\                               0.075\u00b10.006\\      5.00\u00b10.97\\                  39.84\u00b15.48\\\n                                                              26.55\u00b16.09              85.00\u00b115.00        0.531\u00b10.010       35.50\u00b19.00                                0.075\u00b10.040       4.90\u00b11.28                   39.46\u00b17.63\n\n  abnormal n = 2584             29.55\u00b15.09\\                   92.54\u00b111.81\\            0.582\u00b10.076\\       38.13\u00b16.05\\       0.077\u00b10.006\\                              5.55\u00b11.00\\        42.66\u00b15.29\\                 \n                                28.87\u00b16.59                    91.00\u00b116.00             0.575\u00b10.105        38.40\u00b17.65        0.077\u00b10.004                               5.46\u00b11.33         42.44\u00b17.27                  \n\n  HDL                           normal n = 6665               27.63\u00b14.87\\             87.17\u00b111.74\\       0.545\u00b10.076\\      35.53\u00b16.77\\                               0.076\u00b10.006\\      5.08\u00b10.99\\                  40.26\u00b15.55\\\n                                                              26.93\u00b16.24              86.00\u00b116.00        0.538\u00b10.103       36.00\u00b19.00                                0.080\u00b10.005       4.99\u00b11.31                   39.97\u00b17.66\n\n  abnormal n = 1569             29.49\u00b15.08\\                   92.19\u00b111.80\\            0.581\u00b10.077\\       38.00\u00b16.09\\       0.077\u00b10.006\\                              5.55\u00b11.01\\        42.68\u00b15.26\\                 \n                                28.80\u00b16.60                    91.00\u00b116.00             0.573\u00b10.105        38.30\u00b17.70        0.077\u00b10.004                               5.46\u00b11.34         42.50\u00b17.20                  \n\n  At least one MetS component   No n = 4786                   25.39\u00b13.69\\             81.40\u00b19.43\\        0.507\u00b10.077\\      32.39\u00b14.71\\                               0.076\u00b10.005\\      4.62\u00b10.81\\                  37.49\u00b14.71\\\n                                                              24.91\u00b14.59              80.00\u00b113.00        0.556\u00b10.103       37.27\u00b18.00                                0.076\u00b10.006       4.54\u00b11.09                   37.27\u00b16.14\n\n  Yes n = 3448                  28.66\u00b15.03\\                   89.87\u00b111.87\\            0.563\u00b10.077\\       36.94\u00b16.48\\       0.075\u00b10.005\\                              5.31\u00b11.01\\        41.56\u00b15.48\\                 \n                                28.04\u00b16.49                    89.00\u00b115.00             0.556\u00b10.103        37.50\u00b18.40        0.074\u00b10.006                               5.23\u00b11.33         41.43\u00b17.58                  \n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nX--arithmetic mean, SD--standard deviation, Me--median, IQR--interquartile range, BP--blood pressure, TG--triglycerides, HDL--HDL cholesterol, MetS--metabolic syndrome, WC--waist circumference, BMI--body mass index, WHtR--waist-to-height ratio, %BF--body fat percentage, ABSI--a body shape index, BRI--body roundness index, CUN-BAE--Cl\u00ednica Universidad de Navarra-body adiposity estimator\n\nIn men, the largest AUC for elevated BP, was for the CUN-BAE (0.668) and the BMI (0.660). For abnormal glucose concentration, the largest AUC was for the CUN-BAE (0.649) and the WC (0.645) ([Table 3](#pone.0235121.t003){ref-type=\"table\"}). The largest AUC for TGs was for the WC (0.642) and the %BF (0.641). The CUN-BAE, BMI, and WC also had high discriminatory power for at least one MetS component, with an AUC of 0.734, 0.728, and 0.728, respectively.\n\n10.1371/journal.pone.0235121.t003\n\n###### Areas under the curve (AUCs) and cut-off points for anthropometric indices for the prediction of MetS components in men.\n\n![](pone.0235121.t003){#pone.0235121.t003g}\n\n  MetS Components               Indices   AUC            95%CI          p       Sensitivity   Specificity   Youden index   Cut-off points\n  ----------------------------- --------- -------------- -------------- ------- ------------- ------------- -------------- ----------------\n  BP                            BMI       0.660          0.638--0.681   0.000   0.651         0.594         0.245          27.18\n  WC                            0.657     0.636--0.678   0.000          0.590   0.639         0.228         98.00          \n  WHtR                          0.655     0.633--0.676   0.000          0.582   0.647         0.228         0.564          \n  %BF                           0.646     0.624--0.668   0.000          0.575   0.659         0.234         26.10          \n  ABSI                          0.542     0.519--0.565   0.000          0.346   0.715         0.061         0.083          \n  BRI                           0.638     0.616--0.659   0.000          0.538   0.677         0.215         4.855          \n  CUN-BAE                       0.668     0.647--0.960   0.000          0.625   0.630         0.255         28.76          \n  Glucose                       BMI       0.641          0.625--0.659   0.000   0.629         0.601         0.228          28.15\n  WC                            0.645     0.628--0.661   0.000          0.702   0.503         0.206         97.00          \n  WHtR                          0.641     0.624--0.658   0.000          0.666   0.534         0.199         0.561          \n  %BF                           0.630     0.613--0.647   0.000          0.674   0.527         0.202         25.60          \n  ABSI                          0.545     0.527--0.563   0.000          0.653   0.368         0.051         0.080          \n  BRI                           0.624     0.607--0.641   0.000          0.663   0.519         0.182         4.737          \n  CUN-BAE                       0.649     0.633--0.666   0.000          0.613   0.620         0.233         29.89          \n  TG                            BMI       0.639          0.622--0.656   0.000   0.648         0.560         0.208          27.93\n  WC                            0.642     0.625--0.659   0.000          0.600   0.618         0.218         100.00         \n  WHtR                          0.639     0.622--0.656   0.000          0.688   0.513         0.201         0.560          \n  %BF                           0.641     0.624--0.658   0.000          0.690   0.533         0.223         25.70          \n  ABSI                          0.543     0.525--0.561   0.000          0.599   0.469         0.068         0.081          \n  BRI                           0.622     0.605--0.639   0.000          0.605   0.425         0.180         4.859          \n  CUN-BAE                       0.639     0.622--0.656   0.000          0.683   0.525         0.208         29.04          \n  HDL                           BMI       0.607          0.584--0.629   0.000   0.649         0.519         0.168          28.08\n  WC                            0.614     0.59--0.637    0.000          0.613   0.558         0.172         100.00         \n  WHtR                          0.624     0.601--0.647   0.000          0.448   0.735         0.183         0.601          \n  %BF                           0.608     0.586--0.631   0.000          0.693   0.483         0.175         25.90          \n  ABSI                          0.556     0.532--0.580   0.000          0.357   0.676         0.073         0.083          \n  BRI                           0.622     0.599--0.645   0.000          0.685   0.499         0.184         4.803          \n  CUN-BAE                       0.619     0.596--0.641   0.000          0.715   0.473         0.187         29.04          \n  At least one MetS component   BMI       0.728          0.702--0.755   0.000   0.586         0.776         0.362          27.65\n  WC                            0.728     0.701--0.755   0.000          0.740   0.615         0.355         94.00          \n  WHtR                          0.715     0.687--0.742   0.000          0.617   0.656         0.327         0.549          \n  %BF                           0.717     0.689--0.744   0.000          0.575   0.774         0.348         25.90          \n  ABSI                          0.562     0.530--0.593   0.000          0.566   0.526         0.092         0.081          \n  BRI                           0.682     0.653--0.710   0.000          0.658   0.618         0.276         4.612          \n  CUN-BAE                       0.734     0.708--0.761   0.000          0.611   0.759         0.370         28.76          \n\nBP--blood pressure, TG--triglycerides, HDL--HDL cholesterol, MetS--metabolic syndrome, WC--waist circumference, BMI--body mass index, WHtR--waist-to-height ratio, %BF--body fat percentage, ABSI--a body shape index, BRI--body roundness index, CUN-BAE--Cl\u00ednica Universidad de Navarra-body adiposity estimator\n\nIn women, the largest AUCs for elevated BP were for the CUN-BAE (0.691) and the WHtR (0.688) ([Table 4](#pone.0235121.t004){ref-type=\"table\"}). For abnormal glucose and TG concentration, the largest AUC was for the WHtR (0.694 and 0.664, respectively). The WHtR, CUN-BAE, and WC had the highest discriminatory power for at least one MetS component (0.715, 0.714, and 0.712, respectively). In both men and women, the largest AUCs for HDL were for the WHtR (0.624 and 0.632, respectively) and BRI (0.622 and 0.634, respectively).\n\n10.1371/journal.pone.0235121.t004\n\n###### Areas under the curve (AUCs) and cut-off points for anthropometric indices for the prediction of MetS components in women.\n\n![](pone.0235121.t004){#pone.0235121.t004g}\n\n  MetS Components               Indices   AUC            95%CI          p       Sensitivity   Specificity   Youden index   Cut-off points\n  ----------------------------- --------- -------------- -------------- ------- ------------- ------------- -------------- ----------------\n  BP                            BMI       0.681          0.669--0.693   0.001   0.488         0.783         0.271          28.44\n  WC                            0.684     0.672--0.696   0.001          0.564   0.712         0.275         88.00          \n  WHtR                          0.688     0.676--0.700   0.001          0.631   0.649         0.279         0.536          \n  %BF                           0.679     0.666--0.691   0.001          0.567   0.704         0.271         36.80          \n  ABSI                          0.575     0.562--0.589   0.001          0.383   0.727         0.111         0.078          \n  BRI                           0.681     0.669--0.693   0.001          0.635   0.366         0.269         4.931          \n  CUN-BAE                       0.691     0.679--0.703   0.001          0.587   0.701         0.287         40.52          \n  Glucose                       BMI       0.681          0.668--0.694   0.001   0.597         0.679         0.276          28.60\n  WC                            0.691     0.678--0.704   0.001          0.716   0.559         0.276         87.00          \n  WHtR                          0.694     0.681--0.707   0.001          0.612   0.675         0.287         0.564          \n  %BF                           0.662     0.649--0.675   0.001          0.591   0.660         0.250         38.00          \n  ABSI                          0.595     0.581--0.608   0.001          0.624   0.517         0.141         0.076          \n  BRI                           0.686     0.673--0.699   0.001          0.578   0.699         0.278         5.400          \n  CUN-BAE                       0.688     0.675--0.700   0.001          0.635   0.652         0.286         41.56          \n  TG                            BMI       0.636          0.623--0.649   0.001   0.650         0.558         0.208          27.20\n  WC                            0.657     0.644--0.669   0.001          0.641   0.592         0.233         88.00          \n  WHtR                          0.664     0.651--0.676   0.001          0.676   0.562         0.238         0.543          \n  %BF                           0.638     0.626--0.651   0.001          0.671   0.542         0.213         36.14          \n  ABSI                          0.605     0.592--0.618   0.001          0.631   0.528         0.159         0.076          \n  BRI                           0.661     0.648--0.673   0.001          0.731   0.504         0.235         4.910          \n  CUN-BAE                       0.647     0.634--0.659   0.001          0.747   0.476         0.223         39.09          \n  HDL                           BMI       0.610          0.595--0.625   0.001   0.639         0.535         0.174          27.32\n  WC                            0.622     0.607--0.636   0.001          0.625   0.553         0.178         88.00          \n  WHtR                          0.632     0.617--0.647   0.001          0.584   0.608         0.193         55.87          \n  %BF                           0.607     0.593--0.622   0.001          0.713   0.454         0.167         35.30          \n  ABSI                          0.578     0.563--0.594   0.001          0.617   0.500         0.117         0.076          \n  BRI                           0.634     0.619--0.649   0.001          0.652   0.540         0.192         5.082          \n  CUN-BAE                       0.625     0.610--0.640   0.001          0.652   0.544         0.196         40.61          \n  At least one MetS component   BMI       0.702          0.688--0.715   0.001   0.555         0.756         0.311          27.41\n  WC                            0.712     0.699--0.726   0.001          0.547   0.771         0.317         88.00          \n  WHtR                          0.715     0.702--0.728   0.001          0.634   0.689         0.322         0.532          \n  %BF                           0.701     0.688--0.714   0.001          0.589   0.723         0.313         36.14          \n  ABSI                          0.603     0.588--0.618   0.001          0.553   0.597         0.149         0.076          \n  BRI                           0.706     0.692--0.719   0.001          0.618   0.693         0.310         4.934          \n  CUN-BAE                       0.714     0.701--0.727   0.001          0.570   0.759         0.329         40.85          \n\nBP--blood pressure, TG--triglycerides, HDL--HDL cholesterol, MetS--metabolic syndrome, WC--waist circumference, BMI--body mass index, WHtR--waist-to-height ratio, %BF--body fat percentage, ABSI--a body shape index, BRI--body roundness index, CUN-BAE--Cl\u00ednica Universidad de Navarra-body adiposity estimator\n\nFor the BMI, the optimal cut-off point for an early diagnosis of single metabolic disorders (MetS components) was 27.2 kg/m^2^ for both sexes. This point was the lowest for BP for men and TG for women. The optimal cut-off point for WC was 94 cm for men (for at least 1 component of MetS) and 87 cm for women (for glucose). For other indices, optimal cut-off points for the identification of single metabolic disorders in men include WHtR = 0.549, %BF = 25.6, CUN-BAE = 28.76%, BRI = 4.612, and ABSI = 0.080. For women, the optimal cut-off points were WHtR = 0.532, %BF = 35.3, CUN-BAE = 39.09%, BRI = 4.910, and ABSI = 0.076. The discriminatory power of the ABSI was the lowest both in men and women (AUC\\<0.6).\n\nDiscussion {#sec013}\n==========\n\nThe usefulness of anthropometric indices for predicting single metabolic disorders that are MetS components was different depending on sex and the type of metabolic abnormality. In men, the CUN-BAE had the largest AUC, followed by the BMI and WC. In women, the largest AUCs were those of the WHtR, CUN-BAE, and WC. However, the differences between the AUCs of most indices were small. Therefore, the predictive power of most of the indices is similar. In men, indices with a similar predictive power to those with the largest AUCs include the %BF and WHtR. In women, they include the BRI, BMI, and %BF. One reason for the similarity in the predictive power between the indices may be that they are correlated to a large extent \\[[@pone.0235121.ref018]\\]. The ABSI had the lowest discriminatory power for predicting metabolic disorders in both sexes (AUC\\<0.6).\n\nOther authors have reached diverse conclusions concerning the predominance of particular anthropometric indices over other indicators in diagnosing obesity and metabolic disorders. Our results are similar to those obtained in the population of Jordan \\[[@pone.0235121.ref033]\\]. Khader et al. recommend the use of the WHtR for the diagnosis of metabolic disorders defined according to the IDF. However, none of the anthropometric indices included in their analysis (i.e. BMI, WC, WHR, and WHtR) was significantly better than the others in identifying most of these disorders. Similarly, in studies conducted in the Spanish (Caucasian) population, all obesity indices, except for the ABSI, had similar discriminatory power in the prediction of MetS \\[[@pone.0235121.ref034]\\]. When both sexes were analyzed separately, the BMI had the largest AUC in men and the WHtR and BRI had the largest AUCs in women. Similar to the results of our analysis in men, Davila-Batista et al. concluded that the CUN-BAE was the best index for the identification of individuals with hypertension, diabetes, and MetS \\[[@pone.0235121.ref035]\\]. Corbat\u00f3n Anchuelo et al. emphasize the advantages of using the WC and WHtR in the identification of cardiometabolic disorders in women \\[[@pone.0235121.ref016]\\]. In the Chinese population, the best predictors of cardiometabolic disorders in men and women were the BRI and WHtR \\[[@pone.0235121.ref036]\\]. Moreover, these indices were the best predictors of elevated BP and the presence of at least one (any) metabolic disorder in men. Also in men, the BMI and WC were the best indices for the identification of dyslipidemia and MetS. The ABSI had the lowest discriminatory power, in agreement with our findings. The results depend to a large extent on the ethnicity \\[[@pone.0235121.ref037], [@pone.0235121.ref038]\\], sex \\[[@pone.0235121.ref036], [@pone.0235121.ref037]\\], and age of the participants \\[[@pone.0235121.ref039]\\], and also on the indices selected for analysis.\n\nAccording to the National Cholesterol Education Program, Adult Treatment Panel III (NCEP ATP III), the cut-off point for abdominal obesity is 102 cm for men and 88 cm for women \\[[@pone.0235121.ref021]\\]. These values correspond to a BMI of about 30 kg/m^2^. However, according to the IDF criteria, abdominal obesity in Europeans should be recognized when the WC is 94 cm for men and 80 cm for women. In a study on the Saudi population, the optimal cut-off point for the WC for the identification of at least 2 components of MetS was 92 cm for men and 87 cm for women \\[[@pone.0235121.ref040]\\]. In a study conducted in Jordan, the cut-off point for the WC for the identification of single metabolic disorders fluctuated between 88.5 and 91.8 cm in men. Analogically, the cut-off point in women ranged from 84.5 to 88.5 cm \\[[@pone.0235121.ref033]\\]. From these studies and our own observations, we can conclude that optimal cut-off point for the WC is 94 cm for men and 87 cm for women. Therefore, the IDF criteria are more useful for the early identification of metabolic disorders, especially in men. The optimal cut-off point for the WHtR for the identification of MetS was 0.549 for men and 0.532 for women and was only slightly higher than the cut-off point for abdominal obesity (\u22650.5) \\[[@pone.0235121.ref041]\\]. For the CUN-BAE, the cut-off point for the diagnosis of single metabolic disorders was 28.76% for men and 39.09% for women. In the MARK Study, Gomez-Marcos et al. obtained slightly higher cut-off points for the identification of individuals with MetS, according to the criteria of NCEP ATP III (31.22% for men and 41.95% for women) \\[[@pone.0235121.ref034]\\]. An optimal cut-off point for the BMI for the identification of metabolic disorders is 27.2 kg/m^2^ for both sexes. Similar values for MetS were obtained in a study conducted in Israel (27 kg/m^2^) \\[[@pone.0235121.ref042]\\]. Compared with our results, the BMI cut-off values for individual metabolic disorders obtained in studies conducted in Jordan, were slightly lower in men (26.2--27.2 kg/m^2^), but higher in women (27.2--30.0 kg/m^2^) \\[[@pone.0235121.ref033]\\].\n\nConsidering there is a high probability that the indices will have considerable discriminatory power to identify individuals with single metabolic disorders, we recommend using the WHtR and the WC because they are the simplest to measure and interpret. For the WC, the IDF criteria should be preferred, especially in men. We also confirm the usefulness of the CUN-BAE. Despite being based on the BMI, it has the advantage of allowing for the age and sex of participants. In their long-term study on a population of European descent, Vinknes et al. found that the CUN-BAE was more strongly correlated with the risk of cardiovascular diseases and diabetes than the BMI \\[[@pone.0235121.ref043]\\]. The cut-off points for the identification of single metabolic disorders, were slightly lower than the optimal points for the identification \u22652 and \u22653 MetS components in the same population \\[[@pone.0235121.ref018]\\]. Our results allow us to conclude that these anthropometric indices can be used to identify individuals with single metabolic disorders. However, they are more useful for the identification of individuals with at least 2 MetS components \\[[@pone.0235121.ref018]\\].\n\nLimitations {#sec014}\n-----------\n\nThe presence of MetS or its components in individuals with a normal body mass (metabolically obese normal weight---MONW) indicates that not all cases of metabolic disorders are characterized by high anthropometric indices. MetS can be related not only to excessive adipose tissue, but also to the location of this tissue and changes in its functions. Anatomical and/or functional changes in adipose tissue, promoted by a positive energy balance, are responsible for the so-called adiposopathy (\"sick fat\") in genetically susceptible people. Hormonal and immunological reactions caused by adiposopathy can make metabolic disorders more severe, for example by causing dyslipidemia and elevated BP \\[[@pone.0235121.ref044]\\]. Moreover, genetic variation and epigenetic factors play an important role in the pathogenesis of MetS \\[[@pone.0235121.ref045]\\]. Eating habits and physical activity are the main environmental factors conditioning the expression of genes involved in the occurrence of MetS \\[[@pone.0235121.ref046], [@pone.0235121.ref047]\\]. Diets that are rich in fats, especially in saturated fatty acids, with a high glycemic index, and a low fiber content can increase the risk of a MetS. Conversely, diets characterized by a low consumption of sugar, sweets, refined grains, processed meat, and high consumption of fish, legumes, nuts, whole grains, and phytochemical-rich foods decrease the risk of metabolic disorders \\[[@pone.0235121.ref048]--[@pone.0235121.ref050]\\]. Interventions involving physical activity also positively influence each of the MetS components \\[[@pone.0235121.ref046], [@pone.0235121.ref047]\\]. Several genes associated with the ability of skeletal muscles to use lipids have been identified, which helps explain how physical activity affects the concentration of lipids in the blood \\[[@pone.0235121.ref051]\\].\n\nConclusions {#sec015}\n===========\n\nWe recommend the following indices of nutritional status for the identification of the MetS components: CUN-BAE\\>BMI = WC in men and WHtR\\>CUN-BAE\\>WC in women. Except for the ABSI, the diagnostic value of the other indices we analyzed was very similar. Prospective studies are needed to identify those indices in which changes in value predict the development of metabolic disorders best.\n\nIn the diagnosis of metabolic disorders, the cut-off point for the WC should be considered in accordance with the IDF rather than the NCEP ATP III, especially for men.\n\nSupporting information {#sec016}\n======================\n\n###### Data.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nWe are grateful to the members of the PONS project team for their contribution to the study and data sharing.\n\n10.1371/journal.pone.0235121.r001\n\nDecision Letter 0\n\nSanada\n\nKiyoshi\n\nAcademic Editor\n\n\u00a9 2020 Kiyoshi Sanada\n\n2020\n\nKiyoshi Sanada\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n14 Apr 2020\n\nPONE-D-20-07763\n\nAnthropometric indices and cut-off points in the diagnostics of the metabolic disorders\n\nPLOS ONE\n\nDear Professor Suliga,\n\nThank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE's publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.\n\nWe would appreciate receiving your revised manuscript by May 29 2020 11:59PM. When you are ready to submit your revision, log on to <https://www.editorialmanager.com/pone/> and select the \\'Submissions Needing Revision\\' folder to locate your manuscript file.\n\nIf you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.\n\nTo enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: <http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols>\n\nPlease include the following items when submitting your revised manuscript:\n\nA rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled \\'Response to Reviewers\\'.A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled \\'Revised Manuscript with Track Changes\\'.An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled \\'Manuscript\\'.\n\nPlease note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.\n\nWe look forward to receiving your revised manuscript.\n\nKind regards,\n\nKiyoshi Sanada, PhD\n\nAcademic Editor\n\nPLOS ONE\n\nJournal Requirements:\n\nWhen submitting your revision, we need you to address these additional requirements:\n\n1\\. Please ensure that your manuscript meets PLOS ONE\\'s style requirements, including those for file naming. The PLOS ONE style templates can be found at <http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf> and <http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf>\n\n2\\. Please note that outmoded terms and potentially stigmatizing labels should be changed to more current, acceptable terminology. Examples: \"Caucasian\" should be changed to \"white\" or \"of \\[Western\\] European descent\" (as appropriate); \"cancer victims\" should be changed to \"patients with cancer.\" <https://journals.plos.org/plosone/s/submission-guidelines#loc-human-subjects-research>.\n\n3\\. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.\n\n4\\. Your ethics statement must appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please also ensure that your ethics statement is included in your manuscript, as the ethics section of your online submission will not be published alongside your manuscript.\n\n\\[Note: HTML markup is below. Please do not edit.\\]\n\nReviewers\\' comments:\n\nReviewer\\'s Responses to Questions\n\n**Comments to the Author**\n\n1\\. Is the manuscript technically sound, and do the data support the conclusions?\n\nThe manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.\n\nReviewer \\#1: Partly\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n2\\. Has the statistical analysis been performed appropriately and rigorously?\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n3\\. Have the authors made all data underlying the findings in their manuscript fully available?\n\nThe [PLOS Data policy](http://www.plosone.org/static/policies.action#sharing) requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data---e.g. participant privacy or use of data from a third party---those must be specified.\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n4\\. Is the manuscript presented in an intelligible fashion and written in standard English?\n\nPLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.\n\nReviewer \\#1: No\n\nReviewer \\#2: No\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n5\\. Review Comments to the Author\n\nPlease use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)\n\nReviewer \\#1: The authors obtained cut-off points of anthropometric indices in the diagnostics of the metabolic disorders. I think these cut-off points are important. But, there are several problems. I have the following comments.\n\nSeven indices (WC, BMI, WHtR, %BF, CUN-BAE, BRI, and ABSI) were taken into consideration in this study. Why did you chose these indices? Please describe the reason in Introduction. Also, please explain the relationship between these indices and metabolic disorders based on previous studies, simply.\n\nIs the base population of this study 13,172 (Suliga et al. 2019)?\n\nThe authors described \"This analysis was conducted on the data of 12,328 participants\" in this manuscript. How were these 12,328 participants selected? Please describe the process. Please show a flowchart, if possible.\n\nPlease describe how each index was calculated and how components of metabolic syndrome were evaluated.\n\nWho measured the waist circumference? I think the accuracy of the waist circumference measurement is important. Please show reproducibility, if possible.\n\nThere are some unexplained abbreviations in the manuscript (e.g., LDL, IDF, NCEP ATP III, MONW). Please modify.\n\nThe number of normal and abnormal participants in each variable was not shown. Table 1 and 2 are too busy. Why are both the mean and median shown in table 1 and 2? Please modify.\n\nThe value in the text is different from the value in the table. Also, there are \"men/women\" and \"male/female\" in the manuscript. Please check.\n\nReviewer \\#2: Major comments\n\n1\\. There are various types of metabolic syndrome risk. More should be mentioned about which factors are associated with what metabolic risk in the Introduction. You should then add any issues or hypotheses. Moreover What is the difference between the selected factors? Please mention as allometric scaling.\n\n2\\. Throughout the text, data should be considered with significant figures. For example, age does not require a second decimal place.\n\n3\\. Subjects include persons with patients. Please clarify the exclusion criteria for subjects.\n\n4\\. Regarding the criteria for metabolic syndrome, do you need to consider separately the classification based on the reference value and the drug taker? The conditions differ for those who only meet the criteria and those who take medication.\n\n5\\. Title the statistical analysis and summarize it. There is no description of Me \u00b1 IQR. Do you need it?\n\n6\\. The purpose of this study is to compare each factor, so Table 1 needs ANCOVA. At least covariates need age, sex, and smoking status.\n\n7\\. Please state the conclusion concisely.\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n6\\. PLOS authors have the option to publish the peer review history of their article ([what does this mean?](https://journals.plos.org/plosone/s/editorial-and-peer-review-process#loc-peer-review-history)). If published, this will include your full peer review and any attached files.\n\nIf you choose \"no\", your identity will remain anonymous but your review may still be made public.\n\n**Do you want your identity to be public for this peer review?** For information about this choice, including consent withdrawal, please see our [Privacy Policy](https://www.plos.org/privacy-policy).\n\nReviewer \\#1: No\n\nReviewer \\#2: No\n\n\\[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link \\\"View Attachments\\\". If this link does not appear, there are no attachment files to be viewed.\\]\n\nWhile revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, <https://pacev2.apexcovantage.com/>. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at <figures@plos.org>. Please note that Supporting Information files do not need this step.\n\n10.1371/journal.pone.0235121.r002\n\nAuthor response to Decision Letter 0\n\n18 May 2020\n\nReviewer \\#1:\n\nSeven indices (WC, BMI, WHtR, %BF, CUN-BAE, BRI, and ABSI) were taken into consideration in this study. Why did you chose these indices? Please describe the reason in Introduction. Also, please explain the relationship between these indices and metabolic disorders based on previous studies, simply.\n\nThe Introduction has been supplemented with information on indices, as suggested by the reviewer.\n\nIs the base population of this study 13,172 (Suliga et al. 2019)?\n\nThe authors described \"This analysis was conducted on the data of 12,328 participants\" in this manuscript. How were these 12,328 participants selected? Please describe the process. Please show a flowchart, if possible.\n\nThe base population of this study was 13,172. From this group, 844 individuals (6.4%) were rejected due to a lack of complete anthropometric measurements and / or biochemical parameters that were necessary to perform this analysis. Consequently, the study was carried out on 12,328 participants.\n\nPlease describe how each index was calculated and how components of metabolic syndrome were evaluated. Who measured the waist circumference? I think the accuracy of the waist circumference measurement is important. Please show reproducibility, if possible.\n\nIn the text of the study, we have included formulas for calculating the indicators. Information on anthropometric measurements and metabolic factor analyzes has also been provided.\n\nThere are some unexplained abbreviations in the manuscript (e.g., LDL, IDF, NCEP ATP III, MONW). Please modify.\n\nAbbreviations have been explained.\n\nThe number of normal and abnormal participants in each variable was not shown. Table 1 and 2 are too busy. Why are both the mean and median shown in table 1 and 2? Please modify.\n\nThe numbers of participants with normal and abnormal metabolic parameters have been provided. Tables 1 and 2 have been corrected to remove unnecessary information.\n\nThe value in the text is different from the value in the table. Also, there are \"men/women\" and \"male/female\" in the manuscript. Please check.\n\nMale / female \\\"has been corrected to\\\" men / women \\\". The differences in values in the tables and in the text have been explained. The values given in the text have been checked again with the data in the tables. We suppose that the incompatibilities noted by the reviewer may result from the fact that as optimal cut-off points, we ultimately adopted those that were the lowest and were not always points for \\\"at least 1 component of MetS\\\". The result depended, to a large extent, on the risk factor which was most common in the participants of the study. An appropriate explanation has been included in the text.\n\nReviewer \\#2: Major comments\n\n1\\. There are various types of metabolic syndrome risk. More should be mentioned about which factors are associated with what metabolic risk in the Introduction. You should then add any issues or hypotheses. Moreover What is the difference between the selected factors? Please mention as allometric scaling.\n\nWe agree that there are various types of metabolic syndrome risks. However, due to the extensive amount of information, we decided to present only those factors that were the subject of analysis in our work. We have only supplemented them partially, because our introduction is already relatively long.\n\nOur analysis focused on the search for cut-off points for 7 anthropometric indicators and their possible future use in the approach to metabolic disorders. Therefore, in our opinion, it was not advisable to include an allometric approach in this situation.\n\n2\\. Throughout the text, data should be considered with significant figures. For example, age does not require a second decimal place.\n\nThe data were corrected as suggested by the reviewer.\n\n3\\. Subjects include persons with patients. Please clarify the exclusion criteria for subjects.\n\nOur project was an epidemiological study aimed at determining the prevalence of selected non-communicable diseases and their risk factors. Participants volunteered for the study. Therefore, the study group included both healthy people and those who had already begun pharmacological treatment for various metabolic disorders.\n\n4\\. Regarding the criteria for metabolic syndrome, do you need to consider separately the classification based on the reference value and the drug taker? The conditions differ for those who only meet the criteria and those who take medication.\n\nAccording to the definitions of both NCEP ATP III and IDF, not only abnormal values of a given parameter (e.g. blood pressure) are considered as components of the metabolic syndrome, but also those taking medications for this reason are taken into consideration. Therefore, we conducted analyzes for all persons meeting these criteria, guided by the above definitions.\n\n5\\. Title the statistical analysis and summarize it. There is no description of Me \u00b1 IQR. Do you need it?\n\nThe title of the subsection \\\"Statistical Analysis\\\" was on page 4, line 114. The description of statistical analyzes in our opinion does not differ from the applicable standards. Me \u00b1 IQR was removed as suggested by the reviewer.\n\n6\\. The purpose of this study is to compare each factor, so Table 1 needs ANCOVA. At least covariates need age, sex, and smoking status.\n\nThe purpose of this study was the evaluation of the usefulness of selected indices and the determination of optimal cut-off points for the identification of metabolic disorders being the components of metabolic syndrome. Therefore, we did not create models estimating the impact of confounding variables on the risk of metabolic disorders. In order to determine the discriminatory power of anthropometric indices as classifiers, receiver-operating characteristic (ROC) analyzes were performed. Next, in order to compare the discriminatory power of each index, areas under the curve ROC (AUC) were calculated. The AUC comprises a measure of precision of a given index in differentiation between individuals with metabolic disorders or without them , as well as characterizing the probability of a correct classification of a participant to a correct group. To our knowledge, this is the standard way of determining the discriminatory power of anthropometric indices and determining optimal cut-off points. Tables 1 and 2 present only the basic characteristics of the study group. The main purpose of using covariance analysis (ANCOVA) is to check to what extent the average of the analyzed parameters change under the influence of control variables and studied interactions. The use of covariance analysis in this situation would rather change the approach to the research problem raised in this work, creating a completely new problem that can be the subject of different analysis and the purpose of another work.\n\n7\\. Please state the conclusion concisely.\n\nConclusions were corrected as suggested by the reviewer\n\n10.1371/journal.pone.0235121.r003\n\nDecision Letter 1\n\nSanada\n\nKiyoshi\n\nAcademic Editor\n\n\u00a9 2020 Kiyoshi Sanada\n\n2020\n\nKiyoshi Sanada\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n1 Jun 2020\n\nPONE-D-20-07763R1\n\nAnthropometric indices and cut-off points in the diagnosis of metabolic disorders\n\nPLOS ONE\n\nDear Dr. Suliga,\n\nThank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE's publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.\n\nPlease submit your revised manuscript by Jul 16 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at <plosone@plos.org>. When you\\'re ready to submit your revision, log on to <https://www.editorialmanager.com/pone/> and select the \\'Submissions Needing Revision\\' folder to locate your manuscript file.\n\nPlease include the following items when submitting your revised manuscript:\n\nA rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled \\'Response to Reviewers\\'.A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled \\'Revised Manuscript with Track Changes\\'.An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled \\'Manuscript\\'.\n\nIf you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.\n\nIf applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:\u00a0<http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols>\n\nWe look forward to receiving your revised manuscript.\n\nKind regards,\n\nKiyoshi Sanada, PhD\n\nAcademic Editor\n\nPLOS ONE\n\n\\[Note: HTML markup is below. Please do not edit.\\]\n\nReviewers\\' comments:\n\nReviewer\\'s Responses to Questions\n\n**Comments to the Author**\n\n1\\. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the \"Comments to the Author\" section, enter your conflict of interest statement in the \"Confidential to Editor\" section, and submit your \\\"Accept\\\" recommendation.\n\nReviewer \\#1: (No Response)\n\nReviewer \\#2: All comments have been addressed\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n2\\. Is the manuscript technically sound, and do the data support the conclusions?\n\nThe manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n3\\. Has the statistical analysis been performed appropriately and rigorously?\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n4\\. Have the authors made all data underlying the findings in their manuscript fully available?\n\nThe [PLOS Data policy](http://www.plosone.org/static/policies.action#sharing) requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data---e.g. participant privacy or use of data from a third party---those must be specified.\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n5\\. Is the manuscript presented in an intelligible fashion and written in standard English?\n\nPLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.\n\nReviewer \\#1: Yes\n\nReviewer \\#2: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n6\\. Review Comments to the Author\n\nPlease use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)\n\nReviewer \\#1: This manuscript has been properly modified. But I have some minor comments.\n\n1\\) In response to previous review, the authors described \\\" The base population of this study was 13,172. From this group, 844 individuals (6.4%) were rejected due to a lack of complete anthropometric measurements and/or biochemical parameters that were necessary to perform this analysis\\\". This information should be added to Study Design and Participants.\n\n2\\) Page 8 Line 210: \"norm\". \"normal\"? Please check.\n\n3\\) The authors described \"The largest AUC for TGs was for the WC (0.642) and the %BF (0.641)\" in Page 9 Line 227. I think the position of this sentence should be changed. The position before \\\" The CUN-BAE, BMI, and WC also had high discriminatory power for at least one MetS component, with an AUC of 0.734, 0.728, and 0.728, respectively\\\" is better.\n\n4\\) The authors described \"In both men and women, the largest AUCs were for the WHTtR (0.624 and 0.632, respectively) and BRI (0.622 and 0.634, respectively)\" in Page 10 Line 240.\n\nOnly this result (WHtR) was explained for both men and women. Please describe in each paragraph.\n\nIn addition, please modify \"WHTtR\".\n\n5\\) In the text, the ABSI optimal cut-off point for men is 0.079 (Page 11 Line 255). But it is 0.080 in the Table3. Which is correct? Please check.\n\n6\\) The description of WHtR value is not unified.\n\nText: For other indices, optimal cut-off points for the identification of single metabolic disorders in men include WHtR=0.549... etc.\n\nTable3: 56.36 (BP) ... etc.\n\nTable4: 0.536 (BP) ... etc.\n\nPlease unify.\n\nReviewer \\#2: (No Response)\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n7\\. PLOS authors have the option to publish the peer review history of their article ([what does this mean?](https://journals.plos.org/plosone/s/editorial-and-peer-review-process#loc-peer-review-history)). If published, this will include your full peer review and any attached files.\n\nIf you choose \"no\", your identity will remain anonymous but your review may still be made public.\n\n**Do you want your identity to be public for this peer review?** For information about this choice, including consent withdrawal, please see our [Privacy Policy](https://www.plos.org/privacy-policy).\n\nReviewer \\#1: No\n\nReviewer \\#2: No\n\n\\[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link \\\"View Attachments\\\". If this link does not appear, there are no attachment files.\\]\n\nWhile revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool,\u00a0<https://pacev2.apexcovantage.com/>. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at\u00a0<figures@plos.org>. Please note that Supporting Information files do not need this step.\n\n10.1371/journal.pone.0235121.r004\n\nAuthor response to Decision Letter 1\n\n8 Jun 2020\n\nResponse to Reviewers\n\n1\\) In response to previous review, the authors described \\\" The base population of this study was 13,172. From this group, 844 individuals (6.4%) were rejected due to a lack of complete anthropometric measurements and/or biochemical parameters that were necessary to perform this analysis\\\". This information should be added to Study Design and Participants.\n\nThis information has been added as suggested by the reviewer.\n\n2\\) Page 8 Line 210: \"norm\". \"normal\"? Please check.\n\nThe sentence has been corrected.\n\n3\\) The authors described \"The largest AUC for TGs was for the WC (0.642) and the %BF (0.641)\" in Page 9 Line 227. I think the position of this sentence should be changed. The position before \\\" The CUN-BAE, BMI, and WC also had high discriminatory power for at least one MetS component, with an AUC of 0.734, 0.728, and 0.728, respectively\\\" is better.\n\nThe position of this sentence is changed.\n\n4\\) The authors described \"In both men and women, the largest AUCs were for the WHTtR (0.624 and 0.632, respectively) and BRI (0.622 and 0.634, respectively)\" in Page 10 Line 240.\n\nOnly this result (WHtR) was explained for both men and women. Please describe in each paragraph.\n\nIn addition, please modify \"WHTtR\".\n\nWhile editing the text, information about the HDL-cholesterol was mistakenly removed from the above sentence. That is why the meaning changed. It has been corrected.\n\n\"WHTtR\" is modified.\n\n5\\) In the text, the ABSI optimal cut-off point for men is 0.079 (Page 11 Line 255). But it is 0.080 in the Table3. Which is correct? Please check.\n\nThe ABSI optimal cut-off point for men = 0.080 is correct.\n\n6\\) The description of WHtR value is not unified.\n\nText: For other indices, optimal cut-off points for the identification of single metabolic disorders in men include WHtR=0.549... etc.\n\nTable3: 56.36 (BP) ... etc.\n\nTable4: 0.536 (BP) ... etc.\n\nPlease unify.\n\nThe description of WHtR value is unified.\n\n10.1371/journal.pone.0235121.r005\n\nDecision Letter 2\n\nSanada\n\nKiyoshi\n\nAcademic Editor\n\n\u00a9 2020 Kiyoshi Sanada\n\n2020\n\nKiyoshi Sanada\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n10 Jun 2020\n\nAnthropometric indices and cut-off points in the diagnosis of metabolic disorders\n\nPONE-D-20-07763R2\n\nDear Dr. Suliga,\n\nWe're pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.\n\nWithin one week, you'll receive an e-mail detailing the required amendments. When these have been addressed, you'll receive a formal acceptance letter and your manuscript will be scheduled for publication.\n\nAn invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at <http://www.editorialmanager.com/pone/>, click the \\'Update My Information\\' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at <authorbilling@plos.org>.\n\nIf your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they'll be preparing press materials, please inform our press team as soon as possible \\-- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact <onepress@plos.org>.\n\nKind regards,\n\nKiyoshi Sanada, PhD\n\nAcademic Editor\n\nPLOS ONE\n\nAdditional Editor Comments (optional):\n\nReviewers\\' comments:\n\n10.1371/journal.pone.0235121.r006\n\nAcceptance letter\n\nSanada\n\nKiyoshi\n\nAcademic Editor\n\n\u00a9 2020 Kiyoshi Sanada\n\n2020\n\nKiyoshi Sanada\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n11 Jun 2020\n\nPONE-D-20-07763R2\n\nAnthropometric indices and cut-off points in the diagnosis of metabolic disorders\n\nDear Dr. Suliga:\n\nI\\'m pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.\n\nIf your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they\\'ll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact <onepress@plos.org>.\n\nIf we can help with anything else, please email us at <plosone@plos.org>.\n\nThank you for submitting your work to PLOS ONE and supporting open access.\n\nKind regards,\n\nPLOS ONE Editorial Office Staff\n\non behalf of\n\nDr. Kiyoshi Sanada\n\nAcademic Editor\n\nPLOS ONE\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "\n"}
+{"text": "Introduction {#Sec1}\n============\n\nElectrochromic (EC) materials have attracted much attention owing to their potential applications in smart windows, antiglare mirrors, data storage devices, displays, sunroofs, and sunglasses. Various materials can be used as an EC layer, including inorganic metal oxides and organic conducting polymers^[@CR1]--[@CR8]^.\n\nConducting polymers provide benefits such as multiple colors, a fast switching time, and flexibility, but their disadvantages, including relatively nonuniform films, low material stability, and a limited range of colors severely limit their practical applications^[@CR9],[@CR10]^. On the other hand, tungsten trioxide (WO~3~) is a well-known metal oxide owing to its excellent EC performance. WO~3~ with different structures has been prepared by techniques such as hydrothermal process, chemical vapor deposition, thermal evaporation, and sputtering^[@CR11]--[@CR15]^. However, these approaches have drawbacks that restrict the commercial application of WO~3~ EC films, including complicated preparation, high energy consumption, expensive equipment, or the use of toxic and dangerous reagents^[@CR16]^.\n\nMolybdenum oxide is one of the important semiconducting metal oxides and can be used in various applications, including photovoltaic cells, organic light-emitting diodes, gas sensors, hydrogen evolution systems, transistors, and EC devices^[@CR17]--[@CR22]^. However, the coloration efficiency of pure molybdenum oxide used in EC devices is not high. For example, Patil *et al*. found that MoO~3~ used as an EC layer had a coloration efficiency of 34\u2009cm^2^/C^[@CR23]^. A combination of tungsten and molybdenum oxide has been used in EC devices recently. For instance, Mahdavi *et al*. investigated the effect of molybdenum in a WO~3~ thin film prepared by RF magnetron sputtering and obtained a coloration efficiency of 42.5\u2009cm^2^/C^[@CR24]^. Kharade *et al*. synthesized MoO~3~ mixed with WO~3~ using a hybrid physicochemical method and achieved a high coloration efficiency of 121.56\u2009cm^2^/C^[@CR25]^. However, the complicated synthesis method with high cost is a drawback. Consequently, it is urgently necessary to develop immediate, effective, and facile methods to synthesize tungsten oxide films with enhanced EC performance.\n\nIn this study, we report a facile, low-cost method of producing EC thin films based on WO~3~ for smart window applications. In our previous work, we investigated the (NH~4~)~2~WS~4~ precursor annealed at 350\u2009\u00b0C as a hole transport layer in an organic solar cell^[@CR26]^. In this work, we investigate the use of different annealing temperatures to obtain WO~3~, and then we add the (NH~4~)~2~MoS~4~ precursor as a dopant to the (NH~4~)~2~WS~4~ precursor at different concentrations (10, 20, 30, and 40\u2009mM) to obtain optimal EC films. A spin-coating method with an annealing process was applied to obtain EC films with excellent features such as high EC energy efficiency, high coloration efficiency, low cost, excellent chemical stability, fast switching speed, and good adhesion to the substrate.\n\nResults and Discussion {#Sec2}\n======================\n\nThe Raman spectra of the (NH~4~)~2~WS~4~ films annealed at different temperatures are shown in Fig.\u00a0[1(a)](#Fig1){ref-type=\"fig\"}. The Raman peak of the S--W--S stretching mode is weakened for the amorphous phase as the temperature is increased, whereas the peaks corresponding to WO~3~ phases (O--W--O bending and stretching) are strengthened^[@CR27]^. Therefore, the WO~3~ phases are completely decomposed after annealing at 500\u2009\u00b0C, indicating a monoclinic crystal system. Figure\u00a0[1(b)](#Fig1){ref-type=\"fig\"} shows the Raman spectra of the MoO~3~-doped WO~3~ film. The peak at 675\u2009cm^\u22121^ is ascribed to the coordinated oxygen in Mo crystal structure and stretching mode, which confirms that (NH~4~)~2~MoS~4~ was transformed to a MoO~3~ crystal, as indicated by the edge-shared oxygen^[@CR28]^. Moreover, the Raman peaks at 272\u2009cm^\u22121^ is assigned to O=Mo=O wagging modes^[@CR28]^. Fig.\u00a0[1(c)](#Fig1){ref-type=\"fig\"} shows the XRD patterns of (NH~4~)~2~WS~4~ thin films annealed at different temperatures. The pristine WS~4~ shows broad WS~2~ peaks (2\u03b8\u2009=\u200915--35\u00b0) related to weak crystallinity at annealing temperatures below 400\u2009\u00b0C. The broad peak intensity decreases after annealing at 500\u2009\u00b0C, and the peaks corresponding to WO~3~ structure are observed as well. These peaks are quite similar to those reported for monoclinic WO~3~ ^[@CR29]^. Fig.\u00a0[1(d)](#Fig1){ref-type=\"fig\"} confirms the nanocrystallinity of the MoO~3~-doped WO~3~ films. The molybdenum is incorporated into the film, producing a new phase with orthorhombic crystal structure which is similar to the previously reported structures^[@CR24],[@CR30]^. Moreover, the XRD results of the MoO~3~-doped WO~3~ did not show the peaks related to metallic Mo or MoO~3~, suggesting the well-diffusion of molybdenum atoms into WO~3~ crystal structure and substitution of Mo in W sites^[@CR24]^.Figure 1(**a**) Raman spectra of (NH~4~)~2~WS~4~ films annealed at different temperatures, (**b**) Raman spectra of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C, (**c**) XRD pattern of (NH~4~)~2~WS~4~ film annealed at different temperatures, (**d**) XRD pattern of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C.\n\nThe XPS spectra of (NH~4~)~2~WS~4~ films annealed at different temperatures are shown in Fig.\u00a0[2(a)](#Fig2){ref-type=\"fig\"}. As the temperature increases to 500\u2009\u00b0C, the S 2\u2009s and N 1\u2009s peaks related to the (NH~4~)~2~WS~4~ precursor disappear, suggesting the complete decomposition of the (NH~4~)~2~WS~4~ precursor into WO~3~. WO~3~ appears owing to the presence of O~2~ in the air^[@CR26]^. High-resolution views of the W 4\u2009f and O 1\u2009s peaks are shown in Fig.\u00a0[2(b)](#Fig2){ref-type=\"fig\"}. The peak appears at 33.5\u2009eV for the (NH~4~)~2~WS~4~ film annealed at 200\u2009\u00b0C, and the peaks at 35.4, 36.3, and 38\u2009eV are ascribed to W^4+^, W^5+^, and W^6+^\u2009 ^[@CR26]^. The peak of W^5+^ is related to oxygen vacancy^[@CR31]^. As the annealing temperature increases, the peak located at 33.5\u2009eV vanish, confirming the transformation of WS~2~ to WO~3~. The observed shifts of the peaks in the O 1\u2009s and W 4\u2009f spectra toward lower binding energy can be attributed to the emission of photoelectrons from the higher to lower oxidation states of W^[@CR32]^. For the XPS data in the O 1\u2009s region, the peak density related to the oxide phase near 531\u2009eV increases as the annealing temperature increases beyond 300\u2009\u00b0C (Fig.\u00a0[2(b)](#Fig2){ref-type=\"fig\"})^[@CR33]^. The XPS survey scan of the MoO~3~-doped WO~3~ film (Fig.\u00a0[2(c)](#Fig2){ref-type=\"fig\"}) shows additional peaks that are related to Mo 3d and Mo 3p. The XPS peak positions of Mo 3d~3/2~ and Mo 3d~1/2~ are 233 and 236.2\u2009eV, respectively (inset of Fig.\u00a0[2(c)](#Fig2){ref-type=\"fig\"}), which are attributed to pair of orbital spinning of MoO~3~ ^[@CR34]^. The two observed peaks of W 4f~7/2~ and W 4f~5/2~ appear at 35.9 and 38\u2009eV, respectively (Fig.\u00a0[2(d)](#Fig2){ref-type=\"fig\"}). Figure\u00a0[2(e)](#Fig2){ref-type=\"fig\"} shows the high-resolution XPS O 1\u2009s spectrum, in which the oxygen O 1\u2009s peaks are observed at 530.5 and 539.2\u2009eV. The spectra indicate the presence of W, Mo, and O in the as-prepared MoO~3~-doped WO~3~ EC film with oxidation states of +6, +6, and \u22122, respectively^[@CR25]^. All of XPS results and observed peaks are confirmed by previously reported works^[@CR25]--[@CR35]^.Figure 2(**a**) XPS survey spectra of (NH~4~)~2~WS~4~ films annealed at different temperatures, (**b**) high-resolution W 4\u2009f and O 1\u2009s spectra of (NH~4~)~2~WS~4~ films annealed at different temperatures, (**c**) XPS survey spectra of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C with high-resolution Mo 3d spectrum (inset), (**d**) W 4\u2009f and (**e**) O 1\u2009s spectra of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C.\n\nThe atomic ratios of (NH~4~)~2~WS~4~ films annealed at different temperatures are shown in Fig.\u00a0[3(a)](#Fig3){ref-type=\"fig\"}. Those of sulfur (S 2\u2009s) and nitrogen (N 1\u2009s) approach zero as the annealing temperature is increased to 400\u2009\u00b0C. The pie chart in Fig.\u00a0[3(b)](#Fig3){ref-type=\"fig\"} indicates that the oxygen content of the MoO~3~-doped WO~3~ is higher than that of other atoms owing to the annealing process and the presence of oxygen in both the MoO~3~ and WO~3~ structures. It is calculated that the level of doping was 0.76% (30\u2009mM (NH~4~)~2~MoS~4~ into (NH~4~)~2~WS~4~ precursor). These results not only support the formation of an oxide surface layer on the sulfide (NH~4~)~2~WS~4~ backbone, but also represent the functionalization of MoO~3~ on the WO~3~ structure. As shown in the inset of Fig.\u00a0[3(c)](#Fig3){ref-type=\"fig\"}, the work function of the pure WO~3~ thin film increases as the annealing temperature increased to 500\u2009\u00b0C. The work function of WO~3~ at 500\u2009\u00b0C is 4.71\u2009eV (inset of Fig.\u00a0[3(c)](#Fig3){ref-type=\"fig\"}). The increase of work function is attributed to the formation of WO~3~ ^[@CR26]^. In addition, the green curve shown in Fig.\u00a0[3(c)](#Fig3){ref-type=\"fig\"} indicates that the work function of MoO~3~-doped WO~3~ at 500\u2009\u00b0C is 5.02\u2009eV, which is higher than that pure WO~3~. These data suggest better performance of electrochromic in MoO~3~-doped WO~3~ device by facilitating charge transfer. Figure\u00a0[3(d)](#Fig3){ref-type=\"fig\"} shows the valance band maxima (VBMs) of the WO~3~ and MoO~3~-doped WO~3~, which decrease with increasing annealing temperature. However, MoO~3~ doping method increases the VBM owing to changes in the O 1\u2009s states and promoting the transition of intervalance within metal ions^[@CR34],[@CR36]^. These data suggest the enhanced transition of electron in MoO~3~-doped WO~3~ film, resulting in increased coloration efficiency.Figure 3(**a**) Atomic ratios of (NH~4~)~2~WS~4~ films annealed at different temperatures, (**b**) atomic content of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C, (**c**) UPS spectra of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C and (inset) work function of (NH~4~)~2~WS~4~ films annealed at different temperatures, (**d**) ultraviolet photoelectron spectra of (NH~4~)~2~WS~4~ films annealed at different temperatures and (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C.\n\nFigure\u00a0[4](#Fig4){ref-type=\"fig\"} shows FESEM and HRTEM images of the WO~3~ and MoO~3~-doped WO~3~. Both samples have porous and compact surfaces. The observed cracks are attributed to the annealing process (Fig.\u00a0[4(a) and (b)](#Fig4){ref-type=\"fig\"}). Moreover, the MoO~3~-doped WO~3~ exhibits a more uniform than pure WO~3~ (Fig.\u00a0[4(c)](#Fig4){ref-type=\"fig\"}). In addition, porosity measurement was carried out by using MATLAB software whose method was previously reported^[@CR37]^. In order to measure the porosity, the FESEM images of WO~3~ and MoO~3~-doped WO~3~ were converted to binary image (see Figure\u00a0[S1](#MOESM1){ref-type=\"media\"}) and then the percentage of porosity was calculated by following formula:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$P=(1-\\frac{{\\rm{n}}}{{\\rm{N}}})\\times 100$$\\end{document}$$where P is the porosity percent, n is the number of white pixels, and N is the total number of white and black pixels. The percentage of porosities were obtained 58.2 and 75.6% in WO~3~ and MoO~3~-doped WO~3~ film, respectively. The lattice fringes in the HRTEM images demonstrate that the WO~3~ films are crystallized, confirming the XRD results. In addition, the estimated lattice spacings of 0.302 and 0.364\u2009nm are assigned to the monoclinic d-spacing of the (020) plane of WO~3~ and the orthorhombic (002) plane of MoO~3~ (Fig.\u00a0[4(d) and (e)](#Fig4){ref-type=\"fig\"}). Figure\u00a0[4(f)](#Fig4){ref-type=\"fig\"} shows the elemental distributions of W, O, and Mo for the MoO~3~-doped WO~3~ thin film. MoO~3~ atoms not only grew on the WO~3~ surface, but are also diffused into the film. These data indicate that the method has great potential for efficient doping.Figure 4FESEM images of (**a**) (NH~4~)~2~WS~4~ film and (**b**) (NH~4~)~2~WS~4~/(NH4)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C, (**c**) TEM image of MoO~3~-doped WO~3~, (**d**) HRTEM image of MoO~3~-doped WO~3~, (**e**) electron diffraction and (**f**) STEM images and the corresponding STEM--EDX elemental maps of MoO~3~-doped WO~3~.\n\nAccording to the AFM images (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}), the measured roughness of the WO~3~ and MoO~3~-doped WO~3~ is 6.5 and 2.4\u2009nm, respectively. Therefore, the uniformity of film is increased without aggregation caused by doping. The AFM images confirm that the porosity of the film is greater after doping. The higher porosity is expected to improve the EC performance by enhancing the diffusion constant of the intercalating ions through the pore interface^[@CR38]^. In addition, higher roughness and porosity of film can improve the distribution of electrical field during redox process resulting in enhanced electron transfer and ion-insertion, leading to high EC performance^[@CR24],[@CR34],[@CR39],[@CR40]^.Figure 5(**a**) 2D and (**b**) 3D AFM images of (NH~4~)~2~WS~4~ film annealed at 500\u2009\u00b0C, (**c**) 2D and (**d**) 3D AFM images of (NH~4~)~2~WS~4~/(NH~4~)~2~MoS~4~ (30\u2009mM) film annealed at 500\u2009\u00b0C.\n\nFigure\u00a0[6(a)](#Fig6){ref-type=\"fig\"} shows the configuration of the EC cell. The transmittance spectra (wavelength 400--900\u2009nm) of WO~3~ and MoO~3~-doped WO~3~ in the colored and bleached states were measured (Fig.\u00a0[6(b)](#Fig6){ref-type=\"fig\"}). The color was changed to dark blue when a DC voltage of \u22122.5\u2009V was applied across the ITO. After the voltage was changed to +2.5\u2009V, the EC cell returned to the transparent state. The mechanism is thought to be the oxidation and reduction process. Li^+^ ions are inserted into the EC film, leading to reduction of W^6+^ to W^5+^ and the increase in the cathodic current change the color of the film. The change from the colored state to the transparent state is ascribed to oxidation (W^5+^ to W^6+^) due to the changing in redox state of the tungsten ions and the number of electrons (charge) inserted into EC film^[@CR24]^. In the doped EC film, Mo is involved in the reduction/oxidation process (Mo^6+^ to Mo^5+^ and Mo^5+^ to W^6+^), which causes to the enhanced transition of intervalency and electron transition within ions^[@CR34]^. Optical modulation is one of the most important parameters in EC devices and can be defined as \u0394T\u2009=\u2009T~b~ \u2212 T~c~, where T~b~ and T~c~ are the transmittance in the bleached and colored states, respectively, at a particular wavelength^[@CR41]^. The difference in transmission (\u0394T at 675\u2009nm) between the bleached and colored states in WO~3~ and MoO~3~-doped WO~3~ was enhanced from \u0394T~1~\u2009=\u200935% to \u0394T~2~\u2009=\u200949%. This improvement is attributed to a change in the crystal structure of WO~3~ caused by substitution of Mo at W sites and charge transfer between the Mo^5+^ and W^6+^ sites^[@CR24],[@CR42]^. Fig.\u00a0[6(c)](#Fig6){ref-type=\"fig\"} shows the stability of the WO~3~ and MoO~3~-doped WO~3~ in the colored state for several weeks after the voltage is removed. MoO~3~-doped WO~3~ exhibited better memory behaviour in air, in which it relatively retained the colored state very well even after 4 weeks. On the other hand, some parts of the pure WO~3~ EC film became transparent as time passed. Therefore, the memory behaviour of MoO~3~-doped WO~3~ is better than that of pure WO~3~ because of the increase in the diffusion coefficient (D) of Li^+^ ions in MoO~3~-doped WO~3~ during the intercalation process^[@CR24]^. Diffusion coefficient during intercalation process can be calculated by Randles\u2212Sevcik equation^[@CR24],[@CR43]^:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${{\\rm{D}}}^{1/2}={{\\rm{i}}}_{{\\rm{p}}}/[(2.72\\times {10}^{5}){{\\rm{n}}}^{3/2}{{\\rm{AC}}}_{{\\rm{0}}}{{\\rm{\\pi }}r}^{{\\rm{2}}}{{\\rm{v}}}^{1/2}]$$\\end{document}$$where i~p~ is the anodic peak current density, n is the number of electrons transferred during redox process, C~0~ is the concentration of active ions in the electrolyte, \u03bd is the scan rate, and A is the area of the EC film^[@CR24]^. The diffusion coefficient was obtained 1.23\u2009\u00d7\u200910^\u221211^ and 9.42\u2009\u00d7\u200910^\u221211^ for the WO~3~ and MoO~3~-doped WO~3~ EC film, respectively. Figure\u00a0[6(d) and (e)](#Fig6){ref-type=\"fig\"} shows the current-voltage (CV) curves of pure WO~3~ and MoO~3~-doped WO~3~ thin films, which were measured in a 1\u2009M aqueous solution at a scan rate of 50\u2009mV/s. In the cathodic process, the current of pure WO~3~ is higher than that of the MoO~3~-doped WO~3~ thin film. The MoO~3~-doped WO~3~ has a higher conductivity than pure WO~3~ because more defect states are created owing to integration of the two metal oxides, decreasing the energy required to extract the intercalated Li^+^ ions after MoO~3~ doping^[@CR24],[@CR25]^. In addition, the cycling stability of both thin films after 300 cycle steps revealed that the current in the MoO~3~-doped WO~3~ did not change and remained constant compared to that of the pure WO~3~ EC film. Moreover, the CV curves not only indicate a well-crystallized WO~3~ structure in both films, but also confirm the XRD and TEM results.Figure 6(**a**) Structure of as-prepared EC cell, (**b**) transmission spectra of WO~3~ and MoO~3~-doped WO~3~, (**c**) memory behavior of WO~3~ and MoO~3~-doped WO~3~ films during 8 weeks, CV curves of (**d**) WO~3~ and (**e**) MoO~3~-doped WO~3~ EC films and (**f**) coloration efficiency of the different EC films.\n\nThe coloration efficiency (CE), which is an important parameter for EC devices, was calculated as follows^[@CR23],[@CR24],[@CR30]^:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\rm{CE}}=\\frac{{\\rm{\\Delta }}\\mathrm{OD}}{{\\rm{Q}}}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\rm{\\Delta }}\\mathrm{OD}=\\,\\mathrm{log}[\\frac{{{\\rm{T}}}_{{\\rm{b}}}}{{{\\rm{T}}}_{{\\rm{c}}}}]$$\\end{document}$$where \u0394OD is the change in optical density, Q is the charge density, T~b~ is the transmittance of the film in the bleached state, and T~c~ is the transmittance of the film in the colored state^[@CR23],[@CR24],[@CR41]^. The color efficiency of various EC layers and various MoO~3~ concentrations (\u03bb\u2009=\u2009675\u2009nm) is presented in Fig.\u00a0[6(f)](#Fig6){ref-type=\"fig\"}. The CE values of the MoO~3~-doped WO~3~ are higher than that of pure WO~3~ (74.5\u2009cm^2^/C). To determine the optimal concentration of MoO~3~ in WO~3~, the CE was measured for different MoO~3~ concentrations (10, 20, 30, and 40\u2009mM) in WO~3~. As shown in Fig.\u00a0[6(f)](#Fig6){ref-type=\"fig\"}, a much higher CE (128.1\u2009cm^2^/C) was obtained for the MoO~3~-doped WO~3~ with a MoO~3~ concentration of 30\u2009mM. This result indicates that the optimal amount of molybdenum has a crucial role in obtaining high efficiency. Furthermore, the response times of the colored and bleached states for MoO~3~-doped WO~3~ are found to be 3.6 and 4.5\u2009s, respectively, whereas the response times of the colored and bleached states for pure as-prepared WO~3~ are 8 and 9.5\u2009s, respectively. Enhanced EC properties is caused by improvement in extra electron intervalance transfer W^6+^ and Mo^6+^ active sites. In addition, molybdenum after intercalation of Li^+^ ions are more near to the sensitivity of human's vision. The disorder could be increased by random distribution of molybdenum, resulting in the betterment EC properties^[@CR25]^. Therefore, the performance was improved by MoO~3~ doping. Table\u00a0[1](#Tab1){ref-type=\"table\"} compares our results with the reported values through various materials and methods.Table 1Comparison of our work with previously published papers.MaterialMethodSwitching time (t~c~/t~b~)\u0394T (%)Coloration efficiency (cm^2^/C)Ref.Mo-doped WO~3~RF magnetron sputtering---44.342.5[@CR24]MoO~3~/WO~3~Hybrid physicochemical synthesis4.1\u2009s/3.4\u2009s\\~50121.56[@CR25]Nanoparticulate WO~3~Electrodeposition3.7\u2009s/5.2\u2009s88.51137[@CR35]NiO/WO~3~DC magnetron sputtering10\u2009s/20\u2009s5587[@CR40]Ti-doped WO~3~Sol--gel spin-coating---47.5---[@CR44]PANI/WO~3~Electropolymerization9.9\u2009s/13.6\u2009s37.498.4[@CR45]WO~x~ nanorodsLow-temperature ozone exposure11.8\u2009s/20.1\u2009s5733.3[@CR46]WONWS-RGOSolvothermal1.5\u2009s/1.2\u2009s---116.7[@CR47]MoO~3~-doped WO~3~Solution and annealing process3.6\u2009s/4.5\u2009s49128.1Our work\n\nConclusions {#Sec3}\n===========\n\nWO~3~ doped with MoO~3~ was prepared by a facile and low-cost method involving solution and annealing processes. The results indicated that the (NH~4~)~2~MoS~4~/(NH~4~)~2~WS~4~ precursor decomposed to MoO~3~-doped WO~3~ when the film was annealed at 500\u2009\u00b0C in air. The N 1\u2009s and S 2\u2009s emission in the XPS spectrum of the (NH~4~)~2~WS~4~ precursor annealed at 500\u2009\u00b0C, as well as weakening of the S--W--S bond, increased O--W--O bond peaks in the Raman spectra, and the appearance of peaks in the XRD spectra, suggested full decomposition to monoclinic crystalline WO~3~. In addition, (NH~4~)~2~MoS~4~ was added to the (NH~4~)~2~WS~4~ precursor at various concentrations as a dopant, and the resulting films were then annealed at 500\u2009\u00b0C to transform them to MoO~3~-doped WO~3~. The XRD and Raman spectroscopy results confirmed the decomposition to nanocrystalline MoO~3~-doped WO~3~. Moreover, the morphology of the as-prepared films was observed using FESEM and AFM, which showed that the MoO~3~-doped WO~3~ was more uniform and porous than pure WO~3~, suggesting high EC performance. Furthermore, the high doping capability with good distribution of MoO~3~ into WO~3~ was confirmed by HRTEM images. As a result, enhanced EC performance was obtained when the MoO~3~-doped (30\u2009mM) WO~3~ was used as an EC layer. The coloration efficiency was high (CE\u2009=\u2009128.1\u2009cm^2^/C), and the response time was rapid (t~c~\u2009=\u20093\u2009s, t~b~\u2009=\u20094.5\u2009s). These values are much higher than those of pure WO~3~ (CE\u2009=\u200974.5\u2009cm^2^/C, t~c~\u2009=\u20098\u2009s, t~b~\u2009=\u20099.5\u2009s). In conclusion, the MoO~3~-doped WO~3~ prepared by the annealing--solution process is a remarkable candidate for use in high-efficiency, low-cost smart windows that can be efficiently commercialized.\n\nMethod {#Sec4}\n======\n\nPreparation of thin film of MoO~3~-doped WO~3~ {#Sec5}\n----------------------------------------------\n\nFigure\u00a0[7](#Fig7){ref-type=\"fig\"} illustrates the synthesis of the WO~3~ and MoO~3~-doped WO~3~ thin films and fabrication of the EC cell. Indium tin oxide (ITO) substrates were ultrasonically cleaned sequentially with DI water, isopropanol, and acetone, and then dried; they were then treated by ultraviolet ozone for 20\u2009min and maintained there until the start of the spin-coating process. After drying, PtCl~4~ dispersed in isopropanol was coated on one piece of the ITO conductive glass by spin-coating, and the ITO was dried on a hotplate at 250\u2009\u00b0C to evaporate the solvent and chlorine; this sample was used as a counter electrode. Next, (NH~4~)~2~WS~4~ (200\u2009mg) was dissolved in 1\u2009ml of N,N-dimethylformamide, resulting in the formation of a yellowish tungsten sol. Then, a homogenous thin film was prepared by spin-coating the as-prepared solution onto the ITO substrate at 4000\u2009rpm for 60\u2009s. After drying in air, the coated substrates were annealed in a furnace at different temperatures (200, 300, 400, and 500\u2009\u00b0C) for 2\u2009h. A transparent, colorless WO~3~ thin film was obtained at 500\u2009\u00b0C. To form the MoO~3~-doped WO~3~ thin films, (NH~4~)~2~MoS~4~ was added to the (NH~4~)~2~WS~4~ solution separately (at different concentrations, 10, 20, 30, and 40\u2009mM, to determine the optimal amount of doping) to form a homogenous solution. Then, this procedure was repeated to synthesize MoO~3~-doped WO~3~ thin films at 500\u2009\u00b0C.Figure 7Synthesis of EC materials and fabrication of EC cell.\n\nFabrication of the EC cell {#Sec6}\n--------------------------\n\nThe EC device structure for the MoO~3~-doped WO~3~ thin films was glass/ITO/MoO~3~-doped WO~3~/LiClO~4~\u2009+\u2009propylene carbonate (PC)/Pt/ITO/glass. The ITO substrate coated with the MoO~3~-doped WO~3~ thin film acts as a working electrode, and the Pt/ITO-coated conducting glass substrate acts as a counter electrode; the electrodes are assembled to fabricate a sandwich-type EC device. The liquid electrolyte, 1\u2009M lithium perchlorate (LiClO~4~)/PC, was injected into the device through a small hole, which was then sealed with Resibond epoxy glue.\n\nCharacterizations {#Sec7}\n-----------------\n\nX-ray photoelectron spectroscopy (XPS) was performed using an ESCA-3000 (VG Scientific Ltd., England) instrument analyzer under a vacuum better than 1\u2009\u00d7\u200910^\u22125^ mbar using Mg K\u03b1 radiation (1250\u2009eV) and a constant pass energy of 50\u2009eV. The composition of the thin film samples was determined by X-ray diffraction (XRD) analysis (Bruker AXS Model D8 Advance X-ray diffractometer) with a Cu K\u03b1 target having a wavelength 0.1542\u2009nm. Raman spectra (LabRAM HR, Horiba Jobin Yvon, Japan) were obtained at an excitation wavelength of 514\u2009nm. Field-emission scanning electron microscopy (FESEM, Zeiss 300 VP) images were taken at an acceleration voltage of 50\u2009kV. Transmission electron microscopy (TEM) was performed with a JEOL-2100F (Japan) instrument. Contact-mode atomic force microscopy (AFM, XE-100/PSIA) was used to determine the roughness and porosity of the thin films. Cyclic voltammetry (CV) and electrochemical measurements were performed in a quartz electrochemical cell connected to a potentiostat (Ivium 5612, Netherlands). WO~3~ or MoO~3~-doped WO~3~ was used as the working electrode, and a Ag/AgCl electrode and platinum (Pt) wire were applied as the reference and counter electrodes, respectively. The transmittance spectra were measured by a UV--vis spectrophotometer (V-670). The coloration/bleaching switching characteristics of the EC films were recorded as the changes in the transmittance at a wavelength of 675\u2009nm under alternating application of a potential of \u00b12.5\u2009V for 60\u2009s for each state. For the XPS, high-resolution TEM (HRTEM), FESEM, XRD, Raman, and AFM measurements of the MoO~3~-doped WO~3~, samples fabricated using 30\u2009mM of (NH~4~)~2~MoS~4~ in the (NH~4~)~2~WS~4~ precursor were used.\n\nElectronic supplementary material\n=================================\n\n {#Sec8}\n\nSupplementary information\n\nAmirhossein Hasani and Quyet Van Le contributed equally to this work.\n\n**Electronic supplementary material**\n\n**Supplementary information** accompanies this paper at 10.1038/s41598-017-13341-z.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThis research was supported in part by National Research Foundation of Korea (NRF) grants provided by the Korean government (MSIP) (Nos 2015K1A3A1A59073839, 2017H1D8A1030599, 2017K1A3A1A67014432) and in part by Korea Agency for Infrastructure Technology Advancement grant funded by Ministry of Land, Infrastructure and Transport (17IFIP-B133622-01).\n\nA.H. and Q.V.L. contributed equally to this work. A.H. and Q.V.L. performed the synthesized materials and most of characterization. T.P.N., K.S.C., and W.S. performed device characterization. J.K.K. analyzed materials' characterization. H.W.J. and S.Y.K. supervised the experiments and contributed to manuscript preparation. H.W.J. and S.Y.K. initiated and directed the research. S.Y.K. conceived the idea and designed the experiments. All authors were involved in writing the manuscript.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Background\n==========\n\nEffective typing of microorganisms is a prerequisite for establishing epidemiological or phylogenetic links between corresponding isolates. A plethora of different methods has been successfully applied to type and differentiate bacterial strains and clonal groups from each other \\[[@B1]\\]. A critical point to all of these methods is their applicability to answer distinct questions ranging from investigation of outbreaks to establishing rather broad phylogenetic trees of relatedness and arrangement of strains within major clonal complexes. Each method has its respective weaknesses and strengths according to the question(s) addressed and the methodology behind \\[[@B2]-[@B7]\\].\n\nRecently, a new method was introduced using small repetitive elements appearing in a variable number and distributed among the genome of a given species. Accordingly this technique based on a variable number of tandem repeats (VNTR) was named multiple-locus variable-number tandem repeat analysis (MLVA; \\[[@B8]\\]). Initially MLVA was established to differentiate high-risk pathogens such as *Bacillus anthracis*and *Francisella tularensis*\\[[@B9]-[@B11]\\] but has been extended to a numerous number of other bacterial species and scientific questions \\[[@B8],[@B12]\\] including outbreak investigations for pathogenic bacteria \\[[@B4],[@B13]\\].\n\nMLVA was applied also recently to type isolates of *Enterococcus*spp. Its discriminatory power was compared to MLST for a collection of 392 *E. faecium*\\[[@B14]\\] and to macrorestriction analysis in PFGE for 83 *E. faecalis*\\[[@B15]\\]. In both cases it was described that MLVA showed similar and rather concordant discrimination when compared to the respective reference method. Although the selected VNTRs were different between the two species, the overall conclusion would suggest MLVA as a typing method on one hand to discriminate highly enough between strains and on the other hand indicate the possibility to establish rather broad phylogenetic relatednesses. To support this hypotheses and to test the applicability of the established MLVA scheme for *E. faecium*to indicate and differentiate hospital-adapted clonal types appearing in increasing numbers among hospital patients worldwide \\[[@B16]\\], we investigated hospital isolates representing outbreaks and clusters of infections and colonizations from German hospital patients from the last 15 years using MLVA, *Sma*I-macrorestriction analysis in PFGE, and MLST. A collection of 27 *E. faecium*from commensal, animal, and environmental origins was included for reasons of comparison and also typed by MLVA.\n\nResults\n=======\n\nMacrorestriction analysis in PFGE\n---------------------------------\n\nAltogether 58 *E. faecium*were investigated by *Sma*I-macrorestriction analysis originating from 31 German hospitals. Thirty-eight different PFGE types were assigned based on a 90 % similarity cut-off and recommendations described recently \\[[@B17],[@B18]\\]. However, larger clusters of strains possessing at least related patterns (\\> 80 % identity) could be identified (\\[[@B18]\\]; see below and Fig. [1](#F1){ref-type=\"fig\"}).\n\nMLST\n----\n\nAltogether 19 MLST types were assigned. All but one hospital-adapted/outbreak isolates (n = 50/51) of groups II (1996--1999) and III (2004--2006) possessed MLST types (ST) which belong to MLST CC-17 or C1 representing epidemic and hospital-adapted clonal types (Table [1](#T1){ref-type=\"table\"}). Isolate UW6520 revealed ST-314 which belongs to MLST cluster C and not to cluster C1/CC17. Isolates of group I (1991--1995) were quite diverse; only three possessed a MLST type belonging to MLST CC17 (ST-17), the other four were attributed to three different clonal complexes.\n\nMLVA\n----\n\nAltogether 14 MLVA types were assigned. All 51 group II and III isolates possessed MLVA types associated with clonal complex C1 of hospital-adapted types (Table [1](#T1){ref-type=\"table\"}). Cluster assignments of new MLVA types were done by the curator of the MLVA internet pages (J. Top, University Utrecht, Netherlands) based on pairwise similarities, UPMGA clustering and similarities within a minimum spanning tree of relatedness. Four group I VRE belonged to MLVA clonal complex C1, the other three each one to A, B or C (Table [1](#T1){ref-type=\"table\"}, see also \\[[@B14]\\]). The 27 isolates of different non-hospital origins showed ten MLVA types. All but two belonged to MLVA clonal complexes A or B (J. Top, pers. communication). Main type was MT-89 (n = 9; CC-A). Six isolates were none-typeable due to an incapability of an amplification of one of the six VNTR fragments.\n\nConcordance of PFGE, MLVA and MLST results\n------------------------------------------\n\nAltogether 58 hospital VRE (groups I -- III) were typed by all three molecular typing methods. *Sma*I-macrorestriction in PFGE revealed 38, MLST 19, and MLVA 14 different types. PFGE was the most discriminatory method with a discriminatory index of 0.972, MLST the median (0.91) and MLVA the least discriminatory (0.842) (Table [2](#T2){ref-type=\"table\"}). Differences in discriminatory power between those different typing methods are even more obvious and pronounced when results involving only outbreak isolates of groups II and III are compared (PFGE: 0.965; MLST: 0.891; MLVA: 0.814). MLVA appears in all cases as the least discriminatory method which derives from the fact that (i) it revealed the least number of different types and in addition, (ii) more than half of all the investigated VRE belong to only two types MT-1 and MT-159.\n\nSome isolates with identical MLVA types showed different MLST types (Table [1](#T1){ref-type=\"table\"}; Fig. [2](#F2){ref-type=\"fig\"}). But also the opposite was found; isolates with identical MLST types could possess different MLVA types (Table [1](#T1){ref-type=\"table\"}; Fig. [2](#F2){ref-type=\"fig\"}). Isolates harboring the two major MLVA types, MT-1 and MT-159, combine isolates showing a various number of STs. E.g., MT-159 isolates combine *E. faecium*of ST-78, -192, -203, and -283. MT-1 isolates combine *E. faecium*of ST-17, ST-18, ST-280, and ST-282.\n\nResults of MLST partly confirm cluster assignments based on PFGE data, whereas our data did not show a visible concordance between MLVA and PFGE (Fig. [1](#F1){ref-type=\"fig\"}). Many MLVA types were distributed among different main PFGE clusters. Especially the most common types such as MT-1 (n = 16) and MT-159 (n = 14) were evenly distributed among different major PFGE branches. This fact is also reflected in the much lower concordance value for results of MLVA/PFGE (0.863) compared to MLST/PFGE (0.935; Table [2](#T2){ref-type=\"table\"}). Eight of nine ST-203 isolates cluster in a group of related PFGE patterns (\\> 83 % identity) together with two ST-282 VRE (ST-282 is a double-locus variant \\[DLV\\] of ST-203). A single ST-203 isolate clusters in a side branch but at least within a cluster of \\> 75 % identity. These eleven isolates possess three different MLVA types. All eleven ST-18 *E. faecium*cluster in a group with \\> 70 % identical patterns. However, also two ST-78 and a single ST-17 isolates belong to this cluster but are located at both edges of this branch. These altogether 14 VRE possess four MLVA types. In contrast to this, isolates of ST-17 appear distributed among different branches of the PFGE tree.\n\nThe eight group II isolates all possess an identical MLST (ST-117) and MLVA (MT-12) type and an identical or highly related PFGE pattern (\\> 95 % identity; only shown for five isolates). Isolate UW6605 (ST-78) revealed also MT-12 but is clearly different from group II isolates when compared on the basis of its PFGE pattern. However, ST-117 and ST-78 might be phylogenetically related since both MLST types only differ in a single locus.\n\nDiscussion\n==========\n\nMLVA was introduced to molecular bacterial strain typing as a promising alternative to established typing standard methods such as AFLP, MLST and partly PFGE or as an alternative for microorganisms where other common typing techniques could not be applied due to several reasons \\[[@B19]\\]. Therefore it was not possible in all cases to compare its discriminatory power to established standards which is expected to vary also from genus to genus. For nosocomial pathogens, such as *Staphylococcus aureus*, MLVA was introduced as an alternative typing method showing matching results with macrorestriction typing in PFGE and clusters generated by *spa-*typing and MLST \\[[@B4]\\]. For *E. faecium*discriminatory power of MLVA was compared to MLST and AFLP (Amplified fragment length polymorphism) and found to be reliably sufficient as a suitable typing alternative \\[[@B14]\\]. But already when comparing data from its original description in detail, some uncertainties emerged challenging the overall applicability of this method for typing in general including especially investigation of outbreaks or following routes of disseminated outbreak strains. For instance, MT-1 combined 58 isolates represented by eleven STs \\[[@B14]\\]. Since MT-1 is the primary founder and one of the major MLVA types representing the cluster of hospital-adapted, clonal types (MLVA C1) its applicability to investigate single outbreak situations could be questioned. Indeed, this was one of the results of our study: Isolates representing several MTs such as the most common MT-1 and MT-159 neither cluster based on their *Sma*I-macrorestriction patterns in PFGE nor possess all identical or related MLST types (Fig. [1](#F1){ref-type=\"fig\"} and [2](#F2){ref-type=\"fig\"}).\n\nAs shown by our results, discrimination of *E. faecium*based on results of MLVA, MLST and PFGE does not always give concordant results as it was shown for rather clonal populations such as *S. aureus*\\[[@B7]\\] or for an analysis of *E. faecium*within a single hospital over a period of five years \\[[@B20]\\]. Concerning the latter a more detailed comparison of the results of Abele-Horn et al., and ours may be required. These, at the first moment rather contradictory results reflect major differences in the used strain collection (strains from a single hospital vs. 31 hospitals) and correspondingly resulting MLVA and MLST type diversity. A collection of isolates from a single hospital may not be diverse enough to elucidate the non-congruencies we found here between some MLVA and MLST types (e.g., MT-159 absent in \\[[@B20]\\]). The discriminatory power of MLVA in strains recovered from patients at the University Hospital in general, i.e. not only associated with the described outbreaks, revealed a Simpson\\'s diversity index of 0.846 \\[[@B20]\\] and was almost identical to what we calculated for our study population (0.842).\n\nClonal types might be affected by genomic rearrangements to a different extent. Whereas ST-18 or ST-203 isolates showed rather similar *Sma*I-macrorestriction patterns over time and geographical distribution, isolates of ST-17 were rather diverse. The ST-17 isolates were distributed among several subclusters or represent single PFGE patterns (Fig. [1](#F1){ref-type=\"fig\"}). It could be speculated if isolates of ST-17 represent a rather ancient clonal type prone to recombination and thus divergent PFGE patterns for a longer time than ST-203 and ST-18 which could constitute rather recent hospital-adapted clones. Investigation of ST-17 from the early 1990ies as shown in our small collection of group I isolates and in a collection of bacteremia isolates from the mid to the late 1990ies (Werner, unpublished results) supports this hypothesis. The role of mobile and conjugative elements in rearranging bacterial genomes is well-known \\[[@B21]\\]. The genome sequence of *E. faecalis*V583 revealed one of the highest numbers of mobile, integrative and conjugative elements known so far from the bacterial world \\[[@B22]\\]. Similar numbers are suggested for *E. faecium*, however whole genome data for *E. faecium*are still incomplete. Recent comparative genomic hybridizations for a set of 97 *E. faecium*from different origins confirm the role of mobile and conjugative elements for shaping the *E. faecium*genome \\[[@B23]\\]. IS element driven recombination appeared higher in a subset of evolutionary linked epidemic *E. faecium*than in non-epidemic *E. faecium*(based on a comparison of typing results for MLST and PFGE). In this respect PFGE may be \\\"overdiscriminatory\\\" to a certain extent when applied to investigate so-called hospital clade or epidemic *E. faecium*\\[[@B23]\\]. Other factors and impacts that affect distinct clonal types in a different manner could also be discussed \\[[@B24],[@B25]\\].\n\nClearly, MLVA has its advantages when applied to type *E. faecium*. It is a suitable method to identify strains belonging to the complex of hospital-adapted, epidemic clones (MLVA CC-1). The method is quick (intra-day results), easy to perform, comparably cheap, with excellent reproducibility (intra- and inter-laboratory) and allows data transfer and comparison. The comparably long repeat lengths of the VNTR loci used to type *E. faecium*make the assay suitable for agarose gel separation and therefore capable for many standard laboratories. Introduction of fluorescence-labelled primers in future may lead to a combined approach of six single reactions into one or two multiplex PCRs \\[[@B26]\\].\n\nHowever, the advantage of comparably long repeat lengths may lead to an unintentional side-effect. MLVA was primarily established as a typing method to differentiate between strains. The VNTR repeats are, in general, very short and in a range of 6 to 30 bp \\[[@B11],[@B12],[@B27]\\]. The idea behind MLVA typing is mainly based on errors caused by the DNA polymerase resulting in slippage and/or recombination in a somehow expectable manner within these sequences. It could be speculated if repeat lengths of the sizes used to type isolates of *E. faecium*(120 -- 279 bp) may be too long to let these events appear in an essential frequency. So differences between strains would be less detected and the identified changes would only be suitable to establish phylogenetic links between strain types similar to or even less discriminatory than MLST. Indeed, some VNTR fragments showed length of incomplete repeats suggesting recombinational events within the distinct single repeat units. Assignment of VNTR repeat numbers might be rather erroneous in those cases. Analysis of tandem-repeat fragment length was recently reported to provide misleading results about the phylogeny and subspecies affiliation of *Francisella tularensis*isolates. This was found to be due to non-homologous, sequence-different VNTR fragments of identical lengths in unrelated isolates \\[[@B28]\\]. It should be emphasized again that the models predicting stability and flexibility of the distinct genomic repetitive structures used for VNTR analysis in *E. faecium*and other bacteria are rather theoretical and speculative. Nothing is really known about the conditions affecting stability of these structures. Due to this uncertainty, establishing phylogenetic relatedness in appropriate trees by a link via single-locus or double-locus variants of MLVA profiles which might suggest an evolutionary link should be interpreted with caution.\n\nConclusion\n==========\n\nAltogether 58 *E. faecium*mainly from clusters of infections and colonizations and outbreaks among patients from 31 German hospitals were typed by MLVA. Results were compared to results of PFGE and MLST which led to the following conclusions: (i) MLVA is a suitable method to identify isolates of epidemic-virulent *E. faecium*clonal lineages and differentiate them from isolates of other origins; (ii) for isolates of the clonal complex of hospital-adapted strains MLVA is the least discriminatory method when compared to MLST and PFGE and MLVA results are less concordant than MLST results when compared to results of PFGE; and thus (iii) MLVA appears not to be suitable to identify strain types of *E. faecium*in general, e.g. as necessary in cases of outbreaks. The data presented here for *E. faecium*clearly indicate that non-related isolates could possess an identical MLVA type. This would lead to false-positive results being especially critical when MLVA is used to elucidate supposed outbreaks with *E. faecium*within a single or among different hospitals.\n\nMethods\n=======\n\nBacterial strains\n-----------------\n\nMajority of isolates (n = 43) originated from recent outbreaks and clusters of infections and colonizations (CCI) in German hospital patients between 2004 and 2006 (group III). Each isolate represents a single case/patient. We stick to the term CCI since outbreak means cases of infections. Our sample collection also includes cases of colonizations (stool samples) and cases of infections where the role of *E. faecium*is not clear (pneumonia, UTI). According to common outbreak definitions, a cluster is represented by n \u2265 2 cases which were epidemiologically linked. Isolates UW5248-UW5258 (n = 6) originated from a large private laboratory service provider in Southwestern Germany with a representative coverage of hospitals in five federal states (Labor Limbach, Heidelberg, Germany). Six representative isolates were chosen from a larger set of isolates representing clusters in five hospitals \\[[@B29]\\]. The other 37 group III *E. faecium*originated from 14 other German hospitals. Additional eight isolates representing each one an older outbreak sampled all over Germany between 1996--1999 were also included (group II; \\[[@B30]\\]). All seven but one (UW786) older hospital isolates represent sporadic infections or colonizations in hospital patients between 1991--1995 (group I; \\[[@B31]\\]). All hospital *E. faecium*were vancomycin-resistant and possessed the *vanA*gene cluster. For reasons of comparison 27 *E. faecium*isolates from non-hospital origins as well as the *vanA*reference strain BM4147 were also included \\[[@B31],[@B32]\\].\n\nDNA isolation\n-------------\n\nDNA was isolated by a column-based technique as recommended by the manufacturer (DNeasy Tissue Kit, Qiagen, Hilden, Germany). Standard protocol was slightly modified starting with an initial cell wall lysis step of 1 ml overnight culture centrifuged and resuspended in 400 \u03bcl TES buffer (10 mM Tris, 1 mM EDTA, 10 % saccharose) plus 10 mg/ml lysozyme and incubated at 37\u00b0C for 30 min.\n\n*Macrorestriction analysis in PFGE*. Genomic DNA was isolated, digested with restriction endonuclease *Sma*I, and treated as described recently \\[[@B29]\\]. The agarose gel concentration was 1 %, the CHEF-DR III apparatus (BIO-RAD laboratories, Hercules, CA, USA) was used for PFGE. The ramped pulsed times were as follows: 1--11 sec for 15 h and 11--30 sec for 14 h at 14\u00b0C. Relatedness between banding patterns was calculated using a band based similarity coefficient (Dice) and UPMGA clustering (BioNumerics, Applied Maths, Sint-Martens-Latem, Belgium). A composite tree of all 58 hospital VRE of groups I -- III resolved in 18 independent PFGE gels was generated the same way with the exception of increasing the value of the \\\"position tolerance setting\\\" to 1.5 % (default: position tolerance setting 1.0 %; optimization 0.5 %). PFGE types were assigned based on \\>90 % similar patterns and additionally considering recommendations for fragment pattern analyses as described recently \\[[@B17],[@B18]\\].\n\nMLST\n----\n\nPCRs amplifying the seven loci used for MLST were done according to the reference \\[[@B33]\\]. Sequencing reactions were performed according to the manufacturer\\'s recommendations for cycle sequencing of PCR products (Applied Biosystems Germany, Darmstadt). Sequence files were read, evaluated, aligned, and compared to the reference set of alleles using sequencing software Lasergene 6.1. from DNA-STAR (SeqMan 6.1; EditSeq 6.1), TraceEditPro from Ridom \\[[@B34]\\], and via the official MLST webpage \\[[@B33]\\]. Assignment of new MLST types and allocation to their corresponding clonal complexes was done by the curator of the internet service (Dr. R. J. L. Willems, University Utrecht, The Netherlands).\n\nMLVA\n----\n\nMLVA VNTRs were amplified according to Top et al., \\[[@B14]\\] using modifications given at the corresponding website of the University of Utrecht, The Netherlands \\[[@B35]\\]. Pure and high-yield genomic DNA isolated via commercially available kits is especially essential for a successful amplification of VNTR-2. MLVA types were assigned using the corresponding tool at the webpage (see above). For unknown profiles a new MLVA type was assigned by the operator of the internet service (Mrs. J. Top, University Utrecht, The Netherlands).\n\nPhylogenetic analyses\n---------------------\n\nSoftware eBURST v3 was used to establish trees of phylogenetic relatedness between the isolates based on their MLST or MLVA profiles \\[[@B36]\\]. Significance of branching of the calculated trees was evaluated by bootstrap analysis of 1000 computer-generated trees. Clonal complexes are defined as described recently \\[[@B14],[@B16]\\]. Discriminatory indices for and level of congruence between results of MLST, MLVA and PFGE were calculated using Ridom software (Ridom Bioinformatics, Muenster, Germany) based on references given \\[[@B37]-[@B39]\\]. However, the corresponding values should not be interpreted as a stand-alone result. Calculating similarity indices and congruence between methods requires sample collections that are rather unlinked. Our study strain collection is admittedly biased since it involves only hospital *E. faecium*. So we recommend taking this into consideration and comparing those values only within our study strain collection.\n\nAbbreviations\n=============\n\nDLV -- double locus variant; MLST -- Multi-locus sequence typing; MLVA -- multiple-locus variable-number tandem repeat analysis; PFGE -- pulsed-field gel electrophoresis; VNTR -- variable number of tandem repeat.\n\nAuthors\\' contributions\n=======================\n\nGW performed the MLVA typing and cluster analysis, parts of the MLST sequencing and analyses, analyzed macrorestriction results and wrote the manuscript. IK supervised the strain collection including establishing the main data set (contributor\\'s data, resistance profile, standard PCRs, etc.), assorted and allocated isolates for PFGE analyses, and did parts of the MLST sequencing and analyses. WW supervised the study, established contacts to our main contributors of strains, supervised assortment of isolates and critically read the manuscript. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nWe would like to thank all persons, their institutions and microbial laboratories that provided the corresponding strains for our analyses. A special thanks to Dr. A.-M. Fahr and Dr. U. Eigner (Labor Limbach, Heidelberg, Germany) for providing a comprehensive collection of *E. faecium*from various hospitals. A special thanks to Dr. U. Nuebel and B. Strommenger, PhD, for critically reading the manuscript. The excellent scientific support by Mrs. Janetta Top and Dr. R. J. L. Willems, University Utrecht, for assigning new MLVA and MLST types, allocating MLVA types to their corresponding groups, operation of the MLVA webpages \\[[@B35]\\] and support for the MLST internet database \\[[@B33]\\] is highly appreciated. We thank B. Strommenger for sharing her expertise in using Ridom software. A special thanks to Mrs. Carola Konstabel for excellent technical support.\n\nFigures and Tables\n==================\n\n![**Cluster analysis of hospital VRE based on *Sma*I-macrorestriction patterns resolved in PFGE**(see also Tab. 1). The dendrogram was evaluated using Dice coefficient and UPMGA clustering (BioNumercis software). Legend: Groups -- group I isolates from 1991--1995, group II isolates from 1996--1999, group III isolates from 2004--2006; year, year of isolation; hospital, code for hospital. Only five of all eight group II isolates were shown having all identical or highly related banding patterns (\\> 95 % identity; see also \\[30\\])](1471-2180-7-28-1){#F1}\n\n![**Comparison of phylogenetic relatedness between identified MLST (a) and MLVA types (b) for all 58 hospital VRE**. Corresponding trees using eBURST v3 showing clades of an identical MLVA type (MT) for several MLST types (ST) (a) and an identical ST for several MTs (b), respectively. Each line separates MLVA/MLST types differing by a single locus (single locus variants). ST-64 is not part of the given collection and was included for a better illustration of relatedness between MLST types. The MLVA type of ST-64 is not known.](1471-2180-7-28-2){#F2}\n\n###### \n\nCharacteristics of investigated *E. faecium*.\n\n  **Isolate**               **VRE**   **Origin**   **Hospital/**   **year**   **PFGE**   **MLST**   **MLST**           **MLST**   **VNTR**^1^   **MLVA**   **MLVA**   **Reference**\n  ------------------------- --------- ------------ --------------- ---------- ---------- ---------- ------------------ ---------- ------------- ---------- ---------- ---------------\n  **Group III**                                                                                                                                                       \n                                                                                                                                                                      \n  UW5255                    **+**     CCI          L1              2004       2          16         1-2-1-1-1-1-1      C1/17      5-6-3-2-2-3   16         C1         this study\n  UW5250                    **+**     CCI          L2              2004       33         18         7-1-1-1-5-1-1      C1/17      5-6-3-3-2-3   2          C1         this study\n  UW5258                    **+**     CCI          L2              2004       34         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         \\[29\\]\n  UW5248                    **+**     CCI          W1              2004       1          65         1-2-1-20-1-1-9     C1/17      5-7-3-2-2-3   5          C1         \\[29\\]\n  UW5254                    **+**     CCI          G1              2004       23         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW5251                    **+**     CCI          H1              2004       32         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         \\[29\\]\n  UW5920                    **+**     CCI          BB              2004       21         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW6033                    **+**     B/CCI        BB              2005       22         282        7-1-1-1-5-20-1     C1/17      5-7-3-3-2-3   1          C1         this study\n  UW5905                    **+**     B/CCI        F               2004       12         192        15-1-1-1-1-7-1     C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW5903                    **+**     CCI          F               2004       21         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW5910                    **+**     CCI          F               2004       24         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW6316                    **+**     CCI          K1              2005       29         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6035                    **+**     CCI          M1              2005       23         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         \\[29\\]\n  UW5433                    **+**     CCI          M2              2004       4          16         1-2-1-1-1-1-1      C1/17      5-7-3-2-2-3   5          C1         \\[29\\]\n  UW6480                    **+**     CCI          M2              2006       35         17         1-1-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6464                    **+**     CCI          M2              2006       35         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6475                    **+**     CCI          M2              2006       32         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6455                    **+**     CCI          M2              2006       11         192        15-1-1-1-1-7-1     C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6474                    **+**     CCI          M2              2006       8          192        15-1-1-1-1-7-1     C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6324                    **+**     CCI          T               2005       20         203        15-1-1-1-1-20-1    C1/17      5-5-3-3-1-2   215        C1         this study\n  UW6337                    **+**     CCI          T               2005       20         203        15-1-1-1-1-20-1    C1/17      5-5-3-3-1-2   215        C1         this study\n  UW5869                    **+**     CCI          T               2004       6          279        1-5-1-1-1-7-1      C1/17      5-7-1-4-2-3   231        C1         this study\n  UW6321                    **+**     CCI          T               2005       7          280        1-3-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6322                    **+**     CCI          T               2005       22         282        7-1-1-1-5-20-1     C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6323                    **+**     CCI          T               2005       25         283        15-25-1-1-1-20-1   C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6338                    **+**     CCI          U               2005       21         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6343                    **+**     B/CCI        U               2005       20         203        15-1-1-1-1-20-1    C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6151                    **+**     B/CCI        K2              2005       9          17         1-1-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6256                    **+**     CCI          K2              2005       11         192        15-1-1-1-1-7-1     C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6352                    **+**     B/CCI        M3              2005       30         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6397                    **+**     B/CCI        M3              2005       31         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6203                    **+**     B/CCI        S1              2005       13         17         1-1-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6379                    **+**     B/CCI        S1              2005       28         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-3   1          C1         this study\n  UW6494                    **+**     B/CCI        S1              2006       26         78         15-1-1-1-1-1-1     C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6498                    **+**     B/CCI        G2              2006       28         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-2   7          C1         this study\n  UW6500                    **+**     CCI          G2              2006       28         18         7-1-1-1-5-1-1      C1/17      5-7-3-3-2-2   7          C1         this study\n  UW6571                    **+**     B/CCI        P               2006       27         78         15-1-1-1-1-1-1     C1/17      5-7-3-3-1-2   159        C1         this study\n  UW6589                    **+**     CCI          W2              2006       6          17         1-1-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         \\[20\\]\n  UW6605                    **+**     CCI          W2              2006       14         78         15-1-1-1-1-1-1     C1/17      5-7-3-3-1-3   12         C1         \\[20\\]\n  UW6511                    **+**     B/CCI        B-S             2006       17         202        1-1-1-1-1-7-1      C1/17      5-7-3-3-2-2   7          C1         this study\n  UW6517                    **+**     CCI          B-S             2006       17         202        1-1-1-1-1-7-1      C1/17      5-7-3-3-2-2   7          C1         this study\n  UW6528                    **+**     CCI          B-S             2006       17         202        1-1-1-1-1-7-1      C1/17      5-7-3-3-2-2   7          C1         this study\n  UW6520                    **+**     CCI          B-S             2006       5          314        43-5-1-1-12-7-1    C          5-7-1-4-2-3   231        C1         this study\n                                                                                                                                                                      \n  **Group II**                                                                                                                                                        \n                                                                                                                                                                      \n  UW901                     **+**     CCI          D               1996       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW931                     **+**     CCI          G2              1996       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW1505                    **+**     CCI          L4              1997       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW1806                    **+**     CCI          B-F             1998       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW1822                    **+**     CCI          E               1998       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW1823                    **+**     CCI          B-K             1998       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW1827                    **+**     CCI          H2              1998       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n  UW2322                    **+**     CCI          B-U             1999       15         117        9-1-1-1-1-1-1      C1/17      5-7-3-3-1-3   12         C1         \\[30\\]\n                                                                                                                                                                      \n  **Group I**                                                                                                                                                         \n                                                                                                                                                                      \n  70/90                     **+**     I/HS         B-C             1991       38         25         9-3-1-6-1-1-1      C/22       3-2-4-2-1-3   113        B          \\[31\\]\n  U200                      **+**     I/HS         M4              1992       37         8          5-2-1-6-1-7-1      B/26       7-7-3-2-1-3   98         C          \\[31\\]\n  L283                      **+**     I/HS         L3              1993       18         17         1-1-1-1-1-1-1      C1/17      5-5-3-3-2-3   163        C1         \\[31\\]\n  6011                      **+**     I/HS         B-C             1993       36         189        3-2-1-9-1-20-5     A/5        7-4-3-1-1-1   238        A          \\[31\\]\n  18912                     **+**     I/HS         P               1992       3          17         1-1-1-1-1-1-1      C1/17      5-7-3-3-2-3   1          C1         \\[31\\]\n  2513                      **+**     I/HS         S2              1994       10         22         2-3-1-2-1-1-1      C/22       6-7-3-4-2-3   178        C1         \\[31\\]\n  UW786                     **+**     CCI          N               1995       16         307        1-11-1-1-1-1-1     C1/17      5-7-1-4-2-3   231        C1         \\[31\\]\n  BM4147                    **+**     R            Ref.            1986       n.d.       25         9-3-1-6-1-1-1      C/22       5-2-3-2-1-1   95         A          \\[40\\]\n                                                                                                                                                                      \n  **Non-hospital origin**                                                                                                                                             \n                                                                                                                                                                      \n  UW315                     **+**     OS           \\-              1994       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n  UW328                     **+**     OS           \\-              1994       n.d.       n.d.                          \\-         5-7-1-3-2-1   241        C1         \\[31\\]\n  UW440                     **+**     OS           \\-              1994       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n  UW2500                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[32\\]\n  UW2508                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-4-1-3-1-2   242        B          \\[32\\]\n  UW2516                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         N-N-1-3-1-2   n.t.       \\-         \\[32\\]\n  UW2636                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-N-2-2-1-1   n.t.       \\-         \\[32\\]\n  UW2642                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-4-3-1-1-2   173        A          \\[32\\]\n  UW2647                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[32\\]\n  UW2651                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[32\\]\n  UW2655                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         5-N-1-3-1-1   n.t.       \\-         \\[32\\]\n  UW2712                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         3-6-1-4-1-3   243        C          \\[32\\]\n  UW2721                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         6-6-4-2-1-1   244        A          \\[32\\]\n  UW2726                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         5-2-4-3-1-2   245        B          \\[32\\]\n  UW2729                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         4-0-7-2-1-N   n.t.       \\-         \\[32\\]\n  UW2743                    **-**     OS           \\-              1999       n.d.       n.d.                          \\-         5-2-4-3-1-2   245        B          \\[32\\]\n  AW1                       **+**     S            \\-              1991       n.d.       n.d.                          \\-         5-4-1-3-3-2   122        B          \\[31\\]\n  AW2                       **+**     S            \\-              1991       n.d.       n.d.                          \\-         5-4-3-1-1-1   90         A          \\[31\\]\n  AW5                       **+**     S            \\-              1991       n.d.       n.d.                          \\-         5-4-3-1-1-1   90         A          \\[31\\]\n  AW9                       **+**     S            \\-              1991       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n  AW12                      **+**     S            \\-              1991       n.d.       n.d.                          \\-         3-4-4-2-1-2   103        B          \\[31\\]\n  UW4                       **+**     F            \\-              1993       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n  UW7                       **+**     F            \\-              1993       n.d.       n.d.                          \\-         6-4-3-1-1-2   173        A          \\[31\\]\n  UW53                      **+**     M            \\-              1994       n.d.       n.d.                          \\-         4-N-4-2-1-2   n.t.       \\-         \\[31\\]\n  UW68                      **+**     F            \\-              1994       n.d.       n.d.                          \\-         5-N-4-2-1-2   n.t.       \\-         \\[31\\]\n  UW261                     **+**     M            \\-              1994       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n  UW262                     **+**     M            \\-              1994       n.d.       n.d.                          \\-         6-4-3-1-1-1   89         A          \\[31\\]\n\nLegend: Only one isolate per case/patient was allowed. Legend: CC, clonal complex, assigned according to references \\[14,16,41\\]. Origin: CCI, epidemic isolate/cluster of colonizations and infections (n \u2265 2); B, blood stream isolate; I/HS, from single infection or hospital patient stool sample; OS, outpatient stool sample; S, sewage isolate; F, poultry or pig farm; M, poultry meat or pork sample; R, reference isolate; *vanA*, *vanA*glycopeptide resistance genotype; VRE, vancomycin-resistant *E. faecium*; n.t., not typeable; n.d., not determined. ^1^N, no amplification of this locus; MLST scheme: *atpA*, *ddl*, *gdh*, *purK*, *gyd*, *pstS*, *adk*; MLVA scheme: VNTRs 1,2,7,8,9,10.\n\n###### \n\nDiscriminatory indices for and congruence values between PFGE, MLST and MLVA for all 58 hospital *E. faecium*. \\[37--39\\]\n\n         PFGE    MLST    MLVA    Discriminatory index   Confidence interval (95% CI)\n  ------ ------- ------- ------- ---------------------- ------------------------------\n  PFGE           0.935   0.863   0.972                  \\[0.951--0.992\\]\n  MLST   0.935           0.873   0.91                   \\[0.877--0.943\\]\n  MLVA   0.863   0.873           0.842                  \\[0.79--0.893\\]\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nGlobally, noncommunicable diseases (NCDs) accounted for 71% of total deaths. In India, NCDs were estimated to account for 63% of all deaths, and cancer was one of the leading causes (9%).^[@B1]^ Cancer registries are recognized as vital components of national cancer-control programs.^[@B2]^ Publications from developed and developing countries provide updated information on cancer occurrence, trends, and projections.^[@B3]-[@B7]^ In India, the systematic collection of data on cancer has been performed since 1982 by the population-based cancer registries (PBCRs) and hospital-based cancer registries (HBCRs) under the National Cancer Registry Programme (NCRP)--National Centre for Disease Informatics and Research (NCDIR) of the Indian Council of Medical Research (ICMR; ICMR-NCDIR-NCRP), Bengaluru (Appendix). NCRP commenced with the objective of generating reliable data on the magnitude and patterns of cancer. Several NCRP reports on cancer from different registries across India have been published.^[@B8]-[@B9]^\n\nCONTEXT\n-------\n\n**Key Objective**This study reports the cancer incidence, patterns, trends, projections, cancer mortality, and clinical aspects of stage at presentation and treatment for the period 2012-2016 from 28 population-based and 58 hospital-based cancer registries in India under the network of the National Cancer Registry Programme.**Knowledge Generated**There is heterogeneity in cancer incidence (age-adjusted rate per 100,000: males: Osmanabad and Beed, 39.5 *v* Aizawl district, 269.4; females: Osmanabad and Beed, 49.4 *v* Papumpare district, 219.8) and in patterns across India and a lower proportion (\\< 33%) of early stage at presentation for common cancers. The projected incidence of patients with cancer for the year 2020 in India in males is 679,421 and in females is 712,758.**Relevance**Results of this study will help in assessing the status and trends of cancer in India. This will assist local- and national-level stakeholders to implement public health action to control cancer.\n\nPBCRs provide statistics on the occurrence and outcome of cancer in a geographically defined population. They also provide the framework for assessing the control of cancer in the community. HBCRs are concerned with the recording of information on patients with cancer seen in a particular hospital and are mainly used for reviewing clinical performance and the hospital cancer program.^[@B10]^\n\nThis article reports the cancer incidence, patterns, time trends, and mortality from 28 PBCRs for the composite period 2012-2016 across India under ICMR-NCDIR-NCRP. Also, it covers the stage at presentation and type of treatment received by patients with cancer from 58 HBCRs for the period 2012-2016 in India under the network of NCRP. Based on estimates from this period, the projection of patients with cancer in India for the year 2020 is also included.\n\nMETHODS {#s2}\n=======\n\nPresently, there are 36 PBCRs and 236 HBCRs registered under the ICMR-NCDIR-NCRP. However, this article includes data from 28 PBCRs and 58 HBCRs, which were complete with at least 1 year of good-quality data. All neoplasms with a behavior code of 3 as defined by the International Classification of Diseases for Oncology, 3rd Edition, and the International Statistical Classification of Diseases and Related Health Problems (10th revision; ICD-10) were considered reportable and therefore registered in NCRP.^[@B11]^ Cancer registration is a complex process. In India, cancer registration is active wherein trained registry staff go to different sources (hospitals, diagnostic laboratories, vital statistics departments) for collection of data on a standardized core form.^[@B10]^ This is followed by quality control checks, duplicate checks, matching with mortality cases, follow-up of death certificate notifications, and assigning death certificate only (DCO). Patients with cancer who were residents in the registration area for a minimum period of 1 year before the date of diagnosis were included in the registry. Multiple cancer data sources were followed for data collection.^[@B10]^ Quality of the data was maintained per International Association of Cancer Registries/International Agency for Research on Cancer (IACR/IARC) norms.^[@B12],[@B13]^\n\nIncidence and mortality data were retrieved from 28 PBCRs (urban or rural, or both) for the period 2012-2016. Cancer burden measures were calculated as crude rate, age-adjusted rate (AAR) per 100,000 population using world standard population,^[@B14]^ and cumulative risk (probability that an individual will be diagnosed with cancer \\[0- to 74-year-old age group\\] in the absence of any competing cause of death and assuming that the current trends prevail over the time period). Time trends in cancer incidence rate were generated as annual percent change (APC) and considered statistically significant (at *P* \\< .05) from 16 PBCRs (with minimum of 10 years of data) using the Joinpoint regression program, 4.0.1 (National Cancer Institute).^[@B15]^ The years of data for trend analysis varied (11 to 35 years) across the PBCRs, and there were 6 PBCRs with more than 25 years of continuous data.\n\nThe country was categorized into 6 geographic zones based on the location of the PBCRs: North (Delhi, Patiala), South (Hyderabad, Kollam, Thiruvananthapuram, Bangalore, and Chennai), East (Kolkata), West (Ahmedabad urban, Aurangabad, Osmanabad and Beed, Barshi rural, Mumbai, and Pune), Central (Wardha, Bhopal, and Nagpur) and Northeast (NE; Manipur, Mizoram, Sikkim, Tripura, West Arunachal, Meghalaya, Nagaland, Pasighat, Cachar, Dibrugarh, and Kamrup urban). Population denominators were derived from the Census of India, conducted by the Regis-trar General and Census Commissioner of India under the Ministry of Home Affairs, Government of India.^[@B16]^ The census populations of 2001 and 2011 were used to calculate the postcensal population estimates for the years 2012 and 2016 by 5-year age groups and sex. For time trend analysis, the respective 1981, 1991, 2001, and 2011 censuses were taken as the base to estimate inter- or postcensal population by 5-year age groups and sex.^[@B17]^\n\nA standardized patient information form for HBCRs under NCRP and the data collection methods have been explained previously.^[@B18]^ Newly diagnosed and treated patients with cancer (N = 667,666) from 58 HBCRs (with a minimum of 1 year of a complete dataset) in 2012-2016 were pooled and are presented as the relative proportion of clinical stage and treatment. The stage/clinical extent of disease before treatment (excludes previously treated patients with cancer) was classified into localized only, locoregional, distant metastasis, and unknown.^[@B19]^ The completion of planned cancer-directed treatment after diagnosis was classified as surgery, radiotherapy, systemic therapy, and multimodality (combination of surgery and/or radiotherapy and/or systemic therapy).\n\nGood-quality data indices of microscopic verification (MV%) above 75%, DCO% below 20%, other and unspecified sites (O&U%) below 15%, and mortality-to-incidence ratio (M:I%) were calculated for each PBCR and accordingly classified.^[@B12],[@B13]^ NCRP has developed in-house software (PBCR and HBCR Data Management) for data capture, quality checks, duplicates checks (deterministic and phonetic-similar sounding duplicate names), and mortality-incidence matching. The list of errors was sent back to registries for clarifications and corrected at each level. Incidence data for 2012-2016 was used as a reference for projection of patients with cancer in India until 2020 by sex and anatomic site (see Data Supplement for more details).\n\nRESULTS {#s3}\n=======\n\nAverage annual number of patients for all sites of cancer with incidence rate, cumulative risk by sex, and mortality rate according to different regions in India between 2012 and 2016 are given in [Table 1](#T1){ref-type=\"table\"}. PBCRs in the NE showed the highest incidence rate in both sexes. It was observed that Aizawl district had the highest AAR (269.4) and mortality (152.7) rate among males. One of every 4 males in Aizawl district, Papumpare district, East Khasi Hills district, and Kamrup urban are likely to develop cancer in the age group of 0-74 years. One of every 4 females in the Papumpare district and 1 of 5 females in Mizoram state are likely to develop cancer in the age group of 0-74 years.\n\n###### \n\nAnnual Average No. of Patients for All Sites of Cancer With Incidence Rate per 100,000, Cumulative Risk (0-74 age group) and Mortality Rate by Sex and Reporting Year for 28 PBCRs Under NCRP\n\n![](GO.20.00122t1)\n\nOn comparison of AARs for all sites of cancer (ICD10: C00-C97) across the population among males, Aizawl district (269.4) had the highest AAR and was approximately twice the AAR of Delhi PBCR (147.0). East Khasi Hills district of Meghalaya (227.9) had the second highest AAR, followed by Kamrup urban (213.0) and Mizoram PBCR (207.0). Among females, Papumpare district (219.8) had the highest AAR. In the rest of the regions (excluding NE), Delhi had highest AAR (147.0), followed by Thiruvanathapuram district (137.8) among males; Bangalore (146.8) had highest AAR, followed by Delhi (141.0) among females ([Fig 1](#f1){ref-type=\"fig\"}).\n\n![Comparison of all cancer sites' age-adjusted incidence rates (AARs) of all population-based cancer registries, 2012-2016 (International Statistical Classification of Diseases and Related Health Problems, 10th revision: C00-C97). AARs are in blue and crude rates are given in parentheses in red. Thi\\'puram district, Thiruvananthapuram district.](GO.20.00122f1){#f1}\n\nEast Khasi Hills district of Meghalaya had the highest relative proportion of cancers associated with the use of tobacco, with 70.4% and 46.5% for males and females, respectively. The higher proportion of cancers associated with use of tobacco was in the NE states, followed by registries in the West and Central regions ([Fig 2](#f2){ref-type=\"fig\"}).\n\n![Sites of cancer associated with the use of tobacco (%), 2012-2016. Crude rate in red and age-adjusted rates in blue are given in parentheses. Sites of cancer associated with the use of tobacco, anatomic sites (International Statistical Classification of Diseases and Related Health Problems, 10th revision codes), lip (C00); tongue (C01-C02); mouth (C03-C06); pharynx (C10 and C12-C14); esophagus (C15); larynx (C32); lung (C33-C34); urinary bladder (C67).^[@B33]^ Thi'puram district, Thiruvananthapuram district.](GO.20.00122f2){#f2}\n\nAmong males, lung, mouth, esophagus, and stomach were the most common cancer sites. Among females, breast cancer, followed by cervix uteri and ovary cancer, were the most common sites across the PBCRs. Thyroid cancer was the second most common cancer in the PBCRs of Thiruvananthapuram and Kollam, whereas lung cancer was seen in Manipur and Mizoram state. In the NE region, the third most common cancers were stomach and gallbladder (Data Supplement). The decadal changes in leading sites of cancer from 6 older PBCRs (Barshi rural, Bangalore, Bhopal, Chennai, Delhi, and Mumbai) were observed for the first 10 and last 10 years of data (Data Supplement).\n\nThe relative proportion of patients according to clinical extent of disease at the time of diagnosis as seen in the pooled data of 58 HBCRs for common sites of cancer showed that the majority of patients with cancer were diagnosed as locally advanced/locoregional for breast (57.0%), cervix uteri (60.0%), head and neck (66.6%), and stomach (50.8%) cancer. The majority of patients with lung cancer were diagnosed with distant metastasis in males (44.0%) and females (47.6%; [Fig 3](#f3){ref-type=\"fig\"}). The relative proportion of types of cancer-directed treatment received (only at the reporting hospital) according to clinical extent of disease before treatment showed that multimodality was the first choice of treatment (locoregional, 79.5%; localized, 74.4%; distant metastasis, 47.6%; Data Supplement).\n\n![Relative proportion (%) of patients according to clinical extent of disease, 2012-2016 (proportion \\[%\\] may not total 100% because of rounding).](GO.20.00122f3){#f3}\n\nThe estimated APC in cancer AAR for selected anatomic sites of cancer over the time period showed an increase in the incidence rate of all sites of cancer (12 PBCRs in males and 13 PBCRs in females). There was a significant increase in the incidence rate of breast cancer across all PBCRs over the years, except in Nagpur PBCR. There was a significant decrease in the incidence rate of cervical cancer in 10 PBCRs, except in Dibrugarh district and Pune. Lung cancer showed a significant increase in 11 PBCRs among females ([Fig 4](#f4){ref-type=\"fig\"}).\n\n![Annual percentage change (APC) in age-adjusted incidence rates (AAR) over the time period. Calendar years of incidence data for each population-based cancer registry (PBCR) used in trend analysis: Chennai (1982-2016); Mumbai (1982-2015); Bangalore (1982-2014); Barshi rural (1988-2016); Bhopal (1988-2015); Delhi (1988-2014); Kamrup urban (2003-2016); Mizoram state (2004-2016); Dibrugarh district, Sikkim state, Imphal West district; Aurangabad; Thiruvananthapuram (Thi\\'puram) taluk; Nagpur (2005-2016); Kollam district; Pune (2006-2016); PBCRs with small numbers (\\< 10 patients) per year not analyzed. Increase in APC in red, decrease in APC in green. (\\*) APC significantly different from zero; *P* \\< .05. Thi'puram PBCR expanded its coverage to district from the year 2012 onwards. Hence, Thi'puram taluk data was used for trend analysis. Thi'puram taluk, Thiruvananthapuram taluk.](GO.20.00122f4){#f4}\n\nAmong 28 PBCRs, MV%, ranged between 77% (Patiala) and 96.7% (Hyderabad); DCO% ranged between \\< 0.05% (Osmanabad and Beed) and 19.4% (Patiala); O&U% ranged between 1.8% (Hyderabad) and 13.0% (Patiala). M:I% was high in Barshi rural (67.2%) followed by Wardha (59.2%) and Mumbai (56.0%; Data Supplement).\n\nThe projected incidence of patients with cancer in India among males was 679,421 (94.1 per 100,000) and among females 712,758 (103.6 per 100,000) for the year 2020. One in 68 males (lung cancer), 1 in 29 females (breast cancer), and 1 in 9 Indians will develop cancer during their lifetime (0-74 years of age; [Table 2](#T2){ref-type=\"table\"}). The projected 5 most common cancers in 2020 for males (lung, mouth, prostate, tongue, and stomach) constitute 36% of all cancers and for females (breast, cervix uteri, ovary, corpus uteri, and lung) constitute 53% of all cancers (Data Supplement).\n\n###### \n\nProjected Incidence of Cancer Statistics in India, 2020\n\n![](GO.20.00122t2)\n\nDISCUSSION {#s4}\n==========\n\nIndia exhibits heterogeneity in cancer. The incidence rates of Aizawl district were observed to be 7 times and 4 times that of Osmanabad and Beed district PBCRs in males and females, respectively. The highest cancer incidence rate was observed in the NE region (6 PBCRs for males and 4 PBCRs for females) than other areas in the country. The leading sites of cancer in the NE region were nasopharynx, hypopharynx, esophagus, stomach, liver, gallbladder, larynx, lung, breast, and cervix uteri. The NE region lacks required infrastructure with respect to specialized treatment facilities, human resources, as seen by the low 5-year survival of breast, cervix, and head and neck cancer compared with rest of India. A substantial proportion of patients with cancer from the NE region are traveling outside the NE for treatment and cancer care.^[@B20],[@B21]^ Local cultural factors and lifestyle choices may have contributed to the heterogeneity in cancer incidence pattern and differences in India, as was seen in Thailand.^[@B7]^\n\nLung (9 PBCRs), mouth (9 PBCRs), esophagus (5 PBCRs), stomach (4 PBCRs), and nasopharynx (1 PBCR) cancers were the most common cancers in men. Lung cancer was the leading site in metropolitan cities and the southern region, whereas mouth cancer was the leading site in the West and Central regions. Lung cancer and oral/mouth cancer were the most common cancers among males in the Indian subcontinent.^[@B22]^ Cancers of the esophagus, stomach, and nasopharynx were the leading sites in the NE region of India. Here, the cancer incidence pattern is different from the rest of India. There are similarities in the cancer incidence pattern with the Southeast Asian region.^[@B23]^ Overall, these findings on patterns of cancer were similar to previously published reports under NCRP.^[@B8]^\n\nCancer of the breast (19 PBCRs) and cervix uteri (7 PBCRs) were the most common cancers in women. The highest burden of breast cancer was observed in metropolitan cities. There is an increase in the trend of incidence of breast cancer, whereas cervix uteri cancer is on the decline. A steady increase in breast cancer in most of the PBCRs including newer PBCRs, poses a great health challenge to women in India.^[@B24]^ Presently, breast cancer and cervix uteri are the leading sites of cancer among women in India, posing an important public health problem that needs important input from various health and other agencies to tackle.^[@B25]^ A multidisciplinary approach to breast cancer, including awareness programs, preventive measures, screening programs for early detection, and availability of treatment facilities, are vital for reducing both incidence and mortality of cancer in Indian women.^[@B26]^\n\nThe incidence rate of thyroid cancer among women is increasing, and it is most common in the districts of Thiruvananthapuram and Kollam in Kerala. The high burden of thyroid cancer in Kerala could be due to overdiagnosis,^[@B27]^ as was observed even in high-income and low- and middle-income countries.^[@B28]^ AAR in Barshi rural is almost one third of urban PBCRs (males, 50.6 *v* 147.0; females, 61.0 *v* 146.8), and the increase in APC was less compared with urban PBCRs. This needs additional investigation.\n\nThere are cancers of several anatomic sites known to be associated with the use of tobacco.^[@B29]^ Based on PBCR data, almost one third of the cancers were known to be associated with the use of tobacco in India. India state-level disease burden initiative cancer collaborators estimated that tobacco use was the highest contributing risk factor for cancer in India. In India, lung cancer can be attributed to tobacco use and air pollution, which are the leading risk factors.^[@B30]^ Approximately 70% of cancers in India were potentially preventable through modifiable risk factors.^[@B31]^\n\nBecause it is difficult to obtain information on the clinical extent of disease and treatment from PBCRs, the hospital database was used for such analysis.^[@B10]^ The majority of breast and cervix uteri cancers were diagnosed at a locally advanced stage. Chemoradiation was the most common type of treatment of cancer cervix uteri. A multi-institutional study from India on cervix cancer showed significantly better survival with chemoradiation than radiation alone in the locally advanced stage.^[@B18]^ A study from Chennai showed that concurrent chemoradiation for locally advanced cervical cancer resulted in the best disease-free survival.^[@B32]^ Two thirds of the patients with cancer were diagnosed at the locoregional stage for head and neck cancers from HBCRs. Similar to that, a low proportion of patients with head and neck cancer presented in the early stage, and a high proportion (88.1%) were seen in Uttarakhand.^[@B33]^ Multimodality was the most common treatment given for breast and head and neck cancers. A multi-institutional study estimated that 65% of new head and neck cancers with locally advanced disease did not receive the benefit of optimal treatment, resulting in poor survival.^[@B34]^\n\nLess than one fifth of lung and stomach cancers were diagnosed as localized only. Systemic therapy was the most common type of treatment given for lung and stomach cancer. A previous report on HBCR results showed similar findings.^[@B8a]^ A hospital-based study from northern India showed that 90% of patients with lung cancer were diagnosed at an advanced stage of the disease, and there was a delay in diagnostic evaluation and treatment.^[@B35]^ Creating cancer awareness, preventing risk factors, and improving access to care among people would result in downstaging of cancer.\n\nThe measure of validity, MV%, was above 77% for all the PBCRs. Varying patterns of DCO% and M:I% were observed among PBCRs which were dependent on the quality of death registration and certification.^[@B13]^ Efforts to improve the quality are always underway. In some registries, low DCO% (\\< 1%) is due to nonavailability of all-cause mortality data and incomplete/incorrect certification of cause of death. Some registries had an efficient trace back procedure by house visit/phone. Data from PBCRs were regularly published in successive volumes of Cancer Incidence in Five Continents (CI-5) by WHO-IACR/IARC. The incidence data from 15 PBCRs under NCRP (India) were published in Cancer Incidence in Five Continents, Volume XI, by WHO-IACR/IARC.^[@B36]-[@B38]^\n\nThe projected incidence of patients with cancer is higher for females (712,758) than males (679,421) for the year 2020. The projected national cancer incidence burden in 2020 will be 98.7 per 100,000 population (1,392,179 patients) as a conservative estimate. It is assumed that the observed rate of 2012-2016 will remain unchanged until 2020. The time trend in rate was not used to avoid uncertainty in the projection for a populous country like India. NCRP has estimated a slightly higher number of patients with cancer compared with IACR/IARC and GLOBOCAN for all sites of cancer in 2018. This may be because of a difference in methodology and use of recent data (1,392,179 *v* 1,157,294).^[@B22]^ This is the first such attempt in the country and will be further updated on availability of the next data set and census information. The influencing factors, such as risk factors/behavior, case finding procedure, screening program, and improved techniques for detecting patients with cancer, are likely to influence the projected number of patients.\n\nPBCRs in this study covered 100 million average annual person-years, accounting for coverage close to 10% of the population in India. Cancer registration in India faces several challenges because it is not a notifiable disease, posing challenges to data collection.^[@B39]-[@B42]^ The mortality registration system has several gaps, including incomplete and inaccurate certification of cause of death.^[@B43],[@B44]^ Registering through passive notification by health care providers to report cancer occurrence in India would improve the coverage with limited resources. Linking of cancer registry data with Ayushman Bharat,^[@B45]^ mortality databases, and the Hospital Information System would improve cancer registration, follow-up, and outcome data.\n\nThis study provides a framework for assessing the status and trends of cancer in India. This shall guide appropriate support for action to strengthen efforts to improve cancer prevention and control to achieve the National NCD targets and the sustainable development goals.^[@B46],[@B47]^ The data also provide leads to key research questions.\n\nThe authors acknowledge the contribution of the 28 population-based cancer registries and 58 hospital-based cancer registries under the National Cancer Registry Program, which contributed data, and Research Area Panel Expert Group on Cancer.\n\nAUTHOR CONTRIBUTIONS {#s5}\n====================\n\n**Conception and design:** Prashant Mathur, Krishnan Sathishkumar, Meesha Chaturvedi\n\n**Collection and assembly of data:** Prashant Mathur, Krishnan Sathishkumar, Meesha Chaturvedi, Priyanka Das, Kondalli Lakshminarayana Sudarshan, Anish John, Francis Selvaraj Roselind\n\n**Data analysis and interpretation:** All authors\n\n**Manuscript writing:** All authors\n\n**Final approval of manuscript:** All authors\n\n**Accountable for all aspects of the work:** All authors\n\nAUTHORS\\' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST\n========================================================\n\nThe following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO\\'s conflict of interest policy, please refer to [www.asco.org/rwc](http://www.asco.org/rwc) or [ascopubs.org/go/site/misc/authors.html](http://ascopubs.org/go/site/misc/authors.html).\n\nOpen Payments is a public database containing information reported by companies about payments made to US-licensed physicians ([Open Payments](https://openpaymentsdata.cms.gov/)).\n\nNo potential conflicts of interest were reported.\n\nSwetha Acharya, Dillip Kumar Agarwalla, Zarika Ahmed, Gazi Naseem Ahmed, Shiraj Ahmed, B.S. Ajaikumar, P. Anandhi, Anil Kumar Anand, Ramandeep Arora, R.A. Badwe, Manas Ranjan Baisakh, Balamuralikrishna, Debabrata Barmon, Sangita Bhandari, Samir Bhattacharya, M. Bhattacharyya, Arup Bhowmik, T.W. Bhutia, T. Bindhu, Vijay Kumar Bodal, Pradeep Chandrakar, Latha Chaturvedua, Richa Chauhan, Saia Chenkual, S. Chhabra, Lily Chhakchhuak, Vivek Choudhary, Ashok Kumar Das, Rituparna Das, Dhritiman Datta, Biswajit Debbamma, Shiromani Debbarma, S.V. Suryanarayana Deo, Vinay Deshmane, Madhubala Devi, Punyabati Devi, Laishram Purnima Devi, Rajesh Dikshit, H. Dkhar, Biswajit Dubashi, Debanarayan Dutta, Vijay Gadagi, Vinay Gadigi, B. Ganesh, M.S. Ganesh, K.V. Gangadharan, Nitin Gangane, Uma Garg, Preethi Sara George, Sushmita Ghosal, Radha Ghosh, Debasis Gochhait, Ajay Gogia, H.K. Goswami, Ashutosh Gupta, Manish Gupta, Kaveri Hallikeri, Rajesh Harsvardhan, B.H. Srinivas, Vandana Jain, Swarima Jaitley, K.L. Jayakumar, P. Jayalekshmi, Kaling Jerang, G. Jongkey, J. Seios, K. Vijayakumar, D. Kadambari, Geeta Kadayaprath, Ravi Kannan, R. Kapoor, Tejinder Kataria, Mohanvir Kaur, Swarn Kaur, K. Gunaseelan, Shehna A. Khader, V. Khamo, Nazir Ahmad Khan, Sushma Khuraijam, Marla K. Prashanth, Anjali Kolhe, S.S. Koyande, Adishi Kri, Dinesh Kumar, G. Dilip Kumar, Narender Kumar, S. Shiva Kumar, Sunil Kumar, Rajesh Singh Laishram, Lalchhanhimi, K. Lalfakzuala, Lalhlupuii, Doris Lallawmzuali, Jerry Lalrinsanga, Cindy Lalthanpuii, Imliwati Longkumer, M.B. Jayaraman, Gautam Majumdar, Reeni Malik, Swapna Maliye, Indranil Mallick, Syamsundar Mandal, N. Manoharan, Lone Mohammad Maqbool, Aleyamma Mathew, Sanjay Mishra, Janmejay Mohapatra, C.T. Muthukumaran, Rekha G. Nair, Prem Nair, Sunithra Nair, Durgaprasad Nanda, D.K. Nath, Partha Nath, N. Venkatesh, Gamba Padu, Rakesh Pandey, Shashank Pandya, Sangeeta Pankaj, N.S. Panse, Sujith Chayu Patanaik, P.C. Pathy, Jeremy L. Pautu, K. Pavithran, Prasanth Penumade, Priya Darshini Pradhan, Aruna E. Prasad, Punitalal, Kedozeto Punyu, K. Kalpana Raghunath, Mujeeb Rahman, Vinod Raina, Ravi Shankar Rajendra, Ramachandra, C. Ramesh, B.S. Ramesh, Rakesh S. Ramesh, R. Rama Ranganathan, Rita Rani, K. Ratheesan, Sudhir Rawal, Prarthana Roselil, G. Sadashivudu, B.J. Saikia, Projnan Saikia, S.K. Samantara, A. Santa, Anupam Sarma, N. Hariharanadha Sarma, Sunil Kumar Saxena, Pooja Sethi, Anand Shah, B.K. Sharma, Jagannath Dev Sharma, Sudarshan Sharma, Shah Alam Sheikh, Kamalaksh Shenoy, Dinesh Shet, Atul Shrivastava, B.R. Shrivastava, Sushma Shrivastava, Alben Sigamani, Neetu Singhal, Indibor Singh, Rajesh Kumar Singh, Usha Singh, Badarisha R. Sohliya, G.S. Sreenath, U. Srihari, V.G. Sudhakaran, Saina Sunilkumar, Partha Sarathi Sutradhar, R. Swaminathan, A Talukdar, Ritesh Tapkire, Sopai Tawsik, B. Paul Thaliath, K.T. Harichandrakumar, Pampa Ch. Toi, Shyam Tsering, B.B. Tyagi, Ajith Kumar, V.R. Ashok Kumar Vaid, K. Velavan, C.R. Vijay, W.B. Langstieh, Chaitali Wagmere, Meenu Walia, John Zohmingthanga, Eric Zomawia, and B. Zothankima.\n\nNames Listed Alphabetically {#s7}\n===========================\n\nAffiliations. {#s8}\n-------------\n\nNational Centre for Disease Informatics and Research, Indian Council of Medical Research, Bengaluru, India: Prashant Mathur, DNB, PhD; Krishnan Sathishkumar, MSc; Meesha Chaturvedi, MBBS; Priyanka Das, B-Level; Kondalli Lakshminarayana Sudarshan, MSc; Stephen Santhappan, MSc, MPhil; Vinodh Nallasamy, MSc, MPhil; Anish John, BE; Sandeep Narasimhan, MSc; and Francis Selvaraj Roselind, MSc.\n\nPopulation-based cancer registries. {#s9}\n-----------------------------------\n\nDelhi population-based cancer registry (PBCR) -- B.R. Ambedkar Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi (S.V. Suryanarayana Deo, MS, N. Manoharan); Patiala District PBCR - Government Medical College and Rajindra Hospital, Punjab (Vijay Kumar Bodal, MD, Mohanvir Kaur, MD); Hyderabad PBCR - Nizam's Institute of Medical Sciences, Telangana (G. Sadashivudu, MD, U. Srihari); Kollam PBCR - District Cancer Registry Kollam, Regional Cancer Centre, Kerala (Rekha G Nair, MD, P. Jayalekshmi, PhD); Thiruvananthapuram PBCR - Regional Cancer Centre, Thiruvanan-thapuram, Kerala (Aleyamma Mathew, PhD, Preethi Sara George, PhD); Bengaluru PBCR -- Kidwai Memorial Institute of Oncology, Karnataka (Ramachandra, MS, C. Ramesh, PhD, C.R. Vijay, PhD); Chennai PBCR -- Cancer Institute (WIA), Tamil Nadu (R. Swaminathan, PhD, R. Rama Ranganathan, PhD); Kolkata PBCR -- Chittaranjan National Cancer Institute and Saroj Gupta Cancer Centre and Research Institute, West Bengal (Samir Bhattacharya, PhD, Syamsundar Mandal, PhD, Durgaprasad Nanda, MS); Ahmedabad urban PBCR -- Gujarat Cancer & Research Institute, Gujarat (Shashak Pandya, MS, Anand Shah, MD); Mumbai, Aurangabad, Nagpur, Pune PBCRs -- Indian Cancer Society, Maharashtra (Vinay Deshmane, MD, S.S. Koyande); Barshi Expanded (Osmanabad and Beed Districts) PBCR -- Nargis Dutt Memorial Cancer Hospital, Maharashtra (NS Panse); Barshi Rural PBCR -- Tata Memorial Hospital, Maharashtra (R.A. Badwe, MS, Rajesh Dikshit, PhD); Wardha PBCR -- Mahatma Gandhi Institute of Medical Sciences, Maharashtra (Nitin Gangane, MD, Swapna Maliye, S. Chhabra, MD); Bhopal PBCR -- Gandhi Medical College, Madhya Pradesh (Reeni Malik, MD, Atul Shrivastava); Manipur PBCR -- Regional Institute of Medical Sciences, Manipur (Madhubala Devi, MD, Punyabati Devi, MD, Laishram Rajesh, MD, Sushma Khuraijam); Mizoram PBCR -- Civil Hospital, Mizoram (Eric Zomawia, MD, John Zohmingthanga, MD, Lalchhanhimi, MD, Saia Chenkual, MD, B. Zothankima, MD, Lily Chhakchhuak, MBBS, Jerry Lalrinsanga, MBBS); Sikkim PBCR -- Sir Thutob Namgyal Memorial Referral Hospital, Sikkim (T.W. Bhutia, MD, Priya Darshini Pradhan, MD, Sangita Bhandari); Naharlagun (West Arunachal) PBCR -- Timo Riba State Hospital, Arunachal Pradesh (Sopai Tawsik, MD, Shyam Tsering, MD, K.R. Adishi, MD, Gamba Padu, MD); Meghalaya PBCR -- Civil Hospital, Meghalaya (W.B. Langstieh, H. Dkha, MD, B. Sohliya, MD, L Purnima Devi, MD); Nagaland PBCR -- Naga Hospital Authority, Nagaland (V. Khamo, MD, Kedozeto Punyu, MD); Pasighat PBCR -- Bakin Pertin General Hospital, Arunachal Pradesh (Kaling Jerang, PGDCP, G. Jongkey, MD); Tripura PBCR -- Cancer Hospital, Regional Cancer Centre, Tripura (Gautam Majumdar, MD, Arup Bhowmik, MD, Rituparna Das, MD, Shiromani Debbarma); Cachar PBCR -- Silchar Medical College, Assam (Shah Alam Sheikh, MD, R.P. Banik, MD); Dibrugarh PBCR -- Assam Medical College and Hospital, Assam (Projnan Saikia, MD, Zarika Ahmed, MD); Guwahati (Kamrup urban) PBCR -- Dr. Bhubaneswar Borooah Cancer Institute (Regional Institute for Treatment and Research), Assam (Jagannath D. Sharma, MD, Anupam Sarma, MD, Debabrata Barmon, MD, Shiraj Ahmed, MD).\n\nHospital-Based Cancer Registries {#s10}\n--------------------------------\n\nA.J. Hospital & Research Centre, Mangalore (Marla K. Prashanth, MS, Kamalaksh Shenoy, MD); Acharya Harihar Regional Cancer Centre, Cuttack (Dillip Kumar Agarwalla, MD, Janmejay Mohapatra, MD, P.C. Pathy, MD, S K Samantara, MD); Amrita Institute of Medical Sciences and Research, Kochi (Debanarayan Dutta, MD, K. Vijayakumar, MS, Prem Nair, MD, K. Pavithran, MD); Apollo Hospital, Bhubaneswar (Manas Ranjan Baisakh, MD); Asian Institute of Medical Sciences, Faridabad (Neetu Singhal, DNB); Assam Medical College, Dibrugarh (H. K. Goswami, MD); BPS Government Medical College for Women, Sonipat (Uma Garg, MD, Swarn Kaur, MD); Cachar Cancer Hospital, Silchar (Ravi Kannan, MS, Ritesh Tapkire, MS), Cancer Hospital & Research Institute, Gwalior (B.R. Shrivastava, MS); Cancer Institute (WIA), Chennai (R. Rama Ranganathan, PhD, R. Swaminathan, PhD); Chittaranjan National Cancer Institute, Kolkata (Partha Nath, MD, Syamsundar Mandal, PhD); B. Borooah Cancer Institute, Guwahati (M Bhattacharyya, MD, Ashok Kumar Das, MS, B J. Saikia, MD, A Talukdar, MS, Jagannath Dev Sharma, MD); B.R. Ambedkar Institute Rotary Cancer Hospital, New Delhi (Ajay Gogia, MS, Sunil Kumar, MS, S.V. Suryanarayana Deo, MS, N. Manoharan, G.K. Rath, MD); Erode Cancer Centre, Thindal, Erode (K. Velavan, MD); Father Muller Medical College Hospital, Mangalore (Dinesh Shet, MD); Fortis Memorial Research Institute, Gurugaon (Vinod Raina, MD, B. B. Tyagi, MS); Gandhi Medical College, Bhopal (Atul Shrivastava, Sushma Shrivastava); General Hospital, Ernakulam (Balamuralikrishna, DMRT, J. Seios, MD, Sunithra Nair, MBBS); Government Medical College, Thrissur (K.L. Jayakumar, MD, Shehna A. Khader, DNB, M.B. Jayaraman, MD, V.R. Ajith Kumar, DMRT); Govt Arignar Anna Memorial Cancer Hospital & Research Institute, Kanchipuram (P. Anandhi, MD, C.T. Muthukumaran, MBBS); HCG Bangalore Institute of Oncology, Bengaluru (BS Ajaikumar, MD, B.S. Ramesh, MD); HCG NMR Cancer Centre, Hubli (Sanjay Mishra, MD); Indira Gandhi Institute of Medical Sciences, Patna (Sangeeta Pankaj, MD, Rajesh Kumar Singh, MD); Indo-American Cancer Institute & Research Centre, Hyderabad (G. Dilip Kumar, DNB, Sujith Chayu Patanaik, MS, K. Kalpana Raghunath, MBBS, A. Santa, MD); International Cancer Centre, Neyyoor (Prarthana Roselil, MD, V.G. Sudhakaran, MD); Kidwai Memorial Institute of Oncology, Bengaluru (Ramachandra, MS, C. Ramesh, PhD): Mahavir Cancer Sansthan and Research Centre, Patna (Richa Chauhan, MD, Ravi Shankar Rajendra, MD, Rita Rani, DNB, Usha Singh, MD); Malabar Cancer Centre, Kannur (T. Bindhu, K. Ratheesan, Saina Sunilkumar, MD); Mandya Institute of Medical Sciences, Mandya (S. Shiva Kumar, MS, N. Venkatesh, MD); Max Super Specialty Hospital, New Delhi (Anil Kumar Anand, MD, Ramandeep Arora, MD); Max Super Specialty Hospital, Patparganj (Geeta Kadayaprath, MS, Meenu Walia, MD); Medanta Cancer Centre, Gurugaon (Tejinder Kataria, MD, Ashok Kumar Vaid, MD); MES Medical College & Hospital, Perinthalmanna (K.V. Gangadharan, MD, Mujeeb Rahman, MS); Mizoram State Cancer Institute (Civil Hospital), Aizwal (K. Lalfakzuala, MD, Lalhlupuii, MD, Doris Lallawmzuali, MD, Cindy Lalthanpuii, MD, Jeremy L. Pautu, MD, B. Zothankima, MD); Narayana Hrudayalaya Health City, Bengaluru (Alben Sigamani, MD); North East Cancer Hospital & Research Institute, Guwahati (Gazi Naseem Ahmed, MD, Imliwati Longkumer, D.K. Nath, MD); Postgraduate Institute of Medical Education and Research, Chandigarh UT (Sushmita Ghosal, MD, R. Kapoor, MD, Narender Kumar, MD); Pravara Rural Hospital & Rural Medical College, Loni (Vandana Jain, MD, Chaitali Wagmere, MD); Rajiv Gandhi Cancer Institute and Research Centre, New Delhi (Swarima Jaitley, Sudhir Rawal, MS); Regional Cancer Centre, Raipur (Pradeep Chandrakar, MD, Vivek Choudhary), MD; Regional Cancer Centre, Agartala (Dhritiman Datta, MD, Biswajit Debbamma, MD, Partha Sarathi Sutradhar, MD, Gautam Majumdar, MD); Regional Cancer Centre, Thiruvananthapuram (Preethi Sara George, PhD, Aleyamma Mathew, PhD); Regional Cancer Centre Kamala Nehru Memorial Hospital, Allahabad (Radha Ghosh, DNB, B. Paul Thaliath, MD); Regional Cancer Centre, Indira Gandhi Medial College, Shimla (Manish Gupta, MD, Sudarshan Sharma, MD); Regional Institute of Medical Sciences, Imphal (Indibor Singh, MD, Madhubala Devi, MD, Punyabati Devi, MD, Rajesh Singh Laishram, MD); RST Regional Cancer Hospital, Cancer Relief Society, Nagpur (Anjali Kolhe, MD, B.K. Sharma, MD); Rural Development Trust, Bathalapalle (N. Hariharanadha Sarma, MD); Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow (Rajesh Harsvardhan, MD, MD, Rakesh Pandey, MD, Punitalal, MD); SDM College of Dental Sciences and Hospital, Dharwad (Swetha Acharya, MS, Kaveri Hallikeri, MS); Shakuntala Memorial Hospital & Research Centre,, Hubli (Vijay Gadagi, MD, Vinay Gadigi, MD); Sher-I-Kashmir Institute of Medical Sciences, Srinagar (Nazir Ahmad Khan, MD, Lone Mohammad Maqbool); St. Johns Medical Hospital, Bengaluru (Rakesh S Ramesh, DNB); Tata Medical Centre, Kolkata (Indranil Mallick, MD); Tata Memorial Hospital, Mumbai (B. Ganesh, PhD, R.A. Badwe, MS; The Gujarat Cancer & Research Institute, Ahmedabad (Shashank Pandya, MS, Anand Shah, MD); Vydehi Institute of Medical Sciences, Bengaluru (M.S. Ganesh, MS, Aruna E. Prasad, MD); JIPMER, Regional Cancer Centre, Puducherry UT (Biswajit Dubashi, MD, D. Kadambari, MD, Debasis Gochhait, MD, G.S. Sreenath, MD, K. Gunaseelan, K.T. Harichandrakumar, MD, Latha Chaturvedua, MD, Pampa Ch. Toi, MD, Pooja Sethi, MD, Prasanth Penumade, MD, B.H. Srinivas, MD, Sunil Kumar Saxena, MD); Government Medical College, Jammu (Dinesh Kumar, MD, Ashutosh Gupta, MD).\n"}
+{"text": "There was an error in the y-axis label of Figure 2B. The corrected figure is available here:\n\n![](pone.21d1a1f6-35e9-4935-a8d0-387c057f1469.g001){#pone-21d1a1f6-35e9-4935-a8d0-387c057f1469-g001}\n"}
+{"text": "GENOME ANNOUNCEMENT {#h0.0}\n===================\n\n*Bacillus* is one of the most extensively studied microorganisms among plant growth-promoting rhizobacteria (PGPR). Owing to the production of various antibiotics and the formation of stress-resistant endospores, *Bacillus* strains are more amenable to the formulation of commercial products and have been widely used as biocontrol agents ([@B1]). Induced systemic resistance (ISR) elicited by *Bacillus* strains has been recently recognized as one of the key mechanisms by which crops can protect themselves against phytopathogens with the aid of PGPR ([@B2], [@B3]).\n\nThe four *Bacillus* strains sequenced in this study, all purchased from the Korean Collection for Type Cultures (KCTC), were chosen based on the partial results of a systematic screening approach used to search for plant probiotic bacteria. The three type strains, *Bacillus endophyticus* KCTC 13922 ([@B4]), \"*Bacillus\u00a0gaemokensis*\" KCTC 13318 ([@B5]), and *Bacillus\u00a0trypoxylicola* KCTC 13244 ([@B6]), were shown to promote plant growth (*Arabidopsis thaliana*) by volatiles, whereas *Bacillus* sp. strain KCTC 13219 (=b04i-3) triggers ISR in *A. thaliana* against *Pectobacterium carotovorum* (our unpublished data). The type strains *B.\u00a0gaemokensis* JCM 15801 and *B.\u00a0trypoxylicola* NRBC 102646, obtained from the Japan Collection of Microorganisms (JCM) and the NITE Biological Resource Center (NBRC), respectively, were sequenced by other research groups and made available through accession numbers JOTM00000000 ([@B7]) and BCWA00000000 while this paper was being prepared.\n\nGenome sequencing was carried out using the Illumina HiSeq 2000 platform at the National Instrument Center for Environmental Management at Seoul National University (Seoul, Republic of Korea). One hundred one-nucleotide paired reads produced from a library with a fragment size of ca. 500 bp were pretreated using Trimmomatic version 0.32 ([@B8]), and 600-Mb subsamples were randomly extracted from them. After k-mer-based error correction using SGA version 0.10.13 ([@B9]), *de novo* genome assembly was conducted using the A5-miseq pipeline ([@B10]). The assembled sequences were annotated using the RAST server ([@B11]) and Prokaryotic Genome Annotation Pipeline from the NCBI. Biosynthetic genes for secondary metabolites were predicted using antiSMASH 3.0 ([@B12]). The list of strains used in this study, the sequencing and assembly statistics, and the accession numbers are all provided in [Table\u00a01](#tab1){ref-type=\"table\"}. Compared with the previously published records, the genome sequences of KCTC 13318 and KCTC 13244 obtained through this study were shown to be superior in terms of the assembly statistics.\n\nKCTC 13219, isolated from Pu-erh tea, was tentatively named \"*Bacillus nitroreducens*\" by the submitter but did not lead to a proposal for a novel species (J. S. Lee, personal communication). Very recently, the species *Bacillus nitroreducens* sp. nov. was proposed for another unrelated bacterium ([@B13]). Phylogenetic analysis using the 16S rRNA sequence revealed that KCTC 13219 should be classified into the genus *Lysinibacillus* (99.93% similarity with \"*Lysinibacillus fluoroglycofernilyticus*\" cmg86^T^). The genome sequencing results will provide insight into the genomic basis of the versatile *Bacillus* strains and the interspecies interaction between soil bacteria and plants, which can lead to eco-friendly agricultural applications. Additional information is available at <http://genoglobe.kr/kribb/four_bacillus_strains_2016>.\n\nNucleotide sequence accession numbers. {#s1}\n--------------------------------------\n\nThese whole-genome shotgun projects have been deposited at DDBJ/EMBL/GenBank, and the accession numbers for all four genome sequences are listed in [Table\u00a01](#tab1){ref-type=\"table\"}.\n\n###### \n\nSummary of genome sequencing results\n\n  *Bacillus* species                 Draft coverage (\u00d7)[^a^](#ngtab1.1){ref-type=\"table-fn\"}   Genome size (bp)   No. of contigs   G+C content (%)   No. of secondary metabolite biosynthesis gene clusters[^b^](#ngtab1.2){ref-type=\"table-fn\"}   Accession no.\n  ---------------------------------- --------------------------------------------------------- ------------------ ---------------- ----------------- --------------------------------------------------------------------------------------------- ------------------------------------------------------------------\n  *B.\u00a0endophyticus* KCTC 13922^T^    741                                                       5,121,484          42               36.50             11                                                                                            [LTAP00000000](http://www.ncbi.nlm.nih.gov/nuccore/LTAP00000000)\n  *B.\u00a0gaemokensis* KCTC 13318^T^     878                                                       5,616,250          70               35.54             18                                                                                            [LTAQ00000000](http://www.ncbi.nlm.nih.gov/nuccore/LTAQ00000000)\n  *B.\u00a0trypoxylicola* KCTC 13244^T^   9,024                                                     4,347,941          40               35.73             6                                                                                             [LTAO00000000](http://www.ncbi.nlm.nih.gov/nuccore/LTAO00000000)\n  *Bacillus* sp. KCTC 13219          1,244                                                     3,853,058          37               38.31             3                                                                                             [LUFJ00000000](http://www.ncbi.nlm.nih.gov/nuccore/LUFJ00000000)\n\nSubsamples of 600 Mb were randomly taken from pretreated reads and assembled.\n\nPredicted by antiSMASH 3.0.\n\n**Citation** Jeong H, Park S-H, Choi S-K. 2016. Draft genome sequences of four plant probiotic *Bacillus* strains. Genome Announc 4(3):e00358-16. doi:10.1128/genomeA.00358-16\n\nWe would like to thank Jung-Sook Lee (KCTC) for her helpful discussions.\n\nThis work was supported by the Industrial Source Technology Development Program of the Ministry of Trade, Industry, and Energy (grant 10044909), and by the KRIBB Research Initiative Program, Ministry of Science, ICT and Future Planning, Republic of Korea.\n"}
+{"text": "1. Introduction {#sec1-molecules-23-00543}\n===============\n\nIn recent years, cancer incidence rates have risen worldwide. Currently, the health problems caused by cancer treatment are receiving increased attention. Chemotherapy is an important method of therapy in the treatment of malignancies. Different chemotherapeutic agents have different levels of myelosuppression, which may limit the dose that can be used, and can easily result in interruption of chemotherapy. This serious side effect can significantly affect the treatment of malignancies and decrease the quality of a patient's life. At present, granulocyte colony-stimulating factor is used as adjuvant therapy for the reduction of neutropenia. However, granulocyte colony-stimulating factor also has side effects, including bone pain, and flushing \\[[@B1-molecules-23-00543],[@B2-molecules-23-00543],[@B3-molecules-23-00543],[@B4-molecules-23-00543],[@B5-molecules-23-00543]\\].\n\nIn the clinic, extracts from traditional Chinese medicine have received increased attention for their favorable pharmacodynamics effects and low toxicity. In China, burnet root leukopoietic tablets (BRLT) have been widely used due to their beneficial effects on leucopenia caused by tumor-targeting chemotherapy. In previous studies, it was demonstrated that in patients who were given BRLT, the incidence of chemotherapy-induced myelosuppression, was significantly reduced \\[[@B6-molecules-23-00543],[@B7-molecules-23-00543]\\]. BRLT are prepared from *Sanguisorba officinalis* L., a traditional Chinese herb that has been used in China for over 3000 years. BRLT have shown beneficial effects in a variety of disease, including infection, inflammatory, cancer, allergic, and diseases of the central nervous system \\[[@B8-molecules-23-00543]\\]. Previous studies have demonstrated that *Sanguisorba* contains bioactive constituents of saponins and tannins \\[[@B9-molecules-23-00543]\\]. Several clinical reports have demonstrated a pre-protective role of *S. officinalis* on leucopenia \\[[@B10-molecules-23-00543],[@B11-molecules-23-00543]\\]. Moreover, it has been shown that saponins extracted from this plant reduce myelosuppression and 60 Co-\u03b3-irradiation in mice \\[[@B12-molecules-23-00543]\\]. ZGS1 is the main saponin exacted from the famous traditional Chinese medicine \\[[@B13-molecules-23-00543]\\], and has the ability to inhibit apoptosis of TF-1 cells, reduce mouse cytokine production, and promote the survival of focal adhesion kinase (FAK) mouse bone marrow nuclear cells in vitro \\[[@B14-molecules-23-00543],[@B15-molecules-23-00543],[@B16-molecules-23-00543],[@B17-molecules-23-00543]\\]. In our previous study, we showed that ZGS2 is a metabolite of ZGS1 that protects against the appearance of marrow repression during tumor chemotherapy. Therefore, ZGS1 has been chosen as a novel leucopenia drug and needs further research. Currently, limited information is available on ZGS1 and its metabolite ZGS2 in vivo. Therefore, it is of great importance to investigate these compounds and their pharmacokinetics.\n\nThe pharmacokinetic profile of the prototype and their metabolites is a critical estimation of drug candidates. A novel quantitative method for ZGS1 and ZGS2 was established using an Agilent 6460 instrument in this study. This is the first profile study describing the absorption, distribution, and excretion of ZGS1 and ZGS2, and the first time that the pharmacokinetic profile of ZGS1 and its metabolite ZGS2 were investigated. This study provides novel insights for future clinical practice of ZGS1 and ZGS2, which may be promising for the development of ZGS1-based therapeutics in treating leucopenia.\n\n2. Results and Discussion {#sec2-molecules-23-00543}\n=========================\n\n2.1. Method Development {#sec2dot1-molecules-23-00543}\n-----------------------\n\nTo obtain better and interference-free extracts, protein precipitation was used and several solvents were studied. During the precipitation procedure, the use of methanol, acetonitrile, and methanol containing 0.1% FA were investigated. Among the solvents, the addition of 0.1% FA to methanol significantly promoted the recovery of analytes in biological samples. Moreover, the use of the XTERRA MS column gave a sharp peak for ZGS1, ZGS2, and the IS. The chemical structures of all compounds are presented in [Figure 1](#molecules-23-00543-f001){ref-type=\"fig\"}. In the chromatographic separation of analytes, acetonitrile gave a better peak shape and had a lower background when compared to methanol. In comparison with acetic acid and ammonium acetate, the addition of 0.1% FA in methanol significantly enhanced the shapes of the peaks. Therefore, acetonitrile and water (with 0.2% FA) were adopted as the mobile phases. In addition, a gradient elution was chosen to reduce interference of the endogenous substance.\n\nMS parameters were optimized by injecting standard solutions into the MS. For the optimization of ESI conditions for ZGS1, ZGS2, and IS, quadrupole full scans were performed in both positive (+) and negative (\u2212) ion detection modes. Due to the higher sensitivity and a more stable response, the negative mode was more suitable for ZGS1, ZGS2, and the IS. To ascertain precursor ions, ZGS1, ZGS2, and the IS were initially characterized by a tandem mass spectrometry Q1 (full mass spectra) scan. Amproduct scan was used for the selection of suitable product ions, and was used in the MRM mode. Full mass spectra of ZGS1, ZGS2, and the IS, as well as their MS/MS spectra are shown in [Figure 2](#molecules-23-00543-f002){ref-type=\"fig\"}. The MRM transitions of ZGS1, ZGS2, and the IS were *m*/*z* 811.4\u2192602.9, *m*/*z* 603.5\u2192585.3, and *m*/*z* 749.3\u2192471.1, respectively. The fragment and CE were optimized to improve the sensitivity of the detection of the precursor and product ions ([Table 1](#molecules-23-00543-t001){ref-type=\"table\"}). Other mass spectrometry-related parameters, such as ion spray voltage (ISV), capillary temperature (CT), nebulizer gas (GS1), auxiliary gas (GS2), curtain gas (CUR), and collision gas (CAD) were also optimized to achieve a better response.\n\n2.2. Internal Standard {#sec2dot2-molecules-23-00543}\n----------------------\n\nAfter evaluating the retention time, peak shape, response, and MF, \u03b1-hederin was chosen as the IS, which met the requirements of the analysis.\n\n2.3. Method Validation {#sec2dot3-molecules-23-00543}\n----------------------\n\n### 2.3.1. Calibration Curve, LLOQ, Carryover, and Specificity {#sec2dot3dot1-molecules-23-00543}\n\nOver a range of 5--2000 ng/mL, both ZGS1 and ZGS2 showed a good linearity. Calibration curves were evaluated using a quadratic equation with a weighting factor: WF = 1/*x*^2^. The correlation coefficient of all calibration curves was \\>0.99 and the concentration of the samples used for preparation of the calibration curve was within 85--115% of the expected value ([Table 2](#molecules-23-00543-t002){ref-type=\"table\"} and [Table 3](#molecules-23-00543-t003){ref-type=\"table\"}). Representative specificity MS chromatograms for blank and LLOQ plasma samples are shown in [Figure 3](#molecules-23-00543-f003){ref-type=\"fig\"}. When blank plasma sample extracts were injected immediately after injection of the ULOQ sample, no significant carryover was observed. Moreover, no significant interference from endogenous substances was observed, indicating that the specificity of the method was sufficient.\n\n### 2.3.2. Precision and Accuracy {#sec2dot3dot2-molecules-23-00543}\n\nAs shown in [Table 4](#molecules-23-00543-t004){ref-type=\"table\"}, the results obtained from the method validation met the FDA criteria \\[[@B18-molecules-23-00543]\\]. The Intra-day and Inter-day accuracy of all analytes was \\<6%. In addition, the precision of all analytes ranged from 0.81% to 4.48%, indicating that the method had a good precision and accuracy.\n\n### 2.3.3. Recovery and Matrix Effect {#sec2dot3dot3-molecules-23-00543}\n\n[Table 5](#molecules-23-00543-t005){ref-type=\"table\"} summarizes the extraction recovery (ER) and ME of ZGS1 and ZGS2 in rat plasma. The ER of all analytes ranged from 96% to 102%, indicating that the recoveries met the requirements. ME were evaluated by using MF and IS-normalized MF. Among the three evaluated QC levels, the IS-normalized MF ranged from 100% to 107%. These findings indicated that there were no problems in the MF of analytical methods of ZGS1 and ZGS2.\n\n### 2.3.4. Stability {#sec2dot3dot4-molecules-23-00543}\n\nAs shown in [Table 6](#molecules-23-00543-t006){ref-type=\"table\"}, the analytes were stable after incubating at room temperature for at least 24 h (short-term stability), three freeze-thaw cycles, and after storage for 30 days at \u221280 \u00b0C (long-term stability). Thus, these results demonstrated that ZGS1 and ZGS2 are stable at different storage conditions.\n\n### 2.3.5. Dilution Integrity {#sec2dot3dot5-molecules-23-00543}\n\nThe accuracy and precision of the dilution reliability of ZGS1 in plasma were 6.58% and 3.48%, respectively. This indicated that samples have a higher concentration than the ULOQ could predict and quantitatively in analysis after dilution with a blank matrix.\n\n2.4. Pharmacokinetic Study {#sec2dot4-molecules-23-00543}\n--------------------------\n\nThe pharmacokinetic (PK) study of ZGS1 and ZGS2 was performed in ten male SD rats weighing 150--200 g. The PK parameters were evaluated by a non-compartmental model (DAS 3.0 software). PK parameters represented the average values of 5 rats ([Figure 4](#molecules-23-00543-f004){ref-type=\"fig\"} and [Figure 5](#molecules-23-00543-f005){ref-type=\"fig\"}). The half-life (t~1/2~) of tail vein-administrated ZGS1 was 1.338 \u00b1 0.744 h and 1.027 \u00b1 0.057 h for the active metabolite ZGS2. Moreover, t~1/2~ of subcutaneously administered ZGS1 was 6.115 \u00b1 1.92 and 7.935 \u00b1 3.264 h for ZGS2 ([Table 7](#molecules-23-00543-t007){ref-type=\"table\"} and [Table 8](#molecules-23-00543-t008){ref-type=\"table\"}). The relative bioavailability of ZGS1 when administered subcutaneously was 98.82%. These results indicated that administration via subcutaneous injection prolonged the extension time of ZGS1 in vivo when compared to intravenous administration and increased the AUC~0--t~ of the active metabolite ZGS2. The calculation formula of AUC, MRT and t~1/2~ is as follows: $${{AUC}_{({0 - t})} = \\int_{0}^{t^{*}}{Cdt} + \\int_{t^{*}}^{\\infty}{Cdt}\\ t\\frac{1}{2}z = 0.693/K}\\ {{MRT}_{({0 - t})} = \\frac{\\int_{0}^{\\infty}{tcdt}}{\\int_{0}^{\\infty}{cdt}} = \\frac{AUMC}{AUC}}$$\n\n2.5. Tissue Distribution {#sec2dot5-molecules-23-00543}\n------------------------\n\nAdministration by subcutaneous injection was chosen to evaluate the distributions of ZGS1 and ZGS2. The distribution of ZGS1 and ZGS2 in different tissues at 5, 20, and 180 min after subcutaneous administration of ZGS1 were assessed ([Figure 6](#molecules-23-00543-f006){ref-type=\"fig\"} and [Figure 7](#molecules-23-00543-f007){ref-type=\"fig\"}). The standard curves of ZGS1 and ZGS2 in different tissues are shown in [Table 2](#molecules-23-00543-t002){ref-type=\"table\"} and [Table 3](#molecules-23-00543-t003){ref-type=\"table\"}. ZGS1 was rapidly and widely distributed in most organs, except for the brain.\n\nLevels of ZGS1 and ZGS2 were lower in brain tissue, when compared to other tissues, suggesting that ZGS1 and ZGS2 do not easily cross the blood brain barrier (BBB). In addition, the highest levels of ZGS1 and ZGS2 were observed in the intestine, followed by the liver, and kidney at 20 min.\n\n2.6. Excretion Study {#sec2dot6-molecules-23-00543}\n--------------------\n\nThe excretion of ZGS1 and ZGS2 in rat urine, feces, and bile was determined after subcutaneous administration of ZGS1 (10 mg/kg) ([Figure 8](#molecules-23-00543-f008){ref-type=\"fig\"}). The mean recoveries of unchanged ZGS1 were 5.990%, 0.172%, and 0.074% of the 10 mg/kg dose up to 24 h in bile, urine, and fecal samples, respectively.\n\nThese results indicated that ZGS1 was primarily excreted by the bile, and supported the results that ZGS1 was mostly present in rat intestines. However, further studies are required to investigate the in vivo biotransformation of ZGS1.\n\n3. Materials and Methods {#sec3-molecules-23-00543}\n========================\n\n3.1. Chemical, Reagents, and Materials {#sec3dot1-molecules-23-00543}\n--------------------------------------\n\nZGS2 and \u03b1-hederin (internal standard (IS), HPLC \\> 98%) were extracted and purified in-house. ZGS1 (HPLC \\> 98%) was purchased from Raffensi Biotechnology (Chengdu, China). Methanol and acetonitrile (HPLC-MS grade) were purchased from Thermo Fischer Scientific (Fair Lawn, NJ, USA). Formic acid (mass spectrometer grade) was purchased from Aladdin Corporation (Shanghai, China), and Milli-Q ultrapure water was obtained from a Millipore system (Milford, MA, USA).\n\n3.2. LC-MS Conditions {#sec3dot2-molecules-23-00543}\n---------------------\n\nThe liquid chromatography (LC) system used was comprised of a binary pump, a constant temperature column chamber, an auto sampler, and a diode array detector (DAD). An XTERRA MS C~18~ reversed phase column (2.1 \u00d7 50 mm, 5 \u00b5m) was used for the determination of analytics. The temperature of the column was maintained at 40 \u00b0C. The mobile phase A (MA) was water, containing 0.2% formic acid (FA) (*v*/*v*) and mobile phase B (MB) contained acetonitrile. The flow velocity was set to 0.4 mL/min. In 6 min, the gradient elution was set to: 0.0--0.1 min, 10% MB; 0.1--1 min, 10% MB; 1--1.5 min, 10--57% MB; 1.5--2.0 min, 57--68% MB; 2.0--3.2 min, 68--75% MB; 3.2--4.0 min, 75--95% MB; 4.0--4.5 min, 95--10% MB; 4.5--6 min, 10% MB. The injection volume was 2 \u00b5L. For better stability of the samples, the auto-sampler was kept at a temperature of 4 \u00b0C.\n\nAn Agilent 1290 Infinity Rapid Resolution Liquid Chromatography (RRLC) System (Agilent, Lexington, MA, USA), coupled to an Agilent 6460 triple quadrupole mass spectrometer was used for quantitative measurements. Electrospray ionization (ESI) negative and multiple reaction monitoring (MRM) ion mode was used for detection of the ions. The optimal conditions for mass spectrometry were as follows: capillary voltage (CV) 3000 V, gas flow rate 5 L/min, dry gas temperature 300 \u00b0C, sheath gas temperature 250 \u00b0C, dry flow rate 7 L/min, atomizer pressure 45 psi, and instrument use process using nitrogen. For ZGS2, the precursor-product ion pair was *m*/*z* 811.4\u2192602.9, for ZGS1, *m*/*z* 603.4\u2192585.2, and for \u03b1-hederin (IS) *m*/*z* 749.3\u2192471.1 ZGS1, ZGS2, and IS fragments were obtained at \u2212237 V, \u2212330 V, and \u2212320V, respectively. The collision energy (CE) of ZGS1, ZGS2, and IS was \u221227 eV, \u221235 eV, and \u221250 eV, respectively. MassHunter (version B.04) software was used for analysis and data acquisition. Multiple-reaction monitoring was used for further detection. The detailed characteristics are shown in [Table 1](#molecules-23-00543-t001){ref-type=\"table\"}.\n\n3.3. Sample Preparation {#sec3dot3-molecules-23-00543}\n-----------------------\n\nFor preparation of the stock solution (SS), calibration standards, and quality control (QC) samples, ZGS1, ZGS2 as well as \u03b1-hederin were dissolved in methanol to a final concentration of 1 mg/mL. Working solutions of calibration standards and QC samples of ZGS1 and ZGS2 were prepared by dilution with methanol (containing 0.1% FA). Similarly, the \u03b1-hederin IS was further diluted with methanol (containing 0.1% FA) to a final concentration of 800 ng/mL. For ZGS1 and ZGS2, two stock solutions were prepared, one for the calibration standards, the other for QC samples. Stock solutions were stored at 4 \u00b0C.\n\nTo achieve a calibration concentration range from 5 to 2000 ng/mL, including 5, 20, 50, 100, 500, 1000, and 2000 ng/mL, stock and working solutions were added to biological samples. For ZGS1 and ZGS2, QC samples of 10, 800, and 1600 ng/mL were prepared. Working solutions for calibration standards of ZGS1 and ZGS2 and QC samples of ZGS1 and ZGS2 were stored at 4 \u00b0C for future use.\n\n3.4. Preparation of Biological Samples {#sec3dot4-molecules-23-00543}\n--------------------------------------\n\nA total of 20 \u03bcL IS working solution was added to 50 \u03bcL plasma, urine, and bile samples. Then, 330 \u00b5L methanol (containing 0.1% FA) was added to precipitate the proteins in the samples. After vortex-mixing for 5 min, samples were centrifuged at 12,000 rpm for 5 min at 4 \u00b0C. Then, HPLC-MS/MS analysis was used to analyze the supernatant (200 \u00b5L). Tissue samples were homogenized in water at a ratio of 1:4 (*w*/*v*), while pulverized 0.2 g of fecal material was homogenized with water (1:5, *w*/*v*). A total of 50 \u03bcL suspension was analyzed.\n\n3.5. Method Validation {#sec3dot5-molecules-23-00543}\n----------------------\n\nThe methodology used in this study was performed as per FDA guidelines \\[[@B19-molecules-23-00543]\\]. The HPLC-MS/MS analysis for ZGS1 and ZGS2 was validated at a range of 5--2000 ng/mL. Parameters of selectivity, including a lower limit of quantification (LLOQ), matrix effects (ME), precision, accuracy, recovery, stability, and dilution reliability of this method were fully validated.\n\nThe specificity of endogenous substances and analytes in rat plasma was determined by determining the chromatograms of plasma samples in six rats. The standard plasma samples used for creating of the calibration curves were assayed. By plotting the ratio of analyte/IS peak area versus the ratio of the nominal concentration, calibration curves were constructed using linear regression (y = ax + b, weighting factor of 1/*x*^2^). Acceptance criteria for the LLOQ were \u00b120% with a correlation coefficient (r^2^) greater than 0.99. In addition, the acceptable limit of the relative standard deviation (RSD) for each point in the standard curve was \u00b115%.\n\nPrecision and accuracy (Intra-day) assay were evaluated by analyzing replicates (*n* = 6) at QC levels of 10 (Low quality control, LQC), 800 (Middle quality control, MQC) and 1600 (High quality control, HQC) ng/mL, respectively. Moreover, precision and accuracy (Inter-day) from QCs samples were evaluated at 3 consecutive days. Sample precision and accuracy were assessed by RSD and relative error (RE). The maximum acceptable deviation was \u00b115%.\n\nTo determine the extraction recovery, the peak area ratio extracted from 10 ng/mL (LQC), 800 ng/mL (MQC), and 1600 ng/mL (HQC) QC samples as well as the peak areas that ZGS1 and ZGS2 spiked to the blank sample extracts were calculated at the same concentrations. The ME were determined by matrix factor (MF), which presented the ratio of the ZGS1 and ZGS2 peak response in the presence of matrix ions to the analyte peak response in methanol. Experiments were performed in six replicates at 10 ng/mL, 800 ng/mL and 1600 ng/mL levels.\n\nThe stability of ZGS1 and ZGS2 in rat plasma was assessed by analyzing replicates (*n* = 5) at LQC, MQC, and HQC levels. The short-term stability of the standards and QC samples of ZGS1 and ZGS2 was assessed by evaluating the samples after 24 h at room temperature. The freeze--thaw stability was evaluated by subjecting LQC, MQC, and HQC samples to three freeze--thaw cycles prior to extraction. Samples were stored at \u221280 \u00b0C for 30 days. Then, samples were thawed to room temperature, and analyzed by the approach described above. The analytes were considered stable in plasma when the concentrations were within 85--115% of the initial concentration.\n\nTo obtain ZGS1 and ZGS2 samples within the calibration range of 5--2000 ng/mL, samples were diluted using the highest concentration (20,000 ng/L 10-times, the upper limit of quantification (ULOQ)) of the QC sample, which was diluted with plasma.\n\n3.6. Pharmacokinetic Data Analysis {#sec3dot6-molecules-23-00543}\n----------------------------------\n\nTen male Sprague-Dawley (SD) rats weighing 150--200 g were purchased from the Hunan Shrek Laboratory Animal (Changsha, China). All experimental procedures involving animals were reviewed and approved by the Animal Care and Use Committee (ACUC) of Jiangxi University of Traditional Chinese Medicine (Nanchang, China) and were in accordance with the Guide for the Care and Use of Laboratory Animals (GCULA). Previous pharmacodynamic studies have shown that in rats, administration of 10 mg/kg ZGS1 has the best beneficial effects. Therefore, we chose to use this dose for our future studies. The general health status of rats was observed daily. Rats were fasted for 12 h but had with free access to water prior to experiments. Rats were randomly divided into two groups, one group received tail vein injection 10 mg/kg ZGS1, the other group received a subcutaneous injection of 10 mg/kg ZGS1. Approximately 0.15 mL of rat blood samples were collected from the orbital venous plexus at the following time points: pre-dose, 0.083, 0.333, 0.666, 1, 1.333, 1.667, 2, 3, 4, and 6 h. Plasma was prepared by centrifugation at 4000\u00d7 *g* for 5 min at 4 \u00b0C. The supernatant was separated and stored at \u221240 \u00b0C until sample preparation and analysis. Non-compartmental analysis of the data was acquired and processed by using DAS 3.0 software (Chinese Pharmacological Society). Results are expressed as the mean \u00b1 standard deviation (SD).\n\n3.7. Tissue Distribution Study {#sec3dot7-molecules-23-00543}\n------------------------------\n\nTo study the target distribution of ZGS1 and ZGS2, tissue distribution was evaluated. In brief, 15 rats (150--200 g) were randomly divided into three groups. After subcutaneous injection of 10 mg/kg of ZGS1, 11 tissues, including liver, heart, spleen, lung, brain, intestine, stomach, kidney, testis, skeletal muscle, and fat were collected at 5, 20, or 180 min. Tissues were immediately rinsed with normal saline solution, dried, and accurately weighed. Samples were homogenized following the methods described above and stored at \u221240 \u00b0C until LC-MS analysis.\n\n3.8. Excretion Study {#sec3dot8-molecules-23-00543}\n--------------------\n\nFor the excretion study, 12 rats were randomly divided into two groups. One group was used for the urinary and fecal excretion study, whereas the other group was used for evaluating bile excretion. All animals had free access to food and water. Blank samples for urine, feces, and bile were collected before animals were treated. After subcutaneous injection of 10 mg/kg of ZGS1, rats were housed in metabolic cages, and urine and fecal samples were collected at 0--6 h, 6--12 h, 12--24 h, and 24--48 h. Fecal samples were dried at room temperature. After the dry weight of fecal samples and the volume of urine for each collection point were measured, specimens were stored at \u221240 \u00b0C. Rats in the group for bile excretion evaluation were anesthetized with 20% urethane and a cannula was implanted into the bile duct to collect bile. Bile samples were collected at 0--4 h, 4--8 h, 8--12 h, and 12--24 h, and stored at \u221240 \u00b0C.\n\nInvestigating the QC samples in the tissue homogenates was sufficient and equivalent to the peak area measured by the plasma samples under the same condition. Therefore, the above findings indicated that the method of plasma samples can be used to determine the content of ZGS1 and ZGS2 in other tissues.\n\n4. Conclusions {#sec4-molecules-23-00543}\n==============\n\nIn this study, a simple and sensitive HPLC--MS/MS method was developed for the quantification of ZGS1 and ZGS2 in rat biological matrices. The PK, tissue distribution, and excretion of ZGS1 and ZGS2 were evaluated. In this study, we demonstrated that ZGS1 and ZGS2 can rapidly be cleared from rat plasma following tail vein injection and subcutaneous injection of ZGS1. The major organs of ZGS1 and ZGS2 distribution included the intestines, liver, and kidney. The total recovery of ZGS1 in bile, urine, and feces were low, and may potentially be due to in vivo biotransformation. In vivo processing of the ZGS2 metabolite may form the basis of the effect of ZGS1 on leucopenia.\n\nAccording to the novel drug Research and Development requirement of the FDA and the CFDA (China), the PK profile of the prototype and the metabolite is acritical measure in the search for drug candidates. Several studies have focused on PK profile of prototypes and metabolites \\[[@B20-molecules-23-00543],[@B21-molecules-23-00543],[@B22-molecules-23-00543],[@B23-molecules-23-00543]\\]. Previous studies have suggested that, ZGS2 is the main metabolite of ZGS1, and to better understand its beneficial effects in vivo may help to increase the success rate of potential drugs. In our previous study, we indicated that ZGS1 and ZGS2 possessed remarkable activity on myelosuppression mice. To our knowledge, this is the first time that the PK profile of ZGS1 and its metabolite ZGS2 were studied. Therefore, ZGS1 is a potential candidate for the development of novel drugs to treat leucopenia.\n\nThis study was funded by National Natural Science Foundation of China (No. 81560638), Young Scientists Training Project of Jiangxi Province (No. 20142BCB23022), China Postdoctoral Science Foundation (No. 2017M612159), Jiangxi Province 5511 R&D projects (No. 20165BCB19009), Nanchang innovative talent team (No. \\[2016\\]173).\n\n**Sample Availability:** Samples of the compounds ziyuglycoside I and ziyuglycoside II are available from the authors.\n\nShi-Lin Yang, Yu-Lin Feng, Hui Ouyang conceived and designed the experiments; Zhi-Feng Li, Meng-Ying Zhou, Ting Tan, Chen-Cong Zhong analysed the data, wrote the paper, contributed equally to this work; Qi Wang, Ling-Ling Pan and Ying-Ying Luo performed the experiments; Zhi-Feng Li, Meng-Ying Zhou and Hui Ouyang wrote the paper.\n\nThe authors declare no conflict of interest.\n\n![Chemical structures of (**A**) ZGS1; (**B**) ZGS2; and (**C**) \u03b1-hederin (internal standard).](molecules-23-00543-g001){#molecules-23-00543-f001}\n\n![Fragment ion spectra of (**A**) ZGS1; (**B**) ZGS2; and (**C**) \u03b1-hederin (internal standard).](molecules-23-00543-g002){#molecules-23-00543-f002}\n\n![Representative MRM chromatograms of ZGS1, \u03b1-hederin (internal standard \\[IS\\]) and ZGS2 in (**A**) blank plasma sample; (**B**) blank plasma sample spiked with LLQC of ZGS1 (10 ng/mL) and ZGS2 (10 ng/mL), and IS (800 ng/mL); (**C**) rat plasma sample obtained 1h after subcutaneous injection of ZGS1 (10 mg/kg).](molecules-23-00543-g003){#molecules-23-00543-f003}\n\n![Mean plasma concentration-time profiles of ZGS1 and ZGS2 after tail vein administration of ZGS1 (10 mg/kg) (*n* = 5).](molecules-23-00543-g004){#molecules-23-00543-f004}\n\n![Mean plasma concentration-time profiles of ZGS1 and ZGS2 after subcutaneous injection administration of ZGS1 (10 mg/kg) (*n* = 5).](molecules-23-00543-g005){#molecules-23-00543-f005}\n\n![Concentration (mean \u00b1 SD) levels of ZGS1 in rat tissue at 5, 20, and 180 min after subcutaneous injection injection of ZGS1 (10 mg/kg) (*n* = 5).](molecules-23-00543-g006){#molecules-23-00543-f006}\n\n![Concentration (mean \u00b1 SD) levels of ZGS2 in rat tissue at 5, 20, and 180 min after subcutaneous injection of ZGS1 (10 mg/kg) (*n* = 5).](molecules-23-00543-g007){#molecules-23-00543-f007}\n\n![(**A**) Biliary; (**B**) urinary; and (**C**) fecal cumulative excretion profiles of ZGS1 and ZGS2 in rats after subcutaneous injection of ZGS1 (10 mg/kg) (*n* = 5).](molecules-23-00543-g008){#molecules-23-00543-f008}\n\nmolecules-23-00543-t001_Table 1\n\n###### \n\nRetention time (t~R~), optimized MS/MS transitions, fragmentor voltage (FV) and collision energy (CE) for each analyte.\n\n  Analytes           t~R~ (min)   Precursorion   Production   FV (eV)   CE (eV)\n  ------------------ ------------ -------------- ------------ --------- ---------\n  Ziyuglycoside I    2.08         811.4          602.9        237       27\n  Ziyuglycoside II   2.39         603.5          585.3        330       35\n  \u03b1-hederin (IS)     2.32         749.3          471.1        320       50\n\nmolecules-23-00543-t002_Table 2\n\n###### \n\nLinearity and Linear range of ZGS1 calibration curves in different biological matrix.\n\n  Matrix       Calibration Curve    Correlation Coefficient (r^2^)   Linear Range (ng/mL)\n  ------------ -------------------- -------------------------------- ----------------------\n  plasma       y = 0.010x \u2212 0.031   0.999                            5--2000\n  brain        y = 0.001x \u2212 0.003   0.999                            5--2000\n  testis       y = 0.001x + 0.001   0.999                            5--2000\n  spleen       y = 0.001x \u2212 0.009   0.998                            5--2000\n  kidney       y = 0.001x + 0.001   0.999                            5--2000\n  heart        y = 0.001x \u2212 0.007   0.999                            5--2000\n  intestinal   y = 0.001x \u2212 0.006   0.998                            5--2000\n  muscle       y = 0.001x \u2212 0.003   0.999                            5--2000\n  liver        y = 0.001x \u2212 0.003   0.999                            5--2000\n  lung         y = 0.001x \u2212 0.007   0.999                            5--2000\n  stomach      y = 0.001x \u2212 0.005   0.999                            5--2000\n  fat          y = 0.001x \u2212 0.012   0.999                            5--2000\n  bile         y = 0.001x \u2212 0.005   0.998                            5--2000\n  Urine        y = 0.001x \u2212 0.011   0.999                            5--2000\n  feces        y = 0.002x + 0.010   0.999                            5--2000\n\nmolecules-23-00543-t003_Table 3\n\n###### \n\nLinearity and Linear range of ZGS2 calibration curves in different biological matrix.\n\n  Matrix       Calibration Curve    Correlation Coefficient (r^2^)   Linear Range (ng/mL)\n  ------------ -------------------- -------------------------------- ----------------------\n  plasma       y = 0.010x + 0.569   0.996                            5--2000\n  brain        y = 0.001x + 0.057   0.993                            5--2000\n  testis       y = 0.001x + 0.040   0.996                            5--2000\n  spleen       y = 0.001x + 0.043   0.996                            5--2000\n  kidney       y = 0.001x + 0.054   0.992                            5--2000\n  heart        y = 0.001x + 0.065   0.994                            5--2000\n  intestinal   y = 0.001x + 0.039   0.995                            5--2000\n  muscle       y = 0.001x + 0.052   0.992                            5--2000\n  liver        y = 0.001x + 0.059   0.991                            5--2000\n  lung         y = 0.001x + 0.041   0.996                            5--2000\n  stomach      y = 0.001x + 0.042   0.996                            5--2000\n  fat          y = 0.001x + 0.039   0.997                            5--2000\n  bile         y = 0.001x + 0.017   0.998                            5--2000\n  Urine        y = 0.001x \u2212 0.007   0.999                            5--2000\n  feces        y = 0.001x + 0.011   0.999                            5--2000\n\nmolecules-23-00543-t004_Table 4\n\n###### \n\nIntra- and inter-day accuracy and precision of ZGS1 and ZGS2 in rat plasma (*n* = 5).\n\n  Compound           Concentration (ng/mL)   Intra-Day (*n* = 5)   Inter-Day (*n* = 5)          \n  ------------------ ----------------------- --------------------- --------------------- ------ ------\n  Ziyuglycoside I    10                      4.81                  2.42                  6.00   4.12\n  800                1.89                    1.14                  2.41                  2.65   \n  1600               1.23                    3.31                  1.00                  4.48   \n  Ziyuglycoside II   10                      1.28                  4.36                  1.76   3.99\n  800                1.75                    0.81                  2.77                  0.95   \n  1600               1.18                    1.94                  3.58                  1.02   \n\nmolecules-23-00543-t005_Table 5\n\n###### \n\nExtraction recovery and matrix effect of ZGS1 and ZGS2 in rat plasma (*n* = 5).\n\n  Compound           Concentration (ng/mL)   Recovery (%)   RSD (%)   IS Normalized Matrix Factor (%)   RSD (%)\n  ------------------ ----------------------- -------------- --------- --------------------------------- ---------\n  Ziyuglycoside I    10                      96             3.63      105                               5.26\n  800                96                      2.76           104       3.22                              \n  1600               98                      2.42           103       4.27                              \n  Ziyuglycoside II   10                      102            4.09      100                               7.55\n  800                98                      4.60           106       4.24                              \n  1600               100                     3.68           107       2.27                              \n  \u03b1-hederin (IS)     800                     94             2.71                                        \n\nmolecules-23-00543-t006_Table 6\n\n###### \n\nStability data of ZGS1 and ZGS2 in rat plasma (*n* = 5).\n\n  Compound           Concentration   Short-Term Stability for 24 h (25 \u00b0C)   3 Freeze-Thaw Cycles Stability   Long-Term Stability for 2 Weeks                  \n  ------------------ --------------- --------------------------------------- -------------------------------- --------------------------------- ------- ------ ------\n  Ziyuglycoside I    10              3.48                                    5.19                             5.65                              9.65    5.07   9.28\n  800                2.21            1.59                                    3.79                             1.93                              2.87    3.54   \n  1600               2.18            1.09                                    2.53                             1.45                              2.13    2.79   \n  Ziyuglycoside II   10              4.22                                    4.86                             5.95                              12.77   6.84   9.61\n  800                3.86            2.66                                    4.21                             3.83                              3.86    5.07   \n  1600               2.37            3.07                                    2.65                             5.72                              1.81    5.80   \n\nmolecules-23-00543-t007_Table 7\n\n###### \n\nPharmacokinetic parameters of ZGS1 and ZGS2 in rats after tail vein administration of 10 mg/kg ZGS1 (*n* = 5).\n\n  Parameter     Unit        Ziyuglycoside I (Mean \u00b1 SD)   Ziyuglycoside II (Mean \u00b1 SD)\n  ------------- ----------- ----------------------------- ------------------------------\n  AUC~(0--t)~   \u03bcg/L \\* h   534.841 \u00b1 73.569              299.6 \u00b1 28.083\n  MRT~(0-t)~    h           0.421 \u00b1 0.061                 0.575 \u00b1 0.06\n  t~1/2~z       h           1.338 \u00b1 0.744                 1.027 \u00b1 0.057\n\nmolecules-23-00543-t008_Table 8\n\n###### \n\nPharmacokinetic parameters of ZGS1 and ZGS2 in rats after subcutaneous injection administration of 10 mg/kg ZGS1 (*n* = 5).\n\n  Parameter    Unit        Ziyuglycoside I (Mean \u00b1 SD)   Ziyuglycoside II (Mean \u00b1 SD)\n  ------------ ----------- ----------------------------- ------------------------------\n  AUC~(0-t)~   \u03bcg/L \\* h   528.558 \u00b1 91.819              421.161 \u00b1 74.152\n  MRT~(0-t)~   h           2.196 \u00b1 0.173                 2.436 \u00b1 0.263\n  t~1/2~z      h           6.115 \u00b1 1.92                  7.935 \u00b1 3.264\n  Tmax         h           0.333                         0.333\n  Vz/F         L/kg        94.437 \u00b1 25.52                117.41 \u00b1 28.764\n  CLz/F        L/h/kg      10.999 \u00b1 1.839                10.996 \u00b1 2.802\n  Cmax         ng/mL       303.257 \u00b1 93.74               177.652 \u00b1 51.225\n\n[^1]: These authors contributed equally to this paper.\n"}
+{"text": "Modern surgical treatment for breast tumours requires a preoperative diagnosis of malignancy ([Perry, 2001](#bib13){ref-type=\"other\"}). Fine-needle aspiration and core biopsy are widely used for evaluation of palpable and nonpalpable suspicious breast lesions. However, there has been serious concern about malignant tumour cell displacement promoting iatrogenic tumour spread. Tumour cell displacement rates to the needle tract of up to 30% have been reported ([Youngson *et al*, 1995](#bib22){ref-type=\"other\"}; [Diaz *et al*, 1999](#bib4){ref-type=\"other\"}). In theory, tumour seeding into lymphatic or vascular vessels would carry the same risk of axillary lymph node metastases as true lymphatic invasion. To our knowledge, no study has investigated the rate of breast cancer cell seeding to the axillary nodes for fine-eedle aspiration and large gauge needle biopsy procedures. The concept of sentinel lymphadenectomy has been demonstrated to be an accurate staging alternative for breast cancer ([Krag *et al*, 1993](#bib12){ref-type=\"other\"}, [1998](#bib11){ref-type=\"other\"}; [Giuliano *et al*, 1994](#bib8){ref-type=\"other\"}, [1997](#bib7){ref-type=\"other\"}; [Veronesi *et al*, 1997](#bib20){ref-type=\"other\"}; [Veronesi *et al*, 2003](#bib21){ref-type=\"other\"}). The sentinel node (SLN) is as per definition 'the first lymph node that receives afferent lymphatic drainage from a primary tumour'. With the thorough pathologic examination of the SLN, it is possible to detect even early tumour cell spread in a lymph node, which might not have been seen otherwise. The previous analysis carried out by the Austrian Sentinel Node Biopsy Study Group revealed a nonsignificant trend of an increased risk of SLN metastasis after preoperative breast biopsy ([Pichler-Gebhard *et al*, 2002](#bib14){ref-type=\"other\"}). The purpose of this study was to evaluate the impact of preoperative biopsy on the rate of metastasis to the SLN of patients with primary breast cancer.\n\nMATERIALS AND METHODS\n=====================\n\nPatients and data collection\n----------------------------\n\nA total of 2502 consecutive women with primary breast cancer, in whom a SLN procedure was performed, were registered by the multi-centre database project (MCDBP) ([Konstantiniuk *et al*, 2001](#bib10){ref-type=\"other\"}). Patient data from 12 participating departments of the Austrian Sentinel Node Biopsy Study Group were collected prospectively from 1999 onwards, but data from some centres were obtained retrospectively from 1996 onwards. Feasibility and validation of the SLN biopsy method of the participating centres have been demonstrated by the Austrian Sentinel Node Biopsy Study Group previously. Each participating centre had to undergo a learning period, as has been established by means of quality control ([Pichler-Gebhard *et al*, 2002](#bib14){ref-type=\"other\"}). In all, 2328 cases remained after excluding women having received preoperative systemic treatment, patients with multifocal disease and *in situ* carcinomas as the role of SLN biopsy method in this group of patients still needs to be defined.\n\nTreatment methods\n-----------------\n\nBiopsy procedures of palpable masses were carried out by fine-needle aspiration or automated gun, nonpalpable masses, either by sonographically or by stereotactically guided biopsy. The number of FNA or core specimens obtained was at the discretion of the examiner performing the procedure. In cases of nonpalpable lesions, a stereotactic or ultrasound-guided wire localisation of the tumour was performed preoperatively. Lymphatic mapping and SLN dissection (SLND) were performed by using blue dye or radiolabelled colloid, or a combination of both. Blue dye and colloid were injected either around the edge of the lesion or submammillarily. If an excisional biopsy had to be performed for the diagnosis of malignancy, confirmation by the use of frozen section was followed by immediate SLND in the majority of cases. Sentinel lymph node dissection was performed as a second procedure if a permanent section of the excision specimen revealed invasive tumour in frozen section-negative cases or when frozen section was not available. In patients with preoperative diagnosis of malignancy by core needle biopsy, SLND was performed upfront. Sentinel nodes were dissected and sent for frozen section. Standard surgical treatment with quadrantectomy or total mastectomy was completed. Axillary lymph node dissection was performed if frozen section of the SLN identified tumour cells. Tumours were classified as described by the American Joint Committee on Cancer. Sentinel nodes were examined at two-step section levels of the paraffin block separated by 250\u2009*\u03bc*m. One of every pair was stained at each level by H&E and followed by cytokeratin IHC staining with monoclonal anticytokeratin antibodies if the H&E sections did not reveal metastases. Every second corresponding slide was further stained with IHC ([Rudas *et al*, 2002](#bib17){ref-type=\"other\"}).\n\nStatistical analyses\n--------------------\n\nContinuous variables were compared by Student\\'s *t*-test. For comparison of categorical variables, the *\u03c7*^2^ test was used. Risk estimates were carried out by univariate and multivariate logistic regression analysis. All reported *P*-values are results of two-sided tests. A *P*-value equal to or less than 5% was considered statistically significant. The SPSS 10.0.7 statistical software system was used for calculations.\n\nRESULTS\n=======\n\nIn 2079 out of 2328 patients a SLN was found, resulting in an overall identification rate of 89.3% (not stratified according to learning period). Finally, we had 1890 patients aged between 23 and 96 years (median age, 60 years) with complete information on all data evaluable for analyses. In all, 230 (12.1%) patients had a preoperative diagnosis carried out by FNA, 818 (43.3%) by core biopsy *vs* 842 (44.6%) patients without preoperative biopsy.\n\nPatients\\' clinicopathological characteristics are listed in [Table 1](#tbl1){ref-type=\"table\"} Table 1Patients characteristics\u00a0**No biopsyBiopsy**\u00a0\u00a0***N*=842*N*=1048*P*-value**Age (years, mean, s.d.)60.4 (13.0)59.7 (12.9)0.242Tumour size (mm, mean, s.d., range)14.9 (s.d. 8.1, 1.0--75)17.8 (s.d. 8.9, 0.5--80)\\<0.0001\u00a0\u00a0\u00a0\u00a0*Tumour location*\u00a0\u00a00.351\u2003Inner upper100 (11.9%)140 (13.4%)\u00a0\u2003Inner lower54 (6.4%)61 (5.8%)\u00a0\u2003Outer upper464 (55.1%)604 (57.6%)\u00a0\u2003Outer lower135 (16.0%)155 (14.8%)\u00a0\u2003Central89 (10.6%)88 (8.4%)\u00a0\u00a0\u00a0\u00a0\u00a0*Histologic type*\u00a0\u00a00.006\u2003Ductal526 (62.5%)704 (67.2%)\u00a0\u2003Lobular97 (11.5%)138 (13.2%)\u00a0\u2003Ducto- lobular90 (10.7%)73 (7.0%)\u00a0\u2003Other129 (15.3%)133 (12.7%)\u00a0\u00a0\u00a0\u00a0\u00a0*Grading*\u00a0\u00a00.055\u2003GI149 (17.7%)144 (13.7%)\u00a0\u2003GII439 (52.1%)562 (53.6%)\u00a0\u2003GIII254 (30.2)342 (32.6)\u00a0\u00a0\u00a0\u00a0\u00a0*Learning period*\u00a0\u00a0\\<0.0001\u2003Yes237 (28.1%)177 (16.9%)\u00a0\u2003No607 (71.9%)871 (83.1%)\u00a0\u00a0\u00a0\u00a0\u00a0*Palpable*\u00a0\u00a0\\<0.0001\u2003No352 (41.8%)268 (25.6%)\u00a0\u2003Yes490 (58.2%)780 (74.4%)\u00a0\u00a0\u00a0\u00a0\u00a0*Department*\u00a0\u00a0\\<0.0001\u2003175 (8.9%)177 (16.9%)\u00a0\u2003240 (4.8%)72 (6.9%)\u00a0\u20033229 (27.2%)143 (13.6%)\u00a0\u20034188 (22.3%)95 (9.0%)\u00a0\u20035137 (16.3%)338 (32.2%)\u00a0\u2003663 (7.5%)1 (0.1%)\u00a0\u2003713 (1.5%)55 (5.2%)\u00a0\u2003840 (4.8%)103 (9.8%)\u00a0\u2003914 (1.7%)30 (2.8%)\u00a0\u20031010 (1.2%)19 (1.8%)\u00a0\u2003110 (0.0%)13 (1.2%)\u00a0\u20031233 (3.9%)2 (0.2%)\u00a0\u00a0\u00a0\u00a0\u00a0*SLN technique*\u00a0\u00a0\\<0.0001\u2003Blue dye317 (37.6%)184 (17.6%)\u00a0\u2003Radioisotope243 (28.9%)146 (13.9%)\u00a0\u2003Combination282 (33.5%)718 (68.5%)\u00a0. Patients undergoing a preoperative biopsy had larger (mean difference 2.87\u2009mm) and more palpable tumours than the control group. This might be due to the nonavailability of stereotactic procedure in some participating centres. Lymphatic mapping using blue dye alone was more common in the nonbiopsy group. We found more ductal and lobular carcinomas and less ducto-lobular and other subtypes in the biopsy group. Tumour grading was nearly evenly distributed; patients\\' age and tumour location were comparable between both groups. Overall, 7.4% of the patients had SLND performed as a second procedure (3.1% in the biopsy group *vs* 12.8% in the nonbiopsy group).\n\nSentinel lymph node metastases were detected in 641 (33.9%) of 1890 patients. In all, 388 (37.0%) positive SLN were observed in the preoperative biopsy group and 253 (30.0%) in the control group. A total of 553 (29.3%) patients had a positive SLN when conventional H&E staining was performed. In 88 (4.6%) patients the presence of metastases was detected by IHC only ([Table 2](#tbl2){ref-type=\"table\"} Table 2SLN positivity according to biopsy and histological workup\u00a0**No biopsyBiopsyFNACore biopsy**\u00a0***N*=842*N*=1048*N*=230*N*=818**SLN positivity (HE+IHC)253 (30.0%)388 (37.0%)78 (33.9%)310 (37.9%)SLN positivity (HE)213 (25.3%)340 (32.4%)73 (31.7%)267 (32.6%)SLN positivity (IHC)40 (4.7%)48 (4.6%)5 (2.2%)43 (5.3%)).\n\nWith regard to the presence of a positive sentinel node (H&E and IHC staining), univariate regression analysis identified larger tumour size (*P*\\<0.0001), age (*P*=0.002), preoperative biopsy (*P*=0.001), histological type (*P*\\<0.0001), grading (*P*\\<0.0001), palpability of the lesion (*P*\\<0.0001), timing of SLN biopsy (*P*=0.008) and participating centre (*P*=0.003) as significant factors. Preoperative biopsy revealed a hazard ratio of 1.37 (95% CI, 1.13--1.66). Finally, all factors were entered in a multivariate regression model. The impact of each of the evaluated factors is shown in [Table 3](#tbl3){ref-type=\"table\"} Table 3Association between SLN metastases and clinicopathologic factors by univariate and multivariate analysis\u00a0\u00a0\u00a0**UnivariateMultivariate**\u00a0***N*% SLN positiveOdds ratio95% CI*P*-valueOdds ratio95% CI*P*-value**Age (years)\u00a0\u00a00.980.98--0.990.0020.980.97--0.99\\<0.0001Tumour size (mm)\u00a0\u00a01.071.06--1.09\\<0.00011.061.05--1.08\\<0.0001\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Histology*\u00a0\u00a0\u00a0\u00a0\\<0.0001\u00a0\u00a00.0004\u2003Ductal1230331.00\u00a0\u00a01.00\u00a0\u00a0\u2003Ducto-lobular163491.971.42--2.74\\<0.00012.161.48--3.160.0001\u2003Lobular235391.290.97--1.720.0811.120.82--1.540.487\u2003Others262250.680.51--0.930.0160.820.58--1.160.246\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Grading*\u00a0\u00a0\u00a0\u00a0\\<0.0001\u00a0\u00a00.0003\u2003I293180.340.24--0.48\\<0.00010.550.32--0.810.002\u2003II1001360.890.72--1.100.2751.170.93--1.480.177\u2003III596391.00\u00a0\u00a01.00\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Location*\u00a0\u00a0\u00a0\u00a00.710\u00a0\u00a00.310\u2003Inner upper240301.00\u00a0\u00a01.00\u00a0\u00a0\u2003Inner lower115290.960.59--1.560.8631.090.65--1.840.738\u2003Outer upper1068331.190.88--1.620.2521.260.91--1.760.165\u2003Outer lower290381.430.99--2.060.0531.491.01--2.220.044\u2003Central177401.591.06--2.400.0251.390.89--2.170.151\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Palpability*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2003Palpable1270412.782.22--3.49\\<0.00011.771.37--2.29\\<0.0001\u2003Not palpable620201.00\u00a0\u00a01.00\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Preop-biopsy*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2003No842301.000\u00a0\u00a01.00\u00a0\u00a0\u2003Yes1048371.371.13--1.660.0011.090.85--1.400.508\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Learning period*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2003No1476340.950.76--1.200.6731.130.79--1.620.509\u2003Yes414351.00\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Department*\u00a0\u00a0\u00a0\u00a00.003\u00a0\u00a00.001\u20031252371.00\u00a0\u00a01.00\u00a0\u00a0\u20032112421.240.78--1.950.3601.751.04--2.950.036\u20033372290.710.50--0.990.0470.750.50--1.120.157\u20034283411.200.85--1.710.2940.750.47--1.180.210\u20035475280.660.48--0.920.0140.560.38--0.830.004\u2003664330.830.47--1.490.5430.520.24--1.090.084\u2003768320.820.46--1.440.4880.740.39--1.400.354\u20038143431.270.84--1.930.2601.350.84--2.160.216\u2003944360.980.50--1.900.9451.020.45--2.300.958\u20031029340.900.40--2.020.7980.880.3--2.320.795\u20031113310.760.23--2.540.6550.740.20--2.810.664\u20031235250.510.22--1.160.1080.490.19--1.290.150\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*SLN technique*\u00a0\u00a0\u00a0\u00a00.087\u00a0\u00a00.175\u2003Combination100033.51.00\u00a0\u00a0\u00a0\u00a0\u00a0\u2003Blue dye501381.271.01--1.590.0371.200.86--1.680.287\u2003Radiotracer389320.990.77--1.280.9691.350.96--1.900.083\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*SLN timing*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u20031st procedure1750351.00\u00a0\u00a01.00\u00a0\u00a0\u20032nd procedure140230.590.39--0.870.0081.010.63--1.620.969\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0*Study period*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2003Prospective1354341.00\u00a0\u00a01.00\u00a0\u00a0\u2003Retrospective536351.060.86--1.310.5740.970.70--1.380.848. Of the variables which showed a significant correlation with SLN positivity in the univariate analysis, tumour size, tumour grading, histological type, age, palpability and participating centre remained as independent predictors for SLN metastasis, whereas preoperative biopsy failed to show significance. Patients with preoperative breast biopsy had a nonsignificant 1.09-fold (95% CI, 0.85--1.40) increased risk for SLN metastases.\n\nAs seeding to the needle tract during FNA is a very rare event, patients with preoperative FNA might have a different rate of tumour cell seeding than patients with large-gauge needle biopsy ([Roussel and Dalion, 1989](#bib16){ref-type=\"other\"}; [Tabara *et al*, 1991](#bib19){ref-type=\"other\"}; [Bott *et al*, 1999](#bib1){ref-type=\"other\"}). Therefore, further analysis was carried out with FNA and core biopsy as separate categories. Univariate analysis showed a significant correlation with SLN positivity for core biopsy, OR 1.42 (95% CI, 1.16--1.74), whereas no impact for FNA RR 1.19 (95% CI, 0.88--1.63) was found. However, when analysis was adjusted for other relevant factors for axillary node metastasis, preoperative biopsy again failed to show significance, OR 1.10 (95% CI, 0.85--1.43) and OR 1.03 (95% CI, 0.69--1.53), respectively. In addition, subgroup analysis of the risk for occult micro-metastases to the SLN (detected by IHC only) on H&E-negative cases with additional IHC staining also failed to show an increased risk for preoperative biopsy, OR 1.07 (95% CI, 0.69--1.65).\n\nDISCUSSION\n==========\n\nThis study examined the risk of SLN metastases in patients who underwent prior FNA or large-gauge needle core biopsy. Our results do not confirm the findings of a recently published study, performed on a much smaller number of patients ([Hansen *et al*, 2004](#bib9){ref-type=\"other\"}), suggesting a positive association between needle biopsy and SLN metastases. Studies of SLN metastases must take into account all well-known predictive factors for axillary node metastases such as tumour size, clinical and histo-pathological criteria. The significant result of an increased risk for SLN metastases after preoperative biopsy obtained on univariate calculation was no longer present when adjustment for relevant predictors for axillary node metastases was done by means of multivariate analysis. The additional use of IHC for further evaluation of the SLN increases the detection of occult micro-metastases and improves the sensitivity of the SLN procedure ([Czerniecki *et al*, 1999](#bib3){ref-type=\"other\"}). We found a 6.5% conversion rate of H&E-negative patients to lymph node positive in our series. Nevertheless, patients who underwent a preoperative biopsy had no increased risk for a SLN metastases detected by H&E and IHC or IHC alone. Large tumours were more likely to have nodal involvement than smaller tumours; this is in agreement with other findings ([Gann *et al*, 1999](#bib6){ref-type=\"other\"}). Clinical palpability remains to be highly predictive for SLN metastases, as has been demonstrated by other authors ([Silverstein *et al*, 2001](#bib18){ref-type=\"other\"}). In addition to our previous findings, we found high-grade lesions and age to be associated with higher rate of metastases, which is in accordance with other studies ([Gann *et al*, 1999](#bib6){ref-type=\"other\"}). Infiltrating ducto-lobular carcinomas had higher rates of SLN metastasis than other types; this confirmed the surprising findings of our first analysis.\n\nWithout any doubt, the rate of metastases to the sentinel node can be affected by several factors. In order to objectify this study as far as possible, we identified five possibilities of biases, which may have adversely affected the accuracy of our risk assessment.\n\nFirstly, no data on failure of preoperative core biopsy were available in our series. A negative core biopsy would increase the number of cases in the nonbiopsy group, altering the result in favour of a positive impact of the biopsy procedure. Nevertheless, assuming a failure rate of less than 10% would not change our results. Secondly, SLN identification by using blue dye alone might result in a higher false negative and lower detection rate. However, this point was considered and technique of SLN identification was included in the analysis. Thirdly, preoperative hook wire localisation procedure of nonpalpable breast lesions might also cause disintegration of tumours ([Fajardo, 1988](#bib5){ref-type=\"other\"}). To overcome this bias, clinical palpability of the breast lesion was included in the multivariate model. The fourth point is we must emphasise that there were no data on the time interval between the preoperative biopsy procedure and the SLND available in our study. Assuming a decreased incidence of local tumour displacement at increasing intervals between core biopsy and excision, as suggested by other authors ([Diaz *et al*, 1999](#bib4){ref-type=\"other\"}), the impact of the time interval should bias the study in favour of an increased risk for the occurrence of occult SLN metastasis following preoperative biopsy. In addition, the timing of SLND following excision biopsy was identified as a possible confounding factor. And, finally, surgical resection itself might have an impact on tumour cell displacement and tumour cell spread. Previous studies carried out on a small number of patients revealed that manipulation during cancer surgery in human beings might result in tumour cell dissemination into the vascular circulation ([Choy and McCulloch, 1996](#bib2){ref-type=\"other\"}). Excisional biopsy was only performed prior to SLND in patients without preoperative diagnosis of malignancy (non-core biopsy group and patients with negative core biopsy). Assuming additional tumour cell shedding into the lymphatic system will further enhance the impact of preoperative biopsy. This point remains unclear and needs further investigation. We are indeed aware of these possible biases and the retrospective nature of the study while interpreting our data, but considering the large patient population and the power of multivariate analyses the overall impact of the above-mentioned concerns does not alter our results.\n\nThe present data clearly indicate that preoperative biopsy does not increase the risk of metastases to the SLN in patients suffering from breast cancer. There is no evidence for any tumour cell spread to the sentinel node with possible negative impact on the prognosis of breast cancer. In conclusion, preoperative breast biopsy is a safe method and should be used to achieve definitive diagnosis of malignant breast lesions.\n\nWe thank the following members of the Austrian Sentinel Node Biopsy Study Group: Denison U, Felgel-Farnholz C, Gallowitsch HJ, Jakesz R, Pichler-Gebhard B, Poestlberger S, Riegler-Keil M, Rudas M, Selim U, Stengl S, Stierer M, Urbania A, Sevelda P and Wette V.\n"}
+{"text": "INTRODUCTION\n============\n\nRheumatoid arthritis (RA) is a chronic inflammatory disease with enhanced cardiovascular (CV) morbidity and mortality. Patients with RA have an approximately 70% higher risk of myocardial infarction than the general population, which is comparable to the risk in diabetic patients.[@B1] Pooled analysis of CV mortality in 91,916 patients with RA demonstrated a 60% increase in CV mortality in these patients, who were compared with the general population[@B2] and these rates did not change over time.[@B3]\n\nIn addition to the traditional risk factors such as obesity, insulin resistance, and dyslipidemia, the systemic inflammation that accelerates atherogenesis can contribute to the higher risk in RA patients.[@B4][@B5][@B6] Several lines of evidence suggest that a high number of inflamed joints, elevated erythrocyte sedimentation rate, elevated C-reactive protein, and interleukin (IL)-6 confound the development of CV complications in RA,[@B7][@B8] and the recommendations of a European League Against Rheumatism (EULAR) task force emphasize adequate control of disease activity to lower the CV risk in patients with RA.[@B9]\n\nMethotrexate is the first-line disease modifying anti-rheumatic drug (DMARD), and is widely used for monotherapy and combination therapy for RA. It effectively controls immune inflammation in the joint and vascular walls, thereby improving CV and total mortality in RA.[@B10][@B11][@B12] Moreover, other DMARDs to control disease activity in RA patients could have significant effects on patient metabolic profiles,[@B13] although controversy persists with regard to the effects of such therapies on lipids, adipokines, and insulin resistance.[@B14]\n\nInflammatory cytokines, including tumor necrosis factor (TNF)-alpha and IL-6 play central roles in RA pathogenesis; therefore, medications that inhibit TNF-alpha or block IL-6 receptors that are widely used to treat patients with RA refractory to conventional DMARDs include methotrexate. The use of TNF inhibitors is known to reduce the overall incidence of CV events.[@B12][@B15] However, as TNF-alpha is involved in weight loss,[@B16] weight gain is reported after administering TNF-alpha inhibitors. In addition, increases in fat mass, and inconsistent correlations with serum adipokine level changes have been described.[@B17][@B18][@B19]\n\nWith regard to IL-6 receptor inhibition by tocilizumab, another biologic DMARD that blocks an inflammatory cytokine pathway, there has been controversy for a long time on its CV safety. Tocilizumab increased cholesterol level,[@B20] more severely than did adalimumab, a TNF inhibitor.[@B21] Therefore, tocilizumab has been used with caution in patients with high CV risk, particularly patients with dyslipidemia. Following several reassuring data reports on major adverse CV events with tocilizumab compared to other biologic DMARDs,[@B22][@B23][@B24][@B25][@B26] a recent report of a postmarketing randomized controlled trial (RCT) upon the request of the US Food and Drug Administration showed that the risk of major adverse CV events of tocilizumab was comparable to etanercept, with a hazard ratio (HR) of 1.05.[@B26]\n\nWe investigated the effects of tocilizumab on traditional CV risk factors in patients with RA. Among the metabolic changes associated with tocilizumab use, the evidence for dyslipidemia is considered sufficient. Therefore, we focused on changes in body weight and adipokines---the protein mediators released by adipose tissue. A complex relationship between adipokines, RA, and CV diseases has been described, and the best-characterized adipokines are adiponectin, resistin, and leptin.[@B27] Especially, the leptin-to-adiponectin ratio has been reported as an insulin-resistance proxy or arthrogenic index in many studies.[@B28][@B29][@B30]\n\nIn view of all the above mentioned considerations, this study investigated changes in body weight after 24-week treatment with or without tocilizumab in patients with RA. Associations between body-weight changes and treatment responses or baseline body mass indexes (BMIs) were investigated. Finally, we investigated changes in levels of serum adiponectin, leptin, and resistin, as well as their association with weight change.\n\nMETHODS\n=======\n\nStudy design and patients\n-------------------------\n\nThis retrospective study comprised three patient cohorts ([Fig. 1](#F1){ref-type=\"fig\"}). Cohorts I and II were derived from patients with active RA who participated in a RCT and the extension study of tocilizumab treatment ([NCT01211834](http://clinicaltrials.gov/ct2/show/NCT01211834) and [NCT01256736](http://clinicaltrials.gov/ct2/show/NCT01256736)) from October 2009 to November 2013,[@B31] excluding patients without 0- and 24-week weight records. Patients were eligible for the RCT if they had been diagnosed with RA according to the 1987 American College of Rheumatology criteria and had active disease refractory to methotrexate with or without other conventional DMARDs. Active RA was defined by a swollen joint count \u2265 6 from the total of 66, tender joint count \u2265 8 from the total of 68, and serum C-reactive protein level \u2265 1 mg/dL or erythrocyte sedimentation rate (ESR) \u2265 28 mm/h. Cohort I comprised 45 participants who received a placebo and methotrexate for the first 24 weeks, excluding patients who had no serum remaining for further adipokine analysis. Cohort II consisted of 46 participants who were administered tocilizumab (8 mg/kg) and methotrexate for 24 weeks.\n\n![Study population and design. Cohorts I and II were extracted from clinical trial data of tocilizumab in patients with active rheumatoid arthritis. Cohort I comprised 44 patients who received methotrexate for 24 weeks, followed by tocilizumab for 48 weeks. Cohort II comprised 46 patients who received tocilizumab for 72 weeks. Cohort III comprised 41 patients who received tocilizumab with or without methotrexate for 24 weeks on clinical practice.\\\nNSAIDs = nonsteroidal anti-inflammatory drugs.](jkms-35-e155-g001){#F1}\n\nCohort III is an independent cohort of 41 patients with active RA who received tocilizumab therapy from January 2009 to December 2018 and were refractory or intolerable to methotrexate and/or conventional TNF-inhibitor treatment. Active RA was defined as DAS28 \u2265 5.1, or DAS28 \u2265 3.2 and the presence of new bone erosion after 6 months of treatment. All patients had weight data and serum samples for further adipokine analysis at 0 and 24 weeks.\n\nPatients in Cohorts I and II received fixed daily doses of corticosteroids, non-steroidal anti-inflammatory agents, and methotrexate during the initial 24-week RCT. Thereafter, these patients participated in an open extension phase during which they received tocilizumab treatment for 48 weeks, and changes in the doses of corticosteroids, non-steroidal anti-inflammatory agents, and methotrexate were permitted during weeks 24--72 of the study. Therefore, we could evaluate weight changes among patients in Cohort I during the initial 24-week period of methotrexate treatment and the subsequent 48-week period of tocilizumab treatment. In Cohort III, doses of corticosteroids, non-steroidal anti-inflammatory agents, and methotrexate were not fixed.\n\nStudy variables\n---------------\n\nData describing each patient\\'s age, gender, body weight, height, BMI, disease activity score in 28 joints, prednisolone use, and concomitant DMARD use were collected at baseline and at 24 weeks. The patients\\' nutritional status was classified by their baseline BMIs as normal weight (\\< 25 kg/m^2^) and overweight (\u2265 25 kg/m^2^), based on the World Health Organization\\'s criteria. Weight gain was defined as an increase in weight of \u2265 0.1 kg, and weight loss was defined as a decrease in weight of \u2265 0.1 kg. Patient responses to treatment were assessed by the American College of Rheumatology 20 response criteria (ACR20).\n\nWe evaluated levels of adiponectin, leptin, and resistin before and 24 weeks after treatment by using a Magnetic Luminex Multiplex Assay kit (R&D Systems, Inc., Minneapolis, MN, USA) in serum samples from patients in Cohorts I (methotrexate-treated) and III (tocilizumab-treated); serum samples from the patients in Cohort II were not available. We compared pre- and post-treatment body weight and adipokine levels in Cohorts I and III, and analyzed correlations between the adipokine levels and changes in body weight.\n\nStatistical analyses\n--------------------\n\nData distribution was assessed by Kolmogorov-Smirnov tests. Continuous variables are presented as the mean and standard deviation (SD) when they accept normality, and as median with 25th and 75th percentiles when they reject normality. Proportions are presented as numbers with percentages. The independent *t*-test or Mann-Whitney test was used to compare the two groups depending on the distribution of the data. Paired comparison was undertaken by the paired *t*-test or Wilcoxon test, as appropriate. Changes in body weight and adipokine levels from baseline are expressed as mean or median differences with 95% confidence intervals (CIs). The correlation of continuous variables was analyzed by Pearson\\'s or Spearman\\'s correlation coefficient, as appropriate.\n\nStatistical analyses were undertaken in MedCalc Statistical Software, version 18.6 (MedCalc Software bvba, Ostend, Belgium). The graphs were plotted using GraphPad Prism software, version 7.00 for Windows (GraphPad Software, La Jolla, CA, USA). The temporal trends in body weight were examined using the Joinpoint Regression Program, Desktop version 4.7.0.0 (National Cancer Institute; Bethesda, Maryland, USA; <http://surveillance.cancer.gov/joinpoint/>),[@B32] which uses the Bayesian information criterion to generate different numbers of joinpoints that indicate time points at which the rate of occurrence of, or the hazard associated with, an event changes significantly, and it fits separate linear trends in each time segment. For all statistical tests, a value of *P* \\< 0.05 was considered statistically significant.\n\nEthics statement\n----------------\n\nThe present study protocol was reviewed and approved by the Institutional Review Board (IRB) of the relevant hospitals (CBNUH IRB approval No., 2018-03-012; SNUH IRB approval No., 1711-085-901). All patients provided written informed consent prior to study enrollment.\n\nRESULTS\n=======\n\nDemographic characteristics and clinical variables\n--------------------------------------------------\n\n[Table 1](#T1){ref-type=\"table\"} summarizes the participants\\' demographic characteristics and clinical variables.\n\n###### Baseline demographic data and clinical parameters in the patients in the 3 study cohorts\n\n![](jkms-35-e155-i001)\n\n  Characteristics                    Cohort I (n = 44)                                      Cohort II (n = 46)   Cohort III (n = 41)   \n  ---------------------------------- ------------------------------------------------------ -------------------- --------------------- -----------\n  Treatment                          MTX, followed by TCZ                                   TCZ                  TCZ                   \n  Age, yr                            52.8 \u00b1 12.4                                            52.3 \u00b1 10.3          53.4 \u00b1 15.2           \n  Gender, women                      38 (86.4)                                              41 (89.1)            40 (97.6)             \n  Baseline weight, kg                57.6 \u00b1 9.7                                             55.4 \u00b1 7.9           49.0 (45.0, 52.8)     \n  BMI, kg/m^2^                       23.4 \u00b1 3.0                                             22.7 \u00b1 2.9           22.5 \u00b1 8.2            \n                                     Underweight (\\< 20)                                    4 (9.1)              8 (17.4)              10 (48.8)\n                                     Normal (\u2265 20 and \\< 25)                                29 (65.9)            29 (63.0)             17 (41.5)\n                                     Obese (\u2265 25)                                           11 (25.0)            9 (19.6)              4 (9.8)\n  Smoking                            3 (6.8)                                                2 (4.3)              \\-                    \n  Disease duration, yr               7.7 (2.5, 12.9)                                        8.8 (4.60, 16.4)     7.5 (3.8, 16.3)       \n  Baseline DAS28-ESR                 6.2 \u00b1 1.1                                              6.1 \u00b1 0.7            5.1 (4.3, 5.9)        \n  History of TNF inhibitor therapy   5 (11.4)                                               3 (6.5)              22 (53.7)             \n  Concomitant medication                                                                                                               \n                                     Regular glucocorticoids                                37 (84.1)            38 (82.6)             \\-\n                                     Prednisolone (or equivalent) dose^a^ per day, mg/day   5 (5, 7.5)           5 (2.5, 5)            \\-\n                                     MTX                                                    44 (100)             46 (100)              40 (97.6)\n\nData are presented as mean \u00b1 standard deviation or median (25th and 75th percentiles), as appropriate. Proportion was presented as number (%).\n\nMTX = methotrexate, TCZ = tocilizumab, BMI = body mass index, DAS28 = disease activity score in 28 joints, ESR = erythrocyte sedimentation rate, RA = rheumatoid arthritis, TNF = tumor necrosis factor.\n\n^a^Among the patients who used glucocorticoids regularly.\n\nWeight changes between baseline and 24th week\n---------------------------------------------\n\n[Table 2](#T2){ref-type=\"table\"} summarizes the mean weight changes during 24 weeks of treatment in three cohorts. There was no significant weight change in Cohort I, whereas mean weight gain of 0.9 kg (95% CI, 0.2--1.5 kg) was observed in Cohorts II+III (*P* = 0.010). [Fig. 2A](#F2){ref-type=\"fig\"} compares the percentages of patients who gained or lost weight after 24 weeks of treatment in the 3 cohorts. In Cohort I, 40.9% (18/44) of patients gained weight, 15.9% (7/44) of patients maintained their weight, and 43.2% (19/44) of patients lost weight. In Cohorts II and III, 63.0% (29/46) and 61.0% (25/41) gained weight and 28.3% (13/46) and 36.6% (15/41) lost weight, respectively.\n\n###### Weight changes during 24 weeks of treatment in three cohorts\n\n![](jkms-35-e155-i002)\n\n  Variables               Cohort I (MTX)       Cohort II and III (TCZ)\n  ----------------------- -------------------- -------------------------\n  Baseline weight, kg     57.6 \u00b1 9.7           53.4 \u00b1 8.6\n  Weight after 24 weeks   57.4 \u00b1 10.0          54.3 \u00b1 8.4\n  Difference              \u22120.2 (\u22120.8 to 0.5)   +0.9 (0.2 to 1.5)\n  *P* values              0.630                0.010\n\nData are presented as mean with standard deviation and the difference were presented as mean with 95% confidence interval.\n\nMTX = methotrexate-treated group, TCZ = tocilizumab-treated group.\n\n![Weight changes in the study\\'s cohorts. (**A**) Percentage of patients whose weights changed after 24 weeks of treatment in the 3 cohorts. In cohort I, 40.9% of the patients gained weight, 15.9% of the patients\\' weights remained stable, and 43.2% of the patients lost weight following methotrexate treatment. In cohort II, 63.0% of the patients gained weight and 28.3% of the patients lost weight following tocilizumab treatment. In cohort III, 61.0% of the patients gained weight and 36.6% of the patients lost weight following tocilizumab treatment. (**B**) Overall percent changes in body weight in cohort II after 24, 48, and 72 weeks of tocilizumab treatment compared with baseline. The mean (95% confidence interval) percent changes in body weight were 1.3% (0.1%--2.6%) at 24 weeks, 2.2% (0.7%--3.6%) at 48 weeks, and 2.0% (0.4%--3.7%) at 72 weeks.\\\nMTX = methotrexate, TCZ = tocilizumab.](jkms-35-e155-g002){#F2}\n\nThe weight difference before and after the 24-week treatment period did not differ between EULAR responders and non-responders in Cohort I (*P* = 0.747) and Cohorts II+III (*P* = 0.373, independent *t*-test). Furthermore, there was no significant difference in the proportion of weight gained between EULAR responders and non-responders, both in Cohort I (*P* = 0.319, Fisher exact test) and Cohorts II+III (*P* = 0.108, \u03c7^2^ test) ([Table 3](#T3){ref-type=\"table\"}).\n\n###### Subgroup analysis of weight changes according to the treatment response\n\n![](jkms-35-e155-i003)\n\n  Variables                                    Cohort I (MTX)   Cohort II and III (TCZ)                  \n  -------------------------------------------- ---------------- ------------------------- -------------- -------------\n  EULAR treatment response, moderate or good   Yes (n = 8)      No (n = 36)               Yes (n = 75)   No (n = 11)\n  Those who gain weight, No. (proportion)      2 (25%)          16 (44.4%)                35 (47.7%)     8 (72.7%)\n  *P* value                                    0.319^a^         0.108                                    \n\nMTX = methotrexate-treated group, TCZ = tocilizumab-treated group, EULAR = European League Against Rheumatism.\n\n^a^Fisher exact test, otherwise analyzed by \u03c7^2^ test.\n\nIn Cohort I, there was no linear correlation between baseline BMI and weight changes (*r* = 0.135, *P* = 0.384), although there was a weakly negative correlation in the combined analysis of Cohorts II and III (*r* = \u22120.278, *P* = 0.010). There was no significant difference in the proportion of weight gained between those who were underweight and overweight at baseline in Cohort I (*P* = 0. 812) or in Cohort II+III (*P* = 0.089) ([Table 4](#T4){ref-type=\"table\"}).\n\n###### Subgroup analysis of weight changes according to the baseline BMI\n\n![](jkms-35-e155-i004)\n\n  Variables                                 Cohort I (MTX)        Cohort II and III (TCZ)                          \n  ----------------------------------------- --------------------- ------------------------- ---------------------- ----------------\n  Baseline BMI category                     Underweight (n = 4)   Obese (n = 11)            Underweight (n = 28)   Obese (n = 13)\n  Those who gain weight, No. (proportion)   2 (50%)               3 (27.3%)                 17 (60.7%)             4 (30.8%)\n  *P* value                                 0.812^a^              0.089                                            \n\nMTX = methotrexate-treated group, TCZ = tocilizumab-treated group, BMI = body mass index.\n\n^a^Fisher exact test, otherwise analyzed by \u03c7^2^ test.\n\nSerial weight changes for 72 weeks\n----------------------------------\n\nIn Cohort II, mean changes in patient body weight after 24, 48, and 72 weeks of tocilizumab treatment were +0.7 kg (95% CI, 0.0--1.4 kg), +1.2 kg (95% CI, 0.4--2.0 kg), and +1.1 kg (95% CI, 0.2--2.0 kg), respectively, compared with their weight at baseline. In addition, we investigated the overall percent changes in body weight in Cohort II after 24, 48, and 72 weeks of tocilizumab treatment compared with that at the baseline ([Fig. 2B](#F2){ref-type=\"fig\"}). The mean percent changes in the patient body weight from baseline were +1.3% (0.1%--2.6%) at 24 weeks, +2.2% (0.7%--3.6%) at 48 weeks, and +2.0% (0.4%--3.7%) at 72 weeks.\n\nWe analyzed the patients\\' serial weight changes during the initial 24-week treatment period as well as during subsequent treatment with tocilizumab for 48 weeks in Cohorts I and II ([Fig. 3A](#F3){ref-type=\"fig\"}). Patients in Cohort I who received methotrexate only during the initial 24-week treatment period followed by tocilizumab treatment for 48 weeks during the open extension phase of the clinical trial, showed significant weight gain during tocilizumab treatment (mean difference between week 24 and week 72: +1.1 kg; 95% CI, 0.4--1.7 kg; *P* = 0.004). Joinpoint regression analysis of Cohort I showed that the mean body weight started to increase at 24 weeks and plateaued at 56 weeks, thereby creating a significant joinpoint at the 24th week and 56th week ([Fig. 3B](#F3){ref-type=\"fig\"}).\n\n![Patients\\' body weight changes during treatment. (**A**) Weight changes in cohorts I and II over 72 weeks. Body weight was assessed every 4 weeks in the 2 cohorts. (**B**) Joinpoint regression analysis of Cohort I showed that the mean body weight started to increase at 24 weeks when tocilizumab treatment was introduced, and it plateaued at 56 weeks, creating a significant joinpoint at weeks 24 and 56. ^\\*^Indicates that the slope is significantly different from zero at the alpha = 0.05 level.](jkms-35-e155-g003){#F3}\n\nSerum adipokine levels\n----------------------\n\nLevels of adiponectin, leptin, and resistin in serum samples from patients in Cohort I did not change after methotrexate treatment (adiponectin, *P* = 0.081; leptin, *P* = 0.682; and resistin, *P* = 0.070) ([Fig. 4](#F4){ref-type=\"fig\"}), whereas these levels increased significantly in Cohort III following tocilizumab treatment (adiponectin, *P* \\< 0.001; leptin, *P* \\< 0.001; and resistin, *P* \\< 0.01), with median differences of +1,257.2 ng/mL (95% CI, 801.8--1,753.5 ng/mL) for adiponectin, +3,890.0 pg/mL (95% CI, 2,709.7--6,076.1 pg/mL) for leptin, and +826.5 pg/mL (95% CI, 268.7--1,544.1 pg/mL) for resistin ([Fig. 5](#F5){ref-type=\"fig\"}). After tocilizumab treatment, the leptin-adiponectin ratio increased significantly (*P* = 0.015) with a median difference of +2.1 pg/ng (95% CI, 0.42--2.97 pg/ng). The leptin--adiponectin ratio did not increase significantly (*P* = 0.918) following methotrexate treatment (median difference, +0.1 ng/\u03bcg; 95% CI, \u22120.2 to 0.2 ng/\u03bcg).\n\n![Serum adipokine levels after methotrexate treatment. Levels of (**A**) adiponectin, (**B**) leptin, and (**C**) resistin at baseline (0 weeks) and after 24 weeks of methotrexate treatment (cohort I). The serum adiponectin, leptin, and resistin levels did not change significantly (*P* = 0.081 for adiponectin, *P* = 0.682 for leptin, and *P* = 0.070 for resistin). The data presented are the medians with their 25th and 75th percentiles.](jkms-35-e155-g004){#F4}\n\n![Serum adipokine levels after tocilizumab treatment. Serum levels of (**A**) adiponectin, (**B**) leptin, and (**C**) resistin at baseline (0 weeks) and after 24 weeks of tocilizumab treatment (cohort III). The serum adiponectin, leptin, and resistin levels increased significantly (^\\*\\*^*P* \\< 0.01; ^\\*\\*\\*^*P* \\< 0.001). The data presented are the medians with their 25th and 75th percentiles.](jkms-35-e155-g005){#F5}\n\nThe change in adipokine concentration did not showed the correlation with the weight change in adiponectin (rho = \u22120.014, *P* = 0.923 and rho = \u22120.040, *P* = 0.803), leptin (rho = \u22120.282, *P* = 0.052 and rho = \u22120.284, *P* = 0.069) and resistin (rho = \u22120.121, *P* = 0.411 and rho = \u22120.107, *P* = 0.501) in both Cohorts I and III, respectively. In addition, in the pooled analysis of Cohorts I and III, levels of serum adiponectin, leptin, and resistin not differ between the patients who gained weight (n = 43) and those who lost weight (n = 34) (adiponectin, *P* = 0.308; leptin, *P* = 0.321; and resistin, *P* = 0.170).\n\nRegarding disease activity (DAS28ESR) and adipokine concentrations at baseline in pooled cohort of I and III (n = 85), adiponectin (correlation coefficient *r* = \u22120.183, *P* = 0.094) or leptin (*r* = 0.074, *P* = 0.500) concentration did not show a specific correlation with DAS28ESR, but a correlation was observed in resistin concentration (*r* = 0.278, *P* = 0.010). In the subgroup analysis based on treatment response, the difference of adipokine concentrations between the EULAR responder and the non-responder was not significant in both Cohorts I and III ([Table 5](#T5){ref-type=\"table\"}).\n\n###### Subgroup analysis of adipokine changes according to the treatment response\n\n![](jkms-35-e155-i005)\n\n  Variables                                                 Cohort I (MTX)                 Cohort III (TCZ)                                              \n  --------------------------------------------------------- ------------------------------ ------------------------------ ------------------------------ -----------------------------\n  EULAR treatment response, moderate or good                Yes (n = 8)                    No (n = 36)                    Yes (n = 36)                   No (n = 5)\n  Adiponectin, median change (interquartile range), ng/mL   \u2212555.6 (\u22121,016.2 to 52.6)      \u2212171.2 (\u2212941.7 to 380.1)       1,243.6 (626.7 to 2003.3)      2,677.5 (\u2212103.8 to 2,803.1)\n  *P* value                                                 0.533                          0.750                                                         \n  Leptin, median change (interquartile range), pg/mL        \u2212934.9 (\u22121,321.0 to 5.6)       \u2212202.5 (\u22123,530.6 to 3,325.8)   3,745.8 (1,287.3 to 7,864.2)   3,972.0 (734.0 to 7,591.4)\n  *P* value                                                 0.593                          0.780                                                         \n  Resistin, median change (interquartile range) pg/mL       \u2212771.6 (\u22123,531.2 to 2,054.1)   \u22121,084.6 (\u22125,378.6 to 619.4)   1,009.4 (\u221297.9 to 4,082.7)     39.4 (\u2212396.5 to 1,363.1)\n  *P* value                                                 0.573                          0.248                                                         \n\nMTX = methotrexate-treated group, TCZ = tocilizumab-treated group, EULAR = European League Against Rheumatism, Mann-Whitney test (independent samples).\n\nDISCUSSION\n==========\n\nIn this study, we investigated changes in body weight and the adipokine levels in patients with active RA who were or were not treated with tocilizumab; tocilizumab inhibits the binding of IL-6 to its receptor. The patients\\' body weights increased after tocilizumab therapy, even among the non-responders, but no weight gains were observed among the patients who were treated with methotrexate. Therefore, weight gain appeared to be a drug-specific effect. As weight gain after treatment with TNF-alpha inhibitors has not been related to the therapeutic response,[@B33] and weight gain after tocilizumab treatment did not appear to be associated with the therapeutic response in this study.\n\nWhen we looked at the relationship between disease activity and adipokine concentrations at baseline, resistin concentration showed the positive correlation with DAS28ESR in pooled cohort of I and III. However, there was no significant decrease in resistin concentrations even when disease activity declined after tocilizumab therapy. Adipokine levels are affected by several individual characteristics of patients, such as body weight, fat mass, and insulin resistance, as well as RA. Therefore, we investigated the paired data to correct the individual characteristics of the patients that could act as a confounder, and analyzed the correlation of adipokines with EULAR response.\n\nObesity and rheumatoid cachexia have been associated with RA.[@B6] However, the metabolic profile observed in patients with RA does not reflect what has been previously observed in the general population. A lipid paradox has been described, wherein low cholesterol and LDL cholesterol levels in patients with RA are associated with increased CV risks.[@B34][@B35] In addition, a low BMI has been associated with accelerated mortality in patients with RA, namely, the obesity paradox. Weight loss is a strong predictor of death in patients with RA, because RA cachexia rather than obesity has been associated with increased CV mortality in RA.[@B36][@B37] Therefore, it is probable that the weight gain observed after tocilizumab therapy is an additional beneficial effect in RA patients.\n\nIn this regard, adipokine profiles of RA patients differ from the general population in non-inflammatory conditions. Adiponectin is a protein produced by the adipose tissue. In general, low adiponectin levels have been associated with obesity, type 2 diabetes, and atherosclerosis and the role of adiponectin in the metabolic syndrome is clearly anti-inflammatory.[@B13][@B27][@B38] On the other hand, serum adiponectin concentrations are higher among patients with RA compared with those in healthy individuals, and were associated with increased radiographic damage and low amount of visceral fat.[@B33][@B39][@B40] Data on the effects of anti-TNF treatments on adiponectin levels are conflicting, for which the reasons are yet to be explained and might include differences by demographics or ethnicities of patients in those studies.[@B17][@B41][@B42][@B43].\n\nLeptin is produced in the adipose tissue, and it is a product of the *OB* gene.[@B44] Moreover, leptin is a major determinant of obesity and is associated with the development of obesity or CV disease in RA.[@B27][@B33][@B45] Leptin is generally higher in patients with RA than in patients with degenerative arthritis and healthy individuals,[@B46][@B47] and high leptin levels are associated with high RA activity and more aggressive disease course.[@B48][@B49] Thus, leptin might be involved in the RA-related atherosclerosis. Date regarding the effects of anti-TNF treatments on leptin levels are also conflicting,[@B41][@B42] and does not seems to directly modulate leptin levels.[@B27] Resistin was initially implicated in the development of insulin resistance. Although resistin concentrations do not differ between healthy individuals and patients with RA, resistin levels correlated positively with C-reactive protein levels, suggesting a pro-inflammatory role.[@B33][@B47] There is minimal information exists on the vascular effects of resistin and its relationship to CV risk in RA.\n\nRecently, increased adiponectin levels, but not leptin or resistin levels, have been described after tocilizumab therapy,[@B50] although differences in body weight after the administration of tocilizumab were unclear. The study did not limit the use of other RA medications, including steroids; therefore, any changes in weight or the adipokine levels may have been caused by dose reductions of other medications following a good response to tocilizumab therapy. In addition, the study involved European participants, and there may be ethnic differences in Asian populations.\n\nIn this study, the serum adiponectin, leptin, and resistin levels increased after tocilizumab treatment, although they did not increase after methotrexate treatment. The leptin--adiponectin ratio, which is an insulin-resistance proxy,[@B28][@B29] increased after tocilizumab therapy in our study, which led to concerns that CV disease risk factors, including insulin resistance, may deteriorate as the body weight increases following tocilizumab treatment. As with the interpretation of weight gain, the interpretation of the adipokine change should be interpreted differently in patients with RA to general population. The leptin--adiponectin ratio is known to decrease in highly active, therapy-resistant RA and to increase in early or obese RA.[@B6] Therefore, the increase in leptin--adiponectin ratio may reflect a positive situation as opposed to RA cachexia. These findings disagree with findings from a recent study that showed that the use of tocilizumab contributed to a reduction in the leptin--adiponectin ratio in patients with RA in whom TNF inhibitors had failed compared with that in patients who switched to other TNF inhibitors.[@B51] While the patients in that study had already failed to respond to TNF inhibitor treatment, and the investigators compared the effects of tocilizumab and other TNF inhibitors after switching, only half of the patients in our study were exposed to the effects of a TNF inhibitor; therefore, the study populations appear to differ.\n\nAs the changes in the adipokine levels did not differ between the patients who lost or gained weight, we could not attribute the weight gain observed after tocilizumab therapy to increases in the adipokine levels. Indeed, the number of patients was too small to detect statistical differences between subgroups and this is a major limitation of this study. In addition, we do not have information on diet peculiarities and physical activity of participants during the tocilizumab treatment periods, which is another important limitation. Lastly, information on comorbidities such as dyslipidemia, hypertension, and diabetes, which may affect adipokine levels, was not available. To correct individual characteristics of patients that can act as confounders, we used paired data and analyzed the association between changes in disease activity and changes in adipokine concentrations.\n\nAs far as we know, this is the first study to investigate the long-term effects of tocilizumab treatment on body weight among patients with RA. In Cohorts I and II, patients received fixed doses of glucocorticoids and methotrexate, and changes in their body weights were evaluated every 4 weeks, which minimized the risk of influencing weight changes as the steroid doses were reduced in good responders. We could also observe the increases in body weight immediately after the initiation of tocilizumab treatment in Cohort I. In Cohort III, we could observe the change of on body weight in real practice setting, such as adjusted dose of glucocorticoid glucocorticoids and methotrexate if they shows satisfactory response to the tocilizumab treatment. The results of this study, such as the proportion of patients who gained weight in each subgroup, the scale of weight gain, and lack of correlation to treatment effects, could provide the basis to explain drug characteristics to patients when selecting subsequent therapy.\n\nIn conclusion, a statistically significant weight gain occurred among the patients with RA after tocilizumab treatment, which was sustained over the 72-week treatment period. As the tendency to gain weight was not related to the therapeutic response, the changes may have been a feature of IL-6 inhibition itself. Further studies are required to determine the effects of this weight gain on CV outcomes.\n\nThe results of Cohorts I and II on the efficacy and safety of tocilizumab therapy ([NCT01211834](http://clinicaltrials.gov/ct2/show/NCT01211834) and [NCT01256736](http://clinicaltrials.gov/ct2/show/NCT01256736)) were published in a separate paper,[@B31] which did not address the changes in weight or adipokine concentration among participants. No parts of the current manuscript have been copied or published elsewhere. There is a difference in the composition of cohorts, but the contents of the interim analysis of this study were presented as a poster at the 21st Asia Pacific League of Associations for Rheumatology Congress (APLAR) in conjunction with the Australian Rheumatology Association, 8--11 April 2019, Brisbane, QLD, Australia, which can be accessed from <https://doi.org/10.1111/1756-185X.13545> and *International Journal of Rheumatic Diseases*, Volume 22, Issue S3, Pages 110 (2-084). Some biospecimens for this study were provided by the Chungbuk National University Hospital, a member of the National Biobank of Korea.\n\n**Funding:** This research was supported by a research grant from Chungbuk National University Hospital in 2018 and it was partly supported by grants from the Ministry of Science, ICT and Future planning (NRF-2015M3A9B6052011) and (2019M3A9A8065574) and partly supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI14C1277).\n\n**Disclosure:** IA Choi received a research grant from Chungbuk National University Hospital in 2018. She received a research grant from LG Chem and AbbVie Korea, which is not relevant to this study. EY Lee received the grants from the Ministry of Science, ICT and Future planning (NRF-2019M3A9A8065574). EB Lee has received research grants from Pfizer Korea, and has been a consultant for Pfizer. YW Song received a research grant from JW Pharmaceuticals during the clinical trials ([NCT01211834](http://clinicaltrials.gov/ct2/show/NCT01211834) and [NCT01256736](http://clinicaltrials.gov/ct2/show/NCT01256736)). The use of clinical data and serum samples in this study were supported by JW Pharmaceuticals. The present study received grants from the Ministry of Science, ICT and Future planning (NRF-2015M3A9B6052011, 2019M3A9A8065574) and a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI14C1277).\n\n**Author Contributions:** **Conceptualization:** Choi IA, Song YW.**Data curation:** Choi IA, Sagawa A, Song YW.**Formal analysis:** Choi IA.**Methodology:** Choi IA, Sagawa A, Lee EY, Lee EB, Song YW.**Writing - original draft:** Choi IA, Song YW.**Writing - review & editing:** Choi IA, Sagawa A, Lee EY, Lee EB, Song YW.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nEach year, millions of people are exposed to serious health risks due to emerging infectious and communicable diseases. This poses a severe threat to the public health security at local, regional or national level; especially in underdeveloped countries. This is primarily due to the lack of infrastructure for timely collection, reporting, and analysis of epidemic data; non-existent early warning and forecasting systems; inadequate preparedness and emergency response management\u00a0\\[[@CR22]\\].\n\n3.9 billion people in 128 countries are at risk of contracting dengue, with estimated 96 million cases annually. The worldwide incidence of dengue has risen 30-fold in the past 30\u00a0years. Pakistan is among the 110 countries in the world which are badly affected by the mosquito-borne dengue virus. The first outbreak of dengue fever (DF) in Pakistan was confirmed in 1994. The country is currently experiencing among worst-ever dengue outbreaks, recording about 45,000 confirmed cases\u00a0\\[[@CR27]\\]. Even today, an effective dengue vaccine offering balanced protection is still elusive. Unfortunately, existing dengue vaccines are known to have limited efficacy and cure\u00a0\\[[@CR8]\\]. This underscores the critical need of preventing dengue transmission and eventual outbreak by: (i) investigating favorable conditions for the dengue epidemic to occur\u00a0\\[[@CR21]\\]; (ii) plummeting the vector population\u00a0\\[[@CR23]\\]; and averting the vector-human contact\u00a0\\[[@CR7]\\], all of which are perceived as daunting challenges of Epidemiology. Due to the lack of ICT enabled governance, the existing infectious disease surveillance systems in Pakistan are unable to perform epidemiological spread analyses and effective emergency response planning\u00a0\\[[@CR11], [@CR18]\\]. Researchers have started exploring latest techniques like Artificial Intelligence for diagnostic screen of dengue suspects\u00a0\\[[@CR12]\\], but these methods are still in their infancy.\n\nTo mitigate these challenges, we need to develop a reliable *health surveillance* and rapid *emergency response* infrastructure, that monitors and responds to known endemic diseases in the country, and evolves to cater potential new outbreaks. This infrastructure should have the capability to collect spatio-temporal epidemiological data, analyze it using computational methods, forecast possible outbreaks and generate early warnings for rapid emergency response management. Due to the sparsity of data, the dynamics of the vector population and its interaction with the human population is quite difficult to capture. We therefore need to build models that incorporate the vector population, disease transmission and the spread direction of the infectious disease, and that support the application of preemptive strategies and countermeasures.\n\nIn this study we present a simulation approach to analyze and predict the mosquito population density and its consequence on dengue spread. We also simulate the pathogen transmission, and observe the dynamic interaction of human and mosquito population, both of which we use to forecast outbreaks in a spatial environment. Our proposed agent-based simulation framework allows modeling of both human and vector population dynamics, using separate layers. Human agents evolve between different states, from Susceptible to Exposed, Infected and eventually Recovered states while Mosquito agents evolve from Egg to Larva, Pupa and Adult state. Both population layers are spatially distributed and we model the interaction between both layers using our proposed algorithm. In addition, we expose both layers to exogenous variables such a temperature, humidity, rainfall and the permeable water surfaces in the region. From our simulation results we can reproduce predicted dengue cases across age groups and highlight them using spatio-temporal visualizations. We validate our results using data of an existing study of the local region\u00a0\\[[@CR4]\\]. Our proposed framework can be re-used for simulation, visualization and forecasting of any region and aid Public health departments in emergency preparedness.\n\nThe rest of the paper is organized as follows: Sect.\u00a0[2](#Sec2){ref-type=\"sec\"} outlines the background concepts used in this paper and the literature review. Section\u00a0[3](#Sec8){ref-type=\"sec\"} discusses our proposed framework. Section\u00a0[4](#Sec12){ref-type=\"sec\"} provides simulation results and model validation and Sect.\u00a0[5](#Sec13){ref-type=\"sec\"} provides conclusions and future work.\n\nBackground and Literature Review {#Sec2}\n================================\n\nDengue Epidemiology {#Sec3}\n-------------------\n\nThis section provides key concepts used in our modeling approach. Epidemiology is the study of models, causes, effects, risk factors, transmission, spread and outbreaks of infectious diseases in a particular population\u00a0\\[[@CR25]\\]. DENV is transmitted to humans through the bite of infected female Aedes mosquitoes primarily Aedes aegypti and Aedes albopictus. A susceptible mosquito can acquire infection from an infectious person and transmit it further, or human gets infection from the bite of already Infectious Aedes aegypti. It spends its lives around or in side houses, becomes adult and typically flies up\u00a0to 400\u00a0m\u00a0\\[[@CR26]\\]. Adees aegypti life cycle is composed of four stages as shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}. Aedes Aygepti proliferates around 30--32\u00a0$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$^\\circ $$\\end{document}$C. The number of eggs laid per batch depends on the weight of mosquito and other factors. Details of the dengue epidemiology can be viewed at\u00a0\\[[@CR9]\\].Fig. 1.Mosquito to human dengue transmission Fig. 2.Life-cycle of Aedes aegypti\n\nEnvironmental Factors {#Sec4}\n---------------------\n\nVarious meteorological factors also influence the growth of the vector population and the spread of dengue fever. Temperature and humidity influence the incidence of dengue fever by modifying adult feeding behavior, larval development and mosquito survival. The average life span of Aedes mosquito is 25\u00a0days, with a range from one day to 76\u00a0days. Population density of Aedes Aegypti rapidly increases in summers as rising temperatures shortens the incubation period of mosquitoes and they take less time to emerge from eggs to adults. This increase the overall risk of dengue transmission\u00a0\\[[@CR14]\\]. The mortality rate of mosquitoes increases at high temperatures. For temperatures 15\u00a0$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$^\\circ $$\\end{document}$C, the feeding frequency of mosquitoes increases resulting in greater risk of viral transmission. Mosquito life cycle consists of two stages one of which is aquatic i.e they require stagnant water in order to develop and reproduce. Rainfall plays a vital role in the development of mosquitoes and Dengue transmission. There is an increase in dengue cases during and after rainy seasons. Pakistan is unfortunately experiencing severe forms of above stated changes. Relative humidity, temperature and rain remained noteworthy prognosticators of dengue occurrence in Pakistan. Surge of cases occurred from September to October\u00a0\\[[@CR24]\\], as shown in Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}.Fig. 3.Dengue cases in Pakistan (2013--2018)\n\nMathematical Modeling in Epidemiology {#Sec5}\n-------------------------------------\n\nTraditional non-spatial epidemiology models have been used use to represent epidemics of communicable diseases\u00a0\\[[@CR3]\\]. A *Susceptible, Exposed, Infectious, Recovered (SEIR)* is an epidemiological model that describes the transmission process of an infectious disease\u00a0\\[[@CR19]\\]. The individuals enter or leave the compartment based on flow rates defined. There is a decrease in Susceptible individuals when individuals get infection, or they die. Number of individuals in the Exposed compartment increases when they get infection and decreases as soon as they show the symptoms of infection and decreases with the death rate. The infection compartment increases by the infection events of Exposed and decreases by the death and recovery rate. There is an increase in Recovered compartment with the recovery rate and decreases with the natural death. Another extension to the SIS model was proposed in\u00a0\\[[@CR13]\\]. It contains three population components: Humans (H), Vectors (M) and Eggs (E). 'L' represents the number of Latent mosquitoes that are in incubation period, as shown in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}. At first a Susceptible human gets infection from a vector moves to the Infectious compartment. While at infectious state, a human may get recovered and move to the Recovered state or may die. An Infectious human may spread infection to a Susceptible vector which would move to the Latent compartment, where it stays for the time between getting Infection and becoming infectious, called latency period; the vector then moves to the Infectious compartment. A susceptible mosquito if lays eggs increases the population of susceptible vectors, whereas an Infectious mosquito if lays eggs would increase the population of Infectious vectors. It is also possible that an Infectious mosquito lays normal eggs. The main goal of modeling the life cycle of mosquitoes is to estimate the growth of the vector density at a place and time; and the interaction with the hosts to predict the rate of spread of disease. Our proposed approach is based on the foundation of this principle.Fig. 4.Humans, mosquitoes and eggs. S\u00a0=\u00a0Susceptible, E\u00a0=\u00a0Exposed, I\u00a0=\u00a0Infected, R\u00a0=\u00a0Recovered, L\u00a0=\u00a0Latent, H\u00a0=\u00a0Human, M\u00a0=\u00a0Mosquitoes, E\u00a0=\u00a0Egg\n\nAgent-Based Modeling and Simulation {#Sec6}\n-----------------------------------\n\nMost existing mathematical models focus on non-spatial macro-level aspects of the system. Because we are strongly interested in the micro-level aspects of our agents (a level where we can more easily extract rules and behaviors), we therefore use agent-based modelling. An agent-based model consists of autonomous decision-making entities: ***Agents***, each able to perceptively assess situations and make decision according to a set of predefined rules. ABM is decentralized and produces collective behavior by agents communicating and interacting with each other in their environment\u00a0\\[[@CR16]\\]. The interest in using ABM has been recently renewed due to their ability to model complex geospatial structures and interacting networks\u00a0\\[[@CR5]\\].\n\nLiterature Review {#Sec7}\n-----------------\n\nMany researchers have proposed different dengue epidemic simulation models. However, selection of the methodology and platform greatly depends upon the purpose of the study. Jacintho, et al.\u00a0\\[[@CR16]\\] propose an agent-based model of the dengue spread, using the Swarm platform that models the micro-level behavior of agents in the spread and transmission of dengue fever using a rule-based approach, however it lacks GIS based spatial representation. Almeida, et al.\u00a0\\[[@CR1]\\] propose an individual-based approach to model Aedes aegypti population considering vectors, humans and objects as agents, using repast framework. Kang and Aldstadt\u00a0\\[[@CR17]\\] proposed an approach to validate spatially explicit ABM for several specifications of vector-borne disease transmission models, using multiple scale spatio-temporal patterns. Hunter et al.\u00a0\\[[@CR15]\\] developed a data driven agent-based SEIR model that resulted in the emergence of patterns and behaviors that are not directly programmed into the model. Lima, et al. \\[[@CR20]\\] developed DengueME, a collaborative open source platform to simulate dengue disease and its vector's dynamics. It supports compartmental and individual-based models, implemented over a GIS database, that represent Aedes aegypti population dynamics, human demography, human mobility, urban landscape and dengue transmission mediated by human and mosquito encounters. Yein Ling Hii\u00a0\\[[@CR14]\\] proposed a machine learning approach to study dengue fever considering climatic factors: temperature and rainfall using Poisson multivariate regression and validated through multiple statistical models. Guo, et al.\u00a0\\[[@CR10]\\] presented a comparison of various machine learning algorithms for the accurate prediction of dengue in China, and shows that support vector regression achieved a superior performance in comparison with other forecasting techniques assessed in this study.\n\nOur proposed framework uses the mathematical model proposed by \\[[@CR13]\\], as underlying foundation and extend it using an ABM approach, which consists of two population layers: (i) human and (ii) vectors (mosquitoes and eggs). Both layers express the population dynamics, mobility and microscopic behavior of human and vector agents separately, yet provide a common spatial environment for their interactions, to study the spread of disease within the desired spatiotemporal resolution.\n\nProposed Framework {#Sec8}\n==================\n\nThis section discusses the details of our proposed framework which is composed of three layers: (i) Host Layer (Human population); (ii) Vector Layer (Mosquito population); and (iii) Pathogen Layer (Dengue parameters), as shown in Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}. The framework further integrates: (i) GIS based spatial environment for agent distribution and mobility; (ii) time step using a user defined temporal resolution (from Min to Year); (iii) simulation engine to handle change of states and dynamic event processing; (iv) database for importing input data and exporting simulation results; (v) exogenous variables for weather and climate data; and (vi) validation data for model validation. The framework also provide a visualization dashboard for viewing the change of states of the distributed population of agents in real-time (or virtual time a.k.a faster than real-time) and for the graph visualization of the simulation results.Fig. 5.Agent-based simulation framework\n\nHost Layer {#Sec9}\n----------\n\nThis layer concerns with the structure, behavior and the interactions of host agent and allows the modelers to initialize a host population (e.g., human or animals) using variable configurations. The behavior of the host agent is represented by a state chart, as shown in Fig.\u00a0[6](#Fig6){ref-type=\"fig\"} which consists of six states: Susceptible, Exposed, Infected (or Immune), Recovered (or Dead). In this paper we assume the entire host population is initialized in Susceptible state. In the future, we intend to support a distribution of initial states so that we can use more realistic model configurations of a study area. When at *'Susceptible'* state, a host agent receives an incoming message from an infectious mosquito, representing the 'infectious bite', it transits to an *'Exposed'* state. After a certain period, defined by the incubation period parameter (3--8 days\u00a0\\[[@CR28]\\]) either the host will transit to *'Infected'* state or go to immune state (if resolved immune by the immunity probability). When at infected state, an incoming 'bite' message causes a self-loop transition, which is used to transmit the virus to an uninfected mosquito, with a given probability. After a certain duration, defined by the illness duration parameter (20\u00a0days) the host will transit to 'Recovered' state, if resolved true by the survival probability (0.9) or 'Dead' otherwise. We implement the interactions between Host and Vector agents using message passing and a distance based network type\u00a0\\[[@CR2]\\], which implies that a vector is connected with multiple hosts that are situated within a given range (36\u00a0m), and can communicate i.e. bite with infection or acquire an infection through a bite, as shown in Fig.\u00a0[8](#Fig8){ref-type=\"fig\"}. The detailed bite algorithm is presented later in this section.Fig. 6.Host agent (human) Fig. 7.Vector agent (mosquito) Fig. 8.Bite mechanism\n\nVector Layer {#Sec10}\n------------\n\nThis layer concerns with the structure, behavior and the interactions of vector agent. The behavior of the vector agent is represented by a state chart, as shown in Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}. It consists of states: *Egg*, *Larva*, *Pupa* and *Adult* (male or female). A vector agent can be initialized in any of these states. In this paper we assume initialize the vector population with the probability of 0.2 eggs, 0.2 larvae, 0.2 pupas and 0.4 adults (0.5 male and 0.5 female). The vector agents are uniformly distributed over the space, however this distribution can be fine tuned if the data of the concentrations of the mosquitoes nests is available. From egg, the agent moves on to the Larva state between 2--3 days. From larva it takes 4\u00a0days to transit to Pupa state, where it resides for 2\u00a0days and moves to the Adult stage, with a 0.5 probability to be a male or a female\u00a0\\[[@CR6]\\]. The male Aedes aegypti neither makes a bite nor transmits infection and dies after a lifespan of 14\u00a0days and then go to the final state '*dead*'. A female mosquito can enter into the *'Oviposition'* state when it is ready to lay eggs. A susceptible female will lay susceptible eggs whereas an infected female lay eggs with infectious probability\u00a0=\u00a00.3. The spawning of eggs is triggered by the event *'EggSpawn'*. The initial state of an egg (infected or not) depends on whether the female mosquito is in Susceptible or infected state. The lifespan of the female mosquito is assumed to be between 42--56 days. Inspired from the Helmersson's mathematical model\u00a0\\[[@CR13]\\], shown in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, the 'susceptible', 'latent' and 'Infected' states of the female mosquito are implemented. If not initially infected (i.e., *IsInfected = false*), an adult female mosquito enters into 'susceptible' state. It goes to 'latent' state if bites an infected human and acquires the virus (with a transmission probability of 0.18). The latent period is the delay from transmission to infection and is assumed to be 10\u00a0days, after which the agent enters into the 'Infected' state. While at 'susceptible' or 'Infected' states, a female mosquito bites human and have her fill of blood after each meal, at the rate of a bite after 2\u00a0days. Aedes aegypti is an intermittent biter and prefers to bite more than one person during the feeding period, therefore we assume it bites all the connected hosts. This bite is triggered by the event 'bite' at random intervals until the mosquito is dead When a 'bite' occurs the interaction between Host and Vector is implemented as shown in Algorithm 1. It takes VectorID and a list of ConnectedHosts as input. A vector is connected to all the hosts that lie within the range of 36\u00a0m. This algorithm determines infectivity of the vector after the bite using a boolean variable 'Infected'. Line1 assigns the existing state (infected or not) of the vector. Line 2--14 iterates a list of all the connected hosts and evaluate two scenarios as shown in Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}. If the vector is infected it will transmit an *'infectious bite'* through message passing, and cause the host to transit from ***'susceptible'*** to '***exposed'*** state. Else, if the host is infected the vector will acquire infection (with a 0.018 probability), and transit from ***'susceptible'*** to '***Latent'*** state. Otherwise it will just *'bite'* the host without the transmission of any infection. The number of times mosquito will bite human is defined by bite rate\u00a0\\[[@CR13]\\], which is temperature dependent and is calculated as:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ BiteRate=\\frac{\\left( 0.03\\times T\\;+0.66\\right) }{7} $$\\end{document}$$\n\nPathogen Layer {#Sec11}\n--------------\n\nThis layer deals with the modalities of the pathogen under study i.e., 'Dengue' in our case. In this paper the structure and the behaviour of the pathogen is limited only to initialize the serotype (i.e., DENV1, DENV2, DENV3 or DENV4) and key parameters such as infectivity, transmisiability, survivability and incubation period. In future, we aim to extend this layer for dealing with the complex logic of the cross-immunity with different serotypes.\n\nSimulation and Results {#Sec12}\n======================\n\nOur simulation is implemented using AnyLogic University Edition, and performed for a population of 50,000 persons and 1,000 mosquito agents, for a period of 90\u00a0days, on an Intel Core i7-8700 CPU\\@3.20\u00a0GHz, 16.0\u00a0GB RAM, and a 64-bit Windows Operating System. The human and mosquito populations are randomly distributed within a selected region in the city of Islamabad. A simulation run is shown using spatial and temporal visualization in Fig.\u00a0[9](#Fig9){ref-type=\"fig\"} and [10](#Fig10){ref-type=\"fig\"}. Initially the person agent population is susceptible, as soon as it is bitten by infectious mosquito it becomes Exposed (orange), after completing its incubation period of 4--7 days in exposed state it either becomes Infectious (red) or Immune (gray). From Infectious state after completing 4--12 days it either becomes Recovered (green) or Dead (black).Fig. 9.Simulation results - spatial visualization (Color figure online) Fig. 10.Simulation results temporal visualization (Color figure online)\n\nThe simulated results obtained are compared with the actual results of confirmed cases obtained from the dengue outbreak occurred in the local region in 2013. The actual cases were 9,036 out of the total population of 1,257,602 while the simulated cases were 400 out of the total agents' population of 50,000. In order to compare the data results of actual and simulated dengue cases, we calculated the prevalence of both data sets:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\text {Prevalence} = \\frac{\\text {No. of Dengue Cases}}{\\text {Total Population}} \\times 100 \\end{aligned}$$\\end{document}$$For comparison we present a graph containing both cases in Fig.\u00a0[11](#Fig11){ref-type=\"fig\"}. We observe a similar trend in the prevalence of dengue cases between the actual and simulated results. The Root Mean Square Error (RMSE) of both actual and simulated prevalence for the 90\u00a0days is *0.064*. The results show that the dengue transmission is temperature dependent. It is to be noted that with the increase in temperature, the biting rate of Aedes aegypti increases that give rise to the dengue cases.Fig. 11.Comparison of dengue cases -Source of actual data \\[[@CR4]\\]\n\nConclusion and Future Work {#Sec13}\n==========================\n\nIn this paper we presented a framework for modeling, simulation, visualization and forecasting of dengue spread, using an agent-based modeling approach. Our framework incorporates the structure, behaviour and interactions of key entities of disease epidemiology: (i) Host (Humans); (ii) Vector (mosquitoes); and (iii) Pathogen (dengue virus), using a separation of concern through independent layers. Our open-ended framework offers greater flexibility to the modelers for modification and extensibility of the disease under study, and can be used as a research tool by stakeholders (e.g., clinicians, microbiologists and public health professionals) to study the epidemiology of a region. The unique feature of our framework is its ability to model the life-cycle and population dynamics of both host and vector population, while incorporating extended mathematical models for the study of the epidemic spread. In the absence of vector population data, this tool provides means to synthesize vector population, and help improve the understanding of the spread of diseases. In the future, we plan to improve our proposed framework further by adding the mobility layer to incorporate movements, both in the Host and Vector layers. We also aim to extend our framework to support populations of a million agents or more.\n"}
+{"text": "Background {#Sec1}\n==========\n\nNatural language processing {#Sec2}\n---------------------------\n\nNatural Language Processing (NLP) provides a set of computational methods and techniques for automatically extracting and structuring information from free-text documents. NLP research has been successfully applied to free texts for several applications ranging from semantic search to information extraction to text analytics \\[[@CR1]--[@CR3]\\]. The development and availability of biomedical knowledge resources such as the Unified Medical Language System \\[[@CR4]\\], has enabled biomedical NLP to move beyond retrieval and classification to modeling of semantic predicates represented in the literature \\[[@CR5]\\]. Within the clinical domain, NLP systems have been implemented to support pharmaco-vigilance, patient screening, patient narrative summarization, and quality improvement \\[[@CR6]--[@CR13]\\]. The development of text processing pipelines and components specific to clinical text such as the Medical Language Extraction and Encoding System (MedLEE) \\[[@CR14]\\], clinical Text Analysis and Knowledge Extraction System (cTAKES) \\[[@CR15]\\], and Health Information Text Extraction \\[[@CR16]\\] have permitted the analysis of clinical free texts e.g., emergency department notes, radiology reports, etc. using lexical, syntactic, and semantic information \\[[@CR17]\\].\n\nOntologies {#Sec3}\n----------\n\nNLP tools designed to support information extraction routinely use the Web Ontology Language (OWL) to provide a structured way to represent domain content \\[[@CR18]--[@CR21]\\]. In order for an NLP tool to use a given domain ontology, the tool must contain code to parse and interpret the data model represented in the ontology. This creates a close coupling between the ontology and the NLP tool. It is generally not possible to directly share the domain ontology used in one NLP tool with another NLP tool and semantic schematic changes are not easily propagated between tools. To help resolve this issue of incompatibility, a common type system \\[[@CR22]\\] was developed which provides a common framework to create ontologies across a range of clinical domains. Our lab has converted the common type system described by Wu et al. into OWL format and extended its content using the Secondary Use Clinical Element Models (Secondary Use CEMs) \\[[@CR23]\\]. We use this new OWL-base common type system, which we call the Schema Ontology, as the framework to create domain specific ontologies.\n\nThe Schema Ontology can be loaded into Prot\u00e9g\u00e9 \\[[@CR24]\\] or other OWL editors and used as the template for domain ontology creation. Domain ontology creation in this manner, however, requires deep understanding of OWL and an understanding of the structure of the Schema Ontology data model. This creates a potentially burdensome learning curve for those users who simply want to create Schema-Ontology-based domain ontologies and have little training in ontology development. To improve ease of use and support wide-spread adoption of the Schema Ontology, a system which minimizes complexity and allows simple interaction for users is needed. Knowledge Author provides a simple user interface to guide users in development of complex domain ontologies. Furthermore, a domain ontology development tool that supports collaborative editing and has built-in access to the UMLS would speed up the domain ontology creation process. Many OWL editors, such as Prot\u00e9g\u00e9 or NeOn \\[[@CR25]\\] allow user-created plugins to extend their functionality; however, there are no editors that are sufficiently modifiable to support all of this desired functionality. In this paper we introduce Knowledge Author which provides a web-based interface that is simple to use, facilitates domain content development with direct UMLS terminology lookup, and supports collaborative domain content creation.\n\nImplementation {#Sec4}\n==============\n\nTerminology {#Sec5}\n-----------\n\nThe terminology used in the domains of clinical NLP and ontology creation can often vary; however, for the purposes of this paper the following terms are defined as such:*Semantic Schema* -- the target extraction template for an NLP tool.*Atomic Concept* -- a concept found in a standardized terminology such as the UMLS. For example -- PNEUMONIA, TEMPERATURE, COUGH, or IBUPROFEN.*Lexical Variant* -- a lexical variant is another way of phrasing an atomic concept or modifier in the clinical text. This can include synonyms, misspellings and abbreviations. For example -- two lexical variants for TEMPERATURE are \"temperature\" and \"temp\".M*odifier* -- additional information that narrows down, or modifies an atomic concept. Knowledge Author separates the modifiers into two distinct types -- shared and semantic. Shared modifiers are applicable to all concepts (with the exception of \"Person\" concepts which has its own unique set of modifiers such as age, race and gender). Semantic modifiers vary depending on the semantic type associated with the concept. For example, a concept with a semantic type of *Medication* will have a different set of available modifiers than a concept with semantic type *Vital Sign*.*Concept* -- the combination of the atomic concept with its associated modifiers and their lexical variants. For example -- 80\u00a0mg Ibuprofen is comprised of the atomic concept IBUPROFEN and semantic modifiers of *Dosage*: 80\u00a0mg. Lexical variants for IBUPROFEN can include \"Advil\", \"Midol\", \"Motrin\", \"Ibu\", and \"Ibuprofen\", etc. Lexical variants for *Dosage* can include \"80\u00a0mg\", \"80\u00a0mg\", and \"0.08\u00a0g\", etc.\n\nKnowledge Author overview {#Sec6}\n-------------------------\n\nThe overall goal of Knowledge Author is to aid a user in quickly creating a semantic schema, which is the target extraction template for a clinical NLP tool. The semantic schema represents salient concepts of interest to be extracted from the clinical text. It contains a list of atomic concepts, associated modifiers and the lexical variants for those concepts and modifiers. It is the job of the NLP system to extract words and phrases associated with these concepts and modifiers from the clinical text then map this information to the concepts in the semantic schema.\n\nKnowledge Author provides a web-based graphical user interface that guides the user in developing a semantic schema, which is output as an OWL ontology. Knowledge Author standardizes the semantic schema creation process by constraining concept creation to a set of standard semantic types (e.g., Procedure, Medication) and by only allowing the user to assign a pre-defined set of modifiers to the concept. The semantic types and modifiers are based on the Secondary Use Clinical Element Models and the Common Type System (CTS). The Secondary Use CEMS are semantic types and modifier sets used for computerized provider entry and secondary use of clinical data, and the CTS are semantic type sets used for information extraction from clinical text. By adhering to a standardized data model it becomes possible to use the output of Knowledge Author in any NLP system which implements that model.\n\nKnowledge Author also supports the semantic schema creation process by:Providing domain content suggestions through mapping of user-created concepts to concepts in the UMLS Metathesaurus database. This allows the automatic import of synonyms, concept definition, and semantic type into the Knowledge Author interface.Supporting modifier creation through the use of dropdown menus and the filtering of possible modifiers to only those relevant to a given concept type. Dropdown menus are possible because the Knowledge Author data model has a fixed set of allowable modifiers.Allowing the user to store and share their semantic schemas in an organized way.Supporting collaborative development of domain content.\n\nUsing Knowledge Author {#Sec7}\n----------------------\n\nTo illustrate the use of Knowledge Author, the creation of an example semantic schema for carotid stenosis will be walked through. Figure\u00a0[1](#Fig1){ref-type=\"fig\"} illustrates the Knowledge Author workflow to be described below.Fig. 1Illustrates the common set of steps to create a semantic schema using Knowledge Author. It is not required to map a concept to UMLS terminology as the synonyms, definition and semantic type can be entered in manually through the Knowledge Author interface\n\n### Defining a concept {#Sec8}\n\nThe first step in Knowledge Author is to create a concept. Knowledge Author supports creation of two types of concepts: Person and Event. A Person concept can be defined with modifiers such as *birth date, death date, race, age,* and *gender* to facilitate creation of complex concepts such as African American females above 65\u00a0years of age.\n\nThe carotid stenosis use case only has Event concepts. To create the first concept -- aneurysm -- the new concept button \"+\" (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}) was clicked and the concept name, \"aneurysm\", was entered. Upon saving the new concept, the \"Terminology Lookup\" button (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}) becomes available. Clicking that button allows the user to search the UMLS Metathesaurus for the concept name and displays a list of potential matches (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}). For this concept there is a UMLS atomic concept ANEURYSM which we choose. Knowledge Author will now download the definition, synonyms, semantic type and Concept Unique Identifier (CUI) for that atomic concept. All imported information can be changed, deleted, or supplemented as necessary. For the carotid stenosis example, twenty-six of the twenty-eight concepts were able to be mapped to UMLS concepts.Fig. 2Knowledge Author concept creation interface. The large red letters with arrows point out **a**) concept creation button; **b**) terminology lookup button; **c**) shared modifiers; **d**) semantic type; **e**) concept listFig. 3UMLS terminology lookup interface\n\n### Choosing a semantic type {#Sec9}\n\nThe next step is to assign a semantic type to the concept. If the concept is mapped to a UMLS atomic concept, the semantic type for the atomic concept will have already been downloaded and assigned to the concept (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}). If not, the user can manually assign a semantic type. In the context of Knowledge Author, there are two types of modifiers -- shared and semantic. The semantic type determines which type of semantic modifiers can be assigned to the concept.\n\n### Selecting semantic modifiers {#Sec10}\n\nSemantic modifiers are a type of modifier that is associated with specific semantic types. Each semantic type contains a number of possible semantic modifiers based on the Secondary Use CEMs and CTS. Each of the semantic modifiers has, in turn, a number of possible values associated with it. For example, the semantic type *Medication* allows the user to choose from semantic modifiers such as *dosage* or *delivery route.* The *delivery route* modifier has a number of possible values such as *oral* or *intravenous*. Table\u00a0[1](#Tab1){ref-type=\"table\"} lists the 12 semantic types, the modifier classes associated with each semantic type and the number of semantic modifiers associated with each modifier class.Table 1Semantic types, modifier classes, and modifiers available to the userSemantic TypeModifier Class\\# of ModifiersSample of ModifiersAllergy IntoleranceAllergy/Intolerance Type2allergy, intoleranceAllergenunlimitedany drug or food conceptSeverity7mild, moderate, severeAnatomical SiteBody Side3right, left, bilateralBody Laterality33dorsal, medial, superiorDisease DisorderCourse37increased, worsened, maintainedSeverity7mild, moderate, severeEncounterFrom Locationunlimitedhome, ER, SICU, nursing homeTo Locationunlimitedhome, ER, SICU, nursing homeLab/Test/MeasurementAbnormal Interpretation3abnormal, not abnormal, very abnormalDelta Flag8changed, unchanged, increasedLab/Test/Measurement Valueunlimited500\u00a0cc, 100\u00a0kg, 12000 WBCsOrdinal Interpretation35excessive, high, low, positiveMedicationMedication Form27capsule, cream, liquid, tablet, pillMedication Route21inhalation, intradermal, oralMedication Strengthunlimited500\u00a0mgStatus Change8changed, unchanged, increasedDosageunlimited250\u00a0mg, 16 unitsPatient DemographicBirth DateunlimitedDeath DateunlimitedAgeunlimitedGender2First NameunlimitedLast NameunlimitedMiddle NameunlimitedProblemCourse37increased, worsened, maintainedSeverity7mild, moderate, severeProcedure InterventionDelta Flag8changed, unchanged, increasedProcedure Completion3complete, incomplete, N/AProcedure/Intervention DeviceunlimitedProcedure/Intervention Methodunlimitedarthroscopic surgerySign or SymptomCourse37increased, worsened, maintainedSeverity7mild, moderate, severeSocial Risk FactorDelta Flag8changed, unchanged, increasedSocial Risk Qualifier6occasional, frequent, socialSocial Risk Quantityunlimited5 packs, 3 drinksSocial Risk Status5former risk, current riskVital SignAbnormal Interpretation3abnormal, not abnormal, very abnormalDelta Flag8changed, unchanged, increasedOrdinal Interpretation37excessive, high, low, positiveVital Sign Valueunlimited19\u00a0bpm, 86\u00a0%, 101.4\u00a0F\n\nSemantic modifier values can either be chosen from a dropdown list, or for the case of numeric values, entered directly into an editable text box. Some modifiers, such as medication dosage, consist of two numeric value boxes and a dropdown list. The numeric value boxes allow the user to specify a value range, and the dropdown list is for units (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). For example the user could create a concept for 80 to 100\u00a0mg Ibuprofen (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). By leaving one or the other numeric value box empty concepts such as \\>80\u00a0mg Ibuprofen, or \\<80\u00a0mg Ibuprofen can be created. To create a single numeric value such as 80\u00a0mg Ibuprofen, enter the same number into both boxes. For the aneurysm concept created earlier, only the mild form is of interest so the sematic modifier of *severity* is enabled, and the value of *mild* is chosen from the dropdown list.Fig. 4Semantic modifier interface box showing numeric range input boxes with units dropdown list\n\n### Selecting shared modifiers {#Sec11}\n\nA user can also narrow the definition of a concept through the use of shared modifiers. For all Event concepts, Knowledge Author allows the user to specify the *temporality* (whether the concept occurs in the past, present, or future), *certainty* (whether the concept is asserted, negated, or hedged), and *experiencer* (whether the patient or someone else experiences the concept) (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). Several other shared modifiers are also available (Table\u00a0[2](#Tab2){ref-type=\"table\"}). For the carotid stenosis example, a concept for \"no occlusion\" is needed, so a new concept is created and linked to the atomic concept, OCCLUSION, which is then assigned the lexical variant for shared modifier for *certainty: no* from the certainty dropdown list (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). The user could also use shared modifiers to create concepts such as family history of breast cancer or probable chest pain.Table 2Shared modifiers available to the userCategoryShared ModifiersCertaintyDefinite Existence, Definite Negated Existence, Probable Existence, Probable Negated ExistenceExperiencerPatient, Family Member, Donor Family Member, Donor Other Member, Other MemberTemporalityBefore, Before-Overlap, Overlap, AfterContextual AspectContinues, Initiates, Intermittent, Novel, Reinitiates, TerminatesContextual ModalityHypothetical, ConditionalDegreeLittle, MostPermanenceFinite, PermanentFig. 5\"Certainty\" shared modifier dropdown list\n\n### Building a semantic schema {#Sec12}\n\nOnce a concept is created and saved, the \"+\" button is clicked to create a new concept and the process described above is repeated. A concept, once created, is added to the concept list on the left hand side of the Knowledge Author GUI (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}). The concept list can be arranged by the order in which the concepts were created, or by the semantic type they belong to. Once all of the concepts are created, the user can export the semantic schema for use in an NLP system.\n\nExporting data {#Sec13}\n--------------\n\nAs the user works, Knowledge Author saves the user's work to an internal database that is available upon login. Once all of the domain content is entered into Knowledge Author, the user can choose to export the data for use in an NLP system. The \"Export\" button at the top of the application will prompt the user to save the output file to their computer.\n\nThe file exported from Knowledge Author is OWL based and imports and uses the classes defined in the Schema Ontology file. This file contains the semantic categories and modifiers used by the interface as classes. The Schema Ontology is the base ontology file that organizes these classes into appropriate hierarchies. This Schema Ontology file is then imported into every new domain ontology created by Knowledge Author. During the export process, each of the concepts is exported as a subclass of the appropriate semantic category class (i.e., mild aneurysm is a subclass of the \\\"Problem\\\" class found in the Schema Ontology). All of the concept metadata (i.e., synonyms, misspellings, preferred term, CUI from UMLS, etc.) is added as annotation properties to that class. The modifiers are added as restrictions on the concept class (i.e., mild aneurysm has the restriction \\\"hasSemAttribute some Mild_Severity\\\"). Therefore, all of the data gathered by the Knowledge Author user interface is transformed into an ontological representation that can be parsed by a compatible NLP system.\n\nIt is also of note that the Knowledge Author output file can be viewed and modified directly by any OWL editor such as Prot\u00e9g\u00e9. This could be useful for users who want to use the Knowledge Author feature set, such as UMLS terminology mapping, semantic schema management, and dropdown lists, but have a small number of concepts with rare features that are not currently supported in Knowledge Author. Those concepts could be added by hand using the OWL editor.\n\nCollaborative development and semantic schema management {#Sec14}\n--------------------------------------------------------\n\nOver time a user can develop a large number of semantic schemas. Each schema a user creates is saved to the Knowledge Author database and is accessible to the user upon login. The five most recent schemas a user worked on are displayed in the quick launch window. All other schemas can be viewed in a searchable table.\n\nA semantic schema can be designated by the creator as either \"public\" or \"private\". Public schemas can be viewed and edited by anyone using Knowledge Author. This allows multiple users to work on the same schema. It also allows for the creation of a library of public schemas which can be used as the starting point for building a new schema in a similar domain. Private schemas can only be viewed and edited by the original creator.\n\nSoftware tools and specifications {#Sec15}\n---------------------------------\n\nKnowledge Author is a web-based platform written in Java 7 on top of a MySQL database. It runs on an Apache Tomcat 7 Server. The SeaCore \\[[@CR26]\\] framework is used to facilitate the web development. The UMLS terminology is accessed through both the use of a local copy of the UMLS database and the Java based UMLS Terminology Service API 2.0 \\[[@CR27]\\] which queries a remote UMLS Metathesaurus service. The mapping of a user's concept to a UMLS atomic concept uses the UMLS Terminology Service API because of the complexity of performing that operation. The synonyms, definition, and semantic type for a concept are retrieved from the local copy of the UMLS for speed. The OWL API 3.4 \\[[@CR28]\\] is used for converting the semantic schemas to OWL XML.\n\nIntegration with existing NLP tools {#Sec16}\n-----------------------------------\n\nCurrently, only the pyConText \\[[@CR29]\\] NLP system accepts the output from Knowledge Author as input. Work is also underway to integrate cTAKES and a developmental system called Moonstone \\[[@CR30]\\] with the Knowledge Author output.\n\nResults and discussion {#Sec17}\n======================\n\nKnowledge Author standardizes the concept creation process by constraining the semantic types and modifiers that can be assigned to a concept to a discreet set. This enables the use of dropdown lists for assigning modifiers and allows for a standard output format which makes it possible to build NLP systems that use the output directly. We conducted two proof-of-concept studies, using different datasets, to assess the usability of Knowledge Author by demonstrating that (a) *the user interface is sufficiently flexible to allow for the creation of most concepts a user will want to create* and (b) *the output of Knowledge Author can be utilized by an NLP system to produce viable results.*\n\nUser interface flexibility assessment {#Sec18}\n-------------------------------------\n\nWe assessed the flexibility of the Knowledge Author user interface by assembling a dataset of 115 concepts to be created using Knowledge Author. The Additional file [1](#MOESM1){ref-type=\"media\"} contains a full list of the concepts. The concepts were drawn from three disease or procedure areas: pneumonia, colonoscopy quality, and influenza. The concepts were selected to cover a range of complexity and provide a broad view of the types of concepts that can and cannot be created using Knowledge Author.\n\nIn order to assess whether or not the required concepts could be created using Knowledge Author, we considered three degrees of representation: *complete creation*, *partial creation,* and *no creation supported*.\n\nWe observed that 76\u00a0% (87 of 115) of the concepts for the pneumonia, colonoscopy, and influenza use cases could be *completely* created using Knowledge Author. Table\u00a0[3](#Tab3){ref-type=\"table\"} describes the 24\u00a0% (28 of 115) of concepts that could be *partially* created in their entirety (see Additional file [1](#MOESM1){ref-type=\"media\"} for a full list of 115 concepts created). Knowledge Author supported the creation of a very high proportion of \"simple\" concepts (69 of 73), but a lower proportion of \"complex\" concepts (18 of 42) by the knowledge engineer. Complex concepts include compound concepts developed from two semantic types, such as \"lab test positive for influenza\". Knowledge Author supports creation of the concept \"lab test positive\" and \"influenza\" but does not yet support linking the two into a single concept. Knowledge Author, also, does not support creation of concept representing a single atomic concept with a set of modifiers combined with a disjunction, such as \"new or progressive infiltrate\". The four \"simple\" concepts that were not able to be created in Knowledge Author are a result of the required modifiers not being listed in the Knowledge Author data model.Table 3Types and number of concepts that were not able to be created in Knowledge AuthorReason Not CreatedTotal \\# of Concepts% of Total (115)Element or modifier type not found in Schema Ontology2118\u00a0%Relation between concepts missing - could only create separate concepts without linking76\u00a0%\n\nEven though Knowledge Author does not support the creation of some concepts, it is possible to add the desired data by hand outside of Knowledge Author. The Knowledge Author data model allows for the use of the Semantic Web Rule Language (SWRL) \\[[@CR31]\\] rules, even though the Knowledge Author interface itself does not. SWRL is an OWL-based rule language. Through manual editing of the Knowledge Author output file, complex variables can be created by inserting SWRL rules. For modifiers that are not in the data model, it is possible to add the appropriate modifier classes by hand to the Knowledge Author output file. Correctly designed NLP tools that use the Knowledge Author output are able to handle user created classes. Having to add information outside of the Knowledge Author interface is time consuming and as Knowledge Author matures we expect to expand its functionality to cover the vast majority of concepts.\n\nKnowledge Author-powered information extraction evaluation {#Sec19}\n----------------------------------------------------------\n\nWe assessed the viability of the Knowledge Author output for use in clinical NLP by creating a semantic schema for carotid stenosis in Knowledge Author and using it as the target extraction template in the pyConText \\[[@CR32], [@CR33]\\] NLP system.\n\npyConText is a regular-expression, rule-based information extraction system which accepts two files -- one for target concepts and one for associated modifiers. The target file contains regular expressions or lexical variants describing target concepts of interest such as those representing carotid disease. The modifier file contains regular expressions or lexical variants describing the types of modifiers such as *certainty, anatomical location* or *temporality*. A software script was written to automatically marshal the data contained in the Knowledge Author output file into the file format and schema supported by pyConText.\n\nWe selected 34 carotid ultrasound reports from the MT Samples corpus \\[[@CR34]\\] that were used in a previous study \\[[@CR32]\\]. The reports were de-identified and selected at random from the MT Samples corpus. Two physicians independently annotated each report and adjudicated each disagreement with consensus review using an annotation tool called eHOST \\[[@CR35]\\]. Each report was annotated for the targeted finding concepts for carotid stenosis along with the following associated modifiers: *certainty, sidedness,* and *neurovascular anatomy*.\n\nWe applied pyConText using the Knowledge Author semantic schema to the texts and converted its output to Knowtator.xml to be read into eHOST to conduct our error analysis. We computed recall for each type of target and modifiers (the proportion of concept mentions correctly identified from the reference standard) because we are predominately concerned with whether we have enough lexical variants to identify these concepts from free-text.\n\nReasonably high recall was achieved identifying targeted finding concepts (86\u00a0%) and shared modifiers (*certainty:* 91\u00a0%) and high to low recall for the semantic modifiers (*sidedness:* 80\u00a0%, *neurovascular anatomy*: 46\u00a0%) (Table\u00a0[4](#Tab4){ref-type=\"table\"}).Table 4pyConText performance leveraging Knowledge Author knowledge baseConceptTypesTotalCorrectRecallTargetsFindings796886\u00a0%*ModifiersCertainty*111091\u00a0%*Sidedness*413380\u00a0%*Neurovascular Anatomy*411946\u00a0%\n\nThe low recall can be partially attributed to missing cues from the terminology lookup. In particular, many false negatives were due to missing acronyms and abbreviations in the semantic modifier file e.g., \"ICA\" which stands for \"*neurovascular anatomy: Internal carotid artery*\" and \"l\" which stands for \"*sidedness: left*\" which are commonly used in carotid ultrasound reports. Additionally, low recall can be partially attributed to the inability for Knowledge Author to represent ranges of severity for some semantic modifiers e.g., \"70-80\u00a0%\" which indicates significant stenosis. We are actively incorporating this functionality in the system. A manual input of additional acronyms and abbreviations using the Knowledge Author synonym interface and manual input of regular expressions for semantic modifiers using an OWL editor could improve the results. Overall, this result suggests that the Knowledge Author output has the potential to be used by an NLP system to create viable results.\n\nFuture development {#Sec20}\n------------------\n\nWe are continuing to develop Knowledge Author and add new features. Some of the features that we expect to be added in the near future include:Adding constructs that will allow users to link concepts together using relationships (i.e. \"ibuprofen treats pain\") and logical operators.Allowing the user to search a default corpus of de-identified medical records for phrases that would potentially be retrieved for the new concept. This would allow the user to test the accuracy of synonyms and numeric thresholds.Allowing the user to share and collaboratively work on an ontology with a select group of users.\n\nKnowledge Author is the first part of a pipeline that will allow the user to create an NLP schema, annotate documents, process documents using various NLP systems, and analyze the results. We envision an end-to-end system that allows the user to rapidly build custom clinical text queries using a variety of NLP systems. We are actively developing a recommendation module within the pipeline that will suggest new lexical variants for concepts and modifiers from clinical text leveraging active learning methods to improve recall i.e., acronyms and abbreviations observed from development data in real-time. Currently, only the pyConText algorithm uses the output from Knowledge Author. Additional systems are under development.\n\nConclusions {#Sec21}\n===========\n\nKnowledge Author is a new, web-based tool for building a semantic schema of domain content that could be used in an NLP application. It leverages three existing knowledge resources -- the Secondary Use CEMs, CTS, and the UMLS -- to provide the user with relevant information for creation of domain-specific concepts, which allows for rapid semantic schema creation. The output of Knowledge Author can be used directly as input into compatible NLP systems.\n\nAvailability and requirements {#Sec22}\n=============================\n\nKnowledge Author is publically available and can be found at <http://blulab.chpc.utah.edu/KA/>. The user can create an account to access the tool by clicking on the \"Create Account\" link. The data model used by Knowledge Author can be found at <http://blulab.chpc.utah.edu/ontologies/SchemaOntology.owl>. The completed carotid stenosis semantic schema can be found at <http://blulab.chpc.utah.edu/ontologies/schemas/bscuba/carotid_stenosis.owl> and in the Additional file [2](#MOESM2){ref-type=\"media\"}.\n\nAbbreviations {#Sec23}\n=============\n\ncTAKES: clinical Text Analysis and Knowledge Extraction System; CTS: Common Type System; CUI: Concept Unique Identifier; eHOST: extensible Human Oracle Suite of Tools; HiTex: Health Information Text Extraction; MedLEE: Medical Language Extraction and Encoding System; NLP: Natural Language Processing; OWL: Web Ontology Language; Secondary Use CEM: Secondary Use Clinical Element Model; SWRL: Semantic Web Rule Language; UMLS: United Medical Language System; XML: Extensible Markup Language\n\nAdditional files {#Sec24}\n================\n\nAdditional file 1:Full list of use case concepts. (XLSX 12\u00a0kb)Additional file 2:Carotid Stenosis OWL file. (OWL 232\u00a0kb)\n\nWe would like to acknowledge Effective Dynamics for their excellent work programming the system. This work was supported by National Library of Medicine grant LM010964.\n\nAuthors\\' contributions {#FPar1}\n=======================\n\nWWC, MT and WS designed the Knowledge Author interface. WS was the software architect and project manager. WWC provided the vision for the project. MT managed all things related to the Schema and Modifier Ontologies. ET and MT coded the OWL file input and output. FD provided interface design support. YL coded the initial Knowledge Author prototype. MT created the use cases for testing. DM implemented and assessed the proof of concept study leveraging pyConText. MT, WS, DM, and WWC drafted the manuscript. All authors read and approved the final manuscript.\n\nCompeting interests {#FPar2}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "INTRODUCTION\n============\n\nMusculoskeletal disorders are one of the most frequent health problem for people around the world. These disorders have great consequences for public health because of their severe impact on body's disability, health care expenses, sick leave ([@b12-jer-13-2-240]). About 6% of the total healthcare costs ([@b10-jer-13-2-240]) and 30% of all sick-leave days are due to musculoskeletal problems ([@b11-jer-13-2-240]).\n\nRecently, research started to demonstrate a significant relationship between physical inactivity and variety of musculoskeletal disorders and health conditions. Several studies reported that regular aerobic exercise reduce the risk of suffering from various health conditions ([@b7-jer-13-2-240]; [@b8-jer-13-2-240]). With the greater interest of the benefits of physical activity, a more health conscious population expressed interest in exercise. In school surroundings, it is natural that students majoring sports have greater chance to participate in sport activities than nonmajoring students.\n\nIt is well known that regular exercise brings muscle strength and endurance which are two important components of muscular fitness. Diseases and disorders of the musculoskeletal system, such as osteoporosis, osteoarthritis, bone fractures, connective tissue tears, and low back syndrome, are related to physical inactivity and a sedentary lifestyle ([@b4-jer-13-2-240]).\n\nHowever, questions have been raised regarding the incidence of body deformity by participation in sports activities. For example, sport activities with high impact has been found to be a risk factor for sports injuries and deformities of the musculoskeletal system ([@b9-jer-13-2-240]).\n\n[@b1-jer-13-2-240] also demonstrated that body deformities are more prevalent in athletes than in nonathletes and proposed that body deformities develops during adolescence and is probably to be affected by high-impact sports practice.\n\nSpinal deformity alters normal spine biomechanics leading to global balance and possible detriment to quality of life ([@b3-jer-13-2-240]). Imbalances of spinal alignment have been reported to relate to postural instability ([@b5-jer-13-2-240]). Any spinal deformity has been shown to suffer from long-term functional disturbances and earlier onset back pain and disc degeneration than normal individuals. Spinal deformities might be considered to relate to the lower extremities, such as knee or hip flexion and extension positions ([@b6-jer-13-2-240]). However, no study has reported on the relationship between spinal deformities and lower extremities.\n\nTherefore, the purpose of the study was (a) to identify the incidence of each spinal deformity and lower extremity deformity and identify the relationship of the incidence between the deformities and (b) to identify the effects of major and physical characteristics (body mass index \\[BMI\\], gender, age, and dominant arm) on the development of 5 spinal deformities (forward head, protruding abdomen, kyphosis, lordodis, and scoliosis) and 4 lower extremity deformities (genu valgus, genu varus, genu recurvatum, and leg length discrepancy).\n\nMATERIALS AND METHODS\n=====================\n\nParticipants\n------------\n\nOne hundred forty-seven sports major students and 54 nonmajor students participated in this study. Among them 133 students were males and 68 students were females. Physical characteristics of the subjects are listed in [Table 1](#t1-jer-13-2-240){ref-type=\"table\"}.\n\nData collecting\n---------------\n\nAll subjects were screened and individuals were removed from the test if they had medical problems such as lumbago and previous lower extremity fracture. Each subject was asked wear pants and shirts only. No socks were permitted for the data collecting. The subjects were asked to answer each question on the questionnaire including college major, age, gender, height, and weight.\n\nExperimental procedures\n-----------------------\n\nThe visual postural examination of the spine and lower extremity was conducted from anterior, posterior, lateral, and prone positions. From the anterior view, height of the shoulder on both sides was observed. From the posterior view, any sign of scoliosis, genu valgus, genu varus, or genu recurvatum was observed. From the lateral view, any sign of forward head, protruding abdomen, kyphosis, lordosis was observed. From the prone position, observation was performed if there is leg length discrepancy. This procedure was established and validated by [@b13-jer-13-2-240].\n\nStatistical analyses\n--------------------\n\nA logistic regression was used to identify the association of genu valgus, genu varus, genu recurvatum, leg length discrepancy, forward head, protruding abdomen, kyphosis, lordosis, and scoliosis. Independent variables were major and nonmajor, gender, and BMI converted from the height and the weight.\n\nEach spinal and lower extremity deformities was considered as dependent variable in the study. Data were analyzed with IBM SPSS Statistics ver. 20.0 (IBM Co., Armonk, NY, USA) and statistical significance was set at *P*\\<0.05 for all tests.\n\nRESULTS\n=======\n\n[Table 2](#t2-jer-13-2-240){ref-type=\"table\"} shows the result of the relationship of deformity incidence between sport major and nonmajor. According to [Table 2](#t2-jer-13-2-240){ref-type=\"table\"}, the incidences of genu varus (*P*\\<0.00001), genu recurvatum (*P*= 0.007), forward head (*P*\\<0.00001), lordosis (*P*\\<0.00001), and scoliosis (*P*=0.0008) were significantly lower in nonmajor students compared to major students to the amount of 91%, 76%, 90%, 87%, and 73%, respectively.\n\nThe relationship of deformity incidence between gender is presented in [Table 3](#t3-jer-13-2-240){ref-type=\"table\"}. As shown in [Table 3](#t3-jer-13-2-240){ref-type=\"table\"}, the incidences of genu varus (*P*=0.0017), leg length discrepancy (*P*=0.0015), and forward head (*P*=0.005) were significantly greater in women than their counterpart to the amount of 4, 2.3, and 2 times, respectively. However, in genu recurvatum (*P*\\<0.00001), kyphosis (*P*= 0.0001), and lordosis (*P*=0.0024), the incidences were significantly lower in women to the 80%, 91.4%, 91%, respectively.\n\n[Table 4](#t4-jer-13-2-240){ref-type=\"table\"} displays the result of the relationship of deformity incidence according to BMI. As seen in [Table 4](#t4-jer-13-2-240){ref-type=\"table\"}, students in overweight (\\[BMI\\]\\>24.5 kg/m^2^) showed significantly high incidence (4.7 times, *P*=0.0044) in genu varus compared to students with normal BMI, and revealed significantly low incidence in genu recurvatum (*P*=0.0047) and protruding abdomen (*P*=0.0002) to the amount of 64.6% and 79% each in overweight students.\n\nDISCUSSION\n==========\n\nThe result of the study revealed that the incidence of spinal and lower extremity deformities in nonmajoring students showed significantly lower than sport majoring students. This result of the present investigation is consistent with the conclusion by [@b9-jer-13-2-240] and [@b1-jer-13-2-240]. It is believed that adolescent with less musculoskeletal maturity may experience a increased risk for more skeletal damage and structural abnormalities when exposed to intense levels of sports activities.\n\nDeformity incidence in women was significantly high in genu varus, leg length discrepancy, and forward head. On the contrary, the incidences of genu recurvatum, khphosis, and lordosis were significantly lower in women compared to their counterparts. These differences in compensatory mechanisms should be considered when evaluating and planning spinal and lower extremity deformities for further studies.\n\nGenu varum is a physical deformity marked by bowing of the leg. One of the risk factors of this musculoskeletal alignment is stress on the knee joint such as with exercise. [@b2-jer-13-2-240] hypothesized that the stress and strain on a joint caused by regular practicing of an intensive sport during adolescence may lead to a growth deformity. It is speculated that obese students participating sports activities to reduce the body weight may encounter genu varum before weight loss.\n\nCONFLICT OF INTEREST\n\nNo potential conflict of interest relevant to this article was reported.\n\n###### \n\nPhysical characteristics of subjects (n=201)\n\n  Characteristic   Major (n=147)   Nonmajor (n=54)\n  ---------------- --------------- -----------------\n  Sex                              \n  \u2003Male            102             31\n  \u2003Female          45              23\n                                   \n  Age (yr)         22.24\u00b11.47      \n                                   \n  Height (cm)      171.71\u00b10.03     \n                                   \n  Weight (kg)      67.88\u00b112.98     \n\nValues are presented as mean\u00b1standard deviation.\n\n###### \n\nRelationship of deformity incidence between major and nonmajor\n\n  Variable                 95% Confidence interval   Odd ratio   *P* for trend\n  ------------------------ ------------------------- ----------- ---------------\n  Genu valgus              \\-                        \\-          \\-\n  Genu varus               0.033--0.246              0.09        \\<0.0001\n  Genu recurvatum          0.085--0.681              0.24        0.007\n  Leg length discrepancy   0.287--1.034              0.56        0.063\n  Forward head             0.049--0.209              0.10        \\<0.0001\n  Protruding abdomen       0.441--4.441              1.40        0.568\n  Kyphosis                 \\-                        \\-          \\-\n  Lordosis                 0.056--0.302              0.13        \\<0.0001\n  Scoliosis                0.105--0.711              0.27        0.008\n\n###### \n\nRelationship of deformity incidence between gender\n\n  Variable                 95% Confidence interval   Odd ratio   *P* for trend\n  ------------------------ ------------------------- ----------- ---------------\n  Genu valgus              0.130--1.975              0.506       0.327\n  Genu varus               1.282--12.524             4.007       0.017\n  Genu recurvatum          0.005--0.086              0.200       \\<0.0001\n  Leg length discrepancy   1.183--4.591              2.330       0.015\n  Forward head             0.998--4.211              2.050       0.050\n  Protruding abdomen       \\-                        \\-          \\-\n  Kyphosis                 0.021--0.351              0.086       0.001\n  Lordosis                 0.158--0.879              0.373       0.024\n  Scoliosis                0.279--2.090              0.764       0.600\n\n###### \n\nRelationship of deformity incidence according to body mass index (established abnormal when over 24.5)\n\n  Variable                 95% Confidence interval   Odd ratio   *P* for trend\n  ------------------------ ------------------------- ----------- ---------------\n  Genu valgus              0.415--6.342              1.623       0.486\n  Genu varus               1.042--20.824             4.659       0.044\n  Genu recurvatum          0.127--0.987              0.354       0.047\n  Leg length discrepancy   0.704--2.586              1.350       0.366\n  Forward head             0.411--1.543              0.796       0.499\n  Protruding abdomen       0.079--0.558              0.210       0.002\n  Kyphosis                 0.159--1.663              0.514       0.266\n  Lordosis                 0.515--3.168              1.277       0.597\n  Scoliosis                \\-                        \\-          \\-\n"}
+{"text": "INTRODUCTION\n============\n\nMyofibrillar protein (MP) is composed of a variety of structural proteins including myosin, actin, tropomyosin, troponin, \u03b1-actinin, and desmin \\[[@b1-ajas-18-0585]\\]. It can be obtained from muscle cells extracted with a high salt solution (\\>0.6 M). MP has been strongly associated with the binding and water-holding capacity of meat and meat products \\[[@b2-ajas-18-0585]\\]. In addition, rheological properties of myofibrils are changed by electrostatic repulsion in a pH-dependent fashion. Other factors that influence MP gels are sodium concentration and non-protein polymer ingredients, such as hydrocolloids and fibers \\[[@b3-ajas-18-0585]\\].\n\nHydrocolloids including carrageenan, starch, gelatin, and xanthan can functionally improve rheological properties in the food system. They are generally used in comminuted meat products to improve emulsion stability, water/fat binding, and texture, and also can structurally interact with muscle protein. Gelatin can stabilize shrinkage and improve cooking yield due to their gelling and water binding properties \\[[@b4-ajas-18-0585]\\]. Carrageenan improves the textural properties and appearance of sausages \\[[@b5-ajas-18-0585]\\]. A study by Ruiz-Capillas et al \\[[@b6-ajas-18-0585]\\] demonstrated that konjac gel has used as a fat replacer in dry fermented sausages. Lin and Huang \\[[@b7-ajas-18-0585]\\] reported that konjac/gellan mixed gels were acceptable fat replacers in reduced-fat frankfurters, with a resulting positive sensory score and an adequate shelf life. Chin et al \\[[@b8-ajas-18-0585]\\] reported that a desirable gel structure was obtained by the combination of casein and soy protein in MP gels under the proper salt and pH conditions. Locust bean gum interacts synergistically with other hydrocolloids and acts as a thickening agent and a stabilizer in food systems \\[[@b9-ajas-18-0585]\\].\n\nCurdlan is a thermo-gel stable and water-soluble polysaccharide composed of \u03b2-(1,3)-linked glucosidic linkages and glucose, galactose, and mannose in an approximate molar ratio of 87.7:11.0:1.3 \\[[@b10-ajas-18-0585]\\]. It is produced by the microorganism *Alcaligenes faecalis* var. *myxogenes* 10C3. Curdlan is odorless, tasteless, and colorless \\[[@b11-ajas-18-0585]\\]. When aqueous curdlan solution is heated, its gel strength increases with increasing heating temperature and time \\[[@b12-ajas-18-0585]\\]. The adhesiveness and viscosity of emulsified meatballs are increased by the addition of curdlan (0.6%) \\[[@b13-ajas-18-0585]\\], which was effective as fat-replacer in non-fat sausages \\[[@b14-ajas-18-0585]\\]. Curdlan could influence the quality of meat products by improving the water-entrap and replacing the fat in thermal-irreversible gels \\[[@b15-ajas-18-0585]\\]. Finally, curdlan (1.0%) needed to improve texture in meat gels \\[[@b16-ajas-18-0585]\\].\n\nThe previous studies reported the benefits of curdlan in meat products. However, the influence of various concentrations of curdlan on quality of meat products has not been thoroughly studied. Thus, this study carried out to determine the optimum content of curdlan on the rheological properties of pork MP gels and application to low-fat model sausages.\n\nMATERIALS AND METHODS\n=====================\n\nExperiment 1: Interaction of MP and curdlan\n-------------------------------------------\n\n### Materials\n\nSlaughtered pork loins (Landrace\u00d7Yorkshire, grade A, Korea) on the same day were purchased from a local meat market (Samho Co., Gwangju, Korea). Visible fat and connective tissues were removed. The meat was formed into 1 to 2 cm^3^ cubes and stored in a --50\u00b0C freezer until untilized. Curdlan was provided by the Sncfood company (Seoul, Korea).\n\n### Processing of myofibrillar protein gels\n\nThe frozen pork loin was thawed overnight at 4\u00b0C. The MP was adjusted to a concentration of 40 mg/mL after extraction for three times using 0.1 M NaCl in 50 mM phosphate buffer by food mixer (Bowl Rest mixer, Hamilton Beach/Proctor-Silex, Inc., Southern Pines, NC, USA) for 1.5 min each time. The resulting emulsion was washed with 0.1 M NaCl buffer using cheesecloth to remove connective tissue \\[[@b17-ajas-18-0585]\\]. MP, buffer, and curdlan (0.5%, 1.0%, and 1.5%) were mixed in the proportions, and 5 mL of the protein mixture was loaded into vial s which were heated from 20\u00b0C to 80\u00b0C with an incremental temperature increase of 3\u00b0C/min in a water bath (WB-22, Daihan Scientific Co., Seoul, Korea). The cooked MP gels were quickly chilled in an ice water and stored at refrigerator temperature (4\u00b0C\u00b12\u00b0C) until analyzed.\n\n### Evaluation of raw or cooked myofibrillar protein gels\n\nApproximately 3 mL of each sample mixture was loaded into the probe container of the concentric cylinder type rotational rheometer (RC30, Rheotec Messtechnik GmbH, Ottendorf-Okrilla, Germany). The shear rate was increased from 0 to 300/s for 60 s. The shear rate was plotted against shear stress using Excel 2016 software (Microsoft Corporation, Redmond, WA, USA). Cooking loss (CL) of MP gels heated as described above was determined as the average of five different samples and calculated as the weight before cooking compared to the weight after cooking.\n\nGel strength of cooked MP gels was measured by Universal Testing Machine (3344, Instron Corporation, Norwood, MA, USA). The breaking force (gf) of five different samples was determined at a head speed of 500 mm/min by steel drill chuck (33BA 1/2--20, Jacobs chuck, Sparks, MD, USA). The expressed value was the average of the five determinations.\n\nThe three-dimensional structures of MP gels were measured by low-vacuum scanning electron microscopy (LV-SEM) (JSM-6610LV, JEOL Ltd., Tokyo, Japan). Cubed samples (3\u00d73\u00d73 mm^3^) were immersed in 2.5% glutaraldehyde (pH 7) and 0.1 M phosphate buffer solution overnight at 4\u00b0C and then in osmium tetroxide (pH 7) and 0.1 M phosphate buffer for 5 h. After rinsing three times, each sample was dehydrated by immersion for 10 min in a sequentially increasing series of ethanol solutions (50%, 60%, 70%, 80%, 90%, and 100%). Each sample was completely dried by immersion in acetone for 10 min. The dehydrated and dry samples were each gold-coated using a model 108 auto sputter coater (Cressington Scientific Instruments Ltd., Watford, England) prior to LV-SEM microstructure examination.\n\n### Statistical analysis\n\nThe whole experiment was performed in triplicates. Data were analyzed by one-way analysis of variance (ANOVA) followed by Duncan's test using SPSS 20.0 statistical software at a significant level of 0.05 (SPSS Inc., Chicago, IL, USA). A p-value \\<0.05 indicated a significant difference.\n\nExperiment 2: Quality of low-fat sausage prepared with curdlan\n--------------------------------------------------------------\n\n### Materials\n\nPork ham (Landrace\u00d7Yorkshire, Grade A) was purchased for each replication from the aforementioned local market. All visible fat and connective tissue were removed and the remaining meat was ground using meat chopper (M-12S, Hankook Fujee Industries Co., Ltd., Busan, Korea). Curdlan was obtained from the aforementioned source.\n\n### Processing of low-fat model sausages\n\nModel sausages were manufactured with 60% pork, 38% water, 1.3% salt, 0.4% sodium tripolyphosphate, 0.05% sodium erythorbate, 0.25% curing salt and respectively 0%, 0.5%, 1%, and 1.5% curdlan as followed by Lee and Chin study \\[[@b18-ajas-18-0585]\\]. The ground pork ham was mixed with curing ingredients using a cutter (HMC-401, Hanil Electric, Seoul, Korea) for 3 min. The emulsified mixture was stuffed in 45 mL plastic tubes and cooked until reaching 72\u00b0C at the geometric center of each sample. Fully cooked samples were kept in an ice water while they were completely cooled and then stored in the refrigerator (4\u00b0C\u00b12\u00b0C) until used.\n\n### Evaluations of low-fat model sausages\n\n#### i) Chemical composition, pH, and color\n\nMoisture, fat, and protein contents were measured using AOAC methods \\[[@b19-ajas-18-0585]\\]. pH was determined five times from randomly selected samples (n = 5) using the pH-meter (MP-120, Mettler-Toledo, Greifeense, Switzerland) with Inlab 413/IP 67 electrode. Color values (Commission Internationale de l'Eclairage \\[CIE\\]) lightness \\[L\\*\\], redness \\[a\\*\\], yellowness \\[b\\*\\]) were calculated as the average of six determinations made on the internal surface of samples using the color reader (CR-10, Minolta, Tokyo, Japan). Color reader calibrated with white plate standard (L\\* = 95.6, a\\* = 1.0, and b\\* = 0.2). And it characterizes standard illuminant D~65~ at an observation angle of 8\u00b0.\n\n#### ii) Cooking loss (%) and expressible moisture (%)\n\nThe CL (%) was calculated as the weight before cooking compared to the weight after cooking. Expressible moisture (EM, %) was evaluated as previously described \\[[@b20-ajas-18-0585]\\]. Sausage samples (n = 4) were prepared as cubes with an approximate weight of 1.5 g. Prepared samples were wrapped in three pieces of Whatman \\#3 filter paper and centrifuged at 1,660\u00d7g for 15 min in a centrifuge (VS-5500, Vision Science Co., Ltd, Seoul, Korea). The EM was calculated from the weight before and after centrifugation.\n\n#### iii) Textural properties\n\nSausage samples (n = 10) were shaped 13 mm in height and 12.5 mm in diameter were made using a puncturing tool (12.5 mm diameter). Texture profile analysis was done using the aforementioned Instron Universal Testing Machine (Model 3344 testing device, Norwood, MA, USA). Samples were compressed two times until the height was 75% of the original height. A compression probe was used with a 500 N load cell at a crosshead speed of 300 mm/min. Hardness (gf), springiness (mm), gumminess, chewiness, and cohesiveness were expressed as previously detailed \\[[@b21-ajas-18-0585]\\].\n\n### Statistical analysis\n\nOne-way ANOVA followed by Duncan's test was done using the aforementioned SPSS 20.0 statistical software (SPSS Inc., Chicago, IL, USA). Each experiment was performed in triplicate. A p-value \\<0.05 indicated a significant difference.\n\nRESULTS AND DISCUSSION\n======================\n\nEvaluation of myofibrillar protein gel\n--------------------------------------\n\n### Cooking loss and gel strength\n\nThe CL and gel strength of MP mixed gels are shown in [Table 1](#t1-ajas-18-0585){ref-type=\"table\"}. The MP gels formulated with increased levels of curdlan decreased CL (p\\<0.05). During the formation of curdlan gel from a powder, free water moves into curdlan micelles and is bound by hydrogen bonding, resulting in improved water-binding capacity and a strong gel structure in the presence of a protein mixture including MP \\[[@b22-ajas-18-0585]\\]. Funami et al \\[[@b23-ajas-18-0585]\\] reported that heat-stable curdlan gel formed in meat products and hence its structures held water tightly. The gel strength progressively increased with the addition of the increased level of curdlan up to 1.0%. No differences in gel strength of MP get with 1.0% and 1.5% curdlan were observed. The gel of curdlan had much triple helices micelles with junction zone, which could be formed the gel \\[[@b24-ajas-18-0585]\\]. Thus, forming the hydrogen bonded triple helix junction zones between curdlan and MP resulted in the increased gel strength. High-set curdlan gels, which were heated at 80\u00b0C, were progressed with hydrophobic bonding among curdlan molecules \\[[@b22-ajas-18-0585]\\]. Curdlan became swollen during absorbing the water and gave the pressure among the gel matrix, thereby increasing the gel strength \\[[@b25-ajas-18-0585]\\]. Gel strength can be increased by adding hydrocolloids such as konjac flour, which was a typical hydrocolloid for improving the gel strength in MP \\[[@b26-ajas-18-0585]\\]. Since curdlan is also classified as a hydrocolloid, gel strength could be increased by adding curdlan on MP gels.\n\nViscosity\n---------\n\nThe viscosity of MP mixtures with various levels of curdlan is shown in [Figure 1](#f1-ajas-18-0585){ref-type=\"fig\"}. MP mixtures with 0.5% curdlan had low shear stress compared to those with higher concentrations of curdlan. MP mixture without curdlan showed similar value to those with 0.5% curdlan. However, shear stress of MP mixtures containing 1.0% and 1.5% curdlan was higher than that of curdlan-free control. Lo et al \\[[@b27-ajas-18-0585]\\] also reported that apparent viscosity increased with increasing concentration of curdlan, which caused intermolecular cross-linking. These rigid three-dimensional structures had limited to mobile the polymer chain, indicating the increasing the shear stress by shear rate. Lee et al \\[[@b28-ajas-18-0585]\\] reported that several carbohydrates combined well with protein and water, and modified the viscosity of the protein mixture. The viscosity of meatballs was also increased with increased levels of curdlan from 0% to 0.6% \\[[@b13-ajas-18-0585]\\], which was attributed, at least in part, to the sticky properties of curdlan before heating. Water-dissolved curdlan had sticky properties, which was easy to interpenetrate among the three-dimensional meat protein structures, resulting in a bonding tightly and making a firm gel. As viscosity is related to mouthfeel during mastication, these results can attribute to the desirable texture when applied to meat products.\n\nMicrostructure\n--------------\n\nMicrostructures of MP gels with various levels of curdlan are shown in [Figure 2](#f2-ajas-18-0585){ref-type=\"fig\"}. The MP gels generally adopted a globular structure. In the presence of curdlan, the MP microstructure appeared flat and less porous and curdlan might be occupied the MP pores, resulting in a smooth and swollen surface ([Figure 2b, 2c, 2d](#f2-ajas-18-0585){ref-type=\"fig\"}). The MP gels containing 1.0% and 1.5% curdlan ([Figure 2c, 2d](#f2-ajas-18-0585){ref-type=\"fig\"}) showed denser and filling-up appearance, as compared to the control ([Figure 2a](#f2-ajas-18-0585){ref-type=\"fig\"}). Although curdlan didn't affect the thermal stability of meat, it showed a well-mixed structure between curdlan and meat protein after heating \\[[@b23-ajas-18-0585]\\]. Some of the pores were formed because the excess curdlan was not combined with MP and aggregated independently. This might be detrimental to the texture of model sausages formed with 1.5% curdlan. A globular structure was produced in the presence of 0.5% curdlan ([Figure 2b](#f2-ajas-18-0585){ref-type=\"fig\"}), but the structure was less porous compared to the control ([Figure 2a](#f2-ajas-18-0585){ref-type=\"fig\"}). These results agreed with those of Wu et al \\[[@b25-ajas-18-0585]\\], who reported that curdlan molecules filled to the pores around protein matrix and it became a more compact and denser homogenous network through hydrogen bonding, resulting in a high gel strength and exemplary water retention. The natural state of curdlan granule had formed the doughnut-shaped structure, similar with starch structure \\[[@b29-ajas-18-0585]\\]. Rearrangements of the dense structure have been attributed to curdlan and these structures might affect results of gel strength and CL \\[[@b30-ajas-18-0585]\\].\n\nCharacteristics of low-fat model sausage containing various levels of curdlan\n-----------------------------------------------------------------------------\n\n### pH, color, chemical composition, and textural properties\n\npH and color of LFSs are shown in [Table 2](#t2-ajas-18-0585){ref-type=\"table\"}. pH and lightness (L\\*) values did not differ among the treatments (p\\>0.05). However, redness (a\\*) and yellowness (b\\*) values of sausages with curdlan levels lower than 1.0% differ from those with 1.5% curdlan (p\\<0.05). These results indicated that curdlan level influences the color values. Hydrocolloids, such as carboxymethyl cellulose and sodium alginate, were reported to affect redness and yellowness values except for lightness \\[[@b31-ajas-18-0585]\\]. Moisture, fat, and protein contents did not differ among the treatments (p\\> 0.05). Calliari et al \\[[@b32-ajas-18-0585]\\] reported that chicken patties made with 5% polymer from *Agrobacterium radiobacter* k84 showed no differences in moisture and protein contents compared to polymer-free patties. The polymer of *Agrobacterium radiobacter* k84 is predominantly composed of minerals (40%) followed by carbohydrates (35%) and protein (15%) \\[[@b33-ajas-18-0585]\\]. In this study, no differences in the chemical compositions were observed in model sausages with various curdlan levels. Cardoso et al \\[[@b34-ajas-18-0585]\\] reported that carrageenan at 1.0% did not affect the chemical composition of cod frankfurter sausages.\n\nTextural properties of LFSs are shown in [Table 3](#t3-ajas-18-0585){ref-type=\"table\"}. No differences in most textural characteristics of LFSs were observed (p\\>0.05), except for hardness values, which were different among the curdlan levels. The hardness of LFSs with 0.5% and 1.0% curdlan was higher than those without curdlan. By increasing contents of curdlan, the number of junction zone by hydrophobic bonding between curdlan and MP was increased and hence rigid structures appeared \\[[@b15-ajas-18-0585]\\]. However, hardness was decreased with the addition of 1.5% curdlan, partially due to the disrupted protein-protein interaction (p\\<0.05), resulting in a more crumbled texture. Thus, excessive curdlan addition may disturb the cross-linking of the salt soluble protein, resulting in difficulties to form the strong gels \\[[@b25-ajas-18-0585]\\]. Adding 1.0% amidated low methoxyl (ALM) pectin increased the firmness of restructured fish gels, but additional levels of ALM pectin disrupted the interaction with fish meat gels \\[[@b35-ajas-18-0585]\\]. Adhesiveness, chewiness, and gumminess values of emulsified meatballs were increased with increasing levels of curdlan \\[[@b13-ajas-18-0585]\\]. Kappa-carrageenan (0.5% to 2.0%) in low-fat meatballs increased hardness, adhesiveness, chewiness, and gumminess \\[[@b36-ajas-18-0585]\\]. Montero et al \\[[@b31-ajas-18-0585]\\] also reported that kappa-carrageenan and sodium alginate increased hardness, adhesiveness, and cohesiveness in blue whiting gels. Thus, an appropriate addition of hydrocolloid in the manufacture of meat products may improve the textural properties.\n\nCooking loss and expressible moisture\n-------------------------------------\n\n[Figure 3](#f3-ajas-18-0585){ref-type=\"fig\"} shows the result of CL (%) of low-fat model sausages with various curdlan contents. CL of LFSs made with the three levels of curdlan were lower than those without curdlan (p\\< 0.05). This result was supported by Kimura et al \\[[@b37-ajas-18-0585]\\] who reported that the addition of curdlan in sausage increased the product yield. Funami and Nakao \\[[@b16-ajas-18-0585]\\] observed that curdlan retained free and released water during cooking. This partially reflects the formation of hydrophobic interactions and hydrogen as affected by thermal stability \\[[@b38-ajas-18-0585]\\]. Microfibrils, which are from curdlan, had broken the structure at 60\u00b0C and rearranged the structure at the higher temperature by hydrophobic interactions \\[[@b24-ajas-18-0585]\\]. Funami et al \\[[@b23-ajas-18-0585]\\] reported that released water during cooking was decreased as the levels of curdlan increased in minced pork gels. Curdlan was mixed well with meat proteins by chopping progress and the meat proteins were rearranged the curdlan-protein gel matrix during the heating. Increasing these cross-linking in meat gel system may affect to hold the water molecules strongly \\[[@b25-ajas-18-0585]\\]. Thus, CL reduced by adding curdlan since curdlan had the high force to entrap the water within meat protein.\n\nAlthough curdlan can hold water strongly among meat protein molecules, no differences in EM were observed among treatments, regardless of adding curdlan (p\\>0.05) ([Figure 3](#f3-ajas-18-0585){ref-type=\"fig\"}). Generally, water-holding capacity of curdlan was reportedly 3--5 times higher than egg white in minced pork gel \\[[@b23-ajas-18-0585]\\]. Hydration, hydrogen bonds, hydrophobic interactions, and Van der Waals forces between carbohydrates and proteins were thought to affect the water binding ability of protein gels \\[[@b39-ajas-18-0585]\\]. In addition, the triple helix of curdlan was reinforced the interstitial water with hydrogen bond \\[[@b22-ajas-18-0585]\\].\n\nCONCLUSION\n==========\n\nThe MP containing 1.0% and 1.5% curdlan increased the viscosity and gel strength, and CL was decreased in LFSs formulated with all three levels of curdlan. The addition of curdlan into MP or LFS resulted in a reduction in CL and therefore, curdlan improved overall quality characteristics of LFSs by retaining the water. Hardness also increased in low-fat model sausages with 0.5% and 1.0% curdlan. However, the excess addition of curdlan reduced the textural characteristics. Thus, 1.0% of curdlan is considered as an optimal level in improving the physicochemical properties of low-fat model sausages. Further study is needed to determine the mechanism of the negative effect of a high concentration (1.5%) of curdlan on LFSs.\n\n**CONFLICT OF INTEREST**\n\nWe certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.\n\nThis study was financially supported by National Research Foundation (NRF Project \\# 2017 R1D1A1B03035698).\n\n![The viscosity of myofibrillar protein gel with various contents of curdlan.](ajas-18-0585f1){#f1-ajas-18-0585}\n\n![Scanning electron microscopy (\u00d72,000) of myofibrillar protein with various contents of curdlan (a) control, (b) curdlan 0.5%, (c) curdlan 1.0%, (d) curdlan 1.5%.](ajas-18-0585f2){#f2-ajas-18-0585}\n\n![Cooking loss (%) and expressible moisture (%) of low-fat model sausages as affected by different levels of curdlan. ^a,b^ Means with diferent letters differ at p\\<0.05.](ajas-18-0585f3){#f3-ajas-18-0585}\n\n###### \n\nCooking loss and gel strength of myofibrillar protein gel with various content of curdlan\n\n  Items               Curdlan (%)                                                                                                                                                            \n  ------------------- ------------- -------------------------------------------------- -------------------------------------------------- -------------------------------------------------- --------------------------------------------------\n  Cooking loss (%)    Mean          8.46[a](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}   5.78[b](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}   4.07[c](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}   2.64[d](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}\n                      SD            0.25                                               0.24                                               0.87                                               0.25\n  Gel strength (gf)   Mean          260[c](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}    297[b](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}    354[a](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}    385[a](#tfn2-ajas-18-0585){ref-type=\"table-fn\"}\n                      SD            18.0                                               9.83                                               4.41                                               17.2\n\nSD, standard deviation.\n\nMeans (n = 3) having same superscripts in a same row are not different (p\\>0.05).\n\n###### \n\npH and color values of low-fat model sausages as affected by different levels of curdlan\n\n  Items      Curdlan (%)                                                                                                                                                            \n  ---------- ------------- -------------------------------------------------- -------------------------------------------------- -------------------------------------------------- --------------------------------------------------\n  pH value   Mean          6.24[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   6.23[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   6.23[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   6.22[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}\n             SD            0.16                                               0.18                                               0.18                                               0.17\n  CIE L\\*    Mean          74.5[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   73.5[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   73.5[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   74.6[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}\n             SD            0.11                                               0.57                                               1.01                                               0.61\n  CIE a\\*    Mean          9.96[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   10.3[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   9.78[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   8.95[b](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}\n             SD            0.27                                               0.32                                               0.32                                               0.26\n  CIE b\\*    Mean          5.08[b](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   5.36[b](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   5.21[b](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}   6.34[a](#tfn4-ajas-18-0585){ref-type=\"table-fn\"}\n             SD            0.42                                               0.18                                               0.27                                               0.06\n\nSD, standard deviation; CIE, Commission Internationale de l'Eclairage.\n\nMeans (n = 3) having same superscripts in a same row are not different (p\\>0.05).\n\n###### \n\nTextural properties of low-fat model sausages as affected by different levels of curdlan\n\n  Items             Curdlan (%)                                                                                                                                                               \n  ----------------- ------------- --------------------------------------------------- --------------------------------------------------- --------------------------------------------------- ---------------------------------------------------\n  Hardness (gf)     Mean          1,901[b](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}   2,251[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}   2,311[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}   1,764[b](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}\n                    SD            32.0                                                87.9                                                150                                                 177\n  Springiness(mm)   Mean          6.68[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    6.82[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    6.60[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    7.06[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}\n                    SD            1.32                                                0.41                                                1.01                                                0.64\n  Gumminess         Mean          14.9[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    16.7[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    18.3[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    14.4[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}\n                    SD            0.68                                                0.98                                                3.04                                                2.63\n  Chewiness         Mean          98.9[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    109[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}     118[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}     102[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}\n                    SD            27.4                                                2.21                                                30.2                                                27.4\n  Cohesiveness      Mean          7.77[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    7.59[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    7.78[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}    7.79[a](#tfn6-ajas-18-0585){ref-type=\"table-fn\"}\n                    SD            0.47                                                0.55                                                0.37                                                0.25\n\nSD, standard deviation.\n\nMeans (n = 3) having same superscripts in a same row are not different (p\\>0.05).\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nKnee osteoarthritis is a degenerative disease, in the form of a \"wear and tear\" process in the joints as a result of the ageing process and is local \\[[@ref1]\\]. At the molecular level, the imbalance between catabolic and anabolic activities where the primary injury response occurs in joint cartilage resulting in osteoarthritis \\[[@ref2]\\]. The expression of several genes involved in the inflammatory response and cartilage degradation, such as IL-1 and TNF-\u03b1 is regulated predominantly by Nuclear Factor Kappa Beta (NF\u03ba\u03b2). NF\u03ba\u03b2 stimulates the cytokines of TNF-\u03b1 and IL-1\u03b2 which contribute to the inflammatory process in osteoarthritis. NF\u03ba\u03b2 is also needed for the transcription of the Matrix Metalloproteinase (*MMP-13*) gene \\[[@ref3]\\]. *RELA* is a subunit of the NF\u03ba\u03b2 p65 gene which plays an important role in the pathogenesis of osteoarthritis.\n\nThe rapid development of stem cell science has broadened the picture of potential stem cells in the world of research and the medical world, a number of characteristics that stem cells have proven to provide great hope for healing many people who suffer from diseases that are no longer possible to be treated conservatively or operatively, especially degenerative diseases and abnormalities such as trauma, malignancy and so on which also increase dramatically \\[[@ref4]\\].\n\nThis study aims to look at the expression of the *RELA* gene and expression of the *MMP-13* gene and analyse the relationship of *RELA* gene expression with *MMP-13* gene expression after administration of Mesenchymal Stem Cell Wharton Jelly to synoviocytes in vitro.\n\nMaterial and Methods {#sec1-2}\n====================\n\nThis research is a pure experimental study which is divided into 6 treatment groups with 4 number of replications. Group I was a control group of Osteoarthritis (OA) synoviocytes (incubated for 24 hours), group II was a control group of OA synoviocyte cells incubated for 48 hours, group III was a Mesenchymal Stem Cell Wharton Jelly (MSC-WJ) cell group incubated for 24 hours, group IV is a cell group Wharton's Jelly Mesenchymal Stem cell (MSC-WJ) were incubated for 48 hours, the group V is a group of co-culture synoviocyte-MSC-WJ cells were incubated for 24 hours and the group VI is a group of co-culture cell synoviocyte-MSC-WJ were incubated for 48 hours. The number of cells used in each treatment group was 105cells, each for synoviocytes and MSC-WJ cells. Mesenchymal Stem Cell Wharton Jelly comes from IMERI (Indonesian Medical Education and Research Institute), Faculty of Medicine, University of Indonesia. Synoviocyte cells are derived from synovial tissue of patients with grade IV Osteoarthritis who undergo knee joint surgery (Total Knee Replacement) at Dr Hospital. M. Djamil Padang, Indonesia. The synoviocyte cells taken for treatment were the result of 3rd phase cell culture. The samples taken did not use informed consent because the samples used were stored biologically after post-knee joint surgery Osteoarthritis Grade IV. Samples were taken from six patients with male sex aged 40-70 years.\n\nIsolation of OA Primary Cells {#sec2-1}\n-----------------------------\n\nSynovial tissue and search are obtained from OA patients after Total Knee Arthroplasty. Ten samples were used for experiments. Synovial tissue is planted in the well plate with 10% Fetal Bovine Serum (FBS), 1% penicillin/streptomycin and 1% fungizone in Dulbecco's Modified Eagle's Medium (DMEM, Life Technologies) which is planted with an explant planting system. Cells were sub-cultured three times, and and the result of 3rd sub-culture was used for treatment. Each experiment was repeated for three times.\n\nCulture with stem cells with OA primary cells {#sec2-2}\n---------------------------------------------\n\nOA primary cells were cultured to 50--60 % confluence, then cultured together with mesenchymal stem cells Wharton jellies. These cells are observed for 24 hours and 48 hours. Cells are calculated by Haemocytometer with 105cells/well.\n\nPrimary design {#sec2-3}\n--------------\n\nNo. Primer Nucleotide Sequence NM Amplicon, NCBI Accession Size, Number Gene.\n\nMMP-13F 5'-CACTTTATGCTTACTGATGACG-3' NM_002427.3 154 bpMMP-13R 5'-TCCTCGGAGACTGGTAATGG-3' NM_002427.3 154 bpRELA F 5'-CGCATCCAGACCAACAACAA-3' NM_001243984.1 154 bpRELA R 5'-AGATGGGATGAGAAAGGACAGG-3' NM_001243984.1 154 bpHPRT1 5'-CCTGGCGTCGTGATTAGTGAT-3' NM_000194.2 158 bpHPRT1 5'-CCCATCTCCTTCATCACATCTC-3'. NM_000194.2 158 bp\n\nPCR Gradient Amplification {#sec2-4}\n--------------------------\n\nEach gene was replicated with the SYBR Green amplification kit. The PCR program is as follows: Predenaturation 95.0\u00b0C for 30 seconds, 5 seconds denaturation, gradient annealing at 55\u00b0C for 5 seconds for 50 cycles, additional melting curve 65.0\u00b0C-95.0\u00b0C with an increase of 0.5\u00b0C every 5 seconds.\n\nRNA extraction and cDNA synthesis {#sec2-5}\n---------------------------------\n\nRNA was extracted from synovial tissue isolates of grade IV Osteoarthritis patients. RNA isolation using TRIzol\u00ae, Invitrogen Life Technologies. Synthesis of cDNA was performed by using iScript cDNA Synthesis Kit (BioRad, USA) on thermal cycler C1000 (BioRad, USA) Reverse Transcriptase PCR (RT-PCR) devices.\n\nRNA Isolation {#sec2-6}\n-------------\n\nCells that have been treated with stem cells trizol as much as 500 \u00b5l. Then, homogenized and put in a 1.5 ml PCR tube. After that, cells that were given trizol, added 100 \u00b5l of chloroform, then homogenized. Once homogeneous, incubation for 5 min and centrifuged at a speed of 12.000 xg, for 15 minutes at 4\u00b0C. Take a clear layer, then move it to a new tube. Add 250 \u00b5l isopropanol, shake back and forth and let stand for 10 minutes. Followed by 12.000 xg speed centrifuge, for 10 minutes at 4\u00b0C. Pour on a dry, airy tissue. After that, add 500 \u00b5l of ethanol 75%, shake, and centrifuge 7.500 xg speed, for 5 minutes at 4\u00b0C. Dispose of ethanol, dry all for 15 minutes. After that add 25 Rna-se free water, dilute it. Then adjust the concentration of RNA using NanoDrop.\n\ncDNA synthesis {#sec2-7}\n--------------\n\nThe synthesis composition of total cDNA was 5 \u00b5g total RNA, 1x RT buffer 20 pmol oligodT, 4 mM dNTP, 10 mM DTT, 40 U enzyme SuperScript TMII RTase and H~2~O-DEPC with a total volume of 20 \u03bcl. Total cDNA synthesis was carried out at 52\u00b0C for 50 minutes with the work protocol by the manual kit BioRad, USA).\n\nPCR Gradient Amplification {#sec2-8}\n--------------------------\n\nEach gene was replicated with the SYBR Green amplification kits. The PCR program is as follows: Predenaturation 95\u00b0C for 30 seconds, 5 seconds denaturation, gradient annealing at 55\u00b0C for 5 seconds for 50 cycles, additional melting curve 65-95\u00b0C with an increase of 0.5\u00b0C every 5 seconds. Measurement of gene concentration can use two methods, namely the absolute quantification method and relative quantification method. The method used in this study is the relative quantification method of Livak-Schmittgen (2001) or the comparison of Treshhold deltas or plafl methods \\[[@ref5]\\].\n\n\u0394CT experimental = CT target in experiment-CT housekeeping on experiments\n\n\u0394CT control = CT target on control-CT housekeeping on control\n\n\u0394\u0394CT = \u0394CT experimental-\u0394CT control\n\nThe comparison of expression levels is obtained by using the equation: Comparison of gene expression levels = 2\u0394\u0394CT. The measurement of gene concentration in this study uses the LightCycler\u00ae software program which refers to the quantification of the automated Livak formula so that the concentration value of gene expression is obtained in picogram size. HPRT1 gene was a housekeeping gene, and calibrator gene was from the control group.\n\nData analysis {#sec2-9}\n-------------\n\nData will be presented in the form of table and graph, as well as the results of expressions of the RELA gene and MMP-13. P is no data of gene expression of MMP-13 and RELA normality test by using the Shapiro-Wilk test and homogeneity of data with the Levene test. The test decision criteria in the Shapiro-Wilk Test are if the value of p \\> 0.05 then it is said the data is normally distributed, while the test decision criteria in the Levene Test are if the value of p \\> 0.05, then the data is said to be homogeneous. For MMP-13 gene expression or normally distributed homogeneously, then continued with ANOVA Test and Tukey's HSD Post Hoc Test, but for RELA gene expression, the data were not normally distributed and homogeneous, then non-parametric Kruskal Wallis test was carried out and followed by Mann Test Whitney \\[[@ref6]\\]. Data is processed using SPSS 15 statistical analysis.\n\nResearch Ethics Requirements {#sec2-10}\n----------------------------\n\nResearch permits will be submitted to the International Committee of the ethics Faculty of Medicine, the University of Andalas which later if approved will receive a registration number as the approval of the Research Ethics Committee of the Faculty of Medicine, University of Andalas. By the research problems, the research object used is a network sample stored in BioBank's FK Unand network. Before submitting to the Ethics Committee, researchers submitted a letter of approval from the Cancer Study Center and Stem Cell as the rightful owner of a sample of knee osteoarthritis sufferers. In this study, researchers will involve competent experts in the field of orthopaedic surgery.\n\nResult {#sec1-3}\n======\n\nSample Characteristics {#sec2-11}\n----------------------\n\nThe result of synoviocyte isolation from synovial tissue was a fibroblast-shaped cell, cultured in a plate. The morphology of synoviocyte and MSC-WJ presented in [Figure 1](#F1){ref-type=\"fig\"}. while the morphology of synoviocyte co-culture MSC-WJ 24 hour and 48 hour and data on Characteristics of Mesenchymal Stem Cells Wharton Jelly presented in [Figure 2](#F2){ref-type=\"fig\"} and [Figure 3](#F3){ref-type=\"fig\"}.\n\n![Morphology cells A) Cell synoviocyte and (B) MSC-WJ](OAMJMS-7-543-g001){#F1}\n\n![Morphology of synoviocyte co-culture MSC-WJ (A) Co-culture 24 hour and (B) Co-culture 48 hour](OAMJMS-7-543-g002){#F2}\n\n![Data on Characteristics of Mesenchymal Stem Cells Wharton Jelly; (A) Cells MSC-WJ reach confluence. Scale bar: 500 \u00b5M. Photographs of cells taken using a Nikon Ti-S microscope; (B) Data flow cytometry. Forward scatter (FCS) plot&side scatter (SSC) plot. Population gated events (P1): 20,000; (C) Cell surface markers expression: CD73-APC 99.8% and CD105- PerCP-Cy5.5 95%; (D) Cell surface markers expression: CD90-FITC 99.9% and Lin (-) - PE 0.4%](OAMJMS-7-543-g003){#F3}\n\nExpression of RELA and MMP-13 Genes {#sec2-12}\n-----------------------------------\n\nFrom the results of the research obtained, before the analysis, a preliminary test was carried out for the basic assumptions of normality and homogeneity of data. The results of the normality test with the Shapiro Wilk Test obtained a significant value of \u2265 0.05 in all treatment groups, meaning that data was normally distributed.\n\nThis is supported by descriptive analysis with the Skewness ratio value \\< 2 in all treatment groups. To test the variance homogeneity based on the Levene Test is 0.18 \u2265 0.05, this means data on gene research RELA and MMP-13 have the same variant (homogeneous).\n\nFrom RELA gene expression data and MMP-13 gene expression data, the Pearson correlation test was conducted to see the relationship between RELA gene expression and MMP-13 gene expression after Mesenchymal Stem Cell Wharton Jelly administration.\n\nThe relationship between RELA gene expression and MMP-13 gene expression {#sec2-13}\n------------------------------------------------------------------------\n\nTo see the direct relationship between RELA gene expression and MMP-13 gene, Pearson correlation test was conducted which can be seen in [Table 3](#T3){ref-type=\"table\"}.\n\n###### \n\nAnalysis of the relative expression of MMP-13 target genes using the Livak-Schmittgen method (2001) \\[[@ref5]\\]\n\n  Groups                         C ~T~ MMP13 average   C ~T~ average HPRT1   \u0394C ~T~ = MMP13-HPRT1   \u0394\u0394C ~T~ = \u0394C ~T~ treatment - \u0394C ~T~ control   2 ^-^ \u0394\u0394C ~T~\n  ------------------------------ --------------------- --------------------- ---------------------- --------------------------------------------- ---------------\n  Synoviocyte control 24 hours   46.89                 35.89                 11.00                  0.00                                          1\n  MSC-WJ 24 hours                45.89                 34.39                 11.50                  0.50                                          0.70\n  *Co-culture* 24 hours          47,18                 33.34                 13,84                  2.84                                          0.14\n  Synoviocyte control 48 hours   43.09                 33.95                 9.14                   0                                             1\n  MSC-WJ 48 hours                43.41                 33.44                 9.97                   0.83                                          0.56\n  *Co-culture* 48 hours          46.55                 32.26                 14.29                  5.15                                          0.03\n\n###### \n\nAnalysis of relative RELA target gene expression using the Livak-Schmittgen method (2001)\n\n  Groups                         C ~T~ MMP13 average   C ~T~ average HPRT1   \u0394C ~T~ = MMP13-HPRT1   \u0394\u0394C ~T~ = \u0394C ~T~ treatment - \u0394C ~T~ control   2 ^-^ \u0394\u0394C ~T~\n  ------------------------------ --------------------- --------------------- ---------------------- --------------------------------------------- ---------------\n  Synoviocyte control 24 hours   35.72                 33.29                 2.43                   0                                             1\n  MSC-WJ 24 hours                35,64                 32.50                 3.14                   0.71                                          0.61\n  *Co-culture* 24 hours          36.35                 33.34                 3.01                   0.58                                          0.67\n  Synoviocyte control 48 hours   35.94                 33.95                 1.99                   0.00                                          1\n  MSC-WJ 48 hours                38,56                 33.90                 4.66                   2.67                                          0.1 6\n  *Co-culture* 48 hours          37.09                 32.41                 4.68                   2.69                                          0, 15\n\n###### \n\nRelationship between RELA gene expression and MMP-13 gene expression\n\n  Variables   RELA    MMP-13   r^2^    *p*\n  ----------- ------- -------- ------- ------\n  RELA        1       0,662    0,438   0,01\n  MMP-13      0,662   1                \n\nFrom [Table 3](#T3){ref-type=\"table\"} there is a positive relationship, where the lower the level of RELA gene expression, the lower the expression of MMP-13 genes. From the results of the Pearson correlation, RELA gene expression was found to be strongly related to the expression of the MMP-13 gene with r = 0.662 and p = 0.01.\n\nFrom [Figure 4](#F4){ref-type=\"fig\"} it can be seen that the lower the RELA gene expression, the lower the expression of the MMP-13 gene.\n\n![Expression of the RELA and MMP-13 Gene](OAMJMS-7-543-g004){#F4}\n\nDiscussion {#sec1-4}\n==========\n\nThe relationship between RELA gene expression and MMP-13 gene expression {#sec2-14}\n------------------------------------------------------------------------\n\nThe results showed a positive relationship between *RELA* gene expression (NF\u03ba\u03b2 p65) and *MMP-13* gene expression, where the lower the level of *RELA* gene expression, the lower the expression of the *MMP-13* gene. This positive relationship absolutely occurs in synoviocytes affected by osteoarthritis.\n\nThis is in line with several related research results that can explain the results of this study. Over *MMP-13* gene expression is very high in pathological conditions such as in disease rheumatoid arthritis, osteoarthritis and carcinoma. Clinical trials in patients with cartilage damage were found in patients who had high *MMP-13* expression \\[[@ref7]\\]. The results of this study can be used as a new therapeutic target for OA by inhibiting enzymes that play a role in the process of cartilage degradation \\[[@ref8]\\]. Another study also showed that overexpression of *MMP-13* in GMO mice spontaneously would cause cartilage damage; MMP-13 can prevent erosion in joint-prone \\[[@ref9]\\].\n\nAt the time of the inflammatory process, the *RELA* gene, which is one of the families of NF\u03ba\u03b2, is involved in the expression of several genes that play a role in the inflammatory response. The transcription process of NF\u03ba\u03b2 is stimulated by pro-inflammatory cytokines and chemokines. Activation of NF\u03ba\u03b2 will trigger the expression of genes that induce articular joint damage resulting in osteoarthritis. *RELA* is also needed to modulate the immune response. *RELA* is expressed in various cell types, including epithelial cells, endothelial cells and nerve tissue. In general, *RELA* plays a role in the adaptive immune system and destroys pathogens through activation of NF\u03ba\u03b2.\n\nRELA plays a key role in regulating immune responses to infections; incorrect RELA regulation has been linked to cancer, inflammation and autoimmune diseases. RELA is a transcription factor in mammals that controls some important genes in the process of immunity and inflammation. RELA is involved in the expression of several genes that play a role in the inflammatory response, cartilage degradation, cell proliferation, angiogenesis and are predominantly regulated by RELA. The RELA transcription process is stimulated by pro-inflammatory cytokines and chemokines. Activation of RELA will trigger the expression of genes that induce articular joint damage resulting in OA. Besides that RELA controlling the expression of many adaptive genes such as Major Histocompatibility Complex (MHC) and genes important for the regulation of apoptosis. Besides NF\u03ba\u03b2 controlling the expression of many adaptive genes such as Major Histocompatibility Complex (MHC) and genes important for the regulation of apoptosis \\[[@ref10]\\].\n\nRELA is needed in the transcription process of MMP-3 and MMP-13, activation of MMP-3 and MMP-13 requires the RELA gene. Interleukin-1 induces RELA and MMP-13 transcription processes \\[[@ref3]\\]. RELA is one of the NF\u03ba\u03b2 units which plays an important role in NF\u03ba\u03b2 activity \\[[@ref11]\\]. The MMP-13 target is not only in type II collagen, but also damages proteoglycans, type IV and IX collagen, osteonectin in the cartilage. During embryonic development, MMP-13 is expressed in the skeleton \\[[@ref12]\\].\n\nBased on the results of the research that has been done, it can be concluded that there is a strong correlation between *RELA* gene expression and MMP-13 gene expression in osteoarthritis synoviocytes after Mesenchymal Stem Cell Wharton Jelly.\n\nThank you to the Andalas Cancer Research Center and Stem Cell (ACRC) Andalas University and Laboratory and Indonesian Medical Education and Research Institute (IMERI) Faculty of Medicine, University of Indonesia.\n\n**Funding:** This research did not receive any financial support\n\n**Competing Interests:** The authors have declared that no competing interests exist\n"}
+{"text": "1.. Introduction\n================\n\nAs the highest accuracy attitude measurement device, star trackers are capable of providing arcsec level 3-axis attitude and are widely used in many spacecrafts. A star tracker can be treated as a special electronic camera connected to a microcomputer. It can take star images of a part of the sky and identify these stars in the star image. Based on the position information of these identified stars, the attitude of spacecraft can be determined \\[[@b1-sensors-14-04914]--[@b4-sensors-14-04914]\\].\n\nWith the expansion of the application fields, especially on spacecrafts with the capability of rapid and large angle attitude maneuver control, star trackers must still work normally and steadily. Under these highly dynamic conditions, a crucial problem arises: due to the large angular velocity of the spacecraft, the star-spots in the star image will move across many pixels during the exposure time and ultimately form obvious trails. This will affect star detection sensitivity and star location accuracy seriously and result in low attitude accuracy and poor stability. This case can be ameliorated by adjusting the dynamic working parameters, especially the exposure time. Increasing the exposure time means more energy is gathered at each star-spot, which enhances high star detection sensitivity, but on the other hand, increasing the exposure time aggravates the movement of the star-spots and makes it more difficult to locate them. By contrast, reducing the exposure time alleviates the movement effect and reduces star location errors, but at the cost of an energy loss. Therefore, under highly dynamic conditions there exists an optimal exposure time and it is necessary to choose this proper exposure time for star trackers.\n\nAs discussed above, under highly dynamic conditions the exposure time mainly affects two aspects of star trackers: star detection sensitivity and star location accuracy. The exposure time directly determines the total energy and length of each star-spot in the star image, and both of them together affect star detection sensitivity. Star detection sensitivity determines the number of the star-spots, which is an important factor affecting the attitude accuracy and stability of a star tracker. In the past years, some research about star detection sensitivity has been reported. In \\[[@b5-sensors-14-04914]\\] a rough estimation method for star detection sensitivity utilizing the SNR model for static conditions was first reported. Reference \\[[@b6-sensors-14-04914]\\] gives a general expression of star detection sensitivity based on the theory described in \\[[@b5-sensors-14-04914]\\], but under highly dynamic conditions the star-spots model can\\'t use the two-dimensional Gaussian distribution like in static conditions, so that the star detection sensitivity model developed for static conditions is not suitable for dynamic conditions. Reference \\[[@b7-sensors-14-04914]\\] gives a dynamic star-spot imaging model, and obtains the regularity of star detection sensitivity at different angular velocities for a star tracker.\n\nThe movement of the star-spots during the exposure time also increases the difficulty of locating the star-spots and lowers the star location accuracy. The star location accuracy is the primary factor determining the attitude accuracy of a star tracker. In the past, many researchers have concentrated on the exploration of star location errors. Reference \\[[@b8-sensors-14-04914]\\] obtains the star location error of the ideal star-spots in a 5 \u00d7 5 centroiding window by calculation of the effects of various noise components. Reference \\[[@b9-sensors-14-04914]\\] shows an explicit expression of the S-curve systematic error caused by the different positions of the star-spot center in a certain pixel. Reference \\[[@b10-sensors-14-04914]\\] gives a typical model of star location error containing a systematic contribution and a random one. However, all these researches mainly focus on static conditions, and we need to make research in depth the star location errors under highly dynamic conditions.\n\nThis paper presents a method for optimizing the exposure time from the two aspects: star detection sensitivity and star location accuracy, and obtains the optimal exposure times for different angular velocities of a star tracker. This paper is divided into six sections. Following the Introduction, we first introduce the dynamic star-spot imaging model and star detection sensitivity with regard to the exposure time in Section 2. In Section 3, the star location error is deduced based on the error analysis of the sub-pixel centroiding algorithm, and the effect of the exposure time on the star location error is obtained. Combining the analyses in Sections 2 and 3, the overall effect of exposure time on attitude accuracy is obtained in Section 4, and the optimal exposure time is determined with the highest attitude accuracy as the criterion. Night sky experiments with a real star tracker are carried out in Section 5. Conclusions are drawn in the last section.\n\n2.. Star Detection Sensitivity\n==============================\n\n2.1.. Dynamic Star-Spot Imaging Model\n-------------------------------------\n\nUnder static conditions, the angular velocity of the carrier is very low and stars can be assumed to be point sources. The star is projected on the focal plane and this generates a defocused star-spot as shown in [Figure 1](#f1-sensors-14-04914){ref-type=\"fig\"}. The point spread function of the optical system can be described as a two-dimensional Gaussian function, which is the energy distribution shape of the imaged star-spot.\n\nThe static star-spot imaging model is given by \\[[@b11-sensors-14-04914]\\]: $$S(x,y) = \\frac{\\mu_{0}}{{2\\pi\\sigma_{\\textit{pixel}}}^{2}}\\exp\\{ - \\frac{{({x - x_{c}})}^{2} + {({y - y_{c}})}^{2}}{{2\\sigma_{\\textit{pixel}}}^{2}}\\}$$where *\u03bc*~0~ is the total energy that star tracker absorbs from a single star during the exposure time, usually expressed using the signal photoelectrons, (*x~c~*, *y~c~*) is the center of the star-spot, and *\u03c3~pixel~* is the Gaussian radius, determined by the defo**c**used extent of the optical system. The total signal photoelectrons *\u03bc*~0~ of the star-spot is calculated using the photoelectrons transmit model \\[[@b7-sensors-14-04914]\\]: $$\\mu_{0} = \\tau_{A} \\cdot \\tau_{0} \\cdot E_{0} \\cdot 2.512^{- Mv} \\cdot \\frac{\\pi D^{2}}{4} \\cdot T_{e} \\cdot QE \\cdot \\frac{1}{E_{ph}} \\cdot K_{fill}$$where *\u03c4~A~* is the atmospheric transmissivity, *\u03c4*~0~ is the optical transmittance, *E*~0~ = 2.96 \u00d7 10^\u221214^ W/mm^2^, refers to the measured flux (on the Earth in the absence of the atmosphere) of the star with magnitude 0 \\[[@b12-sensors-14-04914]\\], *Mv* is the magnitude, *D* is the optical aperture, *T~e~* is the exposure time, *QE* is the quantum efficiency of the image sensor, *E~ph~* is the average energy of a single photon, and *K~fill~* is the fill factor of the image sensor.\n\nUnder highly dynamic conditions, the traditional two-dimensional Gaussian distribution is not suitable because the star-spots obviously move as shown in [Figure 2](#f2-sensors-14-04914){ref-type=\"fig\"}. A tiny moment during the exposure time is taken out and expressed using the *\u0394T*(*t*). In this moment the star-spot can be expressed approximately using the traditional two-dimensional Gaussian distribution \\[[@b7-sensors-14-04914]\\]. Supposing that the star-spot center coordinates at any time *t* in the exposure period is (*x~c~*(*t*), *y~c~*(*t*)) and the total energy during the *\u0394T*(*t*) is *\u03bc*~0~(*t*), the star-spot model during this moment can be expressed as: $$S(x,y,t) = \\frac{\\mu_{0}(t)}{{2\\pi\\sigma_{\\textit{pixel}}}^{2}}\\exp\\{ - \\frac{{({x - x_{c}(t)})}^{2} + {({y - y_{c}(t)})}^{2}}{{2\\sigma_{\\textit{pixel}}}^{2}}\\}$$\n\nConsidering the angular velocity of the carrier is basically steady and the exposure time is not long, the star-spot trail can be regarded approximately as a beeline. The length of the star-spot trail can be calculated approximately using *l* = *fwT~e~*/*DX*, where *f* is the focal length, *w* is the angular velocity, *T~e~* is the exposure time and *DX* is the size of each pixel. Defining *\u03b8* as the angle between the star-spot trail and the *x* axis and (*x*~0~, *y*~0~) as the star-spot center coordinates at the time *t* = 0, (*x~c~*(*t*), *y~c~*(*t*)) is approximately: $$\\left\\{ {\\begin{matrix}\n{x_{c}(t) \\approx x_{0} + fwt\\cos(\\theta)/DX} \\\\\n{y_{c}(t) \\approx y_{0} + fwt\\sin(\\theta)/DX} \\\\\n\\end{matrix}\\ (0 \\leq t \\leq T_{e})} \\right.$$\n\nThe total signal photoelectrons in (*x*, *y*) can be obtained by adding up the signal photoelectrons of all moments during the exposure time, so the dynamic star-spot imaging model can be obtained from the integral of *S*(*x*, *y*, *t*): $$S(x,y) = \\frac{\\mu_{0}}{{2\\pi\\sigma_{\\textit{pixel}}}^{2}T_{e}}{\\int_{0}^{T_{e}}{\\exp\\{ - \\frac{{({x - x_{c}(t)})}^{2} + {({y - y_{c}(t)})}^{2}}{{2\\sigma_{\\textit{pixel}}}^{2}}\\}}}dt$$\n\nOn the real image plane, the total signal photoelectrons of the star-spot are divided into many pixels. The signal photoelectrons of the pixel (*m*, *n*) can be obtained through the integral of the *S*(*x*, *y*) in the pixel position: $$S(m,n) = \\frac{\\mu_{0}}{{2\\pi\\sigma_{\\textit{pixel}}}^{2}T_{e}}{\\int_{m - 1/2}^{m + 1/2}{\\int_{n - 1/2}^{n + 1/2}{\\int_{0}^{T_{e}}{\\exp\\{ - \\frac{{({x - x_{c}(t)})}^{2} + {({y - y_{c}(t)})}^{2}}{{2\\sigma_{\\textit{pixel}}}^{2}}\\}}}}}\\textit{dtdydx}$$\n\n2.2.. SNR Model\n---------------\n\nStar detection sensitivity *Mv* is a magnitude threshold over which the star can be detected by a star tracker. Under highly dynamic conditions, it can be calculated using the SNR model. The SNR of a pixel is the ratio of its signal photoelectrons to its noise electrons.\n\nFor analysis convenience, we classify the pixels in the star-spot trail into two types. As shown in [Figure 3](#f3-sensors-14-04914){ref-type=\"fig\"}, the pixels in the middle of the trail are classified as the first type, which includes the most photoelectrons. Those pixels next to the first type belong to the second type, whose photoelectrons are just secondary to the first type of pixels. If these two types of pixels can form a connected domain of a certain size, the star-spot can be correctly extracted from the star image. A criterion can be set that the star can be successfully detected if the second type of pixels in the connected domain meet the SNR requirement \\[[@b13-sensors-14-04914]\\].\n\nWe define *S* as the average signal photoelectrons of the second type of pixels, and *\u03b7* as the ratio of the *S* to the total energy *\u03bc*~0~. Taking the CMOS image sensor as an example, the noise electrons *N* of the pixel \\[[@b14-sensors-14-04914]--[@b18-sensors-14-04914]\\] is given as: $$N = \\sqrt{n_{DC}^{2} + n_{S}^{2} + n_{\\textit{PRNU}}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + n_{\\textit{ADC}}^{2}}$$where *n~DC~*, *n~S~*, *n~PRNU~*, *n~TN~*, *n~FPN~*, *n~ADC~* are dark current noise, photon shot noise, photon non-uniformity response noise, temporal noise, fixed pattern noise and analog-to-digital conversion noise, respectively, which are listed in [Table 1](#t1-sensors-14-04914){ref-type=\"table\"}. Therefore the SNR of the second type of pixels and the star detection sensitivity *Mv* can be expressed as: $$\\textit{SNR} = \\frac{S}{N} = \\frac{\\eta\\mu_{0}}{\\sqrt{n_{DC}^{2} + n_{S}^{2} + n_{\\textit{PRNU}}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + n_{\\textit{ADC}}^{2}}} \\geq V_{sn}$$ $$Mv \\leq - \\log_{2.512}(\\frac{V_{sn}^{2} + \\sqrt{V_{sn}^{4} + 4(1 - V_{sn}^{2}\\sigma_{\\textit{PRNU}}^{2})V_{sn}^{2}(n_{DC}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + n_{\\textit{ADC}}^{2})}}{2\\eta(1 - V_{sn}^{2}\\sigma_{\\textit{PRNU}}^{2})(\\tau_{A} \\cdot \\tau_{0} \\cdot E_{0} \\cdot \\pi D^{2} \\cdot T_{e} \\cdot QE)/(4 \\cdot E_{ph})})$$where *S* is the average signal photoelectrons of the second type of pixels, *N* is the average noise electrons of the same pixels, *V~sn~* is the SNR threshold.\n\nDenoting *A* as the Field Of View (FOV) of the star tracker, for the square shaped CMOS image sensor, the average number of the star-spots *N~star~* in the star image is given by \\[[@b12-sensors-14-04914]\\]: $$N_{\\textit{star}} = 6.57 \\cdot e^{1.08Mv} \\cdot \\frac{1 - \\cos(A/\\sqrt{\\pi})}{2}$$which indicates the star detection sensitivity *Mv* is the only factor that affects the number of the star-spots in a given FOV.\n\nSimulations are done to obtain the *\u03b7* values with different trail lengths (*l* = 0\u223c19 pixels). The *\u03b7* is mainly affected by two factors: the length of the star-spot trail and the angle between the star-spot trail and the *x* axis. The angle depends on the direction of the angular velocity. The results with a 45\u00b0 angle are shown in [Table 2](#t2-sensors-14-04914){ref-type=\"table\"}.\n\nThe product of the *\u03b7* and the trail length *l* approximates to a fixed value 0.264 when the length *l* \\> 9 pixels, which is a valid generalization that can be used to calculate the *\u03b7* when *l* \\> 19 pixels. The *Mv* and *N~star~* under different exposure times (*T~e~* = 0\u223c200 ms) and different angular velocities (*w* = 1\u00b0/s, 2\u00b0/s and 4\u00b0/s) are also obtained in [Figure 4](#f4-sensors-14-04914){ref-type=\"fig\"}, with other parameters listed in [Table 3](#t3-sensors-14-04914){ref-type=\"table\"}.\n\nAs shown in [Figure 4a](#f4-sensors-14-04914){ref-type=\"fig\"}, the general regularity is that the star detection sensitivity rises rapidly at first and then slowly approaches a constant with the increasing exposure time. When the exposure time is very short, the star-spot trail is very short, and increasing the exposure time means the gathering of the energy in each pixel, so the SNR and star detection sensitivity rise quickly, but with the continually increase of exposure time, the star-spot moves longer and the total energy spreads out over more pixels. The energy of each pixel begins to increase slowly and finally remains invariable, so that the star detection sensitivity slowly approaches a constant.\n\nThe above analysis indicates that star detection sensitivity has a limit value for each angular velocity. In [Figure 4a](#f4-sensors-14-04914){ref-type=\"fig\"}, star detection sensitivity of the star tracker with a given configuration at the angular velocities of 1\u00b0/s, 2\u00b0/s and 4\u00b0/s are 6.38, 5.65 and 4.91, respectively.\n\nIn addition, it must be noted that the critical exposure times when the star detection sensitivity approaches the limit value are different for different angular velocities. The higher the angular velocity is, the shorter the critical exposure time is. Under highly dynamic conditions, the increase of exposure time has two effects: the increase of the energy at a single pixel and the dispersion of the energy over several pixels. The critical exposure time, denoted as *T~SDS~*, can be regarded as the balance point of these two effects. In case that the exposure time is over the balance point, increasing the exposure time is useless and cannot produce higher star detection sensitivity.\n\n3.. Star Location Accuracy\n==========================\n\n3.1.. Sub-Pixel Centroiding Algorithm\n-------------------------------------\n\nThe sub-pixel centroiding algorithm is utilized to improve star location accuracy. Generally, the optical system of star tracker is slightly defocused, and the star-spots in the star image will occupy several pixels. This will help to calculate the star-spot centroid with sub-pixel accuracy. The sub-pixel centroiding algorithm is given by \\[[@b19-sensors-14-04914]\\]: $$\\begin{matrix}\n{\\overline{x} = \\frac{\\iint\\limits_{C}{xS(x,y)\\textit{dxdy}}}{\\iint\\limits_{C}{S(x,y)\\textit{dxdy}}} = \\frac{\\int_{0}^{T_{e}}{x_{c}(t)dt}}{T_{e}}} \\\\\n{\\overline{y} = \\frac{\\iint\\limits_{C}{yS(x,y)\\textit{dxdy}}}{\\iint\\limits_{C}{S(x,y)\\textit{dxdy}}} = \\frac{\\int_{0}^{T_{e}}{y_{c}(t)dt}}{T_{e}}} \\\\\n\\end{matrix}$$where *C* is a limited area, called centroiding window, *S*(*x*, *y*) is the dynamic star-spot imaging function in [Equation (5)](#FD5){ref-type=\"disp-formula\"}, and (*x~c~*(*t*), *y~c~*(*t*)) is the star-spot center at the time *t* (0 \u2a7d *t* \u2a7d *T~e~*).\n\nFor the dynamic star-spot, the prescriptive center of the whole star-spot trail is the star-spot center at the time *t* = *T~e~*/2: (*x~c~*(*T~e~*/2), *y~c~*(*T~e~*/2)). According to [Equation (4)](#FD4){ref-type=\"disp-formula\"}, the *x~c~*(*t*) and *y~c~*(*t*) are approximately linear functions. Consequently, *x\u0304* \u2248 *x~c~*(*T~e~*/2), *\u0233* \u2248 *y~c~*(*T~e~*/2), and it is feasible to use the sub-pixel centroiding algorithm to determine the star-spot location, even under highly dynamic conditions. In fact, it must be noted that the available image is discrete, and the imaging function *S*(*x*, *y*) is sampled in each pixel. The real formula used to calculate the star-spot centroid is: $$\\overline{x} = \\frac{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{x_{k}I_{k}}}{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}I_{k}},\\overline{y} = \\frac{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{y_{k}I_{k}}}{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}I_{k}}$$where *n* is the number of pixels in the centroiding window, (*x~k~*, *y~k~*) is the geometric center of the *k*th pixel, and *I~k~* is the sampling value of the *k*th pixel. The sampling value *I~k~* is directly proportional to the amount of signal photoelectrons *S~k~*, which can be calculated using [Equation (6)](#FD6){ref-type=\"disp-formula\"}: $$I_{k} = \\lambda S_{k},\\lambda = \\frac{2^{q}}{N_{\\textit{well}}}$$where \u03bb is the ratio, *N~well~* is the full well charge of the image sensor, and *q* is the quantization bits. Therefore, according to [Equation (7)](#FD7){ref-type=\"disp-formula\"}, *I~k~* variance can be expressed as: $$\\sigma_{I_{k}}^{2} = \\lambda^{2}N^{2} = \\lambda^{2}(n_{DC}^{2} + n_{S,k}^{2} + n_{\\textit{PRNU},k}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + n_{\\textit{ADC}}^{2})$$\n\n3.2.. Star Location Error\n-------------------------\n\nUnder highly dynamic conditions, star location error of the star tracker using the sub-pixel centroiding algorithm includes the following error sources: (1)The error from the nonlinearity of the *x~c~*(*t*) and *y~c~*(*t*), which is mainly caused by the variation of the angular velocity \\[[@b20-sensors-14-04914]\\]. Considering the exposure time is usually very short, the angular velocity during that short time can be approximated as a constant in most cases, so this error is not taken into account in this paper.(2)The systematic error caused by using the geometric center of the pixel instead of the energy center in the sub-pixel centroiding algorithm. Under static conditions, it changes like the sine curve with the different position of the star-spot center in a pixel, and is usually called as the S-curve error \\[[@b21-sensors-14-04914]\\].(3)The random error caused by the random noises of the image sensor, which is determined by the range of the random noises and the length of star-spot trail. Generally, more noises and especially the noises at the pixel farther away from the star-spot centroid will aggravate this kind of error.\n\nFor convenience, the following analysis takes the *x\u0304* as an example. According to [Equation (12)](#FD12){ref-type=\"disp-formula\"}, the star location error of the *x\u0304* derives from two sources: *x~k~* and *I~k~*. Assuming that the errors of different pixels are not correlated, the systematic error and the random error of the dynamic star-spot can be derived as: $$\\sigma_{x,2}^{2} = \\left\\lbrack {\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{\\frac{\\partial\\overline{x}}{\\partial x_{k}}\\sigma_{x_{k}}}} \\right\\rbrack^{2} = \\left\\lbrack {\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{\\frac{I_{k}}{I_{0}}\\sigma_{x_{k}}}} \\right\\rbrack^{2}$$ $$\\sigma_{x,3}^{2} = \\underset{k = 1}{\\overset{n}{\\text{\u2211}}}\\left\\lbrack {{(\\frac{\\partial\\overline{x}}{\\partial I_{k}})}^{2}\\sigma_{I_{k}}^{2}} \\right\\rbrack = \\underset{k = 1}{\\overset{n}{\\text{\u2211}}}\\left\\lbrack {{(\\frac{x_{k} - \\overline{x}}{I_{0}})}^{2}\\sigma_{I_{k}}^{2}} \\right\\rbrack$$where *\u03c3~x~*~,2~ is the systematic error resulting from the use of the geometrical center instead of the energy center, *\u03c3~x~*~,3~ is the random error caused by the image sensor noises, and *I*~0~ is the sum of all sampling values in centroiding window.\n\nThe energy center of the *k*th pixel can be calculated as: $$g_{x_{k}} = \\frac{\\iint\\limits_{C_{k}}{xS(x,y)\\textit{dxdy}}}{\\iint\\limits_{C_{k}}{S(x,y)\\textit{dxdy}}}$$where *C~k~* is the region of the *k*th pixel, *S*(*x*, *y*) is the dynamic star-spot imaging function, so the systematic error is written as: $$\\begin{matrix}\n{\\sigma_{x,2}^{2} = \\left\\lbrack {\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{(\\frac{I_{k}}{I_{0}})(x_{k} - \\frac{\\iint\\limits_{C_{k}}{xS(x,y)\\textit{dxdy}}}{\\iint\\limits_{C_{k}}{S(x,y)\\textit{dxdy}}})}} \\right\\rbrack^{2}} \\\\\n{\\sigma_{y,2}^{2} = \\left\\lbrack {\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{(\\frac{I_{k}}{I_{0}})(y_{k} - \\frac{\\iint\\limits_{C_{k}}{yS(x,y)\\textit{dxdy}}}{\\iint\\limits_{C_{k}}{S(x,y)\\textit{dxdy}}})}} \\right\\rbrack^{2}} \\\\\n\\end{matrix}$$\n\nSubstituting [Equation (14)](#FD14){ref-type=\"disp-formula\"} into [Equation (16)](#FD16){ref-type=\"disp-formula\"} yield, the random error of the *x\u0304* can be expressed as: $$\\begin{array}{ll}\n\\sigma_{x,3}^{2} & {= \\underset{k = 1}{\\overset{\\text{n}}{\\text{\u2211}}}\\left\\lbrack {{(\\frac{x_{k} - \\overline{x}}{I_{0}})}^{2}(\\lambda^{2}(n_{DC}^{2} + n_{S,k}^{2} + n_{\\textit{PRNU},k}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + {n^{2}}_{\\textit{ADC}}))} \\right\\rbrack} \\\\\n & {= \\underset{k = 1}{\\overset{\\text{n}}{\\text{\u2211}}}\\left\\lbrack {{(\\frac{x_{k} - \\overline{x}}{I_{0}})}^{2}(\\lambda^{2}(n_{DC}^{2} + n_{RN}^{2} + n_{\\textit{FPN}}^{2} + {n^{2}}_{\\textit{ADC}}) + \\lambda I_{k} + \\sigma_{\\textit{PRNU}}^{2}I_{k}^{2})} \\right\\rbrack} \\\\\n & {= \\frac{\\lambda^{2}(n_{DC}^{2} + n_{RN}^{2} + n_{\\textit{FPN}}^{2} + {n^{2}}_{\\textit{ADC}})}{I_{0}^{2}}\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{({x_{k} - \\overline{x}})}^{2}} \\\\\n & {+ \\frac{1}{I_{0}^{2}}\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{\\lbrack{({x_{k} - \\overline{x}})}^{2}}(\\lambda I_{k} + \\sigma_{\\textit{PRNU}}^{2}I_{k}^{2})\\rbrack} \\\\\n\\end{array}$$\n\nThe random error *\u03c3~x~*~,3~ must be determined computing all the pixels with their noises in the centroiding window. If the centroiding window is set as the traditional rectangular window, the *\u03c3~x~*~,3~ will obviously increase because many unnecessary pixels with fixed noises are computed. Generally, the centroiding window for the dynamic star-spot is set as the Window A in [Figure 5](#f5-sensors-14-04914){ref-type=\"fig\"}. The length of Window A is the same as the star-spot trail, and its width is set to 6*\u03c3~pixel~* pixels because the section plane of the star-spot trail fits Gaussian distribution. Dividing the Window A into many sub-windows like Window B whose width is just a pixel, some approximate expressions for calculating the *\u03c3~x~*~,3~ are obtained as follows: $$\\begin{array}{ll}\n{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{({x_{k} - \\overline{x}})}^{2}} & {\\approx 2\\frac{l_{1}}{\\cos(\\theta)}\\left\\lbrack {{(\\frac{m}{2})}^{2} + {({\\frac{m}{2} - 1})}^{2} + {({\\frac{m}{2} - 2})}^{2}\\cdots\\cdots + 2^{2} + 1^{2}} \\right\\rbrack} \\\\\n & {\\approx 2\\frac{6\\sigma_{\\textit{pixel}}}{\\cos(\\theta)}\\frac{{({m + 1})}^{3}}{12} = \\frac{\\sigma_{\\textit{pixel}}{({l\\cos(\\theta) + 1})}^{3}}{\\cos(\\theta)}} \\\\\n\\end{array}$$where *l* is the length of the star-spot trail, *\u03b8* is the angle between the star-spot trail and *x* axis, *l*~1~ is the width of centroiding window, *m* is the length of the star-spot trail in the *x* direction. As shown in [Figure 5](#f5-sensors-14-04914){ref-type=\"fig\"}, the number of the pixels in the sub-window is *l~1~*/cos(*\u03b8*),and the *m* is *l*cos(*\u03b8*) pixels.\n\nObviously, the *m* is also the number of the sub-windows, and the total sampling value *I~B~* of the sub-window is *I*~0~/*m*: $$\\begin{array}{ll}\n{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{({x_{k} - \\overline{x}})}^{2}I_{k}} & {\\approx 2\\frac{I_{0}}{m}\\left\\lbrack {{(\\frac{m}{2})}^{2} + {({\\frac{m}{2} - 1})}^{2} + {({\\frac{m}{2} - 2})}^{2}\\cdots\\cdots + 2^{2} + 1^{2}} \\right\\rbrack} \\\\\n & {\\approx 2\\frac{I_{0}}{m}\\frac{{({m + 1})}^{3}}{12} = \\frac{I_{0}{({l\\cos(\\theta) + 1})}^{3}}{6l\\cos(\\theta)}} \\\\\n\\end{array}$$\n\nAccording to the dynamic star-spot model, the signal photoelectrons distribution of the sub-window fit approximately the Gaussian distribution with a Gaussian radius *\u03c3~pixel~*/cos (*\u03b8*). The sum of *I~k~*^2^ in the sub-window is approximately *I~B~*^2^/(3.6*\u03c3~pixel~* /cos (*\u03b8*)) statistically: $$\\begin{array}{ll}\n{\\underset{k = 1}{\\overset{n}{\\text{\u2211}}}{({x_{k} - \\overline{x}})}^{2}I_{k}^{2}} & {\\approx 2\\frac{I_{B}^{2}\\cos(\\theta)}{3.6\\sigma_{\\textit{pixel}}}\\left\\lbrack {{(\\frac{m}{2})}^{2} + {({\\frac{m}{2} - 1})}^{2} + {({\\frac{m}{2} - 2})}^{2}\\cdots\\cdots + 2^{2} + 1^{2}} \\right\\rbrack} \\\\\n & {\\approx 2\\frac{I_{B}^{2}\\cos(\\theta)}{3.6\\sigma_{\\textit{pixel}}}\\frac{{({m + 1})}^{3}}{12} = \\frac{I_{0}^{2}{({l\\cos(\\theta) + 1})}^{3}}{21.6\\sigma_{\\textit{pixel}}l^{2}\\cos(\\theta)}} \\\\\n\\end{array}$$\n\nSubstituting [Equations (20)](#FD20){ref-type=\"disp-formula\"}, [(21)](#FD21){ref-type=\"disp-formula\"} and [(22)](#FD22){ref-type=\"disp-formula\"} into [Equation (19)](#FD19){ref-type=\"disp-formula\"}, we get the explicit expression of the random error for the high dynamic star-spot: $$\\begin{array}{ll}\n\\sigma_{x,3}^{2} & {\\approx \\frac{\\lambda^{2}(n_{DC}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + {n^{2}}_{\\textit{ADC}})\\sigma_{\\textit{pixel}}{({l\\cos(\\theta) + 1})}^{3}}{\\cos(\\theta)I_{0}^{2}}} \\\\\n & {+ \\frac{\\lambda{({l\\cos(\\theta) + 1})}^{3}}{6I_{0}l\\cos(\\theta)} + \\frac{\\sigma_{\\textit{PRNU}}^{2}{({l\\cos(\\theta) + 1})}^{3}}{21.6\\sigma_{\\textit{pixel}}l^{2}\\cos(\\theta)}} \\\\\n\\sigma_{y,3}^{2} & {\\approx \\frac{\\lambda^{2}(n_{DC}^{2} + n_{TN}^{2} + n_{\\textit{FPN}}^{2} + {n^{2}}_{\\textit{ADC}})\\sigma_{\\textit{pixel}}{({l\\sin(\\theta) + 1})}^{3}}{\\sin(\\theta)I_{0}^{2}}} \\\\\n & {+ \\frac{\\lambda{({l\\sin(\\theta) + 1})}^{3}}{6I_{0}l\\sin(\\theta)} + \\frac{\\sigma_{\\textit{PRNU}}^{2}{({l\\sin(\\theta) + 1})}^{3}}{21.6\\sigma_{\\textit{pixel}}l^{2}\\sin(\\theta)}} \\\\\n\\end{array}$$\n\nThe *l, I*~0~ and *\u03b8* are three main factors affecting the random error, the *l* and *I*~0~ are directly related to the exposure time, and the *\u03b8* is decided by the direction of angular velocity. The total systematic error and the total random error can be written as: $$\\begin{matrix}\n\\sigma_{2} & {= \\sqrt{\\sigma_{x,2}^{2} + \\sigma_{y,2}^{2}}} \\\\\n\\sigma_{3} & {= \\sqrt{\\sigma_{x,3}^{2} + \\sigma_{y,3}^{2}}} \\\\\n\\end{matrix}$$\n\nThe simulation is carried out to obtain the two errors with different exposure times (*T~e~* = 0\u223c200 ms) and different angular velocities (*w* = 1\u00b0/s, 2\u00b0/s and 4\u00b0/s), with other parameters listed in [Table 3](#t3-sensors-14-04914){ref-type=\"table\"}. Effects of exposure time on the *\u03c3*~2~, *\u03c3*~3~ when the angle *\u03b8* is 45\u00b0are shown in [Figure 6](#f6-sensors-14-04914){ref-type=\"fig\"}.\n\nUnder highly dynamic conditions, the systematic error also changes like the S-curve error under static conditions. The amplitude of the systematic error, as shown in [Figure 6a](#f6-sensors-14-04914){ref-type=\"fig\"}, decreases like an inverse proportion function with the increase of exposure time. This is because the *I*~k~/*I*~0~ is decreasing and the energy center of each pixel becomes nearer to the geometric center with the lengthening of the star-spot trail. The amplitude of the systematic error in the static condition is exp\\[\u22122(\u03c0\u03c3)^2^\\]/\u03c0, which mainly depends on the Gaussian radius \\[[@b9-sensors-14-04914]\\]. For the 0.7 pixel Gaussian radius, it is just about 2 \u00d7 10^\u22125^ pixels which is in agreement with the above simulation. Compared with the random error in [Figure 6b](#f6-sensors-14-04914){ref-type=\"fig\"}, the systematic error is very small and negligible, so that the star location error under highly dynamic conditions mainly depends on the random error caused by the image sensor noises.\n\nFor the random error, it decreases slightly at first when the exposure time is very short, and then keeps rising at an almost fixed speed, which means an inflexion exists. According to [Equation (23)](#FD23){ref-type=\"disp-formula\"}, at a given angular velocity, the random error is affected mainly by two factors: the total energy *I~0~* and the length *l* of the star-spot trail. When the exposure time is very short, the star-spot trail is very short and the total energy of each star-spot is very little. The increase of the exposure time means the gathering of the energy and the rising of the SNR, which results in that the random error decreases slightly at first. However, as the *I~0~* and *l* increase to some extent, the cube of the (*l*cos(*\u03b8*) + 1) increases faster than the square of the *I*~0~, and the derivative of the random error to the exposure time is also almost a positive constant, so the random error begins to rise at a fixed speed. It is obvious that higher the angular velocity is, faster the length *l* increases, higher the speed is. We can denote the exposure time at the inflection point as *T~SLE~*, where the star tracker can achieve the highest star location accuracy.\n\n4.. Exposure Time Optimization\n==============================\n\nAs an attitude measurement device, the main purpose of the star tracker is to provide highly-precision attitude information. Attitude accuracy is the most primary factor for performance evaluation. It can be estimated using the following equation \\[[@b12-sensors-14-04914]\\]: $$E_{\\textit{cross} - \\textit{angle}} = \\frac{A \\cdot E_{\\textit{centroid}}}{N_{\\textit{pixel}} \\cdot \\sqrt{N_{\\textit{star}}}}$$where *E~cross-angle~* is the estimation of the attitude error, *A* is the FOV of star tracker, *E~centroid~* is the average star location error, *N~pixel~* is the pixel array of the image sensor, and *N~star~* is the number of the star-spots in the star image. For a given star tracker, *A* and *N~pixel~* can be regarded as constants, so the attitude accuracy is mainly determined by *E~centroid~* and *N~star~* which are directly related to the exposure time. Hence, we can optimize the exposure time using the highest attitude accuracy as the criterion.\n\nCombining star detection sensitivity and star location error, effect of exposure time on the attitude error is shown in [Figure 7](#f7-sensors-14-04914){ref-type=\"fig\"}. Variation regularity of the attitude error is basically consistent with the star location error. It decreases at first and then increases. The curve also has an inflection point which is just behind the inflection point (*T~SLE~*) of the star location error. According to [Equation (24)](#FD24){ref-type=\"disp-formula\"}, the attitude error is directly proportional to *E~centroid~* and inversely proportional to the square root of *N~star~*. According to the previous analysis, when the exposure time is short, *N~star~* will increase and *E~centroid~* will reduce, so the attitude error decreases slightly at first. Then *E~centroid~* starts to rise when the exposure time pass *T~SLE~*, but the square root of *N~star~* increases more quickly than *E~centroid~*, which leads to the still decreasing of the attitude error. But with the continually increasing of exposure time, *E~centroid~* will get ahead of the square root of *N~star~* and the attitude error will increase. Finally, *N~star~* remains a constant and the increase speed of the attitude error is the same as the star location error.\n\nThe exposure time for the inflection point of the attitude error, denoted as the optimal exposure time, is between *T~SLE~* and *T~SDS~*, which reflects exactly the overall consideration on star detection sensitivity and star location error. It is noticed that when the angular velocity is relatively low, for example the 1\u00b0/s in the figure, the attitude error has a flat variation which means that we can choose a proper exposure time from a wide range, but for high angular velocities, we must set the exposure time carefully to obtain the best performance. The optimal exposure times at different angular velocities are calculated as shown in [Figure 8](#f8-sensors-14-04914){ref-type=\"fig\"}. The result shows the optimal exposure time decreases like the inverse proportion function with the increase of angular velocity.\n\n5.. Night Sky Experiments\n=========================\n\nTo validate the result of the simulations, night sky experiments using a recently developed star tracker were carried out. The star tracker, shown in [Figure 9a](#f9-sensors-14-04914){ref-type=\"fig\"}, has a 14\u00b0 \u00d7 14\u00b0 rectangular FOV with a CMOS image sensor consisting of 2048 \u00d7 2048 pixels, with other parameters as listed in [Table 3](#t3-sensors-14-04914){ref-type=\"table\"}. The night sky experiment was performed at the Xinglong Astronomical Observatory of the Chinese Academy of Sciences. The star tracker was placed on a one dimension rotary table which was mounted on a tripod as shown in [Figure 9b](#f9-sensors-14-04914){ref-type=\"fig\"}. Angular velocities of the rotary table were set to 1\u00b0/s and 2\u00b0/s. The exposure times were set to 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 80, 100, 120, 150, 175 and 200 ms respectively. For each of the 32 conditions, star-spot data of more than 100 star images were collected with the tested star tracker. The average number of the star-spots and the average star location error in each of conditions are calculated as shown in [Figure 10](#f10-sensors-14-04914){ref-type=\"fig\"}.\n\nIn [Figure 10a](#f10-sensors-14-04914){ref-type=\"fig\"}, the general regularity of the average number of star-spots is that it increases fast at first and then levels off slowly with the increasing exposure time. It can be seen from the figure that for the angular velocity of 1\u00b0/s, the average star number is 6.46 at the exposure time of 10 ms, then reaches slowly to the limit value 23.01 at 100 ms. For the angular velocity of 2\u00b0/s, it is 6.78 at 10 ms, reaches the limit value 15.69 at 60 ms. To sum up, the numerical value and general regularity of the average number of star-spots in the experiment are in agreement with the results of the previous simulations.\n\nThe average star location error, as shown in [Figure 10b](#f10-sensors-14-04914){ref-type=\"fig\"}, keeps rising in the range of the exposure time 10 ms to 200 ms. For the angular velocity of 1\u00b0/s, the star location error is 0.24 pixels at 10 ms, and then increases steadily at a fixed speed and reaches 0.82 pixels at 200 ms. For the angular velocity of 2\u00b0/s, it is 0.44 pixel at 10 ms, reaches 1.81 pixel at 200 ms. The experimental results of star location error also agree basically with the simulations.\n\nFinally, the attitude error is calculated as shown in [Figure 11](#f11-sensors-14-04914){ref-type=\"fig\"}. The attitude error at angular velocities of 1\u00b0/s and 2\u00b0/s both have their inflection points, which are the corresponding optimal exposure times with the value of about 30 ms and 20 ms respectively. In previous simulations as shown in [Figure 8](#f8-sensors-14-04914){ref-type=\"fig\"}, the optimal exposure times are 30.7 ms and 17.6 ms for 1\u00b0/s and 2\u00b0/s respectively, which are very close to the experimental results.\n\n6.. Conclusions\n===============\n\nEffects of exposure time on star detection sensitivity and star location accuracy for a star tracker under highly dynamic conditions are analyzed in this paper. On this basis, the effects of exposure time on attitude accuracy and the corresponding optimal exposure times are also obtained. Some simulations and night sky experiments are carried out to validate these effects. The above content can be summarized as follows: (1)Under highly dynamic conditions, star-spots will move across many pixels and form an obvious trail on the image sensor during the exposure time. The length of the star-spot trail is decided by the angular velocity and the exposure time, and the higher the angular velocity is, the larger the exposure time is, so the longer the trail is. This phenomenon has a bad influence on star detection sensitivity and star location accuracy, which result in low star tracker attitude accuracy.(2)Firstly, the model of star detection sensitivity for the high dynamic condition is developed. According to the simulation, the star detection sensitivity rises rapidly at first with the increasing exposure time, and then slowly approaches a limit value. The limit values at different angular velocities are different, and the higher the angular velocity is, the smaller the limit value is. The critical exposure time *T~SDS~* for the limit value are also different, and the higher the angular velocity is, the shorter the *T~SDS~* is.(3)Then this paper establishes the model of star location error for the highly dynamic conditions, and identifies the main error sources through theoretical analysis and simulation. The simulation result shows that when the exposure time is very short, the star-spot trail is short and negligible. Increasing of exposure time means the gathering of the energy, which leads to the rising of SNR and the decreasing of star location error. When the exposure time increases to some extent, the star-spot trail gradually becomes obvious and the star location error starts to rise at a steady speed, and the higher the angular velocity is, the larger the rise speed is. Overall, the star location error has a minimum at the inflection point with the exposure time *T~SLE~*.(4)The estimation of attitude accuracy for highly dynamic conditions is obtained, which is used to determine the optimal exposure time of different angular velocities. The result shows the optimal exposure time is between the *T~SLE~* and the *T~SDS~*, and decreases like an inversely proportional function with the increase of angular velocity.(5)Night sky experiments using a CMOS star tracker have been carried out, and the experimental results validate those of the simulations.\n\nThis work is supported by the National Natural Science Foundation of China (Grant No. 61378052) and the Program for New Century Excellent Talents in University (NCET-10-0037). The authors are grateful for all of the valuable suggestions received during the course of this work.\n\nThe corresponding author XinguoWei, who is responsible for the overall work, proposed the research ideas and prepared the manuscript. The second author Wei Tan has performed the simulations and analysis, while the third author Jian Li has performed the night sky tests using a real star tracker and collected the experimental data. The co-corresponding author Guangjun Zhang, as the research group leader, has provided general guidance during the research and approved the paper.\n\nThe authors declare no conflict of interest.\n\n![Static star-spot imaging process.](sensors-14-04914f1){#f1-sensors-14-04914}\n\n![Dynamic star-spot imaging process.](sensors-14-04914f2){#f2-sensors-14-04914}\n\n![The first and second type of pixels in the star-spot trail.](sensors-14-04914f3){#f3-sensors-14-04914}\n\n![(**a**) Star detection sensitivity *Mv* with different exposure times; (**b**) The average number of star-spots *N~star~* with different exposure times.](sensors-14-04914f4){#f4-sensors-14-04914}\n\n![Detailed description of centroiding window.](sensors-14-04914f5){#f5-sensors-14-04914}\n\n![(**a**) The systematic error *\u03c3*~2~ with different exposure times; (**b**) The random error *\u03c3*~3~ with different exposure times, the right figure is the enlarged view in the range of 0 to 30 ms.](sensors-14-04914f6){#f6-sensors-14-04914}\n\n![The attitude error with different exposure times.](sensors-14-04914f7){#f7-sensors-14-04914}\n\n![The optimal exposure times at different angular velocities.](sensors-14-04914f8){#f8-sensors-14-04914}\n\n![(**a**) The star tracker used in the experiment; (**b**) Night sky experiment setup.](sensors-14-04914f9){#f9-sensors-14-04914}\n\n![(**a**) The average number of the star-spots; (**b**) The average star location error with different exposure times in the experiment.](sensors-14-04914f10){#f10-sensors-14-04914}\n\n![The attitude error with different exposure times in the experiment.](sensors-14-04914f11){#f11-sensors-14-04914}\n\n###### \n\nMain noises of CMOS image sensor.\n\n  **Noise**                              **Formula**\n  -------------------------------------- ---------------------------------------------------------------\n  Dark current noise                     $n_{DC} = \\sqrt{I_{\\textit{dark}}T_{e}}$\n  Photon shot noise                      $n_{S} = \\sqrt{S}$\n  Photon non-uniformity response noise   $n_{\\textit{PRNU}} = \\sigma_{\\textit{PRNU}}S$\n  Quantization noise                     $n_{\\textit{ADC}} = \\frac{N_{\\textit{well}}}{2^{q}\\sqrt{12}}$\n  Fixed pattern noise                    $n_{\\textit{FPN}} = N_{\\textit{FPN}}$\n  Temporal noise                         $n_{TN} = N_{TN}$\n\n###### \n\n*\u03b7* values with different lengths of the star-spot trail.\n\n  **Trail length *l*/pixel**   ***\u03b7* value**   **Trail length *l*/pixel**   ***\u03b7* value**\n  ---------------------------- --------------- ---------------------------- ---------------\n  0                            0.0831          10                           0.0264\n  1                            0.0804          11                           0.0240\n  2                            0.0740          12                           0.0220\n  3                            0.0651          13                           0.0203\n  4                            0.0547          14                           0.0188\n  5                            0.0450          15                           0.0176\n  6                            0.0401          16                           0.0165\n  7                            0.0347          17                           0.0155\n  8                            0.0321          18                           0.0147\n  9                            0.0295          19                           0.0138\n\n###### \n\nSimulation parameters of star tracker with a CMV4000 CMOS image sensor.\n\n  **Parameter**                       **Value**    **Parameter**           **Value**\n  ----------------------------------- ------------ ----------------------- ------------------\n  Focal length                        44.1637 mm   Pixel array             2048 \u00d7 2048\n  Optics aperture                     40 mm        Principal point         (1094.4, 1078.0)\n  Gaussian radius                     0.7 pixel    Pixel size              5.5 \u03bcm \u00d7 5.5 \u03bcm\n  Optical transmittance               0.85         Dark current            125 e/s\n  Atmospheric transmissivity          0.9          Photon non-uniformity   0.01\n  Quantum efficiency \\* Fill factor   0.6          Temporal noise          13 e\n  FOV                                 14\u00b0 \u00d7 14\u00b0    Fixed pattern noise     35 e\n  Quantization bits                   8            Full well charge        13,500 e\n  SNR threshold                       5                                    \n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nPhysic nut (*Jatropha curcas* L., family of Euphorbiaceae) is a common oil seed plant, its seeds contain about 30% of inedible oil, can be found in most tropical and subtropical countries of the world, and can grow up to 5\u2009m height \\[[@B1]\\]. *Jatropha curcas* is a native of Mexico and Central American region and was later introduced in many parts of the tropics and subtropics of the world \\[[@B1]\\]. It has low requirements to soil fertility and can grow under low rainfall conditions \\[[@B2]\\].\n\n*Jatropha curcas* is a multipurpose biodiesel plant adaptable to a wide range of edaphic and climatic conditions \\[[@B3]\\]. It is easily propagated by both generative (direct seeding) and vegetative (stem cuttings) techniques \\[[@B1], [@B3]--[@B5]\\], but plants propagated from stem cuttings do not develop a taproot. The plant only develops thin roots unable to grow deep in to the soil, which makes the plants more susceptible to uprooting by strong wind \\[[@B6]\\]. In agroforestry and intercropping systems, direct seeding is preferable over precultivated *J. curcas* plants, because the taproots of direct seeded plants deeply penetrate the soil layers \\[[@B7]\\], where it can assess extra nutrient resources and compete less with the roots of other companion crops \\[[@B1]\\]. Genetic and environmental factors determine germination rate, speed of germination, and vigor of seed and seedling \\[[@B8]\\]. The germination traits have been found to be complicated, and its inheritance varied among the crops. Speed of germination is genetically correlated to seed dormancy. Genes with additive action conferred high germination speed, while slow germination speed appeared to be induced by genes with pleiotropic effects \\[[@B9]\\]. Selection for quick germination was unsuccessful, and differences between genotypes were regulated by maternal tissue of seed ball \\[[@B10]\\].\n\nCurrently, crop improvement work in this species is limited. Interspecific hybridization has been attempted between different species of *Jatropha* with limited success \\[[@B11], [@B12]\\]. Improvement of *Jatropha* can be done through many breeding options like study of combining ability, heterosis breeding, mutation breeding, interspecific hybridization, and genetic transformation \\[[@B13]\\]. Better understanding and realizing the mode of inheritance of morphoagronomical traits of *J. curcas* in a set of half diallel crosses lead to improve its breeding program \\[[@B14]\\]. The combining ability and heritability estimation is one of the main strategies for heterosis utilization. Diallel cross analysis leads to a fruitful result for identification of genetic parameters regarding combining ability as well as dominance relationship of the parents. However, very little information is available on combining ability and heritability of germination traits in *J. curcas*. The main objective of the present study was to identify high heterotic parental combination in order to develop hybrid variety with good seed germination quality.\n\n2. Materials and Methods {#sec2}\n========================\n\nA greenhouse study was conducted at Universiti Kebangsaan Malaysia (UKM) during January to February 2010. Twenty-eight candidate plus plants (CPPs) were selected on the basis of their agromorphological traits from a population of 487 individuals of 20 accessions originated from nine different origins (Swaziland, Cape Verde, India, Thailand, Vietnam, Indonesia, Borneo, Malaysia, and South Africa) and multiplied by stem cuttings to develop materials for genetic diversity study. Parents of the hybridization program were selected through field evaluation followed by genetic diversity study. Six parents (*P* ~1~: Indonesia, *P* ~2~: Malaysia, *P* ~3~: Borneo, *P* ~4~: Indonesia, *P* ~5~: Cape Verde, and *P* ~6~: South Africa) were mated in a half diallel fashion during July to August 2009. The seeds of 6 \u00d7 6 half diallel population of *J. curcas* were sown directly in polybags (18 \u00d7 10 \u00d7 7\u2009cm) containing sand\u2009:\u2009soil\u2009:\u2009compost in the proportion of 1\u2009:\u20091\u2009:\u20091 at a depth of 3\u2009cm as described by Henning \\[[@B4]\\]. Watering was done three times per week to continue normal growth of seedlings. The experiments were laid in a completely randomized design with three replications (15 seeds in each). Data on different germination parameters were recorded at 24 hours intervals and continued until no further germination occurred. Seed germination criterion was visible protrusion on the soil surface of at least 0.5\u2009cm of the cotyledon and hypocotyls of the seedlings. The seedlings were evaluated as described in Seedling Evaluation Handbook \\[[@B15]\\].\n\nMean germination time (MGT) was calculated based on the equation of Ellis and Roberts \\[[@B16]\\]. Germination index (GI) was calculated as described in the Association of Official Seed Analysts \\[[@B15]\\]. Time to 50% germination (*T* ~50~) was calculated according to the following formula of Coolbear et al. \\[[@B17]\\] modified by Farooq et al. \\[[@B18]\\]: $$\\begin{matrix}\n{\\text{time}{\\,\\,}\\text{of}\\,\\, 50\\%\\,\\,\\text{germination}} \\\\\n{\\quad\\quad = \\frac{\\left\\lbrack {t_{i} + \\left\\{ {\\left( {N/2} \\right) - n_{i}} \\right\\}\\left( {t_{i} - t_{j}} \\right)} \\right\\rbrack}{n_{i} - n_{j}},} \\\\\n\\end{matrix}$$ where *N* is the final number of germination and *n* ~*i*~, *n* ~*j*~ are cumulative number of seeds germinated by adjacent counts at times *t* ~*i*~ and *t* ~*j*~ when *n* ~*i*~ \\< *N*/2 \\< *n* ~*j*~.\n\nVigor index was calculated according to following formula:$$\\begin{matrix}\n{\\text{seedling}{\\,\\,}\\text{vigor}{\\,\\,}\\text{index}{\\,\\,}\\left( \\text{SVI} \\right)} \\\\\n{\u2002 = \\left\\lbrack \\frac{\\text{seedling}{\\,\\,}\\text{length}{\\,\\,}\\left( \\text{cm} \\right) \\times \\text{germination}{\\,\\,}\\text{percentage}}{100} \\right\\rbrack.} \\\\\n\\end{matrix}$$\n\nThe speed of emergence was calculated according to following formula: $$\\begin{matrix}\n{\\text{speed}{\\,\\,}\\text{of}{\\,\\,}\\text{emergence}} \\\\\n{\u2002 = \\left( \\frac{\\text{number}{\\,\\,}\\text{of}{\\,\\,}\\text{seedlings}{\\,\\,}\\text{emerged}{\\,\\,}5{\\,\\,}\\text{days}{\\,\\,}\\text{after}{\\,\\,}\\text{sowing}}{\\text{number}{\\,\\,}\\text{of}{\\,\\,}\\text{seedlings}{\\,\\,}\\text{emerged}{\\,\\,}15{\\,\\,}\\text{days}{\\,\\,}\\text{after}{\\,\\,}\\text{sowing}} \\right)} \\\\\n{\\quad\\quad \\times 100.} \\\\\n\\end{matrix}$$\n\nGermination energy (GE) was determined as the percentage of germinating seeds five days after planting relative to the total number of seeds tested \\[[@B19]\\].\n\nNumber of days taken for first germination was counted from the date of treatment. Numbers of days to first and last germination for each trial were observed. Moreover, measurements of the percentage of germinated seeds were made.\n\nFinal germination percentage (GP) and seedling length (SL) were recorded after 15 days of planting as suggested by Dezfuli et al. \\[[@B20]\\]. For statistical analysis, the data of germination percentage was transformed to arcsin $\\sqrt{(100/X)}$. Experimental data was analyzed by a statistical package SAS, version 9.01 \\[[@B21]\\]. Treatments means were compared using Tukey\\'s test at 5% level of probability \\[[@B22]\\]. Combining ability was analyzed by method II (half diallel population with parents) and model I (fixed effect) of Griffing \\[[@B23]\\].\n\n3. Results {#sec3}\n==========\n\nThe analysis of variance revealed highly significant (*P* \\< 0.01) differences among the parents and their hybrids for all the germination traits indicating the existence of wider variability among the parental genotypes and the hybrids of *J. curcas*. The GCA and SCA variances were highly significant for all germination traits ([Table 1](#tab1){ref-type=\"table\"}). The GCA variances were higher than SCA variances for five traits, namely, number of days to first germination, speed of emergence, germination energy, germination index, and mean germination time. On the other hand, SCA variances were higher than GCA variances for four traits such as germination percentage, time of 50% germination, seedling length, and seedling vigor index ([Table 1](#tab1){ref-type=\"table\"}).\n\nThe highest germination percentage (98.96%) was observed in hybrid combination *P* ~2~ \u00d7 *P* ~4~ ([Table 2](#tab2){ref-type=\"table\"}), followed by the hybrids *P* ~1~ \u00d7 *P* ~2~ (96.46%). The lowest germination percentage (53.43%) was recorded from the hybrid *P* ~5~ \u00d7 *P* ~6~. The highest speed of emergence (70.67) was observed in the parent *P* ~1~ and the lowest (4.80) in *P* ~6~ ([Table 2](#tab2){ref-type=\"table\"}). Hybrids showed higher speed of emergence as compared to their parents and ranges from 3.73 to 69.47 ([Table 2](#tab2){ref-type=\"table\"}). The highest germination energy (53.44) was observed in the parent *P* ~4~ and the lowest (13.27) in *P* ~5~. The hybrid combination *P* ~1~ \u00d7 *P* ~5~ had the highest (80.00) germination energy followed by the hybrids *P* ~1~ \u00d7 *P* ~2~ (76.22). The parents *P* ~1~ and *P* ~5~ showed the highest (6.00) and lowest (2.78) germination index, respectively, but their hybrids showed an increase in germination index compared to their parents ([Table 2](#tab2){ref-type=\"table\"}). The highest (8.63) germination index was found in hybrid combination *P* ~1~ \u00d7 *P* ~5~ and the lowest (2.31) in hybrid *P* ~2~ \u00d7 *P* ~4~. The highest mean germination time was observed (10.18 days after first germination) in parent *P* ~6~ and the lowest (7.26 days) in *P* ~1~ ([Table 2](#tab2){ref-type=\"table\"}). The lowest mean germination time (6.14) was found in the hybrid combination *P* ~1~ \u00d7 *P* ~5~ and the highest (12.48) in *P* ~2~ \u00d7 *P* ~3~.\n\nThe lowest mean germination time was counted in hybrid combinations *P* ~1~ \u00d7 *P* ~2~, *P* ~1~ \u00d7 *P* ~5~, and *P* ~2~ \u00d7 *P* ~5~ ([Table 2](#tab2){ref-type=\"table\"}). The lowest (2.72) and the highest (5.22) *T* ~50~ values were observed in the parents *P* ~4~ and *P* ~6~, respectively. The lowest *T* ~50~ value (1.80) was found in the hybrid combination *P* ~1~ \u00d7 *P* ~5~ and the highest (5.23) in *P* ~4~ \u00d7 *P* ~6~. Most of the hybrids showed lower *T* ~50~ value compared to their parents ([Table 2](#tab2){ref-type=\"table\"}). Out of six parents, *P* ~1~ produced the tallest seedling compared to other parents ([Table 3](#tab3){ref-type=\"table\"}). The tallest (33.33\u2009cm) seedling was observed in the hybrid combination *P* ~2~ \u00d7 *P* ~5~ followed by *P* ~6~ \u00d7 *P* ~6~ ([Table 2](#tab2){ref-type=\"table\"}), while the shortest (19.66\u2009cm) seedling in *P* ~1~ \u00d7 *P* ~4~. The highest seedling vigor index (24.39) was observed in the parent *P* ~4~ and the lowest (14.41) in *P* ~5~ ([Table 2](#tab2){ref-type=\"table\"}). Hybrids showed higher seedling vigor index compared to their parents and ranges from 13.30 (*P* ~3~ \u00d7 *P* ~6~) to 29.78 (*P* ~1~ \u00d7 *P* ~2~) ([Table 2](#tab2){ref-type=\"table\"}).\n\nThe parent *P* ~1~ was the best general combiner for number of days to first germination (\u22120.56\\*\\*), speed of emergence (23.69\\*\\*), germination energy (14.27\\*\\*), germination index (1.42\\*\\*), mean germination time (\u22121.06\\*\\*), and time of 50% germination (\u22120.58\\*\\*), as it showed the highest significant desirable GCA effects ([Table 3](#tab3){ref-type=\"table\"}). The parent *P* ~2~ was the best general combiner for seedling length (1.81\\*\\*) and seedling vigor index (3.26\\*\\*) and *P* ~4~ for germination percentage (3.68\\*\\*). The parent *P* ~6~ was the poor general combiner for all the germination traits except *P* ~4~ for seedling length ([Table 3](#tab3){ref-type=\"table\"}).\n\nThe cross *P* ~1~ \u00d7 *P* ~5~ was the best specific combiner for germination percentage (16.10\\*\\*), speed of emergence (25.17\\*\\*), germination index (2.77\\*\\*), and seedling vigor index (5.91\\*\\*). The best specific combiner for germination energy (30.57\\*\\*), mean germination time (\u22122.20\\*\\*), *T* ~50~ (\u22121.76\\*\\*), and seedling length (3.36) was *P* ~2~ \u00d7 *P* ~5~ ([Table 4](#tab4){ref-type=\"table\"}). The poor specific combiner was the cross combination *P* ~4~ \u00d7 *P* ~5~ ([Table 4](#tab4){ref-type=\"table\"}). On the contrary, the poor specific combiners for most of the characters were *P* ~1~ \u00d7 *P* ~6~, *P* ~2~ \u00d7 *P* ~4~, *P* ~3~ \u00d7 *P* ~4~, *P* ~4~ \u00d7 *P* ~5~, and *P* ~4~ \u00d7 *P* ~6~ ([Table 4](#tab4){ref-type=\"table\"}). The highest negative significant SCA effects (\u22121.12\\*\\*) were observed in the cross combination *P* ~4~ \u00d7 *P* ~5~ for number of days to first germination and the highest positive significant SCA effects (1.13\\*\\*) in *P* ~4~ \u00d7 *P* ~6~ ([Table 4](#tab4){ref-type=\"table\"}).\n\n4. Discussion {#sec4}\n=============\n\nHighly significant mean sum of squares due to general and specific combining ability (GCA and SCA) for all the characters indicates that both additive and nonadditive types of gene action were involved for the expression of these characters. The magnitude of SCA variances was higher than that of GCA variances for germination percentage, time of 50% germination, seedling length, and seedling vigor index, which indicates the predominance of the non-additive gene effects for these characters. The remaining characters, namely, number of days to first germination, speed of emergence, germination energy, germination index, and mean germination time, showed higher GCA variances than their respective SCA variances which indicates the predominance of additive gene effect on the expression of these characters. So, it was evident that both additive and dominant genetic components are important for germination parameter in *J. curcas*. Sadeghian and Khodaii \\[[@B24]\\] also reported additive and nonadditive gene action for germination traits in sugar beet. The GCA effects (*g* ~*i*~) represent the additive nature of gene action. Both the nature (direction or sign) and magnitude of *g* ~*i*~ are important. Besides, performance of the parent *per se* is also considered together with *g* ~*i*~ to select the parent. GCA and SCA variances play significant role in the choice of parents. A parent with higher positive significant GCA effects is considered as a good general combiner.\n\nThe parent *P* ~1~ was the best general combiner for number of days to first germination, speed of emergence, germination energy, germination index, mean germination time, and time of 50% germination. Again, parent *P* ~2~ was the best general combiner for germination percentage, seedling length (cm), and seedling vigor index. The GCA was significantly different among six parents for germination traits in Brassica \\[[@B25]\\]. Seed vigor traits showed significantly higher GCA value in maize and sorghum \\[[@B26]\\]. Highly significant GCA effects were also reported for plant height, collar diameter, and number of leaves in nursery stage of *J. curcas* \\[[@B14]\\]. So, the parent *P* ~1~ and *P* ~2~ can be selected as a good general combiner and could be used in hybridization program for improvement germination quality of *J. curcas* seed.\n\nThe SCA effects denote the highest performance of some specific cross combinations and signify the role of non-additive gene action in the expression of the characters. High SCA effects may arise not only in crosses involving *H* \u00d7 *H* combinations, but also in those involving *H* \u00d7 *H* (Plus), *L* (Plus) \u00d7\u2009\u2009*H* (plus), and from *L* (plus) \u00d7\u2009\u2009*L* (plus). Thus, in practice, some of the low combiners should also be accommodated in hybridization program. Based on SCA effects, it was observed that the cross combination *P* ~1~ \u00d7 *P* ~5~ was found as a best specific combination for number of days to first germination, speed of emergence, germination percentage, germination index, and seedling vigor index. The cross combination *P* ~2~ \u00d7 *P* ~5~ was the best specific combination for germination energy, mean germination time, time of 50% germination, and seedling length and *P* ~2~ \u00d7 *P* ~5~ for number of days to first germination. Smith et al. \\[[@B27]\\] and Thseng and Hou \\[[@B28]\\] also observed similar results for genetic control of germination traits. SCA effect is an important criterion for the evaluation of crosses, and highly significant SCA effects were reported for seedling growth in *J. curcas* \\[[@B14]\\]. The SCA cannot be inherited and is only reflected by the interactions of allele or nonallele from the special parents \\[[@B29]\\].\n\n5. Conclusions {#sec5}\n==============\n\nCombining ability studies involving 6 \u00d7 6 half diallel populations showed both additive and non-additive gene action in the expression of different characters. The magnitude and direction of the significant GCA effects for the six parents provide meaningful information and clue to the future breeding program. The crosses with desirable specific combining ability could be used for exploitation of heterosis in *J. curcas* for germination and seedling growth. For most of the germination parameters, parents behaved poorer than those of hybrids.\n\nThis project is financed by the Universiti Kebangsaan Malaysia, under the Grant UKM-OUP-TK-15-65/2010. The authors would like to thank university authority for financial support.\n\n###### \n\nCombining ability ANOVA for nine germination traits of 6 \u00d7 6 diallel populations in *Jatropha curcas*.\n\n  Sources     df    NDTFG      GP           SE            GE           GI         MGT        *T* ~50~   SL          SVI\n  ----------- ----- ---------- ------------ ------------- ------------ ---------- ---------- ---------- ----------- -----------\n  GCA         5     1.04\\*\\*   178.66\\*\\*   1240.25\\*\\*   642.48\\*\\*   5.21\\*\\*   2.71\\*\\*   1.44\\*\\*   9.80\\*\\*    26.31\\*\\*\n  SCA         15    0.75\\*\\*   238.73\\*\\*   226.92\\*\\*    430.92\\*\\*   2.27\\*\\*   1.97\\*\\*   1.48\\*\\*   10.52\\*\\*   29.84\\*\\*\n  Error       40    0.099      1.88         0.63          0.27         0.02       0.03       0.01       0.70        0.55\n  GCA\u2009:\u2009SCA   ---   1.39       0.75         5.47          1.49         2.30       1.38       0.97       0.93        0.88\n\n\\*\\**P* \\< 0.01.\n\nNDFG: number of days to first germination, SE: speed of emergence, GP: germination percentage (%), GE: germination energy, GI: germination index, MGT: mean germination time (days), *T* ~50~: time of 50% germination, SL: seedling length (cm), and SVI: seedling vigor index.\n\n###### \n\nPerformance of parents and their hybrids for nine germination parameters in *Jatropha curcas*.\n\n  Genotypes           NDTFG   GP      SE      GE      GI     MGT     *T* ~50~   SL      SVI/100\n  ------------------- ------- ------- ------- ------- ------ ------- ---------- ------- ---------\n  Parents                                                                               \n  *P* ~1~             4.00    67.23   70.67   50.00   6.00   7.26    5.00       28.50   19.13\n  *P* ~2~             5.33    66.70   6.53    16.22   3.53   9.73    5.00       25.66   17.12\n  *P* ~3~             5.00    92.76   18.63   50.00   5.20   8.85    4.13       25.83   23.97\n  *P* ~4~             4.33    90.33   31.99   53.44   5.76   8.25    2.72       27.00   24.39\n  *P* ~5~             7.67    58.06   5.90    13.27   2.78   9.83    5.08       24.83   14.41\n  *P* ~6~             6.67    70.33   4.80    13.34   2.89   10.18   5.22       24.50   17.23\n                                                                                        \n  Hybrids                                                                               \n  *P* ~1~ \u00d7 *P* ~2~   4.67    96.46   59.28   76.22   7.63   6.55    2.00       30.83   29.78\n  *P* ~1~ \u00d7 *P* ~3~   5.33    82.10   39.93   33.11   5.61   8.50    2.00       25.83   21.12\n  *P* ~1~ \u00d7 *P* ~4~   4.33    82.60   20.00   60.00   4.66   8.84    4.00       19.66   16.21\n  *P* ~1~ \u00d7 *P* ~5~   4.33    93.10   69.47   80.00   8.63   6.14    1.80       29.33   27.31\n  *P* ~1~ \u00d7 *P* ~6~   4.33    86.96   7.76    43.11   4.37   8.51    4.78       25.66   22.33\n  *P* ~2~ \u00d7 *P* ~3~   5.33    96.40   10.46   56.39   5.21   12.48   4.59       30.33   29.24\n  *P* ~2~ \u00d7 *P* ~4~   5.00    98.96   16.56   16.65   2.31   8.27    4.50       29.16   28.86\n  *P* ~2~ \u00d7 *P* ~5~   5.00    86.90   37.83   73.44   6.35   6.51    2.21       33.33   28.97\n  *P* ~2~ \u00d7 *P* ~6~   5.00    93.00   3.73    36.55   4.36   8.61    5.00       30.83   28.65\n  *P* ~3~ \u00d7 *P* ~4~   5.00    56.10   6.33    13.44   2.63   9.25    5.00       28.83   16.15\n  *P* ~3~ \u00d7 *P* ~5~   5.00    80.33   8.26    53.31   4.19   8.00    4.73       28.33   22.76\n  *P* ~3~ \u00d7 *P* ~6~   5.33    56.23   6.00    23.47   2.59   9.56    5.00       25.66   13.30\n  *P* ~4~ \u00d7 *P* ~5~   4.00    90.00   3.80    30.00   3.83   10.49   5.00       23.66   23.10\n  *P* ~4~ \u00d7 *P* ~6~   6.33    80.66   4.23    16.39   3.73   9.15    5.23       24.66   19.90\n  *P* ~5~ \u00d7 *P* ~6~   4.33    53.43   12.50   26.55   2.46   9.75    4.64       31.66   16.94\n  MSD (0.05)          1.69    7.39    4.29    2.81    0.80   0.91    0.46       4.52    3.99\n\nMSD: minimum significant difference, NDFG: number of days to first germination, SE: speed of emergence, GP: germination percentage (%), GE: germination energy, GI: germination index, MGT: mean germination time (days), *T* ~50~: time of 50% germination, SL: seedling length (cm), and SVI: seedling vigor index.\n\n###### \n\nGCA effects for nine germination traits of 6 \u00d7 6 diallel populations in *Jatropha curcas*.\n\n  GCA effects   NDTFG     GP          SE           GE           GI          MGT         *T* ~50~    SL          SVI/100\n  ------------- --------- ----------- ------------ ------------ ----------- ----------- ----------- ----------- -----------\n  *P* ~1~       \u22120.56\\*   2.02\\*\\*    23.69\\*\\*    14.27\\*\\*    1.42\\*\\*    \u22121.06\\*\\*   \u22120.58\\*\\*   \u22120.38       0.17\n  *P* ~2~       0.03      5.70\\*\\*    \u22120.91\\*      1.67\\*\\*     0.17\\*      0.04        \u22120.12\\*\\*   1.81\\*\\*    3.26\\*\\*\n  *P* ~3~       0.07      \u22120.36       \u22124.99\\*\\*    0.18         \u22120.12       0.49\\*\\*    0.05        \u22120.34       \u22120.39\n  *P* ~4~       \u22120.26     3.68\\*\\*    \u22124.17\\*\\*    \u22124.37\\*\\*    \u22120.36\\*\\*   0.12        \u22120.01       \u22121.17\\*\\*   \u22120.08\n  *P* ~5~       0.32      \u22124.96\\*\\*   \u22120.57        1.44\\*\\*     \u22120.07       \u22120.13       \u22120.08\\*     0.83\\*      \u22120.72\\*\n  *P* ~6~       0.40\\*    \u22126.08\\*\\*   \u221213.05\\*\\*   \u221213.19\\*\\*   \u22121.04\\*\\*   0.54\\*\\*    0.74\\*\\*    \u22120.74\\*     \u22122.26\\*\\*\n  SEd (gi-gj)   0.16      0.68        0.40         0.26         0.07        0.08        0.04        0.42        0.37\n\n\\**P* \\< 0.05; \\*\\**P* \\< 0.01.\n\nNDFG: number of days to first germination, SE: speed of emergence, GP: germination percentage (%), GE: germination energy, GI: germination index, MGT: mean germination time (days), *T* ~50~: time of 50% germination, SL: seedling length (cm), and SVI: seedling vigor index.\n\n###### \n\nSCA effects for nine germination traits of 6 \u00d7 6 diallel populations in *Jatropha curcas*.\n\n  SCA effects         NDTFG       GP           SE           GE           GI          MGT         *T* ~50~    SL          SVI\n  ------------------- ----------- ------------ ------------ ------------ ----------- ----------- ----------- ----------- -----------\n  *P* ~1~ \u00d7 *P* ~2~   0.13        8.82\\*\\*     15.32\\*\\*    20.52\\*\\*    1.53\\*\\*    \u22121.22\\*\\*   \u22121.48\\*\\*   2.07\\*      4.40\\*\\*\n  *P* ~1~ \u00d7 *P* ~3~   0.76\\*      0.50         0.06         \u221221.09\\*\\*   \u22120.2        0.28        \u22121.64\\*\\*   \u22120.79       \u22120.61\n  *P* ~1~ \u00d7 *P* ~4~   0.09        \u22123.03\\*      \u221220.70\\*\\*   10.34\\*\\*    \u22120.90\\*\\*   0.99\\*\\*    0.41\\*\\*    \u22126.12\\*\\*   \u22125.83\\*\\*\n  *P* ~1~ \u00d7 *P* ~5~   \u22120.49       16.10\\*\\*    25.17\\*\\*    24.53\\*\\*    2.77\\*\\*    \u22121.46\\*\\*   \u22121.71\\*\\*   1.55\\*      5.91\\*\\*\n  *P* ~1~ \u00d7 *P* ~6~   \u22120.58\\*     11.08\\*\\*    \u221224.05\\*\\*   2.28\\*\\*     \u22120.52\\*\\*   0.23        0.46\\*\\*    \u22120.56       2.46\\*\\*\n  *P* ~2~ \u00d7 *P* ~3~   0.17        11.12\\*\\*    \u22124.80\\*\\*    14.78\\*\\*    0.65\\*\\*    3.16\\*\\*    0.48\\*\\*    1.53\\*      4.42\\*\\*\n  *P* ~2~ \u00d7 *P* ~4~   0.17        9.65\\*\\*     0.47         \u221220.41\\*\\*   \u22122.01\\*\\*   \u22120.68\\*\\*   0.45\\*\\*    1.19        3.73\\*\\*\n  *P* ~2~ \u00d7 *P* ~5~   \u22120.41       6.22\\*\\*     18.14\\*\\*    30.57\\*\\*    1.74\\*\\*    \u22122.20\\*\\*   \u22121.76\\*\\*   3.36\\*\\*    4.47\\*\\*\n  *P* ~2~ \u00d7 *P* ~6~   \u22120.49       13.44\\*\\*    \u22123.48\\*\\*    8.31\\*\\*     0.72\\*\\*    \u22120.77\\*\\*   0.20\\*      2.42\\*      5.70\\*\\*\n  *P* ~3~ \u00d7 *P* ~4~   0.13        \u221227.16\\*\\*   \u22125.68\\*\\*    \u221222.12\\*\\*   \u22121.40\\*\\*   \u22120.16       0.78\\*\\*    3.01\\*\\*    \u22125.33\\*\\*\n  *P* ~3~ \u00d7 *P* ~5~   \u22120.45       5.72\\*\\*     \u22127.34\\*\\*    11.93\\*\\*    \u22120.13       \u22121.16\\*\\*   0.60\\*\\*    0.51        1.92\\*\n  *P* ~3~ \u00d7 *P* ~6~   \u22120.20       \u221217.27\\*\\*   2.87\\*\\*     \u22123.27\\*\\*    \u22120.77\\*\\*   \u22120.27       0.04        \u22122.60\\*\\*   \u22126.00\\*\\*\n  *P* ~4~ \u00d7 *P* ~5~   \u22121.12\\*\\*   11.34\\*\\*    \u221212.64\\*\\*   \u22126.84\\*\\*    \u22120.25       1.71\\*\\*    0.91\\*\\*    \u22121.33       1.94\\*\\*\n  *P* ~4~ \u00d7 *P* ~6~   1.13\\*\\*    3.12\\*       0.27         \u22125.81\\*      0.61\\*\\*    \u22120.31\\*     0.33\\*      \u22120.76       0.29\n  *P* ~5~ \u00d7 *P* ~6~   \u22121.45       \u221215.470      4.94         \u22121.46        \u22120.94       0.53        \u22120.18       4.24        \u22122.04\n  SEd (Sij-Sik)       0.42        1.81         1.05         0.69         0.20        0.22        0.11        1.11        0.98\n  SEd (Sij-Skl)       0.39        1.68         0.97         0.64         0.18        0.21        0.1         1.03        0.91\n\n\\**P* \\< 0.05; \\*\\**P* \\< 0.01.\n\nNDFG: number of days to first germination, SE: speed of emergence, GP: germination percentage (%), GE: germination energy, GI: germination index, MGT: mean germination time (days), *T* ~50~: time of 50% germination, SL: seedling length (cm), and SVI: seedling vigor index.\n\n[^1]: Academic Editors: D. W. Archer, O. K. Douro Kpindou, and T. Nakazaki\n"}
+{"text": "1. Introduction {#sec1-polymers-12-00573}\n===============\n\nAn open wound is a type of injury that happens when the skin is punctured, torn, or cut. It specifically refers to the epidermis of the skin damaged by a sharp injury in pathology \\[[@B1-polymers-12-00573]\\]. The main factors that promote wound healing include mechanical protection of the localized wound, prevention of infection, absorption of wound exudates, and maintenance of the microenvironment \\[[@B2-polymers-12-00573]\\]. Modern wound dressings are technical products that create an ideal healing environment for wounds \\[[@B3-polymers-12-00573]\\]. Given these requirements, the ideal wound dressing should be designed with certain specialties, including excellent biocompatibility, good mechanical properties as barriers, the ability to create a native extracellular matrix to stimulate cell migration, ability to remove excessive exudates, and sustained drug release properties to prevent bacterial infections and promote tissue structure proliferation and recombination \\[[@B4-polymers-12-00573],[@B5-polymers-12-00573]\\]. No particular dressing has all of these characteristics, which is why the dressing is selected only after strict assessment. Therefore, composite films composed of biocompatible polymer and multifunctional drugs as wound dressings have attracted great attention in the field of wound healing \\[[@B6-polymers-12-00573]\\].\n\nRecently, electrospinning membranes and scaffolds have aroused great interest because of their biomedical applications in wound dressing, such as drug delivery, cell migration scaffolds, and transportation of gas and nutrients \\[[@B7-polymers-12-00573],[@B8-polymers-12-00573]\\]. Compared with the traditional wound dressing, electrospinning nanofibrous membranes have several advantages, including simulation of a natural extracellular matrix due to three-dimensional (3D) reticulate structure \\[[@B9-polymers-12-00573],[@B10-polymers-12-00573]\\], small pore size preventing bacteria invasion \\[[@B11-polymers-12-00573]\\], high porosity to accommodate active ingredients and drug molecular particles \\[[@B12-polymers-12-00573]\\], and high surface area to release loaded drugs effectively \\[[@B13-polymers-12-00573]\\]. Thus, electrospun composite membranes have been widely used in the research areas of wound dressings. Since most electrospinning membranes are usually used to load single drug molecules, the composite membranes cannot achieve multiple drug effects simultaneously, which thus makes it difficult to meet clinical treatment needs \\[[@B14-polymers-12-00573]\\]. Therefore, it is desirable in wound healing research that multiple drugs be loaded via the electrospinning procedure to achieve multi-effects, such as creating an antibacterial environment and promoting cell proliferation.\n\nWinter proposed the concept of a moist healing environment, which creates and sustains the ideal conditions to activate autolytic debridement of the wound, protecting newly formed cells, promoting angiogenesis and epithelial regeneration, reducing pain, and protecting the wound from bacteria and contamination \\[[@B15-polymers-12-00573],[@B16-polymers-12-00573]\\]. Foam, gel, hydrogel, and aerosol are common types of occlusive wound dressings, which maintain the proper humidity in the wound bed. Bacterial cellulose (BC), synthesized by *Acetobacter xylinum*, consists of nanofibers with a three-dimensional structure, bound together by hydrogen of inter-fibers and intra-fibers to form a dried-state or hydrogel \\[[@B17-polymers-12-00573]\\]. With the unique properties of high hydrophilicity, tensile strength, crystallinity, and good biocompatibility, it is a promising material used in wound healing applications. Photodynamic antibacterial chemotherapy (PACT) is a method of producing reactive oxygen species to induce antibacterial effects with visible lights combined with photosensitizer molecules. Porphyrin compounds are commonly used photosensitizers because of their highly active oxygen production. Studies showed \\[[@B18-polymers-12-00573],[@B19-polymers-12-00573]\\] that the combination of protoporphyrin IX (PPIX) and BC membrane has a significant antibacterial effect on *Staphylococcus aureus* and *Escherichia coli.*\n\nIn this work, PPIX and Vaccarin (Vac) particles were, respectively, impregnated into zein/EC fibers by electrospinning and used as an antibacterial layer (ABL) on the top and healing promotion layer (HPL) on the bottom. PETN was immersed into BC culture solution to develop a reinforcement layer (RFL). The above layers were further compounded to obtain multilayer composite (MC) membranes ([Figure 1](#polymers-12-00573-f001){ref-type=\"fig\"}). The morphology, component, and mechanical properties of the prepared membranes were evaluated by scanning electronic microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), and an electronic universal testing machine, respectively. In addition, the hygroscopicity, antibacterial activity, biocompatibility in vitro, and wound healing in vivo of the as-prepared membranes were also assessed. This study paves an innovated way for creating multi-effective composite membranes, which would have a positive effect in the design of next-generation wound dressings.\n\n2. Materials and Methods {#sec2-polymers-12-00573}\n========================\n\n2.1. Materials {#sec2dot1-polymers-12-00573}\n--------------\n\nVaccarin was purchased from Shanghai Shifeng Technology Co., Ltd. (Shanghai, China). Protoporphyrin IX (PPIX, MW = 562.66) was bought from Shanghai Vita Chemical Reagent Co., Ltd (Shanghai, China). Zein (MW = 35,000) was purchased from Sigma-Aldrich (St. Louis, MO, USA) Company. Ethyl cellulose (EC; 6--9 mPa s) was obtained from Aladdin Chemistry Co., Ltd. (Shanghai, China). *Gluconacetobacter xylinus* was cultured by our own lab. Polyester spun-laced nonwovens (40 g/m^2^) were provided by Jiangyin Haiyue Non-woven Fabric Co., Ltd (Wuxi, China). The mouse skin fibroblast cells (L929 cells) were obtained from Shanghai Institutes for Biological Sciences. ICR male mice were gained from Shanghai Slac Laboratory Animal Co., Ltd. (Shanghai, China). All of the chemicals were of analytical grade. All animal experiments were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals (Ministry of Science and Technology of China, 2006) and approved by the animal ethics committees of Jiangnan University (JN. No20180430c0480615).\n\n2.2. Preparation {#sec2dot2-polymers-12-00573}\n----------------\n\n### 2.2.1. Preparation of the ABL and the HPL {#sec2dot2dot1-polymers-12-00573}\n\nThe precursor solution was prepared by dissolving 2 g zein and 2 g EC in 10 mL acetic acid, and the entire solute concentration was maintained at 40% (*w/v*). The solution was stirred intensely at room temperature for 1 h, then, 0.5% PPIX (*w/w*) and 2% Vac (*w/w*) were separately added to obtain the electrospinning solution, which were subsequently stirred for 12 h. The obtained zein/EC/Vac and zein/EC/PPIX electrospinning solutions were, respectively, placed in syringes with a blunt-end metal capillary nozzle with a 1-mm core diameter. The parameters of the electrospinning apparatus were set to a voltage power (18 kV), a constant flow rate at 0.5 mL/h, and a distance of 15 cm between the nozzle and collector. The nanofibers were deposited on the aluminum foil, which covered a circular rotating drum at room temperature.\n\n### 2.2.2. Preparation of the RFL {#sec2dot2dot2-polymers-12-00573}\n\n*G. xylinum* was inoculated to the fermentation medium (D-mannitol 25 g/L, yeast extract 5 g/L, tryptone 3 g/L) in flakes and shaken gently so that the bacteria were evenly distributed in the culture medium. The bacterial cellulose membrane was formed after static incubation for a week at 30 \u00b0C. The inoculated fermentation medium was also added into the petri dishes containing PETN, and then sealed with plastic wrap to culture the BC/PETN membrane. The obtained BC and BC/PETN membrane were purified in 4 wt % NaOH solution to remove any bacteria and culture liquid, followed by rinsing with distilled water and freeze-drying.\n\n### 2.2.3. Preparation of the MC Membrane {#sec2dot2dot3-polymers-12-00573}\n\nThe MC membrane was prepared by using the zein/EC/PPIX nanofibrous membrane as the ABL, the BC/PETN membrane layer as the RFL, and the zein/EC/Vac nanofibrous membrane as the HPL. The BC/PETN membrane was used as a receiver to electrospin the zein/EC/PPIX and zein/EC/Vac solutions onto both sides of the BC/PETN. After hot pressing several times by the roller under the pressure of 1 MPa, the MC membrane was formed.\n\n2.3. Characterization {#sec2dot3-polymers-12-00573}\n---------------------\n\n### 2.3.1. Scanning Electron Microscopy (SEM) {#sec2dot3dot1-polymers-12-00573}\n\nThe microstructure and morphology of the samples were examined by an SEM at an accelerating voltage of 5.0 kV. The composite membranes were cut to expose the cross section, which was stuck on the microscope with a conductive adhesive. Before SEM imaging, a layer of gold was applied to the surface of the dried samples to avoid any sample charging effects. The images were collected at 200\u00d7, 1000\u00d7, or 5000\u00d7 magnification.\n\n### 2.3.2. Fourier Transform Infrared Spectroscopy (FTIR) {#sec2dot3dot2-polymers-12-00573}\n\nA Thermo Fisher Scientific FTIR Nicolet iS10 spectrometer was used to record the FTIR spectrum of the samples in the range of 650--4000 cm^\u22121^, and 16 scans were performed at a resolution of 4 cm^\u22121^.\n\n### 2.3.3. Antibacterial Activity Assessment of the ABL {#sec2dot3dot3-polymers-12-00573}\n\n*Staphylococcus aureus* ATCC-6538 and *Pseudomonas aeruginosa* CMCC(B) 10104, obtained from our laboratory, were cultured with Luria--Bertani medium (LB) on a rotary shaker at 37 \u00b0C and 200 rpm. When growth of bacteria was monitored by measuring the optical density (OD; 0.6--0.7, 600 nm), we inoculated 0.1 mL of bacterial solution evenly on the surface of the culture medium plates. Samples were prepared by cutting the materials into round-shaped disks with diameters of 10 mm. The zein/EC/PPIX membrane was tested as the treated sample, while zein/EC was set as the control, both of which were placed in bacteria-containing agar plates. After illumination for 30 min, plates were incubated under 37 \u00b0C for 24 h in a thermostatic incubator. The formed inhibition zone was recorded and the diameter was measured to calculate the average width.\n\n### 2.3.4. Mechanical Properties of the RFL {#sec2dot3dot4-polymers-12-00573}\n\nAn electronic universal testing machine (KDIII-5) was applied to test the tensile mechanical properties of the BC membranes (dry and wet, respectively) and the BC/PETN composite membranes. The pretension was set as 50 cN, the distance between two clamps that hold the sample was 30 mm, the effective stretching length was 50 mm, and the testing speed was 10 mm/min under the temperature of 25 \u00b0C and relative humidity of 65%. Every sample was tested 5 times, and then the breaking strength, elongation at break, and Young's modulus were calculated.\n\n### 2.3.5. Hygroscopicity Test of the RFL {#sec2dot3dot5-polymers-12-00573}\n\nThe hygroscopicity test was carried out according to the Chinese standard YY/T 0471.1-2004 at room temperature (21 \u00b1 2 \u00b0C) and a relative humidity of 60% \u00b1 15%. To determine the swelling behavior of the BC and BC/PETN membranes, samples were prepared to a size of 2 cm \u00d7 2 cm, weighed (the weight was set as M~0~), and immersed into simulated body fluids at 37 \u00b0C \\[[@B20-polymers-12-00573]\\]. The swelling rate change of all samples and the different swelling environments were measured at a specific time point until the equilibrium as reached by using the following Formula (1): where Q is the swelling ratio, while M~1~ represents the membrane weights at a specific time. The tests were performed in triplicate.\n\n### 2.3.6. In Vitro Biocompatibility Assay of the HPL {#sec2dot3dot6-polymers-12-00573}\n\nThe zein/EC/Vac nanofibrous membrane was plated at the bottom of the 24-well plates after sterilization. L929 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gbico) with 10% fetal bovine serum (FBS; Gbico), and 100 U/L penicillin and 100 \u03bcg/mL streptomycin (Gibco, Carlsbad, CA, USA). The cells were incubated at 37 \u00b0C under an environment with 5% CO~2~. When grown to the logarithmic growth phase, L929 cells were incubated in the 24-well plates with the sterilized zein/EC/Vac nanofibrous membrane by DMEM medium containing 10% FBS. After 24 h of incubation, following with removal of the culture medium with 3--5 times washing in PBS (phosphate buffer saline), PBS with 2.5 % (*w/w*) glutaraldehyde was added at 4 \u00b0C for 4 h. Finally, the solution was removed and the membranes were washed with PBS 3 times, then dried at room temperature. The morphology of the L929 cells growing on the surface of the zein/EC/Vac nanofibrous membrane was observed by SEM.\n\n2.4. In Vivo Animal Experiments {#sec2dot4-polymers-12-00573}\n-------------------------------\n\n### 2.4.1. Establishment of a Skin Wound Model {#sec2dot4dot1-polymers-12-00573}\n\nThirty-six male ICR mice (8 weeks old, 25--30 g) were randomly divided into three groups---sterile gauze group (Control), multilayer composite membrane group (MC), and Nano-Ag dressing group (Nano-Ag), respectively. Mice were fixed, shaved, and sterilized after anesthesia by intraperitoneal injection of 50 mg/kg pentobarbital sodium solution. Sequentially, a round full-thickness skin defect with a diameter of 10 mm was performed on the back of the mice. A digital camera was used to record the shape of the wound, and the diameter of the wound was scaled and compared. Finally, each wound was bandaged with wound dressings from different groups and the mice were fed in different cages. The process of the skin defect preparation is shown in [Figure 2](#polymers-12-00573-f002){ref-type=\"fig\"}.\n\n### 2.4.2. Wound Observation {#sec2dot4dot2-polymers-12-00573}\n\nDigital photographs of the wounds were recorded by a digital camera and wound size was measured on the day of operation and on various postoperative days (3rd, 7th, and 10th). The wound healing rate was calculated by the difference between the original wound area and the wound area on the 3rd, 7th, and 10th postoperative day as a percentage of the original wound area.\n\n### 2.4.3. Histological Analysis {#sec2dot4dot3-polymers-12-00573}\n\nSkin samples were collected after 3rd, 7th, and 10th day of treatment with various dressings (4 mice were sacrificed at each time point in each group), immersed in 4% paraformaldehyde, and paraffin-embedded sections were stained with hematoxylin and eosin (H&E; Baso, Taipei, Taiwan). All of the pathological sections were observed under a light microscope (Leica DM4000B, Leica, Wetzlar, Germany).\n\n2.5. Statistical Analysis {#sec2dot5-polymers-12-00573}\n-------------------------\n\nAll data are presented as mean \u00b1 SD and analyzed using Student's *t*-test or one-way ANOVA followed by Duncan's multiple range test. *p* \\< 0.05 was considered statistically significant.\n\n3. Results and Discussion {#sec3-polymers-12-00573}\n=========================\n\n3.1. Morphological and FTIR Analysis {#sec3dot1-polymers-12-00573}\n------------------------------------\n\nZein, a natural polymer material, has previously been broadly prepared by the electrospinning technique to be used as a drug carrier \\[[@B21-polymers-12-00573],[@B22-polymers-12-00573]\\]. However, it is difficult to maintain the fiber morphology of zein because of its unstable and extremely swellable nature in water, limiting its application as a drug carrier \\[[@B23-polymers-12-00573]\\]. Ethyl cellulose (EC) has high water stability, better general stability, fiber-forming properties during electrospinning, and good biocompatibility, which makes it a great potential for application in wound dressings \\[[@B24-polymers-12-00573]\\]. Therefore, we used these two materials to electrospin, loading them with the drugs Vac and PPIX as the ABL and the HPL, respectively, and then the morphology of the different membranes was evaluated.\n\nThe nanostructure and three-dimensional network structure of the ABL can be clearly seen in [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}a. The nanofibers with an average diameter of 380 \u00b1 92 nm were smooth and continuous without droplets, beads, adhesion, or dissolution. The crystallization and distribution of the nanofibers were uniform and well formed with high spinnability. [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}b shows the planar SEM image of the RFL. The nanofibers from BC were intertwined with the fibers of the PETN nonwoven fabric, and part of the BC nanofibers could intersperse between the PETN fibers, forming a three-dimensional network structure. We concluded that the BC fibers produced by the microorganisms firstly attached to the randomly arranged PETN fibers, and then synthesized a large number of BC nanofibers, which filled the pores and surface of the PETN to form a topological network structure \\[[@B25-polymers-12-00573]\\]. The BC nanofibers were self-assembled and compounded with the nonwovens in the form of a cellulose membrane and a nanofibrous network, in which the BC membrane adhered to the surfaces of the nonwovens and the BC nanofibers interspersed into the voids between the PETN fibers to form a tightly bonded BC/PETN composite membrane \\[[@B25-polymers-12-00573],[@B26-polymers-12-00573]\\]. An SEM image of HPL is shown in [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}c, which demonstrates that the surface of the composite nanofibers was smooth without droplets, beads, adhesion, or dissolution. Zein/EC/Vac fibers with an average diameter of 414 \u00b1 102 nm were irregular, which might be due to the increased surface tension of the mixed solution by Vac, making the jet difficult to differentiate in the electric field. Due to the surface tension of the mixed solution, the composite droplet outside the spinning nozzle was unstable when being drafted, leading to the fiber diameter being uneven and to an increasing fiber discrete degree. In addition, the drug was dissolved in the solvent and existed in the form of ions, which caused charge accumulation on the surface of the jet and divided the jet into smaller jets under relatively high electric field intensity \\[[@B27-polymers-12-00573]\\]. All of the above situations would result in uneven fibers. The SEM image of the cross-section of the multilayer composite membranes shows a three-layer nanofibrous structure with an average thickness of 143 \u00b1 10.35 \u03bcm ([Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}d).\n\n[Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}e displays the FTIR spectrum of the ABL nanofibers. The broad peak at 3294 cm^\u22121^ in the spectra of zein was due to the stretching vibration of N=H and O=H of the protein chain, and the characteristic bands observed at 1647 cm^\u22121^, 1534 cm^\u22121^, and 1230 cm^\u22121^ corresponded to C=O stretching (amide I) and N=H bending vibration (amide II) \\[[@B28-polymers-12-00573]\\]. The spectra of EC illustrate that the peaks at 2972--2870 cm^\u22121^, 1375 cm^\u22121^, and 1055 cm^\u22121^ represent the stretching vibration of --CH~3~, the bending vibration of --CH~3~, and --C--O--C stretching in the cyclic ether \\[[@B29-polymers-12-00573],[@B30-polymers-12-00573]\\]. From the FTIR spectra of PPIX, three obvious characteristic peaks of stretching vibration at 3424 cm^\u22121^ (N--H), 2916 cm^\u22121^ (C--H), and 1708 cm^\u22121^ (C=O) could be observed. The absorption band at 1200--1500 cm^\u22121^ was the skeleton vibration of the porphyrin ring, which could be attributed to the symmetric and asymmetric absorption peaks of the C--H, C=C, and C=N bonds in the pyrrole ring \\[[@B31-polymers-12-00573]\\]. The spectrum of the zein/EC/PPIX composite nanofibers had the characteristic absorption peaks of PPIX, zein, and EC without new absorption peaks, indicating that the components in the composite nanofibers were only physically mixed rather than chemically bonded. [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}f shows the FTIR spectra of the RFL. The characteristic absorption peaks of PETN at 2964 cm^\u22121^, 1714 cm^\u22121^, 1244 cm^\u22121^, and 721 cm^\u22121^ were assigned to the stretching vibration absorption peaks of --CH~2~, --C=O, and --C--O--C in the polyethylene terephthalate macromolecular structure and the bending vibration of the --CH bond on the benzene ring, respectively \\[[@B25-polymers-12-00573]\\]. Due to the large number of hydroxyl groups on the surface of the BC membrane, it presented a strong characteristic absorption band at 3337 cm^\u22121^ \\[[@B32-polymers-12-00573]\\]. From [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}f, a red-shifted hydroxyl absorption peak was observed at 3252 cm^\u22121^ after forming a composite with PETN, while the --CH stretching vibration peak around 2926 cm^\u22121^ was enhanced by the --CH bond in BC and PETN, implying that a good physical composite was formed between BC and PETN. From [Figure 3](#polymers-12-00573-f003){ref-type=\"fig\"}g, the FTIR spectra of the Vac show three characteristic peaks at 3450 cm^\u22121^, 1654 cm^\u22121^, and 1450 cm^\u22121^, which were associated with the stretching vibrations of O--H, C=O, and C=C, respectively \\[[@B33-polymers-12-00573]\\]. The interactions between polymer and drug molecules such as the hydrogen bond, electrostatic adsorption, and hydrophobic effect would increase compatibility while preparing the composite nanocellulose membrane \\[[@B34-polymers-12-00573]\\]. Zein and Vac possessed free hydroxyl and carboxyl groups, generating a hydrogen bond interaction in the composite nanofiber \\[[@B35-polymers-12-00573]\\]. It was noted that no new absorption peak appeared in the FTIR spectrum of the zein/EC/Vac nanofibrous membrane, indicating that only physical mixing existed among the components in the drug-loaded nanofibers. It was testified that the raw materials could maintain their original activity and properties without chemical reaction.\n\n3.2. Antibacterial Activity Assessment of the ABL {#sec3dot2-polymers-12-00573}\n-------------------------------------------------\n\n[Figure 4](#polymers-12-00573-f004){ref-type=\"fig\"}a shows pictures of the inhibition zones obtained by incubating *S. aureus* or *P. aeruginosa* with the ABL membrane. The zein/EC nanofiber membranes displayed no inhibition zone, suggesting that it failed to exert an inhibition effect on *S. aureus* or *P. aeruginosa*, while the ABL membrane displayed a significant inhibition zone after the excitation by illumination. Protoporphyrin IX (PPIX), as one of the common photosensitizers, was shown to exert a significant antibacterial effect on *S. aureus* and *E. coli*. We speculated that the oxygen-free radicals generated by composite membranes loaded with PPIX can cut off the major chemical bonds, such as O--P, C--H, O--H, and N--H bonds of bacteria, then bacteria are decomposed and the ABL exerts an effect of sterilization \\[[@B36-polymers-12-00573]\\]. On the other hand, the PPIX could form singlet oxygen when being excited by illumination, which would destroy the cell membrane structure of the bacteria, reducing the vitality of the cell and inhibiting cell growth \\[[@B31-polymers-12-00573]\\].\n\nFurthermore, we also found that the inhibition zone of the ABL membrane in the *P. aeruginosa* group was significantly bigger than that in the *S. aureus* group (*p* \\< 0.05, [Figure 4](#polymers-12-00573-f004){ref-type=\"fig\"}b). *P. aeruginosa*, also named green pus bacillus and belonging to the category of Gram-negative bacteria, is one of the major causes of hospital infections and can lead to postoperative wound infection \\[[@B37-polymers-12-00573]\\]. *S. aureus*, as the representative bacterium of Gram-positive bacteria, can cause many serious infections and is the most common pathogen in human suppurative infection \\[[@B19-polymers-12-00573]\\]. Compared to Gram-negative bacteria, Gram-positive bacteria contain peptide polysaccharide with a dense reticular structure in the thick cell wall \\[[@B38-polymers-12-00573]\\]. The present study indicates that the PPIX-loaded zein/EC nanofibrous membrane possessed better antibacterial effects against Gram-negative bacteria, which could be relevant to the relatively thick cell wall of Gram-positive bacteria. The exact mechanism still needs further study.\n\n3.3. Mechanical Properties of the RFL {#sec3dot3-polymers-12-00573}\n-------------------------------------\n\nBC is considered a very promising wound dressing \\[[@B39-polymers-12-00573]\\]. However, wet BC is difficult to reprocess and dry BC membranes possess lower mechanical properties because of the porous matrices, showing some restricted application in wound dressings \\[[@B40-polymers-12-00573]\\]. To achieve an ideal material, nonwoven composites with BC have shown more adequate mechanical properties in the literature \\[[@B41-polymers-12-00573]\\].\n\nAs can be seen from [Table 1](#polymers-12-00573-t001){ref-type=\"table\"}, the BC/PETN membrane had a significantly increased elongation at the break in comparison with the dried BC membrane, which was similar to that of the wet BC membrane, suggesting that the BC/PETN membrane had good malleability. This was mainly due to the high elongation at the break of PETN itself \\[[@B42-polymers-12-00573]\\]. Additionally, BC nanofibers with a network structure were filled in between the PETN fibers, and the close entanglement of the fibers greatly improved the mechanical properties of the composite membranes \\[[@B43-polymers-12-00573]\\].\n\nThe dried BC membrane possessed a relatively higher Young modulus and larger rigidity, while the BC/PETN membrane had a decreased Young modulus in contrast to the dried BC membrane, still retaining certain strength. Considering the absorption performance, reprocessing, and easy preservation, the dried BC membrane was set as a layer of the multilayer composite structure for application research on wound dressings. However, the dried BC membrane was easy to break and difficult to preserve or reprocess. Non-woven fabric PETN was added for structural modification during BC membrane formation to modify the mechanical properties of the dried BC membrane with the mechanical effects generated from the entanglement of BC on PETN fibers and the original hydrogen bond interaction among the BC fiber.\n\nThe results revealed that the elongation at the break was largely improved, the flexibility was also increased, and malleability and strength were obtained when the BC membrane was combined with a non-woven fabric such as PETN to form composite materials. It successfully overcame the shortcomings of the dried BC membrane, being light, thin, crispy, and hard to spread. Meanwhile, the BC/PETN membrane was dry, flat, soft, resilient, and easily reprocessed, which laid the foundation for RFL application in wound dressing.\n\n3.4. Hygroscopicity Test of the RFL {#sec3dot4-polymers-12-00573}\n-----------------------------------\n\nModern wound-healing theories confirm that a moist healing environment can provide suitable growth conditions for cells in the damaged skin \\[[@B44-polymers-12-00573]\\]. Meanwhile, an eligible wound dressing not only provides a moist and suitable microenvironment, but also reduces local pain and secondary injury caused by the removal of the dressing for wound healing \\[[@B45-polymers-12-00573]\\]. However, the local wound skin becomes immersed and weakens the defense barrier against bacterial invasion when wound exudate increases. Therefore, timely drainage of wound exudate is needed to promote the healing of the wound. The ability of the wound dressing to absorb or expel wound secretions can be determined by a hygroscopicity test of the wound dressing \\[[@B46-polymers-12-00573]\\].\n\nIn this experiment, the BC/PETN and wet BC membranes were put into SBF solution at 37 \u00b0C, respectively, to test their hygroscopicities within 72 h ([Figure 5](#polymers-12-00573-f005){ref-type=\"fig\"}). Both of the samples had a fast swelling speed within 0.5--1 h, which then gradually decreased. The BC/PETN membrane reached equilibrium at 24 h with a swelling ratio of 883%, indicating a good swellability. Although the wet BC membrane had hygroscopic capacity with a water swelling ratio of 82% at 48 h, its hygroscopic ability was much lower than that of BC/PETN, which was related to the fact that the wet BC membrane contained a large number of water molecules. The good hygroscopicity of the BC/PETN composite membrane was attributed to the rich internal porous structure of BC and a three-dimensional network, resulting in a large specific surface area and a large space for adsorption of water molecules. In addition, the polyhydroxyl structure of BC and the upper and lower BC surface of the composite membrane also contributed to strong hydrophilicity and water swellability \\[[@B47-polymers-12-00573]\\]. Therefore, the BC/PETN composite membrane, as one of the basic structures of a wound dressing, was conducive to the absorption of wound exudate and could meet the requirements as a wound dressing.\n\n3.5. In Vitro Biocompatibility Evaluation of the HPL {#sec3dot5-polymers-12-00573}\n----------------------------------------------------\n\nEvaluation of the cell growth on the zein/EC/Vac nanofibrous membrane was observed through a cell direct contact test. As shown in [Figure 6](#polymers-12-00573-f006){ref-type=\"fig\"}, L929 cells, growing on the surface of the nanofibrous membrane, exhibited favorable morphology in the shape of flat polygons or long spindles, and the cells extended well with a relatively large number of cells. This implies that the zein/EC/Vac nanofibrous membrane would be an ideal material for cell growth, with good biocompatibility.\n\nVaccarin (Vac), an extract of Vaccariae semen, showed properties of promoting specific proliferation and migration of vascular endothelial cells, contributing to neovascularization \\[[@B48-polymers-12-00573]\\]. Vac loaded on the BC membranes as a wound dressing exhibited a beneficial effect due to Vac being released to promote endothelial tissue proliferation \\[[@B33-polymers-12-00573]\\]. Due to the good proangiogenic effect of Vac \\[[@B49-polymers-12-00573]\\], the zein/EC membrane loaded with Vac as the HPL has a good potential for application in the field of wound dressings.\n\n3.6. In Vivo Animal Experiment {#sec3dot6-polymers-12-00573}\n------------------------------\n\n### 3.6.1. Wound Observation {#sec3dot6dot1-polymers-12-00573}\n\nWe conducted an experiment on skin wounds using different dressings ([Figure 7](#polymers-12-00573-f007){ref-type=\"fig\"}a). The Control group showed more severe inflammation, with redness and swelling appearing around the wound on day 3. Both the multi-layer composite wound dressing group (MC group) and the Nano-Ag wound dressing group (Nano-Ag group) promoted a small amount of new epithelial tissue proliferation, while the Nano-Ag group showed some exudate on the wound surface. On day 7, the average wound healing rates of the three groups were 28.35% (Con), 62.15% (MC), and 54.63% (Nano-Ag), respectively. As demonstrated in [Figure 7](#polymers-12-00573-f007){ref-type=\"fig\"}b, the wound contraction in the MC group was remarkably higher than that in the Control and Nano-Ag groups. On day 10, the wound contraction of the MC and Nano-Ag groups reached 92.4% and 81.3% respectively, while the wound of the control group was still obvious.\n\nWe also found that the Nano-Ag dressing adhered to desiccated wound surfaces when the dressing was not removed for a long time, which induced trauma on removal. The BC layer in the MC membrane could maintain the wet healing environment of the wound due to its strong hygroscopicity. Before the dressing was removed, we firstly moistened the dressing with saline to make it reach full absorption and keep it moist, which could effectively avoid the occurrence of reinjury.\n\n### 3.6.2. Histological Analysis {#sec3dot6dot2-polymers-12-00573}\n\nWound healing is a complex biological process involving multiple cells and tissues, including tissue regeneration, epithelialization, and granulation tissue formation \\[[@B50-polymers-12-00573]\\]. The pathological changes of skin wound tissues at different time points can be observed by H&E staining. As shown in [Figure 8](#polymers-12-00573-f008){ref-type=\"fig\"}, on day 3, a great amount of deeply colored inflammatory cells infiltrating and accumulating on the wound surface could be observed in the Control group. However, the MC group exhibited the least inflammatory cells and a small amount of neovascularization (oval, mostly on the wound surface), while some inflammatory cell infiltration and a small amount of neovascularization were observed in the Nano-Ag group.\n\nOn day 7, there were still a large number of inflammatory cells infiltrating the Control group without regular skin structure and no epidermal structure. The MC group showed less inflammatory cells and more new neovascularization and fibroblasts, and thus had basically completed epithelialization. In the Nano-Ag group, part of the epidermis and the basic epidermal structure with a few fibroblasts and abundant new neovascularization could be observed.\n\nOn day 10, the epidermal structure in the Control group was not yet completely formed. The MC and Nano-Ag groups had produced a large amount of collagen fibers on the surface of the wound, forming the structure of the epidermis. Each layer was regular and orderly on the wound surface, similar to normal skin \\[[@B51-polymers-12-00573]\\]. Meanwhile, the granulation tissue had formed and grown into the wound, which was filled with granulation tissue and had begun fibrosis in the MC and Nano-Ag groups.\n\n4. Conclusions {#sec4-polymers-12-00573}\n==============\n\nIn conclusion, an MC membrane---including an ABL (zein/EC/PPIX), an HPL (zein/EC/Vac), and an RFL (BC/PETN)---used as wound dressing material was successfully prepared. Each layer was fully characterized and analyzed, the results of which confirmed that the ABL could effectively inhibit *Staphylococcus aureus* and *Pseudomonas aeruginosa*, and the RFL prepared by the in situ biological composite method possessed a desirable nanonetwork structure and exhibited good flexibility, mechanical properties, and good hygroscopicity. In addition, the HPL favored biocompatibility in vitro. The animal experiment showed that the wound healing rate of mice treated in the MC group reached 92.4% on day 10, which was significantly higher than that of the Control group. Histological examinations demonstrated that active fibroblasts and epithelialization could be observed, and granulation tissue was filled in the wounds to complete the process of tissue remodeling treated by MC and commercial Nano-Ag dressings. Given the general properties of the MC membrane and its satisfactory performance in the mouse wound model, the MC membrane demonstrates a promising future in the design and production of next-generation wound dressing materials.\n\nConceptualization, Q.W. (Qufu Wei); Methodology, Q.W. (Qufu Wei) and Y.Q.; Validation, Y.Q., Y.C. and S.X.; Formal analysis, Y.Q. and Q.W. (Qingqing Wang); Data curation, Q.W. (Qufu Wei) and Y.Q.; Writing-original draft preparation, Y.Q; Writing-review and editing, Q.W. (Qufu Wei); Visualization, W.H.; Supervision, Y.C. and S.X.; Project administration, Q.W. (Qufu Wei) and Y.Q.; Funding acquisition, Q.W. (Qufu Wei) and Y.Q. All authors have read and agreed to the published version of the manuscript.\n\nThis research was funded by the Natural Science Foundation of Jiangsu Province, grant number BK20180628 and Wuxi Science and Technology Development Fund, grant number WX18IIAN011.\n\nThe authors declare no conflict of interest.\n\n![Schematic illustration of the preparation of multilayer composite (MC) membrane by electrospinning. ABL, antibacterial layer; RFL, reinforcement layer; HPL, healing promotion layer; BC, bacterial cellulose.](polymers-12-00573-g001){#polymers-12-00573-f001}\n\n![Process of the skin defect preparation: (**a**) fixed and shaved after anesthesia; (**b**) skin disinfection and mark on skin; (**c**) wound modeling; (**d**) dressing on wound; and (**e**) bandaging.](polymers-12-00573-g002){#polymers-12-00573-f002}\n\n![Scanning electronic microscope (SEM) images of the ABL (**a**), the RFL (**b**), the HPL (**c**), and the cross-section of the MC (**d**); FTIR spectra of the ABL (**e**), the RFL (**f**), the HPL (**g**), as well as raw materials.](polymers-12-00573-g003){#polymers-12-00573-f003}\n\n![(**a**) The images of inhibition zones obtained by incubating *S. aureus* or *P. aeruginosa* with the zein/EC/PPIX nanofibrous membrane (ABL) and the EC/PPIX nanofibrous membrane. (**b**) Diameter statistics of the inhibition zone. \\* represents a statistically significant difference from the *S. aureus* group (*p* \\< 0.05).](polymers-12-00573-g004){#polymers-12-00573-f004}\n\n![Water swelling ratio of the BC/PETN and BC (wet) membranes.](polymers-12-00573-g005){#polymers-12-00573-f005}\n\n![SEM image of L929 cells cultured on the zein/EC/Vac membrane.](polymers-12-00573-g006){#polymers-12-00573-f006}\n\n![(**a**) Photographic images on days 0, 3, 7, and 10 of wounds treated with sterile gauze (Control group), multilayer composite membranes (MC group), and Nano-Ag dressing (Nano-Ag group). (**b**) Analysis of the wound healing rate in each group on days 3, 7, and 10. \\* represents a statistically significant difference from the Control group (*p* \\< 0.05). \\# represents a statistically significant difference from the Nano-Ag group (*p* \\< 0.05).](polymers-12-00573-g007){#polymers-12-00573-f007}\n\n![Histological images on days 3, 7, and 10 of the wounds treated with sterile gauze (Control group), multilayer composite membranes (MC group), and Nano-Ag dressing (Nano-Ag group).](polymers-12-00573-g008){#polymers-12-00573-f008}\n\npolymers-12-00573-t001_Table 1\n\n###### \n\nThe mechanical properties of the BC and BC/PETN membranes.\n\n  Sample          Breaking Strength (MPa)   Elongation at Break (%)   Young Modulus (MPa)\n  --------------- ------------------------- ------------------------- ---------------------\n  Dried BC        7.85 \u00b1 2.07               6.33 \u00b1 1.64               93.6 \u00b1 16.4\n  Wet BC          3.39 \u00b1 0.65\\*             34.39 \u00b1 5.32\\*            19.4 \u00b1 3.27\\*\n  Dried BC/PETN   3.40 \u00b1 0.34\\*             28.88 \u00b1 2.03\\*            11.77 \u00b1 1.90\\*\n\n\\* represents a statistically significant difference from the dried BC group (*p* \\< 0.05).\n"}
+{"text": "Inanimate objects worn and used by health care workers (HCW), such as neckties and stethoscopes, have been shown to be reservoirs for potential pathogens.[@bib1] Of particular concern in the pediatric setting are identity (ID) badges and lanyards. Many pediatric HCWs use them not only for identification but also as a distraction tool during examination or procedures. Children have an increased tendency to place these items in their mouth as HCWs lean over to examine or care for them, with children up to 6\u00a0years old having a hand-to-mouth frequency of 9.5 contacts per hour.[@bib2] This therefore completes the chain of transmission for a potential nosocomial infection.\n\nKotsanas et\u00a0al have demonstrated that ID badges and lanyards worn by HCWs may harbor pathogenic bacteria; however, such colonization has not yet been implicated in a bacterial outbreak.[@bib3] Similar findings have been described in the United Kingdom, which has already implemented a bare below the elbows policy in an attempt to reduce opportunistic pathogen transfer.[@bib4], [@bib5] There is a paucity of comparative data suggesting that ID badges may be similarly contaminated with viral pathogens. However, given the higher incidence of viral infections in pediatrics---up to 50% of preterm infants screened during their hospital stay had viruses detected in their nasopharynx[@bib6]---further evaluation of the viral burden and potential for nosocomial transmission of prevalent viruses are of both clinical and economic significance.\n\nThe principal aim of this study was to determine the contamination rates of ID badges and lanyards worn by pediatric HCWs with common respiratory and gastrointestinal viruses to evaluate their potential as vectors for both nosocomial and patient-to-patient transmission of viruses.\n\nMethods {#sec1}\n=======\n\nA cross-sectional study was performed on 2 occasions (October 2012 and August 2013) on a total of 99 ID badges and lanyards from pediatric HCWs (6 medical students, 24 junior and 34 senior medical staff, and 35 nursing and allied health staff) working on pediatric medical wards or the pediatric emergency department at Monash Children\\'s Hospital, Australia. The study was repeated over sequential years to account for seasonal variability. Sampling was performed in the middle of the day to allow for maximum opportunity for pathogen transfer during morning clinical activities.\n\nOnly lanyards worn around HCW\\'s neck were sampled because these were purported to have maximum contact with patients compared with ID badges fastened to a belt without a lanyard. All lanyards were hospital issued cloth based and 2-cm wide. ID badges were of uniform design and made from white polyvinyl chloride, with additional laminated cards attached to the lanyard behind the ID cards in some staff. One swab sample was collected from each HCW using a flocked nylon nasopharyngeal swab, premoistened with viral transport media (BD, Sparks Glencoe, MD). The use of flocked nasopharyngeal swabs is known to have a sensitivity of 98.4% when compared with nasopharyngeal aspirates[@bib7] and 79%-89% when compared with respiratory washings.[@bib8] The front and back surfaces of ID badges (regardless of the number of badges they carried) and distal 2\u00a0cm of lanyards were sampled with the same swab and placed in universal transport media.\n\nFor comparison, total respiratory and fecal results on pediatric patients in Monash Children\\'s Hospital during January 2012-December 2013 were obtained using the clinical laboratory information system. During the study period, a total of 2,977 respiratory swabs and 2,425 fecal swabs were sent, with a total of 1,688 and 112 positive respiratory and fecal swabs, respectively. Clinical samples taken during this period were used for direct comparison with study results. All laboratory samples were not processed by the study team, but by routine pathology laboratory staff as per protocol subsequently listed and by the enteric virus laboratory at the Murdoch Childrens Research Institute for the enteric virus polymerase chain reaction (PCR) assays.\n\nRespiratory viral testing {#sec1.1}\n-------------------------\n\nHCW ID samples were tested for respiratory viruses using viral PCR technique after an automated DNA-RNA extraction procedure, an identical process to routine clinical samples collected from pediatric patients. The Respiratory Pathogen 12 Assay (AusDiagnostics, Beaconsfield, NSW, Australia) was used to detect RNA or DNA from 9 common respiratory viruses ([Table\u00a01](#tbl1){ref-type=\"table\"} ). Picornavirus results did not differentiate between enteroviruses and rhinoviruses; therefore, they were left as a group.[@bib9] Table\u00a01Respiratory pathogen assay panelRespiratory pathogenInfluenza A virusH1N1 (swine) influenza virusInfluenza B virusRespiratory syncytial virusPicornavirus[\u2217](#tbl1fnlowast){ref-type=\"table-fn\"}Parainfluenza 1 virusParainfluenza 2 virusParainfluenza 3 virusHuman metapneumovirus[^1]\n\nGastrointestinal viral testing {#sec1.2}\n------------------------------\n\nHCW ID badges were tested for rotavirus and norovirus using reverse transcription- polymerase chain reaction (RT-PCR) assays. Viral RNA was extracted from the samples using the QIAamp viral RNA Mini Kit (Qiagen, Limburg, The Netherlands). Norovirus RNA was screened using a nested RT-PCR assay targeting a 266-bp region of the capsid gene.[@bib10] Rotavirus double-stranded RNA was screened using a VP6-targeted RT-PCR assay.[@bib11] Positive RT-PCR products were confirmed by sequence analysis. Pediatric clinical fecal samples were tested for rotavirus using the RIDAQUICK Rotavirus/Adenovirus Combi N1003 (R-Biopharm, Darmstadt, Germany).\n\nData analysis {#sec1.3}\n-------------\n\nResults were summarized as number (%) for categorical variables and analyzed using Fisher *t* test. Statistical significance was defined as *P*\u00a0\\<\u00a0.05. Data were analyzed using Stata 12.0 (StataCorp, College Station, TX). The study was approved as a quality assurance study by the Monash Health Human Research Ethics Committee.\n\nResults {#sec2}\n=======\n\nNinety-nine pediatric HCWs were recruited in this study. Forty-nine HCWs were recruited in October 2012 and 50 in August\u00a02013.\n\nIn our study, across both periods, the rate of positive respiratory ID badge-lanyard samples was 1.0% (95% confidence interval, 0-5.5) ([Table\u00a02](#tbl2){ref-type=\"table\"} ). This compares with an overall 56.4% positive rate in tests ordered on pediatric patients across the study period for respiratory viruses. One ID badge-lanyard swab over the study period was positive for parainfluenza virus type 3. This is in the context of 11 positive parainfluenza virus type 3 results out of 286 pediatric samples in the preceding 60\u00a0days prior to the positive HCW ID result, with no parainfluenza outbreak during that time.Table\u00a02Demographics and results of respiratory and fecal swabsGroupSurveillance samples (HCW)Positive, n (%, 95% CI)Clinical samples (patients)Positive, n (%)Total RPCR, October 2012491 (2, 0.05-10.9)11777 (66)Total RPCR, August 2013500157104 (66)Total SPCR, October 2012490452 (4)Total SPCR, August 2013500503 (6)[^2]\n\nThere were no positive HCWs samples for gastrointestinal viruses ([Table\u00a02](#tbl2){ref-type=\"table\"}). During the overall study period, 4.9% of pediatric fecal samples tested positive for norovirus and rotavirus.\n\nDiscussion {#sec3}\n==========\n\nIn the pediatric population, respiratory tract infections and gastroenteritis are the most common infectious presentations. These viruses are easily transmitted from person to person, and fomites can potentially act as vehicles for transmission. Nosocomial viral infections are common, with the frequency of respiratory syncytial virus infections between 30% and 70% in neonatal wards and 20% and 40% in pediatric wards.[@bib12]\n\nViruses are estimated to have a variable life span on dry inanimate objects, with viability ranging between 6\u00a0hours to 60\u00a0days.[@bib13] Many respiratory viruses, such as respiratory syncytial virus, human parainfluenza viruses, influenza virus, rhinovirus, and coronavirus, are enveloped and have been shown to survive on surfaces of fomites from several hours to several days. Enteric viruses on the other hand are usually nonenveloped and therefore can remain viable for weeks to months.[@bib2] However, our results are based on PCR detection rather than viral culture; hence, viability and infectivity of the aforementioned viruses are difficult to infer. Therefore, whether this translates to an increased risk of transmission and hence increased nosocomial infection rate remains to be seen.\n\nTo our knowledge, this is the first study looking specifically at potential viral contamination of ID badges and lanyards in a pediatric population. Our results demonstrate that they are not likely to be significant vectors in the nosocomial transmission of common respiratory and gastrointestinal viruses. This low contamination rate is despite the study coinciding with the peak season for respiratory viruses in both years and also the tail end of the peak of gastrointestinal viruses, where a higher rate of background infection would imply higher risk of transmission and hence colonization on fomites.\n\nThese results are not congruent with previous studies showing colonization of bacteria in ID badges and lanyards[@bib3], [@bib4], [@bib5]; however, there are certain factors that could explain the difference.\n\nFirst, studies investigating viral survival on environmental surfaces are conducted in artificially created experimental conditions. Bean et\u00a0al demonstrated that influenza viruses isolated from throat swabs could be transferred from nonporous or cloth surfaces to hands for up to 24\u00a0hours or 15\u00a0minutes, respectively, after contamination. However, this is only feasible with viral titers that exceed physiologic levels of viral shedding.[@bib14] Nonetheless, PCR methods used in this study are likely to detect viral loads below the transmission threshold, even assuming that detected RNA-DNA represents viable viruses.\n\nViral titers have also been shown to decline exponentially with time on inanimate surfaces. Therefore, unless swabs were taken immediately or close to initial contamination of the ID or lanyard, the viral titer could have been too low even for detection by PCR. The implication of this may be similar to the way viruses can become inactivated on human skin and can only survive on hands for a few minutes.[@bib15]\n\nFurthermore, respiratory pathogens may be spread via airborne or droplet transmission, but fecal pathogens are predominantly spread via contact transmission. Therefore, unless there is clear surface contact with the lanyard, it is difficult to establish successful transmission, thereby possibly accounting for the lack of any detected fecal nosocomial infections in this study.\n\nOur study was limited by a number of factors. Although the pediatric wards sampled are in close proximity and have a centralized humidity and temperature control systems, the sampling methods do not account for factors such as ultraviolet light exposure, which may affect viral survival. Furthermore, intersample variability caused by time differences between contact and sampling and disinfection of individual lanyards prior to sampling could not be accounted for because of the study method.\n\nThe relatively small sample size of our study limits the applicability to the general population of pediatric HCWs, and confirmation in another setting is warranted. The use of artificially contaminated ID badges and lanyards as positive controls could also be included to further evaluate the validity of the methodology. Nonetheless, our findings in the peak respiratory season suggest that ID badges and lanyards are not significantly contaminated with common respiratory and gastrointestinal viruses and are unlikely to represent major vectors of transmission in pediatric health care settings.\n\nWe thank Julieanne Quinn and Ainsley Swanson.\n\nConflicts of interest: None to report.\n\n[^1]: Picornavirus results did not differentiate between enteroviruses and rhinoviruses; therefore they were left as a group.\n\n[^2]: *CI*, confidence interval; *HCW*, health care worker; *RPCR*, respiratory virus polymerase chain reaction; *SPCR*, stool polymerase chain reaction.\n"}
+{"text": "![](medhist00027-0130.tif \"scanned-page\"){.248}\n"}
+{"text": "Introduction\n============\n\nHepatitis B virus (HBV) is a highly transmissible pathogen infecting humans for more than 1500 years ([@B48]). Despite the availability of a prophylactic vaccine for more than three decades now, HBV continues to pose one of the most prevalent health problems with about 240 million people worldwide chronically infected and accounting for over 600,000 deaths per year ([@B43]). Current therapeutic regimens include pegylated-IFN-\u03b1 (PegIFN-\u03b1) and nucleos(t)ide analogs (NAs). Both types of antiviral treatment are not capable of eliminating the virus and do not establish long-term control of infection after treatment withdrawal in the majority of patients.\n\nHBV is the prototype of the *hepadnaviridae* family and has evolved a distinctive and successful replication strategy, which allows its indefinite persistence in the liver of the infected host. Upon infection of the hepatocyte, the HBV genome translocates to the nucleus, where its relaxed circular, partially double stranded DNA is converted into a covalently closed circular DNA (cccDNA) molecule. The cccDNA is the template for the synthesis of six co-terminal mRNA transcripts ([@B36]; [@B33]). One of the transcripts, termed pre-genomic RNA (pgRNA), is the template for genome replication and encodes for the core and polymerase proteins. Translation of the transcripts occurs in the cytoplasm and the encapsidation of the pgRNA into core particles follows ([@B20]). Inside the core particle, the viral polymerase directs the synthesis of the minus DNA strand of the genome by reverse transcription of the pgRNA template, which then serves as the template for plus DNA strand synthesis. Mature core particles containing DNA genomes are then enveloped and released or cycled back to the nucleus to replenish the cccDNA pool ([@B36]). Persistent infection of HBV relies on the stable maintenance and proper function of the cccDNA pool in the nucleus of the infected hepatocytes. Variable levels of cccDNA can be found in different phases of the natural history of chronic HBV patients. In HBeAg-negative patients or inactive carriers, cccDNA transcription is about 10-fold lower than that in HBeAg-positive patients, while cccDNA levels can be detected in cases with absent or very low viral replication ([@B42]; [@B22]).\n\nAmongst the smallest of all known virus genomes, the 3.2 kb HBV genome contains four partially overlapping open reading frames (ORFs). HBV transcription is regulated by its four promoters and by *cis*-acting viral elements including two enhancers (enhancer I and enhancer II) and a negative regulatory region that depend on host transcription factors for their function ([@B26]). Additionally, a number of epigenetic modifications have been identified which regulate viral replication and viral gene expression. Non-integrated nuclear HBV DNA associates with histones combined with HBcAg to form stable cccDNA minichromosomes ([@B6]). Chromatin condensation of cccDNA is a critical step for the regulation of viral gene expression because it determines the accessibility of DNA to the regulatory transcription factors. The acetylation status of cccDNA bound histones controls HBV replication in a fashion identical to that seen in human genes; hypoacetylation correlates with low viral replication and hyperacetylation leads to increased HBV replication ([@B30]). In addition, both viral and host DNA are known to be targets for methylation in chronic HBV infection suggesting dual effects of methylation as potentially both protective and harmful for the host ([@B14]; [@B40]; [@B17]). This review will provide an overview of how epigenetic factors, including genomic DNA methylation and histone modifications, contribute in HBV persistence and HBV-induced cancer as well as their possible therapeutic implications in chronic HBV infection.\n\nChromatin Organization and Epigenetic Modifications\n===================================================\n\nA number of epigenetic modifications have been recently identified that control viral replication in chronic HBV infection. Chromatin condensation of cccDNA is a critical step in the regulation of viral gene expression because it determines the accessibility of DNA to the regulatory transcription factors. It can be modulated through a variety of mechanisms, including posttranslational covalent modifications of histone tails, ATP-dependent chromatin remodeling events and recruitment of repressor factors on methylated DNA. Methylation is another common cellular defence mechanism in mammalian cells known to silence invading foreign DNA and viral genomes. It permits binding of protein complexes with chromatin-modulating properties and strictly depends on where the methyl group is located. Many host and viral promoters are enriched for CpG dinucleotides and methylation of cytidine leads to gene inactivation. Generally, DNA methylation at the promoter region leads to repression of gene expression, because the 5-methyl-cytosine interferes with the recognition and binding of transcriptional factors, to diminish mRNA transcription ([@B5]).\n\nHistone Acetylation\n===================\n\nHistone modifications are all reversible and mainly localized at the amino-terminal histone tails. They include acetylation, methylation, phosphorylation, sumoylation, ubiquitination, ADP-ribosylation, deamination, and the non-covalent proline isomerization (histone H3) ([@B23]). To further increase the complexity, many of these modifications occur multiple times at the same residue of a histone tail and they influence gene expression patterns by two different mechanisms: (1) histone acetylation, which alters chromatin packaging allowing access to transcriptional machinery; and (2) by histone methylation, which generates interactions with chromatin-associated proteins.\n\nSeveral enzymes catalyze these processes, including histone acetyltransferases (HATs), histone deacetylases (HDACs), histone methylatransfecrases (HMTs), and histone demethylases (HDMTs). HATs and HDACs regulate transcription by selectively acetylating or deacetylating the \ud835\udf00-amino group of lysine residues in histone tails. Histone acetylation induced by HATs promotes chromatin opening and associates with gene transcription while histone hypoacetylation induced by HDACs diminishes the accessibility of the nucleosomal DNA to transcription factors and is associated with gene silencing ([@B4]).\n\nDNA Methylation\n===============\n\nDNA methylation occurs at the 5-methyl cytosine predominantly in the context of CpG dinucleotides with S-adenosyl methionine as the methyl donor ([@B5]). CpG dinucleotides are often found accumulated in conserved regulatory regions (CpG islands) demonstrating their functional importance. The mammalian DNA methylation machinery consists of the DNA methyltransferases (DNMTs), which are responsible for the enzymatic addition of methyl groups, and the methyl-CpG-binding proteins (MBPs), which identify the methylation pattern ([@B12]). The DNMT family includes DNMT1, DNMT2, DNMT3A, DNMT3B, and DNMT3A. DNMT1 maintains the methylation pattern during cell division and methylates hemimethylated CpG islands. DNMT3A and DNMT3B can methylate unmethylated and hemimethylated CpG islands while DNMT2 lacks methyltransferase capabilities but plays a role in methylation of structural RNA ([@B16]).\n\nDNA hypomethylation signifies one of the major DNA methylation states and refers to a relative situation that indicates a decrease from the \"normal\" methylation level. Hypomethylation in transcriptional regulatory regions generally induces gene silencing, either directly, by blocking the binding of transcription factors to their recognition sequences or indirectly, by preventing transcription factors from accessing their target sites through the attachment of MBPs. In turn, methyl-CpG binding domain proteins can recruit histone-modifying and chromatin remodeling complexes, such as HMTs and HDACs, to methylated sites resulting in histone methylation of certain amino acids in a histone ([@B25]). HMTs catalyze the transfer of methyl groups to histones and thus they regulate DNA methylation through chromatin-dependent transcription control. Histone methylation can cause transcription repression or activation depending on the target sites and thus serves in both epigenetic gene activation and silencing.\n\nAcetylation of the cccDna Minichromosome\n========================================\n\nHepatitis B virus cccDNA is formed by histone and non-histone proteins. Hypoacetylation of the cccDNA-associated H3 and H4 histones and the recruitment of the cellular acetyltransferases p300/CREB-binding protein (CBP) and HDAC1 onto cccDNA have been associated with low HBV replication both *in vivo* and *in vitro* ([@B30]). In addition, the presence of class I/II histone deacetylase inhibitors (valproic acid and trichostatin (TSA)) has been shown to induce an increase in HBV replication and the upregulation of cccDNA-bound H4 histones in a HepG2 cell based model ([@B30]) (**Figure [1](#F1){ref-type=\"fig\"}**; **Table [1](#T1){ref-type=\"table\"}**). However, another study that used duck hepatitis B virus (DHBV) *in vitro* showed that a number of HDAC inhibitors, including TSA, suppressed cccDNA transcription and reduced HBV replication in a dose-response manner ([@B45]; [@B22]). Therefore it can be assumed that a cellular function sensitive to HDAC inhibitors is required for HBV RNA transcription.\n\n![**Schematic representation of the chromatin changes on cccDNA in relation to viral replication.** The acetylation status of cccDNA-bound histones and the recruitment of chromatin modifying enzymes onto cccDNA change in response to IFN\u03b1 treatment and HBx status. In high viral replication of the wild type virus or in the absence of IFN\u03b1 treatment, cccDNA-bound histones are hyperacetylated, cccDNA-associated chromatin is in an open configuration and pgRNA is actively transcribed. In cells replicating an HBx mutant and in IFN\u03b1-treated cells, cccDNA-bound histones are hypoacetylated, the recruitment of the p300 acetyltransferase is severely impaired, whereas the recruitment of the histone deacetylases (HDACs) hSirt1 and HDAC1 as well as the polycomb protein enhancer of zeste homolog 2 (Ezh2) is increased. In the absence of HBx, hypoacetylation is accompanied by the recruitment of heterochromatin protein 1 factors (HP1) and SET domain, bifurcated 1(SETDB1). Modified from [@B3].](fmicb-06-01491-g001){#F1}\n\n###### \n\nAn outline of the main research findings on the (A) methylation and (B) acetylation mechanisms involved in HBV infection.\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                                                                                                                                                   Reference\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------- -----------\n  **(A) Findings on HBV DNA methylation**                                                                                                                          \n\n  Methylation of integrated HBV DNA.                                                                                                                               [@B8]\n\n  HBx induces DNMT activity and hypermethylation of tumor suppressor gene promoters via DNMT3A1 and DNMT3A2 methylation.                                           [@B29]\n\n  Island II methylation correlates with low or no HBsAg production. cccDNA methylation correlates with viral gene expression levels.                               [@B38]\n\n  Methylation of integrated and non-integrated liver HBV DNA in islands I (60%) and II (50%).\\                                                                     [@B39]\n  Unmethylated serum HBV DNA.                                                                                                                                      \n\n  cccDNA methylation correlates with HBeAg-positivity in CHB patients and impairs virion productivity.                                                             [@B14]\n\n  HBV DNA is unmethylated in early carcinogenesis and highly methylated in cancer.\\                                                                                [@B13]\n  Methylation of HBcAg and HBsAg genes inhibits their expression.                                                                                                  \n\n  HBx recruits DNMT3A and induces the methylation and transcriptional silencing of IL-4 receptor and metallothionein-1F.                                           [@B47]\n\n  HBx expression correlates with DNMT1 and DNMT3A and hypermethylation of the p16^INK4A^ promoter in CHB patients.                                                 [@B50]\n\n  cccDNA methylation is associated with HBV viremia and aging in cirrhotic CHB patients.                                                                           [@B18]\n\n  HBV DNA methylation correlates with decreased viral replication and gene expression.                                                                             [@B40]\n\n  Increased expression of DNMT3 down-regulates viral protein and pgRNA production.\\                                                                                [@B37]\n  HBV induces DNMT overexpression and correlates with methylation of host CpG islands.                                                                             \n\n  HBV DNA methylation in CHB implicates island I in14%, island II 0.6% and island III 3.7% of cases.\\                                                              [@B17]\n  HBV DNA is unmethylated in CHB and highly methylated in HBV-related cancer.                                                                                      \n\n  HBV DNA is unmethylated in occult HBV.\\                                                                                                                          \n  CpG island I methylation correlates with HCC development.                                                                                                        \n\n  HBV CpG island distribution differs between HBV genotypes.                                                                                                       [@B46]\n\n  HBx induces the hypermethylation of the uPA promoter (via the recruitment of DNMT3A2) leading to liver regeneration impairment.                                  [@B28]\n\n  **(B) Findings on the acetylation of the cccDNA minichromosome**                                                                                                 \n\n  Low HBV replication correlates with cccDNA hypoacetylation and the recruitment of p300/CBP and HDAC1.\\                                                           [@B30]\n  Histone deacetylase inhibitors restore HBV replication.                                                                                                          \n\n  HBx recruitment onto cccDNA correlates with HBV replication and acetyltransferase upregulation.\\                                                                 [@B3]\n  In the absence of HBx, HBV decreased replication correlates with cccDNA hypoacetylation; p300 inhibition; reduced pgRNA, and deacetylase increase.               \n\n  IFN-\u03b1 treatment reduces DHBV acetylation of cccDNA-bound H3K9 and H3K27 histones but has no effect on cccDNA-bound H3K9me3 and H3K27me2 demethylases.\\           [@B22]\n  HDAC inhibitors block DHBV cccDNA transcription but not the long-lasting IFN-\u03b1-induced suppression of cccDNA.                                                    \n\n  IFN-\u03b1 inhibits viral transcription by cccDNA hypoacetylation through active recruitment onto cccDNA of HDAC and of the transcriptional repressor complex.        [@B2]\n\n  IL6 induces cccDNA hypoacetylation and silencing by reducing the binding of transcription factors (HNF1\u03b1, HNF4\u03b1, and STAT3) onto cccDNA.                         [@B27]\n\n  HBx recruitment onto cccDNA activates HBV transcription by counteracting chromatin-mediated transcriptional repression established by SETDB1, HP1 and H3K9me3.   [@B31]\n\n  IFN-\u03b1 represses HBV by reducing active PTMs in cccDNA and that this effect can be recapitulated with the C646 agent (inhibits p300/CBP)\\                         [@B35]\n  The repressive mark H3K27me3 is underrepresented in cccDNA                                                                                                       \n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nDNMT, DNA methyltransferase; uPA, urokinase-type plasminogen activator; CBP, CREB-binding protein; pgRNA, pregenomic RNA; cccDNA, covalently closed circular DNA; DHBV, duck hepatitis B virus; HDAC, histone deacetylase; HNF1\u03b1, hepatocyte nuclear factor 1\u03b1; HNF4\u03b1, hepatocyte nuclear factor 4\u03b1; SETDB1, SET domain, bifurcated 1; HP1, heterochromatin protein 1 factors; and PTM, posttranslational covalent modifications.\n\nHBV replication is known to be strongly inhibited by the administration of IFN-\u03b1, a type I IFN, that engages the IFN-\u03b1/\u03b2 receptor complex to modulate the transcription of the IFN-stimulating genes (ISGs) via the Jak/Stat signaling pathway. Recently it has been shown in HBV-replicating cells and in HBV-infected chimeric uPA/SCID mice that IFN-\u03b1 inhibits cccDNA-driven transcription by targeting the epigenetic control of cccDNA with the involvement of the chromatin remodeling Polycomb Repressive Complex 2 ([@B2]). In response to IFN-\u03b1, cccDNA-bound histones become hypoacetylated and both components of the transcriptional repressor complex, YY1 and Ezh2, and the HDAC1 and hSirt are actively recruited onto cccDNA ([@B2]). However, another study, using a DHBV cell based model demonstrated that IFN-\u03b1 reduced the acetylation of cccDNA-bound H3K9 and H3K27 histone residues but failed to induce cccDNA-bound H3K9me3 and H3K27me2 demethylases ([@B22]). In addition, HDAC inhibitors blocked DHBV cccDNA transcription but did not affect the long-lasting IFN\u03b1-induced suppression of cccDNA transcription ([@B22]). It is possible that the reduced acetylation of H3K9 and H3K27 by IFN-\u03b1 is catalyzed by HDACs that are not susceptible to the HDAC inhibitors used in the DHBV study or that this reduction is due to the disruption of a dynamic acetylation and deacetylation of histone H3 through preventing the recruitment of HATs onto DHBV cccDNA (**Figure [1](#F1){ref-type=\"fig\"}**) ([@B22]). Interestingly, a recent study demonstrated that IFN-\u03b1 represses HBV by reducing active PTMs on the cccDNA minichromosome and that this effect can be recapitulated with treatment with a small epigenetic agent, C646, which specifically inhibits p300/CBP HATs ([@B35]). Nevertheless, these findings suggest that IFN-\u03b1 induces a persistent condition of \"active epigenetic control\" of cccDNA, involving all HBV transcripts that may contribute to the persistent, yet reversible, \"off-therapy\" inhibition of HBV replication. Contrary to IFN-\u03b1, treatment with IL6, another activator of the intracellular Jak/Stat signaling pathway, reduces cccDNA acetylation and transcription but without affecting cccDNA chromatinization ([@B27]). Instead IL6 has been shown to inhibit cccDNA transcription by reducing the binding of essential transcription factors HNF1\u03b1 and HNF4\u03b1 to cccDNA and by redistributing STAT3 binding from the cccDNA to IL6 cellular target genes ([@B27]).\n\nHBx, a pleiotropic regulatory protein, acts as a promiscuous transactivator of viral and cellular promoters and is found in the cytoplasm and the nucleus of infected hepatocytes ([@B11]). HBx activates the transcription of host genes by interacting directly with nuclear transcription factors or by activating various signal transduction pathways in the cytoplasm. In addition to its *trans-* and *cis*-activating roles, HBx protein has been proven to be a potent epigenetic modifying factor in liver tissue. It has been reported to modify chromatin dynamics *in vivo* by favoring the transcription of a number of CREB-regulated genes via the recruitment of the cellular acetyltransferases CBP and p300 to their promoters ([@B11]). HBx is reported to activate HBV transcription by its recruitment onto cccDNA, through recruitment of PCAF/GCN5, p300 and CBP acetyltransferases onto cccDNA and through the inhibition of cellular factors involved in chromatin regulation, such as PP1/HDAC1 complex ([@B3]; [@B31]). On the contrary, in the absence of HBx, HBV replication was suppressed and this decrease correlated with the rapid hypoacetylation of cccDNA-bound histones, the severe impairment of the p300 recruitment and the reduced transcription of pgRNA from cccDNA, whereas the recruitment of the HDACs hSirt1 and HDAC1 preceded ([@B3]). Notably, the IFN-\u03b1-induced cccDNA repression through hSirt and HDAC1 up-regulation was reported to be HBx independent ([@B2]). A recent study showed that in the absence of HBx, HBV silencing associated not only with the deacetylation of histones but also with deposition of repressive chromatin markers (H3K9me2 and H3K9me3), the recruitment of heterochromatin protein 1 factors and the recruitment of SET domain, bifurcated 1 (SETDB1) histone methylatransferases that methylate H3K9 histone ([@B31]). Interestingly, SETDB1 has been shown to be an oncogene and is significantly associated with HCC disease progression ([@B44]).\n\nHBV DNA Methylation\n===================\n\nIn chronic HBV infection, viral DNA methylation has been identified as a novel host defence mechanism, leading to the downregulation of viral gene expression. However, the association of HBV DNA methylation with the methylation of host genes and the development of cancer imply a harmful effect on the host. HBV DNA can be methylated in human tissues in both non-integrated forms ([@B39]) and following integration into the human tissue ([@B8]).\n\nNon-Integrated HBV Genome\n=========================\n\nThe HBV genome contains three predicted CpG islands overlapping the start site of the S gene (island I); the region encompassing enhancer I, the X gene promoter (island II); and the region harboring the Sp1 promoter and the start codon of the P gene (island III) (**Figure [2](#F2){ref-type=\"fig\"}**) ([@B39]; [@B17]). Kaur and his group in France reported a 14% methylation frequency in CpG island I, 0.6% methylation in island II and 3.7% in island III in CHB patients ([@B17]). However, a computation study, reported that 50% of the HBV sequences examined lacked island I, while islands II and III were more conserved across genotypes ([@B46]). The authors argued that conflicting results between HBV methylation studies are due to different HBV genotypes examined. High viral mutation frequencies and high viral replication rates in CHB infection can lead to a great degree of variability in CpG island distribution throughout the viral genome ([@B46]).\n\n![**Schematic representation of the open reading frames (ORFs) of pre-core/core, polymerase (P), surface (S), and X proteins with the genome in open configuration; the four promoters, cp, sp1, sp2, and xp, as well as the main regulatory elements, enhancers I and II (Enh I and Enh II), are indicated.** The positions of the three predicted CpG islands within the HBV genome are shown as shaded boxes. Nucleotide numbering is based on the EcoRI restriction site as position 1.](fmicb-06-01491-g002){#F2}\n\nThe methylation status of island II has been greatly associated with reduced viral gene expression and replication as compared with islands I and III (**Table [1](#T1){ref-type=\"table\"}**). High methylation levels on island II were correlated with absent or low levels of HBsAg production while hypermethylation patterns were also reported in occult HBV cases and in HBsAg negative patients with HCC ([@B39]; [@B17]). In addition to transcriptionally regulatory genes, island II overlaps with the transcriptional control region of HBV cccDNA. Findings on the role of cccDNA methylation in HBV replication are controversial. Studies from several groups in Asia reported cccDNA methylation in human tissue and further showed that it impairs the cccDNA's replication capability and viral gene expression ([@B14]; [@B40]; [@B18]). However, the study of [@B14] included patients with a serum HBV DNA load of more than 10^3^ copies/ml, whilst Vivekandam and his group included liver cancer tissue; neither study made any correlation between cccDNA methylation and viral load, while both studies did not mention the presence of fibrosis. A third study from [@B17] included a small number of liver cirrhosis patients (*n* = 12), in which cccDNA methylation was associated with serum viremia. Contrary to these reports, a study from France, demonstrated that the HBV genome, including island II, is rarely targeted for DNA methylation in liver samples from CHB patients ([@B17]). A Korean study of cirrhotic patients, reported that increased cccDNA methylation is associated with older age ([@B18]). The authors postulated that considering that perinatal transmission is the main mode of HBV infection and the stability of cccDNA, it is possible that methylated cccDNA may be passed to daughter cells after mitotic division, and additional methylation of replenished cccDNA may increase methylation frequency of the cccDNA population in chronic HBV infection ([@B18]). Furthermore, CpG island II is in close proximity to the core gene promoter and enhancer II and its hypermethylation can suppress Pre-C/C gene transcription, and consequently HBeAg expression in CHB patients ([@B14]). HBeAg seroconversion is attributed to hotspot mutations in the precore/BCP region (A1762T/G1764A, G1896A, G1899A) that abrogate or reduce HBeAg secretion ([@B15]). However, in the absence of these mutations cccDNA methylation density was reported to be significantly higher in liver cells of HBeAg(-) patients than in HBeAg(+) patients ([@B14]). Interestingly, within the same HBeAg(+) patient group, the ratio of RC-DNA/cccDNA was lower in the cccDNA methylation positive samples than in the cccDNA methylation negative samples ([@B14]). It is therefore possible that, increased cccDNA methylation correlates with HBeAg clearance possibly via suppressing pre-C/C transcription, which in turn can lead to suppressed viral productivity.\n\nThe specific roles of the DNMTs involved in the HBV methylation processes have not been clearly identified yet. Transfection studies reported that DNMT1, DNMT2 and DNMT3 expression is up-regulated in response to HBV, leading to viral methylation, decreased HBV gene expression, viral replication, and host DNA methylation ([@B38], [@B37]; [@B24]). Co-transfection experiments with HBV and DNMT3 were shown to increase cccDNA methylation and to down-regulation of viral protein and pgRNA production ([@B37]).\n\nIn addition, to stimulate transcription by modifying chromatin dynamics HBx protein has been shown to silence the expression of some genes by increasing the total DNMT activity of the host. [@B28] reported that HBx induced the upregulation of DNMT1, DNMT3A1 and DNMT3A2, and selectively facilitated the regional hypermethylation of the promoters of certain tumor suppressor genes through *de novo* methylation via DNMT3A1 and DNMT3A2 recruitment ([@B37]). In addition, Zheng and his group demonstrated that HBx regulated epigenetic modifications by its physical interaction with DNMT3A, either by promoting DNMT3A recruitment to the promoters of some genes, such as MT1F and IL-4, and thus inducing their hypermethylation and downregulation, or by preventing DNMT3A recruitment to specific genomic loci and thus activating the transcription of genes, such as CDH6 and IGFBP3 ([@B47]). Fernadez and his group showed that most of the HBV genomes, although more methylated than the pre-malignant lesions, retained unmethylated the HBV gp3 gene, which codes for HBx ([@B13]). A recent study showed that the HBx protein impairs the expression of urokinase-type plasminogen (uPA), a serine protease essential for the activation of the hepatocyte growth factor (HGF) that activates hepatic regeneration ([@B28]). HBx-induced uPA inhibition is regulated epigenetically by the hypermethylation of the uPA promoter ([@B28]). In particular, in HBx-expressing cells, the CpG region of the uPA promoter was found to be 99.7% methylated resulting in the hypoactivation of pro-HGF and eventually hampering liver regeneration ([@B28]).\n\nHBV Integration and Hepatocarcinogenesis\n========================================\n\nHCC is the third most common cause of cancer globally and chronic HBV patients have 100-fold greater risk of developing hepatocellular cancer. Ninety percept of HBV-associated liver cancers show integration of the HBV genome within the human genome ([@B13]). The development of HCC in HBV infection involves two major mechanisms: (1) the viral integration in the host's genome causes *cis*-effects that inactivate tumor-suppressor genes and activate oncogenes and (2) the expression of trans-activating HBV proteins, such as the HBx and the PreS2 activators, which disrupt the signal transduction pathways and alter the expression of the infected hepatocyte ([@B32]; [@B28]).\n\nSimilarly to other oncoviruses (HPV16 and HPV18), the HBV genome, is almost unmethylated in the early stages of carcinogenesis while it becomes more methylated in the established HCC ([@B13]). Hepatitis C virus (HCV) contributes to carcinogenesis by inducing regional hypermethylation of CpG islands in the promoter regions of multiple genes ([@B7]). In HBV-induced carcinogenesis, HBx can accelerate hepatocarcinogenesis epigenetically by promoting hypermethylation of tumor suppressor genes by modulating DNMT1 and DNMT3A expression ([@B28]). Moreover, HBx was reported to induce hypermethylation of the E-cathedrin promoter via the activation of DNMT1 *in vitro* ([@B19]). In addition to DNMT expression, high HBx expression has been correlated with the hypermethylation of the promoter of the major tumor suppressor gene p16^INK4A^ and the subsequent reduction in p16 protein expression in non-cancerous tissue but not in HCC tissue suggesting that HBx plays an important role in the early stages of HBV associated HCC ([@B49]).\n\nHCC has been associated with high methylation of CpG island I which overlaps with the HBsAg gene starting site ([@B17]). Another study, confirmed the progressive presence of hypermethylation at the HBV gp2 locus that encodes for S viral proteins in primary liver tumors ([@B13]).\n\nHBV Epigenetic Control of Host's Immune Responses\n=================================================\n\nHost DNMT upregulation by viruses can be a non-specific innate response to infection. HBV has been shown to induce genome-wide DNA methylation changes, including immunoregulatory genes that are active against HBV ([@B1]). In particular, HBV replication was reported to cause the *de novo* methylation and decrease of IL4, which benefits the virus since IL4 expression inhibits HBV replication ([@B47]). In addition, unmethylated CpG dinucleotides have been shown to trigger toll-like receptors expressed in hepatic cells *in vitro*, which in turn can activate the NF-\u03baB pathway that plays a key role in the innate system's ability to inhibit HBV replication ([@B21]). Nevertheless, this potentially protective effect of DNMT upregulation may be offset over time either through viral manipulation of the host methylation machinery or through non-specific methylation of host CpG islands as a result of chronic over expression of DNMTs. For example, latent viruses, such as Epstein-Barr virus (EBV), are maintained in their latent state in part by methylation, suggesting that some viruses have evolved strategies to manipulate host DNMTs to their advantage. HBV viral proteins can lead to DNMT upregulation and eventually to methylation of the host genes, including oncogenes. A recent study by Tropberger and his group showed that transcription and active PTMs in cccDNA are reduced by the activation of an innate pathway, and that this effect can be recapitulated with a small molecule epigenetic modifying agent ([@B35]).\n\nTherapeutic Implications of Epigenetic Mechanisms in HBV Infection\n==================================================================\n\nThe most important goal in HBV research is the development of therapies to eradicate HBV infection. Considering the long half-life of the hepatocytes, the limiting factor in eliminating infection is the clearance of the cccDNA pool from the infected cells. Therefore, interfering with the epigenetic regulation of the cccDNA minichromosome is the most promising therapeutic approach. Experiments in humanized mice and cell culture demonstrated that treatment with IFN-\u03b1 induces cccDNA-bound histone hypoacetylation and the active recruitment of transcriptional co-repressors onto cccDNA ([@B2]). IFN-\u03b1 administration was also shown to reduce binding of STAT1 and STAT2 transcription factors to active cccDNA ([@B9]). Treatment with IFN and other potential cytokines that can activate the cellular response via the epigenetic modifications of cccDNA could mark the episome for selective eradication of infected cells or prevent cccDNA molecules from re-entering into nuclei after mitosis. Epigenetic alterations could also potentially alter cccDNA partitioning into daughter cells. Exploring the molecular mechanisms by which IFN-\u03b1 mediates epigenetic repression of cccDNA transcriptional activity and identifying new molecular determinants can lead to the development of a treatment that would eradicate cccDNA molecules.\n\nIn CHB infection, viral and host DNA is a host defence mechanism to suppress viral replication. Increased expression of DNMTs has been reported in CHB livers that facilitates viral genome methylation and affects protein production and viral replication ([@B14]; [@B37]). Furthermore, host DNA methylation has been shown to be the main mechanism inactivating relevant genes in HCC, suggesting a potential role of strong demethylating agents in the treatment of HCC ([@B10]; [@B34]). However, such potent demethylating treatment could lead to the reactivation of HBV replication. Confirming that methylation of non-integrated HBV genomes can regulate viral replication and cccDNA transcription leads to several significant clinical correlates. Dietary consideration may potentially be important in modulating HBV replication, as dietary deficiencies can limit the liver's ability to methylate HBV. Mouse experiments showed that dietary supplementation with folate, vitamin B12, choline and betaine could lead to host gene methylation ([@B41]). These findings indicate a potential role of methylation in the future treatment of CHB infection.\n\nConcluding Remarks\n==================\n\nIn the last decade our knowledge on epigenetic modifications in viral infection has increased dramatically. DNA methylation and histone modifications have been shown to play important roles in regulating the expression of a variety of HBV genes and viral replication. The holy grail of the future of HBV therapy is the complete elimination of cccDNA of all the infected cells in the host. The association of the cccDNA acetylation changes with viral replication and transcription shows that the dynamic acetylation and deacetylation of cccDNA-associated histones is essential in cccDNA transcription. Additionally, IFN-\u03b1 has been shown to actively decay cccDNA due to acetylation modifications but the exact mechanisms and the role of cell division in cytokine-induced cccDNA elimination remain to be determined. DNA methylation is also being increasingly recognized to play a role in regulation of HBV gene expression. Both integrated and episomal HBV DNA can be methylated in human tissues and high HBV DNA methylation associates with HCC development. Studies on cccDNA methylation have provided conflicting results mainly because the content of cccDNA in the infected hepatocytes is maintained by the *de novo* cccDNA synthesis and possibly due to the different HBV genotypes and experimental approaches applied. HBx has been shown to play a key role in the acetylation and methylation of cccDNA either directly by being recruited onto the cccDNA minichromosome and indirectly by the modulation of epigenetic-associated proteins, including DNMTs and cccDNA-bound histones. Enhancing our understanding of the epigenetic consequences of HBV-host interactions will lead to the identification of novel potential therapeutic targets for selective inhibition of cccDNA transcription and therefore complete eradication of HBV infection. Since the epigenetic processes are reversible they would also provide new molecular determinants by which host and environmental factors can regulate HBV replication and pathogenesis.\n\nConflict of Interest Statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n[^1]: Edited by: *Koichi Watashi, National Institute of Infectious Diseases, Japan*\n\n[^2]: Reviewed by: *Kohji Moriishi, University of Yamanashi, Japan; Masaya Sugiyama, National Center for Global Health and Medicine, Japan*\n\n[^3]: This article was submitted to Virology, a section of the journal Frontiers in Microbiology\n"}
+{"text": "Introduction {#tca13002-sec-0005}\n============\n\nBreast cancer is the most common type of cancer in women across the world, and the incidence is rapidly increasing in China.[1](#tca13002-bib-0001){ref-type=\"ref\"} Metastatic breast cancer (MBC) remains an incurable disease; however, an increasing number of targeted therapies have resulted in ever\u2010improving clinical outcomes. Many studies have shown that clonal evolution of MBC can arise in disease progression or following multiple lines of therapy, leading to treatment failure.[2](#tca13002-bib-0002){ref-type=\"ref\"} Thus, understanding the genomic profile of the tumor is critical for managing MBC, especially with respect to the selection of targeted therapies and when switching regimens. Tumor tissue biopsy in MBC is invasive and often inaccessible (e.g. in bone metastasis). However, as next generation sequencing (NGS) technology has advanced, noninvasive molecular profiling of MBC has become available.\n\nBlood\u2010derived circulating tumor DNA (ctDNA) is reported to be detectable in the plasma of patients with advanced malignancy including BC,[3](#tca13002-bib-0003){ref-type=\"ref\"} acting as a potential noninvasive source to characterize the genomic features of tumors.[4](#tca13002-bib-0004){ref-type=\"ref\"} Many studies on ctDNA have used digital PCR techniques to detect mutations in blood[5](#tca13002-bib-0005){ref-type=\"ref\"}, [6](#tca13002-bib-0006){ref-type=\"ref\"} these techniques are highly sensitive, but molecular profiling information of tumor tissue is still needed. With the recent development of high\u2010throughput DNA sequencing of targeted regions, we are now able to detect and track tumor\u2010specific somatic mutations in cell free DNA (cfDNA) independently.\n\nTo explore whether plasma can be used as \"liquid biopsy\" in Chinese MBC patients, we conducted a pilot study using a commercially available 1021\u2010gene panel tested on plasma samples, paired peripheral blood mononuclear cell (PBMC) samples, and accessible tumor tissue. Herein, we report the genomic profiles and druggable genomic alterations (GAs) in ctDNA from 17 patients with advanced BC during the course of their standard clinical care.\n\nMethods {#tca13002-sec-0006}\n=======\n\nPatient cohort and sample collection {#tca13002-sec-0007}\n------------------------------------\n\nThe study cohort consisted of 17 Chinese MBC patients treated at Sun Yat\u2010Sen Memorial Hospital. This was an observational, non\u2010interventional, retrospective study and was conducted in accordance with recognized ethical guidelines. Written informed consent was obtained from all participants. Patients were treated according to physicians' decisions.\n\nAll patients were diagnosed with pathologically confirmed BC. Staging investigations were performed in all patients with breast ultrasound, computed tomography (CT) and/or magnetic resonance (MR) scan and evaluated according to National Comprehensive Cancer Network guidelines. We used tumor markers, breast ultrasound, CT and/or MR to monitor disease every six months and/or when disease progressed.\n\nThe clinical characteristics of the study cohort are summarized in Table [1](#tca13002-tbl-0001){ref-type=\"table\"}. ER, PR, and HER2 status, as well as Ki67 index were assessed in a single laboratory of the Sun Yat\u2010Sen Memorial Hospital Pathology Department using standard criteria.\n\n###### \n\nThe clinical characteristics of the study cohort\n\n  Patient characteristics (*n*\u2009=\u200917)                \n  ------------------------------------ ------------ -------\n  Age at diagnosis (years)                          \n  Mean\u2009\u00b1\u2009SD                             46\u2009\u00b1\u200911.20  \n  Range                                   28--62    \n  Metastatic sites                                  \n  1                                         9       52.94\n  2                                         5       29.41\n  3                                         3       17.65\n  ER                                                \n  \u2265 10%                                     10      58.82\n  \\< 10%                                    7       41.18\n  PR                                                \n  \u2265 10%                                     3       17.65\n  \\< 10%                                    14      82.35\n  Ki67                                              \n  \u2265 14%                                     13      76.47\n  \\< 14%                                    4       23.53\n  *HER2*                                            \n  Positive                                  2       11.76\n  Negative                                  15      88.24\n  Previous therapy                                  \n  Yes                                       15      88.24\n  No                                        2       11.76\n\nSD, standard deviation.\n\nNext generation sequencing {#tca13002-sec-0008}\n--------------------------\n\nA total of 17 blood samples were collected. Blood was processed within one\u2009hour of sample collection in ethylene\u2010diamine\u2010tetraacetic acid tubes and centrifuged at 3000\u2009rpm for 10 minutes. Plasma was then transferred to new EP tubes and centrifuged at 10\u2009000 rpm to further remove cell debris, and stored at \u221280\u00b0C until DNA extraction. Genomic DNA was extracted from peripheral blood mononuclear cells to generate a reference genome to distinguish germline mutations and single nucleotide polymorphisms (SNPs) for each patient. Archival tumor tissues were also tested if accessible.\n\nTarget region capture and enrichment was conducted based on a 1021\u2010gene panel and a customized library provided by Geneplus\u2010Beijing (Beijing, China). All experimental processes were performed following the manufacturer\\'s protocol under strict quality control and assessment. All of the captured DNA fragments were amplified and pooled to obtain multiplex libraries.\n\nAll of the samples were sequenced with Illumina 2\u2009\u00d7\u200975\u2009bp paired\u2010end reads on an Illumina HiSeq 3000 instrument according to the manufacturer\\'s recommendations using the TruSeq PE Cluster Generation Kit v3 and the TruSeq SBS Kit v3 (Illumina, San Diego, CA, USA).\n\nSequence data analysis {#tca13002-sec-0009}\n----------------------\n\nAfter filtering the adaptor and low\u2010quality sequences from the raw reads, the clean data were mapped to the reference human genome (version hs37d5.fa) aligned with Burrows--Wheeler Aligner (BWA).[7](#tca13002-bib-0007){ref-type=\"ref\"} Somatic small insertions and deletions (indels) and single nucleotide variants (SNVs) were identified using The Genome Analysis Toolkit (<https://www.broadinstitute.org/gatk/>) and MuTeck,[8](#tca13002-bib-0008){ref-type=\"ref\"} and copy number variations (CNVs) were identified using Contra. PyClone[9](#tca13002-bib-0009){ref-type=\"ref\"} was employed to assess the clonal population structure of ctDNA in each patient. The clonal variant allele frequency (VAF) at each time point was analyzed based on the mean allele fraction (MAF) of mutations contained in the cluster with highest cancer cell fraction (CCF).[10](#tca13002-bib-0010){ref-type=\"ref\"}\n\nStatistical analysis {#tca13002-sec-0010}\n--------------------\n\nA log\u2010rank test was used to assess the association between detection of ctDNA and PFS. Correlations between ctDNA level and clinicopathological markers were assessed using Pearson\\'s chi square test. All statistical analyses and visualizations were performed with GraphPad Prism version 6.0 (La Jolla, CA, USA) or R version 3.4.1 with R package pheatmap, ggplot2 (R Foundation for Statistical Computing, Vienna, Austria). All *P* values are two\u2010sided.\n\nResults {#tca13002-sec-0011}\n=======\n\nClinical characteristics of the study cohort {#tca13002-sec-0012}\n--------------------------------------------\n\nSeventeen female patients were enrolled in our study. The average diagnostic age was 46\u2009years. All patients were stage IV. Two patients had primary stage IV BC and were treatment\u2010naive when their blood samples were collected; all other patients had received at least one line of therapy. Of the 17 patients, 10 were ER+/HER2\u2212, 2 were HER2+, and 5 were triple negative BC. The clinical characteristics of the study cohort are summarized in Table [1](#tca13002-tbl-0001){ref-type=\"table\"}.\n\nSomatic mutation profile of circulating tumor DNA (ctDNA) using targeted deep sequencing {#tca13002-sec-0013}\n----------------------------------------------------------------------------------------\n\nTargeted deep sequencing of cfDNA was successfully performed with blood samples collected from the 17 patients. Tumor\u2010specific mutations were identified in cfDNA from the blood samples of all patients (100%), with a median of four somatic mutations per sample (range: 1--9 mutations per sample).\n\nA total of 60 somatic mutations and 1 CNV were detected in the 17 blood samples, with a median MAF of 1.40% (range: 0.06--51.00%). *TP53* (35.29%, 6 patients), and *PIK3CA* (29.41%, 5 patients), were the most frequent mutated genes (Fig [1](#tca13002-fig-0001){ref-type=\"fig\"}), which is consistent with the mutation spectrum of primary tumors.[11](#tca13002-bib-0011){ref-type=\"ref\"} *ESR1* (17.65%) and *PTEN* (17.65%) were the third most frequently mutated genes in our study, with mutation frequencies much higher than those reported based on tumor tissue sequencing in the COSMIC database[12](#tca13002-bib-0012){ref-type=\"ref\"} and other studies (*ESR* 7%, *PTEN* 4%).[13](#tca13002-bib-0013){ref-type=\"ref\"}\n\n![The frequency and distribution of somatic mutations detected in circulating tumor DNA (ctDNA) of 17 metastatic breast cancer (MBC) patients. The clinical characteristics of 17 MBC patients (upper) and ctDNA profiles among the 17 MBC patients (lower). The mutation frequencies of each gene are shown on the right. Somatic mutation type: (![](TCA-10-807-g005.jpg \"image\")) deletion, (![](TCA-10-807-g006.jpg \"image\")) nonsense, (![](TCA-10-807-g007.jpg \"image\")) missense](TCA-10-807-g001){#tca13002-fig-0001}\n\nctDNA profile differs among breast cancer of different hormone receptor status {#tca13002-sec-0014}\n------------------------------------------------------------------------------\n\nWe also compared the mutation profiles of ER positive and negative patients. *PIK3CA* mutations were frequent across different hormone receptor status (30% in ER\u2010positive and 28.57% in ER\u2010negative patients). However, *TP53* occurred in five out of seven (71.43%) ER\u2010negative patients and only 1 out of 10 (10%) ER\u2010positive patient. All of the *ESR1* mutations were detected in ER\u2010positive patients (3 mutations in 3 patients), which is consistent with the tumor tissue sequencing results of other studies (Fig [2](#tca13002-fig-0002){ref-type=\"fig\"}).[11](#tca13002-bib-0011){ref-type=\"ref\"} In addition, we detected *ERBB2* amplification in one patient (P001), whose immunohistochemistry and fluorescence in situ hybridization results were also HER2 positive.\n\n![The distribution of somatic mutations in (**a**) ER\u2010positive and (**b**) ER\u2010negative metastatic breast cancer patients.](TCA-10-807-g002){#tca13002-fig-0002}\n\nConcordance of somatic mutations between synchronous and asynchronous tissue and plasma samples {#tca13002-sec-0015}\n-----------------------------------------------------------------------------------------------\n\nThe reliability of ctDNA sequencing has not been fully established and tumor tissue sequencing remains the golden standard. However, invasive procedures are required to procure biopsy samples of MBC and are often difficult to obtain. In our sample, archival tissue samples of five patients were accessible and sequenced (Fig [3](#tca13002-fig-0003){ref-type=\"fig\"}). In 80% (4/5) of patients, concordant mutations were found in both tissue and plasma samples. Patient P006 had primary stage IV disease, and paired tumor tissue and blood samples were collected at the same time when the primary tumor was surgically removed. In this case sequencing results of ctDNA and tumor tissue were completely concordant (Fig [3](#tca13002-fig-0003){ref-type=\"fig\"}a). Somatic mutations *TP53* ^W91\\*^ and *STK11* ^L290P^ were detected in both the patient\\'s tumor tissue and ctDNA. While in the asynchronous samples from P003, P008 and P017, the mutations identified in tissue and plasma were only partly concordant (Fig [3](#tca13002-fig-0003){ref-type=\"fig\"}b,e) or completely different (Fig [3](#tca13002-fig-0003){ref-type=\"fig\"}c). Only in one patient, P105, who had distant metastases shortly after surgery, the asynchronous tissue and plasma mutations were concordant (Fig [3](#tca13002-fig-0003){ref-type=\"fig\"}d).\n\n![Somatic mutations detected in synchronous and asynchronous tumor tissue and plasma samples. (**a**) Concordance between tissue and circulating tumor DNA (ctDNA) mutations detected in synchronous tumor tissue and plasma samples. (**b--e**) Tissue mutations and ctDNA mutation assays are not completely concordant in asynchronous tumor tissue and plasma samples.](TCA-10-807-g003){#tca13002-fig-0003}\n\nIdentifications of actionable genomic alterations from ctDNA {#tca13002-sec-0016}\n------------------------------------------------------------\n\nWe next sought to characterize druggable mutations in the study cohort. We compared somatic mutations detected in the ctDNA of our study cohort with the identified druggable mutations documented in the National Center for Biotechnology Information ClinVar database and previous reports of tissue sequencing results.[14](#tca13002-bib-0014){ref-type=\"ref\"}, [15](#tca13002-bib-0015){ref-type=\"ref\"} Analysis of ctDNA of the 17 patients revealed that 7 patients harbored a total of 8 druggable somatic mutations and 1 patient had *ERBB2* amplification (Table [2](#tca13002-tbl-0002){ref-type=\"table\"}). The most frequent druggable mutations occurred at two hotspots of the *PIK3CA* gene. One was H1047R (4 mutations in 4 patients) at exon 20 encoding the kinase domain, and the other was E542K (detected in 2 two samples of P013) at exon 9 encoding the helical domain. These two hotspot mutations were reported to activate the phosphatidylinositol\u20103 kinase/protein kinase B/ mammalian target of rapamycin (PI3K/AKT/mTOR) pathway, which diminishes the effects of hormone therapy[16](#tca13002-bib-0016){ref-type=\"ref\"} as well as trastuzumab and lapatinib treatment.[17](#tca13002-bib-0017){ref-type=\"ref\"} However, according to results from BOLERO\u20102, the beneficial effects of the mTOR inhibitor everolimus are maintained regardless of the *PIK3CA* genotype.[18](#tca13002-bib-0018){ref-type=\"ref\"} The remaining three druggable mutations occurred at the *PTEN* gene, namely, p.K144\\*, p.Q219\\*, and p.M134del, which led to the loss of PTEN activity. PTEN can inhibit activity of the PI3K/AKT/mTOR pathway, and *PTEN* gene loss leads to activation of this pathway.[17](#tca13002-bib-0017){ref-type=\"ref\"} Biomarker analyses from BOLERO\u20101 and BOLERO\u20103 trials showed that HER2\u2010positive advanced BC patients with *PTEN* loss in tumors could derive a progression\u2010free survival (PFS) benefit from everolimus.[19](#tca13002-bib-0019){ref-type=\"ref\"}, [20](#tca13002-bib-0020){ref-type=\"ref\"} Other preclinical studies have suggested that targeting mTOR may restore sensitivity to endocrine therapy in hormone receptor\u2010positive advanced BC patients.[21](#tca13002-bib-0021){ref-type=\"ref\"} Overall, druggable mutations were detected in 50% of our patient cohort. All of the druggable mutations detected in our study cohort were related to the PI3K/AKT/mTOR pathway.\n\n###### \n\nDruggable mutations detected in cfDNA of the study cohort\n\n  Patient ID   Druggable mutation   Mutation type   cHGVS             pHGVS                                            Mutant allele frequency (%)   Drug sensitivity not influenced   Resistant drugs\n  ------------ -------------------- --------------- ----------------- ------------------------------------------------ ----------------------------- --------------------------------- -----------------------\n  P001         *PIK3CA*             Missense        c.3140A\\>G        p.H1047R                                         1.60                          Everolimus                        Trastuzumab/lapatinib\n  P004         *PIK3CA*             Missense        c.3140A\\>G        p.H1047R                                         1.20                          Everolimus                        Trastuzumab/lapatinib\n  P008         *PIK3CA*             Missense        c.3140A\\>G        p.H1047R                                         7.90                          Everolimus                        Trastuzumab/lapatinib\n  P008         *PTEN*               Nonsense        c.430A\\>T         p.K144[\\*](#tca13002-note-1002){ref-type=\"fn\"}   36.50                         Everolimus                        Trastuzumab/lapatinib\n  P010         *PTEN*               Nonsense        c.655C\\>T         p.Q219[\\*](#tca13002-note-1002){ref-type=\"fn\"}   39.00                         Everolimus                        Trastuzumab/lapatinib\n  P012         *PTEN*               Deletion        c.402_404delGAT   p.M134del                                        3.4                           Everolimus                        Trastuzumab/lapatinib\n  P013         *PIK3CA*             Missense        c.1035T\\>A        p.N345K                                          3.00                          Everolimus                        Trastuzumab/lapatinib\n  P014         *PIK3CA*             Missense        c.3140A\\>G        p.H1047R                                         2.01                          Everolimus                        Trastuzumab/lapatinib\n\nrepresents Nonsense mutation (a substitution mutation resulting in a termination codon, foreshortening the translated peptide).\n\ncfDNA, circulating free DNA; cHGVS, complementary DNA human genome variations; pHGVS, protein human genome variations.\n\nctDNA level and clinical outcome in metastatic breast cancer patients {#tca13002-sec-0017}\n---------------------------------------------------------------------\n\nWe evaluated the effect of ctDNA mutations on clinical tumor burden and PFS. The VAF of mutations from the major mutated clones (the clone with greatest CCF) was used to access ctDNA levels.[10](#tca13002-bib-0010){ref-type=\"ref\"} Among 17 patients, the ctDNA level varied from 0.06% to 51% (median 2.01%). We found no significant difference between ctDNA level based on the number of metastatic sites or whether there was visceral metastasis (Fig [4](#tca13002-fig-0004){ref-type=\"fig\"}a,b), which may be a result of the limited sample size of our study cohort. However, when we further analyzed ctDNA and PFS (16 of 17 patients had available follow\u2010up data), we found patients with higher than median ctDNA levels (\\> 2.01%) had significantly shorter PFS, and serum tumor marker CA15\u20103 cannot predict PFS of MBC patients (Fig [4](#tca13002-fig-0004){ref-type=\"fig\"}c,d). The median PFS of the high ctDNA level group was less than half the PFS of the low ctDNA level group (138 vs. 386\u2009days; log\u2010rank *P*\u2009=\u20090.02).\n\n![The association of circulating tumor DNA (ctDNA) level between clinical tumor burden and progression\u2010free survival (PFS). (**a**,**b**) The association between ctDNA level and (**a**) visceral metastasis and (**b**) number of metastatic sites. (**c**) A higher ctDNA level was associated with shorter PFS. (![](TCA-10-807-g008.jpg \"image\")) low ctDNA level and (![](TCA-10-807-g009.jpg \"image\")) high ctDNA level. (d) The CA 15\u20103 level did not predict PFS in metastatic breast cancer patients (![](TCA-10-807-g010.jpg \"image\")) CA153\u2010low and (![](TCA-10-807-g011.jpg \"image\")) CA153\u2010high.](TCA-10-807-g004){#tca13002-fig-0004}\n\nDiscussion {#tca13002-sec-0018}\n==========\n\nOur study characterized the mutation profile in a cohort of Chinese MBC patients. We have shown the feasibility of analyzing ctDNA to characterize genomic alterations in MBC. The high frequency (50%) of druggable mutations among the patients suggests that ctDNA is potentially of great clinical utility in the management of MBC. We also showed that a high ctDNA level is associated with poor PFS in MBC patients.\n\nIn our study cohort, ctDNA showed high sensitivity (100%) in plasma derived from MBC patients. In previous reports, ctDNA was detectable in \\> 75% of patients with advanced malignancies.[22](#tca13002-bib-0022){ref-type=\"ref\"} On one hand, the high tumor burden in our MBC patient cohort (47.06% patients had more than one metastatic site) contributed to the high sensitivity. On the other hand, the rapid development of sequencing techniques in recent years has greatly improved the sequencing depth and coverage of cfDNA assays. The commercial panel we applied in our study covered 1021 genes with a target region of 1.1\u2009Mb, which also contributed to the high sensitivity. In 2013, Dawson *et al*. successfully detected ctDNA in 29 out of 30 women (97%) in whom somatic genomic alterations were pre\u2010identified in tumor tissue.[5](#tca13002-bib-0005){ref-type=\"ref\"} However, in our study we achieved high sensitivity without prior knowledge of tissue mutation using a broad\u2010coverage 1021\u2010gene panel. Although only five tissue samples were available because of the difficulties of sampling, we found that tissue mutation and ctDNA mutation is highly concordant in synchronous paired tissue and plasma samples. In contrast, tissue mutation and ctDNA mutation detected in asynchronous tissue and plasma samples were only partly matched or even completely discordant. These findings suggest that ctDNA can reflect real\u2010time tumor mutation profiles and shows potential tumor clonal evolution during disease progression or under the pressure of treatment.\n\nAnother important strength of our study is the high frequency of druggable mutations independently detected in plasma samples without the need for biopsy. In our study cohort, 41.18% of patients (*n*\u2009=\u20097) had druggable mutations; if we include *ERBB2* amplification, actionable genomic alterations were detected in the blood samples of 47.06% of patients (*n*\u2009=\u20098). Druggable genomic alterations in tumor tissue have been investigated by many studies across different cancer types.[23](#tca13002-bib-0023){ref-type=\"ref\"}, [24](#tca13002-bib-0024){ref-type=\"ref\"} In 2017, a large\u2010scale study evaluated druggable mutations in 10 000 metastatic cancer tissue samples of different cancer types.[25](#tca13002-bib-0025){ref-type=\"ref\"} The study revealed that BC ranked third in terms of the prevalence of actionable mutations at 63%, which indicates the importance of genomic profiling in MBC. Our study further shows the utility of ctDNA analysis as a noninvasive method to depict genomic alterations in MBC. The high frequency of druggable mutations, mainly located in *PIK3CA* and *PTEN*, suggests that the PI3K/AKT/mTOR pathway plays an important role in MBC. The PI3K/AKT/mTOR pathway can be targeted by the clinically available mTOR inhibitor everolimus, as shown in BOLERO\u20102. Currently, the PI3K inhibitor buparlisib has shown promising results in penetrating endocrine\u2010resistant HR+/HER2+ MBC in the phase III clinical trial, BELLE\u20102.[26](#tca13002-bib-0026){ref-type=\"ref\"} In the BELLE\u20103 trial, buparlisib plus fulvestrant also showed longer PFS compared to a fulvestrant plus placebo group in HR+/HER2\u2212 MBC.[27](#tca13002-bib-0027){ref-type=\"ref\"} Thus, monitoring the presence and dynamics of these mutations is of clinical importance. Notably, *ESR1* mutations were also detected at a high frequency in ER\u2010positive MBC patients. *ESR1* mutations are commonly detected after therapy for metastatic disease and the presence of *ESR1* mutations indicates the development of endocrine resistance, especially aromatase inhibitors.[28](#tca13002-bib-0028){ref-type=\"ref\"} Schiavon *et al*. reported that patients with *ESR1* mutations in their ctDNAs had substantially shorter PFS on subsequent aromatase inhibitor\u2010based therapy.[29](#tca13002-bib-0029){ref-type=\"ref\"} Given these results, although there is currently no targeted therapy for *ESR1* mutations, they should be carefully considered in disease management.[29](#tca13002-bib-0029){ref-type=\"ref\"}\n\nThe ctDNA level is reported to be associated with PFS in other cancer types, such as lung cancer.[10](#tca13002-bib-0010){ref-type=\"ref\"} The cfDNA tumor fraction is also reported to be associated with survival in MBC.[5](#tca13002-bib-0005){ref-type=\"ref\"}, [30](#tca13002-bib-0030){ref-type=\"ref\"} Our study confirms that a higher ctDNA level is associated with shorter PFS in a small cohort of MBC Chinese patients. Our results shed light on the potential of ctDNA as liquid biopsy to depict genomic alterations and identify druggable mutations, which can complement and substitute multiple biopsies in the management of MBC.\n\nDisclosure {#tca13002-sec-0020}\n==========\n\nNo authors report any conflict of interest.\n\nThis work was supported by a National Natural Science Foundation of China grant (No. 81702630).\n"}
+{"text": "Introduction\n============\n\nLarge-scale cooling water systems are widely used to remove heat from industrial equipment using a heat exchanger in chemical manufacturing facilities, power plants and petroleum refineries ([@B42]). Natural water from a river, lake or sea can be used ([@B36]). In some cooling water systems, the treated refinery wastewater was used ([@B27]). These water systems contain diverse species of microorganisms, organic matters and inorganic salts that enable microbial growth, leading to microbiologically influenced corrosion (MIC) and biofouling ([@B41]; [@B28]). Cooling tower systems typically have a water temperature between 25\u00b0C and 35\u00b0C that is ideal for microbes to grow ([@B29]). Microbes attach to the surfaces to form biofilms by secreting extra polymeric substances (EPS) ([@B12]). These biofilms can cause MIC and biofouling ([@B7]; [@B21]; [@B37]; [@B43]).\n\nMicrobiologically influenced corrosion is a major problem in various industrial sectors, such as water utilities, oil and gas, and power generation ([@B35]; [@B44]; [@B18]). MIC in recirculating cooling water systems causes deterioration of metallic surfaces and reduces the lifetime of the systems ([@B15]). Biofouling on the other hand reduces the heat exchanging efficiency and hinders pipe flows ([@B40]). Thus, these problematic biofilms reduce the reliability and increase the operating cost of the systems ([@B33]).\n\nBiocide dosing is a common approach to combat microorganisms in cooling water systems ([@B4]). Chlorine is a widely used oxidizing biocide due to its low cost and high efficacy ([@B9]; [@B11]). Other non-oxidizing biocides such as tetrakis hydroxymethyl phosphonium sulfate (THPS), glutaraldehyde, isothiazoline and quaternary ammonium compounds are also widely used in cooling towers ([@B32]; [@B16]). In the field, different microbes often live in biofilm communities ([@B31]). Biofilms offer sessile cells protection from harmful environmental conditions and antimicrobial agents ([@B26]). Therefore, a much higher concentration of biocide is required to treat biofilms than to treat planktonic cells due to the various defense mechanisms that biofilms use against antimicrobial agents ([@B30]). The high concentration raises operational and environmental problems. For example, chlorine at a high concentration causes equipment corrosion and is toxic to the environment after discharge ([@B38]). In a recirculating cooling water system, continuous or cyclic biocide dosing is required because biofilms always bounce back ([@B8]). The repeated biocide dosing over time may cause dosage escalation due to biocide resistance. Thus, a more effective biocide treatment with reduced biocide dosages is highly desirable.\n\n[D]{.smallcaps}-amino acids are naturally occurring chemicals. [D]{.smallcaps}-methionine ([D]{.smallcaps}-met), [D]{.smallcaps}-leucine ([D]{.smallcaps}-leu), [D]{.smallcaps}-tyrosine ([D]{.smallcaps}-tyr), and [D]{.smallcaps}-tryptophan ([D]{.smallcaps}-trp) were found to disperse *Bacillus subtilis, Pseudomonas aeruginosa*, and *Staphylococcus aureus* biofilms ([@B23]). [D]{.smallcaps}-amino acids can enhance the efficacy of some existing biocides against corrosive biofilms. Lab tests showed that 1 ppm (w/w) [D]{.smallcaps}-tyr and 100 ppm [D]{.smallcaps}-met individually enhanced the efficacy of low concentrations of THPS and ADBAC (alkyldimethylbenzylammonium chloride) biocides, respectively in the mitigation of the *Desulfovibrio vulgaris* (a sulfate reducing bacterium) biofilm on carbon steel, achieving better efficacies than higher concentrations of THPS and ADBAC ([@B45], [@B46]; [@B17]). [D]{.smallcaps}-tyr at low concentrations (2--5 ppm) were found to enhance ciprofloxacin in the mitigation of anaerobic *P. aeruginosa* biofilms by achieving better efficacies than higher concentrations of ciprofloxacin ([@B19]). It was suggested that a mixture of several [D]{.smallcaps}-amino acids was required to enhance THPS against field biofilm consortia because [D]{.smallcaps}-amino acids used individually showed limited effects ([@B26]). In this work, two mixtures of [D]{.smallcaps}-amino acids containing equimass [D]{.smallcaps}-tyr, [D]{.smallcaps}-met, [D]{.smallcaps}-trp, and [D]{.smallcaps}-leu (labeled as D4) and equimass [D]{.smallcaps}-tyr, [D]{.smallcaps}-met, [D]{.smallcaps}-trp, [D]{.smallcaps}-leu, [D]{.smallcaps}-serine ([D]{.smallcaps}-ser), [D]{.smallcaps}-threonine ([D]{.smallcaps}-thr), [D]{.smallcaps}-phenylalanine ([D]{.smallcaps}-phe), and [D]{.smallcaps}-valine ([D]{.smallcaps}-val) (labeled as D8) were evaluated as biocide enhancers for bleach (active component: NaClO), THPS, and NALCO 7330 (active components: 5-chloro-2-methyl-4-isothiazolin-3-one and 2-methyl-4-isothiazolin-3-one) against a field biofilm consortium on C1018 carbon steel coupons retrieved from a water cooling tower.\n\nMaterials and Methods {#s1}\n=====================\n\nMicrobes and Chemicals\n----------------------\n\nC1018 carbon steel coupons covered with biofilms were retrieved from a water cooling tower in a US chemical manufacturing facility after field exposure of 3 weeks (**Figure [1A](#F1){ref-type=\"fig\"}**). The coupons were shipped overnight in capped vials to minimize deterioration of the biofilms. The biofilm morphology of the consortium on a coupon surface before the removal tests was observed using an SEM (scanning electron microscopy) (Model JSM-6390, JEOL, Tokyo, Japan). The detailed procedure of the coupon preparation for the observation under SEM was reported before ([@B17]). The strip coupons (3\u2033 \u00d7 0.5\u2033 \u00d7 0.06\u2033) were submerged in 30 ml vials with the fluid collected from the tower. NALCO 7330 was provided by the chemical manufacturing facility. [D]{.smallcaps}-amino acids were purchased from Sigma--Aldrich (St. Louis, MO, United States). Other chemicals used in this study were purchased either from Fisher Scientific (Pittsburgh, PA, United States) or Sigma--Aldrich (St. Louis, MO, United States). The biofilm treatment lab tests were conducted aerobically. Before each lab test, the PBS (phosphate buffered saline) buffer solution, tweezers, test tubes, and pipette tips were autoclaved at 121\u00b0C for 20 min. [D]{.smallcaps}-amino acid solutions were sterilized with a 0.22 \u03bcm Stericup filter (Millipore, Bedford, MA, United States). All experiments were conducted at least three times for accuracy.\n\n![Experimental details: coupons with biofilms retrieved from a water cooling tower **(A)**, single-dose batch treatment (biofilms treated with different chemicals in weighing dishes) conducted in a biosafety cabinet **(B)**, and sequential treatment using bleach (biofilms treated with bleach followed by [D]{.smallcaps}-amino acids in weighing dishes) conducted in a biosafety cabinet with a cardboard box cover to avoid UV degradation **(C)**.](fmicb-08-01538-g001){#F1}\n\nEnhanced Non-oxidizing Biocide Treatment against the Biofilm Consortium\n-----------------------------------------------------------------------\n\nBefore the 3-h biofilm removal test, planktonic cells and the field fluid on the coupons were rinsed off using a pH 7.4 PBS buffer solution. Then, coupons were placed in weighing dishes with 100 ml of the PBS buffer with added treatment chemicals for 3 h in a biosafety cabinet at 25\u00b0C (**Figure [1B](#F1){ref-type=\"fig\"}**). Biofilms were treated with non-oxidizing biocides (THPS and NALCO 7330 separately) with and without a [D]{.smallcaps}-amino acid mixture. The test matrix is shown in **Table [1](#T1){ref-type=\"table\"}**.\n\n###### \n\nConditions for [D]{.smallcaps}-amino acid mixture enhancement of non-oxidizing biocides in the 3-h biofilm removal test.\n\n  Parameter            Condition\n  -------------------- -------------------------------------------------------------------------------------------------------------------------\n  Biofilm              Consortium from a water cooling tower\n  Growth time          3 weeks\n  Solution             pH 7.4 PBS buffer solution\n  Treatment method     Non-oxidizing biocides, [D]{.smallcaps}-amino acid mixtures, non-oxidizing biocide + [D]{.smallcaps}-amino acid mixture\n  Treatment duration   3-h exposure to treatment chemicals in a weighing dish\n  Temperature          25\u00b0C\n  Coupon               C1018 carbon steel\n  Assay                Sessile cell counts, CLSM images\n\nAt the end of the 3-h treatment, the coupons were removed to enumerate sessile cells using the most probable number (MPN) method. MPN test kits were purchased from Biotechnology Solutions (Houston, TX, United States). Three liquid culture media, namely the modified Postgate's B (MPB) medium for SRB (sulfate reducing bacteria), the standard bacterial nutrient broth for general heterotrophic bacteria (GHB), and the phenol red dextrose (PRD) medium for acid producing bacteria (APB), were used for MPN. After the biofilm removal test, coupons were rinsed with the PBS buffer solution to remove any loosely attached planktonic cells and treatment chemicals. The biofilm was scrapped off a coupon into a test tube with 10 ml PBS buffer using a small sterile brush. The scraped-off biofilm, the brush and the 10 ml PBS buffer were placed in the test tube. A vortex mixer was used for 30 s to suspend all the sessile cells evenly in the liquid before the liquid was used for MPN serial dilutions and incubation at 37\u00b0C. Each MPN enumeration was repeated twice for reproducibility. The commonly used *t*-test method was applied to obtain the *p*-value for statistical significance.\n\nLive and dead sessile cells in biofilms on coupons were examined under confocal laser scanning microscopy (CLSM) (Model LSM 510, Carl Zeiss, Jena, Germany). The information on the staining procedure was described in a previous work ([@B17]). The ImageJ software (National Institutes of Health, Bethesda, MD, United States) was used to quantify the live and dead sessile cells in CLSM images.\n\nSequential Treatment Using Bleach and Different [D]{.smallcaps}-amino Acid Mixtures\n-----------------------------------------------------------------------------------\n\nThe field coupons covered with biofilms (**Figure [1A](#F1){ref-type=\"fig\"}**) were treated with bleach and [D]{.smallcaps}-amino acid mixtures. The operation was conducted in the biosafety cabinet with a cardboard box cover to prevent UV degradation due to exposure to light (**Figure [1C](#F1){ref-type=\"fig\"}**). Since the bleach can react with [D]{.smallcaps}-amino acids, an abiotic chemical compatibility test was conducted first. Five ppm bleach (i.e., 5 ppm NaClO) was mixed with 10 ppm [D]{.smallcaps}-amino acid mixture in deionized water without inoculation for 3 h at 25\u00b0C. After that, the free chlorine concentration was measured using the \"SenSafe Free Chlorine Water Check\" test strips (Industrial Test Systems, Inc., Rock Hill, SC, United States).\n\nIn the 4-h biofilm removal test, bleach and a [D]{.smallcaps}-amino acid mixture were mixed to treat biofilms for 4 h. In addition to the single-dose batch treatment, a sequential treatment of bleach and a [D]{.smallcaps}-amino acid mixture was tried. During the sequential treatment, in the first 2 h, 5 ppm bleach was applied to treat a coupon. Then the coupon was retrieved and rinsed with the PBS buffer. It was then put into another weighing dish with 100 ml of the PBS buffer containing 50 ppm of a [D]{.smallcaps}-amino acid mixture for another 2 h. **Table [2](#T2){ref-type=\"table\"}** shows the test matrix. After the 4-h biofilm removal tests, the biofilms on coupons were examined under CLSM and the sessile cells on coupons were enumerated using the MPN test kits.\n\n###### \n\nConditions for [D]{.smallcaps}-amino acid mixture enhancement of bleach in the 4-h biofilm removal test.\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Parameter          Condition\n  ------------------ --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Biofilm            Consortium from a water cooling tower\n\n  Growth time        3 weeks\n\n  Treatment method   Bleach, [D]{.smallcaps}-amino acid mixtures, bleach + [D]{.smallcaps}-amino acid mixtures\n\n  Treatment time     Control: 4 h without treatment\\\n                     [D]{.smallcaps}-amino acid mixtures alone: 2 h with 50 ppm [D]{.smallcaps}-amino acid mixtures + 2 h with no treatment Biocide alone: 2 h with 5 ppm bleach + 2 h with no treatment\\\n                     Sequential treatment: first 2 h with 5 ppm bleach followed by another 2 h with 50 ppm [D]{.smallcaps}-amino acid mixtures\\\n                     Biocide mixed with [D]{.smallcaps}-amino acids: 5 ppm bleach + 50 ppm D-amino acid mixture for 4 h Biocide alone: 2 h with 5 ppm bleach + 2 h with 5 ppm bleach\n\n  Temperature        25\u00b0C\n\n  Coupon             C1018 carbon steel\n\n  Assay              Sessile cell counts, CLSM images\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nResults and Discussion\n======================\n\nBiofilm Consortium from the Water Cooling Tower\n-----------------------------------------------\n\n**Figure [2](#F2){ref-type=\"fig\"}** shows the morphology of the biofilm consortium under SEM before testing. The image reveals that it is a mixed-culture consortium with sessile cells of different shapes. The phylogenetic identification of the biofilm consortium shown in **Table [3](#T3){ref-type=\"table\"}** was provided by Ecolyse, Inc. (College Station, TX, United States). The company used bacterial tag-encoded semi-conductor sequencing with 515F-GTGCCAGCMGCCGCGGTAA and 806R-GGACTACHVGGGTWTCTAAT as primers for the analysis. Samples were amplified for semi-conductor sequencing employing a forward primer and a reverse fusion primer. The amplification products were analyzed using eGels (Life Technologies, Grand Island, NY, United States). Sequencing was performed with the Ion Torrent PGM following manufacturer protocols (Life Technologies, Grand Island, NY, United States). The metagenomics data reveal that there were many species in the biofilm consortium. *Achromobacter marplatensis, Lacibacter cauensis*, and *Sphingobacterium* sp. are APB ([@B5]; [@B39]; [@B20]). *Algorimarina* sp. can grow as SRB ([@B22]). *Arenimonas maotaiensis, Staphylococcus* sp., and *Terrimonas* sp. are GHB ([@B6]; [@B47]; [@B13]). The trait of *Alkanindiges* sp., *Flavobacterium* sp., *Hydrogenophaga* sp., and *Novosphingobium* sp. is biodegradation, referring to bacterial utilization of substrates that cannot be used by most of the other bacteria ([@B1]; [@B34]; [@B10]; [@B3]).\n\n![Scanning electron microscopy (SEM) images of the biofilm consortium on 3-week old C1018 coupons retrieved from the water cooling tower. Scale bars are 5 \u03bcm **(A)** and 50 \u03bcm **(B)**.](fmicb-08-01538-g002){#F2}\n\n###### \n\nMetabolic assignments of microbial species (% of population).\n\n  Species                        \\%\n  ------------------------------ -------\n  *Achromobacter marplatensis*   0.432\n  *Acidovorax* sp.               5.405\n  *Afipia* sp.                   0.162\n  *Algorimarina* sp.             0.432\n  *Alkanindiges* sp.             1.189\n  *Arenimonas aquatica*          0.865\n  *Arenimonas maotaiensis*       6.378\n  *Arenimonas* sp.               0.486\n  *Bdellovibrio* sp.             0.378\n  *Cellvibrio* sp.               0.162\n  *Chlamydia* sp.                0.216\n  *Cytophaga* sp.                0.108\n  *Flavobacterium* sp.           8.541\n  *Gemmatimonas aurantiaca*      0.649\n  *Haliscomenobacter* sp.        0.649\n  *Hydrogenophaga* sp.           22.11\n  *Ideonella* sp.                1.514\n  *Lacibacter cauensis*          0.757\n  *Methylophilus* sp.            0.27\n  *Niabella* sp.                 0.649\n  *Novosphingobium* sp.          1.081\n  *Ohtaekwangia* sp.             0.649\n  *Opitutus* sp.                 0.541\n  *Pedobacter daechungensis*     0.595\n  *Rheinheimera* sp.             0.324\n  *Rhizobacteria* sp.            7.405\n  *Rhodobacter* sp.              0.865\n  *Sphingobacterium* sp.         0.324\n  *Sphingomonas* sp.             0.811\n  *Staphylococcus* sp.           0.27\n  *Terrimonas* sp.               3.027\n  *Thermomonas haemolytica*      0.378\n  *Thermomonas* sp.              3.081\n  *Unclassified*                 24.81\n  *Variovorax* sp.               0.757\n  *Xanthomonas* sp.              3.73\n\nRemoval Test Using Two Non-oxidizing Biocides and [D]{.smallcaps}-amino Acid Mixtures\n-------------------------------------------------------------------------------------\n\n**Figure [3](#F3){ref-type=\"fig\"}** shows the sessile cell counts in the biofilm consortium treated with THPS and [D]{.smallcaps}-amino acid mixtures. After the 3-h biofilm removal test, the sessile cell counts on the no treatment control coupon were 7.8 \u00d7 10^4^ cells/cm^2^ APB, 7.8 \u00d7 10^3^ cells/cm^2^ SRB and 6.7 \u00d7 10^5^ cells/cm^2^ GHB, respectively. Fifty ppm D4 alone (*p* = 0.037) and D8 alone (*p* = 0.036) treatments both achieved 1-log GHB sessile cell reduction compared with the no treatment control. Compared with the no treatment control, no APB and SRB sessile cell reductions were observed with the 50 ppm D4 alone or 50 ppm D8 alone treatment. The 15 ppm THPS alone treatment only achieved 1.5-log GHB sessile cell reduction compared with the no treatment control (*p* = 0.032). The cocktail of 15 ppm THPS + 50 ppm D4 led to extra 1-log reduction of APB sessile cell count compared with the 15 ppm THPS alone treatment (*p* = 0.011). The treatment of 15 ppm THPS + 50 ppm D8 achieved extra 1.5-log reductions of APB sessile cell count (*p* = 0.008) and extra 1-log reduction of SRB (*p* = 0.005) and GHB (*p* = 0.004) in sessile cell counts, respectively compared with the 15 ppm THPS alone treatment. The outcome of the 15 ppm THPS + 50 ppm D8 treatment was similar to that of the 30 ppm THPS alone treatment. This means 50% reduction in THPS dosage was achieved. Results here indicate that D8 at the same concentration was more powerful than D4 in the enhancement of THPS against the biofilm consortium.\n\n![Sessile cell counts of the biofilm consortium after the 3-h biofilm removal test using THPS and [D]{.smallcaps}-amino acid mixtures. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g003){#F3}\n\n**Figure [4](#F4){ref-type=\"fig\"}** shows the CLSM biofilm images after different treatments using THPS. There were many live cells (green dots) on the no treatment control coupon (**Figure [4A](#F4){ref-type=\"fig\"}**). Biofilm images after the 50 ppm D4 alone treatment, and the 50 ppm D8 alone treatment are shown in **Figures [4B,C](#F4){ref-type=\"fig\"}**. The 15 ppm THPS alone treatment led to many dead cells (red dots) among live cells (**Figure [4D](#F4){ref-type=\"fig\"}**). After the 15 ppm THPS + 50 ppm D4 treatment (**Figure [4E](#F4){ref-type=\"fig\"}**), there were fewer live cells compared with the 15 ppm THPS alone treatment. After the 15 ppm THPS + 50 ppm D8 treatment (**Figure [4F](#F4){ref-type=\"fig\"}**), there were much fewer live cells. Dead cells were abundant after the 30 ppm THPS alone treatment (**Figure [4G](#F4){ref-type=\"fig\"}**). **Figure [4H](#F4){ref-type=\"fig\"}** shows the numbers of live and dead sessile cells calculated from CLSM images in **Figure [4](#F4){ref-type=\"fig\"}** using the ImageJ software. The results indicate that the mixture of [D]{.smallcaps}-amino acids alone had a very minor ability for the biofilm removal (*p* = 0.032). The cocktail of 15 ppm THPS + 50 ppm D8 achieved a better efficacy than the cocktail of 15 ppm THPS + 50 ppm D4 (*p* = 0.007). The efficacy of the cocktail of 15 ppm THPS + 50 ppm D8 was similar to that of the 30 ppm THPS alone treatment (*p* = 0.108). The CLSM data in **Figure [4](#F4){ref-type=\"fig\"}** supported the MPN data in **Figure [3](#F3){ref-type=\"fig\"}**.\n\n![Confocal laser scanning microscopy (CLSM) images of biofilms after the 3-h biofilm removal test in the pH 7.4 PBS buffer containing: **(A)** no treatment (control), **(B)** 50 ppm D4-1, **(C)** 50 ppm D8-1, **(D)** 15 ppm THPS, **(E)** 15 ppm THPS + 50 ppm D4-1, **(F)** 15 ppm THPS + 50 ppm D8-1, and **(G)** 30 ppm THPS. The ImageJ software calculated numbers of live/dead sessile cells are shown in **(H)**. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g004){#F4}\n\nIn the 3-h biofilm removal test using NALCO 7330 (**Figure [5](#F5){ref-type=\"fig\"}**), the 15 ppm NALCO 7330 alone treatment had a similar efficacy as the 15 ppm THPS alone treatment. The cocktail of 15 ppm NALCO 7330 + 50 ppm D4 led to extra 1-log reduction in GHB sessile cell count in comparison with the 15 ppm NALCO 7330 alone treatment (*p* = 0.006). The 15 ppm NALCO 7330 + 50 ppm D8 resulted in extra 1.5-log reductions of APB (*p* = 0.009) and GHB (*p* = 0.003) sessile cell counts and extra 1-log reduction of SRB (*p* = 0.008) sessile cell count in comparison with the 15 ppm NALCO 7330 alone treatment. The cocktail of 15 ppm NALCO 7330 + 50 ppm D8 achieved a similar efficacy to that achieved by the 30 ppm NALCO 7330 alone treatment. This means that 50 ppm D8 could reduce the biocide concentration by 50%. Results here demonstrate that the [D]{.smallcaps}-amino acid mixtures enhanced these non-oxidizing biocides at lower concentrations by achieving a similar efficacy as higher concentrations of non-oxidizing biocides. The CLSM images in **Figure [6](#F6){ref-type=\"fig\"}** corroborated the sessile cell count results in **Figure [5](#F5){ref-type=\"fig\"}**. With treatments of 15 ppm NALCO 7330 + 50 ppm D8 and 30 ppm NALCO 7330 alone, much fewer live cells were observed compared with those in the 15 ppm NALCO 7330 alone treatment. **Figure [6E](#F6){ref-type=\"fig\"}** shows the numbers of live and dead sessile cells calculated from CLSM images in **Figure [6](#F6){ref-type=\"fig\"}** using the ImageJ software. The cocktail of 15 ppm NALCO 7330 + 50 ppm D8 achieved a better efficacy than the cocktail of 15 ppm NALCO 7330 + 50 ppm D4 (*p* = 0.006). The efficacy for the cocktail of 15 ppm NALCO 7330 + 50 ppm D8 was similar to that for the 30 ppm NALCO 7330 alone treatment (*p* = 0.309). The enhanced non-oxidizing biocides with [D]{.smallcaps}-amino acid mixtures confirmed that a biocidal stress is necessary for [D]{.smallcaps}-amino acids to disperse the biofilm consortium ([@B26]). Results also indicated that non-oxidizing biocides (THPS and NALCO 7330) and [D]{.smallcaps}-amino acids worked synergistically.\n\n![Sessile cell counts of the biofilm consortium after the 3-h biofilm removal test using NALCO 7330 and [D]{.smallcaps}-amino acid mixtures. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g005){#F5}\n\n![Confocal laser scanning microscopy images of biofilms after the 3-h biofilm removal test in the pH 7.4 PBS buffer containing: **(A)** 15 ppm NALCO 7330, **(B)** 15 ppm NALCO 7330 + 50 ppm D4-1, **(C)** 15 ppm NALCO 7330 + 50 ppm D8-1, and **(D)** 30 ppm NALCO 7330. The ImageJ software calculated numbers of live/dead sessile cells are shown in **(E)**. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g006){#F6}\n\nBiofilm Removal Test Using Bleach and [D]{.smallcaps}-amino Acid Mixtures\n-------------------------------------------------------------------------\n\nChlorine is an oxidizing agent that can react with amino acids ([@B14]). In the compatibility test without inoculation, 50 ppm [D]{.smallcaps}-amino acid mixture (D4 or D8) were mixed with 5 ppm bleach (i.e., 5 ppm NaClO) in a dark environment for 3 h at 25\u00b0C. The free chlorine test strips can detect levels of ClO^-^ from 0 to 6 ppm, with a color range from white to dark blue. A white color means 0 ppm ClO^-^. **Figure [7](#F7){ref-type=\"fig\"}** suggests that the [D]{.smallcaps}-amino acids depleted the ClO^-^ after 3 h. Thus, bleach and [D]{.smallcaps}-amino acids should not be dosed together.\n\n![Bleach (containing 5 ppm NaClO) incubated with and without 10 ppm [D]{.smallcaps}-amino acid mixture in water at 25\u00b0C after 3 h: **(A)** bleach alone, and **(B)** bleach + 50 ppm [D]{.smallcaps}-amino acid mixture (D4 and D8 same outcome).](fmicb-08-01538-g007){#F7}\n\nAt the end of the 4-h biofilm removal test, 50 ppm D4 alone (*p* = 0.035), 50 ppm D8 alone (*p* = 0.035), and 5 ppm bleach alone (*p* = 0.029) treatments (exposure to treatment chemicals for 2 h followed by exposure to PBS buffer for another 2 h) all achieved 1 log GHB sessile cell reduction in comparison with the no treatment control (**Figure [8](#F8){ref-type=\"fig\"}**). When 5 ppm bleach and 50 ppm D4 or D8 were mixed to treat for 4 h zero-log reductions in APB, SRB, and GHB sessile cell counts were achieved in comparison with the treatment using 5 ppm bleach alone (lasting 2 h). The results confirmed that the chlorine reaction with [D]{.smallcaps}-amino acids rendered [D]{.smallcaps}-amino acids ineffective. Thus, a sequential treatment method was applied.\n\n![Sessile cell counts of the biofilm consortium after the 4-h biofilm removal test using bleach and [D]{.smallcaps}-amino acid mixtures. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g008){#F8}\n\nIn the sequential treatment, the coupon was first immersed in 100 ml of the PBS buffer containing 5 ppm bleach. After 2 h, it was retrieved and placed into 100 ml fresh PBS buffer containing a 50 ppm mixture of [D]{.smallcaps}-amino acids. The sequential treatment of 5 ppm bleach followed by 50 ppm D4 achieved extra 1-log reductions of SRB (*p* = 0.031) and GHB (*p* = 0.002) sessile cell counts in comparison with the 5 ppm bleach alone treatment. The sequential treatment of 5 ppm bleach followed by 50 ppm D8 achieved extra 1-log reductions of APB (*p* = 0.001), SRB (*p* = 0.021), and GHB (*p* = 0.001) sessile cell counts in comparison with the 5 ppm bleach alone treatment. The 10 ppm bleach alone treatment showed a similar efficacy with the sequential treatment of 5 ppm bleach followed by 50 ppm D8. This means that 50 ppm D8 could reduce the bleach dose by 50%.\n\nThe CLSM analysis in **Figure [9](#F9){ref-type=\"fig\"}** supported the sessile cell count results in **Figure [8](#F8){ref-type=\"fig\"}**. There were numerous live cells on the coupons from the no treatment control, the 50 ppm D4 alone treatment, and the 50 ppm D8 alone treatment as seen in **Figures [9A](#F9){ref-type=\"fig\"}--[C](#F9){ref-type=\"fig\"}**. Many live sessile cells also appeared in the combination treatments using bleach + a [D]{.smallcaps}-amino acid mixture (**Figures [9E,F](#F9){ref-type=\"fig\"}**). Fewer live cells appeared in sequential treatments (**Figures [9G,H](#F9){ref-type=\"fig\"}**) in comparison with the bleach alone treatment (**Figure [9D](#F9){ref-type=\"fig\"}**). The CLSM images also showed the efficacy of the sequential treatment of 5 ppm bleach followed by 50 ppm D8 (**Figure [9H](#F9){ref-type=\"fig\"}**) was close to that of the 10 ppm bleach alone treatment (**Figure [9I](#F9){ref-type=\"fig\"}**). **Figure [9J](#F9){ref-type=\"fig\"}** shows the numbers of live and dead sessile cells calculated from CLSM images in **Figure [9](#F9){ref-type=\"fig\"}**. The sequential treatments of 5 ppm bleach followed by 50 ppm D4 (*p* = 0.030) and 5 ppm bleach followed by 50 ppm D8 (*p* = 0.012) achieved a better efficacy than the 5 ppm bleach alone treatment. The efficacy for the sequential treatment of 5 ppm bleach followed by 50 ppm D8 was similar to that for the 10 ppm bleach alone treatment (*p* = 0.092).\n\n![Confocal laser scanning microscopy images of biofilms after the 4-h biofilm removal test in the pH 7.4 PBS buffer containing: **(A)** no treatment for 4 h, **(B)** 50 ppm D4-1 alone for 2 h + no treatment for 2 h, **(C)** 50 ppm D8-1 alone for 2 h + no treatment for 2 h, **(D)** 5 ppm bleach alone for 2 h + no treatment for 2 h, **(E)** 5 ppm bleach + 50 ppm D4-1 for 4 h, **(F)** 5 ppm bleach + 50 ppm D8-1 for 4 h, **(G)** 5 ppm bleach for 2 h followed by 50 ppm D4-1 alone for 2 h, **(H)** 5 ppm bleach alone for 2 h followed by 50 ppm D8-1 alone for 2 h, and **(I)** 5 ppm bleach for 2 h + another 5 ppm bleach for 2 h (total 10 ppm bleach). The ImageJ software calculated numbers of live/dead sessile cells are shown in **(J)**. Error bars represent standard deviations from 4 independent samples.](fmicb-08-01538-g009){#F9}\n\n**Figure [9](#F9){ref-type=\"fig\"}** shows that 50 ppm D8 showed a better efficacy than 50 ppm D4. This was likely because different microbial species in the biofilm consortium responded to different [D]{.smallcaps}-amino acids. Thus, a mixture of more [D]{.smallcaps}-amino acids was better ([@B26]). Several mechanisms have been suggested to explain why [D]{.smallcaps}-amino acids disperse bacterial biofilms. Researchers speculated that [D]{.smallcaps}-amino acids triggered biofilm disassembly because they replaced the [D]{.smallcaps}-alanine terminus in the peptidoglycan molecules that exist in all bacterial cell walls ([@B23]). An addition of high concentration [D]{.smallcaps}-alanine to the culture medium was found to hinder the efficacy of [D]{.smallcaps}-methionine to enhance THPS against an SRB biofilm on carbon steel ([@B46]). [D]{.smallcaps}-amino acids were also suggested to influence the remodeling of bacteria cell walls ([@B24]; [@B2]). [@B25] found that [D]{.smallcaps}-amino acids inhibited the cell growth and expression of EPS.\n\nConclusion\n==========\n\nIn this work, two different [D]{.smallcaps}-amino acid mixtures (D4 and D8) at 50 ppm enhanced two non-oxidizing biocides (THPS and NALCO 7330) against a field biofilm consortium from a water cooling tower. D8 was found to be more effective than D4 at the same concentration to enhance the biocides. Fifty ppm D8 enhanced 15 ppm THPS and 15 ppm NALCO 7330 by achieving similar efficacies as biocides at 30 ppm. A sequential treatment was tested for bleach due to its reactivity with [D]{.smallcaps}-amino acids. The sequential treatment of 5 ppm bleach followed by 50 ppm D8 achieved extra 1-log reductions in APB, SRB, and GHB sessile cell counts in comparison with the 5 ppm bleach alone treatment. The 10 ppm bleach alone treatment showed a similar efficacy with the sequential treatment of 5 ppm bleach followed by 50 ppm D8. This work demonstrated that 50 ppm D8 reduced the biocide dosages by 50% while achieving the same efficacies.\n\nAuthor Contributions\n====================\n\nConceived and designed the experiments: HA-M and TG. Performed the experiments: RJ and YL. Analyzed the data: RJ and YL. Wrote and polished the paper: RJ, YL, HA-M, and TG.\n\nConflict of Interest Statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe acknowledge the financial support from TOTAL and SABIC.\n\n[^1]: Edited by: *Sara Mar\u00eda Soto, ISGlobal, Spain*\n\n[^2]: Reviewed by: *Devendra Hiraman Dusane, The Ohio State University, United States; Osmar Nascimento Silva, Universidade Cat\u00f3lica Dom Bosco, Brazil*\n\n[^3]: ^\u2020^*These authors have contributed equally to this work.*\n\n[^4]: This article was submitted to Infectious Diseases, a section of the journal Frontiers in Microbiology\n"}
+{"text": "Background {#Sec1}\n==========\n\nSnakes of the genus *Bothrops* account for the majority of snakebites in Latin America \\[[@CR1]\\]. Several signs and symptoms follow *Bothrops* envenomation, which are associated with both local effects -- such as myonecrosis, hemorrhage, edema and dermonecrosis -- and systemic disorders -- characterized by coagulopathy, hemorrhage, hypertension, cardiovascular shock and acute renal failure \\[[@CR2]\\].\n\n*Bothrops* snake venoms trigger a typical local inflammatory response that involves edema and the subsequent mobilization of leukocytes. However, it is still not clear how *Bothrops* snake venoms elicit leukocyte recruitment, which is essential to restore tissue homeostasis and repair the injured sites. Studies on *B. asper* venom suggest that the activation of the complement system (CS) is one of the mechanisms underlying this event \\[[@CR3]\\].\n\n*Bothrops* venoms are a complex mixture of components including phospholipases A~2~, metalloproteases, serine proteases (SPs) and L-amino acid oxidases (LAAOs) that exert different pharmacological and biochemical activities \\[[@CR4]--[@CR6]\\]. There are evidences that the inflammatory response elicited by *Bothrops* toxins is mainly mediated by phospholipases A~2~ and metalloproteases \\[[@CR4], [@CR7]\\]. However, the possible participation of snake venom SPs and LAAOs in the course of the inflammatory response, including activation of the CS, should not be discarded \\[[@CR5]\\].\n\nHuman CS is composed of about 35 to 40 proteins and glycoproteins present in blood plasma or on cell surfaces \\[[@CR8]\\]. The CS performs important biological functions in the maintenance and regulation of immune and inflammatory reactions. Its proteins interact with each other in a highly regulated manner to promote inflammation and destroy invading microorganisms and foreign cells. As the CS has the potential to damage host tissues, regulatory proteins tightly control the activation and activity of this system \\[[@CR9]\\].\n\nThe complement system can be activated via three distinct pathways -- classical (CS-CP), alternative (CS-AP) or lectin (CS-LP) -- depending on the stimulus type. The CS-CP is activated when C1, the first component of the CS cascade, binds to an antibody molecule complexed with an antigen. The CS-AP is activated by spontaneous hydrolysis of the C3 component and subsequent deposition of C3 fragments on activating surfaces \\[[@CR10]\\]. The CS-LP, more recently discovered, is activated through recognition of carbohydrates on the surface of microorganisms by mannose binding lectins \\[[@CR11]\\]. The CS activation entails sequential proteolytic reactions, a process called CS cascade, which generates products with a variety of biological activities such as anaphylaxis, chemotaxis, opsonization, solubilization of immune complexes, and modulation of the immune response \\[[@CR12]\\].\n\nAlthough the effect of snake venoms on the CS activity has already been reported, much remains to be investigated, in particular with regard to the action of SPs and LAAOs \\[[@CR13]--[@CR15]\\]. Snake venom SPs act mainly on components of the coagulation cascade, which in turn leads to a hemostatic imbalance \\[[@CR16], [@CR17]\\]. Flavoxobin, a SP from *Trimeresurus flavoviridis* snake venom, specifically cleaves the human complement protein C3, thus acting as a C3 convertase enzyme \\[[@CR18]\\]. The biological effects of snake venom LAAOs usually proceed through induction of apoptosis, cytotoxicity, and inhibition or induction of both platelet aggregation and microbicidal activity \\[[@CR19]--[@CR23]\\].\n\nTo date, there are no reports on the action of the L-amino acid oxidase isolated from *B. pirajai* venom (BpirLAAO-I) on the complement system. In order to broaden the existing knowledge on the biological properties of *Bothrops* snake venom, the present study aims to assess whether *B. jararacussu* (Bjussu) crude venom and its serine protease (BjussuSP-I), as well as *B. pirajai* (Bpir) crude venom and its LAAO (BpirLAAO-I), modulate the human CS pathways.\n\nMethods {#Sec2}\n=======\n\nSnake venoms and toxins {#Sec3}\n-----------------------\n\nThe lyophilized Bpir and Bjussu crude venoms were acquired from the snake house of Bioagents Bioactive Proteins Ltd. (Batatais, SP, Brazil). BpirLAAO-I was purified by the three-step chromatographic process reported by Izidoro *et al.* \\[[@CR24]\\], which employed Sephadex G-75, Benzamidine-Sephadex, and Phenyl-Sepharose as stationary phases. The purity of the final preparation was higher than 95\u00a0%. The BpirLAAO-I enzymatic activity was determined before performing the experiments. BjussuSP-I was isolated by a three-step chromatographic process that used Sephacryl S-200, Benzamidine Sepharose, and C2/C18 as stationary phases, as described by Menaldo *et al*. \\[[@CR17]\\].\n\nAnimals {#Sec4}\n-------\n\nTwo adult female New Zealand white rabbits, weighing approximately 3.3\u00a0kg, and two adult sheep were obtained from the Central Animal Facility of the University of S\u00e3o Paulo, campus of Ribeir\u00e3o Preto (USP-RP). The rabbits were kept in the vivarium at the School of Pharmaceutical Sciences of Ribeir\u00e3o Preto (FCFRP/USP).\n\nBlood was collected from the rabbit's central ear artery or the sheep's jugular vein into an equal volume of modified Alsever's solution as anticoagulant, and further employed to assay the hemolytic activity of the CS.\n\nNormal human serum (NHS) {#Sec5}\n------------------------\n\nTen milliliter of blood from healthy volunteers of both genders, aged from 20 to 50\u00a0years, were collected in the absence of anticoagulants to obtain normal human serum (NHS). The sera were separated by centrifugation at 500\u2009\u00d7\u2009*g* for ten minutes at 4\u00a0\u00b0C, pooled, aliquoted into polyethylene tubes, and frozen at \u221280\u00a0\u00b0C. NHS pool was employed to evaluate the immunomodulatory activity of Bjussu and Bpir crude venoms and their isolated toxins on the CS activity and the CS-dependent neutrophil chemotaxis.\n\nHuman CS modulation by venoms and toxins {#Sec6}\n----------------------------------------\n\n### Preparation of sheep erythrocytes to assay the CS-CP/LP activity {#Sec7}\n\nSheep erythrocyte suspension was prepared as previously described \\[[@CR17]\\]. Briefly, the anticoagulated sheep blood was diluted in triethanolamine (TEA) buffer containing Ca^2+^ and Mg^2+^ (TEA-Ca^2+^-Mg^2+^ buffer) and centrifuged. After discarding the plasma and buffy coat, the erythrocytes were washed and suspended to a concentration of approximately 5\u00a0% (1.2\u2009\u00d7\u200910^9^ cells/mL). This suspension was mixed with an appropriate dilution of hemolysin and incubated for 15\u00a0min, at 4\u00a0\u00b0C. Final absorbance of the suspension was adjusted to an optical density ranging from 0.7 to 0.8 at 700\u00a0nm (SpectraMax\u00ae Plus spectrophotometer, Molecular Devices, USA).\n\n### Preparation of rabbit erythrocytes to assay the CS-AP activity {#Sec8}\n\nRabbit erythrocyte suspension was prepared as described \\[[@CR17]\\]. Briefly, the anticoagulated rabbit blood was filtered, diluted in an equal volume of a Ca^2+^ and Mg^2+^-chelating solution composed of TEA and EDTA (ethylenediamine tetraacetic acid), and incubated for 15\u00a0min, at 37\u00a0\u00b0C. After washing three times with TEA-Mg^2+^ buffer, the erythrocytes were suspended in modified Alsever's solution supplemented with 0.05\u00a0% sodium azide, in a volume equal to twice the original blood volume. Finally, the erythrocyte suspension was divided into aliquots and stored at 4\u00a0\u00b0C.\n\nPrior to use in the hemolytic assay, the erythrocytes were washed three times with TEA-EGTA-Mg^2+^ buffer \\[EGTA: ethylene glycol-bis (2-aminoethylether)-N,N,N\\',N\\'-tetraacetic acid\\], and suspended in this buffer. Final absorbance of the suspension was adjusted to an optical density ranging from 0.7 to 0.8 by spectrophotometric reading at 700\u00a0nm.\n\n### Measurement of the hemolytic activity of the CS using the kinetic method {#Sec9}\n\nThe residual hemolytic activity of the CS in NHS treated with either crude venoms or isolated toxins was assessed through the kinetic 96-well microassay, as previously described \\[[@CR17]\\]. Bjussu and Bpir crude venoms (3.1-120\u00a0\u03bcg/mL), and the toxins BjussuSP-I (0.6-20\u00a0\u03bcg/mL) and BpirLAAO-I (1.6-50\u00a0\u03bcg/mL) were diluted in appropriate buffers -- TEA-Ca^2+^-Mg^2+^ buffer to assay the CS-CP/LP activity or TEA-EGTA-Mg^2+^ buffer to assay the CS-AP activity -- to the concentrations indicated in parentheses. These samples were incubated with NHS in 96-well microplates for one hour, at 37\u00a0\u00b0C, in a final volume of 200\u00a0\u03bcL. Aliquots of sensitized sheep erythrocytes (CS-CP/LP) or rabbit erythrocytes (CS-AP) were added to the wells and the kinetics of hemolysis was followed by uninterrupted recording of absorbance at 700\u00a0nm for 15\u00a0min (SpectraMax Plus Microplate Reader, Molecular Devices, USA).\n\nThe time-course curve of hemolysis was used to determine the time required to lyse 50\u00a0% of the erythrocytes (t^\u00bd^), which corresponds to the time required for the absorbance to decline to half of its initial value. The t^\u00bd^ value is directly proportional to the percentage of suppression of the hemolytic activity of the CS, which was calculated for each sample concentration, as follows \\[[@CR17], [@CR25]\\]:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ Hemolysis\\ inhibition\\left(\\%\\right)=100\\hbox{-} \\frac{t^{\\frac{1}{2}}\\kern0.10em  control\\  wells\\times 100}{t^{\\frac{1}{2}}\\kern0.15em  sample\\  wells} $$\\end{document}$$\n\nThese percentages of inhibition values were fitted to a linear regression curve to determine the IC~50~ value, which is the venom or toxin concentration that inhibits 50\u00a0% of hemolysis. To obtain accurate IC~50~ values, at least one of the values of percentage of inhibition must be greater than 50\u00a0%.\n\nAssay of CS-dependent neutrophil chemotaxis {#Sec10}\n-------------------------------------------\n\n### Serum treatment {#Sec11}\n\nAliquots of NHS were treated with zymosan or venom as previously described \\[[@CR26]\\]. The amount of crude venom used in the chemotaxis assay was calculated based on the amount of venom that best inhibited the hemolytic activity of CS-CP/LP, 120\u00a0\u03bcg/mL, which corresponds to 3.3\u00a0\u03bcg of venom for each 1\u00a0\u03bcL of NHS. Briefly, Bjussu and Bpir crude venoms (400\u00a0\u03bcg) diluted in TEA-Ca^2+^-Mg^2+^ buffer (final volume of 100\u00a0\u03bcL) were incubated for 40\u00a0min, at 37\u00a0\u00b0C, with 120\u00a0\u03bcL of: NHS, heat-inactivated NHS (56\u00a0\u00b0C, 30\u00a0min), or TEA-Ca^2+^-Mg^2+^ buffer. Zymosan-activated NHS and NHS incubated with TEA-Ca^2+^-Mg^2+^ buffer were used as positive and negative controls, respectively. To inactivate residual complement, the supernatants were collected and heated to 56\u00a0\u00b0C for 30\u00a0min. Finally, the supernatants were diluted 1:5 in Hanks buffered saline solution (HBSS) to be used in the chemotaxis assay.\n\n### Human neutrophils isolation {#Sec12}\n\nHuman neutrophils were isolated from peripheral blood of healthy volunteers using the Ficoll-Hypaque discontinuous density gradient method. Histopaque-1077 was layered over Histopaque-1119, and the blood samples were processed according to the manufacturer\\'s instructions (Sigma Diagnostics, Inc., USA). After washing the cell pellets with HBSS, the concentration of the neutrophil suspension was adjusted to 2\u2009\u00d7\u200910^6^ cells/mL.\n\n### Chemotaxis assay {#Sec13}\n\nThe neutrophil chemotaxis assay was performed using a modified Boyden chamber, as previously described \\[[@CR27]\\]. Briefly, the lower chamber was filled with 200\u00a0\u03bcL of treated NHS and covered with a filter (diameter: 13\u00a0mm, pore size: 3\u00a0\u03bcm, SSWPO1300, Millipore Corp., USA). The upper compartment was filled with 300\u00a0\u03bcL of neutrophil suspension. After 30\u00a0min of incubation at 37\u00a0\u00b0C in humidified air, the filters were removed, fixed in 2-propanol, stained with Harris hematoxylin, dehydrated in 2-propanol, cleared with xylene, and mounted with Entellan\u00ae mounting media (Merck, Germany). The neutrophil migration was determined by the leading front technique, which measures the greatest distance in micrometers crossed by three cells per field with a 100\u00d7 magnification \\[[@CR28]\\]. At least ten fields per filter were examined.\n\nEthics committee approval {#Sec14}\n-------------------------\n\nThe Research Ethics Committee of FCFRP/USP approved the experimental procedures involving human cells and sera, which were registered under CEP/FCFRP n. 125/2008. The Ethics Committee on Laboratory Animal Care and Use at USP-RP approved the animal housing and handling procedures, as well as the experimental protocols involving animal cells, which were registered under CEUA n. 08.1.362.53.0/2008.\n\nStatistical analyses {#Sec15}\n--------------------\n\nExperimental data were analyzed by One-way ANOVA followed by the Bonferroni\\'s post-hoc test, with the aid of the GraphPad Prism Software (version 5.0, GraphPad Software, USA). Values of *p*\u2009\\<\u20090.05 were considered significant.\n\nResults {#Sec16}\n=======\n\nModulation of the hemolytic activity of the CS {#Sec17}\n----------------------------------------------\n\nTo assess whether Bjussu and Bpir crude venoms and the isolated toxins BjussuSP-I and BpirLAAO-I modulate the hemolytic activity of the CS, we measured the residual hemolytic activity of CS in sera treated with these samples, using the kinetic microassay. To conduct separate analysis of the sample effects on CS-CP/LP and CS-AP, we employed sheep erythrocytes suspended in TEA-Ca^2+^-Mg^2+^ buffer and rabbit erythrocytes suspended in TEA-EGTA-Mg^2+^ buffer, respectively. The obtained results are reported below, where the increase in t^\u00bd^ values means suppression of the hemolytic activity of CS.\n\n### Bjussu and Bpir crude venoms selectively suppress the hemolytic activity of the CS {#Sec18}\n\nBjussu and Bpir crude venoms augmented the time required to lyse 50\u00a0% of erythrocytes (t^\u00bd^) values (i.e. diminished the hemolytic activity) for both CS pathways studied, in a concentration-dependent manner. Such increase was statistically significant at concentrations higher than 50\u00a0\u03bcg/mL and 12.5\u00a0\u03bcg/mL for the CS-CP/LP and the CS-AP, respectively (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}-[d](#Fig1){ref-type=\"fig\"}). Bjussu and Bpir crude venoms similarly inhibited the hemolytic activity of the CS-CP/LP, yielding mean IC~50~ values of 91.2\u00a0\u03bcg/mL and 86.9\u00a0\u03bcg/mL, respectively (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}-[f](#Fig1){ref-type=\"fig\"}).Fig. 1Effect of Bjussu and Bpir crude venom on the hemolytic activity of the complement system. This figure depicts the concentration-dependent inhibitory effect of (**a**, **c**, and **e**) Bjussu and (**b**, **d**, and **f**) Bpir crude venoms on the hemolytic activity of the (**a**, **b**, **e**, and **f**) classical and (**c** and **d**) alternative pathways of the complement system. Panels **a** to **d**: Control represents normal human serum incubated with buffer alone. Data are expressed as the mean\u2009\u00b1\u2009standard deviation of the t^\u00bd^ values obtained for each venom concentration, based on three (CS-CP/LP) or two (CS-AP) independent experiments assayed in triplicate. \\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.001, or \\*\\*\\**p*\u2009\\<\u20090.0001 vs. control. Panels **e** and **f**: Linear regression graph, where the X-values represent the amount of (**e**) Bjussu and (**f**) Bpir crude venom (in \u03bcg/mL) and the Y-values represent the mean percentages of hemolytic activity inhibition. The IC~50~ values were calculated from three independent experiments. Bjussu: *Bothrops jararacussu*; Bpir: *Bothrops pirajai*; t^\u00bd^: time required to lyse 50\u00a0% of erythrocytes\n\nIt was not possible to determine the IC~50~ values for the CS-AP because the highest Bjussu and Bpir crude venom concentration tested (120\u00a0\u03bcg/mL) suppressed less than 50\u00a0% of the hemolytic activity: 14.2 and 13.6\u00a0%, respectively. At this concentration, Bjussu and Bpir crude venom reduced the hemolytic activity of the CS-CP/LP by 65.3 and 72.4\u00a0%, respectively. Therefore, Bjussu and Bpir crude venoms inhibited the hemolytic activity of the CS-CP/LP more effectively than they inhibited the hemolytic activity of the CS-AP.\n\n### The isolated toxins BjussuSP-I and BpirLAAO-I inhibit the hemolytic activity of the CS {#Sec19}\n\nIn the range of concentrations tested (0.6-20\u00a0\u03bcg/mL), BjussuSP-I did not significantly alter the t^\u00bd^ values for the CS-CP/LP (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). On the other hand, this toxin significantly augmented the t^\u00bd^ values for the CS-AP at 2.5, 10, and 20\u00a0\u03bcg/mL (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). BpirLAAO-I significantly increased the t^\u00bd^ values for the CS-CP/LP at all concentrations tested (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}), and the t^\u00bd^ values for the CS-AP at the concentrations of 3.1, 12.5, 25, and 50\u00a0\u03bcg/mL (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}).Fig. 2Effect of the toxins BjussuSP-I and BpirLAAO-I on the hemolytic activity of the complement system. This figure depicts the concentration-dependent inhibitory effect of (**a** and **c**) BjussuSP-I and (**b** and **d**) BpirLAAO-I on the hemolytic activity of the (**a** and **b**) classical and (**c** and **d**) alternative pathways of the complement system. Control represents normal human serum incubated with buffer alone. Data are expressed as the mean\u2009\u00b1\u2009standard deviation of the t^\u00bd^ values obtained for each toxin concentration, based on three (CS-CP/LP) or two (CS-AP) independent experiments assayed in triplicate. \\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.001, or \\*\\*\\**p*\u2009\\<\u20090.0001 vs. control. BjussuSP-I: serine protease isolated from *Bothrops jararacussu* crude venom; BpirLAAO-I: L-amino acid oxidase isolated from *Bothrops pirajai* crude venom; t^\u00bd^: time required to lyse 50\u00a0% of erythrocytes\n\nAlthough BjussuSP-I and BpirLAAO-I exerted concentration-dependent effects, it was not possible to determine their IC~50~ values. At the highest concentration tested, BjussuSP-I inhibited the hemolytic activity of the CS-AP by 13.4\u00a0%, while BpirLAAO-I suppressed the hemolytic activity of CS-CP/LP and CS-AP by 24.3 and 12.4\u00a0%, respectively. Therefore, compared with the crude venoms, the isolated toxins weakly inhibit the hemolytic activity of both pathways of the CS.\n\nBjussu and Bpir crude venom-treated sera induce neutrophil chemotaxis {#Sec20}\n---------------------------------------------------------------------\n\nBjussu and Bpir crude venom, the samples that most strongly inhibited the hemolytic activity of CS, were assessed for their ability to modulate the human neutrophil chemotaxis. The venoms alone, i.e. incubated with the CS-CP buffer in the absence of NHS, did not elicit neutrophil migration. The mean distances of migration induced by NHS alone (negative control) were 23.7\u00a0\u03bcm and 23.5\u00a0\u03bcm (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"} and [b](#Fig3){ref-type=\"fig\"}, respectively).Fig. 3Chemotactic response of neutrophils to serum treated with (**a**) Bjussu and (**b**) Bpir crude venom. Normal human serum (NHS) was treated with crude venom, zymosan (positive control), or TEA-Ca^2+^-Mg^2+^ buffer (negative control). Crude venoms were also incubated with TEA-Ca^2+^-Mg^2+^buffer alone or heat-inactivated NHS. Data are expressed as the mean\u2009\u00b1\u2009standard deviation of three independent experiments assayed in duplicate. \\**p*\u2009\\<\u20090.05 or \\*\\*\\**p*\u2009\\<\u20090.0001 vs. negative control. Bjussu: *Bothrops jararacussu*; Bpir: *Bothrops pirajai*\n\nThe Bjussu and Bpir crude venom-treated NHS and zymosan-treated NHS equally induced neutrophil migration (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}). The mean distances of migration induced by Bjussu and Bpir crude venom-treated NHS were 49.7\u00a0\u03bcm and 48.1\u00a0\u03bcm, respectively, and 53.5\u00a0\u03bcm (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}) and 50.4\u00a0\u03bcm (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}) for zymosan-treated NHS. Heat-inactivated NHS treated with Bjussu and Bpir crude venoms slightly stimulated neutrophil migration (28.0\u00a0\u03bcm and 26.7\u00a0\u03bcm, respectively). Thus, pre-treatment of NHS with Bjussu and Bpir crude venoms generate chemotactic factors.\n\nDiscussion {#Sec21}\n==========\n\nSnake venoms consist of complex mixtures of proteins that affect different systems in the human organism, including the CS \\[[@CR1], [@CR14], [@CR15], [@CR29]\\]. Venoms of snakes belonging to the Crotalidae and Viperidae families contain a variety of proteins that modulate the CS activity \\[[@CR30]\\]. These molecules may directly cleave CS proteins such as C2, C3, and C4 and stabilize the C3-convertase of the CS-AP, which in turn amplifies the activation cascade \\[[@CR13], [@CR31], [@CR32]\\]. Consequently, the production of the anaphylatoxins C3a and C5a and other fragments of the CS activation, such as C3b and C4b, is increased \\[[@CR13], [@CR31], [@CR32]\\]. Venoms of snakes from the Elapidae family also contain molecules capable of activating the CS. These venoms reduce the hemolytic activity of the CS and convert C3 into products whose electrophoretic mobilities are distinct from those of the native C3 molecule in human serum \\[[@CR33]\\].\n\nIn this study, treatment of NHS with Bjussu and Bpir crude venoms markedly reduced the hemolytic activity of the CS-CP/LP, which indicates that these venoms are potential modulators of this CS pathway. These findings are consistent with another study showing that the hemolytic activity of serum samples incubated for one hour at 37\u00a0\u00b0C with *B. atrox* venom fractions was reduced by more than 85\u00a0% \\[[@CR34]\\]. In contrast, Bjussu and Bpir crude venoms only weakly inhibited the hemolytic activity of the CS-AP.\n\nRegarding the isolated toxin BpirLAAO-I, this is the first study to demonstrate its negative modulatory effect on the hemolytic activity of the CS-CP/LP and CS-AP. This toxin increased the t^\u00bd^ values for both CS pathways at almost all the concentrations tested. On the other hand, BjussuSP-I selectively inhibited the hemolytic activity of the CS-AP, at least under the assessed conditions.\n\nA recent study has demonstrated that two Bpir SPs, named BpirSP41 and BpirSP27, similarly inhibit the hemolytic activity of the CS-AP, but modulate the CS-CP/LP at different levels: BpirSP41 suppress the hemolytic activity of this CS pathway more strongly than BpirSP27 \\[[@CR17]\\]. Flavoxobin, a SP isolated from *Trimeresurus flavoviridis*snake venom, activates the human CS-AP and leads to the formation of membrane attack complex and to the release of C3a and C5a. This SP was identified as a heterologous C3 convertase that cleaves C3 directly and selectively to form C3b and C3a \\[[@CR18]\\].\n\nThe aforementioned results of the present study were not able to elucidate whether the effects of Bjussu and Bpir crude venoms and the isolated toxins BjussuSP-I and BpirLAAO-I were due to: (i) a prior activation of the CS during pre-incubation of the venom with NHS; (ii) inactivation of CS components; (iii) interference in the CS activation processes; and/or (iv) interaction with CS regulators. Thus, other research strategies are required to unravel the mechanism of action of these *Bothrops* venoms and toxins on the CS.\n\nOne strategy relies on the direct or indirect measurement of possible fragments generated by the CS activation. Induction of neutrophil chemotaxis by activated serum is an indirect method to assess the formation of CS products C3a and C5a, which are potent chemoattractants \\[[@CR35]\\]. In this study, NHS treated with Bjussu and Bpir crude venoms elicited neutrophil migration to levels similar to those elicited by zymosan-treated NHS. Compared with the NHS alone, crude venoms alone did not trigger neutrophil migration. These results led us to conclude that crude venoms induced the cleavage of C3 and C5 components of the CS in NHS, generating the chemotactic fragments C3a and C5a. Previous studies have reported that serum treated with *B. asper* and *Tityus serrulatus* venoms were able to induce rat neutrophil migration. The authors have also proposed that the mechanism of action of venoms involved the generation of CS fragments with chemotactic activity \\[[@CR3], [@CR25]\\].\n\nThe production of the anaphylatoxins C3a and C5a probably plays a role in *in vivo* inflammatory processes that occur after *Bothrops* envenomation, such as edema, coagulopathies, leukocyte mobilization, generation of reactive oxygen species, and secretion of pro-inflammatory cytokines like tumor necrosis factor alpha, interleukin-6, and interleukin-1 \\[[@CR36]\\].\n\nTogether, the results of the hemolytic activity of the CS and the CS-dependent neutrophil chemotaxis demonstrate that Bjussu and Bpir crude venoms activate the CS. Our hypothesis is that the hemolytic activity decrease is due to CS pre-activation during NHS incubation with the venom. At that moment, components of the CS are consumed -- and thus generate CS fragments, including anaphylatoxins C3a and C5a -- which, in turn, diminishes the amount of CS components available to induce hemolysis. In summary, the CS pre-activation by venom reduces the residual hemolytic activity of NHS.\n\nFurthermore, the kinetic microassay employed in this study has proved to be useful for assessing the modulatory effect of snake venoms and toxins on the hemolytic activity of different CS pathways. Compared with the traditional hemolytic assay, the microassay requires smaller volumes of reagents, and shorten the assay length due to simultaneous analysis of many samples. The results reported herein contribute to a better understanding of the mechanism of action of *Bothrops* venoms and toxins.\n\nConclusion {#Sec22}\n==========\n\nBjussu and Bpir crude venoms activate the CS-CP/LP and generate the chemotactic factors C3a and C5a. As the isolated toxins BjussuSP-I and BpirLAAO-I weakly inhibited the human CS hemolytic activity, it is still necessary to perform further studies to isolate and identify the Bjussu and Bpir venom components responsible for the observed effects as well as to clarify the mechanisms responsible for their biological effects.\n\nEthics committee approval {#Sec23}\n-------------------------\n\nThe Research Ethics Committee of FCFRP/USP approved the experimental procedures involving human cells and sera, which were registered under CEP/FCFRP n. 125/2008. The Ethics Committee on Laboratory Animal Care and Use at USP-RP approved the animal housing and handling procedures, as well as the experimental protocols involving animal cells, which were registered under CEUA n. 08.1.362.53.0/2008.\n\nBjussu\n\n:   *Bothrops jararacussu*\n\nBjussuSP-I\n\n:   Serine protease from *Bothrops jararacussu*\n\nBpir\n\n:   *Bothrops pirajai*\n\nBpirLAAO-I\n\n:   L-amino acid oxidase from *Bothrops pirajai*\n\nCS\n\n:   Complement system\n\nCS-AP\n\n:   Alternative pathway of the complement system\n\nCS-CP\n\n:   Classical pathway of the complement system\n\nCS-LP\n\n:   Lectin pathway of the complement system\n\nEDTA\n\n:   Ethylenediamine tetraacetic acid\n\nEGTA\n\n:   Ethylene glycol-bis(2-aminoethylether)-N,N,N\\',N\\'-tetraacetic acid\n\nHBSS\n\n:   Hanks buffered saline solution\n\nIC~50~\n\n:   sample concentration that inhibits 50\u00a0% of hemolysis\n\nLAAO\n\n:   L-amino acid oxidase\n\nNHS\n\n:   Normal human serum\n\nSP\n\n:   Serine protease\n\nTEA\n\n:   Triethanolamine\n\nt^\u00bd^\n\n:   Time required to lyse 50\u00a0% of erythrocytes\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n\nLRA, ARR and SMB contributed equally to the experimental design and writing of this manuscript. LSPC, JCP and ACM participated in the study design. SVS, FAC and LSPC designed the experiments, raised the funds to perform the research and critically discussed the manuscript. All authors read and approved the final manuscript.\n\nThe authors would like to thank the Nucleus for Research on Animal Toxins (NAP-TOXAN-USP-- grant n. 12--125432.1.3), the State of S\u00e3o Paulo Research Foundation (FAPESP -- grant n. 2011/23236-4) and the Coordination for the Improvement of Higher Education Personnel (CAPES) for their funding of this research. Thanks are also due to T. M. Casare-Ogasawara for the technical support and to the Center for the Study of Venoms and Venomous Animals (CEVAP) of UNESP for enabling the publication of this special collection (CNPq process 469660/2014-7).\n"}
+{"text": "Introduction {#s1}\n============\n\nDigital reconstruction of neuronal morphology from light microscopy images provides a powerful technique for the analysis of neural circuits forms and for the investigation of their underlying function [@pone.0084557-Roysam1]. Since the manual reconstruction is very time-consuming, especially for large-scale neuronal analysis, a number of studies have been conducted to develop more efficient computer-assisted approaches to support neurite tracing and neuronal morphology reconstruction [@pone.0084557-Meijering1]--[@pone.0084557-Halavi1].\n\nMethods of neurite tracing roughly fall in three categories. The first category relies on the sequential presentation of serial 2D images. These methods extract the profiles of neural structures in each 2D cross-sectional plane and then connect the results in the third dimension [@pone.0084557-Fiala1]--[@pone.0084557-Myatt1]. However, this may be in trouble when local neurite segments lie parallel to the 2D plane. Resampling the original images along different directions might work but is computationally expensive and difficult to cover all neurites.\n\nThe second category is based on global image segmentation in 3D. These methods first turn the image into a binary form with certain segmentation algorithms and then extract centerlines of foreground areas with a skeletonization algorithm [@pone.0084557-Rodriguez1], [@pone.0084557-Weaver1]. Here, efficient segmentation algorithms, such as thresholding, prove successful for uniformly high quality images, although it is not always the case. Various sophisticated filters have been proposed to enhance feature structures and to improve segmentation and skeletonization. Typically, such filters are based on the analysis of eigenvalues of the Hessian [@pone.0084557-Streekstra1], Jacobian matrix [@pone.0084557-Yuan1], or steerable filters [@pone.0084557-Gonzalez1]. However, these filters require multiple scaling or orientations and are computationally expensive because they operate on the entire image. A few other algorithms, such as voxel coding [@pone.0084557-Vasilkoski1], [@pone.0084557-Chothani1] and voxel scooping [@pone.0084557-Rodriguez2], can trace multiple neurite branches directly from the grayscale images but they operate on a voxel-by-voxel basis and are more dependent on the high quality of pre-processed images.\n\nThe third category explores images only in local regions around the structures of interest. These methods usually start from a seed, which can be located automatically or manually, and trace recursively where neurite branches go in the 3D image. The direction of tracing is determined according to the distribution of local signal and background. A representative algorithm is to use a template that comprises four predefined parallel edge detectors, to determine the direction and boundary of a single branch [@pone.0084557-AlKofahi1]--[@pone.0084557-AbdulKarim1]. More robust algorithms often rely on certain mathematical or graph models. They usually define a cost or energy measure that is based on local image features and curve regularity; then convert the problem into searching for an optimal path or minimum spanning tree between given crucial points [@pone.0084557-Meijering2]--[@pone.0084557-Xie1]. If initial structures or prior knowledge can be obtained, it then becomes optional to fit a snake-based or active contour model to the image data [@pone.0084557-Schmitt1]--[@pone.0084557-Wang1]. A potential problem is that these local tracing algorithms may fail at branch points or crossover regions because of a lack of global information. Thus, a separate branch merger or segment join step is needed to trace whole tree-like structure [@pone.0084557-AlKofahi3]--[@pone.0084557-Choromanska1].\n\nTo reconstruct 3D neuronal morphology, the neurite tracing can be coupled with radius estimation at each node. The rayburst sampling proposed by Wearne and Rodriguez can be used for this [@pone.0084557-Wearne1], [@pone.0084557-Rodriguez3]. It samples the image data in multiple directions using a pre-computed array of unit vectors from a given node. The process of sampling is defined as the simultaneous casting of rays in multiple directions. Each ray grows from a given node until a specified exit criterion is met, which returns a length representing the forward sampling distance. In the original algorithm, a 2D rayburst is run in the cross-sectional plane for each node, and the diameter is computed from the ray lengths. It is worth noticing that to reconstruct the whole dendritic tree, it is necessary to extract the centerline of branches in advance.\n\nThe methods mentioned above are all helpful for the reconstruction of neuronal morphology from light microscopy images. However, these automated reconstructions often contain a substantial number of false segments or short branches, requiring post-correction with significant human effort [@pone.0084557-Luisi1], [@pone.0084557-Peng4]. Thus, to achieve a reasonable balance between fast speed and high accuracy, a rapid reconstruction system incorporating automatic tracing and manual editing is preferable.\n\nIn this paper, we present a practical method for the tracing and reconstruction of 3D neuronal morphology from light microscopy images and implement it in an interactive visualization-assisted freeware system named flNeuronTool. We extend the original rayburst sampling algorithm to a marching fashion, which starts from a single or a few initial seed points and marches recursively forward along neurite branches to trace and perform reconstruction on the whole tree-like structure. We validate and evaluate the method using synthetic data and real datasets from the Digital Reconstruction of Axonal and Dendritic Morphology (DIADEM) challenge [@pone.0084557-Brown1], [@pone.0084557-Liu1]. Then, the system is applied to mouse brain images acquired with the Micro-Optical Sectioning Tomography (MOST) system [@pone.0084557-Li1]--[@pone.0084557-Yang1], to reconstruct single neurons and local neural circuits.\n\nMethods {#s2}\n=======\n\nNeurite Tracing Strategy {#s2a}\n------------------------\n\nAt the typical resolution of light microscope images, each segment of a neurite has a smooth tubular shape, which can be approximated as a generic cylinder with a certain radius [@pone.0084557-Ascoli1]. Moreover, the diameter and direction of a neurite in a local area never changes abruptly, which means that two adjacent segments have a similar radius and their centerlines often form a modest angle. Thus, it is possible to predict the radius and direction of the next segment according to the current presentation [@pone.0084557-Mayerich1].\n\nWe trace individual neurite using a prediction and refinement strategy. As shown in [**Figure 1**](#pone-0084557-g001){ref-type=\"fig\"}, we recursively move a node along the centerline of a neurite. For any given position, the node has three properties: the location C is a point on the centerline; the direction v is a unit vector that represents the tangent of the centerline; and the radius r is the minimum distance from point C to the structure boundary, which is estimated at the orthogonal plane of the direction v. Two consecutive nodes are defined as the parent and child nodes, respectively. If appropriate properties of the parent node have been obtained, we can predict the child node position along direction v. Then, using local images, the properties of the child node can be refined. Recursively, we can trace the whole tree starting from one single or a few of seed points.\n\n![Neurite tracing strategy.\\\nA node is recursively moved along the centerline of a neurite, to reconstruct neurite segments as a sequence of cylinders. At any given position, the node has a location C, a direction v and a radius r.](pone.0084557.g001){#pone-0084557-g001}\n\nCenterline Extraction and Branch Detection {#s2b}\n------------------------------------------\n\nWe use augmented rayburst sampling to detect centerlines and branches in the prediction step. Inspecting 2D rayburst sampling at a given node, we find that the longest rays are always close to the direction of the axis that represents the potential centerline of the local neurite segment, while the shortest rays in the orthogonal direction represent the local radius. Consequently, when extending all of the rays from the position of the current node until their length goes over a specified threshold (approximately 2.0\u223c4.0 times the local radius), those rays that are far away from the axis will reach or exceed the boundary of the neurite, and the corresponding samplings will been terminated. The remaining rays that are still inside the structure will be separated into a few clusters according to their directions, which represent potential branches. Specifically, they are together as a single cluster when the branches are absent. The length threshold of each ray is defined as the sampling distance, which is only related to the local radius but independent of the absolute size of the structural structure. Moreover, to avoid backward tracing, we perform sampling only in the forward direction. As shown in [**Figure 2A**](#pone-0084557-g002){ref-type=\"fig\"}, actual sampling could be restricted to a half circle, whose origin locates at the current node position C, and the half angle ray represents the node direction v.\n\n![Centerline extraction and branch detection using hemisphere sampling.\\\n(A) 2D illustration of the centerline and branch prediction with improved rayburst sampling. (B) 3D illustration of a vector of the hemisphere sampling core in a spherical coordinate system. (C) The sampling state on the hemispherical surface is projected to the orthogonal plane, in which the sampling rays that have the same inclination angle \u03b8 form a circle, and nonzero value samplings (bright blue) suggest possible branches. (D) In the plane, each circle can be mapped into a circumscribed square, and each local center detected (orange) represents a neurite branch.](pone.0084557.g002){#pone-0084557-g002}\n\nThe improved rayburst sampling is introduced for 3D. Here, to detect branches, sampling is restricted to a hemisphere; thus, it is crucial to generate an approximately uniformly distributed sampling core on the hemisphere surface. The random generation or polygon subdivision used in the original algorithm cannot achieve this goal. As shown in [**Figure 2B**](#pone-0084557-g002){ref-type=\"fig\"}, we solve this problem in a spherical coordinate system for which the origin O is located at the current node position C, the z axis coincides with the current node direction v and the orthogonal plane parallels the cross-section that covers the local radius r. For any vector OP with length \u03c1, its endpoint coordinate is P(\u03c1, \u03b8, \u03c6), in which \u03b8 and \u03c6 represent the inclination and azimuthal angles, respectively. We divide the whole hemisphere into M (M\u200a=\u200a7) stacks according to the inclination angle \u03b8, with equal intervals. For each given \u03b8, we divide corresponding small circles into N slices according to the azimuthal angle \u03c6, with equal intervals. The angular distribution is expressed as where the function max selects the maximum value from two given numbers, and the reason for setting the variable N values will be explained below. To sample the original data, vector OP should be converted into the Cartesian form, in which the coordinate of the endpoint P(x, y, z) is expressed as\n\nIt is convenient to set \u03c1 to one for automated normalization. Nevertheless, in some case, where neurite segments are really thin and comparatively straight, one can speed up the sampling in the z direction. Thus, an anisotropic sampling core is optional and has no influence on the sampling form, while the conversion to the z axis is expressed aswhere \u03b5 is defined as the anisotropy coefficient (approximately 1.0\u223c1.1). Then, these vectors could be normalized and used as a per-computed sampling core.\n\nTo perform 3D hemispherical sampling at a given node, we rotate the sampling core to correct the orientation using a transformation matrix, by which we match the current node direction v with the z axis. Then, each vector in the sampling core is iteratively extended with a pre-specified step-length from the current position C until the length goes over the pre-specified sampling distance. Those rays who reach the boundary of the local structure should be terminated early. One could use an adaptive hysteresis threshold to detect the boundary along a sampling ray (supporting information, **Text S1** in **[File S1](#pone.0084557.s001){ref-type=\"supplementary-material\"}**). In the end of the sampling setup, the sampling state whose ray endpoint is still in the inside of the structure is recorded as a value of one, and the remaining is recorded as a value of zero. A possible set of sampling state on the hemispherical surface is projected to the orthogonal plane as shown in [**Figure 2C**](#pone-0084557-g002){ref-type=\"fig\"}, in which the sampling rays that have the same inclination angle \u03b8 form a circle, and nonzero value samplings suggest possible branches.\n\nIn the plane, each circle can be mapped into a circumscribed square with slight deformation, as shown in [**Figure 2D**](#pone-0084557-g002){ref-type=\"fig\"}. Consequently, these sampling squares with different inclination angles \u03b8 and azimuthal angles \u03c6 form a uniform grid. Besides, the number of samplings in the innermost square is only one, and the difference between the adjacent squares is 8, which is the reason for setting the variable N values for the sampling core generation, as described above. In this uniform grid, nonzero value samplings are separated into a few areas for which the local centers could be detected by multiform ways, such as a simple and effective distance transform [@pone.0084557-Sonka1]. In general, each local center represents a potential neurite branch that can be mapped back to the hemisphere surface and is expressed in both spherical coordinates and the Cartesian form. Then, a new node can be located in the branch as the child whose parent is the current node. The distance between it and the parent node is defined as the location distance, which is approximately equal to a diameter of the parent node.\n\nCenterline Refinement and Radius Estimation {#s2c}\n-------------------------------------------\n\nThe location of a node that is predicted by hemispherical sampling may be a bit off the real centerline of local neurite because of holes or other artifacts in light microscopy images, and the radius of predicted child node equals the parent node radius, which should be refined. We run an improved 2D rayburst sampling to perform this step.\n\nAs shown in [**Figure 3A**](#pone-0084557-g003){ref-type=\"fig\"}, a set of unit vectors that are parallel to the sampling plane and uniformly distributed in a circle are used as a sampling core. The sampling starts at the predicted node C. Each ray iteratively extends with a pre-specified step-length from the origin point until the structure boundary is reached. We record the intersection of a ray and the boundary as P. Ideally, the cross-section is a generic ellipse, of which a centroid is simply the average location of all intersections. For real images, to gradually approach to the real position, we iteratively perform multiple sampling in which each calculated centroid is regarded as the origin for the next sampling. This process is terminated when the Euclidean distance \u03b4 between two calculated positions from successive iterations is less than a pre-specified threshold, which is a radius-related value (the coefficient \u03b5\u200a=\u200a0.05). For the last iteration, the lengths of all of the rays are also calculated, of which the mean is regarded to be the estimated radius r of the current node. This process is expressed as where N (N\u200a=\u200a32) is the number of the 2D sampling rays, P~i,\\ j~ represents the intersection position of the ray numbered j and the boundary in the iteration numbered i, and the function dst calculates the Euclidean distance between two given points.\n\n![Centerline refinement and local radius estimation using iterative rayburst sampling.\\\n(A) Centroid refinement and local radius estimation using iterative 2D rayburst samplings in which each calculated centroid is regarded as the origin for the next sampling. (B) 3D illustration of centerline refinement. The sampling described above is iterated in a plane that is orthogonal to the current centerline direction.](pone.0084557.g003){#pone-0084557-g003}\n\nThe segment from the parent node to the current node represents the local centerline. However, a potential problem is that it could no longer be orthogonal to the reconstructed circle because the sampling origin used to refine the centerline and estimate the radius has shifted during the iterations. There is a space angle \u03b4 between the initial direction that was predicted by hemispherical sampling and the refined direction, which suggests a calculation error of the centerline, as shown in [**Figure 3B**](#pone-0084557-g003){ref-type=\"fig\"}. To improve the precision of the centerline and radius refinements, we iterate the step described above in a new plane that is orthogonal to the current direction until the angle between the last direction and the prior direction falls under a pre-specified threshold (\u03b4\u200a=\u200a\u03c0/36), which result in an appropriate precision while preventing unnecessary iteration. Then, the cross-section of neurite is reconstructed as a circle with a radius r that locates at the position C and is approximately orthogonal to the refined centerline.\n\nReconstruction of Neuronal Morphology {#s2d}\n-------------------------------------\n\nThe algorithm described above is implemented in a freeware system named flNeuronTool, which is programmed in C++ and potentially supports multiple platforms. This system allows users to reconstruct and proofread neuronal morphologies in a cooperative 3D interactive visualization-assisted environment (supporting information, **Text S2** in **[File S1](#pone.0084557.s001){ref-type=\"supplementary-material\"}**). The source code and binaries are freely available at <http://sourceforge.net/projects/flneurontool/>.\n\nTo run the tracing algorithm, an initial node has to be created as the seed point. We select one single or a few points that are inside of a neuron and are close to the root of the tree-like structure, and pushed them into a queue as seed points. This can be automatically done, but we do prefer a manual way in a 3D interactive environment, which can significantly reduce false positive errors. At each selected location, two hemisphere samplings that have opposite directions are carried out to detection branches. The rest of tracing process is recursive prediction and refinement as described above. In every recursion, generated nodes are all pushed in the queue, which will be orderly popped as new seed points. When all seed points are run out and the sampling end, one could select a new seed point to reconstruct remaining branches of the tree-like structure. This process iterates until the expected reconstruction is achieved.\n\nDue to imperfect images or excessive seed points, the reconstruction will have a number of redundancy paths and short branches. Here, redundancy paths cover the same neurite, in which each pair corresponding nodes have a similar direction and radius, and cover a more or less overlap area. One could find and merge all such node pairs from the root to endpoints based on whether the Cartesian distance between them is less than any one of them radius [@pone.0084557-Peng3]. The centroid of each pair nodes is calculated to produce a final node location, while radius is averaged. As additional gains, this method uses redundancy paths to generate a more complete and better reconstruction. The length of a branch is defined as the path length that from the last bifurcation to its endpoint. One could remove short branches based on their absolute length and their lengths relative to the local radius [@pone.0084557-Rodriguez2].\n\nExperiments and Results {#s3}\n=======================\n\nParameters Selection {#s3a}\n--------------------\n\nTo validate and evaluate the proposed method, we tested it on synthetic data and real datasets from the DIADEM challenge. Then, we applied the system to mouse brain images acquired with the MOST system to reconstruct single neurons and local neural circuits. All of the experiments were performed on an ordinary computer (Intel Core Duo 2.6 GHz CPU, NVIDIA GeForce 9800 GPU, 4GB RAM, Windows 7). [**Table 1**](#pone-0084557-t001){ref-type=\"table\"} summarizes the sampling parameters for these experiments, in which most parameters remain constant across all experiments, while other parameters need to be adapted for each dataset to produce optimal results.\n\n10.1371/journal.pone.0084557.t001\n\n###### Sampling parameters selection.\n\n![](pone.0084557.t001){#pone-0084557-t001-1}\n\n  Parameters                          Value                                                                                                           Notes\n  ----------------------------- ------------------ -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  3D sampling distance           2.0\u223c4.0 (radius)                    Each ray is terminated when its length is over the threshold. A value too big or too small will lead to the branch detection lost or redundancy. Default value is 3.0.\n  Anisotropy coefficient             1.0\u223c1.1                                                          An anisotropic sampling core can speed up the sampling in the axial direction. Default value is 1.0.\n  Inclination division number           7                                  This value determines the number of 3D sampling rays. The larger value will create a more delicate sampling core, but will require more calculated amount.\n  Sampling step-length             1.0 (voxel)      Each ray is iteratively extending with the step-length from the origin until the exit criterion is satisfied. For anisotropic images, a super sampling should be applied in the low resolution direction.\n  Location distance              1.5\u223c2.5 (radius)                    The distance between the child and the parent node affects the density of nodes in the reconstruction. Default value is 2.0 but no more than the 3D sampling distance.\n  2D sampling ray number                32                                                                                    These rays are used to refine centerline and radius.\n  Iteration exit distance         0.05 (radius)                                                 Centroid refinement is terminated when the distance between two successive locations is less than the threshold.\n  Iteration exit angle                 \u03c0/36                                                     Centerline refinement is terminated when the angle between two successive directions is less than the threshold.\n\nValidation on Synthetic Data {#s3b}\n----------------------------\n\nFirst, we inspected the performance of the branch detection using synthetic tree-like structure data. The original image and the corresponding truth structure were created by VascuSynth [@pone.0084557-Hamarneh1] and were downloaded from website (<http://vascusynth.cs.sfu.ca/>), as shown in [**Figure 4A**](#pone-0084557-g004){ref-type=\"fig\"}. To find the optimal value of the 3D sampling distance for the branch detection, we traced the tree in the image with different sampling distance, which ranges from 1.5 to 4.5 with an increment of 0.5 times the local radius (supporting information, **Text S3** in **[File S1](#pone.0084557.s001){ref-type=\"supplementary-material\"}**). For each tracing, we measured the total length of correct branches and the number of correct bifurcations, and reported the proportions of them to true values. Because the test was run directly on the image without any additional pre-processing or manual editing and used only one seed point that was located at the root of the tree by one mouse click, we could report the computation time for each tracing, as shown in [**Figure 4B**](#pone-0084557-g004){ref-type=\"fig\"}. From the trend in these measurements, it is found that a sampling distance too big or too small would lead to the branch detection lost or redundancy, and the optimal value is about 3.0. Considering the difference of images, one may set this value between 2.0 and 4.0. In general, the value of length is always more than the value of the bifurcations and has smaller change, for which lost branches are relatively short.\n\n![Validation of the branch detection on tree-like structure data.\\\n(A) The original image (voxel size 101\u00d7101\u00d7101) and truth structure (branch length 3325 voxels, bifurcation number 199). (B) The performance analysis for the tracing of the tree structure in the original image with different sampling distance. For each tracing, only one seed point is used. The total length of correct branches and the number of correct bifurcations are measured, and the proportions of them to truth values are reported, as well as the computation time in seconds. (C) The highest Gaussian noised image (SD 30) and the corresponding automated reconstruction (branch length 2036 voxels, bifurcation number 122). (D) The robustness analysis for the tracing of the tree structure in different level Gauss noised images with the same sampling distance (3.0 times local radius).](pone.0084557.g004){#pone-0084557-g004}\n\nThen, we validated the robustness of the branch detection. We used ImageJ (<http://rsb.info.nih.gov/ij/>) to generate a set of images based on the original tree-like structure data with additive Gaussian noise, whose standard deviation (SD) ranges from 5 to 30 with an increment of 5. [**Figure 4C**](#pone-0084557-g004){ref-type=\"fig\"} shows the highest noised image and the corresponding automated reconstruction. For each image, the value of the sampling distance was set to 3.0. Once again, we reported the length of branches, the number of bifurcations and the computation time, as shown in [**Figure 4D**](#pone-0084557-g004){ref-type=\"fig\"}. It is obvious that the tree-like structure becomes more ambiguous when the noise is higher, making the tracing more difficult. However, considering that there is only one seed point for each tracing, the result is acceptable. Overall, low-level noise could lead to more short branches, which must be removed expending additional time. High-level noise tends to cause the early termination of tracing, which is helpful for reducing false positive errors.\n\nEvaluation with DIADEM Dataset {#s3c}\n------------------------------\n\nWe evaluated the proposed method with two real datasets from the DIADEM challenge. Specifically, the two datasets were drosophila olfactory projection (OP) [@pone.0084557-Jefferis1] and mouse neuromuscular projection fibers (NM) [@pone.0084557-Lu2]. The image stacks and their gold standard reconstructions were downloaded from website (<http://www.diademchallenge.org/>).\n\nThe first test was run on 9 image stacks of dataset OP. For each image stack, we traced three times with different sampling distances that range from 2.5 to 3.5, from which the one that had longest path was selected as final reconstruction. The final reconstruction of the first stack is shown in [**Figure 5A**](#pone-0084557-g005){ref-type=\"fig\"}, in which the reconstruction is shifted slightly relative to the voxel data to improve the visibility. Local detail of the reconstruction and the corresponding gold standard are displayed in different colors in the inset, which shows a very small difference between them, especially at several potential topological errors that are indicated with circles. To evaluate the completeness of the reconstructions, we measured their path length, which was reported as the proportion of the length to the gold standard. To quantify the accuracy of automated reconstructions, we compared them to the gold standards by using the DIADEM metric [@pone.0084557-Gillette1]. Let G be the number of nodes in the gold standard, T be the number of nodes in our reconstruction, M be the missing nodes (false negatives) reported by the DIADEM metric, and E be the extra nodes (false positive) in test reconstruction, we define the recall and precision as\n\n![Evaluation with the DIADEM dataset OP.\\\n(A) The original data and the automated reconstruction of the first image stack (voxel size 512\u00d7512\u00d760), which is shifted slightly relative to the voxel data to improve the visibility. Local detail of the reconstruction (orange) and gold standard (cyan) are displayed in the inset, in which potential errors are circled (red). (B) The quantitative measurements of nine image stacks. The DIADEM score, recall and precision are report based on DIADEM metric measurements compare with the gold standards. Specifically, the path length and DIADEM score are approximately 0.95\u00b10.03 (mean \u00b1 SD) and 0.81\u00b10.05.](pone.0084557.g005){#pone-0084557-g005}\n\nThe measurements are shown in [**Figure 5B**](#pone-0084557-g005){ref-type=\"fig\"}. Specifically, the path length and DIADEM score are approximately 0.95\u00b10.03 (mean \u00b1 SD) and 0.81\u00b10.05, respectively, which indicate that the quality of the reconstructions is generally high. In these automated reconstructions, typical errors such as branch breakings and topological connection mistakes could be easily proofread and corrected in the editing mode of flNeuronTool, which requires approximately 5\u223c10 minutes of user intervention per stack.\n\nThe second test was run on dataset NM subset 2. As an example, we reconstructed a piece of volumetric data of 4 image stacks that have been integrated, in which 15 axon fibers start from one of the stacks and continue through the remaining stacks, as shown in [**Figure 6A**](#pone-0084557-g006){ref-type=\"fig\"}. It should be noted that we could not reconstruct the entirety of the axon fibers at a high confidence level because of the absence of appropriate pre-processing. Focusing on the local detail, potential errors in the reconstruction include branch breaking from the high unevenness of the voxel intensity, one of which is indicated with a circle. Because of the differences in data dimensionality and tracing root, the DIADEM metric could not be used as expected to evaluate the quality of the reconstruction. Instead, we reconstructed a single image stack, and compared the result with the reconstruction from NeuronStudio. As shown in [**Figure 6B**](#pone-0084557-g006){ref-type=\"fig\"}, flNeuronTool can create an acceptable automated reconstruction without additional manual editing. In contrast, as shown in [**Figure 6C**](#pone-0084557-g006){ref-type=\"fig\"}, in the reconstruction from NeuronStudio, there are a few errors such as branch breaking or steal that are marked with circles. One possible reason of these differences is that the hemisphere sampling is more or less equivalent to smooth the original image, which could suppress the influence of individual voxels. Indeed, NeuronStudio could create a high-quality reconstruction with appropriate parameters or a small amount of manual editing. However, in our experience, the proofread in 2D projection planes that is adopted by NeuronStudio is not trivial, while the 3D interactive editing adopted by flNeuronTool is more convenient.\n\n![Evaluation with the DIADEM dataset NM subset 2.\\\n(A) An example of 4 image stacks (voxel size 1851\u00d7632\u00d7125) and the reconstruction, in which an error is circled. (B) An image stack (voxel size 512\u00d7512\u00d757) and the reconstruction from flNeuronTool, which is an acceptable result without additional manual editing. (C) The automated reconstruction of the same stack from NeuronStudio, in which potential errors are circled.](pone.0084557.g006){#pone-0084557-g006}\n\nApplication to MOST Image {#s3d}\n-------------------------\n\nWe applied the proposed system to mouse brain images acquired with the MOST system to reconstruct single neurons and local neural circuits. We used two datasets, which are from a green fluorescence labeled [@pone.0084557-Gong1], [@pone.0084557-Yang1] and a modified Golgi-Cox stained [@pone.0084557-Li1], [@pone.0084557-Zhang2] whole mouse brain and provide micron-resolution tomography images at the neurite level. For these original images, pre-processing is necessary, which strongly depends on the details of the neuron labeling. We applied flNeuronTool to pre-processed datasets to reconstruct a few single neurons and local neural circuits.\n\nThe first subset of images was labeled using green fluorescence, which has a 3D resolution of 0.5 \u00b5m\u00d70.5 \u00b5m\u00d72.0 \u00b5m. An example for the reconstruction of a pyramidal neuron is shown in [**Figure 7A**](#pone-0084557-g007){ref-type=\"fig\"}, in which the reconstruction was slightly shifted to improve the visibility. To evaluate the accuracy of the reconstruction, we compared it with the reference reconstruction that was created using commercial software Amira (<http://www.amira.com/>). A part of the details of them is displayed in different colors in [**Figure 7B**](#pone-0084557-g007){ref-type=\"fig\"}, and a few of the differences between them are indicated with circles. In general, the result is acceptable. Specifically, potential errors in the reconstruction such as branch contraction could be due to uneven labeling. The proportion of total length and number of bifurcation of the reconstruction to the reference are 0.96 and 0.94, respectively. The time to achieve the final reconstruction using flNeuronTool is less than 10% of the manual method, which is very attractive for large-scale neuronal image analysis.\n\n![Application the proposed system to MOST fluorescence images.\\\n(A) An example of the original images (voxel size 1728\u00d71088\u00d7176) and the reconstruction of a pyramidal neuron. (B) Detailed comparison of the reconstruction (orange) with the reference reconstruction from Amira (cyan), in which potential errors are circled (red). The proportion of total length and number of bifurcation of the reconstruction to the reference are 0.96 and 0.94, respectively.](pone.0084557.g007){#pone-0084557-g007}\n\nThe second subset of images was stained with the modified Golgi-Cox method, which has a 3D resolution of 0.3 \u00b5m\u00d70.3 \u00b5m\u00d71.0 \u00b5m. We expected to reconstruct local neural circuits from these images, because the density of neurons in this dataset is suitable. However, with our experience, it is very difficult to achieve this goal even by hand because of the complexity of the neuronal morphologies and local connections and the uneven staining. These images exhibit variable contrast, and there is frequent loss in the continuity of neuron. Hence, the best strategy is to reconstruct neuron only when the confidence level is relatively high. An example of the pre-processed images is shown in [**Figure 8A**](#pone-0084557-g008){ref-type=\"fig\"}. The reconstruction from flNeuronTool is shown in [**Figure 8B**](#pone-0084557-g008){ref-type=\"fig\"}, which contains 10 neurons, starting from the somas in the same volumetric subset, as well as 73 separate neurite fragments that have a length of more than 50 voxels. To illustrate the completeness of the neuronal morphologies, we used a sphere to represent a soma location. It is reasonable that the reconstruction contains a few errors, although we have spent approximately an hour on proofreading and correcting them.\n\n![Application the proposed system to MOST Golgi images.\\\n(A) An example of the pre-processed images (voxel size 450\u00d7450\u00d7345). (B) The reconstruction that containing 10 neurons with soma (sphere) and a few separate neurite fragments.](pone.0084557.g008){#pone-0084557-g008}\n\nDiscussion {#s4}\n==========\n\nWe presented an innovative method for tracing and reconstruction of 3D neuronal morphology from light microscopy images. We validated and evaluated the proposed method using synthetic data and the DIADEM datasets, and applied the system to MOST images. The main contribution of this paper is the augmented rayburst sampling algorithm. The original rayburst sampling has been extended to a marching fashion, which is capable of using only local information to perform neurite centerline extraction, branch detection, centroid refinement and radius estimation. The advantages of the method are summarized, as follows. First, all the tracing and reconstruction steps are achieved in a single step, in which the centerline extraction or the extra radius estimation is unnecessary. Second, this method only needs to sample a part of points rather than processing all neuronal voxels one by one. Consequently, the improvement of the reconstruction speed becomes possible, while the influence of random noises is reasonably suppressed so that appropriate computational accuracy could be guaranteed. Third, the sampling is only related to the local radius but not depend on the absolute size of the structural structure, thus, it has the potential to deal with images of different spatial resolutions.\n\nIt is worth mentioning that we implemented the method in the freeware system flNeuronTool. The system incorporates automatic tracing and manual editing of neuron reconstruction into a cooperative 3D interactive visualization-assisted environment, which is a powerful tool for analysis of complex neuronal images. Experimental results showed that it provides an effective means of minimizing the human effort that is required to improve the quality of the final reconstruction. In general, the system can be faster than manual reconstruction and be more accurate than fully automatic tracing.\n\nOne limitation of the proposed method is its poor performance on beaded neurites. For these images, we think that those graph model-based algorithms may be a better choice [@pone.0084557-Peng1]. We also found that it would be unreliable or unnecessary to reconstruct the exact morphology for some very fine dendrites which is about one voxel wide in light microscopy images. In this case, the actual application may only need to trace the centerline of neurites, and those snake-based algorithms may be more appropriate [@pone.0084557-Wang1]. Besides, some computationally expensive but more robust algorithms could be resorted to. For example, it has been shown that the optimal path is useful for automated branch mergers or segment joins [@pone.0084557-Zhao1]. In addition, for a real large-scale dataset, such as the MOST Golgi images that are stained highly unevenly and exhibit variable contrast, a practical pre-processing platform that produces more high quality images for automated tracing and reconstruction is appreciated.\n\nIn summary, the proposed system provides an efficient means for the rapid reconstruction of 3D neuronal morphology from light microscopy images, which achieves a reasonable balance between fast speed and acceptable accuracy. This system is very promising for interactive applications of large-scale neuronal image analysis, such as the reconstruction of single neurons and local neural circuits from MOST images.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Supporting Information for Methods, Experiments and Results.** Text S1, Sampling boundary detection along a sampling ray. Text S2, flNeuronTool: neuron reconstruction system. Text S3, Validation of the reliability of the radius estimation.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nWe thank Shangbin Chen and Tingwei Quan for critical reading of the manuscript, and Guozhi Zhang for his help with language editing. The anonymous reviewers are also acknowledged for the constructive comments.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: HG SZ QL. Performed the experiments: XM. Analyzed the data: XM QL. Contributed reagents/materials/analysis tools: AL JW CY WD. Wrote the paper: XM QL.\n"}
+{"text": "Why was the Agincourt Health and Socio-Demographic Surveillance System set up?\n==============================================================================\n\nThe Agincourt Health and Socio-Demographic Surveillance System (HDSS), located in rural northeast South Africa close to the Mozambique border, provides the foundation for the Rural Public Health and Health Transitions Research Unit of the Medical Research Council (MRC) and University of the Witwatersrand, South Africa (the MRC/Wits-Agincourt Unit). Its origins lie in the university's 'Health Systems Development Unit' that in the early 1990s focused on district health systems development, sub-district health centre networks and referral systems and training of clinically oriented primary health care nurses.[@dys115-B1; @dys115-B2] This was done to guide critical re-organization of the country's health system alongside democratic political change and the dismantling of South Africa's ubiquitous 'Bantustan' system, which had spawned duplicate and inefficient health departments concurrently with systematically marginalizing the rural poor. However, efforts were seriously hampered by the lack of reliable population-based information for programme planning and resource allocation.\n\nInfluenced by Sidney and Emily Kark's vision of 'community-oriented primary care', as well as visits to health and demographic surveillance sites in Bangladesh (Matlab) and Senegal (Niakhar), we established the Agincourt HDSS to address this information gap. In 1992, a baseline census was conducted in 20 contiguous villages chosen for their rural living conditions, limited access to public sector services, underperforming primary care clinics and communities of Mozambican refugees displaced by the civil war.[@dys115-B3; @dys115-B4]\n\nThree phases followed the baseline census: (i) decentralized health systems development that provided a prototype for national policy in response to limited experience in delivering rural services (1993--97),[@dys115-B1]^,^[@dys115-B5] (ii) reorientation to an interdisciplinary health and population research initiative to better understand the dynamics of health, population and social transitions and address serious weaknesses in the rural evidence base (1998--2002)[@dys115-B3; @dys115-B4]^,^[@dys115-B6] and (iii) an established university and MRC-linked field research and training programme supporting multiple investigations into the causes and consequences of critical findings from the HDSS (2004 onwards) ([www.wits.ac.za/academic/health/publichealth/agincourt/](www.wits.ac.za/academic/health/publichealth/agincourt/)).[@dys115-B7; @dys115-B8]\n\nThe Agincourt HDSS was a founding member of the International Network for the Demographic Evaluation of Populations and Their Health (INDEPTH) ([www.indepth-network.org](www.indepth-network.org)) and provides leadership to INDEPTH multi-centre initiatives in adult health and ageing[@dys115-B9] and migration and health.[@dys115-B10]\n\nWhat does it cover now?\n=======================\n\nThe Agincourt HDSS constitutes a platform for research programmes that elucidate causal pathways and test interventions across the life course. [Figure 1](#dys115-F1){ref-type=\"fig\"} outlines the organizational framework, indicating major research themes and links between them. Critical questions relate to (i) the dynamics of rapidly evolving health, population and social transitions---including inequalities between individuals and communities and social and biological explanations, (ii) determinants of vulnerability and resilience along the life course and (iii) implications for policy, programmes and services. Figure 1Research themes supported by the Agincourt HDSS, South Africa\n\nEfforts have been made to deepen observational work, extend a portfolio of intervention research and enhance capacity of the platform to support research training with PhD, post-doctoral and selected masters projects nested within established research areas.\n\nDeepening observational work\n----------------------------\n\nSeveral cohorts are nested within the population under surveillance and generally focus on subgroups at different stages along the life course ([Table 1](#dys115-T1){ref-type=\"table\"}). Table 1Features of cohorts nested within the Agincourt HDSSCohort descriptionSampleCohort nameAimSizeTypeAssessments to date Frequency of follow-upInclusion criteriaSampling unit**OBSERVATIONAL COHORTS***Ntshembo* ('Hope')To measure metabolic disease risk in adolescent girls600ClosedCohort recruited in 2007 Follow-up 2009 and 2012Boys and girls aged 7/8, 11/12 and 14/15 yearsIndividual*Ha Nakekela* ('We care')To measure HIV and non-communicable disease prevalence and their risk factors7428ClosedBaseline in 2010/2011 Repeat planned for 2013Men and women 15 years and older Permanent residentsIndividual*INDEPTH-SAGE* (Survey on global AGEing and adult health)To assess health status, well-being and health seeking behaviour of older persons4509OpenBaseline in 2006 Follow-up in 2010 Repeat planned in 2013Population 50 years and older Permanent residentsIndividual*SEEDS* (Study of the Epidemiology of Epilepsy)To understand the burden of epilepsy, pattern of seizures, and excess mortality in persons with epilepsy310ClosedNine rounds completed Follow-up every 3 monthsAll ages diagnosed with epilepsy following screening in 2008Individual*SUCSES* (Sustainability in Communal Socio-Ecological Systems)To examine (i) household income, resource use, response to shocks on household livelihoods, (ii) how livelihood capital migration and use of natural resources shape household resilience and (iii) how poverty influences child nutrition590 individuals and their householdsClosedRound 1---2010 Round 2---2011 Round 3---2012Sample stratified by gender and age Four individuals selected per age/gender group in nine villagesIndividual (households interviewed)**INTERVENTION COHORTS***Kulani* ('Grow strong')Cluster randomized trial to test school-based interventions to promote social and emotional wellbeing of children 10--12 years988 individuals in 10 schoolsClosedCohort recruited in 2009 Post-intervention follow-up end-2010 Next follow-up in 2013Boys and girls in grades 5 and 6Schools*Swa Koteka* ('We can')Randomized trial to determine whether young women who receive cash transfers conditional on school attendance have a lower incidence of HIV and HSV2 infection over time2900ClosedCohort recruited in 2011/2012 Annual follow-up for 3 yearsYoung women aged 13--20 years, in school grades 8--11, one girl per householdIndividual\n\nThe HDSS contributes to evaluation of national policy at population, household and individual levels. Examples include the following: introduction of the Rotarix\u00ae and pneumococcal conjugate vaccines into the Expanded Programme on Immunization, impact of social support grants (old age pension and child support grant) on the health and well-being of grant recipients and other household members, and the population impact of decentralized delivery of highly active anti-retroviral therapy (HAART) through public and private health systems.\n\nIn addition, a portfolio of work examines household responses to shocks and stresses and the resulting pathways influencing child and adult health and well-being. This includes the care and support roles of older women, intra- and inter-household social connections, use of natural resources and diverse migration and livelihood strategies.\n\nExtending a portfolio of intervention research\n----------------------------------------------\n\nOngoing trials target critical problems affecting the health and well-being of children and adolescents. The *Kulani* ('strength') cluster randomized trial evaluated a primary school-based intervention, developed by a non-governmental organization and conducted by the Provincial Department of Education, to improve psychological and educational outcomes of children aged 10--12 years (in partnership with the University of Oxford, UK, and Soul City Institute for Health and Development Communication, South Africa). The *Swa Koteka* ('we can') multi-level HIV prevention trial aims to reduce HIV transmission in adolescent girls by encouraging girls to remain in high school through a conditional cash transfer (individual randomization) and by influencing gender norms through community mobilization focused on men (village cluster randomization) (with the Universities of North Carolina and San Franciso, USA, and the Wits Reproductive Health and HIV Institute and Sonke Gender Justice, South Africa). The forthcoming *Ntshembo* trial ('hope') aims to reduce the intergenerational transfer of metabolic disease risk through community health worker-delivered interventions to pre-pregnant adolescent girls, which are reinforced during pregnancy and infancy. Formative studies to-date address growth and nutrition, physical activity, body image preferences, beliefs and practices regarding pregnancy, delivery and infant feeding, community food vendors and adolescent health services (with the MRC/Wits Developmental Pathways for Health Research Unit, South Africa and the Universities of Cambridge and Oxford, UK, Ume\u00e5, Sweden, and North Carolina, USA).\n\nEnhancing capacity of the Agincourt research platform\n-----------------------------------------------------\n\nEfforts to ensure rigour and extend analytic possibilities include application of automated measurement techniques to cause-of-death estimation by verbal autopsy,[@dys115-B11; @dys115-B12; @dys115-B13] full 'reconciliation' of in- and out-migrations, follow-up of migrants who depart the study area and recording of extra-household social connections. Linkage of individual records in the HDSS with those from sub-district clinics is based on conventional identifiers (name, date of birth, village, ID, cell number and other household members); fingerprint matching was used to validate these variables in correctly matching clinic patients with their HDSS records.\n\nWhere is the HDSS area?\n=======================\n\nMaps in [Figure 2](#dys115-F2){ref-type=\"fig\"} indicate the location of the Agincourt HDSS in northeast South Africa close to the border with Mozambique ([Figure 2](#dys115-F2){ref-type=\"fig\"}a), the boundary of the study site abutting on the Kruger National Park conservation area ([Figure 2](#dys115-F2){ref-type=\"fig\"}b) and the villages and health and education facilities within the site ([Figure 2](#dys115-F2){ref-type=\"fig\"}c). Figure 2(a) Location of Agincourt HDSS and sub-district within South Africa, near the Mozambique border. (b) Boundary of the Agincourt HDSS study site adjacent to the Kruger National Park, South Africa. (c) Agincourt HDSS study site and sub-district indicating positions of villages and health and education facilities\n\nThe Agincourt HDSS covers an area of 420 km^2^ comprising a sub-district of 27 villages with traditional and elected leadership. Since the democratic transition in 1994, infrastructure development has proceeded but at a rate below expectations: electricity is available in all villages, but the cost is too high for many households; few gravel roads have been tarred within the sub-district; a dam was constructed nearby, but to-date, there is no piped water to dwellings, and sanitation is rudimentary. Every village has a primary school and most a high school; however, the quality of education remains poor.[@dys115-B14] Although almost all children enrol, educational progress is often delayed with few post-secondary opportunities.\n\nThe area is dry in winter (from May to October), with soil more suited to game farming than agriculture. Households generally purchase maize and other foods, supplementing this with home-grown crops and collection of wild foods.[@dys115-B15] South Africa's non-contributory social grant system is a vital source of household income, notably the old age pension[@dys115-B16] and child support grant.[@dys115-B17; @dys115-B18] There are two health centres and six clinics within the sub-district, with three district hospitals 25--60 km away.\n\nWho is covered by the HDSS, and how often have they been followed up?\n=====================================================================\n\nAt baseline in 1992, 57 600 people were recorded in 8900 households in 20 villages[@dys115-B3]; by 2006, the population had increased to \u223c70 000 people in 11 700 households.[@dys115-B8] This increase is partly due to Mozambican in-migrants overlooked in the baseline survey and to a new settlement established as part of the post-*apartheid* government's *Reconstruction and Development Programme*. In 2007, the study area was extended to include the catchment area of a new privately supported community health centre established to provide HIV treatment before public sector roll-out of HAART. By mid-2011, the population under surveillance comprised 90 000 people residing in 16 000 households in 27 villages ([Table 2](#dys115-T2){ref-type=\"table\"}). Households are self-defined as 'people who eat from the same pot of food'. Given sustained high levels of temporary labour migration in southern Africa, we included temporary migrants residing for \\<6 months per year who retain close ties with their rural homes in the HDSS. There have been 17 census and vital event update rounds conducted strictly annually since 2000. Participation is virtually complete, with only two households refusing to participate in 2011. Table 2Features of the Agincourt HDSS population, South AfricaFeature at 1 July 2011No. or %Total population\u2003\u2003\u2003\u2003de facto population (permanent residents)66 876\u2003\u2003\u2003\u2003de jure population (permanent residents + temporary migrants resident \\<6 months/year)90 036Sex ratio (male:female)\u2003\u2003\u2003\u2003de facto population[^a^](#dys115-TF1){ref-type=\"table-fn\"}0.77\u2003\u2003\u2003\u2003de jure population0.92\u2003\u2003\u2003\u2003At birth1.01Percentage under 5 years\u2003\u2003\u2003\u2003de facto population13.74\u2003\u2003\u2003\u2003de jure population11.81Percentage under 15 years\u2003\u2003\u2003\u2003de facto population40.49\u2003\u2003\u2003\u2003de jure population33.59Percentage 65 years and over\u2003\u2003\u2003\u2003de facto population5.95\u2003\u2003\u2003\u2003de jure population4.65Mean household size (based on de jure population)\u2003\u2003\u2003\u2003Total households16 121\u2003\u2003\u2003\u2003Mean no. of members5.58Percentage temporary migrants (no. of temporary migrants/de jure population)[^b^](#dys115-TF2){ref-type=\"table-fn\"}\u2003\u2003\u2003\u2003Male32.67\u2003\u2003\u2003\u2003Female19.31Basic vital statistics199420042009Life expectancy at birth---females72.759.464.4Life expectancy at birth---males68.253.355.7Infant mortality rate per 1000 live births24.739.039.1Under-five mortality rate per 1000 live births---females37.950.751.0Under-five mortality rate per 1000 live births---males38.271.545.0Adult mortality rate---females 15--59 years164.8456.6382.4Adult mortality rate---males 15--59 years276.1539.6505.9Total fertility rate3.712.612.34Death registrationPercent registeredPeriodChildren 0--4 yearsAdults 20 +1992--9615.260.91997--200124.785.52002--0627.893.22007--1045.896.1[^1][^2]\n\nThe LINC (Learning, Information dissemination and Networking with Community) office, responsible for community liaison, has built a 20-year relationship with study communities and their leaders based on mutual trust and respect. Key points of engagement include regular 'community entry' and feedback. Community entry ranges from meetings with civic and traditional leaders to public meetings depending on the nature and sensitivity of research. Annual feedback of HDSS and project findings is to open village meetings; local service providers---health, education, social services and municipality---generally participate, creating a platform for information sharing between these institutions and the community. Selected findings are presented to service providers in the sub-district, local government and provinces.\n\nWhat has been measured, and how have the HDSS databases been constructed?\n=========================================================================\n\nThe primary instrument is a rigorous annual update of resident status and vital events involving every member of the sub-district ([Table 3](#dys115-T3){ref-type=\"table\"}). Well-supervised fieldworkers visit each household and, following verbal consent, interview the most knowledgeable respondent. They verify existing data and record all new events experienced by each household member---pregnancy outcomes, deaths and in- and out-migrations[@dys115-B8]; this is supplemented by a maternity history for in-migrant women aged 15--54 years. Update rounds involve four teams comprising a supervisor and eight fieldworkers (one dedicated to migration reconciliation); teams work from Geographic Information System-based maps that list every dwelling. Quality control measures include fieldworker self-checks, cross-checks and supervisor random checks; forms are then sent to a dedicated quality checker before data entry. Table 3Variables collected during annual resident status and vital events update, Agincourt HDSSMain data itemSpecific information**Household roster**\u2003\u2003\u2003\u2003Village number; dwelling number\u2003\u2003\u2003\u2003Name, surname and genderRecorded for each individual\u2003\u2003\u2003\u2003Date of birthNoted if estimate\u2003\u2003\u2003\u2003Mother's identification and locationVital status and where she lives\u2003\u2003\u2003\u2003Father's identification and locationVital status and where he lives\u2003\u2003\u2003\u2003Relationship of individual to household head\u2003\u2003\u2003\u2003Nationality/refugee statusIf Mozambican, when arrived in sub-district\u2003\u2003\u2003\u2003Months resident in last yearNumber of months resided in rural household\u2003Residence statusMigrant (\\<6 months in area over past year)Permanent (\\>6 months in area over past year)Visitor (not member of household)\u2003\u2003\u2003\u2003Education statusHighest level completed\u2003\u2003\u2003\u2003Pregnancy statusCurrently pregnant or not; expected delivery month\u2003\u2003\u2003\u2003Type of grantNon-contributory social grants received (includes old age, child support)\u2003\u2003\u2003\u2003National ID number**Pregnancy outcome**\u2003\u2003\u2003\u2003Antenatal clinic attendanceNumber of visits\u2003\u2003\u2003\u2003Contraceptive use before/after pregnancy\u2003\u2003\u2003\u2003DeliveryDate, location, name hospital, birth attendant and complications\u2003\u2003\u2003\u2003OutcomeLive birth, stillbirth, abortion and multiple births\u2003\u2003\u2003\u2003Duration pregnancy\u2003\u2003\u2003\u2003InfantGender, birthweight, breastfeeding and birth registration**Death**\u2003\u2003\u2003\u2003Date of deathNoted if estimate\u2003\u2003\u2003\u2003Location of deathHome, clinic, health centre, hospital (+name) and accident site\u2003\u2003\u2003\u2003Maternal deathDeath during pregnancy or delivery or within 42 days\u2003\u2003\u2003\u2003Death registration**Migration**\u2003\u2003\u2003\u2003Details of in- or out-migrantsName, national ID\u2003\u2003\u2003\u2003Move dateNoted if estimate\u2003\u2003\u2003\u2003Place migrated from and to\u2003\u2003\u2003\u2003Main reason for migration\u2003\u2003\u2003\u2003Sector of work for job-related moves**Maternity History**\u2003\u2003\u2003\u2003Full childbirth history of all womenInformation on each child not listed in household roster**Union Status**\u2003\u2003\u2003\u2003Record of all marriages/unionsPartners' details, union start/end dates, civil registration, traditional/civil ceremony, reason for union ending\n\nUsing a locally validated instrument, a dedicated verbal autopsy team (supervisor and four specially trained fieldworkers) interview the closest caregiver of the deceased to establish the probable cause of death.[@dys115-B19] The interview is conducted 1--11 months after a death and then reviewed independently by two medical doctors who assign probable underlying, immediate and contributory causes-of-death. Equivalent diagnoses are accepted as the probable cause-of-death. When diagnoses differ, the physicians confer in an effort to reach a consensus; when this is not possible, the verbal autopsy is reviewed by a third physician blind to the others' assessments.\n\nAdditional modules and 'status observations' gather individual and household data pertinent to the scientific programme ([Table 4](#dys115-T4){ref-type=\"table\"}).[@dys115-B8] These generate information at regular, but less frequent, intervals (e.g. health care utilization, food security and labour participation) or provide limited information on the entire population to inform clinical studies. Table 4Add-on modules and status observations included in update rounds, Agincourt HDSS, 1992--2012![](dys115i1.jpg)\n\nHow have HDSS databases been constructed?\n-----------------------------------------\n\nThe database is designed to store and manipulate data describing the interrelated life histories of all individuals and their households. Data are stored in a relational database, with a schema closely following that of the 'Reference Demographic Surveillance System Data Model'.[@dys115-B20] The database is implemented on the Microsoft SQL Server relational database management platform and hosted in the field site. Integrity constraints, range and type checks and triggers on specific tables check data being entered in a manner that will ensure consistency with existing data. Data are entered using a custom-built data entry frontend and are extracted using custom-written SQL queries. Routine extractions are performed to update basic indicators, whereas tailored data extractions support specific analyses.\n\nKey findings and publications\n=============================\n\nRural South Africa is in the midst of multiple interrelated transitions that have led to marked changes in population structure over two decades ([Figure 3](#dys115-F3){ref-type=\"fig\"}). Fuelled by fast-declining fertility and by the HIV/AIDS epidemic causing an increase in child mortality, the mid-1990s to mid-2000s saw rapid narrowing of the pyramid base.[@dys115-B8] By 2011, this pattern was reversing, partly reflecting the population effects of prevention-of-mother-to-child transmission programmes and stabilizing of fertility trends. Spatial patterns of mortality reflect inequalities between former Mozambican refugees and South African host communities.[@dys115-B21; @dys115-B22] Despite a recent slight upswing, fertility remains at near replacement level.[@dys115-B23] Figure 3Population pyramids of the Agincourt HDSS population: 1994, 2006 and 2011. (a) de jure population, 1 July 1994; population = 66 405. (b) de jure population, 1 July 2006; population = 72 715. (c) de jure population, 1 July 2011; population = 90 036\n\nThere is evidence of increased cardiometabolic disease risk across the life course ([Table 5](#dys115-T5){ref-type=\"table\"}). In children, we find early stunting (one-third of 1--year-olds) and adolescent overweight and obesity (20--25% in older girls) in the same socio-geographic population.[@dys115-B24] Central obesity increased from \u223c15% of girls during puberty to 35% at the end of puberty, indicating elevated risk for metabolic disease that is associated with higher socio-economic status.[@dys115-B24]^,^[@dys115-B25] In adults, evidence for a cardiovascular disease transition is clear despite the massive increase in deaths from AIDS and tuberculosis.[@dys115-B26] High blood pressure and obesity in middle-aged women are at unprecedented levels, fostered by changes in lifestyle, diet and occupation.[@dys115-B27] Table 5Indicators of cardiometabolic risk across the life course 2002--07, Agincourt sub-district, South Africa![](dys115i2.jpg)[^3][^4][^5][^6][^7][^8][^9][^10]\n\nIn rural settings, primary care management of non-communicable disease and associated risk factors is limited,[@dys115-B32] and chronic infectious disease is dominating service development. The rising mortality and risk for non-communicable diseases, notably among older women,[@dys115-B33] alongside pervasive HIV/AIDS mitigated by expanding HAART availability, argue for integrated approaches to community-oriented provision of long-term care.[@dys115-B26] Although women experience greater longevity, they consistently report a poorer quality of life.[@dys115-B34]\n\nMore rural women are migrating for work than ever before.[@dys115-B35] Mortality, mainly from AIDS and tuberculosis, is highest among recently returned migrants of both sexes, imposing high demands on local health services and communities.[@dys115-B36] Older women play key roles supporting child care and schooling while having to meet health care and funeral costs.[@dys115-B17]^,^[@dys115-B37] Pressure on the 'near-old'---women aged \\<60 years and not yet eligible for a pension---can be considerable,[@dys115-B17]^,^[@dys115-B38] whereas self-reported health and quality of life in pension-eligible older women are markedly improved.[@dys115-B39] Infant and child survival are profoundly affected by a mother's death,[@dys115-B22]^,^[@dys115-B40] as is child mobility,[@dys115-B41] and fostering by women pensioners is clearly advantageous.[@dys115-B16; @dys115-B17] Food security among poorer households remains precarious, with harvesting of natural resources acting as an important buffer against 'shocks' such as death of a breadwinner.[@dys115-B42] Among the poorest households, reliance on natural resources is high, no matter what the specific cause of adult death.[@dys115-B28] See the Agincourt website for a listing of publications: [www.wits.ac.za/academic/health/publichealth/agincourt/](www.wits.ac.za/academic/health/publichealth/agincourt/).\n\nFuture analysis plans\n=====================\n\nOngoing work seeks to deepen understanding of health, population and social transitions, their effects at stages along the life course and relationships with livelihood strategies. Analyses derived largely from the HDSS database will highlight inequalities and vulnerable sub-groups to better inform rural health and development policy.\n\nCurrent mortality analyses include trends in adult and child mortality by sex, socio-economic status and cause-of-death; infant survival both before and after a mother's death; the impact of household structure and socio-economic status on child health and survival; and case definitions and estimates of the maternal mortality ratio. Investigation of the fertility transition includes analyses of fertility decline and stall, and patterns of marital and premarital fertility. We are also comparing fertility patterns between the local South African population and former Mozambican refugees. Temporary or circular migration is pervasive in southern Africa. We are examining the association of different migration patterns with fertility outcomes and with mortality trends and causes of death, particularly HIV and non-communicable diseases. Work on the association of child mobility with immunization status is planned. Analyses of natural resource use as a livelihood strategy are ongoing.\n\nThe HDSS provides a platform for research across the life course. Examples of analytic work with adolescents include analyses of body composition and other cardiometabolic disease risk factors and levels, and facilitators and barriers to physical activity. In the populations aged \u226515 years, we are analysing HIV and non-communicable disease risk factors prevalence, as well as their interactions. In older adults, we are analysing the association between self-reported non-communicable diseases and health care utilization.\n\nWork on linking HDSS data with data from Statistics South Africa is underway to validate national census and vital registration data, especially completeness of population count and coverage of demographic events. We plan to inform policy by evaluating coverage and impact of health and social interventions and provide an early warning system to detect crises among vulnerable populations such as rising food insecurity or sudden peaks in mortality.\n\nCost-effectiveness will be a feature of intervention evaluations, and HDSS data can be used to model differential costs and benefits before field trials. For example, work to define the burden of epilepsy in disability-adjusted life years will inform cost-effectiveness analyses of potential interventions to address the treatment gap.\n\nIn several areas---including epidemiological and demographic transitions, physical and cognitive function in older adults and migration and health---comparative and pooled multi-centre analyses as part of INDEPTH initiatives are a major feature.\n\nStrengths and weaknesses\n========================\n\nThe value of the Agincourt HDSS is enhanced by its longevity and context---almost 20 years spanning profound political change and post-*apartheid* economic development and also the full force of the HIV/AIDS epidemic. This provides a robust sampling frame and research infrastructure to support a range of study designs.[@dys115-B8] Verbal autopsies, on all age groups, have been validated locally using district hospital cases as gold standard.[@dys115-B19]\n\nFurther strengths include a university base with a strong graduate training programme; productive partnerships with scientists based in Africa, Asia, UK/Europe and North America; ongoing engagement with village leaders, study communities, local government and non-governmental organizations and public personnel, and increasing collaboration with Statistics South Africa.\n\nEfforts are underway to strengthen the HDSS platform, including real-time electronic data entry, a comprehensive Geographic Information System platform coupled with natural resource and environmental monitoring, and reconciliation of internal migrants. Rural southern Africa is characterized by extensive labour migration that provokes research questions and methodological challenges. The Agincourt data system did not initially track migrants within the study site. We now apply a system of 'migration reconciliation' (MR) to account for migration and strengthen our analytical database. Undertaking retrospective MR from 2000, we have matched around 70% of internal migrations; with prospective MR, we are achieving \\> 90% of matches. We are also following out-migrants to better understand loss to follow-up in our cohorts. Union status was not recorded until 2007, when we retrospectively collected these data on all---with prospective monitoring since then.\n\nAs is the case with many HDSSs, we use a proxy respondent when updating the household roster and vital events. This may reduce the accuracy of some individual-level information (e.g. date of birth) and limits collection of sensitive data such as contraceptive use. We update vital events annually, with the risk of underestimating perinatal and infant mortality, particularly when birth and death occur between consecutive household visits. We limit this possibility by careful probing for pregnancies and births since the last recorded child and since 2006, by asking about pregnancy status of every woman of childbearing age.\n\nAs the number and range of projects increase, with multiple follow-up visits and biomarker measurements, we are concerned about overloading households and the potential for refusal rates to increase in the HDSS, as well as in nested studies. Efforts to ensure full documentation of household recruitments to studies are a priority.\n\nData sharing and collaboration\n==============================\n\nFostering effective collaborations, ensuring cross-site compatibility of common variables and optimizing public access to HDSS data are priorities. The Agincourt data website (<http://www.agincourt.co.za/DataSection/index.htm>) contains full documentation, including questionnaires, data dictionaries and metadata associated with the Agincourt HDSS, as well as an anonymized 10% sample that retains the relational, temporal and data integrity of the full database.[@dys115-B29] Researchers may request a customized data extraction (contact Dr Xavier G\u00f3mez-Oliv\u00e9; <xavier@agincourt.co.za>); details of application procedures with forms are available on the data website. The questionnaires, metadata and '1-in-10' sample database help users to prepare the detailed data request needed for a customized extraction.\n\nThe MRC/Wits-Agincourt Unit participates in data sharing initiatives that yield datasets that can be freely downloaded. The INDEPTH-WHO SAGE study (Study on global AGEing and adult health) is available on the *Global Health Action* and INDEPTH websites (<http://www.globalhealthaction.net/index.php/gha/rt/suppFiles/5302/6049> and <http://www.indepth-network.org/index.php?option=com_content&task=view&id=1215&Itemid=1059>). Agincourt data in the INDEPTH-iShare data repository include a subset of core demographic data exported to a common data model and stored in a central database; this will soon be available for download on its own or with similar data from other INDEPTH centres (<http://www.indepth-ishare.org/>).\n\nFunding\n=======\n\nThe Wellcome Trust, UK (grants 058893/Z/99/A; 069683/Z/02/Z; 085477/Z/08/Z), the National Institute on Ageing of the NIH (grants 1R24AG032112-01 and 5R24AG032112-03), the William and Flora Hewlett Foundation (grant 2008-1840) and the Andrew W Mellon Foundation, USA, and the University of the Witwatersrand and Medical Research Council, South Africa.\n\nBuilding and sustaining a health and socio-demographic surveillance system depends on the support and participation of key stakeholders. We greatly value the involvement of community leaders, study participants and district and provincial managers in health, education and other government sectors. National and international partners help us keep focus and stimulate new directions. The School of Public Health and Faculty of Health Sciences, University of the Witwatersrand, have provided vital support since inception of the Agincourt HDSS, as have the Medical Research Council and Limpopo and Mpumalanga Provinces, South Africa. The Wellcome Trust, UK has provided vital core funding for well over a decade. We acknowledge important contributions from the William and Flora Hewlett Foundation, National Institute on Aging (NIA) of the NIH and the Andrew W Mellon Foundation, USA.\n\n**Conflict of interest:** None declared.\n\nKey messagesThe Agincourt health and socio-demographic surveillance system in rural northeast South Africa, established in 1992 just before socio-political change and emergence of the HIV/AIDS epidemic, captures the dynamics of rapidly evolving health, population and social transitions; these do not conform to the patterns predicted by classical transition theory.The epidemiological transition is well underway in rural South Africa, with evidence of persisting high levels of infectious conditions (mainly HIV and tuberculosis), as well as emerging non-communicable diseases. This poses major challenges for primary health care services, which are ill-equipped to provide quality, integrated long-term care.A dual nutritional burden is characterized by marked stunting of children aged \\<2 years (approaching one-third of children), together with high levels of overweight and obesity in adolescent females (close to 25% by 18 years of age). This indicates elevated risk for metabolic disease in later life and is positively associated with higher socio-economic status.In collaboration with Statistics South Africa, the HDSS is used to validate national census and vital registration data and to analyze causal pathways that drive trends seen nationally.\n\n[^1]: ^a^This reflects disproportionately high male labour migration.\n\n[^2]: ^b^Mostly labour migrants working in towns and cities, in the mining sector or on large agricultural estates.\n\n[^3]: All figures given as % except for body mass index.\n\n[^4]: ^a^At 1 year.\n\n[^5]: ^b^At 2 years.\n\n[^6]: ^c^14--15 years.\n\n[^7]: ^d^18--20 years.\n\n[^8]: ^e^Assessed at Tanner Stage 5: pubertal self-rating scale defined by male genital development and female breast development.[@dys115-B24]\n\n[^9]: ^f^Abdominal obesity: waist circumference \\>102 cm in males and \\>88 cm in females (adults),[@dys115-B27] and \u226594 cm in males and \u226580 cm in females (adolescents).[@dys115-B24]\n\n[^10]: ^g^ABI \u2264 0.9 is associated with other cardiovascular disease indicators such as angina pectoris and carotid stenosis.[@dys115-B31]\n"}
+{"text": "The network models in our study were from existing references. All the underlying data set for our study could be available from the eleven sources listed below. 46. Erd\u0151s P, R\u00e9nyi A. On the strength of connectedness of a random graph. Acta Math. Hungar. 1961;12 (1): 261--267. 47. Barab\u00e1si AL, Albert R. Emergence of scaling in random networks. Science. 1999;286 (5439):509--512. 48. Newman MEJ, Watts DJ. Renormalization group analysis of the small-world network model. Phys. Lett. A. 1999;263:341-346. 49. Watts DJ, Strogatz SH. Collective dynamics of small-world networks. Nature.1998;393:440-442. 50. Batagelj V, Mrvar A. Pajek datasets. \\<URL: <http://vlado.fmf.uni-lj.si/pub/networks/data/>\\>. 2006. 51. Milo R, Itzkovitz S, Kashtan N, Levitt R, Shenorr S, Ayzenshtat I, et al. Superfamilies of designed and evolved networks. Science. 2004;303(5663): 1538-1542. 52. Burt RS. Social Contagion and Innovation: Cohesion versus Structural Equivalence. American Journal of Sociology.1987;92(6):1287-1335. 53. Newman MEJ. Finding community structure in networks using the eigenvectors of matrices. Phys. Rev. E. 2006;74(3): 036104. 54. <https://sparse.tamu.edu/>. 55. Leskovec J, Lang KJ, Dasgupta A, Mahoney MW. Community structure in large networks: Natural cluster sizes and the absence of large well-defined clusters. Internet Mathematics. 2009;6(1):29-123. 56. Leskovec J, Kleinberg J, Faloutsos C. Graph evolution: Densification and shrinking diameters. ACM Transactions on Knowledge Discovery from Data(TKDD). 2007;1(1):2.\n\nIntroduction {#sec001}\n============\n\nThe real world consists of ubiquitous intricate and intertwined networks. Some are tangible, such as traffic networks \\[[@pone.0193827.ref001], [@pone.0193827.ref002]\\], power networks \\[[@pone.0193827.ref003]\\], and financial networks \\[[@pone.0193827.ref004]\\], whereas others are invisible networks that penetrate into every aspect of our lives, such as interpersonal relationship networks \\[[@pone.0193827.ref005], [@pone.0193827.ref006]\\], wireless networks \\[[@pone.0193827.ref007], [@pone.0193827.ref008]\\], and ecological networks \\[[@pone.0193827.ref009]\\]. The expected goal of research into a complex network is to be able to regulate and control it from the outside, achieve the desirable state or performance by injecting outward control signals to some network nodes (called driver nodes), and ultimately achieve the real controllability of a complex network.\n\nMany studies have been conducted on the relationship between network topology and network controllability \\[[@pone.0193827.ref010]--[@pone.0193827.ref013]\\]. Researchers have proposed that all hub nodes with a high degree or betweenness centrality could be chosen as driver nodes \\[[@pone.0193827.ref014]\\]. Jalili et al. \\[[@pone.0193827.ref015]\\] found that an optimum driver node could not always be a hub node. To elucidate the configuration of driver nodes for the optimum network pinning control, a differential evolution method was used. The method worked well, but it was only suitable for undirected networks. Assuming that the objective network had finite-dimensional linear dynamics \\[[@pone.0193827.ref016]\\], the network could be structurally controlled by applying one time-varying input to the power dominating set. In the practical applications with economical and physical constraints, driver nodes can not always be freely selected to be injected to network nodes. In this context, Lo et al. \\[[@pone.0193827.ref017]\\] addressed a geometrical framework for the partial controllability issue of networks by solving an integer linear programme. The approach was also suitable to optimize the complete controllability of networks. The network permeability index provided a quantitative understanding of the challenge of controlling a network partially or completely.\n\nThe dynamics of the network could be expressed as a linear time-invariant system $\\left. \\overset{\u02d9}{x\\left( t \\right.} \\right) = Ax\\left( t \\right) + Bu\\left( t \\right)$, where x(t) is the state vector of the network. Assuming a network has *N* state nodes and P control nodes, A\u03f5*R*^*N*\u00d7*N*^ is the coupled matrix between state nodes, u(t) is the control or input vector forced on the network, and B\u03f5*R*^*N*\u00d7*P*^ is the input matrix. The general approach for the controllability problem $\\left. \\overset{\u02d9}{x\\left( t \\right.} \\right) = Ax\\left( t \\right) + Bu\\left( t \\right)$ is to determine a proper input matrix based on the Kalman rank condition such that the pair (A, B) is controllable \\[[@pone.0193827.ref010]\\]. However, this controllability problem has a large computational load with 2^*N*^ possibilities assuming each node can be either driven or not driven \\[[@pone.0193827.ref011]\\], and this exponential growth is especially rapid when the network size is large. To overcome this difficulty, Liu et al. \\[[@pone.0193827.ref010]\\] introduced the structural controllability concept \\[[@pone.0193827.ref018]\\], which ensured that the Kalman rank condition was verified. They first found that the number of driver nodes for the full controllability of a complex network mainly depended on the network degree distribution. The process controllability of network dynamics was explored by transforming node dynamics into edge switch dynamics \\[[@pone.0193827.ref019]\\] and resulted in similar controllability conclusions to those obtained by Liu et al. \\[[@pone.0193827.ref010]\\].\n\nThe structural controllability methods based on graphical analysis of pair (A, B) for the system $\\left. \\overset{\u02d9}{x\\left( t \\right.} \\right) = Ax\\left( t \\right) + Bu\\left( t \\right)$ \\[[@pone.0193827.ref018]\\] could identify n~D~ for arbitrary directed networks \\[[@pone.0193827.ref010]\\]. Several effective methods have been proposed to identify the minimum number of driver (control) nodes (n~D~), for example, the maximum matching (MM) method \\[[@pone.0193827.ref020]\\], the cavity method \\[[@pone.0193827.ref021]\\], and an extremal optimization (EO) algorithm \\[[@pone.0193827.ref022]\\]. The computational load of determining n~D~ could be effectively reduced based on the MM method, which has the computational complexity of $O\\left( \\sqrt{N}L \\right)$, where L is the number of linked edges between state nodes. EO \\[[@pone.0193827.ref022]\\] was proposed based on the Kalman rank condition to identify n~D~ for the full controllability of directed networks with the computational complexity of O(N^4^*P*^3^). However, this structural controllability framework \\[[@pone.0193827.ref010]\\] is not applicable to undirected networks for the symmetric characteristic of the network matrix or networks with exact link weights \\[[@pone.0193827.ref011], [@pone.0193827.ref023], [@pone.0193827.ref024]\\]. These limitations prompt the development of exact network controllability theory, which is an exact controllability framework for the controllability of complex networks with arbitrary network structures and link weights. It optimizes the complete controllability of networks based on the Popov--Belevitch--Hautus (PBH) rank condition \\[[@pone.0193827.ref025]\\], which is an alternative criterion that is equivalent to the Kalman rank condition \\[[@pone.0193827.ref026]\\]. The PBH controllability method requires the sequential computation of the eigenvalues of the N \u00d7 N matrix A and the rank of the N \u00d7 (N + P) PBH matrix. The computational complexity of the eigenvalue computation of matrix A and the PBH matrix rank is O(*N*^3^), O((*N* + *P*)^3^), respectively \\[[@pone.0193827.ref027]\\]. Thus, the computational complexity of the PBH method is O((*N* + *P*)^3^). The Kalman controllability method does not require an eigenvalue computation. However, it requires the rank computation of the N \u00d7 NP Kalman matrix with the computational complexity of O(*N*^3^*P*^3^), which is larger than that of the PBH controllability method. Representative exact structural controllability methods consist of a maximum multiplicity theory (MMT) \\[[@pone.0193827.ref011]\\] and an effective self-adaptive genetic algorithm (GA) \\[[@pone.0193827.ref028]\\]. n~D~ was computed based on the MMT to be equal to the maximum geometric multiplicity of all eigenvalues of the network \\[[@pone.0193827.ref011]\\], and the computational complexity is O(*N*^2^(logN)^2^). The GA \\[[@pone.0193827.ref028]\\] was studied to identify n~D~ of arbitrary networks, and the computational complexity is O(2m \u00d7 (*N* + *P*)^3^ \u00d7 *l*), where 2m is the population size and *l* is the number of different eigenvalues of the controlled network.\n\nThe authors demonstrated the evolution process of network topology of two networks (Erd\u0151s--R\u00e9nyi (ER) and scale free (SF)) \\[[@pone.0193827.ref028]\\], i.e., the dynamic change of network topology with different number of control nodes being injected to network state nodes. The GA algorithm \\[[@pone.0193827.ref028]\\] exceeded the EO \\[[@pone.0193827.ref022]\\] much in terms of the convergence speed and iterations. However, networks with the more complexity than the scale 150 were not studied in \\[[@pone.0193827.ref028]\\] and the maximum network scale that EO processed was 200 \\[[@pone.0193827.ref022]\\]. And the convergence speed and iterations were still not satisfactory, for example, for ER, with both 150 state nodes and 150 control nodes, and the average degree 5.0, n~D~ converged at the 101st generation after 398.03 s \\[[@pone.0193827.ref028]\\]. The results showed that it remained a challenge for the two algorithms to optimize networks with hundreds of thousands or even larger networks. Additionally, almost all real networks have small-world (SW) properties with a large cluster coefficient and short average distance \\[[@pone.0193827.ref029]\\] (e.g., power grids, transportation networks, and social networks). The addition of the SW network controllability study is also significant for better mimicking reality.\n\nTherefore, based on the PBH rank condition, we propose a parallel quantum genetic algorithm (PAQGA) to more rapidly determine the minimum number of control nodes. The proposed algorithm is suitable for arbitrary networks that comprise both control nodes and state nodes. The simulation results for a series of benchmark networks demonstrate the effectiveness of the algorithm. Furthermore, we demonstrate the relationship between the controllability of a network and its network properties such as network average degree, degree heterogeneity, power-law index, and clustering coefficient.\n\nThe remainder of this paper is organized as follows: In Section 2, we provide a description of the issue in which a network can be controlled through a small amount of control nodes. In Section 3, we introduce the PAQGA for the solution of the minimum number of control nodes to exactly control arbitrary networks. In Section 4, we analyze and discuss the performance and experimental results of the proposed framework by studying popular ER, SW, SF, and some real-life networks. We draw conclusions and suggest future work in Section 5.\n\nProblem definition {#sec002}\n==================\n\nIn this paper, we provide a descriptive definition of the entire controllability problem of a directed weighted network.\n\n**Definition 2.1** \\[[@pone.0193827.ref022]\\]. A network that contains P control nodes and N state nodes can be expressed as a triple tuple G = (V,E,W), where V = *V*~*s*~ \u222a *V*~*c*~, *V*~*s*~ = {*v*~1~,*v*~2~,...,*v*~*N*~} = {*x*~1~,*x*~2~,...,*x*~*N*~} is the set of state nodes, and *V*~*c*~ = {*v*~*N+*1~,*v*~*N+*2~,...,*v*~*N+P*~} = {*u*~1~,*u*~2~,...,*u*~*P*~} is the set of control nodes; E = *E*~*s*~ \u222a *E*~*c*~, *E*~*s*~ \u2208 *V*~*s*~ \u00d7 *V*~*s*~ is the set of the linked edges between state nodes, and *E*~*c*~ \u2208 *V*~*c*~ \u00d7 *V*~*s*~ is the set of linked edges between control nodes and state nodes, where each state node can only be connected to one control node; and W \u2208 *R*^(*N*\\ +\\ *P*)\\ \u00d7\\ (*N*\\ +\\ *P*)^ is the set of edge weights, *w*~*ij*~ = 0 if there is not a link between *v*~*i*~ and *v*~*j*~; otherwise, *w*~*ij*~ (*w*~*ji*~) represents the strength that *v*~*i*~ (*v*~*j*~) could affect *v*~*j*~ (*v*~*i*~), *w*~*ij*~ (*w*~*ji*~) \\> 0 if the direction is i \u2192 j (j \u2192 i). [Fig 1](#pone.0193827.g001){ref-type=\"fig\"} shows an illustration of the definition.\n\n![Example of a directed network, where V~s~ = {x~1~,x~2~,x~3~,x~4~,x~5~,x~6~,x~7~}, V~c~ = {u~1~,u~2~,u~3~}, E~s~ = {(x~1~,x~2~),(x~1~,x~3~),(x~2~,x~4~),(x~2~,x~5~),(x~3~,x~6~),(x~3~,x~7~),(x~4~,x~2~),(x~5~,x~6~),(x~7~,x~7~)}, E~c~ = {(u~1~, x~1~),(u~2~, x~2~),(u~3~, x~3~)}, a~ij~ (i = 1,2,...,7; j = 1,2,...,7) \u2208 V~s~ \u00d7 V~s~, b~ij~(i = 1,2,3; j = 1,2,...,7) \u2208 V~c~ \u00d7 V~s.~, *w*~24~ \\> 0 and *w*~42~ \\> 0.](pone.0193827.g001){#pone.0193827.g001}\n\n**Remark 2.1** \\[[@pone.0193827.ref022]\\]. The set W can be expressed by the representation of a block matrix that contains *A* and B as follows: $$W = \\begin{bmatrix}\nA & B \\\\\n0 & 0 \\\\\n\\end{bmatrix},$$\n\n**Definition 2.2** \\[[@pone.0193827.ref010], [@pone.0193827.ref022]\\]. The network G = (V,E,W) can be represented by $$\\left. \\overset{\u02d9}{x\\left( t \\right.} \\right) = Ax\\left( t \\right) + Bu\\left( t \\right),$$ where **x**(t) = (*x*~1~(*t*),*x*~2~(*t*),...,*x*~*N*~(*t*))^*T*^ is the state vector of the network, A\u03f5*R*^*N*\u00d7*N*^ is the coupled matrix between state nodes, u(t) = (*u*~1~(*t*),*u*~2~(*t*),...,*u*~*P*~(*t*))^*T*^ is the control or input vector forced on the network, B\u03f5*R*^*N*\u00d7*P*^ is the input matrix, and B = {*b*~*ij*~}, *b*~*ij*~ is the weight of a directed link that the input signal *u*~*j*~(j = 1,2,...,P) points to the network state node *x*~*i*~ (*i* = 1,2,...,*N*). For simplicity, hereafter the time symbol (t) will be omitted. [Fig 2](#pone.0193827.g002){ref-type=\"fig\"} shows the equation expression of the network in [Fig 1](#pone.0193827.g001){ref-type=\"fig\"}.\n\n![Corresponding dynamics equation of the network in [Fig 1](#pone.0193827.g001){ref-type=\"fig\"}, where x = {x~1~, x~2~, x~3~,x~4~,x~5~,x~6~,x~7~}, u = {u~1~,u~2~,u~3~}, a~ij~ (i = 1,2,...,7; j = 1,2,...,7) is the connection weight from state node i to state node j, and b~ij~(i = 1,2,3; j = 1,2,...,7) is the connection weight from control node i to state node j.](pone.0193827.g002){#pone.0193827.g002}\n\n**Definition 2.3** \\[[@pone.0193827.ref022]\\]. A control scheme D of a network G = (V,E,W) is determined by the selected control nodes with definite number and their acting position. D could be represented by a binary diagonal matrix as D = diag{*d*~1~,*d*~2~,...,*d*~*P*~}, where *d*~*i*~(*i* = 1,2,...,*P*) is a variable of value zero or one, and *d*~*i*~ = 1 means that the control node *u*~*i*~ is chosen to be a component of the network control strategy; otherwise, *u*~*i*~ is removed together with its associated links.\n\n**Remark 2.2** \\[[@pone.0193827.ref022]\\]. Based on Definition 2.3, a novel control scheme *D*\\* is determined for which a different set of control nodes is selected. Then a novel network topology is generated as *G*\\* = (*V*\\*,*E*\\*,*W*\\*), where *V*\\* \u2208 *V*, *E*\\* \u2208 *E*, and *W*\\* \u2208 *R*^(*N*\\ +\\ *r*)\\ \u00d7\\ (*N*\\ +\\ *r*)^, where r is the number of selected control nodes. Accordingly, the network dynamics are also changed as $$\\overset{\u02d9}{x} = Ax + B^{*}u^{*},$$ where *B*\\* \u2208 *R*^*N*\u00d7*r*^ is the new input matrix that represents the connections between new chosen control nodes and network state nodes, and *u*\\* \u2208 *R*^*r*^ is a time-variable input vector that contains r control nodes.\n\n**Definition 2.4** \\[[@pone.0193827.ref022]\\]. M\u03f5*R*^*P*\u00d7*r*^ is the index set of the selected control nodes, M = {*m*~*ij*~}, and *m*~*ij*~ = 1 means that the *j*~*th*~ chosen control node is *u*~*i*~, *i* = 1,2,...,P, j = 1,2,...,r, r \u2264 P.\n\n**Remark 2.3** \\[[@pone.0193827.ref022]\\]. M is constructed by the nonzero columns of the control scheme *D*\\*. For example, if *u*\\* = {*u*~1~,*u*~2~,*u*~3~} is chosen from a previous control node set *u* = {*u*~1~,*u*~2~,*u*~3~, *u*~4~} to be a new control scheme *D*\\* = *diag*{1,1,1,0}, then M is obtained from this *D*\\* as $$M = \\begin{bmatrix}\n1 & 0 & 0 \\\\\n0 & 1 & 0 \\\\\n0 & 0 & 1 \\\\\n0 & 0 & 0 \\\\\n\\end{bmatrix},$$\n\n**Remark 2.4** \\[[@pone.0193827.ref022]\\]. The evolving input matrix and input vector can be revised as *B*\\* = *BD*\\*M and *u*\\* = M^T^*u*, respectively. Then Eq ([3](#pone.0193827.e007){ref-type=\"disp-formula\"}) can be rewritten as $$\\overset{\u02d9}{x} = Ax + BD^{*}M\\left( {M^{T}u} \\right),$$\n\n[Fig 3](#pone.0193827.g003){ref-type=\"fig\"} shows a simple case that illustrates how a control scheme influences the network topology.\n\n![Illustration of how new control scheme D\\* functions in the network topology.\\\n(a) Original network with seven state nodes and five candidate control nodes. (b) Input matrix changes into B\\* = BD\\*M after choosing u\\* = {u~1~,u~2~,u~3~} as new control nodes. New network has seven state nodes and three control nodes.](pone.0193827.g003){#pone.0193827.g003}\n\n**Remark 2.5.** Based on the PBH rank condition \\[[@pone.0193827.ref025]\\], the network G = (V,E,W) (Eq. ([2](#pone.0193827.e006){ref-type=\"disp-formula\"})) can be steered to any desired state within a finite time, that is, G is fully controllable if and only if $$rank\\left( {\\lambda_{i}I_{N} - A,B} \\right) - N = 0,$$ and the new system (Eq. ([5](#pone.0193827.e009){ref-type=\"disp-formula\"})) is fully controlled if and only if $$rank\\left( {\\lambda_{i}I_{N} - A,BD^{*}M} \\right) - N = 0,$$ is satisfied for each different eigenvalue *\u03bb*~*i*~ of the state matrix A, where *I*~*N*~ \u2208 *R*^*N*\u00d7*N*^ is an identity matrix.\n\nFor an arbitrary network G = (V,E,W), our purpose is to determine the minimum control nodes to guarantee its full control. Based on the above analysis, the controllability problem can be transformed into a single target restricted optimization problem as $$\\min_{D}{\\sum_{j = 1}^{P}{d_{j},}}$$ subject to $$rank\\left( {\\lambda_{i}I_{N} - A,BDM} \\right) - N = 0,{\\forall\\lambda}_{i} \\in eig\\left( A \\right),$$ $$d_{j} = \\left\\{ {0,1} \\right\\},j = 1,2,\\ldots,P,$$ where *A* is the state matrix, *B* is the original input matrix, *D* is the original control scheme, *M* is the indicator matrix that is derived from the nonzero column of *D*, N and *P* are the dimensions of *A* and *B*, respectively, *\u03bb*~*i*~ is the eigenvalue that belongs to *A*, *eig*(*A*) is the set of different eigenvalues of *A*, *d*~*j*~ is the element of *D*, and *d*~*j*~ = 1 when *u*~*j*~ is selected and *d*~*j*~ = 0 otherwise.\n\nSolution framework {#sec003}\n==================\n\nOverview of the QGA {#sec004}\n-------------------\n\nGA is a global optimization method that can optimize problems with multiple parameters to reach near the global optima \\[[@pone.0193827.ref030]--[@pone.0193827.ref033]\\]. However, in some practical applications, it often requires multiple iterations because of the slow convergence speed and prematurity features, and easily falls into the local minima \\[[@pone.0193827.ref034], [@pone.0193827.ref035]\\]. Additionally, for many complex problems, a large population is required to obtain the optimal solution. The convergence of GA mainly depends on the selecting operation, which largely affects the convergence speed. Additionally, its searching capability mainly relies on crossover and mutation operations, which primarily affect the occurrence of the premature phenomenon. Therefore, regarding enhancing GA search performance, the approach used to choose suitable selecting, crossover, and mutation strategies has been always an urgent and pivotal issue in the study and application of GA \\[[@pone.0193827.ref033], [@pone.0193827.ref036]\\].\n\nFor small and medium-sized applications, the solution could be achieved within a tolerance range using GA. However, a gene (typically encoded with a 0--1 string) in a GA chromosome typically delivers certain information, which limits the population diversity. It performs worse in multivariate issues, for example, the controllability study of complex networks, which mostly has complex structures, and large-size nodes and links.\n\nCombining quantum computing and GA, and adopting qubits as the representation of chromosome genes \\[[@pone.0193827.ref037]\\], QGA is a proper intelligent optimization algorithm for solving the network controllability problem \\[[@pone.0193827.ref038]\\]. These QGA qubits cover all possibilities for the linear superposition property of quantum information, which could reduce the algorithm's complexity and promote the achievement of the optimal solution under a smaller population \\[[@pone.0193827.ref037]\\].\n\nIn quantum computing, \\|0\u27e9 and \\|1\u27e9 signify two basic states of microscopic particles. According to the principle of the superposition property, the superposition state of quantum information could be the linear combination of the two basic states \\[[@pone.0193827.ref039]\\], which can be written as $$\\left. \\left| \\varphi \\right. \\right\\rangle\\left. = \\alpha \\middle| {\\left. 0 \\right\\rangle + \\beta} \\middle| 1 \\right\\rangle,\\ \\left| \\alpha \\right|^{2} + \\left| \\beta \\right|^{2} = 1,$$ where \u03b1 and \u03b2 are the state probability amplitudes of a qubit, and \u03b1^2^ and *\u03b2*^2^ are the probability that a qubit changes to be state \\|0\u27e9 and state \\|1\u27e9, respectively. One qubit also can be expressed as $\\begin{bmatrix}\n\\alpha \\\\\n\\beta \\\\\n\\end{bmatrix}$.\n\nAssume the number of optimization variables is n and the population size is 2m. The i~th~ chromosome is denoted by G~i~(i = 1,2,...,m) as $$G_{i} = \\begin{bmatrix}\n\\alpha_{i1} & \\alpha_{i2} & \\ldots & \\alpha_{in} \\\\\n\\beta_{i1} & \\beta_{i2} & \\beta_{in} & \\\\\n\\end{bmatrix},$$ where $\\alpha_{i1}^{2} + \\alpha_{i2}^{2} + \\beta_{i1}^{2} + \\beta_{i2}^{2} = 1,i = 1,2,\\ldots,m$. G~i~ contains two parallel gene chains or individuals (\u03b1~i1~,\u03b1~i2~,...,\u03b1~in~ and \u03b2~i1~,\u03b2~i2~,...,\u03b2~in~). Each individual is a candidate solution of an optimization problem: $${G_{i} = \\begin{bmatrix}\nG_{i1} \\\\\nG_{i2} \\\\\n\\end{bmatrix},\\mspace{9mu}\\left\\{ \\begin{matrix}\n{G_{i1} = \\left\\lbrack {\\alpha_{i1},\\ \\alpha_{i2},..,\\alpha_{ij},..,\\alpha_{in}} \\right\\rbrack} \\\\\n{G_{i2} = \\left\\lbrack {\\beta_{i1},\\ \\beta_{i2},\\ldots,\\beta_{ij},..,\\beta_{in}} \\right\\rbrack} \\\\\n\\end{matrix} \\right.},$$\n\nQGA and enhanced QGA have already been studied to optimize many combinational problems \\[[@pone.0193827.ref040]--[@pone.0193827.ref042]\\]. For example, QGA overmatches classic GA with less complexity and higher performance in 0--1 combinational optimization problems \\[[@pone.0193827.ref039]\\]. Adaptive QGA models were proposed and tested on classical combinational problems, such as knapsack, maxcut and onemax \\[[@pone.0193827.ref038]\\], the multi-aircraft cooperative target allocation problem, and constrained engineering design problems \\[[@pone.0193827.ref043]\\]. However, the time efficiency was not seriously stressed. To increase the speed, a parallel QGA was developed and effectively applied to a knapsack problem \\[[@pone.0193827.ref044]\\]. It divided the entire population into subpopulations on different parallel processors and used the migration rate for the information exchange of these subpopulations. However, the Q-gate rotation was implemented according to a fixed lookup table, which did not take full advantage of the dynamic differences between individuals during the iterating process.\n\nInspired by current achievements, to quickly and efficiently solve the controllability problem of complex networks, we investigated a PAQGA scheme, in which: 1) partial programs of the algorithm are executed in parallel; 2) a set of adaptive Q-gate rotation rules are proposed and adaptive crossover operation are used; and 3) population catastrophe is implemented to accelerate convergence.\n\nWorkflow of the PAQGA for network controllability {#sec005}\n-------------------------------------------------\n\nBased on the above, each control scheme *D* is a diagonal matrix, whose elements on the primary diagonal are either zero or one. Therefore, we adopt the binary mechanism to encode the algorithm chromosome *G*~*i*~(*i* = 1,2,...,*m*), and each binary gene chain can represent a control scheme, as shown in [Fig 4](#pone.0193827.g004){ref-type=\"fig\"}.\n\n![Chromosome encoding in quantum genetic algorithm.](pone.0193827.g004){#pone.0193827.g004}\n\nTo apply PAQGA conveniently, a penalty term *Pen*~*i*~(*D*) is defined to convert the optimization problem described in Eqs ([8](#pone.0193827.e012){ref-type=\"disp-formula\"}), ([9](#pone.0193827.e013){ref-type=\"disp-formula\"}) and ([10](#pone.0193827.e014){ref-type=\"disp-formula\"}) into an unconstrained optimization problem. *Pen*~*i*~(*D*) is used to evaluate the *i*~*th*~ perturbation for a specific control scheme *D*: $${Pen_{i}\\left( D \\right) = \\sigma_{i} \\times \\left( {rank\\left( {\\lambda_{i}I_{N} - A,BDM} \\right) - N} \\right)^{2},i = 1,2,\\ldots,l},$$ where *\u03c3*~*i*~ is the penalty coefficient defined as *\u03c3*~*i*~ = {*c\u03c3*~*i*\u22121~} (*\u03c3*~1~ = 10*P*, *c* \\> 0) is a strictly increasing positive sequence to reduce the calculation burden of minimizing the penalty function, and *l* is the total number of distinct eigenvalues of *A*. The overall penalty of *D* is the sum of *Pen*~*i*~(*D*) expressed as $$Pen\\left( D \\right) = {\\sum_{i = 1}^{l}{{Pen}_{i}\\left( D \\right)}},$$\n\nAccording to the PBH rank condition, when the network G = (V,E,W) is fully controllable, *Pen*(*D*) should be zero and vice versa. Therefore, the fitness function can be achieved by merging the penalty term into the optimization Eqs ([8](#pone.0193827.e012){ref-type=\"disp-formula\"}), ([9](#pone.0193827.e013){ref-type=\"disp-formula\"}) and ([10](#pone.0193827.e014){ref-type=\"disp-formula\"}): $$\\begin{array}{l}\n{f\\left( D \\right) = {\\sum\\limits_{j = 1}^{P}{d_{j} + Pen\\left( D \\right)}}} \\\\\n{\\qquad = {\\sum_{j = 1}^{P}{d_{j} + {\\sum_{i = 1}^{l}{\\sigma_{i} \\times \\left( rank\\left( {\\lambda_{i}I_{N} - A,BDM} \\right) - N \\right)^{2},}}}}} \\\\\n\\end{array}$$\n\nFor the optimization problem, our objective is to minimize f(D) based on the 0--1 integer values of *d*~*j*~ (*j* = 1,2,...,*P*). [Fig 5](#pone.0193827.g005){ref-type=\"fig\"} shows the fitness evaluation of different control schemes on a simple network.\n\n![Illustration of the fitness evaluation for different control schemes on a directed weighted network with self-loop.\\\n(a) Initial network with seven state nodes and five control nodes. Connecting weights are randomly assigned between zero and one. Fitness of initial D = diag{1,1,1,1,1} is f(D) = 5 and the penalty term is zero; thus, the network is entirely controllable. (b) When D changes to new D = diag{1,1,1,0,1}, f(D) is four, the penalty is zero, and the network is still fully controllable. (c) u~4~ is removed from (b). For simplicity, *c* is set to 1, and the penalty term $pen\\left( D \\right) = 10P*{\\sum\\limits_{i = 1}^{l}{{Pen}_{i}\\left( D \\right)}} = 50*1 \\neq 0$ indicates that the network with this topology cannot be fully controlled.](pone.0193827.g005){#pone.0193827.g005}\n\nAfter defining the chromosome representation and fitness function, the network controllability problem can be optimized using the following steps. [Fig 6](#pone.0193827.g006){ref-type=\"fig\"} shows the flow chart of the proposed PAQGA for the optimization problem.\n\n![Flow chart of the PAQGA.](pone.0193827.g006){#pone.0193827.g006}\n\n**Step 1:** The population at the t~*th*~ generation is denoted as $Q\\left( t \\right) = \\left\\{ {G_{1}^{t}{,G}_{2}^{t},\\ldots,G_{m}^{t}} \\right\\},$ $$G_{i}^{t} = \\begin{bmatrix}\n\\alpha_{i1}^{t} & \\alpha_{i2}^{t} & \\ldots & \\alpha_{iP}^{t} \\\\\n\\beta_{i1}^{t} & \\beta_{i2}^{t} & \\beta_{iP}^{t} & \\\\\n\\end{bmatrix},i = 1,2,\\ldots,m,t = 0,2,\\ldots,maxgen - 1,$$ where N is the number of qubits, that is, the number of network state nodes, and *maxgen* is the maximum iterating generation.\n\nInitialize the initial population as $$Q\\left( t_{0} \\right) = \\left\\{ {G_{1}^{0}{,G}_{2}^{0}{,\\ldots,G}_{m}^{0}} \\right\\},$$ where $G_{i}^{0} = \\begin{bmatrix}\n\\alpha_{i1}^{0} & \\alpha_{i2}^{0} & \\ldots & \\alpha_{iP}^{0} \\\\\n\\beta_{i1}^{0} & \\beta_{i2}^{0} & \\beta_{iP}^{0} & \\\\\n\\end{bmatrix},i = 1,2,\\ldots,m$.\n\nAll the quantum states (\u03b1~ik~ and \u03b2~ik~) in the PAQGA are initialized in parallel with the value $\\frac{1}{\\sqrt{2}}$, i = 1,2,...,m, k = 1,2,...,N. Additionally, set *D*~*best*~ = *D*~0~, the iterative generation t = 1, and *\u03c3*~1~ = 10*P*.\n\nWhen we set the initial control scheme D~0~ = *diag*{*d*~*j*~ = 1},*j* = 1,2,...,*P*, the initial best fitness value is f(D~*best*~) = *P* + 0 = *P*. It is easily proved that with *\u03c3*~*i*~ = {*c\u03c3*~*i*\u22121~} (*\u03c3*~1~ = 10*P*, *c* \\> 0), the fitness f(D) \u2264 *P* if and only if the control scheme D always satisfies the constraint Eq ([9](#pone.0193827.e013){ref-type=\"disp-formula\"}). With the initialization f(D~*best*~) = *P*, whenever D~*best*~ is updated by *D*~*i*~, we have f(*D*~*i*~) \\< f(D~*best*~) = *P*, which means *D*~*i*~ meets the constraint Eq ([9](#pone.0193827.e013){ref-type=\"disp-formula\"}). Thus, D~*best*~ always evolves in the feasible region that makes the network entirely controllable.\n\n**Step 2:** Observe the qubits of each individual of Q(*t*~0~) in parallel following the rules in Section 3.3 and obtain the binary strings.\n\n**Step 3:** Evaluate each individual of Q(*t*~0~) in parallel and save the optimal individual as the evolving goal in the next generation.\n\n**Step 4**:\n\nWhile (*maxgen* is not reached), do the following:\n\n1.  Observe the qubit value of each individual of Q(t) in parallel following the rules in Section 3.3.\n\n2.  To increase the diversity of the population and inherit the excellent genes from the previous population, the adaptive crossover operation is performed in parallel in accordance with Section 3.4.\n\n3.  Evaluate each individual of Q(t) in parallel and store the optimal individual as the evolving goal in the next generation.\n\n4.  Perform the Q-gate rotation operation in parallel and obtain the offspring population Q (t+1).\n\nFor each individual, two parallel gene chains update simultaneously. Rotation angle \u03b8 is first computed based on Section 3.5 and then the qubits are updated by Q-gate rotation. The Q-gate is expressed as \\[[@pone.0193827.ref045]\\] $$Q\\text{-}gate = \\begin{bmatrix}\n{cos\\theta} & {- sin\\theta} \\\\\n{sin\\theta} & {\\mspace{14mu}{cos\\theta}} \\\\\n\\end{bmatrix},$$\n\nThe Q-gate rotation operation is $$\\left\\{ \\begin{matrix}\n{\\alpha_{ik}^{t + 1} = cos\\theta*{\\alpha_{ik}^{t}}^{\\prime} - sin\\theta*{\\beta_{ik}^{t}}^{\\prime}} \\\\\n{\\beta_{ik}^{t + 1} = sin\\theta*{\\alpha_{ik}^{t}}^{\\prime} + cos\\theta*{\\beta_{ik}^{t}}^{\\prime}} \\\\\n\\end{matrix} \\right.$$ where *i* = 1,2,...,*m*, k = 1,2,...,*N*, *t* = 1,2,...,*maxgen* \u2212 1, and \u03b8 = \u03b4 \\* s, where \u03b4 is the rotation angle value and s is its sign.\n\n1.  Record *D*~*best*~ and f(D~*best*~).\n\n2.  If the optimal values of the past several successive generations are the same, then perform the parallel population catastrophe operation.\n\nThe fitness function evaluation and the crossover operation are the two most time-consuming steps in the process of the flow execution. Assume that \u03b7 parallel processors are used, the cost is $O\\left( \\frac{m \\times \\left( N + p \\right)^{3} \\times l}{\\eta} \\right)$ and $O\\left( \\frac{m \\times \\left( N + p \\right)^{3} \\times l}{\\eta^{2}} \\right)$, respectively, where *l* is the number of different eigenvalues of the controlled network. Therefore, the computational complexity of the PAQGA is $O\\left( \\frac{m \\times \\left. \\left( N + p \\right. \\right)^{3} \\times l}{\\eta} \\right)$.\n\nObserving operation {#sec006}\n-------------------\n\nEach qubit of the chromosome can be adjusted to be at a stationary state using an observation operation. We adopt a random observing method by running the following pseudocode in parallel:\n\nstart\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0if$\\left( {rank\\left( k \\right) \\geq \\left( \\beta_{ik}^{t} \\right)^{2}} \\right),i = 1,2,\\ldots,m;k = 1,2,\\ldots,N$\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return binary $\\alpha_{ik}^{t} = 1$;\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0else\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return binary $\\alpha_{ik}^{t} = 0$;\n\nend\n\nwhere rand(k) is a random digit. If rand(k) is not less than $\\left( \\beta_{ik}^{t} \\right)^{2}$ (the probability to be state \\|1\u27e9), then the observed value of the qubit $\\alpha_{ik}^{t}$ is 1 and 0 otherwise.\n\nCrossover operation {#sec007}\n-------------------\n\nThe crossover operator is an important operation of GA. Information about individuals can be exchanged using the operation. Subsequently, excellent genes could be reserved for population evolution to move in a better direction. To increase the diversity of the population and improve the optimization performance of PAQGA, the crossover operator is introduced. We obtain novel binary values $\\alpha_{ik}^{\\prime}$ by crossing each binary qubit value \u03b1~*ik*~,i = 1,2,...,m, k = 1,2,...N with corresponding information on the historically best control scheme, that is, *D*~*best*~(*k*,*k*), based on a certain crossover probability. The specific crossover mode is $$\\alpha_{ik}^{\\prime} = \\left\\{ \\begin{matrix}\n{D_{best}\\left( {k,k} \\right),\\ if\\ rand\\left( k \\right) < p_{c}} \\\\\n{\\alpha_{ik},\\ otherwise,\\ i = 1,2,\\ldots,m;k = 1,2,\\ldots,N,} \\\\\n\\end{matrix} \\right.$$ where *i* is the *i*~*th*~ individual, *rand*(*k*) is a random number between \\[0, 1\\], and *p*~*c*~ is the crossover probability. [Fig 7](#pone.0193827.g007){ref-type=\"fig\"} shows a simple crossover example with 10 qubits to explain the rule ([21](#pone.0193827.e040){ref-type=\"disp-formula\"}).\n\n![Simple crossover example.](pone.0193827.g007){#pone.0193827.g007}\n\nIn the early days of population evolution, there existed relatively big differences between individuals. Therefore, the crossover possibility to produce better offspring should have been bigger. Moreover, if we increased the crossover probability at this time, the evolution process would have been accelerated. By contrast, in the late stages of evolution, differences between individuals became smaller as the best solution was approaching. The crossover probability should have been correspondingly diminished to reserve the good genes. We design an adaptive crossover operator as $$p_{c}\\left( i \\right) = \\left\\{ {\\begin{matrix}\n{\\frac{m}{Q_{i}}p_{c0}\\exp\\left( {- \\frac{\\left| {f_{max} - f\\left( G_{i} \\right)} \\right|}{f_{max} - f_{min}}} \\right),\\mspace{9mu} f_{max} \\neq f_{min}} \\\\\n{\\frac{m}{Q_{i}}p_{c0},\\mspace{9mu} f_{max} = f_{min}} \\\\\n\\end{matrix},\\ i = 1,2,\\ldots,m,} \\right.$$ where *p*~*c*~(*i*) is the crossover probability of the *i*~*th*~ current individual, *Q*~*i*~ is the number of those individuals whose fitness is better than that of the historically best individual, *p*~*c*0~ is the initial crossover probability, *f*~*max*~ and *f*~*min*~ are the previous worst fitness and best fitness, respectively, and *f*(*G*~*i*~) is the fitness of the *i*~*th*~ current individual.\n\nWe can observe that *p*~*c*~(*i*) becomes bigger when the control scheme *G*~*i*~ becomes worse and vice versa. Moreover, *p*~*c*~(*i*) is inversely proportional to *Q*~*i*~, which means that if there are not so many good individuals, *p*~*c*~(*i*) should be bigger to produce a greater number of better individuals; otherwise, it should be smaller because the evolving individuals are becoming better. The improved adaptive crossover operation from Eq ([20](#pone.0193827.e030){ref-type=\"disp-formula\"}) is $$\\alpha_{ik}^{\\prime} = \\left\\{ \\begin{matrix}\n{D_{best}\\left( {k,k} \\right),\\ if\\ rand\\left( k \\right) < p_{c}\\left( i \\right)} \\\\\n{\\alpha_{ik},\\ otherwise} \\\\\n\\end{matrix} \\right.,i = 1,2,\\ldots,m;k = 1,2,\\ldots N,$$\n\nA better population is determined after the crossover operation following the pseudocode.\n\nstart\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0obtain fitness(i) in parallel;\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0find f(max) and f(min);\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0obtain pc(i) according to formula ([22](#pone.0193827.e041){ref-type=\"disp-formula\"}) in parallel;\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0obtain new binary population;\n\nend\n\nwhere fitness(i) is the fitness value of the i^*th*^ individual.\n\nRotation angle updating rules {#sec008}\n-----------------------------\n\nLearning from the solid lookup rules \\[[@pone.0193827.ref039]\\], we present a set of adaptive rotation angle updating rules in [Table 1](#pone.0193827.t001){ref-type=\"table\"}. The rotation angle *\u03b8*~*i*~ (*\u03b8*~*i*~ = \u03b4~*i*~ \\* s(*\u03b1*~*i*~,*\u03b2*~*i*~)), *i* = 1,2,...,*m* dynamically varies according to the evolution process.\n\n10.1371/journal.pone.0193827.t001\n\n###### Rotation angle updating rules.\n\n![](pone.0193827.t001){#pone.0193827.t001g}\n\n  D~c~(i)   D~best~(i)   f(D~c~) \\< *f*(D~best~)   s(*\u03b1*~*i*~,*\u03b2*~*i*~)                                                                      \n  --------- ------------ ------------------------- -------------------------------------------------------------------------- ---- ---- ---- ----\n  0         0            false                     0                                                                          0    0    0    0\n  0         0            true                      0                                                                          0    0    0    0\n  0         1            false                     $\\frac{f\\left( D_{c} \\right)}{f\\left( D_{best} \\right)} \\bullet 0.03\\pi$   +1   \u22121   0    \u00b11\n  0         1            true                      $\\frac{f\\left( D_{c} \\right)}{f\\left( D_{best} \\right)} \\bullet 0.01\\pi$   \u22121   +1   \u00b11   0\n  1         0            false                     $\\frac{f\\left( D_{c} \\right)}{f\\left( D_{best} \\right)} \\bullet 0.03\\pi$   \u22121   +1   \u00b11   0\n  1         0            true                      $\\frac{f\\left( D_{c} \\right)}{f\\left( D_{best} \\right)} \\bullet 0.01\\pi$   +1   \u22121   0    \u00b11\n  1         1            false                     0                                                                          0    0    0    0\n  1         1            true                      0                                                                          0    0    0    0\n\nIf D~c~(i) \u2260 D~best~(i), the rotation angle \u03b4~*i*~ is adaptively proportional to $\\frac{f\\left( D_{c} \\right)}{f\\left( D_{best} \\right)}$. If f(D~c~) \\< *f*(D~best~), the angle will be smaller; otherwise, it will be bigger. Initially, a big initial angle is set. As the iteration proceeds, the differences between individuals decrease and \u03b4~*i*~ becomes smaller. In this way, the probability amplitude evolves in the direction of the optimal solution.\n\nPopulation catastrophe {#sec009}\n----------------------\n\nWhen the best individuals in several successive generations are identical, it shows that the algorithm falls into a local minimum. At this moment, catastrophe operations for the current population should be performed to take it out of the constraint and start a new search. Specifically, the successive best individual is retained in the new population Q(t + 1) and the remaining individuals in Q(t + 1) are regenerated as a large disturbance. The pseudocode of the catastrophe operation is as follows:\n\nstart\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0obtain the best individual corresponding to the optimal fitness;\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0keep this best individual;\n\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0rebuild the rest in parallel;\n\nend\n\nThe strategy would prefer that the population eliminate its dull state rather than make it degenerate, which is an effective means to commence a new search.\n\nSimulations and analyses {#sec010}\n========================\n\nWe used the orthodox ER random \\[[@pone.0193827.ref046]\\], SF \\[[@pone.0193827.ref047]\\], SW networks of NW type \\[[@pone.0193827.ref048]\\] and some real-world networks as benchmarks to illustrate the feasibility of the PAQGA for optimizing the controllability of arbitrary networks that encompass control nodes and state nodes. Additionally, we also conducted an analysis of the relationship between the network topology and number of control nodes. ER and SF networks were obtained from the static model \\[[@pone.0193827.ref049]\\] with *N* state nodes and *P* candidate control nodes (N = *P*). Each control node pointed to state nodes with uniform probability and the weights of all edges were randomized between zero and one. SW networks were generated from randomized adding edges \\[[@pone.0193827.ref048], [@pone.0193827.ref049]\\]. The characteristics of random regular networks, ER, SF, and SW networks are illustrated in [Fig 8](#pone.0193827.g008){ref-type=\"fig\"}.\n\n![Characteristics of the addressed networks.\\\nRed stars represent the node in-degree denoted by \u27e8k~in~\u27e9 and the green diamonds represent the node out-degree denoted by \u27e8k~out~\u27e9. (a) Random regular networks with homogeneous degree distribution of \u27e8k~in~\u27e9 = \u27e8k~out~\u27e9 = 4. (b) ER random networks with Poisson degree distribution; the degree heterogeneities rely on the average degree denoted by \u27e8k\u27e9. (c) SF networks with power-law degree distribution, which results in large degree heterogeneities. (d) SW networks with long-tail degree distribution, which decreases much slower than the SF distribution.](pone.0193827.g008){#pone.0193827.g008}\n\nWe define the number of selected control nodes that correspond to the current best control scheme as n~*c*~ and the density of these selected control nodes as N~*c*~, where N~*c*~ = n~*c*~/*N*. The minimum number of selected control nodes after the optimization process is denoted as n~*cm*~, and the minimum control node density is N~*cm*~ = n~*cm*~/*N*. To implement the parallel strategy, we performed the following simulations on eight MATLAB^\u00ae^ workers. The parameters of the PAQGA were set to 2m = 30, maxgen = 100, and *p*~*c*0~ = 0.06. All the following experimental results are the average of 10 independent simulations and the standard deviation is 0.01.\n\nPerformance of the PAQGA {#sec011}\n------------------------\n\nTo show that the optimal solution (*D*~*best*~) at each generation always evolves in the feasible region, that is, *Pen*~*i*~(*D*~*best*~) = 0, we conducted experiments on different networks. All these networks were directed with 100 state nodes and 100 candidate control nodes. The experimental results are shown in [Fig 9](#pone.0193827.g009){ref-type=\"fig\"}. The figure shows that the best fitness quickly converged to the minimum value after approximately the first few generations. The mean current fitness fluctuated dramatically because of the operations of qubit cross, qubit catastrophe, and Q-gate rotation. The penalty was always equal to zero, which means that *D*~*best*~ always met the PBH rank condition throughout the entire optimization process.\n\n![Fitness and penalty curves as a function of iterating generation for (a) ER with \u27e8k\u27e9 = 4.0, (b) SF with \u27e8k\u27e9 = 4.0 and \u03b3 = 2.1, and (c) SW with \u27e8k\u27e9 = 4.0. The red dotted line with a square is the best fitness corresponding to D~best~ at the current generation, the blue dashed line with a circle is the mean fitness of all control schemes at each generation, and the black line with a triangle is the penalty term corresponding to D~best~ at each generation.](pone.0193827.g009){#pone.0193827.g009}\n\nWe compare the performance of PAQGA of optimizing network controllability with that of EO \\[[@pone.0193827.ref022]\\] and adaptive GA \\[[@pone.0193827.ref028]\\] on a list of popular networks and real-life networks. The comparison results are tabulated in [Table 2](#pone.0193827.t002){ref-type=\"table\"}.\n\n10.1371/journal.pone.0193827.t002\n\n###### Performance comparison of PAQGA, GA, and EO on different networks in terms of *n*~*cm*~, the minimum iterating generations, and computational time.\n\nPower-law index of SF networks in these experiments was \u03b3 = 2.1. '/' indicates that the corresponding results were not available for the computational time limit. For data sources, see Supplementary information [S1 Dataset](#pone.0193827.s001){ref-type=\"supplementary-material\"}.\n\n![](pone.0193827.t002){#pone.0193827.t002g}\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  network                                   N/P    \\<k\\>   PAQGA   Adaptive GA \\[[@pone.0193827.ref028]\\]   EO \\[[@pone.0193827.ref022]\\]                                  \n  ----------------------------------------- ------ ------- ------- ---------------------------------------- ------------------------------- ---- ----- --------- ---- ---- ---------\n  ER                                        25     1.5     2       3                                        1.86                            2    30    0.23      3    22   0.45\n\n  ER                                        50     3       3       4                                        4.03                            3    34    4.19      3    22   7.79\n\n  ER                                        100    4       5       5                                        7.38                            5    75    266.26    6    24   121.62\n\n  ER                                        200    6       31      5                                        12.49                           31   128   873.05    31   29   1545.84\n\n  ER                                        300    8       22      11                                       18.31                           25   44    1708.81   /    /    /\n\n  ER                                        500    10      43      17                                       31.82                           /    /     /         /    /    /\n\n  ER                                        1000   16      64      26                                       65.05                           /    /     /         /    /    /\n\n  SF                                        25     1.5     3       5                                        2.36                            3    20    0.43      4    22   0.75\n\n  SF                                        50     3       5       6                                        5.38                            5    32    4.83      5    21   8.37\n\n  SF                                        100    4       10      7                                        8.22                            10   69    278.55    10   21   134.54\n\n  SF                                        200    6       35      5                                        13.06                           35   116   892.16    36   34   1623.13\n\n  SF                                        300    8       83      11                                       19.57                           83   132   1823.12   /    /    /\n\n  SF                                        500    10      192     17                                       34.92                           /    /     /         /    /    /\n\n  SF                                        1000   16      416     26                                       68.04                           /    /     /         /    /    /\n\n  Rhode \\[[@pone.0193827.ref050]\\]          20     2.65    2       5                                        2.76                            2    20    0.52      2    22   0.86\n\n  Maspalomas \\[[@pone.0193827.ref050]\\]     24     3.417   3       6                                        5.56                            3    32    5.24      3    21   8.93\n\n  Michigan\\                                 39     5.667   13      7                                        6.45                            13   69    5.52      14   21   12.62\n  \\[[@pone.0193827.ref050]\\]                                                                                                                                               \n\n  Circuit-s208 \\[[@pone.0193827.ref051]\\]   122    3.126   29      9                                        18.22                           29   116   913.14    30   34   1745.83\n\n  Friend-rev\\                               228    4.01    52      10                                       20.45                           52   121   1201.54   54   45   2733.61\n  \\[[@pone.0193827.ref052]\\]                                                                                                                                               \n\n  Circuit-s420 \\[[@pone.0193827.ref051]\\]   252    3.21    59      13                                       23.65                           59   132   1962.92   /    /    /\n\n  Circuit-s838\\                             512    3.44    119     18                                       39.03                           /    /     /         /    /    /\n  \\[[@pone.0193827.ref051]\\]                                                                                                                                               \n\n  Roget \\[[@pone.0193827.ref050]\\]          1022   4.966   396     27                                       75.66                           /    /     /         /    /    /\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nFrom the columns of *n*~*cm*~, it can be observed that the three algorithms almost converged to the same value, which demonstrates that PAQGA, GA, and EO all had a good ability to determine the optimal control nodes. When the size of the network was small (e.g., N = P \u2264 50), PAQGA took slightly more time than GA and EO to determine the best solution. This paradoxical phenomenon is attributed to the launching of the MATLAB^\u00ae^ distributed server, and the launching time was approximately 2s. However, once the server started, PAQGA showed a greater advantage in processing large-size networks over GA and EO. For example, for the ER network with \u27e8*k*\u27e9 = 6.0 and N = P = 200, PAQGA obtained *n*~*cm*~ at the fifth generation and took 12.49s; for the same network, GA took 873.05s at the 128^th^ generation and, and EO required 29 generations and 1545.84 s. By comparing the computational time, PAQGA saved 98.57% more than GA and 99.19% more than EO.\n\nA parallel version of GA was transformed from the adaptive GA \\[[@pone.0193827.ref028]\\] using *\u03b7* MATLAB\u00ae workers with the computational complexity of $O\\left( \\frac{2m \\times \\left( {N + P} \\right)^{3} \\times l}{\\eta} \\right)$, where 2m is the population size, *l* is the number of different eigenvalues of the controlled network. We compared it with the proposed PAQGA, and the results are shown in [Table 3](#pone.0193827.t003){ref-type=\"table\"}.\n\n10.1371/journal.pone.0193827.t003\n\n###### Performance comparison of PAQGA and parallel GA on different networks using eight MATLAB\u00ae workers in terms of n~cm~, the minimum iterating generations, and computational time.\n\nFor data sources, see Supplementary information [S1 Dataset](#pone.0193827.s001){ref-type=\"supplementary-material\"}.\n\n![](pone.0193827.t003){#pone.0193827.t003g}\n\n  -----------------------------------------------------------------------------------------------------------\n  network                                   N/P    \\<k\\>   PAQGA   parallel GA                       \n  ----------------------------------------- ------ ------- ------- ------------- ------- ----- ----- --------\n  ER                                        25     1.5     2       3             1.86    2     19    2.71\n\n  ER                                        50     3       3       4             4.03    3     28    5.19\n\n  ER                                        100    4       5       5             7.38    5     35    53.31\n\n  ER                                        300    8       22      11            18.31   25    37    89.54\n\n  SF                                        25     1.5     3       5             2.36    3     18    3.52\n\n  SF                                        50     3       5       6             5.38    5     27    6.65\n\n  SF                                        100    4       10      7             8.22    10    34    42.73\n\n  SF                                        300    8       83      11            19.57   83    126   243.12\n\n  Rhode \\[[@pone.0193827.ref050]\\]          20     2.65    2       5             2.76    2     18    3.51\n\n  Maspalomas \\[[@pone.0193827.ref050]\\]     24     3.417   3       6             5.56    3     28    8.32\n\n  Michigan\\                                 39     5.667   13      7             6.45    13    66    11.47\n  \\[[@pone.0193827.ref050]\\]                                                                         \n\n  Circuit-s208 \\[[@pone.0193827.ref051]\\]   122    3.126   29      9             18.22   29    98    220.78\n\n  Friend-rev\\                               228    4.01    52      10            20.45   52    107   275.67\n  \\[[@pone.0193827.ref052]\\]                                                                         \n\n  Circuit-s420 \\[[@pone.0193827.ref051]\\]   252    3.21    59      13            23.65   59    116   295.63\n\n  Circuit-s838\\                             512    3.44    119     18            39.03   119   122   413.41\n  \\[[@pone.0193827.ref051]\\]                                                                         \n\n  Roget \\[[@pone.0193827.ref050]\\]          1022   4.966   396     27            75.66   396   135   511.76\n  -----------------------------------------------------------------------------------------------------------\n\nIt can be inferred from [Table 3](#pone.0193827.t003){ref-type=\"table\"} that the parallel computation (allowing for multiple processors) contributes to the performance of algorithms. However, it is not the only important factor. The computational efficiency of EO, GA, parallel GA and PAQGA could be reflected by their computation complexity. First, the computation of the PBH rank matrix in GA, parallel GA and PAGQA and the Kalman rank matrix in EO is the most time-consuming. This is the main factor affecting their speedability. The rank computation of Kalman matrix takes much more time than that of PBH matrix. Second, PAQGA adopts qubits representation, where each chromosome contains two individuals. This expands the space of feasible solution. And the adaptive Q-gate rotation operation and crossover operation help to improve the algorithm efficiency.\n\nMoreover, comparison tests among the PAQGA, MMT and MM are conducted to observe the performance of the PAQGA on much larger real networks. The experimental results are shown in [Table 4](#pone.0193827.t004){ref-type=\"table\"}.\n\n10.1371/journal.pone.0193827.t004\n\n###### Performance comparison of PAQGA, MMT, and MM on several large real-directed, -weighted and--unweighted networks in terms of N~cm~ and computational time.\n\nFor data sources, see Supplementary information [S1 Dataset](#pone.0193827.s001){ref-type=\"supplementary-material\"}.\n\n![](pone.0193827.t004){#pone.0193827.t004g}\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------\n  network                                class        N/P     PAQGA    MMT \\[[@pone.0193827.ref011]\\]   MM \\[[@pone.0193827.ref010]\\]                     \n  -------------------------------------- ------------ ------- -------- -------------------------------- ------------------------------- -------- -------- --------\n  Coauthorships\\                         Directed\\    1461    0.3436   85.12                            0.3436                          67.03    0.3425   34.16\n  \\[[@pone.0193827.ref053]\\]             weighted                                                                                                         \n\n  SciMet \\[[@pone.0193827.ref054]\\]      Directed\\    2729    0.4251   126.35                           0.4251                          83.29    0.4236   52.91\n                                         unweighted                                                                                                       \n\n  Kohonen \\[[@pone.0193827.ref054]\\]     Directed\\    3772    0.562    173.46                           0.562                           106.62   0.5604   73.85\n                                         unweighted                                                                                                       \n\n  Wiki-Vote \\[[@pone.0193827.ref055]\\]   Directed\\    7115    0.6656   392.44                           0.6656                          228.73   0.6656   167.59\n                                         unweighted                                                                                                       \n\n  P2P-3 \\[[@pone.0193827.ref056]\\]       Directed\\    8717    0.5778   473.19                           0.5778                          279.15   0.5774   206.52\n                                         unweighted                                                                                                       \n\n  P2P-1 \\[[@pone.0193827.ref056]\\]       Directed\\    10876   0.5531   685.78                           0.5531                          359.83   0.552    268.96\n                                         unweighted                                                                                                       \n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nIt can be seen that N~*cm*~ of PAQGA agrees with that of MMT on these real-world networks, and is slightly greater than or equal to that of MM. The experimental results show the efficiency of PAQGA in identifying the minimum control nodes. Nevertheless, PAQGA is at a disadvantage in computational time compared with MMT and MM, although such defect could be improved by adding the number of processors or using computer groups. For example, the cost of calculating N~*cm*~ of Wiki-vote network is reduced to 301.34s with 12 processors.\n\nPAQGA is an intelligent probabilistic optimization algorithm that provides approximate solutions. The optimal solution cannot be guaranteed to be found. Moreover, the optimal solution of a problem is typically unknown in advance. We used the structural controllability theory \\[[@pone.0193827.ref010], [@pone.0193827.ref018]\\] as the benchmark to compute *n*~*cm*~ to test the validation of PAQGA. The results are shown in [Fig 10](#pone.0193827.g010){ref-type=\"fig\"}.\n\n![n~cm~ comparison between the structural controllability theory and PAQGA on (a) ER with \u27e8k\u27e9 = 4.0, (b) SF with \u27e8k\u27e9 = 4.0 and \u03b3 = 2.1, and (c) SW with \u27e8k\u27e9 = 4.0.](pone.0193827.g010){#pone.0193827.g010}\n\nWe can observe that for ER, SF, and SW, the obtained *n*~*cm*~ is the same as the benchmark result, which indicates that the proposed PAQGA was effective in determining the minimum control nodes of complex networks.\n\nDiscussion and analysis of results {#sec012}\n----------------------------------\n\nApplying PAQGA, the optimization results and evolution process of network topology can be achieved. The results are intuitively displayed in [Fig 11](#pone.0193827.g011){ref-type=\"fig\"}.\n\n![PAQGA optimization results and network topology evolution.\\\n(a) Convergence trend of N~c~ of directed ER, SF, and SW with N = P = 100. (b) Initial network topology (at the zeroth generation) of SW with \u27e8k\u27e9 = 6.0. Yellow circles represent the candidate control nodes and green squares represent the state nodes. Selected control nodes are connected to state nodes with a row from circles to squares. Links between state nodes are arrowed lines between squares. (c) Network topology guided by 31 control nodes at the fifth generation. (d) Network topology with 22 control nodes at the first convergence generation (seventh generation).](pone.0193827.g011){#pone.0193827.g011}\n\nFrom [Fig 11(A)](#pone.0193827.g011){ref-type=\"fig\"}, N~c~ of different networks quickly converged to a steady minimum value, which indicates the effectiveness of the PAQGA. For example, *N*~*c*~ of ER with \\<k\\> = 4.0 converged to 0.05 at the fifth generation, which demonstrates that five control nodes were sufficient to maintain network controllability. For SF with \\<k\\> = 6.0 and \u03b3 = 2.1, *N*~*c*~ rapidly decreased to a minimum value of 0.07 at the seventh generation, which means that at least seven control nodes were required to fully control the network. [Fig 11(B)--11(D)](#pone.0193827.g011){ref-type=\"fig\"} together capture the evolution of the SW network with \\<k\\> = 6.0 at the zeroth, fifth, and seventh iteration, and the convergence trend of the control nodes can be acquired.\n\nWe also found that two networks with different \\<k\\> required different *N*~*cm*~. For example, for ER with \\<k\\> = 4.0 and \\<k\\> = 6.0, *N*~*cm*~ of the network with \\<k\\> = 4.0 was 0.05 and with \\<k\\> = 6.0, *N*~*cm*~ = 0.02. Additionally, *N*~*cm*~ of SW networks with \\<k\\> = 4.0 and \\<k\\> = 6.0 was 0.25 and 0.22, respectively. Second, for networks with the same \\<k\\> and different \u03b3, *N*~*cm*~ also differed. For example, consider SF with same \\<k\\> = 6.0, and different \u03b3 = 2.1 and \u03b3 = 3.0. The two networks had *N*~*cm*~ = 0.06 and *N*~*cm*~ = 0.04, respectively. Third, for networks with the same \u03b3 and different \\<k\\>, N~cm~ was also different, which can be determined from [Table 2](#pone.0193827.t002){ref-type=\"table\"}. These results led us to conjecture that *N*~*cm*~ had a relationship with \\<k\\> and \u03b3.\n\nTo confirm our hypothesis, we performed simulations on a set of different networks and plotted *N*~*cm*~ as the function of \\<k\\> and \u03b3. The results are shown in [Fig 12](#pone.0193827.g012){ref-type=\"fig\"}.\n\n![Impact of \\<k\\> and \u03b3 on *N*~*cm*~.\\\n(a) N~cm~ as a function of \\<k\\> with fixed \u03b3. (b) *N*~*cm*~ as a function of \u03b3 with fixed \\<k\\>. Networks are directed with N = P = 500.](pone.0193827.g012){#pone.0193827.g012}\n\nFrom [Fig 12(A)](#pone.0193827.g012){ref-type=\"fig\"}, it is obvious that *N*~*cm*~ of networks with fixed \u03b3 decreased monotonically with \\<k\\> until *N*~*cm*~ became slowly flat. Additionally, the downward trend was of asymptotic exponential dependence, which suggests that the sparse network required more control nodes to maintain full controllability. From [Fig 12(B)](#pone.0193827.g012){ref-type=\"fig\"}, we can observe that N~cm~ with fixed \\<k\\> decreased as \u03b3 increased. The results indicate that N~cm~ may be influenced by the degree heterogeneity, denoted by H, which is the standard deviation of the network node degree distribution \\[[@pone.0193827.ref057]\\]. In this paper, H is defined as $$H = \\left( {\\sum_{i}\\left( {k_{i} - \\left. \\left\\langle k \\right\\rangle \\right)^{2}} \\right)}/N \\right)^{1/2},\\ i = 1,2,\\ldots,N,$$ where k~i~ is the degree of state node i.\n\nTo determine the relationship between N~cm~ and H, we examined N~cm~ as a function of H and obtained the results shown in [Fig 13](#pone.0193827.g013){ref-type=\"fig\"}.\n\n![N~cm~ as a function of H.\\\n(a) N~cm~ as a function of H for ER and SF networks with fixed \u03b3 and variable \u27e8k\u27e9. (b) N~cm~ as a function of H for ER, SF, and SW networks with variable \u03b3 and fixed \u27e8k\u27e9. The networks are directed with N = P = 500.](pone.0193827.g013){#pone.0193827.g013}\n\nFrom [Fig 13(A)](#pone.0193827.g013){ref-type=\"fig\"}, it can be observed that a larger N~cm~ always corresponded to a larger H and smaller \u03b3. [Fig 13(B)](#pone.0193827.g013){ref-type=\"fig\"} shows that the network with a smaller \u27e8k\u27e9 and larger H typically required a larger N~cm~. The results suggest that the larger the differences between node degrees, the more control nodes were required to entirely control the network.\n\nSW networks have the remarkable characteristics of a large clustering coefficient, denoted by C, which represents the overlapping degree of friend circles of two adjacent state nodes and is defined as $$C = \\frac{1}{N}{\\sum_{i}\\frac{E_{i}}{\\frac{1}{2}k_{i}\\left( k_{i} - 1 \\right)}},\\ i = 1,2,\\ldots,N,$$ where i is node i, k~i~ is the number of edges between node i and other nodes, and E~i~ is the number of edges among the k~i~ nodes. For SW networks, N~cm~ may be affected by C. To explore the relationship between N~cm~ and C, we plot N~cm~ as a function of \u27e8k\u27e9 and C,shown in [Fig 14](#pone.0193827.g014){ref-type=\"fig\"}.\n\n![Impact of \\<k\\> and C on N~cm~ of SW networks.\\\n(a) N~cm~ as a function of \\<k\\> with fixed C. When C = 1, the network is fully connected and can be steered to any state with only one controller. (b) N~cm~ as a function of C with fixed \u27e8k\u27e9. Networks are directed with N = P = 500.](pone.0193827.g014){#pone.0193827.g014}\n\nFrom [Fig 14(A) and 14(B)](#pone.0193827.g014){ref-type=\"fig\"}, we can determine that a larger C corresponded to a smaller N~cm~, which indicates that the more interconnected the network, the fewer control nodes were required to control the network. For other networks, such as ER, SF, the conclusion also holds.\n\nConsidering the aforementioned analysis results together, we can determine that for a given network with both control nodes and state nodes: 1) the sparser the network, the more control nodes were required to control it; and 2) the more heterogeneous the network, the more control nodes were required to guarantee its full control. We reflect that the sparse and heterogeneous network is the most difficult for guiding its dynamic evolution (see Tables [2](#pone.0193827.t002){ref-type=\"table\"} and [4](#pone.0193827.t004){ref-type=\"table\"} and Figs [10(A)](#pone.0193827.g010){ref-type=\"fig\"} and [12(B))](#pone.0193827.g012){ref-type=\"fig\"}. The consistency between the results from our approach and from these existing methods \\[[@pone.0193827.ref010], [@pone.0193827.ref011], [@pone.0193827.ref022], [@pone.0193827.ref028]\\] confirms the similarity between them for directed networks, which not only further validate these existing methods, but also reflect the effectiveness of our method.\n\nTo evaluate the controllability of directed networks, the structural controllability framework based on the MM method is still the best for its error-free feature \\[[@pone.0193827.ref011]\\]. Like the MMT, the PAQGA also relies on the eigenvalues and the rank of the network matrix, the computation of which inevitably introduces numerical errors. Further, MM and MMT both surpass PAQGA in computational efficiency in identifying both the minimum set of driver nodes and the number of these driver nodes. However, the PAQGA can have a wider range of applications. For example, the PAQGA is valid for networks containing a number of self-loops with identical or different weights, and networks with bidirectional connections between two nodes. The PAQGA is also applicable to undirected networks, where the structural matrix assumption is slightly violated because of the network symmetry. Further, combined with advantages of computer hardware and the adaptive strategies itself, PAQGA has great room for improvement. Taken together, the PAQGA as an alternative exact structural controllability framework provides us deeper understanding of the controllability of complex networked systems.\n\nConclusions {#sec013}\n===========\n\nIn this paper, we introduced a PAQGA to optimize the controllability of arbitrary networks with control nodes and state nodes under the PBH rank condition. In addition to MATLAB\u00ae workers, more parallel mechanisms can be flexibly embedded in the PAQGA, for which more threads concurrently processing could further promote the time efficiency of generating a solution. Analyses and simulation comparisons demonstrated the effectiveness and applicability of the proposed PAQGA. Furthermore, we found that the minimum control nodes were affected by the network degree distribution, degree heterogeneity, and clustering coefficient. The sparse and heterogeneous network is the most difficult to be fully controlled.\n\nIn our study, the topology that comprises state nodes remained static during the entire evolution process. However, networks normally evolve over time, which manifests as the increasing or decreasing of different nodes and their links. In the future, we will focus on the controllability of dynamic networks. Furthermore, we hope to explore how to use the obtained minimum control nodes to steer an intermediate network to evolve into our desired network considering realistic energy constraints.\n\nSupporting information {#sec014}\n======================\n\n###### Canonical and real-world network datasets for comparison experiments.\n\n(RAR)\n\n###### \n\nClick here for additional data file.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "A diagnosis of cancer can cause great suffering to patients and families. The revealing of the diagnosis of cancer to patients is a key event in their cancer journey. The word cancer is often avoided in these consultations ([Thomsen *et al*, 1993](#bib20){ref-type=\"other\"}). Over recent years, communication and information have increasingly been considered important in helping people with cancer ([Fallowfield *et al*, 1994](#bib11){ref-type=\"other\"}; [Coulter, 1998](#bib5){ref-type=\"other\"}). Research indicates that the vast majority of cancer patients want to be informed of their illness ([Meredith *et al*, 1996](#bib16){ref-type=\"other\"}). Women with ovarian cancer need honest communication that is appropriate to their level of understanding. Communication needs will vary across patient\\'s age, stage of disease and treatment. At present, there are no minimal legal recommendations for documenting such consultations. Furthermore, complaints made by patients often focus on a perceived failure of communication rather than on clinical errors ([Department of Health, 2000](#bib7){ref-type=\"other\"}).\n\nIn its guide 'Good Clinical Practice' ([General Medical Council, 2001](#bib14){ref-type=\"other\"}), the General Medical Council (GMC) states 'a good medical record should contain sufficient information to: identify the patient; support the diagnosis; justify the treatment; document the course and results and promote continuity of care among healthcare providers', and continues, 'Doctors must keep colleagues well informed when sharing the care of patients. Without good notes, this is impossible'.\n\nExperts agree that improved communication between health professionals and cancer patients is essential for the delivery of high-quality care ([Department of Health, 2000](#bib7){ref-type=\"other\"}). As one in three people will be diagnosed with cancer during their lifetime, investigation of this influential consultation is vital to improving and monitoring the service we as clinicians provide. We report here an observational survey of the quality of information relating to giving the diagnosis recorded in the hospital case notes of patients with ovarian cancer.\n\nMATERIAL AND METHODS\n====================\n\nWe reviewed the Hospital records of patients diagnosed with epithelial ovarian cancer in the Mersey Area between 1992 and 1994. The appropriate authorisation for the study was obtained from the individual consultants in the Hospitals audited.\n\nInformation and recurrent themes relating to communications or interviews following the diagnosis of epithelial ovarian cancer were identified in the medical and nursing notes. The themes were: 'diagnosis' words, 'prognosis' words, further treatment, emotional response, patients\\' understanding, information-seeking behaviour and evidence of collusion. Descriptive words were recorded verbatim. Data extraction was performed by three of the authors independently (JMK, DGT and TL); any discrepancies were resolved in discussion. Collusion was defined as any entry recording discussion of the diagnosis or prognosis between relatives and medical or nursing staff where the patient was actively excluded.\n\nThe following demographic data were also extracted from the case notes: age, hospital, postcode, year of surgery, surgeon, stage of disease and debulking achieved, length of survival and preoperative suspicion of ovarian cancer. The International Federation of Gynaecology and Obstetrics (FIGO) stage ([FIGO Cancer Committee, 1986](#bib12){ref-type=\"other\"}) was obtained by review of the operation notes and histology by one of the authors (JMK). We calculated the underprivileged area score (UPAS) using the patients\\' postcode ([Jarman, 1993](#bib15){ref-type=\"other\"}).\n\nData were stored on a statistical software package for the social sciences (SPSS version 10, SPSS Inc., Chicago IL, USA). The frequency of recorded information and themes were compared to the demographic variables using *\u03c7*^2^ test, Fisher\\'s exact test, Mann--Whitney *U* test or Kruskall--Wallis as appropriate. Continuous data are presented as median (range) or mean (s.d.). Significance was set at 1%, taking into account Bonferroni corrections.\n\nRESULTS\n=======\n\nWe identified 359 patients with a histological diagnosis of epithelial ovarian cancer. There were 90 stage 1, 45 stage 2, 132 stage 3, 36 stage 4 patients and 56 patients where it was not possible to stage the disease. The mean age at diagnosis was 62.7 (13.6) years. A total of 331 patients underwent primary surgical debulking and 28 were diagnosed on either cytology of ascites and/or pleural fluid or a pelvic mass on imaging and a raised CA125. Follow-up to 5 years was available for all patients: 78.3% of patients survived 6 months, 55.1% 18 months and 29.1% 5 years. In 34 (9.5%) patients, there was no relevant information recorded in the case notes, and four patients were demented, so the following analyses are presented on the 321 patients where data were available. The diagnosis was recorded in 304 cases (94.7%), prognosis was documented in 66 cases (20.6%), emotional response in 103 (32.1%), patient understanding in 62 (19.3%), information seeking behaviour in 42 (13.1%), further treatment 216 (67.3%) and evidence of collusion with the relatives in 33 cases (10.3%).\n\nExamples from the notes are shown below. 'Findings discussed''Pt informed of probable diagnosis of ovarian neoplasm, grateful for being informed''Doctor thought patient was told of operation findings, pt speaks of cyst and that is all that was mentioned, Dr used words tumour, growth, but not malignant or cancer''Patient does not want family to know her condition she will tell them in her own time''Patient feels it has not registered with her what has happened''Histology explained to patient and mum both state they understand''Spoken with son and daughter told inoperable tumour they will discuss with their father and decide how much to tell their mother'\n\nThe person or persons present at the time of the consultation is show in [Table 1](#tbl1){ref-type=\"table\"}Table 1Person(s) present**Person(s) present*n*%**No-one12739.6Husband/partner6319.6Generic 'family'278.4Not recorded268.1Daughter195.9Other family82.5Son72.2Sister/brother51.6Parents41.2Friend, requested no family20.6Collusion^a^3310.3[^1]. In over one-third of cases (39.6%), the patient was unaccompanied. The husband/partner was the commonest accompanying person (19.6%), who was often present with other members of the family (8.4%). Daughter(s) were more than twice as likely to be present than son(s). In one patient, the consultation was requested without the presence of the family. Collusion between family members and medical/nursing staff was documented in 33 cases (10.3%).\n\nThe diagnostic words and phrases recorded in the notes of the 304 patients are shown in [Table 2](#tbl2){ref-type=\"table\"}Table 2Diagnostic words recorded in the 304 patients where there was an entry**Diagnostic words recorded*n*%**Operation8427.5Diagnosis7925.8Histology6822.2Cancer6019.6Findings4514.7Tumour3511.4Malignant3411.1Cyst123.9Disease, mass103.3Carcinoma, situation72.3Results, procedure, lump, procedure and condition41.3Primary, growth and serious problem31.0Problem, abnormal cells, could not remove/separate, disease remains20.7Cytology, cells, looked odd, microscopic deposits, swelling, not simple, nasty, benign, secondaries, abnormal, neoplasm, tissue diagnosis, metastases, all removed, macroscopic clearance, suspicious, lesion and adherent10.3. A total of 42 separate words or phrases were identified relating to diagnosis, the commonest were generic words such as 'operation', 'diagnosis' and 'histology'. In 60 patients (19.6%), the word cancer was specifically recorded. There were many words that occurred on three or fewer occasions; however, the vast majority were recorded in conjunction with other words in the table such as 'operation' and 'diagnosis'. There were 42 cases of only one word being recorded.\n\nDocumentation of prognosis given to the patient at the time of the consultation was present in only 66 women (20.6%). For these women the words used are shown in [Table 3](#tbl3){ref-type=\"table\"}Table 3Prognosis word recorded in 66 women where there was an entry**Prognosis recorded*n*%**Poor1615.2Palliative65.7Extent of disease43.8Progression, TLC, good outlook, inoperable, unable to remove all32.9Spread of disease, guarded, cure impossible21.9Surgery enough, all being well should have response, uncertainty of response, cure unlikely, advanced, keep at bay, early stage, deterioration, likely prognosis, all away, should be fine, terminal, bleak, extremely poor, too far gone, nature take its course, no further treatment, bad, optimistic, untreatable11.0. There were 31 separate words or phrases used to describe prognosis. The commonest prognosis word recorded was 'poor', recorded on 16 occasions (15.2%). Further treatment was recorded in 216 patients (67.7%), and is detailed for these patients in [Table 4](#tbl4){ref-type=\"table\"}Table 4Further treatment recorded in the 216 women where there was an entry**Further treatment*n*%**None11335.3Chemotherapy4915.3Further treatment278.4Undecided123.7Surgery113.4Palliative/hospice30.9Radiotherapy10.3\u00a0\u00a0\u00a0Total216100. In 113 patients (35.3%), no further treatment was recorded, and in 49 (15.3%) chemotherapy was indicated.\n\nPractitioners recorded the response to diagnosis and/or prognosis on 137 patients. Thematic analysis revealed that these responses fell into one of three categories: emotional response, understanding and information-seeking behaviour.\n\nThe most commonly reported words recorded following initial diagnosis/prognosis were 'anger', 'upset' and 'distressed'. However, there was a continuum of emotional responses recorded from 'feeling positive' and 'indifference' to 'contemplating suicide' and 'wishing to die'. Patients were recorded as being 'frightened', 'depressed' and 'abandoned'. Others were reported as being 'in denial' or 'not unduly concerned'.\n\nPractitioners recorded their perception of the level of understanding of the information provided. Words commonly used were 'accepting', 'understands' and 'realistic'. Records describing the patients understanding were often followed by a brief description of a coping strategy, such as 'patient going to battle on making disease fit round her life, 'taking one day at a time' or 'making the most of the time she has left'.\n\nOne of the frequently reported reactions to the diagnosis/prognosis was to seek more information. A dichotomy of reactions was recorded. Records suggested that some women were not ready to receive further information at the initial session. Entries included 'she does not want to talk about the future and what may happen', 'does not want to ask any more questions' and 'feels it unnecessary to discuss condition'. Other records highlighted the need to seek immediate information, such as 'wanted to ask more questions' and 'asked to see oncologist'.\n\nThere were 145 women (45.2%) over 65 years of age, and 176 women (54.8%) under 65 years. There was a significant reduction in the number of diagnostic words recorded in the over 65\\'s compared to the under 65\\'s (90.3 *vs* 98.3%, *P*=0.002, Fisher\\'s exact test). Collusion was more than three times as common in the over 65\\'s (17.9 *vs* 5.7%, *P*=0.001, Fisher\\'s exact test). Emotional response was more frequently recorded in the under 65\\'s (38.1 *vs* 26.2%, *P*=0.024, *\u03c7*^2^ test). There was no difference in prognosis, understanding or information-seeking behaviour recorded.\n\nThe overall 5-year survival was 29.1%. There was no relation between survival at 6 or 18 months and any themes, categories or variables.\n\nThe UPAS was calculated for each patient using her postcode. The median score was 3.68, range --99.00 to 62.00. There was no relation between UPAS and any themes, categories or variables.\n\nIn 56 of the 359 cases (15.6%), it was not possible to stage the patient because of poor note keeping. In these 56 patients, there was a significant reduction in the diagnostic words being recorded in this group compared to all other stages. A 26.8% had no information at all (*vs* 6.3%, *P*\\< 0.0001) and 41.1% had no diagnostic words (*vs* 9.9%, *P*=0.001).\n\nA general gynaecologist operated on 177 patients (55.1%), a special interest gynaecologist 61 (19.0%), a general surgeon 28 (8.7%), an SpR in gynaecology in 37 (11.5%). A total of 18 patients (5.6%) were not operated on. The frequency with which diagnostic words were recorded in the notes differed significantly depending upon the surgeon who performed the operation: special interest gynaecologists 100%, SpR 97.3%, general gynaecologist 95.5% and general surgeon 89.3% (*P*=0.001, Fisher\\'s exact test).\n\nCollusion occurred significantly more often in women who did not have an operation compared to those who did (*P*=0.0001, *\u03c7*^2^ test). There was no difference in the incidence of collusion according to type of surgeon.\n\nDISCUSSION\n==========\n\nTo our knowledge, this is the first study to specifically investigate information recorded in the hospital case notes following the diagnosis of ovarian cancer. These are unique data with demographic information, 5-year survival and serve to highlight several important issues in the management of ovarian cancer patients. Although data were obtained for this study from patients with epithelial ovarian cancer, the results are likely to be relevant to other cancer patients.\n\nCancerBACUP ([BACUP, 1996](#bib2){ref-type=\"other\"}) and other cancer organisations recommend that patients have a member of family or a close friend present when bad news is given. In our study, over one-third of patients were often told essential information on their own. If a relative was present, it was most commonly the partner/husband. However, this is a group of women with a mean age of 62.7 years and a proportion will be single, divorced or separated or their partner may not be fit enough to travel to hospital. When children were present, daughters were more than twice as likely to be present than sons.\n\nMost studies show that lay populations have a universal dread of cancer ([Fallowfield, 1997](#bib9){ref-type=\"other\"}). The word 'cancer' therefore, with all its connotations and meaning, is stressful for both patient and doctor alike. A doctor\\'s failure to employ accurate terminology squanders an ideal opportunity to correct misconceptions about the disease. Furthermore, euphemisms such as lump, growth, serious problem (as recorded in [Table 3](#tbl3){ref-type=\"table\"}) are confusing and unhelpful. Such euphemisms only serve to reinforce how awful the disease cancer really is as the doctor is unable to use the word 'cancer'. In our study, 42 separate phrases were employed with the term 'cancer' only recorded in 18% of patients. Our study shows that a large proportion of medical staff still uses these ambiguous terms. Moreover, with an average reading age of 91/2 years ([Department for Education and Employment, 1999](#bib6){ref-type=\"other\"}), few patients would have fully understood the full meaning of many of these words.\n\nPatients now want to know the truth ([Meredith *et al*, 1996](#bib16){ref-type=\"other\"}). Despite the advances in information available to patients, many doctors still unwittingly hurt their patients while trying to protect them by withholding information ([Thomsen *et al*, 1993](#bib20){ref-type=\"other\"}). However, few would state that they actively withhold the diagnosis of cancer from their patients. In our study, there was evidence of collusion in over 10% of patients. Active collusion was twice as common in the over 65\\'s, a group who are more vulnerable as they are less likely to question the doctors\\' decisions ([Nordin *et al*, 2001](#bib17){ref-type=\"other\"}). This is despite evidence that the elderly not only want to be given the same information as the young ([Ganz, 1997](#bib13){ref-type=\"other\"}), but also want access to the same radical treatment and same chance of disease cure ([Nordin *et al*, 2001](#bib17){ref-type=\"other\"}).\n\nGeneral and special interest gynaecologists operated on the majority of patients in the study. There were significantly more diagnostic words recorded by special interest gynaecologists compared to general gynaecologists or general surgeons. This is to be expected, as they are more used to dealing with such patients and liaising with multidisciplinary teams.\n\nWe found that the UPAS had no bearing upon any of the information variables we studied. We found this surprising, as anecdotally it appears that patients from higher socioeconomic groups are given more information, partly because they seem better informed and ask more questions. This is confirmed by a review of 16\u2009955 first-time enquirers accessing the CancerBACUP information service, where the users were predominantly middle class ([Boudioni *et al*, 1999](#bib3){ref-type=\"other\"}). It was therefore reassuring to see that all patients in our study were treated the same way.\n\nThe Royal College of Surgeons of England states that Surgeons must, 'Ensure that a record is made by a member of the surgical team of important events and communications to the patient or supporter (for example, prognosis or potential complications) ([The Royal College of Surgeons of England, 2002](#bib19){ref-type=\"other\"}). It was therefore disappointing to find that in 9.5% of patients, there was no recording in the case notes of this consultation. Where a record was made, however, it was not possible to validate the accuracy of the information recorded in the hospital notes in this study. It is entirely possible that the quality of information given was in excess of that recorded in the case notes and that our data simply reflect poor note keeping. We have no way of confirming this. However, since many of the words recorded are evasive and lacking in clear meaning, we suspect that they are indeed a true reflection of the kind of language used with the patients during the consultation. Our data have revealed some fascinating but also disturbing trends, which can only be confirmed and fully examined by a prospective study of information giving, employing taped consultations and qualitative research methods.\n\nEffective clear communication has profound influences on both patients and health-care professionals. It influences the rate of recovery, pain relief and psychological well-being ([Fallowfield *et al*, 1990](#bib10){ref-type=\"other\"}; [Stewart, 1996](#bib18){ref-type=\"other\"}). Moreover, poor communication at the very start of a cancer journey can leave patients anxious, uncertain and generally dissatisfied with their cancer care ([Audit Commission, 1993](#bib1){ref-type=\"other\"}). Good communication skills can be taught, and in the cancer-care arena can be delivered in a variety of ways ([Fallowfield *et al*, 1990](#bib10){ref-type=\"other\"}; [Wilkinson *et al*, 1999](#bib21){ref-type=\"other\"}). As we move towards multidisciplinary team management in all stages of the disease trajectory ([Expert Advisory Group on cancer to the chief medical officers of England and Wales, 1995](#bib8){ref-type=\"other\"}), good communication between individual team members gains importance. Thus, clear written documentation of doctor--patient consultations, through diagnosis, treatment, relapse and terminal illness forms a vital key in total patient management. Our suggested minimum data set that should be recorded in the hospital notes following such a consultation is shown in [Figure 1](#fig1){ref-type=\"fig\"}Figure 1Suggested minimum data set for recording diagnostic conversations.. This recorded information would therefore enable clinicians and other professionals involved in the future care of a patient immediate access to what the patient understands, thus allowing them to build on this knowledge. This would also act as a clear record of the doctor--patient communication at that time in their cancer journey. This will be important in defending litigation cases or as an adjunct when discussing previous treatment with a patient and/or their supporter.\n\nThis study emphasises the importance of high-quality clinical practice coupled with good note keeping, and echoes the recommendations of the Bristol Royal Infirmary Inquiry for accurate note keeping and audio tape recording facilities when an important diagnosis, course of treatment or prognosis is being discussed ([Bristol Royal Infirmary Inquiry, 2001](#bib4){ref-type=\"other\"}).\n\n[^1]: Patient actively excluded from discussion (see text).\n"}
+{"text": "Introduction {#sec1}\n============\n\nLoose\u00a0anagen hair syndrome (LAHS) is characterized by easily extractable anagen hairs that lose the inner and outer root sheath during removal.[@bib1] Patients with uncombable hair syndrome (UHS) have unruly, \"spun-glass\" hair that cannot be combed flat.[@bib2] In 1996, Boyer et\u00a0al[@bib3] reported an instance of LAHS mimicking uncombable hair. In 2005, Lee et\u00a0al[@bib4] reported another instance of LAHS with concurrent features of uncombable hair syndrome. We report another patient with findings suggestive of both syndromes.\n\nCase presentation {#sec2}\n=================\n\nA 10-month-old Middle Eastern girl presented for evaluation of wispy, roughly textured scalp hair. The patient\\'s hair appeared to be growing normally until this rough texture was noted at approximately 6\u00a0months of age. During this first visit at 10\u00a0months of age, areas of roughly textured scalp hair were noted without any appreciable hair loss.\n\nThe patient returned to the clinic at 23\u00a0months of age regarding patchy alopecia most evident on the left side of the scalp. At this time, the patient\\'s scalp hair was noted to have the same wispy, straight, roughly textured hair as at the previous visit but with sparing of the central occipital scalp ([Fig 1](#fig1){ref-type=\"fig\"}, [Fig 2](#fig2){ref-type=\"fig\"}, [Fig 3](#fig3){ref-type=\"fig\"}). Unlike the majority of her scalp hair, the central occipital scalp had curly, softer hair ([Figs 2](#fig2){ref-type=\"fig\"} and [3](#fig3){ref-type=\"fig\"}). During this second visit, it appeared as though the patchy alopecia was secondary to the patient\\'s pulling out her hair.Fig 1A picture taken in the office at 23\u00a0months of age showing unruly, kinked, light-colored hair.Fig 2A picture taken in office at 23\u00a0months of age showing different hair texture in particular areas of the scalp.Fig 3A picture taken in the office at 23\u00a0months of age showing a wool-spun texture over most of the scalp and a different, curly hair texture originating in the occipital scalp.\n\nTwo weeks after the previous evaluation, the patient presented again due to increasing amounts of hair loss. Photos from her mother showed patchy hair loss on the scalp ([Fig 4](#fig4){ref-type=\"fig\"}). The patient\\'s mother also collected a bag of shed hairs for evaluation. No abnormalities of the eyelashes, eyebrows, fingernails, or toenails were noted.Fig 4A picture taken by the patient\\'s mother showing an area of hair loss over the occipital scalp. According to her mother, the patient had been pulling out her hair in this area.\n\nUnder dermatoscopic exam ination, the patient had normal-appearing scalp hair shafts without nodes or trichorrhexis. Her scalp skin appeared normal, without erythema or flaking. Microscopic examination of hairs showed many anagen hairs with the hair bulbs oriented at an acute angle to the hair shaft and a ruffled cuticle that resembled a crumpled sock ([Fig 5](#fig5){ref-type=\"fig\"}). Microscopic examination also showed normal-appearing hair shafts ([Figs 6](#fig6){ref-type=\"fig\"} and [7](#fig7){ref-type=\"fig\"}). No abnormalities were seen pursuant to an examination with polarized filters ([Fig 7](#fig6){ref-type=\"fig\"}). Electron microscopy was not available.Fig 5Easily plucked hair showing anagen bulb and rumpled sock appearance of the shaft.Fig 6Hair sample under simple microscopic examinationFig 7Hair sample under a polarized filter.\n\nThe patient was born at 40\u00a0weeks, and the mother had an uncomplicated pregnancy and vaginal birth. The child has been meeting regular pediatric milestones. The family history is significant for an aunt who has unruly hair. Her parents have no hair issues other than male pattern baldness in the father. The patient has 2 brothers without hair issues.\n\nDiscussion {#sec3}\n==========\n\nOne of the first recorded descriptions of loose anagen hair syndrome was by Zaun in 1987.[@bib5] Patients with LAHS often have thin hair, described as having a tacky or sticky feel, that can be easily extracted from the scalp with traction.[@bib6] Girls between the ages of 2 and 9\u00a0years are most commonly diagnosed with the condition, although it has been proposed that LAHS may be underreported in male patients who are more likely to cut their hair short.[@bib7]^,^[@bib8] Affected patients most commonly have light hair.[@bib1] Both autosomal dominant familial and sporadic forms of syndrome inheritance have been described.[@bib9] Its incidence has been estimated to be 2 to 2.25 occurrences per million per year.[@bib6] Under light microscopy, hairs that are easily pulled from the scalp show ruffling of the hair cuticle, separation of the outer root sheath from the vitreous layer, and grooving of the shaft.[@bib9]^,^[@bib10] Scanning electron microscopy of hair specimens may show ridges that run longitudinally along the shaft, twisting of the hair along its longitudinal axis, triangular or other abnormally shaped cross sections, and a distorted bulb.[@bib1]^,^[@bib8]^,^[@bib9]^,^[@bib11] Recently, a genetic pathway related to loose anagen hair has been described for Noonan syndrome with loose anagen hair and another rasopathy involving the SHOC2 and PPP1CB proteins, respectively.[@bib12]\u00a0In the patient discussed in this case report, the many anagen hairs and clinically apparent hair loss found using light microscopy suggest LAHS. However, the unruly appearance of her hair and the extensive grooving are suggestive of UHS.\n\nIn 1973, Dupre et\u00a0al[@bib13] described *cheveux incoiffables,* meaning \"unmanageable hair,\" and Stround and Mehregan[@bib14] described a spun-glass hair syndrome. This later became known as *uncombable hair syndrome*. Patients have hair that sticks out from the scalp and cannot lie flat. They typically have a light hair color that kinks and twists. This abnormal hair texture is typically noticed before adolescence. This phenotype is often an isolated finding, but it can be found in association with other ectodermal dysplasias manifesting with abnormal teeth, eyes, nails, and sweat glands.[@bib2]^,^[@bib15] Another recent case report details a patient with UHS and congenital anonychia found to have mutations in both the RSPO4 and PADI3 proteins.[@bib16] The differential diagnosis includes pili torti, monilethrix, wooly hair nevus, and loose anagen hair syndrome. The diagnosis of UHS can be distinguished from these syndromes by clinical and microscopic examination.[@bib2]^,^[@bib17] Similar to LAHS, UHS has both inherited and sporadic forms described in the literature. However, unlike LAHS, scalp hair in UHS is not typically fragile, and the syndrome is not associated with hair loss.[@bib2]\n\nIt is unusual for both of these syndromes to present in the same patient, although there are 2 other reports in the literature.[@bib3]^,^[@bib4] Both of the 2 previous cases and this case describe a young girl with clinical and microscopic features of LAHS and UHS simultaneously.\n\nThere are many similarities between this case and the 2 other cases of LAHS with concurrent UHS reported in the literature, although there are some notable differences. In contrast to Lee et\u00a0al,[@bib4] examination of this patient\\'s hair samples with polarized filters did not show light and dark horizontal banding alternating but, rather, symmetric and normally appearing shafts. Lee et\u00a0al note that the banding pattern they described is not characteristic of either LAHS or UHS.\n\nAlthough the incidence of patients meeting the criteria for both of the mentioned conditions is rare, it may be underdiagnosed or underreported in the literature. It is important to consider rare entities and concurrent overlap of syndromes when caring for patients. If criteria for 1 syndrome are met, other diagnostic clues should not be ignored because that may reveal a more complex case. It is possible that there is a link between LAHS and UHS, either genetic or through a shared downstream pathway, but further study and input from the literature are necessary. Increased awareness of patients with concurrent syndromes such as this will allow more providers to recognize and report similar situations, eventually leading to an insight into the mechanism of these diseases.\n\nConclusion {#sec4}\n==========\n\nThe presented case represents an unusual condition characterized by features of both LAHS and UHS. The patient\\'s scalp has patchy hair loss and easily extracted anagen hairs seen under light microscopy. This is characteristic of LAHS. However, the patient also has an unruly texture to her hair, characteristic of UHS. It is important to consider multiple syndromes and rare entities when evaluating a patient.\n\nFunding sources: None.\n\nConflicts of interest: None disclosed.\n"}
+{"text": "![](indmedgaz70882-0012){#sp1 .276}\n\n![](indmedgaz70882-0013){#sp2 .277}\n\n![](indmedgaz70882-0014){#sp3 .278}\n\n![](indmedgaz70882-0015){#sp4 .279}\n"}
+{"text": "INTRODUCTION\n============\n\nEndosonography was initially used in the staging of gastrointestinal tract malignancies.^(^ [@B01] ^)^ In the 1990s, it was adapted for use in bronchial diseases. In patients with lung disease, its uses now include tumor staging; the diagnosis of central (parenchymal) masses; and the detection of mediastinal or hilar lymphadenopathy. Endobronchial ultrasound (EBUS) enables the visualization of lymph node structure, thus allowing the pulmonologist to evaluate and sample lymph nodes. Consequently, minimally invasive staging of lung cancer has advanced considerably. In addition, tumor invasion of the tracheobronchial wall can be assessed more accurately with EBUS than with CT. The accuracy of EBUS in making this distinction is 94%, compared with 51% for CT.^(^ [@B02] ^)^ Lymph node stations 2, 4, 7, 10, and 11 can be sampled by EBUS. If EBUS is combined with esophageal ultrasound, lymph node stations 5, 8, and 9 can also be sampled. Therefore, the combination of esophageal ultrasound and EBUS can be seen as the first and best test in patients with suspected lymph node metastasis. ^(^ [@B03] ^)^ There are studies showing that this is a good alternative to mediastinoscopy.^(^ [@B04] ^,^ [@B05] ^)^\n\nIn patients with malignant disease or granulomatous diseases such as tuberculosis and sarcoidosis, EBUS can contribute to the diagnosis. In a recent meta-analysis, the diagnostic accuracy of EBUS-guided transbronchial needle aspiration (EBUS-TBNA) for sarcoidosis was shown to be 54-93%.^(^ [@B06] ^)^ In tuberculosis, for which EBUS is also diagnostic, EBUS-TBNA has been shown to have a sensitivity of 85%.^(^ [@B07] ^)^\n\nIn the evaluation of airway disease, EBUS has emerged as a technique that has great potential for development. Different diagnostic values for EBUS have been reported in various studies. The aim of this study was to determine the diagnostic value of EBUS-TBNA, its contribution to the diagnosis of different diseases, and the factors that determine the magnitude of that contribution.\n\nMETHODS\n=======\n\nPatients and procedures\n-----------------------\n\nIn this study, we retrospectively analyzed 159 patients in whom EBUS-TBNA was used at our pulmonary medicine clinic between 2010 and 2013. In patients with mediastinal/hilar lymphadenopathy, EBUS-TBNA was performed in order to evaluate the etiology. Lymphadenopathy was defined as a finding of one or more lymph nodes with a short-axis diameter of \u2265 10 mm on CT or high 18F-fluorodeoxyglucose uptake on positron emission tomography/CT. All of the patients evaluated gave written informed consent. Patients fasted for at least 4 h before the procedure. The preparation process included local anesthesia with lidocaine hydrochloride (Xylocaine^(r))^ and sedation with midazolam just prior to the EBUS procedure, which was performed with a convex-probe ultrasound-guided fiberoptic bronchoscope (BF-UC160F-OL8; Olympus Medical Systems, Tokyo, Japan). All EBUS-TBNA procedures were performed by the same pulmonologist. The location and size of the lymph nodes seen during EBUS were recorded. Lymph nodes that appeared to be affected were sampled at least twice. The samples were sent for histopathological assessment. Rapid onsite cytopathological examination was not performed. If EBUS-TBNA did not lead to a diagnosis, the patients underwent mediastinoscopy, open lung biopsy, or transthoracic needle aspiration (TTNA), according to the situation. In patients with pathological diagnosis of chronic granulomatous inflammation, the diagnosis of tuberculosis or sarcoidosis was determined from the EBUS-TBNA results. This determination was made on the basis of the presence of necrosis in the samples; clinical symptoms; the history of contact with tuberculosis cases; microbiological evaluation; tuberculin skin test results; and additional biochemical features. Samples were classified as containing insufficient material if they contained no lymphocytes. The cytology of a sample was classified as benign (normal) if it presented mature lymphocytes or anthracosis, without malignant cells or granulomas. The final diagnosis was based on the cytology, surgical results, or clinical follow-up. We recorded all diagnoses resulting from EBUS-TBNA. For the patients with inconclusive EBUS-TBNA results, the final diagnoses were recorded as those made after surgery, TTNA, or at least 6 months of follow-up (i.e., bacteriological and clinical outcomes). Patients were grouped by final diagnosis: malignant disease; benign disease; or inconclusive (normal cytology/anthracosis or insufficient material).\n\nStatistical analysis\n--------------------\n\nFor statistical analysis of the data, we used the Statistical Package for the Social Sciences, version 15.0 for Windows (SPSS Inc., Chicago, IL, USA). We used descriptive statistics, including mean, standard deviation, and frequency. The lymph node size that supported the EBUS-TBNA-determined diagnosis was calculated via ROC analysis. Values of p \\< 0.05 were considered statistically significant.\n\nRESULTS\n=======\n\nOf the 159 patients evaluated by EBUS-TBNA in the present study, 89 (56%) were male and 70 (44%) were female. The mean age was 54.6 \u00b1 14.2 years for the male patients and 51.9 \u00b1 11.3 years for the female patients. Patients enrolled in the study were similar in terms of gender and age distribution (p = 0.13).\n\nThe EBUS-TBNA-determined diagnoses were as follows: sarcoidosis (in 43 patients); tuberculosis (in 14); malignancy (in 33); and normal cytology/anthracosis (in 58). In addition, EBUS-TBNA resulted in one patient being diagnosed with nocardiosis and another being diagnosed with a cyst in the subcarinal area. In 92 of the 159 patients, the EBUS-TBNA procedure was diagnostic. In 9 patients, the EBUS-TBNA samples were classified as containing insufficient material. Further investigation (surgery, TTNA, or follow-up) of those 9 patients revealed one case of sarcoidosis and one case of malignancy, the remaining seven cases being lost to follow-up. In the analysis of the EBUS-TBNA samples, the cytology was categorized as normal in 58 patients. Further evaluation and assessment of this group yielded the following diagnoses: sarcoidosis (n = 6); tuberculosis (n = 5); malignancy (n = 9); nocardiosis (n = 1); and normal cytology/anthracosis (n = 22). The remaining 15 patients were lost to follow-up, and the final diagnoses were therefore unknown. Of the 159 patients evaluated, 114 (83%) were correctly diagnosed using EBUS-TBNA. That group included those diagnosed with sarcoidosis or tuberculosis, as well as those in which the cytology was categorized as malignant (true positive) or benign (true negative). [Figure 1](#f01){ref-type=\"fig\"} details the distribution of the final diagnoses.\n\nFigure 1.Flow diagram of the diagnoses made by endobronchial ultrasound-guided transbronchial needle aspiration and by other methods. EBUS-TBNA: endobronchial ultrasound-guided transbronchial needle aspiration, TTNA: transthoracic needle aspiration.\n\nOf the 50 patients receiving a final diagnosis of sarcoidosis, 43 (86%) were diagnosed with EBUS-TBNA, as were 14 (74%) of the 19 patients receiving a final diagnosis of tuberculosis. In the sarcoidosis and tuberculosis group, when we considered those 69 cases collectively (as the benign granulomatous disease category), EBUS-TBNA had a diagnostic accuracy of 83%. Among the 159 patients analyzed, the final diagnosis was malignancy in 43. In 33 (77%) of those patients, the diagnosis was based on the EBUS-TBNA findings. Among the remaining 10 patients (i.e., the 23% that were not diagnosed with EBUS-TBNA), there were seven cases of lung malignancy, two cases of hematologic malignancy (plasmacytoma and lymphoma, respectively), and one case of esophageal cancer. Therefore, for malignancy, EBUS-TBNA had a negative predictive value of 92% and an accuracy of 94%. The distribution of the final diagnoses and the diagnostic accuracy of EBUS-TBNA for each are given in detail in [Table 1](#t01){ref-type=\"table\"}. The diagnostic accuracy of EBUS-TBNA was found to be similar for benign and malignant pathologies (p = 0.39). No serious complications (requiring early termination of the procedure) were found to have occurred during the use of EBUS-TBNA.\n\nTable 1.Final diagnoses and diagnostic accuracy of endobronchial ultrasound-guided transbronchial needle aspiration.**Type of diseaseFinal diagnosisDefinitive diagnosis by EBUS-TBNAnn (%)**Malignant4333 (77)Non-small cell lung cancer3226 (81)Small cell lung cancer76 (86)Other^a^41 (25)Benign7259 (82)Sarcoidosis5043 (86)Tuberculosis1914 (74)Other^b^32 (67)[^1]\n\nWhen we evaluated factors that might affect the diagnostic accuracy of EBUS-TBNA, neither age nor gender were found to have any such effect (p = 0.05 and p = 0.43, respectively). The number and size of lymph nodes detected using EBUS-TBNA are given in detail in [Table 2](#t02){ref-type=\"table\"}. In the presence of enlarged mediastinal (N2) lymph nodes, the likelihood of obtaining a diagnosis with EBUS-TBNA increased significantly (p = 0.013). Enlargement of hilar lymph nodes had no such effect (p= 0.065). The ROC analysis, performed in order to assess the contribution of lymph node size to the likelihood of diagnosis, revealed that, for obtaining an EBUS-TBNA-based diagnosis of granulomatous disease or malignancy, a finding of lymph nodes with a short-axis diameter \u2265 16.5 mm had a sensitivity of 60% and a specificity of 76% ([Figure 2](#f02){ref-type=\"fig\"}). For that cut-off value, the positive predictive value was 71%, the negative predictive value was 66%, and the area under the curve was 0.728. When that lymph node size cut-off value was applied, the diagnostic accuracy of EBUS-TBNA was 69% (p \\< 0.001).\n\nTable 2.Lymph node number and size detected by endobronchial ultrasound.**Lymph nodenShort-axis diametermm^a^**2R48.25 (5-35)2L04R5514.5 (5-41)4L1910.1 (4-50)710117 (3-42)10R or 11R4514 (6-35)10L or 11L5715 (4-45)Total281[^2]\n\nFigure 2.ROC curve analysis: when a lymph node size (short-axis diameter) cut-off value of 16.5 mm was applied, the sensitivity and specificity of endobronchial ultrasound (for a diagnosis of granulomatous disease or malignancy) were 76% and 60%, respectively. Diagonal segments are produced by ties.\n\nDISCUSSION\n==========\n\nAlthough EBUS-TBNA is a new method, its use in the diagnosis and staging of lung diseases has become increasingly widespread. With EBUS-TBNA, it is possible to sample a lesion while it is being displayed by the ultrasound probe. This method has been used for lung cancer staging. It has also been shown to be diagnostic for benign lung diseases.^(^ [@B03] ^-^ [@B07] ^)^ In the present study, we found that the overall diagnostic accuracy of EBUS-TBNA (regardless of the etiology) was 83%. Similarly, Choi et al.^(^ [@B08] ^)^ reported that EBUS-TBNA had an overall diagnostic accuracy of 83.9%. Evaluating our cases over the last four years, we found that EBUS-TBNA has been well tolerated and has proven to be a reliable diagnostic method with few complications. The purpose of the present study was to share our experience regarding the diagnostic value of EBUS-TBNA.\n\nIn the diagnosis of sarcoidosis, EBUS-TBNA is often used, because it can preclude the need for conventional TBNA or, in some cases, mediastinoscopy.^(^ [@B04] ^,^ [@B05] ^,^ [@B09] ^)^ The use of EBUS-TBNA guarantees that the sample was taken from the targeted lymph nodes. Therein lies its superiority over conventional TBNA.^(^ [@B10] ^,^ [@B11] ^)^ In addition, EBUS-TBNA is not as invasive as is mediastinoscopy. Various levels of diagnostic accuracy in sarcoidosis have been reported for EBUS-TBNA. A recent meta-analysis found that the reported diagnostic accuracy of EBUS-TBNA in sarcoidosis ranged from 54% to 93%.^(^ [@B06] ^)^ In our study, that level was 86%, which is in agreement with the findings of other studies in the literature. Navani et al.^(^ [@B12] ^)^ evaluated patients with suspected sarcoidosis using EBUS-TBNA and reported a diagnostic accuracy of 88%. When those same patients underwent transbronchial lung biopsy, mucosal biopsy, and BAL via standard bronchoscopy, the diagnostic accuracy increased to 93%. The variability in the reported levels of diagnostic accuracy across studies might reflect the level of experience of the pulmonologist performing the EBUS-TBNA or that of the pathologist, as well as the difference between studies in which rapid onsite cytopathological examination was performed and those in which it was not.\n\nIn another benign disease, tuberculosis, EBUS-TBNA is also used as a diagnostic method. Sun et al.^(^ [@B07] ^)^ and Cetinkaya et al.^(^ [@B13] ^)^ reported the sensitivity of EBUS-TBNA in tuberculosis to be 85% and 79%, respectively,^(^ [@B07] ^,^ [@B13] ^)^ compared with the 74% found in the present study. The diagnostic accuracy of EBUS-TBNA is lower for tuberculosis than for sarcoidosis, and conventional TBNA has a diagnostic accuracy of 65% for tuberculosis.^(^ [@B14] ^)^ Ren et al.^(^ [@B15] ^)^ found that, in patients with suspected mediastinal tuberculous lymphadenitis, standard bronchoscopy had a sensitivity of 18.1% and a specificity of 100%. When the authors added EBUS-TBNA to standard bronchoscopy, those values rose to 80% and 92.3%, respectively.^(^ [@B15] ^)^ Therefore, EBUS-TBNA seems to contribute to the diagnostic accuracy in patients with tuberculous lymphadenitis and mediastinal lymph node enlargement. In the present study, we also found the diagnostic accuracy of EBUS-TBNA to be lower for tuberculosis than for sarcoidosis. The prominent necrosis in paucibacillary tuberculosis could explain that finding. Dhooria et al.^(^ [@B16] ^)^ reported that one can distinguish between tuberculosis and sarcoidosis on the basis of EBUS images. Accordingly, when combined with a positive tuberculin skin test result, the heterogeneous echotexture of the internal structure or necrotic appearance seen via EBUS can diagnose tuberculosis with a specificity of 98% and a positive predictive value of 91%.^(^ [@B16] ^)^\n\nOne recent study reported that EBUS-TBNA has a diagnostic accuracy of 93.9% for malignancy,^(^ [@B08] ^)^ compared with only 77% in the present study. That same study reported that the diagnostic accuracy of EBUS-TBNA was higher for malignant diseases than for benign diseases (93.9% vs. 70.6%, p \\< 0.001). In our study, the diagnostic accuracy of EBUS-TBNA for benign granulomatous disorders was not significantly different from that observed for malignant disease (83% vs. 77%, p \\> 0.05). Memoli et al.^(^ [@B17] ^)^ showed that increased lymph node size supports an EBUS-based diagnosis of malignancy. Abu-Hijleh et al.^(^ [@B18] ^)^ suggested that lymph node size has little effect on diagnostic accuracy or sensitivity of EBUS-TBNA. In contrast, when lymph node size exceeds 20 mm, the negative predictive value decreases. Those authors found that samples taken from lymph nodes \u2265 20 mm in size always contained sufficient material for analysis. The same authors reported that samplings of the 4R, 4L, 7, 10/11R, and 10/11L lymph nodes did not differ in terms of diagnostic accuracy.^(^ [@B18] ^)^ Tedde et al.^(^ [@B19] ^)^ showed that identifying lymphadenopathy in multiple lymph nodes and sampling subcarinal lymph nodes provides a higher diagnostic yield in EBUS-TBNA. Those authors reported that EBUS-TBNA has a diagnostic accuracy of 57% for malignancy. In the present study, lymph nodes with a short-axis diameter \u2265 16.5 mm were found to be more diagnostic in the histopathological evaluation than were those with smaller diameters. We also found that the diagnostic accuracy of EBUS-TBNA was better when there was mediastinal lymph node enlargement than when there was hilar lymph node enlargement. In a study conducted by Cetinkaya et al.,^(^ [@B20] ^)^ the sensitivity of EBUS-TBNA was lower in the region of lymph node 4L than in those of the other lymph nodes. The authors reported that the overall diagnostic accuracy of EBUS-TBNA was not affected by the number of lymph nodes sampled, the number of times a lymph node region was sampled, or lymph node size.^(^ [@B20] ^)^\n\nIn our study sample, we noted that EBUS-TBNA provoked only minor complications, such as mild hypoxia, tachycardia, and minor hemorrhage. None of those complications changed the course of the procedure or led to additional treatment or hospitalization. We can therefore assert that the procedure can be used safely in all adult age groups. Similar studies in the literature have also reported that EBUS-TBNA is safe and causes no severe complications.^(^ [@B08] ^,^ [@B20] ^-^ [@B23] ^)^\n\nIn conclusion, EBUS-TBNA is a proven method in the diagnosis and staging of malignancy. The diagnostic accuracy of this method is also high in benign pathologies such as sarcoidosis and tuberculosis. It is a method that is well-tolerated and minimally invasive, with a low rate of complications and high diagnostic accuracy. The use of EBUS-TBNA should be encouraged as the first procedure of choice for mediastinal and hilar lung pathologies, because it markedly reduces the need for mediastinoscopy.\n\nWe are grateful to Mr. David Chapman for editing the language of the manuscript.\n\nStudy carried out in the Department of Pulmonology, Istanbul Yedikule Chest Diseases and Thoracic Surgery Training and Research Hospital, Istanbul, Turkey.\n\nFinancial support: None.\n\nArtigo Original\n\nValor diagn\u00f3stico da pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico em diferentes doen\u00e7as pulmonares\n\nOrtakoylu\n\nMediha Gonenc\n\n1\n\nIliaz\n\nSinem\n\n1\n\nBahadir\n\nAyse\n\n1\n\nAslan\n\nAsuman\n\n1\n\nIliaz\n\nRaim\n\n2\n\nOzgul\n\nMehmet Akif\n\n1\n\nUrer\n\nHalide Nur\n\n3\n\nDepartment of Pulmonology, Istanbul Yedikule Chest Diseases and Thoracic Surgery Training and Research Hospital, Istanbul, Turkey.\n\nDepartment of Internal Medicine, Istanbul University Istanbul Medical Faculty, Istanbul, Turkey.\n\nDepartment of Pathology, Istanbul Yedikule Chest Diseases and Thoracic Surgery Training and Research Hospital, Istanbul, Turkey.\n\nEndere\u00e7o para correspond\u00eancia: Sinem Iliaz. Istanbul Yedikule Chest Diseases and Thoracic Surgery Training and Research Hospital, Department of Pulmonology, 34020, Zeytinburnu, Istanbul, Turkey. Tel.: 90 212 409-0200. E-mail:\n\nsnmkaraosman\\@gmail.com\n\nObjetivo:\n=========\n\nEndobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA, pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico) \u00e9 um novo m\u00e9todo para o diagn\u00f3stico e estadiamento das doen\u00e7as pulmonares e seu uso est\u00e1 aumentando em todo o mundo. Ela tem sido utilizada como um m\u00e9todo diagn\u00f3stico de c\u00e2ncer de pulm\u00e3o em est\u00e1gios iniciais, e h\u00e1 dados que apoiam sua utiliza\u00e7\u00e3o para o diagn\u00f3stico de doen\u00e7as pulmonares benignas. O objetivo deste estudo foi compartilhar a nossa experi\u00eancia com EBUS-TBNA e discutir seu valor diagn\u00f3stico.\n\nM\u00e9todos:\n========\n\nAnalisamos, retrospectivamente, os resultados relacionados a 159 pacientes submetidos a EBUS-TBNA em nossa cl\u00ednica de medicina pulmonar entre 2010 e 2013. Registramos a localiza\u00e7\u00e3o e o tamanho dos linfonodos visualizados durante EBUS. Os linfonodos suspeitos durante o procedimento foram puncionados ao menos duas vezes. Foram registrados os resultados diagn\u00f3sticos por EBUS-TBNA e, para os casos com diagn\u00f3stico indefinido ap\u00f3s EBUS-TBNA, os diagn\u00f3sticos finais ap\u00f3s investiga\u00e7\u00e3o aprofundada e acompanhamento.\n\nResultados:\n===========\n\nAvaliamos 159 pacientes, dos quais 89 (56%) eram homens e 70 (44%) eram mulheres. As m\u00e9dias de idade foram de 54,6 \u00b1 14,2 anos nos homens e de 51,9 \u00b1 11,3 anos nas mulheres. Dos 159 pacientes avaliados, 115 (84%) foram diagnosticados corretamente por EBUS. A acur\u00e1cia diagn\u00f3stica de EBUS-TBNA foi de 83% para doen\u00e7as granulomatosas benignas e de 77% para doen\u00e7as malignas.\n\nConclus\u00f5es:\n===========\n\nO valor diagn\u00f3stico de EBUS-TBNA tamb\u00e9m \u00e9 alto para patologias benignas, como sarcoidose e tuberculose. Em pacientes com altera\u00e7\u00f5es no mediastino, o uso de EBUS-TBNA deve ser incentivado, principalmente devido \u00e0 redu\u00e7\u00e3o significativa da necessidade de mediastinoscopia.\n\nSarcoidose\n\nTuberculose pulmonar\n\nNeoplasias pulmonares\n\nBroncoscopia\n\nMediastinoscopia\n\nEndossonografia\n\nINTRODU\u00c7\u00c3O\n==========\n\nA endossonografia foi inicialmente utilizada no estadiamento de malignidades do trato gastrointestinal. ^(^ [@B01] ^)^ Na d\u00e9cada de 1990, foi adaptada para utiliza\u00e7\u00e3o em broncopatias. Em pacientes com doen\u00e7a pulmonar, seus usos agora incluem estadiamento tumoral, diagn\u00f3stico de massas (parenquimatosas) centrais e detec\u00e7\u00e3o de linfadenopatia mediastinal ou hilar. O *endobronchial ultrasound* (EBUS, ultrassom endobr\u00f4nquico) possibilita a visualiza\u00e7\u00e3o da estrutura do linfonodo, permitindo assim que o pneumologista avalie e puncione os linfonodos. Em consequ\u00eancia, o estadiamento minimamente invasivo do c\u00e2ncer de pulm\u00e3o tem avan\u00e7ado consideravelmente. Al\u00e9m disso, a invas\u00e3o tumoral da parede traqueobr\u00f4nquica pode ser avaliada com maior acur\u00e1cia por EBUS do que por TC. A acur\u00e1cia do EBUS em fazer essa distin\u00e7\u00e3o \u00e9 de 94%, enquanto a da TC \u00e9 de 51%.^(^ [@B02] ^)^ As esta\u00e7\u00f5es linfonodais 2, 4, 7, 10 e 11 podem ser puncionadas sob orienta\u00e7\u00e3o do EBUS. Se o EBUS \u00e9 combinado com ultrassom esof\u00e1gico, as esta\u00e7\u00f5es linfonodais 5, 8 e 9 tamb\u00e9m podem ser puncionadas. Portanto, a combina\u00e7\u00e3o de ultrassom esof\u00e1gico e EBUS pode ser vista com o primeiro e melhor teste em pacientes com suspeita de met\u00e1stase linfonodal.^(^ [@B03] ^)^ H\u00e1 estudos que mostram que essa \u00e9 uma boa alternativa \u00e0 mediastinoscopia.^(^ [@B04] ^,^ [@B05] ^)^\n\nEm pacientes com doen\u00e7a maligna ou doen\u00e7as granulomatosas como tuberculose e sarcoidose, o EBUS pode contribuir para o diagn\u00f3stico. Em uma recente meta-an\u00e1lise, demonstrou-se que a acur\u00e1cia diagn\u00f3stica de *EBUS-guided transbronchial needle aspiration* (EBUS-TBNA, pun\u00e7\u00e3o aspirativa por agulha guiada por EBUS) para sarcoidose foi de 54-93%.^(^ [@B06] ^)^ Em tuberculose, para a qual o EBUS tamb\u00e9m \u00e9 diagn\u00f3stico, foi demonstrado que EBUS-TBNA apresenta sensibilidade de 85%.^(^ [@B07] ^)^\n\nNa avalia\u00e7\u00e3o das doen\u00e7as das vias a\u00e9reas, o EBUS surge como uma t\u00e9cnica com grande potencial de desenvolvimento. Diferentes valores diagn\u00f3sticos para o EBUS j\u00e1 foram relatados em diversos estudos. O objetivo deste estudo foi determinar o valor diagn\u00f3stico de EBUS-TBNA, sua contribui\u00e7\u00e3o para o diagn\u00f3stico de diferentes doen\u00e7as e os fatores que determinam a magnitude dessa contribui\u00e7\u00e3o.\n\nM\u00c9TODOS\n=======\n\nPacientes e procedimentos\n-------------------------\n\nNeste estudo, analisamos retrospectivamente 159 pacientes nos quais se utilizou EBUS-TBNA em nossa cl\u00ednica de medicina pulmonar entre 2010 e 2013. Nos pacientes com linfadenopatia mediastinal/hilar, realizou-se EBUS-TBNA para avalia\u00e7\u00e3o da etiologia. Considerou-se linfadenopatia o achado de um ou mais linfonodos com di\u00e2metro \u2265 10 mm no menor eixo na TC ou com alta capta\u00e7\u00e3o de 18F fluordesoxiglicose na tomografia por emiss\u00e3o de p\u00f3sitrons/TC. Todos os pacientes avaliados assinaram um termo de consentimento livre e esclarecido. Os pacientes estavam em jejum de no m\u00ednimo 4 h antes da realiza\u00e7\u00e3o do procedimento. O processo de prepara\u00e7\u00e3o incluiu anestesia local com cloridrato de lidoca\u00edna (Xylocaine^(r))^ e seda\u00e7\u00e3o com midazolam imediatamente antes do procedimento de EBUS, que foi realizado com fibrobroncosc\u00f3pio guiado por ultrassom com sonda convexa (BF-UC160F-OL8; Olympus Medical Systems, T\u00f3quio, Jap\u00e3o). Todos os procedimentos de EBUS-TBNA foram realizados pelo mesmo pneumologista. A localiza\u00e7\u00e3o e o tamanho dos linfonodos visualizados durante EBUS foram registrados. Os linfonodos suspeitos durante o procedimento foram puncionados ao menos duas vezes. As amostras foram enviadas para avalia\u00e7\u00e3o histopatol\u00f3gica. N\u00e3o foi realizada citopatologia r\u00e1pida no local. Se EBUS-TBNA n\u00e3o levasse ao diagn\u00f3stico, os pacientes eram submetidos a mediastinoscopia, bi\u00f3psia pulmonar a c\u00e9u aberto ou *transthoracic needle aspiration* (TTNA, pun\u00e7\u00e3o aspirativa transtor\u00e1cica por agulha), de acordo com a situa\u00e7\u00e3o. Em pacientes com diagn\u00f3stico patol\u00f3gico de inflama\u00e7\u00e3o granulomatosa cr\u00f4nica, o diagn\u00f3stico de tuberculose ou sarcoidose foi estabelecido a partir dos resultados de EBUS-TBNA. O estabelecimento desse diagn\u00f3stico foi feito com base na presen\u00e7a de necrose nas amostras, sintomas cl\u00ednicos, hist\u00f3ria de contato com casos de tuberculose, avalia\u00e7\u00e3o microbiol\u00f3gica, resultados do teste tubercul\u00ednico e caracter\u00edsticas bioqu\u00edmicas adicionais. As amostras foram classificadas como contendo material insuficiente se n\u00e3o contivessem linf\u00f3citos. A citologia de uma amostra foi classificada como benigna (normal) se a mesma apresentasse linf\u00f3citos maduros ou antracose, sem c\u00e9lulas malignas ou granulomas. O diagn\u00f3stico final baseou-se na citologia, resultados cir\u00fargicos ou acompanhamento cl\u00ednico. Registramos todos os diagn\u00f3sticos resultantes de EBUS-TBNA. Para os pacientes com resultados de EBUS-TBNA indefinidos, os diagn\u00f3sticos finais registrados foram aqueles feitos ap\u00f3s cirurgia, TTNA ou acompanhamento de no m\u00ednimo 6 meses (isto \u00e9, resultados bacteriol\u00f3gicos e cl\u00ednicos). Os pacientes foram agrupados de acordo com o diagn\u00f3stico final: doen\u00e7a maligna; doen\u00e7a benigna; ou indefinido (citologia normal/antracose ou material insuficiente).\n\nAn\u00e1lise estat\u00edstica\n-------------------\n\nPara a an\u00e1lise estat\u00edstica dos dados, utilizou-se o *Statistical Package for the Social Sciences*, vers\u00e3o 15.0 para Windows (SPSS Inc., Chicago, IL, EUA). Foram utilizadas an\u00e1lises estat\u00edsticas descritivas, como m\u00e9dia, desvio-padr\u00e3o e frequ\u00eancia. O tamanho linfonodal que apoiou o diagn\u00f3stico estabelecido por EBUS-TBNA foi calculado pela an\u00e1lise ROC. Valores de p \\< 0,05 foram considerados estatisticamente significativos.\n\nRESULTADOS\n==========\n\nDos 159 pacientes avaliados por EBUS-TBNA no presente estudo, 89 (56%) eram homens e 70 (44%) eram mulheres. As m\u00e9dias de idade foram de 54,6 \u00b1 14,2 anos para os homens e de 51,9 \u00b1 11,3 anos para as mulheres. Os pacientes inclu\u00eddos no estudo eram semelhantes quanto \u00e0 distribui\u00e7\u00e3o por sexo e idade (p = 0,13).\n\nOs diagn\u00f3sticos estabelecidos por EBUS-TBNA foram os seguintes: sarcoidose (em 43 pacientes); tuberculose (em 14); malignidade (em 33); e citologia normal/antracose (em 58). Al\u00e9m disso, EBUS-TBNA resultou no diagn\u00f3stico de nocardiose em um paciente e no diagn\u00f3stico de cisto na regi\u00e3o subcarinal em outro. Em 92 dos 159 pacientes, o procedimento de EBUS-TBNA foi diagn\u00f3stico. Em 9 pacientes, as amostras de EBUS-TBNA foram classificadas como contendo material insuficiente. Uma investiga\u00e7\u00e3o aprofundada (cirurgia, TTNA ou acompanhamento) desses 9 pacientes revelou um caso de sarcoidose e um caso de malignidade, havendo perda de acompanhamento nos sete casos restantes. Na an\u00e1lise das amostras de EBUS-TBNA, a citologia foi classificada como normal em 58 pacientes. Uma avalia\u00e7\u00e3o aprofundada desse grupo produziu os seguintes diagn\u00f3sticos: sarcoidose (n = 6); tuberculose (n = 5); malignidade (n = 9); nocardiose (n = 1); e citologia normal/antracose (n = 22). Houve perda de acompanhamento dos 15 pacientes restantes, e os diagn\u00f3sticos finais foram, portanto, desconhecidos. Dos 159 pacientes avaliados, 114 (83%) foram diagnosticados corretamente utilizando-se EBUS-TBNA. Esse grupo incluiu aqueles diagnosticados com sarcoidose ou tuberculose, bem como aqueles nos quais a citologia foi classificada como maligna (verdadeiro positivo) ou benigna (verdadeiro negativo). A [Figura 1](#f1){ref-type=\"fig\"} detalha a distribui\u00e7\u00e3o dos diagn\u00f3sticos finais.\n\nFigura 1.Diagrama de fluxo dos diagn\u00f3sticos feitos por pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico e por outros m\u00e9todos. EBUS-TBNA: *endobronchial ultrasound-guided transbronchial needle aspiration*(pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico), TTNA: *transthoracic needle aspiration* (pun\u00e7\u00e3o aspirativa transtor\u00e1cica por agulha).\n\nDos 50 pacientes que receberam diagn\u00f3stico final de sarcoidose, 43 (86%) foram diagnosticados por EBUS-TBNA, assim como o foram 14 (74%) dos 19 pacientes que receberam diagn\u00f3stico final de tuberculose. No grupo sarcoidose e tuberculose, quando considerados esses 69 casos em conjunto (como a categoria doen\u00e7a granulomatosa benigna), EBUS-TBNA apresentou acur\u00e1cia diagn\u00f3stica de 83%. Entre os 159 pacientes analisados, o diagn\u00f3stico final foi de malignidade em 43. In 33 (77%) desses pacientes, o diagn\u00f3stico baseou-se nos achados de EBUS-TBNA. Entre os 10 pacientes restantes (isto \u00e9, os 23% que n\u00e3o foram diagnosticados por EBUS-TBNA), houve sete casos de malignidade pulmonar, dois casos de malignidade hematol\u00f3gica (plasmocitoma e linfoma, respectivamente) e um caso de c\u00e2ncer de es\u00f4fago. Portanto, para malignidade, EBUS-TBNA apresentou valor preditivo negativo de 92% e acur\u00e1cia de 94%. A distribui\u00e7\u00e3o dos diagn\u00f3sticos finais e a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA para cada um deles est\u00e3o detalhadas na [Tabela 1](#t1){ref-type=\"table\"}. A acur\u00e1cia diagn\u00f3stica de EBUS-TBNA se mostrou semelhante para patologias benignas e malignas (p = 0,39). N\u00e3o ocorreu nenhuma complica\u00e7\u00e3o grave (que tenha exigido a interrup\u00e7\u00e3o precoce do procedimento) durante o uso de EBUS-TBNA.\n\nTabela 1.Diagn\u00f3sticos finais e acur\u00e1cia diagn\u00f3stica da pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico.**Tipo de doen\u00e7aDiagn\u00f3stico finalDiagn\u00f3stico definitivo por EBUS-TBNAnn (%)**Maligna4333 (77)C\u00e2ncer de pulm\u00e3o n\u00e3o pequenas c\u00e9lulas3226 (81)C\u00e2ncer de pulm\u00e3o de pequenas c\u00e9lulas76 (86)Outraa41 (25)Benigna7259 (82)Sarcoidose5043 (86)Tuberculose1914 (74)Outrab32 (67)[^3]\n\nQuando analisados fatores que poderiam interferir na acur\u00e1cia diagn\u00f3stica de EBUS-TBNA, nem idade nem sexo mostraram ter esse efeito (p = 0,05 e p = 0,43, respectivamente). O n\u00famero e o tamanho dos linfonodos detectados por EBUS-TBNA est\u00e3o detalhados na [Tabela 2](#t2){ref-type=\"table\"}. Na presen\u00e7a de linfonodos mediastinais (N2) aumentados, a probabilidade de obten\u00e7\u00e3o do diagn\u00f3stico por EBUS-TBNA aumentou significativamente (p = 0,013). O aumento de linfonodos hilares n\u00e3o teve esse efeito (p= 0,065). A an\u00e1lise ROC, realizada para avaliar a contribui\u00e7\u00e3o do tamanho linfonodal para a probabilidade diagn\u00f3stica, revelou que, para obten\u00e7\u00e3o do diagn\u00f3stico de doen\u00e7a granulomatosa ou malignidade com base em EBUS-TBNA, o achado de linfonodos com di\u00e2metro \u2265 16,5 mm no menor eixo apresentou sensibilidade de 60% e especificidade de 76% ([Figura 2](#f2){ref-type=\"fig\"}). Para esse valor de corte, o valor preditivo positivo foi de 71%, o valor preditivo negativo, de 66%, e a \u00e1rea sob a curva, de 0,728. Quando utilizado esse valor de corte para tamanho linfonodal, a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA foi de 69% (p \\< 0,001).\n\nTabela 2.N\u00famero e tamanho dos linfonodos detectados por ultrassom endobr\u00f4nquico.**LinfonodonDi\u00e2metro no menor eixomma**2R48,25 (5-35)2L04R5514,5 (5-41)4L1910,1 (4-50)710117 (3-42)10R ou 11R4514 (6-35)10L ou11L5715 (4-45)Total281[^4]\n\nFigura 2.An\u00e1lise da curva ROC: quando utilizado valor de corte para tamanho linfonodal (di\u00e2metro no menor eixo) de 16,5 mm, a sensibilidade e a especificidade do ultrassom endobr\u00f4nquico (para diagn\u00f3stico de doen\u00e7a granulomatosa ou malignidade) foram de 76% e 60%, respectivamente. Segmentos diagonais s\u00e3o produzidos por tra\u00e7os.\n\nDISCUSS\u00c3O\n=========\n\nEmbora EBUS-TBNA seja um novo m\u00e9todo, seu uso no diagn\u00f3stico e estadiamento das doen\u00e7as pulmonares vem sendo cada vez mais difundido. Com EBUS-TBNA, \u00e9 poss\u00edvel puncionar uma les\u00e3o enquanto a mesma est\u00e1 sendo exibida pela sonda do ultrassom. Esse m\u00e9todo tem sido utilizado para o estadiamento do c\u00e2ncer de pulm\u00e3o. Tamb\u00e9m foi demonstrado que ele \u00e9 diagn\u00f3stico para doen\u00e7as pulmonares benignas.^(^ [@B03] ^-^ [@B07] ^)^ No presente estudo, verificamos que a acur\u00e1cia diagn\u00f3stica global de EBUS-TBNA (independentemente da etiologia) foi de 83%. De forma semelhante, Choi et al.^(^ [@B08] ^)^ relataram que EBUS-TBNA apresentou acur\u00e1cia diagn\u00f3stica global de 83,9%. Avaliando nossos casos ao longo dos \u00faltimos quatro anos, verificamos que EBUS-TBNA tem sido bem tolerada e tem provado ser um m\u00e9todo diagn\u00f3stico confi\u00e1vel e com poucas complica\u00e7\u00f5es. O prop\u00f3sito do presente estudo foi compartilhar nossa experi\u00eancia sobre o valor diagn\u00f3stico de EBUS-TBNA.\n\nNo diagn\u00f3stico de sarcoidose, EBUS-TBNA \u00e9 frequentemente utilizada, pois pode evitar a necessidade de TBNA convencional ou, em alguns casos, de mediastinoscopia.^(^ [@B04] ^,^ [@B05] ^,^ [@B09] ^)^ O uso de EBUS-TBNA garante que a amostra foi colhida dos linfonodos alvo. \u00c9 a\u00ed que reside sua superioridade em rela\u00e7\u00e3o \u00e0 TBNA convencional.^(^ [@B10] ^,^ [@B11] ^)^ Al\u00e9m disso, EBUS-TBNA n\u00e3o \u00e9 t\u00e3o invasiva quanto a mediastinoscopia. Diversos n\u00edveis de acur\u00e1cia diagn\u00f3stica na sarcoidose j\u00e1 foram relatados para EBUS-TBNA. Uma recente meta-an\u00e1lise demonstrou que a acur\u00e1cia diagn\u00f3stica relatada de EBUS-TBNA na sarcoidose variou de 54% a 93%. ^(^ [@B06] ^)^ Em nosso estudo, esse n\u00edvel foi de 86%, o que est\u00e1 de acordo com os achados de outros estudos na literatura. Navani et al.^(^ [@B12] ^)^ avaliaram pacientes com suspeita de sarcoidose utilizando EBUS-TBNA e relataram uma acur\u00e1cia diagn\u00f3stica de 88%. Quando os mesmos pacientes foram submetidos a bi\u00f3psia pulmonar transbr\u00f4nquica, bi\u00f3psia de mucosa e LBA via broncoscopia padr\u00e3o, a acur\u00e1cia diagn\u00f3stica aumentou para 93%. A variabilidade dos n\u00edveis relatados de acur\u00e1cia diagn\u00f3stica entre os estudos poderia refletir o n\u00edvel de experi\u00eancia do pneumologista que realizou EBUS-TBNA ou o do patologista, bem como a diferen\u00e7a entre estudos em que foi realizada citopatologia r\u00e1pida no local e aqueles em que a mesma n\u00e3o foi realizada.\n\nEm outra doen\u00e7a benigna, a tuberculose, EBUS-TBNA tamb\u00e9m \u00e9 utilizada como m\u00e9todo diagn\u00f3stico. Sun et al.^(^ [@B07] ^)^ e Cetinkaya et al.^(^ [@B13] ^)^ relataram a sensibilidade de EBUS-TBNA na tuberculose como sendo de 85% e 79%, respectivamente,^(^ [@B07] ^,^ [@B13] ^)^ enquanto a encontrada no presente estudo foi de 74%. A acur\u00e1cia diagn\u00f3stica de EBUS-TBNA \u00e9 menor para tuberculose do que para sarcoidose, e a TBNA convencional apresenta acur\u00e1cia diagn\u00f3stica de 65% para tuberculose.^(^ [@B14] ^)^ Ren et al.^(^ [@B15] ^)^ verificaram que, em pacientes com suspeita de linfadenite tuberculosa mediast\u00ednica, a broncoscopia padr\u00e3o apresentou sensibilidade de 18,1% e especificidade de 100%. Quando os autores adicionaram EBUS-TBNA \u00e0 broncoscopia padr\u00e3o, esses valores subiram para 80% e 92,3%, respectivamente.^(^ [@B15] ^)^ Portanto, EBUS-TBNA parece contribuir para a acur\u00e1cia diagn\u00f3stica em pacientes com linfadenite tuberculosa e aumento dos linfonodos mediastinais. No presente estudo, tamb\u00e9m verificamos que a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA foi menor para tuberculose do que para sarcoidose. A necrose proeminente na tuberculose paucibacilar poderia explicar esse achado. Dhooria et al.^(^ [@B16] ^)^ relataram que se pode distinguir entre tuberculose e sarcoidose com base em imagens de EBUS. Por conseguinte, quando combinadas com teste tubercul\u00ednico positivo, a ecotextura heterog\u00eanea da estrutura interna ou a apar\u00eancia necr\u00f3tica visualizada por EBUS podem diagnosticar tuberculose com especificidade de 98% e valor preditivo positivo de 91%.^(^ [@B16] ^)^\n\nUm estudo recente relatou que EBUS-TBNA apresenta acur\u00e1cia diagn\u00f3stica de 93,9% para malignidade,^(^ [@B08] ^)^ enquanto no presente estudo a mesma foi de apenas 77%. Esse mesmo estudo relatou que a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA foi maior para doen\u00e7as malignas do que para doen\u00e7as benignas (93,9% vs. 70,6%, p \\< 0,001). Em nosso estudo, a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA para dist\u00farbios granulomatosos benignos n\u00e3o foi significativamente diferente daquela observada para doen\u00e7as malignas (83% vs. 77%, p \\> 0,05). Memoli et al.^(^ [@B17] ^)^ mostraram que o aumento do tamanho linfonodal apoia o diagn\u00f3stico de malignidade baseado em EBUS. Abu-Hijleh et al.^(^ [@B18] ^)^ sugeriram que o tamanho linfonodal tem pouco efeito sobre a acur\u00e1cia diagn\u00f3stica e a sensibilidade de EBUS-TBNA. Em contraste, quando o tamanho linfonodal excede 20 mm, o valor preditivo negativo diminui. Esses autores verificaram que amostras colhidas de linfonodos \u2265 20 mm de tamanho sempre continham material suficiente para an\u00e1lise. Os mesmos autores relataram que pun\u00e7\u00f5es dos linfonodos 4R, 4L, 7, 10/11R e 10/11L n\u00e3o diferiram quanto \u00e0 acur\u00e1cia diagn\u00f3stica. ^(^ [@B18] ^)^ Tedde et al.^(^ [@B19] ^)^ mostraram que a identifica\u00e7\u00e3o de linfadenopatia em m\u00faltiplos linfonodos e pun\u00e7\u00e3o de linfonodos subcarinais proporciona maior rendimento diagn\u00f3stico em EBUS-TBNA. Esses autores relataram que EBUS-TBNA apresenta acur\u00e1cia diagn\u00f3stica de 57% para malignidade. No presente estudo, os linfonodos com di\u00e2metro \u2265 16,5 mm no menor eixo se mostraram mais diagn\u00f3sticos na avalia\u00e7\u00e3o histopatol\u00f3gica do que aqueles com di\u00e2metros menores. Tamb\u00e9m verificamos que a acur\u00e1cia diagn\u00f3stica de EBUS-TBNA foi melhor quando havia aumento dos linfonodos mediastinais do que quando havia aumento dos linfonodos hilares. Em um estudo realizado por Cetinkaya et al.,^(^ [@B20] ^)^ a sensibilidade de EBUS-TBNA foi menor na regi\u00e3o do linfonodo 4L do que nas dos outros linfonodos. Os autores relataram que a acur\u00e1cia diagn\u00f3stica global de EBUS-TBNA n\u00e3o foi afetada pelo n\u00famero de linfonodos puncionados, n\u00famero de vezes que uma regi\u00e3o linfonodal foi puncionada ou tamanho dos linfonodos.^(^ [@B20] ^)^\n\nEm nossa amostra, observamos que EBUS-TBNA provocou apenas complica\u00e7\u00f5es menores, como hip\u00f3xia leve, taquicardia e pequena hemorragia. Nenhuma dessas complica\u00e7\u00f5es alterou o curso do procedimento ou levou a tratamento adicional ou hospitaliza\u00e7\u00e3o. Podemos, portanto, afirmar que o procedimento pode ser utilizado com seguran\u00e7a em todas as faixas et\u00e1rias adultas. Estudos semelhantes na literatura tamb\u00e9m relataram que EBUS-TBNA \u00e9 segura e n\u00e3o causa complica\u00e7\u00f5es graves.^(^ [@B08] ^,^ [@B20] ^-^ [@B23] ^)^\n\nEm conclus\u00e3o, EBUS-TBNA \u00e9 um m\u00e9todo comprovado no diagn\u00f3stico e estadiamento de malignidade. A acur\u00e1cia diagn\u00f3stica desse m\u00e9todo tamb\u00e9m \u00e9 alta em patologias benignas como sarcoidose e tuberculose. \u00c9 um m\u00e9todo bem tolerado e minimamente invasivo, com baixa taxa de complica\u00e7\u00f5es e alta acur\u00e1cia diagn\u00f3stica. O uso de EBUS-TBNA deve ser incentivado como o procedimento de escolha para patologias pulmonares mediastinais e hilares, devido \u00e0 redu\u00e7\u00e3o significativa da necessidade de mediastinoscopia.\n\nAgradecemos ao Sr. David Chapman a revis\u00e3o lingu\u00edstica do manuscrito.\n\nTrabalho realizado no Department of Pulmonology, Istanbul Yedikule Chest Diseases and Thoracic Surgery Training and Research Hospital, Istanbul, Turkey.\n\nApoio Financeiro: Nenhum.\n\n[^1]: EBUS-TBNA: endobronchial ultrasound-guided transbronchial needle aspiration. ^a^Lymphoma (n = 2); plasmacytoma (n = 1); and esophageal cancer (n = 1). ^b^Nocardiosis (n = 2); and cyst (n = 1).\n\n[^2]: Values expressed as median (min-max).\n\n[^3]: EBUS-TBNA: *endobronchial ultrasound-guided transbronchial needle aspiration*(pun\u00e7\u00e3o aspirativa por agulha guiada por ultrassom endobr\u00f4nquico). aLinfoma (n = 2); plasmocitoma (n = 1); e c\u00e2ncer de es\u00f4fago (n = 1). bNocardiose (n = 2); e cisto (n = 1).\n\n[^4]: aValores expressos em mediana (m\u00edn-m\u00e1x).\n"}
+{"text": "![](glasgowmedj75418-0018){#sp1 .210}\n"}
+{"text": "![](transepisoclond76089-0079){#sp1 .75}\n\n![](transepisoclond76089-0080){#sp2 .76}\n\n![](transepisoclond76089-0081){#sp3 .77}\n\n![](transepisoclond76089-0082){#sp4 .78}\n"}
+{"text": "Introduction {#S1}\n============\n\nHepatocellular carcinoma (HCC) is the fifth most common cancer in men and the seventh in women, and is now the second most common cause of cancer-related death worldwide ([@B4]; [@B21]). Patients with advanced HCC have limited treatment options, and chemotherapy provides minimal survival benefit. Lack of knowledge of the mechanisms of HCC is one of the important factors that limit the development of HCC treatment. Therefore, understanding the underlying mechanisms of HCC is still urgently needed to explore novel therapeutic options.\n\nHepatitis B or hepatitis C is the leading risk factor for HCC, and other risk factors include obesity, diabetes and related nonalcoholic fatty liver disease (NAFLD) ([@B49]; [@B30]). Chronic inflammation is closely associated with persistent hepatic injury and concurrent regeneration, leading to sequential development of fibrosis, cirrhosis, and eventually HCC ([@B3]). However, cases of HCC in the absence of inflammation and cirrhosis are not rare in clinics ([@B16]; [@B32]). Most studies have focused on the mechanisms of chronic-inflammation-based HCC, while few studies have been conducted on the molecular basis of non-inflammatory HCC in animal models ([@B33]). The mechanisms of HCC in the absence of inflammation are still undefined and need further investigation.\n\nThe mammalian target of rapamycin (mTOR) pathway is aberrantly up-regulated in up to 50% of HCCs ([@B2]). The mTOR kinase nucleates two distinct protein complexes termed mTOR complex 1 (mTORC1) and complex 2 (mTORC2). The mTOR pathway regulates key cellular functions linked to the promotion of cell growth and metabolism ([@B15]). Various signaling pathways upstream of mTORC1 stimulate its activity through inhibition of the tuberous sclerosis (TSC) 1--TSC2 complex. Disruption of this complex, through the loss of TSC1 or TSC2, results in ectopic activation of mTORC1 ([@B25]). Dysfunction of hepatic TSC1, a suppressor of mTOR signaling, occurs in patients with hepatitis B, hepatitis C or insulin resistance, which is a risk factor for HCC development ([@B26]; [@B40]; [@B48]; [@B17]). [@B29] detected sporadic HCC development in a liver-specific *TSC1* knockout (*LTsc1KO*) mouse model and proposed that the liver damaging, inflammatory cycles of necrosis and regeneration may contribute to the development of HCC in *LTsc1KO* mice. However, in this study, we did not find any obvious histological evidence of inflammation, necrosis and fibrosis in *LTsc1KO* HCC mice. These mice are a good model to investigate the underlying molecular mechanisms of HCC development in the absence of inflammation. We utilized a multiomics strategy including transcriptomics, cytokine proteomics, genomics and metabolomics, to determine the comprehensive molecular basis for without long-term hepatic inflammation, necrosis, or fibrosis.\n\nOur findings demonstrate that inflammation and fibrosis are not prerequisites in HCC development triggered by *TSC1* deficiency. Non-inflammatory HCC developed by *LTsc1KO* displayed molecular characteristics that were similar to those of diethylnitrosamine (DEN)-mediated non-cirrhotic HCC ([@B6]). Mitochondrial and autophagy defects, as well as hepatic steatosis were manifested in HCC. mTORC1 activation on its own regulated a gene network, including DNA-damage-inducible transcript 4 (Ddit4), nuclear protein 1 (Nurp1) and fibroblast growth factor 21 (FGF21), and mTORC1--signal transducer and activator of transcription pathway crosstalk that altered specific metabolic pathways contributed to the development of non-inflammatory HCC.\n\nMaterials and Methods {#S2}\n=====================\n\nGeneration of Mice {#S2.SS1}\n------------------\n\nMice carrying the *Tsc1*^fl^ allele in the FVB background have been previous described ([@B47]). *LTsc1KO* mice were generated by crossing *TSC1*^fl/*fl*^ mice with *Alb-Cre* mice (C57BL/6J background) obtained from Jackson Laboratory (Farmington, CT, United States) ([@B24]). The specificity of recombination was confirmed by PCR using primers flanking the floxed allele. The primer sequences are listed in [Supplementary Table S1](#TS1){ref-type=\"supplementary-material\"}. All procedures involving mice was approved by the Southern Medical University Animal Care and Use Committee. Mice importing, transporting, housing, and breeding were conducted according to the recommendations of \"The use of non-human primates in research.\"\n\nMouse Diets and Treatments {#S2.SS2}\n--------------------------\n\nThe mice were housed in plastic cages at a controlled temperature of 22 \u00b1 1\u00b0C on a 12 h light/12 h dark cycle with lights on from 06:00 to 18:00 h. Standard rodent chow and water were provided *ad libitum* throughout the entire feeding period. All of the animal experiments were approved by the Animal Ethics Committee of Southern Medical University (approval number SYXK 2011-0074) and performed in accordance with animal ethics guidelines and approved protocols. Mice aged 5 months were subjected to normal saline (NS) or rapamycin gavage. For rapamycin treatment, mice were administered by oral gavage 5 mg/kg/day rapamycin until killing at 10 months of age, and the same volume of NS was administered to control mice. All mice were killed prior to their daily feeding.\n\nStatistical Analysis {#S2.SS3}\n--------------------\n\nContinuous variables were compared using independent sample *t*-tests or one-way analysis of variance, with the data expressed as mean \u00b1 SEM. Significant differences are indicated in the figure legends, unless otherwise indicated.\n\nOther detailed experimental procedures are described in [Supplementary Material](#FS1){ref-type=\"supplementary-material\"}.\n\nResults {#S3}\n=======\n\n*LTsc1KO* Mice Developed Spontaneous HCC Without Systemic Glucose Tolerance {#S3.SS1}\n---------------------------------------------------------------------------\n\nWe generated *LTsc1KO* mice by crossing *TSC1*^fl/fl^ with *Alb-Cre* mice. Deletion of the hepatic *TSC1* allele was confirmed by polymerase chain reaction (PCR) of genomic DNA ([Supplementary Figure S1A](#FS1){ref-type=\"supplementary-material\"}). Immunoblot analysis of *LTsc1KO* liver revealed constitutively active mTORC1 signaling in the liver, as indicated by TSC1 and the best-characterized substrates of mTORC1 S6 phosphorylation (pS6). *LTsc1KO* mice displayed a strong reduction in phospho-Ser473-PKB/Akt (p-Akt). As described previously ([@B47]), in the present study, the *LTsc1KO* livers display attenuation of Akt signaling due to a mTORC1-dependent negative feedback loop ([Figure 1A](#F1){ref-type=\"fig\"}). Surprisingly, serum parameters showed that *LTsc1KO* mice displayed unchanged blood glucose, triglyceride, lactate dehydrogenase (LDH), and cholesterol levels relative to *TSC1*^fl/fl^ mice, indicating a normal insulin response in the *LTsc1KO* mice ([Figures 1B--E](#F1){ref-type=\"fig\"}). Compared with the *TSC1*^fl/fl^ mice, *LTsc1KO* mice did not have any decrease in body weight by 6 and 10--14 months.\n\n![LTsc1KO mice developed spontaneous HCC. **(A)** Immunoblot analysis of isolated hepatocytes from *LTsc1KO* and *TSC1*^fl/fl^ mice with TSC1, S6, Akt, and their phosphorylation antibodies to confirm activation of mTORC1. **(B)** Blood glucose levels in 6- and 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice, respectively (*n* = 45, \u00b1SEM). **(C--E)** Serum LDH, cholesterol, and triglyceride levels were observed in 6- and 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice, respectively. **(F)** Liver weights in 6- and 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice (*n* = 45, \u00b1SEM), and liver weight was increased in 10--14-month-old *LTsc1KO* comparing with *TSC1*^fl/fl^ mice (*\\*\\*P* \\< 0.01). **(G)** Liver sections showing histological types of tumor in 10--14-month-old *LTsc1KO* mice. The tumor areas are marked in low magnification and high magnification to show the histological features clearly. **(H)** Cleaved caspase-3 (CC3), p-H2AX, and Ki67 staining of liver sections in 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice. Immunohistochemically positive cells were quantified, and results are shown in the bar graphs (*n* = 19, \u00b1 SEM, *\\*\\*\\*P* \\< 0.001). **(I)** Hepatic CC3 and p53 protein expression in *LTsc1KO* and *TSC1*^fl/fl^ mice.](fcell-08-00225-g001){#F1}\n\nThere were no detectable tumors in the *LTsc1KO* mice by 6 months, while 31 of 45 (68.89%) *LTsc1KO* mice spontaneously developed HCC by 10--14 months. An average of three macroscopic tumors without encapsulation per liver was detected. Multifocal tumors were randomly distributed in all liver lobes ([Supplementary Figure S1B](#FS1){ref-type=\"supplementary-material\"}). Consistently, these *LTsc1KO* mice exhibited increased liver weight by 10--14 months ([Figure 1F](#F1){ref-type=\"fig\"}).\n\nPreviously, tumor developed by *TSC1* deficiency has been considered as benign with a high rate of apoptosis ([@B44]). In the present study, histopathological features of HCC in *LTsc1KO* mice revealed large cells with enlarged and hyperchromatic nuclei and ballooning with weakly eosinophilic staining, which contained numerous microvesicular vacuoles in cytoplasm ([Figure 1G](#F1){ref-type=\"fig\"}). *LTsc1KO* did not induce significant apoptosis, as measured by cleaved caspase 3. Phosphorylated histone-2AX (\u03b3H2AX)-positive hepatocytes, was increased in *LTsc1KO* livers ([Figure 1H](#F1){ref-type=\"fig\"}). p53 level was increased in livers of *LTsc1KO* mice. These data demonstrate that loss of *TSC1* results in dramatic accumulation of p53 in response to the p53-mediated DNA damage-response pathway ([Figure 1I](#F1){ref-type=\"fig\"}).\n\nHepatic Necroinflammation Was Not a Prerequisite for Spontaneous HCC in *LTsc1KO* Mice {#S3.SS2}\n--------------------------------------------------------------------------------------\n\nWe found marked elevation of serum ALT and AST in *LTsc1KO* mice at 10--14 months but not at 6 months, indicating no long-term hepatic damage in *LTsc1KO* mice ([Figure 2A](#F2){ref-type=\"fig\"}). No significant changes in the number of infiltrative lymphocytes were observed in *LTsc1KO* mice as compared with *TSC1*^fl/fl^ mice. F4/80 staining suggested that the number of macrophages (Kupffer cells) remained unchanged in the liver from *LTsc1KO* mice at 6 and 10--14 months ([Figure 2B](#F2){ref-type=\"fig\"}).\n\n![Spontaneous HCC development in *LTsc1KO* mice lacked necroinflammation and fibrosis. **(A)** Serum ALT and AST levels in 6- and 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice (*n* = 29, \u00b1SEM). ALT and AST only increased in 10--14-month-old *LTsc1KO* compared with *TSC1*^fl/fl^ mice (*\\*\\*P* = 0.006, and *\\*P* = 0.046). **(B)** H&E, F4/80 and p-STAT3 immunohistochemical staining of liver sections in 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice. Immunohistochemically positive cells were quantified, and results are shown in the bar graphs (*n* = 19, \u00b1 SEM, *\\*P* \\< 0.05). **(C)** Hepatic levels of inflammatory cytokines in 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice (*n* = 19, \u00b1SEM). **(D)** Immunoblot analysis of hepatic STAT3 and p-STAT3 in *LTsc1KO* and *TSC1*^fl/fl^ mice aged 10--14 months. **(E)** Immunoblot analysis of hepatic STAT3 and p-STAT3 in tumor and non-tumor areas of *LTsc1KO* mice. **(F)** Fibrosis was analyzed by staining liver sections with Sirius red in 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice. **(G,H)** Serum and hepatic GPX3 level in *LTsc1KO* and *TSC1*^fl/fl^ 10--14-month-old mice (*n* = 14, \u00b1SEM).](fcell-08-00225-g002){#F2}\n\nTo explore further whether there was chronic hepatic inflammation in *LTsc1KO* mice, we conducted a Bio-Plex mouse cytokine 23-plex assay to detect simultaneously 23 cytokines, chemokines and growth factors. It has been proved that increased IL-6 production results in activation of the oncogenic transcription factor STAT3 in liver, and TNF-\u03b1 is a major adipose- derived cytokine and potent activator of pro-oncogenic pathways, including mTOR. The cytokine assay revealed no increased expression of inflammatory cytokines, including TNF-\u03b1 and IL-6 ([Figures 2C](#F2){ref-type=\"fig\"} and [Supplementary Figure S1C](#FS1){ref-type=\"supplementary-material\"}). Taken together, these results indicated that, the development of HCC in *LTsc1KO* mice was not dependent on long-term injury and necroinflammation.\n\nWe found that levels of STAT3 phosphorylation decreased in the livers of *LTsc1KO* mice as compared with *TSC1*^fl/fl^ mice, and in tumors as compared with non-tumor areas of the liver in *LTsc1KO* mice ([Figures 2D,E](#F2){ref-type=\"fig\"}). It is worth noting that activity of serum and cellular GPX3 in *LTsc1KO* mice was similar to that in *TSC1*^fl/fl^ mice ([Figures 2G,H](#F2){ref-type=\"fig\"}). We therefore propose that HCC without necroinflammation developed possibly through inactivation of STAT3 and/or other mutations. However, we cannot completely exclude the possibility of additional mechanisms by which mTOR affects STAT3 phosphorylation.\n\nSpontaneous HCC in *LTsc1KO* Mice Displayed Non-cirrhotic and Chemical Carcinogenic Characteristics {#S3.SS3}\n---------------------------------------------------------------------------------------------------\n\nCirrhosis is not always a prerequisite of HCC development and this might particularly apply to the HCC associated with metabolic abnormality. Sirius red staining showed that except for increased blood vessels, no obvious fibrosis was discovered, which demonstrated that hepatic mTORC1 activation did not induce liver fibrosis ([Figure 2F](#F2){ref-type=\"fig\"}). To investigate the molecular characteristics of HCC in *LTsc1KO* mice, we conducted RNA sequencing (RNA-seq) on mice at 6 and 10--14 months. The pattern of gene expression in the livers of *LTsc1KO* mice was also induced in non-cirrhotic DEN-mediated HCC, including *Tff3, Ly6d, Gpc3*, and *Afp* ([Figure 3A](#F3){ref-type=\"fig\"}), and several other fetal genes that were induced in human HCC, such as *Bex2, Spink1*, and *Rnase1*. The change in Tff3 protein was confirmed by western blotting ([Figure 5C](#F5){ref-type=\"fig\"}).\n\n![Spontaneous HCC in *LTsc1KO* mice showed expression of genes commonly induced in chemical HCC. **(A)** qRT-PCR showing increased hepatic mRNA levels for expression of genes that are frequently induced in *LTsc1KO* mice aged 6 and 10--14 months, which were similar to human DEN-mediated HCC, such as *Tff3*, *Gpc3*, and *Ly6d* (*n* = 19, \u00b1SEM). **(B)** Illustration of differentially expressed genes in HCC. Heat-map indicates mean values of differential expression for each gene. **(C)** Histographs of differentially expressed genes by pathway analysis.](fcell-08-00225-g003){#F3}\n\nFurther analysis of the RNA-seq data revealed that the aberrant expression of enzyme genes were mainly responsible for detoxification ([Figure 3B](#F3){ref-type=\"fig\"}), including cytochrome P450 (CYP450), carboxylesterases (Ces), sulfotransferases (Sults), and UDP-glucuronosyltransferases (Ugts). The activities of xenobiotic metabolizing enzymes are required for activation (toxication) of important carcinogens. CYP450 enzymes are key players in the phase-I-dependent metabolism of xenobiotics. We found reduced expression of Cyp1a2, Cyp2b9, Cyp2c50, Cyp2c54, Cyp2c67, Cyp2e1, and Cyp3a16, and increased expression of Cyp2b10 in the livers of *LTsc1KO* mice. Further gene ontology (GO) enrichment analysis revealed that differentially expressed genes were significantly enriched in biological process of chemical carcinogenesis ([Figure 3C](#F3){ref-type=\"fig\"}). We suggest that the first stage of carcinogenesis induced by mTORC1 activation is similar to that induced by chemical carcinogens, which results in cells with altered metabolic responsiveness and a proliferative advantage over the surrounding normal cells.\n\n*LTsc1KO*-Induced Spontaneous HCC Was Also Characterized by Metabolic Disorder {#S3.SS4}\n------------------------------------------------------------------------------\n\nWe checked whether metabolic disorder occurred. Periodic acid-Schiff (PAS) staining confirmed that there was less glycogen in HCC foci than adjacent hepatic tissues ([Figure 4A](#F4){ref-type=\"fig\"}). The metabolome was investigated in livers at 6 and 10--14 months ([Figures 4B,C](#F4){ref-type=\"fig\"}). Consistent with reduction of Ugts1a by RNA-seq, the metabolomic changes were loss of glucuronic acid, with dramatically altered cellular metabolism at 10--14 months compared with 6 months ([Figure 4D](#F4){ref-type=\"fig\"}).\n\n![Characterization of *LTsc1KO* liver-bearing carcinoma by glucuronic acid loss and metabolic disorder. **(A)** Liver sections were stained for glycogen with PAS. **(B)** Representative pictures of the significantly changed metabolites by --log10 *P*-value. **(C)** Metabolic pathway map in *LTsc1KO* mice aged 10--14 months. Green, decrease; red, increase. **(D)** Concentrations of hepatic glucuronic acid and tyrosine in *LTsc1KO* and *TSC1*^fl/fl^ mice aged 6 and 10--14 months (*n* = 19, \u00b1SEM). **(E)** A differential score was calculated for the KEGG pathway. Size of dots indicates the impact of the pathway and the color indicates pathways with a score of -- log10 *P*-value. **(F)** Metabolic shifts in the tricarboxylic acid cycle in *LTsc1KO* mice aged 10--14 month. Green, decrease; gray, not measured.](fcell-08-00225-g004){#F4}\n\nPathway analysis revealed that many related metabolic pathways that influence various biological processes, such as pentose and glucuronate interconversions (carbohydrate metabolism subcategory), and the biosynthesis of ubiquinones ([Figure 4E](#F4){ref-type=\"fig\"}). Integrated omics profiling revealed that carcinogenesis of spontaneous HCC in *LTsc1KO* mice was similar to chemical carcinogenesis. Therefore, HCC induced by *LTsc1KO* was characterized by decreased glucose and TCA cycle disorder ([Figure 4F](#F4){ref-type=\"fig\"}). These findings suggest that specific metabolic changes, commonly seen in human hepatic carcinoma, might contribute to tumor progression in this model.\n\nHepatocarcinogenesis Was Subsequent to Defects Upstream and Downstream of mTORC1 {#S3.SS5}\n--------------------------------------------------------------------------------\n\nRNA-seq analysis identified the dysregulated genes upstream and downstream of mTORC1, including increased mRNA level of *Ragd, Ddit4, Npur1*, and *FGF21* ([Figure 5A](#F5){ref-type=\"fig\"}). Those gene changes were validated by qRT-PCR, and western blotting also revealed lower expression of Ragd, Ddit4, Npur1, and FGF21 proteins in *LTsc1KO* liver ([Figures 5B,C](#F5){ref-type=\"fig\"}), indicating that these proteins may undergo post-transcriptional modification. And no additional mutations in Ddit4, FGF21, and Nupr1 were identified ([Figure 5D](#F5){ref-type=\"fig\"}), indicating that protein mutations were not required for those genes after mTORC1 activation for liver carcinogenesis. While expression of Dddit4, FGF21, and Nupr1 was decreased in tumor compared to non-tumor tissue ([Figure 6A](#F6){ref-type=\"fig\"}). These results indicate that HCC development in *LTsc1KO* mice involves a complex network that requires Dddit4, FGF21, and Nupr1. Therefore, a cohort of mice aged 5 months was treated with rapamycin or NS three times a week for 5 months. Rapamycin-treated *LTsc1KO* livers showed reduced LC3B-II, and increased Ddit4 and Nupr1. NS-treated *LTsc1KO* mice developed HCC at a rate similar to our previous cohort. Blocking mTORC1 significantly rescued the decreased expression of Nupr1 and Ddit4 in *LTsc1KO* liver, indicating that the upstream regulatory network dictates the pathological consequences of dysregulated mTOR signaling in liver carcinogenesis ([Figure 6B](#F6){ref-type=\"fig\"}). Although autophagy-related gene 5 (Atg5) was unchanged, increased levels of LC3B-II (*Map1lc3b-II*) were observed in *LTsc1KO* compared to *TSC1*^fl/fl^ mice, we also found that livers from *LTsc1KO* mice at 6 months age displayed accumulation of the autophagy substrate p62, which is believed to target ubiquitinated proteins and is selectively degraded by autophagy. The abundance of p62 was increased in *LTsc1KO* livers ([Figure 6C](#F6){ref-type=\"fig\"}), indicating that mTORC1 activation by *TSC1* loss caused a defect in autophagy flux. We conclude that long-term activation of mTORC1 by *LTsc1KO*, upstream regulators (Dddit4) of mTORC1 was impaired; meanwhile, expression of downstream genes (Nupr1 and FGF21) of mTORC1 was decreased, resulting in metabolic disorder and mitochondrial defects.\n\n![Liver carcinogenesis subsequent to defects in mTORC1 network, and autophagy in *LTsc1KO* mice. **(A)** qRT-PCR showing hepatic *Nupr1, Ragd, FGF21*, and *Ddit4* mRNA levels in *LTsc1KO* compared with *TSC1*^fl/fl^ mice aged 6 and 10--14 months (*n* = 19, \u00b1SEM). **(B)** Immunoblot analysis of hepatic Akt, p-Akt, S6, p-S6, STAT3, p-STAT3, Ddit4, and Nupr1 in *LTsc1KO* and *TSC1*^fl/fl^ mice aged 10--14 months. **(C)** Immunoblot analysis of hepatic Ragd, FGF21, Atg5, LC3B-II, PGC1-\u03b1, and Tff3 in *LTsc1KO* and *TSC1*^fl/fl^ mice aged 10--14 months. **(D)** Representative images of mutational profile in tumor developed in *LTsc1KO* mice aged 10--14 months, by RNA-seq.](fcell-08-00225-g005){#F5}\n\n![Liver Carcinogenesis Enhanced by the mTORC1--STAT Pathway Crosstalk. **(A)** Immunoblot analysis of liver lysates from tumor and non-tumor areas in *LTsc1KO* mice aged 10--14 months. **(B)** Immunoblot analysis of liver lysates from *LTsc1KO* mice, aged 10 months, treated with vehicle or rapamycin for the preceding 5 months. **(C)** Immunoblot analysis of liver lysates from *LTsc1KO* and *TSC1*^fl/fl^ mice aged 6 months. **(D)** FGF21 immunohistochemical staining of early and late stage liver sections from *LTsc1KO* and *TSC1*^fl/fl^ mice.](fcell-08-00225-g006){#F6}\n\nDysregulation of FGF21 Was a Later Event for Liver Carcinogenesis in *LTsc1KO* Mice {#S3.SS6}\n-----------------------------------------------------------------------------------\n\nThere was no obvious change in expression of FGF21 in *LTsc1KO* mice treated with rapamycin, indicating that altered expression of FGF21 might be a critical event after long-term abrogation of Ddit4 and Nupr1 for HCC development ([Figure 6B](#F6){ref-type=\"fig\"}). Unchanged expression of FGF21 and decreased expression of Ddit4 and Nupr1 were confirmed in *LTsc1KO* mice at 6 months, indicating that FGF21 expression was an essential event during liver carcinogenesis ([Figure 6C](#F6){ref-type=\"fig\"}). FGF21 protein level increases in liver tissues at an early stage, but decreases in liver tissues later when HCC develops ([Figure 6D](#F6){ref-type=\"fig\"}). Loss of FGF21 may play an important role in HCC carcinogenetic transformation during metabolic liver injury in *LTsc1KO* mice. Metabolic remodeling associated with FGF21 expression also requires induction of liver-integrated stress-response-driven Nupr1 ([@B28]). In the present study, Nupr1 was decreased in *LTsc1KO* mice by 10--14 months, indicating its involvement in maintaining mitochondrial defects, thereby contributing to the lipotoxic effects of liver carcinogenesis. mTORC1 negatively regulates hepatic FGF21 expression via peroxisome proliferator-activated receptor-\u03b3 coactivator 1\u03b1 (PGC-1\u03b1) ([@B11]). mTORC1 also controls the transcriptional activity of PGC1-\u03b1, which induces mitochondrial biogenesis in the liver ([@B38]). At 6 months of age, mice without spontaneous HCC revealed increased PGC-1\u03b1 in *LTsc1KO* mice ([Figure 6C](#F6){ref-type=\"fig\"}). Conversely, at 10--14 months of age, PGC-1\u03b1 was markedly reduced in mice with HCC ([Figure 5C](#F5){ref-type=\"fig\"}). Our data also confirm that altered expression of FGF21 and Nupr1 is required during liver carcinogenesis.\n\nCrosstalk Between mTORC1 and the STAT3 Pathway Was Linked to Non-inflammatory HCC in *LTsc1KO* Mice {#S3.SS7}\n---------------------------------------------------------------------------------------------------\n\nDdit4 is a constant modulator of p-Akt in response to growth factors and nutrients ([@B8]), and its expression depends on activation of STAT3 ([@B36]). In the present study, *LTsc1KO* mice demonstrated reduced expression of p-Akt, p-STAT3 and Ddit4 ([Figure 5B](#F5){ref-type=\"fig\"}). Therefore, we posit that Ddit4 might act as a mediator of crosstalk between mTOR and the STAT pathway. Correspondingly, lower expression of p-STAT3 were observed in *LTsc1KO* mice at 6 months, and *LTsc1KO* mice exhibited lower expression of p-STAT3 in tumor compared to non-tumor tissue. Rapamycin-treated *LTsc1KO* livers showed increased p-STAT3 expression compared to livers from NS-treated *LTsc1KO* mice. We therefore propose that the crosstalk between mTORC1 and the STAT3 pathway and/or other mutations was involved in non-inflammatory HCC development.\n\nDiscussion {#S4}\n==========\n\nGiven that mTORC1 regulates glucose homeostasis, lipid metabolism and cell proliferation, it is not surprising that mTOR plays a pivotal role in HCC development, and TSC1/2 mutations define a molecular subset of HCC with aggressive behavior ([@B17]). However, mechanisms of mTORC1 activation-associated HCC are not well understood. It has become necessary to understand the consequences of mTORC1 activation. To address this important issue, we aimed to provide insight into the mechanisms of HCC development in *LTsc1KO* mice, which is analogous to human HCC induced by NAFLD without necroinflammation and cirrhosis.\n\nIn the present study, the histopathology revealed two major types of tumor that were morphologically consistent with clear cell HCC and trabecular HCC in *LTsc1KO* livers. Gege expression analysis demonstrated the expression of many genes in the *LTsc1KO* liver that are also induced in human HCC and showed a specific molecular characteristics similar to non-cirrhotic DEN-mediated HCC ([@B7]). Relationships between steatosis, steatohepatitis, cirrhosis and HCC are not necessarily linear and this pattern possibly applies to HCC arising in non-alcoholic, non-cirrhotic liver disease ([@B1]). In NAFLD-associated HCC, there was a 23--50% prevalence of tumors developing in non-cirrhotic livers in Japan, Italy and the US ([@B45]; [@B9]; [@B31]; [@B37]). Therefore, in patients with NAFLD, HCC can arise in the context of non-cirrhotic liver, suggesting a specific carcinogenic pathway. Here, *LTsc1KO* mice provide us with an optimal model to investigate the underlying molecular mechanisms of steatohepatitic HCC development in the absence of inflammation.\n\nHCC is often initiated by death of hepatocytes, resulting in liver injury followed by inflammation. The production of pro-tumorigenic cytokines including IL-6, which induces STAT3 activation in hepatocytes, eventually promotes compensatory proliferation in hepatocytes that have escaped cell death, and subsequently, tumor development. Hepatic deletion of genes in the PI3K--Akt--mTOR pathway, including Pten and Akt1 in Akt2, induces HCC ([@B18]; [@B43]; [@B23]; [@B42]). For example, hepatic Pten deletion results in hepatic injury and cell death, which activates tumor-initiating cells to induce HCC development ([@B13]). Both hepatic Akt1 and Akt2 deletion mice develop HCC, which is associated with liver injury and inflammation via activated STAT3 and IL-6 expression ([@B42]). [@B29] described the liver damage, inflammation, necrosis and regeneration in HCC developed in *LTsc1KO* mice. The most intriguing and unexpected observation that was uncovered in our study was spontaneous HCC in *LTsc1KO* mice without long-term hepatic injury, and inflammation. The number of infiltrative lymphocytes and macrophages (Kupffer cells), and the level of cytokine remained unchanged, in the *LTsc1KO* mice. Moreover, decreased Akt without IL-6 elevation in *LTsc1KO* mice, due to a negative regulation loop, suggested that mTORC1 activation was independent of Akt for HCC development. We revealed a previously unappreciated role for dysregulated mTORC1 signaling in promoting cancer-initiating events via inhibition of STAT3. Previous research suggested that STAT3 was activated in inflammatory HCC. It has been proved that increased IL-6 production results in activation of the oncogenic transcription factor STAT3 in liver, and TNF-\u03b1 is a major adipose- derived cytokine and potent activator of pro-oncogenic pathways, including mTOR. Hepatocyte-specific ablation of the specific mTORC1 subunit Raptor resulted in elevated IL-6 production, activation of STAT3, and enhanced HCC development ([@B41]). It is worth noting that mice with hepatic deletion of STAT5 are more susceptible to hepatocarcinogenesis, and the hepatic deletion of STAT5 induces STAT3 activity ([@B19]). In the current study, we found that STAT5 is highly expressed (data not shown) and thus may inhibit the expression of STAT3. Perhaps, this is evidence to support our findings that mTORC1 might have a negative regulatory effect on STAT3.\n\nObesity, hepatosteatosis, insulin resistance, and chronic mTORC1 activation are associated with a pronounced increase in HCC risk ([@B5]). Our findings elucidate fundamental biochemical properties displayed by mTORC1 activation to achieve their tumor-promoting effect. There is an increasing appreciation of the fact that oncogenic signaling controls the metabolic reprogramming of cancer cells ([@B46]). mTORC1, a master regulator of cellular metabolism, controls lipogenesis and glucose metabolism, however, the mechanisms and critical factors need to be elucidated. It is unclear whether the altered fatty acid composition observed in the *LTsc1KO* liver contributes to tumorigenesis.\n\nThe transformation of chemicals is important in carcinogenesis the Cyp450 families are key enzymes in tumor transformation, and mediate the metabolic activation of numerous pre-carcinogens, such as Cyp1a2 and Cyp2e1, which degrade xenobiotics, steroids and fatty acids. By integrated omics profiling, we incorporated enzyme transcript levels into stearic acid accumulation. We hypothesized that metabolic disorder accompanied with inactivation of xenobiotic metabolizing enzymes, and causes defective necrosis, apoptosis and autophagy, in addition to chronic mTORC1 activation in promoting anabolic growth and proliferation.\n\nTransformed cells survive by acquiring adaptations that allow mTORC1 to continue signaling and are insensitive to the stress of energy, oxygen and nutrient deprivation, which confers a selective growth advantage. Cancer evolves via a multistep process and molecular events in addition to TSC1 deficiency during tumor development. [@B23] described that phosphorylation of InsR was reduced in TSC1-deficiency tumors, which is upstream of mTORC1. A gene network containing FGF21, Ddit4, and Nupr1 could explain at least in part the transition consistent with activation of an adaptive transcriptional survival program.\n\nDdit4 also named Redd1, regulated in development and DNA damage responses-1, is an mTORC1 inhibitor and critical transducer of the cellular response to energy depletion ([@B39]). Abnormalities of Ddit4-mediated signaling promote tumorigenesis ([@B10]; [@B35]). Moreover, Ddit4 expression depends on the activation of STAT3 ([@B36]). In the present study, Ddit4 was decreased and reversed after rapamycin given, indicating mTORC1 negative feedback on Ddit4. Nupr1, a mitochondrial defect-responsive gene, regulates autophagy induced by the lipotoxic effects of excess fatty acid accumulation in cells ([@B22]). Nupr1 has also been identified as a key regulator and metabolic switch in response to mitochondrial damage during liver cancer progression ([@B27]). In the present study, Nupr1 was decreased in *LTsc1KO* mice by 10--14 months, while mTORC1 negatively regulates hepatic FGF21 expression via PGC-1\u03b1. TSC1 deficiency abrogates FGF21-mediated inhibition of mTORC1 ([@B14]). Overexpression of hepatocyte-specific FGF21 could prevent DEN-induced liver carcinogenesis in transgenic mice ([@B20]). Mice lacking FGF21 develop substantial fatty liver and markedly exacerbated accumulation of liver triglycerides, consistent with impaired fatty acid oxidation ([@B12]). Ddit4 is also an emerging link for crosstalk between mTOR and STAT pathways in HCC development, and highlights that consecutive STAT5 activation may partially protect 4E-BP1 phosphorylation ([@B34]). Insufficiency of Nupr1 in *LTsc1KO* mice results in non-sensitivity to a variety of stressors, including respiratory deficiency. Loss of FGF21 worsens metabolic disorder and contributes to aberrant molecular events, including lipid metabolism, in HCC development. The dysregulated expression of FGF21 is a later and critical event for HCC development. Furthermore, metabolic remodeling associated with FGF21 expression also requires induction of liver-integrated stress-response-driven Nupr1 ([@B28]). Therefore, our data favor the possibility that *LTsc1KO* mice constructed a regulator network and crosstalk between mTOR and STAT3 has been attributed to oxidative, ER stress, mitochondrial dysfunction, and autophagy defects, adding a potential mechanism to induce and maintain constitutive activation of mTORC1, and mediate a metabolic feature for growth advantage ([Figure 7](#F7){ref-type=\"fig\"}).\n\n![Illustration of complicated gene network of consecutive mTORC1 activation in *LTsc1KO* mice. Long-term activation of mTORC1 in *LTsc1KO* mice. Negative upstream regulators of mTORC1 were shut off, such as Dddit4, meanwhile, expression of Nupr1, and FGF21 downstream of mTORC1 were decreased for mitochondrial defects, autophagy defect and oxidative stress. Red arrow, increase; green arrow, decrease.](fcell-08-00225-g007){#F7}\n\nIn summary, we unraveled a comprehensive gene network and obtained mechanistic insights into how mTORC1 activation as a primary driver results in loss of control over four key homeostatic responses in hepatic cells, comprising metabolic disorder, mitochondrial defect, detoxification loss, and autophagy inhibition. mTORC1 activation, Ddit4 and STAT3 phosphorylation inhibition, indicatingdysregulation of Ddit4-mTORC1-STAT3 lop ultimately contributes to development of non-inflammatory HCC ([Figure 8](#F8){ref-type=\"fig\"}). The HCC that developed in *LTsc1KO* mice in this study provides a typical animal model for future investigation into the molecular events underlying carcinogenesis arising from non-cirrhotic liver disease.\n\n![Proposed Mechanism of HCC Development in *LTsc1KO* Mice. mTORC1 activation results in loss of control over four key homeostatic responses in hepatic cells, comprising mitochondrial defects, detoxification loss and autophagy inhibition, and dysregulation of these adaptive responses (necrosis, apoptosis, autophagy, and proliferation) contribute to HCC development. This model give rise to HCC with gene signatures similar to human chemical HCC.](fcell-08-00225-g008){#F8}\n\nData Availability Statement {#S5}\n===========================\n\nAll datasets generated for this study are included in the article/[Supplementary Material](#FS1){ref-type=\"supplementary-material\"}.\n\nEthics Statement {#S6}\n================\n\nAll procedures involving mice was approved by the Southern Medical University Animal Care and Use Committee.\n\nAuthor Contributions {#S7}\n====================\n\nTL, GZ, XX, and YG designed the experiments. TL generated mice and conducted the most experiments, with assistance from LW. LW and SL made the pathological evaluations of the tissue sections. TL and GZ analyzed all the data, completed the experiments for publication, and wrote the paper. YG conceived the project, and secured funding.\n\nConflict of Interest {#conf1}\n====================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This work was supported by the National Key R&D Program of China (2018YFC1106400 and 018YFA0108200), Science and Technology Planning Project of Guangdong Province (2015B020229002), the National Natural Science Foundation of China (81470875, 81701580 and 81600489), the Natural Science Foundation of Guangdong Province (2014A030312013, 2018A030313128 and 2018A030313214), Guangdong key research and development plan (2019B020234003), and Science and Technology Program of Guangzhou (201604020002 and 201803010086).\n\nSupplementary Material {#S9}\n======================\n\nThe Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fcell.2020.00225/full#supplementary-material>\n\n###### \n\nConsequences of mTORC1 Activation by *LTsc1KO*. **(A)** Representative PCR genotyping of *Alb* (left) and *TSC1* (right). Left: lane 1 and 3: *Alb-Cre* positive, lane 2 and 4: *Alb-Cre* negative, lane 5: water, lane 6: DNA maker 100 bp ladder. Right: lane 1: DNA double-chain *TSC1-loxp* site insertion, lane 2 and 4: no *TSC1-loxp* site insertion, lane 3: DNA single-chain *TSC1-loxp* site insertion, lane 5: water, lane 6: DNA maker 100 bp ladder. *Alb-Cre* recombinase is a site-specific enzyme, could catalyze recombination between two *TSC1-loxp* sites, delation the *TSC1* gene. Take together, lane 1 is *Alb-TSC1*^fl/fl^ mice, also refers to *LTsc1KO* mice, lane 2 and 4 is *TSC1 ^+/+^* mice, also refers to C57BL/6j mice, lane 3 is *Alb-TSC1*^fl/+^ mice. **(B)** (Related to [Figure 1](#F1){ref-type=\"fig\"}). Representative pictures of livers in 10--14 months old *LTsc1KO* and *TSC1*^fl/fl^ mice. **(C)** (Related to [Figure 2](#F2){ref-type=\"fig\"}). Hepatic levels of inflammatory cytokines quantified in 10--14-month-old *LTsc1KO* and *TSC1*^fl/fl^ mice, including IL-3, IL-5, IL-12(P40), IL-12(P70), macrophage inflammatory protein (MIP)1\u03b1, MIP1\u03b2, interferon-\u03b3, and monocyte chemotactic protein 1 (*n* = 19, \u00b1SEM).\n\n###### \n\nClick here for additional data file.\n\n###### \n\nPrimer sequences used for RT-PCR.\n\n###### \n\nClick here for additional data file.\n\nmTOR\n\n:   mammalian target of rapamycin\n\nmTORC1\n\n:   Mammalian target of rapamycin complex 1\n\n*LTsc1KO*\n\n:   liver special TSC1 knockdown\n\nHCC\n\n:   hepatocellular carcinoma\n\nNAFLD\n\n:   nonalcoholic fatty liver disease\n\nTSC1\n\n:   tuberous sclerosis-1\n\nDEN\n\n:   diethylnitrosamine\n\nAKT\n\n:   protein kinase B\n\nDdit4\n\n:   DNA-damage-inducible transcript 4\n\nNurp1\n\n:   nuclear protein 1\n\nFGF21\n\n:   fibroblast growth factor 21\n\n\u03b3H2AX\n\n:   Phosphorylated histone-2AX\n\nALT\n\n:   alanine transaminase\n\nAST\n\n:   aspartate transaminase\n\nTNF\n\n:   tumor necrosis factor\n\nGPX3\n\n:   glutathione peroxidase 3\n\nCYP450\n\n:   cytochrome P450\n\nCes\n\n:   carboxylesterases\n\nSults\n\n:   sulfotransferases\n\nUgts\n\n:   UDP-glucuronosyltransferases\n\nAtg5\n\n:   autophagy-related gene 5\n\nPGC-1\u03b1\n\n:   peroxisome proliferator-activated receptor PPAR-\u03b3 coactivator 1\u03b1\n\nSTAT3\n\n:   signal transducer and activator of transcription 3.\n\n[^1]: Edited by: Zhe-Sheng Chen, St. John's University, United States\n\n[^2]: Reviewed by: Xingxiang Pu, Hunan Cancer Hospital, China; Xiaozhuo Liu, University at Buffalo, United States\n\n[^3]: This article was submitted to Molecular and Cellular Oncology, a section of the journal Frontiers in Cell and Developmental Biology\n"}
+{"text": "Background\n==========\n\nThe conserved Notch pathway participates in diverse aspects of animal development, and has been implicated in human diseases and cancers \\[[@B1]-[@B3]\\]. Notch encodes a transmembrane receptor, which, upon ligand binding, undergoes proteolytic processing and releases an intracellular fragment capable of acting as a transcription co-regulator. As both Notch and its ligands (also transmembrane proteins) are widely expressed, their activities need to be tightly regulated. One such important regulation appears to be ligand internalization, which plays a critical role in activating Notch receptors \\[[@B4],[@B5]\\].\n\nNotch ligand internalization utilizes an ubiquitin-dependent endocytic pathway, as two structurally unrelated E3 ubiquitin ligases, *neuralized (neur)*and *mind bomb (dMib)*, can append ubiquitin to DSL (Delta, Serrate, Lag2) ligands \\[[@B6]-[@B13]\\]. Epsin/*lqf (liquid facets)*then recruits the ubiquitinated DSL ligands into clathrin-coated vesicles (CCVs) \\[[@B14]-[@B18]\\]. The scission of these ligand-containing CCVs from the plasma membrane seems critical for Notch activation, as disruption of dynamin function also causes a *Notch*-like defect \\[[@B4],[@B19],[@B20]\\].\n\nAnother relevant factor in *Drosophila*Notch ligand endocytosis is the J-domain protein auxilin \\[[@B21],[@B22]\\]. First identified in mammals, auxilin is known to cooperate with Hsc70 in mediating the disassembly of clathrin triskelia and coat proteins from newly formed CCVs *in vitro*\\[[@B23]\\]. The mammalian genomes contain two distinct auxilin-related genes: *auxilin*and *GAK*, differing in the presence of an Ark (**a**ctin-**r**elated **k**inase) family kinase domain and their tissue distributions \\[[@B23]-[@B25]\\]. GAK contains the Ark domain at the N-terminus and it is ubiquitously expressed. In contrast, auxilin lacks the kinase domain and its expression appears to be neuronal. However, the expression of auxilin in non-neuronal cells has recently been demonstrated \\[[@B26]\\]. Besides uncoating, several other functions during endocytosis have recently been suggested for auxilin family proteins, including facilitating clathrin exchange during coated-pit formation \\[[@B27],[@B28]\\], participating in pit constriction \\[[@B29]\\] and preventing precipitous assembly of clathrin cages \\[[@B26],[@B30]\\]. Furthermore, GAK has been implicated in clathrin-mediated trafficking from the trans Golgi network (TGN) \\[[@B25],[@B31]\\].\n\nWhile it is unclear which of the aforementioned cellular functions are most relevant under physiological conditions, mutations in the sole *Drosophila*auxilin ortholog (*dAux*) clearly disrupt several Notch-dependent processes \\[[@B21],[@B22]\\]. One such process is the patterning of neural tissues, in which the cells destined to become neurons send a Notch-mediated signal to prevent neighboring cells from adopting neuronal fate. In the absence of this lateral inhibition, supernumerary neuronal cells are generated, forming the so-called neurogenic phenotype \\[[@B2]\\]. Consistent with the notion that *dAux*participates in Notch signaling, excessive neurons were seen in both the embryonic CNS and the larval eye discs of *dAux*mutants \\[[@B21],[@B32]\\]. Mosaic analysis showed that the function of dAux during Notch signaling is required in the signal-sending cells, suggesting that it has a role in ligand internalization \\[[@B32],[@B33]\\].\n\nGiven the high degree of conservation of the Notch pathway, it seems reasonable to expect an inhibition of auxilin function in other animal systems would cause *Notch*-like defects. However, as multiple distinct endocytic pathways exist in higher metazoans \\[[@B34]\\], it is possible that different endocytic pathways can functionally substitute for one another in Notch ligand endocytosis. Inhibition of auxilin function by RNA interference in *C. elegans*was shown to disrupt the receptor-mediated uptake of yolk proteins, yet no Notch-related defects were reported \\[[@B35]\\]. RNAi-mediated reduction of GAK function in mammalian cells appeared to deregulate EGF signaling and promote tumorigenesis \\[[@B36]\\]. Tissue-specific inactivation of mouse *GAK*during embryonic development caused severe degenerations in brain, liver, and skin \\[[@B37]\\], but it is unclear whether any of these defects were due to a disruption of Notch function. Thus, while it seems clear that auxilin family proteins are important for animal viability, whether their function in Notch ligand endocytosis is evolutionarily conserved requires additional investigations.\n\nWe have used zebrafish to further assess the roles of auxilin-dependent endocytosis in animal development. Besides being a versatile model organism, zebrafish is suitable for our purpose because the importance of Notch ligand endocytosis has been demonstrated during embryonic neural patterning \\[[@B9]\\]. We show that zebrafish, like mammals but unlike *Drosophila*, contains both *auxilin*and *GAK*. Zebrafish *auxilin*and *GAK*are interchangeable in their abilities to substitute for *dAux*during *Drosophila*Notch signaling, suggesting that they share some cellular functions. However, they have different expression patterns during development, suggesting that these two paralogs are not completely redundant. Morpholino-mediated knockdown of *GAK*function during embryogenesis caused an increase in the formation of neuronal cells and a decrease in the expression of a Notch target gene, supporting our hypothesis that the role of auxilin family proteins in Notch signaling is conserved. Furthermore, we showed that embryos deficient in GAK function had a higher level of programmed cell death in neural tissues, suggesting that GAK is required for the survival of neuronal cells.\n\nResults\n=======\n\nThe zebrafish genome contains both *GAK*and *auxilin*\n-----------------------------------------------------\n\nTo examine the roles of auxilin-related genes during vertebrate development, we first sought to identify *auxilin*or *GAK*in zebrafish genome. Database search revealed that, like mammals, the zebrafish genome contains both *GAK*(*zGAK*, XP_001919224) and *auxilin*(*zAux*, XP_001336673) orthologs, located on chromosome 23 and 6 respectively. This presence of two distinct auxilin-related orthologs appears to be a feature shared by other vertebrates, as the chicken (*Gallus gallus*, XP_424873 and XP_422527) and pufferfish (*Tetraodon nigroviridis*, CAG08624 and CAG11595) genomes both contain *GAK*and *auxilin*. In contrast, arthropods such as *Drosophila*(*melanogaster*and *virilis*, XP_002058717), honeybee (*Apis mellifera*, XP_396906) and flour beetle (*Tribolium castaneum*, XP_967193) have only GAK. *C. elegans*also has only one *auxilin*-related gene, but it lacks both the kinase and the PTEN homologous region \\[[@B35]\\].\n\nThe *zGAK*locus encodes a polypeptide of 1278 amino acids and, like other GAK orthologs, contains an N-terminal Ark kinase, a PTEN domain, a CBM region and a C-terminal J domain (Figure [1A](#F1){ref-type=\"fig\"}). Alignment of protein sequences shows that zGAK is 65% identical to human GAK overall, with the sequence conservations concentrated in the kinase, PTEN and the J-domain. The kinase, PTEN and J-domain of zGAK are 82.4%, 84.7%, and 78.8% identical to the corresponding domains of human GAK respectively. In contrast, the CBM domains of zGAK and human GAK are more divergent, sharing only 31.7% amino acid sequence identity.\n\n![**The genomic organizations and the protein sequences of zebrafish GAK and auxilin**. (A) The *zGAK*locus contains 28 exons and spans 78 kb of genomic DNA. The corresponding cDNA contains a 3834 bp ORF, encoding a polypeptide of 1278 amino acids with an Ark kinase domain (blue), a PTEN-homologous region (orange), a clathrin-binding motif (CBM), and a J-domain (green). Two morpholinos, GAK-MO1 and GAK-MO2, (red boxes) were designed to disrupt the splicing of *zGAK*mRNA by blocking the splice acceptors of exon3 and exon19, respectively. The amino acids deleted by GAK-MO1 injection are boxed in red. The premature stop codon (after Lys679) generated by GAK-MO2 injection is indicated by a red line. (B) The *zAux*locus is smaller, as it contains 17 exons and spans across 40 kb of genomic DNA. The zAux protein, encoded by a 2619 bp ORF, contains a PTEN-related region (orange), CBM, and J-domain (green). (C) A schematic drawing comparing the domain composition of zebrafish auxilin-related genes. The kinase, PTEN, and the J-domains are represented by blue, orange, and green boxes, respectively.](1471-213X-10-7-1){#F1}\n\nThe zAux protein, on the other hand, is an 873 amino acid long polypeptide, which contains a PTEN region, a CBM and a J-domain, but lacks the N-terminal kinase (Figure [1B](#F1){ref-type=\"fig\"}). The protein sequence of zAux is 59.3% identical overall to the human auxilin. Again, the conservations are higher in the PTEN and the J-domain, as these domains share 67.5% and 94.5% amino acid identity respectively. These domains of zAux are also highly related to the PTEN and the J-domain of zGAK (50.8% and 89.4% protein sequence identity respectively), suggesting that *zAux*and *zGAK*are derived from a common ancestral gene. However, while zAux is similar to zGAK, it is more similar to human and mouse auxilin proteins (Figure [2A](#F2){ref-type=\"fig\"}), suggesting that the divergence of GAK and auxilin occurred prior to the divergence of fish and mammals during vertebrate evolution.\n\n![***zGAK*and *zAux*arose from gene duplication**. (A) A phylogenetic analysis of the amino acid sequences of auxilins and GAKs without their kinase domains. The alignment was performed using Geneious software (Biomatters). (B) Alignment of intron positions in the PTEN region of auxilin and GAK orthologs. The residues shared by all are shaded in yellow and the residues conserved in mammals and vertebrates are shaded in orange. The presence of an intron is indicated by a solid line, and the phase is indicated by a number inside the triangle. \\^ indicates an intron that is conserved in honeybee but not in *Drosophila*, whereas \\* indicates an intron that is conserved in fly but not in honeybee. Am: *Apis mellifera*, Dm: *Drosophila melanogaster*, Gg: *Gallus gallus*, Hs: *Homo sapiens*, Mm: *Mus musculus*, and Dr: *Danio rerio*.](1471-213X-10-7-2){#F2}\n\nIn addition to the similarities in protein sequences, the intron/exon organizations among *auxilin*family genes are highly conserved. First, the human, mouse, and zebrafish *GAK*loci all contain an identical number (27) of introns. The *zAux*locus appears to have lost one intron in the CBM during evolution, as it contains 16 introns while the human and mouse auxilins have 17. Furthermore, in the conserved domains, the positions and the phases (0, 1, or 2 in the codons) of these introns can be precisely aligned. For example, the PTEN homologous regions of zebrafish, mouse, and human *GAK*and *auxilin*loci all contain 8 introns (Figure [2B](#F2){ref-type=\"fig\"}). These introns, although different in nucleotide lengths, are located at the same positions in the coding region with identical phases. This high degree of conservation in intron/exon organization strongly bolsters the notion that *GAK*and *auxilin*were derived from a common ancestral gene through gene duplication.\n\nOverexpression of zGAK or zAux causes clathrin aggregations\n-----------------------------------------------------------\n\nTo ask if zebrafish auxilin family proteins have similar cellular functions as their respective mammalian counterparts, we determined the subcellular localizations of zGAK and zAux. Both *zGAK*and *zAux*were tagged with GFP at the N-termini, placed under the control of a *CMV*promoter in *pCS2*, and transiently expressed in HeLa cells. These GFP-tagged fusion proteins are functional, as ectopic expression of either GFP-zGAK or GFP-zAux in *Drosophila*could restore the neurogenic defect and the lethality caused by *dAux*mutations (see below). To reveal clathrin-positive structures, these cells were also stained with a mouse monoclonal antibody against clathrin heavy chain.\n\nAs shown in Figure [3](#F3){ref-type=\"fig\"}, qualitatively different GFP patterns were seen in *pCS2-GFP-zAux*and *pCS2-GFP-zGAK*transfected cells, depending on the expression levels. In cells expressing low levels of GFP-zGAK or GFP-zAux (marked by asterisks), GFP signals were mostly cytosolic and slightly enriched near the perinuclear regions. These perinuclear zGAK- and zAux-positive structures showed overlaps with clathrin, most likely representing the TGN.\n\n![**The subcellular distributions of zebrafish auxilin family proteins**. Spinning disc confocal micrographs of HeLa cells transfected with (A) pCS2-GFP-zGAK and (B) pCS2-GFP-zAux. The cells were also stained for clathrin heavy chain (red). The low GFP-expressing cells are indicated by asterisks. The large GFP- and clathrin-positive aggregates are indicated by arrows. Scale Bar, 10 \u03bcm.](1471-213X-10-7-3){#F3}\n\nIn cells expressing high levels of GFP-zGAK and GFP-zAux, the cytosolic staining and the enrichments near the perinuclear regions could still be seen. However, these cells also contained large GFP-positive aggregates (indicated by solid arrows and arrowheads) that were intensely clathrin-positive. In untransfected cells, clathrin staining had a vesicular appearance throughout the cytosol, but it appeared depleted from these structures in those aggregate-containing cells. We noticed that zGAK seemed more capable of causing these aggregates than zAux, as intense GFP-positive and clathrin-containing structures were readily seen in cells expressing milder levels of GFP-zGAK. It has been shown that endogenous GAK is predominantly cytosolic and shows elevated associations with TGN \\[[@B24],[@B25],[@B31]\\], and over-expression of mammalian auxilin family proteins in HeLa cells can cause the formation of clathrin-containing \\\"granules\\\" \\[[@B24],[@B38]\\]. Thus, our results showed that over-expressed zGAK and zAux are localized similarly within the cells. In addition, their localizations are similar to those of their respective mammalian homologs.\n\nzGAK and zAux are functionally interchangeable in rescuing *dAux*defects\n------------------------------------------------------------------------\n\nTo determine whether there are intrinsic functional differences between zGAK and zAux, we compared their abilities in rescuing the extra photoreceptor defect and the lethality caused by *dAux*mutations. N-terminally GFP-tagged *zGAK*and *zAux*were placed under *UAS*control, and transgenic flies carrying these constructs were generated. We reasoned that if zGAK and zAux function similarly during clathrin-mediated transport, both should be able to supplant dAux function during neuronal differentiation in *Drosophila*eye discs and for animal survival.\n\nIn wild-type eye discs, regular arrays of photoreceptor clusters could be revealed with \u03b1Elav antibody, which labels the nuclei of neuronal cells \\[[@B39]\\]. In contrast, excessive and disorganized Elav-positive cells were seen in *dAux*^*F956X*^mutant tissues (Figure [4A](#F4){ref-type=\"fig\"}) \\[[@B33]\\]. This defect of extra Elav-positive cells could be rescued by an *Act5C-GAL4*-driven expression of the full-length *dAux*(Figure [4B](#F4){ref-type=\"fig\"}) \\[[@B33]\\]. Similar to *dAux*, both *UAS-GFP-zGAK*and *UAS-GFP-zAux*, when driven by *Act5C-GAL4*, could restore the normal Elav staining pattern in *dAux*^*F956X*^mutant clones (Figure [4C](#F4){ref-type=\"fig\"} &[4D](#F4){ref-type=\"fig\"}). We have previously shown that the lethality of *dAux*mutants could be rescued by a ubiquitous expression of full-length *dAux*\\[[@B33]\\]. Similarly, *Act5C-GAL4*-driven *GFP-zGAK*and *GFP-zAux*could both restore the viability of *dAux*mutant flies. These results suggest that zGAK and zAux can both functionally substitute for dAux.\n\n![**Both zGAK and zAux can substitute for *Drosophila*auxilin**. Fluorescent micrographs of *Drosophila*eye imaginal discs stained with \u03b1Elav antibody, which labels the nuclei of the neuronal cells (blue). The expressions of zebrafish and *Drosophila*auxilin genes are shown in green. Homozygous *dAux*^*F956*^\\* mutant tissues are indicated by the absence of the membrane-associated mRFP (red). Regions containing excessive Elav-positive cells are indicated by arrows. All the flies carry *Act5C-GAL4*, *UAS-FLP*on the second chromosome, and other relevant genotypes include: (A) *FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*, (B) *UAS-dAux-GFP; FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*, (C) *UAS-GFP-zGAK*; *FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*, (D) *UAS-GFP-zAux*; *FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*, (E) *UAS-GFP-zGAK*^\u0394*J*^; *FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*, and (F) *UAS-GFP-zAux*^\u0394*J*^; *FRT*^*5-5Z3515*^, *dAux*^*F956\\**^/*FRT*^*5-5Z3515*^, *GMR-src-mRFP*. Scale Bar, 50 \u03bcm.](1471-213X-10-7-4){#F4}\n\nDeletion of the J-domain from *Drosophila*auxilin is known to render them non-functional \\[[@B32],[@B33]\\]. To ask whether the J-domain is essential for both *zAux*and *zGAK*to rescue *dAux*, *UAS-GFP-zGAK*^DelJ^and *UAS-GFP-zAux*^DelJ^were generated. Unlike their full-length counter parts, these J-deletions failed to rescue the extra Elav-positive cell phenotype (Figure [4E](#F4){ref-type=\"fig\"} &[4F](#F4){ref-type=\"fig\"}) and the lethality, suggesting that, like dAux, the J-domain is critical for the functions of these auxilin family proteins.\n\n*zGAK*and *zAux*are differentially expressed during embryonic development\n-------------------------------------------------------------------------\n\nAlthough it is known that mammalian *GAK*and *auxilin*are expressed in different tissues \\[[@B24]\\], their temporal and spatial expressions during development have not been fully investigated. To better understand the requirement of *zGAK*and *zAux*, we examined their expression at various stages of zebrafish development using whole-mount in situ hybridization. Zygotic expressions of *zGAK*and *zAux*, which were also maternally expressed (data not shown), were detected in diverse tissues, but were most prominently associated with neural tissues. Between 8- to 15-somite stages, *zGAK*was broadly expressed in the dorsal region of the embryo, and the expression of *zAux*appeared to be restricted to the bilateral cell clusters on the dorsal side of the embryo (Figure [5A-F](#F5){ref-type=\"fig\"}). As a control, hybridization using sense probes did not yield detectable signals (data not shown). At the 19-somite stage, *zGAK*still expressed broadly in the entire embryo, whereas *zAux*was predominantly found in hindbrain neurons, spinal chord neurons, and otic vesicles (Figure [5G](#F5){ref-type=\"fig\"} and [5H](#F5){ref-type=\"fig\"}). At 24-hour post fertilization (hpf), *zGAK*was enriched in brains, eyes, otic vesicles, and vasculature, and *zAux*expression was detected in hindbrain neurons, spinal chord neurons, otic vesicles and posterior otic capsules (Figure [5I-N](#F5){ref-type=\"fig\"}). The fact that the expression patterns of *zGAK*and *zAux*did not coincide completely suggests that they have overlapping and distinct functions during development.\n\n![**The expression patterns of *GAK*and *auxilin*during zebrafish embryonic development**. (A, D) Lateral, (B, E) dorsal, and (C, F) anterior views of wild-type embryos at the 15-somite stage. (A-C) *zGAK*is expressed broadly in the hindbrain (B), forebrain and eyes (C). (D-F) *zAux*is expressed mostly in neural tissues as described in the main text. (G, H) Dorsal views of 19-somite stage embryos. (G) *zGAK*is still ubiquitously expressed and (H) *zAux*remains specific to bilateral stripes of neural cells. (I, K, L, N) Lateral and (J, M) dorsal views of 24 hpf embryos. (I-K) *zGAK*is seen in the brain, vasculature and otic vesicles. (L-N) *zAux*remains concentrated in bilateral stripes of neural cells. Panels K and N are close-up views of the posterior regions of the embryos shown in (I) and (L), respectively. In all the images, anterior is to the left, and in all the lateral views, dorsal is up. ey, eye; is, intersomitic vessel; lg, lateral line ganglion; nt, neural tube; ot, otocyte; ov, otic vesicle; sag, statoacoustic ganglion; Scn, Spinal cord neuron; tg, trigeminal ganglion; v, vasculature; vd, ventral diencephalon. Scale Bar, 100 \u03bcm.](1471-213X-10-7-5){#F5}\n\nInhibiting *zGAK*function causes neural-specific cell degeneration\n------------------------------------------------------------------\n\nBecause of the structural conservation of the GAK genes from *Drosophila*to human, we decided to focus on zGAK function during embryonic development. To inhibit *zGAK*, we designed an antisense morpholino oligonucleotide (GAK-MO1) to disrupt the normal splicing of *zGAK*mRNA (Figure [6A](#F6){ref-type=\"fig\"}). A control antisense morpholino oligonucleotide (GAK-MO1C), which is a modified sequence of *GAK-MO1*with five nucleotide changes, was also included. Embryos injected with 8.0ng of GAK-MO1 at one-cell stage (hereafter referred as GAK-MO1 morphants) consistently had smaller eyes, enlarged hindbrain, thinner yolk extension, and weak circulation at 36 to 48 hpf (85%, n = 100, see below). In contrast, injection of the same or higher amount of GAK-MO1C caused no developmental defects in embryos (100%, n = 100). To ensure that the splicing on *zGAK*mRNA was disrupted, RNAs isolated from *GAK-MO1*morphants at 6 hpf were analyzed by RT-PCR. As shown in Figure [6A](#F6){ref-type=\"fig\"}, the injection of GAK-MO1 resulted in a transcript lacking exon 3, generating an in-frame deletion in the kinase domain (Figure [1A](#F1){ref-type=\"fig\"}).\n\n![**GAK-MO1 specifically disrupts zGAK splicing**. (A) RT-PCR analysis of total RNA extracted from uninjected and *GAK-MO1*morphant embryos. From normally spliced *zGAK*mRNA, a reaction with primers complementary to exon 2 and 4 (2/4) would yield a band of 178 bp, and a reaction with primers complementary to exon 18 and 20 (18/20) would yield a band of 275 bp. While the 275 bp band seemed unaffected by GAK-MO1 injection, the level of 178 bp band was reduced and a new band of 118 bp appeared in GAK morphants. Sequence analysis of this 118 bp band showed that exon 2 was spliced into exon 4, resulting in an in-frame deletion in the kinase domain. (B) The target sequences of GAK-MO1 and GAK-MO1C were placed in front of mCherry and cloned into pCS2. In vitro transcribed mRNAs were co-injected with morpholino into one-cell stage embryos. Bright field and fluorescent images of injected embryos at 6 hpf are shown. Scale Bar, 200 \u03bcm.](1471-213X-10-7-6){#F6}\n\nMorphologically, the development of *GAK-MO1*morphants appeared normal until the segmentation stage. At the 10-somite stage, GAK-MO1 morphants displayed cell degeneration in the eyes, which became noticeably opaque (data not shown). At the 14-somite stage, degenerating cells persisted in the eyes, began to appear in the hypothalamic and thalamic regions of the brain and the hindbrain (Figure [7E](#F7){ref-type=\"fig\"} and [7F](#F7){ref-type=\"fig\"}), and continued to increase for the remaining duration of somitogenesis. In contrast, these same regions of the uninjected embryos were transparent (Figure [7A](#F7){ref-type=\"fig\"} and [7B](#F7){ref-type=\"fig\"}). In some cases, the degeneration had spread to the developing neural tube of GAK-MO1 morphants. It is noteworthy that cell degeneration was only observed in neuroectodermal tissues, but not in non-neural tissues such as the notochord and somites. From 24 to 28 hpf, thinner yolk extension and weak blood circulation were observed in *GAK-MO1*morphants, as compared to the uninjected embryos at the same stages (Figure [7C, D, G](#F7){ref-type=\"fig\"}, and [7H](#F7){ref-type=\"fig\"}). At 36 hpf, the hindbrain and the tectal ventricles of *GAK-MO1*morphants were enlarged (Figure [7I-L](#F7){ref-type=\"fig\"}), although these embryos still retained the ability to respond to touch (data not shown).\n\n![**Cell degeneration phenotype of *GAK*morphant embryos during development**. (A, E) Lateral and (B, F) dorsal views of (A, B) uninjected and (E, F) *GAK*morphant embryos at 14-somite stage. Visible degeneration in the eyes (E) and the forebrain regions (F) of *GAK*morphants is indicated by white arrowheads. (C, G) Lateral and (D, H) dorsal views of (C, D) uninjected and (G, H) *GAK*morphant embryos at 24 hpf. At this stage, the increased cell degeneration within the brain (arrowheads) and thinner yolk extension in *GAK*morphants is apparent. (I-L) Lateral views of (I, J) uninjected and (K, L) *GAK*morphant embryos at 36 hpf. In all the lateral views, anterior is to the left and dorsal is up. ey, eye; Hb, hindbrain; nt, neural tube; no, notochord; ov, otic vesicles; ye, yolk extension. Scale Bar, 100 \u03bcm.](1471-213X-10-7-7){#F7}\n\nTo ensure that these phenotypes were caused by the disruption of *zGAK*, we tested the binding specificity of GAK-MO1. The complimentary target sequence of GAK-MO1 was placed in front of the mCherry reporter gene \\[[@B40]\\], and the resulting fusion (*GAK-MO1-mCherry*) was cloned into *pCS2*(Figure [6C](#F6){ref-type=\"fig\"}). Similarly, the complimentary sequence for GAK-MO1C was placed in front of mCherry and cloned into *pCS2*as a control. Using these constructs, *GAK-MO1-mCherry*and *GAK-MO1C-mCherry*RNAs were transcribed in vitro and injected along with either GAK-MO1 or GAK-MO1C morpholinos into one-cell stage embryos. Injected embryos at 6 hpf were then analyzed for the protein expression of mCherry. As shown in Figure [6B](#F6){ref-type=\"fig\"}, GAK-MO1, but not GAK-MO1C, was capable of blocking the translation of *GAK-MO1-mCherry*. Conversely, GAK-MO1 had no effect on the expression of GAK-MO1C-mCherry. These results demonstrate a high degree of binding specificity and efficacy of GAK-MO1 to its intended target.\n\nTo further show that the phenotypes of *GAK-MO1*morphants were not due to off-target effects, a second morpholino antisense oligonucleotide (GAK-MO2) was designed to block splicing at a different region of the GAK gene (Figure [1A](#F1){ref-type=\"fig\"}). RT-PCR analysis showed that injection of GAK-MO2 caused the retention of intron 19, resulting in a premature stop codon in the PTEN domain in the *zGAK*transcripts (Figure [1A](#F1){ref-type=\"fig\"} and Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S1A). Embryos injected with 6.0ng of GAK-MO2 exhibited the same phenotypes as those observed with *GAK-MO1*morphants (90%, n= 85; Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S1B and C). Taken together, these results strongly suggest that the developmental defects of *GAK-MO1*and *GAK-MO2*morphants resulted from the specific inhibition of *zGAK*function.\n\nDisrupting *zGAK*function causes excessive neural-specific programmed cell death\n--------------------------------------------------------------------------------\n\nTo further define the extent and understand the cause of the cell degeneration in *GAK*morphants, we performed in situ TUNEL assay to detect apoptotic cells. At the 14-somite stage, a slight increase in apoptotic cell death was observed in the brain regions of *GAK*morphants, as compared to the control embryos (Figure [8A-D](#F8){ref-type=\"fig\"}). At 24 hpf, *GAK*morphants displayed a significantly higher level of cell death in the brain, as well as in the neural tube (Figure [8E-H](#F8){ref-type=\"fig\"}). Similar to the aforementioned cell degeneration, excessive apoptosis in *GAK*morphants was seen only in neural tissues, but not in non-neural tissues. Thus, there appears to be a good correlation in the onset and the extent of these two phenotypes, suggesting that this increase in the programmed cell death is the basis for the cell degeneration in *GAK*morphants.\n\n![**Programmed cell death in wild-type and *GAK*morphant embryos during development**. (A, C) Lateral and (B, D) dorsal views of TUNEL-stained (A, B) wild-type and (C, D) *GAK*morphant embryos at the 14-somite stage. At this stage, the control embryos display no detectable apoptosis, whereas *GAK*morphants have a low level of programmed cell death (indicated by black arrowheads). (E, G) Lateral and (F, H) dorsal views of TUNEL-stained (E, F) wild-type and (G, H) *GAK*morphants at 24 hpf. While the control animal exhibits a low level of apoptosis in anterior brain and the posterior body regions (arrowheads), *GAK*morphants exhibit a high level of apoptotic cell death in the brain, as well as in the neural tube (arrowheads). However, as compared to the control wild-type embryo, no obvious increase in apoptosis was observed in the posterior region of *GAK*morphants (indicated by arrows). In all the lateral views, anterior is to the left and dorsal is up. ey, eye; Hb, hindbrain; MHB, mid-hindbrain boundary; nt, neural tube; s, somites; tb, tailbud. Scale Bar, 100 m.](1471-213X-10-7-8){#F8}\n\nReduction of *zGAK*disrupts the development of specific brain regions\n---------------------------------------------------------------------\n\nTo explore the role of *zGAK*in brain patterning, *GAK*morphants were stained for *krox20, fibroblast growth factor 8 (fgf8)*and *sonic hedgehog (shh)*, whose expressions delineate various regions of the brain. *Krox20*, a zinc finger-containing transcription factor, is expressed specifically in the hindbrain \\[[@B41]\\]. At the 12-somite stage, the pattern and the level of *krox20*expression appeared normal in *GAK*morphants (Figure [9A](#F9){ref-type=\"fig\"} and [9B](#F9){ref-type=\"fig\"}). At 24 hpf, the pattern of *krox20*expression in *GAK*morphants remained similar to that seen in wild type, although the two stripes of *krox20*in the rhombomeres 3 and 5 appeared much closer (Figure [9C](#F9){ref-type=\"fig\"} and [9D](#F9){ref-type=\"fig\"}). This suggests that, while the overall patterning of the brain regions is not significantly affected, the rostral hindbrain is reduced in size in the absence of *zGAK*function.\n\n![**Expression patterns of regional brain markers in wild-type and *GAK*morphant embryos**. (A-D) *krox20*expression in (A, C) uninjected and (B, D) *GAK*morphant embryos at (A, B) 12- to 14-somite stage and (C, D) 24 hpf. At 12- to 14-somite stage, *krox20*expression in the rhombomeres 3 (R3) and 5 (R5) of the control and injected embryos appears comparable. However, at 24 hpf, the spacing between R3 and R5 is dramatically reduced in *GAK*morphants (white arrowhead). (E, F) Lateral and (G, H) dorsal views of *fgf8*expression in (E, G) wild-type and (F, H) *GAK*morphant embryos at 18- to 20-somite stage. The domains of *fgf8*expression in the forebrain and mid-hindbrain boundary (MHB) are up-regulated in *GAK*morphants (black arrows). (I, J) *shh*expression in (I) wild-type and (J) *GAK*morphants at the 18-somite stage. In all the lateral views, anterior is to the left and dorsal is up. d, ventral diencephalon; f, floor plate; os, optic stalk; s, somite; t, telencephalon; dd, dorsal diencephalon; tb, tailbud. Scale Bar, 100 \u03bcm.](1471-213X-10-7-9){#F9}\n\n*Fgf8*, expressed in the forebrain, the MHB (midbrain-hindbrain boundary) and other mesodermal derivatives that include somites and tailbud, has been shown to regulate forebrain development and establish the segmental identity of the hindbrain \\[[@B42]\\]. At the 12-somite stage, the expression of *fgf8*in GAK morphants appeared to be normal (data not shown). In contrast, GAK morphants at the 18-somite stage exhibited higher levels of fgf8 expression in the MHB, optic stalks and forebrain, suggestive of a requirement for GAK in repressing *fgf8*expression in these regions (Figure [9E-H](#F9){ref-type=\"fig\"}). No alteration in *fgf8*expression in the somites and tailbud regions was observed (Figure [9E](#F9){ref-type=\"fig\"} and [9F](#F9){ref-type=\"fig\"}), consistent with the notion that *zGAK*function is required predominantly in neural tissues, but not in non-neural tissues. At the 18-somite stage, *shh*mRNA was detected in the ventral midbrain, hypothalamus, telencephalon and notochord \\[[@B43]\\]. In contrast to *fgf8*, shh expression appeared normal in the *GAK*morphants at this stage (Figure [9I](#F9){ref-type=\"fig\"} and [9J](#F9){ref-type=\"fig\"}).\n\nReduction of *zGAK*impairs Notch signaling\n------------------------------------------\n\nTo ask if zGAK, like dAux, has a role in Notch signaling, we examined the expression of *HuC*and *Her4*mRNAs in *GAK*morphants. *HuC*, the zebrafish homolog of *Elav*, is one of the earliest markers for neuronal cells \\[[@B44]\\]. In 8-somite stage wild-type embryos, *HuC*transcripts are detected in bilateral stripes of neuronal cells along the anterior-posterior axis \\[[@B45]\\]. In *GAK*morphants at similar stage, although the pattern of expression was similar, more *HuC*-positive cells were seen (Figure [10A](#F10){ref-type=\"fig\"} and [10B](#F10){ref-type=\"fig\"}), suggesting that additional cells had adopted neuronal fate. This increase in *HuC*staining persisted in *GAK*morphant embryos at 10-somite stage (Additional file [2](#S2){ref-type=\"supplementary-material\"}: Figure S2A-D). Interestingly, by 24 hpf, *GAK*morphants appeared to have fewer *HuC*-positive cells in the forebrain and midbrain regions (Figure [10C-F](#F10){ref-type=\"fig\"}). However, it is important to note that the reduction in *HuC*-positive cells at this stage coincides with the onset of the elevated cell death in neural tissues (see above).\n\n![**Expression patterns of *HuC*and *Her4*in wild-type and *GAK*morphant embryos**. (A, B) Close-up dorsal views of *HuC*expression in (A) wild-type and (B) *GAK*morphant embryos at the 8-somite stage. Micrographs of a lower magnification are shown in the insets. At this stage, more cells appeared to express *HuC*in *GAK*morphant embryos, suggesting the presence of more neural progenitor cells. (C-F) Lateral views of *HuC*expression in the brain regions of (C, E) wild-type and (D, F) *GAK*morphant embryos at 24 to 28 hpf. A comparison of (C) and (D) shows a reduction in *HuC*-positive cells in the forebrain and midbrain in *GAK*morphants. Similarly, a comparison of (E) and (F) shows the disorganization and reduction of *HuC*-positive cells in the hindbrain in *GAK*morphants. (G-J) Lateral views of *Her4*expression patterns in (G, I) wild-type and (H, J) *GAK*morphants at the (G, H) 8-somite stage or at (I, J) 24 to 28 hpf. At the 8-somite stage, *Her4*expression in *GAK*morphant embryos was significantly reduced, as compared to the wild type. In contrast, the expression of *Her4*in the brain of wild-type and *GAK*morphant embryos at 24 to 28 hpf was comparable. In all the panels, anterior is to the left, and in all the lateral views, dorsal is up. Fb, forebrain; Hb, hindbrain; MHB, mid-hindbrain boundary; Mb, midbrain; mbp, midbrain basal plate; R, rhombomeres; tb, tailbud; te, telencephalon; vd, ventral diencephalon. Scale Bar, 100 \u03bcm.](1471-213X-10-7-10){#F10}\n\n*Her4*, a homolog of Drosophila *E(spl) (Enhancer of split)*gene, is a known target of the Notch pathway and participates in specification of neuronal cells \\[[@B46]\\]. In 8-somite stage wild-type embryos, *Her4*is expressed in bilateral stripes of cells \\[[@B46]\\]. Compared to wild type, the level of *Her4*expression in *GAK*morphants appeared lower (Figure [10G](#F10){ref-type=\"fig\"} and [10H](#F10){ref-type=\"fig\"}), suggesting that the output of the Notch pathway was reduced at this stage. Similar to *HuC*, this decrease in *Her4*expression was still observed in 10-somite stage *GAK*morphant embryos (Additional file [2](#S2){ref-type=\"supplementary-material\"}: Figure S2E-H). At 24 hpf, *Her4*expression in *GAK*morphants and control embryos was not significantly different (Figure [10I](#F10){ref-type=\"fig\"} and [10J](#F10){ref-type=\"fig\"}), suggesting that either GAK is not required for Notch signaling, or GAK and auxilin are redundant for Notch signaling at this stage. Nevertheless, the alterations in the expression of *HuC*and *Her4*at the 8- and 10-somite stages showed that a reduction of GAK function could impair Notch activity in zebrafish.\n\nDiscussion\n==========\n\nZebrafish, like mammals but not invertebrates, has two distinct auxilin-related genes. Like their mammalian counterparts, zebrafish GAK and auxilin differ in the presence of the N-terminal kinase and their respective expression patterns. zGAK has a kinase domain and is ubiquitously expressed during embryonic development. In contrast, the expression of *zAux*, the ortholog without the kinase domain, appears predominantly in the neuronal cells. There are other differences between the two at the protein sequence level. For instance, zGAK, but not zAux, contains a sequence of FGDL at the amino acid position 950, which matches perfectly to the consensus \u03c8G\\[PDE\\]\\[\u03c8LM\\] (\u03c8 is an aromatic residue). This motif, conserved in mammalian GAKs, has been shown to mediate interactions with AP1 adaptor in Golgi-lysosomal trafficking \\[[@B31]\\]. Taken together, these structural differences between GAK and auxilin suggest that their molecular functions may have diverged during evolution. It is notable that inhibiting *zGAK*function causes an increase in apoptotic cell death in the neuroectodermal tissues, where GAK and auxilin are both expressed. This inability of *zAux*to compensate for the *zGAK*knockdown would argue that GAK has a function unique from auxilin. However, we cannot formally exclude the possibility that these neural cells may have a higher demand for the functions of *auxilin*-related genes. In this scenario, the functions of both genes would be needed to prevent the onset of apoptosis; therefore, inhibiting GAK alone would be sufficient to induce extra neural cell deaths.\n\nDespite the structural differences between GAK and auxilin, it seems clear that these two paralogs have overlapping molecular functions. Both GAK and auxilin are required for receptor-mediated endocytosis in HeLa cells, indicating that they act in the same process \\[[@B26]\\]. Consistent with this, we showed that the subcellular localizations of zGAK or zAux in HeLa cells were similar, and over-expression of either in HeLa cells could form clathrin-containing aggregates. More importantly, we showed that over-expression of either zGAK or zAux in *Drosophila*could completely restore the neurogenic defects caused by *dAux*, and this ability to rescue *dAux*absolutely requires their respective J-domains. Together, these results suggest that zGAK and zAux are at least partially redundant.\n\nUnlike zGAK or dAux, zAux does not have the N-terminal Ark kinase. Nevertheless, over-expression of zAux in *Drosophila*could completely restore the defects caused by a strong *dAux*allele. This is not entirely surprising as we have previous shown that over-expression of *dAux*^\u0394K^, a *dAux*with its kinase domain deleted, could rescue the dAux phenotype \\[[@B33]\\]. In fact, over-expression of a fragment consisting of the CBM and J domains alone appears sufficient to restore the function of dAux in Notch \\[[@B32],[@B33]\\]. However, the kinase domain does have a role in GAK\\'s function in Notch as kinase domain-specific disruptions, by either point mutations \\[[@B33]\\] or morpholino-induced mis-splicing (MO1, this study), produce *Notch*-like phenotypes. Still, it does appear that, when expressed at a high level, the kinase activity is not required for the functions of auxilin family proteins.\n\nIn *Drosophila*, auxilin has been shown to participate in Notch signaling by facilitating ligand internalization \\[[@B21],[@B22]\\]. Given that mammals and vertebrates have two *auxilin*-related genes, it is not known if either, neither, or both function in the Notch pathway. The similarities in domain structures and expression patterns suggest that GAK is more likely to have a role in Notch. However, while inactivation of *GAK*affects the formation of multiple tissues in mouse \\[[@B37]\\], it was unclear whether these defects were caused by disrupted Notch signaling. Here we showed that, in *GAK*morphants, the number of *HuC*-positive cells appeared increased, a defect analogous to the neurogenic phenotype. Moreover, we showed that, in *GAK*morphants, the expression of *Her4*, a known Notch target gene, was reduced. Thus, while it is not yet known whether zGAK participates in ligand internalization, these defects in *HuC*and *Her4*expressions suggest that GAK function is also required for Notch signaling in zebrafish. These results provide the first evidence that the requirement for a GAK-dependent endocytic pathway during Notch signaling is evolutionarily conserved. A requirement of the *mindbomb*E3 ligase and epsin in Notch has also been demonstrated in flies \\[[@B7],[@B11],[@B12],[@B15],[@B16]\\], fish \\[[@B9]\\], and mouse \\[[@B8],[@B47],[@B48]\\], which, along with our analysis of zebrafish GAK, suggests that Notch ligand internalization may rely on the same set of endocytic genes.\n\nIt has been demonstrated that a conditional removal of GAK function during mouse brain development causes a significant loss of neural tissues \\[[@B37]\\], although the mechanism is not known. Likewise, our depletion of GAK function during zebrafish embryonic development results in neural-specific cell degeneration. Using TUNEL staining, we showed that this cell degeneration is caused by increased programmed cell death, suggesting that GAK has a role in preventing the apoptosis of neural cells. Thus, although more *HuC*-positive cells were present in *zGAK*-deficient embryos at the 8- and 10-somite stages, fewer *HuC*-positive cells might be expected at later stages because of cell death. Indeed, this was precisely what we observed, as fewer *HuC*-positive cells were seen in *GAK*morphants at 24 hpf. Interestingly, in *mindbomb*mutant embryos, where Notch ligand endocytosis is impaired, no cell degeneration phenotype was observed (J.S., unpublished data). This, along with our observation that *Her4*expression was not significantly reduced at 24 hpf, suggests that the role of GAK in maintaining neural cell survival may be Notch-independent. Taken together, our results suggest that *GAK*has at least two distinct functions during the development of neural tissues: an early role in the patterning of neuronal cells and a later role in maintaining the survival of neuronal cells. Furthermore, human GAK has recently been implicated as a susceptibility gene in familial Parkinson disease \\[[@B49]\\], and the neurodegenerative phenotype observed in GAK morphants certainly supports this conclusion.\n\nIt is noteworthy that the phenotypes of the *GAK*morphants bear a strong resemblance to those of the *\\\"spacehead\\\"*class zebrafish mutants \\[[@B50]\\]. These mutants, isolated from a large-scale screen, are characterized by defects including cell degeneration in the eye and the brain regions, thinner yolk tube, and weak blood circulation \\[[@B50]\\]. As the genes responsible for most of these mutants have not been determined, the phenotypic similarities suggest that *zGAK*may correspond to one of them. If this is indeed the case, it will provide important clues to understand the functions of these genes in maintaining neuronal cell survival.\n\nConclusion\n==========\n\nZebrafish, like mammals but not invertebrates, has two distinct auxilin-related genes, auxilin and GAK. These two genes share some molecular functions, but are not completely redundant, as they are differentially expressed during development. Inhibition of GAK function appears to impair Notch signaling during embryonic neural patterning. This, along with the fact that auxilin has been implicated in *Drosophila*Notch signaling, suggests that the Notch pathway is regulated by a similar set of endocytic factors. In addition, we showed that inhibition of GAK function increases apoptosis in neural tissues, suggesting that GAK has a role in promoting or maintaining the survival of neural cells. As GAK is recently implicated in familial Parkinson disease \\[[@B49]\\], our results should provide a useful model for further understanding the cause of this neurodegenerative disease.\n\nMethods\n=======\n\nEmbryos and morpholino oligoneucleotides injections\n---------------------------------------------------\n\nAll animal procedures were reviewed and approved by the Purdue Animal Care and Use Committee (PUCAC \\#06-111-09). Adult fish and embryos were raised and maintained at 28.5\u00b0C in system water. Embryos were obtained by natural spawning of adult AB strain zebrafish. *zGAK*-specific antisense morpholino oligonucleotides (GAK-MO1, GAK-MO2, GAK-MO1C, and GAK-MO2C) were purchased from Gene Tools (Philomath, Oregon). At the one-cell stage, each embryo was injected with approximately 1 nl volume of morpholino using a Picospritzer III (Parker Hannifin). Embryos were collected at the appropriate stages \\[[@B51]\\] and fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight at 4\u00b0C. Fixed embryos were dechorionated, washed 3 times with PBS, and stored in methanol at -20\u00b0C.\n\n*Drosophila*genetics\n--------------------\n\nFlies were raised at 25\u00b0C on standard food. Mutant clones of *dAux*^*F956X*^, a strong allele with a nonsense mutation deleting the J-domain, were generated as previously described \\[[@B33]\\]. For the lethality assay, a trans-heterozygous combination of *dAux*^*F956X*^/*dAux*^*L78H*^was used to avoid potential unrelated lethal mutations in the background. The *dAux*^*L78H*^allele contains a missense mutation in the kinase domain \\[[@B33]\\]. Immunostaining of eye imaginal discs was performed as previously described \\[[@B52]\\]. Mouse \u03b1Elav 9F8A9 (DSHB, Iowa) was used at 1:100.\n\nMolecular cloning\n-----------------\n\nA 4475 bps-long cDNA containing the entire zGAK ORF in pCCM114 was obtained from OpenBiosystems (ID 2504096). This particular clone has several mutations, including missense mutations at Arg^303^, Tyr^480^, Asp^614^, and a frameshift mutation at Gly^1047^. All were repaired using the QuickChange site-directed mutagenesis kit (Stratagene) and the resulting cDNA was verified by sequencing. For zAux, an Exelixis EST (ID 3410313, OpenBiosystems) containing a partial zAux ORF (missing the N-terminal 696 bps) in pSPORT1 was obtained. A full-length cDNA clone was constructed using RT-PCR products from zebrafish embryonic mRNA and standard cloning techniques, and verified by sequencing.\n\nTo generate pCS2-GFP-zGAK and pCS2-GFP-zAux, the entire zGAK and zAux ORFs were fused in frame to the C-terminus of EGFP, and the resulting fusions were cloned into pCS2 as EcoRI-XhoI fragments. To make zGAK^\u0394J^and zAux^\u0394J^, codons for His^1206^of *zGAK*and His^801^of *zAux*were changed into stops by site-directed mutagenesis. To express zGAK and zAux in *Drosophila*, GFP-tagged full-length or J-domain-deleted cDNAs were cloned into pUAST \\[[@B53]\\].\n\nFor RT-PCR analysis, total RNAs were extracted from embryos using RNeasy Mini Kit (Qiagen) and the RT-PCR reactions were performed using OneStep RT-PCR Kit (Qiagen). Primers GTATGAGGCCCAGGATTTAGGAAG and GTCAGACTCTTCTTTACTGATGGAC were used to examine the splicing at exon 3. Primers GTGCCCAGAAATGCCTCCACTGTC and GCATAACAGGCTGTCGAACCAGGC were used to examine the splicing at exon 19.\n\nHeLa cell manipulation and microscopy\n-------------------------------------\n\nHeLa cells were maintained in DMEM, supplemented with 10% Fetal Bovine Serum under standard conditions. To express GFP-zGAK or GFP-zAux, 70% confluent HeLa cells in each 10cm dish were transfected with a 0.5ml cocktail, which contained 10 \u03bcg plasmid DNA, 25 \u03bcl of Fugene HD (Roche) and serum-free DMEM. Cells were harvested and processed for immunostaining 48 hours after transfection. Mouse \u03b1Chc (Affinity BioReagents) was used at 1:500, and fluorescently-conjugated secondary antibodies (Molecular Probes) were used at 1:100. All images were collected using Olympus BX61 equipped with a Spinning Disc Confocal unit and processed with Photoshop (Adobe).\n\nIn situ hybridization and TUNEL staining\n----------------------------------------\n\nWhole-mount in situ hybridization was performed using digoxigenin-labeled antisense RNA probes and visualized using anti-digoxigenin Fab fragment conjugated with alkaline phosphatase (Roche) as previously described \\[[@B54]\\]. Riboprobes were made from DNA templates, which were linearized and transcribed with either SP6 or T7 RNA polymerases. Embryos were processed and hybridized as previously described \\[[@B54]\\].\n\nWhole-mount in situ TUNEL (terminal deoxynucleotide transferase-mediated dUTP nick-end labeling) was performed using the AP (alkaline phosphatase) In Situ Cell Death Detection Kit (Roche) as previously described \\[[@B55]\\].\n\nAuthors\\' contributions\n=======================\n\nT.B. isolated and analyzed zebrafish *auxilin*and *GAK*cDNA, determined the subcellular localization, and performed the Drosophila rescue experiments. J.L.S. performed the phenotypic analysis of the zebrafish *GAK*morphants. K.K. assisted in the RNA in situ hybridization to determine the expression of zebrafish *auxilin*and *GAK*. H.D. assisted in the molecular cloning of zebrafish *auxilin*and *GAK*. H.C.C. and D.P.S. were responsible for most of the experimental design and the manuscript preparation. All the authors have read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\n**The phenotypes of *GAK-MO2*morphants are similar to those of *GAK-MO1*morphants**. (A) RT-PCR analysis of the total RNA from embryos injected with GAK-MO2. Using the RNA prepared from uninjected embryos, a 2/4 reaction (using primers complementary to exon 2 and 4) yielded a band of 178 bp, and an 18/20 reaction yielded a band of 275 bp and a slightly smaller non-specific band. On the other hand, using the RNA prepared from *GAK-MO2*morphants, the 178 bp band was unaffected. However, the 275 bp band was absent (the non-specific band was unaffected) and a new band of 359 bp appeared. Sequence analysis of this 359 bp band revealed that the injection of GAK-MO2 caused the retention of intron 18, resulting in a frame-shift truncation in the PTEN region (immediately after Lys^679^). (B, C) Lateral views of uninjected and *GAK-MO2*morphants at 36 hpf. In (B), anterior is to the left and dorsal is up, and in (C), anterior is up and dorsal is to the right. ey, eye; Hb, hindbrain; ov, otic vesicles; ye, yolk extension. Scale Bar, 100 \u03bcm.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**Expression patterns of *HuC*and *Her4*in wild-type and *GAK*morphant embryos at the 10-somite stage**. (A, B) Lateral views of *HuC*expression in (A) wild-type and (B) *GAK*morphant embryos at the 10-somite stage. Similar to the 8-somite stage, more cells appeared to express *HuC*in *GAK*morphant embryos at the 10-somite stage, suggesting the presence of more neuronal cells. (C-D) Close-up top views of *HuC*expression in the brain regions (indicated by brackets in A&B) of (C) wild-type and (D) *GAK*morphant embryos. (E, F) Lateral views of *Her4*expression patterns in (E) wild-type and (F) *GAK*morphants at the 10-somite stage. At this stage, *Her4*expression in *GAK*morphant embryos appeared reduced, as compared to the wild type. (G-H) Close-up top views of Her4 expression in the brain regions (indicated by brackets in E&F) of (G) wild-type and (H) *GAK*morphant embryos. In all the panels, anterior is to the left, and in all the lateral views, dorsal is up. Fb, forebrain; Hb, hindbrain; Mb, midbrain; mes, mesencephalon; te, telencephalon; Tg, Trigeminal ganglion; tb, tailbud. Scale Bar, 100 \u03bcm.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nWe thank Dr. Ajay Chitnis for providing reagents. We are grateful to Dr. Donna Fekete for critical reading of this manuscript. This work was supported by American Heart Association Scientist Development Grant and American Cancer Society Research Scholar Grant to H.C. T.B. is partially funded by a Purdue Research Foundation fellowship.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nRespiration is an essential physiologic component for maintaining vital functions and performing physical activities in the human body[@r1]^)^. Enhanced respiration function enables a relatively high efficiency of physiologic competence by increasing the capacity for physical activity[@r2]^)^. On the other hand, pathologic conditions in the respiratory system cause shortness of breath, poor airway clearance, lung distensibility, and so forth[@r3],[@r4],[@r5]^)^. These problems deteriorate the quality of life and are life threatening[@r6]^)^. In particular, pulmonary disease in children can have a harmful influence on physical development related to sensorimotor function and can directly impair cardiopulmonary systems[@r7], [@r8]^)^.\n\nCerebral palsy (CP) is one of the major neurological diseases that cause physical dysfunction such as sensorimotor and respiratory dysfunction[@r9], [@r10]^)^. Motor dysfunction induces paralysis or weakens respiratory muscles, and this gives rise to an abnormal distorted chest, restriction of physical activity, and delayed development of the cardiopulmonary system[@r11]^)^. For recovery of these respiratory symptoms, respiratory training programs have increasingly been view as clinically necessary. Recent studies have revealed that physiological function related to respiration was preserved via walking ability by enabling subjects to perform a variety of vigorous physical activities[@r12], [@r13]^)^. These results suggest that walking ability may be associated with preservation or development of respiratory functions. However, to our knowledge, little previous literature concerning comparison of the effectiveness of respiratory training according to walking ability has been published.\n\nTherefore, in the current study, we attempted to compare differences in respiratory pressure and pulmonary function and the effectiveness of respiratory training depending on walking ability in children with CP.\n\nSUBJECTS AND METHODS {#s2}\n====================\n\nTwenty-five children who suffered from spastic hemiplegic and diplegic cerebral palsy were recruited in this study, according to following inclusive criteria: (1) children with spastic hemiplegic or diplegic cerebral palsy diagnosed by a pediatric neurologist from their brain MR image, (2) communicative disability in language or intellectual problem capable of measuring respiratory function, (3) no psychiatric or neurological symptoms except cerebral palsy, and (4) belonged to levels I, II, and III of the Gross Motor Function Classification System (GMFCS). According to independent walking ability considering the GMFCS level and clinical assessment by a physical therapist with over five years of experience, they were divided into an independent walking group and non-independent walking group. The independent walking group contained subjects classified into GMFCS levels I and II, and the non-independent walking group contained subjects classified into GMFCS level III. Twelve and 13 children with CP belonged to the independent walking group (6 boys, mean age 9.5\u00b12.1) and non-independent walking group (7 boys, mean age 10.5\u00b11.5), respectively. All parents of children provided written informed consent before participation in this study in accordance with the Helsinki declaration, and this study was approved by the local ethics committee.\n\nAll children took part in respiratory training using feedback for 20 minutes per day, 3 times a week for 4 weeks, and received conventional physical therapy for 30 minutes per visit, 2 to 3 times a week, which focused on gross motor tasks and functional activities including independent sitting, walking, stair climbing, and so forth. A SpiroTiger (Idiag AG, Volketswil, Switzerland) was used as the device for biofeedback respiratory training; this device is designed to improve respiratory function by providing feedback on the subject's performance. Children sat on a chair comfortably with their head and trunk straight, took the mouthpiece into their mouth, and placed a nose clip on their nose, ensuring that breathing occurred exclusively through the device. They were instructed to watch the monitor and to inhale when the red bar reached \"in\" on the display panel of the device and they heard a beep sound, whereas they were instructed to exhale when the red bar reached \"out\" on the device and they heard a beep sound. Visual and auditory feedback provided through the device play an important role in restricting the subject's breathing to the threshold value of isocapnia. When symptoms of fatigue or hyperventilation were observed in a subject during respiratory training, training was stopped, and the subject was allowed to rest.\n\nTo evaluate the effect on respiratory function after respiratory training for 4 weeks, respiratory pressure was measured and a pulmonary function test (PFT) was performed. All children were positioned in a chair sitting with their head and trunk straight up and the hip and knee joints flexed at 90\u00b0. Measurements of respiratory pressure and pulmonary function were performed by the same examiner throughout the entire experiment. The respiratory pressure was measured using a MicroRPM Pressure Meter (Micro Direct Inc., Lewiston, ME, USA), which assessed the highest pressure that respiratory muscles are able to generate against an occlusion at the mouth. Children were asked to inhale or exhale against the obstructed mouth piece, with maximal voluntary effort, while keeping the lips sealed tightly around the mouthpiece. In the best of these trials, the values of maximal inspiratory pressure (MIP) and maximal expiratory pressure (MEP) were acquired. The PFT was performed using a spirometer (Vmax 229, SensorMedics, USA). All children were instructed to breathe in and then to breathe out via a mouth piece as deeply and fast as possible. The test was conducted a total of three times with an appropriate rest period for prevention of hyperventilation. In the best of these trials, the values of forced vital capacity (FVC), forced expiratory volume at one second (FEV~1~), and peak expiratory flow (PEF) were acquired.\n\nAll data were analyzed using statistical software, PAWS 18.0 (SPSS, Chicago, IL, USA). The c^2^ test was conducted to examine differences in gender distribution between the two groups. The Independent t-test was performed to compare demographic data (i.e., age, weight, height, chest circumference, heart rate, and respiratory rate) and dependent variables (i.e., MIP, MEP, FVC, FEV~1~, and PEF) between the two groups. The alpha level for statistical significance was set at 0.05.\n\nRESULTS {#s3}\n=======\n\nDemographic information for the independent walking and non-independent walking groups are shown in [Table 1](#tbl_001){ref-type=\"table\"}Table 1.Demographic information of the two groupsIndependent walking groupNon-independent walking groupAge (years)9.3\u00b12.010.8\u00b11.4Gender (M/F)12 (7/5)11 (4/7)Height (cm)134.9\u00b112.0132.8\u00b18.0Weight (kg)35.5\u00b112.429.0\u00b14.8Chest circumference (cm)74.6\u00b110.869.5\u00b14.5Heart rate (beats/min)82.7\u00b19.779.5\u00b17.0Respiratory rate (breaths/min)25.0\u00b12.225.6\u00b13.4Values are expressed as frequencies or means \u00b1 SD. No statistical differences between the two groups were found in terms of age, gender, height, weight, chest circumference, heart rate, and respiratory rate (p\\>0.05).\n\nOf the twenty-five subjects included in this study, two participants dropped out before the posttest in the non-independent walking group. One subject gave up the training due to extremely poor conditions, and another participant failed to participate 2 or 3 times. Therefore, twelve children in the independent walking group and eleven children in the non-independent walking group finished this experiment.\n\nIn comparison of respiratory function at the pretest including respiratory pressure (i.e., MIP and MEP) and the pulmonary function test (FVC, FEV~1,~ and PEF) between the two groups, the independent t-test showed significant differences in MIP, MEP, FVC, and FEV~1~, indicating that the independent walking group had higher respiratory function than the other group in all variables except PEF. In comparison of the changes in respiratory function between the two groups, the independent walking group showed significantly higher improvement of respiratory function in term of MIP, MEP, and FVC (p\\<0.05) ([Table 2](#tbl_002){ref-type=\"table\"}Table 2.Changes in respiratory pressure and pulmonary function between the independent walking group and non-independent walking groupIndependent walking group (n=12)Non-independent walking group (n=11)PretestPosttestChangesPretestPosttestChangesRespiratory pressureMIP (cmH~2~O)36.9\u00b112.342.8\u00b112.95.9\u00b11.921.4\u00b17.4\\*30.6\u00b17.5\\*9.2\u00b12.8\\*MEP (cmH~2~O)48.9\u00b118.555.5\u00b117.86.6\u00b13.129.6\u00b110.4\\*40.2\u00b111.5\\*10.6\u00b13.3\\*Pulmonary functionFVC (\u2113)1.6\u00b10.51.8\u00b10.60.2\u00b10.20.9\u00b10.4\\*1.3\u00b10.3\\*0.4\u00b10.2\\*FEV~1~ (\u2113)1.4\u00b10.51.6\u00b1.060.2\u00b10.30.8\u00b10.3\\*1.2\u00b10.30.4\u00b10.3PEF ( \u2113/sec)2.7\u00b11.42.8\u00b11.50.2\u00b10.81.8\u00b10.72.3\u00b10.60.5\u00b11.0\\*Values represent the results of independent t-tests and are significant at the p\\<0.05 level in comparisons between the CP children in the independent walking group and non-independent walking group.).\n\nDISCUSSION {#s4}\n==========\n\nThe first findings of the current study showed that children with CP who had independent walking ability possessed stronger showed better improvement of respiratory muscles and pulmonary function compared with children without independent walking ability. Respiratory pressure and the PFT have been used to measure the strength of respiration muscles and lung capacity, which is one of routine procedures for measurement of respiratory function, along with PFT. These findings are supported by several previous studies, indicating that children who were capable of more vigorous physical activity showed better respiratory function compared with children without independent walking ability[@r14],[@r15],[@r16]^)^. The low cardiopulmonary capacity in CP is due to restrictive lung dysfunction resulting from limited movement and not because of parenchymal lung dysfunction[@r3], [@r17], [@r18]^)^. Therefore, we deduced that children who were capable of more vigorous physical activity showed better respiratory function.\n\nIn the second findings, the two groups showed improvement of all variables after 4 weeks of respiratory training. However, in comparison of the changes in respiratory function between the two groups, the non-independent walking group showed significantly better improvement of MIP, MEP, and FVC compared with the independent walking group. We think that this might be attributable to the ceiling effect in the independent walking group. In other words, the independent walking group may have already reached the upper level of their capacity in respiratory function, whereas the non-independent walking group may have had more potential for its improvement.\n\nIt is a generalized fact that walking disability can accompany respiratory dysfunction in children with CP, which together result in disturbance of normal motor development and restriction of functional activity in daily life[@r14], [@r19], [@r20]^)^. Elucidation of differences in respiration function and effectiveness of respiratory training according to walking ability is an important clinical issue. Therefore, we expect that our results will be valuable for physical therapists in evaluating respiratory function and applying training programs in children with CP. A limitation of this study is the small sample size due to difficulty in recruiting children with CP. Further study will be required to consider this limitation and other clinical issues, such as differences in respiratory function according to level of walking ability and metabolic equivalence.\n\nThis research was supported by a Daegu University Research Grant (2012).\n"}
+{"text": "Introduction\n============\n\nGene silencing through RNA interference (RNAi) holds great potential for the treatment of various diseases, such as cancer, neurodegenerative disorders and viral infections. Since its basic discovery, RNAi research has advanced rapidly and has already made its way to the clinic, with phase I and II clinical trials under way.^[@bib1],\\ [@bib2]^ RNAi is an evolutionarily conserved sequence-specific post-transcriptional gene silencing mechanism in eukaryotes that results in the degradation of mRNA and subsequently in decrease of protein synthesis.^[@bib3],\\ [@bib4]^ Small double-stranded RNA molecules with perfect sequence complementarity to the target mRNA are processed by Dicer, rendering a mature small interfering RNA (siRNA) to interact with the cytosolic multiprotein RNA-induced silencing complex (RISC). The RISC-incorporated siRNA binds to a complementary mRNA sequence, which leads to cleavage of the target mRNA, followed by degradation through cellular RNases.^[@bib5],\\ [@bib6]^\n\nRNAi delivery methods employ both synthetic small interfering RNAs (siRNAs) that can be added to the cell, and DNA-based vectors that encode short hairpin RNAs (shRNAs), which upon expression are processed intracellularly to siRNAs.^[@bib6]^ Expression of these shRNAs is typically regulated by constitutive strong polymerase III (pol III) promoters (H1 or U6), but their expression can also be made tissue specific or inducible, depending on the choice of promoter.^[@bib7],\\ [@bib8]^ Specificity can be increased by encapsidation of the shRNA into a viral vector with specific tissue tropism. In addition, viral vector-mediated delivery of shRNA-encoding constructs has the advantage of long-term RNAi upon single administration.^[@bib6],\\ [@bib9]^ Recombinant adeno-associated (AAV) viral vectors are excellent vehicles for RNAi-based gene therapy as they allow efficient, safe, long-term gene delivery in a wide range of tissues. AAV is a single-stranded DNA virus with a 4.7-kb genome flanked by inverted terminal repeats (ITRs), which are *cis*-acting elements needed for AAV packaging and DNA replication. Recombinant AAV is devoid of all viral genes except the ITR sequences and is considered a safe vector as it lacks pathogenicity and persists mainly episomally, rather than integrating in the host genome. AAV vectors have great versatility and delivery potential through the availability of multiple serotypes, which can be expanded further by pseudotyping and capsid modification.^[@bib10],\\ [@bib11]^\n\nDespite successful demonstration of RNAi *in vivo* and initiation of clinical trials using RNAi-based therapeutics, safety and delivery of RNAi remain critical issues.^[@bib1],\\ [@bib12]^ Adverse immune responses, off-target effects and saturation of the cellular RNAi processing machinery have been reported, which demonstrates the need for optimization of therapeutic gene-silencing technologies.^[@bib13],\\ [@bib14],\\ [@bib15]^ For gene therapy-based RNAi, this could be achieved by controlling intracellular shRNA expression levels, which however requires cumbersome optimization. An alternative approach is to combine RNAi with gene-silencing technologies with a different mode of action that do not use the same processing pathway as RNAi.\n\nU1 interference (U1i) is a novel post-transcriptional gene-silencing method that inhibits polyadenylation and maturation of pre-mRNAs. The technology is based on a modified U1 small nuclear ribonucleoprotein particle (snRNP). Mammalian wild-type U1 snRNP consists of a 164-nt U1 small nuclear RNA (snRNA) bound by 10 polypeptides, one among which is U1-70K, and it functions as part of the spliceosome in pre-mRNA splicing by hybridizing to the 5\u2032 splice sites of introns using the 10 nucleotides at its 5\u2032 end for recognition. In addition, U1 snRNP has an alternative role in the inhibition of gene expression by binding the 3\u2032 untranslated region of papillomaviruses and certain mammalian genes.^[@bib16],\\ [@bib17],\\ [@bib18]^ Upon base-pairing of the 10 nucleotides at the 5\u2032 end of U1 snRNA to a target sequence in the 3\u2032 untranslated region, polyadenylation of the target pre-mRNA is inhibited through a direct interaction of the U1-70K protein subunit of U1 snRNP with poly(A) polymerase.^[@bib19]^ Inhibition of polyadenylation prevents maturation of pre-mRNA, which is subsequently degraded in the nucleus. By using this naturally occurring mechanism, a gene-specific silencing method based on inhibition of pre-mRNA polyadenylation and maturation is established. Specific modification of the 10 nucleotides at the 5\u2032 end of U1 snRNA leads to base-pairing with the 3\u2032 terminal exon of a desired target gene, resulting in the inhibition of polyadenylation, prevention of maturation, and finally in targeting of pre-mRNA for degradation. This approach has been successfully employed *in vitro* upon transient and stable delivery of DNA-encoding modified U1 snRNA to various cell lines, resulting in the repression of reporter genes as well as endogenous genes.^[@bib20],\\ [@bib21],\\ [@bib22],\\ [@bib23],\\ [@bib24],\\ [@bib25],\\ [@bib26]^\n\nCombining RNAi and U1i gene silencing yields more efficient knockdown of gene expression, as these methods have distinct mechanisms of action in different cellular compartments.^[@bib27]^ RNAi takes place in the cytosol and involves cleavage of mRNA, whereas U1i is effective in the nucleus and functions through prevention of pre-mRNA maturation.^[@bib6],\\ [@bib19]^ A combinatorial approach allows usage of minimal doses of individual components (RNAi and U1i), thereby reducing the chances of toxicity, such as saturation of cellular RNAi pathways. Simultaneous application of siRNA and U1i molecules that target the same gene has been shown to enhance silencing *in vitro*.^[@bib18]^ Furthermore, Abad *et al.*^[@bib27]^ recently demonstrated synergistic enhancement of reporter and endogenous gene silencing by the combined action of RNAi with U1i. However, the effectiveness of these combined silencing methods has yet to be demonstrated *in vivo*.\n\nHere, we employed the firefly luciferase reporter gene to quantify gene silencing by the combined action of RNAi and U1i. First, shRNA and U1i plasmids targeting firefly luciferase were co-transfected *in vitro* in HEK293T cells. Second, the *in vivo* activity of co-transfected shRNA and U1i plasmids was assessed in murine liver using hydrodynamic transfection. Finally, AAV-mediated transduction of murine muscle was employed to demonstrate stable *in vivo* suppression of luciferase expression by combinatorial RNAi and U1i. We show that co-transfection of RNAi with U1i constructs has an additive effect on reporter gene knockdown both *in vitro* and *in vivo*, compared with transfection with either effector alone. In addition, AAV-delivered RNAi and U1i result in prolonged inhibition of luciferase expression in murine muscle. Overall, these data illustrate the effectiveness of RNAi and U1i gene-silencing methods both *in vitro* and *in vivo*, which is of interest for the development of therapeutic applications.\n\nResults\n=======\n\nScreening of shRNA and U1i constructs targeting luciferase\n----------------------------------------------------------\n\nTo evaluate the inhibitory effect of RNAi and U1i on gene expression, we chose the firefly luciferase reporter as a target gene. Knockdown of luciferase was assayed by transfecting shRNA constructs targeting firefly luciferase together with firefly and renilla luciferase reporter plasmids. The ratio of firefly and renilla luciferase signal was used to calculate relative luciferase activity. shLuc1 was highly effective and showed a dose-dependent luciferase knockdown up to approximately 85%, whereas shLuc2 was ineffective ([Figure 1a](#fig1){ref-type=\"fig\"}).\n\nEight U1 interference (U1i) constructs targeting firefly luciferase (L1--L8) were designed and screened for knockdown of firefly luciferase. Co-transfection of L1--L8 constructs with firefly and renilla luciferase reporters demonstrated a mild inhibitory effect of two out of the eight U1i constructs ([Figure 1b](#fig1){ref-type=\"fig\"}). L4 and L5 reduced luciferase expression by 30 and 40%, respectively. We proceeded to test the combined effect of the two mildly effective U1i constructs with other U1i constructs in order to screen for improved efficacy (Supplementary Table 1). The strongest decrease in luciferase expression was observed by the combination of L4 with L5 and reached around 80% (Supplementary Table 1 and [Figure 1c](#fig1){ref-type=\"fig\"}). Next, a bicistronic expression vector was cloned that simultaneously expressed L4 and L5 from one backbone. Transfection with this double U1i plasmid L4/L5 was equally effective in suppressing luciferase activity as co-transfection with L4 and L5 single plasmids, indicating that there is efficient expression of both effectors from one backbone (data not shown). Thus, although the single U1i constructs tested were only mildly effective on their own, they can be combined to reduce luciferase expression to a similar level as the shLuc1 construct.\n\nCo-expression of shRNA and U1i enhances luciferase knockdown *in vitro*\n-----------------------------------------------------------------------\n\nCombining RNAi and U1i gene silencing allows more efficient knockdown of gene expression as they have distinct mechanisms of action in different cellular compartments. Therefore, we co-transfected shLuc1 together with the most effective U1i constructs L4 and L5. Transfection of L4 or L5 combined with control shApoB led to a mild decrease in luciferase expression of approximately 30 and 40%, respectively, comparable to the single U1i constructs ([Figures 1b](#fig1){ref-type=\"fig\"} and [2a](#fig2){ref-type=\"fig\"}). Expression of the control plasmid shApoB together with the combined expression of L4 and L5 or the bicistronic L4/L5 construct suppressed luciferase even more ([Figure 2a](#fig2){ref-type=\"fig\"}). When shLuc1 was co-transfected with control U1C, an 86% reduction in luciferase expression was observed, similar to the inhibition shown with shLuc1 alone ([Figures 1a](#fig1){ref-type=\"fig\"}, [2a and b](#fig2){ref-type=\"fig\"}). We were able to significantly enhance luciferase knockdown when the three luciferase-targeting constructs, shLuc1, L4, and L5, were combined and when shLuc1 was combined with the bicistronic L4/L5 construct, resulting in a 95% reduction in luciferase activity ([Figure 2b](#fig2){ref-type=\"fig\"}). The additive effect of L4 and L5 was shown for several shLuc1 concentrations (Supplementary Figure 1). Hence, co-expression of RNAi and U1i enhances target gene knockdown *in vitro*, consistent with previous data.^[@bib27]^\n\nTransient *in vivo* delivery of shRNA and U1i enhances luciferase knockdown in murine liver\n-------------------------------------------------------------------------------------------\n\nThe *in vivo* silencing activity of RNAi and U1i was evaluated in murine liver by hydrodynamic co-transfection of shLuc1 and bicistronic L4/L5 plasmids. shLuc1 was mixed with L4/L5 and injected together with the firefly luciferase reporter plasmid and a construct expressing secreted alkaline phosphatase (SEAP). Plasma measurements of co-injected SEAP plasmid served to normalize for transfection efficiency. *In vivo* delivery of shLuc1 and L4/L5 plasmids resulted in 85% knockdown of relative luciferase expression at two days post-transfection, which was comparable to luciferase knockdown observed *in vitro* ([Figures 2b](#fig2){ref-type=\"fig\"} and [3](#fig3){ref-type=\"fig\"}). In addition, we observed intermediate suppressive effects of L4/L5 alone (25% in combination with control shApoB) and of shLuc1 alone (61% in combination with control H1/H4) ([Figure 3](#fig3){ref-type=\"fig\"}). These results demonstrate the feasibility of *in vivo* suppression of a target gene by the combined action of RNAi and U1i.\n\nSimultaneous expression of shRNA and bicistronic U1i constructs from a single backbone inhibits luciferase expression *in vitro*\n--------------------------------------------------------------------------------------------------------------------------------\n\nIn order to develop therapeutic applications using RNAi and U1i, durable expression of these gene-silencing mechanisms has to be established. For long-term expression, AAV vectors are excellent delivery vehicles that are capable of transducing a wide range of tissues with high efficiency. Therefore, we proceeded to express shLuc1 and L4/L5 from an AAV backbone. The U1i bicistronic construct L4/L5 was cloned together with shLuc1 in a pro-AAV cloning vector (LL) (Supplementary Figure 2). Each individual effector was expressed from their own promoter and termination sequences. Three additional vectors were made: one that expressed control shApoB and control U1i H1/H4 (AH), a second that expressed anti-luciferase U1i L4/L5 and control shApoB (AL), and a third that expressed anti-luciferase shLuc1 and control U1i H1/H4 (LH). Upon transfection of HEK293T cells, qPCR analysis confirmed that all four shRNA/U1i constructs expressed recombinant U1 snRNA (Supplementary Figure 3a). Furthermore, their expression did not affect cell viability, as determined by the MTT toxicity assay (Supplementary Figure 3b). To assess the *in vitro* inhibitory potential of the shRNA/U1i co-expression constructs, increasing amounts of plasmid were co-transfected together with firefly and renilla reporters. All luciferase-targeting constructs (LH, AL, LL) significantly suppressed luciferase activity compared with the control AH ([Figure 4](#fig4){ref-type=\"fig\"}). Of these, the RNAi and U1i double inhibitory construct (LL) was most effective and reduced luciferase expression by 85%. Although the combined effect of shLuc1 and L4/L5 (LL) appeared to be stronger than the effect of shLuc1 alone (LH), this effect was not significant. This may be due to the fact that all three constructs were expressed from one vector, thereby missing the ability to differentially dose the inhibitory constructs for optimal effect.\n\nAAV-mediated knockdown of luciferase using RNAi and U1i in murine muscle\n------------------------------------------------------------------------\n\nHaving established the *in vitro* efficacy of the shRNA/U1i co-expression constructs, we proceeded to investigate long-term *in vivo* suppression of luciferase in murine muscle by employing AAV serotype 1 viral vectors to deliver shLuc1 and L4/L5. AAV1-expressing luciferase reporter (AAV-Luc) was co-injected intramuscularly with the AAV viral vectors encoding the shRNA/U1i co-expression constructs AAV-AH (shApoB and H1/H4), AAV-AL (shApoB and L4/L5), AAV-LH (shLuc1 and H1/H4), or AAV-LL (shLuc1 and L4/L5). All AAV-shRNA/U1i vectors co-expressed the GFP reporter gene, and full muscle transduction was demonstrated by monitoring GFP expression at 8 weeks post-injection (data not shown). Recombinant U1 snRNA expression was shown for all four AAV-shRNA/U1i vectors in transduced muscle at 8 weeks post-transduction (Supplementary Figure 4). In addition, the absence of muscle damage was demonstrated by creatine phosphokinase (CPK) assay and by hematoxylin and eosin (H&E) staining of frozen muscle sections (data not shown).\n\nLuciferase expression was monitored at 1, 2, 3, 4, 6, and 8 weeks post-transduction ([Figure 5](#fig5){ref-type=\"fig\"}). Overall, luciferase expression increased over time during the first weeks post-transduction with AAV-Luc and AAV-shRNA/U1i, and reached a relatively stable level at 4 weeks, characteristic of AAV expression kinetics.^[@bib28]^ Transduction with AAV-shRNA/U1i targeting firefly luciferase (AAV-LL) reduced luciferase expression significantly by 25--40% compared with the control vector AAV-AH ([Figure 5](#fig5){ref-type=\"fig\"}). This effect was already observed at 1 week post-transduction and remained stable throughout the course of the experiment, demonstrating the effectiveness and stability of luciferase knockdown by combined RNAi and U1i expression *in vivo*. Overall, luciferase suppression *in vivo* was lower than observed *in vitro*; the maximum inhibitory effect of shLuc and L4/L5 (LL) was 40% *in vivo*, versus 85% *in vitro* ([Figures 4](#fig4){ref-type=\"fig\"} and [5](#fig5){ref-type=\"fig\"}). Moreover, we could not observe a significant effect of the U1i constructs alone (AAV-AL), or of the shRNA construct alone (AAV-LH). The fact that we did not observe an intermediate inhibitory effect of these constructs may reflect an overall lower knockdown *in vivo* compared with the effect measured *in vitro* ([Figures 4](#fig4){ref-type=\"fig\"} and [5](#fig5){ref-type=\"fig\"}). In conclusion, we demonstrated long-term stable transduction of murine muscle using AAV delivery vectors and observed sustained *in vivo* knockdown of a luciferase reporter by the combined action of RNAi and U1i.\n\nDiscussion\n==========\n\nDelivery and safety of the therapeutic molecules remain major challenges in the RNAi field. These safety concerns involve unwanted activation of the immune response, off-targeting effects, and saturation of the cellular RNAi machinery due to overexpression of the shRNA.^[@bib13],\\ [@bib14],\\ [@bib15]^ To minimize the risk of these unwanted effects while retaining high target inhibition, we investigated the combined action of RNAi and U1i, two gene-silencing mechanisms with different modes of action. As a proof of concept, we studied the inhibitory effect of RNAi and U1i on luciferase reporter expression, and demonstrated increased luciferase knockdown *in vitro* by shRNA and U1i transfection as well as prolonged suppression *in vivo* using AAV-delivered shRNA/U1i expression cassettes.\n\nCombinatorial RNAi strategies that aim to avert viral resistance (for example, of HIV) by triggering RNAi against multiple targets have been previously developed. These studies often involve expression of several shRNA molecules, which is expected to proportionally increase the toxicity and off-targeting risk associated with mono-RNAi treatments.^[@bib29]^ To circumvent these problems, Hemmings-Mieszczak *et al.*^[@bib30]^ combined siRNAs with antisense oligonucleotides and observed synergistic knockdown effects in cultured cells. Similarly, Jarczak *et al.*^[@bib31]^ mixed shRNAs with hammerhead ribozymes, which overall increased target gene inhibition. The *in vivo* efficacy of these combinatorial approaches nevertheless remains to be demonstrated. Combinatorial use of RNAi with U1i in cell lines has been described previously for both reporter and endogenous genes, and was shown to yield synergistic suppressive effects on target gene expression.^[@bib18],\\ [@bib27]^ We observed a similar effect when transfecting shLuc1 with two U1i constructs L4 and L5, which resulted in 95% inhibition of luciferase activity. Of the individual components, shLuc1 was most effective, yielding \u223c85% knockdown, and the combination of L4 with L5 resulted in \u223c70--80% luciferase inhibition. Previously, increased target knockdown was observed when several U1i constructs were expressed that recognized multiple targets, or when a U1i construct was expressed that recognized repeat regions in the 3\u2032 terminal exon of a target gene.^[@bib22]^ However, the U1i constructs tested here were only mildly effective on their own. When screening the eight U1i constructs (L1--L8), we observed not only inhibition of luciferase expression, but even stimulatory effects on luciferase expression by two of the U1i constructs, L3 and L6; the cause for this is currently unclear. Successful design of U1i constructs seems to depend most strongly on the accessibility of the secondary structure of the target region and may have been suboptimal.^[@bib26]^ In addition, Abad *et al.*^[@bib27]^ observed a milder inhibitory effect of their U1i construct when the target sequence was expressed within the firefly luciferase plasmid, compared with expression in the previously used renilla luciferase construct, and suggested that this decrease in efficacy might reflect the effect of neighboring sequences.\n\nTo prove the feasibility of combinatorial RNAi and U1i, we aimed at demonstrating luciferase knockdown *in vivo* upon delivery of shRNA and U1i constructs in a murine model. U1i has not been established in animal models yet, and this is considered an important step in the development of this gene-silencing mechanism as a therapy.^[@bib32]^ Two approaches were employed, each targeting a different organ; transient luciferase knockdown by plasmid delivery of shLuc1 and L4/L5 to the liver, and their long-term effect by AAV-mediated expression in the muscle. A strong reduction in luciferase expression was shown upon hydrodynamic transfection of murine liver with shLuc1 and the bicistronic L4/L5 construct, and intermediate inhibitory effects were observed after transfection with shLuc1 and L4/L5 alone. These results matched the observations in cell cultures, and prompted us to investigate the long-term silencing ability of combinatorial RNAi and U1i in murine muscle. AAV-mediated transduction of murine muscle is expected to yield high-level long-term expression, and indeed, the AAV-delivered vector co-expressing shLuc1 and double L4/L5 significantly reduced luciferase expression compared with the control AAV vector. However, the effect was relatively mild compared with *in vitro* observations. In addition, no significant luciferase suppression by shLuc1 or L4/L5 alone was measured. Possibly, the lower efficacy is due to the fact that all three effectors were expressed from one backbone, which might result in promoter interference when expressed *in vivo*. Alternatively, co-injection of AAV-Luc with AAV-shRNA/U1i viral vectors may have resulted in transduced cells that express luciferase, but not the shRNA/U1i inhibitors, or *vice versa*. This would lower the overall efficacy of luciferase knockdown. Nevertheless, the inhibitory effect of AAV-delivered shLuc and L4/L5 lasted throughout the course of the 8-week experiment and failed to show any obvious signs of toxicity, as determined by CPK assay and H&E staining. Thus, although the AAV vectors and possibly the AAV dose need to be optimized, we were able to show a proof of principle for the *in vivo* use of combinatorial RNAi and U1i.\n\nFuture research is needed to optimize the inhibitory potential of combined RNAi and U1i *in vivo* when delivered with AAV, and furthermore requires investigation into the potential toxic or off-target effects of U1i gene silencing. With respect to the latter, U1i has the advantage that it only works in the 3\u2032 terminal exon of a target gene, and that secondary structures occlude target sequences.^[@bib22],\\ [@bib26]^ Indeed, Goraczniak *et al.*^[@bib18]^ reported comparable limited off-target effects by siRNA and U1 snRNA-mediated gene silencing in microarray analyses. In addition, the naturally occurring U1 snRNP is highly abundant, with about one million copies present in a typical mammalian nucleus, and therefore overexpression of modified U1i constructs is not expected to interfere with the overall splicing machinery.^[@bib33]^ Furthermore, stable expression of U1i constructs has been demonstrated in stably transduced cell lines, thus indicating the low toxicity of this technique.^[@bib22]^ Finally, optimization and validation of U1i design rules should improve target gene knockdown, which allows use of lower dose of U1i constructs, thus aiding in improving the safety of the U1i gene-silencing technique. In conclusion, this study shows for the first time the *in vivo* application of U1i and furthermore demonstrates the *in vivo* efficacy of combinatorial RNAi and U1i.\n\nMaterials and methods\n=====================\n\nDNA constructs\n--------------\n\nshRNA constructs were made by annealing complementary oligonucleotides and ligating them into the *Bgl*II and *Xho*I site of the pSuper vector (OligoEngine, Seattle, WA, USA). shRNAs expressed from this plasmid are transcribed from a H1 promoter. Two shRNAs were designed that target firefly luciferase (shLuc1 and shLuc2) and two control shRNAs were made that target GFP or human ApoB (shGFP and shApoB). The sequences of the oligonucleotides used in this study are listed in Supplementary Table 2. Eight U1i constructs were designed that target firefly luciferase (L1--L8), and three control U1i constructs were made that target human huntingtin (H1 and H4) or express the wild-type U1 sequence (U1C).^[@bib22]^ The locations of each shRNA and U1i luciferase target sequence are shown in Supplementary Figure 5. U1i constructs were made by annealing complementary oligonucleotides and ligating them into the *Bgl*II and *Bcl*I sites of a pGEM-3Zf+ plasmid (Promega, Madison, WI, USA) encoding the U1 snRNA sequence including promoter and termination sequences.^[@bib26]^ Bicistronic U1i constructs encoding L4/L5 and H1/H4 were made by digestion of L4 and H4 with XbaI, followed by blunting of these digested plasmids. Next, the constructs were digested with EcoRI. The resulting double-digested L4 and H4 were then ligated into the blunt SmaI site and EcoRI site of L5 and H1, respectively. Selected shRNA (shLuc1 and shApoB) and U1i double (L4/L5 and H1/H4) constructs were further subcloned into a pro-AAV vector that encodes CMV-eGFP, using SphI (for shRNA), and MunI and SalI (for double U1i constructs) (Supplementary Figure 2). The resulting AAV-shRNA/U1i plasmids are AH (shApoB/H1/H4), AL (shApoB/L4/L5), LH (shLuc1/H1/H4), and LL (shLuc1/L4/L5). Reporter constructs that were used are pRL-CMV-renilla (Promega), and firefly luciferase pGL4 (Promega) under control of the CMV promoter or the liver-specific AAT promoter in a pGEM-3Zf+ plasmid (Promega). Plasmid encoding AAT-secreted alkaline phosphatase (AAT-SEAP) was made by BaseClear (Leiden, The Netherlands).\n\nAAV vector production\n---------------------\n\nAAV serotype 1 vectors were produced by calcium phosphate-mediated co-transfection in human embryonic kidney (HEK) 293 cells as described previously.^[@bib34]^ Viral batches were produced for AAV-shRNA/U1i constructs AH, AL, LH, and LL and crude lysate was purified by affinity chromatography, diafiltrated in 5% PBS sucrose (v/w) by cross-flow filtration and further concentrated using centricon tubes (YM-100; Millipore, Bedford, MA, USA). The final concentration of 5--10 \u00d7 10^11^ genome copies per ml was determined by qPCR (Applied Biosystems, Foster City, CA, USA) using primers pr59 and pr60 (Supplementary Table 2), amplifying a 70-bp fragment from the CMV promoter region. AAV-CMV-luciferase (AAV-Luc) was produced in insect cells through the use of recombinant baculovirus technology, as described in Smith *et al.*^[@bib35]^\n\nLuciferase assay\n----------------\n\nThe HEK293T cell line was maintained in Dulbecco\\'s modified Eagle\\'s medium (DMEM; Invitrogen, Carlsbad, CA, USA) containing 10% fetal calf serum, 100\u2009U\u2009ml^\u22121^ penicillin, and 100\u2009U\u2009ml^\u22121^ streptomycin at 37\u2009\u00b0C and 5% CO~2~. For luciferase assays, approximately 2.5 \u00d7 10^4^ HEK293T cells were seeded in 96-well plates and transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer\\'s instructions. Cells were co-transfected with 2.5\u2009ng CMV-firefly luciferase plasmid and 0.5\u2009ng pRL-CMV-renilla luciferase plasmid, together with a selection of shRNA (1--250\u2009ng), U1i (100--200\u2009ng), and shRNA/U1i (25--250\u2009ng) constructs. The total amount of transfected DNA was kept constant per experiment by adding vector control plasmid (pSuper, pGEM-3Zf+, or pro-AAV plasmid). Two days after transfection, firefly and renilla luciferase expression was assessed using a luminometer (Spectramax L; Molecular Devices, Sunnyvale, CA, USA) according to the Dual-Luciferase Reporter Assay System (Promega). Relative luciferase activity was calculated as the ratio between firefly and renilla luciferase activities. To correct for between-session variation in replicate experiments, the factor correction program was used.^[@bib36]^ Statistical analysis was performed using one-way analysis of variance, with or without repeated measures, and a Bonferroni *post hoc* comparison. A *P*-value \\<0.05 was considered significant.\n\n*In vivo* transfection of murine liver using hydrodynamic tail vein injection\n-----------------------------------------------------------------------------\n\nAll animal experiments were conducted according to the guidelines of the local animal welfare committee. Six-to-eight-week-old female C57BL/6 mice were anesthetized using intraperitoneal injection with fentanyl-fluanisone-midazolam (FFM) mix and hydrodynamically transfected via the tail vein with plasmids encoding AAT-Luc (2\u2009\u03bcg), AAT-SEAP (0.5\u2009\u03bcg), shRNA (shApoB or shLuc1) (10\u2009\u03bcg), and double U1i (H1/H4 or L4/L5) (40\u2009\u03bcg) in 1.5\u2009ml Ringer\\'s solution (0.9% NaCl, 0.03% KCl, and 0.016% CaCl~2~). Two days after transfection, luciferase was measured in living mice in the Spectrum *In Vivo* Imaging System (IVIS, Caliper Life Sciences, Hopkinton, MA, USA). Mice were anesthetized with 2% isoflurane and injected intraperitoneally with 3\u2009mg -luciferin (Synchem, Felsberg, Germany) in 200\u2009\u03bcl PBS. Regions of interest (ROIs) were traced over the positions of greatest signal intensity on the animal and light intensity was quantified as photons\u2009s^\u22121^\u2009cm^\u22122^\u2009sr^\u22121^. At the same time point, plasma was collected by retro-orbital bleed and secreted alkaline phosphatase (SEAP) was measured using the chemiluminescent SEAP reporter gene assay according to the manufacturer\\'s instructions (Roche Diagnostics GmbH, Mannheim, Germany). Luciferase readings were normalized to SEAP levels to adjust for transfection efficiency. Statistical analysis was performed using one-way analysis of variance and a Bonferroni *post hoc* comparison. A *P*-value \\<0.05 was considered significant.\n\n*In vivo* transduction of murine muscle using AAV\n-------------------------------------------------\n\nSix-to-eight-week-old female Balb/C mice were injected intramuscularly with a mixture of AAV-Luc and AAV-shRNA/U1i (AAV-AH, AAV-AL, AAV-LH, or AAV-LL). Gastrocnemic and adductor muscles were injected with 1 \u00d7 10^10^ gc per limb AAV-Luc and 4.5 \u00d7 10^10^ gc per limb AAV-shRNA/U1i. Luciferase expression in living mice was monitored in the IVIS at 1, 2, 3, 4, 6, and 8 weeks post-transduction, as described above. Statistical analysis was performed using one-way analysis of variance with repeated measures and a Bonferroni *post hoc* comparison. A *P*-value \\<0.05 was considered significant.\n\nWe thank Amaya Abad for technical assistance and Mark Chadwick and Wim Hermens for critically reviewing the manuscript.\n\nSupplementary Information accompanies the paper on Gene Therapy website (http://www.nature.com/gt)\n\nThe authors declare no conflict of interest.\n\nSupplementary Material {#sup1}\n======================\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n![*In vitro* knockdown of luciferase by shRNA and U1i constructs. (**a**) Increasing amounts (1--250\u2009ng) of shRNA constructs targeting firefly luciferase (shLuc1 and shLuc2) were co-transfected with 2.5\u2009ng firefly luciferase and 0.5\u2009ng renilla luciferase plasmid in 96-well plates containing approximately 2.5 \u00d7 10^4^ HEK293T cells. The total amount of transfected DNA was kept constant by adding the pSuper cloning vector. At 2 days after transfection, cells were lysed and firefly and renilla luciferase signals were measured. Relative luciferase activity was calculated as the ratio of firefly and renilla luciferase and plotted relative to the 100% control transfected with 250\u2009ng pSuper. shLuc1 significantly reduced luciferase by 85%, whereas shLuc2 was inactive. Control shRNA plasmids were shGFP and shApoB. (**b**) Co-transfection of 200\u2009ng U1i constructs targeting firefly luciferase (L1-L8) with firefly luciferase and renilla luciferase plasmid. Cells transfected with 200\u2009ng control plasmid U1C were set at 100%. L4 and L5 mildly reduced luciferase expression by 30 and 40%, respectively, compared with U1C. (**c**) Co-transfection of 100\u2009ng L4 and 100\u2009ng L5 significantly improved luciferase inhibition, compared with transfection with 200\u2009ng L4 or L5 alone (*P*\\<0.05). Data are presented as mean of three technical replicates\u00b1s.d.](gt201141f1){#fig1}\n\n![Combination of RNAi and U1i increases luciferase knockdown *in vitro*. (**a**) Co-transfection of 50\u2009ng control shApoB construct or 50\u2009ng shLuc1 construct with 200\u2009ng (or 100+100\u2009ng) U1i plasmid and firefly luciferase and renilla luciferase plasmid. Transfections and analysis were carried out as described in [Figure 1a](#fig1){ref-type=\"fig\"}. Cells transfected with a combination of 50\u2009ng shApoB and 200\u2009ng U1C were set at 100%. Co-transfection of shApoB with L4 and L5 or L4/L5 significantly reduced luciferase activity by \u223c65% compared with the 100% control (*P*\\<0.05). Transfection of shLuc with U1C resulted in a significant 86% knockdown (*P*\\<0.001). (**b**) Co-transfection of 50\u2009ng shLuc1 construct with 200\u2009ng (or 100+100\u2009ng) U1i plasmid and firefly luciferase and renilla luciferase. Note the different scales on the *y* axes in (**a**, **b**). Co-transfection of shLuc1 with L4 and L5 maximally reduced luciferase expression by 95%. Control U1i constructs were U1C, and H1 and H4 that target the human huntingtin sequence. Luciferase activity after co-transfection of shLuc1 with L4 and L5 or L4/L5 was significantly different from co-transfection of shLuc1 with U1C (*P*\\<0.05), and from co-transfection of shLuc1 with H1 and H4 (*P*\\<0.001) or H1/H4 (*P*\\<0.05). Data are presented as mean of three independent experiments +s.d., analyzed using factor correction.^[@bib36]^](gt201141f2){#fig2}\n\n![Hydrodynamic delivery of shRNA and U1i constructs to murine liver inhibits luciferase activity *in vivo*. Mice were hydrodynamically co-transfected with 10\u2009\u03bcg shRNA (shLuc1 or control shApoB) and 40\u2009\u03bcg double U1i (L4/L5 or control H1/H4) plasmid and with 2\u2009\u03bcg firefly luciferase reporter and 0.5\u2009\u03bcg secreted alkaline phophatase (SEAP) plasmid. Both the firefly luciferase and SEAP reporter gene were expressed under control of the liver-specific AAT promoter, as hydrodynamically delivered plasmid DNA localizes mainly to the liver. Two days after transfection, bioluminescence was measured in the IVIS, and luciferase signal was calculated relative to plasma SEAP levels. Murine liver co-transfected with control constructs shApoB and H1/H4 was set at 100%. Co-transfection of shLuc1 with L4/L5 maximally reduced luciferase activity by 84%. However, luciferase knockdown was not significantly different between treatments. Data are presented as mean (*n*=4--8)\u00b1s.e.](gt201141f3){#fig3}\n\n![Co-expression of RNAi and U1i from one vector backbone suppresses luciferase activity *in vitro*. Increasing amounts (25--250\u2009ng) of shRNA/U1i plasmid expressing both shRNA and double U1i constructs were co-transfected with firefly luciferase and renilla luciferase plasmid and analyzed as described in [Figure 1a](#fig1){ref-type=\"fig\"}. The total amount of transfected DNA was kept constant by adding the pro-AAV cloning vector. Cells transfected with 250\u2009ng pro-AAV vector were set at 100%. The control vector AH expressed shApoB and double H1/H4. The U1i vector AL expressed shApoB and double L4/L5. The shRNA vector LH expressed shLuc1 and double H1/H4. The shRNA and U1i combination vector LL expressed shLuc1 and double L4/L5. AL, LH, and LL significantly reduced luciferase activity compared with the AH control vector (*P*\\<0.001) and LL was most effective and led to 85% luciferase knockdown. Data are presented as mean of three independent experiments\u00b1s.d., analyzed using factor correction.^[@bib36]^](gt201141f4){#fig4}\n\n![AAV-mediated *in vivo* knockdown of luciferase using RNAi and U1i. Murine muscle was co-transduced with 1 \u00d7 10^10^ gc AAV-Luc and 4.5 \u00d7 10^10^ gc AAV-shRNA/U1i per hind limb. The gastrocnemic and adductor muscles of both hind limbs were injected. Six mice per treatment were injected, resulting in *n*=12. Luciferase expression was measured in the IVIS at 1, 2, 3, 4, 6, and 8 weeks post-transduction. The control vector AAV-AH expressed shApoB and double H1/H4. The U1i vector AAV-AL expressed shApoB and double L4/L5. The shRNA vector AAV-LH expressed shLuc1 and double H1/H4. The shRNA and U1i combination vector AAV-LL expressed shLuc1 and double L4/L5, and this viral vector significantly reduced luciferase expression on average by 25% during the course of the experiment (*P*\\<0.05). Data are presented as mean (*n*=12)\u00b1s.e.](gt201141f5){#fig5}\n"}
+{"text": "INTRODUCTION\n============\n\nThe acetabular labrum provides a significant functional role and may be important in overall hip preservation \\[[@hnv001-B1]\\]. Labral tears often occur in the setting of bony dysmorphisms such as dysplasia and femoroacetabular impingement (FAI). Arthroscopic surgery for FAI is a less invasive option that typically addresses bony deformities and chondrolabral pathology \\[[@hnv001-B5], [@hnv001-B6]\\]. Patients undergoing labral refixation have better outcomes than those undergoing labral debridement \\[[@hnv001-B7]\\]. When the labrum is deficient and/or irreparable, labral reconstruction has been performed with early successful outcomes \\[[@hnv001-B11]\\]. Current labral repair and reconstruction techniques utilize suture anchor fixation.\n\nSuture anchors provide a common method for fixation of soft tissues to bone. In order to restore labral function including the labral fluid seal effect, the fixation device should be placed on the acetabular rim close to but not violating the articular cartilage or joint \\[[@hnv001-B16], [@hnv001-B17]\\]. Suture anchors placed too far from the articular cartilage can evert or medialize the labrum, compromising its function; anchors placed too close to the articular cartilage can cause iatrogenic cartilage injury. Complications related to suture anchors have been documented, mainly in the shoulder literature including chondrolysis, osteolysis, synovitis, articular cartilage damage and fixation failure \\[[@hnv001-B18]\\]. Consequences from anchor-induced chondral damage may be devastating, especially in young patients, and treatment options may be suboptimal and few (e.g. arthrodesis or arthroplasty). Several studies and systematic reviews have documented complications from hip arthroscopy \\[[@hnv001-B23]\\] however we could find no case reports or case series of anchor-induced chondral damage associated with hip arthroscopy.\n\nThe purpose of this case series is to introduce anchor-induced chondral damage as a formal complication of hip arthroscopy while presenting arthroscopic treatment options, preventative measures and outcomes.\n\nMETHODS\n=======\n\nFive patients were identified among four high-volume arthroscopic hip surgeons (arbitrarily defined as 100+ cases per annum). Inclusion criteria were documentation in operative report of violation (penetration or deformation) of acetabular articular cartilage attributed to one or more suture anchors placed during hip arthroscopic surgery and/or arthroscopic confirmation on initial and/or revision surgery with minimum 1-year follow-up. Exclusion criteria included patients without confirmed documentation and arthroscopic confirmation of anchor-induced chondral violation or \\<1 year follow-up. Retrospective medical record review was performed by each respective surgeon including pre-operative diagnosis, arthroscopic procedure(s), location on acetabular rim by clockface method, number and, where known, type and size of suture anchors, portals used for anchor placement, post-operative outcomes with patient-reported outcome measure where available and type/outcome of any revision surgeries.\n\nCase 1\n------\n\nA 33-year-old female recreational runner underwent hip arthroscopy for cam FAI. Diagnostic arthroscopic evaluation of the central compartment revealed an anterior labral tear at the 1--2 o\\'clock position. Labral repair was performed utilizing two 2.3-mm suture anchors (Osteoraptor, Smith and Nephew, Andover, MA, USA) placed via the mid-anterior portal. After placement of the suture anchors, inspection of the acetabular cartilage revealed focal ballooning of the articular chondral surface adjacent to the labral repair ([Fig. 1](#hnv001-F1){ref-type=\"fig\"}). Post-operative rehabilitation was uneventful with protective weightbearing for 2 weeks. She returned to normal activities after 6 months without symptoms. Two-year follow-up revealed improvement of modified Harris Hip Score Hip (mHHS) (from 67 to 96), Hip Outcome Score Daily Activity Part 1 (from 88 to 100) and Sport Specific Part 2 (from 33 to 86) and post-operative radiographs revealed joint space preservation ([Fig. 2](#hnv001-F2){ref-type=\"fig\"}). Fig. 1.Supine arthroscopic image of right hip from anterolateral portal showing acetabular chondral elevation adjacent to the labral repair with suture anchor fixation. Fig. 2.Post-operative AP radiograph of right hip. Note well-preserved joint.\n\nCase 2\n------\n\nA 16-year-old male high-school swimmer with mixed FAI which was treated with elective arthroscopic rim trimming and femoroplasty. During suture anchor placement via mid-anterior portal for anterosuperior labral refixation, the acetabular articular cartilage was elevated but not penetrated by a 3.5-mm knotless anchor (Push-lock, Arthrex, Naples, FL, USA). The anchor was not removed. The patient had an uneventful post-operative course and eventually resumed asymptomatic competitive swimming and intramural collegiate basketball. His pre-operative non-arthritic hip score (NAHS) was 34 and his latest post-operative NAHS was 89 at 6 years following surgery and radiographs show no joint deterioration or narrowing.\n\nCase 3\n------\n\nA 30-year-old semi-professional bowler with symptomatic cam-pincer FAI underwent elective arthroscopic acetabuloplasty and femoroplasty. Labral refixation was performed via the modified mid-anterior portal \\[[@hnv001-B27]\\] with 3.5-mm suture anchors (Pushlock, Arthrex, Naples, FL, USA) was performed at the 1 and 2 o'clock positions. During suture anchor placement, mild elevation of adjacent acetabular articular cartilage without implant penetration was seen. Sixteen months later, because of recurrent groin pain and index of suspicion, revision hip arthroscopy was performed at that time the suture anchor was partly visible under a 'veil' of thin translucent articular cartilage. Probing revealed no gross motion of the seated implant. A 4-mm unhooded burr was used to remove the encroaching longitudinal side of the anchor, initially at chondral surface level, to a subchondral level 2\u2009mm below the articular surface. After an initial 6 weeks of protected weightbearing, he felt initial subjective improvement. But by 10 months, his pain had returned. He underwent a second revision arthroscopy. Intra-operatively, the previously burr-resected side of the anchor was exposed at the level of the adjacent chondral surfaces ([Fig. 3](#hnv001-F3){ref-type=\"fig\"}) but not loose. Outerbridge grade 2 changes were observed at the anterior apex of the femoral head. Conservative 'unroofing' of marginal cartilage was done with a shaver. The anchor was dislodged using a microfracture awl and subsequently removed with a grasper through an 8-mm arthroscopic cannula without incident ([Fig. 4](#hnv001-F4){ref-type=\"fig\"}). The remaining trough was treated with microfracture chondroplasty ([Fig. 5](#hnv001-F5){ref-type=\"fig\"}). Pain and mechanical symptoms gradually resolved. Pre-operative NAHS of 36 decreased to a nadir of 18 at 10 months post-first revision, and increased to 88 at 2.4 years post-second revision surgery. Post-operative radiographs showed no joint narrowing. He has resumed competitive bowling and limits impact sports. Fig. 3.Supine arthroscopic image from anterolateral portal of left hip during establishment of modified mid-anterior portal (note entry needle and stylet). One can see an exposed hard anchor (arrow) that has penetrated the anterior acetabular articular cartilage. Ac, acetabulum, FH, femoral head. Fig. 4.Arthroscopic image of left hip showing a microfracture awl displacing the embedded but exposed anchor. Fig. 5.Arthroscopic image of left hip after anchor removal and microfracture chondroplasty. Ac, acetabulum, FH, femoral head.\n\nCase 4\n------\n\nA 40-year-old active woman underwent arthroscopic femoral osteoplasty, acetabuloplasty and labral refixation. She complained of more pain following surgery than prior to surgery, had some radiographic medial joint narrowing (initial pre-operative radiographs were unavailable). She elected revision hip arthroscopy, where her anterosuperior labral repair had failed, and she had full thickness articular cartilage flaps and exposed bone. A suture was noted in this area on the articulating surface, and followed to an anchor on the mid-articular surface. The angle of the suture anchor appears that it had entered the acetabular rim at the 1 o'clock position and no distal-based portals had been used in the initial hip arthroscopy. In addition, there was wearing of the articular cartilage on the central femoral head. The patient had the anchor and sutures removed as well as the articular cartilage and labrum debrided. The degree of articular cartilage loss was deemed too excessive to justify microfracture chondroplasty. She subjectively feels 90% of normal and does not feel limited by her hip. At 1.5 years, she had improvement of the IHOT-33 score (from 11.2 to 68.8, with global change rating of 90) and mHHS (from 30.8 to 84.7). Radiographs demonstrate some progression of her medial joint space narrowing.\n\nCase 5\n------\n\nA 41-year-old woman had unrelenting pain and locking following right hip arthroscopy with a 2-anchor labral repair performed by another physician 3 months prior. No distal-based portals had been used in that surgery. Her radiographs showed preserved joint space with residual cam and pincer FAI, which was not treated in her previous procedure. The patient elected revision surgery to address her worsening pain and unaddressed FAI. During revision hip arthroscopy, both of her previous anchors for her labral repair were found to have penetrated her subchondral bone, had become intra-articular, and significantly damaged the cartilage on the acetabulum at the 12 o'clock and 2 o'clock positions ([Fig. 6](#hnv001-F6){ref-type=\"fig\"}). Arthroscopic removal of the two anchors and debridement of the joint was performed. The chondral and labral damage was deemed too excessive to justify labral reconstruction. Her pain continued to worsen and given the extensive damage to the joint, the decision was made to proceed with total hip replacement at 42 years of age. Fig. 6.Supine arthroscopic image of right hip from anterolateral portal showing extensive acetabular chondral damage from one (arrow) of two penetrating hard anchors. Ant, anterior, Sup, superior.\n\nRESULTS\n=======\n\nFive patients (three females) of mean age 32 years (range, 16--41 years) had documented anchor-induced chondral damage with mean 3.5 year (range, 1.5--6.0 years) follow-up. All cases involved various hard suture anchors placed in along the anterosuperior acetabular rim with the 1 o'clock position most commonly involved (four cases). Various portals were utilized for anchor placement (two each via anterior portal and mid-anterior portal, one via modified mid-anterior portal). Two cases of anchor-induced acetabular chondral deformation without frank penetration had successful clinical and radiographic outcomes at 2 and 6 years (Cases 1 and 2), while one case (Case 3) progressed from deformation to chondral penetration with clinical worsening. Of the cases that underwent revision hip arthroscopy for worsening symptoms, all three had confirmed exposed hard anchors which were removed using arthroscopic techniques. Two patients (Cases 3 and 4) have had clinical improvement reflected in increased patient-reported outcome scores and one patient underwent early total hip arthroplasty (Case 5). These findings are summarized in [Table I](#hnv001-T1){ref-type=\"table\"}. Table I.Study findingsPatientAge (years)Anchor typeGuidePortalLocation (O\\'Clock)Chondral damagePre-op PROSPost-op PROSPost-revision PROSPROM133HardStraightMAP2PT6796mHHS216HardStraightMAP1PT3489NAHS330HardStraightmMAP2PT--\\>FT361888NAHS440HardStraightAP1FT3185mHHS541HardStraightAP12, 2FT[^1]\n\nDISCUSSION\n==========\n\nTo our knowledge, this is the first series of this heretofore undocumented complication of hip arthroscopy. Anchor-induced chondral damage of the hip may occur in three potential forms: acetabular chondral deformation, chondral penetration with intra-articular implant exposure and loose anchor implant (whole or in pieces). Indeed, these forms may represent progressive stages. At least three of the cases (Cases 1, 2 and 3) had documented anchor-induced chondral deformation without penetration with resultant ballooning or elevation of acetabular cartilage implying violation of subchondral bone and focal chondral delamination. Of these, two cases demonstrated clinical improvement (Cases 1 and 2) without radiographic hip degeneration at short- and mid-term, while one case experienced clinical deterioration prompting revision arthroscopy with confirmed progression to chondral penetration (Case 3).\n\nRevision hip arthroscopy may be helpful in select patients. Three patients underwent revision hip arthroscopy including anchor removal for persistent or worsening symptoms; of two patients demonstrating clinical improvement, one retained normal radiographic joint appearance (Case 3) and the other had progressive partial medial joint narrowing (Case 4). The one patient who failed revision hip arthroscopy (Case 5) had two exposed anchors and major chondral damage. She ultimately improved with total hip arthroplasty albeit at a young age. It is unknown whether earlier anchor removal may prevent/delay joint deterioration.\n\nTreatment options, many learned from shoulder arthroscopy, depend on whether the offending anchor is hard or soft, threaded or barbed, fully or partially seated at the acetabular rim or loose and causing chondral deformation versus penetration. All errant anchors in this case series were non-threaded non-metallic hard anchors which were fully seated. It seems reasonable to remove an embedded errant anchor immediately in the case of gross chondral penetration or chondral deformation if surgically amenable to such treatment without causing excessive bone removal and/or further chondral damage. Hard anchors with threaded designs facilitate removal by reversing the rotatory direction of insertion. Partially seated barbed anchors may possibly be removed in a retrograde manner using a grasper that fits through an arthroscopic cannula. If the anchor can be disengaged in alignment with the grasper and cannula, it may be extracted through the retained cannula. An alternative is to use a slotted cannula, extracting the grasped anchor in unison with the slotted cannula. (This technique also enables arthroscopic removal of loose or foreign bodies too big to fit through conventional cannulas.) A trephine or 6-mm osteochondral autograft transfer system harvester may be tamped over the anchor and turned in a counterclockwise manner \\[[@hnv001-B28]\\]. This technique, however, may cause extensive osteochondral collateral damage and perhaps should be considered a last resort. Anterograde advancement of a penetrating anchor may permit its removal via the central compartment, perhaps causing less chondral damage. If the offending anchor is not removable, one may burr down the exposed surface however fixation strength may be compromised leading to loss of fixation and/or an intra-articular loose implant. If not overly proud, loose and/or impinging, one may leave the anchor as-is, perhaps modifying post-operative rehabilitation and monitoring (see below).\n\nArthroscopic removal of a fully seated hard anchor that penetrates the articular cartilage may be removed en bloc or piecemeal under hip distraction via the central compartment ([Table II](#hnv001-T2){ref-type=\"table\"}). En bloc removal may require limited debridement of immediately adjacent articular cartilage and the use of a small angled curette or microfracture awl as presented in Case 3. Once excavated from its osteochondral bed, the anchor may be removed with aforementioned arthroscopic techniques. Partial or complete piecemeal removal of a non-metallic hard anchor may be facilitated with a burr. In some cases, an unhooded or retractable sheath burr (Smith and Nephew, Andover, MA, USA) may aid visualization and arthroscopic resection. During burr resection, one may find that the anchor becomes disengaged so that it can be removed. A final option is to burr down the proud anchor to a stable subchondral position. Arthroscopic 'anchor-plasty' was attempted in Case 3, however clinical improvement was temporary. Arthroscopic lavage of the hip to remove debris that could cause synovitis or third body wear is then performed. If damage to the hip is extensive, conversion hip arthroplasty may be performed (Case 5). Table II.Treatment options for anchor-induced chondral penetrationArthroscopic reduction anchor-plasty (only if stable anchor)Burr resection of encroaching region of anchor to subchondral levelArthroscopic anchor removalEn Bloc\u2003Retrograde\u2003\u2003Unscrew threaded anchor\u2003\u2003Arthroscopic grasper\u2003\u2003Trephine over fixated anchor\u2003Anterograde\u2003\u2003Remove once loose in central compartmentPiecemeal (unhooded or retractable sheath burr) and lavageConsider microfracture base of defect\n\nWe have no experience with a detached anchor in the hip but agree with treatment recommendations in the shoulder. If the anchor is intra-articular, we agree with arthroscopic surgery for anchor removal and treatment of any associated chondrolabral pathology. If the anchor is in an intra-capsular but safe extra-articular location and appears stationary on serial MRIs, surgical removal may not be necessary. If not easily amenable to removal, retention of a non-penetrating errant anchor causing chondral elevation may be closely monitored, perhaps altering post-operative rehabilitation and monitoring ([Table III](#hnv001-T3){ref-type=\"table\"}). Until proven otherwise, it seems prudent to recommend treating patients with anchor-induced chondral elevation with prolonged restricted weightbearing, perhaps similar to protocols following microfracture chondroplasty, and to minimize future impact activities. It is also our opinion that relatively frequent post-operative clinic visits supplemented with interval radiographic assessment are merited. A low threshold for revision hip arthroscopy with timely anchor removal may minimize irreversible joint damage from third body wear. Table III.Treatment options for anchor-induced chondral deformationLeave as-is but do not advance furtherRemove if accessible and removal does not cause more damageClose post-operative monitoring (Clinical, radiographic, MRI)Modified post-operative rehabilitation\n\nIf errant anchor placement leads to chondral penetration with visualized exposure of the hard implant, we recommend immediate removal of that implant. If, however, chondral elevation without penetration is detected and the anchor is not readily removable, frequent clinical and radiographic monitoring, perhaps supplemented with magnetic resonance imaging of non-metallic anchors may be reasonable. If clinical worsening, especially if pain with non-impinging hip positions (e.g. extension), even minimal joint narrowing and/or anchor penetration is demonstrated, timely arthroscopic surgery with anchor removal is suggested prior to potentially irreversible osteoarthrosis or intra-articular loose body generation.\n\nSeveral technical pearls may help prevent the errant suture anchor and its potentially devastating consequences. A distal-based arthroscopic working portal can improve the trajectory of anchor placement by better matching the specific skeletal geometry of the acetabular rim \\[[@hnv001-B29]\\]. Compared with the mid-anterior portal, the modified mid-anterior and the distal anterolateral (DALA) portal is progressively more posterior \\[[@hnv001-B30]\\]. Although anchor trajectory improves with these portals, occasional obstruction by the femoral head occurs, especially with the DALA portal \\[[@hnv001-B27]\\]. The modified mid-anterior portal may be the best compromise. However, one case of anchor-induced chondral damage did occur using this portal so no portal may be infallible. Some surgeons may also use percutaneous accessory portals to improve the trajectory for suture anchor placement. Suggested preventative measures based upon our collective experience, preference and opinions are listed in [Tables IV](#hnv001-T4){ref-type=\"table\"} and [V](#hnv001-T5){ref-type=\"table\"}. Table IV.Suggested preventative techniquesDistal portal (e.g. MAP, MMAP, DALA)Clear rim of obstructing capsule and synoviumPilot hole (microfracture awl)Zone-specific drilling (see [Table V](#hnv001-T5){ref-type=\"table\"})Visualize from central compartment during drilling and anchor placementListen for change in drill soundFeel for increase drill resistance if engage subchondral boneExtra caution at 3 o'clock (direct anterior) rim and/or thin articular cartilageRim trimmingSoft (e.g. suture-based) anchorsSmall diameter anchorsShort anchorsIncreasing radius of curvature of curved anchor systems[^2] Table V.Zone-specific chondrophobic acetabular rim drilling[\\*](#hnv001-TF1){ref-type=\"table-fn\"}Anterior zoneDrill path parallel to floorSuperior zoneFluoroscopic divergent drill pathPosterior zoneDivergent drill path\\*\\*[^3]\n\nThe margin of error (i.e. safety angle) for anchor placement increases with rim trimming, smaller diameter, shorter or soft suture-based anchors and greater radius of curvature of curved anchor systems \\[[@hnv001-B31], [@hnv001-B32]\\]. Moreover, the direct anterior or 3 o'clock rim position had the least tolerance to errant drilling/anchor placement \\[[@hnv001-B32]\\]. This case series did not have any anchors placed at 3 o'clock; the 1 o'clock position was most commonly associated with this complication. If an anchor can be acceptably placed at 2 o'clock rather than 3 o'clock, especially after rim trimming in appropriate cases, risk of anchor-induced chondral damage may be lessened. Furthermore, in areas with thin articular cartilage, one may consider small and/or soft anchor options.\n\nAnother preventative measure is sufficient clearance of capsular and synovial tissues from the area of planned anchor fixation. Beyond using a shaver, a radiofrequency wand and/or burr in reverse spin may aid this process, permitting improved arthroscopic visualization for accurate drill hole and anchor placement close to but not in violation of the articular cartilage. A tip is to make a shallow divot with a microfracture awl at the desired rim starting point so that the subsequent drill tip does not wander from its intended starting point.\n\nRim preparation may be done with zone-specific chondrophobic drilling ([Table V](#hnv001-T5){ref-type=\"table\"}) \\[[@hnv001-B27], [@hnv001-B33]\\]. At the superolateral rim or 12 o'clock position, two-dimensional AP fluoroscopic spot imaging can confirm a drill path engaging bone while diverging from the articular cartilage. Desiring suture anchor placement close to but not in violation of the acetabular articular cartilage, the anterior drill hole(s) are made with the drill positioned parallel to the floor. If pelvic orientation has been standardized at the onset of surgery (e.g. using the fluoroscopic templating technique) \\[[@hnv001-B34]\\], anterior chondral damage may be avoided even at the narrow 3 o'clock position. Arthroscopic visualization of the adjacent acetabular cartilage from the central compartment (with hip distraction) while engaging tactile and auditory senses to detect even subtle changes during drill advancement (which might be an indicator of subchondral bone encroachment or penetration) encourages safe anterior suture anchor placement. For the posterior rim, the drill path is inherently chondroprotective when done via the MAP or MMAP; by nature of its posterior vector, drill trajectory diverges from the posterior chondral surface. However, the posterior rim may be particularly thin and posterior wall 'blow-out' with compromised anchor fixation may occur \\[[@hnv001-B27], [@hnv001-B33]\\].\n\nCurved or angled anchor guides, drills and inserters may aid safe anchor placement. A recent study has demonstrated improved divergence from the chondral surface compared with a straight guide \\[[@hnv001-B30]\\]. Interestingly, whereas curved systems may be utilized so as to diverge from the adjacent acetabular cartilage surface, they may be used in a convergent manner (aimed anteriorly) to avoid posterior wall blowout when performing posterior labral repair or reconstruction \\[[@hnv001-B27], [@hnv001-B33]\\].\n\nRecent studies on acetabular safe angles of anchor insertion have demonstrated a wider margin of safety for acetabular suture anchor placement with shorter drill depths, smaller drill diameters, rim trimming and curved guides \\[[@hnv001-B31], [@hnv001-B32]\\]. Hence, future technical and equipment developments may incorporate some or all of these findings.\n\nSoft suture-based deforming anchors offer an alternative to hard non-absorbable or bioabsorbable suture anchors. We have successfully used soft anchors for labral refixation and labral reconstruction. The small diameter (1.4 and 1.5\u2009mm) drills with short drill depths and short deployed implants near the peripheral rim permit anchor placement in desired proximity to the chondral surface while maximizing the margin of safety. Furthermore, the flexible nitinol wire permits drilling via curved guides. By aiming the curved guide away from the acetabular articular cartilage during rim preparation, suture anchor fixation may occur in desired proximity to the osteochondral junction while diverging from the cartilage for safe anchor placement \\[[@hnv001-B27], [@hnv001-B30]\\]. If this anchor detaches or violates the joint, the likelihood of degenerative damage to the femoral head may be less because of the deforming nature of these small soft implants. Finally, computer-assisted navigation may also play a future role in safe drilling and anchor placement.\n\nTwo of the five cases did not have confirmatory documentation of anchor-induced chondral damage (Cases 4 and 5). This may have been from non-recognition or non-reporting. Because of the potential for severe degenerative consequences from this complication, and because modified post-operative rehabilitation and close monitoring may be beneficial, we strongly encourage surgeons to be cognizant of this complication, employ aforementioned techniques to minimize its occurrence, remove offensive anchors when prudent, accurately document and inform the patient.\n\nLimitations\n-----------\n\nLimitations of this case series include the lack of a control group and the inability to apply valid statistical analysis given the small number of patients and inconsistent hip score instruments. Although two patients with chondral deformation are doing well at 2 and 6 years, a larger long-term controlled study is needed to determine the durability of clinical improvement and to substantiate conclusions in this subtype.\n\nCONCLUSION\n==========\n\nAnchor-induced chondral deformation without frank chondral penetration may be treated with close clinical and radiographic monitoring with a low threshold for revision surgery and anchor removal. Chondral penetration should be treated with immediate removal of offending hard anchor implants. Preventative measures include distal-based portals, small diameter and short anchors, removable hard anchors, soft suture-based anchors, curved drill and anchor insertion instrumentation and attention to safe trajectories while visualizing the acetabular articular surface.\n\n[^1]: MAP\u2009=\u2009mid-anterior portal, mMAP\u2009=\u2009modified mid-anterior portal, AP\u2009=\u2009anterior portal, PT\u2009=\u2009partial thickness, FT\u2009=\u2009full thickness, PROS\u2009=\u2009patient-measured outcome score, PROM\u2009=\u2009patient measured outcome measure, mHHS\u2009=\u2009modified Harris Hip Score, NAHS\u2009=\u2009Non-arthritic Hip Score.\n\n[^2]: MAP, mid-anterior portal, MMAP, modified mid-anterior portal, DALA, distal anterolateral accessory.\n\n[^3]: \\*via mid-anterior or modified mid-anterior portal. \\*\\*The posterior-most drill site is the most challenging and should be done before the final drill site. If done as the final drill site, anchor spacing may dictate that far-posterior drilling is done in a region most susceptible to posterior wall fracture with compromised labral refixation.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nCancer is a compendium of perturbed genome functions and is characterized by the deregulation of several genes and their regulatory molecules, including microRNAs (miRNAs) \\[[@R1], [@R2]\\]. In general, miRNAs are 19--24 nucleotides long, noncoding RNA molecules that regulate the expression of 30% of protein-coding genes at the posttranscriptional level \\[[@R3], [@R4]\\]. These miRNAs are transcribed by RNA polymerase II as pre-miRNAs, which are processed by Drosha, to form hairpin-like intermediates called pre-miRNAs, which are approximately 70--100 nucleotides long and have two nucleotide overhangs at their 3\u2032 ends \\[[@R5]--[@R8]\\]. Exportin-5 transports pre-miRNAs out of the nucleus and into the cytoplasm for further processing by dicer \\[[@R9], [@R10]\\], which converts the pre-miRNA into a 19--24 nucleotide long, double-stranded, mature miRNA \\[[@R11]\\]. The guide strand of mature miRNA gets incorporated into a complex to form the RNA-induced silencing complex (RISC), which recognizes the specific target mRNA through complementary base pairing, and consequently, either degrades or inhibits translation of the target mRNA \\[[@R6], [@R12]\\]. Recent reports have shown that miRNAs can play an important role in diverse biological functions such as development, normal cell physiology and pathological conditions like abnormal cell proliferation \\[[@R13]--[@R15]\\] and cancer \\[[@R16], [@R17]\\] including pancreatic cancer (PC).\n\nPancreatic cancer (PC) is the 10^th^ leading cause of cancer-related deaths worldwide and the fourth-leading cause of cancer-related deaths in the United States \\[[@R18]--[@R21]\\]. PC has a five-year survival rate of less than 6% \\[[@R20]\\]. This high mortality rate among PC patients is due to the lack of early symptoms, diagnostic and prognostic markers, metastatic disease at the time of clinical diagnosis, poor response to therapy, and a high rate of recurrence \\[[@R22]\\]. Thus, in order to identify suitable early diagnostic markers or therapeutic targets to combat this disease, there is an urgent need to understand the pathogenesis of PC. Several studies have shown the abnormal expression of miRNAs including miR-21, Let-7b, miR-100, miR-217, and miR-216 in PC and have proposed them as candidates for early diagnosis and potential molecular targets \\[[@R23], [@R24]\\]. However, due to unavailability of early stage PC biopsies from humans, the stage-specific deregulation of miRNA during PC progression is largely unknown.\n\nIn this study, we investigated the expression profile of miRNAs at various stages of PC progression using the Kras^G12D^;Pdx1-Cre mouse model. This model is histopathologically recapitulates human PC \\[[@R25]\\] and enables one to monitor the progression of PC from the onset of precursor lesions to cancer development. Hence, understanding the role of miRNA in PC pathogenesis using this model will subsequently help to develop early detection markers, as well as therapeutic and preventive strategies for PC. Further, this study examined the expression of various components of the miRNA biosynthetic machinery and some of their selected target genes for differentially expressed miRNAs. Findings were extrapolated to human PC tissues and cell lines by investigating the expression of differentially expressed miRNAs as observed in the Kras^G12D^ mouse model. The overall objective of this study was to establish a global differential expression profile of miRNAs during the course of initiation and progression of PC. Moving forward, the identified miRNAs can be used to create novel biomarker(s) for validation studies and therapeutic targets to combat this lethal disease.\n\nRESULTS {#s2}\n=======\n\nPancreatic cancer progression model {#s2_1}\n-----------------------------------\n\nOur previous study using the Kras^G12D^;Pdx1-Cre (KC/floxed Kras^G12D^) PC mouse model showed the presence of pancreatic intraepithelial neoplasia (PanIN) lesions as early as 10 weeks of age; these lesions progressed to pancreatic ductal adenocarcinoma (PDAC) and metastasized to the liver, lungs, and intestines by 50 weeks of age \\[[@R26]\\]. We observed PanIN-I lesions at 10 weeks of age that progressed to PanIN I, II, III lesions at 25 weeks of age, replacing larger portions of the pancreatic parenchyma \\[[@R26]\\]. At 40 weeks of age, the majority of the parenchyma was replaced by PanIN III and extensive desmoplasia, whereas at 50 weeks of age, animals replaced the pancreatic parenchyma with full-blown PDAC, and metastatic spread involving the colon, liver, and lungs in 60--70% of the animals \\[[@R26]\\].\n\nmiRNA microarray {#s2_2}\n----------------\n\nIn order to identify the deregulated miRNAs during PC progression, we used affymetrix miRNA microarray to analyze the global miRNA expression in the pancreas of 25-week old KC animals (Kras^G12D^;Pdx1-Cre) and its age-matched littermate control mice (LSL Kras^G12D^). We chose 25-week old mice as the base line because, at this time point, the mice had only developed PanIN I, II, and III lesions but not PDAC. Hence, this age represented the stage prior to the onset of PC.\n\nDifferentially expressed miRNAs were identified using QC tool software. The miRNA microarray analysis for the KC animals compared to controls revealed that miR-150, miR-494, miR-138, miR-148a\\*, miR-216a, and miR-217 (*p-value* = 0.01) were significantly downregulated (Table [1](#T1){ref-type=\"table\"}), whereas, miR-146b, miR-205, miR-31, miR-192, and miR-21 (*p-value* = 0.01) were significantly upregulated (Table [2](#T2){ref-type=\"table\"}). A majority of the miRNAs were downregulated compared to the number of miRNAs that were upregulated in the KC animals ([Supplementary table S1](#SD2){ref-type=\"supplementary-material\"}). The panel of differentially expressed miRNAs were validated by real-time PCR using TaqMan assays, and the results were consistent with the miRNA microarray data that showed up-regulation of miR-21, miR-221, miR-100 and miR-26a and down-regulation of miR-26b, miR-141, miR-96, miR483-3p, miR-216, and miR-217 in the KC compared to control mice (Figure [1A](#F1){ref-type=\"fig\"}).\n\n###### Top/Significantly downregulated miRNAs in Kras^G12D;^Pdx1-Cre mice\n\n  S.No   microRNA       Fold change     predicted Target genes\n  ------ -------------- --------------- -----------------------------------------------------------------------------------------------------\n  1      miR-150        0.149           c-Myb, Eif4b&e, Ephb2, Elk1, Mcc\n  2      miR-494        0.293           Fgfr2, Cdk6, Nfat5, Ccnt2, IGF1R, Fgf7, Ccnd2, Socs6, Bmpr2,\n  3      miR-138        0.301           Sin3a, Rmnd5a, Thrb, Dek, Rhoc, Nfib, Hif1a, Tjp1\n  4      miR-148a\\*     0.351           Bcl2l11, fbn1, Itga5,9&11, Dgcr8, Snn, Dicer1, E2f3&7, Tgfbr1, wnt1, Itga1\n  5      miR-193        0.351           Kit, Tgfb2, Ets1, Etv6, Tgfbr3, Kras, Fgf1, Tcf4.\n  6      miR-451        0.365           Psmb8, Mex3c, Cab39, Fbxo33, Gpr77\n  7      miR-216a & b   0.406 & 0.447   Jnk2, Esr1, Sp4, Mmp16, Smad7, Tssc1, Tgfbr2, Sox6, Grb2,\n  8      miR-29b        0.409           Eln, Col3a1, Col4a5, Col5a1, Has3, Igf1, Dnmt3a, Camk4, Hdac4, Vcl, Ccnd2, Nkrf, Dicer1, Sp1, Ncoa3\n  9      miR-375        0.550           Max protein (Max)\n  10     miR-217        0.548           Kras, Esr1, Etv6, Dek,\n\n###### Top/Significantly up regulated miRNAs in Kras^G12D;^Pdx1-Cre mice\n\n  S.No   microRNA   Fold change   Some of predicted Target genes\n  ------ ---------- ------------- --------------------------------------------------------------------------------\n  1      miR-146b   18.46         MMP16,\n  2      miR-205    9.89          Cdh11, Cdkn1b, E2F1\n  3      miR-31     9.50          St7, Pdcd11, E2f2, Ret, Dicer1, Pcdha4-12,\n  4      miR-192    8.35          Cdc7, Ercc3, Pim1, Mcm10, Hoxa10, Mad2L1, PRPF38A, Racgap1, and Smarcb1\n  5      miR-194    7.11          IGF1R, Stat1, ITGA1, Sox11, Lats1\n  6      miR-21     5.76          Tgfbr2, Tgfbi, Sox2, Sox5, Sox7, PTEN, TPM1, PDCD4, Maspin, Rasa1&2, Cstf3\n  7      miR-379    5.37          Insr, Igf1r, Gdf6, Eif4g2, Edn1, Nfat5\n  8      miR-214    4.13          Arhgap28, Gnao1, Nr2c1\n  9      miR-541    4.57          Gab1, Dcc, Braf1, Tgfbr3\n  10     miR-199b   3.41          Itga1,3&8, fgf7,10&16, Rb1, Ppp2r2a, Ela1, Ppp2r5e, Fn1, Sp1, Met, Igf1, Zab1,\n\n![The miRNA array revealed several upregulated and downregulated miRNAs in Kras^G12D^; Pdx-1-Cre mice at 25 weeks of age\\\n**A.** Validation of a few upregulated and downregulated miRNAs was confirmed by real-time PCR analysis, using TaqMan assays specific for individual miRNAs. U6 snRNA was used as an internal control for normalization. **B.** Real-time PCR analysis of miRNA biosynthetic pathways in KrasG12D;Pdx1-Cre and LSL- Kras^G12D^ animals at 25 weeks of age using specific primers. Mouse \u03b2-actin was used as an internal control for normalization. The fold change was calculated using the \u0394\u0394^Ct^ method.](oncotarget-06-40295-g001){#F1}\n\nThe downregulation of the majority of miRNAs led us to investigate the expression status of the miRNA biosynthetic machinery during the progression of PC using the KC mouse model. Real-time PCR analysis revealed that expression of Dicer, Drosha, Exportin-5, TRKRA, and TARBP2 were downregulated, while expression of DGCR8 and Ago2 were upregulated in KC mice compared to control littermates (Figure [1B](#F1){ref-type=\"fig\"}).\n\nThe top differentially deregulated miRNAs were further analyzed at 10, 30, 40, and 50 weeks of PC progression. At 10 weeks of age, expression of miR-141 and Let-7b were upregulated, but their expression was not statistically significant. On the other hand, miR-146b, miR-34c, miR-223, miR-195 (*p-value* = 0.031) and miR-216 (*p-value* = 0.063) were downregulated in KC mice compared to control littermates. However, no significant difference was observed in the expression of pancreas-specific miR-217 (Figure [2A](#F2){ref-type=\"fig\"}). At 30 weeks of age, the expression of miR-216 (*p-value* = 0.016), miR-217 (*p-value* = 0.0078), miR-150 (*p-value* =0.023), Let-7b (*p-value* = 0.031,) and miR-96 were significantly downregulated, whereas the expression of miR-146b (*p-value* = 0.0078), miR-205, (*p-value* - 0.0078), miR-21, miR-195 (*p-value* = 0.031), and miR-34c (*p-value* = 0.063) were significantly upregulated in KC animals compared to control animals (Figure [2B](#F2){ref-type=\"fig\"}). At 40 weeks of age, the expression of miR-216, miR-217, miR-223, miR-141, miR-483-3p (*p-value* = 0.031), miR-195, Let-7b (*p-value* = 0.063) and miR-96 were significantly downregulated; on the other hand, the expression of miR-21, miR-205, miR-146b (*p-value* = 0.031), and miR-34c (*p-value* = 0.063) were upregulated in KC mice compared to the control animals (Figure [2C](#F2){ref-type=\"fig\"}). Further, at 50 weeks of age, the expression of miR-216, miR-217, miR-345, miR-141, miR-483-3p, miR-26b, miR-96, Let-7b (*p-value* = 0.01), miR-100, miR-26a and miR-150 (*p-value* = 0.094) were further downregulated in KC animals compared to control mice (Figure [2D](#F2){ref-type=\"fig\"}). The expressions of miR-216 and miR-217 were also progressively reduced in KC mice, but the expressions of miR-21, miR-205, miR-146b, miR-34c, and miR-223 progressively increased (Figure [1A](#F1){ref-type=\"fig\"}, [2A](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"}). At 50 weeks of age, variation in expression of miR-221 was not statistically significant between the KC and control animals (Figure [2D](#F2){ref-type=\"fig\"}). The overall trends of miRNA expression during the mouse PC progression model are shown in Figures [2E and 2F](#F2){ref-type=\"fig\"}.\n\n![Expression profiles of miRNAs during the mouse PC progression in KrasG12D; Pdx1-Cre mice and their contemporary littermate animals\\\n**A, B, C, D.** Expression patterns of miRNA at 10, 30, 40, and 50 weeks of PC progression was analyzed by real-time PCR, using TaqMan assays specific for individual miRNAs. U6 snRNA was used as an internal control for normalization. The fold change was calculated by \u0394\u0394^Ct^ method. **E.** and **F.** Showing trends of miRNA expression during the mouse PC progression.](oncotarget-06-40295-g002){#F2}\n\nValidation of miRNA target genes {#s2_3}\n--------------------------------\n\nIn order to confirm the expression of genes targeted by the differentially expressed miRNAs in PC, first we performed *in silico* analysis using an online bioinformatics program, mirecords/miRDB, which utilizes miRanda, Target scan, Pic Tar, *etc*., to predict several potential targets genes (Table [1](#T1){ref-type=\"table\"} and Table [2](#T2){ref-type=\"table\"}). A few of these target genes, that are important in PC development, were analyzed by Real time PCR at 50 weeks of age. The expression of Kras (5-fold), Galectin-3 (5-fold), EGFR (3.5-fold), TGF\u03b21 (3.5-fold), cMyc (2.8-fold), Fut8 (2.5-fold), and Bcl2 (2-fold) were significantly upregulated, while the expression of PTEN was significantly downregulated in KC animals compared to controls (Figure [2G](#F2){ref-type=\"fig\"}).\n\nExpression patterns of miRNAs in human PC cell lines {#s2_4}\n----------------------------------------------------\n\nIn order to extrapolate the mouse miRNA microarray data to human PC, we next checked the expressions of a few differentially expressed miRNAs in eight human PC cell lines, as well as in the immortalized normal pancreatic ductal cell line (HPDE). Results showed the downregulation of miR-345, miR-96, and Let-7b in the majority of the PC cell lines compared to the HPDE cells (Figure [3A](#F3){ref-type=\"fig\"}), indicating that these miRNAs were expressed in ductal cells rather than in tumor stroma. Also, the expression analysis of miRNA biosynthetic machinery in PC cell lines showed significant downregulation of Drosha, Dicer, and Exportin-5 compared to the HPDE cell line (Figure [3B](#F3){ref-type=\"fig\"}).\n\n![Expression patterns of miRNAs and components of the biosynthetic machinery in human PC tissues and cell lines\\\n**A.** Expression profiles of miRNA in human PC cell lines using TaqMan assays specific for individual miRNA based on real-time PCR analysis. RNU6B was used as an internal control for normalization. The fold change was calculated by using the \u0394\u0394^Ct^ method. **B.** Components of the miRNA biosynthetic pathway in human PC cell lines by using specific primers for each gene by real-time PCR analysis. Human \u03b2-actin was used as an internal control for normalization. The fold change was calculated by using the \u0394\u0394^Ct^ method. **C.** H & E staining of cancer-adjacent normal tissue. **D.** PC tissue and **E.** expression patterns of miRNA in human PC tissue and adjacent normal tissue using real-time PCR using TaqMan assays specific for individual miRNA. RNU6B was used as an internal control for normalization. The fold change was calculated by using the \u0394\u0394^Ct^ method. **F.** Components of the miRNA biosynthetic pathway in human PC tissue and adjacent normal tissue were determined using specific primers to each gene by real-time PCR analysis. Human \u03b2-actin was used as an internal control for normalization. The fold change was calculated by \u0394\u0394^Ct^ method.](oncotarget-06-40295-g003){#F3}\n\nExpression patterns of miRNAs in human PC tissue {#s2_5}\n------------------------------------------------\n\nThe differentially expressed miRNAs that were identified from mouse miRNA array data were further analyzed in eight samples of human PC tissues (Figure [3C](#F3){ref-type=\"fig\"}) and their corresponding eight adjacent normal tissue (Figure [3D](#F3){ref-type=\"fig\"}). The expression of miR-223, miR-483-3p (*p-value* = 0.01), 146b, 205 (*p-value* = 0.001), 221, 21 (*p-value* = 0.023), 195, 34c and miR-26a (*p-value* = 0.0078) were significantly upregulated, whereas the expression of miR-216, miR-141, miR-217, Let-7b (*p-value* = 0.001), and Let-150 (*p-value* = 0.01) were significantly downregulated in human PC tissues as compared to the cancer-adjacent normal tissues (Figure [3E](#F3){ref-type=\"fig\"}). Additionally, the expression of proteins involved in miRNA biosynthesis, such as Drosha, Dicer, and Exportin-5, were analyzed and found to be significantly downregulated in human PC tissues as compared to the cancer-adjacent normal tissues (Figure [3F](#F3){ref-type=\"fig\"}).\n\nEctopic overexpression of Let-7b in human PC cell lines {#s2_6}\n-------------------------------------------------------\n\nThe role of Let-7 family members and their target genes are well known in various cancer \\[[@R27]\\]. Previous studies have shown that restoration of Let-7 results in downregulation of oncogenic Kras, leading to inhibition of cell proliferation and activation of the mitogen-activated protein kinase \\[[@R28]\\]. Similarly, HMGA2 is a direct target of Let-7 family members \\[[@R29]--[@R31]\\]. Our results in the KC mouse model also showed downregulation of Let-7b during PC progression. Further, our *in silico* analysis of Let-7b revealed several target genes like *MUC4, NCOA3, KRAS, HMGA2,* which are critical in PC pathogenesis. Herein, we overexpressed Let-7b and its scramble vectors in two PC cell lines, CD18/HPAF and Capan1, by infection with a lenti-viral supernatant (collected after 48 and 72 hours post transfection of 293FT cells) after mixing with 4 \u03bcg/ml of polybrene and analyzing its effect on downstream targets. The green fluorescent protein (GFP)-positive cells were sorted by Fluorescence Activated Cell Sorting (FACS) (Figure [4A](#F4){ref-type=\"fig\"}); Let-7b overexpression was confirmed by TaqMan assays using real-time PCR (Figure [4B](#F4){ref-type=\"fig\"}). We observed eight-fold and four-fold increase in the expression of Let-7b in Capan 1 and CD18/HPAF cells, respectively, compared to vector control cells (Figure [4B](#F4){ref-type=\"fig\"}). Our results from western blot analysis revealed downregulation of MUC4, KRAS, MSST1, and Cyclin D1, and upregulation of caspase-9 in Let-7b overexpressing CD18/HPAF and Capan1 PC cells compared to vector transfected control cells (Figure [4C](#F4){ref-type=\"fig\"}); \u03b2-actin was used as a loading control. These findings suggest that downregulation of Let-7b plays an important role in PC progression.\n\n![Overexpression of Let-7b in human PC cell lines\\\n**A.** PC cells infected with lentivirus carrying Let-7b miRNA/control, along with GFP as a selection marker, revealed 100% PC cells expressing GFP, indicating overexpression of Let-7b/control vector in PC cell lines. **B.** Validation of Let-7b overexpression in GFP-expressing PC cells was analyzed using the TaqMan assay by real-time PCR revealed significant upregulation of Let-7b in human PC cell lines compared to control vector transfected PC cells. RNU6B was used as an internal control for normalization. The fold change was calculated by using the \u0394\u0394Ct method. **C.** Immunoblotting of protein lysates collected from PC cell lines ectopically overexpressing Let-7b and its control cells revealed downregulation of Kras, MSST1, MUC4 and Cyclin D1, and upregulation of cleaved caspase-9 (target genes predicted by *in silico* analysis). \u03b2-actin was used as a loading control.](oncotarget-06-40295-g004){#F4}\n\nDISCUSSION {#s3}\n==========\n\nDue to unavailability of early stage samples from human PC patients, we utilized the well-characterized spontaneous mouse model (KC) of PC \\[[@R32]--[@R36]\\] to analyze the global miRNA expression profile. Notably, this model histopathologically recapitulates human PC. Using this PC model, we and others have observed PanIN lesions at as early as 10 weeks of age, as well as PDAC with metastasis (in 60--70% of KC animals) to distant organs by 50 weeks of age \\[[@R26], [@R35]\\]. Although high-grade PanINs (PanIN III) were rarely observed before 20 weeks of age, their frequency increased progressively with age (25--50 weeks); however, low- and high-grade PanINs were observed starting at 25 weeks of age. Therefore, the present study analyzed the global miRNA expression profile at 25 weeks of age, when the majority of acinar cells were replaced by high-grade PanINs and marked desmoplasia \\[[@R26]\\].\n\nUsing miRNA microarray analysis, we identified several differentially expressed miRNAs in KC animals compared to control mice, with most of the miRNAs having tumor suppressive functions (Tables [1](#T1){ref-type=\"table\"} and [2](#T2){ref-type=\"table\"}). The microRNA microarray revealed significant downregulation of miR-494 at 25 weeks of PC progression in the KC mouse model. A recent study has also shown significant downregulation of miR-494 in human PC tissues compared to non-tumor tissues, correlating with PC metastasis and decrease free survival in patients. Furthermore, decreased expression of miR-494 is due to loss of SMAD4 in PC. Ectopic expression of miR-494 in PDAC cells resulted in downregulation of FOXM1, inhibition of \u03b2-catenin nuclear translocation, decreased cell proliferation, migration, invasion, and greater sensitivity to gemcitabine \\[[@R37]\\]. The expression of pancreas-specific tumor suppressors miR-217 and miR-216 were unaltered at 10 weeks of age (presence of PanIN-Ia and Ib), but progressively decreased from 25 -- 50 weeks of age as PanIN lesions progressed to PDAC. (Figure [1A](#F1){ref-type=\"fig\"}, [2A](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"}).\n\nIn addition to KC mice, we also observed a significant downregulation of miR-216 and miR-217 in human PC tissue (Figure [3E](#F3){ref-type=\"fig\"}); these results are in agreement with earlier studies on human PC \\[[@R38]--[@R43]\\] that show downregulation of miR-217 in 76.2% (16/21) of PC tissue as well as cell lines \\[[@R43]\\]. Similarly, the expression of miR-216 is significantly downregulated in PC \\[[@R44]\\]. Our *in silico* analysis revealed that miR-216 and miR-217 potentially target many important genes that play critical roles during the pathogenesis of PC (Table [1](#T1){ref-type=\"table\"}); and the downregulation of miR-217 \\[[@R45]\\] and miR-216 \\[[@R44]\\] suggests their potential as tumor suppressors in PC by targeting downstream targets, particularly the Kras oncogene \\[[@R43]\\] and Janus kinase 2 \\[[@R44]\\]. In addition to pancreas-specific miRNAs, we observed significant downregulation of miR-141 and Let-7b in human PDAC and in KC mice from 25--50 weeks of age compared to controls (Figure [1A](#F1){ref-type=\"fig\"}, [2B](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"}, [3E](#F3){ref-type=\"fig\"}). Our human PDAC results were in agreement with earlier reports showing loss of Let-7b and miR-141 expression in PC samples compared to cancer-adjacent normal tissue \\[[@R46], [@R47]\\]. *In vitro* analysis of Let-7b overexpressed HPAF/CD18, and Capan1 PC cell lines showed significant downregulation of MUC4, Kras, MSST1, and Cyclin D1, and upregulation of cleaved caspase-9 (Figure [4C](#F4){ref-type=\"fig\"}). Although the functional analysis of this Let-7b overexpression remains to be determined using our *in vitro* system, however previous studies have shown that overexpression of Let-7b in Capan1 cells resulted in decreased cell proliferation by downregulating Kras, Myc, and HMGA2, and activating mitogen-activated protein kinase \\[[@R28]--[@R31], [@R46], [@R47]\\]. In addition to affecting proliferation, the increased expression of cleaved caspase-9 in both HPAF/CD18 and Capan1 cells suggests it has a pro-apoptotic role that remains to be determined. Further, Kent *et al*. (2009) reported downregulation of miRNA-200 family members, including miRNA-141, in PC cell lines \\[[@R48]\\] and observed a positive correlation between the expression of miR-200 family members and E-cadherin expression, as well as a negative correlation with the zinc-finger E-box-binding homeobox 1 (ZEB1) \\[[@R48], [@R49]\\]. The transcription of miR-200 family members (miR-141 and miR-200c) is suppressed by ZEB1, which activates epithelial differentiation in PC, colorectal, and breast cancer cells. Epithelial-mesenchymal transformation (EMT) activators, such as transforming growth factor beta 2 and ZEB1, have 3\u2032 UTR binding sites for miR-200 family members. Therefore, existence of a feedback loop between ZEB1 and miR-141 serves to stabilize the EMT process and regulate the invasion of cancer cells \\[[@R49]\\].\n\nThe present study also revealed significant downregulation of miR-150 during the progression of PC in both mice and humans (Figure [2A](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"}, [3E](#F3){ref-type=\"fig\"}). These results contradict a previous study on human PC \\[[@R40]\\], possibly attributable to species and/or PC stage variation. Similarly, the expression of miR-345 and miR-96 were significantly downregulated during the progression of PC in the mouse model (Figure [2D](#F2){ref-type=\"fig\"}), which is consistent with previous reports \\[[@R39], [@R40], [@R50]\\]. However, a recent report in colorectal cancer patient tissues has shown significantly methylated mir-345 promoter region compared to the non-cancerous, adenoma, and normal colon tissues; further, ectopic over-expression of mir-345 in colorectal cancer cells significantly reduced cell proliferation by inhibiting the translation of anti-apoptotic BAG3 gene \\[[@R51]\\]. These studies suggest a pro-apoptotic function of miR-345; its downregulation during PC progression may be associated with promoter methylation.\n\nThe miR-26a miRNA inhibits the expression of c-myc, Cyclin D3 and E2, and cyclin-dependent kinases such as CDK4 and CDK6. Also, it stimulates CDK-inhibitors p14(ARF) and p21(CIP1) expression in an EZH2-dependent manner, suggesting an anti-proliferative role \\[[@R52]\\]. Accordingly, we observed significant downregulation of miR-26a in KC mice at 50 weeks of age, although it was upregulated at 25 weeks in KC mice and in human PC tissues compared to normal tissues (Figure [2B, 2D](#F2){ref-type=\"fig\"} and [3E](#F3){ref-type=\"fig\"}). Recently, significant downregulation of miR-26a was also reported in nasopharyngeal carcinoma tissues and cell lines (53); its ectopic expression has been shown to downregulate EZH2 expression, thereby inducing G~1~-phase cell-cycle arrest to decrease cell proliferation and colony formation (53). A decreased expression of miR-483-3p was observed during the progression of mouse PC, whereas it is overexpressed in human PC (Figure [2A](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"} and [3E](#F3){ref-type=\"fig\"}). These results are in agreement with previous reports that show miR-483-3p upregulation in PC tissue compared to the cancer adjacent normal tissue \\[[@R53], [@R54]\\]. Furthermore, they showed that ectopic expression of miR-483-3p significantly inhibited the expression of DPC4/Smad4 protein in PC cell lines, leading to increased cell proliferation and colony formation *in vitro* \\[[@R53], [@R54]\\]. The miRNA microarray also revealed decreased expression of miR-148a and miR-451 in mouse PC, an observation consistent with other reports \\[[@R40], [@R41], [@R55]--[@R58]\\]. Similarly, another study showed that the expression of miR-148a/b and miR-375 were significantly downregulated in the PC developed from p48Cre;Kras^G12D^ transgenic animals compared to the normal controls \\[[@R59]\\].\n\nSeveral studies have demonstrated the upregulation and importance of miR-21 in PC patients. Although the expression of miR-21 expression does not correlate with tumor size, differentiation, nodal status, or T stage, its overexpression was associated with poor patient outcomes and overall survival compared to patients without miR-21 or patients with faint/focal expression in node-negative disease \\[[@R60], [@R61]\\]. Overexpression of miR-21 is associated with both high Ki67 proliferation index and liver metastasis \\[[@R62]\\], whereas inhibition of miR-21 expression in MiaPaCa-2 PC cells resulted in decreased cell proliferation, increased cell death, and increased expression of the tumor suppressor gene like programmed cell death 4 (PCD4) gene \\[[@R63]\\]. In addition, miR-21 has been shown to suppress the activity of the tumor suppressor gene tropomyosin 1 (TPM1), thereby promoting tumorigenesis \\[[@R64]\\]. In line with these reports, our study also indicates the progressive increase in miR-21 and miR-205 from 10 to 50 weeks of age in KC mice compared to the control animals (Figure [1A](#F1){ref-type=\"fig\"} and [2A--2D](#F2){ref-type=\"fig\"}). Similarly, we also observed upregulation of miR-21 and miR-205 in human PC samples compared to cancer- adjacent normal tissue (Figure [3E](#F3){ref-type=\"fig\"}). Several reports have also shown elevated expression of miR-21 and miR-205 in a panel of PC cell lines and tissues compared to the normal controls \\[[@R40]--[@R42], [@R50], [@R55], [@R60], [@R65]--[@R68]\\]. Rieu *et al*. also reported in the KC animals, progressive increases in miR-21 and miR-205 expression from PanIN lesions to full-blown PDAC, with strongest expression of miR-21 in precursor lesions with phenotypic changes in the ducts \\[[@R69]\\]. Further, Rieu *et al*. have shown that expression of miR-21, miR-221, miR-222, and Let-7a increases with PanIN grade, with the strongest expression in the hyperplastic ducts with PanIN 2/3 lesions \\[[@R69]\\]. Similarly, the expression of miR-10, miR-21, miR-100 and miR-155 was shown to increase in p48-Cre/Kras^G12D^ mice when compared to control animals \\[[@R59]\\]. We observed downregulation of miR-146b, miR-34c, and miR-223 at 10 weeks of age; however, their expression increased with the progression of PC in KC animals compared to control animals (Figure [2A](#F2){ref-type=\"fig\"}--[2D](#F2){ref-type=\"fig\"}). We also observed upregulation of these miRNAs in the human PC samples (Figure [3E](#F3){ref-type=\"fig\"}). Notably, the expression of miR-34c is activated by p53 following DNA damage and serves as an important regulator of c-Myc expression, acting downstream to the p38 MAPK/MK2 pathway \\[[@R70]\\].\n\nDeregulation of the miRNA biosynthetic machinery results in differential expression of miRNAs and downstream target genes. Abnormal expression of this machinery is associated with hematological malignancies \\[[@R71]\\], ovarian cancer \\[[@R72]\\], and neuroblastoma \\[[@R73]\\]. In conjunction with these reports, our results also revealed significant downregulation of many miRNA biosynthetic pathway components, such as dicer, TARBP2, and exportin-5, both during the progression of PC in KC animals (Figure [1B](#F1){ref-type=\"fig\"}) and in PC cell lines and tissues compared to normal tissues (Figure [3B and 3F](#F3){ref-type=\"fig\"}). The predicted target genes for differentially expressed miRNAs (Table [1](#T1){ref-type=\"table\"} and Table [2](#T2){ref-type=\"table\"}) such as Kras (5-fold), Galectin-3 (5-fold), EGFR (3.5-fold), TGF\u03b21 (3.5-fold), cMyc (2.8-fold), Fut8 (2.5-fold), and Bcl2 (2-fold) were upregulated. On the other hand, PTEN was observed to be downregulated in KC mice, which highlights the importance of deregulated miRNA expression in the pathology of PC (Figure [2G](#F2){ref-type=\"fig\"}). Both miR-216 and miR-217 act as potential tumor suppressors for PC by targeting the Kras oncogene \\[[@R43]\\].\n\nIn conclusion, the present study reveals the stage-specific (PanIN to PC) unique expression patterns for miRNAs, as well as miRNA biosynthetic machinery, during progression of PC. We have shown that in tumor samples compared to normal samples, the majority of miRNAs (miR-216, miR-217, miR-100, miR-345, miR-141, miR-483-3p, miR-26b, miR-150, Let-7b, Let-195 and miR-96) were downregulated, and few were upregulated (miR-146b, miR-205, miR-31, miR-192, miR-194 21, miR-379, miR-431, miR-541, and miR-199b). Future studies will focus on the analysis of stage-specific miRNA expression in the sera of mice collected during PC progression, as this is expected to provide the unique opportunity to develop early detection biomarkers for PC. Furthermore, restoring the expression of downregulated miRNAs in cancer cells may be performed to promote their de-differentiation and induce their reversion into a benign or normal cell. Overall, the current study offers the unique opportunity to develop early biomarkers for the diagnostic, prognostic and preventive management of PC.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nExperimental animals {#s4_1}\n--------------------\n\nThe B6.129-*Kras^tm4Tyj^* (01XJ6) and B6. FVB-Tg (Ipf1-cre) 1Tuv (01XL5) mice were obtained from the National Cancer Institute (NCI) Mouse Models of Human Cancers Consortium (MMHCC, Frederick, MD, USA). In order to decipher miRNAs in early PC lesions, we obtained the PC progression model by crossing LSL-Kras^G12D^ animals \\[[@R74]\\] with Pdx-1-Cre \\[[@R35]\\]. The animals positive for both Kras and Pdx Cre (Floxed Kras^G12D^) and their contemporary littermates unfloxed Kras (LSL Kras^G12D^) from the F1 litter were sacrificed at 10, 25, 30, 40, and 50 weeks of age (eight animals for each group at each time point). The pancreas was resected, and half of the pancreas was flash frozen in liquid nitrogen (total RNA and miRNA isolation); the other half was fixed in 10% buffered formalin and embedded in paraffin. Serial tissue sections were cut (4 \u03bcM) and stained with hematoxalin and eosin. Throughout the study, the animals were subjected to a 12 hours of dark/light cycle with food and water *ad libitum*. Animal studies were performed in accordance with the United States Public Health Service \"Guidelines for the Care and Use of Laboratory Animals\" under an approved protocol by the University of Nebraska Medical Center\\'s Institutional Animal Care and Use Committee (IACUC).\n\nIsolation of total RNA and miRNAs {#s4_2}\n---------------------------------\n\nTotal RNA and miRNA were isolated from the pancreas of floxed Kras^G12D^ mice positive for Kras;Pdx-1-Cre and unfloxed LSLKras^G12D^ mice, human PC cell lines, tissues and their cancer-adjacent normal tissues (eight tissues each) using mirVana\u2122 miRNA Isolation Kit (Applied Biosystems/Ambion, Austin, TX, USA). The RNA concentration was measured by Nanodrop Spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA), and the quality was analyzed using bioanalyzer (Agilent technologies, Waldbronn, Germany). Samples with good integrity were used for miRNA array, real-time PCR, and TaqMan assays.\n\nAffymetrix microRNA microarray {#s4_3}\n------------------------------\n\nGeneChip^\u00ae^ miRNA Array (Cat. No: 901325) was used to study the global expression profile of miRNAs at 25-week of Kras^G12D^;Pdx-1-Cre mice and LSLKras^G12D^ mice. Two micrograms of total RNA from two floxed Kras^G12D^ (Kras^G12D^;Pdx1-Cre) and two unfloxed LSL-Kras^G12D^ (LSL Kras^G12D^) animals were labeled with the FlashTag Biotin RNA Labeling Kit for Affymetrix GeneChip^\u00ae^ miRNA arrays (Genisphere, Hatfield, PA, USA) and hybridized on Affymetrix GeneChip^\u00ae^ miRNA arrays (Affymetrix, Santa Clara, CA, USA). Hybridization, washing, and scanning of the slides were done according to Affymetrix\\'s recommendations (Fluidics Protocol FS450_0003). Data was extracted from the images, quintile normalized, summarized (median polish) and log2-transformed with the miRNA QC tool software from Affymetrix.\n\nDetection of mature miRNAs by TaqMan assays {#s4_4}\n-------------------------------------------\n\nThe TaqMan miRNA assay kits purchased from Applied Biosystems were used for expression and validation analysis of miRNAs during the progression of PC in both mice as well as human tumor samples. In order to analyze the expression patterns, first, reverse transcription reactions for each miRNA was performed using 1.5 \u03bcl of 10X reverse transcription buffer, 1.0 \u03bcl of MultiScribe\u2122reverse transcriptase (50 U/\u03bcl), 0.15 \u03bcl of 100 mM dNTPs (with dTTP), 0.19 \u03bcl of RNase inhibitor (20 U/\u03bcl) and 4.16 \u03bcl of nuclease-free water. Subsequently, 500 ng of miRNA and 3\u03bcl of 1X looped-primers were added (specific for each microRNA), mixed well and incubated for 30 min at 16\u00b0C followed by 30 min at 42\u00b0C and 5 minutes at 85\u00b0C. The resultant cDNA was diluted to a minimum of 1:15 with nuclease free water and 1.33 \u03bcl was used in real-time PCR.\n\nReal-time PCR was performed on Light cycler 480 II PCR System (PCR System, Roche Applied Science, Indianapolis, IN, USA). During the target amplification step, the AmpliTaq^\u00ae^ Gold DNA polymerase amplifies target cDNA synthesized from the RNA samples using sequence-specific primers from the TaqMan Assay. Real-time PCR reactions based on TaqMan^\u00ae^ reagent chemistry were performed in triplicate, and template controls were run for each assay under the same conditions. End-point PCR was performed in 20 \u03bcl reaction that contained 10 \u03bcl TaqMan 2X Universal PCR Master Mix No AmpErase UNG, 7.67 \u03bcl of nuclease free water, 1.33 \u03bcl diluted RT product ([Minimum1:15]{.ul}) and 1 \u03bcl miRNA-specific PCR probe. The cycling conditions were as follows: 95\u00b0C for 10 minutes followed by 40 cycles consisting of 95\u00b0C for 15 seconds followed by 60\u00b0C for 60 seconds. The miRNA expression levels were normalized to the level of *RNU6B* for human samples and *U6* snRNA for mouse samples to correct for differential cDNA content.\n\nSynthesis of cDNA and real-time PCR {#s4_5}\n-----------------------------------\n\nTotal RNA was isolated, and the cDNA was synthesized by reverse transcription as previously described \\[[@R26]\\]. Briefly, reverse transcription of RNA was performed by adding 10 \u03bcl of (2 \u03bcg) total RNA, 1 \u03bcl of Oligo (dT)12--18 (500 \u03bcg/ml), and 1 \u03bcl of 10 mM dNTP (Invitrogen, Carlsbad, CA), incubating at 65\u00b0C for 5 minutes and immediately chilling on ice. Then, the master mix containing 4 \u03bcl of (5X) first strand RT buffer, 1 \u03bcl of 0.1 M DTT, and 1 \u03bcl of RNaseOUT (RNase Inhibitor) was incubated at 42\u00b0C for 2 min. Finally, 1 \u03bcl (50 units) of SuperScript II RT was added to each tube and incubated at 42\u00b0C for 50 minutes followed by 70\u00b0C for 15 minutes in order to destroy the superscript II RT.\n\nThe real-time primers for all the genes (human and mouse) were designed using Primer 3 software ([Supplementary table S2](#SD1){ref-type=\"supplementary-material\"}). Real-time PCR was performed on Light cycler 480 II PCR System, (Roche Applied Science, Indianapolis, IN, USA). Real-time PCR reactions were performed in triplicate; non-template controls (NTCs) and a standard curve was run for each assay under similar conditions. Real-time PCR was performed in a 10 \u03bcl reaction containing 5 \u03bcl 2X SBYR green Master mix, 3.6 \u03bcl of nuclease free water, 1 \u03bcl diluted RT product (1:10), and 0.2 \u03bcl each of forward and reverse primer (5 pmol/\u03bcl). The cycling conditions used were as follows: 95\u00b0C for 10 minutes, followed by 40 cycles of 95\u00b0C for 15 seconds and 60\u00b0C for 1 minute. Gene expression levels were normalized to the level of \u03b2-actin expression and were reported relative to the expression level in RNA from corresponding normal controls.\n\nGeneration of Let-7b overexpressing human PC cell lines {#s4_6}\n-------------------------------------------------------\n\nThe 293FT cells were transfected with (2 \u03bcg) either with Let-7b over-expressing lentiviral pCDH-CMV-MCS-EF1-copGFP-Let-7b vector or with vector control and along with 10 \u03bcg of pPACKH1 Packaging Plasmid Mix (System Bioscience, CA, USA) using lipofectamine-2000 plus reagents (Invitrogen). The viral supernatant, collected after 48 and 72 hours post-transfection, was filtered and used to infect human PC cells (CD18/HPAF and Capan1) after mixing with 4 \u03bcg/ml of polybrene. The GFP+ cells were sorted by FACS, and Let-7b overexpression was analyzed by TaqMan assay.\n\nImmunoblot assay {#s4_7}\n----------------\n\nProtein lysates were prepared from Let-7b-overexpressing and vector-transfected GFP positive cells, followed by Western blot analysis. Briefly, the cells from 70% confluency cultures were washed twice with PBS and lysed in RIPA buffer (50 mM Tris-HCl, pH 7.4; 0.25% Na-deoxycholate; 1 mM EDTA ;150 mM NaCl; 1% NP-40), supplemented with protease inhibitor cocktail, 5 mM Na3VO4, 5 mM NaF, and 1 mM phenylmethylsulphonyl fluoride. Protein lysates were quantified and resolved using 10% acrylamide/bisacrylamide gels (2% agarose gel for MUC4) and transferred onto a PVDF membrane. The membranes were blocked in 5% nonfat dry milk in PBS for at least 1 hour and then incubated with primary antibodies (anti-Kras, anti-MSST1, anti-MUC4, anti-caspase 9 anti-cyclin D1 and anti-\u03b2-actin) overnight at 4\u00b0C. Then, the membrane was washed (4X 10 min) with phosphate buffered saline and 0.1% Tween 20 (PBST) and probed with 1:5000 diluted horseradish peroxidase-conjugated anti-mouse or anti-rabbit secondary antibodies for 1 hour at room temperature. After washing the blots, the signal was detected with an ECL chemiluminescence kit (Amersham Bioscience, UK).\n\nStatistical analyses {#s4_8}\n--------------------\n\nThe fold change in the miRNA expression was calculated by \u0394\u0394Ct method as a 2^\u2212\u0394\u0394Ct^. U6 and RNU6B were used for normalization in mouse and human samples, respectively. A change of two-fold or more (on a log scale of \u2265 0.3) was considered statistically significant. The miRNA ratios of experimental samples to controls were compared to one using the one-sample non-parametric Wilcoxon sign rank test. The *p*-value was not adjusted for multiple comparisons. SAS software (SAS Institute Inc., Cary, NC) was used for data analysis.\n\nSUPPLEMENTARY TABLES {#s5}\n====================\n\nWe thank Kavita Mallya for technical support and the Molecular Biology Core Facility, UNMC, for miRNA microarray analysis. We thank Melody A. Montgomery for editing this manuscript.\n\nThis work was supported, in parts, by the grants from National Institutes of Health (UO1 CA111294, P50 CA127297 and U54 CA163120) and State of Nebraska ((NE DHHS-LB506).\n\n**CONFLICTS OF INTEREST**\n\nThe authors have no conflict of interest with any company or financial organization.\n"}
+{"text": "1. Introduction {#sec1-antioxidants-08-00205}\n===============\n\nThe *Myrtus* genus, belonging to the Myrtaceae family, comprises about 50 species that are native of the Mediterranean basin. Among those, *M. communis* is an aromatic evergreen perennial sub-shrub (high: 1--3 m) with white flowers (blossoming time: June to July) and dark blue ripe berries \\[[@B1-antioxidants-08-00205]\\]. It is native to Southern Europe, North Africa and West Asia and widespread in the Mediterranean region. Its fruits are consumed either raw or processed in diverse products such as canned fruits, yogurts, beverages, jams and jellies. In addition, there has been a growing interest in the use of berry extracts as ingredients in functional foods and dietary \\[[@B2-antioxidants-08-00205]\\]. Volatile oils, tannins, anthocyanins, fatty acids, sugars, and organic acids such as citric and malic acids are important components of these fruits \\[[@B3-antioxidants-08-00205]\\]. In general, myrtle berries are accepted as being rich in phenolic compounds, which in turn are associated to the fruits claimed health effects, including the prevention of degenerative diseases, such as cancer and cardiovascular diseases \\[[@B4-antioxidants-08-00205]\\]. In fact, phenolic compounds are accepted as potent antioxidants due to their double bonds and hydroxyl groups, being capable of preventing the oxidation of free radicals that may damage physiological molecules cells, such as lipid proteins and DNA \\[[@B5-antioxidants-08-00205]\\]. Many studies have shown a positive relationship between the phenolics content and the antioxidant capacity of fruits and vegetables \\[[@B4-antioxidants-08-00205],[@B5-antioxidants-08-00205],[@B6-antioxidants-08-00205]\\]. Moreover, the regular consumption of fruits and vegetables is believed to prevent oxidative stress events and oxidative-stress related diseases \\[[@B7-antioxidants-08-00205],[@B8-antioxidants-08-00205],[@B9-antioxidants-08-00205]\\].\n\nDue to countless beneficial characteristics of phenolic compounds in human health, research has been intensified, aiming to find fruits, vegetables, plants, agricultural and agroindustrial residues as sources of these bioactive components. Obtaining such compounds often requires many long and costly steps, such as extraction, isolation and identification \\[[@B5-antioxidants-08-00205]\\], and often result in thermal degradation of various bioactive constituents \\[[@B6-antioxidants-08-00205],[@B7-antioxidants-08-00205]\\]. In this context, the development of new extraction methods is one of the major challenges in technological innovation towards the direction of \"Green chemistry\" \\[[@B8-antioxidants-08-00205],[@B9-antioxidants-08-00205]\\]. Among them, ultrasound-assisted extraction (UAE) and microwave-assisted extraction (MAE) are particularly attractive because of their simplicity, low cost of equipment, efficiency in extracting analytes from different matrices and the requirement of low energy, reduced quantity of solvent and/or time consumption \\[[@B10-antioxidants-08-00205]\\], compared to conventional extraction methods, which have several disadvantages, such as the use of volatile and hazardous solvents, the long extraction time and more recovery energy \\[[@B11-antioxidants-08-00205]\\]. The enhancement of the extraction process by ultrasounds is attributed to the disruption of the cell walls, reduction of the particle size and the increased mass transfer of the cell content to the solvent, caused by the collapse of the bubbles produced by acoustic cavitation \\[[@B8-antioxidants-08-00205]\\]. The processing parameters optimization and interpretation of experiments compared to others has been previously done through response surface methodology (RSM) \\[[@B12-antioxidants-08-00205]\\]. This latter has been shown to be a powerful tool in optimizing experimental conditions (factors) to maximize the response. With the experimental results of a response surface design, a mathematical polynomial model, describing the relation between a response (dependent variable) and the considered factors (independent variable), is built. The mathematical model, usually a second-order polynomial model, can be visualized graphically by drawing 2D contour plots or 3D response surface plots \\[[@B13-antioxidants-08-00205]\\]. The model allows determining the optimum value of the independent variables (*X~i~*), as well as those of the dependent ones (*Y*).\n\nPrevious studies focusing on phenolic compounds and/or the antioxidant abilities of myrtle pericarp have been performed with extracts obtained by conventional methods \\[[@B14-antioxidants-08-00205],[@B15-antioxidants-08-00205],[@B16-antioxidants-08-00205]\\], while, to our knowledge, there is no available information on the optimization of ultrasonic procedure for the extraction of phenolic compounds from this matrix, using a safer solvent such as ethanol, which is an organic solvent used in the food and pharmaceutical industries \\[[@B17-antioxidants-08-00205]\\]. Therefore, the present study aimed at the optimization of UAE process parameters using RSM, including ethanol concentration, extraction time, irradiation amplitude and liquid-to-solid ratio, to maximize the content of the extracted phenolics. Levels of phenolic compounds and the antioxidant activity of pericarp *M. communis* extract obtained under the optimum setting parameters (UAE-OPT extract) were compared with those of extracts obtained by microwave-assisted extraction (MAE) and conventional solvent extraction (CSE) methods, using previously established conditions \\[[@B18-antioxidants-08-00205]\\]. Then, the individual phenolic compounds present in the optimized extract obtained by UAE were identified by UHPLC-DAD-ESI-MS^n^.\n\n2. Materials and Methods {#sec2-antioxidants-08-00205}\n========================\n\n2.1. Plant Material {#sec2dot1-antioxidants-08-00205}\n-------------------\n\nThe fruits of *M. communis* were harvested from spontaneous plants in Adakar, Bejaia, located in the northeast of Algeria. The collected samples were identified by the Vegetable Ecological Laboratory of the Algiers University, Algeria and a voucher specimen was deposited at the Herbarium of Natural History Museum of Aix-en-Provence, France, under the voucher number D-PH-2013-37-12. Berries were washed and then dried in a static oven at 40 \u00b0C for one week. Pericarps were separated manually from seeds and further grounded in an electrical grinder (A11Basic, IKA, Retsch, Germany), which was then sieved to obtain a fine powder (\\<250 \u00b5m).\n\n2.2. Extraction of Phenolic Compounds {#sec2dot2-antioxidants-08-00205}\n-------------------------------------\n\n### 2.2.1. Ultrasound Extraction {#sec2dot2dot1-antioxidants-08-00205}\n\nUAE was performed in an ultrasonic apparatus (Vibra cell, VCX 75115 PB, SERIAL No. 2012010971 MODEL CV 334, SONICS, Newtown, Connecticut, USA) with a working frequency fixed at 20 kHz. For extraction, 1 g of the pericarp powder was placed in a 250 mL amber glass bottle containing ethanol. The suspension was exposed to acoustic waves under distinct setting parameters (solvent concentration, irradiation time, ultrasound amplitude and solvent-to-solid ratio). The temperature was maintained constant by circulating external cold water and checking the temperature using a T-type thermocouple \\[[@B5-antioxidants-08-00205]\\]. Indeed, ultrasound is considered a non-thermal technology, since it increases only the local temperature without affecting the surrounding environment \\[[@B19-antioxidants-08-00205]\\]. After extraction, the solution was filtered through a sintered glass filter of porosity 2.\n\nTo determine the effect of ethanol concentration, irradiation time, ultrasound amplitude and solvent-to-solid ratio on the extraction yield of phenolic compounds from myrtle pericarp, RSM was applied with a Box--Behnken Design (BBD) \\[[@B5-antioxidants-08-00205]\\]. This design resulted in the testing of four factors in a single block of 30 sets of test conditions ([Table 1](#antioxidants-08-00205-t001){ref-type=\"table\"}). The constant values for irradiation time, liquid-to-solid ratio and ethanol concentration in the UAE trials were 10 min, 50 mL/g and 50% (*v*/*v*), respectively. $$Y = \\beta_{0} + \\sum_{i = 0}^{k}\\beta_{i}X_{i} + \\sum_{i = 1}^{k}\\beta_{ii}X^{2} + \\sum_{i > 1}^{k}\\beta_{ij}X_{i}X_{j} + E$$ where *X~i~*, *X~j~*, ..., *X~k~* are the independent variables affecting the responses *Y* (the yield of total phenolic compounds); *\u03b2~i~*, *\u03b2~ii~* and *\u03b2~ij~* are the regression coefficients for linear, quadratic and interaction terms, respectively; and k is the number of variables.\n\nThe factor levels were coded as \u22121 (low), 0 (central point or middle) and 1 (high), respectively, according to Equation (2): where *X~i~* is the coded value of the variable *X~i~*; *X~0~* is the value of *X* at the center point; and \u0394*X* is the step change.\n\n### 2.2.2. Microwave-Assisted Extraction {#sec2dot2dot2-antioxidants-08-00205}\n\nPhenolic extracts were obtained using a domestic microwave oven (Samsung MW813ST, Kuala Lumpur, Malaysia) adapted by adding of a condenser \\[[@B18-antioxidants-08-00205]\\]. The apparatus operated at a frequency of 2450 MHz and a maximum output power of 1000 W with a 100 W increment. The size of the heating cavity was 37.5 cm (L) \u00d7 22.5 cm (W) \u00d7 38.6 cm (D). The applied extraction conditions corresponded to those previously optimized \\[[@B18-antioxidants-08-00205]\\]. A volume of 32 mL of 42% ethanol concentration was added to 1 g of pericarp *Myrtus* powder in a flat-bottomed flask. The mixture was irradiated at 500 W for 62 s. The resultant extract was then filtered through a sintered glass filter of porosity 2 and was stored at 4 \u00b0C until further analysis.\n\n### 2.2.3. Conventional Solvent Extraction {#sec2dot2dot3-antioxidants-08-00205}\n\nConventional solvent extract followed the procedure established by Dahmoune et al. \\[[@B5-antioxidants-08-00205]\\]. One gram of myrtle powder was placed in a conical flask, and 50 mL of 50% (*v*/*v*) ethanol were added. After stirring for 2 h, the mixture was filtered through a sintered glass filter of porosity 2 and the extract was stored at 4 \u00b0C until further use.\n\n2.3. Analytical Determinations {#sec2dot3-antioxidants-08-00205}\n------------------------------\n\n### 2.3.1. Total Phenolic and Flavonoid Contents {#sec2dot3dot1-antioxidants-08-00205}\n\nThe total phenolic content (TPC) of the UAE, MAE and CSE extracts was assessed according to the method of George et al. \\[[@B20-antioxidants-08-00205]\\] and expressed as mg of gallic acid equivalent (GAE) per gram of myrtle pericarp powder on dry weight (DW) basis (mg GAE g^--1^ DW). The total flavonoid content was estimated by the aluminum trichloride method according to Quettier-Deleu et al. \\[[@B21-antioxidants-08-00205]\\] and the results were expressed as mg of quercetin equivalent per g of myrtle pericarp powder, on a DW basis.\n\n### 2.3.2. Total Monomeric Anthocyanins and Condensed Tannin Contents {#sec2dot3dot2-antioxidants-08-00205}\n\nTotal monomeric anthocyanin content was determined by the pH-differential method \\[[@B22-antioxidants-08-00205]\\], and the results were expressed as mg cyanidin-3-*O*-glucoside equivalents per g of myrtle pericarp powder on a DW basis. The condensed tannin content was determined by the HCl--vanillin method as described by Aidi Wannes et al. \\[[@B23-antioxidants-08-00205]\\] and the results were expressed as mg catechin equivalents per g of myrtle pericarp powder on DW basis.\n\n### 2.3.3. Antioxidant Activity {#sec2dot3dot3-antioxidants-08-00205}\n\nThe antioxidant activity of all samples was tested by using two different tests, namely 1,1-diphenyl-2-picrylhydrazyl radical (DPPH**^\u25cf^**) scavenging activity and reducing power methods \\[[@B24-antioxidants-08-00205]\\]. DPPH**^\u25cf^** solution (60 \u00b5M) was prepared in absolute methanol and reaction was performed by the adding of 3 mL of this solution to 1 mL of the extracts, during 20 min at 37 \u00b0C in the dark. Thereafter, the absorbance was measured at 515 nm. The inhibition rate of the extracts was calculated according Equation (3). $$\\%\\, Scavenging = \\frac{(A_{control} - A_{extract})}{A_{control}} \\times 100$$ where *A~control~* is the absorbance of DPPH**^\u25cf^** and distilled water *A~sample~* is the absorbance of DPPH**^\u25cf^** and sample extract. \u03b1-tocopherol and BHA (250 \u00b5g/mL) were used as positive controls.\n\nFor reducing power assay, 1 mL of desired dilution was mixed with 2.5 mL of sodium phosphate buffer (0.2 M, pH 6.6) and 2.5 mL of 1% (*m*/*v*) potassium ferricyanide K~3~\\[Fe(CN)~6~)\\], followed by incubation in a water bath at 50 \u00b0C for 20 min and the addition of 2.5 mL of 10% (*m*/*v*) trichloroacetic acid. At last, an aliquot of the resulting solution (1 mL) was added to 5 mL of distilled water and 1 mL of 0.1% (*m*/*v*) of FeCl3\u00b76H2O. Note that this method estimates the ability to reduce Fe^3+^ to Fe^2+^. Antioxidant compounds present in the samples form a colored complex with potassium ferricyanide, trichloroacetic acid and ferric chloride, which is measured at 700 nm.\n\n2.4. Identification of Phenolic Compounds by UHPLC-DAD-ESI-MS^n^ {#sec2dot4-antioxidants-08-00205}\n----------------------------------------------------------------\n\nThe phenolic compounds of the UAE-OPT extract were characterized by UHPLC-DAD-ESI-MS^n^ (DAD: diode array detector; ESI: electrospray ionization) analysis on an Ultimate 3000 (Dionex Co., USA) apparatus equipped with an ultimate 3000 Diode Array Detector (Dionex Co., San Jose, CA, USA) and coupled to a mass spectrometer. Analysis was run on a Hypersil Gold (Thermo Scientific, San Jose, CA, USA) C18 column (100 mm length; 2.1 mm i.d.; 1.9 \u03bcm particle diameter, end-capped) and its temperature was maintained at 30 \u00b0C. The mobile phase was composed of (A) 0.1% of formic acid (*v*/*v*) and acetonitrile (B). The solvent gradient started with 5% of Solvent B, reaching 40% at 14 min and 100% at 16 min, followed by the return to the initial conditions. The flow rate was 0.1 mL min^--1^ and UV--Vis spectral data for all peaks were accumulated in the range 200--700 nm while the chromatographic profiles were recorded at 280, 340 and 530 nm.\n\nThe mass spectrometer consisted of a Thermo LTQ XL (Thermo Scientific, San Jose, CA, USA) ion trap MS (mass spectrometer) apparatus equipped with an ESI source, operating in negative and positive modes, under the pre-established conditions \\[[@B25-antioxidants-08-00205]\\].\n\n2.5. Statistical Analysis {#sec2dot5-antioxidants-08-00205}\n-------------------------\n\nEach extraction trial and all the analyses were carried out in three independent analysis performed in triplicate. The influence of individual factors on the TPC yield (single-factor experiment) was estimated by Analysis of Variance (ANOVA) and Tukey's post hoc test with a 95% confidence level, while data obtained from the BBD and Central Composite Design (CCRD) trials were analyzed through ANOVA for the response variable to evaluate the model significance and suitability. Significant and highly significant levels were set for *p* \\< 0.05 and *p* \\< 0.01, respectively. The John's MacIntosh Product (Version 7.0, SAS, Cary, NC, USA) and Design-Expert (Trial version 10.0, SAS, Cary, NC, USA) software packages were used to construct the BBD and CCRD and to analyze all the results. Principal Component Analysis (PCA) was applied to detect the relationships between contents of phenolic compounds, flavonoids, anthocyanins, tannins, as well as antioxidant activity and their extraction methodologies i.e., UAE, MAE and CSE. All tests were done in triplicate.\n\n3. Results and Discussion {#sec3-antioxidants-08-00205}\n=========================\n\n3.1. Optimization of UAE Conditions {#sec3dot1-antioxidants-08-00205}\n-----------------------------------\n\n### 3.1.1. Modeling and Fitting the Model Using RSM {#sec3dot1dot1-antioxidants-08-00205}\n\nThe experimental design and subsequent response allied to TPC are summarized in [Table 1](#antioxidants-08-00205-t001){ref-type=\"table\"}, with results from TPC recovery varying in the range of 79--235 mg GAE/g DW.\n\nThe least square technique was used to calculate the regression coefficients of the intercept, linear, quadratic, and interaction terms \\[[@B26-antioxidants-08-00205]\\] ([Table 2](#antioxidants-08-00205-t002){ref-type=\"table\"}). Notably, the linear parameters, namely ethanol concentration, irradiation time and liquid--solid ratio (*p* \\< 0.0001), followed by amplitude (*p* = 0.0421) significantly affected the extraction content of phenolic compounds. The quadratic terms *X~2~*^2^ and *X~4~*^2^ were highly significant at the level *p* \\< 0.001, while the *X~1~*^2^ and *X~3~*^2^ terms were insignificant (*p* \\> 0.05). Regarding TPC yield, the interaction of ethanol concentration with amplitude of ultrasound (*X~1~*--*X~3~*) and with liquid to solid ratio (*X~1~*--*X~4~*), and that of irradiation time amplitude of ultrasound (*X~2~*--*X~3~*) were highly significant (*p* \\< 0.0001), followed by irradiation time with liquid-to-solid ratio (*p* = 0.0054), amplitude of ultrasound with liquid-to-solid ratio (*p* \\< 0.0094) and ethanol concentration with irradiation time (*p* = 0.0367). Those significant terms played a dominant role in myrtle pericarp extraction by ultrasound. Indeed, those significant terms played a dominant role in myrtle pericarp extraction by ultrasound. This is justified by the analyses of variance, as represented in [Table 2](#antioxidants-08-00205-t002){ref-type=\"table\"}, with significant *p*-values (*p* \\< 0.0001) for the linear parameters, namely ethanol concentration (*X~1~*), irradiation time (*X~2~*) and liquid--solid ratio (*X~4~*), the quadratic terms *X~2~*^2^ and *X~4~*^2^, and the interaction terms (*X~1~*--*X~3~*), (*X~1~*--*X~4~*) and (*X~2~*--*X~3~*). However, insignificant *p*-values (*p* \\> 0.0001) were obtained for the third linear parameter, amplitude (*X~3~*), the quadratic terms *X~1~*^2^ and *X~3~*^2^ and the interaction terms (*X~1~*--*X~2~*), (*X~2~*--*X~4~*) and (*X~3~*--*X~4~*).\n\nBased on the significant terms, the regression equation for the UAE efficiency was obtained as follows:\n\nNote that the *p*-value can be employed to check the interaction strength between independent factors. From this analysis, *p*-value \\<0.0001 indicated that the response surface quadratic model was significant, which means that the model represented the data satisfactorily. The adjusted coefficient of determination *(R*^2^adj) and the coefficient of determination (R^2^) were 0.9553 and 0.9776, respectively, which implied that the sample variations of 97.76% for the UAE efficiency of myrtle pericarp phenols were attributed to the independent variables, and only 2.24% of the total variations could not be explained by the model, indicating a good degree of correlation between experimental and predict values of the TPC yield. In addition, the low value of coefficient of variance (3.71%) clearly indicated that the model was reproducible and reliable \\[[@B27-antioxidants-08-00205]\\]. All these results indicate that the model could work well for the prediction of TPC in the myrtle pericarp extracts.\n\n### 3.1.2. Response Surface Analysis (RSA) {#sec3dot1dot2-antioxidants-08-00205}\n\nTo provide a better understanding of the interaction between factors, the 3D response surface plot was constructed ([Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}) using Equation (4). The graphs were generated by plotting the response using the z-axis against two independent variables, while keeping the other independent variable at the fixed level. [Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}A--C shows the interactions between the ethanol concentration and each of the three other factors, namely irradiation time, amplitude and liquid-to-solid ratio, respectively, on the recovery of TPC. As shown, an increase of ethanol concentration from 20% to 80% (*v*/*v*), or extraction time from 5 to 10 min resulted in a rapid enhancement of TPC with a maximum of 235.21 mg GAE/g being recovered with an irradiation time of 7.5 min and ethanol concentration of 70% (*v*/*v*).\n\nThe high phenolic content indicates that the mixture ethanol/water at 70% (*v*/*v*) allowed the solubilization of phenolics from *M. communis* pericarp, thus confirming the results of the single factor experiments \\[[@B28-antioxidants-08-00205]\\] that explained the efficiency of the ultrasonic method by the fact that sonication improved the hydration and fragmentation process and hence facilitates the mass transfer of solutes to the extraction solvent. For the extraction yield of TPC performed at fixed extraction time and liquid-to-solid ratio, with varying ethanol concentration and amplitude ([Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}B), it was possible to conclude that maximum recovery (210.05 mg GAE/g) was achieved for 70% (*v*:*v*) of ethanol and an ultrasound amplitude of 35%. This fact can be explained by the larger amplitude ultrasonic wave that promotes the liquid medium to produce more cavitation bubbles, thus resulting in a stronger pressure, capable of destroying the cell wall and accelerating mass transfer \\[[@B29-antioxidants-08-00205]\\]. [Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}C shows an enhancement of TPC that reached a peak value of 230.15 mg GAE/g, for 70% (*v*:*v*) ethanol and about 30 mL/g of liquid-to-solid ratio. A higher ratio corresponds to a greater concentration difference between the exterior solvent and the interior tissues of *Myrtus* pericarp. It prominently prompted the TPC to be rapidly dissolved, which resulted in an increase in the extraction yield. The response surface plot for the significative interactive effect of irradiation time and amplitude of ultrasound on the response value at a fixed ethanol concentration and liquid-to--solid ratio is shown in [Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}D. A higher TPC was obtained with the irradiation time at 10 min and amplitude of 30%; these results confirm those reported in the literature \\[[@B30-antioxidants-08-00205],[@B31-antioxidants-08-00205]\\]. [Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}E shows an interaction between extraction time and the liquid-to-solid ratio (*p* \\< 0.05). The best content (148 mg GAE/g) was found with the solid--liquid ratio of about 30 mL/g and the radiation time of 10 min. The increase of the ethanol proportion required high sonication intensity to generate the cavitation bubbles. However, a higher increase in the liquid-to-solid ratio diminished the supply of ultrasonic energy density and negatively affected the extraction yield.\n\nThe yield of TPC constantly improved with the increase of both amplitude of ultrasound and liquid-to-solid ratio, reaching a maximum when *X~3~* and *X*~4~ became 32% and 20% (*v*/*v*), respectively ([Figure 1](#antioxidants-08-00205-f001){ref-type=\"fig\"}F). Beyond this level, the yield of TPC reduced with the increase of *X*~1~ and *X*~4~. Hence, the interactive effect of *X~3~* and *X~4~* was remarkable. Overall, these results indicate that the TPC extraction yield was more significantly affected (*p* \\< 0.0001) by linear parameters, namely ethanol concentration, irradiation time and liquid-to-solid ratio.\n\n### 3.1.3. Validation and Verification of the Predictive Model {#sec3dot1dot3-antioxidants-08-00205}\n\nAccording to the result of response surface and prediction by this built model, the optimal conditions were thus obtained for the following conditions: ethanol at 70% (*v*/*v*), 7.5 min extraction time, 30% amplitude and a liquid-to-solid ratio of 28 mL/g. To ensure that the predicted result was not biased to the practical value, experimental rechecking was performed using these deduced optimal conditions. The predicted extraction yield of TPC in UAE-OPT was 235.52 \u00b1 9.9 mg GAE/g, that was consistent with the experimental yield of 241.66 \u00b1 12.77 mg GAE/g DW ([Table 3](#antioxidants-08-00205-t003){ref-type=\"table\"}). The results showed no significant difference between the experimental and the predicted values. This strong correlation between experimental and the predicted values indicates that the response of regression model is adequate to reflect the expected optimization for the extraction of antioxidants from *M. communis* pericarp.\n\n3.2. Comparison between UAE, MAE and CSE Methods {#sec3dot2-antioxidants-08-00205}\n------------------------------------------------\n\nRemarkably, the highest TPC was obtained by UAE (241.60 \u00b1 12.77 mg GAE/g). This corresponded to four and three times higher than that obtained by MAE and CSE, respectively, thus indicating that the application of UAE has a positive effect on the extraction of TPC ([Table 3](#antioxidants-08-00205-t003){ref-type=\"table\"}). The highest levels of TPC in UAE-OPT extract was reflected by its higher amounts of flavonoids, anthocyanins and tannins (18.99 \u00b1 1.31 mg QE/g; 25.06 \u00b1 0.36 mg/g; 35.56 \u00b1 0.36 mg CE/g, respectively). These findings are consistent with those reported in the literature \\[[@B32-antioxidants-08-00205]\\] and are mainly attributed to the fact that ultrasound radiation can facilitate mass transfer and accelerate the extracting process so that the extraction of bioactive compounds may be improved. Hence, according to the overall data, it is possible to conclude that the herein optimized UAE process yields higher levels of bioactive compounds in a short time and requires less solvent consumption than MAE and CSE. Note that, in this study, the operating temperature in the UAE-OPT was kept constant at room temperature, excluding any heating effect. This might positively or negatively influence the polyphenols recovery depending on the applied amplitude.\n\nThe antioxidant capacity of the extracts was assessed by DPPH**^\u25cf^** scavenging and ferric reducing antioxidant power assays. The results show that UAE-OPT extract presented higher DPPH**^\u25cf^** scavenging ability (90.71% inhibition) when compared to CSE (88.03% inhibition) and MAE (87.16% inhibition) extracts. The same tendency was also observed for reducing power, since the absorbance at 700 nm for UAE-OPT extract was considerably higher than those obtained for MAE and CSE (0.439 \u00b1 0.006 and 0.429 \u00b1 0.01, respectively). This means that UAE method is more efficient for the recovery of antioxidants than the herein tested microwave and conventional solvent extraction methods, a fact that is probably due to its superior richness in phenolic components, including flavonols \\[[@B33-antioxidants-08-00205]\\] and as evidenced in the following section. This information was also confirmed by PCA analysis. PCA was applied to the extracts (UAE-OPT, MAE and SCE) for phenolic compounds (TPC, flavonoids, anthocyanins and tannins) and antioxidant activity, where the two chosen factors justified 100.0% of total variance. The resulting plots allowed selecting the better extraction method of different compounds of myrtle pericarp, and clearly divided the samples into three groups, depending on the extraction method ([Figure 2](#antioxidants-08-00205-f002){ref-type=\"fig\"}). For PC1, which explains 95.91% of the total variance, the first group showed a positive correlation with PC1, thus confirming that UAE was the best extraction method for phenolic compounds with potent antioxidant activity. The highest correlation was found between antioxidant activity (DPPH^\u2022^ and RP essay) and anthocyanins, hence suggesting that these compounds might have a key influence on the antioxidant capacity of the extracts. The best correlation between the MAE and TPC, flavonoids and tannins were observed in the second group. PC2 explains only better 4.09% of the experimental variability, which could essentially be associated to the CSE method and anthocyanins content (the third group).\n\n3.3. Identification of Phenolics by UHPLC-DAD-ESI-MS^n^ Analysis {#sec3dot3-antioxidants-08-00205}\n----------------------------------------------------------------\n\nThe UAE-OPT extract was analyzed by UHPLC-DAD-ESI-MS^n^ to further elucidate its phenolic profile. The registered chromatogram at 280 nm is shown in [Figure 3](#antioxidants-08-00205-f003){ref-type=\"fig\"} and the UV-Vis and MS^n^ spectral data of eluted peaks are summarized in [Table 4](#antioxidants-08-00205-t004){ref-type=\"table\"}.\n\nAmong the distinct phenolic groups found in the extract, flavonols were the prevalent components. Overall, eleven flavonol glycosides were detected, being myricetin glycosides, namely myricetin-*O*-hexoside and myricetin-*O*-deoxyhexoside (eluted in Peaks 10/11 and 14/15, respectively) the major abundant ones, which probably correspond to myricetin-3-*O*-galactoside and myricetin-3-*O*-rhamnoside, since these are known to be present as main phenolic components in distinct organs of *M. communis* plant \\[[@B34-antioxidants-08-00205],[@B35-antioxidants-08-00205],[@B36-antioxidants-08-00205],[@B37-antioxidants-08-00205],[@B38-antioxidants-08-00205]\\].\n\nBesides the above compounds, four other myricetin glycosides were found in the extract. The compound eluted in Peak 9, showing a \\[M-H\\]^\u2212^ at *m/z* 631, corresponded to myricetin-*O*-galloyl-hexoside, since the main fragments in MS^2^ spectrum were formed by the loss of 152 Da (equivalent to a galloyl moiety) and 332 Da (equivalent to the simultaneous loss of galloyl and hexosyl units). This could possibly correspond to myricetin 3-(6''-*O*-galloyl galactoside), which has been previously reported in leaves \\[[@B32-antioxidants-08-00205],[@B34-antioxidants-08-00205],[@B36-antioxidants-08-00205],[@B38-antioxidants-08-00205]\\] and berries \\[[@B16-antioxidants-08-00205]\\]. In addition, the compounds with \\[M-H\\]^\u2212^ at *m/z* 449 (Peak 13) and at *m/z* 625 (co-eluted in Peak 18) were, respectively, assigned to myricetin-*O*-pentoside and myricetin-*O*-hexosyl-deoxyhexoside, according to their fragmentation pattern, which showed the loss of a pentosyl (132 Da) and deoxyhexosyl plus hexosyl (308 Da) moieties, respectively. In turn, the compound eluted in Peak 19 at 14.6 min with a pseudomolecular ion at *m/z* 569 and fragment ions at *m/z* 485 (equivalent to galloyl ester moiety) and 317 (myricetin) was tentatively assigned to a galloylester of myricetin.\n\nThe three remaining flavonols detected in the UAE-OPT extract were assigned to quercetin and kaempferol derivatives. From those, the compound eluted in Peak 12 was characterized by a \\[M-H\\]^\u2212^ at *m/z* 615 and fragment ions at *m/z* 463 (\u2212152 Da, loss of galloyl group) and 301 (\u2212162 Da, loss of an hexosyl group), and was tentatively assigned to quercetin-*O*-hexoside-gallate on the basis of data reported in the literature \\[[@B39-antioxidants-08-00205],[@B40-antioxidants-08-00205],[@B41-antioxidants-08-00205]\\]. This compound has been already reported in Myrtaceae family, namely in *Eucalyptus* species \\[[@B42-antioxidants-08-00205],[@B43-antioxidants-08-00205],[@B44-antioxidants-08-00205]\\] and two other species from the same family, namely *Myrcia multiflora* extracts \\[[@B45-antioxidants-08-00205]\\] and *Eugenia edulis* \\[[@B46-antioxidants-08-00205]\\]. In addition, the compound eluted in Peak 17 with a deprotonated ion at *m/z* 447 and a base peak fragment ion at *m/z* 301 (\u2212146, equivalent to the loss of a deoxyhexose unit), was identified as quercetin-*O*-deoxyhexoside according to literature data, probably corresponding to quercetin-3-*O*-rhamnoside \\[[@B39-antioxidants-08-00205],[@B40-antioxidants-08-00205]\\]. This last flavonoid was previously detected in pericarp \\[[@B29-antioxidants-08-00205]\\], berries \\[[@B31-antioxidants-08-00205],[@B47-antioxidants-08-00205]\\] and leaves \\[[@B30-antioxidants-08-00205],[@B31-antioxidants-08-00205],[@B35-antioxidants-08-00205]\\] of *M. communis*. Finally, the flavonol eluted in Peak 16 (\\[M-H\\]^\u2212^ at *m/z* 447) presented the main fragment ion at *m/z* 285 in the MS^2^ spectrum, which in turn showed a fragmentation pattern coherent with kaempferol. Based on UV-Vis spectra (UV~max~ at 265 and 353) and MS^n^ spectral data, this compound was assigned to kaempferol-*O*-hexoside.\n\nBesides flavonols, other flavonoids in UAE-OPT extract corresponded to anthocyanins that were eluted from 7.6 min to 9.7 min (Peaks 6--8). Note that, in general, anthocyanins are preferred detected as \\[M\\]^+^ in ESI in the positive mode, while typically they show \\[M-2H\\]^\u2212^ in the negative mode \\[[@B47-antioxidants-08-00205]\\], as represented in [Table 4](#antioxidants-08-00205-t004){ref-type=\"table\"}. Overall, according to UV-Vis and MS^n^ spectral data, these compounds were assigned to delphinidin, petunin and malvidin derivatives. In more detail, the compound in Peak 6 exhibiting a \\[M-2H\\]^\u2212^ at *m/z* 463 and a base peak MS^2^ fragment ion at *m/z* 301 (\u2212162 Da) was assigned to delphinidin-*O*-hexoside by comparison with data reported in the literature \\[[@B48-antioxidants-08-00205],[@B49-antioxidants-08-00205],[@B50-antioxidants-08-00205]\\]. In turn, petunidin-*O*-hexoside and a petunidin-*O*-hexoside derivative were eluted in Peaks 7 and 8, respectively. The first showed a \\[M-2H\\]^\u2212^ at *m/z* 477 and a main MS^2^ fragment ions at *m/z* 315/314 \\[[@B48-antioxidants-08-00205],[@B49-antioxidants-08-00205],[@B50-antioxidants-08-00205]\\] while ions corresponding to petunidin-*O*-hexoside and its hydrated form (at *m/z* 477 and *m/z* 495, respectively) were predominant in MS^2^ spectrum of the latter compound. The petunidin-*O*-hexoside derivative was co-eluted with malvidin-*O*-hexoside (\\[M-2H\\]^\u2212^ at *m/z* 477\u2192329). Note that, except for petunidin-*O*-hexoside, hexosides of delphinidin, petunin and malvidin have already been described in distinct organs of *M. communis*, including pericarp \\[[@B14-antioxidants-08-00205],[@B30-antioxidants-08-00205],[@B31-antioxidants-08-00205],[@B51-antioxidants-08-00205],[@B52-antioxidants-08-00205]\\].\n\nSeveral non-flavonoid compounds could also be observed in UEA-OPT extract, including caffeoyl hexoside, gallic acid and galloyl derivatives. The first (\\[M-H\\]^\u2212^ at *m/z* 341\u2192179, eluted in Peak 1) was the only hydroxycinnamic acid found in the extract. Gallic acid (\\[M-H\\]^\u2212^ at *m/z* 169\u2192125, eluted in Peak 4), has been described in the literature for extracts obtained from the pericarp \\[[@B16-antioxidants-08-00205]\\] berries \\[[@B16-antioxidants-08-00205],[@B30-antioxidants-08-00205]\\] and leaves \\[[@B38-antioxidants-08-00205],[@B52-antioxidants-08-00205]\\].\n\nRegarding galloyl derivatives (typical UV~max~ at 273--276 nm), these enclosed esters of mono- or di-galloyl groups with a hexose or quinic acid unit, or even with myrtucommulone-type groups. In detail, the compound eluted in Peak 2 with a \\[M-H\\]^\u2212^ at *m/z* 331 and corresponding fragments at *m/z* 271, 169, 241, 211, 193 and 125, was assigned to a galloyl hexoside \\[[@B53-antioxidants-08-00205]\\], presumably galloyl-3-*O-\u03b2*-D-galactoside-6-*O*-gallate, since this latter has been previously reported in *M. communis* leaves \\[[@B2-antioxidants-08-00205],[@B38-antioxidants-08-00205]\\]. Besides, two isomers of galloyl quinic acid (\\[M-H\\]^\u2212^ at *m/z* 343\u2192191, 169, 125) could be found in Peaks 3 and 5, while a digalloyl hexoside (\\[M-H\\]^\u2212^ at *m/z* 483\u2192271, 331, 313, 439, 193, 169) and digalloyl quinic acid (\\[M-H\\]^\u2212^ at *m/z* 495\u2192343, 325, 191, 169) were detected as co-eluted compounds in Peak 6. All these galloyl derivatives have been previously detected in *M. communis* leaves \\[[@B35-antioxidants-08-00205],[@B38-antioxidants-08-00205],[@B52-antioxidants-08-00205]\\].\n\nMoreover, four gallomyrtucommulone-type derivatives were found in UAE-OPT extract. All these compounds showed a UV~max~ at 276 nm, and similar fragment ions in MS^n^ spectra, including ions at *m/z* 331, 313, 271 and 211, which are typically formed in galloylhexoside \\[[@B43-antioxidants-08-00205]\\]. Indeed, the ion at *m/z* 331 correspond to the galloyl hexoside moiety, while ions at *m/z* 271 and *m/z* 211 result from the cross-ring fragmentation of the hexose unit in the galloyl hexoside moiety and that at *m/z* 313 can be formed due to the loss of water molecule from the latter. Among these compounds, those eluted in Peaks 20 and 21 (\\[M-H\\]^\u2212^ at *m/z* 583 and 567, respectively) were assigned to gallomyrtucommulone F and gallomyrtucommulone C, in accordance to previous data reported in *M. communis* leaves \\[[@B36-antioxidants-08-00205]\\]. Besides these two compounds, the extract also contained two isomeric unidentified gallomyrtucommulone-type derivatives (MW 468 Da) that presumably vary in their acyl chain regarding those previously identified.\n\n4. Conclusions {#sec4-antioxidants-08-00205}\n==============\n\nThe response surface methodology was successfully employed to optimize total phenolic extraction yield from dried *M. communis* pericarp by non-conventional solvent extraction process, namely using UAE. As compared to MAE and CSE extractions, the proposed UAE method allowed a higher phenolic recovery yield and antioxidant activity with a short working time and a lower solvent consumption. The quantification of the amounts of phenolic compounds in the three types of extract complemented with PCA analysis also allowed concluding that the *M. communis* pericarp extract obtained under optimal UAE experimental conditions contained higher levels of flavonoids, tannins and anthocyanin than the remaining extracts and, particularly, the latter phenolic components could be correlated to its antioxidant activity. According to UHPLC-DAD-ESI-MS^n^ analysis, flavonols, particularly myricetin-*O*-hexoside and myricetin-*O*-deoxyhexoside, were the prevalent phenolic components of UAE-OPT.\n\nN.B.-M. performed the experiments (extraction, optimization and quantification of phenolics and antioxidant activity), analyzed the data and wrote the original draft; L.B.-M. contributed to conceptualization, supervision, data curation, and writing---review and editing; K.M. conceptualized and supervised the research; A.M.S.S. contributed resources and performed writing---review and editing; S.D. contributed to the methodology; S.O.-B. contributed to the investigation; S.M.C. contributed to the supervision, data curation and writing---review and editing.\n\nFoundation for Science and Technology (FCT), the European Union, the National Strategic Reference Framework (QREN), the European Regional Development Fund (FEDER), and Operational Programme Competitiveness Factors (COMPETE), for funding the Organic Chemistry Research Unit (QOPNA) (FCT UID/QUI/00062/2019), through national funds and where applicable co-financed by the FEDER, within the PT2020 Partnership Agreement. Project AgroForWealth (CENTRO-01-0145-FEDER-000001), funded by Centro2020, through FEDER and PT2020, financed the research contract of Susana M. Cardoso.\n\nThe authors have declared no conflict of interest.\n\n![Response surface analysis for the Total phenolic compounds (TPC) with UAE with respect to: (**a**) ethanol concentration and irradiation time; (**b**) ethanol concentration and amplitude; (**c**) ethanol concentration and solvent-to-solid ratio; (**d**) extraction time and amplitude; (**e**) extraction time and solvent-to-solid ratio; and (**f**) amplitude and solvent-to-solid ratio.](antioxidants-08-00205-g001){#antioxidants-08-00205-f001}\n\n![Principal component analysis of phenolic compounds for *M. communis* pericarp with UAE, MAE and CSE. FLAV, flavonoids; ANTHO, anthocyanins; TANN, tannins.](antioxidants-08-00205-g002){#antioxidants-08-00205-f002}\n\n![Chromatographic profile at 280 nm of *M. communis* pericarp extract obtained by UAE extraction at optimized conditions. Numbers in the figure correspond to the eluted UHPLC peaks for which UV and MS data are summarized in [Table 4](#antioxidants-08-00205-t004){ref-type=\"table\"}.](antioxidants-08-00205-g003){#antioxidants-08-00205-f003}\n\nantioxidants-08-00205-t001_Table 1\n\n###### \n\nCentral composite design with the observed responses and predicted values for yield of total phenolic compounds of *Myrtus communis* pericarp using the UAE method.\n\n  -----------------------------------------------------------------------------------------------------------------------------------\n  Run   *X~1~*-Ethanol\\   *X~2~*-Irradiation Time (min)   *X~3~*-Amplitude\\   *X~4~*-Solvent-to Solid Ratio (mL/g)   TPC Recovery\\\n        (%, *v*/*v*)                                      (%)                                                        (mg GAE/g DW)\n  ----- ----------------- ------------------------------- ------------------- -------------------------------------- ----------------\n  1     50                2.5                             50                  25                                     134.93 \u00b1 11.37\n\n  *2*   70                10                              70                  30                                     195.24 \u00b1 0.99\n\n  *3*   30                5                               30                  20                                     105.29 \u00b1 11.72\n\n  4     70                5                               30                  30                                     221.73 \u00b1 3.64\n\n  5     50                7.5                             50                  25                                     200.50 \u00b1 12.82\n\n  6     50                7.5                             50                  15                                     78.90 \u00b1 10.45\n\n  7     50                7.5                             50                  25                                     200.90 \u00b1 28.02\n\n  8     50                7.5                             50                  25                                     200.10 \u00b1 23.11\n\n  9     50                7.5                             90                  25                                     200.02 \u00b1 13.23\n\n  10    70                10                              30                  30                                     214.07 \u00b1 14.66\n\n  11    50                7.5                             10                  25                                     210.72 \u00b1 2.70\n\n  12    50                7.5                             50                  25                                     210.21 \u00b1 15.39\n\n  13    10                7.5                             50                  25                                     170.43 \u00b1 9.38\n\n  14    30                10                              70                  20                                     159.56 \u00b1 10.02\n\n  15    30                10                              70                  30                                     228.39 \u00b1 12.96\n\n  16    50                7.5                             50                  25                                     203.34 \u00b1 16.40\n\n  17    70                10                              70                  20                                     200.42 \u00b1 14.47\n\n  18    70                10                              30                  20                                     185.51 \u00b1 13.34\n\n  19    50                7.5                             50                  35                                     195.94 \u00b1 12.80\n\n  20    50                12.5                            50                  25                                     190.43 \u00b1 15.47\n\n  21    70                5                               70                  30                                     142.96 \u00b1 9.60\n\n  22    30                10                              30                  20                                     118.92 \u00b1 12.08\n\n  23    70                5                               70                  20                                     115.45 \u00b1 16.12\n\n  24    30                5                               30                  30                                     219.12 \u00b1 18.72\n\n  25    30                10                              30                  30                                     180.77 \u00b1 9.38\n\n  26    70                5                               30                  20                                     161.68 \u00b19.42\n\n  27    30                5                               70                  30                                     179.22 \u00b1 9.70\n\n  28    90                7.5                             50                  25                                     235.21 \u00b117.36\n\n  29    50                7.5                             50                  25                                     210.21 \u00b1 16.60\n\n  30    30                5                               70                  20                                     111.38 \u00b1 7.63\n  -----------------------------------------------------------------------------------------------------------------------------------\n\nTPC results are expressed as means \u00b1 standard deviation; GAE, gallic acid equivalent; UAE, ultrasound-assisted extraction; DW, dry weight.\n\nantioxidants-08-00205-t002_Table 2\n\n###### \n\nEstimated regression coefficients for the quadratic polynomial model and analyzes of variance (ANOVA) for the experimental results.\n\n  -----------------------------------------------------------------------------------------------\n  Parameters            Estimated\\     Standard\\   DF ^a^   Sum of\\     F Ratio ^b^   Prob \\> F\n                        Coefficients   Error                Squares                   \n  --------------------- -------------- ----------- -------- ----------- ------------- -----------\n  **Model**                                        14       46971.996   43.8356       \\<0.0001\n\n  **Intercept**         205.032        3.912523                         52.40         \\<0.0001\n\n  **Linear**                                                                          \n\n  *X~1~*-Ethanol        10.99875       1.736676    1        2903.340    37.9327       \\<0.0001\n\n  *X~2~*-Time           14.04375       1.736676    1        4733.446    61.8434       \\<0.0001\n\n  *X~3~*-Amplitude      \u22123.994583      1.736676    1        382.961     5.0035        0.0421\n\n  *X~4~*-Ratio          27.390417      1.736676    1        18005.638   235.2474      \\<0.0001\n\n  **Quadratic**                                                                       \n\n  *X~1~* ^2^            \u22121.526438      1.717541    1        60.454      0.7898        0.3892\n\n  *X~2~* ^2^            \u221211.56144      1.717541    1        3468.113    45.3116       \\<0.0001\n\n  *X~3~* ^2^            \u22120.888938      1.717541    1        20.503      0.2679        0.6128\n\n  *X~4~* ^2^            \u221217.87644      1.717541    1        8291.469    108.3279      \\<0.0001\n\n  **Interaction**                                                                     \n\n  *X~1~*--*X~2~*        5.049375       2.187167    1        407.939     5.3298        0.0367\n\n  *X~1~*--*X~3~*        \u221211.46062      2.187167    1        2101.535    27.4570       \\<0.0001\n\n  *X~1~*--*X~4~*        \u221212.58812      2.187167    1        2535.37     33.1252       \\<0.0001\n\n  *X~2~*--*X~3~*        15.196875      2.187167    1        3695.120    48.2775       \\<0.0001\n\n  *X~2~*--*X~4~*        \u22127.198125      2.187167    1        829.008     10.8312       0.0054\n\n  *X~3~*--*X~4~*        \u22126.580625      2.187167    1        692.874     9.0525        0.0094\n\n  **Lack of fit**                                  10       94.0987     4.550         0.0911\n\n  **Pure error**                                   4                                  \n\n  ***R*^2^**                                                            0.9776        \n\n  **Adjusted *R*^2^**                                                   0.9553        \n\n  **C.V. %**            3.71%.                                                        \n\n  **RMSE**              8.7186                                                        \n\n  **CorTotal ^c^**                                 28       48043.545                 \n  -----------------------------------------------------------------------------------------------\n\n^a^ Degree of freedom; ^b^ the model mean square to error mean square ratio; ^c^ corrected total. DF, degree of freedom; F Ratio, freedom ratio; Prob, probability; C.V., coefficient of variation.\n\nantioxidants-08-00205-t003_Table 3\n\n###### \n\nComparison of extraction yield of polyphenols obtained by optimized ultrasound-assisted (UAE-OPT), microwave-assisted (MAE) and conventional solvent (CSE) methods.\n\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Method             EtOH (%)   Time\\   US amp (w/power)   Liq:sol\\   TPC\\                 Flavonoids\\        Anthocyanin (mg/g)   Tannins\\           DPPH\\               RP\\\n                                (min)                      (mL/g)     (mg GAE/g)           (mgQE/g)                                (mg CE/g)          (%)                 (Abs 700 nm)\n  ------------------ ---------- ------- ------------------ ---------- -------------------- ------------------ -------------------- ------------------ ------------------- -------------------\n  **UAE-OPT**        70         7.5     30                 28         241.66 \u00b1 12.77 ^a^   18.99 \u00b1 1.31 ^a^   25.06 \u00b1 0.36 ^a^     35.56 \u00b1 0.36 ^a^   90.71 \u00b1 0.23 ^a^    0.568 \u00b1 0.002 ^b^\n\n  **MAE**            42         62      500                32         119.59 \u00b1 8.40 ^b^    11.5 \u00b1 0.01 ^b^    5.64 \u00b1 0.06 ^c^      31.70 \u00b1 1.00 ^b^   87.16 \u00b1 0.28 ^b^    0.439 \u00b1 0.006 ^b^\n\n  **CSE**            50         7200                       50         76.40 \u00b1 7.27 ^c^     6.95 \u00b1 0.20 ^c^    6.96 \u00b1 0.72 ^b^      30.70 \u00b1 0.17 ^c^   88.03 \u00b1 1.04 ^b^    0.429 \u00b1 0.001 ^b^\n\n  **BHA**\\                                                                                                                                            26.98 \u00b1 0.69 ^c^\\   1.37 \u00b1 0.03 ^a^\\\n  **\u03b1-tocopherol**                                                                                                                                    17.17 \u00b1 0.4 ^d^     0.53 \u00b1 0.01 ^b^\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nContents of TPC, flavonoids, anthocyans and tannins are means \u00b1 standard deviation. Different letters in the same row indicate significant differences (*p* \\< 0.05) according to the ANOVA test. BHA, butylated hydroxyanisole; CE, catechin equivalents; DPPH, 1,1-diphenyl-2-picrylhydrazyl radical; EtOH, ethanol; GAE, gallic acid equivalents; Liq:sol, liquid-to-solid ratio; QE, quercetin; RP, ferric reducing antioxidant power; US amp, ultrasound amplitude.\n\nantioxidants-08-00205-t004_Table 4\n\n###### \n\nUHPLC-DAD-ESI-MS^n^ data for *M. communis* pericarp extract obtained under optimized UAE conditions.\n\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  No.\\      t~R~ (min)                                                           \u03bbmax\\                                 (*m/z*)                 MS^n^ Ions (*m/z*)                                                   Probable Compound\n  Peak                                                                           (nm)                                                                                                                               \n  --------- -------------------------------------------------------------------- ------------------------------------- ----------------------- -------------------------------------------------------------------- -------------------------------------------\n  1         1.4                                                                  275                                   191 ^a^                 MS^2^\\[191\\]: 173, 127, 111, 93                                      Quinic acid\n\n  341 ^a^   MS^2^\\[341\\]: 179                                                    Caffeoyl-*O*-hexoside                                                                                                              \n\n  2         1.9                                                                  276                                   331                     MS^2^\\[331\\]: 271, 169, 241, 211, 193, 125; MS^3^\\[271\\]: 211, 169   Galloyl-*O*-hexoside\n\n  3         2.1                                                                  273                                   343                     MS^2^\\[343\\]: 191, 169, 125                                          Galloyl quinic acid (isomer 1)\n\n  4         2.3                                                                  271                                   169                     MS^2^\\[169\\]: 125                                                    Gallic acid\n\n  5         6.7                                                                  274                                   343                     MS^2^\\[343\\]: 191, 169, 125                                          Galloyl quinic acid (isomer 2)\n\n  6         7.6                                                                  276, 525                              463 ^a,b^               MS^2^\\[463\\]: 301, 300, 337,315                                      Delphinidin-*O*-hexoside\n\n            495 ^a^                                                              MS^2^\\[495\\]: 343, 325, 191, 169      Digalloyl quinic acid                                                                        \n\n  483 ^a^   MS^2^\\[483\\]: 271, 331, 313, 439, 193, 169; MS^3^\\[271\\]: 211, 169   Digalloyl hexoside                                                                                                                 \n\n  7         8.8                                                                  274, 525                              477 ^b^                 MS^2^\\[477\\]: 315, 314                                               Petunidin-*O*-hexoside\n\n  8         9.7                                                                  274, 525                              647 ^a,b^               MS^2^\\[647\\]: 495, 477                                               Petunidin-*O*-galloyl-hexoside derivative\n\n                                                                                                                       491 ^a,b^               MS^2^\\[491\\]: 329                                                    Malvidin-*O*-hexoside\n\n  9         10.3                                                                 265, 356                              631                     MS^2^\\[631\\]: 479, 299, 317                                          Myricetin-*O*-galloyl-hexoside\n\n  10        11.0                                                                 260, 356                              479                     MS^2^\\[479\\]: 316, 317                                               Myricetin-*O*-hexoside (isomer 1)\n\n  11        11.2                                                                 260, 356                              479                     MS^2^\\[479\\]: 316, 317                                               Myricetin-*O*- hexoside (isomer 2)\n\n  12        11.4                                                                 265, 356                              615                     MS^2^\\[615\\]: 463, 301; MS^2^\\[463\\]: 179, 151                       Quercetin *O*-hexoside- gallate\n\n  13        11.8                                                                 263, 356                              449                     MS^2^\\[449\\]: 316, 317                                               Myricetin-*O*-pentoside\n\n  14        12.1                                                                 261, 351                              463                     MS^2^\\[463\\]: 316, 317                                               Myricetin-*O*-deoxyhexoside (isomer 1)\n\n  15        12.2                                                                 261, 351                              463                     MS^2^\\[463\\]: 316, 317                                               Myricetin-*O*-deoxyhexoside (isomer 2)\n\n  16        12.9                                                                 265, 353                              447                     MS^2^\\[447\\]: 285; MS^3^\\[285\\]: 267, 257, 241                       Kaempferol-*O*-hexoside\n\n  17        13.4                                                                 257, 350                              447                     MS^2^\\[447\\]: 301; MS^3^\\[301\\]: 179, 151                            Quercetin-*O*-eoxyhexoside\n\n  18        14.5                                                                 265, 352                              431 ^a^                 MS^2^\\[431\\]: 271; MS^3^\\[271\\]: 211, 169                            Galloyl derivative\n\n  625 ^a^   MS^2^\\[625\\]: 479, 317                                               Myricetin-*O*-hexosyl-deoxyhexoside                                                                                                \n\n  19        14.6                                                                 273, 350                              569                     MS^2^\\[569\\]: 485, 317                                               Myricetin-*O*-galloyl ester\n\n  20        16.5                                                                 276                                   583                     MS^2^\\[583\\]: 271, 565, 313, 211, 331; MS^3^\\[271\\]: 211, 169        Gallomyrtucommulone F\n\n  21        16.8                                                                 276                                   567                     MS^2^\\[567\\]: 271, 313, 211, 169; MS^3^\\[271\\]: 211, 169             Gallomyrtucommulone C\n\n  22        17.6                                                                 276                                   467                     MS^2^\\[467\\]: 271, 313, 169, 211; MS^3^\\[271\\]: 211, 169             Gallomyrtucommulone-type (isomer 1)\n\n  23        17.8                                                                 276                                   467                     MS^2^\\[467\\]: 271, 313, 169, 211; MS^3^\\[271\\]: 211, 169             Gallomyrtucommulone-type (isomer 2)\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nPeak numbers correspond to those depicted in [Figure 3](#antioxidants-08-00205-f003){ref-type=\"fig\"}; ^a^ co-eluted compounds in a peak fraction; ^b^ the respective \\[M\\]^+^ ions were registered in the positive mode. t~R~, retention time; \u03bbmax, wavelength of maximum absorbance.\n"}
+{"text": "1 Introduction\n==============\n\nCircular RNAs (circRNAs) represent an emerging type of regulatory RNA with 3' and 5' ends covalently join together as 'back-spliced junctions'. circ RNAs have been reported to have multiple functions. Beside serving as miRNA sponges, many circRNAs are important for brain function, synaptic plasticity ([@btz705-B44]; [@btz705-B36]; [@btz705-B46]) and fetal development ([@btz705-B42]). Cell free circRNAs are found stable in saliva ([@btz705-B4]) as well as exosomes ([@btz705-B29]), which makes circRNA a promising diagnosis biomarker. Most circRNAs are originated from circularization of coding gene exons, which leads to the hypothesis that circRNA biogenesis competes with pre-mRNA splicing ([@btz705-B2]). Literature evidences suggest that the biogenesis of each circRNA subset may likely be regulated by different mechanisms ([@btz705-B9]; [@btz705-B11]; [@btz705-B19]; [@btz705-B20]; [@btz705-B28]; [@btz705-B30]; [@btz705-B49], [@btz705-B47], [@btz705-B48]), which supports the existence of multiple subclasses of circRNAs and each with specific roles.\n\nIn this study, we aim to decipher the relationship between the sequence features and circRNA subclasses. To this end, we have developed a computational model to predict whether a genomic locus would generate circRNA based on the presence of the following sequence features within the flanking regions of the circular RNA back-spliced junction sites: CpG islands, enhancers, RNA binding protein (RBP) binding sites, simple repeats, RNA editing sites and DNA self-chains. These features were selected based on the hypothesized circRNA biogenesis mechanisms ([@btz705-B11]; [@btz705-B19]; [@btz705-B20]; [@btz705-B28]; [@btz705-B30]; [@btz705-B49], [@btz705-B47], [@btz705-B48]) as well as sequence features that can potentially participate in biogenesis of non-coding RNAs including CpG islands ([@btz705-B25]) and enhancers regions ([@btz705-B8]; [@btz705-B26]). We have designed a contextual regression model ([@btz705-B31]) that successfully distinguish the circRNA back-spliced junction sites defined by transcriptome sequencing from randomly selected human genome loci in an averagely 72.6% accuracy. Using the feature extraction technique in the contextual regression model ([@btz705-B23]), we found that the examined 21 427 circRNAs can be categorized into 7 groups based on the biogenesis contributing factors. Our analysis supports that multiple biogenesis mechanisms co-exist for different subset of human circRNAs. In particular, we found 79 RBPs were identified to be significantly correlated to circRNA biogenesis. Interestingly, we uncovered a potential new link between circRNA biogenesis and flanking CpG islands, which suggests the potential correlation between DNA methylation and circRNA biogenesis.\n\n2 Materials and methods\n=======================\n\nThe data analysis process of this research is summarized in [Supplementary Figure S1](#sup1){ref-type=\"supplementary-material\"}. First, 55 689 human circRNAs back-spliced junction sites were collected from the database CircNet ([@btz705-B32]) as positive training data, and equal amount of randomly selected locus on HG19 human genome as negative training data. These junction sites and randomly selected locus were then divided into training and testing sets (in a ratio of 7:3) for the contextual regression network model designed to predict whether a randomly selected locus from human genome would generate circRNA. The features of the training set include whether CpG islands, enhancer regions, RBP binding sites, simple repeats, A-to-I RNA editing sites (RNA editing sites for short in the later text) and DNA self-chains present in the upstream and downstream region of the selected locus. After reaching optimum average accuracy, through the application of feature extraction techniques ([@btz705-B23]), we successfully found that the examined 21 427 circRNAs can be categorized into 7 groups based on the presumed biogenesis contributing factors.\n\nThe back-spliced junction sites in the positive training set were selected from the database CircNet ([@btz705-B32]). The data include reported human back-spliced junction sites were collected from 22 recent studies ([@btz705-B1]; [@btz705-B3]; [@btz705-B4]; [@btz705-B5]; [@btz705-B10]; [@btz705-B11]; [@btz705-B12]; [@btz705-B13]; [@btz705-B15]; [@btz705-B20]; [@btz705-B21]; [@btz705-B33]; [@btz705-B36]; [@btz705-B37], [@btz705-B38]; [@btz705-B40]; [@btz705-B49], [@btz705-B47], [@btz705-B48]; [@btz705-B50]) and 465 human transcriptome sequencing datasets were collected from NCBI Sequence Read Archive ([@btz705-B27]). The back-spliced junction sites in each RNA-seq sample were identified using a circRNA discovery pipeline referred as find_circ ([@btz705-B14]; [@btz705-B16]; [@btz705-B33]). The criteria defined in the pipeline hence the detected junction sites met same standards as those in the previous reports, as described in the [@btz705-B33] study was applied. Adhering to the suggestion of recent year comparison study ([@btz705-B16]), we selected the back-spliced junction sites based on the number of previous peer review reports in which the back-spliced junction sites were reported and the number of samples among the 465 collected samples in which the back-spliced junction sites were found meeting the criteria defined in find_circ ([@btz705-B14]; [@btz705-B16]; [@btz705-B34]). Only the circRNAs with the sum of these two numbers \\> 3 were considered as positive training data in this study. Locus of the CpG islands, enhancer regions, simple repeats, RNA editing sites and DNA self-chains on HG19 human genome was collected from UCSC Genome Browser ([@btz705-B7]). Although the locus of the RBP binding sites was collected from the [@btz705-B35] study.\n\n3 Results\n=========\n\nUsing the feature extraction technique in the contextual regression model ([@btz705-B31]), we found that the examined 21 427 circRNAs can be categorized into seven groups based on the biogenesis contributing factors. Our analysis supports that multiple biogenesis mechanisms co-exist for different subset of human circRNAs. In particular, we found 79 RBPs were identified to be significantly correlated to circRNA biogenesis. Interestingly, we uncovered a potential new link between circRNA biogenesis and flanking CpG islands, which suggests the potential correlation between DNA methylation and circRNA biogenesis.\n\n3.1 An interpretable neural network model to predict circRNAs\n-------------------------------------------------------------\n\nWe implemented a contextual regression model to predict circRNA using these features. Instead of letting the neural network learn a function that maps features to target values, our method let the neural network learn a function that maps features to local linear models that best predict the target value from the features, thus generating a model that can both achieve state-of-the-art accuracy like a deep neural network while giving human interpretable quantification of feature contribution ([Fig.\u00a01A](#btz705-F1){ref-type=\"fig\"}). As summarized in [Supplementary Figure S1C](#sup1){ref-type=\"supplementary-material\"}, in this contextual regression model, we used a residual neural network ([@btz705-B17]) that is composed of three layers of FNN (feed forward neural network, [Fig.\u00a01B](#btz705-F1){ref-type=\"fig\"}) as the embedding function to generate the linear models. Batch normalization ([@btz705-B18]) is applied in the input layer to reduce the variance between input batches and make the training process more stable. The output of the embedding function is then dot-producted with the features and fed into a logistic function to output the prediction result. The FNN model is implemented as an operation that maps a vector x to s(Ax+b) where A is an matrix, b is an vector and s is the activation function. Both A and b are first initialized and then trained with tensorflow AdamOptimizer ([@btz705-B24]). As for the parameter setting, all three layers of the FNN have 10 hidden units and tanh as their activation function, batch size is set to 50, max gradient norm to 10, learning rate to 0.0001. The weight matrix of each neural network is initialized with the tensorflow tf.truncated_normal function with standard deviation of 0.05 to prevent vanishing gradient problem. The bias term in each layer is initialized all to value 0. We used cross-entropy as the loss function during training. To make the feature contribution easily interpretable, we applied a Lasso penalty in the form of L1 regularization on the context weight with penalty coefficient 0.0001.\n\n![Illustration of contextual regression. (**A**) Graphic illustration of the mechanism of contextual regression: the features are inputted into a neural network that generates a contextual weight for each feature which represents the importance of the features. Then, the features are then weighted by the corresponding weights to makes an easier separation of samples. In classification or regression tasks, the weighted features are then summed to yield the prediction. (**B**) A graphic demonstration of the FNN parts in the contextual regression model](btz705f1){#btz705-F1}\n\n3.2 Prediction and the feature selection results\n------------------------------------------------\n\nAs summarized in [Supplementary Table S1](#sup1){ref-type=\"supplementary-material\"}, the designed contextual regression model successfully distinguish circRNA back-spliced junction sites defined by transcriptome sequencing data from randomly selected human genome loci, in an average 72.6% accuracy and the area under curve of ROC curve 0.801 ([Supplementary Fig. S2](#sup1){ref-type=\"supplementary-material\"}).\n\nTo extract the informative features, we selected the run with the highest accuracy (Run no. 5) from 10 runs. The difference between the accuracy on the training and testing sets were negligible (72.5 versus 72.8%), which suggested no overfitting. Therefore, we pooled together the training and testing sets in the feature analysis. We selected the most confidently predicted 21 427 circRNAs that have confident scores \\> 0.7 to evaluate the contribution of each feature. The weighted feature contribution (WFC) was then obtained from the model output. The features for each circRNA genesis mechanism (except 'flanking short ALU repeats' (both_have_Alu) and 'binding sites of single RBP' (if_same_RBP) since they only have 1 value for the whole region) were pooled into short (within 1000 bp radius from the circRNA) and long range (1000--2000 bp from the circRNA) by summing the WFC in each category for better data visualization. This resulted in 14 total feature contributions for each circular RNA data point (CpG short range, CpG long range, enhancer short range, enhancer long range, RBP short range, RBP long range, repeats short range, repeats long range, RNA editing short range, RNA editing long range, self-chain short range, self-chain long range, if head and tail both have Alu, if head and tail have the same RBP) that could be treated as a vector with 14 elements. Then, each of these vectors was normalized to the unit length and PCA was applied to the whole data point collection. The top 10 principal components were extracted which explained 98.9% of the total variance. Then a K-mean clustering was applied to separate the processed data into subclasses. Multiple values of *k* were experimented and *k* = 7 was chosen for being the largest *k* that ensured all cosine similarities between cluster centers are \\< 0.5. The significantly contributing features were plotted in the heat map format and confirmed in the original feature data.\n\nAs a result, we found that these circRNAs could be clustered into seven different categories according to their biogenesis factors ([Fig.\u00a02A](#btz705-F2){ref-type=\"fig\"}). To validate that the features with high contribution scores are enriched in each cluster, we calculated the percent enrichment of each feature in each cluster ([Fig.\u00a02B](#btz705-F2){ref-type=\"fig\"}). The enrichment plot supported our clustering result.\n\n![Prediction and feature collection result. The result of the feature collection is summarized in this figure. (**A**) Through the result we found that these circRNAs can be into seven different categories according to their biogenesis factors. The long range features are marked with '\\_l' and the short range ones are marked with '\\_s'. (**B**) Percentage of members in each cluster that contains each of the features](btz705f2){#btz705-F2}\n\nIn Cluster 0, RNA editing sites occurrence within 1000 nucleotide upstream or downstream of the circRNA locus was considered to be the most important factor for the biogenesis of 4576 circRNAs. For the other 4585 circRNAs in Cluster 1, appearance of RNA editing site within range of 1000--2000 nucleotides upstream or downstream of circRNA was considered as the most important biogenesis factor. Similarly, in Cluster 2, existence of RBP binding sites within short, which means within 1000 nucleotides upstream or downstream of the circRNA locus, was considered as the most important biogenesis factor for the 7503 circRNAs. For the 2342 circRNAs in the Cluster 4, occurrence in the long range, which means within range of 1000--2000 nucleotides upstream or downstream of the circRNA locus, of RBP binding sites was suggested as the main biogenesis factor. Short repeat sequence flanking circRNA locus was clustered as the main factor for the 1344 circRNAs in Cluster 3. Biogenesis of 620 circRNAs was linked to flanking CpG islands, whereas 457 circRNAs' biogenesis mechanism appeared to relate to the flanking DNA self-chains. The amount of these seven clusters is summarized in [Supplementary Figure S3](#sup1){ref-type=\"supplementary-material\"}, while a complete list of the circRNAs is available in [Supplementary Additional File S1](#sup1){ref-type=\"supplementary-material\"}. Through examining the input data that forms Clusters 2 and 4, which were associated with RBP, we identified 79 RBPs presumably participate in the biogenesis of these 9845 circRNAs. A complete list of the circRNA locus and associated RBP is available in [Supplementary Additional File S2](#sup1){ref-type=\"supplementary-material\"}. These 79 RBPs (available in [Supplementary Additional File S4](#sup1){ref-type=\"supplementary-material\"}) appear both in the short range (within 1000 nucleotides) in Cluster 2 and long range (from 1000 nucleotides to 2000 nucleotides) in Cluster 4 consistently. Among these RBPs, binding motifs of SRSF1, PUM1, SF3B4, ELAVL1, HNRNPA1, CSDA, SNRPA, RBFOX1 and PABPC1 were found appear in upstream or downstream of thousands of circRNAs in both clusters, as summarized in [Supplementary Table S2](#sup1){ref-type=\"supplementary-material\"}.\n\nHence based on the result of this study, circRNAs can be categorized into multiple different subclasses, each with specific different function and biogenesis mechanism. Result of this research support all of previous discoveries and solidify the idea that multiple biogenesis mechanisms co-exist for different subset of human circRNAs. Result of this study can also contribute to the advance of circRNA detection algorithm. Current mainstream research methods adopted to discover circRNAs are based on detection of back-spliced junction sites spanning reads within transcriptome sequencing data. However, this kind of approaches tends to have high false positive rate. Sensitivity of the junction site detection is also limited by sequencing depth. Had a clear concept of circRNA biogenesis mechanism is available, improved algorithm ruling out sequence base false positive might be able to be developed.\n\nSupplementary Material\n======================\n\n###### \n\nClick here for additional data file.\n\nWe would like to thank Dr Shu Chien for his helpful comments and discussions.\n\nFunding\n=======\n\nThis work has been supported by the National Institutes of Health R01 \\[grant HG009626\\].\n\n*Conflict of Interest*: none declared.\n\n[^1]: The authors wish it to be known that, in their opinion, Chengyu Liu and Yu-Chen Liu authors should be regarded as Joint First Authors.\n"}
+{"text": "Introduction\n============\n\nGastric cancer remains a significant worldwide health problem, as the fourth most common type of cancer and the second leading cause of cancer-related mortality worldwide (one million diagnoses of stomach cancer were made in 2008, with 740,000 related fatalities) ([@b1-mmr-10-06-2898]--[@b4-mmr-10-06-2898]). Although there has been a reduction in stomach cancer incidence in multiple countries, early detection remains the key to a better prognoses. However, in the early stages of gastric cancer, the majority of patients are asymptomatic and thus patients are commonly diagnosed at an advanced stage, leading to a low five-year survival rate (\\<10%) ([@b4-mmr-10-06-2898]). The etiology of gastric cancer, similar to the majority of other types of cancer, remains to be defined, and the susceptibility of the individual to cancer may be altered by a combination of factors, including lifestyle and age, in addition to environmental and genetic aspects ([@b5-mmr-10-06-2898]). For example, consumption of nitrate- or nitrite-rich food (grilled, salted or pickled foods) ([@b6-mmr-10-06-2898]), presence of *Helicobacter pylori* infection ([@b7-mmr-10-06-2898]), an age of \\>60 years and a history of stomach disorders or gastric cancer, have been reported to be possible variables that can lead to gastric cancer ([@b8-mmr-10-06-2898]). By contrast, vitamin C, carotenoids and green tea have been implied to have preventive effects in gastric cancer ([@b9-mmr-10-06-2898]). Furthermore, genetic susceptibility has been extensively investigated as an important contributor to inter-individual variation of gastric cancer risk ([@b10-mmr-10-06-2898]). Accumulation of genetic and epigenetic changes (such as mutation and hypermethylation of tumor suppressor genes) has been confirmed to be involved in the development and progression of gastric cancer. A number of these genes, including p53, APC and c-erbB-2, are not gastric tissue-specific. The gastric tissue-specific genes may serve an essential role in the development and progression of gastric cancer. Thus, investigation of these genes may be useful to improve the understanding of the pathogenesis of gastric cancer, and to develop novel treatments.\n\nThe novel gastrointestinal tract-specific gene GDDR is abundantly expressed in normal gastric mucosae, but is downregulated or completely knocked out in gastric cancer ([@b11-mmr-10-06-2898]). GDDR was originally cloned in our laboratory in 2002, by suppression-subtractive hybridization between the gastric carcinoma tissues and corresponding normal gastric mucosae and the ends-Marathon rapid amplification of cDNA ends ([@b11-mmr-10-06-2898]). GDDR is a stomach-specific secreted protein and is a member of the gastrokine gene family. The GDDR protein is well-conserved and contains one BRICHOS domain with a pair of conserved cysteine residues, and is proposed to function in folding and intracellular transport or secretion ([@b12-mmr-10-06-2898]). It possesses similarities to another gastric foveolar protein termed gastrokine-1 ([@b13-mmr-10-06-2898]), thus GDDR has been renamed gastrokine-2 ([@b14-mmr-10-06-2898]). Functionally, gastrokine-2 protein is involved in the replenishment of the surface lumen epithelial cell layer and maintenance of the mucosal integrity. Previous studies have demonstrated that expression of gastrokine-2 inhibits the proliferation of gastric cancer cells ([@b15-mmr-10-06-2898]) and the progression of gastric cancer *in vivo*, in a trefoil factor 1-dependent manner ([@b16-mmr-10-06-2898],[@b17-mmr-10-06-2898]). Thus, in the present study, the loss of expression of gastrokine-2 protein in human gastric cancer tissue samples was confirmed, and then a functional-grade purified anti-human CD95 (APO/Fas) antibody was used to activate, and an anti-Fas (human, neutralizing, clone ZB4) antibody was used to block the extrinsic pathway following transfection of gastrokine-2. The effects of gastrokine-2 protein on the regulation of gastric cancer cell viability and the underlying mechanism were investigated.\n\nMaterials and methods\n=====================\n\nTissue samples\n--------------\n\nA total of 76 cancer and corresponding normal gastric tissues were collected from the Department of Gastrointestinal Cancer (The Drum Tower Clinical College of Nanjing Medical University, Nanjing, China) between November 2011 and June 2012. The clinicopathological characteristics of the patients with gastric carcinoma are outlined in [Table I](#tI-mmr-10-06-2898){ref-type=\"table\"}. All patients were pathologically confirmed to have gastric adenocarcinoma. The current study was approved by The Ethics Committee of The Drum Tower Clinical College of Nanjing Medical University. All patients or their legal guardians signed an inform consent form prior to participation in the study. Fresh tissue samples were obtained, snap-frozen using liquid nitrogen and stored at \u221280\u00b0C until use.\n\nCell lines and culture\n----------------------\n\nThe SGC-7901 and AGS human gastric cancer cell lines were purchased from the Shanghai Institute of Cell Biology at the Chinese Academy of Sciences (Shanghai, China) and cultured in RPMI-1640 medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (HyClone Laboratories, Logan, UT, USA), 1\u00d710^5^ U/l penicillin and 100 mg/l streptomycin (CC033, Zhongke, Beijing, China)at 37\u00b0C in a humidified atmosphere with 5% CO~2~.\n\nConstruction of expression vector and gene transfection\n-------------------------------------------------------\n\nHuman GDDR cDNA (Invitrogen, Carlsbad, CA, USA) was cloned into *Bam*HI/*Eco*RI restriction sites of the eukaryotic expression vector pcDNA3.1/Myc-His(+) (Invitrogen). Specifically, a primer (5\u2032-GGAATTCTAATGAAAATACTTGTGGCAT-3\u2032) containing a *Bam*HI linker in front of the initial GKN1 Met and 5\u2032-CGGGATCCAACATGAATGTCTGCACAGA-3\u2032 that abolished the GDDR stop codon for PCR amplification of the GDDR open reading frame were used. This PCR amplicon was then cloned into a pcDNA3.1/Myc-His(+) vector. Following sequence confirmation, the vector was termed pcDNA-GDDR. For gene transfection, the cells were subcultured and grown to the logarithmic growth phase then transiently transfected with pcDNA-GDDR or pcDNA3.1 (control) using Lipofectamine 2000 (Invitrogen), according to the manufacturer's instructions. The transfection efficiency was evaluated by a parallel transfection using an EGFP vector (Invitrogen).\n\nReverse transcription-polymerase chain reaction (RT-PCR)\n--------------------------------------------------------\n\nSGC-7901 cells were divided into the control (Con), control vector-transfected (P) and GDDR cDNA-transfected (G) groups. At the end of experiments, total RNA (20--50 \u03bcg) was extracted from SGC-7901 human gastric cancer cells using TRIzol reagent (Invitrogen), and it was reverse transcribed into cDNA using an RNA PCR kit (DRR036A; Takara Bio, Inc., Otsu, Japan), according to the manufacturer's instructions. These cDNA samples were amplified by PCR using a thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) with the following conditions: Initial denaturation at 94\u00b0C for 30 sec; followed by 40 cycles of 95\u00b0C for 5 sec, 65\u00b0C for 30 sec and 72\u00b0C for 30 sec; and a final extension at 72\u00b0C for 10 min. PCR fragments were separated by electrophoresis on a 1.5% agarose gel and visualized with ethidium bromide. Primer sequences (Invitrogen) were as follows: Forward: 5\u2032-GACCCCTTCATTGACCTCAACTACA-3\u2032 and reverse: 5\u2032-GTCCACCACCCTGTTGCTGTAGCCA-3\u2032 for GAPDH; forward: 5\u2032-GTGGCATTTTGGTGGTG-3\u2032 and reverse: 5\u2032-CATTGTTGCTTGGGCTGA-3\u2032 for GDDR; forward: 5\u2032-AGACTGCGTGCCCTGCCAAGA-3\u2032 and reverse: 5\u2032-GGC CTGCCTGTTCAGTAACT-3\u2032 for Fas; forward: 5\u2032-GAGACA GCCAGGAGAAATCA-3\u2032 and reverse: 5\u2032-CCTGTGGAT GACTGAGTA-3\u2032 for bcl-2; and forward: 5\u2032-GACCCGGTG CCTCAGGATGC-3\u2032 and reverse: 5\u2032-GTCTGTGTCCAC GGCGGCAA-3\u2032 for Bax.\n\nProtein extraction and western blot analysis\n--------------------------------------------\n\nSGC-7901 cells were prepared as the Con, P and G groups. At the end of the experiments, total cellular protein was extracted from tissue specimens and gastric cancer cells using a lysis buffer containing 1X Protease Inhibitor Cocktail (Roche Diagnostics GmbH, Mannheim, Germany). Protein concentration was quantified using the Bicinchoninic Protein Assay kit (KeyGEN, China). Equal quantities of protein samples were resolved by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels and electroblotted onto polyvinylidene fluoride membranes (Millipore, Billerica, MA, USA). The membranes were then blocked in 5% non-fat milk overnight, and the following day, membranes were incubated with a rabbit polyclonal anti-GDDR (ab70480, Abcam, Cambridge, UK), rabbit monoclonal anti-Fas (5709-1, Epitomics, Inc., CA, USA), anti-bcl-2 (BS1511, Bioworld, St. Louis, MN, USA), anti-Bax (BS2538 Bioworld) or anti-GAPDH (BSAP0063 Bioworld) for 4 h. Following washing with phosphate-buffered saline (PBS) with Tween-20 four times, and incubation with goat anti-rabbit secondary antibody (GAR0072, LiankeBio, Hangzhou, China) for 2 h at room temperature, the protein bands were visualized using an enhanced chemiluminescence kit (EMD Millipore, Billerica, MA, USA).\n\nFlow cytometry\n--------------\n\nSGC-7901 cells were prepared as Con, P and G groups, and then subjected to evaluation of Fas (also known as CD95) receptor expression. Briefly, cells (1--5\u00d710^5^/100 \u03bcl) were scraped, subsequent to trypsin digestion without EDTA addition, washed twice with ice-cold PBS and the binding buffer, resuspended in the presence of an anti-human CD95 (APO-1/Fas) phycoerythrin (PE) antibody (12-0959, eBioscience, Inc., San Diego, CA, USA) and incubated in the dark for 30 min. The cell suspension was then washed with the binding buffer and resuspended in 200 \u03bcl binding buffer. For each sample, 2\u00d710^4^ events were acquired by a CantoTM Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA). The experiments were conducted in triplicate and repeated three times.\n\nTumor cell viability assay\n--------------------------\n\nSGC-7901 cells were prepared as Con, P and G groups. A G+F group was created by coincubation of G group cells with a functional grade purified anti-human CD95 (APO/Fas; Epitomics, Burlingame, CA, USA) antibody at 5 mg/ml for 24 h. In the G+F+Z group, cells underwent the 48-h GDDR vector transfection plus coincubation with the CD95 (APO/Fas) antibody and an anti-Fas (human, neutralizing, clone ZB4 at 1 mg/ml; Merck Millipore, Darmstadt, Germany)\\] antibody. These cells were seeded into 96-well plates at 5\u00d710^3^ cells/well and grown for up to 72 h. At the end of the experiments, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; KeyGEN, Nanjing, China) at 100 \u03bcg/well was added to the cell culture, and the cells were incubated for another 4 h. A volume of 150 \u03bcl dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO, USA) was added to each well subsequent to removal of the supernatant. After shaking the plate for 20 min on a shaking board, cell viability was assessed by measuring the absorbance at 490 nm using an enzyme-labeling instrument (680 model; Bio-Rad Laboratories, Hercules, CA, USA). The experiments were conducted in quintuplicate and repeated three times. Growth inhibition (IR%) was calculated according to the following formula: IR% = \\[(the absorbance of blank control group-the absorbance of experimental group)/the absorbance of blank control group\\] \u00d7 100.\n\nAnnexin V-fluorescein isothiocyanate (FITC) apoptosis assay\n-----------------------------------------------------------\n\nAn Annexin V-FITC Apoptosis Detection kit with propidium iodide (eBioscience) was used to detect apoptosis. In brief, SGC-7901 cells were prepared as the Con, P and G groups. At the end of experiments, cells were harvested by centrifuging at 2,400 \u00d7 g for 5 min, washed once in PBS, then once in 1X binding buffer, pelleted and resuspended at a concentration of 1\u00d710^6^ in 100 \u03bcl of 1X binding buffer. A volume of 5 \u03bcl Annexin V-FITC was added to the cell solution, followed by incubation for 15 min at room temperature. It was then pelleted, washed with 1X binding buffer, and resuspended in 200 \u03bcl of 1X binding buffer. Next, 5 \u03bcl propidium iodide solution was added to the cells for a 15-min incubation at room temperature in the dark followed by the addition of 300 \u03bcl of 1X binding buffer. A minimum of 10,000 cells were subjected to flow cytometric analysis of the viable, apoptotic and necrotic cell populations. The results were quantified using Cell Quest software with FCS 2.0 files (Flowjo 7.6.5.1, BD Biosciences), according to the manufacturer's instructions.\n\nQuantitation of caspase-3, -8 and -9 activity\n---------------------------------------------\n\nSGC-7901 cells were prepared as the Con, P and G groups. At the end of the experiments, caspase activity was then measured using CaspGLOW Fluorescein Active Caspase-3, CaspGLOW Red Active Caspase-9 and CaspGLOW Red Active Caspase-8 Staining kits (\\#K183, \\#K199 and \\#K198, respectively; BioVision, Inc., Milpitas, CA, USA), according to the manufacturer's instructions. Briefly, cells were resuspended in 300 \u03bcl complete growth medium at a concentration of 1\u00d710^6^/ml, and incubated in a 37\u00b0C incubator for 45 min with the anti-caspase-3, -8 and -9 antibodies. The lysate was centrifuged at 4,800 \u00d7 g for 5 min at 4\u00b0C, washed twice with the ice-cold wash buffer and the activity of caspase-3, -8 and -9 measured using the substrate peptides from the staining kits (FITC-DEVD-FMK, Red-IETD-FMK and Red-LEHD-FMK). The caspase activity was quantified by determining absorbance with the Multiskan Spectrum spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at Ex/Em = 540/570 nm. Analyses were performed in triplicate with at least three separate experiments.\n\nStatistical analysis\n--------------------\n\nAll experimental data were obtained from at least three independent experiments. The results are expressed as the mean \u00b1 standard deviation and were evaluated using one-way analysis of variance followed by Student's t-test. Statistical analysis was performed using the SPSS 13.0 (SPSS, Inc., Chicago, IL, USA) for Windows software. P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\nExpression of gastrokine-2 in human gastric tissues and gastric cancer cell lines\n---------------------------------------------------------------------------------\n\nGastrokine-2 expression was analyzed in 76 primary gastric cancer and corresponding normal tissues using western blot analysis. It was demonstrated that gastrokine-2 protein expression was reduced in 58 (84.0%) of the 76 cancer tissue samples compared with the corresponding gastric mucosal tissue samples ([Fig. 1](#f1-mmr-10-06-2898){ref-type=\"fig\"}). Specifically, gastrokine-2 expression was reduced in 19 (73.07%), 32 (82.05%) and 7 (63.64%) of the 26 diffuse-, 39 intestinal- and 11 mixed-type gastric cancer samples, respectively. Expression of gastrokine-2 protein was indicated to be significantly lower in *H. pylori*-positive patients than the level in *H. pylori*-negative subjects (P\\<0.05; [Table I](#tI-mmr-10-06-2898){ref-type=\"table\"}), however, gastrokine-2 protein expression was not associated with tumor location, depth of invasion, lymph node metastasis, Lauren's classification or tumor stage (P\\>0.05).\n\nGastrokine-2 expression was then analyzed in the two gastric cancer cell lines, and it was demonstrated its expression was absent in the SGC-7901 and AGS cells ([Fig. 1](#f1-mmr-10-06-2898){ref-type=\"fig\"}).\n\nRestoration of gastrokine-2 expression in SGC-7901 gastric carcinoma cells\n--------------------------------------------------------------------------\n\nTo determine the role of gastrokine-2 in gastric cancer cells, pcDNA3.1-GDDR or control pcDNA31 were transiently transfected into SGC-7901 cells. The results demonstrated that pcDNA3.1-GDDR restored gastrokine-2 expression levels in gastric cancer cells ([Fig. 2A](#f2-mmr-10-06-2898){ref-type=\"fig\"}). The altered gene expression was then assessed, and it was indicated that the level of Fas mRNA was significantly upregulated 48 h after gene transfection (P\\<0.05 vs. non-transfected control and vector control; [Fig. 2A](#f2-mmr-10-06-2898){ref-type=\"fig\"}). Fas protein level was also increased, as detected by western blot analysis ([Fig. 2B](#f2-mmr-10-06-2898){ref-type=\"fig\"}) and flow cytometry ([Fig. 3](#f3-mmr-10-06-2898){ref-type=\"fig\"}) with a rabbit monoclonal anti-Fas/CD95 and anti-human CD95 (APO-1/Fas) PE ([Fig. 3](#f3-mmr-10-06-2898){ref-type=\"fig\"}). However, expression of bcl-2 and Bax mRNA and protein was not identified to significantly change from control levels.\n\nRestoration of gastrokine-2 expression reduces tumor cell viability in vitro\n----------------------------------------------------------------------------\n\nFollowing transfection, the altered phenotypes of these gastric cancer cells was evaluated. A cell viability MTT assay was performed, and the results indicated that restoration of gastrokine-2 expression significantly reduced tumor cell viability in the monolayer culture. In brief, the inhibitory rate of G+F was 35.67\u00b15.76 and 58.67\u00b11.78% at 48 and 72 h, respectively. The P and G groups displayed reduced viability of 0.97\u00b13.71 and 3\u00b13.86%, and 13.69\u00b12.29 and 7.72\u00b15.28%, respectively (P\\<0.05). Additionally, the viability of G+F+Z cells was reduced compared with G+F cells (10.46\u00b10.78 vs. 7.14\u00b13.00% at 48 and 72 h, respectively; P\\<0.05; [Fig. 4](#f4-mmr-10-06-2898){ref-type=\"fig\"}).\n\nRestoration of gastrokine-2 expression induces apoptosis in gastric cancer cells\n--------------------------------------------------------------------------------\n\nTo assess the cause of the reduced cell viability, the rate of apoptosis was evaluated. Following 48-h gastrokine-2 transfection, SGC-7901 gastric cancer cells were incubated with functional grade purified anti-human CD95 (APO/Fas) for 24 h. The rate of apoptosis following antibody incubation was 45.89\u00b18.20%, which was significantly higher than the level in cells transfected with gastrokine-2 vector (15.48\u00b17.53%), control vector (12.97\u00b11.99%), and non-transfected controls (5.24\u00b13.71) (P\\<0.05). However, when the cells were coincubated with the two antibodies (CD95 and Fas; G+F+Z cells), the apoptosis rate was 21.71\u00b16.90%, which was significantly reduced compared with the G+F cells (P\\<0.05; [Fig. 5](#f5-mmr-10-06-2898){ref-type=\"fig\"}).\n\nRestoration of gastrokine-2 expression induces activation of caspase-3, -8, and -9\n----------------------------------------------------------------------------------\n\nTo further assess the effect of gastrokine-2 restoration on the induction of apoptosis, the activity of caspase-3, -8 and -9 was determined. The data demonstrated that the relative activity of caspase-3 (7.5\u00b11.04) and caspase-8 (3.09\u00b10.49) was significantly higher in the G+F group compared with cells of the G group (3.58\u00b10.57 and 1.58\u00b10.26, caspase-3 and -8, respectively; P\\<0.05) and parental cell control (1.00\u00b10.12 and 1.00\u00b10.18 for caspase-3 and -8, respectively; P\\<0.01). The relative activity of caspase-3 (4.03\u00b10.55) and caspase-8 (2.23\u00b10.24) was lower in the G+F+Z group compared with the G+F group ([Fig. 6](#f6-mmr-10-06-2898){ref-type=\"fig\"}). Furthermore, the activity levels of caspase-9 were 1.00\u00b10.05, 1.03\u00b10.11, 1.12\u00b10.11, and 1.04\u00b10.17 in the control, G, G+F and G+F+Z groups, respectively, indicating no significant differences (P\\>0.05; [Fig. 6](#f6-mmr-10-06-2898){ref-type=\"fig\"}).\n\nDiscussion\n==========\n\nApoptosis, also known as programmed cell death, is a basic biological process that functions to maintain homeostasis of the human body by removing undesirable cells ([@b18-mmr-10-06-2898]). Apoptosis is controlled by a diverse range of cell signals, which can be classified into two major molecular signaling pathways; the extrinsic and intrinsic pathways ([@b19-mmr-10-06-2898]--[@b22-mmr-10-06-2898]). The extrinsic apoptotic pathway involves binding of the Fas ligand (FasL) to the Fas receptor (FasR; also termed CD95) ([@b23-mmr-10-06-2898],[@b24-mmr-10-06-2898]), a transmembrane protein of the tumor necrosis factor family. This results in formation of the death-inducing signaling complex, which contains the Fas-associated death domain (FADD), caspase-8 and caspase-10. FADD is an adapter complex that recruits and activates caspase-8. Cleaved caspase-8 then induces cleavage and activation of executive caspase-3, and in turn, the activated capase-3 cleaves DNA molecules, leading to apoptosis ([@b25-mmr-10-06-2898]--[@b27-mmr-10-06-2898]). Alternatively, the intrinsic (or mitochondrial) pathway is largely dependent on the bcl-2 family of proteins (such as Bax) to induce cytochrome *c* release from the mitochondria. Cytochrome *c* binds to apoptotic protease activating factor-1, ATP and pro-caspase-9 to form a protein complex known as an apoptosome, in order to activate caspase-3 for induction of apoptosis. Different stimuli activate one of these apoptotic pathways, or both ([@b23-mmr-10-06-2898]--[@b27-mmr-10-06-2898]).\n\nPreviously, it has been demonstrated that the Fas/FasL pathway exerts a central role in induction of apoptosis, and alteration of this pathway has been observed in gastric adenocarcinoma cells ([@b28-mmr-10-06-2898]). Gastric cancer tissues also indicated a downregulation of Fas, but increased FasL expression. Indeed, downregulation of Fas receptor expression in cancer cells can lead to apoptosis resistance and FasL stimulation ([@b29-mmr-10-06-2898],[@b30-mmr-10-06-2898]). However, increased expression of FasL and reduced expression of caspase-3 in gastric cancer cells of the primary foci serve an important role in gastric carcinogenesis ([@b27-mmr-10-06-2898]). FasL has also been implicated in de-differentiation, growth, invasion and metastasis of gastric cancer cells, through the induction of apoptosis in the infiltrating lymphocytes. By contrast, chemical substances derived from the primary foci of gastric cancer tissues and the metastatic microenvironment may inhibit the growth of metastatic cells by enhancing caspase-3 expression levels and decreasing those of FasL ([@b27-mmr-10-06-2898]).\n\nIn the present study, the level of gastrokine-2 protein was reduced, or absent, in the majority of gastric cancer tissues and absent in two gastric cancer cell lines, which is consistent with the results reported by Du *et al* ([@b11-mmr-10-06-2898]). Previous studies have not implicated gastrokine-2 as a putative gastric cancer-specific tumor suppressor gene ([@b11-mmr-10-06-2898]). However, other studies have demonstrated that gastrokine-2 is a secretory peptide of human gastric surface mucous cells ([@b31-mmr-10-06-2898]) and modulates gut epithelial cell proliferation ([@b32-mmr-10-06-2898]). Gastrokine-2 expression has been reported to be attenuated in gastric adenocarcinomas (85% of diffuse and 54% intestinal type tumors) ([@b33-mmr-10-06-2898]), whilst in gastric epithelial cells it has been indicated to be significantly upregulated following eradication of *Helicobacter pylori*, a risk factor for gastric cancer ([@b34-mmr-10-06-2898]). These data indicated that gastrokine-2 may serve an important function in gastric epithelial cell homeostasis and that altered expression of gastrokine-2 protein may contribute to gastric carcinogenesis. Gastrokine-1, another member of the gastrokine family, has been demonstrated to introduce apoptosis in gastric cancer cells mainly through the Fas/FasL pathway ([@b35-mmr-10-06-2898]).\n\nThe current study also demonstrated that restoration of gastrokine-2 protein expression upregulated Fas expression, but there was no significant difference in the expression level of bcl-2 and Bax, indicating that the extrinsic apoptosis pathway serves a role in gastrokine-2-induced gastric cancer cell apoptosis. To verify this, a functional grade purified anti-human CD95 (APO/Fas) antibody was used to promote apoptosis, and an anti-Fas (human, neutralizing, clone ZB4) antibody was used to block this extrinsic pathway. The data indicated that apoptosis was markedly increased in gastric cancer cells transfected with gastrokine-2 and incubated with functional grade purified CD95 (APO/Fas) antibody (48 h, 72 h), but the increase can be reversed by treatment with anti-Fas (human, neutralizing, clone ZB4) antibody. In order to further confirm this hypothesis, the activity of caspase 3, 8, and 9 was analyzed in these groups of gastric cell lines, and it was identified that caspase 3 and 8 in extrinsic apoptosis was activated or inhibited by functional grade purified CD95 (APO/Fas) and anti-Fas (human, neutralizing, clone ZB4) antibodies, respectively. However, caspase 9-related intrinsic apoptotic gene expression was not significantly altered.\n\nIn conclusion, to the best of our knowledge, the data from the current study demonstrated for first time that restoration of gastrokine-2 expression in SGC-7901 gastric cancer cells inhibits cell viability and induces apoptosis. Furthermore, it was demonstrated that apoptosis was induced through activation of the extrinsic apoptosis pathway. Following further investigation, gastrokine-2 may prove to be a potential target for novel molecular therapies for gastric cancer.\n\nThe authors would like to thank Medjaden Bioscience Limited, Hong Kong, China, for their assistance in editing this manuscript.\n\n![Expression of gastrokine-2 in gastric cancer and corresponding normal tissues and cell lines (SGC-7901 and AGS). Tissues specimens were collected and subjected to western blot analysis. ^\\*^P\\<0.01 vs. NT. NT, normal tissue; T, tumor tissue.](MMR-10-06-2898-g00){#f1-mmr-10-06-2898}\n\n![Effects of gastrokine-2 restoration on regulation of Fas, bcl-2 and Bax expression in gastric cancer SGC-7901 cells. Expression levels of Fas, bcl-2 and Bax (A) mRNA and (B) protein were evaluated following transfection of the cells with gastrokine-2 vector for 48 h (G48 h). ^\\*^P\\<0.05 vs. Con and P48 h; ^\u25b2^P\\>0.05 vs. Con and P48 h; ^\\#^P\\>0.05 vs. Con and P48 h. Con, non-transfected SGC-7901 cells; P48 h, control vector; G48 h, gastrokine-2 vector; z, normal gastric tissue.](MMR-10-06-2898-g01){#f2-mmr-10-06-2898}\n\n![Flow cytometric analysis of Fas receptor expression in SGC-7901 cells. The cells were grown and transfected, with or without gastrokine-2 cDNA and then subjected to flow cytometric analysis of Fas expression using an anti-human CD95 antibody. ^\\*^P\\<0.05 vs. Con and P48 h. Con, non-transfected SGC-7901 cells; P48 h, control vector; G48 h, gastrokine-2 vector.](MMR-10-06-2898-g02){#f3-mmr-10-06-2898}\n\n![Effects of gastrokine-2 restoration on gastric cancer viability. SCG-7901 cells were grown and transfected with or without gastrokine-2 cDNA, and then subjected to an MTT assay. Cell viability was reduced in the G+F group (P\\<0.05 vs. Con, P, G and G+F+Z using analysis of variance). Con, non-transfected SGC-7901 cells; P, control vector; G, gastrokine-2 vector; G+F, gastrokine-2 vector + CD95 antibody; G+F+Z, gastrokine-2 vector + CD95 + Fas antibody.](MMR-10-06-2898-g03){#f4-mmr-10-06-2898}\n\n![Effects of gastrokine-2 restoration on regulation of gastric cancer apoptosis. SCG-7901 cells were grown and transfected with or without gastrokine-2 cDNA and then subjected to flow cytometry assay. ^\\*^P\\<0.05 vs. Con, P72 h and G72 h. Con, non-transfected SGC-7901 cells; P72 h, control vector 72-h transfection; G72 h, gastrokine-2 vector 72-h transfection; G+F72 h, gastrokine-2 vector 48-h transfection + CD95 antibody 24-h incubation; G+F+Z, gastrokine-2 vector 48-h transfection + CD95 + Fas antibody 24-h incubation.](MMR-10-06-2898-g04){#f5-mmr-10-06-2898}\n\n![Effects of gastrokine-2 restoration on regulation of caspase activity. SCG-7901 cells were grown and transfected with or without gastrokine-2 cDNA and then subjected to caspase-3, -8 and -9 activity assays. ^\\*^P and ^\u25b2^P\\<0.05 vs. Con, G 72 h and G+F+Z 72 h; ^\\#^P\\>0.05 vs. Con, G72 h and G+F+Z 72 h. Con, non-transfected SGC-7901 cells; G72 h, gastrokine-2 vector 72-h transfection; G+F72 h, gastrokine-2 vector 48-h transfection + CD95 antibody 24-h incubation; G+F+Z, gastrokine-2 vector 48-h transfection + CD95 + Fas antibody 24-h incubation.](MMR-10-06-2898-g05){#f6-mmr-10-06-2898}\n\n###### \n\nGastrokine-2 protein expression in gastric carcinoma.\n\n                            Gastrokine-2 protein expression        \n  ------------------------- --------------------------------- ---- -------\n  Tumor location                                                   0.699\n  \u2003Total                    1                                 4    \n  \u2003Upper                    7                                 23   \n  \u2003Middle                   4                                 10   \n  \u2003Lower                    6                                 21   \n  Depth of invasion                                                0.689\n  \u2003T0 or T1                 2                                 4    \n  \u2003T2                       2                                 3    \n  \u2003T3                       10                                43   \n  \u2003T4                       4                                 8    \n  TNM stage                                                        0.691\n  \u2003N0 (0)                   3                                 10   \n  \u2003N1 (1--6)                6                                 11   \n  \u2003N2 (7--15)               4                                 11   \n  \u2003N3 (\\>15)                5                                 26   \n  Lauren's classification                                          0.187\n  \u2003Intestinal               7                                 32   \n  \u2003Diffuse                  7                                 19   \n  \u2003Mixed-type               4                                 7    \n  Tumor stage                                                      0.667\n  \u2003I                        1                                 4    \n  \u2003II                       5                                 15   \n  \u2003III                      12                                35   \n  \u2003IV                       0                                 4    \n  Anti-*H. pylori* IgG                                             0.039\n  \u2003+                        4                                 36   \n  \u2003\u2212                        14                                22   \n\nTNM, tumor-node-metastasis. Upper, upper one third of the stomach; middle, middle one third of the stomach; lower, lower one third of the stomach.\n"}
+{"text": "Introduction {#cesec10}\n============\n\nRecent global scares involving infectious diseases---such as the release of letters containing anthrax spores in the USA in 2001 and the emergence of severe acute respiratory syndrome (SARS) in 2002---along with the continuing effort to contain highly pathogenic avian influenza A (H5N1) virus have highlighted the need to improve preparedness within Europe for emerging public-health threats.[@bib1], [@bib2], [@bib3], [@bib4], [@bib5], [@bib6] The European Commission is funding several activities intended to improve health security, build capacity, and strengthen preparedness for response to infectious disease emergencies. These ventures include a network of biosafety level 4 (BSL4/P4) laboratories (Euronet-P4), the European Programme for Intervention Epidemiology Training, and [the European Network of Infectious Diseases (EUNID)](http://www.eunid.eu/).[@bib7]\n\nThe 3-year EUNID project began in 2004, and will continue (as the European Network for Highly Infectious Diseases \\[EuroNHID\\]) for a further 3 years, until 2010. EUNID was created to exchange information, share best practice, and improve the connections between national (or regional) centres designated for the care of patients with highly infectious diseases.\n\nEUNID includes national representatives and experts from 16 member states (Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, and the UK). Most are clinicians working in high-level isolation units (HLIU) or centres designated for referral of patients with highly infectious diseases, who have backgrounds in infectious diseases, intensive care, infection control, pulmonary medicine, occupational health, or public health.\n\nThree network meetings were held during the project to reach consensus on specific project objectives,[@bib8], [@bib9], [@bib10] which included defining a highly infectious disease,[@bib11] archiving an inventory of relevant guidelines,[@bib12] defining a high-level isolation unit (HLIU), identifying the key elements in the design and operation of an HLIU, and using these to develop a framework for the design and operation of HLIUs in Europe.\n\nSince highly infectious diseases are uncommon in industrialised countries, there are few prospective randomised controlled trials on their management and a limited evidence base; therefore, EUNID took a consensus-based approach to these tasks. This report presents the framework, which provides practical guidance on the design and operation of HLIUs in Europe to support planning by public-health authorities for the management of highly infectious diseases and preparedness for infectious disease emergencies in Europe.\n\nMethods {#cesec20}\n=======\n\nIn 2003, national public-health authorities in countries that have, or are planning, an HLIU were contacted by the EUNID coordination team, with the help of the European Commission, and asked to suggest (although not to formally endorse) clinicians with expertise in highly infectious disease/HLIU management as national representatives. The team also co-opted additional participants, selected for their experience in one or more aspects of highly infectious disease/HLIU management, including a representative of the group that developed similar guidance in the USA.[@bib13] Thus, the skill mix and expertise of the group represented all key aspects of highly infectious disease/HLIU management (infectious diseases, intensive care, transport, engineering and maintenance, infection control, diagnostic services, occupational health, public health, and unit management).\n\nGroup members presented details of arrangements in their countries for the management of patients with highly infectious diseases, heard (and questioned) expert presentations on HLIU management, and agreed a consensus definition of a highly infectious disease by discussion at the first network meeting in 2005.[@bib8], [@bib11] We then inventoried national and international guidelines, and used these, with relevant legislation and representatives\\' responses to a questionnaire, to identify key elements in the design and operational management of an HLIU, which were agreed by discussion at the network meeting in 2006.[@bib9], [@bib12] These elements formed the basis for a draft framework that incorporated evidence used to support current international and national guidance and legislation. This framework, with a draft definition of an HLIU, was shared throughout the network, and revised to incorporate comments and additional evidence. We used a questionnaire that offered alternative solutions to particular design and management issues (eg, location and staffing of HLIUs, necessity for high efficiency particulate air \\[HEPA\\] filtration) to structure group discussions, clarify preferred options, and agree consensus refinements of the drafts at the third network meeting in 2007.[@bib10] We then prepared this report, and shared it, and subsequent revisions, until the content was agreed by all.\n\nResults {#cesec30}\n=======\n\nThe elements of key importance in the design and operation of HLIUs in Europe are used as subheadings in this report. The framework is based on evidence, where this is available, or, where it is not, represents the collective opinion and current practice of the group.\n\nDefinition of a highly infectious disease {#cesec40}\n-----------------------------------------\n\nEUNID defines a highly infectious disease as an infection that is transmissible from person to person; is life-threatening; presents a serious hazard in the health-care setting and the community; and requires specific control measures (eg, high-level isolation).[@bib11]\n\nDefinition of a high-level isolation unit (HLIU) {#cesec50}\n------------------------------------------------\n\nEUNID defines an HLIU as a health-care facility specifically designed to provide safe, secure, high-quality, and appropriate care, with optimal infection containment and infection prevention and control procedures, for a single patient or a small number of patients who have, or who may have, a highly infectious disease.\n\nFunction, location, and use of HLIUs {#cesec60}\n------------------------------------\n\nEUNID participants were unanimous about the need for provision of HLIU-based care in Europe, since there are some infections that, for the protection of health-care workers, other patients, and the community, require levels of infection control and clinical expertise that cannot easily be provided in any other setting. These include known or suspected infection with a haemorrhagic fever virus (Marburg, Ebola, Lassa, and Crimean-Congo haemorrhagic fever viruses), smallpox or other highly pathogenic orthopoxvirus, emerging highly pathogenic influenza virus or other emerging highly pathogenic respiratory virus, an unknown emerging pathogen, an engineered pathogen or a suspected bioterrorist agent, and any emerging or re-remerging infection considered by national or international authorities to require high-level containment. EUNID concurs with current guidance on extensively drug-resistant and multidrug-resistant tuberculosis, and does not regard provision of HLIU-based care for either infection as essential.[@bib14], [@bib15], [@bib16]\n\nSome countries in Europe---eg, Germany, Sweden, and the UK---have more than one HLIU; others---eg, Estonia, Luxembourg, and Spain---have centres designated for the care of patients with highly infectious diseases but do not have a purpose-built HLIU; Austria has neither, but has a cooperative agreement with centres in neighbouring countries for the transfer and care of patients with highly infectious diseases. Details of the facilities for patients with highly infectious diseases in EUNID-participating countries are available on the project website.[@bib17]\n\nThe advantages of having more than one unit in the same country include shorter, safer, journeys for patients from the referral point to the unit, shorter times for specimen transport from the unit to the nearest BSL4 laboratory, provision of cross-cover (eg, when one unit is closed for maintenance), and the potential for a broader range of specialist research interests. Disadvantages include increased costs, less frequent use, and dispersion of clinical expertise.\n\nHLIUs should be sited so that in-country patient journey and specimen transport times do not exceed 6 h, and should be colocated with a parent tertiary-care facility able to provide appropriate specialist support. A high proportion of patients with highly infectious diseases will have acquired their infections abroad, and since travellers are most likely to return home by air, a country\\'s first HLIU should be sited in, or near to, the population centre nearest to the country\\'s major international airport.\n\nFormal agreements between countries for cross-border transfer of patients and specimens might solve the problem of transport across difficult terrain. Formalised resource-sharing might also be appropriate for countries that cannot provide multiple high-level isolation facilities. Furthermore, integration of civilian and military medical services for care of highly infectious disease patients within a country could reduce unnecessary duplication.\n\nHLIUs are expensive to build, operate, and maintain, and, if designed exclusively for the management of patients with highly infectious diseases, might be used too little to justify their costs. Costs could be offset by the development of flexible-use HLIUs, designed to offer a range of levels of infection containment; staff would then select the containment level appropriate to the patient\\'s condition.[@bib18] Indeed, some countries (eg, Netherlands and Italy \\[[figure](#fig1){ref-type=\"fig\"} \\]) already use their highly infectious disease facility in this way.FigureAn HLIU nearly completed at the National Institute for Infectious Diseases, Rome, ItalyThe HLIU consists of ten one-bed separate units, each one equipped with separate access from the outside, distinct pathways for entering and exiting of personnel, an independent ventilation system, autoclave, and pass-through box. The building is equipped with BSL3 and BSL4 laboratories. Additional 20 single airborne-isolation rooms are located on the second floor for quarantine.\n\nThe primary purpose of the HLIU is to be ready to care for patients with highly infectious diseases, so time must be allocated for training staff and for regular maintenance testing of specialised systems. This function also requires an absolute commitment formalised through a contract, with an effective standard operating procedure, to evacuate any non-highly infectious disease patients and prepare the HLIU for admission of a highly infectious disease patient within 4 h without compromising patient care. It might be preferable to admit non-highly infectious disease patients to an HLIU only when other isolation facilities are not available, and to transfer these patients from the HLIU as soon as practicable.\n\nHLIUs are intended to provide care only for small numbers of patients. Once person-to-person transmission of a highly infectious disease is occurring within a community, HLIU-based care becomes less appropriate, so investment in an HLIU is not an alternative to planning for surge capacity or the provision of alternative care centres for use in major epidemics.[@bib13], [@bib19]\n\nOperational management and clinical care provision in HLIUs {#cesec70}\n-----------------------------------------------------------\n\nThe operational management and clinical care provision that all HLIUs should have are listed in [panel 1](#box1){ref-type=\"boxed-text\"} . Most HLIUs in Europe are directed by a lead clinician (usually an infectious disease physician), and have a designated lead nurse/manager; medical care for patients with highly infectious diseases is provided by specially trained clinical teams. Participation in the team is voluntary, and access to the operational HLIU is limited to essential staff. The team undertakes the routine tasks (eg, phlebotomy, cleaning, food service, linen changing, patient or specimen transport, record keeping) that would usually be undertaken by others (phlebotomists, housekeepers, clerks, porters). This reduces the number of staff exposed to risk, improves safety, and obviates the need to recruit and maintain an HLIU-trained non-clinical team from a high turnover workforce.Panel 1Operational management and clinical care in HLIUs**An HLIU should have:**\u2022A designated HLIU director/lead clinician (usually the most senior and experienced doctor), with overall responsibility for coordination, training, liaison, and communication\u2022A designated HLIU nursing director/lead nurse consultant/nurse manager, responsible for ward management and training of nursing staff\u2022An effective mechanism for succession planning\u2022An HLIU-specific training programme that is mandatory for all staff who will work in, or enter, the HLIU. This programme should have standardised curricula and competencies appropriate for each professional group (eg, doctors, engineers), and should include both initial training and regular refresher training, with an accurate system for recording course attendance and performance[@bib11]\u2022An audit and quality assurance programme, and a system for monitoring adverse events\u2022Sufficient HLIU-trained staff to provide 24-h availability to open and run the unit\u2022A regularly exercised standard procedure for becoming fully operational for management of a patient with a highly infectious disease within 3--4 h\u2022A clear and agreed method for providing cover for any usual duties (eg, anaesthetic list) that staff cannot undertake because they are working in the HLIU\u2022An agreed system of reward and remuneration for HLIU work and training\u2022A controlled access system, so that entry to the HLIU is limited to essential, HLIU-trained staff when it is in use for a patient with a highly infectious disease, and adequate supervision of any partly trained staff or other visitors who might need to enter the unit (eg, security emergency) is ensured\u2022An access log, documenting details of all individuals entering or leaving the unit\u2022An emergency evacuation protocol that is regularly tested\u2022A protocol, compatible with European and national legislation, for ensuring unit security\u2022A high-quality, secure communications systems (eg, scan-safe internet and e-mail, secure radio, emergency-assistance alarms)\u2022An agreed communication strategy, with a designated communication lead with responsibility for coordination with communication experts (eg, press officer) in the parent institution\n\nEUNID participants reported that their highly infectious disease patients had needed a range of interventions, including transfusion of blood/blood products; cardiac, respiratory, and invasive haemodynamic monitoring; radiography; ultrasonography; minor surgical procedures (eg, thoracocentesis); renal dialysis; and mechanical ventilation. The participants also regarded input from critical care clinicians as essential to patient management. HLIUs should therefore be equipped to provide the level of care available in an intensive care unit, and critical care clinicians should routinely train alongside the HLIU team. Support from other specialties (eg, nephrology, paediatrics, cardiology) had also been needed, so specialist clinicians should also be pre-identified and train alongside the HLIU team. Additionally, needs for specialist non-clinical expertise (eg, to repair ventilation systems or maintain near-patient testing equipment) should be assessed, and relevant staff trained appropriately in advance. Partly trained or untrained individuals who need to enter the unit should be escorted and fully supervised throughout their visit by HLIU-trained staff.\n\nOccupational health and safety {#cesec80}\n------------------------------\n\nOccupational health care ([panel 2](#box2){ref-type=\"boxed-text\"} ) is needed to ensure staff fitness to work in the HLIU, to maintain health surveillance, to ensure a rapid, effective, response to any occupational exposure or illness in staff, and to provide the psychosocial support essential in such a demanding environment.Panel 2Occupational health and safety in HLIUs**An HLIU should have:**\u2022Infection-specific and procedure-specific protocols on infection control and prevention, including risk-assessment-based use of personal protective equipment, and a programme for testing these protocols\u2022An occupational health and safety programme, led by a specialist HLIU-trained occupational health physician\u2022An audit and quality assurance programme, and a system for incident reporting and management**HLIU personnel should:**\u2022Have routine pre-employment health checks in accordance with local policy\u2022Have routine pre-employment immunisations, including hepatitis B vaccine and seasonal influenza vaccine, in accordance with local policy\u2022Not be offered HLIU-specific, pre-employment immunisations (eg, smallpox vaccine)\u2022Adhere to written local protocols for active health surveillance, which should be applied to all individuals who enter the unit while a patient with a highly infectious disease is present, or who participate in decontamination, transport, or other procedures, and should continue for a minimum of one disease incubation period after the last possible opportunity for exposure\u2022Have access to confidential psychological and spiritual support\n\nPre-employment smallpox vaccination is not required, given the low probability of a deliberate release of smallpox, the risk of potentially severe adverse events, and the high turnover of staff, but fitness to work assessments should cover contraindications to vaccination.[@bib20], [@bib21] This recommendation does not preclude the vaccination of individuals who may be members of the \"smallpox response teams\" set up in some countries in Europe as part of preparedness planning.[@bib12]\n\nPathology and other diagnostic support services {#cesec90}\n-----------------------------------------------\n\nPathology and diagnostic support services required in an HLIU are shown in [panel 3](#box3){ref-type=\"boxed-text\"} . Some of the isolation units in Europe designated for patients with highly infectious diseases have integral BSL3/4 laboratory facilities; others have access to BSL3 (or BSL4) facilities on the same campus.Panel 3Diagnostic support services in HLIUs**EUNID recommends that:**\u2022All pathology tests on HLIU patients must be undertaken in accordance with relevant European legislation, and national legislation or guidance (eg, in the UK, EU Biological Agents Directive, Control of Substances Hazardous to Health Regulations 2002, and guidance from the Health and Safety Executive[@bib22], [@bib23], [@bib24], [@bib25])\u2022Packaging and transport of patient specimens and infectious substances should conform to all relevant international, European, and national legislation[@bib26]\u2022HLIU protocols for clinical procedures and diagnostic tests should clearly state how any sample obtained should be collected, handled, and transported, state in which laboratory the test should be done, and contain up-to-date contact details for that laboratory\u2022HLIUs should make optimum use of near-patient testing systems and have written protocols for their use and maintenance\u2022HLIUs should make optimum use of laboratory information management systems, including electronic test requesting\u2022HLIUs should have access to portable ultrasonography and radiography services (ideally, digital radiography), with dedicated or pre-identified equipment\u2022Staff who undertake tests on samples from HLIU patients must be fully and appropriately trained in biosafety, and subject to the HLIU occupational health programme or (if employed by a BSL4 laboratory on a distant site) its equivalent\u2022Laboratories that undertake tests on samples from HLIU patients should be appropriately accredited, apply appropriate and valid bio-risk minimisation procedures, participate fully in appropriate external quality assurance schemes, have adequate arrangements for internal quality control and audit, and should keep appropriate records of performance, quality improvement, and staff training[@bib27]\n\nNo consensus was reached on whether auto-analysers in the main (BSL2) hospital laboratory could safely be used for routine haematology and biochemistry on HLIU patients. The view of most participants is that their use for infectious samples is acceptable, provided the analyser is of the \"closed-sampling\" type (which minimises the risk of exposure to patients\\' blood or body fluids), and is operated by trained staff in compliance with written, regularly exercised, protocols for the safe transport, handling, tracking, and disposal of specimens.[@bib28] The strongly held alternative view is that all potentially infectious samples from patients who might be infected with a potentially lethal pathogen (eg, a blood sample from an Ebola patient) should be handled at least under BSL3 conditions.[@bib29]\n\nTransport of patients to the HLIU {#cesec100}\n---------------------------------\n\nTransport arrangements ([panel 4](#box4){ref-type=\"boxed-text\"} ) for patients with highly infectious diseases in Europe vary, as do the legal regulations applicable in each country.[@bib8], [@bib9], [@bib12], [@bib30], [@bib31] Most countries require ambulance crews to be specially trained; some require the use of specialised patient transport equipment (stretcher transport isolators), specially constructed ambulances with controlled ventilation and HEPA filtration, or modified vehicles with additional personal protective equipment for staff. Sweden, Germany, Italy, and the UK have arrangements for, and experience of, national and international aeromedical transport. Transport in stretcher isolators is unpleasant for the patient, and limits patient care, so might be unsuitable for critically ill patients.[@bib32], [@bib33] High-level personal protective equipment (eg, impervious suit, powered air purifying respirator) can be used for short shifts only (less than 4 h), which limits the length of journeys that can be undertaken, and the suits, motors, and filters are easily damaged. In the past 40 years, patients with symptomatic haemorrhagic fever virus infections have travelled by road, by public airline, and by unmodified air casualty transport without the detection of a secondary case,[@bib34], [@bib35], [@bib36] so that it is now believed that almost all such patients can be managed by transport in standard ambulances, with appropriate personal protective equipment for staff. International guidance on epidemic-prone acute respiratory disease and recommendations on the management of suspected smallpox do not require patient transport by modified ambulance.[@bib20], [@bib37], [@bib38] Therefore, EUNID does not consider that ambulances with controlled ventilation systems are essential, but recognises that, in some countries, national authorities regard them as desirable, and so have invested in their use. Formal arrangements, which might involve cooperation between international agencies, non-governmental organisations, and civilian and military medical services, could help to overcome the challenges of arranging an international aeromedical evacuation.Panel 4Transport of patients to the HLIU**EUNID recommends that:**\u2022The decision to transport a patient to the HLIU must be based on expert clinical risk assessment\u2022The appropriate transport mode and vehicle type should be determined by expert risk assessment\u2022Staff and others who might be exposed to the patient during the journey should wear appropriate personal protective equipment, as determined by expert risk assessment, and should be subject to the same health surveillance after the journey as potentially exposed HLIU staff\u2022The ambulance and any fixed equipment used must be able to be effectively decontaminated (by wiping, spraying, or fogging with an effective disinfectant, according to national policy)\u2022The ambulance crew must be trained in the protocol for transport of patients with highly infectious diseases\u2022The HLIU should have an external, securable, area for ambulance parking and decontamination, and procedures for safe decontamination of ambulance equipment, including safe storage before decontamination\u2022The HLIU should have an admission route from the ambulance area to the unit entrance that can be controlled and secured, and that is wide enough to permit transfer of patient, staff, and equipment\u2022An ambulance used to transport a highly infectious disease patient should not be returned to normal use until the vehicle (and any fixed equipment in it) has been decontaminated\n\nVentilation systems, air handling, and airborne infection isolation {#cesec110}\n-------------------------------------------------------------------\n\nControlled ventilation systems ([panel 5](#box5){ref-type=\"boxed-text\"} ) reduce the risk of infection with obligate airborne pathogens.[@bib14], [@bib44], [@bib45], [@bib46], [@bib47] Although most highly infectious diseases are not primarily transmitted by the airborne route, the risk of opportunistic airborne transmission of SARS and influenza infections, through exposure to droplet nuclei during aerosol-generating procedures, might be increased in poorly ventilated environments. Therefore, WHO suggests that airborne isolation precautions could be used when aerosol-generating procedures are undertaken on these patients, and also recommends that inpatients with an acute respiratory disease caused by a novel pathogen with potential for high public-health impact should be managed in airborne isolation until routes of transmission have been clarified.[@bib37], [@bib44], [@bib45], [@bib46], [@bib47], [@bib48], [@bib49], [@bib50] Airborne transmission of smallpox can occur; therefore, all guidelines recommend airborne infection isolation for suspected cases.[@bib12], [@bib20], [@bib37], [@bib51] European legislation on containment measures for patients who have, or who may have, an infection with a group 3 or group 4 biological agent has been interpreted in supplemental legislation in some countries as requiring airborne infection isolation for all these infections, which include viral haemorrhagic fevers.[@bib22], [@bib51], [@bib52], [@bib53] Furthermore, flexible-use HLIUs may be used to care for patients with tuberculosis, measles, or chickenpox (all of which require airborne infection isolation) or to isolate laboratory technicians after occupational exposure to a pathogen of unknown transmissibility. EUNID therefore recommends that HLIUs be designed for airborne infection containment.[@bib15], [@bib17], [@bib39], [@bib40], [@bib41], [@bib42] Panel 5Ventilation systems and air handling in HLIUs**EUNID recommends that:**\u2022The HLIU ventilation system is independent of the other building heating, ventilation, and air conditioning systems\u2022Each patient room should have an anteroom\u2022Air flows and pressure gradients within the HLIU run from the cleanest to the most contaminated areas; with the patient room at negative air pressure relative to adjacent areas, and a suggested differential pressure gradient of more than 15 Pa between patient room and anteroom and between anteroom and the rest of the unit, and an effective ventilation rate of at least 12 air changes per h in the patient room\u2022Air from the HLIU is not recirculated, and exhaust air is vented 100% to the outside of the building\u2022Exhaust air is discharged at a site and distance from the building that minimises the risk of contamination of occupants of the building (eg, by down-draught into open windows) and the community\u2022HEPA filtration of exhausted air is preferable, and, if there is any possible risk of re-entry of exhaust air or of human exposure to exhausted air, obligatory\u2022HEPA filtration of supply air may be considered\u2022HEPA filters are appropriately protected by pre-filters, housed correctly, and sited for ease of safe access for maintenance[@bib39], [@bib40], [@bib41], [@bib42], [@bib43]\u2022HLIU ventilation systems are designed to fail safe, and to minimise cross-contamination in the event of system failure in the unit or elsewhere on the site (eg, built-in redundancy---HLIU ventilation system with dual fans each capable of exhausting 100% air; air flow shutdown system independent of site system to protect against unwanted shutdown after alarm elsewhere on site; interlocking supply and exhaust systems, so that supply fan is prevented from running if exhaust fans fail)\u2022HLIU ventilation systems are connected to an emergency back-up power source\u2022Commissioning of ventilation systems includes functional (\"in-use\") testing\u2022Ventilation systems incorporate current best practice performance checking tools (eg, visual pressure check gauges, audible alarms); have a schedule for planned preventive maintenance, and that HLIUs have written protocols for performance-checking the ventilation system that comply with all relevant European and national regulations and current standards of best practice\n\nDirectional airflow, where air flows from less contaminated (\"clean\") areas to more contaminated (\"dirty\") areas contains infection at source. In mechanically ventilated systems this is achieved by engineering pressure gradients by manipulating supply and exhaust air so that the patient room is maintained at negative air pressure relative to adjacent areas. Minimum recommended air pressure differentials for airborne isolation rooms vary from 2\u00b75 Pa to 30 Pa. There is no direct evidence to link increases in the pressure gradient above 10 Pa with reduction in risk of infection. Some countries (eg, Sweden, Japan, and Australia) require that airborne infection isolation rooms have an anteroom,[@bib39], [@bib43], [@bib54] whereas other countries do not, but recognise that anterooms may increase ventilation system efficiency.[@bib40] Anterooms also provide a controlled environment in which to put on and remove personal protective equipment, prepare clinical equipment, and store immediately necessary supplies.\n\nHigh, effective ventilation rates reduce the concentration of contaminating airborne particles by dilution, since contaminated air is replaced by clean air. Standards have evolved through convention, coupled with application of experimental work on concentration decay rates of airborne particles: the relation between incremental increases in air change rates and reduction in infection risk has not been assessed.[@bib39], [@bib40] Recent guidelines define an adequately ventilated airborne infection isolation room (or airborne precaution room) as one that has more than 12 air changes per h, although newer HLIUs have rates that substantially exceed this.[@bib14], [@bib37], [@bib39], [@bib41]\n\nSome guidelines recommend that ventilation system design should ensure that clean air flows from parts of the room where carers are likely to work, across the infectious source, and into the exhaust.[@bib41] This arrangement might not be achievable at lower air exchange rates, and others have suggested, in a full review of isolation room design, that the best strategy is to achieve effective mixing and the highest contaminant dilution rate consistent with maintaining thermal comfort.[@bib39] Achieving high air exchange rates and maintaining higher-range pressure differentials requires rooms designed and constructed for \"air envelope tightness\"---ie, with as little unplanned leakage as possible---with controlled leakage paths, and specially designed doors. To maximise source containment, exhaust registers (air outlet points) should be located as close to the patient\\'s head as possible.\n\nEuropean legislation on worker protection specifies a mandatory set of containment measures, including HEPA filtration, for laboratories working with group 3 or group 4 biological agents, but is less prescriptive about containment in patient isolation facilities, requiring only that containment measures be selected from the list of measures mandated for laboratories.[@bib22], [@bib52], [@bib53] In some European countries this has been interpreted as mandating HEPA filtration of both supply and exhaust air in high-level isolation facilities. Consequently, some isolation facilities in Europe have HEPA-filtered supply air, despite a lack of evidence to support the need for it.\n\nIn an outbreak in Germany, smallpox virus was thought to have been carried on air currents, from a window in an isolation room, up the outside of the building, and into other clinical areas, where patients became infected.[@bib51] Building regulations and codes now require that ventilation systems are designed to prevent re-entry of exhaust air, and that exhausts are not located near areas that might be populated, or within the building\\'s air recirculation zone. Exhaust air from HLIUs must be 100% exhausted to the outside, and not recirculated. Although HEPA filters in theory provide additional protection, their use increases the complexity and cost of the system. HLIUs in France, Germany, Italy, Sweden, and the UK have HEPA-filtered exhaust air, as do newly built isolation wards in Denmark, Finland, and Estonia; isolation units in Greece and Ireland do not.\n\nSafe management of clinical waste {#cesec120}\n---------------------------------\n\nThe greatest hazard from clinical waste ([panel 6](#box6){ref-type=\"boxed-text\"} ) is percutaneous exposure to contaminated needles or other sharp objects.[@bib55] Workers can also be at risk of airborne infection: an outbreak of tuberculosis in waste workers exposed in an industrial setting has been reported, and even small-scale compaction of waste can generate pathogen-containing aerosols.[@bib56], [@bib57] Panel 6Safe management of clinical waste in HLIUs**EUNID recommends that:**\u2022Procedures for management of clinical waste must comply with all relevant European and national legislation\u2022Unit procedures should incorporate waste-reduction measures (eg, removal of packaging from equipment in a \"clean\" area before transfer of the equipment into the clinical area)\u2022Solid waste should be decontaminated before disposal; autoclaving (with verification testing) is the preferred method\u2022Procedures for handling and disposal of liquid waste should be determined by risk assessment. Suggested methods include: direct disposal into the dirty drain system, autoclaving after solidification, or chlorine decontamination before disposal\n\nThe European Hazardous Waste Directive and related regulations set the framework for the management and disposal of clinical (medical) waste in Europe. Waste from the management of patients with a known or suspected infectious disease, where the causal pathogen or toxin is present in the waste, must be identified, separately packaged, and incinerated.[@bib58] Related national legislation varies somewhat from country to country. Increasingly, environmental legislation discourages incineration on-site in health facilities, making it necessary to transport waste to industrial incineration sites. Some countries in Europe have national standards that require that clinical waste be treated to destroy infectious agents before the waste is removed from the site where it was generated; others, however, have defined and identified \"low-risk\" clinical waste, and categorised it as non-hazardous, although universal waste-handling precautions may be more likely to ensure safety.[@bib40], [@bib55]\n\nSolid wastes generated in an HLIU require decontamination before disposal. Most HLIUs achieve this by autoclaving. The process is easily controllable, and hospital engineers are usually experienced in equipment maintenance and validation. However, autoclaving is time consuming, and energy inefficient. Large quantities of clinical waste are generated during the care of patients with highly infectious diseases, especially when critically ill: a 4-day admission of a single patient with suspected Lassa fever in an HLIU in the UK generated three industrial skip loads of double-bagged clinical waste. One HLIU has a protocol for removing all packaging and protective wrapping from disposable and other equipment in a clean area, before taking the equipment into the clinical area. Packaging and other discarded materials (eg, information leaflets) are disposed of as domestic waste (which may include recycling), cutting the quantity of waste needing autoclaving by about a half. A similar process led to savings of 25--30% in energy expenditure in a hospital in Saudi Arabia.[@bib59] Alternatives to autoclaving include microwave treatment, compaction combined with chemical treatment, which reduces the risks of aerosols but requires the use of chemicals that pose their own risks, and grinding combined with ozone injection, which is unsuitable for waste that might contain group 4 pathogens.[@bib60], [@bib61], [@bib62] Local use of special treatment processes has disadvantages: a scarcity of maintenance expertise, greater system complexity, deterioration of equipment through infrequent, intermittent use, and difficulties in obtaining spare parts could all contribute to reduced system functionality and safety.\n\nHealth-care workers are at risk of infection from exposure caused by splash, spillage, or aerosol generation during handling or disposal of infected fluids, and there is a theoretical risk for engineers or others working on drainage or sewage systems within or close to the unit. Guidance in the USA recommends disposal of liquid waste (eg, blood, urine, vomit) without pretreatment by pouring into the sanitary sewer.[@bib13] A unit in the UK converts all liquid waste from the patient to a gel by use of specially formulated absorbent crystals that are added as degradable sachets to disposable containers/urinals immediately before use. The gel is then managed as solid waste by autoclaving. This procedure avoids contamination of the drainage system with high concentrations of viruses, and protects health-care workers from the risk of spillage. The unit uses this method for peritoneal dialysate (which can contain high concentrations of virus, even if not blood-stained[@bib63]), and for fluids collected when setting up haemodialysis equipment or rinsing the blood compartment at the end of the session,[@bib64] but haemodialysis ultrafiltrate (likely to contain no, or very low, concentrations of virus because particles of more than 7 nm diameter will be retained by the filtration membrane[@bib64], [@bib65], [@bib66], [@bib67]) is disposed of directly into the dirty drains system. Alternative decontamination methods include chemical disinfection, although the chemicals used might be ineffective in the presence of organic matter and are potentially hazardous to health-care workers, and heat treatment, which requires storage of wastes until treated and needs complex equipment and engineering controls.\n\nEUNID participants did not reach a consensus on the need to decontaminate liquid waste before disposal. Most would recommend direct disposal of excreta into the dirty drains system without additional safety measures (eg, pre-addition of hypochlorite), because any significant concentrations of virus would be rapidly diluted within the hospital wastewater system and degraded by agents used for routine cleaning, and the theoretical risk to workers undertaking repairs could be managed by using properly trained, supervised, and equipped maintenance staff. A few participants advise more caution, and would recommend that untreated body fluids (including urine, faeces, haemodialysis ultrafiltrate) should not be disposed of into the dirty drain system, arguing that their approach provides maximum protection and would also be of benefit in the event of the emergence of a new, highly infectious pathogen in the period before the pathogen\\'s transmission and capacity for environmental survival were fully understood. However, the group agreed unanimously that treatment of water resulting from washing or showering by unit personnel (or an ambulant, continent, patient) was not required.\n\nDecontamination of equipment and environmental hygiene {#cesec130}\n------------------------------------------------------\n\nHLIU cleaning and decontamination protocols ([panel 7](#box7){ref-type=\"boxed-text\"} ) should cover patients\\' likely primary diagnoses and health-care-associated infections (eg, *Clostridium difficile*, meticillin-resistant *Staphylococcus aureus*).Panel 7Decontamination of equipment and environmental hygiene in HLIUs**EUNID recommends that:**\u2022HLIU equipment should be selected with decontamination in mind\u2022If an item of equipment cannot safely be decontaminated for reuse, a disposable alternative should be selected\u2022The HLIU should maintain an inventory of all unit equipment, which states the usual method of decontamination/disposal of each item (eg, bed linen, smallpox---double leak proof bag, autoclave or incinerate on site; bed linen, viral haemorrhagic fever---alginate bag, leak proof container, hot wash laundry cycle; bed frame---wipe down with approved hospital detergent/disinfectant, clean with water, dry)\u2022Staff undertaking cleaning or decontamination should wear appropriate personal protective equipment and be appropriately trained[@bib11], [@bib18], [@bib40]\u2022Patient care equipment (eg, mechanical ventilator, pressure-control mattress) can be decontaminated according to standard national/local protocols. Large, complex equipment that has been contaminated might require decontamination on site before disassembly, and a fumigation method may sometimes be appropriate.[@bib13] In each case, a risk assessment should determine whether staff can safely disassemble the equipment before the separate parts are autoclaved\u2022Standard national/local hospital protocols are used for cleaning and decontamination of environmental surfaces (eg, hospital detergent/disinfectant designed for general housekeeping purposes for routine cleaning of clinical area; hypochlorite 1% solution for heavily contaminated clinical areas)\u2022Formaldehyde fumigation of clinical areas is not necessary\u2022Spillages of blood and other potentially infectious material should be promptly cleaned and decontaminated\n\nEnvironmental survival of pathogens for which HLIU-based care is recommended is such that extraordinary methods of decontamination will not usually be necessary; an exception might be the need to use a fumigation method to decontaminate large, complex equipment.[@bib13], [@bib18], [@bib40]\n\nHLIU design and construction requirements {#cesec140}\n-----------------------------------------\n\nThe design and construction requirements for HLIUs are listed in [panel 8](#box8){ref-type=\"boxed-text\"} and [panel 9](#box9){ref-type=\"boxed-text\"} . Since HLIUs are specifically designed to provide high-level infection control and best possible patient care without compromising the safety of carers or the community, the design and construction values must be specified to ensure this. We suggest that planners seek early advice from clinicians and managers in existing HLIUs. Unit design should also minimise the stressful effects of isolation on the patient, and provide a good working environment for staff.Panel 8Overall design and construction requirements for HLIUs**The HLIU should:**\u2022Be colocated with a tertiary referral hospital, either in an entirely separate building on the same site, or as a separate unit (with a secure, controllable, entrance/exit route) within the hospital\u2022Be designed and constructed in compliance with relevant European, national, and local regulations and codes\u2022Have a separate, securable, entrance; double, inter-locking doors are ideal\u2022Be designed internally so that movement of \"clean\" and \"contaminated\" staff, patients, and equipment through the unit ensures segregation of \"clean\" and \"dirty\" areas\u2022Have integral autoclave facilities, or safe access to pre-identified, dedicated autoclave facilities\u2022Have an integral BSL3 or equivalent laboratory, or access to one in close proximity on the same campus\u2022Have an adequate storage area for large equipment\u2022Have adequate storage space for supplies of personal protective equipment, pharmaceuticals (including controlled drugs), and clinical supplies\u2022Have a sealable area for decontamination of large equipment\u2022Have a designated area for handling and packaging clinical waste\u2022Have an area for the temporary safe-keeping of deceased patients, large enough to contain and decontaminate trolleys, sealable coffins, and other mortuary equipment\u2022Have a staff rest area\u2022Have a staff office area\u2022Have staff changing and showering facilities\u2022Have a decontamination shower\u2022Be designed and constructed for ease of cleaning and decontamination (eg, seamless floors and walls, solid horizontal surfaces)\u2022Have building, electrical, ventilation, and other systems that are designed and constructed for easy and safe access for maintenance\u2022Have a safe, securable pathway for emergency evacuation of staff and patients\u2022Have a connection to an emergency power-generating system\u2022Have standard life-safety systems (eg, automatic sprinkler systems) that are compliant with current European, national, and local regulation, with an independent airflow shutdown systemPanel 9Design and construction requirements for high-level patient isolation rooms**Patient isolation rooms in the HLIU should:**\u2022Be large enough to contain the specialist equipment (eg, mechanical ventilator, haemofiltration machine, monitoring equipment) needed for critical care, and to allow free movement by staff wearing personal protective equipment\u2022Have a self-closing door, with well-fitted, durable, door seals\u2022Have a non-hand operated wash basin for clinical use in the patient\\'s room\u2022Have a non-hand operated wash basin for clinical use in the anteroom\u2022Have an en-suite bathroom (toilet, hand basin, and shower); entrance to bathroom should be from within the patient\\'s room, not the anteroom\u2022Have a system for visually monitoring the patient and room from the outside, which is flexible enough to maintain patient privacy, dignity, and safety\u2022Have a high-quality patient--clinician communication system\u2022Have a high-quality clinician--clinician communication system\u2022Have an emergency alarm system so that help can be summoned immediately if need be\u2022Have an anteroom (or equivalent designated area) large enough to store immediately necessary personal protective equipment and clinical supplies (eg, intravenous fluids and tubing, syringes, dressings, specimen containers)\u2022Have an adequate area for packaging clinical specimens and for decontaminating outer specimen containers\u2022Be designed and constructed for ease of cleaning and decontamination\u2022Have sealed windows\u2022Be designed and constructed to be as airtight as possible (ie, monolithic ceilings, tightly fitting doors and windows, door grill designed for a controlled air path); other design features that enhance the functionality of the ventilation system may also be desirable (eg, interlocking door system, with clinician controlled override function, which ensures that the patient room--anteroom door and anteroom--corridor door cannot both be opened at the same time)\n\nConclusion {#cesec150}\n==========\n\nHLIUs are designed to provide optimum medical care for patients with highly infectious diseases, while at the same time protecting health-care workers, other patients, and the wider community from infection. The renewed interest in biodefence-related research, fuelled by fear of bioterrorism, has meant that more laboratories are working with group 4 pathogens than at any time in the past 50 years, which has increased the likelihood that occupationally acquired infections requiring HLIU-based care will occur.[@bib68] In 2004, for example, a laboratory technician in the USA was admitted to an HLIU for observation after occupational exposure to Ebola virus, and a Russian laboratory worker died of occupationally acquired Ebola haemorrhagic fever.[@bib69] Laboratory personnel working on newly emergent infections can also be at risk of infection, and might, if diagnosis and containment is delayed, transmit infection to others.[@bib70] Infections caused by hazardous pathogens or by newly emerging infections of unknown transmissibility are likely to occur, or be imported, too infrequently in Europe to generate widespread experience or confidence in their management, but sufficiently often enough to justify the maintenance of centres of expertise where they can be most effectively and safely managed.[@bib71], [@bib72], [@bib73], [@bib74], [@bib75]\n\nThis framework for the design and operation of these centres in European countries was developed by consensus by a multinational, multidisciplinary expert group. Strongly opposed positions emerged on only two issues: the management of routine clinical samples and the disposal of liquid clinical waste. In each case, the options expressed represent current practice in at least one of the isolation facilities for patients with highly infectious diseases in Europe; in neither case is there evidence to determine which option is the safest. The more cautious positions rely on a stricter application of the precautionary principle, or have been required by a more conservative interpretation, at national level, of European legislation, but are more costly and complex to implement, and are not recommended by similar clinical guidance elsewhere.[@bib13], [@bib28]\n\nThere are other areas of uncertainty, including the optimum air exchange rate, optimum mode of patient transport, and the necessity for HEPA filtration of exhaust air. Future research could provide definitive answers, and could also provide innovative design and engineering solutions that reduce costs, simplify maintenance, maximise energy efficiency, or improve health-care worker or patient comfort without compromising safety.\n\nHealth-care systems, clinical services and training, and health and safety legislation vary between countries, and it should not be expected, and may not be desirable, that risk assessments, risk management solutions, and operational policies for HLIUs in Europe will be completely standardised. For public-health authorities and individual countries to develop and maintain effective, flexible, high-quality services, and improve regional capacity to respond to health threats, a collaborative approach to the design, operation, maintenance, audit, and oversight of these specialist services is required. This approach should be taken from member state to member state, or more broadly across the European Union in a way that makes the most of existing expertise.\n\nEuroNHID will expand on EUNID\\'s work (intended to complement, rather than duplicate, the work of the European Centre for Disease Prevention and Control), and will seek the opinion of the European Union of Medical Specialists on this framework. Furthermore, EuroNHID will explore, through the Public Health Executive Agency, other mechanisms by which this framework might be adopted as a standard for HLIUs in Europe.\n\nSearch strategy and selection criteria {#cesec160}\n======================================\n\nData published up to December, 2007, were obtained by searches of PubMed, Medline, and ProMed, and from review of the references listed in retrieved articles. Search terms included \"patient isolation\", \"airborne transmission\", \"infection control\", \"occupational health\", \"patient transfer\", \"ventilation\", \"waste management\", \"decontamination\", and \"hospital design and construction\", and \"Lassa fever\", \"Ebola\", \"haemorrhagic fever\", \"Marburg virus\", \"haemorrhagic fever, Crimean\", \"smallpox\", \"monkey pox\", \"SARS virus\", \"extensively drug resistant tuberculosis\", and \"avian influenza\". Abstracts of articles in English, French, German, Italian, and Spanish languages were read and considered, and we also reviewed relevant national guidelines, supplied (and if necessary translated) by EUNID participants.[@bib12] No data restrictions were placed on our searches.\n\n**Conflicts of interest**\n\nWe declare that we have no conflicts of interest.\n\nAcknowledgments\n===============\n\nWe thank Carla Nisii and Ramona Iacovino for their support throughout the project and Philip W Smith for scientific advice. This work was partly supported by the EC grant EUNID (2003207), and by the Ministero della Salute, Italia-Ricerca Corrente, Istituti di Ricovero e Cura a Carattere Scientifico.\n\nMembers of the EUNID Working Group are: Agoritsa Baka, National Health Operations Centre, Ministry of Health, Athens, Greece; Hans-Reinhard Brodt, J W Goethe University, Frankfurt am Main, Germany; Philippe Brouqui, CHU Nord AP-HM, Marseille, France; Per Follin, Link\u00f6ping University, Link\u00f6ping, Sweden; Ida E Gj\u00f8rup, University of Copenhagen, Herlev, Denmark; Ren\u00e9 Gottschalk, Public Health Office, Frankfurt am Main, Germany; Margaret M Hannan, University College of Dublin, Dublin, Ireland; Robert Hemmer, Centre Hospitalier de Luxembourg, Barbl\u00e9, Luxembourg; Ilja M Hoepelman, University Medical Center Utrecht, Utrecht, Netherlands; Boo Jarhall, Link\u00f6ping University, Link\u00f6ping, Sweden; Kuulo Kutsar, National Public Health Institute, Tallinn, Estonia; Helena C Maltezou, Hellenic Center for Disease Control and Prevention, Athens, Greece; Magda Campins Marti, Hospital Universitario Vall d\\'Hebron, Barcelona, Spain; Kristi Ott, West Tallinn Central Hospital, Tallinn, Estonia; Renaat Peleman, University Hospital of Gent, Gent, Belgium; Christian Perronne, Unit\u00e9 des Maladies Infectieuses--H\u00f4pital Universitaire Raymond Poincar\u00e9, Paris, France; Gerard Sheehan, University College of Dublin, Dublin, Ireland; Heli Siikam\u00e4ki, Helsinki University Central Hospital, Helsinki, Finland; Peter Skinhoj, Rigshospitalet, Copenhhagen, Denmark; Antoni Trilla, University of Barcelona, Barcelona, Spain; Norbert Vetter, Otto Wagner Spital, Wien, Austria.\n\nOther members of EUNID project are Philippe Bossi (France), William Powderly (Ireland), Kamal Mansinho (Portugal).\n\nWe declare that we have no conflicts of interest.\n\n[^1]: Other members listed at end of paper\n"}
+{"text": "1. Introduction {#sec1-viruses-10-00434}\n===============\n\nMosquito-borne Zika virus (ZIKV; genus *Flavivirus*) originated in a sylvatic cycle of transmission between arboreal *Aedes* mosquitoes and wildlife, including non-human primates, in Africa \\[[@B1-viruses-10-00434],[@B2-viruses-10-00434]\\]. Between its discovery in 1947 and 2006, the virus was detected in a limited range within the tropical belt of Africa and Asia, and human cases during this time were primarily attributable to spillover from the sylvatic cycle \\[[@B1-viruses-10-00434]\\]. However, in 2007, an epidemic of ZIKV occurred on Yap island, where the virus was most likely transmitted by *Ae.* (*Stegomyia*) *hensilii* \\[[@B3-viruses-10-00434]\\]. In 2013, ZIKV invaded French Polynesia and then the Americas, where it initiated explosive spread among humans via the urban vector *Ae.* (*Stegomyia*) *aegypti* \\[[@B1-viruses-10-00434],[@B2-viruses-10-00434],[@B4-viruses-10-00434]\\]. At this time, the link between Zika virus infection of pregnant women and Zika congenital syndrome, particularly the manifestation of microcephaly, was first established \\[[@B5-viruses-10-00434]\\].\n\nThe introduction of ZIKV to the New World raised concern that the virus would spill back into a sylvatic cycle in its new range \\[[@B1-viruses-10-00434],[@B6-viruses-10-00434]\\]. We recently demonstrated, via mathematical modeling, that ZIKV does have high potential for such spillback in Brazil \\[[@B6-viruses-10-00434]\\]. A sylvatic ZIKV cycle would preclude eradication of the virus from the Americas via human vaccination and undermine ZIKV control efforts \\[[@B1-viruses-10-00434],[@B6-viruses-10-00434]\\]. The spillback of YFV, another flavivirus, which was introduced into the New World centuries ago via the slave trade, illustrates the potential for flaviviruses to establish novel sylvatic cycles and the risks such cycles pose \\[[@B7-viruses-10-00434],[@B8-viruses-10-00434]\\]. YFV circulates between non-human primates and canopy-living mosquitoes, particularly *Haemagogus* and *Sabethes*, in the Americas \\[[@B8-viruses-10-00434]\\]. Spillover from this cycle is currently driving an unprecedented outbreak of yellow fever among humans in Brazil \\[[@B9-viruses-10-00434]\\].\n\nOne of the caveats of our modeling study was that more data was urgently needed on susceptibility of New World primates and sylvatic mosquitoes to ZIKV \\[[@B6-viruses-10-00434]\\]. Subsequently, we and others have demonstrated that New World monkeys are susceptible to ZIKV in the laboratory \\[[@B10-viruses-10-00434],[@B11-viruses-10-00434],[@B12-viruses-10-00434]\\] and that free-living monkeys are becoming infected with ZIKV in Brazil \\[[@B9-viruses-10-00434],[@B13-viruses-10-00434]\\]. However, to our knowledge, the vector competence of sylvatic New World mosquitoes for ZIKV has not yet been tested. To help close this knowledge gap, we tested the vector competence of *Sabethes cyaneus*, which inhabits neotropical forests from Belize south to Argentina (<http://www.mosquitocatalog.org/taxon_descr.aspx?ID=15874>) for ZIKV.\n\n2. Materials and Methods {#sec2-viruses-10-00434}\n========================\n\n2.1. Virus Strains and Cell Lines {#sec2dot1-viruses-10-00434}\n---------------------------------\n\n*Ae. albopictus* (C6/36) cells were maintained in minimum essential medium (MEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM [l]{.smallcaps}-glutamine, 2 mM non-essential amino acids (all from Gibco, Life Technologies, Grand Island, NY, USA), and 0.05 mg/mL gentamycin (Invitrogen, Life Technologies, Grand Island, NY, USA) at 32 \u00b0C, 5% CO~2~, and 85% relative humidity (RH). African green monkey kidney (Vero) cells (CCL-81) were maintained in DMEM supplemented with 5% FBS and Penicillin/Streptomycin (P/S; 100 Units/mL and 100 \u03bcg/mL respectively) at 37 \u00b0C, 5% CO~2~, and 85% RH. Zika virus (strain Mex-1-7 obtained from the University of Texas Medical Branch's World Reference Center for Emerging Viruses and Arboviruses) was used to infect mosquitoes and mice. ZIKV MEX 1-7 (Genbank accession. No. KX247632) was originally isolated from *Ae*, *aegypti* from Chiapas (Mexico) in 2015, and subsequently passaged on Vero (4x), C6/36 (1x), and Vero (3x) and C6/36 (2x) cells.\n\n2.2. Mosquito Strains {#sec2dot2-viruses-10-00434}\n---------------------\n\nAlthough the particular species *Sa. cyaneus* is not a known vector of arboviruses, it was chosen for this study because there is a stable colony available, and at least a subset of the individuals of this colony are willing to feed on mice. The Maj\u00e9 strain of *Sa. cyaneus* was isolated by Dr. Woodbridge Foster on Isla de Maj\u00e9, Lago Bayano in eastern Panama in 1988 and maintained continuously at The Ohio State University until 2016, when a portion of the colony was transferred to New Mexico State University. *Sa. cyaneus* were maintained at 28 \u00b0C and 80% RH in BugDorm cages (30 \u00d7 30 \u00d7 30 cm, Bioquip, Rancho Dominguez, CA, USA) provisioned with branches that served as mating locations, artificial tree holes containing water for oviposition, and two separate Erlenmeyer flasks with a cotton ball wick, one each containing water or 20% sucrose. The mosquitoes were allowed to feed weekly on one investigator's (I.A. Hansen) hand for 15--30 min for blood meal (per approved NMSU IRB application 15308). Eggs were transferred from the artificial tree holes into rearing. Larvae were fed with dry cat food pellets (Special Kitty, Walmart, Bentonville, AR, USA) *ad lib*. See [Supplementary Video 1](#app1-viruses-10-00434){ref-type=\"app\"} for a detailed protocol for maintenance of *Sa. cyaneus*. An *Ae. aegypti* strain from Salvador, Brazil at the F7 generation was used as a benchmark for vector competence; these were maintained as previously described \\[[@B14-viruses-10-00434]\\]. Both species were reared in a common incubator at the University of Texas Medical Branch for the infection studies at 27 \u00b0C \u00b1 1 \u00b0C, 16:8 L:D, 80% RH and maintained in this incubator after feeding on infected mice.\n\n2.3. Mouse Infections and Mosquito Feeding {#sec2dot3-viruses-10-00434}\n------------------------------------------\n\nWe have previously shown that ZIKV is more infectious to mosquitoes that imbibe the virus from a live, viremic host than from an artificial bloodmeal \\[[@B14-viruses-10-00434]\\]. Thus we infected A129 mice lacking the interferon (IFN) alpha and beta-receptor with ZIKV as described in detail previously \\[[@B14-viruses-10-00434]\\]. Briefly, three-week-old mice were injected intraperitoneally with 1 \u00d7 10^5^ pfu/mouse ZIKV. To facilitate mosquito feeding, mice were anesthetized via intraperitoneal injection with 100 mg/kg ketamine and 10 mg/kg xylazine in a volume of 100 \u00b5L Cartons of *Sa. cyaneus* or *Ae. aegypti* were allowed to feed on mice one day or two dpi for 30 min, after which they were cold-anesthetized and engorged mosquitoes were separated and returned to normal maintenance conditions. Serum was collected from each infected mouse after mosquito feeding and the virus titer was determined using a focus-forming assay as previously described \\[[@B14-viruses-10-00434]\\]. The titers of these samples were 5.38 log~10~ pfu/mL at 1 dpi and 6.83 log~10~ pfu/mL at 2 dpi, respectively. All animal procedures and manipulations were approved by the UTMB Institutional Animal Care and Use Committee (IACUC) (protocol \\#170851 approved 2 February 2018).\n\n2.4. Mosquito Incubation and Virus Quantification {#sec2dot4-viruses-10-00434}\n-------------------------------------------------\n\nA subsample of 5 to 9 *Sa. cyaneus* was collected on 3, 4, 5, 7, and 14 dpf from the batch of mosquitoes fed on the mice at 1 dpi and 3, 4, 5, 7, 14, and 21 dpf from the batch of mosquitoes fed on the mice at 2 dpi. A subsample of 10 *Ae. aegypti* was collected on 14 and 21 dpf from each of the batches of mosquitoes fed on mice 1 and 2 dpi. To collect saliva, individual mosquitoes were cold-anesthetized and legs were removed and retained in a microfuge tube containing a steel ball bearing and 500 \u00b5L of homogenization media, consisting of 500 \u00b5L of DMEM (Gibco, Grand Island, NY, USA) supplemented with 2% FBS, 1% P/S (Gibco) and 2.5 \u00b5g/mL amphotericin B (Gibco). Mosquitoes were affixed to a glass slide using mineral oil and the proboscis was inserted into a sterile micropipette tip containing 8 \u00b5L FBS and allowed to expectorate for 30 min. The expectorate was then transferred into 100 \u00b5L of homogenization media. Finally, the mosquito body was placed into a microtube containing 500 \u00b5L homogenization media and a steel ball bearing. Bodies and legs were triturated for 5 min at 26 Hz in a TissueLyser II (Qiagen, Venio, The Netherlands) and clarified by centrifugation.\n\nVirus titer was determined in each sample via plaque-forming unit assay on C6/36 cells using anti-ZIKV hyperimmune serum against ZIKV strain MR-766 at a 1:5000 dilution, and peroxidase-labeled goat anti-mouse secondary antibody (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD, USA) at a 1:1000 dilution following previously described methods \\[[@B15-viruses-10-00434]\\]. The units of resulting titers are given as log~10~ pfu/sample per Hanley et al. \\[[@B16-viruses-10-00434]\\]. All statistical analysis was conducted using JMP Pro 13 (SAS Institute Inc., Cary, NC, USA).\n\n3. Results {#sec3-viruses-10-00434}\n==========\n\nFeeding rates for *Sa. cyaneus* were low, with 21% feeding on the mouse at 1 dpi and 28% feeding on the mouse at 2 dpi; in contrast, more than 85% of *Ae. aegypti* fed on the mice on each day. Of 69 engorged *Sa. cyaneus,* ZIKV was detected in only one individual ([Table 1](#viruses-10-00434-t001){ref-type=\"table\"}), albeit in all body compartments sampled (body, legs, and saliva). This mosquito had fed on a mouse at day 2 dpi; titers increased in the mice by approximately tenfold between day 1 and day 2 pi (5.38 log~10~ pfu/mL to 6.83 log~10~ pfu/mL). Moreover the one infected *Sa. cyaneus* had been allowed to incubate for 21 days, the maximum time period utilized in this study. The small sample size of infected *Sa. cyaneus* precluded statistical analysis of the effects of the experimental manipulations (mouse dpi or mosquito dpf) on infection or a statistical comparison with *Ae. aegypti*.\n\nIn contrast, *Ae. aegypti* showed high levels of ZIKV infection, dissemination, and transmission ([Table 2](#viruses-10-00434-t002){ref-type=\"table\"}). Among mosquitoes fed on the mouse at day 2 pi, 100% showed infection and dissemination by day 14 and 70% showed potential transmission. A nominal logistic analysis was used to test the effects of mouse dpi, mosquito dpf, and their interaction on the percent *Ae. aegypti* infected, disseminated or transmitting. The interaction between mouse dpi and mosquito dpf, as well as mosquito dpf independently, did not have a significant effect on any of these outcomes (*p* \\> 0.05 for all comparisons). Mouse dpi, which reflects virus titer, did have a significant effect on percent mosquitoes infected (DF = 1, chi squared = 13.68, *p* = 0.0002), the percent mosquitoes in which ZIKV disseminated (DF = 1, chi squared = 21.02, *p* \\< 0.0001), and the percent of mosquitoes with potential transmission from saliva (DF = 1, chi squared = 13.68, *p* = 0.0002), with higher percentages of ZIKV positive bodies, legs, and saliva among mosquitoes fed on the mouse at 2 dpi.\n\nThe titer of ZIKV in the body and legs of the single infected *Sa. cyaneus* fell squarely within the range of ZIKV titers in the bodies and legs of infected *Ae. aegypti* ([Table 3](#viruses-10-00434-t003){ref-type=\"table\"}). However ZIKV titer in the saliva of *Sa. cyaneus* was almost two orders of magnitude lower than the ZIKV titer in the saliva of *Ae. aegypti* ([Table 3](#viruses-10-00434-t003){ref-type=\"table\"}).\n\n4. Discussion {#sec4-viruses-10-00434}\n=============\n\nHere we have shown for the first time that *Sa. cyaneus* is a competent vector for a New World strain of ZIKV, although its competence is orders of magnitude lower than that of *Ae. aegypti*. The minimum extrinsic incubation period (EIP) of ZIKV in *Sa. cyaneus* measured in this study was between 15 and 21 days, substantially longer than the EIP of ZIKV in *Ae. aegypti* measured in this and other studies \\[[@B2-viruses-10-00434],[@B14-viruses-10-00434]\\]. Although it must be considered preliminary due to the low number of *Sa. cyaneus* infected, this estimate of EIP is consistent with the long EIP of YFV in forest-living New World mosquitoes. Studies conducted in the 1930s and 1940s have shown that the minimum EIP required for various *Haemogogus* species to transmit YFV to monkeys was 13 days, much longer than for *Ae. aegypti* \\[[@B17-viruses-10-00434],[@B18-viruses-10-00434]\\]. These same studies also showed that the number of mosquitoes capable of transmission increased dramatically between 13 and 21 days post-feeding. Similarly, Couto-Lima et al. \\[[@B19-viruses-10-00434]\\] reported a minimum YFV EIP of 14 days in *H. leucolaenas* and *S. albiprivus*. Galindo et al. \\[[@B20-viruses-10-00434]\\] found that the EIP of YFV in *H. mesodentatus gorgasi*, *H. mesodentatus mesodentatus*, *H. equinus* and *S. chloropterus* was \u226426--27 days, but 26 days was the earliest time post-feeding tested. Our data are also in line with studies of the competence of sylvatic *Aedes* species in Africa for ZIKV. Diagne et al. \\[[@B21-viruses-10-00434]\\] fed four *Aedes* species on artificial bloodmeals containing six strains of ZIKV and monitored virus presence in the body, legs, and saliva at 5, 10, and 15 dpf. They found that only two species, *Ae. (Fredwardsius) vittatus* and *Ae. (Stegomyia) luteocephalus*, expectorated detectable virus, and this occurred only at 15 dpf.\n\nSeveral caveats to the current study must be acknowledged. First, the *Sa. cyaneus* utilized have been maintained in a colony for decades, whereas the *Ae. aegypti* were only seven generations removed from wild type. Maintenance in a colony can impact vector competence for flaviviruses \\[[@B22-viruses-10-00434]\\]. However, studies that utilized field collected *Haemagogus* nonetheless showed a long EIP for YFV \\[[@B17-viruses-10-00434],[@B18-viruses-10-00434],[@B19-viruses-10-00434]\\], suggesting that our observations may reflect the actual dynamics of ZIKV in *Sa. cyaneus*. We have efforts ongoing to collect *Sabethes* and *Haemagogus* mosquitoes in the forest canopies in and around Manaus, Brazil in order to establish collections of field-caught and low-generation colony mosquitoes for vector competence testing. Second, a relatively small number of *Sa. cyaneus* fed on infected mice. This species is extremely difficult to maintain in colony and does not readily feed on mice; even under our tailored feeding protocol, a maximum 28% feeding efficiency was observed. Third, we utilized only one strain of ZIKV. We and others have previously shown that ZIKV strains differ in their ability to infect, disseminate and be transmitted in saliva in *Ae. aegypti* \\[[@B14-viruses-10-00434],[@B23-viruses-10-00434]\\] and *Ae. (Stegomyia) albopictus* \\[[@B24-viruses-10-00434]\\], and it therefore seems likely that strains may also vary in their ability to infect *Sa. cyaneus*. Nonetheless, we do know that the ZIKV strain used in the current study is quite infectious for both *Ae. aegypti* (this study) and *Ae. albopictus* \\[[@B24-viruses-10-00434]\\]. Finally, the ZIKV titers achieved in A129 mice were several orders of magnitude higher than those reported for humans, i.e., \\[[@B25-viruses-10-00434],[@B26-viruses-10-00434],[@B27-viruses-10-00434]\\] or New World non-human primates \\[[@B10-viruses-10-00434],[@B11-viruses-10-00434],[@B12-viruses-10-00434]\\], although these comparisons are indirect as most of the human and non-human primate studies quantified viremia via measurement of the viral genome rather than infectious virus. Generally, the percentage of mosquitoes infected is positively correlated with arbovirus titer in the bloodmeal \\[[@B28-viruses-10-00434]\\].\n\nDespite these limitations, the current study offers the first view of vector competence and EIP of any New World sylvatic mosquito vector for ZIKV. EIP is a critical parameter for models of arbovirus transmission \\[[@B29-viruses-10-00434]\\], Although we were not able to measure the EIP~50~ of ZIKV in *Sa. cyaneus* due it its low overall competence \\[[@B30-viruses-10-00434]\\], nonetheless, the data generated here can be used to refine models of ZIKV spillback \\[[@B6-viruses-10-00434]\\]. While *Sabethes cyaneus* has not itself been implicated as a vector of any arbovirus, other *Sabethes* species have been shown to carry the flaviviruses YFV \\[[@B31-viruses-10-00434],[@B32-viruses-10-00434],[@B33-viruses-10-00434],[@B34-viruses-10-00434],[@B35-viruses-10-00434]\\], dengue virus \\[[@B36-viruses-10-00434]\\], and Ilheus virus \\[[@B37-viruses-10-00434]\\] as well as the rhabdovirus Xiburema virus \\[[@B38-viruses-10-00434]\\]. *Sabethes* species are highly abundant in the forest canopy \\[[@B39-viruses-10-00434],[@B40-viruses-10-00434],[@B41-viruses-10-00434],[@B42-viruses-10-00434]\\], occur in forested areas of large cities \\[[@B43-viruses-10-00434],[@B44-viruses-10-00434],[@B45-viruses-10-00434]\\], have long lifespans in nature that exceed the 21 day EIP measured here \\[[@B46-viruses-10-00434]\\], and have been collected via human land catch \\[[@B47-viruses-10-00434],[@B48-viruses-10-00434]\\]. Together, these traits suggest that species in this genus have high potential act as bridge vectors \\[[@B49-viruses-10-00434]\\] for spillback from humans to wildlife.\n\nWe offer warm thanks to Woodbridge Foster for providing us with *Sabethes cyaneus* as well as information on the history of the colony, to Stacy Rodriguez, NMSU, for assistance with mosquito colony maintenance, and to Julia Vulcan, NMSU, for production of the rearing video.\n\nSupplementary materials can be found at <http://www.mdpi.com/1999-4915/10/8/434/s1>.\n\n###### \n\nClick here for additional data file.\n\nConceptualization, I.A.H. and K.A.H.; Data curation, A.K.K. and S.R.A.; Formal analysis, K.A.H.; Funding acquisition, N.V., S.C.W. and K.A.H.; Investigation, A.K.K., S.R.A., J.A.P. and R.Y.; Methodology, S.R.A., I.A.H. and K.A.H.; Project administration, K.A.H.; Resources, S.R.A., J.A.P. and I.A.H.; Supervision, N.V., S.C.W. and I.A.H.; Validation, K.A.H.; Visualization, I.A.H.; Writing---Original draft, K.A.H.; Writing---Review & editing, A.K.K., S.R.A., J.A.P., R.Y., N.V., S.C.W., I.A.H. and K.A.H.\n\nThis research was funded by NIH 1R15AI113628-01, R24AI120942, and NIH ICIDR 1U01AI115577-01; the concept for this study was shaped by the NSF-funded Research Coordination Network Infectious Disease Evolution Across Scales.\n\nThe authors declare no conflict of interest.\n\nviruses-10-00434-t001_Table 1\n\n###### \n\nInfection, dissemination, and potential transmission of ZIKV in *Sabethes cyaneus* fed on infected A129 mice.\n\n  Mosquito Day Post-Feeding   Mouse Day Post-Infection   ZIKV Titer (log~10~ pfu/mL)   N   No. (%) Infection   No. (%) Dissemination   No. (%) Transmission\n  --------------------------- -------------------------- ----------------------------- --- ------------------- ----------------------- ----------------------\n  3                           1                          5.38                          5   0                   0                       0\n  4                           1                          5.38                          5   0                   0                       0\n  5                           1                          5.38                          5   0                   0                       0\n  7                           1                          5.38                          5   0                   0                       0\n  14                          1                          5.38                          5   0                   0                       0\n  3                           2                          6.83                          7   0                   0                       0\n  4                           2                          6.83                          7   0                   0                       0\n  5                           2                          6.83                          7   0                   0                       0\n  7                           2                          6.83                          7   0                   0                       0\n  14                          2                          6.83                          7   0                   0                       0\n  21                          2                          6.83                          9   1 (11.1)            1 (11.1)                1 (11.1)\n\nviruses-10-00434-t002_Table 2\n\n###### \n\nInfection, dissemination, and transmission of ZIKV in *Aedes aegypti*.\n\n  Mosquito Day Post-Feeding   Mouse Day Post-Infection   ZIKV Titer (log~10~ pfu/mL)   N    No. (%) Infection   No. (%) Dissemination   No. (%) Transmission\n  --------------------------- -------------------------- ----------------------------- ---- ------------------- ----------------------- ----------------------\n  14                          1                          5.38                          10   3 (30)              1 (10)                  0 (0)\n  21                          1                          5.38                          10   4 (40)              2 (20)                  2 (20)\n  14                          2                          6.83                          10   10 (100)            10 (100)                7 (70)\n  21                          2                          6.83                          10   10 (100)            10 (100)                7 (70)\n\nviruses-10-00434-t003_Table 3\n\n###### \n\nZIKV titer in infected tissues from *Sabethes cyaneus* and *Ae. aegypti*.\n\n  Species              Mosquito Day Post Feeding   Mouse Day Post Infection   Mean Body Titer \\* (log~10~ pfu/mL) \u00b1 1SE (N)   Mean Legs Titer \\* (log~10~ pfu/mL) \u00b1 1SE (N)   Mean Saliva Titer \\* (log~10~ pfu/mL) \u00b1 1SE (N)\n  -------------------- --------------------------- -------------------------- ----------------------------------------------- ----------------------------------------------- -------------------------------------------------\n  *Sabethes cyaneus*   21                          2                          5.4 (1)                                         4.4 (1)                                         1.1 (1)\n  *Aedes aegypti*      14                          1                          4.5 \u00b1 0.2 (3)                                   3.2 (1)                                         NA (0)\n  21                   1                           5.1 \u00b1 0.2 (4)              4.4 \u00b1 0.4 (2)                                   2.8 \u00b1 1.0 (2)                                   \n  14                   2                           5.2 \u00b1 0.2 (10)             4.4 \u00b1 0.4 (10)                                  2.8 \u00b1 0.3 (7)                                   \n  21                   2                           5.7 \u00b1 0.1 (10)             4.7 \u00b1 0.3 (10)                                  2.9 \u00b1 0.5 (7)                                   \n\n\\* Mean values are calculated from samples with detectable virus.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nParkinson\\'s disease (PD) is a progressive neurodegenerative disease characterized by motor symptoms, including bradykinesia, tremor, and muscle rigidity \\[[@B1]\\]. The neuropathological hallmark of PD is intracellular protein inclusions of misfolded *\u03b1*-synuclein, termed Lewy bodies (LB) \\[[@B2]\\]. Lewy pathology is present in the central nervous system (CNS), but can also be found throughout the peripheral autonomic nervous system (PNS) \\[[@B3]\\] and has been detected in the enteric nervous system up to 20 years prior to diagnosis \\[[@B4]\\]. Symptoms of autonomic dysfunction including constipation, decreased saliva production, urinary dysfunction, orthostatic hypotension (OH), and sweating abnormalities are common in PD patients \\[[@B5], [@B6]\\].\n\nResearch in skin abnormalities in PD patients is gaining increasing interest, especially since Lewy pathology (aggregated alpha-synuclein) can be detected in autonomic nerve fibers in most patients using simple punch biopsies \\[[@B7]\\]. Also, a significant loss of autonomic nerve fibers, including fibers with vasomotor function, is seen in many patients \\[[@B8]\\], and it has been reported that PD patients have altered skin blood flow regulation resulting in cold limbs \\[[@B9]\\]. However, our understanding of autonomic skin dysfunction in PD is still very limited.\n\nA recent study used infrared thermography and demonstrated that PD patients display abnormal thermal responses of the skin when exposed to cold stress test (CST), i.e., immersion of one hand in cold water and subsequent thermographic imaging of the recovery of skin temperature \\[[@B10]\\]. The authors reported reduced cooling and reduced thermal overshoot in the contralateral nonimmersed hand (NIH), as well as reduced recovery rates (RRs) in the immersed hand (IH) in the PD group compared to healthy controls (HCs). However, the authors only investigated thermographic data from the distal phalanx of the third finger \\[[@B10]\\], and the study has so far not been replicated.\n\nThe study aimed to determine thermographic differences in the skin post-CST between HC and PD patients on hands, feet, and trunk and to correlate findings with symptoms and signs of dysautonomia.\n\n2. Materials and Methods {#sec2}\n========================\n\n2.1. Ethics Statement {#sec2.1}\n---------------------\n\nThe study was approved by the Central Denmark Region Committee on Health Research Ethics (no. 1-10-72-277-16) and conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent.\n\n2.2. Subjects {#sec2.2}\n-------------\n\nIn this cross-sectional study, 22 PD patients and 19 age- and sex-matched HCs were included between September and December 2017. PD patients fulfilled the diagnostic criteria of PD \\[[@B11]\\]. The PD patients were recruited from an earlier in-house project, while HCs were recruited via newspaper advertisement.\n\nInclusion criteria were as follows: age 50--90 years, ability to give informed consent, and fulfillment of diagnostic criteria (PD patients). Exclusion criteria were as follows: any neurological disease (except PD), significant medical disease of any kind (liver, kidney, heart, connective tissue, or endocrinological disease), known peripheral atherosclerotic disorder or intake of vasoregulatory substances (including catecholamines, phosphodiesterase inhibitors, calcium sensitizers, vasopressors, and digoxin), fever, infection in the past 14 days, and alcohol or drug abuse. Patients refrained from eating or drinking at least one hour prior to examinations. One PD patient was excluded after having eaten within 30 minutes before thermography. Two PD patients had restricted data due to technical issues, clinical findings (edema), and/or partial digit amputation. One PD patient was subsequently excluded, since he had eaten 30\u2009min prior to thermography, leaving 21 PD patients in the study. The PD patients were studied on medication, and the L-dopa-equivalent daily dose (LEDD) was calculated.\n\n2.3. Thermography {#sec2.3}\n-----------------\n\nInfrared thermography was performed in all participants using a FLIR E60bx infrared camera (FLIR Systems, Inc. Wilsonville, Oregon, USA). A standardized examination setting was applied, which adhered closely to recommendations for infrared imaging in medicine \\[[@B12], [@B13]\\]. All subjects were studied in a 10\u2009m^2^ darkened room between 9 am and 2 pm. All CSTs were performed with the patient sitting. Temperature and air humidity were monitored; room temperature was kept at 23\u2009\u00b1\u20091\u00b0C. The recorded air humidity was 55\u2009\u00b1\u200915%.\n\nThree separate thermographic sessions were performed after acclimatization for at least 15 minutes and on the same day during a period of maximum 2 hours: (1) Regular anterior, posterior, and lateral (right and left) full-body images were acquired at baseline. (2) Baseline and post-CST images were acquired of the hands. CST was performed with the right hand immersed into cold water (3\u2009\u00b1\u20091\u00b0C) for two minutes. Water temperature was selected based on the only previous thermography study in patients with PD \\[[@B10]\\]. Water temperature was monitored using a digital thermometer prior to and during CST. Subsequently, images were obtained at 0, 2, 4, 6, 8, and 10 minutes postimmersion, comparable to previously published data \\[[@B10]\\]. (3) Baseline and post-CST images were acquired of the feet (procedure identical to CST of hands). Standardized regions of interest (ROIs) were used for analysis ([Figure 1](#fig1){ref-type=\"fig\"}). During all three sessions, the camera was mounted on a tripod. Images were obtained with the patient placed at a standardized position in the room 300\u2009cm (full body), 100\u2009cm (hands), and 100\u2009cm (feet) from the standardized position of the camera tripod. Standardized camera settings were applied: emissivity of 0.98 (human skin) \\[[@B14]\\], mean humidity of 55%, and mean reflected temperature of 20\u00b0C. There were no sources of airflow near or directly at the subject. Hands and feet were protected from direct contact with the cold water by a plastic glove/close-fitting plastic bag on the immersed hand/foot, respectively.\n\n2.4. Clinical Assessment {#sec2.4}\n------------------------\n\nMotor symptoms and disease stage of PD patients were scored using the MDS-UPDRS-III \\[[@B15]\\] and Hoehn & Yahr (H&Y) scales \\[[@B16]\\]. In all participants, blood pressure, distal blood pressure, and orthostatic hypotension test were performed. Distal blood pressures, measured by use of the strain gauge method on toes, served to exclude any subjects presenting abnormal values indicative of decreased peripheral blood pressure. Olfactory function was tested using the Sniffin\\' Sticks 16-item identification test \\[[@B17]\\]. Nonmotor symptoms were evaluated with the nonmotor symptoms questionnaire (NMSQuest) \\[[@B18]\\] and scales for outcome in Parkinson\\'s disease---autonomic (SCOPA-AUT) \\[[@B19]\\]. Sleep behavior symptoms were scored with the REM sleep behavior disorder screening questionnaire (RBDSQ) \\[[@B20]\\], and cognition was evaluated with the Montreal Cognitive Assessment scale (MoCA) \\[[@B21]\\].\n\n2.5. Data Analyses and Statistics {#sec2.5}\n---------------------------------\n\nThermographic images were analyzed using dedicated software, PMOD (PMOD Technologies LLC, Z\u00fcrich, Switzerland). Analyses were performed blinded to clinical (PD/HC) category. For each hand (NIH and IH), 13 regions of interest (ROI) were defined ([Figure 1](#fig1){ref-type=\"fig\"}). On the feet, 7 ROIs (immersed foot (IF) and nonimmersed foot (NIF)) were defined. On whole-body images at baseline, 3 ROIs on each leg/lateral side of the lower extremities and 4 ROIs on each upper extremity/back were defined.\n\nTo facilitate comparison to the study by Antonio-Rubio et al. \\[[@B10]\\], the skin of the dorsum of the third finger distal phalanx was included as a ROI. The main outcome was the calculated recovery rates (RRs) of the CSTs. These RRs were compared between groups and served as the main outcome when differentiating between groups. RRs were calculated as follows \\[[@B10]\\]:$$\\begin{matrix}\n{\\text{RR} = \\frac{T_{x\\text{post}} - T_{0\\text{post}}}{T_{\\text{baseline}} - T_{0\\text{post}}} \\cdot 100\\%,} \\\\\n\\end{matrix}$$in which *T* is the calibrated temperature (Celsius), and *x* indicates the time point of 2, 4, 6, 8, or 10 minutes postimmersion. Furthermore, thermal symmetry was calculated and equals the numeric value of the temperature difference between sides for each ROI.\n\nStatistical analysis was carried out in GraphPad Prism (GraphPad Software, San Diego, CA, USA). For group comparisons, unpaired *t*-tests were applied for age, BMI, and baseline temperature data; Fisher\\'s exact test was applied for sex (and other categorical variables), and nonparametric Mann--Whitney test was used for olfactory scores, MoCA, questionnaires, thermal symmetry, and RR. *Post hoc* statistical outlier analysis was applied to RR data. Post-CST group differences in RR over time were evaluated using two-way repeated-measures ANOVA. Group differences in RR were also interrogated using multivariate linear regression with correction for age, sex, and BMI. Correlations between thermal parameters and clinical parameters were interrogated using linear regression.\n\n3. Results {#sec3}\n==========\n\nDemographic data are listed in [Table 1](#tab1){ref-type=\"table\"}. The groups were well matched on sex and age, but the controls had slightly lower average BMI although not significant (*p*\u2009=\u20090.08). All participants were able to complete the CST. Representative thermal images of hands are shown in [Figure 2](#fig2){ref-type=\"fig\"}. Clinical data and baseline thermography results are shown in [Table 2](#tab2){ref-type=\"table\"}.\n\nAt baseline, thermal asymmetry was seen only on the 1^st^ distal finger phalanx (more asymmetric in HC) and on the 2^nd^ toe and lower dorsal foot (more asymmetric in PD). In CST, after correcting for BMI, sex, and age and after excluding two significant outliers in the PD group (*p*\u2009=\u20090.0021; ROUT test, GraphPad Prism), significant between-group differences were seen on all five fingers at the *T*~6~, *T*~8~, and *T*~10~ time points, except for *T*~6~ on the 1^st^ and 3^rd^ fingers. The 5^th^ finger *T*~10~ RR was 27% lower in the PD group compared to the HC group (mean RR\u2009\u00b1\u2009SD: 51\u2009\u00b1\u200918% vs. 70\u2009\u00b1\u200923%, *p*\u2009\\<\u20090.001).\n\nThermal recovery and differences between groups after hand CST are shown in [Figure 3](#fig3){ref-type=\"fig\"}. Recovery is shown for the dorsal skin on the distal phalanx of the 3^rd^ finger for comparison with previously published results by Antonio-Rubio et al. \\[[@B10]\\], and on the 5^th^ distal phalanx dorsal skin, which showed the most significant between-group difference in our data.\n\nNo significant differences were seen in CST of the foot in any of the investigated ROIs (*p*\u2009\\>\u20090.05). The intermalleolar and upper foot ROIs were not analyzed, since the thermal images showed too large variance in water level and coverage of ROIs.\n\nThe RR showed positive correlation with BMI on the 3^rd^ and 5^th^ phalanx at all time points in the HC group (distal 5^th^ phalanx, 6\u2009min, *p*\u2009=\u20090.002 *R*^2^\u2009=\u20090.437), but not in the PD group. In the PD group, no significant correlations were seen between thermal baseline and CST parameters and age, UPDRS-III, H&Y, LEDD, SCOPA total score, RBDSQ score, or with diastolic or systolic blood pressure drops in the OH test (*p*\u2009\\>\u20090.05 in all tests).\n\n4. Discussion {#sec4}\n=============\n\nThe study results show significantly altered thermal skin response in PD patients compared to healthy control subjects when exposed to CST, especially on the fingers. The decreased response showed increasing significance over time. At baseline, thermal asymmetry was seen only on the 1^st^ distal finger phalanx (more asymmetric in HC) and on the 2^nd^ toe and lower dorsal foot (more asymmetric in PD). No significant between-group differences were seen in baseline or post-CST values of the feet, and no correlations were seen between questionnaire and autonomic data and thermal responses in the PD group.\n\nA previous study reported data from only the 3^rd^ finger distal phalanx, which showed decreased thermal recovery 6\u2009min postimmersion in PD patients compared to HC (RR: 29\u2009\u00b1\u200917% vs. 55\u2009\u00b1\u200928%, *p*\u2009=\u20090.002) \\[[@B10]\\]. The recovery rates in our HC group were in general in agreement with this previous study (6\u2009min post-CST, 3^rd^ finger 49%, 5^th^ finger 56%), whereas our PD cases seemed to show faster recovery (6\u2009min post-CST, 3^rd^ finger 36%, 5^th^ finger 41%). Thus, our PD patients showed a more normal response compared to the previously studied group.\n\nIt is unclear what could have caused this difference, but our patients may have been at an earlier disease stage. They had shorter disease duration (4.8\u2009y vs. 5.9\u2009y), lower LEDD (315 vs. 619\u2009mg), lower total SCOPA-AUT (13 vs. 18), and lower frequency of orthostatic hypotension (17% vs. 36%). Also, our HC group was carefully age-matched, whereas the HC group of the previous study was significantly younger (*p*\u2009=\u20090.04). In both studies, the thermography was carried out in the \"on\" state, so effects of dopaminergic replacement therapy probably did not affect the results. In summary, it is therefore possible that the previously reported thermal response to CST \\[[@B10]\\] applies to later stage PD patients, whereas the thermal response may be closer to normal at earlier stages of PD. Both the baseline and post-CST data showed considerable overlap between the PD and HC groups, and thermography therefore probably does not have diagnostic applicability---at least not with the currently implemented protocol. Of note, the recovery rate after CST varied slightly on different fingers. We do not have an explanation for this variation, and it could be a spurious finding.\n\nThe only published paper with PD patients only explored the hands after CST \\[[@B10]\\]. We wanted to exploratively study other regions. We had expected to see similar results on the CST of the feet, given that cold feet are a common complaint among PD patients and that the axons to the feet are longer than those innervating the hands. Also, previous studies have shown a drastic reduction of epidermal and piloerector nerve fibers to the distal lower extremities \\[[@B22]\\]. For these reasons, we wanted to obtain data from many different body regions to explore if other regions, including on the trunk and proximal extremities, were affected. Nevertheless, we did not see any difference in post-CST responses on the feet between PD and HC. We also did not see any thermal asymmetry on the proximal extremities and trunk ROIs, which is perhaps less surprising given that the density of epidermal, piloerector, and probably also arteriolar nerve fibers has found to be normal on the trunk of PD patients \\[[@B22]\\].\n\nWe found no correlation between baseline or post-CST thermal parameters and disease duration or motor symptom severity, LEDD, autonomic symptom burden (SCOPA-AUT), or the decline in blood pressure on tests of orthostatic hypotension. We only had three patients with OH, so it is possible that our patient sample was too healthy in this regard and that inclusion of a group of later stage cases with a higher prevalence of OH might have disclosed a significant association between thermal temperature responses and blood pressure regulation. Of note, limited correlation is also seen between MIBG heart scintigraphy values and OH parameters \\[[@B6]\\], suggesting that the cause of OH may be a complicated interplay between damage to the peripheral and central nervous systems.\n\nThis study has several limitations. The study sample size was only modest resulting in limited statistical power. However, the sample size is sufficient to conclude that thermography cannot reliably separate PD patients from controls at the individual level with the currently implemented protocol. It would have been preferable to have included a number of later stage PD patients with more progressed disease and involvement of the peripheral autonomic nervous system. The present study does not allow us to firmly conclude that thermal skin responses do not correlate with symptoms or other objective measures of autonomic dysfunction. It would also have been preferable to include skin biopsies with quantification of the amount of alpha-synuclein pathology as well as of the reduction in functional nerve fibers to arterioles. It is conceivable that a correlation may have been seen between skin pathology and denervation, assessed by histology, and the thermal skin responses. Finally, nutritional status can affect thermographic results of the skin. We did correct for BMI in our analyses, but more comprehensive indicators of nutritional status might be preferable.\n\n5. Conclusion {#sec5}\n=============\n\nWe have shown that PD patients display an attenuated thermal recovery response to cold stress test of the hands but not of the feet, which agrees with one single previously published study. The thermal parameters did not show correlations with clinical or other autonomic parameters, which may have been caused in part by our inclusion of a fairly early stage population of PD patients. Thermography may have value as a research tool to study skin involvement in PD, and further studies should investigate the relationship between pathological thermal skin responses and autonomic denervation of the skin.\n\nData Availability\n=================\n\nData can be made available by contacting the corresponding author.\n\nConflicts of Interest\n=====================\n\nThe authors declare that they have no conflicts of interest.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\nSupplementary Figure 1: recovery rate on hands in patients with Parkinson\\'s disease (PD) and healthy control subjects (HC). Data are shown for each finger phalanx region at 4 (T4), 6 (T6), 8 (T8), and 10 (T10) minutes postimmersion. Only regions with significant between-group difference are shown.\n\n###### \n\nClick here for additional data file.\n\n![Schematic illustration of analyzed regions of interest (ROIs). Proportions are not accurate.](PD2020-2349469.001){#fig1}\n\n![Thermographic images. A series of images from a healthy control ((a)--(c), RR\u2009=\u200975.0%) and a PD patient ((d)--(f), RR\u2009=\u200931.9%). Images at baseline (a, d), *T*~0~ postimmersion (b, e), and *T*~6~ postimmersion (c, f) are shown. All images are scaled 10--35\u00b0C. *Note.* The reduced recovery of temperature in the PD patient on the right immersed hand (IH).](PD2020-2349469.002){#fig2}\n\n![Cold stress test. Thermal recovery on the phalanx of the 3^rd^ finger (a) and the phalanx of the 5^th^ finger (b). Absolute temperature\u2009\u00b1\u2009SEM.](PD2020-2349469.003){#fig3}\n\n  Demographics\u2009                                    PD patients (*n*\u2009=\u200921)   Healthy controls (*n*\u2009=\u200919)   *p* value\n  ------------------------------------------------ ------------------------ ----------------------------- --------------\n  Age, years: mean\u2009\u00b1\u2009SD (range)                    66.7\u2009\u00b1\u20097.5 (52--84)      66.1\u2009\u00b1\u20095.4 (54--74)           0.8\n  Gender, M/F                                      13/8                     12/7                          \\>0.99\n  BMI, kg/m^2^: mean\u2009\u00b1\u2009SD                          26.6\u2009\u00b1\u20093.7               24.5\u2009\u00b1\u20093.5                    0.080\n  UPDRS, median (range)                            20 (8--33)               NA                            NA\n  H&Y stage, I/II/III                              7/12/2                   NA                            NA\n  Duration of disease, months: mean\u2009\u00b1\u2009SD (range)   57.9\u2009\u00b1\u200934.8 (17--149)    NA                            NA\n  Worse hemibody, right/left/symmetrical           13/8/0                   NA                            NA\n  LEDD, mg: median (range)                         315 (15--1155)           NA                            NA\n  SCOPA-AUT, mean\u2009\u00b1\u2009SD                             12.5\u2009\u00b1\u20096.8               3.6\u2009\u00b1\u20093.4                     **\\<0.0001**\n  RBDSQ, mean\u2009\u00b1\u2009SD                                 3.9\u2009\u00b1\u20093.1                0.7\u2009\u00b1\u20091.2                     **0.0002**\n\nNA: not available.\n\n  Examinations, baseline                                               PD patients (*n*\u2009=\u200921)   Healthy controls (*n*\u2009=\u200919)   *p* value\n  -------------------------------------------------------------------- ------------------------ ----------------------------- --------------\n  Arterial hypertension, yes/no                                        11/10                    9/10                          \\>0.999\n  Orthostatic hypotension^1^, yes/no                                   3/18                     5/14                          0.442\n  Tympanic temperature, \u00b0C: mean\u2009\u00b1\u2009SD                                  36.7\u2009\u00b1\u20090.3               36.6\u2009\u00b1\u20090.2                    0.195\n  Subjective hyposmia, yes/no                                          19/2                     2/17                          **\\<0.0001**\n  Olfactory function, Sniffin\\' sticks; median (IQR)                   7 (5.5--10)              12 (11--14)                   **\\<0.0001**\n  Baseline temperature, 3^rd^ finger distal phalanx, \u00b0C: mean\u2009\u00b1\u2009SD     27.8\u2009\u00b1\u20093.7               29.2\u2009\u00b1\u20093.1                    0.1814\n  Baseline temperature, 5^th^ finger distal phalanx, \u00b0C: mean\u2009\u00b1\u2009SD     27.4\u2009\u00b1\u20093.7               28.9\u2009\u00b1\u20093.2                    0.1749\n  Thermal symmetry^2^, 1^st^ finger distal phalanx, \u00b0C: median (IQR)   0.22 (0.12--0.63)        0.52 (0.32--1.0)              **0.029**\n  Thermal symmetry^2^, lower dorsal foot, \u00b0C: median (IQR)             1.13 (0.49--1.98)        0.46 (0.33--0.69)             **0.002**\n  Thermal symmetry^2^, 2^nd^ toe, \u00b0C: median (IQR)                     0.55 (0.32--1.89)        0.14 (0.03--0.63)             **0.012**\n\n^1^Orthostatic hypotension: decline of \u226520/10 in mmHg in systolic/diastolic blood pressure. ^2^Thermal symmetry\u2009=\u2009temperature difference between contralateral ROIs, calculated from all ROIs on trunk/extremities. Only ROIs with significant baseline symmetry differences are listed.\n\n[^1]: Academic Editor: H\u00e9lio Teive\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nAPP plays a central role in the pathogenesis of both sporadic and familial AD. Indeed, *APP* mutations that alter APP processing either protect from sporadic AD or cause familial AD; additionally, mutations in genes that regulate APP processing -such as *PSENs* and *BRI2/ITM2B*- cause FAD, FBD and FDD \\[[@R1]-[@R12]\\].\n\nAs briefly mentioned above, mutations in *BRI2/ITM2B* cause the AD-like autosomal dominant FBD and FDD \\[[@R5], [@R7]\\]. FBD is characterized by the early onset of personality changes, memory and cognitive deficits, spastic rigidity, and ataxia \\[[@R5]\\]. FDD patients present early onset cataracts, deafness, progressive ataxia and dementia \\[[@R7]\\]. BRI2 is a type II membrane protein of 266 amino acids that is cleaved at the C terminus into a peptide of 23 amino acids (Bri23) plus a membrane-bound mature BRI2 (mBRI2) product \\[[@R13], [@R14]\\]. In FBD patients, a point mutation at the stop codon of *BRI2* results in a read-through of the 3\u2032-untranslated region and the synthesis of a BRI2 molecule containing 11 extra amino acids at the COOH terminus. Cleavage by convertases generates a normal mBRI2 plus a longer peptide, the ABri peptide. FDD is caused by a10- nucleotide duplication before the stop codon of the *BRI2* gene, which leads to the synthesis of a longer (277 amino acids) mutant protein \\[[@R7], [@R15]\\]. Convertase-mediated processing of the Danish mutant protein generates a longer C-terminal fragment, called ADan, and a normal mBRI2 polypeptide. Both ABri and ADan are deposited as amyloid fibrils. Of note, ADan deposits together with APP-derived Amyloid (A\u03b242) peptides forming wide-spread amyloid angiopathy in the small blood vessels and capillaries of the cerebrum, choroid plexus, cerebellum, spinal cord, and retina \\[[@R15]\\]. Overall, FBD and FDD patients present cognitive dysfunctions and neuropathology including neurodegeneration, amyloid, and neurofibrillary tangles \\[[@R7], [@R15]-[@R17]\\], which are similar to those of Alzheimer\\'s patients.\n\nKnock-in mice models of FDD and FBD (FDD~KI~ and FBD~KI~ mice) showed that the mutant BRI2 proteins are mainly targeted for degradation, leading to a loss of mBRI2 function and, consequently, increased APP processing. Of note, loss on mBRI2 and increased APP processing was also detected in brain lysates from FDD and FBD patients. These alterations in APP processing, and not amyloid lesions, mediate memory and synaptic plasticity deficits caused by *BRI2/ITM2B* mutations. In fact, synaptic and memory deficits or FDD~KI~ mice were reduced by inhibition of \u03b2--cleavage of APP, which generates the fragments \u03b2-CTF and sAPP\u03b2 \\[[@R18]\\], while they were worsened by inhibition of \u03b3-secretase, which cleaves \u03b2-CTF into A\u03b2 and AID/AICD \\[[@R6], [@R19]-[@R27]\\].\n\nThese findings suggest that increases in \u03b2-CTF can be neurotoxic and predicts that reducing \u03b3-cleavage of APP has pathogenic consequences, while enhancing clearance of \u03b2-CTF is therapeutically advantageous. Several data support the first hypothesis: 1) loss of \u03b3-secretase in the mouse brain induces neurodegeneration, memory and synaptic plasticity deficits; 2) Presenilins mutations associated with FAD cause a loss of \u03b3-secretase function \\[[@R28]-[@R36]\\]; 3) sub-chronic administration of GSIs impairs normal cognitive function in APP transgenic mice \\[[@R37]\\]; 4) the GSI Semagacestat exacerbated cognitive deficits and impaired activities of daily living in human AD patients \\[[@R38]\\].\n\nThe evidence that reduction of \u03b3-secretase activity impairs cognitive functions is consistent with a negative effect of \u03b2-CTF. Yet, \u03b3-secretase cleaves other type I trans-membrane proteins, including the APP-like Protein 1 and -2, Neuregulin-1 and Notch \\[[@R39]-[@R44]\\] and cognitive functions deficits prompted by \u03b3-secretase inhibition could be caused by decreased processing of any combination of \u03b3-secretase substrates.\n\n\u03b3-secretase is a multi-molecular complex comprising the catalytic subunits PSEN1 or PSEN2 and three accessory proteins: Anterior Pharynx-Defective 1 (Aph1), Nicastrin and Presenilin Enhancer Protein 2 (PEN2) \\[[@R45]-[@R47]\\]. Humans have two *APH1* genes (*APH1A* and *APH1B*) \\[[@R48], [@R49]\\]; rodents have three because of a duplication of *Aph1B* that gave rise to the *Aph1C* gene \\[[@R50], [@R51]\\]. Aph1A-containing \u03b3-secretase complexes are essential for Notch processing, while APP and Neuregulin-1 are preferred substrates of \u03b3-secretase complexes containing either Aph1B or Aph1C; hence, *Aph1BC^\u2212/\u2212^* mice show increased \u03b2-CTF but decreased A\u03b2 peptide due to reduced \u03b3-processing of \u03b2-CTF \\[[@R50]-[@R53]\\] and allow studying the consequence of inactivating \u03b3-processing of APP limiting the confounding effects of inhibition of processing of other \u03b3-secretase substrates.\n\nPharmacological evidence suggests a synaptic-toxicity of \u03b2-CTF. Here, we have tested this hypothesis genetically. If increases in \u03b2-CTF prompt learning and memory deficits, *Aph1BC^\u2212/\u2212^* mice may show deficits similar to those observed in FDD~KI~ mice; additionally, deletion of *Aph1BC* may worsen the defects of FDD~KI~ mice. On the contrary, if Danish mice develop learning and memory defects due to over-production of A\u03b2, the *Aph1BC^\u2212/\u2212^* mutation will ameliorate learning and memory deficits of FDD~KI~ mice.\n\nRESULTS {#s2}\n=======\n\nYoung mice carrying the FDD mutation and deletion of *Aph1B/C* show mild learning and memory deficits {#s2_1}\n-----------------------------------------------------------------------------------------------------\n\nMice were first tested at four months of age for anxiety-like behavior on the elevated rero maze. We analyzed the percentage of time spent in the open areas of the elevated zero maze during the 5-min testing period. While FDD~KI~*/Aph1BC^\u2212/\u2212^* mice spent more time in the open areas on average than mice of the other genotypes, one-way ANOVA revealed no significant effect of genotype, F(3, 60) = 1.94, *p* = 0.1331. Twelve animals (3 WT, 2 FDD~KI~, 4 FDD~KI~*/Aph1BC^\u2212/\u2212^*, and 3 *Aph1BC^\u2212/\u2212^*) fell off the open areas of the maze during testing and were excluded from the data analysis. In addition, two mice (1 FDD~KI~ and 1 FDD~KI~*/Aph1BC^\u2212/\u2212^*) were excluded due to technical problems with video tracking encountered during testing (Figure [1](#F1){ref-type=\"fig\"}).\n\n![Elevated Zero Maze test on mice at 4 months of age\\\nData are expressed as means \u00b1 S.E.M. No significant effect of genotype was found.](oncotarget-07-11923-g001){#F1}\n\nNext, mice were assayed for general locomotor activity levels and anxiety-like behavior in the open field. Four mice (1 FDD~KI~ and 2 FDD~KI~*/Aph1BC^\u2212/\u2212^*) were dropped from the statistical analysis due to persistent tracking errors caused by their light-colored fur. Since the video tracking system was not able to track these mice consistently, they were not included in the subsequent experiments. In addition, one FDD~KI~ mouse was excluded from the analysis due to a one-time tracking error during testing. Analysis of the mean distance traveled during the 10-min testing period by two-way ANOVA found a significant main effect for day, F(2, 138) = 82.23, *p* \\< 0.0001, indicating habituation to the box over the three-day testing period, and a significant main effect for genotype, F(3, 69) = 3.46, *p* \\< 0.05, but no significant interaction between genotype and day, F(6, 138) = 1.15, *p* = 0.3387. *Post-hoc* comparisons (Dunnett\\'s) showed that the mean distance traveled by FDD~KI~ animals were significantly greater than that traveled by WT mice on the first (*p*\\<0.01) and second (*p*\\<0.05) days (Figure [2A](#F2){ref-type=\"fig\"}). Analysis of the time spent traveling at speed greater than 50 mm/s yielded significant main effects for day, F(2, 138) = 136.2, *p* \\< 0.0001, and for genotype, F(3, 69) = 3.97, *p* \\< 0.05, but no significant day \u00d7 genotype interaction, F(6, 138) = 0.95, *p* = 0.4583, and showed that FDD~KI~ mice spent more time traveling at speed greater than 50 mm/s than WT mice on all three days (*p*\\<0.001, Day 1; *p*\\<0.05, Days 2 and 3, Dunnett\\'s) (Figure [2B](#F2){ref-type=\"fig\"}). Analysis of the mean time spent in the center of the open field showed a significant main effect for day, F(2, 138) = 7.09, *p* \\< 0.01, and a significant day \u00d7 genotype interaction, F(6, 138) = 3.41, *p* \\< 0.01, while the main effect for genotype was close to significance, F(3, 69) = 2.47, *p* = 0.0692 (Figure [2C](#F2){ref-type=\"fig\"}). FDD~KI~ mice spent more time in the arena center than WT mice on the first two days (*p*\\<0.05, Dunnett\\'s). Analysis of the number of entries into the arena center showed significant main effects for day, F(2, 138) = 37.67, *p* \\< 0.0001, and genotype, F(3, 69) = 4.20, *p* \\< 0.01, and a near significant interaction between genotype and day, F(6, 138) = 2.04, *p* = 0.0650. *Post-hoc* comparisons (Dunnett\\'s) revealed that FDD~KI~ mice entered the arena center significantly more than did WT mice on the first (*p*\\<0.001) and second (*p*\\<0.05) days (Figure [2D](#F2){ref-type=\"fig\"}). Overall these data indicate that FDD~KI~ mice were generally more active and less anxious than WT, *Aph1BC^\u2212/\u2212^* and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice, while no differences were detected between WT, *Aph1BC^\u2212/\u2212^* and FDD~KI~*/Aph1BC^\u2212/\u2212^* animals.\n\n![Open Field test on mice at 4 months of age\\\nData are expressed as means (\u00b1 S.E.M.) during the 10-min testing period over 3 days. **A.** Total distance traveled. **B.** Amount of time in which the animal ambulated at speed greater than 50 mm/s. FDD~KI~ mice were more active than WT mice, especially on Day 1. **C.** Amount of time the animal spent in the center of the arena (20 cm \u00d7 20 cm). **D.** Total number of entries into the arena center. FDD~KI~ mice entered and spent more time in the arena center on Days 1 and 2 than WT mice. \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01, \\*\\*\\* *p* \\< 0.001, WT *vs*. FDD~KI~.](oncotarget-07-11923-g002){#F2}\n\nFollowing the open field test, mice were tested in the MWM for spatial reference memory. One WT and one FDD~KI~ mouse were dropped from the experiments permanently due to severe bite wounds and persistent tracking errors caused by its light fur color, respectively. In addition, one FDD~KI~*/Aph1BC^\u2212/\u2212^* mouse died before the task was completed. As shown in Figure [3A](#F3){ref-type=\"fig\"} and [3B](#F3){ref-type=\"fig\"}, the visible platform task conducted prior to the reference memory task revealed no significant differences among the four genotype groups in path length traveled, F(3, 66) = 1.43, *p* = 0.2427, or swim speed, F(3, 66) = 0.60, *p* = 0.6177, indicating that none of the mutant mice had any visual or motor deficits relative to WT control at this age. Figure [3C](#F3){ref-type=\"fig\"} depicts the mean path length traveled by animals during the acquisition phase of the hidden platform task. Two-way ANOVA revealed a significant main effect for day on path length, F(5, 330) = 68.20, *p* \\< 0.0001, indicating animals\\' acquisition of reference memory for the platform location. There was also a significant interaction between day and genotype, F(15, 330) = 2.18, *p* \\< 0.01, while no significant main effect for genotype was found, F(3, 66) = 2.49, *p* = 0.0676. Tukey\\'s comparisons revealed some differences among the genotypes on the first day, with *Aph1BC^\u2212/\u2212^* mice traveling a significantly larger distance than WT mice (*p*\\<0.05) and FDD~KI~ mice (*p*\\<0.001), and with FDD~KI~*/Aph1BC^\u2212/\u2212^* mice than FDD~KI~ mice (*p*\\<0.01). On the probe trial conducted two days after the last acquisition session, the analysis of the percentage of time spent in the four quadrants revealed a significant main effect for quadrant, F(3, 198) = 50.71, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 66) = 0.55, *p* = 0.6469, or significant quadrant \u00d7 genotype interaction, F(9, 198) = 1.15, *p* = 0.3293 (Figure [3D](#F3){ref-type=\"fig\"}). While the percentage of time spent in the target quadrant is the most popular measure of probe trial performance \\[[@R54]\\], counting the number of times the animal crosses a small area surrounding the former platform position (counter crossings) provides more information on the spatial accuracy with which the exact location of the platform has been encoded \\[[@R55]\\]. One-way ANOVA did not reveal a significant effect of genotype on the number of counter crossings in the target quadrant, F(3, 66) = 2.00, *p* = 0.1233 (Figure [3E](#F3){ref-type=\"fig\"}). However, a separate unpaired t-test showed a significant difference between WT and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice (*p* = 0.0177, Figure [3F](#F3){ref-type=\"fig\"}). There were no significant differences in the average proximity to the original platform location among the genotype groups, F(3, 66) = 0.75, *p* = 0.5282 (Figure [3G](#F3){ref-type=\"fig\"}). Thus, at 4 months of age, when present together the FDD mutation and the *Aph1BC* deficiency mildly interfered with spatial long-term memory by compromising its accuracy.\n\n![The Morris Water Maze task shows a mild deficit in accuracy of spatial long-term memory 4 month-old FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice\\\nData are expressed as means \u00b1 S.E.M. **A.-B.** Performance on the visible platform task. The visible platform task indicates that there were no significant differences among the genotypes in path length traveled (A) or swim speed (B). **C.** Acquisition of spatial reference memory in the hidden platform task. Mean path lengths across 6 daily trials are shown. Significant differences among the genotypes were found only on the first day: \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01, \\*\\*\\* *p* \\< 0.001. **D.-F.** Performance on the 60-s probe trial given 2 days after the last acquisition session. (D-F) Performance on the 60-s probe trial given 2 days after the last acquisition session. (D) Percentage of time spent in the four quadrants. \\*\\* *p* \\< 0.01, \\*\\*\\* *p* \\< 0.001, \\*\\*\\*\\* *p* \\< 0.0001 (E) Number of counter crossings in the target quadrant (Quadrant 4). \\* *p* \\< 0.05 (F) A separate unpaired t-test showed a significant difference between WT and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice (\\* *p* = 0.0177). **G.** Average proximity to the former platform location. **H.** Performance in the reversal learning task with a new platform location. Mean path length across 6 daily trials are shown. **I.-K.** Performance on the 60-s probe trail given 2 days after the last reversal learning session. (I) Percentage of time spent in the four quadrants. \\*\\*\\*\\* *p* \\< 0.0001 (J) Number of counter crossings in the target quadrant (Quadrant 1). (K) Average proximity to the former platform location.](oncotarget-07-11923-g003){#F3}\n\nIn the reversal learning task, in which the location of the hidden platform was moved to the quadrant opposite to the original target quadrant, all the genotypes performed the task in a similar manner during the acquisition phase (Figure [3H](#F3){ref-type=\"fig\"}), with ANOVA showing a significant main effect for day, F(5, 330) = 102.5, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 66) = 0.70, *p* = 0.5577, or day \u00d7 genotype interaction, F(15, 330) = 0.71, *p* = 0.7770. On the probe trial given two days after the last reversal learning session, no differences were found among the genotypes in the percentage of time spent in the quadrants (Figure [3I](#F3){ref-type=\"fig\"}), with a significant quadrant main effect, F(3, 198) = 140.2, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 66) = 0.38, *p* = 0.7681, or quadrant \u00d7 genotype interaction, F(9, 198) = 0.55, *p* = 0.8376. Similarly, no differences were found among the genotypes in the number of counter crossings in the target quadrant (Figure [3J](#F3){ref-type=\"fig\"}), F(3, 66) = 0.95, *p* = 0.4230, or the average proximity to the target (Figure [3K](#F3){ref-type=\"fig\"}), F(3, 66) = 0.47, *p* = 0.7029.\n\nThe same cohorts of mice were again assessed for spatial reference memory in the Morris water maze at 7-8 months of age. To determine whether genetic manipulations affect learning and memory in an aging dependent manner, it is customary to use different cohorts of animal to perform a task (such as the Morris water maze) at different ages. In this manner, the results are not influenced by learning and/or habituation to the tasks that may occur when the same cohort of mice are subjected to the same task more than once. However, a longitudinal analysis is ethically and economically preferable since it reduces the number of experimental animals. In addition, it eliminates possible confounding effects due to genetic variability between different cohorts of mice. Finally, testing if mice forget, during aging, tasks learned as young adults, mirrors what happens in AD patients. In this new Morris water maze test, the number of daily trials was reduced to three to alter the difficulty of the task. One WT mouse was dropped since it had become too weak to swim. The visible platform task revealed no significant differences among the genotypes in path length traveled, F(3, 65) = 1.36, *p* = 0.2638 (Figure [4A](#F4){ref-type=\"fig\"}), or swim speed, F(3, 65) = 0.27, *p* = 0.8488 (Figure [4B](#F4){ref-type=\"fig\"}). In the reference memory task, as shown in Figure [4C](#F4){ref-type=\"fig\"}, no significant differences among the genotypes were found during acquisition, with two-way ANOVA showing a significant main effect for day, F(4, 260) = 12.41, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 65) = 0.81, *p* = 0.4932, or day \u00d7 genotype interaction, F(12, 260) = 0.95, *p* = 0.4957. The probe trial conducted two days after the last acquisition session did not reveal any differences among the genotypes in the percentage of time spent in the four quadrants (Figure [4D](#F4){ref-type=\"fig\"}), with a significant quadrant main effect, F(3, 195) = 115.4, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 65) = 1.18, *p* = 0.3257, or quadrant \u00d7 genotype interaction, F(9, 195) = 1.25, *p* = 0.2670. There was no significant effect of genotype on the number of counter crossings in the target quadrant, F(3, 65) = 1.67, *p* = 0.1828 (Figure [4E](#F4){ref-type=\"fig\"}), or the average proximity to the target, F(3, 65) = 2.18, *p* = 0.0990 (Figure [4F](#F4){ref-type=\"fig\"}). Thus, no significant differences were found among the genotypes in the acquisition or retention of spatial reference memory at 7-8 months of age, and the small deficits in counter crossing observed in 4 month-old FDD~KI~*/Aph1BC^\u2212/\u2212^* mice was not seen again at 7-8 months of age, possibly due to learning/habituation to the task.\n\n![Normal spatial memory in 7-8 month-old mice\\\nData are expressed as means \u00b1 S.E.M. **A.-B.** Performance on the visible platform task. There were no significant differences among the genotypes in path length traveled (A) or swim speed (B). **C.** Acquisition of spatial reference memory in the hidden platform task. Mean path lengths across 3 daily trials are shown. No significant differences were found among the genotypes. **D.-F.** Performance on the 60-s probe trial given 2 days after acquisition. (D) Percentage of time spent in the four quadrants. \\*\\*\\* *p* \\< 0.001, \\*\\*\\*\\* *p* \\< 0.0001. (E) Number of counter crossings in the target quadrant (Quadrant 2). (F) Average proximity to the former platform location. No significant differences were found among the genotypes in any of the measures.](oncotarget-07-11923-g004){#F4}\n\nFollowing the completion of the probe trial at 7-8 months, mice were subjected to a working memory task in the eight-arm radial arm water maze. Mice were initially given four consecutive daily acquisition trials with a 15 seconds inter-trial interval. However, it became apparent after six days of training with this procedure that most mice lacked the physical strength required for maintaining their performance in the maze during the four consecutive acquisition trials, and that their performance was not improving after six days (data not shown). Accordingly, the procedure was modified such that mice would be given a 6-min inter-trial interval in the holding cage, and that three, instead of four, acquisition trials would be given before the retention trial. After a break on the seventh day, mice were tested for five more days with this new procedure and the results pertaining to these last five days are shown in Figure [5](#F5){ref-type=\"fig\"}. Repeated measures one-way ANOVA (RM one-way ANOVA) found a significant main effect for trial in WT \\[F (2.065, 33.04) = 18.37, *p* \\< 0.0001)\\], FDD~KI~ \\[F (2.320, 37.12) = 19.88, *p* \\< 0.0001)\\] and *Aph1BC^\u2212/\u2212^* \\[F (2.312, 39.31) = 22.14, *p* \\< 0.0001)\\], but not in FDD~KI~*/Aph1BC^\u2212/\u2212^* mice \\[F (2.033, 28.46) = 3.218, p=0.0542)\\]. A *post hoc* multiple comparison of the mean of each trial to the mean of every other trial (Tukey\\'s multiple comparisons test) indicated that performance by WT and FDD~KI~ mice improved significantly between Trials 1 and 2 (*p* \\< 0.05 for WT and *p* \\< 0.01 for FDD~KI~), Trials 1 and 3 (*p* \\< 0.001 for WT and *p* \\< 0.01 for FDD~KI~) and Trials 1 and R (*p* \\< 0.0001 for both genotypes). The performance of *Aph1BC^\u2212/\u2212^* mice improved significantly between Trials 1 and 3, 1 and R (*p* \\< 0.0001) as well as 2 and 3 (*p* \\< 0.05),and 2 and R (*p* \\< 0.01). On the other hand, this *post hoc* multiple comparison analysis was not performed for the FDD~KI~*/Aph1BC^\u2212/\u2212^* mice since the RM one-way ANOVA analysis did not find a significant main effect for trial in these mice. Overall, these observations indicated that the magnitude of improvement in performance was the smallest in FDD~KI~/*Aph1BC^\u2212/\u2212^* mice, compared to WT, FDD~KI~, or *Aph1BC^\u2212/\u2212^* mice. This difference may reflect working memory impairment in FDD~KI~*/Aph1BC^\u2212/\u2212^* mice, or might have resulted from the fact that FDD~KI~/*Aph1BC^\u2212/\u2212^* mice actually made fewer errors than mice of the other genotypes on the first trial, thereby making the gain smaller.\n\n![The 8-arm Radial Arm Water Maze task detects possible working memory impairment in FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice at 7-8 months of age\\\nData are expressed as means \u00b1 S.E.M. **A.** Number of errors on 3 daily acquisition trials averaged across the last 3 test days. The performance by WT, FDD~KI~, and *Aph1BC^\u2212/\u2212^* mice improved across trials to a greater degree, compared to FDD~KI~*/Aph1BC^\u2212/\u2212^* mice. \\**p* \\< 0.05, \\*\\**p* \\< 0.01, \\*\\*\\**p* \\< 0.001, \\*\\*\\*\\* *p* \\< 0.0001. **B.** Number of errors on trials 2, 3 and retention (R) given 30 min after the third acquisition trial. No significant differences were found among the genotypes.](oncotarget-07-11923-g005){#F5}\n\nThe FDD and *Aph1BC^\u2212/\u2212^* mutations induce mild deficits in spatial long-term memory in middle age {#s2_2}\n--------------------------------------------------------------------------------------------------\n\nAt 12 months of age, another reference memory task was given. For this task, the number of daily trials was further reduced to two per day, and the length of the probe trial was reduced to 30 s. In addition, the visible platform task was given after the completion of the probe trials. One WT mouse that displayed extensive and persistent thigmotaxis during the reference memory task was excluded from the data analysis. Also, one *Aph1BC^\u2212/\u2212^* mouse was eliminated because it had developed rectal prolapse. Again, the visible platform task did not find any differences among the genotypes in path length traveled, F(3, 63) = 0.81, *p* = 0.4911 (Figure [6A](#F6){ref-type=\"fig\"}), or swim speed, F(3, 63) = 0.81, *p* = 0.4908 (Figure [6B](#F6){ref-type=\"fig\"}). In the reference memory task, mice of all the genotype learned the location of the hidden platform in a similar manner (Figure [6C](#F6){ref-type=\"fig\"}). ANOVA found a significant main effect for day, F(7, 441) = 11.84, *p* \\< 0.0001, but no significant genotype main effect, F(3, 63) = 0.10, *p* = 0.9576, or day \u00d7 genotype interaction, F(21, 441) = 0.75, *p* = 0.7773. On the probe trial given two days after the last acquisition session, no significant differences were found among the genotypes in the analysis of the percentage of time spent in the quadrants (Figure [6D](#F6){ref-type=\"fig\"}), with a significant main effect for quadrant, F(3, 189) = 35.30, *p* \\< 0.0001, but no significant genotype main effect, F(3, 63) = 1.21, *p* = 0.3123, or quadrant \u00d7 genotype interaction, F(9, 189) = 0.65, *p* = 0.7571. However, the analysis of the number of counter crossings in the target quadrant revealed a significant effect of genotype, F(3, 63) = 5.80, *p* \\< 0.01 (Figure [6E](#F6){ref-type=\"fig\"}). *Post-hoc* comparisons (uncorrected Fisher\\'s LSD) revealed that WT mice crossed the counter in the target quadrant significantly more often than FDD~KI~ mice (*p* \\< 0.001), *Aph1BC^\u2212/\u2212^* (*p* \\< 0.01) and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice (*p* \\< 0.05). There was no significant effect of genotype on the average proximity to the former platform location, F(3, 63) = 0.79, *p* = 0.5065 (Figure [6F](#F6){ref-type=\"fig\"}). Thus, at 12 months of age, the FDD mutation and the *Aph1BC* deficiency both mildly interfered with spatial long-term memory by compromising its accuracy.\n\n![Mild deficit in accuracy of spatial long-term memory in 15 month-old FDD~KI~, *Aph1BC*^\u2212/\u2212^ and FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice\\\nData are expressed as means \u00b1 S.E.M. **A.-B.** Performance on the visible platform task. There were no significant differences among the genotypes in path length traveled (A) or swim speed (B). **C.** Acquisition of spatial reference memory in the hidden platform task. Mean path lengths across 2 daily trials are shown. No significant differences were found among the genotypes. **D.-F.** Performance on the 30-s probe trial given 2 days after acquisition. **D.** Percentage of time spent in the four quadrants. \\*\\*\\* *p* \\< 0.001, \\*\\*\\*\\* *p* \\< 0.0001. **E.** Number of counter crossings in the target quadrant (Quadrant 1). WT mice crossed the target counter significantly more than FDD~KI~ or *Aph1BC^\u2212/\u2212^* mice. \\*\\*\\* *p* \\< 0.01, WT *vs*. FDD~KI~, \\*\\* *p* \\< 0.01, WT *vs*. *Aph1BC^\u2212/\u2212^*; \\* *p* \\< 0.05, WT *vs*. FDD~KI~/*Aph1BC^\u2212/\u2212^*. **F.** Average proximity to the former platform location.](oncotarget-07-11923-g006){#F6}\n\nMice were tested for spatial working memory in Morris water maze at 15 months of age. As shown in Figure [7A](#F7){ref-type=\"fig\"} and [7B](#F7){ref-type=\"fig\"}, the visible platform task given at this age found no significant effect of genotype on path length traveled, F(3, 64) = 0.37, *p* = 0.7743, or swim speed, F(3, 64) = 0.66, *p* = 0.5797. In the working memory task, in which the location of the hidden platform was changed daily, we measured the path-length to target (Figure [7C](#F7){ref-type=\"fig\"}). RM one-way ANOVA found a significant main effect for trial in WT \\[F (1.984, 29.76) = 5.004, *p* = 0.0136)\\], but not in FDD~KI~ \\[F (1.948, 31.17) = 1.822, *p* = 0.1793)\\], *Aph1BC^\u2212/\u2212^* \\[F (1.943, 31.09) = 2.749, *p* = 0.0809\\] and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice \\[F (1.655, 26.48) = 2.180, *p=*0.14)\\]. A *post hoc* multiple comparison of the mean of each trial to the mean of every other trial (Tukey\\'s multiple comparisons test) indicated that only the performance of WT mice improved significantly between Trials 1 and 3 (*p* \\< 0.05). This test suggests that, at 15 months of age, the FDD mutation and the *Aph1BC* deficiency both mildly interfered with working memory.\n\n![Mild deficit in working memory in FDD~KI~, *Aph1BC*^\u2212/\u2212^ and FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice at 15 months of age\\\nData are expressed as means \u00b1 S.E.M. **A.-B.** Performance in the visible platform task. There were no significant differences among the genotypes in path length traveled (A) or swim speed (B). **C.** Path length traveled on the 3 daily trials of the working memory task averaged across the last 3 test days. WT mice significantly reduced the path length traveled to reach the platform between trial 1 and 3 (\\* *p* \\< 0.05), while mice of the other 3 genotypes did not.](oncotarget-07-11923-g007){#F7}\n\nLearning and memory deficits are caused by the Aph1BC deficiency and the FDD mutation at old age (18-19 months) {#s2_3}\n---------------------------------------------------------------------------------------------------------------\n\nAt 18-19 months of age, mice were given a series of tests. The first test, conducted at 18-19 months of age, evaluated mice in the two-trial Y-maze test task for spatial recognition memory. One FDD~KI~/*Aph1BC^\u2212/\u2212^* mouse had died before reaching this age. Figure [8A](#F8){ref-type=\"fig\"} depicts the mean number of arm entries during the 5-min test trial, which is an index for animals\\' total activity levels. ANOVA found a small but significant effect of genotype, F(3, 63) = 2.77, *p* =0.0490 \\< 0.05, and Fisher\\'s LSD comparison test showed that *FDD~KI~* mice were more mobile than WT and *FDD~KI~/Aph1BC^\u2212/\u2212^* mice (*p* \\< 0.05). Fisher\\'s test was used here because, despite the significant overall genotype effect, no significant differences among the genotypes were detected by corrected comparison tests. Given the significant genotype effect on the number of total arm entries, the percentage of entries into each arm, instead of raw numbers of arm entries, was used to analyze animals\\' preference for the novel arm vs. the known arm. As shown in Figure [8B](#F8){ref-type=\"fig\"}, the analysis of the percentage of entries into the novel and known arms found a significant main effect for arm, F(1, 63) = 43.34, *p* \\< 0.0001, but no significant genotype main effect, F(3, 63) = 0.78, *p* =0.5076, or arm \u00d7 genotype interaction, F(3, 63) = 1.96, *p* =0.1295. Sidak\\'s multiple comparisons between the arms revealed that the novel arm was entered significantly more than the known arm by WT mice and FDD~KI~ mice, but not by FDD~KI~*/Aph1BC^\u2212/\u2212^* mice or *Aph1BC^\u2212/\u2212^* mice, and that the level of statistical significance was much higher for WT mice (*p* \\< 0.0001) than for FDD~KI~ mice (*p* \\< 0.01). As can be seen in Figure [8C](#F8){ref-type=\"fig\"}, the analysis of the time spent in the novel and known arms found a significant main effect for arm, F(1, 63) = 29.63, *p* \\< 0.0001, as well as a significant interaction between arm and genotype, F(3, 63) = 2.82, *p* \\< 0.05, while showing no significant main effect for genotype, F(3, 63) = 0.43, *p* = 0.7345. *Post-hoc* comparisons across the genotypes (Dunnett\\'s) revealed that WT mice spent significantly more time in the novel arm than did FDD~KI~*/Aph1BC^\u2212/\u2212^* mice (*p* \\< 0.05). In addition, comparisons between the arms (Sidak\\'s) showed that WT mice spent significantly more time in the novel arm than in the known arm (*p* \\< 0.0001). FDD~KI~ mice also spent significantly more time in the novel arm than in the known arm, but to a much smaller degree (*p* \\< 0.05) than WT mice. By contrast, the difference in time spent between the novel and known arms were not statistically significant for FDD~KI~*/Aph1BC^\u2212/\u2212^* or *Aph1BC^\u2212/\u2212^* mice. In sum, the results of the two-trial Y-maze task demonstrated that WT mice and, to a lesser extent, FDD~KI~ mice distinguished between the novel and familiar arms better than FDD~KI~*/Aph1BC^\u2212/\u2212^* or *Aph1BC^\u2212/\u2212^* mice after a 1-h retention interval, indicating that the *Aph1BC* deficiency, and to a lesser extent the FDD mutation, leads to impairment of short-term spatial recognition memory.\n\n![The two-trial Y-maze test showed mild deficit of short-term spatial recognition memory in FDD~KI~, *Aph1BC*^\u2212/\u2212^ and FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice at 18-19 months of age\\\nData are expressed as means \u00b1 S.E.M. **A.** Total number of arm entries. FDD~KI~ mice made significantly more arm entries than WT or FDD~K*I*~*/Aph1BC^\u2212/\u2212^* mice (\\**p* \\< 0.05). **B.** Percentage of entries into the novel (N) and known (K) arms. WT and, to a smaller degree, FDD~KI~ mice entered the novel arm significantly more than the known arm, while FDD~KI~*/Aph1BC^\u2212/\u2212^* or *Aph1BC^\u2212/\u2212^* mice did not (\\*\\**p* \\< 0.01; \\*\\*\\*\\* *p* \\< 0.0001). **C.** Time spent in the novel and known arms. WT mice and, to a smaller degree, FDD~KI~ mice spent significantly more time in the novel arm than in the known arm, while FDD~KI~*/Aph1BC^\u2212/\u2212^* or *Aph1BC^\u2212/\u2212^* mice did not (\\**p* \\< 0.05; \\*\\*\\*\\* *p* \\< 0.0001). In addition, there was a significant difference in time spent in the novel arm between WT and FDD~KI~/*Aph1BC^\u2212/\u2212^* (\\**p* \\< 0.05).](oncotarget-07-11923-g008){#F8}\n\nNext we analyzed the general locomotor activity levels in the open field at 18-19 months. There was a main effect for day in the total distance traveled, F(2, 126) = 85.01, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 63) = 2.03, *p* = 0.1186, or interaction between day and genotype, F(6, 126) = 1.83, *p* = 0.0977 (Figure [9A](#F9){ref-type=\"fig\"}). There was a main effect for day in the amount of time in which the animal ambulated at speed greater than 50 mm/s, F(2, 126) = 94.25, *p* \\< 0.0001, but no significant genotype main effect, F(3, 63) = 2.24, *p* = 0.0919, or day \u00d7 genotype interaction, F(6, 126) = 1.62, *p* = 0.1466 (Figure [9B](#F9){ref-type=\"fig\"}). We detected a significant main effect for day, F(2, 126) = 10.66, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 63) = 0.09, *p* = 0.9667, or day \u00d7 genotype interaction, F(6, 126) = 0.66, *p* = 0.6770 in the mean time spent in the center of the arena (Figure [9C](#F9){ref-type=\"fig\"}). The total number of entries into the arena center showed significant main effect for day, F(2, 126) = 35.27, *p* \\< 0.0001, but no significant genotype main effect, F(3, 63) = 1.51, *p* = 0.2212, or day \u00d7 genotype interaction, F(6, 126) = 0.45, *p* = 0.8419 (Figure [9D](#F9){ref-type=\"fig\"}).\n\n![Open Field test on mice at 18-19 months of age\\\nData are expressed as means (\u00b1 S.E.M.) during the 10-min testing period over 3 days. **A.** Total distance traveled. **B.** Amount of time in which the animal ambulated at speed greater than 50 mm/s. **C.** Amount of time the animal spent in the center of the arena (20 cm \u00d7 20 cm). **D.** Total number of entries into the arena center. No significant differences were found among the genotypes in any of the measures.](oncotarget-07-11923-g009){#F9}\n\nFollowing the open field test, the spontaneous alternation test was conducted in the Y-maze to assess animals\\' spatial working memory. Analysis of the total number of arm entries during the 5-min testing period showed no significant effect of genotype was detected, F(3, 62) = 0.36, *p* = 0.7811 (Figure [10A](#F10){ref-type=\"fig\"}). Since the minimum number of arm entries needed for a complete alternation is three, six mice that had less than three arm entries in total (2 WT, 1 FDD~KI~, 1 FDD~KI~/*Aph1BC^\u2212/\u2212^*, and 2 *Aph1BC^\u2212/\u2212^*) were not included in the analysis of the percentage of alternations, which showed no significant effect of genotype, F(3, 57) = 1.26, *p* = 0.2969 (Figure [10B](#F10){ref-type=\"fig\"}).\n\n![Y-maze spontaneous alternation test at 18-19 months of age\\\nData are expressed as means \u00b1 S.E.M. **A.** Total number of arm entries. **B.** Percentage of alternations. Six mice with less than three arm entries are not included in (B). No significant differences were found among the genotypes in either measure.](oncotarget-07-11923-g010){#F10}\n\nNext, we used the elevated zero maze to test for anxiety-like behavior. As shown in Figure [11](#F11){ref-type=\"fig\"}, while *Aph1BC^\u2212/\u2212^* mice spent more time in open areas than mice of the other genotypes on average, the genotype effect did not reach statistical significance, F(3, 60) = 1.396, *p* = 0.2527. Three animals (2 FDD~KI~/*Aph1BC^\u2212/\u2212^* and 1 *Aph1BC^\u2212/\u2212^*) fell off the open areas of the maze during testing and were excluded from the data analysis.\n\n![Elevated Zero maze test on mice at 18/-9 months of age\\\nMean (\u00b1 S.E.M.) percentage of time spent in the open areas of the elevated zero maze. There were no significant differences among the genotypes.](oncotarget-07-11923-g011){#F11}\n\nMice were then evaluated for spatial working memory in the 6-arm radial arm water maze. RM one-way ANOVA found a significant main effect for trial in WT \\[F (2.025, 32.40) = 28.12, *p* \\< 0.0001)\\], FDD~KI~ \\[F (2.082, 33.32) = 13.32, *p* \\< 0.0001)\\], *Aph1BC^\u2212/\u2212^* \\[F (2.437, 38.99) = 4.542, *p* = 0.012)\\] and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice \\[F (2.738, 41.07) = 4.680, *p=*0.0081)\\] (Figure [12A](#F12){ref-type=\"fig\"}). A *post hoc* multiple comparison of the mean of each trial to the mean of every other trial (Tukey\\'s multiple comparisons test) indicated that performance by WT mice improved significantly between Trials 1 and 2 (*p* \\< 0.01), 1 and 3 (*p* \\< 0.001), 1 and R (*p* \\< 0.0001) as well as between Trials 2 and R (*p* \\< 0.01). The FDD~KI~ mice improved between Trials 1 and 3 and 1 and R as well as between Trials 2 and 3 and 2 and R (*p* \\< 0.01). As for FDD~KI~/*Aph1BC^\u2212/\u2212^* and *Aph1BC^\u2212/\u2212^* mice, their performance improved significantly between Trials 1 and R (*p* \\< 0.05). However, it is worth noting that *Aph1BC^\u2212/\u2212^* mice actually made fewer errors than mice of the other genotypes on the first trial, thereby making the gain smaller. On the other hand, the performance by FDD~KI~*/Aph1BC^\u2212/\u2212^* mice significantly improved between Trials 1 and 2,1 and R, but the degree of significance was much smaller (*p* \\< 0.05) as compared to WT and FDD~KI~ mice.\n\n![Mild spatial working memory deficits in 18-19 month-old FDD~KI~, *Aph1BC*^\u2212/\u2212^ and FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice\\\nData are expressed as means \u00b1 S.E.M. (A) Number of errors on 3 daily acquisition trials averaged across days 6-9 of testing. (B) Number of errors on trials 2, 3 and retention (R) given 30 min after the third acquisition trial. The visible platform task showed no significant differences among the genotypes in path length traveled (C) or swim speed (D). \\**p* \\< 0.05, \\*\\**p* \\< 0.01, \\*\\*\\**p* \\< 0.001, \\*\\*\\*\\* *p* \\< 0.0001.](oncotarget-07-11923-g012){#F12}\n\nWe also compared the number of errors made by mice of the four genotypes at Trials 2, 3 and R (Figure [12B](#F12){ref-type=\"fig\"}). Ordinary one-way ANOVA found significant effect of genotype at Trials 3 \\[F (3, 63) = 2.751, *p* = 0.05)\\] and R \\[F (3, 63) = 3.616, *p* = 0.00178)\\], but not 2 \\[F (3, 63) = 0.9047, *p* = 0.4439)\\]. A *post hoc* multiple comparison of the mean of each genotype to the mean of every other genotype (Tukey\\'s multiple comparisons test) indicated that WT mice committed significantly fewer errors than FDD~KI~*/Aph1BC^\u2212/\u2212^* mice both at Trials 3 (*p* \\< 0.05) and R (*p* \\< 0.01). The differences between WT *vs.* FDD~KI~, WT *vs. Aph1BC^\u2212/\u2212^*, FDD~KI~*/Aph1BC^\u2212/\u2212^ vs.* FDD~KI~ and FDD~KI~*/Aph1BC^\u2212/\u2212^ vs. Aph1BC^\u2212/\u2212^* were not statistically significant. Altogether, these observations indicated that the degree of improvement in performance across trials during acquisition and retention was the greatest for WT mice, followed by FDD~KI~, *Aph1BC^\u2212/\u2212^* and FDD~KI~/*Aph1BC^\u2212/\u2212^* mice in the order named. These results further indicate that the FDD mutation and deletion of *Aph1BC* induces spatial working memory deficits. These data are also in accordance with the results of a previous study showing that the *Aph1BC^\u2212/\u2212^* mutation induces spatial working memory deficits in mice of the F2 generation of the C57BL/6J--129/Ola hybrids \\[[@R53]\\].\n\nAfter the radial arm water maze test, mice were assessed for possible visual or motor deficits in the visible platform task. Ten animals (2 WT, 1 FDD~KI~, 5 FDD~KI~*/Aph1BC^\u2212/\u2212^*, and 2 *Aph1BC^\u2212/\u2212^*) were mistakenly sacrificed prematurely after the radial arm water maze and before the visual platform task. As shown in Figure [12C](#F12){ref-type=\"fig\"} and [12D](#F12){ref-type=\"fig\"}, the visible platform task did not reveal any visual or motor deficits at this age. There was no significant effect of genotype on path length traveled, F(3, 53) = 1.13, *p* = 0.3467, or swim speed, F(3, 53) = 2.19, *p* = 0.0999.\n\nFinally, mice were tested for contextual and cued fear memory using the fear conditioning paradigm. As is indicated in Figure [13A](#F13){ref-type=\"fig\"} and [13B](#F13){ref-type=\"fig\"}, in the test for contextual fear memory conducted 24 h after conditioning, WT mice exhibited freezing behavior significantly more than mice in the other genotype groups. ANOVA revealed a significant effect of genotype on the mean percentage of freezing during the last 3 min of the test session, F(3, 53) = 5.45, *p* \\< 0.01, and *post-hoc* comparisons (Tukey\\'s) indicated that WT mice froze significantly more than FDD~KI~ (*p* \\< 0.01), *Aph1BC^\u2212/\u2212^* (*p* \\< 0.01), and FDD~KI~*/Aph1BC^\u2212/\u2212^* (*p* \\< 0.05) mice (Figure [13A](#F13){ref-type=\"fig\"}). As can be seen in Figure [13B](#F13){ref-type=\"fig\"}, the analysis of the time course of the percentage of freezing in 1-min bins during the contextual test also showed a significant main effect for genotype, F(3, 53) = 4.75, *p* \\< 0.01, in addition to a significant main effect for time, F(4, 212) = 20.55, *p* \\< 0.0001, while finding no significant time \u00d7 genotype interaction, F(12, 212) = 0.89, *p* = 0.5557. Tukey\\'s multiple comparison test further indicated that WT mice froze significantly more than FDD~KI~ mice in the last three minutes (*p* \\< 0.01 for the fourth min; *p* \\< 0.05 for the third and fifth min), more than *Aph1BC^\u2212/\u2212^* mice in the first minute (*p* \\< 0.05) and the last three minutes (*p* \\< 0.01 for the third and fourth min; *p* \\< 0.05 for the fifth min), and more than FDD~KI~*/Aph1BC^\u2212/\u2212^* mice in the third minute (*p* \\< 0.05). In the test for cued fear memory conducted 24 h after the contextual test, no significant effect of genotype was found on the percentage of freezing during the presentation of the tone CS, F(3, 53) = 2.27, *p* = 0.0910 (Figure [13C](#F13){ref-type=\"fig\"}). However, as can be seen in Figure [13D](#F13){ref-type=\"fig\"}, which shows the time course of freezing during tone presentation, while WT mice remained frozen throughout the 3-min presentation of the tone, FDD~KI~ mice, which were initially as frozen as WT mice, became more mobile towards the end of this period, and FDD~KI~*/Aph1BC^\u2212/\u2212^* and *Aph1BC^\u2212/\u2212^* mice froze less than WT mice for the entire period. Yet, the analysis of the time course showed a significant main effect for time, F(2, 106) = 6.97, *p* \\< 0.01, but no significant main effect for genotype, F(3, 53) = 2.27, *p* = 0.0910, or time \u00d7 genotype interaction, F(6, 106) = 2.09, *p* = 0.0603. Nevertheless, the time \u00d7 genotype interaction approached significance, and when one-way ANOVA was performed on the percentage of freezing during the first, second, and third 1-min time periods separately, while no significant genotype effect was detected in the first minute, F(3, 53) = 2.47, *p* = 0.0720, or the second minute, F(3, 53) = 1.34, *p* = 0.2712, there was a significant genotype effect in the third minute, F(3, 53) = 3.25, *p* \\< 0.05. Dunnett\\'s comparison test further revealed a significant difference between WT and FDD~KI~*/Aph1BC^\u2212/\u2212^* mice (*p* \\< 0.05), as well as between WT and *Aph1BC^\u2212/\u2212^* mice (*p* \\< 0.05), in the last minute of tone presentation. Lastly, a day after the completion of the test for cued fear memory, all the experimental mice were evaluated for shock sensitivity. As shown in Figure [13E](#F13){ref-type=\"fig\"}, mice of all the genotypes reacted to electric shock at the four intensity levels in a similar manner, with ANOVA showing a significant main effect for shock level, F(3, 159) = 97.70, *p* \\< 0.0001, but no significant main effect for genotype, F(3, 53) = 1.60, *p* = 0.2010, or shock level \u00d7 genotype interaction, F(159, 159) = 1.45, *p* = 0.1709. While the amygdala is required and sufficient for fear conditioning to discrete sensory cues such as a tone, rapid formation of conditioned fear responses to environmental context is additionally mediated by the hippocampus \\[[@R56]-[@R59]\\]. Thus, our results indicate that both the FDD and *Aph1BC^\u2212/\u2212^* mutations interfere with mnemonic information processing at the level of the hippocampus for the formation of long-term contextual fear memory at this age, while the *Aph1BC^\u2212/\u2212^* mutation additionally has adverse effects on fear memory formation in the amygdala.\n\n![Deficit in long-term contextual fear memory in FDD~KI~, *Aph1BC*^\u2212/\u2212^ and FDD~KI~/*Aph1BC*^\u2212/\u2212^ mice at 18-19 months of age\\\nData are expressed as means \u00b1 S.E.M. **A.** Percentage of freezing during the last 3-m period of the contextual test. WT mice froze significantly more than mice of the other genotypes. \\**p* \\< 0.05, WT *vs*. FDD~KI~/*Aph1BC^\u2212/\u2212^*; \\*\\* *p* \\< 0.01 WT *vs*. FDD~KI~ and WT *vs*. *Aph1BC^\u2212/\u2212^*. **B.** Time course of freezing behavior during the contextual test in 1-m time bins. There was a significant genotype main effect (*p* \\< 0.01), with WT freezing more than the other genotypes. \\**p* \\< 0.05, \\*\\* *p* \\< 0.01, WT *vs*. FDD~KI~; \u00a3 *p* \\< 0.05, \u00a3\u00a3 *p* \\< 0.01, WT *vs*. *Aph1BC^\u2212/\u2212^*; \\# *p* \\< 0.05, WT *vs*. FDD~KI~*/Aph1BC^\u2212/\u2212^*. **C.** Percentage of freezing during the 3-m tone presentation in the altered context in the cued test. WT mice froze more than mice of the other genotypes, although not significantly. **D.** Time course of freezing behavior during tone presentation in the cued test. Only WT mice remained frozen for the entire period. **E.** Sensitivity to varying intensity levels of foot shock. Mice of all the genotypes reacted to shock at the four intensity levels in a similar manner.](oncotarget-07-11923-g013){#F13}\n\nDISCUSSION {#s3}\n==========\n\nThe clinical symptoms of AD include progressive loss of memory, thinking and language skills, as well as other behavioral changes. Short-term memory is the first to fail in AD patients. These clinical symptoms are accompanied by neuronal degeneration. Although it is widely believed that the disease is precipitated by the insurgence of brain lesions, such as A\u03b2 amyloid plaques and neurofibrillary tau tangles, whether these alterations of tau and A\u03b2 metabolism cause the debilitating clinical symptoms of AD and FDD is still unclear.\n\nBecause of the uncertainty concerning the pathogenic biochemical mechanisms of AD and related dementias, and considering that to improve the quality of life of AD patients -and caregivers- we need to either reverse or slow down the progression of the clinical symptomatology, we focused our analysis on learning and memory in our animal model of disease. In this study, the same cohort of mice of WT, FDD~KI~, FDD~KI~/*Aph1BC^\u2212/\u2212^*, and *Aph1BC^\u2212/\u2212^* mice was assessed longitudinally for possible deficits in learning and memory from \\~4 to \\~19 months of age. FDD~KI~/*Aph1BC^\u2212/\u2212^* mice presented mildly compromise accuracy of spatial long-term memory and working short-term memory impairments at 4 and 8 months of age, respectively. Deficits in spatial and working memory were, at later ages, observed also in single mutant FDD~KI~ and *Aph1BC^\u2212/\u2212^* mice. Analysis of older animals showed that spatial recognition memory and long-term contextual fear memory were impaired in FDD~KI~, FDD~KI~/*Aph1BC^\u2212/\u2212^*, and *Aph1BC^\u2212/\u2212^* mice. Overall, this study shows that; 1) the *Aph1BC^\u2212/\u2212^* mutation does not rescue memory deficits in *FDD~KI~* mice, as would have been expected if A\u03b2 played a pathogenic role in *FDD~KI~* mice; 2) in contrast, the FDD mutation and the *Aph1BC^\u2212/\u2212^* deletion cause similar behavioral deficits in spatial memory and appear to have a small additive negative effect, at least on spatial long-term memory and working short-term memory, that is more pronounced at younger ages. Performing the two-trial Y-maze and the fear conditioning tasks in younger mice could unveil whether the FDD mutation and the *Aph1BC* deletion have also a negative additive effect on spatial recognition memory and long-term contextual fear memory. These observations are consistent with a pathological role of \u03b2-CTF, which is augmented in FDD~KI~ mice due to increased production and in *Aph1BC^\u2212/\u2212^* mice due to reduced turnover.\n\nWe have previously shown that inhibition of \u03b2-processing of APP rescues memory and synaptic impairments of FDD~KI~ mice acutely and transiently, suggesting that increased \u03b2-cleavage of APP, perhaps in the synaptic cleft, during synaptic events leading to LTP and memory acquisition, leads to memory/synaptic deficits in FDD~KI~ mice \\[[@R25], [@R26]\\]. This evidence, together with the data showing memory deficits in *Aph1BC^\u2212/\u2212^* mice, suggest that either increasing the rate of production (like in FDD~KI~ mice) or decreasing the rate of clearance (as in *Aph1BC^\u2212/\u2212^* mice) of \u03b2-CTF during synaptic transmission might lead to cognitive impairments. Future experiments will have to test whether these hypotheses are correct and whether the additive adverse effects on spatial long-term memory in young mice of the FDD~KI~ and the *Aph1BC^\u2212/\u2212^* mutations are due to a synergistic effect on \u03b2-CTF levels transiently produced during synaptic transmission.\n\nFDD~KI~ and *Aph1BC^\u2212/\u2212^* mice show a few distinct phenotypes. In fear conditioning tests, mice carrying the *Aph1BC* deletion (FDD~KI~*/Aph1BC^\u2212/\u2212^* and *Aph1BC^\u2212/\u2212^* mice) showed a mild impairment of cued fear memory, a task completely dependent on the functional integrity of the amygdala, as compared to FDD~KI~ and WT animals. This result is not surprising since several differences exist between FDD~KI~ and *Aph1BC^\u2212/\u2212^* mice. First, the Aph1B/C deficiency causes a reduced clearance of \u03b2-CTF that leads to accumulation of \u03b2-CTF and a concomitant reduction in the \u03b2-CTF metabolites AID/AICD and A\u03b2. On the contrary, in FDD~KI~ mice the loss of mBRI2 leads to an increased production of all these APP metabolites. Second, in *Aph1BC^\u2212/\u2212^* mice processing of another \u03b3-secretase substrate, Neuregulin-1 \\[[@R53]\\] is reduced. Thus, behavioral deficits caused by the deletion of *Aph1B* and *Aph1C*, but not by the FDD mutation, could be attributed to reduction in AID/A\u03b2 and/or reduction of processing of Neuregulin-1.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nSubjects {#s4_1}\n--------\n\nAll behavioral experiments were conducted by using male littermates of the F2 generation of the C57BL/6J--129 hybrid mice as subjects. Mice were generated and maintained at the Animal facility of the Albert Einstein College of Medicine. Four genotypes, *Aph1BC^\u2212/\u2212^*, *FDD~KI~*, *FDD~KI~/Aph1BC^\u2212/\u2212^*, and wild type (WT), of F2 mice (n=11-19 per genotype) were evaluated for behavior. *Aph1BC^\u2212/\u2212^* mice have been previously described \\[[@R51]\\]. *FDD~KI~* mice carried one mutant and one wild type *BRI2/ITM2b* allele \\[[@R19]\\]. Upon weaning, all mice were implanted with electronic chips (PharmaSeq, Monmouth Junction, NJ) subcutaneously on the tail for identification purposes, and their identity was regularly checked during testing periods. Animals were group-housed in plastic cages with *ad libitum* access to food and water in a temperature- and humidity-controlled animal care facility with a 12-h light/12-h dark cycle. All experimental procedures were in accordance with the National Institutes of Health guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) at the Albert Einstein College of Medicine in animal protocol number 20130509.\n\nBehavioral experimental procedures {#s4_2}\n----------------------------------\n\nAll mice were extensively handled prior to the start of behavioral testing. All behavioral testing was conducted during the light cycle. On each testing day, animals were transported to a behavioral testing suite in their home cages and allowed to acclimate for at least 30 min prior to the start of testing. The experimenter was not blind to the genotypes of the animals in tests conducted at 4, 7-8, 12 and 15 months of age but was made blind in tests conducted at 18-19 months. All measurements were taken automatically by video tracking software.\n\nElevated zero maze {#s4_3}\n------------------\n\nMice were assessed for anxiety-like behavior on the elevated zero maze initially at 4 months of age and again at 18-19 months. The zero maze (Stoelting, Wood Dale, IL) consisted of an annular platform (inner diameter 50 cm, width 5 cm) elevated to 50 cm above the ground level, divided equally into four quadrants. Two opposite quadrants were enclosed by walls (15 cm high) on both the inner and outer edges of the platform (closed areas), while the remaining two opposite quadrants were open without walls (open areas). Light levels over the maze were kept constant at approximately 50 lx in the open areas and 30 lx in the closed areas. Mice were placed individually in a closed quadrant and allowed to explore the maze freely for 5 min. The behavior of mice was monitored using a video camera, and their movements were analyzed with a video tracking system (ANY-maze, Stoelting). The percentage of time spent in the open and closed areas was used as measures of anxiety-like behavior, with larger time in the open arms indicating lower levels of anxiety.\n\nOpen field {#s4_4}\n----------\n\nThe open field test was conducted to assess animals\\' general locomotor activity, exploratory behavior, and anxiety-like behavior at 4 and 18-19 months of age. The open field apparatus (Stoelting) consisted of a square open field (40 cm \u00d7 40 cm) surrounded by opaque walls (35 cm high) and was dimly lit with a single light bulb directly above the apparatus, which illuminated the arena at approximately 5 lx in the center and 9 lx in corners. Each mouse was placed in the center of the open field box and allowed to explore the box freely for 10 min. The total distance traveled and the number of entries into, and the time spent in, the center of the arena (20 cm \u00d7 20 cm) were recorded with the ANY-maze video tracking system. This was repeated for three consecutive days to assess how animals would habituate to the increasingly familiar environment.\n\nMorris water maze {#s4_5}\n-----------------\n\nMice were tested in the Morris water maze for spatial reference memory at 4, 7-8, and 12 months and for spatial working memory at 15 months of age. The water maze consisted of a circular tank (120 cm in diameter) filled with water made opaque with nontoxic white paint and maintained slightly above the room temperature (25 \u00b1 2\u00b0C). All water maze tasks involved the animal finding a circular platform (10 cm in diameter) submerged in the water in order to escape from the water. On each trial, the mouse was released into the water facing the wall of the pool and allowed to swim freely in the pool to find the platform for the maximum of 60 s. Once the animal located the platform, it was allowed to stay on it for 15 s. Mice that did not locate the platform within 60 s were guided to the platform and allowed to stay on it for 15 s. After 15 s on the platform, the animals were removed from the pool, gently dried with paper towels, and placed in a single holding cage under a heat lamp until the next trial. A video tracking system (HVS 2020 and 2014; HVS Image, Mountain View, CA) was used to measure parameters such as the distance traveled, escape latency, swimming speed, the percentage of time spent in the quadrants, the number of counter crossings, and the average proximity to the platform location. The experimenter\\'s position was maintained at the southeast (SE) corner of the room far from the tank for all the water maze tasks conducted at different ages. The experimenter was visible to the animal but remained stationary.\n\nVisible platform task {#s4_6}\n---------------------\n\nAt each age that mice were tested in the water maze, either before (4 and 7-8 months) or after (12, 15 and 18-19 months) the memory task, a visible platform task, in which the platform was made visible by attaching a small flag (7 cm \u00d7 5 cm) to it, was conducted to examine whether mice had any visual, motor, or motivational deficits at that particular age. At 4, 7-8, 12, and 15 months of age, two or three daily sessions, with three or four trials a day, were given, in which both the platform location and the starting position were changed in a semi-random manner between trials to ensure that the animal was using the proximal cue (i.e., flag) to locate the platform. At 18-19 months, two three-trial sessions were given in a single day. The distance traveled to the platform (path length) and swimming speed were measured by the HVS video tracking system. The data collected from the last session were used for data analysis.\n\nReference memory task {#s4_7}\n---------------------\n\nThe Morris water maze hidden platform task was performed to assess spatial reference memory at 4, 7-8, and 12 months of age. In this task, the platform was hidden 1 cm below the water level, and distal visual cues were placed on the walls surrounding the pool. The location of the hidden platform remained constant across the acquisition sessions, while the starting position was varied in a pseudo-random manner between trials within each session. The distance traveled by the mouse to reach the platform was recorded by the HVS video tracking system. At 4 months, mice received two daily sessions of three trials each with an inter-trial interval of 6-10 min and an inter-session interval of 3 h for six consecutive days. The platform was located at the center of the fourth quadrant between the center and the northwest (NW). Following a probe trial given two days after the last acquisition session of this initial reference memory task, mice received a reversal learning task for another six days, in which the location of the hidden platform was moved to the quadrant opposite to the original target quadrant (i.e., the first quadrant). At 7-8 months, the number of daily trials was reduced to three, with an inter-trial interval of 6-10 min, and the task was run for five consecutive days. The platform was placed at the center of the second quadrant between the center and the southeast (SE). Mice were tested again at 12 months and received two trials per day with an inter-trial interval of 6-10 min for eight consecutive days. The platform position was at the center of the first quadrant between the center and the northeast (NE). Two days after the last acquisition session, a single probe trial was given, during which the platform was removed from the pool, and each mouse was released from the quadrant opposite to the trained platform location and allowed to search the pool for 60 s (4 and 7-8 months) or 30 s (12 months). The time spent in the target quadrant, where the platform had been located prior to its removal, the number of crossings of a circular area encompassing the original platform location (counter: 2 \u00d7 platform diameter), and the average proximity to the former platform location were measured to assess the animal\\'s spatial reference memory for the location of the hidden platform.\n\nWorking memory task {#s4_8}\n-------------------\n\nMice were tested in the Morris water maze for working memory at 15 months of age. In the working memory task, the position of the hidden platform varied from day to day but remained in the same place throughout the trials of a given day. The starting position was pseudo-randomly changed from trial to trial within a given day. On each day, four trials (1 cued trial+ 3 test trials) were given. On the cued trial, each mouse was placed on the platform for 20s, after which it was removed from the platform to a single holding cage, where it remained for 5 min. After 5 min, three test trials were given, with an inter-trial interval of 6-8 min. On each test trial, the animal was allowed to swim freely to find the hidden platform for the maximum of 60 s. Any mouse not locating the platform within 60 s was guided to the platform and allowed to stay on it for 15 s. Mice were tested for 10 days, with a one-day break after the seventh day. The distance traveled for each mouse was averaged over the last three days of testing and used for statistical analysis.\n\nRadial arm water maze {#s4_9}\n---------------------\n\nMice were tested in the radial arm water maze for spatial working memory at 7-8 and 18-19 months of age. An eight-arm radial arm water maze (each arm 8 cm in width and 38 cm in length) was used at 7-8 months. A six-armed radial arm water maze (each arm 20 cm in width and 30 cm in length) was used at 18-19 moths. The radial maze was placed into the same water tank as used for the Morris water maze, filled with opaque water (25 \u00b1 2\u00b0C). The height of the walls of the maze was 8 cm above the water level. Distal visual cues were placed on the walls of the testing room. A clear submerged platform (square, 8 cm \u00d7 8 cm for the 8-arm maze; circular, 10 cm in diameter for the 6-arm maze) was placed at the end of one of the arms. In the working memory task, the platform would remain in the same arm for all trials on a given day but its location was changed pseudo-randomly from day to day. On each trial (maximum time 60 s), the mouse was released from one of the non-goal arms and allowed to swim freely to locate the platform. The release arm was pseudo-randomly changed from trial to trial. A mouse was charged with one error each time it entered an arm other than the goal arm or did not enter any arm for 15 s. The trial continued for 60 s or until the mouse ascended the platform. If a mouse did not locate the platform within 60 s, it was guided to the platform. The mouse was removed after 15 seconds on the platform. After three acquisition trials (inter-trial interval 6-8 min), the mouse was placed in a single holding cage for 30 min, after which it was given a fourth retention trial. The error scores for each mouse was averaged over the last three days of testing and used for statistical analysis. Mice were tested until the mean number of errors over three days for WT mice reached performance criteria (2.5 on Trial 3).\n\nY-maze {#s4_10}\n------\n\nMice were tested in the Y-maze for spatial recognition memory and spatial working memory at 18-19 months. The Y-maze consisted of three arms of equal length (35 cm) and width (5 cm), which were interconnected at 120\u00b0 and enclosed by walls (10 cm high). The inside of the arms were identical, providing no intra-maze cues. The maze was placed under a bright fluorescent light and was surrounded by distal visual cues.\n\nTwo-trial test {#s4_11}\n--------------\n\nThe two-trial test test was conducted to assess short-term spatial recognition memory at 18-19 months. During the first trial (training trial), one of the arms of the maze was blocked, and mice were placed into one of the remaining arms of the maze (start arm) and allowed to explore the unblocked two arms for 10 min. After a 1-hr inter-trial interval, the blocked arm was opened (novel arm), and mice were placed in the start arm and allowed to explore freely all three arms of the maze for 5 min (test trial). The number of entries into and the amount of time spent in each arm were registered by the ANY-maze video tracking system. The relative position of the novel vs. known arms (i.e., left or right) was counterbalanced within each genotype to reduce place preference effects. This test takes advantage of the innate tendency of mice to explore novel unexplored areas (e.g., the previously blocked arm). Mice with intact recognition memory will prefer to explore a novel arm over the familiar arms, whereas mice with impaired spatial memory will enter all arms randomly.\n\nSpontaneous alternation test {#s4_12}\n----------------------------\n\nTen to thirteen days after the two-trial test, the spontaneous alternation Y-mazetest was conducted to assess spatial working memory. Mice were released to the center of the Y maze with all three arms open and allowed to explore freely for 5 min. The number and the sequence of arm entries were recorded by a video tracking system (ANY-maze). The dependent variables were activity, defined as the number of arms entered, and percent alternation, which was calculated as the number of alternations (entries into three different arms consecutively) divided by the total possible alternations (i.e., the number of arms entered minus 2) and multiplied by 100. For efficient alternation, mice need to use working memory to maintain an ongoing record of most recently visited arms, and a mouse with impaired working memory cannot remember which arm it has just visited and shows decreased spontaneous alternation accordingly.\n\nFear conditioning {#s4_13}\n-----------------\n\nMice were tested for contextual and cued fear memory at 18-19 months. Fear conditioning was conducted in a mouse conditioning chamber (18 cm \u00d7 20 cm \u00d7 28 cm) with a metal grid floor, lit with a single house light and enclosed within a sound-attenuating cubicle (Coulbourn Instruments, Whitehall, PA). The floor grid was connected to a shocker (Coulbourn Instruments) for the delivery of an electric foot shock, which was to be used as an unconditioned stimulus (US). The chamber was also equipped with a speaker connected to an amplifier for the delivery of a pure tone (2.8 kHz, 85 dB), which served as a conditioned stimulus (CS) for cued fear conditioning. The same conditioning chamber was used for testing for contextual fear memory, while, in testing for cued fear memory, the chamber was altered with a rectangular partition placed at a diagonal, wall and floor covers with novel texture, and a novel (vanilla) scent. On the first day, mice were individually placed in the conditioning chamber, and the house light was immediately turned on. One hundred and twenty seconds later, animals were presented with a continuous tone for 30 s, at the end of which an electric shock (0.6 mA) was delivered through the floor grid for 2 s and co-terminated with the tone. Mice remained in the chamber for another 30s before being removed to the home cage. Approximately 24 hr after the training session, animals were tested for contextual fear memory. The mouse was placed in the same conditioning chamber as had been used for training and observed for freezing behavior in the absence of any shock or tone for 5 min. The last 3-min period constituted testing for contextual fear memory. Approximately 24 hr after the test for contextual fear memory, mice were tested for cued fear memory. The animal was placed in the modified chamber (novel environment) and observed for freezing behavior for 2 min. After 2 min, the animal was presented with the tone CS continuously for 3 min, during which time it was again observed for freezing behavior. The last 3-min period with the tone presentation constituted testing for cued fear memory. In both tests, each animal\\'s movements were recorded and the percentage of freezing was calculated by FreezeFrame software (Coulbourn Instruments).\n\nAt the completion of the test, mice were assessed for possible genotype effects on shock sensitivity. A sequence of single foot shocks was delivered to animals placed in the same chamber used for fear conditioning at four intensity levels (0.1, 0.2, 0.4, and 0.6 mV) in the ascending order. There were two presentations at each shock intensity level, with a 20-s inter-stimulus interval. At each intensity level, the animal\\'s behavior was evaluated using the following scale to determine the threshold to each of these behavioral responses (0=no response, 1=ambulation, 2=flinch, 3=hop, 4=run, and 5=jump).\n\nStatistical analysis {#s4_14}\n--------------------\n\nStatistical analysis of most data was performed by analysis of variance (ANOVA), with one between-subjects factor (genotype) and, when appropriate, a within-subjects factor (e.g., day). When significant effects were found, the data were further analyzed by *post hoc* comparison tests (Tukey\\'s, Sidak\\'s, Dunnett\\'s, or Fisher\\'s LSD). The level of significance was set at *p* \\< 0.05. Statistical analyses were carried out using the Prism software (GraphPad, La Jolla, CA).\n\nWe thank Bart De Strooper and Lutgarde Serneels for providing the *Aph1B/C^\u2212/\u2212^* mice.\n\n**CONFLICTS OF INTEREST**\n\nThe AECOM has a patent on the commercial use of FDD~KI~ mice. Luciano D\\'Adamio is a co-inventor on this patent. AECOM has licensed the patent to Remegenix, a company of which Luciano D\\'Adamio is a co-founder and a Board member. As a co-founder Luciano D\\'Adamio owns 35% of Remegenix. The patent and the licensing only covers commercial use of the mice and does not pose any obstacle to distribution of the mice to academic laboratories. There are no further patents, products in development or marketed products to declare. This does not alter the authors\\' adherence to all the Oncotarget\\'s policies on sharing data and materials, as detailed online in the guide for authors.\n\n**GRANT SUPPORT**\n\nThe work was funded by the following grants: R01AG052286, D\\'ADAMIO, LUCIANO (PD/PI), R01AG041531, D\\'ADAMIO, LUCIANO (PD/PI), R01AG033007, D\\'ADAMIO, LUCIANO (PD/PI).\n\n**Author contributions**\n\nF.B. planned, performed and analyzed all the experiments; K.I. planned and conducted some of the experiments with F.B. and analyzed the data; D.D.P. helped generating the mice; L.D. conceived the experiments and produced the mouse strains, K.I., F.B. and L.D. prepared the figures and wrote the paper. All authors reviewed the manuscript.\n"}
+{"text": "Introduction {#s1}\n============\n\n*Barley stripe mosaic virus* is the type species of the genus *Hordeivirus*. Barley stripe mosaic virus (BSMV) infects barley, wheat, and oats under natural conditions, and numerous other monocots and dicots by artificial inoculation ([@CIT0006]; [@CIT0016]). BSMV has recently been isolated from 750-year-old barley grains found near the Nile River ([@CIT0054]), and consists of a large number of strains that were early subjects of phenotypic and host-range studies ([@CIT0035]). BSMV has a positive-sense tripartite RNA genome (RNA\u03b1, -\u03b2, and -\u03b3) that encodes seven major proteins ([@CIT0006]; [@CIT0016]). RNA\u03b1 directs synthesis of the \u03b1a protein, which functions as the methyltransferase/helicase subunit of the replicase \\[RNA-dependent RNA polymerase (RdRp)\\]. RNA\u03b2 encodes \u03b2a \\[coat protein (CP)\\], and an overlapping triple gene block (TGB) sequence encoding three major movement proteins (TGB1, TGB2, and TGB3) that are expressed from two subgenomic RNAs. RNA\u03b3 serves as the mRNA for the \u03b3a RdRp polymerase subunit, and the \u03b3b protein, which functions as a suppressor of RNA silencing and a modulator of host defences ([@CIT0016]).\n\nThe BSMV TGB1 protein is a multifunctional 58kDa protein with RNA-binding, RNA helicase, and ATPase activities ([@CIT0011]; [@CIT0058]). The C-terminal region contains seven conserved helicase motifs, and mutations within one or more of these motifs have been shown to be involved in enzymatic and movement functions and RNA-binding activities ([@CIT0024]; [@CIT0006]). TGB2 and TGB3 are trans-membrane proteins that are integrated in the endoplasmic reticulum bilayer ([@CIT0039]). Cell-to-cell movement of BSMV does not require the CP, and this property has permitted isolation of nucleoprotein complexes composed of the TGB1 protein and viral genomic and subgenomic RNAs ([@CIT0006]). The TGB1 protein interacts directly with the TGB3 protein and indirectly with the TGB2 protein to form heterologous complexes required for co-localization of the TGB proteins at the plasmodesmata (PD) and BSMV cell-to-cell movement ([@CIT0024]; [@CIT0006]; [@CIT0016]; [@CIT0016]). Our recent results have shown that TGB1 proteins function in eliciting resistance to BSMV strains that are unable to infect *Brachypodium distachyon* inbred lines containing the *Bsr1* resistance gene. In the case of the North Dakota 18 (ND18) strain TGB1 protein, amino acid residues at positions 390 and 392 are critical for TGB1 protein genetic interactions with *Bsr1* and for inducing resistance responses ([@CIT0009]; [@CIT0009]).\n\nAlthough the studies mentioned above provide valuable insights into BSMV cell-to-cell movement processes, little information is available about post-translational biochemical events, such as phosphorylation, that may function in regulating intercellular macromolecular trafficking. Movement protein (MP) phosphorylation was first demonstrated to be important for virus cell-to-cell movement during investigations with the tobacco mosaic virus (TMV) 30kDa MP, and this study provided important approaches for subsequent MP analyses ([@CIT0008]). Several proteins may function in TMV movement, because a cell-wall-associated kinase ([@CIT0008]), a PD-associated protein kinase (PAPK1) ([@CIT0026]), an endoplasmic reticulum-associated kinase ([@CIT0019]), and protein kinase CK2 (formerly known as casein kinase II; [@CIT0014]) have been shown to be involved in *in vitro* and *in vivo* phosphorylation of the 30kDa protein ([@CIT0008]). Moreover, mimicking MP phosphorylation by negatively charged amino acids inhibited MP transport through PD and delayed TMV and potyvirus spread in *Nicotiana tabacum* ([@CIT0060]; [@CIT0019]). More recently, phosphorylation of the MPs of other viruses, such as tomato mosaic virus ([@CIT0021]; [@CIT0033]), potato leafroll virus ([@CIT0030]), Abutilon mosaic virus ([@CIT0023]), brome mosaic virus ([@CIT0001]), apple chlorotic leaf spot virus ([@CIT0051]), and cucumber mosaic virus ([@CIT0034]), has been shown to either enhance or inhibit virus movement during infection.\n\nAlthough little information is available about the roles of phosphorylation in the movement processes of TGB MPs, potato virus X TGB1 protein is efficiently phosphorylated by *N. tabacum* protein kinase CK2 ([@CIT0037]). Furthermore, tyrosine phosphorylation regulates the functions of potato mop-top virus TGB3 protein because substitution of tyrosine residues in two phosphorylation domains enhances interactions between the TGB3 and TGB2 proteins and inhibits virus cell-to-cell movement ([@CIT0050]). In addition, the N-terminal half of the TGB1 protein of Poa semilatent virus (PSLV), a hordeivirus closely related to BSMV, has been reported to be phosphorylated *in vitro* by casein kinase 1 (CK1), protein kinase A (PKA), and protein kinase C (PKC)-like kinases present in *N. benthamiana* cell-wall fractions. In the case of PSLV, phosphorylation of an internal domain decreases RNA-binding activity and homologous protein--protein interactions, but experiments to determine whether these activities affect movement have not been conducted ([@CIT0031]). Here, we present the first evidence that BSMV TGB1 protein is phosphorylated *in vitro* and *in vivo* by the host protein kinase CK2. Our biochemical and molecular approaches demonstrated that Thr-401 in the TGB1 C-terminal region is a major phosphorylation site of the TGB1 protein, and that the Thr-395 residue serves as a CK2 docking domain. Mutational analyses of these residues indicate that a phosphorylation-dependent mechanism is involved in BSMV local and systemic infections in monocots and dicots.\n\nMaterials and methods {#s2}\n=====================\n\nPlant growth conditions {#s3}\n-----------------------\n\n*N. benthamiana* plants were grown in a climate chamber at 23 to 25 \u00b0C with a 14/10h light (\\~75 mmol m^--2^ s^--1^)/dark photoperiod as described previously ([@CIT0064]). Barley (Yangpi 8), wheat (Yangmai 11), and *B. distachyon* Bd21 plants were grown in a greenhouse until the two-leaf stage, and then inoculated and transferred to a climate chamber at the same temperature and light regimen as above until evaluated at 5 to 12 d post-inoculation (dpi).\n\nConstruction of infectious clones of BSMV Xinjiang (~XJ~BSMV) strain {#s4}\n--------------------------------------------------------------------\n\nGenomic (g) RNAs of ~XJ~BSMV strain were extracted from purified virus with Trizol (Life Technologies) and used to prime reverse transcription of the gRNAs with primer BS32 as described previously ([@CIT0064]; [@CIT0009]). BSMV \u03b1, \u03b2, and \u03b3 cDNAs were amplified with the primer pairs XJ-1/BS32, XJ-2/BS32, and XJ-3/BS32, respectively ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online), and inserted into the pMD20-T vector (Takara) to generate pT7-\u03b1~XJ~, pT7-\u03b2~XJ~, and pT7-\u03b3~XJ~. Site-specific mutagenesis was carried out with a QuikChange Site-Directed Mutagenesis kit (Agilent Technologies) to make alanine (A), aspartic acid (D), and glutamic acid (E) substitutions for ~XJ~TGB1 protein residues 395 and 401 in pT7-\u03b2~XJ~ with the corresponding primer pairs ([Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online). These clones and all those described below were verified by DNA sequencing (Tsingke Biotech, Beijing).\n\nTo engineer ~XJ~BSMV derivatives for agroinfiltration, full-length cDNAs were amplified from pT7-\u03b1~XJ~, pT7-\u03b2~XJ~ (or the site-specific pT7-\u03b2~XJ~ TGB1 mutants), and pT7-\u03b3~XJ~ clones with the primer pairs CH-10/BS-26, CH-11/BS-26, and CH-12/BS-26, respectively ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)). The cDNAs then were inserted between the *Stu*I and *Bam*HI sites of pCass4-Rz ([@CIT0002]), and the resulting clones and site-specific mutants with the 395 and 401 residue substitutions were designated pCa-\u03b1~XJ~, pCa-\u03b2~XJ~, pCa-\u03b3~XJ~, and pCa-\u03b2~XJ~TGB1 mutants.\n\nMechanical inoculation of *in vitro* transcripts and agroinfiltration of BSMV derivatives {#s5}\n-----------------------------------------------------------------------------------------\n\nThe pT7-\u03b1~XJ~, pT7-\u03b2~XJ~, and pT7-\u03b3~XJ~ plasmids were linearized with *Spe*I or *Bam*HI and used as templates for T7 RNA polymerase (Promega) *in vitro* transcription of capped infectious RNAs ([@CIT0045]). The \u03b1, \u03b2, and \u03b3 *in vitro* transcripts were mixed in equal amounts with FES inoculation buffer (0.1M glycine, 0.06M potassium phosphate, 1% sodium pyrophosphate decahydrate, 1% bentonite, 1% celite, pH 8.5) and used immediately for inoculation of 7- to 10-d-old barley, wheat, and *B. distachyon* Bd21. Plasmids pCa-\u03b1~XJ~, pCa-\u03b2~XJ~, and pCa-\u03b3~XJ~ were maintained in *Agrobacterium tumefaciens* strain EHA105 and infiltrated into the lower side of *N. benthamiana* leaves as described previously ([@CIT0064]).\n\nImmunoprecipitation {#s6}\n-------------------\n\nImmunoprecipitation assays were carried out with minor modifications of a published protocol ([@CIT0047]). *N. benthamiana* leaf sections were harvested at 5 d after agroinfiltration and proteins were extracted in 2 vols (w/v) of GTEN buffer \\[10% (v/v) glycerol, 25mM Tris/HCl (pH 7.5), 1mM EDTA, 150mM NaCl, 10mM dithiothreitol, 2% (w/v) polyvinylpolypyrrolidone, 1% Protease Inhibitor Cocktail (Roche)\\]. Extracted complexes were stirred for 30min and centrifuged at 12 000*g* for 30min at 4 \u00b0C. The supernatants were first incubated with TGB1 protein antibody at 4 \u00b0C for 6h, and then mixed with protein G--agarose (Millipore) beads to enrich the ~XJ~TGB1 protein. The immunoprecipitation products were washed five times with immunoprecipitation buffer \\[GTEN buffer with 0.15% (v/v) NP-40, 0.5mM dithiothreitol\\] and evaluated by Western blotting with anti-TGB1 and anti-phosphothreonine antibody (\u03b1-pT; Millipore) at 1:500 dilutions.\n\nConstruction of protein expression vectors for *in vitro* phosphorylation, glutathione S-transferase (GST), and His-tagged pull-down assays {#s7}\n-------------------------------------------------------------------------------------------------------------------------------------------\n\nTo engineer ~XJ~TGB1 C-terminal His-tagged (~XJ~TGB1-6His) fusion proteins, the ~XJ~TGB1 open reading frame (ORF) was amplified from the plasmid pT7-\u03b2~XJ~ with the primer pair TGB1-NdeIF/TGB1-XhoIR ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)), and integrated into the *Nde*I and *Xho*I sites of the pET-30a(+) expression vector (Novagen) to generate pET-~XJ~TGB1. The TGB1 mutant derivatives of the Thr-395 and Thr-401 residues were engineered using a QuikChange Site-Directed Mutagenesis kit with corresponding primer pairs for the TGB1 ORFs ([Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)).\n\nThe \u03b1-catalytic subunit of CK2 from *N. benthamiana* (NbCK2\u03b1, GenBank accession no. KJ748371) ([@CIT0004]) and barley (HvCK2\u03b1, GenBank accession no. AB252049) ([@CIT0020]) was isolated from host RNA by reverse transcription (RT)-PCR amplification with the primer pairs NbCK2-NdeIF/NbCK2-XhoIR and HvCK2\u03b1-NdeIF/HvCK2\u03b1-SalIR, respectively ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)). Primers for amplifying NbCK2\u03b1 were based on the flanking sequence of the *N. tabacum* CK2\u03b1 ORF (GenBank accession no. AF374474) ([@CIT0048]). The PCR products of the NbCK2\u03b1 and HvCK2\u03b1 genes were inserted into the *Nde*I/*Xho*I, and *Nde*I/*Sal*I sites of pET-30a(+), respectively, to generate pET-NbCK2\u03b1 and pET-HvCK2\u03b1.\n\nFor GST pull-down of ~XJ~TGB1 and ~XJ~TGB3 proteins, the ~XJ~TGB3 ORF was amplified via PCR and ligated in frame to the C terminus of the GST ORF at the *Bam*HI and *Xho*I restriction sites of the plasmid pGEX-KG ([@CIT0012]) to yield the pGEX-~XJ~TGB3 plasmid.\n\nExpression and purification of recombinant proteins from *Escherichia coli* BL21 cells {#s8}\n--------------------------------------------------------------------------------------\n\nThe pET plasmids designed for expression of the C-terminal His-tagged ~XJ~TGB1 proteins, the ~XJ~TGB1 protein mutants, and the HvCK2\u03b1 and NbCK2\u03b1 proteins were transformed into *E. coli* strain BL21(DE3) by standard procedures ([@CIT0049]). Each transformant was grown overnight at 37 \u00b0C in 3ml of LB medium containing kanamycin (100 \u03bcg ml^--1^), transferred into 1 litre of fresh LB/kanamycin medium and grown to a density of 0.4 OD~600~. Recombinant protein expression was induced by the addition of 0.2mM isopropyl \u03b2-[d]{.smallcaps}-1-thiogalactopyranoside and the cells were shaken for 16 additional hours at 18 \u00b0C. Recombinant proteins were extracted from the bacteria and purified by Ni-affinity chromatography according to the manufacturer's instructions (Bio-Rad) and evaluated by 12.5% SDS-PAGE.\n\n*In vitro* phosphorylation assays {#s9}\n---------------------------------\n\nSoluble cytoplasmic protein extracts of healthy *N. benthamiana* leaves were used for *in vitro* kinase assays according to the protocol described by [@CIT0013] and [@CIT0059]. Phosphorylation reactions were performed with the *N. benthamiana* soluble protein extracts or with purified recombinant NbCK2\u03b1 and HvCK2\u03b1 subunits. Assays were performed with 1 \u03bcg of *N. benthamiana* soluble protein extracts, or 0.1 \u03bcg of recombinant CK2\u03b1, and 1 \u03bcg of purified ~XJ~TGB1 protein or its mutants in a final volume of 10 \u03bcl of 25mM Tris/HCl (pH 7.4), 10mM MgCl~2~, and 1 \u03bcl \\[\u03b3-^32^P\\]ATP or GTP (10 \u03bcCi, \\~3000 Ci mmol^--1^; Perkin Elmer). Selected reactions were carried out in the presence or absence of heparin, and various concentrations of unlabelled ATP or GTP. Negative controls contained no TGB1 protein or 500 and 1000ng of bovine serum albumin. After incubation at 30 \u00b0C for 30min, the reactions were terminated by addition of 2.5 \u03bcl of 5\u00d7 SDS buffer, and the samples were subjected to 12.5% SDS-PAGE. The gels were dried with a Model 583 Gel Dryer (Bio-Rad) and phosphorylated proteins were visualized by autoradiography.\n\nMass spectrometry analysis {#s10}\n--------------------------\n\nPhosphorylated ~XJ~TGB1 and unphosphorylated ~XJ~TGB1 proteins were digested with trypsin at 37 \u00b0C overnight. The digested peptides were then analysed by Q-Exactive liquid chromatography tandem mass spectrometry (LC-MS/MS) (Thermo Scientific) at the Mass Spectrometry Facility at China Agricultural University. The data were searched against the NCBI database using Mascot software with a 1% false discovery rate.\n\nFluorescence and confocal microscopy {#s11}\n------------------------------------\n\nGreen fluorescent protein (GFP) or red fluorescent protein (RFP) fluorescence in epidermal cells of *N. benthamiana* was observed with an Olympus confocol FV1000 microscope. GFP and RFP were excited at 488 or 546nm, respectively, with an argon laser. Images were recorded with an Olympus camera and processed using an Olympus Fluoview version 3.0 Viewer. In addition, cell-to-cell movement assays in epidermal cells of barley and *N. benthamiana* were observed with a BX53+DP72 fluorescence microscope (Olympus) and images were manipulated with the cellSens Entry programs.\n\nElectrophoretic mobility shift assays (EMSA) {#s12}\n--------------------------------------------\n\nRNAs for binding assays were transcribed *in vitro* in the presence of digoxigenin (DIG)--UTP (Roche) and the DIG-labelled transcripts were purified to remove the DNA templates ([@CIT0061]). Phosphorylation reactions were performed in a final volume of 5 \u03bcl containing phosphorylation assay buffer and purified NbCK2\u03b1, and different amounts of recombinant ~XJ~TGB1 protein as described above. Negative controls consisted of samples lacking recombinant ~XJ~TGB1 protein or 500 and 1000ng of bovine serum albumin. EMSA binding comparisons were performed by adding increasing amounts of protein to 50ng of purified RNA in binding buffer \\[50mM Tris/HCl (pH 7.5), 10mM MgCl~2~, 1mM EDTA\\] in a final volume of 20 \u03bcl. The binding reaction mixtures were incubated on ice for 30min and subjected to electrophoresis on 1% (w/v) non-denaturing agarose gels in 0.5\u00d7 TBE buffer. RNA--protein complexes were transferred to a Hybond N^+^ nylon membrane (GE Healthcare) via a pump suction filter, and the RNA was cross-linked to the membrane by two 60 s cycles at 0.12 J in a Bio-Link crosslinker (Vilber Lourmat). Mobility shifts of the DIG-labelled RNAs were detected with DIG-- alkaline phosphatasealkaline phosphatase Fab fragments (Roche), and the blots were developed with a 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium chloride substrate solution (Amresco).\n\nGST pull-down {#s13}\n-------------\n\nFor co-expression of GST--~XJ~TGB3 with ~XJ~TGB1--His or the ~XJ~TGB1~T395A/T401A~--His fusion proteins, relevant plasmids were co-transformed into *E. coli* BL21(DE3). Cells were harvested by low-speed centrifugation and disrupted by vortexing in TB buffer \\[20mM Tris/HCl (pH 7.3), 500mM NaCl\\] in the presence of glass beads. The GST fusions and bound TGB proteins were purified by glutathione--Sepharose affinity chromatography and elution with glutathione (Pharmacia).\n\nResults {#s14}\n=======\n\nConstruction and sequence analysis of infectious clones of the BSMV Xinjiang strain {#s15}\n-----------------------------------------------------------------------------------\n\nSeveral BSMV field strains from China collected in our laboratory have broader host ranges than the more extensively studied ~ND~BSMV and Type BSMV (~TY~BSMV) strains. To evaluate the diversity of the more virulent BSMV strains ([@CIT0009]), we constructed infectious clones of the ~XJ~BSMV strain ([@CIT0062]) under the bacteriophage T7 or double cauliflower mosaic virus 35S promoters ([@CIT0045]; [@CIT0064]) ([Fig. 1A](#F1){ref-type=\"fig\"}). The infectivity of *in vitro* transcripts synthesized from linearized pT7-\u03b1~XJ~, pT7-\u03b2~XJ~, and pT7-\u03b3~XJ~ plasmids was tested by mechanical inoculation to barley, wheat, and *B*. *distachyon* Bd21. Inoculated plants consistently developed chlorotic stripes and mosaic symptoms typical of BSMV infections on upper uninoculated leaves by 6--7 dpi ([Fig. 1B](#F1){ref-type=\"fig\"}) and the efficiency of infectivity in barley, wheat, and *B*. *distachyon* Bd21 was 70--80, 80--90, and 50--60%, respectively (also see [Supplementary Table S6](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)). Agroinfiltration was used to initiate infections of *N. benthamiana* because only 10--30% of the plants became infected when using *in vitro* transcripts as inocula. *Agrobacterium* harbouring the plasmids pCa-\u03b1~XJ~, pCa-\u03b2~XJ~, and pCa-\u03b3~XJ~ were infiltrated into the basal sides of the leaves. Newly emerging leaves developed mild mottling symptoms at 7--9 d after agroinfiltration ([Fig. 1C](#F1){ref-type=\"fig\"}), and the efficiency of infectivity was increased to approximate 90%. RT-PCR and Western blot analysis verified the infectivity of the ~XJ~BSMV infectious clones in the monocot hosts ([Fig. 1B](#F1){ref-type=\"fig\"}) and in *N. benthamiana* ([Fig. 1C](#F1){ref-type=\"fig\"}).\n\n![Diagram of ~XJ~BSMV strain infectious clones and cereal and dicot host infectivity test results. (A) Illustration of ~XJ~BSMV infectious clones under the T7 promoter or double cauliflower mosaic virus 35S promoter as described previously ([@CIT0064]; [@CIT0009]). (B) Infectivity assays with *in vitro*-synthesized gRNAs of barley, wheat, and *B*. *distachyon* Bd21. (C) Agroinfiltration was used to initiate infections of *N. benthamiana*. Typical chlorotic stripes and mosaic symptoms (top) of BSMV appeared on emerging uninoculated leaves by 7--9 dpi. Upper uninoculated leaf tissue was harvested at 12 dpi, and the relative BSMV RNA and CP amounts were evaluated by RT-PCR (middle) and Western blotting with the antibody against BSMV CP (bottom). BSMV RNA\u03b3 was detected by RT-PCR with the primer pair BS-10/BS-32 ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)). (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0001){#F1}\n\nFor comparisons of ~XJ~BSMV genomic variation with other published BSMV strains, the pT7-\u03b1~XJ~, pT7-\u03b2~XJ~ and pT7-\u03b3~XJ~ cDNAs were sequenced. The results indicated that ~XJ~RNA\u03b1 (GenBank accession no. KJ746471), ~XJ~RNA\u03b2 (GenBank accession no. KJ746472), and ~XJ~RNA\u03b3 (GenBank accession no. KJ746473) consisted of 3789, 3222, and 2793 nt, respectively, and shared nucleotide identities of 95.2--99.8% ([Supplementary Table S3](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online), 94.1--99.2% ([Supplementary Table S4](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online), and 87.9--98.8% ([Supplementary Table S5](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online) with RNAs \u03b1, \u03b2, and \u03b3 of other BSMV strains. These results suggest that substantial diversity may exist among BSMV strains.\n\nThe ~XJ~TGB1 protein is phosphorylated *in vitro* and *in vivo* {#s16}\n---------------------------------------------------------------\n\nIn order to explore phosphorylation *in vitro*, the full-length ~XJ~TGB1 protein was expressed as a C-terminal His-tagged fusion protein and purified to near homogeneity by Ni-affinity chromatography ([Fig. 2A](#F2){ref-type=\"fig\"}). A soluble protein kinase is known to be present in tobacco species ([@CIT0013]), and hence the purified ~XJ~TGB1 protein was first assayed for phosphorylation using *N. benthamiana* protein extracts as a kinase source. The first control reaction containing \\[\u03b3-^32^P\\]ATP and cytoplasmic extracts without the ~XJ~TGB1 protein resulted in no distinct labelled products ([Fig. 2B](#F2){ref-type=\"fig\"}, lane 1). The corresponding control with the ~XJ~TGB1 protein and \\[\u03b3-^32^P\\]ATP alone suggested that the ~XJ~TGB1 protein was not autophosphorylated *in vitro* ([Fig. 2B](#F2){ref-type=\"fig\"}, lane 2). In contrast, when both the cytoplasmic extracts and the ~XJ~TGB1 protein were present, a radioactive phosphorylated product co-migrated with the ~XJ~TGB1 protein ([Fig. 2B](#F2){ref-type=\"fig\"}, lane 3). These results provide evidence that the ~XJ~TGB1 protein is phosphorylated *in vitro* with \\[\u03b3-^32^P\\]ATP by a soluble kinase in the *N. benthamiana* extracts. The vast majority of protein kinases use ATP as an exclusive phosphate donor, whereas CK2 can effectively use either ATP or GTP ([@CIT0032]); hence we carried out phosphorylation comparisons with GTP to obtain clues about the identity of the kinase involved in ~XJ~TGB1 phosphorylation. Autoradiography of the phosphorylated products revealed a single intense radiolabelled band when either \\[\u03b3-^32^P\\]ATP or \\[\u03b3-^32^P\\]GTP was used as a phosphoryl donor ([Fig. 2B](#F2){ref-type=\"fig\"}, lanes 3 and 4), implying that ~XJ~TGB1 is phosphorylated by a CK2-like kinase in *N. benthamiana.*\n\n![Phosphorylation of the ~XJ~TGB1 protein *in vitro* and *in vivo*. (A) Coomasie Brilliant Blue (CBB) staining of recombinant ~XJ~TGB1 protein purified from *E. coli* cells. Molecular weight markers (Fermentas) are indicated on the left side of the gel. (B) *In vitro* phosphorylation of purified ~XJ~TGB1 protein by cellular kinases present in healthy *N. benthamiana* extracts in the absence or presence of \\[\u03b3-^32^P\\]ATP or \\[\u03b3-^32^P\\]GTP. After the phosphorylation reactions, the TGB1 proteins were separated by 12.5% SDS-PAGE and the incorporated radioactivity was analysed by autoradiography. Reaction mixtures lacking ~XJ~TGB1 protein or *N. benthamiana* protein extracts served as negative controls. The CBB staining in the lower panel indicates that similar amounts of the ~XJ~TGB1 protein were present in each *in vitro* phosphorylation reaction. (C) *In vivo* phosphorylation of ~XJ~TGB1 protein in *N. benthamiana* by Western blotting with \u03b1-TGB1 polyclonal antibodies and \u03b1-threonine antibodies. A mock agroinfiltration lacking ~XJ~RNA\u03b2 was used as a negative control and molecular weight markers (Thermo Scientific) were used to estimate the size of the ~XJ~TGB1 protein. (D) *In vivo* phosphorylation of ~XJ~TGB1 protein immunoprecipitated (IP) from *N. benthamiana* was analysed as in [Fig. 2C](#F2){ref-type=\"fig\"}. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0002){#F2}\n\nTo determine whether the ~XJ~TGB1 protein is phosphorylated *in vivo*, ~XJ~BSMV-infiltrated *N. benthamiana* leaves were harvested, concentrated by immunoprecipitation, and subjected to Western blot analysis with an anti-phosphothreonine (\u03b1-pT) antibody. The \u03b1-pT results revealed a labeleld protein with an electrophoretic mobility corresponding to the 58kDa ~XJ~TGB1 protein ([Fig. 2C](#F2){ref-type=\"fig\"} and [D](#F2){ref-type=\"fig\"}, top, lane 3). Similar-sized bands were not observed in mock-inoculated leaves, or leaves infiltrated with only pCa-\u03b1~XJ~ and pCa-\u03b3~XJ~ ([Fig. 2C](#F2){ref-type=\"fig\"} and [D](#F2){ref-type=\"fig\"}, top, lanes 1 and 2). Western blot analyses using the ~XJ~TGB1 protein antibody confirmed that the \u03b1-pT-reacting protein co-migrated with ~XJ~TGB1, and revealed a 45kDa immunospecific band that we suspect is a degradation product ([Fig. 2C](#F2){ref-type=\"fig\"} and [D](#F2){ref-type=\"fig\"}, bottom, lane 3). Collectively, these experiments provide convincing evidence that the ~XJ~TGB1 protein is phosphorylated by a soluble CK2-like kinase *in vitro* and *in vivo*.\n\nCK2 is able to phosphorylate the ~XJ~TGB1 protein {#s17}\n-------------------------------------------------\n\nTo predict potential ~XJ~TGB1 protein phosphorylation kinases and sites, we analysed the ~XJ~TGB1 protein sequence with the GPS 2.1 program tool ([@CIT0063]) and the Scansite Motif Scanner online server (<http://scansite.mit.edu>) ([@CIT0043]). The GPS 2.1 program was set at a HIGH threshold (false-positive rates of 2% for Ser/Thr kinases) and the Scansite was set at a high stringency level to predict potential phosphorylation targets in the ~XJ~TGB1 protein. According to the conserved CK2 phosphorylation site motif \\[(S/T)-X-X-(D/E); [@CIT0036]\\], the GPS 2.1 predictions suggested that three residues in the ~XJ~TGB1 protein (Ser-69, Thr-395, and Thr-401) are potential CK2 phosphorylation sites, whereas the Scanner only indicated that the Thr-401 site is phosphorylated by CK2 ([Supplementary Fig. S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online). For comparison with ND18, the ~XJ~TGB1 amino acids Ser-69, Thr-395, and Thr-401 correspond to the ND18 TGB1 (~ND~TGB1) Ser-70, Thr-396, and Thr-402 residues, respectively.\n\nBased on the GTP results and phosphorylation predictions, we suspected that a CK2-like kinase might be responsible for phosphorylation of the ~XJ~TGB1 protein. Therefore, the CK2\u03b1 subunits from *N. benthamiana* (NbCK2\u03b1) and barley (HvCK2\u03b1) were cloned, and the C-terminal His-tagged fusion proteins were purified from *E. coli* cells ([Fig. 3A](#F3){ref-type=\"fig\"}). Subsequent *in vitro* phosphorylation assays revealed that the purified NbCK2\u03b1 and HvCK2\u03b1 proteins efficiently phosphorylated the recombinant ~XJ~TGB1 protein in the presence of \\[\u03b3-^32^P\\]ATP ([Fig. 3B](#F3){ref-type=\"fig\"}). CK2 is highly sensitive to heparin inhibition ([@CIT0033]), and our results confirmed that the levels of phosphorylation were reduced proportionally with increasing heparin concentrations ([Fig. 3C](#F3){ref-type=\"fig\"}). To further test the kinase specificity, we used NbCK2\u03b1 to compare ~XJ~TGB1 and TMV P30 MP ([@CIT0014]) protein phosphorylation in the presence of \\[\u03b3-^32^P\\]ATP and \\[\u03b3-^32^P\\]GTP ([Fig. 3D](#F3){ref-type=\"fig\"}). In both cases, the NbCK2\u03b1 phosphorylation assays resulted in the presence of highly intense bands that co-migrated with the ~XJ~TGB1 and P30 proteins, and the kinase exhibited similar activities in the presence of both ATP and GTP ([Fig. 3D](#F3){ref-type=\"fig\"}, lanes 3 and 4, and 6 and 7). Moreover, the absence of radioactive bands in reactions lacking the NbCK2\u03b1 protein confirmed that the 58kDa ~XJ~TGB1 and the P30 proteins are not autophosphorylated and the lack of radioactivity in the reactions without substrate proteins also indicated that the NbCK2\u03b1 protein is not self-phosphorylated ([Fig. 3D](#F3){ref-type=\"fig\"}, lanes 2 and 5). In addition, tests were carried out in the presence of Mn^2+^, Mg^2+^, and Ca^2+^ to assess the cation specificity of CK2 phosphorylation ([@CIT0042]). These results revealed that NbCK2\u03b1 exhibited similar phosphorylation intensities for ~XJ~TGB1 and P30 in the presence of either Mn^2+^ or Mg^2+^ ([Fig. 3E](#F3){ref-type=\"fig\"}, lanes 2 and 3, and 7 and 8), and that ^32^P incorporation was negligible in reactions containing Ca^2+^ ([Fig. 3E](#F3){ref-type=\"fig\"}, lanes 4 and 9). All of the *in vitro* phosphorylation data with the TGB1 protein were consistent with several published biochemical properties of CK2 ([@CIT0014]; [@CIT0013]).\n\n![*In vitro* phosphorylation of ~XJ~TGB1 protein by recombinant CK2 kinase. (A) SDS-PAGE analysis of NbCK2\u03b1 and HvCK2\u03b1 purified from *E. coli* BL21 cells. (B) *In vitro* phosphorylation of ~XJ~TGB1 protein with the NbCK2\u03b1 and HvCK2\u03b1 recombinant proteins and negative controls lacking the kinases. (C) Effects of heparin on *in vitro* phosphorylation of ~XJ~TGB1 protein. Phosphorylation levels were reduced with increasing amount of heparin. (D) Ability of NbCK2\u03b1 to use both ATP and GTP as phosphate donors. (E) Divalent metal ion specificity of NbCK2\u03b1 and the TMV-MP (P30) proteins. The CBB-stained proteins at the bottom of panels (B)--(E) are as indicated as in [Fig. 2B](#F2){ref-type=\"fig\"}. (F) Co-localization of the GFP:~XJ~TGB1 and DsRed:NbCK2\u03b1 proteins in *N. benthamiana* leaf cells. Single localization of GFP:~XJ~TGB1 and DsRed:NbCK2\u03b1 proteins are indicated at the top of the panels. Bars, 50 \u03bcm. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0003){#F3}\n\nFurthermore, to explore the potential co-localization of the ~XJ~TGB1 protein and NbCK2\u03b1 *in planta*, we conducted co-expression experiments with GFP:~XJ~TGB1 and DsRed:NbCK2\u03b1 fusion proteins in *N. benthamiana* cells. Confocal microscopy revealed that GFP:~XJ~TGB1 was distributed in both the cytoplasm and nucleus, whereas DsRed:NbCK2\u03b1 was present primarily in the nucleus. The co-localization of the two proteins indicated that the ~XJ~TGB1 protein and NbCK2\u03b1 interact in both the nucleus and cytoplasm in some manner ([Fig. 3F](#F3){ref-type=\"fig\"}, overlay channel). Taken together, the presented data demonstrate that XJTGB1 is phosphorylated by CK2 *in vitro* and *in planta*.\n\nThr-401 is the major ~XJ~TGB1 protein site for CK2 phosphorylation {#s18}\n------------------------------------------------------------------\n\nTo identify the phosphorylation sites of CK2, the purified ~XJ~TGB1 protein was phosphorylated by NbCK2\u03b1 *in vitro* with unlabelled ATP, and the gel-purified phosphorylated and unphosphorylated ~XJ~TGB1 proteins were separated by PAGE and digested with trypsin. The trypsin digestion products were analysed by Q-Exactive LC-MS/MS. The analysis showed that 87.9% of the TGB1 protein amino acid sequence was covered, and revealed that the phosphorylated and unphosphorylated proteins differed in a unique monophosphorylated peptide (^399^GE**[T]{.ul}**DETEKNIAFTVDTVR^416^) with a 2103.9362 *m*/*z* peak corresponding to a neutral precursor ion lacking phosphoric acid (97.9769Da). Based on the observed masses of the phosphorylated and unphosphorylated y~16~ fragment ions ([Fig. 4A](#F4){ref-type=\"fig\"}), we conclude that the phosphorylated ~XJ~TGB1 residue is located at Thr-401.\n\n![Thr-401 is the major ~XJ~TGB1 protein site for CK2 phosphorylation. (A) LC-MS/MS analysis of ~XJ~TGB1 protein phosphorylation by NbCK2\u03b1. The absence of phosphoric acid (97.9769Da) on the y~16~ ion fragment demonstrates that Thr-401 is a phosphorylation site for CK2 kinase. (B) Identification of the phosphorylation sites in ~XJ~TGB1 protein mutants by *in vitro* phosphorylation with HvCK2\u03b1 and NbCK2\u03b1. The radioactive intensities of the ~XJ~TGB1 protein and its phosphorylation mutants indicate the extent of radiolabelling with \\[\u03b3-^32^P\\]ATP. CBB-stained proteins at the bottom of the panels (B) and (C) are as indicated in [Fig. 2B](#F2){ref-type=\"fig\"}. (C) Phosphorylation comparisons of selected ~XJ~TGB1 protein mutants with wt ~XJ~TGB1 protein to confirm that Thr-401 is the major phosphorylated residue. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0004){#F4}\n\nTo verify the GPS 2.1 and Scanner predictions ([Supplementary Fig. S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)) and LC-MS/MS analysis of ~XJ~TGB1 protein, we replaced one or both of the Thr-395 and Thr-401 residues with alanine residues to produce ~XJ~TGB1~T395A~, ~XJ~TGB1~T401A~, and ~XJ~TGB1~T395A/T401A~ phosphorylation-deficient mutants. To mimic the phosphorylation state of the ~XJ~TGB1 protein, Thr-395 and Thr-401 residues were substituted with aspartic acid (D) or glutamic acid (E) residues to produce the ~XJ~TGB1~T395D~, ~XJ~TGB1~T395E~, ~XJ~TGB1~T401D~, and ~XJ~TGB1~T401E~ mutants. *In vitro* phosphorylation comparisons of the wild-type (wt) ~XJ~TGB1 protein, and the Thr-395 and Thr-401 mutants were performed with HvCK2\u03b1 and NbCK2\u03b1, respectively. Compared with the wt ~XJ~TGB1 protein ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 1), the phosphorylation level of the ~XJ~TGB1~T395A~ mutant protein was reduced partially ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 2). However, both the ~XJ~TGB1~T395D~ and the ~XJ~TGB1~T395E~ mutant proteins incorporated slightly larger amounts of ^32^P label than the wt ~XJ~TGB1 protein ([Fig. 4B](#F4){ref-type=\"fig\"}, compare lane 1 with lanes 3 and 4), suggesting that the positive charges imparted by the aspartic acid and glutamic acid residues increased the kinase efficiency. As anticipated, the phosphorylation intensities of the ~XJ~TGB1~T401A~ ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 5) and ~XJ~TGB1~T401E~ ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 7) proteins only showed faint shadows. We also observed similar reductions during incorporation into the ~XJ~TGB1~T401D~ protein ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 6). We believe that Thr-395 may have been phosphorylated, and this is supported by negligible incorporation into the three double mutants ([Fig 4B](#F4){ref-type=\"fig\"}, lane 8; [Fig. 4C](#F4){ref-type=\"fig\"}, lanes 1--3). Thus, our interpretation of these results is that Thr-401 is a major phosphorylation site and that Thr-395 is a minor phosphorylation target.\n\nBased on the data above and previous analyses of 308 CK2 phosphorylation sites for other proteins ([@CIT0036]), we propose that Thr-395 is a docking site for CK2 because phosphorylation of this residue enhanced kinase activity at Thr-401. To evaluate this hypothesis, we used a double mutant, ~XJ~TGB1~T395A/T401A~, for phosphorylation. The results revealed extremely low, if any, ^32^P incorporation into other TGB1 residues ([Fig. 4B](#F4){ref-type=\"fig\"}, lane 8; [Fig 4C](#F4){ref-type=\"fig\"}, lane 3). To determine whether there was 'off-site' targeting of other residues within ~XJ~TGB1 at the proposed Thr-395 docking site, we engineered the ~XJ~TGB1~T395D/T401A~ and ~XJ~TGB1~T395E/T401A~ double mutants, which could not be phosphorylated at Thr-401, and both double mutants had negligible levels of ^32^P incorporation ([Fig. 4C](#F4){ref-type=\"fig\"}, lanes 1 and 2). In summary, these results support the hypothesis that Thr-395 functions as a docking residue and that Thr-401 is the major phosphorylation site within the kinase motif of the ~XJ~TGB1 protein.\n\nMutations of the ~XJ~TGB1 protein phosphorylation site affect BSMV local and systemic infections of dicots and monocots {#s19}\n-----------------------------------------------------------------------------------------------------------------------\n\nTo determine whether the ~XJ~TGB1 mutants affected systemic movement in *N. benthamiana*, leaves were infiltrated with *Agrobacterium* harbouring the pCa-\u03b1~XJ~, wt pCa-\u03b2~XJ~, or individual TGB1 mutant derivatives, and the pCa-\u03b3~XJ~ clones. Three independent experiments revealed that only wt ~XJ~TGB1 and the \u03b2~XJ~-TGB1~T395A~ mutant, which exhibited slightly lower phosphorylation levels than the wt ~XJ~TGB1 protein, were able to establish systemic infections at 10 dpi ([Fig. 5A](#F5){ref-type=\"fig\"}, lanes 2 and 3). None of the remaining infiltrations containing single or double ~XJ~TGB1 mutants developed mosaic symptoms or invaded the upper uninoculated leaves as assessed by the absence of CP accumulation ([Fig. 5A](#F5){ref-type=\"fig\"}, lanes 4--11).\n\n![Mutants affecting phosphorylation of the ~XJ~TGB1 protein have host-specific effects on systemic infectivity. (A) Symptoms of *N. benthamiana* elicited after infiltration with an *Agrobacterium* mixture harbouring pCa-\u03b1~XJ~, pCa-\u03b3~XJ~, and pCa-\u03b2~XJ~ or its phosphorylation site mutants. Upper uninfiltrated leaf tissues were harvested and photographed at 10 dpi (top). CP ELISA (middle) and RNA\u03b3 RT-PCR amplification (bottom) were monitored to estimate the infectivity levels. (B, C) Systemic symptoms appearing in barley (B) and wheat (C) after inoculation with pT7-\u03b1~XJ~ and pT7-\u03b3~XJ~ *in vitro* transcripts mixed with pT7-\u03b2~XJ~ and various phosphorylation mutant transcripts. Leaves were photographed at 14 dpi (top) and all experiments were repeated three times. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0005){#F5}\n\nFor systemic infectivity on monocots, barley and wheat leaves were co-inoculated with *in vitro* transcripts of pT7-\u03b1~XJ~, pT7-\u03b2~XJ~ (wt or mutant derivatives), and pT7-\u03b3~XJ~. Visual observations, ELISA and RT-PCR analyses demonstrated that the barley and wheat plants had similar systemic infections on upper uninoculated leaves at 14 dpi after wt and mutant ~XJ~TGB1 inoculations ([Fig. 5B](#F5){ref-type=\"fig\"}, [C](#F5){ref-type=\"fig\"}). In the case of \u03b2~XJ~-TGB1~T395A~, a milder infection phenotype was noted on the systemically infected cereal leaves, compared with those of the wt \u03b2~XJ~ infections ([Fig. 5B](#F5){ref-type=\"fig\"}, [C](#F5){ref-type=\"fig\"}, lanes 2 and 3), but the aspartic acid mutant \u03b2~XJ~-TGB1~T395D~ was unable to infect the plants systemically ([Fig. 5B](#F5){ref-type=\"fig\"}, [C](#F5){ref-type=\"fig\"}, lane 4). In contrast, when inoculated with the \u03b2~XJ~-TGB1~T395E~ mutant, which more effectively mimics phosphorylation of threonine residues, the cereal plants exhibited CP accumulation levels that were similar to wt BSMV ([Fig. 5B](#F5){ref-type=\"fig\"}, [C](#F5){ref-type=\"fig\"}, lane 5). Moreover, the \u03b2~XJ~-TGB1~T401E~ mutant elicited systemic infections, albeit with slightly lower CP levels ([Fig. 5B](#F5){ref-type=\"fig\"}, [C](#F5){ref-type=\"fig\"}, lane 8), but the levels were higher than those with \u03b2~XJ~-TGB1~T401A~ ([Supplementary Table S6](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online). However, the \u03b2~XJ~-TGB1~T395A/T401A~, \u03b2~XJ~-TGB1~T395D/T401A~, and \u03b2~XJ~-TGB1~T395E/T401A~ double mutants were unable to infect either barley or wheat ([Fig. 5B](#F5){ref-type=\"fig\"}, lanes 9--11).\n\nTaken together, these results demonstrated that the ~XJ~TGB1 phosphorylation site mutations generally reduced the infection efficiencies in dicots and monocots, and that the mutants had more dramatically compromised systemic movement phenotypes in *N. benthamiana*. To summarize, the T395A, T395E, T401A, and T401E mutants exhibited systemic movement in barley and wheat, whereas only the T395A mutant is able to establish systemic infections in *N. benthamiana*.\n\nWe next conducted experiments to evaluate the cell-to-cell movement profiles of the Thr-395 and Thr-401 mutants in *N. benthamiana* and barley. For this purpose, *N. benthamiana* leaves were infiltrated with *Agrobacterium* containing pCa-\u03b1~XJ~, pCa-\u03b2~XJ~ and its mutant derivatives, and a pCa-\u03b3~XJ:GFP~ construct that harbours a \u03b3b:GFP reporter gene to assess cell-to-cell movement. In barley, leaves were co-inoculated with T7 transcripts of the RNA\u03b1, RNA\u03b2 derivatives, and RNA\u03b3-\u03b3~XJ:GFP~ transcripts. Epidermal cells of the leaves were observed by confocal microscopy at 3 dpi and compared with inoculated controls lacking the RNA\u03b2.\n\nThe localized movement in infiltrated *N. benthamiana* leaves generally reflected the systemic infection phenotypes elicited by the mutants ([Fig. 6A](#F6){ref-type=\"fig\"}). In *N. benthamiana*, the wt \u03b2~XJ~ and \u03b2~XJ~-TGB1~T395A~ mutant both exhibited cell-to-cell movement encompassing several cells at 3 dpi ([Fig. 6A](#F6){ref-type=\"fig\"}), as expected due to their ability to elicit systemic infections. The remaining mutants usually developed fluorescence in a single cell or rarely in two to three adjacent cells ([Fig. 6A](#F6){ref-type=\"fig\"}). Hence, the localized movements of the mutants correlated reasonably well with their systemic movement patterns in *N. benthamiana*. In barley leaves, most of the fluorescence at 3 dpi appeared in mesophyll cells, but in this case, the virus had to traverse only two to three cell layers before encountering the closely aligned parallel vasculature ([Fig. 6B](#F6){ref-type=\"fig\"}). Hence, the \u03b2~XJ~-TGB1~T395A~, \u03b2~XJ~-TGB1~T395E~, and \u03b2~XJ~-TGB1~T401A~ mutants that established systemic infections in barley and wheat needed to negotiate only a limited number of mesophyll cells to reach the vascular elements for systemic spread, whereas movement through a larger number of cells was required to reach the dicot vasculature. The other mutants, \u03b2~XJ~-TGB1~T395D~, \u03b2~XJ~-TGB1~T401D~, and \u03b2~XJ~-TGB1~T395A/T401A~, exhibited more limited cell-to-cell movement compared with \u03b2~XJ~-TGB1~T395A~, \u03b2~XJ~-TGB1~T395E~, and \u03b2~XJ~-TGB1~T401A~, but could spread to a few adjacent cells. However, these three mutants were unable to invade the upper cereal leaves. Hence, these results demonstrate that phosphorylation activities at Thr-395 and Thr-401 differentially affecte systemic movement in monocot versus dicot hosts and suggest that at least some of the host-specific results may be a consequence of the vasculature architecture of these hosts.\n\n![Effects of ~XJ~TGB1 protein phosphorylation on ~XJ~BSMV cell-to-cell movement in *N. benthamiana* and barley. (A) Fluorescence in *N. benthamiana* leaves at 3 dpi with an *Agrobacterium* mixture of pCa-\u03b1~ND~, pCa-\u03b3~ND:GFP~, and pCa-\u03b2~XJ~ or the pCa-\u03b2~XJ~ mutant derivatives. The total bacterial concentrations for infiltration were OD~600~ of 0.08. (B) Fluorescence in barley leaves at 3 dpi with *in vitro* transcripts of RNA\u03b1 and RNA\u03b3:GFP plus wt ~XJ~RNA\u03b2 or the ~XJ~RNA\u03b2 phosphorylation site mutant derivatives. Bars represent 500 \u03bcm. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0006){#F6}\n\nPhosphorylation promotes virus infection of ~XJ~BSMV by enhancing TGB1 and TGB3 protein interactions {#s20}\n----------------------------------------------------------------------------------------------------\n\nPrevious studies have shown that BSMV spreads from cell to cell through the co-ordinated actions of TGB proteins, which co-localize at the cell wall in close association with PD, during cell-to-cell movement in monocots and dicots ([@CIT0006]). Our results shown above demonstrated that interference with CK2 phosphorylation at the ~XJ~TGB1 Thr-395 and Thr-401 sites affected ~XJ~BSMV local and systemic movement. TGB1 is a multifunctional protein that engages in homologous interactions and formation of a ribonucleoprotein complex containing viral genomic and subgenomic RNAs ([@CIT0024]). Therefore, we used three approaches to identify ~XJ~TGB1 protein functions affected by CK2 phosphorylation.\n\nTo determine whether the RNA-binding affinity of the ~XJ~TGB1 protein changed upon phosphorylation, we first used purified wt ~XJ~TGB1 and the double mutant ~XJ~TGB1~T395A/T401A~ proteins in EMSA with DIG-labelled RNA transcripts ([Fig. 7A](#F7){ref-type=\"fig\"}). The results clearly showed that both proteins bound almost all of the available RNA at 250ng, indicating that the mutant protein did not affect RNA-binding activities ([Fig. 7A](#F7){ref-type=\"fig\"}, panels 1 and 2). Next, to determine phosphorylation effects directly, the ~XJ~TGB1 protein was incubated in phosphorylation assay buffer containing unlabelled ATP, with and without the addition of purified NbCK2\u03b1, and EMSA assays were performed to compare the abilities of the non-phosphorylated and phosphorylated ~XJ~TGB1 proteins to bind the labelled RNA transcripts ([Fig. 7A](#F7){ref-type=\"fig\"}, panels 3 and 4). In addition, the RNA-binding activities of the phosphorylated wt ~XJ~TGB1 and mutant ~XJ~TGB1~T395A/T401A~ proteins were compared, but the ~XJ~TGB1~T395A/T401A~ protein was found to retain almost the same level of RNA binding as the wt ~XJ~TGB1 protein ([Fig. 7A](#F7){ref-type=\"fig\"}, panels 5 and 6). Hence, phosphorylation appeared to have little, if any, effect on the RNA-binding activities of the ~XJ~TGB1 protein.\n\n![Effect of ~XJ~TGB1 protein phosphorylation mutants on its functions. (A) Comparison of RNA binding by the phosphorylated native ~XJ~TGB1 and double-mutant ~XJ~TGB1~T395A/T401A~ proteins. (B) GST affinity chromatography comparisons of the ~XJ~TGB1 and double-mutant ~XJ~TGB1~T395A/T401A~ proteins with the GST:~XJ~TGB3 protein. The concentrations of the TGB proteins were similar in the experiments, but the ~XJ~TGB1~T395A/T401A~ protein had approximate 40% TGB3 protein-binding efficiency compared with the wt ~XJ~TGB1 protein. The illustrated binding result is typical of three independent experiments. (C) Co-localization of TGB proteins. Confocal laser-scanning microscopy observation of *N. benthamiana* leaf epidermal cells co-infiltrated with mixtures of *Agrobacterium* harbouring GFP:~XJ~TGB1 or the GFP:~XJ~TGB1~T395A/T401A~ mutant derivatives and the pGD-TGB2 and RFP:TGB3 plasmids. Bars, 50 \u03bcm. (This figure is available in colour at *JXB* online.)](exbotj_erv237_f0007){#F7}\n\nTo evaluate the possible role of phosphorylation in heterologous interactions of ~XJ~TGB1 and ~XJ~TGB3 proteins, experiments were carried out with His-tagged ~XJ~TGB1 and its mutant ~XJ~TGB1~T395A/T401A~ fusion proteins in co-expressions with the GST:~XJ~TGB3 protein in *E. coli* BL21 cells. Both ~XJ~TGB1 and ~XJ~TGB1~T395A/T401A~ proteins were retained to some extent on the affinity columns by the GST:~XJ~TGB3 protein ([Fig. 7B](#F7){ref-type=\"fig\"}), but our three experiments consistently showed that the ~XJ~TGB1~T395A/T401A~ protein bound the TGB3 protein approximate 40% less effectively than the ~XJ~TGB1 protein. These results thus suggested that the mutant ~XJ~TGB1~T395A/T401A~ may weaken TGB1:TGB3 protein interactions and result in impaired cell-to-cell movement functions of ~XJ~BSMV.\n\nIn additional attempts to ascertain whether the compromised TGB1:TGB3 protein interactions or CK2 phosphorylation affected localization of TGB proteins in plant cells, co-localization assays were performed by transient expression of the three TGB proteins via agroinfiltration in *N. benthamiana*, and GFP and RFP localizations were evaluated at 2 dpi by confocal laser-scanning microscopy. The results revealed that TGB2 and TGB3, the ~XJ~TGB1~T395A/T401A~ protein, and the wt ~XJ~TGB1 protein had similar TGB localization patterns ([Fig. 7C](#F7){ref-type=\"fig\"}). Taken together, we conclude that phosphorylation promotes ~XJ~BSMV infection by enhancing the interactions between the ~XJ~TGB1 and ~XJ~TGB3 proteins, although this effect was not sufficient to substantially alter the TGB localization patterns visible by confocal microscopy.\n\nDiscussion {#s21}\n==========\n\nReversible phosphorylation and dephosphorylation of proteins have regulatory roles in a wide range of cellular processes, including cell signalling transduction ([@CIT0038]), protein subcellular localization ([@CIT0041]), and protein--protein ([@CIT0056]) and protein--nucleic acid interactions ([@CIT0052]). Numerous proteins with distinct phosphorylation sites have been investigated, and protein kinases affecting a wide range of cellular responses have been characterized ([@CIT0044]; [@CIT0057]; [@CIT0005]; [@CIT0010]). A growing body of evidence now shows that viral proteins with different functions are phosphorylated by various protein kinases during infection. These include CK2, PKA, PKC, and CK1 protein kinases ([@CIT0025]; [@CIT0030]; [@CIT0031]; [@CIT0013]), and among these kinases, CK2 phosphorylation effects on infectivity have been most extensively studied.\n\nProtein kinase CK2, a conserved Ser/Thr kinase existing in almost all eukaryotes, phosphorylates proteins with a consensus phosphorylation site motif (S/T-D/E-X-E/D, where X is any residue) ([@CIT0057]). Increasing evidence indicates that CK2 protein phosphorylation has important roles in plant growth and development ([@CIT0040]), and that CK2 also regulates virus infection processes, including virion assembly, cell-to-cell and long-distance movement, and interactions between viral proteins and other host proteins ([@CIT0017]; [@CIT0041]). In addition to the MP phosphorylation effects mentioned in the Introduction, phosphorylation of cucumber mosaic virus, cucumber necrosis virus, and turnip yellow mosaic virus RdRp has substantial effects on virus replication ([@CIT0022]; [@CIT0053]; [@CIT0053]; [@CIT0018]; [@CIT0017]). Ser/Thr phosphorylation has also been suggested to affect the CP functions of several plant viruses. For example, phosphorylation of potato virus A (PVA) CP by host CK2 inhibits viral RNA binding *in vitro*, and mutation of a major phosphorylation CP site generates a PVA variant that is defective in cell-to-cell and long-distance movement ([@CIT0015], [@CIT0014]). In addition, phosphorylation of the cauliflower mosaic virus CP precursor at several sites by CK2 is important for virus infectivity and symptom development ([@CIT0007]). Phosphorylation of the bamboo mosaic virus CP by CK2 also regulates cell-to-cell movement by modulating RNA binding ([@CIT0013]).\n\nAlthough phosphorylation of the PSLV N-terminal portion of the TGB1 protein has been reported ([@CIT0031]), but the results were not extended to evaluate the roles of phosphorylation in PSLV movement processes. Our results now demonstrate that the BSMV TGB1 protein is phosphorylated by CK2 *in vitro* and *in planta*, and that the phosphorylation events affect virus movement. Although the prediction programs we used suggest that the ~XJ~TGB1Thr-395 residue in the ^395^ [T]{.ul}DYD^398^ site is more likely to be a conserved CK2 phosphorylation site than the Thr-401 (^401^ [T]{.ul}DET^404^) site ([@CIT0036]), it is noteworthy that Thr-401 localizes within an acidic residue-rich region (^399^GE[T]{.ul}DETEK^**406**^) that may be more favourable for phosphorylation ([@CIT0003]; [@CIT0046]) than the Thr-395 residue. Thus, based on the *in vitro* phosphorylation assays of the TGB1 mutant derivatives, as well as results derived from LC-MS/MS analysis, we conclude that Thr-401 is the major TGB1 phosphorylation site and that Thr-395 functions as a CK2 docking site and has a more limited phosphorylation role. To the best of our knowledge, this is the first report showing that a plant viral protein, which can be phosphorylated by CK2, has a CK2 docking site adjacent to the phosphorylation site.\n\nTo determine the effects of ~XJ~TGB1 Thr-395 and Thr-401 phosphorylation on virulence, point mutations were introduced into the ~XJ~BSMV clone. Infectivity results in the monocot and dicot hosts revealed that the mutant derivatives differed in their systemic movement phenotypes. For example, the T395A mutant was the only derivative able to infect *N. benthamiana*, barley, and wheat systemically, but the T395E, T401A, and T401E mutants also systemically infected the monocot hosts ([Fig. 5](#F5){ref-type=\"fig\"}). Our results suggest that the mutants may be compromised by partial disruption of phosphorylation and dephosphorylation dynamics in TGB1 in ways that contribute to diminished cell-to-cell movement. However, the amino acid structures of the T395A, T395D, T395E, T401D, and T401E substitutions are not entirely consistent with this simplistic model, as it is obvious that the substituted amino acids differ in the sizes of their side chains and their charges. Moreover, we cannot exclude the possibility that 'off-site' phosphorylation of Thr-395 or phosphorylation by kinases other than CK2 may be elicited by the substitutions and that these events may contribute to protein modifications that affect local and vascular movement.\n\nHordeivirus TGB1 proteins are multifunctional and contain two positively charged regions rich in lysine (K) and arginine (R) residues at the N-terminal half of the protein and a C-terminal region consisting of seven conserved motifs (I, IA, II, III, IV, V, and VI) ([@CIT0016]). The Thr-395 and Thr-401 sites are located between domain IV and V of the TGB1 protein and are not included in the most highly conserved regions. The hordeivirus TGB1 proteins have multiple ssRNA- and dsRNA-binding sites, and hence mutagenesis of single or closely associated ~XJ~TGB1 protein sites may not have obvious effects on RNA-binding activities *in vitro*.\n\nWe have shown previously that the BSMV TGB1 protein is the major protein component of ribonucleoprotein complexes involved in BSMV cell-to-cell movement. TGB1 also participates in interactions of TGB1 and TGB3 proteins during intra- and intercellular virus movement, and functions in TGB1:TGB3 interactions that recruit the TGB2 protein during transport through PD to adjacent cells ([@CIT0016]). These interactions are critical for movement because TGB3 serves as a bridge to direct TGB co-localization at the cell wall and to establish close associations with the PD ([@CIT0016]). Even though both CK2 phosphorylation sites (T395A/T401A) in ~XJ~TGB1 were mutated simultaneously, the mutant TGB1 protein did not elicit obvious changes in subcellular localization patterns ([Fig. 7C](#F7){ref-type=\"fig\"}). However, the ~XJ~TGB1~T395A/T401A~ protein did reduce binding affinity to the TGB3 protein compared with the wt TGB1 protein. Therefore, our results provide evidence that CK2 phosphorylation of TGB1 affects BSMV movement by modulating TGB1:TGB3 protein interactions.\n\nIn summary, our results shown here provide evidence showing that phosphorylation of the BSMV ~XJ~TGB1 protein by CK2 at C-terminal residues affects cell-to-cell movement and the systemic infection phenotype. Moreover, the mutant results are compatible with a model whereby modulation of TGB1:TGB3 interactions contribute to phenotypic differences with in BSMV movement in monocot and dicot hosts. Compared with the TGB1 proteins of other hordeiviruses ([Supplementary Fig. S2A](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1), available at *JXB* online), Thr-401 but not Thr-395 is conserved in PSLV TGB1 and BSMV TGB1. This implies that the proposed docking site function of Thr-395 may be unique for BSMV phosphorylation. Furthermore, multiple sequence alignments of TGB1 proteins ([Supplementary Fig. S2B](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1)) showed that the Thr-395 and Thr-401 sites are highly conserved among six sequenced BSMV strains, suggesting that phosphorylation of the TGB1 protein is required during infection of all BSMV strains.\n\nSupplementary data {#s22}\n==================\n\nSupplementary data are available at *JXB* online.\n\n[Supplementary Fig. S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Phosphorylation predictions of the ~XJ~TGB1 protein by the GPS 2.1 program (A) and the Scansite Motif Scanner online server (B).\n\n[Supplementary Fig. S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Alignment of the TGB1 proteins of the hordeiviruses (A) and among six sequenced BSMV strains (B).\n\n[Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Primers used in construction and analysis of biologically active BSMV Xinjiang cDNA clones.\n\n[Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Primers used for site-specific mutagenesis of Xinjiang RNA\u03b2 clones.\n\n[Supplementary Table S3](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Sequence alignment of Xinjiang strain RNA\u03b1 with different BSMV strains.\n\n[Supplementary Table S4](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Sequence alignment of Xinjiang strain RNA\u03b2 with different BSMV strains.\n\n[Supplementary Table S5](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Sequence alignment of Xinjiang strain RNA\u03b3 with different BSMV strains.\n\n[Supplementary Table S6](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv237/-/DC1). Systemic infectivity efficiency of ~XJ~BSMV TGB1 phosphorylation mutants on dicot and monocot hosts.\n\n###### Supplementary Data\n\nWe thank Professor Yau-Heiu Hsu (National Chung Hsing University) for kindly providing the details of the protocol for *in vitro* phosphorylation assays, Professor A. L.N. Rao (University of California, Riverside) for providing the pCass4-Rz vector, Professors Dongtao Ren, Qun He, and Huiqiang Lou (China Agricultural University) for helpful suggestions and constructive criticism, and Dr Zhen Li and Jingqiang Zhang (The Mass Spectrometry Facility, CAU) for technical assistance in LC-MS/MS. This work was supported by the National Natural Science Foundation of China (31270184) and the Project for Extramural Scientists of SKLAB (2012SKLAB06-02).\n\nBSMV\n\n:   barley stripe mosaic virus\n\nCP\n\n:   coat protein\n\nDIG\n\n:   digoxygenin\n\ndpi\n\n:   d post-inoculation\n\nEMSA\n\n:   electrophoretic mobility shift assays\n\nGFP\n\n:   green fluorescent protein\n\nLC-MS/MS\n\n:   liquid chromatography tandem mass spectrometry\n\nMP\n\n:   movement protein\n\nORF\n\n:   open reading frame\n\nPD\n\n:   plasmodesmata\n\nPSLV\n\n:   Poa semilatent virus\n\nRdRp\n\n:   RNA-dependent RNA polymerase\n\nRFP\n\n:   red fluorescent protein\n\nRT-PCR\n\n:   reverse transcription-PCR\n\nTGB\n\n:   triple gene block\n\nTMV\n\n:   tobacco mosaic virus.\n"}
+{"text": "\n"}
+{"text": "Background:\n===========\n\nThe single-leg hop for distance (SLHOP) is a commonly used functional performance measure to determine return-to-play readiness. However, it is unknown if functional performance at return-to-play can predict future patient outcomes. Therefore, the purpose of this study was to determine if SLHOP performance and symmetry at return-to-play following a lower extremity injury would predict lower extremity patient-reported outcomes six months post-injury.\n\nMethods:\n========\n\nTwo-hundred thirty-six adolescent athletes (15.7\u00b11.4 years, 171.1\u00b17.6 cm, 70.3\u00b115.3 kg) were recruited for this study. If an individual sustained a non-surgical lower extremity injury during their sports season and missed three days of sports participation they were eligible for follow-up functional and patient outcome assessment Thirty-two adolescent athletes were eligible and completed the follow-up testing (15.0\u00b11.1 years, 166.5\u00b14.9 cm, 67.3\u00b110.4 kg; 14 American football, 8 volleyball, 4 girls' basketball, 6 boys' basketball athletes). The SLHOP was completed by affixing a tape measure to the ground and instructing the participant to hop on one leg as far as possible. Each participant performed three trials per leg at return-to-play (alternating legs). To be counted as a successful trial the participant maintained postural stability for 2 seconds after the landing and the contralateral leg was not allowed to touch the ground. The Pediatric Patient-Reported Outcomes Measurement Information System (PROMIS) Lower Extremity scale was completed six months after return-to-play as the patient-reported outcome measure. A stepwise multiple linear regression was conducted to predict PROMIS Lower Extremity scale score from SLHOP distance on the injured and uninjured leg (cm), and limb symmetry (%).\n\nResults:\n========\n\nAt step one of the analysis, injured limb SLHOP distance was entered into the regression equation and was significantly related to the PROMIS Lower Extremity scale F(1,31)=20.8, p\\<.001. The multiple correlation coefficient was 0.79, indicating approximately 62% of the variance of the PROMIS Lower Extremity scale scores could be accounted for by the injured limb SLHOP distance. Specifically, for every increase in one cm of SLHOP performance there would be an increase of 0.6 points on the PROMIS Lower Extremity scale. Uninjured limb SLHOP distance and limb symmetry did not enter into the equation at step two of the analysis (p\\>.05).\n\nConclusion:\n===========\n\nThe SLHOP conducted at time of return-to-play following a lower extremity injury can predict lower extremity patient-reported outcomes six months later. This indicates that clinicians may be able to use the SLHOP following an injury to not only determine return-to-play readiness but a successful long-term outcome.\n"}
+{"text": "Introduction {#S1}\n============\n\nA significant portion of the chromosomal material is compartmentalized into 'topologically associated domains' (TADs), typically encompassing several hundred kilobases of linear genome folded upon itself with regulatory proteins including cohesion and the multifunctional CCCTC-binding factor CTCF constraining expression of TAD-associated genes^[@R1]--[@R4]^. TAD-like structures in brain^[@R5],\\ [@R6]^ were implicated in the genetic risk architecture of psychiatric disease^[@R7]^ but regulatory mechanisms remain unexplored. Here, we report that neuronal maintenance of a subset of very large 'superTADs' critically requires *Set-domain-bifurcated 1* (*Setdb1*/*Eset/Kmt1e*) histone H3-lysine 9 methyltransferase^[@R8]^. *Kmt1e/Setdb1* is important for prenatal development and pup survival^[@R9]^ and broadly regulates retroelement suppression and transcriptional silencing in stem cells^[@R10]--[@R12]^. Little is known about its essential functions in differentiated cells including neurons. Cell type-specific 3D genome, CTCF, DNA methylation and histone modification profilings, in conjunction with targeted epigenomic editing and conditional mutagenesis uncovered a Setdb1-dependent 'shield' protecting genomes from excess CTCF binding and unique locus-specific epigenomic vulnerabilities triggering higher order chromatin collapse on a megabase-scale.\n\nResults {#S2}\n=======\n\nLocus-specific TAD disintegration in Setdb1 mutant neurons {#S3}\n----------------------------------------------------------\n\nTo explore higher order chromatin in *Setdb1*-deficient brain, we generated a mouse line for CK-Cre driven recombination in postnatal forebrain neurons with exon 3 deletion, frameshift and premature stop upstream of the critical Tudor, methyl-CpG binding (MBD) and catalytic SET domains. *CK-Cre^+^ Setdb1^2lox/2lox^* mutant, in comparison to *CK-Cre^\u2212^Setdb1^2lox/2lox^* control mice showed grossly normal brain cytoarchitecture. However, adult mutants consistently showed a reduction in brain weight, without premature death or neuronal loss as assessed by flow cytometry-based nuclei counts and COMET nuclear DNA damage assays ([Figure 1a](#F1){ref-type=\"fig\"}, [Supplementary Figure 1](#SD1){ref-type=\"supplementary-material\"}). We conducted in situ Hi-C^[@R1]^, or genome-scale DNA-DNA proximity mappings in formalin-fixed, restriction-digested, religated NeuN^+^ neuronal nuclei collected by fluorescence-activated sorting (to avoid signal contribution from non-neuronal nuclei) in adult *CK-Cre^+^Setdb1^2lox/2lox^* and *CK-Cre^\u2212^Setdb1^2lox/2lox^* cortex ([Figure 1B](#F1){ref-type=\"fig\"}). However, chromosomal contact mappings at 40kb resolution (N=2 in situ Hi-C libraries/genotype, with 250--300M aligned reads/library, [Figure 1b](#F1){ref-type=\"fig\"} and [Supplementary Figure 2a](#SD1){ref-type=\"supplementary-material\"}) showed that mutant neurons were not affected by a generalized disorganization of the 3D genome. For example, length distributions and numbers of autosomal TADs, assessed by TADtree^[@R1]^, were indistinguishable and minimally different between genotypes, with \\~200kb median length as expected for mammalian genomes^[@R1]^ ([Figure 1c](#F1){ref-type=\"fig\"}, [Supplementary Figure 2b](#SD1){ref-type=\"supplementary-material\"}). We then assessed longer-range chromosomal contacts, spanning \\>200kb of linear genome. We identified genome-wide 110 long-range loop contacts affected in mutant neurons (DESeq2 P\\<0.05). Unexpectedly, the large majority, 84/110 or 76%, represented clustered, locus-specific 'loop aggregates' showing massive weakening, or complete loss, after neuronal *Setdb1* ablation ([Figure 1d](#F1){ref-type=\"fig\"}, [Supplementary Figure 2c](#SD1){ref-type=\"supplementary-material\"}). These included a singular hotspot on chromosome 18 with remarkable chromosome-wide enrichment (1Mb sliding window (1Mb^sw^) Poisson test P=1.2\u00d710^\u221224^, [Figure 1d,e](#F1){ref-type=\"fig\"}), fully engulfing the *clustered Protocadherin* (cPcdh) locus harboring 77 genes including 58 cell adhesion molecules linearly arranged as three gene clusters (*Pcdh* \u03b1, \u03b2,\u03b3), regulating neuronal connectivity^[@R13],\\ [@R14]^. Closer inspection of the wildtype *cPcdh* domain revealed multiple small (\\~100kb length) cluster-specific subTADs nested into a massive superTAD encompassing at least 1.2Mb of linear genome. TADtree analyses confirmed that this superTAD completely disintegrated after *Setdb1* ablation, leaving behind only subTAD remnants in mutant neurons ([Figure 1e](#F1){ref-type=\"fig\"}, [Supplementary Figure 2d](#SD1){ref-type=\"supplementary-material\"}). Additional loci, including distal portions of chromosomes 5 and 7, showed partial loss of large-sized TADs ([Figure 1d](#F1){ref-type=\"fig\"}, [Supplementary Figure 2e](#SD1){ref-type=\"supplementary-material\"}). Insulation, measuring the strength of physical segregation of neighboring chromosomal sequences, informs about functional compartmentalization of chromatin^[@R15]^. We quantified contact insulation across multiple DNA-DNA contact scales or 'bands' ([Figure 1F](#F1){ref-type=\"fig\"}) defined by increasing genomic distance^[@R16]^. Indeed, wildtype neurons showed very strong insulation scores at the fringes of the *cPcdh* locus. However, insulation of corresponding sequences in mutant neurons was dramatically weakened across multiple bands ([Figure 1f](#F1){ref-type=\"fig\"}). Therefore, multiple computational approaches, including (i) TADtree, (ii) long-range contact mapping and (iii) insulation analysis reveal structural disintegration of the superTAD*^cPcdh^* after *Setdb1* ablation.\n\nSetdb1 shields neuronal genomes from excess CTCF occupancy {#S4}\n----------------------------------------------------------\n\nHow could neuronal *Setdb1* ablation trigger such highly localized alterations in chromosomal conformations? To explore the role of Setdb1-regulated repressive histone methylation, we charted the Setdb1 product, trimethyl-histone H3-lysine 9 (H3K9me3), in NeuN^+^ and, for comparison, NeuN^\u2212^ nuclei sorted from adult *Cre^+^Setdb1^2lox/2lox^* and *CK-Cre^\u2212^ Setdb1^2lox/2lox^* cortex. DiffRep-based analysis with 1kb^sw^ revealed that 75% of 2,021 differentially H3K9me3-tagged sequences were hypomethylated in mutant neurons. These deficits were specific, because ChiP-seq profiling for open chromatin-associated acetyl-H3-lysine 27 (H3K27ac) showed that 96.4% of 1,112 differentially tagged sequences were hyperacetylated in *Setdb1*-deficient neurons ([Figure 2a](#F2){ref-type=\"fig\"}, [Supplementary Tables 1,2](#SD1){ref-type=\"supplementary-material\"}). Furthermore, *cPcdh* emerged genome-wide as top scoring locus for H3K9me3 hypomethylation (1Mb^sw^: H3K9me3, 35-fold enrichment, observed/expected 21/0.61, Poisson test P=3.19e^\u221225^) ([Figure 2b](#F2){ref-type=\"fig\"}), with the densely concentrated H3K9me3 deficit readily visible in whole chromosome 18 browser views ([Figure 2c](#F2){ref-type=\"fig\"}). In contrast, NeuN^\u2212^ nuclei sorted from the same cortical specimens were only minimally affected ([Figure 2a](#F2){ref-type=\"fig\"}, [Supplementary Figure 3](#SD1){ref-type=\"supplementary-material\"}, [Supplementary Tables 3,4](#SD1){ref-type=\"supplementary-material\"}).\n\nWe then analyzed motifs in H3K9me3 hypomethylated sequences in *Setdb1*-deficient neurons. Strikingly, 3/5 top scoring motifs matched to the transcriptional regulator and key 3D genome organizer CTCF, including the CTCFL/BORIS paralog (HOMER enrichment^[@R17]^, P\\<10^\u221260^) ([Figure 2d](#F2){ref-type=\"fig\"}, [Supplementary Table 5](#SD1){ref-type=\"supplementary-material\"}). Furthermore, we uncovered in published Setdb1 ChIP-seq data from stem cells and CD19^+^ B lymphocytes significant CTCF motif enrichment ([Supplementary Figure 4a](#SD1){ref-type=\"supplementary-material\"}, [Supplementary Tables 6,7](#SD1){ref-type=\"supplementary-material\"}). Of note, other types of H3K9 methyltransferase, including G9a/Glp, completely lacked CTCF motif enrichment ([Supplementary Figure 4b](#SD1){ref-type=\"supplementary-material\"}). We therefore predicted altered CTCF occupancy in the *Setdb1*-deficient neuronal genome. Strikingly, ChIP-seq on NeuN^+^ from adult *CK-Cre^+^Setdb1^2lox/2lox^* and *CK-Cre^\u2212^ Setdb1^2lox/2lox^* cortex showed that 99.4% (3059/3078) of sequences with altered CTCF binding represented up-regulated and *de novo* peaks ([Figure 2e](#F2){ref-type=\"fig\"}, [Supplementary Table 8](#SD1){ref-type=\"supplementary-material\"}), including many promoters and enhancers ([Supplementary Figure 5](#SD1){ref-type=\"supplementary-material\"}). There was extreme over-representation for CTCF motifs (HOMER enrichment P\\<10^\u22121000^) ([Figure 2f](#F2){ref-type=\"fig\"}, [Supplementary Table 9](#SD1){ref-type=\"supplementary-material\"}) independent of filter conditions ([Supplementary Figure 6](#SD1){ref-type=\"supplementary-material\"}, [Supplementary Table 10](#SD1){ref-type=\"supplementary-material\"}), affecting cis-regulatory elements and H3K9me3 hypomethylated sequences ([Supplementary Table 11](#SD1){ref-type=\"supplementary-material\"}). Therefore, Setdb1 shields mature neuronal genomes from excess CTCF occupancy at cryptic binding sites. Of note, *cPcdh* again emerged as top scoring locus genome-wide (1Mb^sw^, CTCF NeuN^+^ up-peaks: 18.7-fold enrichment, Poisson test P=1.32e^\u221221^, [Figure 2g,h](#F2){ref-type=\"fig\"}). Additional localized enrichments of excess CTCF occupancies matched to loci on chromosomes 5 and 7 affected by loss of long-range chromosomal contacts and H3K9 hypomethylation ([Figure 2g](#F2){ref-type=\"fig\"}).\n\nGiven that CTCF---a key regulator of higher order chromatin including domain insulation^[@R18]^ ---is upregulated at thousands of positions in the *Setdb1*-deficient neuronal genome, we assessed genome-wide domain insulation in the in situ Hi-C datasets from our mutant and wildtype cortical neurons. We first focused on CTCF *de novo* peaks, filtered for (i) proximity to TAD boundary (20% of total TAD length) and (ii) vicinity (\u00b1100kb) of sequences with altered H3K9me3 after *Setdb1* ablation. Of these, 52--57% of de novo CTCF showing stronger insulation scores in mutants across 8/9 insulation bands, covering 80--1040kb contact distance ([Supplementary Figure 7](#SD1){ref-type=\"supplementary-material\"}). At sites with conserved CTCF peaks, insulation scores showed very minimal differences between mutant and control neurons ([Supplementary Figure 8](#SD1){ref-type=\"supplementary-material\"}). Therefore, excess of CTCF on a genome-wide scale conveys a subtle shift towards increased insulation strength in mutant neurons, with the notable exception of Setdb1-sensitive superTADs affected by structural collapse and loss of insulation. Our findings, in conjunction with recent genome-scale studies reporting loss of domain insulation in glioma cells due to decreased CTCF binding^[@R19]^, suggest that spatial architectures of chromosomes are highly sensitive to bidirectional changes in CTCF occupancies. Next, we explored alterations in A/B compartments, defined as multi-Mb chromosomal segments representing 'A' open/'B' condensed chromatin tending to interact with other loci sharing similar levels of chromatin accessibility^[@R2]^. Because A/B compartments are defined on a continuum^[@R2]^ (as opposed to a biphasic signal), we quantified 'compartment-ness' from the intrachromosomal contact matrices generated by HiC-Pro at 100kb bin resolution (see Online Methods). Of note, the total number of A/B-specific compartment bins was only minimally different between genotypes. Strikingly, however, 6032/11048 or 54% of 'A' and 8468/12977 or 65% of 'B' bins had higher 'compartment-ness' scores in mutant compared to wildtype neurons (Fisher's exact P \\<10^\u2212210^) ([Supplementary Figure 9a](#SD1){ref-type=\"supplementary-material\"}). However, *Setdb1*-sensitive superTADs did not follow this genome-wide trend, as exemplified by the weakened 'B' signal at *cPcdh* in mutant neurons ([Supplementary Figure 9b](#SD1){ref-type=\"supplementary-material\"}).\n\nWhich molecular mechanisms contribute to the CTCF excess at the H3K9me3 hypomethylated sites? Of note, the majority of *cPcdh* sequences affected show coordinate increases in CTCF binding and histone hyperacetylation ([Figure 2h](#F2){ref-type=\"fig\"}), suggesting a shift towards open/permissive chromatin states. To this end, alterations in DNA cytosine methylation---reducing CTCF's DNA affinity^[@R20],\\ [@R21]^ via interaction with the seventh of CTCF's 11 zinc fingers^[@R22]^---could play a key role, because Setdb1 functions as upstream regulator for DNA methylation^[@R23]^. To explore, we quantified by bisulfite sequencing (bis-seq)^[@R24]^ levels of cytosine ^m^C5 methylation with 43 PCR amplicons targeting 13 *cPcdh* sites including cortical and striatal NeuN^+^ and NeuN^\u2212^ nuclei from adult *CK-Cre^+^Setdb1^2lox/2lox^* and *CK-Cre^\u2212^ Setdb1^2lox/2lox^* brain, plus cerebellar tissue as additional control, comprising 46 individual samples altogether ([Supplementary Table 12](#SD1){ref-type=\"supplementary-material\"}). As expected, ^m^C5 in non-neuronal nuclei and cerebellum remained unaltered. However, in *Setdb1*-deficient neurons, intergenic and promoter sequences affected by excess/de novo CTCF showed significant ^m^C5 deficits. In contrast, ^m^C5 levels were extremely low while CTCF peaks were very robust, independent of genotype, at sites harboring strong cPcdh enhancers elements including HS5-1 and HS16^[@R25]--[@R29]^ ([Figure 3](#F3){ref-type=\"fig\"}, [Supplementary Tables 12,13](#SD1){ref-type=\"supplementary-material\"}). Therefore, excess CTCF in *Setdb1*-deficient neuronal genomes is associated with 'open' chromatin state conversion at cryptic CTCF binding sites. This includes reduced DNA methylation levels, and weakening of regulatory mechanism designed to prevent excess CTCF binding.\n\nTAD-specific regulation of gene expression {#S5}\n------------------------------------------\n\nTo explore whether disintegration of superTAD*^cPcdh^* affects gene expression, we mapped transcriptomes and neuronal H3K27ac in *Cre^+^Setdb1^2lox/2lox^* and *CK-Cre^\u2212^ Setdb1^2lox/2lox^* cortex. Consistent with Setdb1's repressor function^[@R8]^, the majority of transcripts altered in mutant were up-regulated (208/321) ([Supplementary Table 14](#SD1){ref-type=\"supplementary-material\"}). Importantly, 20% of the entire pool of *Setdb1*-sensitive genes located to *cPcdh,* affected Protocadherins and non-Protocadherins, resulting in a unique, 543-fold enrichment on a genome-wide scale (Poisson test P=2.32e^\u221289^; 1Mb^sw^ sliding windows applied to N=208 transcripts) ([Figure 4a,b](#F4){ref-type=\"fig\"}). Similarly, among the 1070 sequences with histone hyperacetylation in mutant neurons (DiffRep 1kb^sw^, adj. P\\<0.05, [Supplementary Table 2](#SD1){ref-type=\"supplementary-material\"}), the *cPcdh* locus was uniquely affected with 96-fold enrichment on a genome-wide scale (1MB^sw^ observed/expected 38/0.42, Poisson test P=6.06e^\u221260^) ([Figure 4a](#F4){ref-type=\"fig\"}).\n\nTo confirm that such extremely locus-specific accumulation of up-regulated transcripts and H3K27ac hyperacetylated sequences is indeed driven by neuronal ablation of *Setdb1*, we reintroduced full length *Setdb1*, via a *CK*-*Setdb1* transgene^[@R30]^, into the conditional mutant line ([Supplementary Figure 10](#SD1){ref-type=\"supplementary-material\"}). Parallel testing of four genotypes, or *CK-Cre^+^ Setdb1^2lox/2lox^* and *CK-Cre^\u2212^ Setdb1^2lox/2lox(2lox/wt)^* each with and without *CK-Setdb1* (N=6 mice/genotype) confirmed complete rescue with return to baseline expression for 33/43, or 76% of \u03b1, \u03b2 and \u03b3 *Protocadherins* and of additional (non-Protocadherin) genes in *cPcdh* ([Figure 4c](#F4){ref-type=\"fig\"}, [Supplementary Table 15](#SD1){ref-type=\"supplementary-material\"}). To further test whether *cPcdh's* unique vulnerability is specific for *Setdb1* deletions in postnatal neurons, we profiled adult cortical and striatal transcriptomes after *CK-Cre* ablation of neuronal *G9a/Glp*, encoding a H3K9 methyltransferase complex essential for normal brain function^[@R31],\\ [@R32]^. In addition, we profiled transcriptomes from embryonic day E15.5 Nestin-Cre^+^,*Setdb1^2lox/2lox^* cortex. None of these various transcriptome sets showed local enrichment at *cPcdh* ([Supplementary Figure 11](#SD1){ref-type=\"supplementary-material\"}). Therefore, Setdb1 exerts unique transcriptional control across the *cPcdh* domain specifically in mature neurons. However, this regulatory layer is not representative for other types of H3K9 methyltransferase, or for the prenatal *Setdb1*-deficient brain.\n\nOf note, 31/53 S-type (single-neuron stochastically expressed) *Protocadherin* \u03b1, \u03b2 and \u03b3 genes, critically important for neuronal diversity and connectivity^[@R33],\\ [@R34]^, showed increased expression after neuronal *Setdb1* ablation ([Figure 4b,c](#F4){ref-type=\"fig\"}). We studied S-type expression patterns by in situ hybridization providing single cell resolution, using probes specific for individual S-type *Pcdha1, Pcdha8, Pcdhb22* and *Pcdhga7*. Strikingly, brain sections from *Setdb1* conditional mutants but none of the three control genotypes including transgenic rescue showed massively increased numbers of robustly stained neurons diffusely distributed across cortical layers II-VI and hippocampus ([Figure 4C](#F4){ref-type=\"fig\"}). Cerebellar cortex, which in contrast to forebrain is lacking CK-Cre expression, remained unaffected in conditional mutant brain ([Figure 4c](#F4){ref-type=\"fig\"}, [Supplementary Figures 12--15](#SD1){ref-type=\"supplementary-material\"}). Therefore, S-type single neuron stochastic constraint is severely compromised in *Setdb1*-deficient neurons, contributing to up-regulated expression at the *cPcdh* locus. Given the critical importance of orderly *cPcdh* expression---including single-cell stochastic constraint of S-type *Pcdh* genes---for neuronal morphology and connectivity^[@R13],\\ [@R33],\\ [@R34],\\ [@R49]^, we quantified spine densities and diameters from layer III apical dendrites from Setdb1*^wt/wt^* and *Setdb1^2lox/2lox^* mice, crossed into a conditional line expressing membrane-bound GFP (GFP-F)^[@R50]^ for Golgi-like labeling after low-titer AAV8*^hSYN1-CreGFP^* delivered to adult PFC. Indeed, spines from *Setdb1*-deficient neurons showed 40--50% increased density, and overall decreased size ([Supplementary Figure 1g](#SD1){ref-type=\"supplementary-material\"}), providing a morphological correlate for dysregulated *cPcdh* expression.\n\nBalanced facilitative and repressive conformations at cPcdh {#S6}\n-----------------------------------------------------------\n\nNext, we wanted to gain deeper mechanistic insight into the molecular mechanisms mediating the unique position of the *cPcdh* locus within the Setdb1-sensitive transcriptome and epigenome space. Because CTCF associates with RNA polymerase subunits^[@R35]^ and transcriptional activators^[@R36],\\ [@R37]^, CTCF-upregulation at H3K27ac-hyperacetylated S-type *Pcdh* \u03b1/\u03b2/\u03b3 promoters in *Setdb1*-deficient neurons could facilitate expression, including loss of single cell-stochastic constraint. However, promoter-bound CTCF alone is not sufficient to up-regulate transcription because from genome-wide 63 genes with excess CTCF around the transcription start site, only transcripts within the *cPcdh* locus were increased. Of note, promoter-enhancer loopings furnished by the CTCF-cohesin scaffolding complex contribute to transcriptional regulation of *cPcdh* genes^[@R20]^ and therefore, excess CTCF occupancy in cPcdh sequences from *Setdb1*-deficient neurons could trigger alterations in higher order chromatin. Indeed, excessive CTCF binding at the *cPcdh* locus of *Setdb1* mutant neurons was not limited to promoters, because multiple CTCF peaks emerged *de novo* in intergenic DNA upstream from \u03b1, and within the \u03b3 cluster (peaks A-C in [Figure 2h](#F2){ref-type=\"fig\"}). Importantly, these *de novo* peaks were surrounded by broad \\>100--200kb stretches of H3K9me3-tagged chromatin that underwent significant 'shrinkage' after neuronal *Setdb1* ablation (labeled 'R1' and 'R2' in [Figure 5a](#F5){ref-type=\"fig\"}). Importantly, 'R1' and 'R2' marked the anchor regions of massive bundles of long-range chromosomal conformations in wildtype neurons. Thus, densely spaced H3K9me3-tagged 'R1' loopings, emanating from 100--200kb wide blocks of repressive chromatin upstream of Pcdh\u03b1 genes, radiated towards many sites within *cPcdh,* even reaching the distal-most Pcdh\u03b3 sequences. However, these long-range loopings became completely dissolved upon structural disintegration of the superTAD*^cPcdh^* ([Figure 1e](#F1){ref-type=\"fig\"}). Among H3K9me3-tagged conformations lost after neuronal *Setdb1* deletion were multiple loopings interconnecting R1 and R2 with two DNAse I hypersensitive enhancer elements, HS16- and HS5-1, previously shown to broadly facilitate *cPcdh* expression^[@R25]--[@R29]^. These defects in HS16/HS5-1 bound long-range contacts were highly specific, because mutant neurons fully maintained shorter-range loopings from *Protocadherin* gene promoters to HS16/HS5-1 enhancers within the subTADs ([Figure 1e](#F1){ref-type=\"fig\"}). We confirmed these Hi-C findings, including specific weakening of long-range R1---HS16, R1---R2 and R2---HS5-1 and preservation of shorter-range contacts, in neuron-specific chromosome conformation capture (3C) PCR assays from adult mutant and control cortex ([Figure 5b](#F5){ref-type=\"fig\"}, [Supplementary Figure 16a](#SD1){ref-type=\"supplementary-material\"}). These studies, taken together, would suggest that in wildtype, HS16 and HS5-1 enhancer sequences are 'locked' into H3K9me3-tagged repressive chromatin. Upon *Setdb1* deletion, loss of R1/R2 repressive loop formations could release the 'epigenomic brake', thereby shifting the balance from repressive towards facilitative contacts furnished by HS16 and HS5-1-bound promoter-enhancer loopings, thereby triggering increased expression across the *cPcdh* locus ([Figure 5c](#F5){ref-type=\"fig\"}). To test this hypothesis, we transfected NG108 neuroblastoma cells with small RNA guided (sgRNA) Cas9-SunTag protein scaffolds^[@R38]^ designed to load ten copies of the potent transcriptional activator, VP64, onto single HS16 sites ([Figure 5d](#F5){ref-type=\"fig\"}). Therefore, such type of HS16^Cas9-SunTag(10xVP64)^ 'epigenomic superactivation' could, like the loss of R1---HS16 and R2---HS5-1 repressive loopings after *Setdb1* ablation, increase transcription at multiple positions across the entire 1Mb *cPcdh* locus via promoter-enhancer contacts and other mechanisms. Dual labeled cells (*sgRNA*-*dCas9-10xGCN4^'SunTag'^-blue fluorescent protein* and *scFv^(recognizing\\ GCN4\\ epitope)^-superfold GFP-VP64*^[@R38]^) were compared to controls expressing exactly the same types of vectors but without the sgRNA. Indeed, HS16 epigenomic superactivation was associated with increased expression of 3/6 c*Pcdh* transcripts (pre-selected for consistent baseline expression in neuroblastoma cells), closely mimicking the transcriptional phenotype in *Setdb1* mutant cortex ([Figure 5e](#F5){ref-type=\"fig\"}).\n\nConserved regulation of human and mouse superTAD^cPCDH^ {#S7}\n-------------------------------------------------------\n\nThe linear arrangement of \u03b1, \u03b2 and \u03b3 clusters with S- and C-type *Protocadherin* genes is highly conserved across vertebrate genomes^[@R13]^. We showed that higher order chromatin, including broad \\>100--200kb stretches of intergenic Setdb1-regulated H3K9me3-tagged sequence associated with repressive loop bundles, critically regulates transcription across the *cPcdh* locus. We asked whether such types of 3D genome conformations, just like the linear genome, could be conserved across mammalian lineages. To examine, we generated in situ Hi-C interaction matrices in human glutamatergic neurons differentiated from induced pluripotent stem cell-derived neural precursors by controlled expression of Neurogenin 2, and compared the 3D genome map to wildtype mouse cortex NeuN^+^ nuclei (of which \\>80% are from excitatory neurons). Indeed, TAD landscapes surrounding the *cPcdh/PCDH* locus (*Mm10* chr. 18; *Hg19* chr. 5) showed startling similarities between mouse cortex NeuN^+^ nuclei and human neurons, including complete preservation of cluster-specific subTADs nested into a large Mb-scale superTAD*^cPCDH^*. In addition, human and mouse neuronal chromatin exhibited highly similar-shaped H3K9me3 landscapes, including the broadly stretched aforementioned Setdb1-sensitive 'R1' at the superTAD's 5' end and 'R2' around the 5'end of the \u03b3 cluster ([Figure 6a](#F6){ref-type=\"fig\"}). We were surprised to discover that 'R1' near-perfectly matched a risk haplotype (*chr5:140,023,664-140,222,664*) of the Psychiatric Genomics Consortium^[@R39]^. This haplotype (no. 108 in reference^[@R39]^, referred to as 'PGC' hereafter) significantly contributes, independently from another 107 loci genome-wide, to schizophrenia heritability^[@R39]^, with a small INDEL as the lead polymorphism (rs111896713 chr5:*140,143,664*^[@R39]^). This risk polymorphism matched to robust Setdb1 peaks conserved in human and mouse cells including brain, but 'replaced' by *de novo* CTCF peaks upon *Setdb1* ablation ([Figure 6a](#F6){ref-type=\"fig\"}, [Supplementary Figure 16b,d](#SD1){ref-type=\"supplementary-material\"}). Therefore, we predicted that higher order chromatin organization at these positions will be highly conserved in human brain cells, and specifically in neurons, with long-range loopings radiating from \\~200kb R1 towards *cPCDH* promoter and enhancers primarily anchored in chromatin at and around the Setdb1 peak. To explore, we surveyed with 40kb resolution the *cPCDH*-bound chromosomal contacts in our in situ Hi-C datasets generated from human neurons and their isogenic neural precursors cells (NPC) and NPC-differentiated astrocytes. Strikingly, 40kb bins within PGC showed a step-wise progression in contact intensities with *cPCDH* sequences, culminating in massively increased contact frequencies at the bin harboring a robust Setdb1 peak ('PGC-3' in [Figure 6b](#F6){ref-type=\"fig\"}). This effect was pronounced in neurons and NPC, while corresponding loopings were much weaker or missing altogether in our contact maps from astrocytes, indicating strong cell type-specific regulation of local 3D genome architectures ([Figure 6b](#F6){ref-type=\"fig\"}). Next, we asked whether these within-PGC haplotype differences in *cPCDH* interaction frequencies translate into differential repressive potential. To this end, we introduced small guide RNAs (sgRNA) into two stable NPC lines, expressing 1. dCas9-KRAB fusion protein tethering KAP1 (KRAB-associated protein 1)-Setdb1 repressor complex^[@R8],\\ [@R40]^, or 2. dCas9-VP64 to dock the VP64 activator at different positions within the schizophrenia risk haplotype. We then measured expression levels for S-type \u03b3Pcdh genes expressed in NPC at comparatively high levels at baseline (data not shown). Interestingly, KRAB recruited to sequences close to the Setdb1 peak at the risk haplotype's lead polymorphism ('PGC-3' in [Figure 6c](#F6){ref-type=\"fig\"}), was consistently associated in 3/3 experiments with a robust multifold decrease in expression of *PCDHGB6* (but not *PCDHGA3*). In contrast, dCas9-KRAB docked to a non-Setdb1 binding site ('PGC-2' in [Figure 6c](#F6){ref-type=\"fig\"}) or to a scrambled control sequence ('Scr' in [Figure 6c](#F6){ref-type=\"fig\"}) remained ineffective and did not suppress *PCDHGB6* and *PCDHGA3* expression ([Figure 6c](#F6){ref-type=\"fig\"}). Of note, VP64 epigenomic editing at 'PGC-3' and neighboring 'PGC-2' was associated with increased expression of a subset of *cPCDH* genes ([Figure 6c](#F6){ref-type=\"fig\"}). These findings, taken together, suggest that repressive effects on Protocadherin gene expression are specific for loop-bound KRAB positioned at intergenic 'PGC-3' sequences upstream of the *cPCDH* gene clusters.\n\nIt is remarkable that KRAB---a critical module in KRAB-zinc finger proteins (KRAB-ZNF/Zfp) important for sequence-specific docking of the KAP1-Setdb1 repressor complex^[@R8],\\ [@R40]^---inhibits *cPCDH* expression via long-range loopings, bypassing 644kb linear genome in case of *PCDHGB6* ([Figure 6c](#F6){ref-type=\"fig\"}). Therefore, intergenic R1/risk haplotype-bound Setdb1 is likely to function as key transcriptional regulator at the *cPCDH* locus. These intergenic sequences harbor in the ENCODE database matching peaks for Setdb1, KAP1 and multiple KRAB-ZNF proteins including ZNF 274^[@R41]^ and ZFP143 ([Figure 6a](#F6){ref-type=\"fig\"}). Importantly, ZFP143 recognition sequences emerged as top scoring zinc finger motifs enriched at sites with excessive CTCF binding in *Setdb1*-deficient mouse neurons ([Figure 6d](#F6){ref-type=\"fig\"}, [Supplementary Tables 9,10](#SD1){ref-type=\"supplementary-material\"}). ZNF143, like CTCF and cohesion considered a key organizer for the 3D genome^[@R42],\\ [@R43]^, co-assembles with positive and negative regulators of transcription depending on local chromatin context^[@R44]^. Unsurprisingly then, ZNF143 at the *cPCDH* locus occupies in addition to R1 repressive chromatin also promoters and HS16/HS5-1 enhancers ([Figure 6a](#F6){ref-type=\"fig\"}). Therefore, alterations in ZNF143 supply, affecting facilitative and repressive chromatin, could destabilize *cPcdh* expression. Indeed, small RNA-mediated *Zfp143* knock-down in mouse neuroblastoma cells was associated with decreased expression of multiple *cPcdh* genes ([Figure 6d](#F6){ref-type=\"fig\"}). Our studies, taken together, suggest that 1. regulatory 3D genome architectures at the *cPCDH* locus are highly conserved between mouse and human, 2. include SETDB1-KRAB-ZNF143 and CTCF as key organizers of local repressive and facilitative chromosomal conformations, 3. 'bundles' or 'aggregates' of Setdb1-dependent long-range repressive loopings radiating from intergenic DNA ('R1', 'R2') function as 'epigenomic brakes' for transcriptional control, counterbalancing facilitative shorter range promoter-enhancer contacts ([Figure 6e](#F6){ref-type=\"fig\"}).\n\nDiscussion {#S8}\n==========\n\nNeuronal *Setdb1* ablation triggers structural disintegration of megabase-scale TADs, including the *Protocadherin \u03b1/\u03b2/\u03b3* locus as the only Setdb1-sensitive TAD harboring a gene cluster. TADs affected in mutant neurons showed shrinkage of broadly stretched H3K9me3-tagged chromatin, in conjunction with localized hotspots of excess and *de novo* CTCF binding. Setdb1-regulated long-range repressive *cPcdh* loopings were highly enriched in neurons as compared to their isogenic precursors, and carried DNA polymorphisms conferring liability for schizophrenia. 3D genome conformations at *cPCDH* could have even broader relevance for neuropsychiatric disease, given that *SETDB1* microdeletions and structural variants are associated with neurodevelopmental delay^[@R51],\\ [@R52]^, with CpG hypermethylation reported for orthologous CTCF binding sites within the *PCDH* gene cluster in Down syndrome (trisomy 21) including the mouse model^[@R53]^, and *cPcdh* DNA promoter methylation linked to depression and anxiety^[@R54]--[@R56]^. Furthermore, mice exposed to chronic variable stress, a preclinical paradigm frequently implied in psychiatric disease^[@R57]^, show hyperexpression of \u03b2 *Protocadherin genes* ([Supplementary Figure 17](#SD1){ref-type=\"supplementary-material\"}).\n\nWe show that Setdb1, maintaining high levels of DNA methylation and low levels of histone acetylation at sequences in close vicinity or partial overlap with potential CTCF binding sites, critically shields neuronal genomes from uncontrolled CTCF docking at thousands of cryptic binding sites genome-wide. However, upon neuronal *Setdb1* ablation, the shield becomes defunct, triggering collapse of vulnerable TADs. Remarkably, our findings on excessive *cPcdh* CTCF occupancies and increased *cPcdh* expression and the resulting increase in spine densities in *Setdb1*-deficient neurons are perfect opposites to the previously reported *decreases* in *cPcdh* expression and spine densities after neuronal *Ctcf* ablation^[@R58]^. Likewise, switching 'reverse-forward' strand orientations in CTCF binding sequences^[@R25]^ could disrupt promoter-enhancer loopings and broadly dampen transcription across \u03b1/\u03b2/\u03b3 clusters (figure 2d in reference^[@R25]^). These findings strongly point to delicate regulatory mechanisms governing chromosomal conformations, with genomes excessively populated by CTCF, including the *Setdb1*-deficient neurons of the present study, showing disintegration of higher order chromatin in locus-specific manner. The structural collapse of the *Setdb1*-sensitive superTADs was highly specific, given that the genome-wide excess of CTCF binding in mutant neurons triggered a genome-wide *increase* in insulation strength and 'compartment-ness'.\n\nFuture work will clarify whether neuronal Setdb1 overexpression^[@R30],\\ [@R59]^, or loss of other proteins assigned with regulation of *cPcdh* expression, including DNMT3b cytosine methyltransferase^[@R34]^ and SMCHD1^[@R60]^ and WIZ^[@R61]^ repressors, could trigger TAD-specific 3D genome changes in neurons. We note that Setdb1 is primarily located towards the 5' and 3' ends of superTAD*^cPcdh^* ([Figure 6a,e](#F6){ref-type=\"fig\"}, [Supplementary Figure 18](#SD1){ref-type=\"supplementary-material\"}). In any case, the TAD-specific phenotypes in *Setdb1* mutant neurons point towards unexplored modular complexities in the regulatory mechanisms governing the 3D genome. Thus, rewiring or disintegration of specific TAD units may not be exclusive to chromosomal microdeletion and --duplication events^[@R62],\\ [@R63]^, because as shown here, loss of Setdb1 function triggers the disintegration of highly select subset of neuronal TADs. With each chromosome furnishing hundreds of TAD-like structures, it will be an exciting and challenging task to dissect 'TAD-by-TAD' and in cell-type specific manner, the multilayered mechanisms governing locus-specific higher order chromatin in highly differentiated brain cells.\n\nONLINE METHODS {#S9}\n==============\n\nHuman Stem Cell Lines {#S10}\n---------------------\n\nAll work with human induced pluripotent stem cell lines has been approved by the Institutional Review Board of the Mount Sinai School of Medicine, in accordance with Mount Sinai's Federal Wide Assurances (FWA\\#00005656, FWA\\#00005651) to the Department of Health and Human Services. No new stem cell lines had been generated for the work presented here. Informed consent had been obtained from all participating subjects. See [Supplementary Methods](#SD1){ref-type=\"supplementary-material\"} for differentiation into neural pregenitors, glutamatergic neurons and astrocytes.\n\nAnimal studies {#S11}\n--------------\n\nAll animal work was approved by the Institutional Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai. Mice were held under specific pathogen-free conditions with food and water being supplied ad libitum in an animal facility with a reversed 12 h light/dark cycle (light off at 7:00 am) under constant conditions (21 \u00b1 1\u00b0C; 60% humidity). All animals were group housed (2--5/cage).\n\nGeneration of Setdb1 conditional mutant and rescue mice {#S12}\n-------------------------------------------------------\n\n*Setdb1^2lox/2lox^* mice were generated by Ozgene, Australia. In brief, two loxP sites were inserted to endogenous *Setdb1* locus (*Setdb1* ENSMUSG00000015697; SET domain, bifurcated 1, *MGI:1934229*) flanking exon 3. To generate conditional *Setdb1* knockout mice, *Setdb1*^2lox/2lox^ mice were first crossed with *CK-Cre^+/\u2212^* transgenic mice to generate *Setdb1^2lox/+^*, *CK-Cre^+/0^* heterozygous, which were further crossed with *Setdb1^2lox/2lox^* mice to generate *Setdb1^2lox/2lox^*, *CK-Cre^+/0^* homozygous. Cre recombinase mediated excision of *Setdb1* exon 3 causes frame shift and generates a stop codon at the new junction of exon 2 and exon 4, and results in early termination of *Setdb1* translation. Gender and age matched littermates with genotype *Setdb1^2lox/2lox^*, *CK-Cre^0/0^* were used as controls with wildtype SETDB1 levels. *CK-Setdb1* transgenic mice, described previously^[@R64]^, express full-length mouse *Setdb1* cDNA driven by CK promoter in postnatal and adult mouse forebrain. To generate *Setdb1* rescue mice, the *CK-Setdb1* transgene was introduced into *Setdb1^2lox/2lox^* conditional knockout background and *CK-Setdb1*^+/0^, *Setdb1^2lox/2lox^* mice were crossed with *Setdb1^2lox/+^*, *CK-Cre^+/0^* to generate *Setdb1^2lox/2lox^*, *CK-Cre^+/0^*, *CK-Setdb1*^+/0^ rescue mice. Gender and age matched littermates with genotype *Setdb1^2lox/2lox^*, *CK-Cre^0/0^*, *CK-Setdb1*^0/0^ were used as wildtype controls. All genetically engineered lines were backcrossed to the C57BL6/J line for at least 10 generations.\n\nSee [Supplementary Methods](#SD1){ref-type=\"supplementary-material\"} for information on Nestin-Cre conditional mutagenesis, Mendelian survival ratios, Histology, Comet-assay and RNA quantifications including RNA-seq.\n\nChromatin assays {#S13}\n----------------\n\nChromatin assays (ChIP-seq, in situ Hi-C, 3C-PCR) and RNA-seq were conducted in young adult mice, at 3 months of age (\u00b1 2 weeks). [Supplemental Tables 1--4, 8 and 12](#SD1){ref-type=\"supplementary-material\"} provide additional information for each chromatin assay, including number of animals and sex ratios.\n\n### Nuclei preparation, immunotagging and fluorescence-activated sorting {#S14}\n\nFor fluorescence-activated nuclei sorting, nuclei were extracted from mouse cerebral cortex or human prefrontal cerebral cortex (control, PFC, male, PMI 17) and [anterior cingulate cortex]{.ul} (control, ACC, female, PMI 27) as described before^[@R65]^. In brief, brain tissue was homogenized in hypotonic lysis solution, purified by ultra-centrifugation, and then re-suspended in 1 ml DPBS containing 0.1% BSA, 1:1000 Anti-NeuN antibody, clone A60, Alexa Fluor\u00ae488 conjugated (EMD MILLIPORE CORP MAB377X). Samples were incubated for at least 45 min by rotating in the dark at 4\u00b0C. DAPI was added before FACS to label all the nuclei. Sorting was done on Flow Cytometry Center at Mount Sinai. Nuclei were separated into NeuN+ and NeuN- population and then pelleted for following applications. For XChIP, 3C and *in situ* HiC experiments, 10 minutes of 1% formaline fixation at room temperature was incorporated right after brain homogenization. Cross-linking was quenched by incubating with 125 mM glycine. Nuclei were then purified, stained and sorted as described above.\n\n### ChIP-seq {#S15}\n\nNative immunoprecipitation (NChIP) was performed as described^[@R65]^. In brief, NeuN^+^ (neuronal) and NeuN^\u2212^ (non-neuronal) nuclei were pelleted post-FACS and then resuspended in 300 ul of MNase digestion buffer (10 mm Tris, pH 7.5; 4 mm MgCl2; and 1 mm Ca^2+^), digested with 3ul of MNase (0.2 U/\u03bcl) for 5 min at 28\u00b0C to obtain mono-nucleosomes. Reaction was stopped with 50mM EDTA, pH 8. Nuclei were swollen to release chromatin after addition of hypotonization buffer (0.2 mm EDTA, pH 8, containing PMSF, DTT, and benzamidine). Chromatin was incubated with anti-H3K9me3 (Abcam AB8898) and anti-H3K27ac (Active Motif, \\#39133) antibodies overnight at 4\u00b0C. The DNA-protein-antibody complexes were captured by Protein AG Magnetic Beads (Thermo Scientific^\u2122^ 88803) by incubating at 4\u00b0C for 2 hours. Magnetic beads were then washed with low-salt buffer, high-salt buffer, and TE buffer. DNA was eluted from the beads, and treated with RNase A followed by proteinase K digestion. DNA was purified by phenol-chloroform extraction and ethanol precipitation.\n\nFor XChIP on crosslinked preparations, formaldehyde-fixed NeuN^+^ nuclei after FACS were resuspended in lysis buffer containing 0.1%SDS, sonicated ([Bioruptor\u00ae Plus sonication device]{.ul}, Diagenode) at the 'high' setting for 30 minutes on ice. The size of DNA fragmentation was between 100 bp to 500 bp with an average size of 300bp. Chromatin was then incubated with anti-CTCF (EMD Millipore, \\# 07729) or anti-SETDB1 (Santa Cruz H-300X \\#\\#sc-66884 X) or anti-SETDB1 (Thermofisher 5H6A12 \\#MA5-15722) and captured with Protein AG Magnetic Beads. After washing and elution, DNA was incubated at 65\u00b0C overnight for reverse cross-linking, followed by RNase A, proteinase K treatment and DNA precipitation. Setdb1 occupancies were measured by conventional ChIP-PCR. For CTCF ChIP-seq library preparation, ChIP DNA was end repaired (End-it DNA Repair kit; Epicentre) and A tailed (Klenow Exo-minus; Epicentre). Adaptors (Illumina) were ligated to the ChIP-DNA (Fast-Link kit; Epicentre) and then PCR amplified using Illumina TruSeq ChIP Library Prep Kit. Library DNA with expected size (NChIP, \\~275bp; XChIP, 350bp to 500bp) was selected by Pippin and submitted to New York Genomic Center and sequenced with Illumina HiSeq 2000, 75bp, paired end. The ChIP-seq data was first checked for quality using the various metrics generated by FastQC (v0.11.2). Raw sequencing reads were then aligned to the mouse mm10 genome (or Hg19 for human) using default settings of Bowtie (v2.2.0). Uniquely mapped reads were retained and the alignments were subsequently filtered using the SAMtools package (v0.1.19) to remove duplicate reads. Differential analysis between mutant and control samples was performed using diffReps with window size 1000 bp and moving step size 500 bp, and FDR\\<5% as significance cutoff^[@R66]^ and data visualized on the genome using the Integrative Genomics Viewer (IGV) program^[@R67]^. For H3K27ac and CTCF ChIPseq, peak-calling was performed using MACS (v2.1.1) with a FDR cutoff of 0.05. Gene Ontology enrichment of annotated genes, with significant hits from diffReps within gene bodies or within 3Kb around transcriptional starting sites, was further analyzed using DAVID Functional Annotation Bioinformatics tools (Resources 6.7, National Institute of Allergy and Infectious Diseases, NIH). Significant hits from diffReps for decrease in H3K9me3 and increase in CTCF ChIPseq were subjected to motif analysis, using the Homer package (v4.8.3) at default settings^[@R68]^. Manhattan plots for genome-wide differential epigenetic profiling of conditional mutants and controls were constructed after the genome was divided into non-overlapping 1Mb bins, including the 1Mb bin spanning the clustered *Pcdh* genes chr18:36,870,001-37,870,000, mm10. The number of occurrences of each signal was tabulated within each bin. The probability of the number of occurrence of each signal per 1Mb bin was then modeled using a Poisson distribution with the maximum likelihood estimator for the lambda parameter given by the calculated mean number of occurrences. The Poisson models for each signal were used to calculate the probability of occurrence of the signal observed in every 1Mb bin (including the *Pcdh* bin).\n\n### Chromosome Conformation Capture (3C) {#S16}\n\n3C was performed using standard protocols with minor modifications^[@R69]^. In brief, nuclei were fixed and extracted from mouse cerebral cortex and FACS sorted as described above. NeuN positive (neuronal) nuclei were then pelleted and digested with Hind III restriction enzyme (NEB) at 37\u00b0C overnight, washed, and treated with T4 DNA ligase at room temperature for 4 hr. 3C DNA was then incubated at 65\u00b0C overnight for reverse crosslinking followed by DNA purification and precipitation. 3C primers were listed Primers are listed in [Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}. Sequence-verified PCR products were measured semi-quantitatively with UVP Bioimaging system/Labworks 4.5 software. Neighboring primers at *B2m* gene locus was used for normalization.\n\n### in situ Hi-C including bioinformatical analyses {#S17}\n\nNuclei were fixed and extracted from mouse cerebral cortex and human postmortem anterior cingulate cortex and sorted into NeuN^+^ (neuronal) and NeuN^\u2212^ (non-neuronal) populations, which were then processed using an *in situ* Hi-C protocol^[@R70]^, with minor modifications. Briefly, the protocol involves a restriction digest of the cross-linked chromatin within intact nuclei, followed by biotinylation of the strand ends, re-ligation, sonication and size selection for 300--500bp fragments, followed by standard library preparation for Illumina sequencing. The resulting data were mapped, filtered, and normalized using HiC-Pro^[@R71]^ (v2.7.8) and visualized on the Washington University Epigenome Browser. To explore localized enrichments in the in situ Hi-C datasets, we tabulated for each 40kb bin along chromosome 18 the number of long-range interactions greater than 200kb that were disrupted (ie. significantly decreased in conditional CK-Cre mutants versus control, as detected using DESeq2 at P\\<0.05). The probability of observing the number of disrupted interactions at each bin was then modeled using a Poisson distribution with maximum likelihood of mean (0.165) calculated from the data. Topological associating domains (TAD) were predicted using TADtree^[@R70]^ using the 20kb Hi-Cpro data as input with the following parameter settings: maximum size of TAD in bins (S) = 60; maximum number of TADs in each tad-tree (M) = 10; boundary index parameter (p) = 6; boundary index parameter (q) = 24; balance between boundary index and squared error in score function (gamma) = 500; number of TADs to use (N) = 400 (chr18) or 700 (chr5).\n\nIn addition, initial processing of the raw 2\u00d7125bp read pair FASTQ files was performed using the HiC-Pro analysis pipeline. In brief, HiC-Pro performs four major tasks: aligning short reads, filtering for valid pairs, binning, and normalizing contact matrices. HiC-Pro implements the truncation-based alignment strategy using Bowtie v2.2.3^[@R72]^, mapping full reads end-to-end or the 5' portion of reads preceding a GATCGATC ligation site that results from restriction enzyme digestion with MboI followed by end ligation. Invalid interactions such as same-strand, dangling-end, self-cycle, and single-end pairs are not retained. Binning was performed in 40kb and 100kb non-overlapping, adjacent windows across the genome and resulting contact matrices were normalized using iterative correction and eigenvector decomposition (ICE) as previously described^[@R73]^.\n\nStarting with the 20 kb resolution intra-chromosomal contact matrices generated by HiC-Pro, we first generated 100 kb resolution contact matrices by summing the interaction frequencies of the 20 kb bins within each 100 kb bin. We next generated the corresponding log2(observed/expected) matrices, where the observed/expected values are the ratio of the contact values of each interaction bin to the average contact values of all interaction bins the same distance apart. The Pearson's correlation matrices were then calculated from the log2(observed/expected) matrices, and PCA was performed on them. The first principal component (PC1) was then used to differentiate the compartments. When the first principal component value was positively correlated to gene density and gene expression (we found that the first principal component always correlates with both gene density and gene expression in the same direction), bins with positive PC1 values are assigned as compartment A while bins with negative PC1 values are assigned as compartment B. Conversely, when PC1 is negatively correlated to gene density and gene expression, bins with negative PC1 values are assigned as compartment A while bins with positive PC1 values are assigned as compartment B.\n\nHigher-resolution topologically associated domain (TAD) calls were made following the procedure described by Dixon, et al.^[@R74]^ using the directionality index (DI) metric. DI was calculated using raw interaction counts between 40kb or 100kb bins and respective window sizes of 2Mb or 5Mb to capture observed upstream or downstream interaction bias of genomic regions. A Hidden Markov model (HMM) was then trained to infer true bias states. TADs were defined by pairing adjacent regions of inferred downstream or upstream bias states.\n\nTo identify significantly enriched interactions involving a bin of interest, the expected interaction counts for each interaction distance were estimated by calculating the mean of all intrachromosomal bin-bin interactions of the same separation distance throughout the entire ICE-normalized contact matrix. We estimated the probability of observing an interaction between a bin-of-interest and some other bin by calculating the expected interaction between those two bins divided by the sum of all expected interactions between the bin-of-interest and all other intrachromosomal bins. We then calculated the p-value of observing the observed number of interaction counts or more between the bin-of-interest and some other bin using a binomial test where the number of successes was defined as the observed interaction count, the number of tries as the total number of observed interactions between the bin-of-interest and all other intrachromosomal bins, and the success probability as the probability of observing the bin-bin interaction estimated from the expected mean interaction counts. To control false discovery rate, the R package, *qvalue*, was used to estimate q-values from the calculated binomial p-values.\n\nThe Insulation Analysis was performed with reference to Crane et al^[@R75]^ and Vietri Rudan et al^[@R76]^. Briefly, using the 20kb resolution HiC matrix, we calculated (at each 20kb bin) the average interaction frequency of the chromosomal bins within a certain distance band. The normalized \"insulation\" score along the chromosome for each band was then calculated as the log2 ratio of average interaction frequency at each 20kb bin to the average of all 20kb bins in that band. Regions along the chromosome that display a dip/valley/minima of normalized insulation values represent regions of reduced interactions, and can be interpreted as TAD boundaries or regions of high local insulation.\n\n### DNA methylation {#S18}\n\nTargeted bis-seq was utilized for fine-mapping of methylation patterns in *cPcdh* candidate sequences. Genomic DNA (1 \u03bcg) was bisulfite-converted using the EpiTect Bisulfite Kit (Qiagen). Primers were designed in MethPrimer^[@R77]^ and bisulfite-converted DNA was amplified and multiplexed by high throughput PCR using a Fluidigm AccessArray instrument. PCRs were performed in duplicate and the duplicates pooled. Primers are listed in [Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}. The library was diluted to a final concentration of 10pM with 35% of PhiX library. Paired-end reads (250 bp) were generated with an Illumina MiSeq sequencer. Fastq files were generated by the MiSeq sequencer. After trimming for low-quality bases (Phred score\\<30), Illumina and Fluidigm adaptors and reads with a length \\<40 bp with TrimGalore, the reads were aligned to the mouse genome (mm10) using Bismark^[@R78]^ using the following settings: -D 50 -R 10 \\--score_min L,0,-0.6. Since the sequences are PCR-based, reads were not deduplicated. Methylation calling was performed using Bismark extractor^[@R78]^. Net methylation was assessed when the coverage was at least 100X and reported by CpG and averaged across amplicon. Graphical representations of random samples of 50 sequenced DNA fragments were generated using R. Briefly, using CpG context output files generated by Bismark methylation extractor^[@R78]^ which reports CpG methylation status for each individual sequenced DNA fragments (taking into account paired-end reads), methylation patterns for each DNA fragment were reconstructed based on the coordinates of the covered CpGs and their methylation status. Off-target reads mapping outside the amplicon coordinates were discarded. After random sampling of 50 sequences using the R sample function, their methylation patterns were plotted using the R plot function. Only CpGs present in the reference genome were represented and sequences were represented on positive strand. Circles represent consecutive CpGs, with each line being a unique DNA fragment. White circles are unmethylated CpGs and black circles are methylated CpGs.\n\nTo ensure sufficient library complexity, we pooled two PCRs for each amplicon. The median coverage per library ranged from 1000X to 2600X ([Supplementary Figure 19A](#SD1){ref-type=\"supplementary-material\"}). Since we used a PCR based targeted bis-seq approach, the sequence start points are constrained for each amplicon. Therefore the constrained start sites generated duplication levels which should not be treated as technical duplicates nor removed by bioinformatics deduplication. These duplication levels will be reflected in the distribution of specific sequences that are identified using start sites and therefore not informative to assess the library complexity. Therefore, since some level of randomness of DNA methylation patterns in a cell population is present, to estimate the library complexity, we assessed the number of distinct methylation patterns (i.e. specific sequences based on C\\>T conversion) observed for each amplicon in a given library. Because of biological duplicates (genuinely distinct DNA molecules with the same methylation pattern), this metric provides highly stringent coverage information. In addition, to take into account sequencing errors, only methylation patterns representing more than 1% of the total reads covering a given amplicon were counted. Overall, the median of methylation patterns per amplicon was 5 (range:1--12.5), with 81.2% of the amplicons with less than 5 methylation patterns covering fully methylated (\\>80% of methylation) or fully unmethylated (\\<20%) regions, which are expected to show lower randomness level ([Supplementary Figure 19B](#SD1){ref-type=\"supplementary-material\"}).\n\nLenti-shRNA knockdown of Zfp143 {#S19}\n-------------------------------\n\nMouse shRNA lentiviral particles targeting Zfp143 (4 unique 29mer target-specific shRNA, 1 scramble control) were purchased from Origene (TL502149V), with sequences in the shRNA expression cassettes are verified by the manufacturer to correspond to the target gene Zfp143 (Gene ID 20841) with 100% identity, to produce 70% or more gene expression knock-down provided a minimum transfection efficiency of 80%. NG108 cells (NG108-15 \\#108CC15, Vendor: ATCC, Organism: Mus musculus (neuroblastoma); Rattus norvegicus (glioma)) were seeded in 12-well plates 12 hours before viral transduction. 72 hours after transduction, cells were lysed with 500ul of Trizol, total RNA was extracted, reverse transcribed and *Pcdh* gene expression quantified by real-time PCR. Primers are listed in [Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}.\n\ndCas9-SunTag(10xVP64) epigenomic editing {#S20}\n----------------------------------------\n\nThe CRISPR/dCas9_Suntag_VP64 two-plasmid system^[@R79]^ was used for genomic editing on *Pcdh* locus in the NG108 neuronal cell line. For pLV-U6-sgRNA-CK-dCas9-10xGCN4-BFP construct, lentiviral backbone and dCas9 cassette were cloned from plasmid Lenti-dCAS-VP64_Blast (Addgene, \\#61425), CamK-II promoter (1293 bp) was cloned from plasmid pAAV-CamKII-hChR2(T159C)-p2A-ETFP-WPR (a gift from Javier Maeso), and 10XGCN4-P2A-BFP cassette was cloned from pHRdSV40-dCas9-10xGCN4_V4-P2A-BFP(Addgene, \\#60903). The U6-sgRNA cassette was inserted upstream of CamK-II promoter. For the pLV-CK-scFv-GCN4-sfGFP-VP64-GB1-NLS construct, the lentiviral backbone and CK promoter cloned as described above. The scFV-GCN4-sfGFP-RsrII-GB1-NLS (1875 bp) sequence was cloned from plasmid pHR-scFv-GCN4-sfGFP-GB1-NLS-dWPRE (Addgene, \\#60906), and the VP64 cassette was cloned from plasmid pHRdSV40-scFv-GCN4-sfGFP-VP64-GB1-NLS (Addgene, \\#60904). All plasmid construction was done by the VectorBuilder team from Cyagen Biosciences. Three sgRNAs ([Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}) were designed to target mouse HS16 site at the *Pcdh* locus. For transfection, control cells received a two-vector system pLV-CK-dCas9-10xGCN4-BFP lacking the sgRNA cassettes and pLV-CK-scFv-GCN4-sfGFP-VP64-GB1-NLS; Cell cultures for HS16 superactivation were transfected with pLV-U6-sgRNA-CK-dCas9-10xGCN4-BFP plasmid that included the sgRNA cassettes targeting the HS16 site, and pLV-CK-scFv-GCN4-sfGFP-VP64-GB1-NLS. 72 hours after transfection, cells were harvested and sorted by FACS to collect BFP and GFP double positive cells. Total RNA was then extracted, reverse transcribed and *Pcdh* gene expression quantified by real-time PCR. Primers are listed. Primers are listed in [Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}.\n\ndCas9-KRAB epigenomic editing {#S21}\n-----------------------------\n\nHuman neural precursor cells (NPCs) were maintained at high density, grown on growth factor reduced Matrigel (BD Biosciences) coated plates in NPC media (Dulbecco's Modified Eagle Medium/Ham's F12 Nutrient Mixture (ThermoFisher Scientific), 1x N2, 1x B27-RA (ThermoFisher Scientific) and 20 ng ml^\u22121^ FGF2 (R & D Systems, 233-FB-10) and split 1:3 every week with Accutase (Millipore, Billerica, MA, USA). The hiPSC-NPC line 553-S1-1, as previously described and validated^[@R80],[@R81]^ was used in all NPC editing experiments. *Generation of stable dCas9-KRAB NPCs*: 3.5\u00d710^6^ NPCs per well were seeded onto growth factor reduced Matrigel coated 6-well plates in NPC media. The following day lentiviruses generated as above using either the lentiviral vectors dCas9:VP64-T2A-puro and dCas9:KRAB-T2A-puro were added and cultures were spinfected (1 hour, 1000xg, 25\u00b0C). Following spinfection, plates were transferred to a cell culture incubator for 3 hours. Media was then removed and replaced with fresh NPC media. The following day, fresh NPC medium containing 1 \u03bcg/ml puromycin (Sigma, \\#P7255) was added and cells were maintained in NPC medium containing 1 \u03bcg/ml puromycin for the remainder of the experiment. Stable NPC lines were validated via FACS using Cas9-AF488 antibody (Cell Signaling Technologies, 5uL/1\u00d710^6^ cells \\#34963S). *sgRNA design and cloning:* The sgRNAs were designed using the Optimized CRISPR Design tool (See URLs) at the genomic regions of interest. Guide RNAs were selected based on their specific locations at decreasing distances from region of interest as well as strand specificity and lack of predicted off targets. Synthesic oligonucleotides ([Supplementary Table 16](#SD1){ref-type=\"supplementary-material\"}) were annealed (95\u00b0C for 5 min, ramp down to 25\u00b0C at 5\u00b0C per minute), diluted 1:100 and then ligated into BsmB1 digested lentiGuide-dTomato. *NPC lentiviral transduction and FACS*: 100,000 dCas9-KRAB NPCs per well were seeded onto growth factor reduced Matrigel coated 24-well plates in NPC medium containing 1 \u03bcg/ml puromycin. The following day lentiviruses with scrambleD sgRNAs and pooled sgRNAs targeting PGC-1, PGC-2 and PGC-3 region were added to cultures in the presence of polybrene. 48 hours after transduction, NPC cells were FACS sorted and live cell population with dTomato signal were collected directly into Trizaol LS (Thermo Fisher, 10296028). Total RNA was extracted for RT-PCR.\n\nData Availability {#S22}\n-----------------\n\nAll next generation sequencing data for genome-scale analysis in this publication have been deposited in NCBI's Gene Expression Omnibus^[@R82]^ and are accessible through GEO Series accession number GSE99363 (See URLs). All other data discussed are included in the publication and available from the Authors upon request.\n\nURLs {#S23}\n----\n\nOptimized CRISPR Design tool, <http://crispr.mit.edu/>\n\nGSE99363, <https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE99363>\n\nSupplementary Material {#S24}\n======================\n\nSupported by National Institutes of Health, U.S.A. (NIH) grants R01MH106056 (S.A.), P50MH096890 (E.N.), R01MH101454, NIA U01P50AG005138-30-1 and U01AG046170 (K.B.), R01AG050986, R01MH109677, (P.Rou.), R01NS091574 (A.S.), and NIH training and fellowship awards 1F30MH113330 (P.Ra.) and T32-AG049688 (S.C.). Additional support was provided by a Grant-in-Aid for AMED-CREST, AMED, Japan and the Japan-U.S. Brain Research Cooperation Program to T.Y., the Veterans Affairs Merit grant BX002395 (P.Rou.), the Brain & Behavior Research Foundation (Y.J., P.Rou.), the Alzheimer's Association (P.Rou.), the New York Stem Cell Foundation (K.B.) and the Brain Research Foundation (S.A.).\n\n**Competing Financial Interest**: The Authors declare no conflicts of interest.\n\n**Author Contributions**: Performed experiments: Y.J., P.Raj., T.H., B.K., B.H., S-M.H., B.J., L.K., R.P., S.C., C.D., C.P., J.TCW, B.S. Conceived and designed experiments: Y.J., S.A. Performed statistical analyses: Y.J.. Bioinformatics and genomic analyses: E.L., P.Raj., W.L., P.Rou., L.S.. Contributed materials: A.S., B.R. (G9a/Glp transcriptome data); B.L., E.N. (mouse stress model/transcriptome data). Supervised the DNA methylation analysis: B.T.. Supervised research: H.M., K.B., T.Y., L.S., S.A.. Wrote the paper: Y.J., Y-H. E.L., B.T., L.S., S.A. with contributions from co-authors.\n\n![3D genomes in *Setdb1*-deficient cortical neurons\\\n(**a**) (Left) Conditional *Setdb1* ablation with loxP sites surrounding exon 3. Recombination results in frame shift and premature stop (TGA) upstream of Tudor, methyl-CpG-binding (MBD) and catalytic SET domains. (Right) Setdb1 immunoblot (histone H3 loading control) (complete blot shown in [Supplementary Figure 1c](#SD1){ref-type=\"supplementary-material\"}) and RNA-seq from adult *CK-Cre^+^ Setdb1^2lox/2lox^* mutant (K), in comparison to *CK-Cre^\u2212^Setdb1^2lox/2lox^* (WT) cortex. **(b)** (Left) Flow cytometry-based sorting of adult cortex NeuN immunotagged nuclei. (Right) genome-scale *in situ Hi-C* contact matrix from WT and KO NeuN^+^ nuclei. (**c**) TAD numbers per autosome for mutant (KO) and wildtype (WT) NeuN^+^, (N=2/genotype). (**d**) Manhattan plot summarizing loss of long-range DNA loop contacts bypassing \\>200kb linear genome in KO compared to WT NeuN^+^. Notice localized aggregates of densely spaced loop losses on chromosomes 5, 7 and 18. (**e**) *in situ* HiC 2Mb window showing chromosome 18 conformations in KO and WT at position marked by red arrow in Manhattan blot in panel D, with TADs called (TADtree) in both genotypes marked gray. Large 'superTAD' called in WT but lost in KO shown as red line. (**f**) Contact insulation map for 3Mb window centered on *cPcdh* locus. (Top) Heatmaps from WT and KO cortical neurons for 9 bands, from 0--80kb to 920--1040kb distance. KO shows loss of superTAD*^cPcdh^* insulation. (Bottom) Two representative insulation bands reveal Setb1-sensitive insulation zones in KO neurons aligned with excess CTCF peaks, as indicated.](nihms882517f1){#F1}\n\n![Histone modification and CTCF landscapes in *Setdb1*-deficient neuronal nuclei\\\n(**a**) DiffRep counts (1kb^sw^) for H3K9me3 methylation and H3K27ac acetylation (ChIP-seq) from KO compared to WT adult cortex, for NeuN^+^ and NeuN^\u2212^. H3K9me3 \\>1.5-fold; H3K27ac \\>2-fold; FDR P\\<0.05. (**b**) Manhattan plot with linear representation of autosomes, showing localized enrichments (1MB^sw^) for H3K9me3 hypomethylation in KO. Top-scoring chromosome 18 *cPcdh* locus corresponds to site affected by loss of long-range loop bundles ([Figure 1d](#F1){ref-type=\"fig\"}). (**c**) Mouse total chromosome 18 (mm10; merged fastQ N=3 animals) H3K9me3 landscape for NeuN^+^ KO and WT, with \\~1.2 Mb (chr18:36,691,575-37,938,923) *cPcdh* locus flagged. Scale bar, 10 Mb. (**d**) CTCF motif (red) enrichment in sequences H3K9me3 hypomethylated in KO. (**e**) (Left) FACS plots showing separation of crosslinked NeuN^+^ from NeuN^\u2212^ nuclei (adult cortex) for cell-type specific CTCF ChIP-seq. (Right) Mutant NeuN^+^ showed 3059 CTCF up- and only 19 CTCF down-regulated sequences, affecting primarily inter- and intragenic DNA (\\>2-fold KO/WT NeuN^+^ nuclei, N=4 animals/group, FDR P \\<0.05. (**f**) Dramatic CTCF motif (red) enrichment among the 3059 CTCF-up sequences. (**g**) Autosomal genome Manhattan plot showing localized clustering of CTCF up sequences in KO neurons, with *cPcdh* as top ranking locus (1Mb^sw^). (**H**) (Top) *cPcdh* CTCF landscapes in KO and WT NeuN^+^ as indicated. Significantly up-regulated (KO\\>WT) promoter-bound and intergenic CTCF sequences marked separately. Scale bar, 100 kb. (Bottom) Coordinate increase of (red) CTCF and (green) H3K27ac at selected *cPcdh* promoters and intergenic DNA in KO neurons. In contrast, robust peaks at baseline, independent of genotype, at DNaseI hypersensitive HS5-1(HS5-1a+HS5-1b).](nihms882517f2){#F2}\n\n![DNA methylation profiling at the *cPcdh* locus\\\n(**a**) (Top) red tick marks for 11 *cPcdh* promoters and 2 intergenic sequences (A,B) with excess/*de novo* CTCF occupancy in KO neurons; black tick marks for HS5-1 and HS16 with robust CTCF peaks in both KO and WT (see also [Figure 2h](#F2){ref-type=\"fig\"}). (Bottom) Averaged 5mC DNA methylation levels (green-red=0--100%) of 43 amplicons representing the set of 13 regulatory sequences show in Top panel. Bis-seq data were averaged across 47 DNA samples from cortical and striatal NeuN^+^ and NeuN^\u2212^, and cerebellar homogenate. Downward arrows: 18/43 bis-seq amplicons show ^m^C5 deficit in *Setb1-*deficient neurons. 0/43 show increase (P\\<0.5--0.1/amplicon) (see [Supplementary Table 13](#SD1){ref-type=\"supplementary-material\"} for details on quantification). (**b**) Representative bis-seq example from *Pcdha8* amplicon no.2 capturing 10 CpG sites. Score cards from 50 randomly selected DNA molecules: circles black/white methylated/not methylated. (**c**) Quantification of bis-seq amplicons expressed as %methylated. \\*,\\*\\*P\\<0.05(0.01) unpaired one-tailed t-test. Each symbol represents 1 sample from 1 animal. Note methylation deficits specifically for cortical (CX) and striatal (Str) NeuN^+^ from Setdb1-deficient (k) neurons, compared to wildtype (w).](nihms882517f3){#F3}\n\n![Transcriptional dysregulation at the *cPcdh* locus\\\n(**a**) (Left) genome-wide RNAseq heatmap, blue-yellow range show average levels of expression (log 2), for transcripts with significant (FDR P\\<0.05) difference in expression of KO compared to WT PFC. (Right) Gene Ontology of differentially expressed genes (FDR P\\<0.05) highlight Setdb1-dependent regulation of *cPcdh* cell adhesion genes. (Middle) Manhattan plots for autosomal genome (mouse chromosomes 1--19), showing singular enrichment (1MB^sw^) for (Top) upregulated transcripts and (Bottom) histone hyperacetylated chromatin at chromosome 18 *cPcdh* locus, as indicated. (**b**) (Top) Whole chromosome 18 view on cortex RNA-seq, merged FastQ N=2KO (light purple) and N=2WT (dark purple). (Bottom) representative RNAseq tracks for first exon of specific *Pcdh\u03b1, Pcdh\u03b2,* and *Pcdh\u03b3* genes. Scale bar, 500bp. Notice increased expression primarily from non-C type *Pcdh* genes with stochastic expression pattern (S-type), while C-type *Pcdh* genes remain unaffected. (**c**) Top: *Pcdh* and non-*Pcdh* transcripts tested in adult cortex from mutant and transgenic rescue mice and their respective controls by qRT-PCR and *in situ* hybridization as indicated. Bottom: Transgenic rescue for representative S-type *Pcdh*, shown by ISH from middle layers of lateral cerebral cortex (scale bar, 50 \u03bcm). Box plots (1^st^/3^rd^ quartile, median, whiskers^min,max^) summarizing qRT-PCR in PFC of WT, TG (*CK-Setdb1^+^* transgenic line), KO and RC (*CK-Setdb1^+^* transgenic rescue of conditional *CK-Cre^+^, Setdb1^2flox/2lox^* mutants). N=6/group, \\*\\*\\*P\\<0.001. *Pcdha8*, t ~(WT/KO)~ = 9.59, t ~(KO/RC)~=8.33; *Pcdhb8*, t ~(WT/KO)~=11.07, t ~(KO/RC)~=10.03; *Pcdhga8*, t ~(WT/KO)~=10.07, t ~(KO/RC)~=7.39. One-way ANOVA, Bonferroni corrected. Additional ISH gene expression data are shown in [Supplementary Figures 12--15](#SD1){ref-type=\"supplementary-material\"}.](nihms882517f4){#F4}\n\n![Epigenomic editing at the cPcdh locus\\\n(**a**) *cPcdh* locus and surrounding sequences \\~2Mb of mouse chr. 18, including TADs called (TADtree) and H3K9me3 tracks for KO and WT. Notice 'shrinkage' of broadly (\\>100--200kb) stretched 'R1' and 'R2' blocks of H3K9me3-tagged chromatin in KO neurons. (**b**) Overview on cell-type specific 3C-PCR, cropped agarose gels showing specific loop products for c*Pcdh* and *B2m* control. No lig=3C without DNA ligase, L=100bp DNA ladder. Dot graphs summarizing 3C-PCR (mean\u00b1S.E.M.; 1dot=1animal) *cPcdh* loop1,2,3 as indicated. All data normalized to *B2m* 3C, N~(Loop1)~=3/group, \\*P~(Loop1)~=0.05, Mann Whitney, one-tailed; N~(Loop2)~=4/group, \\*P~(Loop2)~=0.014,Mann Whitney, two-tailed. Loop defects in KO include A/R1(de novo CTCF peak A in R1)-HS16 and A/R1-B/R2 (de novo CTCF peak B in R2). In contrast, shorter-range *Pcdha8* promoter-HS5 enhancer loop is maintained in KO neurons. Complete gels shown in [Supplementary Figure 16c](#SD1){ref-type=\"supplementary-material\"}. (**c**) Summary presentation of 3C-PCR. (**d**) dCas9-SunTag superactivation of HS16 *cPcdh* enhancer with U6-sgRNA cassette upstream of *CK-dCas9-10xGCN4^epitope^-BFP* cassette, and *CK-svFv-sGFP-VP64* cassette on separate vector. Representative FACS sort shows dually labeled BFP^+^GFP^+^ NG108 cells. NC=negative control (**e**) RT-PCR quantification (mean\u00b1S.E.M.; 1dot=1cell culture or animal) of *Pcdha3*, *Pcdha8*, *Pcdhb16*, *Pcdhgb2* and *Pcdhgb8* transcripts (black arrows in panel d mark genomic positions), normalized to *Gapdh* RNA. (Top) BFP^+^GFP^+^ NG108 cells with (HS16/VP64) and without (VP64) sgRNA^HS16^ cassette. (Bottom) adult KO and WT PFC. N=4 VP64/3 hs16/vp64 (NG108 cells), \\*P~(~*~Pcdha8~*~)~=0.0268, \\*P~(~*~Pcdha11~*~)~=0.0437, \\*P~(~*~Pcdhgb8~*~)~=0.0126, unpaired t test, one-tailed; N=6/group (mice), \\*\\*P~(~*~Pcdha3~*~)~=0.002, \\*\\*P~(~*~Pcdha8~*~)~=0.002, \\*P~(~*~Pcdha11~*~)~=0.026, \\*\\*P~(~*~Pcdhgb8~*~)~=0.0022, Mann Whitney, two-tailed. See also [Supplementary Figure 16A](#SD1){ref-type=\"supplementary-material\"} for additional 3C-PCR loop quantifications.](nihms882517f5){#F5}\n\n![Regulatory mechanisms at human and mouse TAD*^cPCDH^*\\\n(**a**) (Top) Neuronal *in situ Hi-C* interaction matrices, and H3K9me3 landscapes, for \\~2Mb of mouse and human *cPCDH*, including superTAD spanning across \u03b1,\u03b2,\u03b3 clusters. (Bottom left) Setdb1 peaks in mouse embryonic stem cell and lymphocytes match to de *novo* CTCF peak in *Setdb1* KO neurons. (Bottom right) 'PGC', Psychiatric Genomics Consortium risk haplotype *chr5:140,023,664-140,222,664* with lead polymorphism rs111896713 matching to Setdb1, KAP1 and ZNF143 peaks. Note epigenomic similarities of 'R1' (mouse) and 'PGC' (human). (**b**) Neural progenitor cell (NPC) differentiation into neurons and astrocytes, with phenotypic markers as indicated. Scale bar, NPC (neuron/astrocyte) 100 (50) \u03bcm. Conformations for three representative 40kb bins from 200kb 'PGC' haplotype, with bin harboring the index polymorphism ('PGC-3') showing dramatically increased *cPCDH* contact. (**c**) Dot graphs show *cPCDH* gene expression in epigenomically edited NPC, with *PCDHGB6* (but not *PCDHGA8*) transcript decreased by sgRNA-guided dCas9-KRAB in 3/3 experiments. dCas9-VP64 elicits increased expression of a subset of *Protocadherin* transcripts. Scr, scrambled control. All data normalized to 18srRNA, shown as fold change. N=6--9/group, P \\< 0.05, Mann Whitney, two-tailed. (**d**) ZNF-specific motif enrichments in CTCF-up sequences. Dot graphs (1dot/cell culture) summarize expression of specific *Pcdh\u03b1, \u03b2 and \u03b3* genes after *shRNA*-induced Zfp143 knock-down in NG108 neuroblastoma cells. Unpaired t test, two-tailed, N=3 per treatment. \\*P\\<0.05 (**e**) Schematic summary of TAD*^cPcdh^* epigenomic architectures in WT and KO neurons. Loss of repressive long-range contacts in KO shifts the balance towards facilitative shorter range promoter-enhancer loopings.](nihms882517f6){#F6}\n\n[^1]: equal contribution\n\n[^2]: Dr. Hartley's current affiliation is VL39 Inc, Cambridge MA 02142.\n"}
+{"text": "Background {#Sec1}\n==========\n\nStage IV, the point in tumor progression in which cancer spreads beyond the primary site and regional lymph nodes and is found in other organs, is the cancer stage that most often leads to patient mortality \\[[@CR1]\\]. The tumor's microenvironment plays a critical role in tumor growth and the development of metastasis where the interaction between tumor cells and the associated stroma and cellular components modulates the tumor's progression and patient prognosis. Recently, the acellular 4D lung model has successfully mimicked the development of metastasis \\[[@CR2]\\]. It is named the 4D model because of its perfusion of tumor nodules that allows it to change over time and grow in the 3D space. Findings from the 4D model suggest that the only component of tumor microenvironment that is important to show tumor progression is an intact natural matrix \\[[@CR2]\\].\n\nThe acellular 4D lung model is created by removing all of the cells from a rat heart and lung block \\[[@CR3], [@CR4]\\]. This natural lung matrix maintains its three-dimensional architecture, including perfusable vascular beds and preserved airways. The matrix is composed of collagen, proteoglycans, and elastic fibers that preserve the architecture of airways and capillaries. A unique feature of the matrix is that this composition is preserved among species in the distal airways \\[[@CR5]\\]. Furthermore, the basement membranes of the alveolar septa are preserved after decellularization in this model \\[[@CR3]\\]. The acellular 4D lung model shows that when tumor cells are placed into the trachea, they form perusable nodules in the lung matrix \\[[@CR6]\\]. Moreover, the model allows tumor cells to secrete proteins that are more similar those found in lung cancer patients than the same tumor cells grown on a petri dish \\[[@CR7]\\]. The acellular 4D lung model mimics metastasis, with the placement of all tumor cells in the left lung lobes and perfusion of the model in the bioreactor through the pulmonary artery. In order for the tumor cells to enter the right lung, the cells would need to leave the epithelial space in the left side, enter the vasculature, and enter the other epithelial space on the right side. Over time, this process occurred as metastatic lesions formed in the right lung and grew over time in the 4D model \\[[@CR2]\\]. There are significant differences in the spatial organization of the tumor cells where the primary tumor grew in a pattern along the airway and the metastatic lesion formed in a distribution that is consistent with cancer distributed along the vasculature. The model's unique vascular channel allowed dead cells as well as live circulating tumor cells (CTC) to enter the vasculature. The CTC showed differences in behavior and gene expression compared to those cells initially placed in the model. The CTC took longer to attach to the petri dish than the parental cells placed in the model and they stayed alive in supernatant with decreased expression of integrin beta 4 (ITGB4) \\[[@CR8]\\]. In addition, CTCs were resistant to chemotherapy \\[[@CR9]\\]. There is no difference in the number of live CTC from the 4D model when they are placed in the petri dish, with or without Cisplatin, while the same dose for the parental cells (2D) placed in the model showed a significant reduction of live cells \\[[@CR9]\\]. Previous studies also show that the CTC form metastatic lesions in the 4D model \\[[@CR2]\\].\n\nA major drawback of this acellular 4D model has been the lack of cellular components that are found in a patient's tumor microenvironment. These studies suggest that the natural matrix architecture is the only component necessary for complex perfusable nodules to form, the creation of CTC, and ultimately metastatic lesion formation. However, one could argue that this phenomenon is simply due to the artificial creation of an acellular environment. Thus, in this study, we show that the ex vivo 4D lung model can mimic the metastatic process in a normal cellular environment. We postulate that non-small cell and small cell lung cancer cell lines as well as breast cancer cell lines will grow in the model and form a primary tumor, CTC, and metastatic lesions.\n\nMethods {#Sec2}\n=======\n\nAll of the animal experiments were carried out in accordance with all applicable laws, regulations, guidelines, and policies governing the use of laboratory animal in research. The Institutional Animal Care and Use Committee (IUCAC) at the Houston Methodist Research Institute approved the protocols for animal experiments.\n\nRat lung isolation {#Sec3}\n------------------\n\nWe harvested the lung--heart block from 4 to 6\u00a0week old male Sprague-Dawley rats as previously described \\[[@CR6]\\]. Briefly, we euthanized Sprague-Dawley rats with ketamine (100\u00a0mg/mL) and xylazine (10\u00a0mg/mL) and performed bilateral thoracotomy to open the thoracic cavity. We injected 2\u00a0mL of heparin (1000\u00a0units/mL, Sagent Pharmaceuticals, IL, USA) into the right ventricle of heart, removed the rib cage and injected 20\u00a0mL of heparinized Phosphate Buffered Saline (12.5 Units/mL) in the right ventricle after placing an 18-gauge needle (McMaster Carr, USA) in the left ventricle as a vent. The superior vena cava and inferior vena cava were cut and the lungs were flushed again with 20\u00a0mL of heparinized PBS. Next, we divided the trachea at the level of the thyroid, the branches of the aorta at the arch, and the descending aorta at the level of the hemiazygos vein. The heart-lung block was then separated from the esophagus and the rest of the rat body. We performed ventriculotomy to expose the right and left ventricles and placed a custom-made prefilled 18-gauge stainless steel needle (McMaster Carr, USA) through the right ventricle into the main pulmonary artery. This was secured with a 2--0 silk tie (Ethicon, San Angelo, TX, USA). We also placed a female luer bulkhead (Cole-Parmer, IL, USA) in the left ventricle and secured it with a 2--0 silk tie. We flushed the pulmonary artery cannula with heparinized PBS and placed it in a container containing heparinized PBS.\n\nCell culture {#Sec4}\n------------\n\nHuman cancer cell lines A549, H1299, H460, H69, H446, SHP-77, MCF7, and MDAMB231 were obtained from ATCC (Manassas, VA, USA). These cell lines were grown in BD T175 cell culture flasks in RPMI 1640 medium (Hyclone, UT, USA) supplemented with 10% fetal bovine serum (Gibco, USA) and antibiotics (100\u00a0IU/mL penicillin, 100\u00a0mg/mL streptomycin, and 0.25\u00a0mg/mL amphotericin; Gibco, USA) at 37\u00b0C in 5% CO2. Once the cells were 85% confluent, they were washed with PBS and subjected to trypsinization using 0.25% trypsin (Gibco, USA) to collect the cells from flasks. The cells were washed with medium and finally suspended in 50\u00a0mL of complete culture medium. Approximately 15 million cells were used to seed the lung matrix.\n\nBioreactor for the cellular 4D lung {#Sec5}\n-----------------------------------\n\nA simplified small closed-system bioreactor was set up in an incubator for the lung cell culture (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). We used a custom-designed 500-mL glass bottle with three holes in the cap fitted with a female luer thread-style panel (Cole-Parmer, USA): one for the pulmonary artery cannula, one for the trachea cannula, and one for the circulation of medium from the bottle. To obtain a controlled flow through the pulmonary artery, the cap was connected to a 3-way stopcock (Smith Medical, Dublin, OH, USA). The bottle was filled with 200\u00a0mL of complete medium that was circulated through the oxygenator tubing to prevent air bubbles.Fig. 1Cellular ex vivo 4D lung model in a bioreactor without any tumor cells (control). Bioreactor showing connection from the lung to oxygenator and pump in cell culture incubator (**a**). The lung after 15\u00a0days in the bioreactor is intact (**b**). H&E staining shows intact alveoli, bronchus and vasculature (**c**) with some apoptotic cells. CD34 staining shows the endothelial lining in intact vasculature (**e**) and Movat pentachrome staining represents the components of extracellular matrix (**f**)\n\nBefore seeding the human lung cancer cells into the lung matrix, the trachea was cannulated using an 18-gauge needle, and the scaffold was fixed to the bioreactor bottle in a hanging position. To modify it for the metastasis model, we tied the right main bronchus with a silk tie that was left there for the entire experiment and placed it in the bioreactor. The cells diluted in 50\u00a0mL complete media were seeded into the left lung lobes through the tracheal cannula via a sterile syringe fed by gravity. We perfused the scaffold at a flow rate of 6\u00a0mL/min. The culture medium in the bottle was changed every day to ensure the nutrients were optimal for cell growth, and the CTC were spun down and counted. We grew the cells on the matrix for 10--15\u00a0days. The lung matrix was carefully removed from the bioreactor bottle, maintaining sterile conditions, and a lobectomy performed under a culture hood by tying the anatomic lobe with a 2--0 silk tie and resecting it on different days.\n\nImpact of cellular lung on tumor growth on 2D {#Sec6}\n---------------------------------------------\n\nWe plated 500,000 GFP labeled of A549, H460 and H1299 cells in 6-well plate, with different lung lobes from cellular and acellular rat lungs in cell culture incubator. After 48\u00a0h, cells were trypsinized and GFP positive cells were counted using fluorescence activated cell sorting (FACS) with same gating parameters (HMRI NIR Ariall, USA). All the cells were plated from same stock of cells on same day to avoid any bias with the GFP intensity. We determined the percentage of cells that that GFP+ cells in each sample. We compared the groups using student t-test and used *p*\u2009\\<\u20090.05 as significant.\n\nHistology and immunohistochemistry {#Sec7}\n----------------------------------\n\nAfter lobectomy, the lung tissues were placed in 10% paraformaldehyde and shipped to the Pathology Core Laboratory at The Methodist Hospital Research Institute for further processing. Briefly, the tissues were fixed in 10% formalin overnight, processed, and embedded in paraffin. Hematoxylin and eosin (H&E) staining and immunohistochemistry were performed for human mitochondria to identify human cancer regions. The embedded tissues were cut into 4-\u03bcm slides and dewaxed; antigen retrieval was performed with antigen-unmasking solution (H-3300; Vector Laboratories, Burlingame, CA) in a steamer for 20\u00a0min. Slides were cooled for 20\u00a0min at room temperature, washed in PBS, and stained with H&E, Movat Pentachrome (American MasterTech Scientific, CA, USA), Human Anti-mitochondria (Abcam, MA, USA), CD34 and other markers following the standard protocol \\[[@CR10]\\]. Expert board-certified pathologists examined stained slides, and images were captured using a microscope (EVOS, Fisher Scientific, USA). The metastatic lesions per high power field were determined by averaging the number of human mitochondria positive tumor cells per high power field (40X) of 10 areas.\n\nResults {#Sec8}\n=======\n\nCellular rat lung model {#Sec9}\n-----------------------\n\nThe cellular ex vivo lung was intact in the bioreactor for up to 15\u00a0days (Fig. [1b](#Fig1){ref-type=\"fig\"}). The control tissue's histology shows a marginally intact matrix and lung with some apoptotic and non-viable cells (Fig. [1c](#Fig1){ref-type=\"fig\"} and [d](#Fig1){ref-type=\"fig\"}). The lung tissue showed the lung with bronchovascular bundles, alveolar tissue and overlying pleural tissue. The peribronchiolar tissue areas have collections of histiocytes resembling non-necrotizing granulomas (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"} and [d](#Fig1){ref-type=\"fig\"}). The alveoli are degenerating and many non-viable cells are seen along with scattered intra-alveolar macrophages. IHC of human mitochondria showed no positive staining, as there were no human cells seeded. Furthermore, staining the lung with Movat pentachrome highlights the matrix architecture showing the presence of collagen, proteoglycans and elastin (Fig. [1f](#Fig1){ref-type=\"fig\"}). The immunohistochemistry of CD34 showed its presence in vessels (Fig. [1e](#Fig1){ref-type=\"fig\"}).\n\nPrimary tumor in the cellular model {#Sec10}\n-----------------------------------\n\nAfter 10--15\u00a0days of incubation, in the presence of cellular components, human NSCLC, SCLC, and breast cancer cells formed microscopic tumor nodules on a native rat lung though no gross nodules were visualized (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}, [e](#Fig2){ref-type=\"fig\"} and [i](#Fig2){ref-type=\"fig\"}). These tumor cells showed a distinctive morphology unique to cell type. All of the NSCLC cells (A549, H1299, and H460) formed the solid pattern of a primary tumor in a bronchocentric distribution, but they differ in histology based on subtype. A549, an adenocarcinoma cell line, formed a focal acinar pattern and was compatible with human lung adenocarcinoma (Fig. [2b](#Fig2){ref-type=\"fig\"}, [c](#Fig2){ref-type=\"fig\"} and [d](#Fig2){ref-type=\"fig\"}). The tumor cells were enlarged epithelioid cells with enlarged nuclei containing prominent nucleoli and scattered atypical mitotic cells (Fig. [2d](#Fig2){ref-type=\"fig\"}). The Anti-Human mitochondrial stain highlights the cytoplasmic stain in tumor cells, but not the peribroncholar histocyte aggregates (Fig. [2b](#Fig2){ref-type=\"fig\"}). H1299 cell lines also formed primary tumors with a solid growth pattern (Fig. [2f](#Fig2){ref-type=\"fig\"}, [g](#Fig2){ref-type=\"fig\"} and [h](#Fig2){ref-type=\"fig\"}) and resemble poorly differentiated NSCLC. The cells resembled enlarged epithelioid, large central nuclei with large centrally placed nucleoli. Some of the nuclei are vesicular with scattered atypical mitosis, but this is not a definitive characteristic of squamous or adenocarcinoma (Fig. [2g](#Fig2){ref-type=\"fig\"} and [h](#Fig2){ref-type=\"fig\"}). H460 cell lines also formed tumors on the cellular model with brocho/bronchiolar centric distribution (Fig. [2j](#Fig2){ref-type=\"fig\"}, [k](#Fig2){ref-type=\"fig\"} and [l](#Fig2){ref-type=\"fig\"}). These tumor cells showed a solid pattern of tumor growth with vague glandular formation and a similar pattern to adenocarcinoma (Fig. [2k](#Fig2){ref-type=\"fig\"} and [l](#Fig2){ref-type=\"fig\"}). Tumor cells appear similar to A549 cells in the cellular model.Fig. 2Non small cell lung cancer (NSCLC) cells on the cellular 4D lung model. Primary tumors were formed in a bronchocentric fashion on the cellular lung upon A549 (**a**--**d**), H1299 (**e**--**h**) and H460 (**i**--**l**) cell seeding though the trachea. Human mitochondrial IHC staining clearly shows the presence of a tumor formed by human lung cancer cells in a rat lung after 12\u00a0days of culture in a bioreactor. Low and high power H&E staining indicates the presence of a microscopic tumor with distinctive histology based on type of cell placed in the model. A549 cells formed a focal acinar pattern with enlarged nuclei and scattered atypical mitotic cells (**c** and **d**). H1299 cells formed a solid tumor and resemble poorly differentiated cells with scattered atypical mitosis (**g** and **h**). H460 cells formed a solid tumor with vague glandular formation (**k** and **l**)\n\nIn addition to NSCLC cell lines, we placed small cell lung carcinoma (SCLC) cell lines (H69, H446, SHP77) on the 4D cellular lung model (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}, [e](#Fig3){ref-type=\"fig\"} and [i](#Fig3){ref-type=\"fig\"}). H69 cell lines showed the primary tumor in the broncho/bronchiolar centric space (Fig. [3b](#Fig3){ref-type=\"fig\"}, [c](#Fig3){ref-type=\"fig\"} and [d](#Fig3){ref-type=\"fig\"}) with a peculiar pattern of high-grade basaloid malignancy, relatively scant cytoplasm, dark-hyperchromatic nucei, lack of nucleoli, abundant mitosis, and karyorrhectic debris in the background. The tumor was arranged in a nest and there was a vague rosette/acinar appearance (Fig. [3c](#Fig3){ref-type=\"fig\"} and [d](#Fig3){ref-type=\"fig\"}). H446 cells grew on the cellular model and formed a collection of tumors in a bronchiolar centric distribution and intra-alveolar (Fig. [3f](#Fig3){ref-type=\"fig\"}, [g](#Fig3){ref-type=\"fig\"} and [h](#Fig3){ref-type=\"fig\"}). Cytomorphologic features were typical of small cell carcinoma similar to H69 features. SHP77 cells also formed the primary tumor on the cellular model and appeared more like poorly differentiated carcinoma within alveolar spaces (Fig. [3j](#Fig3){ref-type=\"fig\"}, [k](#Fig3){ref-type=\"fig\"} and [l](#Fig3){ref-type=\"fig\"}). The tumor cells have large nuclei with some cells having nucleoli. Numerous mitotic cells with associated karyorrhetic debris were also visible (Fig. [3k](#Fig3){ref-type=\"fig\"} and [l](#Fig3){ref-type=\"fig\"}).Fig. 3Small cell cancer (SCLC) cells on the cellular 4D lung model. Primary tumors formed in a bronchocentric fashion on the cellular lung upon H69 (**a**--**d**), H446 (**e**--**h**) and SHP77 (**i**--**l**) cell seeding though no distinct nodules appeared. Human mitochondrial IHC staining clearly shows the presence of a tumor formed by human lung cancer cells in a rat lung after 10\u00a0days of culture in a bioreactor. Low and high power H&E staining indicates the presence of microscopic tumor with distinctive histology based on the cell type seeded. H69 cells formed tumors similar to high-grade basaloid malignancy with vague rosette/acinar appearance (**c** and **d**). H446 cells formed a collection of tumors, typical of small cell carcinoma (**g** and **h**). SHP77 cells formed a tumor more like a poorly differentiated carcinoma within alveolar spaces (**k** and **l**)\n\nBreast cancer cell lines (MCF7 and MDAMB231) also colonized on the cellular 4D lung model and formed microscopic tumor nodules (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"} and [e](#Fig4){ref-type=\"fig\"}). An ER/PR positive MCF7 cell formed a primary tumor in the left lobes that resembled a solid gland carcinoma in a broncho-bronchiolar centric distribution (Fig. [4b](#Fig4){ref-type=\"fig\"}, [c](#Fig4){ref-type=\"fig\"} and [d](#Fig4){ref-type=\"fig\"}). The tumor cells forming the gland were medium to large epithelioid cells with large nuclei, some vesicular and some with prominent nucleoli (Fig. [4c](#Fig4){ref-type=\"fig\"} and [d](#Fig4){ref-type=\"fig\"}). The morphologic features were similar to metastatic high-grade breast carcinoma (ductal). MDAMB231 cells grown on the cellular model formed large confluent primary tumors with poorly differentiated malignant carcinoma (Fig. [4f](#Fig4){ref-type=\"fig\"}, [g](#Fig4){ref-type=\"fig\"} and [h](#Fig4){ref-type=\"fig\"}). There were loosely cohesive cells with rare acinar formation. Most of the malignant cells have spindle cell morphology with atypical mitosis, looking like metaplastic carcinoma (Fig. [4g](#Fig4){ref-type=\"fig\"} and [h](#Fig4){ref-type=\"fig\"}).Fig. 4Breast cancer cell lines on the cellular 4D lung model. Primary tumors were formed in a bronchocentric fashion on cellular lung upon MCF7 (**a**) and MDAMB231 (**e**) cell seeding though no distinct nodules appeared. Human mitochondrial IHC staining clearly shows the presence of a tumor formed by human breast cancer cells in a rat lung after 15\u00a0days of culture in a bioreactor. Low and high power H&E staining indicates the presence of a microscopic tumor with distinctive histology based of the cell type seeded. MCF7 cells formed solid gland carcinoma with large epithelioid cells and prominent nucleoli (**c** and **d**). MDAM231 cells formed large confluent primary tumors with poorly differentiated malignant carcinoma having spindle cell morphology (**g** and **h**)\n\nCirculating tumor cells {#Sec11}\n-----------------------\n\nWe seeded the cellular model with GFP tagged tumor cells (H1299 and H460). We found GFP positive cells in the circulation and plated them in 96 well plates. These GFP tagged circulatory cells floated and survived in the media for a longer time compared to respective 2D cells. We visualized the viable GFP tagged cells attached to the surface and dividing, under a fluorescent microscope after 7\u00a0days (Fig.\u00a0[5i](#Fig5){ref-type=\"fig\"} and [j](#Fig5){ref-type=\"fig\"}).Fig. 5Circulatory tumor cells. We seeded GFP labeled H1299 or H460 tumor cells on the cellular 4D lung model, collected CTC and plated in 96 well plates. Fluorescent microscopy shows CTC from H1299 (**a**) or H460 (**b**) attached to the plate and actively undergoing mitosis\n\nMetastasis to contralateral lung {#Sec12}\n--------------------------------\n\nThe NSCLC, SCLC and breast cancer cell lines formed metastatic lesions that were positive for human mitochondrial cells in the contralateral lung in a vascular distribution (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}). Unlike the primary tumor, where most of the tumors grew in bronchial distribution, the metastatic lesions were uniformly distributed in the lung. However, there were differences in the number of cells per HPF based on the cell type. Among NSCLC cells, A549 formed the least amount of metastatic lesions (0.9\u2009\u00b1\u20090.35 cells per HPF), while H1299, a metastatic cell line, formed the maximum number of metastatic lesions (3.6\u2009\u00b1\u20091.03 cells per HPF). H460 cells showed 1.6\u2009\u00b1\u20090.88 cells per HPF (Fig. [6a](#Fig6){ref-type=\"fig\"}--[d](#Fig6){ref-type=\"fig\"}). Among SCLC cell lines, H69 formed the highest number of metastatic lesions (12.9\u2009\u00b1\u20093.49 cells per HPF) as compared to SHP-77 and H446 (4.4\u2009\u00b1\u20092.09 cells per HPF and 2.2\u2009\u00b1\u20090.57 cells per HPF) (Fig. [6e](#Fig6){ref-type=\"fig\"}--[h](#Fig6){ref-type=\"fig\"}). MDAMB231, a metastatic breast cancer cell line, formed more metastatic lesions than MCF7 cells (3.2\u2009\u00b1\u20091.15 cells per HPF vs. 17.4\u2009\u00b1\u20095.18 cells per HPF) (Fig. [6i](#Fig6){ref-type=\"fig\"}--[k](#Fig6){ref-type=\"fig\"}).Fig. 6Metastatic lesion formation. Among NSCLC, H1299, a metastatic p53 mutant cell formed significantly more metastatic lesions as compared to A549 cells (**a**; **b** vs **d**) per high power field (HPF). H460 cells also formed metastatic lesions (**a** and **c**). Among SCLC cell lines, H69 formed more metastatic lesions than H446 and SHP-77 (**e**--**h**). In breast cancer cells, MDAMB231, a triple negative cell line, formed significantly more metastatic lesions than MCF-7 cells (**i**--**k**)\n\nImpact of cellular lung on tumor growth {#Sec13}\n---------------------------------------\n\nCellular lung significantly decreased cell growth on 2D of H1299 (*p*\u2009=\u20090.01, Fig.\u00a0[7a](#Fig7){ref-type=\"fig\"}), H460 (*p*\u2009=\u20090.03, Fig. [7b](#Fig7){ref-type=\"fig\"}) and A549 (*p*\u2009=\u20090.0002, Fig. [7c](#Fig7){ref-type=\"fig\"}) compared to control. In addition, there was significantly lower number of tumor cells in the presence of cellular lung compared to acellular lung for H1299 (*p*\u2009=\u20090.01) and A549 (*p*\u2009=\u20090.0009). There were significantly less A549 cells with acellular lung compared to control (*p*\u2009=\u20090.002).Fig. 7Impact of cellular matrix on tumor growth. We plated GFP labeled H1299 (**a**), H460 (**b**) and A549 (**c**) on a petri dish alone (control) or with acellular matrix (acellular) or cellular matrix (cellular). There was significant decrease in tumor growth with cellular matrix compared to control for all three cell lines\n\nDiscussion {#Sec14}\n==========\n\nTumor models play an important role in understanding tumor progression, metastasis, and cancer therapeutics. To date, there are several 2D/3D in vitro, in vivo and ex vivo cancer models in use, but each one has considerable shortcomings to mimic the human cancer progression \\[[@CR11], [@CR12]\\]. We previously developed an acellular ex vivo 4D lung model for cancer metastasis \\[[@CR2]\\]. The model's main criticism was that it lacked the normal lung tissue component, and thus the metastasis seen in the acellular ex vivo 4D model could be an artifact of not having cellular components. Thus, we pursued the creation of the cellular model\u00a0(Additional file [1](#MOESM1){ref-type=\"media\"}). We discovered that like the acellular model, the cellular 4D model can mimic tumor progression in 2\u00a0weeks and allows for the isolation of a high number of CTC that are the main driver for metastasis. In this study, we successfully grew both NSCLC and SCLC as well as breast cancer cells in the model with all of them forming metastatic lesions.\n\nThe cellular lung matrix model is a complex system that supports the growth of human tumor cells in a natural lung microenvironment. The cellular 4D lung model could provide a better model for cancer studies than existing 3D/2D and acellular lung models due to improved cell-cell interactions, cell-ECM interactions and presence of cell populations and a structure that resembles the in vivo environment. Similar to the acellular 4D lung model, the control native lung (devoid of tumor cells) maintains its matrix architecture and leads to tumor development with nutrient perfusion through the vasculature. The major difference lies in the presence of lung cellular components, which the acellular 4D lung model lacks. In both the acellular and cellular model, the tumor grows in an airway centric fashion due to seeding tumor cells through the trachea. Most of the NSCLC tumor cells showed similar histology and tumor cell morphology in the acellular and cellular lung model. The main difference between the cellular and acellular model was the primary tumor volume. We often visualized gross tumor nodules in the acellular model after 2\u00a0days while they were not readily visible on the cellular model. Furthermore, on pathologic examination, the nodules were much smaller in the cellular model than the acellular model. We also observed this difference when tumor cells were co-cultured with cellular lung on a petri dish (2D). Several possibilities might account for this difference. One issue is the limited nutrients for the cancer cells in the cellular model. We placed a similar amount of complete media in the cellular and acellular model but the cellular model has a normal cellular component that leaves fewer nutrients for the primary tumor cells. Another possibility is that the normal lung environment may play a tumor suppressive role. We have shown that normal lung fibroblasts can inhibit tumor progression \\[[@CR10]\\]. Thus, it is possible that cancer coordinates with intercellular interactions that are present in normal tissues and disrupt the normalizing cues from the microenvironment, and in turn, the microenvironment evolves to accommodate the growing tumor \\[[@CR13]--[@CR15]\\].\n\nCompared to the isolation of CTC from in vivo models \\[[@CR16]\\], the process of isolating the CTC is simpler with the ability to tag tumor cells with GFP and sort them with FACS analysis. The GFP-tagged CTC were viable when they were placed on a petri dish similar to the CTC isolated from the acellular model. Moreover, like cells isolated in the acellular model, the CTC were in the supernatant for several days prior to attaching to the petri dish and showed a different phenotype than the parental cells. This feature supports the concept that these cells are a unique phase of tumor development.\n\nAmong the NSCLC, the presence of the cellular component impacts the efficacy of metastatic lesion formation. In the acellular model, despite forming fewer CTC, H460 cells tended to form more metastatic lesions compared to H1299 \\[[@CR2]\\] but in the cellular model, there were more metastasis with H1299 compared to H460. In both models, the A549 cells have the least amount of metastatic lesion formation. This deviating behavior may be due to the difference in kinetics of metastatic tumor formation between the models with the presence of an intact tumor microenvironment.\n\nThis model also allows for the rapid growth of small cell lung cancer cell lines to form nodules and eventual metastasis. Most of these SCLC cell lines grow slowly and float (less adherent) in vitro which makes them difficult to study \\[[@CR17]\\]. For most in vitro studies, these cells have to be grown in suspension, which does not mimic the growth in patients with small cell lung cancer. When we placed these cells in our model, they formed a primary tumor, CTC and metastatic lesions within 10--14\u00a0days. This rapid growth and tumor progression can provide a much-needed new model in understanding the biology of small cell lung cancer.\n\nFurthermore, we were able to show growth of breast cancer cell lines in this model. Breast cancer is the most common cancer in women and the lung is one of the sites of metastasis. We have successfully grown breast cancer cell lines in the acellular model \\[[@CR18]\\] and now we have been able to grow them in the cellular model. Like the NSCLC and SCLC cell lines, the breast cancer cell lines form a primary tumor and metastatic lesions. The MDAMB231, triple negative cell line showed more metastatic cells in contralateral lung lobes than MCF7 cells. This correlates with the aggressiveness of triple negative breast cancer compared to ER\u2009+\u2009PR+ breast cancer in patients.\n\nConclusion {#Sec15}\n==========\n\nThe cellular lung model is an advanced cancer model that mimics the crucial phases of tumor growth---primary tumor, CTC formation and metastatic lesions---in the presence of normal lung tissue. This supports the concept that the natural matrix scaffold is the only component necessary to mimic metastasis and that cellular components modulate the process of metastasis. This ex vivo model can be used to study the complex mechanism of tumor metastasis.\n\nAdditional file\n===============\n\n {#Sec16}\n\nAdditional file 1:Schematic diagram illustrating the major steps in ex vivo 4D Lung model creation. (JPG 70 kb)\n\n3D\n\n:   Three dimensional\n\n4D\n\n:   Four dimensional\n\nATCC\n\n:   American Type Culture Collection\n\nCTC\n\n:   Circulating tumor cells\n\nEx vivo\n\n:   Out of the living\n\nGFP\n\n:   Green fluorescent protein\n\nH&E\n\n:   Haematoxylin and eosin\n\nIACUC\n\n:   Institutional animal care and use committee\n\nIHC\n\n:   Immunohistochemistry\n\nITGB4\n\n:   Integrin subunit beta 4\n\nIU\n\n:   International unit\n\nmg\n\n:   Milligram\n\nmL\n\n:   Milliliter\n\nNSCLC\n\n:   Non-small cell lung cancer\n\nPBS\n\n:   Phosphate-buffered saline\n\nSCLC\n\n:   Small cell lung cancer\n\n**Electronic supplementary material**\n\nThe online version of this article (10.1186/s12885-018-4358-x) contains supplementary material, which is available to authorized users.\n\nWe thank Anna Saikin for editing the language of the manuscript.\n\nFunding {#FPar1}\n=======\n\nDr. Kim received grant support from the Second John W. Kirklin Research Scholarship, American Association for Thoracic Surgery, Graham Research Foundation, Houston Methodist Specialty Physician Group Grant and Michael M. and Joann H. Cone Research Award. The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nAll materials described in the manuscript, including all relevant raw data will be freely available to any scientist wishing to use them for non-commercial purposes.\n\nDKM, RM, KAP, MPK made substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; DKM, RM, KAP, MPK has been involved in drafting the manuscript or revising it critically for important intellectual content; DKM, RM, KAP, MPK has given final approval of the version to be published. Each author should have participated sufficiently in the work to take public responsibility for appropriate portions of the content; and DKM, RM, KAP, MPK agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. DKM, RM, KAP, MPK read and approved the final manuscript.\n\nEthics approval {#FPar3}\n===============\n\nAll of the animal experiments were carried out in accordance with all applicable laws, regulations, guidelines, and policies governing the use of laboratory animal in research. The Institutional Animal Care and Use Committee (IUCAC) at the Houston Methodist Research Institute approved the protocols for animal experiments (AUP-0716-0037).\n\nCompeting interests {#FPar4}\n===================\n\nThe authors declare that they have no competing interests in relation to this work.\n\nPublisher's Note {#FPar5}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nInfectious diseases of the lower respiratory tract are exacerbated by the lack of sufficient treatment options, which is underlined by the 3.2 million reports of death in 2015 (World Health Organization; WHO). Although less frequently observed in wealthy countries, the increasing number of nosocomial infections constitutes a substantial public health problem worldwide^[@CR1]^. Treatment of lower respiratory tract infections becomes particularly problematic, if the invasive microorganisms are resistant to standard or reserve antibiotic therapy as observed with increasing frequency for infections triggered by *Pseudomonas aeruginosa* (*P. aeruginosa*), one of the most common pathogens responsible for severe nosocomial infections^[@CR2]^. In fact, the most recent global priority list of antibiotic-resistant bacteria to guide research, discovery, and development of new antibiotics from the WHO, states carbapenem-resistant *P. aeruginosa* as one of the three most critical pathogens for which new treatment options are urgently required^[@CR3]^.\n\nOne promising alternative to antibiotic therapy might be a cell-based immunotherapy approach applying phagocytes to enhance pulmonary immunity. As of yet, generating the therapeutically required amount of immune cells from peripheral blood or other sources remains challenging. In contrast to somatic cells, human-induced pluripotent stem cells (hiPSC), with their unlimited potential for proliferation and differentiation, may---in principle---enable this therapeutic scenario. In this line, hematopoietic differentiation of human iPSC has been proven feasible^[@CR4]--[@CR6]^ and thus, has been proposed as a promising strategy for future cell-based treatment approaches. However, clinical translation of hiPSC-derived hematopoiesis remains hampered by (i) insufficient knowledge about in vivo functionality and (ii) lack of therapeutically required quantities of effector cells.\n\nConsidering phagocytes, and especially alveolar macrophages as critical regulators in the maintenance of lung homeostasis and pulmonary immunity^[@CR7]--[@CR9]^, here we evaluate the therapeutic potential of iPSC-derived macrophages (iPSC-Mac) for the treatment of pulmonary infections caused by *P. aeruginosa*. We show that human iPSC can be efficiently differentiated into primitive macrophages in scalable suspension culture, and that the differentiation and harvesting process is transferable to industry-compatible stirred bioreactor systems. More importantly, we demonstrate that pulmonary transplantation of iPSC-Mac can prevent the onset of acute respiratory *P. aeruginosa* infection and rescue mice from established pulmonary infections and severe respiratory insufficiency.\n\nResults {#Sec2}\n=======\n\nMass production of human macrophages in stirred bioreactors {#Sec3}\n-----------------------------------------------------------\n\nAlthough the generation of different mature hematopoietic cell types from PSC has been proven successful using classical two-dimensional (2D) differentiation cultures^[@CR4],[@CR10]--[@CR13]^, these systems do not allow for the generation of iPSC-derived cells in clinically relevant quantities. Thus, we developed a suspension-based (3D), continuous (4D) hematopoietic differentiation protocol, suitable for process upscaling in industry-compatible stirred tank bioreactors^[@CR14],[@CR15]^. Using a well-characterized hiPSC line (hCD34iPSC16)^[@CR16]^, we induced the formation of myeloid cell forming complexes (MCFCs) from embryoid bodies (EBs) in small-scale suspension culture on an orbital shaker (Supplementary Fig.\u00a0[1a](#MOESM1){ref-type=\"media\"} and [1b](#MOESM1){ref-type=\"media\"}). After 10--15 days, MCFCs continuously produced iPSC-Mac that could be harvested weekly for up to 3 months. Generated iPSC-Mac exhibited a clear surface marker profile of CD45^+^CD11b^+^CD14^+^CD163^+^CD34^\u2212^TRA1-60^\u2212^, although freshly collected cells comprised a minor population of CD45^+^/CD11b^+^/CD14^\u2212^/CD163^\u2212^ immature myeloid cells (Supplementary Fig.\u00a0[1c](#MOESM1){ref-type=\"media\"} and d). Following terminal differentiation for 7 more days, iPSC-Mac represented a homogenous population of CD45^+^CD11b^+^CD14^+^CD163^+^CD34^\u2212^TRA1-60^\u2212^ cells with classical macrophage-like morphology and efficiently phagocytosed fluorescently labeled *E. coli* particles (Supplementary Fig.\u00a0[1e-g](#MOESM1){ref-type=\"media\"}).\n\nWe then translated the suspension-based differentiation into stirred tank bioreactors using an industry-compatible system (DASbox Mini Bioreactor System)^[@CR17]^ previously applied for the efficient cultivation of human iPSC and their differentiation into cardiomyocytes^[@CR18]^ (Fig.\u00a0[1a, b](#Fig1){ref-type=\"fig\"} and Supplementary Fig.\u00a0[2a](#MOESM1){ref-type=\"media\"}). From day 10 onwards, weekly harvest of iPSC-Mac from the bioreactors showed an increase in cell yield over time, reaching a stable production of \\~1--3\u2009\u00d7\u200910^7^ iPSC-Mac per week as early as in week 3, which was maintained for more than 5 weeks in two independent bioreactor runs (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}, Supplementary Movie\u00a0[1](#MOESM4){ref-type=\"media\"}). Efficient generation of iPSC-Mac in both bioreactor experiments was reflected by the weekly increase in biomass, particularly during the first days after full medium refreshment. Dissolved oxygen (DO) and pH monitoring revealed expected zigzag-like patterns typical for repeated batch cultures. Notably, all process parameters showed maintenance of repetitive patterns after reaching the steady state of macrophage production around d15--20, confirming the overall stability of the process (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"} and Supplementary Fig.\u00a0[2b](#MOESM1){ref-type=\"media\"}). This finding was further supported by stable values for glucose, lactate, lactate dehydrogenase, and osmolality determined weekly parallel to macrophage harvests (Supplementary Fig.\u00a0[2c](#MOESM1){ref-type=\"media\"}). Similarly, secretion of cytokines/chemokines associated with the activation of macrophages, such as IL2, IL6, IL8, MCP1, TNF, and IFN\u03b12, was detected from the first harvest (week 2) onwards (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}) and corresponded with the appearance of CD45^+^ iPSC-Mac. MCFCs cultivated in the bioreactor sustained their morphology during the entire process and continuously generated iPSC-Mac of typical morphology and a CD45^+^CD14^+^ surface marker profile (Fig.\u00a0[1f](#Fig1){ref-type=\"fig\"}).Fig. 1Continuous generation of human iPSC-Mac in stirred tank bioreactors. **a** Scheme of hematopoietic differentiation of human iPSC in stirred tank bioreactors (DASbox system). **b** Representative pictures of DASbox bioreactor filled with floating MCFCs (left). Images of the 8-blade impeller (right). **c** Individual cell counts of viable macrophages produced in bioreactors (*n*\u2009=\u20092 of independent bioreactor runs, mean\u2009\u00b1\u2009SD.). **d** Representative macrophage harvest counts (upper graph) and corresponding data from continuous process monitoring (biomass, temperature, pH, and dissolved oxygen (DO) level) for the entire cultivation phase of one 42-day bioreactor run. **e** Analysis of human cytokines in the medium from the bioreactor (technical duplicates, mean\u2009\u00b1\u2009SD). **f** Representative light microscopy of macrophage generation analyzed at harvests 1--5. First row shows brightfield images of non-filtered medium samples from bioreactors (scale bar 500\u2009\u00b5m). Second row shows brightfield images of freshly harvested macrophages separated from MCFCs by sedimentation (scale bar 50\u2009\u00b5m). Third and fourth row show cytospin (scale bar 50\u2009\u00b5m) and flow cytometric analysis of iPSC-Mac (gray line: respective isotype control, red line: CD45, blue line CD14)\n\nDetailed characterization of bioreactor-derived iPSC-Mac {#Sec4}\n--------------------------------------------------------\n\nDetailed characterization of unsorted, bioreactor-derived, and freshly harvested iPSC-Mac revealed a highly pure CD45^+^CD11b^+^CD163^+^CD14^+^CD34^\u2212^TRA1-60^\u2212^ phenotype, and a typical morphology after adherence to tissue culture plates (Fig.\u00a0[2a, b](#Fig2){ref-type=\"fig\"}). Comparison of iPSC-Mac to non-differentiated hiPSC and peripheral blood mononuclear cells (PBMC)-derived macrophages (PBMC-Mac) by unsupervised, hierarchical heatmap clustering of whole transcriptomes revealed close proximity of iPSC-Mac and PBMC-Mac when compared to iPSC (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). Analysis of genes associated with pluripotency and activation of innate immune response, confirmed efficient differentiation of iPSC into macrophage-like cells (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}). Importantly, genes associated with macrophage function such as the toll-like receptors (TLR) 1 and 4, CD14, or components of the NF-\u03baB signaling pathway (gene ontology activation of innate immune response (GO: 0002218)) were significantly upregulated in iPSC-Mac and PBMC-Mac versus iPSC, whereas the opposite could be observed when analyzing genes associated with pluripotency (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}). As expected, upregulated genes in iPSC-Mac versus pluripotent hiPSC were assigned to CD14^+^ monocytes, CD33^+^ myeloid cells and whole blood, including key genes, such as *CYBB*, *CSFR1*, *CCR1*, or the formyl peptide receptor 1 (*FPR1*).Fig. 2Characterization of iPSC-Mac derived from stirred tank bioreactors. **a** Representative flow cytometric analysis of iPSC-Mac derived from bioreactor differentiation (pre-gated on viable cells in forward and sidewards scatter (FSC/SSC) analysis, gray line: respective isotype control, green line: cell surface marker). **b** Representative pictures of either brightfield (upper image, scale bar 200\u2009\u00b5m) or cytospin preparations (lower image, scale bar 100\u2009\u00b5m) of terminally differentiated iPSC-Mac derived from the bioreactor. **c**--**g** Transcriptome analysis of iPSC, iPSC-derived macrophages (iPSC-Mac) and PBMC-derived macrophages (PBMC-Mac) (*n*\u2009=\u20092, biological replicates). **c** Unbiased, hierarchical heatmap clustering. **d** Heatmap of differentially regulated genes (*p*\u2009\\<\u20090.001) associated with activation of innate immune response (GO:0002218 left) and pluripotency-associated genes (right). **e** Volcano plot showing differentially expressed genes between PBMC-Mac and iPSC-Mac (FDR\u2009=\u20090.5). **f** Gene set enrichment analysis (GSEA) from iPSC-Mac versus PBMC-Mac shows enrichment of genes related to yolk sac-derived macrophages. The gene set was derived from\\>5-fold upregulated genes of murine yolk sac (YS)-derived macrophages compared to bone marrow-derived macrophages. NES -- normalized enrichment score. **g** Heatmap of \\>5-fold regulated genes associated with YS macrophages shows predominantly upregulation in iPSC-Mac compared to PBMC-Mac\n\nBesides these global similarities between iPSC-Mac and PBMC-Mac, direct comparison between the two cell types revealed several differentially expressed genes (DEG) (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}). Here, PBMC-Mac showed higher expression of HLA II molecules (*HLA-DQs*, *HLA-DRs*), as well as genes associated with CD56^+^ natural killer (NK) cells, such as granzyme A, K, and H. On the other hand, iPSC-Mac showed significant higher levels of extracellular matrix proteins, such as *CYR61*, *COL3A1*, *KRT 8/18/19* as well as several members of the insulin-like growth factor axis. Interestingly, gene set enrichment analysis (GSEA) revealed a clear enrichment of genes associated with yolk sac-derived macrophages in iPSC-Mac (Fig.\u00a0[2f](#Fig2){ref-type=\"fig\"}), including the DNA-binding protein inhibitor ID1, transforming growth factor-beta 2 (*TGFB2*) and insulin-like growth factor 2 (*IGF2*) (Fig.\u00a0[2g](#Fig2){ref-type=\"fig\"}), suggesting a transcriptional phenotype of iPSC-Mac similar to primitive macrophages from the yolk sac.\n\nTaken together, our iPSC-Mac show close transcriptional similarities to PBMC-Mac, however, with a distinct signature towards primitive macrophages.\n\nAntimicrobial activity of bioreactor-derived iPSC-Mac {#Sec5}\n-----------------------------------------------------\n\nNext, bioreactor-derived iPSC-Mac were evaluated for their in vitro antimicrobial activity. Scanning electron microscopy (SEM) at different time points after incubation of iPSC-Mac with fluorescently labeled latex beads revealed typical changes in the overall cell morphology early after the stimulus and efficient phagocytic uptake of beads over time (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). IPSC-Mac also phagocytosed GFP-labeled *P. aeruginosa* (PAO1) at 37\u2009\u00b0C in comparable efficiency to PBMC-Mac, whereas no phagocytosis was observed at 4\u2009\u00b0C, confirming active phagocytosis (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"} and Supplementary Figure\u00a0[3a](#MOESM1){ref-type=\"media\"}). To gain insights into the ability of iPSC-Mac to remodel their transcriptome towards the signature of activated macrophages, we performed whole-transcriptome analysis of iPSC-Mac and PBMC-Mac before and after contact with *P. aeruginosa* (PAO1). Hierarchical cluster analysis of genes associated with inflammatory response (Fig.\u00a0[3c](#Fig3){ref-type=\"fig\"}) and innate immune response (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}) showed marked upregulation of cytokines (e.g. *IL23A, TNF, IL1A/B, IL6*), chemokines (e.g. *CCL3, CCL4, CCL20, CXCL2, CXCL3*) and molecules involved in NF\u03baB signaling in iPSC-Mac as well as PBMC-Mac in response to pathogen contact. Although both macrophage types showed characteristic transcriptional activation after stimulation, iPSC-Mac showed a more pronounced response (Fig.\u00a0[3c, d](#Fig3){ref-type=\"fig\"}; Supplementary Fig.\u00a0[3b](#MOESM1){ref-type=\"media\"}). This was further supported by analysis of \\>5-fold upregulated genes after pathogen contact showing 136 genes upregulated in iPSC-Mac compared to only 40 upregulated genes in PBMC-Mac (Fig.\u00a0[3e](#Fig3){ref-type=\"fig\"}). PBMC-Mac shared 29 (72.5%) of those upregulated genes, including important regulators of innate immunity, such as the immune-responsive gene 1 (*IRG1*), *IL6*, *CXCL1,2* and *3*, *IL1A* and *B*, *TNF*, with iPSC-Mac. Gene ontology enrichment analysis of genes only upregulated in PBMC-Mac revealed highest scores for GOs associated with activation of *IL10*, *GM-CSF*, *IL17*, and *IL12*, whereas genes upregulated in iPSC-Mac revealed enrichment in GOs associated with regulation of transcription, as well as acute inflammatory response. To investigate the quality of the antimicrobial response, we next analyzed the top 100 upregulated genes in iPSC-Mac or PBMC-Mac following pathogen challenge. Here, GO enrichment analysis for molecular function and biological processes, revealed similar GOs associated with chemokine and cytokine activity (Fig.\u00a0[3f](#Fig3){ref-type=\"fig\"}) as well as inflammatory response (Fig.\u00a0[3g](#Fig3){ref-type=\"fig\"}) in both cell types. Taken together, both, iPSC-Mac and PBMC-Mac demonstrate overall similar characteristic transcriptional changes after pathogen contact, but with a more pronounced response in iPSC-Mac.Fig. 3Antimicrobial activity of iPSC-Mac generated in stirred tank bioreactors. **a** Scanning raster electron microscopy depicting phagocytosis of latex beads (green) by iPSC-Mac (red) at different time points. **b** Rate of phagocytosis by terminally differentiated iPSC-Mac (blue) and PBMC-Mac (green) after 2\u2009h of incubation with GFP-labeled *P. aeruginosa* (PAO1) at 4 or 37\u2009\u00b0C (*n*\u2009=\u20093 of biological replicates, mean\u2009\u00b1\u2009s.e.m, two-way ANOVA with Sidak's multiple comparisons test, ns denotes not significant, representative data and gating strategy provided in Supplementary Figure\u00a0[3a](#MOESM1){ref-type=\"media\"}). **c**, **d** Heatmap of commonly regulated genes in iPSC-Mac and PBMC-Mac associated with (**c**) inflammatory response (GO:0006954; *\u03c3*/*\u03c3*~max~\u2009=\u20090.13, *P*\u2009\\<\u20090.05) and (**d**) innate immune response (GO:0045087, *\u03c3*/*\u03c3*~max~\u2009=\u20090.13, *P*\u2009\\<\u20090.05) after 1\u2009h exposure to PAO1. **e** Venn-diagramm of \\>5-fold upregulated genes in iPSC-Mac (blue) and PBMC-Mac (green) after pathogen contact. **f**, **g** Top-ranked GOs associated with **f** biological processes and **g** molecular function (according to human gene atlas) of top 100 upregulated genes after 1\u2009h pathogen contact (without pre-selection, EnrichR)\n\niPSC-Mac prevent respiratory infections by *P. aeruginosa* {#Sec6}\n----------------------------------------------------------\n\nTo evaluate the therapeutic efficacy of bioreactor-derived iPSC-Mac in vivo, we utilized the humanized mouse model C;129S4-*Rag2*^*tm1.1Flv*^ *Csf1*^*tm1(CSF1)Flv*^ *Csf2*/*Il3*^*tm1.1(CSF2,IL3)Flv*^ *Thpo*^*tm1.1(TPO)Flv*^ *Il2rg*^*tm1.1Flv*^ Tg(SIRPA)1Flv/J (hIL-3/GM-CSF KI), an immunodeficient strain with impaired alveolar macrophage development and an established model to study pulmonary infections^[@CR19]--[@CR21]^. Given its frequent application in several murine infections models, including pulmonary infections, we used the *P. aeruginosa strain* PAO1 for our infection experiments^[@CR22],[@CR23]^.\n\nWe first established an acute pulmonary infection model in hIL-3/GM-CSF KI mice by intratracheal installation of PAO1. In line with the current animal welfare regulations, we were able to identify a suitable dose of *P. aeruginosa*, which leads to a prominent infection in mice represented by a disease score of 6--8 in hIL-3/GM-CSF KI mice (0: unaffected, 10: moribund) within 24\u2009h post infection without leading to mortality of the animals. Applying the afore mentioned protocol, hIL-3/GM-CSF KI mice were infected with the *P. aeruginosa* strain PAO1 and co-administered with 4\u2009\u00d7\u200910^6^ bioreactor-derived, non-purified iPSC-Mac in the same instillation process (solely *PAO1* infected mice served as controls). The infection course was closely monitored and mice were killed 24\u2009h post infection for end analysis (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}).Fig. 4Pulmonary infection and simultaneous macrophage transplantation in hIL-3/GM-CSF KI mice. **a** Scheme of pulmonary transfer of *P. aeruginosa* (PAO1) and simultaneous transplantation of iPSC-Mac (PiMT) into hIL-3/GM-CSF KI mice. **b** Rectal temperature and disease score 24\u2009h post infection of hIL-3/GM-CSF KI mice infected with PAO1 (infected) or infected and transplanted (infected\u2009+\u2009PiMT). Disease score and temperatures measured before the experiment served as control values (*n*\u2009=\u20096 animals/group, mean\u2009\u00b1\u2009s.e.m). **c** Change in body weight after 24\u2009h. Values are normalized to the respective weights before infection (*n*\u2009=\u20096 animals/group, mean\u2009\u00b1\u2009s.e.m). **d** Lung function measured by head-out body plethysmography. Values measured pre infection served as control values (*n*\u2009=\u20096 animals/infected and infected\u2009+\u2009PiMT groups, *n*\u2009=\u200912 animals for control values, mean\u2009\u00b1\u2009s.e.m). **e** Colony forming units (CFU) of PAO1 per left lung after 24\u2009h (*n*\u2009=\u20096 animals/group, mean\u2009\u00b1\u2009s.e.m). **f** Images of bronchio-alveolar lavage fluid (BALF) samples and BALF absorbance at 650\u2009nm (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **g** Flow cytometric analysis of BALF and lung. Percentage of mouse granulocytes (determined as GR1^+^ cells) in BALF and Lung (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m, exemplarily data and gating strategy provided in Supplementary Figure\u00a0[4c](#MOESM1){ref-type=\"media\"}). **h** Right lung histology. Left: Two representative images of infected (left) and infected\u2009+\u2009PiMT (right) mice. Scale bars: 500\u2009\u00b5m upper row and 100\u2009\u00b5m lower row. Right: histological scoring (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **i** Representative diagram of BALF and lung of one animal per group stained with hCD45. (\\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.01, \\*\\*\\**p*\u2009\\<\u20090.001, \\*\\*\\*\\**p*\u2009\\<\u20090.0001, ns denotes not significant; statistical significances were assessed using one-way ANOVA with Dunnett's multiple comparisons test (**b**, **d**, **g**) or Student's *t*-test (**c**, **e**, **f**, **h**))\n\nInfected hIL-3/GM-CSF KI mice displayed clinical symptoms already 4\u2009h post infection as indicated by an elevated disease score of 3.5\u2009\u00b1\u20090.9, which gradually increased over time up to 6.1\u2009\u00b1\u20090.4 after 24\u2009h. In addition, a decrease in rectal temperature to 35.0\u2009\u00b1\u20090.3\u2009\u00b0C and body weight loss of 2.6\u2009\u00b1\u20090.5\u2009g was observed in infected animals 24\u2009h post infection (all mean\u2009\u00b1\u2009s.e.m., *n*\u2009=\u20096). In contrast, animals receiving a simultaneous pulmonary iPSC-Mac transplantation (PiMT) showed only very mild symptoms of infection (Fig.\u00a0[4b, c](#Fig4){ref-type=\"fig\"} and Supplementary Fig.\u00a0[4a](#MOESM1){ref-type=\"media\"}). This was in line with normal rectal temperature of 37.1\u2009\u00b1\u20090.2\u2009\u00b0C, a very low disease score of 0.6\u2009\u00b1\u20090.2 and only little loss of body weight of \u22120.89\u2009\u00b1\u20090.2\u2009g 24\u2009h post infection in PiMT-treated animals (all mean\u2009\u00b1\u2009s.e.m., *n*\u2009=\u20096) (Fig.\u00a0[4b, c](#Fig4){ref-type=\"fig\"}). Similar beneficial effects of PiMT were demonstrated by head-out body-plethysmography to measure lung function^[@CR24]^. Whereas infected animals showed a decrease in tidal volume as well as expiratory and inspiratory time and an increase in breathing rate, animals from the infected\u2009+\u2009PiMT group showed normal values for all parameters analyzed (Fig.\u00a0[4d](#Fig4){ref-type=\"fig\"} and Supplementary Fig.\u00a0[4b](#MOESM1){ref-type=\"media\"}). As a consequence of PiMT, transplanted mice showed significantly reduced bacterial numbers in the lung 24\u2009h post infection compared to their non-transplanted controls (Fig.\u00a0[4e](#Fig4){ref-type=\"fig\"}). Moreover, erythrocytes were present only in bronchoalveolar lavage fluid (BALF) from infected mice not receiving PiMT (Fig.\u00a0[4f](#Fig4){ref-type=\"fig\"}). Lung inflammation was furthermore indicated by an increase in murine GR1^+^ granulocytes in the BALF and lungs of infected animals, but not infected\u2009+\u2009PiMT or control mice (Fig.\u00a0[4g](#Fig4){ref-type=\"fig\"} and Supplementary Figure\u00a0[4c](#MOESM1){ref-type=\"media\"}). In addition, histological evaluation revealed massive granulocyte infiltration, severe hemorrhage and alveolar edema in infected mice (score: 13.7\u2009\u00b1\u20090.3), whereas lungs from infected\u2009+\u2009PiMT animals showed merely slight changes (score: 2.0\u2009\u00b1\u20091.2, both mean\u2009\u00b1\u2009s.e.m., *n*\u2009=\u20093) (Fig.\u00a0[4h](#Fig4){ref-type=\"fig\"} and Supplementary Fig.\u00a0[4d](#MOESM1){ref-type=\"media\"}). Reduced inflammation in infected\u2009+\u2009PiMT animals was associated with the detection of hCD45^+^ cells in lung and BALF as well as presence of macrophages in histological sections of the right lung (Fig.\u00a0[4i](#Fig4){ref-type=\"fig\"} and Supplementary Fig.\u00a0[4e](#MOESM1){ref-type=\"media\"}).\n\nPiMT rescues mice from severe respiratory infections {#Sec7}\n----------------------------------------------------\n\nAfter demonstrating the efficacy of iPSC-Mac in simultaneous infection experiments, we evaluated a clinically more relevant therapeutic PiMT treatment approach. In these experiments, hIL-3/GM-CSF KI mice were infected intra-pulmonary with *P. aeruginosa* and closely monitored for 3--4\u2009h until first disease symptoms developed (determined by a disease score \u22655). Subsequently, mice received 4\u2009\u00d7\u200910^6^ iPSC-Mac (infected\u2009+\u2009PiMT) only after the infection-related symptoms were readily manifested. Instead of PiMT, infected control mice received PBS only (infected) (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}).Fig. 5Pulmonary infection and therapeutic transplantation of iPSC-Mac in hIL-3/GM-CSF KI mice. **a** Scheme of pulmonary infection with *P. aeruginosa* (PAO1) and therapeutic transplantation of iPSC-Mac (PiMT) derived from bioreactors into hIL-3/GM-CSF KI mice. **b** Disease score of hIL-3/GM-CSF KI mice infected with PAO1 (infected, red), infected and transplanted (infected\u2009+\u2009PiMT, blue) and control mice receiving PBS twice (control, black). Left: Disease score over time. Right: Disease score 24\u2009h post infection (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **c** Infected mice (red dot) and infected\u2009+\u2009PiMT mice (blue dot) 24\u2009h after infection. **d** Representative trajectory of mouse activity 24\u2009h post infection for hIL-3/GM-CSF KI mice infected with PAO1 (infected, red), infected and transplanted (infected\u2009+\u2009PiMT, blue) and control mice receiving PBS twice (control, black) analyzed by video documentation and manual tracking. **e** Rectal temperature analyzed 24\u2009h post infection (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **f** Change in body weight after 24\u2009h. Values are normalized to the respective weights before infection (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **g** Colony forming units (CFU) of PAO1 per lung after 24\u2009h (*n*\u2009=\u20093 animals/group, mean\u2009\u00b1\u2009s.e.m). **h** Images of BALF samples. **i** Levels of human cytokines in BALF of infected\u2009+\u2009PiMT mice (*n*\u2009=\u20094: 2 animals and 2 technical duplicates, mean\u2009\u00b1\u2009s.e.m). **j** Lung histology images of infected (left), infected\u2009+\u2009PiMT (middle) and control (right) mice. Scale bars: 500\u2009\u00b5m upper row and 100\u2009\u00b5m lower row. (\\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.01, \\*\\*\\**p*\u2009\\<\u20090.001, \\*\\*\\*\\**p*\u2009\\<\u20090.0001, ns denotes not significant; statistical significances were assed using one-way ANOVA with Dunnett's multiple comparisons test (**b**, **e**, **f**) or Student's *t*-test (**g**))\n\nIn our infection and treatment schedule, therapeutic PiMT-treated mice showed a decrease in the disease score as well as a normalization of rectal temperature comparable to non-infected animals already within 4--8\u2009h post treatment. In contrast, infected mice receiving PBS showed pronounced disease progression over time (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). Of note, 24\u2009h post infection, infected mice showed marked disease symptoms, whereas animals that received a therapeutic PiMT showed a significantly reduced disease score (1.8\u2009\u00b1\u20090.2 infected\u2009+\u2009PiMT vs 8.1\u2009\u00b1\u20090.2 for infected animals, mean\u2009\u00b1\u2009s.e.m., *n*\u2009=\u20093) (Fig.\u00a0[5b, c](#Fig5){ref-type=\"fig\"}). In addition, the elevated disease score in infected mice was associated with clearly restricted activity when compared to control animals and mice receiving the therapeutic PiMT (representative trajectory shown for one animal each, Fig.\u00a0[5d](#Fig5){ref-type=\"fig\"}). Efficiency of therapeutic PiMT was further documented by normalized rectal temperature and a tendency towards maintained body weight values 24\u2009h post infection (Fig.\u00a0[5e, f](#Fig5){ref-type=\"fig\"}). Moreover, a profound reduction in lung bacterial numbers in infected\u2009+\u2009PiMT mice was observed (Fig.\u00a0[5g](#Fig5){ref-type=\"fig\"}). Recapitulating our findings in the simultaneous transplant model, BALF of animals from the infected\u2009+\u2009PiMT group showed reduced erythrocyte levels compared to infected, non-transplanted controls (Fig.\u00a0[5h](#Fig5){ref-type=\"fig\"}). This observation was accompanied by the detection of important pro-inflammatory human cytokines, such as hIL6, hIL8, hINF\u03b12, hMCP-1, and TNF only in animals receiving a PiMT, whereas human cytokines could not be detected in control or infected only animals (Fig.\u00a0[5i](#Fig5){ref-type=\"fig\"} and Supplementary Fig.\u00a0[5a](#MOESM1){ref-type=\"media\"}). In contrast, levels for murine pro-inflammatory cytokines (mIL1b, mIL6, mMCP-1, mTNF) were highly elevated in infected only animals, whereas almost normal levels could be detected in PiMT-treated animals (Supplementary Fig.\u00a0[5b](#MOESM1){ref-type=\"media\"}). Inflammation was furthermore assessed by lung histology sections. Here, infected mice showed several areas of massive granulocytic infiltration, severe hemorrhage and alveolar edema, which were barely detectable in mice treated with iPSC-Mac from the bioreactor (Fig.\u00a0[5j](#Fig5){ref-type=\"fig\"}).\n\nDiscussion {#Sec8}\n==========\n\nIn the present study, we introduce an immunotherapy approach using iPSC-Mac to enhance innate pulmonary immunity and combat bacterial infections. As clinically relevant numbers of macrophages can hardly be produced from somatic cell sources, we demonstrate the use of human iPSC and their scalable differentiation in industry-compatible bioreactors for generating substantial quantities of functional macrophages.\n\nThe strong and rapid therapeutic effect observed in our proof-of-concept study is of particular importance for a broad applicability of iPSC-Mac to target bacterial infections and quick clinical implementation. We evaluated the antimicrobial activity of iPSC-Mac against *P. aeruginosa*. However, a phagocyte-based immunotherapy approach might allow for a broad spectrum of applications in a number of different infection scenarios caused by other Gram-negative or -positive bacteria, such as *Streptococcus pneumoniae* or *Staphylococcus aureus*. Moreover, this approach might also be used to target pathogens associated with implant infections, which meanwhile pose an immense health and economical problem. This being said, new treatment options are of specific clinical relevance considering the increasing numbers of pathogens, which are refractory to standard or reserve antibiotic therapy^[@CR25]--[@CR27]^.\n\nAs a future clinical prerequisite, iPSC-Mac generated with our bioreactor platform represent a highly pure population without implementation of further FACS or MACS-based enrichment strategies and shared phenotypic, transcriptional and functional hallmarks with PBMC-Mac. Despite these similarities, iPSC-Mac exhibited a stronger response to pathogen stimulation as demonstrated by whole transcriptome analysis, further favoring their application as a cellular therapeutic for the treatment of bacterial infections.\n\nFor our immunotherapy approach, we employed an acute infection model using hIL-3/GM-CSF KI mice, an immunodeficient mouse strain with impaired alveolar macrophages development^[@CR19]^. In this model, treatment of acute *P. aeruginosa-*induced pneumonia either by simultaneous or, even more important, therapeutic PiMT was highly effective within a short time frame. These proof-of-concept experiments were performed with a high cell dose of 4\u2009\u00d7\u200910^6^ iPSC-Mac per single intra-pulmonary application. However, even with this maximum cell dose, no apparent adverse events were observed in either treatment scenario. This suggests that transplantation of iPSC-Mac is well tolerated and safe, concurring with observations from pulmonary macrophage transplantation studies using macrophages from different stem cell sources^[@CR28]--[@CR31]^. Nevertheless, more studies are needed to address potential immunological side effects induced by the transplanted macrophages. Similarly, longer follow-up, detailed posology as well as intense preclinical safety and toxicology studies will be required prior to clinical translation.\n\nConsidering a body weight of 25\u2009g per mouse, clinical translation of iPSC-Mac for the treatment of respiratory infections would require a maximal cell dose of 1\u2009\u00d7\u200910^10^ iPSC-Mac for a complete lung of a 60\u2009kg patient. Even without further process optimization, this would translate to a 40--60\u2009l production scale, which is in principle feasible with current bioreactor technologies^[@CR14]^. Continuous process monitoring of our bioreactor-based differentiation revealed a substantial increase in biomass especially in the first 2--3 days of the repeated batch culture. This was followed by partial recovery of dissolved oxygen levels and a profound drop in pH in the last days before medium refreshment. These observations suggest the presence of process-limiting factors and highlight the potential for further process optimization. This is achievable by cultivation at higher cell densities combined with the application of perfusion systems and feedback control of oxygen, pH, and other process parameters, which can multiply cell yields of hPSC and their progenies^[@CR32]^.\n\nGiven the impressive in vivo functionality of iPSC-Mac in the acute infection model, a potentially important extension of this work would be to evaluate the therapeutic efficacy of a macrophage-based immunotherapy in chronic infections with *P. aeruginosa*, as frequently observed in patients suffering from chronic obstructive pulmonary disease (COPD)^[@CR33]^ or cystic fibrosis (CF)^[@CR34]^. As previously shown for endogenous alveolar macrophages^[@CR35]^, iPSC-Mac might be induced towards a vaccine-induced macrophage (ViM) phenotype, which upon intra-pulmonary administration could protect mice from *P. aeruginosa* infections. This might be of particular interest in life-threatening, chemotherapy associated infections. Whereas in our \"*protective*\" regimen, we have used the simultaneous administration of iPSC-Mac and *P. aeruginosa*, administration of iPSC-derived ViM prior the infection would provide additional insights in the protective properties. Along this line, further long-term studies are needed to assess potential long-term pulmonary engraftment and sustained functionality of the transplanted iPSC-Mac. In our transplantation scenarios, iPSC-Mac were still detected at 24\u2009h, however, later time points were not investigated. Earlier studies using the same mouse model demonstrated long-term pulmonary engraftment of cord blood- or iPSC-derived macrophages, whereas no cells could be detected in other organs associated with homing of macrophages^[@CR28],[@CR29]^. Although the open lung microenvironment may serve as an instructive signal for the local engraftment of cells, the ancestry of the individual macrophages subset also has to be considered in more detail. Previous studies suggested a primitive fingerprint of murine and human iPSC-derived macrophages^[@CR30],[@CR36],[@CR37]^, which is further supported by our microarray analysis. Although our cells seem to have a primitive origin, previous studies revealed a similar adaption potential of iPSC-Mac and CD34^+^-Mac, highlighting a similar tissue plasticity^[@CR28]^. Clearly, further studies are needed to evaluate whether iPSC-Mac can also provide a long-term antimicrobial defense and integrate into the lung environment after the infection is solved.\n\nTaken together, we here introduce an immunotherapy approach employing iPSC-Mac to target bacterial airway infections. To translate this concept to clinical application, either allogeneic or genetically modified universal iPSC^[@CR38]^ may be employed to produce suitable macrophages in a cost-effective way by industrial scale bioreactor systems as an off-the-shelf product. While our proof-of-concept study is done in an acute infection mouse model, a first application of an iPSC-Mac-based therapy (iMATH) approach in humans may be envisioned in CF patients suffering from chronic *P. aeruginosa* infections. Given the defined clinical background of such patients, first-in-human-studies may be performed in a homogenous patient population before this approach is translated to patients suffering from acute and life-threatening respiratory infections. Further development of this technology, including process upscaling and the generation of additional hematopoietic cell types as well as assessment in other preclinical models may allow for innovative cell-based treatment strategies for a wide variety of diseases.\n\nMethods {#Sec9}\n=======\n\niPSC cultivation {#Sec10}\n----------------\n\nHuman iPSC (hCD34iPSC16) have been previously generated from mobilized peripheral blood CD34^+^ cells^[@CR16]^ and were culture on irradiated murine embryonic fibroblasts (MEF) in human iPSC medium (knockout DMEM, 20% knockout serum replacement, 1\u2009mM [l]{.smallcaps}-glutamine, 1% NEAA, 1% penicillin/streptomycin (all Invitrogen), 0.1\u2009mM \u03b2-mercaptoethanol (Sigma-Aldrich), supplemented with 10\u2009ng per mL bFGF (PeproTech). Cells were splitted every 7 days using 0.2% collagenase IV for 10--15\u2009min. Basic-FGF was omitted from the maintenance medium for the last 3--5 days prior to EB formation. Human iPSC cultures have been regularly screened for mycoplasma contamination by PCR.\n\nHematopoietic differentiation in suspension culture {#Sec11}\n---------------------------------------------------\n\nTo induce hematopoietic differentiation in suspension, \\~80% confluent hiPSC cultures on day 7--10 after passaging were harvested using 0.2% collagenase IV for 10--15\u2009min. EBs were formed by pooling collected cell clusters from 3 wells of a 6-well plate per 1 well of a 6-well suspension plate in 3\u2009mL of hiPSC medium without bFGF, but supplemented with 10\u2009\u00b5M Y-27632 (Tocris). After 5 days incubation on an orbital shaker (Celltron, Infors HT) at 85 revolutions per minute (rpm), largest EBs (\\>200\u2009\u00b5m diameter) were selected manually with a binocular or by sedimentation properties (sedimentation for 10--15\u2009min) and transferred to a new 6-well plate containing 3\u2009mL differentiation medium I (X-VIVO 15 (Lonza), 1% penicillin-streptomycin (Life Technologies), 1\u2009mM [l]{.smallcaps}-glutamine, and 0.05\u2009mM \u03b2-mercaptoethanol (Sigma-Aldrich)) supplemented with 25\u2009ng per mL IL3 and 50\u2009ng per mL M-CSF. Medium was refreshed every 6--7 days. From day 10 to 15 onwards, iPSC-Mac were collected from the medium once a week.\n\nHematopoietic differentiation in stirred tank bioreactors {#Sec12}\n---------------------------------------------------------\n\nThe bioreactor (DASbox Mini bioreactor system, Eppendorf) was setup and calibrated as previously described^[@CR39]^. In brief, the 250\u2009mL glass vessel was equipped with an 8-blade impeller (60\u00b0 pitch) and probes for online monitoring of biomass (Aber Instruments), pH, DO as well as control of temperature. Calibration was performed in 120\u2009mL chemically defined X-VIVO 15 (Lonza).\n\nFor hematopoietic differentiation in the bioreactor, iPSC were expanded to 20 6-well plates, maintained in human iPSC-medium for 4--7 days and then cultivated for 3 days in the absence of bFGF. EB formation was performed equivalent to suspension cultures. After 5 days, EBs were selected by sedimentation properties (sedimentation for 10--15\u2009min) to exclude debris, single cells as well as small cell clusters and subsequently transferred to equilibrated bioreactors. Cells were cultivated in differentiation medium I at 37\u2009\u00b0C with constant headspace-gassing at 3\u2009L per hour (21% O~2~; 5% CO~2~) and stirring at 50\u2009rpm. To monitor MCFCs integrity and macrophages formation, 1\u2009mL samples were collected 1--2 times per week via the sampling port without interrupting the culture process. Differentiation medium I (X-VIVO 15 (Lonza), 1% penicillin-streptomycin (Life Technologies), 1\u2009mM [l]{.smallcaps}-glutamine and 0.05\u2009mM \u03b2-mercaptoethanol (Sigma-Aldrich)) supplemented with 25\u2009ng per mL IL3 and 50\u2009ng per mL M-CSF) was manually replaced every 6--7 days with optional fed of 20\u2009mL after 3--4 days. Macrophages were collected weekly by separation from MCFC via sedimentation (4--5\u2009min) and subsequent filtering of the medium through a 100\u2009\u00b5m strainer (PluriSelect). Retained MCFCs were returned to the bioreactor. Macrophages were collected from filtered medium via centrifugation at 300\u00d7*g* for 4\u2009min.\n\nData from online monitoring were processed using Microsoft Excel 2016 and GraphPad Prism 6. Supernatant was analysed for concentration of glucose and lactate using YSI 2700 select biochemistry analyser, for osmolarity using Osmomat 300 (Gonotec) and for concentration of lactate dehydrogenase according to manufacturer's instruction (MAK066, Sigma) using a microplate reader (Paradigm, Beckman Coulter).\n\nTerminal differentiation {#Sec13}\n------------------------\n\nFor further maturation, cells freshly collected from MCFCs were cultured in differentiation medium II (RPMI1640 medium supplemented with 10% fetal calf serum (FCS), 2\u2009mM [l]{.smallcaps}-glutamine, 1% penicillin-streptomycin) containing 50\u2009ng per mL hM-CSF for at least 7 days.\n\nGeneration of PBMC-derived macrophages {#Sec14}\n--------------------------------------\n\nDonation of peripheral blood and isolation of peripheral blood mononuclear cells (PBMC) was approved by the local ethical committee (Ethics Committee Hannover Medical School). All healthy donors gave written informed consent. PBMC were isolated from the peripheral blood of healthy volunteers by gradient centrifugation using Biocoll Separating Solution (40\u2009min, 400\u00d7*g*; Biochrome, Billerica, MA). Subsequently, cells were cultured in RPMI1640 medium supplemented with 10% fetal calf serum, 2\u2009mM [l]{.smallcaps}-glutamine, 1% penicillin-streptomycin (all Invitrogen), and hIL-3 and hM-CSF (50\u2009ng per mL each, PeproTech) for 1 week. After this, PBMC-Mac were cultivated for further 3--4 days in differentiation medium containing 50\u2009ng per mL M-CSF only.\n\nFlow cytometry {#Sec15}\n--------------\n\nFlow cytometric analysis of myeloid cells was performed as described^[@CR16],[@CR40]^. For macrophages, PBS supplemented with 10% FCS was used to prevent unspecific binding. Cells were stained with 0.5--1\u2009\u00b5l of the respective antibody and analyzed with a FACScalibur cytometer (Beckton & Dickinson, Heidelberg, Germany) and analyzed with FlowJo software (TreeStar, Ashland, OR). Antibodies were purchased from eBioscience: hTRA-1-60-PE (Cat-No: 12-8863-80), hCD11b-APC (Cat-No: 17-0118-41), hCD14-PE (Cat-No: 12-0149-42), hCD163-APC (Cat-No: 17-1639-41, hCD16-FITC (Cat-No: 11-0168-41), hCD34-FITC (Cat-No: 11-0349-41) and isotype-controls: mouse-IgG1a-PE (Cat-No: 12-4714-41), FITC (Cat-No: 11-4714-41) or APC (Cat-No: 17-4714-41), and rat-IgG2a-PE (Cat-No: 12-4321-81). Antibodies from Biolegend San Diego, CA, United States: hCD86-APC (Cat-No: 305411), hCD66b-FITC (Cat-No: 305104), or hCD45-PE (Cat-No: 304007).\n\nFor flow-cytometric analysis of mouse lung and BALF, samples were fixed using 4%PFA. After this, samples were incubated with 1\u2009\u00b5l\u2009fc receptor blocking antibodies (CD16/CD32, eBioscience, Cat-No: 14-0161-81) for 20\u2009min to prevent unspecific binding and finally stained with 1\u2009\u00b5l of the respective antibody for 1\u2009h at 4\u2009\u00b0C in the dark. Used antibodies were purchased from Biolegend San Diego, CA, United States: hCD45-PeCy7, and eBioscience: mGR1-eFluor450. Samples were analyzed at a BD LSR II Flow Cytometer (Beckton & Dickinson, Heidelberg, Germany) and analyzed with FlowJo software (TreeStar, Ashland, OR).\n\nCytospin preparation {#Sec16}\n--------------------\n\nA total of 20,000--50,000 cells were spun on glass slides at 600\u00d7*g* for 7\u2009min and stained for 5\u2009min in 0.25% May-Gr\u00fcnwald and 20\u2009min in 0.4% Giemsa stain modified solution (Sigma).\n\nPhagocytosis assays {#Sec17}\n-------------------\n\nThe phagocytic activity of iPSC-Mac derived from suspension or adherent cultures and terminally differentiated on tissue-culture plates was assessed by flow cytometry. Therefore, 1\u2009\u00d7\u200910^5^ cells were incubated with pHrodo\u2122 Red *E. coli* BioParticles\u00ae Conjugate (MolecularProbes/Thermo Fisher Scientific, Schwerte, Germany) or medium for 2\u2009h at 37\u2009\u00b0C or 4\u2009\u00b0C as a negative control. After incubation, the cells were put on ice for 10\u2009min. Analysis was performed using a Beckman Coulter FC500 flow cytometer.\n\nThe functional capacity of iPSC-Mac and PBMC-Mac to phagocytose vital *P. aeruginosa* was assessed using GFP-PAO1 (wild-type *P. aeruginosa* PAO1 tagged with green fluorescent protein (GFP) by Tn*7* transformation^[@CR41]^, (kindly provided by Thomas Bjarnsholt, University of Copenhagen) and compared to PBMC-derived Macrophages (PBMC-Mac). For the phagocytosis assay, 1\u2009\u00d7\u200910^5^ iPSC-Mac were incubated for 2\u2009h with 6\u2009\u00d7\u200910^5^ CFU GFP-PAO1 at 37\u2009\u00b0C or 4\u2009\u00b0C as a negative control. Medium controls were treated similar. After incubation, cells were put on ice for 10\u2009min and subsequently fixed with 2% paraformaldehyde (PFA) solution for 30\u2009min. Analysis was performed using a Beckman Coulter FC500 flow cytometer.\n\nElectron microscopy {#Sec18}\n-------------------\n\nMacrophages were grown on 1\u2009cm diameter round coverslips. Latex beads (1\u2009\u00b5m diameter) were added to cells at 4\u2009\u00b0C and adhesion of latex beads to the surface of the macrophages was allowed for 5\u2009min. Following samples were washed with cold PBS and warmed up with fresh culture medium to 37\u2009\u00b0C. Phagocytosis was allowed for up to 1\u2009h and samples were fixed at different time points using 1.5% Paraformaldehyde, 1.5% Glutaraldehyde in 150\u2009mM HEPES, pH 7.35. Samples were then dehydrated using an increasing methanol series. Critical point drying was performed using a CPD030 critical point dryer (Balzers, Lichtenstein) following manufacturer instructions. Coverslips were then sputtered with gold (Sem Coating System, Polarion) and SEM was carried out using a Philips SEM 505 (Eindhoven, The Netherlands).\n\nBacterial culture {#Sec19}\n-----------------\n\nFor experiments *P. aeruginosa* laboratory strain *PAO1*^[@CR22]^ was taken from a stock culture kept at \u221280\u2009\u00b0C and grown in Luria Broth (LB) overnight. After washing with sterile PBS the desired infectious dose was extrapolated from a standard growth curve. For the determination of the actual dosage, inoculates were serially plated on LB agar plates via the drop-plate method^[@CR42]^ and CFU determined after 16--18\u2009h incubation at 37\u2009\u00b0C.\n\nCollection of microarray samples {#Sec20}\n--------------------------------\n\nTerminally differentiated human iPSC-Mac or human PBMC-Mac were seeded on 24-well plates (500,000 cells/well) and cultured overnight. On the next day, cells were washed three times with PBS. Subsequently, *P. aeruginosa* laboratory strain PAO1 in RPMI medium without antibiotics (MOI10) was centrifugated onto the cells (600\u00d7*g*) and incubated at 37\u2009\u00b0C. Cells with medium only served as non-infected controls. After 1\u2009h cells were de-attached, washed and resuspended in RNA lysis buffer. RNA isolation was performed with RNAeasy micro Kit (Quiagen) according to manufacturer's instructions. Human iPSC samples were obtained after sorting of iPSC for TRA-1-60^+^ to separate iPSC from feeder cells.\n\nMicroarray experiments {#Sec21}\n----------------------\n\nThe Microarray utilized in this study represents a refined version of the Whole Human Genome Oligo Microarray 4\u2009\u00d7\u200944\u2009K v2 (Design ID 026652, Agilent Technologies), called '054261On1M' (Design ID 066335) developed in the Research Core Unit Transcriptomics (RCUT) of Hannover Medical School. Microarray design was created at Agilent's eArray portal using a 1\u2009\u00d7\u20091\u2009M design format for mRNA expression as template. All non-control probes of design ID 026652 have been printed five times within a region comprising a total of 181,560 Features (170 columns\u2009\u00d7\u20091068 rows). Four of such regions were placed within one 1M region giving rise to four microarray fields per slide to be hybridized individually (Customer Specified Feature Layout). Control probes required for proper Feature Extraction software operation were determined and placed automatically by eArray using recommended default settings.\n\nA quantity of 30\u2009ng of total RNA was used to prepare aminoallyl-UTP-modified (aaUTP) cRNA (Amino Allyl MessageAmp\u2122 II Kit; \\#AM1753; Life Technologies) as directed by the company (applying one-round of amplification). The labeling of aaUTP-cRNA was performed by use of Alexa Fluor 555 Reactive Dye (\\#A32756; LifeTechnologies).\n\ncRNA fragmentation, hybridization, and washing steps were carried out as recommended in the 'One-Color Microarray-Based Gene Expression Analysis Protocol V5.7', except that 500\u2009ng of each fluorescently labeled cRNA population were used for hybridization.\n\nSlides were scanned on the Agilent Micro Array Scanner G2565CA (pixel resolution 3\u2009\u00b5m, bit depth 20). Data extraction was performed with the 'Feature Extraction Software V10.7.3.1' using the extraction protocol file 'GE1_107_Sep09.xml', except that 'Multiplicative detrending' algorithm was inactivated.\n\nMeasurements of on-chip replicates (quintuplicates) were averaged using the geometric mean of processed intensity values of the green channel, 'gProcessedSignal' (gPS) to retrieve one resulting value per unique non-control probe. Single features were excluded from averaging, if they (i) were manually flagged, (ii) were identified as Outliers by the Feature Extraction Software, (iii) lie outside the interval of '1.42 \u00d7 interquartile range' regarding the normalized gPS distribution of the respective on-chip replicate population, or (iv) showed a coefficient of variation of pixel intensities per Feature that exceeded 0.5.\n\nAveraged gPS values were normalized by quantile normalization approach first. Subsequently, values were additionally processed by global linear scaling: All gPS values of one sample were multiplied by an array-specific scaling factor. This factor was calculated by dividing a 'reference 75th percentile value' (set as 1500 for the whole series) by the 75th percentile value of the particular Microarray to be normalized ('Array I' in the formula shown below). Accordingly, normalized gPS values for all samples (microarray datasets) were calculated by the following formula:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\rm normalized}\\,{\\rm gPS}_{{\\rm Array}\\,{\\rm i}} = {\\rm gPS}_{{\\rm Array}\\,{\\rm i}}\\,{\\times}\\left( {1500\\,{\\rm per}\\,75{{\\rm th}}\\,{\\rm percentile}_{{\\rm Array}\\,{\\rm i}}} \\right).$$\\end{document}$$\n\nFinally, a lower intensity threshold (surrogate value) was defined based on intensity distribution of negative control features. This value was fixed at 15 normalized gPS units. All of those measurements that fell below this intensity cutoff were substituted by the respective surrogate value of 15.\n\nNormalized microarray data of all non-control features were imported into Omics Explorer software v3.2 (Qlucore) under default import settings for Agilent One Color mRNA Microarrays, except that any normalization option was deselected. Accordingly, data processing steps during import were: (1) log base 2 transformation, (2) baseline transformation to the median.\n\nHeatmap clustering analysis and generation of GO-based heatmaps were performed in Omics Explorer. Top 100 upregulated genes were calculated using the RCUTAS tool (V1.7; Hannover Medical School) and processed using Venny 2.1 (<http://bioinfogp.cnb.csic.es/tools/venny>). Gene ontology analysis of biological processes, molecular function and cell type classification based on the human gene atlas were conducted using Enrichr (<https://amp.pharm.mssm.edu/Enrichr>). Gene set enrichment analysis was performed using GSEA (v3.0; Broad Institute). Gene set for YS macrophages was obtained from Takata et al. using \\>5-fold upregulated genes compared to BMDM^[@CR37]^ and converted using OrthoRetriever (v1.2; <http://lighthouse.ucsf.edu/orthoretriever/>). Volcano plots were visualized using Perseus (v.1.5.6.0; <http://www.perseus-framework.org>; FDR\u2009=\u20090.05; s0\u2009=\u20090.1). Microarray data were deposited under accession number [E-MTAB-5436](https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-5436/) in the ArrayExpress database ([www.ebi.ac.uk/arrayexpress](http://www.ebi.ac.uk/arrayexpress)).\n\nCytokine secretion assays {#Sec22}\n-------------------------\n\nIn order to analyze the secretion of human cytokines in bioreactor samples or human and murine cytokines BALF samples, Luminex\u00ae analysis with a Cytokine Human 14-Plex Panel or Mouse 5 -Plex Panel (Millipore, Schwalbach, Germany) was performed as described before^[@CR40]^. Data were acquired on a Luminex-200 System and analyzed with the Xponent software v.3.0 (Life Technologies).\n\nAnimal maintenance and infection {#Sec23}\n--------------------------------\n\nAll animal experiments were approved by the local animal welfare committee (\"Nieders\u00e4chsisches Landesamt f\u00fcr Verbraucherschutz und Lebensmittelsicherheit/ LAVES\") and performed according to their guidelines.\n\nhIL-3/GM-CSF KI mice (C;129S4-*Rag2*^*tm1.1Flv*^ *Csf1*^*tm1(CSF1)Flv*^ *Csf2*/*Il3*^*tm1.1(CSF2,IL3)Flv*^ *Thpo*^*tm1.1(TPO)Flv*^ *Il2rg*^*tm1.1Flv*^ Tg(SIRPA)1Flv/J) mice^[@CR19]^ were obtained from the Jackson Laboratory and housed in the central animal facility of Hannover Medical School. Immunodeficient mice were maintained under pathogen free conditions in individually ventilated cages (IVC) with free access to food and water.\n\nFor the experiments we used male and female animals between 10 and 26 weeks of age. The animals were randomized between the groups to ensure a homogenous mixture of male/female animals and different ages. Only animals with a disease score of \u22641 were included in the study, no animals were excluded from the analysis. For infection with *P. aeruginosa* or pulmonary iPSC-Mac transplantation, anesthetized (ketamine/midazolam) mice were instilled via the trachea after oral intubation. For the simultaneous infection experiments, iPSC-Mac (4\u2009\u00d7\u200910^6^/animal) and PAO1 (0.2\u2009\u00d7\u200910^5^ CFU/per animal) were resuspended in PBS and mixed in a total volume of 60\u2009\u00b5l. For solely infected mice the same CFU was applied in 60\u2009\u00b5l PBS. To avoid phagocytosis before instillation, cell/bacteria mixes were kept on ice all the time. For the therapeutic PiMT experiments, hIL-3/GM-CSF KI mice were infected with 0.3\u2009\u00d7\u200910^5^ CFU PAO1 (in a volume of 30\u2009\u00b5l PBS) after anesthesia with ketamine/midazolam. Control mice received the same volume PBS. After 4\u2009h, mice were anesthetized by isoflurane inhalation and the second instillation with 50\u2009\u00b5l PBS or 4\u2009\u00d7\u200910^6^ iPSC-Mac in PBS was performed. Control mice again received the same volume PBS. iPSC-Mac were freshly harvested from suspension cultures after harvest \u22652. The disease score of the animals was assessed using a scoring matrix^[@CR43]^. The investigators were not blinded during the experiments and analysis. After 24\u2009h, animals were killed and end analysis was performed.\n\nTrajectory of mouse activity was analyzed by video documentation and manual tracking. The video was recorded using a handheld mobile camera in sterile housing and stabilized by optical flow methods. Mice motion was then tracked with the help of manually annotated keypoints on the animals' heads for a time slot of 22\u2009s. Total trajectory is given by accumulating all keypoints.\n\nMurine lung function {#Sec24}\n--------------------\n\nNon-invasive head-out spirometry investigating 14 lung function parameters was performed on conscious restrained mice^[@CR24]^. Mice were positioned in glass inserts, their breathing causes air to flow through a pneumotachograph. A pressure transducer creates an electrical signal, which is analysed using NOTOCORD HEM software (Version 4.2.0.241, Notocord Systems SAS, Croissy Sur Seine,France). The parameters of tidal volume (measured in mL), expiratory time, inspiratory time, time of inspiration\u2009+\u2009expiration, relaxation time (all measured in ms) and the flow at 50% of the expiratory tidal volume (EF50) were selected to characterize murine lung function during infection.\n\nBroncho-aleveolar lavage {#Sec25}\n------------------------\n\nBronchoalveolar lavage was performed by cannulating the murine trachea post-mortem. The right lung was rinsed with 1\u2009mL of PBS for three times.\n\nMeasurement of hemoglobin levels in BALF {#Sec26}\n----------------------------------------\n\nFresh BALF was used for photometric analysis of hemoglobin. After this, BALF samples were centrifuged and supernatants were stored at \u221280\u2009\u00b0C for Luminex analysis. Pellets were fixed and stained for flow cytometry.\n\nLung bacterial numbers {#Sec27}\n----------------------\n\nThe right lungs of the killed mice were ligated, resected, and homogenized with a tissue homogenizer (Polytron PT 1200, Germany). Total bacterial numbers were assessed from serial dilutions of the homogenates, which were cultured on Luria--Bertani plates using the drop plate method^[@CR42]^.\n\nHistology {#Sec28}\n---------\n\nAt the dedicated time points the animals were killed. Right lungs were filled with OCT buffer and fixed in neutral buffered 4% PFA for 3 days at 4\u2009\u00b0C. For control animals, left lungs were used. Tissues were trimmed according to the RITA-Guidelines^[@CR44]^, dehydrated (Shandon Hypercenter, XP) and subsequently embedded in paraffin (TES, Medite). Sections (2--3\u2009\u03bcm thick, microtom Reichert-Jung 2030) were deparaffinized in xylene and H&E stained according to standard protocols. Blinded evaluation (Axioskop 40, Zeiss microscope) and histological scoring of the sections was performed as described before^[@CR45]^ by a trained pathologist.\n\nStatistics {#Sec29}\n----------\n\nGraphPad Prism 6 and 7 was applied to perform unpaired Student's *T*-test or analysis of variance (ANOVA). For experiments *n*\u2009\\>\u20093, mean\u2009\u00b1\u2009SD is given, for experiments *n*\u2009\\<\u20093, mean\u2009\u00b1\u2009s.e.m. is plotted. For all experiments, individual datapoints are shown. Asterisks denote: \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009*\\<*\u20090.01; \\*\\*\\**P*\u2009*\\<*\u20090.001; \\*\\*\\*\\**P*\u2009\\<\u20090.0001.\n\nStudy approval {#Sec30}\n--------------\n\nFor the isolation of peripheral blood mononuclear cells: healthy donors gave written informed consent according to the local ethical committee at Hannover Medical School (Ethics Committee Hannover Medical School). For animal experiments: all animal experiments were approved by the animal welfare committee of lower Saxony (\"Nieders\u00e4chsisches Landesamt f\u00fcr Verbraucherschutz und Lebensmittelsicherheit/ LAVES\") and performed according to their guidelines.\n\nElectronic supplementary material\n=================================\n\n {#Sec31}\n\nSupplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Movie\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThese authors contributed equally: Mania Ackermann, Henning Kempf, Antje Munder, Nico Lachmann.\n\nElectronic supplementary material\n=================================\n\n**Supplementary Information** accompanies this paper at 10.1038/s41467-018-07570-7.\n\nThe authors thank Doreen L\u00fcttge, Theresa Buchegger, Yannik D\u00fcnow, Annika Franke, Gerhard Preiss, and Silke Hedtfeld (all Hannover Medical School) for assistance. Moreover, we are thankful to Christina Kropp and Caroline Halloin for assistance with bioreactor setup and maintenance, Janita L\u00fchrs, Laura Boge, and Sabine Schild (all Fraunhofer ITEM, Hannover) for performance of the phagocytosis as well as Andreas P\u00f6sch (Leibniz University Hannover) for help with image analysis. The authors also thank Oliver Dittrich-Breiholz and Heike Schneider (Research Core Unit Transcriptomics, Hannover Medical School) for performing microarray analysis as well as M. Ballmaier and C. Struckmann (Core Unit Cell Sorting, Hannover Medical School) for FACS. The authors thank Thomas Bjarnsholt, University of Copenhagen, for kindly providing the GFP-tagged PAO1. In addition, the authors thank B. T\u00fcmmler (Hannover Medical School) for discussions and scientific input to the infection experiments. The authors would like to thank Michael Morgan (Hannover Medical School) for review of the manuscript. The authors would also like to thank all donors for donating blood. This work was supported by grants from the Deutsche Forschungsgemeinschaft (Cluster of Excellence REBIRTH; Exc 62/1 to T.M., R.Z.,\u00a0U.M., A.S., and N.L. as well as DFG LA 3680/2-1 and ZW64/4-1), the Else-Kr\u00f6ner-Fresenius-Stiftung (EKFS 2013_A24 to T.M., EKFS; 2015_A92 to N.L. and EKFS; 2016_A146 to M.A.), the German Ministry for Education and Science (BMBF 13N12606,13N14086 and iMACnet 01EK1602A), StemBANCC (support from the Innovative Medicines Initiative joint undertaking under grant 115439-2, whose resources are composed of financial contribution from the European Union \\[FP7/2007-2013\\] and EFPIA companies' in-kind contribution), TECHNOBEAT (European Union H2020 grant 668724), the Joachim Herz Stiftung (to N.L., H.K., and M.A.), and MHH Hannover (HiLF grant to M.A. and H.K. as well as and support from the young academy to N.L.). M.P.K.\u00a0and A.S. were\u00a0funded by the SFB738 \"Optimization of conventional and innovative transplants\".\n\nM.A., H.K., R.Z., A.M., and N.L. designed the study, wrote the paper, and performed experiments. C.H., M.H., A.R.H., K.B., J.W.S., K.H., M.P.K., S.G., and C.F. performed experiments and analyzed data. A.S., T.M., S.W., and D.J., K.S., U.M. provided conceptual advice, discussed results, and edited the manuscript.\n\nThe authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information files or from the corresponding author on reasonable request. Microarray data have been deposited in the ArrayExpress database under accession number [E-MTAB-5436](https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-5436/).\n\nCompeting interests {#FPar1}\n===================\n\nPart of this work is included in a patent application. M.A., H.K., T.M., R.Z., and N.L. are authors of the patent application (European patent application number PCT/EP2018/061574) entitled \"Stem-cell derived myeloid cells, generation and use thereof\". The priority date of the application is 04.05.2017. All the remaining authors declare no competiting interests.\n"}
+{"text": "![](indmedgaz73198-0104){#sp1 .601}\n\n![](indmedgaz73198-0105){#sp2 .602}\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nEndometriosis is a common condition that affects as many as 10%--20% of women of reproductive age.[@ref1] Endometrial implants are observed most commonly on the ovaries, and in the anterior and posterior cul-de-sacs, broad ligaments, and uterosacral ligaments.[@ref2] Commonly, most patients have pelvic pain. However, urinary tract complications are uncommon, occurring in approximately 1--2% of affected patients.[@ref3] Medical treatment of fibrotic tissue secondary to endometriosis is ineffective. Laparoscopy is a reasonable choice for treatment of severe pelvic endometriosis, assuming that the surgeon is highly skilled in this technique as laparoscopic radical hysterectomy.[@ref4] We report a patient with hydronephrosis by severe left urinary tract endometriosis and asymptomatic loss of renal function that was initially managed laparoscopically.\n\nCase report {#sec1-2}\n===========\n\nThis case report was approved by Nasu red cross hospital institutional review board. The informed consent of this case report and associated images was obtained from the patient for the publication.\n\nA 49-year-old woman, gravida 2, para 0, came to our hospital for the screening of uterine cervical cancer once a year. Her menstrual cycle is regular. She chronically had symptoms of pelvic pain, dysmenorrhea and hypermenorrhea. T2-weighted magnetic resonance imaging (MRI) showed a 8 cm uterine fibroid, a 7 cm left ovarian endometrioma and uterosacral ligament endometriosis infiltrating the rectal wall ([Fig. 1](#F1){ref-type=\"fig\"}). Coronal section T1-weighted MRI with fat-saturation after contrast medium enhancement showed a left hydroureter ([Fig. 1](#F1){ref-type=\"fig\"}). There was no ascites. Her serum level of both cancer antigen (CA) 125; 115.0 U/ml and 19-9; 76.1 U/ml was elevated. Computed tomography (CT) of the abdomen revealed the left ureter and hydronephrosis ([Fig. 1](#F1){ref-type=\"fig\"}). 3D volume rendered images created from CT urography showed left proximal ureteral obstruction ([Fig. 1](#F1){ref-type=\"fig\"}). The renogram (^99m^Tc-MAG3) showed a residual renal function of 14% in the left kidney and 86% in the right kidney ([Fig. 1](#F1){ref-type=\"fig\"}). Endometriotic lesion was suspected of causing periureteral compression. Cystourethroscopy showed external compression 2 cm above the ureteral orifice. There was no evidence of disease within the urinary tract itself. A ureteral catheter was inserted with mild difficulty.\n\n![Colonel T2-W1 MRI shows (A) endometrioma and strong adhesion between posterior wall of uterus and rectum (hatched arrow) and (B) the left hydroureter (white arrow). (C) CT scan axial view shows hydronephrosis (white arrow) in left kidney. (D) 3D volume rendered images created from Computed tomography urography doesn't show the imaging of left ureter, however detects left renal cortex. (E) In ^99m^Tc-mercaptoacetyltriglycine (^99m^Tc-MAG3) renogram, the effective renal plasma flow of the left kidney markedly decreased and showed no peak curve. (F) Laparoscopic view demonstrating left ureter (white arrow) above the level of the uterosacral ligament (hatched arrow), which is infiltrated by endometriosis.](GMIT-6-214-g001){#F1}\n\nThe preoperative diagnosis was left ovarian endometrioma with deep infiltrating endometriosis and a leiomyoma.\n\nThe patient was evaluated and discussed by a multidisciplinary team including urology. The proposed treatment was a laparoscopic approach to perform adhesiolysis, liberation and resection of the uterus and appendages, the rectovaginal septum nodule. When it is difficult to preserve left ureter laparoscopically, left nephrectomy would be additionally performed. The patient was informed of the risks of the procedure and she received informed consent. She was made aware of the possible need for nephrectomy, depending upon the extent of disease.\n\nDuring laparoscopy, we found a frozen pelvis with significant adherence syndrome. The left ureter was entrapped by an endometriotic nodule that also affected the left uterosacral ligament ([Fig. 1](#F1){ref-type=\"fig\"}). The left uterosacral ligament had dense adhesions which were separated by blunt dissection. The retroperitoneal space was opened above the ureter and ureterolysis was performed by blunt dissection. The uterosacral ligament and a pararectal space were identified. The ureter was pushed downward and laterally. The uterine artery was also identified to prevent injury during ureterolysis. Ureterolysis was performed to a point below the uterine artery, but this creation of a ureteral tunnel had to be stopped because of injury to the descending uterine vessels. The bleeding was stopped by bipolar electrocauterization. It was extremely difficult to divide the space of where left ureter was involved into deep infiltrating endometriosis because of the intensive adhesion because there are two reasons why we estimated that its endometriosis might be regressed by the bilateral oophorectomy and needed to avoid ureteral injury.\n\nWe also resected the left endometrioma and uterus including leiomyoma. The resected specimen was delivered through the opened vaginal vault. The incised vaginal vault was closed with laparoscopically placed absorbable sutures. The patient was hospitalized for a total of seven days and her subsequent clinical course was excellent.\n\nPostoperatively, the ureteral catheter was removed 2 months postoperatively. Then, CT scan revealed the improvement of hydronephrosis, so called mild hydronephrosis, compared to preoperation ([Fig. 2](#F2){ref-type=\"fig\"}). While mild hydronephrosis was left, that was not enlarged after 1 years of follow-up.\n\n![CT scan axial view improves hydronephrosis (white arrow) in left kidney.](GMIT-6-214-g002){#F2}\n\nDiscussion {#sec1-3}\n==========\n\nDeep infiltrating endometriosis is defined as lesions penetrating surrounding tissue by 5 mm or more[@ref5] and is likely of multifactorial origin. It can affect both genital and extragenital organs. Of these, the most frequently involved are the large intestine and urinary tract. Patients usually present with multiple symptoms including abdominal pain, dysmenorrhea, chronic pelvic pain, dyspareunia, dyschezia and other digestive and urinary symptoms. The goal of treatment is to improve the patient's quality of life.\n\nThe surgical approach is often complex for various reasons, including the intrinsic difficulty of treating endometriosis and the previous (often laparoscopic) surgeries resulting in adhesions and anatomic distortion. This requires the participation of expert laparoscopic surgeons with wide experience in this disease. In our case, there was hypermenorrhea and dysmenorrhea. On the other hand, there were no symptoms as any urinary trouble despite of loss of renal function. Jadoul et al reported that the risk of loss of renal function in cases of ureteral endometriosis is 11.5%.[@ref6] Whereas, there are some cases reported with silent ureteral involvement in the context of deep infiltrating endometriosis,[@ref7][@ref8] as occurred in our patient. When renal function is below 14%, there is a risk of pyelonephritis and renovascular hypertension, and nephrectomy is recommended. Our patient had a left renal function of 14%. It is important to call the attention of professionals who treat these patients with deep infiltrating endometriosis to the possibility of silent ureteral involvement, using imaging tests to diagnose this eventuality and resolving the obstruction surgically before loss of renal function. Generally, MRI is very useful for guiding laparoscopy, and fat-saturation MRI is more accurate in the diagnosis of pelvic endometriosis.[@ref9][@ref10] Our patient had irregular infiltration of the pelvic sidewall with strong enhancement, indicating peritoneal endometriosis, a finding confirmed at laparoscopy. Deeply penetrating endometriosis generally requires surgery. Chapron and Dubuisson reported that laparoscopic surgery is efficacious for the treatment of deep infiltrating endometriosis.[@ref4] However, they also cautioned that retroperitoneal infiltration may substantially modify the normal anatomic relationships and, hence, laparoscopy must be performed with great care. In fact, they could extend the ureterolysis below the uterine artery and completely remove deep infiltrating endometriosis. In further severe cases, if stenosis is evident but limited to the ovarian fossa and distal ureter can be preserved, then ureteral resection and end-to-end anastomosis can be performed over a ureteral stent. Surgical approach is changed if the ureteral stenosis is close to the vesicoureteral junction and is extensive. The ureter is resected before the area of disease, and the proximal end is reimplanted into the bladder. However, advanced surgical techniques that are comparable with laparoscopic radical hysterectomy are required.\n\nIn our case, we created a pararectal space to identify the uterosacral ligament, and the uterine artery was also carefully identified to prevent injury. However, we could not extend the ureterolysis below the uterine artery as their procedure. We did not try to remove whole deep infiltrating endometriosis for the purpose of avoiding nephrectomy because we inferred that deep infiltrating endometriosis would be diminished by the influence of bilateral ovarian resection which means that reproductive hormone is lacked such as menopause. Optionally, that patient didn't hope to take any medication after surgery though we could choose to resect only left endometrioma and use hormone therapy such as dienogest after surgery. Consequently, we technically made an effort to preserve the left urinary tract to avoid nephrectomy.\n\nIn conclusion, our case illustrates that even in the absence of symptoms, deep infiltrating endometriosis may seriously affect surrounding structures, in particular the urinary tract. MRI and the renogram may help to diagnose pelvic endometriosis and the renal function before surgery. While laparoscopy may be adequate in the hands of a skilled operator, chronic inflammation with fibrosis may not be amenable to this less invasive technique. In patient without fertile preservation, we do not dare to try to resect whole endometriosis completely because endometriosis left would be degraded by the bilateral oophorectomy that decrease the level of reproductive hormone such as menopause.\n\nConflict of interest: The authors have no conflict of interest relevant to this article.\n"}
+{"text": "Presentation slides available from: <http://www.integratedcarenetwork.org/Sweden2008/slides/01-00-edgren.ppt>\n"}
+{"text": "1. Introduction {#s1}\n===============\n\nBenchmarking is a notoriously difficult task. Benchmarks are often created by the creators of the tools being benchmarked, resulting in biased comparisons favoring their tool. Benchmarking can be an inefficient process, as the tool being benchmarked may need changes to collect certain performance metrics. Even once that effort is undertaken, benchmarks are often run a few times and then forgotten, quickly becoming obsolete. However, benchmarks can spur progress as tool developers have an objective metric to maximize or minimize.\n\nPrior work benchmarking neural simulators and neuromorphic hardware has focused on low-level neural performance. For example, Sharp and Furber ([@B18]) showed that SpiNNaker can simulate a recurrent network of leaky integrate-and-fire neurons with similar firing rates and inter-spike intervals as the NEST neural simulator, but around six times faster. Stromatias et al. ([@B22]) showed that SpiNNaker\\'s power consumption varies between 15 and 37 Watts (0.5--0.8 Watts per chip) depending on the number of neurons being simulated. Goodman and Brette ([@B14]) showed that Brian simulated a randomly connected network of 4000 leaky integrate-and-fire neurons twice as fast as an equivalent Matlab implementation, but around three times slower than a C implementation. In all of these cases, none of the networks elicited activity that could be directly related to experimentally recorded data on a behavioral task.\n\nIn contrast, Ehrlich et al. ([@B8]) and Br\u00fcderle et al. ([@B4]) have presented a set of benchmarks that target the FACETS neuromorphic system through the PyNN Python package. These benchmarks include an attractor-based memory model, a model of self-sustained AI states, and a Synfire Chain, all of which are directly related to neuroscientific experiments. We aim to build on this line of research and provide an unbiased, efficient, and future-proof set of benchmarks that focuses on high-level functional performance using Nengo instead of PyNN. We have previously shown that Nengo is an order of magnitude faster than the software simulators that PyNN targets (Bekolay et al., [@B2]), and have recently implemented backends that target neuromorphic hardware.\n\nIn this study, we propose that the Nengo test suite can serve as an unbiased, efficient, and future-proof set of benchmarks for neuromorphic systems. Nengo is a rigorously tested software package for building and simulating large-scale neural models that can perform cognitively relevant tasks. It provides a high-level API (frontend) that can express large-scale models concisely and in a platform-independent manner. Several Nengo-compatible simulators (backends) have been developed that can run Nengo models on diverse platforms, including neuromorphic hardware. Nengo\\'s separation between frontend and backend, as well as its extensive test suite, provide standardized comparisons between different neuromorphic implementations with respect to functional performance.\n\nThe end result is a set of benchmarks that are written once but run on all backends, making new benchmarks easy to implement. Nengo\\'s primary goal is to build large-scale functional models, and for this reason its codebase contains tests of functional performance; it is therefore not intentionally biased in favor of any particular backend. While there is effort involved in creating a Nengo backend for a particular neuromorphic system, the reason to build such a backend is to leverage the large-scale modeling interface provided by Nengo; the ability to collect benchmarks is automatically provided to any Nengo backend. Finally, since the test suite exists to ensure that Nengo continues to function correctly, and is run whenever Nengo is changed using continuous integration systems, it will continue to grow and be run frequently, rather than becoming obsolete in the future.\n\nSeveral Nengo backends exist or are in active development. In this paper, we run benchmarks on five backends. The reference, distilled, and Brainstorm backends run on general purpose computers, and focus on aspects other than speed. The OpenCL backend aims to be a fast backend that still works on general purpose computers and can take advantage of graphical processing units (GPUs). The SpiNNaker backend uses SpiNNaker neuromorphic hardware to simulate models in real time for natural interaction with physical sensors and actuators.\n\nIn the subsequent sections, we detail the architecture of Nengo\\'s frontend and backend, and describe what any backend is required to implement. We describe Nengo\\'s testing framework, including explanations of the test fixtures used to collect and visualize benchmarks. We then list the metrics that are collected in this paper, and give further details on the backends benchmarked. Finally, we show and discuss the results of collecting those metrics for each backend.\n\n2. Background {#s2}\n=============\n\nThere are two key features of Nengo (from version 2.0 onward) that enable rapid benchmarking of neuromorphic systems. The first is a decoupling of model creation and simulation, resulting in a platform independent frontend interface to any backend that implements a certain set of requirements. The second is a functional test suite that ensures Nengo can be used to create models that solve cognitive tasks. The test suite makes ample use of test fixtures to enable data collection while testing on arbitrary backends.\n\n2.1. Nengo architecture\n-----------------------\n\nNengo has a strict separation between frontend and backend. The frontend exposes a modeling interface that uses Python to define models concisely. Backends are responsible for transforming those frontend objects into code that can be run on the target platform. While each backend must be exposed through Python, this requirement does not significantly limit flexibility in the backend. A backend can be implemented in C and exposed through Python bindings, or can be run as a separate process managed by the Python backend, with data transmitted to Python through sockets or other inter-process communication protocols.\n\n### 2.1.1. Frontend interface\n\nNengo contains five frontend objects that validate and store symbolic information about the neural model to be simulated. All neural models defined with Nengo are built with these five building blocks, including Spaun, a model that performs eight cognitive tasks with 2.5 million neurons that use visual input to produce motor output (Eliasmith et al., [@B11]). A backend only needs to be aware of these five objects to run Nengo models, and therefore run the benchmarks discussed in this paper.\n\nThe primary abstraction in Nengo is the *Ensemble*, which is a group of neurons. The activity of these neurons is generally taken to be a distributed representation of a numerical state vector; for example, a group of 100 neurons might represent an agent\\'s location in three-dimensional space, and as the agent moves, the pattern of neural activity will change accordingly. Two parameters are mandatory: the number of neurons, and the dimensionality of the vector being represented. There are several optional parameters that affect how the neurons represent the vector space; encoders map the vector space into currents to be injected in the neurons, maximum firing rates can be specified for each neuron, the type of neuron model can be specified, and so on.\n\nThe *Node* provides a structure for all non-neural aspects of a model. *Nodes* can provide input to a system, collect output values, interface with physical sensors and actuators, or provide any other computation needed in a model. In the case of the benchmarks presented in this paper, *Nodes* are used to provide input signals to *Ensembles*.\n\nThe *Connection* connects two objects (e.g., *Ensembles* or *Nodes*) together. The two objects being connected are the only mandatory parameters. The synapse model filtering the connection, a function applied to the vector communicated across the connection, and one or more learning rules can optionally be specified.\n\nThe *Probe* provides the main mechanism for data collection during run time by denoting that a particular quantity in the simulation should be recorded at a particular rate. The object to be probed must be specified. The attribute of that object, a sampling rate, and a synapse model for filtering can optionally be specified.\n\nAn important distinction between typical parameters in a neural simulation and those specified in Nengo is that Nengo parameters can be stochastic, and are not guaranteed to be supported by every backend. The vast majority of numeric parameters, for example, are often specified as probability distributions; an ensemble\\'s maximum firing rate defaults to a uniform distribution, but this could be set to a Gaussian distribution or a discrete set of options and associated probabilities. Other parameters, most notably an ensemble\\'s neuron model, may not be supported by a particular backend, which may raise an error. Even if a particular neuron model is supported, it may be implemented differently by each backend; Nengo does its best to approximate the high-level specification regardless of the neuron model that is actually used.\n\nFinally, the *Network* is a container for the other four objects. The only parameter on the network is an optional integer seed; setting this should make all of the random factors in the model deterministic, which is important for testing and debugging models, but is not used for benchmarking. The network is also responsible for maintaining a network-specific set of default parameters for the four other objects. This results in shorter and less error-prone model creation scripts.\n\nAn example script showing how Nengo can express a functional network with 500 neurons and 6 connections (representing 50,000 connection weights and 100 direct current injection sites) in under 20 lines of code can be seen in the Appendix.\n\nIt is important to note that, although the Nengo frontend is designed to make large-scale networks in relatively few lines of code, it does not impose any constraints that would preclude the creation of detailed small-scale networks. While we will only show examples of functional connection between ensembles, it is also possible to make direct neuron-to-neuron connections; functional and direct connections can coexist in the same network, as is commonly done in situations requiring direct inhibitory currents. With both types of connections, it is possible to create any network topology in Nengo.\n\n### 2.1.2. Backend requirements\n\nThe role of the backend is to take a single network, which contains ensembles, nodes, connections, and probes, and construct the backend-specific objects necessary to implement the model specified by that network. That implementation is exposed to Python through a *Simulator* object, which has three required methods and one required attribute.\n\nThe first method is `__init__`, which is a special Python method for initializing objects. This method must accept a network as its first argument. It can then accept optional arguments depending on the capabilities and requirements of the backend; the reference backend, for example, accepts `dt`, which is the length of each timestep. The purpose of `__init__` is to set up the low-level system that implements the high-level objects contained in the provided network. In the reference backend, this involves sampling from the distributions in all of the parameters of all objects, solving for decoding weights and connection weights, setting up data structures for probed data, and so on.\n\nThe second method is `run`, which advances the simulation by the number of seconds passed in as a required argument. Whatever low-level structure was created in `__init__` advances forward for that many real or simulated seconds, and importantly, any probed data is sampled according to its sample rate. In the reference backend, this advances the simulation by `time/dt` timesteps, where `dt` is fixed at the start of the simulation and cannot change; other backends may have variable-length timesteps.\n\nThe third method is `trange`, which returns a sequence of times that correspond to the times at which data was probed. In the reference backend, `trange` always returns a straightforward sequence of increasing multiples of `dt`; other backends may simulate at variable rates, though the sequence of times must increase monotonically.\n\nFinally, the simulator object implemented by each backend must have an attribute called `data`, which exposes a dictionary-like interface to the data being probed over the course of the simulation. The data dictionary should, at a minimum, provide a mapping from probe instances to the probed data; in other words, probe instances are keys and the probed data are values.\n\nThese three methods and one attribute make up the public interface that Nengo tests expect from backends. While the existence of these methods and attribute are necessary for every Nengo model, backends are not required to implement all of the neuron models or advanced features that are available in all other backends. Backends are responsible for informing the user when a model cannot be implemented on that backend. The compliance metric (see Section 3.1) explicitly tracks what types of models can be implemented by each backend. The details of how each backend implements the high-level frontend objects are out of the scope of this paper.\n\n2.2. Functional testing\n-----------------------\n\nAs of September 2015, Nengo\\'s full test suite contains 643 tests. Many tests are unit tests that ensure the frontend API operates as expected. 289 tests construct a simulator instance, run the simulator, and test the output of that model (as exposed through `trange` and `data`) and therefore can be considered \"functional\" tests. For the remainder of this paper, we will focus only on these 289 functional tests.\n\nUnlike traditional software testing, there can be significant variability in many aspects of a Nengo model. Many model parameters, for example, can be randomly generated; other aspects of a model, such as injected noise, are necessarily random. Noise is a fundamental property of neuromorphic systems. The accuracy of any large-scale model is dependent on many factors, including the number of neurons used to implement a particular task. For these reasons, functional tests can only ensure that the backend implements the system described by the frontend well enough. Each test must determine what \"well enough\" means for that particular network. While this introduces some subjectivity to testing, we believe this is an acceptable consequence of being able to test across multiple backends. When adapting these functional tests for benchmarking, we record the actual accuracy rather than ensuring that it is within some tolerances.\n\nNengo\\'s test suite employs the `pytest` testing framework[^1^](#fn0001){ref-type=\"fn\"}. `pytest` enables expressive testing of pluggable components (such as backends) through what are called \"test fixtures.\" Fixtures are exposed to the test suite as arguments that can be provided to any test function. `pytest` inspects the function signature, and passes an appropriate value to the function. This allows for boilerplate code to be specified once, and run for any test that might use it. A demonstration of how these fixtures are used in test scripts can be seen in the Appendix.\n\n### 2.2.1. Test fixtures\n\nNengo\\'s test suite defines several test fixtures to support collecting and visualizing performance metrics on multiple backends.\n\n`Simulator` allows tests to be run with multiple backends. As detailed in Section 2.1.2, each backend must expose a simulator class that accepts a network. Every functional test uses this test fixture rather than an actual class so that the backend-specific simulator class is used when the test is run. Each backend\\'s test suite loads the tests defined in Nengo, but replaces the implementation of this fixture with their own fixture returning that particular backend\\'s simulator class.\n\n`plt`, `logger`, and `analytics` allow tests to save artifacts from test runs. These artifacts are how accuracy and speed metrics are collected, and how the figures in this paper are generated. The `plt` fixture exposes the Matplotlib (Hunter, [@B16]) `pyplot` interface to a test function. Test functions can then analyze and plot the data generated in a simulation (exposed from a simulator\\'s `data` attribute) to visualize the activity in a network. Figures are saved in a specified or backend-specific directory. The `logger` fixture exposes a logging interface to save arbitrary text to a specified or backend-specific directory; these are useful for summary statistics that can inspected manually. For large amounts of data requiring more analysis, the `analytics` fixture exposes an interface to save arbitrary data (in the form of NumPy arrays) to specified or backend-specific directories.\n\n`analytics_data` enables comparative testing between two or more runs of the same test. The runs might represent multiple versions of the same backend---one before a speedup and one after---or multiple entirely different backends. The `analytics_data` fixture provides the results from those two test runs to a single test function, allowing for comparisons between the two results saved with the `analytics` fixture. See the Appendix for a concrete example of using the `analytics_data` fixture.\n\n3. Methods {#s3}\n==========\n\nBenchmarking involves collecting performance metrics for two or more comparable systems. We collect three metrics in this study, as a proof of concept that the Nengo test suite can be used to collect meaningful benchmarks for neural simulators and neuromorphic hardware.\n\n3.1. Metrics collected\n----------------------\n\n### 3.1.1. Compliance\n\nCompliance is the number of tests that a particular backend passes successfully, relative to the reference backend, which must pass all tests. While this metric can give an indication of how many features a particular backend supports, it does not take into account how commonly used a particular feature is; for this reason, a backend that has relatively low compliance can still be useful in many situations.\n\n### 3.1.2. Accuracy\n\nAccuracy measures how well a particular backend implements a desired model. The actual accuracy metric depends on the desired model, but in simple cases can be the root mean squared error (RMSE) between a desired signal and the actual signal decoded from an ensemble.\n\n### 3.1.3. Speed\n\nSpeed measures the amount of time it takes for a backend to run a model. Primarily, we are concerned with the amount of time taken in `run` calls; however, we also separately measure how long each backend spends building each model (i.e., the time spent in `__init__`), as models must be built before being run.\n\nIt should be noted that during a `run` call, each backend incurs some overhead to manage the simulation in addition to actually moving the simulation forward. The amount of overhead depends on the backend; backends making use of hardware other than the CPU (like the SpiNNaker and OpenCL backends) are likely to incur more overhead than those using the same CPU core as the Nengo frontend. However, we include the overhead in our speed benchmarking because it is not possible to distinguish overhead from actual run time in a backend-agnostic way.\n\n3.2. Test models\n----------------\n\nUnlike the compliance metric, accuracy and speed are only collected for tests that define an accuracy metric (i.e., those that use the `analytics` fixture). In this study, we focus on four such tests that address central functioning aspects of a wide variety of large-scale brain models (see Eliasmith and Anderson, [@B10]; Eliasmith, [@B9]).\n\nEach test has several parameters that can be varied in order to benchmark a wide variety of cases (see the Appendix for an example). We test each model with parameters typically used in large-scale models, but make these parameterized models available at <https://github.com/ctn-archive/bekolay-fnme2015> for those who wish to explore additional cases.\n\n### 3.2.1. Communication channel chain\n\nIn this model, five ensembles, comprised of 100 leaky integrate-and-fire (LIF) neurons each (500 total), are connected in series, with a communication channel (i.e., the identity function) computed across each connection. This model, therefore, attempts to faithfully communicate an input signal to the last ensemble in the chain. The input signal used is a static vector. This model tests how robust each backend is to the noise introduced by representing a vector space in the activity of spiking neurons.\n\n### 3.2.2. Two-dimensional product\n\nIn this model, the scalar product is computed from a two-dimensional ensemble comprised of 100 LIF neurons. Since the product is non-linear, we use a space-filling curve (the Hilbert curve; Hilbert, [@B15]) as the input signal to ensure that we sample the entire two-dimensional space. This model tests each backend\\'s ability to compute a non-linear function, albeit a low-dimensional one. Despite the low dimensionality, two-dimensional products are frequently used in large-scale models; it is the primary operation needed to support symbol-like processing using neuron-like elements, for example (Eliasmith, [@B9]).\n\n### 3.2.3. Controlled oscillator\n\nIn this model, a three-dimensional ensemble is recurrently connected such that the first two dimensions continuously traverse a limit cycle (i.e., they implement a cyclic attractor), and the third dimension controls the speed and direction of the oscillation. As input, we provide an initial stimulus to start the oscillation, and provide a control signal to cause oscillation at 2, 1, 0, --1, and --2 Hz. Negative frequencies indicate oscillations in the reverse direction[^2^](#fn0002){ref-type=\"fn\"}. In total, the model is comprised of 600 LIF neurons. This model tests each backend\\'s ability to stably implement a dynamical system, which is required for many cognitive functions like working memory and motor control.\n\n### 3.2.4. Basal ganglia sequence\n\nIn this model, brain structures known collectively as the basal ganglia are constructed from 4900 LIF neurons, and are organized such that they iterate through a repeating set of actions. This model has been used to investigate action selection and learning (Stewart et al., [@B19]) and the switching time between actions has been closely mapped to human decision making (Stewart et al., [@B20]). This benchmark tests the ability to construct this model, and evaluates the time needed to transition between actions.\n\nWe also test an alternate version of the basal ganglia sequence model in which some \"passthrough\" nodes are pruned from the model. Passthrough nodes are nodes that collect signals from several sources, and pass them to other objects unchanged. They act as hubs that group together related signals to reduce the number of connections that must be made to a group of related objects. Since they do no processing, they can be difficult to deal with in backends that are designed to simulate neurons quickly. The basal ganglia makes liberal use of passthrough nodes, so we also test a version of the model with most passthrough nodes removed.\n\nWe have used relatively small models run for short times in order to run many iterations on all backends. However, all of these models could be made significantly larger by increasing the number of neurons used (i.e., adjusting the `n_neurons` parameter on ensembles), and increasing the dimensionality of the signals represented in the model. Additionally, existing tests of the semantic pointer architecture within Nengo use significantly more neurons than the models presented here, and could be adapted into benchmarks.\n\n3.3. Backends tested\n--------------------\n\nWe collected the three benchmark metrics on five Nengo backends.\n\n### 3.3.1. Reference (`nengo`)\n\nThe reference backend is designed to run quickly on any general purpose computer by using NumPy (Van Der Walt et al., [@B23]) for fast vectorized computations. It is included with the Nengo frontend as `nengo.Simulator`. The reference backend offers the most features, but does not target specialized hardware.\n\n### 3.3.2. Distilled (`nengo_distilled`)\n\nThe distilled backend is intended as a teaching tool, and as a template for building new backends. It can also run on any general purpose computer, and also uses NumPy for fast vectorized computations, but does not aim to implement all of the features implemented by the reference backend, and omits some optimizations that obfuscate code. Therefore, the distilled backend is easier to read and suitable for learning about Nengo, but is expected to be slower in terms of run time.\n\n### 3.3.3. OpenCL (`nengo_ocl`)\n\nThe OpenCL backend uses the Open Computing Language (OpenCL; Stone et al., [@B21]) to run Nengo models on many different computing devices, including graphical processing units (GPUs), and field-programmable gate arrays (FPGAs). In contrast to the distilled backend, it is designed to run Nengo models as quickly as possible, using fast general purpose computing devices like GPUs and any optimizations available, at the cost of code readability.\n\n### 3.3.4. Brainstorm software (`nengo_brainstorm`)\n\nThe Brainstorm backend is a software implementation of a new neuromorphic chip based partly on Neurogrid (Benjamin et al., [@B3]) currently in development by the Brains in Silicon lab at Stanford University[^3^](#fn0003){ref-type=\"fn\"}. The software backend does not aim for speed; instead, it attempts to emulate the proposed hardware in order to test its applicability for large-scale neural models. If the emulated hardware can perform well, then it follows that actual hardware will also perform well, but will be much faster.\n\n### 3.3.5. SpiNNaker hardware (`nengo_spinnaker`)\n\nThe SpiNNaker backend (Mundy et al., [@B17]) targets the eponymous neuromorphic hardware developed by Furber et al. ([@B12]). In contrast to the Brainstorm backend, this backend targets physical neuromorphic hardware, and therefore is concerned both with accuracy and speed.\n\nUnlike software backends, considerable effort is taken to translate a Nengo model into something that can run on a SpiNNaker board. Fortunately, SpiNNaker can be reprogrammed as it is composed of a large collection of chips, each with 18 ARM processing cores. Notably, this allows SpiNNaker to take advantage of Nengo\\'s encoding and decoding capabilities, greatly reducing the amount of RAM needed (Mundy et al., [@B17]). The SpiNNaker backend accomplishes this by using decoded values to determine changes to the input currents of each neuron, rather than using purely spike-based transmission. However, neuromorphic hardware that can only communicate through spikes can be made to interact with Nengo through explicit encoding and decoding processes on the chip, as was done in Galluppi et al. ([@B13]), or by using Nengo on the host machine to explicitly generate spike patterns that are communicated to the device, as was done in Choudhary et al. ([@B5]). However, these methods may introduce additional noise in the simulation, resulting in inaccuracies.\n\n3.4. Benchmarking environment\n-----------------------------\n\nAll benchmarks were run 50 times on a server running Debian sid. The server was configured with two Intel Xeon E5-2650 CPUs clocked at 2.00 GHz, four Nvidia Tesla C2075 GPUs, and 64 GB of RAM. A single 48-chip SpiNN-5 board was connected directly to a 1 Gb/s Ethernet port on the Supermicro X9DRG-QF motherboard for the SpiNNaker benchmarks.\n\nThe versions of software used for benchmarking can be found in Table [1](#T1){ref-type=\"table\"}.\n\n###### \n\n**Software versions used for benchmarking**.\n\n  **Package**          **PyPI name**           **Version number**\n  -------------------- ----------------------- -----------------------------------------\n  Nengo                `nengo`                 Development version, commit `7d2d24145`\n  Distilled backend    `nengo_distilled`       0.1.0\n  OpenCL backend       `nengo_ocl`             0.1.0\n  Brainstorm backend   Not available on PyPI   Development version, commit `ac3cfa708`\n  SpiNNaker backend    `nengo_spinnaker`       0.2.4\n  NumPy                `numpy`                 1.8.1\n\n*The PyPI name is the unique identifier used for installing the Python package through the Python Package Index (PyPI). Packages can be found by visiting <https://pypi.python.org/pypi/>\\<PyPI name\\>*.\n\n4. Results {#s4}\n==========\n\n4.1. Compliance\n---------------\n\nTable [2](#T2){ref-type=\"table\"} gives compliance results for all backends. Compliance has been separated into tests that use a backend\\'s default neuron type (leaky integrate-and-fire in all of the backends tested), and tests that are parameterized by neuron type. Of the non-reference backends, the OpenCL backend has the highest compliance. The distilled backend has lower compliance than OpenCL, which follows from its stated goal of being simple but not necessarily feature-rich. The Brainstorm backend fails one more test than the distilled background; however, it is still an early experimental test bed, and is not necessarily indicative of the final Brainstorm hardware backend being less compliant than other backends. Finally, the SpiNNaker backend passes the fewest tests.\n\n###### \n\n**Compliance of the five backends**.\n\n  **Backend**                       **Non-parameterized tests passed**   **LIF tests passed**   **Other tests passed**\n  --------------------------------- ------------------------------------ ---------------------- ------------------------\n  Reference (`nengo`)               165 (100%)                           31 (100%)              93\n  OpenCL (`nengo_ocl`)              147 (89.0%)                          29 (93.5%)             29\n  Distilled (`nengo_distilled`)     95 (57.6%)                           15 (48.4%)             15\n  Brainstorm (`nengo_brainstorm`)   94 (57.0%)                           15 (48.4%)             15\n  SpiNNaker (`nengo_spinnaker`)     88 (53.3%)                           7 (22.6%)              0\n\n*Compliance is the number of functional tests passed by a particular backend. We have divided these tests into tests parameterized by neuron type, and those not parameterized. Non-parameterized tests are either designed to work with only one neuron type, or use leaky integrate-and-fire (LIF) neurons, which are implemented on all backends. For parameterized tests, we have separated the compliance for those tests using LIF neurons, which all backends implement, and for any additional neuron types implemented. Since there are 31 parameterized tests, the maximum compliance is 31 times the number of neuron types implemented. The reference backend implements three additional neuron types, while OpenCL, Distilled, and Brainstorm implement one additional neuron type, and SpiNNaker only implements LIF neurons*.\n\nIt should be noted that these test suites are being run on the distilled, Brainstorm, and SpiNNaker backends for the first time in this study. In other words, these results are the first objective interrogation of each backend\\'s feature set. In many cases, tests failed by backends are tests of features recently added to Nengo, such as stochastic processes for injecting current noise. As such, we expect compliance on all backends to rise quickly as backend developers implement these new features. The OpenCL backend has higher compliance in large part because it is developed by the same group that develops the reference backend. However, there are some notable features missing from some backends, such as learning through plasticity rules applied to neuron-to-neuron connections, and implementation differences, such as one timestep delays on connections with no synaptic filter, that may remain even when backends are brought up to date with the reference implementation. It is likely that tests will be rewritten in the future to allow some implementation differences if they do not affect simulation accuracy.\n\n4.2. Accuracy\n-------------\n\nFigure [1](#F1){ref-type=\"fig\"} gives accuracy results for each backend on the chained communication channel model. The boxplot shows that all five backends can implement this model accurately, despite five layers of processing that each introduce noise. The Brainstorm and distilled backends have the least variability and the SpiNNaker backend has the most variability, though the median RSME is the same across all backends. One driver of these differences is in how each backend handles ensembles operating near the edge of their representational range (i.e., the radius of the ensemble). The outlier with highest RMSE represents a model instance in which the static vector target was at the extreme of the representational range; the SpiNNaker backend\\'s relatively high RMSE on this example indicates that it may not perform as well as other backends in this situation. The SpiNNaker backend uses signed fixed point numbers with 16 digits before the decimal point, and 15 digits after the decimal point; accuracy could be improved by optimizing the backend\\'s internal calculations to use as many digits as possible. For all backends, accuracy could be improved by increasing the representational range and increasing the number of evaluation points generated when solving for decoding weights.\n\n![**Chained communication channel model**. **(A)** One example instance of the model run on the reference backend; the decoded output of the five chained two-dimensional ensembles during the first 12 ms of the simulation are shown. Each ensemble reaches approximately the correct value (indicated with the solid black line) but with slight delays depending on their position in the chain. **(B)** The aggregated accuracy results for 50 trials of each backend on the chained communication channel model. The observed metric is the root mean-squared error (RMSE) between the decoded output of the last ensemble in the chain and the correct value (which is randomly selected on each trial), during the last 100 ms of the 500 ms simulation. The box shows the median and interquartile range, while the whiskers extend to the farthest data point within 1.5 times the interquartile range; outliers are shown above and below the whiskers.](fnins-09-00380-g0001){#F1}\n\nFigure [2](#F2){ref-type=\"fig\"} gives accuracy results for the two-dimensional product model. In this case, while all of the backends perform well, the SpiNNaker backend is less accurate than the other four backends. Its performance, however, is still well within acceptable ranges for the two-dimensional product model. Again, it is likely that performance at the extremes of the representational range is responsible for the SpiNNaker backend\\'s reduced accuracy, though we have not investigated this in detail. This benchmark also suggests that the reference and OpenCL backends have more variability than the Brainstorm and distilled backends.\n\n![**Two-dimensional product model**. **(A)** One example instance of the model run on the reference backend; the decoded input represented by the two-dimensional ensemble is shown. The input signal is designed to fully explore the two-dimensional vector space in the range \\[\u22121, 1\\]. **(B)** The decoded product represented by an ensemble downstream of the ensemble representing the input. It closely matches the correct product (indicated with the solid black line). **(C)** The aggregated accuracy results for 50 trials of each backend on the two-dimensional product model. The measured metric is the root mean-squared error (RMSE) between the decoded output of the ensemble representing the product and the correct product, during the whole range of inputs provided over 5 simulated seconds (after a 500 ms wait period to allow the input ensemble to reach the correct initial state). The box shows the median and interquartile range, while the whiskers extend to the farthest data point within 1.5 times the interquartile range; outliers are shown above and below the whiskers.](fnins-09-00380-g0002){#F2}\n\nFigure [3](#F3){ref-type=\"fig\"} gives accuracy results for the controlled oscillator model. The reference and OpenCL backends perform well here, as do the distilled and Brainstorm backends. The median accuracy of the SpiNNaker is the same as the distilled and Brainstorm backends, but model instances with less accuracy than the median perform poorly compared to other backends. In this example, differences in how each backend implements synaptic filtering are more pronounced than in previous examples, as the recurrent connection is responsible for the dynamics of the oscillation. Relatively large variability in the reference, OpenCL and SpiNNaker backends suggests that their synaptic filtering implementations should be examined for potential accuracy improvements.\n\n![**Controlled two-dimensional oscillator model**. **(A)** One example instance of the model run on the reference backend; the power spectrum is shown for each 2 s period, colored to indicate the target frequency given by the control signal. High power is seen only at the frequency we attempt to induce, for each of the five test frequencies. **(B)** The aggregated accuracy results for 50 trials of each backend on the two-dimensional product model. The measured metric is the similarity between the power spectrum (calculated using the Fast Fourier Transform or FFT) of the decoded output of one of the two oscillatory dimensions in the ensemble, and the ideal power spectrum, computed on a pure sine wave oscillating at the induced frequency. Similarity is computed using a normalized dot product across all frequencies. The box shows the median and interquartile range, while the whiskers extend to the farthest data point within 1.5 times the interquartile range; outliers are shown above and below the whiskers.](fnins-09-00380-g0003){#F3}\n\nFigure [4](#F4){ref-type=\"fig\"} gives accuracy results for the basal ganglia sequence model. For the version of the model with passthrough nodes, all of the backends perform similarly, except for the SpiNNaker backend. All other backends have a transition time around 43 ms, but the SpiNNaker backend has a median transition time around 51 ms; its interquartile range is also significantly larger than that of other backends. However, its performance for the version of the model with passthrough nodes removed is a closer match to the other backends. Its median transition time is around 47 ms, and its interquartile range is indistinguishable from the other backends. The large difference between the two versions of this model for the SpiNNaker backend indicates that node-to-ensemble and ensemble-to-node connections introduce additional delays that are not present on other backends. This result is to be expected, as nodes can execute arbitrary code and are therefore difficult to simulate in real time with special purpose hardware.\n\n![**Basal ganglia sequence model**. **(A)** One example instance of the model run on the reference backend; the similarity between the current representation in working memory and the six possible actions is shown (note that in the full benchmark, 10 items are used). The model quickly progresses from one action to the next, cycling back at the end of the six items. The point at which the model switches from selecting one action to another action (the transition times) are indicated with dashed gray lines. **(B)** The aggregated accuracy results for 50 trials of each backend on the basal ganglia sequence model, with all passthrough nodes included. The observed metric is the mean time to transition from one action to the next. The model is run for 4 s, allowing for approximately 90 transitions, depending on the transition time. The box shows the median and interquartile range, while the whiskers extend to the farthest data point within 1.5 times the interquartile range; outliers are shown above and below the whiskers. **(C)** The aggregated accuracy results for 50 trials of each backend on the basal ganglia sequence model, with most passthrough nodes pruned. Lines have the same meaning as in **(B)**.](fnins-09-00380-g0004){#F4}\n\n4.3. Speed\n----------\n\nFigures [5](#F5){ref-type=\"fig\"}, [6](#F6){ref-type=\"fig\"} show the build and run speeds for all backends, respectively. While build time is not critical to optimize, it is worth noting that the reference backend has consistently fast builds. The OpenCL backend is also usually fast, but can be slow for certain types of networks (e.g., the chained communication channel model). The distilled and Brainstorm backends---which have a very similar build process---become slow when dealing with moderately sized models (e.g., the basal ganglia sequence model, which contains 4900 neurons). The SpiNNaker backend incurs an unavoidable cost in setting up the hardware (determining the placement of computational resources, and generating routes to connect resources), though it builds the largest model faster than the distilled and Brainstorm backends.\n\n![**Build time for each model on each backend**. Each bar represents the mean build time across 50 instances of each model; error bars are bootstrapped 95% confidence intervals. BG Sequence ^\\*^ refers to the basal ganglia sequence model in which passthrough nodes have been removed.](fnins-09-00380-g0005){#F5}\n\n![**Run speed for each model on each backend**. Each bar represents the mean run speed across 50 instances of each model; error bars are bootstrapped 95% confidence intervals. Run speed is measured as the simulation time plus overhead, relative to real time; e.g., a value of two indicates that calling `run` for a number of seconds takes the model two times that number of real seconds to complete the `run` call. BG Sequence ^\\*^ refers to the basal ganglia sequence model in which passthrough nodes have been removed.](fnins-09-00380-g0006){#F6}\n\nThe run time results are of the utmost importance when evaluating backends for different applications, especially those that require real-time interaction, such as robotics. In general, all backends perform adequately for the three smaller models. The OpenCL backend is significantly slower for small models, like the product and controlled oscillator models. SpiNNaker, on the other hand, is slower than other backends for these models because its goal is to run in real-time even if it is possible to run faster.\n\nThe most important result is for the largest model, the basal ganglia sequence. In this model, the distilled and Brainstorm backends performed poorly, though neither has speed as a primary goal. The reference backend also did not perform well, operating at nearly six times slower than real-time. The OpenCL backend performed better, operating at around three times slower than real-time. The best performance was seen by the SpiNNaker backend, however, which operated at around 1.2 times real-time (including overhead) on this moderately-sized model. Since the SpiNNaker board always runs at real-time, this means that around one-sixth of the time taken for the `run` call is overhead when running this model for 10 s.\n\nWhile passthrough nodes were removed in the basal ganglia sequence model for improved accuracy, it is interesting to note that speed is also impacted in this version of the model. While the distilled and Brainstorm backends are unaffected, both the SpiNNaker and OpenCL backends run faster when passthrough nodes are removed. This result indicates that passthrough nodes contribute to SpiNNaker\\'s overhead. The reference backend, on the other hand, runs slower, indicating that the computational costs of the additional connections introduced when removing passthrough nodes are greater than the costs of the passthrough nodes.\n\n5. Discussion {#s5}\n=============\n\nThe most important finding from these benchmarks is that significant speedups can be gained by running models on specialized hardware and GPUs, with little to no cost in accuracy. The conditions under which the OpenCL backend performs slower than the reference backend (e.g., the product and controlled oscillator models) warrant closer inspection to determine the source of the slowdown.\n\nDifferences in accuracy benchmarks are also important for individual backends. The improved accuracy for the SpiNNaker backend when passthrough nodes are removed from the basal ganglia model has resulted in ongoing work to automatically remove passthrough nodes in the build phase of the SpiNNaker backend. Similarly, we will investigate instances of the product model in which the SpiNNaker backend is less accurate than other backends, though we suspect that its lack of floating point hardware may be responsible. We also plan to investigate why the reference and OpenCL backends have higher variability than other backends on the simple feedforward models.\n\nIt should be noted that although there are clear relative speed differences between backends, these models are very small---each would only be a small component of a real large-scale model. Spaun, for example, includes multiple copies or larger versions of all four of the models benchmarked here in its 2.5 million neuron network. Additionally, in this study, we run these models for very small amounts of time---the longest simulation runs for 10 simulated seconds. Real models run for longer amounts of time to gather data comparable to data gathered in neuroscientific experiments. However, we expect that the results would be similar in larger models run for longer amounts of time. The distilled and Brainstorm backends should perform slowly, the OpenCL backend should run faster than the reference backend, and SpiNNaker will continue to run at real-time.\n\nAbove and beyond the benchmarking results, however, we believe that the primary contribution of this study is to provide evidence that Nengo has a tested, stable, productive frontend that can target multiple backends, and therefore provides an attractive platform for benchmarking neuromorphic backends, and other neural simulators. While this is similar to the goal of other projects, most notably PyNN (Davison et al., [@B6]), Nengo is unique in focusing on large-scale functional simulations, rather than attempting to support detailed single-neuron models (though this capability is still possible in Nengo). Topographica (Bednar, [@B1]) plays a similar role, in that it can interact with multiple neural simulators with a high-level API, but its API focuses specifically on models of topographic maps and other sensory pathways, rather than focusing on a wider variety of functions and dynamics more generally.\n\nOne inherent weakness of using Nengo as a standard platform for benchmarking neuromorphic systems is that new benchmarking capabilities may take a long time to be standardized and developed. If one wishes to add a new metric, such as power consumption (as was done in Stromatias et al., [@B22]), we must first come to a consensus on a suitable interface to this information through Nengo. Once consensus is reached, it must be implemented and tested on all backends before benchmarks can be written using that quantity. Despite this limitation, we believe that it is possible to use Nengo to collect power consumption information, and plan to implement energy efficiency comparisons in future work.\n\nWhile we cannot claim that Nengo solves all of the complications that arise in benchmarking, we believe that it improves upon the three major issues identified in the introduction. First, benchmarking neuromorphic hardware with Nengo is less biased than hardware-specific benchmarks, because the purpose of Nengo is to make functionally interesting large-scale models. These models are typically built and tested in the reference backend, leaving little opportunity to introduce hardware-specific optimizations that can bias benchmarks.\n\nSecond, there is less wasted effort in using Nengo for benchmarking, as the only requirement for using Nengo is to develop a backend. Building a Nengo backend for a piece of neuromorphic hardware gives access to all Nengo models, which is reason enough to develop a backend. The ability to collect functionally relevant benchmarks comes \"for free\" once the backend exists.\n\nThird, these benchmarks are more likely to remain up-to-date because backend developers themselves do not have to implement or update these benchmarks; they will be implemented to test Nengo\\'s capabilities as a neural simulator, as an extension of the existing test suite. Once implemented, backends that implement the features that are used in a particular benchmark should not have to modify their own code to run those benchmarks on their own hardware.\n\nIn summary, we have implemented four benchmark models using the same testing framework used for Nengo\\'s test suite, and gathered benchmark results for five backends, including an OpenCL backend that targets GPUs and a backend that targets SpiNNaker neuromorphic hardware. All five backends were able to implement the models accurately, with the OpenCL and SpiNNaker backends simulating the largest model much faster than the reference backend. We believe that these benchmarks can be easily expanded upon to develop a suite of benchmarks that can be run by any neuromorphic hardware or neural simulator with an associated Nengo backend. Running these benchmarks in a common environment and cleanly visualizing the results could catalyze the development of neural systems that implement large-scale functional models efficiently and accurately.\n\n5.1. Data sharing\n-----------------\n\nAll of the software packages discussed in this paper are available online, except for the Brainstorm backend, which has not yet been publicly released (see Table [1](#T1){ref-type=\"table\"}). All of the benchmarks presented here, as well as the scripts used to generate the presented figures, are available at <https://github.com/ctn-archive/bekolay-fnme2015>. The results of running the benchmarks are not included in the repository; however, they can be downloaded separately at <http://dx.doi.org/10.6084/m9.figshare.1496569> to replicate the figures in this paper.\n\nAuthor contributions {#s6}\n====================\n\nTB adapted the models for benchmarking, ran the benchmarks, wrote the text of the paper, and prepared all of the figures. TS wrote initial versions of the benchmark models and edited text. CE oversaw all research activities, and edited text.\n\nFunding {#s7}\n=======\n\nNSERC Graduate Fellowships, NSERC Discovery (grant 261453), ONR (N000141310419) AFOSR (FA8655-13-1-3084), Canada Research Chairs, and Canadian Foundation for Innovation Ontario Innovation Trust.\n\nConflict of interest statement\n------------------------------\n\nThe authors of this paper are three of the seven co-founders of Applied Brain Research (ABR, Inc.), which is the sole copyright holder of Nengo and the distilled backend. While ABR reserves the right to commercialize Nengo, it is freely available for all non-commercial purposes. As of the writing of this paper, ABR has not sold any licenses to Nengo, but has used it for other revenue-generating projects. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe thank Andrew Mundy for his help running the SpiNNaker backend and providing feedback on drafts of this paper. We thank Jan Gosmann for the initial version of the product benchmark models, and for conceptualization and implementation of the `analytics` and `analytics_data` fixtures. We thank Xuan Choo for administering the server on which the benchmarks were run, Eric Hunsberger for help in running the benchmarks on the OpenCL backend, and James Bergstra for implementing most of the OpenCL backend. We thank the Brainstorm group at Stanford University for making their experimental hardware emulation backend available to us. Finally, we thank our two reviewers for their helpful feedback and recommendations.\n\n^1^Available at <http://pytest.org/>.\n\n^2^There is some evidence that neural oscillators such as locomotive central pattern generators accomplish forward and backward locomotion with the same neural mechanism (Duysens and Van de Crommert, [@B7]). While this does not mean that all neural oscillators must be able to operate in the reverse direction, we believe that the capability to traverse the limit cycle forward and backward is advantageous, and in the current design requires no additional neural resources.\n\n^3^Some details available at <http://brainstorm.stanford.edu/projects/>.\n\nChained communication channel model\n===================================\n\nimport numpy as np import nengo from nengo . utils . numpy import rmse \u00a0 def test_cchannelchain(Simulator, analytics, plt): \u00a0\u00a0\u00a0\u00a0\\# Parameters that can be varied to investigate extreme cases \u00a0\u00a0\u00a0\u00a0dims = 2 \u00a0\u00a0\u00a0\u00a0layers = 5 \u00a0\u00a0\u00a0\u00a0n_neurons = 100 \u00a0\u00a0\u00a0\u00a0synapse = nengo . Lowpass(0.01) \u00a0 \u00a0\u00a0\u00a0\u00a0with nengo . Network() as model: \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0hypersphere = nengo . dists. UniformHypersphere() \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0value = hypersphere . sample(dims, 1) . ravel() \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0stim = nengo . Node(value) \u00a0 \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0ens = \\[nengo . Ensemble(n_neurons, dimensions=dims) \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0for \\_ in range(layers)\\] \u00a0 \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0nengo . Connection(stim, ens\\[0\\]) \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0for i in range(layers - 1): \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0nengo . Connection(ens\\[i\\], ens\\[i\\\\, + \\\\, 1\\], synapse=synapse) \u00a0 \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0p_input = nengo . Probe(stim) \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0p_outputs = \\[nengo . Probe(ens\\[i\\], synapse=synapse) \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0for i in range(layers)\\] \u00a0 \u00a0\u00a0\u00a0\u00a0sim = Simulator(model) \u00a0\u00a0\u00a0\u00a0sim . run(0.5) \u00a0 \u00a0\u00a0\u00a0\u00a0for p_output in p_outputs: \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0plt . plot(sim . trange(), sim . data\\[p_output\\]) \u00a0\u00a0\u00a0\u00a0plt . plot(sim . trange(), sim . data\\[p_input\\], color='k', linewidth=1) \u00a0\u00a0\u00a0\u00a0plt . ylabel('Decoded output') \u00a0\u00a0\u00a0\u00a0plt . xlabel('Time (s)') \u00a0 \u00a0\u00a0\u00a0\u00a0last = p_outputs\\[-1\\] \u00a0\u00a0\u00a0\u00a0decoding_rmse = rmse(value, sim . data\\[last\\]\\[sim . trange() \\> 0.4\\] \u00a0\u00a0\u00a0\u00a0analytics . add_data('rmse', decoding_rmse) \u00a0 def test_compare_cchannelchain(analytics_data, plt): \u00a0\u00a0\u00a0\u00a0rmses = \\[d\\['rmse'\\] for d in analytics_data\\] \u00a0\u00a0\u00a0\u00a0plt . bar(np . arange(len(rmses)), rmses, align='center') \u00a0\u00a0\u00a0\u00a0plt . ylabel(''RMSE'')\n\n[^1]: Edited by: Jonathan C. Tapson, Western Sydney University, Australia\n\n[^2]: Reviewed by: Johannes Partzsch, Technische Universit\u00e4t Dresden, Germany; Daniel Neil, Eidgen\u00f6ssische Technische Hochschule Z\u00fcrich, Switzerland\n\n[^3]: This article was submitted to Neuromorphic Engineering, a section of the journal Frontiers in Neuroscience\n"}
+{"text": "Microbial Biotechnology (2016) 9(5), 652--657\n\n**Funding Information**\n\nResearch on polymer biotechnology at the lab of Auxi Prieto is supported by grants from the European Union\\'s Seventh Framework Programme and for research, technological development and demonstration under grant agreement no. 311815, and the Horizon 2020 research and innovation Programme under grant agreement no 633962 and no. 679050. We also acknowledge the Community of Madrid (P2013/MIT2807) and the Spanish Ministry of Economy (BIO201344878R, BIO2014\u201061515\u2010EXP).\n\nGlobal warming, market and production capacity, the key drivers for selecting the main players {#mbt212393-sec-0001}\n==============================================================================================\n\nThe unparalleled success of plastics shows no signs of slowing down, and total plastics consumption could grow from the current 250\u00a0000 kilotonnes per year to about 1 million kilotonnes by the end of this century. The Intergovernmental Panel\u00a0on Climate Change trajectory to 2050 for stabilization of atmospheric greenhouse gas (GHG) concentrations at 450\u00a0ppm CO~2~ requires an 80% reduction in emissions compared to the 1990 level (<http://www.ipcc.ch/publications_and_data/ar4/wg3/en/contents.html> and <http://unfccc.int/paris_agreement/items/9485.php>). The vast majority of plastics in current production are derived from crude oil, thus, besides the resistance to biodegradability, their GHG emissions are an environmental issue. Considering climate change as a grand societal challenge, establishing a market of bio\u2010based plastics is seen as virtuous in terms of Bio\u2010economy, due to their bio\u2010based carbon content (Philp *et\u00a0al*., [2013](#mbt212393-bib-0010){ref-type=\"ref\"}).\n\nConsequently, the demand for bio\u2010based polymers is growing worldwide. The chemical industry is continuing to produce traditional plastics but, in parallel, it is developing a broad range of bio\u2010based, biodegradable or non\u2010biodegradable, and/or compostable plastics. In this scenario, we should distinguish between two main groups of players; on the one hand, the new bioplastics based on renewable resources, either monomers or full biopolymers. Some of these are biodegradable plastics and are produced by fermentation or chemical processes, while others are produced by a combination of biotechnological and chemical processes. Examples are polyhydroxyalkanotes (PHA), polylactic acid (PLA), succinic acid and 1,3\u2010propanediol based polymers, etc. The second group comprises the drop\u2010in bio\u2010based polymers, chemically identical to their petrochemical counterparts but they have, at least partially, a component of biological origin such as the bio\u2010based polyethylene terephtalate (PET) or polyethylene (PE). The term drop\u2010in was initially used for biofuels whose specifications allow market applications with existing infrastructure and avoid important investments. Drop\u2010in plastics are non\u2010biodegradable materials, obtained from renewable raw materials that present identical technical properties to their fossil counterparts.\n\nBy 2019, worldwide production capacity of bio\u2010based plastics will grow by over 400%, or from 1.7 million tonnes in 2014 to 7.8 million tonnes in 2019 in absolute terms. The market is clearly being dominated by drop\u2010in bio\u2010based polymers and non\u2010biodegradable polymers. Drop\u2010in bio\u2010based polymers such as PET and PE lead this category. Bio\u2010based PET is the overall market leader and is expected to grow quickly, from 35.4% in 2014 to 76.5% in 2019. Consequently, the bio\u2010based non\u2010biodegradable polymers market is expected to grow strongly. PET production capacity was around 600\u00a0000 tonnes in 2014 and is projected to reach about 7 million tonnes by 2020, using bio\u2010ethanol from renewable sources. PET is currently 20% bio\u2010based and produced from bio\u2010based monoethylene glycol and terephthalic acid (TPA). The huge bio\u2010based market demand is mainly the result of the 'Plant PET Technology Collaborative' comprising Coca Cola, Ford, Heinz, Nike and Proctor & Gamble (<http://www.plasticstoday.com/five-major-us-brands-collaborating-plant-based-pet/69477799417544>). This consortium and other retailers are assisting several pre\u2010commercial enterprises to develop commercial processes for bio\u2010based TPA, which is currently still petro\u2010based but subject to ongoing research. Most of these processes target p\u2010xylene, the petrochemical precursor for TPA (Smith, [2015](#mbt212393-bib-0013){ref-type=\"ref\"}).\n\nRecently, PET 100% bio\u2010based polymer alternatives with potentially similar applications based on furanic building blocks are being developed such as Poly\u2010ethylene\u2010furanoate (PEF) and polytrimethylene furandicarboxylate (PTF). Avantium patented the technology to convert biomass into FDCA (2,5\u2010Furandicarboxylic acid), which is the monomer used to produce the bio\u2010based polyester PEF. DuPont and ADM plan to make PTF by reacting fructose\u2010derived furandicarboxylic methyl ester with 1,3\u2010propanediol (<https://cen.acs.org/articles/94/i12/BASF-Avantium-combine-bio-polyester.html>).\n\nExact figures for worldwide biodegradable bioplastics production and capacities are difficult to determine. They are mostly based on estimates, and these are constantly changing due to the rapid growth of the bioplastics industry. According to the study of the Nova\u2010Institute (<http://www.bio-based.eu/market_study/media/files/15-12-03-Bio-based-Building-Blocks-and-Polymers-in-the-World-short-version.pdf>), following the bio\u2010based drop\u2010in polymers, PHA shows the second fastest market growth rate since 2014 with projections to 2020. PHA production capacity was around 35\u00a0000 tonnes in 2014 and is projected to reach about 100\u00a0000 tonnes by 2020. PHAs are biodegradable biocompatible polyesters, which accumulate as granules in the cytoplasm of many bacteria under unbalanced growth conditions. PHA diversity (recently described as 'PHAome') has increased rapidly due to the increasing diversity of monomers, homopolymers, random and block copolymers, functional and graft polymers, molecular weights and combinations of the above. Moreover, successful manipulation of \u03b2\u2010oxidation in *Pseudomonas* spp. has provided control over the above diversity (Chen *et\u00a0al*., [2015](#mbt212393-bib-0003){ref-type=\"ref\"}). These species produce mainly medium\u2010chain\u2010length PHAs (mcl\u2010PHAs), which are promising thermo\u2010elastomers because they can be further modified by inserting functional groups in the side\u2010chains. Functionalized mcl\u2010PHAs provide modified mechanical and thermal properties, and consequently have new processing requirements and highly diverse potential applications in emergent fields such as biomedicine. However, process development and sample availability are still limited due to the toxicity of some precursors and current low productivity, which hinder investigation. Conversely, improved mutant strains designed through systems biology approaches and co\u2010feeding strategies with low\u2010cost substrates may contribute to the widespread application of these biopolymers (Prieto *et\u00a0al*., [2016](#mbt212393-bib-0011){ref-type=\"ref\"}).\n\nGrowth rates of polybutylene succinate and PLA are also impressive; their production capacities are expected to increase very significantly between 2014 and 2020. PLA production capacity was around 200\u00a0000 tonnes in 2014 and is projected to reach about 450\u00a0000 tonnes by 2020. PLA is used for a broad spectrum of applications, ranging from fibre products to bottles and packaging foams. PLA can not only be used in its pure form but also several derived co\u2010polyesters make up the core of the biodegradable polymer markets as disposable, short\u2010life products. Thin\u2010walled goods bags or bio\u2010waste make up two\u2010thirds of the total consumption today and this share is expected to grow further until 2020 (Castro\u2010Aguirre *et\u00a0al*., [2016](#mbt212393-bib-0002){ref-type=\"ref\"}).\n\nFinally, starch\u2010based materials are also pushing into new markets like, for example, coffee capsules or aquaculture items. Functional products such as barrier packaging and various biodegradable outdoor uses are currently low volume but have a significant potential for market breakthrough.\n\nOver the past decade, we have gained a better understanding of the molecular mechanisms and regulatory processes underlying the biotechnological synthesis of bio\u2010based polymers and monomers. Knowledge in this field is the foundation for metabolic\u2010 and protein\u2010engineering approaches to improve economic\u2010production efficiency and produce tailor\u2010made polymers with highly applicable material properties (Chung *et\u00a0al*., [2015](#mbt212393-bib-0004){ref-type=\"ref\"}).\n\nBiodegradability versus durability and the issue of waste management {#mbt212393-sec-0002}\n====================================================================\n\nThe durability of plastics was originally viewed as a virtue; however, this durability has created environmental problems, and led to the early research and development of the first biodegradable plastics. Pure aromatic polyesters like PET are traditionally considered as quite insensitive to any hydrolytic degradation. Under drastic chemical conditions (e.g. sulphuric acid at 150\u00b0C) hydrolysis of such polyesters can be used for recycling purposes.\n\nDue to the requirement of these energy\u2010consuming conditions for chemical recycling, the application of polyester hydrolases has been proposed as an environmentally friendly alternative for PET recycling (Wei *et\u00a0al*., [2014](#mbt212393-bib-0015){ref-type=\"ref\"}). Polyester hydrolases from several fungi and bacteria have already been evaluated for their activity against PET fibre surfaces with applications in the textile and detergent industries. Efficient hydrolysis of synthetic polyesters requires reaction temperatures close to their glass transition temperature (Tg) (approximately 75\u00b0C for PET). Thermophilic microorganisms are therefore an interesting source of thermostable hydrolases capable of degrading aromatic polyesters. All polyester hydrolases from actinomycetes described so far are members of the \u03b1/\u03b2 hydrolase fold superfamily of enzymes, similar to the well\u2010known PHA depolymerases for the degradation of PHA (Knoll *et\u00a0al*., [2009](#mbt212393-bib-0008){ref-type=\"ref\"}). In the past decade, a growing number of enzymes capable of hydrolysing synthetic polyesters have been identified from thermophilic actinomycetes of *Thermobifida* and *Thermomonospora* species (Wei *et\u00a0al*., [2014](#mbt212393-bib-0015){ref-type=\"ref\"}). They demonstrate versatile hydrolytic activity against both soluble and insoluble substrates, including PET and other aromatic polyesters. Their thermal stability at temperatures between 50 and 60\u00b0C make these enzymes applicable to the surface modification of polyester films and fibres. However, we still require more detailed knowledge of the factors influencing thermal stability and substrate binding to engineer polyester hydrolases with a superior ability to hydrolyse synthetic polyesters, opening up new avenues for protein evolution using rational design strategies. Microbial cutinases and enzymes with applications to surface modification of PET also play a key role in this field. To develop enzymes with further enhanced activity will require a better understanding of the interaction of the enzyme with the substrate regarding factors such as sorption, movement on the polymer surface, and the role of hydrophobins or binding modules (Guebitz and Cavaco\u2010Paulo, [2008](#mbt212393-bib-0007){ref-type=\"ref\"}).\n\nVery recently, a mesophilic bacterium that degrades and assimilates PET has been isolated by Yoshida *et\u00a0al*. ([2016](#mbt212393-bib-0017){ref-type=\"ref\"}). These authors screened natural microbial communities exposed to PET in the environment, and thus isolated a novel bacterium, *Ideonella sakaiensis* 201\u2010F6, which is able to use PET as its major energy and carbon source. When grown on PET, this strain produces two enzymes capable of hydrolysing PET and the reaction intermediary, mono(2\u2010hydroxyethyl) TPA. Both enzymes are required to enzymatically convert PET efficiently into its two environmentally benign monomers, TPA and ethylene glycol. These are interesting bioprospecting strategies to isolate microorganisms for assimilating these recalcitrant plastics and, therefore, pave the way for *in\u00a0vivo* strategies for plastic mineralization. In this context, plastics are the most abundant form of marine debris, with documented impacts on some marine environments, although the influence of plastic on microbial communities in ocean ecosystems is poorly understood. Amaral\u2010Zettler and her co\u2010workers collected plastic marine debris at multiple locations in the North Atlantic and analysed the attached microbial communities. They found a diverse microbial community of heterotrophs, autotrophs, predators and symbionts; a community that they referred to as the 'Plastisphere' (Zettler *et\u00a0al*., [2013](#mbt212393-bib-0019){ref-type=\"ref\"}). These findings highlight the potential niche for plastic\u2010degrading bioprospecting strategies.\n\nIn contrast to drop\u2010in plastics, biopolymers like PHA are, by definition, biodegradable, and so their application as commodity products becomes increasingly attractive in view of the need to avoid recalcitrant oil\u2010based polymers. When exposed to the microbial flora present in a given environment (e.g. in soil or water), biopolymers are fully degraded and mineralized to CO~2~ and H~2~O. Secreted depolymerases and hydrolases attack the biopolymer backbone, leading to low\u2010molecular\u2010mass degradation products, which can then be taken up by the microbial cell and used as carbon and energy sources (Knoll *et\u00a0al*., [2009](#mbt212393-bib-0008){ref-type=\"ref\"}). The biodegradability of these materials has been widely demonstrated at lab scale but the biodegradation of these upcoming bioplastics (bio\u2010based polymers and their blends) needs to be demonstrated at pilot scale in waste management plants.\n\nResearch shows PLA is absorbed in animals and humans and, hence, it is extensively used in biomedicine. The degradation of the polymer in animals and humans is thought to occur via non\u2010enzymatic hydrolysis. Several enzymes can degrade the polymer, including proteinase K, pronase and bromelain. However, few have been characterized with regard to microbial degradation of the polymer. PLA is also readily degraded in compost (Castro\u2010Aguirre *et\u00a0al*., [2016](#mbt212393-bib-0002){ref-type=\"ref\"}).\n\nWaste management of bio\u2010based plastic will be one of the most important issues in terms of the circular economy. In an ideal 2020 city scenario, in which the bioplastic would be subject to sorting instructions, biodegradable plastics could either be collected with other packaging, residual waste or with organic waste. However, any of these waste management channels could have drawbacks, either during the separation process at origin and/or in the biodegrading process in waste management plants. Biodegradable plastics and their blends are not always compostable plastics; biodegradable plastics are mineralizable due to the action of microorganisms and enzymes. Through this process, the materials are converted into carbon dioxide, methane, water and biomass. However, compostable plastics conform to the official standards according to EN 13432, which specifies they must be compostable only under specific conditions (temperature, humidity level, time) in the composting system, and should lack toxic side\u2010effects for water, soil, plants or living organisms. To deal with this standardization in the biodegradation process, scientists should address composting solutions based on microbial communities with complementary hydrolytic activities for a broad substrate range of bio\u2010based polymers and their additives. Otherwise, the advantages of biodegradable plastics will be futile. Concerning the separation of waste at origin, the different labels indicating how the bioplastics should be collected and sorted might be extremely confusing for the consumer. There is a clear need for standardized labelling and sorting instructions, which should be regulated in a coordinated way by policymakers and material producers (more information at <http://www.pro-e.org/Fact-sheet-on-bioplastics.html>).\n\nUnexploited feedstock of international origin for bio\u2010based plastics production {#mbt212393-sec-0003}\n===============================================================================\n\nIt is widely accepted that the price of the carbon source is one of the main factors affecting the cost of bio\u2010based plastics, influencing the sustainability of production processes. However, the choice of a suitable carbon source is not a clear\u2010cut issue and there is debate around the merits of using plant\u2010based biomass normally used as a food source for the production of non\u2010food chemical products, such as fuels. This debate has promoted the search for other sources, such as the use of waste. There are some bio\u2010wastes that share more or less homogeneous compositions (e.g. glycerol, vegetal oils, corn steep liquor, molasses, milk, meat), which can be managed using various biological processes. The use of industrial and agricultural by\u2010products could require extensive purification, equalling or even surpassing the energy demand of cost\u2010intensive agricultural feed\u2010stocks. Also, as waste complexity increases, so too does their disposal and/or recycling, requiring the integration of different technologies to achieve more efficient processes. In the particular case of bio\u2010wastes, technologies are integrated through the concept of bio\u2010refinery. In fact, the production of bio\u2010based and biodegradable polymers from renewable resources is an area of intense research and industrial activity (Koutinas *et\u00a0al*., [2014](#mbt212393-bib-0009){ref-type=\"ref\"}).\n\nHowever, a significant portion of waste is poorly biodegradable and cannot be easily converted into new added\u2010value chemicals or polymers by microorganisms or other biological processes because it contains mixtures of very complex compounds and some toxic pollutants, which are highly recalcitrant to degradation. In this context, the use of plastic wastes has recently been proposed as a novel second\u2010generation carbon source for biotechnology, supported by the analogy to the mega\u2010developments in lignocellulosic biotech. This idea is supported by the fact that carbon\u2010rich polymers have a relatively simple and well\u2010defined composition compared with biomass, and they are also extremely abundant, which will empower the recycling industry to a qualitatively new dimension (Wierckx *et\u00a0al*., [2015](#mbt212393-bib-0016){ref-type=\"ref\"}). In 2010, plastics comprised 12.4% (31 million tonnes) of total municipal solid waste in the United States (United States Environmental Protection Agency (USEPA), [2011](#mbt212393-bib-0014){ref-type=\"ref\"}). Only 7.6% of that plastic waste was recovered for recycling, leaving 28.7 million tonnes of plastic for landfill or incineration. In Europe, 25 million tonnes of plastic waste is generated yearly, of which some 60--80% is managed in this way with the remaining 20--40% being materially recycled (<http://www.plasticseurope.org/documents/document/20150227150049-final_plastics_the_facts_2014_2015_260215.pdf>). Although still in their infancy, bioprospecting strategies are being developed to isolate microorganisms able to assimilate recalcitrant plastics; this will yield potential drivers to design bio\u2010based up\u2010cycling activities from plastic wastes.\n\nAnother example of insufficiently exploited bio\u2010wastes can be found in the municipal and commercial wastes and the sludge derived from urban water treatment. These potential raw materials contain significant reusable carbon fractions that are suitable for revalorization processes. According to the European Energy Agency, the use of municipal waste as a resource could cut GHG emissions by 62 million tonnes of carbon dioxide equivalents by 2020 compared with 2008 (European Environment Agency (EEA), [2011](#mbt212393-bib-0006){ref-type=\"ref\"}). Thermochemical conversion techniques, other than incineration/combustion, such as gasification and pyrolysis are becoming widely accepted as suitable alternatives. These gasification processes can convert any carbonaceous material into a synthesis gas (or syngas), which is predominantly composed of hydrogen, carbon monoxide and carbon dioxide. Syngas can serve directly as a fuel in gas engines, or it can be used to produce hydrogen, methanol, converted to liquid fuels via Fischer--Tropsch reactions or applied in fermentation processes to produce biofuel, chemicals and bioplastics. Recently, gasification and pyrolisis of different waste streams has attracted scientific and industrial attention, including the use of petrochemical plastic waste to produce fermentable syngas. (Drzyzga *et\u00a0al*., [2015](#mbt212393-bib-0005){ref-type=\"ref\"}). Research related to the fermentative production of chemicals from syngas has greatly increased in recent years, recognizing the potential of using biological means to convert CO and CO~2~ to chemicals. On the one hand, the drawbacks of this technology include low rates of bacterial cell growth, transference of CO and H~2~ from the gas phase to the liquid phase of the fermentation medium, and low rates of bioproduction. On the other hand, the advent of efficient and inexpensive methods to sequence complete genomes and omic\u2010technologies have shed light on the metabolic and regulatory pathways of several syngas\u2010fermenting microorganisms. Also concerning gas fermentation, one of the most challenging strategies is the production of biopolymers and other bio\u2010based products from solar energy and carbon dioxide by artificial photosynthesis, converting sunlight to electricity and H~2~ *via* water electrolysis. An autotrophic H~2~\u2010oxidizing bacterium fixes CO~2~ in dark conditions. The assimilated CO~2~ is stored in bacterial cells as PHA. Compared with natural photosynthesis of a fast\u2010growing cyanobacterium, the artificial photosynthetic system has much higher energy efficiency and greater yields of bio\u2010based products (Yu, [2014](#mbt212393-bib-0018){ref-type=\"ref\"}). Progress is underway in these kinds of multidisciplinary strategies, combined with the latest systems biology tools, which are being applied to these microorganisms to enhance polymer or monomer production from gaseous C1 compounds via genetic and metabolic modification. Meanwhile, methods such as medium optimization and reactor design are being pursued to enhance polymer production from existing and new CO/CO~2~\u2010fermenting microorganisms.\n\nFinally, mixed microbial cultures have been identified as promising processes for the production of some monomers and PHA from certain waste, such as sludge from water\u2010treatment plants. This is because these systems are based on the use of open cultures, which can use a wide variety of complex nutrient\u2010rich substrates, and, as opposed to most pure cultures, PHA storage is not induced by nutrient limitation. This is particularly advantageous when using industrial feedstock waste containing compounds of undefined composition (Serafim *et\u00a0al*., [2008](#mbt212393-bib-0012){ref-type=\"ref\"}). However, the possibility of using the facilities already existing in wastewater treatment plants to produce PHA should be explored and evaluated as a cost\u2010effective technology. The future of this technology might be envisioned with the synergistic development and integration of omic\u2010systems biology tools (including technologies for population profiling and analysis) and analytical tools for studying molecular interactions in microbial populations. The interaction mechanisms of complex mixed microbial cultures could be thoroughly explained and understood, and the rational engineering of these microbial communities might thus pave the way for a host of intriguing applications in bio\u2010waste revalorization to biopolymers.\n\nConcluding remarks {#mbt212393-sec-0004}\n==================\n\nIn summary, climate change, market demand and production capacity have selected drop\u2010in plastics as the leaders of the bio\u2010based polymers market. BioPET is set to become the bio\u2010based polymer with the biggest production capacity in upcoming years while biodegradable plastics like PHA and PLA will experiment an upward market trend until 2020. Research will provide improved strains designed through synthetic and systems biology approaches; furthermore, the use of low\u2010cost substrates will contribute to the widespread application of these biopolymers. The availability of high\u2010throughput experimental tools and quantitative analysis techniques currently facilitates the design of more robust metabolic engineering strategies aiming to enhance and/or tailor this bio\u2010based plastic/monomer production. Exploitation of complex bio\u2010wastes will support the development of a bio\u2010based plastic market. Durability of bio\u2010based drop\u2010in plastics may impede the up\u2010cycling of these materials in terms of circular economy and sustainability. Combined biodegradation and bioprospecting strategies have emerged to address this issue. In line with these developments, waste management of bio\u2010based plastics needs to be standardized and regulated by governmental policymakers.\n\nConflict of interest {#mbt212393-sec-0005}\n====================\n\nThe authors declare no conflict of interest.\n"}
+{"text": "In plasma based atomic spectrometry, the sample introduction has long been considered as the Achilles' heel[@b1]. The development and improvement of sample introduction systems have been one of the main challenges for scientists working in this field[@b2]. Conventional pneumatic nebulization as the most common sample introduction device suffers from low nebulization efficiency (\\<5%), high sample uptake rates (1--2\u2009mL min^-1^) and serious matrix interferences, which is problematic when sample volumes are small or sample with complex matrices[@b3][@b4]. Later technologies such as micro-nebulizers and electrothermal vaporization (ETV) providing higher sample introduction efficiency with reduced sample consumption have indeed attracted some interests[@b5][@b6][@b7]. However, these technologies still have some shortcomings, such as expensive equipment required in nebulizer with the increased risk of clogging[@b8][@b9] and serious memory effects on ETV[@b10]. In recent years, electrolyte-cathode discharge (ELCAD) has proven to be successful for solution analysis such as solution-cathode glow discharge (SCGD) developed by Hieftje *et al*.[@b11][@b12][@b13][@b14] and liquid-film dielectric barrier discharge (LFDBD) developed by Zhu *et al*.[@b15]. These liquid discharge devices offer advantages of small size, elimination the use of a nebulizer, low sample/power consumption, and cost effectiveness. However, most of the ELCAD are operated with flowing solution as one electrode, which necessitates the use of a sampling pump in the system[@b13]. Moreover, a known behavior about ELCAD is that the elemental emission intensity strongly depends on the pH of the solution, which restricted their applications[@b11][@b16].\n\nIn the year 2004, Cooks *et al*. developed a new ambient surface sampling/ionization technique for mass spectrometry, termed desorption electrospray ionization (DESI)[@b17]. In the DESI technique, a pneumatically assisted electrospray jet is directed toward the sample surface to facilitate desorption/ionization. The DESI technique has several outstanding advantages over traditional techniques, including ambient sampling, high analysis speed, and direct sample analysis with little need for sample preparation/preseparation. Under the tide of the new scientific revolution, a mount of plasma-based techniques have been developed for ambient ionization methods[@b18][@b19][@b20][@b21][@b22][@b23][@b24][@b25]. Nevertheless, all of these techniques focus on the analysis of organic compounds with mass spectrometry. Few works have been reported in atomic spectrometry with plasma based surface sampling techniques. This is mainly due to the fact that the excitation energy and gas temperature of the plasma tail plume was too low to excite the atomic spectral information.\n\nIn chemical analysis, the introduction of sample matrix sometimes can bring unexpected effects. A typical example is the matrix-assisted laser desorption ionization mass spectrometry (MALDI/MS)[@b26][@b27], which was developed for the analysis of molecules with large molecular weights. To obtain a signal, the analyte is mixed with a matrix compound. Most of the laser energy is absorbed by the matrix and then \"transferred\" to the analyte, which can be eventually ionized by the process. Therefore, the matrix is one of the most important factors in the success of the laser desorption technique. Recently, a matrix assisted ionization vacuum (MAIV) method was also developed. Compared to MALDI, laser is not required in vacuum of MAIV for ion formation[@b28][@b29].\n\nIn this work, matrix assistance was utilized to couple with plasma atomization emission spectrometry for surface sampling for the first time. An appropriate matrix candidate should provide analyte auxiliary energy for atomization and excitation, and be absorbable to liquid analytes. Since filter paper (FP) is extremely simple and has already been successfully used in sample introduction, such as paper spray with DESI and MAIV[@b30][@b31][@b32], it has been selected in this work as surface sampling matrix. This new method, termed matrix-assisted plasma atomization emission spectrometry (MAPAES), is simple in concept and instrument design, and is also easy for system operation. In most cases, MAPAES allows analytes present at matrix surfaces to be analyzed without any sample pretreatment, which simplified the analytical process. Furthermore, this method enables fast analysis with little sample consumption (sampling at 1-\u03bcL level, less than 1\u2009min per sample). In addition, the detection can be accomplished through optical emission, which enables simultaneous detection of multiple elements. These advantages make the proposed MAPAES attractive as a potential miniaturized AES system for *in situ* and high-throughput elemental analysis.\n\nResults\n=======\n\nThe optical emission process of MAPAES\n--------------------------------------\n\nThe schematic diagram and photograph of the experimental setup are shown in [Fig. 1(a,b)](#f1){ref-type=\"fig\"} respectively. The tail plume of the home-built microwave-induced plasma jet was directly in contact with the ceramic sample plate surface at an angle of \\~60\u00b0, forming an elliptic plasma region on the surface. Due to the thermal effect of the plasma, the sample solution adsorbed in the FP quickly evaporated even before touched by the plasma tail plume. Therefore, liquid samples were allowed to be detected immediately after being spotted on the FP without waiting for the FP to dry. After a few seconds treatment with plasma tail plume, the FP substrate was carbonized with its color turned black. To ensure an approximately same degree of carbonization, the carbonizing position was kept at about 2\u2009mm away from the visible plasma region and the duration was fixed for 8\u2009s. When the carbonized FP substrate being ablated by the plasma tail plume, it was instantaneously overheated and blazed up, which induced the sample atom released from the substrate, and the released sample atoms were simultaneously excited by the plasma in the ablation region. The ablation process was less than 2\u2009s. The measurement time is about 15\u2009s and the whole analysis process is within 1\u2009min for a single sample.\n\nThe carbonization process was essential. Compared with noncarbonized fresh FP, the carbonized FP combusted much more quickly during ablation, which accelerated the sample atomization/excitation process, thus producing more intense and narrower pulse optical signals. In addition, combustion of carbonized FP emitted less visible light, which partly reduced the elevation of the spectral baseline in the visible region. Furthermore, after carbonization, the small piece of paper (2.5\u2009mm\u2009\u00d7\u20091.5\u2009mm, L\u2009\u00d7\u2009W) can be completely burned out one-time as a whole, which minimized the nonuniformity of combustion and fluctuation of signals caused by manual operation, and therefore helped to improve the reproducibility of the method.\n\nFor quantitative analysis, the size of filter paper is important. As peak height was used as analytical signal in this work, the whole filter paper must completely burn out in one time to give a single pulse signal. Therefore the filter paper must be limited in a small size. However, the absorptive capacity of filter paper decreases with its size reduction. It is also difficult to detect sample with very small size of filter paper, particularly for samples in low concentrations. Therefore, a compromised filter paper size of 2.5\u2009mm\u2009\u00d7\u20091.5\u2009mm, which can effectively absorb 1\u2009\u03bcL liquid analyte and can be completely burned out in one time, was selected.\n\nEffect of FP sample matrix\n--------------------------\n\nTo test the effect of FP sample matrix in the atomization emission process, two kinds of sample plates were used for comparison --- one was ceramic chip baseplate with precut FP pasted on it for sample adsorption, the other was clean ceramic chip itself as the sample substrate. Blank solution (5% HNO~3~) and the solution spiked with an 11-elements standard (4\u2009\u03bcg mL^\u22121^ of Au, Cr, Cu, Eu, In, Mn, Ni, Rh, Y and 0.8\u2009\u03bcg mL^\u22121^ of Ba, Sr) were introduced onto both of the two substrates respectively: 3\u2009\u03bcL on the clean ceramic chip surface, while 1\u2009\u03bcL on the FP substrate. Due to the low absorbency of the slippy ceramic surface for liquid solution, air drying for about 2\u2009h was performed only when clean ceramic chip was used as the substrate to prevent sample diffusion by plasma gas flow. The blank/samples spotted on the ceramic substrates were first ablated by the plasma tail plume, and then the FP substrates, so that the difference of the spectra can be distinguished easily. [Figure 2](#f2){ref-type=\"fig\"} shows the typical emission spectra over the wavelength region between 260 and 460\u2009nm. When the plasma tail plume touched the sample spot on the ceramic chip, no atomic spectral lines belong to analytes could be detected on the spectrum ([Fig. 2(b)](#f2){ref-type=\"fig\"}). The emission spectrum in [Fig. 2(b)](#f2){ref-type=\"fig\"} is almost the same as the background emission ([Fig. 2(a)](#f2){ref-type=\"fig\"}), with only several molecular emissions attributed to OH bands (band heads at 281 and 306\u2009nm), and the N~2~ second positive system (band heads at 337, 357, 380, and 405\u2009nm) found, indicating that the analytes were not excited by the plasma tail plume. The emission spectrum of ablating the clean FP ([Fig. 2(c)](#f2){ref-type=\"fig\"}) shows three CN bands, with band heads located around 358, 388, and 420\u2009nm. The CN molecular bands originate from the recombination of carbon and nitrogen atoms in the plasma. The carbon atoms are from the cellulose fiber of the filter paper and the nitrogen atoms are from the surrounding air. Along with the CN bands, atomic lines of Mg I, Mg II, Na I, Ca I, Ca II, Fe I, Al II, Sr II and Ba II were also found. These atomic and ionic emission lines originated from the ash content of the FP. The spectrum of the 11-element sample with FP as substrate is shown in [Fig. 2(d)](#f2){ref-type=\"fig\"}. Apart from the lines shown in [Fig. 2(c)](#f2){ref-type=\"fig\"}, atomic emission lines of all these 11 testing elements can be found in [Fig. 2(d)](#f2){ref-type=\"fig\"}, including Au I (328.068\u2009nm), Cr I (357.869, 425.435 and 428.972\u2009nm), Cu I (324.754, 327.396\u2009nm), Eu II (381.967, 412.970 and 420.505\u2009nm), In I (451.131\u2009nm), Mn I (279.482, 403.076 and 403.307\u2009nm), Ni I (300.249, 341.476 and 352.454\u2009nm), Rh I (343.489, 369.236 and 365.799\u2009nm), Y II (324.228, 360.073, 371.030 and 377.433\u2009nm), Sr II (407.771\u2009nm) and Ba II (455.403\u2009nm). Therefore, we can make a conclusion that FP substrate was necessary for the detection of these elements through surface sampling --- no FP, no signal. With FP as substrate, the sample spot is subject to both radiation heating by the plasma jet and the combustion heating by burning FP during ablation. The combustion of the FP made up for the energy deficiency of the plasma jet, which effectively promoted the atomization/excitation processes.\n\nThe small precut FP matrix also helps to maintain the sample amount in unit surface and is suitable for sample storage. After sample being deposited on the FP, it can be detected immediately or allowed to dry (at least four hours) for storage. In the latter case, these dried sample FPs can be stacked and stored in resealable plastic bags or plastic containers. The stability of analytes in dried sample spots was evaluated by exposing samples (4\u2009ng Cu, Rh, Y, In and 0.8\u2009ng Ba) sealed in plastic bags at room temperature (22--28\u2009\u00b0C) for up to 30 days. As shown in [Fig. 3](#f3){ref-type=\"fig\"}, the emission intensities of all the five nonvolatile elements were very stable. The FP matrix appears to provide some protection against sample degradation. As a result, FP can be considered as an ideal method for keeping sample stable over time and facilitating transportation of samples.\n\nThe blank correction process\n----------------------------\n\nCombustion of FP inevitably gives out lots of visible light, and the instability of combustion contributes to a source of variation, resulting in remarkable elevation of spectral baseline in the visible region and large fluctuations in the emission of blank. This variation significantly deteriorates the accuracy and precision of the entire analytical process. Emission intensity at peak position can't be regarded as peak height for quantitative determination due to the irregular undulations of baseline. Therefore, blank correction is of great importance in quantitative analysis. Since Eu has a secondary sensitive line 412.970\u2009nm located in the visible region, it was selected as an example for testing. [Figure 4(a)](#f4){ref-type=\"fig\"} shows the emission spectra of 4\u2009\u03bcg mL^\u22121^\u2009Eu and blank emission in the region between 412.2 to 413.8\u2009nm. The wavelength of 412.361\u2009nm is adjacent to the peak area of Eu 412.970\u2009nm. The temporal signal profiles of blank emission at 412.970\u2009nm and 412.361\u2009nm were recorded simultaneously for six measurements and the results are shown in [Fig. 4(b)](#f4){ref-type=\"fig\"}. It can be found that fluctuations of blank emissions at the two neighboring wavelengths (412.970\u2009nm and 412.361\u2009nm) follow exactly a same trend during whole measurement cycles. And the signal at the two wavelengths have similar intensities, the emission intensity at 412.361\u2009nm is around 95% of that at 412.970\u2009nm. Therefore, the emission intensity at 412.361\u2009nm can be regarded as the blank data for 412.970\u2009nm. The blank correction can be performed by first recording the emission signals at both 412.970\u2009nm and 412.361\u2009nm ([Fig. 5(a,b)](#f5){ref-type=\"fig\"}), and then the net intensity of 412.970\u2009nm atomic line can be derived by subtracting the corresponding blank data at 412.361\u2009nm ([Fig. 5(c)](#f5){ref-type=\"fig\"}). As shown in [Fig. 5(c)](#f5){ref-type=\"fig\"}, the blank correction process effectively eliminated the interferences of visible light caused by combustion of FP and significantly improved the reproducibility of analysis. Blank correction was done for every element quantitatively analyzed in this work. The analytical lines and the corresponding blank correction wavelengths for the 8 elements tested are listed in [Table 1](#t1){ref-type=\"table\"}.\n\nAnalytical performance and figures of merit\n-------------------------------------------\n\nThe analytical performance of our MAPAES technique was evaluated under optimal conditions (microwave power: 150\u2009W, argon flow rate: 300\u2009mL/min). The net emission peak was used as analytical signal throughout this work. Linear correlation coefficients (R) for calibration curves of the 8 analytes were all better than 0.99. Limits of detection (LOD) were calculated using the definition 3\u2009s/m, where s is the standard deviation of 11 blank measurements and m is the slope of the calibration plot. The LOD values of the 8 elements range from 1.6 to 59\u2009ng mL^\u22121^ corresponding to absolute mass LODs of 1.6 to 59\u2009pg in 1\u2009\u03bcL sample solution. The detection limits of the proposed MAPAES system have also been compared with other OES systems (cf. [Table 1](#t1){ref-type=\"table\"}). It can be seen that, with little sample consumption, the detection limits for this new source are comparable to or one order higher than those of ICPAES and MIPAES. In addition, repeatability, expressed as relative standard deviation from 10 replicates, ranged from 2.5% to 5.6% for analyte concentrations at 0.8\u2009\u03bcg mL^\u22121^ (Ba) and 4\u2009\u03bcg mL^\u22121^ (others). [Figure 6](#f6){ref-type=\"fig\"} shows the temporal profile of emission signal from 10 consecutive determinations of an 8-element standard solution. These results indicate that this new system has acceptable sensitivity and repeatability.\n\nTo validate the proposed method, the MAPAES system was employed to determine Cu, Mn, Ba and Ni in standard reference materials of simulated natural water samples (GBW(E)080397, GBW(E)080406, and GBW(E)080405) and a real water sample (river water). In comparison with conventional ICPAES and MIPAES, MAPAES requires less sample pretreatment. Pneumatic nebulizers as the most common sample introduction device for ICP and MIP, can only sampling clean solution without any precipitate. However, MAPAES is able to analysis turbid liquid samples directly. The real water sample can be added onto the sample plate for direct analysis without filtration or acidification. The analytical results of each element in these standard reference materials and the water sample are summarized in [Tables 2](#t2){ref-type=\"table\"} and [3](#t3){ref-type=\"table\"}. The analytical results obtained with the proposed method are well in agreement with the certified values. In addition, real sample recovery was also tested with spiked 500\u2009ng mL^\u22121^ of Cu, Mn, Ba and Ni (97.6--102.4%), and satisfactory results were obtained, confirming that the proposed method is valid for analysis of environmental water samples.\n\nDiscussion\n==========\n\nA novel method for surface sampling elemental analysis was developed based on the matrix-assisted plasma atomization emission spectrometry. With the assistance of filter paper, multiple elements on its surface can be detected directly by the plasma tail plume with good sensitivity. It can be concluded that the FP matrix provides a number of essential functions: (1) offering energy to assist the atomization/excitation process by combusting itself, (2) as an absorbing matrix for analytes, preventing the sample solution/powder diffusion or splash, (3) without complex sample pretreatment, (4) possibly direct detection of elements using plasma tail plume through surface sampling, (5) stabilizing analytes in dried sample spots at room temperature, thus providing considerable flexibility for sample collection and transportation.\n\nThe present method provides a new way for atomic spectrometry and offers several advantages. First, MAPAES enables simultaneously fast analysis (less than 1\u2009min per sample) for multiple elements without any sample pretreatment, which is promising for high-throughput analysis. Second, analysis by MAPAES consumes only low sample volume (1\u2009\u03bcL), a benefit for samples with limited amount. In addition, MAPAES eliminates the use of a sample flow system. The whole system is small in size, simple in structure, and cost-efficient, which provides a possibility to design a portable, miniature MAPAES instrument for field analysis.\n\nMethods\n=======\n\nThe MAPAES system\n-----------------\n\nThe home-built microwave-induced plasma source was in a Surfatron structure, which employed surface wave for energy propagation. The microwave coaxial cavity was in a cylinder-shape and made of copper with a fused-silica tube (1.0\u2009mm i.d., 6\u2009mm o.d., 200\u2009mm long) centered axially. The rf power from the solid-state microwave generator (2450\u2009MHz, maximum power of 150\u2009W) was input into the cavity via a standard N-type connector. Both inner chamber length and gap length of Surfatron were adjustable to get the highest microwave coupling efficiency. By easily rubbing the internal surface of the fused-silica tube with a slender metal wire, discharge was started and a stable, cone-shaped plasma jet was generated. The stability of the microwave induced plasma was examined over a period of 6\u2009hours. As can be seen from [Figure SI-1](#S1){ref-type=\"supplementary-material\"}, there is very limited fluctuation in background signals during this time period.\n\nCommercial quantitative filter paper (ashless grades, grade 44, Whatman, \u03a670\u2009mm) was used as the sample substrate after being cut into small pieces (2.5\u2009mm\u2009\u00d7\u20091.5\u2009mm, L\u2009\u00d7\u2009W). Alumina ceramic chip (50\u2009mm\u2009\u00d7\u200920\u2009mm\u2009\u00d7\u20091\u2009mm, L\u2009\u00d7\u2009W\u2009\u00d7\u2009T) was chosen as the baseplate due to its high heat resistance (\\>1000\u2009\u00b0C) and opaqueness. Sample plate was fabricated by pasting 10 pieces of the precut FP into two columns on the ceramic baseplate by homemade PVA glue ([Fig. 1(c)](#f1){ref-type=\"fig\"}).\n\nThe microwave cavity was mounted onto a vertical rotating stage, which allowed selection of impact angles of 0\u00b0 to 90\u00b0. The argon stream, controlled by a mass flow controller (D07-19B, Beijing Sevenstar Electronics CO., LTD), was fed through the fused-silica tube at a flow rate of 0.2--2.0\u2009L min^\u22121^. The length of plasma jet extending beyond the fused-silica tube was 2--10\u2009mm. For analysis, the sample plate was placed on a goniometer stage, which was mounted on a 3D translational stage to allow precise alignment of the plasma jet with the sample column. The analyte was detected *in situ* on the surface of the sample plate. By manually translating the ceramic baseplate, the analyte contained FP was ablated by plasma tail plume successively. The light emitted from the atomic emission process was picked up and focused into an optical fiber by a 5\u2009mm diameter collimate lens and the emission spectra were recorded by an Avaspec four channel fiber optic spectrometer system (AvaSpec-2048-4-DT). The liner CCD-array detector with 2048-element pixels was able to simultaneously detect the whole spectrum from 200 to 825\u2009nm. An integration time of 40\u2009ms, an average of two scans were set for the CCD to ensure a reasonable signal-to-noise ratio. This spectrometer has 8 functions for signal collection, which enables simultaneously quantitative analysis of at most 8 elements.\n\nReagents and standards\n----------------------\n\nAll reagents used in this work were of at least analytical grade. Ultrapurified water (18.2\u2009M\u03a9 cm^\u22123^) obtained from a UP water purification system was used throughout this work. Stock solutions (1000\u2009mg L^\u22121^) of Au, Ba, Cr, Cu, Eu, In, Mn, Ni, Rh, Sr, and Y were purchased from the National Research Center for Standard Materials (NRCSM) of China. Working standards were prepared from diluting stocking solution with nitric acid (5%).The discharge gas, high-purity argon (99.99%) was provided by Qiao Yuan Gas Company. Reference materials, including simulated natural water samples GBW(E)080397, GBW(E)080406, and GBW(E)080405, were used to validate the accuracy of our method. A water sample was collected from Funan River in Chengdu city.\n\nSafety considerations\n---------------------\n\nElectrical shock may happen on igniting the discharge with a slender metal wire. Extreme care and precaution should be taken to prevent electrical shock through the use of electrically insulating gloves. Personal protective equipment, such as microwave protective clothing and safety glasses, are highly recommended to prevent microwave radiation.\n\nAdditional Information\n======================\n\n**How to cite this article**: Yuan, X. *et al*. Matrix-Assisted Plasma Atomization Emission Spectrometry for Surface Sampling Elemental Analysis. *Sci. Rep.* **6**, 19417; doi: 10.1038/srep19417 (2016).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nThe authors thank the National Recruitment Program of Global Experts (NRPGE) and the Hundred Talents Program of Sichuan Province (HTPSP) for funding this project. They are also grateful to the support from the Research Center of Analytical Instrumentation of Sichuan University for providing all the devices and materials demanded in this work.\n\n**Author Contributions** X.Y., X.Z. and Y.D. conceived the original idea. X.Y. designed and executed the experiment under the guidance of Y.D. Finally, X.Y., X.Z., X.L., Z.Z. and Y.D. analyzed the results and wrote the paper.\n\n![(**a**) Schematic diagram of the MAPAES system, (**b**) photographs of the experimental apparatus, and (**c**) sample plates.](srep19417-f1){#f1}\n\n![Emission spectra of a blank solution (5% HNO~3~) and solution spiked with an 11-element standard (Ba, Sr at 0.8\u2009\u03bcg mL^\u22121^ and Au, Cr, Cu, Eu, In, Mn, Ni, Rh, Y at 4\u2009\u03bcg mL^\u22121^) on ceramic and filter paper substrates.\\\n(**a**) Blank emission on the ceramic substrate; (**b**) emission spectrum of the 11-element standard on ceramic substrate; (**c**) Blank emission on the filter paper substrate; (**d**) emission spectrum of the 11-element standard on filter paper substrate.](srep19417-f2){#f2}\n\n![Stability of emission intensity of Cu (324.754\u2009nm), Rh (343.489\u2009nm), Y (371.030\u2009nm), Ba (455.403\u2009nm) and In (451.131\u2009nm).\\\nError bars in the figure represent standard deviations of the results (n\u2009=\u200910). The concentrations of Cu, Rh, Y and In were 4\u2009\u03bcg mL^\u22121^, Sr was 0.8\u2009\u03bcg mL^\u22121^, with sample volume of 1\u2009\u03bcL.](srep19417-f3){#f3}\n\n![(**a**) The CCD responses to the 412.970\u2009nm Eu atomic line and blank emission. (**b**) Temporal emission profiles of blank emission at 412.970\u2009nm and 412.361\u2009nm for 6 consecutive measurements.](srep19417-f4){#f4}\n\n![Blank correction for the emission intensity of Eu at 412.970\u2009nm with respect to the blank emission obtained at 412.361\u2009nm.\\\n(**a**) The emission intensity recorded at 412.970\u2009nm for 5\u2009\u03bcg mL^\u22121^\u2009Eu; (**b**) the blank emission; (**c**) the net emission intensity after performing blank correction.](srep19417-f5){#f5}\n\n![Temporal emission profiles of 10 consecutive determinations of In (451.131\u2009nm), Y (371.030\u2009nm), Ni (352.454\u2009nm), Rh (343.489\u2009nm), Eu (412.970\u2009nm), Cu (324.754\u2009nm), Mn (403.076) and Ba (455.403\u2009nm).\\\nThe concentrations of In, Y, Ni, Rh, Eu, Cu and Mn were 4\u2009\u03bcg mL^\u22121^, Ba was 0.8\u2009\u03bcg mL^\u22121^, with sample volume of 1\u2009\u03bcL.](srep19417-f6){#f6}\n\n###### Analytical characteristics of elemental determination with MAPAES system and compared with other optical emission spectrometric systems.\n\n  element   wavelength(nm)   blank correctionwavelength(nm)   present workRSD(%)    present workLOD(ng mL^\u22121^)   present workLOD (pg)   present worklinear range(\u03bcg mL^\u22121^)   ICP-AESLOD (ng mL^\u22121^)   MIP-AESLOD (ng mL^\u22121^)   APGD LOD(ng mL^\u22121^)   ELCAD LOD(ng mL^\u22121^)\n  --------- ---------------- -------------------------------- -------------------- ---------------------------- ---------------------- ------------------------------------- ------------------------ ------------------------ --------------------- ----------------------\n  Ba        455.403          454.384                          4.1                              1.6                       1.6                          0.02--1                         8[@b6]                  41[@b33]                  --                 6.9[@b34]\n  Cu        324.754          324.993                          2.5                              7.6                       7.6                          0.05--5                       1.5[@b35]                2.1[@b36]               77[@b37]               31[@b38]\n  Eu        412.970          412.361                          4.1                               19                        19                          0.1--10                       0.34[@b39]                   --                     --                     --\n  In        451.131          450.835                          5.5                               59                        59                          0.5--30                           --                       --                  40[@b40]                  --\n  Mn        403.076          402.844                          2.7                               11                        11                          0.05--5                         1[@b6]                 9.2[@b33]               100[@b41]              30[@b38]\n  Ni        352.454          352.679                          5.6                               41                        41                          0.2--20                        65[@b6]                 6.2[@b36]              1800[@b37]              11[@b38]\n  Rh        343.489          343.785                          5.6                               15                        15                          0.1--15                       120[@b42]                1.8[@b43]                  --                     --\n  Y         371.030          370.866                          4.6                               42                        42                          0.5--20                       3.4[@b44]                    --                     --                     --\n\n###### Determination of different elements in the certified reference materials by the present system.\n\n  sample          element  certified value(\u03bcg mL^\u22121^)   measured value(\u03bcg mL^\u22121^)\n  -------------- --------- ---------------------------- ---------------------------\n  GBW(E)080397      Cu     10.00\u2009\u00b1\u20090.20                 9.96\u2009\u00b1\u20090.08\n  GBW(E)080406      Mn     5.00\u2009\u00b1\u20090.10                  4.85\u2009\u00b1\u20090.14\n  GBW(E)080405      Ni     5.00\u2009\u00b1\u20090.20                  5.02\u2009\u00b1\u20090.15\n\n###### Analytical results for four elements (Cu, Ba, Mn and Ni) in river water sample.\n\n  sample         element   added (ng mL^\u22121^)   detected[a](#t3-fn1){ref-type=\"fn\"} (ng mL^\u22121^)   recovery (%)\n  ------------- --------- ------------------- ------------------------------------------------- --------------\n  River water      Cu              0                                 nd                               --\n  \u00a0                500         488\u2009\u00b1\u200912                             97.6                        \n  Ba                0             nd                                 --                         \n  \u00a0                500         512\u2009\u00b1\u200922                             102.4                       \n  Mn                0             58                                 --                         \n  \u00a0                500         549\u2009\u00b1\u200930                             98.2                        \n  Ni                0             nd                                 --                         \n  \u00a0                500         505\u2009\u00b1\u200913                             101.0                       \n\n^a^Mean\u2009\u00b1\u2009SD (n\u2009=\u200910); nd: not detectable.\n"}
+{"text": "When a tooth fails to erupt into the arch beyond its chronological age limit is said to be impacted. Impacted teeth do not erupt on its own; they should be brought into occlusion or extracted based on the underlying cause. Even though there are many reasons behind impactions, regularly impacted teeth are seen with insufficient arch length and space into which they erupt. Impaction of third molars, canine and premolars are more common in the maxillary arch whereas central and lateral incisors are seen rarely.\\[[@ref1][@ref2]\\]\n\nOdontomas generally look like small, solitary or multiple radio-opaque lesions found on routine radiographic examinations.\\[[@ref3]\\] Odontomes may cause disturbances in the eruption of permanent teeth or retention of primary teeth, which leads to impaction of permanent teeth. Odontomes are associated with permanent teeth mostly and rarely seen with deciduous teeth.\n\nCase Report {#sec1-1}\n===========\n\nAn 18-year-male patient by name Vinith came with a complaint of generalized spacing in his upper front teeth.\n\nOn clinical examination, the case was diagnosed as Angles class I molar relation on class I skeletal base with retained deciduous lateral incisor 52, with missing right side lateral incisor and spacing in the upper anterior teeth and mild crowding in the lower teeth. Patient had no significant medical and dental history. The central incisor was drifted slightly toward unerupted lateral incisor and lack of space for the lateral incisor to erupt. On palpation, there were no signs of bulge either in the labial vestibule or on palatal vault.\n\nOn radiographic examination, odontome and retained deciduous tooth was seen along with horizontally impacted lateral incisor \\[[Figure 1](#F1){ref-type=\"fig\"}\\]. The central incisor was shifted toward the right side and occupied half of the lateral incisor space. The crown of the lateral incisor was hitting the root of the central incisor. Same side lingual, opposite side buccal technique with two intraoral periapical radiographs confirmed the presence of impacted lateral incisor on the palatal side.\n\n![Pretreatment osteoprotegerin and lateral cephalogram](JPBS-7-728-g001){#F1}\n\nThe treatment options available were extraction of impacted lateral incisor and going for a prosthetic implant by creating sufficient space for the implant placement and residual space closure with the help of fixed orthodontic treatment.\n\nSecond option was a fixed partial denture by closing the spaces in the upper front teeth, and the third option was to extract the deciduous tooth and bringing the impacted tooth into occlusion, but it is time-consuming. The patient was more interested in saving the tooth instead of going for artificial teeth.\n\nThe treatment plan was to bring the impacted lateral incisor into occlusion and close the spaces in the upper arch. Odontome and retained deciduous lateral incisor was surgically excised \\[Figures [2](#F2){ref-type=\"fig\"} and [3](#F3){ref-type=\"fig\"}\\].\n\n![Pretreatment intra oral view](JPBS-7-728-g002){#F2}\n\n![Smile and buccal view showing missing lateral incisor](JPBS-7-728-g003){#F3}\n\nAn McLaughlin, Bennett, Terevisi fixed appliance with 0.022 slot was bonded, 0.016 NiTi was used for initial leveling and aligning. The space was maintained for the impacted lateral incisor with the archwire sleeve. The central incisors were moved to the left side in order to prevent root resorption, as the lateral incisor was lying in closer proximity to the root of central incisor, and to regain the space for lateral incisor. Space closure was done on 0.19 \u00d7 0.25 stainless steel wire, and to provide stability for the adjacent tooth while bringing the impacted tooth into occlusion. Once the considerable amount of space obtained for the lateral incisor to bring into occlusion, under local anesthesia surgical exposure was done on palatal side \\[[Figure 4](#F4){ref-type=\"fig\"}\\]. After elevation of the mucoperiosteal flap, an adequate amount of bone was removed by creating a channel for the tooth to come out with the help of rotary cutting instruments. A lingual button was bonded onto the palatal aspect of the lateral incisor. Ligature wire with bull eye holes were attached to the lingual button and tied to the main arch wire. Mucoperiosteal flap was closed and sutured.\n\n![Surgical exposure of palatally placed impacted lateral incisor](JPBS-7-728-g004){#F4}\n\nOnce wound got healed, 0.012 NiTi wire was used as piggyback wire and engaged it to the ligature wire and tightly tied to the rigid stainless steel main archwire. As NiTi has shape memory effect, it started moving the tooth into occlusion, this was seen periodically through radiographs. We have increased the force slightly in further visits by increasing the dimension of the wire to 0.014 and 0.016 NiTi. Once the tooth was seen on the labial aspect, 0.016 \u00d7 0.22 TMA wire was used with a coil to bring the tooth into occlusion. Elastic chain was used simultaneously to move the tooth distally.\n\nOnce the tooth came closer, 0.016 NiTi was placed directly into the bracket of lateral incisor and brought it into occlusion \\[[Figure 5](#F5){ref-type=\"fig\"}\\]. It took 18 months to bring the horizontally impacted lateral incisor into occlusion. A panoramic radiograph shows the dilacerated lateral incisor with uprighted teeth and with proper space closure \\[[Figure 6](#F6){ref-type=\"fig\"}\\]. The stabilized teeth were retained with upper and lower fixed lingual retainer [Figure 5](#F5){ref-type=\"fig\"}. Impacted lateral incisor into occlusion.\n\n![Impacted lateral incisor into occlusion](JPBS-7-728-g005){#F5}\n\n![Posttreatment osteoprotegerin and lateral cephalogram](JPBS-7-728-g006){#F6}\n\nDiscussion {#sec1-2}\n==========\n\nThe patient wanted to retain the tooth instead of going for extraction and was very much cooperative throughout the treatment. The treatment was done on closed flap technique to control and reduce infection. The tooth was placed deep inside near the root apex of central incisor. The lingual button was bonded onto the palatal side of the impacted tooth because the labial side was facing towards the nasal floor. Many precautions were taken during bonding because the chances of debonding were high on the lingual aspect compared to the labial side, due to its irregular morphology. Lingual button was bonded on blood free field in order to prevent debonding. Controlled amount of force was applied carefully with the help of ligature wire in every visit to prevent from debonding. Stepwise bull eye loops were given and attached to the lingual button. As the tooth gets closer, the ligature wire was cut and the loop adjacent to the tooth was tied to the NiTi wire. The slow continuous force from the NiTi wire brought the tooth into occlusion. Complete space closure and a broad smile was achieved within a span of 18 months \\[[Figure 7](#F7){ref-type=\"fig\"}\\].\n\n![Posttreatment frontal and smile view](JPBS-7-728-g007){#F7}\n\nConclusion {#sec1-3}\n==========\n\nImpacted teeth can be treated easily in the mixed dentition with the periodic extraction of retained deciduous tooth. If the impacted tooth was in favorable position and time was not a constraint impacted tooth can be brought into normal occlusion with orthodontic tooth movement.\n\nFinancial support and sponsorship {#sec2-1}\n---------------------------------\n\nNil.\n\nConflicts of interest {#sec2-2}\n---------------------\n\nThere are no conflicts of interest.\n"}
+{"text": "![](indmedgaz73515-0028){#sp1 .228}\n"}
+{"text": "Introduction\n============\n\nHumans are complex supra-organisms composed of various endosymbionts that stem from all three domains of life including bacteria, archaea, and eukarya in addition to their own cells. Body habitats that are considered \"hot spots\" for microbial colonization include the skin, oral cavity, gut and urogenital tract [@B1],[@B2]. However, *in utero*, the fetus is completely sterile with colonization beginning postnatally. It is this developmental property that enables for the exploitation of germ-free animals, which are devoid of microbes on or within their body, in order to determine the functional properties of host endogenous microbiota.\n\nIn fact, the gastrointestinal (GI) tracts of both humans and conventionally raised mice harbor upwards of 10^14^micro-organisms with levels increasing along the cephalocaudal axis. Temporo-spatially the organization of bacterial cohorts differs with the largest densities residing in the large intestine, with the caecum acting as a fermentation chamber where upwards of 10^11^-10^12^bacteria/gram luminal contents ferment otherwise indigestible polysaccharides leading to the production of short chain fatty acids. The predominant bacteria groups found in human caecal fluid stems from the *E. coli* and *Lactobacillus-Enterococcus* groups that represent 50% of the caecal bacterial ribosomal Ribonucleic Acid (rRNA) whereas, *Bacteroides* (*Bacteroides, Porphyromonas* and *Prevotella* spp.) and *Clostridium* groups (*Clostridium*, *Eubacterium* and *Ruminococcus*spp.) represent 13% of caecal bacterial rRNA [@B3]. Similarly, the murine caecal microbiota establishes gradually during early postnatal life, and its complexity increases with age until a mature community is reached by 4-6 weeks of age predominately comprising the *Bacteroides* and *Lactobacillus* genera and the *Clostridium coccoides* group [@B4]. The microbiota residing along the alimentary canal takes advantage of a continuous supply of nutrients and optimal temperature while playing a pivotal role in host physiology, including nutrient processing and generation, affecting energy homeostasis, education of the immune system, and fortifying the intestinal barrier both directly and indirectly [@B5]. One of the mechanisms underlying this host-microbe mutualistic relationship is the reciprocal impact of host and microbial cells on each other\\'s gene expression programs [@B6],[@B7]. In particular, the endogenous microbiota acts as an environmental factor impacting the expression of thousands genes in the host epithelium [@B8],[@B9], and this is a function of its composition [@B9],[@B10]. However, the impact of the microbiota on the intestinal gene expression signature at the messenger RiboNucleic Acid (mRNA) level may have thus far been under-evaluated due to a lack of studies linking gut microbiota to epigenetic changes in gene expression particularly, via micro-RiboNucleic Acids (miRNA).\n\nMiRNAs are 20-22 nucleotide, single-stranded, non-coding RNA molecules involved in post-transcriptional gene regulation. Nascent miRNA exist as large hairpin-loop precursor structures that undergo several stages of enzymatic processing. Precursor miRNA molecules are first generated in the nucleus and then exported into the cytosol where they are processed by the enzyme Dicer to form shorter duplexes, with one of the two single-stranded molecules being incorporated as part of the molecular machinery involved in post-transcriptional gene regulation while the other, passenger strand (usually indicated with \\*), is short-lived and rapidly degraded. The association between the single-stranded miRNA molecule and the enzymatic complex RNA Induced Silencing Complex (RISC) lends to translational repression, or cleavage of the targeted mRNA via complementary base pairing to the three prime untranslated region (3\\'UTR) of their target mRNAs, with the degree of complementarity dictating the fate of the target [@B11]. What has come to light in recent years is that miRNAs can also induce the up-regulation of gene expression through interactions with genes that contain complementary binding sequences in their promoter regions [@B12]. To date, 1,048 miRNAs have been annotated in humans and 672 in mice (miRBase release 16, 2010) [@B13] with the true number suggested to be well over 1,000 miRNAs that are encoded in the mammalian genome [@B14]. Indeed, it is estimated that these short non-coding RNA molecules regulate up to 50% of the transcriptome (protein encoding mRNAs) [@B15], however, the true breadth of their potential lies in the fact that each miRNA can have hundreds of targets [@B16] and in retrospect, multiple miRNAs can have the same mRNA targets. These properties of miRNAs suggest that a single miRNA can potentially influence multiple biological pathways [@B17]. In fact, miRNAs whose expression is tissue and developmentally regulated [@B18], have been shown to affect a broad range of biological processes in plants and animals including; development, differentiation, cell proliferation, apoptosis [@B19], regulation of innate immunity [@B20] and defense from viruses and pathogens [@B21],[@B22].\n\nWhilst few studies have investigated the mammalian intestinal miRNA signature, a recent analysis of the global porcine micronome demonstrated the expression of 332 miRNAs along the intestinal tract with region-specific expression along the longitudinal gut axis [@B23]. In line with these findings, upwards of 200 known mature miRNAs and 122 miRNA\\* species were identified in colorectal cell lines [@B24] with some found in following clinical studies to have a greater affinity for expression in specific regions and most expressed globally in the human GI tract [@B25]. Intestinal miRNAs have experimentally proven biological roles ranging from the regulation of neonatal nutrient metabolism [@B26] to the control of intestinal fluid and electrolyte transport [@B27] and permeability [@B28], besides affecting intestinal epithelial cell differentiation [@B29] and maturation [@B30].\n\nThe intestinal miRNA signature has been found to be deregulated in various disease states. MiRNAs can display both oncogenic or tumor suppressive effects in several types of cancers [@B31], and recently 11 miRNAs were found to be differentially expressed in the sigmoid colon of patients with active ulcerative colitis (UC) versus healthy controls [@B32], with effects on secretion of pro-inflammatory chemokines [@B32].\n\nIn addition, both plants and animals differentially express miRNAs following sensing of pathogen-associated molecular patterns (PAMPs). For instance, bacterial flagellin-induced upregulation of miR-393 in *Arabidopsis thaliana* participates in the regulation of the host defense system [@B22]. In animals, specific miRNAs are induced in response to various bacterial components, such as lipopolysaccharide (LPS) in monocytes [@B33] and to viral infection such as in Hepatitis B and C [@B34]. Moreover, miR-155 is upregulated in gastric epithelial cells following *Helicobacter pylori*infection [@B35]. All of these changes resulted in downstream regulation of the immune response.\n\nIt has been recently suggested that the onset of several intestinal diseases including Inflammatory Bowel Disease (IBD) and Irritable Bowel Syndrome (IBS), are caused by both deregulation of the intestinal barrier function and by microbial factors [@B36],[@B37], but how the two intertwine to affect such conditions is not well understood. The intestinal barrier is a multi-tiered line of defense localized at the interface between the external environment and internal milieu and comprises physical, chemical and receptor-mediated pathogen sensing components [@B38]. The endogenous gut microbiota is an important constituent of the barrier in that it not only participates in the formation of the physical and chemical barrier via pathogen exclusion, antimicrobial peptide secretion, and immuno-modulation, but also acts as a vector of change by modulating the mRNA expression of a number of genes involved in intestinal barrier function [@B9],[@B7]. However, the epigenetic basis of these interactions is yet to be elucidated as there is a lack of studies evaluating modulation of host miRNAs in response to symbiotic microorganisms. Intriguingly, legumes miRNAs are modulated during the establishment and maintenance of the rhizobia symbiosis in root nodules [@B39]. Though, it is unknown if this is also true for animals who live in a symbiotic relationship with complex microbial communities at various body sites such as the intestine.\n\nWe used germ-free and conventionally raised mice to investigate the impact of the endogenous microbiota on the global expression of caecal miRNAs *in vivo*. We show that the murine miRNA signature in the caecum is comprised of several variously expressed species and that it is indeed affected by the presence of the microbiota. Moreover, we show that several of the putative mRNA targets of the modulated miRNAs encode for genes known to be involved in the regulation of the intestinal barrier function, including glycosylation enzymes, junctional proteins, proteins found in the mucus layers and genes involved in immune regulation.\n\nMaterials and Methods\n=====================\n\nAnimals\n-------\n\nSwiss Webster male mice were used according to the Regulations of the Animals for Research Act in Ontario and the Guidelines of the Canadian Council on Animal Care. Animal study design and procedures were approved by the animal ethics committee at the University of Toronto (Animal Use Protocol Number: 20008318). Five germ-free and five conventionally raised mice, 6 weeks of age, were obtained from Taconic Farms (Germantown, NY), sacrificed via cervical dislocation and then dissected in sterile conditions. Upon sacrifice, the entire caecum was immediately excised and caecal contents were collected. Caecal tissues were further cleaned with sterile 0.9 % NaCl, divided into two halves longitudinally, snap-frozen in liquid nitrogen and stored at -80\u00b0C until further processing. Caecal contents were immediately fixed in 4% paraformaldehyde and used to confirm the germ-free status of the animals by Fluorescence *in situ* Hybridization with the EUB338 5\\'-Cy3 labeled 16S rRNA probe specific for all bacteria (5\\'/5Cy3/GCT GCC TCC CGT AGG AGT-3\\') (Integrated DNA Technologies), as previously described [@B40].\n\nRNA extraction\n--------------\n\nSmall RNA-containing total RNA was extracted from one-half segment of the caecum from both germ-free (n = 5) and conventionally-raised (n = 5) mice, using miRVANA^TM^ miRNA Isolation Kit (Ambion, Austin, TX, USA), as per the manufacturer\\'s instructions, eluted in 100 \u00b5l of RNAse-free water and stored at -80\u00b0C. Recovered total RNA concentration and purity were spectrophotometrically assessed using Thermoscientific\\'s Nanodrop 1000 Spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA) and ranged between 2.13-2.16 and 2.01-2.12, respectively. RNA integrity was confirmed by denaturing agarose gel electrophoresis.\n\nGlobal microRNA expression profiling\n------------------------------------\n\ncDNA was synthesized from 1 \u00b5g of total RNA (n = 3 per group) using the Taqman\u00ae MicroRNA Reverse Transcription Kit in conjunction with Rodent Megaplex\u2122 Primer Pools according to the manufacturer\\'s protocol (Applied Biosystems, CA, USA). The Rodent Megaplex\u2122 Primer Pools contains two sets of microRNA-specific RT primers, pools A and B, that enable for the RT of 375 microRNAs/6 species-specific controls and 210 microRNAs/6 species-specific controls, respectively. Separate reactions were run for pools A and B in Applied Biosystems\\' GeneAmp\u00ae PCR System 2700 Thermocycler. Global microRNA expression profiling was conducted by TaqMan quantitative PCR using Applied Biosystems\\' Taqman\u00ae Rodent MicroRNA Array Set v2.0 (Taqman Low Density Arrays, TLDAs) that comprise two microfluidic cards (plates A and B) containing a total of 384 Taqman\u00ae Assays per card (some of which are duplicate probes). cDNA products from the Megaplex RT pools set were independently assessed on both microfludic cards with plate A enabling for the simultaneous quantification of 375 microRNA targets/6 controls while plate B, 210 microRNA targets/6 controls. The protocol suggested by the manufacturer was followed. Real-Time PCR was performed using an Applied Biosystems 7900 HT Real-Time PCR system and default thermal-cycling conditions for 384-wells Taqman Low Density Arrays.\n\nReal-Time PCR validation of individual microRNAs expression\n-----------------------------------------------------------\n\nTen ng of total RNA (n=5 per group) was reverse transcribed with the Taqman\u00ae MicroRNA Reverse Transcription Kit and primers specific for miR-455 (Assay ID: 002455) and the endogenous control snoRNA135 genes (Assay ID: 001230) (Applied Biosystems) in Applied Biosystems\\' GeneAmp\u00ae PCR System 2700 Thermocycler according to the manufacturer\\'s protocol. Real time PCR was then conducted using undiluted cDNA, TaqMan MicroRNA Assays (miR-455 assay ID: 002455 and snoRNA135 assay ID: 001230) and the TaqMan 2X Universal PCR Master Mix, No AmpEraseUNG^a^ (Applied Biosystems) in a 10 \u00b5l PCR reaction. Each reaction was run in triplicates in a 384-well optical plate in Applied Biosystems\\' 7900 HT Real-Time PCR machine using the 9600 emulation mode with an initial hold at 95\u00b0C for 10 minutes followed by 40 cycles of 95\u00b0C for 15 seconds, and 60\u00b0C for 60 seconds. Results were expressed as fold change between germ-free and conventional mice as calculated by \u0394\u0394Ct method [@B41] after normalization to sno-135 gene, which was shown to be equally expressed in the caeca of the two groups of mice by the TLDA experiments. Significance of differential gene expression was assessed with the Mann-Whitney test using GraphPad Prism 5 Software (La Jolla, CA, USA).\n\nAnalysis of TLDA data\n---------------------\n\nRaw data were pre-processed in SDS 2.3 for individual plates and then concurrently for all plates in SDS RQ Manager 1.2 (Applied Biosystems) for the generation of Ct (Cycle threshold) values. A pre-selection filter was applied to all miRNA TLDA data to reduce noise in the dataset and to reduce the severity of the multiple testing adjustment. To this end, any miRNA not meeting both of the following criteria were removed from further analysis: a) Presence in all three of either conventional samples or the germ-free group; b) Presence in at least one of the conventional samples to enable the \u0394\u0394Ct method to be used for normalization. Following pre-filtering, all duplicate probes for the same miRNA species on the same plate were averaged and the mean Ct value was utilized for further analysis. Data were then normalized by \"columnwise mean\" normalization, such that the target miRNA is normalized to the mean Ct of all miRNA for each sample, a method that has been suggested as an improvement for high-throughput miRNA Quantitative PCR (qPCR) [@B42] where the mean abundance of hundreds of targets may be more stable than any endogenous control across samples and experimental groups. To identify differentially expressed genes, the empirical Bayes-moderate t-test was used as implemented in the LIMMA R package [@B43]. For this approach, missing Ct values were assumed to be unknown rather than imputed to 40, to avoid creating a bimodal distribution of Ct values, which would violate the assumption of the t-test. False discovery rate was calculated by the method of Benjamini and Hochberg, as implemented in the R package *multtest* (Pollard *et al.,* v.1.22.0). Supervised heatmaps were created using the R package *gplots*. Clustering in the heatmaps is based on complete linkage and Euclidean as the distance metric, using default setting for the *hclust* (hierarchical cluster) function in R [@B44].\n\nAnalysis was performed in the R language and environment for statistical computing (R Development Core Team, 2008, v2.8.1) [@B45].\n\nMiRNAs target prediction\n------------------------\n\nTo investigate the relationship between selected miRNAs of interest and the genes that they potentially target, we mapped them into a miRNA network using NAViGaTOR ver. 2.2 [@B46]. We first used high precision miRNA:target relationships in mouse - taken from the TargetScan Conserved Targets (Conserved_Sites_Context_Scores.txt Release 5.1) [@B47],[@B48],[@B49] or PITA TOP database predictions (PITA_targets_mm9_0\\_0_TOP.tab.gz, May 2010 download) [@B50]. Previous work examining miRNA:target relationships suggests that both PITA and TargetScan provide high quality interactions suitable for the construction of an interaction network [@B51]. Genes identified by this first analysis were then filtered based on their inclusion in an intestinal barrier gene set to assess the potential impact of differentially expressed miRNAs on the intestinal barrier function. A subset of 527 genes important in maintenance of the intestinal barrier function were identified and classified according to function - mainly physical, chemical and pathogen sensing components as per Cummings J. H., *et al*. [@B38] ([Additional file 5](#SM5){ref-type=\"supplementary-material\"}: Table S5). Identified miRNA target genes were filtered by the intestinal barrier set prior to being mapped into the miRNA network. This reduced the number of initial miRNA target genes of the 11 miRNAs with predicted targets from 2,755 in the general setting to 34 present in the intestinal barrier setting.\n\nAnalysis of miRNA potential targets biological function\n-------------------------------------------------------\n\nTo further understand the functions of gene targets of miRNAs with altered expression in this study, we examined all 2,755 gene targets of the 11 miRNAs prior to the filtering step. Using the Panther Classification System Version 7.0 [@B52],[@B53], we examined the over- or under- representation of our miRNA target genes compared to a universe consisting of all genes listed as miRNA targets in the PITA Top Targets or TargetScan Conserved Targets (as discussed above). Categories examined include Gene Ontology Classes: Biological Processes, Molecular Function, Cellular Component as well as Pathway Analysis and Protein Class Analysis.\n\nResults\n=======\n\nDifferential expression of miRNAs in the caecum of germ-free and conventionally raised mice\n-------------------------------------------------------------------------------------------\n\nTo assess if the caecal miRNA expression signature is associated with the presence of the endogenous microbiota, we examined small RNA-containing total RNA extracted from the caecum of germ-free (n=3) and conventionally raised mice (n=3). Each sample was independently run on two different Taqman Low Density Arrays (Plates A and B), which combined allow for the analysis of 585 mature miRNAs. Fifty-seven percent of the targeted miRNAs were detectable (Ct\\<35) in the caeca of both germ-free and conventionally raised mice ([Additional file 1](#SM1){ref-type=\"supplementary-material\"}: Table S1). Ranked mean abundance of miRNAs was similar for both groups of mice (Spearman R=0.74, P\\< 0.0001, 95% CI= 0.68 to 0.78) (Figure [1](#F1){ref-type=\"fig\"}).\n\nWe found 18 transcripts differentially expressed between the two groups, including both up- and down-regulated miRNAs with a fold change (germ-free vs. conventional) ranging between 0.2 and 4.6 (Table [1](#T1){ref-type=\"table\"}) (False Discovery Rate = 0.2). These correspond to 16 unique miRNAs, including mmu-miR-351 and rno-miR-351 - two sequences conserved in mouse and rat - while Y1 is a rat endogenous small RNA. A second rat miRNA species, rno-miR-664, was found to be significantly up-regulated by 2.85-fold in germ-free samples. The murine homolog, mmu-miR-664, is not represented on the TLDA plates that we used. Sequence analysis using miRviewer [@B54] demonstrates that miR-664-1 shows sequence similarity in rat, mouse and horse with a greater conservation amongst rat and mouse. Therefore, it is likely that the measured signal is biologically reliable and derives from cross-reaction of the rno-miR-664 TaqMan assay with the homologous murine miRNA species. Five of the sixteen transcripts correspond to passenger miRNA (miRNA\\*) sequences. Up-regulation of miR-455 in germ-free versus conventional mice was confirmed in a separate experiment using gene-specific single-well TaqMan PCR and RNA from the caecum of five mice per group (fold change germ-free versus conventional=1.7, Mann-Whitney test p=0.0079).\n\nAnalysis of miRNA conservation and their genomic contexts revealed that all of the significantly differently expressed miRNAs belong to various families and cluster separately in terms of their genomic locations with the exception of miR-351 in which both rat miRNA (rno-miR-351), and murine miRNA (mmu-miR-351) were found to belong to the same mir-351 family, based on sequence conservation [@B13].\n\nFurthermore, supervised hierarchical clustering analysis using the 18 differentially expressed transcripts demonstrated intra-group similarities in miRNA expression with inter-group variation in miRNA expression (Figure [2](#F2){ref-type=\"fig\"}), showing that the caecal miRNA signatures cluster according to the presence or absence of the endogenous microbiota.\n\nExperimentally verified and predicted mRNA targets of the differentially expressed miRNAs\n-----------------------------------------------------------------------------------------\n\nBased on Tarbase V5.0 [@B55] and miRecords V2.0 [@B56], two freely available databases that provide a repository of information pertaining to experimentally validated miRNA targets in several animal species, plants and viruses, six (miR-133a, miR-672, miR-183, miR-148a, miR-145, miR-150) of the sixteen differentially expressed miRNAs have experimentally verified mRNA targets ([Additional file 2](#SM2){ref-type=\"supplementary-material\"}: Table S2). Of these mRNA targets, seven (Serum response factor (*Srf*), Ras homolog gene family, member A (*Rhoa*), Cell division cycle 42 homolog (*S. cerevisiae*) (*Cdc42*), Peroxiredoxin 6 (*PRDX6*), Homeo box A9 (*Hoxa9*), Vascular endothelial growth factor A (*Vegfa*), and Myeloblastosis oncogene (*Myb*) were detected with a signal intensity higher than 150 in microarray experiments analyzing gene expression in C57BL/6 mice caeca (n=2) (Gene Expression Omnibus [@B57] dataset GSE1133 [@B58]). Therefore, based on the miRNA/mRNA co-expression criterion for target validation [@B59], these genes are good candidates for microbiota-dependent expression modulation via miRNA.\n\nHowever, each miRNA species is likely to have multiple physiologically relevant targets, most of which are unknown, and several algorithms can be employed for their *in silico* identification. To predict targets for the 16 miRNAs found to be differentially expressed between germ-free and conventional mice, we adopted a conservative approach. We extracted target predictions from 2 sources: Probability of Interaction by Target Accessibility (PITA) [@B50] and TargetScan [@B47],[@B48],[@B49]; thus, our target prediction sets are based on several criteria including conservation of target binding sites and the degree of accessibility of the three prime untranslated regions of the mRNA target. Two thousand seven hundred and fifty-five unique genes were found to map as targets of these miRNAs as predicted by both algorithms ([Additional file 3](#SM3){ref-type=\"supplementary-material\"}: Table S3). These genes were mapped to PANTHER database [@B52] to assess their group descriptors. We first considered the Gene Ontology classifications and found our gene list to be significantly enriched in several categories: (1) biological processes including development, cell communication, signal transduction (all at p\\<0.0001), among others; (2) molecular functions including DNA, transcription factor, protein binding (all at p\\<0.0001), among others; and (3) cellular components such as actin cytoskeleton (p\\<0.0001) ([Additional file 4](#SM4){ref-type=\"supplementary-material\"}: Table S4). Next, we found our targets enriched in the Wnt signaling pathway, angiogenesis, transforming growth factor-beta (TGF-\u03b2) and cadherin signaling pathways and in the transcription factors protein class followed by enzyme modulator and ribosomal proteins (all at p\\<0.0001), among others ([Additional file 4](#SM4){ref-type=\"supplementary-material\"}: Table S4). Several of the 2,755 global putative targets of the differentially expressed miRNAs were found to map to diverse components of the intestinal barrier function when assessing gene ontologies. These include; (1) Biological Processes such as cell-cell adhesion (p\\<0.0001), immune system processes (p\\<0.001) encompassing antigen processing and presentation (p\\<0.01) and defense response to bacterium (p\\<0.2) amongst others; (2) Molecular Functions including structural constituents of cytoskeleton (p\\<0.0001) and (3) Cellular Components such as actin cytoskeleton (p\\<0.0001), MHC protein complex (p\\<0.01), cell junction (p\\<0.01) as well as other factors ([Additional file 4](#SM4){ref-type=\"supplementary-material\"}: Table S4). Similarly, identification of the inflammation mediated by chemokine and cytokine signaling pathway (p\\<0.001) during functional analysis of *in silico* data demonstrate potential implications of microbial-dependent miRNA regulation on the intestinal barrier function.\n\nMiRNA-dependent regulation of the intestinal barrier function by the endogenous microbiota\n------------------------------------------------------------------------------------------\n\nConsidering the results described above, and the fact that endogenous microbes play an important role in the creation and maintenance of the intestinal barrier we decided to investigate the potential impact of the microbiota-responsive miRNAs on the intestinal barrier function via analysis of intestinal barrier specific gene targets. We first compiled a gene set containing 527 genes involved in the regulation of the intestinal barrier function, as explained in Materials and Methods, which we called the \"Intestinal Barrier Gene Set\". The list of genes included in this gene set is provided as ([Additional file 5](#SM5){ref-type=\"supplementary-material\"}: Table S5). We then filtered the 2,755 target genes by this pre-defined gene set and established all miRNA:target interactions for the remaining genes. Using NAViGaTOR (Network Analysis, Visualization and Graphing Toronto) ver. 2.2 [@B46], a scalable, network analysis and visualization system, we mapped the miRNA network linking our identified miRNA of interest in order to examine the micronome, as described before [@B51] (Figure [3](#F3){ref-type=\"fig\"}). miR-487b did not have any intestinal barrier targets as per the algorithms employed and therefore it does not appear in the figure. Thirty-four intestinal barrier-related genes were found to be among the potential targets of the intestinal miRNAs the expression of which depends on the endogenous microbiota, and fifteen of these were identified by both algorithms. These include genes involved in glycosylation, cell-cell junction formation, the mucus layer and genes involved in immune regulation particularly MHC I and II proteins amongst others. Closer inspection of this miRNA interaction network reveals several genes co-targeted by the miRNAs identified as differentially expressed between the caecal miRNA signatures in germ-free and conventional mice. Formin 1 (FMN1) is co-targeted by 2 miRNAs down-regulated in the germ-free mice (miR-351 and miR-467a) as well as one up-regulated miRNA (miR-145). Other genes appearing to be co-targeted by multiple differentially expressed miRNAs are: Cadherin 5 (*Cdh5*), UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 5 (*Galnt5*), poliovirus receptor-related 1 (*Pvrl1*), fascin homolog 1, actin bundling protein (*Fscn1*), Cingulin (*Cgn*), glucosaminyl (N-acetyl) transferase 1, core 2 (*Gcnt1*) and UDP-N-acetyl-alpha-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase 7 (*Galnt7*). Of the thirty-four predicted intestinal barrier genes targets, twenty had been previously found to be expressed in the caeca of C57BL/6 mice (n=2), in microarray experiments with a hybridization signal higher than 150 (ATP-binding cassette, sub-family B (MDR/TAP) member 9 (*Abcb9*), Nicastrin (*Ncstn*), Spermidine/spermine N1-acetyl transferase 1 (*Sat1*), Desmoglein 3 (*Dsg3*), UDP-Gal:betaGlcNAc beta 1,4- galactosyltransferase, polypeptide 1 (*B4galt1*), Leucine aminopeptidase 3 (*Lap3*), beta-1,3-galactosyl-O-glycosyl-glycoprotein beta-1,6-N-acetylglucosaminyl\u00adtransferase (*Gcnt1*), CMP-N-acetylneuraminate-beta-1,4-galactoside alpha-2,3-sialyltransferase (*St3gal3*), Junction plakoglobin (*Jup*), Aminopeptidase puromycin sensitive (*Npepps*), UDP-N-acetyl-alpha-D-galactosamine: polypeptide N-acetylgalactosaminyl\u00adtransferase 7 (*Galnt7*), Plakophilin 1 (*Pkp1*), (alpha-N-acetyl-neuraminyl-2,3-beta-galactosyl-1,3)-N-acetylgalactosaminide alpha-2,6-sialyltransferase 6 (*St6galnac6*), Carcinoembryonic antigen-related cell adhesion molecule 1 (*Ceacam1*), Formin-1 (*Fmn1*), Prostasin (*Prss8*), glycoprotein-N-acetylgalactosamine 3-beta-galactosyltransferase 1 (*C1galt1*), UDP-Gal:betaGlcNAc beta 1,4-galactosyltransferase, polypeptide 5 (*B4galt5*), UDP-Gal:betaGlcNAc beta 1,4- galactosyltransferase, polypeptide 2 (*B4galt2*), Myosin, light polypeptide kinase (*Mylk*) (Gene Expression Omnibus [@B57] dataset GSE1133 [@B58]). This list is conservative, since not all of the genes shown in Figure [3](#F3){ref-type=\"fig\"} had microarray probes.\n\nSeventy one percent of the genes included in our gene set were not considered by PITA and TargetScan due to the low conservation of the 3\\'UTR in homologous genes. Moreover, this analysis did not incorporate passenger miRNAs because they are not considered in PITA and TargetScan databases. Alternatively, when using MicroCosm targets [@B13] to map these, only miR-let7g\\* had targets remaining after filtering with the intestinal barrier gene set. These are: C1GALT1-specific chaperone 1 (*C1galt1c1*), Claudin-7 (*Cldn7*), Histocompatibility 2, class II antigen A, beta 1 (*H2-Ab1*), Pancreatitis-associated protein (*Pap*), Phospholipase A2, group XIIA (*Pla2g12a*), Phospholipase A2, group IB (*Pla2g1b*), Spermidine synthase (*Srm*), Thimet oligopeptidase 1 (*Thop1*), Toll-like receptor-11 (*Tlr11*) and Toll-like receptor-13 (*Tlr13*).\n\nFinally, in order to substantiate the hypothesis that gut commensals impact the intestinal barrier via miRNA expression modulation, we crossed-matched our global list of intestinal barrier genes with genes previously identified to be differentially expressed in the jejunal mucosa of intestinal-specific Dicer knock-out mice [@B60]. The result of this analysis provides experimental evidence that miRNAs indeed impact on barrier-related gene expression, with potential repercussions on its function. Of particular interest are intestinal barrier genes from our list that were found to be experimentally perturbed (up- or down-regulated) by the conditional knock-out of Dicer [@B60]. They include: glycosylation enzymes, immuno-inflammatory response genes, components of MHC I and II, junctional proteins, mucus layer associated proteins and defense response proteins, including antimicrobial peptides and Pathogen Associated Molecular Pattern (PAMP) responsive elements. Although, this analysis shows that miRNAs affect genes that comprise the intestinal barrier, in order to further establish a nexus between microbial induced modulation of miRNAs, which in turn affects barrier function, we combined putative barrier related gene targets of the microbial dependent miRNAs (Figure [3](#F3){ref-type=\"fig\"}) with the Dicer knock-out mice data [@B60]. Among the genes differentially expressed in the absence of miRNAs in the jejunal mucosa, we found that seven intestinal barrier related genes were either up-regulated (glycoprotein-N-acetylgalactosamine 3-beta-galactosyltransferase 1 (*C1galt1*), myosin, light polypeptide kinase (*Mylk*), Aminopeptidase puromycin sensitive (*Npepps*), UDP-N-acetyl-alpha-D-galactosamine: polypeptide N-acetylgalactosaminyl\u00adtransferase 7 (*Galnt7*) and Prostasin (*Prss8*)), or down-regulated (protein kinase C zeta isoform a (*Prkcz*), beta-1,3-galactosyl-O-glycosyl-glycoprotein beta-1,6-N-acetylglucosaminyltransferase (*Gcnt1*)). More specifically, since these seven genes are *in silico* targets of the miRNAs modulated by the absence or presence of the microbiota (germ-free versus conventional mice) and experimental evidence points to a role of epigenetic regulation of these genes via miRNAs (Dicer knock-out study) it reiterates a potential novel mechanism of host-microbial cross-talk via microbial dependent regulation of miRNAs that may translate into effects on the host with respect to regulation of the intestinal barrier function.\n\nDiscussion\n==========\n\nGene expression modulation is one of the mechanisms underlying the cross-talk between gut endogenous microbiota and host epithelium, and therefore plays a critical role in intestinal homeostasis. Here we show that the presence of the microbiota in the murine intestinal tract, particularly in the caecum, also associates with a distinctive miRNA signature, supporting a role for gut endosymbionts in post-transcriptional regulation of gene expression. Few studies have looked at the relative expression of miRNAs along the cephalocaudal axis of the healthy gut, particularly with respect to the passenger miRNA\\* forms, and to the best of our knowledge, this is the first study examining miRNA expression in the murine caecum. We found a characteristic micronome in the caecum, with 334 miRNA species expressed in this region in both germ-free and conventional mice; of these, 74 are miRNA\\* forms. While miRNAs\\* are thought to correspond to the rapidly degraded strand of the miRNA duplex, there is evidence suggesting that they may play a so-far unrecognized role within cells [@B24], and in fact may act in a similar fashion to guide strand miRNAs in terms of abundance and gene regulation [@B61].\n\nComparison of the global murine miRNA signature along several intestinal loci, namely the small and large intestinal mucosa, as well as our caecal analysis allows for certain parallels to be drawn. Some of the miRNAs belonging to the 15 miRNAs/miRNA families most expressed in the jejunal and colonic mucosa [@B60] are also expressed in the caeca of both germ-free and conventional mice ([Additional file 6](#SM6){ref-type=\"supplementary-material\"}: Table S6) and 3 (miR-192, miR-378, miR-29a) of the 15 miRNAs most highly expressed (based on Ct values) in the caecum are also expressed in both the jejunum and colon [@B60]. Although diverse genes cannot be compared by Ct values, sorting allows for a qualitative measure of the relative level of gene expression and to identify genes that do or do not display a regional expression pattern within the intestine. miR-143 and miR-145 were part of the top expressed miRNAs in common between the jejunum and caecum, and are found in the same genomic cluster (\\<10kb distance from one another on chromosome 18). miR-200b was found in common as a highly expressed miRNA within both the large intestine and caecum. Interestingly, other miRNAs with sequence similarities to miR-200b were also found to be highly expressed, including miR-200a in the colon, and miR-200c in the caecum, suggesting that members of the miR-8 family play an important physiological role in distal intestinal regions. On the other hand, 19 miRNAs were found to be expressed in the caecum of conventional mice (Ct\\<35) but not in the jejunal or colonic mucosa based on the absence of sequence read data [@B60] in either of the two regions ([Additional file 6](#SM6){ref-type=\"supplementary-material\"}: Table S6), suggesting they may be restricted to the caecum. Intergroup comparisons between germ-free and conventional mice illustrate a relatively high degree of similarity between the top miRNAs expressed in germ-free and conventional caeca with all 15 of the miRNAs with the lowest Ct values overlapping between the two groups ([Additional file 6](#SM6){ref-type=\"supplementary-material\"}: Table S6). Moreover, there is a general concordance between the murine caecal micronome and the human intestinal micronome, which incorporates the caecum. Juxtaposing data on the 13 most highly constitutively expressed miRNAs in both the terminal ileum and colon (caecum, transverse colon, sigmoid colon and rectum) from pinch biopsy samples of healthy adults [@B25] with the murine caecal miRNA signature of conventional and germ-free mice, revealed that, five miRNAs (miR-143, miR-192, miR-200b, miR-200c and miR-24) found in the intestines of humans were amongst the top 15 mostly highly expressed miRNAs in the murine caecum based on Ct values ([Additional file 6](#SM6){ref-type=\"supplementary-material\"}: Table S6). Moreover, miR-19b which was found to have a 3.2-fold higher expression in the caecum versus the terminal ileum from biopsied samples in humans, was also found, based on our aforementioned analysis, to be part of the 15 most highly expressed miRNAs in the murine caecum (germ-free and conventional mice) but not within the jejunal or colonic mucosa, suggesting this miRNA may exert a more profound effect within the caecum. Although our analysis of miRNA expression between intestinal regions in the mouse is qualitative and cannot be used to determine fold differences in expression, it gives merit into using the mouse as a model organism to investigate intestinal miRNAs as certain parallels can be found in humans.\n\nThe expression of these miRNAs may be under genetic and environmental control. The latter is particularly important in the case of the intestine where the epithelium engages in a continuous cross-talk with the luminal microbes. Here we show that indeed the endogenous microbiota contributes to the physiological miRNA signature in the caecum, which results in 16 miRNAs being differentially expressed between germ-free and conventionally raised mice. Moreover, of these, miR-133a and miR-467a were found to be caecal specific miRNAs when compared with the jejunum and colonic mucosa and miR-145 was a non-selectively expressed miRNA with high levels of expression along the intestine, insinuating a role for microbial control of both regional specific and globally expressed miRNAs that may transcend the boundaries of the caecum. Host miRNA modulation has been so far observed in response to pathogenic insults including prions [@B62], viruses such as Hepatitis B and C [@B34] and influenza virus [@B63], bacteria such as *Helicobacter pylori* [@B64], *Francisella novicida* [@B65] and Gram negative bacteria LPS [@B66], the yeast *Candida albicans* [@B67] or parasites such as *Cryptosporidium parvum* [@B68] and *Toxoplasma gondii* [@B69]. To our knowledge, the only host-microbe symbiotic relationship associated to miRNA modulation in the host is the legume-rhizobium symbiosis [@B39]. While we used whole-thickness caeca, and therefore could not establish the cellular origin of our measured signals, a previous study revealed that at least nine human miRNAs (hsa-miR-145, hsa-miR-150, hsa-miR-133a, hsa-miR-148a, hsa-miR-183, hsa-let-7g\\*, hsa-miR-181a\\*, hsa-miR-21\\*, hsa-miR-27b\\*) that have sequence homology with our differentially expressed murine miRNA are indeed expressed in colorectal cell lines [@B24], suggesting that the intestinal epithelial monolayer is susceptible of responding to the endogenous symbionts or their products, by miRNA modulation.\n\nTo date limited information is available on the biological role of these miRNAs; however, several of the miRNAs found to be differentially expressed in this study are known to be altered in cancer states. miR-148a, which we found to be expressed more highly in conventional mice, was found to be more highly expressed in tumor samples versus normal colonic epithelium [@B24], while miR-133a and miR-145, which we found to be more highly expressed in germ-free mice, were shown to exhibit significantly higher levels of expression in normal versus tumor tissues [@B70],[@B71],[@B72]. There is a general consensus in the literature that endogenous gut microbes can alter colon cancer susceptibility and germ-free rats were found to develop less and smaller tumors than their conventional counterparts when using a protocol that induces colorectal cancer [@B73]. This was attributed to enhanced anticancer immune response. In our study, a novel pathway may be proposed that incorporates microbe signaling to the host and can alter the expression of tumor-suppressors or oncogenes post-transcriptionally via miRNA regulation. Indeed, miR-145 and miR-133a were both predicted by multiple algorithms to target Fascin-1 (FSCN1) (Figure [3](#F3){ref-type=\"fig\"}), a gene involved in actin cytoskeleton assembly, the down-regulation of which was experimentally found to explain the tumor suppressive effects of miR-145 and miR-133a in bladder, esophageal squamous cell and breast carcinomas [@B71],[@B74],[@B75].\n\nHowever, inferring a microbiota-dependent physiological role for differentially modulated miRNA species depends on the identification of their mRNA targets in the caecum. Six of the endogenous microbiota-dependent miRNAs were experimentally proven in previous studies to target various genes, some of which are expressed in the caecum. These genes are categorized in various Gene Ontology classes including development, DNA binding, protein binding, transcription as well as signaling pathways including Wnt receptor signaling suggesting that the microbiota may be an additional factor controlling these functions. These findings are also in line with our PANTHER analysis where experimentally validated targets that are co-expressed in the caecum also map to some of the same functions of the targets predicted *in silico*.\n\nMoreover, PANTHER, TargetScan and PITA findings collectively reinforce the role that gut bacteria play in organization of the actin cytoskeleton and gut angiogenesis, both previously shown to be affected by gut bacteria at the transcriptional (mRNA) level [@B9],[@B7],[@B76] suggesting that the impact of gut bacteria on specific pathways is many-sided. Particularly, in terms of angiogenesis, global pathway analysis of targets affected by the microbiota-dependent miRNAs illustrate effects on angiogenesis including the process of angiogenesis (p= 4.56E-13) itself and the Vascular Endothelial Growth Factor (VEGF) signaling pathway (p=7.67E-3). Although both the microbiota [@B76] and miRNAs [@B77] have been independently shown to affect vascularization,*in silico* findings in this study establish a possible link between them, and demonstrate a potential mechanism in which the molecular dialogue between gut bacteria and the host is carried out to affect these functions. Gut bacteria are important in the formation of the intestinal vascular network during postnatal development [@B76] and miRNAs in general are known to be developmentally regulated. Since the gut microbiota gradually establishes during postnatal life, it is possible for the two processes to intertwine. Indeed, a recent study showed that exposure to LPS from endogenous *E. coli* in the developing gut of the murine neonate, results in toll-like receptor-4 mediated expression of miR-146a and subsequent down-regulation of interleukin-1 receptor-associated kinase 1 (IRAK1) and the creation of an immunologically tolerant environment [@B78]. Future studies could examine the postnatal expression pattern of the differentially expressed miRNAs and of their target genes.\n\nSeveral studies have shown microbial dysbiosis and miRNA deregulation to be important culprits in a number of digestive diseases, including irritable bowel syndrome [@B36],[@B28] and ulcerative colitis [@B79],[@B32]. Though, it is not understood if and how the two associate to impact these conditions. Based on our findings, we suggest that deregulation of the microbial composition in digestive diseases may at least partially affect the miRNA expression signature, and in turn influence the associated pathologies. One line of evidence involves miR-455, which in our study was found to be up-regulated in the caecum of germ-free mice, and found to target heat-shock factor 1 (*hsf1*) based on bioinformatics analysis. *Hsf1* attenuates the effects of experimentally induced colitis in mice models via indirectly inhibiting the production of pro-inflammatory cytokines, cellular apoptosis and cell adhesion molecule induction [@B80]. Although the authors did not take into account both the microbiota and miRNAs in these mice models we speculate that altered microbial composition in these disease states may affect miRNAs that in turn impact on *hsf1* with potential repercussions on gastrointestinal disease states.\n\nIn both a healthy situation and disease state one of the primary lines of defense in the gastrointestinal tract is the intestinal barrier, of which the gut microbiota is a critical component. Though, at the same time gut microbes act as a regulator of the barrier function at the mRNA level, by impacting the expression of several genes. Recently, genes regulating the intestinal barrier were found to be differentially expressed in the jejunum of intestinal-specific Dicer knock-out mice, highlighting a role for intestinal miRNAs in the regulated expression of intestinal barrier genes [@B81]. In line with this finding, we found that a number of genes included in our intestinal barrier gene set are indeed regulated post-transcriptionally in Dicer knock-out mice and therefore depend on miRNAs. Interestingly, these genes are also the potential targets of gut microbiota-dependent miRNAs (Figure [3](#F3){ref-type=\"fig\"}). These were identified despite a stringent approach excluding 376 of our intestinal barrier genes which are not reported in the PITA and TargetScan databases. Further supporting the existence of an intestinal barrier regulatory network involving miRNAs and the gut microbiota, some of the intestinal barrier genes targeted by our selected miRNAs were found to be up- or down-regulated in Dicer 1-deficient mice versus controls [@B60], suggesting that the microbiota can indirectly impact on the intestinal barrier post-transcriptionally via miRNA regulation. Though, it is important to note that in this study we used whole thickness tissues in order to obtain a comprehensive evaluation of the intestinal miRNA signature response to the commensals.\n\nPhysiologically, the basis of this dialogue has yet to be established; nonetheless, an emerging concept incorporates the utilization of toll-like receptors (TLRs) as potential mediators. For example, miR-147 was found to respond to LPS stimulation of TLR4 in murine peritoneal macrophages, resulting in an attenuated release of Interleukin-6 (IL-6) and Tumour Necrosis Factor alpha (TNF-\u03b1) [@B82]. Moreover, miR-146a was also reported to dampen the inflammatory response upon up-regulation through PAMP activated TLRs [@B83],[@B33]. In turn, these studies show applicability of microbial alterations in miRNA which can impact the barrier function. Therefore, it seems plausible that TLRs which are localized at the interface between the microbiota and the molecular machinery of host cells may be a potential facilitator of this communication.\n\nIn summary, this study shows that the murine caecum expresses a large variety of miRNAs, sixteen of which exhibit differential expression in the presence or absence of the endogenous microbiota. Therefore, gut bacteria may impact on intestinal gene regulation not only at the transcriptional level but also post-transcriptionally; thus, contributing to intestinal homeostasis through fine-tuning gene expression. By modulating miRNAs, the gut microbiota may affect a much larger number of genes than so far expected, particularly in a disease situation where the microbiota composition is altered towards less desirable species. In this perspective, abnormally expressed miRNAs could be considered novel therapeutic targets.\n\nThe authors would like to thank Wen Su and Stefania Arioli for help with dissections and fluorescence *in situ* hybridization. This study was funded by the J. P. Bickell Fdn Medical Research Program. Computational analyses were in part supported by Ontario Research Fund (GL2-01-030), Canada Foundation for Innovation (CFI \\#12301 and CFI \\#203383), and Canada Research Chair Program. NS was in part supported by an Ontario Graduate Scholarship (OGS).\n\nSupplementary Material\n======================\n\n###### \n\nTable S1. Detection levels of 336 mature miRNAs expressed in the caecum of conventionally raised and germ-free mice.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable S2. Experimentally validated microRNA targets.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable S3. List of putative targets of the differentially expressed miRNAs.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable S4. Panther enrichment analysis of the putative targets of differentially expressed miRNAs.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable S5. Intestinal barrier gene set.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable S6. Top expressed miRNAs.\n\n###### \n\nClick here for additional data file.\n\n![**Caecal global microRNA expression is correlated in germ-free and conventional mice**. Scatter plot depicting the relationship between global miRNA expression levels of 336 miRNAs in germ-free (GF) and conventional (Conv) caecal samples (n=3/group) as assessed by qRT-PCR and TLDA plates A and B. Data are presented as mean delta Ct values for each miRNA (mean delta Cts represent the average delta Ct for all three samples/group, with each individual Ct normalized by mean expression value normalization procedure). Of the 585 miRNA species analyzed, 336 remained after filtering and normalization, as explained in Materials and Methods and were used in subsequent analyses.](ijbsv08p0171g01){#F1}\n\n![**Clustering of caecal microRNAs expression profiles in germ-free and conventionally raised mice**. The profiles of 18 transcripts including 16 microRNAs significantly differently expressed (p\\<0.05, FDR\\<20%) between germ-free and conventional caecal samples were visualized using a supervised heatmap (complete linkage and Euclidean distance metric). Expression values range from +2.5 log~2~ to -2.5 log~2~ of \u2206Ct values normalized using mean expression value normalization with positive values (red) indicating higher expression and negative values (green) indicating lower expression in germ-free versus conventional mice. Dendrograms indicate the correlation between groups of samples and genes. Samples are in columns and transcripts in rows.](ijbsv08p0171g02){#F2}\n\n![**Differentially expressed microRNAs impact on the intestinal barrier**. MicroRNAs significantly differentially expressed between germ-free and conventional mice are represented in this diagram by circles, with the colour corresponding to degree of differential microRNA expression in germ-free and conventional samples. Expression values range from +2.5 log~2~ to -2.5 log~2~ of \u2206Ct values with positive values (red) indicating higher expression, and negative values (green), indicating lower expression in germ-free versus conventional mice. Putative intestinal barrier gene targets as identified by the algorithms TargetScan and PITA are represented by triangles. MicroRNAs with a greater number of intestinal barrier targets are symbolized with a larger circle size. Intestinal barrier gene targets that are predicted by both algorithms are indicated by thicker lines. Differentially expressed microRNAs\\* were not included in the diagram as they are not present in both prediction databases. miR-487b did not have any intestinal barrier targets as per the algorithms and therefore was excluded from the figure.](ijbsv08p0171g03){#F3}\n\n###### \n\n**Differentially expressed microRNA in germ-free versus conventional mice**. Real-time PCR analysis of the global expression of microRNAs in the caecum of germ-free (GF) (n=3) versus conventional (CONV) mice (n=3). Fold-change and statistical significance were calculated after mean expression value normalization. Statistical significance is based on Bayes-moderated t-test with a FDR of 20%. In total 18 transcripts were found to be differentially expressed between the two groups with 16 unique mature microRNAs (Y1 is a rat endogenous small RNA). Fold change corresponds to the ratio of mean expression of the microRNA in GF mice to mean expression in CONV mice. Values \\< 1 indicate lower expression in GF (down-regulated in GF mice compared to CONV mice), whereas values \\> 1 indicate higher expression in GF (up-regulated in GF mice compared to CONV mice). Genomic locations and corresponding microRNA sequences (5\\' - 3\\') are based on miRBase version 16. Homology of microRNA sequences between diverse species is derived from miRviewer (last updated November 9, 2008).\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  miRNA              Fold\\         Adjusted\\   Genomic Location                   Mature miRNA Sequence\\              Homology\n                     GF vs. Conv   P-values                                       (5\\' - 3\\')                         \n  ------------------ ------------- ----------- ---------------------------------- ----------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------\n  mmu-miR-21\\*       0.20          0.06        chr11: 86397569-86397660 \\[-\\]     56 -CAACAGCAGUCGAUGGGCUGUC - 77     NA\n\n  rno-miR-351        0.31          0.06        chrX: 139999130-139999210 \\[-\\]    16 -UCCCUGAGGAGCCCUUUGAGCCUGA- 40   *Mmu, Rno*\n\n  mmu-miR-351        0.33          0.20        chrX: 50406432-50406530 \\[-\\]      16 -UCCCUGAGGAGCCCUUUGAGCCUG- 39    *Mmu, Rno*\n\n  mmu-miR-487b       0.45          0.12        chr12: 110965543-110965624 \\[+\\]   50 -AAUCGUACAGGGUCAUCCACUU- 71      *Mmu, Rno, Hsa, Ptr, Mml, Str, Bta, Laf*\n\n  mmu-miR-467a -1\\   0.47          0.06        chr2: 10397973-10398045 \\[+\\]\\     10 -UAAGUGCCUGCAUGUAUAUGCG- 31\\     *Mmu*\n  mmu-miR-467a -2\\                             chr2: 10400425-10400507 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -3\\                             chr2: 10405305-10405387 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -4\\                             chr2: 10407762-10407844 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -5\\                             chr2: 10410226-10410308 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -6\\                             chr2: 10412675-10412757 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -7\\                             chr2: 10415137-10415219 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -8\\                             chr2: 10417607-10417689 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -9\\                             chr2: 10420020-10420102 \\[+\\]\\     15 -UAAGUGCCUGCAUGUAUAUGCG- 36\\     \n  mmu-miR-467a -10                             chr2: 10424900-10424982 \\[+\\]      15 -UAAGUGCCUGCAUGUAUAUGCG- 36      \n\n  mmu-miR-27b\\*      0.54          0.16        chr13: 63402020-63402092 \\[+\\]     7 -AGAGCUUAGCUGAUUGGUGAAC- 28       NA\n\n  mmu-miR-148a       0.58          0.12        chr6: 51219811-51219909 \\[-\\]      61 -UCAGUGCACUACAGAACUUUGU- 82      *Mmu, Hsa, Ptr, Ppy, Mml, Oga, Cpo, Ocu, Opr, Bta, Cfa, Eca, Eeu, Fca, Laf, Tbe, Mlu, Mdo, Gga, Xtr*\n\n  mmu-miR-145        1.52          0.17        chr18: 61807479-61807548 \\[-\\]     7 -GUCCAGUUUUCCCAGGAAUCCCU - 29     *Mmu, Rno, Hsa, Ptr, Ppy, Mml, Mim, Oga, Cpo, Ocu, Opr, Sar, Str, Bta, Cfa, Dno, Eca, Eeu, Ete, Fca, Tbe, Mlu, Mdo, Oan, Dre, Gac, Tru, Xtr*\n\n  mmu-miR-183        1.56          0.17        chr6: 30119668-30119737 \\[-\\]      6 -UAUGGCACUGGUAGAAUUCACU- 27       *Mmu, Rno, ,Hsa, Ptr, Ppy, Mml, Mim, Oga, Opr, Sar, Str, Bta, Cfa, Dno, Eca, Ete, Fca, Tbe, Mlu, Mdo, Oan, Gga, Dre, Gac, Ola, Tni, Tru, Xtr, Cin*\n\n  mmu-miR-133a -1\\   1.61          0.12        chr18: 10782907-10782974 \\[-\\]\\    43 -UUUGGUCCCCUUCAACCAGCUG- 64\\     miR-133a-1: *Mmu, Hsa, Ptr, Cpo, Ocu, Opr, Sar, Str, Bta, Cfa, Dno, Eca, Eeu, Ete, Fca, Tbe, Gga, Dre, Gac, Ola, Tru*\\\n  mmu-miR-133a-2                               chr2: 180133084-180133187 \\[+\\]    59 -UUUGGUCCCCUUCAACCAGCUG- 80      miR-133a-2: *Mmu, Hsa, Ptr, Gga, Dre*\n\n  mmu-miR-150        1.66          0.15        chr7: 52377127-52377191 \\[+\\]      6 -UCUCCCAACCCUUGUACCAGUG- 27       *Mmu, Rno, Hsa, Ptr, Ppy, Mml, Mim, Opr, Sar, Str, Bta, Cfa, Eca, Fca, Tbe, Oan, Dre, Ola, Xtr*\n\n  Y1-4386739.B       2.02          0.15        NA                                 NA                                  NA\n\n  mmu-miR-672        2.74          0.06        chrX: 101311514-101311613 \\[-\\]    25 -UGAGGUUGGUGUACUGUGUGUGA- 47     *Mmu, Rno, Mim, Oga, Cpo, Ocu, Str, Dno, Eca, Ete, Cin*\n\n  mmu-miR-181a-1\\*   2.79          0.17        chr1: 139863032-139863118 \\[+\\]    54 -ACCAUCGACCGUUGAUUGUACC- 75      NA\n\n  rno-miR-664 -1\\    2.85          0.06        chr18: 47881354-47881412 \\[+\\]\\    38 -UAUUCAUUUACUCCCCAGCCUA- 59\\     miR-664-1: *Mmu, Rno, Eca*\\\n  rno-miR-664-2                                chr13: 101253993-101254051 \\[+\\]   38 -UAUUCAUUUACUCCCCAGCCUA- 59      miR-664-2: *Rno, Laf*\n\n  mmu-miR-455        3.00          0.15        chr4: 62917885-62917966 \\[+\\]      54 -GCAGUCCACGGGCAUAUACAC- 74       *Mmu, Rno, Hsa, Ptr, Ppy, Mml, Mim, Oga, Cpo, Ocu, Opr, Str, Bta, Cfa, Eeu, Ete, Fca, Laf, Mlu, Mdo, Gga, Dre, Gac, Ola, Tni, Tru, Xtr*\n\n  mmu-miR-138\\*      4.43          0.12        chr9: 122591994-122592092 \\[+\\]    61 -CGGCUACUUCACAACACCAGGG- 82      NA\n\n  mmu-let-7g\\*       4.60          0.10        chr9: 106081171-106081258 \\[+\\]    63 -ACUGUACAGGCCACUGCCUUGC- 84      NA\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n[^1]: \\# Present address: Department of Biostatistics, Harvard School of Public Health, Building 1, Room 412C, 655 Huntington Avenue, Boston, Massachusetts 02115, USA\n\n[^2]: \\* Present address: The Biomedical Research Centre, Department of Medical Genetics, Faculty of Medicine, The University of British Columbia, Vancouver, Canada\n"}
+{"text": "Introduction {#s1}\n============\n\nAs an essential but frequently less available nutrient for plant growth, phosphorus (P) is taken up by plants as orthophosphate (H~2~PO~4~ ^\u2212^, Pi) mainly through Pi transporters and driven by a proton gradient generated by plasma membrane H^+^-ATPases [@pone.0047726-UllrichEberius1]. In native soil solution, Pi concentration is always less than 10 \u00b5M because it is easily bound by either soil organic matter or minerals [@pone.0047726-Holford1], [@pone.0047726-Marschner1], [@pone.0047726-Raghothama1], [@pone.0047726-Schachtman1]. Meanwhile, the Pi concentration in the cytoplasm of plant cells is generally greater than 10 mM [@pone.0047726-Mimura1]. Therefore, plants must have specialized transporters to transport Pi from soil solution to plant cells against a large concentration gradient at the root-soil interface. Recent genome sequence analysis and experimental evidence indicated that plants contained a wide variety of Pi transporter families, including Pht1, Pht2, Pht3, Pht4, which were defined by protein sequence, structure, locations and functions [@pone.0047726-Raghothama1], [@pone.0047726-Schachtman1], [@pone.0047726-Liu1], [@pone.0047726-Guo1], [@pone.0047726-Knappe1], [@pone.0047726-Picault1], [@pone.0047726-Rausch1].\n\nAmong the Pi transporter families in plants, Pht1 family is most widely studied. All the members in the Pht1 family have the same predicted structure, including 12 putative membrane-spanning domains, hydrophilic N- and C-terminals, a hydrophilic loop between transmembrane segments (TM) six and seven, a putative glycosylation site in TM10 and two cytoplasmic phosphorylation sites [@pone.0047726-Karandashov1]. Since cloning of the first *Pht1* family gene from Arabidopsis [@pone.0047726-Muchhal1], many *Pht1* genes have been isolated from a number of plant genomes, including Arabidopsis [@pone.0047726-Mudge1], graminaceous species [@pone.0047726-Ai1], [@pone.0047726-Paszkowski1], [@pone.0047726-Schnmann1], [@pone.0047726-Schnmann2], [@pone.0047726-Nagy1], solanaceous species [@pone.0047726-Liu2], [@pone.0047726-Nagy2], [@pone.0047726-Chen1] and legumes [@pone.0047726-Wu1], [@pone.0047726-Liu3], [@pone.0047726-Javot1], [@pone.0047726-Harrison1]. Most Pi transporter genes in the Pht1 family are expressed in roots, while a few are expressed in aerial parts, including leaves, stems, cotyledons, tubers, flowers, grains and seeds [@pone.0047726-Mudge1], [@pone.0047726-Ai1], [@pone.0047726-Nagy1], [@pone.0047726-Karthikeyan1], [@pone.0047726-Rae1], implying their potential involvement in Pi internal translocation. Pi transporter genes in the Pht1 family from Arabidopsis and rice, containing 9 and 13 members, respectively, have been comprehensively studied and well characterized [@pone.0047726-Liu1], [@pone.0047726-Mudge1], [@pone.0047726-Paszkowski1], [@pone.0047726-Ai2]. All results indicate that there are distinct functions and different responses to P deficiency among *Pht1* family genes.\n\nIn Arabidopsis, eight of nine Pi-transporter genes are expressed in roots. Fusion of the promoter regions from these genes with the GUS reporter gene indicates that four of them are highly expressed in the root epidermis and the expression is enhanced by P deficiency. Additionally, some members are expressed in shoot tissues, such as pollen grains, and thereby implying a wider role in Pi uptake and remobilization [@pone.0047726-Mudge1]. In rice, nine out of thirteen *Pht1* transporter genes are expressed in both Pi-deprived roots and leaves. The transcript levels of *OsPT2, OsPT3, OsPT6* and *OsPT7* are significantly enhanced by P deficiency in roots. The expressions of *OsPT1* and *OsPT8* are abundant in both roots and leaves at two P levels [@pone.0047726-Ai2], [@pone.0047726-Jia1]. For legumes, Pi-transporters in the Pht1 family in *Medicago truncatula* have been well studied. Among them, *MtPT1, MtPT2, MtPT3* and *MtPT5* are highly expressed in Pi-deprived roots, but less with addition of high Pi [@pone.0047726-Liu3]. In *Lotus japonicus*, 3 Pi transporters genes in Pht1 family have been isolated [@pone.0047726-Maeda1].\n\nSoybean (*Glycine max* (L.) Merr.) is one of the most widely grown leguminous crops in the world. However, soybean production is limited by various environmental factors, especially by low P availability in soils [@pone.0047726-Bureau1]. It might help us to find some new approaches to improve the P efficiency of soybean through understanding the detailed characteristics of *Pht1* genes. Compared to the *Pht1* genes in Arabidopsis and rice, much less work has been done in soybean. Recently, two members from the soybean Pht1 family (*GmPT1* and *GmPT2*) were reported to be plasma membrane proteins, and complementation analysis in yeast indicated that they might be constitutively expressed low affinity Pi transporters [@pone.0047726-Wu1]. Therefore, these two Pi transporters might not be critically involved in response to P deficiency. The discovery and functional characterization of Pi transporters responsible for plant Pi uptake under limited P conditions should be more important to provide a clearer understanding of how plants coordinate the use of Pi to support growth and development.\n\nPlant growth in soils can also be constrained by other nutrients rather than P, and therefore plants might evolve adaptive mechanisms to cope with multiple nutrient stresses. However, only a few studies have reported on interactions among P and other nutrients. For example, coincident potassium (K) and P deprivation induced the transcriptions of a MAP kinase gene, transcription factors and nutrient transporters in tomato [@pone.0047726-Wang1]. After long term low P treatment, some iron (Fe) and sulfur (S) transporters were up regulated in Arabidopsis [@pone.0047726-Misson1]. Cross-talks between P and Fe have also been demonstrated in rice [@pone.0047726-Wasaki1]. The expression of *AAR6* was up-regulated by N, P, or K deprivation, indicating that shared nutrient signaling transduction pathways might exist in higher plants [@pone.0047726-Coello1]. As the most responsive gene family to Pi starvation, *Pht1* family genes are likely to be involved in those shared pathways. But up to date, there have been no reports about regulation of *Pht1* family genes by nutrients other than P.\n\nIn this study, 14 Pi transporter genes from the *Pht1* family (*GmPTs*) were isolated from the soybean genome. A yeast Pi uptake-defective mutant was used to characterize the Pi uptake kinetics of GmPTs through radioactive ^33^P uptake analysis. Tissue specificity and regulation by P as well as N, K and Fe starvation of all 14 *GmPTs* were analyzed through quantitative RT-PCR (qRT-PCR).\n\nResults {#s2}\n=======\n\nSoybean Pi Transporter Genes in the *Pht1* Family {#s2a}\n-------------------------------------------------\n\nA search of the Phytozome soybean genome database (<http://www.phytozome.net/soybean>) yielded a total of 14 sequences identified as being related to high-affinity Pi transporters. According to the recommended nomenclature for plant Pi transporters (<http://www.botanik.uni-koeln.de/bucher_ppnomenclature.html>), these 14 identified genes were named as *GLYma;Pht1;1* through *GLYma;Pht1;14*, with the order being based on their chromosome locations. For simplification, the Pi transporters are called *GmPT1* through *GmPT14* in this report ([Table 1](#pone-0047726-t001){ref-type=\"table\"}). BLAST analysis against the Pfam database (<http://www.sanger.ac.uk/resources/databases/pfam.html>) showed that they all belonged to the MFS family. The full-length cDNAs and amino acid sequences of these 14 Pi transporters were available in the Phytozome website. All 14 GmPTs exhibited a high degree of homology and were comparable in length, calculated molecular weight and theoretical pI value ([Table 1](#pone-0047726-t001){ref-type=\"table\"}). The distribution of *GmPTs* on soybean chromosomes is uneven and involves only 8 of 20 chromosomes. Chromosome 10 had the highest number of *GmPTs* genes with 4, followed by chromosome 20 with 3 ([Table 1](#pone-0047726-t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0047726.t001\n\n###### Members of Pi transporter genes in the Pht1 family from soybean.\n\n![](pone.0047726.t001){#pone-0047726-t001-1}\n\n  Gene name    Accessionnumber    Locus tag.     aa     kD      pI\n  ----------- ----------------- --------------- ----- ------- ------\n  *GmPT1*         FJ814697       Glyma02g00840   533   58.47   8.31\n  *GmPT2*         FJ814696       Glyma03g31950   539   59.27   8.52\n  *GmPT3*         FJ814701       Glyma07g34870   516   58.31   8.64\n  *GmPT4*         JQ518269       Glyma10g00950   533   58.42   8.54\n  *GmPT5*         FJ814694       Glyma10g04230   521   57.31   8.63\n  *GmPT6*         FJ814693       Glyma10g33020   502   55.44   9.28\n  *GmPT7*         FJ814695       Glyma10g33030   536   58.73   8.34\n  *GmPT8*         FJ814700       Glyma13g08720   519   57.64   8.56\n  *GmPT9*         FJ814698       Glyma14g28780   525   58.24   8.33\n  *GmPT10*        FJ814699       Glyma14g36650   529   58.19   7.91\n  *GmPT11*        JQ518270       Glyma19g34710   539   59.27   8.25\n  *GmPT12*        FJ814692       Glyma20g02660   506   56.78   8.63\n  *GmPT13*        FJ789662       Glyma20g34610   536   58.63   7.63\n  *GmPT14*        JQ518271       Glyma20g34620   527   58.11   8.93\n\nTheir accession number, locus tag on chromosomes, the respective numbers of amino acids (aa), the calculated molecular mass in kiloDaltons (kD), and theoretical pI value are given.\n\nLike Pht1 transporters in other plant species, the *PT* genes in soybean encode proteins with comparable predicted secondary structures, with each being characterized by 12 hydrophobic domains presumably spanning the plasma membrane. Amino acid sequences of soybean Pi transporter homologs were aligned and compared with each other, and conserved amino acid residues were found ([Figure S1](#pone.0047726.s001){ref-type=\"supplementary-material\"}). We calculated the relatedness of soybean Pht1 transporters using the DNAMAN computer program, with the results suggesting that the percentage of identical protein sequences ranges from 48% to 99% ([Table S1](#pone.0047726.s004){ref-type=\"supplementary-material\"}). The highest percentage of amino acid sequence identity was found between GmPT2 and GmPT11 (99.27%), followed by GmPT6 and GmPT14 (98.23%). The lowest identity was observed between GmPT3 and GmPT10 (47.63%). GmPT3 and GmPT12 had a very high amino acid sequence identity of 93.89%, but both proteins had comparatively low amino acid sequence identity with all the other 12 Pht1 Pi transporter proteins.\n\nPhylogenetic Analysis of GmPTs {#s2b}\n------------------------------\n\nCombining GmPT1 through GmPT14 with Pi transporter protein sequences in rice, Arabidopsis and *Medicago truncatula,* we constructed a phylogenetic tree using neighbor-joining analysis in the MEGA 4.1 program ([Figure 1](#pone-0047726-g001){ref-type=\"fig\"}). The phylogenetic results demonstrated that like the other three species, soybean GmPTs were divided into two distinct groups with strong bootstrap support. Group I contained 12 of the 14 GmPTs, 9 of 10 MtPTs, 7 of the 9 AtPhts and 11 of the 13 OsPTs. Group I could be further divided into two subgroups: I-1 and I-2. Subgroup I-1 contained 9 GmPTs, 7 MtPTs, 7 AtPhts and 10 rice Pi transporters. It is interesting to find that in subgroup I-2, GmPT8, GmPT9 and GmPT10 were clustered with the mycorrhiza-specific OsPT11 from rice [@pone.0047726-Paszkowski1] and MtPT4 from *Medicago truncatula* [@pone.0047726-Liu3], [@pone.0047726-Javot1], and no Arabidopsis AtPTs, implying possible roles for these three GmPTs in Arbuscular Mycorrhizal (AM) symbiosis. Group II only included a few Pi transporters, a pair from soybean (GmPT3 and GmPT12), a pair from Arabidopsis (AtPht1;8 and AtPht1;9), MtPT6 from *Medicago truncatula* and a pair from rice (OsPT9 and OsPT10). Both group I and group II contained Pht1 family Pi-transporters from all four species, indicating that divergence of Pi-transporters preceded divergence of dicots and monocots. By exploring the intron (the noncoding sequence between two coding sequences within a gene) and exon (the protein-coding region in the DNA) structures of GmPTs, we found the genes from the same group had the similar intron/exon structures, such as GmPT3 and GmPT12 in group II, their structures were totally different from those in Group I by characterized with three exons and two long introns, indicating the divergence of structures and functions of different GmPTs ([Figure S2](#pone.0047726.s002){ref-type=\"supplementary-material\"}).\n\n![Phylogenetic tree of soybean, Arabidopsis, rice and Medicago plant Pi transporter proteins in Pht1 family.\\\nTransporters and corresponding plant species are as follows: rice (*Oryza sativa*), OsPT1 through OsPT13 [@pone.0047726-Paszkowski1]; Arabidopsis (*Arabidopsis thaliana*), AtPht1;1 through AtPht1;9 [@pone.0047726-Mudge1]; Medicago (*Medicago truncatula*) MtPT1 through MtPT6 [@pone.0047726-Grunwald1] and other four PT proteins obtained in Phytozome (<http://www.phytozome.net/medicago>), soybean (*Glycine Max*), GmPT1 through GmPT14 (this work).](pone.0047726.g001){#pone-0047726-g001}\n\nAnalysis of Pi Transport Activities of GmPTs in a Yeast Strain Defective in Pi-uptake {#s2c}\n-------------------------------------------------------------------------------------\n\nTo analyze and compare the Pi transport activities of all the GmPTs, their coding regions were separately cloned into a yeast expression vector (p112A1NE), under the control of the yeast alcohol dehydrogenase promoter. The constructs were separately introduced into a yeast Pi transport mutant MB192, which lacks the function of the high-affinity Pi transporter gene *PHO84*. An empty vector was also transformed to be used as a control.\n\nBecause the Pht1 transporters are members of the H^+^/Pi symporter family, the pH dependence of GmPTs during Pi transport was assayed by measuring the optical density of the yeast cell lines at pH values ranging from 4 to 8. The pH optima for most of the yeast mutant cells carrying p112-GmPTs was 6, in comparison to the wild type, in which the pH optima ranged from 4 to 6. Therefore, the pH value was set as 6 in the subsequent studies.\n\nThe yeast MB192 strain grows poorly when supplied with limiting amounts of Pi. As shown in [Figure 2](#pone-0047726-g002){ref-type=\"fig\"}, under low P conditions (20 \u00b5M), all the *GmPT* transformants grew much better than the control, suggesting that GmPTs can complement the yeast Pi-uptake mutant under Pi-limiting conditions ([Figure 2](#pone-0047726-g002){ref-type=\"fig\"}).\n\n![Complementation of a yeast inorganic phosphate (Pi) transport mutant by *GmPTs* genes.\\\nYeast MB192 cells harboring either an empty expression vector (control) or the candidate Pht1 ORF(open reading frame), transformants were grown in YNB medium to an OD600 \u22480.8, then washed by 3% glucose with centrifugation at 1500 g, 4\u00b0C, and suspended in phosphate free YNB medium to OD600 \u22481.0. Different number cells (5\u00d710^5^, 5\u00d710^4^, 5\u00d710^3^) were applied to Pi-limiting medium (20 \u00b5M, pH 6.0) then incubated at 30\u00b0C for 3 d.](pone.0047726.g002){#pone-0047726-g002}\n\nTo make sure the effects of the expressions of *GmPT* genes in yeast on the growth of transformed yeast cells are Pi-deficiency dependent, but not just because of the expressions of *GmPT* genes, we analyzed the kinetic growth of Yp112-GmPTs transformants, wild type and empty vector control in normal YNB liquid medium with 2 mM Pi, and found that there were no significant growth differences between empty vector control and all the 14 Yp112-GmPTs transformants ([Figure S3D](#pone.0047726.s003){ref-type=\"supplementary-material\"}), But under certain low P conditions, Yp112-GmPTs transformants always grew better than the empty vector control ([Figure S3A](#pone.0047726.s003){ref-type=\"supplementary-material\"}, B, C), indicating the complementation of GmPTs was really Pi-deficiency dependent.\n\nFurthermore, in order to determine the kinetic properties of GmPTs,^33^P was employed in Pi-uptake experiments using transformed yeast cells. A Line weaver-Burk plot indicated that Pi uptake mediated by 12 *GmPT* genes followed Michaelis-Menten kinetics ([Figure 3](#pone-0047726-g003){ref-type=\"fig\"}) with the estimated *Km* values ranged from 25.70\u00b11.63 \u00b5M to 116.30\u00b110.00 \u00b5M (mean\u00b1SE) ([Table 2](#pone-0047726-t002){ref-type=\"table\"}), while only GmPT6 and GmPT10 showed no difference when Pi concentration in the growth medium was less than 0.1 mM. The results from kinetic studies above demonstrated that most of the GmPTs were high-affinity Pi transporters.\n\n10.1371/journal.pone.0047726.t002\n\n###### Kinetic parameter estimates of GmPTs-mediated inorganic phosphate (Pi) transport.\n\n![](pone.0047726.t002){#pone-0047726-t002-2}\n\n  Yeast             *Km*(\u00b5M)     *Vmax*(\u00b5M/g \\[dry weight\\]\u00b7min)\n  -------------- -------------- ---------------------------------\n  Yp112-GmPT1     67.30\u00b115.60             298.30\u00b124.40\n  Yp112-GmPT2     44.00\u00b112.90             295.00\u00b120.90\n  Yp112-GmPT3      30.00\u00b18.60             226.70\u00b116.30\n  Yp112-GmPT4     65.30\u00b113.30             282.30\u00b143.28\n  Yp112-GmPT5      25.70\u00b11.63             231.30\u00b118.80\n  Yp112-GmPT6     153.00\u00b144.00            281.70\u00b137.60\n  Yp112-GmPT7     105.30\u00b120.00            247.00\u00b124.50\n  Yp112-GmPT8      46.00\u00b15.60              231.50\u00b15.50\n  Yp112-GmPT9      88.00\u00b18.90              240.30\u00b17.10\n  Yp112-GmPT10    231.00\u00b115.60            400.00\u00b125.00\n  Yp112-GmPT11     79.00\u00b18.30             271.30\u00b145.00\n  Yp112-GmPT12    45.30\u00b114.53             245.30\u00b115.60\n  Yp112-GmPT13     32.30\u00b11.60             265.70\u00b132.50\n  Yp112-GmPT14    116.30\u00b110.00            373.00\u00b114.10\n  Yp112           127.00\u00b18.50              154.70\u00b13.60\n\n*Km* and *Vmax* for yeast strain MB192 expressing the indicated *GmPTs* or carrying the empty expression vector (control) were determined at pH 6.0. The GmPTs mediated ^33^Pi uptake velocities, calculated according to their total Pi transport, following the Michaelis-Menten kinetics equation. Values shown are means \u00b1 SE for three independent experiments.\n\n![Kinetic analysis of inorganic phosphate (Pi) uptake in yeast.\\\nThe non-linear regression of total Pi uptake by strain Yp112-GmPTs versus external Pi concentration at pH 6 were used to estimate the apparent *Km* value for Pi uptake. All the results were calculated from the three independent experiments.](pone.0047726.g003){#pone-0047726-g003}\n\nExpression Patterns of *GmPTs* as Regulated by P Availability in Different Tissues {#s2d}\n----------------------------------------------------------------------------------\n\nAccording to the nucleic acid sequences ([Text S2](#pone.0047726.s010){ref-type=\"supplementary-material\"}), we designed gene-specific primers for the 14 *GmPTs* genes ([Table S2](#pone.0047726.s005){ref-type=\"supplementary-material\"}). Tissue specificity and P response of the *GmPTs* were investigated using qRT-PCR. The main expression pattern under high P conditions was consistent with the results from the two soybean transcriptome atlases ([Table S3](#pone.0047726.s006){ref-type=\"supplementary-material\"}) [@pone.0047726-Libault1], [@pone.0047726-Severin1], such as the expression levels of most *GmPTs* were low, and roots had the highest number of *GmPTs* genes to express in, indicating the important roles of Pi transporters in Pi transport and absorption in roots; *GmPT5* mainly expressed in flowers, implying possible functions in Pi transport for source to sink ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}).\n\n![Spatial Expression pattern analysis for the 14 *GmPTs* as related to P availability.\\\nPlants were grown on low P (added 5 \u00b5M P as KH~2~PO~4,~ open bars) and high P (added 500 \u00b5M P as KH~2~PO~4,~ closed bars) conditions. Young leaves (YL), roots (R), stems (S) and flowers (F) were sampled 18 days after treatment initiation, and young pods (P) and seeds (SE) were sampled 29 days after treatment initiation. Each bar is the mean of three biological replications with standard error. Note: different scales are used in the graphs; asterisks indicate significant differences of *GmPTs* expression in certain tissues under low P and high P conditions in *t*-tests.](pone.0047726.g004){#pone-0047726-g004}\n\nThe expression of *GmPLDZ* (Glyma20g38200) which has been well demonstrated as a Pi starvation induced gene [@pone.0047726-Li1], was also analyzed presently. The results found that the *GmPLDZ* expression was highly enhanced under low P conditions, especially in leaves ([Figure 5A](#pone-0047726-g005){ref-type=\"fig\"}), proving that the P treatments in the present study were sufficient for analyzing the responses of Pi starvation induced genes. Like Pi transporters in other plant species, the expressions of *GmPTs*, except *GmPT10*, were highly induced by P deficiency ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}). Seven, including *GmPT1*, *GmPT2*, *GmPT3*, *GmPT4, GmPT7 GmPT8* and *GmPT12*, were predominantly expressed in roots. *GmPT5* and *GmPT14* were mainly expressed in flowers and stems, respectively. *GmPT9* and *GmPT13* were highly induced in roots and stems, and young leaves and roots, respectively. *GmPT6* and *GmPT11* were expressed highly in young leaves, roots and stems ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}). In addition, Pi starvation induced *de novo* synthesis of *GmPT3* and *GmPT4* in roots and *GmPT11* in young leaves, roots and stems. In contrast, *GmPT10* was expressed in stems, flowers and seeds at high P level. Finally, low or even undetectable expressions of *GmPTs* were observed in pods ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}).\n\n![Expression of P, N, K or Fe responsive genes to different nutrient stresses.\\\n14-day old soybean seedlings were treated with N (\u2212N), K (\u2212K) and Fe (\u2212Fe) deficiencies (see Experimental procedures for details). Seedlings grown under normal solution were used as controls (CK, added 500 \u00b5M P as KH~2~PO~4~). The expression levels in shoots and roots were analyzed by quantitative real-time PCR. Soybean gene *PLDZ* (Glyma20g38200) was used as low P responsive gene (A), *NiR* (Glyma02g14910) for low N treatment (B), *HAK* (Glyma3g42480) for low potassium treatment (C) and *IRT* (Glyma07g34930) for low Fe treatment (D), respectively. Each bar was the mean of three biological replications with standard error.](pone.0047726.g005){#pone-0047726-g005}\n\nResponses of *GmPTs* to Nitrogen, Potassium and Iron Deficiency {#s2e}\n---------------------------------------------------------------\n\nTo examine the potential induction of *GmPT* genes by other nutrient deficiencies, transcript abundance was assessed by qRT-PCR in soybean plants separately grown in nutrient solution deficient in nitrogen (N), potassium (K) or iron (Fe) for 14 days. Three marker genes, including *GmHAK* (Glyma3g42480, a K transporter) and *GmIRT* (Glyma07g34930, an iron transporter) which had been known respectively enhanced by K or Fe deficiency [@pone.0047726-Gierth1], [@pone.0047726-Eide1], and *GmNiR* (Glyma02g14910, a nitrite reductase gene) which was repressed by N deficiency [@pone.0047726-Wang2], were used to monitor the treatmental conditions. As expected, the expression of *GmHAK*, *GmIRT* and *GmNiR* was significantly regulated by K, Fe or N deficiency, respectively ([Figure 5B, C, D](#pone-0047726-g005){ref-type=\"fig\"}).\n\nThe expressions of *GmPT* genes in response to N, K, or Fe deficiency were shown in [Figure 6](#pone-0047726-g006){ref-type=\"fig\"}. The 14 Pi transporter genes had very low expressions in leaves and roots under normal conditions. All of the gene expression levels were modified by one or more of the three stresses in different tissues. Under N deficiency conditions, the expressions of 10 genes were up-regulated more than 2-fold. Among them, *GmPT6, GmPT7, GmPT8, GmPT11* and *GmPT14* were increased in leaves, and *GmPT12* was in roots. N deficiency enhanced the expressions of *GmPT2, GmPT3, GmPT5* and *GmPT13* in both leaves and roots. K deficiency enhanced the expressions of 8 *GmPTs* genes. *GmPT3* was up-regulated in both leaves and roots, while the expressions of the other 7 *GmPTs*, including *GmPT1, GmPT2, GmPT4, GmPT5, GmPT8, GmPT9* and *GmPT10,* were only enhanced in roots. Under Fe deficiency, *GmPT3, GmPT12* and *GmPT13* were up-regulated. The expression levels of the remaining *GmPT* genes were not altered significantly under Fe deficiency. Interestingly, *GmPT3* was the only one gene with expression up-regulated by N, P or K deficiency simultaneously in both leaves and roots ([Figure 6](#pone-0047726-g006){ref-type=\"fig\"}), suggesting that *GmPT3* might be involved in an universal regulation network in response to multiple nutrient deficiencies.\n\n![Responses of *GmPTs* to different nutrient stresses.\\\nTen-day old soybean seedlings were treated with N (\u2212N), K (\u2212K) and Fe (\u2212Fe) deficiencies (see Experimental procedures for details). Seedlings grown under normal solution were used as controls (CK, added 500 \u00b5M P as KH~2~PO~4~). Asterisks indicated the significant differences of *GmPTs* expression between nutrients deficient stresses and normal conditions in Student's t-tests.](pone.0047726.g006){#pone-0047726-g006}\n\nDiscussion {#s3}\n==========\n\nPhosphorus is particularly critical for legume growth due to the huge demands for protein and oil synthesis, as well as, for biological N~2~ fixation [@pone.0047726-SnchezCaldern1]. However, low P availability is a worldwide constraint for crop growth in most soils. Therefore, to understanding Pi uptake and translocation in soybean can dramatically facilitate soybean adaptation to the low P soils. Phosphate transporter genes in the *Pht1* family have been reported playing important roles in Pi uptake and translocation in plants [@pone.0047726-Mudge1], [@pone.0047726-Paszkowski1], [@pone.0047726-Schnmann1], [@pone.0047726-Nagy1], [@pone.0047726-Chen1], [@pone.0047726-Liu3], [@pone.0047726-Nagy3], [@pone.0047726-Rausch2]. To date, information on this family is limited in soybean. In the current study, we identified and isolated 14 members of PTs in the Pht1 family from the soybean genome, and characterized their functions in Pi uptake using yeast mutants defective in high affinity Pi transport. The expression patterns in terms of tissue specificity, responses to P deficiency and other nutrient stresses were also analyzed.\n\nSoybean Pi Transporter Genes in *Pht1* Family {#s3a}\n---------------------------------------------\n\nLike *Pht1* family genes in other species, soybean PTs in the Pht1 family exhibited a high degree of homology, characterized by high protein identity ([Table S1](#pone.0047726.s004){ref-type=\"supplementary-material\"}) and similar hydrophobic domains presumably spanning the plasma membrane ([Figure S1](#pone.0047726.s001){ref-type=\"supplementary-material\"}). Interestingly, phylogenetic analysis of GmPT1 through GmPT14, along with Pi transporter protein sequences from *Medicago truncatula*, Arabidopsis, and rice, showed a similar pattern of evolutionary divergence within dicotyledons (Arabidopsis, Medicago and soybean) and monocotyledons (rice). These phylogenetic relationships among the GmPTs ([Figure 1](#pone-0047726-g001){ref-type=\"fig\"}), may reflect biochemical and functional differences among different types of PTs.\n\nWith well-supported bootstrap values, GmPT8 and GmPT9 were clustered with the mycorrhiza-specific OsPT11 from rice [@pone.0047726-Paszkowski1] and MtPT4 from *Medicago truncatula*. Since OsPT11 is specifically induced by AM symbiosis, and MtPT4 is critical for AM symbiotic Pi transport and development, the absence of MtPT4 will lead to the inability of the fungus to proliferate within roots, and termination of symbiosis [@pone.0047726-Javot1]. We speculate that, given the phylogenetic proximity, GmPT8 and GmPT9 are likely to play important roles in the symbiosis between soybean and AM fungus. GmPT10 was classified into the same subgroup with GmPT8 and GmPT9, though with a weaker bootstrap value than that between GmPT8 and GmPT9, implying a possible role for GmPT10 in symbiosis with AM fungi, but its functions might be different from those of GmPT8 and GmPT9 in AM roots.\n\nGmPT3 and GmPT12, which had relatively lower protein sequence identity with the other GmPTs ([Table S1](#pone.0047726.s004){ref-type=\"supplementary-material\"}), were classified into the group II ([Figure 1](#pone-0047726-g001){ref-type=\"fig\"}) of *Pht1* genes. By exploring gene structures, we found these two genes shared a similar and slightly unusual structure with a relatively large intron in the coding sequence and more exons than the other GmPTs, which had only one or two exons ([Figure S2](#pone.0047726.s002){ref-type=\"supplementary-material\"}). These structural differences might be related to the special functions of GmPT3 and GmPT12 differing from other GmPTs. Recent reports have indicated that the size of exon and intron, and their intergenic distance are correlated with gene expression levels and expression breadth in soybean [@pone.0047726-Woody1]. The relationship between gene structures and functions of GmPTs needs to be further analyzed.\n\nFunctions of GmPTs in Pi Uptake {#s3b}\n-------------------------------\n\nThree systems have been well accepted to functionally characterize plant Pi transporters, including the complementation of yeast mutants defective in Pi transport, monitoring Pi uptake using *Xenopus laevis* oocytes and ectopic expression of Pi transporters in plant suspension cells [@pone.0047726-Ai1], [@pone.0047726-Rae1], [@pone.0047726-Jia1], [@pone.0047726-Daram1], [@pone.0047726-Mitsukawa1], [@pone.0047726-Miller1]. The yeast mutants defective in Pi transporters are most widely used for characterizing the kinetic properties of many PTs from different plants. With complemented yeast mutants, kinetic studies allow for quantifying the affinity of transporters [@pone.0047726-Ai1], [@pone.0047726-Daram1], [@pone.0047726-Leggewie1]. However, few whole Pht1 families in crops have been functionally analyzed [@pone.0047726-Ai1], [@pone.0047726-Jia1]. The yeast Pi transport mutant MB192, which lacks function of the high-affinity Pi transporter gene *PHO84*, has been used to analyze Pi transporters from several crop species [@pone.0047726-Jia1], [@pone.0047726-Yadav1]. In the present study, we used the MB192 yeast system to work on kinetic analysis of all 14 GmPTs, and found that 12 of 14 GmPTs acted as high-affinity Pi transporters ([Table 2](#pone-0047726-t002){ref-type=\"table\"}, [Figure 3](#pone-0047726-g003){ref-type=\"fig\"}), implying their functions in high-affinity uptake and transport of Pi from low P soils. However, among the five reported *Pht1* family genes in *Medicago truncatula*, only MtPT5 showed high affinity for Pi uptake [@pone.0047726-Liu3], [@pone.0047726-Harrison1]. The possible reason why we could identify so many high-affinity Pi transporters here using the yeast system is that due to the existence of native Pi transport systems, the expression of high-affinity Pi transporters in yeast cells might display altered transport properties. Some high-affinity Pi transporters, such as AtPHT1;1 and HvPHT1;1, were not able to complement the yeast mutant. We also found that among the tested GmPTs, although GmPT6 and GmPT10 could complement the Pi absorption of the yeast mutant MB192 to a certain extent ([Figure 2](#pone-0047726-g002){ref-type=\"fig\"}), but their Km values were higher than that of the tested mutant based on the kinetic analysis ([Figure 3](#pone-0047726-g003){ref-type=\"fig\"}, [Table 2](#pone-0047726-t002){ref-type=\"table\"}). Therefore, we still characterized them as low-affinity Pi transporters. However, there is not a defined Km value for categorizing the Pi affinity of PTs, and the external Pi concentrations used here were limited, further works are needed to test weather these two transporters have dual-affinity for Pi uptake.\n\nCharacterization of the *GmPTs* in Response to P Availability {#s3c}\n-------------------------------------------------------------\n\nRecently, two soybean transcriptome atlases providing a record of high-resolution soybean gene expressions have been reported by two groups using next generation sequencing technique [@pone.0047726-Libault1], [@pone.0047726-Severin1]. The expression patterns of the 14 *GmPTs* with normal Pi supply in the present study are well consistent with those atlases. Most of the *GmPTs* showed low expressions in most tissues, and some were undetectable ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}) [@pone.0047726-Libault1], [@pone.0047726-Severin1]. Differently, we also found that *GmPT10* was mainly expressed in stems, flowers and seeds in high P conditions, implying it might be a low-affinity Pi transporter functioning in internal Pi translocation. The differences among our results and the previously reported transcriptome atlases might be due to the different sampling stages and growth systems. In addition to work on normal growth conditions, we tested the responses of the 14 *GmPTs* to Pi starvation, and found that most of the *Pht1* genes had elevated transcript levels under low P conditions, while a few of them were differentially expressed between roots and aerial parts of the plants as previously reported [@pone.0047726-Muchhal1], [@pone.0047726-Mudge1], [@pone.0047726-Schnmann1], [@pone.0047726-Liu3], [@pone.0047726-Rae1], [@pone.0047726-Daram1], [@pone.0047726-Leggewie1]. Thirteen *GmPTs* were induced or increased under low P conditions in soybean ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}), indicating their roles in soybean adaptation to low P availability. Six *GmPTs* were mainly expressed in roots, and the others were differently expressed in stems, flowers and seeds. As a whole, these results suggest pivotal functions for *GmPTs* in both Pi acquisition and translocation. Low or no expressions of *GmPTs* were detected in pods ([Figure 4](#pone-0047726-g004){ref-type=\"fig\"}), indicating rare Pi exchange occurred or other Pi transporters function in this organ. The differential expressions of individual genes in different plant organs imply that there might be additional functions for the respective genes other than Pi uptake.\n\nInvolvement of *GmPTs* in N, K and Fe Signals {#s3d}\n---------------------------------------------\n\nPrevious reports strongly suggest that there might be crosstalks among ion signals in response to different nutrient stresses, the expressions of ion transporters might be involved in a process that facilitates mineral nutrient homeostasis [@pone.0047726-Grossman1], [@pone.0047726-Hammond1], [@pone.0047726-Kobae1], [@pone.0047726-Wang3]. Wang *et al* [@pone.0047726-Wang3] used a high-density array from tomato roots to test the expression profiling of 1,280 genes in N, K or Fe deficiency, and found the expressions of some Pi, K and Fe transporter genes were up-regulated by all three nutrient deficiencies. This suggests some transporter genes might be involved in the coordinated and coregulated uptake of these essential nutrient elements. Nitrogen, P, K, and Fe deficiencies are well known as important limiting factors to agricultural production [@pone.0047726-Kochian1], but no research has been reported on the responses of Pi transporters to N, K or Fe deficiency. Therefore, studying expressions of *GmPTs* genes under nutrient deficiencies other than P deficiency could reveal possible functions of Pi transporters in multiple mineral nutrient homeostasis.\n\nThe rice *PHO1* gene family has been well documented to play important roles in Pi translocation from roots to shoots. Furthermore, one of the *PHO1* family genes was up-regulated by N starvation [@pone.0047726-Secco1], implying crosstalks exsiting between P and N signals in rice. Recent work on the *NLA* (N limitation adaptation) gene in Arabidopsis indicates that there is an antagonistic crosstalk between N and P deficiency [@pone.0047726-Kant1]. It has been reported that *NLA* is involved in adaptive responses to low N conditions, where *nla* mutant plants display abrupt early senescence. Further analysis found the two suppressors of the *nla* mutation can impart the *nla* mutant phenotype to *NLA* wild type plants, and these suppressors have been proved to be the Pi transport-related genes, PHF1 and PHT1.1. In addition, *NLA* expression is regulated by the low-Pi induced microRNA miR827, suggesting that Pi transporters could directly affect plant N nutrition. In the present study, we found that most of the soybean *PTs* genes were significantly up-regulated under N deficiency, indicating that at least some of the Pi transporters are involved in N signaling pathways in soybean. Increased expression of a high-affinity Pi transporter gene was also found under low K in barley [@pone.0047726-Smith1], showing the existence of interactions between P and K signals. This is also supported by our results in which 8 *GmPTs* were regulated by K deficiency ([Figure 6](#pone-0047726-g006){ref-type=\"fig\"}). The underlying molecular mechanisms and pathways involved in these crosstalks need to be further researched.\n\nIron is the most studied nutrient element for interactions with P due to their strong precipitation [@pone.0047726-Dalton1]. It is well accepted that P and Fe deficiency have similar effects on the differentiation of epidermal cells, and subsequently affect root growth, including lateral root and root hair formation [@pone.0047726-LpezBucio1], [@pone.0047726-Nacry1], [@pone.0047726-Schmidt1]. In Arabidopsis, primary root growth was inhibited by P deficiency due to Fe toxicity in the root tip [@pone.0047726-Ward1]. But no P and Fe interactions related to Pi transporters has been reported. Here we found that the expressions of 5 *GmPTs* highly responded to Fe deficiency ([Figure 6](#pone-0047726-g006){ref-type=\"fig\"}), and thus provided the first evidence that Pi transporter genes might be involved in interactions between P and Fe signaling.\n\nAll together, we conclude that there are strong interactions among N, P, K and Fe signals, and Pi transporter genes might be involved in cross-talks for sensing the changes of N, P, K and Fe status in soybean.\n\nMaterials and Methods {#s4}\n=====================\n\nComputational Identification of the Pht1 Family Members of Pi Transporters in Soybean {#s4a}\n-------------------------------------------------------------------------------------\n\nThe nucleic acid sequences of all 9 and 13 members of PTs in the Pht1 family in Arabidopsis [@pone.0047726-Mudge1] and rice (*Oryza sativa*) [@pone.0047726-Paszkowski1], respectively, were used as query sequences to BLAST search the Phytozome soybean genome database (<http://www.phytozome.net/soybean>). The predicted sequences revealed there were in total 14 members of the *Pht1* gene family in soybean. The nucleic acid and amino sequences of the 14 Pi transporters were retrieved from the Phytozome website ([Text S1](#pone.0047726.s009){ref-type=\"supplementary-material\"}, [S2](#pone.0047726.s010){ref-type=\"supplementary-material\"}). Protein molecular weights and theoretical pI values were calculated using Compute pI/Mw tool (<http://www.expasy.org/tools/pi_tool.html>). Sequence identity was performed using DNAMAN version 6.0 (Lynnon Biosoft Company). Sequence alignment was performed with ClustalW [@pone.0047726-Thompson1]. The membrane-spanning domains of GmPT1 through GmPT14 were predicted by TopPred (<http://www.cbib.u-bordeaux2.fr/pise/toppred.html>). A phylogenetic tree based on entire protein sequence alignments using ClustalW was constructed by the neighbor-joining method with 1000 bootstrap replicates in the MEGA 4.1 program (<http://www.megasoftware.net/mega4/mega41.html>). Complete deletion was used to deal with gaps or missing data in sequences. The distance between sequences was estimated after Poisson correction. The distance between sequences was estimated after Poisson correction. Primers in [Table S4](#pone.0047726.s007){ref-type=\"supplementary-material\"} and [Table S5](#pone.0047726.s008){ref-type=\"supplementary-material\"} were used to amply the ORF and DNA sequences of *GmPTs*, respectively, the gained ORF and DNA sequences were used to construct gene structures through Gene Display Server (<http://gsds.cbi.pku.edu.cn/index.php>).\n\nYeast Manipulations {#s4b}\n-------------------\n\n*Saccharomyces cerevisiae* MB192 (*MATa* pho3--1 pho84::HIS3 ade2 leu2-3, 112 his3-532, trp1-289 ura3-1, 2 can1) defective in the high-affinity Pi transporter gene *PHO84* by insertion of an *HIS3* DNA fragment [@pone.0047726-BunYa1], and the expression vector, p112A1NE, were used to functional complementation assay of *GmPTs* following the protocol described previously. The full coding regions of each of the 14 *GmPTs* were produced by RT-PCR. PCR primers were designed to introduce unique restriction sites at the 5\u2032 and 3\u2032 ends of the genes ([Table S4](#pone.0047726.s007){ref-type=\"supplementary-material\"}). The cDNA amplicons were cloned into the yeast expression plasmid p112A1NE to create Yp112-GmPTs. According to Dohmen *et al* [@pone.0047726-Dohmen1], these constructs and empty vector (as control) were transformed into the MB192. Yp112-GmPTs transformants were tested for their ability to complement the growth defect of MB192. Transformed yeast strains grew in yeast nitrogen base liquid (YNB) medium to the logarithmic phase (when the absorbance at 600 nm was 0.8 at 30\u00b0C), and then were harvested and washed (centrifugation at 1500 g, 4\u00b0C, 5 min) three times with Pi-free YNB medium (containing an equivalent concentration of potassium chloride rather than potassium phosphate). The collected pellets were suspended in the same medium and incubated at 30\u00b0C until absorbance at 600 nm was 1.0, as preparation for the next experiments. For pH-dependent Pi uptake experiments, different extracellular pH values in the range of 4.0--7.0 were used. For ^33^P uptake experiments in yeast, about 1 mg fresh yeast cell samples were used following the modified method as previously described [@pone.0047726-Ai1]. Eight different concentrations of Pi (10 \u00b5M, 20 \u00b5M, 60 \u00b5M, 100 \u00b5M, 200 \u00b5M, 300 \u00b5M, 500 \u00b5M and 1000 \u00b5M) were used to assay Pi uptake by intact *Saccharomyces cerevisiae* cells by the addition of 2 \u00b5L of \\[^33^P\\]orthophosphate to 50 \u00b5L aliquots of cells, were incubated with shaking at 30\u00b0C for 5 min, Pi uptake was stopped by the addition of 1 mL cooled tris-succinate, harvested, then suspended in 25 mM Tris-succinate (pH 6) solution, and washed three times (centrifugation at 1500 g, 4\u00b0C, 5 min) with 3% glucose. Radioactivity was measured by a Beckman LS 6500 Scintillation Counter. The data were analyzed using the software SIGMAPLOT (v10.0) to determine the *Km*. For growth experiments on YNB solid medium, different numbers of transformant cells (5\u00d710^5^, 5\u00d710^4^, 5\u00d710^3^) calculated by blood cell counting plate methods were plated on pH 6.0 Pi limited (20 \u00b5M) YNB agar plates and culture at 30\u00b0C, 3 days. There were three independent biological experiments for each measurement. For growth experiments in YNB liquid medium, 100 \u00b5L cells in the logarithmic phase (OD600\u22480.90) were subjected to 3.5 mL YNB liquid medium with different Pi concentrations and incubated at 30\u00b0C, OD 600 were measured every 5 hours up to 25 hours.\n\nPlant Materials and Growth Conditions {#s4c}\n-------------------------------------\n\nSoybean cv. HN66 was employed in this study. For expression analysis of *GmPTs* in different tissues responding to P availability, soybean plants were nutrient solution cultured in a greenhouse. One week after germination, the seedlings were transplanted into the full-strength nutrient solution containing 500 \u00b5M KH~2~PO~4~, 3000 \u00b5M KNO~3~, 2000 \u00b5M Ca(NO~3~)~2~, 250 \u00b5M MgSO~4~, 25 \u00b5M MgCl~2~, 12.5 \u00b5M H~3~BO~3~, 1 \u00b5M MnSO~4~, 1 \u00b5M ZnSO~4~, 0.25 \u00b5M CuSO~4~, 0.25 \u00b5M (NH~4~)~6~Mo~7~O~24~ and 25 \u00b5M Fe-Na-EDTA. The seedlings were grown for 10 days till the first trifoliate leaves fully expanded and then treated with two P supplies (5 \u00b5M and 500 \u00b5M P). At 18 d after treatment, roots, stems, leaves and flowers were separately harvested. At 29 d after treatment, young pods and seeds were separately harvested. All tissue samples were stored at \u221280\u00b0C for RNA extraction.\n\nTo elucidate the responses of the *GmPTs* to the other nutrient deficiencies, ten-day old seedlings precultured in full-strength nutrient solution as described above were treated under \u2212N, \u2212P \u2212K, and -Fe conditions for 14 days, respectively. For \u2212N treatment, KNO~3~, Ca(NO~3~)~2~ and (NH~4~)~6~Mo~7~O~24~ was replaced by 1500 \u00b5M K~2~SO~4~, 2000 \u00b5M CaSO~4~ and 0.25 \u00b5M Na~2~MoO~4~, respectively. For -K treatment, KNO~3~ and KH~2~PO~4~ was replaced by 1500 \u00b5M Ca(NO~3~)~2~ and 500 \u00b5M NaH~2~PO~4~, respectively. For -Fe treatment, Fe-EDTA was fully withdrawn. Plants continuously grown in the full-strength nutrient solution were sampled as control (CK). Each treatment had three biological replicates. Leaves and roots were separately collected for total RNA extraction and qRT-PCR analysis.\n\nQuantitative Real-time RT-PCR Analysis {#s4d}\n--------------------------------------\n\nFor qRT-PCR analysis, the soybean housekeeping gene *TefS1* (encodes the elongation factor EF-1a: X56856) was used as a reference gene. The optimal primer sequences for *GmPTs* and *TefS1* were listed in [Table S2](#pone.0047726.s005){ref-type=\"supplementary-material\"}. Total RNA was extracted from soybean plants using RNAiso\u2122 Plus reagent (TaKaRa) according to the manufacturer's instructions. RNA samples were treated with RNase-free DNaseI (Invitrogen) to remove the contaminating genomic DNA before synthesizing the first strand cDNA using the MMLV-reverse transcriptase (Promega, USA) according to the protocol from the supplier. qRT-PCR was carried out in a 20 \u00b5L volume containing 2 \u00b5L 1\u223650 diluted reverse transcription product, 0.2 \u00b5M primers, and 10 \u00b5L SYBR\u00ae Premix EX Taq\u2122 (TaKaRa). All the reactions were done on a DNA Engine Opticon 2 Continuous Fluorescence Detection System (MJ Research Inc., Waltham, MA). Reaction conditions for thermal cycling were: 95\u00b0C for 1 min, 40 cycles of 95\u00b0C for 15 s, 58--60\u00b0C for 15 s, and 72\u00b0C for 30 s. The annealing temperature (58--60\u00b0C) was adjusted to suit the amplification of individual *GmPTs*. Fluorescence data were collected during the cycle at 72\u00b0C.\n\nData Analysis {#s4e}\n-------------\n\nAll the data were analyzed statistically using Microsoft Excel 2003 (Microsoft Company, USA) for calculating mean and standard error. Comparisons of gene expressions in different tissues or responses to different nutrient deficiencies were performed using the student *t*-test in the Microsoft Excel 2003.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Alignment of amino acid sequences of the Pht1 family phosphate transporters in soybean.** Sequence alignment was performed with the ClustalW program [@pone.0047726-Thompson1]. Identical and similar amino acids are shaded in black and grey, respectively. The membrane spanning domains of GmPTs predicted by TopPred (<http://www.cbib.u-bordeaux2.fr/pise/toppred.html>) are under lined and numbered by roman numerals (I--XII).\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Schematic diagram of intron/exon structure of** ***GmPT*** **genes.** The thin line represents the introns and the open boxes indicate the exons of the respective genes.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Kinetic growth profiles of yeast transformants.** Yeast strains, including WT, empty vector control or Yp112-GmPTs transformants, grew in the logarithmic phase (OD600\u22480.90), then 100 \u00b5L different yeast cells were subjected to 3.5 mL YNB liquid medium with different Pi concentrations and incubated at 30\u00b0C, OD 600 were measured every 5 hours up to 25 hours. The Pi concentrations were selected according to their *Km* values.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Percentage of protein sequences identity among the 14 soybean phosphate transporters in Pht1 family.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Genes and gene-specific primers used for quantitative real-time PCR experiments.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Expression pattern of soybean GmPTs in the common used tissues in reference 37, 38. The numbers presented in the table are normalized Illumina-Solexa reads number coming from the according experiment.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Primers used to generate the expression vectors in yeast complementary assays (restriction site sequences are underlined).**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Primers used to generate DNA sequences of GmPTs.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Protein sequences of GmPTs.**\n\n(TXT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Nucleic acid sequences of** ***GmPTs.***\n\n(TXT)\n\n###### \n\nClick here for additional data file.\n\nWe thank Ms. Xinxin Li and Mr. Chengchen Li for sampling assistance. Drs. Xiurong Wang, Jiang Tian and Jinxiang Wang for helpful discussions and comments.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: LQ YXG JZ HL. Performed the experiments: LQ YXG LYC RKL MG. Analyzed the data: LQ YXG JZ TW. Wrote the paper: LQ YXG GX JZ HL. Revised the manuscript: JZ GX TW HL.\n"}
+{"text": "Case report {#ajum00223-sec-0001}\n===========\n\nA 31\u2010year old female, G2P0, presented to the fetal medicine unit at 38 weeks gestation with a history of reduced fetal movements. Ultrasound confirmed fetal death. The amniotic fluid volume was norma; biometry was suggestive of undetected fetal growth restriction, and no fetal anomalies were identified. Targeted placental ultrasonography revealed numerous hypo\u2010echogenic lesions scattered throughout most of the substance of the placenta ([Figures 1](#ajum00223-fig-0001){ref-type=\"fig\"} & [2](#ajum00223-fig-0002){ref-type=\"fig\"}). The standard protocol for investigating a perinatal death[^1^](#ajum00223-bib-0001){ref-type=\"ref\"} was followed. Amniotic fluid was collected and sent for microarray analysis, which ruled out an underlying chromosomal abnormality. The parents consented to a full post mortem and placental examination.\n\n![Numerous hypeoechic lesions in the placenta.](AJUM-18-157-g001){#ajum00223-fig-0001}\n\n![Numerous hypeoechic lesions in the placenta.](AJUM-18-157-g002){#ajum00223-fig-0002}\n\nThe post mortem examination revealed a growth restricted but otherwise normally appearing female infant. There was evidence of chronic hypoxia and congestive cardiac failure. No developmental malformations were identified.\n\nThe placental disc was noted to be smaller than expected for the gestational age, but otherwise, externally, appeared essentially normal. Multiple, firm, cream to dark\u2010red well\u2010circumscribed lesions, measuring between 3 to 35 mm in diameter, were identified when the placenta was sliced ([Figures 3](#ajum00223-fig-0003){ref-type=\"fig\"} & [4](#ajum00223-fig-0004){ref-type=\"fig\"}). These lesions involved 60--70% of the entire placental disc volume, and were identified as multiple chorangiomas on histopathology ([Figures 5](#ajum00223-fig-0005){ref-type=\"fig\"} & [6](#ajum00223-fig-0006){ref-type=\"fig\"}). Calcification and infarction was evident in some of the tumours. Additionally, multiple foci of chronic lymphohistiocytic villitis, affecting terminal and basal villi, were identified. In the absence of a cultured organism, this was classified as chronic villitis of unknown etiology (VUE).\n\n![Cut surface of the placenta, demonstrating multiple pale to dark red chorangiomas.](AJUM-18-157-g003){#ajum00223-fig-0003}\n\n![Cut surface of the placenta, demonstrating multiple pale to dark red chorangiomas (indicated by black arrow).](AJUM-18-157-g004){#ajum00223-fig-0004}\n\n![Low and Medium power magnification show well circumscribed vascular lesions (chorangiomas) surrounded by placental tissue.](AJUM-18-157-g005){#ajum00223-fig-0005}\n\n![Low and Medium power magnification show well circumscribed vascular lesions (chorangiomas) surrounded by placental tissue.](AJUM-18-157-g006){#ajum00223-fig-0006}\n\nDiscussion {#ajum00223-sec-0002}\n==========\n\nA stillbirth is a heartbreaking outcome for any family but even more devastating is our failure to properly investigate such a death. Professor Ruth Fretts from Harvard Medical School called it the 'Stillbirth Scandal' in a recent RANZCOG publication.[^2^](#ajum00223-bib-0002){ref-type=\"ref\"} Many stillbirths that are thought to be 'unexplained' are actually incompletely investigated. When the unthinkable happens, we should be striving to find answers. It is tempting to think all answers will come from high end testing, but sometimes the answers lie in front of our eyes, if we would only take the time to have a careful look. During the formal scan to confirm a fetal death, we advocate that a detailed study should be mandatory. Circumstances permitting, one should aim to specifically, document the fetal presentation and placental site, search for fetal and placental anomalies, perform biometry and measure the amniotic fluid volume. Following a stillbirth, many parents often opt for limited investigations, and in these cases, the findings from this ultrasound scan, together with the external examination of the baby performed following delivery, will form the basis of perinatal loss counselling. In our case, targeted placental ultrasonography, together with corroboratory information obtained from the post mortem and placental histopathology examinations identified the cause of death. Had these parents not agreed to investigations, in the absence of a detailed ultrasound study, this stillbirth too had the potential of being classified as yet another unexplained loss.\n\nChorangiomas are analogous to haemangiomas in other parts of the body, and can be observed in 0.5% of placentas.[^3^](#ajum00223-bib-0003){ref-type=\"ref\"} They develop after the first trimester, and whilst small solitary lesions are often inconsequential, the larger and multiple lesions are of concern, and necessitate closer pregnancy surveillance by a specialist. These tumours essentially function as arteriovenous shunts, causing a prolonged and persistent disruption of fetal cardiac function. Consequently, affected fetuses exhibit a varied tendency towards high output cardiac failure, which is not only associated with the development of hydrops and an increased risk of subsequent fetal/neonatal demise, but, may also predispose to abnormal neurodevelopment in childhood. Fetal growth restriction, anaemia and hypoxia are thought to occur because of preferential shunting of blood through the low\u2010pressure system within the chorioangioma thereby depriving functioning villi of adequate blood flow and nutrient exchange.[^4^](#ajum00223-bib-0004){ref-type=\"ref\"} The development of chorioangiomas is associated with high altitude pregnancies, multiple pregnancies and other conditions associated with a low oxygen tension, suggesting the influence of vascular growth factors induced by hypoxia. This case highlights the importance of submitting the placenta for pathological evaluation as the findings clearly have management implications for subsequent pregnancies.\n\nIn this case two major placental insults were identified: multiple chorioangiomas and chronic lymphohistiocytic villitis. Individually both of these conditions can cause a compromised, under\u2010perfused, feto\u2010placental circulation that can result in a small placenta, growth\u2010restricted baby and ultimately fetal demise. It should be noted that whilst the recurrence risk of chorioangiomas in a future pregnancy is low, VUE might reoccur in up to 17% of cases.[^5^](#ajum00223-bib-0005){ref-type=\"ref\"}\n\nA final learning point that can be drawn from this case is the reminder that decreased fetal movement is associated with a less\u2010than\u2010optimal outcome in approximately 26% of pregnancies, with unsuspected growth restriction being the most common finding at evaluation.[^6^](#ajum00223-bib-0006){ref-type=\"ref\"} Women presenting with decreased movements, who are scanned, need a comprehensive ultrasound examination, which should include placental sonography. In doing so, we have the potential to avoid a stillbirth by not missing the important message that both mother and baby are trying to make.\n"}
+{"text": "Tables of Links\n===============\n\n  TARGETS\n  -------------------------\n  **Ion channels**\n  hERG, (Kv11.1) channels\n\n  LIGANDS       \n  ------------- ----------------\n  Cisapride     Risperidone\n  Dofetilide    \\(DL\\) Sotalol\n  E-4031        Terfenadine\n  Haloperidol   Thioridazine\n\nThese Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in <http://www.guidetopharmacology.org>, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson *et\u2009al*., [@b29]) and are permanently archived in the Concise Guide to PHARMACOLOGY 2013/14 (Alexander *et\u2009al*., [@b3]).\n\nIntroduction\n============\n\nCompounds exhibiting blocking properties at the voltage-gated K^+^ channel, hERG (K~v~11.1), are associated with a higher risk for the life-threatening cardiac arrhythmia called Torsades de pointes (TdP). Their arrhythmic properties are attributed to their ability to cause a lengthening of repolarization of ventricular cells resulting from an inhibition of the I~Kr~ delayed rectifier repolarizing current. However, a large majority of drugs showing both this electrophysiological property and QT prolongation ability have never caused TdP in human (De Ponti *et\u2009al*., [@b11]; Redfern *et\u2009al*., [@b31]). In spite of a great deal of research over the last decade, it remains difficult to discriminate with assurance whether compunds genuinely have a ventricular arrhythmic profile from *in vitro* models (Abi-Gerges *et\u2009al*., [@b2]; Nalos *et\u2009al*., [@b25]; Di Veroli *et\u2009al*., [@b12]; Yang *et\u2009al*., [@b39]). Alternatively, use of *in vivo* models demonstrated that the short-term variability (STV) of ventricular repolarization biomarkers (left ventricular monophasic action potential, QT interval) could constitute a stronger surrogate of TdP liability than QT prolongation. However, this surrogate seems to require conditions of decreased repolarization reserve, such as chronic atrio-ventricular block (Thomsen *et\u2009al*., [@b34]) or blockade of I~Ks~ channels (Lengyel *et\u2009al*., [@b22]) that are not reached in regulatory studies (ICH S7B) conducted in healthy animals. Consequently, few arrhythmic hERG blockers have been shown to increase beat-to-beat repolarization STV in healthy animals.\n\nDespite this, we have used this parameter in the present study as a reference biomarker to explore whether the Coumel\\'s triangle concept (Coumel, [@b8]) was applicable to arrhythmic hERG blockers. According to this concept, the generation of arrhythmias requires three conditions or factors. The first one is a substrate. In the case of arrhythmic hERG blockers, it is widely accepted that ventricular repolarization constitutes a substrate favourable to TdP (Antzelevitch, [@b4]). The second one is a trigger factor and the third one is a modulator factor. The autonomic nervous system (ANS) has been demonstrated as representing this modulator factor in various kinds of arrhythmias (Podrid *et\u2009al*., [@b30]; Coumel, [@b9]; Leenhardt *et\u2009al*., [@b21]). In the case of hERG blockers, we demonstrated that thioridazine-induced QT prolongation was modulated by the sympathetic nervous system (Champeroux *et\u2009al*., [@b6]) as sympatho-activation is able to fully reverse its QT prolongation properties. Interestingly, expression of gens of the ether-a-go-go (EAG) family has been reported not only in the heart, but also in brain neurons (Vandenberg *et\u2009al*., [@b36]). In humans, expression of KCNH2-1a, a gene coding for a hERG channel subunit present in cardiac tissues, has been reported also in the hippocampus and cerebral cortex. Its level of expression was similar to that of a brain-specific isoform named KCNH2-3.1 (Huffaker *et\u2009al*., [@b19]). Both isoforms generate inward rectifying K^+^ currents blocked by E-4031, a selective hERG blocker. Both isoforms play a role in neuronal excitability and firing rate through hyperpolarizing properties. Conversely, E-4031 was reported to cause a depolarizing effect and increase neuronal firing in rat cortical neurons transfected with the KCNH2-3.1 isoform suggesting that hERG blockers might increase neurotransmitter release. Interestingly, mutations characterized by a loss of function of hERG channels have been linked to both the long QT syndrome type 2 (LQTS2) and a higher incidence of seizures in patients (Omichi *et\u2009al*., [@b26]; Zamorano-Le\u00f3n *et\u2009al*., [@b40]). Taken together, these data suggest that hERG-blocking drugs could induce both cardiac (delayed repolarization) and potentially neuronal effects through depolarizing properties.\n\nIf such interactions with neuronal activities exist and contribute to the cardiac adverse effect of arrhythmic hERG blockers, they would be expected to be mediated by the ANS. However, little is known about the presence of neuronal ERG channels in the two limbs of the ANS or in brain areas involved in the autonomic control except that ERG channels are expressed in rat peripheral sympathetic ganglia (Shi *et\u2009al*., [@b33]). The present study was designed to search for functional evidence supporting the hypothesis of interactions of hERG blockers with cardiac autonomic control. Experiments were conducted in healthy beagle dogs and cynomolgus monkeys, the two main species used in preclinical regulatory QT studies required for phase I clinical trials. In the present work, we report a detailed analysis of dofetilide effects in cynomolgus monkeys, results in dogs being reported as Supporting Information. Indeed, this hERG blocker offers in this species good conditions to study any link between possible interactions with the ANS and ventricular arrhythmias as this agent is able to induce TdP with a high incidence in cynomolgus monkeys.\n\nMethods\n=======\n\nAll animal care and experimental procedures complied with the 2010/63/UE animal welfare European directive and were approved by ethical review (ethical committee n\u00b0 CEEA-111). All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals (Kilkenny *et\u2009al*., [@b20]; McGrath *et\u2009al*., [@b24]). A total of 24 beagle dogs and 12 cynomolgus monkeys were used in the experiments described here.\n\nECG recordings in conscious animals\n-----------------------------------\n\nAdult beagle dogs (10--15\u2009kg; CEDS, Mezilles, France) and cynomolgus monkeys (3--7\u2009kg; originated from Mauritius and purchased from Bioprim, Bazi\u00e8ge, France) were instrumented with radio telemetry transmitters (Data Sciences International, Saint Paul, MN, USA) as described elsewhere (Champeroux *et\u2009al*., [@b7]). After left thoracotomy, one electrode was sutured directly to the left ventricular epicardium near the apex while the second electrode was sutured to the pericardium above the right atrium to approximate a limb Lead II ECG. Analgesic treatment with buprenorphine/meloxicam was given before surgery and continued for a minimum of 2 days to alleviate any post-operative pain. A minimum period of 3 weeks was allowed for recovery from the surgery. Animals were housed in individual stainless steel cages for telemetry recordings. Out of recording periods, animals were housed in pens by groups of six animals at maximum. Environmental parameters were recorded continuously and maintained within a fixed range, room temperature at 15--21\u00b0C and 45--65% relative humidity. The artificial day/night cycle was 12\u2009h light and 12\u2009h darkness with light on at 07:30\u2009h. Drinking water was provided *ad libitum*. Solid diet (300\u2009g) was given daily in the morning. All dosing with drugs was performed between 15:00 and 15:30\u2009h. ECGs were recorded continuously for a minimum of 2\u2009h before dosing up to 24\u2009h post dose. ECG waveforms were continuously recorded at a sampling rate of 500\u2009Hz using the ART\u2122 acquisition software release 4.2 (Data Sciences International). Cardiac conduction times including QT interval were calculated from a beat-to-beat analysis using internal software developed in RPL (RS/1 programming language, RS/1 release 6.3, Applied Materials, Cambridge MA, USA), Good Laboratory Practices (GLP) validated. Validation of correct location of cardiac wave markers was performed according to a standardized procedure, which covered the whole 24\u2009h period. In most cases, the percentage of errors in location of the end of the T wave was less than 3%. Experiments were conducted following a randomized cross-over study design (vehicle vs. test items) in groups of six animals (three males and three females).\n\nSTV\n---\n\nThe beat-to-beat STV of QT interval, STV~QT~, was calculated from 30 successive beats as described elsewhere (Thomsen *et\u2009al*., [@b34]) using the following formula: . Beat-to-beat STV of heart rate (HR), STV~HR~, was calculated according to the same principle.\n\nHR and QT interval high-frequency (HF) rhythms analysis\n-------------------------------------------------------\n\nHR and QT interval values were re-sampled at a 1\u2009Hz frequency. Analysis of rhythms was derived from a spectral analysis of the HR signal. This consisted of applying a fast Fourier transform on successive and continuous 256\u2009s sequences of the 1\u2009Hz re-sampled HR signal. HF rhythm power in bpm^2^ or ms^2^ was derived from integration of power spectrum within the 0.1--0.5\u2009Hz frequency band (Pagani *et\u2009al*., [@b28]). HF rhythm power was calculated as 1\u2009h reduced mean values (Champeroux *et\u2009al*., [@b7]).\n\nHF oscillation amplitude analysis\n---------------------------------\n\nMaximum and minimum beat-to-beat HR and QT interval values were derived from analysis of successive 10\u2009s ECG sequences, that is the maximum period of HF rhythms (0.1\u2009Hz). Amplitude of HF oscillations was calculated from the difference between maximum and minimum HR or QT interval values noted within each 10\u2009s sequence (see Supporting Information Fig. [S1](#sd1){ref-type=\"supplementary-material\"}, for a typical example of the determination of HF oscillation amplitude).\n\nArrhythmia analysis\n-------------------\n\nPresence of arrhythmias was investigated throughout the whole post-dosing 24\u2009h period using an internal software development in RPL (RS/1 programming language, RS/1 release 6.3, Applied Materials), GLP validated. Qualification of arrhythmias was performed manually following the Lambeth conventions (Curtis *et\u2009al*., [@b10]).\n\nAutonomic blockade\n------------------\n\nComplete suppression of autonomic control was achieved with hexamethonium (20\u2009mg\u00b7kg^\u22121^, i.v.).\n\nData analysis\n-------------\n\nData were processed and analysed using GLP-validated RS/1 computer procedures (release 6.3, Applied Materials). Individual results were calculated as 1\u2009h reduced mean values (mean calculated from 50% of discrete values centred on the median) derived from these discrete values for all parameters (QT interval, HR, STV, HF rhythms). Mean results derived from these individual values were reported as mean \u00b1 SEM. 256\u2009s samples with presence of arrhythmias were excluded from mean calculations and statistical comparisons for all parameters. Drug-induced effects on QT interval, HR, STV~QT~, STV~HR~ and HF rhythms were compared with their respective vehicle using an [anova]{.smallcaps} for repeated measurements. In case of a significant 'treatment effect' (*P* \u2264 0.05) or significant 'treatment \u00d7 time' interaction (*P* \u2264 0.05), the analysis was continued by Fisher\\'s least significant difference test at each time point. Effects on arrhythmia incidences were compared between groups using a non-parametric Wilcoxon test for paired data.\n\nMaterials\n---------\n\nDofetilide was purchased from Sequoia Research Product, Ltd (Pangbourne, UK). Hexamethonium was purchased from Sigma-Aldrich (Saint Quentin, France). These compounds were dissolved in saline (vehicle).\n\nResults\n=======\n\nWe performed most of the analyses over very short sample periods of 10\u2009s. This duration corresponds to a period of 0.1\u2009Hz for HF autonomic rhythms of HR. In cynomolgus monkeys, individual inspection of ECG traces show that beat-to-beat HR and QT interval variability is very moderate when analysed over very short 10\u2009s periods (Figure\u2009[1](#fig01){ref-type=\"fig\"}A). QT interval variability was markedly increased after dosing with dofetilide (Figure\u2009[1](#fig01){ref-type=\"fig\"}B). The Poincar\u00e9 plot representation (Figure\u2009[1](#fig01){ref-type=\"fig\"}C) drawn from these two 10\u2009s sequences confirms the dofetilide-induced increase in QT interval STV described earlier with this agent in other species and models (Thomsen *et\u2009al*., [@b34]). Moreover, an increase in beat-to-beat HR variability is also clearly visible in parallel to QT interval variability. In this typical example (Figure\u2009[1](#fig01){ref-type=\"fig\"}B), QT interval variability seems dependent on fast rhythmic oscillations of the HR through the well-known inverse rate dependence between beat-to-beat HR and QT interval (Holzgrefe *et\u2009al*., [@b17]). Moreover, changes in QT interval during oscillations may influence subsequent beat-to-beat HR values and elicit correcting HR adaptation in dofetilide-treated (1\u2009mg\u00b7kg^\u22121^, p.o.) animals (Supporting Information Fig. [S2](#sd1){ref-type=\"supplementary-material\"}).\n\n![Typical beat-to-beat HR and QT interval variability over a 10\u2009s period recorded in the same cynomolgus monkey 1\u2009h before (A) and 40\u2009min after dosing with dofetilide (B, 1\u2009mg\u00b7kg^\u22121^, p.o.). (C) The corresponding Poincar\u00e9 plot (QT~n+1~ vs. previous QT~n~) showing an increase in beat-to-beat QT interval variability after dosing with dofetilide.](bph0172-2878-f1){#fig01}\n\nDofetilide-induced STV~QT~ is closely associated with QT interval HF rhythmic oscillations\n------------------------------------------------------------------------------------------\n\nDofetilide induced, as expected, a large and sustained increase in QT interval (Figure\u2009[2](#fig02){ref-type=\"fig\"}A). Note that the large increase in QT interval seen from 5\u2009h after dosing is related to the nycthemeral cycle of HR, light being off 4.5\u2009h after dosing. This agent caused in parallel a large increase in STV~QT~ (Figure\u2009[2](#fig02){ref-type=\"fig\"}B). Interestingly, the STV~QT~ increase appears not to be an exclusive consequence of QT prolongation as STV~QT~ returned to baseline 12\u2009h after dosing while QT interval value remained close to its highest level at the same time. Moreover, STV~QT~ changes seen up to 12\u2009h after dosing with dofetilide were more closely correlated to changes in HF QT interval rhythms power and HF QT interval oscillations amplitude than with QT or QTc interval changes (Supporting Information Fig. [S3](#sd1){ref-type=\"supplementary-material\"}). The rhythmic feature of dofetilide-induced STV~QT~ was first confirmed from power spectral analysis of beat-to-beat QT changes. Indeed, this analysis reveals the rhythmic aspect of QT interval variability (Figure\u2009[2](#fig02){ref-type=\"fig\"}C). These oscillations correspond to HF rhythms (\\>0.1\u2009Hz). The time course of QT interval HF oscillations fits well with that of STV~QT~. These results provide a first argument suggesting that dofetilide-induced STV~QT~ might have a rhythmic feature. Amplitude of QT interval HF oscillations was close to 40\u2009ms at the peak of effect (Figure\u2009[2](#fig02){ref-type=\"fig\"}D) instead of 10\u2009ms in vehicle-treated animals.\n\n![Effect of dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.) on QT interval and short-term QT interval variability in cynomolgus monkeys. (A) Dofetilide-induced QT prolongation. (B) Dofetilide also caused a marked increase in STVQT. (C) Power spectral analysis reveals QT interval HF (\\>0.1\u2009Hz) rhythmic oscillations in dofetilide treated animals. (D) amplitude of QT interval HF oscillations. The time course of changes in QT interval HF rhythmic oscillations fits with those of STVQT. Data are presented as mean values \u00b1 SEM; *n* = 6. \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, significantly different from vehicle.](bph0172-2878-f2){#fig02}\n\nDofetilide increased the HF oscillations in HR\n----------------------------------------------\n\nWhile dofetilide had almost no effect on mean HR in cynomolgus monkeys (Figure\u2009[3](#fig03){ref-type=\"fig\"}A), it caused a marked increase in beat-to-beat HR STV (STV~HR~, Figure\u2009[3](#fig03){ref-type=\"fig\"}B). As for STV~QT~, power spectral analysis of HR variability reveals that dofetilide-induced STV~HR~ has a rhythmic feature within the same HF band (Figure\u2009[3](#fig03){ref-type=\"fig\"}C). Indeed, amplitude of the HF oscillations in HR was markedly increased, close to 70\u2009beats per min at the peak of effect (Figure\u2009[3](#fig03){ref-type=\"fig\"}D), when compared with vehicle-treated animals. The time course of HF oscillations in HR fits with that of QT interval HF oscillations and STV~QT~. As HF oscillations in HR are classically attributed to rhythmic activity of the parasympathetic nervous system (Pagani *et\u2009al*., [@b28]), these results show that dofetilide enhanced endogenous parasympathetic rhythms of HR.\n\n![Effects of dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.) on HR and short-term HR variability in cynomolgus monkeys. (A) Dofetilide was quite devoid of effect on mean HR. (B) In parallel, dofetilide increased beat-to-beat STV~HR~. (C) Power spectral analysis of HR variability reveals HF (\\>0.1\u2009Hz) rhythmic oscillations of HR in dofetilide-treated animals. (D) Amplitude of HR HF oscillations. Data are presented as mean values \u00b1 SEM; *n* = 6. \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, significantly different from vehicle.](bph0172-2878-f3){#fig03}\n\nAutonomic blockade by hexamethonium prevents dofetilide-induced STV~QT~\n-----------------------------------------------------------------------\n\nWe used hexamethonium, a ganglioplegic agent, to investigate the involvement of the ANS in dofetilide effects on STV~QT~ and HF rhythms. This agent blocks both the parasympathetic and sympathetic limbs of the ANS. Hexamethonium alone fully suppressed the endogenous HF parasympathetic rhythms of HR (Figure\u2009[4](#fig04){ref-type=\"fig\"}A). Duration of its effect was 6\u2009h in cynomolgus monkeys. Hexamethonium had moderate effect on mean HR levels and QT interval (Supporting Information Fig. [S4](#sd1){ref-type=\"supplementary-material\"}). Comparisons of the effects of dofetilide in the absence and the presence of hexamethonium show that increases in endogenous parasympathetic rhythms of HR (Figure\u2009[4](#fig04){ref-type=\"fig\"}B), in QT interval HF oscillations (Figure\u2009[4](#fig04){ref-type=\"fig\"}C) and STV~QT~ (Figure\u2009[4](#fig04){ref-type=\"fig\"}D) were fully suppressed for the whole duration of autonomic blockade. These results confirm that the effects of dofetilide on HR HF rhythms, QT interval HF oscillations and STV~QT~ have an autonomic component. However, comparison of dofetilide in the presence of hexamethonium versus hexamethonium alone revealed a residual source of QT interval variability (STV~QT~) that was resistant to autonomic blockade (Figure\u2009[4](#fig04){ref-type=\"fig\"}E). Magnitude of this residual source of QT interval variability caused by dofetilide is very much lower than that seen under intact autonomic control. This QT interval variability cannot be explained by a residual source of HR variability as STV~HR~ was dramatically reduced following autonomic blockade both in the presence or absence of dofetilide (Figure\u2009[4](#fig04){ref-type=\"fig\"}F). The experiments with dofetilide combined with hexamethonium in beagle dogs lead to the same findings and conclusions (Supporting Information Fig. [S5](#sd1){ref-type=\"supplementary-material\"}).\n\n![Effects of autonomic blockade by hexamethonium (20\u2009mg\u00b7kg^\u22121^, i.v., HEX) on dofetilide-induced HR and short-term QT interval variability in cynomolgus monkeys. (A) Hexamethonium alone fully abolished endogenous HF oscillations in HR for 6\u2009h, compared with vehicle (VEH). During this period, this ganglioplegic agent markedly reduced dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o., DOF) induced increases in HF oscillations in HR (B), QT interval HF oscillations (C) and STVQT (D). (E) A residual QT interval variability (STVQT) insensitive to autonomic blockade was still visible in the presence of dofetilide plus hexamethonium versus hexamethonium alone. (F) On the other hand, dofetilide did not cause any further beat-to-beat STV~HR~ under autonomic blockade when compared with hexamethonium alone. Data are presented as mean values \u00b1 SEM; *n* = 6. *P* \u2264 0.05, \\*\\**P* \u2264 0.01, significantly different from vehicle or hexamethonium alone.](bph0172-2878-f4){#fig04}\n\nDofetilide enhances QT interval HF oscillations through HR HF oscillation amplitude-dependent and -independent mechanisms\n-------------------------------------------------------------------------------------------------------------------------\n\nWe built the relationship between the amplitudes of the QT interval oscillation and HF oscillations in HR (HF QT/HF HR relationship) using the same methodological approach as described earlier for the QT/RR relationship (Holzgrefe *et\u2009al*., [@b16]). Briefly, this analysis consists of sorting pairs of QT interval and HR HF oscillation amplitudes by increasing increments of 5\u2009bpm. Then, mean QT interval HF oscillation amplitude was calculated for each 5\u2009bpm increment. In vehicle-treated animals, this relationship revealed that the amplitude of the QT interval HF oscillation progressively increased with the amplitude of HF oscillations in HR (Figure\u2009[5](#fig05){ref-type=\"fig\"}A). Dofetilide had two distinct effects on this relationship (Figure\u2009[5](#fig05){ref-type=\"fig\"}B). First, dofetilide caused a shift of the whole relationship towards higher amplitude of HF oscillations in HR (horizontal shift). By prolonging this relationship towards greater HF oscillations in HR, dofetilide increases amplitude of QT interval HF oscillations. Furthermore, it enhanced amplitude of QT interval oscillations according to a mechanism independent of the amplitude of HF oscillations in HR (vertical shift) as it increased amplitude of QT interval HF oscillations for fixed values of amplitude of HF oscillations in HR. Similar findings were seen in beagle dogs (Supporting Information Fig. [S6](#sd1){ref-type=\"supplementary-material\"}).\n\n![(A) Relationship between the amplitudes of the QT interval oscillation and the HF oscillations in HR (HF QT/HF HR). (A) QT interval HF oscillation amplitude progressively increased with the amplitude of HF oscillations in HR in vehicle-treated animals. (B) Dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.,) caused a horizontal shift (1) of the whole relationship towards higher amplitude of HF oscillations in HR. It also enhanced amplitude of QT interval oscillations according to a mechanism independent from amplitude of HF oscillations in HR (2: vertical shift). The relationship was built from all data collected between 1 and 10\u2009h after dosing, that is the period during which dofetilide caused a marked QT prolongation and increase in HF oscillations in HR. Data are presented as mean values \u00b1 SEM; *n* = 6. \\*\\**P* \u2264 0.01, significantly different from vehicle.](bph0172-2878-f5){#fig05}\n\nHF oscillations contribute to arrhythmic properties of dofetilide in cynomolgus monkeys\n---------------------------------------------------------------------------------------\n\nDofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.) induces ventricular arrhythmias in all cynomolgus monkeys offering good conditions for evaluation of the role of HF oscillations on arrhythmic properties of this molecule. In dogs, dofetilide is also able to trigger ventricular arrhythmias, but not in all animals, making it more difficult to obtain valid statistical comparisons from small-sized groups. During the whole period of suppressed HF oscillations, hexamethonium prevented ventricular premature beats following dofetilide dosing in five out of six animals (Figure\u2009[6](#fig06){ref-type=\"fig\"}A), and in all animals for more severe ventricular arrhythmias (doublets, triplets, ventricular tachycardia episodes, Figure\u2009[6](#fig06){ref-type=\"fig\"}B). Moreover, four out of six animals exhibited TdP episodes with dofetilide, whereas no TdP was observed following dosing with dofetilide in the presence of hexamethonium. It should be pointed out that hexamethonium prevented dofetilide-induced arrhythmic events in a way independent of an elevation of HR levels. Conversely, hexamethonium slightly decreased HR in the presence of dofetilide during the first 4\u2009h when compared with dofetilide alone, that is when the incidence of arrhythmias is already high with dofetilide (Supporting Information Fig. [S7](#sd1){ref-type=\"supplementary-material\"}). We also noted that time course of STV~QT~ changes were correlated with arrhythmia incidence in terms of start of occurrence and duration (Supporting Information Fig. [S8](#sd1){ref-type=\"supplementary-material\"}). In the absence of autonomic blockade, dofetilide-induced TdP episodes were initiated by a premature ventricular beat that occurred just after a deceleration phase within HF oscillations in HR and just before the onset of the TdP episodes (Figure\u2009[6](#fig06){ref-type=\"fig\"}C). Taken together, these results highlight the role of HF autonomic oscillations and their importance for the mechanisms for triggering ventricular arrhythmias.\n\n![Hexamethonium (HEX: 20\u2009mg\u00b7kg^\u22121^, i.v.) prevented dofetilide-induced (DOF: 1\u2009mg\u00b7kg^\u22121^, p.o.) ventricular arrhythmias during the whole period of suppression of HF oscillations in cynomolgus monkeys. Data are reported as individual total number of ventricular arrhythmic events counted in each animal (*n* = 6), N, number of arrhythmic event per animal. (A) Premature ventricular beats (PVB). (B) Doublets + triplets + ventricular tachycardia (VT) + TdP episodes. \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, significantly different from dofetilide alone. (C) Example of HF oscillations in HR characterized by fast successive deceleration/acceleration phases preceding a PVB occurring just after a deceleration phase and preceding the onset of a TdP episode in a cynomolgus monkey (the same animal as Figure\u2009[1](#fig01){ref-type=\"fig\"}) dosed with dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.).](bph0172-2878-f6){#fig06}\n\nHR HF rhythm enhancement by several arrhythmic hERG blockers in beagle dogs\n---------------------------------------------------------------------------\n\nWe retrospectively analysed data collected in our laboratory in beagle dogs with 13 other hERG channel blockers. This analysis shows that enhancements of endogenous HF oscillations in HR were also found in beagle dogs with dofetilide and with other hERG blockers such as cisapride, haloperidol, risperidone, DL sotalol, terfenadine and thioridazine (Figure\u2009[7](#fig07){ref-type=\"fig\"}A). On the other hand, the other hERG blockers had either no effect or reduced endogenous HF oscillations in HR in beagle dogs (Figure\u2009[7](#fig07){ref-type=\"fig\"}B). In parallel, few compounds were able to cause statistically significant increases in STV~QT~ in the group of hERG blockers causing enhancement of HF oscillations in HR (dofetilide and terfenadine only, Figure\u2009[7](#fig07){ref-type=\"fig\"}C) while STV~QT~ was rather decreased among the second group of hERG blockers, in particular with drugs causing the largest decrease in HF oscillations in HR (Figure\u2009[7](#fig07){ref-type=\"fig\"}D).\n\n![Maximum enhancement in the amplitude of HF oscillations in HR (HF HR) or changes in STVQT in two groups of hERG blockers tested at various dose levels in beagle dogs. (A) All hERG blockers of group I enhanced HR HF oscillation amplitude. (B) In the second group (II), hERG blockers rather caused a reduction in endogenous HF oscillations in HR. (C) Only two hERG blockers of the group I (dofetilide and terfenadine) were found to cause statistically significant increases in STVQT. (D) Compounds in group II causing the largest reduction of HF oscillations also induced the largest decrease in STVQT. All drugs were given orally: CIS, cisapride; CIP, ciprofloxacin; DOF; dofetilide; EBA, ebastine; HAL, haloperidol; MOX, moxifloxacin; NIC, nicardipine; PHE, phenytoin; RIS, risperidone; SOT, DL sotalol; TER, terfenadine; THI, thioridazine; VER, verapamil. Figures close to abbreviations indicate tested dose levels in mg\u00b7kg^\u22121^ p. o. Values were determined relative to vehicle at peak of effect. Data are presented as mean values \u00b1 SEM; *n* = 6 per compound. \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, significantly different from vehicle.](bph0172-2878-f7){#fig07}\n\nDiscussion\n==========\n\nAutonomic origin of dofetilide-induced STV~QT~ in healthy animals\n-----------------------------------------------------------------\n\nInitially, beat-to-beat variability of repolarization was assessed from beat-to-beat STV derived from left ventricular monophasic action potential duration, that is STV~LVMAPD~, in anaesthetized chronic atrio-ventricular block dogs (Thomsen *et\u2009al*., [@b34]). Dofetilide was reported as increasing STV~LVMAPD~ and causing TdP in this model. TdP occurrence was associated with an increase in STV~LVMAPD~. Later, STV~QT~ was derived from QT interval in healthy conscious beagle dogs leading to the conclusion that this biomarker could be considered as better predictor than QTc prolongation for TdP liability (Lengyel *et\u2009al*., [@b22]). However, until now, no link between the ANS and STV~QT~ has been described under conditions of QT prolongation by arrhythmic hERG blockers. On the other hand, STV~APD~ derived from action potential duration (APD) was found to be increased by arrhythmic hERG blockers in several models devoid of any interactions with the ANS such as isolated cardiomyocytes (Oros *et\u2009al*., [@b27]; Abi-Gerges *et\u2009al*., [@b1]) and isolated perfused heart (Hondeghem *et\u2009al*., [@b18]) leading to the conclusion that drug-induced increased STV~APD~ is a consequence of beat-to-beat repolarization instability because of particular electrophysiological properties (named thereafter electrophysiological instability). Our work does not call this concept into question, especially from those drawn from *in vitro*/*ex vivo* models. However, the data we present here demonstrate that *in vivo* STV~QT~ increases by dofetilide in cynomolgus monkeys and beagle dogs are mainly dependent on autonomic HF oscillations in HR and do not simply reflect an erratic instability of the QT interval exclusively because of a source of electrophysiological instability. Worthy of note, we found a residual source of QT interval variability after autonomic blockade that can be attributed to the electrophysiological instability described from *in vitro*/*ex vivo* studies. Moreover, this electrophysiological instability may enhance the variability of QT interval dependent on the ANS. However, if sources of electrophysiological instability have a role, our results demonstrate that they require the presence of HF autonomic oscillations to induce changes in STV~QT~ in healthy animals. Considering the frequency band of these oscillations, the HF parasympathetic oscillations of HR play a major role in hERG blocker-induced beat-to-beat ventricular repolarization variability.\n\nDofetilide enhances QT interval HF oscillations through HR HF oscillation amplitude-dependent and -independent mechanisms\n-------------------------------------------------------------------------------------------------------------------------\n\nStudy of the relationship between HF oscillations in HR and QT interval oscillations demonstrated that QT interval oscillations physiologically increase with amplitude of HR oscillations. This relationship is not surprising as QT interval duration is highly dependent on HR, even during very short periods of 10\u2009s (Holzgrefe *et\u2009al*., [@b17]). Consequently, important changes within these short periods of 10\u2009s of beat-to-beat HR are expected to cause noticeable changes in beat-to-beat QT interval duration, that is in QT interval HF oscillations. Study of the relationship between QT interval oscillations and HF oscillations in HR demonstrates that the HF oscillations in HR were, by themselves, a source of beat-to-beat ventricular repolarization variability. Dofetilide has two distinct effects on this relationship between QT interval oscillations and HF oscillations in HR. First, it prolongs this relationship towards higher amplitude of HR oscillations, and consequently, causes an increase in amplitude of QT interval HF oscillations through this relationship. As these HR oscillations are driven by the ANS, these data demonstrate that dofetilide directly interacts with the autonomic control of HR. Second, dofetilide causes a vertical shift of QT interval oscillations towards higher amplitudes independently from amplitudes of the HF oscillations in HR. This vertical shift cannot be simply attributed to QT prolongation properties as, moxifloxacin, another hERG blocker devoid of effect on ventricular repolarization STV in chronic AV block dogs (Thomsen *et\u2009al*., [@b35]) or in cynomolgus monkeys (Supporting Information Fig. [S9](#sd1){ref-type=\"supplementary-material\"}), was unable to cause such vertical shift despite marked QT prolongation properties (Supporting Information Fig. [S10](#sd1){ref-type=\"supplementary-material\"}). These data support the hypothesis that the mechanism of this effect is likely to be a consequence of an interaction between a second source of variability and the endogenous QT interval variability generated by the ANS. As previously proposed, this second source of variability could be the beat-to-beat ventricular repolarization electrophysiological instability, visible with dofetilide under autonomic blockade. This interaction could result from this source of beat-to-beat ventricular repolarization instability with autonomic HR oscillations throughout the dependence of QT interval towards HR. Reciprocally, our results also show that QT interval changes during HR oscillations in dofetilide treated animals may influence subsequent beat-to-beat HR values and elicit correcting HR adaptation, causing, in turn, an increase in QT interval variability through rate-dependent mechanisms. This latter mechanism involving correcting HR adaptation is in accordance with abnormalities of QT/RR relationship restitution described first by Fossa ([@b13]) with arrhythmic hERG blockers.\n\nApplying to Coumel\\'s triangle\n------------------------------\n\nThe increase in HF oscillations in HR reflects an enhancement of alternate cycles of very short-term deceleration and acceleration processes over very short periods of less than 10\u2009s. Classically, HF rhythms of HR are dependent on alternate withdrawal and activation of the parasympathetic pathway. Oscillatory short-long-short (SLS) cycle length sequences were described as preceding the onset of premature ventricular beats and TdP episodes in patients with acquired prolonged ventricular repolarization (Locati *et\u2009al*., [@b23]), as we found in cynomolgus monkeys with dofetilide. Such pro-arrhythmic property of SLS sequences is still routinely used in the clinic in patients with unexplained syncopes or rhythm disorders using electrical programmed stimulations (Wellens *et\u2009al*., [@b38]). Likewise, acceleration from an initial slow rate was also reported to facilitate early after-depolarization in canine myocardial M cells and Purkinje fibres under hERG blockade via a mechanism involving intracellular calcium loading (Burashnikov and Antzelevitch, [@b5]). All these findings demonstrated that fast deceleration and acceleration processes constitute sources of triggering of ventricular arrhythmias. Interestingly, QT prolongation induced by hERG blockers was proposed to constitute a substrate for occurrence of premature ventricular beats in particular during acceleration phases (Fossa, [@b13]). Increases in HR HF rhythms have also been described in patients with congenital LQT (Shamsuzzaman *et\u2009al*., [@b32]). Independently from this latter study, STV~QT~ was found to be increased in patients with LQTs (Hinterseer *et\u2009al*., [@b15]). Taken together, our results and already published findings strongly support that an enhancement of fast deceleration and acceleration processes carried by endogenous HR parasympathetic oscillations could constitute a primary trigger factor leading secondarily to creation of favourable conditions for expression of cellular trigger factors such as re-entries or early after-depolarizations because of confluence of impaired repolarization in a context of QT prolongation.\n\nIn other words, these data support the concept of Coumel\\'s triangle of arrhythmogenesis. To develop life-threatening arrhythmia such as TdP, arrhythmic hERG channel blockers would require (i) an electrophysiological substrate, that is APD prolongation and instability (APD***~i~***); (ii) a primary trigger factor, that is enhancement of HF oscillations in HR; and (iii) a modulator factor, the ANS, which modulates both APD and endogenous HF oscillations. The combination of these three ingredients leads to the increase of ventricular repolarization STV (STV~QT~), therefore creating favourable conditions for expression of cellular triggering factors such as re-entries or early after-depolarizations. Figure\u2009[8](#fig08){ref-type=\"fig\"} proposes an adaptation of Coumel\\'s triangle as applied to dofetilide in cynomolgus monkeys that also serves to summarise the previous discussions.\n\n![Applying Coumel\\'s triangle of arrhythmogenesis to the actions of dofetilide. To develop life-threatening arrhythmia such as TdP, this agent would require (i) an electrophysiological substrate, that is APD prolongation and APD~i~; (ii) a primary trigger factor, that is enhancement of HR HF oscillations; and (iii) a modulator factor, the ANS which modulates both APD and endogenous HF oscillations. The combination of these three ingredients leads to increasing of ventricular repolarization STV (STVQT), so creating favourable conditions for expression of cellular triggering factors such as re-entries or early after depolarizations. In the context of these drug-induced arrhythmias, STVQT is a biomarker of the arrhythmic threshold above which probability for arrhythmia occurrence is high. PVB, premature ventricular beats.](bph0172-2878-f8){#fig08}\n\nPredictive value of HF oscillations in HR\n-----------------------------------------\n\nWe found that all hERG blockers able to enhance endogenous HF oscillations in HR in beagle dogs (group I) were compounds already reported as causing TdP (Redfern *et\u2009al*., [@b31]; Vieweg *et\u2009al*., [@b37]). In the second group of hERG blockers, only moxifloxacin was described as responsible for TdP at an extremely low incidence (Haverkamp *et\u2009al*., [@b14]). Consequently, this biomarker may be of relevance in safety assessment, in particular for new drug candidates exhibiting hERG-blocking properties. However, many mechanisms are susceptible to enhance endogenous HF parasympathetic oscillations in HR, such as baroreflex activation, central effects and parasympathomimetic properties. Consequently, taken alone, this biomarker cannot thus be considered as a specific surrogate for TdP liability. On the other hand, an enhancement of HF oscillations in HR could be considered as an early and sensitive surrogate for TdP liability when associated with evidence of hERG-blocking properties, such as QT prolongation. This way of indicating TdP liability is more sensitive than STV~QT~ assessment. Indeed, enhancement of HF oscillations in HR was easily detected with several arrhythmic hERG blockers, sometimes at low dose levels like thioridazine or with low TdP incidence arrhythmic drugs such as risperidone (Vieweg *et\u2009al*., [@b37]), while increases in STV~QT~ reached statistically significant levels only with a few drugs (dofetilide and terfenadine) in healthy animals. Finally, analysis of HF QT oscillations amplitude in clinical studies exploring QT variability in patients with long-QT syndrome (Hinterseer *et\u2009al*., [@b15]) could offer interesting opportunities to check whether this biomarker is more sensitive than STV~QT~ in human to reveal increased QT variability issues.\n\nIn conclusion, the present work offers a new insight by demonstrating a link between the arrhythmic profile of dofetilide and interaction with cardiac autonomic control. This concept could be very helpful as it may allow refinement of the current preclinical safety pharmacology strategy in this area.\n\nWe acknowledge the contribution of D Bouard, A Contamine and C Roubinet for their technical assistance. Jerome Thireau holds a CNRS position and was supported by Fondation de France (Project SYNAPTOCARD N\u00b0201300038586).\n\nAuthor contributions\n====================\n\nResearch studies were designed by P. C. Experiments were conducted by S. J., C. L.-B. and A. M. Data analysis and interpretation were performed by P. C. The paper was written by P. C., J. T. and J. Y. L. G., and was critically evaluated by all authors. J. T. proposed application to Coumel\\'s triangle.\n\nConflict of interest\n====================\n\nThe authors state no conflict of interest.\n\nANS\n\n:   autonomic nervous system\n\nAPD\n\n:   action potential duration\n\nEAG\n\n:   ether-a-go-go\n\nERG\n\n:   ether-a-go-go related gene\n\nGLP\n\n:   Good Laboratory Practices\n\nhERG\n\n:   human ether-a-go-go related gene\n\nHF\n\n:   high-frequency\n\nHR\n\n:   heart rate\n\nLQTS\n\n:   long QT syndrome\n\nSLS\n\n:   short-long-short\n\nSTV\n\n:   short-term variability\n\nTdP\n\n:   Torsades de pointes\n\nSupporting Information\n======================\n\nAdditional Supporting Information may be found in the online version of this article at the publisher\\'s web-site:\n\n###### \n\n**Figure\u2009S1** Typical example of determination of HF oscillations amplitude. Amplitude of HF oscillations is calculated from the difference between maximum and minimum HR values noted within a 10\u2009s sequence, that is the maximum period of HF rhythms (0.1\u2009Hz). The same principle is followed for determination of QT interval HF oscillations amplitude.\n\n**Figure\u2009S2** Regression curves between changes in beat-to-beat QTn or QTn\u22121 interval versus HR (HRn) drawn from the typical example of the Figure\u20091B (dofetilide, 1\u2009mg\u00b7kg^\u22121^, p.o.) in cynomolgus monkeys. This figure shows a better correlation between QT intervals (QTn\u22121) preceding subsequent HR values than concomitant beat-to-beat QT intervals (QTn) and HR (HRn) values. This analysis supports that changes in QT interval might influence subsequent beat-to-beat HR values and elicit correcting HR adaptation in dofetilide-treated (1\u2009mg\u00b7kg^\u22121^, p.o.) animals. This possible mechanism could contribute towards the enhancement of HF HR oscillations.\n\n**Figure\u2009S3** Regression curves between changes in STV~QT~ and HF QT interval rhythms power (A, HF QTp), HF QT interval oscillations amplitude (B, HF QTa), QT interval (C, QT) and QT interval corrected by Holzgrefe\\'s probabilistic method (D, QTcH). This figure shows that changes in STV~QT~ recorded for 12\u2009h after dosing with dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.) were more closely correlated to changes in HF QT interval rhythms power and HF QT interval oscillations amplitude than with QT or QTc interval changes. Correlations curves were drawn from mean changes calculated in relation to vehicle over the period ranging from 1 to 12\u2009h after dosing with dofetilide. R, regression coefficient; *P*, significance level.\n\n**Figure\u2009S4** (A) Hexamethonium (20\u2009mg\u00b7kg^\u22121^, i.v.) had moderate effects on HR related to suppression of cardiac autonomic control of HR, that is lowering HR during the daylight period and inhibition of the nocturnal bradycardia during the first hours of the night period. (B) QT interval was increased in parallel to lowering HR. Filled squares, vehicle; open squares, hexamethonium alone. Data are presented as mean values \u00b1 SEM (*n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with vehicle).\n\n**Figure\u2009S5** Effects of autonomic blockade by hexamethonium (20\u2009mg\u00b7kg^\u22121^, i.v., H) on dofetilide-induced HR and QT interval variability in beagle dogs. (A) Hexamethonium alone fully abolished endogenous HR HF oscillations for 4\u2009h when compared with vehicle (V). During this period, this ganglioplegic agent markedly reduced dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o., D) induced increases in the HF oscillations in HR (B), QT interval HF oscillations (C) and STV~QT~ **(**D). (E) a residual QT interval variability (STV~QT~) insensitive to autonomic blockade was still visible in the presence of dofetilide plus hexamethonium versus hexamethonium alone. (F) On the other hand, dofetilide did not cause any further beat-to-beat HR variability (STV~HR~) under autonomic blockade when compared with hexamethonium alone. Data are presented as mean values \u00b1 SEM (*n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with vehicle or hexamethonium alone).\n\n**Figure\u2009S6** (A) Relationship between QT interval and HR HF oscillations amplitudes (HF QT/HF HR) in vehicle beagle dogs (filled squares) over a 24\u2009h period: QT interval HF oscillation amplitude progressively increased with HR interval HF oscillation amplitude in vehicle-treated animals. (B) dofetilide (open squares) caused a horizontal shift (1) of the whole relationship towards higher amplitude of HF oscillations in HR. It also amplified amplitude of QT interval oscillations according to a mechanism independent from amplitude of HF oscillations in HR (2: vertical shift). The relationship was built from all data collected between 2 and 6\u2009h after dosing, that is the period during which dofetilide caused a marked QT prolongation and increase in HF oscillations in HR. Data are presented as mean values \u00b1 SEM (*n* = 6, \\*\\**P* \u2264 0.01, when compared with vehicle).\n\n**Figure\u2009S7** (A) Time course of arrhythmic events induced by dofetilide alone (1\u2009mg\u00b7kg^\u22121^, p.o., red circles) versus dofetilide in hexamethonium (20\u2009mg\u00b7kg^\u22121^, i.v., blue circles) treated cynomolgus monkeys. Data are reported as number of arrhythmic events (premature ventricular beats, doublets, triplets, ventricular tachycardia and TdP episodes). (B) Time course of HR changes induced by dofetilide alone (1\u2009mg\u00b7kg^\u22121^, p.o., red circles) versus dofetilide in hexamethonium-treated (20\u2009mg\u00b7kg^\u22121^, i.v., blue circles) animals. The analysed period of 6\u2009h corresponds to the duration of effects of hexamethonium on HF HR oscillations amplitude and the period, during which incidence of arrhythmias was analysed referring to the Figure 6 of the main paper. This figure shows that hexamethonium prevented dofetilide-induced arrhythmic events in a way independent of an elevation of HR levels. Conversely, hexamethonium slightly decreased HR in the presence of dofetilide during the first 4\u2009h when compared with dofetilide alone, that is when the incidence of arrhythmias is already high with dofetilide. Data are presented as median values (A: *n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with dofetilide alone, non-parametric Mann--Whithney *U*-test) and as mean values \u00b1 SEM (B: *n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with dofetilide alone).\n\n**Figure\u2009S8** Time course of STV~QT~ changes (left axis, blue circles) versus arrhythmic events (right axis, black circles) induced by dofetilide (1\u2009mg\u00b7kg^\u22121^, p.o.) in cynomolgus monkeys. This figure shows that dofetilide-induced arrhythmic events started to occur as soon as STV~QT~ began to increase. Moreover, arrhythmic events disappeared 12\u2009h after dosing when STV~QT~ returned to baseline. Data are presented as median values for arrhythmia incidence (*n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with vehicle, non-parametric Mann--Whithney *U*-test) and mean values \u00b1 SEM for STV~QT~ (*n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with vehicle). No arrhythmia was seen in vehicle-treated animals.\n\n**Figure\u2009S9** (A) Moxifloxacin (90\u2009mg\u00b7kg^\u22121^, p.o.) had no effect on HR (A) in cynomolgus monkeys. It caused a large QTc prolongation (B, probabilistic method) and a moderate lowering of HF oscillations in HR (C). Moxifloxacin did not change STV~QT~ (D). Filled squares: vehicle. Open squares: moxifloxacin. Data are presented as mean values \u00b1 SEM (*n* = 6, \\**P* \u2264 0.05, \\*\\**P* \u2264 0.01, when compared with vehicle).\n\n**Figure\u2009S10** Moxifloxacin (90\u2009mg\u00b7kg^\u22121^, p.o.) had no effect on the relationship between amplitudes of the QT interval and the HF oscillations in HR in cynomolgus monkeys. The relationship was built from all data collected between 2 and 12\u2009h after dosing, that is the period during which moxifloxacin caused a marked and stable QT prolongation. Data are presented as mean values \u00b1 SEM (*n* = 6). *P* \\> 0.05, when compared with vehicle.\n\n[^1]: Laboratoire INSERM U1046, Physiologie et M\u00e9decine Exp\u00e9rimentale, C\u0153ur et Muscles, CHU A. de Villeneuve, 371 Avenue du doyen G. Giraud, 34295 Montpellier cedex 05, Universit\u00e9s Montpellier 1 et Montpellier 2.\n"}
+{"text": "Introduction {#s1}\n============\n\nAutoantibodies, as detected by the indirect immunofluorescence assay (IIFA) on HEp-2 cells (IIFA HEp-2), are recognised as important diagnostic markers in a plethora of autoimmune diseases, in particular the systemic autoimmune rheumatic diseases (SARD).[@R1] Although somewhat dated by today's standards, members of the American College of Rheumatology (ACR) prepared an evidence-based guideline for the usefulness of the HEp-2 IIFA results for diagnostic and prognostic purposes and also for meeting diagnostic criteria.[@R2] That guideline was based on reactivity with nuclear antigens as detected by IIFA on rodent tissue or HEp-2 cells. More recently, the IIFA on HEp-2 cells was reinforced as the gold standard for autoantibody screening in SARD.[@R3]\n\nInterestingly, the HEp-2 IIFA test reveals much more information than the mere absence or presence of autoantibodies, that is, the level of antibody as well as the HEp-2 IIFA pattern. Based on titration or appropriate evaluation of the fluorescence intensity, the antibody level can be determined and this information has general concordance with the clinical relevance of the test result. Indeed, higher antibody levels are better associated with SARD and have an increased likelihood to identify the autoantigen in follow-up testing.[@R4] The importance of the level of autoantibodies is also recognised in the ACR guideline as well as by the recommendations issued by the European Autoimmunity Standardization Initiative (EASI) and the International Union of Immunologic Societies (IUIS) Autoantibody Standardization Subcommittee.[@R2]\n\nThe HEp-2 IIFA pattern may also reveal clinically relevant information. This information is not restricted to giving direction to follow-up testing for antigen-specificity, but, for instance, the centromere pattern is included in the classification criteria for systemic sclerosis,[@R8] while the nuclear dense fine speckled pattern is reported to be more prevalent in apparently healthy individuals as compared with patients with SARD.[@R9] To harmonise the names and descriptions of the distinct HEp-2 IIFA patterns, an ordered classification taxonomy was proposed.[@R10] This proposal was subsequently elaborated on by the International Consensus on ANA Patterns (ICAP), initiated in parallel to the 12th International Workshop on Autoantibodies and Autoimmunity (2014) held in Sao Paulo, Brazil. During this workshop, a consensus was reached on the nomenclature and definitions of 28 HEp-2 IIFA patterns. Each HEp-2 IIFA pattern was ascribed an alphanumeric code from AC-1 to AC-28.[@R11] The consensus nomenclature for each pattern and representative images were also made available online at the ICAP website (<http://www.ANApatterns.org>).\n\nIn addition to the nuclear patterns, important cytoplasmic and mitotic patterns may also be observed in HEp-2 IIFA analysis. Although reporting non-nuclear patterns is considered clinically relevant,[@R7] for various jurisdictional reasons there is no clear-cut consensus viewpoint on reporting non-nuclear patterns as a negative or positive test.[@R12] With the understanding that the term 'Antinuclear antibody (ANA) test' may be inappropriate to designate a test that also addresses autoantibodies to antigens in the cytoplasm and mitotic apparatus, an alternative name, anticellular antibodies, was suggested in the EASI/IUIS recommendations.[@R7] Recent publications from ICAP have preferred the term HEp-2 IIFA as it covers the whole spectrum of patterns that can be observed when using the HEp-2 cells as substrate.[@R13]\n\nOriginally, the HEp-2 IIFA patterns were associated with diseases, but it was anticipated that many of these associations are only valid if the antigen-specificity was confirmed by follow-up testing. In subsequent ICAP workshops, it was agreed that the disease associations should be replaced by clinical relevance. In this current paper, we present the consensus on the clinical relevance of the distinct HEp-2 IIFA patterns as achieved by consecutive workshops and discussions among the executive ICAP members.\n\nMaterials and methods {#s2}\n=====================\n\nFor discussion about the structure of clinical relevance templates were prepared for AC-2 (LECA), AC-3 (JD) and AC-5 (MS). This formed the basis of a guideline for description of each AC pattern (EC). Of highest importance, it was agreed that the information should be objective and helpful for the clinician, the pattern--antigen associations should be put in the right clinical context and information should be evidence-based.\n\nIn preparation for the third ICAP workshop in Kyoto (2016), composition of the clinical relevance documents was started for the nuclear patterns (JD, LECA, MS), cytoplasmic patterns (CAvM, EKLC) and mitotic patterns (MH, TM). As far as already available, the documents were commented on by the ICAP executive board and, after appropriate adjustment, discussed with the workshop participants. The feedback from participants mainly focused on the structure of the information provided, on the required level of detail and the format of recommended follow-up testing.\n\nIn anticipation of the fourth ICAP workshop in Dresden (2017), the set of clinical relevance documents was completed for all patterns. Further comments from the ICAP executive board were included. The resulting documents were individually discussed with the workshop participants for nuclear (JD), cytoplasmic (CAvM) and mitotic (MH) patterns. Besides several substantive comments, there was general agreement that the information should be provided in tabular format at two distinct levels. The first level should contain information on relevant follow-up testing in the respective clinical context, the recommended follow-up tests should be commercially available and detailed test characteristics should not be given because of potential geographic and jurisdictional differences. Information based on case reports or small patient cohorts, as well as information on possible follow-up testing that is only available in specialised research laboratories, should only be provided in the second level information.\n\nTables for nuclear, cytoplasmic and mitotic patterns were prepared for first and second level information (JD). These tables were commented by the ICAP executive board and finalised by JD. Of note, since the starting point of the tables on clinical relevance is the HEp-2 IIFA pattern and not the clinically suspected disease, the tables do not list all autoantibodies related to the respective disease.\n\nResults {#s3}\n=======\n\nNuclear HEp-2 IIFA patterns {#s3-1}\n---------------------------\n\nTo date, a total of 15 nuclear HEp-2 IIFA patterns have been described, that is, AC-1--AC-14 and AC-29. [Table 1](#T1){ref-type=\"table\"} summarises the clinical relevance of these patterns.[@R8] Since AC-29 was only recently described,[@R14] the advice for follow-up testing for autoantibodies to topoisomerase I (Scl-70) in case of clinical suspicion of systemic sclerosis is also added as a note to the clinical relevance of AC-1. In particular, disease-specific immunoassays, like autoimmune liver disease profile, inflammatory myopathy profile, systemic sclerosis profile, are often only available in specialty clinical laboratories.\n\n###### \n\nNuclear HEp-2 IIFA patterns\n\n  Code                                                                                                                                                                                                                                                                                                                                                                                                                               AC pattern---clinical relevance                                                                                                Refs\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------ ------\n  AC-1                                                                                                                                                                                                                                                                                                                                                                                                                               HOMOGENEOUS                                                                                                                    \n  Found in patients with SLE, chronic autoimmune hepatitis or juvenile idiopathic arthritis                                                                                                                                                                                                                                                                                                                                                                                                                                                                         \n  If SLE is clinically suspected, it is recommended to perform a follow-up test for anti-dsDNA antibodies, alone or in combination with dsDNA/histone complexes (nucleosomes/chromatin); anti-dsDNA antibodies are included in the classification criteria for SLE                                                                                                                                                                   [@R15], [@R16]                                                                                                                 \n  If chronic autoimmune hepatitis or juvenile idiopathic arthritis is suspected, follow-up testing is not recommended because the respective autoantigens revealing the AC-1 pattern are not completely defined                                                                                                                                                                                                                      [@R17]                                                                                                                         \n  Notes: Although autoantibodies to Topoisomerase I (formerly Scl-70) may be reported as nuclear homogeneous, they typically reveal a composite AC-29 HEp-2 IIFA pattern; as such, clinical suspicion of SSc may warrant follow-up testing for reactivity to this antigen.                                                                                                                                                           [@R14], [@R18]                                                                                                                 \n  Although AC-1 is the most prevalent pattern in chronic autoimmune hepatitis, other HEp-2 IIFA patterns may occur, but also for these patterns the autoantigens are not completely defined.                                                                                                                                                                                                                                         [@R19]                                                                                                                         \n  AC-2                                                                                                                                                                                                                                                                                                                                                                                                                               DENSE FINE SPECKLED                                                                                                            \n  Commonly found as high titer HEp-2 IIFA-positive in apparently healthy individuals or in patients who do not have a systemic autoimmune rheumatic disease (SARD)                                                                                                                                                                                                                                                                   [@R9]                                                                                                                          \n  The negative association with SARD is only valid if the autoreactivity is confirmed as being directed to DFS70 (also known as LEDGF/p75) and if no other common ENA is recognized                                                                                                                                                                                                                                                  [@R20], [@R21]                                                                                                                 \n  Both in apparently healthy individuals as well as patients who do not have a SARD the AC-2 pattern may be caused by autoantibodies to other antigens than DFS70                                                                                                                                                                                                                                                                    [@R22]                                                                                                                         \n  Note: Confirmatory assays for anti-DFS70 antibodies may be available only in specialty clinical laboratories.                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  AC-3                                                                                                                                                                                                                                                                                                                                                                                                                               CENTROMERE (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details                   \n  Commonly found in patients with limited cutaneous SSc, and as such included in the classification criteria for SSc                                                                                                                                                                                                                                                                                                                 [@R8], [@R15], [@R23]                                                                                                          \n  In combination with Raynaud phenomenon, the AC-3 pattern is prognostic for onset of limited cutaneous SSc                                                                                                                                                                                                                                                                                                                          [@R15], [@R23]                                                                                                                 \n  Strongly associated with antibodies to CENP-B; especially in case of low titers, confirmation by an antigen-specific immunoassay is recommended to support the association with limited cutaneous SSc; the CENP-B antigen is included in many routine ENA profiles                                                                                                                                                                 [@R15]                                                                                                                         \n  The AC-3 pattern is also apparent in a subset of patients with PBC; these patients often have both SSc as well as PBC                                                                                                                                                                                                                                                                                                              [@R15]                                                                                                                         \n  AC-4                                                                                                                                                                                                                                                                                                                                                                                                                               FINE SPECKLED                                                                                                                  \n  Present to a varying degree in distinct SARD, in particular SjS, SLE, subacute cutaneous lupus erythematosus, neonatal lupus erythematosus, congenital heart block, DM, SSc, and SSc-AIM overlap syndrome                                                                                                                                                                                                                          [@R15]                                                                                                                         \n  If SjS, SLE, subacute cutaneous lupus erythematosus, neonatal lupus erythymatosus, or congenital heart block is clinically suspected, it is recommended to perform follow-up tests for anti-SS-A/Ro (Ro60) and anti-SS-B/La antibodies; in most laboratories these antigens are included in the routine ENA profile                                                                                                                [@R15]                                                                                                                         \n  Autoantibodies to SS-A/Ro are part of the classification criteria for SjS (the criteria do not distinguish between Ro60 and Ro52/TRIM21)                                                                                                                                                                                                                                                                                           [@R25]                                                                                                                         \n  If SSc, AIM, or to a lesser extend SLE, is clinically suspected, it is recommended to perform follow-up tests for detecting autoantibodies to Mi-2, TIF1\u03b3, and Ku; these antigens are typically included in disease specific immunoassays (i.e., inflammatory myopathy profile\\*)                                                                                                                                                  [@R26]                                                                                                                         \n  Autoantibodies to Mi-2 and TIF1\u03b3 are associated with DM; autoantibodies to TIF1\u03b3 in patients with DM, although rare in the overall AC-4 pattern, is strongly associated with malignancy in old patients                                                                                                                                                                                                                            [@R26], [@R27]                                                                                                                 \n  Autoantibodies to Ku are associated with SSc-AIM and SLE-SSc-AIM overlap syndromes                                                                                                                                                                                                                                                                                                                                                 [@R26]                                                                                                                         \n  Notes: Anti-SS-A/Ro (Ro60) and AIM-specific autoantibodies may be undetected in HEp-2 IIFA-screening.                                                                                                                                                                                                                                                                                                                              [@R28]                                                                                                                         \n  AC-5                                                                                                                                                                                                                                                                                                                                                                                                                               LARGE/COARSE SPECKLED (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details)       \n  Present to a varying degree in distinct SARD, in particular SLE, SSc, MCTD, SSc-AIM overlap syndrome, and UCTD (i.e, patients with rheumatic symptoms without a definite SARD diagnosis)                                                                                                                                                                                                                                           [@R29]                                                                                                                         \n  If SLE is clinically suspected, it is recommended to perform follow-up tests for anti-Sm and anti-U1RNP antibodies; these antigens are commonly included in the routine ENA profile; anti-Sm antibodies are included in the classification criteria for SLE                                                                                                                                                                        [@R16], [@R30], [@R31]                                                                                                         \n  If SSc is clinically suspected, it is recommended to perform a follow-up test for anti-RNApol III antibodies (e.g., SSc profile\\*); the anti-RNApol III antibodies are included in the classification criteria for SSc                                                                                                                                                                                                             [@R8]                                                                                                                          \n  If MCTD is clinically suspected, it is recommended to perform a follow-up test for anti-U1RNP antibodies; the antigen is commonly included in the routine ENA profile; anti-U1RNP antibodies are included in the diagnostic criteria for MCTD                                                                                                                                                                                      [@R32]                                                                                                                         \n  If the SSc-AIM overlap syndrome is clinically suspected, it is recommended to perform follow-up tests for anti-U1RNP and anti-Ku antibodies; these antigens are included in the routine ENA profile (U1RNP), or in disease specific immunoassays (Ku, i.e., inflammatory myopathy profile\\* and SSc profile\\*)                                                                                                                     [@R26], [@R33]                                                                                                                 \n  In non-SARD individuals in the general population, the presence of the AC-5 pattern is not associated with the autoantigens mentioned above and most often concerns low antibody titers                                                                                                                                                                                                                                                                                                                                                                           \n  AC-6                                                                                                                                                                                                                                                                                                                                                                                                                               MULTIPLE NUCLEAR DOTS                                                                                                          \n  Found in a broad spectrum of autoimmune diseases, including PBC, AIM (DM), as well as other inflammatory conditions                                                                                                                                                                                                                                                                                                                [@R34]                                                                                                                         \n  If PBC is clinically suspected, it is recommended to perform follow-up tests for anti-Sp100 (and PML/Sp140) antibodies; in particular anti-Sp100 antibodies have the best clinical association with PBC and have added value, especially when associated with AMA; the Sp100 (and PML-Sp140) antigen is included in disease specific immunoassays (ie, liver profile\\*)                                                            [@R17], [@R35], [@R36]                                                                                                         \n  If DM is clinically suspected, it is recommended to perform a follow-up test for anti-MJ/NXP-2 antibodies; these anti-MJ/NXP-2 antibodies are highly specific for AIM, are found in up to one third of patients with juvenile DM, and have been reported to be associated with malignancies in adult AIM patients; the antigen is included in disease specific immunoassays (i.e., inflammatory myopathy profile\\*)                [@R37]                                                                                                                         \n  AC-7                                                                                                                                                                                                                                                                                                                                                                                                                               FEW NUCLEAR DOTS (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details)            \n  The AC-7 pattern has low positive predictive value for any disease                                                                                                                                                                                                                                                                                                                                                                 [@R40], [@R41]                                                                                                                 \n  Antigens primarily localized in the dots include p80-coilin and SMN complex; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                               [@R42], [@R43]                                                                                                                 \n  AC-8                                                                                                                                                                                                                                                                                                                                                                                                                               HOMOGENEOUS NUCLEOLAR (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details)       \n  Found in patients with SSc, SSc-AIM overlap syndrome, and patients with clinical manifestations of other SARD                                                                                                                                                                                                                                                                                                                      [@R44]                                                                                                                         \n  If limited cutaneous SSc is clinically suspected, it is recommended to perform a follow-up test for anti-Th/To antibodies; the antigen is included in disease specific immunoassays (ie, SSc profile\\*)                                                                                                                                                                                                                            [@R44], [@R45]                                                                                                                 \n  If SSc-AIM overlap syndrome is clinically suspected, it is recommended to perform a follow-up test for anti-PM/Scl antibody reactivity; the antigen may be included in the routine ENA profile and is included in disease specific immunoassays (i.e., inflammatory myopathy profile\\* and the SSc profile\\*); in general, anti-PM/Scl antibodies yield a diffuse nuclear fine speckled staining in addition to the AC-8 pattern   [@R46]                                                                                                                         \n  Other antigens recognized include B23/nucleophosmin, No55/SC65, and C23/nucleolin, but the clinical significance of these autoantibodies is not well established; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                                                         \n  Notes: Although some anti-Th/To antibody immunoassays are commercially available, technical issues relating to the limited sensitivity of these immunoassays should be taken in to consideration.                                                                                                                                                                                                                                  [@R44], [@R47]                                                                                                                 \n  AC-9                                                                                                                                                                                                                                                                                                                                                                                                                               CLUMPY NUCLEOLAR                                                                                                               \n  Found in patients with SSc                                                                                                                                                                                                                                                                                                                                                                                                         [@R48]                                                                                                                         \n  If SSc is clinically suspected, it is recommended to perform a follow-up test for anti-U3RNP/fibrillarin antibodies; the antigen is included in disease specific immunoassays (i.e, SSc profile\\*)                                                                                                                                                                                                                                 [@R48]                                                                                                                         \n  If confirmed as anti-U3RNP/fibrillarin reactivity by immunoassay, the clinical association is with diffuse SSc, increased incidence of pulmonary arterial hypertension, skeletal muscle disease, severe cardiac involvement, and gastrointestinal dysmotility                                                                                                                                                                      [@R23], [@R48]                                                                                                                 \n  Among SSc patients, anti-U3RNP/fibrillarin antibodies are most commonly found in African American and Latin American patients                                                                                                                                                                                                                                                                                                      [@R48], [@R49], [@R51]                                                                                                         \n  Notes: Although some anti-U3RNP/fibrillarin immunoassays are commercially available, technical issues relating to the limited sensitivity of these immunoassays should be taken into consideration.                                                                                                                                                                                                                                [@R24]                                                                                                                         \n  AC-10                                                                                                                                                                                                                                                                                                                                                                                                                              PUNCTATE NUCLEOLAR                                                                                                             \n  The AC-10 pattern can be seen in various conditions, including SSc, Raynaud's phenomenon, SjS, and cancer                                                                                                                                                                                                                                                                                                                          [@R52]                                                                                                                         \n  If the AC-10 pattern is observed in the serum of patients with conditions mentioned above, follow-up testing for anti-NOR90(hUBF) antibodies is to be considered; the antigen is included in disease specific immunoassays (i.e. SSc profile\\*)                                                                                                                                                                                    [@R54], [@R55]                                                                                                                 \n  While AC-10 is associated with anti-RNApol I antibodies, these antibodies almost always coexist with anti-RNApol III antibodies which reveal the AC-5 pattern; therefore, if SSc is clinically suspected, it is recommended to perform a follow-up test for anti-RNApol III antibodies (See also AC-5); specific immunoassays for anti-RNApol I antibodies are currently not commercially available                                [@R52], [@R53], [@R57]                                                                                                         \n  AC-11                                                                                                                                                                                                                                                                                                                                                                                                                              SMOOTH NUCLEAR ENVELOPE                                                                                                        \n  The AC-11 pattern is infrequently found in routine autoantibody testing and has been described in autoimmune-cytopenias, autoimmune liver diseases, linear scleroderma, APS, and SARD; current information on clinical associations is based mainly on case reports and small cohorts                                                                                                                                              [@R58]                                                                                                                         \n  Antigens recognized include lamins (A, B, C) and LAP-2; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                                    [@R58]                                                                                                                         \n  AC-12                                                                                                                                                                                                                                                                                                                                                                                                                              PUNCTATE NUCLEAR ENVELOPE (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details)   \n  Found in patients with PBC, as well as patients with other autoimmune liver diseases and SARD                                                                                                                                                                                                                                                                                                                                      [@R61]                                                                                                                         \n  If PBC is clinically suspected, it is recommended to perform a follow-up test for anti-gp210 antibodies; the antigen is included in disease specific immunoassays (ie, extended liver profile\\*)                                                                                                                                                                                                                                   [@R62]                                                                                                                         \n  Other antigens recognized include p62 nucleoporin, LBR, and Tpr; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                           [@R65]                                                                                                                         \n  AC-13                                                                                                                                                                                                                                                                                                                                                                                                                              PCNA-like (see [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} for further details)                   \n  The AC-13 pattern has formerly been considered highly specific for SLE, but this specificity is debated                                                                                                                                                                                                                                                                                                                            [@R69], [@R70]                                                                                                                 \n  If SLE is clinically suspected, it is recommended to perform a follow-up test for anti-PCNA antibodies; the antigen is included in several routine ENA profiles                                                                                                                                                                                                                                                                    [@R69]                                                                                                                         \n  Recent studies with antigen-specific immunoassays show clinical associations also with SSc, AIM, RA, HCV, and other conditions                                                                                                                                                                                                                                                                                                     [@R70]                                                                                                                         \n  AC-14                                                                                                                                                                                                                                                                                                                                                                                                                              CENP-F-like                                                                                                                    \n  The majority of sera exhibiting the AC-14 pattern are from patients with a diversity of neoplastic conditions (breast, lung, colon, lymphoma, ovary, brain); paradoxically, the frequency of the AC-14 pattern in patient cohorts with these malignancies is low                                                                                                                                                                                                                                                                                                  \n  The AC-14 pattern is also seen in inflammatory conditions (Crohn's disease, autoimmune liver disease, SjS, graft-versus-host disease); current information on clinical associations is based mainly on case reports and series of cases                                                                                                                                                                                                                                                                                                                           \n  Possible associations only hold if the reactivity to CENP-F is confirmed in an antigen-specific immunoassay; current information on clinical associations is based mainly on case reports and series of cases; specific immunoassays for this autoantibody are currently not commercially available                                                                                                                                [@R74]                                                                                                                         \n  AC-29                                                                                                                                                                                                                                                                                                                                                                                                                              *TOPOI-like*                                                                                                                   \n  The AC-29 pattern is highly specific for SSc, in particular with diffuse cutaneous SSc and more aggressive forms of SSc                                                                                                                                                                                                                                                                                                            [@R14], [@R18], [@R23]                                                                                                         \n  If SSc is clinically suspected, it is recommended to perform a follow-up test for anti-Topoisomerase I (formerly Scl-70) antibodies; the anti-Topoisomerase I antibodies are included in the classification criteria for SSc and the antigen is included in routine ENA profiles                                                                                                                                                   [@R8], [@R23], [@R79]                                                                                                          \n\n\\*Availability of the inflammatory myopathy profile, the SSc profile and the (extended) liver profile may be limited to specialty clinical laboratories.\n\nAIM, autoimmune myopathy;AMA, antimitochondrial antibodies;APS, antiphospholipid syndrome;CENP, centromere-associated protein;DFS, dense fine speckled;DM, dermatomyositis;ENA, extractable nuclear antigens;HCV, hepatitis C virus;IIFA, indirect immunofluorescence assay;LAP, lamin-associated polypeptide;LBR, lamin B receptor;LEDGF, lens epithelial derived growth factor;NOR, nucleolus organiser region;NXP, nuclear matrix protein;PBC, primary biliary cholangitis;PCNA, proliferating cell nuclear antigen;PML, promyelocytic leukaemia;PM/Scl, polymyositis-scleroderma;RA, rheumatoid arthritis;RNApol, RNA polymerase;RNP, ribonucleoprotein;SARD, systemic autoimmune rheumatic diseases;SLE, systemic lupus erythematosus;SMN, survival of motor neuron;SSc, systemic sclerosis;SjS, Sj\u00f6gren's syndrome;TIF, transcription intermediary factor;TRIM, tripartite motif;Tpr, translocated promoter region; UCTD, undifferentiated connective tissue disease; dsDNA, double stranded DNA;hUBF, human upstream binding factor.\n\n10.1136/annrheumdis-2018-214436.supp1\n\nFor six nuclear HEp-2 IIFA patterns (AC-3, 5, 7, 8, 12 and 13), additional information about clinical relevance is summarised in [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"}. Although some assays for anti-CENP-A antibodies are commercially available, these antibodies are included in [online supplementary table S1](#SP1){ref-type=\"supplementary-material\"} because the majority of sera revealing the AC-3 pattern are also reactive with CENP-B. In contrast to CENP-A, CENP-B is included in many routine extractable nuclear antigens profiles.\n\nCytoplasmic HEp-2 IIFA patterns {#s3-2}\n-------------------------------\n\n[Table 2](#T2){ref-type=\"table\"} summarises the clinical relevance of the nine cytoplasmic HEp-2 IIFA patterns, that is, AC-15--AC-23.[@R26] It is recognised that the distinction between AC-19 (dense fine speckled) and AC-20 (fine speckled) can be challenging. Moreover, within the spectrum of anti-tRNA synthetase antibodies, not all produce an HEp-2 IIFA pattern and only some anti-Jo-1 antibodies are considered to give the AC-20 pattern, while the other anti-tRNA synthetase antibodies (EJ, KS, OJ, PL-7 and PL-12) are more likely to reveal the AC-19 pattern. Solid information on the pattern of two additional anti-tRNA synthetase antibodies (Ha and Zo) is lacking. Overall, the relation between these two cytoplasmic HEp-2 IIFA patterns and the distinct anti-tRNA synthetase antibodies is subject to further discussion. In clinical practice, the complete spectrum of the anti-tRNA synthetase antibodies should be determined irrespective of the subtype of cytoplasmic speckled pattern, that is, AC-19 or AC-20.\n\n###### \n\nCytoplasmic HEp-2 IIFA patterns\n\n  Code                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 AC pattern-----clinical relevance                                                                                          Refs\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ -------------------------------------------------------------------------------------------------------------------------- ------\n  AC-15                                                                                                                                                                                                                                                                                                                                                                                                                                                                                FIBRILLAR LINEAR (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)        \n  Found in patients with AIH type 1, chronic HCV infection, and celiac disease (IgA isotype); rare in SARD                                                                                                                                                                                                                                                                                                                                                                             [@R17]                                                                                                                     \n  If AIH type 1 is clinically suspected, it is recommended to confirm reactivity with smooth muscle antibodies (IgG isotype), typically detected by IIFA on rodent tissue (liver, stomach, kidney); anti-smooth muscle antibodies are included in the international criteria for AIH type 1                                                                                                                                                                                            [@R17],[@R80]                                                                                                              \n  F-actin is the main target antigen of anti-smooth muscle antibodies in AIH type 1; autoantibodies to F-actin are of more clinical importance than antibodies to G-actin                                                                                                                                                                                                                                                                                                              [@R81]                                                                                                                     \n  Notes: Although anti-F-actin immunoassays are commercially available, technical issues relating to the sensitivity of these immunoassays should be taken into consideration.                                                                                                                                                                                                                                                                                                                                                                                                                                    \n  AC-16                                                                                                                                                                                                                                                                                                                                                                                                                                                                                FIBRILLAR FILAMENTOUS (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)   \n  Found in various diseases, but AC-16 is not typically found in SARD                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             \n  Antigens recognized include cytokeratins 8, 18, & 19, tubulin, and vimentin; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                                                                                                                                                                                            \n  AC-17                                                                                                                                                                                                                                                                                                                                                                                                                                                                                FIBRILLAR SEGMENTAL (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)     \n  Found very infrequently in a routine serology diagnostic setting                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n  Antigens recognized include \u03b1-Actinin and Vinculin; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n  AC-18                                                                                                                                                                                                                                                                                                                                                                                                                                                                                DISCRETE DOTS (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)           \n  Autoantibodies revealing the AC-18 pattern have been reported in distinct SARD and in a variety of other diseases; their prevalence in unselected or specified disease cohorts has not been thoroughly studied                                                                                                                                                                                                                                                                       [@R84]                                                                                                                     \n  Antigens recognized include GW-body (Processing or P body) antigens (Ge-1/Hedls, GW182, and Su/Ago2) and endosomal antigens (EEA1, CLIP-170, GRASP-1, and LBPA); specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                                                                                                                                                                                                        \n  Notes: Autoantibodies to GW-bodies and endosomes may yield slightly different HEp-2 IIFA patterns.                                                                                                                                                                                                                                                                                                                                                                                   [@R84], [@R85]                                                                                                             \n  AC-19                                                                                                                                                                                                                                                                                                                                                                                                                                                                                DENSE FINE SPECKLED (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)     \n  Found in patients with SLE and the anti-synthetase syndrome (a subset of AIM), interstitial lung disease, polyarthritis, Raynaud's phenomenon, and mechanic's hands; these features may occur in various combinations or as an isolated manifestation, especially interstitial lung disease                                                                                                                                                                                          [@R33], [@R86], [@R87]                                                                                                     \n  If SLE is clinically suspected, follow-up tests for antibodies to ribosomal P phosphoproteins (P0, P1, P2, C22 RibP peptide) are recommended; these antigens may be included in the routine ENA profile                                                                                                                                                                                                                                                                                                                                                                                                         \n  Anti-RibP antibodies have been associated in some studies with neuropsychiatric lupus, and in childhood-onset SLE with autoimmune hemolytic anemia                                                                                                                                                                                                                                                                                                                                   [@R86], [@R88], [@R89]                                                                                                     \n  If AIM, in particular the anti-synthetase syndrome, is suspected, it is recommended to perform follow-up tests for antibodies to tRNA synthetases; antigens are included in disease specific immunoassays (ie, inflammatory myopathy profile\\*)                                                                                                                                                                                                                                      [@R26], [@R33]                                                                                                             \n  If AIM, in particular necrotizing myopathy, is suspected, it is recommended to perform follow-up tests for anti-SRP antibodies; the antigen is included in disease specific immunoassays (ie, inflammatory myopathy profile\\*)                                                                                                                                                                                                                                                       [@R26]                                                                                                                     \n  Notes: The fine distinction between AC-19 and -20 may depend on HEp-2 substrates and/or antibody concentration; antibodies to both RibP as well as tRNA synthetases may be undetected in HEp-2 IIFA-screening.                                                                                                                                                                                                                                                                                                                                                                                                  \n  AC-20                                                                                                                                                                                                                                                                                                                                                                                                                                                                                FINE SPECKLED                                                                                                              \n  Found in patients with the anti-synthetase syndrome (a subset of AIM), interstitial lung disease, polyarthritis, Raynaud's phenomenon, and mechanic's hands; these features may occur in various combinations or as an isolated manifestation, especially interstitial lung disease                                                                                                                                                                                                  [@R33], [@R90]                                                                                                             \n  Autoantibodies associated with the AC-20 pattern are primarily reported for the anti-Jo-1 antibody, which recognizes histidyl-tRNA synthetase; since AC-20 is not specific for Jo-1, it is recommended to perform a follow-up test for anti-Jo-1 antibodies; the antigen is included in the routine ENA profile, as well as in disease specific immunoassays (i.e., inflammatory myopathy profile\\*); the anti-Jo-1 antibodies are included in the classification criteria for AIM   [@R91], [@R92]                                                                                                             \n  Notes: The fine distinction between AC-19 and -20 may depend on HEp-2 substrates and/or antibody concentration; antibodies to Jo-1 may be undetected in HEp-2 IIFA-screening.                                                                                                                                                                                                                                                                                                                                                                                                                                   \n  AC-21                                                                                                                                                                                                                                                                                                                                                                                                                                                                                RETICULAR/AMA                                                                                                              \n  Commonly found in PBC, but also detected in SSc, including PBC-SSc overlap syndrome and PBC-SjS overlap syndrome                                                                                                                                                                                                                                                                                                                                                                     [@R93]                                                                                                                     \n  If PBC is clinically suspected it is recommended to perform a follow-up test for AMA, historically detected by IIFA on rodent tissue (liver, stomach, kidney); these autoantibodies are primarily directed to the PDH complex, and in particular the E2-subunit (PDH-E2); the antigen is included in disease specific immunoassays (i.e., liver profile\\*) as well as in some routine ENA profiles                                                                                   [@R93], [@R94]                                                                                                             \n  Additional antigens recognized include the E1\u03b1 and E1\u03b2 subunits of PDH, the E3-binding protein of PDH, and the 2-OGDC; these antigens are only included in extended disease specific immunoassays (i.e., extended liver profile\\*)                                                                                                                                                                                                                                                   [@R93], [@R94]                                                                                                             \n  AC-22                                                                                                                                                                                                                                                                                                                                                                                                                                                                                POLAR/GOLGI-like (see [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)        \n  Found in small numbers of patients with a variety of conditions                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n  Antigens recognized include giantin/macrogolgin and distinct golgin molecules; specific immunoassays to detect autoantibodies directed to specific Golgi antigens are currently not commercially available                                                                                                                                                                                                                                                                           [@R85]                                                                                                                     \n  AC-23                                                                                                                                                                                                                                                                                                                                                                                                                                                                                RODS and RINGS (see ([online supplementary table S2](#SP2){ref-type=\"supplementary-material\"} for further details)         \n  Most commonly found in HCV patients who have been treated with pegylated interferon-\u03b1/ribavirin combination therapy, but autoantibodies revealing the AC-23 patterns were undetected prior to treatment; as the use of interferon-\u03b1/ribavirin in HCV treatment is decreasing, the frequency and clinical associations of the AC-23 pattern may change                                                                                                                                [@R98]                                                                                                                     \n  Specific immunoassays to detect autoantibodies directed to specific Rods and Rings antigens, for instance IMPDH2, are not commercially available                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n  Note: Presence of the AC-23 pattern depends on the HEp-2 cell substrate.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        \n\n\\*Availability of the inflammatory myopathy profile, the SSc profile and the (extended) liver profile may be limited to specialty clinical laboratories.\n\nAIH, autoimmune hepatitis; AIM, autoimmune myopathy; AMA, anti-mitochondrial antibodies; APS, antiphospholipid syndrome; Ago, argonaute protein; CENP, centromere-associated protein; CLIP, class II-associated invariant chain peptide; DFS, dense fine speckled; DM, dermatomyositis; EEA, early endosome antigen; ENA, extractable nuclear antigens; HCV, hepatitis C virus; IFA, indirect immunofluorescence assay; LAP, lamin-associated polypeptide; LBR, lamin B receptor; LEDGF, lens epithelial derived growth factor; NOR, nucleolus organizer region; NXP, nuclear matrix protein; PBC, primary biliary cholangitis; PCNA, proliferating cell nuclear antigen; PML, promyelocytic leukaemia; PM/Scl, polymyositis-scleroderma; RA, rheumatoid arthritis; RNApol, ribonucleic acid polymerase; RNP, ribonucleoprotein; SARD, systemic autoimmune rheumatic diseases; SLE, systemic lupus erythematosus; SMN, survival of motor neuron; SRP, signal recognition protein; SSc, systemic sclerosis; SjS, Sj\u00f6gren's syndrome; TIF, transcription intermediary factor; TRIM, tripartite motif; Tpr, translocated promoter region; dsDNA, double stranded deoxyribonucleic acid; hUBF, human upstream binding factor; tRNA, transfer ribonucleic acid.\n\n10.1136/annrheumdis-2018-214436.supp2\n\nFor seven cytoplasmic HEp-2 IIFA patterns (AC-15--AC-19, AC-22 and AC-23), more detailed information is provided in [online supplementary table S2](#SP2){ref-type=\"supplementary-material\"}. In particular, for AC-16--AC-18, the clinical associations are quite diverse, depending on the antigen recognised. Overall, the clinical associations provided are primarily based on antigen-specific immunoassays and not on the HEp-2 IIFA pattern as such.\n\nMitotic HEp-2 IIFA patterns {#s3-3}\n---------------------------\n\nThe clinical relevance of the five mitotic patterns is summarised in [table 3](#T3){ref-type=\"table\"},[@R102] with more detailed information in [online supplementary table S3](#SP3){ref-type=\"supplementary-material\"}. As for the cytoplasmic patterns, clinical associations for the mitotic patterns are primarily based on antigen-specific immunoassays and not on the HEp-2 IIFA pattern as such.\n\n10.1136/annrheumdis-2018-214436.supp3\n\n###### \n\nMitotic HEp-2 IIFA patterns\n\n  Code                                                                                                                                                                                                                                                                                                         AC pattern---clinical relevance                                                                                          Refs\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------------------------------ ------\n  AC-24                                                                                                                                                                                                                                                                                                        CENTROSOME (see [online supplementary table 3](#SP3){ref-type=\"supplementary-material\"} for further details)             \n  The AC-24 pattern has low positive predictive value for any disease                                                                                                                                                                                                                                                                                                                                                                   \n  Within the spectrum of the SARD, the AC-24 pattern is found in patients with Raynaud's phenomenon, localized scleroderma, SSc, SLE and RA, either alone or in combination with other SSc-associated antibodies;                                                                                              [@R102]                                                                                                                  \n  Antigens recognized include \u03b1-enolase, \u03b3-enolase, ninein, Cep-250, Mob1, PCM-1/2, pericentrin; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                       [@R104], [@R106]                                                                                                         \n  AC-25                                                                                                                                                                                                                                                                                                        SPINDLE FIBERS (see [online supplementary table 3](#SP3){ref-type=\"supplementary-material\"} for further details)         \n  The AC-25 pattern has low positive predictive value for any disease                                                                                                                                                                                                                                          [@R109]                                                                                                                  \n  Found very infrequently in a routine serology diagnostic setting                                                                                                                                                                                                                                             [@R109]                                                                                                                  \n  Antigen recognized includes HsEg5; specific immunoassays for this autoantibody, or other spindle fiber targets, are currently not commercially available                                                                                                                                                     [@R110], [@R111]                                                                                                         \n  AC-26                                                                                                                                                                                                                                                                                                        NuMA-like                                                                                                                \n  Approximately one-half of the patients with the AC-26 pattern have clinical features of a SARD (SjS, SLE, UCTD, limited SSc, or RA); the AC-26 pattern is also observed in patients with organ-specific autoimmune diseases and less frequently in non-autoimmune conditions, especially when in low titer   [@R109], [@R111]                                                                                                         \n  Found very infrequently in a routine serology diagnostic setting                                                                                                                                                                                                                                             [@R109]                                                                                                                  \n  Antigens recognized include NuMA, centrophilin, SP-H antigen and NMP-22; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                             [@R115]                                                                                                                  \n  AC-27                                                                                                                                                                                                                                                                                                        INTERCELLULAR BRIDGE (see [online supplementary table 3](#SP3){ref-type=\"supplementary-material\"} for further details)   \n  The AC-27 pattern has low positive predictive value for any disease                                                                                                                                                                                                                                          [@R116]                                                                                                                  \n  Found very infrequently in a routine serology diagnostic setting                                                                                                                                                                                                                                             [@R117]                                                                                                                  \n  Antigens recognized include, among other, CENP-E, CENP-F, TD60, MSA36, KIF-14, MKLP-1, MPP1/KIF20B, and INCENP; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                      [@R116], [@R118], [@R119]                                                                                                \n  AC-28                                                                                                                                                                                                                                                                                                        MITOTIC CHROMOSOMAL (see [online supplementary table 3](#SP3){ref-type=\"supplementary-material\"} for further details)    \n  The AC-28 pattern has low positive predictive value for any disease                                                                                                                                                                                                                                                                                                                                                                   \n  Found very infrequently in a routine serology diagnostic setting                                                                                                                                                                                                                                             [@R120]                                                                                                                  \n  Antigens recognized include DCA, MCA1, and MCA5; specific immunoassays for these autoantibodies are currently not commercially available                                                                                                                                                                     [@R120]                                                                                                                  \n\nCENP, centromere-associated protein; Cep, centrosomal protein; DCA, dividing cell antigen; IIFA, indirect immunofluorescence assay; INCENP, inner centromere protein; KIF, kinesin family; MCA, mitotic chromosomal antigen; MKLP, mitotic kinesin-like protein; MPP, M-phase phosphoprotein; MSA, mitotic spindle apparatus; NMP, nuclear matrix protein; NuMA, nuclear mitotic apparatus; PCM, pericentriolar material; RA, rheumatoid arthritis; SARD, systemic autoimmune rheumatic diseases; SLE, systemic lupus erythematosus; SSc, systemic sclerosis; SjS, Sj\u00f6gren's syndrome; UCTD, undifferentiated connective tissue disease.\n\nDiscussion {#s4}\n==========\n\nIn the current paper, we present the ICAP consensus on the clinical relevance of 29 HEp-2 IIFA patterns defined by ICAP.[@R11] The consensus on clinical relevance is defined in the clinical context of the patient, that is, suspected disease, and includes recommended follow-up testing within the spectrum of antigen-specificities that are commercially available. Obviously, if follow-up testing identifies the antigen, the clinical relevance can be further refined.[@R123]\n\nDefining the clinical relevance of HEp-2 IIFA patterns in the context of disease manifestations is meant to be an important tool for the clinician in the diagnostic work-up of patients suspected of SARD. Unfortunately, good data on the association between HEp-2 IIFA patterns and the distinct diseases are lacking, probably due to reasons summarised below. There are several reasons for not finding a perfect association between HEp-2 IIFA patterns and diseases. First, pattern assignment in clinical laboratories is rather inconsistent as shown by external quality assessments.[@R14] This is exactly the reason why ICAP was initiated: the consensus on nomenclature and definitions of HEp-2 IIFA patterns allows to align pattern description across laboratories. Also, the integration of computer-aided immunofluorescence microscopy (CAIFM) may further improve the consistency in pattern assignments.[@R126] As such, it is promising that several companies involved in CAIFM have declared their intention to accommodate to the ICAP classification. Second, even apparently healthy individuals may have autoantibodies as detected by the HEp-2 IIFA. Such autoantibodies, being either innocent bystander antibodies or predictive antibodies, may still be present on development of SARD and interfere with the SARD-related pattern. Interestingly, the pattern best associated with apparently healthy individuals is the nuclear dense fine speckled pattern (AC-2), but this association only holds if the specificity is confirmed as monospecific for DFS70.[@R20] Third, the HEp-2 IIFA patterns may slightly differ depending on the cellular substrate used. For this reason, the ICAP website contains for each pattern multiple pictures taken from different brands of HEp-2 slides. Fourth, diseases like systemic lupus erythematosus and autoimmune inflammatory myopathies may be associated with distinct autoantibodies, each associated with a distinct HEp-2 IIFA pattern. If the autoantigens are ill defined, as is the case, for instance, in autoimmune hepatitis, only the most prevalent patterns are included. Altogether, it is evident that, with the exception of the centromere pattern (AC-3), all patterns are to be confirmed by antigen-specific immunoassay for a solid association with the respective autoimmune diseases.\n\nWhile consensus statements have been generated for all 29 HEp-2 IIFA patterns, and it is highly recommended to report patterns,[@R7] it is anticipated that laboratories may restrict their reports to the so-called 'competent level' patterns (<http://www.ANApatterns.org>).[@R133] Although, for instance, the nucleolar patterns may not be reported as distinct entities (AC-8, AC-9 and AC-10), all three subtypes represent autoantibodies reactive with antigens associated with systemic sclerosis, either alone or in combination with autoimmune inflammatory myopathies. Follow-up testing, therefore, anyhow involves the systemic sclerosis multiparameter assay including all the relevant autoantibodies. Traditionally, only nuclear HEp-2 IIFA patterns have been considered as a true positive HEp-2 IIFA test, and this is most likely related to the time-honoured terminology 'Antinuclear Antibody Test',[@R12] but it is evident from this report that even for nuclear HEp-2 IIFA patterns, the clinical associations are quite diverse. In particular, the nuclear dense fine speckled pattern (AC-2) seems to have an inverse association with SARD.[@R9] On the other hand, the cytoplasmic HEp-2 IIFA patterns, and to a lesser extent the mitotic patterns, are also clinically relevant and may demand dedicated follow-up testing in daily clinical practice. Therefore, the ICAP executive board advocates that information on HEp-2 IIFA patterns should be reported to the clinician and should also be incorporated in diagnostic and classification criteria instead of the simple assignment '*ANA-positive*'.[@R135]\n\nAlthough the HEp-2 IIFA has been considered the gold standard for autoantibody detection in SARD,[@R3] the limitations of this assay are understood.[@R136] Indeed, up to 35% of healthy controls may be positive if a screening dilution of 1/40 is used.[@R139] Therefore, in the EASI/IUIS recommendations, it is advocated that each laboratory verifies that the screening dilution is defined by a cut-off set at the 95th percentile.[@R7] However, by taking into account that the HEp-2 IIFA nowadays is ordered by a wide spectrum of clinical disciplines,[@R1] the number of clinically unexpected positive results, that is, positive test results with no clinical evidence of an associated autoimmune disease, is ever increasing and may even equal the likelihood of a clinically true-positive result.[@R140] A study performed in a community setting concluded that many patients with a positive ANA test are incorrectly given a diagnosis of systemic lupus erythermatosus and sometimes even treated with toxic medications.[@R142] These arguments are used to introduce a gating strategy in order to restrict test-ordering to those cases that have a sufficiently high pretest probability for having a SARD. However, it can also be argued that patients with a low pretest probability should be tested using the HEp-2 IIFA in order to prevent true cases, especially those with very early disease manifestations, from being missed. This is a paradigm shift to disease prediction and prevention.[@R143] In this strategy, the HEp-2 IIFA could be integrated in multianalyte 'omic' profiles for case finding and establishing an early diagnosis and preventing severe complications.[@R143] Obviously, it is anticipated that the added value of the HEp-2 IIFA in this approach can be increased by incorporating information on both patterns as well as titres in combination with well-directed advices on follow-up testing.\n\nAlthough the current consensus on the clinical relevance of HEp-2 IIFA patterns has come across after extensive discussion and debate within the ICAP executive board as well as with the workshop participants, the information provided is not based on a systematic review or meta-analysis of the existing literature. Because of the short history of ICAP, being founded in 2014, inclusion of older literature might have been hampered by potential differences in pattern nomenclature and definitions. For instance, the nuclear dense fine speckled (AC-2) and topo I-like (AC-29) patterns were previously often considered homogeneous, speckled or even mixed patterns. The centromere pattern (AC-3) or the cytoplasmic reticular/AMA (AC-21) patterns, on the other hand, are examples that probably have been less prone to change in pattern definition over time. The universal use of the ICAP nomenclature and pattern definitions, both in daily clinical practice as well as in the scientific literature, may enable systematic reviews in the future, and may well fine-tune current consensus based on expert opinions only.\n\nIn conclusion, the consensus statements on clinical relevance should be readily available to clinicians and this will enable further harmonisation of test-result interpretation with respect to HEp-2 IIFA patterns. Obviously, clinicians should be aware of the clinical suspicion for the respective patient, and therefore should order specific tests accordingly, also taking into account the anticipation of prevalence of HEp-2 IIFA negative (AC-0)[@R13] results in SARD. The information on clinical relevance of HEp-2 IIFA patterns is intended to support the decision strategy of the clinician. Information presented in the [online supplementary tables 1--3](#SP1 SP2 SP3){ref-type=\"supplementary-material\"} is primarily intended to be used for complex cases in the consultation of the laboratory specialist by the clinician. Depending on various jurisdictional regulations, follow-up testing can be automated in predefined algorithms which eventually will shorten the diagnostic delay. Eventually, appropriate integration of HEp-2 IIFA pattern information may help to better define disease criteria and even enable a paradigm shift in the pretest probability paradox.\n\nWe thank all the workshop participants for their constructive comments and fruitful discussions.\n\n**Handling editor:** Josef S Smolen\n\n**Contributors:** All authors actively participated in the respective workshops in Kyoto and Dresden. They also participated in the discussions of the executive ICAP committee. The draft of the manuscript was made by JD and was commented on by all authors. Final discussions have taken place at the international autoimmunity meeting in Lisbon. Required amendments were made by JD and approved by all authors.\n\n**Funding:** The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.\n\n**Competing interests:** The ICAP committee is funded by unrestricted educational grants by several in vitro diagnostics companies (for details see [www.anapatterns.org/sponsors.php](http://www.anapatterns.org/sponsors.php)). JD has received lecture fees from Euroimmun and Thermo Fisher. MJF is a consultant to Inova Diagnostics and Werfen International; none of the other authors declare any competing interest.\n\n**Patient consent for publication:** Not required.\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n\n**Data sharing statement:** No additional data are available.\n"}
+{"text": "Background\n==========\n\nMale infertility accounts for about half of all infertility cases and may arise from a variety of factors. One known cause is the retention of excess cytoplasm around the midpiece due to an arrest in spermiogenesis and incomplete cytoplasmic extrusion \\[[@B1]\\]. This is now known as excess residual cytoplasm (ERC). In comparison to the typical cytoplasmic droplet (CD) found in ejaculated human spermatozoa, ERC contains elevated levels of cytoplasm enzymes that produce pathological amounts of reactive oxygen species (ROS) \\[[@B2]\\]. The high ROS levels may then result in oxidative stress (OS). ERC ultimately affects sperm motility \\[[@B3]\\], morphology \\[[@B2]\\] and fertilization potential \\[[@B4]\\], thereby leading to male infertility. The purpose of this review is to compare CDs to ERC, to describe ERC's relevance to human reproduction, and to clarify its assessment and clinical importance.\n\nReview\n======\n\nThe cytoplasmic droplet\n-----------------------\n\nRetzius first identified the normal CD in 1909, the significance of which has eluded scientists ever since \\[[@B5]\\]. It is a familiar element of mammalian spermatozoa, though many studies in the past have focused primarily on CDs in non-human species. The differences between human and non-human mammals became apparent once researchers drew attention to human CDs. Unlike those of domestic species (i.e. boars, rams, goats, etc.), CDs of normal human spermatozoa are still present after ejaculation \\[[@B6]\\]. For this reason, CDs are not considered detrimental to proper sperm function.\n\nStructure\n---------\n\nMammalian CDs are surrounded by a cell membrane and contain cytoplasm that houses the cytosol and cytoskeletal network \\[[@B7]\\]. Hermo and coworkers \\[[@B8]\\] conducted studies on rat CDs, and in these species, the droplet comprises of lamellae and small vesicles unlike structural components of the Golgi apparatus and endoplasmic reticulum. The CDs of mature non-human mammalian sperm are found at the distal end of the midpiece. Mature human spermatozoa are similar to that of other mammals in possessing a CD at the midpiece. However, the human CD is more proximal, located at the neck as opposed to the end of the annulus \\[[@B9]\\] (Figure [1A](#F1){ref-type=\"fig\"}). Due to the mitochondrial helix that forms around the core of the axoneme-outer dense fibers complex, the midpiece has a large diameter relative to the rest of the cell \\[[@B10]\\].\n\n![**CD vs. ERC Structure.** Illustration comparing (**A**) spermatozoa with typical cytoplasmic droplets (CD) and (**B**) spermatozoa presenting with excess residual cytoplasm (ERC). Also listed are specific causes of spermiogenesis arrest that can lead to ERC.](1477-7827-10-92-1){#F1}\n\nThe mammalian CD is approximately 2 \u03bcm in diameter \\[[@B11]\\] and is comprised of lipids, lipoproteins, RNAs, and a variety of hydrolytic enzymes \\[[@B12]\\]. It also contains ion channels and golgi-derived vesicles \\[[@B10]\\].\n\nThe midpieces of human spermatozoa contain osmotically sensitive \"midpiece vesicles\" (MPVs) as confirmed by Cooper and colleagues \\[[@B9]\\]. MPVs were previously considered distinct from CDs and characteristic of immature sperm \\[[@B13]\\]. However, Cooper and colleagues suggested that MPVs and CDs are one and the same as neither withstands air-drying well.\n\nManifestation\n-------------\n\nAs a result of spermatogenesis and epididymal maturation, male germ cells differentiate to become fully functional spermatozoa. Spermiogenesis, the final stage of spermatogenesis, transforms haploid spermatids to free testicular (non-motile) spermatozoa \\[[@B14]\\]. It is during the last (\"maturation\") phase of spermiogenesis that the CD is formed \\[[@B15]\\].\n\nThis is accomplished through the actions of Sertoli cells by a process known as \"cytoplasmic extrusion\" \\[[@B16]\\], which occurs before sperm are transported to the epididymis \\[[@B17]\\]. Cytoplasmic extrusion, along with various other maturation steps, is essential for allowing zona binding capacity and fertilization potential of spermatozoa \\[[@B16]\\].\n\nIn the tubular lumen of the testis, Sertoli cells extrude and phagocytose most of the germ cell cytoplasm as \"residual bodies\", the remnant of which becomes the CD \\[[@B17]\\]. In most non-human mammalian species, the droplet migrates down the tail and is finally shed around the time of ejaculation \\[[@B6]\\]. However, the exact mechanism of this movement has not yet been clarified \\[[@B5]\\]. Retention of CDs on these non-human ejaculated spermatozoa has been associated with infertility \\[[@B18],[@B19]\\]. Human spermatozoa, however, normally retain a small amount of the droplet around the midpiece after spermiogenesis.\n\nPhysiology\n----------\n\nBecause they possess only a small amount of cytoplasm, spermatozoa lack the organelles required for osmoregulation \\[[@B20]\\]. Therefore, spermatozoa rely on their extracellular environment for osmoregulation. The midpiece is an ideal location for the CD; this is the major site of water influx and cell volume regulation, which is important when spermatozoa face hypo-osmotic challenges (e.g. cervical mucus osmolality) \\[[@B21]\\].\n\nRegulatory volume decrease (RVD) has been shown to occur in spermatozoa under hypotonic conditions \\[[@B22]\\], which compensates for any swelling during and after ejaculation \\[[@B6]\\]. This occurs because CDs contain osmolyte channels that facilitate osmosis. K^+^ and Cl^-^ channels have also been found at the neck and along the midpiece \\[[@B22]\\].\n\nThe epididymis is responsible for osmolyte loading via regulatory volume increase (RVI) \\[[@B12]\\]. The amount of osmolytes present is vital for proper sperm function; if adequate, motility will be \"forward progressive\" for successful transit through the female reproductive tract \\[[@B6]\\]. In the case of low osmolyte loading or presence of cryoprotectants or an osmolyte channel blocker (e.g. quinine), spermatozoa swell and forward progression is thwarted \\[[@B23],[@B24]\\].\n\nDroplets in non-human mammalian spermatozoa can induce swelling and flagellar angulation \\[[@B25]\\], which inhibit progressive motility and are associated with infertility. Human spermatozoa, on the other hand, retain their CDs with no negative effects, because there is no angulation or coiling effect on the flagellum under naturally hypotonic conditions \\[[@B6]\\].\n\nChen and coworkers \\[[@B26]\\] have reported that aquaporin 3 (AQP3) is a water channel in human spermatozoa necessary for RVD. Their research suggests that AQP3 balances reduction in sperm motility due to any cell swelling that may occur.\n\nThe CD is home to a variety of enzymes and receptors. K\u00f6hn and colleagues \\[[@B27]\\] identified angiotensin-converting enzyme (ACE) to be present in the CD, a membrane-bound enzyme that is released during capacitation. In this study, it was shown that less ACE is present among more motile spermatozoa, signifying a negative correlation between ACE level and sperm maturity.\n\n15-lipoxygenase (15-LOX) and components of the ubiquitin-dependent proteolytic pathway are prominent in the CDs of mammalian spermatozoa \\[[@B28]\\]. 15-LOX may be responsible for the removal of CDs from spermatozoa and are thought to participate in epididymal sperm maturation and formation of the midpiece and mitochondrial sheath. The components of the ubiquitin-proteasome pathway are believed to assist in spermiogenesis and organelle degradation. Though this study was performed on boar spermatozoa, 15-LOX and ubiquitin components were detected in human spermatozoa as well.\n\nCalreticulin (CRT) and the inositol 1,4,5-trisphosphate receptor (IP3R) have also been found within vesicles of the CD \\[[@B29]\\]. Both are implicated in calcium level oscillations during hyperactivation and the acrosome reaction. Ropporin is yet another protein found in both the CD and flagellum that is suggested to have a role in regulating sperm motility and the acrosome reaction by binding to the amphipathic helix region of A-kinase anchoring proteins (AKAPs) \\[[@B30]\\].\n\nMiranda-Vizuete and colleagues \\[[@B31]\\] have reported the localization of sperm-specific thioredoxin (Sptrx) in the CD of human spermatozoa. Sptrx was shown to behave as a reductant, possibly to correct wrong disulfide pairings during sperm tail formation. Failure in Sptrx expression has been linked to dysplasia of the fibrous sheath (DFS) \\[[@B31]\\], or \"stump tail syndrome\" \\[[@B32]\\], which can cause severe asthenozoospermia or sperm immotility.\n\nAmong the assortment of CD enzymes are creatine kinase (CK), lactic acid dehydrogenase (LDH), superoxide dismutase (SOD), and glucose-6-phosphate dehydrogenase (G6PDH) \\[[@B15]\\], all of which are involved in the energy metabolism of the CD. Defective CDs harbor higher levels of these metabolic enzymes, which can impair overall sperm function.\n\nThe World Health Organization (WHO) considers CDs to be defects when they are larger than one-third of the sperm head size \\[[@B33]\\] (Table [1](#T1){ref-type=\"table\"}). This ERC arises due to the premature arrest of spermiogenesis \\[[@B9]\\]. Unlike most mammalian species, human spermatozoa are unable to modify any residual cytoplasm that may exist during epididymal maturation or at ejaculation \\[[@B34]\\]. Compared to the typical CD, ERC possesses a surplus of the aforementioned enzymes, which can negatively affect sperm function and lead to male infertility (Figure [2](#F2){ref-type=\"fig\"}).\n\n###### \n\nComplete CD and ERC comparison\n\n  \u00a0                                  **Cytoplasmic droplet**                                                         **Excess residual cytoplasm**\n  ---------------------------------- ------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  *Location*                         Proximal part (neck) of midpiece                                                Along midpiece\n  *Structure*                        Cytoplasm smaller than 1/3 sperm head size, 2 \u03bcm in diameter                    Cytoplasm larger than 1/3 to 1/2 sperm head size\n  *Caused by*                        Cytoplasmic extrusion via Sertoli cell phagocytosis                             Spermiogenesis arrest and interruption of cytoplasmic extrusion\n  *Function/Consequences*            Physiological -- Regulatory volume decrease and ROS production                  Physiological -- Elevated levels of cytoplasm enzymes; Pathological -- Incomplete maturation (\"dysmature\"), oxidative stress, lipid peroxidation, and apoptosis\n  *Identification/Testing Methods*   Air-drying (does not survive) and binary image analysis via NADH/NBT staining   Air-drying (survives), binary image analysis via NADH/NBT staining, immunofluorescence, immunoblotting, ROS markers, aniline blue chromatin staining, and Sptrx screening\n\nDifferences between cytoplasmic droplets (CDs) and excess residual cytoplasm (ERC) with regards to location, structure, cause, function/consequences, and identification/testing methods, and hallmarks.\n\n![**Problems Associated with ERC.** Illustration detailing the pathological effects of excess residual cytoplasm (ERC), including peroxidative damage to the sperm membrane, DNA damage, mitochondrial dysfunction, and impaired sperm function within the female reproductive tract.](1477-7827-10-92-2){#F2}\n\nExcess residual cytoplasm\n-------------------------\n\nUntil recently, the concept of ERC was not as widely accepted throughout the scientific community. Instead, a variety of words were used indiscriminately with regards to both normal and abnormal cytoplasmic residue \\[[@B5]\\]. This discrepancy was settled in a study conducted by Cooper and colleagues \\[[@B9]\\] that helped bring this distinction to the forefront. Their research demonstrated a major difference; ERC survives the air-drying techniques used for human seminal smears whereas CDs do not. ERC may span the entire midpiece area although the human CD is found at the neck of the midpiece (Table [1](#T1){ref-type=\"table\"}).\n\nSince ERC-bearing spermatozoa are unable to complete maturation \\[[@B5]\\]**,** some may consider sperm bearing ERC as \"immature\". However, the term \"dysmature\" is deemed more appropriate \\[[@B35]\\]. \"Immature\" is mostly used to describe normal spermatozoa that have yet to undergo epididymal maturation after release from the testis. \"Dysmature\", on the other hand, describes spermatozoa that have experienced an interruption or cessation of spermiogenesis and/or epididymal maturation.\n\nManifestation\n-------------\n\nSpermatozoa retain excess cytoplasm due to a disruption of spermiogenesis as evidenced by high levels of cytoplasmic enzymes \\[[@B2]\\] (Table [1](#T1){ref-type=\"table\"}). This process has yet to be detailed, but there are various explanations as to how this may occur.\n\nFor example, suppression of FSH and/or androgens (primary hormone regulators of spermatogenesis) can cause spermiogenesis arrest \\[[@B36]\\] (Figure [1B](#F1){ref-type=\"fig\"}). D'Souza and coworkers \\[[@B37]\\] demonstrated that administration of exogenous 17-beta-estradiol led to reduced FSH and intratesticular testosterone levels, thereby leading to an absence of tubulobulbar complexes.\n\nThe apically located tubulobulbar complex is an actin-based structure that anchors late spermatids to the Sertoli cell and indirectly produces the residual body during cytoplasmic extrusion \\[[@B38]\\]. During formation of the tubulobulbar complex, there is a considerable loss of cytoplasm in the spermatid, implicating these complexes in Sertoli cell phagocytosis \\[[@B39]\\]. Disruption of tubulobulbar complex formation is therefore associated with retention of excess cytoplasm around the midpiece.\n\nCyclin-dependent kinase 16 (CDK16), which is highly expressed in the brain and testis, has also been associated with spermiogenesis. Mikolcevic and coworkers \\[[@B40]\\] determined that CDK16 deficiency in mice is correlated with morphologically defective spermatozoa possessing malformed heads and ERC (Figure [1B](#F1){ref-type=\"fig\"}). The precise role of CDK16 in spermatozoa maturation is still unclear.\n\nRecently, there have been reports associating organophosphorus pesticide (OP) exposure and reduced sperm function in people and laboratory animals. OPs, such as dichlorvos (DDVP), are used as pesticides to shield crops and homes from insect attack. A study conducted by Okamura and colleagues \\[[@B41]\\] revealed a higher incidence of cytoplasm in rat spermatozoa with DDVP exposure (Figure [1B](#F1){ref-type=\"fig\"}). Malathion, another OP, has been shown to increase cytoplasm in rat spermatozoa. Researchers suggest that such pesticides trigger an early arrest of spermatozoa maturation, which may be alleviated by reducing OP exposure.\n\nStudies have suggested the association of ERC with varicocele presence \\[[@B42]\\]. They each cause an increase in ROS production, and both idiopathic and varicocele-related male infertility have been correlated with impaired cytoplasmic extrusion \\[[@B43]\\]. Infertile men with varicocele were shown to have the highest percentage of sperm presenting with ERC. The mechanism by which varicocele and ERC are related has yet to be elucidated.\n\nSmoking has also been shown to impair cytoplasmic extrusion and sperm function \\[[@B44]\\]. The exact pathophysiology is uncertain, but smoking may have effects on Sertoli and/or Leydig cell function \\[[@B45]\\] and/or oxidative balance in the testis \\[[@B46]\\]. Since smoking may be associated with a lifestyle of other poor health decisions, such as alcohol/drug abuse and suboptimal diet, confounding factors may further affect a couple's fertility \\[[@B44]\\].\n\nPathology\n---------\n\nERC has many health implications spanning a wide range of disorders (Table [1](#T1){ref-type=\"table\"}). Problems primarily arise due to the elevated levels of key enzymes found within the cytoplasm itself (Figure [2](#F2){ref-type=\"fig\"}).\n\nNormally, spermatozoa produce a low level of ROS from their mitochondria \\[[@B35]\\]. Physiological levels of ROS trigger and modulate tyrosine phosphorylation, for instance, which elicits vital functions like capacitation and the acrosome reaction \\[[@B34]\\].\n\nWhen ROS levels are raised, as is the case with ERC, spermatozoa are limited in their ability to eradicate the surplus. This is typically due to the presence of extra electron transport chains in the plasma membrane, unfamiliar oxidases, or oxidoreductases that promote xenobiotic production \\[[@B34]\\]. Physiological levels of antioxidants, in its limited availability, cannot counteract this excess production of ROS. In fact, electron leakage from the sperm mitochondrial electron transport chain is considered to be a major source of ROS generation in defective sperm \\[[@B47]\\]. Sperm motility has been shown to decrease as mitochondrial production of ROS increases, subsequently inducing DNA damage.\n\nAs a result of high ROS levels, spermatozoa may assume a state of oxidative stress, characterized by damage to both mitochondrial and nuclear DNA \\[[@B48]\\] along with peroxidative damage to the sperm plasma membrane \\[[@B49]\\]. It is presumed that peroxidative damage is associated with higher activities of CK and G6PDH \\[[@B50]\\]. The ROS not only affects the abnormal spermatozoa that generates it, but it can also damage normal spermatozoa as well \\[[@B51]\\].\n\nIn terms of energy metabolism, ERC contains higher levels of G6PDH, resulting in greater NADPH production. By means of the hexose monophosphate shunt, G6PDH produces NADPH and controls the levels of glucose flux and NADPH availability \\[[@B15]\\]. Spermatozoa use this shunt for its supply of electrons for ROS generation. NADPH is a substrate for ROS-generating NADPH oxidases \\[[@B15]\\]. The superoxide anion, O~2~\u2009**\u00b7**\u2009^-^, is one common ROS that is converted to H~2~O~2~ via SOD, another enzyme commonly found in ERC \\[[@B52]\\]. H~2~O~2~ molecules, if left to accumulate, may then undergo homolytic cleavage to form two hydroxyl free radicals (OH\u2009**\u00b7**\u2009^-^), highly reactive electrophiles that cause oxidative damage to cells \\[[@B53]\\] (Figure [2](#F2){ref-type=\"fig\"}). Due to their high phospholipid content and relatively low cytoplasmic volume, spermatozoa are especially susceptible to this condition \\[[@B34]\\]. LDH, also involved in the maintenance of spermatozoa energy metabolism, has not been shown to be directly injurious to the cells in higher levels \\[[@B54]\\].\n\nPrevious studies have indicated that sperm mitochondria release cytochrome C in response to ROS stimulation \\[[@B55]\\], which activates a signaling cascade involving caspase 3 and 9 that ultimately leads to sperm apoptosis. Elevated ROS levels have been associated with increased instances of apoptosis \\[[@B56]\\]. Because caspase 3 activation is localized to the midpiece, Weng and coworkers \\[[@B57]\\] have suggested that apoptotic mechanisms may stem from the midpiece cytoplasm and then operate in the nucleus.\n\nUnder physiological conditions, normal ROS levels have been correlated with reduced DNA damage via chromatin cross-linking \\[[@B58]\\]. When ROS levels increase, the subsequent oxidative stress induces DNA damage \\[[@B59]\\]. DNA repair in spermatozoa and the ability of sperm to undergo apoptosis deteriorates during late spermatogenesis \\[[@B15]\\]. Both mitochondrial and nuclear DNA can be affected, the former especially susceptible due to insufficient protection against ROS attack \\[[@B35]\\]. This may not pose as great a risk since sperm mitochondria are discarded in the oocyte following fertilization \\[[@B60]\\]. However, any damaged nuclear DNA will be incorporated into the zygote. Even if peroxidative membrane damage occurs, damaged DNA can still be transferred into the embryo \\[[@B61]\\]. Therefore, the oocyte is responsible for correcting any DNA damage or inducing apoptosis (i.e. embryonic loss) before the first cleavage. Any lingering errors following these preventative measures may affect the eventual development and overall health of the offspring.\n\nERC can also affect how a spermatozoon functions within the female reproductive system. Oxidative stress caused by ERC can impair sperm motility and function, affecting capacitation and fertilization \\[[@B62]\\]. Furthermore, it may render the sperm plasma membrane unable to respond to intracellular calcium signals, impeding sperm-oocyte fusion \\[[@B63]\\]. This may involve reduced plasma membrane fluidity \\[[@B64]\\] in addition to changes to membrane-bound enzyme activity (e.g. ion channels) \\[[@B65]\\]. Defective enzyme expression may affect fertilization in that an abundance of 15-LOX in semen can cause early acrosomal exocytosis \\[[@B66]\\], possibly hampering zona binding and penetration \\[[@B67]\\]. Oxidative stress in human spermatozoa impairs both fertilization potential \\[[@B68]\\] and embryonic/fetal development \\[[@B69]\\]. It can lead to higher instances of miscarriage \\[[@B70]\\], and offspring morbidity, including childhood cancer \\[[@B71]\\].\n\nIdentification\n--------------\n\nThere are a variety of known methods for ERC analysis (Table [1](#T1){ref-type=\"table\"}). Microscopy can be used to estimate the size of the cytoplasm around a sperm head (Figure [3](#F3){ref-type=\"fig\"}). CDs larger than one-third of the sperm head size are classified as ERC \\[[@B33]\\].\n\n![**CD and ERC under Microscope.** True examples of cytoplasmic droplets (**A**, **B**) and excess residual cytoplasm (**C**, **D**) in human spermatozoa as examined by microscopy. This image was modified from Cooper and colleagues \\[[@B9]\\] with permission, \u00a9 Oxford Journals.](1477-7827-10-92-3){#F3}\n\nAn image analysis for ERC has been developed via midpiece staining to generate binary images \\[[@B2]\\]. The staining, which renders the entire midpiece blue-black, uses NADH and nitroblue tetrazolium (NBT) as an electron donor and acceptor, respectively. Once the binary images are created, morphometric analysis is performed in order to gauge the extent of excess cytoplasm retention. Though this technique is more time-consuming than biochemical assays, it provides information on a cellular level that may be used as selection criteria for intracytoplasmic sperm injection (ICSI) during assisted reproduction.\n\nRago and colleagues \\[[@B1]\\] have demonstrated the presence of two estrogen receptors within ERC, ER-alpha and ER-beta. The presence of the ER-beta form in abnormal sperm and its absence in normal spermatozoa suggest a role for estrogens in sperm maturation. These receptors can be monitored using immunofluorescence and immunoblotting techniques.\n\nThe defenselessness that sperm mitochondrial DNA exhibits under ROS attack makes it a sensitive marker for examining oxidative stress \\[[@B35]\\] as the DNA damage can be accessed through the TUNEL assay, for example. To assess the extent of DNA damage, further screening (e.g. aniline blue chromatin staining, which tests for histone integrity) should be performed \\[[@B72]\\].\n\nScreening of Sptrx status in ERC-bearing spermatozoa may provide useful information to clinicians, in particular, on sperm tail pathologies \\[[@B31]\\]. Therefore, Sptrx screening may facilitate the diagnosis of patients affected with DFS.\n\nInterventions and treatments\n----------------------------\n\nCK, which is sperm-specific and expressed in the testis, is a marker of cell maturity. Expression of the CK-M isoform, also known as HspA2 \\[[@B73]\\], is associated with changes in the sperm plasma membrane as well as cytoplasmic extrusion \\[[@B16]\\]. CK levels can be tested with immunochemistry while CK-M/HspA2 antiserum is used to indicate maturity by highlighting the ERC \\[[@B74]\\].\n\nIn order to manage oxidative stress, a variety of antioxidant treatments exist to reduce ROS levels by becoming radicals themselves \\[[@B69]\\]. This therapy could be used on patients presenting with a high OS status \\[[@B72]\\]. Glutathione and selenoproteins are example of such antioxidants. alpha-Tocopherol, a form of vitamin E \\[[@B75]\\], and vitamin C may be used together *in vivo* to prevent peroxidative damage \\[[@B76]\\]. A cofactor of the pyruvate dehydrogenase complex, alpha-lipoic acid, has also been shown to significantly reduce reactive oxygen metabolites *in vivo*\\[[@B77]\\]. Ascorbate and catalase have been shown to reduce ROS (specifically H~2~O~2~) *in vitro* as well \\[[@B78]\\].\n\nBecause ERC and subsequent peroxidative damage to the sperm plasma membrane can compromise sperm-oocyte fusion, ICSI can be used as an intervention to bypass this obstacle. Nevertheless, other defects may still persist, such as nuclear DNA damage, preventing proper embryonic/fetal development \\[[@B71]\\].\n\nMoreover, varicocelectomy has been shown to improve cytoplasmic extrusion and reduce excess ROS production in men with ERC \\[[@B42]\\]. This topic merits further investigation to assess the efficacy of this and other treatments.\n\nConclusions\n===========\n\nThe CD is a morphological feature unique to human spermatozoa and absent in other non-human mammals. It has functional significance and is involved in hyperactivation, capacitation, and the acrosome reaction. An abnormally large amount of cytoplasm around the sperm midpiece has pathological implications and is then termed \"ERC\" based on size and function. Having distinguished CDs from ERC, future studies should attempt to measure the incidence rate of ERC in human males and explain why human spermatozoa tend to retain CDs after maturation. Additionally, the pathophysiology of ERC must be described in greater detail with regards to male infertility. Furthermore, prospective studies should evaluate the efficacy of the various treatments for ERC. There is a lot to be learned about this particular morphological defect. Clarification is vital for better assessing male infertility in a clinical setting.\n\nAbbreviations\n=============\n\nERC: Excess residual cytoplasm; CD: Cytoplasmic droplet; ROS: Reactive oxygen species; OS: Oxidative stress; MPV: Midpiece vesicle; RVD: Regulatory volume decrease; RVI: Regulatory volume increase; AQP3: Aquaporin 3; ACE: Angiotensin-converting enzyme; 15-LOX: 15-lipoxygenase; CRT: Calreticulin; IP3R: ; AKAP: A-kinase anchoring protein; : Inositol 1,4,5-trisphosphate receptor; Sptrx: Sperm-specific thioredoxin; DFS: Dysplasia of the fibrous sheath; CK: Creatine kinase; LDH: Lactate acid dehydrogenase; SOD: Superoxide dismutase; G6PDH: Glucose-6-phosphate dehydrogenase; WHO: World health organization; CDK16: Cyclin-dependent kinase 16; OP: Organophosphorus pesticide; DDVP: Dichlorvos; NBT: Nitroblue tetrazolium; ICSI: Intracytoplasmic sperm injection.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nAKR studied this topic and drafted this manuscript. AA served as a mentor and reviewer of this article. MVDL revised this manuscript. SSDP conceived of the review, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.\n"}
+{"text": "![](indmedgaz71842-0061){#sp1 .222}\n\n![](indmedgaz71842-0062){#sp2 .223}\n\n![](indmedgaz71842-0063){#sp3 .224}\n\n![](indmedgaz71842-0064){#sp4 .225}\n\n![](indmedgaz71842-0065){#sp5 .226}\n\n![](indmedgaz71842-0066){#sp6 .227}\n\n![](indmedgaz71842-0067){#sp7 .228}\n"}
+{"text": "Introduction {#s1}\n============\n\n*Campylobacter* continues to be the most commonly reported cause of bacterial gastroenteritis in the European Union (EU). In total there were 198,252 confirmed cases of campylobacteriosis in 2009, giving an overall crude incidence rate (CIR) of 45.6 per 100,000 population [@pone.0028490-EFSA1].\n\nWhile a range of risk factors for infection with *Campylobacter* have been identified, the most common is the handling and/or consumption of undercooked poultry, in particular chicken. According to a recent opinion of the European Food Safety Authority\\'s (EFSA) Biological Hazards (BIOHAZ) Panel, 50 to 80% of human cases of campylobacteriosis may be attributed to the chicken reservoir [@pone.0028490-EFSA2]. This report also outlined the considerable underreporting of cases of campylobacteriosis and suggested that no less than 2 million and up to 20 million cases of clinical campylobacteriosis occur per year in the EU.\n\nIn 2008, a baseline survey on *Campylobacter* in broiler batches and carcasses in the EU was undertaken [@pone.0028490-EFSA3]. The prevalence at community level of *Campylobacter*-colonized broiler batches was reported to be 71.2% and the prevalence of *Campylobacter*-contaminated broiler carcasses was found to be 75.8%. The colonization of broiler flocks with *Campylobacter* is therefore a significant food safety issue and a reduction in the number of contaminated poultry products entering the food chain would reduce the negative impact on public health.\n\nIt is thought that the reduction of *Campylobacter*-contaminated poultry meat can be achieved most effectively by implementing on-farm control measures [@pone.0028490-ACMSF1].\n\nExtensive research into the most important source of *Campylobacter* in poultry production units has been carried out. Vertical transmission has previously been implicated [@pone.0028490-Buhr1], [@pone.0028490-Buhr2], [@pone.0028490-Cox1], [@pone.0028490-Pearson1]. Carryover to subsequent flocks as a result of inadequate disinfection of broiler houses has also been identified as a risk factor [@pone.0028490-Petersen1]. Horizontal transmission from the surrounding environment to the broiler house either *via* the farm workers or other vectors such as wild birds, vermin and house flies, is considered to be the most likely source of contamination [@pone.0028490-Newell1], [@pone.0028490-Bull1]. Risk factors for infection may vary from country to country because of differing farming practices and associated climatic conditions. Therefore, efforts to understand the relative importance of each potential source and transmission route of *Campylobacter* infection on-farm, continues to have an important role in extending our understanding of the epidemiology of this important pathogen.\n\nIn 2009, there were 1,808 confirmed cases of campylobacteriosis in Ireland [@pone.0028490-HPSC1]. The prevalence of *Campylobacter*-colonized broiler batches in Ireland in 2008 was 83.1% and of that *Campylobacter*-contaminated broiler carcasses accounted for 98.3% [@pone.0028490-EFSA3]. In order to implement effective control measures and reduce the prevalence of *Campylobacter* in Irish poultry products, the most significant transmission routes must be identified. The aim of this study was to investigate the occurrence of *Campylobacter* in a subset of intensively reared Irish poultry flocks and to identify sources of *Campylobacter* in each farm environment. Molecular sub-typing methods were used to identify the genotypes present in Irish broiler farms and to shed further light on possible transmission routes of *Campylobacter* in poultry farms.\n\nMaterials and Methods {#s2}\n=====================\n\nDescription of Farms in this Study {#s2a}\n----------------------------------\n\nThree housed broiler flocks (denoted as broiler flocks 1, 2 and 3), located on three different farms (denoted as broiler farms 1, 2 and 3) in different geographical locations of Ireland, were studied throughout their 6--7 week life span. The flocks were screened between November 2006 and September 2007. Farm 1 consisted of three poultry house units sampled from November 2006 to January 2007. Farm 2 consisted of two poultry house units that were sampled between April 2007 and May 2007. Farm 3 consisted of one poultry house unit and was sampled between July 2007 and September 2007. Broiler flock sizes ranged from approximately 18,000 to 34,000 birds per house. Three breeder flocks, located on breeder farm 1 (denoted as breeder farm 1, flocks 1--3), supplied chicks for broiler farm 1. Broiler farm 2 was supplied by breeder farms 2 and 3. Broiler farm 3 was supplied by two flocks from breeder farm 1 (denoted as breeder farm 1, flocks 4 and 5), and one flock from both breeder farms 4 and 5. Each breeder flock comprised approximately 5,000 birds per house.\n\nSample Collection {#s2b}\n-----------------\n\nSamples were collected approximately every 14 days from (i) the cleaned and disinfected broiler houses prior to chick placement, (ii) the chickens and (iii) the environments inside and outside the broiler houses. The breeder flocks supplying each broiler house were also tested. Samples were transported to the laboratory at 4\u00b0C in a cool box and processed on the same day.\n\n### Breeder flocks {#s2b1}\n\nSixty fresh faecal samples (5 pooled faecal samples each containing 12 fresh faecal droppings) were collected from the floor of the house of the corresponding breeder farms supplying broiler hatching eggs for each flock in this study.\n\n### Broiler flocks {#s2b2}\n\nOne-day old broiler chicks were tested by enriching the faeces-soiled paper that lined the crates used to transport the birds from the hatchery to the broiler house (100 birds per crate). Faecal samples were then taken at regular intervals throughout the rearing period of each flock (at days 14, 28 and 42 approximately). Sixty individual fresh faecal samples were collected from the broiler house floor and combined (as outlined previously).\n\n### Environmental samples {#s2b3}\n\nSamples from walls, floors, structural support columns, feed and water dispensers along with concrete aprons (the concreted area outside the front of the house) were taken using sterile swabs pre-moistened with 10 ml Maximum Recovery Diluent (MRD, Lab M Ltd., Bury, UK). An area of 0.1 m^2^ of the object\\'s surface was chosen for sampling, and swabbing continued outside this area until either the entire surface was sampled or the swab was dry. Two litres of water supplying the broiler house drinkers were also taken. Aspiration of air samples onto one *Campylobacter* blood-free selective agar base (CCDA, CM0739, Oxoid, Cambridge, UK), supplemented with cefoperazone and amphotericin B (CCDA selective supplement) (Oxoid), and one Tryptone Soya Agar (TSA, Oxoid Ltd., Cambridge, UK) plate was performed using a Sampl\\'air MK2 double agar plate sampler (AES Laboratoire Groupe, Combourg, France). Two air samples (500 l\u00d72) were taken from inside and one directly outside each broiler house on each sampling occasion (at days -1, 14, 28 and 42). Occasionally, samples were taken opportunistically inside and outside the broiler house environment ([Table 1](#pone-0028490-t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0028490.t001\n\n###### *Campylobacter* sources during the sampling period across farms.\n\n![](pone.0028490.t001){#pone-0028490-t001-1}\n\n  Flock          Sample         Number of positive samples/number of samples tested                                                            \n  ------- -------------------- ----------------------------------------------------- --------- ----------------------- ----------------------- ---------------------------\n  **1**         Broilers                                \\-                            0/48^c^        **3/5** ^c^             **5/5** ^c^               **5/5** ^c^\n           Adjacent Broiler 1                           \\-                              \\-             0/2^c^                **2/2** ^c^               **2/2** ^c^\n           Adjacent Broiler 2                           \\-                              \\-           **2/2** ^c^             **2/2** ^c^                   \\-\n                  Air                                   0/6                             \\-             **1/6**                   0/6                       0/6\n                 Puddle                                 0/1                             \\-             **1/2**                   0/1                       0/2\n                  Soil                                  0/2                             \\-               0/2                     0/2                     **1/2**\n                 Other           0/20^ap,\\ cl\\ co,\\ d,\\ f,\\ fe,\\ fl,\\ h,\\ w,\\ wa^       \\-      0/7^ap,\\ h,\\ i,\\ wa^    0/9^ap,\\ h,\\ i,\\ wa^      0/11^ap,\\ h,\\ i,\\ wa^\n  **2**         Broilers                                \\-                            0/48^c^          0/5^c^                  0/5^c^                  **5/5** ^c^\n            Adjacent Broiler                            \\-                              \\-             0/2^c^                  0/2^c^                      \\-\n                  Soil                                  0/2                             \\-               0/2                   **1/2**                     0/2\n                 Other             0/22^a,\\ ap,\\ co,\\ d,\\ fe,\\ fl,\\ p,\\ w,\\ wa^         \\-        0/10^a,\\ ap,\\ wa^     0/13^a,\\ ap,\\ i,\\ wa^   0/12^a,\\ ap,\\ p,\\ r,\\ wa^\n  **3**         Broilers                                \\-                            0/48^c^          0/5^c^                **5/5** ^c^               **5/5** ^c^\n                  Soil                                  0/2                             \\-               0/2                     0/2                     **1/2**\n                 Other           0/24^a,\\ ap,\\ cl\\ co,\\ d,\\ fe,\\ fl,\\ p,\\ w,\\ wa^       \\-      0/12^a,\\ ap,\\ p,\\ wa^   0/12^a,\\ ap,\\ p,\\ wa^       0/10^a,\\ ap,\\ wa^\n\nBold type indicates positive result; -, not tested; \\*, approximate day of sampling; ^a^, air; ^ap^, apron; ^cl^, clothing; ^co^, support columns; ^d^, drinkers; ^f^; fan, ^fe^; feeder, ^fl^; floor, ^h^; horse faeces; ^i^, insects; ^p^, puddles; ^r^, rodent faeces; ^w^, walls; ^wa^, water; ^c^, composite samples.\n\nIsolation of *Campylobacter* {#s2c}\n----------------------------\n\nTo determine the presence/absence of *Campylobacter*, samples were examined by direct plating and/or enrichment culture methodologies based on the Horizontal Method for Detection and Enumeration of *Campylobacter* spp. (ISO 10272-1:2006) and the Detection and Semi-quantitative Enumeration of Thermotolerant *Campylobacter* spp. (ISO 17995:2005). Isolation of emerging *Campylobacter* species, based on the method previously described [@pone.0028490-Lynch1], was also performed for faecal samples. Following incubation, five suspect colonies were randomly selected from plates and subcultured to obtain pure colonies.\n\nFor the lined transport crates, 60 papers from each flock were collected at random and divided into six piles consisting of ten papers on top of one another. Each pile of ten papers was then aseptically cut into 8 strips. Each strip was enriched in 200 ml *Campylobacter* Enrichment Broth (CEB, Lab M Ltd., Bury, UK) supplemented with 5% (v/v) lysed horse blood (TCS Biosciences, Buckingham, UK) and cefoperazone, vancomycin, trimethoprim and cyclohexamide (CVTC supplement, Lab M Ltd., Bury, UK). Forty-eight composite samples, each consisting of ten strips, from each flock were tested in this way. Swabs were enriched in 100 ml of CEB supplemented with 5% (v/v) lysed horse blood and CVTC. Water samples were filtered using 0.45 \u00b5m filters (Millipore, Billerica, MA., USA), which were then enriched in sterile 30 ml containers containing 20 ml of CEB supplemented with 5% (v/v) lysed horse blood and CVTC. For the air samples, the agar from each TSA plate was aseptically removed and enriched in 100 ml CEB supplemented with 5% (v/v) lysed horse blood and CVTC, while each CCDA plate was incubated directly. Faecal samples from horses, crushed flies and beetles were enriched with CEB supplemented with 5% (v/v) lysed horse blood and CVTC using a 1\u223610 ratio of sample to broth. Water from puddles was enriched with an equivalent volume of double-strength CEB supplemented with 5% (v/v) lysed horse blood and CVTC.\n\nBiochemical confirmation {#s2d}\n------------------------\n\nPresumptive *Campylobacter* isolates were confirmed using standard biochemical procedures including a Gram lysis test (3% \\[w/v\\] KOH, Sparks Lab Supplies, Dublin, Ireland), oxidase test (Oxoid, Cambridge, UK), catalase test (3% H~2~O~2~, Sigma Aldrich, St. Louis, MO, USA) and [l]{.smallcaps}-alanine aminopeptidase test (Aminopeptidase Test, Fluka Biochemika, Buchs, Switzerland). A latex agglutination test was also used (DrySpot Campylobacter Test Kit, Oxoid, Cambridge, UK). Following biochemical confirmation, isolates were stored at \u221220 and \u221280\u00b0C on Protect cryobeads (Technical Service Consultants Ltd., Heywood, Lancashire, UK) containing 80% \\[v/v\\] glycerol, for further analysis.\n\nGenotyping {#s2e}\n----------\n\n### Multiplex Polymerase Chain Reaction (mPCR) {#s2e1}\n\nDNA purification was carried out using a commercial kit (Wizard Genomic DNA Purification Kit, Promega, Madison, WI, USA). Isolates were identified to species level as described previously [@pone.0028490-Wang1]. Amplicon sizes were determined by comparison with a molecular weight marker (50 bp ladder, Promega) following migration on a 1.5% \\[w/v\\] agarose gel.\n\n### flaA-SVR sequencing {#s2e2}\n\nSequencing of an internal 321-bp fragment of the flagellin A short variable region (SVR), was performed using the primer pair Fla242FU \\[5\u2032-CTA TGG ATG AGC AAT T (AT) A AAAT-3\u2032\\] and Fla625RU \\[5\u2032-CAA G (AT) C CTG TTC C (AT)A CTG AAG-3\u2032\\], as previously described by Meinersmann *et al.* [@pone.0028490-Meinersmann1]. All *flaA*-derived amplicons were purified using a QIAquick PCR Purification kit (Qiagen GmbH, Hilden, Germany) and sequenced commercially by MWG Biotech (Ebersberg, Germany). The nucleotide sequences were deposited in the internet accessible *flaA*-SVR sequence database (<http://hercules.medawar.ox.ac.uk/flaA/>) and SVR allele numbers were assigned by sequence comparisons against the existing *flaA*-SVR sequences. The DNA trace files were submitted for confirmation of novel *flaA*-SVR alleles.\n\n### Multi Locus Sequence Typing (MLST) {#s2e3}\n\nInternal fragments of seven gene targets were amplified by PCR and their nucleotide sequences determined with primers and reaction conditions in accordance with the published MLST scheme for *C. jejuni* and *C. coli*. Allele identification followed by sequence type (ST) and clonal complex (CC) assignments were done by interrogation of the *Campylobacter* MLST database for each isolate (<http://pubmlst.org/campylobacter>).\n\nData Analysis {#s2f}\n-------------\n\nTrimmed *flaA*-SVR sequences (including STs) were imported into BioNumerics, Version 6.1 (Applied Maths, Sint-Martens-Latem, Belgium). A dendrogram was created using the unweighted-pair group method with arithmetic mean (UPGMA).\n\nResults {#s3}\n=======\n\nPresence of *Campylobacter* on poultry farms {#s3a}\n--------------------------------------------\n\n### Broiler breeders {#s3a1}\n\nAll broiler breeder flocks supplying each rearing farm were found to be colonized with both *C. jejuni* and *C. coli* ([Table 2](#pone-0028490-t002){ref-type=\"table\"}). A total of 27 *C. jejuni* and 19 *C. coli* were isolated from breeder farm 1 (containing flocks 1, 2 and 3). Two *C. jejuni* and 42 *C. coli* were isolated from breeder farm 2. One *C. coli* isolate and 22 *C. jejuni* isolates were recovered from breeder farm 3. There were 6 *C. jejuni* and 14 *C. coli* isolates recovered from breeder farm 1, flock 4, along with 7 *C. jejuni* and 12 *C. coli* from breeder farm 1, flock 5. A total of 13 *C. jejuni* and 7 *C. coli* were isolated from breeder farm 4, while 26 *C. jejuni* and 25 *C. coli* were recovered from breeder farm 5.\n\n10.1371/journal.pone.0028490.t002\n\n###### *Campylobacter* species recovered from all farms tested.\n\n![](pone.0028490.t002){#pone-0028490-t002-2}\n\n  Farm                       Day                       Source          Sample            Species (no.)\n  ----------- --------------------------------- --------------------- -------- ----------------------------------\n  1                          \\-                   Breeder 1 Flock 1    Faeces    *C. jejuni* (6), *C. coli* (2)\n                             \\-                   Breeder 1 Flock 2    Faeces    *C. jejuni* (6), *C. coli* (9)\n                             \\-                   Breeder 1 Flock 3    Faeces   *C. jejuni* (15), *C. coli* (8)\n                             13                        Broiler         Faeces           *C. jejuni* (10)\n                             13                  Adjacent Broiler 2    Faeces           *C. jejuni* (3)\n                             13                  Environment Inside     Air             *C. jejuni* (2)\n                             13                  Environment Outside   Puddle           *C. jejuni* (1)\n                             32                        Broiler         Faeces           *C. jejuni* (9)\n                             32                  Adjacent Broiler 1    Faeces           *C. jejuni* (1)\n                             32                  Adjacent Broiler 2    Faeces           *C. jejuni* (8)\n                             42                        Broiler         Faeces           *C. jejuni* (26)\n                             42                  Adjacent Broiler 1    Faeces           *C. jejuni* (5)\n                             42                  Environment Outside    Soil            *C. jejuni* (1)\n  **Total**    *C. jejuni* (93) *C. coli* (19)                                 \n  2                          \\-                       Breeder 2        Faeces   *C. jejuni* (2), *C. coli* (42)\n                             \\-                       Breeder 3        Faeces   *C. jejuni* (22), *C. coli* (1)\n                             27                  Environment Outside    Soil            *C. jejuni* (5)\n                             41                        Broiler         Faeces           *C. jejuni* (24)\n  **Total**                                                                     *C. jejuni* (53) *C. coli* (43)\n  3                          \\-                       Breeder 4        Faeces   *C. jejuni* (13), *C. coli* (7)\n                             \\-                       Breeder 5        Faeces   *C. jejuni* (26), *C. coli* (25)\n                             \\-                   Breeder 1 Flock 4    Faeces   *C. jejuni* (6), *C. coli* (14)\n                             \\-                   Breeder 1 Flock 5    Faeces   *C. jejuni* (7), *C. coli* (12)\n                             27                        Broiler         Faeces           *C. jejuni* (23)\n                             41                        Broiler         Faeces           *C. jejuni* (26)\n                             41                  Environment Outside    Soil            *C. jejuni* (1)\n  **Total**                                                                     *C. jejuni* (102) *C. coli* (58)\n\n-, not applicable.\n\n### Broiler chickens {#s3a2}\n\n*Campylobacter* could not be cultured from the transport crate paper liners on the day the chicks arrived at the rearing house. All three broiler flocks were found to be contaminated with *Campylobacter* by the end of the rearing period. *Campylobacter jejuni* was the only species isolated from all three poultry flocks under investigation ([Table 2](#pone-0028490-t002){ref-type=\"table\"}). Faecal samples from broiler flock 1 were contaminated with *Campylobacter* on days 13, 32 and 42. Faecal samples from broiler flock 2 were negative until the final sampling day (day 41). Broiler flock 3 was found to be contaminated with *Campylobacter* on days 27 and 41.\n\n### Presence of Campylobacter in the poultry farm environment {#s3a3}\n\nOn all three farms, *Campylobacter* could not be detected in the empty poultry house structure/environment after the cleaning and disinfection procedure had been carried out. All samples taken from the external environment of the cleaned house, including soil, air, the concrete apron and horse faeces were negative for *Campylobacter* ([Table 1](#pone-0028490-t001){ref-type=\"table\"}).\n\nDuring the rearing period of each flock, the environment was found to be contaminated with *Campylobacter* ([Table 1](#pone-0028490-t001){ref-type=\"table\"}). In the case of broiler farm 1 (day 13), an air sample taken inside the rearing house, an adjacent broiler flock and a puddle outside the house were sampled and found to be contaminated with *Campylobacter*. On day 32, two adjacent broiler flocks tested positive for *C. jejuni*. On day 42, one adjacent broiler house had been depopulated. *Campylobacter jejuni* was isolated from the remaining adjacent flock and was also recovered from an environmental soil sample. In the case of broiler farm 2, *C. jejuni* was isolated from a soil sample taken on day 27. On broiler farm 3, *C. jejuni* was recovered from a soil sample taken on day 41.\n\nMolecular sub-typing and spatio-temporal tracking of *Campylobacter* isolates {#s3b}\n-----------------------------------------------------------------------------\n\nA representative subset of *Campylobacter* isolates was chosen from each farm and characterized by DNA sequence analysis of the SVR containing region of the *flaA* gene (*flaA*-SVR), to determine if same-species isolates from different sources could be distinguished. [Figure 1](#pone-0028490-g001){ref-type=\"fig\"} shows the distribution of allele types discovered in relation to the source of each isolate. A total of 22 *flaA*-SVR alleles were identified in the 101 isolates investigated.\n\n![Distribution of *flaA*-SVR alleles detected according to source, (asterisks signify new *flaA*-SVR alleles).](pone.0028490.g001){#pone-0028490-g001}\n\n[Figures 2](#pone-0028490-g002){ref-type=\"fig\"}, [3](#pone-0028490-g003){ref-type=\"fig\"} and [4](#pone-0028490-g004){ref-type=\"fig\"} show the spatio-temporal distribution of *flaA*-SVR alleles identified on each sampling day across each of the three broiler farms. This sub-typing analysis revealed a diverse and weakly clonal population structure of *C. jejuni*, with multiple subtypes present throughout the lifecycle of each flock. *Campylobacter* isolates originating from each set of breeders and the faeces from their respective progeny presented with non-identical *flaA*-SVR DNA sequences.\n\n![Schematic diagram of the *flaA*-SVR alleles detected on farm 1.](pone.0028490.g002){#pone-0028490-g002}\n\n![Schematic diagram of the *flaA*-SVR alleles detected on farm 2.](pone.0028490.g003){#pone-0028490-g003}\n\n![Schematic diagram of the *flaA*-SVR alleles detected on farm 3.](pone.0028490.g004){#pone-0028490-g004}\n\nIn the case of farm 1, *flaA*-SVR allele number 16 was detected on day 13 in faeces from the broiler flock, faeces from adjacent broiler flock 2, and in a puddle located outside the house ([Figure 2](#pone-0028490-g002){ref-type=\"fig\"}). On day 32, allele number 16 continued to be identified in the broiler and adjacent broiler flock 2 faeces. On the same day, allele number 36 was detected in faeces from adjacent broiler flock 1. On day 42, allele number 36 could be detected in faeces from adjacent broiler flock 1, and was also now identified in broiler flock 1 and in an environmental soil sample. Allele numbers 8 and 16 were also detected in the broiler faeces.\n\nIn the case of farm 2, allele number 1137 was identified in an environmental soil sample taken on day 27 ([Figure 3](#pone-0028490-g003){ref-type=\"fig\"}). *Campylobacter* was not isolated from broiler flock 2 faeces until the final sampling day, when allele number 9 was detected.\n\nIn the case of farm 3, allele numbers 8, 36 and 816 were detected in the broiler faeces on day 27 ([Figure 4](#pone-0028490-g004){ref-type=\"fig\"}). On day 41, alleles 8 and 36 were again identified. Allele 26 was also found to be present in a soil sample outside the house on this day.\n\nDiversity of MLST Sequence Types (STs) and Clonal Complexes (CCs) {#s3c}\n-----------------------------------------------------------------\n\nIn order to substantiate *flaA*-SVR findings, a subset of isolates from each farm was chosen for MLST analysis.Eleven STs were identified among the 35 isolates (34 *C. jejuni*, 1 *C. coli*) chosen. Five STs were assigned to more than one isolate while six STs were assigned to single isolates. [Figure 5](#pone-0028490-g005){ref-type=\"fig\"} shows that ST 257 was the most frequently detected sub-type (in 15/35 isolates), followed by ST 48 (in 6/35 isolates) and ST19 (in 4/35 isolates). ST 21 and ST 45 were assigned to two isolates each. ST 583, ST 51, ST 1922, ST 1744, ST 4223 and ST 4224 were assigned to single isolates. Two novel sequence types (ST 4223 and ST 4224) were identified and submitted to the *C. jejuni* MLST database (<http://pubmlst.org/campylobacter/>).\n\n![Distribution of STs detected according to source (asterisks signify new STs).](pone.0028490.g005){#pone-0028490-g005}\n\nAll 11 STs were grouped into 6 CCs. The two novel sequences could not be grouped into a defined CC. The largest CC was found to be CC 257 (consisting of 16 isolates), followed by CC 21 and CC 48 (6 isolates each). In total, 80% (28/35) of isolates analyzed by MLST were grouped into one of these three CCs. Isolates grouped in CC 257 were found to have originated from a variety of sources ([Figure 6](#pone-0028490-g006){ref-type=\"fig\"}).\n\n![Distribution of CCs and STs according to source.](pone.0028490.g006){#pone-0028490-g006}\n\n[Figure 7](#pone-0028490-g007){ref-type=\"fig\"} depicts the phylogenetic relationship between the 22 *flaA*-SVR types and the 11 STs identified during the study. Nine cluster genotypes were observed using a threshold genetic similarity of 98% as a cut-off coefficient value.\n\n![Comparison of *flaA*-SVR sequences of *Campylobacter* isolates from breeder and broiler farms and farm environments.\\\nA total of 101 isolates were included. Strain *Campylobacter jejuni* NCTC11168 was included as a control strain in the pairwise analysis. The arrow indicates the 98% similarity cut-off point. *flaA*, *flaA*-SVR allele number; ST, sequence type; CC, clonal complex; Nt, not tested; NA, not assigned; -, not applicable.](pone.0028490.g007){#pone-0028490-g007}\n\nDiscussion {#s4}\n==========\n\nDespite extensive research, the definitive sources of infection and routes of transmission of *C. jejuni* in the poultry reservoir remain to be fully elucidated.\n\nThe purpose of this study was to identify sources of *Campylobacter* in intensively reared broiler flocks in Ireland.\n\nMolecular sub-typing techniques have previously shown the population structure of *C. jejuni* to be highly diverse and weakly clonal [@pone.0028490-Dingle1], [@pone.0028490-Meinersmann2]. This study confirmed that genetic diversity also exists among Irish *C. jejuni* strains isolated from broiler flocks, adjacent flocks and the farm environment.\n\nThe occurrence of vertical transmission of *Campylobacter* in poultry has been a controversial issue. The isolation of *Campylobacter* species from the reproductive tract of broiler breeders has been reported [@pone.0028490-Buhr2], [@pone.0028490-Cox1] and transmission from breeder hens to broiler chickens has been suggested [@pone.0028490-Cox2]. Conflicting reports have also been published suggesting that vertical transmission is unlikely or of little importance [@pone.0028490-JacobsReitsma1], [@pone.0028490-Sahin1]. In this study, while breeder flocks were found to be colonized with *Campylobacter*, molecular characterization confirmed the isolates to be of different *flaA*-SVR and ST types, suggesting that vertical transmission was not involved in the colonization of these broiler flocks with *Campylobacter*. This work is in agreement with a related study conducted by Patriarchi *et al.*, where none of the genotypes identified in breeder flocks were subsequently identified on any of the broiler farms [@pone.0028490-Patriarchi1]. Colonization of chickens with *Campylobacter* usually occurs between 3 and 5 weeks of age, and once infected, prevalence in a flock can often be close to 100% [@pone.0028490-JacobsReitsma2]. Interestingly, *Campylobacter* was not detected in any of the three broiler flocks in this study before day 13 of the rearing cycle. A possible factor contributing to the delay in colonization of chickens with *Campylobacter* is the presence of protective maternal antibodies in young chicks [@pone.0028490-Sahin2], [@pone.0028490-Cawthraw1].\n\nIn this study, *Campylobacter* was not isolated from broiler flock 2 until after the process of partial depopulation or thinning had been carried out. The practice of thinning has previously been reported as an important risk factor for *Campylobacter* colonization of residual birds [@pone.0028490-Allen1], [@pone.0028490-Hue1], [@pone.0028490-Hald1]. It has been shown *in vitro* that the presence of the neurotransmitter noradrenaline stimulates the growth and motility of *C. jejuni* [@pone.0028490-Cogan1]. As a result of triggering the release of noradrenaline, the stressful thinning process could be expected to contribute to rapid growth of the bacterium in the avian gastrointestinal tract leading to increased shedding of *Campylobacter* by birds, and the subsequent rapid spread of the bacteria.\n\nChallenges in maintaining biosecurity during the thinning process can result in cross-contamination from environmental sources. Allen *et al.* reported the isolation of *Campylobacter* from transport vehicles, equipment, personnel and the farm driveways prior to the thinning process [@pone.0028490-Allen1]. Using *flaA*-SVR and MLST subtyping methods, Patriarchi *et al.* also identified the practice of partial depopulation as a potential source and route of flock contamination on Irish broiler farms. Molecular evidence of the role of transport crates in introducing *Campylobacter* spp. into the broiler house was also reported [@pone.0028490-Patriarchi1]. In the case of farm 1 in this study, a *C. jejuni* isolate identified as *flaA*-SVR allele type 36 was isolated from an adjacent broiler flock prior to thinning. On the final sampling day, this *flaA*-SVR allele type was again isolated in the adjacent broiler flock, and was also identified in broiler flock 1 and in an environmental soil sample. Improved biosecurity measures in relation to the depopulation process may contribute to the prevention or delayed colonization of chickens with *Campylobacter*.\n\nPrevious studies have reported that poultry strains are frequently found to be genetically distinct from environmental isolates [@pone.0028490-Nesbit1], [@pone.0028490-Petersen2]. In the case of broiler farm 2, *C. jejuni flaA*-SVR allele types isolated from broilers and environmental soil samples were confirmed as non-identical. However, identical *C. jejuni* strains were identified from air, soil, water puddles and chickens on broiler farms 1 and 3, suggesting that transfer of campylobacters between these environments may be occurring. However, it is not possible based on the epidemiological data presented here to establish the direction of a given exchange. Bi-directional movement of *Campylobacter* between sources cannot be ruled out and has been implicated previously [@pone.0028490-Stern1]. In the case of two broiler flocks studied here, contamination of the farm environment was not detected until after the chickens had become infected, highlighting the broilers as a possible source of environmental contamination. MLST data from flock 1 further demonstrate that genetically identical strains can be isolated from broiler faeces and environmental samples. On day 13, ST 257 was found to be present in 2 broiler houses and in a water puddle outside the house. This ST was identified in the broiler flock on all sampling days. On day 42, ST 19 and ST 45 were also detected. These results reflect the findings of previous spatio-temporal studies on broiler farms, where different STs were identified in chicken faeces as the rearing period progressed [@pone.0028490-Bull1].\n\nDuring 2007, New Zealand experienced a 50% decline in the rate of campylobacteriosis notifications and hospitalisations [@pone.0028490-Sears1]. This decline followed the introduction of voluntary and regulatory interventions to reduce contamination of poultry with *Campylobacter* species and was sustained in 2008 and 2009. A number of other countries have reported a reduced incidence of campylobacteriosis infections following the implementation of poultry-focussed control strategies [@pone.0028490-Hofshagen1], [@pone.0028490-Stern2], [@pone.0028490-Hansson1], [@pone.0028490-Rosenquist1]. Various interventions were employed in each country however all strategies included strengthening on-farm biosecurity and monitoring the prevalence of *Campylobacter*-positive flocks. The implementation of similar measures in poultry farms in Ireland could contribute to a reduction in human campylobacteriosis infection rates and lead to improved public health protection.\n\nThe population structure of Campylobacter isolates from broiler farms in Southern Ireland was determined, (based on these data), to be weakly clonal. Such genetic diversity complicates the challenge of managing Campylobacter species population dynamics within the poultry farm environment. Nevertheless, our data point to the importance of applying more than one sub-typing method as part of our epidemiological studies to carefully describe this dynamic process.\n\nIt is reasonable to conclude that there are multiple sources from which *Campylobacter* can be transmitted on broiler farms. Following their introduction into broiler flocks, the spread of *Campylobacter* is influenced by various host and environmental factors, such as biosecurity measures in place, farming practices, the immune status of the chickens, and the presence of other animals on the farm. A combined protocol individually targeting each potential source and route of transmission is required as a logical approach to effectively reduce the colonization of broiler chickens with *Campylobacter* in a step-wise fashion.\n\nThe authors wish to acknowledge the cooperation of the poultry farmers who participated in the study.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This work was funded by Safefood (04-RESR-04). [www.safefood.eu](http://www.safefood.eu). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: EOM JB DB SF. Performed the experiments: EOM. Analyzed the data: EOM. Contributed reagents/materials/analysis tools: JB DB PW SF. Wrote the paper: EOM.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nSchizophrenia is a chronic mental disorder which involves disturbances of thought, perception, affect, and social behavior \\[[@B1]\\]. It is ranked among the top ten causes of disability worldwide with mortality rates two times as high as in the general population \\[[@B2]--[@B4]\\]. Although accidents and suicide account for just a portion of this high mortality rate, more than two-thirds of the rate is caused by different forms of comorbid physical diseases such as cardiovascular and metabolic syndrome \\[[@B3]\\]. The emergence of these physical diseases has been linked to poor lifestyle (smoking, sedentary lifestyle, and poor diet) and cardiovascular risk factors \\[[@B5]\\].\n\nMetabolic syndrome (MetS) is a group of disorders which includes central obesity, dyslipidemia, hypertension, abnormal glucose homeostasis, proinflammatory state, and prothrombotic state \\[[@B6], [@B7]\\]. The prevalence of MetS seems to be common and rising in the general population possibly due to high intake of calories and a lifestyle of inactivity, irrespective of definition criteria \\[[@B8], [@B9]\\]. It is, however, worth considering that some population groups and, for that matter, particular patient groups even have an increased propensity of developing MetS. Several studies have convincingly indicated that schizophrenic patients are twice at risk of developing MetS compared to the general population \\[[@B8], [@B10], [@B11]\\]. Before the introduction of antipsychotics, there were increasing reports of glucose dysregulations (i.e., glucose intolerance, insulin resistance, and hyperglycaemia) by researchers in schizophrenic patients \\[[@B12]\\]. Moreover, several recent studies have reported increased prevalence of hyperglycaemia, insulin resistance, and impaired fasting glucose tolerance in first episode, antipsychotic-na\u00efve schizophrenic patients compared to healthy control subjects \\[[@B11]\\].\n\nConsequently, treatment with antipsychotics in people with schizophrenia tends to worsen or even exacerbate these MetS and, thus, contribute to the increased prevalence from 40% to 60% versus 30% in the general population \\[[@B13], [@B14]\\]. The atypical antipsychotics are generally known to possess greater propensity of causing these metabolic abnormalities than the typical antipsychotics. Concerning the metabolic dysregulation that accompanied treatment with antipsychotic medications, several studies have recommended a baseline and continuous monitoring (every 6 months) of fasting blood glucose, glycated haemoglobin, and fasting lipid profile to be done in routine clinical practice with all atypical antipsychotics \\[[@B15]--[@B17]\\].\n\nHowever, in resource constraint health systems as we find in psychiatric hospitals in Ghana, such monitoring is not routinely undertaken, hence the difficulty in determining the prevalence of these metabolic dysregulations. There is a need for current data to serve as a basis for decision-making in terms of routine monitoring for metabolic dysregulations among these patients by healthcare providers in Ghana. Therefore, this study sought to determine the prevalence of MetS among patients with schizophrenia.\n\n2. Methods {#sec2}\n==========\n\n2.1. Study Design and Setting {#sec2.1}\n-----------------------------\n\nThis was a comparative cross-sectional study which took place at the Out-Patient Department of the Psychiatry Unit of Komfo Anokye Teaching Hospital, (KATH) over a period of five (5) months (April to August 2016). KATH is a 1200-bed facility located in Kumasi in the Ashanti region of Ghana with a projected population of 4,780,380 according to the 2010 census, accounting for 19.4% of Ghana\\'s total population. The Ashanti region is centrally located in the middle belt of Ghana. It lies between longitudes 0.15W and 2.25W, and latitudes 5.50N and 7.46N. The region is divided into 27 districts. The Psychiatry Unit, which is a department under Internal Medicine Directorate, serves both in- and out-patients. About 150 schizophrenic patients are attended to every month on out-patient basis at the Psychiatry Unit.\n\n2.2. Study Population and Subject Selection {#sec2.2}\n-------------------------------------------\n\nThis study comprised 112 newly diagnosed antipsychotic-na\u00efve and 236 patients on antipsychotics who were receiving treatment at the psychiatric unit. These participants were randomly selected into the study. Laboratory investigations which include fasting blood glucose and lipid levels \\[triglycerides (TG), low density lipoprotein (LDL), very low density lipoprotein (VLDL), and total cholesterol (TC) as well as measurements of blood pressure and anthropometric measurements (height, weight, waist, and hip circumference)\\] were carried out.\n\n2.3. Inclusion Criteria {#sec2.3}\n-----------------------\n\nPatients with confirmed diagnosis of schizophrenia according to the International Classification of Diseases (ICD-10) criteria and attending the clinic for review as scheduled by their medical doctors, who were on either monotherapy or dual therapy (fluphenazine-based combination therapies) of antipsychotics, and aged 18 years and above were recruited randomly for the study.\n\n2.4. Exclusion Criteria {#sec2.4}\n-----------------------\n\nIn-patients and patients with known history of diabetes or hypertension and who were on mood stabilizers and antidepressants/benzodiazepines at the time of enrollment in the study were excluded from the study.\n\n2.5. Biochemical Analysis {#sec2.5}\n-------------------------\n\nAbout 5 ml of venous blood sample was collected after an overnight fast (12--16 hours) and dispensed into fluoride oxalate tubes and vacutainer plain tubes for separation into plasma and serum, respectively. In the laboratory, the samples were centrifuged at 4000 revolution/minute for 5 min within 30 min of collection and stored at -80\u00b0C. Assay parameters included fasting plasma glucose (FPG), triglycerides (TG), total cholesterol (TC), low density lipoprotein (LDL), and high density lipoprotein (HDL) cholesterol using the COBAS INTEGRA^(R)^ 400*plus*automated chemistry analyzer.\n\n2.6. Anthropometric Measurement {#sec2.6}\n-------------------------------\n\nAnthropometric measurements including height (measured without shoes) and weight (to nearest 0.1 kg in light clothing) were taken. Calculation of the Body Mass Indices (BMI) was done by dividing weight (kg) by height squared (m^2^). With respect to the BMI, there were four groupings of subjects: underweight (BMI \\< 19 Kg/m^2^), normal (BMI between 19 and 24.9 Kg/m^2^), overweight (BMI between 25 and 29.9 Kg/m^2^), and obese (BMI \\> or = 30 Kg/m^2^). The waist circumference (in centimeter) was measured midway between the lowest rib and the above iliac crest using the Gulick II spring-loaded measuring tape (Gay Mill, WI). The hip circumference was also measured (in centimeters) as the widest circumference around the buttocks. The waist and hip circumference as well as other parameters were used for the calculation of waist to hip ratio (WHR), body adiposity (BAI) \\[[@B18]\\], and visceral adiposity indices (VAI) \\[[@B19]\\].\n\n2.7. Blood Pressure Measurement {#sec2.7}\n-------------------------------\n\nBlood pressure (BP) was recorded after 5 minutes of rest with the subject being in the seated position using manual and an automated sphygmomanometer placed on the subject\\'s right arm. This was measured three times, and the average reading was recorded. Individuals were deemed hypertensive if one or more of these features occur: they were taking antihypertensive medications, they self-reported a diagnosis of hypertension, systolic pressure reading was above 140 mm Hg, or diastolic pressure reading was above 90 mm Hg.\n\n2.8. Definition of Metabolic Syndromes {#sec2.8}\n--------------------------------------\n\nThe three most commonly used definitions of metabolic syndrome, as used in this study, are described below.\n\n2.9. World Health Organization (WHO) Criteria {#sec2.9}\n---------------------------------------------\n\nThe WHO criteria involve the presence of diabetes mellitus, insulin resistance, or impaired glucose tolerance and any two of the following: (1) body mass index (BMI) \u2265 30 kg /m^2^ and/or waist: hip ratio \\> 0.85 (female), \\> 0.90 (male); (2) blood pressure, systolic \u2265 140 or diastolic \u2265 90 mmHg or on medication; (3) triglyceride \u2265 1.7 mmol/L and/or HDL-C \\< 1.01 mmol/L (female), \\< 0.91 mmol/L (male) \\[[@B20]\\].\n\n2.10. National Cholesterol Education Program, Adult Treatment Panel III (NCEP ATPP III) Criteria {#sec2.10}\n------------------------------------------------------------------------------------------------\n\nAccording to the NCEP ATP III, individuals with metabolic syndrome should have at least three of the following: (1) abdominal obesity (waist circumference \\> 88 cm for women or \\> 102 cm for men); (2) raised triglyceride (\u2265 1.7 mmol/L); (3) low HDL cholesterol (\\< 1.0 mmol/L in women or \\< 0.9 mmol/L in men); (4) high blood pressure (systolic BP \u2265 130 mmHg or diastolic BP \u2265 85 mmHg or treatment of hypertension); and (5) raised fasting glucose (\u2265 6.1 mmol/L) \\[[@B21]\\].\n\n2.11. International Diabetes Federation (IDF) Criteria {#sec2.11}\n------------------------------------------------------\n\nAccording to the IDF criteria, metabolic syndrome is diagnosed if there is central obesity (waist \\> 80 cm for women or circumference \\> 90 cm for men) in addition to any two (2) of the following four (4) factors: (1) triglyceride level \u2265 1.7 mmol/L; (2) HDL cholesterol \\< 1.29 mmol/L for women or \\< 1.03 mmol/L for men; (3) blood pressure \u2265 130 /85 mmHg or treatment of previously diagnosed hypertension; and (4) fasting blood glucose (FBG) \u2265 5.6 mmol/L or previously diagnosed type 2 diabetes \\[[@B20]\\].\n\n2.12. Ethical Consideration {#sec2.12}\n---------------------------\n\nThe investigations were approved by the Committee on Human Research Publications and Ethics (CHRPE) at School of Medical Sciences, KNUST, Ghana, and the Psychiatry Unit of KATH. Written informed consent was obtained for all participants to maintain their right to self- determination from their visitors/relatives.\n\n2.13. Data Analysis {#sec2.13}\n-------------------\n\nThe data obtained were analyzed using Statistical Package for Social Sciences SPSS (version 23.0), statistical packages for Windows. Descriptive statistics that were computed included frequencies and percentages for all categorical variables in addition to means, standard deviations, and ranges for all normally distributed continuous variables, while median and interquartile range were used for continuous variables that were not normally distributed. Differences between groups were carried out using independent sample*t*-test for continuous variable while Chi-square test was used for categorical variables. Univariate binary logistic regression model was used to detect factors associated with MetS among schizophrenia patients. Statistical significance was determined as p value \\<0.05.\n\n3. Results {#sec3}\n==========\n\nSubjects in the treatment group were slightly older (with median age of 36 years) than the treatment-na\u00efve group (median age = 32.5 years) but the difference was not significant (p=0.2222). Majority of the study participants were unemployed (55.5%), were single (58.9%), completed Junior High School (45.3%), and had no family history of psychiatric illness (69.1%). Also, several of the study subjects had no history of alcohol intake and smoking (79.1%). Within the antipsychotic-treated group, about half of these subjects (51.3%) had been on a particular medication continuously for more than 12 months**\\[[Table 1](#tab1){ref-type=\"table\"}\\]**.\n\nAs shown in [Table 2](#tab2){ref-type=\"table\"}, prevalence of MetS was significantly prevalent in subjects on antipsychotic treatment compared to the treatment-na\u00efve patients for all the classifications (p\\<0.05). Based on the NCEP/ATP III criteria, 17.8% of the patients on treatment had MetS (score \u22653) compared to 6.2% of the treatment-na\u00efve patients with an odd ratio of 11.3 (p=0.0001). Using the IDF criteria, 30.3% of patients on treatment compared to 9.8% of antipsychotic-na\u00efve patients had developed MetS (OR = 12.6, p\\<0.0001). Similarly, the WHO criteria revealed that 26.2% of the treated patients in comparison with 8.0% of newly diagnosed patients had MetS with an odd ratio of 15.5 (p\\<0.0001).\n\nThe prevalence of MetS was higher in subjects on atypical antipsychotics as compared to those on typical medications for all the classifications, though no significant difference was observed (p\\>0.05). Furthermore, the risk of developing MetS increased with increasing score of MetS for the all classifications (p\\>0.05).\n\nIt was observed that the odds of MetS increase steadily with increasing bioscore of MetS for the individual classifications. Also, the prevalence of MetS was generally higher in patients on the monotherapy than in those on the dual therapy in all the classifications. There was, however, no statistical difference observed (p\\>0.05). With respect to the NCEP/ATP III criteria, there were increasing reduced odds of MetS observed with patients on dual therapy with reference to those on monotherapy (OR=0.6). However, no statistical significant difference was observed (p=0.643). Similar observations were seen with the IDF and WHO criteria**\\[[Table 4](#tab4){ref-type=\"table\"}\\].**\n\nUsing NCEP/ATP III criteria as dependent variable, obesity-WHR, obesity-WHtR, raised fasting blood sugar, raised TC, raised triglycerides, and decreased HDL were significantly associated with increased risks of MetS based on logistic regression model. When the IDF criteria were used as a dependent outcome, age range 50-59 years, obesity-BMI, obesity-WHR, obesity-WHtR, raised FPG, raised TC, raised TG, and raised LDL-C were significant predictors of MetS. Using the WHO criteria as dependent variables, age range 40-49 years, age \u226560 years, obesity-BMI, obesity-WHR, raised TG, and reduced HDL-C were statistically significantly associated with increased risk of MetS**\\[**[S1](#supplementary-material-1){ref-type=\"supplementary-material\"} and [S2](#supplementary-material-1){ref-type=\"supplementary-material\"}**\\]**.\n\n4. Discussion {#sec4}\n=============\n\nThe prevalence of MetS seems to be common and rising in the general population possibly due to high intake of calories and a lifestyle of inactivity, irrespective of definition criteria \\[[@B9], [@B10]\\]. Similar observation was made when the overall prevalence of MetS in this study was compared to the prevalence rate reported from previous study in the general Ghanaian population. Thus, the overall prevalence of MetS among schizophrenic patients in Ghana as determined by IDF (23.6%), WHO (20.4%), and NCEP ATP III (14.1%) in this study was higher compared to the general Ghanaian population prevalence rate (IDF = 7.8%, NCEP ATP III = 3.9% and WHO = 2.2%) as determined by Owiredu et al. \\[[@B23]\\]. Moreover, a comparison of the overall prevalence from this study to that obtained among psychiatric patients in Ghana shows a consistent rise in the prevalence of MetS among these patients over the years regardless of definition criteria (IDF = 23.6% versus 15.5%; WHO = 20.4% versus 13.5%; NCEP ATP III = 14.1% versus 11.5%) \\[[@B23], [@B24]\\]. Furthermore, the MetS prevalence observed in this study was in agreement with a range of prior reports published regionally and worldwide, although there is a considerable variation in criteria and methodology among the published reports \\[[@B13], [@B16], [@B25]--[@B27]\\]. However, studies about prevalence rate of MetS are very few in Africa, with most of the studies originating from the southern and northern part of Africa \\[[@B16], [@B28]--[@B32]\\].\n\nMoreover, the prevalence of MetS was significantly higher among schizophrenic patients on antipsychotic treatment than the treatment-na\u00efve group, which concurs with findings from Saddichha et al. \\[[@B11]\\] who reported a high prevalence of MetS among antipsychotic-treated group compared with matched healthy control group. Findings from Grover et al. \\[[@B33]\\]and Shakeri et al. \\[[@B34]\\] also reported similar findings that MetS was highly prevalent among patients treated with antipsychotics, hence increasing their risk of developing cardiovascular diseases. The smaller prevalence observed in the newly diagnosed treatment-na\u00efve group from the current study further implies that exposure to antipsychotics contributes largely, but not singularly, to the development of cardiometabolic disorders among schizophrenics as well as mentally ill patients.\n\nAmong the antipsychotic medications, numerous studies have indicated higher prevalence of MetS associated with atypical antipsychotic use as compared to the typical medications \\[[@B5], [@B35]\\]. This is not consistent with the present study; though the current study showed a higher prevalence of MetS in patients on atypical antipsychotics for all the classification criteria compared to the typical drugs, no statistical significant difference was observed \\[[Table 3](#tab3){ref-type=\"table\"}\\]. The higher metabolicogenic ability of atypical antipsychotics may be due to their affinity for a wide range of receptor systems (serotonergic, adrenergic, cholinergic, and histaminergic receptors) other than the traditional dopaminergic activity \\[[@B5]\\]. Several reports have also shown that some atypical antipsychotics have greater propensity of causing metabolic syndrome compared to other medicines within the same class \\[[@B37]--[@B39]\\]. Furthermore, studies by Shirzadi and Ghaemi and Lieberman reported that clozapine and olanzapine treated subjects were at increased risk of weight gain, dyslipidaemias, and hyperglycaemia when compared with other atypical antipsychotics as well as typical medicines. \\[[@B15], [@B39]\\]. Therefore, these reports have provided undeniable evidence of the metabolicogenic ability of olanzapine as well as clozapine and permit the inclusion of a black-box warning in prescriptions of these medicines. It is, however, worth noting that strict monitoring of metabolic parameters is paramount for all schizophrenic patients irrespective of antipsychotic used.\n\nCurrently, antipsychotic polypharmacy is commonly seen in clinical practice for the management of schizophrenia in most countries as well as Ghana \\[[@B40]\\]. This practice is not explicitly recommended in most clinical practice guidelines exemplified by National Institute for Health and Care Excellence (NICE), Ghana Standard Treatment Guidelines 2010, and World Federation of Societies of Biological Psychiatry \\[[@B41]\\]. However, most clinicians use combination therapies in treatment resistant cases or when unsatisfactory response to standard dose of monotherapy is observed \\[[@B40]\\]. In spite of this, studies that compare the efficacy of mono- and combination therapies have been inconclusive \\[[@B40]\\]. Furthermore, several cross-sectional studies have reported increased prevalence of metabolic effects and cardiovascular mortality associated with antipsychotic cotreatment \\[[@B42], [@B43]\\]. Contrary to these findings, the present study revealed a higher prevalence of MetS associated with monotherapy. Since this study did not consider simultaneous initiation of cotreatment at the beginning of the data collection, it is possible that the adverse effects observed may be related to the direct effect of specific antipsychotic rather than the combination therapy.\n\nNevertheless, it is unclear whether reports from those previous studies are related to cohort effect, as patients who were selected for antipsychotic combination therapy were psychiatrically and physically sicker than those on monotherapy. Also, such studies mostly considered clozapine- based therapies rather than fluphenazine-based combination therapy as used in this present study \\[[@B44]\\]. Moreover, this finding could also be explained by the fact that there are more rapid metabolizers in treatment resistant patients. Thus they might experience fewer side effects. Also they might smoke more or use more cannabis, both affecting metabolic syndrome rates. Consequently, it is difficult to assume such risks for all other combination therapies. It is therefore important that longer-term studies of sufficient methodological quality and sample size be conducted to provide sufficient evidence to support the efficacy, long-term safety, mortality, and cost of antipsychotic cotreatment.\n\nThe association between age and development of MetS has been widely reported. The prevalence of MetS increases with age in the general population and a similar trend had been generally observed in patients with schizophrenia \\[[@B17], [@B34]\\]. Conversely, some studies have also revealed a peak of MetS rates in the third, fourth, or fifth decade of life with subsequent decline in MetS rates in later life \\[[@B27], [@B45]\\]. The present study revealed significantly increased risk of MetS in both younger (\\<49 years) and older (\\>50 years) age groups as defined by the IDF and WHO classifications. Furthermore, a critical assessment of the significant odds of developing MetS per the IDF and WHO classifications showed that the risk of MetS increased exponentially with increasing age. The increased risk of MetS observed in the younger age group may be attributed to the fact that patients within this group may have a greater predisposition to developing MetS despite being exposed to antipsychotics. Thus, the exposure to antipsychotics may be thought of as quickening MetS onset instead of precipitating it de novo. Also, the significant risk observed in the older ages suggests that, in those groups, age may be a more important risk factor for comorbid MetS as against the neuroleptic prescribed.\n\nFrom the study, the prevalence of MetS was found to be higher in female subjects compared to the male counterparts but was not statistically significant according to the NCEP ATP III criteria. However, significantly higher prevalence rate was observed in the IDF and WHO classifications. Also, females were approximately 3-4 times at risk of developing MetS compared to the males. These findings agree with a study by You-Kyung et al., \\[[@B46]\\] who found a significantly higher prevalence rate of MetS in female subjects compared to males according to the IDF classification. However, Sugawara et al. \\[[@B47]\\] together with other fewer studies \\[[@B13], [@B48]\\] have reported slightly higher MetS prevalence rate in men or no significant differences in MetS rates across sexes. The higher prevalence of MetS in females may be attributed to a more sedentary lifestyle, obesity, and age-related physical and hormonal changes \\[[@B49]\\].\n\nSeveral studies have suggested that schizophrenia is an independent risk factor for diabetes due to disease-related stress and poor lifestyle of patients \\[[@B3], [@B50]\\]. Also, studies have shown that exposure to antipsychotics further predisposes patients to hyperglycaemia, new onset diabetes, and diabetic ketoacidosis in schizophrenic patients with higher risk associated with atypical antipsychotics than the typical medicines \\[[@B51]--[@B53]\\]. Logistic regression model indicates that hyperglycaemia (FPG \\> 6.4 mmol/L) was significantly associated with higher risks of MetS among patients treated with antipsychotics according to the IDF criteria \\[Table 6\\]. These findings concur with previous studies that have identified high prevalence of hyperglycemia and risk of MetS in patients on antipsychotics \\[[@B51], [@B54]\\]. Weight gain may be the most obvious signal of MetS and is usually the most distressing issue among patients, thus contributing to noncompliance to antipsychotic medications \\[[@B55]\\]. Increased abdominal adiposity as measured by waist circumference is important causative risk factor for MetS and a prerequisite for the diagnosis of MetS by the IDF criteria. Results from the logistic regression in this study showed that obesity as defined by BMI, WHR, and WHtR as per the NCEP ATP III, WHO, and IDF criteria were significant predictors of MetS. This shows that both atypical and typical antipsychotics have an almost equal propensity in inducing central obesity (i.e., per definition using WHR) in patients on medication. This is in accordance with reports by Allison et al. and Bobes et al.\\[[@B14], [@B56]\\] which revealed that typical antipsychotics and, to a greater extent, atypical medicines are associated with metabolic disturbances and weight gain.\n\nDyslipidaemia is an important component of the metabolic syndrome and occurs along with glucose dysregulation and weight gain in patients treated with antipsychotics. Logistic regression model revealed that dyslipidaemia (defined as increased TC, TG, and LDL as well as reduced HDL-C) in patients on antipsychotic treatments was associated with increased risk of MetS in all the classifications, though not statistically significant. This agrees with a study by Owiredu et al. \\[[@B23]\\] which showed that hypertriglyceridaemia and reduced HDL cholesterol were significant risk factors for metabolic syndrome associated with antipsychotic use, especially atypical medicines. In a study to explore the association between hyperglycaemia and antipsychotic use, Haupt and Newcomer \\[[@B57]\\] suggested that dyslipidaemia as well as increased abdominal obesity impairs glucose metabolism leading to treatment-related hyperglycaemia and type 2 diabetes mellitus. This could therefore explain the significant finding of diabetes observed in patients on antipsychotics treatments in this study.\n\nNotwithstanding, findings in this study are comparable to reports from numerous studies. The study faced some limitations that may affect the generalizability of the results. First, as the study subjects were patients at a single psychiatric hospital, the results obtained from their examination cannot be generalized for the entire population of schizophrenic patients in Ghana. Secondly, variables that can impact MetS, such as physical condition, diet, severity of symptoms, duration of illness, physical activity level, and socioeconomic status, were not adequately evaluated.\n\n5. Conclusion {#sec5}\n=============\n\nThe prevalence of MetS among the psychotic-treated patients in this study is triple that previously reported in nonpsychotic individuals. MetS was also prevalent among patients on atypical and monotherapy than those on typical and dual antipsychotic medications. Regular monitoring of cardiometabolic parameters should be an important therapeutic objective in the management of these patients.\n\nThe authors would like to acknowledge the management and staff at Psychiatry Unit, KATH, for allowing the authors to carry out this work in their departments.\n\nData Availability\n=================\n\nThe data used to support the findings of this study are included within the article.\n\nConflicts of Interest\n=====================\n\nThe authors declare that there are no conflicts of interest regarding the publication of this paper.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\n**S1: Table**5: logistic regression analysis of biochemical parameters predicting risk of MetS among studied subjects.**S2: Table6**: logistic regression analysis of age and anthropometric indices predicting risk of MetS among studied subjects.**S3: Table7**: prevalence of MetS and components of MetS defined by the NCEP/ATP III, IDF, and WHO criteria stratified by drugs.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nSociodemographic characteristics of study participants.\n\n  **Sociodemographics**                     **Psychotic-treated**   **Treatment-na\u00efve**   **\u03c7** **2,df**   **p-value**\n  ----------------------------------------- ----------------------- --------------------- ---------------- -------------\n  **Age (years)**                           36.0 (27.0 -45.0)       32.5 (24.8 - 45.3)    \u2009                0.222\n  **Age groups (years)**                    \u2009                       \u2009                     6.5,3            0.091\n  \\< 40                                     143(60.6)               76(67.9)              \u2009                \u2009\n  40 -- 49                                  61(25.9)                16(14.3)              \u2009                \u2009\n  50 -- 59                                  20(8.5)                 11(9.2)               \u2009                \u2009\n  \u2265 60                                      12(5.1)                 9(8.0)                \u2009                \u2009\n  **Gender**                                \u2009                       \u2009                     \u2009                0.647\n  Male                                      115(48.7)               58(52.0)              \u2009                \u2009\n  Female                                    121(51.3)               54(48.0)              \u2009                \u2009\n  **Marital status**                        \u2009                       \u2009                     10.4,3           **0.015**\n  Single                                    139(58.9)               56(50.0)              \u2009                \u2009\n  Married                                   79(33.5)                47(4.0)               \u2009                \u2009\n  Divorced                                  14(5.9)                 2(1.8)                \u2009                \u2009\n  Widowed                                   4(1.7)                  7(6.2)                \u2009                \u2009\n  **Employment status**                     \u2009                       \u2009                     0.60,2           0.740\n  Unemployment                              131(55.5)               65(58.0)              \u2009                \u2009\n  Employment                                99(41.9)                43(38.3)              \u2009                \u2009\n  Student                                   6(2.5)                  4(3.6)                \u2009                \u2009\n  **Educational level**                     \u2009                       \u2009                     10.0,4           **0.040**\n  Primary                                   12(5.1)                 11(9.8)               \u2009                \u2009\n  JHS                                       107(45.3)               56(50.0)              \u2009                \u2009\n  SHS                                       59(25.0)                18(16.1)              \u2009                \u2009\n  Tertiary                                  50(21.2)                18(16.1)              \u2009                \u2009\n  None                                      8(3.4)                  9(8.0)                \u2009                \u2009\n  **Family history of Psychotic illness**   \u2009                       \u2009                     0.206            \n  Yes                                       73(30.9)                27(24.1)              \u2009                \u2009\n  No                                        163(69.1)               85(75.9)              \u2009                \u2009\n  **History of alcohol and smoking**        \u2009                       3.2,3                 0.365            \n  Alcohol                                   20(8.5)                 9(8.0)                \u2009                \u2009\n  Smoking                                   6(2.5)                  0(0.0)                \u2009                \u2009\n  Alcohol and smoking                       22(9.3)                 9(8.0)                \u2009                \u2009\n  None                                      188(79.7)               94(84.0)              \u2009                \u2009\n  Duration of Treatment                     \u2009                       \u2009                     \u2009                \u2009\n  \\< 3 months                               23(9.8)                 \u2009                     \u2009                \u2009\n  3 -8 months                               44(18.6)                \u2009                     \u2009                \u2009\n  9 -- 12 months                            48(20.3)                \u2009                     \u2009                \u2009\n  \\>12 months                               121(51.3)               \u2009                     \u2009                \u2009\n\ndf: degree of freedom; *\u03c7*2: chi-square; n: frequency; JHS: Junior High School; SHS: Senior High School; EPS: extrapyramidal symptoms; p\\<0.05 is statistically significant.\n\n###### \n\nMetS among psychotic-treated and treatment-na\u00efve studied subjects.\n\n  **Variables**            **Total**   **Psychotic-treated**   **Treatment-na\u00efve**   \u2009                \u2009\n  ------------------------ ----------- ----------------------- --------------------- ---------------- --------------\n  **Met / NCEP/ATP III**   \u2009           \u2009                       \u2009                     \u2009                \u2009\n  0                        75(21.6)    26(11.0)                49(43.8)              1                \u2009\n  1                        136(39.1)   93(39.4)                43(38.4)              4.1(2.2-7.4)     **0.0013**\n  2                        88(25.2)    75(31.8)                13(11.6)              10.9(5.1-23.2)   **\\<0.0001**\n  \u22653                       49(14.1)    42(17.8)                7(6.2)                11.3(4.5-47.6)   **0.0001**\n  **Met / IDF**            \u2009           \u2009                       \u2009                     \u2009                \u2009\n  0                        47(13.5)    16(6.7)                 31(27.7)              1                \u2009\n  1                        123(35.3)   71(30.3)                52(46.4)              2.7(1.3 -5.3)    **0.006**\n  2                        96(27.6)    78(32.7)                18(16.1)              8.4(3.8-18.5)    **\\<0.0001**\n  \u22653                       82(23.6)    71(30.3)                11(9.8)               12.5(5.2-30.0)   **\\<0.0001**\n  **Met / WHO**            \u2009           \u2009                       \u2009                     \u2009                \u2009\n  0                        65(18.7)    20(8.5)                 45(40.2)              1                \u2009\n  1                        122(35.1)   75(31.8)                47(42.0)              3.6(1.9-6.8)     **\\<0.0001**\n  2                        90(25.9)    79(33.5)                11(9.8)               16.2(7.1-36.8)   **\\<0.0001**\n  \u22653                       71(20.4)    62(26.2)                9(8.0)                15.5(6.5-37.2)   **\\<0.0001**\n\nNCEP ATP III: National Cholesterol Education Program, Adult Treatment Panel III; IDF: International Diabetes Federation; WHO: World Health Organization; MetS: metabolic syndrome; cOR: crude odds ratio; CI: confidence interval.\n\n###### \n\nMetS among studied subjects treated with atypical and typical antipsychotics.\n\n  **Variables**        **Total**   **Atypical**   **Typical**   \u2009               \u2009\n  -------------------- ----------- -------------- ------------- --------------- -------\n  Met / NCEP/ATP III   \u2009           \u2009              \u2009             \u2009               \u2009\n  0                    12(7.5)     6(5.7)         6(11.1)       1               \u2009\n  1                    65(41.3)    41(39.1)       24(44.5)      1.7(0.5-5.9)    0.522\n  2                    58(36.2)    40(38.1)       18(33.3)      2.2(0.6-7.8)    0.316\n  \u22653                   24(15.0)    18(17.1)       6(11.1)       3.0(0.6-12.9)   0.157\n  Met / IDF            \u2009           \u2009              \u2009             \u2009               \u2009\n  0                    8(5.0)      4(3.8)         4(7.4)        1               \u2009\n  1                    50(31.5)    26(24.8)       24(44.5)      1.1(0.1-9.0)    1.000\n  2                    56(35.2)    44(41.9)       12(22.2)      3.7(0.8-16.9)   0.099\n  \u22653                   45(28.3)    31(29.5)       14(25.9)      2.2(0.5-10.2)   0.421\n  Met / WHO            \u2009           \u2009              \u2009             \u2009               \u2009\n  0                    12(7.6)     6(5.7)         6(11.1)       1               \u2009\n  1                    53(33.3)    34(32.0)       19(35.2)      1.8(0.5-6.3)    0.512\n  2                    55(34.6)    36(34.0)       19(35.2)      1.9(0.5-6.7)    0.341\n  \u22653                   39(24.5)    29(44.3)       10(18.5)      2.9(0.8-11.1)   0.157\n\nNCEP ATP III: National Cholesterol Education Program, Adult Treatment Panel III; IDF: International Diabetes Federation; WHO: World Health Organization; MetS: metabolic syndrome; OR: odds ratio; CI: confidence interval; p\\<0.05 is statistically significant.\n\n###### \n\nMetS among studied subjects on dual and mono antipsychotic therapy.\n\n  **Variables**        **Total**   **Dual therapy**   **Monotherapy**   **OR (95**%**CI)**   **p-value**\n  -------------------- ----------- ------------------ ----------------- -------------------- -------------\n  Met / NCEP/ATP III   \u2009           \u2009                  \u2009                 \u2009                    \u2009\n  0                    26(10.9)    14(18.1)           12(7.5)           1                    \u2009\n  1                    93(39.5)    27(35.1)           66(41.5)          0.4(0.1-0.9)         **0.034**\n  2                    75(31.9)    18(23.4)           57(38.9)          0.3(0.1-0.7)         **0.007**\n  \u22653                   42(17.7)    18(23.4)           24(15.1)          0.6(0.2-1.7)         0.643\n  \u2009                    \u2009           \u2009                  \u2009                 \u2009                    \u2009\n  Met / IDF            \u2009           \u2009                  \u2009                 \u2009                    \u2009\n  0                    16(6.7)     8(10.4)            8(5.0)            1                    \u2009\n  1                    71(30.3)    22(28.6)           49(30.8)          0.5(0.2-1.4)         0.159\n  2                    78(32.7)    22(28.6)           56(35.2)          0.4(0.1-1.12)        0.138\n  \u22653                   71(30.3)    25(32.4)           46(28.9)          0.6(0.2-1.6)         0.393\n  \u2009                    \u2009           \u2009                  \u2009                 \u2009                    \u2009\n  Met / WHO            \u2009           \u2009                  \u2009                 \u2009                    \u2009\n  0                    10(8.4)     8(10.4)            12(7.5)           1                    \u2009\n  1                    38(31.9)    22(28.6)           54(33.9)          0.6(0.2-1.7)         0.418\n  2                    40(33.6)    24(31.2)           55(34.6)          0.7(0.2-1.8)         0.431\n  \u22653                   31(26.1)    24(31.2)           38(23.9)          0.9(0.3-2.7)         1.000\n\nNCEP ATP III: National Cholesterol Education Program, Adult Treatment Panel III; IDF: International Diabetes Federation; WHO: World Health Organization; MetS: metabolic syndrome; OR: odds ratio; CI: confidence interval.\n\n[^1]: Academic Editor: Hugo Schnack\n"}
+{"text": "Introduction {#s1}\n============\n\nThere is an ongoing need for novel drugs that are highly effective in the treatment of cancer, drug resistant bacteria, fungal infections, emerging viruses and parasitic protozoan infections. Historically, natural products have provided the basis for the majority of new drugs and the bioactive properties of a wide variety of flora is reflected in their continued roles in the traditional healthcare systems of many cultures [@pone.0035953-Newman1]. The successful use of plants in traditional medicine and modern natural products research has meant that there is a renewed interest in exploiting various aspects of the underlying bioactivities.\n\nStill widely practiced in the modern era, Traditional Chinese Medicine (TCM) can trace its origins back thousands of years, to the dawn of civilisation in China. TCM theory is based on experiences of the effects of the medicines through documented trials to establish knowledge of the use of plants (approximately 5000 species) for the treatment of many diseases [@pone.0035953-Zhu1]. This TCM background has provided the basis for the discovery of several therapeutic agents, including the anticancer agents indirubin [@pone.0035953-Hoessel1], camptothecin [@pone.0035953-Wall1] and harringtonine [@pone.0035953-Perdue1]; aretemisinin (antimalarial) [@pone.0035953-Acton1] and ephedra (central nervous system stimulant) [@pone.0035953-Chen1].\n\nCompounds isolated from plant preparations are largely the products of plant metabolism, however, microorganisms living in symbiosis with plants also produce bioactive compounds. Taxol and camptothecin are examples of anticancer compounds synthesized by both plants and endophytes [@pone.0035953-Puri1], [@pone.0035953-Stierle1]. In their natural environment the internal tissues of plants are frequently colonized by numerous different microorganisms, termed endophytes. Many endophytes produce secondary metabolites which confer major ecological benefits to their host plants including plant growth promotion [@pone.0035953-Compant1], enhanced resistance to various predators and pathogens [@pone.0035953-Arnold1], [@pone.0035953-Mejia1], and increased drought resistance [@pone.0035953-Kannadan1]. Consequently, endophyte derived metabolites may also cause the observed bioactivity and associated beneficial health claims of the TCM host plants [@pone.0035953-Strobel1].\n\nLow molecular weight endophytic secondary metabolites demonstrate a large degree of structural diversity with the largest and most important groups of compounds including the polyketides, amino acid derived compounds, and terpenes [@pone.0035953-Keller1]. To date genetic methods have been used to screen for biosynthetic pathways involved in secondary metabolism. Genetic screening for microbial natural product genes has largely focused on the detection of the polyketide and non-ribosomal peptide synthesis pathways [@pone.0035953-Cox1], [@pone.0035953-Mankelow1].\n\nPolyketides are produced by many fungi, bacteria, plants and marine organisms [@pone.0035953-Keller1]. This large family of structurally diverse natural products have already found widespread application as pharmaceuticals including rapamycin (immunosuppressant), erythromycin (antibiotic), lovastatin (anticholesterol drug), and epothilone B (anticancer drug). Polyketides are biosynthesized by large multimodular enzyme complexes termed polyketide synthases (PKSs) which catalyse the polymerisation of acyl-CoA thioester building blocks [@pone.0035953-Fischbach1]. PKSs are typically categorized based on their number of subunits (single or multiple) and their mode of synthesis (iterative or modular). The best characterized group of PKSs are the type I PKSs with a single enzyme complex responsible for the biosynthesis of the polyketide backbone. They can be either modular as found in bacteria, or iterative as found in fungi [@pone.0035953-Staunton1]. Extension of the polyketide backbone requires three core PKS domains: an acyltransferase (AT) domain, acyl carrier proteins (ACP) domain and a ketosynthase (KS) domain. The modular structure of type I PKSs and their assembly of polyketides, resembles the organisation of non-ribosomal peptide synthetases (NRPSs) and the production of non-ribosomal peptides.\n\nComplex oligopeptides are formed by NRPSs through the linear condensation of proteinogenic and non-proteinogenic amino acids. NRPSs are involved in the biosynthesis of compounds including penicillin and cephalosporin (antibiotics), cyclosporine A (anti-inflammatory and immunosuppressive activities, and the endophyte-produced fusaricidin (antibiotic). NRPSs are composed of a series of modules, each containing the catalytic units required for the recognition, activation and peptide bond formation of the growing peptide chain by adenylation (A), thiolation (T), and condensation (C) domains, respectively [@pone.0035953-Mootz1]. NRPSs are found in bacteria and fungi feature the three core domains arranged in the order C-A-T [@pone.0035953-Mootz1].\n\nWhile mechanistically distinct, the highly conserved regions within the KS and A domains are logical targets for the detection of PKS and NRPS biosynthesis genes, respectively. This facilitates the design of degenerate oligonucleotide primers, suitable for genetic screening of secondary metabolism pathways [@pone.0035953-Sauer1], [@pone.0035953-Burns1], [@pone.0035953-Ehrenreich1].\n\nThe location of host plants can affect endophyte populations [@pone.0035953-Gore1], [@pone.0035953-Hoffman1] and for drug discovery potential it is imperative to investigate TCM plants sampled from their natural environment or those cultivated for medicinal purposes. Considering this, plant collections from Yunnan province were selected for two main reasons. Firstly, Yunnan is recognized for its large botanical diversity [@pone.0035953-Xiwen1] and secondly, it is a source of many pharmaceutical herbs. TCM herbs with reported anticancer activity were chosen in an effort to increase the likelihood of discovering endophyte produced anticancer compounds [@pone.0035953-Zhu1].\n\nThe aim of this study was to assess the use of genetic screening for the identification of plant leads for endophyte natural product investigations. The biosynthetic potential of endophytes was considered via the detection of endophyte PKS and NRPS biosynthesis genes. Total DNA from thirty selected anticancer TCM herbs was genetically screened using four PCR primers sets. This study has identified a subset of TCM herbs found to harbor endophytes with bioactive potential, and these plants can be targeted in future investigations. The use of this genetic screen increases the likelihood of isolating endophytes which biosynthesize novel pharmaceuticals independently of the host plant.\n\nMethods {#s2}\n=======\n\nSelection, collection and storage of TCM herbs {#s2a}\n----------------------------------------------\n\nYunnan province is situated in the southwestern region of the People\\'s Republic of China. There is a strong temperature gradient across the province and it contains areas of permanent snow as well as rainforests with temperatures perennially above 10\u00b0C [@pone.0035953-Xiwen1]. Thirty plant tissue samples ([Table 1](#pone-0035953-t001){ref-type=\"table\"}) were collected during spring in April.2007 from sites within and surrounding Kunming City. No specific permits were required, however, the collections were acknowledged and supported by the Yunnan Provincial government. Potential anticancer TCM herbs were selected based on scientific and traditional literature, their maturity for use as TCMs and their availability for collection at the time of collection. Following collection, samples were placed in sterile bags and stored in the dark at 4\u00b0C until processed. Processing was completed as soon as possible following receipt of samples, usually within 24 h.\n\n10.1371/journal.pone.0035953.t001\n\n###### Traditional Chinese medicinal herbs collected for genetic screening.\n\n![](pone.0035953.t001){#pone-0035953-t001-1}\n\n  Host Plant Species                                                Location           Plant Sample                      Pharmacological or TCM Activity                              Reference\n  ---------------------------------------------------------- ----------------------- ----------------- -------------------------------------------------------------------- ------------------------------\n  *Aconitum carmichaeli* Debx.                                     Countryside         Leaf and root         Sedative, analgesic, antiinflammatory, anticancer, toxic            [@pone.0035953-Li1]\n  *Actinidia chinensis* Planch.                                Medicinal herb farm     Root and stem               \"Heat clearing\\\", antimutagenic, anticancer                   [@pone.0035953-Lee1]\n  *Agrimonia pilosa* Ledeb. var. japonica (miq.) Nakai         Medicinal herb farm     Above ground                Haemostatic, antiinflammatory, antiinfective                  [@pone.0035953-Zhu1]\n  *Artemisia capillaris* Thunb.                               Medicinal herb garden    Leaf and stem    \"Heat clearing\\\", antihepatotoxic, antibacterial, antiinflammatory       [@pone.0035953-Zhu1]\n  *Astilbe rivularis* Buch.-Ham.                              Medicinal herb garden     Whole plant                                 Antiviral                                [@pone.0035953-Rajbhandari1]\n  *Belamcanda chinensis* (L.) DC.                               Medical herb farm          Root                 Antifungal, antiviral, antiinflammatory, diuretic                [@pone.0035953-Zhu1]\n  *Bletilla striata* (Thunb.) Reichp.                           Medical herb farm        Root/bulb                           Haemostatic, anticancer                             [@pone.0035953-Zhu1]\n  *Cephalotaxus fortunei* Hook. F.                                 Countryside         Leaf and stem                                Anticancer                                  [@pone.0035953-Huang1]\n  *Clematis hexapetala* Pall.                                  Medicinal herb farm         Root                             Antibacterial, antifungal                            [@pone.0035953-Zhu1]\n  *Coptis chinensis* Franch.                                   Medicinal herb farm         Root          Antibacterial, antiinflammatory, affects central nervous system         [@pone.0035953-Zhu1]\n  *Dendrobium nobile*, Lindl.                                      Countryside             Stem                             Anticancer, antimutagenic                         [@pone.0035953-Miyazawa1]\n  *Digitalis purpurea* Ehrh.                                  Medicinal herb garden        Root                              Cardiotonic, anticancer                         [@pone.0035953-LopezLazaro1]\n  *Eleutherococcus senticosus* (Rupr. & Maxim) Maxim           Medicinal herb farm         Root                              Anticancer, expectorant                           [@pone.0035953-Davydov1]\n  *Leonurus heterophyllus* Sweet                               Medicinal herb farm     Above ground                      Anticancer, \"invigorates blood\\\"                       [@pone.0035953-Huang1]\n  *Lithospermum erythrorhizon* Sieb.et Zucc.                   Medicinal herb farm         Root                    Anticancer, antibacterial, antiinflammatory                   [@pone.0035953-Zhu1]\n  *Lonicera japonica* Thunb.                                       Countryside             Leaf                                   Antiinfective                                  [@pone.0035953-Zhu1]\n  *Lycium chinense* Mill                                      Medicinal herb garden   Branch and leaf                     Anticancer, immune stimulating                        [@pone.0035953-Huang1]\n  *Pinellia ternata* (Thunb.) Breit.                               Countryside             Root             Anti-emetic, mucolytic, anticancer (cervical), discutient            [@pone.0035953-Li1]\n  *Pinellia pedatisecta* Schott                                Medicinal herb farm         Root                                 Toxic, anticancer                                [@pone.0035953-Lee1]\n  *Paris polyphylla* Smith var. *chinensis* (Franch.) Hara     Medicinal herb farm     Root and leaf                 Antiinflammatory, detoxicant, anticancer                    [@pone.0035953-Lee1]\n  *Rheum palmatum* L.                                          Medicinal herb farm         Root                 Anticancer, detoxicant, antiinflammatory, laxative               [@pone.0035953-Zhu1]\n  *Rubia cordifolia* L.                                            Countryside         Leaf and root                       Antiinflammatory, antiviral                           [@pone.0035953-Ho1]\n  *Salvia przewalskii* Maxim                                   Medicinal herb farm         Root                    Anticancer, immune stimulating, antioxidant                   [@pone.0035953-Lu1]\n  *Salvia yunnanensis* C.H. Wright                             Medicinal herb farm         Root                           Anticancer, immune stimulating                         [@pone.0035953-Lu1]\n  *Senecio scandens* Buch.-Ham. Ex D.                              Countryside         Leaf and root         Antiinflammatory, anticancer, detoxicant, antibacterial            [@pone.0035953-Huang1]\n  *Solanum nigrum* L.                                              Countryside         Leaf and root                              Antimicrobial                                 [@pone.0035953-Huang1]\n  *Sophora davidii* (Franch) Kom. Ex Pavol.                   Medicinal herb garden    Leaf and root                       Anticancer, antiinflammatory                       [@pone.0035953-NiceData1]\n  *Sophora japonica* L.                                            Countryside         Leaf and stem                                Hemostatic                                  [@pone.0035953-Huang1]\n  *Taxus baccata* L.                                               Countryside         Leaf and stem                                Anticancer                                  [@pone.0035953-Wani1]\n  *Thalictrum foliolosum* DC                                  Medicinal herb garden    Leaf and root               Anticancer, antiinfective, antihypertensive                [@pone.0035953-NiceData1]\n\nSurface sterilisation of plant samples {#s2b}\n--------------------------------------\n\nSections of each plant were carefully washed in running water to remove external soil and debris. Samples were examined to exclude those that showed symptoms of disease or superficial damage. In a method adapted from Cankar et al. [@pone.0035953-Cankar1], plant material was surface sterilized by immersion in 30% hydrogen peroxide for 30 min, followed by rinsing in sterile Milli-Q filtered water (Millipore) and washing with 70% ethanol for 1 min [@pone.0035953-Cankar1]. To confirm that the surface sterilisation process was successful, a section of each sterilized plant was placed into an aliquot of heart infusion broth and incubated at 30\u00b0C for three days before examining the media for any growth of plant surface-associated contaminating microorganisms.\n\nDNA extraction {#s2c}\n--------------\n\nThe different tissue types collected from the one plant were processed together for the total DNA extract. Total genomic DNA was extracted from the surface sterilized plant samples using the Power Plant DNA isolation kit (MoBio Laboratories). Extracted DNA consisted of host plant nuclear and plastid DNA and any endophyte microbial DNA. Extracted DNA was run on an agarose gel and NanoDrop (Thermo Scientific) readings taken to assess concentration and purity. Working stocks of DNA extracts were diluted to 1--10 ng/\u03bcL for PCR assays.\n\nPolymerase Chain Reaction {#s2d}\n-------------------------\n\nDNA extracts from TCM plants were screened using PCR primers that targeted various plant, bacterial and fungal genes. The PCR targets and primers employed in this investigation (NRPS, PKS, and the plant RUBISCO genes) are listed in [Table 2](#pone-0035953-t002){ref-type=\"table\"}. Typical gene-specific PCR reactions were carried out in 20 \u00b5l reaction volumes that contained 1\u00d7reaction buffer (Bioline), 2.5 mM MgCl~2~, 0.2 mM dNTPs (Bioline), 10 pmol each of forward and reverse primers (Sigma-Aldrich, MO, USA), 0.2 U *Taq* DNA polymerase (Bioline) and 1--10 ng of DNA template. Thermal cycling was performed in a MyCycler Thermal Cycler (BioRad), with an initial denaturation at 94\u00b0C for 5 min, followed by 30 cycles of 94\u00b0C for 30 s, 55\u00b0C for 30 s, and 72\u00b0C for 1 min, followed by a final extension step at 72\u00b0C for 7 min. Degenerate PCRs were also performed in 20 \u03bcL reaction volumes containing the same reaction mix as for the non-degenerate primers except that 25 pmol of forward and reverse primers were used and the number of amplification cycles was increased to 35. PCR amplicons were separated by agarose gel electrophoresis in TAE buffer (40 mM Tris-acetate, 1 mM EDTA, pH 7.8), visualized by ethidium bromide staining (0.5 \u00b5g /ml) and the GelDoc UV transilluminator (BioRad).\n\n10.1371/journal.pone.0035953.t002\n\n###### PCR primers and reaction conditions used in the screening of total DNA extracts from TCM plants.\n\n![](pone.0035953.t002){#pone-0035953-t002-2}\n\n  Target                                             Primer      Primer Sequence (5\u2032\u21923\u2032)      Product size (bp)   Altered[\\*](#nt101){ref-type=\"table-fn\"} reaction conditions         Altered[\\*](#nt101){ref-type=\"table-fn\"} thermocycling conditions                   Reference\n  ------------------------------------------------- --------- ------------------------------ ------------------- -------------------------------------------------------------- -------------------------------------------------------------------------------- ------------------------------\n  Plant RUBISCO enzyme                               rbcL-F       tgtcaccacaaacagarackaa            \u223c1400                                                                                                                                                         [@pone.0035953-vanderBank1]\n                                                     rbcL-R    caaaattaaatmsgatctctttccatac                                                                                                                                                                      \n  Adenylation domain, Bacterial NRPSs                 MTF2         gcnggyggygcntaygtncc             \u223c1000                                                                                         94\u00b0C for 10 s, 52\u00b0C for 30 s, 72\u00b0C for 30 s                       [@pone.0035953-Neilan1]\n                                                      MTR2           ccncgdayttnacytg                                                                                                                                                                            \n  Ketosynthase domain, Bacterial PKSs (Type 1)         DKF         gtgccggtnccrtgngyytc           \u223c650--700                                                                                                                                                         [@pone.0035953-Moffitt1]\n                                                       DKR         gcgatggayccncarcarmg                                                                                                                                                                          \n  Adenylation domain, Fungal NRPSs                   RJ016-R        arrtcnccngtyttrta               \u223c300                                                                         94\u00b0C for 30 s, 50\u00b0C for 30 s, 60\u00b0C for 1 min; final extension: 60\u00b0C for 10 min     [@pone.0035953-Johnson1]\n                                                     RJ016-F          tayggnccnacnga                                                                                                                                                                             \n  Ketosynthase domain, Fungal PKSs (Type 1)            LC1        gayccimgittyttyaayatg             \u223c700              Primers (pmol): LC1, 50; LC2c, 8; LC3, 24; LC5c, 16                         94\u00b0C for 30 s, 56\u00b0C for 10 s, 72\u00b0C for 30 s                       [@pone.0035953-Bingle1]\n                                                      LC2c         gticcigticcrtgcatytc                                                                                                                                                                          \n                                                       LC3         gcigarcaratggayccica             \u223c700                                                                                                                                                         \n                                                      LC5c         gtigaigticrtgigcytc                                                                                                                                                                           \n  Adenylation domain, Streptomycete NRPSs              A3F       gcstacsysatstacacstcsgg            \u223c700                  10 pmol each primer, 0.25 mM dNTPs, 10% DMSO                           94\u00b0C for 30 s, 58\u00b0C for 1 min, 72\u00b0C for 2 min                    [@pone.0035953-AyusoSacido1]\n                                                       A7R         sasgtcvccsgtscggtas                                                                                                                                                                           \n  Ketosynthase domain, Streptomycete PKS (Type 1)      K1F         tsaagtcsaacatcggbca           \u223c1200\u22121400               20 pmol each primer, 0.25 mM dNTPs, 10% DMSO                             94\u00b0C for 30s, 55\u00b0C for min, 72\u00b0C for 3 min                     [@pone.0035953-AyusoSacido1]\n                                                       M6R         cgcaggttscsgtaccagta                                                                                                                                                                          \n\nAltered from standard PCR procedures described.\n\nA spiked PCR targeting the conserved plant RUBISCO gene was used to ensure the absence of PCR inhibitors in the DNA samples. This procedure involved two PCRs per DNA sample, the first containing both sample DNA and an inhibitor free control DNA and the second reaction contained only sample DNA. Lack of amplification in control-spiked reactions was likely to be due to PCR inhibitors in the sample DNA. In this event, DNA was diluted to an extent where the PCR reaction was not inhibited. The second PCR reaction, containing only sample DNA, was used to ensure that DNA was not diluted beyond the PCR detection limits.\n\nDNA sequencing {#s2e}\n--------------\n\nPCR amplicons, of the correct size, were purified directly by ethanol precipitation or gel purification using the UltraClean GelSpin DNA Extraction Kit (MoBio Laboratories) in accordance with the manufacturer\\'s instructions.\n\nAmplicons were sequenced using the PRISM BigDye\u2122 cycle sequencing system v3.1 on an ABI PRISM 373 DNA Sequencer (Life Technologies, Carlsbad, USA). PCR products that contained multiple sequences, as observed by multiple peaks in the sequencing chromatograms, were cloned using the TOPO cloning vector (Invitrogen Life Technologies, Carlsbad, USA), transformed into *Escherichia coli* DH5\u03b1 and a clone library established according to the manufacturer\\'s instructions. Thereafter, colony PCR screening and DNA sequencing of cloned inserts was performed to characterize single amplicons. Sequences were deposited into GenBank.\n\nPhylogenetic analysis of KS and NRPS fragments {#s2f}\n----------------------------------------------\n\nSequence data were compared to the NCBI dataset on GenBank with the BLASTX algorithm (<http://blast.ncbi.nlm.nih.gov/Blast.cgi>). The amino acid substrates recognized by the A domain binding pockets were predicted using the NRPSpredictor analysis tool available at (<http://www-ab.informatik.uni-tuebingen.de/toolbox/index.php?view=domainpred>) [@pone.0035953-Rausch1]. Multiple sequence alignments of KS and A domain protein sequences of PKS or NRPS pathways, respectively, were generated using ClustalX [@pone.0035953-Thompson1] ([Figure S1](#pone.0035953.s001){ref-type=\"supplementary-material\"}). Gaps and positions with ambiguities were excluded from the phylogenetic analysis. The ProtTest program was used to determine the most appropriate substitution model for the dataset [@pone.0035953-Abascal1]. Phylogenetic analysis was performed using the maximum likelihood methods [@pone.0035953-Guindon1]. Node support values were determined by the approximate Likelihood-Ratio Test (aLRT).\n\nResults {#s3}\n=======\n\nIdentification of genes encoding bioactivity in TCM plants {#s3a}\n----------------------------------------------------------\n\nThe total DNA extracted from thirty TCM plants was expected to contain a mixture of plant and endophyte genes. The DNAs were free from PCR inhibitors, as evidenced by successful PCR amplification of the plant RUBISCO gene. Degenerate PCRs were used to detect putative NRPS and PKS gene sequences originating from bacterial and fungal endophytes in the DNA extracts. Specifically, the PCR screens were successful in targeting endophyte sequences with four streptomycete KS, six streptomycete A, ten fungal KS and three fungal A domain genes detected ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). Some of the amplified PKS and NRPS PCR products could be directly sequenced, however, many contained a mixture of amplicons that needed to be cloned into a TOPO cloning vector. In such cases, at least ten transformants were randomly selected for direct colony PCR and the products sequenced. Amplification of bacterial and fungal A domains and KS domains was confirmed via sequencing and BLASTX (translated) analysis ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). BLASTN (nucleotide) analysis of the PCR products did not reveal any significant matches to the nucleotide database, indicating that the sequences were novel at this level. Endophyte-derived genes were amplified from DNA extracted from different types of plant tissues including the roots, leaves, stems and bark ([Table 1](#pone-0035953-t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0035953.t003\n\n###### PKS and NRPS genes identified with degenerate PCR primers.\n\n![](pone.0035953.t003){#pone-0035953-t003-3}\n\n  Gene                    Plant          Sequence identifier                           BLASTX Match                            Identity (%)    Accession number   Predicted binding pocket (amino acid substrate)[1,2](#nt102){ref-type=\"table-fn\"}\n  ---------------- -------------------- --------------------- -------------------------------------------------------------- ---------------- ------------------ -----------------------------------------------------------------------------------\n  Fungal PKS          *Bl. striata*              BS1                      Elsinochrome PKS, *Elsinoe fawcettii*               124/161 (77%)        ABU63483                                              ND\n  Fungal PKS          *C. fortunei*              CF1              PKS involved in melanin production, *Bipolaris oryzae*      189/210 (90%)        BAD22832                                              ND\n  Fungal PKS        *L. heterophyllus*           LH1           non-reduced type PKS for melanin pigment, *Ascochyta rabiei*   232/236 (98%)        ACS74449                                              ND\n  Fungal PKS        *L. heterophyllus*           LH2                  Uncharacterized PKS, *Botryotinia fuckeliana*           221/221 (100%)       AAR90249                                              ND\n  Fungal PKS          *Lo. japonica*             LJ1                 Putative non-reduced type PKS, *Penicillium* sp.         193/207 (93%)        ABQ85550                                              ND\n  Fungal PKS           *P. ternata*              PT1                 Conidial pigment PKS, *Verticillium albo-atrum*          197/211 (93%)        EEY14472                                              ND\n  Fungal PKS          *S. scandens*              SS1                            PKS 1, *Glarea lozoyensis*                    198/220 (90%)        AAN59953                                              ND\n  Fungal PKS          *S. scandens*              SS2                     Putative melanin PKS, *Penicillium* sp.              205/243 (98%)        ACJ13039                                              ND\n  Fungal PKS          *S. scandens*              SS3                            PKS 1, *Glarea lozoyensis*                    205/243 (84%)        AAN59953                                              ND\n  Fungal PKS           *T. baccata*              TB1                      Elsinochrome PKS, *Elsinoe fawcettii*               205/243 (84%)        ABU63483                                              ND\n  Fungal NRPS       *Pi. pedatisecta*           PPe1                    MicC synthetase, *Planktothrix rubescens*              20/34 (58%)         CAQ48260                                         No prediction\n  Fungal NRPS       *Pi. pedatisecta*           PPe2                    AerB synthetase, *Planktothrix rubescens*              56/95 (58%)         CAQ48266                                         No prediction\n  Fungal NRPS       *Pi. pedatisecta*           PPe3                  Arthrofactin synthetase, *Burkholderia ghumae*           65/95 (68%)       YP_002908545                                       No prediction\n  Bacterial PKS       *D. purpurea*              DL1                Nystatin synthase (NysC) *Streptomyces halstedii*         103/174 (59%)        AAF71776                                              ND\n  Bacterial PKS      *Pa. polyphylla*           PPol1                   NapC synthase *Streptomyces hygroscopicus*            157/218 (72%)        ABB86421                                              ND\n  Bacterial PKS      *Pa. polyphylla*           PPol2                   NapC synthase *Streptomyces hygroscopicus*            166/235 (70%)        ABB86421                                              ND\n  Bacterial PKS     *Pi. pedatisecta*           PPe4                 Arthrofactin synthetase, *Pseudomonas sp*. MIS38          47/63 (74%)         BAC67535                                              ND\n  Bacterial NRPS      *D. purpurea*              DL2            Peptide synthetase PhsB, *Streptomyces viridochromogenes*      83/130 (63%)        CAJ14037                                       DVEHLSLID- (pro)\n  Bacterial NRPS      *D. purpurea*              DL3                    Ansamitocin PKS, *Actinosynnema pretiosum*            121/231 (52%)        AAM54075                               -AFALACGM- (val/leu/ile/abu/iva)\n  Bacterial NRPS    *L. heterophyllus*           LH3                Amino acid adenylation domain *Acidovorax avenae*          18/19 (94%)        YP_972054                                         DVWNIGLI(thr)\n  Bacterial NRPS    *L. heterophyllus*           LH4                 Pyoverdine synthetase, *Pseudomonas fluorescens*         154/181 (85%)        AAF40219                                        CVWHFGRI (glu)\n  Bacterial NRPS      *B. chinensis*             BC1            Amino acid adenylation domain, *Rhodococcus erythropolis*     192/195 (98%)      ZP_04387369                                   DATFAGGI (leu/ile/val)\n  Bacterial NRPS      *B. chinensis*             BC2            Amino acid adenylation domain, *Rhodococcus erythropolis*     224/228 (98%)      ZP_04387369                                   DATFAGGI (leu/ile/val)\n\nIn silico prediction of the amino acid substrate recognized by putative NRPS fragments [@pone.0035953-Rausch1].\n\nND\u200a=\u200a not done.\n\nTen unique fungal KS domain sequences (\u223c700 bp) were amplified from total DNA extracts of *Bl. striata*, *C. fortunei*, *L. heterophyllus*, *Lo. japonica*, *P. ternata*, *S. scandens,* and *T. baccata* ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). The amplified sequences possessed between 76% and 100% amino acid identity to known fungal PKSs. Specifically, sequence LH2 shared 100% sequence identity with an uncharacterized PKS from *Botryotinia fuckeliana.* Other fungal PKS amplicons were similar to genes involved in the production of polyketides, such as melanin and the toxic pigment elsinochrome.\n\nScreening of total TCM plant DNA successfully amplified \u223c300 bp fragments of the fungal NRPS A domain. Of all the DNAs screened only DNA from *Pi. pedatisecta* produced NRPS gene fragments (PPe1, PPe2, PPe3). BLASTX analysis revealed that the predicted translated DNA sequences were homologous to bacterial NRPSs, with sequence similarity between 58% and 68%. Prediction of the eight amino acids that form the A domain active site, and the amino acids that bind to the active residues, was not successful possibly indicating that the algorithm is not sufficiently robust to predict fungal NRPS module specificity [@pone.0035953-Keller2].\n\nBacterial directed primers were also used to screen for PKS and NRPS genes [@pone.0035953-AyusoSacido1]. Degenerate primers targeting the genes coding for the KS domain of actinobacterial PKSs amplified a total of four unique amplicons from the plant DNA extracts, including *D. purpurea* (one), *P. polyphylla* (two), and *P. peditasecta* (one). Predicted translation of the gene fragments was afforded through the BLAST program and subsequent peptide sequence analysis established that three of the fragments were similar to sequences of PKSs involved in the biosynthesis of the antifungal compound nystatin (DL1, 59%) and the anticancer and antibiotic compound geldanamycin (PPo1, 72% and PPo2, 70%). The sequence of PPe4 was shorter (\u223c235 bp) and had 74% similarity to arthrofactin synthetase from *Pseudomonas* sp. MIS38 ([Table 3](#pone-0035953-t003){ref-type=\"table\"}).\n\nScreening of the TCM plant DNA extracts for bacterial NRPS A domain gene sequences revealed six unique amplicons from the extracts of *D. purpurea*, *L. heterophyllus* and *B. chinense*. The translated sequences of five of these amplicons showed high sequence similarity, 57--98%, to bacterial NRPS synthetases ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). Importantly, amplicons BC1 and BC2 showed significant similarity (98%) to an adenylation domain of an uncharacterized *Rhodococcus erythropolis* NRPS. Of relevance to drug discovery, the product amplified from *D. purpurea* (DL3) had 65% sequence similarity to the ansamitocin synthase from *Actinosynnema pretiosum*. Unlike the PKS amplicons, which were all similar to published *Streptomyces* sp. sequences, the NRPS PCR amplicons were similar to sequences of diverse bacteria, including the Proteobacteria and different genera of the Actinobacteria group ([Table 3](#pone-0035953-t003){ref-type=\"table\"}).\n\nThe deduced translated A domains sequences were analysed to determine if they contained an eight residue binding pocket, and the amino acid that was likely to be bound by the pocket. Predictions were successful for the bacterial NRPS sequences including fragments BC1 and BC2 that are likely to possess the same amino acid binding pocket, and could bind either a Leu, Ile or Val residue ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). Analysis of the two fragments detected from the endophytes of *L. heterophyllus*, LH3 and LH4, were shown to possess different binding pockets and thus would be involved in the addition of different amino acids to the growing peptide chain produced by the NRPS ([Table 3](#pone-0035953-t003){ref-type=\"table\"}).\n\nPhylogenetic distribution of PKS and NRPS genes from TCM endophytes {#s3b}\n-------------------------------------------------------------------\n\nIn an attempt to investigate the homology and novelty of amplicons from the genetic screen, the deduced amino acid sequence of the KS and A domain gene fragments were used in multiple sequence alignments and the phylogenies reconstructed. The unrooted phylogenetic tree of fungal PKSs ([Figure 1](#pone-0035953-g001){ref-type=\"fig\"}) revealed the presence of two main clades that correspond to the two largest structural classes of fungal polyketides: reduced and unreduced. All fungal KS domain sequences amplified in this study belonged to the clade of PKSs which synthesize unreduced PKs. The proteins were closely grouped to PKSs involved in the biosynthesis pathways of various pigments including melanin (LH1, PT1), other pigments from *Penicillium* sp. and *Elsinoe fawcetti* (CF1, LJ1, SS2), or putative PKSs derived from endophytes and other fungal symbionts (BS1, LH2, SS1, SS3, PT1). The gene products of SS1 and SS3 formed a highly supported clade and are possible homologs of the same biosynthetic pathway.\n\n![Phylogenetic analysis of putative fungal type I PKSs in TCM herbs.\\\nEvolutionary relationships were determined by maximum likelihood analysis using the LG substitution model. Branch lengths indicate inferred divergence of amino acid sequences. Numbers adjacent to the nodes indicate aLRT support, with support values \\>50% considered to be significant. Accession numbers for sequences obtained from GenBank are indicated. The scale bar represents 0.2 amino acid changes.](pone.0035953.g001){#pone-0035953-g001}\n\nPhylogenetic analysis was performed on the inferred amino acid sequences of NRPS A domains ([Figure 2](#pone-0035953-g002){ref-type=\"fig\"}). The three NRPS A domains sequences identified in this study were amplified from the extract of *Pi. pedatisecta* and although each sequence was unique, the sequences (PPe1, PPe2 and PPe3) were clustered in the phylogenetic tree indicating that they most probably have a common phylogenetic ancestor and/or similar function. This cluster of fragments also grouped with cyanobacterial NRPSs. Related NRPSs included those involved in the synthesis of a range of bioactive compounds; MicC in the synthesis of the pentapeptide microginin in *Planktothrix rubescens* [@pone.0035953-Rounge1], AerB in the synthesis of the aeruginosin protease inhibitors in *Microcystis viridis* [@pone.0035953-Ishida1], and OciA in the synthesis of cyanopeptolin [@pone.0035953-Rounge2]. Despite the fact that fungal specific PCR primers were used, these A domain sequences had greater similarity to NRPSs of bacterial origin with bacterial and fungal NRPSs clustered separately in the phylogenetic tree ([Figure 2](#pone-0035953-g002){ref-type=\"fig\"}).\n\n![Phylogeny of fungal NRPSs.\\\nRelationships inferred by maximum likelihood analysis of fungal NRPS adenylation domain fragments using the WAG substitution model. Branch lengths indicate inferred divergence of amino acid sequences. Numbers adjacent to nodes indicate aLRT support, with support values \\>50% considered significant. Bacterial and fungal reference sequences are included in the analysis and the accession numbers for these sequences obtained from GenBank are indicated. The scale bar represents 0.2 amino acid changes.](pone.0035953.g002){#pone-0035953-g002}\n\nPhylogenetic analysis of inferred bacterial KS domains revealed that the endophyte sequences clustered with KS domains of Actinobacteria ([Figure 3](#pone-0035953-g003){ref-type=\"fig\"}). Sequences PPol1 and PPol2 were closely related to each other and also clustered with PKS fragments involved in the biosynthesis of the anticancer antibiotic geldanamycin, from *Streptomyces hygroscopicus* [@pone.0035953-Rascher1]. The DL1 sequence clustered within a clade containing KS domains, involved in the tailoring of geldanamycin and herbimycin in *S. hygroscopicus,* and an uncharacterized KS domain from *Pseudonocardia autotrophica*. The sequence PPe4 was not included in the bacterial PKS multiple sequence alignment and phylogenetic tree, because it was significantly shorter than the other fragments (\u223c80 amino acids).\n\n![Phylogeny of bacterial type I PKSs.\\\nRelationships inferred by maximum likelihood analysis of bacterial PKS ketosynthase domain fragments, using the LG substitution model. Branch length indicates inferred divergence of amino acid sequences and numbers adjacent to the nodes indicate aLRT support, with support values \\>50% considered significant. The accession numbers for all sequences obtained from GenBank are indicated. The scale bar represents 0.2 amino acid changes.](pone.0035953.g003){#pone-0035953-g003}\n\nThe phylogenetic tree containing bacterial NRPS A domain sequences demonstrated two main clusters from the Actinobacteria and Proteobacteria ([Figure 4](#pone-0035953-g004){ref-type=\"fig\"}). Sequences BC1 and BC2 were closely related and clustered with an uncharacterized NRPS biosynthesis pathway from *Rhodococcus erythropolis*. Although the function of the *R. erythropolis* NRPS is unknown, it also grouped with A domain sequences from other Actinobacteria. Sequences LH3 and LH4 were related to NRPS sequences from the Proteobacteria, specifically *Pseudomonas fluorescens* and *Acidovorax avenae*, respectively. The sequences DL3 grouped with the clade of bacterial A domain sequences, however, had low homology to known NRPSs.\n\n![Phylogeny of bacterial NRPS sequences.\\\nRelationships inferred by maximum likelihood analysis of bacterial NRPS adenylation domain fragments using the LG substitution model. Branch lengths indicate inferred divergence of amino acid sequences. Numbers adjacent to the nodes indicate aLRT support, with support values \\>50% considered significant. The accession numbers for all sequences obtained from GenBank are indicated. The scale bar represents 0.5 amino acid changes.](pone.0035953.g004){#pone-0035953-g004}\n\nDiscussion {#s4}\n==========\n\nNatural product drug discovery investigations are supported by a vast number of candidate systems and environments available for study, presenting the necessity to optimise the targets chosen for investigation. Ethnopharmocological approaches have been affirmed by the discovery of novel bioactive compounds from traditional medicinal plants [@pone.0035953-Hoessel1], [@pone.0035953-Acton1], [@pone.0035953-Castillo1]. Additionally endophytes are capable of producing a wide variety of natural products, contributing to the overall bioactivity of the host plant. Thus, they are considered to be a valuable source of novel bioactive compounds and an avenue to increase the breadth of natural product studies [@pone.0035953-Strobel2]. The present study took advantage of the immense knowledge of TCM plants to focus on plants of pharmaceutical interest. This preliminary study screened the metabolic potential of endophytes using PKS and NRPS genes as a proxy for endophytic production of bioactive compounds. The use of these genes reflects the noted involvement of PKSs and NRPSs in the production of many natural products [@pone.0035953-Sauer1], [@pone.0035953-Wawrik1], [@pone.0035953-Moffitt1]. This screen has the potential to be applied to the endophytes of other TCM plant species or growth climates.\n\nPKS and NRPS genes are appropriate targets for detection of small molecule biosynthesis systems [@pone.0035953-Sauer1], [@pone.0035953-Wawrik1], [@pone.0035953-Moffitt1]. Genetic screening results demonstrated the presence of at least one NRPS/PKS biosynthetic pathway in 36% (11/30) of the plant DNA extracts. Furthermore, PKS genes were detected more frequently in the total DNA samples (30%, 9/30) than NRPS genes (13%, 4/30). PKS and NRPS biosynthesis genes have previously been detected in fungal and bacterial isolates from ecological niches similar to those inhabited by endophytes. Fungal and bacterial isolates of sponges (18% and 13.8%, respectively) and marine (90%) and freshwater (65%) cyanobacteria were positive for PKS and NRPS genetic screens [@pone.0035953-Ehrenreich1].\n\nDue to the fact that plants are host to diverse endophyte populations, it was speculated that many of the total DNA extracts would contain PKS and NRPS genes. However, genetic screening for endophyte biosynthesis genes has not previously been attempted using total plant DNA extracts. In this study, DNA samples often contained only one PKS/NRPS sequence, inferring that there may be few biosynthesis genes present in the endophytes of the sampled plants. It is probable that only a few endophytes biosynthesize compounds that increase plant viability, and these endophytes may dominate the population, and any PCR results. Additionally, many samples failed to reveal any endophyte PKS or NRPS genes. This indicates that either endophytes within the samples did not contain PKS or NRPS biosynthesis genes or that the biosynthesis gene sequences were too divergent to be amplified by the reaction described. Alternatively, endophytes involved in the biosynthesis of bioactive compounds in these plants may have been absent from the tissue type used for the DNA extractions [@pone.0035953-Bayman1]. Instead endophytes could inhabit an alternate host tissue and export bioactive compounds for storage in the tissue used medicinally [@pone.0035953-Marler1].\n\nActinobacteria, in particular *Streptomyces* sp., are significant secondary metabolite producers, synthesising a functionally diverse class of bioactive compounds, often through PKS and NRPS pathways [@pone.0035953-Moore1], [@pone.0035953-Weber1]. Although Actinobacteria are only a subset of the total bacterial population, sequences related to non-actinobacteria NRPS genes were also detected using the actinobacterial primers ([Figure 4](#pone-0035953-g004){ref-type=\"fig\"}). This is possible, due to the conserved gene regions within the KS and A domains [@pone.0035953-JenkeKodama1]. However, the type I PKS and NRPS primers used in this study cannot amplify all PKS or NRPS genes, suggesting that these primers underestimate endophytes\\' KS and NRPS gene diversity within the plants. This limitation is in part due to the absence of divergent sequences in the sequence databases and primers which are not designed to detect atypical PKS or NRPS genes. Consequently, some interesting pathways may be overlooked with this approach. Sequence-based approaches will improve from advances in our understanding of the molecular genetics of PKS and NRPS biosynthesis.\n\nPhylogenetic analysis of putative fungal KS domains showed that sequences clustered into two main groups ([Figure 1](#pone-0035953-g001){ref-type=\"fig\"}). PKSs which synthesize partially reduced polyketides, such as lovastatin and T-toxin, were clustered separately from those PKSs that synthesize unreduced polyketides, examples of which include aflatoxin and melanin. This division is supported by previous phylogenetic investigations of fungal PKSs [@pone.0035953-Kroken1], [@pone.0035953-Nicholson1]. All ten sequences from the TCM samples were within the cluster of PKSs that synthesize unreduced polyketides, specifically, pigments. Fungal pigments show a range of bioactivities such as elsinchrome, which possesses phytotoxic activities [@pone.0035953-Liao1], and the napthopyrones, which possess antimicrobial activities [@pone.0035953-Brimble1], [@pone.0035953-Bianchi1]. It can be concluded that the detection of pigment-type biosynthesis genes indicates the biosynthetic potential of TCM endophytes. The putative fungal KS domain fragments identified; BS1, CF1, LH1, LH2, SS1, SS3 and TB1, showed close relationships with KS domains of symbiotic fungi, including endophytes and mycobionts of lichens ([Figure 1](#pone-0035953-g001){ref-type=\"fig\"}). These amplified KS domains may encode biosynthetic pathways unique to symbiotic fungi and thus be involved in the host-symbiont relationship.\n\nPhylogenetic analysis of the fungal A domains demonstrated that the sequences were clustered with NRPSs from bacteria ([Figure 2](#pone-0035953-g002){ref-type=\"fig\"}). Some sequences showed similarity to bacterial PKS sequences involved in the production of cyclic lipopeptides, such as arthrofactin synthetases ([Table 3](#pone-0035953-t003){ref-type=\"table\"}). Cyclic lipopeptide production by *Pseudomonas fluorescens* was shown to play a key role in biofilm formation, motility and the antifungal activity of strain SBW25. Thus the detected fungal A domains could be involved in the biosynthesis of novel fungal peptides. The amplified A domains were related to NRPS pathways involved in the synthesis of compounds with a range of bioactivities; microginin, an angiotensin-converting enzyme inhibitor [@pone.0035953-Okino1], anabaenopeptin, a serine protease inhibitor [@pone.0035953-Murakami1], and aeruginosin, a thrombin inhibitor [@pone.0035953-Kodani1]. Protease inhibitor-type compounds are significant because of the biochemical processes they control. For example serine protease inhibitors possess cell cytotoxic activities and are used for the treatment of a wide variety of human diseases [@pone.0035953-Leung1].\n\nSequences identified from the bacterial NRPS screen were phylogenetically grouped with diverse bacterial NRPS biosynthesis pathways. The sequences LH3 and LH4 were grouped with proteobacterial NRPSs involved in the synthesis of the siderophores pyoverdine and enterobactin [@pone.0035953-Visca1], [@pone.0035953-Liu1]; antibiotics, massetolide A and virginiamycin [@pone.0035953-Namwat1], [@pone.0035953-deBruijn1]; and the biosurfactant, arthrofactin [@pone.0035953-Morikawa1]. It appears that these putative A domain genes could be involved in the synthesis of novel and biologically active peptides, which in turn add to the medicinal activity of the host plants. Gene fragments BC1 and BC2 were related to an A domain of an uncharacterized NRPS pathway of *Rhodococcus erythropolis*. While this specific NRPS gene is uncharacterized, *Rhodococcus* spp. are known to contain a large set of NRPS and PKS genes with their biotechnological significance arising from their existence in harsh environmental niches [@pone.0035953-McLeod1], [@pone.0035953-deCarvalho1]. It is proposed that the products of this NRPS may assist endophytes to survive inside the plant. The putative A domain sequence DL3 failed to cluster with known NRPSs and therefore may represent divergent NRPSs, for the production of unique peptide-derived compounds.\n\nOf interest in the phylogenetic analysis of bacterial PKS fragments, is that fragments PPol1 and PPol2 grouped with various enzymes of *Streptomyces hygroscopicus* involved in geldanamycin biosynthesis. This suggests that the endophyte genes could encode pathways for the production of similar anticancer antibiotics [@pone.0035953-Stebbins1]. The detection of endophyte associated biosynthesis genes in *P. polyphylla* highlights the fact that this plant is a good candidate for further endophyte investigations involved in isolation endophyte and characterisation of their associated bioactive compounds. Endophytes offer sustainable production of plant associated natural products, and have undeniable application in the development of compounds in drug discovery [@pone.0035953-Kusari1].\n\nIn summary, TCM herbs are host to diverse bacteria and fungi with the potential to synthesize secondary metabolites which would contribute to the plant\\'s chemical composition. In this particular study, genetic screening revealed endophyte-derive PKS and NRPS fragments with putative roles in the biosynthesis of secondary metabolites with a wide array of biological activities. This screening allows further investigations to focus on plants that host endophytes with the greatest potential to produce polykletide and nonribosomal peptide-based bioactive substances. To further investigate links between PKS/NRPS genes and bioactivity, elucidation of the full biosynthetic pathways is required from endophytes. Future studies of the plant candidates may lead to harnessing endophytes, their genetic resources, and associated compounds which are valuable leads for drug discovery investigations, as well as understanding how endophytes contribute to the bioactivity of medicinal plants. In addition, it remains to be determined whether the PKS and NRPS genes identified in this study are functionally dedicated to specialized endophyte-based activities.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Multiple sequence alignments.** The multiple sequence alignments of KS domains protein sequences, from PKS pathways, and A domain protein sequences, from NRPS pathways, were generated using ClustalX.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\nThe authors would like to thank Dr. Troco K. Mihali (University of New South Wales), Yanli Zhang and Yajuan Chen (Kunming Medical University) for advice and helpful discussions. We would also like to acknowledge Bin Qiu (Yunnan Institute of Materia Medica) for assistance with collection of some of the plant samples.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**KIM and DMYS were supported by the Australia-China Special Fund for S&T Cooperation (CH070067) <https://grants.innovation.gov.au/ISL/Pages/Doc.aspx?name=ChinaFund.htm>. CQ was financially supported by the National Natural Science Foundation of China (No.30811120009) and the International Cooperation Foundation of Yunnan Province of China (No.30811120009). BAN was supported by a grant and fellowship from the Australian Research Council (FF0883440) <http://www.arc.gov.au/>. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: KIM CQ DMYS BAN. Performed the experiments: KIM CQ. Analyzed the data: KIM BAN. Contributed reagents/materials/analysis tools: CQ BAN. Wrote the paper: KIM BAN. Assisted in organising trip to China: KIM DMYS. Assisted in organising plant collection: CQ.\n"}
+{"text": "**(See the Major article by Schwartz et al on pages 2059--70.)**\n\nThe implementation of rotavirus vaccines in national immunization programs in \\~100 countries (some with phased, subnational introductions) has substantially reduced the disease burden of rotavirus, the leading cause of severe childhood gastroenteritis worldwide \\[[@CIT0001], [@CIT0002]\\]. Early introductions in high- and middle-income countries confirmed the large public health impact anticipated based on the high efficacy (85--98%) of the vaccines against severe rotavirus gastroenteritis observed in clinical trials in these settings. Africa has led the implementation of rotavirus vaccines in low-income settings, with nearly three-quarters of African countries routinely vaccinating against rotavirus. Emerging data from many African countries has shown a substantial impact of vaccination on reducing diarrhea hospitalizations and deaths, which is particularly encouraging given concerns about the somewhat moderate rotavirus vaccine efficacy (50--64%) observed in clinical trials in low-income countries \\[[@CIT0003]\\]. Despite this substantial progress, however, 57% of all children globally still lack access to rotavirus vaccines. In particular, vaccine implementation has lagged in Asia, where less than one-third of countries, including many with large birth cohorts, have implemented national rotavirus vaccination \\[[@CIT0004]\\]. Additional evidence on the health benefits of rotavirus vaccination from Asian countries will encourage the further adoption of vaccines in the region \\[[@CIT0005]\\].\n\nIn this issue of *Clinical Infectious Diseases*, Schwartz and colleagues report an interesting interrupted time-series analysis of data collected over a 15-year period that examines the population-level impact of rotavirus vaccination in Bangladeshi children. Using data for children residing in villages monitored through a health and demographic surveillance system (HDSS), they examined the impact of rotavirus vaccination administered through a rotavirus vaccine donation program in this HDSS population, following the completion of earlier clinical trials. Because data on both diarrhea hospitalizations and the population under surveillance were accurately captured in the HDSS, these analyses avoid the potential biases from changes in referral patterns or catchment populations that could affect the interpretation of data on the vaccine impact if using hospital-based surveillance alone. Appropriately---given differences in the timing of the vaccine introductions, vaccine coverages, and baseline rates of diarrhea hospitalization, which likely reflect differences in healthcare-seeking behavior and access---they conducted separate analyses for children from villages in icddr,b service areas (ISA) versus government service area (GSA).\n\nThere were 2 different time-series models---Model 1, defined *a priori*, and Model 2, defined after an initial examination of the data---used to examine this HDSS data. A comparison of results from the 2 models is complicated by the fact that only a subset of the population used for Model 1 (ie, the population from cluster-randomized villages that did not receive the vaccine in the trial) was used for Model 2. Thus, it is hard to determine to what extent the differences in the results from the 2 models are due to differences in the analytic approaches used, versus differences in the underlying populations. Despite these issues and the post hoc definition of Model 2, the authors present compelling reasons for greater reliance on the results of this model. First, home visits by field staff to encourage treatment for diarrheal episodes during the individually randomized rotavirus vaccine trial conducted during 2007--2009 in the ISA \\[[@CIT0006]\\] were likely responsible for an increase in the overall healthcare-seeking behavior for diarrhea, since no similar change was seen in the GSA over the same period. If this increased healthcare-seeking behavior was sustained during the later period of routine vaccine use, it would tend to artificially increase the postvaccine rotavirus hospitalization rates and, thus, diminish the measured impact of vaccinations. Secondly, the inclusion of data from control villages for the period during the rotavirus vaccine cluster randomized trial \\[[@CIT0007]\\] allowed for the inclusion of more contemporary data in the prevaccine baseline and for improved analytic power, because of the availability of 2 additional years of prevaccine baseline data.\n\nOverall, while there were some differences between the results from the 2 models and only the results from Model 2 reached statistical significance, analyses using both models showed a decreasing trend in rotavirus gastroenteritis hospitalizations during the period of routine rotavirus vaccine use, compared with the prevaccine baseline. Several lines of evidence support that this decline was attributable, at least in part, to the rotavirus vaccinations. First, while time-series analyses are susceptible to confounding by other interventions or factors that might affect the incidences of diarrhea if temporally related to the timing of vaccine implementation, the lack of declines in rotavirus-negative gastroenteritis hospitalizations argues against a nonspecific effect. Secondly, greater declines in rotavirus gastroenteritis hospitalization rates were seen in the ISA compared to the GSA, which correlates with the greater rotavirus vaccine coverage achieved during the routine vaccine use period in the ISA versus the GSA. Finally, the observed 39% overall decline in rotavirus hospitalization rates among children \\<2 years of age in the ISA in the routine vaccine use period is consistent with the decline expected, given the approximately 60--70% vaccination coverage achieved and the 40--60% vaccine efficacy seen in the vaccine trials in Bangladesh and similar settings.\n\nBangladesh has been a global leader in research documenting the health burden of rotavirus and the potential benefits of vaccination. It is the only country in the world with more than 3 decades of continuous and systematic active surveillance data, coupled with laboratory testing; the latest figures show that approximately two-thirds of childhood diarrhea hospitalizations are attributable to rotavirus \\[[@CIT0008]\\]. Trials of both the multinational rotavirus vaccines---RotaTeq (Merck and Co) and Rotarix (GlaxoSmithKline)---have been conducted in Bangladesh and have shown efficacy/effectiveness consistent with that in other developing countries \\[[@CIT0006], [@CIT0007]\\]. A cost-effectiveness analysis showed that rotavirus vaccination would substantially reduce mortality, illness, and the costs associated with rotavirus vaccine in Bangladesh; would be highly cost-effective if supported through a subsidy from Gavi, the Vaccine Alliance; and can be cost-effective without a vaccine subsidy, depending on the vaccine price \\[[@CIT0009]\\]. Furthermore, a recent analysis showed that in Bangladesh, where there is limited hospital bed availability and fierce competition for beds, a reduction in rotavirus gastroenteritis inpatients because of vaccination would make more beds available for other patients with childhood morbidities and, indirectly, improve their treatment and outcomes \\[[@CIT0010]\\]. The analysis by Schwartz and colleagues showing the population-level impacts of rotavirus vaccination further extends and reaffirms the vast evidence of potential benefits from rotavirus vaccination in Bangladesh.\n\nThe experience from Bangladesh will also be valuable for policymakers in other Asian countries that have similar rotavirus epidemiologies and burdens in their deliberations around rotavirus vaccine implementation. It is encouraging that the 2 countries with the largest childhood populations in Southeast Asia---India and Pakistan---have both implemented rotavirus vaccination in a phased manner in their national immunization programs over the past 2--3 years. The governments of Bangladesh and Nepal have also recommended national rotavirus vaccination, and these countries have been approved for funding support from Gavi, the Vaccine Alliance, for vaccine purchases. However, the implementation of rotavirus vaccination has been delayed by global supply shortages for both the rotavirus vaccines from the multinational companies---RotaTeq (Merck, West Point, PA) and Rotarix (GlaxoSmithKline, Rixensart, Belgium). Promisingly, 2 new rotavirus vaccines, manufactured in India---ROTAVAC (Bharat Biotech, Hyderabad, India) and RotaSIIL (Serum Institute of India, Pune, India)---were prequalified by the World Health Organization in 2018 and can now be procured with financial support from Gavi, the Vaccine Alliance.\n\nThe Indian-made rotavirus vaccines have shown efficacy similar to the multinational rotavirus vaccines in developing countries \\[[@CIT0011]\\], and a recent analysis for 3 low-income countries, including Bangladesh, showed that implementation of the Indian-made rotavirus vaccines will have a substantial public health benefit and that they are highly cost-effective \\[[@CIT0014]\\]. The availability of multiple cost-effective rotavirus vaccines and the rapidly growing evidence of their public health impact in routine programmatic use should accelerate the implementation of rotavirus vaccines in Asia and globally, thereby achieving the full public health potential of this life-saving intervention.\n\n***Disclaimer.***The finding and conclusions in this report are those of the authors and do not necessarily represent the official position of the Bill & Melinda Gates Foundation, nor the US Centers for Disease Control and Prevention.\n\n***Potential conflicts of interest.***Both authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.\n"}
+{"text": "\n"}
+{"text": "Introduction {#s1}\n============\n\nX-chromosome inactivation (XCI) has evolved in female mammals to compensate for sex-chromosome dosage differences by suppressing gene expression from one X chromosome, and to render all cells as functionally monosomic for the X chromosome, a process considered to be important for normal embryonic development (Monk and Harper, [@B22]; Penny et al., [@B36]). The current understanding of XCI during early embryogenesis largely originates from studies in mice, partially because XCI occurs during a very early developmental window when the embryo is accessible. In female mice, dosage compensation takes place within a continual cycle of XCI and reversal, also known as X-chromosome reactivation (XCR) (Lee and Bartolomei, [@B16]). Inactivation of the paternal X chromosome occurs progressively during the first days of post-fertilization cleavage until the morula stage (Okamoto and Heard, [@B27]; Kalantry et al., [@B11]; Namekawa et al., [@B26]). At the late blastocyst stage, XCR is observed within cells from the inner cell mass (ICM) that will form the embryo proper (i.e., two active X chromosomes are present in these cells), whereas cells of the trophectoderm (TE), which will form the placenta, maintain imprinted inactivation of the paternal X chromosome (Mak et al., [@B19]; Okamoto et al., [@B28]; Patrat et al., [@B32]). At the onset of gastrulation, embryonic lineage cells randomly undergo XCI again without a parent-of-origin bias, and once random X-inactivation is initiated, all the progeny cells maintain the same X-inactivation status (Lyon, [@B18]). In contrast to embryonic lineage cells, extra-embryonic cells (i.e., TE cells) maintain the silenced paternal X chromosome throughout embryogenesis. Meanwhile, XCR can be induced during reprogramming of differentiated cells toward pluripotency by nuclear transfer, cell fusion, or ectopic expression of reprogramming factors (for a review, see Pasque and Plath, [@B31]; Payer, [@B33]; Vallot et al., [@B43]). Recently, it was shown that during the early period of reprogramming in hybrid cells between human fibroblasts and mouse embryonic stem cells, human nuclei undergo a loss of *XIST* and XCI-associated histone marks from the inactive X chromosome to accomplish XCR, although some regions on the X chromosome are refractory to reprogramming (Cantone et al., [@B3]).\n\nHowever, species-specific differences are found in the patterns of X-inactivation and reactivation in mammals. Early human and rabbit embryos have different XCI initiation strategies compared with mice (Okamoto et al., [@B29]; Deng et al., [@B4]). Furthermore, in these species, *XIST* was not imprinted, and both X chromosomes remained active in the ICM and TE of blastocysts. In addition, several studies found that the paternal X chromosome does not undergo imprinted X inactivation in human embryos and in extra-embryonic tissues (Skuse et al., [@B40]; Skuse, [@B39]; Moreira de Mello et al., [@B23]; Penaherrera et al., [@B35]; Tachibana et al., [@B41]). Similarly, XCI does not occur in bovine blastocysts. Early studies found upregulation of X-linked genes such as *XIAP, G6PD*, and *HPRT* in female bovine blastocysts, despite strong *XIST* expression (Gutierrez-Adan et al., [@B8]; Peippo et al., [@B34]; Wrenzycki et al., [@B46]; Morton et al., [@B24]). Allelic expression analysis of the X-linked polymorphic *MAOA* gene showed preferential inactivation of the paternal X chromosome in bovine fetal placentae at approximately 100 days of gestation (Xue et al., [@B47]); however, more definitive studies with earlier stage embryos have not been reported. A microarray study using pooled *in vitro*-produced bovine blastocysts of known sexes showed that X-linked transcripts were mostly upregulated (Bermejo-Alvarez et al., [@B2]), which indicated that XCI does not occur at the blastocyst stage. Studies on XCI in non-rodent mammalian species demonstrated that X inactivation in mice might not apply to other species; therefore, determining the similarities and differences in XCI among mammalian species is very important.\n\nGlobal gene expression studies using bovine pre-implantation embryos are scarce, and investigations into X-inactivation/reactivation using transcriptomic data from early embryos are scarcer still. This is mostly because of the major technical barrier in dealing with a limited number of embryos in a minuscule volume. In this study, we generated RNA-seq data derived from *in vitro* bovine male and female blastocysts to identify and characterize gender-specific expression patterns of genes from the X chromosome and autosomes. To investigate differences between embryos of different origins, we included male and female somatic cell nuclear transfer (SCNT) embryos in the RNA-seq analysis. Furthermore, our RNA-seq data were obtained from single male and female blastocysts, which enabled analysis of individual blastocysts to investigate the variability in X chromosome gene expression profiles. Knowledge regarding XIC in non-rodent pre-implantation embryos is extremely limited; therefore, our study using bovine embryos will provide insights into the evolutionary and molecular aspects of X inactivation and reactivation that occurs in early mammalian embryos, including humans.\n\nMaterials and methods {#s2}\n=====================\n\nGeneration of bovine blastocyst samples and sexing\n--------------------------------------------------\n\nThis study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Livestock Research Institute of Korea. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Korea Research Institute of Bioscience and Biotechnology.\n\nThe procedures for generation of bovine IVF and SCNT blastocysts were described in detail elsewhere (Kwon et al., [@B15]). Briefly, single donor cell was injected into enucleated oocyte by a micromanipulator with an inverted microscope (Leitz), and each donor-oocyte complex was fused by an Electro Cell Manipulator 2001 (BTX). The fused eggs were activated 4 h after. Blastocysts were generated 6--8 days post-NT. The quality of each blastocyst was assessed by Hoechst staining, and only high quality embryos with 60--80 blastomeres were chosen for transcriptomic analysis. We used male and female bovine ear skin fibroblasts as donor cells which were passaged three times before SCNT.\n\nFor generation of sham SCNT blastocysts, 18--22 h post-IVF, the zygote with two parental pronuclei was chosen for manipulation (Park et al., [@B30]). To exactly mimic the physical damage of enucleation, zona pellucida was partially ripped and the polar body and a part of the underlying ooplasm were removed using a micropipette without touching either male or female pronucleus. After 2 h of incubation, the reconstructed oocytes were activated using 5 \u03bcM ionomycin (Sigma) for 5 min, followed by treatment with 2.5 mM 6-dimethyl-aminopurine (DMAP, Sigma) in CR1aa culture media supplemented with 0.3% BSA for 3.5 h. The oocytes were then *in vitro* cultured to the blastocyst stage.\n\nSexes of IVF or sham-SCNT blastocysts were determined by PCR with Y-specific primers (BY; 5\u2032-CTCAGCAAAGCACACCAGAC-3\u2032 and 5\u2032-GAACTTTCAAGCAGCTGAGGC-3\u2032) and bovine-specific primers (BSP; 5\u2032-TTTACCTTAGAACAAACCGAGGCAC-3\u2032 and 5\u2032-TACGGAAAGGAAAGATGACCTGACC-3\u2032) as previously reported (Rattanasuk et al., [@B37]). One-tenth volume of genomic DNA extracted from single blastocysts was amplified by PCR using AccuPower PCR PreMix (Bioneer) and PCR product was resolved on 2% agarose gel.\n\nTranscriptome amplification of single blastocysts by pico-profiling\n-------------------------------------------------------------------\n\nTranscriptomic materials were extracted from total 35 blastocysts (6 male IVF, 6 male SCNT, 6 male sham, 6 female IVF, 6 female SCNT, 5 sham) and male/female donor cells and amplified by the pico-profiling method. The pico-profiling procedure was described in detail elsewhere (Min et al., [@B21], [@B20]). Briefly, from each bovine blastocyst, poly-A tailed RNAs were extracted using Dynabeads mRNA DIRECT kit (Invitrogen) and reverse transcribed using 200 units of SuperScript III (Invitrogen). Pico-profiling was done using random primer harboring MlyI restriction enzyme site. The pico-profiled cDNA fragments were amplified using 5\u2032 anchor primer by PCR with 20 cycles of 94\u00b0C for 2 min, 70\u00b0C for 5 min. Adapters were removed from amplicons by overnight digestion with MlyI restriction enzyme to produce double strand cDNA fragments suitable for Illumina NGS library generation.\n\nPreparation of NGS libraries\n----------------------------\n\nIllumina NGS libraries were generated using TruSeq DNA Sample Preparation kit (Illumina) according to the supplier\\'s guide with several minor modifications. End-repair was performed using the whole pico-profiling amplicons (20 \u03bcl) by incubating with 25 \u03bcl End Repair Mix (Illumina) and 5 \u03bcl DW at 30\u00b0C for 2 h. The reactions were purified with AMPure XP beads (Beckman) and DNA was eluted in 12.5 \u03bcl DW. Then, 12.5 \u03bcl A-Tailing Mix (Illumina) and 5 \u03bcl DW were added into each sample, and the mixtures were incubated at 37\u00b0C for 2 h. Next, 1 \u03bcl barcoded adapter was added into each 3\\' end adenylated DNA sample with 2.5 \u03bcl Ligation Mix (Illumina) and 6.5 \u03bcl DW, and the mixtures was overnight incubated at 16\u00b0C. Ligates were purified twice using AMPure XP bead (Bechman) and eluted into 30 \u03bcl DW. For size selection, adapter ligated DNA samples were mixed with 10 \u03bcl loading solution (Sage Science) and loaded onto Pippin Prep (Sage Science). DNA samples were enriched by 18 cycles of PCR reaction comprising 5 \u03bcl size selected DNA fragments, 25 \u03bcl PCR Master Mix(Illumina), 1 \u03bcl PCR Primer Cocktail (Illumina), and 19 ul DW. Finally, enriched DNA fragments were purified and sequenced using HiSeq2500. Male and female samples were separately pooled and sequenced using two flow cells of HiSeq2500 system.\n\nBioinformatic analyses\n----------------------\n\nRaw reads from HiSeq2500 (100 bp, PE) were preprocessed using \"trim_galore\" to remove Illumina adapter sequences and low quality bases, and the trimmed reads were aligned on *Bos taurus* UMD3.1 (NCBI) with TopHat2 (Trapnell et al., [@B42]; Kim et al., [@B12]). We followed the Tuxedo suit pipeline (TopHat, Cufflinks, and CummeRbund) with default parameters for mapping, expression estimation and differential expression (DE) analysis (Trapnell et al., [@B42]). For the estimation of gene expression levels, Cufflinks with -G option was used to calculate the abundance of only known transcripts due to the incomplete genomic annotation of the bovine genome.\n\nFor sliding window analyses, each chromosome was binned into 1 megabase windows. FPKM values of genes in each window were summed up, and relative expression levels against mean expression levels of male IVF blastocysts were calculated. In order to plot the calculated relative expression, FPKM values were smoothened by averaging 30 windows moving along each chromosome. All plots in this study were generated using in-house R scripts, Origin (OriginLab), or Excel (Microsoft).\n\nResults {#s3}\n=======\n\nExperimental scheme and validation of pico-profiling-sequencing method through a pilot experiment\n-------------------------------------------------------------------------------------------------\n\nWe extracted mRNA and genomic DNA from single bovine IVF blastocysts (IVF-BLs) for pico-profiling of transcripts (Min et al., [@B21], [@B20]) and gender identification, as illustrated in Figure [1A](#F1){ref-type=\"fig\"}. Complementary DNA (cDNA) amplicons from six male and six female IVF-BLs were used for RNA-seq library construction. First, we evaluated the pico-profiling-sequencing (Pip-seq) method in a preliminary experiment, using total RNA from a pool of four IVF-BLs that was divided into four parts for Pip-seq library construction (Figure [1B](#F1){ref-type=\"fig\"}). The sequencing results showed that the distributions of normalized counts in the four replicates were comparable (Figure [1C](#F1){ref-type=\"fig\"}), and the correlation was higher among the replicates (*r* = 0.98) compared with those between other IVF-BLs (*r* = 0.86--0.90; Figure [1D](#F1){ref-type=\"fig\"}). In addition, we found that the replicates had significantly less variable FPKM (fragments per kilobase of exon per million mapped fragments) values for housekeeping genes such as *GAPDH* and *ACTB* compared to those found in a separate RNA-seq dataset derived from single blastocysts (Figure [1E](#F1){ref-type=\"fig\"}). These results demonstrated that Pip-seq is a reliable method of cDNA amplification for use with minute samples, such as individual preimplantation embryos. In addition to IVF-BLs, we included SCNT blastocysts (SCNT-BLs) in our analysis, which were generated from adult ear skin fibroblasts under standardized SCNT conditions (Kwon et al., [@B14]). As a reference, we included sham nuclear transfer blastocysts (sham NT-BLs) that mimicked SCNT-BLs because they underwent the same nuclear transfer procedure except that they retained their intact genomic material (Figure [1F](#F1){ref-type=\"fig\"}). From Pip-seq, we obtained 37 transcriptomes from 35 blastocysts and two donor cells (see Supplementary Table [S3](#SM5){ref-type=\"supplementary-material\"} for the details of samples and sequencing results). We found that *XIST* was expressed abundantly in female blastocysts, and male-specific genes such as *UTY, DDX3Y*, and *EIF2S3Y* were expressed exclusively in male blastocysts (Supplementary Figure [S1](#SM1){ref-type=\"supplementary-material\"}).\n\n![**Schematics of the sample preparation and pico-profiling processes for deep sequencing. (A)** Preparation of bovine blastocyst samples for sexing and pico-profiling. Blastocysts were derived by *in vitro*-fertilization (IVF) or somatic cell nuclear transfer (SCNT). Adult ear skin fibroblasts were used as donor cells to obtain SCNT embryos. Genomic DNA and mRNA were extracted from single IVF and sham nuclear transfer (NT) blastocysts for sexing and pico-profiling. **(B--E)** Validation of the pico-profiling method. In **(B)**, there is an overview of the experimental strategy for validation using four identical replicates of pooled blastocyst samples. In **(C)** and **(D)**, sequencing results are presented. Box plot in **(C)** shows the distribution of total gene expression levels \\[log~2~(FPKM)\\] for each replicate. Colored stripes indicate gene density within a specific expression level. Scatter plots in **(D)** show the correlation between replicates. For comparison, other non-technical replicates (i.e., male IVF and female IVF blastocyst samples) are included. Pearson correlation coefficients (r) are denoted on each plot. In E, expression levels of housekeeping genes are compared between replicate groups (red) and other independent sample groups (gray). Note significantly less variability in the expression levels of the replicates compared to that found in other sample groups. **(F)**, Generation of sham NT embryos. Sixteen to 18 h after IVF, a zygote with maternal and paternal pronuclei was manipulated to remove the polar bodies plus a portion of cytoplasm underneath them, followed by a standard nuclear transfer protocol to develop to the blastocyst stage.](fgene-08-00042-g0001){#F1}\n\nX-linked genes were upregulated in bovine female blastocysts\n------------------------------------------------------------\n\nWe calculated the mean expression levels of individual genes in each group of blastocysts. Plotting the relative mean expression levels for each group to the mean expression levels for the male IVF group against chromosome, we found that X-linked genes were upregulated specifically in the female blastocyst groups (Figure [2A](#F2){ref-type=\"fig\"}). In a comparison of the mean expression levels between male and female X-linked genes, we found a clear distinction between the sexes in each blastocyst group, in which female blastocysts had high and variable expression, whereas male blastocysts had relatively low and constant expression (Figure [2B](#F2){ref-type=\"fig\"}). The mean female to male X-linked gene expression ratios were estimated as 1.8, 1.5, and 1.9 in the IVF, SCNT, and sham groups, respectively (Figure [2C](#F2){ref-type=\"fig\"}). The almost two-fold higher abundance of X-linked gene transcripts indicated that both maternal and paternal X chromosomes are active in bovine female blastocysts, despite the presence of *XIST* transcripts, and suggested a late onset of XCI, similar to that observed in humans (Okamoto et al., [@B28]).\n\n![**Upregulation of X-chromosome genes in bovine female blastocysts. (A)** Box plots showing relative gene expression levels in FPKMs (fragments per kilobase of exon per million mapped fragments) of IVF, SCNT, and sham NT blastocysts plotted from chromosome 1 to X (excluding the Y chromosome) against the mean FPKM of male IVF blastocysts. X chromosomes are distinguished by higher mean expression levels (red brackets). Upper panel, females; lower panel, males. **(B)** Mean FPKMs of X-chromosome genes (X-linked genes) in individual embryos (dots) from IVF, SCNT, and sham NT groups. Donor cells (cell) are included. Blue, male; red, female; black line, group mean FPKM. **(C)**, Relative expression of X-linked genes in female blastocysts (black) compared to that in male blastocysts (gray). Error bars, standard deviation. **(D)** Comparison of chromosome-scale X-linked gene expression patterns using sliding window analysis between male and female samples, and between IVF and SCNT groups. Mean expression of X-linked genes in IVF and SCNT groups are plotted in order of genomic location. The location of the *XIST* gene is indicated (arrow).](fgene-08-00042-g0002){#F2}\n\nPlotting X-linked gene expression along the chromosome, we found that the expression patterns were very similar in all blastocyst groups (sham groups omitted), with an overall higher level of expression in the female groups, which indicated that upregulation of X-linked genes in female blastocysts does not occur locally, but is chromosome-wide (Figure [2D](#F2){ref-type=\"fig\"}). Notably, female SCNT-BLs exhibited a lower profile compared to that found in female IVF-BLs, which is consistent with the lower FPKMs of X-linked genes found in female SCNT-BLs compared with that found in female IVF-BLs (Figure [2C](#F2){ref-type=\"fig\"}). In addition, although the expression levels of X-linked genes differed between male and female SCNT-BLs, their expressions were very similar in male and female donor cells (Supplementary Figure [S2](#SM2){ref-type=\"supplementary-material\"}), which suggested sexual differentiation in transcriptomes during SCNT development.\n\nThe distribution of female-to-male ratios of X-linked genes was female-biased in both IVF-BLs and SCNT-BLs\n----------------------------------------------------------------------------------------------------------\n\nNext, we examined the distribution of female-to-male (F:M) ratios of the expression levels of X-linked genes in IVF-BLs and SCNT-BLs (Figure [3A](#F3){ref-type=\"fig\"}). Overall, we found that the distribution was female-biased in both groups. However, in contrast to the SCNT group, donor cell populations had a relatively balanced distribution, which suggested that the symmetric F:M distribution pattern of donor cells drifted to an asymmetric, female-biased pattern in SCNT-BLs, as in IVF-BLs. Non-parametric analysis using the Mann-Whitney (MW) test and the two-sample Kolmogov-Smirnov (KS) test showed that distributions of the F:M ratios of X-linked genes with FPKMs \\> 1 were significantly different in SCNT vs. donor cells, and in IVF vs. donor cells (both *p* \\< 0.05), but not between IVF vs. SCNT groups (MW test: *p* = 0.860).\n\n![**Distribution of female-to-male ratios of X-chromosome gene expression levels in bovine blastocysts. (A)** Distribution of female-to-male (F:M) ratios of expression levels of X-chromosome genes (X-linked genes) in bovine IVF (black) and SCNT (red) blastocysts, and donor cells (blue). The dotted line (cyan) denotes an F:M ratio of 1 \\[log~2~(F:M) ratio = 0\\]. Bin size, 0.2 in log~2~(FPKM). **(B)** Volcano plots showing mean F:M levels of individual X-linked genes against *p*-values and fold changes in each blastocyst group. Genes with different F:M fold changes are denoted in different colors. **(C)** Comparison of expression (mean FPKM \u00b1 STD) in autosomal (A, blue) and X-chromosome genes (X, orange) between males and females in IVF, SCNT, and sham-NT blastocysts. The X:A ratio that denotes the FPKM level in each blastocyst is provided in the lower panel; the mean X:A ratios (\u00b1 STD) of individual groups are denoted below. Donor cells are included for reference. **(D)** Distribution of female-to-male ratios in autosomal (A, black) and X-chromosome genes (X, red) in IVF and SCNT blastocysts. A dotted line (cyan) denotes an F:M ratio of 1 \\[log~2~(F:M) ratio = 0\\]. Bin size, 0.2 in log~2~(FPKM). Donor cell samples are included for reference.](fgene-08-00042-g0003){#F3}\n\nThe numbers of X-linked genes expressed in IVF-BLs, SCNT-BLs, sham NT-BLs, and donor cells were 668, 666, 628, and 479, respectively. We used volcano plots to display the mean F:M levels of individual X-linked genes in each blastocyst group (Figure [3B](#F3){ref-type=\"fig\"}), which showed that the majority of X-linked genes were upregulated in female blastocysts. Notably, however, the SCNT group still had a considerable number of X-linked genes that were expressed at higher levels in males. Proportions of X-linked genes that were highly expressed (F:M \\> 1) in female blastocysts were 78.7, 66.7, 78.1, and 51.2%, and those with an F:M ratio \\> 2 were 46.7, 36.2, 33.1, and 24.2% in IVF-BLs, SCNT-BLs, sham NT-BLs, and donor cells, respectively.\n\nExpressions levels of autosomal genes are balanced between bovine female and male blastocysts\n---------------------------------------------------------------------------------------------\n\nWe compared the mean expression levels of X-linked genes and autosomal genes (A-genes) in individual blastocysts, and found that the expressions of A-genes were relatively constant in all blastocysts, irrespective of the different X-linked gene expressions observed between the sexes (Figure [3C](#F3){ref-type=\"fig\"}). Furthermore, we found that the male IVF group had a mean X:A ratio of 0.49 compared with a mean of 1.01 found in the female IVF group, revealing a significant difference between the sexes (Wilcoxon signed rank test; *p* = 6.0e-09). We found X:A ratios of 0.51 and 0.80 in male and female SCNT-BLs, respectively, which were similar to those found in IVF-BLs, although slightly lower in female SCNT-BLs compared with those in female IVF-BLs.\n\nWe compared the distributions of F:M ratios of A-gene expression levels in IVF and SCNT blastocysts, and donor cell groups (Figure [3D](#F3){ref-type=\"fig\"}). The F:M distribution was not skewed for A-genes in these groups, centering on an F:M ratio of 1 (log~2~ ratio = 0). Using an MW test, we found that the F:M distributions between X- and A-genes were significantly different between the IVF and SCNT groups (*p* \\< 0.002), but not in the donor cells (*p* \\> 0.05). In addition, examining autosomal differentially expressed genes (aDEGs; *p* \\< 0.05) between the sexes, we found that the number of male-high or female-high aDEGs was similar in both IVF (51.0 vs. 49.0%, respectively; *n* = 2,020) and SCNT (48.8 vs. 51.2%; *n* = 3,051) groups. This probably indicated that X chromosomes impose an extensive transcriptional regulation on A-genes in both IVF and SCNT blastocysts, as previously suggested (Bermejo-Alvarez et al., [@B2]).\n\nFeatures of X-linked differentially expressed genes and in the de-repression of X-linked genes\n----------------------------------------------------------------------------------------------\n\nA heatmap of X-linked gene expression revealed a gender-specific pattern in blastocysts, with higher levels found in female blastocysts (Figure [4A](#F4){ref-type=\"fig\"}). The heatmap pattern of female SCNT-BLs differed from that of male SCNT-BLs, and resembled that observed in female IVF-BLs. In addition, we found that male and female donor cells had similar X-linked gene expression patterns, but neither the pattern observed in female SCNT-BLs nor that found in male SCNT-BLs was similar to the pattern found in donor cells. Figure [4B](#F4){ref-type=\"fig\"} shows a cluster analysis of X-linked gene expression, in which each cluster presents the pattern of change in gene expression between donor cells and SCNT-BLs, using the pattern seen in IVF-BLs as a reference. The genes that belong to each category are listed in the Supplementary Table [S1](#SM3){ref-type=\"supplementary-material\"}.\n\n![**Features of X-chromosome gene expression in IVF and SCNT blastocysts. (A)** Heatmap of relative X-chromosome gene (X-linked gene) expression in individual IVF and SCNT blastocysts vs. the mean expression level of the male IVF blastocyst group. Donor cells are included for reference. Genes are hierarchically clustered. **(B)** *k*-mean clustering (*k* = 9) of X-chromosome genes among donor cell, SCNT, and IVF blastocyst groups. The frequency of each cluster is denoted in the panel. Upper panel, males; lower panel, females. **(C)** Venn diagram showing the overlap between 182 IVF X-linked differentially expressed genes (xDEGs, *p* \\< 0.05; orange) and 156 SCNT xDEGs (green). **(D--E)** Scatter plots of xDEGs in IVF **(D)** and SCNT **(E)** blastocyst groups. Most xDEGs are highly expressed in females, and high-expressing xDEGs in males (male-high) are partly denoted with gene symbols. **(F--H)** Respective location **(F)** and expression **(G--H)** of 29 X-linked genes in bovine IVF and SCNT blastocysts specifically expressed in human and mouse testes. In **(F)** the relative positions and distances of genes on the X chromosome are indicated by \"lollipops.\" Genes are categorized by expression level where a FPKM \\< 0.1 indicates a none-to-weak level of expression, a FPKM \\> 0.1 indicates weak-to-moderate expression, and a FPKM \\> 0.5 (red) indicates moderate-to-high expression. Comparison of the expression of 29 testes-specific genes between female IVF and SCNT groups **(G)**, and between males and females **(H)**.](fgene-08-00042-g0004){#F4}\n\nNext, we detected and evaluated differentially expressed X-linked genes (xDEGs; *p* \\< 0.05) between the sexes (Figure [4C](#F4){ref-type=\"fig\"}). In the IVF group, we found 182 xDEGs, of which 94% (171/182) were highly expressed in female blastocysts (Supplementary Table [S2](#SM4){ref-type=\"supplementary-material\"}). We found that only a small proportion (6%) of xDEGs, including *GYG2, CD99, GDPD2, ZBED1, XIAP, TM9SF2*, and LOC100848605 were highly expressed in male IVFs (Figure [4D](#F4){ref-type=\"fig\"}). Of these genes, *ZBED1* and *CD99* are located in the pseudoautosomal region 1 (PAR1) and therefore, are present on both the X and Y chromosomes. We propose that this male-biased expression resulted from partial spreading of XCI in females (Johnston et al., [@B10]). Furthermore, *GYG2* is known to have a short truncated version on the Y chromosome in humans (Zhai et al., [@B48]).\n\nBy contrast, in the SCNT group, we found 156 xDEGs (Supplementary Table [S2](#SM4){ref-type=\"supplementary-material\"}), of which 18% (28/156) were highly expressed in males (Figure [4E](#F4){ref-type=\"fig\"}). This proportion was larger than that observed in the IVF group, and it would interesting to determine whether these genes reside at loci that are refractory to X-chromosome reprogramming. We found that 93 xDEGs were common to IVF-BLs and SCNT-BLs (Figure [4C](#F4){ref-type=\"fig\"}), and among them, *XIAP, TM9SF2*, and LOC100848605 were identified to be male-high. The regulatory mechanism involved in the expression of these male-high X-linked genes is currently unknown.\n\nCertain male-specific X-linked genes are expressed specifically in the testes, the majority of which are expressed predominantly or exclusively at both pre- and post-meiotic stages (Wang et al., [@B45]; Mueller et al., [@B25]). Using previously reported data (Wang et al., [@B44]) as a reference, we selected 68 X-linked genes that have testes-specific expression in humans and mice, and identified annotations for 29 of these genes in the bovine genome browser (Figure [4F](#F4){ref-type=\"fig\"}). We found that these genes are not localized at specific loci or regions, but are distributed over the entire X chromosome. In our blastocyst samples, we found that more than half of these 29 genes were expressed at an FPKM \\> 0.1 (Figure [4G](#F4){ref-type=\"fig\"}), and that their expression levels were 3--5 times higher than that found in female blastocysts (Figure [4H](#F4){ref-type=\"fig\"}). Therefore, our findings indicated that upregulation of the X chromosome or XCR leads to a de-repression of a large number of X-linked genes, including testes-specific genes, which are not necessary at the blastocyst stage. Furthermore, the data suggest a randomness and non-selectiveness in the de-repression of X-linked genes during XCR.\n\nX chromosome genes were similarly expressed in inner cell mass cells and trophectoderm cells\n--------------------------------------------------------------------------------------------\n\nBlastocysts contain cells of two different lineages, the ICM cells and trophectoderm (TE) cells. These two groups of cells might, as in mouse blastocyst (Mak et al., [@B19]; Okamoto et al., [@B28]; Patrat et al., [@B32]), have different strategies to express X-linked genes: ICM cells having both maternal and paternal active X chromosomes, and TE cells having active maternal and inactive paternal X chromosomes. To test whether this was the case for bovine blastocysts, female IVF blastocysts (*n* = 19) were dissected physically into two parts, the 'IT', containing both ICM and TE cells (almost 1:1 ratio in cell number), and the TE-only part, and subjected to RNA-seq separately. We found that the mean expression levels of X-linked genes were similar in the IT and TE cells (1.000 \u00b1 0.098 and 1.074 \u00b1 0.137, respectively; Figure [5A](#F5){ref-type=\"fig\"}). If the paternal X had been imprinted in the TE part, the expression level of X-linked genes would have been higher in the IT part than in the TE part. The similarity in X-linked gene expressions between the IT and TE cells argued against the imprinted inactivation of paternal X. The relative expressions of X-linked genes in the IT to those in the TE in respective blastocysts were 0.945 on average, ranging from 0.80 to 1.18 (Figure [5B](#F5){ref-type=\"fig\"}). There was no difference in the expression level of *XIST* at *p* \\< 0.01 level between the IT and TE parts (Figure [5C](#F5){ref-type=\"fig\"}), which represented further evidence against the imprinted X inactivation in TE cells. We compared the mean expression level of X-linked genes and A-genes in each part, and found that the mean X:A ratios were not different between IT and TE samples (0.967 \u00b1 0.102 and 1.008 \u00b1 0.089, respectively; *p* = 0.227; Figure [5D](#F5){ref-type=\"fig\"}), and similar to the mean X:A ratio of whole female IVF blastocysts (1.01 \u00b1 0.29) (Figure [3C](#F3){ref-type=\"fig\"}). Thus, we failed to find evidence for imprinted inactivation of paternal X chromosome or a preferential inactivation of either of the parental X chromosomes in TE lineage cells; therefore, we interpreted the result as indicating that both maternal and parental X chromosomes are regulated similarly in ICM and TE lineage cells in bovine blastocysts.\n\n![**X-chromosome gene expression in inner cell mass and trophectoderm lineage cells of female IVF blastocysts. (A)** The mean expression levels of X-linked genes in different lineage cells of the blastocyst. Female IVF blastocyst was physically split using a syringe needle into the \"IT\" part containing both ICM and TE cells and the TE-only part and separately subjected to RNA-seq. Right panel, the mean expression levels of X-linked genes in the IT and TE cells. **(B)** The relative expression of X-linked genes in IT to that in TE part in each blastocyst. **(C)** The mean *XIST* expression levels in read counts. Error bars, standard deviations. **(D)** The mean X:A ratios in IT (blue) and TE (red) cells. AVG, average IT/TE ratio (see also dotted red line).](fgene-08-00042-g0005){#F5}\n\nDiscussion {#s4}\n==========\n\nWe found that the expression of X-linked genes in female IVF-BLs were almost twice that found in male IVF-BLs (Figure [2C](#F2){ref-type=\"fig\"}). In addition to differences in the global expression levels of X-linked genes, the respective expression patterns were different in some clusters between female and male IVF-BLs, as evident from the heatmap (Figure [4A](#F4){ref-type=\"fig\"}). We interpreted these findings as being caused by differences in X-chromosome dosage between male and female blastocysts, which implied that there is no mechanism for dosage compensation in bovine blastocysts. Our interpretation is in agreement with the observation that the lack of dosage compensation for the most highly expressed X-linked genes is tolerated not only in early mouse embryos (Namekawa et al., [@B26]), but also in embryonic stem cells (Elling et al., [@B6]; Leeb and Wutz, [@B17]). The resulting chromosome copy-number difference results in proteomic and metabolomic differences between the sexes (Gardner et al., [@B7]), which renders pre-implantation embryos of either sex more sensitive (or conversely, resistant) to essential signals for survival from the reproductive tract and also to numerous environmental factors, such as population density, famine, season, and stress, which are known to influence the sex ratio in mammals (Kruuk et al., [@B13]; Zorn et al., [@B50]; James, [@B9]). As shown in the heatmap of X-linked gene expression, SCNT-BLs displayed a contrasting sexual dimorphism, as did IVF-BLs, which demonstrated that the transcriptomic differences among X-linked genes is significantly more prominent between the sexes than between blastocyst groups of different origins (e.g., IVF vs. SCNT). Therefore, if the bovine blastocyst has an innate gender-specific response or behavior in the reproductive tract during the peri-implantation period, the response of SCNT-BLs may not differ from that found in gender-matched IVF-BLs.\n\nIn addition, we found that the imprinted X chromosome derived from donor cells was reactivated in female SCNT-BLs. An important observation was that, possibly due to insufficient XCR, the measured values of X-linked gene expression in SCNT-BLs consistently sat between the respective measures of donor cells and IVF-BLs: (1) the mean FPKMs of X-linked genes (1.0, 1.5, and 1.8 in female donor cells, female SCNT-BLs, and female IVF-BLs, respectively; Figure [2C](#F2){ref-type=\"fig\"}); (2) the proportion of X-linked genes having an F:M ratio \\> 1 (51.2, 66.7, and 78.7%, respectively; Figure [3B](#F3){ref-type=\"fig\"}); (3) the intermediate profile for X-linked gene expression levels across the chromosome observed using a sliding window analysis (Figure [2D](#F2){ref-type=\"fig\"}); and (4) the mean FPKMs of X-linked genes relative to A-genes (0.64, 0.80, and 1.01, respectively; Figure [3C](#F3){ref-type=\"fig\"}). We propose that this \"in-between\" ranking reflects the existence of chromosomal domains, or certain cell lineages, which are resistant to XCR in female SCNT-BLs. An alternative explanation, which is mutually non-exclusive to the theory of XCR-resistant domains, is that female SCNT-BLs might comprise a mosaic of cells that have different X-chromosome states; for example, one active and one inactive X chromosome (XaXi), such as that found in human pluripotent stem cell populations (Silva et al., [@B38]; Dvash et al., [@B5]; Anguera et al., [@B1]). Whatever the reason for the observed intermediate expression of X-linked genes in female SCNT-BLs, our findings indicated that reprogramming is incomplete at the mid-blastocyst stage, and might occur variably among blastocysts and even among cells in a blastocyst.\n\nOwing to XCR in SCNT embryos, the imprinted X chromosome that is inherited from the female donor cell becomes the active X chromosome. This occurs in female SCNT-BLs, but not in male SCNT-BLs. XCR may be another process that facilitates reprogramming of the X chromosome, which might be used by female SCNT embryos because of the exceptional reconfiguration of the imprinted X chromosome. Conversely, XCR might be a burden to female SCNT embryos, which causes a serious delay in other concurrent and linked events of reprogramming. As shown by the X-linked gene expression profiles (Figure [4A](#F4){ref-type=\"fig\"}), XCR might differentiate female SCNT-BLs from male SCNT-BLs, which would inevitably establish a sexual dimorphism in male and female SCNT-BLs, and may affect their cloning efficiency. However, because sham NT-BLs were comparable to their IVF counterparts, we concluded that manipulation and other nuclear transfer procedures do not interfere with the XCR process in female SCNT-BLs.\n\nWe found that upregulation of X-linked genes does not entail upregulation of A-genes in female blastocysts. If it had been true, the distribution of the F:M ratios of A-genes would have been skewed toward females, similar to that found for X-linked genes. However, we found that all three blastocyst groups, which derived differently, exhibited a well-balanced F:M ratio for A-gene expression (Figure [3D](#F3){ref-type=\"fig\"}). We detected 2,020 aDEGs from the comparison of male and female IVF blastocysts. These aDEGs were well-balanced in number between the sexes, forming 7% (1030 male-high and 990 male-high aDEGs) of \\~14,300 transcripts expressed (FPKM \\> 0.1) in IVF blastocysts. This result differs from a previous result reporting an array-based analysis of global gene expression in bovine IVF blastocysts (Bermejo-Alvarez et al., [@B2]). In that study, out of \\~9300 transcripts expressed in bovine blastocysts, 12% (\\~1100) were detected as female-high aDEGs and 17% (\\~1600) as male-high aDEGs; this difference appeared large enough to shift the A-gene F:M ratio off-center. Moreover, when the lists of DEGs was compared with our DEGs, only 42 out of 99 annotated DEGs (\\~42%) overlapped. The discrepancy could be explained by a report that the RNA-seq and microarray platforms can yield different DEG results (Zhang et al., [@B49]). In addition, there may be other explanations for the discrepancy; for example, different sample sizes, the need for cDNA amplification, sperm-sorting-based embryo sexing, and advanced data analytical tools. The limited availability of genomic materials for deep sequencing means that the study of the transcriptome of mammalian preimplantation-stage embryos remains under-investigated. Continued research effort will provide sufficient embryo transcriptome data for verification. In addition, the early embryo-specific Pip-seq method, by lowering significantly the current technical constraints that hamper transcriptomic analysis of early mammalian embryos, will likely have a role in these future studies.\n\nAuthor contributions {#s5}\n====================\n\nYK led the project and supervised the study. BM and YK designed the experiments and interpreted the results. JP provided IVF and SCNT embryos. KJ pico-profiled blastocyst samples. BM and YK jointly performed bioinformatic analyses. BM and YK wrote the paper. All authors read and approved the final manuscript.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer MC and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.\n\nThis work was supported by grants from the National Research Foundation of Korea (015R1A2A2A01003117), the National Research Foundation-Science Research Center program (2011-0030049), the KIOM program (K16130), and the KRIBB program.\n\nSupplementary material {#s6}\n======================\n\nThe Supplementary Material for this article can be found online at: <http://journal.frontiersin.org/article/10.3389/fgene.2017.00042/full#supplementary-material>\n\n###### \n\n**X- or Y-linked gene expression patterns**.\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**X-chromosome wide gene expression pattern in donor cells**.\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**k-mean clustering of X-genes**.\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**xDEGs between male vs. female blastocysts**.\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Sequencing depth and mapping rates**.\n\n###### \n\nClick here for additional data file.\n\n[^1]: Edited by: Rafael Franco, University of Barcelona, Spain\n\n[^2]: Reviewed by: Montserrat Corominas, University of Barcelona, Spain; Andrea Cerase, European Molecular Biology Laboratory, Italy\n\n[^3]: This article was submitted to Epigenomics and Epigenetics, a section of the journal Frontiers in Genetics\n"}
+{"text": "\n"}
+{"text": "Introduction\n============\n\nThe safety of contrast agent in echocardiography was reported and there is a study showing that these agents have a good safety profile in both cardiac and abdominal ultrasound applications.[@B1] Here we introduced a simple diagnostic approach to coronary artery fistula with contrast agent during echocardiography and it helped us to access the diagnosis because of the typical diastolic flow in pulmonary artery.\n\nCase\n====\n\nA 46-year-old female was admitted to the hospital with progressive dyspnea. She had been suffering from dyspnea on exertion for 5 years but could not be diagnosed. No abnormal findings were observed on an initial chest posterioaneterior view and electrocardiogram. However cardiac echocardiography revealed abnormal small turbulent flow in the main pulmonary artery ([Fig. 1A](#F1){ref-type=\"fig\"}), and Doppler revealed diastolic-dominant flow ([Fig. 1B](#F1){ref-type=\"fig\"}). Because there was no evidence of a right-to-left shunt, pulmonary parenchymal disease, or hypersensitivity to perflutren, which are contrast agent contraindications, we administered the Definity\u00ae contrast agent (Bristol-Myers Squibb Medical Imaging, North Billerica, MA, USA) in real echocardiographic mode (infusion rate of 3.0 mL/min mixed with normal saline, 50 mL) using a Vivid 7 (GE Ultrasound, Horten, Norway) ultrasound system and found unusual whitish flow in the main pulmonary artery during the diastolic phase ([Fig. 1C and D](#F1){ref-type=\"fig\"}). Under the impression that this was a coronary artery fistula, we performed aortic computed tomography (CT) and revealed two huge right coronary artery fistulas in the main pulmonary artery ([Fig. 2](#F2){ref-type=\"fig\"}). Finally she underwent a surgical correction.\n\nDiscussion\n==========\n\nCoronary artery fistulas account for only 0.4% of congenital heart defects[@B2] and approximately 50% of pediatric coronary vasculature anomalies. The incidence of coronary artery fistula in the overall population is estimated to be about 0.002%, and 20% of patients with a congenital coronary artery fistula have other concomitant car-diac anomalies, most frequently aortic and pulmonary atresia and patent ductus arteriosus. Tetralogy of Fallot has also been reported.[@B3]-[@B5] Congenital fistulas often arise from the right coronary artery system and the majority enter the right ventricle, right atrium, superior vena cava, coronary sinus, or pulmonary arteries.[@B6]\n\nMany patients are asymptomatic; however, an awareness of these fistulas is important because they have been associated with various clinical features including chest pain or heart failure in young patients.[@B7] Coronary artery fistulas have been diagnosed with aortography,[@B8] coronary angiography,[@B9] and coronary CT.[@B7] Although there is a case of colour Doppler assessment of a coronary fistula,[@B10] we cannot easily confirm these fistulas with echocardiography. In this study, we used a simple diagnostic approach for a coronary artery fistula using a contrast agent, which aided with the diagnosis, because of the typical diastolic whitish flow in the pulmonary artery. The safety of the contrast agent (including Definity\u00ae) has been reported, and Wei et al.[@B1] showed that these agents have a good safety profile in both cardiac and abdominal ultrasound applications. The incidence of severe adverse reactions to ultrasound contrast agents is no greater and may be lower than that reported for contrast agents commonly used in other cardiac imaging tests.[@B1] The flow of a coronary artery fistula can be noted easily during the diastolic phase by contrast infusion, so the coronary artery fistula can be approached in a more direct way.\n\nThe authors have no financial conflicts of interest.\n\n![Echocardiographic findings of a coronary fistula with Doppler and contrast agent. Parasternal short axis view showing abnormal, small turbulent flow in the main pulmonary artery (A), and Doppler revealed diastolic-dominant flow (B, white arrow heads). Systolic and diastolic flow in the pulmonary artery were shown with contrast infusion (C and D), and an unusual whitish flow was noted in the main pulmonary artery during the diastolic phase (D, white arrow heads).](kcj-42-205-g001){#F1}\n\n![Computed tomographic findings of the coronary fistula. Aortic computed tomography revealed two huge right coronary artery fistulas in the main pulmonary artery (white arrows).](kcj-42-205-g002){#F2}\n"}
+{"text": "\n"}
+{"text": "Introduction {#s1}\n============\n\n*Pseudomonas aeruginosa* is a highly heterogeneous species whose members show various levels of pathogenicity towards plants and animals [@pone.0060575-He1], [@pone.0060575-Rahme1]. It is widely distributed in the environment and is a normal human commensal. *P. aeruginosa* is naturally resistant to many drugs and its capacity to form biofilms makes it very difficult to eradicate, particularly in chronically infected cystic fibrosis (CF) patients [@pone.0060575-Davies1], [@pone.0060575-Lipuma1].\n\nDue to the recent increase in multidrug resistant bacteria, phage therapy is being considered as a therapeutic alternative to antibiotics [@pone.0060575-Kutter1], [@pone.0060575-Abedon1]. It has been used since soon after the 1917 discovery of bacteriophages by Felix d\\'Herelle at the Pasteur Institute in Paris, mostly in countries of Eastern Europe. Although controlled studies were seldom available, its efficacy has been clearly demonstrated in numerous cases [@pone.0060575-Chanishvili1], [@pone.0060575-Miedzybrodzki1]. Phages have the capacity to reach bacteria trapped inside biofilms such as those that form in the lung of cystic fibrosis patients [@pone.0060575-Brussow1], [@pone.0060575-Alemayehu1]. Therapeutic cocktails such as the \"pyophage\" formulation widely used in Georgia to treat purulent skin, wound and lung infections, contain many different phages against each of their target pathogens (which are *staphylococcus*, *streptococcus*, *proteus*, *Escherichia coli* and *P. aeruginosa*, for pyophage). Each set must be modified twice a year to retain its ability to lyse a large proportion of the target species [@pone.0060575-Kutter1], [@pone.0060575-Kutateladze1]. *P. aeruginosa* bacteriophages are numerous, and current knowledge of their diversity shows that they are distributed in at least 7 genera of purely lytic phages (T7-like, \u03a6KMV-like, LUZ24-like, N4-like, PB1-like, \u03a6KZ-like, JG004-like) in addition to a similar number of temperate genera [@pone.0060575-Ceyssens1], [@pone.0060575-Pirnay1]. Within each genus, phages with a variety of different host spectra are observed, in part at least reflecting differences in their tail-associated adhesins [@pone.0060575-Ceyssens1], [@pone.0060575-Krylov1], [@pone.0060575-Kwan1]. Yet the combination of phages used to treat different types of infections remain empirical and usually target not more than 85% of the isolates necessitating the development of a \"Sur-mesure\" strategy [@pone.0060575-Pirnay1]. For this strategy, at least one virulent phage for a patient\\'s given otherwise recalcitrant bacterial strain is identified from a collection, or even *de novo* isolated when none is available, and a therapeutical preparation is freshly made. A major challenge is the regular emergence of bacteria resistant to any given phage, which necessitates the ongoing isolation of new phages or variants of phages targeting different hosts or host receptors.\n\nThe mechanisms of bacterial resistance to phages are diverse, driving co-evolution of bacteria and bacteriophages [@pone.0060575-Pleteneva1], [@pone.0060575-Hyman1], [@pone.0060575-Stern1], [@pone.0060575-Koskella1]. They involve not only the inhibition of the adsorption of phages on the bacteria but also the injection of DNA, the degradation of phage DNA by restriction enzymes and the recently-identified bacterial immunity system called CRISPR-Cas [@pone.0060575-Hyman1], [@pone.0060575-Labrie1]. CRISPRs (which stand for Clustered regularly interspaced short palindromic repeats) store copies of fragments of foreign DNA called spacers which, together with the Cas proteins, are used to target incoming DNA, leading to their inactivation [@pone.0060575-Pourcel1], [@pone.0060575-Barrangou1], [@pone.0060575-Sorek1]. In *P. aeruginosa* recent experiments demonstrated a CRISPR-Cas mediated interference against specific temperate phages [@pone.0060575-Cady1]. The majority of clinical strains of *P. aeruginosa* harbour prophages and it has been suggested that they may play a role in mucoid conversion in CF patients [@pone.0060575-Miller1]. Interestingly it was shown that some phages belonging to the Mu-like genus carry genes that inactivate the CRISPR-Cas system, illustrating the complex arm-race between phages and bacteria [@pone.0060575-BondyDenomy1].\n\nIn past years, we have investigated the diversity of over 600 *P. aeruginosa* isolates using a Multiple Variable Number of Tandem Repeats (VNTR) assay (MLVA) [@pone.0060575-VuThien1], [@pone.0060575-Sobral1]. Many clusters were found, some of them containing more than 10 non-epidemiologically related isolates recovered in multiple countries and representing clonal complexes. Indeed clones are seemingly emerging in relation to antibiotic resistance or in relation to certain diseases such as CF, in a population that is considered as panmictic [@pone.0060575-VuThien1], [@pone.0060575-Romling1], [@pone.0060575-Scott1]. We recently showed that many multidrug resistant strains belonging to international clones could be lysed *in vitro* by bacteriophages [@pone.0060575-Larche1].\n\nHere, we sought to investigate the spectrum of phage-susceptibility in a selection of *P. aeruginosa* strains representative of the species diversity in CF patients, and to characterize the genera and other properties of phages that are particularly successful clinically, as seen in the Republic of Georgia. To look for resistance mechanisms other than lack of relevant surface receptors, we examined the CRISPR-Cas system pattern and showed that it does not play a primary role in resistance to these various lytic phages.\n\nMaterials and Methods {#s2}\n=====================\n\nEthics statement {#s2a}\n----------------\n\nThe present project is in compliance with the Helsinki Declaration (Ethical Principles for Medical Research Involving Human Subjects). Strains were collected from sputum as part of the patients\\' usual care, without any additional sampling. The ethic committee \"Comit\u00e9 Consultatif pour la Protection des Personnes dans la Recherche Biom\u00e9dicale (CCPPRB) Ile-De-France\" was consulted, specifically approved this study and declared that patient informed consent was not needed.\n\nStrains {#s2b}\n-------\n\nClinical *P. aeruginosa* strains were isolated from CF patients in the context of longitudinal surveys [@pone.0060575-VuThien1], [@pone.0060575-Sobral1]. The letter C refers to CF, the first digit corresponds to a center and the second digit to the strain index in this center. They were genotyped by MLVA as described previously [@pone.0060575-VuThien1] and their characteristics will be reported in more details (Llanes et al. JAC 2013 in press). Forty seven strains were selected for phage isolation and 18 additional ones were used for CRISPR-Cas investigation. Strains UCBPP-PA14 and PAO1 were purchased from the \\\"Collection de l\\'Institut Pasteur\\\" (CIP, Paris, France) and C50 was a gift of U. R\u00f6mling (Karolinska Institutet, Sweden).\n\nBacteriophages {#s2c}\n--------------\n\nThe pyophage preparation was obtained from the Eliava Institute (Tbilisi, Georgia). It is a mixture of phages directed against five bacterial species that predominate in purulent infections, including *P. aeruginosa*, and is used for phage therapy [@pone.0060575-Kutter1], [@pone.0060575-Kutateladze1].\n\nCulture of bacteria for phages isolation {#s2d}\n----------------------------------------\n\nOne colony of *P. aeruginosa* was added to 20 mL Luria Broth (LB) and incubated overnight at 37\u00b0C, the culture was centrifuged at 2500 g for 10 min and the bacteria were concentrated 10 times into saline magnesium (SM) buffer (50 mM Tris-HCL pH 7.5, 100 mM NaCl, 8.1 mM MgSO4, 0.01% gelatin) (density of approximately 15 OD~600\\ nm~).\n\nIsolation of phages from pyophage {#s2e}\n---------------------------------\n\nSeveral dilutions of pyophage were prepared in SM buffer and 10 \u00b5L of each dilution were incubated for 15 min at room temperature with 50 \u00b5L bacterial suspension before adding 4 mL soft agar (0.7%) LB medium and pouring immediately onto LB agar plates. After overnight incubation at 37\u00b0C, a single plaque was picked and purified by several successive platings on host bacteria. Purified phages were amplified on agar plates by infecting 10^8^ bacteria with 10^6^ phages (MOI\u200a=\u200a0.01) and incubating at 37\u00b0C for 8 hours. A double agar layer plaque assay was used to titrate phage suspensions with10 \u00b5L each of different phage dilutions.\n\nProphage induction {#s2f}\n------------------\n\nTo induce prophages, 20 mL of an exponential bacterial culture (0.6 OD~600\\ nm~) was centrifuged and resuspended in 5 mL of 100 mM MgSO~4~. Bacterial cells were transferred onto a Petri dish and placed under a germicide UV light for 30 s before adding 5 mL of fresh LB medium. Irradiated bacteria were then cultured until the absorbance at 600 nm dropped and were pelleted by centrifugation at 2500 g for 10 min at 4\u00b0C. The pH of the supernatant was neutralized with 0.1N NaOH and bacteriophage presence was tested by spotting 10 \u00b5L onto at least four bacterial strains from the panel of 47. When no lysis was observed the assay was repeated with two more strains.\n\nIsolation of new phages {#s2g}\n-----------------------\n\nNew bacteriophages were isolated from sewage samples after overnight enrichment on several bacterial strains. For this, the sewage sample was clarified by centrifugation at 4000 rpm for 20 min and filtered through 45 \u00b5m pore size membranes (Sarstedt, Marnay, France). Then, 5 mL of filtrate were incubated overnight with selected bacterial strains (200 \u00b5L of a tenfold concentration of an overnight culture) and 10 mL of LB medium at 37\u00b0C. About 100 \u00b5L of chloroform were added to the culture and bacteria debris were removed by centrifugation at 2500 g for 10 min. Ten microliters of the supernatant were spotted onto different bacterial strains including the strain used for phage enrichment. Phages were recovered from lysis zones and purified by two cycles of plating on the enrichment bacterial strain. Following the proposal by Kropinski et al. [@pone.0060575-Kropinski1], the full name of a bacteriophage will include vB for virus of bacteria, and indication of the host species and the virus family. For example the first phage isolated on *P. aeruginosa* strain C1-14 in Orsay will be called vB_Pae_M-C1-14~Or01~. \\\"M\\\" indicates that the phage is a myovirus.\n\nPhage host range {#s2h}\n----------------\n\nIn order to determine phage host range, 10 \u00b5L of concentrated phages (titer\u226510^9^ \"plaque forming units\" PFU per mL) were spotted on stationary growing bacteria and plaque assays were used to check the sensitivity of bacteria. This assay was repeated three times. For those testing positive, their efficiency of plating was determined by spotting 5 \u00b5L of phages at ten-fold serial dilutions (10^\u22121^--10^\u22128^) onto the bacterial lawn. Similarly, phages cocktails prepared by mixing 10^7^ PFU of each phage, were tested using serial dilutions.\n\nElectron microscopy {#s2i}\n-------------------\n\nPhages were amplified by culturing with selected bacterial strains on fresh LB agar plates for 8 hours at 37\u00b0C, and the soft agar was recovered in 4 mL of SM buffer and a few drops of chloroform. After centrifugation at 2500 g for 10 min, the supernatant was filtrated through 0.45 \u00b5m pore size membranes, and the phages were pelleted by ultracentrifugation (Beckman Coulter SW41 rotor) at 260,000\u00d7g for 2 hours at 4\u00b0C. Phages were purified by two washings with 100 mM ammonium acetate pH 7 and the pellet was finally resuspended in 100 \u00b5L SM buffer. Then 3 \u00b5L of concentrated bacteriophages were spotted on carbon-coated grids and were allowed to adsorb for 5 min. Phages were stained by adding a drop of 2% potassium phosphotungstate (pH 7) for 30sec, and excess sample was removed by carefully touching the side of the grid with filter paper. The grid was then visualized using an EM208S transmission electron microscope (FEI, Eindhoven, The Netherlands) operating at 80 kV.\n\nDNA preparation and restriction enzyme digestion {#s2j}\n------------------------------------------------\n\nFor rapid DNA purification, phages were amplified by culturing on fresh LB agar plates for 8 hours at 37\u00b0C. Then 5 mL of a 10 mM Tris PH 7.5, 10 mM MgSO~4~ solution were added to the plate followed by overnight incubation at 4\u00b0C. The buffer was transferred to a 50 mL tube and bacteria debris were pelleted by centrifugation at 2500 g for 10 min at 4\u00b0C. A mixture of 0.2 mL 2 M Tris pH 7.5, 0.4 mL 0.5M EDTA, 0.2 mL 10% SDS and 10 \u00b5L diethylpyrocarbonate was added to 4 mL of supernatant. Following incubation at 65\u00b0C for 30 min, the tube was cooled on ice and 1 mL 5M KOH was added. After 1 hour incubation in ice, centrifugation was performed at 25000 g for 20 min at 4\u00b0C. The DNA contained in the supernatant was precipitated with 2 volumes of absolute ethanol, pelleted by centrifugation, washed twice with 70% ethanol, dried and dissolved in 0.4 mL 10 mM Tris pH 7.5, 1 mM EDTA. For restriction enzyme analysis, 10 \u00b5L bacteriophage DNA was digested according to the manufacturer\\'s recommendations and analysed on a 0.8% agarose gel in 0.5\u00d7 TB buffer. In some cases, digestion was carried out overnight to ensure total digestion. The restriction enzymes were *Eco*RI, *Hind*III, *Bss*HII, *Bgl*II, *Sau*3A, *Alu*I, *Hae*II.\n\nPhage whole genome sequencing {#s2k}\n-----------------------------\n\nPhage DNA was prepared using the rapid protocol described above and the quality was checked by restriction enzyme digestion, followed by gel analysis. Then 5 \u00b5g were sent to the CNRS sequencing facility in Gif sur Yvette (IMAGIF) for preparation of a library and sequencing on the Illumina platform (Illumina Genome Analyzer IIx). Paired-end 75 bp reads were produced, from mean insert length around 300 bp. The BioNumerics tools (Applied Maths, Sint-Martens-Latem, Belgium) were used for assembly and comparison of genome sequences. For the *de novo* assembly, two algorithms were applied to produce contigs, namely Velvet [@pone.0060575-Zerbino1] and Ray [@pone.0060575-Boisvert1]. Both are implemented in the 6.6 version of the BioNumerics software, in the \\\"Power Assembler\\\" module. A single contig was obtained without any gap and the distribution of reads along this contig was analysed. After alignment with published sequences large regions of deletion were verified by PCR amplification using primers selected in flanking sequences. Annotated sequences were deposited at EMBL-EBI under accession number HE983845 for phage vB_PaeM_C2-10_Ab01, and HE983844 for phage vB_PaeP_C1-14_Or01. Additional information is available at the web site <http://bacteriophages.igmors.u-psud.fr>.\n\nAmplification and sequencing of CRISPR and of *cas* genes {#s2l}\n---------------------------------------------------------\n\nTwo CRISPR-Cas systems have been observed in sequenced genomes of *P. aeruginosa*. The type I--F (*Yersinia*) system or CASS3 subtype [@pone.0060575-Makarova1] found in strain PA14 includes two CRISPRs with 28 bp repeats. They are referred to as NC_6483_2 or CRISPR1 with repeat TTTCTTAGCTGCCTATACGGCAGTGAAC and NC_006483_3 or CRISPR2 with repeat TTTCTTAGCTGCCTACACGGCAGTGAAC [@pone.0060575-Grissa1]. They are associated with a set of *cas* genes, from among which *cas*1 was selected for amplification. The primers used for *cas*1 and CRISPR PCR amplification are shown on [Table 1](#pone-0060575-t001){ref-type=\"table\"}. In the unpublished draft genome sequence of strain 2192 [@pone.0060575-Mathee1] (NZ_AAKW00000000, available from the Broad Institute web site [www.broad.mit.edu/annotation/genome/pseudomonas_group/MultiHome.html](http://www.broad.mit.edu/annotation/genome/pseudomonas_group/MultiHome.html)) the type I--F CRISPR-Cas system is present together with another system composed of three CRISPRs with a 29 bp repeat and another set of *cas* genes. This system is called type I--E (*Escherichia*) or CASS2 [@pone.0060575-Makarova2]. Cse3_2192 was amplified using primers Pseudomonas cse3 For and Pseudomonas cse3 Rev described by Cady et al. [@pone.0060575-Cady2].\n\n10.1371/journal.pone.0060575.t001\n\n###### Primers used for PCR amplification in the present study.\n\n![](pone.0060575.t001){#pone-0060575-t001-1}\n\n  Primers                             Sequence 5\u2032 to 3\u2032                   Reference\n  ------------------------ --------------------------------------- -----------------------\n  CRISPRPaer-F                      CTTGACGACCTATGTGGCAG                 This study\n  CRISPRPaer-R                      GTCGCTGCAAAAAGAGGGTT                 This study\n  CRISPR2Paer-F                     TTTTCGTCTGTGTGAGGAGC                 This study\n  CRISPR2Paer-R                     AGCAAGTTACGAGACCTCGA                 This study\n  CRISPR3Paer-F                     ATTTCCAGGAGCGGCGCGAG                 This study\n  CRISPR3Paer-R                     GATCACGCCACTGGTCGTAG                 This study\n  PaerCas1_F                       GACATTTCTCCCAGCGAACTGA                This study\n  PaerCas1_R                       CTTCTTCGGTCAGTAGATGCTC                This study\n  Pseudomonas *cse3* For    ATGTACCTGACCAGACTGACCCTTGATCCTCGCAGCG   [@pone.0060575-Cady2]\n  Pseudomonas *cse3* Rev        GGCTCAGCAGGCCACAGCCGAAAGCCTTG       [@pone.0060575-Cady2]\n\nPCR products were sequenced by Eurofins MWG biotech after purification using the DNeasy Blood & Tissue kit (QIAGEN). To classify the spacers in CRISPRs, the CRISPRtionary tool at the CRISPRcompar web site was used (<http://crispr.u-psud.fr/CRISPRcompar/>) [@pone.0060575-Grissa2]. A dictionary of spacers was produced for CRISPR1 in the orientation based on repeat sequence TTTCTTAGCTGCCTATACGGCAGTGAAC such as described by Cady et al. [@pone.0060575-Cady2].\n\nResults {#s3}\n=======\n\nPurification of bacteriophages from pyophage on selected *P. aeruginosa* strains {#s3a}\n--------------------------------------------------------------------------------\n\nThe complex commercial Eliava \"pyophage\" formulation has a very broad host range, built up over the years, but the details of its composition are unknown. To isolate bacteriophages from pyophage and to test their host ranges, 47 *P. aeruginosa* strains were selected from among 325 strains from French CF patients, on the basis of their distribution into different MLVA clusters ([Fig. S1](#pone.0060575.s001){ref-type=\"supplementary-material\"}). Two to three members each of 14 MLVA clusters were chosen in order to check whether a similar pattern of phage susceptibility would be observed within each cluster. In addition three reference laboratory strains, PA14, PAO1 and C50, were also used. In about 90% of the strains, the presence of prophages or pyocins could be demonstrated following UV activation and spot testing onto 4 to 6 different strains among the panel of 47 (data not shown). To isolate phages, the pyophage preparation was titered at 10^\u22121^, 10^\u22122^, 10^\u22123^ and 10^\u22124^ dilution on a subgroup of 14 CF strains, distributed into the different MLVA clusters, and on the reference strains PAO1 and C50. On C50, C3-16, C7-12, C7-25, C8-20 and C9-6 no plaques were seen whereas plaques of different sizes and morphologies were observed on the other strains. A total of 10 plaques were recovered from 10 different bacterial strains (C1-14, C1-15, C2-10, C3-20, C5-13, C8-5, C8-13, C8-17, PAO1, Tr60) and purified by three successive platings on the same strain. Different plaque sizes and shapes were observed and were characteristic for a phage on a given strain (three examples are shown on [Fig. 1](#pone-0060575-g001){ref-type=\"fig\"}). All phages produced clear plaques after two rounds of infection on a strain. On PAO1, four phages produced very large plaques surrounded by a halo, which kept enlarging with time, suggesting the presence of a strong depolymerase activity ([Fig.1A](#pone-0060575-g001){ref-type=\"fig\"}).\n\n![Morphology of plaques produced by three different phages of the pyophage A) P1-15~pyo~ on PAO1, B) P1-14~pyo~ on C1-14, C) PTr60~pyo~ on Tr60.](pone.0060575.g001){#pone-0060575-g001}\n\nThe restriction profiles of phage genomic DNA were investigated using several enzymes. *EcoR*I, *Hind*III, *Sau3*A and *Alu*I were able to restrict all of the samples, although full digestion of P1-14~pyo~ was not observed with *EcoR*I and *Hind*III. The four phages producing a large halo had the same profile with different enzymes (data not shown) and only one, P1-15~pyo~, was retained for further analysis. Finally, six phages showing different *Hind* III restriction profiles were selected ([Fig. 2](#pone-0060575-g002){ref-type=\"fig\"}). Sequencing of several genomic fragments and evaluation of the genome size by analysis of restriction enzyme patterns, suggested that P1-15~pyo~ is a \u03a6KMV-like phage, P8-13~pyo~ is an N4-like phage, P2-10~pyo~, P3-20~pyo~ and PTr60~pyo~ are LUZ24-like phages, and P1-14~pyo~ is a PB1-like phage ([Fig. S2](#pone.0060575.s002){ref-type=\"supplementary-material\"} and data not shown) [@pone.0060575-Ceyssens1]. In agreement with these results, electron microscopy analysis showed that phage P1-14~pyo~ with a 72 nm diameter head and a 140--142 nm long tail belonged to the *Myoviridae* family, whereas P1-15~pyo~, P8-13~pyo~ and P2-10~pyo~ with isometric heads of respectively 63--65 nm, 72 nm and 58--60 nm diameter belonged to the *Podoviridae* ([Fig. 3](#pone-0060575-g003){ref-type=\"fig\"}). The morphology of P3-20~pyo~ and PTr60~pyo~ was similar to that of podovirus P2-10~pyo~ (data not shown).\n\n![Restriction analysis of phage DNA.\\\nThe DNA of each of six phages from pyophage was digested by *Hind*III and electrophorezed on a 0.8% agarose gel. From 1 to 6: P1-15~pyo~, P8-13~pyo~, P2-10~pyo~, P3-20~pyo~, PTr60~pyo~, P1-14~pyo~. MW is a size marker. On the side are indicated the sizes of three DNA fragments. The DNA of P1-14~pyo~ is not totally digested.](pone.0060575.g002){#pone-0060575-g002}\n\n![Electron microscopy observation of four phages from pyophage: P8-13~pyo~, P1-14~pyo~, P1-15~pyo~, P2-10~pyo~.](pone.0060575.g003){#pone-0060575-g003}\n\nVirulence of the pyophage-isolated phages {#s3b}\n-----------------------------------------\n\nWe evaluated the host range and virulence of the six pyophage-derived phages toward the full collection of 47 *P. aeruginosa* strains representative of the diversity in CF patients and towards the three reference strains, by spotting a concentrated suspension onto a bacterial lawn. [Table 2](#pone-0060575-t002){ref-type=\"table\"} shows the bacteriophage virulence profile. No clear lysis zone could be seen even with high phage concentrations on eleven strains (C50, C5-2, C7-6, C7-12, C7-25, C7-26, C8-7, C8-14, C8-15, C8-20, C9-6) and only a turbid zone or trace growth was observed on 3 additional strains (C3-16, C5-13, C9-5) in good agreement with the lack of susceptibility to the pyophage cocktail presented in the previous paragraph. The other strains showed different susceptibilities to the six phages, mostly independent of their MLVA genotypes. In four cases, strains with the same MLVA genotype showed the same or highly similar phage susceptibility (C8-14 and C8-15, C10-3 and C10-4, C7-25 and C7-26, C8-7 and C8-20) but they corresponded to strains from epidemiologically related cases. In other instances, strains with the same MLVA genotype originating from patients in different hospitals displayed different phage susceptibility profiles (C50 and C10-5; C3-13, C5-13 and C10-1; C1-7 and C3-16; Tr60, C7-11 and C8-12).\n\n10.1371/journal.pone.0060575.t002\n\n###### Susceptibility of 50 *P. aeruginosa* strains to 6 pyophage-derived phages, and CRISPR content.\n\n![](pone.0060575.t002){#pone-0060575-t002-2}\n\n                                                            N4-like                \u03a6-KMV-like   PB1-like   LUZ24-like   type I--F   type I--E                 \n  ------------------------------------------- ----------------------------------- ------------ ---------- ------------ ----------- ----------- ---- ---- ---- ----\n  **PAO1** [a](#nt101){ref-type=\"table-fn\"}    C[b](#nt102){ref-type=\"table-fn\"}       C           \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  PA14                                                        \\-                       \\-          C           \\-          \\-           T       \\+   \\+   \\+   \\-\n  **C50**                                                      T                       \\-          \\-          \\-          \\-          \\-       \\+   \\+   \\+   \\-\n  **Tr60**                                                    \\-                       C           C           C            C           C       \\+   \\+   \\+   \\-\n  C1-1                                                         C                       C           T           C           \\-          \\-       \\+   \\+   \\+   \\-\n  **C1-2**                                                     T                       T           T           T            T           T       \\-   \\-   \\-   \\-\n  C1-3                                                         T                       \\-          C           C            C           C       \\-   \\-   \\-   \\-\n  C1-4                                                        \\-                       \\-          \\-          C            C           C       \\-   \\-   \\-   \\-\n  C1-7                                                         C                       \\-          C           T           \\-           T       \\+   \\+   \\+   \\-\n  C1-11                                                       \\-                       \\-          \\-          C            C           T       \\+   \\+   \\+   \\-\n  **C1-14**                                                    C                       C           C           C            T           T                     \n  **C1-15**                                                   \\-                       C           \\-          C           \\-          \\-       \\+   \\+   \\+   \\-\n  C2-3                                                         T                       \\-          \\-          C            C           T       \\+   \\+   \\+   \\-\n  **C2-10**                                                   \\-                       \\-          \\-          C            C           C       \\-   \\-   \\-   \\-\n  C3-13                                                        C                       T           C           T           \\-          \\-       \\+   \\+   \\+   \\+\n  C3-14                                                        C                       \\-          \\-          C            C           C       \\-   \\-   \\-   \\-\n  **C3-16**                                                    C                       \\-          \\-          \\-          \\-          \\-       \\+   \\+   \\+   \\+\n  C3-17                                                       \\-                       \\-          C           C            C           C       \\-   \\-   \\-   \\+\n  **C3-20**                                                    T                       \\-          \\-          C            C           C       \\-   \\-   \\-   \\-\n  C4-12                                                       \\-                       T           \\-          C            C           T       \\-   \\-   \\-   \\-\n  C4-17                                                        T                       T           \\-          C            C           C       \\+   \\+   \\+   \\-\n  C5-2                                                        \\-                       \\-          \\-          \\-          \\-          \\-       \\+   \\+   \\+   \\-\n  C5-12                                                       \\-                       \\-          T           T            T           T       \\+   \\+   \\+   \\-\n  **C5-13**                                                   \\-                       T           \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\+\n  C7-6                                                        \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  C7-11                                                       \\-                       \\-          C           C            C           C       \\+   \\+   \\+   \\-\n  **C7-12**                                                   \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  C7-22                                                       \\-                       \\-          \\-          C            C           C       \\-   \\-   \\-   \\-\n  **C7-25**                                                   \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  C7-26                                                       \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  **C8-5**                                                     T                       T           \\-          T            T          \\-       \\+   \\+   \\+   \\-\n  C8-7                                                        \\-                       \\-          \\-          T           \\-          \\-       \\-   \\-   \\-   \\-\n  C8-12                                                       \\-                       \\-          C           T            T           T       \\+   \\+   \\+   \\-\n  **C8-13**                                                    T                       \\-          C           C            C           C       \\+   \\+   \\+   \\-\n  C8-14                                                       \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  C8-15                                                       \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  **C8-17**                                                   \\-                       C           \\-          C           \\-          \\-       \\+   \\+   \\+   \\-\n  C8-19                                                       \\-                       \\-          \\-          C            T           T       \\+   \\+   \\+   \\-\n  **C8-20**                                                   \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  C9-2                                                        \\-                       \\-          \\-          C           \\-          \\-       \\-   \\-   \\-   \\-\n  C9-5                                                        \\-                       T           \\-          T           \\-          \\-       \\-   \\-   \\-   \\-\n  **C9-6**                                                    \\-                       \\-          \\-          \\-          \\-          \\-       \\-   \\-   \\-   \\-\n  **C9-11**                                                    C                       C           C           C            C          \\-       \\+   \\+   \\+   \\-\n  C9-17                                                       \\-                       \\-          T           \\-           T          \\-       \\+   \\+   \\+   \\-\n  C10-1                                                        C                       \\-          \\-          \\-          \\-          \\-       \\+   \\+   \\+   \\+\n  C10-2                                                        C                       \\-          C           C            C           C       \\+   \\+   \\+   \\-\n  C10-3                                                        C                       \\-          \\-          \\-          \\-           T       \\-   \\-   \\-   \\-\n  C10-4                                                        C                       \\-          \\-          \\-          \\-           T       \\-   \\-   \\-   \\-\n  C10-5                                                        C                       \\-          T           T            C           T       \\+   \\+   \\+   \\-\n  C10-6                                                        T                       \\-          \\-          T            C           C       \\-   \\-   \\-   \\+\n\nin bold are shown strains used to purify phages from pyophage.\n\nC means clear lysis zone and T means turbid lysis zone.\n\nIsolation and whole genome sequencing of two new phages {#s3c}\n-------------------------------------------------------\n\nDifferent strains that were not susceptible to pyophage were used to enrich for phages in sewage water from Orsay (France) and Abidjan (Ivory Coast) and one from each place was selected for further characterization. One phage isolated in Orsay using C8-15, and thereafter amplified on C1-14, was called \"vB_PaeP_C1-14_Or01\" abbreviated as P1-14~Or01~. It is a podovirus as shown by EM analysis, with a 60 nm head and a short tail ([Fig. 4](#pone-0060575-g004){ref-type=\"fig\"}). Whole-genome sequencing was performed using the Illumina technology and a single contig of 45,469 nucleotides was assembled. Interestingly, when we analysed the reads coverage along this contig we noticed the presence of high peaks with one defined end on one strand ([Fig. S3](#pone.0060575.s003){ref-type=\"supplementary-material\"}). Two peaks could be assigned to the genome ends (labelled RE and LE Fig S3A) and 8 peaks were distributed along the genome. At the 8 sites we found a short conserved sequence which consensus is GTCATAGTAC ([Fig. S3B](#pone.0060575.s003){ref-type=\"supplementary-material\"}).\n\n![Electron microscopy observation of the new phages P1-14~Or01~ and P2-10~Ab01~.](pone.0060575.g004){#pone-0060575-g004}\n\nUpon alignment of the P1-14~Or01~ genome with sequenced genomes present in Genbank, the closest phages were \u03a6MR299-2 with 97% average nucleotide similarity [@pone.0060575-Alemayehu1], PaP3 with 95% average similarity [@pone.0060575-Tan1], and LUZ24 with 90% average similarity [@pone.0060575-Ceyssens2]. The differences were not evenly distributed but were concentrated into several regions with 84 to 94% identity with the closest phage genome \u03a6MR299-2. An approximately 3000 bp fragment showed 97% similarity with phage PaP3 and 86% similarity with phage \u03a6MR299-2. It encompassed 3 hypothetical open reading frames (orf10 to orf13), two of which (orf12 and orf13) encode constituents of the phage particle. Another region of about 660 bp showed 91% identity with \u03a6MR299-2 and 92% identity with LUZ24 and encompassed orf17 encoding a putative phage particle protein. Finally a 600 bp region with 84% identity with \u03a6MR299-2 and 96% identity with PaP3 encoded orf25, a putative phage constituent protein. [Figure 5](#pone-0060575-g005){ref-type=\"fig\"} shows the gene organisation of P1-14~Or01~ and a comparison with that of phages \u03a6MR299-2 and LUZ24. A 182 bp sequence similar to the LUZ24 long terminal repeat sequence could be assembled on one end. Three tRNA genes (tRNA^Asn^, tRNA^Asp^ and tRNA^Pro^) were present in P1-14~Or01~ and phiMR199-2 compared to four in PaP3 and two in LUZ24. Phage \u03a6MT299-2 is lacking a 734 bp region, as compared to P1-14~Or01~, encompassing three putative genes (orf58 to orf60 in P1-14~Or01~). No genes potentially involved in lysogeny (integrase, recombinase, repressor, or excisionase) were detected in this phage.\n\n![Alignment of P1-14~Or01~, MR199-2 and LUZ24 genomes.\\\nPutative open reading frames (ORF) are shown with arrows. In red are shown the tentative cluster of genes encoding structural proteins and in yellow the genes involved in DNA replication. tRNA genes are shown with triangles. Boxes with different shades of grey represent degrees of similarities between genomes.](pone.0060575.g005){#pone-0060575-g005}\n\nThe second new phage recovered from a sewage water sample from Abidjan by enrichment with pyophage-resistant strain C7-6 produced large clear plaques on this strain, and on C9-5 but with a low efficiency. The highest degree of growth was seen on C2-10 and C9-11. It is a myovirus as shown by EM examination, with a 85 nm diameter head and a 142 nm long tail ([Fig. 4](#pone-0060575-g004){ref-type=\"fig\"}). This phage, which was amplified on C2-10, was named \"vB_PaeM_C2-10_Ab01\" abbreviated as P2-10~Ab01~. Genome sequencing produced a single contig of 92,777 bp showing 93% average sequence similarity with JG004 [@pone.0060575-Garbe1], and 91% average similarity with PaKP1 [@pone.0060575-Debarbieux1] and PaP1 (HQ832595) which are JG004-like phages. Eleven tRNA genes were found (tRNA^Gln^, tRNA^Arg^, tRNA^Lys^, tRNA^Leu^, tRNA^Ile^, tRNA^Asp^, tRNA^Cys^, tRNA^Pro^, tRNA^Gly^, tRNA^Phe^, tRNA^Glu^) compared to twelve in JG004 and PaKP1 (the same eleven plus tRNA^Asn^). Alignment of P2-10~Ab01~, PaKP1 and JG004 genomes ([Fig. 6](#pone-0060575-g006){ref-type=\"fig\"} A and B) emphasized the mosaic composition of the genomes and showed that the structural genes have the highest degree of conservation. The gene organization was similar, apart from short orfs which were differentially annotated in the three genomes. PJG4_036 and PJG4_070 encode two putative homing endonucleases which were absent from PaKP1 and from P2-10~Ab01~, as confirmed by PCR using flanking primers (data not shown).\n\n![Alignment of P2-10~Ab01~, JG004 and PAKP1 genomes.\\\nThe genomes are separated into two sections (A and B) of approximately 46 kb. The legend is that of [Figure 4](#pone-0060575-g004){ref-type=\"fig\"}.](pone.0060575.g006){#pone-0060575-g006}\n\nThe susceptibility of 20 *P. aeruginosa* strains to the two newly isolated phages was investigated by spotting serial dilutions of phages ([Table 3](#pone-0060575-t003){ref-type=\"table\"}). P1-14~Or01~ differs from the three LUZ24-like phages isolated from the pyophage preparation by it virulence for C5-12 and by a low level of growth on C8-15 ([Table 2](#pone-0060575-t002){ref-type=\"table\"} and [Table 3](#pone-0060575-t003){ref-type=\"table\"}). Interestingly P1-14~Or01~ systematically produced both clear and turbid plaques on C2-10, even after purification of a single plaque, whereas only clear plaques were seen on C1-14, C5-12 and C8-15. P2-10~Ab01~ produced plaques on C7-6 and C9-5 although with approximately 10 times lower efficiency as compared to growth on C2-10 and C9-11. Several strains which did not support growth of the pyophage-derived phages were not susceptible to the new phages (C5-2, C7-12, C7-25, C8-20). Partial 16S RNA gene amplification and sequencing was performed on DNA from these strains confirming that they belonged to the *P. aeruginosa* species.\n\n10.1371/journal.pone.0060575.t003\n\n###### Susceptibility of 17 *P. aeruginosa* strains to two newly isolated phages.\n\n![](pone.0060575.t003){#pone-0060575-t003-3}\n\n                       P1-14~Or01~                P2-10~Ab01~\n  ------- -------------------------------------- -------------\n  PAO1     t[a](#nt103){ref-type=\"table-fn\"} ++       \\-\n  PA14                      \\-                        t++\n  C50                       t+                        \\-\n  Tr60                     c+++                       t++\n  C1-14                    c+++                       c+\n  C2-10                    c+++                      c+++\n  C3-16                    t++                        t+\n  C3-20                    t++                        \\-\n  C5-2                      \\-                        \\-\n  C5-12                    c+++                       \\-\n  C7-6                      t+                        c+\n  C7-12                     \\-                        \\-\n  C7-25                     \\-                        \\-\n  C8-7                      \\-                        t++\n  C8-14                     t+                        \\-\n  C8-15                     c+                        \\-\n  C8-20                     \\-                        \\-\n  C9-5                      t+                        c+\n  C9-6                      \\-                        \\-\n  C9-11                     t+                       c+++\n\n\\\"t\\\" - turbid plaques; \\\"c\\\" - clear plaques; \\\"-\\\" - no growth; \\\"+\\\" - degree of growth.\n\nSusceptibility of bacteria to phage cocktails {#s3d}\n---------------------------------------------\n\nThree phages mixtures were prepared, respectively composed of the 6 pyophage-derived phages (6 ph cocktail), the two newly isolated phages P1-14~Or~ and P2-10~Ab01~ (2 ph cocktail) and the combination of 8 isolated phages (8 ph cocktail) together representing 5 genera ([Table 4](#pone-0060575-t004){ref-type=\"table\"}). All mixtures contained 10^7^ PFU/mL of each phage. Susceptibility of bacteria toward pyophage and toward 6 ph and 8 ph cocktails were similar except for strains C7-6 and C9-5 which were lysed by P2-10~Ab01~ present in 6 ph and 8 ph. The trace growth seen for C8-15 using P1-14~Or~ alone was not observed in the 2 ph and 8 ph cocktails. The concentration of phages in the pyophage was probably lower than in the other cocktails explaining the lower level of growth on different strains. All the strains that did not support the growth of pyophage were equally not infected by the 6 ph cocktail. Altogether these results demonstrate that the association of several phages does not affect the activity of a given phage.\n\n10.1371/journal.pone.0060575.t004\n\n###### Bacteria susceptibility to four phage-cocktails.\n\n![](pone.0060575.t004){#pone-0060575-t004-4}\n\n           Pyophage   6 ph[a](#nt104){ref-type=\"table-fn\"}   2 ph[b](#nt105){ref-type=\"table-fn\"}   8 ph[c](#nt106){ref-type=\"table-fn\"}\n  ------- ---------- -------------------------------------- -------------------------------------- --------------------------------------\n  PAO1       c+++                     c+++                                   t++                                    c+++\n  PA14       c+++                     c+++                                    \\-                                    c+++\n  Tr60        c+                      c+++                                   c+++                                   c+++\n  C50         \\-                       \\-                                     \\-                                     \\-\n  C1-2       c+++                     c+++                                   c+++                                   c+++\n  C1-14      c+++                     c+++                                    t+                                    c+++\n  C2-10       c+                      c+++                                   c+++                                   c+++\n  C3-16       \\-                       \\-                                     \\-                                     \\-\n  C3-20       c+                      c+++                                   t++                                    c+++\n  C5-2        \\-                       \\-                                     \\-                                     \\-\n  C5-12       t+                      c+++                                   c+++                                   c+++\n  C7-6        \\-                       \\-                                     c+                                     c+\n  C7-12       \\-                       \\-                                     \\-                                     \\-\n  C7-25       \\-                       \\-                                     \\-                                     \\-\n  C8-5        t+                       t+                                     t+                                     t+\n  C8-7        \\-                       \\-                                     \\-                                     \\-\n  C8-14       \\-                       \\-                                     \\-                                     \\-\n  C8-15       \\-                       \\-                                     \\-                                     \\-\n  C8-20       \\-                       \\-                                     \\-                                     \\-\n  C9-2       c+++                     c+++                                   c+++                                   c+++\n  C9-5        t+                       t+                                    c+++                                   c+++\n  C9-6        \\-                       \\-                                     \\-                                     \\-\n  C9-17      c+++                     c+++                                   c+++                                   c+++\n\n6 ph is a mixture of the 6 pyophage-isolated phages.\n\n2 ph is a mixture of phages P1-14~Or01~ and P2-10~Ab01~.\n\n8 ph is a combination of 6 ph and 2 ph.\n\nCRISPR and *cas*1 gene analysis {#s3e}\n-------------------------------\n\nIn order to investigate the potential role of the CRISPR-Cas system in resistance to infection by these phages, we first searched for the presence of the type I--F (*Yersinia*) system in the selected *P. aeruginosa* strains. We used primer sequences localised in the *cas*1 gene because this gene has been shown to be essential for the activity of the CRISPR-Cas system. We also used primers localised on both sides of the CRISPR1 and CRISPR2 loci in PA14 ([Fig. S4](#pone.0060575.s004){ref-type=\"supplementary-material\"}). In 24 strains, including PA14 and C50, PCRs were positive for all three loci ([Table 2](#pone-0060575-t002){ref-type=\"table\"}). The CRISPRs size varied from one strain to the other. To analyse the DR and spacer diversity, CRISPR1 was sequenced in all these strains and in 18 additional ones from the same collection, resulting in sequences for a total of 264 spacers ([Table S1](#pone.0060575.s007){ref-type=\"supplementary-material\"}). Common spacers were observed in strains belonging to clusters defined by MLVA ([Fig. S5](#pone.0060575.s005){ref-type=\"supplementary-material\"} and [S6](#pone.0060575.s006){ref-type=\"supplementary-material\"}). A search for sequences homologous to spacers by BLAST in Genbank gave a 100% match for 40% of the spacers, corresponding mostly to temperate phages or prophages but in a few cases to sequences of plasmids, pyocins or pathogenicity islands. A given CRISPR often contained several fragments of a particular phage. In order to compare the CRISPR of our strains with those described in the literature we produced a combined dictionary of spacers found in the study of Cady et al. [@pone.0060575-Cady2] and the spacers described in the present study ([Table S2](#pone.0060575.s008){ref-type=\"supplementary-material\"}). Twenty six of our strains possessed spacers that were previously found in other strains.\n\nWe also investigated the presence of the type I--E (*Escherichia*) CRISPR-Cas by amplification using primers in the *cas*1 gene. Six strains possessed the type I--E (*Escherichia*), three of which had also the type I--F (*Yersinia*) ([Table 2](#pone-0060575-t002){ref-type=\"table\"}). Interestingly, there was an inverse relationship between the presence of a CRISPR-Cas system and absence of growth of our phages.\n\nDiscussion {#s4}\n==========\n\nThe pyophage preparation {#s4a}\n------------------------\n\nThe pyophage formulation is continuously enriched by new phages and variants of these in response to changes in pathogen targets and it is not clear how many species are present [@pone.0060575-Abedon1]. The collection of strains we have used both to purify phages from pyophage and to test their virulence spectrum is representative of the diversity of clinical (CF) *P. aeruginosa* in French hospitals. The most susceptible *P. aeruginosa* strains in this collection were C1-14 and C2-10. A large variety of plaque morphologies could be observed on these strains. The present work allowed the identification of 6 phages showing different restriction enzyme profiles and belonging to four genera [@pone.0060575-Ceyssens1]. P1-15~pyo~, a \u03a6KMV-like phage, apparently produces an important depolymerase activity which may be of interest to lyse bacteria inside biofilms [@pone.0060575-Hughes1]. Three phages belonging to the LUZ24-like genus show the largest virulence spectra, with few differences between them, although their restriction profiles were rather different. Other phage species may be present in the pyophage but may have been missed on our collection of strains, because of their lower concentration, inability to target the particular strains being used or pure chance. For example, a JG004-like phage was isolated from pyophage in studies involving a large set of CF strains from Children\\'s Hospital in Seattle, as well as phages from 3 of the genera observed above but with different host ranges (E. Kutter, unpublished). Further work including bulk sequencing will be needed to fully characterize the pyophage cocktail (Mzia Kutateladze and Marina Goderdzishvili, work in progress).\n\nIsolation and sequencing of new phages {#s4b}\n--------------------------------------\n\nFrom Orsay sewage water a new phage, P1-14~Or01~, was isolated which could lyse to some extent the pyophage-resistant strain C8-15. Its genome is close to that of phage PAP3 isolated in China [@pone.0060575-Tan1] and to that of \u03a6MR299-2 isolated in Ireland [@pone.0060575-Alemayehu1]. They all belong to the LUZ24-like genus but show substantial variability. PAP3 was reported to be capable of lysogenization (by integration into a tRNA gene rather than by the usual integrase), but we did not find any such evidence for P1-14~Or01~, nor did investigators studying LUZ24 and its other relatives. We observed a very important relative amount of sequencing reads at 8 positions along the genome and detected the consensus sequence GTACTATGAC at each of these positions. This sequence is very close to the consensus sequence 5\u2032-TACTRTGMC-3\u2032 corresponding to single-strand DNA interruptions in phage tf of *P. putida* [@pone.0060575-Glukhov1]. Such nicks were never described in *P. aeruginosa* LUZ24-like phages although the site is present in all of the sequenced genomes reported so far. We believe that the DNA preparation method and/or the sequencing technique utilized in the present study is a convenient way to reveal the existence of nicks in phage genomes.\n\nPhage P2-10~Ab01~ isolated from sewage water in Abidjan could lyse two pyophage-resistant strains, C7-6 and C9-5. Its genome is related to those of JG004 and other phages from Poland, France, Japan and Germany, representing a genus we did not detect in pyophage using our strains. Comparison of these phage genomes shows an important degree of mosaicism although the overall arrangement and gene distribution is similar. P2-10~Ab01~ and PAKP1 apparently lack two genes encoding putative homing endonucleases which were observed in phage JG004. JG004 orf36 encodes a 219 amino acids HNH endonuclease which shows similarities with homing endonucleases possessing a C-terminal AP2 DNA binding domain, found in many plants and some bacterial and viral proteins [@pone.0060575-Magnani1]. Genes of this family are present in different bacteriophages including *Salmonella* phage Felix01 [@pone.0060575-Whichard1] and *P. chlororaphis* phiKZ-like phage 201phi2-1 (orf35) [@pone.0060575-Thomas1]. Orf70 encodes a 258 amino acid putative endonuclease with a N-terminal GIY-YIC catalytic domain similar to that of the SegA--E proteins of phage T4 [@pone.0060575-Sharma1] and also present in many bacteriophages. Homing endonucleases can transpose and duplicate themselves and can move via lateral gene transfer [@pone.0060575-Gimble1]. Their abundance in some genomes such as that of *Xanthomonas oryzae* phage Xp10 genome [@pone.0060575-Yuzenkova1] or phage T4 [@pone.0060575-Edgell1] suggests that they play a role in the evolution of phages. They influence the distribution of sequences flanking their insertion site, whether this site is specific or not.\n\nP2-10~Ab01~ and P1-14~Or01~ were able to grow on some strains that were not infected by the pyophage-derived phages. This may be due to the existence of a specific receptor in the case of phage P2-10~Ab01~ which belongs to a new genus. In the case of P1-14~Or01~ belonging to the LUZ24-like genus as three pyophage-derived phages, it might be due to a mutation in the tail binding domain.\n\nNew phages are continuously isolated for *P. aeruginosa,* and their capacity to target various clinical strains is tested *in vitro* and in animal models [@pone.0060575-Alemayehu1], [@pone.0060575-Morello1], [@pone.0060575-Vieira1]. Our first attempt to obtain phages that could lyse strains that do not support pyophage infection led in part to the isolation of phages which belong to different species or genera. Most new phages belong to already-recognized genera even though they may have new ranges of host specificities. From our observations we suggest that an efficient cocktail of lytic phages for treatment of CF patients should contain phages belonging to different genera such as proposed previously [@pone.0060575-Karumidze1]. However it is not clear if the virulence of different phages in a cocktail will add up or if on the contrary different phage types may not be compatible [@pone.0060575-Chan1]. Our own observations suggest that phages behave similarly alone or in a cocktail. Work done with *E. coli* T4-like phages demonstrates a large variety of host ranges inside a single group [@pone.0060575-ChibaniChennoufi1], [@pone.0060575-Kutter2], and a quite broad variation of host specificity have also been observed between members of some of the genera of *Pseudomonas* phages (E. Kutter, unpublished studies of phages targeting a set of 200 CF strains and phages targeting a set of 100 dog-ear strains). In cocktails containing multiple members of the same genus with different host ranges, there is a substantial possibility for recombination between such phages to generate new host specificities. It is quite probable that this occurs on a regular basis when pyophage is produced and used.\n\nMechanisms of resistance and the role of the CRISPR-Cas system {#s4c}\n--------------------------------------------------------------\n\nThirteen out of 47 *P. aeruginosa* strains, including members of clone C frequently isolated from CF patients, were not lysed by the pyophage preparation or by each phage individually. Some were also not susceptible to a much larger collection of phages from different geographical origins and belonging to other genera such as \u03a6kZ-like phages (V. Krylov personal communication). Loss or significant alteration of receptors or overproduction of exopolysaccharides such as alginates may represent general mechanisms of resistance [@pone.0060575-Koskella1]. The existence of a particular restriction system is also possible. These strains were not genetically related, were not mucoid, and were susceptible to most antibiotics. The lung of CF patients is colonized very early by multiple bacterial species which coexist with bacteriophages inside a biofilm [@pone.0060575-Tejedor1]. This might create a favourable environment for evolution of bacteria towards broad phage resistance although this was not observed experimentally either here or by others [@pone.0060575-Koskella1].\n\nWe showed that, in the tested strains, the presence of a CRISPR-Cas system was not associated with resistance to a particular lytic phage. On the contrary, it appears that a CRISPR-Cas system was absent from the strains that could not support multiplication of any tested phages. However, in the light of the recent discovery in some Mu-like phages of genes that inactivate the CRISPR-Cas systems it is not possible to completely rule out a role of this system in controlling lytic phage infection [@pone.0060575-BondyDenomy1]. None of the phages present in this study belong to the Mu-like genus and we did not find a candidate gene in the configuration described by Bondy-Denomy et al. in the two sequenced genomes. Cady et al. analysed the prevalence of the two CRISPR-Cas systems in *P. aeruginosa* and showed that the 132 viral spacers with matches in sequence databases matched temperate bacteriophage/prophages capable of inserting into the host chromosome; none matched extrachromosomally replicating lytic *P. aeruginosa* bacteriophages [@pone.0060575-Cady2]. They later reported that the type I--F (*Yersinia*) CRISPR region of *P. aeruginosa* strain UCBPP-PA14 plays a role in immunity against specific temperate phages [@pone.0060575-Cady1]. In the strains used here, the majority of spacers with homologues in databases also match temperate phages. We observed some matches with chromosomal sequences such as pyocins and genes of pathogenicity islands, both of which often appear to have their origins in temperate phages. The CRISPR-Cas system might thus be used to inhibit reactivation of prophages or transposable sequences. We confirmed that the majority of clinical *P. aeruginosa* strains are lysogenic (often multilysogenic), and that spontaneous activation of these phages is frequent, as previously reported [@pone.0060575-Miller1], [@pone.0060575-Holloway1]. In *E. coli* and in *Streptococcus pyogenes* the CRISPR-Cas system is also involved in the control of lysogenisation and prophage induction [@pone.0060575-Edgar1]. To understand the role of the CRISPR-Cas system it will be necessary to further explore its function during the establishment and control of lysogeny.\n\nConclusion {#s5}\n==========\n\nOur results emphasize the difficulty in preparing a phage cocktail that could be efficient on the whole species, particularly given the wide diversity of the *P. aeruginosa* species. Several bacterial strains cannot support growth of all lytic phages investigated in this study and phages from other collections, representing a total of 6 different genera, which suggests that they may have developed new resistance mechanisms. These strains will need to be analysed in detail and their genomes fully sequenced to try and identify the involved mechanisms, including possible prophage-based ones, and look for ways of bypassing them.\n\nSupporting Information {#s6}\n======================\n\n###### \n\n**Minimum spanning tree representation of the clustering of 325** ***P. aeruginosa*** **strains from CF patients.** The 50 strains selected for phage isolation are colored in red.\n\n(PPT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Restriction pattern of 5 pyophage-derived phages using** ***Eco*** **RI and** ***Hind*** **III.** The fragment sizes were measured using the BioNumerics software. Lanes 1 to 6, *Hind*III-digested lambda DNA used as size marker, P1-15~pyo~, P8-13~pyo~, P2-10~pyo~, P3-20~pyo~, PTr60~pyo~. Double or triple bands were evaluated by analyzing the band intensity.\n\n(PPT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Distribution of nicks along the genome of phage P1-14~Or01~.** A) Sequence read coverage as observed in BioNumerics. B) Alignment of sequences at the nick site. The consensus motif is shown in red. The \\<\\> symbol in the center of the motif indicates the end of all the sequencing reads.\n\n(PPT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Analysis of the** ***Yersinia*** **CRISPR-Cas system.** PCR amplification of A) CRISPR1, B) CRISPR2 and C) *cas*1 in 19 selected strains and reference strain PA14.\n\n(PPT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Schematic representation of CRISPR1 spacer organization in clusters of strains.** The CRISPR is oriented with the leader on the left, corresponding to the growing end where the more recently added spacers are found. The alignment is the output of a CRISPRtionary analysis after the Re-annote Spacers function has been activated.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Comparison of MLVA clustering and CRISPR 1 content.** On the left is shown a dendrogram produced from MLVA data. Colors indicate strains whose CRISPR1 possesses common spacers. On the right are shown the spacer organization in the four larger clusters.\n\n(PPT)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Sequence of 264 CRISPR1 spacers and distribution into the 42 tested strains.**\n\n(XLS)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Combined list of spacers from Cady et al.** [@pone.0060575-Cady2] **and from the present study.**\n\n(XLS)\n\n###### \n\nClick here for additional data file.\n\nWe are grateful to Mzia Kutateladze and Marina Goderdzishvili (Tbilisi, Georgia) for the pyophage and to Annick Jacq (Orsay) for her expertise and support. We thank Julien Navarro and Christophe Tourterel for their help with genome sequence analysis, and Danielle Jaillard and Luvidine Houel at the Electron Microscopy platform of University Paris Sud for performing EM examination. We thank Victor Krylov and Elena Pleteneva for testing their phage collection on some strains. This work has benefited from the facilities and expertise of the high throughput sequencing platform of IMAGIF (Centre de Recherche de Gif - [www.imagif.cnrs.fr](http://www.imagif.cnrs.fr)).\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: GL GV CP. Performed the experiments: CE YB HVT AC SL. Analyzed the data: CE EK GV CP. Contributed reagents/materials/analysis tools: GL HVT. Wrote the paper: CE EK GV CP.\n"}
+{"text": "![](indmedgaz72907-0011){#sp1 .77}\n\n![](indmedgaz72907-0012){#sp2 .78}\n"}
+{"text": "Introduction {#s1}\n============\n\nMolecular hydrogen (H~2~) is an important feedstock for the synthesis of chemicals and fertilizers, and it also has great potential as a clean carrier of energy for renewable fuel technologies. However, conventional means for industrial-scale H~2~ production such as steam reformation of natural gas fall short of the environmental criteria now needed for sustainable fuels and chemicals [@pone.0015491-Turner1]. The use of hydrogenase enzymes offers a promising alternative to traditional H~2~ production technologies.\n\nHydrogenases catalyze the redox interconversion of protons and hydrogen gas (2H^+^ +2e^\u2212^ H~2~) using unique transition metal cofactors by which the enzymes are classified. Since \\[FeFe\\] hydrogenases more rapidly and preferentially evolve H~2~ than \\[NiFe\\] hydrogenases [@pone.0015491-Adams1], [@pone.0015491-Ghirardi1], they are more desirable for H~2~ production technologies. Unfortunately, these prolific H~2~ producing enzymes are also easily inactivated by oxygen.\n\nThe \\[FeFe\\] hydrogenase active site cofactor, termed the H-cluster, is a complex iron-sulfur cluster that is stabilized by carbon monoxide and cyanide ligands as well as a dithiol bridging molecule [@pone.0015491-Nicolet1], [@pone.0015491-Nicolet2], [@pone.0015491-Peters1]. H-cluster assembly and active hydrogenase expression require at least three accessory proteins called the HydE, HydF, and HydG maturases [@pone.0015491-Posewitz1]. Recent investigations have provided valuable insights regarding H-cluster synthesis. For example, tyrosine was first implicated as an essential substrate for hydrogenase activation [@pone.0015491-Kuchenreuther1], and subsequent work revealed that this amino acid is likely the source for the carbon monoxide and cyanide ligands [@pone.0015491-Driesener1], [@pone.0015491-Shepard1]. A cationically charged channel has also been identified through which the H-cluster cofactor is inserted into the hydrogenase apoenzyme [@pone.0015491-Mulder1], [@pone.0015491-Mulder2], possibly from the HydF maturase [@pone.0015491-Czech1], [@pone.0015491-McGlynn1]. Despite these advances, however, the concerted functionality of the maturases and complete H-cluster biosynthetic pathway have yet to be elucidated [@pone.0015491-Nicolet3]. In addition to the challenges imposed by the complexity of these enzymes and the maturation process, most work with hydrogenases and their maturases must be done in strictly anaerobic environments. The reduced nature of the H-cluster and accessory iron-sulfur clusters (ISCs) makes them susceptible to damage by O~2~ oxidation.\n\nResearch groups have overcome these challenges and have used hydrogenases for energy conversion at the laboratory scale in several different applications. Protein complexes attached to solid-state devices have evolved H~2~ using electrons activated by light [@pone.0015491-Krassen1], [@pone.0015491-Reisner1]. Photoelectrochemical fuel cells with hydrogenases adsorbed to cathodic carbon electrodes have preferentially evolved H~2~ and perform similarly to fuel cells that use platinum catalysts [@pone.0015491-Hambourger1]. Also, synthetic metabolic pathways assembled with purified enzymes have converted sugars to H~2~ and CO~2~ at high yields [@pone.0015491-Woodward1], [@pone.0015491-Zhang1]. Despite the successful development of these hydrogenase technologies, their commercial realization and sustainability will require large quantities of active protein. Various microbial systems have been engineered for producing native and heterologous \\[FeFe\\] hydrogenases [@pone.0015491-Adams2], [@pone.0015491-Akhtar1], [@pone.0015491-Kamp1], [@pone.0015491-King1], [@pone.0015491-Laffly1], [@pone.0015491-Sybirna1], [@pone.0015491-vonAbendroth1], but active enzyme yields are generally less than 1 mg\u00b7L^\u22121^ of culture. Also, few recombinant DNA tools exist for effective overexpression of proteins in organisms that naturally harbor \\[FeFe\\] hydrogenases.\n\n*Escherichia coli* has several advantages that make it desirable for hydrogenase production. The bacterium does not contain a native \\[FeFe\\] hydrogenase that needs to be knocked out to simplify analytical measurements, it is capable of anaerobic respiration, and heterologous expression techniques for this microbe are well-established. Active hydrogenase production using *E. coli* systems has been demonstrated, with total yields comparable to the best reported. However, specific activities of purified hydrogenases from these systems are significantly lower than activities of hydrogenases isolated from their native hosts, likely because of incomplete enzyme maturation [@pone.0015491-Akhtar1], [@pone.0015491-King1], [@pone.0015491-Laffly1].\n\nIn this report, we describe the high-yield production of active \\[FeFe\\] hydrogenases using the maturases native to *Shewanella oneidensis* along with *E. coli* BL21(DE3) *\u0394iscR*, a strain previously engineered for improved synthesis of iron-sulfur (Fe--S) proteins [@pone.0015491-Akhtar1]. Following expression with our optimized protocol, both the *C. reinhardtii* HydA1 and *C. pasteurianum* CpI hydrogenases were isolated by *Strep*--Tactin affinity chromatography and characterized using activity assays and Fourier transform infrared (FTIR) spectroscopy.\n\nResults {#s2}\n=======\n\nRecombinant Protein Expression Concurrent with Anaerobic Metabolism and Growth {#s2a}\n------------------------------------------------------------------------------\n\nCo-expression of the \\[FeFe\\] hydrogenases and the maturases was induced under strictly anoxic conditions at an optical density (OD~600~) of \u223c0.4. To facilitate anaerobic metabolism, glucose (0.5% w/v) and the electron acceptor fumarate (25 mM) were added to the complex growth medium. Aerobic growth rates were exponential (0.45 hr^\u22121^), while anaerobic growth rates were linear and eventually ceased after 24 hr at final OD~600~ measurements ranging from 1.5 to 3.0 ([Fig. 1A](#pone-0015491-g001){ref-type=\"fig\"}). Substrate limitations and acetate accumulation may have contributed to the slowed anoxic growth. Without glucose addition, the culture density did not increase during the anaerobic incubation period. This lack of growth resulted in a lower cell concentration, which thus decreased the total amount of hydrogenase produced per culture volume.\n\n![*E. coli* growth and anaerobic expression of heterologous active \\[FeFe\\] hydrogenases.\\\nAll data are for cultures of *E. coli* strain BL21(DE3) *\u0394iscR*, and both iron and cysteine were included in the growth medium. (Fig. 1A) Optical density at 600 nm (shown on a logarithmic scale) of cultures during aerobic (orange circles) and anaerobic (blue circles) growth phases for cells containing the pACYCDuet-1--*hydGX*--*hydEF* and pET-21(b) *shydA1\\*--Strep-tag II* plasmids. The pH of culture media (\u00d7) was also measured. Data for cultures with cells containing the pET-21(b) *shydA--Strep-tag II* plasmid instead of pET-21(b) *shydA1\\*--Strep-tag II* were similar and are not shown. (Fig. 1B) Cell lysate-based hydrogenase activities (\u00b5mol MV reduced\u00b7min^\u22121^\u00b7mg^\u22121^ total protein) for active CpI (red squares) and HydA1 hydrogenase (green triangles) were determined using the methyl viologen reduction assay. Data are the average for n\u200a=\u200a3 cultures examined \u00b1 standard deviations. (Fig. 1C) SDS-PAGE analysis for the soluble fractions of final cell lysates after the anoxic co-expression of HydA1 or CpI and the HydE, HydF, and HydG maturases: (Lane 1) the molecular weight markers are from the Mark12^TM^ protein ladder (Invitrogen); (Lane 2) soluble cell lysate protein content for *E. coli* strain BL21(DE3) *\u0394iscR* following expression of no heterologous proteins from recombinant DNA plasmids; (Lane 3) co-expression of only the HydE, HydF, and HydG maturases; (Lane 4) co-expression of HydE, HydF, HydG, and HydA1--*Strep*-tag II; and (Lane 5) co-expression of HydE, HydF, HydG, and CpI--*Strep*-tag II.](pone.0015491.g001){#pone-0015491-g001}\n\nActive hydrogenase levels were determined in samples taken during the anaerobic growth period by measuring the methyl viologen reduction activities of cell lysates. This assay for hydrogenase-catalyzed H~2~ uptake enabled us to identify the conditions for optimal active enzyme production. For both the HydA1 and CpI hydrogenases, maximal H~2~ uptake activities were observed after 20--24 hr of anaerobic incubation ([Fig. 1B](#pone-0015491-g001){ref-type=\"fig\"}), and the four heterologous proteins had accumulated to become abundant proteins, based on SDS-PAGE analysis ([Fig. 1C](#pone-0015491-g001){ref-type=\"fig\"}). Increased rates of cell death accompanied by protein degradation could explain the modest decrease in activity after 24 hr, as anoxic growth appeared to cease at this time. Only minimal amounts of methyl viologen reduction (less than 1% of the maximal activities) were observed when either the hydrogenase alone or only the three maturases were expressed. Thus, methyl viologen reduction could be specifically attributed to active \\[FeFe\\] hydrogenase in the cell lysates and not the maturases or native *E. coli* \\[NiFe\\] hydrogenases.\n\n\\[FeFe\\] hydrogenases and each of the three maturases require ISCs in order to function [@pone.0015491-Brazzolotto1], [@pone.0015491-Rubach1], and recombinant overexpression of the four Fe--S proteins likely creates increased demand for ISC assembly. We therefore investigated the benefits of supplementing potentially limiting substrates as well as using the mutant *\u0394iscR* strain engineered for improved production of proteins harboring ISCs [@pone.0015491-Akhtar1]. Expression with the *\u0394iscR* strain improved active hydrogenase production 2--10 fold, with a greater benefit for CpI production ([Fig. 2](#pone-0015491-g002){ref-type=\"fig\"}). Addition of both iron (2 mM ferric ammonium citrate) and cysteine (2 mM) to the culture medium resulted in a further 5--10 fold increase in hydrogenase activity. Neither additive individually improved hydrogenase activation to the same level, although cysteine supplementation led to a moderate improvement in active CpI expression when using the *\u0394iscR* strain. The cooperative effect of iron and cysteine suggests that both substrates are limiting when overexpressing proteins harboring ISCs. We also tested the addition of other relevant hydrogenase maturation substrates to the medium, along with the iron and cysteine. These substrates were *S*-adenosyl methionine (SAM; 2 mM), tyrosine (2 mM), and methionine (2 mM). Adding these individually or in combination did not increase hydrogenase activities (data not shown). Overall, expression with the *\u0394iscR* strain receiving iron and cysteine supplementation enhanced HydA1 activities 25-fold and CpI activities 100-fold.\n\n![Effects of iron and cysteine supplementation as well as the *iscR* deletion on active hydrogenase expression.\\\nIron (2 mM ferric ammonium citrate) and cysteine (2 mM) were added to cultures as indicated. Methyl viologen reducing activities (\u00b5mol MV reduced\u00b7min^\u22121^\u00b7mg^\u22121^ total protein) of active \\[FeFe\\] hydrogenase in cell lysates from *E. coli* strains BL21(DE3) (gray bars) and BL21(DE3) *\u0394iscR* (black bars). Hydrogenase activities were measured after 16--18 hrs of anaerobic HydA1 expression (A) and CpI expression (B). Hydrogenase activities for the *\u0394iscR*-derived samples are indicated above the respective columns, and all activities are the average for n\u200a=\u200a2 cultures \u00b1 standard deviations.](pone.0015491.g002){#pone-0015491-g002}\n\nBiophysical Characterization of Purified \\[FeFe\\] Hydrogenases {#s2b}\n--------------------------------------------------------------\n\nBased on the results presented in [Figure 2](#pone-0015491-g002){ref-type=\"fig\"}, HydA1 and CpI were expressed using the conditions identified for maximal hydrogenase activities and were subsequently isolated to high purity with *Strep*-Tactin affinity chromatography. The elution fractions containing active hydrogenase were identified using the methyl viologen reduction assay as well as by SDS-PAGE analysis. Generally, 70--90% of the total activity present in the cell lysates was recovered in the elution fractions. The purification yields for HydA1 and CpI were 30 mg\u00b7L^\u22121^ of culture and 8 mg\u00b7L^\u22121^ of culture, respectively ([Table 1](#pone-0015491-t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0015491.t001\n\n###### Biophysical characterization of purified \\[FeFe\\] hydrogenases produced in *Escherichia coli*.\n\n![](pone.0015491.t001){#pone-0015491-t001-1}\n\n                                                                 HydA1      CpI\n  ------------------------------------------------------------ --------- ----------\n  **H~2~ Uptake (MV)** [\u2020](#nt102){ref-type=\"table-fn\"}         251\u00b149     476\u00b139\n  \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^                                               \n  **H~2~ Evolution (MV)** [\u2020](#nt102){ref-type=\"table-fn\"}      641\u00b188    1087\u00b1146\n  \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^                                               \n  **H~2~ Evolution (PetF)** [\u2020](#nt102){ref-type=\"table-fn\"}     41\u00b14      90\u00b110\n  \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^                                               \n  **Iron Content**                                              4.5\u00b10.2   13.2\u00b11.3\n  Fe atoms\u00b7peptide^\u22121^                                                   \n  **Purification Yield**                                         30\u00b111    7.9\u00b10.8\n  mg\u00b7L^\u22121^ of culture                                                    \n\nData are the average from n\u200a=\u200a3--6 cultures examined \u00b1 standard deviations.\n\nSpecific activities of purified hydrogenases were determined using methyl viologen (MV) or PetF ferredoxin as the electron donating or accepting substrates.\n\nBoth specific activities and protein-bound iron content were determined for the purified HydA1 and CpI enzymes, and the results are summarized in [Table 1](#pone-0015491-t001){ref-type=\"table\"}. H~2~ uptake rates as well as H~2~ evolution rates are higher for CpI in all cases, and both HydA1 and CpI contained \u223c70% of the maximum amount of iron (6 and 20 iron atoms are expected for these hydrogenases, respectively). We also used the purified *Synechocystis* \\[2Fe--2S\\] ferredoxin PetF (50 \u00b5M) instead of methyl viologen as the electron donating substrate. Since ferredoxin proteins are the native electron transport substrates for \\[FeFe\\] hydrogenases, the catalytic rates when using PetF are more relevant for future design efforts to engineer microbial H~2~ production systems. We observed a K~m~ of 20 \u00b5M for PetF with CpI when using sodium dithionite (DTH) as the source of electrons for ferredoxin reduction. Reduced PetF, however, supported significantly lower H~2~ evolution rates compared to methyl viologen. Lower catalytic rates when using a \\[2Fe--2S\\] ferredoxin have been previously observed when compared to rates observed when a \\[4Fe--4S\\] ferredoxin was used [@pone.0015491-Moulis1].\n\nFTIR Spectroscopic Analysis of Isolated \\[FeFe\\] Hydrogenases {#s2c}\n-------------------------------------------------------------\n\nFTIR spectroscopy was used to analyze the purified HydA1 ([Fig. 3A](#pone-0015491-g003){ref-type=\"fig\"}) and CpI hydrogenases ([Fig. 3B](#pone-0015491-g003){ref-type=\"fig\"}) in both the as-isolated state as well as following treatment with exogenous CO. The spectra for both as-isolated hydrogenases clearly show peaks representing CN^\u2212^ and CO vibrational stretches, indicating the presence of fully assembled H-clusters. Based on previous reports for each of these enzymes [@pone.0015491-Chen1], [@pone.0015491-Silakov1] as well as other \\[FeFe\\] hydrogenases [@pone.0015491-Roseboom1], these spectra also indicate that the as-isolated hydrogenases have a mixture of H-clusters in both the oxidized (H~ox~) and reduced (H~red~) states. The presence of DTH in the elution buffer was essential to prevent hydrogenase inactivation during purification, and this additive was likely the cause for the mixture of H-cluster redox states. The CO inhibition studies confirmed the presence of the H-cluster cofactors, as the CO and CN^\u2212^ vibrational modes shifted as expected after exogenous CO binding to the H-cluster.\n\n![Fourier transform infrared spectra of heterologous \\[FeFe\\] hydrogenases produced in *E. coli*.\\\nInfrared spectra are for 100--200 \u00b5M of the (A) HydA1 and (B) CpI hydrogenases. Both enzymes were examined [@pone.0015491-Turner1] in their as-isolated state as well as [@pone.0015491-Adams1] following treatment with exogenous CO. Vibrational energies (in cm^\u22121^) for the H-cluster CO and CN^\u2212^ ligands are indicated in each spectrum. The wavenumber ranges for terminal CN^\u2212^ (*\u03bd*(CN^\u2212^)), terminal CO (*\u03bd*(CO)), and bridging CO (*\u03bd*(\u00b5-CO)) vibrational modes are shown above the spectra. Scale bars shown at 1870 cm^\u22121^ represent a difference of 0.5 milliabsorbance units.](pone.0015491.g003){#pone-0015491-g003}\n\nDiscussion {#s3}\n==========\n\nBy implementing several changes for heterologous \\[FeFe\\] hydrogenase expression, we achieved HydA1 and CpI yields more than 10-fold higher than previously reported for these enzymes ([Table 2](#pone-0015491-t002){ref-type=\"table\"}). Enabling concurrent anaerobic cell growth and T7 RNA polymerase induction was essential for increased heterologous protein production. In the absence of anoxic cell growth (i.e. incubation without glucose added to the medium), anaerobic hydrogenase and maturase accumulation levels were noticeably lower, as indicated by both hydrogenase activity assays and SDS-PAGE analysis (data not shown). This decrease in protein expression could be expected since translation is energy intensive due to the high entropic demands, and the rates of ATP synthesis are dramatically reduced under non-respiring conditions.\n\n10.1371/journal.pone.0015491.t002\n\n###### Comparing this work to previous reports for the production of the *C. reinhardtii* HydA1 hydrogenase.\n\n![](pone.0015491.t002){#pone-0015491-t002-2}\n\n  Microbial host                *C.r.* HydA1 Variant           Specific Activity[1](#nt103){ref-type=\"table-fn\"}   Purification Yield[2](#nt104){ref-type=\"table-fn\"}   Iron Content[3](#nt105){ref-type=\"table-fn\"}               Ref.\n  --------------------- ------------------------------------- --------------------------------------------------- ---------------------------------------------------- ---------------------------------------------- -------------------------------\n  *C. reinhardtii*                      HydA1                                       730\u00b1146                                               0.04                                               NR                            [@pone.0015491-Kamp1]\n                                        HydA1                                         935                                                0.001                                            3.9\u00b10.3                         [@pone.0015491-Happe1]\n  *S. oneidensis*               *Strep-*tag II--HydA1                               740\u00b156                                                0.5                                             6.1\u00b10.1                        [@pone.0015491-Sybirna1]\n  *C. acetobutylicum*           HydA1--*Strep-*tag II                                 625                                                 1.0                                                NR                        [@pone.0015491-vonAbendroth1]\n                                HydA1--*Strep-*tag II                                 760                                                 0.1                                                NR                           [@pone.0015491-Girbal1]\n  *E. coli*                     HydA1--*Strep-*tag II                                 150                                                 1.0                                                NR                            [@pone.0015491-King1]\n  ***E. coli***          **HydA1--** ***Strep-*** **tag II**                      **641\u00b188**                                           **30\u00b111**                                        **4.5\u00b10.2**                            **This Work**\n\nSpecific activities for H~2~ production rates are expressed in units of \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^ of HydA1.\n\nPurification yields are mg of HydA1 isolated per liter of cell culture.\n\nIron content is expressed in units of Fe atoms per HydA1 peptide; NR, not reported.\n\nAlong with increased yields, the purified hydrogenases are also highly active and contain a properly assembled H-cluster based on *in vitro* enzymatic activities as well as FTIR spectroscopic analyses. Previous studies using *E. coli* as an expression host for heterologous \\[FeFe\\] hydrogenase production reported specific activities much lower than those measured for the same protein purified from its native host ([Table 2](#pone-0015491-t002){ref-type=\"table\"}), likely due to incomplete hydrogenase maturation as well as possible loss of activity during purification [@pone.0015491-Akhtar1], [@pone.0015491-King1], [@pone.0015491-Laffly1]. The CpI and HydA1 enzymes produced with our optimized system have specific activities similar to those of the respective wild-type enzymes isolated from their native hosts [@pone.0015491-Kamp1], [@pone.0015491-Adams3]. The HydA1 hydrogenase isolated from *C. reinhardtii* was shown to evolve H~2~ with rates 650--950 \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^ [@pone.0015491-Kamp1], [@pone.0015491-Happe1], while CpI from *C. pasteurianum* was shown to evolve H~2~ with rates of 1100--5500 \u00b5mol H~2~\u00b7min^\u22121^\u00b7mg^\u22121^, depending on the assay conditions [@pone.0015491-Adams1], [@pone.0015491-Peters1], [@pone.0015491-Adams2], [@pone.0015491-Adams3]. While it appears that all HydA1 enzymes were active, we estimate that 20--40% of the CpI enzymes were not. Considering that CpI requires additional accessory iron-sulfur clusters, it is possible that this enzyme is more difficult to mature compared to HydA1. The iron contents for both hydrogenases that we examined were measured to be \u223c70% of that expected, which is consistent with the observation of lower-than-expected specific activity. Alternatively, experimental error associated with measuring iron quantities and protein concentration could also account for the lower-than-expected iron content, and some of the hydrogenases may have been inactivated during the purification process. The latter of these theories agrees with the 70% recovery of total hydrogenase activity during purification, which is further discussed below.\n\nLike the \\[FeFe\\] hydrogenases, HydE, HydF, and HydG also require ISCs. Thus, the benefits of the engineered *\u0394iscR* strain along with iron and cysteine supplementation could be expected since not one, but four Fe--S proteins must be overexpressed [@pone.0015491-Akhtar1]. The more pronounced benefit of the *\u0394iscR* strain for active CpI expression ([Fig. 2B](#pone-0015491-g002){ref-type=\"fig\"}) might be explained by the enzyme\\'s requirement for N-terminal ISCs. Unlike the algal HydA1 hydrogenase, CpI has three \\[4Fe--4S\\] clusters and one \\[2Fe--2S\\] cluster that participate in electron transfer to the H-cluster cofactor [@pone.0015491-Peters1]. As transcription of the *E. coli isc* operon is deregulated in the *\u0394iscR* strain [@pone.0015491-Schwartz1], higher expression of the corresponding native ISC proteins likely enhances the assembly, installation, and/or repair of these four accessory ISCs. The benefit of cysteine supplementation for active CpI expression in the *\u0394iscR* strain further supports this hypothesis, as increased *in vivo* levels of the cysteine desulfurase IscS may improve cysteine utilization for ISC biosynthesis.\n\nWith our methods and a single purification step, we obtained greater than 70% recovery of the *in vivo* hydrogenase activity. During our experimentation, we identified several factors that affected the overall efficacy of the purification process such as the necessity for DTH during protein purification. When buffers did not contain fresh DTH, both HydA1 and CpI more rapidly deactivated, in agreement with previously reported observations [@pone.0015491-Laffly1], [@pone.0015491-Happe1]. We also used a commercial lysis buffer (BugBuster Master Mix) to produce the lysates as this approach is simpler (given the anaerobic requirements) than alternative methods such as sonication and homogenization. High-yield expression was also beneficial for efficient recovery, since hydrogenase concentrations in the lysates (estimated to be 5--25 \u00b5M) are then higher than the K~d~ for *Strep*-tag II:*Strep*-Tactin adsorption (1 \u00b5M). The affinity tag location was another important factor as observed in other studies [@pone.0015491-Girbal1]. While the presence of a C-terminal affinity tag did not seem to negatively affect the solubility or catalytic properties of the hydrogenases, we could not produce soluble CpI with an N-terminal affinity tag. Moreover, HydA1 with an N-terminal affinity tag had a 50% lower specific activity, as indicated by both methyl viologen-based activity assays (data not shown). Traditional methods for isolating HydA1 without an affinity tag have combined multiple purification steps, and the majority of the hydrogenase activity (80--90%) was lost during the procedures [@pone.0015491-Kamp1], [@pone.0015491-Happe1]. Also, immobilized metal ion affinity chromatography (IMAC) could cause detrimental interactions between protein metal clusters and the resin [@pone.0015491-Soboh1], and high salt concentrations are generally used to recover the bound protein. *Strep*-Tactin affinity chromatography may also be more favorable for purifying metalloproteins compared to multi-step chromatography or IMAC. The *Strep*-Tactin approach involves a single chromatography step for efficient recovery of pure protein. Moreover, since buffer exchanges are not required to deplete high salt concentrations, the preparation of concentrated hydrogenase samples for FTIR spectroscopic analysis is simpler and less time consuming.\n\nThe production of two structurally different \\[FeFe\\] hydrogenases using heterologous maturases illustrates the versatility of this expression system. Infrared spectroscopic data confirm the presence of CO and CN^--^ ligands, indicating that both HydA1 and CpI contain an intact H-cluster identical to that of the protein produced in the native organisms and assembled by the native maturases. Despite the evolutionary diversity of \\[FeFe\\] hydrogenases, H-cluster biosynthesis and hydrogenase maturation appear to be highly conserved. Our results also underscore the modularity of the microbial world and the potential for dramatic evolutionary change through DNA exchange. It thus seems likely that the HydE, HydF, and HydG maturases from *S. oneidensis* could also activate other \\[FeFe\\] hydrogenases of interest (e.g. hydrogenases from *Thermatoga maritima* [@pone.0015491-Schut1] and *Desulfovibrio vulgaris* [@pone.0015491-Laffly1]). One advantage of using the *S. oneidensis* maturases is the similarity between *Shewanella and Escherichia*. For example, high yields and soluble expression of HydE, HydF, and HydG were observed ([Fig. 1C](#pone-0015491-g001){ref-type=\"fig\"}), even without codon optimizing the maturase genes.\n\nThe effectiveness of *E. coli* for inducible expression of heterologous proteins along with the variety of commercial recombinant DNA expression tools make this organism more desirable than others (e.g. *Clostridia*) for the large-scale production of hydrogenases for a variety of applications. In this work, we illustrate this advantage via the facile production of hydrogenases for IR spectroscopic analysis. Such analytical techniques generally require large quantities of hydrogenase per sample (\\>500 \u00b5g) at concentrations greater than 5 mg\u00b7mL^\u22121^. With our system, sufficient quantities of HydA1 and CpI hydrogenase for multiple IR spectroscopic measurements can be obtained from a single 250 mL culture. In comparison, isolation of HydA1 from its native host requires 8 L of culture and multiple purification steps to produce enough hydrogenase for one IR spectroscopic measurement [@pone.0015491-Kamp1].\n\nConclusions {#s3a}\n-----------\n\nThe commercialization of technologies that use \\[FeFe\\] hydrogenases will most certainly require economical approaches for producing these complex oxygen-sensitive enzymes. Furthermore, much remains unknown about H-cluster biosynthesis and the hydrogenase maturation process. We provide a new, effective protocol for producing these enzymes to greatly facilitate both technology development and hydrogenase biochemistry research. The methods we describe could also be extended for producing other enzymes associated with anaerobic metabolism such as \\[NiFe\\] hydrogenases and nitrogenases.\n\nMaterials and Methods {#s4}\n=====================\n\n\\[FeFe\\] Hydrogenase and Maturase Expression Constructs {#s4a}\n-------------------------------------------------------\n\nThe *C. reinhardtii hydA1* and *C. pasteurianum hydA* genes were used for expression of the HydA1 and CpI \\[FeFe\\] hydrogenases, respectively. Both genes were previously codon-optimized for expression in *E. coli* [@pone.0015491-Boyer1]. The coding sequencing for a C-terminal *Strep*-tag II\u00ae extension (IBA GmbH) with a two residue linker (5\u2032-SAWSHPQFEK-3\u2032) was added by PCR amplifying the hydrogenase genes from the plasmids pY71 *shydA1\\** [@pone.0015491-Kuchenreuther1] and pK7 *shydA* [@pone.0015491-Boyer1]. PCR products were then cloned into the pET-21(b) expression vector (Novagen). The plasmid pACYCDuet-1--*hydGX*--*hydEF* [@pone.0015491-Kuchenreuther1] was used for expression of the *S. oneidensis* \\[FeFe\\] hydrogenase maturases HydE, HydF, and HydG. Multiple cloning sites I and II contain the *hydGX* and *hydEF* nucleotide sequences, respectively. The *hydX* sequence (Accession code AAN56899) is a part of the *S. oneidensis* \\[FeFe\\] hydrogenase operon and encodes a soluble protein with no identified functions. The *petF* gene from *Synechocystis sp.* PCC 6803 was PCR amplified from the pK7 expression vector and cloned into the pET-21(b) plasmid [@pone.0015491-Boyer2]. All expression constructs were confirmed by DNA sequencing and transformed into the *E. coli* strains BL21(DE3) (Novagen) and BL21(DE3) *\u0394iscR* by selection with the appropriate antibiotics.\n\nRecombinant Expression and Purification of Active Hydrogenases {#s4b}\n--------------------------------------------------------------\n\n*E. coli* strains BL21(DE3) and BL21(DE3) *\u0394iscR* contained the pACYCDuet-1--*hydGX*--*hydEF* plasmid and one of the two pET-21(b) hydrogenase plasmids. Cells were grown in LB Miller growth medium supplemented with kanamycin (40 mg\u00b7L^\u22121^; only when using the *\u0394iscR* strain), chloramphenicol (25 mg\u00b7L^\u22121^), ampicillin (100 mg\u00b7L^\u22121^), 0.5% w/v glucose (\u223c25 mM), and 100 mM MOPS/NaOH (final pH of medium was 7.4). The *\u0394iscR* strain contains a chromosomal substitution of the *iscR* gene with another gene conferring resistance to kanamycin [@pone.0015491-Akhtar1]. 10--50 mL cultures were grown for investigating the effects of cell strains and substrates, while 50--250 mL cultures were grown for hydrogenase purification work. Initially, all cultures were grown aerobically at 25\u00b0C until an OD~600~ of 0.3--0.5. They were then moved into an anaerobic glove box (Coy Laboratory Products) containing 98% N~2~ and 2% H~2~ prior to IPTG-based T7 RNA polymerase induction and heterologous protein expression. While ferric ammonium citrate (2 mM) was added to the growth medium prior to inoculation, both cysteine (2 mM) and sodium fumarate (25 mM) were added with IPTG (0.5 mM) within the anaerobic glove box. Cultures were sealed and incubated at 25\u00b0C for 16--24 hours following induction.\n\nFor investigating media formulations and protein expression by different strains, cells from 1 mL of culture were pelleted at 4,000\u00d7g and resuspended in 100 \u00b5L of anaerobic BugBuster\u00ae Master Mix lysis solution (Novagen) containing an additional 25 mM Tris/HCl (pH 8.0), 25 mM KCl, 3 mM sodium dithionite (DTH), 1 mM dithiothreitol (DTT), 2% v/v glycerol, 0.1% v/v Tween 20, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 2 \u00b5M resazurin as an oxygen indicator. After cell lysis (incubation at 25\u00b0C for 20 min), lysates were clarified by centrifugation at 14,000\u00d7g. Hydrogenase activities in cell lysates were measured using the methyl viologen reduction assay described below. Total protein content of lysates was determined using a commercial assay (Bio-Rad) based on the method of Bradford [@pone.0015491-Bradford1], and the extent of heterologous protein expression was visualized using polyacrylamide gel electrophoresis with SDS-PAGE gels (Invitrogen).\n\nHydrogenase purifications were carried out while maintaining strict anaerobic conditions. After centrifugation and lysis as described above, approximately 1 mL of *Strep*-Tactin\u00ae Superflow\u00ae high capacity resin (IBA GmbH) was used per 50 mL of cell culture for purification. Wash and elution buffers contained the above lysis buffer additives excluding the BugBuster Master Mix and PMSF, and active hydrogenase was eluted with 2.5 mM D-desthiobiotin. Elution fractions were evaluated for active hydrogenase using the methyl viologen reduction assay, and fractions with high activity were pooled. Protein concentrations were measured with the Bradford assay, and hydrogenase iron content was measured using a ferrozine-based colorimetric assay [@pone.0015491-Fish1]. Hydrogenase samples for IR spectroscopic studies were anaerobically concentrated to \u223c100 \u00b5M using a 10 mL stirred cell and a 5 kD MWCO membrane (Amicon). Hydrogenase samples were not frozen prior to characterization and spectroscopic analysis.\n\nHydrogenase Activity Assays {#s4c}\n---------------------------\n\nHydrogenase activities were measured at 37\u00b0C in both the H~2~ consumption and H~2~ evolution directions using methods previously described [@pone.0015491-Boyer1]. Generally, 1--25 ng of hydrogenase was tested. H~2~ uptake was quantified with a methyl viologen reduction assay. 200 \u00b5L assay solutions contained 50 mM Tris/HCl (pH 8.0) and 2 mM methyl viologen. Absorbance was measured at 578 nm for 1--2 min following addition of lysate or purified hydrogenase. Methyl viologen reduction rates were adjusted by subtracting background activities, which were generally less than 1% of the hydrogenase activities. An extinction coefficient for reduced methyl viologen of 9.78 mM^\u22121^\u00b7cm^\u22121^ was used to calculate H~2~ oxidation rates, in which 2 moles of methyl viologen are reduced per mole of H~2~ consumed. Hydrogen production was quantified using DTH-reduced methyl viologen (5 mM) or the ferredoxin PetF from *Synechocystis* (50 \u00b5M) as the electron donating substrate. 9.5 mL glass vials contained 1 mL of 100 mM MOPS/KOH buffer (pH 6.8), 100 mM KCl, 25 mM DTH, and either methyl viologen or ferredoxin. Upon hydrogenase addition, vials were sealed and the headspace was sparged with 100% N~2~ for 2 min. H~2~ quantities in the headspace were then quantified after 15--30 min of incubation using a ShinCarbon ST 100/120 mesh column (Resteck) with a Hewlett Packard 6890 gas chromatograph (Hewlett Packard). For PetF ferredoxin production, the *petF* gene from *Synechocystis sp.* PCC 6803 was first cloned from the pK7 plasmid [@pone.0015491-Boyer1] into the pET-21(b) vector, and subsequently transformed into BL21(DE3) *\u0394iscR*. Both PetF expression and purification using ammonium sulfate precipitation followed by anion exchange chromatography were carried out as previously described [@pone.0015491-Boyer2], [@pone.0015491-Nakamura1].\n\nFourier Transform IR Spectroscopy {#s4d}\n---------------------------------\n\nInfrared spectra were measured using a Bruker IFS/66s FTIR spectrometer interfaced to a home-built stopped-flow drive system. The IR sample cuvette and drive system were maintained inside an anaerobic chamber (O~2~\\<1.1 ppm) as previously described [@pone.0015491-Thorneley1]. The sample cuvette was maintained at 25\u00b0C with a calibrated path length of 47.6 \u00b5m. For IR spectroscopic measurements, one drive syringe contained the protein sample. A second syringe contained either the elution buffer without any protein or buffer saturated with exogenous carbon monoxide. Spectra were measured at 4 cm^\u22121^ resolution from 1000 sample scans, and the average spectrum was improved with a background correction.\n\nThe authors wish to thank Junko Tanno and Hiroshi Takami of Fujeribio, Inc. as well as Patrik Jones for their assistance with supplying the *E. coli* strain BL21(DE3) *\u0394iscR*.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This work was funded by the United States Department of Energy BioEnergy Science Program. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: JMK SPC JRS. Performed the experiments: JMK CSG-S ASB SJG. Analyzed the data: JMK CSG-S ASB SJG. Wrote the paper: JMK JRS. Discussed results and commented on the manuscript: JMK CSG-S ASB SJG SPC JRS.\n"}
+{"text": "Introduction\n============\n\nDiabetes is a type of metabolic disease characterized by hyperglycemia resulting from either defective insulin secretion, insulin action or the two ([@b1-br-0-0-921]). The most prevalent type of diabetes is type 2 diabetes mellitus (T2DM), which is one of the leading causes of morbidity globally, as well as the third-highest risk factor for premature mortality ([@b2-br-0-0-921]).\n\nIn Mexico, T2DM has led mortality rates since 2005 and today it represents the leading cause of death in the country ([@b3-br-0-0-921],[@b4-br-0-0-921]) with a prevalence of 11.8% ([@b5-br-0-0-921]). Costs associated with medical treatment of T2DM are \\~450 million dollars annually ([@b6-br-0-0-921]), while \\~75% of diagnosed patients do not observe adequate glycemic control even with medical assistance. These factors make T2DM a critical concern for the Mexican State\\'s public health and research systems.\n\nWhile genetic factors causing T2DM have not yet been sufficiently defined, they are currently under extensive study. Numerous T2DM-associated genes present as single nucleotide polymorphisms (SNPs) whose frequencies vary among different populations. Genetic variations associated with either pharmacological targets or drug metabolism are of particular interest, as different responses to pharmacological treatments may be explained by the presence of a genetic mutation or the combination of various genotypes ([@b7-br-0-0-921]).\n\nThe initial pharmacological treatments for T2DM are oral hypoglycemiants (OH), with sulfonylureas (SU) and metformin (ME) being the two most commonly administered in developing countries. SU are a type of oral OHs, which inhibit ATP-sensitive potassium channels (K~ATP~), thus inducing glucose-independent insulin release by the \u03b2-pancreatic cells ([@b8-br-0-0-921]). However, not all T2DM patients respond to the anti-diabetic action of SU (primary failure) and those who initially respond adequately may experience a decrease in its efficacy following the years (as variable as 1--2 to 10 or more years) of treatment (secondary failure) ([@b9-br-0-0-921],[@b10-br-0-0-921]). SUs are frequently combined with ME, a drug which reduces hepatic glucose production and insulin resistance ([@b11-br-0-0-921]). It has been observed that the short-term reduction of glycated hemoglobin A1c (HbA1c) is similar in SU and ME monotherapies ([@b12-br-0-0-921]), and that this drug combination reduces HbA1c more efficiently than SU alone ([@b13-br-0-0-921]). When compared, observations on the adjuvant effects of SU/ME-based treatments are inconclusive; while certain authors have associated hypoglycemic events of different severity and weight increase to SUs, either alone or in combination with ME ([@b14-br-0-0-921]), others have reported no effect on body weight when combining SUs with ME ([@b15-br-0-0-921]). However, due to their low cost and accessibility, SUs (alone or combined with ME) remain the most frequent first-line T2DM treatment in the world, particularly in developing countries, such as Mexico ([@b16-br-0-0-921],[@b17-br-0-0-921]). The UK Prospective Diabetes Study demonstrated that only 25% of patients achieved glycemic control of \\<7% in HbA1c over a nine-year follow-up period of monotherapy on either SU or ME ([@b10-br-0-0-921]).\n\nOf the genetic polymorphisms that have already been reviewed extensively and whose clinical implications have already been analyzed ([@b18-br-0-0-921],[@b19-br-0-0-921]), there are a number, which appear to be well suited for pharmacogenetic studies. Various SNPs have been reported among K~ATP~-channel encoding genes \\[potassium voltage-gated channel subfamily J member 11 (*KCNJ11*) and ATP binding cassette subfamily C member 8 (*ABCC8*)\\] as the therapeutic target of SU ([@b20-br-0-0-921]). Many of these genes are associated with T2DM predisposition or progression, as well as with SU response variability. A specific SNP frequency may vary between different populations; therefore, it is important to evaluate and compare its distribution among different human populations, in order to better understand whether there is an association between drug response variability, patients\\' glycemic control and genetic architecture.\n\nIn the present study, the frequencies of three pharmacologically important SNPs are described in a Mestizo Mexican (MM) population and, in order to compare these with other reported populations, the distribution of *KCNJ11* rs5219 (E23K), *ABCC8* rs757110 (S1369A) and rs1799854 (\u22123C/T) is presented in MM T2DM patients. The aim of the present study is to increase the understanding of the genetic characteristics of specific populations. This may facilitate with elucidating the causes of therapeutic failure and the findings may also be extrapolated and/or compared to other populations in order to improve treatment options and patient management.\n\nMaterials and methods\n=====================\n\n### Patient selection and study design\n\nThis study was observational and included 247 T2DM patients recruited from July 2014 to October 2016 from two health centers: 145 from Centro de Salud Portales and 102 T2DM patients from Centro de Salud Mixcoac, both located in Mexico City\\'s Benito Ju\u00e1rez Health Jurisdiction (Mexico). Out of the total adult patient population, 165 were females while 82 were males. Patient eligibility criteria were as follows: Self-proclaimed MM ancestry of at least three generations; age between 18 and 90 years; individuals diagnosed with T2DM according to the American Association of Diabetes criteria ([@b1-br-0-0-921]); individuals taking OHs, the SU glibenclamide alone or combined with ME, for at least 3 months. All participants were enrolled in their health centers after providing written informed consent. Patient clinical history, anthropometrics and biochemical characteristics were obtained from their clinical records (a summary of this data is presented in [Table I](#tI-br-0-0-921){ref-type=\"table\"}).\n\n### Genotyping\n\nGenomic DNA was obtained by taking 6 ml peripheral blood through arm phlebotomy in glass EDTA-tubes (Vacutainer^\u00ae^; BD Biosciences, Franklin Lakes, NJ, USA) from each patient. Patients were fasted at the time of blood sampling. Genomic DNA was isolated from 200 \u00b5l total blood, using the UltraClean^\u00ae^ BloodSpin^\u00ae^ DNA Isolation kit (Mo Bio Laboratories, Inc., Carlsbad, CA, USA), and the DNA-extraction protocol was performed according to the manufacturer\\'s instructions. DNA quality was achieved by electrophoresis in Invitrogen 1% agarose gels (UltraPure\u2122 Agarose; Thermo Fisher Scientific, Inc., Waltham, MA, USA), at 100 V over 50 min. DNA concentration was measured by spectrophotometry (Jenway 7305; Cole-Parmer Ltd., Staffordshire, UK). The *KCNJ11* rs5219 (E23K), *ABCC8* rs757110 (S1369A) and rs1799854 (\u22123C/T) polymorphisms were determined using 20 ng total genomic DNA per reaction, by allelic discrimination via quantitative polymerase chain reaction (qPCR) using a ViiA\u2122 7 Real-Time PCR system and TaqMan^\u00ae^ SNP assays (Applied Biosystems; Thermo Fisher Scientific, Inc. Waltham, MA, USA) using the standard cycling conditions as follows: An initial denaturation stage 95\u00b0C for 10 min, 40 cycles of denaturation 95\u00b0C for 15 sec, and annealing at 60\u00b0C for 1 min, and post read stage or final extension at 60\u00b0C for 30 sec.\n\n### Statistical analysis\n\nThe SNP frequencies of *KCNJ11* rs5219, ABCC8 rs757110 and rs1799854 were determined through direct counting, the remaining analyses were performed using SPSS 23.0 for Windows (IBM Corp., Armonk, NY, USA). The SNP frequencies of the patients from both health centers were compared by performing \u03c7^2^ test of independence. In addition, biochemical and anthropometric data were compared using two-way analysis of variance (ANOVA). The total frequencies for each SNP were compared with those of other populations using the \u03c7^2^ test of independence. P\\<0.05 with 95% CI was considered to indicate a statistically significant difference.\n\nResults\n=======\n\n### Mexican-Mestizo sample characteristics\n\nNo statistically significant differences were identified after comparing biochemical and anthropometric data from the individuals from the two health centers (data not shown). Within the whole sample, there were approximately twice as many females as there were males (162 females and 83 males); 61.5% of the patients were overweight (55.7% of females and 73.8% of males); \\~25% of the subjects presented with grade I obesity (27.9% in females and 20.0% in males). The mean period since the first T2DM diagnosis was 10 years, with a mean of 7.55\u00b10.15% for HbA1c and 141.86\u00b14.06 mg/dl for fasting plasma glucose (FPG). A summary of anthropometrics and biochemical characteristics of our subjects is presented in [Table I](#tI-br-0-0-921){ref-type=\"table\"}.\n\nNo significant differences were observed in genotypic and allelic frequency distribution between the two health centers (data not shown). The genotyping frequencies of the three SNPs in the whole sample were in Hardy-Weinberg equilibrium ([Table II](#tII-br-0-0-921){ref-type=\"table\"}). The obtained allelic and genotypic frequencies were compared with those reported for other non-diabetic populations using data from the 1000 Genomes Project ([@b21-br-0-0-921]) and from literature on T2DM patients; the comparative of each SNP is presented in [Tables III](#tIII-br-0-0-921){ref-type=\"table\"}--[V](#tV-br-0-0-921){ref-type=\"table\"} ([@b22-br-0-0-921]--[@b38-br-0-0-921]). For the present study, studies were included that involved populations of T2DM subjects that were comparable with the current sample.\n\n### SNP comparison with other populations\n\n*KCNJ11* rs5219 (E23K) allelic and genotypic frequencies in MM individuals were not identified to be different between all of the T2DM populations that were compared \\[predominantly European (EUR) and Asian\\]. To the best of our knowledge, there were only two other studies from Mexican rs5219 allelic frequencies, which demonstrated no difference compared with the MM population: T2DM Mestizo from Yucatan and T2DM Mayans. In the 1000 Genomes Project populations, MM frequencies were only statistically different from African individuals (P=2.17E-12 and P=2.31E-19 for allelic and genotypic frequencies, respectively) and admixed American-Colombians from Medellin (P=0.002 and P=7.20E-05 for allelic and genotypic frequencies, respectively; [Table III](#tIII-br-0-0-921){ref-type=\"table\"}).\n\nMM *ABCC8* rs757110 (S1369A) allelic and genotypic frequencies behave similarly to rs5219. The only statistical differences were observed when compared with African individuals (P=1.45E-13 and P=1.72E-23 for allelic and genotypic frequencies, respectively) and American-Colombians from Medellin (P=6.13E-04 and P=2.52E-06 for allelic and genotypic frequencies, respectively). Currently, to the best of our knowledge, there are no other studies regarding allelic or genotypic frequencies of this SNP in a Mexican population ([Table IV](#tIV-br-0-0-921){ref-type=\"table\"}).\n\n*ABCC8* rs1799854 (\u22123C/T) SNP allelic and genotypic frequencies from MM were significantly different from all of the compared T2DM populations (primarily European and Asian). In comparison to non-diabetic populations from the 1000 Genomes Project, MM allelic and genotypic frequencies were significantly different from African individuals, East Asians, the majority of EUR individuals, and other admixed Americans from Colombia, Peru, Puerto Rico and individuals of Mexican ancestry from Los Angeles. Only in Iberic EUR and South Asian individuals did the 1000 Genomes Project report non-significant differences in allelic and genotypic frequencies compared with MM ([Table V](#tV-br-0-0-921){ref-type=\"table\"}). Currently, to the best of our knowledge, there are no other studies on the frequencies of these SNPs among Mexican populations of non-diabetic or diabetic individuals.\n\nDiscussion\n==========\n\nIn the present study, in the MM from Mexico City, rs5219 allelic and genotypic frequencies were not different from the majority of the reported populations of healthy and diabetic individuals among other ethnic groups, except for the African and Colombian subjects; the same observation applies to rs757110 ([Tables III](#tIII-br-0-0-921){ref-type=\"table\"} and [IV](#tIV-br-0-0-921){ref-type=\"table\"}). It appears that the distribution of these polymorphisms prevails among the majority of populations, while the ancestral allele is most frequent among African individuals. The results from the Colombian population may reflect the ancestral admixture history of the country, where the African component is widely spread across the Pacific and Caribbean regions ([@b39-br-0-0-921]--[@b41-br-0-0-921]).\n\nIn Mexican populations, rs5219 is the only SNP reported for K~ATP~-coding genes, exhibiting no differences between the alleles of healthy and T2DM subjects, even when the geographical and ethnic profile of the three Mexican samples were markedly different: The healthy volunteers were Mestizo from the South East ([@b22-br-0-0-921]), the T2DM group had Mayan Amerindian ancestry ([@b23-br-0-0-921]) and our group (MM) was formed by Mestizo individuals, primarily from the central area of Mexico.\n\nTo the best of our knowledge, this is the first report of *ABCC8* rs757110 and rs1799854 SNPs in Mexican populations, and more specifically, in diabetic Mexican patients. rs757110 allelic and genotypic frequencies behaved similarly to rs5219 SNP, which was to be expected as it is well known that these SNPs form a haplotype. Yet, rs1799854 SNP frequencies seem extremely different to those reported by the 1000 Genomes Project and the literature on various diabetic populations ([Table V](#tV-br-0-0-921){ref-type=\"table\"}). A comparison between allelic and genotypic frequencies in the investigated group and those populations reported by the 1000 Genomes Project, demonstrated differences between all but three samples: South Asian, European and Iberic. The Iberic population may have affected the comparison result of the EUR sample, as it was included in the EUR group in the 1000 Genome Project data. This was consistent with the fact that other EUR diabetic populations from different ethnic origins were statistically different when compared with the MM group (UK, Croatia and Poland). These results may reflect the Hispanic admixture dting from Mexico\\'s colonial past; however, it is interesting and unexpected that this was not observed in the other Latin-American populations reported. It was demonstrated by the results of the previous Mexican health and nutrition national survey (ENSANUT 2012) that diabetes diagnosis and first level medical attention have improved considerably in Mexico ([@b5-br-0-0-921]), yet the percentage of treated patients actually maintaining adequate glycemic control remains poor. If this lack of control is genetic, at least partially, the differences observed in Mexican genetic SNP frequencies may significantly contribute to explaining treatment failure. This first screening may facilitate with understanding where focus and investigations are required to establish whether genetic structure and pharmacological failure are associated.\n\nIt would be interesting to observe whether a healthy control group of MM individuals from central Mexico behaves the same as diabetic subjects, and to analyze groups of different ethnicity within the country to establish whether those results are consistent with or different from the present study.\n\nThe SNPs, rs5219 (E23K) and rs757110 (S1369A) form a haplotype, where K23/A1369 has been identified as a risk genotype ([@b35-br-0-0-921],[@b42-br-0-0-921],[@b43-br-0-0-921]). While electrophysiological studies have demonstrated that channels containing the K allele and K23/A1369 are less sensitive to ATP inhibition ([@b44-br-0-0-921],[@b45-br-0-0-921]), the SU response of these polymorphic channels appears to be different depending on the drug. For example, it has been shown that K23/A1369 channels are more sensitive to SU gliclazide, yet these same channels are less sensitive to inhibition by SUs, such as tolbutamide and glimepiride, while glibenclamide demonstrated no significant inhibition difference in any haplotype ([@b45-br-0-0-921],[@b46-br-0-0-921]). In another study, K allele carriers exhibited significantly higher secondary treatment failure than E allele homozygous ([@b29-br-0-0-921]) treated with SU and ME; however, Dawed *et al* ([@b19-br-0-0-921]) demonstrated that secondary failure on patients treated with a combination of SU and ME, carrying the K allele polymorphism of rs5219 may be more involved with diabetes progression than with SU response ([@b19-br-0-0-921]).\n\nIn Mexico, the most common first level treatment for T2DM combines ME with glibenclamide administration ([@b5-br-0-0-921]). In this first analysis, whose objective was mainly SNP frequency description, the authors included patients receiving glibenclamide or ME either as a mono- or combined therapy, as the SNP distribution is not affected by treatment. In future studies, the aim will be to investigate clinical implications of using ME-only treated patients as a control group to distinguish the ME effect.\n\nThe *ABCC8* rs1799854 polymorphism has been associated with a predisposition for diabetes ([@b25-br-0-0-921],[@b47-br-0-0-921],[@b48-br-0-0-921]). In terms of pharmacogenetics, certain studies have found that the TT genotype may be associated with an increase in HbA1c and triglyceride levels in SU-treated diabetics ([@b33-br-0-0-921],[@b49-br-0-0-921]), and deemed partially determinant of hyperglycemia-cardiovascular risk factor in rs1799854 heterozygotes ([@b37-br-0-0-921]). However, other authors report no significant associations when analyzing FPG and BMI ([@b28-br-0-0-921],[@b50-br-0-0-921]). Even when results in the case of this SNP are contradictory, the difference in the allelic and genotypic frequencies observed in our sample in comparison with the majority of reported cases, indicates a considerable requirement to evaluate their implications in diabetes progression and drug response in the MM population.\n\nThe SNP frequencies of rs5219 and rs757110 polymorphisms appear to be well conserved among the majority of populations, including Mexicans, there is not yet a clear association between them and the pharmacological effects. The present results may contribute to future studies to clarify whether there is a real association.\n\nThe aim of the present study was to describe the genetic architecture of three pharmacogenetically important SNPs of ABCC8 and KCNJ11 within an MM population. To the best of our knowledge, this study is the first to report allelic and genotypic frequencies of ABCC8 rs757110 and rs1799854 SNPs in an MM population. Diabetes is a major concern in Mexico, and current pharmacological treatment is considered insufficient, as shown by the latest national health survey. Therefore, understanding the characteristics of our population is a priority for elucidating a viable hypothesis to improve our knowledge of this complex pathology. It is known that interindividual responses to SU are affected by clinical factors, such as baseline glucose levels, disease duration, \u03b2-cell function and insulin resistance levels ([@b51-br-0-0-921]). However, multiple gene interaction may explain the marginal impact of each individual SNP, indicating, the necessity to construct an interaction model.\n\nThe present study was conducted as part of the Universidad Nacional Aut\u00f3noma de M\u00e9xico (National Autonomous University of Mexico-UNAM) Biological Sciences PhD postgraduate academic program fulfillment, and it was supported by the Programa de Apoyo a Proyectos de Investigaci\u00f3n e Innovaci\u00f3n Tecnol\u00f3gica (Support Program for Research and Technological Innovation Projects-PAPIIT) as the IN218216 Direcci\u00f3n General de Apoyo a Personal Acad\u00e9mico (General Directorate for Academic Personnel Support-DGAPA)-UNAM project. The authors would like to thank Mar\u00eda de los \u00c1ngeles Granados-Silvestre for her technical assistance and training in the use of the qPCR equipment and sample processing; Professor Marta Menjivar-Hiraeta for the access granted to her facility and equipment to perform qPCR experiments. To Dr Isaac Gonz\u00e1lez Romero (Centro de Salud Portales), Dr Mario Alberto Tinoco Centeno and Mr. Jos\u00e9 Antonio Rojas (Centro de Salud Mixcoac) for providing his technical expertise for sample collection.\n\nMM\n\n:   Mestizo Mexican\n\nT2DM\n\n:   type 2 diabetes mellitus\n\nSNP\n\n:   single nucleotide polymorphism\n\nSU\n\n:   sulfonylurea\n\nK~ATP~\n\n:   ATP-sensitive potassium channel\n\nME\n\n:   metformin\n\n###### \n\nDemographic baseline disease characteristics.\n\n  Parameter                        Total (n=247)   Female (n=165)   Male (n=82)\n  -------------------------------- --------------- ---------------- --------------\n  Age (years)                      60.43\u00b10.75      60.05 \u00b10.91      61.24\u00b11.34\n  Weight (kg)                      70.87\u00b11.06      69.25\u00b11.23       74.40\u00b11.98\n  Body mass index (kg/m^2^)        29.51\u00b10.37      30.26\u00b10.45       27.89\u00b10.62\n  Diabetes diagnosis (years)       10.27\u00b10.61      9.57\u00b10.64        11.92\u00b11.34\n  Triglycerides (mg/dl)            199.06\u00b18.02     199.59\u00b19.44      197.90\u00b115.24\n  Glycated hemoglobin A1c (%)      7.55\u00b10.15       7.53\u00b10.19        7.60\u00b10.26\n  Fasting plasma glucose (mg/dl)   141.86\u00b14.06     144.55\u00b15.22      135.76\u00b15.99\n  Cholesterol (mg/dl)              194.82\u00b12.71     197.24\u00b13.45      189.39\u00b14.12\n\nValues are presented as means \u00b1 standard error of the mean.\n\n###### \n\nGenotypic frequencies of the analyzed polymorphisms.\n\n  SNP       Frequency   \\%      P-value\n  --------- ----------- ------- ---------\n  E23K                          0.69\n  \u00a0\u00a0GG      93          37.7    \n  \u00a0\u00a0GA      112         45.3    \n  \u00a0\u00a0AA      42          17.0    \n  \u00a0\u00a0Total   247         100.0   \n  \u22123C/T                         0.55\n  \u00a0\u00a0CC      125         50.6    \n  \u00a0\u00a0CT      98          39.7    \n  \u00a0\u00a0TT      24          9.7     \n  \u00a0\u00a0Total   247         100.0   \n  S1369A                        1.63\n  \u00a0\u00a0AA      75          30.4    \n  \u00a0\u00a0AC      131         53.0    \n  \u00a0\u00a0CC      41          16.6    \n  \u00a0\u00a0Total   247         100.0   \n\nP\\<3.84 using the \u03c7^2^ test for Hardy-Weinberg equilibrium.\n\n###### \n\nPotassium voltage-gated channel subfamily J member 11 rs5219 (E23K) allelic and genotypic frequency comparison.\n\n  Author (year)                        Population          T       n      G       N      p\u00a5                                                     TT      n     TG      n     GG      n     p\u00a5                                                     Refs.\n  ------------------------------------ ------------------- ------- ------ ------- ------ ------------------------------------------------------ ------- ----- ------- ----- ------- ----- ------------------------------------------------------ ---------------------\n  1000 Genomes Project                                                                                                                                                                                                                           \n  Auton *et al*                        MM                  0.603   298    0.397   196    --                                                     0.377   93    0.453   112   0.17    42                                                           ([@b21-br-0-0-921])\n  \\(2015\\)                             AFR                 0.977   1291   0.023   31     2.17E-12^[a](#tfn3-br-0-0-921){ref-type=\"table-fn\"}^   0.956   632   0.047   27    0.003   2     2.31E-19^[a](#tfn3-br-0-0-921){ref-type=\"table-fn\"}^   \n                                       EAS                 0.662   667    0.338   341    0.387                                                  0.429   216   0.466   235   0.105   53    0.386                                                  \n                                       SAS                 0.604   591    0.396   387    0.988                                                  0.38    186   0.448   219   0.172   84    0.997                                                  \n                                       EUR                 0.647   651    0.353   355    0.52                                                   0.4     201   0.495   249   0.105   53    0.406                                                  \n                                       EUR IBS             0.617   132    0.383   82     0.839                                                  0.374   40    0.486   52    0.14    15    0.815                                                  \n                                       AMR                 0.707   491    0.293   203    0.121                                                  0.496   172   0.424   147   0.081   28    0.084                                                  \n                                       AMR CLM             0.803   151    0.197   37     0.002^[a](#tfn3-br-0-0-921){ref-type=\"table-fn\"}^      0.638   60    0.33    31    0.032   3     7.20E-05^[a](#tfn3-br-0-0-921){ref-type=\"table-fn\"}^   \n                                       AMR MXL             0.594   76     0.406   52     0.897                                                  0.328   21    0.531   34    0.141   9     0.54                                                   \n                                       AMR PEL             0.682   116    0.318   54     0.243                                                  0.435   37    0.494   42    0.071   6     0.092                                                  \n                                       AMR PUR             0.712   148    0.288   60     0.104                                                  0.519   54    0.385   40    0.096   10    0.086                                                  \n  Type 2 diabetic                                                                                                                                                                                                                                \n  Hernandez-Escalante *et al* (2014)   AMR YCN             0.633   164    0.367   95     0.662                                                  --      --    --      --    --      --    --                                                     ([@b22-br-0-0-921])\n  Lara-Riegos *et al* (2015)           AMR MYN             0.654   75     0.346   40     0.449                                                  --      --    --      --    --      --    --                                                     ([@b23-br-0-0-921])\n  He *et al* (2008)                    EAS (China)         0.62    124    0.38    76     0.805                                                  0.35    35    0.54    54    0.11    11    0.34                                                   ([@b24-br-0-0-921])\n  Yokoi *et al* (2006)                 EAS (Japan)         0.614   1954   0.386   1226   0.873                                                  0.384   610   0.462   734   0.155   246   0.959                                                  ([@b25-br-0-0-921])\n  Holstein *et al* (2009)              EUR (Germany)       0.604   116    0.396   76     0.988                                                  0.385   37    0.437   42    0.177   17    0.975                                                  ([@b27-br-0-0-921])\n  Gloyn *et al* (2001)                 EUR (UK)            0.593   427    0.407   293    0.885                                                  0.369   133   0.447   161   0.183   66    0.97                                                   ([@b28-br-0-0-921])\n  Sesti *et al* (2009)                 EUR (Italy)         0.642   674    0.358   376    0.569                                                  0.385   202   0.514   270   0.101   53    0.338                                                  ([@b29-br-0-0-921])\n  Ragia *et al* (2012)                 EUR (Grece)         0.668   235    0.332   117    0.339                                                  0.455   80    0.426   75    0.119   21    0.423                                                  ([@b30-br-0-0-921])\n  Javorsky *et al* (2012) (Slovakia)   EUR                 0.599   121    0.401   81     0.954                                                  0.366   37    0.465   47    0.168   17    0.984                                                  ([@b31-br-0-0-921])\n  Klen *et al* (2014)                  EUR (Slovenia)      0.622   194    0.378   118    0.783                                                  0.378   59    0.487   76    0.135   21    0.76                                                   ([@b32-br-0-0-921])\n  Nicolac *et al* (2009)               EUR (Croatia)       0.607   277    0.393   179    0.954                                                  0.382   87    0.452   103   0.167   38    0.997                                                  ([@b33-br-0-0-921])\n  Chistiakov *et al* (2009)            EUR (Russia)        0.496   128    0.504   130    0.128                                                  0.217   28    0.558   72    0.225   29    0.045^[a](#tfn3-br-0-0-921){ref-type=\"table-fn\"}^      ([@b34-br-0-0-921])\n  Sokolova *et al* (2015)              WAS (East Russia)   0.649   1926   0.351   1042   0.501                                                  0.428   635   0.442   656   0.13    193   0.647                                                  ([@b35-br-0-0-921])\n\nP\\<0.05 (95% confidence interval); p\u00a5, c^2^ test of independence; MM, type 2 diabetic Mestizo Mexican (1000 Genomes Project third release); AFR, African individuals (1000 Genomes Project third release); EAS, East Asian individuals (1000 Genomes Project third release); SAS, East South Asian individuals (1000 Genomes Project third release); EUR, European individuals (1000 Genomes Project third release); IBS, Iberian population in Spain; AMR, Admixed American; CLM, Colombians from Medellin, Colombia; MXL, Mexican ancestry from Los Angeles, USA; PEL, Peruvians from Lima, Peru; PUR, Puerto Ricans from Puerto Rico (1000 Genomes Project third release); YCN, type 2 diabetic from Mestizo Yucatan; MYN, T2D Mayan.\n\n###### \n\nATP binding cassette subfamily C member 8 S1369A allelic and genotypic frequency comparison.\n\n  Author (year)             Population                T       n      G       n      p\u00a5                                                     TT      n     TG      n     GG      n     p\u00a5                                                     Refs.\n  ------------------------- ------------------------- ------- ------ ------- ------ ------------------------------------------------------ ------- ----- ------- ----- ------- ----- ------------------------------------------------------ ---------------------\n  1000 Genomes Project                                                                                                                                                                                                                      \n  Auton *et al*             MM                        0.569   281    0.431   213                                                           0.304   75    0.53    131   0.166   41                                                           ([@b21-br-0-0-921])\n  \\(2015\\)                  AFR                       0.975   1289   0.025   33     1.45E-13^[a](#tfn4-br-0-0-921){ref-type=\"table-fn\"}^   0.953   630   0.044   29    0.003   2     1.72E-23^[a](#tfn4-br-0-0-921){ref-type=\"table-fn\"}^   \n                            EAS                       0.639   644    0.361   364    0.195                                                  0.395   199   0.488   246   0.117   59    0.331                                                  \n                            SAS                       0.585   572    0.415   406    0.819                                                  0.354   173   0.462   226   0.184   90    0.652                                                  \n                            EUR                       0.648   652    0.352   354    0.252                                                  0.404   203   0.489   246   0.107   54    0.238                                                  \n                            IBS                       0.617   132    0.383   82     0.49                                                   0.374   40    0.486   52    0.14    15    0.567                                                  \n                            AMR                       0.693   481    0.307   213    0.069                                                  0.47    163   0.447   155   0.084   29    0.030^[a](#tfn4-br-0-0-921){ref-type=\"table-fn\"}^      \n                            CLM                       0.793   149    0.207   39     6.13E-04^[a](#tfn4-br-0-0-921){ref-type=\"table-fn\"}^   0.606   57    0.372   35    0.021   2     2.52E-06^[a](#tfn4-br-0-0-921){ref-type=\"table-fn\"}^   \n                            MXL                       0.586   75     0.414   53     0.808                                                  0.328   21    0.516   33    0.156   10    0.932                                                  \n                            PEL                       0.682   116    0.318   54     0.098                                                  0.435   37    0.494   42    0.071   6     0.041                                                  \n                            PUR                       0.678   141    0.322   67     0.111                                                  0.462   48    0.433   45    0.106   11    0.061                                                  \n  Type 2 diabetics                                                                                                                                                                                                                          \n  Zhang *et al* (2007)      EAS (China)               0.565   130    0.435   100    0.954                                                  0.33    38    0.47    54    0.2     23    0.676                                                  ([@b38-br-0-0-921])\n  Yokoi *et al* (2006)      EAS Japan                 0.592   1884   0.408   1296   0.742                                                  0.358   570   0.468   744   0.174   276   0.655                                                  ([@b25-br-0-0-921])\n  Klen *et al* (2014)       EUR (Slovenia)            0.619   193    0.381   119    0.471                                                  0.378   59    0.481   75    0.141   22    0.536                                                  ([@b32-br-0-0-921])\n  Nicolac *et al* (2009)    EUR (Croatia)             0.607   277    0.393   179    0.585                                                  0.395   90    0.425   97    0.18    41    0.301                                                  ([@b33-br-0-0-921])\n  Sokolova *et al* (2015)   East Russia (West Asia)   0.623   1763   0.377   1065   0.436                                                  0.393   556   0.46    651   0.146   207   0.414                                                  ([@b35-br-0-0-921])\n\nP\\<0.05 (95% confidence interval); p\u00a5, c^2^ test of independence; CDMX, type 2 diabetics from Mestizo Mexico City (all of the following are 1000 Genomes Project third release): AFR, African; EAS, East Asian; SAS, South Asian; EUR, European; IBS, Iberian population in Spain; AMR, Admixed American; CLM, Colombians from Medellin, Colombia; MXL, Mexican ancestry from Los Angeles, USA; PER, Peruvians from Lima, Peru; PUR, Puerto Ricans from Puerto Rico.\n\n###### \n\nATP binding cassette subfamily C member 8 \u22123C/T Allelic and genotypic frequency comparison.\n\n  Authors (year)            Population      C       n      T       n      p\u00a5                                                     CC      n     CT      n     TT      n     p\u00a5                                                     Refs.\n  ------------------------- --------------- ------- ------ ------- ------ ------------------------------------------------------ ------- ----- ------- ----- ------- ----- ------------------------------------------------------ ---------------------\n  1000 Genomes Project                                                                                                                                                                                                            \n  Auton *et al*             MM              0.704   348    0.296   146    --                                                     0.506   125   0.397   98    0.097   24    --                                                     ([@b21-br-0-0-921])\n  \\(2015\\)                  AFR             0.862   1139   0.138   183    0.006^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      0.741   490   0.241   159   0.018   12    7.96E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   \n                            EAS             0.449   453    0.551   555    2.41E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.198   100   0.502   253   0.3     151   2.20E-06^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   \n                            SAS             0.681   666    0.319   312    0.724                                                  0.481   235   0.401   196   0.119   58    0.865                                                  \n                            EUR             0.58    583    0.42    423    0.067                                                  0.33    166   0.499   251   0.171   86    0.031^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      \n                            EUR IBS         0.612   131    0.388   83     0.17                                                   0.393   42    0.439   47    0.168   18    0.169                                                  \n                            AMR             0.464   322    0.536   372    5.35E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.225   78    0.478   166   0.297   103   1.30E-05^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   \n                            AMR CLM         0.516   97     0.484   91     0.006^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      0.266   25    0.5     47    0.234   22    0.001^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      \n                            AMR MXL         0.422   54     0.578   74     5.10E-05^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.172   11    0.5     32    0.328   21    1.35E-07^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   \n                            AMR PEL         0.312   53     0.688   117    1.91E-08^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.071   6     0.482   41    0.447   38    3.15E-14^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   \n                            AMR PUR         0.567   118    0.433   90     0.044^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      0.346   36    0.442   46    0.212   22    0.022^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      \n  Type 2 diabetics                                                                                                                                                                                                                \n  He *et al* (2008)         EAS (China)     0.41    82     0.59    118    2.50E-05^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.14    14    0.54    54    0.32    32    1.03E-08^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   ([@b24-br-0-0-921])\n  Yokoi *et al* (2006)      EAS (Japan)     0.474   1507   0.526   1673   8.91E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.233   371   0.481   765   0.286   454   2.90E-05^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   ([@b25-br-0-0-921])\n  Matharoo *et al* 2013     SAS India       0.568   227    0.433   173    0.044^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      0.405   81    0.325   65    0.27    54    0.006^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      ([@b36-br-0-0-921])\n  Gloyn *et al* (2001)      EUR (UK)        0.464   412    0.536   476    5.35E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.191   85    0.545   242   0.264   117   3.50E-06^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   ([@b28-br-0-0-921])\n  Nicolac *et al* (2009)    EUR (Croatia)   0.489   223    0.511   233    0.002^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^      0.197   45    0.583   133   0.219   50    1.30E-05^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   ([@b33-br-0-0-921])\n  Dworacka *et al* (2007)   EUR (Poland)    0.45    36     0.55    44     2.54E-04^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   0.25    10    0.4     16    0.35    14    6.00E-06^[a](#tfn5-br-0-0-921){ref-type=\"table-fn\"}^   ([@b37-br-0-0-921])\n\nP\\<0.05 (95% confidence interval); p\u00a5, c^2^ test of independence; CDMX, type 2 diabetics from Mestizo Mexico City (all of the following are 1000 Genomes Project third release): AFR, African; EAS, East Asian; SAS, South Asian; EUR, European; IBS, Iberian population in Spain; AMR, Admixed American; CLM, Colombians from Medellin, Colombia; MXL, Mexican ancestry from Los Angeles, USA; PER, Peruvians from Lima, Peru; PUR, Puerto Ricans from Puerto Rico.\n"}
+{"text": "Actinomycetes are free living, saprophytic, filamentous bacteria, and a major source for the production of antibiotics ([@B1]). They are found in soil, fresh water and marine water environments ([@B2]). Actinomycetes provided many important bioactive compounds of high commercial value and screened for new bioactive substances ([@B3]). These bacteria are an important group of microorganisms due to their ability to produce a wide array of secondary metabolites, such as antibiotics, antitumor agents, immunosuppressive agents, cosmetics, vitamins, nutritional materials, herbicides, pesticides, anti-parasitic agents and enzymes ([@B1], [@B4], [@B5]). Around 23000 bioactive secondary metabolites produced by microorganisms have been reported. Over 10000 of these compounds are produced by actinomycetes, representing 45% of all bioactive microbial metabolites discovered. Among actinomycetes, around 7600 compounds are produced by *Streptomyces* species ([@B6]). As the frequency of novel bioactive compounds obtained from terrestrial actinomycetes decreased, it had been emphasized that actinomycetes from marine sediments might be valuable for the isolation of novel strains which could potentially yield a broad spectrum of secondary metabolites ([@B7]-[@B9]). However, it has been resolved whether actinomycetes forms part of the marine microbial community of sediment samples originated from terrestrial environments and was simply carried out to sea in the form of resistant spore ([@B10]).\n\nIt has been reported that marine actino-mycetes not only have several new species, but also have plenty novel structures with potent bioactivities ([@B11]). Several novel bioactive compounds were discovered from aquatic actinomycetes for example rifamycin from *Micromonospora*sp. ([@B12]); salinosporamide-A, an anticancer metabolite from *Salinispora* sp. ([@B13]) (Feling, 2003); marinomycins from *Marinophilus* sp. ([@B14]); abyssomicin-C from *Verrucosispora* sp. and marinopyrroles from *Streptomyces* sp. ([@B12], [@B15]). The appearances of multidrug resistant pathogenic strains caused substantial morbidity and mortality especially among the elderly and immunocompro-mised patients. To overcome this situation, there is an interest to improve or discover novel class antibiotics that have different mechanisms of action worldwide ([@B16]).\n\nAccording to incomplete statistics, the number of novel compounds obtained from marine actinomycetes in the 21^st^ is more than twice of the last century. This study was focused on the actinomycetes of marine sediments collected from the Caspian Sea. For the first time, an effort was made to screen different marine sediments which are a large unscreened and diverse ecosystem for the isolation of potent antibiotic producing actino-mycetes.\n\nMaterials and Methods\n=====================\n\n**Sample collection**\n\nSamples were collected from the sediments of Caspian Sea at the depths of 5-10 m by Van vein grab (0.2 m^2^). Two sampling stations were located along of the Caspian Sea with the following latitudes 36\u00b043\\'N and 36\u00b044\\'N. The surface of each grab sample was aseptically collected and processed within 30 minutes.\n\n**Sample treatment**\n\nThe samples were subjected to physical pretreatment method in order to facilitate the isolation of actinomycetes. Heat treatments were performed by holding sediment samples in a water bath (Memmert) at 50 \u00b0C for 60 minutes. All samples were diluted with sterile 0.9 % saline prior to inoculation in triplicate onto isolation plates ([@B11]).\n\n**Isolation of** **actinomycetes**\n\nActinomycetes were isolated by serial dilution method from sediments ([@B17]). Stock solution was prepared by diluting 1 g of sediment in 9 ml of sterile saline water and shaking well by using a vortex mixer (IKA). From the stock solution, 1 ml was used to prepare the final volume of 10^-2^ and 10^3^ by serial dilution method. Samples were inoculated on Starch Casein Agar (SCA) (composition: soluble starch: 10 g, K~2~HPO~4~: 2 g, KNO~3~: 2 g, casein: 0.3 g, MgSO~4~.7H~2~O: 0.05 g, CaCO~3~: 0.02 g, FeSO~4~.7H~2~O: 0.01 g, agar: 15 g, filtered sea water: 1000 ml and pH: 7.0\u00b10.1), Yeast Extract Malt Extract Agar (ISP2) (Composition: yeast extract: 4 g, malt extract: 10 g, dextrose: 4 g, agar: 15 g, filtered sea water: 1000 ml and pH: 7.3) and Kuster\\'s Agar (composition: glycerol: 10 g, casein: 0.3 g, KNO~3~: 2 g, K~2~HPO~4~: 2 g, soluble starch: 0.5 g, asparagine: 0.1 g, FeSO~4~.7H~2~O: 0.01 g, CaCO~3~: 0.02 g, MgSO~4~.7H~2~O: 0.05 g, agar: 15 g, filtered sea water: 1000 ml and pH: 7.0\u00b10.1). Each medium was supplemented with 25 \u00b5g ml^\u22121^ nystatin to minimize contamination with fungi and 10 \u03bcg ml^\u22121^ nalidixic acid to minimize contaminant growth ([@B11], [@B18]). Plates were incubated for 7 to 20 days at 28 \u00b0C. Then the colonies with a tough or powdery texture, dry or folded appearance and branching filaments with or without aerial mycelia were sub-cultured on slants SCA ([@B19]). Until further use, the slants were kept in cold room at 4 \u00b0C ([@B6]).\n\n**Preliminary screening for antibacterial activity using cross-streak method**\n\nIsolated strains MN1 to MN44 were inoculated onto nutrient agar plates by streak in the center. The plates were incubated at 28 \u00b0C for 3 days. Six bacteria including *Staphylococcus aureus*ATCC 25923, *Bacillus subtilis*PTCC 1156, *Escherichia coli*PTCC 1533, *Pseudomonas* *aeruginosa* PTCC 1074, *Salmonella* *typhi* PTCC 1609 and *Klebsiella* *pneumonia* were used as test organisms. A pure colony of test bacteria was transferred into fresh nutrient broth and incubated at 37 \u00b0C for 24 hours until the visible turbidity and density equal to that of 0.5 McFarland. After adjusting the turbidity, sterile cotton swab was dipped into the bacterial suspension and streaked perpendicular to the antagonist on the agar medium. The plates were incubated at 37 \u00b0C for 24 hours. The microbial inhibitions were observed by determining the diameter of the inhibition zones.\n\n**Extraction of antimicrobial compounds**\n\nThe selected antagonistic actinomycetes were inoculated into 100 ml of actinomycete isolation broth (composition: glycerol: 5.0 g, sodium propionate: 4.0 g, sodium caseinate: 2.0 g, K~2~HPO~4~: 0.5 g, asparagine: 0.1 g, MgSO~4~\u00b77H~2~O: 0.1 g, FeSO~4~\u00b77H~2~O: 1.0 mg, water: 1000 ml and pH: 8.0\u00b10.1) and incubated in orbital shaker at 28 \u00b0C and 190 rpm for 7 days. To extract the antimicrobial compounds, the cultures were filtered then centrifuged (Sigma) at 6000 rpm, 10 minutes ([@B1]). The supernatant was transferred aseptically into a screw capped bottle and stored at 4 \u00b0C for further assay.\n\n**Secondary screening of antibacterial activity using disk diffusion method**\n\nThe antimicrobial activities of those extracts were tested against different test organisms by using agar disc diffusion method as described by Kirby-Bauer with modification ([@B20]).\n\nLate exponential phase of the test bacteria were prepared by inoculating 1% (v/v) of the cultures into the fresh Muller-Hinton broth (Merck) and incubating on an orbital shaker at 37 \u00b0C and 100 rpm overnight. Before using the cultures, they were standardized with a final cell density of approximately 10^8^ cfu ml^-1^. Muller-Hinton agar (Merck) were prepared and inoculated from the standardized cultures of the test organisms then spread as uniformly as possible throughout the entire media. Sterile paper discs (6 mm diameter, Padtan, Iran) were impregnated with 30 \u00b5l of the extracts then allowed to dry. The impregnated disc was introduced on the upper layer of the seeded agar plate and incubated at 37 \u00b0C for 24 hours.\n\nThe antibacterial activities of the extracts were compared with known antibiotic tetracycline (30 \u00b5g/disc) as positive control and ethyl acetate (30 \u00b5l/disc) as negative control. Antibacterial activity was evaluated by measuring the diameter of inhibition zone (mm) on the surface of plates and the results were reported as Mean \u00b1 SD after three repeats ([@B21]). The potential actinomycetes isolates were selected from the primary and secondary screening then characterized by morphological and physiological methods for further studies.\n\n**Exoenzymatic assay**\n\nThe potential actinomycetes isolates were screened for hydrolytic exoenzymatic activities including amylase and protease. These tests were conducted on Yeast Extract Malt Extract Agar (ISP2) containing 1% soluble starch for amylolytic activity and 1% skimmed milk for proteolytic activity.The presence of amylase was visualized by decolorized halo around the culture due to starch digestion. Proteolytic activity was observed by clearing of the milk around the colony ([@B6]).\n\nResults\n=======\n\nActinomycete colonies were readily isolated from marine sediments on SCA and ISP 2, but the growth on Kuster\\'s agar was very poor. The macroscopic appearance of the isolates showed leathery and powdery colonies in Starch Casein Agar (SCA) ([Fig. 1](#F1){ref-type=\"fig\"}). A total of 44 actinomycetes were isolated from the sediments of Caspian Sea. Among the isolated actinomycetes, 24 strains showed antibacterial activities against at least one of the tested bacteria using preliminary screening. The results were summarized in [table 1](#T1){ref-type=\"table\"}. Of all the 24 isolates, seven best antagonistic actinomycetes isolates were selected for further studies. The potential isolates are MN2, MN3, MN38, MN39, MN40, MN41, and MN44. The morphological characteristics of the selected isolates are shown in [table 2](#T2){ref-type=\"table\"}. The strains were gram positive, filamentous with long spore chain except the MN39 which produced single spore. For antibacterial activities assay, the crude extracts of the potential isolates were subjected for secondary screening using disc diffusion methods ([Fig. 2](#F2){ref-type=\"fig\"}). The results demon-strated that among the tested bacteria, the crude extracts exhibited highest antibacterial activity against *B. subtilis* and *S. aureus* ([Table. 3](#T3){ref-type=\"table\"}). The values obtained for activity of MN38 and MN39 strains show great activity against all test bacteria whereas the growth inhibitory of the crude extract of MN3 strain showed low activityagainst *K. pneumonia* ([Table 3](#T3){ref-type=\"table\"}). The results revealed that MN44, MN38 andMN39 were active against *S. aureus*; MN41, MN38 against *E. coli*; MN2, MN3 and MN44 against *P. aeruginosa*; MN2, MN39, MN40 and MN44 against *K. pneumonia*; MN38, MN39 and MN44 against *B. subtilis*; MN40 and MN44 against *S. typhi*. Physiological and biochemical characteristics indicate that all isolates showed the ability of starch and protein hydrolysis.\n\n![Morphological appearance of MN38 isolates (A) and MN39 (B) on Starch Casein Agar](ijmcm-2-064-g001){#F1}\n\n![Antibacterial activities of the crude extract of MN38 isolates (3) and MN44 (4) against S. aureus (A) and B. subtilis(B) using disc diffusion method](ijmcm-2-064-g002){#F2}\n\n###### \n\nPreliminary screening of actinomycetes for antimicrobial activity using cross-streak method\n\n  Isolates   **Test bacteria**                       \n  ---------- ------------------- ---- ---- ---- ---- ----\n  MN1        \\-                  \\+   \\-   \\+   \\-   \\+\n  MN2        \\+                  \\+   \\+   \\+   \\+   \\+\n  MN3        \\+                  \\+   \\+   \\+   \\+   \\+\n  MN6        \\-                  \\+   \\-   \\+   \\-   \\+\n  MN8        \\+                  \\-   \\-   \\+   \\-   \\+\n  MN13       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN16       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN17       \\+                  \\+   \\-   \\+   \\-   \\-\n  MN18       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN19       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN21       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN23       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN27       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN30       \\+                  \\+   \\-   \\-   \\-   \\+\n  MN31       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN32       \\-                  \\+   \\-   \\+   \\-   \\-\n  MN33       \\+                  \\+   \\-   \\+   \\-   \\+\n  MN35       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN36       \\-                  \\+   \\-   \\+   \\-   \\+\n  MN38       \\+                  \\+   \\+   \\+   \\+   \\+\n  MN39       \\+                  \\+   \\+   \\+   \\+   \\+\n  MN40       \\+                  \\+   \\+   \\+   \\+   \\+\n  MN41       \\+                  \\+   \\+   \\+   \\+   \\+\n  MN44       \\+                  \\+   \\+   \\+   \\+   \\+\n\n###### \n\nMorphological and physiological characteristics of potent actinomycetes isolates\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------\n  **characteristics**    **Isolates**                                                                                                         \n  ---------------------- ---------------------- -------------------- --------------------- ------------------- -------------- --------------- --------------\n  Aerial mycelia         Green                  Gray                 Gray                  White               Gray           White           White\n\n  Reverse color          Yellow-brown- orange   Yellow-brown-green   Yellow-brown- green   Yellow-brown- red   Yellow-brown   Yellow-brown    Yellow-brown\n\n  Spore-bearing hyphae   Verticillate (MV)      Simple\\              Simple\\               Simple\\             Simple\\        Verticillate\\   Simple\\\n                                                (R)                  (F)                   (RA)                (R)            (BIV)           (R)\n\n  Starch hydrolysis      \\+                     \\+                   \\+                    \\+                  \\+             \\+              \\+\n\n  Casein hydrolysis      \\+                     \\+                   \\+                    \\+                  \\+             \\+              \\+\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------\n\nMV):Monoverticillus, (R): Straight - Rectus, (F): Flexibilis, (RA): Retinaculum-Apertum, (BIV): Biverticillus-: Negative; +: Positive.\n\n###### \n\nAntibacterial activity of the potential actinomycetes isolates using Kirby--Bauer disk diffusion method\n\n  **Test Strain**   **Zone of growth inhibition (mm)**                                                                      \n  ----------------- ------------------------------------ ---------- ----------- ---------- ---------- ---------- ---------- ----------\n  *E. coli*         13.0\u00b11.4                             17.0\u00b10.7   11.0\u00b11.4    23.5\u00b10.7   14.0\u00b11.4   13.0\u00b11.1   20.7\u00b11.5   18.6\u00b11.1\n  *P. aeruginosa*   18.0\u00b10.7                             20.0\u00b10.7   15.5\u00b10.7    11.0\u00b11.4   12.5\u00b10.7   18.0\u00b10.5   10.0\u00b11.5   NE\n  *S. aureus*       10.5\u00b10.7                             13.0\u00b11.4   20.0\u00b10.4    20.5\u00b10.7   12.5\u00b10.7   20.3\u00b11.5   12.0\u00b10.4   24.0\u00b11.0\n  *B. subtilis*     12.0\u00b10.5                             11 0\u00b10.6   27.0 \u00b10.7   23.0\u00b11.4   13.0\u00b10.7   22.0\u00b11.7   13.0\u00b11.4   23.6\u00b10.6\n  *S. typhi*        14.0\u00b11.4                             14.0\u00b11.4   16.0\u00b11.4    16.0\u00b11.4   18.0\u00b11.4   20.0\u00b11.4   11.0\u00b11.7   19.3\u00b10.6\n  *K. pneumonia*    17.0\u00b10.4                             7.0\u00b11.4    15.0\u00b11.4    24.0\u00b11.4   17.0\u00b11.4   18.0\u00b11.4   10.0\u00b11.4   14.3\u00b10.6\n\nTet.: Tetracycline (30\u00b5g/disk)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 NE: No Effect\n\nDiscussion\n==========\n\nCurrently, the incidence of multidrug resistant organisms is increasing and compromising the treatment of a growing number of infectious diseases. As a result, there is an urgent need for developing new drugs which are effective against current antibiotic resistant pathogens. Actino-mycetes have been proven as a potential source of bioactive compounds and the richest source of secondary metabolites ([@B22]). The isolation of antibacterial compounds from the freshwater environment is of interest to isolate novel bioactive actinomycetes. Actinomycetes form 10% of the total bacteria colonizing marine aggregates. Marine habitat has been proven as an outstanding and fascinating resource for innovating new and potent bioactive producing microorganisms. Only very few reports are available on the occurrence and distribution of antagonistic actinomycetes in the marine environment. Recent investigations indicate the tremendous potential of marine actinomycetes, particularly *Streptomyces* species as a useful and sustainable source of new bioactive natural products ([@B1]). The present study was aimed to isolate actinomycetes from marine environment and screen them for the production of secondary metabolites. The medium was supplemented with nystatin to eliminate the fungal contamination. The same method was previously done by Sambamurthy and Ellaiah (1974) that used amphotericin B as antifungal agent ([@B23]). The production of antibiotic substance is dependent on sea water ([@B24]). In this study also, the SCA, ISP 2 and Kuster\\'s agar were prepared using sterile sea water. Okazaki and Okami (1972) observed that *Streptomyces* species showed efficient antagonistic activity ([@B25]). Holt *et al*. (1994) identified the isolated actinomycetes based on the colony morphology and gram staining ([@B26]). We have identified the actinomycetes by the presence of powdered colonies on the surface of agar plate. Actinomycetes are gram positive and filamentous in nature. Kokare*et al.* (2004) have stated the filamentous nature of actinomycetes which are gram positive ([@B27]). During the screening of the novel secondary metabolites, isolated actinomycetes showing more activity against gram positive bacteria than gram negative bacteria were often encountered. This was similar to the findings of another study ([@B27]). It has been found that estuarine actinomycetes, which remained largely ignored, show promising antibacterial activities ([@B28], [@B29]). In the present study, 44 actinomycetes were isolated from sediments and 24 showed wide range of inhibition zone in the primary screening. More yield of crude extract was produced with ethyl acetate solvents. The reason for the increased yield was due to the lack of water and complete miscibility in organic solvents (ethyl acetate) of the growth supernatant. Thus, the results of this investigation revealed that the marine actinomy-cetes collected from the sediments of Caspian Sea might be a potent source of novel antibiotics. It is anticipated that isolation, characterization and study of actinomycetes can be useful for the discovery of novel species of bacteria producing bioactive compounds.\n\n**Conflict of interest:**Non declared.\n"}
+{"text": "![](edinbmedj74401-0066){#sp1 .157}\n\n![](edinbmedj74401-0067){#sp2 .158}\n\n![](edinbmedj74401-0068){#sp3 .159}\n\n![](edinbmedj74401-0069){#sp4 .160}\n\n![](edinbmedj74401-0070){#sp5 .161}\n\n![](edinbmedj74401-0071){#sp6 .162}\n\n![](edinbmedj74401-0072){#sp7 .163}\n\n![](edinbmedj74401-0073){#sp8 .164}\n\n![](edinbmedj74401-0074){#sp9 .165}\n\n![](edinbmedj74401-0075){#sp10 .166}\n\n![](edinbmedj74401-0076){#sp11 .167}\n\n![](edinbmedj74401-0077){#sp12 .168}\n\n![](edinbmedj74401-0078){#sp13 .169}\n\n![](edinbmedj74401-0079){#sp14 .170}\n\n![](edinbmedj74401-0080){#sp15 .171}\n\n![](edinbmedj74401-0081){#sp16 .172}\n\n![](edinbmedj74401-0082){#sp17 .173}\n\n![](edinbmedj74401-0083){#sp18 .174}\n\n![](edinbmedj74401-0084){#sp19 .175}\n\n![](edinbmedj74401-0085){#sp20 .176}\n\n![](edinbmedj74401-0086){#sp21 .177}\n\n![](edinbmedj74401-0087){#sp22 .178}\n\n![](edinbmedj74401-0088){#sp23 .179}\n\n![](edinbmedj74401-0089){#sp24 .180}\n"}
+{"text": "Conflict of interest: None.\n\nFunding source: None.\n\nDear Editor,\n\nSince December 2019, an outbreak of Coronavirus disease 2019 (COVID\u201019) that started in Wuhan, China, has been spreading across the world and was assessed as a pandemic from March 11, 2020, by the World Health Organization (WHO). Several manifestations related to COVID\u201019 have been reported, but small data about skin manifestations in these patients is available. In Mexico, COVID\u201019 alert began on February 27, and to date no associated clinical skin manifestations have been reported in our country. Here is our experience with two patients.\n\nCase 1 {#ijd14958-sec-0002}\n======\n\nA previously healthy 33\u2010year\u2010old male arrived to the Emergency Room (ER) on March 26, 2020, with an 11\u2010day history of fever, diarrhea, altered mental status with aggressiveness, and a 4\u2010day maculopapular skin rash affecting the trunk and upper and lower limbs. Three days previous to arrival, typhoid fever was diagnosed, and he was treated with ciprofloxacin and metamizole. No cough or dyspnea was reported.\n\nDue to a current measles epidemic outbreak in Mexico, the patient was considered suspicious and evaluated by the Neurology Department, diagnosing acute infectious meningoencephalitis. Laboratory findings included significant anemia (hemoglobin 12.7\u00a0mg/dl), elevated lactic dehydrogenase (4268.7\u00a0UI/l), and normal chest x\u2010ray. However, IgM and IgG for measles were negative. Due to persistent skin rash, biopsy of maculopapular rash was performed to exclude erythema multiforme; histologic slides showed nonspecific mild to moderate dermatitis with isolated areas with interphase dermatitis and isolated apoptotic bodies (Fig.\u00a0[1](#ijd14958-fig-0001){ref-type=\"fig\"}).\n\n![(a) Macular erythematous confluent nonpruritic rash affecting upper trunk. (b) Histological images show some isolated areas with interphase dermatitis and lymphoepitheliosis (Hematoxylin and eosin, \u00d7100)](IJD-9999-na-g001){#ijd14958-fig-0001}\n\nOn the sixth day after admission, the\u00a0patient showed progressing acute respiratory distress syndrome, and molecular assays of respiratory panel were performed with a negative result. Nasopharyngeal swabs were positive for SARS\u2010COV\u20102, and chest CT showed characteristic ground\u2010glass opacities in both lungs. The patient had a fulminant course in the ICU and died on the 15th day.\n\nCase 2 {#ijd14958-sec-0003}\n======\n\nA 36\u2010year\u2010old male with the diagnosis of human immunodeficiency virus (HIV), referred from another health institution, arrived to the ER on April 14, 2020, with 7\u2010day history of asthenia, adynamia, and diarrhea. Skin examination showed a 4\u2010day heterogeneous dermatosis that started in the lower limbs and was characterized by confluent erythematous macules forming larger plaques followed later with pruritic papules spreading upward to the dorsolumbar region (Fig.\u00a0[2](#ijd14958-fig-0002){ref-type=\"fig\"}). Two days later, he developed a holocranial headache, nasal congestion, dry cough, and fever up to 39.5\u00a0\u00b0C.\n\n![(a, b) Macular symmetrical, heterogeneous dermatosis affecting limbs with small pruritic plaques that spread upwards to posterior trunk (c)](IJD-9999-na-g002){#ijd14958-fig-0002}\n\nLaboratory findings included lymphopenia 1.7\u00a0\u00d7\u00a010^3^\u00a0U/l, elevated erythrocyte sedimentation rate (ESR) 11\u00a0mm/h, elevated ferritin (400\u00a0ng/ml), elevated D\u2010dimer (846), elevated lactic dehydrogenase (227.4\u00a0UI/l), and\u00a0elevated C\u2010reactive protein (5.04\u00a0mg/dl). Nasopharyngeal swabs were positive for SARS\u2010COV\u20102, but chest x\u2010ray and CT chest were normal. The patient remained stable with mild symptoms and was discharged for home isolation.\n\nBefore the\u00a0COVID\u201019 outbreak, exanthems have not been described for other coronaviruses, and recently some skin manifestations associated with this virus have been described, including transient livedo reticularis,[^1^](#ijd14958-bib-0001){ref-type=\"ref\"} erythematous rash (like our cases), widespread urticaria, chickenpox\u2010like vesicles,[^2^](#ijd14958-bib-0002){ref-type=\"ref\"} and skin rash with petechiae (dengue\u2010like),[^3^](#ijd14958-bib-0003){ref-type=\"ref\"} but none of these descriptions included the skin manifestations as a previous presentation of common associated COVID\u201019 respiratory symptoms,[^4^](#ijd14958-bib-0004){ref-type=\"ref\"} and none included skin biopsy. Indisputably, we need further information to understand how this virus affects skin and describe all the dermatological manifestations of this emerging disease.\n\nWe wish to express our appreciation to Dr. Hiram J. Serrano\u2010Ort\u00edz for the technical support.\n"}
+{"text": "The ethics requirements of the committees that approved this study, and the partnership agreement made with school administrators, require that the data be kept by the authors. Therefore, all data are available from the corresponding author, Dr. Javier G. Ogembo (<javier.ogembo@umassmed.edu>), University of Massachusetts Medical School, for researchers who meet the criteria for access.\n\nIntroduction {#sec001}\n============\n\n> *\"If you want the truth*, *ask a child...\"*\n>\n> ---Portion of Danish proverb (Sandheden skal man h\u00f8re fra b\u00f8rn)\n\nRegistration of a child's birth is one of the more fundamental, far reaching, and troublesome, steps for securing the health and happiness of a society \\[[@pone.0149925.ref001]\\]. Registration is noted as a vital, basic human right. It allows children to be counted and acknowledged by a government, and is the first line of protection for shielding children from underage labor, marriage, prostitution, trafficking, and military conscription \\[[@pone.0149925.ref002], [@pone.0149925.ref003], [@pone.0149925.ref004]\\]. Registration, *and increasingly having a physical birth certificate*, is also vital to ensure that children receive access to healthcare or vaccination \\[[@pone.0149925.ref005], [@pone.0149925.ref006], [@pone.0149925.ref007]\\], and is required for many children to enter schooling, travel, gain employment; and therefore is directly tied to their own, and their communities,' health and future development.\n\nAt the same time, despite need to formally recognize and record child births, this is one of the more persistently problematic issues affecting many developing regions \\[[@pone.0149925.ref008], [@pone.0149925.ref009], [@pone.0149925.ref010]\\]. Recent findings by UNICEF \\[[@pone.0149925.ref007]\\] estimate that over 30% of the world's children---56% in sub-Saharan Africa where we conducted the present study---are not registered. Even fewer have birth certificates. Low registration levels have been attributed to a number of reasons, such as lack of infrastructure, transportation, sociocultural factors causing delay or avoidance, or even lack of basic motivation in parents; and has been met with a number of on the ground assessments and implementations \\[[@pone.0149925.ref009], [@pone.0149925.ref011], [@pone.0149925.ref012], [@pone.0149925.ref013]\\]. However, despite much effort in study and policy implementations, it is also a well established finding that current approaches have often been disappointingly ineffective at substantially increasing numbers of children registered at birth, or even later in their development \\[[@pone.0149925.ref013]\\]. This has led to calls for new answers regarding why children's births have not been recorded by parents and policy solutions.\n\nInterestingly, one avenue that has not been considered is that of asking children *what they themselves* know about the need, the process and the improvement in certification of their own births. The importance of such study might be framed in several ways. It may be illuminating to discover what children do know about their rights and necessities regarding registration because it does play such an important role regarding their opportunities and health, or if children's understanding differs materially from that of adults. This fits a growing realization in social and policy research, especially in Western countries, that children have a voice that should be considered \\[[@pone.0149925.ref014], [@pone.0149925.ref015]\\], as well as findings that children's own knowledge, or parent-child communication, may play a key role in improving programs related to demographics and health \\[[@pone.0149925.ref016]\\]. At the same time, with older children who may soon be specifically impacted by requirements for birth certificates in order to continue with schooling or employment, it may also be enlightening to determine whether they are aware of these impending requirements or of their rights, which also may contribute to policy improvement.\n\nPerhaps most important, children may also be a particular prescient window into the motivations of their parents. In our recent study of parent attitudes in the same region \\[[@pone.0149925.ref013]\\], we argued that the current ineffectiveness of registration policy might in fact often be attributed to a disconnect between what parents say in a survey or in studies used for policy planning regarding what impedes them from registering, and their true motivations, which may often be driven more by informed, personal indifference. Parent answers may also be driven by a desire to explain-away previous inaction by producing a list of acceptable factors that absolve the parent. A confirmation bias may also lead parents to believe the difficulties and issues that they list, although these may not bear out objective measurement. This may create a self-fulfilling inability to resolve low registration via policy implementations, which target these same factors. On the other hand, it is a well-regarded truism that children are often surprisingly insightful regarding the motivations and actions of their parents; and often with a candor avoided by adults \\[[@pone.0149925.ref017]\\]. To paraphrase a classic truism \\[[@pone.0149925.ref018]\\]: Children seldom misquote. More often, they state what parents *should have said* about the issues affecting their households, community, or---potentially---their children's health.\n\nThis is the goal of this paper. We introduce findings from a survey of children's understanding regarding registration of their births, as well as their ideas regarding policy decisions and motivations of their parents. This was conducted as one aspect of a larger analysis, previously reported in \\[[@pone.0149925.ref013]\\], in which we assessed the adult population in an area of rural Kenya, argued to provide a specifically representative example of under-registration of births. We will briefly review the background situation and our previous study findings regarding parent attitudes/answers toward registration of births. We then introduce results from our analysis of these parents' children, targeting late elementary/junior high age students. We conclude by connecting the findings to the wider adult population and by considering policy suggestions derived from the ideas and awareness of these students. While---it is important to note---this paper does not seek to offer the systematic representative sampling of registration totals, as put forth in our previous work, we hope that the present manuscript can serve as a valuable extension and alternative viewpoint for this topic.\n\nReview: Kenya and Registration/Certification of Births {#sec002}\n======================================================\n\nKenya marks an intriguing case for discussion of child registration and birth certificates \\[13 for review\\]. On one hand, it possesses many advantages placing it towards the head of the curve for developing nations. Since the founding of the modern state in the beginning of the past century it has enjoyed stable growth \\[[@pone.0149925.ref019]\\] and now has a relatively stable democracy and emerging economy as well as infrastructure \\[[@pone.0149925.ref020]\\]. These improvements are also supported by a population largely motivated to achieve socioeconomic progress and by a stabilizing government which has enacted multiple reforms in public and private sectors, with the country largely expected to attain middle-income status by 2030 \\[[@pone.0149925.ref021]\\]. At the same time, when it does come to registering children, historical rates have been quite low, with issues similar to many other developing areas.\n\nFrom the founding of the modern country in 1904, to the mid 1980s, registration in Kenya hovered between 30--49% of the population \\[[@pone.0149925.ref022]\\]. As the country has moved toward industrialization, there have been attempts to increase registration and issuing of birth certificates. This has included a mix of infrastructural implementations, including educational programs targeting parent awareness, as well as establishment of devoted registration services, and decentralization to the community level \\[[@pone.0149925.ref013], [@pone.0149925.ref023], [@pone.0149925.ref024]\\]. Most notably for the present study, in 2010 the Ministry of Education introduced a requirement that all primary school children must have a birth certificate upon seeking admission to public and private schools, or, especially for those students already enrolled in classes, before registering for national high school entrance examination \\[[@pone.0149925.ref025]\\].\n\nRecent totals have shown positive improvement, with roughly 60 percent of children under five now having registered births (2008/09 Kenya Demographic and Health Survey \\[[@pone.0149925.ref026]\\]). Especially the tying of registration to entry into formalized schooling has shown recent anecdotal impact \\[[@pone.0149925.ref013]\\]. However, totals are still far below the \"universal\" target set by the government. Even more, only 24 percent of children had birth certificates \\[[@pone.0149925.ref026]\\]. This is often attributed to the continuing existence of a large percentage of the population outside of major urban centers \\[[@pone.0149925.ref027]\\], contributing to a widening gap, \\~10--20 percentage points lower than more urban regions \\[[@pone.0149925.ref026], [@pone.0149925.ref028]\\]. The pairing of low certification with demands by government for school documentation has also had a secondary impact on children in these areas, delaying or precluding some from continuing studies or from securing employment \\[[@pone.0149925.ref013]\\].\n\nIn 2010, the government therefore created a new program for Universal Birth Registration (UBR), targeted to the Millennium 2030 program. In this project, the Civil Registration Department partnered with Plan (<http://plan-international.org/>), one of the world's largest child centered community development organizations, to undergo an initiative aimed at increasing registration through enhancing accessibility, efficiency and community awareness. This also had the goal of targeting previously under-registered areas by surveying stakeholders with the goal of identifying what factors had impeded registration or certification of births.\n\nKwale, Kenya: findings from a representative study of parents {#sec003}\n-------------------------------------------------------------\n\nThe above program served as the basis for our previous study \\[[@pone.0149925.ref013]\\], and will also set the context for the present assessment. In collaboration with the Kenyan government and Plan Kenya we conducted a representative sampling (between November 2011 and March 2012) of the adult population at six test sites spread across a region of Kenya specifically selected by the government and Plan as typifying the previous issues with UBR. The designated region, Kwale County, lies in the Southeastern coastal portion of Kenya on the Indian Ocean, bordering Tanzania to the South, and near to the coastal city of Mombasa to the North. The area is a center for small scale farming with a population of roughly 650,000 people (KNBS, 2010), divided into six wards ([Fig 1](#pone.0149925.g001){ref-type=\"fig\"}) with a mix of tribal and ethnic groups typical for many regions in Kenya \\[[@pone.0149925.ref029]\\]. The area was selected by the government and Plan Kenya because it also has one of the lowest registration rates in the country (37.3% in 2012) \\[[@pone.0149925.ref026]\\], and was thus argued to provide an important basis for planning future policy.\n\n![Kwale county Kenya.\\\nSchool testing sites for present study denoted by white circles; towns used in previous testing of adult population \\[[@pone.0149925.ref013]\\] denoted by black circles.](pone.0149925.g001){#pone.0149925.g001}\n\nThis survey revealed birth registration levels even lower than government estimates for the same Coast Province region \\[[@pone.0149925.ref026]\\], with 46.5% of parents claiming to have registered all of their children and 7.9% reporting that all children had birth certificates. In addition, we found a number of infrastructural or socioeconomic factors which had significant impact on registration. These included parent education, occupation, rural vs. urban environment, and understanding of the registration process itself. This was coupled with other factors---cost, travel, lack of awareness---mentioned by many parents as reasons why they had not registered, but which, upon further investigation, did not appear to have a significant impact on actual registration action. In fact, as noted above, a key finding was that parents did appear to have sufficient information, resources, and access to successfully register children. Rather, we argued that the most persuasive impediment appeared to be a conscious weighing by parents of perceived benefits. This was coupled with a decision not to take immediate action with registering young children because they did not perceive reasons sufficient to outweigh the small costs. This outcome was then in turn also tied to school registration, which was noted by many parents to be a main impetus for finally taking action when children were close to entering secondary schooling. However, this led to a gap from infancy to junior high where children were not registered.\n\nWe concluded that new approaches were needed to motivate, or to modify, behavior of parents. For example, we suggested tying registration to other programs such as vaccination, which would provide an immediate reward to parents. Even better, we advocated avoiding the need for parent decisions via implementation of hospital birth or other top-down initiatives, whereby representatives (such as hospital staff or government) took responsibility of registration from parents, and with the argument that, by better understanding parent psychology rather than merely focusing on limitations, we may find meaningful improvement.\n\nStudy Design and Methodology {#sec004}\n============================\n\nThe present study builds from the above findings, focusing on the children of the previously surveyed parents. The study was administered between November 2011 and March 2012 to a convenience sample of all students in three grades (six-eight) in six schools within the same Kwale region ([Fig 1](#pone.0149925.g001){ref-type=\"fig\"}). Schools were selected from the target region with the goal of assessing those areas previously surveyed in the adult population, while also achieving geographic distribution across all six wards of Kwale. Grades six to eight (roughly ages 12--16) were selected in order to target students who might be more aware of the activities and issues in their community, as well as those who would be specifically preparing to take the entrance examinations for high school following grade eight. This age (11+) has also been shown to align with individuals who have moved beyond earlier cognitive development stages, which can impact understanding or responding to survey questions \\[[@pone.0149925.ref017], [@pone.0149925.ref030]\\].\n\nTo administer the survey, an introduction letter was sent to the schools through the respective District Education Offices to inform head teachers of our project and to request support and permission. A consent letter was also sent to parents to get their permission to survey their children. Before taking the survey, students were informed regarding the study purpose, and their written consent was obtained (see Ethics for more details). Following consent, surveys were conducted with each student by a trained surveyor. These individuals were selected from among volunteers working within each community and instructed by the study authors in data collection. Surveyors visited each school on an appointed day and met with students, working in small groups, with the surveyor reading the questions and students independently writing answers. Questionnaires were originally written in English, however survey questions were also offered in Kiswahili, and students were allowed to answer in the language they found most comfortable. Students were also checked with each question regarding their understanding, and were given unlimited time to make responses. These procedures, as well as the avoidance of negatively formulated questions or overly large number of possible choices, followed suggested best practice for surveys of children \\[[@pone.0149925.ref015], [@pone.0149925.ref017], [@pone.0149925.ref031]\\]. Surveys took about twenty minutes.\n\nSurvey questions and Goals {#sec005}\n--------------------------\n\nThe survey consisted of a paper-based list of questions divided into four main sections: (1) first this included a number of questions eliciting basic demographic information, including age of students and grade level. This was followed by (2) a section assessing students' possession of a birth certificate as well as their awareness of the time and conditions whereby it was received. (Note that we targeted possession of an actual certificate rather than the more general question of birth registration because of its importance for future school enrollment). This also included a question relating to awareness of the possession of certificates among siblings and peers, followed by (3) a section assessing student knowledge of purpose and procedures for birth registration/birth certificates. Based on these questions we were interested in assessing what were the basic rates of registration/certification among students, especially those preparing for the high school entrance examination, and if they differed materially from our adult data. We were also interested in students' awareness of the purpose and procedures for registration in order to determine whether any notable trends could be discovered among answers and possession of a certificate. Finally, (4) we included two questions eliciting students' own ideas regarding what government or others could do to increase rates of registration/certification in the students' region as well as how we might motivate their parents.\n\nEthics Statement {#sec006}\n----------------\n\nAll studies were approved by the Northeastern University Institutional Review Board and Plan Kenya Nairobi before being conducted. As noted above, the studies made use of an informed consent procedure for parents, school officials, and the individual students. All parties were made aware of the study purpose, procedures, as well as the intended use of the data. Both students and parents were also informed that the answers would be kept anonymous, and that the study was completely voluntary, could be stopped at any time, and that participation would have no bearing on class grading or other consequences. Written consent was obtained from all parties before beginning.\n\nStatistical analysis {#sec007}\n--------------------\n\nStatistical analyses were carried out with SPSS (IBM corp., version 22.0). Categorical data were summarized using frequency counts and percentages. Continuous data was assessed using means.\n\nComparison between survey sites (individual schools) showed no significant differences in demographic or grade level distributions, thus all sites were combined in the following analyses. Because the individual schools can be considered to represent different communities, however, which may have meaningful differences in their responses to birth certification, we have also provided between-school comparisons. These were analyzed via univariate Chi square tests. Results are reported in the accompanying tables. Factors affecting possession of a birth certificate were also first analysed by univariate Chi square tests, followed by a binomial logistic regression model, with possession of a certificate as the dependent variable and all factors that had been significant in the univariate analyses included as predictors. In line with the exploratory nature of this study, significance level was set to 5% (two sided) and we performed neither adjustment for multiple testing nor imputation of missing values. We also stress again that, given the convenience sampling technique used for the study, results should primarily be considered as articulating factors impacting this specific group of children. For broader inference-making regarding specific totals, we direct readers to our previous paper \\[[@pone.0149925.ref013]\\].\n\nResults {#sec008}\n=======\n\nAll students (and their parents) within the contacted classes agreed to participate, resulting in a final dataset of 503 (239 female; 264 male) out of an estimated 15,780 total households within the region (\\~3.1%) \\[[@pone.0149925.ref013]\\]. Mean age of students was 14.6 years (*SD* = 1.9). Students were also evenly distributed among the three grade levels: grade six (31.6%, *n* = 159), seven (33.8%, *n* = 170), and eight (34.6%, *n* = 174). The majority of participants were Muslim (58.1%), followed by Protestant (31.6%) and Catholic (10.3%).\n\nRates regarding possession of a birth certificate are shown in [Table 1](#pone.0149925.t001){ref-type=\"table\"}. Overall, 42.9% of students said that they possessed a certificate, while 6% did not know the answer. When students who had positively answered the above question were asked when they had received the certificate, half of the respondents (50.5%) clustered in the most recent three year period of 2012 to 2010---or when children were on average 11.6 to 13.6 years of age---while 15.7% received the certificate before 2009 and the remainder did not know. Of those children with siblings who were less than 18 years old (86.1%, *N* = 433, Mean number of siblings = 3.5, *SD* = 1.9), 33.9% were positive that their siblings all had certificates, while 4.4% were not sure of the answer.\n\n10.1371/journal.pone.0149925.t001\n\n###### Possession of birth certificate among students, Kwale Kenya (% answer distribution).\n\n![](pone.0149925.t001){#pone.0149925.t001g}\n\n                                                                                         Individual schools[^a^](#t001fn002){ref-type=\"table-fn\"}                                      \n  ----------------------------------------------------------------------------- -------- ---------------------------------------------------------- ------ ------ ------ ------ ------ ----------------------------\n  **Have birth certificate?**                                                                                                                                                          \n  Yes                                                                           42.90%   49                                                         29     74.7   37.6   50.7   18.7   67.9 (10, *N* = 503)\\*\\*\\*\n  No                                                                            51.1     41                                                         63.4   24     55.3   45.3   77.3   \n  Don\\'t know                                                                   6        10                                                         7.5    1.3    7.1    4      4      \n  **When did you get certificate?** [^c^](#t001fn004){ref-type=\"table-fn\"}                                                                                                             \n  2012                                                                          19       22.4                                                       11.1   17.9   28.1   7.9    35.7   39.1 (25, *N* = 216)\\*\n  2011                                                                          22.2     34.7                                                       3.7    21.4   21.9   18.4   28.6   \n  2010                                                                          9.3      10.2                                                       7.4    7.1    3.1    18.4   7.1    \n  2009                                                                          2.3      2                                                          0      5.4    0      2.6    0      \n  Before 2009                                                                   15.7     12.2                                                       29.6   21.4   9.4    10.5   7.1    \n  Don\\'t know                                                                   31.5     18.4                                                       48.1   26.8   37.5   42.1   21.4   \n  **Siblings have birth certificate?** [^d^](#t001fn005){ref-type=\"table-fn\"}                                                                                                          \n  Yes                                                                           33.9     33.7                                                       29.1   66.2   10     49.2   17.5   67.7 (10, *N* = 433)\\*\\*\\*\n  No                                                                            61.7     60.2                                                       65.1   30.9   85.7   50.8   76.2   \n  Don\\'t know                                                                   4.4      6                                                          5.8    2.9    4.3    0      6.3    \n\n*Notes*: All respondents were elementary school age children in grades six to eight. Percentages account for individuals who did not answer specific questions. Answer rate for all questions was \\> 97%.\n\n^a^ Comparison between survey sites (individual schools) showed no significant differences in demographic or grade level distributions. Provided for information purposes only.\n\n^b^ \\* and \\*\\*\\* indicate statistical significance at the .05 and .001 levels, respectively. Comparisons assessed via Chi square. First number in parentheses refers to degrees of freedom. Second number indicates total sample size.\n\n^c^ Question addressed to only students who answered 'yes' to having a certificate (*N* = 216).\n\n^d^ Question addressed to only students who answered 'yes' to having a sibling (*N* = 433). Average number of siblings = 3.5 (*SD* = 1.9).\n\nKnowledge of certificate importance and procedures {#sec009}\n--------------------------------------------------\n\nAs shown in [Table 2](#pone.0149925.t002){ref-type=\"table\"}, when asked to explain the definition of a birth certificate, the majority (61.5%) reported that a birth certificate is a document showing vital statistics or birth information of a child. This was followed by 17.7% who said that it was a document used for school exam registration, and then by a number of other answers (a document used in the \"future,\" identification (ID) or citizenship, healthcare) all \\< 10%. When asked whether having a certificate was necessary, 93.2% said 'yes'. When asked why one needed a certificate, most (31.4%) said that it was required by law, followed by necessity for being recognized as a citizen, and that it was needed for school and exam registration. In response to how soon a child should get a certificate, answers were spread rather evenly among 0--4 months (the correct answer via current Kenyan regulations), before one year of age, and after one year.\n\n10.1371/journal.pone.0149925.t002\n\n###### Understanding of birth certificate purpose and procedures among students.\n\nKwale, Kenya.\n\n![](pone.0149925.t002){#pone.0149925.t002g}\n\n                                                                                       individual schools [^a^](#t002fn002){ref-type=\"table-fn\"}                                                           \n  --------------------------------------------------------------------------- -------- ----------------------------------------------------------- ------ ------ ------ ------ --------------------------- -----------------------------\n  **What is a birth certificate?** [^c^](#t002fn004){ref-type=\"table-fn\"}                                                                                                                                  \n  Doc. with vital/birth statistics                                            61.50%   51.0                                                        67.7   68.0   42.4   77.3   64.0                        126.0 (30, *N* = 503)\\*\\*\\*\n  Doc. for school registration                                                17.7     17.0                                                        16.0   17.3   23.5   13.3   16.0                        \n  Doc. used in future                                                         6.3      7.0                                                         1.1    1.3    22.4   4.0    0                           \n  Identification (ID)/citizenship                                             5.1      14.0                                                        1.1    10.7   3.5    5.3    2.7                         \n  Used for healthcare                                                         1.2      5.0                                                         0      0      1.2    0      0                           \n  Other                                                                       3.8      5.0                                                         7.4    2.7    7.1    0      13.3                        \n  Don\\'t know                                                                 4.2      1.0                                                         6.4    0      0      0      4.0                         \n  **Why do you need a cert.?**                                                                                                                                                                             \n  Required by law                                                             31.4     23.2                                                        49.5   25.3   49.4   12.0   26.7                        116.9 (20, N = 500)\\*\\*\\*\n  To be recognized as citizen                                                 27.2     32.3                                                        30.8   32.0   18.8   21.3   28.0                        \n  School examination                                                          21.7     19.2                                                        16.5   38.7   18.8   17.3   22.7                        \n  Identification                                                              17.7     23.2                                                        2.2    1.3    11.8   48.0   22.7                        \n  Don\\'t know                                                                 0.8      2.0                                                         1.1    2.7    1.2    1.3    0                           \n  **How soon should you get cert.?**                                                                                                                                                                       \n  0--4 months                                                                 31.1     36.0                                                        25.8   18.7   42.4   40.5   21.6                        28.9 (15, *N* = 501)\\*\n  Before one year old                                                         29.1     25.0                                                        37.6   34.7   29.4   21.6   25.7                        \n  After one year                                                              29.7     29.0                                                        24.7   37.3   22.4   31.1   36.5                        \n  Don\\'t know                                                                 10.0     10.0                                                        11.8   9.3    5.9    6.8    16.2                        \n  **Where can you get certificate?**                                                                                                                                                                       \n  Chief/Assistant chief                                                       43.3     51.0                                                        26.9   38.7   40.0   37.3   68.0                        107.4 (25, *N* = 503)\\*\\*\\*\n  Community health worker                                                     19.5     19.0                                                        21.5   17.3   18.8   25.3   14.7                        \n  Nurses/doctors                                                              14.7     17.0                                                        31.2   4.0    16.5   9.3    5.3                         \n  Birth attendants                                                            11.1     4.0                                                         4.3    26.7   8.2    25.3   2.7                         \n  Volunteer organizations                                                     7.0      5.0                                                         9.7    10.7   10.6   1.3    4.0                         \n  Don\\'t know                                                                 3.4      4.0                                                         6.5    2.7    5.9    1.3    5.3                         \n  **Who responsible for getting cert.?**                                                                                                                                                                   \n  Parents/relatives                                                           31.4     23.0                                                        12.9   26.7   12.9   62.7   60.0                        156.3 (35, *N* = 503)\\*\\*\\*\n  Hospital                                                                    19.3     18.0                                                        21.5   18.7   23.5   12.0   21.3                        \n  Church/mosque                                                               16.3     17.0                                                        30.1   5.3    29.4   6.7    4.0                         \n  Government/chief                                                            10.6     15.0                                                        7.5    14.7   8.2    8.0    9.3                         \n  Midwife                                                                     10.3     10.0                                                        15.1   12.0   14.1   5.3    4.0                         \n  Community organization                                                      5.0      5.0                                                         8.6    16.0   7.1    0      0                           \n  School                                                                      3.4      5.0                                                         4.3    4.0    3.5    1.3    1.3                         \n  Don\\'t know                                                                 1.4      7.0                                                         0      2.7    1.2    4.0    0                           \n  **Where did you get information?** [^d^](#t002fn005){ref-type=\"table-fn\"}                                                                                                                                \n  Chief                                                                       32.0     39.8                                                        10.2   55.4   12.0   33.1   26.6                        151.7 (40, *N* = 610)\\*\\*\\*\n  School                                                                      26.4     17.7                                                        26.6   1.4    33.7   15.3   45.6                        \n  Hospital                                                                    23.9     18.6                                                        29.7   29.7   16.3   20.2   10.1                        \n  Family                                                                      16.3     12.4                                                        12.5   13.5   16.3   20.2   7.6                         \n  Community                                                                   5.2      0.9                                                         9.4    0      10.9   2.4    2.5                         \n  Siblings\\' birth                                                            3.9      2.7                                                         5.5    0      4.3    4.0    5.1                         \n  Church/Mosque                                                               2.6      4.4                                                         2.3    0      3.3    1.6    0                           \n  Midwife                                                                     2.4      1.8                                                         3.1    0      2.2    3.2    1.3                         \n  Don\\'t know                                                                 0.9      1.8                                                         0.8    0      1.1    0      1.3                         \n  **Know other children with cert.?**                                                                                                                                          69.2 (5, *N* = 503)\\*\\*\\*   \n  Yes                                                                         61.6     64.0                                                        47.3   61.3   34.1   85.3   84.0                        \n  No                                                                          38.3     36.0                                                        52.7   38.7   65.9   14.7   16.0                        \n\n*Notes*: All respondents were elementary school children in grades six to eight. Percentages account for individuals who did not answer specific questions. Answer rate for all questions was \\> 97%.\n\n^a^ Comparison between survey sites (individual schools) showed no significant differences in demographic or grade level distributions. Provided for information purposes only.\n\n^b^ \\* and \\*\\*\\* indicate statistical significance at the .05 and .001 levels, respectively. Comparisons assessed via Chi square. First number in parentheses refers to degrees of freedom. Second number indicates total sample size.\n\n^c^ Notable answers in 'Other' category included: general physical description, general notion of \\\"importance,\\\" or use for securing employment (all \\< 1%).\n\n^d^ Respondent could give more than one answer.\n\nWhen asked where one could file for or receive a certificate, the most noted response was chiefs/assistant chiefs---the community level representatives of local government. This was followed by community health workers, nurses/doctors, or hospital birth attendants. Regarding the individual who should take responsibility to ensure that a child gets a birth certificate, 31.4% noted parents or relatives, 19.3% said hospitals, followed by churches/mosques, or some portion of local/national government. Regarding where they had received their information, the highest number mentioned chiefs (32%), followed by schools, hospitals, and relatives or parents. Over half (61.6%) knew other children who had certificates.\n\nFactors significantly associated with possession of birth certificate {#sec010}\n---------------------------------------------------------------------\n\nWe then looked at the question of which of the above factors had significant ties to a student's possession of their own birth certificate. Results of the Chi square analyses, are shown on the left side of [Table 3](#pone.0149925.t003){ref-type=\"table\"}. Among the population of respondents who had answered either 'yes' or 'no' to having a certificate (n = 473), differences were found between grade levels, with the highest rates of certification found among students in class eight. Differences were also found between religions, with Muslim students showing lower rates, as well as between those who reported that their siblings did/did not have their own certificates. Notably, we found 70.3% certificate possession for those whose siblings also had certificates versus 33.6% for those whose siblings did not.\n\n10.1371/journal.pone.0149925.t003\n\n###### Knowledge and demographic factors significantly tied to possession of birth certificate among students in Kwale, Kenya.\n\nUnivariate analyses (left) and Binomial Logistic Regression (right).\n\n![](pone.0149925.t003){#pone.0149925.t003g}\n\n                                        \\% have certificate   Univariate Analyses         Binomial Logistic Regression [^a^](#t003fn002){ref-type=\"table-fn\"}          \n  ------------------------------------- --------------------- --------------------------- --------------------------------------------------------------------- ------ ----------------\n  **Grade level**                                             36.2 (2, *N* = 473)\\*\\*\\*   40.7\\*\\*\\*                                                                   \n  eight                                 63.9                                              \\-\\-\\--                                                               1      \\-\\-\\--\n  seven                                 32.9                                              36.9\\*\\*\\*                                                            0.17   \\[0.10--0.30\\]\n  six                                   38.6                                              22.8\\*\\*\\*                                                            0.24   \\[0.13--0.43\\]\n  **School**                                                  58.9 (5, *N* = 473)\\*\\*\\*   29.3\\*\\*\\*                                                                   \n  Kinango                               75.7                                              \\-\\-\\--                                                               1      \\-\\-\\--\n  Lukore                                54.4                                              4.6\\*                                                                 0.41   \\[0.18--0.93\\]\n  NdoHivyo                              52.8                                              5.1\\*                                                                 0.38   \\[0.16--0.88\\]\n  Mwambalazi                            40.5                                              4.3\\*                                                                 0.36   \\[0.14--0.94\\]\n  Kikoneni                              31.4                                              8.3\\*\\*                                                               0.26   \\[0.11--0.65\\]\n  KwaKadogo                             19.4                                              27.6\\*\\*\\*                                                            0.08   \\[0.03--0.21\\]\n  **Sibling has certificate?**                                56.5 (1, *N* = 473)\\*\\*\\*   38.0\\*\\*\\*                                                                   \n  Yes                                   70.3                                              \\-\\-\\--                                                               1      \\-\\-\\--\n  No                                    33.6                                              \\-\\-\\--                                                               0.2    \\[0.12--0.34\\]\n  **Religion**                                                9.8 (2, *N* = 473)\\*\\*      4.3                                                                          \n  Catholic                              54.2                                                                                                                           \n  Protestant                            53.9                                                                                                                           \n  Muslim                                39.5                                                                                                                           \n  **Know other children with cert.?**                         5.0 (1, *N* = 473)\\*        1.9                                                                          \n  Yes                                   49.7                                                                                                                           \n  No                                    39.1                                                                                                                           \n  **Definition of cert.**                                     16.7 (6, *N* = 473)\\*       4.4                                                                          \n  Doc. of ID/citizenship                83.3                                                                                                                           \n  Doc. used for healthcare              50.0                                                                                                                           \n  Doc. with vital/birth statistics      45.4                                                                                                                           \n  Doc. used in \\\"future\\\"               46.7                                                                                                                           \n  Other (general card or important)     38.9                                                                                                                           \n  Doc. for school/exam registration     36.9                                                                                                                           \n  Don\\'t know                           45.0                                                                                                                           \n  **Source of information**                                   14.0 (7, *N* = 473)\\*       5.1                                                                          \n  Relatives                             56.0                                                                                                                           \n  Church                                55.6                                                                                                                           \n  Chief                                 51.7                                                                                                                           \n  Hospital                              48.5                                                                                                                           \n  Midwife/ sibling birth                40.0                                                                                                                           \n  School                                37.0                                                                                                                           \n  Community                             34.1                                                                                                                           \n  Don\\'t know                           23.5                                                                                                                           \n\n*Notes*: All respondents were elementary school students in grades six to eight.\n\n^a^ Regression model (right side) performed on all factors shown in left column which had shown significance in individual univariate analyses. All factors treated as categorical. Model significant at *p* \\< .001. \u03c72 (24, *N* = 473) = 165.97, Nagelkerke *R*2 = .40, correctly predicted cases = 74.2%.\n\n^b^ \\*\\*\\* significant at p \\< .001, \\*\\* p \\< .01, \\* p \\< .05.\n\n^c^ Odds ratios and CIs are shown for significant factors only, and show comparison to topmost category in group.\n\nOdds ratios should not be interpreted as approximated relative risk. Estimated odds as shown in the Table will be closer to 1 than the ratio change of all odds (which cannot reliably be estimated via this approach)\n\nAmong knowledge and awareness questions, significant difference was also found regarding definition of a birth certificate. Notably, students who answered that certificates were primarily a \"document used for ID (government identification) or citizenship\" possessed certificates at the highest rate (83.3%), whereas 36.9% (lowest) of those who answered that certificates were primarily \"documents used for exam or school registration\" themselves had certificates. We also found significant difference regarding primary source of information. Here, the prominent difference appeared to be use of schools or community, with students using these avenues reporting low rates of certificate possession, compared to higher rates certificate possession among students mentioning hospitals, chiefs, churches and relatives/parents. Those who did not know any source in turn showed the lowest rates of certificate possession (23.5%). Follow-up comparison of source of information and ability of students to give correct information about certificate definition/procedures, did not reach significance. Significant difference was also found for those who knew other children with certificates versus those who did not.\n\nWe also found significant difference between schools. However again, univariate Chi square comparisons showed no demographic differences that would explain these patterns. Comparison for knowledge and understanding questions between schools did show significant differences. These are reported in the right sides of Tables [1](#pone.0149925.t001){ref-type=\"table\"} and [2](#pone.0149925.t002){ref-type=\"table\"}. Here as well, patterns were not overly instructive. For example, regarding purpose of a certificate, students from Kinango---with the highest percentage of students possessing certificates---answered that certificates are used to register for school at a relatively higher rate than many other schools. Whereas students in schools with lower percentages of students possessing certificates (e.g., Kwakadogo and Mwambalazi) were more likely to say that a certificate is used for ID. These patterns did not hold for other schools, while essentially the same patterns were found for all schools for other answers, regardless of percentage of students possessing certificates. Significant difference between schools was also found for knowing others who had a certificate. Here again however, Kwakadogo, for example, with the lowest percentage of students with certificates, showed an answer of 'yes' to knowing others at the second highest rate (84%), while Kinango (highest percentage of students with certificates) showed relatively low rates of knowing others with certificates.\n\nThe one factor that did appear to show both a significant and a *meaningful* pattern of differences was again source of information. Students from schools with the lowest percentages of individuals possessing certificates mentioned schools as their primary information source---Kwakadogo (45.6% of time), Mwambalazi (33.7%). This compared to lower mention of schools as an information source among students at schools with higher percentage of students possessing certificates (e.g., Kinango, 1.4%; Ndohivyo, 15.3%). On the other hand, mention of chiefs was lower among schools whose students showed lowest rates of possessing certificates---Kikoneni, Mwambalazi---compared to schools with higher percentages of students in possession of certificates (see [Table 2](#pone.0149925.t002){ref-type=\"table\"}).\n\nFinally, to better understand the combined contribution of the above factors, we conducted a binary logistic regression. This again used all of the factors (listed on the left side of [Table 3](#pone.0149925.t003){ref-type=\"table\"}) which had shown significance in regards to possession of birth certificates as predictor variables (treated as categorical data). The model results are reported on the right side of [Table 3](#pone.0149925.t003){ref-type=\"table\"}, with Odds Ratios and 95% Confidence Intervals also shown for factors that were found to be significant in the model (indicated by asterisks accompanying the Wald statistics). The decision was made to use a logistic regression and Odds Ratios (ORs), rather than prevalence rates or relative risk (RR), following suggestion \\[[@pone.0149925.ref032], [@pone.0149925.ref033]\\] that in non-controlled sampling cases where there is high outcome prevalence (above 10%), ORs are a more appropriate summary measure with a more tenable assumption of homogeneity across all individuals in a sample population. Note therefore that the estimates should not be interpreted as approximated risk.\n\nThe model (*p* \\< .001, *\u03c7*^*2*^ \\[24, *N* = 473\\] = 165.97, Nagelkerke *R*^2^ = 40%, correctly predicted cases = 74.2%) indicated that only three factors---grade level, school, and whether or not one's sibling had a certificate---remained significant predictors. Odds ratios indicated that students in grade eight had significantly higher odds of possessing a birth certificate when compared to lower grade levels, while having a sibling with a certificate was associated with higher likelihood of the student him/herself possessing a certificate.\n\nStudents' suggestions for improving certification of their own births {#sec011}\n---------------------------------------------------------------------\n\nStudents' own ideas for improving certification rates are reported in [Table 4](#pone.0149925.t004){ref-type=\"table\"}. Regarding how the government could improve registration, student answers---which were given in a free style with no prearranged answer choices---showed a rather even distribution between the themes of (1) direct action, (2) infrastructure and (3) education. The most often mentioned suggestion (16.7% of students) was that the government should in some way take up the full responsibility from parents and therefore complete the process automatically while minimizing parent involvement. This typically involved the suggestion that the government dispatch some official representative (chief or other National level official) to visit children or parents in order to complete the registration process. Following this, 15.2% said that the government should provide more education or information for parents, and 12.7% said that the government should do more to enforce the existing laws or make official checks at schools or villages to ensure that all children were registered; 11.6% also noted that the government should give *non financial* assistance through helping with transportation or paperwork.\n\n10.1371/journal.pone.0149925.t004\n\n###### Students' suggestions for how to improve birth registration and tie to possession of own certificate, Kwale, Kenya.\n\n![](pone.0149925.t004){#pone.0149925.t004g}\n\n                                                                     \\% total respondents [^a^](#t004fn002){ref-type=\"table-fn\"}   \\% who possess certificate   *X*^*2*^ Answer x cert. possession\n  ------------------------------------------------------------------ ------------------------------------------------------------- ---------------------------- ------------------------------------\n  **What should government do to raise birth registration?**                                                                                                    \n  Register automatically instead of parents.                         16.7%                                                         36.7                         21.2 (8, *N* = 473)\\*\\*\n  Improve education about registration                               15.2                                                          50.0                         \n  Enforce the law.                                                   12.7                                                          62.7                         \n  Assist parents (non financial).                                    11.6                                                          38.2                         \n  Increase hospital births.                                          8.7                                                           65.9                         \n  Improve access or ease of registration.                            7.8                                                           43.2                         \n  Reduce cost.                                                       5.3                                                           32.0                         \n  Don\\'t know.                                                       7.4                                                           34.3                         \n  **How would you encourage your parents to register your birth?**                                                                                              \n  Explain importance for school/exams.                               32.1                                                          40.1                         18.1 (8, *N* = 473)\\*\n  Explain importance (general).                                      20.3                                                          58.3                         \n  Ask third party to talk to them.                                   9.7                                                           56.5                         \n  Explain importance for ID/citizenship.                             8.7                                                           29.3                         \n  Take action: do the registration for them.                         7.8                                                           51.4                         \n  Explain importance for getting job.                                4.9                                                           30.4                         \n  Explain importance for following law.                              1.7                                                           37.5                         \n  Don\\'t know.                                                       9.5                                                           42.2                         \n\n*Notes*: All respondents were elementary school students in grades six to eight.\n\n^a^ Percentages account for individuals who did not answer specific questions. Answer rate for all questions was \\> 97%.\n\n^b^ \\* and \\*\\* indicate statistical significance at the .05 and .01 levels, respectively.\n\nComparisons assessed via Chi square. First number in parentheses refers to degrees of freedom. Second number indicates total sample size.\n\nOther notable ideas were that the government should make moves to increase birth in hospitals/health centers, either through encouragement or by building more infrastructure (8.7%). Looking across responses, a common theme was also the need to connect services directly to the community level. This idea was also commonly tied with the hiring, funding and dispatch of representatives charged with the task of visiting the community, school or household, and in order to check, register or educate every member of the community area. Answer to this question also showed significant connection to possession of a birth certificate. Chi square analysis ([Table 4](#pone.0149925.t004){ref-type=\"table\"}) showed that those who mentioned importance of enforcement had certificates 62.7% of the time. This was also true for the strategy of increasing hospital births (65.9% having certificates). On the other hand, those who mentioned the need for the government to take responsibility, or to reduce costs, were less likely to have a certificate (36.7 and 32%, respectively).\n\nFinally, when asked how they themselves could encourage their parents to register or get certificates ([Table 4](#pone.0149925.t004){ref-type=\"table\"}), the majority of students mentioned that they would communicate some aspect of its importance as an impetus for parent action. Among such answers, 32.1% explicitly said that they would stress need of registration when taking school examinations, while others said they would stress importance for obtaining an ID or citizenship, future employment, or legal necessity. On the other hand, 9.7% said that they would arrange for their parents to talk with a third party (chief, head teacher), and 7.8% said that they would themselves act to pursue registration or directly ask their parents to accompany them. Chi square analysis of this question showed significant relation to possession of a certificate. Those who gave answers stressing general importance, asking for third party communication or taking action themselves had highest rates, while those who stressed tie to jobs or school were lower.\n\nDiscussion {#sec012}\n==========\n\nThis study considered students' awareness of the purpose and procedures for certification of their own births. This was pursued because these children represent the front line of the ongoing under-registration of children within developing regions, and will themselves soon be required to have birth certificates to continue schooling. Thus, it was argued that investigating what they know about registration, as well as awareness of their own rights, protections or empowerment, may afford an important addition to the understanding of the decisions of their community and parents. Our findings do provide important evidence that may help to further articulate this topic.\n\nFirst, we actually found higher rates of certificate possession among students in this study, with just under half (43%) having birth certificates. While again care should be taken when considering these totals, this compares to only 8% of households in our previous study reporting all children with certificates, and 24.9% found in the 2008/09 Demographic and Health Survey \\[[@pone.0149925.ref026]\\] for the Coast Province region, of which Kwale is a part. The reason for this difference appeared to be the older age of this population, consisting of the three grades directly preceding the high school entrance examination. The students also showed low incidence of \"Don't know\" responses, which can---in cases of high incidence, and in addition to simply not knowing---also indicate problems in survey methodology or question understanding, especially with children \\[[@pone.0149925.ref017]\\].\n\nWhen we did look to the breakdown between grade levels there was significant difference in rates of students having certificates, with the highest totals among those in grade eight directly preceding examination. Students also showed much higher rates in the three years (2010--2012) before the study and falling off to almost no certification before this period. This trend of late registration also could be seen when asking about students' younger siblings, who had certificates at a lower rate (34%).\n\nLooking to other structural factors that correlated to whether or not children had certificates, we again found essentially the same issues as were found in our previous study and which also articulated a school-centered interpretation. For example, there were significant differences between Muslim and Protestant/Catholic populations, with the former showing lower rates, as well as between schools. These differences between religion as well as regions also coincided with similar findings in the adult population. However when we look at the breakdown between factors, no noticeable trends were discovered for any factor that would point to why children did not have certificates.\n\nThe same argument can be made for differences between schools. Although significant, consideration of other contextual aspects did not indicate any one reason that could be attributed to different rates of registration. In turn, when we look to our previous adult dataset we see that in fact there is not a correspondence between the highest registered schools and relative rates of registration in their wider regions. Rather, this appeared to be mainly tied to parent personal belief in importance of registration itself. This is also supported by significant correspondence between the students' certification and the rates of certification of their younger siblings within the same household.\n\nAs discussed in our previous paper, this present situation whereby parents do not take action unless there is a pending need or positive benefit leads to the present issue witnessed in Kenya and many other developing regions of a significant time gap wherein children go without documentation. This same time period---in the present case roughly fourteen years---is also the ages whereby a child would most benefit from the health and social protections afforded by being recorded and identified by the government. Even more, despite need of birth certification for these children in order to continue their education, almost half still did not have a birth certificate by grade eight, foreshadowing the secondary issues endemic in developing regions from not having a record of children's birth.\n\nStudents' awareness of issues affecting their possession of certificates {#sec013}\n------------------------------------------------------------------------\n\nRegarding what children themselves know about registration or need of birth certificates, here the most noticeable finding may be the simple fact that children appeared highly informed regarding the need and procedures for securing a certificate. More than 90% could articulate a definition of a birth certificate, with the majority explaining that it contained vital and/or birth statistics.\n\nInterestingly, this number of correct respondents was actually higher among students than their parents, where only 44.5% could articulate the definition of a certificate. Even more, most students (\\~93%) could definitively answer whether or not they themselves had a certificate, and could also answer for their siblings. Over half of students could also give examples of other children who had certificates. For comparison, our survey of parents showed that 13% responded that they did not know whether their children's births had been registered, and 75% said that they did not know whether they had obtained a birth certificate. This difference might of course be interpreted as the product of a number of factors. It is possible that this reflects the, possibly unfounded, confidence of youth---note that we did not check whether they actually did have a birth certificate. Parents may also have had lower actual awareness than was found among children---a rather intriguing predicament. More plausibly, adults may have had higher embarrassment or unwillingness to give a candid answer, especially in negative registration cases \\[[@pone.0149925.ref017]\\], thus again pointing to the knowledge that might be gained from asking children about activity of their parents.\n\nThe majority (93%) also knew that certificates were necessary, while there also appeared to be preexisting official means of educating children about certificates, with children listing either chiefs---the local head of government---or schools and hospitals as an information source. Interestingly, this mention of health workers and local government as primary sources coincides with our finding among parents. This finding might also be interpreted to support the conclusion that only about 25% of students had received information from their parents or from their own experience of the births of their siblings, suggesting that there was not extensive parent-child communication regarding this topic.\n\nWhen looking to answers regarding why one needed a certificate, there also appeared to be a lack of clarity, with a roughly even breakdown of answers between the given choices of either its basic legal requirement, its need in order to be recognized as a citizen, or for exam registration. This suggests a slight difference in understanding or awareness from parents, where most were able to articulate that registration or certification were necessary in order for identification or citizenship. However, among parents we also did find essentially the same number claiming the basic importance for school enrollment. Children also did not seem to know when one should get a certificate with even distribution between the given choices, and did not appear to have a strong awareness of who was responsible, with only a third claiming their parents. On the other hand, children did appear more knowledgeable about the procedure for obtaining a certificate, with the majority listing chiefs/assistant chiefs---the correct answer in most cases. The remainder mostly mentioned nurses or health attendants who also are traditional means of registration, especially in hospital births.\n\nThe findings raise another interesting angle regarding children and their parents. While children appeared to be informed about the basic need and purpose of registration, they did not appear highly knowledgeable about the role that should be played by their parents in securing registration. Our findings regarding knowledge or awareness questions might be taken as an indicator of an information gap in communication between parents, community and children when considering actual choice to pursue certification. Again, children without certificates were more likely to say that the purpose of a certificate was for school exam registration and less likely to frame its importance around citizenship or identification than were those presently with certificates. This might be read to suggest that children and potentially parents who had not certified were only now becoming aware of registration or certificate importance as a result of school involvement. This is also supported by our finding of significant difference in source of information, with those not registered mentioning schools as their primary source of information. While those who were registered were much more likely to mention chiefs or central government representatives, who presumably might frame registration importance along children's rights, policy planning or citizenship. The same pattern was also found in our study of the adult population, in which a significant distribution was found regarding source of information and successful registration of all children, and where those parents who listed chiefs or health facilities registered at a much higher rate than those who listed schools as their primary source.\n\nThe above finding also leads to a handful of compelling questions. First, it may be that some communities had better established structures for communication or collaboration between citizens, government and children whereby they were made aware of the need for registration from the moment of birth. In addition, it is potentially important to discover whether the source of knowledge or awareness was the result of the actions of government representatives, parents or another third party. Alternatively, the finding of awareness of the chief as a source of information among children also raises the question of why these children became informed. That is, was this knowledge passed directly to children, or learned via their parents? An answer aligning especially to the former might suggest that having informed children, who can then better interact regarding this topic, may lead to quicker action by parents. In turn, these findings might also suggest that gains could be made from facilitating discussion between children and parents. Previous research shows that effective parent-child communication (for example regarding discussion of sexual issues) is associated with improved outcomes in children and adolescents \\[[@pone.0149925.ref016]\\], and might also work to motivate caregivers.\n\nWhat do students suggest? {#sec014}\n-------------------------\n\nThe results also give important insights regarding how children suggest we might increase registration and possession of birth certificates. Beginning with the specific question of what could be done to motivate their parents, children most often argued for better articulation, to parents, of registration's purpose or use. The factor that most children said they would mention was school registration, followed by identification, employment or basic importance. Interestingly, there also did appear to be a trend whereby children who were themselves not yet registered explicitly noted that they would stress the need of having a certificate for the purpose of citizenship or identification. This occurred at about double the rates of those who were not registered. At the same time, those who were already registered also noted more often that instead of communicating directly with parents they would put their parent in touch with an authority figure, chief or other government representative. These findings do suggest that children may have been quite aware of the underlying issues raised by our previous adult study and a prescient potential source for strategic planning.\n\nWhen asked to give strategy ideas that could be used by the government, students again seemed quite in tune with the motivations or potential actions of their parents. Many argued for direct action from the government, and essentially for authorities to take the responsibility from parents. This was often framed along the lines of a need to fund, hire and send staff directly to the village, school, or the door of the parent. On the other hand, the other most often mentioned means were for officials to visit each village and provide more information, education, direct paperwork assistance and for them to ensure that each child was registered. This finding matches the argument made in our previous paper, where we suggested that the best course of action was for direct government involvement and action in place of relying on the motivation of the parent.\n\nIt is also informative to note what answers *were not* often given. While many students mentioned the need for building registration centers in their local areas or otherwise connecting services more directly to parents, only 7.8% argued for bureaucratic or structural developments to make the process more easy, and only 5% mentioned cost. As we discussed in our previous paper, while prevalent wisdom for increasing certification frames policy along the lines of cost or access for parents, the present low rates of registration even following programs to alleviate these issues suggests that this does not have the desired impact. Rather, the best case may be structural change that removes deliberation from the parent.\n\nChildren seemed to share this intuition. Equally interesting was our finding of some students (8.2%), again with no prompting, who explicitly mentioned that the government should increase birth in hospitals or build more health centers to improve rates of registration. This suggestion, which might provide an automatic means of registering and certifying children as part of the official course of hospital admittance, was given as a major suggestion from our previous paper. Our finding in this paper of likelihood among those students who were registered to suggest either hospitalization or direct enforcement, is further evidence for the intuition of children.\n\nConclusion and Suggestions for Policy Planning and Research {#sec015}\n===========================================================\n\nWe will conclude with a short discussion of potential application of these findings to policy or health research, and some suggestions. Although not necessarily representative of all children, these results do raise important implications for policy and study in Kenya, East Africa, or in similar developing regions. First, the results do specifically lend support to our previous arguments for policy implementations. We again had argued that a policy that can automate procedures or reduce responsibility or deliberation from parents may mark the most impactful solution. This is because parents in previously under-registered regions, *as reiterated here by their children*, do not appear motivated by the relative ease, cost or awareness of registration services. Rather they appear to take action only when they perceive an immediate benefit. In our present survey the contextual factors for when and why children were or were not registered appeared to center on parents waiting to register until it was immediately needed for school registration, and highlighting the need for different approaches that may minimize this decision.\n\nIn turn, when we look to what these children themselves suggested as possible solutions, a notable portion also argued for coupling registration to community centered programs that have immediate perceived benefit to themselves or their parents. Children also noted that such a strategy should include either directly registering a child for the parents, appointing individuals to visit each and every household in a village, or increasing hospital births. All three factors, especially entering children into a hospital system, were also stressed in our previous paper and do appear to be intuited by students.\n\nStudents also appeared quite cognizant of the deliberation and lack of sufficient motivation by their parents. Our respondents routinely stressed that parents do not find benefit until school becomes an issue, do not register younger children. They also suggested educational programs which stress earlier benefits such as need for child health or rights/identity protection.\n\nCoupling children's' ideas to policy planning, one of the most obvious approaches would again be hospitalization for child birth, which was shown in our adult study to be the largest determinant of registration and to provide a structure allowing automatic registration services. Despite the advocacy of this approach also by a notable number of students, however, wide-scale hospitalization may not yet be practical in Kenya or other developing regions. Therefore we suggested again other structures that could provide similar automatic registration, such as child vaccination. As also noted in our last paper, this could also be accomplished through use of recent advances in ICT and mobile technology, which has shown promise in many similar areas \\[[@pone.0149925.ref008], [@pone.0149925.ref034]\\]. Interestingly, this structure for a representative to visit every home within a village in order to confirm or process registration was also an idea also specifically articulated by many children.\n\nSecond, when considering education services for informing parents, there also appears room for an important tweaking of present approaches. While it may seem intuitive to tie registration to schooling, because it would impact the most children even throughout rural or under-reported areas, this may be counterproductive. This is so because it appears to cause parents to wait until the last moment. As argued before, there may be more merit in reshaping education so that it does clearly articulate the immediate benefits, to children *and to parents*. This might even be done at the expense of downplaying latter educational or economic motives. This finding also appeared well intuited by children.\n\nFinally, findings suggest that taking heed of ideas of children, especially as it pertains to the motivations of their parents, may itself play an important role in policy planning. They may be more candid or realistic sources for uncovering the actual factors that motivate the decisions or behaviors of the adult population. It should of course be explicitly noted that we are not advocating a shift of the burden for policy direction from adults to kids. Nor are we suggesting that this population of children in Kenya can speak for other countries' children, who may of course have quite different circumstances.\n\nThe present study also does of course come with caveats. Children represent a challenging population for survey-based studies, especially regarding a tendency to not understand procedures or questions. See \\[[@pone.0149925.ref015], [@pone.0149925.ref031], [@pone.0149925.ref035]\\] for examples from Kenya. Although this study did use the utmost care in its procedures, following established practice for children, readers should be mindful when considering the findings. The study was also a convenience sample, albeit of a targeted population. Thus, care should be taken in making inferences regarding other children or other countries' populations. We again in fact suggest against using this paper's findings as de facto estimates of total rates---i.e., of birth certification. These are better addressed through our earlier adult study. However, especially in a case such as birth registration which does directly touch the lives of students, it might be argued that directly asking for their insights or opinions should at least be one aspect of research.\n\nThe authors would like to thank Plan Kenya for the opportunity to conduct this study as a component of their Universal Birth Registration initiative. This study was also made possible by technical support from Nokia Corporation and Plan Finland. The discussion, suggestions and opinions expressed in the paper, however, are the sole responsibility of the authors. The authors will make the data discussed in this study as well as the full survey results available upon written request to the corresponding authors.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: RGW JW HN JGO. Performed the experiments: RGW JW HN JGO. Analyzed the data: MP GOO BKN. Wrote the paper: MP RGW BKN JGO.\n"}
+{"text": "Introduction {#coy046s2}\n============\n\nThe measurement of biochemical and haematological parameters can serve as a valuable tool for evaluating and monitoring the health of wild reptile populations ([@coy046C45]; [@coy046C11]). However, a major obstacle to conducting wildlife health assessments is a lack of baseline data against which new data can be compared ([@coy046C10]). Without an understanding of typical species-specific (or taxon-specific) variation in biochemical and haematological parameters, researchers are unable to identify potential effects of disease, injury, pollutants or other changing environmental conditions ([@coy046C35]). This issue is particularly important for regions with large numbers of endemic species that are experiencing rapid change, such as the Galapagos Archipelago.\n\nThe Galapagos Archipelago has nine endemic species of lava lizard in the South American genus *Microlophus* (Family: Tropiduridae) ([@coy046C4]; [@coy046C32]). *Microlophus bivittatus* (Fig. [1](#coy046F1){ref-type=\"fig\"}), which is restricted to the island of San Crist\u00f3bal and nearby Islote Lobos, inhabits xeric, low-elevation areas, such as coastal scrubland and rocky beaches. This lizard is listed as near threatened ([@coy046C38]) and is particularly vulnerable to predation by feral cats ([@coy046C12]; [@coy046C13]). At present, no blood chemistry or haematology data have been acquired for the genus *Microlophus*, which complicates efforts to evaluate or monitor the health of wild lava lizard populations. In this study, we report baseline biochemical and haematological values, along with basic physiological measures, for the San Crist\u00f3bal lava lizard.\n\n![Female individual of a San Crist\u00f3bal Lava Lizard (*Microlophus bivittatus*).](coy046f01){#coy046F1}\n\nMethods {#coy046s3}\n=======\n\nEthics statement {#coy046s4}\n----------------\n\nThis study was conducted on San Crist\u00f3bal Island in the Gal\u00e1pagos archipelago of Ecuador as part of a population health assessment authorized by the Gal\u00e1pagos National Park Service (Permit No. PC-23-17, PI Carlos A. Valle, PhD) and approved by the Universidad San Francisco de Quito ethics and animal handling protocol. All handling and sampling procedures were consistent with standard vertebrate protocols and veterinary practices.\n\nCapture and marking {#coy046s5}\n-------------------\n\nLizards were captured in locations within 2 km of the Galapagos Science Center on San Cristobal Island. Twenty-six lizards (*N*~male~ = 13, *N*~female~ = 13) were captured along a walking trail to Punta Carola (0\u00b053\u203232\u2033 and 89\u00b036\u203243\u2033) on 5 July 2017. On the following day, 21 lizards (*N*~male~ = 12, *N*~female~ = 9) were captured along Tijeretas trail (0\u00b053\u203234\u2033and 89\u00b036\u203232\u2033).\n\nEach lizard was captured by noose, placed in a cloth bag, and transported to a temporary field laboratory for processing (see below). Afterwards, to avoid recapturing the same individuals, each was marked on its dorsum with a spot of non-toxic, red acrylic paint used previously in other studies ([@coy046C8]; [@coy046C33]; [@coy046C41]) and known to wash or rub off within 10 days of application (C.A. Valle, personal observation). Because the study was conducted outside of the breeding season and the paint fades quickly, the marking was unlikely to affect the behaviour, fitness or survival of the lizards.\n\nMorphological measurements and body temperature {#coy046s6}\n-----------------------------------------------\n\nA flexible measuring tape was used to determine snout-vent length (SVL) and total length (TL). Head length (HL), head width (HW), head depth (HD), forelimb length (FLL) and hindlimb length (HLL) measurements were made using digital calipers. Body weight was measured with a spring scale (Pesola^\u00ae^). An EBRO^\u00ae^ Compact J/K/T/E Thermocouple Thermometer was used to obtain all temperature readings (model EW-91 219--40; Cole-Parmer, Vernon Hills, IL, USA). Core body temperatures were recorded from the cloaca using the probe T PVC epoxy tip 24GA, usually within five minutes of capture. Finally, sex was determined on the basis of external sexual dimorphism: males are larger and have two white bands on each side of the body, while females are smaller with orange coloration on the ventral--frontal area ([@coy046C48]).\n\nBlood sample collection and handling {#coy046s7}\n------------------------------------\n\nWithin five to ten minutes of capture, each lizard was manually restrained while 0.03--0.2 ml of blood was obtained from the coccygeal vein or the heart using a heparinized 31--29 gauge needle attached to a 1.0-ml syringe. The blood was immediately divided into several samples: (1) for making blood films on clean glass microscope slides, (2) for filling microhaematocrit tubes and (3) for loading the lactate strip or the CG8+ iSTAT cartridge within 5 min of sample collection. Any remaining blood was stored on ice and then maintained at \u221220\u00b0C in a freezer at the Gal\u00e1pagos Science Center for future analysis. In some cases, we were only able to extract enough blood to prepare slides and fill microhaematocrit tubes.\n\nBiochemistry parameters {#coy046s8}\n-----------------------\n\nBlood gas, electrolyte and biochemistry results were obtained using an iSTAT Portable Clinical Analyzer (Heska Corporation, Fort Collins, CO, USA) with CG8+ cartridges. The iSTAT is a portable, hand-held, battery-operated electronic device that measures a wide variety of blood gas, chemistry and haematology parameters using only a few drops (0.095 ml) of whole, non-coagulated blood. The following parameters were measured: anion gap, TCO~2~, haematocrit, haemoglobin, sodium (Na), chloride (Cl), potassium (K), ionized calcium (iCa) and glucose. Mean corpuscular haemoglobin concentration (MCHC) was also calculated \\[Hb (g/L)/Htc (%) \\* 100\\].\n\nHaematology {#coy046s9}\n-----------\n\nHeparinized whole blood was stored on ice immediately after collection. Time-sensitive analyses were completed on the day of sampling. Haematocrit was determined using high-speed centrifugation of blood-filled microhaematocrit tubes. Differential white blood cell counts were made by examining 100 white blood cells on a peripheral smear stained with Diff-Quick stain ([@coy046C11]).\n\nStatistical analysis {#coy046s10}\n--------------------\n\nWe calculated standard summary statistics of all parameters for each sex before testing for differences in morphometrics, biochemistry and haematology between the sexes using *t-*tests. All statistical analyses were run using IBM SPSS.^\u00ae^v23 with a standard \u03b1 level of 0.05.\n\nResults {#coy046s11}\n=======\n\nMorphometrics and body temperature {#coy046s12}\n----------------------------------\n\nMales were larger than females in all morphometric measurements. There were significant differences between sexes in SVL (*P* \\< 0.0001), TL (*P* \\< 0.0001), HL (*P* \\< 0.0001), HW (*P* \\< 0.0001), HD (*P* \\< 0.0001), FLL (*P* \\< 0.0001) and HLL (*P* \\< 0.0001) (see Table [1](#coy046TB1){ref-type=\"table\"}). The mean body mass for males was 23 g (14.5--31 g) and 10 g (6.5--14 g) for females. This difference was also significant (*P* \\< 0.0001). Table 1:Morphometric measurement of San Crist\u00f3bal Lava Lizards (*Microlophus bivittatus*) separated by sexMorphometricsMaleFemale*n* = 25*n* = 22TL (mm)195.7156.4(30.3)(19.8)127.0--237.096.0--182SVL (mm)80.664.7(5.3)(3.5)69.6--90.057.0--71.8HL (mm)16.013.8(0.9)(0.9)13.7--17.411.6--16.0HW (mm)12.610.2(0.96)(0.9)11.1--14.77.6--11.8HD (mm)10.18.20.80.88.4--11.87.1--10.5FLL (mm)33.226.22.41.729.0--39.024.0--29.0HLL (mm)53.441.83.92.546.0--60.037.0--47.0[^1]\n\nThe mean internal body temperature for males (*n* = 20) was 34.37\u00b0C (range 32.4\u00b0C--36.1\u00b0C). For females (*n* = 17), it was 33.14\u00b0C (range 29.3\u00b0C--36.2\u00b0C).\n\nBiochemistry parameters {#coy046s13}\n-----------------------\n\nSummary statistics of the biochemistry data are provided in Table [2](#coy046TB2){ref-type=\"table\"}. The Cl values of 20 individuals exceeded the maximum detectable value of the iSTAT (140 mmol/l). While calculating means for Cl, we used 140 mmol/l for the samples that exceeded the upper limit of the reportable range. Thus, our procedure likely underestimates the mean Cl level for both males and females in the population. Table 2:Blood biochemical values of San Cristobal Lava Lizards (*Microlophus bivittatus*) separated by sexAnalyteMaleFemaleNa (mmolol/l)*n* = 23*n* = 9165.77168.26(6.99)(3.66)149--180160--171K (mmol/l)*n* = 22*n* = 94.263.51(1.90)(0.97)2--8.12--4.9Cl (mmol/l)*n* = 21*n* = 9138.44138.52(2.87)(3.12)131--140132--140iCa (mmol/l)*n* = 23*n* = 91.561.67(0.14)(0.25)1.31--1.91.4--2.24tCO2 (mmol/l)*n* = 23*n* = 99.1710.44(2.55)(2.50)6--157--14Glucose (mg/dl)*n* = 23*n*= 9272.30241.22(45.96)(19.79)199--401218--273Lactate (mmol/l)*n* = 24*n* = 2217.0817.04(2.06)(2.62)13.6--22.28.8--21.1Total protein (mg/dl)*n* = 22*n* = 218.108.77(1.06)(2.07)5--10.25--12Haemoglobin (g/l)*n* = 22*n* = 912.1410.26(1.65)(1.33)8.5--14.37.8--12.2Haematocrit (%)*n* = 24*n* = 2133.3927.80(5.90)(7.85)23--438--39MCHC (g/l)*n* = 21*n* = 9(37.82)(44.22)7.4115.227.94--54.7825.67--75[^2]\n\nWe looked for sex differences between all the biochemistry analytes but did not find differences in lactate (*P* = 0.95), total protein (*P* = 0.2), Na (*P* = 0.32), K (*P* = 0.27), Cl (*P* = 0.94), iCa (*P* = 0.28), tCO~2~ (*P* = 0.21) and MCHC values (*P* = 0.26). Significant differences were found in glucose values (*P* = 0.011), in haematocrit values (*P* = 0.009) and in haemoglobin (*P* = 0.005). In all three of these cases, values in males were higher than those of females (Fig. [2](#coy046F2){ref-type=\"fig\"}).\n\n![Haematocrit, haemoglobin and glucose box plots for males and females of the San Cristobal lava lizard (*M. bivittatus*). Significant differences were found between sexes. Haematocrit (*P* = 0.009), haemoglobin (*P* = 0.005), glucose (*P* = 0.011).](coy046f02){#coy046F2}\n\nHaematology {#coy046s14}\n-----------\n\nLeucocyte values are presented in Table [3](#coy046TB3){ref-type=\"table\"}. No sex differences were detected for lymphocytes (*P* = 0.49), heterophils (*P* = 0.57) or monocytes (*P* = 0.79). We also noted the presence of an intraerythrocytic haemoparasite while examining the blood slides of one female (Fig. [3](#coy046F3){ref-type=\"fig\"}). Table 3:Leucocyte counts of San Cristobal Lava Lizards (*Microlophus bivittatus*) separated by sexCell typeMaleFemale*n* = 25*n* = 22Lymphocyte (%)86.0484.81(5.43)(6.61)70--9370--92Monocyte (%)7.167.45(3.14)(4.54)3--160--20Heterophil (%)5.46.18(4.98)(4.42)1--220--14Eosinophil (%)1.080.32(1.58)(0.65)0--60--2Basophil (%)0.640.54(1.22)(1.22)0--50--5[^3]\n\n![Photographs of selected San Crist\u00f3bal lava lizard (*Microlophus bivittatus*) blood cells stained with Diff-Quick stain at 100\u00d7. (**a**) Heterophil. (**b**) Monocyte. (**c**) The black arrow indicates an eosinophil and the red arrow indicates a lymphocyte. (**d**) Intraerythrocytic haemoparasite.](coy046f03){#coy046F3}\n\n {#coy046s35}\n\n### Brief descriptions of leucocytes {#coy046s15}\n\n*Lymphocytes*: Round cells with a round basophilic nucleus, a variable nucleus to cytoplasm ratio and usually non-visible cytoplasm (Fig. [3](#coy046F3){ref-type=\"fig\"}).\n\n*Monocytes*: Large round to amoeboid cells, nucleus semi-lobed with less condensed chromatin and a pale blue cytoplasm (Fig. [3](#coy046F3){ref-type=\"fig\"}).\n\n*Heterophils*: Round with eosinophilic granules and a bilobed (sometimes four lobules) nucleus (Fig. [3](#coy046F3){ref-type=\"fig\"}).\n\n*Eosinophils*: Large amoeboid to round cells with rather purple conspicuous granules and round nuclei (Fig. [3](#coy046F3){ref-type=\"fig\"}).\n\n*Basophils*: Small round cells with purple granules obscuring the nucleus.\n\nDiscussion {#coy046s16}\n==========\n\nOur results provide the first data on the morphometrics, body temperature, biochemistry and haematology of San Crist\u00f3bal lava lizards (*M. bivittatus*). Such baseline data on the morphology and physiology of wild animals is crucial to understanding various aspects of a population's biology and health ([@coy046C30]) and can be especially important in evaluating the cause of morbidity events and population declines.\n\nMorphometrics can be helpful in evaluating cases of disease or environmental stress, as optimal body condition can be calculated using body weight and size ([@coy046C20]; [@coy046C34]). Similarly, the typical ranges of body temperatures and heart rates of a species are critical metrics useful to veterinary professionals during medical interventions (e.g. anaesthesia and surgery). These parameters are known to vary with handling stress in reptiles (e.g. [@coy046C40]), so the values reported here should be used only as approximations.\n\nThe measurement of biochemical and haematological parameters can serve as a valuable tool for evaluating and monitoring the health of wild reptile populations ([@coy046C45]; [@coy046C11]). To determine the significance of changes in biochemical and haematological values associated with factors such as disease, injury, pollutants or starvation, it is essential to establish species-specific (or at least taxon-specific) normal values for parameters of interest ([@coy046C35]). The sole published haematological study of a neotropical ground lizard (*Tropidurus torquatus*) reported only erythrocyte values and morphology ([@coy046C43]), and no data have previously been reported for *Microlophus*.\n\nIn general, the biochemistry values reported in Table [2](#coy046TB2){ref-type=\"table\"} were very similar to those of other lizards ([@coy046C19]; [@coy046C49]; [@coy046C17]; [@coy046C27]; [@coy046C31]; [@coy046C37]; [@coy046C16]; [@coy046C35]). Because the iStat cannot measure chloride concentrations greater than 140 mmol/l and 20 individuals exceeded this limit, we suspect that chloride levels are high in clinically healthy individuals. We note that similarly high values exceeding 140 mmol/l have been reported in other lizards such as *Pogona vitticeps and Iguana iguan*a ([@coy046C22]; [@coy046C27]).\n\nFor lactate values, we found higher levels than reported by [@coy046C24] in *Sceloporus torquatus*, [@coy046C25] in *S. grammicus*, and [@coy046C35] for marine iguanas (*Amblyrhynchus cristatus)*. Very few studies have reported lactate levels in squamate reptiles ([@coy046C35]; [@coy046C24], [@coy046C25]), and most evaluated lactate production after exercise in diving reptiles ([@coy046C3]; [@coy046C42]; [@coy046C26]; [@coy046C50]). In squamates, blood lactate levels may be affected by altitude ([@coy046C24], [@coy046C25]), exercise ([@coy046C7]; [@coy046C23]; [@coy046C44]) and handling ([@coy046C35]). Lactate levels may also be influenced by glucose and oxygen levels ([@coy046C23]; [@coy046C5]), but apparently there is little dependence on body temperature ([@coy046C6]). In mammal medicine, hyperlactatemia usually correlates with disease severity and mortality ([@coy046C53]) and with capture myopathy and muscle damage in birds ([@coy046C9]). Little is known, however, about its clinical use in reptile medicine and further study is needed.\n\nGlucose levels can vary in reptiles depending on nutritional or environmental conditions ([@coy046C11]). In *M. bivittatus* glucose was significantly lower in females than males. Such sex-based variation was not identified in a study of the agamid genus Calotes ([@coy046C15]), and the prevalence of this pattern in other lizard families, including Tropiduridae, is unknown. Because glucose metabolism in reptiles is similar to that in mammals, a known difference between human sexes may provide insight ([@coy046C2]; [@coy046C39]). Compared to men of the same age, healthy women have lower skeletal muscle mass and higher adipose tissue mass, more circulating free fatty acids, and higher intramyocellular lipid content, all of which might promote insulin resistance in women relative to men ([@coy046C39]). In addition, insulin sensitivity is greater in women as a result of higher glucose disposal by skeletal muscles ([@coy046C39]). The mechanisms for facilitated glucose homoeostasis are unclear, but could be due at least in part to the effect of circulating oestrogen ([@coy046C39]). Although a similar process may explain our findings in *Microlophus*, it is important to note that [@coy046C18] found that male tortoises had higher glucose levels than females, raising questions about whether such patterns are conserved across amniotes.\n\nBlood cell counts and morphology can vary greatly among species of reptile, even among members of the same genus ([@coy046C45]; [@coy046C30]). In addition, numerous factors, including age, sex, environment, season, presence of environmental stressors, parasite load, nutritional state and restraint, can complicate the evaluation of haematological data in reptiles ([@coy046C11]). Therefore, published reference intervals provide only a baseline for interpretation, and veterinarians need to be aware of these factors to accurately interpret and correlate haematological and clinical findings in reptile patients ([@coy046C45]).\n\nCellular morphology of *M. bivittatus* is similar to that reported in the closely related families Iguanidae, Corytophanidae, and Liolaemidae. As in other lizards that have been studied \\[e.g. *Cyclura nubila*, ([@coy046C1]); *Amblyrhynchus cristatus* ([@coy046C35]); *Salvator* (*Tupinambis*) *merianae* ([@coy046C47]); *Pogona vitticeps* ([@coy046C22])\\], the highest leucocyte population was the lymphocytes.\n\nWe could not analyse total erythrocyte counts due to logistical constraints. However, haematocrit, haemogoblin and MCHC were run as references for red cell values. Lizard haematocrits tend to be higher than those of other reptiles ([@coy046C11]), and lava lizard values resembled those reported in other lizards. Differences were found between sexes; males had higher values, a phenomenon previously reported in reptiles ([@coy046C27]). It is important to know normal values and differences between sexes to identify possible disease events. For example, low haematocrit may be due to anaemic processes (haemorrhagic, haemolytic or decreased red cell production), but a higher than normal value may be due to haemoconcentration, usually as a result of dehydration. Haemoconcentration also sometimes involves altitude ([@coy046C29]; [@coy046C24], [@coy046C25]) or body size and age ([@coy046C21]). Anaemia has been reported in free-living birds and mammals, but is also often associated with traumatic events or ill health (e.g. severe dehydration or nutritional stress, toxicity, or high blood loss due to parasites ([@coy046C51]). However, a decrease in haematocrit has also been observed in birds during 'normal' reproduction. Specifically, haematocrit has been found to drop routinely between pre-laying and egg-laying, possibly due to plasma vitelogenin levels, estradiol effects on erythropoiesis, nutritional stress and temperature ([@coy046C51]). Similar processes may drive the differences observed in lizards.\n\nThe presence of hemoparasites in wild reptiles is common and is usually considered non-pathogenic ([@coy046C45]). The fact that just one individual had hemoparasites (an unknown species of haemogregarine) is surprising. It might be due to limitations of sample size, parasite prevalence, and susceptibility or resistance to infection of particular individuals ([@coy046C36]). A study in wall lizards in Portugal with a similar sample size reported more than 50% prevalence using microscope diagnostics ([@coy046C36]), suggesting that *M. bivittatus* might actually have low prevalence. Because hemoparasites can infect multiple species of sympatric reptile ([@coy046C36]), the relatively low reptile diversity of San Cristobal might limit hemoparasites on the island.\n\nOne caveat about our study is that the iSTAT unit used to analyse blood chemistry was designed as a fast and efficient way to determine blood gas and chemistry parameters in humans. For this reason, future work should involve corroborating these results with methods that are more commonly used for reptile work in veterinary laboratories ([@coy046C46]; [@coy046C28]), which was impractical in our study due to export constraints and logistical challenges.\n\nIn sum, our study represents an initial step toward producing a foundation for future health assessment work on *Microlophus* lava lizards, a small and interesting group of reptiles in the Gal\u00e1pagos. Because there are several closely related species on various islands in the Archipelago, future studies should determine the extent to which the parameters measured here vary across species.\n\nWe wish to thank the following people for their support and assistance: GSC staff: Carlos Mena, Stephen Walsh, Philip Page, Sofia Tacle, Sylvia Sotamba and Karla Vasco. DPNG staff: Galo Quezada, Daniel Lara-Solis, Jorge Carri\u00f3n- Tacuri, Ingrid Jaramillo, Gabriel V\u00e1squez and Maryuri Y\u00e9pez. In addition, we thank the Gal\u00e1pagos National Park (GNP) for the request and trust granted for sampling, and Gal\u00e1pagos Science Center (GSC) for the logistic support during the study and The Heska Corporation. Finally, we thank to the GNP rangers from the GNP technical office at San Crist\u00f3bal for help and assistance in collecting data, special thanks to the ranger Jeffreys M\u00e1laga.\n\nFunding {#coy046s17}\n=======\n\nThis research was authorized by the Gal\u00e1pagos National Park Service and was conducted with support of the Heska Corporation, the Gal\u00e1pagos Academic Institute for the Arts and Sciences (GAIAS)---Universidad San Francisco de Quito (USFQ) and the Gal\u00e1pagos Science Center-USFQ/University of North Carolina---Chapel Hill.\n\n[^1]: Values represent means, standard deviation (parentheses) and range (minimum--maximum). TL (tail), SVL (snout-vent length), HL (head length), HW (head width), HD (head diameter), FLL (front leg length), HLL (hind leg length).\n\n[^2]: Values represent *n*, means, standard deviation (parenthesis) and range (minimum--maximum). Sample sizes vary because some females were too small to extract the amount of blood required for all analyses.\n\n[^3]: Values represent means, standard deviation (parenthesis) and range (minimum--maximum).\n"}
+{"text": "Diabetes mellitus is a complex metabolic disorder, the 5^th^ leading cause of mortality worldwide resulting in more than 4 million deaths annually. Recent reports have predicted a 150% increase in occurrence in the next 20 years, with a major burden on medicinal care due to its devastating complications[@b1]. Diabetic nephropathy (DN) is a common complication of diabetes, with 25--45% of patients developing renal fibrosis and progressing to end stage renal disease[@b2]. There is no cure for DN and therapeutic efforts are focused on limiting loss of renal function and associated symptoms[@b3]. During the last fifteen years, significant advances have been made concerning the mechanisms underlying initiation and progression of chronic kidney disease. The capacity of renal mesangial cells to undergo remodelling and acquire fibroblastic plasticity was first suggested by studies from our laboratory that identified the recapitulation of ontogenic gene expression profiles in experimental models of diabetic nephropathy and in patients[@b4] ([Supplementary Fig. S1](#S1){ref-type=\"supplementary-material\"}). Subsequent studies have extensively characterised the role of Transforming Growth Factor \u03b21 (TGF\u03b21) in mediating these change however despite significant efforts in this area, therapeutic interventions have yet to demonstrate clinical efficacy. New paradigms are emerging from recent studies elucidating the instructive role of TGF\u03b2 during embryonic development, coupled with the identification of parallel processes in adult tissues[@b5].\n\nCell fate specification is a progressive process of diversification through which a cell, by undergoing profound changes in gene expression and regulation, takes its role within a defined context. On the other hand, cell fate conversion is considered a process by which a cell can change its phenotype and acquire a new and distinct \"altered\" fate[@b6]. While the first process is pivotal during development, the latter is increasingly recognized as fundamental not only during embryogenesis but also in numerous disease states[@b7][@b8]. A cell must acquire a plastic phenotype in order to properly adapt and respond to environmental stimuli. These adapting processes involve and are controlled by the interplay between microRNAs, transcription factors (TFs), and epigenetic modifiers that work in concert to determine cell fate.\n\nHuman Mesangial Cells (HMCs) are a specialized type of microvascular pericyte[@b9] anchored to the glomerular membrane. Due to their intrinsic nature, these cells are highly plastic and responsive to the surrounding microenvironment. These responsive mechanisms often result in detrimental processes being triggered by extracellular stimuli that can lead to the destruction of the complex glomerular and renal ultrastructure[@b10]. Frequently, these alterations result in a change in cytoskeletal-mediated contractility, reflected in dynamic focal adhesions[@b11]. Perhaps this is best evidenced by the apparent alterations in actin dynamics mediated by TGF\u03b2 and CTGF, reflecting changes in contractility both *in vivo* and *in vitro*; previous studies carried out in our laboratory have demonstrated that HMCs acquire plasticity triggered by hyperglycaemia and growth factors[@b4] which greatly enhances the expression of a family of microRNAs, the miR302 family. The miR302 family is composed of 4 members, miR302a/b/c/d that are transcribed as a single polycistronic cluster[@b12]; this cluster is prominently expressed in Embryonic Stem Cells (ESCs) while its expression is decreased during differentiation and commitment[@b13]. It has been previously shown that miR302 can promote iPSC (induced Pluripotent Stem Cells) generation[@b14][@b15][@b16] and its expression is directly regulated by the stemness factors, Oct4 and Sox2, and Nanog[@b17][@b18].\n\nThe primary validated target of miR-302 is the Transforming Growth Factor (TGF)-\u03b2 Type II receptor (T\u03b2RII)[@b14]. TGF\u03b2 has a prominent role in triggering Epithelial to Mesenchymal Transition (EMT) during embryogenesis and activating parallel processes in disease[@b19]. Strikingly, miR302 is able to control and impact on the TGF\u03b2 pathway in an extensive and context-dependent manner; it has a well-established role in promoting the acquisition of pluripotency by targeting the T\u03b2RII, thus blocking the activation of the pathway[@b12][@b14], however it can also propagate and promote its activation in ESCs by levelling the expression of LEFTY1[@b20].\n\nThe identification of embryonic stem cell specific miRNAs led to the widely accepted hypothesis that interplay between specific microRNAs and their repressed targets controls both the maintenance of stemness and the specification of cell types[@b21][@b22]. Increasingly, parallel processes in pathogenesis are recognised as critical mediators of damage and repair. Specifically, the potential role of the miR302 and Let-7 families in both these processes have been recently established by our group and others[@b4][@b23].\n\nIn addition to these processes, a third level of regulation, involving the remodelling of the chromatin environment, has emerged as numerous studies have demonstrated that cell type specific regulatory genes can be identified by specific histone marks[@b24]. Among others, changes in the methylation status of histone H3 have been associated with stemness, cell specification and numerous diseases[@b25][@b26]. The methylation of histone H3 on the lysine 4 (H3K4) and and 26 (H3K26) are generally associated with active transcription[@b27], whereas permissive promoters are enriched with both active (H3K4) and repressive marks (H3K27) and considered to exist in a \"poised\" state[@b27][@b28]. Central to these processes is the Polycomb Repressive Complex 2 (PRC2), which contains EZH2 (Enhancer of zeste homolog 2), a histone methyltransferase that catalyzes the trimethylation of H3K27[@b29], mediating gene repression, and additional core components EED, SUZ12 and RBBP4/RbAp48/NURF55.\n\nIn the present study we investigated the role of miR302 in regulating mesangial plasticity and explore the idea that partial reprogramming of mesenchymal cells leads to the acquisition of a \"poised\" state that may be manipulated for therapeutic repair.\n\nResults\n=======\n\nOverexpression of miR302a/b/c/d in Human Mesangial Cells\n--------------------------------------------------------\n\nHMCs were seeded at a very low confluency and then incubated for 48\u2009h with a polycistronic lentiviral vector encoding all four members of the miR302 family and a Green Fluorescent Protein (GFP)-reporter. At 7 days post transduction all cells demonstrated clear GFP expression ([Fig. 1A](#f1){ref-type=\"fig\"}). The expression of the miR302d was analysed by RealTime PCR as a readout of the level of expression of miR302 in the cells. 7 days post lentiviral transduction, a marked increase in miR302d expression ([Fig. 1B](#f1){ref-type=\"fig\"}) was observed, indicating successful transduction. RNA and protein were extracted at various time points in order to investigate the phenotypic changes caused by the miR302 overexpression system. One of the best-characterised targets of the miR302 family is the T\u03b2RII[@b4][@b14]. T\u03b2RII is involved in EMT and its activation leads to the phosphorylation of Smad2 and Smad3 resulting in their translocation from the cytoplasm to the nucleus[@b30]. We verified the effective downregulation of T\u03b2RII by both RNA ([Fig. 1C](#f1){ref-type=\"fig\"}) and protein analysis ([Fig. 1D](#f1){ref-type=\"fig\"}). As expected, the receptor is repressed throughout all time points. miR302 is the most important and abundant microRNA present in human ESCs (hESCs)[@b12]. Since its promoter can be directly bound and regulated by Oct4[@b17][@b18] and various reports have highlighted its ability to induce Oct4 expression[@b16], we investigated whether miR302 overexpression caused the acquisition of a pluripotent phenotype in HMCs. RealTime PCR analysis for Oct4 and Nanog was performed at various time points ([Supplementary Fig. S2](#S1){ref-type=\"supplementary-material\"}) showing no expression of either transcription factors. Noticing the appearance of rounded granulated colonies between 14 and 21 days post-transduction, we hypothesised that these colonies originated from HMCs cells successfully reprogrammed toward pluripotency. After picking, HMC-derived colonies were cultured on matrigel under stem cell-like conditions for up to 21 days. These colonies were able to attach to the coated plates and proliferate. Although they became bigger and tried to divide ([Supplementary Fig. S3A](#S1){ref-type=\"supplementary-material\"}), no obvious hallmarks of a pluripotent phenotype were observed nor was there any significant change in Oct4 expression ([Supplementary Fig. S3B](#S1){ref-type=\"supplementary-material\"}). Taken together, these results demonstrated the successful transduction of HMCs with miR302 lentivirus and its ability to block T\u03b2RII expression. Moreover, they showed that miR302 upregulation alone is not sufficient to reprogram HMCs to pluripotency.\n\nmiR302 upregulates Snail expression in HMCs\n-------------------------------------------\n\nHaving verified the lack of pluripotency in miR302-HMCs, we proceeded to analyse the phenotype acquired by the cells. Interestingly, increased expression of Snail (or SNAI1) was consistently observed at 3 days and 7 days post transduction in miR302 overexpressing cells, although some variability in levels were apparent, likely reflecting the heterogeneous and asynchronous nature of the cell populations at these time points. ([Figure 2A,B](#f2){ref-type=\"fig\"}, quantified in [Supplementary Fig. S4A](#S1){ref-type=\"supplementary-material\"}). We verified that this upregulation was not due to a nonspecific effect by analysing the resulting Snail expression in an additional arbitrarily chosen cell type ([Supplementary Fig. S5A](#S1){ref-type=\"supplementary-material\"}) after miR302 overexpression. This was a particularly unexpected result as Snail is widely regarded as the most important transcription factor involved in driving EMT as a result of T\u03b2RII/T\u03b2RI activation. No increased expression of Snail was apparent in HMCs following TGF\u03b2 treatment ([Supplementary Fig. 5B](#S1){ref-type=\"supplementary-material\"}) and additionally, they also exhibited downregulation of miR302 ([Supplementary Fig. 5C](#S1){ref-type=\"supplementary-material\"}), supporting the idea that Snail is not regulated by TGF\u03b2 in this context.\n\nTo rule out the possibility that Snail expression was linked to an alternative mechanism of activation of the TGF\u03b2 Receptor I (T\u03b2RI) we treated miR302-HMCs with a potent and specific inhibitor of the receptor[@b31], SB431543, 7 days post miR302 transduction. After 7 and 14 days of treatment with the inhibitor ([Fig. 2C](#f2){ref-type=\"fig\"}, quantified in [Supplementary Fig. S4B,C](#S1){ref-type=\"supplementary-material\"}) the cells maintained a high level of Snail protein expression, comparable to the standard culture condition ([Fig. 2C](#f2){ref-type=\"fig\"}). Similarly, Snail mRNA expression is maintained throughout the 2 weeks of treatment with the inhibitor ([Supplementary Fig. S4D](#S1){ref-type=\"supplementary-material\"}). A recent study from the Daley laboratory demonstrated that Snail is paradoxically required in order to successfully reprogram fibroblasts to pluripotency[@b32]. They showed how Snail promotes the acquisition of plasticity by binding the promoter of the Let-7 family of microRNAs, thus causing its downregulation. This hypothesis is particularly fascinating when considering the reciprocal relationship that occurs between miR302 and Let-7[@b13]. In order to verify whether this mechanism is valid in this context, we measured the levels of two members of the Let-7 family, Let-7a and Let-7c, 7 days post transduction. Although the levels were lower in miR302 cells when compared to scramble cells, no statistically significant difference was detected ([Fig. 2D](#f2){ref-type=\"fig\"}), demonstrating that Snail in this context does not suppress Let-7 expression.\n\nThese results demonstrated, for the first time, that miR302 drives Snail expression in HMCs, during the acquisition of plasticity.\n\nEZH2 upregulation in miR302-HMCs is independent from Snail activation\n---------------------------------------------------------------------\n\nAs Snail plays a role in enhancing the acquisition of pluripotency in fibroblasts[@b32] and pathogenic plasticity during renal failure[@b33], we analysed the expression of other proteins known to be involved in these processes. We verified that Slug (or SNAI2), another member of the Snail family is similarly upregulated within the first 2 weeks post transduction ([Fig. 3A](#f3){ref-type=\"fig\"}, quantified in [Supplementary Fig. S6A](#S1){ref-type=\"supplementary-material\"}). Intriguingly, miR302-HMCs also exhibited enhanced expression of EZH2 coincident with Snail upregulation. EZH2 is the catalytic subunit of the PRC2 Complex and it catalyses the trimethylation of the lysine 27 on histone H3[@b34]; by doing so the PRC2 complex can repress gene transcription through the control of promoter access[@b29]. EZH2 has been widely studied as a protein implicated in cancer and metastatic progression and its upregulation is linked to an aggressive phenotype and to enhanced proliferation by cancer cells[@b35]. Interestingly, in miR302-HMCs, EZH2 showed the same expression pattern as Snail, being significantly upregulated within the first 2 weeks post transduction, while its expression subsequently decreases ([Fig. 3A](#f3){ref-type=\"fig\"}, quantified in [Supplementary Fig. S6A](#S1){ref-type=\"supplementary-material\"}). Of note, the H3K27 tri-methylation mark, a repressive mark produced by EZH2 activity, was unexpectedly decreased, irrespective of EZH2 upregulation. Moreover, no difference was detected across all the time points for SUZ12, another core component of the PRC2 complex ([Fig. 3A](#f3){ref-type=\"fig\"}, quantified in [Supplementary Fig. S6A](#S1){ref-type=\"supplementary-material\"}), suggesting a specific effect on EZH2 by miR302. Finally, we established the correct localization in the nuclear fraction of EZH2 and Snail ([Supplementary Fig. S7](#S1){ref-type=\"supplementary-material\"}). Interestingly, in miR302-HMCs the downregulation of T\u03b2RII by miR302 overexpression did not affect the localisation of Smad3 in the nuclei ([Supplementary Fig. S7](#S1){ref-type=\"supplementary-material\"}), suggesting again a TGF\u03b2 independent effect of miR302 in HMCs. A recent paper from the Liang group, demonstrated that Snail expression in cancer cells is able to upregulate EZH2 by inhibiting miR101 expression[@b36]. As EZH2 and Snail expression appeared to be tightly linked, and taking into account the fact that Smad3 was still present regardless of the silencing of the TGF\u03b2 type II receptor, we decided to knock down EZH2, Snail and Smad3 in miR302-HMCs using esiRNA. At 7 days post transduction, when the HMCs showed a high level of expression of both EZH2 and Snail we knocked down Snail, EZH2 and Smad3 respectively, while a scrambled off-target sequence was used as a control. Interestingly, only the Smad3 knockdown resulted in the downregulation of EZH2 suggesting that the 2 proteins may form a cooperative, regulatory complex. Knock down of EZH2 does not affect Snail and vice versa ([Fig. 3B](#f3){ref-type=\"fig\"}). SUZ12 was used as an unrelated control for off-target effect. Taken together, these results demonstrated that miR302 regulates EZH2 and Snail expression independently.\n\nmiR302-HMCs do not acquire a cancer-like phenotype\n--------------------------------------------------\n\nSince cancer cells expressing high levels of EZH2 display a higher rate of proliferation[@b37][@b38] we performed an MTT assay on miR302-HMCs. A first set of experiments was carried out by seeding miR302-HMCs at 7 days post transduction ([Fig. 4A,B](#f4){ref-type=\"fig\"}) finding, as expected, an increased rate of cell proliferation. Importantly, miR302 cells return to a normal proliferation rate once the expression of EZH2 started to drop ([Fig. 4C,D](#f4){ref-type=\"fig\"}). In order to understand whether the increased proliferation of HMCs was linked to EZH2 activity we treated the cells at the time of seeding with DZnep, an inhibitor of EZH2 catalytic activity[@b39]. By blocking the activity of EZH2 the cells no longer demonstrated any changes in proliferation ([Fig. 4E](#f4){ref-type=\"fig\"}).\n\nIncreased proliferation, together with upregulation of both EZH2 and Snail are accepted hallmarks of tumourigenesis and metastastic progression[@b36], so we therefore investigated whether HMCs transduced with miR302 acquired a cancer-like phenotype. A well-known transcription factor involved in tumour progression is NF\u03baB. NF\u03baB is interesting for a number of different reasons; it represents not only an important link between cancer and inflammation[@b40] but it is also known for its role as Snail activator promoting cancer aggressiveness[@b41]. Intriguingly, analysis of NF\u03baB expression by Western Blot ([Fig. 5A](#f5){ref-type=\"fig\"}, quantified in [Supplementary Fig. S8](#S1){ref-type=\"supplementary-material\"}), found that it is decrased in miR302-HMCs compared to scramble transduced HMCs. Moreover, no change in TGF\u03b2 expression was detected at any of the analysed time points ([Fig. 5C](#f5){ref-type=\"fig\"}).\n\nIn hESCs, miR302 inhibits tumorigenicity by controlling G~1~-S cell cycle transition and promoting p16/Ink4a upregulation[@b42], a well-known tumour suppressor protein[@b43]. By doing so, the members of this family protect the pluripotent nature of hESCs, preventing them from giving rise to cancer formation. We analysed p16/Ink4a expression by RealTime PCR ([Fig. 5B](#f5){ref-type=\"fig\"}) and found, in line with the literature, that miR302 overexpression caused an upregulation of p16/Ink4a expression.\n\nFinally, we investigated the involvement of miR302 in promoting/preventing apoptosis. miR302 has been shown to cause apoptosis in cancer cells[@b44], while in normal cells it does not affect the apoptotic pathway[@b42]. To stimulate the activation of the apoptotic pathway in HMCs we treated the cells with etoposide, a potent anticancer drug capable of initiating a program of apoptosis[@b45]. As shown in in [Fig. 5D](#f5){ref-type=\"fig\"}, no difference was detected among the miR302, scramble and non-transduced groups regarding the appearance of the cleaved form of the caspase 3, while p53 upregulation is slightly diminished in miR302 and scramble cells compared to the non-transduced ones. This consistency, among the different groups, reflects the lack of an aggressive phenotype in miR302-HMCs.\n\nThese results are particularly important as they confirm that, although miR302-HMCs acquire a more-plastic phenotype, this process, in healthy cells, is not linked to a cancer-like state, raising the possibility that it can be used to manipulate cell fate and identity.\n\nmiR302-HMCs acquire plasticity that facilitates *de novo* assembly of ZO-1 containing junctions\n-----------------------------------------------------------------------------------------------\n\nTaking these results together, miR302-HMCs exhibit a higher degree of plasticity, evidenced by Snail and EZH2 expression and their higher proliferation rate, when compared to scramble or control (non-transduced) HMCs.\n\nIn order to evaluate if this \"positive\" plasticity can be manipulated to push the cells toward a different, more epithelial phenotype, we devised a protocol employing a combination of two different inhibitors of GSK-3\u03b2 and of the EZH2 activity respectively. DZnep was chosen as previous experiments carried out in our laboratory demonstrated a facility to increase E-cadherin expression in epithelial cells ([Supplementary Fig. S9](#S1){ref-type=\"supplementary-material\"})[@b46]. Moreover, our group has demonstrated a role for EZH2 in supressing E-cadherin expression in the presence of Smad3[@b46] (Manuscript under submission). Similarly, GSK-3\u03b2 inhibition has been shown to stabilize epithelial junctions making ESCs more \"epithelial\"[@b47]. 12 days post miR302 transduction, when the cells exhibit the highest level of Snail and EZH2 expression, HMCs were plated on matrigel with the addition of the inhibitors. Within days of switching to a different microenvironment, miR302-HMCs started to acquire a more cobblestoned shape, ([Fig. 6A](#f6){ref-type=\"fig\"}), while scramble cells stopped proliferating and died. To assess the nature of these cobblestoned cells we stained the miR302-HMCs for Zonula Occludens (ZO)-1, a protein contained in the tight junction. The miR302-HMCs, demonstrated *de novo* tight junction formation (white arrows, [Fig. 6C](#f6){ref-type=\"fig\"}), indicating that the cells have lost their characteristic scattered phenotype and formed connections with the surrounding cells. Interestingly, expression of the transcriptional repressor Snail was silenced in both miR302 and scramble cells after plating on matrigel ([Fig. 6B](#f6){ref-type=\"fig\"}). Finally, miR302-HMCs showed a statistically relevant upregulation of CDH1 (E-cadherin) expression at mRNA level ([Fig. 6D](#f6){ref-type=\"fig\"}), in particular when treated with DZnep; overexpression of miR302 alone is not sufficient to trigger CDH1 expression.\n\nThese results demonstrated for the first time the *de novo* expression of epithelial junctions components by human mesangial cells demonstrating the potential of utilising such an approach for reprogramming in chronic renal disease.\n\nDiscussion\n==========\n\nIn the context of chronic renal diseases, TGF\u03b2 plays an important role in directing cellular damage in the DN milieu[@b48]. The TGF\u03b2 superfamily is composed of activins, nodal, Bone Morphogenetic Proteins (BMPs), Growth and Differentiation Factors (GDFs) and anti-M\u00fcllerian hormone (AMH). They act by regulating various developmental and homeostatic processes and are involved in numerous human diseases[@b49] as well as being considered a viable therapeutic target; nevertheless their pleiotropic signalling has resulted in limited success in targeting their activity. New insights provided herein suggest that miR302 plays a critical role in regulating epigenetic phenomena that control cell fate.\n\nDiabetes mellitus is a complex metabolic disorder characterised by persistent hyperglycaemia; emerging evidence indicates that multiple factors involved in the aetiology of diabetes can alter epigenetic mechanisms and regulate susceptibility to microvascular complications. Particularly important is the role of the PRC2 methyltransferase, EZH2. EZH2 is the catalytic subunit of the PRC2 complex which catalyses the trimethylation of lysine 27 on the histone H3 and mediates genes silencing[@b29]. It is involved in repressing the E-cadherin promoter during EMT and its expression correlates with invasiveness and aggressiveness in multiple types of cancers[@b35]. In the present study we investigated the role that miR302 plays in acquired plasticity in primary human mesangial cells. HMCs are significantly affected during renal diseases and these alterations lead to the expansion of mesangial compartment and extracellular matrix deposition[@b9][@b50]. Our findings demonstrated an unexpected role for the miR302 family as part of a regulatory loop that controls cellular plasticity.\n\nPrevious analysis carried out in our laboratory demonstrated that miR302 is upregulated in HMCs undergoing a process of acquired, pro-fibrotic plasticity[@b4]. This might seem counterintuitive considering that its upregulation is triggered by CTGF stimulation, which is widely accepted to work in concert with TGF\u03b2 to activate fibrosis[@b51] and impacts directly on the TGF\u03b2 pathway. Intrigued by these findings, we investigated the role that miR302 plays in HMC plasticity. During miR302 mediated differentiation, we found persistently high levels of Snail even after a decline in mRNA. In contrast, scramble control cells always exhibited low levels of mRNA and no protein was detectable. This likely reflects the dynamic nature of the reprogramming process where cells are likely to be asynchronous. In addiction, overexpression of miR302 in HMCs causes a strong upregulation of Slug, another member of the Snail family. Snail and Slug are both considered master regulators of the EMT process. This observation is important, since normally miR302 overexpressing cells are prevented from activating Smad3 and Smad2 upon TGF\u03b2 stimulation[@b4]. Nevertheless, increased Snail expression is driven by miR302 in various cell types, consistent across both epithelial and mesenchymal phenotypes ([Fig. 1](#f1){ref-type=\"fig\"} and [Supplementary Fig. S5](#S1){ref-type=\"supplementary-material\"}), pointing toward a TGF\u03b2 independent upregulation of the two TFs triggered by miR302. These two transcription factors have been recently implicated in iPSC generation[@b32][@b52][@b53]. In particular, Snail is required during the initial phase of reprogramming in order for the cells to acquire a higher degree of plasticity, by supressing Let-7 expression[@b32], while Slug is pivotal in the later phases, although acting independently of its role as an EMT-transcription factor[@b52]. Interestingly, miR302 and Let-7 have essentially opposing roles during embryonic development, where miR302 identifies pluripotent stem cells whereas Let-7 is highly expressed in differentiated cells[@b13].\n\nEMT involves distinct phenotypic changes through which a polarized epithelial cell gains a scattered mesenchymal phenotype, increasing its motility and invasion properties. A recent paper from the group of Nieto[@b33] points out that within injured renal parenchyma, tubular epithelial cells undergo a partial EMT, gaining a hybrid phenotype, in which they acquire Snail expression while maintaining their epithelial hallmarks. According to the authors, this partial EMT of the renal epithelial cells is able to trigger the release of inflammatory and fibrotic cytokines, such as TGF\u03b2, in the microenvironment of the injured kidney leading to the development of fibrosis. Interestingly, HMCs overexpressing miR302 in an *in vitro* healthy context do not become more fibrotic, as indicated by the lack of NF\u03baB expression and TGF\u03b2 mRNA. As Snail has been implicated in repressing the expression of Let-7, we analysed the level of expression of Let-7a and c, finding no correlation at early time points post transduction ([Fig. 2D](#f2){ref-type=\"fig\"}). This is perhaps not surprising as the processes of iPSC generation and acquisition of plasticity by HMCs, although having some common ground, remain substantially different. The finding that Smad3 is present and phosphorylated in the nuclei of miR302-HMCs, despite the silencing of the receptor, is intriguing. This suggests that Smad3 is required by cells as an adaptor for phenotypic transition, as its knock down results in the decreased levels of EZH2. Indeed, other TGF\u03b2 family members have been proposed to signal in a similar manner; for example Smads likely direct Jmjd3 to target genes in ESCs and during differentiation via direct interaction in response to nodal/activin and facilitating target gene de-repression during endodermal specification[@b54]. Furthermore, this points to a functional association between the two proteins, even though the expression relationship is not reciprocal as EZH2 knockdown does not affect Smad3 expression. This result supports recent findings from our laboratory that demonstrate a direct interaction between EZH2 and Smad3[@b46]. Recent work from the Kaji group, demonstrated that the activated forms of Smad3 and Smad2 are not only required during iPSC generation but also promote the direct reprogramming of cells into a different type[@b55]. Moreover, the use of the SB431542, an inhibitor of the T\u03b2RI, during somatic cell reprogramming, enhances the acquisition of pluripotency and in the absence of TGF\u03b2 stimulation causes an increase in the level of Smad3/2 phosphorylation, results in line with our observations.\n\nHMCs were found to increase their rate of proliferation with EZH2 overexpression. It has been proposed that EZH2 may play a role independent of its activity in the PRC2 complex[@b56]. In the present study we confirmed that EZH2 expression is upregulated in miR302-HMCs however the methylation of lysine 27 of histone H3 is decreased ([Fig. 3](#f3){ref-type=\"fig\"}). Interestingly, we recently demonstrated a direct interaction between Smad3 and EZH2 during neuroepithelial differentiation[@b57]. Silencing of the type II TGF\u03b2 receptor by miR302 likely perturbs the interaction with EZH2 while the addition of DZnep erases the H3K27me3 mark, leading to the derepression of epithelial genes as evidenced by *de novo* detection and expression of ZO-1 and E-cadherin in miR302-HMCs. As EZH2 and Snail show the same trend of expression we expected to find a reciprocal interaction between the two of them; however we clearly demonstrated that the lack of EZH2 does not impair Snail expression and vice versa. Nevertheless, these two proteins co-operate to control the acquisition of plasticity in HMCs, as evidenced by *de novo* expression of ZO-1 and E-cadherin, suggestive of MET.\n\nIn summary, the complex signalling systems that regulate the acquisition of plasticity in mesangial cells have clear parallels with the processes controlling pluripotency and are likely mediated by a small number of transcriptional master regulators including Smad3 and Snail. Increased miR302 is sufficient to drive cells towards a highly plastic \"hybrid\" state in which they are susceptible to microenvironmental cues and differentiation. Further understanding of the intricate interplay between TGF\u03b2 receptors, Smad3, Snail and EZH2 will not only increase our knowledge of epigenetic and pathogenic processes in diverse diseases such as diabetes and cancer, but will point towards new therapeutic approaches focused on the exploitation and control of cellular plasticity ([Fig. 7](#f7){ref-type=\"fig\"}).\n\nMaterial and Methods\n====================\n\nCell culture, viral transduction, TGF\u03b2 and inhibitor treatments of HMCs\n-----------------------------------------------------------------------\n\nPrimary HMCs (Human Mesangial Cells; Lonza) were maintained in MCDB-131 medium (Gibco) while Human Kidney Cells (HKC8) were maintained in regular DMEM (Gibco), both supplemented with 10% (v/v) heat-inactivated Fetal Bovine Serum (FBS, Life Technologies), 10\u2009mM L-glutamic acid (Gibco) and 100\u2009U/ml penicillin and 100\u2009\u03bcg/ml streptomycin (Gibco). For TGF\u03b2 stimulation, HMCs and HKC8 were serum starved for 24\u2009h. Recombinant human TGF\u03b21 (Promokine) was used at a final concentration of 5\u2009ng/ml and cells were stimulated for 24\u2009h, 48\u2009h and 72\u2009h. Protein and RNA were harvested at each time points. 3-Deazaneplanocin A (DZNep, Calbiochem) was added 1\u2009h before TGF\u03b2 at a final concentration of 5\u2009\u03bcM.\n\nFor viral transduction, HMCs and HKC8 were seeded at 5\u2009\u00d7\u200910^4^ cells/well. The following day cells were transduced with a lentivirus carrying the human miR302a/b/c/d cluster (System Biosciences) or a scramble sequence (pGiPz). The media was changed 48\u2009h post transduction and replaced with DMEM-F12 (Sigma) supplemented with 10% v/v FBS, 10\u2009Mm L-glutamine, 100\u2009U/ml penicillin and 100\u2009\u03bcg/ml streptomycin. The successfully transduction was verified through GFP expression. Non-transduced HMCs and HKC8 maintained in DMEM-F12 were used as a positive control.\n\nThe treatment with the SB431542 (Calbiochem) compound was carried out at 7 days post transduction for either 7 days (day 14 post transduction) or 14 days (day 21 post transduction) in mir302-HMCs scramble infected HMCs and non-treated HMCs. The final concentration was 10\u2009\u03bcM, media was changed every other day, no splitting was performed during the inhibitor treatment. Protein and RNA were extracted at day 7 and 14 post treatment.\n\nDifferentiation protocol\n------------------------\n\n7 days after viral transduction miR302 and scrambled-infected HMCs were adapted to KnockOut\u2122 Serum Replacement (Thermo Fisher Scientific) media for 5 days. At day 13 post-transduction the cells were trypsinized and plated on matrigel in the following media: DMEM-F12, 20% KnockOut\u2122 Serum Replacement, 0.01\u2009mM non-essential amino acid (NEAA, Gibco), 100\u2009U/ml penicillin, 100\u2009\u03bcg/ml streptomycin, 0.05\u2009mM 2-mercaptoethanol (Sigma), 25-ng/ml bFGF (BioLegends) plus and minus 0.1\u2009\u03bcM DZNep. The cells were maintained in culture for up to a week.\n\nMatrigel coating\n----------------\n\nGrowth-factor reduced matrigel (BD Bioscience) was prepared by diluting the matrigel in a ratio of 1:10 with cold serum-free media. The matrigel was washed three times in the plate and then the plate was incubated with the lid off for 30\u2009minutes at room temperature inside the laminar hood. The plate was then incubated at 37\u2009\u00b0C for 1\u2009hour and washed one time with serum-free media before cells were seeded onto the plate.\n\nPicking and culturing of HMC-colonies\n-------------------------------------\n\nAfter their initial appearance, colonies morphologically resembling ES colonies were picked and subcultured. Briefly, individual colonies were subcultured on matrigel in ES cell medium comprising Murine Embryonic Fibroblasts (MEFs) conditioned media, 20% KnockOut\u2122 Serum Replacement, 0.01\u2009mM non-essential amino acid, 100\u2009U/ml penicillin, 100\u2009\u03bcg/ml streptomycin, 0.05\u2009mM 2-mercaptoethanol, 25-ng/ml bFGF. The colonies were cultured for up to 21 days.\n\nAnalysis of mRNA expression\n---------------------------\n\nTotal RNA was extracted from HMCs at days 3, 7, 14, 21 and 28 post-transduction according to the manufacturer' s instruction (TRIzol\u00ae, Life Technologies). cDNA was synthesized starting from 500\u2009ng of RNA as previously described[@b58]. Quantitative RealTime PCR was performed using TaqMan assays for T\u03b2RII, Snail and CDH1. Data was normalized using 18\u2009S as endogenous control.\n\nmicroRNA analysis\n-----------------\n\nTotal RNA was extracted from HMCs at day 7 post-transduction according to the manufacturer's instruction (TRIzol\u00ae, Life Technologies). cDNA was synthesized using TaqMan qRT-PCR assay for miR302\u2009d, Let-7c and Let-7a. Data was normalized using RNU6B as endogenous control.\n\nWestern Blot analysis and nuclear-cytoplasmic fractionation\n-----------------------------------------------------------\n\nTotal protein was isolated from HMCs at day 3, 7, 14, 21 and 28 using RIPA buffer (Tris-HCL, NaCl, NP-40, Na-deoxycholate, and Ethylenediaminetetraacetic acid (EDTA)) supplemented with 1\u2009mM NaF, 1\u2009mM Na~2~VO~4~, 50\u2009mM phenylmethylsulphonylfluoride (PMSF) and protease and phosphatase inhibitor cocktails (Sigma). Nuclear-cytoplasmic fractionation of HMCs was achieved by suspending the cells in an isotonic buffer containing 10\u2009mM HEPES, 1.5\u2009mM MgCl2, 10\u2009mM KCl, 0.5\u2009mM Dithiothreitol (DTT) and protease inhibitor cocktails, left on ice for 15\u2009min to separate the nuclei from the cytoplasm and then centrifuged for 1\u2009min at 12,000\u2009\u00d7\u2009g at 4\u2009\u00b0C. Nuclear pellets were then resuspended in a solution containing 20\u2009mM HEPES, 25% v/v Glycerol, 0.42\u2009M NaCl, 1.5\u2009mM MgCl2, 0.2\u2009mM EDTA, 05\u2009mM PMSF, 0.5\u2009mM DTT and protease inhibitor. The solution was then incubated for 30\u2009min on a shaker at 4\u2009\u00b0C and subsequently centrifuged. Total protein concentration was determined by Bradford Assay (Biorad). The following primary antibodies were used: Snail, Slug, T\u03b2RII, GAPDH, NF\u03baB, caspase 3, Smad3/P (all from Cell Signalling Technologies), E-Cadherin (BD pharmingen), EZH2 (kind gift from the Bracken Laboratory, Trinity College Dublin, Ireland), SUZ12 (Cell Signalling), p53 (Calbiochem), \u03b2-actin (Sigma), Histone 3 and H3K27 (all from Abcam).\n\nImmunofluorescence staining\n---------------------------\n\nHMCs were fixed with 4% paraformaldehyde for 10\u2009minutes, followed by 1\u2009h incubation at room temperature with rabbit polyclonal anti-ZO-1 (Invitrogen) primary antibody diluted in 5% Goat Serum (Sigma). Goat anti-rabbit Texas Red (Invitrogen) was used as secondary antibody and incubated for 1\u2009h at room temperature. Nuclei were counterstained with DAPI (4,6 diamidino-2-phenylindole, Sigma).\n\nProliferation and apoptosis assay\n---------------------------------\n\nAt day 7 and 21 post-transduction, HMCs were seeded onto a 96-well plate at a concentration of 500 cells/well. Non-transduced HMCs were seeded at the same concentration, as a control. 10\u2009\u03bcl of MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide, Sigma) diluted in ddH~2~O (5\u2009mg/ml) was added to each well, and the plate was incubated at 37\u2009\u00b0C for 4\u2009hours. Following incubation, 100\u2009\u03bcl of dimethyl sulphoxide (DMSO, Sigma) was added to each well. Optical Density (OD) was read at 570\u2009nm and 650\u2009nm with an ELISA reader (Molecular Devices). In one set of experiments cells were treated at the time of the seeding with 5\u2009\u03bcM DZnep.\n\nTo trigger apoptosis, HMCs were incubated with 50\u2009\u03bcM of etoposide (Merck) for 24\u2009h as follow. At day 7 post transduction miR302, scramble and non-transduced cells were seeded at confluency of 50%. The day after the cells were gently washed with PBS (Sigma) and fresh media was added along with 50\u2009\u03bcM of etoposide. The cells were lysed in RIPA buffer after 24\u2009h.\n\nGene Knock down using the esiRNA technology\n-------------------------------------------\n\nEndoribonuclease-prepared siRNA (esiRNA) were prepared as described by Heninger and Buchholz[@b59], with a few modifications. In brief, cDNA was isolated from NT2 cells and used as a template from which to generate T7 RNA polymerase promoter-flanked cDNA fragments, through two rounds of PCR amplification. These fragments represented the region of mRNA against which the esiRNAs would target genes of interest. These fragments were subsequently transcribed *in vitro* to generate double-stranded RNA (dsRNA) of the target amplicons, which were purified to isolate only larger (\\>250\u2009bp) RNA molecules. Following purification, dsRNA was digested using a Ribonuclease III (RNase III) mutant with a single amino acid substitution (E38A), which was produced in *Escherichia coli*[@b60] to generate a heterozygous population of \\~21\u2009nt siRNAs, capable of targeting multiple regions of target mRNA. Post-digestion, the siRNA fragments were purified using a QIAGEN RNeasy kit using high ethanol concentrations. See [table 1](#t1){ref-type=\"table\"} for a complete list of primers.\n\nmiR302, scramble (Firefly luciferase) and non-transduced HMCs as control, were transfected using Lipofectamine\u00ae 2000 (Invitrogen) according to manufacturer's instruction 7 days post viral transduction. 48\u2009h after transfection the cells were lysate for protein extraction.\n\nStatistical analysis\n--------------------\n\nThe reported values were represented as the mean\u2009\u00b1\u2009standard error of mean. One-way Anova and Bonferroni post-test were used to calculate the statistical significance for the analysis of more than two categories, Student's t-test was used in the charts were only 2 categories are presented (\\*P\u2009\\<\u20090.05, \\*\\*P\u2009\\<\u20090.01, \\*\\*\\*P\u2009\\<\u20090.001, \\*\\*\\*\\*P\u2009\\<\u20090.0001). All the analyses were performed with PRISM5.02 (GraphPad Software Inc, La Jolla CA, USA).\n\nAdditional Information\n======================\n\n**How to cite this article**: De Chiara, L. *et al*. miR302 regulates SNAI1 expression to control mesangial cell plasticity. *Sci. Rep.* **7**, 42407; doi: 10.1038/srep42407 (2017).\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Data\n\nThis study was supported by the Science Foundation Ireland (SFI), the Irish Research Council (IRC), and the Fighting Blindness Foundation.\n\nThe authors declare no competing financial interests.\n\n**Author Contributions** L.D.C. designed and performed experiments, analysed data and wrote the paper; D.A and A.W. performed experiments; G.O. helped in analysing and interpreting data; G.C. helped in manuscript evaluation and gave conceptual advices; J.C. supervised development of work, helped in data interpretation and edited the paper.\n\n![Overexpression of miR302a/b/c/d in Human Mesangial Cells silences the TGF\u03b2 type II receptor.\\\n(**A**) Representative images of GFP^+^ HMCs 7 days post transduction with miR302 (left panel) and scramble virus (right panel). Original magnification x100. 7 days post transduction, HMCs showed a marked upregulation of miR302d expression when compared to scramble infected cells (**B**). As a result of miR302 overexpression T\u03b2RII expression is decreased at mRNA (**C**) and protein (**D**) level, across all the analysed time points. Graph B and C and panel (**D**) are representative of 3 independent experiments. (SCR: scramble; NT: Non-treated cells; GFP: Green Fluorescent Protein; T\u03b2RII: TGF\u03b2 Receptor II; RQ: Relative Quantification normalized on 18\u2009S for T\u03b2RII or RNU6B for miR302d; \\*\\*P\u2009\\<\u20090.01).](srep42407-f1){#f1}\n\n![miR302 increases the expression of Snail during the acquisition of plasticity.\\\n(**A**) Demonstrates increased expression of Snail in miR302-HMCs compared to scramble HMCs. During the first week following miR302 overexpression in HMCs, Snail is upregulated at both mRNA (**A**) and protein level (**B**). Snail expression is not blocked by the addition to the culture media of the SB431545 (**C**), a potent inhibitor of TGF\u03b2 Receptor I (top panel) compared to standard culture conditions (bottom panel). Given the reciprocal nature of miR302 and Let-7 family expression and considering that Snail has been recently reported to bind and repress the promoter of the Let-7 family we analysed the level of expression of two of the components of this family, Let-7a and Let-7c, noticing no difference 7 days post transduction (**D**). All the figures are representative of 3 independent experiments. (\\*P\u2009\\<\u20090.05).](srep42407-f2){#f2}\n\n![miR302 upregulates EZH2 independently from Snail expression, whereas Smad3 and EZH2 demonstrate co-dependence.\\\nThe expression of the PRC2 components was assessed at days 7, 14 and 21 post transduction (**A**). No changes were observed in SUZ12 across all the time points whereas EZH2 follows the same pattern as both Snail and Slug, another member of the Snail family. Of note, the trimethylation of lysine 27 of histone H3, which is readout of the EZH2 activity, is unexpectedly turned off. EZH2, Snail and Smad3 were knocked down using esiRNA technology in miR302-HMCs, demonstrating that EZH2 and Snail are independently upregulated by miR302 overexpression in HMCs (**B**). In contrast the Smad3 knock down results in decreased expression of EZH2 suggesting cooperativity between EZH2 expression and Smad3 (**B**). (NT: non-transfected cells; H3K27me3: trimethylation of the lysine 27 on the histone H3). All panels are representative of 3 independent experiments.](srep42407-f3){#f3}\n\n![EZH2 expression correlates with increased proliferation of miR302-HMCs.\\\nmir302, scramble and non-transduced HMCs were seeded at a low confluency and an MTT assay was performed at 7 days (**A**) and 21 days (**B**) post transduction. Increased proliferation was observed 7 days post transduction correlating with the level of expression of EZH2 (**B**), while the proliferative advantage of miR302-HMCs is lost (**C**) when EZH2 expression decreases (**D**). This was confirmed by treating the cells with DZnep at the time of seeding (**E**); by ablating the activity of EZH2, the cells are no longer able to proliferate. All data are representative of 4 independent experiments performed separately. (DZnep: 3-Deazaneplanocin A, inhibitor of the EZH2 catalytic activity; \\*\\*P\u2009\\<\u20090.01; \\*\\*\\*P\u2009\\<\u20090.001).](srep42407-f4){#f4}\n\n![miR302-HMCs acquire plasticity without gaining a tumorigenic phenotype.\\\nTypical markers of cancer have been investigated in order to gain more insight into miR302-HMCs phenotype. NF\u03baB, a potent driver of carcinogenesis and inflammation, is dramatically downregulated in miR302-HMCs (**A**), while p16, a well-known tumour suppressor protein is maintained in miR302-HMCs (**B**). Despite the upregulation of Snail, no TGF\u03b2 increase is detected at any of the time points analysed (**C**). To test whether miR302-HMCs were protected against apoptosis, we treated the cells with 50\u2009\u03bcM of etoposide for 24\u2009h (**D**). No difference is detected in the expression of p53 and the cleaved form of caspase 3 in miR302, scramble and non-treated cells. (\\*P\u2009\\<\u20090.05). All panels are representative of 3 independent experiments.](srep42407-f5){#f5}\n\n![miR302-HMCs form ZO-1 containing junctions when cultured on matrigel.\\\n(**A**) Representative picture of normal, scramble and miR302 cells plated on matrigel. Original magnification x100. Mir302-HMCs show *de novo* expression of ZO-1 protein (**C**, white arrow), while they lose the expression of Snail, a marker of mesenchymal cells (**B**). (CTRL: control, HKC8s treated with TGF\u03b2 for 24\u2009hrs). Furthermore they show upregulation of CDH1 (E-cadherin) expression at mRNA level (**D**), in particular when treated with DZnep (CTRL: control, HKC8). Original magnification x630. (DAPI: 4\u2032,6-diamidino-2-phenylindole). (\\*\\*\\*P\u2009\\<\u20090.001).](srep42407-f6){#f6}\n\n![Working hypothesis.\\\nUpon miR302 overexpression, HMCs acquired an enhanced plasticity, becoming highly responsive to the surrounding microenvironment (plastic hybrid cells). During this process the cells upregulate the expression of both Snail and EZH2. At this critical tipping point if these cells are placed under the appropriate stimuli, they can acquire epithelial characteristics or, potentially, other kind of phenotypes (question marks). This very dynamic process can be potentially blocked or reversed.](srep42407-f7){#f7}\n\n###### Primers sequences for esiRNAs synthesis.\n\n  Gene name    Primer F (target gene only)   Primer R (target gene only)\n  ----------- ----------------------------- -----------------------------\n  SNAI1           TTTACCTTCCAGCAGCCCTA          CCAGGCTGAGGTATTCCTTG\n  EZH2            GAGGACGGCTTCCCAATAAC          GGAGCTGGAGCTATGATGCTA\n  FFlucGL3         CGGATTACCAGGGATTTC            CCTCAGAAACAGCTCTTC\n  SMAD3           ACAAGGTCCTCACCCAGATG          TGGACTGTGACATCCCAGAA\n"}
+{"text": "Background\n==========\n\nHuman immunodeficiency virus (HIV) reached all 31 provinces in China by 1998; the first indigenous Chinese acquired immunodeficiency syndrome (AIDS) case was detected in 1989 \\[[@B1]\\]. An assessment by the Chinese Ministry of Health, the Joint United Nations Program on HIV/AIDS and the World Health Organization suggested that approximately 740,000 were HIV infections in China at the end of 2009 \\[[@B2]\\], while the cumulative number of HIV infections actually diagnosed by the end of 2009 was 326,000 in China \\[[@B2]\\]. It was reported that 44.1% of HIV infections were not aware of their infection \\[[@B2]\\]. There were 13715 newly-diagnosed HIV infections in 2009 in China, 55.7% of which occurred through sexual transmission, and it was estimated that there were actually 48000 new HIV infections in China in 2009, 74.7% occurring through sexual transmission \\[[@B2]\\]. Heilongjiang Province, in northeast China, is still an area of low HIV/AIDS incidence and prevalence, but HIV infection has been rapidly spreading to the general population \\[[@B3]\\]. The incidence rate increased from 0.16 per 100,000 in 2004 to 0.33 per 100,000 in 2006, to 0.91 per 100,000 in 2009 in Heilongjiang Province. 66.5% people who were newly diagnosed and reported in 2009 acquired their infection through sexual transmission in Heilongjiang Province.\n\nTesting for HIV is an important component of an effective AIDS prevention and control campaign \\[[@B4],[@B5]\\]. Studies indicated that HIV-infected persons reduced high-risk behaviour substantially when they became aware of their infection \\[[@B6],[@B7]\\]. However, a major barrier to effective prevention and control efforts in China is that most infected people are not aware of their serostatus \\[[@B1],[@B8]\\]. Although willingness to undergo testing and taking part in testing are two different dimensions along a continuum of HIV prevention, willingness to be tested for HIV before taking concrete action is an important step in the HIV prevention process \\[[@B9]\\]. In China, studies about willingness to take HIV tests focused primarily on high-risk groups such as sex workers \\[[@B10]\\], rural-to-urban migrants \\[[@B11]\\], and commercial blood donors \\[[@B12]\\]. There have been few studies in the general population about willingness to take HIV tests. One study in Guizhou Province, China, investigated the acceptance of voluntary counselling and testing (VCT) in VCT centre among adults (18--45\u2009years) \\[[@B5]\\]. However, our study focused on the willingness to undergo testing if free HIV tests were provided at the nearest health care facility, which offered greater accessibility and convenience than a free HIV test in a distant VCT centre.\n\nThis study, in Heilongjiang Province, examined (a) the level of willingness of residents to participate in free HIV testing; (b) differences in the willingness between urban and rural residents; (c) the effect of demographic characteristics, HIV-related knowledge and stigma on willingness; and (d) the main reasons for participants' disinclination to be tested. The outcome of the study should help researchers to determine the feasibility of instituting HIV testing, and to identify factors and reasons for participants' disinclination to be tested.\n\nMethods\n=======\n\nParticipants and procedures\n---------------------------\n\nThe study population, aged 15--69\u2009years, was drawn from rural (September 2007) and urban areas (April 2008) using stratified sampling methods. First, two cities, Qiqihaer and Mudanjiang representing a middle socioeconomic level in urban areas (The population density is relatively high, usually more than 100,000 people; people mainly engage in non-agricultural industries), were selected from a total of seven cities. Then two communities of a middle socioeconomic level were selected respectively in the two cities, and all residents of selected communities were invited to participate. Fuyu and Dongning counties with a middle socioeconomic level in rural areas (The population density is relatively scarce; people mainly engage in agricultural production) was selected. Then four villages with a middle socioeconomic level were selected respectively in Fuyu and Dongning counties, and all residents of selected four villages were invited to participate. The residents were informed of the survey about their willingness to undergo free HIV testing at their nearest health care facility via broadcast media and newspapers. After providing informed consent, trained interviewers went door-to-door to invite people and participants complete an anonymous questionnaire (Additional file [1](#S1){ref-type=\"supplementary-material\"}) in a separate room at home. The interviewers provided assistance to some individuals with limited literacy by reading the questionnaire. It took about 15\u2009minutes for participants to complete the questionnaire. A total of 4050 rural and urban residents were approached.,Forty-eight declined to participate. Finally, 4002 were recruited and complete the survey, and the response rate was 98.8%. The Committee on Human Research of Harbin Medical University approved this study.\n\nMeasures\n--------\n\nIn addition to demographic information, the questionnaire consisted of four other sections: knowledge of HIV/AIDS (fifteen items) \\[[@B13],[@B14]\\], public attitudes regarding HIV/AIDS (three items) \\[[@B14]\\], willingness to participate in a free HIV test (one item), and reasons for not being willing to take a free HIV test (one item).\n\nHIV-related knowledge\n---------------------\n\nThe fifteen HIV-related knowledge questions included five items about HIV/AIDS transmission modes, six items about common HIV/AIDS transmission misconceptions and the other four items associated with willingness to participate in free HIV tests in previous studies \\[[@B15]\\], respectively are AIDS being a contagious disease, an apparently healthy person may being a carrier for HIV, not having a vaccine to protect against HIV, AIDS not being curable, with \"do not know\" responses scored as incorrect \\[[@B15],[@B16]\\]. The total score for HIV knowledge, knowledge of HIV transmission modes and HIV transmission misconceptions was the total number of correct responses.\n\nPublic stigma towards people living with HIV/AIDS (PLWHA)\n---------------------------------------------------------\n\nPublic stigma, as described by Corrigan and Watson, is the stigmatizing attitudes or reactions of the general population regarding persons with HIV and their family members \\[[@B17]\\]. Three items were adopted from previous research to reflect public stigma \\[[@B14]\\]. These items included whether participants were willing to work with PLWHA, accept family members with HIV/AIDS, or have their children study with PLWHA. The stigma towards PLWHA was scored according to the number of responses indicating stigma.\n\nWillingness to participate in free HIV testing\n----------------------------------------------\n\nWillingness to participate in free HIV testing was measured using responses to the question: \"If you were offered a free HIV test, with complete secrecy, would you wish to accept it?\" In this analysis, \"no\" and \"do not know\" responses were combined.\n\nStatistics analysis\n-------------------\n\nData were input using Epidata 3.02. Statistical analysis was performed using SAS Software 9.1. Differences among urban residents and rural residents were evaluated by the Student *t*-test or Chi-square test. Univariate log binomial regressions were performed to assess the relationship between each variable and willingness to participate in free HIV tests. Variables in univariate analysis with a significance level\u2009\\<\u20090.10 were included in the multivariate analysis. Multivariate log binomial regressions adjusted for basic demographic characteristics (i.e., age, gender) were performed to examine the association of HIV-related knowledge and willingness to participate in free HIV tests. *P*\u2009\\<\u20090.05 was considered statistically significant. We analysed the data separately for urban respondents and rural respondents, because there are large cultural and economic differences between urban areas and rural areas in China.\n\nResults\n=======\n\nThe 2018 urban respondents (mean age\u2009\u00b1\u2009standard deviation (SD): 39.1\u2009\u00b1\u200913.4\u2009years) consisted of 47.1% male and 52.9% female. The 1984 rural respondents (mean age: 39.6 \u00b113.5\u2009years) consisted of 49.4% male and 50.6% female. The majority of urban respondents and rural respondents were currently married or cohabitation (74.2%, 83.4%), Han ethnicity (82.6%, 78.7%), lower than 2000 RMB Yuan household per-capita monthly income (88.7%, 89.4%), and had lived in the city for more than two years (83.9%, 92.6%). There were large differences in education levels between urban respondents and rural respondents (P\u2009\\<\u20090.001). The vast majority of the urban respondents had completed middle school (90.6%), whereas just over half of the rural respondents had done so (53.9%). Demographic characteristics are listed in Table\u2009[1](#T1){ref-type=\"table\"}.\n\n###### \n\nDemographic profile of the study subjects\n\n  **Characteristic**                             **Urban residents**   **Rural residents**   ***P*value**         \n  --------------------------------------------- --------------------- --------------------- -------------- ------ ---\n  **Gender**                                            0.150                                                     \n  \u2003Male                                                  936                  47.1               950        49.4   \u00a0\n  \u2003Female                                               1053                  52.9               973        50.6   \u00a0\n  **Ethnicity**                                         0.002                                                     \n  \u2003Han                                                  1647                  82.6               1516       78.7   \u00a0\n  \u2003Minorities                                            346                  17.4               413        21.3   \u00a0\n  **Age (years)**                                       0.134                                                     \n  \u200315--20                                                138                   6.9               148        7.8    \u00a0\n  \u200321--50                                               1372                  68.5               1241       65.6   \u00a0\n  \u200351--69                                                492                  24.6               504        26.6   \u00a0\n  **Marital status**                                   \\<0.001                                                    \n  \u2003Single                                                415                  21.1               284        14.7   \u00a0\n  \u2003Married or cohabitating                              1461                  74.2               1612       83.4   \u00a0\n  \u2003Divorced or widowed                                   92                    4.7                36        1.9    \u00a0\n  **Education level**                                  \\<0.001                                                    \n  \u2003Primary school and lower                              187                   9.4               892        46.1   \u00a0\n  \u2003Middle School                                         571                  28.8               882        45.5   \u00a0\n  \u2003High school and more                                 1226                  61.8               163        8.4    \u00a0\n  **Employment**                                       \\<0.001                                                    \n  Unemployed                                             186                   9.4               268        13.8   \u00a0\n  Employed                                              1785                  90.6               1668       86.2   \u00a0\n  **Length of time lived in city**                     \\<0.001                                                    \n  \\< 2\u2009years                                             317                  16.1               142        7.4    \u00a0\n  \u2265 2\u2009years                                             1654                  83.9               1777       92.6   \u00a0\n  **Household per capita income (CYN/month)**           0.756                                                     \n  Low income level (\\<1000)                             1283                  65.6               1266       66.8   \u00a0\n  Medium income level (1000--2000)                       451                  23.1               429        22.6   \u00a0\n  High income level (\u22652000)                              221                  11.3               201        10.6   \u00a0\n\nPrimary school and lower: illiterate or semi-illiterate and elementary school;\n\nMiddle school: middle school;\n\nHigh school and more: high school, technical school, 3-Year college, undergraduate and above.\n\nHIV-related knowledge\n---------------------\n\nTwenty-eight (0.7%) subjects, who were rural residents, had not heard of HIV/AIDS. The majority of urban residents (82.4%) and rural residents (81.1%) were aware that HIV infection was a contagious disease. For urban residents, the mean scores for the total HIV-related knowledge, HIV/AIDS transmission modes and HIV/AIDS transmission misconceptions respectively was 10.8 (SD\u2009=\u20093.5), 3.8 (SD\u2009=\u20091.4) and 3.8 (SD\u2009=\u20091.9). However, for rural residents, the scores for the total HIV-related knowledge (8.9; SD\u2009=\u20093.8), HIV/AIDS transmission modes (3.5; SD\u2009=\u20091.6), transmission misconceptions (3.1; SD\u2009=\u20092.0) and awareness of respective HIV-related knowledge were all significantly lower than those among urban participants (all *P* \\<0.05), except for the knowledge that AIDS is a contagious disease. Awareness of five items about HIV/AIDS transmission modes were ranged from 56.6% to 86.8% among all participants, while knowledge of all six items about common HIV/AIDS transmission misconceptions ranged from 21.7% to 76.2%. Details are shown in Table\u2009[2](#T2){ref-type=\"table\"}.\n\n###### \n\nHIV/AIDS-related knowledge, Stigma towards people with HIV and willingness to participate in free HIV testing in the study subjects\n\n  **Knowledge items**                                                                          **Urban residents**   **Rural residents**   ***P*value**          \n  ------------------------------------------------------------------------------------------- --------------------- --------------------- -------------- ------- ---------\n  **Total knowledge score(mean, SD )**                                                                10.8                  (3.5)              8.9        (3.8)   \\<0.001\n  Is AIDS a contagious disease? (yes)                                                                 1644                  82.4               1599       81.1     0.285\n  Can an apparently healthy person be a carrier for HIV? (yes)                                        1479                  74.1               1174       59.6    \\<0.001\n  At present, is there a vaccine to protect against HIV? (no)                                         1085                  54.3               753        38.2    \\<0.001\n  At present, is AIDS curable? (no)                                                                   1393                  70.1               996        50.6    \\<0.001\n  **Mode of transmission score(mean, SD )**                                                            3.8                  (1.4)              3.5        (1.6)   \\<0.001\n  1\\. Can having only one uninfected sex partner reduce the risk of HIV transmission? (yes)           1221                  61.4               1112       56.6     0.001\n  2\\. Can use of condoms reduce the risk of transmission of HIV? (yes)                                1424                  71.3               1264       64.3    \\<0.001\n  3\\. Can sharing contaminated needles and syringes cause AIDS?                                       1609                  80.6               1522       77.3     0.011\n  4\\. Can receiving blood and blood products lead to AIDS? (yes)                                      1735                  86.8               1583       80.3    \\<0.001\n  5\\. Can a pregnant woman transmit HIV to her baby? (yes)                                            1624                  81.4               1406       71.4    \\<0.001\n  **Misconceptions of transmission score(mean, SD )**                                                  3.8                  (1.9)              3.1        (2.0)   \\<0.001\n  1\\. Can mosquito bites transmit HIV? (no)                                                            739                  37.0               427        21.7    \\<0.001\n  2\\. Can eating food together transmit HIV? (no)                                                     1390                  69.6               1084       55.1    \\<0.001\n  3\\. Can coughing or sneezing lead to AIDS? (no)                                                     1353                  67.8               1060       53.8    \\<0.001\n  4\\. Can shaking hands, hugging, or kissing cause AIDS? (no)                                         1470                  73.6               1276       64.7    \\<0.001\n  5\\. Can sharing the same office cause AIDS? (no)                                                    1523                  76.2               1236       62.7    \\<0.001\n  6\\. Can using the same toilet, shower, or swimming pool transmit HIV? (no)                          1097                  54.9               976        49.5    \\<0.001\n  **Stigma score(mean, SD )**                                                                          0.6                  (0.9)              0.8        (0.9)   \\<0.001\n  1\\. Is willing to work together with PLWHA?                                                          519                  26.3               650        33.8    \\<0.001\n  2\\. Is willing to accept family members with HIV/AIDS?                                               189                   9.6               192        10.0     0.706\n  3\\. Is willing to have their children study together with PLWHA?                                     454                  23.0               648        33.8    \\<0.001\n  **willingness to participate in free HIV testing**                                                  1438                  73.0               1516       78.8    \\<0.001\n\nPublic stigma towards PLWHA\n---------------------------\n\nOverall, 26.3% of urban respondents and 33.8% of rural respondents thought that PLWHA should be kept away from their colleagues (*P*\u2009\\<\u20090.001); 9.6% and 10.0%, respectively, were unwilling to accept family members with HIV/AIDS (*P*\u2009=\u20090.706); and 23.0% and 33.8%, respectively, would not allow their children to study with PLWHA (*P*\u2009\\<\u20090.001). Mean score on stigma towards PLWHA among urban participants was 0.8 (SD\u2009=\u20090.9), significantly higher than that among rural participants (0.6; SD\u2009=\u20090.9) (*P*\u2009\\<\u20090.001).\n\nWillingness to participate in free HIV tests\n--------------------------------------------\n\nThe proportions of participants who were willing to participate in free HIV testing among urban participants (73.0%) was significantly lower than that among rural participants (78.8%; *P*\u2009\\<\u20090.001). Univariate and multivariate log binomial regression analyses to identify factors associated with willingness to accept free HIV testing are shown in Table\u2009[3](#T3){ref-type=\"table\"}.\n\n###### \n\nUnivariate and multivariate analysis of factors associated with willingness to participate in free HIV testing\n\n  **Variable**                                             \u00a0                        **Urban residents**    **Rural residents**                              \n  -------------------------------------------------------- ------------------------ ---------------------- ------------------------ ----------------------- -----------------------\n  **Gender**                                               Male                     1.00                   1.00                     1.00                    1.00\n  \u00a0                                                        Female                   0.95(0.90--1.00)       0.99(0.94--1.05)         0.93(0.89--0.98) \\*\\*   0.94(0.90--0.98) \\*\\*\n  **Ethnicity**                                            Han                      1.00                   \u00a0                        1.00                    1.00\n  \u00a0                                                        Minorities               0.99(0.92--1.06)       \u00a0                        0.94 (0.88--1.00) \\*    0.97 (0.92--1.03)\n  **Age (Years)**                                          15--20                   1.00                   1.00                     1.00                    1.00\n  \u00a0                                                        21--50                   1.10(0.97--1.24)       1.13(0.99--1.28)         1.24(1.10--1.40) \\*\\*   1.14 (0.99--1.32)\n  \u00a0                                                        51--69                   1.04(0.91--1.18)       1.09(0.95--1.27)         1.21(1.06--1.37) \\*\\*   1.13 (0.98--1.32)\n  **Marital status**                                       Single                   1.00                   1.00                     1.00                    1.00\n  \u00a0                                                        Married/cohabitating     0.95(0.89--1.01)       0.94(0.88--1.01)         1.10(1.02--1.18) \\*     1.05(0.97--1.15)\n  \u00a0                                                        Divorced/widowed         0.87(0.74--1.02)       0.84(0.71--0.99) \\*      1.17 (1.00--1.37) \\*    1.07(0.91--1.27)\n  **Education**                                            Primary school or less   1.00                   \u00a0                        1.00                    \u00a0\n  \u00a0                                                        Middle school            0.94 (0.85--1.04)      0.89(0.81-0.89)\\*        1.02 (0.97--1.07)       0.99(0.95--1.04)\n  \u00a0                                                        Over high school         0.99 (0.91--1.09)      0.90(0.83-0.99)\\*        1.05 (0.97--1.14)       0.98(0.90--1.05)\n  **Employment**                                           Unemployed               1.00                   \u00a0                        1.00                    \u00a0\n  \u00a0                                                        Employed                 1.06 (0.96--1.17)      \u00a0                        0.94(0.88--1.00)        \u00a0\n  **Time in city**                                         \\< 2\u2009years               1.00                   \u00a0                        1.00                    \u00a0\n  \u00a0                                                        \u2265 2\u2009years                0.95 (0.89--1.02)      \u00a0                        0.95 (0.88--1.03)       \u00a0\n  **Household per-capita income (CNY/month)**              \\<1000                   1.00                   1.00                     1.00                    1.00\n  1000--2000                                               1.09(1.02--1.16) \\*\\*    1.07(1.01--1.14)\\*     1.00 (0.95--1.06)        0.98 (0.93--1.03)       \n  \u22652000                                                    1.03 (0.94--1.12)        1.01(0.92--1.10)       0.93 (0.85--1.01)        0.92 (0.85--0.99) \\*    \n  **HIV/AIDS Knowledge**                                   \u00a0                        \u00a0                      \u00a0                        \u00a0                       \n  Is AIDS a contagious disease?                            1.10(1.02--1.19) \\*      0.99(0.92--1.08)       1.11 (1.03--1.19) \\*\\*   1.02(0.95--1.10)        \n  Can an apparently healthy person be a carrier for HIV?   1.14(1.06--1.22) \\*\\*    1.12(1.03--1.21)\\*\\*   1.13 (1.08--1.19) \\*\\*   1.07(1.01--1.13) \\*     \n  At present, is there a vaccine to protect against HIV?   1.05 (0.99--1.10)        1.00(0.94--1.06)       1.10 (1.05--1.15) \\*\\*   0.99(0.94--1.04)        \n  At present, is AIDS curable?                             1.02 (0.96--1.08)        0.95(0.89--1.02)       1.11 (1.06--1.17) \\*\\*   1.02 (0.97--1.08)       \n  **Awareness of transmission modes score**                1.05(1.03--1.07) \\*\\*    1.02(1.00--1.05)       1.05(1.03--1.06)\\*\\*     1.03(1.01--1.06)\\*\\*    \n  **Awareness of transmission misconceptions score**       1.03(1.02--1.05) \\*\\*    1.02(1.00--1.04) \\*    1.03 (1.02--1.04) \\*\\*   1.01(0.99--1.02)        \n  **Stigma attitude score**                                1.00(0.97--1.03)         \u00a0                      1.02(1--1.05)            \u00a0                       \n  **Total Knowledge score**                                1.02(1.01--1.03) \\*\\*    \u00a0                      1.02(1.01--1.03) \\*\\*    \u00a0                       \n\nRR: Relative risk (95% CI); \\* P \\<0.05; \\*\\*P \\<0.01.\n\nCrude RR: Relative risk of univariate analysis.\n\nAdjusted RR: Basic demographic characteristics (i.e., age, gender) and variables that were associated with willingness to accept free HIV testing in univariate analysis among urban residents at the 0.10 significance level were put stepwise in a multivariate log binomial regression model.\n\nThe univariate analysis showed that greater total knowledge about HIV/AIDS was significantly associated with greater willingness among urban residents (RR\u2009=\u20091.02, 95%CI: 1.01--1.03, *P*\u2009\\<\u20090.001) and among rural residents (RR\u2009=\u20091.02, 95%CI: 1.01--1.03, *P*\u2009\\<\u20090.001). Stigma was not associated with willingness among urban participants (RR\u2009=\u20091.00, 95%CI: 0.97--1.03, P\u2009=\u20091.000), while there was a trend for an association with willingness among rural participants (RR\u2009=\u20091.02, 95%CI: 1.01--1.03, *P*\u2009\\<\u20090.067).\n\nResults of multivariate regression analysis among urban participants adjusted for age, gender, education, marital status, and income showed that a medium income level (RR\u2009=\u20091.07, 95%CI: 1.01--1.14, *P*\u2009=\u20090.021), having a greater knowledge of HIV transmission misconceptions (RR\u2009=\u20091.02, 95%CI: 1.00--1.04, *P*\u2009=\u20090.021) and being aware that an apparently healthy person can be a carrier for HIV (RR\u2009=\u20091.12, 95%CI: 1.03--1.21, *P*\u2009=\u20090.007) was significantly associated with greater willingness to accept a free HIV test (Table\u2009[3](#T3){ref-type=\"table\"}). However, a higher education level (middle school vs. primary school or less: RR\u2009=\u20090.89, 95%CI: 0.81--0.98, *P*\u2009=\u20090.017; high school vs. primary school or less: RR\u2009=\u20090.90, 95%CI: 0.83--0.99 *P*\u2009=\u20090.032), and being divorced or widowed (RR\u2009=\u20090.84, 95%CI: 0.71--0.99, *P*\u2009=\u20090.042) were significantly associated with a decreased willingness to accept a free HIV test (Table\u2009[3](#T3){ref-type=\"table\"}). Multivariate regression analysis among rural participants adjusted for age, gender, education, ethnicity, marital status, and income showed that having a greater knowledge of HIV transmission modes (RR\u2009=\u20091.03, 95%CI: 1.01--1.06, *P*\u2009=\u20090.013), and being aware that an apparently healthy person could be a carrier for HIV (RR\u2009=\u20091.07; 95%CI: 1.01--1.13 *P*\u2009=\u20090.019) were significantly associated with greater willingness to participate in HIV testing. However, being female (RR\u2009=\u20090.94, 95%CI:16 0.90--0.98 P\u2009=\u20090.006) and high income level (RR\u2009=\u20090.92, 95%CI: 0.85--0.99 P\u2009=\u20090.034) were significantly associated with lower willingness to participate in HIV testing.\n\nThe main reasons for not being willing to accept free HIV tests among urban participants included believing that they were unlikely to have been exposed to HIV (38.7%), not having enough time (19.7%), fear of stigma (14.7%), and not wanting to know the result (11.1%). Reasons were similar among rural participants: 58.3%, 17.6%, 17.0%, and 15.0%, respectively. Other reasons given were \"do not trust the result to be confidential,\" \"old age\" (i.e., believed oneself too old to be important to health), and \"worry about infection in the course of testing.\"\n\nDiscussion\n==========\n\nThe results of our study showed that the willingness of urban residents (73.0%) to undergo HIV testing was significantly lower than that among rural residents (78.8%). Willingness to participate in free HIV testing were slightly lower than those found in studies carried out among high-risk populations in China (from 78.0% to 94.0%) \\[[@B12],[@B18]-[@B21]\\]. Willingness to undergo HIV testing in this study (78.8% rural; 73.0% urban) was greater than that of participants in the only published study of the general public, aged 18--45\u2009years, carried out between October 2005 and February 2006 in Guizhou Province, China (43.5%) \\[[@B22]\\]. The difference may be due to time trends. With the further promotion of HIV knowledge and relevant policy, people may be more willing to accept a free HIV test. Simultaneously, the difference may also be due to the possibility of undergoing HIV testing in the nearest health care facility, which offers greater accessibility and convenience, compared with accepting a free HIV test in a VCT centre in the study in Guizhou Province. There are relatively few VCT centres in China, especially in rural areas, and they are often at a distance from the area of residence, thus visiting a VCT centre incurs costs in terms of travel and time. A study of Chinese sex workers found that participants were more willing to be tested near their workplaces \\[[@B23]\\]. In September 2006, to reduce the number of persons with undiagnosed HIV infection, the United States Centers for Disease Control issued recommendations to implement HIV screening as part of routine medical care for all persons aged 13--64\u2009years in all health care settings, based on ensured privacy and confidentiality and referral to prevention and clinical services as necessary \\[[@B24]\\]. However, further studies are needed to find suitable protocols and sites of HIV testing among high-risk groups and the general population, such as combining HIV testing with other health examinations, and integrating VCT services into existing health centres in China.\n\nHIV-related knowledge awareness was lower than that found in studies among high-risk populations in Heilongjiang Province \\[[@B25],[@B26]\\]. In our study, overall HIV-related knowledge awareness among rural residents was lower than that among urban residents. Similar results were found in studies in 2005 in Hubei Province and Liaoning Province, China \\[[@B27],[@B28]\\]. This indicates a need for further education about HIV in rural areas in China. Meanwhile, While 45.0% of urban respondents and 38.0% of rural respondents could correctly identify all five items regarding important modes of HIV/AIDS transmission, only 19.8% of urban respondents and 9.3% of rural respondents could identify all six common HIV/AIDS transmission misconceptions (*P*\u2009\\<\u20090.001). Similar results were found in the study in southwest China in 2003 \\[[@B29]\\] and in Xinjiang in 2006 \\[[@B30]\\]. Low awareness about HIV/AIDS transmission misconceptions highlights the need for further education about these misconceptions.\n\nUnivariate analysis among both urban and rural residents showed that greater total HIV/AIDS knowledge was significantly associated with greater willingness to participate in free HIV testing, which suggests that the wider and more accurate an individual's HIV-related knowledge, the more willing he or she will be to accept a free HIV test. In our study, stigma was not associated with willingness to undergo testing in urban areas, while there was a trend toward an association in rural areas. Liu's path analysis in a rural area of China found that stigmatizing attitudes were only indirectly associated with intention to disclose HIV serostatus through perceived stigma \\[[@B31]\\]. Further study is needed to document the relationship between willingness to participate in free HIV tests and other stigma indicators.\n\nMultivariate regression analysis of willingness to participate in HIV testing among both urban and rural residents showed that the awareness that an apparently healthy person can be a carrier of HIV was significantly associated with greater willingness to participate in HIV testing. This finding is consistent with Sarker's study in Burkina Faso \\[[@B32]\\] and Gage's study in Uganda \\[[@B33]\\]. There was no significant difference between male and female among urban residents. However, males were more willing to participate in free tests than females among rural residents, similar to results found in the study by Liu among rural residents in China \\[[@B20]\\]. It is possible that rural females have less access to health information in China. Urban residents with lower education level were more willing to participate in free HIV testing. This result is consistent with the findings among the general population in the United States \\[[@B34]\\]. Urban residents with better knowledge of HIV transmission misconceptions were more willing to participate in free HIV testing. Possible explanations are that misconceptions about modes of HIV transmission could heighten the fear of infecting another person or being infected via daily life contact, thus reducing willingness to participate HIV testing. In contrast, rural residents who had better knowledge of HIV transmission modes were more willing to participate in free HIV testing. It may be that knowledge of HIV transmission modes strengthens the perception of infection risk, thus increasing willingness to participate in HIV testing. We conclude that HIV/AIDS education needs to be improved to increase willingness to undergo HIV testing among the general population in Heilongjiang, with particular emphasis on HIV transmission modes in rural areas, and on dispelling misconceptions about HIV transmission modes in urban areas.\n\nThe most common reason for unwillingness to accept free HIV tests was participants' belief that they had not personally been exposed to HIV. The results suggest that when people are conscious of their risk for HIV, they are more willing to participate in HIV testing. Another reason was that respondents were not willing to know their test results. This suggests that the respondents thought they could not benefit from awareness of their serostatus. Therefore, programs to improve knowledge about HIV treatment should be carried out, so that people know that early and compliant treatment can improve the quality and length of HIV carriers' and AIDS patients' lives, and mother-to-child transmission of HIV can be prevented \\[[@B35]-[@B38]\\]. Efforts should also be promoted to inform people about Chinese HIV-related policies, such as the \"Four Free and One Care\" policy \\[[@B39]\\]. The policy includes free antiretroviral drugs for those who cannot afford to pay, testing, prevention of mother-to-child transmission, and schooling of orphans \\[[@B40]\\].\n\nThe study has some limitations. First, although the study results indicates willingness to test in the general population are high, not all who are willing to test for HIV will actually do so over time. Therefore, further studies as to the circumstances under which residents will actually accept testing are needed (e.g., education about HIV-related knowledge and policies, suitable styles and sites of HIV testing, privacy and confidentiality protection, and clinical services). Secondly, our questionnaire might not reflect all aspects of stigma or bias, as we only studied the association of public stigmatising attitudes and willingness to participate in HIV testing. Thirdly, as the sample was not a perfect random sample, some selection bias may exist. another limitation of the analysis are non-response bias, however the relatively high response rate (98.8%) might minimise this bias. Finally, owing to the cross sectional nature of this study, these data should be interpreted as associations rather than implying causality. Despite the limitations, our findings provide valuable information for HIV test. The findings highlight that education on HIV/AIDS needs to be improved to increase willingness for HIV testing among the general public, especially emphasising knowledge of HIV transmission modes in rural areas and countering misconceptions of HIV transmission modes in urban areas.\n\nConclusions\n===========\n\nOur results indicate the willingness of general residents in Heilongjiang, Northeast China, to participate in free HIV testing in the nearest health care setting is high, and is higher in rural residents than in urban residents. However, further studies are needed to find suitable protocols and sites of HIV testing among high-risk groups and the general population. Our results indicate that knowledge of HIV transmission misconceptions are associated with willingness to accept free HIV testing among urban residents, while knowledge of HIV transmission modes are associated with willingness to accept free HIV testing among rural residents. Therefore, we believe that HIV/AIDS education needs to be improved to increase willingness to undergo HIV testing among the general population in Heilongjiang, with particular emphasis on HIV transmission modes in rural areas, and on dispelling misconceptions about HIV transmission modes in urban areas.\n\nCompeting interests\n===================\n\nWe declare no competing interests.\n\nAuthors' contributions\n======================\n\nYuan Lili was responsible for data acquisition, analysis, interpretation, drafting and production of the final manuscript. Li Xin was responsible for study design, data acquisition, analysis, interpretation. Shi Jingli was responsible for study design, date acquisition. Jiang Liying was responsible for date acquisition, analysis, interpretation. Zhang Chundi was responsible for study design, acquisition of date. Yang Xiujing was responsible for study design, acquisition of date. Zhao Yashuang was responsible for study design, analysis, interpretation, drafting and production of the final manuscript. All authors read and approved the final manuscript.\n\nFunding\n=======\n\nThis study was supported by Health Department of Heilongjiang Province, China (2006--308).\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2334/12/256/prepub>\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\nHealth Survey Questionnaire.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis study was supported by Health Department of Heilongjiang Province, China (2006--308). We would like to thank colleagues and graduates from Qiqihaer Medical College and Mudanjiang Medical College in Heilongjiang Province for their participation in data collection.\n"}
+{"text": "1 INTRODUCTION {#SEC1}\n==============\n\nUnderstanding the association between protein occupancy and the target gene expression is essential to study the mechanism of gene regulation. The first step is to identify the protein-binding sites in the genome. Current chromatin immunoprecipitation (ChIP) technologies coupled with microarray hybridization (ChIP-chip) or parallel DNA sequencing (ChIP-seq) enable the identification of transcription factor-binding sites *in vivo* in the genome-wide scale. Many computational methods have been developed to detect transcription factor occupancy from ChIP-chip and ChIP-seq data, which we refer to as peak calling. Making choices of program parameters and choosing significance thresholds to accurately control genome-wide false positives (FPs), however, is often difficult. We have previously developed a statistical method (Zhang, [@B17]) that can precisely control the expected number of genome-wide FP peak calls in the context of correlated multiple comparisons. With FPs controlled, it is further desired to improve the peak calling methods to reduce false negative instances.\n\nGene regulation is a complex process that usually involves the cooperation of multiple transcription factors, which may interact to form a regulatory module that binds to a DNA segment to regulate their target gene\\'s expression. Certain histone modifications may also play crucial roles in regulatory mechanisms (Heintzman *et al.*, [@B13]; Muller *et al.*, [@B14]). The binding potential of a target transcription factor, therefore, can be partially learned from the co-binding proteins and features that jointly participate in a regulatory module within the cell. As the prevalent ChIP technologies increase the need for analyzing ChIP-chip and ChIP-seq data, several studies have made efforts to utilize multiple biological features into ChIP data analysis. For example, methods have been developed for segmenting the genomic regions into active intervals of interest (Day *et al.*, [@B10]; Du *et al.*, [@B11]). Few methods for peak calling, however, have incorporated the joint effects of related biological features while detecting binding sites for a target protein. Datta and Zhao ([@B9]) have proposed a method that uses a log-linear model to infer co-binding associations between two or more transcription factors. Their method is a post-processing algorithm that takes the *P*-values from existing peak calling algorithms as input, but cannot incorporate general types of data.\n\nIn this article, we propose a novel yet rigorous method that accounts for the co-binding information and relevant biological features to detect DNA occupancy of a target protein from ChIP data. Without assuming distributions of the related biological features, which we call supporting tracks, we first use a logistic regression model to describe the correlation between the binding of the target protein and the supporting tracks. The output is the probability of each position being potentially occupied by the target protein, predicted by the supporting tracks. We then introduce a varying threshold method to call significant peaks from the ChIP data of interest, where the threshold for each probe is adjusted by its predicted probability of protein binding. Our approach is similar to that of a Bayesian method that incorporates prior knowledge of protein binding into the analysis. Different from Bayesian methods, we still control the family wise FP rate, or false discovery rate (FDR) (Benjamini and Hochberg, [@B2]), at a user-specified level. Our varying threshold method, called PASS2, is a generalization from the PASS algorithm (Zhang, [@B17]), and is related with conditional test in statistics (Cox and Hinkley, [@B8]).\n\nUsing simulation studies and real datasets of GATA1 binding in a mouse erythroid regulation study (Cheng *et al.*, [@B7]), we show that the proposed method can identify many more GATA1-binding sites than using the target ChIP data alone, when the FDR is controlled at a common level. Our study shows that the proposed framework is robust with respect to irrelevant supporting data added to the model. The additional binding sites detected by incorporating the related biological features are potentially real GATA1-binding sites, many of which are either experimentally verified or enriched near RefSeq Genes (Pruitt *et al.*, [@B15]). To our best knowledge, the proposed method is the only algorithm that incorporates multiple sources of information in peak calling to improve the power of detecting weak protein-binding signals, and simultaneously, our method controls a user-specified FP level adjusting for millions of correlated comparisons.\n\n2 METHODS {#SEC2}\n=========\n\nWe assume the ChIP data of interest and the supporting tracks are generated by independent experiments. We first map all tracks of data onto a common coordinate represented by tiling probes. Here, we use the term 'probe' to represent a short genomic interval (\u2264100 bp) that corresponds to the probes used in ChIP-chip experiments. A probe for ChIP-seq experiments and for other types of data can be arbitrarily defined. To convert each track of data into probe statistics, we calculate a standardized value (*t*~*i*~) at each probe *i* by taking the average value of the original data (*x*~*j*~) within the probe interval, and dividing the average value by its SD. The SD is calculated as the SD (\u03c3) of all data in the track divided by the square root of the number of data points (*w*) within the probe interval:\n\nFor computational efficiency and model robustness, we further apply discretization methods to convert the continuous values of supporting tracks into ordinal bins. We then apply a logistic regression model to compute the binding probability of each probe being occupied by the target protein, using the binned supporting data as predictors and a list of known or highly probable binary binding sites of the protein as the response. We use permutation to evaluate the relevance of supporting tracks, and we discard insignificant tracks. We finally calculate a probe-specific threshold for each probe according to the calculated binding probabilities. We use the probe-specific thresholds to call significant peaks from the ChIP data of interest. A flow chart of our method is shown in [Figure 1](#F1){ref-type=\"fig\"}. Fig. 1.Flow chart of the proposed method.\n\n2.1 Discretization methods {#SEC2.1}\n--------------------------\n\nTo efficiently represent a large number of unique values generated by the genome-wide arrays, we categorized continuous values of supporting tracks into bins to reduce the size of data matrix. Binning data also improves the robustness of binding site prediction by reducing the effects of extreme values in the supporting data. We applied different discretization techniques and compared their effects on the performance of peak calling. Unsupervised discretization methods such as equal-width or equal-frequency methods, and clustering algorithms, require a specified number of bins for discretization. We evaluated the performance of unsupervised methods using different number of bins. We also applied an entropy-based discretization method that utilizes the known binding sites in a supervised manner to determine an optimal number of bins and assignment of bins based on information content maximization (Fayyad and Irani, [@B12]).\n\n### 2.1.1 Equal-width and equal-freq method {#SEC2.1.1}\n\nlet *k* denote the number of bins, the 'Equal-Width' method partitions the range of continuous values into *k* intervals of equal width. The 'Equal-Freq' method, on the other hand, assigns an approximately equal number of continuous values in each bin.\n\n### 2.1.2 Clustering method {#SEC2.1.2}\n\na *k*-means clustering algorithm is used to assign all continuous values into *k* bins. For *k* = 2, the minimum and the maximum values in a supporting track are used as the cluster centroids. For *k* \\> 2, the initial *k* cluster centroids are the values, including the minimum and maximum values that partition the data into (*k* \u2212 1) bins of equal width.\n\n### 2.1.3 Entropy method {#SEC2.1.3}\n\ngiven a sorted array of continuous values *S* and a corresponding array of binding status of the target protein, the method finds a best cut point *T* that partitions the range of *S* into two non-overlapping intervals. A cut point *T* is the midpoint between two contiguous data points in the sorted array. For each candidate cut point *T*, the data are divided into two subsets on each side of T. The class entropy of a subset *S*~*j*~, where 'class' refers to the binding status, is defined as here, *c* denotes the number of classes (*c* = 2), and *p*~*i*~ denotes the proportion of data points in *S*~*j*~ that belong to class *i*. The entropy of a bi-partition of *S* at cut point *T* is then defined as the weighted average of the class entropies of subset *S*~1~ and *S*~2~: where \\|*S*\\| denotes the size of array *S*. The partition that minimizes Ent(*S*, *T*) over all cut points *T* is selected as the best partition.\n\nDefine the information gain of a partition at cut point *T* as Our goal is to find a best split of data that maximizes the information gain. The algorithm is applied recursively within each partitioned interval to find sub-partitions, until some stopping criteria are satisfied. The algorithm uses the minimum description length (MDL) as a criterion for accepting or rejecting a given partition. The partition by a cut point *T* is accepted if and only if Gain(*S*\u2032, *T*) \\> MDL(*S*\u2032), where MDL(*S*\u2032) is calculated as here, *S*~1~\u2032 and *S*~2~\u2032 denote a best bipartition of *S*\u2032, and *k*, *k*~1~, *k*~2~ (= 1 or 2) denote the number of distinct classes in *S*\u2032, *S*~1~ and *S*~2~, respectively. The algorithm stops when no more partitions satisfying the MDL constraints can be found.\n\nThe supporting tracks were binned individually while running the above four discretization methods. To use the entropy-based discretization, each track of data was sorted, and a list of known or predicted binding sites was mapped to the probe coordinate and used as the class label *c* of each data point. We only evaluated the cut points at the boundary between two classes (boundary of binding sites), because the cut point *T* that minimizes the average class entropy Ent(*S*, *T*) is always a value between two data points of different classes in a sorted array (Fayyad and Irani, [@B12]).\n\n2.2 Prediction of potential binding {#SEC2.2}\n-----------------------------------\n\nTo combine information from the supporting tracks and predict the binding locations of a target protein, we fit a logistic regression model between a vector of binary indicators *Y* = (*Y*~1~, *Y*~2~,..., *Y*~*n*~) denoting the known binding status of the target protein at *n* probes, and a data matrix *X* of *m* supporting tracks. Each column of *X*, denoted by *X*~*i*~ = (*X*~*i*1~, *X*~*i*2~,..., *X*~*in*~)\\', contains the converted bin values of the *i*-th supporting track at *n* probes, for *I* = 1,..., *m*. The vector *Y* can be constructed from experimentally verified binding events, previous studies of the target protein occupancy or computationally detected sites from the current ChIP data at a stringent threshold. When computationally detecting binding sites as the responses in training data, a stringent FDR should be controlled so that the fitted regression model will not be strongly biased towards FPs. We suggest using a FDR no larger than the FDR allowed at the end of peak-calling. For example, if 10% FDR is allowed at final peak calling, then the initial peak-calling for training should be controlled at 10% FDR. The rationale is that, even if training peaks are FPs, they are still allowed at final peak calling.\n\nWe model the binding events *Y* of the target protein at *n* probes as independent Bernoulli random events with parameter \u03c0, where \u03c0 denotes an *n*-dim vector of binding probabilities. We model \u03c0 as a function of supporting tracks *X* via a logistic link function as and hence where \u03b2~0~ denotes the baseline binding coefficient and \u03b2~*i*~ denotes the additional effect of binding contributed by the supporting data *X*~*i*~.\n\nThe parameters \u03b2 = (\u03b2~0~,..., \u03b2~*m*~) were estimated by maximum likelihood estimation. Since the data points are generally correlated, we used permutation test to evaluate the empirical significance of the effect of each supporting track *X*~*i*~, and we removed insignificant tracks from the model at 0.01 significance level.\n\n2.3 Varying thresholds for peak calling {#SEC2.3}\n---------------------------------------\n\nGiven the effects \u03b2 of supporting tracks, we can calculate a vector \u03c0 of binding probabilities of the target protein at all probes in the ChIP data. We use the same notation \u03c0 to denote the binding probability of probes in the ChIP data, for which significant peaks are to be called, where the \u03c0 used in the previous section corresponds to the training probes that are used to fit the logistic regression model.\n\nWe generalized the original PASS algorithm (Zhang, [@B17]) to utilize the binding information predicted by the supporting tracks. By default, we assume that the probe values in the ChIP data have been standardized to follow a normal distribution at unbound regions. A probe (or a window of probes) is called statistically significant if its statistic (or the average statistic of a window) is above a threshold. Intuitively, if a probe has a higher probability to be occupied, we may reduce its peak-calling threshold so that the probe is easier to be called and hence reduce the chance of false negatives. The key is to choose the probe-specific thresholds according to the predicted binding probabilities of probes, and at the same time control the overall number of FPs at a user-specified level.\n\nThe difficulty of controlling the overall FP rate in peak calling arises from the fact that the probe values are locally strongly correlated. The original PASS algorithm (Zhang, [@B17]) uses a de-clumping method to compensate the positive correlations among probes, such that the total number of FPs after de-clumping can be simply computed as a summation of the FP rate of individual probes. We slightly modified the de-clumping method as follows: we now call a probe significant if and only if its statistic is above a threshold and is the maximum among probes within its neighborhood. As a result, for all probes within any local interval, at most one probe can be called significant (assuming no ties) under this rule, and hence the positive correlation among probes is compensated. All theoretical treatments stated in the original PASS paper (Zhang, [@B17]) still hold true under this new de-clumping scheme. Ignoring local interference of negative correlations created by de-clumping, and based on the fact that a probe being significant by chance is rare, we can approximate the expected total number of FPs in the entire ChIP data by summing over the FP rate of individual probes. It further holds true that the family wise FP rate of peak calling, with or without de-clumping, remain unchanged (Zhang, [@B17]), and hence our method can control both family wise FP rate and FDR.\n\nTo utilize the predicted binding probabilities, and to control the overall number of FPs at a desired level \u03bb, we choose probe-specific thresholds as follows. For each probe *i*, we calculate a threshold *t*~*i*~ such that probe *i* has probability \u03b1~*i*~ = min(1, \u03bb\u03c0~*i*~/\\|\u03c0\\|) to be a FP by chance after de-clumping, where \\|\u03c0\\| denotes to the summation of elements in \u03c0. This can be done by calculating the significance of a range of thresholds at probe *i*, using the importance-sampling algorithm proposed in PASS (Zhang, [@B17]). We then use interpolation to compute *t*~*i*~ that yields \u03b1~*i*~.\n\nTo use the varying thresholds *t*~*i*~ in peak calling, we call a probe significant (occupied by the protein) if its test statistics is greater than or equal to *t*~*i*~ and is the maximum within its neighborhood (500 bp by default). The expected total number of FPs of all probes can therefore be approximated as\n\nAs a result, we can control the overall number of FP calls at a user-defined level \u03bb. At the same time, we gain a substantial amount of power in detecting genuine protein-binding sites by using a liberal threshold at probes that are likely to be occupied, as suggested by the supporting data. To control family wise FP rate at level \u03b1 ~FWER~, we let \u03bb = \u2212log(1 \u2212 \u03b1~FWER~) and calculate varying thresholds with respect to \u03bb at individual probes. To control FDR at \u03b1 level, we use a step-down approach as follows: (i) start at \u03bb = \u03b1, we calculate the varying thresholds and report all peaks passing the thresholds; (ii) we increase \u03bb to \u03bb = *k*\u03b1, where *k* denotes the total number of peaks detected in previous iterations, and we recalculate the varying thresholds and call more peaks; and (iii) we repeat step 2 until no more peaks can be found.\n\n2.4 ChIP data and supporting data {#SEC2.4}\n---------------------------------\n\nThe target ChIP data and the supporting data are generated from the same mouse erythroid cell line with restoration of GATA1 function (G1E-ER4). We applied our method to two real studies, one ChIP-chip data and one ChIP-seq data, to detect GATA1 occupancy in the mouse genome. GATA1 is a transcription factor that regulates erythroid genes. The ChIP-chip data was obtained from a hybridized GATA1 ChIP sample to the NimbleGen HD2 tiling array for the mouse genome (mm8 assembly). The ChIP-seq data was obtained from a different GATA1 biological sample sequenced by Illumina GAII technology, with 23 million 36 bp reads uniquely mapped to the mouse genome (mm8 assembly) (Cheng *et al.*, [@B7]).\n\nWe further obtained four additional datasets that are related with GATA1 binding and are used as supporting tracks: (i) DNase-seq open chromatin data from F-Seq (Boyle *et al.*, [@B4], [@B5]); (ii) ChIP-seq data of TAL1 (also known as SCL) that often co-binds with GATA1, LDB1 and LMO2 to form a multi-protein complex (Wadman *et al.*, [@B16]); (iii) ChIP-seq data of trimethylation of lysine 4 of histone H3 (H3K4me3) associated with active promoters (Heintzman *et al.*, [@B13]); and (iv) ChIP-seq data of trimethylation of histone H3K27 (H3k27me3) associated with down-regulation (Muller *et al.*, [@B14]). The data tracks used in this study contained 883 758 probes in a previously reported 66 Mb region on mouse chromosome 7. This region contains a large number of experimentally verified GATA1-binding sites (Cheng *et al.*, [@B6]), and hence servers as a good example to demonstrate our method.\n\nThe ChIP-seq data contained discrete counts of short reads mapped to consecutive positions in the genome. To apply our method, we converted the GATA1 ChIP-seq read counts to probe statistics as follows: (i) calculate the sum of read counts according to a defined probe coordinate; (ii) with top 1% probes of large read counts removed, model the background distribution of read counts by a negative binomial distribution, NB(*r*, *p*), and estimate the two parameters *p* and *r* by maximum likelihood estimation; (iii) compute the *P*-value of each probe from the assumed background distribution and converted the *P*-values to *Z*-scores. Our method allows the user to specify the probe size and 'tiling' resolution. In general, larger distance between probes (e.g. 100 bp) may lose ChIP-seq signals and therefore reduce the power of peak detection. Conversely, smaller distance between probes (e.g. 1 bp) will not compromise mapping resolution, but could be computationally intensive. Probe size should not be too small so that it contains enough tag counts. By default, we use a probe size of 30 bp tiled at every 10 bp distance for ChIP-seq data, and we recommend using a window of 2--5 probes to call peaks. Users can change these values in our program.\n\n3 RESULTS {#SEC3}\n=========\n\n3.1 Simulation study {#SEC3.1}\n--------------------\n\nWe first performed simulation study to evaluate the power and the robustness of our method. We generated random ChIP-chip data containing 883 758 probes, among which 300 randomly selected probes corresponded to protein-binding sites. For probes at the binding and non-binding sites, we simulated the probe intensities from a normal distribution with mean 8 and 0, respectively, and variance 1. To further introduce correlation among probes, each probe value was replaced by a weighted average within its 500 bp window, calculated as where *w* = 1 when *i* = *j*, *w* = 0.8 when *d*(*i*, *j*) \u2264 125 bp and *w* = 0.4 when *d*(*i*, *j*) \\> 125 bp.\n\nThe signals in the simulated data ranged from \u22125.28 to 8.13 with mean 0.0046 and variance 1.02. This is comparable with the ChIP-chip HD2 data (ranged from \u221211.77 to 8.29 after normalization). Since the simulated binding sites were randomly placed, they were independent with the four supporting tracks. We therefore can evaluate the impact of using irrelevant supporting tracks in peak calling. In particular, we did not remove the unrelated supporting tracks when calculating the binding probabilities, and we checked whether using the irrelevant information can increase the number of FPs by our method. We used the 'Equal-Freq' method to discretize each supporting data into *k* = 7 bins, and we called significant peaks at 10% FDR level. We further compared our method with an existing method, MPeak (Zheng *et al.*, [@B19], trimming *P*-value set at 1e-05) on the same datasets.\n\nAs shown in [Figure 2](#F2){ref-type=\"fig\"}, the estimated FDR (number of FPs/number of detected peaks) level of our method was accurately controlled at the specified 10% level, and the FDR level remained invariant before and after incorporating irrelevant supporting information. In each simulated dataset, our method detected an average of 240 (out of 300) true peaks. As we expected, the number of peaks detected before and after using irrelevant data remained almost unchanged (first column in [Fig. 3](#F3){ref-type=\"fig\"}). This result indicates that our method can accurately control FPs with and without using additional information, and the method is robust to irrelevant data added to the model. In comparison, MPeak reported an average of 210 peaks per dataset, among which only 135 were true peaks, pertaining to \u223c36% FPs. Fig. 2.Comparison of FDRs in 10 simulated datasets. Our method with and without using supporting tracks (prior) is controlled at 10% FDR level. We cannot specify FDR level for MPeak. Fig. 3.(**A**) FDR and (**B**) detection power comparison of our method before (black) and after (colored) using supporting tracks at different levels of concordance. Box-plot of 10 datasets at each concordance level is shown.\n\nWe next evaluated the power of our method by introducing correlation between the binding sites and the supporting data. Out of 300 binding sites in each simulation study, *x*% binding sites were randomly placed at locations of previously identified GATA1-binding sites from real GATA1 ChIP-chip HD2 data (Cheng *et al.*, [@B7]). The remaining (100 \u2212 *x*)% binding sites were placed at random locations. We varied *x* from 0% to 100% with a 20% increment. For each *x* value, we generated 10 ChIP-chip datasets using the method described above. Since the four supporting tracks were significantly correlated with GATA1 binding, they were correlated with the simulated binding sites at varying levels.\n\nWe applied our method to each simulated dataset, with insignificant supporting features removed by permutations, except for *x* = 0%, and we called significant peaks at 10% FDR level. As shown in [Figure 3](#F3){ref-type=\"fig\"}A, we observed that the actual FDR level for each simulated datasets remained around 10% or less at varying concordance levels. As shown in [Figure 3](#F3){ref-type=\"fig\"}B, after incorporating supporting features in peak calling, our method detected many more (up to an additional 20%) true peaks that were missed by conventional methods. The gain of power is an increasing function with respect to the absolute correlation between the protein-binding data and the supporting data.\n\n3.2 Performance of discretization methods {#SEC3.2}\n-----------------------------------------\n\nTo evaluate the peaks detected by our method in real data analysis, we constructed a positive set of 99 true GATA1 occupied segments validated by qPCR in G1E-ER4 cells in a previous study (Zhang *et al.*, [@B18]). The ChIP data used to identify those 99 validated peaks (VPs) was different from the ChIP-chip HD2 data analyzed in this article, and hence the VPs do not necessarily show strong signals in our HD2 data. In fact, we observed that the ChIP-chip HD2 intensity of the 99 VPs ranged from 0.32 to 4.07, where the entire ChIP-chip HD2 signals ranged from \u22125.76 to 4.07. The wide range of signals observed from the VPs provides us a good reference set to evaluate our method. We further partitioned the VPs into three groups: (i) high VPs: 53 true sites with probe intensities \\>2 in ChIP-chip HD2 data; (ii) medium VPs: 24 true sites with probe intensities between 1.5 and 2; and (iii) low VPs: 22 true sites with probe intensities \\<1.5. We also constructed a negative set of 83 FPs-binding sites that failed the previous validation of qPCR (Cheng *et al.*, [@B6]; Zhang *et al.*, [@B18]). Some of these 83 FPs, however, showed large signals in our ChIP-chip HD2 data and hence may be weak GATA1-binding sites. In addition to those positive and negative validation peaks, we also compared our results with results obtained using existing methods on the same ChIP-chip and ChIP-seq data (Cheng *et al.*, [@B7]).\n\nWe first evaluated the impacts of discretization methods on peak calling. We conducted independent peak-calling experiments on the ChIP-chip HD2 data using various discretization methods on the four supporting tracks. We first ran the PASS program (Zhang, [@B17]) to detect peaks at 10% FDR level. We then fit the output to a logistic regression model with the four supporting tracks as covariates. The following methods were used to discretize the supporting data: (i) round each probe value to the nearest integer ('Round'); (ii) calculate the average value (*t*~*i*~) of a 1000 bp window of each probe and round the value to integer ('SmoothRound'); (iii) use the four methods ('Equal-Width', 'Equal-Freq', 'Clustering' and 'Entropy') described in [Section 2](#SEC2){ref-type=\"sec\"}.\n\nAs shown in [Figure 4](#F4){ref-type=\"fig\"}, when evaluated at the same FDR level, 'Entropy' and 'Equal-Freq' outperformed other methods in terms of the number of detected peaks, medium VPs and peaks overlapping with previously identified ChIP-seq peaks (Cheng *et al.*, [@B7]). 'Equal-Width' performed the worst, but still slightly outperformed the conventional method. The results of Round, SmoothRound and Clustering were better than the conventional method, but worse than Entropy and Equal-Freq methods. Our results indicated that a proper choice of discretization method is important, as it have significant impacts on peak-calling results. Fig. 4.Comparison of peak-calling performance using six different discretization methods on supporting tracks: Equal-Width, Equal-Freq, Clustering (*k* = 2 \u223c 8), Entropy, Round and SmoothRound. The traditional peak-calling method without using supporting tracks (no prior) is shown as a comparison. (**A**) The total number of detected peaks. (**B**) The total number of detected true peaks with medium intensity (1.5--2.0). (**C**) The total number of detected peaks overlapping with previously identified ChIP-seq peaks.\n\n3.3 Novel peaks detected by supporting tracks {#SEC3.3}\n---------------------------------------------\n\nBy incorporating the 4 GATA1-related supporting tracks, we identified a total of 125 novel GATA1-binding sites, of which 66 sites were identified from the ChIP-chip data and 63 sites were identified from the ChIP-seq data ([Table 1](#T1){ref-type=\"table\"}). Before using supporting tracks, 14 out of 24 VPs with medium intensity were detected by the original PASS program at 10% FDR level. Two additional VPs with medium intensity lied within 400 bp of the PASS detected peaks. After using supporting data, our method captured four (out of the remaining eight) more VPs with medium intensity. The other four missing VPs were hard to detect from the ChIP-chip HD2 data and were also missed by the previous ChIP-chip analysis (Cheng *et al.*, [@B7]). They were only detected from the ChIP-seq data (Cheng *et al.*, [@B7]). Table 1.Predicted GATA1-binding sites enrichment in true binding regions, RefSeq Gene and predicted peaks by Cheng *et al.* ([@B7])DatasetNumber of peaksNumber of high VPsNumber of medium VPsNumber of low VPsTotal VPsNumber of FPChIP-chip peaksChIP-Seq peaksUnion of peaksRefSeq genes5324229983311[^a^](#TF1){ref-type=\"table-fn\"}780[^a^](#TF1){ref-type=\"table-fn\"}890[^a^](#TF1){ref-type=\"table-fn\"}1176ChIP-chip peaks\u2003PASS---no prior31751140654251221292198/149\u2003PASS2---additional660404119334443/38\u2003MPeaks147458053214211214591/74\u2003TMAL (L1)13940915011349713784/41ChIP-seq peaks\u2003PASS---no prior554451613744177463467325/186\u2003PASS2---additional63000005353536/33[^1][^2]\n\nFor VPs of high intensity in the ChIP-chip data, our method with and without using supporting tracks performed equally---51 out of 53 VPs overlapped with our detected peaks. For VPs of weak intensity, none were detected by our method with or without using supporting tracks. From the ChIP-seq data, however, we detected 13 VPs (60%) of low intensity. Although we found more GATA1-binding sites from the ChIP-seq data than from the ChIP-chip data, it is worthy of noting that the true peaks found in ChIP-chip data were not completely captured by the ChIP-seq data, and vice versa.\n\nTo further evaluate the novel peaks found by the supporting tracks, we compared our results with peaks found in a previous study (Cheng *et al.*, [@B7]). Of the 66 additional ChIP-chip peaks, 50% (33/66) overlapped with the previously identified ChIP-seq peaks, 65% (43/66) overlapped with 38 RefSeq Genes. Similar results were observed in the 63 additional ChIP-seq peaks. We also compared our results with MPeak and TAMALPAIS (Bieda *et al.* [@B3]). The two programs, however, do not directly provide FDR control. For MPeak, we used its *P*-value in the output to compute FDR assuming independence between tests. At 10% FDR threshold, MPeak detected 147 peaks when their *P*-value cutoff is around 1.6e-05. For TAMALPAIS, we used the most stringent threshold (L1) and detected 139 peaks. Compared to PASS2, MPeak and TAMALPAIS (L1) were both conservative.\n\nFor the effects of the four supporting tracks, as shown in [Table 2](#T2){ref-type=\"table\"}, the open chromatin, H3k4me3 and TAL1 all had positive effects on GATA1 binding, where H3K27me3 as a repressor had negative effect on GATA1 binding. We observed that the effect of the open chromatin data and the GATA1-binding status in the ChIP-chip was different from that in the ChIP-seq data. The inconsistency may be attributable to the difference in noise levels and the bias in ChIP-chip and ChIP-seq experiments. Table 2.Estimated effects and *P*-values of the supporting tracksChIP-chipChIP-seq\u03b2*P*-value\u03b2*P*-valueIntercept\u22121.25e+010\u22121.14e+010Open chromatin2.42e\u2212026.73e\u2212016.70e\u2212015.26e\u221248H3K27me3\u22122.63e\u2212019.78e\u221206\u22121.51e+001.25e\u221212H3K4me31.79e+001.36e\u22121061.18e+009.45e\u221234TAL17.99e\u2212012.46e\u2212671.51e+003.26e\u2212246[^3]\n\nAmong the 554 ChIP-seq peaks that were originally detected by PASS, three peaks were redefined by our method at shifted nearby locations ([Fig. 5](#F5){ref-type=\"fig\"}A). The raw ChIP-seq data around the three sites showed double-peak distributions. As suggested by the supporting data, the new binding sites defined by PASS2 were more likely to be occupied by GATA1.We show in [Figure 5](#F5){ref-type=\"fig\"}B one example of the novel peaks found by our method. The GATA1 occupied segment in this region was validated by qPCR and had a max ChIP-chip intensity of 1.77. The region, however, was missed by a previous study using both ChIP-chip HD2 data and ChIP-seq data (Cheng *et al.*, [@B7]). The original PASS program also missed this region. This GATA1 occupied segment is located within gene 'Tjp1', which is depleted of H3K27me3 but enriched with open chromatin and H3K4me3 signals. After incorporating the related feature tracks, we recovered this true GATA1-binding site using the proposed method. Fig. 5.(**A**) Venn diagram of the ChIP-chip and ChIP-seq peaks identified with and without the supporting tracks. (**B**) An example of a novel GATA1-occupied segment within Tjp1 identified by our method. It was missed by previous HD2 ChIP-chip and ChIP-seq analysis (Cheng *et al.*, [@B7]). This region also shows depleted H3K27me3 and enriched H3K4me3 signals. Horizontal black lines indicate signal means.\n\n4 DISCUSSION {#SEC4}\n============\n\nWe introduced a new statistical method to improve the power of detecting protein binding in ChIP data by combining additional biological features. The proposed method not only improves the sensitivity of peak calling than traditional methods, but also precisely controls a user-specified level of FP rate or FDR. The additional sites detected by our method are those regions with medium- or low-binding signals, which cannot pass the genome-wide statistical significance control. After taking into account of the correlation between the binding sites and biologically related supporting features, regions coincide with strong supporting signals will become detectable by our method.\n\nUsing both simulation and real data analysis, we observed that our method can effectively detect 20% more true binding sites than traditional peak-calling methods. Under all scenarios we tested, our method also precisely controlled the proportion of FP calls at 10% FDR level. We further observed that the proposed method is robust to irrelevant data tracks added to the model.\n\nOur method does neither assume any distributions on the supporting data nor it attempts to estimate data distributions empirically. Such assumptions and estimation procedures may introduce unwanted bias and uncertainty in peak calling. Our method is flexible to incorporate any types of biological information that overlap with the ChIP regions. By converting continuous data into bins, the binding probabilities computed from our logistic regression model will be robust to outliers and extreme values. Our analysis showed that a proper discretization method applied to supporting tracks can also have significant impact on peak calling results. Instead of applying discretization to individual supporting tracks (univariate discretization), multi-variate discretization methods that take into account of the interaction among features could be used to capture the missing patterns from univariate approaches (Bay, [@B1]).\n\nThe real data analysis we performed in this study was based on ChIP-chip data and a ChIP-seq data in 66 Mb region on the mouse chromosome 7. This region contained a substantial amount of GATA1- and TAL1-binding sites compared to other regions (Cheng *et al.*, [@B6]). This is a focal region for investigating the interaction between proteins and is a best region to test our method. We used GATA1 peaks detected at a stringent threshold to train the model parameters to avoid using many FPs in fitting the logistic regression model. It is possible that the fitted regression model may be biased towards strong binding sites. The underlying assumption of our method, therefore, is that the fitted model from strong peaks will have predictive power to weak binding sites. The worst scenario will occur when weak binding sites have completely opposite supporting data distribution compared to that of strong binding sites, in which case the fitted regression model will predict against the weak binding sites. This is a common issue that will occur to all methods that rely on training data, if the training data and the testing data are heterogeneous. In our method, we attempted to alleviate this problem by repeating the process of peak-calling, fitting regression and peak-calling again, iteratively, such that if weak binding sites are detected at some iteration, their information will be included in the training of regression model, and a new iteration of peak-calling will be applied. The novel peaks found by our method was a set of candidate peaks of weaker binding events. It is still unclear whether the binding strength of a transcription factor can affect gene regulation. The weak peaks detected by our method therefore provide a source of information for investigating this association. The logistical regression model we fitted in this region can be further applied to predict genome-wide GATA1 occupancy.\n\nThere are currently a large number of computational methods developed for detecting protein--DNA interactions in ChIP experiments. Most methods did not provide a rigorous means to control FP detections. The PASS algorithm (Zhang, [@B17]) solved this problem in the context of correlated multiple comparisons. The detected statistically significant binding intervals, however, may not correspond well to the real biological binding sites. The method proposed in this study attempts to improve both sensitivity and specificity of peak calling by incorporating biologically related information. The proposed framework is flexible in terms of accommodating various types of data as supporting tracks, and is also flexible in terms of the methods used at each step of the algorithm. Instead of using a logistic regression model to predict binding of the target protein, statistical or machine-learning classifiers can be used to measure the potential of protein binding at each probe from the supporting data. We then fit the predicted binding potentials into the varying threshold framework to determine probe-specific thresholds.\n\nWith the advancement of next-generation sequencing technologies, researchers are now able to generate a huge amount of data of various biological features of interest, including ChIP data for multiple transcription factors, histone modifications, nucleosome positioning and RNA-seq. These feature tracks are usually highly correlated and jointly provide valuable information for answering some of the fundamental questions in gene regulation. The proposed method is an example of integrating such information to increase the power and specificity in peak detection.\n\n*Funding*: National Human Genome Research Institute (grants R01 DK065806 to Y.Z. and HG002238 to K.C., in part).\n\n*Conflict of Interest*: none declared.\n\n[^1]: ^a^Computationally identified peaks by Cheng *et al.* ([@B7]).\n\n[^2]: VPs, validated peaks by q-PCR (Cheng *et al.*, [@B6]; Zhang *et al.*, [@B18]); RefSeq Genes, RefSeq genes from the UCSC browser in the 66 Mb region on chromosome 7 in the mouse genome (mm8). Overlapping entries are merged. The overlapping intervals between RefSeq genes and the detected peaks have *P*-value \\< 0.05 from 100 permutations.\n\n[^3]: \u03b2 is the regression coefficient in the logistic regression model.\n"}
+{"text": "Background\n==========\n\nNatural killer (NK) cells comprise 5--20% of peripheral blood mononuclear cells (PBMC) in humans and play a fundamental role in the defense against viral infections, as well as in tumor surveillance, and help shape adaptive immune responses through their production of cytokines \\[[@B1]-[@B3]\\]. NK cells are traditionally identified as CD3^neg^, CD14^neg^, CD19^neg^ lymphocytes expressing CD56 (neural cell adhesion molecule) and NKp46 \\[[@B4]\\], although not all human NK cells express NKp46 \\[[@B1]\\]. NK cells are typically characterized into two main subsets; CD56^dim^CD16^+^ NK cells, which comprise approximately 90% of circulating NK cells and are considered mature, and CD56^bright^CD16^neg/dim^ NK cells, which represent approximately 10% and are considered immature \\[[@B1],[@B5]\\]. A diverse array of activating and inhibitory receptors controls their function. Upon activation, NK cells secrete IFN\u03b3 and other cytokines and kill susceptible target cells \\[[@B2]\\].\n\nInitial studies of NK cells during Human Immunodeficiency Virus (HIV-1) infection found significantly reduced absolute numbers of CD3^neg^CD56^+^CD16^+^ NK cells with a concomitant increase in CD3^neg^CD56^neg^CD16^+^ cells \\[[@B6]-[@B8]\\]. In HIV-1-infected subjects, CD3^neg^CD56^neg^CD16^+^ cells are described as having decreased expression of activating receptors (i.e. NKp30 and NKp46) and increased expression of inhibitory receptors (i.e. LIR-1 and inhibitory KIR), and have poor cytolytic, proliferative, and cytokine-producing capabilities \\[[@B8]-[@B14]\\]. It has been hypothesized that the expansion of this defective CD3^neg^CD56^neg^CD16^+^ population might be one mechanism by which HIV-1 subverts the NK cell response.\n\nExpansion of CD3^neg^CD56^neg^CD16^+^ cells has also been observed in other infectious diseases (i.e. hantavirus and chronic hepatitis C virus (HCV) infection) \\[[@B15],[@B16]\\], as well as in ocular myasthenia gravis \\[[@B17]\\] and dermatomyositis \\[[@B18]\\]. Common to each of these pathological conditions is immune activation. Indeed, chronic immune activation is a hallmark of untreated HIV-1 disease and results in accelerated immunosenescence \\[[@B19]\\]. In healthy subjects, one study has suggested that NK cells might proliferate and die more rapidly than do T cells however this requires further investigation \\[[@B20]\\]. We recently demonstrated that CD57 is a marker of terminally differentiated NK cells \\[[@B21]\\], and that during acute cytomegalovirus (CMV) infection NKG2C^hi^-expressing NK cells acquire expression of CD57 \\[[@B22]\\]. Importantly, treatment with antiviral drugs in HIV-1 and HCV infection or immunosuppressants for myasthenia gravis and dermatomyositis decrease the frequency of CD3^neg^CD56^neg^CD16^+^ cells over the course of months to levels found in healthy subjects \\[[@B11],[@B16]-[@B18]\\].\n\nThere remain many unanswered questions regarding the phenotype, function, and origin of CD3^neg^CD56^neg^CD16^+^ cells in healthy individuals, and how they compare to the expanded population found during chronic infectious diseases such as HIV-1 infection. We recently demonstrated that including CD7 as an additional positive NK cell marker is an effective method for studying non-classical NK cell subsets \\[[@B23]\\]. Using CD7, we expand our collective knowledge about the phenotype and function of CD56^neg^CD16^+^ cells in healthy and HIV-1-infected subjects. Based upon previous observations that chronic viremia is associated with an increased frequency of CD56^neg^CD16^+^ cells, we tested the hypothesis that CD7 would refine the population of NK cells defined as CD56^neg^CD16^+^ cells and provide evidence that CD7^+^CD56^neg^CD16^+^ NK cells are mature NK cells.\n\nResults\n=======\n\nCD56^neg^CD16^+^ cells are a mixed population of myeloid and NK cells that is expanded during chronic HIV-1 infection\n---------------------------------------------------------------------------------------------------------------------\n\nPBMC from HIV-1-negative and HIV-1-infected subjects (Table\u00a0[1](#T1){ref-type=\"table\"}) from the OPTIONS (early infection) and SCOPE (chronic infection) cohorts were gated on CD3^neg^CD14^neg^CD19^neg^ cells and the three major NK cell subsets were assessed: CD56^bright^CD16^neg^, CD56^+^CD16^+^, and CD56^neg^CD16^+^ (Additional file [1](#S1){ref-type=\"supplementary-material\"}). A significant decrease in the frequency of CD56^bright^CD16^neg^ and CD56^+^CD16^+^ NK cells with a significant increase in the frequency of CD56^neg^CD16^+^ cells was observed in HIV-1-infected subjects, particularly during chronic HIV-1 infection (Figure\u00a0[1](#F1){ref-type=\"fig\"}A). These results are in agreement with previous studies describing CD56^neg^CD16^+^ cells as highly expanded during chronic HIV-1 (and HCV) infections \\[[@B8],[@B12],[@B16],[@B24]\\].\n\n###### \n\n**Early and chronic HIV-1-infected patient viral loads and CD4**^**+**^**T cell counts**\n\n   **Patient ID**   **Cohort**   **Antiretroviral status**   **Viral load**   **CD4**^+^**T cell count**\n  ---------------- ------------ --------------------------- ---------------- ----------------------------\n      **647**        OPTIONS               None                  73988                   594\n      **683**        OPTIONS               None                  95903                   480\n      **722**        OPTIONS               None                  124070                  522\n      **730**        OPTIONS               None                   2185                   594\n      **792**        OPTIONS               None                  128274                  281\n      **804**        OPTIONS               None                   4812                   800\n      **830**        OPTIONS               None                  14207                   850\n      **858**        OPTIONS               None                  56133                   466\n      **876**        OPTIONS               None                  149012                  343\n      **1049**        SCOPE                None                  336630                  218\n      **1070**        SCOPE                None                  365000                  331\n      **1091**        SCOPE                None                  59935                   288\n      **1208**        SCOPE                None                  17641                   1054\n      **1217**        SCOPE                None                  19933                   281\n      **1311**        SCOPE                None                  56566                   647\n      **1335**        SCOPE                None                  50738                   501\n      **1538**        SCOPE                None                  47025                   612\n      **1566**        SCOPE                None                  34418                   853\n      **1571**        SCOPE                None                  21734                   381\n      **1587**        SCOPE                None                  13589                   469\n      **1588**        SCOPE                None                  25552                   891\n      **1596**        SCOPE                None                  23364                   621\n      **1597**        SCOPE                None                  12593                   498\n      **1606**        SCOPE                None                  24458                   678\n      **1654**        SCOPE                None                  91597                   441\n      **1662**        SCOPE                None                  236000                  584\n      **1679**        SCOPE                None                  145465                  288\n      **4015**        SCOPE                None                  63400                   714\n\n![**Distribution of NK cell subsets. A**, Frequency of classically defined NK cell subsets within the CD3^neg^ CD14^neg^ CD19^neg^ population of healthy donors ((\u2022) N, n\u2009=\u200963) and early- ((\u2610) E, n\u2009=\u20099) and chronically- (() C, n\u2009=\u200919) HIV-1-infected subjects. **B**, Frequency of CD3^neg^ CD14^neg^ CD19^neg^ cells expressing CD7 within the CD56^bright^CD16^neg^, CD56^+^CD16^+^ and CD56^neg^CD16^+^ subsets in healthy ((\u2022) N, n\u2009=\u200963) and early- ((\u2610) E, n\u2009=\u20099) and chronically- (() C, n\u2009=\u200919) HIV-1-infected subjects. **C**, Frequency of CD56^neg^CD16^+^ cells within the CD3^neg^ CD14^neg^ CD19^neg^ population (left axis) and of CD3^neg^ CD14^neg^ CD19^neg^ cells expressing CD7 within the CD56^neg^CD16^+^ subset (right axis) of CMV-negative (closed circles) and CMV-positive (open circles) healthy donors. The median and 25th and 75th percentile are indicated on each graph.](1742-4690-10-158-1){#F1}\n\nA recent gene expression study of CD56^neg^CD16^+^ cells, traditional CD56^+^CD16^+^ NK cells, and myeloid cells found that the transcriptome of CD56^neg^CD16^+^ cells was more similar to myeloid cells than to traditional CD56^+^CD16^+^ NK cells \\[[@B25]\\]. Therefore, we hypothesized that CD56^neg^CD16^+^ cells are a mixed population of myeloid and NK cells that could be distinguished using CD7 \\[[@B23]\\]. In HIV-1-negative controls, less than 40% of CD56^neg^CD16^+^ cells co-express CD7, indicating the CD56^neg^CD16^+^ population is not a homogeneous population of NK cells (Figure\u00a0[1](#F1){ref-type=\"fig\"}B). During early HIV-1 infection, there is a significant expansion in the frequency of CD7^+^CD56^neg^CD16^+^ NK cells that appears to persist into chronic HIV-1 infection; however, CD56^neg^CD16^+^ cells remain a mixed myeloid and NK cell population (Figure\u00a0[1](#F1){ref-type=\"fig\"}B). In the absence of CD7 gating we did not observe the expansion of CD56^neg^CD16^+^ cells during early infection (Figure\u00a0[1](#F1){ref-type=\"fig\"}A). Persistent HIV-1 viremia appears to be important for the accumulation of CD7^+^CD56^neg^CD16^+^ NK cells \\[[@B8],[@B12],[@B16],[@B24]\\]. No significant difference was observed in the frequency of CD7^+^CD56^neg^CD16^+^ NK cells in healthy, non-HIV infected individuals, including both cytomegalovirus-seropositive and cytomegalovirus-seronegative subjects (Figure\u00a0[1](#F1){ref-type=\"fig\"}C). Taken together, CD56^neg^CD16^+^ cells are a mixed population of cells, a subset of which are CD7^+^ NK cells that are expanded during persistent HIV-1 viremia.\n\nCD7^+^CD56^neg^CD16^+^ and CD7^+^CD56^+^CD16^+^ NK cells are phenotypically similar in healthy donors; however, HIV-1 infection significantly alters the phenotype of both NK cell subsets\n------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nIn the absence of CD7 gating, CD56^neg^CD16^+^ cells appear to have no or very low levels of NKp30 and granzyme B expression. However, when the CD7^+^CD56^neg^CD16^+^ NK cells are gated and compared to mature CD7^+^CD56^+^CD16^+^ NK cells, little difference between the cell subsets was observed (Additional file [2](#S2){ref-type=\"supplementary-material\"}). As these data are in disagreement with previous reports \\[[@B12],[@B26]\\], we sought to further characterize CD7^+^CD56^neg^CD16^+^ NK cells compared to CD7^+^CD56^+^CD16^+^ NK cells in healthy and HIV-1-infected individuals.\n\nWe assessed the expression of Killer cell Immunoglobulin-like Receptors (KIRs), C-type lectin-like receptors (NKG2A, NKG2C and CD161), and Natural Cytotoxicity Receptors (NCRs; NKp30, and NKp46), as well as receptors typically found on myeloid cells (CD33, CD13, and HLA-DR) in HIV-1-negative donors (Figure\u00a0[2](#F2){ref-type=\"fig\"}). No significant difference was observed in the frequency of KIR2DL1- and KIR2DL3-bearing cells between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells, although CD7^+^CD56^neg^CD16^+^ NK cells had a significantly lower frequency of KIR3DS1/DL1-positive NK cells. The density of KIR2DL1 and KIR2DL3 as measured by mean fluorescent intensity (MFI) was significantly lower on CD7^+^CD56^neg^CD16^+^ NK cells, whereas no significant difference in MFI of KIR3DS1/DL1 was observed between the NK cell subsets (Figure\u00a0[2](#F2){ref-type=\"fig\"}A-C). With regard to the C-type lectin-like receptors, similar frequencies of both NK cell subsets expressed NKG2A and NKG2C. While both NK cell subsets had similar densities of NKG2C expression, the density of NKG2A was significantly lower on CD7^+^CD56^neg^CD16^+^ NK cells (Figure\u00a0[2](#F2){ref-type=\"fig\"}D-E). No significant differences were observed in the frequency of NKp30 and NKp46 between the NK cell subsets (Figure\u00a0[2](#F2){ref-type=\"fig\"}G-H). In contrast to previous studies, NKp30 and NKp46 were expressed at significantly greater densities on CD7^+^CD56^neg^CD16^+^ NK cells compared with CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[2](#F2){ref-type=\"fig\"}G-H). By including all CD7^+^CD56^+^ NK cells (i.e. CD56^br^CD16^neg^, CD56^br^CD16^dim^, and CD56^+^CD16^+^) in the analyses as was done by Mavilio *et al.*\\[[@B12]\\], the MFI of NKp30 and NKp46 expression remained significantly greater in CD7^+^CD56^neg^CD16^+^ NK cells compared with total CD7^+^CD56^+^ NK cells and CD7^+^CD56^+^CD16^+^ NK cells (Additional file [3](#S3){ref-type=\"supplementary-material\"}). However, total CD7^+^CD56^+^ NK cells showed a trend (p\u2009=\u20090.070) toward greater NKp46 density compared to CD7^+^CD56^+^CD16^+^ NK cells (Additional file [3](#S3){ref-type=\"supplementary-material\"}B). No significant differences were observed in the frequency or density of CD161 expression between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells (Figure\u00a0[2](#F2){ref-type=\"fig\"}F). A prior study of CD56^neg^CD16^+^ NK cells indicated a significantly greater expression of the inhibitory receptor LIR-1 on CD56^neg^CD16^+^ compared to CD56^+^CD16^+^ NK cells \\[[@B12]\\]. However, using CD7 to further delineate NK cells, we observed that LIR-1 density on CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells was not significantly different (Figure\u00a0[2](#F2){ref-type=\"fig\"}I). Indeed, the apparent higher levels of LIR-1 expression on CD56^neg^CD16^+^ cells was likely due to the contamination with CD7^neg^CD56^neg^CD16^+^ myeloid cells that have five-fold greater LIR-1 density than CD7^+^ NK cells. Importantly, CD7^neg^CD56^neg^CD16^+^ cells did not express any of these NK cell-associated receptors. However, high frequencies of CD7^neg^CD56^neg^CD16^+^ cells did express CD33, CD13, and HLA-DR, markers that are classically expressed by myeloid cells (Figure\u00a0[2](#F2){ref-type=\"fig\"}J-L). Recently, Bigley *et al.* observed that a subset of CD123+ plasmacytoid dendritic cells express CD7 \\[[@B27]\\]. Although approximately 80% of the CD7^neg^CD56^neg^CD16^+^ cells expressed CD123, no expression of CD123 was observed on CD7^+^CD56^+^CD16^+^ or CD7^+^CD56^neg^CD16^+^ NK cells (Figure\u00a0[2](#F2){ref-type=\"fig\"}M). These results are in agreement with Bigley *et al.* who reported that CD7^+^\u00a0CD123^+^ plasmacytoid dendritic cells were CD16-negative \\[[@B27]\\]. The phenotypic profile of CD7^neg^CD56^neg^CD16^+^ cells is largely consistent with non-classical CD14^neg^CD16^+^ monocytes and a subset of dendritic cells (DCs) designated slanDCs \\[[@B28]\\] that fall within the lymphoid gate based on light scattering properties.\n\n![**Phenotypic characterization of CD7**^**+**^**CD56**^**+**^**CD16**^**+**^**and CD7**^**+**^**CD56**^**neg**^**CD16**^**+**^**NK cells and CD7**^**neg**^**CD56**^**neg**^**CD16**^**+**^**cells in healthy donors.** CD3^neg^CD14^neg^CD19^neg^ cells were gated for CD56^+^ and CD56^neg^ cells and plotted against CD7 to identify CD7^+^ NK cells and CD7^neg^ monocyte or DC-like cells. The percentage of CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells and CD7^neg^CD56^neg^CD16^+^ monocyte or DC-like cells expressing each receptor was determined (n\u2009=\u20096). **A**-**I**, Expression of NK cell-associated receptors on the different cell subsets. **J**-**M**, Expression of myeloid-associated markers on the different cell subsets. The median and 25th and 75th percentile are indicated on each graph.](1742-4690-10-158-2){#F2}\n\nTo determine the impact of HIV-1 infection on the phenotype of CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells, KIR, C-type lectin-like receptors, and NCRs were assessed in 16 chronically HIV-1-infected subjects and 8 healthy controls (Figure\u00a0[3](#F3){ref-type=\"fig\"}). HIV-1-infected subjects trended towards a reduced frequency of NKp30- and NKp46-positive CD7^+^CD56^+^CD16^+^ NK cells compared to healthy controls. However, the frequency of NKp30- and NKp46-expressing CD7^+^CD56^neg^CD16^+^ NK cells and the density of NKp30 and NKp46 expression on CD7^+^CD56^neg^CD16^+^ NK cells were significantly lower in HIV-1-infected subjects. Importantly, there were no significant differences in the frequencies of cells expressing NCRs or the density of the NCRs between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in the HIV-1-infected subjects (Figure\u00a0[3](#F3){ref-type=\"fig\"}A and B). Assessment of KIR2DL3 and KIR3DL1/DS1 did not reveal any significant differences between subsets of NK cells or the HIV-1 infection status of the subject (Figure\u00a0[3](#F3){ref-type=\"fig\"}C and D). A significantly lower frequency of CD7^+^CD56^neg^CD16^+^ NK cells within HIV-1-infected subjects expressed NKG2A, while the density of NKG2A expression was significantly greater on CD7^+^CD56^+^CD16^+^ NK cells in HIV-1-infected subjects (Figure\u00a0[3](#F3){ref-type=\"fig\"}E). The frequency of NK cells and density of the activating receptor NKG2C is elevated in CMV-infected individuals \\[[@B22],[@B29],[@B30]\\], a co-infection that is highly prevalent (\\>98% in the SCOPE cohort \\[[@B31]\\]) in HIV-1-infected subjects. To this end, it was not unexpected to observe that the frequency of NKG2C-bearing cells was significantly greater on both subsets of NK cells in HIV-1-infected subjects compared to healthy controls (Figure\u00a0[3](#F3){ref-type=\"fig\"}F). Furthermore, the density of NKG2C expression was significantly greater on CD7^+^CD56^+^CD16^+^ NK cells in HIV-1-infected compared to uninfected individuals. However, no significant differences were observed between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cell subsets within the HIV-1-infected subjects (Figure\u00a0[3](#F3){ref-type=\"fig\"}F). The increased frequency and density of NKG2C^+^ NK cells in HIV-1-infected subjects likely represents the higher prevalence of cytomegalovirus infection or reactivation in these subjects \\[[@B31]\\]. Taken together these results indicate that HIV-1 infection has a significant impact on the overall NK cell phenotype. However, with the exception of NKp30 and NKp46, CD7^+^CD56^neg^CD16^+^ NK cells do not appear to be significantly altered compared to CD7^+^CD56^+^CD16^+^ NK cells within HIV-1-infected subjects.\n\n![**Assessment of CD7**^**+**^**CD56**^**+**^**CD16**^**+**^**and CD7**^**+**^**CD56**^**neg**^**CD16**^**+**^**NK cells in healthy (n\u2009=\u20098) and chronically (n\u2009=\u200916) infected HIV-1 donors. A** and **B**, The frequency and mean fluorescence intensity (MFI) of NKp30 **(A)**, NKp46 **(B)**, KIR2DL3 **(C)**, KIR3DL1/DS1 **(D)**, NKG2A **(E)** and NKG2C **(F)** within CD7^+^CD56^neg^CD16^+^ and CD7^+^CD56^+^CD16^+^ NK cells. The median and 25th and 75th percentile are indicated on each graph.](1742-4690-10-158-3){#F3}\n\nCD7^+^CD56^neg^CD16^+^ NK cells are mature NK cells\n---------------------------------------------------\n\nTo determine whether differences in maturation existed between CD7^+^CD56^neg^CD16^+^ and CD7^+^CD56^+^CD16^+^ NK cells, we assessed the expression of CD16, CD57, Siglec-7, CD62L, and CXCR3. A significantly greater frequency of CD7^+^CD56^+^CD16^+^ NK cells in HIV-1-infected subjects expressed CD57 compared to healthy subjects. Although CD57 expression on CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in healthy controls was not significantly different, the frequency of CD57^+^CD7^+^CD56^neg^CD16^+^ NK cells in HIV-1-infected subjects was significantly less compared to CD57^+^CD7^+^CD56^+^CD16^+^ NK cells (\\~44% versus \\~54% respectively; Figure\u00a0[4](#F4){ref-type=\"fig\"}A). When we separated healthy subjects by CMV serology status, we did not observe a significant difference in CD57 expression between the NK cell subsets (data not shown). CD16 expression increases with NK cell maturation \\[[@B21]\\]. We observed that early HIV-1 infection resulted in a significant increase in CD16 expression on both subsets of NK cells compared to healthy controls (Figure\u00a0[4](#F4){ref-type=\"fig\"}B left half). Because the studies were performed at different times, we were unable to directly compare the density of CD16 expression (based on MFI) in early and chronic HIV-1 infection. However, in contrast to early infection, CD16 expression did not appear to differ significantly on CD7^+^CD56^+^CD16^+^ or CD7^+^CD56^neg^CD16^+^ NK cell subsets in chronically HIV-1-infected subjects compared to healthy controls (Figure\u00a0[4](#F4){ref-type=\"fig\"}B, right half). Recently, HIV-1 infection was reported to induce a rapid and sustained decrease in Siglec-7 expression on NK cells, which is associated with impaired function \\[[@B11]\\]. We observed that greater than 90% of CD7^+^CD56^+^CD16^+^ NK cells in healthy donors expressed Siglec-7, whereas approximately 50% of this NK cell subset expressed Siglec-7 in HIV-1-infected subjects (Figure\u00a0[4](#F4){ref-type=\"fig\"}C). Furthermore, we found that a lower frequency of CD7^+^CD56^neg^CD16^+^ NK cells expressed Siglec-7 in both healthy controls and HIV-1-infected subjects, but the loss of Siglec-7 on CD7^+^CD56^neg^CD16^+^ NK cells of HIV-1-infected subjects was significantly greater (Figure\u00a0[4](#F4){ref-type=\"fig\"}C). During NK cell maturation from an immature CD56^bright^CD16^neg^ to a mature CD56^+^CD16^+^ phenotype, CD62L and CXCR3 expression are decreased \\[[@B32]-[@B34]\\]. In accordance with a mature phenotype, CD7^+^CD56^neg^CD16^+^ NK cells have similar CD62L expression compared to CD7^+^CD56^+^CD16^+^ NK cells, which is significantly lower than that expressed by CD7^+^CD56^bright^CD16^neg^ immature NK cells (Figure\u00a0[4](#F4){ref-type=\"fig\"}D). Similarly, less than 5% of both CD7^+^CD56^neg^CD16^+^ and CD7^+^CD56^+^CD16^+^ NK cells express CXCR3 (Figure\u00a0[4](#F4){ref-type=\"fig\"}E). By comparison, approximately 70% of CD7^+^CD56^bright^CD16^neg^ immature NK cells express CXCR3, and HIV-1 infection is associated with a significantly lower frequency of CXCR3^+^ CD7^+^CD56^bright^CD16^neg^ immature NK cells compared to healthy controls (Figure\u00a0[4](#F4){ref-type=\"fig\"}E). In healthy controls, CXCR3 is expressed by a significantly lower frequency of CD7^+^CD56^neg^CD16^+^ compared to CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[4](#F4){ref-type=\"fig\"}E). HIV-1 infection was associated with a significant decrease in the frequency of CXCR3^+^CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[4](#F4){ref-type=\"fig\"}E). Taken together, CD7^+^CD56^neg^CD16^+^ and CD7^+^CD56^+^CD16^+^ NK cells are mature NK cell subsets with at least a fraction being terminally differentiated CD57^+^ NK cells; however, HIV-1-infection significantly alters the differentiation of both NK cell subsets.\n\n![**Assessment of maturation in NK cell subsets of healthy and HIV-1-infected subjects. A**, Frequency of CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells expressing CD57 in healthy (n\u2009=\u200962) and early (n\u2009=\u20099) and chronically (n\u2009=\u200919) HIV-1-infected subjects. **B**, The MFI of CD16 expression was assessed on CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in healthy (n\u2009=\u200952) and early (n\u2009=\u20099) HIV-1-infected subjects (left half), and in healthy (n\u2009=\u200910) and chronically HIV-1-infected subjects (n\u2009=\u200919) (right half). The healthy controls and early HIV-1 subjects were stained and analyzed in a different study with an anti-CD16 antibody conjugated to a different fluorophore than the healthy controls and chronic HIV-1-infected subjects. Therefore, early and chronic HIV-1-infected subjects cannot be directly compared with regard to mean fluorescent intensity. **C**, The frequency of CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells expressing Siglec-7. **D-E**, The frequency of CD7^+^CD56^bright^CD16^neg^, CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells expressing **D)** CD62L and **E)** CXCR3. The same healthy (n\u2009=\u200910) and chronic HIV-1-infected subjects (n\u2009=\u200919) used in 4B were used to assess Siglec-7, CD62L and CXCR3. The median and 25th and 75th percentile are indicated on each graph. ( = healthy; = HIV-1-infected).](1742-4690-10-158-4){#F4}\n\nHIV-1 infection has a significant impact on the function of both NK cell subsets\n--------------------------------------------------------------------------------\n\nGranzyme B and perforin are important effector molecules expressed by mature NK cells. In healthy controls, no statistically significant differences in the frequency of NK cells expressing granzyme B or the amount of granzyme B in the NK cell subsets was observed, although a trend toward a decreased frequency of CD7^+^CD56^neg^CD16^+^ NK cells expressing granzyme B was observed (Figure\u00a0[5](#F5){ref-type=\"fig\"}A). In contrast, the frequency of granzyme B-positive NK cells and the amount of granzyme B expression were significantly lower in CD7^+^CD56^neg^CD16^+^ NK cells of HIV-1-chronically-infected subjects. No statistically significant differences were observed between healthy control and HIV-1-chronically-infected subjects in granzyme B expression within the CD7^+^CD56^+^CD16^+^ or CD7^+^CD56^neg^CD16^+^ NK cell subsets (Figure\u00a0[5](#F5){ref-type=\"fig\"}A). A significantly lower amount of perforin was observed in CD7^+^CD56^neg^CD16^+^ NK cells of both healthy controls and HIV-1-chronically-infected subjects compared with CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[5](#F5){ref-type=\"fig\"}B). The frequency of cells expressing perforin in both CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cell populations was significantly lower in HIV-1-chronically-infected subjects. Furthermore, the amount of perforin in CD7^+^CD56^neg^CD16^+^ NK cells was significantly lower compared to CD7^+^CD56^+^CD16^+^ NK cells in HIV-1-chronically-infected subjects (Figure\u00a0[5](#F5){ref-type=\"fig\"}B). Taken together, significantly reduced granzyme B and perforin expression in CD7^+^CD56^neg^CD16^+^ NK cells of HIV-1-chronically-infected subjects might indicate this NK cell subset has recently degranulated in response to an encounter with target cells.\n\n![**Perforin and granzyme B expression in CD7**^**+**^**CD56**^**+**^**CD16**^**+**^**and CD7**^**+**^**CD56**^**neg**^**CD16**^**+**^**NK cells of healthy and chronically infected HIV-1 subjects. A**, Granzyme B expression in healthy donors and chronically infected HIV-1 subjects. **B**, Perforin expression in healthy donors and chronically infected HIV-1 subjects. Healthy subjects n\u2009=\u200910; chronic HIV-1-infected subjects n\u2009=\u200917. The median and 25th and 75th percentile are indicated on each graph. ( = healthy; = HIV-1-infected).](1742-4690-10-158-5){#F5}\n\nCD56^neg^CD16^+^ NK cells have been described as anergic and having poor cytokine and cytotoxic function *in vitro*\\[[@B9],[@B10],[@B12]\\]. Indeed, in both healthy controls and HIV-1-infected subjects, CD7^+^CD56^neg^CD16^+^ NK cells expressed significantly less IFN\u03b3 compared to CD7^+^CD56^+^CD16^+^ NK cells following K562 target cell stimulation (Figure\u00a0[6](#F6){ref-type=\"fig\"}A-B). IL-12 plus IL-18 stimulation did induce IFN\u03b3 secretion from CD7^+^CD56^neg^CD16^+^ NK cells in healthy controls, albeit less than CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[6](#F6){ref-type=\"fig\"}A). CD7^+^CD56^+^CD16^+^ NK cells from HIV-1-chronically-infected subjects had a significantly lower frequency of IFN\u03b3-secreting cells in response to either K652 target cells (p =\u20090.033) or IL-12 plus IL-18 stimulation (p\u2009=\u20090.011) (Figure\u00a0[6](#F6){ref-type=\"fig\"}A-B). CD7^+^CD56^neg^CD16^+^ NK cells from HIV-1-chronically-infected subjects had a significantly higher frequency of IFN\u03b3^+^ cells (p\u2009=\u20090.013) compared to healthy controls in the unstimulated condition. IFN\u03b3 responses induced by K562 target cells in CD7^+^CD56^neg^CD16^+^ NK cells from healthy and HIV-1-chronically-infected subjects were negligible. CD7^+^CD56^neg^CD16^+^ NK cells from HIV-1-chronically-infected subjects failed to produce IFN\u03b3 following cytokine stimulation (Figure\u00a0[6](#F6){ref-type=\"fig\"}B). In contrast to IFN\u03b3 responses, no significant differences in the frequencies of CD107a-expressing CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells were observed following K562 target stimulation in healthy controls. Indeed, both subsets of NK cells were capable of degranulating in response to target cell stimulation (Figure\u00a0[6](#F6){ref-type=\"fig\"}C). In HIV-1-chronically-infected subjects, CD7^+^CD56^neg^CD16^+^ NK cells had a significantly higher frequency of CD107a expression in the unstimulated condition (Figure\u00a0[6](#F6){ref-type=\"fig\"}D). In contrast to their CD7^+^CD56^+^CD16^+^ NK cell counterpart, CD7^+^CD56^neg^CD16^+^ NK cells from HIV-1-chronically-infected subjects did not significantly degranulate following K562 stimulation (Figure\u00a0[6](#F6){ref-type=\"fig\"}C-D). A comparison of CD107a^+^ cells in CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in healthy and HIV-1-chronically-infected subjects did not result in any statistically significant differences. Taken together, CD7^+^CD56^neg^CD16^+^ NK cells from healthy controls were capable of degranulating following target cell stimulation. However, HIV-1 infection resulted in a significant defect in IFN\u03b3 secretion in both NK cell subsets and in degranulation in CD7^+^CD56^neg^CD16^+^ NK cells compared to CD7^+^CD56^+^CD16^+^ NK cells. Interestingly, HIV-1 infection also resulted in a higher frequency of IFN\u03b3 secretion and degranulation by CD7^+^CD56^neg^CD16^+^ NK cells in unstimulated conditions.\n\n![**Assessment of IFN and CD107a expression by CD7**^**+**^**CD56**^**+**^**CD16**^**+**^**and CD7**^**+**^**CD56**^**neg**^**CD16**^**+**^**NK cells following*in vitro*stimulation.** A-B IFN\u03b3 expression was compared between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in **(A)** healthy and **(B)** chronic HIV-1-infected donors following incubation in media alone, IL-12 plus IL-18, or K562 target cells. C-D, CD107a expression was compared between CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells in **(C)** healthy and **(D)** chronic HIV-1-infected donors following incubation in media alone or K562 target cells. A total of 10 healthy and 19 chronic HIV-1-infected donors were assessed. A minimum of 200 cells and 70 cells were analyzed for IFN\u03b3 and CD107 expression in CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells, respectively, of healthy donors. A minimum of 550 cells and 100 cells were analyzed for IFN\u03b3 and CD107 expression in CD7^+^CD56^+^CD16^+^ and CD7^+^CD56^neg^CD16^+^ NK cells, respectively, of HIV-1-infected donors. The median and 25th and 75th percentile are indicated on each graph. ( = CD7^+^CD56^+^CD16^+^ NK cells; = CD7^+^CD56^neg^CD16^+^ NK cells).](1742-4690-10-158-6){#F6}\n\nIncreased CD95 expression by CD7^+^CD56^neg^CD16^+^ NK cells\n------------------------------------------------------------\n\nGranzyme B and perforin expression are lower in unstimulated CD7^+^CD56^neg^CD16^+^ NK cells compared to CD7^+^CD56^+^CD16^+^ NK cells (Figure\u00a0[5](#F5){ref-type=\"fig\"}). Furthermore, unstimulated CD7^+^CD56^neg^CD16^+^ NK cells have significantly higher basal CD107a expression compared to CD7^+^CD56^+^CD16^+^ NK cells, particularly in HIV-1-infected subjects (Figure\u00a0[6](#F6){ref-type=\"fig\"}). These data suggest that CD7^+^CD56^neg^CD16^+^ NK cells might represent an activated subset of NK cells that have recently engaged a target. Previous studies have shown that resting NK cells express a basal level of the apoptosis-inducing receptor CD95 on their cell surface and that upon activation either through Fc receptor \\[[@B35]\\] or cytokine stimulation \\[[@B36]\\], CD95 is significantly upregulated on NK cells. Kottilil *et al.* recently demonstrated that CD95 expression is increased on NK cells in HIV-1-infected subjects predominantly within the CD56^+^CD16^+^ NK cell subset \\[[@B37]\\]. Indeed, we observed that CD95 is significantly increased on both subsets of NK cells in HIV-1-chronically-infected subjects compared to healthy controls, but is particularly increased on CD7^+^CD56^neg^CD16^+^ NK cells in both healthy controls and HIV-1-chronically-infected subjects (Figure\u00a0[7](#F7){ref-type=\"fig\"}). These data suggest NK cells are activated during HIV-1 infection and that CD7^+^CD56^neg^CD16^+^ NK cells might represent NK cells that have recently engaged a target cell.\n\n![**Assessment of NK cell activation in healthy and chronically HIV-1-infected subjects.** CD95 expression on CD7^+^CD56^neg^CD16^+^ compared to CD7^+^CD56^+^CD16^+^ NK cells in healthy (n\u2009=\u200911) and chronically HIV-1-infected subjects (n\u2009=\u200919). The median and 25th and 75th percentile are indicated.](1742-4690-10-158-7){#F7}\n\nDiscussion\n==========\n\nThe lineage and biological role of CD3^neg^CD56^neg^CD16^+^ cells has remained poorly understood. The lack of a truly NK cell-specific marker and changes in cell phenotype during HIV-1 infection, as well as the potential overlap of phenotypic and functional properties of NK cells with other immune cell subsets, creates difficulties in studying rare and non-classical NK cell subsets. In humans, NK cells are most commonly identified as CD3^neg^CD56^+^ lymphocytes with functionally distinct subsets being further defined by CD16 expression. However, CD16 is expressed on many other immune cells, including granulocytes and a subset of dendritic cells (DCs), designated slanDCs \\[[@B28]\\]. Importantly, slanDCs do not express CD56; however, they do express CD13, CD33, HLA-DR, and high surface expression of CD16, which is the phenotype we observed on the CD7^neg^CD56^neg^ population of cells. Therefore, caution is needed in interpreting the frequencies and functions of CD56^neg^ NK cells. Here, we built upon our previous work using CD7 as an informative marker of NK cells and demonstrated that CD3^neg^CD56^neg^CD16^+^ cells are a mixed population of CD7^+^ NK cells and CD7^neg^ myeloid cells present in both healthy and HIV-1-infected subjects. Previously published microarray analyses support this observation as the CD3^neg^CD56^neg^CD16^+^ population exhibited a transcriptome more closely related to myeloid cells than traditional CD3^neg^CD56^+^CD16^+^ NK cells \\[[@B25]\\]. Our study has provided a more detailed understanding of the phenotype, function, and possible origin during chronic HIV-1 infection, of CD7^+^CD56^neg^CD16^+^ NK cells in healthy and HIV-1-infected subjects.\n\nLittle is known regarding the phenotype and function of CD56^neg^CD16^+^ cells in healthy subjects and whether these cells are similar to those observed in chronic viral infections. In HIV-1-negative subjects, we observed minor alterations in KIR, NCR, or C-type lectin-like receptors between the two NK cell subsets. In contrast to a previous report indicating a reduced density of NKp30 on CD56^neg^CD16^+^ cells in healthy donors \\[[@B38]\\], we observed that CD7^+^CD56^neg^CD16^+^ NK cells actually had increased NKp30 expression. Functionally, CD7^+^CD56^neg^CD16^+^ NK cells had impaired cytokine responses following K562 target cell or IL-12 plus IL-18 stimulation; however, they degranulated (as measured by CD107a-expression) similar to CD7^+^CD56^+^CD16^+^ NK cells following K562 target cell stimulation. By removing CD7^neg^ non-NK cells, we demonstrated that CD7^+^CD56^neg^CD16^+^ NK cells do not have as significantly altered of a phenotype as previously shown. While our results confirm that CD56^neg^ NK cells do exist in healthy donors, using CD7 provided for a precise determination of the frequency of CD56^neg^ NK cells present in the blood of healthy donors and allowed for analyses of their phenotype and functional characteristics.\n\nNumerous studies have found decreases in both frequency and absolute numbers of NK cells defined as CD3^neg^CD56^+^ with a concomitant increase in a population of CD3^neg^CD56^neg^CD16^+^ cells in HIV-1 infection \\[[@B6]-[@B8]\\]. Our results confirm the expansion of a CD56^neg^ NK cell population; however, this population is not homogenous (\\<70% were CD7^+^CD56^neg^CD16^+^ NK cells). KIR expression was not significantly altered on either CD7^+^CD56^+^ or CD7^+^CD56^neg^ NK cells during HIV-1 infection compared to healthy donors. While HIV-1 infection was not associated with significant alterations in the frequency of cells expressing NKp30 or NKp46 in the CD7^+^CD56^+^ NK cell population, we did observe significant reductions in density of these receptors on CD7^+^CD56^neg^ NK cells. During HIV-1 infection, CD7^+^CD56^neg^ NK cells appeared to be mature; however, there were fewer CD57^+^ terminally differentiated CD7^+^CD56^neg^ compared to CD7^+^CD56^+^ NK cells. We recently demonstrated that CD57 is a marker of terminally differentiated NK cells \\[[@B21]\\], and that CD57^+^NKG2C^hi^ NK cells take weeks to develop *in vivo* following cytomegalovirus infection \\[[@B22]\\]. One possible explanation for the expansion of this CD7+CD56neg NK cell subset is that during HIV-1 infection, the already rapid turnover of NK cells \\[[@B20]\\] is further accelerated, preventing this subset of NK cells from terminally differentiating and gaining CD57 expression. Furthermore, our observation that CD7^+^CD56^neg^ NK cells express more of the apoptosis-inducing receptor CD95 further suggests these cells might undergo apoptosis prior to becoming terminally differentiated.\n\nAn alternative hypothesis is that the expanded population of CD7^+^CD56^neg^CD16^+^ NK cells during HIV-1 infection represents mature NK cells that arise at least in part when CD7^+^CD56^+^CD16^+^ NK cells engage target cells. NK cells are capable of killing multiple target cells and this results in a reduced, but never complete, loss of perforin and granzyme B \\[[@B39]\\]. Here, decreased granzyme B and perforin expression and increased expression of CD107a in the absence of *ex vivo* stimulation within the CD7^+^CD56^neg^CD16^+^ NK cells, particularly of HIV-1-infected subjects, suggests that CD7^+^CD56^neg^CD16^+^ NK cells may have recently engaged target cells *in vivo*. This is further supported by the increased expression of CD95 on CD7^+^CD56^neg^CD16^+^ NK cells, indicating these cells are more activated than their CD7^+^CD56^+^CD16^+^ NK cell counterparts.\n\nImportantly, the frequency of CD7^+^CD56^neg^CD16^+^ NK cells decreases *in vivo* during antiretroviral treatment \\[[@B9],[@B11],[@B40]\\] and *in vitro* with the addition of IL-2 \\[[@B12]\\]. Indeed, IL-2 has been shown to help recover perforin and granzyme B expression in NK cells that have killed target cells \\[[@B39]\\]. Furthermore, one study suggests that IL-2 may be beneficial in restoring CD56^+^CD16^+^ NK cells as they observed an increase in CD56^+^CD16^+^ NK cells and a reduction in CD56^neg^CD16^+^ NK cells *in vivo*\\[[@B41]\\]. The *in vivo* efficacy of IL-2 plus antiretroviral therapy requires further exploration. Another interesting approach is the combinatorial use of statins, particularly hydrophilic statins, with cytokine therapies. Statins are known to protect against oxidative stress \\[[@B42]\\] and hydrophilic statins do not inhibit NK cell cytotoxicity \\[[@B43]\\]. Such therapies may afford an individual increased control of viral replication prior to initiating antiretroviral therapy and provide increased clearance of viral reservoirs during antiretroviral therapy or in the event of drug resistance. Future studies investigating the effects of cytokine and statin therapies in chronic inflammatory diseases are needed to determine whether they help maintain a healthy NK cell repertoire.\n\nConclusion\n==========\n\nTaken together, CD56^neg^CD16^+^ NK cells are a mixed population of CD7^+^ NK cells and CD7^neg^ myeloid cells present at a low frequency in healthy donors irrespective of CMV serology status and expanded in HIV-1 viremic subjects. Our results indicate CD7^+^CD56^neg^CD16^+^ NK cells are activated, mature NK cells that may have recently engaged target cells. Further *in vivo* studies are needed to evaluate treatments for maintaining NK cell functionality in HIV-1 viremic subjects through the use of cytokine and statin therapies.\n\nMethods\n=======\n\nHuman subjects\n--------------\n\nDensity gradient centrifugation n over Ficoll-Paque (GE Healthcare) was used to obtain PBMC from leukocyte concentrates of healthy volunteers (Stanford Blood Center). Additionally, PBMC were obtained from participants in two San Francisco-based HIV-1-infected cohorts, OPTIONS that recruits subjects during early HIV-1 infection \\[[@B44]\\] and SCOPE that recruits subjects during chronic HIV-1 infection \\[[@B45]\\]. Table\u00a0[1](#T1){ref-type=\"table\"} summarizes the subjects used from each HIV-1 cohort and their respective CD4^+^ T cell counts and viral loads. All persons gave informed consent to participate in this study, and the University of California, San Francisco Committee on Human Research approved this study.\n\nPhenotypic and functional characterization of cell subsets\n----------------------------------------------------------\n\nWith the exception of the following antibodies, all fluorophore-labeled antibodies used for phenotypic analysis have been previously described \\[[@B23]\\]: phycoerythrin (PE)-conjugated anti-Siglec 7 (BioLegend) and PE-conjugated anti-LIR-1 (Beckman Coulter). All stains were performed in the presence of 100 \u03bcg/mL human IgG to block Fc receptors. No significant differences in isotype-matched control Ig staining were observed between CD7^+^CD56^+^CD16^+^ NK cells and CD7^+^CD56^neg^CD16^+^ NK cells. Cells were analyzed by flow cytometry with a four-laser LSR-II instrument (BD Biosciences, San Jose, CA) as previously described \\[[@B23]\\] and data analyses were carried out using FlowJo flow cytometric analysis software version 9.3.1 (Tree Star, Ashland, OR).\n\nNK cell stimulation and intracellular cytokine staining\n-------------------------------------------------------\n\nNK cell stimulation with K562 target cells or IL-12 plus IL-18, as well as intracellular IFN\u03b3 staining, was performed as previously described \\[[@B23]\\].\n\nStatistical analysis\n--------------------\n\nStatistical analyses were performed with GraphPad Prism software (GraphPad Software). The nonparametric Mann--Whitney *U* test was used to compare between-group distributions with a significance threshold set at *p*\u2009\\<\u20090.05.\n\nAbbreviations\n=============\n\nNK cell: Natural killer cell; HIV-1: Human immunodeficiency virus-1; DC: Dendritic cell; PBMC: Peripheral blood mononuclear cells.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nJMM designed and performed experiments, analyzed results, made figures, and wrote the manuscript. SL-V. designed and performed experiments and assisted in writing the manuscript. VAY performed experiments and analyzed results. SGD, FMH, and JNM provided reagents and critical revision of the manuscript. LLL and DFN. participated in study design, discussion of the results, and provided critical editing of the manuscript. All authors read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\n**Gating strategy to identify NK cell subsets in a representative healthy donor.** Single, live cells were gated on lymphocytes based on forward and side scatter parameters. CD3^neg^, CD14^neg^, and CD19^neg^ cells were gated and used to identify classically defined NK cells using CD56 and CD16 expression. To eliminate any potential contaminating myeloid cells, CD7 was assessed on each subset of classically defined NK cells (CD56^br^CD16^neg^ (green), CD56^dim^CD16^+^ (red) and CD56^neg^CD16^+^ (teal)). All three subsets contained CD7^neg^ cells; however, the CD56^neg^CD16^+^ subset contained the highest proportion of CD7^neg^ non-NK cells. Overlaying the CD7^neg^ non-NK cells onto the CD7^+^ NK cells indicates the high overlap within the subsets and the usefulness of CD7 as an additional marker of NK cells.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**CD7 gating allows precise identification of NK cells.** CD56^+^CD16^+^ (red) or CD56^neg^CD16^+^ (blue) cell subsets were assessed for NKp30 or granzyme B expression without gating on CD7^+^ cells (left panels) or after gating on CD7^+^ NK cell subsets (right panels).\n\n###### \n\nClick here for file\n\n###### Additional file 3\n\n**Comparison of NK cell gating strategies on NKp30 and NKp46 expression.** (A) NKp30 and (B) NKp46 expression were assessed on NK cells defined in three ways; (1) CD7^+^CD56^+^CD16^+^ NK cells, (2) total CD7^+^CD56^+^ NK cells inclusive of CD56^bright^CD16^neg^, CD56^dim^CD16^neg^ and CD56^dim^CD16^pos^ NK cells and (3) CD7^+^CD56^neg^CD16^+^ NK cells.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis work was supported by the National Institute Of Allergy And Infectious Diseases (grant P01-AI64520 to DFN and LLL; grant AI068129 to LLL; P30-AI27763 and R24-AI067039 to JNM; and P01-AI071713 to FMH and DFN). This research was supported, in part, by the Department of Health and Human Services funding under NIH Grant number 5T32HL007185 to JMM. SL-V was supported by a Cancer Research Institute/Irvington Institute postdoctoral fellowship. LLL is an American Cancer Society Professor. We thank the study participants who made this study possible. We also thank Gerald Spotts and Lisa Loeb for help in patient sample acquisition.\n"}
+{"text": "The past decade has seen big changes in the field of genomics, not only in terms of advances in technology, but also with regard to the views on sharing the data generated \\[[@B1],[@B2]\\]. Open data has become the buzzword of this age. No one can deny that this openness and willingness to share genomic data, both published and (perhaps more importantly) unpublished, has resulted in remarkable progress. However, when it comes to unpublished genomic data, this openness can also leave the data generators vulnerable. The community needs to balance the benefits of data sharing against the interests of the data owners, and usually the process works well.\n\nThe genomics community has measures in place to protect the data owners--data are often released under embargoes (of varying lengths, but usually not longer than 24 months) and data owners can also publish a 'statement of intent', i.e. outline the specific analyses they plan to undertake, when they release the data. There are also community norms--specifically the Bermuda rules \\[[@B3]\\], and the Fort Lauderdale \\[[@B4]\\] and Toronto \\[[@B5]\\] agreements--to help researchers navigate this rather sensitive issue. However, embargoes are not indefinite and neither does it seem fair to indefinitely prohibit specific analyses. It is also worth clarifying that the agreements mentioned above are so-called gentlemen's agreements, they are not law, and their utility depends on goodwill and communication within the community, not unlike attribution and the way scientists use citation to give credit.\n\nThe key words, as we see it, are community and communication. The researchers in the field are essentially in the same boat--they could be the data generators in one case and data users in another. Without communication the boat is likely to capsize. The data generators need to be clear in their intentions and in specifying any conditions that the data are released under, and the data users need to inform the data generators and seek permission to use the data if appropriate. Perhaps also there is a need to have enforceable guidelines in place rather than relying on gentlemen's agreements? The US National Institutes of Health (NIH) have already taken a step in this direction and have recently released a draft policy on the sharing of genomic data \\[[@B6]\\], which, if approved, will be applicable to all researchers who receive NIH funding. The guidelines cover, amongst other topics, the issue of when to release data; for raw sequence data from non-human organisms, the specified deadline is within 6 months of submission to an approved data repository.\n\nA question that follows is--whose responsibility is it to ensure that appropriate permission has been acquired to include the analysis of unpublished genomic data in a manuscript? Does the responsibility lie with the authors or the reviewers or with the journal editors? In our experience, such issues have usually been brought to light during the review process, but given the extensive amounts of data being generated, neither reviewers nor editors can be expected to be aware of the requirements for the use of each and every genome sequence. *BMC Genomics* has recently published a study by Zhao *et al*. \\[[@B7]\\], including an analysis of 103 fungal genomes. After publication it became apparent that some of these genomes were unpublished, and the authors had not informed the data owners of their intent of publishing an analysis of these genomes. Given this situation, we and the authors, in consultation with the data owners, agreed that a correction \\[[@B8]\\], whereby the authors would remove specific genomes from the analysis, was the appropriate way to proceed. In fact, only two of the disputed genomes were specifically under embargo, but after discussion with the data generators the authors agreed to remove from the analysis not only the embargoed genomes, but also an additional seven yet unpublished genomes.\n\nData generators, data users and journal editors all have a role to play in ensuring that the interests of all involved parties are protected, and as we have mentioned, the key to this is communication. We feel the ultimate responsibility should lie with the data user; it is up to them to ensure that they are aware of (and adhere to) any conditions set by the data generators. The latter could also make it easier for the data users by ensuring that the necessary information is readily available.\n\nThis is not to say that a journal has no responsibility however; a journal can increase awareness of the requirements in a field by incorporating guidance into their policies or instructions for authors. BioMed Central's editorial policies \\[[@B9]\\] now include a section on the use of unpublished genomic data: \"Authors using unpublished genomic data are expected to abide by the guidelines of the Fort Lauderdale and Toronto agreements. Based on broadly accepted scientific community standards, the key requirement for the third parties using genomic data is to contact the owners of unpublished data (i.e., the principal investigator and sequencing center) prior to undertaking their research, to advise them about their planned analyses.\" A journal is also, of course, responsible for taking the appropriate action when problems such as those exemplified by this case arise. Additionally, journal editors can facilitate communication between the concerned parties and help them arrive at a mutually satisfactory solution. Finally, a journal can instigate discussion on a topic or issue by bringing them to light--as we are doing by publishing this editorial.\n\nCompeting interests\n===================\n\nThe authors are employees of BioMed Central.\n\nAuthors' contributions\n======================\n\nBoth authors contributed to this editorial. Both authors read and approved the final text.\n\nAcknowledgements\n================\n\nWe would like to thank John Colbourne, Scott C. Edmunds, Amye Kennall, Elizabeth Moylan and Brian Oliver for their feedback and encouragement.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nThe most common etiology of spontaneous excess cortisol production is Cushing\\'s disease due to hypersecretion of pituitary adrenocorticotropic hormone (ACTH) by a corticotroph adenoma \\[[@B1]\\]. Cushing\\'s disease is a rare disorder, with an annual incidence of 1--3 per million and a female\u2009:\u2009male ratio of 3\u2009:\u20091, that is responsible for significant morbidity and mortality resulting from cardiovascular complications, infections, and psychiatric disturbances \\[[@B2]\\]. It is now recognized that, in a small proportion of cases, Cushing\\'s disease is accompanied by an elevated prolactin (PRL) level. For example, Lado-Abeal and colleagues identified a mildly elevated serum PRL in 20% of women with Cushing\\'s disease and menstrual abnormalities \\[[@B3]\\], and both Mahler et al. \\[[@B4]\\] and Sherry et al. \\[[@B5]\\] have presented a patient with Cushing\\'s disease and symptomatic hyperprolactinemia. Although Cushing\\'s disease with hyperprolactinemia due to mixed ACTH- and PRL-secreting adenomas occurs rarely, elevated preoperative PRL levels in Cushing\\'s disease are of diagnostic significance. Despite the abundance of published data concerning various aspects of the clinical course and treatment of Cushing\\'s disease, very few investigations have focused on the characteristics of patients with Cushing\\'s disease and hyperprolactinemia, and there is a paucity of studies comparing the clinical features of patients with Cushing\\'s disease and hyperprolactinemia and those with Cushing\\'s disease but no hyperprolactinemia. This is particularly the case for patients in China, about whom very little information is available concerning clinical outcomes following transsphenoidal surgery.\n\nThe aims of our investigation were to describe the prevalence and clinical features of Cushing\\'s disease in patients treated with transsphenoidal surgery at our hospital in China and to evaluate the differences between patients with Cushing\\'s disease and hyperprolactinemia and those with Cushing\\'s disease alone. We therefore performed a retrospective analysis of patients with Cushing\\'s disease, with or without concomitant hyperprolactinemia, to clarify the clinical features and confirm the efficacy of transsphenoidal surgery for treatment of the disease. This study provides novel insights that extend our current knowledge and will help to improve the diagnosis and treatment of patients with Cushing\\'s disease and hyperprolactinemia.\n\n2. Materials and Methods {#sec2}\n========================\n\n2.1. Patients {#sec2.1}\n-------------\n\nThis was a retrospective study of patients diagnosed with Cushing\\'s disease, who had been treated by primary transsphenoidal surgery between January 2002 and June 2011, at the Department of Neurosurgery, Shandong Provincial Hospital Affiliated to Shandong University, Jinan, Shandong, China. The diagnosis of Cushing\\'s disease was made on the basis of the medical history, symptoms, the levels of serum hormones, immunohistochemical staining, computer-aided tomographic (CT) imaging, and magnetic resonance imaging (MRI). Patients were divided into two groups, according to the preoperative hormonal levels and immunohistochemical results: a CD group (patients with elevated cortisol only) and a CD + PRL group (patients with both elevated cortisol and PRL levels). Due to the retrospective nature of this study, written informed consent was deemed not to be required. Oral informed consent was obtained from each participant, and the data were analyzed anonymously. The Ethics Committee of the Shandong Provincial Hospital Affiliated to Shandong University approved the research protocols.\n\nPatients were included in the study if all the following criteria were satisfied: classical signs and symptoms of Cushing\\'s disease; increased serum cortisol levels; loss of diurnal rhythm of serum cortisol; increased 24-hour urinary free cortisol levels; normal or slightly elevated plasma ACTH levels; lack of suppression of urinary free cortisol and serum cortisol following oral 2\u2009mg dexamethasone loading; suppression of urinary free cortisol and serum cortisol after oral 16\u2009mg dexamethasone (urinary free cortisol and/or serum cortisol suppressed by \\>50% of the baseline value following oral administration of 2\u2009mg dexamethasone every 6 hours for 48 hours); an adenoma staining positively for ACTH; the presence of a pituitary mass on MRI; treatment by surgery alone; and availability of a complete set of medical records. An additional inclusion criterion required for the CD + PRL group was an adenoma staining positive for PRL.\n\nThe exclusion criteria for the CD group were patients with hypersecretion of two or more hormones; adenomas with positive immunohistochemical staining for two or more hormones; adrenal-derived Cushing\\'s syndrome; ectopic Cushing\\'s syndrome; and other potential causes of an elevated cortisol level. The exclusion criteria for the CD + PRL group were patients with hypersecretion of two or more hormones (except for cortisol and PRL); adenomas with positive immunohistochemical staining for two or more hormones (except for ACTH and PRL); and other causes of elevated cortisol and PRL levels.\n\nMicroscopically assisted transsphenoidal pituitary tumor resection surgery was performed by two experienced neurosurgeons (Yuan-ming Qu and Guang-ming Xu), and the resection was considered subtotal or complete in all enrolled patients. In order to accurately assess the effects of surgery, only patients treated by surgery alone were included in this study.\n\n2.2. Data Extracted from the Medical Records {#sec2.2}\n--------------------------------------------\n\nThe data collected included the age and gender of the patient, past medical history, clinical presentation, preoperative neurological, endocrine, and ophthalmological status, and postoperative outcomes. The time from the onset of symptoms and signs to the date of diagnosis (time to diagnosis) was noted.\n\n2.3. Investigations {#sec2.3}\n-------------------\n\nThe tumor size was measured using MRI. A microadenoma was defined as a tumor with a diameter of 10\u2009mm or less, whereas a macroadenoma was defined as a tumor with a diameter of more than 10\u2009mm \\[[@B6]\\]. Representative T1-weighted coronal MRI scans showing an example of a microadenoma and a macroadenoma are shown in [Figure 1](#fig1){ref-type=\"fig\"}. Cavernous sinus invasion was suspected when adenoma tissue was located lateral to the cavernous portion of the internal carotid artery on MRI. Definite diagnosis could be made when the medial wall of the cavernous sinus was found to be destroyed during surgery by invasion \\[[@B7]\\]. The Knosp grade from the radiology findings was classified for degree of invasion. Pituitary imaging was also performed three months after surgery.\n\nEndocrinological assessment was carried out in all patients, including measurements of the serum levels of cortisol, PRL, growth hormone (GH), thyroid stimulating hormone (TSH), luteinizing hormone (LH), follicle stimulating hormone (FSH), testosterone, estradiol, thyroxine, and progesterone. In this study, cortisol was measured using an electrochemiluminescent immunoassay kit (Roche Diagnostics GmbH, Mannheim, Germany) with a functional sensitivity of \\<8.5\u2009nmol/L and intra-assay and interassay variation coefficients of \\<10% and \\<15%, respectively. The normal range for cortisol is 171--536\u2009nmol/L. The serum level of PRL was measured using a commercially available chemiluminescence kit (Beckman Coulter, Inc., Brea, CA, USA). The normal range for PRL is 2.64--13.13\u2009ng/mL in males, 3.34--26.72\u2009ng/mL in premenopausal females, and 2.74--19.64\u2009ng/mL in postmenopausal females. The analytical sensitivity for PRL was 0.25\u2009ng/mL, and the intra-assay and interassay variation coefficients were \\<10% and \\<15%, respectively. Serum levels of cortisol and PRL were evaluated in all patients on three occasions: 3 days after surgery, approximately 3 months after surgery, and 12 months after surgery.\n\nThe resected tumor fragments were fixed in 10% buffered formalin, stained with hematoxylin and eosin (H&E) using routine histochemical methods, and embedded in paraffin blocks. Immunohistochemical analysis was also performed on specimens, using specific antibodies against ACTH (rabbit polyclonal antibody; Maxin Co. Ltd., Fujian, China), GH (rabbit polyclonal antibody; Maxin Co. Ltd., Fujian, China), PRL (rabbit polyclonal antibody; Maxin Co. Ltd., Fujian, China), FSH (mouse monoclonal antibody; Maxin Co. Ltd., Fujian, China), LH (mouse monoclonal antibody; Maxin Co. Ltd., Fujian, China), and TSH (mouse monoclonal antibody; Maxin Co. Ltd., Fujian, China). An Olympus BX53 upright fluorescence microscope (Olympus Co. Ltd., Shinjuku, Japan) was used to view and record the images. Examples of images obtained using these pathological techniques are shown in [Figure 2](#fig2){ref-type=\"fig\"} (original magnification \u00d7400).\n\n2.4. Postoperative Complications {#sec2.4}\n--------------------------------\n\nPostoperative complications were divided into major and minor categories. A major complication was considered to be a disease or injury diagnosed after surgery that was potentially fatal and/or resulted in a permanent defect. Less serious adverse events, including diabetes insipidus (DI) and hyponatremia, were counted as minor complications. Urine volume measurements and serum electrolyte analyses were carried out for all patients. Serum levels of sodium were measured immediately before surgery and on days 1, 3, and 7 after surgery; hyponatremia was diagnosed when the serum sodium level was \u2264135\u2009mmol/L, and, when present, its onset, frequency, and time course were monitored. Central DI includes both transient DI and permanent DI \\[[@B8]\\]; transient DI is diagnosed when hypotonic polyuria (40\u2009mL/kg body weight daily) ensues soon after surgery and is usually self-remitting within a few days. Since permanent DI requires assessment of endocrine function several months after surgery, the incidence of DI in this study may only represent transient DI.\n\n2.5. Follow-Up {#sec2.5}\n--------------\n\nAll patients were followed up as outpatients for three months after surgery and then at annual or biannual intervals. Follow-up information for patients not attending our department was obtained by contacting the patients or their relatives by telephone or mail. Normalization was defined as a return of hormone levels to within the normal range without the requirement of any further treatment. Recurrence was diagnosed from the clinical features, hormonal function, or identification of residual tumor, based on the results of hormone and MRI examinations performed two years after surgery.\n\n2.6. Statistical Analysis {#sec2.6}\n-------------------------\n\nQuantitative data are reported as the mean \u00b1 standard deviation (SD), and qualitative data are expressed as percentages. Quantitative and nonparametric data, such as mean age, time to diagnosis, and tumor size, were compared by Student\\'s *t*-tests or Mann-Whitney *U* tests. Chi-square tests were used to compare qualitative data, including gender, invasiveness, the presence of pituitary apoplexy, tumor classification, remission rate, and recurrence rate. Statistical comparisons of endocrine outcomes at different follow-up times were adjusted for multiple testing using a Bonferroni correction. Differences between groups in endocrine outcome at the same follow-up time were assessed by Student\\'s *t*-test. Correlations between variables were evaluated using Pearson or Spearman correlation coefficients. Analyses were performed using Statistical Package for Social Sciences Version 16.0 (Chicago, IL, USA). All statistical tests were two-sided, and *P* \\< 0.05 was considered statistically significant.\n\n3. Results {#sec3}\n==========\n\n3.1. Baseline Characteristics {#sec3.1}\n-----------------------------\n\nA total of 84 patients were enrolled in the study, with a mean age of 33.60 \u00b1 12.70 years (range: 17--75 years); more than half the patients were female (*n* = 59; 70.2%). The demographics and tumor characteristics of the two groups of patients are summarized in Tables [1](#tab1){ref-type=\"table\"} and [2](#tab2){ref-type=\"table\"}.\n\nOf the 84 patients enrolled, 48 (15 males and 33 females) were categorized into the CD group and 36 (10 males and 26 females) into the CD + PRL group. At the time of diagnosis, the patients in the CD + PRL group (mean age: 30.28 \u00b1 14.23 years; range: 17--75 years) were significantly younger than those in the CD group (mean age: 36.08 \u00b1 10.91 years; range: 23--65 years) (*P* = 0.037). Furthermore, maximal adenoma size was larger for patients in the CD + PRL group (mean diameter: 2.44 \u00b1 1.32\u2009cm; range: 0.6--6.0\u2009cm) than for those in the CD group (mean diameter: 1.44 \u00b1 1.05; range: 0.5--4.0\u2009cm) (*P* \\< 0.001). The mean time to diagnosis did not differ significantly between the CD group (25.32 \u00b1 37.98 months; range: 0.1--180 months) and the CD + PRL group (33.68 \u00b1 34.96 months; range: 0.3--120 months), and there were also no significant differences in the incidences of invasiveness or pituitary apoplexy between the two groups. The patients were classified according to the degree of invasion of the cavernous sinus according to Knosp classification, and the two groups showed significant differences with the CD + PRL group showing a tendency for higher grades reflecting higher rates of invasion (*P* = 0.005).\n\n3.2. Tumor Classification {#sec3.2}\n-------------------------\n\nIn the CD group, 47.9% of the patients had microadenomas, and 52.1% had macroadenomas. The corresponding values for patients in the CD + PRL group were 5.6% and 94.4%, indicating that macroadenomas were more prevalent in the CD + PRL group than in the CD group (*P* \\< 0.001).\n\n3.3. Preoperative Clinical Manifestations {#sec3.3}\n-----------------------------------------\n\nThe most commonly presenting clinical manifestations in the two groups were menstrual disorders, headaches, and dizziness ([Table 3](#tab3){ref-type=\"table\"}). Menstrual disorders and visual field defects were more common in patients in the CD + PRL group than in patients in the CD group (*P* = 0.027 and 0.021, resp.), whereas the incidences of progressive obesity and hypertension were much lower in patients in the CD + PRL group than in patients in the CD group (*P* = 0.009 and *P* \\< 0.001, resp.).\n\n3.4. Endocrine Function before and after Surgery {#sec3.4}\n------------------------------------------------\n\nPreoperative and postoperative serum cortisol and PRL levels were available for all patients ([Table 4](#tab4){ref-type=\"table\"}). Elevated preoperative levels of serum cortisol and PRL were detected in patients in the CD + PRL group, and these patients had lower cortisol levels and higher PRL levels than those in the CD group (*P* \\< 0.05 for both comparisons). The cortisol level of patients in the CD group decreased from 833.87 \u00b1 235.75\u2009nmol/L before surgery to 336.87 \u00b1 267.56\u2009nmol/L at 3 days, 320.64 \u00b1 213.54\u2009nmol/L at 3 months, and 332.01 \u00b1 229.81\u2009nmol/L at 12 months after surgery. The preoperative and postoperative PRL levels in the CD group were all within the normal range. In the CD + PRL group, the cortisol level decreased from 673.24 \u00b1 65.53\u2009nmol/L before surgery to 375.83 \u00b1 187.51\u2009nmol/L at 3 days, 385.24 \u00b1 186.57\u2009nmol/L at 3 months, and 368.33 \u00b1 170.87\u2009nmol/L at 12 months after surgery. Furthermore, the PRL levels of this group decreased from 233.63 \u00b1 188.06\u2009ng/mL before surgery to 72.63 \u00b1 66.94\u2009ng/mL at 3 days, 74.51 \u00b1 62.58\u2009ng/mL at 3 months, and 77.43 \u00b1 70.38\u2009ng/mL at 12 months after surgery. There were no significant differences in the postoperative cortisol levels between the CD and CD + PRL groups, but the postoperative PRL levels in the CD + PRL group remained markedly higher than those in the CD group.\n\nAmong the patients with microadenomas, the normalization rates (based on the postoperative hormone levels at three months) in the CD and CD + PRL groups were 95.7% and 100%, respectively ([Figure 3(a)](#fig3){ref-type=\"fig\"}). For patients with macroadenomas, the corresponding normalization rates were 88.0% and 38.2%, respectively. Overall, normalization of hormonal levels was achieved in 44 patients (91.7%) in the CD group, but only 15 patients (41.7%) in the CD + PRL group (*P* \\< 0.001). In the CD + PRL group, normalization of both cortisol and PRL levels occurred in 41.7% of the patients ([Figure 3(b)](#fig3){ref-type=\"fig\"}).\n\n3.5. Postoperative Complications {#sec3.5}\n--------------------------------\n\nNo deaths or major complications were observed, although some minor complications did occur, including transient DI and hyponatremia. Transient DI was observed in 26 of 48 patients (54.2%) in the CD group and in 18 of 36 patients (50.0%) in the CD + PRL group. Hyponatremia occurred in 50% of the patients in both groups, with nadir serum sodium levels in the CD group (113--134\u2009mmol/L) similar to those in the CD + PRL group (109--130\u2009mmol/L). There were no significant differences in the incidences of transient DI and hyponatremia between the two groups.\n\n3.6. Follow-Up {#sec3.6}\n--------------\n\nThe median postoperative duration of hospitalization was 5 days (range: 3--15 days). Overall, 90% of the patients were discharged by postoperative day 5. The median duration of follow-up for the entire group was 45 months (range: 13--121 months). Tumor recurrence was significantly less frequent in the CD group (4 patients, 8.3%) than in the CD + PRL group (13 patients; 36.1%) (*P* \\< 0.001). All recurrences arose within the first two years after surgery.\n\n4. Discussion {#sec4}\n=============\n\nCushing\\'s disease, caused by excessive ACTH secretion from tumorous pituitary corticotrophs, is a potentially life-threatening endocrine condition \\[[@B9]\\]. The first-line treatment for Cushing\\'s disease is pituitary surgery to control excessive hormone production and improve pituitary function \\[[@B10]\\]. Cases of pituitary adenoma associated with increased production of both ACTH and PRL, causing apparent Cushing\\'s disease and hyperprolactinemia, are extremely rare \\[[@B5], [@B11]\\]. The present study has evaluated the clinical features and endocrine and surgical outcomes, over a relatively long follow-up period, of a large series of patients who underwent transsphenoidal surgery for Cushing\\'s disease with or without hyperprolactinemia.\n\nThe majority of our patients were diagnosed in their thirties and forties, with those in the CD + PRL group significantly younger at diagnosis than those in the CD group. In addition, our results revealed that the mean maximal diameter of the adenoma varied between the two groups, consistent with a previous report that patients with Cushing\\'s disease alone usually present with small tumors \\[[@B12]\\]. The CD + PRL group also had a tendency for higher Knosp grades so the adenoma was more likely to have invaded the cavernous sinus space. Approximately half of the patients in the CD group were found to have microadenomas, and the other half were found to have macroadenomas, a finding similar to that of others \\[[@B13]\\]. Although recent advances in imaging technologies, including MRI, can detect relatively small lesions within the pituitary gland, these methods can localize microadenomas in only 60--80% of patients with Cushing\\'s disease, since most of the lesions are very small \\[[@B8], [@B14]\\]. In our study, the average diameter of the adenoma was less than 5 to 6\u2009mm, in agreement with previous investigations \\[[@B8], [@B15]\\]. We also found that nearly 95% of patients in the CD + PRL group had macroadenomas; this higher incidence may be related to the higher levels of PRL.\n\nMenstrual disorders and visual field defects were more common in patients in the CD + PRL group (compared with the CD group), whereas the incidences of progressive obesity and hypertension were much lower. Hypersecretion of PRL has a variety of manifestations, including menstrual disorders in women \\[[@B16]\\], and our findings concurred with those of previous investigations \\[[@B17]\\]. Pituitary compression by a tumor may present with visual field defects; furthermore, patients in the CD + PRL group were much more likely to have macroadenomas, which are more commonly associated with the development of neurological deficits and visual disturbances. ACTH has antagonistic effects on protein and lipid metabolism that can lead to uncontrolled hypertension and other serious complications \\[[@B18]\\]. Overall, the clinical symptoms observed in our study agreed with those of previous reports in the literature \\[[@B19]--[@B21]\\]. Of the patients in the CD + PRL group who had hormonal symptoms, most were related to plurihormone production.\n\nElevated preoperative cortisol and PRL levels were detected in patients in the CD + PRL group, with PRL levels improving after surgery; this is consistent with the findings of Ratliff and Oldfield \\[[@B22]\\] and Wynne et al. \\[[@B23]\\]. We also found that patients in the CD + PRL group with high PRL levels often had a relatively low serum cortisol level. Yamaji and colleagues reported PRL elevation in 23% of patients with Cushing\\'s disease \\[[@B24]\\], while Caufriez and coworkers found that 91% of patients with Cushing\\'s disease had preoperative elevation of PRL levels that normalized after transsphenoidal surgery \\[[@B25]\\]. Although Wynne et al. reported that the predictive value of preoperative PRL levels in Cushing\\'s disease was limited \\[[@B23]\\], we found that patients in the CD group were more likely to achieve normalization of endocrine function. It was notable that, in the CD + PRL group, the normalization rate for patients with macroprolactinomas was significantly lower than that for patients with microprolactinomas. Since complete resection of microadenomas was possible, it is perhaps not surprising that the normalization rates for patients with microadenomas exceeded 95% for both groups.\n\nOur results, together with those of previous investigations, indicate that surgery is a highly effective treatment for Cushing\\'s disease, and transsphenoidal surgery is considered the definitive treatment method \\[[@B26]\\]. Thus, the discovery of patients with a plurihormonal tumor depends on the hormonal levels as well as pathological identification \\[[@B27], [@B28]\\]. Plurihormonal adenomas are defined as those that secrete two or more hormones that differ in chemical composition, immunoreactivity, and biologic effects \\[[@B29], [@B30]\\]. In every patient in the CD + PRL group, a mixed tumor with ACTH- and PRL-positive cells was discovered by pathological investigations following surgery.\n\nTranssphenoidal surgery represents a relatively safe procedure for resecting tumors underlying Cushing\\'s disease. However, even in the most experienced hands, postoperative complications are unavoidable. The occurrences of postoperative transient DI and hyponatremia correlate with intraoperative manipulation of the pituitary stalk and posterior lobe, multiple incisions in the gland, and partial hypophysectomy \\[[@B31]\\]. Since pituitary exploration may damage the normal gland, gland manipulation is always kept to the minimum necessary to identify and remove the adenoma. The rate of postoperative transient DI reported in the literature varies widely \\[[@B32], [@B33]\\]. Although there was no significant difference in the incidence of DI between the two groups in our study, relatively higher rates of DI were observed in our patients overall, likely due to an increased awareness regarding this potential complication. There was also no significant difference in the incidence of hyponatremia between the two groups. Future prospective studies of the impact of transsphenoidal surgery on transient DI and electrolyte abnormalities are merited.\n\nWide variations in surgical outcomes and recurrence rates have been reported, depending on the tumor characteristics, the surgeon\\'s experience, and the duration of the follow-up. The recurrence rate in the CD group was 8.3%, lower than that reported by others \\[[@B34], [@B35]\\], whereas it was 36.1% in the CD + PRL group. The lower recurrence rate in the CD group may reflect long-term remission. The elevated rate of recurrence in the CD + PRL group is an important finding, which suggests that careful long-term follow-up is needed in this subset of patients, even if surgery is considered successful.\n\nMixed ACTH and PRL adenomas often manifest themselves as coexisting Cushing\\'s disease and prolactinoma \\[[@B5]\\]. Several hypotheses have been proposed to explain the coexistence of two or more hormones within an adenoma. First, neoplastic transformation caused by stimulation could result in different cell types \\[[@B36]\\]. Second, different hormones could be derived from distinct cell types within the borders of a single adenoma \\[[@B4], [@B37]\\]; for example, two separate tumors growing in a small space could become conjoined as a single mass.\n\nOur study is not without its limitations. First, this was a retrospective cohort study in a single institution; prospective, multicenter studies are merited to extend our observations. Second, the follow-up period was relatively short for some patients, limiting the evaluation of the real recurrence rates. Hence, long-term postoperative follow-up of a similar cohort of patients is needed to better assess the effects of surgery and treatment outcomes.\n\n5. Conclusion {#sec5}\n=============\n\nIn summary, the current study has compared characteristics between patients with both Cushing\\'s disease and hyperprolactinemia and those with Cushing\\'s disease alone. We found that patients in the CD + PRL group were characterized by younger age, larger tumor size, more clinical manifestations (including a higher incidence of menstrual disorders caused by PRL secretion), lower endocrine normalization rate, and more frequent recurrence rate, compared with patients with Cushing\\'s disease alone. Our observations extend our knowledge of the characteristics of patients with Cushing\\'s disease and hyperprolactinemia and may provide insights that will help to improve the diagnosis and treatment of the disease.\n\nthe authors would like to thank the entire staff of the Department of Neurosurgery of the Provincial Hospital Affiliated to Shandong University for their help.\n\nConflict of Interests\n=====================\n\nThe authors declare no conflict of interests regarding the publication of this paper.\n\nAuthors' Contribution\n=====================\n\nCheng Huan and Chao Lu contributed equally to this work.\n\nFunding\n=======\n\nThis work was supported by the Chinese National Natural Science Foundation (81372473) and Shandong Provincial Scientific and Technological Development Projects Foundation (2010GSF10225).\n\n![Representative T1-weighted coronal MRI scans. (a) Microadenoma (arrow). (b) Macroadenoma expanding into the suprasellar cistern (arrow).](IJE2014-919704.001){#fig1}\n\n![Surgical pathology analyses of the tumors. (a) H&E staining of tumor tissue from a patient in the CD group. (b) H&E staining of tumor tissue from a patient in the CD + PRL group. (c) Immunohistochemistry: cells of the adenoma from a patient in the CD group were immunoreactive to antibodies against ACTH. (d) Immunohistochemistry: cells of the adenoma from a patient in the CD + PRL group were immunoreactive to antibodies against ACTH (d1) and PRL (d2).](IJE2014-919704.002){#fig2}\n\n![The rate of hormonal normalization in patients with Cushing\\'s disease. (a) Comparisons of the normalization rates (%) after transsphenoidal surgery between microadenomas and macroadenomas, for patients in the CD and CD + PRL groups. (b) Distribution of the hormonal normalization rate in the CD + PRL group. \"H\" indicates hormone levels above normal after surgery; \"N\" indicates hormone levels in the normal range after surgery.](IJE2014-919704.003){#fig3}\n\n###### \n\nPreoperative demographic characteristics and tumor sizes of the patients in the two groups.\n\n  Variable                     CD (*n* = 48)   PRL + CD (*n* = 36)   *P*\n  ---------------------------- --------------- --------------------- ---------\n  Age (years)                  36.08 \u00b1 10.91   30.28 \u00b1 14.23         0.037\n  Time to diagnosis (months)   25.32 \u00b1 37.98   33.68 \u00b1 34.96         0.177\n  Tumor size (cm)              1.44 \u00b1 1.05     2.44 \u00b1 1.32           \\<0.001\n\nData are expressed as the mean \u00b1 SD. *P* values were determined using a Student\\'s *t*-test or a Mann-Whitney *U* test, as appropriate. *P* \\< 0.05 was considered statistically significant.\n\n###### \n\nGender, tumor invasiveness, and the presence of pituitary apoplexy in the two groups of patients.\n\n  Variable               CD (*n* = 48)   PRL + CD (*n* = 36)   *P*\n  ---------------------- --------------- --------------------- -------\n  Gender                 \u2009               \u2009                     \u2009\n  \u2003Male                  15 (31.3%)      10 (27.8%)            0.461\n  \u2003Female                33 (68.8%)      26 (72.2%)            \n  Invasion               \u2009               \u2009                     \u2009\n  \u2003+                     11 (22.9%)      12 (33.3%)            0.289\n  \u2003\u2212                     37 (77.1%)      24 (66.7%)            \n  Apoplexy               \u2009               \u2009                     \u2009\n  \u2003+                     9 (18.8%)       6 (16.7%)             0.805\n  \u2003\u2212                     39 (81.3%)      30 (83.3%)            \n  Classification Knosp   \u2009               \u2009                     \u2009\n  \u20030                     25 (52.1)       7 (19.4)              0.005\n  \u20031                     7 (14.6)        9 (25.0)              \n  \u20032                     6 (12.5)        7 (19.4)              \n  \u20033                     8 (16.7)        8 (22.2)              \n  \u20034                     2 (4.2)         5 (13.9)              \n\nData are expressed as *n* (%). \"+\" and \"\u2212\" indicate the presence and absence, respectively, of the characteristic in the group. *P* values were determined using a chi-square test or a Mann-Whitney *U* tests. *P* \\< 0.05 was considered statistically significant.\n\n###### \n\nPreoperative clinical manifestations in the two groups of patients.\n\n  Symptom                CD (*n* = 48)   PRL + CD (*n* = 36)   *P*\n  ---------------------- --------------- --------------------- ---------\n  Headache/dizziness     27 (56.3%)      24 (66.7%)            0.333\n  Progressive obesity    21 (43.8%)      6 (16.7%)             0.009\n  Vomiting               4 (8.3%)        4 (11.1%)             0.720\n  Visual impairment      19 (39.6%)      18 (50.0%)            0.341\n  Visual field defects   6 (12.5%)       12 (33.3%)            0.021\n  Menstrual disorders    16/28 (57.1%)   22/26 (84.6%)         0.027\n  Galactorrhea           10/28 (35.7%)   14/26 (53.8%)         0.180\n  Sexual dysfunction     4/10 (40.0%)    2/20 (10.0%)          0.141\n  Polyuria/polydipsia    12 (25.0%)      8 (22.2%)             0.767\n  Hypertension           40 (83.3%)      10 (27.8%)            \\<0.001\n  Diabetes mellitus      5 (10.4%)       6 (16.7%)             0.523\n\nData are expressed as *n* (%). *P* values were determined using a chi-square test. *P*\\< 0.05 was considered statistically significant.\n\n###### \n\nThe levels of cortisol (nmol/L) and PRL (ng/mL) measured during follow-up of the patients in the two groups.\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------\n  Group        Hormone           Before surgery        3 days after surgery   3 months after surgery   12 months after surgery   *F*       *P*\n  ------------ ----------------- --------------------- ---------------------- ------------------------ ------------------------- --------- ---------\n  CD\\          Cortisol          833.87 \u00b1 235.75^\\*^   336.87 \u00b1 267.56        320.64 \u00b1 213.54          332.01 \u00b1 229.81           54.098    \\<0.001\n  (*n* = 48)                                                                                                                               \n\n  PRL          13.77 \u00b1 6.87^\u25b3^   11.12 \u00b1 6.71^\u25b3^       12.93 \u00b1 6.57^\u25b3^        12.66 \u00b1 6.20^\u25b3^          1.345                     0.305     \n\n                                                                                                                                           \n\n  PRL + CD\\    Cortisol          673.24 \u00b1 65.53        375.83 \u00b1 187.51        385.24 \u00b1 186.57          368.33 \u00b1 170.87           30.713    \\<0.001\n  (*n* = 36)                                                                                                                               \n\n  PRL          233.63 \u00b1 188.06   72.63 \u00b1 66.94         74.51 \u00b1 62.58          77.43 \u00b1 70.38            18.64                     \\<0.001   \n  --------------------------------------------------------------------------------------------------------------------------------------------------\n\nData are expressed as the mean \u00b1 SD. \u2217 indicates a significant difference in the cortisol level between the two groups at the same time point (*P* \\< 0.05). \u25b3 indicates a significant difference in the PRL level between the two groups at the same time point (*P* \\< 0.05). Each *P* value shown in the Table is for comparative analysis, within the CD or PRL + CD group, between values before surgery and those 3 days after surgery.\n\n[^1]: Academic Editor: Ludwik K. Malendowicz\n"}
+{"text": "Introduction {#s1}\n============\n\nHormones co-ordinate developmentally programmed processes as well as responses that are induced by external stimuli. Although different hormones govern distinct biological processes, the final outcome is often the result of complex interactions among multiple hormone pathways. Deciphering the molecular mechanism of this crosstalk has been an important trend in plant hormone research ([@CIT0039]; [@CIT0050]). This includes studies of auxin and jasmonate (JA), which carry out many indispensable functions throughout a plant's life cycle. Auxin is necessary for the proper development of embryos, roots, and shoots, and is also well known for its role in gravitropism and phototropism ([@CIT0039]; [@CIT0030]). JA, on the other hand, is best known for its role in plant resistance to insects and fungal pathogens ([@CIT0017]; [@CIT0050]).\n\nMany cases of crossover between JA and auxin signaling pathways have been detected, at the level of both gene expression and metabolism ([@CIT0036]; [@CIT0050]). For example, JA inhibits primary root growth in Arabidopsis through repressing the expression of *PLETHORA* (*PLT1*) and *PLT2*, which are key transcription factors (TFs) of the auxin-regulated root meristem specification and maintenance ([@CIT0005]). JA can promote auxin biosynthesis through transcriptional activation of *ANTHRANILATE SYNTHASE a1* (*ASA1*) and *ASB1* encoding enzymes in the [l]{.smallcaps}-tryptophan biosynthetic pathway, which provides precursors for auxin biosynthesis ([@CIT0043]). Transcriptional regulation of two auxin biosynthetic genes, *YUCCA8* and *YUCCA9*, is also linked to JA-induced auxin biosynthesis ([@CIT0015]). On the other hand, auxin can induce the expression of JA biosynthetic genes ([@CIT0047]), and, in flowers, auxin-responsive TFs, AUXIN RESPONSE FACTOR 6 (ARF6) and ARF8, promote fertility through activation of JA biosynthesis ([@CIT0044]). Auxin-inducible acyl amidosynthetases, Gretchen Hagen 3.3 (GH3.3), GH3.5, and GH3.6, participate in adventitious root initiation by modulating JA homeostasis ([@CIT0012]).\n\nMolecular perception of auxin and JA is linked to transcriptional regulation of downstream genes through 26S proteasome-dependent protein degradation mechanisms ([@CIT0046]; [@CIT0039]). The two mechanisms bear a remarkable resemblance to each other, including nuclear-residing three-component receptor complexes consisting of a hormone ligand (auxin or JA), an F-box protein, and a transcriptional repressor ([@CIT0007]; [@CIT0045]; [@CIT0046]; [@CIT0036]). In the absence of the hormones, transcription is repressed by repressor proteins, Aux/IAA (for auxin) or JASMONATE ZIM-DOMAIN (JAZ) (for JA). Rising levels of auxin or JA promote recruitment of Aux/IAA or JAZ by the SKP1/Cullin/F-box (SCF) E3 ubiquitin ligase complex, which is shared by the two hormones, except for the F-box protein which is unique for each hormone (TIR1/AFB for auxin and COI1 for JA). The ensuing events of ubiquitination and degradation of Aux/IAA or JAZ proteins unleash TFs for transcription of auxin- or JA-responsive genes. Protein interaction studies have established the strict structural requirements for the auxin and JA molecules to be able to act as ligands for binding and formation of the tertiary receptor complexes ([@CIT0045]; [@CIT0046]). For auxin, the free carboxylic acid form, indole-3-acetic acid (IAA), and its structurally related receptor agonists, and, for JA, an amino acid-conjugated form, jasmonoyl-isoleucine (JA-lle), and its close structural variants were identified as the most effective ligands. Genetic evidence supports such conclusions ([@CIT0042]; [@CIT0041]).\n\nBecause IAA and JA-Ile have direct signaling roles with global consequences, their levels must be precisely regulated. Along with biosynthesis and catabolism, conjugation to amino acids is a common way of controlling hormone levels by converting active signaling forms to inert forms for storage, transport, or degradation. Amino acid conjugates of auxin and JA occur in most plant species investigated so far ([@CIT0030]; [@CIT0050]). The conjugation step is catalyzed by members of the GH3 family proteins ([@CIT0013]; [@CIT0041]). Several enzymes in Group II of the 19 Arabidopsis GH3 enzymes catalyze IAA conjugation, whereas JA-Ile is formed by JASMONATE RESISTANT1 (JAR1) in Group I of the same family ([@CIT0042]; [@CIT0041]). The reverse reaction of IAA conjugation to amino acids is catalyzed by enzymes belonging to the IAA amidohydrolase (IAH) family consisting of seven members in Arabidopsis that have previously been studied with respect to auxin metabolism ([@CIT0001]; [@CIT0009]; [@CIT0030]). Of the seven members, IAA-LEU RESISTANT1 (ILR1), ILR1-like (ILL) 1, ILL2, and IAA-ALA RESISTANT3 (IAR3) were found to have catalytic activity towards various IAA--amino acid conjugates ([@CIT0028]). Both *in vitro* and genetic evidence support ILR1, the founding member of the IAH family, to be an IAA-Leu hydrolase ([@CIT0001]), and IAR3, the most evolutionarily conserved member ([@CIT0004]), to hydrolyze mainly IAA-Ala ([@CIT0009]). No *in vitro* activity against IAA conjugates was observed with ILL3, ILL5, or ILL6 ([@CIT0028]).\n\nSigns of connection between *IAH* genes and JA signaling were first detected with the expression of *JR3*, later reported to be identical to *IAR3* ([@CIT0009]), which was found to be induced by JA and wounding ([@CIT0048]). Recently, three independent groups have reported the role of IAR3 and ILL6 in JA metabolism ([@CIT0053]; [@CIT0002]; [@CIT0052]). The first among these studies showed that a *Nicotiana attenuata* homolog of IAR3 hydrolyzed JA-Ile *in vitro*, and that when the gene was silenced in *N. attenuata* the endogenous JA-Ile level increased ([@CIT0053]). The *ILL6* gene was picked up as the top candidate regulator of the JA pathway by a novel gene expression network analysis ([@CIT0002]). The knock-out mutants of *ill6* displayed decreased capacity to release isoleucine from exogenously applied radioisotope-labeled JA-Ile and increased endogenous JA-Ile levels in wounded leaves, consistent with the role of ILL6 as a JA-Ile hydrolase. Recombinant IAR3 and ILL6 proteins expressed in bacteria were able to hydrolyze JA-Ile; IAR3 additionally hydrolyzed 12-hydroxy-JA-Ile (12OH-JA-Ile) ([@CIT0052]), a major JA metabolite formed by oxidative JA-Ile catabolism ([@CIT0023]). Consistently, 12OH-JA-Ile levels were increased in both *iar3* and *ill6* mutants ([@CIT0052]; [@CIT0025]).\n\nHere we describe the further characterization of ILL6, IAR3, and a third JA-inducible IAH family member, ILR1, for their function in JA and auxin metabolism and signaling. Catalytic activities of recombinant proteins were tested for a variety of different JA and IAA conjugates. ILR1, ILL6, and IAR3 were overexpressed in Arabidopsis and the resulting plants were analyzed for their endogenous hormone profile and for phenotypes associated with JA and auxin signaling. Higher order mutants were created to examine functional redundancies between IAHs. Subcellular localization was determined for ILL6 and IAR3 proteins. These results support a hormone crosstalk model where wound-inducible amidohydrolases simultaneously regulate the levels of JA and auxin to co-ordinate stress responses.\n\nMaterials and methods {#s2}\n=====================\n\nPlant material, growth conditions, and treatments {#s3}\n-------------------------------------------------\n\n*Arabidopsis thaliana* ecotype Col-0 was used as the wild type (WT) for all experiments, except for [Fig. 2D](#F2){ref-type=\"fig\"}--[F](#F2){ref-type=\"fig\"} where Wassilewskija (WS) was used instead. The *ill6-2iar3-5* double mutant was made by a genetic crossing between the T-DNA insertion lines CS852193 (*ill6-2*) ([@CIT0002]) and SALK_069047 (*iar3-5*) obtained from the Arabidopsis Biological Resource Center. *ilr1-1*, *ilr1-5* ([@CIT0038]), *iar3-2*, *iar3-2ilr1-1* ([@CIT0009]), and *ILR1-OE* ([@CIT0009]) seeds were kindly provided by Dr Bonnie Bartel (Rice University). A complete list of oligonucleotide primers used for genotyping is given in [Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) available at *JXB* online. Plants were grown in environmental growth chambers maintained at 22 \u00b0C with a photoperiod of 16h light (100--120 \u03bcE m^\u22122^ s^\u22121^). Hormone treatments for gene expression studies were carried out either by spraying an indicated amount of hormones evenly onto the surface of fully expanded leaves of soil-grown plants or by growing seedlings on plates containing the hormone. Wounding was conducted on fully expanded rosette leaves of 4- to 5-week-old soil-grown plants by crushing across the midrib twice using a hemostat with serrated tips ([@CIT0016]).\n\nChemicals {#s4}\n---------\n\n(\u00b1)-Jasmonic acid, methyl jasmonic acid (MeJA), coronatine, IAA, IAA-Ala, and d~5~-IAA were purchased from Sigma-Aldrich. JA--amino acid conjugates, 12OH-JA, 12OH-JA-Ile, 12COOH-JA-Ile, 12-*O*-Glc-JA, and 12-*O*-Glc-JA-Ile, were synthesized according to protocols described previously ([@CIT0027]; [@CIT0042]; [@CIT0008]; [@CIT0022]; [@CIT0019]). IAA-Leu was synthesized according to methods described for JA conjugates ([@CIT0042]) except that tetrahydrofuran was replaced with acetonitrile as the solvent.\n\n12-Sulfonyl-JA-Ile sodium salt (12-*O*-SO~3~Na-JA-Ile) was synthesized as follows. To a stirred mixture of 12OH-JA-Ile methyl ester (29mg, 0.085 mmol) in dry pyridine (6ml), a solution of sulfur trioxide pyridine complex (135mg, 0.85 mmol) in dry pyridine (6ml) was added and stirred for 2h at room temperature. A 45ml aliquot of methanol:water (2:1) was added and the mixture was neutralized by 1M KOH. The volatile component of the reaction mixture was removed under reduced pressure to give an oil, which was purified using Si gel column chromatography (30g, chloroform:methanol:acetic acid 80:20:0.1) to produce 12-sulfonyl JA-Ile methyl ester \\[27mg, 0.062 mmol (2ml), 73%\\]. 12-Sulfonyl JA-Ile methyl ester (27mg, 0.062 mmol) in ethanol (2ml) was stirred with an aqueous solution of 1M NaOH (0.7ml) for 2h at room temperature. Following neutralization with Amberlite IR-120, the volatile component of the reaction mixture was removed under reduced pressure to give an oil, which was purified using Si gel column chromatography (20g, methanol:chloroform:acetic acid 7:3:0.1) to produce 12-sulfonyl-JA-Ile sodium salt (12-*O*-SO~3~Na-JA-Ile, 25mg, 0.012 mmol, 92%). Physical data for 12-*O*-SO~3~Na-JA-Ile are as follows. Electrospray ionization (ESI)-MS *m/z* (rel. int., %) 419 (15), 418 (100, \\[M-Na\\]^\u2212^), 344 (22). High-resolution (HR)-ESI-MS: *m/z* 418.1544 \\[M-Na\\]^\u2212^ (calcd. for C~18~H~28~NO~6~S; 418.1541). ^1^H-Nuclear magnetic resonance (NMR) (270 MHz, D~2~O) \u03b4: 5.37 (2H, m), 4.15 (1H, dd, *J*=6.6, 9.9 Hz), 3.90 (2H, t, *J*=6.4 Hz), 2.49 (1H, m), 2.36-1.73 (10H, m), 1.12 (1H, m), 0.81 (3H, d, *J*=6.9 Hz), 0.75 (3H, t, *J*=7.3 Hz).\n\nTransgenic plants overexpressing IAR3 and ILL6 {#s5}\n----------------------------------------------\n\nConstruction of binary vector constructs for *in planta* overexpression of *ILL6* (At1g44350) (*ILL6-OE*) and *IAR3* (At1g51760) (*IAR3-OE*) was carried out by amplifying the full-length open reading frame (ORF) of each gene by reverse transcription--PCR (RT--PCR) from a total RNA from wounded WT leaves using the primer pairs indicated in [Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online. The resulting PCR fragments were cloned into a pBI121-derived pBITS vector ([@CIT0024]) using restriction enzyme sites *Xba*I and *Xho*I for *ILL6* or *Xba*I and *Bam*HI for *IAR3*. The resulting constructs with the respective genes placed behind the *Cauli\ufb02ower mosaic virus* (CaMV) 35S promoter (*35S:ILL6* or *35S:IAR3*) were transformed into the C58C1 strain of *Agrobacterium tumefaciens*. *Agrobacterium tumefaciens* harboring each construct was transformed into Arabidopsis using a floral dip method. Seeds harvested from the resulting plants (T~1~) were screened for resistance to kanamycin (50 \u03bcg ml^\u22121^). A total of 55 and 69 seedlings each from *35S:ILL6* and *35S:IAR3* that survived the kanamycin selection were tested for transgene expression by quantitative (q)RT--PCR and for JA levels. Two selected lines---one each from *ILL6-OE* and *IAR3-OE*---were further propagated eventually to obtain T~3~ homozygous lines.\n\nSubcellular localization studies {#s6}\n--------------------------------\n\nVector constructs were made for each of the two proteins, IAR3 and ILL6, with cyan fluorescent protein (CFP) fused to either their N- or C-terminus. ORFs of IAR3 and ILL6 with fused CFP sequence were amplified using an overlapping PCR method with the primers indicated in [Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online. The N-terminal fusions had CFP inserted behind the putative 23 and 24 amino acid signal peptide sequences of IAR3 and ILL6, respectively. The C-terminal fusion for IAR3--CFP preserved the IAR3's ER retention 'KDEL' motif at the very end, whereas ILL6--CFP did not have any ER retrieval sequence at the end. The amplified fragments were cloned into a Gateway binary expression vector, pGWB2 ([@CIT0032]). The resulting constructs were transiently expressed in *Nicotiana benthamiana* leaves by syringe infiltration of the *Agrobacterium* C58C1 strains harboring each construct ([@CIT0022]). A second strain of *Agrobacterium* containing a *CYP94B3-mRFP* construct with previously demonstrated ER localization ([@CIT0025]) was co-infiltrated. After 48h of infiltration, fluorescent images were acquired using a Leica TCP SP8 confocal microscope.\n\nRecombinant proteins and *in vitro* hydrolysis assays {#s7}\n-----------------------------------------------------\n\nORFs of ILL6 and IAR3 lacking 24 and 23 amino acids, respectively, of predicted N-terminal signaling sequences were PCR amplified ([Supplementary Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online) and cloned in between the *Eco*RI and *Sal*I sites of the pGEX-6P-1 vector (GE Healthcare) to obtain *pGEX-ILL6* and *pGEX-IAR3*, which placed a glutathione *S*-transferase (GST) tag on the N-terminus of ILL6 or IAR3. Sequence-verified *pGEX-ILL6* and *pGEX-IAR3* were each transformed into *Escherichia coli* strain BL21 (DE3). The *pGEX-ILR1* plasmid was obtained from Dr Bonnie Bartel (Rice University) ([@CIT0028]), and was transformed into the same BL21 (DE3) strain. Protein expression was induced at mid-log phase by adding 0.5mM isopropyl-\u03b2-[d]{.smallcaps}-1-thiogalactopyranoside (IPTG) and culturing for another 16--20h at 16 \u00b0C. GST\u00b7Bind\u2122 Kits (Novagen) were used to purify GST-fused proteins following the manufacturer's instructions. The purity of recombinant proteins estimated by SDS--PAGE was \\>90%.\n\nA previously described *in vitro* hydrolysis assay procedure ([@CIT0028]) was adopted to test enzymatic activities of the purified GST--ILL6, GST--IAR3, and GST--ILR1 proteins. A typical reaction mixture consisted of 50mM TRIS-HCl (pH 8.0), 1mM MnCl~2~, 1mM dithiothreitol, 5 \u03bcg of puri\ufb01ed protein, and the indicated amounts of substrates in a 25 \u03bcl reaction volume. GST proteins purified from an *E. coli* strain transformed with an empty pGEX-6P-1 vector were used as a control. The reaction was carried out at 28 \u00b0C for the indicated times and was terminated by adding 75 \u03bcl of stop solution consisting of 70% aqueous methanol and 0.5% acetic acid spiked with 0.25 \u03bcM dihydro-jasmonic acid (dhJA) and 0.5 \u03bcM d~5~-IAA as internal standards. A 5 \u03bcl aliquot of the centrifugation-cleared supernatant was directly injected for LC-MS.\n\nRNA analysis {#s8}\n------------\n\nRNA was extracted using TRIzol reagent (Invitrogen) from 50--100mg of frozen tissue ground into a fine powder. First-strand cDNA was synthesized from 1 \u03bcg of total RNA treated with DNase I (Qiagen) using oligo(dT)~20~ primers and Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega). A 5ng aliquot of the resulting cDNA was used as template for subsequent PCR steps. Phusion *Taq* polymerase (Thermo Scientific) and Taq-Pro Red Complete (Denville) were used as DNA polymerases for cloning purposes and for semi-quantitative RT--PCR experiments, respectively. qRT--PCR was performed on a CFX96 Touch\u2122 real-time PCR detection system (Bio-Rad) using SsoFast\u2122 EvaGreen^\u00ae^ Supermix (Bio-Rad) as per the manufacturer's instructions. Relative transcript abundance was calculated after normalization with *ACTIN8* (AT1G49240) as an internal reference gene. For time course experiments, fold change compared with the 0h time point of the WT was plotted. All other graphs show transcript levels relative to *ACTIN8*.\n\nMetabolite analysis by mass spectrometry {#s9}\n----------------------------------------\n\nHormone metabolites were quantified using ultra-performance liquid chromatography--tandem MS (UPLC-MS/MS) (Acquity/Xevo TQ-S system, Waters) based on methods described previously ([@CIT0025]). Characteristic MS transitions were monitored using multiple reaction monitoring in ESI positive mode for IAA (*m/z*, 176\\>130), d~5~-IAA (181\\>135), and ESI negative mode for JA (*m/z*, 209\u219259), dhJA (211\u219259), 12OH-JA (225\u219259), 12COOH-JA (239\u219259), 12-HSO~4~-JA (305\u219297), 12-*O*-Glc-JA (387\u2192207), JA-Ile (322\u2192130), \\[^13^C~6~\\]JA-Ile (328\u2192136), 12OH-JA-Ile (338\u2192130), 12COOH-JA-Ile (352\u2192130), 12-HSO~4~-JA-Ile (418\u2192130), and 12-*O*-Glc-JA-Ile (500\u2192130). Data analysis was carried out using MassLynx 4.1 and TargetLynx software (Waters).\n\nResults {#s10}\n=======\n\nTranscripts of *ILR1*, *ILL6*, and *IAR3* are induced by JA or wounding but not by IAA {#s11}\n--------------------------------------------------------------------------------------\n\nEarlier gene expression studies have identified members of the *IAH* gene family as being induced by wounding or JA treatment ([@CIT0048]; [@CIT0002]; [@CIT0052]). To gain a more comprehensive understanding of the JA-inducible nature of the *IAH* transcripts, we conducted a time course qRT--PCR analysis of all seven *IAH* genes ([Supplementary Fig. S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). Mature rosette leaves of 4-week-old Arabidopsis plants were either mechanically wounded or sprayed with a 100 \u03bcM MeJA solution, and sampled at 0, 0.5, 1, 4, 8, and 12h after treatment. Transcripts of three genes, *IAR3*, *ILL6*, and *ILR1*, were strongly induced by MeJA. The induction was transient, rising within a few hours of treatment and peaking at 4h, following typical 'early gene' kinetics ([@CIT0008]). Wounding also induced expression of *IAR3*, *ILL6*, and *ILR1*, and, additionally, *ILL5*, which is a pseudogene in the Col-0 ecotype ([@CIT0028]; [@CIT0052]). No transcript increase by either treatment was detected in *coi1-1* ([Supplementary Fig. S1B, D](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)), indicating a strict requirement for the functional JA perception pathway. The promoter regions of all three genes contained canonical MYC2-binding *cis*-elements (or G-box) according to the Arabidopsis Gene Regulatory Information Server (<http://arabidopsis.med.ohio-state.edu/>), and earlier microarray data indicated that their expression is regulated by MYC2 and MYC3 ([@CIT0011]; [@CIT0006]; [@CIT0034]).\n\nWe next tested if their expression can be induced by IAA treatment ([Supplementary Fig. S1E](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). Unlike the two IAA-responsive marker genes, *IAA5* and *GH3.1*, which were strongly induced by IAA, none of the *IAH* transcripts responded to IAA treatment. This is in contrast to the *GH* family genes encoding the IAA--amino acid-conjugating enzymes, which were strongly induced by IAA ([@CIT0013]; [@CIT0041]), showing that although IAH enzymes function in auxin metabolism, their gene expression is not controlled by IAA, but rather by JA.\n\nSubstrate specificities of recombinant ILR1, IAR3, and ILL6 proteins {#s12}\n--------------------------------------------------------------------\n\nTo test their catalytic activities, ILR1, IAR3, and ILL6 were expressed in *E. coli* as N-terminal GST fusions. Conditions for the *in vitro* hydrolysis assay were first determined by incubating purified recombinant proteins with varying concentrations of JA-Ile substrate for different times, and measuring the amount of released free JA ([Supplementary Fig. S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). A linear increase of JA was maintained for at least 3h for both GST--IAR3 and GST--ILL6 ([Supplementary Fig. S2A](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)) under the chosen assay conditions (see the Materials and methods). GST--IAR3 displayed at least several times higher hydrolytic activity compared with GST--ILL6 under all tested conditions, while purified GST protein alone showed no significant hydrolytic activity ([Table 1](#T1){ref-type=\"table\"}). GST--ILR1 did not show any detectable activity against JA-Ile substrate even at relatively high substrate concentrations, even though the enzyme was highly active against the IAA-Leu substrate ([Supplementary Fig. S2B, C](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)). Next, the three proteins were tested against 10 different conjugates of JA ([Table 1](#T1){ref-type=\"table\"}). Among the 10 were six with different amino acids (-Ile, -Leu, -Val, -Gln, -Thr, -Phe) and four with modification at the C~12~ position of JA-Ile (12OH-, 12COOH-, 12-*O*-Glc-, 12-HSO~4~-). GST--IAR3 hydrolyzed all six amino acid-substituted conjugates comparably well. Although hydroxylation at C~12~ had no effect on GST--IAR3 activity, further modifications to 12COOH-, 12-*O*-Glc-, or 12-HSO~4~- almost completely eliminated the cleavage activity.\n\n###### \n\nSubstrate specificity of recombinant amidohydrolase enzymes\n\n  Substrate           GST--IAR3           GST--ILL6    GST--ILR1     GST\n  ------------------- ------------------- ------------ ------------- ----------\n  JA-Ile              3563.6\u00b1110.5^*a*^   512.3\u00b11.4    25.3\u00b11.1      0.4\u00b10.4\n  JA-Leu              5066.8\u00b1213.9        236.9\u00b11.0    881.8\u00b1138.2   5.2\u00b11.9\n  JA-Val              3093.1\u00b1107.5        87.7\u00b12.7     107.5\u00b16.5     2.4\u00b10.9\n  JA-Gln              2672.7\u00b1169.1        80.8\u00b14.1     158.0\u00b17.0     23.0\u00b11.1\n  JA-Thr              4016.0\u00b1155.9        132.3\u00b13.5    36.0\u00b11.7      10.7\u00b11.1\n  JA-Phe              3812.7\u00b167.0         76.5\u00b16.4     1798.0\u00b158.9   21.8\u00b13.8\n  12OH-JA-Ile         2672.7\u00b1646.1        128.6\u00b117.7   59.8\u00b11.6      72.8\u00b12.4\n  12COOH-JA-Ile       207.7\u00b14.3           3.5\u00b10.0      7.8\u00b10.2       7.9\u00b10.4\n  12-*O*-Glc-JA-Ile   47.1\u00b14.2            4.3\u00b11.3      5.2\u00b16.6       6.7\u00b16.1\n  12-HSO~4~-JA-Ile    277.9\u00b126.0          3.0\u00b10.4      1.4\u00b11.5       1.1\u00b10.8\n  IAA-Ala^*b*^        27515.5\u00b1438.7       25.4\u00b14.1     284.0\u00b116.7    19.9\u00b11.8\n  IAA-Leu             78.9\u00b13.4            11.7\u00b14.0     2581.0\u00b178.7   8.6\u00b11.9\n\n^*a*^ Specific activity (pmol h^\u22121^ mg^\u22121^ of protein) was determined by quantifying JA, 12OH-JA, 12COOH-JA, 12-*O*-Glc-JA, 12-HSO~4~-JA, and IAA after 2h incubation of 5 \u03bcg of purified proteins with 4 \u03bcM of the indicated conjugate substrates. Each value represents the mean \u00b1SD of three replicates.\n\n^*b*^ For IAA conjugates, total reaction time was 0.5h.\n\nGST--ILL6 displayed the highest activity toward JA-Ile even though its activity was an order of magnitude lower than that of GST--IAR3. Its relative activity also varied more widely between substrates compared with GST--IAR3. For example, its activity decreased 4- to 6-fold when the -Val, -Gln, and -Phe conjugates of JA were used instead of JA-Ile. GST--ILL6 was able to hydrolyze 12OH-JA-Ile, but, as with GST--IAR3, further modifications at the C~12~ position eliminated the activity. GST--ILL6 showed essentially no activity against IAA-Ala or IAA-Leu, making ILL6 a specialized enzyme for JA conjugates. Even though GST--ILR1 did not show any activity for JA-Ile, it was active against JA-Leu and JA-Phe. This stark difference in preference for JA-Leu over JA-Ile was somewhat surprising, but a similar strict preference of ILR1 for -Leu over -Ile conjugates was observed with the IAA conjugates ([@CIT0001]).\n\nOverall, these *in vitro* studies show that IAR3 has broad substrate specificity for conjugates of both JA and IAA, whereas ILL6 is more specific for JA conjugates. Even though ILR1 also showed dual substrate specificity for IAA and JA conjugates, its lack of activity against major JA conjugates, i.e. JA-Ile and 12OH-JA-Ile, predicts its limited role in JA homeostasis.\n\nIn vivo function of ILR1, IAR3, and ILL6 in JA metabolism {#s13}\n---------------------------------------------------------\n\n*In planta* overexpression was employed to study the enzymatic function of ILL6, IAR3, and ILR1 *in vivo*. Stably transformed lines of Arabidopsis plants expressing the *ILL6* or *IAR3* genes under the control of the constitutive CaMV35S promoter, designated as *ILL6-OE* and *IAR3-OE*, respectively, were generated ([Supplementary Figs S3, S4](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). Among 55 T~1~ lines of *ILL6-OE* resistant to antibiotic selection, eight lines (lines 5, 19, 23, 42, 46, 47, 63, and 69) had markedly increased *ILL6* transcripts. The same eight lines were also severely reduced in JA-Ile content, consistent with increased JA-Ile hydrolysis by ILL6 overexpression. *IAR3-OE* lines also displayed strong correlation between the increased *IAR3* transcripts and reduced JA-Ile levels, and were established following a similar procedure to *ILL6-OE* ([Supplementary Fig. S4](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)). Two lines each from *ILL6-OE* (lines 5 and 47) and *IAR3-OE* (lines 4 and 42) were further propagated to generate T~3~ homozygous lines. Reduced JA-Ile levels were maintained in these homozygous lines, with levels in each representative line comparable with one another for both *ILL6-OE* and *IAR3-OE* progeny ([Supplementary Fig. S5](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)).\n\nA detailed time course JA profiling was carried out on one each of these homozygous *ILL6-OE* (line 5) and *IAR3-OE* (line 42) lines, along with a previously reported *ILR1-OE* line ([@CIT0009]) ([Fig. 1](#F1){ref-type=\"fig\"}). JA-Ile was strongly reduced in both *ILL6-OE* and *IAR3-OE* plants throughout the time course ([Fig. 1A](#F1){ref-type=\"fig\"}, D). *ILL6-OE* performed surprisingly well considering the lower hydrolytic activity (10 times less than GST--IAR3) that GST--ILL6 had shown in the *in vitro* assays ([Table 1](#T1){ref-type=\"table\"}). In addition to JA-Ile, there was a major reduction in the 12OH-JA-Ile level in both *ILL6-OE* and *IAR3-OE* ([Fig. 2B](#F2){ref-type=\"fig\"}, E, H), which was consistent with their *in vitro* enzyme assay results. 12OH-JA, the product of 12OH-JA-Ile hydrolysis, in both *ILL6-OE* and *IAR3-OE* was similar to (before 2h) or lower than (4h) the level in the WT ([Fig. 1C](#F1){ref-type=\"fig\"}, F). The increased rate of 12OH-JA-Ile hydrolysis and decreased pool size of the precursor (12OH-JA-Ile) were likely to have contributed to this mild decrease in 12OH-JA. Similarly, no significant changes were observed in the JA level in these plants ([Supplementary Fig. S6](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). *ILR1-OE* did not show any change in the major JA metabolite profile, consistent with the *in vitro* assay results.\n\n![Time course of JA metabolite accumulation in wounded leaves of the WT, *ILL6-OE*, *IAR3-OE*, and *ILR1-OE*. Mechanically damaged leaves at the indicated times after wounding were analyzed for JA-Ile (A, D, G), 12OH-JA-Ile (B, E, H), and 12OH-JA (C, F, I) contents using UPLC-MS/MS. Asterisks denote a significant difference at *P*\\<0.05; Student's *t*-test. Each data point represents the mean \u00b1SD of three biological replicates.](exbotj_erv521_f0001){#F1}\n\nNext, we analyzed loss-of-function mutants to test if the endogenous JA profiles would show opposite trends to the overexpressors ([Fig. 2](#F2){ref-type=\"fig\"}). To combat gene redundancy problems, double homozygous mutants, *iar3-5ill6-2* ([Supplementary Fig. S7](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online) and *iar3-2ilr1-1* ([@CIT0009]), were analyzed along with the single mutants. JA-Ile contents in wounded leaves of the single mutants, *iar3-5* and *ill6-2*, were not different from the WT, as noted in earlier reports ([@CIT0052]; [@CIT0025]), but the levels were significantly higher (*P*\\<0.05) in the *iar3-5ill6-2* double mutant compared with the WT or the two single mutants, indicating that the JA-Ile turnover was strained in this double mutant ([Fig. 2A](#F2){ref-type=\"fig\"}). The impact on 12OH-JA-Ile was even greater, displaying close to 2-fold increases in the single mutants and up to 3-fold increases in the double mutant compared with the WT levels ([Fig. 2B](#F2){ref-type=\"fig\"}). The trend was reversed with 12OH-JA; that is, reduced in each single mutant and still more reduced in the double mutant ([Fig. 3C](#F3){ref-type=\"fig\"}). On the other hand, the JA profile in *iar3-2ilr1-1* was not significantly different from that of *iar3-2*, indicating that there was no additive contribution by *ilr1-1* mutation ([Figs. 3D](#F3){ref-type=\"fig\"} [, F](#F3){ref-type=\"fig\"}).\n\n![Loss-of-function mutants of IAR3 and ILL6, but not ILR1, hyperaccumulate JA-Ile and 12OH-JA-Ile. Endogenous JA-Ile (A, D), 12OH-JA-Ile (B, E), and 12OH-JA (C, F) contents were quantified from wounded (2h) and unwounded leaves of 4-week-old plants. Asterisks denote a significant difference at *P*\\<0.05 compared with the WT (\\*) or *iar3-5* (\\*\\*); Student's *t*-test. Each data point represents the mean and SD of three biological replicates.](exbotj_erv521_f0002){#F2}\n\n![Overexpression of amidohydrolases impacts JA and IAA marker gene expression. (A and B) qRT--PCR analysis of *JAZ8* expression in unwounded and wounded (2h) leaves of the WT, *ILL6-OE* (line 5), *IAR3-OE* (line 42), and *ILR1-OE*. Fold change relative to the unwounded WT transcript level is displayed. (C and D) *IAA5* expression in 9-day-old seedlings grown on MS medium containing the mock treatment, 50 \u03bcM IAA-Ala, or 30 \u03bcM IAA-Leu. Expression levels relative to *ACTIN8* are displayed. Error bars denote the SD of three biological replicates. Asterisks indicate a significant difference at *P*\\<0.05 compared with the WT; Student's *t*-test.](exbotj_erv521_f0003){#F3}\n\nTogether, these results show that ILL6 and IAR3 act redundantly *in vivo* to metabolize JA-Ile and 12OH-JA-Ile, whereas ILR1 has a limited role in overall JA metabolism. A previous study ([@CIT0037]) revealed elevated levels of IAA-Ala and IAA-Leu in a *ilr1iar3ill2* triple mutant, establishing the *in vivo* function of ILR1 and IAR3 in hydrolyzing endogenous IAA--amino acid conjugates.\n\nIAH overexpression oppositely impacts expression of JA- and IAA-responsive genes {#s14}\n--------------------------------------------------------------------------------\n\nConstitutive activation of the JA-Ile turnover pathway in *ILL6-OE* and *IAR3-OE* is expected to have negative signaling consequences on JA-Ile-regulated gene expression. On the other hand, liberation of IAA from the amino acid conjugates by overexpression of *IAR3* and *ILR1* is expected to activate IAA-responsive gene expression. We tested this hypothesis in *IAR3-OE*, *ILL6-OE*, and *ILR1-OE* plants using a JA-Ile-responsive and an IAA-responsive marker gene, *JAZ8* ([@CIT0040]) and *IAA5* ([@CIT0010]), respectively. *JAZ8* transcripts increased \\>20-fold by wounding in WT leaves ([Fig. 3A](#F3){ref-type=\"fig\"}, B). *JAZ8* transcripts in *IAR3-OE* and *ILL6-OE* also increased upon wounding, but the induction was less than half that in the WT ([Fig. 3A](#F3){ref-type=\"fig\"}). This is correlated with the strongly depleted JA-Ile in these lines ([Fig. 1A](#F1){ref-type=\"fig\"}, D). Consistent with the unaltered JA-Ile levels in *ILR1-OE* ([Fig. 1G](#F1){ref-type=\"fig\"}), *JAZ8* gene expression levels were not changed compared with the WT in *ILR1-OE* ([Fig. 3B](#F3){ref-type=\"fig\"}). To test auxin response, the WT and the overexpressing lines were grown on Murashige and Skoog (MS) medium supplemented with IAA-Ala or IAA-Leu, or mock solutions. *IAA5* transcript levels increased in both the IAA-Ala treated WT and *IAR3-OE* plants compared with mock treatment, but the increase was much greater in *IAR3-OE* ([Fig. 3C](#F3){ref-type=\"fig\"}), consistent with the explanation that a higher amount of IAA is generated by the constitutive hydrolysis of IAA-Ala in *IAR3-OE*. Similarly, *IAA5* transcripts were strongly induced in *ILR1-OE* grown on IAA-Leu plates, consistent with ILR1's enzymatic function as an IAA-Leu hydrolase ([Fig. 3D](#F3){ref-type=\"fig\"}). *IAA5* gene expression was not induced in *ILL6-OE* by either IAA-Ala or IAA-Leu ([Fig. 3C](#F3){ref-type=\"fig\"}, D), consistent with the lack of *in vitro* hydrolytic activity of ILL6 against IAA conjugate substrates ([Table 1](#T1){ref-type=\"table\"}). Similar results were obtained using *JAR1* and *GH3.1* as additional marker genes, corroborating the above observations ([Supplementary Fig. S8](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). These gene expression results show that JA-Ile hydrolysis by IAR3 and ILL6 can attenuate JA-responsive gene expression and, at the same time, IAA-Ala and IAA-Leu hydrolysis by IAR3 and ILR1 can activate IAA-responsive gene expression in plants.\n\nILL6-OE and IAR3-OE display JA-deficient phenotypes {#s15}\n---------------------------------------------------\n\nWe next investigated whether biochemical and molecular changes caused by ectopic expression of IAHs result in any visible plant symptoms. Among the eight *ILL6-OE* lines selected from the T~1~ screening ([Supplementary Fig. S3](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online), two independent lines (lines 5 and 47) displayed partial fertility defects, which were heritable in the following generations ([Fig. 4A](#F4){ref-type=\"fig\"}, B). Over half of the siliques in these plants were underdeveloped, with few or no seeds inside ([Supplementary Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)). Developing flowers of these plants exhibited short anther \ufb01laments and delayed dehiscence ([Fig. 4B](#F4){ref-type=\"fig\"}), which are typical symptoms of JA-Ile deficiency in Arabidopsis ([@CIT0003]). Increased JA-Ile turnover by overexpression of CYP94B enzymes created similar reproductive defects ([@CIT0021], [@CIT0025]). Auxin also plays a role in flower development through JA, but JA biosynthesis is downstream of auxin signaling ([@CIT0031]). Unlike *ILL6-OE*, no apparent fertility defect was observed with *IAR3-OE* or *ILR1-OE* ([Fig. 4A](#F4){ref-type=\"fig\"}, B). Lack of flower phenotype in *IAR3-OE* was not entirely unexpected given the relatively milder depletion of JA-Ile in *IAR3-OE* compared with *ILL6-OE* ([Fig. 1](#F1){ref-type=\"fig\"}; [Supplementary Fig. S5](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)).\n\n![*ILL6-OE* and *IAR3-OE* plants display JA-deficient phenotypes. (A and B) Photographs of representative stems and flowers of the WT, *ILL6-OE* (line 5), *IAR3-OE* (line 42), and *ILR1-OE*. Underdeveloped siliques and short anther filaments are observed for *ILL6-OE*. Scale bar=1mm. (C--E) Photo images and root length measurements of 9-day-old seedlings grown vertically on MS medium supplemented with 20 \u03bcM jasmonic acid (C--E) or 0.5 \u03bcM coronatine (COR) (D and E). Data show the mean and SD (*n* \\>18). Asterisks denote a significant difference compared with the WT at *P*\\<0.05; Student's *t*-test. (This figure is available in colour at *JXB* online.)](exbotj_erv521_f0004){#F4}\n\nIn roots, both *ILL6-OE* and *IAR3-OE* exhibited resistance to exogenous JA-induced growth inhibition ([Fig. 4C](#F4){ref-type=\"fig\"}, D). Mutations in JA perception or signaling, defects in the conjugation step to produce JA-Ile (i.e. *jar1*), or increased turnover of JA-Ile (e.g. *CYP94B3* overexpression) can make plants resistant to exogenous JA ([@CIT0029]; [@CIT0042]; [@CIT0046]; [@CIT0021]). In *ILL6-OE* and *IAR3-OE*, increased hydrolysis of JA-Ile is likely to have made plants appear insensitive to exogenous JA. *IAR3-OE* and *ILL6-OE* roots were fully sensitive to exogenous coronatine, a structural mimic of JA-Ile and a potent agonist of the COI1--JAZ receptor system, indicating that the observed insensitivity to exogenous JA was not due to any defect in JA perception or signaling ([Fig. 4E](#F4){ref-type=\"fig\"}).\n\nOverexpression of ILR1 and IAR3 confers hypersensitivity to exogenous IAA conjugates whereas ILL6 antagonizes the process {#s16}\n-------------------------------------------------------------------------------------------------------------------------\n\nAn auxin-induced primary root growth inhibition assay was employed to examine auxin-related phenotypes. *ILL6-OE*, *IAR3-OE*, *ill6-2*, and *iar3-5* plants were grown together with WT control on MS medium supplemented or not with IAA (5 \u03bcM) or IAA-Ala (50 \u03bcM) ([Fig. 5A](#F5){ref-type=\"fig\"}, B). All genotypes grown on plain MS medium had similar root lengths, but their root growth was uniformly inhibited by IAA inclusion in the medium, indicating fully functional IAA perception and signaling pathways in all genotypes. However, when seedlings were grown on IAA-Ala, clearly visible variations in root lengths between genotypes appeared. WT root growth was inhibited by IAA-Ala. This could mostly be prevented by knocking-out IAR3 (*iar3-5*) and enhanced by overexpressing IAR3, consistent with IAR3's dominant function in plant sensitivity to IAA-Ala ([@CIT0009]). The results were similar with *ilr1-5* and *ILR1-OE* on IAA-Leu plates ([Fig. 5C](#F5){ref-type=\"fig\"}), demonstrating the dominant function of ILR1 in the plant response to IAA-Leu ([@CIT0001]). Surprisingly, *ILL6-OE* roots were insensitive to IAA-Ala, reaching the same length as *iar3-5* roots ([Fig. 5B](#F5){ref-type=\"fig\"}). This was unexpected because ILL6 lacked any substrate specificity for IAA-Ala, and, even if it did have, it should make plants more sensitive to IAA-Ala, not insensitive ([Table 1](#T1){ref-type=\"table\"}). Furthermore, the insensitivity of *ILL6-OE* to exogenous auxin conjugates was also apparent with IAA-Leu, with their roots reaching the length of the IAA-Leu-insensitive *ilr1-5* roots ([Fig. 5C](#F5){ref-type=\"fig\"}).\n\n![Impacts of genetic manipulation of IAHs on plant sensitivity to exogenous IAA conjugates. (A--C) Photographs and root length measurements of 9-day-old WT, *ILL6-OE* (line 5), *IAR3-OE* (line 42), *ILR1-OE*, *ill6-2*, *iar3-5*, and *ilr1-5* plants grown on MS medium supplemented or not with 5 \u03bcM IAA, 50 \u03bcM IAA-Ala, or 30 \u03bcM IAA-Leu. Data show the mean and SD (*n* \\>18). (D and E) qRT--PCR analysis of *IAA5* and *GH3.1* expression in leaves of 4-week-old WT and *ILL6-OE* plants sprayed (2h) with the mock, 5 \u03bcM IAA, or 50 \u03bcM IAA-Ala solutions. Relative abundance compared with *ACTIN8* is displayed. Error bars indicate the SD of three biological replicates. Asterisks denote a significant difference compared with the WT at *P*\\<0.05 (\\*) or *P*\\<0.001 (\\*\\*); Student's *t*-test. (This figure is available in colour at *JXB* online.)](exbotj_erv521_f0005){#F5}\n\nWe tested whether the *ILL6-OE*-conferred resistance to IAA conjugates is observed in leaves. For this, expression of IAA response marker genes, *IAA5* and *GH3.1*, was measured 2h after spraying the mature rosette leaves of WT and *ILL6-OE* plants with solutions containing the mock treatment, IAA (5 \u03bcM), or IAA-Ala (50 \u03bcM). Both *IAA5* and *GH3.1* transcripts in the WT increased significantly upon IAA treatment and, to a lesser extent, upon IAA-Ala treatment compared with the mock treatment ([Fig. 5D](#F5){ref-type=\"fig\"}, E). In *ILL6-OE*, expression of the two genes was also induced by IAA, but not by IAA-Ala, reproducing a loss of responsiveness to IAA-Ala similar to that seen in the roots. The mechanism whereby ILL6 interferes with plant perception of IAA conjugates in *ILL6-OE* is unclear, but it is does not occur through transcriptional down-regulation of *IAR3* or *ILR1* genes because the expression of *IAR3* and *ILR1* genes was not impaired in the *ILL6-OE* leaves ([Supplementary Fig. S9A--C](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). In addition, we obtained indirect evidence that ILL6 could bind to IAA conjugates. As noted earlier, when GST--ILL6 is incubated with JA-Ile in *in vitro* hydrolysis assays, it hydrolyzes JA-Ile. However, when IAA-Ala was added along with JA-Ile, the JA-Ile hydrolysis by GST--ILL6 was strongly inhibited ([Supplementary Fig. S9D](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)), indicating that IAA-Ala was somehow interfering with GST--ILL6's activity against JA-Ile, most probably through competition for the active site. It is possible then, that, in *ILL6-OE* plants, the highly abundant ILL6 proteins are competing with the relatively small number of endogenous IAR3 or ILR1 proteins for IAA conjugate substrates, preventing their hydrolysis by IAR3 or ILR1. In this case, exceeding concentrations of IAA conjugates in the root assay medium may eventually saturate the ILL6-binding sites and restore the root sensitivity to IAA conjugates. Supporting this hypothesis, a dose-dependent inhibition of root growth was observed with the *ILL6-OE* seedlings grown on increasing concentrations of IAA-Ala ([Supplementary Fig. S9E](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1)).\n\nILL6 and IAR3 are localized to the ER {#s17}\n-------------------------------------\n\nThe ER has been the predicted site of subcellular location for several IAHs based on their primary sequence features ([@CIT0030]). IAR3 has both the putative N-terminal signal sequence and the C-terminal 'KDEL' ER retrieval signature. ILL6, on the other hand, lacks the C-terminal ER retrieval motif but has the predicted N-terminal cleavable signal sequence. To determine experimentally the subcellular localization of IAR3 and ILL6 proteins, vector constructs with CFP-fused IAR3 and ILL6 were generated. Two versions of the constructs were tested for each protein where CFP was joined to either the N- or C-terminus of each protein, designated as CFP--IAR3, IAR3--CFP, CFP--ILL6, and ILL6--CFP. For N-terminal fusions, CFP was inserted behind the predicted signal sequence of IAR3 and ILL6, and, for the C-terminal fusions, CFP was inserted in front of the 'KDEL' of IAR3, whereas CFP was joined to the very end of the ILL6 protein. The resulting four constructs were transiently expressed in *N. benthamiana* leaves by syringe infiltration of *Agrobacterium* harboring each construct. A second culture of *Agrobacterium* carrying the CYP94B3--mRFP (monomeric red fluorescent protein) construct was co-infiltrated to be used as an ER marker ([@CIT0025]). [Figure 6](#F6){ref-type=\"fig\"} displays laser scanning confocal microscopy images taken after 2 d of infiltration. All four CFP-fused IAR3 and ILL6 proteins illuminated the characteristic ER network which perfectly overlapped with the CYP94B3--mRFP signals. Thus, we conclude that both IAR3 and ILL6 proteins are targeted to the ER, making the ER the probable site of IAA and JA conjugate hydrolysis.\n\n![Subcellular localization of IAR3 and ILL6. Confocal images of tobacco leaf epidermal cells co-expressing CFP-fused IAR3 or ILL6 (left panels of A--D) with CYP94B3--mRFP (middle panels). CFP was fused to either the N- or C-terminus of IAR3 and ILL6, and designated as CFP--IAR3, IAR3--CFP, CFP--ILL6, and ILL6--CFP. Merges show co-localization (right panels of A--D). Scale bars=10 \u03bcm.](exbotj_erv521_f0006){#F6}\n\nDiscussion {#s18}\n==========\n\nOur results show that IAH enzymes contribute to the homeostasis of both auxin and JA in plants, providing a potential mechanism to regulate simultaneously the two hormone levels, which bear opposite signaling consequences of switching one on while switching off the other ([Fig. 7](#F7){ref-type=\"fig\"}).\n\n![IAA and JA conjugation and deconjugation have opposite signaling consequences. IAA-Ala and IAA-Leu are hydrolyzed by IAR3 and ILR1, respectively, and the released free IAA activates IAA-responsive gene expression through the AUX/IAA--SCF^TIR1/AFB^ system. JA-Ile and 12OH-JA-Ile are hydrolyzed by either IAR3 or ILL6. As a result, JA-Ile-responsive gene expression through the JAZ-SCF^COI1^ system is attenuated. Genes encoding IAHs, AUX/IAA, and JAZ are under positive feedback regulation (dashed lines) along with other biosynthetic enzymes and transcriptional regulators.](exbotj_erv521_f0007){#F7}\n\nExpression of the three genes was up-regulated by JA, but not by IAA, meaning that they are more likely to function under physiological conditions that induce JA biosynthesis. Mechanical tissue damage and insect herbivory are known to induce the otherwise low levels of JA in leaves dramatically. The function of ILL6 and IAR3 in these circumstances is probably to hydrolyze excess JA-Ile and thereby attenuate prolonged activation of stress responses. The potentially harmful effects of JA-Ile overaccumulation, however, were not apparent in *ill6iar3.* This is probably because JA-Ile eventually gets turned over via other pathway(s), such as the CYP94-mediated \u03c9-oxidation pathway ([@CIT0020]; [@CIT0021]; [@CIT0014]). Higher order mutants that further block JA-Ile turnover will be useful to study the physiological effects of unrestrained JA-Ile accumulation.\n\nRegarding the physiological impacts of IAA conjugate hydrolysis by IAHs, IAA conjugate content in mature Arabidopsis leaves is quite low ([@CIT0026]) and is expected to be a relatively minor source for bulk free IAA in leaves. However, IAA conjugate hydrolysis may become more important for localized IAA increase in specific cell types (e.g. in undamaged meristematic regions of plants) where they could contribute to post-stress emergence of new organs. In addition to wounding or insect herbivory, IAHs may play a role based on a normal developmental program. Tissue-specific and developmentally controlled expression of *ILR1* and *IAR3* genes was studied using transgenic promoter--reporter lines of Arabidopsis ([@CIT0037]). Both *ILR1* and *IAR3* were expressed in flowers, particularly in pollen, where interplay between auxin and JA signaling is required for fertility ([@CIT0031]). *IAR3* is also expressed in roots where collaborative actions of auxin and JA repress root meristem activity and growth ([@CIT0005]). Indeed, the root was where the strongest morphological effects of IAR3 and ILL6 overexpression were observed. Tissue-specific expression of *ILL6* has not been studied in detail; however, an available public microarray resource (<http://bar.utoronto.ca/>) indicates its expression in flowers, which coincides with reduced fertility of *ILL6-OE* plants. The remaining three members of IAHs (especially ILL2 which has high sequence homology to IAR3), even though they are not inducible by JA or wounding, could also potentially function as JA amidohydrolases in tissues where they are expressed. Catalytic activities of these other IAHs on JA conjugates remain to be tested.\n\nA second aspect of auxin and JA pathway crosstalk mediated by IAHs is through overlapping substrate specificities of IAHs for both auxin-- amino acid and JA--amino acid conjugates. Cross-regulation between different hormone signaling pathways by means of transcriptional regulation is common, but examples of crosstalk at the metabolic level are relatively rare. Purified recombinant GST--IAR3 enzyme showed the broadest substrate specificity among the three IAHs, being active against both IAA and JA conjugates, and, in addition, catalyzing the cleavage of all six amino acid-substituted conjugates of JA with similar efficiencies ([Table 1](#T1){ref-type=\"table\"}). Somewhat differently from a previous report ([@CIT0052]), hydroxylation at the C~12~ position of JA-Ile did not diminish the hydrolysis by GST--IAR3. On the other hand, further modification of the -OH group by oxidation (to form -COOH), glucosylation, or sulfation abolished their hydrolysis by GST--IAR3 ([Table 1](#T1){ref-type=\"table\"}). The versatility of IAR3 is also shown by its ability to hydrolyze synthetic abscisic acid--amino acid conjugates *in vitro* ([@CIT0049]).\n\nGST--ILR1 displayed more stringent substrate specificities compared with GST--IAR3, with the most notable difference being its inability to use JA-Ile, making it a more specific enzyme for auxin conjugates. GST--ILR1 showed appreciable activity toward conjugates other than JA-Ile (e.g. JA-Phe and JA-Leu), but these conjugates occur at much lower levels than JA-Ile in Arabidopsis seedlings and wounded leaves ([@CIT0042]; [@CIT0022]). JA-Leu and JA-Phe were also ineffective in inhibiting root elongation ([@CIT0042]) and were not effective as ligands to promote COI1 interaction with JAZs ([@CIT0046]). Consistently, knocking-out or overexpressing ILR1 had no measurable impact on the endogenous JA profile or visible JA-deficient phenotypes. However, we cannot rule out the possibility that ILR1's function as a JA conjugate hydrolase could become important in certain conditions, cell types, or plants species, where these rarer JA conjugates accumulate to higher concentrations ([@CIT0051]).\n\nBoth *in vitro* and *in vivo* results are consistent with ILL6 being a JA-specific amidohydrolase. Even though GST--ILL6's *in vitro* hydrolytic activity for JA-Ile was an order of magnitude lower than that of GST--IAR3, overexpression of ILL6 in plants (*ILL6-OE*) resulted in a severe depletion of JA-Ile. This stark difference between *in vitro* and *in vivo* activities illustrates the importance of proper contextual requirements for optimal ILL6 activity whether that is protein folding, cofactors, or post-translational modification. Surprisingly, even though GST--ILL6 enzyme displayed no *in vitro* hydrolytic activity against IAA-Ala or IAA-Leu, overexpression of ILL6 made plants strongly insensitive to these IAA conjugates, suggesting the potential function of ILL6 as a negative regulator of IAA--amino acid hydrolysis *in vivo*. IAA perception and signaling in *ILL6-OE* was normal, and no sign of changes in *IAR3* or *ILR1* transcription was detected to explain *ILL6-OE*'s insensitivity to IAA conjugates. We found out that the JA-Ile-hydrolyzing activity of GST--ILL6 was strongly inhibited by the addition of IAA-Ala (a non-suitable substrate for ILL6) to the reaction mixture along with JA-Ile substrates ([Supplementary Fig. S9D](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1) at *JXB* online). This suggests that even though ILL6 cannot hydrolyze IAA-Ala, it may still bind IAA-Ala, which, by an unknown mechanism, interferes with the JA-Ile hydrolysis by ILL6. It can be speculated that in the *ILL6-OE* seedlings, high levels of ILL6 proteins may compete with the less abundant native IAR3 or ILR1 proteins for IAA conjugate substrates, preventing their hydrolysis by IAR3 or ILR1, and thus making plants appear resistant to the IAA conjugates. Our data showing restoration of *ILL6-OE*'s sensitivity to the increasing concentrations of IAA-Ala are consistent with this hypothesis. More study is needed to investigate this novel function of ILL6 as a negative regulator of IAA conjugate hydrolysis.\n\n12OH-JA that forms as a result of 12OH-JA-Ile hydrolysis can be considered an inactive by-product of the JA catabolic pathway; however, specific signaling roles for 12OH-JA have been reported in other plant species ([@CIT0055]; [@CIT0033]; [@CIT0035]). Presently, hydrolysis of 12OH-JA-Ile is the only experimentally proven metabolic pathway to produce 12OH-JA in higher plants ([Fig. 7](#F7){ref-type=\"fig\"}). However, the close to 50% of the WT level of residual 12OH-JA in the *iar3ill6* double mutant ([Fig. 2](#F2){ref-type=\"fig\"}) indicates the existence of separate metabolic routes to make 12OH-JA. Given the low expression level of the remaining members of *IAH* genes in wounded leaves and the lack of 12OH-JA-Ile hydrolytic activity by ILR1, a contribution of other members seems less likely. Recently, a monooxygenase from rice blast fungus was reported to catalyze direct 12-hydroxylation of JA ([@CIT0035]), but such an enzyme has yet to be identified in plants.\n\nOne question regarding IAH's contribution to the attenuation of JA signaling is: 'What is the net effect of the JA-Ile hydrolysis to JA, which could potentially be conjugated back to JA-Ile by JAR1?' We do not have direct evidence as to whether or how much of the hydrolyzed JA is recycled back to synthesize JA-Ile. However, the contribution by this recycling to the overall JA-Ile level is expected to be far less than JA-Ile hydrolysis, resulting in a net decrease of JA-Ile over time. This is partly because only a fraction of bulk JA, typically \\~10%, is converted to JA-Ile in leaves. JA-Ile-deficient phenotypes of *IAR3-OE* and *ILL6-OE* plants are consistent with JA-Ile signal being destroyed through hydrolysis rather than being futile. Differential partitioning between subcellular compartments may also play a role in physical separation of the cleaved JA from JA-Ile. In contrast to the cytosolic location of the JA-Ile-conjugating enzyme, JAR1 ([@CIT0018]), we have provided evidence that the hydrolases are localized to the ER, together with the CYP94 enzymes catalyzing the oxidative turnover of JA-Ile ([@CIT0025]). The close proximity of the hydrolases to the hydroxylases may facilitate their collaborative turnover of JA-Ile. The exact topology of these proteins remains to be determined. Based on the established topology of many ER-localized P450s, CYP94s were assumed to face the cytosolic face of the ER ([@CIT0025]). However, IAR3 which has both an N-terminal signal sequence and a C-terminal 'KDEL' ER retrieval signature may have different topology from CYP94s or ILL6 which lacks the C-terminal ER retrieval motif. Strong resistance to exogenous JA displayed by the overexpressing plants suggests effective competition by the ER-residing hydrolases with the CYP94s and the nuclear-residing receptor complexes for the JA-Ile substrate. This indicates rapid interchange of JA-Ile between different subcellular compartments where these enzymes can have access to it. The situation may be analogous to the IAA conjugates, where ER-localized IAHs are hydrolyzing and activating the IAA conjugates made by the cytosolic GH3 enzymes ([@CIT0054]; [@CIT0030]).\n\nAdditional studies are needed to understand the biological function of IAA and JA conjugate hydrolysis and the advantage provided by the simultaneous regulation of auxin and JA levels by IAHs. This may be too subtle to be detected easily. However, strong evolutionary conservation across higher plants of the amidosynthetase and amidohydrolase family genes as well as the ubiquitous occurrence of amino acid conjugates of auxin and JA ([@CIT0030]; [@CIT0050]) bear witness to their importance for plant survival in nature.\n\nSupplementary data {#s19}\n==================\n\nSupplementary data are available at *JXB* online.\n\n[Table S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). *ILL6-OE* plants have reduced fertility.\n\n[Table S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Primers used in this study.\n\n[Figure S1](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Arabidopsis *IAH* family gene expression.\n\n[Figure S2](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). *In vitro* hydrolysis activities of GST--IAR3, GST--ILL6, and GST--ILR1.\n\n[Figure S3](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Generation of *ILL6-OE* plants.\n\n[Figure S4](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Generation of *IAR3-OE* plants.\n\n[Figure S5](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). JA-Ile in the homozygous T~3~ *ILL6-OE* and *IAR3-OE* plants.\n\n[Figure S6](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Jasmonic acid and 12COOH-JA-Ile accumulation in IAH-overexpressing lines.\n\n[Figure S7](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Molecular characterization of *iar3-5ill6-2*.\n\n[Figure S8](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). *JAR1* and *GH3.1* gene expression in *IAR3-OE* and *ILL6-OE*.\n\n[Figure S9](http://jxb.oxfordjournals.org/lookup/suppl/doi:10.1093/jxb/erv521/-/DC1). Possible mechanism of *ILL6-OE*-conferred insensitivity to IAA conjugates.\n\n###### Supplementary Data\n\nWe acknowledge the technical assistance of Vinit Shanbhag, Hannah Wahl, Taylor Wagner, and Benjamin Boeding. We appreciate helpful comments by Dr Gregg Howe (Michigan State University) and Dr John Browse (Washington State University). We thank Dr Bonnie Bartel (Rice University) for *ilr1-1*, *ilr1-5*, *iar3-2*, *ilr1-1iar3-2*, and *ILR1-OE* seeds and the *pGEX-ILR1* construct, and the Arabidopsis Biological Resource Center for *ill6-2* and *iar3-5* seeds. We thank MU Cytology Core and Metabolomics Core for their assistance with confocal microscopy and mass spectrometry. This work was supported by funds from the Food for the 21st Century Program, University of Missouri (to AJK) and the UM System Research Board Grant (to AJK).\n\n[^1]: Editor: Robert Hancock, The James Hutton Institute\n"}
+{"text": "Background\n==========\n\nA steady increase in the prevalence of obesity has been observed worldwide. Increased caloric intake and decreased physical activity are major contributors. Obesity is often accompanied by activation of inflammatory signaling pathways and alterations in systemic and local levels of inflammatory cytokines \\[[@B1]\\]. Many of these cytokines, including tumor necrosis factor- \u03b1 (TNF-\u03b1), interleukin-6 (IL-6), interleukin-10 (IL-10), and monocyte chemotactive protein-1 (MCP-1) have profound effects on metabolism \\[[@B1]\\]. The adipose tissue is directly involved in the chronic inflammatory state that is observed in obesity. Cells of innate immune system, such as macrophages \\[[@B2]\\], mast cells \\[[@B3],[@B4]\\], and neutrophils \\[[@B5]\\] accumulate in the adipose tissue in obesity. Moreover, cells of adaptive immunity, including regulatory T cells \\[[@B6],[@B7]\\], CD8^+^T cells \\[[@B7]\\], and natural killer T cells \\[[@B8]\\] are also involved in adipose tissue inflammation in obesity.\n\nAlthough the details of the underlying mechanisms that provoke inflammation in the adipose tissue in obesity remain largely elusive, adipocyte injury and death play a central role \\[[@B9]\\]. Based on ultrastructural and immunohistochemical alterations, it was proposed that many adipocytes in white adipose tissue of obese rodents and humans demonstrate characteristic features of necrosis, but not apoptosis \\[[@B9]\\]. A recent study, however, showed that both mitochondrial- and death receptor-mediated caspase activation and adipocyte apoptosis were increased in the adipose tissue of obese humans and diet-induced obese mice \\[[@B10]\\]. Regardless of the pathway(s) involved, metabolically challenged adipocytes are more prevalent and accompanying inflammatory reactions are more severe in visceral than in subcutaneous fat in obesity \\[[@B3],[@B11]\\]. There is also a remarkable heterogeneity in the severity of adipose tissue inflammation in different visceral fat depots in obese subjects \\[[@B3],[@B12]\\].\n\nAlthough much emphasis has been placed on the understanding of molecular, biochemical, and immunological mechanisms involved in adipose tissue inflammation in obesity, little is known about immune responses taking place in the adipose tissue during the advanced stages of obesity. In the present study, we examined the epididymal fat of mice with long-standing obesity and determined: 1) the density and distribution of dead adipocytes, macrophages, and mast cells throughout the fat depot, 2) gene expression levels of several adipocytokines, 3) the prevalence of apoptosis, and 4) the relation among apoptosis, adipocytokine gene expression, and the degree of inflammatory cell infiltration. We showed that reduced mass of the epididymal fat in mice with long-standing obesity is accompanied by divergent distribution of crown-like structures, apoptotic cells, mast cells, and macrophages leading to diminished adipocytokine gene expression.\n\nMethods\n=======\n\nExperimental animals\n--------------------\n\nWe followed the \\'Principles of laboratory animal care\\' established by the National Institutes of Health. The Institutional Animal Care and Use Committee of the University of Miami specifically approved this study (IACUC protocol number 08-245). Blood samples and tissues were from a recent study \\[[@B3]\\]. Male C57BL/6 mice were purchased from Taconic (Hudson, NY) and acclimated for two weeks before the beginning of the study. Mice were randomized to either a high-fat diet (HFD) consisting of 60% calories from fat (D12492) or a low-fat diet (LFD) consisting of 10% calories from fat (D12450B) at the age of 6 weeks and were sacrificed when 6 months old (Research Diets Inc., St. Louis, MO). Mice were weighed with a Scout Pro balance SP202 (Ohaus, Pine Brook, NJ). Organ weights were measured with a Sartorius ED124S Analytical Balance (Sartorius, Bohemia, NY).\n\nGlucose, insulin and cholesterol assays\n---------------------------------------\n\nAfter overnight fasting (15 hours), blood was sampled from the tail of unanesthetized mice. Blood glucose concentrations were measured using a Contour glucometer (Bayer, Tarrytown, NY). Serum insulin concentrations were measured by immunoassay following manufacturer\\'s instructions (Crystal Chem, Downer Grover, IL). Serum cholesterol concentrations were determined by Cholesterol LiquiColor enzymatic assay (Stanbio Laboratory, Boerne, TX). Homeostasis Model of Assessment - Insulin Resistance (HOMA-IR) was calculated using the formula: fasting glucose (mg/dl) \u00d7 fasting insulin (mU/L)/405.\n\nThe anatomy of the epididymal fat in the mouse\n----------------------------------------------\n\nThe epididymal fat is an anatomically distinct and metabolically active abdominal fat depot widely used to study adipose tissue biology in rodents \\[[@B2]-[@B12]\\]. The paired epididymal fat is attached caudally to the ipsilateral testis and epididymis and extends proximally toward the diaphragm (Figure [1](#F1){ref-type=\"fig\"}). The epididymal fat incorporates spermatic blood vessels and accompanies them medially to the retroperitoneum \\[[@B13]\\]. Using spermatic blood vessels as a landmark, the epididymal fat can be divided into three zones: 1) medial zone, located around the spermatic artery and its main branches, 2) caudal zone, attached to the testis and epididymis, and 3) rostral zone, the loose end proximal to the medial zone. We quantified the density of mast cells and crown-like structures (CLS) in all three zones. A crown-like structure consists of several macrophages enveloping an adipocyte \\[[@B9]\\]. Protein and gene expression was measured only in the caudal and rostral zones that showed the greatest difference in the density of mast cells and CLS (see below).\n\n![**The epididymal fat in the male mouse**. To demonstrate abdominal organs, the anterior abdominal wall was removed and the left epididymal fat (1) was deflected to the left. The epididymal fat was divided into a caudal, medial, and rostral zone relative to the location of spermatic blood vessels (2). The left testis (3), the urinary bladder (4), a small bowel loop (5), and attached mesenteric fat (6) are also shown.](1476-511X-10-198-1){#F1}\n\nHistology\n---------\n\nTissues were fixed in Carnoy\\'s fixative and embedded in paraffin. Five micron-thick sections were cut, baked overnight at 60\u00b0C, deparaffinized in xylene, rehydrated in a graded ethanol series, and then stained. To demonstrate mast cells, toluidine blue staining was carried out by briefly submerging tissue sections in 0.1% aqueous toluidine blue (EMS, Hatfield, PA) \\[[@B3]\\]. In addition, the esterase activity of mast cells was detected using naphthol AS-D chloroacetate as substrate as previously described \\[[@B3]\\].\n\nImmunohistochemical staining for macrophage F4/80\n-------------------------------------------------\n\nAs previously described, following quenching endogenous peroxidase activity and heat-induced epitope retrieval, deparaffinized tissue sections were blocked with rabbit serum and sequentially incubated with a monoclonal antibody against macrophage F4/80 (1:100; AbD Serotec, Raleigh, NC) or isotype control, an appropriated secondary antibody, avidin-biotin complex, and 3,3\\'-diaminobenzidine \\[[@B3]\\].\n\nQuantitation of mast cells and CLS\n----------------------------------\n\nToluidine blue-stained sections were used to count mast cells. F4/80-stained sections were used to count CLS. Mast cells and CLS were counted in 20 high-power fields (400X) in the caudal, medical, and rostral zones of the epididymal fat. Density was expressed as mast cells or CLS per mm^2^of tissue section. Light microscopic images were acquired using a Leica DMLB microscope with a Leica DFC420 C color camera (Leica, Bannockburn, IL).\n\nImmunofluorescence staining for cleaved caspase-3\n-------------------------------------------------\n\nAntigen retrieval was carried out in 10 mM citrate buffer, pH 6.0, heated to 95-100\u00b0C for 15 minutes in a microwave. Tissue sections were blocked with 5% goat serum for 30 minutes and then sequentially incubated with rabbit anti-mouse cleaved caspase-3 antibody for 60 minutes (1:200; Cell Signaling, Danvers, MA), Alexa Fluor 594 goat anti-rabbit IgG for 60 minutes (1:1,000; Invitrogen, Carlsbad, CA), and DAPI (4\\',6-diamidino-2-phenylindole) for 5 minutes (0.1 \u03bcg/ml, Invitrogen, Carlsbad, CA). All incubations took place in a humid chamber at room temperature. Sections were rinsed with PBS between incubations.\n\nElectron microscopy\n-------------------\n\nAfter immersion fixation in 0.1 M phosphate buffer containing 2% glutaraldehyde and 2% paraformaldehyde, pH 7.4, adipose tissue samples were post-fixed in 0.1 M phosphate buffer containing 1% osmium tetroxide for 2 hours at 4\u00b0C. Samples were then dehydrated in a graded acetone series, infiltrated and embedded in Spurr\\'s epoxy resin. Tissue sections were cut at 50 nm (Reichert Ultracut S microtome), retrieved onto 300 mesh copper grids, and contrasted with uranyl acetate and lead citrate. Sections were examined using a Morgagni 268 transmission electron microscope (FEI, Hillsboro, OR) and images were acquired with an AMT Advantage 542 CCD camera system (Advanced Microscopy Techniques, Danvers, MA).\n\nProtein extraction and immunoblotting\n-------------------------------------\n\n100 mg adipose tissue was homogenized in 0.5 ml ice-cold lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 5 mM EDTA, 1 mM EGTA (Boston BioProducts, Ashland, MA) supplemented with a cocktail of protease inhibitors (Complete Mini, Roche Applied Science, Indianapolis, IN). Homogenates were agitated on a rotator for 1 hour at 4\u00b0C followed by centrifugation at 13,000 rpm for 20 minutes at 4\u00b0C. The middle phase of lysate was subjected to polyacrylamide gel electrophoresis (NuPAGE Bis-Tris system, Invitrogen, Carlsbad, CA). The following antibodies were used for immunoblotting: cleaved caspase-3 (1:1,000; Cell Signaling, Danvers, MA), mMCP6 (1:1,000; R&D Systems, Minneapolis, MN), and GAPDH (1:10,000; Abcam, Cambridge, MA). Signal intensities were measured using ImageJ (NIH, Bethesda, MD) and target protein-to-GAPDH ratio was used for statistical analysis.\n\nQuantitative real-time PCR\n--------------------------\n\nTotal RNA was isolated from 50 mg of adipose tissue using the RNeasy kit (Qiagen, Valencia, CA) and treated with DNAse I (Applied Biosystems, Carlsbad, CA). cDNA was synthesized using the reverse transcriptase reaction (High-Capacity cDNA Reverse Transcription Kits, Applied Biosystems) following manufacturer\\'s protocol. Primers for polymerase chain reactions (PCR) were from Applied Biosystems; adiponectin (catalogue \\# Mm00456425_m1), leptin (Mm00434759_m1), TNF-\u03b1 (Mm00443258_m1), IL-10 (Mm00439614_m1), IL-6 (Mm00446190_m1), MCP-1 (Mm00441242_m1), mMCP6 (Mm01301240_g1), F4/80 (Mm00802529_m1), and 18S (Hs03928990_G1). Quantitative gene expression by real-time PCR was performed using 40 ng of cDNA amplified with TaqMan Universal PCR Master Mix (\\#4352042) and reactions run using universal cycling conditions on a StepOnePlus real-time PCR System (Applied Biosystems). Samples were analyzed in triplicate and were normalized to the housekeeping gene, 18S. To analyze relative quanitation (RQ), the \u0394\u0394CT (threshold cycle) method was used \\[[@B14]\\]. The RQ results were expressed as the fold change in gene expression relative to expression levels in the caudal epididymal fat of lean mice. *Statistics*Results are presented as mean \u00b1 SEM. Unpaired Student\\'s t test was used to assess for statistically significant differences between groups. Comparisons between multiple groups were carried out using one-way analysis of variance (ANOVA) with Tukey post-hoc analysis. GraphPad Prism software (5.0a) was used for calculations (GraphPad Software, La Jolla, CA).\n\nResults\n=======\n\nMice with long-standing diet-induced obesity had a lower epididymal fat mass\n----------------------------------------------------------------------------\n\nCompared with mice fed on a LFD (n = 10), HFD-fed mice (n = 10) demonstrated significantly higher fasting blood glucose, serum insulin and cholesterol concentrations, and HOMA-IR (Table [1](#T1){ref-type=\"table\"}). In addition, HFD-fed mice had significantly higher body weights and larger livers and inguinal subcutaneous fat depots (Table [1](#T1){ref-type=\"table\"}). However, the epididymal fat of diet-induced obese mice weighed less than that of lean controls (Table [1](#T1){ref-type=\"table\"}).\n\n###### \n\nMetabolic characteristics of mice\n\n  Variables                    obese (n = 10)   lean (n = 10)   *p*value\n  ---------------------------- ---------------- --------------- ----------\n  Body weight (g)              48.7 \u00b1 1.0       36.7 \u00b1 0.7      \\< 0.001\n                                                                \n  Epididymal fat weight (g)    0.56 \u00b1 0.02      0.86 \u00b1 0.04     \\< 0.001\n                                                                \n  Inguinal fat weight (g)      1.33 \u00b1 0.05      0.60 \u00b1 0.03     \\< 0.001\n                                                                \n  Liver weight (g)             3.05 \u00b1 0.19      1.81 \u00b1 0.06     \\< 0.001\n                                                                \n  Blood glucose (mmol/l)       8.44 \u00b1 0.72      4.27 \u00b1 0.33     \\< 0.001\n                                                                \n  Serum insulin (ng/ml)        2.3 \u00b1 0.2        0.6 \u00b1 0.1       \\< 0.001\n                                                                \n  HOMA-IR                      25.3 \u00b1 3.7       3.4 \u00b1 0.6       \\< 0.001\n                                                                \n  Serum cholesterol (mmol/l)   3.99 \u00b1 0.21      3.31 \u00b1 0.16     \\< 0.001\n\nIn obese mice, epididymal fat mass correlated inversely with liver mass and body weight\n---------------------------------------------------------------------------------------\n\nIn lean mice, there was a trend for positive correlation between the mass of the epididymal fat and the mass of the liver (Figure [2A](#F2){ref-type=\"fig\"}). Moreover, the mass of the epididymal fat correlated positively with body weight and the mass of the inguinal subcutaneous fat (Figure [2B-C](#F2){ref-type=\"fig\"}). In obese mice, however, the mass of the epididymal fat correlated inversely with the mass of the liver (Figure [2D](#F2){ref-type=\"fig\"}). Moreover, there was a trend for an inverse correlation between the mass of the epididymal fat and the body weight (Figure [2E](#F2){ref-type=\"fig\"}). There was no correlation between the mass of the epididymal fat and that of the inguinal subcutaneous fat (Figure [2F](#F2){ref-type=\"fig\"}).\n\n![**The body weight and the weights of the epididymal and inguinal subcutaneous fat depots**. Relations between the body weight and the weights of the epididymal and inguinal fat depots in lean (A-C) and longstanding obese (D-F) mice are shown. The weight of the epididymal fat correlated inversely with the weight of the liver and the body weight in long-standing obese mice.](1476-511X-10-198-2){#F2}\n\nMast cells were distributed differentially in the epididymal fat\n----------------------------------------------------------------\n\nIn lean mice, there was a trend for higher mast cell density in caudal (0.4 \u00b1 0.1 cells per mm^2^) than in medial (0.2 \u00b1 0.1 cells per mm^2^) and rostral (0.1 \u00b1 0.1 cells per mm^2^) zones (Figures [3A-D](#F3){ref-type=\"fig\"}). In obese mice, however, the density of mast cells increased substantially and progressively from caudal (1.4 \u00b1 0.2 cells per mm^2^) to rostral (32.2 \u00b1 4.4 cells per mm^2^) (Figures [3E-H](#F3){ref-type=\"fig\"}). Although obesity was accompanied by a significant increase in mast cells in all three zones (Figures [3I-K](#F3){ref-type=\"fig\"}), the greatest increase was observed in the rostral epididymal fat where a 230-fold increase was noted (Figure [3L](#F3){ref-type=\"fig\"}). Ultrastructural examination confirmed the presence of frequent mast cells in the rostral epididymal fat of obese mice (Figures [3M-O](#F3){ref-type=\"fig\"}). Mast cells were intermingled with macrophages in the expanded interstitial space between adipocytes (Figures [3N,O](#F3){ref-type=\"fig\"}).\n\n![**Differential distribution of mast cells in the epididymal fat**. The distribution of mast cells in the epididymal fat from lean (A-D) and diet-induced obese (E-H) mice (n = 10 per group) are shown. The density of mast cells in the rostral, medial, and caudal epididymal fat was compared between lean and obese mice (I-L). Ultrastructural examination revealed frequent mast cells (arrows) intermingled with macrophages (arrowheads) in the expanded interstitial space between adipocytes (stars) in obese mice (M-O). Scale bars: 50 \u03bcm or as indicated. \\*\\*\\* *p*\\< 0.001.](1476-511X-10-198-3){#F3}\n\nMouse mast cell protease-6 (mMCP6) protein expression was increased in the epididymal fat of obese mice\n-------------------------------------------------------------------------------------------------------\n\nWestern blot analysis demonstrated that mMCP6 expression was higher in the epididymal fat of obese vs. lean mice (Figures [4A,B](#F4){ref-type=\"fig\"}). Although mMCP6 protein expression was higher in rostral than in caudal epididymal fat of obese mice, the difference was not significant in lean mice (Figures [4C,D](#F4){ref-type=\"fig\"}). In addition, there was a positive correlation between mMCP6 and F4/80 gene expression in the epididymal fat, especially in obese mice (Figures [4E-H](#F4){ref-type=\"fig\"}).\n\n![**Mouse mast protease-6 (mMCP6) expression in the epididymal fat**. Western blot analysis of mMCP6 protein expression in the rostral and caudal epididymal fat from lean (n = 4, white bars) and diet-induced obese (n = 4, black bars) mice is shown (A-D). Correlations between mMCP6 and macrophage-specific F4/80 gene expression are also demonstrated (E-H). \\* *p*\\< 0.05, \\*\\* *p*\\< 0.01.](1476-511X-10-198-4){#F4}\n\nCLS were distributed differentially in the epididymal fat\n---------------------------------------------------------\n\nIn lean mice, CLS were more prevalent in rostral (0.8 \u00b1 0.2 CLS per mm^2^) than in medial (0.3 \u00b1 0.1 CLS per mm^2^) and caudal (0.03 \u00b1 0.03 CLS per mm^2^) epididymal fat (Figures [5A-D](#F5){ref-type=\"fig\"}). Similarly, the density of CLS was higher in rostral (92.6 \u00b1 6.2 CLS per mm^2^) than in medial (24.7 \u00b1 3.4 CLS per mm^2^) and caudal (6.7 \u00b1 0.8 CLS per mm^2^) epididymal fat of obese mice (Figures [5E-H](#F5){ref-type=\"fig\"}). Obesity was accompanied by a substantial increase in CLS in all three zones (Figures [5I-K](#F5){ref-type=\"fig\"}). Ultrastructural examination of the epididymal fat of lean mice demonstrated a delicate interstitium separating adipocytes (Figure [5L](#F5){ref-type=\"fig\"}). In obese mice, however, macrophages and mast cells accumulated in the expanded interstitium and often enveloped adipocytes forming CLS (Figure [5M](#F5){ref-type=\"fig\"}). Furthermore, many macrophages contained numerous lipid droplets (Figure [5N](#F5){ref-type=\"fig\"}), while some were multinucleated (Figure [5O](#F5){ref-type=\"fig\"}).\n\n![**Differential distribution of crown-like structures (CLS) in the epididymal fat**. The densities of CLS in the epididymal fat from lean (A-D) and diet-induced obese (E-H) mice (n = 10 per group) are shown. The density of CLS in the rostral, medial, and caudal epididymal fat was compared between lean and obese mice (I-K). Ultrastructural examination demonstrated a delicate interstitium in the epididymal fat from lean mice (L). In obese mice, the expanded interstitial space contained macrophages (arrowheads), many of which enveloped scattered adipocytes forming CLS (stars) (M). While many macrophages were packed with lipid droplets (N), some were multinucleated (O). Scale bars: 50 \u03bcm or as indicated. \\* *p*\\< 0.05, \\*\\* *p*\\< 0.01, \\*\\*\\* *p*\\< 0.001.](1476-511X-10-198-5){#F5}\n\nReduced leptin, adiponectin, IL-6, and MCP-1 gene expression accompanied apoptosis in the epididymal fat of obese mice\n----------------------------------------------------------------------------------------------------------------------\n\nOn ultrastructural examination of the epididymal fat of obese mice, frequent macrophages and mast cells were found abutting lipid cores with no evidence of adipocyte cytoplasm indicating loss of adipocytes (Figure [6A](#F6){ref-type=\"fig\"}). Rarely observed in the epididymal fat of lean mice, apoptosis, detected by immunostaining for cleaved caspase-3, was frequent in the epididymal fat of obese mice, especially in the rostral zone (Figures [6B-E](#F6){ref-type=\"fig\"}). Western blot analysis confirmed increased cleaved caspase-3 protein expression in the rostral, and to a lesser degree in the caudal, epididymal fat of obese mice (Figures [6F,G](#F6){ref-type=\"fig\"}). Compared to lean controls, leptin, adiponectin, IL-6, and MCP-1 gene expression was reduced in the epididymal fat of obese mice (Figures [6H-K](#F6){ref-type=\"fig\"}). There was also a trend for lower leptin and adiponectin expression in the rostral than in the caudal epididymal fat of obese mice (Figures [6H, I](#F6){ref-type=\"fig\"}).\n\n![**Adipocytokine gene expression and apoptosis in the epididymal fat**. Ultrastructural examination of the epididymal fat from diet-induced obese mice demonstrated frequent macrophages (arrowheads) and mast cells (arrow) abutting lipid cores (star) with no evidence of the cytoplasm of adipocytes (A). The expression of cleaved caspase-3 was determined by immunostaining (arrows, B-E) and Western blot analysis (F-G). Leptin, adiponectin, IL-6, and MCP-1 gene expression was determined in the caudal and rostral epididymal fat from lean (white bars) and obese (black bars) mice (H-K). Scale bars: 50 \u03bcm or as indicated. \\* *p*\\< 0.05, \\*\\* *p*\\< 0.01, \\*\\*\\* *p*\\< 0.001.](1476-511X-10-198-6){#F6}\n\nDiscordance between TNF-\u03b1 and IL-10 gene expression and immune cell accumulation in the epididymal fat of obese mice\n--------------------------------------------------------------------------------------------------------------------\n\nIn the epididymal fat of obese mice, the densities of mast cells and CLS were 22- and 14-fold, respectively, higher in the rostral vs. caudal zone (Figure [7A,B](#F7){ref-type=\"fig\"}). Similarly, F4/80 and mMCP6 gene expression levels were higher in the rostral vs. caudal zone (Figures [7C,D](#F7){ref-type=\"fig\"}). Although obesity was accompanied by increased TNF-\u03b1 and IL-10 gene expression levels in the epididymal fat, there was no significant difference between the caudal and the heavily infiltrated rostral zones (Figures [7E,F](#F7){ref-type=\"fig\"}). In fact, expression levels of TNF-\u03b1 and IL-10 genes relative to those of F4/80 and mMCP6 genes were significantly lower in the heavily infiltrated rostral vs. caudal epididymal fat (Figures [7G-J](#F7){ref-type=\"fig\"}).\n\n![**TNF-\u03b1 and IL-10 gene expression in the epididymal fat**. The distribution of CLS and mast cells and the expression of F4/80, mMCP6, TNF-\u03b1 and IL-10 genes in the caudal and rostral epididymal fat from lean (white bars) and diet-induced obese (black bars) mice are shown (A-F). TNF-\u03b1 and IL-10 gene expression relative to F4/80 and mMCP6 gene expression is also shown (G-J) \\* *p*\\< 0.05, \\*\\* *p*\\< 0.01, \\*\\*\\* *p*\\< 0.001.](1476-511X-10-198-7){#F7}\n\nIn the heavily infiltrated rostral epididymal fat of obese mice, positive correlations among TNF-\u03b1, IL-6, MCP-1, and IL-10 gene expression levels were lost\n-----------------------------------------------------------------------------------------------------------------------------------------------------------\n\nPositive correlations were noted among expression levels of TNF-\u03b1, IL-6, MCP-1, and IL-10 genes in the epididymal fat of lean mice (Figures [8A,B](#F8){ref-type=\"fig\"}). In obese mice, positive correlations persisted in the caudal epididymal fat (Figure [8C](#F8){ref-type=\"fig\"}). However, in the heavily infiltrated rostral epididymal fat of obese mice, there were no statistically significant correlations between gene expression levels of these cytokines (Figures [8D,E](#F8){ref-type=\"fig\"}).\n\n![**Interrelationship between inflammatory cytokines in the epididymal fat**. The correlations among expression levels of TNF-\u03b1, IL-10, IL-6, and MCP-1 genes in the caudal and rostral epididymal fat from lean and diet-induced obese mice are shown.](1476-511X-10-198-8){#F8}\n\nDiscussion\n==========\n\nWhile immune responses in the adipose tissue of mice fed on a high-fat diet over a relatively short period of time are studied extensively, little is known about advanced stages of adipose tissue inflammation in long-standing obesity. Here, we examined the effects of long-term high-fat feeding on adipocytokine gene expression in the epididymal fat of mice with long-standing obesity and correlated the findings with the extent of apoptosis and inflammatory cell infiltration. We showed that extensive loss of adipocytes in this fat depot was associated with reduced fat mass and diminished leptin, adiponectin, IL-6, and MCP-1 gene expression levels. We also showed that cell death and inflammatory cell infiltration occurred predominantly in the rostral zone of the epididymal fat of obese mice. In addition, we found an increase in the expression levels of TNF-\u03b1 and IL-10 genes in the epididymal fat of obese mice. However, TNF-\u03b1 and IL-10 expression levels relative to those of F4/80 and mMCP-6 was significantly lower in the heavily infiltrated rostral than in caudal epididymal fat.\n\nThere are structural and functional differences between the adipose tissue found in different anatomical locations. For example, subcutaneous and visceral fat differ in immune cell composition, insulin sensitivity, lipolysis, expression of glucose transporters, and obesity-associated adipocyte death and attendant inflammation \\[[@B3],[@B15]-[@B17]\\]. Biological and immunological differences also exist among various visceral fat depots. For example, we have shown that the prevalence of both mast cells and macrophage CLS are higher in epididymal and perinephric fat than in mesenteric fat of diet-induced obese mice \\[[@B3]\\]. Emerging evidence indicates that there is also substantial heterogeneity in the structure and function of superficial and deep abdominal subcutaneous fat \\[[@B18]\\]. Here, we show for the first time that there are distinct zones (regions) within an abdominal fat depot that exhibit different level of susceptibility to metabolic challenges. Although the mechanisms underlying adipocyte injury and death in obesity are under investigation, it has been proposed that exaggerated metabolic demands trigger stress responses in adipocytes that could lead to cell injury and death \\[[@B19]\\]. In addition, the disequilibrium between oxygen demand and supply in expanding adipose tissues in obesity can lead to local hypoxia, which might have a role in the pathogenesis of adipocyte dysfunction, altered adipokine expression, and adipose tissue inflammation \\[[@B20]\\]. Based on these observations, it could be argued that, within an abdominal fat depot, there are functionally diverse populations of adipocytes that show different levels of tolerance to metabolic challenges. Alternatively, adipose tissue expansion in obesity might not be uniform throughout the same depot leading to lower oxygen tension in areas of fat depots undergoing a more rapid expansion. Considering the severity of pathological changes in the present study, we elected not to determine the degree of hypoxia in the caudal and rostral epididymal fat of obese mice.\n\nConsistent with a previous report, the loss of the epididymal fat mass in obese mice was inversely related to the weight of the liver \\[[@B11]\\]. Therefore, it could be argued that lipids released from dead adipocytes, if not locally metabolized by phagocytes, would possibly find their way to the liver and/or other fat depots. Alternatively, the diminished ability of a smaller epididymal fat depot containing many dysfunctional or dead adipocytes would result in energy storage in the liver and/or other fat depots.\n\nInflammatory cytokines play a significant role in the pathogenesis of insulin resistance \\[[@B1]\\]. IL-6 is reported to induce insulin resistance in the liver, the skeletal muscle, and 3T3-L1 adipocytes \\[[@B21]-[@B23]\\]. While IL-6 secretion from explanted adipose tissue from obese human subjects is well documented \\[[@B24]\\], studies comparing IL-6 protein and/or gene expression between lean and obese subjects are rather sparse. Nevertheless, a recent study showed that IL-6 gene expression was increased in the epididymal fat of mice fed on a high-fat diet (42% calories from fat) for 6 weeks \\[[@B10]\\]. Similarly, the perigonadal adipose tissue of 8-week old obese KK-*A*^*y*^mice, a model of genetic obesity, demonstrated elevated IL-6 gene expression levels \\[[@B25]\\]. Loss-of function and gain-of-function mutations showed that MCP-1 play a role in macrophage accumulation and insulin sensitivity in the adipose tissue \\[[@B26]\\]. MCP-1 was overexpressed in white adipose tissue of mice fed on a high-fat diet (56% calories from fat) for 12 weeks \\[[@B26]\\]. In addition, compared to mice on control diet, MCP-1 gene expression was greater in the epididymal fat of 20-week old mice after 16 weeks of high fat feeding (45% calories from fat) \\[[@B2]\\]. In contrast to these studies, here we used mice fed on a diet with higher fat content (60% calories from fat) for a longer period of time (20 weeks) and showed a reduction in IL-6 and MCP-1 gene expression in the epididymal fat. Strissel et al. randomized 5-week-old C57BL/6 mice to either a low-fat diet (10% calories from fat) or a high-fat diet (60% calories from fat) for up to 20 weeks and compared adipocytokine gene expression in the epididymal fat of obese mice of different age \\[[@B11]\\]. They observed a trend for decreased MCP-1 and IL-6 gene expression between weeks 16 and 20 of high-fat feeding \\[[@B11]\\]. Since IL-6 and MCP-1 are secreted mainly by the stromal vascular fraction of the adipose tissue \\[[@B27]\\], we argue that reduced IL-6 and MCP-1 gene expression in the epididymal fat of mice with long-standing obesity is a manifestation of reduced pro-inflammatory activity of immune cells in the adipose tissue.\n\nA potent anti-inflammatory cytokine, IL-10 is secreted by macrophages, dendritic cells, mast cells, and subsets of B- and T-lymphocytes \\[[@B28]\\]. Recent evidence suggests that IL-10 has favorable effects on insulin sensitivity and adipocyte function. Hyperinsulinemic-euglycemic clamp studies showed that IL-10 prevents IL-6-induced insulin resistance in the liver and the skeletal muscle in mice \\[[@B21]\\]. In addition, IL-10 decreased MCP-1 production and ameliorated TNF-\u03b1-induced inhibition of insulin-stimulated glucose uptake by 3T3-L1 cells \\[[@B29]\\]. Consistent with previous reports, we found increased IL-10 and TNF-\u03b1 gene expression in the epididymal fat of obese mice \\[[@B30],[@B31]\\]. However, IL-10 and TNF-\u03b1 gene expression relative to those of F4/80 and mMCP6 was reduced in the rostral epididymal fat where macrophages and mast cells were abundant. Since IL-10 and TNF-\u03b1 are expressed and released mainly from the stromal vascular fraction of the adipose tissue \\[[@B27]\\], these findings provide further evidence for altered inflammatory activity of immune cells in the epididymal fat depot of mice with long-standing obesity.\n\nAdiponectin and leptin are cytokines produced mainly by adipocytes. Adiponectin promotes insulin sensitivity and, by modulating innate and adaptive immune responses, exerts potent anti-inflammatory effects \\[[@B32]\\]. Consistent with previous reports, here we found reduced adiponectin gene expression in the epididymal fat of obese mice \\[[@B33],[@B34]\\]. However, our finding of reduced leptin gene expression in the epididymal fat of mice with long-standing obesity is not consistent with previous reports \\[[@B20],[@B35]\\]. We argue that high numbers of dysfunctional or dead adipocytes in a fat depot could lead to the reduced expression of proteins that are primarily synthesized by adipocytes. In fact, the loss of adipocytes, at least in part by apoptosis, in the epididymal fat of obese mice in the present study was so severe that it led to a decrease in fat mass, a finding that has also been observed by others \\[[@B11]\\]. Consistent with these findings, we recently reported reduced adiponectin and leptin gene expression in the visceral and inguinal subcutaneous fat of aP2-nSREBP-1c transgenic mice, a model of lipodystrophy \\[[@B36]\\]. Since mice have several distinct fat depots in various anatomical locations, the contribution of diminished leptin and adiponectin gene expression in the epididymal fat to their circulating concentrations in obese mice remains to be determined. However, considering paracrine actions of leptin and adiponectin, it could be argued that decreased leptin and adiponectin expression would affect the behavior of adipocytes and cells of the stromal vascular fraction, including immune cells, present in the adipose tissue. This is of particular interest considering the differences in relative TNF-\u03b1 and IL-10 expression between the caudal and the rostral epididymal fat of obese mice in the present study.\n\nThe underlying mechanisms that render adipocytes susceptible to metabolic challenges and provoke immune responses in the adipose tissue remain largely elusive. Zonal (regional) distribution of metabolically challenged adipocytes and associated immune responses in a single fat depot could serve as a useful model for studying the underlying mechanisms. Furthermore, an understanding of the mechanisms involved in the downregulation of immune cell activity in inflamed adipose tissue in obesity might help identify molecular targets of therapeutic relevance.\n\nAbbreviations\n=============\n\nCLS: Crown-like structures; HFD: High-fat diet; HOMA-IR: Homeostatic model assessment-insulin resistance; IL-6: Interleukin-6; IL-10: Interleukin-10; LFD: Low-fat diet; MCP-1: Monocyte chemotactive protein-1; mMCP6: Mouse mast cell protease-6; TNF-\u03b1: Tumor necrosis factor-\u03b1.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nMMA, MAR, BN, LMO and AJM researched data and reviewed/edited manuscript. KBJ and GM contributed to discussion and reviewed/edited manuscript. PZ researched data. AP and JR reviewed/edited manuscript. AN designed experiments, researched data, and wrote manuscript.\n\nAcknowledgements\n================\n\nThis study was made possible by generous funds from the Katz Family Foundation. Authors wish to express their gratitude to Christian Faul and Pedro Salas at the University of Miami for critically reviewing the manuscript and providing insightful comments.\n"}
+{"text": "THIS ARTICLE focuses on how the environment influences the ability of people with aphasia to communicate successfully. Aphasia intervention has traditionally focused on the language and communication of the individual and their close communication partners. A new trend in aphasiology, informed by the World Health Organization\\'s (WHO\\'s) International Classification of Functioning, Disability and Health (ICF; [@R50]), recognizes the importance of the environment in enabling or disabling the functioning and participation of people with aphasia.\n\nThe ICF defines *Environmental Factors* as \"those factors in the physical, social, and attitudinal environment in which people live and conduct their lives\" ([@R50], p. 10). Researchers have begun to identify how environmental factors influence the ability of people with aphasia to communicate ([@R23], [@R24]; [@R36]) and importantly, which environmental factors can be modified to improve communication and participation ([@R25]; [@R28]; [@R40]; [@R43]). These research endeavors inform the understanding of the kinds of environments that create barriers to or facilitators of communication for people with aphasia. However, this research is focused on understanding how the environment impacts the *present* communication and participation of people with aphasia. The chronic nature of aphasia necessitates that researchers and clinicians also consider how people\\'s circumstances and, therefore, their communication needs and participation needs change over time. This, in turn, requires clinicians to consider the many different environments that impact the communication and participation of people with aphasia in the long term and the consequences on their future health and well-being.\n\nThe importance of exploring the interplay between aphasia, the environment, and long-term health and well-being was highlighted when the first author met Hank.[1](#fn1){ref-type=\"fn\"} They were both members of a volunteer aphasia advocacy group that met once every two months to develop services and provide a support network for people living with aphasia. During the four years they worked together, Hank gave several presentations to community groups and to students about his stroke and living with aphasia. Through listening to these presentations, as well as through social conversation with Hank, the author got to know him well. His story raises important questions about the cumulative impact of the environment on the communication, participation, as well as health and well-being of people with aphasia over time. Hank\\'s story may cause clinicians to reconsider how they conceptualize the impact of aphasia over time. Questions about long-term impacts of environmental factors are explored in detail later.\n\nIn Part 1 of this article, with Hank\\'s permission, the first author recalls his story and how his life changed and evolved in the 15 years since the onset of aphasia. In Part 2, the authors explore how two different conceptual frameworks---the ICF ([@R50]) and the Social Determinants of Health model (SDH; [@R8])---account for the changes to Hank\\'s health and well-being over time. Finally, Part 3 includes a review of the research literature to explore the evidence for a relationship between aphasia and the SDH.\n\nPART 1: HANK\\'S STORY\n=====================\n\nHank, a white Australian man, was in his 40s and married with four children at the onset of aphasia. He worked full-time for a large national company negotiating complex sales and contracts and his wife, Beth, worked part-time in retail. Although he often travelled interstate on business, he was working in his home town of Melbourne when he suddenly collapsed and was rushed to the nearest hospital. Subsequently, Hank was taken to one of the large publically funded tertiary hospitals that serve Melbourne\\'s population of 4.5 million people.\n\nHis recollections of those days in hospital were vague; however, he recalled the doctors telling him that he had had three strokes in that first week. His only other memories were that he could hardly talk or walk. After several months of inpatient rehabilitation, he was discharged home. At the time, he had an expressive and receptive aphasia and a persistent right-sided weakness but was able to walk with a stick.\n\nDespite the improvements he had made, his aphasia meant that he could not go back to his former job. As a result, he lost the enjoyment and challenge of work that he loved, he lost contact with work colleagues, and he lost his substantial income. The financial implications were serious. All four children were attending a private school. The eldest child finished her final year of high school, but the family could no longer afford the school fees for the younger children, who had to move to the local government-funded secondary school. Thus they lost contact with families they had known in the school community for years.\n\nHank reported that the stress of all the changes was terrible, and he and his wife separated within 2 years. Hank moved into a private rental on his own. He was offered work through a disability employment agency, which entailed packing envelopes in a mailroom alongside people with intellectual disabilities. He reported attending for a while, but ultimately he found it too depressing, so he left.\n\nWith only a disability pension as income, Hank found living on his own too expensive, so he moved into accommodations where the costs of rent and bills are shared. He said he did not like living in a shared house, and the house was not adapted to suit his needs, as there were no rails on the stairs or in the bathroom. Furthermore, he said his aphasia made it difficult for him to find the right person to live with, and he was worried that his housemate would take advantage of him. As he was living on a disability pension, he was eligible for public housing. He was looking forward to having his own place but there was a five-year waiting list. Around that time, Hank was diagnosed with depression.\n\nIn talking about how he managed every day, Hank said that he was okay but that he still needed occasional help with communication. For example, when his car needed to be fixed, he reported that he did not really understand what the mechanic was talking about; therefore, he did not really know what the problem was with his car and was worried that he had been charged too much to get it fixed. He also said he needed help with his computer; his e-mail often did not work and he could not work out how to fix it.\n\nAfter Hank did not attend the aphasia group for two months, the author learned that he had fallen on the stairs at home and broken his ankle. He was admitted to hospital and had another short period of inpatient rehabilitation. Back in his share house, he said his mobility was worse but he was now the highest priority on the public housing list. The following month, he moved into his own single-level, one-bedroom unit.\n\nReflecting on the events that occurred in the 15 years following Hank\\'s stroke provides a unique insight into one Australian man\\'s experience of living with aphasia. The immediate consequences of his stroke and aphasia, such as being unable to return to his former job, would be familiar to clinicians and researchers. But Hank\\'s story suggests that there are long-term consequences of aphasia consequences of aphasia. In other words, there are secondary consequences of the initial consequences of aphasia that are not immediately apparent. In Hank\\'s case, one could speculate that the stroke and aphasia contributed directly to his economic challenges and indirectly to the demise of his marriage, both of which, in turn, contributed to his depression. Furthermore, it is probable that his financial difficulties and separation meant Hank had to live in a shared house where he felt unsafe. One could even surmise that these consequences contributed to the deterioration of his health, culminating in his broken ankle.\n\nHank\\'s story is unique, but it is not an isolated example. It raises questions that have broader implications, including the essential question addressed in this article: Is there a conceptual framework that can explain the sequence of challenging events that can follow stroke and aphasia in cases like Hank\\'s? In the following sections, the authors explore how two different conceptual frameworks could account for Hank\\'s story and the implications each might have for reducing barriers to long-term health and participation for people with aphasia.\n\nPART 2: APPLYING DIFFERENT CONCEPTUAL MODELS TO HANK\\'S STORY\n=============================================================\n\nApplying the ICF to Hank\\'s story helps to understand the current consequences\n------------------------------------------------------------------------------\n\nThe ICF is a biopsychosocial framework of health and health-related conditions ([@R50]). It has been used in speech--language pathology to describe communication and swallowing disability ([@R30]; [@R48]). The ICF consists of four components: Body Functions and Structures (Impairments), Activities and Participation (Activity Limitations and Participation Restrictions), Environmental Factors (Barriers and Facilitators), and Personal Factors, all of which contribute to a person\\'s experience of functioning, disability, and health ([@R50]).\n\nThe ICF provides a way to conceptually understand the different factors that influence a health condition such as aphasia. A speech--language pathologist (SLP) working with Hank could apply the ICF to Hank\\'s situation to capture the wide-ranging factors influencing his aphasia. For example, the clinician could describe his language deficits (the Impairment) or describe the consequences of the language deficits in terms of his communication Activity Limitations and Participation Restrictions. These might include the range of communication Activity Limitations related to his interpersonal relationships, such as difficulty expressing feelings, difficulty understanding intent, and difficulty adjusting to a change in topic ([@R14]). Furthermore, the SLP could consider how these different communication Activity Limitations combine to result in Participation Restrictions, such as difficulty in maintaining his spousal relationship.\n\nThe ICF framework also provides a useful framework to think about how the communication environment might be creating barriers to or facilitating Hank\\'s ability to communicate and participate. Consideration of environmental factors might prompt the SLP to wonder if Hank\\'s important communication partners, such as his wife and daughters, have the knowledge and skills to communicate with Hank. Finally, the ICF\\'s Personal Factors component might prompt the SLP to consider if Hank\\'s personality, his knowledge of stroke and aphasia, and his ability to cope with such a dramatic loss might also be influencing his ability to communicate. Some of these factors are depicted in Figure [1](#F1){ref-type=\"fig\"}. The ICF framework can be applied to explore the factors influencing Hank\\'s aphasia at any point in time after his stroke. Figure [2](#F2){ref-type=\"fig\"} provides a hypothetical example of the factors that may have influenced Hank\\'s experience of aphasia many years after his stroke.\n\n![The factors hypothesized to influence Hank\\'s communication activity and participation in the months after his stroke as conceptualized by the ICF. Adapted from *International Classification of Functioning, Disability and Health*, by World Health Organization, 2001, Geneva, Switzerland: Author.](tland-37-85-g001){#F1}\n\n![The factors hypothesized to influence Hank\\'s communication activity and participation many years after his stroke as conceptualized by the ICF. Adapted from *International Classification of Functioning, Disability and Health*, by World Health Organization, 2001, Geneva, Switzerland: Author.](tland-37-85-g002){#F2}\n\nIn summary, the ICF ([@R50]) provides a conceptual framework to describe Hank\\'s functioning and disability at a particular moment in time. However, it does not provide a way to conceptualize how current level of functioning and disability, such as difficulty participating in relationships or difficulty reading complex information, may influence health and well-being over time. Is there another conceptual model that offers guidance in regard to potential long-term consequences?\n\nApplying the SDH to Hank\\'s story helps determine the possible long-term consequences\n-------------------------------------------------------------------------------------\n\n### The social determinants of health\n\nThe SDH model provides a different conceptualization of health: one that identifies the social factors that influence health and well-being in the long term ([@R45]). The SDH may provide a better account of the factors contributing to Hank\\'s dramatic change in circumstances from being a successful executive, living with his wife and children with financial means and several social networks, to being an unemployed recipient of a disability pension, living in shared accommodations where he does not feel comfortable or safe. The SDH model, as depicted in Figure [3](#F3){ref-type=\"fig\"}, states that social factors (such as the socioeconomic and political context, and a person\\'s subsequent social position) directly influence a person\\'s material, social, psychological, and biological circumstances, which in turn influence a person\\'s health and well-being ([@R45]). According to the SDH, it is not simply that disease or injury causes poor health and well-being; rather, it is the complex interplay of social factors that determine the likelihood that an individual will experience disease or injury that leads to poor health and well-being.\n\n![The Social Determinants of Health model. Amended from Solar & Irwin (2007). Reproduced with permission from *Closing the Gap in a Generation: Health Equity Through Action on the Social Determinants of Health,* by Commission on Social Determinants of Health, 2008. Final report of the Commission on Social Determinants of Health. Geneva, Switzerland: Author.](tland-37-85-g003){#F3}\n\nA wealth of evidence world wide demonstrates how different social factors influence population health and well-being ([@R8]). For example, mortality rates for both men and women steadily increase as poverty increases; this pattern is evident both within individual countries and across different countries ([@R8]). Mortality rates are related to ethnicity, education, and income. For example, indigenous Australians live approximately 16 years less than non-indigenous Australians ([@R8]). In the United States, men with a low education live on average 6.5 years less than men with university degrees ([@R11]). Similarly, women on low incomes live on average 5 years less than affluent women ([@R11]). As [@R11] states, some differences in population health, such as health differences due to genetic factors, may be unavoidable. These differences are called *health inequalities*. However, if the variations are avoidable, like the variations in health described earlier, then they are no longer considered to be health inequalities but *health inequities* ([@R11]). Health inequities occur because of the way society is structured and the way in which resources are distributed within society ([@R8]).\n\nThe SDH model proposes that there are *structural determinants of health* such as the socioeconomic and political context that influence social position in society. The socioeconomic and political context is proposed to influence a person\\'s access to education, income, and occupation. The socioeconomic and political context may also have differential impact on people, depending on their gender and ethnicity. These structural determinants are hypothesized to effect *intermediary determinants of health* such as a person\\'s material circumstances, psychosocial factors, and behaviors. The health system in which an individual finds himself or herself is also considered an intermediary determinant of health because it influences a person\\'s access to health services, such as the specific criteria around access to health care and health care resources ([@R45]). Apart from an individual\\'s behaviors and biological factors, all the other factors in the SDH model are environmental factors. Some factors such as material circumstances and psychosocial networks are part of a person\\'s immediate environment, whereas factors such as the socioeconomic context, the political context, and the health care system are part of a person\\'s broader environment. Can the SDH model be applied to individuals such as Hank, to explain the series of events that occurred to him after his stroke and aphasia?\n\n### Applying the SDH to Hank\\'s story\n\nThe SDH is a cyclical model that depicts how an individual\\'s health and well-being are influenced by structural and intermediary determinants of health. Therefore, in applying the SDH to Hank\\'s story, the authors acknowledge that there were structural and intermediary determinants of health already operating that had influenced Hank\\'s health and well-being prior to his stroke. These included the prevailing Australian socioeconomic and political context at the time (e.g., Australia is a stable democracy; it is a wealthy, developed nation; free public health care is available), as well as Hank\\'s prestroke social position (e.g., white male, high school educated, high status occupation, and high income), his preexisting material circumstances (e.g., whether or not he could afford safe comfortable housing, healthy food, adequate clothing), the degree of social cohesion (e.g., the safety of his neighborhood), psychosocial circumstances (e.g., his social networks, his stress levels), behavioral circumstances (e.g., the extent to which he ate well, slept well, and exercised), and biological factors (e.g., his preexisting medical conditions). The application of SDH to conceptualize the consequences of aphasia over time begins at Hank\\'s discharge from formal rehabilitation services approximately 3 months after his stroke and consider the possible interplay of social determinants of health from this time on. Although all social determinants influence health and well-being, the particular social determinants thought to play a key role in influencing Hank\\'s health and well-being are highlighted in Figure [4](#F4){ref-type=\"fig\"}.\n\n![Some of the factors hypothesized to influence Hank\\'s participation as conceptualized by the Social Determinants of Health. Adapted from Solar & Irwin (2007).](tland-37-85-g004){#F4}\n\n### How might structural determinants influence Hank\\'s long-term health and well-being?\n\nThe SDH leads to reflection on how structural determinants of health may have influenced Hank\\'s health and well-being after his stroke, subsequent aphasia, and reduced mobility. That is, were there any factors in the environment related to government and reflected in Australian macroeconomic, social, and/or health policies that impacted Hank\\'s health and well-being? Did Australian cultural and societal norms and values regarding stroke and disability affect Hank? Two structural determinants that influence a person\\'s social position, income, and occupation are considered later.\n\nHank\\'s stroke, aphasia, and reduced mobility meant that he could no longer participate in a high-status, high-income job. His inability to resume his prestroke occupation had many consequences, including a dramatic reduction in income. However, government policy did play a role. Although considerably less than the income he had previously received, Hank received some income through a government-funded disability pension. This pension also entitled him to a concession on the cost of some essential services, such as electricity. The overall value of this pension would determine the extent to which it served as a protective factor for Hank\\'s health and well-being.\n\nThere was also a government service to support people with acquired disabilities to find employment. However, it is important to consider whether there were any environmental factors operating within these government services that created barriers for Hank, given his aphasia. Specifically, did the people responsible for finding Hank employment after his stroke understand what aphasia is? Were they able to communicate with him in a way that revealed his competence (see [@R27])? It is unknown if the disability employment staff had an understanding of aphasia and provided Hank with a supportive communicative environment or not. Instruction in aphasia and supportive communication may have benefited their assessment of his capabilities and the work opportunities he was offered.\n\n### How might intermediary determinants influence Hank\\'s long-term health and well-being?\n\nThe SDH proposes that social position, indicated by occupation and income, has a direct bearing on intermediary determinants of health such as psychosocial factors, material circumstances, and access to health care. The SDH may be useful when considering the long-term impact of these environmental factors on Hank\\'s health and well-being. The authors aim to initiate a conversation within aphasiology regarding the long-term impact of aphasia that can inform future research, service planning, and delivery, and resource allocation.\n\n#### Psychosocial factors\n\nThe SDH model suggests that the fall in Hank\\'s social position as a result of his loss of occupation and income may have resulted in a range of negative psychosocial consequences. These might include living with the stress of debt, the loss of his work-related social network, the loss of the school social network, and the breakdown of his marriage. It is also important to consider the interactions among aphasia, the environment, and these psychosocial consequences. Whereas the environmental supports provided to Hank were unknown, the SDH can prompt clinicians to consider what they might need to be. For example, if there were a communicatively accessible financial counselling service, Hank and his wife would have been able to get financial advice and support in a way that Hank could understand. This support might have been in the form of supporting Hank to continue to be involved in making financial decisions and/or giving him the opportunity to be involved in appointing someone to assist with these decisions.\n\nSimilarly, as Hank and his wife experienced their relationship deterioration, communicatively accessible relationship counselling services might have meant that they would have been able to get the psychological and emotional support that they needed. The consequences of losing his job and his income also meant that Hank lost important social networks. Environmental factors related to the knowledge, skills, and attitudes of family, friends, school colleagues, as well as work colleagues become critical. Having communication partners with the skills and knowledge to support a person with aphasia in conversation ([@R44]) and having opportunities to develop new social networks (see Howe, this issue) may have provided Hank with a communicatively supportive social network. The loss of social networks alone can be detrimental to feelings of belonging and overall health and well-being ([@R10]).\n\n#### Material circumstances\n\nAnother important intermediary determinant of health is a person\\'s material circumstances, such as housing ([@R45]). This prompts a number of important questions around how a range of environmental factors, such as the accessibility of legal services and the availability of advocacy services, may have contributed to Hank\\'s financial situation and the quality of his housing. Although Hank did not mention it explicitly in his story, thinking about the relationship between aphasia and the social determinants of health may prompt clinicians to consider what, if any, communicative support Hank received during the complex, legal process of divorce and how this might have affected the amount of money he received as part of the financial settlement. This is important because it would have had implications on many aspects of his life, including the quality of rental accommodation he could afford. A better financial settlement might have made the difference between living in a place where he felt safe or in a place where he did not feel safe.\n\nAnother environmental factor relates to the presence of advocacy services. There are no advocacy services for people with aphasia in Melbourne. If Hank had had access to an advocate, he might have been able to find affordable and acceptable accommodations for himself. Alternatively, an advocate might have been able to help him find a shared house where he felt both physically safe and psychologically comfortable. An advocate might also have supported Hank when more complex communication situations arose such as understanding his rights and responsibilities as a tenant, negotiating with his housemates on bills, negotiating with the landlord about getting stair-rails in place, and negotiating car repairs with the mechanic.\n\n#### The health care system\n\nAnother intermediary determinant of health is the health care system itself ([@R45]). The SDH provides a way of explaining how access or a lack of access to the health care system can serve to protect a person\\'s health or further compromise it. Health care policies determine who is able to access health care and the cost of this access. Given that Hank was in receipt of a disability pension and living in Australia, he would have received free transport to hospital, free hospital care, and free inpatient rehabilitation. A publically funded health care system facilitates access for all Australians to basic health care. However, on his admission to hospital with a broken ankle, the SDH model prompts clinicians to consider whether there were any environmental factors that could have influenced Hank\\'s access to this health care given that he had a preexisting aphasia.\n\nA recent metasynthesis of the environmental factors that influence health care for people with communication disabilities based on observations in Melbourne hospitals ([@R36]) identified that there are no systems in place to detect patients with preexisting communication disabilities, and there are no systems in place to equip health care providers with the knowledge, skills, and resources to support people with communication disabilities to participate in their health care. It is possible that health care staff were not aware that Hank had aphasia, that the SLPs were not aware of Hank\\'s admission, and that staff did not modify the information they provided him about the assessment and treatment of his broken ankle. Ineffective communication might have placed Hank at risk of a preventable adverse event in hospital ([@R20]) and might have undermined an optimal recovery ([@R46]).\n\nIn summary, the SDH model prompts consideration of the effect of aphasia and the environment beyond communicating in a particular activity or participating in a certain event. It encourages consideration of how aphasia, as a chronic condition, may make an individual more vulnerable to a set of conditions, such as poorer housing and fewer social supports, which, in turn, place him or her at greater risk of poorer health and well-being over time. For clinicians and researchers, the SDH identifies some environments that may be particularly important to target in order to enhance and protect the long-term health and well-being of people with aphasia. These include government policies, employment, income, education, psychosocial networks, and health care environments. The SDH also prompts new questions in terms of how aphasiologists define and measure the success of aphasia interventions. Finally, it highlights the need for support services for people living with aphasia when life circumstances change. Below, the authors review the research evidence to explore the relationship between aphasia and the structural and intermediary social determinants of health.\n\nPART 3: IS THERE EVIDENCE OF A RELATIONSHIP BETWEEN APHASIA AND SDH?\n====================================================================\n\nThis section provides an overview of the research evidence in relation to aphasia, the communicative environment, and SDH. The authors have not identified any studies that apply the SDH model to aphasia and its consequences. However, as described later, some literature exists examining aphasia and these health and well-being determinants.\n\nExploring the structural determinants of health and aphasia\n-----------------------------------------------------------\n\nWithin the SDH model, structural determinants of health refer specifically to \"interplay between the socioeconomic--political context, structural mechanisms generating social stratification and the resulting socioeconomic position of individuals\" ([@R45], p. 28). As such, the authors begin by discussing the evidence regarding the citizenship experiences of people with aphasia, and the environmental factors that influence their access to government services. The structural determinants of health both influence and are influenced by an individual\\'s socioeconomic position. The most important indicators for socioeconomic position are occupational status, level of education, and income level ([@R45]). To further illustrate the impact of aphasia on these indicators of socioeconomic position, the following provides an overview of the literature in relation to the influence of aphasia and the environmental factors that influence access to education.\n\n### Aphasia, civic engagement, and access to government agencies\n\nCitizenship is \"a concept which encompasses connection to wider society, rights and responsibilities, and the capability for exerting power and influence\" ([@R31], p. 187). People with aphasia largely define citizenship in terms of community involvement, although some suggest a broader definition involving dealing with government agencies ([@R31]).\n\nFindings from interviews with people with aphasia suggest that their desire for civic involvement is driven by a wish to engage in activities beyond the home, and by a desire to act as agents for change for both themselves and others in the populations they represented ([@R31]). An investigation by [@R23] found that the environmental factors that influenced the community participation of people with aphasia included (a) awareness of aphasia; (b) opportunity for participation; (c) familiarity; (d) availability of extra support for communication; (e) communication complexity; (f) message clarity; and (g) time available for communication. However, little is known about the relationship between environmental factors and participation, or the impact of environmental level interventions on the participation of people with aphasia.\n\nThe ability of people with aphasia to engage with government agencies has been reported to be fraught with challenges. As an example, people with aphasia have been found to experience difficulties accessing services through Centrelink ([@R4]), the public interface of the Australian Government\\'s Department of Human Services responsible for the provision of social security payment to, among others, people with disabilities. In interviews about their experiences with this government agency, people with aphasia reported that additional time, accessible interactions, information and processes, and service relationships with individualized approaches were needed to facilitate their engagement, but that none of these things was currently being provided.\n\n### Aphasia and access to education\n\nThe notion of access to education is important in addressing both equity within the education system, and opportunities that seek to grant equal opportunities in employment beyond course completion ([@R41]). There has been very little research on the environmental factors that influence access to education for people with aphasia. The number of people with aphasia who make the decision to either enter or return to higher or further education following the onset of aphasia is unknown.\n\nExisting literature on the experiences of people with aphasia in the educational sector describes both experiences in which the challenges of trying to access education lead to withdrawal (e.g., [@R38]) and those in which academic success was achieved despite these challenges (e.g., [@R6]; [@R38]). That is, in addition to the person\\'s language impairment, the findings reported in the literature revealed environmental barriers experienced by people with aphasia in the educational setting. These barriers can grossly be categorized into two areas: (1) poor awareness of aphasia and its consequences within the education sector and (2) a lack of responsive and flexible systems to support the needs of people with aphasia in educational settings. These categories are explored in further detail later.\n\n### Poor awareness of aphasia and its consequences within the education sector\n\nTo meet the needs of people with aphasia in an educational context, their needs must first be understood clearly by the educational provider. When staff have a lack of awareness or understanding about a student\\'s disability, students are more likely to have negative educational experiences ([@R22]). Facilitating an understanding of aphasia within educational settings is likely to be made difficult by a poor public awareness and understanding of the aphasia in the community ([@R7]). The \"invisible\" nature of aphasia may mean the obligation of disclosure and seeking accommodations lies with the student with aphasia ([@R35]), a process that has been reported as difficult for individuals with a variety of disabilities and may be made more difficult for people with aphasia as a result of their linguistic deficits.\n\n### Lack of responsive and flexible systems to support the needs of people with aphasia in educational settings\n\nBecause of the high-level language demands in educational settings and the communication difficulties inherent to aphasia, people with aphasia may be limited in their capacity to participate in formal education without learning support ([@R6]; [@R38]). Aphasia can restrict a person\\'s ability to engage with the process of education, which is normally conducted in the spoken and written language modalities, including their ability to meet linguistically focused outcome assessment requirements.\n\n[@R38] identified a number of specific teaching methods that contribute to educational success for people with aphasia. However, it has been reported that these methods are not commonly employed. Exploration of education service provision to people with aphasia by [@R26] in the United Kingdom identified a number of small-scale examples of collaboration between SLP services and adult education providers to meet the needs of people with aphasia; however, challenges related to conflicting philosophies were identified. Despite this, the value of collaborations between SLPs and academic staff was also identified by [@R6]. They suggest that the model and location of SLP services should enable people with aphasia to seek support from SLPs as their educational circumstances and needs change.\n\nExploring the intermediary determinants of health and aphasia\n-------------------------------------------------------------\n\nThe structural determinants of health for people with aphasia, as discussed earlier, operate through a series of intermediary determinants of health. Those structural determinants can be considered to influence these intermediary determinants. The main categories of intermediary determinants of health as outlined in the SDH model are material circumstances, social cohesion, psychosocial factors, behaviors, and biological factors. The following section reviews the evidence regarding environmental factors that influence access to interpersonal relationships for people with aphasia and access to the health care system for people with aphasia.\n\n### Aphasia and interpersonal relationships\n\nWithout the speed or ease of prestroke verbal output, are people with aphasia able to maintain and create new social relationships? Language has been described as the \"currency\" of relationships ([@R38], p. 44). Thus, the loss of language with aphasia deprives the individual of one of the fundamental ways to maintain relationships ([@R21]). There is substantial evidence to support the assertion that aphasia has a profound impact on interpersonal relationships, described as a psychosocial factor within the SDH. The effects of aphasia have been described in regard to many different kinds of relationships, including intimate relationships, relationships with children, and relationships with friends.\n\nThere are no data available on the number of marital relationships that break down following one person acquiring aphasia. There are some limited studies on the perceptions and experiences of spouses of people with aphasia that suggest that aphasia impacts spousal relationships in positive and negative ways ([@R33]; [@R34]). Recent research also indicates that many spouses of people with aphasia experience third party disability that may further compromise their ability to support the relationship. For example, family members of people with aphasia are at risk of developing depression ([@R16], [@R17]), deterioration in their own health ([@R18]), changes to their own social relationships ([@R15]), and recreational activities ([@R29]).\n\nPositive and supportive intimate relationships are a powerful contributor to living successfully with aphasia ([@R5]). Access to supportive and meaningful relationships was one of the seven themes identified in a qualitative meta-analysis of interview data from people with aphasia, family members, and SLPs ([@R5]). The authors concluded that their study reinforced \"the idea that living successfully with aphasia can only occur within the context of love, acceptance, friendship, and support from others\" (p. 146). This research supports the idea that access to and participation in meaningful personal relationships may serve to buffer people with aphasia against some of the negative impacts of aphasia.\n\nRelationships with children appear more resilient to the impact of aphasia. In a study of 83 people with chronic aphasia, 71% reported that they had the same amount of contact with their children following the stroke ([@R21]). This is in keeping with evidence from the broader stroke population (e.g., [@R1]). However, less is known about the relationships between younger children and their parents, when parents still have caring responsibilities (e.g., [@R19]).\n\nOutside family relations, evidence suggests that aphasia affects the nature and quality of interactions between people with aphasia and their friends ([@R38]). This manifests in reduced social networks and/or reduced quality of social relationships ([@R9]; [@R12]; [@R13]). [@R21] found that 64% of 83 people with aphasia interviewed reported reduced interactions with friends and 30% reported having no close friendships whatsoever. This loss of friendship is uniquely attributed to the presence of aphasia rather than the stroke more generally. A survey of people with aphasia in the United States found that 75% felt others avoided them because of their communication difficulty ([@R42]). The evidence to date creates a picture of shrinking social networks, while the nature of aphasia further restricts opportunities and ability to seek emotional and informational support ([@R12]). It is still possible for people with aphasia to develop and maintain friendships, but this requires the \"two way hard work of friendship\" as well as creativity and resourcefulness ([@R39], p. 354).\n\n### Aphasia and access to health care\n\nRecent research has addressed the factors that influence the success and failure of communication between patients (including those with aphasia and other communication disabilities) and health care providers across the continuum of health care ([@R3]). When patients and health care providers manage to establish shared meaning, positive health care outcomes are likely ([@R3]). Conversely, communication breakdowns within the health care sphere lead to negative health outcomes, increased length of hospital stay, higher rates of readmission, increased costs, an increase in negative events, and a reduction in patient satisfaction ([@R2]; [@R47]).\n\nWithout access to appropriate and relevant communication support, people with aphasia may experience a reduction in their capacity to express their health care needs and actively participate in decisions regarding their own health care ([@R37]). Poor communicative access within health care can lead to damaging and negative experiences for people with aphasia, including a reduction in the reported satisfaction with the health care experience ([@R49]), an increased risk of inappropriate or inadequate service provision ([@R20]), and an increased likelihood of the occurrence of adverse events ([@R2]).\n\nCONCLUSION\n==========\n\nThe SDH model offers clinicians and researchers a way to conceptualize the cumulative effect of inaccessible environments on the long-term health and well-being of people with aphasia. The authors are not suggesting that Hank\\'s story is typical for people who acquire aphasia, nor is his story rare. Aphasiologists have much to learn regarding how SDH can enrich approaches to aphasia assessment, decision making around the aims and type of the interventions offered, and measurement of intervention effects.\n\nFor SLPs, the SDH may challenge traditional views of the clinician\\'s role in the lives of people with aphasia. Viewing the impact of aphasia decades after the onset emphasizes the importance of understanding the impact of many different environments on people with aphasia. The SDH suggests that SLPs need to work beyond the level of the individual with aphasia in formal rehabilitation settings, to address the vast range of systems, services, and policies that have the potential to damage or enhance the long-term health and well-being of people with aphasia (see [@R32]).\n\nHank\\'s life following the onset of aphasia is one example of how many different facets of the environment appeared to influence his health and well-being for the long term. His story illustrates the disconnect between how an individual might present initially in a health care setting (e.g., with a clinical diagnosis of aphasia) and the foreseeable and unforeseeable impacts that aphasia may have on that person\\'s health and well-being in the future. Indeed Hank\\'s story serves to highlight the powerful (at times devastating) domino effect that can result from changes in one aspect of the model (i.e., biological factors), which turns a present-day health condition into a long-term social problem that impacts all aspects of life. Given the chronic nature of aphasia and that limited public health care resources are typically frontloaded to acute care and rehabilitation within the first year poststroke, the SDH poses interesting questions on how clinicians might serve the needs of people with aphasia by recognizing and reducing the environmental barriers to full participation in life to optimize health and wellbeing for the long term.\n\nPseudonym.\n\nAll the authors have indicated that they have no financial and no nonfinancial relationships to disclose.\n"}
+{"text": "INTRODUCTION\n============\n\nThe continuing emergence of drug resistant variants of the human immunodeficiency virus (HIV) calls for new tools to combat this epidemic ([@B1]). Current clinical treatment of HIV targets essential enzymatic activities associated with the reverse transcription of the viral genome and the protease activity required for maturation of protein components of the virus. The most effective control of the disease has been achieved with combination therapy targeting these pathways simultaneously. The inclusion of additional targets in a combination drug regimen is highly desirable for potent and long-term disease management.\n\nThe ribonuclease H (RNase H) activity associated with HIV-reverse transcriptase (HIV-RT) degrades the viral RNA genome in RNA/DNA hybrids ([@B2]), and has been identified as a potential target for antiretroviral therapy as it is required for virus infectivity ([@B3]); yet there are no antiRNase H agents in clinical use. Few inhibitors of HIV-1 RNase H were identified until the transition of testing methods from gel-based techniques to fluorescent assays amenable to high-throughput screening (HTS) ([@B4; @B5; @B6; @B7; @B8]). The most widely used assay was developed by Parniak and co-workers ([@B6]) and utilizes a two label, molecular beacon strategy ([@B9]) in which the RNA strand is labeled with a 3\u2032-terminal fluorophore (fluorescein, F) and a DNA strand with a quencher (dabcyl, Q) at the 5\u2032-terminus ([Scheme 1](#F7){ref-type=\"fig\"}). Scheme 1.Representation of fluorescent RNase H assay using a dual label system employing a fluorophore and a quencher.\n\nFluorescent studies involving nucleic acids most often utilize luminescent tags such as fluorescein or rhodamine sometimes in combination with a quencher such as 4-(dimethylaminoazo)benzene (dabcyl). However, these probes can be perturbing to the processes under investigation due to the steric bulk or non-polar groups they introduce to DNA and RNA. Through our own studies on inhibitors of HIV-1 RNase H, we have observed that the 5\u2032-dabcyl quencher substantially reduces the catalytic efficiency of RNase H for its RNA/DNA substrate (SI).\n\nThe use of intrinsically fluorescent nucleotide base analogs offers a more conservative approach to fluorescence labeling of nucleic acids. Using phosphoramidite solid-phase synthesis ([@B10]), a fluorescent nucleotide can be incorporated at any position on an oligomer without the use of linkers or postsynthetic modifications. In addition, the fluorescence intensity of many base modified nucleotides are responsive to changes in their microenvironment, making them reporters for nucleic acid structure and dynamics ([@B11]). Despite these advantages, few fluorescent nucleobase analogs have found widespread use as a tool for molecular biology, 2-aminopurine ([@B12]) and the pteridine ([@B13]) base analogs being historically the most important. Recent work from Tor and others have elegantly demonstrated the value and utility in the design and discovery of new fluorescent nucleobases ([@B14; @B15; @B16; @B17]). Despite these and other recent advances, there remains a paucity of intrinsically fluorescent cytidine analogs that demonstrate responsiveness to their microenvironment and state of hybridization ([@B18]) thus motivating this work.\n\nThe success of a fluorescent nucleobase analog as a useful reporter resides in its ability to form proper Watson--Crick base pairs, thermostability in double stranded nucleic acids, changes in fluorescence with different hybridization states, recognition of nucleic acid processing/binding enzymes, and its fluorescence intensity. The cytidine analogs pyrrolocytidine (pC, [Figure 1](#F1){ref-type=\"fig\"}) with 6-methyl substitution (MepC) ([@B19; @B20; @B21]) satisfies the above criteria as a fluorescent reporter, though its fluorescence intensity, as manifested by a modest quantum yield (\u03a6), lags behind competing chemistries ultimately making it a less sensitive probe. One of our laboratories has shown that the low quantum yield of MepC can be remedied by substituting the C-6 position with an aromatic group without any penalties on sensitivity or base pairing fidelity ([@B22; @B23; @B24]). Figure 1.The structure of cytidine (C) and unsubstituted pC, with numbering, compared to PhpC. R = ribose or 2\u2032-deoxyribose.\n\nWe have now synthesized the ribonucleoside of 6-phenylpyrrolocytidine (PhpC, [Figure 1](#F1){ref-type=\"fig\"}) and found that, like its DNA and PNA homologs, it ranks among the brightest C-analog luminiophores. The fluorescence of PhpC is sensitive to its microenvironment, and displays a \u03bb~max~ excitation (360 nm) and \u03bb~max~ emission (466 nm) significantly red-shifted from the absorption of other biomolecules. The combination of these properties makes PhpC unique among fluorescent nucleobase analogs and has enabled its use as a fluorescent reporter group in an RNase H assay. Maintaining the same RNA sequence as the traditional FQ-based assay ([@B6]) (RNA-1, [Table 1](#T1){ref-type=\"table\"}), we replaced a single rC nucleoside with PhpC close to the 3\u2032-terminus of the RNA/DNA hybrid, PhpC-1, where it exhibits low fluorescence (quenched state) as depicted in [Scheme 2](#F8){ref-type=\"fig\"}. Treatment with HIV-1 RT RNase H generates an RNA tetranucleotide ([@B25]) bearing PhpC, and quickly dissociates from its DNA complement with a concomitant dramatic increase in fluorescence emission. Table 1.Oligonucleotides used in RNase H assays and *T*~m~ dataNameSequence (5\u2032--3\u2032)[^a^](#TF1){ref-type=\"table-fn\"}*T*~m~ (\u00b0C)[^b^](#TF2){ref-type=\"table-fn\"}DNA-1RNA-2RNA-1GAU CUG AGC CUG GGA GCU65.178.5PhpC-1GAU CUG AGC CUG GGA G\\***C**U64.778.8PhpC-2GAU CUG AGC \\***C**UG GGA GCU67.580.8[^1][^2] Scheme 2.Reporting via a single fluorescent nucleotide analogs. \\*C = 6-phenylpyrrolocytidine (PhpC).\n\nMATERIALS AND METHODS\n=====================\n\n5\u2032-*O*-(4,4\u2032-dimethoxytrityl)-5-iodouridine ([@B2])\n---------------------------------------------------\n\nUridine (2.44 g, 10 mmol) was added to a mixture of iodine (2.54 g, 10 mmol) and silver sulfate (3.12 mg, 10 mmol) in methanol (200 ml) at room temperature (r.t.). The mixture was stirred for 8 min, after which time the yellow solid was removed by filtration and the filtrate was evaporated under reduced pressure. The residue was co-evaporated with 2 : 1 MeOH/H~2~O (50 ml) to remove iodine (which sublimed away) and crystallized from water to give pure 5-iodouridine as a white solid (4.44 g, 60%). 5-Iodouridine (2.22 g, 6 mmol) was dissolved in pyridine (30 ml) and 4,4\u2032-dimethoxytritylchloride (2.64 g, 7.8 mmol) was added in four portions during a 4 h period. After an additional 2 h of stirring at r.t. the reaction was quenched by adding 1 ml of 5% NaHCO~3~ for 15 min. The solution was concentrated under reduced pressure, redissolved in 50 ml of dichloromethane (DCM) and washed with a solution of 5% NaHCO~3~ (30 ml). The product was further extracted with two 25 ml portions of DCM, and the combined organic layer was dried over Mg~2~SO~4~ and evaporated to leave the crude product. The residue was then purified by silica gel column chromatography using an elution gradient with DCM/MeOH/TEA (98 : 1 : 1 to 96 : 3 : 1). Fractions containing **2** were evaporated to give a white foam (3.72 g, 55%). ^1^H NMR (400 MHz, CDCl~3~): \u03b4 = 8.14 (s, 1 H), 7.21--7.42 (m, 11 H), 6.84 (m, 4 H), 6.31 (dd, ^1^*J* = 6.8 Hz, ^2^*J* = 5.9 Hz, 1 H), 4.55 (m, 1 H), 4.07 (m, 1 H), 3.79 (s, 6 H), 3.78 (m, 1 H), 3.36--3.44 (m, 2 H), 2.46--2.64 (m, 1 H), 2.30--2.33 (m, 1 H) p.p.m. ESI-TOF (*m/z*) 695.4 (M + Na).\n\n5\u2032-*O*-(4,4\u2032-dimethoxytrityl)-5-(phenylethynyl)uridine ([@B3])\n--------------------------------------------------------------\n\nIn a 50 ml Schlenk flask, 5\u2032-*O*-(4,4\u2032-dimethoxytrityl)-5-iodouridine (0.67 g, 1 mmol) was mixed with dry THF (15 ml), phenylacetylene (6 mmol), Et~3~N (4 mmol) and was then deoxygenated with N~2~ for 5 min. Subsequently, Pd(PPh~3~)~4~ (0.1 mmol) and CuI (0.2 mmol) was added and the mixture was stirred in the dark for 36 h at r.t. The solution was then filtered, concentrated to dryness, redissolved in CH~2~Cl~2~ (50 ml) and washed sequentially with 5% NaHCO~3~ (30 ml), 5% EDTA (30 ml) and 5% NaHCO~3~ (30 ml) and dried over Na~2~SO~4~. The solvent was evaporated and the resulting residue was purified by silica gel column chromatography using a gradient of hexanes/acetone/TEA (78 : 20 : 2 to 40 : 58 : 2) to provide product **3** as a pure compound in 60% yield (0.39 g). ^1^H NMR (400 MHz, CDCl~3~) \u03b4 = 8.15 (s, 1H, 6-H); 7.05--7.42 (m, 14H, ArH), 6.74--6.78 (m, 4H, ArH), 5.80 (d, ^3^*J* = 4.0 Hz, 1H, H-1\u2032), 4.44, 4.34, 4.30 (3m, 3H each, H-2\u2032, H-3\u2032, H-4\u2032), 3.66 and 3.65 (s, 6H, 2 OMe), 3.44(d, 1H, ^2^*J* = 12.0 Hz, ^3^*J* = 4.0 Hz, CH~A~H~B~-5\u2032), 3.29 (d, 1H, ^2^*J* = 12.0 Hz, CH~A~H~B~-5\u2032). ^13^C NMR(100 MHz, CDCl~3~) \u03b4 = 164.89, 158.71, 152.42, 144.68, 142.24, 135.92, 135.74, 131.84, 130.24, 130.12, 128.21, 128.09, 127.12, 122,99, 113.52, 100.68, 93.56, 90.62, 87.09, 84.99, 81.27, 76.12, 71.28, 71.28, 63.36, 55.33. HRMS (EI) *m/z* calcd\\[Na\\]: 669.2913; found: 669.2924.\n\n5\u2032-*O*-(4,4\u2032-dimethoxytrityl)-6-phenylfuranouridine ([@B4])\n-----------------------------------------------------------\n\nCompound **3** (0.645 g, 1 mmol) was dissolved in deoxygenated acetone (10 ml), then AgNO~3~ (0.1 mmol) was added. The reaction mixture was stirred at r.t. in the dark for 24 h and monitored by TLC (5--15% MeOH/DCM) until reaction completion. After filtration and concentration of the solution in vacuum, the solid residue was purified by silica gel column chromatography using an elution gradient of hexanes/EtOAc/TEA (50 : 50 : 2 to 15 : 85 : 2) to provide **4** (0.450 g, 70%). ^1^H NMR (400 MHz, CDCl~3~) \u03b4 8.90 (s, 1H, 4-H); 7.24--7.64 (m, 14H, ArH), 6.81--6.84 (m, 4H, ArH), 6.50 (d, 1H, ^3^*J* = 2.0 Hz, H-1\u2032), 5.77 (s, 1H, 5-H), 5.56 (m, 1H, H-2\u2032), 4.48 (m, 1H, H-3\u2032), 34.39 (m, H, H-4\u2032), 3.72, 3.74 (2s, 6H, 2 OMe), 3.55 (m, 2H, H-5\u2033). ^13^C NMR (100 MHz, CDCl~3~): \u03b4 171.38 (C-7a); 158.60(C-2), 158.55(C-6), 155.04 (C~arom~), 143.91 (C~arom~), 137.35 (C-4) 135.61, 135.18 (2 C~arom~), 130.16, 129.99 (2 CH~arom~), 129.35(C~arom~), 128.70, 128.34, 128.20, 128.08 (4 CH~arom~), 126.98 (C~arom~), 124.57, 113.32 (2 CH~arom~), 108.40 (C-5), 98.10 (C-4a), 93.23 (C-1\u2032), 87.00 (*C*--Ph(Ph--OCH~3~)~2~) 83.77 (C-4\u2032), 75.67 (C-2\u2032), 68.72 (C-3\u2032), 61.14 (\u2212*C*H~2~O), 60,29 (C-5\u2032), 55.45 (2*C*H~3~OAr). ESI-TOF (*m/z*) 669.2 (M + Na).\n\n5\u2032-*O*-(4,4\u2032-dimethoxytrityl)-6-PhpC ([@B5])\n--------------------------------------------\n\nA pressure bottle with compound **4** (0.64 g, 1.0 mmol), concentrated NH~3~OH (4 ml) and MeOH (6 ml) was stirred at 55\u00b0C for 24 h. Progress of the reaction was followed by TLC (*R*~f~ 0.38 in 10 : 1 DCM/MeOH). After completion of the reaction (24 h), the yellow solid was filtered and then dissolved in CH~2~Cl~2~ and dried over Na~2~SO~4~ and the solvent was removed and the residue was further dried under vacuum overnight (0.57 g, 60%). ^1^H NMR (400 MHz, CDCl~3~) \u03b4 = 11.57 (s, 1H), 8.83 (s, 1H), 7.57 (d, 2H), 7.37 (d, *J* = 7.5, 3H), 7.33--7.06 (m, 13H), 7.03 (*t*, *J* = 7.55, 1H), 6.75 (dd, *J* = 8.8, 6.9, 4H), 6.13 (s, 1H), 5.52 (s, 1H), 4.48 (s, 1H), 4.46 (d, *J* = 5.8, 1H), 4.32 (d, *J* = 5.8, 1H), 3.69, 3.58 (m, 2H), 3.63 (d, *J* = 10.2, 6H). ^13^C NMR (101 MHz, CDCl~3~) \u03b4 = 171.40, 159.27, 158.88, 158.82, 155.38, 144.40, 140.89, 136.54, 136.20, 135.71, 130.48, 130.30, 130.26, 128.90, 128.62, 128.38, 127.25, 125.50, 113.65, 113.43, 111.17, 97.22, 93.44, 87.22, 83.90, 76.23, 69.36, 61.73, 60.64, 55.37, 55.34. HRMS (EI) *m/z* calcd for C~38~H~36~N~3~O~7~: 645.2475; found \\[M + H\\]: 646.2585.\n\n5\u2032-*O*-(4,4\u2032-Dimethoxytrityl)-2\u2032-*O*-*tert*-butyldimethylsilyl-6-PhpC ([@B6])\n-----------------------------------------------------------------------------\n\nTo a 25 ml round-bottom flask containing 315 mg (0.48 mmol) of compound **5** dissolved in 2 ml of dry DMF, TBDMS-Cl (88 mg, 0.58 mmol) and imidazole (82 mg, 1.2 mmol) was added and stirred overnight. The formation of the 2\u2032-*O*-silyl and 3\u2032-*O*-silyl products were monitored by TLC (*R*~f~ values of 0.56 and 0.13, respectively in 1 : 3 EtOAc/DCM) until the starting material was consumed. The reaction was worked-up in 5% NaHCO~3~, filtered over MgSO~4~ and evaporated to dryness. The 2\u2032-*O*-silyl regioisomer was obtained by silica gel column chromatography using a gradient of hexanes-ethyl acetate from 1 : 3 to 1 : 1 keeping 1% TEA in the solvents. The remaining crude 2\u2032- and 3\u2032-regioisomer were mixed for 2 h in 5 ml of pyridine and three drops of water to form an equal mix of isomers. From this mix, the 2\u2032-regioisomer was again purified by silica gel and both purified fractions were evaporated over high vacuum to give 213 mg of a yellow powder (58% yield). ^1^H NMR (500 MHz, DMSO) \u03b4 = 11.76 (s, 1H), 8.75 (s, 1H), 7.61 (d, *J* = 7.6, 2H), 7.47--7.10 (m, 16H), 6.92 (d, *J* = 4.1, 4H), 5.78 (s, 1H), 5.41 (s, 1H), 5.20 (d, *J* = 6.1, 1H), 4.37 (s, 1H), 4.15 (s, 1H), 4.11 (d, *J* = 8.3, 1H), 3.70 (d, *J* = 6.6, 7H), 3.45 (dd, *J* = 9.9, 58.4, 2H), 0.89 (s, 9H), 0.15 (s, 3H), 0.10 (s, 3H) p.p.m. ^13^C NMR (126 MHz, DMSO) \u03b4 = 160.38, 158.89, 158.84, 154.53, 144.68, 139.67, 136.79, 136.67, 136.35, 135.72, 130.82, 130.70, 130.54, 130.31, 129.70, 129.53, 128.74, 128.58, 127.53, 125.43, 125.31, 114.10, 114.01, 109.63, 97.04, 92.40, 92.28, 86.88, 82.06, 81.95, 77.58, 77.46, 68.37, 61.76, 55.76, 55.62, 26.45, 26.41, 18.66, \u22124.12, \u22124.20 p.p.m. ESI-TOF (*m/z*) 782.4 (M + Na).\n\n5\u2032-*O*-(4,4\u2032-Dimethoxytrityl)-2\u2032-*O*-*tert*-butyldimethylsilyl-3\u2032-*O*-(2-cyanoethyldiisopropyl-phosphoramidite)-6-PhpC ([@B7])\n------------------------------------------------------------------------------------------------------------------------------\n\nCompound 6 (213 mg, 0.28 mmol) was sublimed in 3 ml of distilled benzene over dry ice on high vaccum overnight, purged with argon, and dissolved in freshly distilled THF. Dry diisopropylethylamine (0.2 ml, 1.12 mmol) and diisopropylamino-(2-cyanoethyl)-phosphoramidic chloride (63 \u00b5l, 0.28 mmol) were added and the reaction was stirred for 2 h and monitored by TLC (5% methanol in DCM). At the end of this time, the reaction was washed with 5% NaHCO~3~, dried over sodium sulfate and loaded on a silica gel column neutralized in 2.5% triethylamine in hexane. The purified phosphoramidite diastereomers were eluted in 6 : 4 haxane-ethyl acetate, rotary evaporated and sublimed in 3 ml of distilled benzene overnight to yield a yellow foam (240 mg, 0.25 mmol). ^31^P NMR (200 MHz, ACN) \u03b4 = 151.3 (s, 1P), 149.5 (s, 1P) p.p.m. ESI-TOF (*m/z*) 982 (M + Na).\n\nOligonucleotide synthesis\n-------------------------\n\nSolid-phase synthesis of oligonucleotides was carried out on an Applied Biosystems 3400 DNA synthesizer using standard protocols ([@B26]). Coupling yields for PhpC were comparable to standard 2\u2032-*O*-TBDMS ribonucleonucleoside phosphoramidites based on analytical denaturing PAGE (SI). Cleavage from the solid support was carried out in a 3 : 1 mixture of NH~4~OH:EtOH at r.t for 48 h, and removal of the 2\u2032-*O*-silyl protecting groups was achieved by treatment with distilled triethylammonium trihydrofluoride for 48 h. The crude, deprotected oligonucleotides were precipitated in ice-cold butanol and 3M sodium acetate and quantitated by their UV absorbance at 260 nm on a Cary-300 UV-VIS spectrophotometer (Varian Inc.). Crude products were then analyzed and purified by denaturing PAGE (7 M urea) and visualized by UV shadowing using 254 nm light for non-fluorescent oligonucleotides, and 365 nm light for fluorescent oligonucleotides. Full-length products were excised from the gels using a sterile surgical blade, and eluted in sterile water. The eluted products were desalted by size-exclusion chromatography on G-25 Sephadex (GE Healthcare) and quantitated.\n\nThermal denaturation and circular dichroism studies\n---------------------------------------------------\n\nFor UV--Vis thermal denaturation experiments oligonucleotides were analyzed in 1 ml solutions at 1 \u00b5M concentrations in a buffer of 10 mM sodium phosphate at pH 7.0 and 50 mM NaCl. Double-stranded oligonucleotides were annealed in equimolar amounts by heating to 95\u00b0C and then slowly cooling to r.t. on a heating block. Thermal denaturation experiments were performed on a Cary-300 UV--Vis spectrophotometer (Varian Inc.) equipped with a 6 \u00d7 6 cell changer and Peltier temperature controller. Samples were heated from 10\u00b0C to 95\u00b0C at a rate of 0.5\u00b0C/min and the change in absorbance was measured every 1.0\u00b0C. The melting temperature (*T*~m~) values were obtained by the baseline (alpha) method, and defined as the point when the mole fraction of duplex was equal to 0.5. These values represent the averages of at least three independent experiments. Fluorescent thermal denaturation plots were obtained in a similar fashion on a Cary Eclipse fluorescent spectrophotometer equipped with a multicell Peltier temperature controller and automated polarization accessories. Measurements were carried out in 1 cm \u00d7 1 cm quartz cells in 2 ml volumes and 1 \u00b5M concentration of oligonucleotides\n\nCircular dichroism (CD) spectra were obtained using a Jasco J-800 spectropolarimeter. Samples were prepared in the same fashion as with *T*~m~ experiments. Scans were performed in triplicate at 20\u00b0C at a rate of 50 nm/min from 350 to 190 nm. The data were averaged, corrected against a blank and smoothed using the Spectra Manager CD software provided by the manufacturer.\n\nRNase H assays\n--------------\n\nWild-type p66/p51 HIV-1 RT was a generous gift from Dr M. G\u00f6tte (McGill), prepared by his group as described earlier ([@B2]). RNA substrates were 5\u2032-radiolabeled with \u03b3-^32^P ATP by T4 polynucleotide kinase (Fermentas) using the manufacturer's recommended procedure. The DNA and 5\u2032-^32^P labeled RNA strands were combined in a 1.2 : 1 ratio and annealed by heating to 95\u00b0C followed by slow cooling to r.t. HIV-1 RT (2.5 nM final concentration) was incubated for 10 min at 37\u00b0C in RNase H reaction buffer (50 mM Tris--HCl, pH 7.8, 60 mM KCl, 5 mM MgCl~2~, 0.1 mM DTT and 0.01% Tween-20). The reactions were initiated by the addition of duplexed RNA/DNA substrate to a concentration of 50 nM. Aliquots were removed at various times as indicated in [Figure 2](#F2){ref-type=\"fig\"} and stopped by the addition of an equal volume of loading buffer (98% deionized formamide, 1 mg/ml bromophenol blue and 1 mg/ml xylene cyanol) followed by heat inactivation at 95\u00b0C for 5 min. Cleavage products were resolved on a 16% denaturing polyacrylamide gel and visualized by autoradiography. Figure 2.5\u2032-^32^P-labeled PAGE assay for HIV-1 RT RNAase H activity. Arrowheads indicate substrate (**s**) and major cleavage product (**c**).\n\nFluorescent RNase H assays of HIV-1 RT RNase H activity was monitored by the changes in emission of PhpC on a Cary Eclipse fluorescent spectrophotometer equipped with a multicell Peltier temperature controller and automated polarization accessories. These assays were run in 2 ml volumes in identical buffer conditions as the gel-based assays, but were initiated by the addition of MgCl~2~ (5 mM final concentration) in order to monitor fluorescence prior to cleavage by HIV-1 RT. Substrate concentrations varied from 1 to 0.2 \u00b5M and enzyme concentrations were changed to 25 or 5 nM, respectively. The fluorescence intensity was monitored every 15 s. PhpC RNase H assays were monitored at an excitation wavelength of 360 nm (5 nm bandwidth) and emission wavelength of 465 nm (5 nm bandwidth). The fluorescein-RNA/dabcyl-DNA pair was monitored at an excitation wavelength of 485 nm (5 nm bandwidth) and emission wavelength of 520 nm (5 nm bandwidth). Fluorescence polarization (FP) RNase H assays were run under identical conditions as stated above.\n\nFluorescent RNase H assays on 96-well microplates\n-------------------------------------------------\n\nRNase H activity of HIV-1 RT was monitored on a Gemini XS and M5 spectrofluorometers by Molecular Devices in 96-half well opaque plates. FP measurements in 96-well microplates were performed on a BioTek Synergy 4. Experiments were run under identical buffer conditions as gel-based assays in a final volume of 100 \u00b5l. Reactions were initiated by the addition of a 50 \u00b5l solution containing HIV-1 RT and MgCl~2~. The concentration of substrate was varied from 1 \u00b5M to 10 nM, and enzyme concentrations varied from 25 to 5 nM, respectively. Excitation and emission wavelengths for PhpC were 360 and 465 nm, respectively, and 485 and 520 nm for the dabcyl/fluorescein assay. To determine *K*~m~, *V*~max~ and *k*~cat~, the initial velocity of RNase H cleavage was measured under a fixed concentration HIV-1 RT (1 nM) and the substrate concentration was varied from 10 to 900 nM. The IC~50~ of DHBNH, a sample of which was generously provided by Dr M.A. Parniak, was determined on two separate days and was run in triplicate using 200 nM substrate and 1.5 nM HIV-1 RT in 1% DMSO. Eight concentrations of DHBNH were dispensed by two-fold serial dilutions starting from 50 \u00b5M of the inhibitor.\n\nRESULTS AND DISCUSSION\n======================\n\nPhysical properties of RNA containing PhpC\n------------------------------------------\n\nPhpC-containing oligoribonucleotides were obtained by conventional phosphoramidite chemistry in a similar manner as deoxy-PhpC ([@B22; @B23; @B24]) with the 2\u2032-OH group protected as the *tert-*butyldimethylsilyl ether ([@B27],[@B28]) ([Scheme 3](#F9){ref-type=\"fig\"}). Thermal denaturation studies of these sequences hybridized to complementary DNA (DNA-1) and RNA (RNA-2) were carried out ([Table 1](#T1){ref-type=\"table\"}). PhpC-1 showed comparable binding affinity to unmodified controls, whereas the PhpC at a central position, PhpC-2, displayed a 2\u00b0C increase in thermal stability that is ascribed to greater \u03c0-stacking, which is consistent with the increased hypochromicity observed in the thermal denaturation profiles (SI), and previous reports of PhpC in PNA and DNA ([@B22],[@B24]). The PhpC insert was also shown to be exceptionally well-tolerated in RNA/RNA duplexes by CD spectroscopy (SI), as the CD spectra overlapped nearly identically with the native strands in shape and intensity. It is noteworthy that despite the changes in global conformation that accompany the transition for B-form (dsDNA) through hybrid duplexes (RNA:DNA) to A-form (dsRNA) that the modified *C5*-face of cytosine is well accommodated based on *T*~m~ and CD measurements. Although there is a scarcity of data reporting on conformation for *C5*-modified pyrimidines in dsRNA or hybrid RNA:DNA duplexes, *T*~m~ data indicate that modest modifications such as halogens ([@B29]) and alkyl/alkynyl ([@B30]) may be modestly stabilizing, as we have observed for PhpC. Scheme 3.Synthesis of 6-Phenylpyrrolocytidine phosphoramidite ([@B7]). a. I~2~, AgSO~4~, MeOH. b. DMT-Cl, py. c. phenylacetylene, DMF, Pd(PPh~3~)~4~, CuI, NEt~3~. d. AgNO~3~, acetone. e. NH~4~OH, MeOH, 55\u00b0C. f. TBDMS-Cl, imidazole, DMF. g. (*i*Pr)~2~NP(Cl)O(CH~2~)~2~CN, THF, N*i*Pr~2~Et.\n\nFluorescence properties of RNA containing PhpC\n----------------------------------------------\n\nThe fluorescence parameters of PhpC as a free nucleoside, and when incorporated in short oligomers, single-stranded 18-mers and duplexed to DNA or RNA are given in [Table 2](#T2){ref-type=\"table\"}. The PhpC nucleoside ([@B1]) and PhpC nucleoside monophosphate ([@B2]) displayed similar fluorescence intensity and quantum yields of 0.31 and 0.29, respectively, a 12-fold increase from a reported 6-methylpyrrolocytidine ([@B31]) value (0.023). Thus, it appears that the phosphate group does not play a role on the fluorescence intensity. Comparing the trinucleotides, G-PhpC-U ([@B3]), C-PhpC-U ([@B4]), U-PhpC-U ([@B5]) and ([@B6]) A-PhpC-U the fluorescence intensity and quantum yield is dependent on the nature of neighboring bases, with **3** (neighboring guanine) showing a reduced quantum yield (\u03a6) as was reported earlier for MepC ([@B31],[@B32]). Theoretical studies on MepC suggested that base-stacking interactions can cause diminution of \u03a6 by affecting the strength of the oscillator for the fluorescence transition ([@B31]), which is consistent with our observations, although quenching of the fluorescence by intrastrand electron transfer cannot be discounted. The U--PhpC--U trinucleotide had the highest observed quantum yield of the series (\u03a6 = 0.41), which is likely due to shielding of the fluorophore from solvent (dynamical quenching) while not introducing other compensating non radiative deactivation pathways. Interestingly, the 18-mer single stranded PhpC-1 had the same fluorescence intensity and quantum yield as the trinucleotide **3**, where PhpC had the same nearest neighbors. This trend did not continue with PhpC-2 (\u03a6 = 0.13) and its analogous trinucleotide **4** (\u03a6 = 0.24). In this case, it appears that PhpC is quenched in the longer single-stranded RNA sequence by 50%, which may be due to better stacking interactions in a central position; however, length and base composition may play a greater role for shorter PhpC containing oligonucleotides that are conformationally more flexible. Table 2.Fluorescent properties of PhpC and oligonucleotides containing single PhpC insertsEntryNameFluorescence intensity[^a^](#TF3){ref-type=\"table-fn\"}Quantum yield[^b^](#TF4){ref-type=\"table-fn\"} \u03a6FP[^c^](#TF5){ref-type=\"table-fn\"}1PhpC510.31\\<0.0225\u2032-PO~4~-PhpC470.29\\<0.023G-PhpC-U360.110.034C-PhpC-U1040.240.035U-PhpC-U1340.410.036A-PhpC-U106n.d.0.037PhpC-1350.110.218PhpC-1:DNA-17.80.0350.289PhpC-1:RNA-29.5n.d.0.3010PhpC-2380.130.2211PhpC-2:DNA-1800.200.2712PhpC-2:RNA-269n.d.0.28[^3][^4][^5][^6]\n\nOligonucleotides containing pC analogs have consistently shown fluorescence quenching upon duplex formation ([@B22; @B23; @B24],[@B33; @B34; @B35]). The fluorescence of PhpC-1 hybridized to DNA ([@B8]) and RNA ([@B9]) also showed \u223c75% quenching. An increase in fluorescence as the duplexes denature could be monitored by thermal denaturation profiles, and the mid-point of the transition corresponds to the *T*~m~ determined by UV absorbance (SI). The fluorescence emission when PhpC-2 bound to DNA ([@B11]) and RNA ([@B12]) complementary strands surprisingly increased. This is unexpected based on our previous work with modified oligonucleotides, and to our knowledge is the first example of a pC analog showing increased fluorescence emission with hybridization. These observations are consistent with Thompson's recent study that proposed that the changes in quantum yield of pC are not only dominated by base stacking, which is responsible for much of the changes in fluorescence for 2-AP ([@B31]), but possibly by collisional deactivation from solvent. We observed a transition in the fluorescent thermal denaturation curve of single-stranded PhpC-1 (SI) and PhpC-2 (not shown) indicating that some quenching in single strands is contributed by base stacking. The dramatic quenching of fluorescence normally observed during duplex formation and specific hydrogen bonding to guanine may be due to base pairing mediated electron transfer, which leads to non-radiative relaxation of the electronically excited state. Currently, this explanation awaits experimental support. It is clear from these results and the limited studies on MepC, that there are a variety of factors that affect the fluorescence of PhpC including the sequence context (neighboring bases) and hybridization state along with microsequence effects.\n\nRNase H assays\n--------------\n\nThe PhpC-containing RNA were determined to be excellent substrates for HIV-1 RT RNase H by standard 5\u2032-^32^P-label/PAGE assay ([Figure 2](#F2){ref-type=\"fig\"}), unlike the FQ system, which compromises enzymatic activity (SI). Furthermore, no aberrant cleavage products were generated as was previously observed with deoxy-MepC inserts on the polypurine tract of HIV-1 RT ([@B36]).\n\nSubsequently, the RNase H activity was monitored fluorimetrically. Duplex **8** proved to be remarkably responsive and gave the largest fluorescence change (14-fold) between the hybrid duplex and the cleavage product ([Figure 3](#F3){ref-type=\"fig\"}). This was easily followed in kinetic or end-point mode. The increase in fluorescence was so dramatic that it is visible to the naked eye ([Figure 3](#F3){ref-type=\"fig\"}, inset). We compared the RNase H cleavage of PhpC-1, to the classic RNA-3\u2032-fluorescein/DNA-5\u2032-dabcyl (FQ) assay ([@B6]) ([Figure 4](#F4){ref-type=\"fig\"}) by running them simultaneously at 1 \u00b5M substrate concentrations. We found that the FQ assay has higher fluorescence signal due to the greater brightness of fluorescein, but the PhpC-1 assay generated cleavage products with greater relative fluorescence compared with the intact substrates (14- versus 8-fold). As observed in gel-based assays, the PhpC insert did not slow cleavage as did the FQ substrate resulting in a substantial gain in analysis time. The end point of cleavage (90% completion) was achieved in \u223c20 min for 8 versus 130 min for FQ. Figure 3.Fluorescence emission spectra of PhpC nucleotide (1, black), PhpC-1 single stranded (2, red), PhpC-1:DNA-1 duplex (3, blue) and PhpC-1:DNA-1 after RNase H cleavage by HIV-1 RT (4, green). Visual changes in fluorescence of the same solutions under UV~365~ light, at 1 \u00b5M concentration (inset). Figure 4.Comparison the change in fluorescence emission intensity for the PhpC-1 RNase H assay (red curve) compared to the classic Fluorescein/Dabcyl assay (black curve). Both substrates were treated with HIV-1 RT RNase H activity simultaneously in identical conditions at a concentration of 1 \u00b5[M]{.smallcaps} scanning at the \u03bb~max~ excitation and emission for both fluorescein (485/520 nm) and PhpC (360/465 nm).\n\nWhile a labeled DNA strand was found to be fluorimetrically responsive to the cleavage of its unlabeled RNA complement (data not shown), an advantage of incorporation of the fluorescent nucleoside into the RNA strand is that it gets processed by RNase H into fragments that are much smaller than the hybrid duplex. The substantial change in mass/size of the fluorescent moiety may be detected by changes in FP ([@B33]). As seen in [Table 1](#T1){ref-type=\"table\"}, FP values increase with increasing molecular weight as expected for molecules undergoing less Brownian motion. RNase H mediated cleavage of PhpC-1:DNA was readily monitored in real-time by FP ([Figure 5](#F5){ref-type=\"fig\"}) in a similar fashion to the RNase H assay developed by Pfizer ([@B7]). Figure 5.Monitoring the RNase H activity of HIV-1 RT on PhpC-1:DNA-1 by fluorescence polarization (black curve). The inset shows the fluorescence intensity with the excitation and emission filters parallel (blue curve) and perpendicular (green curve).\n\nFor application of this assay to screen for inhibitors of RNase H activity, it is advantageous to monitor polarization as well as fluorescence intensity. FP can discern molecules that quench the fluorophore producing false hits ([@B7]). With the versatility of PhpC we have demonstrated that both fluorescence intensity (which reports on hybridization/base stacking) and FP (which reports on the size of the molecule containing the fluorophore) can be monitored simultaneously using only one probe.\n\nRNase H assays on 96-well plates\n--------------------------------\n\nAlthough fluorescent spectrophotometers are far more sensitive, it is desirable to adapt assays to fluorescent microplate readers that are amenable to HTS. Since duplex **8** was the most fluorescent and responsive substrate for RNase H activity, it was tested in 96-well plate spectrofluorometers by fluorescence intensity ([Figure 6](#F6){ref-type=\"fig\"}) and FP (SI). Using identical reaction conditions used in the cuvettes (1 \u00b5M substrate and 25 nM HIV-1 RT), [Figure 6](#F6){ref-type=\"fig\"} shows very good signal to noise and response to RNase H cleaveage (10-fold increase). We later tested the limits of sensitivity of the assay and found we could still monitor RNase H activity down to 10 nM substrate concentration (60 nM in FP mode), which is comparable to ^32^P PAGE assays Figure 6.Fluorescence HIV-1 RT RNase H assay of 1 \u00b5[M]{.smallcaps} PhpC-1:DNA-1 monitored by a 96-well plate spectrofluorimeter. Colored triangles represent reactions in three individual wells, with the red curve representing their average. The average of three controls without enzyme is represented by the black curve.\n\nWe determined the kinetic parameters *K*~m~ (54 \u00b13.3 nM), *V*~max~ (0.31 nmols/min) and a *k*~cat~ of 6/min of duplex **8** for HIV-1 RT RNase H (SI). Since this assay will ultimately be used to screen potential inhibitors of RNase H, a known inhibitor of HIV-1 RT RNase H, namely DHBNH, was tested in this assay. Our single-label assay demonstrated that DHBNH was able to inhibit RNase H activity in a dose-dependent manner with an IC~50~ of 5 \u00b5M, which is consistent with previous results ([@B37]).\n\nCONCLUSIONS\n===========\n\nThis work demonstrates the significant advantages of base-modified nucleosides, such as PhpC, compared with traditional fluorophores: one single PhpC insert can act as a sensitive reporter group that is non-disruptive to the structure and enzymatic activity. Although fluorescein is a substantially brighter luminophore, the PhpC-based assay for RNase H offers several advantages. The responsiveness, rapidity and ease (single label versus dual) of the RNase H assay has been improved. The fluorescence provided by PhpC is sufficient to compete with gel-based techniques on the basis of sensitivity and the assay can be adapted to multiwell plate format for HTS. This was also possible because the HIV-1 RT RNase H cleavage product, the tetranucleotide, showed remarkably greater fluorescence than even the free nucleoside. Also uncovered in this study, was the surprising and unprecedented observation that oligomer PhpC-2 showed an increase in fluorescence intensity upon duplex formation. Together, these observations indicate that PhpC is significantly different from MepC and warrants further investigation on the affects of sequence length and composition on changes in fluorescence. Overall, this report lays the framework for a sensitive and rapid assay for RNase H activity and inhibitors thereof. We feel that PhpC will contribute to the growing repertoire of useful fluorescent nucleobase analogs such as tC\u00b0, which also shows red-shifted fluorescence and a high quantum yield (\u03a6 = 0.3) but whose fluorescence is insensitive to duplex formation ([@B38]).\n\nSUPPLEMENTARY DATA\n==================\n\n[Supplementary Data](http://nar.oxfordjournals.org/cgi/content/full/gkp1022/DC1) are available at NAR Online.\n\nFUNDING\n=======\n\nThe Natural Sciences and Engineering Research Council of Canada is acknowledged for funding for this work through the discovery grants program (RHEH, MJD).\n\n*Conflict of interest statement*. None declared.\n\nSupplementary Material\n======================\n\n###### \\[Supplementary Data\\]\n\nWe are grateful to the Natural Sciences and Engineering Research Council of Canada for financial support in the form of grants (RHEH and MJD). We thank Dr Matthias G\u00f6tte for providing the HIV-1 RT and Dr Michael A. Parniak for the RNase H inhibitor.\n\n[^1]: ^a^PhpC indicated by \\*C.\n\n[^2]: ^b^*T*~m~ values represent the average of at least 3 independent experiments within 1\u00b0C.\n\n[^3]: ^a^One micro molar samples measured in 10 m[M]{.smallcaps} phosphate buffer (pH 7.0) and 50 m[M]{.smallcaps} NaCl at 25\u00b0C, \u03bb~ex~ = 360 nm; \u03bb~emm~ = 465 nm\n\n[^4]: ^b^See Supporting Information for Quantum Yield determination.\n\n[^5]: ^c^FP measured in identical conditions as footnote a.\n\n[^6]: n.d. signifies not determined.\n"}
+{"text": "Introduction {#S0001}\n============\n\nOrofacial pain is a prevalent and debilitating pain that may be particularly distressing due to the special psychological and emotional context of this body area.[@CIT0001] A formalin injection into the orofacial region is considered to be the only animal model of persistent pain in the trigeminal region.[@CIT0002] It has been demonstrated that carbachol as a non-selective cholinergic receptor agonist activates 20% of orexin neurons through the M3 muscarinic receptors and subsequent activation of non-selective cation channels.[@CIT0003] Carbachol also inhibits a very small population of orexin neurons (1%).[@CIT0003] Orexin neurons are localized exclusively in the lateral hypothalamus (LH) and project to almost all of the brain regions except the cerebellum.[@CIT0004],[@CIT0005] The rodent ventral tegmental area (VTA) contains both orexin receptors, orexin-1 (OX1) and orexin-2 (OX2) receptors [@CIT0006],[@CIT0007] with particularly high levels of OX2 receptor.[@CIT0007] Orexin potentially activates catecholaminergic neurons in the brainstem, including the VTA dopaminergic (DA) neurons and locus coeruleus noradrenergic neurons.[@CIT0006],[@CIT0008] Approximately 55% of the neurons located in VTA are dopaminergic neurons.[@CIT0009] Both DA neurons and GABAergic neurons of the VTA receive massive afferents from orexinergic neurons.[@CIT0010],[@CIT0011] In this regard, electrophysiological evidence indicated that orexin activates both DA and non-DA cells of the VTA via direct postsynaptic mechanisms.[@CIT0006] The involvement of VTA in the rewarding process has received much attention, and certain studies have demonstrated that the neural circuitry of reward also processes the motivational-affective dimension of pain and responds to aversive stimuli, such as noxious stimuli.[@CIT0012],[@CIT0013] This is because the rewarding system shares large neural networks with the pain processing system.[@CIT0014],[@CIT0015] Dopamine neurons located in the VTA region project to the prefrontal cortex, nucleus accumbens (NAc), basolateral amygdala, hippocampus, and other regions that participate in the modulation of pain 16--19.\n\nDopaminergic neurotransmission is mediated by five receptor subtypes (D1-D5) in the central nervous system (CNS). Among the five subtypes, the D1 and D5 dopamine receptors are classified as members of the D1-like family, while the D2-like dopamine receptor subfamily consists of D2, D3 and D4 dopamine receptors. These dopamine receptor subtypes have different affinity to various ligands.[@CIT0020] In our previous studies, we demonstrated that D1- and D2-like dopamine receptors within VTA and NAc are involved in the reduction of analgesia induced by chemical stimulation of LH in the tail-flick test as a model of acute pain.[@CIT0021],[@CIT0022] Pain assays in rodent studies revealed that the DA system has a more potent role in tonic pain, such as the formalin or writhing test as apposed to brief phasic pain stimuli, such as tail-flick or hot plate tests.[@CIT0023] These findings suggest that the neural systems involved in the mediation of tonic pain are not exactly the same as those that mediate phasic pain. Accordingly, in the current study, we induced orofacial pain in rats and assessed whether D1- and D2-like dopamine receptors within the VTA participate in antinociception are induced by LH chemical stimulation. In other words, the aim of this study was to suggest the LH-VTA neural circuitry in which LH participates in orofacial pain modulation.\n\nMaterials and Methods {#S0002}\n=====================\n\nExperimental Design {#S0002-S2001}\n-------------------\n\nIn this study, rats were assigned to two separate control groups including normal saline and formalin control groups ([Table 1](#T0001){ref-type=\"table\"}). Among these, formalin group received 50 \u00b5L formalin into upper lip and normal saline-control group received 50 \u00b5L saline into upper lip, then returned to the plexiglass box for assessing animal's pain-related behaviors. To evaluate the effect of chemical stimulation of LH on pain modulation, three different solutions of carbachol (62.5, 125 and 250 nM; CAR/Control) or it^'^s vehicle (saline; SAL/Control) were microinjected into the LH of rats and subjected to orofacial formalin test. To evaluate the effect of D1- or D2-like dopamine receptor antagonist on antinociception induced by carbachol injections, SCH-23390 (0.25, 1 or 4 \u00b5g/0.3 \u00b5L saline in each group; SCH/CAR) or Sulpiride (0.25, 1 or 4 \u00b5g/0.3 \u00b5L DMSO 12% in each group; SUL/CAR) were microinjected, respectively, in the VTA, 5 min prior intra-LH microinjection of carbachol (250 nM). Saline is solvent of SCH-23390 and DMSO is solvent of Sulpiride. We used these vehicles for evaluation of the effect of injection volume and to compare their effects with respective drugs. Rats were then subjected to orofacial formalin test. Additionally, to investigate the effect of D1 or D2-like dopamine receptor antagonist alone on the nociceptive behaviors, animals received the maximum dose of SCH-23390 (4 \u00b5g/0.3 \u00b5L saline; SCH/Saline) or Sulpiride (4 \u00b5g/0.3 \u00b5L DMSO 12%; SUL/DMSO) respectively, into the VTA 5 min prior microinjection of intra-LH saline (0.3 \u00b5L) and subjected to orofacial formalin test ([Table 1](#T0001){ref-type=\"table\"}). In this study, the experimenter was blind to treatment condition and protocol.Table 1A Table Depicting All Control and Experimental Groups in the StudyGroup NamesMicroinjection into the Lateral Hypothalamus (0.5 \u03bcL/rat)Microinjection into the Ventral Tegmental Area (0.3 \u03bcL/rat)SurgeryIntactSubcutaneous injection of 50 \u03bcL formalin 1% into the upper lip\\-\\--Sham operated\\--Surgery and recovery periods (5 to 7 days)Saline-control\\\n(SAL/Control)Saline-Carbachol (CAR)Carbachol (250 nM)-Vehicles\\\n(VEHs)SalineSaline or DMSO 12%Carbachol-control groups\\\n(CAR/Control)Carbachol (250 nM)Saline or DMSO 12%SCH23390-treated groups\\\n(SCH/CAR)\\\nand (SCH/Saline)Carbachol (250 nM)SCH23390 (0.5 \u00b5g)Carbachol (250 nM)SCH23390 (1 \u00b5g)Carbachol (250 nM)SCH23390 (4 \u00b5g)SalineSCH23390 (4 \u00b5g)Sulpiride-treated groups\\\n(SUL/CAR)\\\nand (SUL/DMSO)Carbachol (250 nM)Sulpiride (0.5 \u00b5g)Carbachol (250 nM)Sulpiride (1 \u00b5g)Carbachol (250 nM)Sulpiride (4 \u00b5g)DMSO 12%Sulpiride (4 \u00b5g)\n\nAnimals {#S0002-S2002}\n-------\n\nIn this study, the ninety-one adult male albino Wistar rats weighing 230--250 g were purchased from Pasteur Institute (Tehran, Iran). The animals were randomly chosen and assigned in different experimental groups. Animals were kept in controlled temperature (22 \u00b1 2\u00b0C), humidity (47%), and 12-hour light/dark cycle with free access to food and water during the whole period of experiments. All experiments were carried out by following the guidelines outlined in National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 80--23, revised 1996) and were confirmed by the Research and Ethics Committee of Shahid Beheshti University of Medical Sciences, Tehran, Iran (IR.SBMU.PHNS.REC.1397.031).\n\nSurgical Preparation {#S0002-S2003}\n--------------------\n\nAnimals were anesthetized with an intraperitoneal injection of xylazine 2% (10 mg/kg) and ketamine 10% (100 mg/kg) mixture and were placed in a stereotaxic apparatus (Stoelting, USA). Two stainless steel guide cannulae (23-gauge) were unilaterally inserted 1 mm above the LH and VTA. According to the Paxinos and Watson rat brain atlas,[@CIT0024] the coordinate for the LH was AP = 2.65 \u00b1 0.15 mm caudal to the bregma, Lat = \u00b11.3 mm lateral to the midline and DV = 8.6 mm ventral from the skull surface, and for the VTA: AP = 4.8 \u00b1 0.15 mm caudal to the bregma, Lat = \u00b1 0.8 mm lateral to the midline and DV = 8.3 mm ventral from the skull surface. Guide cannulae were then anchored to the skull surface using two stainless steel screws and dental acrylic cement. After surgery and before the initiation of behavioral experiments, animals were allowed to recover for 5--7 days.\n\nDrugs and Drug Administration {#S0002-S2004}\n-----------------------------\n\nCarbachol as a non-selective cholinergic receptor agonist (Carbamylcholine chloride; Sigma-Aldrich, St. Louis, Missouri, USA) was dissolved in sterile normal saline. D1-like dopamine receptor antagonist, SCH-23390, (R)-(+)-7-Chloro-8-hydroxy-3-methyl-1-phenyl-2, 3, 4, 5-tetrahydro-1H-3-benzazepine hydrochloride (Tocris Bioscience, Bristol, UK) was dissolved in sterile normal saline. D2-like dopamine receptor antagonist, Sulpiride, (S)-5-aminosulfonyl-N- \\[(1- ethyl-2 pyrrolidinyl) methyl\\]-2-methoxybenzamide (Tocris Bioscience, Bristol, UK) was dissolved in dimethyl sulfoxide (DMSO 12%), Sigma-Aldrich, Germany. Formalin 2.5% was prepared by diluting 37% formaldehyde (Merck, Germany) with sterile physiological saline (0.9%). All drugs or vehicle solutions were freshly prepared on the day of the experiment and infused slowly over 60 s. All microinjections were conducted in freely moving animals via a stainless-steel injector (30-gauge needle; 1 mm longer than guide cannula) which was connected to a 1-\u00b5L Hamilton syringe via a polyethylene tube (PE-20).\n\nOrofacial Formalin Test {#S0002-S2005}\n-----------------------\n\nThe orofacial formalin test was conducted using a plexiglass box (30 \u00d7 30 \u00d730 cm) with a mirror angled at 45\u00b0 below the surface of the box to observe the nociceptive behaviors of rats.[@CIT0002] Before to test initiation, an acclimation period of about 30 min was considered to allow animals to be stabilized in a new environment. Animals were subcutaneously injected with 50\u03bcL of 1% formalin into the upper lip just lateral to the nose using a 29-gauge injection needle. Formalin was injected into the upper lip, ipsilateral to the cannula placement into the rats' brain. Rats then immediately returned to the box and the amount of time that rats spent face rubbing with the ipsilateral paw was measured and considered as the index of nociception. In our experiments, we divided the recording time into 15 blocks of 3 min for a time-course analysis. Subcutaneous formalin injection induces biphasic nociceptive responses. Face rubbing activity measured between 0 and 3 min after subcutaneous formalin injection, considered as the \ufb01rst (early or acute) phase and those were measured between 15 and 33 min after formalin injection considered as the second (late or chronic) phase.\n\nHistological Verification {#S0002-S2006}\n-------------------------\n\nImmediately after completion of the experiments, animals were deeply anesthetized with an intraperitoneal injection of ketamine and xylazine and the following perfusion with 0.9% normal saline and 10% formaldehyde solution, rats were then sacrificed, and their brains were carefully removed. Transverse brain sections with 50-\u03bcm thickness were prepared and the location of the guide cannula tips was identified using the Paxinos and Watson rat brain atlas.[@CIT0024] Only the animals with correct cannulae placements were chosen for final data analysis ([Figure 1](#F0001){ref-type=\"fig\"}).Figure 1Coronal brain sections show the microinjection sites in the (left panel) lateral hypothalamus (\u25cbSaline; \u25cfCarbachol; \u25b2Misplacement), and (right panel) ventral tegmental area (\u25cbSaline; \u25a1DMSO; \u25cfSCH-23390; \u25a0Sulpiride;\u25b2Misplacement).**Abbreviations:** D3V, dorsal 3^rd^ ventricle; LV, lateral ventricle; CPu, caudate putamen (striatum); ic, internal capsule; cc, corpus callosum; DA, dorsal hypothalamic area; CPu, caudate putamen (striatum); D3V, dorsal 3^rd^ ventricle; DA, dorsal hypothalamic area; mt, mammillothalamic tract; PeF, perifornical nucleus; PeFLH, perifornical part of lateral hypothalamus; MTu, medial tuberal nucleus; f, fornix; VMH, ventromedial hypothalamic nucleus; pc, posterior commissure; 3V, 3^rd^ ventricle; mL, medial lemniscus; Rad, radiatum layer of the hippocampus; fr, fasciculus retroflexus; SNR, substantia nigra, reticular part; str, superior thalamic radiation; VTA, ventral tegmental area; ML, medial mammillary nucleus, lateral part; PBP, parabrachial pigmented nucleus of the VTA; SuM, supramammillary nucleus; MCLH, magnocellular nucleus of the lateral hypothalamus; scale bar = 1 mm.\n\nData Analysis {#S0002-S2007}\n-------------\n\nThe obtained pain scores were expressed as mean\u00b1 SEM (standard error of mean). Data analyses were performed by commercially available software GraphPad Prism^\u00ae^ 6.0. The repeated measures two-way analysis of variance (ANOVA) followed by Bonferroni's post-*hoc* test was used to assess the effects of time and treatment (formalin injection) on the nociceptive scores (3-min blocks during a 45-min period; [Figure 2A](#F0002){ref-type=\"fig\"}). To evaluate the carbachol dose-response, we used one-way ANOVA followed by Dunnett's comparison tests and three different doses of carbachol were compared with saline as a vehicle ([Figure 2B](#F0002){ref-type=\"fig\"}). It is also noted to evaluate the face rubbing time during both early and late phases separately, one-way ANOVA followed by Newman-Keuls multiple comparison tests were conducted to compare all groups with each other ([Figure 3A](#F0003){ref-type=\"fig\"} and [Figure 4A](#F0004){ref-type=\"fig\"}). Furthermore, to estimate the effective dose 50% (ED50), the best fitted line to represent the data on scatter plot was drawn and the estimated ED50 values were mathematically calculated based on three different doses of SCH-23390 or Sulpiride (0.25, 1, 4 nM). In this study, the effect size was also calculated separately for early and late phases by dividing mean difference between the experimental and control groups by the standard deviation of the population from which the different treatment groups were taken. P \\< 0.05 was considered statistically significant.Figure 2(**A**) The time course of face rubbing as the nociceptive responses immediately after subcutaneous injection of 1% formalin or normal saline into the orofacial region. Following formalin injection into the upper lip a significant increase in face rubbing time spent compared to respective 3-min block in normal saline group was observed. (**B**) The effect of microinjection of different solutions of carbachol into the LH on formalin-induced orofacial nociception. Intra-LH microinjection of carbachol (62.5, 125 and 250 nM/rat) attenuated face rubbing time spent (sec) during both phases of formalin orofacial nociception in a dose-dependent manner. Each point represents the mean \u00b1 SEM for 6--8 rats in each group. \\*P \\< 0.05, \\*\\*P \\< 0.01 and \\*\\*\\*P \\< 0.001 compared to previous 3-min block. ^\u2020^P \\< 0.05, ^\u2020\u2020^P \\< 0.01 and ^\u2020\u2020\u2020^P \\< 0.001 compared to respective 3-min block in normal saline group. \\*\\*P \\< 0.01 and \\*\\*\\*P \\< 0.001 compared to vehicle (Saline) group.Figure 3(**A**) Effect of intra-VTA administration of SCH23390 (D1-like dopamine receptor antagonist) on the antinociception induced by chemical stimulation of LH using carbachol. Intra-VTA administration of SCH23390 (0.25, 1 and 4 \u03bcg/0.3 \u03bcL saline) significantly attenuated antinociception induced by intra-LH microinjection of carbachol (250 nM/rat) during both early and late phases of formalin-induced orofacial nociception. (**B**) A log dose--response curve of the effect of intra-VTA administration of different solutions of SCH-23390 (0.25, 1and 4 \u00b5g/0.3 \u03bcL saline) on carbachol-induced antinociception during the early compared to that of late phase of formalin-induced orofacial nociception. The effective dose (ED50) of SCH-23390 in the late phase (0.33 \u00b5g) was saliently less than that in the early phase (1.78 \u00b5g). Each point represents the mean \u00b1 SEM for 7--8 rats in each group. \\*P \\< 0.05, \\*\\*P \\< 0.01 and \\*\\*\\*P \\< 0.001 compared to saline-carbachol group. ^+^P \\< 0.05 and ^++^P \\< 0.01 compared to another group.**Abbreviation:** ns, non-significant.Figure 4(**A**) Effect of intra-VTA injection of Sulpiride (D2-like dopamine receptor antagonist) on the LH stimulation-induced antinociception during orofacial formalin nociception. Intra-VTA administration of different doses of Sulpiride (0.25, 1 and 4 \u03bcg/0.3 \u03bcL DMSO 12%) dose-dependently attenuated antinociception produced by intra-LH microinjection of carbachol (250 nM/rat) during both early and late phases of formalin-induced orofacial nociception. (**B**) A log dose--response curve of the effect of intra-VTA administration of different solutions of Sulpiride (0.25, 1and 4 \u00b5g/0.3 \u03bcL DMSO 12%) on antinociception produced by carbachol during the early compared to that of late phase of formalin-induced orofacial nociception. The effective dose (ED50) of Sulpiride in the late phase (0.67 \u00b5g) was obviously less than that in the early phase (1.93 \u00b5g). Each point represents the mean \u00b1 SEM for 7--8 rats in each group. \\* P \\< 0.05, \\*\\*P \\< 0.01 and \\*\\*\\*P \\< 0.001 compared to DMSO-carbachol group ^+^P \\< 0.05 and ^+++^P \\< 0.001 compared to another group.**Abbreviation:** ns, non-significant.\n\nResults {#S0003}\n=======\n\nEffect of Chemical Stimulation of LH Using Carbachol Microinjection on the Formalin-Induced Orofacial Nociception {#S0003-S2001}\n-----------------------------------------------------------------------------------------------------------------\n\nThe time course of face rubbing as the nociceptive responses following injection of formalin or normal saline (50 \u00b5L) into the orofacial region is presented in [Figure 2A](#F0002){ref-type=\"fig\"}. Repeated measures two-way analysis of variance (ANOVA) followed by Bonferroni's post-*hoc* test was used to assess the effects of time and treatment (formalin injection) on the nociceptive scores (3-min blocks during a 45-min period treatment effect: F (1120) =657.4, P\\<0.0001; time effect: F (14,120) =17,65 P\\<0.0001; treatment and time interaction: F (14, 120) =16.3, P\\<0.0001\\].\n\nFollowing injection of formalin into the upper lip, a biphasic pattern was observed while there were no significant differences in face rubbing time spent among several consecutive time points in normal saline-treated animals. As it has been shown in [Figure 2B](#F0002){ref-type=\"fig\"}, one-way ANOVA followed by Dunnett's multiple comparison demonstrated that intra-LH carbachol microinjection (125 or 250 nM) dose-dependently decreased the face rubbing time during the first \\[F (3,29) = 7.752, P = 0.0007; left panel\\] and second \\[F (3,29) = 41.48, P \\< 0.0001; right panel\\] phases of orofacial formalin test. The low concentration of carbachol (62.5 nM) had no significant effect on face rubbing time during both phases of formalin nociception.\n\nEffects of Intra-VTA Administration of D1-Like Dopamine Receptor Antagonist, SCH-23390 on the Antinociception Induced by Intra-LH Microinjection of Carbachol {#S0003-S2002}\n-------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nOur results showed that intra-VTA administration of D1-like dopamine receptor antagonist, SCH-23390, could significantly increase face rubbing activity during early \\[F (3, 30) = 5.195, P=0.0058; [Figure 3A](#F0003){ref-type=\"fig\"}, left panel\\] and late F (3, 30) = 42.35, P\\<0.0001; [Figure 3A](#F0003){ref-type=\"fig\"}, right panel\\] phases of orofacial formalin nociception compared to saline-carbachol group. As demonstrated in [Figure 3A](#F0003){ref-type=\"fig\"}, both 1 \u00b5g and 4 \u00b5g solutions of SCH-23390 were able to decrease antinociceptive effects of carbachol (250 nM) in both phases of orofacial formalin test. In addition, unpaired Student's *t*-test demonstrated that, 4 \u00b5g solution of SCH-23390 could completely block the antinociception of carbachol, so that there was no difference between SCH/CAR and SCH/Saline groups in both early \\[t (13) = 0.4259, p= 0.6771\\] and late \\[t (13) = 0.4259, p= 0.677, p= 0.4363\\] phases of the orofacial formalin test. The lowest concentration of SCH-23390 (0.25 \u00b5g) had no significant effect on face rubbing time compared to the saline-carbachol group during both early and late phases of formalin-induced orofacial nociception. Moreover, Eta-squared as an estimated measure of effect size during the early phase (\u03b72 = 0.37) was smaller than the late phase (\u03b72 = 0.82) which represents the more prominent role of SCH-23390 in the elimination of intra-LH carbachol-induced antinociception in late phase in comparison with the early phase. Moreover, as it has been depicted in [Figure 3B](#F0003){ref-type=\"fig\"}, comparison of the effective dose of SCH-23390 in the elimination of intra-LH carbachol-induced antinociception between the early and late phases of orofacial formalin test showed that the antinociceptive effect of carbachol was reversed by a lower dose of SCH-23390 in the late phase (ED50=0.33) than the early phase (ED50=1.78) of orofacial formalin test.\n\nEffects of Intra-VTA Administration of D2-Like Dopamine Receptor Antagonist, Sulpiride on the Antinociception Induced by Intra-LH Microinjection of Carbachol {#S0003-S2003}\n-------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAccording to our results, intra-VTA administration of D2-like dopamine receptor antagonist, Sulpiride could significantly decrease carbachol-induced antinociception during early \\[F (3, 31) = 7.614, P=0.0007; [Figure 4A](#F0004){ref-type=\"fig\"}, left panel\\] and late \\[F (3, 31) = 31.57, P \\<0.0001; [Figure 4A](#F0004){ref-type=\"fig\"}, right panel\\] phases of orofacial formalin nociception compared to saline-carbachol group. As it has been depicted in [Figure 4A](#F0004){ref-type=\"fig\"}, the lowest concentration of Sulpiride (0.25 \u00b5g) had no significant effect on face rubbing time compared to the saline-carbachol group while both 1 \u00b5g and 4 \u00b5g solutions of Sulpiride dose-dependently decreased antinociceptive effects of carbachol (250 nM) in both phases of orofacial formalin nociception. Additionally, unpaired Student's *t*-test demonstrated that 4 \u00b5g solution of Sulpiride could completely block the antinociception induced by intra-LH carbachol microinjection, so that there was no difference between SUL/CAR and SUL/Saline groups in both early \\[t (13) = 0.1335, p= 0.8959\\] and late \\[t (13) = 0.8032, p= 0.4363\\] phases of orofacial formalin test. The effect size of Sulpiride during the early phase (\u03b72 = 0.45) was smaller than the late phase (0.77) indicating a more prominent role of Sulpiride in elimination of intra-LH carbachol-induced antinociception in late phase compared with the early phase. [Figure 4B](#F0004){ref-type=\"fig\"} depicts the log dose--response curve of changes in face rubbing time in Sulpiride injected groups compared to CAR/Control group. The log dose-response curve demonstrated that the antinociceptive effect of carbachol was reversed by a lower dose of Sulpiride in the late phase (ED50=0.67) than the early phase (ED50=1.93) of orofacial formalin test.\n\nDiscussion {#S0004}\n==========\n\nThe results of the current study illustrated that D1- and D2-like dopamine receptors within the VTA are involved in antinociceptive effects induced by chemical stimulation of LH during both early and late phases of orofacial formalin test. The major findings of this study were as follows: (i) Chemical stimulation of LH using carbachol dose-dependently attenuated biphasic orofacial pain induced by formalin injections in the orofacial region of rats. (ii) Intra-VTA administration of D1-like dopamine receptor antagonist (SCH-23390) dose-dependently, decreased analgesia produced by intra-LH carbachol injections during both early and late phases of orofacial formalin test. (iii) Intra-VTA injection of D2-like dopamine receptor antagonist (Sulpiride), dose-dependently, attenuated analgesia produced by intra-LH injection of carbachol during both phases of orofacial formalin nociception. (iv) The roles of D1- and D2-like dopamine receptors of VTA in intra-LH carbachol-induced antinociception were noticeably more prominent in the late phase of orofacial formalin nociception in comparison with the early phase.\n\nIn the current study and other similar studies, it has been shown that chemical stimulation of LH using carbachol attenuates formalin nociception in a dose-dependent manner.[@CIT0025]--[@CIT0029] Orexin neurons in the LH receive massive cholinergic input. So carbachol as a potent muscarinic cholinergic receptor agonist is able to increase the activity of these neurons.[@CIT0003] It has been established that neuroanatomical connections of OX neurons with some spinal and supra-spinal areas, that are known to be part of descending pain modulatory circuitry, are involved in pain modulation.[@CIT0030] According to the result of the current study, VTA is involved in LH stimulation-induced antinociception. Although VTA plays a key role in the reward system,[@CIT0031] It has been confirmed that the reward system shares certain significant neural networks with the pain processing system.[@CIT0014],[@CIT0032] Accordingly, Morgan and Franklin, in 1990 reported that, lesions of VTA neurons abolished morphine-induced analgesia in an animal model of tonic pain.[@CIT0033] In this regard, it has been demonstrated that electrical stimulation of VTA after the nociceptive stimulus facilitates the analgesic process.[@CIT0034]\n\nWe previously demonstrated that OX1and OX2 receptor functions in the VTA are essential for reward processing in the conditioned place preference model of rats.[@CIT0035] The rodent VTA contains both OX1 and OX2 receptors.[@CIT0006],[@CIT0036] Our results showed that the antinociception produced by LH stimulation is mediated by D1- and D2-like dopamine receptors of VTA. It has been reported that OX-A activates phospholipase C- and protein kinase C-mediated Ca^2+^signaling in DA neurons of the VTA and this effect may serve as the mechanism by which OX exerts its behavioral and psychostimulant effects.[@CIT0008] Neuroanatomical and electrophysiological evidences indicating the presence of OX1 and OX2 receptors in DA and non-DA neurons in VTA which demonstrates that both cell types of the VTA are potential targets for orexinergic actions. Both DA neurons and non-DA neurons, such as, GABAergic neurons of the VTA receive massive afferents from orexinergic neurons and are excited by these neurons. Electrophysiological evidence indicated that orexin activates both DA and non-DA cells of the VTA via direct postsynaptic mechanisms.[@CIT0006] Dopamine neurons of the VTA region project to the prefrontal cortex, NAc, basolateral amygdala, hippocampus, and other regions that participate in the modulation of pain.[@CIT0016]--[@CIT0019] It has been proven that LH orexin projections to VTA increase firing in NAc projecting dopamine neurons.[@CIT0037] Dopaminergic VTA projections to NAc have been largely associated with the pain modulatory role of LH's orexinergic neurons.[@CIT0016],[@CIT0038] On the other hand, the results of a study demonstrated thatchronic increase of GABAergic neuronal activity in VTA region suppresses VTA dopaminergic neuronal activity which is responsible for negative affective aspects of neuropathic pain.[@CIT0039] Therefore, it seems that the both of DA cells and non-DA cells of VTA are responsible for the pain modulatory role of LH, but finding the exact contribution of these neurons requires more extensive studies. In this respect, the results of our recent study demonstrated that orexin receptors within the NAc participate in pain modulatory role of the LH in both early and late phases of orofacial formalin test.[@CIT0040] In addition, the results of an animal study demonstrated that the administration of dopamine receptor antagonists into the NAc led to attenuation of antinociception produced by intra-VTA injection of morphine or Substance P.[@CIT0023] Therefore, it seems that antinociceptive effect of VTA dopaminergic system in part is associated with VTA projecting neurons to the NAc. Lapirot et al in 2011 documented that D2-like dopamine receptors are predominantly located within superficial medullary dorsal horn adjacent to trigeminal nociceptive fibers and activating these receptors inhibits both formalin and capsaicin-evoked pain behavior. Activation of D2-like dopamine receptors in this area also inhibits C-fiber-evoked action potential firing of trigeminal wide dynamic range (WDR) neuron.[@CIT0041] In the current study, we demonstrated that D1- and D2-like dopamine receptors within the VTA are involved in the pain modulatory role of LH stimulation in a model of tonic orofacial pain. Additionally, we previously reported in two separate studies that D1- and D2-like dopamine receptors within VTA and NAc are potentially involved in LH-induced analgesia in the tail-flick test as a model of acute phasic pain.[@CIT0021],[@CIT0022] It has been found that dopamine is differentially involved in the mediation of antinociception in phasic and tonic pain conditions. Pain assays in rodent studies revealed that the DA system has a more potent role in tonic pain, such as that shown in the formalin or writhing test than brief phasic pain stimuli, shown in the tail flick or hot plate tests.[@CIT0023] However, the precise mechanisms served by the DA system to alleviate both phasic and tonic pain should be more extensively evaluated.\n\nA typical behavioral response to subcutaneous formalin injection into the orofacial region is biphasic pain response, a short-lasting early phase and a prolonged late phase.[@CIT0042],[@CIT0043] In the present study, the role of D1- and D2-like dopamine receptors of VTA in intra-LH carbachol-induced antinociception was noticeably more prominent in the late phase of orofacial formalin nociception in comparison with the early phase. This discrepancy might stem from the different mechanisms involved in early and late phases of formalin nociception.[@CIT0044] Formalin injection into orofacial receptive field evoked activity in A\u03b4 and C fibers as well as spinal and trigeminal nociceptive neurons and provokes biphasic orofacial pain and concentration-dependent tissue damage.[@CIT0002] The first phase of formalin nociception is attributed to the activation of primary afferent neurons and induced by the activation of the C-fibers. This is while the second phase seems to be dependent on the inflammatory reaction in the peripheral tissue and combination of functional changes in the dorsal horn which leads to peripheral and central sensitization of primary and secondary trigeminal nociceptive neurons.[@CIT0002],[@CIT0043],[@CIT0045]\n\nIn this regard, it has been reported that the action of some analgesics is different in the early and the late phase of formalin nociception.[@CIT0046],[@CIT0047]\n\nIn this study, we tried to focus on the role of D1- and D2-like dopamine receptors within the VTA in antinociception induced by LH stimulation using carbachol. However, this study has some limitations that should take into consideration for interpreting the data. First of all, the excitation of OX neurons in the LH by carbachol could be confirmed using c-fos immunoreactivity method. Secondly, anterograde tracing is also a valuable research method, which could be used to trace the dopaminergic axonal projections to the VTA. Thirdly, the contribution of each D1-/D2-like dopamine receptors of VTA alone, by antinociceptive effects of LH should be evaluated. Taken together, we suggest that chemical stimulation of the LH by using carbachol activates the VTA dopaminergic neurons possibly via activation of LH orexin projecting neurons, and so, it participates as a neural circuitry that modulates tonic orofacial pain. The contribution of this neural circuitry to orofacial pain perception and modulation will help us to understand tonic pain pathophysiology. Moreover, understanding the critical role of certain brain regions such as LH or VTA in pain modulation as well as, the mechanisms in which these areas participate in analgesia offer an alternative approach to the development of more efficient pain therapies with focusing on these brain areas.\n\nThe authors would like to thank the Dental Research Center, School of Dentistry, Shahid Beheshti University of Medical Sciences, and also the Neuroscience Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran, for their cooperation in carrying out this study.\n\nDisclosure {#S0005}\n==========\n\nThe authors report no conflicts of interest in this work.\n"}
+{"text": "Peaks and valleys in protein translation prepare one set of immune cells for duty, according to findings from Lelouard et al. With the ebb and flow of translation, antigen-presenting dendritic cells (DCs) adjust the origin of their antigens.\n\nFigure 1Overall translation levels (black bars) peak several hours after DCs are activated and then decline.\n\nDCs are professional immunity activators that alert T cells to the presence of invaders by displaying antigens on their surface. This display is ramped up by a maturation program initiated when a DC encounters inflammatory stimuli such as pathogenic components. Scientists have identified an abundance of transcriptional changes that take place during maturation. The new results show that a boost in translation is necessary to put the new transcripts into action.\n\nTranslation peaked in DCs about 4 hours after their activation. During this stretch, translation---and the PI3K/AKT/mTOR signaling pathway that activated it---was needed for maturation-associated changes in the DCs. These changes include producing T cell-activating cytokines and activating the antigen-presenting machinery.\n\nAfter 4 to 8 hours, protein synthesis levels declined, due at least in part to proteasome-mediated cleavage of the eIF4GI translation initiation factor. At 16 hours, overall translation levels were even lower than they were before activation.\n\nSome transcripts, however, escaped the translation shutdown. As also occurs during stress conditions, eIF4GI cleavage initiated an unusual translation pathway that bypasses the need for a 5\u2032 cap on transcripts. Cap-independent translation, which favors the synthesis of antiapoptotic proteins, made mature DCs resistant to apoptosis-inducing drugs. This pathway might thus help activated DCs survive the stress of their heavy new transcription and translation loads while they search for and then activate T cells.\n\nAs DCs switched translation pathways, they also changed the source of antigens presented on MHC class I molecules. During the early, translation-heavy stage, antigens were derived from pieces of newly synthesized proteins; when the authors blocked translation, antigen presentation was limited. But later on, translation inhibitors did not interfere with MHC class I presentation.\n\nThe authors speculate that once DCs have matured, their antigens are mainly derived from exogenous sources, such as pathogens. Although normally presented on MHC class II molecules, exogenous peptides can also crossover to the class I pathway. An alternative possibility is that late antigens come from a pool of stored, presynthesized self-peptides.\n\nReference:\n\nLelouard, H., et al. 2007. J. Cell Biol. 179:1427--1439. 18166652\n\n[^1]: <lebrasn@rockefeller.edu>\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\n*Clostridium difficile* is a Gram-positive, spore-forming anaerobic bacterium. It is responsible for primary and recurrent antibiotic-associated diarrhea and pseudomembranous colitis \\[[@B1]\\]. The recent emergence of hypervirulent strains has caused a rapid increase in the incidence of*C. difficile*infection (CDI) worldwide. Two exotoxins,*Clostridium difficile* toxin B (TcdB) and toxin A (TcdA), are key virulent factors of CDI \\[[@B2], [@B3]\\]. TcdA is highly cytotoxic, inducing the damage of intestinal epithelial cells and the release of inflammatory cytokines as well as trigging inflammatory and immune response \\[[@B4]--[@B6]\\]. Previous studies showed that TcdA could activate ERK2 and p38 MAP kinases in human monocytic cells and induce apoptotic cell death through ER stress \\[[@B7]\\].\n\nHigh Mobility Group Box1 (HMGB1), the first identified member of the HMGB family, highly conserved in evolution, is described originally as a nuclear DNA-binding protein \\[[@B8]--[@B10]\\]. HMGB1 was identified as an important extracellular mediator of inflammation \\[[@B11], [@B12]\\]. Within the nucleus, HMGB1 maintains chromosomal structure and regulates DNA damage responses \\[[@B13]\\]. However, under a variety of stressful situations, HMGB1 is translocated to the cytosol, and is released into the extracellular coordinating inflammation, immunity, and other local cellular processes \\[[@B14]\\]. The recent discovery of extracellular HMGB1 as a proinflammatory mediator by TcdA-induced acute inflammation and intestinal damage already has been reported in our laboratory previously \\[[@B15]\\].\n\nThe endoplasmic reticulum (ER) that plays an essential role in multiple cellular processes encompasses about half the total membrane area and one-third of the newly translated proteins in a typical eukaryotic cell \\[[@B16], [@B17]\\], and it is an organelle that plays an essential role in multiple cellular processes. Live cells start a homeostatic signaling network named unfolded protein response (UPR), involving three stress transducer proteins, namely, protein kinase RNA-like ER kinase (PERK), inositol-requiring protein 1*\u03b1* (IRE1*\u03b1*), and activating transcription factor 6 (ATF6), each of which activates its own unique cascade of downstream events to regulate metabolism and survival of cells \\[[@B18]\\].\n\nThe apoptotic properties of TcdA have been confirmed in different kinds of cell lines \\[[@B5], [@B6]\\]. And TcdA-induced ER stress reactions also have been clearly described in cells \\[[@B19], [@B20]\\]. In the paper, the potential role of HMGB1 involved in TcdA-induced ER stress was identified which could reveal its underlying mechanism making a contribution to the development of CDI therapies.\n\n2. Materials and Methods {#sec2}\n========================\n\n2.1. Cell Culture and TcdA Purification {#sec2.1}\n---------------------------------------\n\nCT26 (the murine colon adenocarcinoma cell line) was obtained from American Type Culture Collection (Manassas, USA) and cultured in Dulbecco\\'s Modified Eagle\\'s medium (GIBCO, Carlsbad, USA) containing 10% fetal bovine serum (GIBCO), 100\u2009U/mL penicillin, 100\u2009*\u03bc*g/mL streptomycin, 2\u2009mM L-glutamine, and 1\u2009mM sodium pyruvate (GIBCO).\n\nThe full-length wild-type recombinant TcdA plasmid was kindly provided by Dr. Feng (University of Maryland at Baltimore, Baltimore, USA). TcdA was expressed and purified according to the protocol reported by Sullivan et al. \\[[@B21]\\]. The highly purified recombinant TcdA appeared as a single band on sodium dodecyl sulfate polyacrylamide gels (data not shown). rHMGB1 was purchased from Uscn Life Science Inc. (Wuhan, China).\n\n2.2. Cell Rounding Assay {#sec2.2}\n------------------------\n\nCT26 cells were seeded in 96-well plates (1 \u00d7 10^4^ cells/well) and then were treated with TcdA (10\u2009ng/mL) for 4\u2009h or pretreated with 100\u2009*\u03bc*M glycyrrhizin (30\u2009min) before TcdA exposure. Cell rounding was visualized by phase-contrast microscopy. Each concentration was tested in triplicate for overall cell rounding, and the experiments were repeated three times.\n\n2.3. Western Blot Analysis {#sec2.3}\n--------------------------\n\nTotal protein extracts were prepared and separated on 12% SDS-PAGE and electrophoretically transferred to nitrocellulose filter membranes. The membranes were blocked with 5% bovine serum albumin. And then the membranes were incubated overnight at 4\u00b0C with the indicated primary antibodies. The dilutions of antibodies were prepared as follows: HMGB1 (CST number 3935, 1\u2009:\u20091000), BSA (Santa Cruz sc-50528, 1\u2009:\u20091000), PERK (CST number 3192, 1\u2009:\u20091000), CHOP (CST number 2895, 1\u2009:\u20091000), IRE1*\u03b1* (CST number 3294, 1\u2009:\u20091000), ATF6 (Abcam ab62576, 1\u2009:\u20092000), Bcl~2~ (Santa Cruse number KO112, 1\u2009:\u2009500), and *\u03b2*-actin (CST number 5125, 1\u2009:\u20091000). After washing, membranes were treated with horseradish peroxidase-conjugated secondary antibodies (1\u2009:\u20092000; CST, number 7074, USA) and then the bands were visualized using the enhanced chemiluminescence kit (Thermo, number 34080, USA).\n\n2.4. Statistical Analysis {#sec2.4}\n-------------------------\n\nUnless indicated, experiments were repeated at least three times. Data were expressed as the mean \u00b1 SEM. Data were analyzed using Prism v.5.03 (GraphPad Software, San Diego, USA). Statistical significance was assessed by one-way analysis of variance (ANOVA) followed by Turkey\\'s test or two-way repeated-measures ANOVA. *P* \\< 0.05 was considered as statistically significant.\n\n3. Results {#sec3}\n==========\n\n3.1. TcdA Exposure Induces HMGB1 Release from CT26 Cells {#sec3.1}\n--------------------------------------------------------\n\nThe effect of TcdA on CT26 cells was examined by cell rounding assay as evidenced by morphological changes and survival inhibition of cells.\n\nCT26 cells treated with TcdA were examined, which showed that TcdA induced cell rounding in a dose dependent manner, with 60% cell rounding observed after exposure to 1\u2009ng/mL TcdA and 100% cell rounding after exposure to 10\u2009ng/mL for 4\u2009h ([Figure 1(a)](#fig1){ref-type=\"fig\"}). And the results showed that, after exposure to 10\u2009ng/mL TcdA, the morphology of CT26 cells changed from fusiform (control) to rounding ([Figure 1(b)](#fig1){ref-type=\"fig\"}).\n\nTo measure HMGB1 secretion in response to TcdA, CT26 cells were cultured in the presence of 10\u2009ng/mL TcdA and the medium was collected at the indicated times. Western blot analysis showed that the release of HMGB1 induced by TcdA in medium was increased in a time-dependent manner after 12\u2009h of exposure ([Figure 1(c)](#fig1){ref-type=\"fig\"}).\n\n3.2. Exogenous rHMGB1 Induces ER Stress {#sec3.2}\n---------------------------------------\n\nTo determine whether HMGB1 is involved in ER stress, rHMGB1 was used to verify the assumption. CT26 cells were incubated with 1\u2009ng/mL rHMGB1 and were collected at different time points (0, 4, 8, 12, 16, and 24\u2009h). IRE1, ATF6, and PERK branches were detected using western blot. As shown in [Figure 2(a)](#fig2){ref-type=\"fig\"}, the expressions of the ATF6 and PERK in cells were markedly elevated in a time-dependent manner at 12\u2009h of rHMGB1 exposure, in contrast to those of the PBS group, and the expression levels of PERK were detected after 4\u2009h and continued increasing until the end of the experiment; the content of ATF6 was enhanced to the maximum value at 12\u2009h and did not recover at 24\u2009h, whereas the protein expression of IRE1*\u03b1* had no change.\n\nFurthermore, to investigate the involvement of HMGB1 in ER stress, glycyrrhizin, the HMGB1 inhibitor \\[[@B22]\\], was added to the CT26 cells to inhibit activity of rHMGB1. After exposure to rHMGB1 and glycyrrhizin for 12\u2009h, cells were collected and proteins were extracted to evaluate the expression level of IRE1*\u03b1*, ATF6, and PERK by western blot. As shown in [Figure 2(b)](#fig2){ref-type=\"fig\"}, as expected, the ATF6 and PERK proteins expressions were decreased in mixture group (the mixture of 20\u2009*\u03bc*M glycyrrhizin and rHMGB1), in contrast to those of the rHMGB1 group. No significant difference was observed between the glycyrrhizin group and the mixture group, which was similar to that of PBS group, whereas the IRE1*\u03b1* protein had not significantly changed in different groups.\n\nAnd the statistical analysis showed that the level of the ATF6 and PERK proteins in rHMGB1 treatment group had significantly increased in contrast to those of PBS (*P* \\< 0.0001). In contrast, there was no significant difference in the expression of IRE1*\u03b1*.\n\n3.3. Glycyrrhizin Alleviates TcdA-Induced Cell Damage {#sec3.3}\n-----------------------------------------------------\n\nHMGB1 is released from cells exposed to TcdA and further induces cell damage. So we pretreated the cells with glycyrrhizin, the HMGB1 inhibitor \\[[@B22]\\], prior to TcdA exposure to inhibit the activity of the subsequently secreted HMGB1 and further observed the relationship between HMGB1 and TcdA-induced cell damage. As shown in [Figure 3(a)](#fig3){ref-type=\"fig\"}, CT26 cells were treated with different concentration of glycyrrhizin, prior to 10\u2009ng/mL TcdA exposure; 100\u2009*\u03bc*M glycyrrhizin could completely inhibit the toxicity of TcdA, whereas glycyrrhizin alone did not induce cell rounding and the cells showed normal morphology. CT26 cells were exposed to 10\u2009ng/mL TcdA for different times; 50% cell rounding appeared after exposure for 2\u2009h and 100% cell rounding appeared after exposure for 4\u2009h, whereas there was no cell rounding after 4\u2009h and only 50% cell rounding occurred at 24\u2009h when cells were pretreated with 100\u2009*\u03bc*M glycyrrhizin before exposure to TcdA ([Figure 3(b)](#fig3){ref-type=\"fig\"}).\n\n3.4. Glycyrrhizin Pretreatment Affects TcdA-Induced ER Stress {#sec3.4}\n-------------------------------------------------------------\n\nFurther experiments were carried out to investigate if HMGB1 was involved in endoplasmic reticulum stress induced by*Clostridium difficile* toxin A. Treating the cells with glycyrrhizin prior to TcdA exposure which inhibited the activity of the subsequently secreted HMGB1. CT26 cells were incubated with glycyrrhizin (100\u2009*\u03bc*M) for 30 minutes before TcdA exposure and extracted proteins to determine the level of IRE1*\u03b1*, ATF6, and PERK by western blot.\n\nAs shown in [Figure 4](#fig4){ref-type=\"fig\"}, ER stress markers were expressed after treatment of 10\u2009ng/mL TcdA. The expression level of ATF6 and PERK proteins significantly increased, compared with pretreated glycyrrhizin group. The same result was observed in rHMGB1 group. By contrast, without TcdA exposure, the glycyrrhizin or PBS group\\'s proteins expression significantly was decreased to normal levels, whereas the IRE1*\u03b1* protein was not significantly changed when the CT26 was treated with TcdA or pretreated glycyrrhizin.\n\nThe statistical analysis showed that there is notable difference between the expression levels of the ATF6 and PERK proteins in TcdA treatment group and those of the glycyrrhizin pretreatment group (*P* \\< 0.0001).\n\n3.5. HMGB1 Involved Apoptotic ER Stress {#sec3.5}\n---------------------------------------\n\nIn order to investigate the involvement of HMGB1 in TcdA-induced apoptotic ER stress, two experiments were designed and two key mediators of apoptosis ER stress marker, CHOP and Bcl~2~, were detected.\n\nrHMGB1 was used to verify the involvement of rHMGB1 in apoptotic ER stress. As shown in [Figure 5(a)](#fig5){ref-type=\"fig\"}, after exposure to rHMGB1 or the mixture of rHMGB1+glycyrrhizin for 12\u2009h, proteins were extracted from the collected cells to evaluate the expression level of CHOP and Bcl~2~. CHOP protein expression was obvious upregulation which was treated with rHMGB1. In contrast, CHOP proteins expression significantly decreased in the mixture group. Bcl~2~ protein had significantly decreased in rHMGB1 group, compared with PBS group, and the Bcl~2~ expressions in the mixture group were not significantly different from those in the PBS group.\n\nFurther experiments were carried out to investigate if HMGB1 involved apoptotic ER stress by TcdA. CT26 cells were incubated with glycyrrhizin (100\u2009*\u03bc*M) for 30 minutes before TcdA exposure and extracted proteins to determine the level of CHOP and Bcl~2~. As expected, the CHOP, a key mediator of apoptotic ER stress, was significantly increased after treatment with 10\u2009ng/mL TcdA ([Figure 5(b)](#fig5){ref-type=\"fig\"}), compared with glycyrrhizin pretreatment group. The expression of Bcl~2~ in glycyrrhizin pretreatment group was apparently reduced, compared with that of the control. By contrast, without TcdA exposure, no significant difference of the CHOP and Bcl~2~ expression was observed between the glycyrrhizin and PBS group. Our results revealed that HMGB1 could be involved in apoptotic ER stress induced by TcdA.\n\n4. Discussion {#sec4}\n=============\n\n*Clostridium difficile* toxin A is one of the major virulence factors of*C. difficile*infection (CDI) and has been proved to induce apoptotic cell death through ER stress pathway by previous work \\[[@B19], [@B20]\\]. HMGB1 is stabilized nucleosomal structure and facilitates gene transcription in intracellular \\[[@B8], [@B10]\\]. Moreover, HMGB1 is massively released extracellularly and plays a cytokine-like function \\[[@B12], [@B13]\\]. Early studies have shown that HMGB1 is released from intestinal cell and involved in toxin-induced inflammation \\[[@B15]\\].\n\nThe results of our study demonstrated that HMGB1 is involved in endoplasmic reticulum stress induced by*Clostridium difficile* toxin A. And it proved that HMGB1 was released from the nucleus to culture medium after exposure of TcdA for 12\u2009h. Glycyrrhizin can attenuate the activity of HMGB1. The results showed that glycyrrhizin pretreatment delayed the onset of TcdA-induced cell rounding ([Figure 3(b)](#fig3){ref-type=\"fig\"}). These results imply that HMGB1 is probably involved in the cytotoxic and cytopathic effects of TcdA. Glycyrrhizin could modulate the ATF6 and PERK signaling pathways, leading to upregulation of the ATF6, PERK, and CHOP expressions and suppressing endoplasmic reticulum stress. The inhibitory effect of glycyrrhizin on ER stress is another possible mechanism by which glycyrrhizin prevents HMGB1-related cell damage and ER stress. Overall, it is supposed that late stage of release of HMGB1 or activity of HMGB1 has potential effects on ER stress.\n\nTcdA could induce apoptosis of cells including intestinal epithelial cells \\[[@B4], [@B5]\\]. To investigate the interaction of HMGB1 and TcdA-induced apoptotic ER stress, we successfully demonstrated that rHMGB1-induced ER stress could be able to increase the level of CHOP and decrease the Bcl~2~ protein expression, which is similar to the effect of toxin A ([Figure 5(a)](#fig5){ref-type=\"fig\"}), suggesting that the CHOP and Bcl~2~ proteins are involved in cell death induced by TcdA or rHMGB1. CHOP and Bcl~2~ are the key mediators involved in ER stress induced apoptosis. CHOP mediates cell death primarily through two mechanisms, alteration of the transcription of genes involved in apoptosis and oxidative stress \\[[@B23], [@B24]\\].\n\nIn summary, our data suggest that HMGB1 plays a role in endoplasmic reticulum stress induced by*Clostridium difficile* toxin A. It hints that HMGB1 can possibly be a potential candidate for therapies of CDI, which might represent a new approach in the development of new drugs for CDI.\n\nThe authors would like to thank Dr. Hanping Feng for kindly providing the plasmid encoding recombinant TcdA. This study was supported by International Cooperation Programme of Department of Science and Technology of Guangdong Province (no. 2015A050502016) and Project of Natural Science Foundation of Guangdong Province (no. 2016A030313457).\n\nCompeting Interests\n===================\n\nThe authors declare that they have no conflict of interests with the contents of this paper.\n\nAuthors\\' Contributions\n=======================\n\nJi Liu and Yi Ma contributed equally to this paper.\n\n![TcdA induces the release of HMGB1 from CT26 cells. (a) CT26 cells were treated with different concentrations of TcdA for 4\u2009h, and the rate of cell rounding was calculated. (b) CT26 cells were exposed to the medium (cell control) or the TcdA for 4\u2009h. The percentage of cells affected (cell rounding) was observed under a phase-contrast microscope. (c) CT26 cells were exposed to 10\u2009ng/mL TcdA for the indicated time intervals, and HMGB1 levels in the culture medium were detected by western blot analysis using BSA as a loading control. ^*\u2217\u2217\u2217*^ *P* \\< 0.001.](BMRI2016-4130834.001){#fig1}\n\n![Glycyrrhizin prevents HMGB1-induced ER stress. (a) Protein expressions of IRE1*\u03b1*, ATF6, and PERK in CT26 cells treated with HMGB1 for the indicated time were measured by western blot. (b) rHMGB1 combined with glycyrrhizin was incubated with CT26 cells for 12\u2009h and the protein expressions of IRE1*\u03b1*, ATF6, and PERK were detected by western blot analysis. Actin was used as the loading control. Data represent the mean of three independent experiments. ^*\u2217\u2217\u2217*^ *P* \\< 0.001.](BMRI2016-4130834.002){#fig2}\n\n![Glycyrrhizin prevents TcdA-induced cell rounding. CT26 cells were treated with glycyrrhizin and TcdA, and cell rounding rate was measured. (a) CT26 cells were pretreated with different concentrations of glycyrrhizin, followed by 10\u2009ng/mL TcdA for 4\u2009h, and cell rounding was quantified. (b) CT26 cells were treated with glycyrrhizin 30\u2009min before TcdA exposure, and the rate of cell rounding was determined.](BMRI2016-4130834.003){#fig3}\n\n![HMGB1 involves TcdA-induced ER stress. CT26 cells were pretreated with glycyrrhizin 30\u2009min before TcdA exposure. And then CT26 cells were treated by TcdA for 12\u2009h and the protein expressions of IRE1*\u03b1*, ATF6 and PERK were detected by western bolt analysis. Actin was used as the loading control. Data represent the mean of three independent experiments. ^*\u2217\u2217\u2217*^ *P* \\< 0.001.](BMRI2016-4130834.004){#fig4}\n\n![Glycyrrhizin reduces HMGB1-induced apoptotic ER stress. CT26 cells were pretreated with glycyrrhizin 30\u2009min before HMGB1 exposure. And then CT26 cells were treated by HMGB1 for 12\u2009h and collected for analysis. The proteins levels change of CHOP and Bcl~2~ were examined by western blotting. (a) Glycyrrhizin prevented HMGB1-induced apoptotic ER stress. (b) Glycyrrhizin reduced TcdA-induced apoptotic ER stress. Actin was used as the loading control. Data represent the mean of three independent experiments. ^*\u2217\u2217\u2217*^ *P* \\< 0.001.](BMRI2016-4130834.005){#fig5}\n\n[^1]: Academic Editor: Jiazhang Lian\n"}
+{"text": "Background {#sec001}\n==========\n\nThe Global Hepatitis Health Sector Strategy is aiming for \"elimination of viral hepatitis as a public health threat\" by 2030 \\[[@pntd.0005842.ref001]\\], while enhanced elimination efforts for hepatitis are also promoted under the broader remit of global Sustainable Development Goals (SDGs) \\[[@pntd.0005842.ref002]\\]. This is an enormous challenge for hepatitis B virus (HBV) given the estimated global burden of 260 million chronic carriers, of whom the majority are unaware of their infection \\[[@pntd.0005842.ref003]\\] ([Fig 1](#pntd.0005842.g001){ref-type=\"fig\"}).\n\n![The hepatitis B virus (HBV) cascade.\\\nDiagrammatic representation of the total burden of HBV infection and the subsets of individuals who are diagnosed (orange), linked to care (green), engaged with care (blue), on treatment (light purple), and have suppressed viremia (dark purple). An estimate of the proportion of cases undiagnosed versus diagnosed (91% versus 9%, respectively) is based on the WHO fact sheet \\[[@pntd.0005842.ref003]\\]. The proportion who flow from each pool to the next is otherwise represented by a question mark, as these numbers are not represented by robust data.](pntd.0005842.g001){#pntd.0005842.g001}\n\nWe here present HBV within the framework for neglected tropical diseases (NTDs) \\[[@pntd.0005842.ref004]\\] in order to highlight the ways in which HBV meets NTD criteria and to discuss the ways in which the NTD management paradigm could be used to strengthen a unified global approach to HBV elimination \\[[@pntd.0005842.ref005]\\]. The major burden of morbidity and mortality from HBV is now borne by tropical and subtropical countries \\[[@pntd.0005842.ref006]\\]. Many African populations epitomize specific vulnerability to HBV \\[[@pntd.0005842.ref007]\\], so we here focus particular attention on Africa, both through focus on the existing published literature and through presentation of a unique data set of opinion and experience (see [S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}). However, the themes we represent are transferable to other low- and middle-income settings and are relevant on the global stage.\n\nCurrent strategies for HBV control {#sec002}\n==================================\n\nRobust preventive vaccines have been rolled out in Africa since 1995 as a component of the Expanded Programme on Immunization (EPI). Traditionally, most vaccine campaigns have relied upon monovalent HBV vaccines (for summary, see <https://doi.org/10.6084/m9.figshare.5242303.v1>). For adults with chronic infection and evidence of ongoing liver damage, a daily dose of suppressive antiviral therapy using nucleot(s)ide analogues ([Table 1](#pntd.0005842.t001){ref-type=\"table\"}) is successful at mediating viremic suppression in the majority of cases, reducing complications and diminishing spread. Antiviral therapy does not commonly result in cure, due to the persistence of DNA in the hepatocyte nucleus, in the form of both cccDNA and integrated HBV DNA, but interferon (IFN)-based therapy can increase rates of clearance.\n\n10.1371/journal.pntd.0005842.t001\n\n###### Drug therapy used to treat HBV.\n\nCosting is based on the International Medical Products Price Guide: <http://mshpriceguide.org/en> (data accessed May 2017. Price for lamivudine (3TC)---South Africa Department of Health; Price for tenofovir (TDF)---Supply Chain Management Project; price for HBV immunoglobulin (HBIG)---Sudan Medicins Sans Frontieres). WHO essential medicines: <http://who.int/medicines/publications/essentialmedicines/EML_2015_FINAL_amended_NOV2015.pdf?ua=1>.\n\n![](pntd.0005842.t001){#pntd.0005842.t001g}\n\n  Drug name                                        Drug class                                   Potency against HBV[\\*](#t001fn002){ref-type=\"table-fn\"}   Resistance                                                                         Severe adverse effects                                                           Safe in pregnancy?   Use in children                                                                          Use as part of combined ART?   WHO \"essential medicine\"   Monitoring             Cost (International Medical Products Price Guide)\n  ------------------------------------------------ -------------------------------------------- ---------------------------------------------------------- ---------------------------------------------------------------------------------- -------------------------------------------------------------------------------- -------------------- ---------------------------------------------------------------------------------------- ------------------------------ -------------------------- ---------------------- ---------------------------------------------------\n  **Tenofovir (TDF)**                              Nucleotide reverse transcriptase inhibitor   \\+                                                         Rare                                                                               Lactic acidosis, hepatitis, renal injury, bone demineralization                  Yes                  \\>12 years for HBV[\\*\\*](#t001fn003){ref-type=\"table-fn\"}                                Yes                            Yes                        LFTs, renal function   US\\$3.91/month\n  **Entecavir (ETV)**                              Nucleoside reverse transcriptase inhibitor   ++                                                         \\<10% at 3 years. Increased in 3TC resistance                                      Lactic acidosis, steatosis                                                       Not known            From age 2 years                                                                         No                             Yes                        LFTs, FBC              Not listed\n  **Lamivudine (3TC)**                             Nucleoside reverse transcriptase inhibitor   \\+ (potentially limited by resistance)                     50% at 3 years. Best-recognized mutations are in YMDD motif in viral polymerase.   Lactic acidosis, hepatomegaly and steatosis, pancreatitis                        Yes                  From birth                                                                               Yes                            Yes                        LFTs, FBC              US\\$1.43/month\n  **Interferon (IFN)**                             Biologic response modifier                   \\+ (genotype dependent)                                    No                                                                                 Anorexia, diarrhea, flu-like symptoms, neurotoxicity, seizures, hepatotoxicity   No                   Not recommended in children (\\>18 years only)[\\*\\*\\*](#t001fn004){ref-type=\"table-fn\"}   N/A                            Yes                        LFTs, FBC, TFTs        Not listed\n  **HBV immuno-globulin (HBIG) for prophylaxis**   Biologic response modifier                   ++                                                         N/A                                                                                Abdominal pain, buccal ulceration, chest pain                                    Yes                  From birth                                                                               N/A                            No                         N/A                    US\\$38.02/dose\n\n**Abbreviations:** ART, antiretroviral therapy (for HIV infection); FBC, full blood count; HBV, hepatitis B virus; LFT, liver function test; N/A, nonapplicable; TFT, thyroid function test; YMDD, tyrosine-methionine-aspartic acid-aspartic acid motif.\n\n\\*Potency against HBV is defined as + or ++ to differentiate between agents with lower and higher suppressive capacity, respectively.\n\n\\*\\* British National Formulary (<https://www.bnf.org/>) states tenofovir can be prescribed for HIV in infants \\>2 years, but data for HBV treatment are lacking.\n\n\\*\\*\\* British National Formulary (<https://www.bnf.org/>) states Peg-interferon-alpha can be prescribed for chronic hepatitis C virus (HCV) in infants \\>5 years, but data for HBV treatment are lacking. <https://www.medicinescomplete.com/mc/bnfc/current/>\n\nPrevention of mother to child transmission (PMTCT) can be improved through a combination of routine antenatal screening, antiviral drugs during the latter stages of pregnancy, and HBV vaccination to the baby starting at birth. Where resources permit, HBV immunoglobulin (HBIG) can further reduce the risk of vertical transmission.\n\nDespite the efficacy of these strategies in managing or preventing individual cases, these interventions do not currently offer a route to global HBV eradication, due to a shortage of investment and resources, the large pool of undiagnosed cases, lack of routine diagnostic screening, the high cost of IFN and HBIG, the lack of a curative therapy, substantial gaps in drug and vaccine coverage, and the potential for increasing drug resistance \\[[@pntd.0005842.ref008]\\].\n\nApplication of NTD criteria to HBV {#sec003}\n==================================\n\nWe have applied the WHO criteria for NTDs to HBV \\[[@pntd.0005842.ref004]\\] and refer to case studies and experience from our own clinical practice ([S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}) to illustrate how HBV in Africa fulfills NTD criteria. Some of the factors underpinning the neglect of this infection are summarized in [Table 2](#pntd.0005842.t002){ref-type=\"table\"}.\n\n10.1371/journal.pntd.0005842.t002\n\n###### Summary of factors potentially contributing to the neglect of investment in hepatitis B virus (HBV) clinical care, research, advocacy, and education.\n\n![](pntd.0005842.t002){#pntd.0005842.t002g}\n\n  Factors contributing to HBV neglect\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  \u2022 **Stigma and discrimination** leading to lack of patient voice ([S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 5, 6, 7) \\[[@pntd.0005842.ref009]\\].\n  \u2022 **Silent infection**, which may never be diagnosed and is not apparent to onlookers (contributes to large pool of undiagnosed infection).\n  \u2022 **Poverty**, leading to lack of patient voice, lack of public profile, and underrepresentation ([S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 4, 5, 7).\n  \u2022 **Complacency** that ongoing deployment of existing resources and approaches (e.g., suppressive antiviral therapy and vaccination) is sufficient to bring about elimination \\[[@pntd.0005842.ref010]\\].\n  \u2022 **High burden in low-/middle-income countries** \\[[@pntd.0005842.ref006]\\], where investment is not a priority.\n  \u2022 **Lack of public/media representation**; no \"high profile\" cases.\n  \u2022 **HBV is \"eclipsed\" by higher profile infections** such as HIV and malaria.\n  \u2022 **Poor education and knowledge** among patients, the public, and healthcare workers ([S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 6, 8, 9) including underrecognition of the global burden of infection.\n  \u2022 **Lack of existing investment** \\[[@pntd.0005842.ref011],[@pntd.0005842.ref012]\\] contributing to a cycle of underinvestment ([Fig 2](#pntd.0005842.g002){ref-type=\"fig\"}).\n  \u2022 **Lack of development of infrastructure** through which to provide education, prevention, diagnosis, and treatment and as a way to collect robust data.\n  \u2022 **Poor-quality data** (poor understanding of epidemiology and risk factors, little recognition of the impact of stigma, lack of assessment regarding feasibility of interventions, etc.). ([S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 7, 8)\n  \u2022 **Lack of major dedicated funding agencies**.\n\n![Resource gap in research funding allocations and academic publications for hepatitis B virus (HBV), hepatitis C virus (HCV), HIV, and malaria.\\\nPanels A/C: funding data from the United States National Institutes for Health (NIH) estimated funding for research, condition, and disease categories 2013--2018 (\\*projected figures for 2017 and 2018), available at <https://report.nih.gov/categorical_spending.aspx>, downloaded June 2017. For the projected funding allocation for 2018, HCV will receive 2.3-fold HBV funding, malaria 4.8-fold, and HIV 66.8-fold. Research into \"malaria\" and \"malaria vaccine\" are subdivided in the source data set but have been pooled in this graphic. Panels B/D: We recorded the number of publications listed on NCBI PubMed based on the search terms \"HIV,\" \"HBV,\" \"HCV,\" and \"malaria\" for each year from 2007--2016. Example search string for HBV publications in 2016: (HBV\\[Title\\]) AND (\\\"2016/01/01\\\"\\[Date---Publication\\]: \\\"2016/12/31\\\"\\[Date---Publication\\]). Data are represented as absolute numbers (panels A and B) and the proportion of the whole (panels C and D). For hepatitis delta virus (HDV), funding allocation data are not available, and we identified \\<25 publications/year (range 7--23).](pntd.0005842.g002){#pntd.0005842.g002}\n\nNTDs \"primarily affect populations living in tropical and subtropical areas\" {#sec004}\n----------------------------------------------------------------------------\n\nAlthough HBV is endemic globally, the bulk of morbidity and mortality is now borne by low-/middle-income countries in tropical and subtropical regions \\[[@pntd.0005842.ref006],[@pntd.0005842.ref013]\\]. In Africa, many populations are particularly vulnerable due to coendemic HIV infection and other coinfections, host and viral genetic factors, poverty, and lack of education and infrastructure \\[[@pntd.0005842.ref007]\\]. In this setting, HBV has been eclipsed by the more acute and tangible health crisis of human immunodeficiency virus (HIV); only now in the era of antiretroviral therapy (ART) is it reemerging as a visible threat \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; case 2\\]. One illustration of this shift is the increase in deaths from HBV-related liver cancer over time that contrasts a reduction in AIDS deaths \\[[@pntd.0005842.ref014]\\].\n\nNTDs \"disproportionately affect populations living in poverty and cause...morbidity and mortality, including stigma and discrimination\" {#sec005}\n---------------------------------------------------------------------------------------------------------------------------------------\n\nHBV is part of a cycle of poverty, with a high burden of morbidity and mortality in young adults. The economic burden on individual families can be particularly catastrophic in low- and middle-income settings \\[[@pntd.0005842.ref015]\\], although robust data are lacking for Africa. In resource-poor settings, lack of education and scarce healthcare resources impinge on successful diagnosis and monitoring as well as failure to control symptoms where relevant. Stigma and discrimination are often invisible but can be potent and highly relevant challenges to the success of scaling up interventions for prevention, diagnosis, and treatment \\[[@pntd.0005842.ref009]\\] \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 1, 4, 5, 6, 7, 9\\].\n\nNTDs are \"immediately amenable to broad control, elimination, or eradication by applying...public health strategies\" {#sec006}\n--------------------------------------------------------------------------------------------------------------------\n\nWe already have an armamentarium of strategies with which to tackle HBV prevention and treatment ([Fig 3](#pntd.0005842.g003){ref-type=\"fig\"}). In order to be widely and robustly deployed, these approaches should interlink with existing resources and infrastructure wherever possible \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; case 2\\].\n\n![A package of interventions to move towards elimination of hepatitis B virus (HBV) infection as a public health threat.\\\nSuggested measures are aligned with WHO interventions for neglected tropical diseases (NTDs).](pntd.0005842.g003){#pntd.0005842.g003}\n\nNTDs are \"relatively neglected by research---i.e., resource allocation is not commensurate with the magnitude of the problem\" {#sec007}\n-----------------------------------------------------------------------------------------------------------------------------\n\nCompared with other blood-borne viruses, namely, HIV and hepatitis C virus, which infect substantially lower numbers \\[[@pntd.0005842.ref007]\\], HBV has attracted far fewer research resources, and this gap may actually be widening over time \\[[@pntd.0005842.ref011],[@pntd.0005842.ref012]\\] ([Fig 2](#pntd.0005842.g002){ref-type=\"fig\"}). A recent report of infectious diseases research investment in United Kingdom institutions documents that 0.7% of total expenditure is for HBV, compared to 3.0% for HCV, 13.9% for malaria, and 17.5% for HIV \\[[@pntd.0005842.ref012]\\]. HBV mortality (887,000 deaths/year \\[[@pntd.0005842.ref003]\\]) is now twice that of malaria (429,000 deaths/year \\[[@pntd.0005842.ref016]\\]), but malaria receives nearly 5-fold more funding ([Fig 2](#pntd.0005842.g002){ref-type=\"fig\"}). Hepatitis delta virus (HDV), a copassenger with HBV, can drive aggressive liver disease but is almost completely overlooked in terms of resource allocation. Moreover, development of clinical programs for hepatitis testing and treatment are fragmented in comparison to the progressive infrastructure that has emerged to tackle HIV \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; case 7\\].\n\nRecommendations based on NTD framework {#sec008}\n======================================\n\nEven for an organism that is not officially recognized as an NTD, there is much to be learnt from the NTD paradigm that could accelerate progress in tackling HBV. The ethos of combining several public health strategies and integrating care for different diseases is captured by the approach advocated for NTDs \\[[@pntd.0005842.ref004]\\] and is also a helpful strategy for HBV. Particularly in the African subcontinent, where other NTD models have had significant impact \\[[@pntd.0005842.ref017]\\], using this framework for HBV could promote awareness, leverage advocacy and resources, and promote integration of HBV prevention and treatment into existing HIV infrastructure \\[[@pntd.0005842.ref005]\\].\n\nIn the following section, we use suggested interventions for NTDs to discuss briefly how these are pertinent to reducing---and ultimately eliminating---HBV infection as a public health threat.\n\n\"Intensified case management\" {#sec009}\n-----------------------------\n\nBased on the significant numbers of individuals lost at every step of the \"cascade\" from diagnosis through to successful treatment and prevention ([Fig 1](#pntd.0005842.g001){ref-type=\"fig\"}), enhanced efforts are needed to promote linkage through care pathways. Enhanced HBV testing is crucial to facilitate entry into clinical care, allowing treatment to reduce the risk of onward spread, including underpinning PMTCT \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; case 8\\]. Initially, this may rely on using existing diagnostic platforms (based on serology), but investment is required in developing and rolling out new approaches, including molecular testing strategies that are more sensitive, provide enhanced data (e.g., detection of drug resistance), and are fast enough to enable point-of-care testing. This can often be transferred from technology that has been initially developed for the diagnosis of other diseases. Definitive curative therapy for HBV remains a crucial aspiration, as elimination using existing strategies is not realistic during the time frames set by SDGs \\[[@pntd.0005842.ref010]\\]. Specifically, a therapeutic vaccine harnessing successful immune responses to boost immune clearance would provide a huge leap forward in tackling the existing burden of chronic infection \\[[@pntd.0005842.ref018]\\].\n\nThe role and significance of stigma associated with HBV infection in Africa is largely unreported in the literature. However, individual testimony leaves no doubt that this is a significant barrier to diagnosis and clinical care \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 5, 6\\]. Gaining a better understanding of the extent and nature of stigma and discrimination in different populations is a crucial first step, in parallel with enhanced efforts to educate patients, health care workers, and the public.\n\n\"Preventive chemotherapy\" {#sec010}\n-------------------------\n\nAlthough antiviral therapy for HBV is generally regarded as treatment rather than prevention, in the majority of cases, it renders individuals aviremic, preventing onward transmission. Antiviral therapy for HBV ([Table 1](#pntd.0005842.t001){ref-type=\"table\"}) should be made accessible, ideally capitalizing on the supply chains and distribution infrastructure that have been developed for HIV (and/or other prevalent infections, such as tuberculosis and malaria) \\[[@pntd.0005842.ref005]\\]. Research efforts are still required to identify prognostic factors that predict differential response to therapy and allow tailoring of care.\n\nPMTCT can progressively become a realistic goal by expanding access to antenatal diagnostics, simple treatment interventions such as maternal tenofovir during trimester 3, and HBV vaccination for all babies, with the first dose delivered at birth \\[[@pntd.0005842.ref008]\\]. Vaccination remains a cornerstone of prevention, but more work is needed to investigate the most effective catch-up immunization strategies to reduce the burden of HBV infection at a population level \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; cases 3, 4, 8\\].\n\n\"Sanitation and hygiene\" {#sec011}\n------------------------\n\nAlthough this category of interventions is conventionally applied to reducing food- and waterborne infections, we here broaden our interpretation to include other aspects of prophylaxis. Safety and security of medical supplies has increasingly improved to reduce nosocomial transmission of blood-borne viruses over recent decades \\[[S1 Supporting Information](#pntd.0005842.s001){ref-type=\"supplementary-material\"}; case 3\\]. However, sterile practices need to be more widely promoted and guaranteed to assure the safety of other procedures such as scarification, tattoos, piercings, and circumcision that may occur in community settings. Provision of condoms alongside education regarding safe sex, particularly for high-risk groups such as sex workers and men who have sex with men, is another important strategy for prevention.\n\nConclusions {#sec012}\n===========\n\nElimination of HBV infection has gained status within international health and development agendas but is a complex clinical and public health challenge that currently lacks proportionate multilateral commitment from pharma, government, commissioners, funders, and the research community. The many parallels with other NTDs are clearly exemplified by vulnerable populations of the African subcontinent. By viewing HBV---as well as its partner in coinfection, HDV---within the NTD framework, we can improve approaches to reducing the burden of disease and move towards eventual elimination.\n\nSupporting information {#sec013}\n======================\n\n###### This document contains supporting information to corroborate the view that hepatitis B virus (HBV) can helpfully be represented within the framework set out for neglected tropical diseases (NTDs) by the World Health Organization (WHO) \\[[@pntd.0005842.ref001]\\].\n\nThis is in line with aims stated within Sustainable Development Goals (SDGs) \\[[@pntd.0005842.ref002]\\]. Complementary evidence gathered from patients, researchers, and healthcare workers from different locations in Africa illustrates the ways in which HBV infection meets the criteria for NTDs. These scenarios (designated cases 1 to 9 and presented geographically in order from south to north) contribute important insights into how the NTD paradigm can be helpful in informing strategies to improve diagnosis, treatment, and prevention of HBV infection, with the ultimate goal of eliminating infection as a public health threat.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n[^1]: MA has received research funding from Gilead.\n"}
+{"text": "INTRODUCTION {#s01}\n============\n\nAs one of the oldest and most important predatory arthropods, the centipede has a fossil record that extends back 420 million years ([@bUndheim2011]). Approximately 3 300--3 500 centipede species have been found, with distribution worldwide and in most provinces of China ([@bRong2015]). Centipede venom, which is secreted from venom glands in the first pair of limbs ([@bEdgecombe2007]), is essential for survival, not only for subduing and killing prey but also for defense against predators.\n\nAnimal venom has long been considered a rich source of pharmacological and novel therapeutics ([@bKalia2015]; [@bSmith2013]; [@bZhang2015]). Furthermore, dried centipedes have been used medicinally for centuries, as shown in ancient Chinese medical records. Recently, an increasing number of studies have shown that centipede venom contains various functional components, including a rich reservoir of structural and pharmacological peptides ([@bHakim2015]; [@bUndheim2015], [@bUndheim2016]). In addition, because of their excellent chemical and pharmacological activities, particularly as neurotoxins and ion channel inhibitors, centipede toxins have received further attention ([@bLiu2012]; [@bYang2012], [@bYang2013], [@bYang2015]). Several antimicrobial peptides and specific toxins have also been identified in centipede venom ([@bChen2014]; [@bHou2013]; [@bPeng2010]; [@bYang2012]). Interestingly, centipede toxins are expressed outside the venom gland and are involved in gene recruitment processes ([@bZhao2018a]). These venom peptides have significant chemical, thermal, and biological stability, which enable researchers to adapt their functions for therapeutic use.\n\nTherefore, centipede venom research is of great interest for investigating putative toxins. These toxins can act on a range of molecular targets, including voltage-gated sodium (Na~V~), potassium (K~V~), and calcium (Ca~V~) channels ([@bLiu2012]; [@bYang2012]). However, biochemical studies on centipede toxins are not nearly as extensive as studies on other venomous animals, such as snakes, spiders, and scorpions ([@bUndheim2016]), and complete data on centipede venom toxins, peptides, and protein sequences are currently limited to a small number of species ([@bHakim2015]; [@bUndheim2016]). One potential reason is that most centipede species are considered too small to obtain enough venom for activity testing or high-throughput drug screening. Omics analysis of venom or venom glands is one approach for probing toxin molecular diversity. Specifically, to identify new putative proteins and enable comparison across species, large-scale sequencing of a broad array of centipede venom should be applied to further confirm the complexity of venom ([@bGonzalez-Morales2014]; [@bLiu2012]; [@bRong2015]).\n\nPrevious centipede research has mainly focused on *Scolopendra mutilans* ([@bZhao2018a]), and occasionally on *S. subspinipes subspinipes*, *S. viridis*, and *S. dehaani* ([@bLiu2012]). To date, however, no comprehensive research has been reported on the new pharmaceutical centipede, *S. mojiangica* ([@bWang1997]), which is used as a substitute medicinal material in traditional Chinese medicine. Therefore, a fully integrated approach combining transcriptomics and proteomics is essential for understanding the differences among pharmaceutical centipedes, including venom composition and toxin diversity. Here, in-depth proteotranscriptomic analyses (combined proteomic and transcriptomic analyses) were used to study centipede venom, and the protein/peptide composition of the dissected venom gland from *S. mojiangica* was described. Complete comparative analyses of the protein compounds and toxin distribution in the venom or venom gland of *S. mojiangica* and *S. mutilans* were also presented based on RNA-Seq and MS datasets.\n\nMATERIALS AND METHODS {#s02}\n=====================\n\nAnimals and ethics {#s03.01}\n------------------\n\nAdult *S. mojiangica* (both sexes) were collected from Mojiang (N23\u00b027\\', E101\u00b041\\'), Yunnan Province, China. All centipede (*S. mojiangica*) studies were reviewed and approved by the Animal Care and Use Committee of Puer University (ACUP. 531068520180126, approved on 17 September 2018).\n\nVenom collection and sample preparation {#s04.02}\n---------------------------------------\n\nThe venom of *S. mojiangica* was collected as per our previous method. Briefly, a 3 V alternating current (AC) was used to stimulate the venom glands in the first pair of centipede limbs ([@bLiu2012]). The venom samples were stored at \u221220 \u00b0C until use. A 300 mg *S. mojiangica* venom sample was solubilized in 3 mL of Tris-HCl buffer. The venom solution was then loaded on a Sephacryl S-100HR (HiprepTM26/60, 71-1247-00-EG, GE Healthcare, USA) gel filtration column with a flow rate of 0.5 mL/h. Thirteen peaks (named P1\u221213) were obtained from this procedure (Supplementary Figure S1).\n\nThe proteins/peptides contained in the venom were pre-denatured with 500 \u00b5L of 25 nmol/L NH~4~HCO~3~ and separated with a 3 kDa cut-off ultrafiltration tube. The low molecular weight (\\<3 kDa) proteins/peptides were collected and desalinated before peptidomic analysis. Proteins/peptides with molecular weights greater than 3 kDa were applied to SDS-PAGE gels for separation. One half of each sample was mixed with extraction buffer (0.25% acetic acid and protease inhibitor cocktail) and disrupted with a sonicator (Hielscher Ultrasound Technology, Germany). To further separate these samples, 12% gel with protein ladder (Thermo, ref. 26614, USA) SDS-PAGE was used, followed by staining with GelCode Blue Stain (Thermo ref. 24592, USA) and destaining with Milli-Q water (Millipore, USA). We excised six bands from each lane for in-gel trypsin digestion. Samples were extracted with 100% acetonitrile, desalinated, lyophilized, and stored at \u221280 \u00b0C until further electrospray ionization-tandem mass spectrometry (ESI-MS/MS) analysis.\n\nRNA extraction, sequencing, and transcriptome analysis {#s05.01}\n------------------------------------------------------\n\nA total of 260 mg of venom gland was preserved in liquid nitrogen after collection from *S. mojiangica* until use. RNA extraction and cDNA library construction were performed according to our previous work ([@bZhao2014a], [@bZhao2014b]). cDNA from the *S. mojiangica* venom gland was sequenced using the Illumina HiSeq\u2122 2000 (USA), and the short-read assembly program SOAPdenovo-Trans (v1.03) was run with default parameters to complete *de novo* transcriptome assembly. Overlaps with certain lengths and connected paired-end reads were combined in the program to form contigs. The sequence clustering software TGICL was used to splice sequences and remove redundant sequences to produce the complete assembly of contigs of each sample ([@bPertea2003]), and the longest possible non-redundant unigenes were produced. The TGICL parameters were the same as the parameters used in our previous work ([@bZhao2014b]).\n\nHPLC fractionation and mass spectrometry {#s06.03}\n----------------------------------------\n\nAfter in-gel digestion, candidate fractionation samples were loaded onto an EASY-nLC HPLC system (Thermo Fisher Scientific, USA) equipped with a binary rapid separation nano-flow pump and ternary loading pump. Mobile phase eluent A (0.1% TFA contained in ddH~2~O) and mobile phase eluent B (ACN/ddH~2~O/TFA 90/10/0.08% (v/v/v)) were used. Samples were applied to a Thermo Scientific EASY loading column (2 cm\u00d7100 \u00b5m, 5 \u00b5m \u2212C18, USA) by the auto-sampler and analytical column (75 \u00b5m\u00d7100 mm, 3 \u00b5m \u2212C18), respectively, with a flow rate of 250 nL/min. With linear stepwise gradients (0\u2032\u22125% B, 5\u2032\u22125% B, 12.5\u2032\u221220% B, 62.5\u2032\u221270% B, 63.5\u2032\u221299% B, 65\u2032\u221299% B, 66\u2032\u22125% B and 72\u2032\u22125% B), we separated the peptides with the column. Starting at 20% eluent B, 1.25 mL/5 min of each fraction was collected and lyophilized.\n\nWe selected the data-dependent mode of the Q Exactive instrument (Thermo Finnigan, USA), which then switched between full scan MS and MS/MS acquisition automatically. Based on the predictive automatic gain control (AGC) of the previous full scan, we accumulated 3\u00d710^6^ target value ions and acquired 70 000 (m/z 200) resolution of full scan MS spectra (m/z 300\u22121 800) in the Orbitrap. In addition, 15 s was set as the dynamic exclusion value. We isolated and fragmented the 10 most intense multiply charged ions (z\u22652) sequentially by higher-energy collisional dissociation (HCD) with a fixed resolution of 17 500 (m/z 200) and an injection time of 60 ms for the MS2 scanning method. The mass spectrometric conditions were as follows: 2 kV spray voltage, no sheath and auxiliary gas flow, 250 \u00b0C heated capillary temperature, 27 eV normalized HCD collision energy, and 0.1% underfill ratio. A total of 1\u00d710^5^ counts was set as the ion selection threshold for MS/MS.\n\nData processing and bioinformatics analysis {#s07.04}\n-------------------------------------------\n\nUsing Proteome Discoverer (version 1.4), RAW data files were produced. Mascot v2.2 was used as the search tool to generate peak lists in our transcriptome database. Trypsin was chosen as an enzyme, and two missed cleavages were allowed. The MS/MS search criteria were as follows: MS polypeptide tolerance 2\u00d710^4^ mg/m\u00b3 and MS/MS mode 0.1 Da. The aminomethylation of cysteine was statically modified and the oxidation of methionine was dynamically modified. High confidence peptides were used for protein identification, generating a 1% false discovery rate (FDR) threshold. Only unique peptides with high confidence were used for protein identification.\n\nAll unigenes in our centipede database were annotated with BLASTX and searched against known databases, as presented in our previous study ([@bZhao2014a], [@bZhao2014b], [@bZhao2018a]). Unigenes were aligned with high-priority databases and annotated with a given description instead of aligning with a low-priority database. Gene Ontology (GO) annotation was carried out using the Blast2GO ([@bConesa2005]) software suite v2.5.0. In these searches, the BLASTX cut-off was set to 1e^-6^. The BLAST tool was used to search the toxin database and annotate the toxin with Tox-Prot in UniProtKB (02 February 2019, 6 822 sequences) and the animal toxin database platform ATDB ([@bHe2008]), with the toxins then verified by phylogenetic analyses. The grouped sequences were aligned using MUSCLE v3.8.31 ([@bEdgar2010]). MrBayes 3.2.7 was used for phylogenetic analyses with maximum likelihood. The values were estimated by ultrafast bootstrap using 10 000 iterations. The resulting trees were analysed with MEGA 7 ([@bKumar2016]), which was also used to automatically plot expression values and detection in venom.\n\nComparative expression analysis was performed as follows: comparison of RNA-Seq data of venom glands of various species was performed using Bowtie v0.12.7 ([@bLangmead2009]) and TopHat v2.0.6 ([@bTrapnell2009]) for mapping. Gene expression values were calculated from the expected number of fragments per kilobase of transcript sequence per million base pairs sequenced (FPKM) ([@bTrapnell2010]). The FPKM values for genes from every tissue were determined by rSeq ([@bJiang2009]). The graphs and statistical analyses were performed using GraphPad Prism v5.0 (La Jolla, USA) and R v3.3.2. Here, *P*\\<0.05 was considered statistically significant.\n\nInsect bioassays and hemolytic assays {#s08.01}\n-------------------------------------\n\nInsect bioassays were performed according to the method in [@bYang2012]. Freeze-dried crude venom powder was dissolved in insect saline (concentrations in deionized water: 140 mmol/L NaCl, 5 mmol/L KCl, 4 mmol/L NaHCO~3~, 1 mmol/L MgCl~2~, 0.75 mmol/L CaCl~2~, 5 mmol/L HEPES) and injected into grasshoppers (*Locusta migratoria manilensis*; mass 700--900 mg) and mealworms (*Tenebrio molitor* larvae; mass 190--210 mg). Ants (*Tetramorium spp*., adults; mass 35--55 mg) were fed with same venom.\n\nUsing human, mouse, and rabbit red blood cells (RBCs), hemolytic activity was assayed as described previously ([@bLiu2012]; [@bZhao2018b]). Briefly, serial dilutions of the samples were incubated with washed RBCs (3%) at 37 \u00b0C for 30 min and then centrifuged. The resulting supernatant was measured at an absorbance of 540 nm. Maximum hemolysis was determined by adding 1% Triton X-100 to the cell samples.\n\nRESULTS {#s09}\n=======\n\nPhylogeny of scolopendrid centipedes and isolation of venom gland {#s10.01}\n-----------------------------------------------------------------\n\nOriginal Chinese medicinal centipedes include *S. mutilans*, *S. multidens*, *S. mojiangica*, and *S. negrocapitis* ([@bWang1997]). Here, we studied the novel substitutional pharmaceutical centipede, *S. mojiangica*, with comparative analysis of active molecules. *Scolopendra mojiangica* showed a relatively close relationship to *S. negrocapitis*, *S. mutilans*, and *S. multidens* ([Figure 1A](#Figure1){ref-type=\"fig\"}), though a smaller body size than *S. mutilans*, *S. dehaani*, and *S. multidens*. Similar to other species, it also uses venom to attack prey and in defense.\n\n![Proteomic and transcriptomic analyses of new pharmaceutical centipede](zr-41-2-138-1){#Figure1}\n\nThe protocol for isolating venom glands from *S. mojiangica* was described in our previous study ([@bLiu2012]). Healthy adult centipedes (*n*=280) without injury were selected, and the venom glands were dissected from their first pair of limbs. After that, 3 V AC was used to stimulate the venom gland and ensure that more toxins were included, so that proteome coverage could be improved. The isolated venom glands were then further processed ([Figure 1B](#Figure1){ref-type=\"fig\"}). A portion of each sample was used to obtain the proteome by SDS-PAGE analysis. Protein bands from the venom gland were excised for in-gel digestion and subjected to ESI-MS/MS analysis. The remaining portion of each sample was used to extract RNA, followed by RNA-Seq analysis of the transcriptome.\n\nProteomic analysis of venom components {#s11.01}\n--------------------------------------\n\nA total of 246 proteins were identified in *S. mojiangica* at 95% coverage by ESI-MS/MS analysis (Supplementary Table S1; [Figure 2A](#Figure2){ref-type=\"fig\"}). In the proteome, 73.6% of proteins (*n*=181) were cellular components and 19.1% of proteins (*n*=47) were unknown functional proteins, which were putative venom toxins. Only 18 proteins were identified as toxin-like proteins, including neurotoxins, K^+^ channel inhibitors, and blarina toxins ([Figure 2B](#Figure2){ref-type=\"fig\"}; [Table 1](#Table1){ref-type=\"table\"}). Although we obtained more proteins in *S. mojiangica* than in *S. mutilans* and *S. viridis* with proteomic analysis, the detected toxin-like proteins in *S. mojiangica* represented a lower percentage of total proteins than those identified in *S. mutilans* in our previous study ([Figure 2C](#Figure2){ref-type=\"fig\"}). In the venom proteome, most of the identified proteins showed a molecular weight of less than 50 kDa, similar to the proteome of *S. mutilans* ([Figure 2D](#Figure2){ref-type=\"fig\"}). Thus, the centipedes contained notably small functional molecules for potential pharmaceutical use, as expected. Based on peptide detection, 23.2% of proteins consisted of six or more unique peptides (Supplementary Figure S2). In addition, the more enriched the peptides assembled into proteins, the more comprehensive was the proteome obtained.\n\n###### Toxin-like proteins/peptides identified from venom proteome of *S. mojiangica* centipede\n\n  Sequence ID          GenBank accession No.   Sequence description                                Category                 Peptides   E-Value     MW (kD)   Calc. pI   FPKM\n  -------------------- ----------------------- --------------------------------------------------- ------------------------ ---------- ----------- --------- ---------- -----------\n  ScoMo_singlet48841   AT0003236               Blarina toxin precursor (EC 3.4.21.-)               Blarina toxin            9          1.00E-37    21.61     4.15       92.54\n  ScoMo_singlet50899   AT0003766               Mucrofibrase-5 precursor (EC 3.4.21.-)              Mucrofibrase-5           11         4.00E-16    14.40     9.93       3 454.74\n  ScoMo_singlet71394   AT0002263               Pseudechetoxin-like protein precursor               Pseudechetoxin           276        9.00E-42    28.74     9.86       7 195.57\n  ScoMo_contig2076     gi\\|429840589           K+ channel inhibitor                                Channel inhibitor        617        4.00E-164   62.76     9.15       1.37\n  ScoMo_singlet78309   AT0000117               Latisemin precursor                                 Latisemin                412        2.00E-22    20.89     7.96       0.00\n  ScoMo_contig4762     AT0003236               Blarina toxin precursor (EC 3.4.21.-)               Blarina toxin            108        1.00E-44    28.58     6.5        15 173.32\n  ScoMo_singlet45908   AT0003741               Thrombin-like enzyme contortrixobin (EC 3.4.21.-)   Serine proteinase        109        1.00E-41    44.94     5.08       1 685.57\n  ScoMo_singlet67462   AT0000120               Pseudecin precursor                                 Pseudechetoxin           66         5.00E-32    23.71     8.91       14 111.58\n  ScoMo_singlet72573   AT0000552               Hopsarin-D (EC 3.4.21.6)                            Hopsarin-D               93         1.00E-121   85.15     6.53       132.70\n  ScoMo_singlet76606   AT0000554               Trocarin precursor (EC 3.4.21.6)                    Trocarin                 38         3.00E-138   84.92     6.17       60.34\n  ScoMo_singlet25641   AT0000552               Hopsarin-D (EC 3.4.21.6)                            Hopsarin-D               46         5.00E-20    27.21     4.6        184.53\n  ScoMo_singlet69905   AT0000554               Trocarin precursor (EC 3.4.21.6)                    Trocarin                 14         4.00E-107   40.69     5.28       1 245.366\n  ScoMo_singlet57737   AT0003404               Zinc metalloproteinase fibrolase (EC 3.4.24.72)     Metalloproteinase        20         4.00E-16    35.21     8.13       48.71\n  ScoMo_singlet8256    AT0000762               Alpha-latrocrustotoxin                              Alpha-latrocrustotoxin   10         0           50.48     6.79       136.27\n  ScoMo_singlet68890   AT0000552               Hopsarin-D (EC 3.4.21.6)                            Hopsarin-D               13         5.00E-75    42.03     7.88       161.84\n  ScoMo_singlet7846    gi\\|392295725           Omega-slptx-ssm2a neurotoxin precursor              Neurotoxin               11         8.00E-36    8.56      4.93       16 647.01\n  ScoMo_singlet55496   gi\\|501293796           Cathepsin L                                         Cathepsin L              180        1.00E-155   37.30     6.35       2.83\n  ScoMo_singlet39956   AT0000554               Trocarin precursor (EC 3.4.21.6)                    Trocarin                 12         4E-09       4.64      3.79       5.55\n\n![Profiles of *S. mojiangica* proteome](zr-41-2-138-2){#Figure2}\n\nTranscriptomic analysis of venom components {#s12.02}\n-------------------------------------------\n\nWe acquired 43 381 437 clean reads assembled into 132 597 contigs from the venom gland using the Trinity program. As a result, the transcriptome data consisted of 107 642 putative gene objects (all unigenes) ranging from 101 bp to 9 184 bp, with an average length of 423 bp. The number of unigenes larger than 500 bp was 24 219. The largest unigenes were 9 184 bp in size, and the N50 of the unigenes was 214 bp (Supplementary Figure S3 and Table S2).\n\nFor comparative analysis, the venom gland transcriptome from *S. mojiangica* showed many transcripts (*n*=46 571) with high similarity to those of *S. mutilans*. Notably, however, most transcripts showed low similarity between the two centipede species ([Figure 3A](#Figure3){ref-type=\"fig\"}). In the transcriptomic expression analysis, the read count of each transcript in *S. mojiangica* and *S. mutilans* showed biases for gene expression, with higher expressed transcripts in *S. mojiangica* ([Figure 3B](#Figure3){ref-type=\"fig\"}). Functional annotation analyses of these transcripts were combined with Blast searching and phylogenetic analyses to obtain toxin-like unigenes. In total, 410 toxin-like transcripts were identified in the transcriptome of *S. mojiangica*, more than that identified in *S. mutilans* (342 transcripts). Furthermore, these transcripts were divided into 34 categories, mainly consisting of alpha-latrocrustotoxin, delta-latroinsectotoxin, ion channel inhibitors, and alpha-latrotoxin ([Figure 3C](#Figure3){ref-type=\"fig\"}).\n\n![Identification of toxins from transcriptome of venom gland in centipedes](zr-41-2-138-3){#Figure3}\n\nComparative determination of centipede toxins {#s13.04}\n---------------------------------------------\n\nAs expected, we identified 34 kinds of toxin-like unigenes (*n*=342) from the transcriptome of *S. mutilans* using the same annotation method as that of *S. mojiangica* ([Figure 4A](#Figure4){ref-type=\"fig\"}). In total, 11 of these toxin-like unigenes encoded the most transcripts in the two centipedes. With gene expression analyses, most toxin-like unigenes showed no differential expression between *S. mojiangica* and *S. mutilans*, except for four toxin-like unigenes (i.e., alpha-latrotoxin, hopsarin-D, metalloproteinase, and trocarin) ([Figure 4B](#Figure4){ref-type=\"fig\"}).\n\n![Comparison of toxin-like molecules distributed in centipedes *S. mojiangica* and *S. mutilans*](zr-41-2-138-4){#Figure4}\n\nFinally, we determined the toxicity and performed crude isolation of the centipede venom. The crude centipede venom exhibited strong insecticidal action ([Figure 5A](#Figure5){ref-type=\"fig\"}), and the crude venom had a similar potency as the venom of *S. mutilans*. The crude venom and its fractions eluted from the S-100HR column (Supplementary Figure S1; [Figure 5B](#Figure5){ref-type=\"fig\"}) showed hemolytic activity. The elution of peak 1 (P1) showed high hemolytic activity on human RBCs when 1 mg/mL protein/peptide was incubated for 4 h. In contrast, peaks 3, 5, and 6 (P3, P5, and P6) had lower hemolytic activity than that of P1 and crude venom.\n\n![Insecticidal activity of crude centipede venom](zr-41-2-138-5){#Figure5}\n\nDISCUSSION {#s14}\n==========\n\nDue to long-term evolutionary fine-tuning, venom toxins exhibit high specificity and potency for molecular targets that are not often found in natural or synthetic small molecules, and thus animal toxins are valuable pharmacological tools ([@bKing2011], [@bKing2013]). There are many cases in which venom toxin has been used as a pharmacological molecule, e.g., snake venom, dried toad skin secretions (Chan Su), tarantula venom, and cobra venom used as traditional Ayurvedic, Chinese, Mexican, and Central and South American medicines, respectively ([@bHarvey2014]; [@bKing2011]). These traditional medicines have been used to treat arthritis, gastrointestinal ailments, asthma, polio, multiple sclerosis, rheumatism, severe pain, and trigeminal neuralgia, or as a diuretic anesthetic and anti-cancer agent. Centipede venom has different biomedical properties and represents a vast reservoir of toxins, similar to venom from other animals. Due to its origins in one of the oldest venomous arthropods, centipede venom displays excellent activities and good prospects for drug development ([@bUndheim2016]; [@bZhang2015]). Importantly, the centipede is a traditional Chinese medicine with an application history of more than 2 000 years ([@bChen1999]; [@bZhao2018a]). In China, pharmaceutically applied centipedes include *S. mutilans*, *S. multidens*, *S. dehaani*, and *S. negrocapitis*, with *S. mojiangica* ([@bWang1997]) very occasionally used as a substitute. Our results showed that the venom toxicity of this centipede is strong in comparison to that of *S. mutilans*, a commonly used centipede in medicine.\n\nIn our previous study, the centipede showed diverse protein or peptide components, with the most abundant toxins in the venom and torso tissues found to be more highly expressed than other active molecules using our method ([@bLiu2012]; [@bZhao2018a]). Here, based on proteomic detection, we showed that the toxin-like proteins in *S. mojiangica* accounted for a lower percentage of total proteins than that in *S. mutilans*. However, there was a similar constitution and quantity of toxin transcripts in these two centipedes. We used high-throughput ESI-MS/MS and RNA-Seq technology to investigate the diversity of novel venom proteins, especially low-abundance peptides/proteins not detected using conventional methods ([@bSavitski2005]). Most of the detected proteins were identified as potentially active molecules with low molecular weights and unknown functions. In addition, each detected protein contained at least six peptides in the proteome dataset. The proteomic results for *S. mojiangica* were very similar to the protein detection results for *S. mutilans*. More than 400 toxin-like proteins/peptides were identified by transcriptome analysis in the centipede, but not detected in the proteome. Thus, most putative toxins in centipede venom may have low levels of expression in *S. mojiangica* and *S. mutilans*. In conclusion, centipede venom contains a surprising variety of toxin-like proteins/peptides.\n\nRegarding toxin distribution, based on transcriptomic analysis, we identified more toxin transcripts in *S. mojiangica* than in *S. mutilans*. Most toxins did not show significantly differential expression between *S. mojiangica* and *S. mutilans*, including that of ion channel inhibitors and serine proteinases. The centipede *S. mojiangica* demonstrated higher gene expression of metalloproteinase, trocarin, hopsarin-D, and alpha-latrotoxin compared to *S. mutilans*. Therefore, *S. mojiangica* could be substituted for *S. mutilans* in medical use. These results indicate that *S. mojiangica* venom could be a rich source of pharmacologically and medically useful compounds.\n\nUsually, we can obtain approximately 0.2--0.5 mg of crude venom from a single adult *S. mutilans* centipede over a period of two weeks. However, one adult *S. mojiangica* yielded less than 0.1 mg of crude venom in the same period. Therefore, it was difficult to study the venom components, including their pharmaceutical activity or medicinal application. In addition to the current annotation methods of centipede toxins, our results revealed that a wide variety of toxin-like active molecules were expressed in the venom gland by combining Blast alignment with the existing toxin databases and phylogenetic reconstruction of toxin relationships. Theoretically, this method may produce false positives, especially for proteins with low abundance and expression when using high-throughput proteomic and transcriptomic analyses with ESI-MS/MS and RNA-Seq technology. However, we used previously established approaches to maximize the search for functional proteins. Our results provide good evidence that the use of this substitute medicinal centipede is an appropriate medical option. Importantly, our data provide important clues to improve the use of the centipede as a traditional Chinese medicine.\n\nCONCLUSIONS {#s15}\n===========\n\nHere, we used omics techniques to determine the profiles of venom components and toxin-like molecules in a new pharmaceutical centipede, *S. mojiangica*. We performed in-depth proteomic analysis of venom and deduced full-length protein sequences by combining proteome and transcriptome databases. We obtained more than 400 toxin-like molecules with potent activity. With gene expression and inter-species comparative analysis, we identified a broad and diverse composition of toxin-like molecules, which may play key roles in the functions of centipede venom. Our results indicate that this centipede is valuable for medicinal use and drug development, like other centipede species. Furthermore, our methods could improve the application of the centipede as a traditional Chinese medicine.\n\nSUPPLEMENTARY DATA {#s16}\n==================\n\n###### \n\nSupplementary data to this article can be found online.\n\n###### \n\nClick here for additional data file.\n\nCOMPETING INTERESTS {#s17}\n===================\n\nThe authors declare that they have no competing interests.\n\nAUTHORS\\' CONTRIBUTIONS {#s18}\n=======================\n\nF.Z., W.H.L., and Y.Z. conceived the study. F.Z. wrote the paper and performed data and bioinformatics analyses. Z.C.L., J.Y.L., X.Q.L., T.L., J.R.Z., F.Z., G.L., and P.Y.C. performed all experiments. All authors read and approved the final version of the manuscript.\n\nACKNOWLEDGMENTS {#s19}\n===============\n\nWe thank the members of our research groups for providing technical assistance and participating in discussions.\n\n[^1]: \\#Authors contributed equally to this work\n"}
+{"text": "Introduction and Motivations {#Sec1}\n============================\n\nThe current approach to forecasting modelling consists of simulating explicitly only the largest-scale phenomena, while taking into account the smaller-scale ones by means of \"physical parameterisations\". All numerical models introduce uncertainty through the selection of scales and parameters. Additionally, any computational methodology contributes to uncertainty due to discretization, finite precision and accumulation of round-off errors. Finally the ever growing size of the computational domains leads to increasing sources of uncertainties. Taking into account these uncertainties is essential for the acceptance of any numerical simulation. Numerical forecasting models often use Data Assimilation methods for the uncertainty quantification in the medium to long-term analysis. Data Assimilation (DA) is the approximation of the true state of some physical system at a given time by combining time-distributed observations with a dynamic model in an optimal way. DA can be classically approached in two ways: as variational DA \\[[@CR16]\\] and as filtering \\[[@CR5]\\]. In both cases we seek an optimal solution. The most popular filtering approach for data assimilation is the Kalman Filter (KF) \\[[@CR15]\\]. Statistically, KF seeks a solution with minimum variance. Variational methods seek a solution that minimizes a suitable cost function. In certain cases, the two approaches are identical and provide exactly the same solution \\[[@CR16]\\]. However, the statistical approach, though often complex and time-consuming, can provide a richer information structure, i.e. an average and some characteristics of its variability (probability distribution). During the last 20\u00a0years hybrid approaches \\[[@CR11], [@CR18]\\] have become very popular as they combine the two approaches into a single taking advantage of the relative rapidity and robustness of variational approaches, and at the same time, obtaining an accurate solution \\[[@CR2]\\] thanks to the statistical approach. In this paper, in order to achieve the accuracy of the KF solution and reduce the execution time, we use Recurrent Neural Networks (RNN). Today the computational power of RNN is exploited for several application in different fields. Any non-linear dynamical system can be approximated to any accuracy by a Recurrent Neural Network, with no restrictions on the compactness of the state space, provided that the network has enough sigmoidal hidden units. This is what the Universal Approximation Theorem \\[[@CR12], [@CR20]\\] claims. Only during the last few years, the DA community is starting to approach machine learning models to improve the efficiency of DA models. In \\[[@CR17]\\], the authors combined Deep Learning and Data Assimilation to predict the production of gas from mature gas wells. They used a modified deep LSTM model as their prediction model in the EnKF framework for parameter estimation. Even if the *prediction* phase is speed up due to the introduction of Deep Learning, this only partially affects the whole *prediction*-*correction* cycle which is still time-consuming. In \\[[@CR9]\\], the authors presented an approach for employing artificial neural networks (NNs) to emulate the local ensemble transform Kalman filter (LETKF) as a method of data assimilation. Even if the Feed Forward NN they implemented is able to emulate the DA process for the time window they fixed, when they need to assimilate observations in new time steps, it still needs the *prediction*-*correction* cycle and this affects the execution time which is just 90 times faster than the reference DA model. To further speed up the process, in \\[[@CR8]\\] the authors combined the power of Neural Networks and High Performance Computing to assimilate meteorological data. These studies, alongside others discussed in conferences and still under publication, highlight the necessity to avoid the *prediction*-*correction* cycle by developing a Neural Network able to completely emulate the whole Data Assimilation process. In this context, we developed a Neural Assimilation (NA) as a Coupled Neural Network made of two RNNs. NA captures the features of a Data Assimilation process by interleaving the training of the two component RNNs on the forecasting data and the observed data. That is, the two component RNNs are trained on forecasting and observed data respectively with additional inputs provided by the interaction of these two. This NA network emulates the KF and runs much faster than the KF *prediction*-*correction* cycle for data assimilation. In this paper we develop the NA architecture and proved its equivalence to the KF. The equivalence between NA and KF is independent from the structure on the RNNs. In this paper we show results we obtained employing two Long short-term memory (LSTM) architectures for the two RNNs. Then we employ the NA model to a practical problem in predicting of oxygen (and drugs) diffusion across the Blood-Brain Barrier (BBB) \\[[@CR1]\\] to justify its correctness and efficiency.\n\nThis paper is structured as follows. In Sect.\u00a0[2](#Sec2){ref-type=\"sec\"} the Data Assimilation problem is described. The Neural Assimilation is introduced in Sect.\u00a0[3](#Sec3){ref-type=\"sec\"}, where we investigate the accuracy of the introduced method and we present a theorem demonstrating that the novel model is consistent with the KF result. Experimental results are provided in Sect.\u00a0[4](#Sec4){ref-type=\"sec\"}. Conclusions and future works are summarised in Sect.\u00a0[5](#Sec5){ref-type=\"sec\"}.\n\nData Assimilation {#Sec2}\n=================\n\nData Assimilation (DA) is the approximation of the true state of some physical system at a given time by combining time-distributed observations *o*(*t*) with a dynamic model $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\dot{x}=\\mathcal {M}(x,t)$$\\end{document}$ in an optimal way. DA can be classically approached in two ways: as variational DA \\[[@CR3]\\] and as filtering. One of the best known tools for filtering approach is the Kalman filter (KF) \\[[@CR15]\\]. We seek to estimate the state *x*(*t*) of a discrete-time dynamic process that is governed by the linear difference equation$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} x(t)=M\\ x(t-1)+w_t \\end{aligned}$$\\end{document}$$with an observation *o*(*t*):$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} o(t)=H\\ x(t)+v_t \\end{aligned}$$\\end{document}$$Note that *M* and *H* are discrete operators. The random vectors $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$w_t$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$v_t$$\\end{document}$ represent the modeling and the observation errors respectively. They are assumed to be independent, white-noise processes with normal probability distributions$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} w_t \\sim \\mathcal {N}(0,B_t), \\qquad v_t \\sim \\mathcal {N}(0,R_t) \\end{aligned}$$\\end{document}$$where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$B_t$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$R_t$$\\end{document}$ are covariance matrices of the modeling and observation errors respectively. All these assumptions about unbiased and uncorrelated errors (in time and between each other) are not limiting, since extensions of the standard KF can be developed should any of these not be valid \\[[@CR5]\\]. The KF problem can be summarised as follows: given a background estimate *x*(*t*), of the system state at time *t*, what is the best analysis *z*(*t*) based on the current available observation *o*(*t*)?\n\nThe typical assimilation scheme is made up of two major steps: a *prediction* step and a *correction* step. At time *t* we have the result of the previous forecast, *x*(*t*) and the result of an ensemble of observations *o*(*t*). Based on these two vectors, we perform an analysis that produces *z*(*t*). We then use the evolution model to obtain a prediction of the state at time $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t+1$$\\end{document}$. The result of the forecast at the *prediction* step is denoted with $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$x(t+1)$$\\end{document}$$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} x(t+1)=M z(t), \\end{aligned}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} B_{t+1}=M \\left( (1-K_{t} H )B_{t} \\right) M^T, \\end{aligned}$$\\end{document}$$and becomes the background for the next *correction* time step:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} K_{t+1}=B_{t+1} H^T (H B_{t+1} H^T+R_{t+1})^{-1}, \\end{aligned}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} z(t+1)=x(t+1)+K_{t+1} \\left( o(t+1)-H x(t+1)\\right) , \\end{aligned}$$\\end{document}$$We observe that, in case the observed data are defined in the same space of the state variable, the operator $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$H_t$$\\end{document}$ in ([2](#Equ2){ref-type=\"\"}) is the identity matrix and the Eqs.\u00a0([6](#Equ6){ref-type=\"\"})--([7](#Equ7){ref-type=\"\"}) can be simplified becoming:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} K_{t+1}=B_{t+1} ( B_{t+1} +R_{t+1})^{-1}, \\end{aligned}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} z(t+1)=x(t+1)+K_{t+1} \\left( o(t+1)- x(t+1)\\right) , \\end{aligned}$$\\end{document}$$Due to the high computational cost in updating the covariance matrices $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$B_t$$\\end{document}$ by Eq.\u00a0([5](#Equ5){ref-type=\"\"}), it in operational DA, is often used to assume $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$B_t=B_{t+1}$$\\end{document}$ $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\forall t$$\\end{document}$. This assumption leads to a model which is also called Optimal Interpolation \\[[@CR16]\\].\n\nStatistically, KF seeks a solution with minimum variance. This approach, though often complex and time-consuming, can provide a rich information structure (often richer than information provided by variational DA), such as an average and some characteristics of its variability (probability distribution). In order to maintain the accuracy of the KF solution and reduce the execution time, we introduce, in the next section, a Neural Assimilation (NA) which is a network representing KF but much faster than a KF *prediction*-*correction* cycle.\n\nNeural Assimilation {#Sec3}\n===================\n\nFor a fixed time window $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$[t_0, t_1]$$\\end{document}$ and a fixed discretization time step $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\varDelta t$$\\end{document}$, let *x*(*t*) still denote the forecasting result at each time step $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t \\in [t_0, t_1]$$\\end{document}$. Let *o*(*t*) denotes an observation of the state value (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). As it does not affect the generality of our study, we are assuming here the observed data defined in the same space of the state variable, i.e. the operator $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$H_t$$\\end{document}$ in ([2](#Equ2){ref-type=\"\"}) is the identity matrix.Fig. 1.Available data in the fixed time window.\n\nGiven the data sets $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\{x(t)\\}_{t\\in [t_0,t_1]}$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\{o(t)\\}_{t\\in [t_0,t_1]}$$\\end{document}$, the Neural Assimilation (NA) is a Coupled Neural Network (for temporal processing) as shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}, where:a first forecasting network NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_F$$\\end{document}$ is a Recurrent Neural Network trained on forecasting data *x*(*t*) with an additional input provided by a second forecasting network NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_O$$\\end{document}$ trained on observed data *o*(*t*);a second forecasting network NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_O$$\\end{document}$ is a Recurrent Neural Network trained on observed data *o*(*t*) with an additional input provided by a first forecasting network NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_F$$\\end{document}$.\n\nA fundamental feature of each network is that it contains a feedback connection, so the activations can flow round in a loop. That enables the networks to do temporal processing and learn sequences with temporal prediction. The form of NA is a RNN with the previous set of hidden unit activations feeding back into the network along with the inputs.Fig. 2.Neural Assimilation\n\nNote that the time *t* is discretized, with the activations updated at each time step. The time scale might correspond to any time step of size appropriate for the given problem. A delay unit given by the network NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_F$$\\end{document}$ needs to be introduced to hold activations in NN$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$_O$$\\end{document}$ until they are processed at the next time step and vice versa. As for simple architectures and deterministic activation functions, learning will be achieved using similar gradient descent procedures to those leading to the back-propagation algorithm for feed forward networks.\n\nThe NA scheme is made up of two major steps: a *pre-processing* step and a *training* step. During the *pre-processing* step, the data set is normalized considering the information we have about the error estimations and the error covariance matrices introduced in ([3](#Equ3){ref-type=\"\"}). We consider, to normalise, the inverse of the error covariance matrices so that, data with big covariance/variance are assumed with a small weight \\[[@CR5], [@CR16]\\]. We pose$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\bar{x}(t)=B_t^{-1} x(t)\\quad \\text {and} \\quad \\bar{o}(t)=R_t^{-1}o(t). \\end{aligned}$$\\end{document}$$The computed vectors $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\bar{x}(t)$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\bar{o}(t)$$\\end{document}$ are the data used in the *training* step:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\bar{o}(t) = f_{O_O} \\left( W_{HO_O} h(t-1) \\right) \\end{aligned}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} h(t)=f_H\\left( W_{IH} \\bar{x}(t-1) +W_{HH} h(t -1)\\right) \\end{aligned}$$\\end{document}$$ $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\bar{x}(t) = f_{O_F} \\left( W_{HO_F} h(t) \\right) \\end{aligned}$$\\end{document}$$where the vectors $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\bar{x}(t-1)$$\\end{document}$ are the inputs, the matrices $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$W_{IH}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$W_{HH}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$W_{HO_F}$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$W_{HO_O}$$\\end{document}$ are the four connection weight matrices, and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$f_H$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$f_{O_F}$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$f_{O_O}$$\\end{document}$ are the hidden and outputs unit activation functions. The state of the dynamical system is a set of values that summarizes all the information about the past behaviour of the system that is necessary to provide a unique description of its future behaviour, apart from the effect of any external factors. In this case the state is defined by the set of hidden unit activations *h*(*t*). The Back propagation Through Time for this algorithm is a natural extension of standard back propagation that performs gradient descent on a complete unfolded network (\\[[@CR21]\\],\u00a0Chapter\u00a05 of \\[[@CR6]\\]). If the NA training sequence starts at time $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t_0$$\\end{document}$ and ends at time $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t_1$$\\end{document}$, the total cost function is simply the sum over time of the standard error function *C*(*t*) at each time-step:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} C_{total}=\\sum _{t=t_0}^{t_1} C(t) \\end{aligned}$$\\end{document}$$where$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} C(t)= \\frac{1}{2} \\sum _{k=1}^n \\left( (\\bar{o}_{k}(t-1)-h_{k}(t-1))^2 + (\\bar{x}_{k}(t)- h_{k}(t))^2\\right) \\end{aligned}$$\\end{document}$$and *n* is the total number of training samples. The gradient descent weight updates have contributions from each time-step \\[[@CR19]\\]:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\varDelta w_{ij} = - \\eta \\frac{\\partial C_{total}(t_0,t_1) }{\\partial w_{ij}} = - \\eta \\sum _{t=t_0}^{t_1} \\frac{\\partial C(t) }{\\partial w_{ij}} \\end{aligned}$$\\end{document}$$where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\eta $$\\end{document}$ is the learning rate \\[[@CR14]\\]. The constituent partial derivatives $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\frac{\\partial C(t) }{\\partial w_{ij}}$$\\end{document}$ have contributions from the multiple instances of each weight$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$w_{ij}\\in \\left\\{ W_{IH} ,W_{HH}, W_{HO_O} , W_{HO_F} \\right\\} $$\\end{document}$$and depend on the inputs and hidden unit activations at previous time steps. The errors now have to be back-propagated through time as well as through the network \\[[@CR23]\\].\n\nWe prove that the output function *h*(*t*) of the NA model corresponds to the solution of Kalman filter with fixed covariance matrices, i.e. in its Optimal Interpolation version \\[[@CR16]\\]. The following result held.\n\nTheorem 1 {#FPar1}\n---------\n\nLet *h*(*t*) be the solution of NA given by Eqs.\u00a0([10](#Equ10){ref-type=\"\"})--([16](#Equ16){ref-type=\"\"}) and let *z*(*t*) denote the solution of the KF algorithm as defined in ([9](#Equ9){ref-type=\"\"}). We have$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} h(t)=z(t), \\quad \\forall t\\in [t_0,t_1] \\end{aligned}$$\\end{document}$$**Proof:** Due to the definition of the $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$L^2$$\\end{document}$ norm, the loss function in ([15](#Equ15){ref-type=\"\"}) can be written as$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} C(t)=\\Vert \\bar{o}(t-1)-h(t-1) \\Vert _2^2+ \\Vert \\bar{x}(t)- h(t) \\Vert _2^2 \\end{aligned}$$\\end{document}$$then, from Eq.\u00a0([1](#Equ1){ref-type=\"\"}), and except for the numerical errors that will be introduced later as already included in the data sets, the ([18](#Equ18){ref-type=\"\"}) can be written as:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} C(t)=\\Vert \\bar{o}(t-1)-h(t-1) \\Vert _2^2+ \\Vert M\\ \\bar{x}(t-1)- M\\ h(t-1) \\Vert _2^2 \\end{aligned}$$\\end{document}$$From the properties of the $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$L^2$$\\end{document}$ norm, the ([19](#Equ19){ref-type=\"\"}) can be written as$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} C(t)=&(\\bar{o}(t-1)-h (t-1))^T (\\bar{o}(t-1)-h(t-1)) \\nonumber \\\\&+\\,(M \\bar{x}(t-1)- M h(t-1) )^T(M \\bar{x}(t-1)- M h(t-1) ). \\end{aligned}$$\\end{document}$$To minimise this loss function, we compute the gradient$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\nabla _{h(t-1)} C(t)= 2 (\\bar{o}(t-1)-h(t-1)) + 2 M^T (M \\bar{x}(t-1)- M\\ h(t-1) ) \\end{aligned}$$\\end{document}$$where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$M^T$$\\end{document}$ denotes the Adjoint operator of the linear operator *M* \\[[@CR7]\\] and we pose $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\nabla _{h(t-1)} C(t)=0$$\\end{document}$, then we have:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} 2 h(t-1) = \\bar{o}(t-1)+ \\bar{x}(t-1) \\end{aligned}$$\\end{document}$$From the definition of $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\bar{x}$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\bar{o}$$\\end{document}$ in ([10](#Equ10){ref-type=\"\"}), the ([22](#Equ22){ref-type=\"\"}) gives:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} h(t-1) \\left( B_{t-1}+R_{t-1}\\right) = R_{t-1} x(t-1) + B_{t-1} o(t-1) \\end{aligned}$$\\end{document}$$Then, adding and subtracting the quantity $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$B_{t-1} x(t-1)$$\\end{document}$ and merging the common factors, the ([23](#Equ23){ref-type=\"\"}) become$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} h(t-1) \\left( B_{t-1}+R_{t-1}\\right) = x(t-1) \\left( B_{t-1}+R_{t-1}\\right) +B_{t-1}\\left( o(t-1)-x(t-1)\\right) \\end{aligned}$$\\end{document}$$Finally, posed $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$Q_{t-1}=B_{t-1} \\left( B_{t-1}+R_{t-1}\\right) ^{-1}$$\\end{document}$, the ([24](#Equ24){ref-type=\"\"}) gives:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} h(t-1)= x(t-1) +Q_{t-1}\\left( o(t-1)-x(t-1)\\right) \\end{aligned}$$\\end{document}$$which is the expression of the KF solution $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$z(t-1)$$\\end{document}$ in ([9](#Equ9){ref-type=\"\"}) for the time step $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t-1$$\\end{document}$ and for the case of observed data defined in the same space of the state variable (i.e. $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$H=I$$\\end{document}$ and *I* is the identity matrix). $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$Q_{t-1}$$\\end{document}$ is the Kalman gain matrix in\u00a0([8](#Equ8){ref-type=\"\"}).\n\nThe Eq.\u00a0([25](#Equ25){ref-type=\"\"}) in Theorem\u00a0[1](#FPar1){ref-type=\"sec\"} represents a condition to assume that NA is consistent with KF.\n\nIn Sect.\u00a0[4](#Sec4){ref-type=\"sec\"}, we validate the results provided in this section. We also show that the employment of NA alleviates the computational cost making the running less expensive.\n\nExperimental Results {#Sec4}\n====================\n\nIn this section we provide experimental results that demonstrate the applicability and efficiency of NA. In our experiment, the NA is implemented by adopting Long short-term memory (LSTM) architecture for the two RNNs. The reason we use LSTMs is that they are suitable to contain information outside the normal flow of the recurrent network so it is easier to plug two networks together. Also, LSTMs allow to preserve the error that can be backpropagated through time and layers which is a very important point for discrete forecasting models. A description of the NA we implemented is provided in Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}.Fig. 3.Implementation of Neural Assimilation\n\nThe test case we consider is a numerical model to predict the oxygen diffusion across the Blood-Brain Barrier (BBB). Nevertheless the model can be used for any drugs by replacing the diffusion constant and the initial and boundary conditions \\[[@CR1]\\]. The Blood-Brain Barrier protects the central nervous system, controls the entry of compounds into the brain by restricting access for blood borne compounds and facilitates access for nutrients. This protection makes it difficult to provide therapeutic compounds to brain cells when they are affected by brain diseases as Alzheimer, Autism \\[[@CR13]\\]. The BBB is composed of endothelial cells connected by tight junctions. The main mechanisms allowing the transport of drugs across the membrane are passive transport, carrier-mediated transport, receptor-mediated transcytosis, and adsorption-mediated transcytosis \\[[@CR22]\\]. The passive transport mechanism is the easiest method of drug transport for lipophilic and low molecular size molecules. It means a simple diffusion across any membrane without application of energy and carrier proteins. Opioids and steroids are examples of drugs which can be passively diffused \\[[@CR4]\\]. Assuming that the main transport mechanism is through passive diffusion, the initial three-dimensional space problem can be reduced to a one-dimensional space problem. In fact, passive diffusion involves many simplifications as no reaction term, uniform movement in all directions and an overall diffusion constant. Therefore, a 1D partial differential equation (PDE) as ([26](#Equ26){ref-type=\"\"}) with one initial condition and two boundary conditions is an accurate model for this problem \\[[@CR22]\\] where 0 corresponds to the location at which the blood meets the Blood-Brain Barrier and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$L=400$$\\end{document}$\u00a0nm is the real average thickness of the Blood-Brain Barrier.$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} \\left\\{ \\begin{array}{l} \\frac{\\partial x}{\\partial t}=D \\frac{\\partial ^2x}{\\partial y^2} \\\\ x(0,y)=x_{0,y} \\\\ x(t,0)= x_{t,0}\\\\ x(t,L)=x_{t,L} \\end{array}\\right. \\end{aligned}$$\\end{document}$$where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t\\in [0,10\\,\\mathrm{{ms}}]$$\\end{document}$ (*ms* denotes microsecond) and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y\\in [0,L]$$\\end{document}$. We consider that, at time 0 there is no oxygen, then $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ x_{0,y}=0 $$\\end{document}$. Moreover, for our boundary conditions, we consider that we have a constant concentration of oxygen in the bloodstream and that at the interface of the barrier and the brain tissue all oxygen will be consumed $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ x_{t,0}= 0.02945$$\\end{document}$\u00a0L/L\u00a0blood and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ x_{t,L}=0 $$\\end{document}$. We assume the diffusivity of oxygen through the Blood-Brain Barrier to be $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$3.24*10^{-5}$$\\end{document}$ c$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$m^2$$\\end{document}$/s \\[[@CR1]\\].\n\nEquation\u00a0([26](#Equ26){ref-type=\"\"}) is discretised by a second order central finite difference in space with $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\varDelta y=8$$\\end{document}$\u00a0nm and a backward Euler method in time with $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\varDelta t=0.1$$\\end{document}$\u00a0ms:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ -Fx_{i-1}^n+(1+2F)x_i^n-x_{i+1}^n=x_{i-1}^{n-1} $$\\end{document}$$where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$F=D\\frac{\\varDelta t}{\\varDelta y^2}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$i=1,\\dots ,50$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$n=1,\\dots , 100$$\\end{document}$. As we know that it does not affect the generality of our study, in this paper we show results of NA using observed data *o*(*t*) provided in \\[[@CR1]\\] by the analytical solution of ([26](#Equ26){ref-type=\"\"}) for the oxygen diffusion. The model can be used for any drugs by replacing the diffusion constant and the initial and boundary conditions. Data sets for observed data can be found in <http://cheminformatics.org/datasets/>. The NA code and the pre-processed data can be downloaded using the link <https://drive.google.com/drive/folders/1C_O-rk5wyqFsG5U-T7_vugBOddTPmOlY?usp=sharing>.\n\nThe NA network has been trained using the $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$85\\%$$\\end{document}$ of the data and tested on the remaining $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$15\\%$$\\end{document}$. Figure\u00a0[4](#Fig4){ref-type=\"fig\"} shows the value of the Loos function for training and testing the forecasting network.Fig. 4.Values of the loss function.\n\nNA has been compiled as a sequential neural network with just one LSTM layer of 48 units using as loss function the mean squared error one and as optimiser the Adam one. Weights are automatically initialised by Keras using:Glorot uniform for the kernel weights matrix for the linear transformation of the inputs;Orthogonal for the linear transformation of the recurrent state. Fig. 5.Temporal evolution of the concentration at (a) $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm and (b) $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=35$$\\end{document}$\u00a0nm\n\nFigure\u00a0[5](#Fig5){ref-type=\"fig\"} shows the temporal evolution of the concentration at $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}a) and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=35$$\\end{document}$\u00a0nm (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}b). The accuracy of the NA results is evaluated by the absolute error$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} e_{NA}(t,y)=\\vert z(t,y)-h(t,y) \\vert \\end{aligned}$$\\end{document}$$and the mean squared error$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} MSE(h(t,y)) = \\frac{\\Vert z(t,y)-h(t,y) \\Vert _{L^2}}{\\Vert z(t,y) \\Vert _{L^2}} \\end{aligned}$$\\end{document}$$where *z*(*t*,\u00a0*y*) is the solution of KF performed at each time step. Table\u00a0[1](#Tab1){ref-type=\"table\"} shows values of absolute error computed every 10 time steps. We can see that the order of magnitude of the error is between $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$e{-}07$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$e{-}04$$\\end{document}$. The corresponding values of mean squared error are $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE(h(t,y))=1.31e{-}07$$\\end{document}$ for $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE(h(t,y))=8.16e{-}08$$\\end{document}$ for $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=35$$\\end{document}$\u00a0nm where $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t\\in [0,0.10\\,\\mathrm{{ms}}]$$\\end{document}$. Figure\u00a0[6](#Fig6){ref-type=\"fig\"} shows the comparison of the KF result and the NA result for the temporal evolution of the concentration at each point of the BBB we are modelling. Values of execution time are provided in Table\u00a0[2](#Tab2){ref-type=\"table\"}. The values are computed as mean of execution times from 100 runnings. We can observe that the time for *prediction* in NA is 1000 faster than the *prediction* with KF.Table 1.Error computed every 10 time steps at (a) $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm and (b) $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=35$$\\end{document}$\u00a0nmTime step *t*$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$e_{NA}(t,y)$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$e_{NA}(t,y)$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=35$$\\end{document}$\u00a0nm000107.05e\u2212046.11e\u221204204.17e\u2212044.88e\u221204304.29e\u2212041.91e\u221204401.52e\u2212046.05e\u221207502.51e\u2212049.11e\u221205603.40e\u2212051.11e\u221204704.13e\u2212051.05e\u221204804.72e\u2212057.35e\u221205901.11e\u2212041.89e\u2212051001.60e\u2212043.18e\u221205 Table 2.Execution time for 100 time steps and all the distancesExecuting time (s)Neural Assimilation (training)121.47Neural Assimilation (prediction)0.117Kalman filter (prediction)138\n\nFig. 6.Comparison between Data Assimilation (KF) and the Neural Network version for $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$t\\in [0,10\\,\\mathrm{{ms}}]$$\\end{document}$ (*ms* denotes microsecond) and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y\\in [0,400\\,\\mathrm{{nm}}]$$\\end{document}$.\n\nFinally, Table\u00a0[3](#Tab3){ref-type=\"table\"} shows the values of mean square forecasting error:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} MSE^{F}(x(t,y)) = \\frac{\\Vert x(t,y)-o(t,y) \\Vert _{L^2}}{\\Vert o(t,y) \\Vert _{L^2}} \\end{aligned}$$\\end{document}$$and mean square assimilation error:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\begin{aligned} MSE^{NA}(h(t,y)) = \\frac{\\Vert h(t,y)-o(t,y) \\Vert _{L^2}}{\\Vert o(t,y) \\Vert _{L^2}} \\end{aligned}$$\\end{document}$$computed with respect observations *o*(*t*,\u00a0*y*). The values of the errors in the assimilation results present a reduction of approximately one order of magnitude with respect to the error in forecasting data.Table 3.Mean square error forecasting error $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE^F$$\\end{document}$ and mean square assimilation error $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE^{NA}$$\\end{document}$ computed every 10 time steps at $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nmTime step *t*$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE^{NA}(h(t,y))$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$MSE^{F}(x(t,y))$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$y=12$$\\end{document}$\u00a0nm05.03e\u2212031.00e\u221202105.48e\u2212031.00e\u221202205.34e\u2212039.99e\u221203305.06e\u2212039.95e\u221203404.88e\u2212031.00e\u221202504.80e\u2212031.00e\u221202604.78e\u2212031.00e\u221202704.80e\u2212031.00e\u221202804.83e\u2212031.00e\u221202904.87e\u2212031.00e\u2212021004.90e\u2212031.00e\u221202\n\nConclusions and Future Works {#Sec5}\n============================\n\nWe introduced a new neural network for Data Assimilation (DA) that we named Neural Assimilation (NA). We proved that the solution of NA is the same of KF. We tested the validity of the provided theoretical results showing values of misfit between the solution of NA and the solution of KF for the same test case. We provided experimental results on a realistic test case studying oxygen diffusion across the Blood-Brain Barrier. NA is trained on both forecasting and observed data and it is used for *predictions* without needing a correction given by the information provided by observations. This allows to avoid the *prediction*-*correction* cycle of a Kalman filter, and it makes the assimilation process very fast. We show that the time for *prediction* in NA is 1000 faster than the *prediction* with KF. An implementation of NA to emulate variational DA \\[[@CR10]\\] will be developed as future work. In particular, we will focus on a 4D variational (4DVar) method \\[[@CR5]\\]. 4DVar is a computational expensive method as it is developed to assimilate several observations (distributed in time) for each time step of the forecasting model. We will develop an extended version of NA able to assimilate set of distributed observations for each time step and, then, able to emulate 4DVar.\n\nThis work is supported by the EPSRC Centre for Mathematics of Precision Healthcare EP/N0145291/1 and the EP/T003189/1 Health assessment across biological length scales for personal pollution exposure and its mitigation (INHALE).\n"}
+{"text": "Key Points\u00a0for Decision Makers {#FPar11}\n==============================\n\nImmunization delivery costs are a necessary component of high-quality cost-effectiveness models, and are also used to inform resource mobilization for immunization programs.Our study provides estimates produced via meta-regression analyses that can help improve resource mobilization and planning in situations where empirical cost data are unavailable or of low quality.\n\nBackground {#Sec1}\n==========\n\nRoutine immunization is critical to achieving 14 of the 17 Sustainable Development Goals (SDGs) adopted by countries in 2015 to \"ensure prosperity for all\" \\[[@CR1]\\]. To monitor progress towards these goals, to plan for the financial sustainability of immunization programs, and to improve program coverage and equity, decision-makers need to know how much immunization programs cost beyond the cost of the vaccine. Non-vaccine immunization delivery unit cost estimates are essential for resource mobilization and planning for routine health systems, which have required the development of new financing mechanisms to support the increasing costs of country immunization programs \\[[@CR2]\\]. Moreover, delivery cost estimates can help to identify and evaluate strategies to improve efficiency in vaccine service delivery \\[[@CR3]\\]. Cost estimates can significantly influence the cost-effectiveness estimates used to allocate resources at the country level, particularly as service delivery can be the main driver of immunization program costs \\[[@CR4], [@CR5]\\]. In addition, low- and middle-income countries (LMICs) often require an updated immunization program costing to apply for support from Gavi, the Vaccine Alliance \\[[@CR3], [@CR6]\\]. Despite the multiple uses of accurate cost estimates, many LMICs have not conducted an empirical study of immunization services costs, or rely on estimates that are highly uncertain, unreliable, or old. While this evidence gap has been narrowed by recent efforts to improve the production and collation of immunization costing data---with 37 LMICs represented in available costing studies \\[[@CR7]--[@CR9]\\]---these efforts would need to be expanded substantially in order to supply all countries with up-to-date and high-quality cost estimates.\n\nThe objective of this study was to produce standardized country-level estimates of immunization delivery unit costs for all countries meeting the World Bank's LMIC classification (136 total) \\[[@CR10]\\], via an evidence synthesis of available data on immunization delivery costs \\[[@CR9], [@CR11]\\]. Using these data, we fit a Bayesian meta-regression model for routine childhood (i.e., under-five) vaccination program delivery unit costs. We present estimates of immunization delivery cost per dose by year, by cost category (labor, supply chain, capital, and other service delivery costs), and by country.\n\nMethods {#Sec2}\n=======\n\nWe developed a Bayesian meta-regression model to predict immunization delivery unit costs, excluding vaccines and supplies such as syringes and safety boxes, in LMICs, according to World Bank country income classification in 2019 \\[[@CR10]\\]. In this section, we describe the data used to fit the model, modeling methods, and alternative regression specifications.\n\nData Extraction {#Sec3}\n---------------\n\nWe relied on a publicly available database describing immunization delivery costs in LMIC settings---the Immunization Delivery Cost Catalogue (IDCC) maintained by the Immunization Costing Action Network (ICAN) \\[[@CR9], [@CR11]\\]. The IDCC is an online web catalog and downloadable Excel spreadsheet of immunization delivery cost evidence in LMIC settings, which describes the results of a systematic review of published and grey literature available between January 2005 and March 2019. The search of the peer-reviewed literature included six major electronic databases, including EconLit, Embase, Medline (via PubMed), NHS Economic Evaluation Database (NHS-EED), Web of Science, and World Health Organization (WHO) Global Index Medicus. Search terms included three categories of keywords--- \"immunization\" AND \"cost\" AND \"delivery\"---translated into the query language of each database. All resources with full-text availability in English, French, or Spanish, conducted in LMIC settings, that included a form of delivery unit cost data from primary data collection were included. In addition to extracting relevant contextual and methodological information from each study, the IDCC presents cost estimates in 2016 US dollars \\[[@CR9], [@CR11]\\].\n\nFrom the IDCC, we identified studies that reported delivery cost per dose of routine (i.e., fixed facility) vaccine delivery, and which specified the cost categories included in the estimate. We excluded 39 studies that did not report a cost per dose or for which a cost per dose could not be calculated, studies that did not define the cost categories included in the cost per dose, and studies that focused solely on the costs of vaccines delivered through outreach or through supplementary immunization activities (SIAs), i.e., mass vaccination campaigns.\n\nFor the identified studies, we extracted estimates of the routine delivery cost per dose for childhood immunization, defined as vaccination of children under 5\u00a0years of age. We also extracted study-specific contextual information, including the number of sampled sites, when reported by a study, whether the study examined programmatic (i.e., full) cost per dose or single antigen (i.e., the incremental delivery costs of a single new vaccine) cost per dose, economic and/or financial cost per dose, and cost category cost per dose. An economic cost analysis estimates the annualized value of capital investments and the value of donated goods and labor time; whereas, a financial cost estimate is based on the financial outlay for capital equipment and excludes the value of donated goods and services \\[[@CR7]\\]. If any information reported in the IDCC was unclear for this analysis, we confirmed the information in the original studies. For the observed cost per dose outcome, we extracted the observation with the highest level of granularity available from a given study. For example, if the total vaccine delivery cost per dose was also reported as individual cost categories (i.e., labor, supply chain, capital, and other service delivery), we extracted these data as independent observations and excluded the total vaccine delivery cost per dose. \"Supply chain\" includes costs for cold chain equipment, vehicles, transport, and fuel; \"other service delivery\" includes costs for program management (i.e., supervision and monitoring), training, social mobilization, and disease surveillance \\[[@CR6], [@CR7]\\].\n\nWe compiled data on covariates potentially associated with immunization delivery unit costs: the year of data collection (*Year*), a proxy for time-varying characteristics not captured by other model covariates; log number of doses in the routine vaccination schedule \\[log(*Doses*)\\], an indicator of the economies of scale for vaccine delivery; log gross domestic product (GDP) per capita \\[log(*GDP*)\\], an indicator of country price levels; reported diphtheria-tetanus-pertussis third dose coverage (*DTP3*), an indicator of routine health system capacity and overall coverage of the immunization program; and log total country population \\[log(*Pop*)\\], an indicator of service delivery volume \\[[@CR12], [@CR13]\\]. Model covariates log(*Doses*), log(*GDP*), and log(*Pop*) were compiled for the year of data collection for each study. By using a log transformation of specified covariates, we assumed these explanatory factors relate to the outcome on the multiplicative scale rather than linearly (for example, a doubling in per capita GDP produces a fixed increase in the outcome). The number of doses in the routine vaccination schedule was derived by reviewing country immunization schedules \\[[@CR14]\\]. As historical vaccination schedules were not available for all calendar years, we assumed that the schedule for the most recently available year prior to the reported base year of the study applied.\n\nPrediction Model {#Sec4}\n----------------\n\nWe used a Bayesian meta-regression model to regress immunization delivery unit costs against country-level and study-level explanatory variables (Table [1](#Tab1){ref-type=\"table\"}). Continuous variables \\[*Year*, log(*Doses*), log(*GDP*), *DTP3*, and log(*Pop*)\\] were standardized to mean zero and unit standard deviation before fitting the regression model. We adopted an analytic model that allowed the synthesis of cost estimates that included different combinations of cost categories. Under this approach, a regression equation with a separate cost intercept was specified for each of four cost categories: *Labor* ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{l}$$\\end{document}$), *Supply chain* ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{sc}$$\\end{document}$), *Service delivery* ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{sd}$$\\end{document}$), and *Capital* ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{c}$$\\end{document}$). Each regression equation estimated the cost per dose for a given cost category as a function of the covariates in Table [1](#Tab1){ref-type=\"table\"}:Table 1Model covariatesCovariate nameDescriptionLabor indicator ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${I}_{l}$$\\end{document}$)Labor cost category included in cost per dose\u2009=\u20091; otherwise\u2009=\u20090Supply chain indicator ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${I}_{sc}$$\\end{document}$)Supply chain cost category included in cost per dose\u2009=\u20091; otherwise\u2009=\u20090Service delivery indicator ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${I}_{sd}$$\\end{document}$)Service delivery cost category included in cost per dose\u2009=\u20091; otherwise\u2009=\u20090Capital indicator ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${I}_{c}$$\\end{document}$)Capital cost category included in cost per dose\u2009=\u20091; otherwise\u2009=\u20090YearStudy yearEconCost type: financial\u2009=\u20090; economic\u2009=\u20091; undefined\u2009=\u20092SingleAntigens included: full vaccine program\u2009=\u20090; single antigen\u2009=\u20091log(Doses)Number of doses in the routine vaccination schedule in study year, loggedlog(GDP)GDP per capita in study year, loggedDTP3DTP3 coverage in study yearlog(Pop)Country population size in study year, logged*DTP3* diphtheria-tetanus-pertussis third dose coverage, *GDP* gross domestic product, *Pop* population$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{l}=\\mathrm{exp}\\left({\\beta }_{{0}_{l}}+{\\beta }_{1}*{Year}_{i}+{\\beta }_{2}*{Econ}_{i}+{\\beta }_{3}*{Single}_{i}+{\\beta }_{4}*{\\mathit{log}\\left(Doses\\right)}_{i}+{\\beta }_{5}*{DTP3}_{i}+{\\beta }_{6}*{\\mathit{log}\\left(GDP\\right)}_{i}+{\\beta }_{7}*{\\mathit{log}\\left(Pop\\right)}_{i}\\right)$$\\end{document}$$$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{sc}=\\mathrm{exp}\\left({\\beta }_{{0}_{sc}}+{\\beta }_{1}*{Year}_{i}+{\\beta }_{2}*{Econ}_{i}+{\\beta }_{3}*{Single}_{i}+{\\beta }_{4}*{\\mathit{log}\\left(Doses\\right)}_{i}+{\\beta }_{5}*{DTP3}_{i}+{\\beta }_{6}*{\\mathit{log}\\left(GDP\\right)}_{i}+{\\beta }_{7}*{\\mathit{log}\\left(Pop\\right)}_{i}\\right)$$\\end{document}$$$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{sd}=\\mathrm{exp}\\left({\\beta }_{{0}_{sd}}+{\\beta }_{1}*{Year}_{i}+{\\beta }_{2}*{Econ}_{i}+{\\beta }_{3}*{Single}_{i}+{\\beta }_{4}*{\\mathit{log}\\left(Doses\\right)}_{i}+{\\beta }_{5}*{DTP3}_{i}+{\\beta }_{6}*{\\mathit{log}\\left(GDP\\right)}_{i}+{\\beta }_{7}*{\\mathit{log}\\left(Pop\\right)}_{i}\\right)$$\\end{document}$$$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${\\widehat{c}}_{i}^{c}=\\mathrm{exp}\\left({\\beta }_{{0}_{c}}+{\\beta }_{1}*{Year}_{i}+{\\beta }_{2}*{Econ}_{i}+{\\beta }_{3}*{Single}_{i}+{\\beta }_{4}*{\\mathit{log}\\left(Doses\\right)}_{i}+{\\beta }_{5}*{DTP3}_{i}+{\\beta }_{6}*{\\mathit{log}\\left(GDP\\right)}_{i}+{\\beta }_{7}*{\\mathit{log}\\left(Pop\\right)}_{i}\\right)$$\\end{document}$$\n\nAn additional equation related these cost category unit costs to the combination of cost categories included in each empirical study estimate, with $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${tc}_{i}$$\\end{document}$ representing the mean estimate of the delivery cost per dose for a given study:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${tc}_{i} ={\\widehat{c}}_{i}^{l}*{I}_{l}+{\\widehat{c}}_{i}^{sc}*{I}_{sc}+{\\widehat{c}}_{i}^{sd}*{I}_{sd}+{\\widehat{c}}_{i}^{c}*{I}_{c}$$\\end{document}$$\n\nThese $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${tc}_{i}$$\\end{document}$ values were used to parameterize a Gamma likelihood function for the observed data ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${y}_{i}$$\\end{document}$), where the shape parameter $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\alpha$$\\end{document}$ described the residual variance:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${y}_{i} \\sim Gamma\\left(\\alpha ,\\frac{\\alpha }{{tc}_{i}}\\right)$$\\end{document}$$\n\nThis specification assumed variance proportional to $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${tc}_{i}$$\\end{document}$. We assumed informative prior distributions for all regression coefficients, which were assumed to follow a normal distribution centered at zero with a standard deviation of 1 \\[[@CR15]\\]. The shape parameter $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\alpha$$\\end{document}$ was assumed to follow a half-Cauchy distribution centered at zero with a standard deviation of 5 \\[[@CR16]\\]. The prediction model was estimated in R software \\[[@CR17]\\] using an adaptive Hamiltonian Monte Carlo algorithm using the Stan software package, version 2.20.0, with four chains of 5000 iterations. The first 2500 iterations were discarded (burn-in period), yielding 10,000 posterior draws for analysis \\[[@CR18], [@CR19]\\]. Stan model diagnostics were utilized to determine any problems encountered by the sampler, and the potential scale reduction factor (i.e., Rhat) for all parameters was evaluated to determine that the model had successfully converged.\n\nThe fitted prediction model was used to generate both economic and financial programmatic delivery cost per dose estimates (i.e., as opposed to generating single antigen cost per dose estimates) for each LMIC for 2009--2018. To generate these estimates, we predicted values from the fitted model, with covariates values specific to each country and year. Estimates included all cost categories and are presented in 2018 US dollars, inflated to 2018 values in local currency using the country consumer price index, and converted to 2018 US dollars using market exchange rates \\[[@CR12]\\]. Global and regional cost per dose estimates were calculated as population-weighted averages of individual country estimates. We tested predictive performance by comparing model predictions to the observed cost per dose matched to country and year.\n\nSensitivity Analysis: Alternative Regression Specifications {#Sec5}\n-----------------------------------------------------------\n\nWe estimated several alternative regression specifications as robustness checks. First, we re-estimated the model having excluded two outlier observations with cost per dose estimates less than \\$0.01 as a robustness check. Second, we adopted weakly informative prior distributions, i.e., all predictors were assumed to follow a Normal distribution centered at zero with a standard deviation of 10 instead of 1. Finally, we adopted non-informative prior distributions for parameters.\n\nResults {#Sec6}\n=======\n\nData {#Sec7}\n----\n\nA total of 52 routine delivery cost per dose estimates (i.e., excluding vaccine costs) from 29 studies covering 24 countries were included in the analysis \\[[@CR20]--[@CR48]\\]. For example, a study may have included both economic and financial cost per dose estimates, which were both included in this analysis. Of the 29 studies, 13 were undertaken in low-income country settings and 16 in middle-income country settings, as classified by 2019 World Bank income level \\[[@CR10]\\]. The observed cost per dose ranged from \\$0.66 to \\$9.45 (focusing on studies that included all cost categories). Thirty-four unit cost estimates (65%) could not be disaggregated into separate cost categories; the remaining observations could be disaggregated into unique categories, resulting in a total of 119 unique observations included in the analysis. When examining the total cost per dose by cost categories included, 92% of observations included labor, 98% supply chain, 63% service delivery, and 63% capital. Table [2](#Tab2){ref-type=\"table\"} provides summary information on the empirical studies used in the analysis. For the observed dataset, the mean and standard deviation (in parentheses) of the continuous explanatory variables were 12 (3) for *Doses*, \\$1550 (\\$1300) for *GDP*, 0.88 (0.09) for *DTP3*, and 57,100,000 (207,000,000) for *Pop*.Table 2Summary characteristics for immunization delivery unit cost per dose dataEstimates (*n*)Reported cost per dose\u00a0Total cost per dose only34\u00a0Total\u2009+\u2009cost categories18^a^Income^b^\u00a0Low income32\u00a0Lower middle income16\u00a0Upper middle income4Antigens costed\u00a0Single antigen28\u00a0Full vaccination program24Cost type\u00a0Economic27\u00a0Financial11\u00a0Undefined14^a^There were 52 total cost per dose observations with reported base years between 2001 and 2017, but the 18 cost per dose estimates that could be disaggregated into cost categories brought the total analyzed observations from 52 to 119^b^Low income: gross national income (GNI) per capita of \\$1025 or less; lower middle income: GNI per capita of \\$1026--\\$3995; upper middle income: GNI per capita of \\$3996--\\$12,375 \\[[@CR12]\\]. Costs in US dollars\n\nRegression Model {#Sec8}\n----------------\n\nTable [3](#Tab3){ref-type=\"table\"} reports point estimates and standard errors for regression coefficients and other model parameters. While certain cost category intercepts, the single antigen indicator, and the coefficients on log(*Pop*) and DTP3 were statistically significantly different from zero, the coefficients for other non-intercept predictor variables were not significant.Table 3Results for regressions of routine childhood delivery unit cost per dose on predictorsVariableMean coefficientCost category intercepts\u00a0Labor0.08 (0.23)\u00a0Supply chain\u2009\u2212\u20090.97 (0.26)\u00a0Service delivery\u2009\u2212\u20090.07 (0.28)\u00a0Capital\u2009\u2212\u20091.16 (0.31)Predictors\u00a0Year\u2009\u2212\u20090.14 (0.13)\u00a0Economic cost indicator\u2009\u2212\u20090.06 (0.14)\u00a0Single antigen indicator\u2009\u2212\u20090.49 (0.22)\\*\u00a0log(Doses)0.01 (0.13)\u00a0log(GDP per capita)0.19 (0.13)\u00a0log(Pop)\u2009\u2212\u20090.30 (0.12)\\*\u00a0DTP3 coverage0.29 (0.12)\\*Gamma dispersion parameter\u00a0Alpha1.08 (0.07)Continuous predictors were standardized to mean zero and unit standard deviation; thus, fitted coefficients for continuous variables \\[e.g., log(Doses)\\] represent the increase in log cost per dose observed for a 1.0 standard deviation increase in the variable. Values in parentheses represent standard errors*DTP3* diphtheria-tetanus-pertussis third dose coverage, *GDP* gross domestic product, *Pop* population^\\*^Significant at 5% level\n\nWe assessed in-sample fit by comparing observed versus predicted values for the study sample (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). Figure\u00a0[1](#Fig1){ref-type=\"fig\"} includes six subplots: (A) includes all observations; (B) includes observations in which all four cost categories were included in the total cost per dose; (C) includes observations that contained the labor cost category only; (D) includes observations that contained the supply chain category only; (E) includes observations that contained the service delivery cost category only; and (F) includes observations that contained the capital cost category only.Fig. 1Comparison of predicted cost per dose and published literature cost per dose for routine childhood vaccine delivery. The original costs per dose represent 119 observations across 24 countries for reported base years between 2001 and 2017. The predicted costs per dose are matched to the country and year of each observation\n\nWe calculated first differences to describe the percentage difference in the cost per dose associated with specified changes in individual predictors, holding others fixed. These values represent posterior means, and values in parentheses represent equal-tailed 95% credible intervals. More recent studies reported lower cost per dose estimates, controlling for other covariates. The routine cost per dose was estimated to decline by 4.6% (\u2212\u20094.2%, 13.0%) for each additional calendar year, increase by 0.2% (\u2212\u20093.7%, 3.9%) for each additional dose added to the routine vaccination schedule, increase by 21.1% (\u2212\u20095.3%, 52.5%) when GDP per capita was two times greater, decline by 14.1% (3.0%, 24.0%) when population size was two times greater, and increase by 3.4% (0.7%, 6.0%) for each percentage point increase in DTP3.\n\nEstimated Costs Per Dose for All LMICs {#Sec9}\n--------------------------------------\n\nFor the year 2018, the population-weighted average economic cost per dose for routine delivery of childhood vaccines was estimated to be \\$1.87 (95% uncertainty interval \\$0.64--4.38) across all LMICs. By individual cost category, the programmatic economic cost per dose for routine delivery of childhood vaccines was \\$0.74 (\\$0.26--1.70) for labor, \\$0.26 (\\$0.08--0.67) for supply chain, \\$0.22 (\\$0.06--0.57) for capital, and \\$0.65 (\\$0.20--1.66) for other service delivery costs. By income level, the average predicted programmatic, economic cost per dose was \\$1.41 (\\$0.52--3.16) for low-income countries, \\$1.36 (\\$0.44--3.32) for lower middle-income countries, and \\$2.59 (\\$0.82--6.38) for upper middle-income countries. Figure\u00a0[2](#Fig2){ref-type=\"fig\"} presents the country-level cost per dose estimates by GDP per capita and World Bank income level for 136 LMICs. Table [4](#Tab4){ref-type=\"table\"} presents the programmatic, economic cost per dose by each stratification by world region and income level.Fig. 2Predicted economic cost per dose in 2018 for routine childhood vaccine delivery by GDP per capita and World Bank income level for 136 LIC and middle-income countries. *GDP* gross domestic product, *LIC* low-income countries, *LMIC* lower middle-income countries, *UMIC* upper middle-income countriesTable 4Predicted economic cost per dose (US\\$) in 2018 for routine childhood vaccine delivery by cost category and world region/income levelTotal cost per doseLabor cost per doseSupply chain cost per doseService delivery cost per doseCapital cost per doseRegion\u00a0Africa\\$1.49 (\\$0.57--3.31)\\$0.59 (\\$0.23--1.28)\\$0.21 (\\$0.07--0.49)\\$0.52 (\\$0.17--1.23)\\$0.17 (\\$0.06--0.43)\u00a0Americas\\$2.61 (\\$0.87--6.33)\\$1.02 (\\$0.35--2.42)\\$0.37 (\\$0.11--0.94)\\$0.91 (\\$0.27--2.44)\\$0.31 (\\$0.09--0.81)\u00a0Eastern Mediterranean\\$1.86 (\\$0.71--4.05)\\$0.73 (\\$0.29--1.56)\\$0.26 (\\$0.09--0.62)\\$0.65 (\\$0.22--1.52)\\$0.22 (\\$0.07--0.52)\u00a0Europe\\$3.51 (\\$1.21--8.18)\\$1.38 (\\$0.50--3.19)\\$0.50 (\\$0.15--1.25)\\$1.22 (\\$0.38--3.06)\\$0.41 (\\$0.12--1.07)\u00a0Southeast Asia\\$1.35 (\\$0.40--3.48)\\$0.53 (\\$0.16--1.34)\\$0.19 (\\$0.05--0.52)\\$0.47 (\\$0.13--1.26)\\$0.16 (\\$0.04--0.45)\u00a0Western Pacific\\$2.07 (\\$0.59--5.47)\\$0.81 (\\$0.24--2.10)\\$0.29 (\\$0.07--0.81)\\$0.73 (\\$0.18--2.09)\\$0.24 (\\$0.06--0.69)Income level\u00a0Low income\\$1.41 (\\$0.52--3.16)\\$0.56 (\\$0.20--1.24)\\$0.20 (\\$0.07--0.47)\\$0.49 (\\$0.16--1.17)\\$0.16 (\\$0.05--0.41)\u00a0Lower middle income\\$1.36 (\\$0.44--3.32)\\$0.54 (\\$0.18--1.28)\\$0.19 (\\$0.06--0.50)\\$0.47 (\\$0.14--1.21)\\$0.16 (\\$0.04--0.43)\u00a0Upper middle income\\$2.59 (\\$0.82--6.38)\\$1.01 (\\$0.34--2.44)\\$0.37 (\\$0.10--0.95)\\$0.91 (\\$0.25--2.39)\\$0.30 (\\$0.08--0.81)Countries included in each World Health Organization (WHO) region are low- and middle-income countries according to World Bank income level in 2019 \\[[@CR12]\\]\n\nThe set of country-specific economic cost estimates for delivery cost per dose can be found in Appendix Table A (see the \"Supplementary Appendix\" in the electronic supplementary material). Figure\u00a0[3](#Fig3){ref-type=\"fig\"} presents the predicted programmatic, economic cost per dose for childhood vaccine delivery by year for six example countries selected for differences in region and income level. The predicted cost per dose shows a decreasing trend on average over time. Within individual cost categories, financial cost per dose estimates did not differ substantially from the economic cost per dose predictions; however, financial cost observations generally reported fewer cost categories, notably capital. The set of country-specific financial cost estimates for delivery cost per dose can be found in Appendix Table B. Appendix Table C further provides the median and interquartile range globally, and by region and income level.Fig. 3Predicted programmatic, economic cost per dose for routine childhood vaccine delivery by year. Armenia\u2009=\u2009European region, lower middle income; Haiti\u2009=\u2009region of the Americas, low income; Lebanon\u2009=\u2009Eastern Mediterranean region, upper middle income; Tanzania\u2009=\u2009African region, low income; Thailand\u2009=\u2009Southeast Asian region, upper middle income; Vietnam\u2009=\u2009Western Pacific region, lower middle income\n\nSensitivity Analysis: Alternative Regression Specifications {#Sec10}\n-----------------------------------------------------------\n\nWith the first alternative regression specification, excluding two outliers produced small changes in most coefficients and large changes in a few (Appendix Table D) and resulted in cost per dose estimates that were 16% higher on average (Appendix Table E). Adopting weakly informative (Appendix Table F) or non-informative (Appendix Table G) priors did not substantially change regression results (i.e., less than 10% average change in cost per dose estimates).\n\nDiscussion {#Sec11}\n==========\n\nFor the year 2018, the average economic cost per dose across all LMICs for routine delivery of childhood vaccines was estimated to be \\$1.87 (95% uncertainty interval \\$0.64--4.38), excluding vaccine costs. By country income classification, the average cost per dose was \\$1.41 (\\$0.52--3.16) for low-income countries, \\$1.36 (\\$0.44--3.32) for lower middle-income countries, and \\$2.59 (\\$0.82--6.38) for upper middle-income countries. These estimates are consistent with the empirical estimates reported in the ICAN IDCC \\[[@CR11]\\], where studies including all cost categories averaged \\$1.41 for low-income countries and \\$4.02 for lower middle-income countries (there was only one empirical estimate, \\$1.87, for upper middle-income countries). While the population-weighted average predicted cost per dose estimates were similar for low- and lower middle-income countries, Fig.\u00a0[2](#Fig2){ref-type=\"fig\"} indicates that costs were predicted to be higher for richer countries overall. Costs were also predicted to be lower for more populous countries (e.g., Ethiopia and Nigeria), compared to similar countries with smaller populations.\n\nThese predicted cost per dose estimates can be useful for cost-effectiveness analyses when country-level costs are unavailable, highly uncertain, or old. For example, instead of using neighboring country data or regional data when primary cost data are unavailable---or creating estimates based on expert opinion alone---these modeled costs provide another alternative to use, based on a broader set of data. While recent costing reference cases provide concrete guidance for implementing and reporting costing studies \\[[@CR49], [@CR50]\\], we are unlikely to have the resources to conduct empirical costing and/or cost-effectiveness analysis for all questions and settings of interest. Therefore, a key strategy for improving the availability of costing data is to determine how and where we can leverage insights from specific studies to understand a general theme. Country-specific costs modeled within a Bayesian meta-regression framework provide a broad indication of immunization delivery costs that may be useful when accurate local data are unavailable.\n\nThe regression results showed several relationships that might be expected between predictors and immunization delivery costs. The statistically significant relationships were population size and DTP3; higher population size (a proxy for higher service volume at the site level) was associated with lower cost per dose, while DTP3 (a proxy for overall coverage of the immunization program) was associated with higher cost per dose. While the reason for the significant relationship with population size is unclear, it could be related to economics of scale, i.e., the increasing scale of an immunization program results in cost savings through both efficiency gains and spreading fixed costs over a larger population. Additionally, we did not investigate alternative measures of service volume, such as under-five population, which may impact the results. However, we expect that the direction of this relationship is likely to remain the same, as the higher service volume to lower delivery costs relationship has also been found in previous studies \\[[@CR8]\\]. Higher DTP3 being associated with higher cost per dose may be due to increasing marginal costs with higher vaccine coverage levels, or some other feature of higher-coverage programs that leads to higher costs. Greater GDP per capita (a proxy for country price levels) was also associated with higher cost per dose. Additionally, we found no relationship between number of doses in the routine immunization schedule and unit cost, although we might expect to observe a lower cost per dose due to economies of scale. We also found the sign estimated for some coefficients was different to what might have been expected (e.g., the negative signs on *Year* and the economic cost indicator, *Econ*). If interpreted directly, these findings would suggest lower costs (in real terms) in more recent years, holding other covariates constant, and economic costs which are lower than financial cost estimates. However, the statistical precision of these two findings (as expressed by the *p* values) is low, and consequently little weight should be placed on the point estimates. In addition, the direction and magnitude of the coefficients in the model could be driven by unobserved characteristics of the included studies (e.g., the costs of inputs or the data collection approach) that are correlated with both the predictor variables and the unit cost. For this reason, all of these relationships should be viewed as correlations with the predicted cost per dose and should not be interpreted causally.\n\nThere are several limitations to this analysis. First, the studies we included in the analysis were heterogeneous in terms of the scope of costing, site selection, data collection methods, and the level of detail with which results were reported. This heterogeneity in costing study methods and reporting has been observed by prior reviews \\[[@CR51]--[@CR55]\\]. We attempted to deal with this by adopting an analytic strategy that allowed for differences in the cost categories and types of costs (i.e., financial vs. economic) reported. However, it is unlikely our approach fully reconciled all methodological differences between studies. For this reason, the residual variance of the meta-regression model---which was substantial---will reflect not only sampling uncertainty but also non-sampling error due to methodological heterogeneity, such as inconsistent definitions of cost categories and cost types. This could induce omitted-variable bias, if methodological differences were correlated with regression variables, or if cost estimates were biased systematically. For example, it has been observed that costing studies may over- or underestimate costs due to the costing approach used (i.e., gross costing vs. micro-costing \\[[@CR56]\\]) or may underestimate costs due to the exclusion of relevant intervention cost categories \\[[@CR57]\\]. In our analysis, we saw that excluding outliers changed the magnitude, although not the direction, of several regression coefficients, as shown in the results of Appendix Table C. All included studies were reported from the immunization provider perspective, and excluded caregiver/beneficiary time and transportation costs to receive vaccination. As the definition of this provider perspective may have varied across studies, we defined cost categories to include the same elements in order to improve comparability. However, for categories that included both recurrent and investment costs, the treatment of investment costs (e.g., useful life estimates, annuitization) may differ between individual costing studies. A second limitation stems from the assumption that routine childhood vaccine delivery costs are similar regardless of individual vaccine product. This assumption was necessitated by the format in which data were reported, yet in reality there may be differences in delivery costs, particularly in the case of injection versus oral vaccines that involve differences in training and delivery, or where multiple doses are delivered in a single immunization visit. Furthermore, as routine immunization schedules were not available for all historical years, country-specific time trends in the number of doses in the immunization schedule (represented by the *Doses* variable) may be less accurate for earlier years. A third limitation relates to sample size. While we were able to include a large number of studies in this meta-regression analysis, the average number of observations within each of these studies (i.e., number of sites) was small. A costing study that relies on a small number of sampled sites may produce results that are not representative, if there is likely to be large variation across sites \\[[@CR58]\\]. In addition, the 24 countries included in the dataset were not a representative sample of all LMICs. The countries represented by the sampled studies were generally of lower GDP per capita (average \\$1600 vs. \\$4000) and higher DTP3 (average coverage 88% vs. 85%) compared to the 136 LMICs for which we provide modeled estimates. While we adjusted for these factors in the meta-regression model, the imperfect overlap of sample and population adds to the uncertainty in country-level unit costs estimated for countries with higher GDP per capita, lower population size, and lower DTP3. Therefore, estimates will be less reliable for countries with combinations of covariates not included in the analyzed sample, as shown by the estimates for countries such as Tuvalu and Palau (Appendix Table A).\n\nIn light of the issues described above, this analysis does not fully resolve the evidence gap created by the limited number of immunization costing studies that are published (limited relative to the number of settings and policy questions where they would be useful), and the estimates we report inherit many of the limitations of the empirical studies they are based on. Additional primary data collection on delivery costs of programs is necessary. However, the need to make policy choices based on imperfect information is unavoidable, and the estimates we report provide an additional evidence source for analysts missing this important input to their analysis.\n\nImmunization delivery costs are a necessary category of high-quality cost-effectiveness models, and are also used to inform resource mobilization for immunization programs. Using observed costs per dose from 29 studies in 24 countries, our study provides estimates for 136 LMICs for the years 2009--2018 produced via meta-regression analyses that can help improve resource mobilization and planning in situations where empirical cost data are unavailable or of low quality. This methodology has the potential to be applied to many other areas of health care for LMICs in which it is unlikely that primary data cost studies would be conducted for all settings/services of interest.\n\nElectronic supplementary material\n=================================\n\n {#Sec13}\n\nBelow is the link to the electronic supplementary material.Supplementary file1 (DOCX 44 kb)\n\nEarlier versions of this paper were presented at the 2019 International Health Economics Association World Congress, and we received valuable comments from conference participants. We also received valuable comments from Jane Kim, Annette \u00d6zaltin, and St\u00e9phane Verguet.\n\nAP and NM conceptualized and designed the study. AP did the analysis, drafted the initial manuscript, and approved the final manuscript as submitted. KV, ECD, CS, SCR, LB, and NM developed the methodology, critically reviewed the analysis, reviewed and revised the manuscript, and approved the final manuscript as submitted.\n\nThe data that support the findings of this study are openly available in the Immunization Delivery Cost Catalogue (IDCC) at <https://immunizationeconomics.org/ican-idcc/>, reference number \\[[@CR11]\\].\n\nThis work was supported by the Bill & Melinda Gates Foundation (OPP1158709).\n\nThe authors, Allison Portnoy, Kelsey Vaughan, Emma Clarke-Deelder, Christian Suharlim, Stephen C. Resch, Logan Brenzel, and Nicolas A. Menzies, declare that they have no competing interests.\n\nThe programming code is available from the corresponding author on reasonable request.\n"}
+{"text": "S\u00e3o Paulo, November 17, 2014\n\nDear Editor\n\nThe fine structure of apoptotic HeLa cells from cultures contaminated with mycoplasma in early and in advanced stages of the cell demise process differs from those so far described in apoptotic cells. The observed changes are enhanced after exposure of the cells to staurosporine. At low microscopic magnifications cells that have apparent normal cytoplasm and nuclei, actually may be harbouring cystic-like profile(s) of parasitic origin in an altered cytoplasm. The membranes of the transitional elements of the endoplasmic reticulum (TER) appear fragmented in irregular branching stripes of the smooth component of the TER ([Fig. 1](#f01){ref-type=\"fig\"}, white asterisks in L delimited area). The concentration of the rough endoplasmic reticulum (RER) membranes is less than in normal HeLa cells. Near to the smooth ER tubule-saccular elements lie groups of 50 nm microvesicles aside stacked, thin, various sized profiles of Golgi saccules ( \\] ). The 50 nm microvesicles bud off mainly from the periphery of the stacked Golgi elements ([Fig 1](#f01){ref-type=\"fig\"} thin arrow heads inside line U) and also from the extremities of smooth ER tubules ([Fig. 1](#f01){ref-type=\"fig\"} small arrows). Small groups of compacted microvesicles are noted in cells still maintaining normal nuclear appearance (not shown). With the start of chromatin condensation progressively larger compact microvesicular clusters are formed. These attain sizes larger (Fig. 2) than those of the clusters of microvesicles derived from the fragmentation of Golgi saccules seen in mitotic (LUCOCQ *et al.*, 1989; SESSO *et al.*, 1999) and in apoptotic (SESSO *et al.*, 1999) cells. Contemporaneously two major cytoplasmic alterations may be noted in contaminated cells namely when treated with staurosporine. Occasionally, both deformations appear in the same cell. One, is progressive cytoplasmic loss by formation at the cell periphery of blebs that separate from the inner cytoplasm (Figs. 3 and 4) or by localized detachment of sectors of the peripheral cytoplasm with various forms and sizes (not shown). In some cells, the remainder thin, cytoplasm with few mitochondria and rough ER profiles surrounds the nucleus in a ring-like form. Such small cells are noted in heavily contaminated samples. Some of the cells exhibit sectors of the cytoplasm with a reticulated appearance. Such net-like regions are composed by various sized tubular and ellipsoidal, apparently empty profiles. It is unclear if the smooth membranes that compose these regions with reticulated aspect, may have derived from the Golgi apparatus. The shape and size of the empty spaces correspond to those from villus-like formations seen free close to and emerging from the cell surface. In contaminated cells namely after staurosporine treatment the free villus - like forms are seen sprouting from the cell surface and also free nearby. Spheroidal 50-100nm (thin arrow in Fig. 5) profiles with inner structure identical to that of villus-like elements are consistently proximate to the fake villi. Vestiges of what can be remnants of the villus-simile structures and/or of the parasite itself are seen in these spaces (Figs. 6 and 7).\n\nFigs. 1**7** are a synopsis of fine structural observations in some 50 samples of staurosporine (0.1 \u00b5M - 2.0 \u00b5M for 2 - 24h) and 10 control samples of HeLa cell cultures contaminated with mycoplasma. [Fig. 1](#f01){ref-type=\"fig\"} is from an apparent normal cell at low microscopic exam. Figs 2, 6 and 7 and 3 and 4 are from apoptotic cells exposed to staurosporine 0.5 \u00b5M/3h and 0.1 \u00b5M/6h, respectively. Fig. 5 is from a non apoptotic cell exposed to staurosporine 0.1 \u00b5M/4h. Bars -1 \u00b5M. [**Fig. 1**](#f01){ref-type=\"fig\"} - HeLa cell-from a heavily contaminated cell culture devoid of staurosporine where apoptosis occurred. Normal nucleus (N). Smooth, branched ER profiles (extremities of thin white lines departing from the central white asterisks in the area delimited by the lower (L) dotted line. Conglomerates of 50 nm (lower and upper right part of the figure). Part of the microvesicles bud off from smooth ER membranes (arrow head) and from tubular elements (small arrows). The majority of the grouped microvesicles pinch off from the extremities of stacked Golgi saccules (thin arrow heads at the periphery of the frontally exposed dense ellipsoidal-like profile delimited by U)\\]. Stacked, thin Golgi saccular profiles of various lengths that often appear dense ( \\] ). **Fig. 2** - Large, compact conglomeration of various sized microvesicles with predominance of the ones with 50 nm. Groups of former dilated stacked Golgi saccules (black asterisks) are interspersed among the microvesicles. A large such dilated saccule contains remnants of a parasite or of material of parasitic origin (MPO) (empty triangle). **Fig. 3** - The cytoplasm of this apoptotic cell is delimited in two concentric major regions. The peripheral one is partitioned into adjacent blebs containing predominantly membranes of the ER. The inner part of the cytoplasm contains clustered, swollen mitochondria. **Fig. 4** - Dismantling of the peripheral apoptotic cytoplasm by a contemporaneous detachment of the previously formed bleb regions. **Fig. 5** - Non apoptotic infected (cysts of parasitic origin, upper arrows) cell, with an abnormally elongated thin cytoplasm. The free villus-like structures often appear curved aside spheroidal 50-100 nm (thin arrow) elements. They assemble in various degrees in the sectors of the contaminated cells apoptotic or not undergoing progressive dismantling of the peripheral cytoplasm. **Fig. 6** - Most of the cytoplasm above the apoptotic nucleus is occupied by membrane bound void spaces of various sizes and forms. The mitochondria are unusually dense, a common occurrence in contaminated cells, apoptotic or not. Adherent to the limiting membrane of the space indicated (open triangle) a fluffy material possibly derived from MPO. **Fig. 7** - Sector of an apoptotic cytoplasm with net-like membranal arrangement as in Fig. 6. Compact mass of parasitic material in a cystic-like form (open triangle). The upper, middle and vertical white lines indicate remnants of parasitic origin that were not completely removed by the processing of the cells. Part of an elongate membrane bound space possibly still fully occupied by MPO is indicated by the lower line that branches in two.\n\nIt is yet undetermined whether the early structural changes expressed by foci of assembled microvesicles at the transitional endoplasmic reticulum-Golgi interface is an exclusive type of membranal alteration preceding overt apoptosis in mycoplasma infected cells.\n\nAll cytoplasmic membrane bound organelles as peroxisomes, lysosome-endosomes and the Golgi apparatus derive from microvesicles that bud off from the ER. The ER is also mobilized by promoters of cellular stress (references in DOLAI & ADAK, 2014). The here described structural deviation of the ER-Golgi interface from the normal condition may represent more than only a mycoplasma induced alteration of the programmed cell death mechanism. It is speculative, whether this initial accumulation of microvesicles in cells with typical normal nuclei is part of a general forewarning mechanism of cell defence. In a less intense cell stress than that occurring here the observed changes of the TER could eventually pass undetected under the transmission electron microscope.\n\nThe mycoplasmas fine structures of our samples are identical (NIR-PAZ *et al.*, 2002; KORNSPAN *et al.*, 2010;) and similar (EDWARDS & FOGH, 1960; HUMMELER *et al.*, 1965; TAYLOR-ROBINSON *et al.*, 1991) to those from various mycoplasmas strains seen in cultures and in infected cells.\n"}
+{"text": "Problem {#sec0001}\n=======\n\nThe COVID-19 pandemic has created difficulties with execution of traditional didactic curricula. Programs have restricted in-person gatherings and increased use of work-from-home models. Essential staff involved in execution of didactics have also been subject to furlough. The pandemic has directly affected facilitator health and time for preparing and delivering didactics; this impact has been highly variable across regions and institutions.\n\nDespite these challenges, it is imperative that programs continue to deliver quality didactic education. The Accreditation Council for Graduate Medical Education (ACGME) has issued statements indicating that \"programs should continue to provide education to residents/fellows, when feasible, utilizing remote conferencing technology, web-based resources, and other innovative tools.[@bib0001]\" Program directors need innovative approaches to ensure that didactics are held regularly, are high quality, and engaging to learners at a time of significant stress.\n\n\"Megaflip\" is a novel form of open-access national didactics. These sessions are multicenter educational experiences using materials from The Organization of Neonatal-Perinatal Medicine Training Program Directors National Neonatology Curriculum, a standardized, peer-reviewed curriculum utilizing flipped classrooms.[@bib0002] In the National Neonatology Curriculum flipped classrooms model, learners are assigned prework, consisting of short videos and reading materials in preparation for the \"classroom\" session. The classroom session focuses on case-based knowledge application rather than knowledge acquisition.\n\n\"Megaflip\" sessions are delivered via the Zoom (San Jose, Calif) platform. A small group of facilitators (2--3) hosts the Megaflip and interested programs request a link to participate. The participating programs provide a local facilitator to lead their own fellows in large group and breakout discussions. This facet of \"Megaflip\" allows the faculty team to highlight important interinstitutional practice differences and expands the peer-learning opportunity across programs. The Zoom platform overcomes many of the challenges of a large group interactive session by allowing learners to virtually raise their hand, answer yes/no questions with the click of a button, utilize the chat feature, and participate in smaller group discussions via the breakout room feature.\n\nOutcomes to Date {#sec0002}\n================\n\nThere have been 2 pilot sessions with 131 fellows from 16 neonatology and pulmonary fellowship programs across the United States. We have thus far received 22 survey responses regarding the utility of Megaflips and the impact of COVID-19 on fellow education.\n\nPrior to their Megaflip experience, only 4.5% of respondents had previously participated in didactics with fellows from other programs. Seventy-three percent felt the amount of active discussion was better than a typical educational session and 91% of respondents felt this model had good or great effectiveness for learning. No fellows reported technical difficulties with the online platform.\n\nNext Steps {#sec0003}\n==========\n\nAs expertise and comfort with the online format grows, we anticipate that more programs will host and facilitate Megaflip sessions. Expanding the use of Megaflips on a national level will allow increased education across medical disciplines and content areas. We plan to continue this model beyond the pandemic given the opportunity for multicenter collaboration, reduced burden on faculty facilitators, and expanding fellow exposure to center-specific innovations and practice patterns.\n\nThe authors have no conflicts of interest to disclose.\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nHistory of hypothermia dates back to Hippocratic era and in modern history to the 1950\\'s when elective hypothermia (28--32\u00b0C) was practiced during general anesthesia for brain and heart protection.\\[[@ref1]\\] In 1960\\'s Peter Safar, father of modern cardio-pulmonary resuscitation (CPR) started using hypothermia, in post-cardiac arrest scenarios.\\[[@ref1]\\] The initial enthusiasm in hypothermia was followed by subdued ebb in time zone of 1960--1980, due to complications arising from hypothermia, that is, patients were cooled longer and deeper which resulted in bleeding and septic complications. From 1980 to 2000 a lot of work on experimental animals showed that hypothermia leads to neuro-protection.\\[[@ref2]\\] Two randomized controlled trials (RCT) which were published in 2002 showed the benefit of hypothermia in post-cardiac arrest situations, which changed the world regarding the use of hypothermia.\\[[@ref3][@ref4]\\] Therapeutic hypothermia (TM) is now-a-days popularly known as targeted temperature management (TTM). TTM is a unique modern era therapeutic modality for salvaging neurological tissue viability in critically ill patients.\n\nDefinitions {#sec1-2}\n===========\n\nThe normal body temperature in healthy individuals (measured in the oral cavity) is 36.8\u00b0C \u00b1 0.4\u00b0C, with normal diurnal variations of 0.5\u00b0C.\\[[@ref5]\\] Rectal temperatures are usually 0.4\u00b0C higher than oral readings.\\[[@ref5]\\] The temperature of the blood measured with aid of pulmonary artery catheter is the accepted gold standard for \"true\" core temperature.\\[[@ref6]\\] The lower esophageal temperature is the most rapid and accurate, noninvasive method of measurement of core temperature and is close to gold standard. The rectal temperature and bladder temperature also closely reflect core temperature. Clinically, tympanic temperature, which measures radiating heat from the tympanic membrane, is often used as a surrogate for deep brain temperature.\\[[@ref5]\\]\n\n {#sec2-1}\n\n### Hypothermia {#sec3-1}\n\nHypothermia is defined as core temperature of \\<36.0\u00b0C.\\[[@ref5]\\] This can be further classified as mild, moderate or severe. Temperature\\'s range of 33--36\u00b0C is referred as mild hypothermia, moderate hypothermia is temperature in range of 28--32\u00b0C and deep hypothermia is temperature \\<28\u00b0C.\n\n### Induced hypothermia {#sec3-2}\n\nIt is reduction of a patient\\'s core body temperature below 36.0\u00b0C which is intentional.\\[[@ref5]\\]\n\n### Therapeutic hypothermia {#sec3-3}\n\nIt is induced hypothermia with side effects such as shivering, being controlled or suppressed.\\[[@ref5]\\] This term therapeutic hypothermia is incomplete as it does not include interventions intended to maintain temperature near 37\u00b0C. The new term TTM was coined which covered all the deficiencies in the above term. It comprises of three distinct phases:\\[[@ref7]\\]\n\nInduction phase\n\nThis phase is characterized by change from the current temperature to a lower temperature which is the desired target temperature. This phase lasts for 60--80 min as cooling has to be initiated rapidly.\n\nMaintenance phase\n\nIn this phase patient is kept at desired target temperature for a prolonged period of time. This phase lasts for 24--28 h with minimal or no fluctuations in temperature.\n\nRewarming phase\n\nIn this patient is slowly rewarmed to near normal temperature range from target temperature. The rewarming has to be done very slowly with side effects of vasodilatation caused by rewarming on various organ systems. In patients with traumatic brain injury (TBI), rate of rewarming should be very slow, that is, 0.1--0.2\u00b0C h and 0.25\u00b0C/h in patients with postcardiac arrest.\n\nThe side-effects of each phase of TTM are described in [Table 1](#T1){ref-type=\"table\"}.\n\n###### \n\nSide effects as per phases of TTM\n\n![](IJCCM-19-537-g001)\n\nMethods of Cooling {#sec1-3}\n==================\n\nConvection, conduction, radiation, and evaporation are four common mechanisms of heat loss. All methods of cooling basically aim at increasing convective or conductive heat loss. The convective heat loss is most common of all the mechanisms as it accounts for 20--30% of heat loss. A list of cooling methods is illustrated in [Figure 1](#F1){ref-type=\"fig\"}.\n\n![Methods of cooling](IJCCM-19-537-g002){#F1}\n\nMechanism of Action of Hypothermia {#sec1-4}\n==================================\n\nWhenever there is neuronal injury either directly or indirectly it leads to increased free oxygen radicals formation, excess glutamate release, ion pump dysfunction ultimately leading to excess calcium influx and excess leakage through capillaries leading to cytotoxic edema. Hypothermia leads to neuroprotection basically by blunting all the four mechanisms.\\[[@ref8]\\] Apart from these for every 1\u00b0C fall in temperature, the cerebral metabolism falls by 8%, which in turn leads to decrease in cerebral metabolic requirement of oxygen (CMRO~2~) and hence decreased cerebral blood flow.\n\nPhysiology of Thermoregulation {#sec1-5}\n==============================\n\nThermoregulation is an important aspect of human homeostasis. Humans have been able to adapt to a great diversity of climates. High temperatures pose serious stresses for the human body, placing it in great danger of injury or even death. If the heat production is greater than heat loss; the heat concentration will increase and the body temperature will rise. If the heat production is less than heat loss; the heat concentration will decrease and the body temperature will drop. In hot conditions; the sweat glands under the skin secrete sweat, which travels up the sweat duct, through the sweat pore and onto the surface of the skin. This causes heat loss by evaporation; however, a lot of essential water is lost. The tiny muscles under the surface of the skin called erector pili muscles, relaxes so follicles are not erect. These flat hairs increase the flow of air next to the skin increasing heat loss by convection. The arterioles vasodilatation occurs; this is the process of relaxation of smooth muscle in arteriole walls allowing increased blood flow through the artery. The flow of blood, into the superficial capillaries in the skin increasing heat loss by convection and conduction. In contrast in cold conditions, the sweat stops being produced. The arterioles, carrying blood from superficial capillaries under the surface of the skin towards the warmer core of the body vaso-constrict. The muscles also receive messages from the thermoregulatory center of the brain (the hypothalamus) to cause shivering.\n\nEffects of Hypothermia on Various Organ Systems {#sec1-6}\n===============================================\n\n {#sec2-2}\n\n### Central nervous system {#sec3-4}\n\nHypothermia is known for its neuro-protective action; for every 1\u00b0C fall in temperature, cerebral metabolism falls by 8%.\\[[@ref8]\\] The most problematic of all the side-effects is shivering. The shivering in turn leads to increased cerebral metabolism, which in turn leads to increased cerebral blood flow and thus increase in intracranial pressure (ICP). Shivering starts at 35.5\u00b0C and stops at 33.5\u00b0C. The shivering is easy to control in patients on mechanical ventilation but is problematic to control in spontaneously breathing patients. A range of drugs is available to control shivering viz., pethidine, fentanyl, dexamethasone etc.\\[[@ref5]\\]\n\n### Cardiovascular system {#sec3-5}\n\nHypothermia has varied effect on cardiovascular system.\n\nHeart rate\n\nThe heart rate decreases with fall in temperature. The decrease in heart rate leads to decrease in myocardial oxygen demand hence myocardial contractility increases, as measured by systolic function. As temperature decreases below 35.5\u00b0C, sinus bradycardia occurs and as core temperatures falls below 32\u00b0C, the heart rate further decreases to around 40--45 beats/min.\\[[@ref5]\\] The increase in heart rate artificially through administration of chronotropic drugs or pacing wire, leads to decrease in myocardial contractility.\\[[@ref9]\\]\n\nCardiac output\n\nThe decrease in heart rate during mild to moderate hypothermia decreases cardiac output (CO) by 25--40%.\\[[@ref5]\\] The decrease in metabolic rate is equal to or greater than the decrease in CO.\n\nSystemic vascular resistance\n\nThe hypothermia induced vasoconstriction of peripheral arteries or arteriole leads to an increase systemic vascular resistance (SVR), which in turn leads to rise in blood pressure (by \u226510 mm Hg). This vasoconstrictive effect is absent or negligible in cerebral circulation.\\[[@ref10][@ref11][@ref12][@ref13][@ref14]\\] The increase in SVR could lead to increased afterload of injured heart but patients who have reperfusion following cardiac arrest develop a systemic inflammatory response syndrome. In this condition an increase in SVR due to vasoconstriction will be beneficial, and will also increase coronary perfusion.\n\nBlood pressure\n\nThe blood pressure is the product of CO and SVR.\\[[@ref5]\\] A fall in heart rate leads to decrease in CO whereas hypothermia induced vasoconstriction leads to increased SVR. The net rise in blood pressure (by \u226510 mm Hg) is due to more pronounced effect of hypothermia induced increase in SVR.\\[[@ref15][@ref16]\\]\n\nCentral venous pressure\n\nThe central venous pressure which symbolizes increased preload due to vasoconstriction of peripheral arterioles and arteries as a result of hypothermia.\\[[@ref15][@ref16]\\]\n\nElectrocardiographic changes\n\nCharacterized by prolonged PR interval, widening of the QRS complex, increased QT interval, finally resulting into Osborne waves.\\[[@ref5]\\] Electrocardiographic changes, as described above, do not require treatment, and at a temperature of 32\u00b0C the heart rate of 40/min is perfectly normal. The increase in heart rate artificially through the administration of chronotropic drugs or pacing wire, leads to decrease in myocardial contractility.\\[[@ref9]\\] Atropine is ineffective in this condition. The risk of arrhythmia is low as long as the core temperature is above 30\u00b0C. If the temperature is brought below 30\u00b0C then the risks of arrhythmia increases. Atrial fibrillation (AF) is most common arrhythmia seen in moderate to deep hypothermia, but this rhythm can change to ventricular tachycardia (VT) or ventricular fibrillation (VF) if temperature is allowed to go below 28\u00b0C.\\[[@ref5]\\] If temperature is below 28\u00b0C and chest compressions are performed then it can easily be converted from AF to VF.\\[[@ref17]\\] At temperature below 28\u00b0C myocardium is least sensitive to antiarrhythmic drugs and to defibrillation. These problems are least encountered above 30\u00b0C, hence target should be keep the temperature above that.\\[[@ref18][@ref19][@ref20][@ref21]\\]\n\n### Volume status {#sec3-6}\n\nHypovolemia is a common feature which occurs at the time of initiation of hypothermia and is multifactorial. Hypothermia leads to increased venous return, which in turn leads to decreased release of antidiuretic hormone along with increased release of atrial natriuretic factor thus contributing to cold diuresis.\\[[@ref22][@ref23][@ref24]\\] This is accompanied by tubular dysfunction, which also leads to volume loss and thus hypovolemia. Thus careful attention at time of initiation of hypothermia should be paid to fluid balance and volume status of these patients.\n\n### Metabolic changes {#sec3-7}\n\nAt the time of initiation of hypothermia, hyperglycemia occurs due to decrease insulin sensitivity and reduced insulin secretion by pancreatic islet cells.\\[[@ref25]\\] Similarly during induction phase, loss of volume and intracellular shifts lead to loss of electrolytes leading to hypokalemia, hyponatremia along with hypomagnesaemia. During rewarming phase the insulin gets functional and thus leads to hypoglycemia. There is also shift of electrolytes from intracellular compartment to extracellular compartment during rewarming phase and thus leads to rise in K^+^ levels. Strict monitoring of sugar and electrolytes is warranted.\n\n### Gastrointestinal {#sec3-8}\n\nTargeted temperature management decreases gastrointestinal motility and eventually patients will require pro-kinetics to avoid delays in enteral feeding. Serum amylase and liver enzymes are commonly raised. The metabolic acidosis occurs as a result of increased production of free fatty acids, ketones and glycerol and increase in lactate concentrations.\\[[@ref26]\\]\n\n### Ventilation {#sec3-9}\n\nAs hypothermia leads to decreased metabolic demand, this in turn leads to decreased carbon dioxide production. Normal mechanical minute ventilation at this stage leads to hypocapnia and cerebral vasoconstriction which will further aggravate brain injury.\\[[@ref5]\\] The minute ventilation should be decreased to avoid developing hypocapnia during induction phase. Another point to consider is that value of blood gases are temperature dependent. The analyzers warm blood to 37\u00b0C before analysis and patient is actually hypothermic so both PaO~2~ and PaCO~2~ are overestimated.\n\n### Coagulation abnormalities {#sec3-10}\n\nCoagulation abnormalities are witnessed at temperature \\<33\u00b0C whereas platelet dysfunction are noted at temperature \\<35\u00b0C.\\[[@ref27][@ref28][@ref29][@ref30][@ref31][@ref32][@ref33]\\] The coagulation cascade is dependent on enzymes; due to hypothermia the enzymes are inactivated which in turn affects the coagulation cascade.\\[[@ref5]\\] The risk of clinically significant bleeding is too low despite of coagulation defects.\n\nCurrent Evidence and Practices in Critical Care {#sec1-7}\n===============================================\n\nWe used the search titled \"therapeutic hypothermia\" in PubMed search engine and applied filters: RCT, systematic reviews and meta-analyses in English literature and human subjects 1192 articles were retrieved. A detailed summary is described in [Figure 2](#F2){ref-type=\"fig\"}.\n\n![Schematic presentation of literature search strategy .TBI\\*: Traumatic brain injury, HIE\\*: Hypoxic ischemic encephalopthy](IJCCM-19-537-g003){#F2}\n\nThe inclusion criteria were conditions for which hypothermia is commonly being used in critical care settings such as for neuroprotection viz. In post-cardiac arrest, TBI, stroke, meningitis, acute liver failure (ALF) and other conditions such as hypoxic ischemic encephalopathy (HIE), spinal cord injury, trauma and myocardial infarction (MI). A detailed list is described in [Figure 2](#F2){ref-type=\"fig\"}. The references from identified studies were also searched for additional citations.\n\nWe also specifically looked for articles featuring therapeutic hypothermia and post-cardiac arrest in English literature and human subjects. Around 2839 articles were found which are shown in [Figure 3](#F3){ref-type=\"fig\"}. In the year 2002, two articles published in NEJM changed the world regarding the use of TTM in the post-cardiac arrest and there have been a plethora of articles post 2002 as shown in [Figure 3](#F3){ref-type=\"fig\"}. There has been International Liason Committee on Resuscitation (ILCOR) updates, Cochrane reviews and meta-analyses in concerned area. In the following review we will be discussing current evidence and practices of TTM in critical care settings.\n\n![Use of therapeutic hypothermia and postcardiac arrest studies in literature](IJCCM-19-537-g004){#F3}\n\n {#sec2-3}\n\n### Outside hospital cardiac arrest {#sec3-11}\n\nIn early 1960\\'s, the use of CPR increased survival for patients who had cardiovascular collapse. There was a gradual improvement in the number of survivors as more and more medical staff got trained and above all increased availability of defibrillators. Most patients who had return of spontaneous circulation (ROSC) did not survive to leave the hospital and even if they did so, that too in a neurologically devastated state due to the neuronal injury, which begins within minutes after cardiac arrest. The neurological injury was so severe that it led to research into this area so that more and more people can be neurologically restored. In this research, only one thing that came out to be feasible and functional was the use of therapeutic hypothermia.\n\nSignificant positive studies\n\nHypothermia after cardiac arrest trial\n\nThis trial was conducted at nine European centers and overall included 273 patients with out-of-hospital cardiac arrest (OHCA) due to VF or VT. The use of therapeutic hypothermia, that is, temperature of 32--34\u00b0C led to improved survival at 6 months (59% in hypothermic group vs. 45% in normothermic group; *P* = 0.009) and better neurologic outcome (55% in hypothermic group vs. 39% in normothermic group; *P* = 0.046).\\[[@ref4]\\]\n\nBernard *et al*.\n\nIn contrast to multicenteric hypothermia after cardiac arrest (HACA) trial, this one was a single center trial, conducted in Australia. This trial included overall 77 patients with out-of-hospital cardiac arrest with presenting rhythm as VF. In this study, 49% of the patients in hypothermic group (33\u00b0C) survived with a good neurologic outcome as compared to 26% of normothermic group patients (*P* = 0.046).\\[[@ref3]\\]\n\nThe Cochrane review\n\nIncluded five RCT (Bernard *et al*. 2002; HACA *et al*. 2002; Hachimi-Idrissi *et al*. 2001; Laurent *et al*. 2005; Mori *et al*. 2000) which included a total of 481 patients. The Cochrane review summarizes that mild therapeutic hypothermia in patients with ROSC after cardiac arrest seemed to improve survival and neurologic outcome.\\[[@ref34]\\]\n\nInternational Liason Committee on Resuscitation 2010 statement\n\nTherapeutic hypothermia was advocated by ILCOR, thereafter adopted by European Resuscitation Council, American Heart Association, and by the National Institute for Health and Clinical Excellence. Three important facts inferred from ILCOR statement are: One that patients with out-of-hospital cardiac arrest (OHCA) with presenting rhythm VF/VT should be cooled. Patients with in-hospital cardiac arrest (IHCA) or presenting rhythm as a systole or pulseless electrical activity (PEA) may be cooled. Third point is that the temperature of 32--34\u00b0C should be targeted, and duration of hypothermia should be at least for 12--24 h.\\[[@ref35][@ref36]\\]\n\nCriticism of cooling and negative studies\n\nHypothermia after cardiac arrest trial\n\nThis trial had a very low inclusion rate that is, 8%, of all the 3551 screened only 275 patients were randomized. This low inclusion rates of around 8% make it difficult to generalize results to usual clinical practice. An average of just over one patient per week from nine centers was recruited over a period of 5 years. The study was abandoned because of a variety of reasons, which included slow recruitment, lack of funding and most importantly, there was no predefined power calculation. The level of Glasgow Coma Score (GCS) before randomization was not reported and withdrawal of care was not standardized leading to potential biases to the primary outcome measures of neurological outcome and death. The reporting of adverse effects was also inconsistent, making it difficult to assess the harm from this treatment.\\[[@ref4]\\]\n\nBernard *et al*.\n\nIn the trial by Bernard *et al*. the patients in normo-thermia group, that is, 37.4\u00b0C were actually having temperatures of 35.5\u00b0C on admission in emergency and these patients were passively warmed in Intensive Care Unit (ICU) settings. It was rise in temperature of 2\u00b0C in normo-thermia group that was causing more harm than the benefit of therapeutic hypothermia in 32\u00b0C group.\\[[@ref3]\\]\n\nCochrane review\n\nThe Cochrane review included trials by HACA group, Bernard *et al*., Hachimi-Idrissi *et al*., Laurent *et al*. and Mori *et al*. which were all conducted in time period between 2001 and 2005. The shortcomings of HACA trial are described above. In two trials, that is, by Hachimi Idrissi and Mori *et al*. there was baseline difference between groups. Similarly in trials by Bernard *et al*. and Laurent *et al*. there was no description of sequence generation or blinding.\\[[@ref34]\\]\n\nTrial sequential analysis by Nielsen *et al*.\n\nThe benefits and harms of the TTM were systematically evaluated, taking into account the risk of systematic bias and random errors. The treatment effects were quantified using trial sequential analysis, which reduced the risk of type I errors in contrast to cumulative meta-analysis.\\[[@ref37]\\] This trial concluded that there was a lack of firm evidence of a beneficial effect of TTM and that too the quality of evidence was quite low.\n\nNielsen *et al*. 2013\n\nIt was a multicenteric trial conducted over a period of 3 years in between 2010 and 2013 at 36 centers both in Europe and Australia.\\[[@ref38]\\] The inclusion criteria included patients with OHCA with presenting rhythm either as VF/VT or asystole (80% of patients had shockable rhythm, that is, VF, 12% had asystole and 8% PEA). The patients were randomized to receive TTM to either 33\u00b0 or 36\u00b0C, for at least 28 h and were continued on fever-reduction methods for 72 h postarrest. Post 72 h, a neurologist who was blinded to initial treatment recommended continued care or withdrawal based on predictors of poor neurological outcome, that is, presence of myoclonus, bilateral absence of N20 response on SSEP, absence of motor response or pupillary response. TTM to 33\u00b0C did not improve outcome in any of the measurable way. 50% of the patients (33\u00b0C group), as compared to 48% of the patients (36\u00b0C group) died (Hazard ratio with 1.06; 95% confidence interval \\[CI\\], 0.89--1.28; *P* = 0.51). Similarly at 6 months follow-up, 54% of the patients (33\u00b0C group) had died or had poor neurologic function, as compared to 52% of patients (36\u00b0Cgroup) (Risk ratio, 1.02; 95% CI, 0.88--1.16; *P* = 0.78). When the analysis was restricted only to the 80% of subjects with shockable rhythms, that is, VF, then also there was no benefit of TTM. The comparison between the three major trials (HACA, Bernard *et al*. and Nielsen *et al*.) is shown in [Table 2](#T2){ref-type=\"table\"}.\n\n###### \n\nSummary of postcardiac arrest trials\n\n![](IJCCM-19-537-g005)\n\nThe trial by Nielsen *et al*. had some shortcomings: Higher dose of hypothermia was used in previous TTM trials; there was about 4--4.5\u00b0C average temperature difference between two groups, as opposed to 3\u00b0C in the present trial. In this trial, the patient population included patients with shockable rhythms and nonshockable rhythms as compared to previous trials which included patients with shockable rhythms only. Despite therapeutic hypothermia\\'s proven physiologic benefits and success in experimental animals, the same success couldn\\'t be extrapolated in human settings or real life scenarios. This proves that healthy experimental animal is different from real life individual with multiple comorbidities. Above all there have been advancements in the field of critical care in last 20 years and improvements in patient care may have reduced the potential benefits of any single intervention.\n\nIn patients with OHCA when the presenting rhythm is either VF/VT, these patients should be cooled as early as possible, that is, within 4 h of arrest not after 8 h.\\[[@ref39]\\] The patients have to be cooled for at least 24--48 h. The trial by Nielsen *et al*. found similar outcomes (survival/neurological) with a new regimen targeting a temperature of 36\u00b0C. A word of caution is that we should not abandon therapeutic hypothermia in favor of strict fever management on the basis of one study, until all relevant issues have been satisfactorily addressed. Most importantly the patient should not be allowed to develop a fever for at least 72 h post-cardiac arrest.\n\n### Prehospital induction of hypothermia {#sec3-12}\n\nAround 1359 adults with prehospital cardiac arrest (583 with VF and 776 without VF) were randomized into two groups - intervention and control.\\[[@ref40]\\] Patients in the intervention group received prehospital cooling by infusing 2 L of 4\u00b0C normal saline following ROSC. The intervention reduced the time to achieve a temperature of \\<34\u00b0C by about 1 h. The intervention was neither associated with improved neurological status nor survival. Overall, the patients in the intervention group had more re-arrest in prehospital setup than control (26% vs. 21% respectively; *P* = 0.008), as well as increased diuretic use and pulmonary edema. The use of prehospital cooling reduced the time to reach a temperature of 34\u00b0C; it did not improve survival or neurological status.\n\nDebaty *et al*. conducted a randomized, multicenter study in three prehospital emergency medical services and four critical care units in France, to study the impact of intra-arrest therapeutic hypothermia (IATH) on neurological injury and inflammation following OHCA.\\[[@ref41]\\] The OHCA patients, irrespective of the initial rhythm, received either an infusion of cold saline and external cooling during cardiac arrest (IATH group) or TH was started after hospital admission (hospital-cooling group). The primary endpoint was serum neuron-specific enolase (NSE) concentrations at 24 h. The secondary end-points included interleukin-6 (IL-6), IL-8, and IL-10 concentrations and clinical outcome. 245 patients were included, 123 were analyzed in the IATH group and 122 in the hospital-cooling group. Levels of NSE and inflammatory biomarkers were not different between the two groups (median NSE at 24 h: IATH 96.7 \u03bcg/l (interquartile range (IQR): 49.9--142.8) vs. hospital cooling 97.6 \u03bcg/l (IQR: 74.3--142.4), *P* = 0.64). No difference in survival and cerebral performance were found at 1-month. The major criticism of this study was the patient population selected in this study was 80% of the patients with presenting rhythm as asystole and the NSE levels in this study were too high. Thus, the results of this study cannot be generalized as the results of this trial pertain to a specific patient population who already has a poor prognosis.\\[[@ref42]\\]\n\n### Use of therapeutic hypothermia after in-hospital cardiac arrest {#sec3-13}\n\nA multicenteric, prospective cohort study was conducted in a total of 538 hospitals which included 67,498 patients who had ROSC after IHCA.\\[[@ref43]\\] The basic aim of this study was to evaluate, that in how many patients post IHCA therapeutic hypothermia was initiated and if so whether the desirable target temperature was achieved or not. Therapeutic hypothermia was initiated in total of 1,367 of the total 67,498 patients (2.0%) only. Target temperature of 32--34\u00b0C was not achieved in 44.3% of these patients within 24 h and 17.6% were overcooled. Younger age (*P* \\< 0.001), arrest in a non-ICU location (*P* \\< 0.001), on a weekday (*P* = 0.005), and in a teaching hospital (*P* = 0.001) were associated with an increased likelihood of therapeutic hypothermia being initiated. Post IHCA, therapeutic hypothermia is initiated rarely and that too if initiated, the target temperature was commonly not achieved.\n\n### Condition associated with cerebral edema {#sec3-14}\n\nTraumatic brain injury\n\nTherapeutic hypothermia as per its mechanism of action by decreasing CMRO~2~ is effective in reducing ICP. 16 RCT evaluating favorable neurological outcomes and fifteen evaluating survival were evaluated by a jury comprising of experts in field of critical care from various critical care societies, that is, SCCM, ESICM etc.\\[[@ref7]\\] The spectrum of head injury varied in these trials from diffuse brain edema, diffuse axonal injury to subdural hematoma, extradural hematoma, and contusion. The first two conditions are treated medically while the latter conditions are treated surgically. Hence, hypothermia in each of these conditions is combined with adjuvant modality. The outcome depends upon underlying spectrum of head injury, the basic treatment given and whether it is given within appropriate time limit or not matters the most. The depth and duration of hypothermia applied in each of these trials have varied widely, as have the use of medical or surgical modalities. The better results were achieved in centers with expertise in applied hypothermia. Five published meta-analyses indicate a trend toward improved neurologic outcome and mortality when hypothermia was used, but definitive statistical significance is lacking.\\[[@ref7][@ref44]\\] The recommendation from the TTM group is as follows:\n\nTTM may be considered in patients with TBI and increased ICPEarly initiation, duration of \\>48 h, and slow rewarming with tight monitoring of ICP appear to be of paramount importancePatients hypothermic on admission should possibly be maintained hypothermic or very slowly rewarmedHowever, despite superior ICP control, the favorable neurologic outcome could not be achieved in all patients.\n\nThe current data neither support nor discourage\\'s the use of TTM in patients with TBI.\n\nStroke\n\nMajority of stroke patients are spontaneously breathing patients and any attempt of cooling will lead to shivering. The shivering will lead to increased cerebral metabolic requirement and hence further aggravate brain injury. Thus mechanical ventilation becomes mandatory, if hypothermia has to be initiated in these patients. The current published literature contains no RCT to support or refute an assertion of TTM benefit.\\[[@ref7]\\] Furthermore, the use of early thrombolysis has itself changed the outcomes in patients suffering from acute ischemic stroke. Whether TTM as adjunctive therapy would further improve the outcome is questionable. The use of therapeutic hypothermia will increase the need for endotracheal intubation and hence mechanical ventilation. As we all know both intubation and mechanical ventilation are associated with harmful effects, these must be acknowledged too before planning a trial.\n\nAcute liver failure with cerebral edema\n\nTherapeutic hypothermia alone has been used as a modality to reduce ICP and mortality in patients with ALF. It has also been used as a bridge for patients waiting orthotropic liver transplant (OLT) and has been used to decrease ICP during OLT. There is a case series suggesting a favorable effect of TTM in patients with ALF with cerebral edema.\\[[@ref7]\\] No RCT exists till date in this field. This is an area of research where RCT can be planned regarding the use of TTM alone or in combination with hepatic dialysis strategies in patients with ALF.\n\nBacterial meningitis\n\nA multicenteric, randomized clinical trial was conducted at 49 ICU in France, over a period of 2 years. Ninety-eight patients with bacterial meningitis with GCS score of \\<8 for \\<12 h were randomized.\\[[@ref45]\\] All patients received appropriate antibiotics and adequate organ support. The patients in hypothermic group were cooled to 32--34\u00b0C for 48 h; on the other hand control group received standard care. After inclusion of 98 patients, the trial was terminated early because of excess mortality in the hypothermic group (25 of 49 patients \\[51%\\]) vs. the control group (15 of 49 patients \\[31%\\]; relative risk, 1.99; 95% CI, 1.05--3.77; *P* = 0.04). In severe bacterial meningitis use of mild-moderate hypothermia did not improve outcome but on the other hand proved to be more harmful.\n\n### Spinal cord injury {#sec3-15}\n\nDue to high incidence of motor vehicle accidents the incidence of spinal cord injuries is on a rise. No RCT exists to support or criticize TTM as effective treatment.\\[[@ref7]\\] Animal studies and little human success point to the importance of designing and executing an RCT evaluating benefit of TTM for this common and devastating condition.\n\n### Myocardial infarction {#sec3-16}\n\nHypothermia was sought as therapy in patients with MI as it decreases myocardial oxygen demand. With the advent of thrombolytic therapy, percutaneous coronary intervention and surgery, the outcome in patients with MI has improved so much over the last two decades.\\[[@ref7]\\] In current scenario the use of any cooling intervention must not delay the proven effective treatments. Furthermore, with advancement in fields of medicine, mortality is now so rare that proof of further reduction in mortality would likely require a very large RCT.\n\n### Trauma {#sec3-17}\n\nSpontaneous hypothermia is common posttrauma. The cause of hypothermia in this scenario is multifactorial. The injury itself leads to heat loss viz., bleeding which is accompanied by iatrogenic heat loss, which is due to exposure of injured part, cold environment and pumping of cold intravenous fluids to these patients. No RCT exists that support cooling of trauma patients and it is a known fact that uncontrolled hypothermia is associated with substantially worse outcomes.\\[[@ref7]\\] The current practice at present aims at prevention of hypothermia, that is, preventing heat loss via conductive, convective mechanisms.\n\n### Hypoxic-ischemic encephalopathy \\[[Table 3](#T3){ref-type=\"table\"}\\] {#sec3-18}\n\n###### \n\nSummary of evidence\n\n![](IJCCM-19-537-g006)\n\nHypoxic ischemic encephalopathy from asphyxial insults is associated with high mortality and long-term neurodevelopmental disability especially in infants and children. The injury is two staged. A certain amount of damage results from acute, primary neuronal death. This often is followed by a secondary, delayed period of neuronal loss. This secondary injury provides a therapeutic window in which further damage might be prevented. In a recent Cochrane review; the data from eleven RCT comprising 1505 term and late preterm infants were summarized.\\[[@ref46]\\] The review concludes that TTM was of benefit to term newborns with HIE. Death or major disability in form of neuro-developmental disability were all reduced in term newborns born with HIE.\n\nConclusion {#sec1-8}\n==========\n\nTargeted temperature management in the ICU is a promising multifaceted therapy for quite a few medical conditions. In patients with OHCA, if the presenting rhythm is either VF or VT these patients should be cooled as early as possible for at least 24--48 h. The use of TTM in IHCA is less widely practiced and prehospital use of hypothermia has not shown any benefit. TTM is generally accepted treatment for HIE in newborns. TBI represents heterogeneous patient population with heterogeneous treatments offered hence TTM alone has not shown any benefit. Future studies should focus on optimizing hypothermic treatment and assess its value in other clinical settings for the full benefit of patients.\n\n {#sec2-4}\n\n### Financial support and sponsorship {#sec3-19}\n\nNil.\n\n### Conflicts of interest {#sec3-20}\n\nThere are no conflicts of interest.\n"}
+{"text": "With your kind patronage and full support, *Journal of Traditional and Complementary Medicine* (JTCM) is one-year-old now. We already have five issues with 50 excellent articles in the past year. The number of visitors to the journal website is more than 130,000 since October 10, 2011 indicating the quality of the published papers. JTCM continues to grow and is getting stronger.\n\nJTCM strives to promote exchange of recent novel research findings in traditional and complementary medicine worldwide, and provide a convenient platform for sharing knowledge of traditional medicine. Towards this, JTCM is going to cooperate with Wolters Kluwer Health -- Medknow from 2013. In this collaboration, JTCM will be promoted to international academic community via more active and efficient strategies, and will provide novel, innovative, original, and important knowledge of traditional medicine. The papers published thus will be read and cited by a larger audience. JTCM is already indexed in Google Scholar, and the application for Medline is under processing. Hopefully, we could obtain good result very soon.\n\nWe sincerely hope that you continue to support JTCM, and look forward to your future participation and contribution.\n"}
+{"text": "1. Introduction {#sec1-ijms-17-01080}\n===============\n\nMyelodysplastic syndromes (MDS) are a group of bone marrow disorders characterized by dysplastic maturation of bone marrow cells, reduced levels of mature myeloid cells in peripheral blood, and increased risk of transformation to acute myeloid leukemia (AML) \\[[@B1-ijms-17-01080],[@B2-ijms-17-01080]\\]. MDS can be divided into four different subsets based on the international prognostic scoring system (IPSS), and based on this scoring patients can be classified as low-risk (low and intermediate 1 IPSS classes) and high-risk (intermediate 2 and high risk IPSS classes) \\[[@B3-ijms-17-01080]\\]. The bone marrow in low-risk MDS is characterized by increased apoptosis, whereas high-risk patients are characterized by accumulation of apoptosis-resistant blasts in bone marrow and, eventually, peripheral blood.\n\nThe immune system is postulated to play an important role in the pathophysiology of MDS, at least in a subset of the patients \\[[@B4-ijms-17-01080],[@B5-ijms-17-01080],[@B6-ijms-17-01080],[@B7-ijms-17-01080]\\]. Low-risk MDS is dominated by proinflammatory cells, while immunosuppressive cells (regulatory T cells and myeloid-derived suppressor cells) \\[[@B8-ijms-17-01080],[@B9-ijms-17-01080]\\] are more important in high-risk disease where they possibly provide immune evasion for MDS blasts and aid in transformation to AML. The bone marrow environment is important for normal maturation of bone marrow progenitor cells, and signaling through direct cell contact, as well as soluble mediators, are both necessary for correct maturation of progenitor cells \\[[@B10-ijms-17-01080],[@B11-ijms-17-01080]\\]. Thus, investigation of systemic (i.e., serum or plasma levels) of soluble mediators may reveal information regarding the pathophysiology in MDS, e.g., immunoregulation, as well as regulation of hematopoiesis.\n\nSystemic cytokine profiles (plasma or serum levels) in MDS patients have been investigated in relatively few previous studies \\[[@B12-ijms-17-01080],[@B13-ijms-17-01080],[@B14-ijms-17-01080]\\]. Kornblau et al. \\[[@B12-ijms-17-01080]\\] investigated serum levels of cytokines and chemokines for AML and MDS patients. They found major similarities regarding cytokine levels between these two groups, but large differences compared to healthy controls. Feng et al. \\[[@B13-ijms-17-01080]\\] found increased plasma levels of tumor necrosis factor \u03b1 (TNF\u03b1), interleukin 6 (IL-6), CCL3, interleukin 1 receptor \u03b1 (IL-1ra), hepatocyte growth factor (HGF), and CCL4 in MDS compared to patients with aplastic anemia (AA) and healthy controls. Epidermal growth factor (EGF) and CXCL5 were then lower in AA as well as MDS compared to controls, whereas only CCL2 was lower in aplastic anemia when compared to MDS. A third study by Pardanani et al. \\[[@B14-ijms-17-01080]\\] found increased plasma levels of a wide range of cytokines/chemokines in MDS patients compared to healthy controls. Pardanani et al. \\[[@B14-ijms-17-01080]\\] did not find any significant differences between high- and low-risk MDS, while Feng et al. \\[[@B13-ijms-17-01080]\\] observed increased CCL4 levels and decreased levels of CCL5, CXCL5, cluster of differentiation 40 ligand (CD40L), vascular endothelial growth factor (VEGF), and EGF in high-risk MDS compared to low-risk disease. This discrepancy calls for further exploration into the cytokine profiles in low- and high-risk MDS. Additionally, these studies did not combine a broad cytokine/chemokine profiling with investigation of biologically different, but functionally related, mediators (i.e., soluble adhesion molecules, proteases, protease inhibitors). For these reasons we have investigated the serum levels of a wide range of biologically diverse, but functionally interacting, soluble mediators (including several cytokines) in a large group of unselected MDS patients.\n\n2. Results {#sec2-ijms-17-01080}\n==========\n\n2.1. The Serum Levels of Single Mediators Show a Wide Variation between Myelodysplastic Syndromes (MDS) Patients {#sec2dot1-ijms-17-01080}\n----------------------------------------------------------------------------------------------------------------\n\nWe compared the serum cytokine levels for MDS patients and a group of healthy adults. Our controls were younger than the MDS patients. However, Feng et al. \\[[@B13-ijms-17-01080]\\] could not detect any age-dependent variation in the systemic levels for any of the cytokines included in our present study, and this has later been confirmed by others for a large number of these cytokines \\[[@B15-ijms-17-01080]\\]. It is not known whether systemic levels of soluble adhesion molecules, matrix metalloproteases, or tissue inhibitors show any age-dependent variation. For these reasons [Table 1](#ijms-17-01080-t001){ref-type=\"table\"} only includes the comparison of systemic cytokine levels for MDS patients and healthy adults, whereas the comparisons for the other mediators are not shown. We used the Mann--Whitney *U* test for these statistical analyses. The patients showed increased levels of several interleukins (IL-5 *p* \\< 0.001, IL-6 *p* = 0.013, IL-8/CXCL8 *p* \\< 0.001, IL-13 *p* = 0.001) and chemokines (CCL3 *p* \\< 0.001, CXCL10 *p* \\< 0.001), whereas they showed decreased levels for one interleukin (IL-10 *p* \\< 0.001), two chemokines (CCL5 *p* = 0.010, CXCL5 *p* \\< 0.001), two immunoregulatory cytokines (interferon \u03b3 (IFN\u03b3) *p* \\< 0.001, CD40L *p* = 0.013), and one growth factor (thrombopoeitin (TPO) *p* \\< 0.001). Most of these differences were still significant after Bonferroni correction (*p* \\< 0.0015). Thus, the differences in systemic mediator profiles include a wide range of biologically different cytokines.\n\nIt can be seen from [Table 1](#ijms-17-01080-t001){ref-type=\"table\"} that a relatively wide variation range was detected for many of the cytokines compared with the healthy controls. The systemic levels of many soluble adhesion molecules, matrix metalloproteases, and tissue inhibitors of matrix proteases also showed a wide variation between the patients.\n\nMost of the significant differences noted above from comparing all MDS patients with controls were also detected when comparing high- and low-risk patients separately with the healthy adults (data not shown). Only four mediators varied differently for high- and low-risk MDS patients; CXCL11 was significantly decreased only for low-risk patients compared to controls (*p* = 0.014) whereas CD40L (*p* = 0.003) and EGF (*p* = 0.008) were decreased only for the high-risk patients. Thus, for the majority of cytokines similar differences were detected when high- and low-risk MDS patients were compared with healthy individuals.\n\nPrevious studies have described that age can influence the systemic mediator levels also for patients with myeloid malignancies \\[[@B16-ijms-17-01080]\\]. However, only CXCL5 (*n* = 49. *r* = 0.343, *p* = 0.017), MMP-3 (*r* = 0.422, *p* = 0.003), TIMP-2 (*r* = 0.362, *p* = 0.011), and TIMP-4 (*r* = 0.506, *p* \\< 0.001) showed significant correlations with age for our MDS patients. MMP-1, MMP-3, MMP-8, MMP-9, TIMP-3, and P-selectin all showed positive correlation to both absolute neutrophil and lymphocyte counts (data not shown).\n\nWe could not detect any correlation between age and the systemic levels of soluble adhesion molecules, MMPs or TIMPs for our healthy controls, and for this reason we used the comparison between MDS patients and healthy controls to illustrate that these mediator levels also showed a considerable variation between patients. The following differences were then observed: The patient P-selectin levels were significantly lower than the corresponding levels in the healthy controls (median control levels 83,100 pg/mL, range 43,800--112,000 pg/mL; *p* = 0.001), whereas the patient levels of intercellular adhesion molecule 1 (ICAM-1) (median control level 146,000 pg/mL, range 76,100--329,000 pg/mL, *p* \\< 0.001) and vascular cell adhesion molecule 1 (VCAM-1) (median control level 652,000 pg/L, range 342,000--1,580,000, *p* \\< 0.001) were increased. Furthermore, the patient levels did not show any significant correlation with age for any of these soluble adhesion molecules.The patient levels were significantly lower than corresponding levels in healthy controls for MMP-2 (median control level 180,000 pg/mL, range 54,400--248,000 pg/mL, *p* \\< 0.001), MMP-3 (median control level 39,500 pg/mL, range 4600--160,000 pg/mL, *p* \\< 0.001), MMP-8 (median control level 6380 pg/mL, range 1708--12,100 pg/mL, *p* = 0.039), and MMP-9 (median control level 115,000 pg/mL, range 42,900--333,000 pg/mL, *p* \\< 0.001). In contrast, the patients showed increased MMP-7 levels compared with the controls (median control level 2060 pg/mL, range 944--8930 pg/mL, *p* \\< 0.001). Only MMP-3 showed a correlation with age for the MDS patient group.\n\n### 2.1.1. Serum Mediator Profiles Differ Slightly between High- and Low-Risk MDS Patients {#sec2dot1dot1-ijms-17-01080}\n\nMDS is a heterogeneous disease and the inflammatory state seems to differ between low- and high-risk patients \\[[@B5-ijms-17-01080],[@B17-ijms-17-01080],[@B18-ijms-17-01080]\\]. We, therefore, compared serum levels of cytokines, growth factors, adhesion molecules, and proteases/protease inhibitors between the high- and low-risk MDS patient groups ([Table 2](#ijms-17-01080-t002){ref-type=\"table\"}). Patients with high-risk MDS had lower serum levels of EGF (Mann--Whitney *U* test, *p* = 0.011), CD40L (*p* = 0.006), CCL5 (*p* = 0.001), CCL11 (*p* = 0.012), and CXCL5 (*p* = 0.004) compared to the low-risk patients ([Figure 1](#ijms-17-01080-f001){ref-type=\"fig\"}). The levels of MMP-1 (*p* = 0.003) and MMP-9 (*p* = 0.010), as well as the inhibitor TIMP-2 (*p* = 0.006), were also lower in high-risk patients ([Figure 1](#ijms-17-01080-f001){ref-type=\"fig\"}). However, CCL5 was the only mediator which retained significance using Bonferroni-corrected *p*-values (*p* \\< 0.0013). Differences found based on high- and low-risk were confirmed when comparing patients with blast counts below and above five percent, with the exception of CCL11 and MMP-9, which were present in similar levels between the two groups (EGF *p* = 0.002, CD40L *p* = 0.001, CCL5 *p* = 0.004, CCL11 *p* = 0.213, CXCL5 *p* = 0.001, MMP-1 *p* = 0.002, MMP-9 *p* = 0.059, and TIMP-2 *p* = 0.001). Thus, high-risk MDS patients show decreased serum levels of a subset of biologically heterogeneous soluble mediators compared to low-risk patients.\n\n### 2.1.2. Platelet Counts, Platelet-Derived Mediators and Serum Levels {#sec2dot1dot2-ijms-17-01080}\n\nPlatelets contain several mediators that can be released during ex vivo handling of the samples \\[[@B19-ijms-17-01080]\\]. Platelet counts were available only for 30 of our patients at the time of sampling. We observed significant correlations between platelet counts and serum levels only for CD40L (Spearman correlation, *n* = 30, *r* = 0.689, *p* \\< 0.001), CCL5 (*r* = 0.569, *p* = 0.001), CXCL5 (*r* = 0.371, *p* = 0.044), VEGF (*r* = 0.610, *p* \\< 0.001), TPO (*r* = 0.435, *p* = 0.016), and P-selectin (*r* = 0.394, *p* = 0.031) when investigating the whole group which was including both high- and low-risk patients. However, the platelet counts for the low-risk patients were relatively high (median 175 \u00d7 10^9^/L, range 26--566) and when these 20 patients were investigated separately we observed significant correlations between peripheral blood platelet counts and CD40L (*r* = 0.751, *p* \\< 0.001), CCL5 (*r* = 0.652, *p* = 0.002), VEGF (*r* = 0.565, *p* = 0.009), TPO (*r* = 0.516, *p* = 0.020), and P-selectin (*r* = 0.677, *p* = 0.001). No statistically significant correlations were detected when the 10 high-risk patients were examined separately (median 66 \u00d7 10^9^/L, range 17--206). These observations suggest that ex vivo platelet release may have a major impact on the serum levels for these mediators.\n\n### 2.1.3. The Overall Serum Mediator Profile Shows a Wide Variation in MDS Patients and Cannot Distinguish between Low- and High-Risk Patients {#sec2dot1dot3-ijms-17-01080}\n\nTo further visualize the differences in serum mediator profiles for our MDS patients we did unsupervised hierarchical clustering analyses. The mediators that showed strong correlations between serum levels and platelet counts (CD40L, CCL5, VEGF, and TPO) were left out from these analyses together with those mediators that showed detectable levels in less than 10 samples. We first did an analysis investigating serum levels of 37 soluble mediators from both patients and healthy controls ([Figure 2](#ijms-17-01080-f002){ref-type=\"fig\"}). The individuals then grouped into two major subsets/clusters; all of the healthy individuals were included in the upper cluster and formed a distinct sub-cluster. The upper cluster also included a heterogeneous group of MDS patients. This analysis clearly illustrates that there is a wide variation in systemic cytokine profiles for MDS patients and a minority of patients even shows a profile very similar to healthy individuals. However, an interesting observation is that five out of six patients intermingled between the controls are either RARS or RA with very good prognostic cytogenetics. A major difference between the healthy individuals and patients was caused by compounds in the left major cluster of mediators particularly IFN-\u03b3, MMP-2, MMP-3 (leftmost cluster) and CXCL5, MMP-1, and MMP-9. The lower cluster was mainly defined by higher levels of IL-1\u03b1, IL-5, and IL-13 (rightmost cluster). Finally, the frequency of high-risk patients differed between the upper (patients 38, 26, 25, 43, and 40) and lower patient clusters (patients 47, 11, 9, 6, 21, 18, 46, 45, and 19) (four out of 43 in the upper cluster versus nine out of 25 in the lower cluster; \u03c7-square test, *p* = 0.0158). A similar distribution of high-risk patients between, and within, two main clusters was also seen when the clustering analysis only included the patients (data not shown).\n\nFinally, we performed a clustering analysis based only on five mediators that showed significant differences between high- and low-risk patients, leaving out CCL5, CXCL5, and CD40L to exclude a possible influence from ex vivo handling of samples ([Figure 3](#ijms-17-01080-f003){ref-type=\"fig\"}). The patients were then separated into two main clusters, and even though the use of this selection of biomarkers were able to group high-risk patients to some extent (four out of 29 in the upper cluster versus 10 out of 19 in the lower cluster; \u03c7-square test, *p* = 0.0148) high-risk patients did not form separate sub-clusters. Taken together with the overall mediator analysis in [Figure 2](#ijms-17-01080-f002){ref-type=\"fig\"} these last results show that there is a wide variation between MDS patients with regard to their systemic mediator profile, and this variation seems without strong relation to the high/low-risk classification.\n\n### 2.1.4. Systemic Mediator Profiles Show Only Minor Variation over Time {#sec2dot1dot4-ijms-17-01080}\n\nWe could compare the serum levels of soluble mediators at the time of diagnosis and 12 months later for 24 unselected patients (16 low-risk and eight high-risk). Two of the low-risk patients then experienced an increase in bone marrow blasts (1% versus 6.5% and 3% versus 9%, respectively), while the other 14 had stable disease. Two of the high-risk patients transformed to AML before 12 months. One of the low-risk patients and five of the high-risk patients were treated with 5-azacitidine in the 12 months period. When investigating all 24 patients together, the majority of the serum markers were not significantly altered, the exceptions being decreased levels in the 12-month sample of leptin (Wilcoxon's signed ranks, *n* = 24, *p* = 0.005), MMP-2 (*p* = 0.018), TPO (*p* = 0.026), and ICAM-1 (*p* = 0.040). When the low-risk patients were examined alone decreased leptin levels (*n* = 15, *p* = 0.008) could still be detected and in addition VCAM-1 levels were significantly increased (*p* = 0.007), whereas analysis of the high-risk patients alone showed increased MMP-3 levels (*n* = 8, *p* = 0.012). Thus, the systemic mediator profiles showed only minor variation over time.\n\n3. Discussion {#sec3-ijms-17-01080}\n=============\n\nThe cytokine network is important in orchestrating immune responses, and previous studies suggest that this network is dysregulated in MDS. However, the cytokine network interacts with several other and biologically different immune mediators, e.g., soluble adhesion molecules and the protease system \\[[@B20-ijms-17-01080],[@B21-ijms-17-01080]\\]. Our study is the first to analyze the systemic (i.e., serum) profile of such a broad range of biologically-heterogeneous soluble mediators in a relatively large group of unselected MDS patients. Previous studies have shown conflicting results regarding systemic cytokine levels in MDS. We observed a wide variation of the soluble mediator profile in serum from MDS patients and this variation showed no strong association with the established prognostic classification of MDS patients.\n\nOur healthy controls were younger than the MDS patients, but two previous studies failed to show any age-dependent differences between healthy individuals for a majority of the cytokines that they examined \\[[@B13-ijms-17-01080],[@B15-ijms-17-01080]\\]. One of these studies compared the levels of all cytokines included in our present study in different age groups \\[[@B13-ijms-17-01080]\\]. To the best of our knowledge it is not known whether systemic levels of soluble adhesion molecules, MMPs, and TIMPs correlate with age. We, therefore, solely compared serum cytokine levels for the MDS patients and our healthy controls in [Table 1](#ijms-17-01080-t001){ref-type=\"table\"}.\n\nThe levels of several soluble adhesion molecules and MMPs differed between our patients and healthy controls and these adhesion molecules/MMPs did not show any correlation with age neither for our patients (MMP-3 being the only exception), nor for the healthy controls. However, despite this, we would emphasize that these results have to be interpreted with great care.\n\nOnly serum samples were available for our study, and platelets may then become activated and release soluble mediators ex vivo during sample preparation with activation of the coagulation system \\[[@B19-ijms-17-01080],[@B22-ijms-17-01080]\\]. A significant correlation between serum mediator levels and peripheral blood platelet counts would then be expected if this ex vivo release caused a significant contribution to the serum levels, and this was observed only for a minority of the mediators examined in our study. On the other hand, it would not be surprising if the true systemic in vivo levels of platelet-associated mediator levels were correlated with the peripheral blood platelet counts, and this is also supported by studies of plasma samples \\[[@B19-ijms-17-01080]\\]. Whether one should leave out or include serum levels of platelet-derived mediators in analyses of serum mediator profiles is, thus, controversial. For the clustering analysis in the present study we either left out those mediators showing very strong correlations with platelet counts ([Figure 2](#ijms-17-01080-f002){ref-type=\"fig\"}) or we left out all showing significant correlations ([Figure 3](#ijms-17-01080-f003){ref-type=\"fig\"}).\n\nMMPs and their inhibitors should be regarded as immunoregulatory mediators and show several interactions with the chemokine/cytokine system \\[[@B21-ijms-17-01080]\\]. Very few previous studies have investigated MMP expression in MDS \\[[@B23-ijms-17-01080],[@B24-ijms-17-01080],[@B25-ijms-17-01080],[@B26-ijms-17-01080]\\], and these studies examined the intracellular expression or the release during in vitro culture of bone marrow mononuclear cells. In contrast to the high release of MMP-2 by bone marrow mononuclear cells described by Ries et al. \\[[@B23-ijms-17-01080]\\] and Travaglino et al. \\[[@B25-ijms-17-01080]\\] we observed decreased levels of MMP-2 in serum from MDS patients, and our present results thereby suggest that MMP-2 release by non-myeloid cells and/or MMP-2 degradation/absorption are more important for the systemic levels than the release by immature myeloid cells. Our observation that MMP-1/3/9 serum levels correlate with neutrophil and lymphocyte peripheral blood counts are also consistent with the hypothesis that the serum MMP levels reflect release by various cells. Former observations regarding MMP-9 are heterogeneous; Messingerov\u00e1 et al. \\[[@B26-ijms-17-01080]\\] observed increased MMP-9 levels in MDS patients with del(5q), and these levels normalized after treatment with lenalidomide, Travaglino et al. \\[[@B25-ijms-17-01080]\\] observed normal levels in MDS and decreased levels in patients with AML, while Ries et al. \\[[@B23-ijms-17-01080]\\] found decreased MMP-9 compared to healthy controls. These observations are consistent with our present observations that there is a wide distribution range for MDS patients with regard to MMP-9 levels. Travaglino et al. \\[[@B25-ijms-17-01080]\\] observed inverse correlations between the frequency of bone marrow blasts and release of MMP-2, as well as MMP-9 by myeloid cells. Observations by Iwata et al. \\[[@B24-ijms-17-01080]\\] also suggested reduced potential for MMP-9 induction in monocytes from MDS patients with excess of blasts. These previous observations may explain our present observation of lower MMP-2/3/9 serum levels in high-risk patients. Finally, MMP-7 serum levels were increased in MDS patients and showed no correlation to peripheral blood cell counts. MMP-7 is involved in cleavage of Fas ligands \\[[@B27-ijms-17-01080]\\], and the increased levels may, thus, affect immune surveillance or regulation of apoptosis in MDS patients.\n\nThe results from the two previous studies of soluble adhesion molecules in MDS are conflicting; S\u00fcdhoff et al. \\[[@B28-ijms-17-01080]\\] found no difference between MDS patients and healthy controls, whereas Passam et al. \\[[@B29-ijms-17-01080]\\] reported increased systemic levels of both VCAM-1 and ICAM-1 and this increase was most pronounced in high-risk disease. Both studies described significant correlations between these adhesion molecule levels and peripheral blood monocyte counts, and Passam et al. \\[[@B29-ijms-17-01080]\\]. in addition observed an association between high levels and adverse prognosis. We also detected increased serum levels of VCAM-1 and ICAM-1 in MDS, and higher median levels in high-risk disease, but we could not detect any association with prognosis based on IPSS scores. Of note, the overlap between low- and high-risk groups for these adhesion molecules was greater than between MDS patients and controls, and a lack of significant association is, thus, not likely to be caused by smaller sample size.\n\nWe observed increased levels of CXCL8/IL-8 and CXCL10 in MDS compared to controls in our cohort of unselected patients, and these chemokines clustered together in [Figure 2](#ijms-17-01080-f002){ref-type=\"fig\"}. CXCL8/IL-8 is (i) a chemoattractant, primarily for neutrophils; and (ii) a pro-angiogenic factor \\[[@B30-ijms-17-01080]\\]. CXCL8 has previously been described as the quantitatively-dominant pro-angiogenic factor in AML \\[[@B31-ijms-17-01080]\\], found to be increased in MDS and AML \\[[@B32-ijms-17-01080]\\]. Further, release of CXCL8/IL8 has been found to be increased when AML blasts are co-cultured with osteoblasts \\[[@B33-ijms-17-01080]\\]. CXCL8/IL-8 binds the receptor CXCR2 and the inhibition of this axis is proposed as a potentially specific therapeutic target in MDS and AML \\[[@B32-ijms-17-01080]\\]. Increased CXCL10 levels have also been detected in previous studies and has been suggested as an independent poor prognostic factor \\[[@B14-ijms-17-01080]\\]; our present results describing no difference between low- and high-risk patients defined by the IPSS score is also consistent with the hypothesis that CXCL10 levels have an independent prognostic impact. CXCL10 binds to the CXCR3 chemokine receptor and acts as (i) a chemoattractant, particularly for Th1 and Tc1 effector lymphocytes \\[[@B34-ijms-17-01080],[@B35-ijms-17-01080],[@B36-ijms-17-01080]\\]; (ii) an angioregulatory mediator \\[[@B37-ijms-17-01080]\\]; and (iii) a regulator of hematopoiesis \\[[@B38-ijms-17-01080]\\]. The previously-described increased expression of CXCR3 mRNA both in peripheral blood and bone marrow mononuclear cells further support a role of CXCL10 in the pathogenesis of MDS \\[[@B39-ijms-17-01080]\\].\n\n4. Materials and Methods {#sec4-ijms-17-01080}\n========================\n\n4.1. Patients {#sec4dot1-ijms-17-01080}\n-------------\n\nThe study was approved by the Regional Ethics Committee (REK vest; REK number 3.2008.409), and samples were collected after written informed consent. Serum samples were collected at the time of diagnosis for 49 patients (35 low-risk and 14 high-risk) ([Supplementary Table S1](#app1-ijms-17-01080){ref-type=\"app\"}), and a second sample was collected one year later for 24 of them (16 low-risk, eight high-risk). The male/female ratio for the MDS patients were 3.1:1 and the median age was 74 years (54--93). Serum samples from 23 healthy individuals were also collected (male/female ratio 0.6:1, median age 38 years (26--68)). Haukeland University Hospital (Bergen, Norway) is the primary hospital for a defined geographic area and our patient population represents an unselected/consecutive group of MDS patients from this single institution. Patients were risk-stratified according to the international prognostic scoring system (IPSS) \\[[@B3-ijms-17-01080]\\], and patients with classification as low and intermediate-1 IPSS were grouped together as low-risk, while intermediate-2 and high IPSS classification are referred to as high-risk. Three patients who scored as intermediate-1 IPSS, were included in the high-risk group as patient one (patient 18, see [Supplementary Table S1](#app1-ijms-17-01080){ref-type=\"app\"}) showed clinical progression with increasing BM blast counts (to 9%) and increasing cytopenias following sampling, the second (patient 25) showed a rapid increase in bone marrow blast counts (to 18%) within 10 weeks after sampling and the third patient (patient 40) had a major discrepancy to international prognostic scorin system revised (IPSS-R) (score 4.0---intermediate risk) and increasing bone marrow blast counts (to 6%) within 15 weeks of sampling.\n\n4.2. Sample Preparation and Analysis {#sec4dot2-ijms-17-01080}\n------------------------------------\n\nPeripheral blood samples were left to coagulate at room temperature for 60 min before centrifugation (1300\u00d7 *g* for 10 min) and subsequent serum collection. Samples were stored at \u221280 \u00b0C until the levels of 47 soluble mediators were determined by Luminex analyses (R&D Systems, Minneapolis, MN, USA): (i) the eleven interleukins IL-1\u03b1, IL-1\u03b2, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70(IL-12), IL-13, IL-17, and interleukin 1 receptor antagonist (IL-1ra); (ii) the nine chemokines CCL-2/3/4/5/11, CXCL8/IL-8 and CXCL-5/10/11; (iii) the eight growth factors granulocyte colony-stimulating factor (G-CSF), granulocyte macrophage colony-stimulating factor (GM-CSF), VEGF, TPO, EGF, fibroblast growth factor basic (FGF-B), HGF, and leptin; (iv) the three immunomodulatory cytokines IFN-\u03b3, CD40L, and TNF-\u03b1; (v) the four soluble adhesion molecules P-selectin, E-selectin, ICAM-1, and VCAM-1; and (vi) the eight matrix metalloproteinases (MMP)-1/2/3/7/8/9/12/13 together with four tissue inhibitors of metalloproteinases (TIMP)-1/2/3/4. Six of these mediators (FGF-B, GM-CSF, IL-2, IL-12, IL-17, and MMP-13) were undetectable for more than 90% of samples, and were excluded from the analyses. Samples from 48 MDS patients (34 low-risk and 14 high-risk) and 20 healthy individuals were analyzed for cytokine/chemokine and growth factor levels, whereas samples from all 49 patients and 23 controls were analyzed for protease and adhesion molecule levels; TIMP levels were not analyzed for the controls.\n\n4.3. Statistical and Bioinformatical Analyses {#sec4dot3-ijms-17-01080}\n---------------------------------------------\n\nStatistical analyses were performed using IBM SPSS (IBM Corp., Armonk, NY, USA) and Graph Pad Prism (GraphPad Software, La Jolla, CA, USA, used for analysis of paired samples). Mann--Whitney *U* test was used when comparing non-parametric values. Paired samples were analyzed using Wilcoxon's signed ranks test. Correlation analysis for nonparametric values was performed using Spearman's test. *p*-Values below 0.05 were considered significant. Bioinformatical analyses were performed using J-Express (MolMine AS, Bergen, Norway) \\[[@B40-ijms-17-01080]\\]. For unsupervised hierarchical clustering all values were median variance standardized and log(2)-transformed. Complete linkage was used as linkage method, and for distance measured, the Euclidian correlation was used.\n\n5. Conclusions {#sec5-ijms-17-01080}\n==============\n\nTo conclude, we investigated the systemic/serum profile of 47 biologically different soluble immune mediators (i.e., cytokines/chemokines, soluble adhesion molecules, proteases and their inhibitors) for a cohort of unselected MDS patients, and we describe: (i) differences between MDS patients and healthy individuals for several of these biologically different mediators but, at the same time, a considerable heterogeneity within the MDS cohort with regard to their serum profile of soluble immune mediators; (ii) this wide variation is seen both for patients with low- and high-risk disease, (iii) low- and high-risk patients differ only for a minority of these mediators; and (iv) this mediator profile is relatively stable over time in the individual patient, and this seems to be true even for patients with disease progression requiring treatment with hypomethylating agents. Based on our present results we suggest that the possible prognostic or predictive impact of the systemic profile of soluble immune mediators should be further investigated in future clinical studies.\n\nSupplementary materials can be found at <http://www.mdpi.com/1422-0067/17/7/1080/s1>.\n\n###### \n\nClick here for additional data file.\n\n\u00d8ystein Bruserud conceived and designed the experiments; Kristin Paulsen Rye performed the experiments; Astrid Olsnes Kittang, Kristoffer Sand and Annette Katharina Brenner analyzed the data; Astrid Olsnes Kittangand \u00d8ystein Bruserud contributed reagents and materials; Astrid Olsnes Kittang, Kristoffer Sand, Annette Katharina Brenner and \u00d8ystein Bruserud wrote the paper. All authors have read and approved the final version of the manuscript.\n\nThe authors declare no conflict of interest.\n\n![Serum levels of soluble mediators in myelodysplastic syndromes (MDS) patients, a comparison of patients with low-risk (squares) and high-risk (triangles) disease. The figure presents the overall results for those eight mediators showing statistically significant differences between the two groups. The Mann--Whitney *U* test was used for comparison; *p*-values \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01. CCL5 was the only mediator with a *p*-value lower than *p* = 0.0013, and considered significant following Bonferroni correction.](ijms-17-01080-g001){#ijms-17-01080-f001}\n\n![Unsupervised hierarchical clustering of MDS patients and healthy controls. Parameters where 10 or more sample values were above lowest detectable value were included in the clustering analysis. Parameters closely correlated to platelet counts were also excluded from cluster analysis (cluster of differentiation 40 ligand (CD40L), CCL5, vascular endothelial growth factor (VEGF), and thrombopoietin (TPO)). All values were median variance standardized and log(2)-transformed prior to Euclidian distance clustering. Red-to-green gradient (see lower left part of the figure) represents expression levels above or below the median serum levels for each individual parameter, respectively. The classification systems that are used in the right part of the figure with regard to low/high risk myelodysplastic syndromes (MDS) based on international prognostic scoring system (IPSS) (Risk), World Health Organization classification (WHO), and cytogenetic abnormalities (Cytogen) are also explained in the lower part of the figure. Risk: low-risk MDS patients---blue, high-risk MDS patients---red, healthy controls---white. WHO subclasses: refractory cytopenia (RC)---turquoise, refractory anemia with ring sideroblasts (RARS)---magenta, refractory cytopenia with multilineage dysplasia (RCMD)---green, refractory cytopenia with multilineage dysplasia and ring sideroblasts (RCMD-RS)---yellow, refractory anemia with excess blasts-1 (RAEB-1)---orange, RAEB-2---red, data not available (NA)---white. Cytogenetics: very good---magenta, good---green, intermediate---yellow, poor---orange, very poor---red, NA---white.](ijms-17-01080-g002){#ijms-17-01080-f002}\n\n![Unsupervised hierarchical clustering analysis of high- and low-risk MDS patients. The analysis was based on five mediators that showed statistically significant differences between the two MDS subsets; we excluded the three mediators CCL5, CXCL5, and CD40L that also showed significant differences but in addition showed significant correlations with platelet counts. All values were median variance standardized and log(2)-transformed prior to Euclidian distance clustering. Red-to-green gradient represents expression levels above or below the median serum levels for each individual parameter, respectively. Each row represents one individual, and low-risk MDS patients are marked with blue and high-risk MDS patients with red.](ijms-17-01080-g003){#ijms-17-01080-f003}\n\nijms-17-01080-t001_Table 1\n\n###### \n\nSerum levels of soluble mediators; a comparison between myelodysplastic syndromes (MDS) patients and healthy individuals.\n\n  Mediator                                                                   MDS Patients               Healthy Individuals     *p*-Value\n  -------------------------------------------------------------------------- -------------------------- ----------------------- --------------------------------\n  *Cytokines*                                                                                                                   \n  IL-1ra                                                                     761 (182--10,311)          852 (503--1700)         **\\<0.001**\n  IL-1\u03b1                                                                      5.04 (bd--241)             n/a                     n/a\n  IL-1\u03b2 \\*                                                                   bd (bd--6.99)              n/a                     n/a\n  IL-4                                                                       bd (bd--51.4)              bd (bd--29.5)           ns\n  IL-5                                                                       4.17 (bd--16.3)            bd (bd--1.57)           **\\<0.001**\n  IL-6 \\*                                                                    5.85 (bd--88.7)            3.47 (bd--5.23)         0.013\n  IL-10 \\*                                                                   bd (bd--22.8)              3.87 (2.62--5.33)       **\\<0.001**\n  IL-13                                                                      bd (bd--362)               bd (bd--66.7)           **0.001**\n  *Chemokines*                                                                                                                  \n  CCL2 \\*                                                                    276 (bd--1570)             268 (146--632)          ns\n  CCL3                                                                       113 (bd--6370)             55.2 (bd--137)          **\\<0.001**\n  CCL4 \\*                                                                    66.0 (bd--6320)            64.7 (14.0--150)        ns\n  CCL5                                                                       ad (134--ad)               ad (5060--ad)           0.010\n  CCL11                                                                      136 (bd--283)              114 (30.4--500)         ns\n  CXCL5                                                                      566 (bd--5700)             1 590 (425--4200)       **\\<0.001**\n  CXCL8 \\*                                                                   41.0 (8.10--1220)          14.5 (1.69--37.8)       **\\<0.001**\n  CXCL10 \\*                                                                  59.0 (7.00--590)           17.1 (9.16--57.6)       **\\<0.001**\n  CXCL11 \\*                                                                  40.5 (bd--2690)            bd (bd--167)            ns\n  *Immunomodulatory cytokines*                                                                                                  \n  TNF\u03b1                                                                       2.01 (bd--24.9)            3.17 (bd--17.8)         ns\n  IFN\u03b3 \\*                                                                    bd (bd--6.97)              2.94 (bd--3.57)         **\\<0.001**\n  CD40L \\*                                                                   3940 (264--27,300)         7560 (4470--22,600)     0.013\n  *Growth factors*                                                                                                              \n  EGF \\*                                                                     73.4 (bd--283)             80.1 (40.2--222)        ns\n  VEGF \\*                                                                    86.5 (bd--275)             84.2 (10.4--264)        ns\n  TPO                                                                        467 (110--2058)            779 (174--1057)         **\\<0.001**\n  HGF                                                                        275 (31.0--1605)           385 (175--545)          ns\n  G-CSF \\*                                                                   33.9 (bd--140)             25.6 (bd--71.3)         ns\n  Leptin                                                                     4750 (622--148,000)        10,500 (1400--46,600)   ns\n  *Adhesion molecules; serum levels in MDS patients*                                                                            \n  E-Selectin                                                                 28,400 (6770--107,000)     ICAM-1                  272,000 (84,100--589,000)\n  P-Selectin                                                                 56,200 (18,100--126,000)   VCAM-1                  1,220,000 (552,000--4,540,000)\n  *Matrix metalloproteinases and inhibitors; serum levels in MDS patients*                                                      \n  MMP-1                                                                      3570 (184--14,800)         MMP-12                  8.50 (bd--38.3)\n  MMP-2                                                                      45,800 (36,000--62,500)    TIMP-1                  166,000 (91,000--257,000)\n  MMP-3                                                                      12,500 (2720--17,800)      TIMP-2                  108,000 (57,400--248,000)\n  MMP-7                                                                      4810 (993--31,100)         TIMP-3                  27,400 (bd--77,400)\n  MMP-8                                                                      4260 (350--35,600)         TIMP-4                  2080 (1110--4170)\n  MMP-9                                                                      8110 (1040--74,700)                                \n\nAll mediators are given as pg/mL. The Mann--Whitney *U* test is used for comparison between the two groups for each mediator, and the corresponding *p*-values still significant after Bonferroni correction are given in bold (*p*-values \\< 0.0015). Fibroblast growth factor basic (FGF-B), granulocyte macrophage colony-stimulating factor (GM-CSF), interleukin 2 (IL-2), IL-12, IL-17, and matrix metalloproteinase 13 (MMP-13) are not included in the table as all or most serum levels of these mediators were undetectable. Serum levels of IL-1\u03b1, IL1-\u03b2, MMP-12, and tissue inhibitors of metalloproteinases 1-4 (TIMP1-4) were only investigated for MDS patients. ad; above detection level; bd, below detection level; n/a, not available; ns, not significant; tumor necrosis factor (TNF); interferon (IFN), cluster of differentiation 40 ligand (CD40L); epidermal growth factor (EGF); vascular endothelial growth factor (VEGF); thrombopoietin (TPO); hepatocyte growth factor (HGF); granulocyte colony-stimulating factor (G-CSF); intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1). \\* A previous study could not detect any age-dependent differences for these mediators except for CD40L that showed higher levels in elderly individuals, while G-CSF and CCL2 showed lower levels in elderly individuals \\[[@B15-ijms-17-01080]\\].\n\nijms-17-01080-t002_Table 2\n\n###### \n\nSerum levels of soluble mediators; a comparison between low- and high-risk MDS patients.\n\n  Mediator                                     Low-Risk MDS                     High-Risk MDS                    *p*-Value\n  -------------------------------------------- -------------------------------- -------------------------------- -----------\n  *Cytokines*                                                                                                    \n  IL-1ra                                       665 (318--4610)                  1032 (182--10,300)               ns\n  IL-1\u03b1                                        2.75 (bd--241)                   23.9 (bd--203)                   ns\n  IL-1\u03b2                                        bd (bd--6.99)                    0.87 (bd--4.04)                  ns\n  IL-4                                         bd (bd--46.0)                    bd (bd--51.4)                    ns\n  IL-5                                         bd (bd--16.3)                    4.55 (bd--16.1)                  ns\n  IL-6                                         5.55 (bd--39.2)                  5.85 (bd--88.7)                  ns\n  IL-13                                        bd (bd--292)                     103 (bd--362)                    ns\n  *Chemokines*                                                                                                   \n  CCL2                                         276 (bd--1570)                   268 (146--632)                   ns\n  CCL3                                         113 (bd--6370)                   55.2 (bd--137)                   ns\n  CCL4                                         66.0 (bd--6320)                  64.7 (14.0--150)                 ns\n  CCL5                                         ad (2010--ad)                    ad (134--ad)                     **0.001**\n  CCL11                                        136 (bd--283)                    114 (30.4--500)                  0.012\n  CXCL5                                        566 (bd--5700)                   1590 (425--4200)                 0.004\n  CXCL8                                        41.0 (8.1--1220)                 14.5 (1.69--37.8)                ns\n  CXCL10                                       59.0 (7.0--590)                  17.1 (9.16--57.6)                ns\n  CXCL11                                       40.5 (bd--2690)                  bd (bd--167)                     ns\n  *Immunomodulatory cytokines*                                                                                   \n  TNF\u03b1                                         4.02 (bd--17.8)                  bd (bd--24.9)                    ns\n  IFN\u03b3                                         bd (bd--3.35)                    bd (bd--6.97)                    ns\n  CD40L                                        5830 (918--27,300)               2260 (264--14,700)               0.006\n  *Growth factors*                                                                                               \n  EGF                                          115 (bd--283)                    bd (bd--144)                     0.011\n  VEGF                                         91.1 (17.1--191)                 69.1 (bd--275)                   ns\n  TPO                                          497 (110--1770)                  439 (154--2060)                  ns\n  HGF                                          290 (31.0--1610)                 273 (91.7--634)                  ns\n  G-CSF                                        32.4 (bd--89.0)                  38.7 (bd--140)                   ns\n  Leptin                                       4560 (622--54,700)               4750 (786--148,000)              ns\n  *Adhesion molecules*                                                                                           \n  E-Selectin                                   28,600 (6770--107,000)           24,700 (8310--59,700)            ns\n  P-Selectin                                   57,200 (18,100--126,000)         54,800 (21,400--94,800)          ns\n  ICAM-1                                       239,000 (84,100--503,000)        303,000 (118,000--589,000)       ns\n  VCAM-1                                       1,180,000 (553,000--4,540,000)   1,380,000 (870,000--2,770,000)   ns\n  *Matrix metalloproteinases and inhibitors*                                                                     \n  MMP-1                                        3820 (425--ad)                   777 (184--ad)                    0.003\n  MMP-2                                        45,900 (37,300--62,500)          44,700 (36,000--62,200)          ns\n  MMP-3                                        12,500 (5240--ad)                12,100 (2720--ad)                ns\n  MMP-7                                        4710 (993--28,400)               6100 (1640--31,100)              ns\n  MMP-8                                        4260 (438--35,600)               4100 (350--24,100)               ns\n  MMP-9                                        12,500 (1070--74,700)            5510 (1040--15,100)              0.010\n  MMP-12                                       8.20 (bd--38.3)                  8.50 (bd--16.2)                  ns\n  TIMP-1                                       167,000 (98,400--257,000)        135,000 (91,000--245,000)        ns\n  TIMP-2                                       119,000 (62,700--248,000)        102,000 (57,400--158,000)        0.006\n  TIMP-3                                       34,300 (bd--77,400)              9180 (bd--77,300)                ns\n  TIMP-4                                       2120 (1110--3500)                1880 (1280--4170)                ns\n\nAll mediators are given as pg/mL. The Mann--Whitney *U* test is used for comparison between low-risk and high-risk MDS patients for each mediator and the corresponding *p*-values still significant after Bonferroni correction are given in bold (*p*-values \\< 0.0013). FGF-basic. IL-2, IL-12, IL-17, and MMP-13 are not included in the table as all or most serum levels of these mediators were undetectable. ad, above detection level; bd, below detection level; ns, not significant.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "1. Introduction {#Section1}\n===============\n\nDengue Fever (DF) associated with Dengue Hemorrhagic Fever (DHF) and Dengue Shock Syndrome (DSS) accounts for about 300 million cases reported annually all across the world [@B1]. In 2016, India observed around 1,11,880 cases and 227 deaths as per data published by the National Vector Borne Disease Control Programme [@B2]. Besides the fact that dengue is emerging as major viral threat in the country, very few genomic sequences, specific to Indian sub-continent, have been reported [@B3]-[@B6]. Dengue virus has been reported as one of the fastest mutating viruses [@B7],[@B8] hence to address the increasing disease burden every year, a periodic knowledge of its whole genome sequence is necessary in order to update regional transmission pattern and clinical complexities.\n\nFor successful whole genome sequencing of any dengue virus serotype using sanger based method, availability of highly specific and sensitive primers covering whole genome is a critical determinant, as there exists spatiotemporal genomic diversity in individual serotypes leading to auxiliary categorization in lineages and genotypes [@B9]-[@B15]. Also, evidences of concurrent infections of multiple dengue virus serotypes in a single host adds further to the complexity as there is 60 to 75% similarity level in dengue virus serotypes at amino acid level [@B16]-[@B20]. During designing and selection of primers these complexities needs to be over ruled. Considering the challenges of an appropriate primer design for PCR based sequencing of whole genome of dengue virus subtypes, many of the PCR-based whole genome sequencing methods, as reported in the literature, may not function efficiently for every study.\n\nPresent paper reports, a simplified yet highly accurate Sanger based sequencing protocol of dengue virus subtypes which has broader applicability. To accomplish this, various bioinformatics tools and programmes were applied to utilize available dengue virus genome sequences data with high level of filtering and optimization. A nearly conservative primer pairs for each dengue virus subtype covering whole genome has been developed with ideal properties and high level of target region specificity and no cross-reactivity to other subtypes.\n\n2. Material and Methods {#Section2}\n=======================\n\n2.1. Primer Designing {#Section2.1}\n---------------------\n\nA methodology was employed for identifying genomic regions which were not repetitive and which showed high degree of evolutionary conservation in dengue virus subtype genome. The same design algorithm was employed for primer designing for each serotype. Also, it was ensured that developed primer sets have ideal characteristics of primer pairs, covering amplicon size of 500 - 1000 nucleotide bases and have no cross reactivity to the other than target region of same subtype genome and also no cross reactivity to other subtype genomes. Steps were taken to ensure primers do not have target in host genomes. To accomplish this, available genome sequencing data of each dengue virus serotype, various bioinformatics tools, computer programs and primer design tools along with several well defined filtering steps were used. The approach was iterative so that best primer pairs, for whole genome sequencing were designed. The process of primer design flow is described as flowchart in Figure [1](#F1){ref-type=\"fig\"}.\n\n2.2. Genome Selection {#Section2.2}\n---------------------\n\nFor selecting evolutionary conservative genomic regions for primer design multiple sequence alignment (MSA) of available whole genome sequences of a dengue subtype was used. Selection of genomes was a key step as selective filters were required to minimize effect of geographical and time scale variations in dengue viruses. Inclusion of too many genomes would have resulted in very few conservative regions, hence to optimize this, geographical and time scale filters were applied. Whole genomes of a dengue virus subtype reported from countries India, Pakistan, Nepal, Bangladesh and Myanmar (from China also in case of DENV4) isolated in period from January year 2000 to July 2015, were downloaded from NCBI virus variation database [@B21]. By applying these filters, a total of 14, 17, 8, and 5 complete genomes available for DENV-1, DENV-2, DENV-3 & DENV-4 respectively, were identified whose GenBank accession numbers are provided in the Supplementary file (Table [S1](#SM0){ref-type=\"supplementary-material\"}).\n\n2.3. Screening of Conserved regions {#Section2.3}\n-----------------------------------\n\nFor each serotype, pairwise local Multiple Sequence Alignment (MSA) was performed using multithreaded program MAFFT (v 7.221) option L-GSSi of above selected genomes [@B22]. Obtained MSA output was then analyzed in MEGA tool (version 6.0) for identification of conserved genomic sites (constant and singleton sites) [@B23]. Conserved regions of length 15 bases or more were screened using programs written in Perl programming language. Here we define conserved region a continuous array of conserved genomic sites only.\n\n2.4. Serotype specificity of conserved regions {#Section2.4}\n----------------------------------------------\n\nThe serotype specificity of obtained conserved regions was checked by using Basic Local Alignment Search Tool (BLAST). Search method \\'megaBLAST\\' was opted against Non-Reundant (NR) database of the remaining DENV serotypes [@B24]. Sequences which showed hits (more than five) to other serotypes in BLAST output were not taken further in primer designing process. Sequences which showed no or less than five hits were then taken for primer design.\n\n2.5. Designing Primers from Short sequences {#Section2.5}\n-------------------------------------------\n\nSelection and robust optimization process was performed for these short sequences to obtain desired primer pair candidates. It was ensured that only those primer pairs were selected which have 35% or more GC content, GC lock at 3 prime end, absence of self-annealing or primer dimer possibility with the pair, absence of secondary priming sites, and Melting temperature in a range of 62 to 67\u02daC. It was also ensured that 3\\' region of primer is completely from selected conserved sequence so that it remains a highly conserved region. The expected melting temperature (Tm) was calculated using MWG operon primer design tool and care was taken that difference of Tm of forward and reverse primer do not exceed by 5\u02daCelsius. Also, it was ensured that the amplicon length using primer pair remains within limits of 800-1200 nucleotides. Primer pairs were rejected on the basis of possible primer-dimer formation if a pair had five or more consecutive complementary nucleotides. To ensure complete coverage and end-to-end inclusion, primers for terminal regions were identified to those derived from reference genome with ideal characteristics of good primers but no checks for specificity and complementing them with a highly specific primer so that it forms an specific primer pair. For the purpose of genome positioning and for terminal region primer selection, top most hit of multiple alignment was selected for dengue virus serotypes DENV1 to DENV4 (GenBank accession numbers JN903578, JX475906, GQ466079 & JQ922560 respectively; listed in Supplementary Table [S2](#SM0){ref-type=\"supplementary-material\"}). The designed primers were then commercially manufactured by m/s Eurofins Genomics Ltd, Bengaluru, India.\n\n2.6. Empirical testing of designed primers employing Reverse Transcriptase- polymerase Chain Reaction (RT-PCR) assay {#Section2.6}\n--------------------------------------------------------------------------------------------------------------------\n\nSanger Sequencing of DENV serotypes was done using designed primers. In one of the earlier studies [@B25], whole genome sequencing of DENV-3 isolated from direct clinical serum sample and its C6/36 cell line passage culture sample was performed and reported (GenBank Accession numbers KU216209 & KU216208). The whole genome primers designed for DENV-3 were verified in that study; however sequence and method of primers designed is being reported here. Since no clinical sample positive for DENV-1, 2 and 4, was available during the study period, hence RNA extracted from DENV1-4 positive control virus mix supplied in Dengue virus type detection kit (m/s CDC, Atlanta, USA) was used. The mix contained heat\u2010inactivated DENV\u20101 Haw (Hawaii), DENV\u20102 NGC (New Guinea C), DENV\u20103 H87, and DENV\u20104 H241 strains for which whole genome sequence data was already available in NCBI database (GenBank accession numbers: DENV-1 Haw: EU848545, DENV2 - NGC: AF038403, DENV3 - H87: M93130 and DENV4 - H241: AY947539). The mapping of the primers designed with these corresponding DENV strains referred above is shown in Supplementary table (S3). Using of CDC-positive control virus mix also gave an advantage of protocol applicability if ever concurrent infections by multiple serotypes was to be reported.\n\nRNA was extracted using the QIAamp Viral RNA kit (m/s Qiagen, CA) followed by RT-PCR Step using the One Step superscript -III RT-PCR kit with platinum Taq polymerase (m/s ABI, USA) according to the manufacturer\\'s protocol. For testing each of the sets of designed primers; 19 fragments sets of DENV -1, 18 fragments sets of DENV - 2 and 18 fragment sets of DEN -4 (Total 55 sets; including both forward and reverse primers), a working solution of 1\u00b5l (10pmol) was used. The thermal cycling conditions was as follows: Reverse transcription at 55 \u00b0 C for 40 minutes, Denaturation at 94\u00b0 C for 3 minutes, further denaturation of 35 cycles at 94\u00b0 C for 30 seconds, 50\u00b0 C for 1 minutes and 68 \u00b0 C for 1 minutes and a final extension at 68 \u00b0 C for 5 minutes. After amplification, 5\u00b5l of each amplicons were run in 2% Agarose gel and bands were obtained. The expected Molecular weight (MW) for each amplified product was approximately 800-1000 nucelotide bases while those amplified with primer sets covering the start and stop points had MW of approximately 450-550 bases (Figure [2](#F2){ref-type=\"fig\"}).\n\n2.7. Amplicon Sequencing {#Section2.7}\n------------------------\n\nThe total 55 amplicons were then subjected to purification using PCR purification kit (QIA quick PCR purification kit, Qiagen, CA). 10\u00b5l of purified cDNA\\'s were eluted and subjected to cycle sequencing protocol (Big Dye Terminator v3.1 Ready Reaction -24 mix kit; m/s ABI, USA) using 1 \u00b5l of forward primer. Finally to remove any unincorporated dNTP\\'s, the cycle sequenced products were subjected to Dye Ex 2.0 spin kit (m/s Qiagen, CA). These were then subjected to 3130xl Genetic Analyzer (m/s ABI, USA) for obtaining their sequence information.\n\n2.8. Contig assembly {#Section2.8}\n--------------------\n\nSeqScape v2.6 (Applied Biosystems, USA) was used for quality analysis and optimization of sequenced fragments, contig assembly, reference based mapping of fragments, analysis of variations at nucleotide and amino acid level. NCBI Reference genomes for dengue virus subtypes were used for reference based assembly (DENV1 to DENV4: NC_001477, NC_001474, NC_001475 & NC002640 respectively).\n\n2.9. Integrative analysis {#Section2.9}\n-------------------------\n\nIntegrative analysis of primer nucleotide sequences, published genome sequence of DENV serotypes present in positive control virus mix, RT-PCR reaction outcome, sequencing data quality were performed to understand effect of nucleotide mismatch of primer sequence with the published genome of sample whose sequencing is done. Also, to estimate threshold of nucleotide mismatch in total and in the 3 prime end till which good quality sequencing results are obtained. Derived mismatch threshold values can be utilized in insilico testing of primers. Mapping of designed primers to available genome sequence of dengue virus strains present in positive dengue virus control mix was performed using NCBI - BLAST with relaxed search algorithm. Number of variations in nucleotide bases of primers compared to available genome was computed. Nucleotide variations of a pair of primer compared to published sequence were computed and then mapped with RT-PCR outcome and also sequencing quality of reads. Similar analysis were performed for forward primer only of pair, 3 prime end of primer pairs and 3 prime end only of forward primer.\n\n2.10. In-silico testing of primers using e-PCR {#Section2.10}\n----------------------------------------------\n\nFor In-silico testing of designed primer pairs, e-PCR tool standalone version was employed {Ref e-PCR}. For all primer pairs of each serotype which were given as query set, two separate searches were performed to two different genome data sets. First set had all genome sequences of that serotype for which primers have to be tested, submitted till 31st July 2015 and isolated in any of South Asian countries (Countries: Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan and Sri Lanka). In second set, whole Asian countries was taken as sample isolation source with submission time same i.e. 31st July 2015. Decision of geographical region filters was based on fact that the design constraints for primers were so that it has high probability of returning good quality sequencing results of dengue virus strains isolated from Asian or south Asian region. Nucleotide mismatch threshold level derived from integrative analysis were taken as input parameter for e-PCR program.\n\n3. Results {#Section3}\n==========\n\n3.1. Primer pairs designing and testing {#Section3.1}\n---------------------------------------\n\nThrough exhaustive computational analysis, a total of 144 primers (72 forward and 72 reverse) were designed covering the four serotypes (listed in suppl table [S2](#SM0){ref-type=\"supplementary-material\"}).\n\nOf this, 55 primer pairs were used in amplification of DENV 1,2 and 4 and 50 reactions gave results (16 of DENV1, 18 of DENV2 and 16 of DENV4) (Figure [2](#F2){ref-type=\"fig\"}). Average read length was 386 nucleotide bases and 82% bases were of high quality (Phred score \\> 20). No sequencing results were obtained for primers 12^th^, 18^th^ and 19^th^ of DENV1 and 18^th^ pair of DENV4.\n\nThe 17 primer pairs used in whole genome sequencing of DENV 3 isolated from patient serum and from passaged culture (as performed in earlier study)[@B25], showed average read length of 544 bases for serum sample and 560 bases for passaged sample. 87% of nucleotide bases were of high quality (Phred score \\>20) for clinical isolate while 79% of total bases were of high quality for cell culture passage. Genome coverage of both was the same valued to 1.6x.\n\n3.2. Integrative analysis results {#Section3.2}\n---------------------------------\n\nIntegrative analysis results suggested possible explanations for negative results to an extent. 18th primer pair for DENV1 had 13 nucleotide mismatches with published sequence which includes 3 mismatches in 3 prime end. Reverse primer of 18th pair for DENV4 had no matching sequence to published genome sequence as published sequence was shorter at 3 prime end compared to from which primer was designed. There were good quality results for either primer of a pair having upto 4 mismatches, primer pair having 8 mismatches, 3 prime end of a either primer of a pair having 1 mismatch and 3 prime end of primer pair combined having 2 mismatch to mapping region of published genome. These mismatch values were then considered threshold levels to be taken up as input for insilico e-PCR based testing (Supplementary file; Table [S3](#SM0){ref-type=\"supplementary-material\"}).\n\n3.3. e-PCR testing results {#Section3.3}\n--------------------------\n\nAmplification was predicted for all whole genomes as of July 2015 of both South Asian and Asian regions using e-PCR for designed primers. The analysis was performed to predict amplification by designed primers for both set of genomes for all serotypes. Output of reverse e-PCR was then further analysed using Perl programs. Genome sequences for which 15 or more primer pairs predicted amplification were then looked. Threshold 15 pairs was taken to avoid exclusion of genome sequences due to non-mapping of terminal regions primer pairs (Table [1](#T1){ref-type=\"table\"}).\n\n4. Discussion and Conclusion {#Section4}\n============================\n\nPresent paper discusses designing and testing of primer pairs for Sanger based whole genome sequencing of dengue virus serotypes using computational algorithm. These designed primers have ability to amplify reported genomes from Asian and South Asian region, as reported till July, 2015. Instead of designing primers from a single genome, an algorithm to extract relevant sequence information from multiple genomes was taken into consideration. Previous studies also supported that multiple alignment based methods in which identified conserved regions are taken up for designing primers show better efficiency compared to methods in which primers are derived from single gene or genome [@B25],[@B26]. In addition to identifying conserved regions, the paper also focusses on unique selection of only those conserved regions which showed high degree of dengue virus serotype specificity. The algorithm focuses to minimize effect on sensitivity of primers due to nucleotide variations present in dengue virus subtypes at geographical region and time scale levels. At the same time, it also serves to maximize specificity of primers so as to be efficient in cases of concurrent infections of multiple subtypes.\n\nIn case of rapidly evolving organisms such as RNA viruses, the available primers may become ineffective due to risk of insufficient complementation among primers and target sequence. Hence during primer designing it needs to be assured that there remains sufficient complementarity. Magnitude and threshold of complementarity of primers to target sequence so as to get a positive experimental outcome is usually not well defined, herein an attempt to address this by performing an integrative analysis is shown. The analysis suggests that up to eight nucleotide level mismatches in a primer pair with target genome sequence, can also lead to good quality sequencing results. However, nucleotide mismatches more than four and importantly mismatches in 3 prime end region in primers may be a reason for low quality or no sequencing results. These derived values were then given as input parameters in In-silico amplification prediction using e-PCR tool. The prediction results were highly promising and of good applicability to the commonly circulating subtypes with maximum prediction in DENV3 followed by DENV2 subtype. Moderate prediction for DENV4 subtype and least for DENV1 subtype was observed. For pathogens having well identified genotypes or with multiple serotypes, reported methodology has high applicability at individual genotype or serotype level.\n\nSupplementary Material {#SM0}\n======================\n\n###### \n\nSupplementary figures and tables.\n\n###### \n\nClick here for additional data file.\n\nAuthors are grateful to the Director-General, Indian Council of Medical Research, New Delhi, India for providing financial support to undertake the study and to the Director, Desert Medicine Research Centre, Jodhpur for encouragements and support. The study was supported by the funds provided by Indian Council of Medical Research, India (Biomedical Informatics Center). <http://www.icmr.nic.in/>. The funders had no role in study design, data collection and analysis and decision to publish or preparation of the manuscript. Authors are also thankful to Center for Disease Control, Atlanta, USA for allowing us to use Positive control virus mix of Dengue virus type detection kit for validation experiments.\n\n![Primer design process flow chart](jgenv06p0034g001){#F1}\n\n![Dengue virus genome bands obtained as fragments on a 2% Agarose gel. (A) DENV-1 genome bands amplified using 19 sets of primers displayed numerically from lane L1 through L19. (B).DENV-2 genome bands amplified using 18 sets of primers displayed numerically from lane L1 through L18. (c). DENV-4 genome bands amplified using 18 sets of primers displayed numerically from lane L1 through L18.](jgenv06p0034g002){#F2}\n\n###### \n\nReverse e-PCR results, with designed primers as query set and Genomes set as database.\n\n  Dengue virus subtype   Genomes predicted / Total Genomes (South Asia region, till July 2015)   Genomes predicted / Total Genomes (Asia region, till July 2015)\n  ---------------------- ----------------------------------------------------------------------- -----------------------------------------------------------------\n  DENV1                  3 /22                                                                   7/936\n  DENV2                  19/23                                                                   191/483\n  DENV3                  31/31                                                                   276/281\n  DENV4                  4 /6                                                                    15/34\n\n[^1]: Competing Interests: The authors have declared that no competing interest exists.\n"}
+{"text": "Introduction {#S1}\n============\n\nPodocytes are critical elements of the glomerular filtration barrier (GFB), a complex vascular unit in the glomerular capillaries that performs plasma ultrafiltration^[@R1]^. Podocytes are highly differentiated cells which form long primary and secondary extensions that wrap around the outside of the capillary loops. Between their interdigitating foot processes the slit diaphragm is formed which is a key component of the GFB^[@R2]^. Recent genetic and cell biological studies highlighted the critical importance of podocytes in the development of glomerular diseases^[@R3]--[@R11]^. Still, a critical barrier in understanding the mechanistic details of glomerular pathology is the technical limitation to study the GFB in its native environment. Due to the lack of *in vivo* data, there are still significant gaps in our understanding of podocyte dynamics and motility and their link to albuminuria and glomerulosclerosis. However, this insight would be needed for the development of new treatment strategies.\n\nBecause of the dynamic and complex structure of the GFB, the application of high resolution imaging tools may provide the long missing *in vivo* technology for podocyte research. MPM is a revolutionary, minimally invasive optical sectioning technique which allows the imaging of the mouse kidney^[@R12]^ including glomeruli *in vivo*^[@R13]--[@R15]^. Mouse genetic tools permit the cell-specific expression of multicolor fluorescent proteins^[@R16],[@R17]^ including in podocytes for imaging applications^[@R18]--[@R22]^. Also, genetic tagging and lineage tracing are powerful tools to study cell fate *in vivo*. Therefore we aimed to develop a new visual approach to directly visualize and track the two glomerular epithelial cell types, podocytes and PECs *in vivo* in the intact mouse kidney using serial MPM over time.\n\nThe motility of podocytes and PECs and their role in renal pathologies and nephron regeneration are debated. According to the classic view podocytes are terminally differentiated cells and cannot repair themselves by means of cell division^[@R23],[@R24]^. However, podocytes can proliferate in a limited number of conditions^[@R25]^. Lately the new research field of renal stem cells has also challenged the stationary podocyte concept by reports that podocytes can be replaced by PECs which migrate along the Bowman's capsule^[@R26]^ and express progenitor cell markers^[@R27]^. Visual techniques to track podocytes and PECs *in vivo* may aid research on these topics and the future development of novel progenitor or stem cell-based therapeutic approaches to renal injury. To test the utility of our new MPM imaging approach, two disease models were used: UUO and adriamycin nephropathy. UUO is a widely used animal model to study progressive renal disease and tubulointerstitial fibrosis^[@R28]^. UUO pathology features injury of PECs and damage of the tubuloglomerular junction leading to the formation of atubular glomeruli^[@R29]^. The rapid remodeling of the Bowman's capsule observed after UUO^[@R29]^ suggested that this model may be ideally suited to image PEC and podocyte motility in kidney injury. Adriamycin nephropathy is a model resembling human focal segmental glomerulosclerosis^[@R30]^.\n\nResults {#S2}\n=======\n\nCharacterization of Podocin-GFP (Pod-GFP) mice {#S3}\n----------------------------------------------\n\nPreviously we visualized podocytes based on their typical position and shape by a negative labeling technique in wild type animals using systemic injections of Lucifer Yellow, a freely filterable dye^[@R14]^ ([Supplementary Fig. 1a](#SD1){ref-type=\"supplementary-material\"}). However, this precluded their tracking during damage, migration or replacement. Therefore, we generated a Pod-GFP mouse model in which podocytes express membrane-targeted GFP, while all other cells express the red fluorescent protein Tomato. ([Supplementary Fig. 1b](#SD1){ref-type=\"supplementary-material\"}). GFP fluorescence outlined a single cell layer and was most intense closely around the capillary loops compared to the cell body, due to the membrane-rich foot process region (Inset, [Supplementary Fig. 1b](#SD1){ref-type=\"supplementary-material\"}). GFP expressing podocytes co-localized with cells identified by the negative labeling technique confirming the validity of our previous technique ([Supplementary Fig. 1c](#SD1){ref-type=\"supplementary-material\"}, and [Supplementary Movie 1](#SD2){ref-type=\"supplementary-material\"}). Co-staining for GFP and synaptopodin, a podocyte differentiation marker, confirmed podocyte-specific expression of GFP ([Supplementary Fig. 1d--f](#SD1){ref-type=\"supplementary-material\"}).\n\nMPM imaging of podocyte migration {#S4}\n---------------------------------\n\nCompared to the lack of labeling in control wild-type mice ([Fig. 1a](#F1){ref-type=\"fig\"}) podocyte-specific GFP expression in Pod-GFP mice is demonstrated in a low-power MPM image of a kidney 3 weeks after UUO ([Fig. 1b](#F1){ref-type=\"fig\"}). Starting 1--2 weeks after UUO MPM imaging visualized podocytes that formed sporadic multi-cellular clusters on the glomerular tuft ([Fig. 1c](#F1){ref-type=\"fig\"}) with short projections developing towards the Bowman's space and into the remainder of the proximal tubule fragment ([Fig. 1d](#F1){ref-type=\"fig\"}). Occasionally, detached single podocytes were found further downstream in the lumen of proximal tubules in a few filtering nephrons, indicating podocyte shedding from these projections ([Fig. 1e](#F1){ref-type=\"fig\"}). However, the most common morphological change in the UUO model was the appearance of podocytes in the parietal layer of the Bowman's capsule. In most glomeruli podocyte projections propagated to the PEC layer after a contact between podocytes and PECs was established. These contact points developed either via continuous transition at the vascular pole ([Fig. 1f](#F1){ref-type=\"fig\"}) or by forming bridges across the Bowman's space close to the urinary pole ([Supplementary Fig. 2a](#SD1){ref-type=\"supplementary-material\"}, and [Supplementary Movie 2](#SD3){ref-type=\"supplementary-material\"}) or anywhere around the glomerular tuft ([Fig. 1g](#F1){ref-type=\"fig\"}). The podocyte projections appeared to grow further along the PEC layer towards the remodeled (sealed) glomerulotubular junction and in some glomeruli to the proximal tubule fragment ([Supplementary Fig. 2a--c](#SD1){ref-type=\"supplementary-material\"}). Three to four weeks after UUO a continuous layer of GFP-expressing (GFP+) cells were found along the parietal Bowman's capsule and numerous cell bridges of podocyte origin were formed between the visceral and parietal cell layers ([Fig. 1g](#F1){ref-type=\"fig\"}). Statistical analysis ([Fig. 1h](#F1){ref-type=\"fig\"}) showed that \\<5 weeks after UUO the percentage of glomeruli in which \\>50% of the parietal layer was covered by GFP+ cells was 19\u00b14% (a total of n=895 glomeruli analyzed from n=43 mice) compared to 3\u00b12% in age-matched control mice (n=239 glomeruli from n=19 mice, p\\<0.05). The percentage of glomeruli with \\>50% parietal GFP+ cell coverage increased continuously with time after UUO ([Fig. 1i](#F1){ref-type=\"fig\"}) (R=0.846, n=48 mice, p\\<0.05). [Supplementary Fig. 2b](#SD1){ref-type=\"supplementary-material\"} demonstrates the rare phenomenon of GFP+ cell projections along the parietal layer of Bowman's capsule in control healthy (non-UUO) kidneys. Importantly, we never observed solitaire or isolated clusters of GFP+ cells in the parietal layer. These cells always had continuous physical connections to the glomerular tuft suggesting that they originated from propagating visceral podocytes. Immunolabeling of GFP and the podocyte marker podocin 9 weeks after UUO found expression of podocin in GFP+ cells in the PEC layer, further suggesting the migration of podocytes in this model ([Supplementary Fig. 2d--f](#SD1){ref-type=\"supplementary-material\"}).\n\nTo confirm that GFP+ cells in the parietal layer developed by the migration of visceral podocytes rather than podocin gene activation in PECs (hence Cre-mediated GFP expression), we repeated the experiment with tamoxifen-inducible Pod-GFP mice (iPod-GFP mice). There was no GFP expression outside the visceral podocyte layer in iPod-GFP mice after tamoxifen induction ([Fig. 1i](#F1){ref-type=\"fig\"}, [Fig. 2a](#F2){ref-type=\"fig\"}). However, \\>2--4 weeks after UUO, iPod-GFP mice developed the same glomerular phenotype as described above for the constitutive Pod-GFP model, including podocyte clustering ([Fig. 2a](#F2){ref-type=\"fig\"}), the development of visceral-to-parietal podocyte projections (bridging), and podocyte migration to the parietal layer ([Fig. 2b--c](#F2){ref-type=\"fig\"}). Statistical analysis ([Fig. 1i](#F1){ref-type=\"fig\"}) showed that similarly to the constitutive Pod-GFP model, the percentage of glomeruli with \\>50% parietal GFP+ cell coverage in iPod-GFP mice increased continuously with aging after UUO (R=0.972, n=296 glomeruli analyzed from n=5 mice, p\\<0.05), although the phenomenon had apparently different dynamics in the two models.\n\nIn addition, to study the effects of a different injury, Pod-GFP mice were treated with a single high dose of adriamycin. Similarly to the UUO model, adriamycin treatment within 4 days induced the development of sporadic podocyte clusters ([Fig. 2d](#F2){ref-type=\"fig\"}, [Supplementary Movie 3](#SD4){ref-type=\"supplementary-material\"}) and the appearance of GFP+ cells in the parietal layer close to the vascular pole transition, and in continuum with the visceral podocyte clusters ([Fig. 2e](#F2){ref-type=\"fig\"}). Histological analysis ([Fig. 2f](#F2){ref-type=\"fig\"}) revealed the significantly higher percentage of glomeruli with podocyte clusters 4 days after adriamycin treatment (7\u00b12%, from a total of n=1241 glomeruli, analyzed from n=4 mice) compared to 0\u00b10% in age-matched control mice (n=1460 glomeruli in n=4 mice, p\\<0.05). Albuminuria was measured in the same groups and confirmed that glomerular damage developed after adriamycin treatment ([Supplementary Fig. 3](#SD1){ref-type=\"supplementary-material\"}).\n\nSerial *in vivo* MPM imaging of the same glomerulus {#S5}\n---------------------------------------------------\n\nTo better understand the dynamics of podocyte migration we developed serial MPM imaging of the same glomerulus over time in the intact Pod-GFP mouse kidney *in vivo*, once in every 24 hours. Glomeruli were identified based on a laser-induced mark placed close to the glomerulus. Z-sectioning was performed at each time point and the z-stacks were compared afterwards. Serial MPM imaging was instrumental in depicting the dynamics and the dramatic changes in the morphology of podocyte projections which often occurred within 24 hours of the previous imaging session. We found evidence for the migration of podocytes away from the multi-cellular clusters, e.g. around the urinary pole into the remainder of the proximal tubule ([Fig. 3a--b](#F3){ref-type=\"fig\"}, [Supplementary Movie 4](#SD5){ref-type=\"supplementary-material\"}). Also, the growth of GFP+ cell projections in the parietal layer was evident by the increasing length and width of the projections ([Fig. 3c--d](#F3){ref-type=\"fig\"}, [Supplementary Movie 5](#SD6){ref-type=\"supplementary-material\"}). In the few nephrons which retained residual glomerular filtration, shedding of podocytes into the proximal tubule was observed ([Supplementary Fig. 4](#SD1){ref-type=\"supplementary-material\"}).\n\nTracking single podocytes in Podocin-Confetti (Pod-Confetti) mice {#S6}\n-----------------------------------------------------------------\n\nBy using Pod-Confetti mice with podocyte specific expression of either membrane-targeted CFP, nuclear GFP, cytosolic YFP or cytosolic RFP (Confetti construct^[@R17]^) and *in vivo* MPM imaging, we were able to identify and track over time individual podocytes based on their specific color ([Fig. 4a](#F4){ref-type=\"fig\"} and [Supplementary Movie 6](#SD7){ref-type=\"supplementary-material\"}). As with Pod-GFP and iPod-GFP mice, UUO induced the bridging of podocytes over the Bowman's space to the parietal layer ([Fig. 4b](#F4){ref-type=\"fig\"}). Adjacent projections in the parietal layer were often labeled by a different color suggesting that multiple visceral podocytes migrate to the parietal Bowman's capsule. Statistical analysis of Confetti+ projections ([Fig. 4c](#F4){ref-type=\"fig\"}) showed that the majority of projections had either a visible, monochromatic bridge between the visceral and parietal layers or the parietal Confetti+ cell matched the color of the closest visceral podocyte (81% in isolated projections, 73% in multiple projections). In addition, serial MPM imaging of the same glomerulus in the UUO kidney ([Fig. 4d--e](#F4){ref-type=\"fig\"}, [Supplementary Movie 7](#SD8){ref-type=\"supplementary-material\"}) provided visual evidence for the appearance of new visceral podocytes within 24h. Also, MPM imaging of Pod-Confetti mice allowed the visualization of the interdigitating podocyte foot processes ([Fig. 4f](#F4){ref-type=\"fig\"}) as described in other genetic approaches^[@R18],[@R19]^.\n\nMPM imaging of PEC migration {#S7}\n----------------------------\n\nTo further elucidate the dynamic interaction between podocytes and PECs, we used *in vivo* serial MPM of PEPCK-GFP mice which express GFP in proximal tubule segments controlled by the PEPCK promoter ^[@R31]^. MPM imaging of PEPCK-GFP mouse kidneys *in vivo* found strong GFP expression in some, but not all PECs in addition to proximal tubule cells, but not in the glomerular tuft ([Fig. 5a](#F5){ref-type=\"fig\"}). Immunofluorescence co-localization experiments confirmed GFP expression in some PECs ([Supplementary Fig. 5](#SD1){ref-type=\"supplementary-material\"}). Occasionally we observed a GFP expressing PEC at the vascular pole of the glomerulus ([Fig. 5b--c](#F5){ref-type=\"fig\"}), which formed projections propagating into both the visceral and parietal layers ([Supplementary Movie 8](#SD9){ref-type=\"supplementary-material\"}).\n\nInterestingly, the strong GFP expression helped to visualize for the first time the presence of extremely thin nanotubules connecting PECs with the glomerular tuft after UUO ([Fig. 5d](#F5){ref-type=\"fig\"}). The width of these cell-to-cell connectors was uniform (\\<350 nm,) and they bridged over long distances between PECs and podocytes (\\>25 \u00b5m) across the entire Bowman's space. Nanotubule connections between PECs were also observed (not shown). The presence of PEC-to-podocyte nanotubules was quite common in the UUO model with an average of 1--4 per glomerulus. Time-lapse MPM imaging revealed ([Supplementary Movie 9](#SD10){ref-type=\"supplementary-material\"}) that the nanotubules were steady over time regardless of variations in hemodynamics, suggesting that these structures are indeed nanotubules rather than long primary cilia. In addition, entire cell bodies appeared occasionally along the length of the nanotubules ([Supplementary Fig. 2c](#SD1){ref-type=\"supplementary-material\"}) suggesting that nanotubules may participate in podocyte migration.\n\nDiscussion {#S8}\n==========\n\nHere we report the development of a new direct visual approach to track migrating podocytes and PECs over time *in vivo* in the same glomerulus of the intact kidney using serial MPM. This represents an important technical advance in nephrology research to study the dynamics and interactions of different glomerular cell types in the intact glomerular environment. The tracking of cell migration and alterations in glomerular morphology may be combined with quantitative MPM imaging of key glomerular functions as we reported recently^[@R32]--[@R34]^. The highly complex and dynamic portrayal of the structure and function of the same glomerulus during the course of a disease is one of the greatest values and advantages of this new MPM imaging technique. Accordingly, its first applications resulted in several new biological and anatomical discoveries. Serial MPM imaging provided visual evidence for podocyte motility and migration to the parietal Bowman's capsule in the intact mouse kidney *in vivo*, which were very robust after UUO. MPM imaging allowed us to visualize the development of multi-cellular podocyte clusters after UUO and adriamycin treatment, and the highly dynamic nature of the migration of the same podocytes and PECs over several days after UUO. Moreover, a new anatomical discovery is the first visualization of nanotubules connecting the visceral (podocytes) and parietal epithelial layers (PECs) which may participate in cell-to-cell communication and cell migration. The combination of serial MPM with mouse models, which allow the identification and tracking of single podocytes over time, will be very useful in future work to study mechanisms of glomerular disease and repair.\n\nThe feasibility of performing *in vivo* MPM imaging of the glomerulus in the intact mouse kidney was demonstrated before ^[@R13]--[@R15],[@R34]^, but this is the first study which utilized serial MPM *in vivo* to track over time the fate of podocytes and PECs that were individually marked using fluorescent lineage tags. Serial MPM complements the powerful classical genetic cell lineage tracing tools^[@R35]^ and can be combined with functional studies, e.g. measurement of the glomerular permeability of macromolecules as demonstrated by our lab recently^[@R32],[@R33]^. Altogether, serial MPM is a significant new addition to the existing arsenal of tools to study podocytes and other glomerular cells^[@R14],[@R18],[@R34]^.\n\nMPM imaging provided important visual clues on podocyte migration, consistent with the importance of podocyte motility and the intact actin cytoskeleton in GFB function as established by a number of recent studies^[@R3]--[@R11]^. Signs of podocyte clustering and/or migration were observed not only in the robust UUO model ([Fig. 1](#F1){ref-type=\"fig\"}--[2c](#F2){ref-type=\"fig\"}), but also in adriamycin nephropathy ([Fig. 2d--f](#F2){ref-type=\"fig\"}) and rarely in control kidneys ([Supplementary Fig. 2b](#SD1){ref-type=\"supplementary-material\"}), suggesting that these are general phenomena not restricted to UUO, and may also be involved in the maintenance of the intact GFB. Serial MPM imaging over several days was instrumental to visualize the robust dynamics of podocyte and PEC migration and remodeling ([Fig. 3](#F3){ref-type=\"fig\"}--[5](#F5){ref-type=\"fig\"}, [Supplementary Movies 4](#SD5){ref-type=\"supplementary-material\"}, [5](#SD6){ref-type=\"supplementary-material\"}, [7](#SD8){ref-type=\"supplementary-material\"}, [8](#SD9){ref-type=\"supplementary-material\"}). It should also be noted that protrusion and migration of these cluster cells into the proximal tubule occur in the absence of filtration ([Fig. 3a--b](#F3){ref-type=\"fig\"}), while podocyte detachment and shedding occur in the presence of mechanical forces ([Fig. 1e](#F1){ref-type=\"fig\"}, [Supplementary Fig. 4](#SD1){ref-type=\"supplementary-material\"}). Based on the present findings, podocyte migration contributes to the pathological remodeling of the glomerulotubular junction and the Bowman's capsule after UUO^[@R29]^. Our data are consistent with the previously described presence of podocytes in the parietal Bowman's capsule in certain pathologies^[@R36],[@R37]^ and rarely in the healthy kidney^[@R38]^.\n\nWith longer periods of UUO, a complete GFP+ cell lining of the parietal Bowman's capsule, and a progressive increase in the percentage of glomeruli with parietal GFP+ cell coverage were observed ([Fig. 1g--i](#F1){ref-type=\"fig\"}). These findings suggest that podocyte clustering and migration are constantly ongoing phenomena, at least in UUO. Although mice with inducible Cre-mediated GFP expression (iPod-GFP mice) showed the same general phenotype as the constitutive Pod-GFP model further confirming podocyte migration ([Fig. 1](#F1){ref-type=\"fig\"}--[2](#F2){ref-type=\"fig\"}), the parietal GFP+ cell coverage developed faster in Pod-GFP versus iPod-GFP mice ([Fig. 1i](#F1){ref-type=\"fig\"}). This was likely due to either imperfect Cre induction^[@R17]^ in iPod-GFP mice, or PEC--to--podocyte transformation in the constitutive Pod-GFP mice. Also, in Pod-Confetti mice with constitutive Cre expression it is possible that a change in podocyte color (\"flipping\") occurs even after the initial Cre recombination^[@R39]^. Therefore, the future use of inducible Cre models is required for cell lineage tracing experiments, while constitutive Cre reporter models may be used for short-term serial MPM imaging of cell migration.\n\nAnother finding was the propagation of PECs observed in the PEPCK-GFP mouse model. We noticed Cre activity in some, but not all PECs in addition to the cells of the proximal tubule ([Fig. 5a](#F5){ref-type=\"fig\"}, [Supplementary Fig. 5](#SD1){ref-type=\"supplementary-material\"}). Consistent with recent reports on the replacement of podocytes by PECs^[@R26],[@R27]^ serial MPM imaging was able to visualize PEC propagation into both the visceral and parietal layers at the vascular pole of the glomerulus ([Fig. 5b--c](#F5){ref-type=\"fig\"}, [Supplementary Movie 8](#SD9){ref-type=\"supplementary-material\"}).\n\nThe visualization of nanotubules interconnecting PECs and podocytes ([Fig. 5d](#F5){ref-type=\"fig\"}) is an exciting anatomical discovery that was made possible by the new highly fluorescent mouse models and the application of MPM *in vivo*. The (patho)physiological relevance of these physical connectors between PECs and podocytes that bridge over long distances within the Bowman's space is yet to be established. Nanotubules may provide an alternative to the vascular pole migration hypothesis put forward by recent reports^[@R26],[@R27]^ and constitute a guiding structure for cell migration directly over the Bowman's space ([Supplementary Fig. 2c](#SD1){ref-type=\"supplementary-material\"}), or could be a unique form of cell-to-cell communication. Nanotubules were reported to allow the unidirectional transfer of specific molecules and organelles between cells^[@R40]^.\n\nIn conclusion, serial MPM imaging of the intact kidney *in vivo* using novel fluorescent lineage tagged mouse models is an exciting new tool to study the development of kidney disease. This imaging approach solves a critical technical barrier in podocyte research and allows the dynamic portrayal of the fate and function of various glomerular cell types *in vivo*. The results of the first applications of this technique may help to change current paradigms in glomerular biology and support the emerging new concept, which emphasizes the dynamic rather than static nature of podocytes, PECs and the entire glomerular environment. Serial MPM imaging has tremendous potential and utility in the future development and testing of new regenerative therapeutic approaches.\n\nMaterials & Methods {#S9}\n===================\n\nAnimals {#S10}\n-------\n\nBoth male and female C57BL6/J mice at the age of 4--12 weeks were used randomly. Four new fluorescent reporter mouse models were generated: (i) Pod-GFP by crossing mice expressing Cre recombinase under the control of the podocin promoter^[@R21]^ and animals with a Tomato^floxSTOP^EGFP sequence in the ROSA 26 locus (B6.129(Cg)-*Gt(ROSA)26Sor^tm4(ACTB-tdTomato,-EGFP)Luo^*)^[@R16]^. This results in an expression of Tomato in all cells except in podocytes, which express GFP. To study the effect of different backgrounds those mice were used in a C57BL6 background, a FvB background or the F1 generation crossed between these backgrounds. (ii) iPod-GFP by crossing mice expressing tamoxifen-inducible improved Cre recombinase under the control of the podocin promoter (podocin-iCreER^T2^mice^[@R22]^, kind gift from Dr. Farhad R. Danesh, Baylor College of Medicine) and animals with the Tomato^floxSTOP^EGFP construct mentioned above. (iii) PEPCK-GFP by crossing mice expressing Cre recombinase under the control of the phosphoenolpyruvate-carboxykinase promoter (PEPCK-Cre, kind gift from Dr. Volker Haase, Vanderbilt University)^[@R31]^ and animals with the Tomato^floxSTOP^EGFP construct mentioned above. (iv) Pod-Confetti by crossing mice expressing the R26R-Confetti construct^[@R17]^ and podocin-Cre mice^[@R21]^ resulting in the expression of either membrane-targeted CFP, nuclear GFP, cytosolic YFP or cytosolic RFP in podocytes. All animals were purchased from the Jackson laboratory (Bar Harbor, ME) or otherwise noted. All animal protocols were approved by the Institutional Animal Care and Use Committee at the University of Southern California or by local government authorities in Germany (LANUV NRW) under the license 8.87--50.10.31.08.049.\n\nTamoxifen induction {#S11}\n-------------------\n\n4 weeks-old mice received Tamoxifen chow (40 mg/kg body weight, Harlan Laboratories, Indianapolis, IN) for 2 weeks, followed by a 2 week washout period before additional procedures were performed.\n\nUUO {#S12}\n---\n\nBetween 3 and 8 weeks of age the animals were anaesthetized with isoflurane, and after a midline laparotomy the left ureter was exposed and ligated three times. Successful ligation was confirmed by the hydronephrotic distension of the kidney at the time of imaging.\n\nAdriamycin nephropathy {#S13}\n----------------------\n\nPod-GFP animals received a single dose of 25mg/kg Adriamycin (Doxorubicin) via retroorbital injection and were imaged, sacrificed and their tissues collected 4--6 days later. Successful induction of albuminuria was confirmed by measuring the urinary albumin/creatinine ratio (Exocell, Philadelphia, PA) between days 0--6 after adriamycin treatment.\n\nMultiphoton imaging {#S14}\n-------------------\n\nThe animals were anaesthetized with a combination of ketamine (100 mg/kg) and xylazine (10 mg/kg). A tracheal tube was placed to facilitate breathing and the right carotid artery and/or jugular vein was cannulated for dye infusion. A 70kDa Texas red dextran or Alexa 594 bovine serum albumin was injected to label the vasculature. In some experiments Lucifer Yellow (Invitrogen) was injected as iv bolus to visualize glomerular filtration. The left kidney was exteriorized through a flank incision and the animal placed on the microscope stage as described before^[@R34]^. Body temperature was maintained with a homeothermic blanket system (Harvard Apparatus). The images were acquired using a Leica TCS SP5 multiphoton confocal fluorescence imaging system with a 63x Leica glycerine-immersion objective (NA 1.3) powered by a Chameleon Ultra-II MP laser at 860 nm (Coherent) and a DMI 6000 inverted microscope's external nondescanned detectors. Short pass filters (680 nm for blue and red, 700 nm for green and yellow), dichroic mirrors (cut off at 515 nm for green and yellow, 560 nm for blue and red) and bandpass filters were specific for detecting CFP/GFP/YFP/RFP emission (473nm/514nm/545nm/585nm) (Chroma). The images of iPod-GFP animals were acquired using the external detectors of an inverted Zeiss LSM710 NLO multiphoton confocal fluorescence microscope powered by a Chameleon Ultra-II MP laser at 860 nm and a 40x Zeiss water-immersion objective (NA1.2).\n\nSerial survival imaging of the same glomerulus {#S15}\n----------------------------------------------\n\nAfter anaesthesia the left kidney was exteriorized via a small cut in the left flank below the kidney to avoid sutures right above the kidney afterwards. To avoid invasive vascular access surgeries, the dyes were administered by retroorbital injections. The animal was transferred to the microscope stage and the exteriorized kidney was placed into a kidney cup. An area of the kidney suitable for imaging was identified and the position of the kidney was noted for identical placement on the following days. After acquiring z-stacks of the glomerulus a small distant area in the field of view was marked by shortly focusing the laser beam on this area with high power. This maneuver generated an easy to find highly fluorescent spot (reference point) which remained there for 3--5 days. The position of the mark relative to the glomerulus of interest was documented. After imaging, the kidney was placed back into the retroperitoneum and the flank cut was closed with two layers of sutures. This procedure was repeated 24, and 48 hours later by removing the sutures and exteriorizing the kidney again. With this technique we were able to subsequently find approximately 70% of the glomeruli that were marked in the first imaging session. Z-stacks of marked glomeruli were acquired with identical imaging settings as the day before.\n\nThe potential toxicity of laser excitation and fluorescence to the cells were minimized by using a low laser power and high scan speeds to keep total laser exposure as small as possible. The usual image acquisition consisted of only one z-stack per glomerulus (\\<1 minute) per 24 hours which resulted in no apparent cell injury. Serial imaging once every 24 hours for up to 72 hours after the first imaging session and the associated multiple, dorsal abdominal surgeries did not cause visible tissue adhesions or renal fibrosis.\n\nZ-stack analysis {#S16}\n----------------\n\nZ-stacks from different time points were aligned using StereoMovie Maker (<http://stereo.jpn.org>) to visualize changes over time in a side-by-side fashion, see Supplementary movies.\n\nImmunohistochemistry {#S17}\n--------------------\n\nAnimals were perfused with ice cold PBS into the left ventricle followed by ice cold 4% PFA for 2 minutes each. Antigen retrieval was performed on paraffin sections with a heating step for 8 min at 95\u00b0C in a citrate buffer using a pressure cooker. The sections were incubated with the following primary antibodies (1:100 dilution unless stated otherwise) overnight: anti-synaptopodin (mouse, Progen, Heidelberg, Germany), anti-podocin (1:250, rabbit, Sigma-Aldrich, St. Louis, MO), anti-claudin-1 (1:250, rabbit, Sigma), anti-villin (mouse, Immunotech, Chicago, IL), anti-GFP (chicken, Aves, Tigard, OR) and anti-GFP (rabbit, Invitrogen) followed by incubation with the secondary antibodies conjugated with Alexa 488/594. Confocal fluorescence microscopy was performed using the same Leica TCS SP5 microscope as above.\n\nStatistical analysis {#S18}\n--------------------\n\nData are expressed as average \u00b1 s.e.m. and were analyzed non-blinded using Student's t-test or linear regression (R value) as indicated. P\\<0.05 was considered significant. For the statistical analysis of visceral to parietal projections in Confetti-UUO mice, multiple projections were defined as projections in which parietal Confetti+ cells had direct contact with other projections (n=167 isolated projections, n=217 multiple projections, from a total of n=112 glomeruli analyzed from 4 mice).\n\nSupplementary Material {#SM}\n======================\n\nThis work was supported in part by the US National Institutes of Health grant DK64324, by the American Diabetes Association grant 1-11-BS-121, by the University Kidney Research Organization to J.P-P, by Deutsche Forschungsgemeinschaft SFB635 to T.B, and by US National Institutes of Health grant DK076077 to K.S. M.J.H. was supported by Deutsche Forschungsgemeinschaft grant HA-6212. We thank Matthias Hammerschmidt and Hans-Martin Pogoda (Institute of Developmental Biology, University of Cologne) for granting access to their imaging facility.\n\n**Contributions**\n\nJ.P-P. conceived the study. J.P-P., M.J.H., and J.L.B. designed the experiments, coordinated the study and analyzed the data. J.P-P., M.J.H. and T.B. wrote the manuscript. M.J.H., J.L.B., L.L., K.V., B.S. and K.S. performed the experiments.\n\n**Conflict of interest/disclosure**\n\nNone of the authors has any conflicts of interest to declare.\n\n![MPM imaging *in vivo* reveals signs of podocyte migration in the intact kidney in the model of unilateral ureteral obstruction (UUO) in Pod-GFP mice\\\n(a) In a wild-type mouse kidney the glomerular podocytes appear as dark, unlabeled cells around capillaries (arrows). Plasma is labeled red with Alexa594-Albumin. Lucifer yellow (yellow) was injected iv to label the filtration (Bowman's) space. (b) Low-power image of glomeruli after UUO shows restriction of GFP fluorescence to podocytes in Pod-GFP mice. Early podocyte changes that appear after UUO include the development of podocyte cell clusters (c) and their projections protruding (arrow) into the lumen of the proximal tubule (d). (e) Detached, individual podocytes can be seen in the lumen of the proximal tubule. GFP+ cells in the parietal layer appear to develop by visceral podocytes migrating to the parietal layer via the vascular pole transition (f) or directly over the Bowman's space. (g) Thin connections appear to bridge the parietal and visceral layers of the Bowman's capsule (arrows). Continuous GFP+ cell coverage in the parietal layer (g) develops later, \\>3--4 weeks after UUO. Scale bars are 20\u00b5m. Magnification is the same on all panels except panel b. (h) Percentage of glomeruli in which \\>50% of the parietal layer is covered by GFP+ cells in age-matched control mice (n=19) and 2--5 weeks after UUO (n=43). \\*: P\\<0.05. Data are presented as mean \u00b1 s.e.m. (i) Linear fit of data points (Pod-GFP in black (n=48 mice), iPod-GFP in red (n=5 mice) shows the percentage of glomeruli with \\>50% parietal GFP+ cell coverage.](nihms530039f1){#F1}\n\n![Confirmation of podocyte clustering and migration in iPod-GFP mice after UUO (a--c) and podocyte clustering in the model of adriamycin nephropathy in Pod-GFP mice (d--f)\\\nIn mice with tamoxifen-induced podocyte GFP expression (iPod-GFP) podocyte clustering was observed \\<2 weeks after UUO (a, arrow). After \\>4 weeks following UUO, podocyte projections to the parietal Bowman's capsule developed either at the urinary pole invading the glomerulo-tubular junction (b, arrow), or at the vascular pole transition (c, arrowheads), or anywhere in between the poles (c, arrow) giving rise to parietal GFP+ cells. Plasma was labeled red with Alexa594-Albumin. MPM imaging of the intact Pod-GFP mouse kidney after adriamycin treatment revealed the presence of podocyte clusters around capillaries (d, arrowhead) and the development of GFP+ cell coverage in the parietal layer which was continuous with the podocyte clusters via the vascular pole transition (e, arrow). Scale bar is 20\u00b5m for all panels. (f) Summary of histological analysis showing the percentage of glomeruli with podocyte clusters in age-matched control mice and 4 days after adriamycin treatment (n=4 each). \\*: P\\<0.05. Data are presented as mean \u00b1 s.e.m.](nihms530039f2){#F2}\n\n![Serial *in vivo* MPM imaging of the same glomerulus in Pod-GFP mice after UUO over time, once in 24 hours\\\n(a--b) Podocytes from a hypercellular area at the urinary pole of a collapsed, non-filtering glomerulus (a, arrow) appear to migrate away from the capillary tuft to form a projection into the remainder of the proximal tubule (b, arrow, 24 h later). (c--d) Demonstration of the increased size (growth) of the GFP+ cell projections in the parietal layer (d, arrows) 24 hours after the previous imaging session (c). Plasma is labeled red with Alexa594-Albumin. Scale bar is 20\u00b5m for all panels.](nihms530039f3){#F3}\n\n![Identification and tracking of single podocytes in the multi-color Pod-Confetti mouse model using *in vivo* MPM imaging\\\n(a) Podocytes are labeled in one of the four colors, either by membrane-targeted CFP, nuclear GFP, cytosolic YFP or cytosolic RFP. Labeled dextran (plasma dye) is shown in grayscale. (b) A CFP-labeled (blue) visceral podocyte is shown bridging over the Bowman's space to the parietal layer and in direct contact with other parietal cells of the same color (blue arrow). Other adjacent parietal Confetti+ cells are labeled with RFP (red arrow) or GFP (green arrow). (c) Analysis of the Confetti+ cell projections between the visceral and parietal layers shown for both single, isolated projections and for multiple projections in which parietal Confetti+ cells had direct contact with other projections (n=167 single and n=217 multiple projections from 112 glomeruli in 4 mice). (d--e) Serial MPM imaging of the same glomerulus shows the appearance of a nuclear GFP-labeled podocyte around a glomerular capillary (e) compared to 24h prior (d). Arrows indicate the same glomerular tuft region). d--e are projection images of 8 different confocal z-planes. (f) Cell morphology of podocytes including the interdigitating foot processes between two adjacent CFP and YFP-labeled podocytes are visualized in great detail (fixed, frozen tissue section). Scale bar is 20\u00b5m for all panels except panel f (1\u00b5m).](nihms530039f4){#F4}\n\n![MPM imaging of PEC migration *in vivo* in the PEPCK-GFP mouse model after UUO\\\n(a) Cell-specific GFP expression in the proximal tubule and in some, but not all PECs. All other cells are labeled with Tomato. (b--c) Serial MPM imaging of the same glomerulus shows a GFP expressing PEC at the vascular pole of the glomerulus that forms projections propagating into both the visceral and parietal layers. The length of the green projections increased between the time points of 0h (b), and 48 h (c). (d) Four green nanotubules (arrows) are visualized in the intact glomerulus, which appear to connect the PEC layer with the glomerular tuft. Scale bar is 20\u00b5m for all panels.](nihms530039f5){#F5}\n"}
+{"text": "Introduction {#Sec1}\n============\n\nQuestionnaires for health-related quality-of-life (HRQoL) measurement are important for several reasons. First, they may be used to compare the mean level of different patient groups with respect to physical, mental and social health. A researcher may want to find out whether these patient groups have different needs with respect to, for example, therapy or medication or whether different adaptations of their environment are in order so as to improve their conditions of living. Second, HRQoL questionnaires are also important for the measurement of mean change----either progress or deterioration----of such groups due to, for example, therapy. The researcher's interest then lies in the effectiveness of therapy with respect to HRQoL. Third, the total score a patient obtains on an HRQoL questionnaire may be used to diagnose this patient's general level of physical health and psychological well-being, for example, so as to be able to estimate the budget needed for his/her treatment during a particular period.\n\nTo effectively measure HRQoL, we argue that an instrument must meet two requirements. The first requirement is that it is clear what the instrument measures: one overall dimension of HRQoL or several dimensions reflecting different aspects of HRQoL. If the instrument measures one dimension, one can use the total score on all items to obtain an impression of the overall level of HRQoL. If the instrument measures multiple dimensions, it may be recommendable to determine total scores on subsets of items (e.g., domain scores), each reflecting a particular aspect of HRQoL (e.g., HRQoL with respect to physical, psychological, and social limitations) and then assess individuals or compare groups on a profile of scores. These two cases may be characterized as unidimensional and multidimensional measurement.\n\nThe second requirement is that the psychometric properties of the items are known and found sufficient. One important psychometric item property is the item's location on the scale that quantifies the HRQoL aspect of interest. For example, patients are likely to experience fewer problems when engaging in activities like bathing and dressing than in more demanding activities such as shopping and travelling. The items concerning bathing and dressing require a lower level of physical functioning than the other two items. Thus, bathing and dressing are located further to the left (at a lower level of the scale) than shopping and travelling. A good diagnostic HRQoL instrument contains items of which the locations are widely spread along the scale. Such a scale allows for measurement at varying levels of physical functioning and may be used, for example, for assessing mean differences between groups, mean change due to therapy, and individual patients' levels of physical functioning.\n\nAnother important psychometric item property is the item's discrimination power. This is the degree to which the item distinguishes patients with relatively low psychological well-being levels from patients with relatively high psychological well-being levels. The higher the discrimination power the higher the item's contribution to reliable measurement (\\[[@CR1]\\], pp. 101--124). A good diagnostic instrument has items with high discrimination power that each contributes effectively to reliable measurement of patients at different locations along the scale. Such an instrument picks up differences between groups, effects of therapy, and individual levels of activity limitation.\n\nItem response theory (IRT) models \\[[@CR1]\\] are becoming more popular as statistical tools for scale construction in the HRQoL context. IRT can be used effectively to investigate the dimensionality of an instrument and the psychometric properties of its constituent items. The goal of this study is to discuss one particular class of IRT models known as *nonparametric* IRT models \\[[@CR2]--[@CR5]\\], and to argue that this class in particular provides a general and flexible data analysis framework for studying the dimensionality of a set of polytomously scored items (with dichotomously scored items as special cases) and ascertaining ordinal scales for the measurement of HRQoL aspects which contain items that have varying locations and sufficient discrimination power. Over the past few years, nonparametric IRT models already have been used occasionally for constructing HRQoL scales; see \\[[@CR6]--[@CR9]\\].\n\nOur point of view is that, given that the researcher has formulated desirable measurement properties, (s)he should construct his/her scale by means of an IRT model that is as general as possible while satisfying the desired measurement properties. Examples of such properties are that the items measure the same dimension, that the measurement level is at least ordinal, and that measurement values are reliable. An HRQoL researcher who has constructed and pre-tested a questionnaire consisting of, say, 40 items is not served well when his/her data are analyzed by means of an IRT model that is unnecessarily restrictive, the result of which is that, say, half of the items are discarded. We will argue that the most general IRT model that serves one's purposes well, often (but not always) is a nonparametric IRT model.\n\nMany questionnaires are used for assessing differences in HRQoL between groups, change due to therapy, and individual patients' scale levels, and a general nonparametric IRT model then is the perfect choice for analyzing one's data. Nonparametric IRT models have several advantages over more-restrictive *parametric* IRT models \\[[@CR1]\\]: Nonparametric IRT models (1) are based on less-restrictive assumptions, thus they allow more items into the scale while maintaining desirable measurement properties; (2) offer diverse tools for HRQoL analysis that give ample information about the dimensionality of the data and the properties of the items; and (3) provide patient measurement values and item location and discrimination values, which have an interpretation that is close to intuition and therefore easy to interpret for users of HRQoL scales. For computerized adaptive HRQoL testing, more-restrictive parametric IRT models such as the Rasch \\[[@CR10]\\] model (dichotomous items) and the generalized partial credit model \\[[@CR11]\\] and the rating scale model \\[[@CR12]\\] (polytomous items) are more appropriate than nonparametric models.\n\nParametric IRT models have been used more than nonparametric IRT models, especially in psychological and educational measurement, and also in HRQoL research (e.g., \\[[@CR13]--[@CR15]\\]). A reason for this may be that nonparametric IRT models were developed later than parametric IRT models. See \\[[@CR16]--[@CR18]\\] for reviews of nonparametric IRT.\n\nThis paper is organized as follows. First, we explain assumptions of IRT and compare parametric IRT and nonparametric IRT. Second, we discuss methods and software from nonparametric IRT that can be used for analyzing the polytomous item scores obtained from HRQoL questionnaires. Third, we use this software to analyze data from the World Health Organization Quality-of-Life Scale (WHOQOL-Bref) \\[[@CR19]\\]. The results are compared to those obtained by means of a parametric IRT model. Finally, we provide recommendations for HRQoL researchers on how to use nonparametric IRT methods for analyzing their data.\n\nDefinitions and assumptions {#Sec2}\n===========================\n\nIRT models are suited for the analysis of multi-item questionnaire data that typically result from HRQoL questionnaires. The data are discrete scores representing the responses of *N* respondents to *J* items (items are indexed *j*; *j\u00a0*=\u00a01, \\..., *J*). Many HRQoL questionnaires use items that have three or more ordered answer categories represented by three or more ordered scores, also called polytomous item scores. For simplicity, we assume that all items have the same number of ordered answer categories; this number is denoted by *M*\u00a0+\u00a01. Let *X*~*j*~ be the random variable representing the discrete score on item *j*, and let the items be scored *X*~*j*~\u00a0=\u00a00,\\...,*M*. For example, for an item asking whether one is satisfied with one's sleep, score 0 may represent '*very much dissatisfied*' and score *M* may represent '*very much satisfied*', and the intermediate scores represent intermediate levels of satisfaction. With dichotomous item scoring, *X*~*j*~\u00a0=\u00a00,1, and possible intermediate satisfaction levels are not quantified separately but collapsed into the score categories 0 and 1. Usually, researchers summarize the *J* item scores for each patient by the total score (sometimes referred to as the sum score), which is formally defined as $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ X_ + = \\sum\\nolimits_{j = 1}^J {X_j } $$\\end{document}$. Total score *X*~+~ is an estimate of a patient's true score *T*; this is the expectation of *X*~+~ across independent replications of the measurement procedure (\\[[@CR20]\\], pp. 29--30).\n\nIRT models distinguish observable or manifest variables such as item score *X*~*j*~ and total score *X*~+~ from latent variables. These latent variables play the role of summaries of the behavior that is described by the responses to the items. Sometimes, latent variables are interpreted as if they were causal agents driving responses to items and individual differences between patients. We will also use the distinction between latent and manifest variables in our examples. Thus, we assume that patients are characterized by either one latent HRQoL attribute (meaning that measurement is unidimensional) or different HRQoL attributes (meaning that measurement is multidimensional) which together represent the patient's latent physical, mental or social health. For example, assume that measurement is unidimensional and that the latent variable is mental health or psychological well-being. Psychological well-being then is an unobservable state in each patient, and inferences about it are made on the basis of the manifest responses reported by patients in reaction to the items in an HRQoL questionnaire. Latent variables are denoted by notation *\u03b8*. If measurement is unidimensional, the IRT model contains one latent variable*\u03b8*, and if it is multidimensional multiple latent variables are needed. We only consider unidimensional IRT models here; see \\[[@CR21]\\] for a discussion of multidimensional IRT models.\n\nSeveral families of IRT models for polytomous item scores have been proposed (e.g., \\[[@CR22]\\]). The family of graded response models (GRMs; \\[[@CR23]\\]) is suitable for analyzing ordered item scores collected by means of polytomous response scales \\[[@CR22], [@CR24], [@CR25]\\]. Suppose the item 'How much do you enjoy life?' has four ordered answer categories running from 'Not at all' (score 0) to 'Very much' (score 3); thus *X*~*j*~\u00a0=\u00a00,\\...,3 and *M\u00a0*=\u00a03*.* For this item, GRMs conceptualize the response process by means of four conditional response probabilities, called item-step response functions (ISRFs) and denoted by *P*(*X*~*j*~\u00a0\u2265\u00a0*m*\\|*\u03b8*)=*P*~*jm*~(*\u03b8*) (*m\u00a0*=\u00a00,\\...,*M*). It may be noted that for *m\u00a0*=\u00a00 we have that P~*j*0~(*\u03b8*)\u00a0=\u00a01 by definition, irrespective of the latent variable level; thus, this response probability is uninformative about the response process and may be ignored. The ISRF describes the relationship between expressing at least a particular minimum level of enjoying life (i.e., having at least a score of *m* on the example item) and the latent variable of psychological well-being (*\u03b8*).\n\nFigure\u00a0[1](#Fig1){ref-type=\"fig\"}a shows an example of the ISRFs of two items having four answer categories each (i.e., *M\u00a0*=\u00a03). The solid lines denote the ISRFs for one item and the dashed lines denote the ISRFs for the other item. Two things are noteworthy. First, the ISRFs of different items may intersect but based on the cumulative character of the definition of the ISRFs, the ISRFs of the same item cannot intersect. Second, the ISRFs have been drawn as monotone curves with rather irregular shapes. Such shapes are typical of a nonparametric GRM \\[[@CR24]\\] and found not to prevent a set of items from having favourable measurement properties, as we will see shortly. Figure\u00a0[2](#Fig2){ref-type=\"fig\"}b also shows monotone ISRFs but now having smooth S-shapes, typical of a parametric GRM. This parametric GRM nearly has the same measurement properties as the nonparametric GRM, but because it assumes smooth ISRFs (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b) instead of irregular ISRFs (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a), it is more restrictive and often leads to the rejection of more items (and thus shorter scales) than its nonparametric counterpart.Fig.\u00a01Examples of item step response functions of (**a**) a nonparametric item response model and (**b**) a parametric item response modelFig.\u00a02Examples of (**a**) flat ISRFs and (**b**) non-monotone ISRFs\n\nGRMs have the next three assumptions as a point of departure. The first assumption is unidimensionality (UD); that is, each model assumes that one latent variable *\u03b8* summarizes the variation in the *J* item scores in the questionnaire. Assumption UD implies that respondents can be ordered meaningfully by means of a single number. The second assumption is local independence (LI); that is, if we condition on *\u03b8*, the *J* item scores are statistically independent. An implication of LI is that in a subgroup of patients who have the same *\u03b8* value, all covariances between item scores are 0. The third assumption is monotonicity (M); that is, the ISRFs are each assumed to be monotone increasing functions in *\u03b8* (see Figs.\u00a0[1](#Fig1){ref-type=\"fig\"}a and b). Applied to the latent variable of psychological well-being (*\u03b8*) and the items from the WHOQOL-Bref, assuming UD, LI and M means that we hypothesize that (1) only psychological well-being drives responses to items and has a systematic effect on individual differences in item scores and total scores (UD); (2) given a fixed level of psychological well-being, relationships between concrete aspects of psychological well-being such as represented by the items 'How much do you enjoy life?' and 'Are you able to concentrate?' are explained completely (i.e., the covariance between these items, conditional on *\u03b8*, equals 0) (LI); and (3) the higher the level of psychological well-being, the higher the probability that one enjoys life and is able to concentrate.\n\nParametric and nonparametric graded response models {#Sec3}\n===================================================\n\n*Parametric Graded Response Model.* The ISRFs of the parametric GRM \\[[@CR26]\\] are defined by logistic functions that have the following parameters:*\u03b4*~*jm*~: the location parameter of the *m*th ISRF of item *j* (i.e., *P*~*jm*~(*\u03b8*)) on the scale of *\u03b8*;*\u03b1*~*j*~: the slope parameter or 'discrimination power' of item *j*.\n\nThe meaning of these parameters is explained after the ISRF of the parametric GRM is introduced. This ISRF is defined as\n\n$$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ P_{jm} (\\theta ) = \\frac{{\\exp [\\alpha _j (\\theta - \\delta _{jm} )]}} {{1 + \\exp [\\alpha _j (\\theta - \\delta _{jm} )]}},\\;\\;\\;\\;j = 1{\\text{ }},...,{\\text{ }}J;\\;m = 1,...,M. $$\\end{document}$$\n\nFigure\u00a0[1](#Fig1){ref-type=\"fig\"}b shows the logistic ISRFs of two items (*M\u00a0*=\u00a03). The two items are denoted by *j* (solid ISRFs) and *k* (dashed ISRFs). For each item, the three location parameters are also shown. For item *j*, by definition we have that *\u03b4*~*j*1~\u00a0\\<\u00a0*\u03b4*~*j*2~\u00a0\\<\u00a0*\u03b4*~*j*3~, and for item *k* by definition *\u03b4*~*k*1~\u00a0\\<\u00a0*\u03b4*~*k*2~\u00a0\\<\u00a0*\u03b4*~*k*3~. An ISRF's location parameter is the value of *\u03b8* for which the probability of having an item score of at least *m* equals .5: that is, *P*~*jm*~(*\u03b8*)\u00a0=\u00a0*P*~*km*~(*\u03b8*)\u00a0=\u00a0.5, *m\u00a0*=\u00a01,\\..., *M*.\n\nFigure\u00a0[1](#Fig1){ref-type=\"fig\"}b also shows that the slopes of the ISRFs of the same item are equal (mathematically, they must be equal or the ISRFs would intersect; this is impossible given the cumulative definition of the ISRFs), but also that the ISRFs of item *j* are steeper than the ISRFs of item *k*. Steepness of slopes is evaluated as follows. For ISRF *m* of item *j*, consider the point with coordinates (*\u03b4*~*jm*~, .5). This is the point in which the slope of a logistic ISRF is steepest, and this steepest slope is taken to be typical for the whole ISRF. Parameter *\u03b1*~*j*~ expresses this maximum steepness (but is not exactly equal to it). In the example in Fig.**\u00a0**[1](#Fig1){ref-type=\"fig\"}b, we have that *\u03b1*~*j\u00a0*~\\>\u00a0*\u03b1*~*k*~.\n\n*Nonparametric Graded Response Model.* Instead of choosing a parametric function, nonparametric IRT models typically define order restrictions on the ISRFs. The nonparametric GRM, better known as the monotone homogeneity model (MHM) for polytomous items \\[[@CR27], [@CR4]\\], assumes UD, LI, and M: that is, for any pair of *\u03b8*s, say, *\u03b8*~*a*~ and *\u03b8*~*b*~, the MHM assumes that\n\n$$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ P_{jm} (\\theta _a ) \\le P_{jm} (\\theta _b ),\\;\\;\\;\\;{\\text{whenever }}\\theta _a < \\theta _b . $$\\end{document}$$\n\nThus, the ISRF is monotone non-decreasing in *\u03b8*; see Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a for examples of ISRFs that are monotone but not logistic. This assumption says that a higher level of psychological well-being induces a higher probability of obtaining at least an item score of *m* (i.e., a higher item score). The ISRFs of different items can have any monotone form and be very different. Requiring monotone ISRFs only is less restrictive than requiring monotone logistic ISRFs; thus, the MHM is a more general model for describing the data than the GRM (henceforth, we call the nonparametric GRM by its better known name (in fact, acronym) MHM, and the parametric GRM simply the GRM).\n\nUnlike the GRM, the MHM does not provide numerical estimates of the latent variable *\u03b8* . Instead, the MHM allows that total score *X*~+~ orders patients stochastically on latent variable *\u03b8* in almost all practical measurement situations \\[[@CR28]\\]. This means that, for two total scores *X*~+~ denoted *v* and *w*,\n\n$$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ E(\\theta |X_ + = v) \\le E(\\theta |X_ + = w),\\;\\;\\;\\;{\\text{for }}0 \\le v < w \\le J, $$\\end{document}$$\\[[@CR4]\\]. This inequality says that as the total score increases the mean *\u03b8* also increases (or stays the same). Thus, groups of patients that have higher total scores, on average also have higher latent variable values. This result may not seem spectacular at first sight, but it (1) ascertains an ordinal scale for patient measurement (2) using only observable total scores (without requiring the actual estimation of *\u03b8*). For the psychological well-being example, if the MHM fits the data, ordering patients by means of the total score by implication orders them on the latent variable *\u03b8*.\n\nAlso, the MHM does not provide numerical estimates of the item parameters *\u03b4* and *\u03b1*. Instead, a distinction can be made between drawing information about item functioning from estimates of the complete ISRFs and item parameters typical of the MHM. Estimates of the complete ISRFs provide much information about the exact relationship between the item scores and the latent variable \\[[@CR16], [@CR29]\\]. ISRFs that are relatively flat or fail to be monotone can be studied in much detail so as to reveal why they dysfunction.\n\nFigure\u00a0[2](#Fig2){ref-type=\"fig\"}a shows three relatively flat ISRFs of a hypothetical item (*M\u00a0*=\u00a03). This item does not distinguish low *\u03b8* and high *\u03b8* patients well. It may be noted that the ISRFs do not all need to be flat simultaneously, but given that they cannot intersect if one ISRF is flat others are likely to be relatively flat as well and the item as a whole contributes little to the reliable ordering of patients on *\u03b8*.\n\nFigure\u00a0[2](#Fig2){ref-type=\"fig\"}b shows three non-monotone ISRFs of another hypothetical item. Each shows relative good distinction between low *\u03b8* values and between above-average *\u03b8* values but bad distinction just below the middle of the scale. Again, non-intersection of the ISRFs of the same item implies that often several ISRFs simultaneously show such disturbing non-monotonicities. For the example item one may conclude that when the questionnaire contains few items, which are effective in the high *\u03b8* area, this item may be retained to cover this area even though this would be at the expense of measurement quality just below the middle of the scale.\n\nFor each item, the MHM framework provides *M* location parameters and a scalability coefficient, which provides information about item discrimination. The location parameters are the proportions of the population of interest, which have at least a score *m* on item *j*, and which are denoted by *\u03c0*~*jm*~, *m\u00a0*=\u00a01, \\..., *M*. For the same item, due to non-intersection of ISRFs we have that *\u03c0*~*j*1~\u00a0\u2265\u00a0 \\...\u00a0\u2265\u00a0*\u03c0*~*jM*~, whereas in the GRM item location parameters are ordered oppositely, *\u03b4*~*j*1~\u00a0\u2264\u00a0\\...\u00a0\u2264\u00a0*\u03b4*~*jM*~. The item scalability coefficient *H*~*j*~ (e.g., \\[[@CR2]\\], pp. 148--153; \\[[@CR18]\\], chap. 4) summarizes the discrimination power of an item across its *M* ISRFs. The numerical *H*~*j*~ value is determined by the interplay of the slope and the location of the ISRFs of all *J* items and the distribution of the latent variable *\u03b8* \\[[@CR30]\\], and it expresses how well item *j* separates patients given the ISRFs of item *j* relative to the other items' ISRFs and the distribution of *\u03b8*. Mathematically, holding constant this distribution and the location of all items' ISRFs, coefficient *H*~*j*~ is higher when the slopes of ISRFs of item *j* are steeper \\[[@CR18]\\].\n\nNonparametric versus parametric graded response models {#Sec4}\n======================================================\n\n*Relationships Among Models*. The MHM shares assumptions UD, LI and M with the GRM, but the MHM is less restrictive with respect to the shape of the ISRFs. Thus, the MHM is more general than the GRM or, equivalently, the GRM is a special case of the MHM. This hierarchy implies that, if the GRM fits the data, by implication the more general MHM also fits but if the MHM fits the data, this does not imply that the GRM also fits. Fit of the GRM then needs to be investigated separately. Because of the hierarchical relationship, for any data set the MHM fits as least as many items as the GRM (Table\u00a0[1](#Tab1){ref-type=\"table\"}).Table\u00a01Comparison of monotone homogeneity model (MHM) and graded response model (GRM)Nonparametric IRT (MHM)Parametric IRT (GRM)Restrictiveness of modelsLow; many items admitted to the scaleHigh; fewer items admitted to the scaleInterpretation of parametersIntuitively appealingMore-complicatedParameters(*typical range*)\u00a0\u00a0\u00a0\u00a0Person level*T*, *X*~+~*\u03b8*(\u22123\u00a0\u2264\u00a0*\u03b8\u00a0*\u2264\u00a03)\u00a0\u00a0\u00a0\u00a0ISRF location*\u03c0*~*jm*~\u03b4~*jm*~(\u22123\u00a0\u2264\u00a0*\u03b4*~*jm*~\u00a0\u2264\u00a03)\u00a0\u00a0\u00a0\u00a0ISRF discrimination*H*~*j*~*\u03b1*~*j*~(0.5\u00a0\u2264\u00a0*\u03b1*~*j*~\u00a0\u2264\u00a02.5)Data analysisExploratory, data as point of departureConfirmatory, model as null hypothesisApplicationsComparing groups\\\nMeasuring change\\\nDiagnosing patientsComparing groups\\\nMeasuring change\\\nDiagnosing patients\\\nConstructing item banks\\\nAdaptive testing\n\n*Patient and Item Parameters*. The MHM and the GRM provide the following patient and item parameters (also, see Table\u00a0[1](#Tab1){ref-type=\"table\"}):For patient measurement, the MHM uses total score *X*~+~ to order patients on latent variable *\u03b8*. Because total score *X*~+~ has an easy interpretation and, moreover, in many IRT models *X*~+~ and *\u03b8* tend to correlate extremely high suggesting a strong linear relationship \\[[@CR31]\\], total score *X*~+~ may be preferred in practice. Total score *X*~+~ is the sum of the rating scale scores on the *J* items, whereas estimates of *\u03b8* are expressed on a logit scale, which does not have a straightforward interpretation for users of HRQoL scales. In general, the ordinal relationship of *\u03b8* with total score *X*~+~ (which can be approximated well by a linear relation) enables users to switch between scales, and use the one that suits their goals best.Because item location *\u03b4*~*jm*~ is expressed on the same scale as latent variable *\u03b8*, it also has an interpretation in logits. For many users, proportion *\u03c0*~*jm*~, the proportion of patients who have at least an item score of *m*, has an easier interpretation.Item discrimination *\u03b1*~*j*~ gives the maximum slope of the logistic ISRF irrespective of the locations of the other ISRFs of item *j* and the other items in the questionnaire, and irrespective of the distribution of *\u03b8*. Thus, information on ISRF slopes is absolute in the sense that a particular *\u03b1*~*j*~ value does not provide information on the item's suitability for measurement in a particular group (characterized by a particular distribution of *\u03b8*) by means of a set of *J* items (characterized by particular location and slope parameters). On the other hand, item scalability coefficient *H*~*j*~ depends explicitly on the interplay of the distribution of *\u03b8*, the spread of locations of the ISRFs, and the slopes of the ISRFs. In particular, keeping two of these factors fixed, *H*~*j*~ tends to increase in the third. This dependence on the distribution of *\u03b8* and the item properties informs the researcher precisely how well item *j* separates patients with low and high *\u03b8* values in the particular group of patients under consideration using the particular set of items. The difference between absolute slope information (*\u03b1*~*j*~) and relative slope information (*H*~*j*~) is illustrated as follows.\n\nTwo data sets of size *N\u00a0*=\u00a05,000 and five items (*J\u00a0*=\u00a05) (*M\u00a0*=\u00a03 for each item) were generated using the item parameters in Table\u00a0[2](#Tab2){ref-type=\"table\"}. The first data set came from a hypothetical clinical population with *\u03b8* \u223c N(\u22122,1) (i.e., low psychological well-being level), and the second data set came from a hypothetical healthy population with \u03b8\u00a0\u223c\u00a0N(1,1) (i.e., high psychological well-being level). It may be noted that for each of the five items, *\u03b1*~*j\u00a0*~=\u00a01.4 by definition, irrespective of ISRF location parameters and the *\u03b8* distribution. The *H*~*j*~ values in the clinical group were computed and found to range from .36 to .39. In the healthy group, the *H*~*j*~ values were found to be smaller for items 1, 4, and 5 (*H*~1\u00a0~=~\u00a0~.26, *H*~4\u00a0~=~\u00a0~*H*~5\u00a0~=~\u00a0~.25). In the clinical group, the ISRFs of these three items were located more closely to the middle of the *\u03b8* distribution **(**Fig.\u00a0[3](#Fig3){ref-type=\"fig\"} shows this for Item 4) such that higher *H*~*j*~ values resulted, but in the healthy group the ISRFs were located further in the lower tail of the *\u03b8* distribution (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}) resulting in lower *H*~*j*~ values (.3 is considered minimally acceptable; \\[[@CR2]\\], chap. 5). Thus, the items 1, 4, and 5 are well suited for measurement in the clinical group but not in the healthy population, despite their overall discrimination power expressed by *\u03b1*~*j*~\u00a0=\u00a01.4, for *J\u00a0*=\u00a01, \\..., 5. For item 3, *H*~3~ was a little higher in the healthy group because the second and third ISRFs discriminate particularly well at higher ranges of *\u03b8*. For item 2, the location parameters of the ISRFs were widely spread across the *\u03b8* distribution, resulting in good discrimination both at lower and higher levels of *\u03b8* in both distributions. \"We conclude that item scalability coefficient *H*~*j*~ has the advantage that it takes the item (and not the individual ISRF) as a unit and depends simultaneously on the *\u03b8* distribution, the slopes of the ISRFs, and the spread of the locations of the ISRFs. Thus, *H*~*j*~ informs the researcher whether item *j* discriminates well in the group under consideration using the particular set of items, whereas item discrimination *\u03b1*~*j*~ provides information about the discrimination power irrespective of the patient group under consideration and the item properties of item *j* and the other items in the scale.Table\u00a02Example of item parameters of the graded response model (GRM), and Item *H* Values in two different populations with normally distributed latent traitsItem *j*Item parameters GRM*H*~*j*~*\u03b1*~*j*~*\u03b4*~*j*1~*\u03b4*~*j*2~*\u03b4*~*j*3~*\u03b8*\u223cN(--2,1)*\u03b8*\u223cN(1,1)11.4--2.4--2.2--1.00.360.2621.4--4.0--2.01.00.370.3931.4--1.02.02.50.390.4241.4--3.0--2.0--1.00.360.2551.4--2.5--2.0--1.50.360.25*Note*: *H*~*j*~ values are based on simulated data for sample size *N*\u00a0=\u00a05,000Fig.\u00a03Three ISRFs of the same item (Item 4) relative to two different distributions of the latent variable\"\n\n*Confirmatory and exploratory data analysis*. In general, before IRT models are accepted as reasonable descriptions of the data their goodness-of-fit to these data must be investigated and assessed. In general, goodness-of-fit research is different for parametric and nonparametric IRT (with the GRM and the MHM as special cases, respectively). In a parametric IRT analysis the model often serves as null-hypothesis and it is tested whether this null-hypothesis must be rejected or may be supported by the data. Nonparametric IRT analysis in general takes the data as point of departure and (1) instead of positing a unidimensional or multidimensional latent variable structure analyzes the data to find its true dimensionality, and (2) instead of positing a logistic or other functional shape estimates the ISRFs from the data so as to diagnose the items' functioning \\[[@CR16], [@CR18]\\]. This research strategy renders nonparametric IRT a more flexible data-analysis tool than parametric IRT. One could also characterize this distinction as confirmatory (parametric IRT) versus exploratory (nonparametric IRT) (Table\u00a0[1](#Tab1){ref-type=\"table\"}).\n\n*Application of parametric and nonparametric IRT*. If a nonparametric IRT model such as the MHM fits the data, the result is a scale on which patients can be ordered by means of the total score *X*~+~. This total score has a strong linear correlation with latent variable *\u03b8*. Such a scale suffices in many applications. Examples are the comparison of groups, the measurement of change due to therapy, and the establishment of the patient's psychological well-being level as low, medium, or high (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The practical advantage of nonparametric IRT models over parametric IRT models is that the scales they produce contain more items thus reducing the risk of wasting items that have non-logistic but monotone ISRFs that discriminate well in (part of) the group under consideration (e.g., Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}b). Such items contribute well to reliable measurement. In addition, rejection of such items may also harm the coverage of the latent attribute.\n\nIf a fitting parametric IRT model is obtained for a set or a subset of the items, one has a parsimonious description of the item characteristics, and one can use the estimated item parameters to scale the items, and the estimated *\u03b8*s as interval level measures to locate patients on this scale. If a large set of items, also known as an item bank \\[[@CR32]\\], is available, and if a parametric IRT model fits the item bank, parametric IRT models have the advantage that the patient's *\u03b8* can be assessed using different sets of items from the item bank. This may be useful for the measurement of change when change is so large that the set of items that was used initially no longer captures the higher or lower *\u03b8* levels needed for the second measurement, thus necessitating the use of other items. Another application of parametric IRT is computerized adaptive testing (CAT), which selects items that match the patient's *\u03b8* level well from a huge item bank so as to optimize accuracy of *\u03b8* measurement. In principle, CAT requires different item sets for different *\u03b8* values (Table\u00a0[1](#Tab1){ref-type=\"table\"}).\n\nNonparametric IRT analysis in practice {#Sec7}\n======================================\n\n*Software for nonparametric IRT analysis*. Several programs are available for data analysis using nonparametric IRT but not each can handle polytomous item scores. We briefly discuss the programs MSP (\\[[@CR33]\\]; also see \\[[@CR4]\\]) and TestGraf98 \\[[@CR34]\\]. Both programs are used regularly and, together, they provide the researcher with a clear and informative picture of (1) the dimensionality of the data, (2) the (lack of) monotonicity of the ISRFs, and (3) estimates of item locations and item discrimination.\n\nProgram *MSP* uses the MHM as the main analysis model but another nonparametric model not discussed here is also included in the program. Basically the program consists of two parts. The first part of MSP is an algorithm for exploring the dimensionality of the data (\\[[@CR2]\\], chap. 5; \\[[@CR4]\\], chap. 5). The algorithm uses item scalability coefficient *H*~*j*~ to select items. Items that are related to the same latent variable *\u03b8* are selected one by one on the basis of their *H*~*j*~ value. Suppose that one latent variable drives the responses to one subset of the items, another latent variable drives the responses to another subset of items, and so on. Then, the algorithm selects mutually exclusive clusters of items each of which is driven by a different latent variable.\n\nThe second part of MSP provides several statistical tools for exploring the shape of the ISRFs \\[[@CR29]\\]. This is most useful after the dimensionality of the data has been ascertained. For example, due to their strong positive tendency the non-monotone ISRFs in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}b may have an *H*~*j*~ value that is high enough for the items to be selected in a unidimensional cluster, but the non-monotonicity also may distort parts of the ordinal scale defined by the items in the cluster. MSP estimates ISRFs by means of a number of discrete points that are connected to form a jagged 'curve'. Figure\u00a0[4](#Fig4){ref-type=\"fig\"}a shows four discrete ISRFs of the same item (i.e., *M*\u00a0=\u00a04) each estimated by means of eleven points. The researcher can manipulate the number of estimated points. If more points are estimated from the same data more details of the ISRF become visible (i.e., bias is reduced) but because for each point fewer data are available, accuracy decreases. In statistics, this is known as the bias-accuracy trade-off, and it is advisable to try several options to reach a good decision. MSP tests observable deviations from monotonicity for significance. Figure\u00a0[4](#Fig4){ref-type=\"fig\"}a also shows a 'mean' ISRF (dashed curve**)**, which we may call the item score function (ISF) and which is not standard output of MSP. This function must be monotone nondecreasing.Fig.\u00a04(**a**) Discrete estimates of four ISRFs (solid curves; from MSP) of Item 16 ('Satisfied doing daily activities?', from Physical Health and Well Being domain), and the corresponding discrete item-score function (dashed curve), (**b**) also estimated as a continuous curve (from Testgraf98)\n\nProgram *TestGraf98* \\[[@CR34]\\] can be used for studying the shape of the ISF. Unlike MSP, TestGraf98 produces continuous estimates of response functions (\\[[@CR35]\\]; here, only the ISFs); and like MSP, TestGraf98 shows graphical displays of these estimates that can be manipulated with respect to bias and accuracy, and also here it is advisable to try several options. The quality of the decision can be improved by using the confidence envelopes for the estimated ISFs for statistical testing. Figure\u00a0[4](#Fig4){ref-type=\"fig\"}b shows an example of an estimated ISF (solid curve) and its confidence envelopes (dashed curves), which were estimated by means of TestGraf98.\n\n*Research strategies for nonparametric IRT analysis.* MSP provides a method for investigating the dimensionality of the data, and MSP and TestGraf98 both can be used to investigate assumption M. The investigation of dimensionality and monotonicity serves to identify the items that together constitute an ordinal patient scale for the same latent variable.\n\nFor investigating dimensionality, MSP offers the researcher the possibility to set a positive lower bound *c* on *H*~*j*~. Under the MHM, the lowest admissible value is *c\u00a0*=\u00a0.0; MSP's default is *c\u00a0*=\u00a0.3 (\\[[@CR2]\\], chap. 5; \\[[@CR4], [@CR33]\\]). This default value ascertains a lower bound on the overall discrimination power of the items (but researchers are free to choose a higher value) and, as a result, item clusters consist only of sufficiently discriminating items that measure the same latent variable. Thus, MSP aims to produce unidimensional scales that allow accurate patient measurement.\n\nTestGraf98 estimates the ISFs (e.g., Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}b) by means of the nonparametric regression method known as kernel smoothing (e.g., \\[[@CR36]\\], chap. 2; \\[[@CR35]\\]). The availability of confidence envelopes for the continuous ISF estimates provides detailed information of (lack of) monotonicity for each item. TestGraf98 provides these estimates irrespective of the dimensionality of the data. Thus, a good research strategy is to first investigate item-set dimensionality by means of MSP and then use MSP and TestGraf98 to study the ISRFs and the ISFs in dimensionally distinct clusters. See \\[[@CR37]\\] for another method for assessing the shape of these curves.\n\nA real-data example: The World Health Organization Quality-of-Life Scale {#Sec9}\n========================================================================\n\nThe WHOQOL-Bref was developed for assessing individuals' perception and feelings of their daily life. The questionnaire starts with two items, which ask for global estimates of one's quality of life, and then continues with 24 items covering four domains: (a) physical health and well-being (seven items); (b) psychological health and well-being (six items); (c) social relations (three items); and (d) environment (eight items). The two general items were left out of the analysis. In agreement with their numbering in the WHOQOL-Bref, the other 24 items were numbered from 3 to 26. Examples of items are:Do you have enough energy for daily life? (physical domain)How much do you enjoy life? (psychological domain)How satisfied are you with your personal relationships? (social domain)How safe do you feel in your daily life? (environmental domain)\n\nEach item uses a five-point rating scale (i.e., *X*~*j*~\u00a0=\u00a00, \\..., 4 **)**; the higher the item score, the better one's quality of life on the specific domain covered by the item.\n\nThe data were collected by undergraduate psychology students of Tilburg University as part of a course Research Practical in the academic year 2005--2006. Students were instructed to strive for a sample of participants equally distributed across both sexes and the following age categories: 30--39, 40--49, 50--59, and more than 60\u00a0years. The final sample consisted of *N*\u00a0=\u00a0589 respondents from the Dutch population. Of these respondents, 55% were women, mean age was 55.2\u00a0years (SD\u00a0=\u00a014.6), 32% had completed community college or university, 36% had completed vocational school, 20% had high school at most, and 12% had only elementary school or less.\n\n*N*\u00a0=\u00a055 cases had missing item scores. Missing values were estimated using two-way imputation. Comparable to an analysis-of-variance layout, this method uses both a person effect and an item effect for estimating a missing score (for details, see \\[[@CR38]\\]). MSP and TestGraf98 were used to analyze these data and construct one or more scales, thus illustrating the possibilities of the MHM. For the sake of comparison, we ran a principal component analysis and a GRM scale analysis on the data.\n\nResults {#Sec10}\n=======\n\nSample statistics of item and scale scores {#Sec11}\n------------------------------------------\n\nTable\u00a0[3](#Tab3){ref-type=\"table\"} shows that the mean item scores ranged from 2.58 (Item 20: 'Satisfied with sex life?') to 3.46 (Item 25: 'Moving around well?'). The mean *X*~+~ scores were 21.04 (physical domain), 17.01 (psychological domain), 8.52 (social domain), and 24.27 (environmental domain). Correlations between the domain scores ranged from .37 (between physical and social domains) to .51 (between physical and environmental domains). Table\u00a03Results from MSP item selection procedure (Item clusters, item *H*~*j*~ values, and total *H*), and Item *H*~*j*~ values and total *H* for each content domainMSP item selection procedure*H*~*j*~ per content domain*c*\u00a0=\u00a0.3*c*\u00a0=\u00a0.4*j*Mean1212345*Physical health and well-being*3Distraction due to pain^a^3.04.59.59.4010Experiencing energy^a^2.98.43.53.4615Satisfied with sleep2.66.22------------.2825Moving around well3.46.36.43.4116Satisfied doing daily activities2.84.41.56.524Need medical treatment for daily functioning^a^3.17.59.59.4317Satisfied work capacity2.89.40.57.52Scale value21.04.43*Psychological health and well-being*5Enjoying life2.66.34.42.377Being able to concentrate2.80.32----------.2918Satisfied with yourself2.95.41.48.4511Acceptance physical appearance3.23.33----------.3526Experiencing negative feelings^a^2.72.30----------.346Life meaningful2.66.30----------.37Scale value17.01.36*Social relations*19Satisfied relationship with other people3.06.34.50.5020Satisfied with sex life2.58.30.42.4221Satisfied support from others2.88.28--.40.40Scale value8.52.44*Environment*8Feeling safe in daily life3.08.30----------.3322Satisfied living conditions3.13.43----------.4312Enough financial resources3.08.33.53.4223Satisfied getting adequate health care2.87.29--.52.3613Availability information needed in daily life3.06.34.53.4014Opportunities leisure2.90.34.49.399Healthy environment2.88.29------------.3124Satisfied with transport in daily life3.28.34.52.40Scale value24.27.38^a^Reversely scored items\n\nDimensionality analysis {#Sec12}\n-----------------------\n\n*Principal components analysis.* Dimensionality was explored by means of a principal components analysis using polychoric correlations. The ratio of the first to the second eigenvalue of the polychoric correlation matrix was 8.428/1.986\u00a0=\u00a04.24. A ratio of 4:1 is taken as evidence of considerable strength of the first dimension (e.g., \\[[@CR39]**\\]**). The first factor explained 32.4% of the variance. A confirmatory factor analysis of the four a priori domain scales of the WHOQOL-bref improved fit over the one-factor model (*P*\u00a0\u2264\u00a00.001). However, the factors correlated from .50 (physical and social domain) to .79 (psychological and social domain). The explorative factor analysis in conjunction with the confirmative factor analysis justifies the assumption of a general HRQoL dimension underlying each scale.\n\n*Monotone homogeneity model analysis.* Next, MSP was used treating all 24 items as a fixed scale. The MHM does not allow negatively correlating items in one scale. Item 4 ('Need medical treatment for daily functioning?') and Item 20 ('Satisfied with sex life?') correlated negatively but not significantly (*P\u00a0*\\>\u00a0.05); thus all 24 items were used for analysis. The item *H*~*j*~ values (not tabulated) ranged from .21 (Item 15: 'Satisfied with sleep?') to .40 (Item 22: 'Satisfied with living conditions?' and Item 10: 'Enough energy for everyday life?'). The total-scale *H* coefficient was equal to .30. The results suggest that the items tend to cover one latent HRQoL aspect, which, however, induces only weak general association between the items**.**Thus, in addition to this common aspect it seems reasonable to also look for more-specific HRQoL aspects that are covered by subsets of items.\n\nDimensionality was investigated by means of the MSP search algorithm using several *c* values, starting with .3 (default) and then increasing *c* with steps of .05 in each next analysis round. We only report results for *c\u00a0*=\u00a0.3 and *c\u00a0*=\u00a0.4 (other values did not reveal interesting results). For *c\u00a0*=\u00a0.3, one scale consisting of 18 items (*H\u00a0*=\u00a0.35) and one scale consisting of 2 items (*H\u00a0*=\u00a0.59) were found (Table\u00a0[3](#Tab3){ref-type=\"table\"}). The four remaining items were not selected because their *H*~*j*~ values were under .3 (i.e., .22 and .28, .29, and .29). The 18-item scale had a rather heterogeneous content. The 2-item scale asked about distraction due to pain (Item 3) and the need for medical treatment for daily functioning? (Item 4). Thus, their high scalability may be explained by the use of palliative medicines.\n\nFor *c\u00a0*=\u00a0.4, five scales were found consisting of 6, 2, 3, 2, and 3 items, respectively (Table\u00a0[3](#Tab3){ref-type=\"table\"}). The first scale consisted of items from the physical domain and the psychological domain. The other scales consisted of items from one domain. Scale 2 again covered Item 3 and Item 4. Scales 3 and 4 covered environmental-domain aspects. Scale 5 contained all social-domain items.\n\nThus, for default *c\u00a0*=\u00a0.3, 18 of the 24 items were selected in one scale. The pattern of item selection for higher *c* values such as *c\u00a0*=\u00a0.4 showed that the item set progressively crumbled into many smaller scales while other items remained unselected. Sijtsma and Molenaar (\\[[@CR4]\\], pp. 80--86; see also \\[[@CR7]\\] \\[[@CR40]\\]) argued that this typical pattern of results gives evidence that the 18-item set constitutes a unidimensional scale. The total-scale *H* equaled .35, giving evidence of weak scalability (\\[[@CR2]\\], p. 185). Most of the item *H*~*j*~ values were between .3 and .4, also suggesting a weak relationship with the latent variable (\\[[@CR2]\\], p. 185).\n\nFinally, it was investigated whether the four a priori identified item domains could be considered as separate scales. Table\u00a0[3](#Tab3){ref-type=\"table\"} (last column) shows that the total-scale *H* values ranged from .36 (environmental domain) to .44 (social domain). Thus, based on Mokken's classification of scales \\[[@CR2]\\] the four a priori item domains constituted weak to medium scales. Two items had *H* values just smaller than *c\u00a0*=\u00a0.3 (i.e., Item 15 (*H*~15~\u00a0=\u00a0.28): 'Satisfied with sleep?' and Item 18 (*H*~18~\u00a0=\u00a0.29): 'Satisfied with yourself?'). The content domains may be considered as unidimensional clusters of items measuring distinct aspects of HRQoL, each of which are related to a more general underlying HRQoL construct. Because of their conceptual clarity, the remaining analyses were done on the a priori defined item domains.\n\nMonotonicity assessment in each item domain {#Sec13}\n-------------------------------------------\n\nFor each item domain, MSP was used to assess the ISRFs' shapes. First, ISRFs were estimated accurately (i.e., many cases were used to estimate separate points of the ISRFs) but at the expense of possible bias (i.e., only few points were estimated). Second, ISRFs were estimated with little bias (i.e., many points were estimated) but at the expense of accuracy (i.e., few cases were used to estimate each point).\n\nFor the physical domain, the first analysis (high accuracy, more risk of bias) revealed four items of which one or more ISRFs showed minor violations of monotonicity, but none of these violations were significant (5% level, one-tailed test, because only sample decreases are tested as violations; increases support monotonicity). The second analysis (more inaccuracy, less bias) revealed that for all seven items one or more ISRFs showed one or more local decreases, but none them were significant. Figure\u00a0[5](#Fig5){ref-type=\"fig\"}a shows the local, nonsignificant decreases in the ISRFs for Item 3 ('Distraction due to pain?').Fig.\u00a05Four ISRFs of Item 3 ('Distraction due to pain?', from Physical Health and Well Being domain) showing nonsignificant violations of assumption M and rejected by the GRM: (**a**) Results from MSP (including the ISF); (**b**) results from Testgraf98 (ISF and confidence envelopes); (**c**) and results from Multilog7.0 (GRM)\n\nFor the psychological domain, for both analyses (i.e., high accuracy versus little bias) two items were found which had ISRFs showing significant local decreases. For example, Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}a shows for Item 7 ('Being able to concentrate?') that the estimate of ISRF *P*~72~(*\u03b8*)\u00a0=\u00a0*P*(*X*~7~\u00a0\u2265\u00a02\\|*\u03b8*) (the second curve from the top) has several local decreases, the largest of which was significant (5% significance level, *P*\u00a0=\u00a00.019). The estimate of ISRF *P*~73~(*\u03b8*)\u00a0=\u00a0*P*(*X*~7~\u00a0\u2265\u00a03\\|*\u03b8*) shows two small decreases; they were not significant. For the social domain and the environmental domain, no significant violations of the monotonicity assumption were found. It can be concluded that assumption M holds for each of the four scales.Fig.\u00a06Four ISRFs of Item 7 ('Being able to concentrate?') showing significant violations of assumption M and rejected by the GRM: (**a**) Results from MSP (including the ISF); (**b**) results from Testgraf98 (ISF and confidence envelopes); and (**c**) results from Multilog7.0 (GRM)\n\nNext, TestGraf98 was used to investigate assumption M for the ISFs. Several sample sizes were used for estimating curve fragments of the ISFs and balancing the bias-accuracy trade-off. Figure\u00a0[5](#Fig5){ref-type=\"fig\"}b shows the estimated ISF of Item 3 ('Distraction due to pain?'). The confidence envelopes show that the local decrease of the estimated ISF can be ignored safely. For the estimated ISF of Item 7 ('Being able to concentrate?'), Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}b suggests a violation of assumption M for high latent variable levels (i.e., *\u03b8\u00a0*\\>\u00a02). With MSP the user specifies the minimum number of observations used for estimating each point of an ISRF, but Testgraf98 controls the bias-accuracy trade-off by means of a bandwidth parameter, also to be specified by the user but without being able to control the number of observations for estimating separate curve fragments. The effect may be that, in particular at the lower end and higher end of the scale, the ISF is estimated very inaccurately. Combining the results from MSP (no significant decreases of the ISRFs at the higher end of the *\u03b8* scale) and Testgraf98 (a smooth monotone increasing ISF in the middle of the *\u03b8* scale), we may conclude that assumption M holds for item scores. Thus, the expected item score monotonically increases in the latent variable.\n\nComparison of the MHM with the GRM {#Sec14}\n----------------------------------\n\nThe GRM was fitted using Multilog7.0 \\[[@CR41]\\]. Estimation problems occurred for item scoring 0--4 because some score categories were (almost) empty. This was resolved by combining scores 0 and 1 into score 0, and re-scoring the items 0--3. Table\u00a0[4](#Tab4){ref-type=\"table\"} provides the estimated slope parameters (*\u03b1*~*j*~) and the three location parameters (*\u03b4*~*jm*~, *j\u00a0*=\u00a01, 2, 3), and also the estimated item *H*~*j*~ coefficients (also in Table\u00a0[3](#Tab3){ref-type=\"table\"}, last column) and the nonparametric location parameters (*\u03c0*~*jm*~, *j*\u00a0=\u00a01, 2, 3). The location parameters indicate that the items are relatively popular (highly endorsed).Table\u00a04Results of monotone homogeneity model (MHM) scale analysis and estimated item parameters from the graded response model (GRM)MHMGRM ^b^*jH*~*j*~*\u03c0*~*j1*~*\u03c0*~*j2*~*\u03c0*~*j3*~*\u03c0*~*j4*~*\u03b1*~*j*~*\u03b4*~*j1*~*\u03b4*~*j2*~*\u03b4*~*j3*~*Physical health and well-being*3Distraction due to pain^b^.40.99.93.72.401.14\u22122.71\u22120.990.4910Experiencing energy^a^.46.98.95.79.451.97\u22122.24\u22120.630.5615Satisfied with sleep.28.99.95.70.340.85\u22122.09\u22120.791.7825Moving around well.41.99.96.91.611.23\u22123.08\u22122.24\u22120.4116Satisfied doing daily activities.52.98.93.72.213.96\u22121.59\u22120.580.874Need medical treatment for daily functioning^a^.43.98.92.75.231.23\u22122.82\u22121.290.2417Satisfied work capacity.52.99.98.77.203.58\u22121.56\u22120.690.81Scale value.43*Psychological health and well-being*5Enjoying life.371.00.98.61.071.93\u22122.79\u22120.381.977Being able to concentrate.29.99.98.77.200.82\u22124.270.681.6918Satisfied with yourself.451.00.93.64.151.99\u22122.85\u22120.971.1211Acceptance physical appearance.35.99.98.80.451.08\u22123.93\u22121.560.2226Experiencing negative feelings^a^.341.00.97.61.081.11\u22122.84\u22120.601.906Life meaningful.371.00.96.62.231.90\u22122.71\u22120.351.94Scale value.36*Social relations*19Satisfied relationship with other people.50.99.96.82.292.52\u22122.21\u22121.100.9720Satisfied with sex life.40.97.88.58.151.32\u22123.19\u22120.901.3821Satisfied support from others.42.99.97.72.201.38\u22121.88\u22120.341.62Scale value.44*Environment*8Feeling safe in daily life.331.00.98.78.321.12\u22124.12\u22121.370.8422Satisfied living conditions.43.98.90.68.341.96\u22122.71\u22121.270.6312Enough financial resources.42.99.95.70.441.84\u22122.22\u22120.730.2123Satisfied getting adequate health care.36.99.97.68.231.32\u22122.84\u22120.931.3313Availability information needed in daily life.40.99.95.72.211.64\u22123.21\u22120.980.6014Opportunities leisure.39.99.98.75.341.61\u22121.94\u22120.690.609Healthy environment.31.99.98.83.321.02\u22123.95\u22120.891.4324Satisfied with transport in daily life.40.99.97.86.451.66\u22122.77\u22121.550.17Scale value.38^a^\u00a0Reversely scored items; ^b^\u00a0For the GRM analysis, items were recoded by collapsing item scores 0 and 1 into item score 1\n\nSeveral methods are available for assessing the fit of the GRM, but many are problematic and no generally accepted standard goodness-of-fit method for the GRM is presently available (\\[[@CR42]\\] pp. 85--89). We investigated goodness-of-fit of the GRM by means of posterior predictive assessment \\[[@CR43]\\], which can provide graphical and numerical evidence about model fit. Model fit was investigated at the level of items. The most interesting results were found for items from the physical and psychological domains. Results are only discussed for these domains. For the physical domain, Item 3 ('Distraction due to pain?') and Item 7 **(**'Being able to concentrate') showed significant misfit (*P\u00a0*\\<\u00a0.01). In Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}c, the curve made up by dots connected by straight line pieces represents the estimated nonparametric ISF of Item 3, the solid curve represents the expected ISF of Item 3 under the GRM, and the dotted curves represent the 95% confidence envelopes. The GRM rejects this item. Thus, modeling the jagged pattern of the estimated ISF by means of logistic ISRFs having the same slopes would do injustice to the data. However, it is noteworthy that Item 3 has good measurement properties (Table\u00a0[4](#Tab4){ref-type=\"table\"}: e.g., *H*~3~\u00a0=\u00a0.40) under the more general MHM, and from this model's perspective it might be retained in the scale.\n\nItem 7 ('Being able to concentrate?') was a popular item (Table\u00a0[4](#Tab4){ref-type=\"table\"}; *\u03c0*~*j*1~\u00a0=\u00a0.99, *\u03c0*~*j*2~\u00a0=\u00a0.98; also, see Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}a, upper two ISRFs). As a result, the GRM could not be estimated accurately; item parameters were estimated very inaccurately (standard errors \\>.25). Figure\u00a0[6](#Fig6){ref-type=\"fig\"}c shows evidence of misfit at \u03b8\u00a0\\>\u00a02, for which the observed ISF fell outside the 95% confidence interval. This means that the GRM gives biased results for \u03b8\u00a0\\>\u00a02. The nonparametric estimates of the ISRFs were monotone. This provides evidence that the MHM adequately fitted Item 7. However, *H*~7\u00a0~=~\u00a0~.29, which is rather low. A reason to keep this item in the scale is that it may help measuring differences at the lower and middle ranges of the *\u03b8* scale, which are the most relevant ranges for measuring HRQoL.\n\nSummary of the scale properties {#Sec15}\n-------------------------------\n\nThe WHOQOL-Bref is most often used in scientific research (e.g., epidemiological studies and clinical trails) and by health professionals (e.g., to assess treatment efficacy) \\[[@CR19]\\]. The nonparametric MHM analyses revealed that the scales have adequate properties for comparing groups on the underlying HRQoL aspects. Each of the four domains of the WHOQOL-Bref constitutes a unidimensional scale, each scale measuring a different aspect of HRQoL in addition to a weak common HRQoL attribute. This justifies reporting both separate domain scores and possibly an overall HRQoL score. The scalability results showed that the domain scales are weak to moderate, with scalability coefficients *H* ranging from .36 to .44. The test-score reliabilities of the four domain scores were .82, .76, .66, and .81, respectively. The rank correlations between sum score *X*~+~ and the estimated *\u03b8* from the GRM varied from .91 ('physical health') to .96 ('environment') (Pearson correlations ranged from .94 ('physical health') to .99 ('environment')). Thus, *X*~+~ and estimated *\u03b8* carry nearly the same rank order (and numerical) information. This interesting result further justifies the use of the nonparametric MHM for scale analysis, and the use of *X*~+~ for (at least) ordinal measurement of persons.\n\nBecause the item-score distributions were severely skewed to the left, the lower response categories 0 and 1 were ineffective for HRQoL measurement in the general population. The locations of the ISRFs for the higher response categories 2, 3, and 4 were well spread along the *\u03b8* scale. The ISRFs' discrimination power as reflected by the *H*~*j*~ values often was in the weak to medium range. Thus**,** the higher response categories are modestly informative across a wide range on the *\u03b8* scale.\n\nThe relatively short WHOQOL-Bref may also be considered for use as a tool for assessing HRQoL at the individual level in clinical and medical settings. For example, the WHOQOL-Bref may be used to evaluate whether a patient's HRQoL has improved after taking medication. An interesting feature of a fitting IRT model is that psychometric properties can be evaluated conditionally on the latent variable. For example, the measurement error of *X*~+~ can be evaluated at different values of the latent variable. TestGraf98 provides graphical information about the standard error of measurement based on the MHM model. For example, for the physical-health domain Testgraf98 estimated a standard error of measurement ranging from 2.8 for *X*~+~\u00a0\u2264\u00a020 to 1.8 for X~+~\u00a0\u2265\u00a028. Thus, to be significant at the 5% significance level differences between two observed *X*~+~ scores have to be larger than $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ 2.8 \\times 1.96 \\times \\sqrt 2 \\approx 8 $$\\end{document}$ for *X*~+~\u00a0\u2264\u00a020, and larger than $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{mathrsfs}\n\\usepackage{upgreek}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}$$ 1.8 \\times 1.96 \\times \\sqrt 2 \\approx 5 $$\\end{document}$ for X~+~\u00a0\u2265\u00a028 (e.g., see \\[[@CR44]\\], p. 209). This is a substantial standard error of measurement relative to the length of scale. This relatively large measurement error appears to be consistent with the observed *H*~*j*~ values, which indicate weak to moderate scalability. For the other three content domains, the standard error of measurement was also substantial. Thus, caution has to be exercised when drawing conclusions about differences and changes in individual levels of HRQoL based on observed *X*~+~ scores from the WHOQOL-Bref and any other HRQOL measure---see \\[[@CR45]\\].\n\nDiscussion {#Sec16}\n==========\n\nThis study explained how the nonparametric monotone homogeneity model contributes to the construction of scales for the measurement of HRQoL. The MHM is more general than parametric IRT models \\[[@CR24]\\], such as the much-used parametric graded response model \\[[@CR26]\\] but also the partial credit model \\[[@CR13]--[@CR15]\\] and the generalized partial credit model \\[[@CR11]\\]. Hemker et\u00a0al. \\[[@CR46]\\] showed that all known parametric IRT models for polytomous items are special cases of the nonparametric MHM. This means that any item set satisfying the requirements of a parametric IRT model for polytomous items also satisfies the requirements of the nonparametric MHM. Given the greater generality and flexibility of the nonparametric MHM, which results in longer scales, and because *X*~+~ and estimated *\u03b8* carry the same rank order information (based on the approximate stochastic ordering property of *\u03b8* given *X*~+~), the nonparametric MHM is highly suited for person measurement.\n\nIn an HRQoL context, often little is known about the psychometric properties of new questionnaires. A typical nonparametric MHM analysis explores the dimensionality of the data by capitalizing on model assumptions such as monotonicity (MSP), and studies the shapes of the ISRFs and the ISFs in order to learn more about the (mal-)functioning of individual items (MSP and TestGraf98). This results in scales on which groups can be compared and changes monitored without making unduly restrictive assumptions about the data.\n\nThe properties of any IRT model only hold for the application at hand when the model fits the data. In case of misfit, the structure of the model does not match the structure of the data. One cause of misfit is that the data are multidimensional while the model assumes unidimensionality. Another cause of misfit is that the real ISRFs may not be monotone or that they are monotone but fail to have the logistic shape assumed by many parametric models. Other causes of misfit, such as a multiple-group structure as in differential item functioning (e.g., \\[[@CR47]\\]) or person misfit \\[[@CR48]\\] were not considered here.\n\nWhen the MHM fits the data, the researcher may decide to also investigate goodness of fit of the GRM or other parametric IRT models for polytomous items. The choice of a parametric model may be based on the flexibility of the model. For example, the partial credit model only has item location parameters but assumes the slopes of the response functions to be the same within and between items, whereas the generalized partial credit model also allows for varying slope parameters between items, just as the GRM. If one pursues a parametric IRT analysis, misfit may be a good reason to resort to a nonparametric IRT model and still have an ordinal patient scale. If CAT is pursued, one of the parametric models is a better option provided the model fits the data well. In an HRQoL context, CAT indeed could prove to be successful because patients have a definitive interest in providing truthful answers (in the educational context, in which CAT originated, CAT requires that items be kept secret. This requires item banks often containing hundreds of calibrated items). As a result, in HRQoL measurement CAT presently meets with a growing interest (e.g., \\[[@CR49]--[@CR54]\\]).\n\nThis article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.\n\nGRM\n\n:   Graded response model\n\nHRQoL\n\n:   Health-related quality-of-life\n\nIRF\n\n:   Item response function\n\nIRT\n\n:   Item response theory\n\nISF\n\n:   Item score function\n\nISRF\n\n:   Item-step response function\n\nLI\n\n:   Local independence\n\nM\n\n:   Monotonicity\n\nMHM\n\n:   Monotone homogeneity model\n\nUD\n\n:   Unidimensionality\n\nWHOQOL-Bref\n\n:   World Health Organization Quality-of-Life Scale\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\nFine needle aspiration and cytology (FNAC) has become a commonly used technique in the clinical management of TN.[@ref1] It is considered as a useful diagnostic aid as well as a screening test in this condition with an overall accuracy of around 95% in detection of thyroid cancers.[@ref2] In the process of evaluation of nodular lesions of thyroid, FNAC remains the gold standard and is definitely the procedure of choice. However, FNAC has its shortcomings and diagnostic drawbacks like any other test. Major issues are the gray-zoned cases i.e. cases suspicious for malignancy.[@ref3]\n\nIt has been reported in different studies that CD44 and Galectin-3 can serve as prospective markers to spot and effectively pick malignant transformed thyrocytes preoperatively. An accurate and finer technique that can be implicated to thyroid nodules being planned for excision is the detection of these immunomarkers on cytology specimens by immunocytology.[@ref4],[@ref5]\n\nGalectin-3, plays a pivotal role in processes like cell proliferation, cell-cell interaction, and cell-matrix adhesion mainly via binding to glycoproteins. Especially nuclear Gal-3 is apparently linked with normal cell proliferation.[@ref6]-[@ref8]\n\nConsidering the shortcomings of traditional FNAC during diagnosis of thyroid neoplasms and the need of a simple test based on an amalgamation of FNAC and different markers for improving preoperative diagnosis of thyroid nodules, Galectin-3 was employed as a marker in 60 patients who presented with solitary thyroid nodules.\n\nMETHODS {#sec1-2}\n=======\n\nA cross sectional (descriptive and comparative), includes 60 cases of solitary thyroid nodules was conducted at the Departments of Pathology King Edward Medical University/Mayo Hospital and PGMI/General Hospital Lahore. The Purposive convenient sampling technique was used to complete this study in two years i.e. from June 2012 to July 2014.\n\nThe ethical committee of the University of Health Sciences, Lahore, approved the study protocol. After taking the consent of the patient, FNAC was done and two smears and a cellblock were prepared. The FNA smears and cellblocks slides were stained with Haematoxylin and Eosin. Formalin fixed cellular aspirate specimen (cell block) was placed in an automated tissue processor (Model RH-12 EP SAKURA Fine Technical Company Limited, Tokyo, Japan) for 17 hours. After processing, tissue embedding was done using paraffin wax. Paraffin embedded cellblocks were kept for Galectin-3immunostaining.\n\nImmunohistochemical staining of the sections prepared from the cellblocks was done.[@ref9] A labeled antibody is used to identify the antigen of interest in tissue sections. Peroxidase-anti peroxidase immune complex method or the biotin-avidin immunoenzymatic technique is used. Antigen retrieval is carried out to increase the sensitivity of the technique.[@ref10]\n\nThe cells that displayed a brownish nuclear/cytoplasmic stain were considered immunopositive.[@ref11] A positive immunoreactivity was considered when 10% of the tumor cells stained positively for the antibody.[@ref12] Scoring of the smears for Galectin-3 reactivity was done as follows:[@ref9]\n\nNo staining of tumor cells (0).+1 \\<10% stained cells.+2 10-50% stained cells.+3 \\>50% stained cells.\n\nRESULTS {#sec1-3}\n=======\n\nOut of these 60 patients, the maximum number of cases 44/60 (73%) had a history of duration of disease of 10-20 months ([Table-I](#T1){ref-type=\"table\"}). It was found that maximum number of malignant cases (follicular carcinoma + papillary carcinoma) 27/37 (72.9%) were also falling in the group of 10-20 months duration of symptoms ([Table-I](#T1){ref-type=\"table\"}).\n\n###### \n\nAssociation of Galectin-3 expression with benign and malignant Thyroid nodules.\n\n             *Histopathological diagnoses*   *Total*   *%age*   \n  ---------- ------------------------------- --------- -------- -----\n  Positive   13                              23        36       60\n  Negative   24                              0         24       40\n  Total      37                              23        60       100\n\nP-value \\<0.001 (highly significant).\n\nAll the patients presented with solitary thyroid nodule. The mean size of the nodule was 46.02cm^3^ \u00b1 41.66. The minimum and maximum nodule size was two and 168cm three cms respectively. Number of cases having a size of \\<20cm^3^ (22/60) was same as the number of cases (22/60) having a size in the group of 20-60cm^3^ and was 36% of the total cases.\n\nMost of the malignant case 4/9 (44%) of follicular carcinomas and 12/28 (42%) of papillary carcinomas had a size that fell in the group 20-60cm^3^ while most of the benign cases of follicular adenomas 9/23 (39%) were falling in the group of \\<20cm^3^ ([Fig.1](#F1){ref-type=\"fig\"}).\n\n![Photomicrograph **A.** FNAC follicular neoplasm Bethesda category IV (H&E stain 200 X), **B.** histological section of follicular carcinoma showing capsular invasion (H&E stain 400 X), **C.** FNAC papillary carcinoma thyroid showing prominent nuclear features (H&E stain 400 X), and **D.** histological section of papillary carcinoma H&E stain 200 X](PJMS-33-726-g001){#F1}\n\nThe categorization of the smears after cytological diagnosis was done in accordance with The Bethesda System for reporting thyroid cytopathology. 30/60 (50%) aspirates were placed in category IV, 26/60 (43.33%) in categories V and 4/60 (6.67%) patients were placed in category VI. It was found that Category IV was the most common and category VI was the least common category ([Fig.1](#F1){ref-type=\"fig\"}).\n\nHistological diagnosis done on H&E stained sections prepared from the resected nodules showed that among 60 cases of thyroid lesions there were 28/60 (46.6%) cases of papillary carcinoma. Cases of follicular adenoma were 23/60 (38.3%) and 9/60 (15%) cases of follicular carcinoma were reported. However, no case of medullary or anaplastic carcinoma was diagnosed in our study ([Fig.1](#F1){ref-type=\"fig\"}). This was taken as a gold standard.\n\nGalectin-3immunostaining of the sections prepared from the cell blocks was negative in 23/23 benign lesions showing that it was 100% negative in follicular adenomas. Out of 37 malignant cases 24/37 (65%) lesions showed Positivity for this immunomarker while 13/60 (35%) showed negativity for the immunostain. Galectin-3 helped in the preoperative diagnosis of 100% (23/23) follicular adenomas, 75% (21/28) papillary carcinomas and 33.3% (3/9) follicular carcinomas. Overall it helped in reaching a conclusive diagnosis in 78% (47/60) thyroid cytologies ([Table-III](#T2 T3){ref-type=\"table\"}).\n\n###### \n\nExpression of Galectin-3 staining in different histologic lesions.\n\n  *Galectin-3*   *Histological*   *Total*        \n  -------------- ---------------- --------- ---- ----\n  Positive       7                6         23   36\n  Negative       21               3         0    24\n  Total          28               9         23   60\n\n###### \n\nPredictive value of Galectin-3 in solitary Thyroid nodules.\n\n                        *Histopathological diagnosis*   *Total*                                                 \n  --------------------- ------------------------------- ----------------------------- ------------------------- ----\n  Galectin-3 Staining   Positive                        True positive(a) 24(40%)      False positive(b) 0(o%)   24\n  Negative              False negative(c) 13(21.67%)    True negative(d) 23(38.33%)   36                        \n  Total                 37(61.6%)                       23(38.3%)                     60(100%)                  \n\n***Note:*** Sensitivity =64.87%, Specificity=100%, Positive predictive value=100%, Negative predictive value=63.89%, Accuracy rate=78.33%, Confidence interval = 95% Spearman correlation .644 (*p* --value \\<0.001)\n\nThe expression was more common in Papillary carcinoma as compared to follicular carcinoma because out of 21/28 PC (75%) showed positive reaction and only 7/28 (25%) showed a negative reaction while only 3/9 (33%) follicular carcinoma showed positive staining for Galectin-3 antibody and 6/9 (67%) remained negative ([Table II](#T2){ref-type=\"table\"}-[III](#T3){ref-type=\"table\"} and [Fig.2](#F2){ref-type=\"fig\"}).\n\n![Photomicrograph showing negative control for Galectin-3staining in thyroid follicular cells on FNAC 200X, Photomicrograph showing Positive control for Galectin-3staining in Anaplastic large cell lymphoma 400X, Photomicrograph showing strong positive Galectin-3staining in thyroid follicular cells on FNAC and Photomicrograph showing weak positive Galectin-3staining in thyroid follicular cells on FNAC.](PJMS-33-726-g002){#F2}\n\nThe positivity of Galectin-3 was 65% in malignant cases as compared to benign cases (0%). The difference calculated came out to be highly significant p \\<0.001 ([Table II](#T2){ref-type=\"table\"}-[III](#T3){ref-type=\"table\"}).\n\nThe sensitivity and specificity of Galectin-3 in detecting malignant transformed thyrocytes on cytological specimen obtained by FNAC were 64.87% and 100% respectively. The PPV and NPV of Galectin-3 staining were 100% and 63.89% respectively. Moreover, diagnostic accuracy of Galectin-3staining was 78.33%. The confidence interval was calculated to be 95% ([Table-III](#T3){ref-type=\"table\"}).\n\nDISCUSSION {#sec1-4}\n==========\n\nNumber of cases (benign and malignant) 22/60 having a size \\<20 cm^3^ was the same as the number of cases with a size of 20-60 cm^3^. Similarly maximum number of papillary carcinomas 12/28 as well as follicular carcinomas 4/9 also fell in the group of 20-60cm.[@ref3] These results obtained in our research were corresponding to other studies.[@ref13]-[@ref15] Our findings differ slightly with Basharat et al, who found in their study the size range of thyroid nodules to be 9 cm in greatest dimension.[@ref1]\n\nNevertheless no noteworthy association between size and duration of growth and the presence of malignancy was found. The findings are consistent with findings of other studies[@ref16],[@ref17] who reported that size of nodule and interval of growth of nodule are not helpful for predicting or excluding thyroid malignancy.\n\nVarious reporting systems for thyroid cytology have been adopted for the last three decades but none of them was related to the prognosis of disease and patient's outcome. These reporting schedules were least informative due to variability of sensitivity and least reproducibility. The introduction of the new simplified Bethesda system for reporting Thyroid cytopathology into six categories logically relates to the prognosis of thyroid diseases and increases the reproducibility of diagnosis.[@ref18] Bethesda Cytopathology Reporting system can help with a better patient's outcome due to proper clinical management of thyroid swellings and saves patients from unnecessary thyroid surgery. The use of standardized categorical systems for FNAC reporting can make results easier to understand for clinicians and give clear indications for therapeutic action.[@ref19]-[@ref21]\n\nAdopting the above mentioned reporting system the inclusion criteria in our study were, the cytopathology categories IV, V and VI. In our study the most frequent was the category IV with 30/60 (50%) showing follicular neoplasm or showing a risk factor for a follicular neoplasm with specification for H\u00fcrthle cell (oncocytic) type. The next commonly seen patients belonged to category V 26/60 (43.3%), showing morphology indicating a risk of malignancy including papillary carcinoma, medullary carcinoma, metastatic carcinoma, or lymphoma. Category VI showing malignant cytology was the least commonly seen class with 4/60 (7%). This class comprised of papillary thyroid carcinoma, medullary thyroid carcinoma, poorly differentiated carcinoma, undifferentiated (anaplastic) carcinoma, carcinoma with mixed features, squamous cell carcinoma, metastatic carcinoma and lymphoma.\n\nOn histopathology there were 23/60 (38.3%) cases of follicular adenomas. 28/60 (46.6%) cases were of papillary carcinoma while 9/60 (15%) cases were of follicular carcinoma. None of the tumor revealed the histology of anaplastic or medullary carcinoma.\n\nAmongst these, the most frequent one was papillary carcinoma and it was followed by follicular adenoma. The incidence of higher number of papillary carcinoma than follicular carcinoma is correspondent with other excerpts in the literature.[@ref22],[@ref23] This also shows a relationship with the actuality that Papillary carcinomas comprise 80% while follicular carcinomas 10-15% of all thyroid malignancies.[@ref24]\n\nThe outcome of a study conducted in, 2008 shows that the immunohistochemical expression of Galectin-3 can serve as a highly specific marker of malignancy. Nevertheless its sensitivity in differentiating malignant from benign thyroid neoplasms is somewhat less.\n\nIn our study, Galectin-3 was 100% negative in follicular adenomas because all 23 cases were not reactive on staining with this antibody. Out of 37 malignant cases 24/37 (65%) lesions showed positivity for this immunomarker while 13/60 (35%) showed negativity for the immunostain. With a 65% positivity of Galectin-3 in malignant lesions and 0% positivity in benign cases the difference is highly significant and the *p-* value came out to be \\< 0.001 which was highly significant.\n\nThe negativity does not mean that the cytopathology report should be considered 100% benign because some of the malignant neoplasms were missed by Galectin-3 test and showed a false negative reactivity. Technical issues can be accredited to some of these diagnostic malfunctions at any rate. These can be improved partially by additional specific guidance and committed workshops on the use of Galectin-3 expression testing.\n\nHowever, a positive expression was always seen in malignant lesions suggesting that positivity of Galectin-3 is a strong indicator for the presence of malignancy in a thyroid aspirate and negates the chance of it being a benign lesion as 0% false positive results were observed in the present study.\n\nThe expression was more common in Papillary carcinoma as compared to follicular carcinoma because out of 28 PC, 21(75%) showed positive reaction and only 7 (25%) showed negative reaction while out of 9 follicular carcinoma only 3 (33%) showed positive staining for Galectin-3 antibody while 6 (67%) remained negative. In our research, the positivity of Galectin-3 showed a disseminate cytoplasmic/nuclear staining in most cases of PTC. This included both the classical and follicular variant. The finding is consistent with the study of Aiad et al. according to which Galectin-3 expression was markedly lower in FC as compared to PC.[@ref11]\n\nThe follicular thyroid carcinomas also demonstrated the positive expression of Galectin-3 by immunohistochemistry but much less as compared to PTC. Our findings are consistent with Fischer and Asa.[@ref24]\n\nIn the present study Galectin-3 helped in the preoperative diagnosis of 100% (23/23) follicular adenomas on FNAC. The result is consistent with a study of Raggio et al., (2010) which also claimed that the immunostain provided a help in the cytological diagnosis of 100% (31/31) follicular adenomas.[@ref25] As far as papillary thyroid carcinoma is concerned our study aided in the diagnosis of 75% (21/28) smears of PCs. Coming to the pre-operative diagnosis of follicular carcinoma the present study was of a help in 33% (3/9) thyroid aspirates.\n\nAccording to this study the sensitivity of Galectin-3was found to be 64.87% and specificity of staining was found to be 100%. The negative predictive value was found to be 63.89% while the positive predictive value was 100%. The accuracy rate was recorded as 78.33%. The results were quite similar to the results of a study conducted by Herraiz et al.[@ref13] where the sensitivity and specificity of Galectin-3 was found to be 60% and 100% respectively.\n\nCONCLUSION {#sec1-5}\n==========\n\nWe recommend Galectin-3 as a supplementary immunostain to be used in in difficult cytologic smears and for differentiation of benign from malignant thyroid lesions especially a hyperplastic papilla from PTC and FA from the follicular variant of PTC. We propose the use of Galectin-3 immunostaining in preoperative FNAC of thyroid nodule more than ever to evaluate the indeterminate cytology and to avoid unnecessary aggressive surgical interference in benign lesions.\n\nAuthor's Contribution {#sec2-5}\n---------------------\n\n**AM** was the main author who did her MPhil research on this topic.\n\n**MHB** supervised the research.\n\n**IUQ** supervised the thesis.\n\n**AM** is responsible and accountable for the accuracy or integrity of the work.\n"}
+{"text": "INTRODUCTION\n============\n\nExposure of feedlot animals to hot or cold stress reduces growth and feed efficiency ([@b2-ajas-29-3-436]). Cold climate during winter and exposure to cold stress induces metabolic acclimatization, resulting in decreased animal performance and production efficiency ([@b29-ajas-29-3-436]; [@b4-ajas-29-3-436]). Temperature stress seemingly changes metabolic and endocrinal status of animals. For example, cattle exposed to acute cold have increased concentrations of plasma corticosteroids ([@b1-ajas-29-3-436]) and circulating nonesterified fatty acid (NEFA) concentrations ([@b5-ajas-29-3-436]; [@b21-ajas-29-3-436]).\n\nLow environmental temperature can cause cold stress, which affects the immune system and thus animal health. The effect of a cold environment on the immune response has been studied in other than cattle. For example, pigs in a cold environment caused increase in serum adrenocorticotropic hormone and cortisol concentrations coincident with tissue cytokine mRNA levels ([@b10-ajas-29-3-436]). Cold stress increases *in vivo* proinflammatory cytokine gene expression, including interleukin mRNA, in chicken ([@b13-ajas-29-3-436]). However, limited data are available on the effect of a cold environment on the immune response in cattle.\n\nTemperatures on the Korean peninsula increased by approximately 2\u00b0C during 1992 through 2004 due to a global warming ([@b14-ajas-29-3-436]), and hotter summers are expected in the future. In addition, global warming also caused colder winters in South Korea. Therefore, animal performance should have been greatly impacted by cold stress, as cattle sheds in Korea are generally of the open type. Little information is available on the effects of temperature on growth performance and the metabolic and immunological responses in Korean cattle steers. This study was performed to examine whether ambient temperature affects growth, feed efficiency, blood metabolites, and immune cell populations in Korean cattle steers.\n\nMATERIALS AND METHODS\n=====================\n\nAnimals and feeding trial\n-------------------------\n\nAll experimental procedures involving animals were approved by the Seoul National University Institutional Animal Use and Care Committee (SNUIAUCC), Republic of Korea, and were conducted in accordance with the Animal Experimental Guidelines provided by SNUIAUCC.\n\nEighteen Korean cattle steers, with a mean age of 10 months and a mean weight of 277 kg, were used. The study was conducted at the University Animal Farm of the College of Agriculture & Life Sciences on the Pyeongchang campus of Seoul National University, South Korea. Experimental farm was covered by roof, and thus animal was protected from the rain. Doors were installed at both of barn, and animal was raised indoor. All steers were fed a growing stage- concentrate diet at approximately 1.5% of body weight using the DeLaval Alpro automatic feeding station (DeLaval, Tumba, Sweden) for 8 weeks. Feeding period 1 (P1) was from March 7 to April 3 and period 2 (P2) was from April 4 to May 1. The formula of concentrate diet is shown in [Table 1](#t1-ajas-29-3-436){ref-type=\"table\"}. Timothy hay was fed *ad libitum*. Water was freely provided. The chemical composition of the feed is shown in [Table 1](#t1-ajas-29-3-436){ref-type=\"table\"}. Daily feed intake of the concentrate diet was automatically recorded on an online Alpro system computer. An equal amount of hay was fed to each animal twice a day (8 AM and 6 PM), and the residual hay was weighed before the morning feeding. Samples of concentrate and hay were collected weekly and stored until analysis. Body weight was measured at 9 AM before feeding on the start date, and 4 and 8 weeks after feeding.\n\nChemical composition analyses\n-----------------------------\n\nThe chemical compositions (dry matter, crude protein, crude fat, ash, Ca, and P) of the concentrate diet and timothy hay were determined using an AOAC method (1990). The neutral detergent fiber and acid detergent fiber contents of the hay were analyzed using the sequential method with the ANKOM200 Fiber Analyzer (Ankom Technology Corp., Macedon, NY, USA) and reagents, as described by [@b24-ajas-29-3-436].\n\nBlood collection and temperature measurements\n---------------------------------------------\n\nBlood was collected three times; on March 7, April 4, and May 2, 2014, after a 9 h fast by jugular venipuncture into non-heparinized vacutainers (20 mL; Becton-Dickinson, Waltham, MA, USA) and ethylenediaminetetraacetic acid (EDTA)-treated vacutainers (20 mL). Some of the whole blood collected in the EDTA-vacutainers was used for the immune cell analysis. Serum and plasma were separated by centrifugation at 1,500\u00d7*g* at 4\u00b0C for 15 min. The plasma and serum were subsequently stored at \u221280\u00b0C until analysis.\n\nAmbient and climate temperatures and relative humidity inside and outside the barn were recorded at 1 h intervals using three HOBO data loggers (Onset Computer Corp., Bourne, MA, USA), and monthly average values of minimum, mean, and maximum temperatures and humidity were calculated using the daily data.\n\nBlood analysis\n--------------\n\nThe analytical reagents for glucose, triglyceride (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), cholesterol, total protein, glutamic oxaloacetic transaminase (GOT), glutamic pyruvate transaminase (GPT), creatinine, blood urea nitrogen, calcium, magnesium, and phosphorus were purchased from JW Medical (Seoul, Korea). The analytical reagents for the NEFA, phospholipid, and beta-hydroxy butyrate analyses were purchased from Wako Pure Chemical (Osaka, Japan). The analytical reagents for complements C3 and C4 were purchased from Sekisui (Osaka, Japan). All of these parameters were analyzed using an automated chemistry analyzer (Hitachi 7180, Tokyo, Japan). Plasma cortisol was analyzed using a cortisol salivary HS enzyme-linked immunosorbent assay kit (cat. no. SLV4635; DRG Instruments, Marburg, Germany). The intra and interassay coefficients of variation for the cortisol kit were 4.0% and 4.6%, respectively, using bovine plasma samples.\n\nImmune cell analysis\n--------------------\n\nA 100 \u03bcL aliquot of whole blood was treated with 1 mL BD lysing buffer (Becton-Dickinson, USA) and incubated for 3 min at room temperature to analyze the immune cell characteristics. Total cell numbers were counted with a TC10 Automated Cell Counter (Bio-Rad, Hercules, CA, USA). Then, the granulocyte: lymphocyte ratio, identified by flow cytometry using the forward scatter (FSC) and side scatter (SSC) settings ([@b15-ajas-29-3-436]), was calculated as a percentage of total cells. Leukocytes (5\u00d710^5^ cells), stained with anti-bovine CD14-FITC and MHC class II DQ-PE antibodies for monocytes and B cells, respectively and CD4-Alexa647, CD8-PE, and CD25-FITC antibodies (AbD Serotec, Raleigh, NC, USA) for T cells, respectively, for 15 min at 4\u00b0C, were analyzed by flow cytometry (Becton-Dickinson, USA) and FlowJo software (TreeStar, Ashland, OR, USA). Briefly, lymphocytes and monocytes were gated based on FSC-SSC setting and CD14^+^MHC class II DQ^+^ cells were analyzed as monocytes and CD14^\u2212^MHC class II DQ^+^ as B cells. T cells were analyzed with CD4 and/or CD8 positive among lymphocytes and then activated T cells were identified as CD25^+^ cells.\n\nStatistical analysis\n--------------------\n\nAll data are expressed as means\u00b1standard error of mean. Differences in weather data among months were analyzed using one-way analysis of variance (ANOVA). Changes in the growth performance and blood parameters over time were analyzed by repeated-measures ANOVA with the Tukey--Kramer post-hoc test. A p\\<0.05 was considered significant. All statistical tests were performed using R Studio for Windows software package (R Studio, Boston, MA, USA).\n\nRESULTS AND DISCUSSION\n======================\n\nClimate conditions\n------------------\n\nExperimental farm was covered by roof, and thus animal was protected from the rain. Doors were installed at both of barn, and animal was raised indoor. Thus, mild wind may affect wind-chill temperature. However, minimum indoor temperature was 1.0\u00b0C and 6.2\u00b0C on March and April, which are not within severe cold stress range. Thus, wind factor may be a minor climate variable. Mean (8.7\u00b0C) and minimum (1.0\u00b0C) indoor ambient temperature of March was lower (p\\<0.001) than that (13.0\u00b0C and 6.2\u00b0C) of April, respectively. But, maximum temperature of March (20.6\u00b0C) was not different from that of April (21.4\u00b0C). Mean (4.6\u00b0C), maximum (10.6\u00b0C), and minimum (\u22121.7\u00b0C) climate temperature of March was lower (p\\<0.001) than that (9.8\u00b0C, 15.8\u00b0C, 3.4\u00b0C) of April, respectively. Mean, maximum, and minimum relative humidity of March was not different from those of April ([Table 2](#t2-ajas-29-3-436){ref-type=\"table\"}).\n\nGrowth performance\n------------------\n\nDaily intake of concentrate and forage was higher (p\\<0.001) during April (concentrate, 3.97 kg and forage, 3.71 kg) than March (concentrate, 3.54 kg and forage, 3.34 kg). Average daily gain was higher (p\\<0.001) during April (1.38 kg/d) than March (1.13 kg/d). Feed efficiency during April (0.18 kg gain/kg feed) was higher (p\\<0.02) than March (0.16 gain/kg feed) ([Table 3](#t3-ajas-29-3-436){ref-type=\"table\"}). In this study, indoor ambient temperature affected both the growth and feed efficiency of Korean cattle steers. Weight gain and feed efficiency were lower during the colder month of March compared to April. Other studies have reported similar results; colder temperatures resulted in lower feed efficiency ([@b9-ajas-29-3-436]; [@b8-ajas-29-3-436]) and average daily gain ([@b4-ajas-29-3-436]). Cold stress has been categorized as \"mild\" (0\u00b0C to \u22126.7\u00b0C), \"moderate\" (\u22127.2\u00b0C to \u221213.9\u00b0C), and \"severe\" (\\<\u221213.9\u00b0C) under dry winter cattle coat conditions ([@b12-ajas-29-3-436]). [@b19-ajas-29-3-436] suggested that animal susceptibility to stress varies more for cold stress than for heat stress dependent on breed, stage of animal's life cycle. Ambient temperatures well above the estimated lower critical value affect feedlot cattle growth performance ([@b27-ajas-29-3-436]; [@b28-ajas-29-3-436]). Studies have suggested that energy requirements increased during winter or when animals are under cold stress due to the need to increase resting heat production to maintain body temperature by shivering or other thermogenic process ([@b30-ajas-29-3-436]; [@b7-ajas-29-3-436]). Thus, the lower feed efficiency during March compared to April may have been due to the increased energy utilization for heat production at a colder ambient temperature. In results of our study, both climate condition and age effects may be compounded because climate effects were compared at different growing period. Generally, the younger animal has higher feed efficiency. But, feed efficiency was higher at period 2 in older age. Thus, climate condition may be a major effect, and age effect may be minor since only one month difference between 9.7 and 10.6 months of age.\n\nBlood cortisol and metabolites\n------------------------------\n\nBlood NEFA concentrations were higher during the colder month of March than those during April and May (p\\<0.05; [Figure 1](#f1-ajas-29-3-436){ref-type=\"fig\"}), suggesting that lipolysis may occur at colder ambient temperatures to generate heat for maintaining body temperature. However, cortisol concentrations were not changed. These results suggest that the decrease in NEFA concentration may not be regulated by cortisol. Blood NEFA concentrations in dairy cattle often increase when feed intake cannot support their energy requirements, requiring the mobilization of NEFA by lipolysis of fat depots to support energy demand ([@b3-ajas-29-3-436]). [@b5-ajas-29-3-436] reported that NEFA concentrations increase under low temperature conditions. Calves in a cold environment (4.7\u00b0C) also have higher NEFA concentrations than those in a warm indoor environment (15.5\u00b0C; Nonnecke et al., 2006). Taken together, increased NEFA concentrations during colder weather may help generate energy to maintain body temperature and growth. Stress is generally accompanied by an increase in cortisol concentration, and cortisol concentrations increase under cold conditions ([@b16-ajas-29-3-436]; [@b22-ajas-29-3-436]). In our study, the minimum temperature (1.0\u00b0C) during March may not have been sufficiently cold to stimulate cortisol secretion. Generally, amount of diet feed consumed significantly affects blood metabolites and hormone concentrations. Therefore, we collected blood after 9 h fasting to minimize such diet variation among animals. [@b20-ajas-29-3-436] and [@b25-ajas-29-3-436] reported that cortisol concentrations were increased during fasting. Thus, cortisol concentrations at fasting may be higher than that at feeding stage. Our study showed no significant difference in fasting cortisol concentrations between April and May. We cannot exclude that fasting condition masks temperature responses.\n\nIn the present study, blood concentrations of phospholipids, HDL, LDL, and cholesterol were lower during March and April than during May (p\\<0.05; [Figure 2](#f2-ajas-29-3-436){ref-type=\"fig\"}). In rat study, cold stress decreased serum total cholesterol and HDL concentrations ([@b23-ajas-29-3-436]). Consistent with our study, cold-stressed broilers have lower serum total cholesterol, HDL, and LDL concentrations ([@b6-ajas-29-3-436]). Collectively, our results suggest that ambient temperature may affect lipoprotein metabolism.\n\nIn our study, blood concentrations of glucose, TG, and total protein were not different (p\\>0.05) among months ([Figures 1](#f1-ajas-29-3-436){ref-type=\"fig\"} and [2](#f2-ajas-29-3-436){ref-type=\"fig\"}). Blood concentrations of urea nitrogen, GOT, GPT, phosphorus, and albumin were higher (p\\<0.05) during May than during March ([Figure 2](#f2-ajas-29-3-436){ref-type=\"fig\"}). Changes in these parameters due to cold stress or low temperature have not been reported previously.\n\nImmune cell populations\n-----------------------\n\nInflammatory responses are often associated with the induction of acute-phase reactants, such as the C3 and C4 compartments of the complement system ([@b11-ajas-29-3-436]). In our study, blood C4 levels were higher (p\\<0.05) during March than April or May, whereas C3 level did not differ ([Figure 3](#f3-ajas-29-3-436){ref-type=\"fig\"}). Our results suggest that the C4 complement factor may be induced more during the colder March temperatures than other months. The functional significance of this change remains to be elucidated.\n\nThe granulocyte: lymphocyte ratio and B cell and monocyte populations did not differ among months. It is well known that glucocorticoids released during stressful events often modulate the immune system, including increases in the neutrophil: lymphocyte ratio and changes in the number of monocytes ([@b26-ajas-29-3-436]). Neutrophil: lymphocyte ratio changes occurred under acute and inflammatory conditions, such as weaning stress and tissue injury ([@b17-ajas-29-3-436]). It is probable that the ambient temperatures during March in the present study were not low enough to induce such changes.\n\nThe numbers of CD4^+^, CD8^+^, and CD4^+^CD25^+^ T cells were higher (p\\<0.05), and the number of CD8^+^CD25^+^ T cells was lower (p\\<0.05), in March than in May ([Figure 3](#f3-ajas-29-3-436){ref-type=\"fig\"}). These results suggest that the number of helper T cells, cytotoxic T cells, and regulatory T cells are likely modulated during the colder month of March compared to May. T cells were increased significantly in the spleen of acute cold-stressed mice, whereas CD4+ and CD8+ cell populations were minimally increased ([@b18-ajas-29-3-436]). Whether these changes in blood T cell populations are directly associated with ambient temperature must be clarified in cattle.\n\nIn conclusion, the minimum ambient temperature was lower in March (1\u00b0C) than in April (6\u00b0C). Daily weight gain and feed efficiency were lower in March than in April, suggesting that ambient temperature affects growth and feed efficiency in Korean cattle steers. Higher circulating NEFA concentrations during March compared to April suggest that lipolysis may occur at colder temperatures to generate heat to maintain body temperature. Blood CD4+, CD8+ and CD4+CD25+ T cell populations were higher in March than in May, suggesting that ambient temperature affects blood T cell populations.\n\nThis study was supported by a grant from the Bio-industry Technology Development Program (No. 313020-04-1-HD030), Ministry of Agriculture, Food, and Rural Affairs, Republic of Korea, and by the Next Generation BioGreen 21 Program (No. PJ01114001), Rural Development Administration, Republic of Korea.\n\n**CONFLICT OF INTEREST**\n\nWe certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.\n\n![Serum cortisol and lipid metabolite concentrations in Korean cattle steers over several months. Values are means+standard error. Minimum ambient temperature (MT) is shown in the top-left panel. Mean values with different letters differ significantly (p\\<0.05). NEFA, non-esterified fatty acid; HDL, high density lipoprotein; LDL, low density lipoprotein.](ajas-29-3-436f1){#f1-ajas-29-3-436}\n\n![Serum concentrations of glucose, blood urea nitrogen (BUN), total protein (TP), glutamic oxaloacetic transaminase (GOT), glutamic pyruvate transaminase (GPT), albumin, and phosphorus in Korean cattle steers over several months. Minimum ambient temperature (MT) is shown in the top-left panel. Values are means+standard error. Mean values with different letters differ significantly (p\\<0.05).](ajas-29-3-436f2){#f2-ajas-29-3-436}\n\n![Serum concentrations of complement factors, the granulocyte: lymphocyte ratio, and monocyte, T cell, and B cell populations in Korean cattle steers over several months. Minimum ambient temperature (MT) is shown in the top-left panel. Values are means+ standard error. Mean values with different letters differ significantly (p\\<0.05).](ajas-29-3-436f3){#f3-ajas-29-3-436}\n\n###### \n\nIngredients of concentrate diet and composition of experimental diets for Korean cattle steers\n\n  Items                                                                  \\% DM\n  ---------------------------------------------------------------------- -------\n  Ingredients of concentrate diet                                        \n  \u2003Ground corn                                                           15.8\n  \u2003Ground wheat                                                          18.0\n  \u2003Salt                                                                  0.88\n  \u2003Molasses                                                              5.50\n  \u2003Wheat bran                                                            3.00\n  \u2003Corn flour                                                            5.00\n  \u2003Rice bran                                                             3.00\n  \u2003Cottonseed hulls                                                      1.50\n  \u2003Palm kernel meal                                                      10.0\n  \u2003Ammonium chloride                                                     0.15\n  \u2003Rapeseed meal                                                         2.22\n  \u2003Dried distilled grain- soluble                                        9.38\n  \u2003Condensed molasses- soluble                                           1.50\n  \u2003Corn gluten feed                                                      8.50\n  \u2003Limestone                                                             3.30\n  \u2003Copra meal                                                            10.0\n  \u2003Porphyry                                                              2.00\n  \u2003Mineral/vitamin premix[1](#tfn2-ajas-29-3-436){ref-type=\"table-fn\"}   0.25\n  \u2003Total                                                                 100.0\n  Chemical composition                                                   \n  \u2003Concentrate diet                                                      \n  \u2003\u2003DM                                                                   87.8\n  \u2003\u2003CP                                                                   14.5\n  \u2003\u2003Fat                                                                  3.23\n  \u2003\u2003Ash                                                                  7.80\n  \u2003\u2003Ca                                                                   1.43\n  \u2003\u2003P                                                                    0.50\n  \u2003Timothy hay                                                           \n  \u2003\u2003DM                                                                   90.7\n  \u2003\u2003CP                                                                   7.84\n  \u2003\u2003Fat                                                                  2.13\n  \u2003\u2003Ash                                                                  8.05\n  \u2003\u2003Ca                                                                   0.24\n  \u2003\u2003P                                                                    0.20\n  \u2003\u2003ADF                                                                  35.5\n  \u2003\u2003NDF                                                                  59.3\n\nDM, dry matter; CP, crude protein; ADF, acid detergent fiber; NDF, neutral detergent fiber.\n\nMineral and vitamin premix contained vitamin A 2,650,000 IU, vitamin D~3~ 530,000 IU, vitamin E 1,050 IU, niacin 10,000 mg, Mn 4,400 mg, Zn 4,400 mg, Fe 13,200 mg, Cu 2,200 mg, iodine 440 mg, and Co, 440 mg/kg of additive Grobic-DC (provided from Bayer Health Care, Leverkusen, Germany).\n\n###### \n\nMean, maximum, and minimum values of ambient temperatures, climate temperatures, and relative humidity at March and April of 2014\n\n  Items                      March[1](#tfn4-ajas-29-3-436){ref-type=\"table-fn\"}   April[2](#tfn5-ajas-29-3-436){ref-type=\"table-fn\"}   SE     p\n  -------------------------- ---------------------------------------------------- ---------------------------------------------------- ------ ---------\n  Ambient temperature (\u00b0C)                                                                                                                    \n  \u2003Mean                      8.70                                                 13.0                                                 0.88   \\<0.001\n  \u2003Maximum                   20.6                                                 21.4                                                 0.38   0.58\n  \u2003Minimum                   1.00                                                 6.24                                                 0.48   \\<0.001\n  Climate temperature (\u00b0C)                                                                                                                    \n  \u2003Mean                      4.60                                                 9.80                                                 0.66   \\<0.001\n  \u2003Maximum                   10.6                                                 15.8                                                 0.81   \\<0.001\n  \u2003Minimum                   \u22121.7                                                 3.4                                                  0.64   \\<0.001\n  Relative humidity (%)                                                                                                                       \n  \u2003Mean                      61.6                                                 63.7                                                 1.43   0.47\n  \u2003Maximum                   87.0                                                 88.0                                                 0.97   0.63\n  \u2003Minimum                   37.3                                                 40.3                                                 1.98   0.47\n\nSE, standard error.\n\nMarch 7 to April 3 (4 weeks).\n\nApril 4 to May 1 (4 weeks).\n\n###### \n\nGrowth performance of Korean cattle steers at March and April of 2014\n\n  Items                                                           March[1](#tfn7-ajas-29-3-436){ref-type=\"table-fn\"}   April[2](#tfn8-ajas-29-3-436){ref-type=\"table-fn\"}   SE      p\n  --------------------------------------------------------------- ---------------------------------------------------- ---------------------------------------------------- ------- ---------\n  Age (month)                                                     9.7                                                  10.6                                                 0.13    \n  Body weight[3](#tfn9-ajas-29-3-436){ref-type=\"table-fn\"} (kg)   309                                                  347                                                  4.60    \\<0.001\n  Feed intake (DM)                                                                                                                                                                  \n  \u2003Total daily feed intake (kg/d)                                 6.88                                                 7.68                                                 0.10    \\<0.001\n  \u2003Daily concentrate intake (kg/d)                                3.54                                                 3.97                                                 0.06    \\<0.001\n  \u2003Daily forage intake (kg/d)                                     3.34                                                 3.71                                                 0.05    \\<0.001\n  Average daily gain (kg/d)                                       1.13                                                 1.38                                                 0.04    \\<0.001\n  Feed efficiency (gain/feed)                                     0.16                                                 0.18                                                 0.004   0.02\n\nSE, standard error; DM, dry matter.\n\nMarch 7 to April 3 (4 weeks).\n\nApril 4 to May 1 (4 weeks).\n\nInitial body weight was 277.1 kg. Body weight was recorded at April 4 and May 2, respectively.\n\nN = 18.\n"}
+{"text": "Quantifying cell-to-cell heterogeneity in the tissue context {#Sec1}\n============================================================\n\nIntra-tumor heterogeneity is currently accepted as a hallmark of cancer, being present in virtually all tumor traits \\[[@CR1]\\]. Sensitive molecular techniques developed in the last few years have allowed a detailed genetic and phenotypic deconvolution of intra-tumor heterogeneity. These include genome-wide analysis of bulk tumor samples to describe evolutionary trajectories in relapsed tumors and genomic divergence between primary tumors and metastases \\[[@CR2]-[@CR4]\\], as well as single-cell genomic profiling \\[[@CR2],[@CR5]\\]. However, despite methodological improvements in the molecular characterization of single cells, the accurate interpretation of intra-tumor heterogeneity requires the inference of cell-to-cell variability within a particular tissue context, which can only be directly assessed by *in situ* analysis. Microenvironmental constraints within spatially restricted areas of a tumor can exert differential selective pressures, leading to the manifestation and the selection of different phenotypes and particular genotypes. For instance, different oxygen levels, the presence of inflammatory cells, or the physical interaction with extracellular components in different parts of a tumor \\[[@CR6]-[@CR8]\\] can influence cellular phenotypes and contribute to different trajectories in the evolution of a tumor \\[[@CR2]\\]. Therefore, the accurate interpretation of cellular phenotypic and genomic heterogeneity requires tissue-context specificity \\[[@CR9]-[@CR11]\\].\n\n**IFISH: Immunofluorescence***in situ***hybridization***In situ* fluorescence-based detection of proteins, DNA, and RNA enables the simultaneous detection of multiple markers in single cells by epifluorescence, confocal, or multispectral imaging technology. Combining both immunofluorescence and fluorescence *in situ* hybridization (IFISH) allows multiplexing to detect both genomic and phenotypic traits at the single cell level \\[[@CR10]\\]. This approach captures cell-to-cell variations missed in cell population analyses while preserving specific microenvironmental contexts. As an *in situ* analysis, IFISH allows the spatial mapping of individual cells to measure topological heterogeneity. Visualizing topological heterogeneity can have important implications in predicting treatment response, as well as tailoring treatment to suit the diverse cell populations observed within a tumor \\[[@CR10]\\]. However, these *in situ* studies require the analysis of multiple markers in thousands of cells, are very time-consuming, and their reproducibility could be influenced by variability between users \\[[@CR12]\\]. Therefore, there is an urgent need for the development of objective analytical tools that minimize scoring subjectivity and facilitate the quantification of multiple traits in single cells while preserving context specificity. The implementation of these tools in both basic and translational research will advance our understanding of tumor biology and will facilitate biomarker discovery and validation.\n\n**GoIFISH**:**quantifying tumor heterogeneity in IFISH images** For application to IFISH, accurate segmentation at the nuclear, membrane and spot level are critical for subsequent analysis, which often interrogates clonal populations or evaluates relationships between protein and genomic expression. Objective integration of protein expression and copy number requires not only accurate segmentation, but also the separation of normal cells from tumor cells, and appropriate background subtraction associated with auto-fluorescence. Very few existing softwares allow manual alterations of small inaccuracies in cell segmentation and often incorrect cell classification results cannot be changed. Visual scoring by a trained observer (e.g. pathologist) is the gold standard for detecting cells and automated image analysis systems developed to complement pathologist scoring require user validation \\[[@CR12]\\].\n\nTo address these challenges, we have developed a semi-automated system which provides users with an automated starting point in segmentation and can readily accept user input to improve the segmentation result. GoIFISH is able to segment nuclei, locate and estimate spots, detect membranes, measure morphological and intensity properties and classify cells. GoIFISH is a versatile method that allows researchers to determine and quantify, for instance, the amplification status of single locus within cells, together with the detection of phenotypic markers present in different subcellular locations. It preserves the tissue context specificity and provides coordinates for the topological mapping of each cell. Simple topology maps can be displayed to illustrate spatial variations within an image with respect to two given stains. GoIFISH allows users to analyze FISH, IF or IFISH images containing a maximum of 5 markers, of which one must be the nuclear marker DAPI. We validated our software in a pilot HER2+ breast cancer cohort of 10 samples and compared its performance with existing softwares.\n\n**Related approaches** Several general image processing methods including CellProfiler \\[[@CR13]\\], Icy \\[[@CR14]\\], OMERO \\[[@CR15]\\], ImageJ \\[[@CR16]\\], CellTracker \\[[@CR17]\\] and ImageM \\[[@CR18]\\] have been developed for the quantitative analysis of images, and the capabilities of each software are described in Table [1](#Tab1){ref-type=\"table\"}. Table 1**Comparison between softwares available for cell segmentationPropertyGoIFISHColumbusCellProfilerImageJImageMCellTracker**CostOpen-SourceProprietaryOpen-SourceOpen-SourceOpen-Source on requestOpen-SourceReal-Time UpdateYesYesNoNoNoYesBackground Intensity SubtractionYesTuning but no subtractionCorrect Illumination CalculateFlatten IlluminationNoNoA-Priori knowledge required (eg. cell size or segmentation methods)NoNoYesNoNoNoNuclear SegmentationYesYesYesYesManualYesMembrane SegmentationYesYesYesYesManualYesSpot DetectionYesYesYesYesYesNoBatch ProcessingCmdLine OnlyYesYesBatchCommand PluginNoNoVisualise Batch Segmentation ResultsYesNoYesNoNoNoManual EditingYesNoNoNoYesYesCell Specific InformationYesYesYesYesYesYesSummary ReportYesYesYesYesYesYesTopology or summary MapsYesNoYesTopit PluginNoNo\n\nOMERO is a platform for the storage and annotation of microscopy images \\[[@CR15]\\], and Icy and ImageJ have been developed as general platforms for image analysis \\[[@CR14],[@CR16]\\]. All three are dependent on the development of plug-ins for specific applications from its user-base. OMERO currently does not have automated algorithms for image segmentation, and is dependent on user input for the delineation of cell boundaries or other regions of interest. ImageJ and Icy have a series of plugins for nuclear segmentation, membrane segmentation and spot detection, however these three processes are often disjoint and will require user effort to collate these results. We have used MATLAB as the platform for developing GoIFISH due to its wide user-base, strong image analysis capabilities and comprehensive data analysis features.\n\nSoftwares with specific cell segmentation capabilities include Columbus, CellProfiler \\[[@CR13]\\], ImageM \\[[@CR18]\\] and CellTracker \\[[@CR17]\\]. CellProfiler and Columbus are programs which specialize in the segmentation of cells from the cytoplasmic and nuclear level, down to the subcellular or genomic level. These have been developed primarily for high-throughput analysis of cells in culture, and often are based on assumptions about the regularity of size and morphology within the cell population of interest. These softwares are optimized to have minimal segmentation errors in cell-culture images, however, may not be directly applicable to real tissue.\n\nImageM \\[[@CR18]\\] is a software developed for detection and counting of RNA signals using a semi-automated approach. Users can refine results, however, to extract features on a per nucleus or cell basis, manual delineation of regions of interest is required. CellTracker \\[[@CR17]\\] has been developed primarily for the live-tracking of cells, and a semi-automated approach to nuclear and cytoplasmic segmentation is also applied.\n\nCurrently available segmentation softwares are not tailored for the segmentation of IFISH images, due to the heterogeneous nature of tumor populations and complex tissue structure. This has motivated the development of GoIFISH to perform accurate nuclear, membrane and spot detection, while allowing the user the freedom to manually edit segmentation outputs (Figure [1](#Fig1){ref-type=\"fig\"}). This is critical as a starting point for the analysis of tumor subpopulations, and the extraction of biologically relevant features from the images. Figure 1**Overview of** **GoIFISH**, **including computational capabilities.** **(A)** The GoIFISH graphical user interface **(B)** Outputs from GoIFISH including nuclear and membrane segmentation, spot detection, manually edited segmentation, and topology maps **(C)** Cell classification. After cell segmentation, cells are labelled as tumor (green) or stromal (yellow) by the user. Cells are automatically classified using the marked cells as training data.\n\nAvailability of GoIFISH {#Sec5}\n=======================\n\nGoIFISH is freely available at www.sourceforge.net/projects/goifish/ under the GNU General Public License version 2. GoIFISH is written in MATLAB and all source code is provided to allow analysis on both command line and through the Graphical User Interface (GUI) (Additional file [1](#MOESM1){ref-type=\"media\"}). It is dependent on OMERO Bio-formats for the conversion of images to the correct file format for loading \\[[@CR19]\\]. All images used in this study are available at the given link. The GUI has been created into a stand-alone program operable on Windows and Mac OS systems following the installation of the appropriate version of MATLAB Compiler Runtime (v7.14), provided at the given link.\n\nThe user-guide for the software is included as Additional file [2](#MOESM2){ref-type=\"media\"}. The types of images for optimal use in GoIFISH are described in Table [2](#Tab2){ref-type=\"table\"}. Note that due to memory constraints, there are size limitations of 12 megapixels for comfortable use in the GUI however, this limitation can be overcome using the GoIFISHWrapper in the MATLAB environment. Table 2**Optimal properties of images for analysis in** **GoIFISHGUIGoIFISHWrapper (Command Line)**File Format.mat or.tiff.tiff.zvi Use bio-formats to convert other formatsImage Size12 megapixels for comfortable useTheoretically unlimitedNumber of stainsUp to 5. Must include DAPIUnlimited but must include DAPINumber of CellsUnder 1000 for comfortable useTheoretically unlimitedCell Size (px)Optimally 60x magnification (2500-10000 px), 20x and 40x magnificationalso available (250-3000 px)\n\nThe GoIFISH workflow {#Sec6}\n====================\n\nThe following paragraphs give a detailed overview of the GoIFISH workflow. An overview of the capabilities of GoIFISH, including the user interface is shown in Figure [1](#Fig1){ref-type=\"fig\"}.\n\n**Step 1: Loading and preprocessing data**Images can be loaded as a.mat file containing a cell array or a.tiff file into the GUI. After successful loading, the first image in the series will be presented (Figure [1](#Fig1){ref-type=\"fig\"}A). These are automatically adjusted to ensure 1% of the image is saturated at lower and higher intensities, which is more suited for nuclear or cytoplasmic images but may saturate spots. The 'RangeScale' option is recommended in these scenarios. Brightness and contrast of each image can be adjusted for auto-fluorescence, to ensure an optimal dynamic range and to prevent saturation of the image. This is critical for good segmentation results. Note that the image adjustment will improve the user experience and segmentation results, however, the intensities can be measured from the raw image for comparative quantitation.\n\nGoIFISH allows background intensity adjustment, both using a single global intensity value, or on a cell specific level for nuclei. The user draws 'background' regions using the paintbrush tool, from which the mean background intensity is calculated. This is subtracted from either raw intensities, which is recommended for comparisons between samples, or from an adjusted image. A per cell nuclear background adjustment is also available, whereby a background intensity is calculated at a margin of 2-6 pixels from the edge of each segmented nucleus. This will be computed automatically for all nuclear stains, and will account for local variation in background intensity.\n\nTo finalize the preprocessing stage the user will need to indicate the stain type in each channel, of which one must be DAPI, and the magnification of the image. 60x, 40x and 20x magnifications are permitted, however higher resolutions are recommended for the accurate detection of spots. Following this, a quick preprocessing step is applied to assign the stain type to each image. After this, segmentations on either the DAPI channel alone or all channels using default parameters can be performed (Figure [1](#Fig1){ref-type=\"fig\"}B).\n\n**Step 2: Nuclear segmentation** The foreground or cellular portion of the image is automatically detected by Otsu thresholding \\[[@CR20]\\] on a combined image of entropy and intensity, which is used as a mask in cell segmentation. Nuclei are segmented using an iterative H-minima transformed watershed \\[[@CR21]\\], where the local intensity depth of pixels under a given threshold is suppressed, and a watershed is applied. Fragments attained from each step are classified as either optimal, undersegmented, or oversegmented. Cells with optimal properties are selected, and the remaining image is subjected to segmentation at a lower threshold. We have developed an approach to mitigate oversegmentation by joining neighboring fragments according to their morphological features (see [Material and methods](#Sec25){ref-type=\"sec\"}). There is also an option of performing a seeded-watershed \\[[@CR22]\\] for images with a small number of cells or poor contrast at cell boundaries, whereby the user indicates the cell locations with a series of spots. A H-minima watershed is then applied using this information.\n\n**Step 3: Membrane segmentation** Following segmentation, the user has the option of narrowing down the population of interest using a minimum size threshold, or edit the borders manually (see Step 5). It is recommended that the user checks the output of the nuclear segmentation before proceeding as membrane segmentation and spot detection are dependent on the nuclear map generated.\n\nMembrane segmentation is performed by combining a Voronoi segmentation of nuclei with the intensity information of the image. High intensity edges within the image are set as local maxima to ensure segmentation occurs along these edges. Fragments are then merged based on their location with respect to the nuclear segmentation. This result can be further refined using active contours, such as Chan-Vese Segmentation \\[[@CR23]\\] or Localized Segmentation \\[[@CR24]\\].\n\n**Step 4: FISH detection** For single spot detection, a Laplacian of Gaussian filter is applied to the image to determine candidate spots. Spots which are also local maxima in the gradient image are selected. The user has an option of entering an expected size threshold (Default minimum spot size of 15 pixels for 60x images), and a preferred intensity threshold. An optional morphology classifier (see [Material and methods](#Sec25){ref-type=\"sec\"}) can be applied to differentiate between spots and artefacts in images with low contrast signal, such as signals from centromere 17.\n\n**Step 5: Manually editing segmentations**GoIFISH provides a toolbox to manually edit segmentations if needed. Manual editing of the segmentation output can be easily achieved by drawing a border between cells with the 'scissors' tool, oversegmented cells can be 'glued' together, and artefacts can be 'trashed'. Regions can be 'painted' or 'erased'. All operations are terminated by right clicking, and pressing the 'escape' exits a particular editing mode.\n\n**Step 6: Post segmentation processing**Segments from each channel need to be mapped to the DAPI channel in order to construct a matrix of features, using the 'Update' function. An error will appear if there are inconsistencies in the segmentation, such as two nuclei mapping to one membrane. Following successful mapping, the user can generate heatmaps and topology maps to visualize staining variations within the image (Figure [1](#Fig1){ref-type=\"fig\"}B) and perform cell classification (Figure [1](#Fig1){ref-type=\"fig\"}C).\n\nCells in the image are classified by support vector machine into 4 possible cell types. The user labels candidate cells for each class (for example, to separate fibroblasts from tumor cells), and applies the classification. This classification is based on morphological parameters but can also include intensity information. The classification result can be manually corrected if inconsistencies appear.\n\n**Step 7: Output from GoIFISH**The output from GoIFISH can be saved as a series of images and a.csv file with cell specific measurements. These include intensity measurements, including raw and background adjusted intensities, morphologial parameters such as area, perimeters, axis lengths, the location of the centroid of each nuclei, the cell label if classification is performed and copy numbers in the case of spot detection. Data can be downloaded into the MATLAB environment, or saved as a progress.mat file where processing can be resumed in another session.\n\n**GoIFISHWrapper: Combining Steps 1-4**GoIFISH provides a wrapper for batch analysis, which is implemented via command line. The user simply provides the filepath of interest and edits a Parameter File which contains information such as the stains used, the magnification of the image and the segmentation parameters. The results are automatically saved as progress.mat files which can be loaded into the GUI for segmentation editing, background selection and cell labelling. Other benefits of running the wrapper include unlimited number of stains and the analysis of larger images. However, it is recommended that each image does not exceed a resolution of 12 megapixels if the user wishes to edit cells in the GUI. In this circumstance, it is recommended that the image is sectioned into a number of smaller constituent images which are analyzed independently.\n\nGoIFISH performance and benchmarking {#Sec15}\n====================================\n\nThe performance of GoIFISH was compared to two state-of-the-art image analysis systems, the proprietary Columbus software from Perkin-Elmer, and the open-source CellProfiler from the Broad Institute \\[[@CR13]\\]. All three image analysis softwares can detect nuclei, FISH signals, cytoplasm staining, and report morphological properties including size, and intensities. Details of all statistical analyses including code to reproduce plots are contained in the Additional file [3](#MOESM3){ref-type=\"media\"}.\n\nComparing **GoIFISH** to existing automated methods {#Sec16}\n---------------------------------------------------\n\nColumbus has an intuitive interface, real-time feedback, automatic detection of approximate cell sizes, with very little image processing knowledge required to operate the system (Table [1](#Tab1){ref-type=\"table\"}). On the other hand, CellProfiler has the benefits of wider functionality as its open-source nature allows its user base to develop and maintain specialized functions. However, it requires a-priori knowledge about the images and different segmentation methods, which may take the user a long time to develop an optimal segmentation pipeline. The parameters used to segment images in these two programs are described in [Material and methods](#Sec25){ref-type=\"sec\"}.\n\nGoIFISH was tested in two scenarios. In order to perform fair benchmarking, images from 10 samples were run in GoIFISH on the default settings (see [Material and methods](#Sec25){ref-type=\"sec\"}). In addition, we tested its capabilities for improvement with user input.\n\nAn example of the segmentation output from all three softwares is shown in Figure [2](#Fig2){ref-type=\"fig\"}A. GoIFISH, using its default segmentation parameters, demonstrated precision and recall in nuclear and spot detection which are comparable to both Columbus and CellProfiler (Figure [2](#Fig2){ref-type=\"fig\"}B, Additional file [4](#MOESM4){ref-type=\"media\"}: Figure S1B, Mean F-Score: 0.68,*N*=10). From visual inspection of the segmentation output (Figure [2](#Fig2){ref-type=\"fig\"}A), Columbus does not perform nuclear segmentation when the borders are not well defined. CellProfiler applies edges to ensure segments are within the defined range of cell diameters, resulting in segmentation inaccuracies. GoIFISH with default parameters outperformed existing softwares in membrane detection (Figure [2](#Fig2){ref-type=\"fig\"}B, Mean F-Score 0.86, N=10) with results very similar to the manually edited GoIFISH result. All methods demonstrated high precision as membranes are detected around nuclei, but varying degrees of recall (Additional file [4](#MOESM4){ref-type=\"media\"}: Figure S1B). We observed that Columbus treats membranes with high intensity as cytoplasmic regions (Figure [2](#Fig2){ref-type=\"fig\"}A). Figure 2**Performance benchmarking against Columbus and CellProfiler.** **(A)** Example of nuclear, membrane and spot segmentation using CellProfiler, Columbus, and GoIFISH manually corrected segmentations. **(B)** Average F-Scores for nuclear segmentation, membrane detection, spot detection (including clusters) in 10 sample images. **(C)** Differences in perimeter-area ratio of segmented nuclei and membranes compared to the gold standard. Values closer to 1 indicate similar morphology to the gold standard.\n\nIn Cent17 and *HER2* detection GoIFISH with default parameters surpassed the two existing methods (Mean F-Scores: 0.69, 0.83 for cent17 and *HER2* respectively). Columbus has high recall but poor precision, reflecting a higher false positive rate. It should be noted that the manually curated samples had variable F-score, which may be a subsequent propagation error from nuclear segmentation.\n\nWhile precision and recall assess the presence of an object, they provide no information on how accurate the morphology of the segmentation is. For instance, encroachment errors with slight misplacement of cell boundaries will have no effect on the F-Score. Therefore, the perimeter-area ratio was measured for each cell and compared to the gold standard as an assessment of whether the correct shape was detected. Each individual spot in Figure [2](#Fig2){ref-type=\"fig\"}C represents the average difference in perimeter-area ratio for one particular image. Points centred around 1 indicate very little variation in shape compared to the gold standard. In both nuclear and membrane segmentation, GoIFISH values were closer to 1 compared to Columbus and CellProfiler, indicating that the morphology was better represented.\n\nTime-benchmarking was performed between GoIFISH and the newest version of CellProfiler (v2.1.0) (Table [3](#Tab3){ref-type=\"table\"}). Columbus operates on a server and thus a direct comparison was not applicable. Timings were performed on a 2x2.4 GHz Quad-Core Intel Xeon processor with 6GB RAM, on 5 candidate samples with an increasing number of cells. GoIFISH performs comparably to CellProfiler when the image contains well defined cells, but the processing time increased with greater image complexity. For example, processing image 6370 took the longest as it has a high number of cells with invasive phenotype and poorly defined cell boundaries. All segmentations can be conducted within 1.2 minutes, which despite being longer than CellProfiler, is still sufficiently short for mainstream use. Table 3**Timing comparisons between** GoIFISH **and CellProfilerSample74616361743576196370**Approximate Number of Cells206080100120CellProfiler Time (s)2625293032GoIFISH Time (s)2627404273\n\nCorrelating **GoIFISH** output with visual interpretation {#Sec17}\n---------------------------------------------------------\n\nWhile precision-recall testing allows reliable assessment of segmentation accuracy, the obtained data must be reflective of the biology to draw valid conclusions. Automated scoring of protein intensities and spot areas were compared with visual pathologist scoring on a single-cell level, with a total of 355 cells scored for ER staining, membrane HER2 intensity and cent17 and *HER2* copy number.\n\nIn a first analysis, we assessed immunofluorescence of ER and HER2 to determine (1) whether the distribution of cell intensities within an image is reflective of the semi-quantitative scoring by a trained observer and (2) whether two cells with the same scoring in two different images are directly comparable. All methods detecting the nuclear ER stain showed a correlation between the semi-quantitative scoring and intensity measurements, however, samples were not shown to be directly comparable to each other. As an example, a 'positive' cell in sample 6361 was on similar intensity to a 'moderately' stained cell in sample 6370 (Figure [3](#Fig3){ref-type=\"fig\"}A). Figure 3**Correlation between measured intensities and classification by a pathologist.** Comparison of semi-quantitative scoring by a pathologist with CellProfiler, Columbus and GoIFISH quantitation in **(A)** ER staining, **(C)** HER2 membrane completeness using mean intensities **(E)** HER2 membrane intensity using coefficient of variance. Combined distributions of quantitations across all 10 samples for **(B)** ER positivity, **(D)** HER2 membrane completeness using mean intensities and **(F)** HER2 membrane intensity using coefficient of variance. Note that GoIFISH nuclear adjusted intensities were used in **(A, B)**, GoIFISH background adjusted intensities were used in **(C, D)** and GoIFISH raw intensities used in **(E, F)**.\n\nANOVA in conjunction with Tukey's range test was performed on samples with 'negative' expression to determine whether the baseline means are directly comparable to each other. Out of the 45 possible pairwise comparisons, Columbus had 26 pairwise comparisons, CellProfiler had 13 and GoIFISH had only 10 pairwise comparisons which showed a significant difference in baseline mean (*p*\\<0.05). Sample 7360 had a negative intensity after background correction using GoIFISH, but in practice would be assigned a value of 0 which would further lower the number of significant differences. Using a per nucleus specific background subtraction method in GoIFISH, 'positive' samples become comparable to each other across all images (Figures [3](#Fig3){ref-type=\"fig\"}A, Additional file [5](#MOESM5){ref-type=\"media\"}: Figure S2A). Figure [3](#Fig3){ref-type=\"fig\"}B illustrates the right ordering and statistical difference between each class, demonstrating that the method can reproduce quantitatively the visual scoring (T-test, *p*\\<0.01 between all categories).\n\nThe same analysis was applied to HER2 membrane staining to determine whether the intensity could recapitulate the membrane completeness in cells. HER2 protein assessment in a clinical setting often uses patterns of staining to guide subsequent treatment \\[[@CR25]\\]. Cells are classified as having 'negative' membrane staining, 'complete' positive staining or 'incomplete' positive staining. Both GoIFISH after background subtraction and Columbus demonstrated a step-wise increase in intensity with highest intensity observed in complete membranes (Figure [3](#Fig3){ref-type=\"fig\"}C,D). Combining all cells, a statistical difference at the correct order of classes was observed in only the GoIFISH background adjusted and manually edited samples (Figure [3](#Fig3){ref-type=\"fig\"}D, Additional file [5](#MOESM5){ref-type=\"media\"}: Figure S2D).\n\nThe coefficient of variation was also computed as a second metric of differentiating between 'complete' and 'incomplete' membranes. It is expected that 'complete' membranes have a lower variation than 'incomplete' membranes. In both GoIFISH and Columbus, an increased coefficient was observed in the samples with broken membranes compared to the samples with complete membranes, this was however only statistically significant in GoIFISH but not in Columbus (Figure [3](#Fig3){ref-type=\"fig\"}E,F, Additional file [5](#MOESM5){ref-type=\"media\"}: Figure S2C,D).\n\nIn a second analysis, we assessed how accurately an automated system can estimate the number of spots within each cell based on the area measured. In 9 samples (Sample 7619 was excluded due to overexposure of the channel and thus high false positive rate), we assessed the correlation between manually counted copy number per cell with the automatically detected spot area (Figure [4](#Fig4){ref-type=\"fig\"}). In samples where the distinct number of copies cannot be counted due to amplification, a value of 22 was assigned (which was one more than the upper observable limit of 21 spots). Most relationships appeared to be linear for centromere 17, with the exception of measurements by CellProfiler. The estimates of *HER2* spots by GoIFISH and Columbus showed a strong linear correlation with manually curated copy numbers, which plateaued at a value of approximately 400 pixels. Similar gradients were shown for centromeric and *HER2* spots, and in both cases a linear regression will approximate an area of 20 pixels per spot. Figure 4**Correlating copy number with spot area at a single cell level.** Correlation between manually counted spots and automated detection of spot area using Columbus, CellProfiler and GoIFISH. A linear relationship between copy number and area is observed until 21 spots per cell, after which individual spots are no longer discernable by eye. Darker regions indicate a higher population of cells with similar properties.\n\nEffects of user variability on **GoIFISH** outputs {#Sec18}\n--------------------------------------------------\n\nUser input may influence background intensity correction, cell segmentation results and cell classification. GoIFISH has been developed with a number of strategies to minimize the effects of inter-user variation.\n\nTo determine the variation in background selection between users, measurements were made by two trained observers and one untrained observer who was given background selection-guidelines (see User Guide in Additional file [2](#MOESM2){ref-type=\"media\"}). All three observers demonstrated 88% or greater correlation with the gold standard (Figure [5](#Fig5){ref-type=\"fig\"}A), demonstrating that our guidelines are sufficient for an 'untrained observer' to attain a similar scoring as a 'trained observer'. Figure 5**Robustness of** GoIFISH **performance with user input.** **(A)** Inter-user variation in background intensity selection from three observers (two trained) when compared to the Gold Standard. **(B)** Heatmap of coefficient of variation of background intensity in all sample images. Larger shape indicates a higher mean. **(C)** Inter-user variation in area of segmented cells (nuclear and membrane) and **(D)** corresponding intensity of HER2 membranes and ER. **(E)** Permutation testing of accuracy of classifier using either only morphological information or morphological and intensity information for myopepithelial-luminal discrimination (left) and lymphocyte-stromal-tumor discrimination (right).\n\nTo address the issue of background heterogeneity within an image, a trained observer selected four different background regions within each image to compare intensities. These values are reported as coefficients of variation (Figure [5](#Fig5){ref-type=\"fig\"}B), where the size of each box proportional to the mean reported intensity. In most images, a low variation of 10% or less was observed, with the exception of 6361 and 7916 which had high auto-fluorescence and overexposure respectively. The greatest variation was observed in the DAPI channel, which is a general DNA marker used to assess the quality of the sample. In practice, DAPI intensities are rarely measured for quantitative analysis. The other stains are more selective and specific for a particular protein or locus of interest, and have demonstrated greater stability in background intensities.\n\nManual editing of segmentation results is also prone to user subjectivity. To reduce both this effect and manual labor, GoIFISH was designed with a toolbox which minimizes the amount of clicks or mouse-drawing performed by the user. For example, the merging of cells requires two clicks of the mouse, and the segmentation of overlapping cells requires one line to be drawn. These features ensure that the morphology and boundary of cells are consistent in each image irrespective of the user.\n\nTo determine the effectiveness of these tools, two independent scorers manually edited 50 missegmented cells across the 10 test images. The nuclear and cytoplasmic areas were measured, alongside nuclear ER intensity and HER2 membrane intensity. Nuclear segmentation was consistent between the two scorers (*r*=0.86, Figure [5](#Fig5){ref-type=\"fig\"}C), however a number of cells were considered to be larger by Scorer 1 than Scorer 2. The discrepant cells were determined to be mitotic, phenotypically characterised by the appearance of two nuclei in the DAPI channel yet sharing the same membrane in the HER2 channel, and considered as one cell by Scorer 1 but as two cells by Scorer 2. The cyotplasmic areas had lower correlation between the two observers (*r*=0.79), which can be attributed to the HER2 status of the cell. In the absence of a well defined HER2 membrane, the shape is open to interpretation, accounting for the greater variation in the HER2- cells than the HER2+ cells (HER2+ only: *r*=0.83).\n\nDespite the differences in morphology of the segmented cells, 99% correlation was observed in the raw recorded intensities (*r*=0.99 for both ER and HER2, Figure [5](#Fig5){ref-type=\"fig\"}D), demonstrating that intensity measurements are robust to differences in cell segmentation between users.\n\nFinally, we tested how the performance of the cell classifier depends on the size of the training set and cellular features in two different cell classification scenarios: (1) to differentiate myoepithelial from luminal cells and (2) to differentiate lymphocytes and stroma from tumor (Figure [5](#Fig5){ref-type=\"fig\"}E). Two images representing these two scenarios were labelled by a trained observer. Training sets of increasing size (starting from 2 cells) were created by randomly sampling the number of required cells, and where possible an equal number of cells from each class were selected. To determine the 95% confidence interval for classifier accuracy, 500 permutations of the training set for each size were used to predict the labels within an image. Our results demonstrate that using morphological parameters alone, the accuracy approaches 70% for myoepithalial-luminal discrimination, and 80% for lymphocyte-stromal-tumor discrimination. With the addition of stain information, the accuracy approached 95% and 100% accuracy respectively. The average accuracy of the classifier increases with a larger number of labelled cells, however, if well-chosen, high classification accuracy can still be attained with a training set of under 10 cells.\n\nVisualizing the cellular diversity within an image {#Sec19}\n--------------------------------------------------\n\nFinally, we visualized the cellular diversity within the images, first by looking at IF staining to compare HER2 and ER intensities and also by comparing genomic expression with protein expression of HER2 (Figure [6](#Fig6){ref-type=\"fig\"}). A global intensity cut-off of 50 for ER after nuclear adjustment, and 300 for background adjusted HER2 staining was implemented to define cells as positive or negative. These values were chosen from global assessment of the distribution of staining across all 10 samples (Additional file [5](#MOESM5){ref-type=\"media\"}: Figure S2B,D). These cutoffs yield 4 classes as shown in Figure [6](#Fig6){ref-type=\"fig\"}A. This procedure was applied to both the training set of cells, which were scored by a pathologist, and the remaining cells within the same image, which formed the test set. Figure 6**Visualizing cellular heterogeneity.** **(A)** Comparison of cellular diversity in IFISH images by applying a global intensity threshold of 300 for HER2 and 50 for ER on the gold standard cell distribution, the training distribution and test set. **(B)** Classification of cells in Sample 6370 based on HER2 and ER intensity and the corresponding topology map. The color indicates the relative ratio between the two stains. **(C)** Classification and topology map applied to HER2 protein expression (cyan) and *HER2* spot area (yellow). **(D)** Classification of cells in Sample 6361 into luminal or myoepithelial cells based on morphology, and topology map based on HER2 and ER staining.\n\nBoth the training and test set showed resemblance, however samples 6361 and 7619 differed from the gold standard distribution. These differences may be due to the arbitrary setting of one intensity threshold, and the subjectivity in visual scoring where background intensities between samples are seldom taken into consideration.\n\nFigure [6](#Fig6){ref-type=\"fig\"}B illustrates the spatial distribution of the classified cells in sample 6370 and the corresponding topology map which displays the relative ratio between HER2 and ER. Most of the cells which were considered as double negative are identified morphologically as stroma, and display low intensity in both proteins. Cells which were classified as double positive demonstrated subtle cell-to-cell variations which would not otherwise be observable with a strict threshold. The same analysis was applied to HER2 protein and *HER2* spots, with a cutoff placed at *HER2* spot area of 60 pixels which is roughly equivalent to 3 spots. Most cells exhibited a HER2 protein intensity increase with a spot area increase (shown in green), however, a subset with high expression of HER2 but a relatively lower spot area was also present.\n\nThe scoring of cells is also dependent on the tumor region selected for analysis (Figure [6](#Fig6){ref-type=\"fig\"}D). Sample 6361 was considered to be clinically HER2+, however a ductal carcinoma *in situ* rather than an invasive component of the tumor was selected for this analysis. As a result, weak HER2 and ER intensities were attained, as shown in the topology map, compared to sample 6370.\n\nConclusions {#Sec20}\n===========\n\n**Segmentation of complex tissue components improves with manual correction**There are many challenges with *in situ* analysis of molecular features in tissue sections. Tissues display complex compositions of organic structures such as epithelial elements, vasculature, lymphatic components, nerves and supportive tissue including different types of fibers. There is morphological diversity between cell types and their organization, and in cancer this is even more pronounced. For automatized image analysis, accurate segmentation of cellular components is crucial to avoid misleading estimates of markers. Overlapping or closely spaced nuclei can easily be interpreted as one, and cells with major sectioning artefacts need to be discarded.\n\nMany automatized approaches have been developed to address these issues, and Columbus and CellProfiler are two state of the art softwares used to benchmark the performance of GoIFISH. These have been developed for general applicability to a number of biological scenarios, of which cell-based culture is their main strength. As a result, the application to tissue-sections with the heterogeneous morphology may have resulted in the poorer results observed. In particular, CellProfiler did not perform at a similar level as Columbus and GoIFISH: this may be attributed to the need to optimize parameters in a segmentation pipeline before batch processing, whereas both Columbus and GoIFISH automatically calculates parameters for each image. In addition, GoIFISH provides user-friendly options to manually correct inaccurate segmentations and remove artefacts. As shown in Figures [2](#Fig2){ref-type=\"fig\"}A and Additional file [4](#MOESM4){ref-type=\"media\"}: Figure S1B, this correction step is crucial for improving precision and recall.\n\nInter-user subjectivity is a large issue in the field of pathology and there is potential of introducing user bias in the background selection and segmentation editing steps. To mitigate this, our editing toolbox is designed to minimize the amount of manual cell-outlining required, and guidelines have been included in the user-manual for background selection. We have demonstrated that these measures are effective in minimizing inter-user variation, with similar intensity measurements for both background and cell staining reported by different scorers (Figure [5](#Fig5){ref-type=\"fig\"}).\n\n**Image analysis is dependent on image quality**The quality of the segmentation and marker recognition is highly dependent on the quality of the samples attained. For formalin-fixated paraffin-embedded tissue sections, there are variables including fixation type, fixation duration and tissue processing that differ from patient to patient and between laboratories \\[[@CR26]\\]. For fluorescence analysis in general, some tissue composites induce more autofluorescence than others making \"true\\\" staining difficult to quantify. Tissue sections from patient samples have preprocessing steps which cannot be controlled for at the same level of precision as fixation of cell lines can. This explains the variation in sample-to-sample fluorescence intensities as illustrated in Figure [3](#Fig3){ref-type=\"fig\"}, [5](#Fig5){ref-type=\"fig\"}B, [6](#Fig6){ref-type=\"fig\"}B, despite imaging all sections under similar conditions.\n\nThe images used in this study were of high quality but still exhibited artefacts that confounded segmentation results, which is reflective of the challenges faced in image acquisition and analysis. Low intensity of a spot marker compared to the background was observed in sample 6361, resulting in poor detection using all three methods. Overexposure of a channel will increase the false positive rate, as seen in image 7619, and the presence of background artefacts in the DAPI channel will affect segmentation accuracy, as seen in sample 7350. Background adjustment, manual editing and the application of classifiers are strategies GoIFISH uses to address these issues. GoIFISH allows users to select background regions to ensure baseline intensities are comparable across samples, and performs per nucleus background adjustment to remove local variations in auto-fluorescence. This is necessary for comparability of cells across samples. To assist in accurate segmentation, a morphological classifier was applied to centromere 17 detection to remove confounding effects. In addition, manual user input in GoIFISH rectified most difficulties encountered during segmentation.\n\n**Contribution of the software on measuring intra-tumor heterogeneity**Accurate segmentation on the nuclear, membrane and spot level are essential for the extraction of biologically meaningful features from cells. GoIFISH has demonstrated comparable segmentation to CellProfiler and Columbus in nuclear segmentation, and has outperformed them in membrane and spot detection. GoIFISH is capable of segmenting membranes when weakly positive or incomplete, allowing for subsequent objective analysis of intensity-based features.\n\nTo address the complexity of tissue composition and its impact on prognosis \\[[@CR27]\\] we have also included a cell-type classifier based on morphology and intensity. By marking a few segmented cells, all cells with a similar morphology are identified with high accuracy, particularly if intensity information is also included. To illustrate the importance of quantitative analysis in the correct cellular context, we have included a pre-invasive part of a tumor in our analysis (Sample 6361). As shown in Figure [6](#Fig6){ref-type=\"fig\"}, the clinically reported HER2 positivity was not detected. These cells are luminal epithelial and myoepithelial cells, rather than invasive neoplastic cells. In downstream analyses, the added categorical knowledge will ensure these would not be directly compared to invasive cells. The extraction of features from a sample can be multi-dimensional, making visualization of heterogeneity a difficult task. We have included simple topology maps that overlap two stains of interest, allowing visualization of both heterogeneity across cells and their spatial relationships.\n\n**Summary**GoIFISH has been developed to segment high magnification images with combined genomic and phenotypic traits, combining the analysis of nuclei, membranes and spots into a single easy-to-operate system. Thus, GoIFISH allows the objective quantification of the morphological, genomic and phenotypic heterogeneity often observed in tumor IFISH images. Application of quantitative approaches like GoIFISH on large sample collections will lead to profound insights into the impact tumor heterogeneity has on disease progression, and may uncover evolutionary pathways explaining the development of resistance.\n\nMaterial and methods {#Sec25}\n====================\n\nA sample set of HER2 positive breast cancers {#Sec26}\n--------------------------------------------\n\nThis study was conducted in compliance with the Declaration of Helsinki and was approved by the regional ethics committee (REK S-06495b).\n\nHuman tissue samples were collected following protocols approved by the institutional review board of Oslo University Hospital Radiumhospitalet (IRB 2006-53). We used 10 formalin-fixed paraffin-embedded (FFPE) primary tumors from HER2+ breast cancer patients. In this work, we performed IFISH by combining the immunodetection of HER2 protein (expressed in the cell membrane) and Estrogen Receptor *\u03b1* (ER *\u03b1*, located in the nuclei) with the detection of *HER2* and centromere 17 (cent17) copy number, following a protocol previously described \\[[@CR10]\\]. FFPE samples were dewaxed and hydrated in series of ethanol. Heat-induced antigen retrieval was performed in citrate buffer (pH 6) followed by pepsin digestion. After the immunostaining of HER2 and ER at room temperature in a humidifier, tissue slides were hybridized with *HER2* and centromere 17 probes at 37C\u00b0. overnight. Post wash was carried out in SSC (saline-sodium citrate) buffers with different stringency, before air drying and mounting media with DAPI was added. Image acquisition was carried out in an epifluorescence microscope. One randomly selected area per tumor was photographed in a Zeiss Axioplan 2 microscope equipped with an Axio Cam MRM CCD camera and Axio Vision software. The experimental methods are explained in greater detail in the Additional file [3](#MOESM3){ref-type=\"media\"}.\n\nAnalysis pipelines for CellProfiler, Columbus and**GoIFISH** {#Sec27}\n------------------------------------------------------------\n\nColumbus provides 4 nuclear, 4 cytoplasmic and 3 spot detection methods. These were first tested visually to determine the best candidate methods, which were then quantitatively compared with CellProfiler and GoIFISH in terms of precision and recall. The best results from each image were then used for direct comparison (see Additional file [3](#MOESM3){ref-type=\"media\"}).\n\nA CellProfiler analysis pipeline was constructed with the following parameters: Nuclei Segmentation was performed using two class Otsu Global Thresholding, and diameter of objects restricted to 20-120 pixels. Clumped cells were separated using the propagation method. Membrane detection was performed based on the propagation method, using the combination of the distance to the nuclei and intensity gradient to select the membrane. The spot signals were enhanced and masked to nuclear regions. Spots were detected using 'RobustBackgroundPerObject', limited to a diameter between 5 and 40 pixels, with clumped objects separated based on intensity.\n\nThe default GoIFISH pipeline performed shape optimised nuclear segmentation with intensity suppression between 10 to 30% and fragments of size less than 500 pixels discarded. The output nuclear map was applied to HER2 membrane detection and spot detection. Detected Centromere 17 spots were run through a morphological classifier to minimise effects of autofluorescence.\n\nMetrics for performance evaluation {#Sec28}\n----------------------------------\n\nWe compared computational approaches to a manually segmented 'gold standard'. For each of the 10 images, nuclei, membranes and spots were outlined manually in the maximum projected image. Spots were counted through 15-21 z-stacks. In total, 355 individual cells were scored for membrane completeness, nuclear positivity and copy number. We then benchmarked the computational outputs of GoIFISH with the gold standard using the following panel of quality criteria. We define *N*~*t*~ as the number of correctly segmented cells, *N*~*under*~ and *N*~*over*~ as the number of under and over segmented cells, *N*~*FP*~ as the number of false positives which do not appear in the gold standard image, and *N*~*FN*~ as the number of false negatives (See Additional file [4](#MOESM4){ref-type=\"media\"}: Figure S1A). Precision is defined as *P*=*N*~*t*~/(*N*~*t*~+*N*~*FP*~+*N*~*over*~)Recall: *R*=*N*~*t*~/(*N*~*t*~+*N*~*FN*~+*N*~*under*~)the F-Score is the harmonic mean of precision and recall: *F*=2*P*\u00b7*R*/(*P*+*R*), which is a measure of how well a cell can be detected, and whether the number of cells present closely resemble the true value.\n\nMitigating oversegmentation {#Sec29}\n---------------------------\n\nWe have developed and implemented a technique to mitigate oversegmentation. For a fragment *f* classified as oversegmented, we list all neighboring fragments **g**={*g*~1~,...,*g*~*N*~} within a radius of *r* pixels. Each union *c*^*j*^=*f*\u222a**g**^*j*^ of *f* with a subset of neighboring fragments **g**^*j*^\u2286**g** is a potential extension of *f* into a full cell. The index *j* ranges over all members of the powerset of **g**, i.e., the set of all its subsets. To select the best extension *c*^\u2217^, we first score all possible extensions *c*^*j*^ by a function *S*~*F*~(*c*^*j*^) that compares a feature *F* of the combined fragments to the average of that feature in the individual fragments: $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}\n $$ S_{F}\\left(c^{j}\\right) = F\\left(c^{j}\\right) - \\frac{1}{n_{j}} \\sum\\limits_{i=1}^{n_{j}} F\\left({c^{j}_{i}}\\right), \\; \\text{where } n_{j} = |c^{j}|,  $$ \n\\end{document}$$\n\nand then sum all feature-wise scores into a final score: $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}\n$S\\left (c^{j}\\right) = \\sum S_{F}\\left (c^{j}\\right)$\n\\end{document}$, which we use to select the optimal extension *c*^\u2217^ as $$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}\n $$ c^{*} = \\mathop{\\text{argmax}}\\limits_{j} \\ S\\left(c^{j}\\right).  $$ \n\\end{document}$$\n\nCurrently we use two morphological features for scoring fragments: Solidity and deviation from theoretical area. If **g**=*\u2205* or if there is no positive maximum of the score, then *c*^\u2217^=*f*.\n\nClassifiers {#Sec30}\n-----------\n\nClassifiers have been implemented in a number of segmentation steps, including nuclei detection and spot detection, to minimise errors. GoIFISH only uses linear classifiers to reduce overfitting to data.\n\n**Nuclear detection**Nuclei detection performs linear discriminant analysis based on a training set of 153 fragments from 5 images attained from nuclear segmentation performed at different depths. A total of 10 morphological properties including solidity, area, perimeter, axis lengths, axis ratios, circularity, area-perimeter ratio, and deviation from theoretical area and perimeter were measured. Each fragment was scored for oversegmentation, undersegmentation or optimal shape.\n\n**Spot detection**A spot classifier was constructed using segmentation output from 2 training images containing 67 spot candidates and placed into a linear discriminant analysis. Features extracted for the classifier include solidity, area, perimeter, axis ratios, circularity, area-perimeter ratio deviation from theoretical area and perimeter, mean intensity and minimum intensity. Three classes were assigned to each 'spot': optimal, too small or too large.\n\n**Cell classification**Classification of cells after segmentation is performed using a one vs all Support Vector Machine with linear kernel. Training data is generated on the spot from information supplied from the user. Features can either be morphological (area, perimeter, solidity, axis lengths and eccentricity) or contain extra information from the other channels, such as spot area or intensity.\n\nAdditional files {#Sec34}\n================\n\nAdditional file 1**GoIFISH** **software.** Source code for use directly in MATLAB. Binary files for use outside of MATLAB (For Mac and Windows OS) and the required MATLAB Compiler Runtime, are available for download at www.sourceforge.net/projects/goifish/ due to size constraints.\n\nAdditional file 2**GoIFISH** **user manual.** A user guide for the operation of GoIFISH.\n\nAdditional file 3**Supplementary information.** Supplementary information includes detailed experimental methods for the staining of images, and all code for the generation of plots and statistical analyses in R.\n\nAdditional file 4**Figure S1.** Precision and recall in cell segmentation. (A) Illustration of the metrics used for Precision-Recall Testing (B) Precision Recall Plots for nuclear segmentation, HER membrane detection, centromere 17 spot detection and *HER2* cluster detection.\n\nAdditional file 5**Figure S2.** Correlation between intensity and pathologist scoring. (A) ER staining intensity using GoIFISH raw, background adjusted and manually edited nuclear adjusted intensities compared to semi-quantitative pathologist scored ER intensity for each individual image. (B) Combined distribution of ER intensities across all samples using the methods described in (A) (C) Distribution of GoIFISH raw, GoIFISH manually edited mean HER2 intensities and GoIFISH manually edited coefficient of variation compared to pathologist scored membrane completeness for each individual image. (D) Distribution of HER2 intensities or coefficient of variation across all samples using the methods described in (C).\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n\nHGR and FM planned and outlined the study. AT developed and implemented the software. IHR, VA, HGR provided data, scored images manually, and tested the software. \u00c5H collected samples. AT, IHR, VA, HGR and FM wrote the paper. AT and IHR contributed equally and are joint first authors. HGR and FM are co-last and co-corresponding authors. All authors read and approved the final manuscript.\n\nWe would like to thank Prof. Kornelia Polyak and Prof. Anne-Lise B\u00f8rresen-Dale for support and insightful discussions. AT and FM would like to acknowledge the support of The University of Cambridge, Cancer Research UK and Hutchison Whampoa Limited. IHR and HGR would like to acknowledge support from Helse S\u00f8r-\u00d8st, The Norwegian Cancer Association, the K. G. Jebsen Foundation and Radiumhospitalets legater. VA would like to thank the Breast Cancer Research Foundation and the National Cancer Institute (grant U54CA143798).\n"}
+{"text": "Over the past decade, Web\u00a02.0 applications, defined as internet-based tools that allow users to generate and manage content, have demonstrated the potential to revolutionize the search of up-to-date health information by users and the respective dissemination by health professionals \\[[@CR1]\\]. Considered the 'social Web', this updated web-based platform facilitates personable interactions between individuals, whether sharing educational or entertainment information (e.\u2009g. YouTube, Instagram, blogs, podcasts) or networking (e.\u2009g. Facebook, Twitter, Google Plus) \\[[@CR2]\\]. Traditional health promotion tools, such as television or radio broadcasts, newspapers or pamphlet distribution, used in mass media campaigns have demonstrated positive influences on health behaviours \\[[@CR3]\\]. By incorporating these technological advances of the 'social Web' into traditional health promotion tools, health teams can develop local or national primary, secondary or tertiary prevention campaigns that improve dissemination of health information to available users. This potential benefit of the use of social media technology in campaigns has been demonstrated in a\u00a0review of case studies \\[[@CR4]\\].\n\n'Social mobilizers', on the other hand, are key to implementing local or national health initiatives in the field and building rapport with community members \\[[@CR5]\\]. Although no standard definition of community mobilization exists \\[[@CR6]\\], this notion refers to facilitating change through educating community members to increase their knowledge and skill sets, empowering active participation in programmed activities, and forming partner collaborations at local or national levels \\[[@CR7], [@CR8]\\]. When highly trained in project objectives and activities, these community health workers, who understand the local cultural beliefs or customs, can reach marginalized communities and teach accurate health information \\[[@CR6]\\]. When joined with interdisciplinary health teams, they can tailor the health message to be culturally appropriate, increase interpersonal communication that empowers community members, and promote optimal health outcomes \\[[@CR9]\\]. It is anticipated that community members will use recommended primary prevention methods, such as understanding risk factors or immunizations, which mitigate disease transmission. This was observed during the Ebola epidemic in Western Africa in 2014, where 'social mobilizers' visited homes, interacting with individuals and families to raise community awareness and reduce stigma associated with the transmission of the Ebola virus \\[[@CR10]\\].\n\nAlthough 'social mobilizers' play a\u00a0significant role in the field, there is limited research that describes the role of medical students when using the dual strategies of community mobilization and social media. Medical students, from their health professions education, understand the influence of the social determinants of health as well as the basic anatomy, physiology and epidemiology of disease pathology and transmission. They are creative and highly skilled in the use of social media technology, such as Facebook, Google Plus, Instagram, Twitter, or YouTube, to develop catchy slogans, hashtags, or taglines that are popular and tailored to affected communities. Thus, they may be trained and serve as essential team members in the 'roll out' of health educational outreach programmes during disaster relief efforts.\n\nWe describe how African medical students in Sierra Leone and Guinea applied dual community mobilization and social media approaches during the Ebola epidemic. We also provide reasons that support the inclusion of medical students, due to their training in basic and clinical sciences and skills in Web\u00a02.0 applications, on intra- or multi-disciplinary health teams during relief programmes.\n\nThe case of the Ebola virus {#Sec1}\n===========================\n\nThe emergence of Ebola virus disease in Guinea in March 2014, followed by Liberia and Sierra Leone \\[[@CR11]\\], provided optimal conditions for disease transmission. This resulted from the three neighbouring countries having fluid borders; high human travel between rural, peri-urban, and urban regions or countries; and collapse of environmental and health systems during post-civil war rebuilding \\[[@CR12]\\]. Initially, the rapid communicability and high mortality rates of Ebola virus disease presented multiple challenges to clinicians, health facilities, and national health sectors. Suboptimal isolation facilities for Ebola virus disease were widespread. Health facilities were understaffed, with less than an estimated 0.1\u00a0physicians per 1,000 population in these three respective countries \\[[@CR13]\\]. These burdened health care workers encountered complex challenges in managing the increased number of suspected patients with Ebola virus disease while diagnosing those with other infectious diseases of similar clinical presentation, such as cholera, dengue, malaria, or typhoid fever.\n\nInternational public health authorities, such as the World Health Organization (WHO), distributed posters and pamphlets \\[[@CR14]\\], and local Ministries of Health and Sanitation conducted radio and television broadcasts to increase community awareness about Ebola virus disease \\[[@CR15], [@CR16]\\]. Widespread public media messages emphasized adherence to standard protocols, basic hygiene policies, and quarantine measures or *cordons sanitaires*, in addition to eliminating specific cultural practices that facilitated physical contact with infected individuals. Understanding the cultural role of oral communication in Liberia, for example, 'social mobilizers' entered local communities to channel accurate health information about Ebola virus disease \\[[@CR17]\\]. It is essential that health interventions, such as infection control practices, are precisely aligned with cultural practices to maximize community acceptance \\[[@CR18]\\]. In the presence of this 'intervention--culture' disconnect, some authorities observed cultural backlash from community members with escalating levels of distrust and fear of relief responses to Ebola virus disease, including stigma associated with strict isolation procedures. Thus, lessons learned were the key intersections among the identification of cultural beliefs, appraisal of existing health systems, use of evidence-based practices by interdisciplinary teams, and household- and community-level participation \\[[@CR19], [@CR20]\\]. Understanding the intertwining roles of these constructs may serve as a\u00a0foundation to establish and strengthen cultural-based, standardized protocols for infection control during relief programmes, such as for Ebola virus disease.\n\nUtilized health educational strategies in a\u00a0case study {#Sec2}\n======================================================\n\nDue to the temporary closure of educational institutions in Sierra Leone and Guinea, including one medical school in Sierra Leone and three medical schools in Guinea, medical students of all academic years, forming part of the International Federation of Medical Students' Associations (IFMSA), initiated the Kick Ebola Out campaign in August\u00a02014 \\[[@CR21]\\]. Since medical students learn the basic and clinical sciences with community and clinical rotations with patients in health institutions, their knowledge and skill sets are more advanced than other community health workers. Through didactic sessions, they were trained by specialists of the respective Ministries of Health and Sanitation on updated knowledge about Ebola virus disease and how to dispel common myths or misinformation. Representing multiple ethnicities, they were also instructed on appropriate and culturally sensitive communication strategies to use in community campaigns, including disseminating health messages through social media. By using dual health promotional strategies of community mobilization and social media, medical students could address two main objectives of the campaign (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The first objective was to increase overall knowledge of medical students and community members about Ebola virus disease transmission, clinical presentation and preventive measures. The second objective was to reduce community perceptions related to Ebola-associated stigma or fear of disease transmission for susceptible individuals or recent survivors.Fig. 1Dual health promotional strategies of community mobilization and social media used by medical students during the Ebola virus disease outbreak in Sierra Leone and Guinea in 2014\n\nUsing community mobilization strategies, groups of medical students, who were led by at least one advanced medical student, conducted home visits to numerous communities, distributed chlorine and soap, and provided accurate health information about Ebola virus disease. By visiting media outlets, such as television and radio stations, they educated audiences about Ebola virus disease and dispelled myths without provoking alarm or stigma to transmission of the Ebola virus. They discouraged handshaking as well as certain cultural practices, such as washing or touching the bodies of patients infected with Ebola virus disease. Still, failure to abide to the established quarantine protocol due to fear or anxiety, lack of trust in orthodox medical practice, low literacy levels, and extreme poverty in communities challenged Ebola containment measures \\[[@CR22]\\].\n\nThe advantage of integrating social media, such as Facebook and Twitter, in the Ebola virus educational campaigns was twofold. First, these Web\u00a02.0 applications facilitated rapid communication among medical students of multiple African countries and other geographic regions. Professional networks were solidified, accurate Ebola virus disease information was identified, and health promotional tools (\\@KickEbolaOut) and an educational mobile telephone app (KickEbolaOut) were created and shared. Second, by using this virtual interface to disseminate key health facts and graphics on Ebola virus disease, community members could read up-to-date health messages and, in turn, educate family members. For example, Google and Twitter movements increased when newsworthy findings were promoted on media outlets, such as when new cases of Ebola virus disease were identified in the United States \\[[@CR23]\\].\n\nOne limitation of this campaign was the possible distrust or fear of community members, related to the stigma associated with Ebola virus disease, which influenced their ability to participate in educational activities during home visits. Medical students reported that due to their identity as health professional students, they were positively received and welcomed by community members. A second limitation was misinformation that spread quickly through social media, because it was not possible to validate or verify the quality of millions of Ebola tweets or posts \\[[@CR24]\\]. To combat this challenge, African medical students spearheaded the development of Twitter handles, Facebook pages and posts, and hashtags, which enabled prompt dissemination of accurate messages from reputable health authorities, including WHO (\\@WHO), Centers for Disease Control and Prevention (\\@CDCgov), and M\u00e9decins Sans Fronti\u00e8res (\\@MSF). Because social media technology spans across all age groups and socioeconomic strata, this veritable tool in health promotion can significantly aid health teams in disseminating accurate information among health care providers and community members during outbreak relief efforts.\n\nConclusions {#Sec3}\n===========\n\nAs described in this report, medical students can serve as human resources for health through their integral role in disseminating health information and achieving health programme objectives through the use of community mobilization and social media strategies. Due to their knowledge and skills in medicine, public health and Web\u00a02.0 applications, they can critically evaluate evidence-based health messages for accuracy, utilize cultural-sensitive communication strategies to relay health messages and develop creative messages that can be publicized on radio and television broadcasts and social media for communities. The power and structure of the IFMSA national member organizations worldwide provide an effective pathway and approach to facilitate Ministry of Health and Sanitation training of medical students on key clinical and communication skills. By using these dual health promotion strategies, medical students can join intra- or multi-disciplinary health teams and establish innovative health programmes for primary, secondary or tertiary health initiatives at local or national levels.\n\nThe role of medical students should be recognized as an indispensable health educator in community-based health promotion and education programmes. Due to their vibrancy, passion, and expertise in the use of social media strategies, they serve as key team members to facilitate broad diffusion of health messages that keep community members educated and empowered for their optimal health decisions. Although academically prepared with their health professions education in the classroom, they may have limited opportunities to apply this acquired knowledge and skill set in the community setting. Therefore, the future integration of early practical community experiences utilizing social media strategies throughout the health professions education may strengthen academic training of medical students in public health, as well as provide them with a\u00a0specialized toolkit to use when leading local or national health initiatives, such as the relief programmes during the Ebola outbreak.\n\nAcknowledgements {#FPar104}\n================\n\nWe admire the dedicated Kick Ebola Out collaborative campaign efforts by IFMSA members of the Sierra Leone Medical Students' Association (SLeMSA), Guinean Medical Students' Association (AEM) (Hadiatou Diallo, President), and other participating nations.\n\nFunding {#FPar5}\n=======\n\nNone.\n\nConflict of interest {#FPar6}\n====================\n\nH.J.\u00a0Chapman, A.E.\u00a0Tade, V.J. Animasahun and A.\u00a0Naveed state that there are no conflicts of interest.\n"}
+{"text": "Background {#Sec1}\n==========\n\nIt is indisputable that an elevated blood pressure (BP) is one of the most important risk factors for cardiovascular disease, heart failure, strokes, chronic kidney disease, peripheral arterial disease, and atrial fibrillation \\[[@CR1]\\]. Many factors including the genetics, life style, medications, obesity, and dietary factors contribute to the development of hypertension in a complicated way.\n\nEpidemiological studies have revealed that excessive salt intake plays a major role in elevating the BP in the global population, leading to increased cardio- and cerebrovascular morbidity and mortality \\[[@CR2]--[@CR4]\\]. Also, many studies have demonstrated that salt itself has an impact on the aortic stiffness and coronary heart disease in a BP-independent manner \\[[@CR3], [@CR5]\\]. The global mean sodium intake in 2010 was about 4\u00a0g/d (salt \u2248 10\u00a0g/d) in adults, and is especially higher in Asian regions \\[[@CR4]\\]. In that year, the Korean population aged 20 and over had a high estimated sodium intake of 5.21 (4.98--5.48) g/d. WHO (World Health Organization) recommends a sodium reduction of \\<2\u00a0g/d in adults \\[[@CR6]\\] and also the current hypertension guidelines recommend reducing the sodium intake as a way of a lifestyle modification \\[[@CR7]\\].\n\nIf a high salt intake of the population has such a big impact on the global health, it is of importance to figure out what factors contribute to that high salt intake, which obviously is much more than needed. We were interested in the salt-taste sensitivity threshold, which meant how sensitively an individual perceives a salt-taste. When an individual has a high threshold for a salt-taste, it might be possible to expect that he or she would intake more salt than one who had a low threshold. There is little data about the relationship among the salt-taste sensitivity, amount of salt intake, and prevalence or incidence of hypertension. Even the limited data available has had inconsistent results \\[[@CR8]--[@CR10]\\]. Therefore, we tried to figure out if there were any differences in the salt-taste sensitivity threshold and amount of salt intake, estimated by the urinary sodium excretion, between normotensive and hypertensive groups.\n\nMethods {#Sec2}\n=======\n\nStudy population {#Sec3}\n----------------\n\nWe assessed 289 patients, aged 18\u00a0years or older, who visited our clinic for high BP from March 2013 to August 2016. The patients with diabetes mellitus (*n*\u00a0=\u00a08), pregnancy (*n*\u00a0=\u00a021), an impaired renal function (serum creatinine \\>1.5\u00a0mg/dL), and secondary hypertension such as hyperaldosteronism, Cushing's syndrome, renal artery stenosis and thyroid dysfunction were excluded. Also, the patients whose ABPM (*n*\u00a0=\u00a012) or taste threshold (*n*\u00a0=\u00a047) data were missing, and who did not stop their medication (*n*\u00a0=\u00a02) were excluded. Finally, a total of 199 patients were included in the analysis. We checked thyroid hormones and plasma renin activity, serum aldosterone levels to exclude secondary hypertension. They were divided into control (normotension) and hypertension groups according to their ABPM results. They did not take any hypertensive medications. If they started to take hypertensive medications recently, they were included after cessation of the drug at least for 1 week. We compared the serum level of the sodium (Na), potassium (K), renin activity, aldosterone, blood urea nitrogen (BUN), and creatinine (Cr) between the two groups. This study was approved by the institutional review board of Daegu Catholic Medical Center.\n\nThe definition of hypertension {#Sec4}\n------------------------------\n\nHypertension was diagnosed by 24-h ABPM (TM-2430, A&D company, Tokyo, Japan), that is, an average daytime systolic BP of \u2265135\u00a0mmHg or diastolic BP of \u226585\u00a0mmHg. The patients were checked for nocturnal BP dipping patterns and were divided into 4 groups: reverse-dipper (rise in the nocturnal BP), non-dipper (\\<10%), dipper (10\\~20%), and extreme-dipper (\\>20%). The morning BP surge was calculated as the morning BP (after awakening, average of 3 consecutive BP readings) minus the nocturnal BP (average 3 consecutive BP readings centered on the lowest nocturnal BP).\n\nThe measurement of the salt-taste threshold {#Sec5}\n-------------------------------------------\n\nThe salt-taste threshold was assessed, using graded solutions of saline as shown in Table\u00a0[1](#Tab1){ref-type=\"table\"}. We employed a whole-mouth gustatory test procedure to test the solutions. Each level of solution was applied to the tongue evenly (1\u00a0ml of the solution with a 5\u00a0ml syringe). Then, the subject held the solution for a few seconds and swallowed it. The subject was then asked about the taste of the solutions. Starting from the lowest concentration, the detection threshold was defined as the point at which the subject perceived a different taste from that of the distilled water. The recognition threshold was defined as the lowest concentration at which the subject recognized the solution as being salty. We went back to a lower concentration and repeated the test again if the threshold could not be determined. The subjects rinsed their mouth with water whenever the subject tested a different concentration of the solution. We performed the same test once again another day in patients with an unusual test result (*n*\u00a0=\u00a014).Table 1The concentration of the salt-taste solutionConcentration levels of the taste solutionSodium chloride (g/mL)10.0048820.0097730.0195340.0390650.0781360.1562570.3125080.6250091.25000102.50000115.000001210.000001320.00000\n\nThe 24-h urinary sodium and potassium excretion {#Sec6}\n-----------------------------------------------\n\nWe collected 24-h urine samples and measured the sodium excretion in it to estimate the amount of sodium intake. The patients were instructed to collect self-urine samples for 24\u00a0h after discarding the first morning urine. The urine samples were kept in a refrigerator at home and brought to the laboratory.\n\nTransthoracic echocardiography {#Sec7}\n------------------------------\n\nThe left ventricular (LV) chamber diameters were acquired in the parasternal long-axis view and left atrial (LA) chamber diameters in the parasternal short-axis view. The LV mass was calculated using a 2D--guided M-mode method where the following parameters were included: interventricular septum, posterior wall thickness, and LV internal diameter at end-diastole. The LA volume was calculated using a disk summation technique. The ejection fraction (EF) was measured using the biplane method of disks and M-mode of the LV in the parasternal long-axis view. For the analysis, we chose one method, the M-mode method for convenience. The LV mass index (LVMI) was calculated by dividing the LV mass by the body surface area (BSA) and the LA volume index (LAVI) by dividing the LA volume by the BSA.\n\nStatistical analysis {#Sec8}\n--------------------\n\nThe differences in the baseline characteristics of the patients in each group were compared with the use of a Pearson's chi-square test for categorical variables. A Student's *t*-test was used for a comparison of the continuous variables. To assess the relationship among the continuous variables, correlation and regression analyses were used. Spearman correlation method was used to assess the relationship between ordinal variable and continuous variable. The one-way analysis of variance (ANOVA) test was used to compare the data among the taste threshold groups and BP groups. The Bonferroni's method was used as a post hoc analysis. Multiple regression analysis was used to assess the relationship between the LVMI and the salt taste threshold. The values are expressed as the mean\u00a0\u00b1\u00a0standard deviation. *P* values of \\<0.05 were considered significant. All analyses were conducted using IBM SPSS statistics software (version 19.0, Chicago, IL, USA).\n\nResults {#Sec9}\n=======\n\nPatient characteristics {#Sec10}\n-----------------------\n\nA total of 199 patients were included in our study and their characteristics are shown in Table\u00a0[2](#Tab2){ref-type=\"table\"}. The proportion of males was significantly higher in the hypertensive group than control group. The age, BMI, and BSA were similar between the two groups. The serum sodium and potassium levels were similar between the two groups. Also, the other laboratory test results, including the BUN, Cr, plasma renin activity and aldosterone levels, were similar between the two groups.Table 2The patient characteristicsVariablesControlHTN*p*Patient number66133Male, n (%)22 (33.3%)71 (53.4%)0.010Age, years53.6\u00a0\u00b1\u00a012.551.7\u00a0\u00b1\u00a010.90.273Height, cm162.4\u00a0\u00b1\u00a09.5165.1\u00a0\u00b1\u00a08.90.054Weight, kg66.0\u00a0\u00b1\u00a012.968.8\u00a0\u00b1\u00a011.50.115BMI, kg/m^2^24.8\u00a0\u00b1\u00a03.225.2\u00a0\u00b1\u00a03.20.492BSA, m^2^1.72\u00a0\u00b1\u00a00.211.77\u00a0\u00b1\u00a00.180.073Serum sodium, mEq/L140.0\u00a0\u00b1\u00a02.4139.7\u00a0\u00b1\u00a02.60.390Serum potassium, mEq/L4.47\u00a0\u00b1\u00a00.374.36\u00a0\u00b1\u00a00.360.057Blood urea nitrogen, mg/dL13.2\u00a0\u00b1\u00a03.813.1\u00a0\u00b1\u00a03.50.858Serum creatinine mg/dL0.74\u00a0\u00b1\u00a00.150.78\u00a0\u00b1\u00a00.160.150Urinary sodium, mEq/d117.9\u00a0\u00b1\u00a057.2140.9\u00a0\u00b1\u00a059.80.011Urinary potassium, mEq/d52.2\u00a0\u00b1\u00a020.850.4\u00a0\u00b1\u00a024.20.638Urinary sodium/potassium2.46\u00a0\u00b1\u00a01.073.03\u00a0\u00b1\u00a01.180.003Plasma renin activity, ng/ml/h1.85\u00a0\u00b1\u00a01.531.83\u00a0\u00b1\u00a01.410.936Serum aldosterone, pg/ml89.6\u00a0\u00b1\u00a068.979.4\u00a0\u00b1\u00a055.10.278*HTN* hypertension, *BMI* body mass index, *BSA* body surface area\n\nTransthoracic echocardiography {#Sec11}\n------------------------------\n\nThe summaries of the echocardiographic data are shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}. The LV systolic function was normal or nearly normal in the entire study population. The LVMI correlated with the daytime systolic/diastolic BP (*p*\u00a0=\u00a00.001, *r*\u00a0=\u00a00.244 and *p*\u00a0=\u00a00.011, *r*\u00a0=\u00a00.183, respectively). The average LVMI in the hypertensive group was significantly higher than that in the control group. When the LVMI was analyzed separately by the gender, it did not significantly differ between the control and hypertension groups in males (128.8\u00a0\u00b1\u00a020.4 vs. 133.5\u00a0\u00b1\u00a025.6, respectively, *p*\u00a0=\u00a00.435) as well as females (114.6\u00a0\u00b1\u00a028.2 vs. 121.0\u00a0\u00b1\u00a023.5, respectively, *p*\u00a0=\u00a00.21). There was no correlation between the LVMI and the recognition threshold of salt taste, adjusted by systolic BP (partial correlation coefficient\u00a0=\u00a00.117, *p*\u00a0=\u00a00.103). When the subjects were divided into three groups according to the recognition threshold of salt, the LVMI differed significantly between groups (Table\u00a0[3](#Tab3){ref-type=\"table\"}). In post hoc analysis, there was significant difference in the LVMI between the subjects with threshold of 7 and subjects with threshold 8 or more (*p*\u00a0=\u00a00.030). But there was marginal significance between the subjects with threshold of 7 and the subjects with 6 or less (*p*\u00a0=\u00a00.071). The other echocardiographic parameters, including the LV dimension, LA dimension, and LAVI did not differ between the control and hypertension groups.Table 3The echocardiographic measurementsVariablesControlHTN*p*Recognition threshold of salt taste*p*\\<7 (*n*\u00a0=\u00a095)7 (*n*\u00a0=\u00a071)\\>7 (*n*\u00a0=\u00a033)LVEF, %64.5\u00a0\u00b1\u00a06.063.6\u00a0\u00b1\u00a05.70.29663.9\u00a0\u00b1\u00a06.164.4\u00a0\u00b1\u00a05.363.0\u00a0\u00b1\u00a06.00.542LVEDD, mm49.4\u00a0\u00b1\u00a04.750.5\u00a0\u00b1\u00a05.10.12849.9\u00a0\u00b1\u00a05.150.1\u00a0\u00b1\u00a04.750.8\u00a0\u00b1\u00a05.20.681LVESD, mm31.8\u00a0\u00b1\u00a04.232.8\u00a0\u00b1\u00a04.20.11032.3\u00a0\u00b1\u00a04.432.3\u00a0\u00b1\u00a03.933.1\u00a0\u00b1\u00a04.20.577LAD, mm36.7\u00a0\u00b1\u00a04.637.9\u00a0\u00b1\u00a04.60.07337.0\u00a0\u00b1\u00a04.337.7\u00a0\u00b1\u00a04.938.3\u00a0\u00b1\u00a05.00.353LVMI, g/m^2^119.4\u00a0\u00b1\u00a026.5127.6\u00a0\u00b1\u00a025.30.036123.6\u00a0\u00b1\u00a024.8121.4\u00a0\u00b1\u00a021.9135.4\u00a0\u00b1\u00a034.10.030LAVI, ml/m^2^27.0\u00a0\u00b1\u00a07.728.1\u00a0\u00b1\u00a08.10.38227.4\u00a0\u00b1\u00a07.628.1\u00a0\u00b1\u00a07.927.9\u00a0\u00b1\u00a08.90.843*LVEF* left ventricular ejection fraction, *LVEDD/LVESD* left ventricular end diastolic/systolic dimension, *LAD* left atrial dimension, *LVMI* left ventricular mass index, *LAVI* left atrial volume index\n\nThe 24-h ambulatory blood pressure monitoring {#Sec12}\n---------------------------------------------\n\nAn average daytime BP in the hypertension group was not so high (Table\u00a0[4](#Tab4){ref-type=\"table\"}). Only 31 patients had an average systolic daytime BP of \u2265155\u00a0mmHg or higher. The BP difference between the two groups was about 20\u00a0mmHg for the systolic pressure. The degree of nocturnal dipping was a little higher in the hypertension group than the control group, but there was no statistical significance. The non-dipper group contributed to 50.0% of the control group and 46.6% of the hypertension group. Nocturnal dipping was not related to the status of hypertension (*p*\u00a0=\u00a00.459). Further, the degree of morning BP surge was similar between the two groups. When we divided the subjects into three groups according to the recognition threshold of salt taste, the ABPM data did not differ among groups (Table\u00a0[4](#Tab4){ref-type=\"table\"}).Table 4The 24-h ambulatory blood pressure monitoring dataVariablesControlHTN*p*Recognition threshold of salt taste*p*\\<7 (*n*\u00a0=\u00a095)7 (*n*\u00a0=\u00a071)\\>7 (*n*\u00a0=\u00a033)Mean SBP, mmHg125.0\u00a0\u00b1\u00a06.7145.1\u00a0\u00b1\u00a012.1\\< 0.001138.8\u00a0\u00b1\u00a013.9136.5\u00a0\u00b1\u00a013.6141.7\u00a0\u00b1\u00a016.10.205Mean DBP, mmHg78.8\u00a0\u00b1\u00a05.792.5\u00a0\u00b1\u00a09.5\\< 0.00188.0\u00a0\u00b1\u00a010.886.7\u00a0\u00b1\u00a09.790.7\u00a0\u00b1\u00a011.80.194SBP at day, mmHg127.2\u00a0\u00b1\u00a06.5148.1\u00a0\u00b1\u00a011.2\\< 0.001141.7\u00a0\u00b1\u00a013.7139.4\u00a0\u00b1\u00a013.5143.4\u00a0\u00b1\u00a015.60.339DBP at day, mmHg79.8\u00a0\u00b1\u00a04.494.8\u00a0\u00b1\u00a09.1\\< 0.00190.0\u00a0\u00b1\u00a010.988.6\u00a0\u00b1\u00a09.691.7\u00a0\u00b1\u00a011.60.367SBP at night, mmHg117.9\u00a0\u00b1\u00a010.3134.8\u00a0\u00b1\u00a016.50.001129.7\u00a0\u00b1\u00a017.7127.1\u00a0\u00b1\u00a015.9132.2\u00a0\u00b1\u00a015.50.331DBP at night, mmHg73.1\u00a0\u00b1\u00a07.485.1\u00a0\u00b1\u00a012.0\\< 0.00181.7\u00a0\u00b1\u00a012.779.1\u00a0\u00b1\u00a011.483.7\u00a0\u00b1\u00a011.10.154MBPS, mmHg18.7\u00a0\u00b1\u00a015.619.1\u00a0\u00b1\u00a013.60.48519.5\u00a0\u00b1\u00a012.619.0\u00a0\u00b1\u00a015.617.2\u00a0\u00b1\u00a016.10.721Nocturnal dipping0.1990.527Reverse-& non-dipper, n (%)43 (65.2%)74 (55.6%)56 (58.9%)39 (54.9%)22 (66.7%)Dipper & extreme-dipper, n (%)23 (34.8%)59 (44.4%)39 (41.1%)32 (45.1%)11 (33.3%)*HTN* hypertension, *S/DBP* systolic/diastolic blood pressure, *MBPS* morning BP surge\n\nThe 24-h urinary sodium and potassium excretion {#Sec13}\n-----------------------------------------------\n\nThe 24-h urinary sodium excretion and Na/K ratio were significantly higher in the hypertensive group compared to the normotensive group, but the 24-h urinary potassium excretion did not differ between the two groups (Table\u00a0[2](#Tab2){ref-type=\"table\"}). The 24-h urinary sodium excretion was correlated to the average systolic BP during the daytime and nighttime (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}, p=0.002, *p*\u00a0=\u00a00.018, respectively). The 24-h urinary sodium excretion was higher in the male than female group (155.9\u00b167.3 vs. 113.5\u00b144.0\u00a0mEq/d, respectively, *p* \\< 0.001). When the 24-h urinary sodium excretion was analyzed separately by the gender, it was correlated to the BP only in females and not in males (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}, p\\<0.001, *p*\u00a0=\u00a00.825, respectively). There was no correlation between the degree of nocturnal dipping and the urinary sodium excretion (Spearman's rho=0.076, *p*=0.294). After adjusting the systolic BP, the LVMI was correlated to the 24-h urinary sodium excretion (*p*=0.032, *r*=0.156).Fig. 1The relationship between the systolic blood pressure (BP) and 24-h urinary sodium excretion **a** and comparison by gender (**b**)\n\nThe salt-taste threshold {#Sec14}\n------------------------\n\nThe salt-taste thresholds did not significantly differ between the control and hypertension groups (Table\u00a0[5](#Tab5){ref-type=\"table\"}), and also between genders. More than two-thirds of the study population (69.8%) corresponded to a recognition threshold of 6 and 7 (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). Also, there were no correlations between the detection and recognition thresholds and the daytime systolic BP (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}). The 24-h urinary sodium excretion did not show any correlation to the detection and recognition threshold of salt (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). When we divided the patients into 3 groups according to the degree of the BP status, the recognition threshold of salt did not differ (*p*\u00a0=\u00a00.918 in one-way ANOVA). Also, the recognition threshold of salt did not correlate with the early morning BP surge (Spearman's rho\u00a0=\u00a0\u22120.080, *p*\u00a0=\u00a00.26) and was not associated with dipping patterns; extreme dipper, dipper, non-dipper, and reverse dipper (Spearman's rho\u00a0=\u00a0\u22120.007, *p*\u00a0=\u00a00.925).Table 5The salt-taste thresholdVariablesControlHTN*p*Detection threshold5.65\u00a0\u00b1\u00a00.875.67\u00a0\u00b1\u00a00.960.900Recognition threshold6.56\u00a0\u00b1\u00a01.036.59\u00a0\u00b1\u00a00.970.823*HTN* hypertension Fig. 2Histogram of the salt taste threshold Fig. 3The relationship between the detection (**a**) and recognition (**b**) threshold of salt and daytime systolic blood pressure (BP) Fig. 4The relationship between the 24-h urinary sodium excretion (**a**) and 24-h urinary sodium/potassium ratio (**b**) and recognition threshold of salt\n\nDiscussion {#Sec15}\n==========\n\nIn our study, both the detection and recognition threshold of salt had no relationship to the 24-h urinary sodium excretion, which is an estimate of the sodium intake. The systolic and diastolic BPs positively correlated with the 24-h urinary sodium excretion, and that relationship persisted in females, but not in males. Further, the salt-taste thresholds in the hypertensive group did not differ from those in the control group.\n\nSalt taste-threshold and hypertension {#Sec16}\n-------------------------------------\n\nThere are some controversies about the relationship between the salt-taste threshold and BP status. Some studies have reported that hypertensive patients have an elevated recognition threshold for salt \\[[@CR11]--[@CR14]\\], but other studies have not noted any difference in the recognition threshold between normotensive and hypertensive persons \\[[@CR15], [@CR16]\\]. Because most of those studies were conducted in a small population, Fischer et al. assessed the relationship in a quite large study population with over 2000 people \\[[@CR8]\\]. They reported that there was no significant association between hypertension and the salt taste intensity, and the salt taste intensity of males was less strong than that of females. Similarly, in our study, the detection and recognition thresholds for salt did not differ between the control and hypertension groups. It is probably because people with a recognition threshold level of 6 and 7 accounted for nearly 70% of the total study population, which meant that it may be hard to discern the differences between the groups. Although there are still controversies about the association between the salt-taste sensitivity and blood pressure, to date, there is little evidence to say that hypertensive patients have a higher salt-taste threshold than normotensive people.\n\nSalt recognition threshold and urinary sodium excretion {#Sec17}\n-------------------------------------------------------\n\nOther studies concerning the relationship between the amount of salt intake and salt-taste sensitivity showed that there was no significant association, but their study populations were healthy volunteers \\[[@CR17]--[@CR19]\\]. In agreement with their studies, no correlation was observed between the 24-h urinary sodium excretion and recognition threshold. One explanation may be that the salt-taste sensitivity itself did not seem to affect the individual's habitual salt intake. This may be because the majority of the source of an individual's sodium intake is determined by food processing steps not by discretionary salt use \\[[@CR20]\\]. Various dietary sodium sources such as the discretionary use of cooking and table salt, inherent salt in food, added salt in processing, etc. comprise an individual's total sodium intake. If more than two thirds of the total intake of salt is determined regardless of an individual's preference for salt, it is a natural result that the salt-taste sensitivity does not affect a person's total sodium intake. Moreover, there is no consistent evidence about the association between the salt preference and actual salt intake \\[[@CR16], [@CR18], [@CR21], [@CR22]\\]. In this regard, only an individual's effort to reduce the sodium intake has a critical limitation.\n\nUrinary sodium excretion and the blood pressure {#Sec18}\n-----------------------------------------------\n\nThere are lots of epidemiological, migratory and experimental studies that have shown a positive relation between the amount of salt intake and the BP \\[[@CR23], [@CR24]\\]. Similar to the INTERSALT (International Cooperative Study on Salt, Other Factors, and Blood Pressure) study, the 24-h urinary sodium excretion was significantly higher in the hypertensive group than control group in our study. Migrants to high salt intake areas had a higher BP and greater increase in the BP along with an increasing age than did their relatives left behind \\[[@CR25], [@CR26]\\]. Moreover, there are many studies showing that a reduction in the salt intake in a population led to a decrease in the BP in the community \\[[@CR27], [@CR28]\\].\n\nIn most epidemiological studies about salt intake, males tend to have a higher urinary sodium excretion than females \\[[@CR29]\\]. The exact reasons for the difference are not known but in part, there is a higher amount of food intake and higher body weight in males \\[[@CR30]\\]. Our data are in agreement with these findings. The daytime systolic BP had a positive correlation to the urinary sodium excretion in females, but not in males. One of the reasons may be that a female's blood pressure is more sensitive to salt loading than a male's. Kojima et al. observed a BP reduction only in females when the dietary sodium intake was dramatically decreased \\[[@CR31]\\]. Also, He et al. reported greater BP responses in females compared to males in dietary sodium intervention \\[[@CR32]\\]. Females tend to have a lower body weight than males and especially during the postmenopausal period, the change in the composition of the sexual hormones, like estrogen, and testosterone, may also affect the salt sensitivity to the BP response \\[[@CR33]\\].\n\nLimitations {#Sec19}\n-----------\n\nThis study had several limitations to consider. First, the taste thresholds were distributed relatively in a small range so it was hard to tell whether the differences between the groups existed or not. Therefore, a larger study is needed to evaluate the accurate relationship between the taste threshold and BP. Second, the weakness of the 24-h urine collection method places a heavy burden on the patients. It is not easy to collect the total 24-h urine correctly for socially active persons. Unfortunately, we could not check the 24-h urine creatinine to see if the urine samples were adequately collected. Further, an individual's daily sodium intake varies day by day. Ideally, several 24-h urine collections are needed to estimate an individual's accurate sodium intake. Because of the difficulty in collecting 24-h urine samples, many studies have inferred the 24-h urinary sodium excretion using Tanaka's method or Kawasaki's method from spot urine sample, or using a food frequency questionnaire \\[[@CR34], [@CR35]\\]. Thus far, 24-h urine collections have been known to be the gold standard for evaluating the 24-h urinary sodium excretion as compared to the various method using spot urine samples. Third, we obtained urine samples after explaining the impact of the salt intake on hypertension. That could influence the patient's salt intake habits. In Korea in 2014, the average sodium intake per capita was 3.9\u00a0g/d (salt equivalent 9.72\u00a0g/d) \\[[@CR36]\\]. However, in our study, the average 24-h urinary sodium excretion was 3.1\u00a0g (133.2\u00a0mEq)/d, that is a much lower sodium intake amount. That was inevitable in order to conduct the study and to receive a written consent from patients. Finally, this was a single center study and had small study population. Therefore, multicenter and large scaled studies are required to precisely evaluate the relationship among the salt-taste threshold, 24-h urinary sodium excretion, and BP.\n\nConclusions {#Sec20}\n===========\n\nThe salt-taste threshold might not be related to the BP status as well as the 24-h urinary sodium excretion.\n\nThis research received research funding from by The Society of Korean Hypertension (2014).\n\nFunding {#FPar1}\n=======\n\nThis research received research funding from by The Society of Korean Hypertension (2014).\n\nAvailability of data and materials {#FPar2}\n==================================\n\nThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\n\nCYK analyzed and interpreted the patient data. YSL designed this study and collected patient data and was a major contributor in writing the manuscript. MKY performed the salt-taste examination. All authors read and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar3}\n==========================================\n\nThis study was approved by the institutional review board of Daegu Catholic Medical Center.\n\nConsent for publication {#FPar4}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar5}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar6}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "![](medphysj74060-0025){#sp1 .212}\n\n![](medphysj74060-0026){#sp2 .213}\n\n![](medphysj74060-0027){#sp3 .214}\n\n![](medphysj74060-0028){#sp4 .215}\n\n![](medphysj74060-0029){#sp5 .216}\n\n![](medphysj74060-0030){#sp6 .217}\n\n![](medphysj74060-0031){#sp7 .218}\n"}
+{"text": "![](indmedgaz72368-0030){#sp1 .94}\n\n![](indmedgaz72368-0031){#sp2 .95}\n\n![](indmedgaz72368-0032){#sp3 .96}\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nBreast cancer is a leading cause of mortality and morbidity all over the world. In 2008, close to 1.4 million cases were diagnosed with breast cancer worldwide \\[[@B1]\\]. The incidence varies among different populations with high rates seen in developed countries compared to developing countries \\[[@B2], [@B3]\\]. In general, breast cancer rates are highest in white European and lowest in east Asian populations \\[[@B1], [@B4], [@B5]\\]. The estimated incidence rate for women living in the south-east Asia region of World Health Organization\\' is 26.1 per 100000 population and this figure is 89.7 for women living in Western Europe \\[[@B1]\\]. The established risk factors of breast cancer are, mainly, early age at menarche, late age at menopause, nulliparity, number of live birth, and age at first live birth. Contrary to the large variations seen in incidence between population of Europe and Asia, the prevalence of the established risk factors is not very much different between the two populations \\[[@B5], [@B6]\\]. A higher risk of breast cancer among American and European women has been blamed for so-called \"western lifestyle\" characterized \\[[@B7], [@B8]\\] by the combination of early menarche, decreased parity, delayed childbearing, and a sedentary lifestyle. Studies of migrants have confirmed the relative importance of environment and lifestyle in the etiology of breast cancer \\[[@B9]--[@B11]\\]. The so-called \"western lifestyle\" is now very common in Asian countries such as Japan, Korea, Taiwan, and Hong Kong and is spreading fast in the economic booming region of East Asia. In addition to the major differences in magnitude of rates between Asians and Europeans, there is a distinct difference in the shape of age specific rates between the two populations. In Asian population, the age specific incidence curve peaks at 45--50 and then plateaus and even slightly decreases so that the rates after 60 years are less than or close to the rates at age group 45--55. In European population, the age specific curve increases steadily with no change of pace around age 50 and the increase continues up to age 80 with a peak around 65 years. A lower risk and the distinct pattern of age specific incidence rates among the Asian population (even in countries with great extent of similar lifestyle with western population such as Japan, Taiwan, and Hong Kong) have been a challenging issue among epidemiologists and cancer scientists to the extent that some have labeled breast cancer in the population of Asia as a different disease \\[[@B12]\\]. A recent symposium in Montreal, Canada, specifically addressing the same topic, highlighted the younger age at onset as one of the chief characteristics of breast cancer natural history in Asian population \\[[@B12]\\]. What contributes to the peculiar phenomenon of younger age at onset has been the subject of inertest to epidemiologists, and it has been hypothesized that a cohort effect among Asian population causes this phenomenon. The aim of this study was to use the incidence rates reported in the Cancer Incidence in Five Continents for the two populations of Asia and Europe in order to address the nature of the age specific rate differences between the two populations using age-period-cohort analysis.\n\n2. Material and Methods {#sec2}\n=======================\n\n2.1. Patients and Population {#sec2.1}\n----------------------------\n\nRegistered cases of female breast cancer and corresponding person years were ascertained from the *CI5plus, Cancer Incidence in Five Continents Annual Dataset* (an online data repository of International Agency on Research on Cancer, IARC) for 29 registries in Europe and 9 registries in Asia for a duration from 1953 to 2002 \\[[@B13]\\]. Cases from Europe included cases registered for the period from 1953 to 2002 and cases from Asia included cases registered for the period from 1963 to 2002. Cases and their corresponding person years were pooled for each population to make two distinct populations, referred to hereafter as Asian and European.\n\n2.2. Statistical Analysis {#sec2.2}\n-------------------------\n\nDescriptive and analytical approaches were used to analyse the data. Age specific rates were addressed both cross-sectionally (period wise) and longitudinally (generational birth cohort wise). For period wise age specifics, the rates were constructed and described based on five years period (1955, 1960, 1965,... 2000). For the cohort wise, age specific rates were estimated from age-period-cohort analysis constructed over each five years cohort. In addition, the trends in rates expressed as annual percentage change (APC) and their 95% confidence intervals were estimated using the age-period-cohort model.\n\nFor analytical part, the age-period-cohort model was used. For this, the periods and cohorts were constructed in intervals of 5 years. The period included 8 intervals for Asian population and 10 intervals for European population. The cohort included 19 intervals for Asian population and 21 intervals for European population. The five-year age groups were truncated to age more than 25 years with the last interval (85 and over) included all cases more than 85 years (13 five-year age groups were constructed). The constrained generalized linear model (CGLM), the most utilized approach in the epidemiology literature dealing with age-period-cohort analysis, was used. For this purpose, a log-linear model with the general form that includes *a* (age), *p* (period), and *c* (cohort) was applied as follows: $$\\begin{matrix}\n{\\text{Log}\\left\\{ {\\lambda\\left( {a,p} \\right)} \\right\\} = f\\left( a \\right) + g\\left( p \\right) + h\\left( c \\right),} \\\\\n\\end{matrix}$$ where *a*, *p*, and *c* represent the mean age, period, and cohort and *f*, *g*, and *h* are parametric functions fitted to the data. In this model, in addition to estimating the main effect of age, other components contributing to magnitude of rates specially the secular changes of rate across study periods and birth cohorts are estimated. The secular change or net drift corresponds, interchangeably, to hazard due to period or cohort, and it has been used to estimate the annual percent changes of rates over a period of time \\[[@B14], [@B16]\\]. As the purpose of our study was to tackle the difference between the two populations\\' age specific rates, it was assumed that mainly the cohort effect explains the changing of rates across aging intervals during the study period in both populations. With this assumption, the model estimates the age function presented as the log of the age specific rates for the reference cohort (longitudinal age specific or age specifics across cohorts) and the cohort effect as log of rate ratio relative to a reference cohort while period effect constrained to be zero on average with zero slopes. The estimated logs of age specific rates were transformed to rate scale (number per 100000 population) for better realization. In the model, the cohort born during 1970 was considered as the reference cohort and the period of 1970 was considered as the reference period. The longitudinal age specific raters were estimated and reported for cohort born on 1885, 1910, 1930, 1950, and 1970. For details of the modeling please refer to \"age-period-cohort models for the Lexis diagram\" by Carstensen \\[[@B19]\\]. Data were analyzed using the R 2.14.1 statistical software utilizing Epi 1.1.9 package (R Development Core Team, 2009).\n\n3. Results {#sec3}\n==========\n\nA total of 236,851 cases of breast cancer registered in the 29 European registries and a total of 188,630 cases registered in the 9 Asian registries were included in the analysis ([Table 4](#tab4){ref-type=\"table\"} presents details of the included registries for the two populations).\n\n3.1. Descriptive Approach {#sec3.1}\n-------------------------\n\nThere was a constant increasing of rates for both populations during the last 50 years with an estimated annual percent of change 1.03 (with 95% CI of 1.029, 1.031) for Asians and 1.016 (95% CI of 1.015, 1.017) for Europeans. The incidence rates across all age groups in Europeans were higher than Asians, especially in older age groups. During the study period, the magnitude of rates increased for both populations for each succeeding five-year period for all age groups. The shape of the age specific rates (period wise) showed basic differences between the two populations. For Asian population, the age specific rates for all periods peaked around 50 years and then decreased and plateaued afterward ([Figure 1](#fig1){ref-type=\"fig\"}). For the European population, the age specific rates increased up to the last age group for periods ending 1985 and for the periods after 1985, the age specific rates peaked between 55 to 75 years and then slightly decreased ([Figure 1](#fig1){ref-type=\"fig\"}).\n\nThe fitting of the age-period-cohort model to data indicated that the model that included all the main effects (age, period, and cohort) has the greatest reduction of deviance, indicating the best model to explain the observed rates in both populations; [Table 1](#tab1){ref-type=\"table\"} presents the goodness of fit of the models along with their parameters. There were cohort effects present in incidence rates of both populations during the study period; however, the cohort effects in Asians were much stronger than European. In the Asian population, the rate ratios presenting the cohort effects ranged from a low of 0.06 (95% CI 0.05, 0.08) for those born in 1870 to 0.94 (95% CI, 0.93, 0.96) for those born in 1965. In the European population, the rate ratios presenting the cohort effect ranged from 0.33 (95% CI, 0.32, 0.35) for cohort born in 1865 to 1.03 (95% CI, 1.02, 1.04) for the cohort born in 1965 ([Table 2](#tab2){ref-type=\"table\"} and [Figure 2](#fig2){ref-type=\"fig\"}). There were residual period effects in the Asian population around 1975 (rate ratio of 0.89 and 95% CI 0.86, 0.92) and 1985 (rate ratio of 1.12 95% CI of 1.09, 1.13), [Figure 2](#fig2){ref-type=\"fig\"}. For both populations, the estimated age specific rates expressed as longitudinal age specific indicated the same pattern for both populations; the age specific rates increased sharply before the age of 50, and the increase slowed down pace with the last age groups (over 75 years), still the groups with highest incidence rates. The pattern of longitudinal age specific rate is presented in [Figure 2](#fig2){ref-type=\"fig\"} along with the other effects, cohorts, and periods. The estimated longitudinal age specific rates (in an increment of 20 years) and their corresponding confidence intervals are presented numerically in [Table 3](#tab3){ref-type=\"table\"} and graphically in [Figure 3](#fig3){ref-type=\"fig\"}. As [Table 3](#tab3){ref-type=\"table\"} and [Figure 3](#fig3){ref-type=\"fig\"} indicate, the estimated age specific rates have steadily increased in all age groups for both populations but the increase is more in Asians compared to Europeans. There is a large difference in the magnitude of rates between the two populations in early cohort (1890) when they are compared with the most recent cohort (1970), [Table 3](#tab3){ref-type=\"table\"} and [Figure 3](#fig3){ref-type=\"fig\"}. The difference in age specific rates between early and late cohorts is indicative of the cohort effects that cause the distinct pattern of age specific rates observed between the two populations. In addition, comparing the magnitude of the cohort effects between the two populations ([Figure 2](#fig2){ref-type=\"fig\"}), they indicate that, though, the cohort effects are decreasing along succeeding cohorts for both populations, but the decrease in cohort effects in Asians is far larger than those of Europeans. This difference in decreasing rates of Cohort effects between the two populations indicates that both populations may experience similar rates in the future if there are no other major changes to the underlying cause of the disease in future years.\n\n4. Discussion {#sec4}\n=============\n\nOur study proved a steady increase of breast cancer rate with similar pace during the last 50 years for both populations. We demonstrated that there is no difference between the patterns of age specific rates between the two populations when rates are measured as longitudinal age specific rates. It was demonstrated that a strong cohort effect contributes to the differences in pattern of age specific rates between the two populations. The difference in breast cancer rates with low rates for Asians versus high rates for Europeans has been documented since registries in Asia started reporting population rates \\[[@B20], [@B21]\\]. While several studies have demonstrated marked differences in magnitude and the pattern of age specific rates among different countries of Europe and Asia, no study systematically and collectively has addressed the age specific rate differences in the two populations as our study did. An overall increasing trend of morbidity from breast cancer has been reported for all populations of the world and the increase has been attributed to ageing and increasing median age of women \\[[@B16], [@B20], [@B22]--[@B25]\\]. Our study showed a very similar increase of incidence between the two populations. This similarity in slope of increase indicates that despite the fact that the two populations are basically different in terms of culture, ethnicity, lifestyle, and social attributes, the breast cancer epidemic enforces its own pace of epidemic projection.\n\nAny increase in incidence of breast cancer rates is due to either changing of risk factors or implementation of mass screening (especially mammographic screening). Either of the two can affect both magnitude and pattern of age specific rates. The difference in the pattern of the age specific rates between the two populations is well recognized and several studies have addressed this discrepancy; a study comparing the shape of the age specific rates between Taiwanese and Caucasian American reported that the age specific rates of breast cancer differed between the two populations and the study concluded that the difference is due to a cohort effect presented in Taiwanese \\[[@B27]\\]. Another study comparing breast cancer rates among populations of Singapore and Sweden attributed the difference in rates to a large cohort effect and concluded that this effect will decrease in future generations causing similar incidence rates between the two population in coming decades \\[[@B28]\\]. In addition to comparative studies, it has been demonstrated that the pattern of age specific rates for the populations of Japan, Korea, China, Singapore, Thailand, and Philippine has changed in the recent years attributing this change to changing of life style toward more westernization and implementation of mammographic screening \\[[@B29]\\]. The pattern and magnitude of age specific rates of breast cancer have been affected by screening mammography specially in the European population, and this effect has been mainly presented as increase of incidence in the age group of 50 to 70 \\[[@B30]--[@B32]\\]. This is compatible with our finding as it was demonstrated in [Figure 1](#fig1){ref-type=\"fig\"} that age specific rates for European population increased in that age group 50 to 70 and the change happened after 1985 when the wide spread use of mammographic screening started \\[[@B31], [@B33]\\]. Mammographic screening started in Asia in late 1990 \\[[@B34], [@B35]\\] and its effect on the shape of age specific incidence cannot be assessed in our study. Compatible with previous studies comparing the age specific rates of breast cancer between population of Europe and Asia, our study proved that a large cohort effect in Asian population rates plays a major role in the differences in age specific rates between the two populations. The age specific rates can be defined both cross-sectionally (period wise) and longitudinally (cohort wise). If there are no cohort or period effects, the two definitions will show similar magnitude and pattern of age specific rates. The distinct pattern of age specific rates (cross-sectional rates) between the two populations is in fact due to the cohort effect that was demonstrated in our study. In the other world, what contributes to the observed pattern of period wise age specific rates in Asian population is the additive nature of the cross-sectional definition of age specific rates in the presence of decreasing cohort effects. Since the rates for two sequential age groups come from two different cohorts, when the older cohort has lower risk compared to younger cohort, the cross sectional patterns of age specifics will decrease. The strong cohort effect that exists in the Asian breast cancer is responsible for a distinct pattern of cross-sectional age specifics seen in the Asian population.\n\nIn applying the CGLM model, the choice of constraint (period or cohort) is based on an external knowledge of the underlying cause of change of rates in a population. Our choice of cohort instead of period as constraint was based on previous studies that attributed the changes of the rates to a cohort effect \\[[@B16], [@B22], [@B37]\\]. The period effect has been mainly attributed when change of health policy (e.g., introduction of more sensitive detection techniques or availability of certain diagnostic procedures) causes the changing of rates. In the light of the nature of breast cancer risk factors that are mainly of hormonal and sociobehavioral nature, their change would translate on cohort effect than on period effect.\n\nThis study enjoyed data of adequate quality as the data were utilized from the registries that met acceptable degree of validity and reliability to be published in the International Agency in Research on Cancer (IARC) official report. In addition, the quality of the data for both populations is comparable owing to the efforts and quality assurances and control protocols that IARC requires for different registries contributing data to the Cancer Incidence in Five Continents reports; one may ask. We would have been able to draw the same conclusion on analyzing just the data of Asian population without the need of European population. The main reason for using European for comparison was that the fact that the specificity attributed to Asian breast cancer age specific rates has been defined as it has contrasted to the well-sestablished breast cancer epidemiology in Europe and western countries.\n\nThe methodology we used is a very established and routine way of analyzing rates at population level when the data of calendar time exist. The age-period-cohort analysis has been a major tool in the hands of demographers and, in recent decades well utilized by epidemiologists. The methodology, while very common in use, suffers major problem especially when it is utilized to attribute the underlying cause of changes of a rate to period versus cohort (the nonidentifiability problem). In our study, this problem was not a concern as our assumption was that the nature of risk factors in breast cancer would translate into cohort effect other than period effect.\n\n5. Conclusion {#sec5}\n=============\n\nIt was concluded that no differences in the pattern of age specific rates exist between the two populations when the age specifics are measured cohort wise, and the difference seen in the period wise age specific rate is due to a strong cohort effect present in the Asian population rates.\n\n![The cross-sectional age specific rates for both populations.](ISRN.ONCOLOGY2013-429862.001){#fig1}\n\n![Estimated effects and their 95% confidence intervals from the age-period-cohort model (breast cancer, age group \\>25 years, period 1953--2002, and cohort born on 1865--1970 for both European (red) and Asian (blue) populations). Curve in the left represents the estimated age specific rates for the reference period 1970. The middle curve shows the rate ratios of cohort relative to the reference cohort (1970), cohort effect. The rightmost curves show the rate ratios of period constrained to be zero on average with zero slopes or the residual period effect.](ISRN.ONCOLOGY2013-429862.002){#fig2}\n\n![The graphical presentation of the estimated rates for different generations for both populations (red for Asian and blue for European).](ISRN.ONCOLOGY2013-429862.003){#fig3}\n\n###### \n\nSummary statistics of age-period-cohort model for breast cancer rates in European and Asian populations\\*.\n\n  Model               Asia   Europe                          \n  ------------------- ------ --------- ------- ----- ------- -------\n  Age                 514    11764.7   \u2009       644   45047   \u2009\n  Age-drift           513    1920.7    0.000   643   8227    0.000\n  Age-cohort          509    1628.1    0.000   639   4809    0.000\n  Age-period          509    1546.2    0.000   639   8118    0.000\n  Age-period-cohort   505    1308.7    0.000   635   4377    0.000\n\n\\*The model fitted sequentially.\n\n\\*\\*Degree of freedom.\n\n###### \n\nThe magnitude of cohort effects for the two populations.\n\n  Year of birth   Rate ratio (95% confidence intervals)   \n  --------------- --------------------------------------- -------------------\n  1865            \u2217\u2217                                      0.33 (0.32, 0.35)\n  1870            0.06 (0.05, 0.08)                       0.35 (0.33, 0.36)\n  1875            0.07 (0.06, 0.09)                       0.36 (0.35, 0.37)\n  1880            0.08 (0.07, 0.10)                       0.37 (0.36, 0.38)\n  1885            0.09 (0.08, 0.11)                       0.38 (0.37, 0.40)\n  1890            0.11 (0.10, 0.12)                       0.40 (0.39, 0.41)\n  1895            0.13 (0.12, 0.14)                       0.41 (0.40, 0.42)\n  1900            0.15 (0.14, 0.16)                       0.43 (0.42, 0.44)\n  1905            0.18 (0.17, 0.19)                       0.45 (0.43, 0.46)\n  1910            0.22 (0.20, 0.23)                       0.47 (0.46, 0.49)\n  1915            0.25 (0.24, 0.27)                       0.51 (0.50, 0.53)\n  1920            0.28 (0.27, 0.30)                       0.57 (0.55, 0.58)\n  1925            0.31 (0.29, 0.33)                       0.64 (0.62, 0.65)\n  1930            0.36 (0.34, 0.38)                       0.73 (0.71, 0.75)\n  1935            0.43 (0.41, 0.46)                       0.83 (0.81, 0.86)\n  1940            0.54 (0.51, 0.57)                       0.93 (0.90, 0.96)\n  1945            0.64 (0.61, 0.68)                       1.00 (0.97, 1.03)\n  1950            0.73 (0.70, 0.77)                       1.04 (1.02, 1.07)\n  1955            0.82 (0.78, 0.85)                       1.06 (1.04, 1.08)\n  1960            0.88 (0.86, 0.91)                       1.05 (1.04, 1.07)\n  1965            0.94 (0.93, 0.96)                       1.03 (1.02, 1.04)\n  1970            1.00                                    1.0\n\n\\*\\*No data for Asian population.\n\n###### \n\nThe estimated longitudinal age specific rates and corresponding 95% confidence intervals for the two populations for the cohort born in 1880, 1910, 1930, 1950, and 1970.\n\n  \u2009        Age group              1880                   1910                   1930                   1950                   1970\n  -------- ---------------------- ---------------------- ---------------------- ---------------------- ---------------------- ----------------\n  Europe   25                     2.3 (2.2, 2.3)         2.9 (2.8, 3.0)         4.4 (4.3, 4.5)         6.3 (6.2, 6.5)         6.1 (5.9, 6.3)\n  30       7.4 (7.3, 7.6)         9.5 (9.3, 9.7)         14.5 (14.3, 14.8)      20.9 (20.6, 21.1)      20.0 (19.6, 20.5)      \n  35       20.6 (20.3, 20.9)      26.4 (26.1, 26.7)      40.4 (40.0, 40.8)      58.0 (57.6, 58.5)      55.6 (54.2, 57.0)      \n  40       42.6 (41.9, 43.2)      54.5 (54.0, 55.0)      83.5 (82.8, 84.2)      119.9 (119.2, 120.7)   114.9 (111.8, 118.1)   \n  45       68.2 (67.3, 69.2)      87.4 (86.7, 88.0)      133.8 (133.0, 134.6)   192.2 (191.0, 193.3)   184.1 (178.9, 189.5)   \n  50       92.3 (91.0, 93.7)      118.2 (117.4, 119.1)   181.1 (180.1, 182.1)   260.1 (258.1, 262.1)   249.2 (241.9, 256.7)   \n  55       114.1 (112.5, 115.7)   146.1 (145.2, 147.0)   223.8 (222.8, 224.7)   321.3 (318.7, 324.0)   307.9 (298.9, 317.2)   \n  60       134.9 (133.1, 136.8)   172.7 (171.6, 173.8)   264.6 (263.3, 265.8)   379.9 (376.5, 383.5)   364.1 (353.5, 374.9)   \n  65       157.8 (155.7, 159.9)   202.0 (200.9, 203.1)   309.4 (307.9, 311.0)   444.4 (440.2, 448.6)   425.8 (413.5, 438.4)   \n  70       183.6 (181.3, 186.0)   235.1 (233.9, 236.3)   360.1 (357.9, 362.2)   517.1 (512.0, 522.2)   495.5 (481.1, 510.2)   \n  75       212.4 (209.7, 215.1)   271.9 (270.5, 273.4)   416.5 (413.5, 419.5)   598.1 (592.0, 604.3)   573.1 (556.4, 590.3)   \n  80       244.1 (241.2, 247.1)   312.6 (311.0, 314.3)   478.8 (475.2, 482.5)   687.6 (680.5, 694.8)   658.9 (639.5, 678.8)   \n  85       279.7 (276.1, 283.4)   358.2 (355.1, 361.3)   548.6 (543.0, 554.3)   787.8 (778.2, 797.6)   754.9 (732.1, 778.4)   \n                                                                                                                              \n  Asia     25                     0.4 (0.4, 0.5)         1.2 (1.1, 1.2)         1.9 (1.8, 2.0)         4.0 (3.8, 4.2)         5.4 (5.2, 5.7)\n  30       1.6 (1.4, 1.8)         4.2 (4.0, 4.4)         6.9 (6.7, 7.1)         14.2 (13.9, 14.6)      19.4 (18.6, 20.2)      \n  35       4.6 (4.0, 5.3)         12.0 (11.7, 12.4)      19.9 (19.4, 20.3)      40.9 (40.3, 41.4)      55.6 (53.1, 58.3)      \n  40       9.0 (7.8, 10.4)        23.7 (23.0, 24.4)      39.1 (38.3, 39.9)      80.4 (79.5, 81.3)      109.5 (103.9, 115.4)   \n  45       13.1 (11.4, 15.1)      34.4 (33.5, 35.4)      56.8 (55.8, 57.8)      116.8 (115.4, 118.3)   159.1 (150.6, 168.2)   \n  50       16.1 (13.9, 18.5)      42.3 (41.2, 43.4)      69.7 (68.6, 70.8)      143.3 (141.0, 145.7)   195.2 (184.4, 206.7)   \n  55       18.7 (16.2, 21.5)      49.1 (47.9, 50.2)      80.9 (79.9, 81.9)      166.5 (163.3, 169.7)   226.7 (214.2, 239.9)   \n  60       21.3 (18.5, 24.5)      55.9 (54.7, 57.2)      92.3 (90.9, 93.6)      189.7 (185.3, 194.3)   258.4 (244.2, 273.5)   \n  65       24.2 (21.0, 27.8)      63.6 (62.4, 64.8)      104.9 (103.2, 106.6)   215.6 (210.4, 221.0)   293.7 (277.4, 310.8)   \n  70       27.6 (24.0, 31.8)      72.7 (71.4, 74.0)      119.9 (117.3, 122.5)   246.5 (239.8, 253.4)   335.7 (316.6, 356.0)   \n  75       32.0 (27.8, 36.8)      84.2 (82.6, 85.7)      138.9 (135.1, 142.7)   285.6 (276.9, 294.6)   388.9 (366.0, 413.3)   \n  80       37.7 (32.8, 43.4)      99.2 (97.3, 101.3)     163.7 (158.7, 168.8)   336.7 (325.4, 348.3)   458.5 (430.7, 488.0)   \n  85       44.9 (39.0, 51.8)      118.2 (113.8, 122.8)   195.0 (186.3, 204.1)   401.0 (382.2, 420.7)   546.1 (508.8, 586.1)   \n\n###### \n\nDetail information about the included registries.\n\n  \u2009        Registry code                                       Registry name      Starting year of reporting data   Duration of contributing data (in years)\n  -------- --------------------------------------------------- ------------------ --------------------------------- ------------------------------------------\n  Europe   20800                                               Denmark            1953                              49\n  24600    Finland                                             1953               49                                \n  25001    France, Bas-Rhin                                    1975               27                                \n  25002    France, Calvados                                    1978               24                                \n  25003    France, Doubs                                       1978               24                                \n  25004    France, Haut-Rhin                                   1988               14                                \n  25005    France, Herault                                     1988               14                                \n  25006    France, Isere                                       1979               23                                \n  25007    France, Somme                                       1983               19                                \n  25008    France, Tarn                                        1983               19                                \n  27603    Germany, Saarland                                   1970               32                                \n  38002    Italy, Florence                                     1985               17                                \n  38007    Italy, Parma                                        1978               24                                \n  38008    Italy, Ragusa province                              1983               19                                \n  38009    Italy, Romagna                                      1988               14                                \n  38010    Italy, Torino                                       1985               17                                \n  38012    Italy, Lombardy, Varese province                    1978               24                                \n  38020    Italy, Modena                                       1988               14                                \n  52802    The Netherlands, Eindhoven                          1973               29                                \n  57800    Norway                                              1953               49                                \n  75200    Sweden                                              1958               44                                \n  75602    Switzerland, Geneva                                 1970               32                                \n  75605    Switzerland, St. Gall-Appenzell                     1983               19                                \n  82603    UK, England, Merseyside, and Cheshire               1975               27                                \n  82604    UK, England, north western                          1979               23                                \n  82605    UK, England, Oxford                                 1985               17                                \n  82609    UK, England, Birmingham, and west Midlands region   1979               23                                \n  82610    UK, England, Yorkshire                              1983               19                                \n  82620    UK, Scotland                                        1975               27                                \n                                                                                                                    \n  Asia     34400                                               China, Hong Kong   1983                              19\n  39203    Japan, Miyagi prefecture                            1975               27                                \n  39206    Japan, Osaka prefecture                             1963               39                                \n  39208    Japan, Yamagata prefecture                          1983               19                                \n  60801    Philippines, Manila                                 1983               19                                \n  70200    Singapore: Chinese                                  1968               34                                \n  70200    Singapore: Malay                                    1968               34                                \n  76401    Thailand, Chiang Mai                                1968               34                                \n\n[^1]: Academic Editors: G. Ferrandina and O. Hansen\n"}
+{"text": "Background {#Sec1}\n==========\n\nGlobally, hepatocellular carcinoma (HCC) is the sixth most common malignancy and the third leading cause of cancer-related deaths \\[[@CR1]\\]. In the Pacific region, China accounted for 84.6% of HCC incidence and 86.3% of HCC mortality with Hepatitis B virus (HBV) infection as the most common cause \\[[@CR2]\\].\n\nAlthough liver transplantation (LT) is an excellent therapeutic choice for HCC as the patients who received LT have the highest chance of cure among all other therapies, the organs shortage is still a main challenge \\[[@CR3]\\]. For this, several selection criteria of HCC candidates for LT were proposed, of them, Milan and Hangzhou criteria are the most recommended tools by the Chinese Society of Organ Transplantation \\[[@CR4], [@CR5]\\]. The high rates of HCC recurrence after LT which have been reported to be 8--30% \\[[@CR3], [@CR6]\\] remains an important cause of death for HCC patients.\n\nAlpha-fetoprotein (AFP) is a potential biomarker for early diagnosis and prediction of HCC recurrence. High level of preoperative AFP, which can be seen in approximately 60% of the HCC patients, is a risk factor for HCC recurrence and can be used to define at-risk HCC patients \\[[@CR7], [@CR8]\\]. In addition to AFP, there are well recognized preoperative risk factors which reflect the biological behavior of HCC and closely associated with post-LT HCC recurrence including vitamin K absence-II and neutrophil-to-lymphocyte ratio \\[[@CR9], [@CR10]\\]. The predominant morphological factors that show correlation with higher rates of HCC recurrence after LT include; tumor number and size \\[[@CR11], [@CR12]\\].\n\nThese risk factors were employed in a collective fashion to establish different predictive models for HCC recurrence, however unavailability of an effective, validated and reliable model to stratify patients preoperatively of HCC recurrence makes a unified practice across different countries out of reach. So, careful preoperative risk stratification of HCC recurrence is not only crucial for better management, but also very helpful to define a risk-based prioritizing strategy for selection of HCC candidates for LT.\n\nIn this multicenter study, we established a preoperative predictive model for early recurrence of HCC after LT. This model could be used as an adjuvant tool beside the conventional selection criteria to predict postoperative prognosis at a personal level more accurately.\n\nMethods {#Sec2}\n=======\n\nThe design of this study followed the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD) Statement \\[[@CR13]\\]. This study has been approved by the Scientific Committee of the China Liver Transplant Registry (<http://www.cltr.org>) which is in accordance with ethical guidelines of Helsinki Declaration 1975, as revised in 2013. Written informed consent was obtained. Data of 1512 consecutive patients who underwent LT were retrospectively recalled from the prospectively maintained database of (CLTR) from 2015 January to 2019 February. Inclusion criteria were \\[[@CR1]\\] adult patients with age\u2009\u2265\u200918 \\[[@CR2]\\] pre-operative radiologically diagnosed HCC depending on guidelines of the current guidelines of American Association for the Study of Liver Diseases (AASLD) \\[[@CR14]\\] \\[[@CR3]\\] no history of previous LT or combined hepatorenal transplantation \\[[@CR4]\\] patients who survived at least 3\u2009months after the date of surgery \\[[@CR5]\\] no incidental HCC \\[[@CR6]\\] all the clinical and laboratory data required for the analysis are available. After applying the inclusion criteria, 748 patients were involved in the final analysis and divided randomly into training (*n*\u2009=\u2009486) and validation (*n*\u2009=\u2009262) cohorts. Data collection were performed by independent researchers blinded to statistical analysis. The collected clinicopathological variables included; age, gender, diabetes and hypertension, body mass index (BMI), presence of hepatitis B virus (HBV) infection, cirrhosis, Model for End-stage Liver Disease (MELD), Child score, neoadjuvant therapy (i.e. transarterial chemoembolization (TACE), radiofrequency ablation (RFA) and hepatectomy), donor type, donor death cause. The Pre-LT characteristics of HCC were obtained from radiological assessment (mainly CT, MRI), including the total tumor diameter, largest tumor diameter, number of nodules and the last pre-LT measurements of AFP. Post-LT features of HCC were obtained from the pathology reports including lymphovascular invasion and tumor differentiation according to the modified Edmondson score \\[[@CR15]\\]. Data of survival and recurrence, including death cause, last follow-up dates, recurrence and death dates. Milan, Hangzhou criteria and AFP model were calculated \\[[@CR4], [@CR5], [@CR16]\\]. The last censoring date of this study was 21st February 2019.\n\nOutcome and definitions {#Sec3}\n-----------------------\n\nThe primary outcome of this study is 2-year recurrence rate of HCC after LT. The recurrence was considered as 'early' if the time from LT to recurrence was \u22642\u2009years \\[[@CR17]\\]. Time to recurrence was calculated from the date of LT surgery to the date of recurrence diagnosis or last follow-up. The corresponding overall survival (OS) was also calculated from the date of LT surgery to date of death or last follow-up.\n\nThe selection criteria for LT included in this study are Milan and Hangzhou criteria. The former required the absence of distance metastasis and macrovascular invasion and included patients with a single nodule \u22645\u2009cm or\u2009\u2264\u20093 nodules (each nodule \u22643\u2009cm) \\[[@CR4]\\], while the latter required the absence of macrovascular invasion and included patients with (a) total tumor size\u22648\u2009cm, (b) total tumor diameter\u2009\\>\u20098\u2009cm, well moderate tumor differentiation and preoperative AFP level\u2009\u2264\u2009400\u2009ng/mL, concurrently \\[[@CR5]\\].\n\nAFP model is a binary tool incorporated largest tumor size, number of nodules and pre-LT AFP (at 100 and 1000\u2009ng/mL) with a cut-off value of two points to discriminate patients within and outside the AFP model patients \\[[@CR16]\\].\n\nThe death was defined as HCC- related death if there is an evidence of HCC recurrence post-LT or documented metastasis and/or vascular invasion otherwise it was considered as HCC-unrelated death. The last censoring date or the date of events (recurrence and death) were considered after following up all patients.\n\nPostoperative management and follow up {#Sec4}\n--------------------------------------\n\nGenerally, postoperative immunosuppressants consisted of calcineurin inhibitors and steroids. Steroids were withdrawn within 3\u2009months. Follow up of the patients were done every 3--6\u2009months during the first 2\u2009years post-LT. During follow up time, in addition to AFP measurement, abdominal computed tomography (CT) scan and magnetic resonance imaging (MRI) were performed according to their indications.\n\nStatistical analysis {#Sec5}\n--------------------\n\nStatistical analysis was performed using Stata MP 14. Categorical data were reported as values and percentages and compared using Fisher's exact test or Chi-Square test. Continuous data were reported as mean\u2009\u00b1\u2009SD or median (interquartile range \\[IQR\\]) and compared with Student's T-test or rank sum test according to their distribution respectively. Recurrence and survival probabilities were estimated by Kaplan-Meier (KM) methods and compared using the log-rank test (Mantel-Cox). Univariable and multivariable Cox regression analyses for factors affecting post-LT HCC recurrence were performed by Cox proportional hazards regression models. Variables after a univariable analysis with a *P*-value \\<\u20090.05 were included in the multivariable analysis, and the final model was constructed by stepwise backward selection (Wald test). Notebaly,the potential cut-off values were estimated in accordance with previous studies in the literature and using the Akaike information criterion (AIC), the cut-off points with the lowest AIC values were selected to be included in the final model. Proportional-hazards assumption was assessed by the Schoenfeld test and by visual assessment of log-log survival curves. The training and validation cohorts were compared. The discriminatory performance of 5--8 Model, Milan, Hangzhou, and AFP model was calculated and compared using Harrell's C and Somers' D statistics \\[[@CR18]\\] and also was assessed visually via KM curves. Moreover, the competing risk analysis was also performed to evaluate the cumulative incidence of HCC related and unrelated deaths \\[[@CR19]\\]. A two-tailed *p*-value of \\<\u20090.05 indicates a statistically significance difference.\n\nResults {#Sec6}\n=======\n\nA total of 748 patients were included in the study with a mean age of 51.6\u2009\u00b1\u20098.6 and 89.6% were male. HBV infection was the most common cause (96%) and cirrhosis was found in 84%. The 486 patients of the training cohort had similar characteristics to the 262 patients of validation cohort without any significant differences as summarized in (Table\u00a0[1](#Tab1){ref-type=\"table\"}). For the training and validation cohorts, the median post-LT follow-up was 338\u2009days, (IQR: 205--673\u2009days) and 416.5\u2009days, (IQR:205--672\u2009days), respectively. The 2-year OS was 82.6% (95%CI:0.77--0.87) vs 81.9% (95% CI: 0.74--0.87), *p*\u2009=\u20090.870(Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a). Recurrence of HCC was observed in 12.8% (62 of 486) vs 16.8% (44 of 262) at a median of 11.3\u2009months (IQR: 5.9--21.2, months) vs 12.0\u2009months IQR:6.2--20.5\u2009months) after LT. The 2-year overall HCC recurrence was 17.1% (95% CI:0.13--0.22) vs 25.9% (95% CI: 0.19--0.36), *p*\u2009=\u20090.180 (Fig. [1](#Fig1){ref-type=\"fig\"}b). Table 1Baseline characteristics of training and validation cohortsVariableTraining cohort (n\u2009=\u2009486)Validation cohort(n\u2009=\u2009262)*P*-valueAge (years)^a^51.6\u2009\u00b1\u20098.651.5\u2009\u00b1\u20098.70.890Gender (Male/Female) ^b^439 (90.3)/47 (9.7)231 (88.1)/31 (11.8)0.356Diabetes (Yes/No) ^b^62 (12.8)/424 (87.2)40 (15.3)/222 (84.7)0.340Hypertension (Yes/No) ^b^48 (9.9)/438 (90.1)33 (12.6)/229 (87.4)0.254BMI^a^24.4\u2009\u00b1\u200912.923.3\u2009\u00b1\u20093.80.162Cause (HBV/Non-HBV) ^b^462 (95.1)/24 (4.9)256 (97.7)/6 (2.3)0.078Cirrhosis (Yes/No) ^b^411 (84.6)/75 (15.4)217 (82.8)/45 (17.2)0.535Pre-LT AFP (ng/mL) ^b^\u2003\u2009\u2264\u200910164 (33.7)94 (35.9)0.621\u200310--200167 (34.4)95 (36.3)\u2003200--100082 (16.9)35 (13.4)\u2003\u2009\\>\u2009100073 (15.0)38 (14.5)Total tumor diameter (cm)^b^\u2003\u2009\u2264\u20095283 (58.2)148 (56.5)0.880\u20035.1--8104 (21.4)57 (21.8)\u2003\u2009\\>\u2009899 (20.4)57 (21.8)Largest tumor diameter (cm)^b^\u2003\u2009\u2264\u20094309 (63.6)162 (61.8)0.296\u20034--683 (17.1)56 (21.4)\u20036.1--842 (8.6)15 (5.7)\u2003\u2009\\>\u2009852 (10.7)29 (11.1)Tumor number (single/multiple) ^b^296 (60.9)/190 (39.1)151 (57.6)/111 (42.3)0.384MELD^c^12 \\[8--21\\], (6--51)11 \\[9--19\\], (6--44)0.503Child score^c^7 \\[5--9\\], (5--14)7 \\[5--9\\], (5--14)0.473Neoadjuvant therapy^b^\u2003TACE (yes/no)195 (40.1)/291 (59.9)113 (43.1)/149 (56.9)0.425\u2003RFA (yes/no)85 (17.5)/401 (82.5)49 (18.7)/213 (81.3)0.680\u2003Hepatectomy (yes/no)83 (17.1)/403 (82.9)41 (15.7)/221 (84.4)0.616\u2003Donor type ^b^ DBD/DCD/DBCD142 (29.2)/174 (35.8)/170 (35.0)70 (26.7)/103 (39.3)/89 (34.0)0.609Donor death cause ^b^\u2003Trauma233 (47.9)138 (52.7)0.404\u2003CVA190 (39.0)94 (35.9)\u2003Tumor29 (6.0)10 (3.8)\u2003Anoxia17 (3.5)13 (4.9)\u2003Others17 (3.5)7 (2.7)Differentiation ^b^\u2003well72 (14.8)42 (16.0)0.884\u2003moderate338 (69.6)178 (67.9)\u2003poor76 (15.6)42 (16.0)Vascular invasion (yes/no) ^b^113 (23.3)/373 (76.8)74 (28.2)/188 (71.8)0.132Milan (in/out) ^b^259 (53.3)/227 (46.7)125 (47.7)/137 (52.3)0.145Hangzhou (in/out) ^b^390 (80.3)/96 (19.8)197 (75.2)/65 (24.8)0.108AFP model (in/out) ^b^301 (61.9)/185 (38.1)167 (63.7)/95 (36.3)0.626Post-LT mortality (Died/alive) ^b^55 (11.3)/431 (88.7)31 (11.8)/231 (88.2)0.833Post-LT recurrence (yes/no) ^b^62 (12.8)/424 (87.2)44 (16.8)/218 (83.2)0.131Time to recurrence (months)^c^11.3 \\[5.9--21.2\\], (0.2--47.0)12.0 \\[6.2--20.5\\], (1.1--44.8)0.863Follow-up (days)^c^388 \\[205--673\\], (92--1428)416.5 \\[205--672\\], (92--1363)0.802Note: *BMI* Body mass index, *HBV* Hepatitis B virus infection, *AFP* Alpha-fetoprotein, *MELD* Model for End-Stage Liver Disease, *LT* Liver transplantation, *TACE* Transarterial chemoembolization, *RFA* Radiofrequency ablation, *DBD* Donation after brain death, *DCD* Donation after circulatory death, *DBCD* Donation after brain death followed by circulatory death, *CVA* Cerebrovascular accident, ^a^Mean\u2009\u00b1\u2009SD, ^b^ number (percentage), ^c^ (median, \\[IQR, interquartile range\\]),(range) Fig. 1The 2-year recurrence and overall survival rates in the training (**a**) and validation(**b**) cohorts\n\nFactors affecting post-LT HCC recurrence {#Sec7}\n----------------------------------------\n\nFactors of post-LT HCC recurrence in the training cohort were identified by univariable Cox regression analysis (Table\u00a0[2](#Tab2){ref-type=\"table\"}). The preoperative factors that associated with post-LT HCC recurrence included pre-LT TACE, pre-LT AFP, total tumor diameter (cm), the largest tumor diameter (cm) and the number of nodules at all tested cut-off values. The postoperative factors included; vascular invasion, poorly differentiated tumor grade. Other factors including recipient age, gender, MELD score, number of HCC nodules, use of other neoadjuvant therapies (i.e. hepatectomy, RFA) did not associate with post-LT HCC recurrence. We emphasize that the tumor diameters and the number of nodules were obtained from the last imaging before transplantation by which Milan and Hangzhou criteria were calculated. On multivariable analysis, only pre-LT AFP, largest tumor size and tumor number were the preoperative predictors found to be associated with increasing the risk of post-LT HCC recurrence. While the postoperative predictors were the only presence of vascular invasion and poorly differentiated tumor grade. Table 2Univariable cox analysis of risk factors for early recurrence of HCCVariable2-year recurrence rateHRSEz*P*\u2009\\>\u2009z95% CIAge (52/\u226452\u2009years)17.6/16.40.980.250.070.9420.59--1.63Gender (male/female)17.9/9.21.510.780.800.4260.55--4.16Diabetes (yes/no)19.4/17.10.790.340.550.5850.34--1.84Hypertension (yes/no)24.7/16.31.540.591.140.2530.73--3.25BMI (25/\u226425)18.5/16.40.990.280.050.9590.561.73CTP class(A/B/C)19.5/14.9/15.30.940.160.380.7050.67--1.31MELD (12/\u226412)13.7/20.20.810.210.810.4160.49--1.35Cirrhosis (yes/no)17.414.21.240.470.570.5660.59--2.62Pre-LT AFP (ng/mL) (Reference, \u226410)\u200310--20019.53.571.652.750.0061.44--8.83\u2003201--100021.84.652.293.110.0021.77--12.23\u2003\u2009\\>\u2009100039.29.034.204.72\\< 0.0013.62--22.49Total tumor diameter (cm) (Reference, \u22645)\u20035--821.93.201.193.130.0021.55--6.64\u2003\u2009\\>\u2009852.68.892.886.75\\<\u20090.0014.7116.76Largest tumor diameter (cm) (Reference, \u22644)\u20034--621.82.370.902.280.0231.13--4.98\u20036--828.23.671.553.080.0021.61--8.39\u2003\u2009\\>\u2009874.810.493.317.45\\<\u20090.0015.65--19.46Nodules number (single/multiple)12.3/24.61.883142.48742332.450.0141.13--3.13Pre-LT TACE (yes/no)23.3/13.02.030.532.730.0061.22--3.39Pre-LT RFA (yes/no)13.6/18.00.790.290.650.5190.39--1.61Pre-LT hepatectomy (yes/no)25.0/15.51.590.491.520.1280.88--2.90Vascular invasion (yes/no)43.0/9.84.311.125.62\\<\u20090.0012.59--7.17Differentiation (Reference, well)\u2003moderate14.91.520.670.950.3440.64--3.59\u2003poor26.12.811.352.160.0311.10--7.19\u2003Milan criteria (out/in)33.8/4.87.682.785.64\\<\u20090.0013.78--15.62\u2003Hangzhou criteria (out/in)57.5/9.17.241.917.52\\<\u20090.0014.32--12.13\u2003AFP model (out/in)35.0/7.95.801.735.89\\<\u20090.0013.23--10.41Note: *BMI* Body mass index, *CTP* Child-Turcotte-Pugh, *AFP* Alpha-fetoprotein, *MELD* Model for End-Stage Liver Disease, *LT* Liver transplantation, *TACE* Transarterial chemoembolization, *RFA* Radiofrequency ablation\n\nDevelopment of the 5--8 score {#Sec8}\n-----------------------------\n\nPreoperative factors with *p*-value \\<\u20090.05 on univariable analysis were then used in the multivariable model. A Cox regression analysis was then performed with backward selection to conduct a multivariable analysis of clinicopathologic factors that associated significantly with post-LT HCC recurrence. Preoperative independent factors of post-LT HCC recurrence were utilized to construct the 5--8 score including (1) pre-LT AFP (at the following cut-off values: 10--200, 201--1000, and\u2009\\>\u20091000\u2009ng/mL), (2) the largest diameter of tumor (at the following cut-off values: 4--6,6.1--8 and\u2009\\>\u20098), and (3) number of nodules (single vs multiple). It is important to note that the model with the lowest (AIC), was chosen as the final model for the risk score. The multivariable HR of these factors derived from the Cox regression model were rounded to the nearest integer, then used to calculate the simplified 5--8 score. To calculate the score for each patient, the individual points for each of the three variables can be added together giving a minimum point of 0 and a maximum point of 24(Table\u00a0[3](#Tab3){ref-type=\"table\"}). Table 3Multivariate Cox analysis of risk factors for early recurrence of HCCVariableHRSEzP\u2009\\>\u2009z95% CI\u03b2 CoefficientPointsPre-LT AFP (ng/mL)\u2003\u2009\u2264\u2009101----------0\u200310--2003.121.512.290.0221.18--8.251.293\u2003200--10003.621.692.770.0061.46--9.051.144\u2003\u2009\\>\u200910006.222.923.89\\<\u20090.0012.48--15.621.836Largest tumor diameter (cm)\u2003\u2009\u2264\u200941----------0\u20034--62.170.832.030.0431.01--4.570.772\u20036--83.091.322.640.0081.34--7.161.133\u2003\u2009\\>\u2009812.824.437.39\\<\u20090.0016.52--25.242.5513Nodules number\u2003Single1----------0\u2003Multiple2.540.713.360.0011.48--4.380.9330--5: low risk, 6--8 medium risk, \\>\u20098 high risk\n\nPrediction of HCC recurrence risk by the 5--8 score {#Sec9}\n---------------------------------------------------\n\nBased on the 5--8 score, patients were then accurately stratified according to their risk of recurrence into three categories; the low-risk group had a score of 0 to 5, the medium-risk group had a score of 6--8, the high-risk group had a score of \\>\u20098(Table [3](#Tab3){ref-type=\"table\"}). The most common group being low risk group \\[*n*\u2009=\u2009253(52.1%)\\] then medium-risk group \\[*n*\u2009=\u2009129(26.5%)\\] and high-risk group (*n*\u2009=\u2009104(21.4%)\\]. The risk of 2-year HCC recurrence was increased significantly from low to high-risk group as shown by KM curves and log-rank test. The 2-year HCC recurrence rate was 4.5% (95% CI:0.02--0.09), 20.0% (95% CI:0.12--0.34) and 51.4% (95% CI:0.36--0.73) in the low, medium and high-risk group respectively (overall log-rank *p*\u2009\\<\u2009\u00a00.001))(Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a). The corresponding 2- year OS was 92.5% (95% CI: 0.86--0.96), 82.9%(95% CI:0.72--0.90) and 57.2% (95% CI: 0.42--0.70), respectively (overall log-rank *p*\u2009\\<\u20090.001) (Fig. [2](#Fig2){ref-type=\"fig\"}b). According to the 5--8 score, patients before transplantation with AFP level of \u226410\u2009ng/mL and single nodule with the largest tumor diameter of \\<\u20094\u2009cm on radiological assessment would be categorized in the low-risk group, in contrast to patients who have pre-LT AFP of \\>\u20091000(ng/mL) and largest tumor diameter of \\>\u20098\u2009cm, will be categorized in high-risk category. There was no deviation from the proportional hazard assumption according to the visual inspection of log-log survival curves and the Schoenfeld test (*p*\u2009=\u20090.708). Also, for prediction HCC recurrence in the training cohort, visual assessment of the KM curves showed good discrimination between the three 5--8 model prognostic subgroups. Moreover, the Harrell's C and Somers' D of 5-8score were 79%(95% CI:0.73--0.86) and 59%(95% CI:0.40--0.72) (Table\u00a0[4](#Tab4){ref-type=\"table\"}). Based on the competing risk analysis, the 2-year cumulative incidence of mortality, while controlling for the risk of HCC-related death, was 5.2%(95% CI: 0.02--0.11), 12.9%(95% CI: 0.06--0.22) and 35.0% (95% CI: 0.22--0.49, overall *p*\u2009\\<\u20090.001) in patients with low, medium and high 5--8 score (Fig. [2](#Fig2){ref-type=\"fig\"}c). Furthermore, 2- year cumulative incidence of HCC-unrelated death which not related to HCC recurrence were, 2.4%(95% CI: 0.01--0.05),4.2% (95% CI: 0.01--0.10) and 7.9%(95% CI: 0.03--0.16), (overall *p*\u2009=\u20090.120) in patients with low, medium and high 5-8score (Fig. [2](#Fig2){ref-type=\"fig\"}d). Fig. 2In the training cohort and according to the 5--8 model, the 2-year recurrence rates (**a**) and overall survival rates(**b**). The cumulative incidence of HCC-related deaths (**c**) and HCC-unrelated deaths(**d**) as assessed by the competing risk analysis Table 4Accuracy of the 5--8 model for predicting the risk of HCC early recurrence in the training and validation cohort compared with Milan, Hangzhou, and AFP modelTraining cohortValidation cohortHarrell's C (95% CI)Somer's D (95% CI)Harrell's C (95% CI)Somer's D (95% CI)5--8 model0.79 (0.73--0.86)0.59 (0.40--0.72)0.74 (0.66--0.82)0.49 (0.32--0.74)Milan0.72 (0.67--0.76)0.43 (0.35--0.58)0.65 (0.59--0.71)0.30 (0.17--0.45)Hangzhou0.72 (0.65--0.78)0.43 (0.31--0.61)0.61 (0.54--0.69)0.23 (0.07--0.40)AFP model0.72 (0.66--0.77)0.43 (0.33--0.60)0.68 (0.60--0.75)0.35 (0.20--0.53)\n\nComparison of the 5--8 model with Milan and Hangzhou criteria {#Sec10}\n-------------------------------------------------------------\n\nThe 2-years recurrence rate for patients meeting and exceeding Milan criteria, was 4.8 and 33.8% respectively (*p*\u2009\\<\u20090.001) (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a), while it was 9.1% vs 57.5% in patients meeting and exceeding Hangzhou criteria respectively, (*p*\u2009\\<\u20090.001) (Fig. [3](#Fig3){ref-type=\"fig\"}d). We further analyzed the risk-stratified patients of 5--8 score for the patients who were fulfilling and exceeding either Milan or Hangzhou criteria. Among 259 patients who were fulfilling Milan criteria (53.2%), the risk of 2-year HCC recurrence according to 5--8 score was 2.6% (95% CI: 0.01--0.07) in the low risk group, 8.5% (95% CI: 0.03--0.29) in medium- risk group and 23.6% (95%CI: 0.06--0.95) in the high-risk group (overall *p*\u2009=\u20090.009) (Fig. [3](#Fig3){ref-type=\"fig\"}b). While for 227 patients who are exceeding Milan criteria (46.7%), the risk of 2-year HCC recurrence was 11.5% (95% CI: 0.05--0.28),29.0% (95% CI:0.16--0.51) and 55.0%(95% CI: 0.38--0.79), respectively (overall *p*\u2009\\<\u20090.001) (Fig. [3](#Fig3){ref-type=\"fig\"}c). Likewise, for the 390 patients within Hangzhou criteria (80.3%), the risk of 2-year HCC recurrence according to 5--8 score was 3.0% (95% CI:0.01--0.07) in the low risk group, 16.3% (95% CI:0.09--0.31) in medium- risk group and 24.6% (95% CI: 0.12--0.49) in the high-risk group (overall *p*\u2009\\<\u20090.009) (Fig. [3](#Fig3){ref-type=\"fig\"}e). While for 96 patients who are exceeding Milan criteria (19.8%), the risk of 2-year HCC recurrence was 26.2% (95% CI: 0.08--0.82), 37.2%(95% CI: 0.16--0.87) and 79.3% (95% CI:0.52--1.21), respectively (overall *p*\u2009\\<\u20090.001) (Fig. [3](#Fig3){ref-type=\"fig\"}f). For prediction of HCC recurrence, Harrell's C and Somers' D of 5-8score were 79%(95% CI:0.73--0.86) and 59%(95% CI:0.40--0.72) in the training cohort compared with 72%(95% CI:0.67--0.76) and 43%(95% CI:0.35--0.58) for Milan criteria and 72%(95% CI:0.65--0.78) and 43%(95% CI:0.31--0.61) for Hangzhou criteria (Table [4](#Tab4){ref-type=\"table\"}). Fig. 3In the training cohort, 2-year recurrence rates according to Milan criteria (**a**), according to 5--8 model in patients fulfilling Milan criteria (**b**) and in patients exceeding Milan criteria (**c**). Two-year recurrence rates according to Hangzhou criteria (**d**), according to the 5--8 model in patients fulfilling Hangzhou criteria (**e**) and in patients exceeding Hangzhou criteria (**f**)\n\nComparison of the 5--8 model with AFP model {#Sec11}\n-------------------------------------------\n\nFor patients exceeding and fulfilling the AFP model, the 2-year rates of HCC recurrence were 35.0 and 7.9%, respectively (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}a). Further analysis for the risk-stratified patients of 5--8 score for the patients who were within and outside the AFP model showed that among 301 patients who were within AFP model (62%), the risk of 2-year HCC recurrence was 4.0% (95% CI: 0.02--0.09) in the low-risk group, 19.6% (95% CI: 0.09--0.43) in medium- risk group (*p*\u2009=\u20090.006) (Fig. [4](#Fig4){ref-type=\"fig\"}b). While for 185 patients who are exceeding AFP model (38.1%), the risk of 2-year HCC recurrence was 9.8% (95% CI: 0.02--0.39),20.8% (95% CI: 0.11--0.41), and 51.4%(95% CI: 0.36--0.73), respectively (overall *p*\u2009=\u20090.006) (Fig. [4](#Fig4){ref-type=\"fig\"}c). Fig. 4In the training cohort, the 2-year recurrence rates according to AFP model (**a**), according to the 5--8 model for patients within the AFP model (**b**) and in patients outside AFP model (**c**). In the validation cohort, the 2-year recurrence rates (**d**) according to the 5--8 model\n\nValidation of the 5--8 model {#Sec12}\n----------------------------\n\nAs mentioned above, there were no significant differences in baseline characteristics among the training and validation cohort. In the validation cohort, the median post-LT follow-up was 416.5\u2009days \\[IQR:205--672\\]. The 5--8 model could also accurately stratify patients into low, medium and high-risk prognostic subgroups with 2-year HCC recurrence rates of 10.8% (95% CI: 0.06--0.21), 31.7% (95% CI: 0.17--0.59) and 62.4% (95% CI: 0.39--1.00) respectively, (overall log-rank *p*\u2009\\<\u20090.001) (Fig. [4](#Fig4){ref-type=\"fig\"}d). The 5--8 score also achieved a good performance in post-LT recurrence risk prediction in the validation cohort according to the visual assessment of the resulting KM curves and the Harrell's C and Somers' D of 74%(95% CI:0.66--0.82) and 49%(95% CI:0.32--0.74) compared with 65%(95% CI:0.59--0.71) and 30%(95% CI: 0.17--0.45) for Milan criteria and 61%(95% CI:0.54--0.69), 23%(95% CI:0.07--0.40) for Hangzhou criteria (Table [4](#Tab4){ref-type=\"table\"}).\n\nDiscussion {#Sec13}\n==========\n\nRecurrence of HCC after LT remains a major obstacle and associated with an unfavorable prognosis \\[[@CR20]\\]. Several independent risk factors for post-LT HCC recurrence have been identified including number and size of tumors on preoperative imaging studies \\[[@CR11], [@CR12], [@CR21]\\] and pre-LT serum levels of AFP at different cut-off points: 10,200,1000\u2009ng/mL \\[[@CR22]--[@CR25]\\]. However, lack of agreement about accurate, reliable and robust validated tool especially for prediction of early recurrence of HCC in HBV predominant population, make appropriate risk stratification and doctor-patient communication challenging.\n\nIn the present study, 786 patients with HCC diagnosed by imaging who underwent deceased donor LT from centers distributed throughout the whole China were involved and the 5--8 score was developed and validated. Our predictive model is a simple and reliable tool that showed excellent stratification of HCC patients into three risk subgroups; low-, medium- and high-risk with predicted 2-year recurrence ranging from 5% in low-risk to 20.0% in the medium risk and 51% in the high-risk category. For the 253 patients (52.1%) in the low-risk group, the 2- year OS was 93%, significantly superior to the 83 and 57% of the medium- and high-risk groups.\n\nSimilar to our study, previous risk models \\[[@CR16], [@CR26]--[@CR28]\\] have attempted to employ pre-LT AFP, largest tumor diameter and the number of nodules to predict the post-LT HCC recurrence, but the characteristics of their study population (western participants with predominantly HCV) was different. Our model is mainly based on the data of patients with HCC occurring in patients with HBV infection (96%) which is the most common cause of HCC in China. The high predictive accuracy of our preoperative model \\[Harrell's C was 79%(95% CI: 0.73--0.86) and 74%(95% CI:0.66--0.82), respectively, in the training and validation cohorts\\] stems from the utilization of three preoperative factors with more accurate cut-off points that were strongly associated with HCC recurrence: pre-LT AFP(10--200, 201--1000, and\u2009\\>\u20091000\u2009ng/mL), largest tumor diameter(4--6,6.1--8,\\>\u20098\u2009cm), and the number of nodules (single vs multiple). Unlike previous studies, our model end-point was recurrence at 2-year which is the discriminative cut-off value for early and late recurrence of HCC, however, the precise cut-off may require genetic/genomic analyses recurrence. Early recurrence results from preexisting tumor cells while late recurrence or de novo tumor mainly arising due to new malignant clones \\[[@CR29]--[@CR31]\\]. Using the cut-off values in our study was a clinical decision based on the expert's assessment and they achieved the lowest AIC values so better fit the model.\n\nFurthermore, one of the main advantage of our score over previous mentioned models is its ability to discriminate three subgroups of HCC recurrence(i.e low,medium,high), while other models only classify HCC patients into two risk subgroups of HCC recurrence, high and low risk. The categorization defining only two subgroups at risk is not valuable, or at least less practical. For example, tumour recurrence risk divided into low (\\<\u20098%) and high (\\>\u200950%) will produces a big 'grey zone' through 8--50% of medium-risk individuals \\[[@CR32]\\]. Our score solved this drawback by stratify the individuals at risk into three subgroups of risk of recurrence, low,medium and high. This can be translated clinically into an excellence of individuals selections for LT, a more reasonable organs allocation taking into account the donor offer, besides, and the opportunity to stratify risk for the development of upcoming adjuvant treatments \\[[@CR33]\\].\n\nOne of the most interesting results comes when we compared our model with the conventional selection criteria by which our model can stratify patients within and outside the Milan and Hangzhou criteria. For instance, the subgroup of patients (25%) who were recognized as high risk of recurrence by Milan (out Milan), they carry a low risk according to our score with 2-year recurrence probability of 12%. While the subgroup of patients(4%) who were recognized as low risk of recurrence by Milan (in Milan), they carry a high risk according to our score with 2-year recurrence probability of 24%. Similarly, the subgroup of patients (12%) who were recognized as low risk of recurrence by Hangzhou (in Hangzhou), they carry a high risk according to our score with 2-year recurrence probability of 25%. The addition of our model to the conventional selection criteria may, therefore, allow us to capture accurately the patients with a higher risk of recurrence who were traditionally considered as the lowest risk group. External validation of this preoperative adjuvant model is mandatory to help in more accurate selection of HCC patients for LT and to avoid the high post-LT recurrence probabilities.\n\nMoreover, one of the advantages of our model is when it was compared to the AFP model, about (11 and 32%) of patients recognized as high risk of recurrence by AFP model, but they had a low and medium risk according to our score with 2-year recurrence probability of 10, 21% respectively. Also, about 23% of the patients were defined as a low risk of recurrence by the AFP model, but a medium risk of recurrence was revealed by our score with 2-year recurrence probability of 20%. Another advantage of our model is the outcome of the competing risk analysis, the rates of HCC-unrelated death were similar (*p*\u2009=\u20090.120) and the rates of HCC-related death which mainly due to HCC recurrence were significantly different (*p*\u2009\\<\u20090.001), indicating that the 5--8 model is a powerful tool to pick up the HCC recurrence but not other causes of deaths and this explains the differences in survival rates according to the 5--8 model.\n\nAt this point, we need a further explanation for the clinical application of our model. The patients presented preoperatively with a single nodule, diameter of \u22644\u2009cm and serum AFP of \u226410\u2009ng/mL will fall in the low-risk group by getting the 5--8 score of 0 and this will effectively exclude the probability of post-LT HCC recurrence, so these patients can go directly for LT. However,The patients with largest nodule diameter of \\>\u20098\u2009cm and serum AFP of \\>\u20091000\u2009ng/mL will be categorized in the high-risk category (Table [3](#Tab3){ref-type=\"table\"}), so LT should be excluded or neoadjuvant therapy and close surveillance until the tumor biology and morphology could be brought down to a safer level. However, patients who belong to the medium HCC recurrence risk category (e.g. single nodule, diameter of 8\u2009cm and serum AFP of 700\u2009ng/mL), a careful selection for LT based on a personalized assessment and pre-LT downstaging therapy should be considered then early administration of mTOR inhibitor postoperatively are strongly recommended \\[[@CR34]\\].\n\nOur study has some limitations. First, its retrospective nature so we designed a prospective study to confirm the clinical utility of our model. Second, the characteristics of our study population (predominantly HBV infected male Chinese patients) so external validation in HCV predominant, non-Chinese populations is required to confirm the reliability of the 5--8 model. Third, the patients with vascular invasion were not excluded from our study. Fourth,some parameters, such as microvascular invasion and data regarding response to pre-LT neoadjuvant therapies are incomplete in the CLTR database and this precludes us from comparing our model with other prominent risk scores. Lastly, although the follow-up time was enough to pick up the events, it is relatively short in comparison with other studies.\n\nConclusions {#Sec14}\n===========\n\nIn conclusion, a preoperative predictive risk score for early recurrence of HCC after LT was constructed using 3 variables (pre-LT AFP, largest tumor diameter, and the number of nodules). Our model accurately predicts early recurrence of HCC at an individual level in patients with HBV-cirrhosis. It could effectively classify HCC patients into subgroups with a low, medium and high risk of recurrence and can potentially be used to guide therapeutic decisions and facilitate risk communication between surgeons and patients. We believe that our model can play a complementary role in the selection of HCC patients for LT and prediction of early recurrence alongside the conventional selection criteria. Prospective validation of our model will be an important step to verify its clinical utility.\n\nAFP\n\n:   Alpha-fetoprotein\n\nBMI\n\n:   Body mass index\n\nCI\n\n:   Confidence interval\n\nCLTR\n\n:   China Liver Transplant Registry\n\nCVA\n\n:   Cerebrovascular accident\n\nDBCD\n\n:   Donation after brain death followed by circulatory death\n\nDBD\n\n:   Donation after brainstem death\n\nDCD\n\n:   Donation after circulatory death\n\nHBV\n\n:   Hepatitis B virus\n\nHCC\n\n:   Hepatocellular carcinoma\n\nHR\n\n:   Hazard ratio\n\nIQR\n\n:   Interquartile range\n\nLT\n\n:   Liver transplantation\n\nMELD\n\n:   Model for End-Stage Liver Disease\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nWe would like to thank China Liver Transplantation Registry (CLTR) for their help in the data acquisition and extraction. We also extend our thanks to the transplant centers from Mainland China for their major contribution to the CLTR database.\n\nAA, XX, SZ designed the study, HX, LZ, XX, and SZ give the administrative support, XW, LL, ZS, and HG provide the study materials and collect the data, AA and XX analyse the data and interpret the results. All authors read and approved the final manuscript.\n\nThis work supported by National S&T Major Project (Grant number:2017ZX10203205); Medical Science and Technology Project of Zhejiang Province (Grant number:2016146968) and China Postdoctoral Science Foundation (Grant number:2017\u2009M612014). The above funding bodies had no further role in the study design; in the collection, analysis and interpretation of data; in the writing of the article or in the decision to submit the article for publication.\n\nThe data that support the findings of this study are available from the China Liver Transplant Registry but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the China Liver Transplant Registry.\n\nThis study approved by the Scientific Committee of the China Liver Transplant Registry (<http://www.cltr.org/>) which is in accordance with ethical guidelines of Helsinki Declaration 1975, as revised in 2013. Every participant had received a written informed consent before being included in this study.\n\nNot applicable.\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "More than 170 years after the discovery of skeletal remains of giant vertebrates preserved in Australian caves[@b1], and later in dry lakebeds[@b2], the cause of their demise is still debated. Most scientific studies are directed at evaluating the two dominant explanations for extinction: human impact or climate change[@b3]. As dating methods have improved, chronologies for human dispersal across Australia and for the last surviving megafauna have become more accurate and more precise[@b4][@b5][@b6]. Although the date of initial human arrival on mainland Australia remains uncertain, populations were established over most of the continent by at least 47\u2009ka (refs [@b6], [@b7]). Although many elements of the Australian megafauna (those animals \\>45\u2009kg body mass) lack firm extinction timelines, last appearance dates for taxa that occur most frequently in the fossil record are between 50 and 40\u2009ka (refs [@b4], [@b5], [@b8]), indicating a temporal overlap between humans and megafauna[@b9]. The climate of Australia was drying gradually between 60 and 40\u2009ka, but neither the rate nor magnitude of change were more severe than during earlier Pleistocene climate shifts[@b10][@b11]. The lack of evidence for unprecedented climate change between 60 and 40\u2009ka and survival of megafauna during earlier more extreme climate fluctuations[@b12] implies that climate change is unlikely to be the sole cause of megafaunal extinction, leaving human agency more likely to have been the decisive factor, with modest additional stress from increasing aridity potentially a contributing factor[@b13][@b14]. Human colonizers may stress megafaunal populations by predation or by altering habitats so that dietary resources are reduced, with recent studies emphasizing the impacts of even modest hunting, especially for long-lived large vertebrates with low reproductive rates[@b15][@b16]. Although the argument for human predation contributing to widespread extinction in Australia is on firm grounds from ecological theory, little supporting evidence is available in the form of kill sites or unambiguous human modification of skeletal remains despite 50 years of systematic archaeological fieldwork. This lack of evidence has been used to argue against human predation as a cause of megafaunal extinction[@b17]. The absence of kill sites in Australia compared with the Americas has been postulated to be a consequence of the much earlier date for Australian extinctions, diminishing the probability of their preservation[@b4][@b18][@b19]. Here we provide direct evidence from sites across most of the continent that humans preyed on at least one element of the Australian megafauna, harvesting eggs of *Genyornis newtoni*, an extinct, 200\u2009kg flightless bird, and leaving diagnostic burnt eggshell fragments as evidence of their activities. In the largest series of dates collated for any Australian megafaunal species, we demonstrate that burnt *Genyornis* eggshell only occur during a window between 53.9 and 43.4\u2009ka, and likely before 47\u2009ka, documenting megafaunal predation by humans as they dispersed across the continent.\n\nResults\n=======\n\n*Genyornis* eggshell\n--------------------\n\nFragments of *Genyornis* eggshell are found in recently deflated sand dunes where the birds nested, with morphological features differentiating *Genyornis* eggshell from those of *Dromaius* (emu)[@b20]. We analysed *Genyornis* eggshell from nearly 2,000 localities across Australia; none is clearly younger than 50\u00b15\u2009ka, whereas *Dromaius* eggshell are commonly found in the same regions from \\>100\u2009ka to the present[@b8]. Field studies in ten regions across the continent ([Fig. 1](#f1){ref-type=\"fig\"}) yielded \\>200 collections that contain variably blackened *Genyornis* eggshell, frequently blackened at only one end of the fragment, suggestive of irregular heating patterns.\n\nDiagnostic patterned burning\n----------------------------\n\nTo test whether blackened eggshell is diagnostic of high temperatures, we sub-sampled partially blackened fragments for amino-acid analysis. We found that the blackened ends had been heated sufficiently to decompose all amino acids, with decreasing decomposition away from the burnt end, eventually reaching amino-acid concentrations similar to those in unburnt fragments ([Fig. 2](#f2){ref-type=\"fig\"}). Such a strong decomposition gradient can only be accomplished if the blackened end was briefly in contact with a localized high-heat-source, likely an ember; rapid amino-acid decomposition requires temperatures \u2265500\u2009\u00b0C (ref. [@b21]). Graded amino-acid decomposition is apparent in transects across other variably blackened fragments, whereas fully blackened eggshell are devoid of amino acids, consistent with sustained high heat over the entire fragment ([Supplementary Figs 1 and 2](#S1){ref-type=\"supplementary-material\"}).\n\nIn most collections with burnt eggshell, we found examples of partially and wholly burnt fragments, some burnt so severely that their natural curvature was flattened, and occasionally reversed, as well as many visually unburnt fragments. These collections often form a tight cluster \\<3\u2009m in diameter without other eggshell nearby, suggesting the source was a single, or small number of eggs. We conclude that natural wildfire could not produce such steep thermal gradients within and between nearby eggshell fragments, as it requires an untenable combination of circumstances. Rather, these characteristics are most consistent with humans harvesting one or more eggs from a nest, making a fire and presumably cooking the egg. For the same reason that it is possible to boil water in a paper cup over a fire without burning the cup, cooking an egg in a manner that does not cause the egg to explode, will not char the eggshell. Records of traditional Aboriginal cooking of emu eggs describe a relatively slow cooking of the eggs, either wrapped in vegetation or in hot ashes in a hole dug in the ground for that purpose, from which the egg would be removed and rotated or shaken frequently, then repositioned[@b22]. After cooking, we presume that eggshell fragments were discarded randomly in and around the fire. We find similarly burnt *Dromaius* eggshell in unambiguous late Holocene archaeological contexts along with other burnt food debris ([Supplementary Fig. 3](#S1){ref-type=\"supplementary-material\"}), and burnt ostrich eggshell is found in similar archaeological settings in Africa[@b23][@b24].\n\nBurnt *Genyornis* eggshell fragments are most common in coastal sand dunes of Western Australia (WA). Of 567 collections from four WA regions that contained *Genyornis* eggshell, 192 included burnt fragments, with much smaller proportions in the Darling (1 of 189 collections) and around Lake Eyre, the driest sector of the continent (1 of 542 collections; [Supplementary Data 1](#S1){ref-type=\"supplementary-material\"}). Although the Willandra (A, GA) and Spencer Gulf (WL, PB) regions each had few collections, burnt fragments occurred relatively frequently (2 of 6 and 10 of 23, respectively; [Supplementary Fig. 5](#S1){ref-type=\"supplementary-material\"}).\n\nDating burnt eggshell\n---------------------\n\nTo evaluate whether burnt *Genyornis* eggshell only coincide with a human presence in the landscape, we obtained absolute and relative dates using optically stimulated luminescence (OSL), radiocarbon (^14^C) and amino-acid racemization (AAR). Seven sites with burnt *Genyornis* eggshell in a stratigraphic context and eight other stratified sites with unburnt *Genyornis* eggshell have been dated by OSL ([Table 1](#t1){ref-type=\"table\"}). The cumulative sum of the individual OSL ages and their uncertainties for collections with burnt eggshell yield an aggregate median age of 47.5\u2009ka and a range from 53.9 to 43.4\u2009ka ([Fig. 3a](#f3){ref-type=\"fig\"}), where the upper and lower bounds correspond to the 16th and 84th percentiles of the aggregate distribution (roughly equivalent to \u00b11\u03c3 in a normal distribution). Eight other collections with *Genyornis* eggshell that lack burnt fragments, but for which AAR indicates they are among the youngest in each region, have a median OSL age of 51.5\u00b14.9\u2009ka ([Table 1](#t1){ref-type=\"table\"}).\n\n*Genyornis* eggshell were dated directly by ^14^C using accelerator mass spectrometry following rigorous pretreatment. However, as in other carbonate media from terrestrial settings, eggshell calcite is subject to slow diffusion of younger carbon from its surroundings. For samples older than 40\u2009ka, small amounts of young carbon will result in apparent ages significantly younger than their true age[@b25][@b26]. For example, we obtained finite ^14^C ages \\<46\u2009ka on *Genyornis* eggshell from two collections dated \u226570\u2009ka by OSL and/or AAR ([Table 1](#t1){ref-type=\"table\"}), despite \\>50\u2009ka background dates in standards. This demonstration of exchange with younger carbon suggests that all ^14^C dates \\>40\u2009ka should be regarded as minimum ages. Unburnt *Genyornis* eggshell from 13 collections with burnt fragments, including all regions in [Fig. 1](#f1){ref-type=\"fig\"}, were dated by ^14^C ([Table 1](#t1){ref-type=\"table\"}). Calibrated ^14^C ages for two of the Warroora sites are significantly younger than their corresponding OSL ages (no overlap at \u00b11\u03c3; Table 1) and are considered anomalously young as a result of carbon exchange. For the remaining 11 samples, all but 1 have minimum calibrated ^14^C ages \u226544\u2009ka. Considered collectively, calibrated \u00b11\u03c3 age ranges for the nine finite and two non-finite ages constrain a likely minimum calendar age for burnt *Genyornis* to 47.5\u2009ka ([Fig. 3b](#f3){ref-type=\"fig\"}), although somewhat younger ages cannot be conclusively ruled out. These dates refine earlier estimates for *Genyornis* extinction of 50\u00b15\u2009ka (ref. [@b8]) to 47.5\u00b12.5\u2009ka.\n\nEggshell relative age is constrained by AAR in intracrystalline protein residues[@b27] isolated from physically and chemically cleaned *Genyornis* eggshell (*n*=3,877)[@b28]. The amino acid isoleucine epimerizes to its non-protein diastereomer alloisoleucine at a rate dependent on temperature, with their ratio (A/I) reflecting time and the effective diagenetic temperature for each sample. For collections buried \u22652\u2009m and not subjected to any other heat source, effective diagenetic temperature is set by the integrated mean annual temperature (MAT) since the egg was laid[@b28]. However, the extra energy imparted to fire-heated eggshell accelerates racemization. To minimize this effect, we select only visually unburnt fragments for AAR analysis, and we analyse multiple fragments from each collection containing burnt fragments. Yet, even visually unburnt fragments often exhibit variably high A/I values, presumably because they were heated at temperatures sufficient to accelerate racemization but too low to create black carbon. Cooking temperatures \u2264100\u2009\u00b0C for \\<1\u2009h will not blacken eggshell or significantly raise A/I, whereas brief transient heating \\>150\u2009\u00b0C but well below 500\u2009\u00b0C accelerates racemization without blackening the eggshell. To prevent unblackened but fire-heated samples from biasing our interpretations, we screen our results, limiting summary plots to those collections for which the two (or more) lowest A/I differ by no more than 0.02\u2009A/I units, thereby excluding heated (higher) A/I values; for a few collections, we used the lowest A/I if that was the only analysis within 1 \u03c3 of the region\\'s mean A/I. We measured A/I in over 550 *Genyornis* eggshell fragments from 84 collections that also contain burnt fragments. In all, 63 of the 84 collections met our screening criteria, from which we compute regionally averaged A/I ([Supplementary Data 2](#S1){ref-type=\"supplementary-material\"}), and compare those to their corresponding regional MAT in [Fig. 4](#f4){ref-type=\"fig\"}. The close approximation to a simple second-order polynomial regression (*r*^2^=0.99) is consistent with the exponential dependence of racemization rate on temperature, based on kinetics derived in ref. [@b28], and the high correlation coefficient is consistent with a similar age for all 63 collections across all regions ([Fig. 1](#f1){ref-type=\"fig\"}). Age differences between regions \\>5\u2009ka would in almost all instances significantly lower the correlation coefficient. However, the exact form of the trend line cannot be predicted *a priori* because of uncertainties in the magnitude of the glacial-age temperature depression for each region. Consequently, the A/I--MAT relation is a necessary, but not sufficient condition to confirm that all collections of burnt *Genyornis* eggshell are of the same age.\n\nTemporal distribution of burnt *Dromaius* and *Genyornis* eggshell\n------------------------------------------------------------------\n\nTo further test whether burnt eggshell is diagnostic of human predation, we utilize the temporal distribution of collections containing burnt *Genyornis* and *Dromaius* eggshell derived from AAR analyses. If human predation is the sole cause of variably burnt *Genyornis* eggshell fragments, then similarly burnt *Dromaius* eggshell should first appear in the record \u223c50\u2009ka, occur continuously to the present, but never occur before human arrival. In our WA collections, where burnt eggshell of both taxa are most common, burnt *Dromaius* eggshell first appear in sites with AAR indicative of 50\u00b15\u2009ka, indistinguishable within stated uncertainties to the dates associated with burnt *Genyornis* eggshell, remain frequent in collections through to near-modern time, but are not present in any collection \\>55\u2009ka in the four WA regions ([Fig. 5a](#f5){ref-type=\"fig\"}) or in any of the other regions. Similarly, none of the *Genyornis* eggshell collections from WA that predate 50\u00b15\u2009ka contain burnt fragments ([Fig. 5b](#f5){ref-type=\"fig\"}); the increasing percentage of burnt *Genyornis* eggshell in the lowest three A/I bins is consistent with human predation leading to *Genyornis* extinction.\n\nDiscussion\n==========\n\nWe found no *in situ* hearths or *in situ* stone artefacts directly associated with burnt *Genyornis* eggshell, or with similar-age burnt *Dromaius* eggshell, and only rarely in association with pre-Holocene, post-45\u2009ka burnt *Dromaius* eggshell. This is expected because the alkaline dune sediments that preserve eggshell carbonate also degrade charcoal[@b29], and transient cooking fires in a sandy substrate leave no baked clays. Few regions provided *Genyornis* eggshell in stratigraphic sections where their association with artefacts could be securely evaluated; most samples were collected from the floors of deflation hollows. Lithic artefacts (commonly) and hearthstones (occasionally) are found in deflation hollows among surface scatters of burnt *Genyornis* eggshell, but because both occur as deflationary lags, temporal association cannot be demonstrated. However, the presence of hearthstones confirms that fire-using humans were in the same landscape, despite the lack of preserved hearths. At Garnpung ([Fig. 1](#f1){ref-type=\"fig\"}), one of the few sites with burnt *Genyornis* exposed in a stratigraphic section, an *in situ* hearthstone stratigraphically below the horizon containing burnt *Genyornis* eggshell ([Supplementary Fig. 4](#S1){ref-type=\"supplementary-material\"}) demonstrates a temporal overlap with humans.\n\nThe time interval during which *Genyornis* became extinct (50\u00b15\u2009ka)[@b8] coincides with the interval when humans were consuming its eggs (53.9 to 43.4\u2009ka), suggesting that predation contributed to the bird\\'s extinction. This is also the same interval when the dietary intake of *Dromaius* underwent a dramatic reduction in the proportion of C4 grasses, with C4 dietary elements remaining reduced through to the present[@b30]. An explanation for the sudden shift in *Dromaius* diet remains obscure, but the loss of palatable C4 grasses across the arid zone would have placed additional stress on *Genyornis* survival, as their diet always included some C4 grass elements[@b30].\n\nOur interpretation of burnt eggshell places greatest reliance on the chronological patterns. Burnt *Genyornis* eggshell only occur in a narrow temporal window between 54 and 43\u2009ka defined by dated collections containing burnt eggshell across the arid zone, and this window coincides with both the extinction of the species[@b8] and the dispersal of people across Australia[@b6][@b7]. Furthermore, the oldest similarly burnt *Dromaius* eggshell are dated to the same time window, persist to contemporary time, but are absent from our extensive collections dated between 55\u2009ka and \\>100\u2009ka. The range of burn patterns found in clusters of *Genyornis* eggshell is most consistent with humans scattering eggshell fragments of consumed eggs in and around transient cooking fires, and the strong thermal gradients required to explain the observed burn patterns are incompatible with a wildfire cause. OSL dating limits the timing of *Genyornis* egg predation to no older than 53.9\u2009ka, or younger than 43.4\u2009ka, and likely no younger than 47\u2009ka, based on the ^14^C dates (see also [Supplementary Data 3](#S1){ref-type=\"supplementary-material\"} for details). AAR correlations expand to 63 the number of dated collections from ten regions across the continent with burnt *Genyornis* eggshell. Our data provide compelling evidence that humans not only dispersed rapidly across Australia\\'s well-watered landscapes, but also deep into its arid interior at or before \u223c47\u2009ka, preying on at least one element of the megafauna, the eggs of the giant bird, *Genyornis newtoni.* We hypothesize that human predation on *Genyornis* eggs likely contributed to the birds\\' extinction, with the harvesting of their eggs decreasing *Genyornis* reproductive success. Predation, combined with widespread changes in ecosystem composition throughout its range[@b30], very likely caused *Genyornis* extinction by \u223c47\u2009ka.\n\nMethods\n=======\n\nAmino-acid analyses\n-------------------\n\nEggshell are mechanically cleaned by grinding to remove surface impurities and the outer portion of each eggshell. For *Dromaius* eggshell, which has a tripartite structure, the outer two layers are mechanically removed. After grinding, an additional 33% of the remaining eggshell mass is removed by the stoichiometric addition of 2\u2009N HCl, with the reaction driven to completion *in vacuo*. Cold 7\u2009N HCl (spiked with the non-protein amino-acid norleucine (6.25 \u00d7 10^\u22125^ mol l^\u22121^, to enable absolute concentrations for each amino acid to be determined) is added to a \u223c15\u2009mg subsample of the cleaned fragment in proportion to sample mass. The resultant solution is flushed with N~2~, sealed and heated at 110\u2009\u00b0C for 22\u2009h to hydrolyse protein residues. Amino acids are separated by automated high-performance liquid chromatography (Agilent 1100/1200 HPLC) utilizing ion-exchange and post-column derivitization with o-phthalaldehyde. Amino-acid concentrations are derived by comparing peak areas of individual amino acids to the area of the norleucine spike. The proportion of [D]{.smallcaps}-alloisoleucine to [L]{.smallcaps}-isoleucine (A/I) is based on peak--height ratios. All samples are analysed at least twice. A natural standard (ILC-G[@b31]) is analysed daily to monitor instrumental precision.\n\nOptically stimulated luminescence\n---------------------------------\n\nThe dated sites are well bleached, finely stratified aeolian deposits lacking any visible evidence of post-depositional bioturbation or other mixing. Sediment samples for OSL dating were collected by driving a 7 \u00d7 25\u2009cm^2^ stainless steel tube horizontally into a cleaned vertical face stratigraphically related to levels with burnt *Genyornis* eggshell. Bulk sediment from 30\u2009cm diameter around the sampled site was collected for U, Th and K concentrations, which were measured by neutron activation analysis and delayed neutron activation (NAA/DNA); K was calculated from measurements of K~2~O by X-ray fluorescence. Radioisotope activities for U, Th and K were also measured by high-resolution gamma spectrometry and subsequently converted to concentrations; these data confirmed secular equilibrium in the U and Th decay chains. Cosmic ray dose rates were calculated using the data of ref. [@b32], making allowance for site altitude, geomagnetic latitude and time-averaged thickness of sediment overburden. Alpha-particle irradiation from radioisotopes within the etched quartz grains was assumed to be 10% of the external activity, and the efficiency with which alpha-particle irradiation induced OSL (*a*) was assumed to be 0.05\u00b10.02. Long-term water content was estimated from measured values and reconstructions of the landscape history and topographic position, with uncertainties sufficient to accommodate all likely possibilities.\n\nIn the laboratory, 90--125\u2009\u03bcm quartz grains were isolated from each sediment sample under low-intensity red and orange light. OSL measurements were performed on \u223c5--6\u2009mg of etched quartz attached by silicone oil to the central 7\u2009mm diameter of each of 128 stainless steel discs. The OSL signal was measured on an Elsec Type 9010 automated reader with 500\u00b180\u2009nm stimulation, and ultraviolet emissions detected by an EMI 9235QA photomultiplier tube optically filtered by one UG 11 and one U-340 filter and P determined by the 'Australian slide\\' using a linear plus single saturating exponential fit (scale factor=1.00). More complete details are in ref. [@b33]. The advantage of multiple-grain OSL over single-grain OSL for these stratified low-dose-rate sediments is that the use of many grains effectively eliminates small-scale dose heterogeneity by averaging out the slight grain-to-grain differences in beta dose.\n\nRadiocarbon dating\n------------------\n\nAfter mechanical cleaning to remove all secondary carbonate, eggshell mass was further reduced by stoichiometric addition of 2\u2009N HCl to remove 75% of the remaining eggshell mass. CO~2~ was evolved from the cleaned eggshell fragment, purified and converted to graphite at the Laboratory for AMS Radiocarbon Preparation and Research, University of Colorado Boulder, and measured at the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory, University of California Irvine. Conventional radiocarbon dates were calibrated using Calib 7.1 and SHcal13 (ref. [@b34]).\n\nAdditional information\n======================\n\n**How to cite this article:** Miller, G. *et al.* Human predation contributed to the extinction of the Australian megafaunal bird *Genyornis newtoni* \u223c47\u2009ka. *Nat. Commun.* 7:10496 doi: 10.1038/ncomms10496 (2016).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nSupplementary Figures 1-5 and Supplementary References\n\n###### Supplementary Data 1\n\nField ID, Region (Fig. 1), and location of all collections containing burnt *Genyornis* eggshell.\n\n###### Supplementary Data 2\n\nExtent of isoleucine epimerization (A/I) for each collection containing burnt *Genyornis* eggshell that has been analyzed from the ten regions shown in Fig. 1. In most collections some fragments were heated sufficiently to accelerate racemization, but not hot enough to blacken the eggshell. To minimize the bias introduced by unrepresentative A/I we limited the averaging to samples with at least two fragments that difference by no than 0.02 A/I units above the lowest A/I in the collection. For a few collections we used the lowest A/I if that was the only analysis within 1 \u03c3 of the region\\'s mean A/I. Collections for which there were no samples within 0.02 A/I units of the lowest, or within 1 \u03c3 of the region\\'s mean are considered \\\"unresolved\\\" and were not utilized for any statistical analyses (\\\"1\\\" in the \\\"unresolved\\\" column). Shading is used to aid in differentiating different collections within a region. Samples that were so heated that all amino acids had decomposed are given an A/I of 99999.\n\n###### Supplementary Data 3\n\nDetails of sites from which multiple dating methods have been applied. \\*Youngest of 9 OSL dates from the Genyornis-bearing unit at this site (supplementary ref. 3)\n\nThis study was supported by the Archaeology and Archaeometry Program, and the Geosciences Directorate of the US National Science Foundation (BCS-0914821, ATM-0502632, ATM-0082254), and the Australian Research Council (F00103660, A00102515). S. Webb (Bond University) collected the burnt *Genyornis* samples from Lake Eyre and participated in field seasons. The data reported in this paper are tabulated in the Supplementary Materials and Methods and archived in the NOAA paleoclimate database. We thank the generosity of station owners/managers and Aboriginal elders for assistance in the field and for permission to access and collect on their lands. M. Bunce, J. Dortch and R. Gillespie provided helpful comments on earlier drafts. We thank J. Hellstrom (Melbourne University, Australia) and R. Chen (NTU, Taiwan, China) for ongoing research using Uranium decay to date eggshell, and D.G. Questiaux for OSL data analysis.\n\n**Author contributions** G.M. and J.M. initiated the study, with contributions from all authors. G.M., J.M., A.B., M.S., M.F., C.F., H.J., D.W., P.C. and R.H. were responsible for collecting samples, G.M., N.S., S.L. and S.D. provided the geochronology.\n\n![Regions across Australia from which diagnostic burnt *Genyornis* eggshell were collected.\\\nMap of Australia showing the ten regions from which burnt *Genyornis* eggshell fragments have been collected (solid circles) and one other region where other key samples were collected (open circle). A: Arumpo Station, NSW; GA: Garnpung Station, NSW; D: Lower Darling River, NSW; WL: Wallaroo, SA; PB: Port Broughton, SA; P: Port Augusta region, SA; E: Lake Eyre, SA; Q: Sites on and around Quobba Station, WA; GN: Sites on and around Gnaraloo Station, WA; W: Sites on and around Warroora Station, WA; B: Bullara, Ningaloo, and Cardabia stations, WA.](ncomms10496-f1){#f1}\n\n![Characterizing blackened *Genyornis* eggshell.\\\n(**a**) Excavated variably burnt and unusually broken *Genyornis* eggshell from the Wood Point site, Port Broughton, Spencer Gulf, SA (PB, [Fig. 1](#f1){ref-type=\"fig\"}). (**b**) *Genyornis* eggshell fragment from region W ([Fig. 1](#f1){ref-type=\"fig\"}), blackened only at one end, with locations of samples for amino-acid analysis (**c**). (**c**) Concentrations of the stable amino acids glutamic acid (Glu), valine (Val) and leucine (Leu) are reduced in direct proportion to the degree of visual blackening, with samples 3--6 exhibiting only slightly lower concentrations than in unheated fragments of the same egg; \u00b11\u03c3 uncertainties (7%) based on duplicate analyses. Unheated fragments of the same egg exhibit 10% inter-eggshell variability for the same amino acids (6 fragments analysed). Less stable amino acids show similar patterns. Transects through other similarly blackened as well as fully blackened fragments are given in [Supplementary Figs 1 and 2](#S1){ref-type=\"supplementary-material\"}.](ncomms10496-f2){#f2}\n\n![Luminescence and radiocarbon dating of burnt *Genyornis* eggshell.\\\n(**a**) Cumulative probability age distribution derived by summing the individual OSL Gaussian probability distributions for seven sites with burnt *Genyornis* eggshell ([Table 1](#t1){ref-type=\"table\"}). The median age is 47.5\u2009ka, with the light dashed lines representing the 16th and 84th percentiles of the aggregate distribution, similar to \u00b11\u03c3 in a Gaussian distribution. (**b**) Calibrated age ranges (\u00b11\u03c3) for 11 ^14^C dates on unburnt *Genyornis* eggshell associated with burnt fragments ([Table 1](#t1){ref-type=\"table\"}), omitting the two dates from Warroora (W, [Fig. 1](#f1){ref-type=\"fig\"}) that are significantly younger than OSL dates from the same sites (no overlap at \u00b11\u03c3; Table 1). Vertical heavy dashed line represents the age that most closely satisfies the ^14^C dates (with one exception), suggesting that burnt *Genyornis* eggshell is unlikely to be younger than 47.5\u2009ka. Unshaded region in **a** represents the most likely age range for burnt *Genyornis* eggshell, although a range including the lightly shaded region cannot be confidently excluded. ^14^C dates are plotted with increasing minimum age upward; letters refer to regions located on [Fig. 1](#f1){ref-type=\"fig\"}.](ncomms10496-f3){#f3}\n\n![Extent of isoleucine epimerization in *Genyornis* eggshell compared to their current regional temperatures.\\\nAverage A/I in *Genyornis* eggshell from 63 collections with burnt fragments recovered from all ten regions ([Fig. 1](#f1){ref-type=\"fig\"}) for which a secure A/I could be determined, plotted against their current (1960-1990 AD) mean annual temperature (MAT), each with its \u00b11\u03c3 uncertainty ([Supplementary Data 2](#S1){ref-type=\"supplementary-material\"}). Letters refer to regions in [Fig. 1](#f1){ref-type=\"fig\"}, numbers in parenthesis are the number of collections from each region, and the number of different eggshell fragments analysed in each region that contribute to the mean A/I value. The close fit to a simple polynomial regression with a trend towards ever-higher A/I for higher regional MATs, an expected result due to the exponential dependency of racemization rate on temperature, is consistent with, although not in itself proof of, all samples being of the same age. Age differences in excess of 5\u2009ka in almost all cases would result in a significantly poorer polynomial fit.](ncomms10496-f4){#f4}\n\n![Temporal distribution of burnt *Dromaius* and *Genyornis* eggshell in collections from Western Australia.\\\nThe lowest A/I in each collection is used to characterize the age of all fragments, to exclude fragments with accelerated racemization resulting from heating by cooking fires. (**a**) Percentage of all WA *Dromaius* collections that contain burnt eggshell, binned in 0.04\u2009A/I units. (**b**) Percentage of all WA *Genyornis* collections that contain burnt eggshell binned in 0.04\u2009A/I units. (**c**) Total number of WA *Dromaius* collections characterized by AAR and binned in 0.04\u2009A/I units. (**d**) Total number of WA *Genyornis* collections characterized by AAR and binned in 0.04\u2009A/I units. The percentage of collections containing burnt *Genyornis* eggshell (**b**) increases as the number of *Genyornis* collections (**d**) decreases, reflecting the difficulty of eliminating localized heating effects on the measured A/I. The likely extinction window for *Genyornis* is 0.54\u00b10.04\u2009A/I units. Isoleucine racemizes 16% faster in *Dromaius* eggshell relative to *Genyornis*[@b28]; the first appearance of burnt eggshell fragments is at the same age for both taxa. *X* axis scales are identical in all panels.](ncomms10496-f5){#f5}\n\n###### Primary geochronological data for sites with burnt *Genyornis* eggshell.\n\n  **Region**                                                                                             **Site**              **MAT (**\u00b0C**)**   **OSL\u00b11\u03c3 (ka)**   ^**14**^**C (conv)\u00b11\u03c3**   **Cal BP** **\u22121\u03c3**   **Cal BP +1\u03c3**   **Avg A/I\u00b11\u03c3 (n)**\n  ------------------------------------------------------------------------------------------------------ -------------------- ------------------ ----------------- ------------------------- -------------------- ---------------- --------------------\n  *Geochronology for sites with burnt* *Genyornis* *eggshell*                                                                                                                                                                      \n  \u2003B                                                                                                     Cardabia Sliver             24.3                \u00a0               44,110\u00b11,430               45,974             48,753         0.67\u00b10.01 (4)\n  \u2003W                                                                                                     Fabulous                    23.4            53.6\u00b14.4             40,060\u00b1860                43,008             44,410         0.66\u00b10.01 (3)\n  \u2003W                                                                                                     Z-Blowout                   23.4            47.9\u00b12.3             40,850\u00b1960                43,504             45,145         0.57\u00b10.01 (3)\n  \u2003W                                                                                                     Upper 12-Mile               23.4            45.3\u00b13.5            45,360\u00b11,660               47,445            \\>50,000        0.67\u00b10.01 (4)\n  \u2003GN                                                                                                    Small Slot                  23.2                \u00a0                 \\>44,350                \\>48,040              \u00a0            0.58\u00b10.02 (10)\n  \u2003Q                                                                                                     Borrow Pit                  23.0                \u00a0                 \\>41,800                \\>45,380              \u00a0            0.57\u00b10.01 (4)\n  \u2003Q                                                                                                     Sunset-1                    23.0            47.2\u00b13.5            48,790\u00b13,640               47,444            \\>50,000        0.58\u00b10.01 (6)\n  \u2003E                                                                                                     Williams Point              21.1            50.8\u00b13.3            44,630\u00b11,520               46,517             49,283         0.51\u00b10.02 (26)\n  \u2003PB                                                                                                    Wood Point                  17.9            55.0\u00b15.0            42,400\u00b11,760               44,094             47,419         0.37\u00b10.01 (19)\n  \u2003A                                                                                                     Outer Arumpo                17.1                \u00a0                39,830\u00b1840                42,864             44,220         0.34\u00b10.01 (6)\n  \u2003GA                                                                                                    Garnpung                    17.1            43.1\u00b11.6            42,010\u00b11,100               44,317             46,310         0.34\u00b10.01 (9)\n  \u2003WL                                                                                                    Wallaroo                    17.0                \u00a0               42,630\u00b11,190               44,752             47,035         0.35\u00b10.01 (3)\n  \u2003D                                                                                                     Perry Sandhills             16.8                \u00a0               42,600\u00b11,900               44,630             47,132         0.34\u00b10.01 (7)\n  \u00a0                                                                                                      \u00a0                            \u00a0                  \u00a0                     \u00a0                      \u00a0                  \u00a0                  \u00a0\n  *Geochronology for sites with* *Genyornis* *lacking burnt fragments, but with low A/I for their MAT*                                                                                                                             \n  \u2003B                                                                                                     Ningaloo                    24.3                \u00a0                45,060\u00b1930                47,438             49,427         0.67\u00b10.03 (3)\n  \u2003W                                                                                                     11-Mile                     23.4            47.3\u00b14.9                  \u00a0                      \u00a0                  \u00a0            0.68\u00b10.01 (8)\n  \u2003Q                                                                                                     Sunset-2                    23.0            52.1\u00b14.8                  \u00a0                      \u00a0                  \u00a0            0.67\u00b10.03 (11)\n  \u2003E                                                                                                     Williams Point              21.0                \u00a0                42,830\u00b1880                45,194             46,874         0.51\u00b10.02 (26)\n  \u2003E                                                                                                     Williams Point              21.0                \u00a0               45,440\u00b11,210               47,776             49,877         0.51\u00b10.02 (26)\n  \u2003E                                                                                                     Hunt Peninsula              21.0                \u00a0                42,930\u00b1720                45,410             46,792         0.49\u00b10.02 (3)\n  \u2003E                                                                                                     North Harbour               21.0            59.1\u00b13.9             40,800\u00b1580                43,773             44,863         0.52\u00b10.02 (4)\n  \u2003S                                                                                                     Geny Heaven                 19.1                \u00a0                42,250\u00b1360                43,329             45,871         0.37\u00b10.01 (3)\n  \u2003S                                                                                                     Coopers Dune                19.1            44.7\u00b12.1                  \u00a0                      \u00a0                  \u00a0            0.39\u00b10.01 (25)\n  \u2003S                                                                                                     Mystery Is. Sp1             19.1            55.5\u00b12.3             43,770\u00b1630                46,240             47,664         0.46\u00b10.03 (16)\n  \u2003S                                                                                                     Mystery Is. Sp2             19.1            47.3\u00b11.7                  \u00a0                      \u00a0                  \u00a0            0.51\u00b10.08 (5)\n  \u2003S                                                                                                     Flinders North              19.1            55.0\u00b13.0                  \u00a0                      \u00a0                  \u00a0            0.43\u00b10.01 (4)\n  \u2003D                                                                                                     Nialia Lake                 17.1                \u00a0               46,600\u00b12,600              \\>50,000              \u00a0            0.35\u00b10.01 (4)\n  \u2003D                                                                                                     Tandou; Buffy Sand          17.1                \u00a0                46,000\u00b1790                48,768            \\>50,000        0.43\u00b10.01 (4)\n  \u2003D                                                                                                     Tandou; Double Red          17.1                \u00a0                43,230\u00b1590                45,790             47,005         0.39\u00b10.01 (2)\n  \u2003D                                                                                                     Kangaroo Lake               17.1                \u00a0               50,330\u00b11,300              \\>49,000              \u00a0            0.47\u00b10.01 (5)\n  \u2003D                                                                                                     Menindee                    17.1            51.0\u00b14.6                  \u00a0                      \u00a0                  \u00a0            0.56\u00b10.04 (12)\n  \u2003D                                                                                                     Lake Victoria               16.8                \u00a0                 \\>44,400                \\>47,637              \u00a0            0.41\u00b10.01 (4)\n  \u00a0                                                                                                      Mean OSL age                 \u00a0              51.5\u00b14.9                  \u00a0                      \u00a0                  \u00a0                  \u00a0\n  \u00a0                                                                                                      \u00a0                            \u00a0                  \u00a0                     \u00a0                      \u00a0                  \u00a0                  \u00a0\n  *Geochronology for sites with* *Genyornis* *known to be \\>70\u2009ka*                                                                                                                                                                 \n  \u2003B                                                                                                     George\\'s Dune              24.3                \u00a0                42,390\u00b1110                  \u00a0                  \u00a0             0.88\u00b10.02 6)\n  \u2003W                                                                                                     Lower 12-Mile               23.4              89\u00b18               45,040\u00b1140                  \u00a0                  \u00a0            1.00\u00b10.01 (3)\n\nTwo sites known by AAR (B, W) and/or OSL (W) to be more than 70 ka, returned ^14^C ages similar to ages on burnt *Genyornis* eggshell that are known to be much younger (OSL), documenting the limitations of ^14^C dating of eggshell carbonate that is more than 40 ka. A/I, D-alloisoleucine to L-isoleucine; Avg, average; MAT, mean annual temperature; OSL, optically stimulated luminescence; Cal BP, calibrated years before present; A: Arumpo Station, NSW; B: Bullara, Ningaloo, and Cardabia stations, WA; D: Lower Darling River, NSW; E: Lake Eyre, SA; GA: Garnpung Station, NSW; GN: Sites on and around Gnaraloo Station, WA; PB: Port Broughton, SA; Q: Sites on and around Quobba Station, WA; WL: Wallaroo, SA; W: Sites on and around Warroora Station, WA.\n\nSites are ordered in each panel by their current MAT. ^14^C dates calibrated with Calib 7.1 and SHcal13.\n\n[^1]: Deceased.\n"}
+{"text": "1. Introduction {#sec1-sensors-17-02762}\n===============\n\nTactile sensing is an important mode for both human and robots to perceive the environment. In the past decades, researchers have developed many different tactile sensors for robots \\[[@B1-sensors-17-02762],[@B2-sensors-17-02762],[@B3-sensors-17-02762],[@B4-sensors-17-02762]\\], and the core part of those tactile sensors is to detect the contact and contact force, or force distribution over the fingertip area. For example, a successfully commercialized sensor is the tactile sensor array from Pressure Profile Systems, which measures the normal pressure distribution over the robot fingertip, with a spatial resolution of 5 mm. The sensor has been applied to multiple commercialized robots, including the PR2 robot, and Barrett hands, and it successfully assisted common robotic tasks, such as contact detection and gripping force control. With the force measurement from the fingertip tactile sensors, a robot is much less likely to break delicate objects. The contact detection and localization also refine the robots' performance in grasping, and in-hand manipulation.\n\nHowever, tactile sensing can be much more informative for robots. Considering humans as an example, although we are not able to precisely measure contact force, humans can get abundant information from tactile sensing, such as the objects' shape, texture, materials, and physical properties including mass, compliance, roughness, friction and thermal conductivity \\[[@B5-sensors-17-02762],[@B6-sensors-17-02762]\\]. Tactile sensing is an important part for humans' closed-loop manipulation as well. With touch, we can know whether the cup in the hand is going to slip, and thus adjust the gripping force accordingly; we can know whether a USB connector is going to be correctly plugged into the socket because we get the feedback of the impedance force. Those tasks, however, are still challenges for robots because they are not yet able to fully apply tactile sensing. Robots need more than force feedback, while the existing sensors do not obtain enough tactile information for the robots. Thus, we ask the question: what kind of tactile signal is needed for the robots to better understand the environment?\n\nWe try to address the question with the answer that geometry sensing is equally important as the force sensing for robots. To better measure the geometry, a deformable surface, and high spatial resolution sensing are required. With the measurement of high-resolution geometry, the robot will be able to learn more about the objects' shape and texture. Moreover, the dynamic interaction between the soft sensor and the environment can reveal more physical properties of the object being contacted, such as the slipperiness and compliance.\n\nIn this paper, we discuss the development of a tactile sensor called GelSight, which measures high-resolution geometry, and which can be used to infer local force and shear. The sensor uses a deformable elastomer piece as the medium of contact, and an embedded camera to capture the deformation of the elastomer surface. The high resolution 3D geometry of the contact surface can be reconstructed from the camera images. When the sensor surface is painted with small black markers, the motion of the markers provides information about both normal force and shear force. The vision-based design of the sensor also makes the hardware accessible and the installation much simpler, and the software for processing the raw data easier to develop by using the algorithms in computer vision. With the help of GelSight, a robot can easily capture the detailed shape and texture of the object being touched, which makes the touch-based object or material recognition much easier. Research also shows that GelSight can also help a robot to sense multiple physical properties of the objects.\n\nThe first GelSight prototype was developed in 2009 by Johnson and Adelson \\[[@B7-sensors-17-02762]\\]. Its high resolution capabilities were further demonstrated in \\[[@B8-sensors-17-02762]\\]. Unlike other optically based approaches, GelSight works independently of the optical properties of the surface being touched. The ability to capture material-independent microgeometry is valuable in manufacturing and inspection, and GelSight technology has been commercialized by a company (GelSight Inc., Waltham, MA, USA). However, the requirements for metrology are quite different than those of a robot finger.\n\nFor a robot finger, it is necessary that the sensor be compact enough to be mounted on an end effector. At the same time, there is no need for micron-scale resolution. We have developed sensors that are compact, yet have resolution far exceeding that of human skin. In this paper, we will firstly demonstrate the basic working principle and software development of the GelSight, and then we will introduce the design and fabrication processes of different sensor versions. Following that, we will show some experiment results on the robotic fingertip GelSight's performance on measuring geometry and force. For the force measurement, since the sensor displacement, which is directly measured by GelSight, is in the nonlinear relationship to the contact force, we use deep neural networks to measure the contact force from the learned examples, and show some preliminary results of the force measurement with basic shapes. In the last section, we will briefly introduce some examples of GelSight helping robots with multiple tasks.\n\n2. Related Work {#sec2-sensors-17-02762}\n===============\n\nSome thorough reviews of the existing tactile sensors in the past decades are given by \\[[@B1-sensors-17-02762],[@B2-sensors-17-02762],[@B3-sensors-17-02762],[@B4-sensors-17-02762]\\]. The sensors use different sensing principles, such as resistance, capacitance, piezoelectricity, optic component, or magnetics. The major focus of the sensors has been measuring the pressure distribution, or contact force and location, over a certain area. For the tactile sensors applied on robots, most of the sensors are designed for fingertip or gripper end-effector (an example is \\[[@B9-sensors-17-02762]\\]), which measures the force or contact during grasping; some other sensors designed for body (an example is \\[[@B10-sensors-17-02762]\\]), which detect the contact over a much larger area and are commonly used for contact or collision detection during the robot motion.\n\nAmong the tactile sensors, the optical sensors based on vision stand out because they are of easier wiring and fabrication processes, and can mostly provide a relatively high spatial precision in locating contact area. The vision-based tactile sensors usually use a deformable body, either a piece of rubber or a fluid balloon, as the sensing medium, and apply a camera to capture the deformation of the medium. In most cases, the deformation is indirectly measured from other visual cues, such as the deformation of the pre-printed patterns on the medium. In the 1980s, Schneiter and Sheridan \\[[@B11-sensors-17-02762]\\] and Begej \\[[@B12-sensors-17-02762]\\] used an optical fiber to capture the light reflection on silicone surface, and used cameras to capture the change of optic signal from the optical fibers. The deformation of the medium would cause a change in the light reflection. A successful example is Tar and Cserey \\[[@B13-sensors-17-02762]\\], which has already been commercialized. The sensor used a hollow hemispherical rubber dome as the contact medium, the dome has a reflective inside surface. The sensor uses three receivers in the bottom to measure the reflective light from the deformed dome, thus estimating the three-axis contact force. However, the spatial measurement is not available with the sensor.\n\nAnother popular design for a vision-based sensor is to print marker patterns on or in the sensing medium, and track the motion of the markers from the embedded camera. Some examples of the designs include \\[[@B14-sensors-17-02762],[@B15-sensors-17-02762],[@B16-sensors-17-02762],[@B17-sensors-17-02762]\\]. Ferrier and Brockett \\[[@B14-sensors-17-02762]\\] developed an analytical model to calculate the value and position of the point-contact force on a round shaped fluid finger. However, in most cases, an analytical model is hard to build and is restrained by many contact requirements. The nonlinearity of the elastomer material, and the complicated contact condition with multiple contact points or contact surface greatly increases the difficulty. Kamiyama et al. \\[[@B15-sensors-17-02762]\\] developed a large flat sensor with two layers of markers in the cover rubber, and they used a simplified mechanical model to calculate normal and shear force during the point contact condition. In their later work \\[[@B18-sensors-17-02762]\\], they scaled down the sensor design to a robot finger tip. They used experimental methods to calibrate the contact force. Chorley et al. \\[[@B16-sensors-17-02762]\\] designed another hemispherical sensor with markers close to the surface, and Cramphorn et al. \\[[@B19-sensors-17-02762]\\] introduced an improved design by adding the core-shaped fingerprint on the sensor surface. The sensor is 3D printed, which makes it much easier to reproduce. Ward-Cherrier et al. \\[[@B20-sensors-17-02762]\\] showed the sensor can be adjusted to different 3D shapes and applied on robot grippers. TacTip could sensitively discriminate the contact, and can be used to extract the approximate edge of the contact surface. The force was not measured, but the edge detection successfully helped a robot to localize contact and follow the contours \\[[@B21-sensors-17-02762]\\]. Ito et al. \\[[@B17-sensors-17-02762]\\] designed a sensor that had a hemispherical shape filled with translucent water, and the marker patterns were printed on the surface. The sensor responded to contact with both marker motion and the change of filling water's reflection, and they built an analytical model to estimate the shapes when the contact shape is simple. They also showed in \\[[@B22-sensors-17-02762]\\] that, by counting the markers that were 'stuck' on the surface or 'lagged behind', the sensor could estimate partial slip stage. However, their sensors are too large in volume to be properly applied on robots.\n\nThe high-resolution measurement is still largely under-exploited. Maheshwari and Saraf \\[[@B23-sensors-17-02762]\\] offered one possible solution: they proposed an electro-optical tactile sensor that used a surface film made of metal and semiconducting nanoparticles that converted pressure into optical signals. The sensor is able to sense the pressure distribution at a resolution similar to human fingers.\n\nHere are some major challenges for making the desired robotic tactile sensors:Measurement of shear force. Only some of the tactile sensors are able to measure shear force as well as normal force, while the shear force is very important in multiple robotic tasks, such as estimating object states in grasping, estimating surface friction of the objects through touch.Detecting contact area. Most of the sensors focus on the situation of point contact, which means they are designed to measure the location and magnitude of the contact force. However, in many tactile tasks, the contact is an area, or multiple areas, instead of a single point. Sensors that can detect more tactile information based on the contact area are desired.Hardware optimization. If the sensors can be used on the robots, they must be small in size, easy on wiring, and offer real-time feedback. Some sensors have delicate structures and good signals, but the design is either too bulky or too complicated to be mounted on robot hands.Fabrication challenge. A major work in the research of the tactile sensing is to develop a method to fabricate the sensors, which is usually nontrivial. Unfortunately, most of the fabrication methods are not well shared---it is hard for another lab to duplicate the complicated fabrication procedures. TacTip offers a good example where the 3D printing methods are open sourced. In other cases, devices have been commercialized and are therefore available via purchase. Two good examples are the Optoforce sensor (OptoForce Kft., Budapest, Hungary) \\[[@B13-sensors-17-02762]\\] and the BioTac sensor (SynTouch Inc., Montrose, CA, USA) \\[[@B24-sensors-17-02762]\\]: the researchers founded startups to improve the product design and produce the sensor commercially, so that other robotic labs have direct access to the sensors.\n\nThe GelSight sensor can precisely locate the contact area, and even measure the high-resolution shapes. It also can be used to estimate shear force and slip state. Apart from the sensing principles, we also introduce our effort in improving the compact design of the sensor and the fabrication process, so that the sensor can be easily applied on robot grippers. We also try to encourage other labs to reproduce the sensor by simplifying the fabrication process, and sharing the 3D printing files and other fabrication details.\n\n3. Principle of GelSight {#sec3-sensors-17-02762}\n========================\n\n3.1. Overview {#sec3dot1-sensors-17-02762}\n-------------\n\nThe GelSight sensor was initially designed to measure the 3D shape and texture of the target surface, as firstly introduced by Johnson and Adelson \\[[@B7-sensors-17-02762]\\]. The device consists of a slab of transparent elastomer covered with a reflective coating membrane. When an object is pressed on the elastomer, the membrane distorts to take on the shape of the object's surface, but with consistent reflectance. When viewed from behind, the membrane appears as a relief replica of the surface. We place a camera to record the image of this relief, using illumination from light sources at different directions. A photometric stereo algorithm \\[[@B25-sensors-17-02762]\\] is then used to reconstruct the depth map of surface. [Figure 1](#sensors-17-02762-f001){ref-type=\"fig\"}a,b shows an example of an Oreo cookie being pressed against the elastomer, while the reflective membrane takes on the shape of the Oreo's surface. We reconstruct the surface from the shaded image, which is rendered in [Figure 1](#sensors-17-02762-f001){ref-type=\"fig\"}c. The spatial resolution of the sensor, when optimized for resolution, can reach 1--2 microns. In the case of compact GelSight devices designed for robot fingers, the spatial resolution is typically in the range of 30--100 microns.\n\nPhotometric stereo is a technique in computer vision for estimating the surface normals of objects by observing the object under different lighting conditions. Theoretically, when the surface reflectance and light properties are known, the reflection under lights from three separate directions can be calculated. For GelSight, the reflectance of the coating membrane is known and consistent, and the lighting condition is well controlled: different LED arrays are arranged at different directions. Images corresponding to the multiple directions can be recorded sequentially, or separated by using different color channels. By combining the shading from three or more directions, we estimate the surface normals on each pixel of the shaded image, and then integrate the surface normal to get the 3D shape of the surface. [Figure 2](#sensors-17-02762-f002){ref-type=\"fig\"} shows the schematic and pictures of the GelSight device introduced in \\[[@B7-sensors-17-02762]\\].\n\nPhotometric stereo provides a height map, i.e., it describes the displacement of the membrane surface along the *z*-axis. During contact, the membrane also moves tangentially. We can measure the tangential displacements (in *x* and *y*) by printing markers on the membrane and tracking their displacement over time \\[[@B26-sensors-17-02762],[@B27-sensors-17-02762]\\]. We paint the black dots, either evenly or arbitrarily distributed, between the elastomer base and the reflective membrane. Their motion in the camera image directly corresponds to the lateral displacement of the elastomer's surface, and can indicate contact force and torque, as demonstrated in [Figure 3](#sensors-17-02762-f003){ref-type=\"fig\"}b: different kinds of force or torque will make different patterns of motion field, and the magnitude of the marker motion is roughly proportional to the amount of force or torque. The distribution of the marker motion also tells information about slip or incipient slip states.\n\nWe have developed multiple versions of GelSight devices, with similar working principles but different dimensions, illumination arrangements, elastomer shapes and coatings. They have different resolution, speeds, and aim at varied working conditions and applications. Those designs of GelSight are introduced in [Section 4](#sec4-sensors-17-02762){ref-type=\"sec\"}. Lately, we developed a compact version for robot grippers, called fingertip GelSight sensors \\[[@B28-sensors-17-02762],[@B29-sensors-17-02762]\\]. The sensors are compact in volume, and can provide near real-time signal feedback. The elastomer sheets on these sensors are much thinner (under 2.5 mm), and are of domed shape to allow a greater range of contact conditions.\n\n3.2. Material {#sec3dot2-sensors-17-02762}\n-------------\n\nIn this section, we introduce how to choose the materials for the sensor's sensing elastomer and the reflective coating.\n\nFor choosing the elastomer base for the sensor, there are several factors that need to be considered: optical transparency, stretchability, hardness, robustness, and complexity to fabricate. We require the elastomer to be optically clear, and stretchable enough to yield to the object shapes, but the hardness could be varied according to different applications. We typically use elastomer with Shore A values between 5 and 20. The common materials we use for making the elastomer are thermoplastic elastomers (TPEs) such as styrenic block copolymers, and silicone rubbers. TPEs are cast by melting, while the silicone rubbers are two-part liquids that cross-link into elastomers.\n\nIn general, we wish the sensor elastomer base to be soft enough so that it is sensitive to small contact force. However, if the sensor is to be used in heavy-loaded environments, a harder elastomer base is desired to ensure the elastomer will be sensitive in the measuring range. At the same time, the hardness of the elastomer is restricted to the available elastomer material. As an example, the elastomer in the fingertip GelSight sensor (introduced in [Section 4.2](#sec4dot2-sensors-17-02762){ref-type=\"sec\"}) is a clear silicone that is close to a neo-Hookean solid, with the coefficient $\\mu$ of 0.145 MPa, and the hardness is close to Shore 00-45. In practice, the sensor with this hardness works well for moderate contact perception and robotic manipulation tasks. The minimum perceivable force of the sensor, when contacting different types of objects, is mostly less than 0.05 N. We use an ATI Nano-17 force/torque sensor (ATI Industrial Automation, Apex, NC, USA) to calibrate the force measurement of GelSight (as shown in Figure 14a, and GelSight performs more sensitively than the ATI sensor in detecting small contact forces. The detailed results are shown in [Table 1](#sensors-17-02762-t001){ref-type=\"table\"}.\n\nThe coating skin is a thin layer of silicone mixed with pigment. We choose pigment with different reflective properties for different sensor designs or applications. Examples of the commonly used matte and semi-specular coatings are shown in [Figure 4](#sensors-17-02762-f004){ref-type=\"fig\"}. A semi-specular coating is sensitive to small variations on the surface normal, and a matte coating is favorable for accurate measuring of the general shapes. The different reflective properties requires different optical designs. However, in all cases, the pigment should be made of fine particles, and the coating should be thin in order to reveal the fine details on the object. We use bronze flake or aluminum flake pigment for the semi-specular coating, and fine aluminum powder for the matte coating.\n\n3.3. Algorithm for Measuring Shape {#sec3dot3-sensors-17-02762}\n----------------------------------\n\nWe model the surface of the sensor with a height function $z = f\\left( x,y \\right)$, so that the surface normal is $\\mathbf{N}\\left( x,y \\right) = (\\frac{\\partial f}{\\partial x},\\frac{\\partial f}{\\partial y}, - 1)$ (The surface normal is in this form because the surface is the zero level surface of a scalar field $\\phi\\left( x,y,z \\right) = f\\left( x,y \\right) - z$, and which is normally equal to $\\nabla\\phi$). Given the lighting and surface reflection are evenly distributed, the shading depends only on the local surface normal, which means that the cast shadows or internal reflections are neglected. Under this assumption, under the single light source, the light intensity at $\\left( x,y \\right)$ can be calculated as $I\\left( x,y \\right) = R{(\\frac{\\partial f}{\\partial x},\\frac{\\partial f}{\\partial y})}$. The reflectance function *R* models both the lighting environment and the surface reflectance. Considering the lighting sources from multiple directions, which can be inferred as the different channels in the red-green-blue (RGB) image $\\mathbf{I}$ when the lights are of different colors, we have $$\\mathbf{I}\\left( x,y \\right) = \\mathbf{R}\\left( {\\frac{\\partial f}{\\partial x}\\left( x,y \\right),\\frac{\\partial f}{\\partial y}\\left( x,y \\right)} \\right),$$ where $$\\mathbf{I}\\left( x,y \\right) = \\left( \\!\\begin{matrix}\n{I_{1}\\left( x,y \\right)} & {I_{2}\\left( x,y \\right)} & {I_{3}\\left( x,y \\right)} \\\\\n\\end{matrix}\\! \\right),\\mathbf{R}\\left( p,q \\right) = \\left( \\!\\begin{matrix}\n{R_{1}\\left( x,y \\right)} & {R_{2}\\left( x,y \\right)} & {R_{3}\\left( x,y \\right)} \\\\\n\\end{matrix}\\! \\right).$$\n\nWe use different lights of red, green and blue colors from different directions, so that $I_{1}\\left( x,y \\right),I_{2}\\left( x,y \\right)$ and $I_{3}\\left( x,y \\right)$ correspond to the shading images under the lights from a single direction. We assume $\\mathbf{R}$ is a consistent function for one sensor.\n\nThe reflectance functions $\\mathbf{R}$ are in the nonlinear relationship to $(\\frac{\\partial f}{\\partial x},\\frac{\\partial f}{\\partial y})$, and we are interested in the inverse function---a function that maps observed intensity to geometry gradients. To do this, we build a lookup table. The lookup table is three-dimensional, and each entry contains a gradient and a first-order approximation of the reflectance functions in the neighborhood near the gradient. In a simplified version, we only record the gradient at each entry space and match the gradient reading to the closest entry. The lookup table is generated by a calibration procedure, when a small sphere with known diameter is pressed on the sensor, as introduced in [Section 3.5](#sec3dot5-sensors-17-02762){ref-type=\"sec\"}.\n\nAfter getting $(\\frac{\\partial f}{\\partial x},\\frac{\\partial f}{\\partial y})$ using $\\mathbf{R}^{- 1}$, we reconstruct the heightmap $z = f\\left( x,y \\right)$ by integrating the normals. It can be precisely calculated using re-weighted least squares approach (IRLS) \\[[@B31-sensors-17-02762]\\], as introduced in \\[[@B8-sensors-17-02762]\\], or can be solved using Poisson equation:$$\\nabla^{2}f = g,$$ where $$g = \\frac{\\partial}{\\partial x}(\\frac{\\partial f}{\\partial x}) + \\frac{\\partial}{\\partial y}(\\frac{\\partial f}{\\partial y}).$$\n\nWe solve the Poission Equation ([3](#FD3-sensors-17-02762){ref-type=\"disp-formula\"}) using fast Poisson solver with discrete sine transform (DST), thus we can get a fast computation on the heightmap reconstruction, and it is possible to run the algorithm online \\[[@B32-sensors-17-02762]\\].\n\n3.4. Algorithm for Measuring Marker Motion {#sec3dot4-sensors-17-02762}\n------------------------------------------\n\nApart from getting the heightmap, we also calculate the motion of the markers from the images. In the initial GelSight frame $Frm_{0}$, where nothing is in touch with the sensor, we locate the markers and record their positions $\\mathbf{u}_{0}$; in the following frames, we also calculate the markers' positions $\\mathbf{u}$, so the motion is $d\\mathbf{u} = \\mathbf{u} - \\mathbf{u}_{0}$.\n\n[Figure 5](#sensors-17-02762-f005){ref-type=\"fig\"} shows the procedure of locating markers from $Frm$: in the beginning, we record the initial frame $Frm_{0}$, and get the image's background $I_{0}$ using low-pass Gaussian filter on $Frm_{0}$, where the black markers and high-frequency noise are removed. For a given image $Frm$, we calculate the differential image $dI = Frm - I_{0}$, which shows only the change in the image. In $dI$, the pixels where the sensor surface has a big curvature and large surface normal will have a large color intensity, but the black markers are dark. We use a threshold on $dI$ to get the black area *m*, which correspond to the area of markers. Then, we calculate the centroids of each connected area in *m*, and consider them as the markers' positions $\\mathbf{u}$.\n\n3.5. Calibration for Shape Measurement {#sec3dot5-sensors-17-02762}\n--------------------------------------\n\nThe sensor calibration aims to find the mapping between pixel intensities of the GelSight images and the surface normal or geometry. In other words, it builds the lookup table of the function $\\mathbf{R}^{- 1}\\left( \\mathbf{I} \\right)$, which is the reverse function of Equation ([1](#FD1-sensors-17-02762){ref-type=\"disp-formula\"}). Each GelSight sensor is slightly different in construction, so that, for measuring the shape, the calibration is mandatory.\n\nDuring the calibration, we press a small ball on GelSight, and record the image. Examples of the GelSight images for different sensors are shown in [Figure 6](#sensors-17-02762-f006){ref-type=\"fig\"}. Since the ball is of known shape, we are able to know the surface normal at every pixel. We choose a ball because the hemispherical shape theoretically contains surface normals in all directions. From the image, we can get the correspondence of the pixel intensity $I\\left( x,y \\right)$ to the surface normal $(\\frac{\\partial f}{\\partial x},\\frac{\\partial f}{\\partial y},1)$. In practice, instead of using the direct intensity *I*, we use the relative intensity $dI = I - I_{0}$ as introduced in [Section 3.3](#sec3dot3-sensors-17-02762){ref-type=\"sec\"}, in order to eliminate the influence of the spatial inhomogeneity in the illumination. We make the lookup table $\\mathbf{R}^{- 1}$ as a three-dimensional matrix, corresponding to the three color channels of the input images. For the Fingertip GelSight devices, which use a USB webcam and output the color images in 8 bits, we select the dimension of $\\mathbf{R}^{- 1}$ as $80 \\times 80 \\times 80$.\n\nTo further reduce the influence of the markers, noise, and inhomogeneity of the light conditions on the calibration image, we record the multiple images of the ball pressed on various positions on the sensor surface, and make the final look-up table from the average values of all images. For the intensities that are not captured in the images, we use a linear interpolation to fill in the blanks.\n\n3.6. Marker Motion for Force Measurement {#sec3dot6-sensors-17-02762}\n----------------------------------------\n\nThe markers' motion represents well the contact force in two ways: the pattern of the motion field indicates the type of the force or torque, and the magnitude of the motion is roughly proportional to the force \\[[@B26-sensors-17-02762],[@B27-sensors-17-02762]\\]. Examples of the different motion field patterns are shown in [Figure 3](#sensors-17-02762-f003){ref-type=\"fig\"}: under the normal force, the markers spread outwards from the contact center; under the shear force, the markers all move in the shear direction; under the in-plane torque, which means the torque axis is perpendicular to the surface, the markers move in a rotational pattern. When there is a combination of multiple force types, the motion field can be approximated as a linear combination of the individual forces.\n\nThe force-deformation curves of the fingertip GelSight (when elastomer not glued on sensor) are shown in [Figure 7](#sensors-17-02762-f007){ref-type=\"fig\"}. In [Figure 7](#sensors-17-02762-f007){ref-type=\"fig\"}a, the gap between the loading and unloading curves is caused by the viscoelasticity. [Figure 7](#sensors-17-02762-f007){ref-type=\"fig\"}b shows that when the shear load is small, the force is linear to the load displacement; when the load increases, partial slip or slip occurs, which stops the shear force from growing. [Figure 7](#sensors-17-02762-f007){ref-type=\"fig\"}c is from the same shear experiment, and it shows that the average motion magnitude of the markers within the contact area is proportional to the shear force, regardless of whether slip or partial slip occurs. In fact, the linear relationship remains even before force reaches equilibrium.\n\nOn the other hand, the linear relationship between force and marker motion only exists when the contact geometry remains unchanged. When contacting an object of another shape, there is still a linear relationship between the force and marker motion, but the parameters change. To achieve a shape independent relationahip between force and displacement, we turn to machine learning. A method to measure the force when contacting unknown geometries is to apply Convolutional Neural Networks (CNN) on the GelSight image. CNNs have proved very effective in extracting features from the nonlinear and high-dimensional data, like images. The GelSight output is similar to normal computer images in that it is also high-dimensional, noisy, and highly nonlinear. It can be expected that the CNN can also pick up the useful features regarding to contact force from GelSight data.\n\n3.7. Marker Motion for Slip Detection {#sec3dot7-sensors-17-02762}\n-------------------------------------\n\nThe GelSight sensor can detect signals related to slip and incipient slip by analyzing the distribution of the markers' motion. When the markers in the peripheral contact area moves significantly less than the markers in the central or along the shear directions, it is very likely that slip has occurred or is going to occur soon. The rule holds for the GelSight sensor with both flat and curved surface. The physical foundation of this measurement is that, when slip going to occur under a soft contact condition, it starts from the peripheral contact area, which results in the relative displacement between the sensor surface and the object in the area, but the central area is still 'stuck' to the contact surface until the overall slip occurs. Ito et al. \\[[@B22-sensors-17-02762]\\] used the similar feature to detect slip with a vision-based soft sensor.\n\nExamples of how the displacement field changes as the shear load increases are shown in [Figure 8](#sensors-17-02762-f008){ref-type=\"fig\"} (experiments done with the fingertip GelSight sensor with flat sensor elastomer piece). The degree of partial slip can be inferred from the inhomogeneity degree of the marker displacement distribution. We use the entropy to describe the inhomogeneity degree of the displacement field. The entropy of a histogram X is $$H\\left( X \\right) = - \\int_{X}p\\left( x \\right)\\log p\\left( x \\right)dx.$$\n\nThe entropy *H* increases as the partial slip degree increases, shown in the last column in [Figure 8](#sensors-17-02762-f008){ref-type=\"fig\"}. To prevent the slip occurrence, a possible way is to set a warning threshold on *H*.\n\nApart from translational slip, rotational slip is also commonly seen in grasping. Similarly, partial rotational slip occurs before the total slip occurs, and it starts from the peripheral area. The partial slip can be inferred from the inhomogeneous distribution of the markers' rotational angles along the rotational center, where the markers in the peripheral area make much less angles. [Figure 9](#sensors-17-02762-f009){ref-type=\"fig\"} shows examples of how the marker motion and rotation angle change as the in-plane torque increases.\n\nThe marker motion analysis for detecting slip or partial slip is most effective for contacting the objects with flat surfaces or small curvatures, when the contact surface is large enough to contain multiple markers, and the object motion can be barely detected by the motion of contact geometry. For objects with complicated shapes or obvious textures, slip can be more easily detected by comparing the motion of the marker and the motion of the shapes. This is equivalent to comparing the shear or rotation motion of the sensor surface and the motion of the object. In reality, the success rate for slip detection on general objects could be largely increased when combining the multiple cues. Detailed experiments are shown in \\[[@B29-sensors-17-02762]\\].\n\n4. Design and Fabrication of GelSight {#sec4-sensors-17-02762}\n=====================================\n\n4.1. The Desktop GelSight Sensor {#sec4dot1-sensors-17-02762}\n--------------------------------\n\nThe first GelSight sensor Johnson and Adelson \\[[@B7-sensors-17-02762]\\] is shown in [Figure 2](#sensors-17-02762-f002){ref-type=\"fig\"}: the sensor is in a shape of a cubic box with the side length of 30 cm, and is made of standard building frames and dark acrylic boards. The sensing elastomer is placed on the top, and the reflective coating of the elastomer is semi-specular, made from bronze flake pigment. The reflective coating has a sharp response to surface textures. A machine vision camera is placed in the bottom. Three LED arrays, covered by the filters of red, green, and blue, are placed under the sensor surface, making the lights illuminate at the elastomer with an angle of 60\u00b0. The LEDs have lenses on the top, so that the lights getting out of the LEDs are in small angles. The sensing range of the sensor is about 60 mm \u00d7 40 mm, which is only a small part in the central of the sensor's top surface.\n\nJohnson et al. \\[[@B8-sensors-17-02762]\\] propose an improved version of the sensor with better spatial resolution. For the sensor coating, they choose a fine and diffuse surface made of silver powder with particle size of one micro and in spherical shape. The diffuse reflectance makes the coating have a more uniform response to a broader range of surface normals, thus the sensor could capture a more accurate geometry. The diffuse coating makes more distinctive reflection under a side lighting conditions, which means that the incoming lights are expected to be nearly parallel to the sensor surface. To do this, the researchers mount the LEDs on the side of a clear glass plate, and mount the sensing elastomer in the central part of the plate, as shown in [Figure 10](#sensors-17-02762-f010){ref-type=\"fig\"}a. In this case, the lights from LEDs are guided by the glass plate and is in an approximately parallel angle when entering the elastomer. The researchers arrange six LEDs around the glass plate, all in white color. The six LEDs light up asynchronously, and the camera captures six images under each LED lighting. The design increases the precision of the measurement of surface normal. With the same optical design, the authors made a bench configuration sensor ([Figure 10](#sensors-17-02762-f010){ref-type=\"fig\"}b) and a portable configuration ([Figure 10](#sensors-17-02762-f010){ref-type=\"fig\"}e), and both sensors can get a high resolution geometry of the contact surface. However, the asynchronous setting of the lights takes longer time to obtain one geometry measurement, so that the sensor can not be used to measure real-time dynamic tactile information.\n\n4.2. Fingertip GelSight Sensors {#sec4dot2-sensors-17-02762}\n-------------------------------\n\nA qualified tactile sensor for robot hand should be small in volume and should be able to acquire online signal and processing. For the GelSight sensor, the major design challenge is to reduce the size of the illumination system while maintaining the illumination quality, which means the lights should enter the sensing elastomer in the concentrative and desired angles. However, on the other hand, for most of the robotic tasks, the required precision of the geometry measurement is not very high.\n\nIn this section, we introduce two versions of the fingertip GelSight sensors that are designed for robot grippers. The illumination of the sensors are of different colors, and the embedded camera is a commercialized USB web-camera (C310 from Logitech, Lausanne, The Switzerland). The camera can be easily connected to a computer through USB, and capture images at 30 Hz with a resolution as high as 1920 \u00d7 1080 (we mostly use the resolution 640 \u00d7 480 because the driver is more widely supported). Therefore the sensors could measure the online tactile images and reconstruct the 3D geometry. (Note, however, that this sensor, like other USB two sensors, uses H.264 encoding, which introduces significant image latency). The sensing field is a 18 mm \u00d7 14 mm area in the center of the sensor, and the spatial resolution is around 20 to 30 microns. We also make the sensing elastomer on the sensor in the domed shape, which enables the contact over a wider set of conditions.\n\nThe first version is built by Li et al. \\[[@B30-sensors-17-02762]\\], and is designed for a Baxter Robot gripper (Rethink Robotics, Boston, MA, USA) ([Figure 11](#sensors-17-02762-f011){ref-type=\"fig\"}a,e). The sensor is close to a cubic shape, with the compact size of 35 mm \u00d7 35 mm \u00d7 60 mm. The design uses perpendicular acrylic plate sets to guide the lights from the top to the sensing elastomer in the bottom, as shown in [Figure 11](#sensors-17-02762-f011){ref-type=\"fig\"}b. The sensing elastomer has semi-specular coating to reveal the details and small fluctuations on the object's surface. After the internal reflection in the acrylic plate, the lights will be at the angles close to the parallel direction of the elastomer surface. The LEDs on the four sides of the sensors are in the form of a line array (1 \u00d7 8), and are of four colors: red (R), green (G), blue (B) and white (RGB). The hue and saturation of each pixel indicates the direction (yaw angle) of the surface normal, since the light sources from different directions are different in color, and the intensity corresponds to the magnitude (pitch of surface normal). The parts of the sensor are displayed in [Figure 11](#sensors-17-02762-f011){ref-type=\"fig\"}d: the support structure is 3D printed with non-transparent materials, which holds the camera and the acrylic guiding plates, as well as providing the mounting structure to the robot gripper. The clear acrylic plates are cut by a laser cutter, and one side is manually ground to 45\u00b0 to change the light direction. The LEDs are manually soldered into compact arrays, and glued to the top side of the support, just in front of the acrylic guiding plates.\n\nThis sensor has two major deficiencies: firstly, although the sensor is very sensitive to small curvatures, the measurement of surface normal is not precise because of the non-parallel illumination and semi-specular coating of the gel. The size reduction decreases the illumination quality. Secondly, the fabrication of this sensor is over-complicated. Too much accurate operation is required by manual work, so that the product is hard to be standardized, and the fabrication is time-consuming.\n\nIn 2017, Dong et al. \\[[@B29-sensors-17-02762]\\] proposed another version of the fingertip GelSight sensor ([Figure 12](#sensors-17-02762-f012){ref-type=\"fig\"}a), with largely improved precision of shape measurement and simplified fabrication process. The new sensor is of approximately the same size and spatial resolution with the previous fingertip sensor, but is in a hexagonal shape, and has a new illumination system using LEDs with three colors (RGB). The new LEDs (Osram Opto Semiconductor Standard LEDs-SMD, Sunnyvale, CA, USA) have small collimating lenses, and the emitted light are within a viewing angle of 30\u00b0. They are tightly arranged into $2 \\times 4$ arrays with a customized circuit boards, and are tilted at the angle of 71\u00b0 to the sensing elastomer surface from the sides. The LED and elastomer are supported by a semitransparent tray, which homogenizes the LED light while allowing a high transmission rate. The sensor coating is a matte surface, and the matte coating as well as the entire illumination system favor a more accurate surface normal measurement.\n\nMost of this sensor's parts are 3D printed with a Formlab 2 printer (Formlabs Inc., Somerville, MA, USA), and the clear acrylics are cut by a laser cutter. The parts are shown in [Figure 12](#sensors-17-02762-f012){ref-type=\"fig\"}c. The base part, a semitransparent hexagonal plastic tray and camera spacer, is printed with clear resin. The tray is filled with a piece of clear acrylic sheet, and the sensing elastomer is glued on the acrylic sheet. The camera is mounted on the top of the hexagonal tray to capture the deformation of the elastomer. A support shell, 3D printed with 'tough' resin, covers the circuits and lights, as well as providing the mounting structure to mount the sensor to a WSG-50 robot gripper. The drawings for the 3D printing parts are open-sourced at \\[[@B33-sensors-17-02762]\\]. The fabrication of the new sensor is highly standardized, and manual labor is significantly reduced.\n\n4.3. Fabrication of the Sensor Elastomer {#sec4dot3-sensors-17-02762}\n----------------------------------------\n\nThe sensing elastomer used on GelSight is made of two parts: the clear elastomer sheet as the base, and the reflective membrane. The elastomer sheet can be made from many kinds of silicone or thermoplastic elastomer (TPE), as long as it is transparent and has good deformability. However, the reflective membrane is essential for the sensor's signal quality. For making good signal quality, the membrane must be fine, uniform, thin, smooth, firm, and light-blocking. The fineness and thickness of the membrane will influence the resolution of the shape measurement, and if the membrane is not uniform or smooth, the tactile images will contain noise from the bumps on the surface. Finally, we must glue the elastomer on the sensor's supporting plate to reduce the residue force during the contact. The glue we commonly use is a double-side adhesive silicone sheet.\n\nThe typical fabrication of the sensing elastomer contains three steps: making the transparent elastomer base, printing the markers, and painting the reflective coating over the markers and base. To make the elastomer, we choose some commercialized polymer that comes in the fluid phase. We pour the liquid material in the mold, either a flat tray or a concave mold, depending on what kind of sensor we need. Then, we demold the elastomer after it solidifies. The elastomer we use the most is the XP-565 from Silicones, Inc. (High Point, NC, USA). It comes in two liquid parts, and after mixing the two parts at some specific portion, we pour the silicone in the molds, and it solidifies within several hours under room temperature, or within 15 min in an oven at the temperature of 100 \u00b0C. The hardness of the elastomer can be adjusted by changing the mixture portion of part A and B.\n\nWe paint the markers on the elastomer using water transfer paper from Lazertran. We firstly print the designed marker patterns on the transfer paper using a laser printer, and then lay the printed side on the elastomer, and wet the paper from the back. The printed markers will transfer to the elastomer after we peel off the transfer paper.\n\nThere are three kinds of coating pigments that we use: the bronze flake and aluminum flake for semi-specular coating, and 1 $\\mathsf{\\mu}$m spherical aluminum powder for matte coating. Examples of the different coatings are shown in [Figure 4](#sensors-17-02762-f004){ref-type=\"fig\"}. The different coatings are good for different applications, and they require different illumination systems. There are two kinds of methods to make the coating: one is using the airbrush to spray the dissolved pigment on the elastomer base, and the other one is to directly brush the pigment powder on the elastomer base.\n\nThe semi-specular paint coating is made by airbrush. Then, we disperse the metal flake pigment in some silicone paint base (Cirius from ArtMolds) and organic solvent (such as toluene, d-limonine, methyl ethyl ketone (MEK)), and use an airbrush to evenly spray the solution on the elastomer surface, layer by layer. The airbrush can make the coating thin and even, but it takes a very long time for a fine piece of work, and we must pay close attention to removing the dust stuck in the coating. Brushing the pigment powder directly is much easier, but it only applies for pigment that comes as dry powders. Those powders will naturally stick on the silicone base. We use a make-up brush to evenly spread the powders on the surface, and then coat another thin protective silicone layer on the top. The simplest way to coat the layer is to pour the diluted fluid silicone on the painted elastomer.\n\n5. Evaluation {#sec5-sensors-17-02762}\n=============\n\nIn this section, we take the new compact GelSight sensor mentioned in \\[[@B29-sensors-17-02762]\\] as an example to evaluate the sensor's performance in estimating object shapes and contact force.\n\n5.1. Evaluation of Shape Measurement {#sec5dot1-sensors-17-02762}\n------------------------------------\n\nFor evaluating GelSight's geometry measurement, we firstly evaluate the sensor's precision in measuring the surface normal. We use the same images for calibration: the GelSight images of a small ball ($d = 3.96$ mm) pressed on the sensor. The ground truth of the surface normals can be calculated from the known geometry of the ball, and we measure the surface normal using the calibrated lookup table, as introduced in [Section 3.3](#sec3dot3-sensors-17-02762){ref-type=\"sec\"} and [Section 3.5](#sec3dot5-sensors-17-02762){ref-type=\"sec\"}. Examples of the results are shown in [Figure 13](#sensors-17-02762-f013){ref-type=\"fig\"}a. In the figures, we compared the measured values and the ground truth, of the pitch and yaw angles of the surface normal. In the $x - y$ plane of the sensor, the pitch angle can be considered as the geometry's spatial gradient, and the yaw angle can be considered as the gradient direction. The figures show that the measured value are correlated with the ground truth, especially when the gradient is low or medium; the measurement of yaw angle, or the gradient direction, is much more precisely measured. The figures also indicate that there is some spatial variance of the sensor's measurement, but can be accepted for an approximate measurement. [Figure 13](#sensors-17-02762-f013){ref-type=\"fig\"}b gives some examples of the reconstructed geometry of common objects. The ground truth of the geometry is hard to get, but the reconstructed 3D structures capture both the overall shape and local textures of the objects.\n\n5.2. Evaluation of Force Measurement {#sec5dot2-sensors-17-02762}\n------------------------------------\n\nThe GelSight sensors measure force and torque using the displacement field of the markers on the surface. The motion of the markers correspond to the lateral displacement of the sensor's surface, and the deformation of the elastomer can infer contact force. However, the relationship between the markers' motion and the contact force or shape is nonlinear when the contact geometry is unknown. For general situations, we use deep neural network to measure the contact force and in-plane torque from the GelSight images. In this section, we evaluate the fingertip GelSight sensor's ability in measuring force and torque with Convolutional Neural Network (CNN) in basic cases. In this preliminary experiment, we train the force measurement neural network with the data of GelSight contacting objects with some basic shapes, including spheres, cylinders and flat plane, and then test the network's performance on measuring the force when contacting other similar but unseen objects The forces and torque we try to measure is the normal force, shear force and direction, and in-plane torque (the torque along the *z*-axis in [Figure 14](#sensors-17-02762-f014){ref-type=\"fig\"}a).\n\nThe experimental setup is shown in [Figure 14](#sensors-17-02762-f014){ref-type=\"fig\"}a: a fingertip GelSight sensor is mounted on an ATI Nano-17 force/torque sensor on a fixed table, and we manually push or twist different objects against the GelSight sensor with different force amounts and directions, so that the force and in-plane torque on the GelSight surface is equal to the load on the ATI sensor, and we use the measurement from the ATI sensor as the ground truth. The objects include six balls (diameters from 12 mm to 87 mm), five cylinders (diameters from 10 mm to 70 mm), and two flat surfaces of different rigid materials. The total size of the training dataset is around 28,815. We only use the data in the loading process to reduce the influence of viscoelasticity.\n\nThe CNN model for measuring force and torque is adjusted from VGG-16 net \\[[@B34-sensors-17-02762]\\], pre-trained on the computer vision dataset ImageNet \\[[@B35-sensors-17-02762]\\]. We replace the network's last fully-connected layer with an output layer of four neurons, corresponding to the forces and torques in four axes (Fx, Fy, Fz, Tz). The input of the network is the three-channel difference image of the current GelSight image and the initial image, when nothing is in contact. We train the network with the mean squared error loss function for the regression problem. To test the model, we use the GelSight data of contacting three novel objects: a ball ($d =$ 25 mm), a cylinder ($d =$ 22 mm), and a flat plane. The test set contains 6705 GelSight images under different forces and torques.\n\nThe comparison between the output of the neural network and ground truth from the force/torque sensor is summarized in [Figure 14](#sensors-17-02762-f014){ref-type=\"fig\"}b--e. The coefficient of determination ($R^{2}$) and root mean square error (RMSE) for the results of three different objects are also listed in the figure. The plots show that the model output of forces and torques by GelSight sensor are correlated to the ground truth measured by the force torque sensor. For the force measurements, $R^{2}$ is higher than 0.9. The results also show that the GelSight measurement of force can be robust regardless of the geometry of the contact objects.\n\nHowever, for measuring the contact force and torque, the CNN methods have some deficiencies---that it can not well generalize the force-related information from the GelSight images---which costs most of the measuring error. On the one hand, the prediction from CNN highly relies on the training data, and to make robust measurement, the training data should contain the contact cases with objects of all shape categories and all contact conditions. On the other hand, for the given training set and test set, the measurement is still influenced by the contact geometry. We found that if we align the measurement in the chronological order, it well matches the ground truth qualitatively at all times, but the measurement at some entire contact sequences is worse than the others. In general, the GelSight images within the same contact sequence is of similar contact conditions, and the geometry is similar, while the CNN may make some biased measurement towards the certain contact geometry. In the long run, for making a good force measurement with GelSight, we should either collect a more comprehensive dataset (simulation methods could be applied), or choose some other methods that can better extract force information from GelSight images.\n\n6. Application {#sec6-sensors-17-02762}\n==============\n\nWith the high-spatial resolution of shape measurement, GelSight is proved to be effective in many robotic tasks that are very challenging for traditional tactile sensors.\n\nLi and Adelson \\[[@B38-sensors-17-02762]\\] show that GelSight can recognize materials from textures with very high precision. [Figure 15](#sensors-17-02762-f015){ref-type=\"fig\"}a gives some examples of the GelSight images when pressing on different fabrics. Although the fabrics are in the same object category with similar physical properties, GelSight can discriminate them by recognizing the high-resolution texture of the textile type. Additionally, because GelSight can measure the shapes of the objects, the geometry of the folds on the fabrics can be used to infer other properties such as thickness or stiffness.\n\nAt the same time, GelSight can sense other object properties as well. Yuan et al. \\[[@B37-sensors-17-02762],[@B39-sensors-17-02762]\\] show that GelSight can measure the hardness of objects from a dynamic image sequence when the sensor is pressed against the object. The sensor records how the shape and force of the sensor surface changes as the contact proceeds. Objects with different hardness may make different deforming patterns, since softer objects have more deformation in geometry in the contact process. [Figure 15](#sensors-17-02762-f015){ref-type=\"fig\"}b shows some examples. By using neural networks, the GelSight sensor can estimate hardness of the objects with unknown shapes, under a loosely controlled contact condition, while the contact force and trajectory are unknown.\n\nGelSight can assist with multiple manipulation tasks as well. As indicated in [Section 3.7](#sec3dot7-sensors-17-02762){ref-type=\"sec\"}, the GelSight sensor can estimate the slip and incipient slip state from the stretching of the surface. Additionally, it can detect slip by measuring the relative movement of the objects on the sensor. Dong et al. \\[[@B29-sensors-17-02762]\\] presents experiments of robot gripper successfully detecting or predicting slip from multiple cues when grasping different kinds of objects. Li et al. \\[[@B30-sensors-17-02762]\\] shows the usage of GelSight in the USB insertion task: the GelSight sensor measures the USB pattern on the connector, thus calculating the in-hand position of the connector. The robot then refines the manipulation position using the in-hand position, which increases the success rate of the inserting task. Izatt et al. \\[[@B40-sensors-17-02762]\\] further shows that GelSight can serve as a supplemental sensing modal to picture a much more precise point-cloud of the area within the grippers.\n\n7. Conclusions {#sec7-sensors-17-02762}\n==============\n\nTactile sensing is an important modality for robots, and one of the major challenges for robotic tactile sensing is to develop sensor hardware that can obtain adequate tactile information for multiple perception and manipulation tasks. We provide a possible solution by developing a novel tactile sensor, GelSight, which can measure high-resolution geometry, as well as information about force and shear. In this paper, we review the principles, algorithms, design, fabrication process and performance evaluation of some different versions of the GelSight sensor. Apart from measuring geometry and force, GelSight can also indicate other information, such as slip or incipient slip. The fingertip version of GelSight has been successfully applied on robotic grippers, and the new design makes the sensor fabrication and the data accessibility much more convenient. With the tactile information provided by GelSight sensor, a robot will be able to perform much better in multiple tasks related to both perception and manipulation.\n\nThis work was supported by grants from the National Science Foundation, the Toyota Research Institute, MIT Lincoln Labs, and NTT Communication Science Laboratories.\n\nW. Yuan organized the paper and wrote the manuscript; S. Dong performed the experiments for force measurement and proof-read the manuscript; E. Adelson revised the manuscript and ensured conciseness.\n\nE. Adelson holds shares of GelSight Inc. (Waltham, MA, USA).\n\n![(**a**) a cookie is pressed against the skin of an elastomer block; (**b**) the skin is distorted, as shown in this view from beneath; (**c**) the cookie's shape can be measured using photometric stereo and rendered at a novel viewpoint \\[[@B7-sensors-17-02762]\\] (Got copyright permission from IEEE).](sensors-17-02762-g001){#sensors-17-02762-f001}\n\n![(**a**) basic principle of the Gelsight and the desktop design introduced in \\[[@B7-sensors-17-02762]\\]. There are four main components for the GelSight sensor: an sensing elastomer piece with the opaque reflective membrane on top, supporting plate, LEDs which provide illumination, and the camera in the bottom to capture the shaded images under the illumination from different directions; (**b**) shows the picture of the sensor, and (**c**) shows the arrangement of the LEDs and camera when viewing from the top.](sensors-17-02762-g002){#sensors-17-02762-f002}\n\n![(**a**) an example pattern of printed markers on the GelSight. In the figure, the elastomer is for the fingertip GelSight sensor \\[[@B30-sensors-17-02762]\\] (Got copyright permission from IEEE), with the dimension of 25 mm \u00d7 25 mm \u00d7 2 mm, the markers have an average interval of 1.1 mm; (**b**) the markers make diverse motion field patterns under different kinds of forces or torques. The magnitude of the markers' motion is roughly proportional to the force/torque value.](sensors-17-02762-g003){#sensors-17-02762-f003}\n\n![Three kinds of elastomer coatings: semi-specular coating painted by bronze flake and aluminum flake paint, and matte coating by aluminum powder. In the second and third row, the three pieces of elastomer are pressed against a ball with diameter of 6 mm, but the in the 3rd row, the elastomer is illuminated by light from side direction.](sensors-17-02762-g004){#sensors-17-02762-f004}\n\n![The procedure of segmenting and locating the markers in the GelSight images. (**a**) the initial GelSight frame *Frm*~0~ when noting is in touch; (**b**) the low-pass Gaussian filtered image *I*~0~ from *Frm*~0~, where only the color background is left; (**c**) a GelSight frame *Frm* when the sensor is touching a cylinder; (**d**) the image *dI* after subtracting the initial background *I*~0~ from the current frame *Frm* (color is normalized for display purpose); (**e**) the mask of the markers after thresholding on *dI*. The red dots denotes the centroids of the markers.](sensors-17-02762-g005){#sensors-17-02762-f005}\n\n![Calibration images (part) with different GelSight sensors. During calibration, a ball/ball array is pressed on the sensor, and the image intensity change corresponds to the surface normal of the ball. (**a**) the calibration image of the side-illuminated GelSight device \\[[@B8-sensors-17-02762]\\] (Got copyright permission from Association for Computing Machinery), under the single-colored illumination from one direction; (**b**) the calibration image of the first fingertip GelSight sensor \\[[@B28-sensors-17-02762]\\] (Got copyright from MIT), where the illumination has four colors; (**c**) the calibration image of the improved fingertip GelSight \\[[@B29-sensors-17-02762]\\] (\u00a92017 IEEE), where the illumination has three colors. For (**b**,**c**), we press the ball at different locations, and take the average color change value to surface normal.](sensors-17-02762-g006){#sensors-17-02762-f006}\n\n![(**a**) when an indenter pressed on the GelSight surface in normal direction, the force is in linear relationship with the indenting depth, but the unloading curve is different from the loading curve due to the viscoelasticity; (**b**) when an indenter moves on GelSight in the shear direction, the shear force is initially proportional to the shear displacement of the indenter, but then grows slowly and comes to a constant value when partial slip or slip occur; (**c**) under the shear force, the average marker displacement is proportional to overall shear force, regardless of whether it is in the shear stage or slip stage.](sensors-17-02762-g007){#sensors-17-02762-f007}\n\n![The change of the marker displacement field with the increase of shear force. The degree of partial slip also increases as the force grows, and it starts from the peripheral contact area, which results an inhomogeneity in the marker displacement field. We can measure the inhomogeneity with the entropy of the markers' displacement, which keeps increasing. The experiment is conducted with a fingertip GelSight sensor \\[[@B30-sensors-17-02762]\\] whose surface is flat, (Figures adapted from \\[[@B27-sensors-17-02762]\\] (Got copyright permission from IEEE)).](sensors-17-02762-g008){#sensors-17-02762-f008}\n\n![The change of the marker displacement field and rotational angle with the increase of in-plane torque. The contact surface is flat. When the torque grows large, partial rotational slip or full slip occurs, starting from the peripheral area. The partial rotational slip results in the inhomogeneous distribution of the markers' rotational angle. The experiment is conducted with a fingertip GelSight sensor \\[[@B30-sensors-17-02762]\\] (Got copyright permission from IEEE) whose surface is flat, and the object in contact is a flat surface.](sensors-17-02762-g009){#sensors-17-02762-f009}\n\n![The GelSight device with matte lcoating and side illumination \\[[@B8-sensors-17-02762]\\] (Got copyright permission from Association for Computing Machinery). (**a**) the schematic of the optical system: six LEDs with a single color (white) is evenly distributed on the side of the supporting glass, and the sensor elastomer is in the central area of the supporting plate; (**b**) the bench setting of the design: a digital single-lens reflex (SLR) camera and macro lens is used to capture the high-resolution images on the sensor surface; (**c**) the captured image of human skin. (**d**) the reconstructed geometry of the skin from (**c**); (**e**) the portable design of the sensor.](sensors-17-02762-g010){#sensors-17-02762-f010}\n\n![Fingertip GelSight sensor introduced in \\[[@B28-sensors-17-02762],[@B30-sensors-17-02762]\\] (Got copyright permission from MIT and IEEE): the illumination has four colors (red, green, blue and white), and the entire sensing elastomer through the guiding plates are made of clear acrylic boards. The sensing elastomer coating in this design is the semi-specular coating. (**a**) picture of the sensor; (**b**,**c**) schematic of the sensor; (**d**) the parts of the sensor before assembling; (**e**) integration of the sensor into the Baxter Robot's gripper.](sensors-17-02762-g011){#sensors-17-02762-f011}\n\n![The new GelSight fingertip sensor introduced in \\[[@B29-sensors-17-02762]\\] (\u00a92017 IEEE). The sensor has three LED arrays of different color to illuminate the elastomer surface from a tilted angle of 71\u00b0, and a USB web cam to capture real-time image from the sensor. (**a**) the picture of the sensor; (**b**) the schematic diagram; (**c**) the parts of the sensor; (**d**) the sensor mounted on a WSG-50 parallel gripper (Weiss Robotics, Ludwigsburg, Germany).](sensors-17-02762-g012){#sensors-17-02762-f012}\n\n![Evaluation on the shape measurement of fingertip GelSight (Figures adapted from \\[[@B29-sensors-17-02762]\\] (\u00a92017 IEEE)). (**a**) the measurement of surface normal, when there the sensor is contacting a small ball. The plots compare the measured surface normal value and the ground truth, where the red line marks equal relation. The pitch can be considered as the magnitude of geometry gradient of the contact surface, and the yaw can be considered as the planar direction of the gradient; (**b**) examples of reconstructed 3D geometry when contacting different objects.](sensors-17-02762-g013){#sensors-17-02762-f013}\n\n![Evaluation on the force measurement of fingertip GelSight with simple but unseen objects. (**a**) experiment setup, where the GelSight is fixed on a ATI Nano-17 force-torque sensor, which is used to measure ground truth. In the experiments, we manually contact the GelSight surface with different objects and different loads. (**b**--**e**) experiment results of the force torque measurement with different unseen objects, including a ball, a cylinder and a flat plane. The gray dashed line denotes the 1:1 plot where the measurement meets the ground truth.](sensors-17-02762-g014){#sensors-17-02762-f014}\n\n![Examples of GelSight applications. (**a**) Li et al. \\[[@B30-sensors-17-02762]\\] (Got copyright permission from IEEE): a robot inserts a USB cable into the socket, while GelSight image helps the robot to locate the in-hand position of the USG plug; (**b**) GelSight can differentiate difference fabrics from their textures (images are enhanced for display purpose). The shape of the foldings can also reveal other material information, such as the thickness, stiffness \\[[@B36-sensors-17-02762]\\] (Got copyright permission from IEEE); (**c**) When pressing on a deformable object \\[[@B37-sensors-17-02762]\\] (Got copyright permission from IEEE), the GelSight sequence can infer hardness information from the change of the object shapes and contact force.](sensors-17-02762-g015){#sensors-17-02762-f015}\n\nsensors-17-02762-t001_Table 1\n\n###### \n\nThe minimum distinguishable force of the fingertip GelSight (Setup shown in Figure 14a, using the shape measurement and marker measurement respectively. The sensor elastomer has a neo-Hookean coefficient of 0.145 MPa. Since we use an ATI Nano-17 force/torque sensor to calibrate GelSight force measurement, and minimum distinguishable normal force for the setup is 0.05 N, we are not able to get finer measurement of the minimum force when using marker measurement.\n\n  Contact Surface Type       Rigid           Soft (Shore 00-10)                      \n  -------------------------- --------------- ----------------------- --------------- -----------------------\n  Contact area               30 mm${}^{2}$   Flat (\\>2 cm${}^{2}$)   30 mm${}^{2}$   Flat (\\>2 cm${}^{2}$)\n  Using shape measurement    \\<0.05 N        \\<0.05 N                \\<0.05 N        0.08 N\n  Using marker measurement   \\<0.05 N        \\<0.05 N                \\<0.05 N        \\<0.05 N\n"}
+{"text": "INTRODUCTION\n============\n\nIt is a fascinating thought that the single cell zygote contains all the information required for the development of the adult organism. Understanding how this information is encoded and deciphered is a major uncompleted scientific challenge. A group of genes known as homeobox genes has emerged as important master regulators of development. These genes have been highly conserved throughout evolution. They are expressed during embryonic development in a highly co-ordinated manner and continue to be expressed in virtually all tissues and organs throughout adult life.\n\nHomeobox (*Hox)* genes were discovered following the observation of two striking mutations in the fruit fly, *Drosophila melanogaster*. In the *antennapedia* mutation the antennae are changed into legs, whereas in the bithorax mutation, the haltere (a balancing organ on the third thoracic segment) is transformed into part of a wing. These changes were described as homeotic transformations from the Greek word homeosis, signifying a change of a complete body structure into another. *Drosophila* geneticists devised the term 'homeotic selector gene' to encapsulate the concept that a master regulatory gene could control the development of each segment of the fly. Subsequently *Drosophila* was found to contain a cluster of genes consisting of the *bithorax* complex with three homeobox genes (*Ubx, Abd-A*, and *Abd-B*) and the *antennapedia* complex with five homeobox genes (*Lab, Pb, Dfd, Scr and Antp*). The relationship between the chromosomal arrangement of *Hox* genes and the localisation of their expression was established by Lewis in 1978. In effect, these genes specify positional identity of the body segments of the fly along the anterior-posterior axis.[@b1]\n\nEVOLUTION OF *HOX* GENES\n========================\n\nHomeobox genes are present in the genomes of all animals which have so far been mapped as well as in the genomes of plants and fungi, indicating that the origins are ancient and precede the divergence of these kingdoms. Plants, fungi and unicellular animals do not, however, have clustered homeobox genes. Shortly after the origins of animals the primordial homeobox gene duplicated to form a protohox cluster of two genes which are still present in *cnidara* such as *hydra* ([Figure 1](#fig1){ref-type=\"fig\"}). Sponges do not have clustered homeobox genes, suggesting that this duplication occurred before the divergence of the parazoa. This is also reflective of the very simple body structure of sponges compared to other multicellular animals.\n\n![A representative dendrogram illustrating the evolution of *Hox* clusters. *Hox* gene clusters are thought to have developed by a process of duplication and divergence from a primordial homeobox gene estimated to have arisen about 1,000 million years ago.](umj7501-023-f1){#fig1}\n\nThe nematode *Caenorhabditis elegans* has a single cluster of at least five homeobox genes.[@b2] *Amphioxus* is a vertebrate-like chordate which has a notochord and segmental muscles derived from somites but does not develop a true vertebral column. It has only one *Hox* cluster which contains ten *Hox* genes and this cluster is regarded as being homologous to the ancestral cluster from which all vertebrate *Hox* clusters were derived. Two duplication events, early in vertebrate evolution, resulted in the four clusters seen in mammals and birds. Loss of some of the *Hox* genes in each cluster has also occurred with the result that not every primordial vertebral *Hox* gene is represented in each of the four clusters. Interestingly in some fish, such as zebrafish, a further duplication has occurred resulting in seven clusters.[@b3]\n\nHox GENES IN VERTEBRATES\n========================\n\nThe vertebrate counterparts of the *bithorax/antennapedia* cluster are the *Hox* genes, usually found in four clusters (reviewed by Duboule [@b4]). In man the four *HOX* gene clusters (A-D) are located on different chromosomes, at 7p15, 17q21.2, 12q13, and 2q31. Each cluster consists of 13 paralog groups with nine to eleven members assigned on the basis of sequence similarity and relative position within the cluster. A high degree of homology is evident between the human *HOX* genes and the *Hom-C* genes of *Drosophila*, ([Figure 2](#fig2){ref-type=\"fig\"}). Thus the human paralog groups 1--8 are more closely related to antennapedia (Antp), with groups 9--13 more closely related to *abdominal-B* (abd-B).\n\n![Conservation between the *HOM-C* and *HOX* gene clusters.\\\nThe four *Hox* gene clusters found in mammals are conserved from the *Drosophila Hom-C* complex in terms of nucleotide sequence and colinear expression. During embryonic development, the genes are expressed in a pattern that correlates with the chromosomal positioning, depicted here for human and mouse. The 3\u2032 genes are expressed both earlier and more anteriorly than the 5\u2032 genes.](umj7501-023-f2){#fig2}\n\nHox STRUCTURE\n=============\n\nMammalian *Hox* genes are small, each containing only two exons and a single intron which varies from less than 200 bases to several kilobases ([Figure 3](#fig3){ref-type=\"fig\"}). The homeobox is always present within the second exon in *Hox* genes and shows a high degree of homology among these genes, especially within paralog groups. The structures of non-Hox homeobox genes are more variable, frequently having the homeobox bridging an exon splicing site.\n\n![*HOX* gene/protein structure and mutations found in limb malformation.\\\n(A) *HOX* genes consist of two exons and one intron. Exon 2 contains a 180-nucleotide sequence, termed the homeobox, that encodes a 60-amino acid helix-turn-helix motif, termed the homeodomain, which has DNA-binding activity.\\\n(B) Mutations in *HOXA13* and *HOXD13* are found in disorders of limb formation, such as hand--foot--genital syndrome (HFGS), synpolydactyly (SPD), and brachydactyly.](umj7501-023-f3){#fig3}\n\nHox proteins have an acidic tail at the C-terminus and a pentamer upstream of the homeodomain that binds the TALE (three amino acid loop extension) proteins which act as cofactors. The homeodomain is a highly conserved motif of 60 amino acids. The function of the homeodomain was suggested by its similarity to the sequence of several prokaryotic gene regulatory proteins which contain a helix-turn-helix DNA binding motif. The homeodomain can be divided into three helical regions. Helix 3 contacts the major groove of DNA while helices 1 and 2 lie above the DNA.[@b5] Further contact of the homeodomain to the DNA is made by the sequence which proceeds helix 1, the N-terminal arm. The binding of Hox cofactors, (*exd* in *Drosophila*, *Meis* or *Pbx* in mammals) increases the stability of *Hox-*DNA binding.\n\nHox GENES AND DEVELOPMENT\n=========================\n\nThe order of expression of *HOX* genes within a cluster is co-ordinated during development, so that the low number, 3\u2032 genes, are expressed more anteriorly and earlier than the high number, 5\u2032 genes. During embryogenesis, cells require positional information to ensure that uncommitted cells differentiate into tissue appropriate for its location within the developing embryo. Thus groups of cells, known as functional domains, become committed to form body structures such as limbs and organs. There is growing evidence that it is the combination of *Hox* genes expressed within the functional domains along the AP axis which results in specifying the development of structures within these domains. The possible mechanisms by which this occurs have been reviewed by Kmita and Duboule.[@b6] In both Drosophila and man the spatial patterning corresponds to the relative position on the chromosome, thereby conforming to the \"principle of colinearity\".\n\nIn the developing vertebrate *Hox* genes are first expressed during early gastrulation at a stage when the embryo generates its major body axis.[@b7] In a pattern which correlates with the spatial expression of *Hox* genes, 3\u2032 genes are expressed earlier than 5\u2032 and as the embryo develops more progressively 5\u2032 genes are expressed. This pattern is termed \"temporal colinearity\" and is evident in other models of development such as haematopoiesis.\n\n*HOX* GENES AND LIMB DEVELOPMENT\n================================\n\n*Hox* genes define patterns of development in vertebrate limbs. In the chick, at least 23 *Hox* genes are expressed during limb development, with *Hoxa9* expressed in the proximal part of the limbs where the humerus or femur develop. *Hoxa9, Hoxa10* and *Hoxa11* are expressed in the forelimb where the radius and ulna (or tibia and fibula) develop. *Hoxa9* to *Hoxa13* are expressed in the wrist (or ankle) and the digits. A similar pattern of expression was found for the *Hoxd* genes whereas the expression of the *Hoxc* cluster was more complex. These observations illustrate that complicated networks of gene expression are involved in organ development, and suggest that functional redundancy among the *Hox* genes may mask the effects of under-expression or mutations in individual *Hox* genes. However a number of abnormalities in human limb formation have been described and recently these have been linked to specific *Hox* genes.\n\nSynpolydactyly (SPD), a rare, dominantly inherited limb malformation with a distinctive combination of syndactyly (fusion of digits) and polydactyly (extra digits), is caused by mutations in *HOXD13.* SPD typically consists of 3/4-finger and 4/5-toe syndactyly, with a duplicated digit in the syndactylous web. Affected family members often show variable expression of the disorder due to incomplete penetrance. The molecular basis of SPD was identified during a study of affected individuals in an isolated Turkish village.[@b8] The SPD locus was mapped to chromosome 2q31, where the *HOXD* gene cluster is located.[@b9] In normal individuals exon 1 of *HOXD13* contains an imperfect trinucleotide repeat sequence encoding a 15-residue polyalanine tract, and in subsequent studies each affected family displayed an expansion of this repeat, resulting in 7, 8 or 10 additional residues being expressed, see [Figure 3B](#fig3){ref-type=\"fig\"}.\n\nBrachydactyly, in which there is shortening of the digits, is rare in patients who are homozygous for SPD.[@b10] Two patients out of 128 screened for unselected congenital limb abnormalities requiring reconstructive surgery, were found to have a novel mutation within the *HOXD13* homeodomain (Ile314Leu). In further investigations specific mutations in *HOXD13* were linked with different combinations of limb disorders.[@b11]\n\nHypodactyly, a semi-dominant syndrome of loss of digit development, has been studied in mice. Animals with homozygous hypodactyly have a profound deficit in digital arch formation associated with a deletion in exon 1 in *Hoxa13*.[@b12] This leads to a translational frame-shift resulting in the loss of wild-type Hoxa13 protein and the production of a novel, stable protein in the limb buds of mutant mice. Mortlock and Innis have linked hypodactyly to a strikingly similar human disorder -- hand-foot-genital syndrome (HFGS), which differs from SPD because the deformities of the hands and feet are fully penetrant, bilateral and symmetrical, and uniform in their severity.[@b13]\n\nThe first *HOXA13* mutation associated with HFGS was a nonsense mutation in exon 2 which leads to the conversion of a tryptophan residue in the homeodomain to a stop codon, truncating the protein by 20 amino acids.[@b13] Some patients with HFGS also harbour expansions of the polyalanine tract of HOXA13, similar to those found in HOXD13 of SPD.[@b14],[@b15] A missense mutation in exon 2 is associated with an exceptionally severe form of HFGS.[@b14] Two unrelated boys had deletions at 2q24.1-q31 and 2q31.1-q32.2, regions that include *HOXD3* and *HOXD13*, associated with severe limb and genital abnormalities.[@b16] Other patients, in whom the entire *HOXD* cluster is deleted, have a mild SPD phenotype attesting to the inherent redundancy in the *HOX* gene network.\n\n*HOX* GENES AND LUNG DEVELOPMENT\n================================\n\nLung development is dependent upon the coordinated expression of a large number of genes in a manner tightly controlled both in time and space. Expression studies in fetal human and rodent lungs have demonstrated high expression of 3\u2032 *Hox* genes in clusters A and B.[@b17], [@b18] There is a marked decrease in expression of most of these genes as lung development progresses suggesting that they are involved in the early stages of lung morphogenesis, such as airway branching. However some *Hox* genes, for example *Hoxa5*, continue to be expressed at high levels throughout development and may be required for pulmonary maturation.[@b19]\n\nAbnormal expression of *HOX* genes is associated with several congenital lung abnormalities e.g. *HOXB5* is over-expressed in both bronchopulmonary sequestration [@b20] and congenital cystic adenomatoid malformation.[@b18] These disorders are characterised by deregulated patterns of morphogenesis in primordial lung tissue. Persistent high levels of *HOXB5* expression, beyond the early stages of lung development, result in primitive lung morphology. Altered patterns of *HOX* gene expression have also been demonstrated in several acquired disorders including emphysema, primary pulmonary hypertension and lung carcinomas.[@b21],[@b22]\n\nMurine models in which *Hox* gene expression has either been reduced or deleted provide strong evidence for the role of these genes in structural development of the respiratory system and regulation of pulmonary surfactant production. The degree of branching morphogenesis is decreased following reduction in *Hoxb5* levels by antisense oligonucleotides.[@b23] Furthermore, *Hoxa5* knock-out mice develop to full term but die in the early neonatal period due to tracheal occlusion, reduced expression of surfactant proteins and lung pathology similar to surfactant-deficient respiratory distress syndrome in preterm human neonates.[@b24]\n\n*HOX* GENES AND LEUKAEMIA\n=========================\n\nMultiple *HOX* genes of clusters A, B and C, but not D, are expressed in haematopoietic stem cells. Down-regulation or many *HOX* genes occurs as cells within a given lineage differentiate. For example, Care *et al*. demonstrated that peripheral T lymphocytes which were stimulated to proliferate using phytohaemagglutinin showed a rapid induction wave of *Hox* genes from *Hoxb1* to *Hoxb9*, *i.e.* in the 3\u2032 to 5\u2032 direction.[@b25]\n\nPerturbation of the process of cell differentiation by reciprocal chromosomal translocations can lead to the development of leukaemia. Such translocations lead to the creation of fusion genes, and may involve individual *HOX* genes or regulators of *HOX* gene activity. Thus translocations involving t\\[(7;11)(p15;p15)\\] or t\\[(2;11)(q31;p15)\\] have been described in which the *HOXA9* or *HOXD13* genes, respectively, are fused with the NUP98 nucleoporin gene in rare cases of acute myeloid leukaemia (AML). More frequently rearrangements of the mixed-lineage leukaemia gene *MLL1*, a positive regulator of cell specific *HOX* gene expression, have been found associated with aggressive acute leukaemias in both children and adults. Both types of translocation lead to gain of function, affecting the normal processes of differentiation of the pluripotent stem cells or the committed lymphoid or myeloid progenitors by deregulating the *HOX* gene expression patterns.\n\nRearrangements involving *MLL* and its 39 partner genes identified to date, are associated with approximately 5% of patients suffering from AML and 22% of those with acute lymphoblastic leukaemia (ALL).[@b26] To investigate the t\\[(11;19)(p22;q23)\\] translocation which gives rise to the MLL-ENL fusion protein, commonly found in infant acute leukaemias of both myeloid and lymphoid lineage, Horton *et al,* established a tetracycline-regulable system of MLL-ENL expression in primary haematopoietic cells.[@b27] Utilising a real-time quantitative PCR system [@b28] they were able to measure the expression of all 39 murine *Hox* genes and showed for the first time that reduced *Hox* gene expression is specific to loss of MLL-ENL and is not a consequence of differentiation. They concluded that MLL-ENL is required to initiate and maintain immortalisation of myeloid progenitors and may contribute to the development of leukaemia by aberrantly sustaining the expression of a *\"Hox* code *\"* consisting of *Hoxa4* to *Hoxa11*.\n\n*HOX* GENES AND CANCER\n======================\n\nNumerous studies have been undertaken to examine the differences in *HOX* gene expression between normal and neoplastic tissue, but the functional relationship with the malignant phenotype has remained elusive as reviewed by Abate-Shen.[@b29] Some investigators have explored the postulate that *Hox* genes expressed during embryogenesis but down-regulated during adult life are re-expressed in neoplasia-the so called \"oncology recapitulates ontology\" hypothesis. During embryogenesis a fine balance exists between cell proliferation and differentiation which is essential for normal development of the fetus. In contrast in cancer the balance between the two processes goes awry as reviewed by Grier *et al*.[@b30]\n\nNeoplastic growth in mammary epithelial cells is associated with increased expression of human growth hormone (hGH). Utilising human mammary carcinoma cells, Zhang and colleagues found that hGH production increased the expression and transcriptional activity of *HOXA1*.[@b31] Furthermore overexpression of *HOXA1* in mammary carcinoma cells resulted in up-regulation of Bcl-2, an anti-apoptotic factor, and increased total cell numbers. Interestingly *HOXA1* also enhanced anchorage-independent cell proliferation and caused oncogenic transformation of the cells, rendering them capable of aggressive tumour formation. Furthermore overexpression of *HOXA1* abrogated the response of the mammary carcinoma cells to daunorubicin. Taken together these observations serve to exemplify the effect of overexpression of a single gene *HOXA1*, and indicate that changes of expression of multiple *Hox* genes may substantially dysregulate cellular processes in neoplasia.\n\nEpithelial ovarian cancers (EOCs) arise from the simple epithelium lining the ovarian surface. Major EOC subtypes show morphological features that resemble m\u00fcllerian duct-derived epithelia of the reproductive tract. Recently Cheng and colleagues presented strong evidence that lineage infidelity of epithelial ovarian cancers is controlled by *HOXA* genes that specify regional identity in the reproductive tract.[@b32] They found that the *HOX* genes which normally regulate m\u00fcllerian duct differentiation are not expressed in normal ovarian surface epithelium, but are expressed in EOC subtypes according to the pattern of m\u00fcllerian-like differentiation of the cancers. Furthermore overexpression of *Hoxa9, Hoxa10* and *Hoxa11* gave rise to papillary tumours resembling serous, endometrioid-like and mucinous-like EOCs respectively. These observations support the contention that alteration of expression of genes in the *HOX* network that controls the patterning of the reproductive tract could explain the morphological heterogeneity of EOCs.\n\nCONCLUSIONS\n===========\n\nAttempts to understand the role of *HOX* genes in both normal and abnormal development and malignant transformation will be enhanced by the identification of their upstream regulators and downstream target genes. Whereas *MLL* fusion genes have provided some useful insights in the molecular mechanisms involved in leukaemogenesis much work remains to be done to identify specific gene products involved in the *HOX* network which might ultimately become feasible targets for therapeutic intervention. More research is also needed to explore the role of *HOX* genes in developmental processes.\n\nBased on an Invited Lecture \"Physiology and Pathophysiology of *HOX* genes in embryonic development\" given by Professor Terry Lappin to the British Association for Perinatal Medicine at their Annual Scientific Meeting in Belfast on 9th September, 2005.\n\nThis work was supported by the Research and Development Office of the Health and Personal Social Services in Northern Ireland, the Northern Ireland Leukaemia Research Fund and the Elimination of Leukaemia Fund. The authors have no conflict of interest.\n\nD[oes]{.smallcaps} the British public, or even the medical profession, fully realize that the nation owes that magnificent institution, the British Museum, to the liberality of a doctor? Its true begetter was Sir Hans Sloane, a fashionable physician of the eighteenth century... . Hans Sloane, who was of Scottish descent, was born in County Down in 1660. Even in boyhood he collected specimens, and the taste grew upon him till it became the ruling passion of his life. Natural history led him to medicine, which in those spacious days comprehended all science within itself... . With a rapidity that seems enviable to us whose professional lot lies in more arduous times, Sloane was elected a Fellow of the Royal Society in 1685, and was admitted to the Fellowship of the College of Physicians in 1687. In that year there came to him an offer to go to Jamaica as physician to the Duke of Albemarle, who had been appointed Governor of that island... . Within eighteen months the Duke died, and his physician\\'s nominal occupation was gone. Sloane\\'s real occupation, however, had been the gathering of materials for the museum which was his lifework. He returned to England in 1689, loaded with the spoils of his expeditions ... and became a highly prosperous physician. The Court and the aristocracy, we are told, had the \"greatest confidence in his prescriptions.\" Queen Anne took counsel of him; George the Second made him the keeper of the royal constitution; George the First had previously made him a baronet and appointed him Physician-General to the Army. The University of Oxford gave him its doctor\\'s degree in 1701, and he was President of the College of Physicians for sixteen years. He was appointed Secretary of the Royal Society in 1693, and succeeded Isaac Newton in the Presidency of that body in 1727... . Throughout his life Sloane went on adding to his museum, and he accumulated a vast collection, which included books, manuscripts, pictures, medals, and coins, as well as objects of natural history. He retired from practice in 1721, and died in 1753 at the age of 93 leaving in his will directions that his museum, which was valued at \u00a350,000, should be offered to the nation for the sum of \u00a320,000. The offer was accepted by Parliament, and the collection formed the nucleus of the British Museum, which was opened to the public in 1759. During the greater part of his professional life Sloane lived in Bloomsbury Square, close to the site of the future British Museum. Towards the end of his life he retired to Chelsea, where be had purchased a manor house and land, which is now covered by the stately mansions of the Cadogan estate. One of his daughters became the wife of the second Lord Cadogan, and the physician\\'s own name is perpetuated in Sloane Street and Hans Place. If Sloane was wealthy, he was also liberal. He gave the Apothecaries\\' Society their famous Physic Garden at Chelsea; he took part in the establishment of the Foundling Hospital, and he was never deaf to any deserving appeal made in the name of charity.\n\n(*BMJ* 1905;ii:1225)\n"}
+{"text": "Background {#Sec1}\n==========\n\nThe dome-shaped macula is a distinct entity characterized by convex elevation of the macula in high myopia \\[[@CR1]\\]. This elevation of the macula is often negligible on fundus examination or ultrasonography, but can be clearly detected via an optical coherence tomography (OCT) examination \\[[@CR1]\\]. Enhanced depth imaging (EDI) OCT examination revealed that the elevation is the result of a relatively thick sclera under the macula \\[[@CR2]\\]. In addition, eyes with dome-shaped macula exhibited thicker choroid than eyes those without \\[[@CR2]\\].\n\nAlthough tomographic features of dome-shaped macula have been investigated \\[[@CR2], [@CR3]\\], paucity of knowledge exists regarding the clinical significance of this peculiar entity. Several investigators suspected possible association of dome-shaped macula with foveal detachment \\[[@CR1], [@CR4]\\] and choroidal neovascularization \\[[@CR2], [@CR3]\\]. One previous study suggested that macular elevation may prevent retinoschisis and retinal detachment related to macular hole \\[[@CR5]\\]. A more recent study by Liang et al. suggested that a dome-shaped macula was associated with an increased incidence of foveal detachment and a decreased incidence of myopic foveoschisis \\[[@CR6]\\]. However, potential associations between ocular disorders and dome height have not yet been fully elucidated.\n\nEDI-OCT is a technique that facilitates better visualization of deep ocular tissue, including choroid and sclera \\[[@CR2], [@CR7], [@CR8]\\]. The purpose of the present study was to investigate the detailed structural profile of the dome-shaped macula using EDI-OCT and to evaluate its association with myopic macular disorders. We focused on the association between dome height and such disorders.\n\nMethods {#Sec2}\n=======\n\nThis retrospective observational study was performed at a single center according to the tenets of the Declaration of Helsinki. The study was prospectively approved by the Institutional Review Board of the Samsung Medical Center.\n\nWe retrospectively reviewed the medical records of patients with degenerative myopia. A computerized search was performed using the term \"myopi-\" and \"degen-\"in records of patients who were examined at the vitreo-retinal clinic of the Samsung Medical Center.\n\nTo be included in this study, all subjects were required to have undergone a comprehensive ophthalmologic examination that included measurements of best-corrected visual acuity (BCVA), manifest refraction, slit-lamp biomicroscopy, and fundus examination. In eyes that had undergone cataract surgery or refractive surgery, refractive error before the surgery was taken for analysis. Only eyes with greater than \u2212\u20096.0 diopters of myopia were included. The BCVAs were transformed to a logMAR (minimal angle of resolution) scale for analysis. All patients had also had fundus photographs and EDI-OCT examination.\n\nThe horizontal and vertical EDI-OCT crosshair scans centered at the center of the fovea were routinely conducted in those eyes using spectral domain OCT (Spectralis, Heidelberg Engineering GmbH, Heidelberg, Germany). The EDI image was obtained either via the conservative method by pushing the instrument close to the eye \\[[@CR8]\\] until year 2010 or via Spectralis enhanced depth imaging mode by pressing the conversion button provided in the Spectralis software afterwards. To improve visualization, the values of 50 to 100 scans were averaged for each section.\n\nThe scleral thickness and choroidal thickness was measured manually for all the included eyes via the Heidelberg Eye Explorer software (version 1.7.0.0). Since a previous study demonstrated the possible overestimation of tissue thickness in measurements based on 1:1 pixel images \\[[@CR9]\\], all the measurements were performed through 1:1\u2009\u03bcm images. Scleral thickness was defined as the distance between the inner and outer border of the sclera. The measurement line was drawn as a perpendicular line between the two borders. The mean value from horizontal and vertical scans was used for analysis. Choroidal thickness was defined as the distance from the hyperreflective line of the Bruch's membrane to the chorio-scleral interface. The measurement line was also drawn as a perpendicular line between Bruch's membrane and the chorio-scleral interface. The subfoveal choroidal thickness was measured based on both horizontal and vertical scans. The measurements were also performed at 500\u2009\u03bcm in the superior, inferior, nasal, and temporal directions from the foveal center. The average choroidal thickness of all six measurements was defined as the central choroidal thickness. When the choroidal thickness was immeasurable due to severe choroidal thinning or atrophy, the thickness was assumed to be 1\u2009\u03bcm.\n\nThe confirmative diagnosis of dome-shaped macula was made based on the result of height measurement. The height and diameter of the dome region were measured for eyes with dome-shaped macula. A line connecting the border of the dome-shaped sclera was drawn and the length of the line was defined as the diameter of the dome base (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). Another perpendicular line which starts at the top of the dome and ends at the measurement line for the diameter of the dome base was drawn. The length of this second line was defined as the height of the dome (Fig. [1](#Fig1){ref-type=\"fig\"}). The maximum height of the dome among the measurements based on horizontal and vertical scans was used for analysis. Two different cut-off values were used. Based on previous experience, we could identify the definite shape of dome-shaped macula on OCT when the height of dome exceeds approximately 50\u2009\u03bcm. Thus, we determined to diagnose dome-shaped macula when more than 50\u2009\u03bcm of the height of the dome was noted in either a horizontal or a vertical scan. A similar cut-off value was used in the previous studies \\[[@CR3], [@CR10]\\]. To evaluate possible differences in results due to cut-off values, 150\u2009\u03bcm was set as another cut-off value. Both cut-off values were used to compare the differences in incidence of macular disorders between eyes with and without dome-shaped macula. All the measurements were performed with the verification of two examiners (J.H.K. and J.K.). Additional analysis was performed to evaluate whether the 50\u2009\u03bcm cut-off value is appropriate to diagnose dome-shaped macula. To perform analysis, two examiners blinded to the height of the dome independently and intuitively evaluated the presence of dome-shaped macula based on OCT images. The presence of a dome-shaped macula was confirmed when the two examiners both noted the feature. The receiver operating characteristic curve for dome height was obtained, and the optimal cut-off value was determined using Youden's index. After that, the optimal cut-off value was compared with 50\u2009\u03bcm. Fig. 1Representative vertical optical coherence tomography of the 1:1\u2009\u03bcm setting of an eye with a dome-shaped macula. **a** Dome-shaped macula without inferior staphyloma (dome height\u2009=\u200985\u2009\u03bcm, dome diameter\u2009=\u20092167\u2009\u03bcm, subfoveal scleral thickness\u2009=\u2009442\u2009\u03bcm). **b** Dome-shaped macula with inferior staphyloma (dome height\u2009=\u2009234\u2009\u03bcm, dome diameter\u2009=\u20092716\u2009\u03bcm, subfoveal scleral thickness\u2009=\u2009561\u2009\u03bcm). Both images have a short black line (measurement line for dome height) and a long black dotted line (measurement line for dome diameter). The outer scleral border is marked by white arrowheads. Note obvious subfoveal scleral thickening in the eye with inferior staphyloma\n\nThe association of the height of the dome with the diameter of the dome, spherical equivalent, and BCVA was analyzed. The association of the diameter of the dome with the spherical equivalent and BCVA was additionally analyzed. The incidence of macular disorders, including choroidal neovascularization, myopic foveoschisis, and macular hole was compared in eyes with dome-shaped macula and eyes without dome-shaped macula. The diagnosis of myopic foveoschisis was made when retinoschisis involved the fovea. Subgroup analysis was performed for eyes with dome-shaped macula. In eyes with dome-shaped macula, the height of the dome and the diameter were compared between eyes with and without choroidal neovascularization, myopic foveoschisis, and macular hole.\n\nBased on the findings in the fundus photograph and vertical OCT scan, eyes were divided into dome-shaped macula with inferior staphyloma and dome-shaped macula without inferior staphyloma (Fig. [1](#Fig1){ref-type=\"fig\"}). The presence of inferior staphyloma was diagnosed when the steep slope of the retinal/choroidal image plane in the inferior direction was observed in the vertical section of the OCT image. The fundus photographs were used as adjunctive data to identify findings known to be associated with inferior staphyloma, such as inferonasal tilting of the disc, and degenerative changes evident in the inferior macula. The distribution of eyes within the two groups was compared in eyes with dome-shaped macula and eyes without dome-shaped macula. In eyes with dome-shaped macula, the subfoveal scleral thickness was additionally compared between eyes with and without inferior staphyloma.\n\nStatistical analyses were performed with the commercially available software package (SPSS ver. 18.0 for Windows; SPSS Inc., Chicago, IL, USA). Comparisons of scleral thickness and choroidal thickness were performed using independent samples *t* test. Prevalence of macular disorders between eyes with and without dome-shaped macula was compared using chi-square test or Fisher's exact test. The associations of the height and the diameter of the dome with the spherical equivalent and BCVA were analyzed using Pearson's correlation analysis. Comparisons of the height of the dome in eyes with and without macular disorders were performed using a Mann--Whitney *U* test. The distribution of eyes in the staphyloma group was compared using a Chi-square test. A *P*-value less than 0.05 was considered significant.\n\nResults {#Sec3}\n=======\n\nThe database included a total of 144 patients. Among them, 154 eyes of 99 patients met the inclusion criteria. The other patients were excluded for the following reasons: the refractive error was lesser than \u2212\u20096.0 D, preoperative refractive error could not be verified in cases with a history of cataract surgery or refractive surgery, or the EDI-OCT image was not available. Of those patients, eight eyes of six patients were additionally excluded because the scleral thickness could not be accurately measurable due to a relatively thick choroid. Eventually, data of 147 eyes of 93 patients were included in the result analysis.\n\nThe number of eyes with a dome with measurable height was 97. Among these, 70 eyes were intuitively diagnosed as having a dome-shaped macula on OCT. The distribution of the height of the dome in 70 eyes is shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a. The optimal cut-off value was 48.5\u2009\u03bcm, which was determined using the receiver operating characteristic curve and Youden's index (Fig. [2](#Fig2){ref-type=\"fig\"}b, area under the curve\u2009=\u20090.988, sensitivity\u2009=\u20090.971, specificity\u2009=\u20090.963). Through agreement analysis, 48.5 um was found to have almost perfect agreement when compared to the 50-um cut-off value we first used (Cohen's kappa coefficient: 0.933, *P*\u2009\\<\u20090.001). A confirmative diagnosis of dome-shaped macula with a dome height over 50\u2009\u03bcm was made in 60 eyes (40.8%) of 42 patients (24 horizontal scans and 55 vertical scans) with degenerative myopia. Fig. 2**a** Among the 97 eyes with a measurable dome, the bar graphs show the dome height distribution in the 70 eyes first diagnosed with the dome-shaped macula. **b** The receiver operating curve was used to obtain the optimal cut-off value (asterisk \\[\\*\\]\u2009=\u200948.5\u2009\u03bcm)\n\nEighteen patients exhibited bilateral dome-shaped macula. The mean age of the 42 patients was 55.1\u2009\u00b1\u200913.0\u2009years (mean\u2009\u00b1\u2009standard deviation). In eyes with dome-shaped macula, the mean spherical equivalent was \u2212\u200913.7\u2009\u00b1\u20094.8 and the mean BCVA was 0.54\u2009\u00b1\u20090.42. Subfoveal scleral thickness and central choroidal thickness was 401.2\u2009\u00b1\u200998.9\u2009\u03bcm and 37.6\u2009\u00b1\u200921.4\u2009\u03bcm, respectively.\n\nMyopic choroidal neovascularization, myopic foveoschisis, and macular holes were noted in 10 eyes (16.7%), 9 eyes (15.0%), and 6 eyes (10.0%), respectively. Retinal detachment related to a macular hole was observed in one of the 6 eyes with a macular hole. The height of the dome in that eye was 132\u2009\u03bcm. Among eyes with dome-shaped macula without other myopic macular complications (e.g., choroidal neovascularization, myopic foveoschisis, or macular hole), two of 35 eyes (5.7%) had macular serous detachment (i.e., foveal detachment). Table\u00a0[1](#Tab1){ref-type=\"table\"} compares characteristics of eyes with dome-shaped macula and eyes without dome-shaped macula. Table 1Comparisons of characteristics of the eyes with and without dome-shaped maculaCharacteristicEyes with dome-shaped macula\\\n(*N*\u2009=\u200960)Eyes without dome-shaped macula (*N*\u2009=\u200987)*P*-valueSpherical equivalent, diopters\u221213.7\u2009\u00b1\u20094.8\u221212.1\u2009\u00b1\u20093.90.022\\*Best-corrected visual acuity0.54\u2009\u00b1\u20090.420.43\u2009\u00b1\u20090.450.132\\*Subfoveal scleral thickness, \u03bcm401.3\u2009\u00b1\u200998.9315.7\u2009\u00b1\u200993.5\\<\u20090.001\\*Central choroidal thickness,\u03bcm37.6\u2009\u00b1\u200921.444.6\u2009\u00b1\u200927.30.082\\*Macular complications, n (%)\u2003Choroidal neovascularization10 (16.7%)20 (22.9%)0.409^\u2020^\u2003Myopic foveoschisis9 (15.0%)16 (18.4%)0.660^\u2020^\u2003Macular hole6 (10.0%)2 (2.3%)0.063^\u2020\u2020^\\* Statistical significance was tested using independent samples *t* test^\u2020^Statistical significance was tested using chi-square test^\u2020\u2020^Statistical significance was tested using Fisher's exact test\n\nDome-shaped macula was not noted in 87 eyes (59.2%) of 64 patients, and the mean age of patients was 58.1\u2009\u00b1\u200911.4\u2009years. In eyes without dome-shaped macula, the mean spherical equivalent was \u2212\u200912.1\u2009\u00b1\u20093.9 and the mean BCVA was 0.43\u2009\u00b1\u20090.45. Subfoveal sclera thickness and central choroidal thickness were 315.7\u2009\u00b1\u200993.5\u2009\u03bcm and 44.6\u2009\u00b1\u200927.3\u2009\u03bcm, respectively. Choroidal neovascularization, myopic foveoschisis, and macular holes were noted in 20 eyes (22.9%), 16 eyes (18.4%), and 2 eyes (2.3%), respectively. Although eyes with dome-shaped macula had a greater degree of myopia than the eyes without dome-shaped macula (*P*\u2009=\u20090.022), BCVA was not different between the two groups (*P*\u2009=\u20090.132). BCVA was not different between the two groups when analyzed in eyes without macular disorders (*P*\u2009=\u20090.692). The incidence of choroidal neovascularization, myopic foveoschisis, and macular holes was not different between the two groups (*P*\u2009=\u20090.409, *P*\u2009=\u20090.660, and *P*\u2009=\u20090.063, respectively). The subfoveal sclera was significantly thicker in eyes with dome-shaped macula (*P*\u2009\\<\u20090.001). However, difference in central choroidal thickness between the two groups was not significant (*P*\u2009=\u20090.082).\n\nIn comparison, using 150\u2009\u03bcm as a cut-off value, 17 eyes were diagnosed with dome-shaped macula. Among them, two eyes (11.8%) were found to have choroidal neovascularization and two eyes (11.8%) were found to have a macular hole. Myopic foveoschisis was not observed. The incidences of choroidal neovascularization (*P*\u2009=\u20090.525) and macular hole (*P*\u2009=\u20090.232) were not different between eyes with (*n*\u2009=\u200917) and without (*n*\u2009=\u2009104) dome-shaped macula. Although eyes with dome-shaped macula exhibited a relatively lower incidence of myopic foveoschisis than eyes without it, the difference was not significant (*P*\u2009=\u20090.078).\n\nIn the eyes with dome-shaped macula defined using a cut-off value of 50\u2009\u03bcm, the mean height of the dome was 126.5\u2009\u00b1\u200969.4\u2009\u03bcm (53\u2009\u03bcm to 345\u2009\u03bcm) and the mean diameter of the dome base was 2862.1\u2009\u00b1\u2009794.9\u2009\u03bcm (1567\u2009\u03bcm to 4886\u2009\u03bcm). There was a significant positive correlation between the height of the dome and the diameter of the dome base (*P*\u2009\\<\u20090.001, r\u2009=\u20090.470). Also, the height of the dome was positively correlated with subfoveal scleral thickness (*P*\u2009=\u20090.003, r\u2009=\u20090.378). However, the height of the dome was not correlated with BCVA or spherical equivalent (*P*\u2009=\u20090.489 and *P*\u2009=\u20090.479, respectively). Also the diameter of the dome base was not correlated with BCVA and spherical equivalent (*P*\u2009=\u20090.096, and *P*\u2009=\u20090.285, respectively). The association of the height of the dome and the diameter of the dome base with BCVA and spherical equivalent were not significant when analyzed in eyes without macular disorders (*P*\u2009\\>\u20090.05).\n\nIn the analysis of the eyes with dome-shaped macula, the heights of the dome in eyes with and without choroidal neovascularization were 127.7\u2009\u00b1\u200970.5\u2009\u03bcm and 126.2\u2009\u00b1\u200969.9\u2009\u03bcm, respectively. The values were 78.6\u2009\u00b1\u200920.6\u2009\u03bcm and 134.9\u2009\u00b1\u200971.6\u2009\u03bcm in eyes with and without myopic foveoschisis, respectively, and 164.0\u2009\u00b1\u2009104.6\u2009\u03bcm and 122.3\u2009\u00b1\u200964.4\u2009\u03bcm in eyes with and without macular holes, respectively. The height of the dome in eyes without myopic foveoschisis was significantly greater than in eyes with myopic foveoschisis (Table\u00a0[2](#Tab2){ref-type=\"table\"}, *P*\u2009=\u20090.009). The maximum height of the dome among the nine eyes with myopic foveoschisis was 114\u2009\u03bcm. The difference in the height of the dome was not significant in eyes with or without choroidal neovascularization and macular holes (*P*\u2009=\u20090.835 and *P*\u2009=\u20090.324). Again, the difference in the height of the dome was not significant in eyes with or without epiretinal membrane, vitreomacular traction, or inferior staphyloma (Table [2](#Tab2){ref-type=\"table\"}). In addition, the correlation of dome diameter with the presence of complications showed significance only in myopic foveoschisis (Table [2](#Tab2){ref-type=\"table\"}, *P*\u2009=\u20090.017). In eyes with and without dome-shaped macula, inferior staphyloma was noted in 12 eyes (20.0%) and 15 eyes (17.2%), respectively. The incidence of inferior staphyloma was not different in eyes with or without dome-shaped macula (*P*\u2009=\u20090.671), and prominent thickening of subfoveal sclera was noted regardless of the type of staphyloma. In eyes with dome-shaped macula, subfoveal scleral thicknesses in eyes with and without inferior staphyloma were 448.8\u2009\u00b1\u200994.5\u2009\u03bcm and 395.0\u2009\u00b1\u200998.7\u2009\u03bcm, respectively (*P*\u2009=\u20090.209). Table 2Comparisons of the height of the dome and the diameter of the dome between eyes with and without myopic macular complications, as the subgroup analysis of the eyes with dome-shaped maculaMacular complicationWith complication (n)Without complication (n)***p***-value\\***Comparisons of the height**\u2003Choroidal neovascularization127.7\u2009\u00b1\u200970.5\u2009\u03bcm (10)126.2\u2009\u00b1\u200969.9\u2009\u03bcm (50)0.835\u2003Myopic foveoschisis78.6\u2009\u00b1\u200920.6\u2009\u03bcm (9)134.9\u2009\u00b1\u200971.6\u2009\u03bcm (51)0.009\u2003Macular hole164.0\u2009\u00b1\u2009104.6\u2009\u03bcm (6)122.3\u2009\u00b1\u200964.4\u2009\u03bcm (54)0.324\u2003Epiretinal membrane110.8\u2009\u00b1\u200938.5\u2009\u03bcm (10)129.6\u2009\u00b1\u200973.9\u2009\u03bcm (50)0.439\u2003Vitreomacular traction111.9\u2009\u00b1\u200950.4\u2009\u03bcm (11)129.7\u2009\u00b1\u200972.9\u2009\u03bcm (49)0.446\u2003Inferior staphyloma128.3\u2009\u00b1\u200982.6\u2009\u03bcm (12)126.2\u2009\u00b1\u200968.3\u2009\u03bcm (48)0.941**Comparisons of the diameter**\u2003Choroidal neovascularization2877.1\u2009\u00b1\u2009419.9\u2009\u03bcm (10)2789.9\u2009\u00b1\u2009696.2\u2009\u03bcm (50)0.905\u2003Myopic foveoschisis2499.2\u2009\u00b1\u2009303.1\u2009\u03bcm (9)2969.3\u2009\u00b1\u2009645.7\u2009\u03bcm (51)0.017\u2003Macular hole2920.3\u2009\u00b1\u2009369.7\u2009\u03bcm (6)2896.5\u2009\u00b1\u2009652.8\u2009\u03bcm (54)0.605\u2003Epiretinal membrane2724.4\u2009\u00b1\u2009494.3\u2009\u03bcm (10)2933.7\u2009\u00b1\u2009650.1\u2009\u03bcm (50)0.415\u2003Vitreomacular traction2805.7\u2009\u00b1\u2009359.9\u2009\u03bcm (11)2919.8\u2009\u00b1\u2009674.8\u2009\u03bcm (49)0.748\u2003Inferior staphyloma2843.0\u2009\u00b1\u2009449.3\u2009\u03bcm (12)2912.8\u2009\u00b1\u2009668.5\u2009\u03bcm (48)0.824\\*Statistical significance was tested using Mann-Whitney *U* test\n\nDiscussion {#Sec4}\n==========\n\nThis study focused on identifying the association between dome-shaped macula and representative myopic complications. Dome-shaped macula was noted in approximately 40% of the eyes with degenerative myopia. In a manner consistent with the previous studies \\[[@CR1], [@CR5], [@CR6]\\], a dome-shaped macula was more frequently identified in vertical scans than in horizontal scans and the degree of myopia was greater in eyes with a dome-shaped macula than in those without. A definite association between dome-shaped macula and visual acuity was not verified. The subfoveal scleral thickness was significantly greater in eyes with dome-shaped macula than those without dome-shaped macula. However, there was no difference in choroidal thickness. Although the incidence of macular disorders was not different in the two groups, significantly lower incidence of myopic foveoschisis was noted in eyes with a relatively high dome.\n\nThe original definition of dome-shaped macula which was established by Gaucher and associates is a \"convex elevation of macula.\" \\[[@CR1]\\] There is no definite guideline on the exact degree of elevation in diagnosing dome-shaped macula. Coco and associates \\[[@CR5]\\] defined a height greater than 250\u2009\u03bcm of the dome as a \"clear intrusion,\" whereas more than 50\u2009\u03bcm of inward bulge of the retinal pigment epithelium in the OCT was defined as dome-shaped macula in the other studies \\[[@CR3], [@CR6]\\]. In the present study, dome-shaped macula was diagnosed when the height of the dome exceeded 50\u2009\u03bcm.\n\nThe height of the dome was relatively lower than that in previous studies \\[[@CR2], [@CR3], [@CR6]\\]. Possible explanations for this discrepancy are as follows: first, the difference in definition of dome-shaped macula may have an influence. In this study, only a small amount of scleral elevation was considered to be dome-shaped macula. We suspect that the definition of \"convex elevation of macula\" in previous studies might be stricter than our definition \\[[@CR2]\\]. Second, the possible disparity between measurements based on 1:1\u2009\u03bcm images and 1:1 pixel images may also have some influence. A previous study demonstrated the overestimation of tissue thickness via measurement based on horizontally compressed 1:1 pixel images when the measurement line is not vertical \\[[@CR9]\\]. Considering various scleral curvatures of posterior poles in eyes with degenerative myopia \\[[@CR11]\\], straight vertical measurement of the dome height may not be possible in many eyes. Measurements based on a 1:1\u2009\u03bcm setting might reflect the real height of the dome more accurately.\n\nIn the present study, two different types of dome-shaped macula have been noted. The first type is a classic dome-shaped macula that shows a radially symmetric shape (Fig. [1](#Fig1){ref-type=\"fig\"}a). The second type is associated with a variation in the inner scleral contour as a result of inferior staphyloma (Fig. [1](#Fig1){ref-type=\"fig\"}b). Maruko and associates evaluated the scleral thickness in eyes with tilted disc syndrome. The superior edge of staphyloma in the eyes included in that study was located at the fovea. In a manner similar to scleral thickening in eyes with a dome-shaped macula, the subfoveal sclera in the included eyes was thicker than that of other areas \\[[@CR12]\\]. The scleral profile in the representative OCT images presented by Maruko et al. was similar to that of second type dome-shaped macula of the present study. Although both types of dome-shaped macula in the current study exhibit a thickening of the subfoveal sclera \\[[@CR2]\\], it is not certain whether these two types of dome-shaped macula share the same etiology. In the current study, the height of the dome and subfoveal scleral thickness was not different between eyes with and without inferior staphyloma. This indicates that the presence of inferior staphyloma may not influence the shape of the dome. The possible difference in the influence of retinal and choroidal profiles between the two types of dome-shaped macula warrants further investigation.\n\nDome-shaped macula has been associated with visual impairment \\[[@CR1], [@CR13]\\]. In the present study, there was no difference in BCVA between eyes with a dome-shaped macula and eyes without, even in eyes without macular disorders. Also, the association between BCVA and the height of the dome was not significant in eyes with dome-shaped macula.\n\nPreviously, Coco and associates postulated that a dome-shaped macula may prevent retinoschisis at the top of the bulge \\[[@CR5]\\]. In a study by Liang et al., the incidence of fovea-involving retinoschisis was significantly lower in eyes with a dome-shaped macula than in eyes without \\[[@CR6]\\]. In the present study, we first investigated whether dome height is associated with retinal disorders. We found that dome height and dome diameter were significantly lower in eyes with myopic foveoschisis. The similarity in this result among these two parameters is likely due to the positive correlation between dome height and dome diameter. It is likely that dome-shaped macula has some valid role in preventing myopic foveoschisis probably by releasing the traction on the fovea when the dome is high enough. The maximum height of the dome among the eyes with myopic foveoschisis was 114\u2009\u03bcm. Thus, it is likely a dome height greater than 114\u2009\u03bcm may have some preventive effect on the development of myopic foveoschisis. Further studies with a different cohort are warranted to confirm the finding. There were discrepancies between the results of the current study and those of the previous studies regarding the association between dome-shaped macula and other myopic macular complications. In the present study, the mean height of the dome was not different between the eyes with or without choroidal neovascularization. However, in the study by Ellabban and associates, the mean height of the dome was significantly lower in eyes with choroidal neovascularization than in eyes without it \\[[@CR3]\\]. The incidence of choroidal neovascularization was relatively high, and ranging up to about 40%, in the eyes in their study. We suspect that the eyes included in the study and the definition of choroidal neovascularization would be different in the two studies. Coco and associates hypothesized that dome-shaped macula may also prevent retinal detachment related to macular holes based on the result that none of their 3 patients with macular holes developed retinal detachment \\[[@CR5]\\]. However, our finding does not support this postulation. The history of retinal detachment related to macular holes was noted in one eye with 132\u2009\u03bcm of height of dome. Dome-shaped macula may have some effect similar to a macular buckling procedure by pushing the retina towards the vitreous cavity. However, the effect of dome-shaped macula is likely limited because all our patients had no more than 350\u2009\u03bcm of height of the dome whereas representative OCT images after macular buckling procedure \\[[@CR14], [@CR15]\\] show that the elevation induced by buckling is greater than several times the retinal thickness. A further large-scale study may provide a clear answer for the association of the height of the dome with the incidence of choroidal neovascularization, macular holes, and retinal detachment.\n\nThere has been some debate regarding the definition of dome-shaped macula. Beyon and Chu believe that the dome-shaped macula may be a feature of a complicated staphyloma type or a variant of inferior staphyloma and suggested to use the term \"sclera compression maculopathy \\[[@CR16]\\].\" However, Spaide and Imamura claimed that thickening of the sclera under the fovea is a unique finding distinct from ordinary staphyloma \\[[@CR17]\\]. More recently, Coco and associates classified dome-shaped macula as a posterior staphyloma and inferior staphyloma groups. Posterior staphyloma was noted in 48 of 68 eyes (70.6%) whereas inferior staphyloma was noted in the other 20 eyes (29.4%) \\[[@CR5]\\]. However, the possible difference in scleral structural profile between the two groups was not investigated because they used a previous version of OCT that had limited penetration into the deep tissue. In the present study, the incidence of inferior staphyloma was not different between the eyes with dome-shaped macula and the eyes without dome-shaped macula, and a prominent thickening of the subfoveal sclera was noted regardless of the type of staphyloma. In eyes with dome-shaped macula, we could verify relative thickening of sclera at the subfoveal lesion regardless of the type of staphyloma. The mean subfoveal scleral thickness of eyes without inferior staphyloma was 395\u2009\u03bcm, approximately 80\u2009\u03bcm thicker than the thickness of eyes without dome-shaped macula. We believe that these findings may support the suggestion made by Spaide and Imamura \\[[@CR17]\\]. However, one interesting finding is that the mean subfoveal scleral thickness of eyes with inferior staphyloma was approximately 54\u2009\u03bcm thicker than eyes without inferior staphyloma in eyes with dome-shaped macula. Although the difference was not statistically significant, an association between inferior staphyloma and dome-shaped macula could be suspected.\n\nThick choroid was suggested as one of the possible reasons for dome-shaped macula \\[[@CR1]\\]. According to the report by Imamura and associates, the choroidal thickness was minimally thicker in eyes with dome-shaped macula when compared to that of eyes without dome-shaped macula \\[[@CR2]\\]. The possible reason of the variation in elevation would be of scleral origin. Eyes with dome-shaped macula exhibited relatively thinner choroid than eyes without dome-shaped macula in the present study. We often observed extreme thinning of the choroid at the top of the bulge in eyes with dome-shaped macula. In the present study, eyes with dome-shaped macula had a relatively greater degree of myopia than eyes without dome-shaped macula. In our experience, there is a considerable regional variation in choroidal thickness around the subfoveal area in some of the eyes with degenerative myopia. For example, since eyes with degenerative myopia usually exhibit a very thin choroid, increase in the subfoveal choroidal thickness could be noted when a large choroidal vessel was located at subfoveal area. This variation may partially influence the different results within the studies. One recent study suggested an association between thin choroid and foveal serous retinal detachment in eyes with inferior staphyloma \\[[@CR18]\\]. Further studies to verify the influence of dome-shaped macula on choroid would be of value.\n\nIn the present study, macular serous detachment (foveal detachment) without other macular complication was identified only in two eyes (5.7%). There was a considerable variation in the incidence of foveal detachment associated with dome-shaped macula in the studies. Although rates of 52.1 to 66.7% incidence were reported by several investigator groups \\[[@CR1], [@CR13], [@CR19]\\], the incidence was only 1.8 to 10.2% in other studies \\[[@CR3], [@CR5], [@CR6], [@CR10]\\]. Liang et al. reported the incidence of 1.8% in their study \\[[@CR6]\\]. The exact reason for this variation is not certain. One possible explanation is the different subjects' characteristics, including ethnic differences among the studies. The incidence was relatively low in Asian subjects \\[[@CR3], [@CR6], [@CR10]\\]. In addition, it has been reported that foveal detachment in dome-shaped macula may spontaneously resolve \\[[@CR20], [@CR21]\\]. It is possible that spontaneous resolution accounts for the differences in frequency observed. To date, none of the previous studies evaluated the true population-based prevalence of serous detachment in eyes with dome-shaped macula. An accurate prevalence rate requires confirmation based on further population-based studies.\n\nIn the study of Ellabban et al., the three-dimensional shape of the dome-shaped macula was evaluated \\[[@CR3]\\], and classified into two subtypes. One was a band-shaped ridge within the staphyloma and the other was a dome-shaped convexity within the staphyloma. In the former subtype, a flat RPE line was usually observed in horizontal OCT scan images, whereas a convex configuration was observed on both vertical and horizontal OCT scan images in the latter subtype. In some patients, different subtypes of dome-shaped macula were observed in both eyes. In the present study, convex elevation of sclera was more frequently noted in vertical OCT scans. In approximately half of the patients, the elevation was noted in only vertical OCT scans, suggesting these patients may have had a 'band-shaped ridge within the staphyloma' subtype of dome-shaped macula.\n\nThe definition of dome-shaped macula has not yet been firmly established. Several different definitions of dome-shaped macula, including the convex elevation of macula \\[[@CR1], [@CR2]\\], an inward bulge in the macular retina, retinal pigment epithelium, and choroid \\[[@CR5]\\], or more than 50\u2009\u03bcm of elevation of the inward bulge \\[[@CR3], [@CR10]\\] have been used by different investigator groups. However, these definitions were merely arbitrary and the proper threshold value for diagnose dome-shaped macula has not been investigated. In the present study, we attempted to verify whether the cut-off value 50\u2009\u03bcm is appropriate to diagnose dome-shaped macula. In our study, 85.7% of eyes with dome heights between 50 and 60\u2009\u03bcm were diagnosed to have dome-shaped macula, whereas the proportion was only 61.5% in eyes with dome heights between 40 and 50\u2009\u03bcm. Although the proportion was 100% in cases with dome heights of at least 60\u2009\u03bcm, using this cut-off value may miss cases with smaller dome-shaped maculae. The optimal cut-off value obtained using the receiver operating characteristic curve was 48.5\u2009\u03bcm and almost perfect agreement was obtained when compared to the 50-\u03bcm cut-off value used in our study, i.e., a value of 50\u2009\u03bcm may be appropriate for diagnosis.\n\nThis study has several limitations. First, in addition to its retrospective nature, patients were included based on a computerized search using several specific terms. Thus, not all patients who might have met the inclusion criteria were included. Second, because this study was performed at a tertiary referrer center, it is possible that the characteristics of dome-shaped macula found in this study may be partially different from characteristics of dome-shaped macula in ordinary degenerative myopia.\n\nConclusion {#Sec5}\n==========\n\nIn conclusion, localized thickening and convex elevation of sclera were observed in eyes with dome-shaped macula, and its prevalence was about 40% among people with degenerative myopia. The presence of a high and large dome seemed to have a preventive effect on the development of myopic foveoschisis. Further studies may reveal a more detailed profile of dome-shaped macula and its clinical significance.\n\nOCT\n\n:   Optical Coherence Tomography\n\nEDI\n\n:   Enhenced depth imaging\n\nBCVA\n\n:   Best-corrected visual acuity\n\nlogMAR\n\n:   Minimal angle of resolution\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nGeun Woo Lee and Jae Hui Kim contributed equally to this work.\n\nNot applicable.\n\nGWL and JHK designed the study, prepared the data, analyzed data, and wrote the paper JK prepared the data. SWK (the corresponding author) designed the study, analyzed data. All authors have read and approved the manuscript.\n\nNo funding was received for the work.\n\nThe data used to support the findings of this study are available from the corresponding author upon request.\n\nThis study was approved by the Institutional Review Board of the Samsung Medical Center, which waived the written informed consent because of the study's retrospective design and was conducted in accordance with the tenets of the Declaration of Helsinki.\n\nNot applicable.\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nOne of the most fundamental components of health care quality is the patient safety \\[[@B1]\\]. The patient safety is a priority for every health care system which follows the ensuring and improving of the quality of health care \\[[@B2]\\] and also is one of the main concerns of all health care systems \\[[@B3], [@B4]\\]. The principle of \"no harm\" in the Hippocratic Oath is a confirmation of this issue \\[[@B5]\\]. The patient safety is known as to avoid injuries to the patients or occurring unexpected adverse events of health care processes \\[[@B6]\\].\n\nAdverse events and medical errors are the main factors endangering the patient safety which are the most important problems of all health care systems, and all of these systems try to reduce their resulted injuries \\[[@B5]\\]. Statistics show that from 2.9% to 16.6% of hospital inpatients suffer from treatment-related adverse events \\[[@B7], [@B8]\\]. On the other hand, the rate of medical errors throughout the world is so high that is accounted for one of the five major causes of deaths \\[[@B9]\\].\n\nOne of the most common types of medical errors is medication error which because of their high prevalence and potential risks to patients, the rate of medication errors is considered as an indicator to determine the level of the patient safety in hospitals \\[[@B3], [@B9], [@B10]\\]. Medication error has been defined as the preventable misuse of the drugs \\[[@B11], [@B12]\\].\n\nIn the patient safety management issues, from the 1990s onwards, determining the factors influencing the occurrence of errors as well as the barriers to reports on occurred errors has been focused instead of only detection and prevention of errors \\[[@B13]\\]. The occurrence of medication errors among nurses is more than that among other health care professionals \\[[@B14]\\]. The results of studies have shown that among approximately 44000 to 98000 annual deaths due to medical errors, 7000 deaths have been related to the medication errors \\[[@B9]\\]. Reporting on the medication errors can cause maintaining the safety of patients as well as providing a valuable database of information to prevent medication errors in the future \\[[@B15]\\].\n\nSeveral studies have been conducted on determining the barriers to report on medication errors. Tol and colleagues in their study on a group of nurses conducted at a teaching hospital concluded that the managerial factors were the most important barriers to report on medication errors \\[[@B16]\\]. The results of Wakefield and colleagues\\' study on 1300 nurses showed that using medication errors for reprimanding nurses by their supervisors was one of the factors affecting the refusal of reporting on medication errors \\[[@B17]\\]. Based on the results of a study conducted in Japan, the main cause of occurring medication errors among newly graduated nurses was related to the wrong administration of intravenous drugs which, in turn, was due to their little knowledge of pharmacology \\[[@B18]\\].\n\nThe researchers of the present study in their searches found that most of the researches had been focused only on identifying the causes of medication errors and few researches had conducted to determine the factors influencing the refusal of reporting on medication errors or barriers to reporting on these errors. Because determining the factors influencing not reporting on the medication errors can result in taking appropriate managerial and individual measures to increase the rate of reporting on medication errors and finally decrease the occurrence of these errors, the present survey aimed to study the factors influencing not reporting on medication errors from the nurses\\' viewpoints in Abbasi Hospital of Iran.\n\n2. Methods {#sec2}\n==========\n\nThis was a cross-sectional, descriptive analytical study conducted in 2012 in which all nurses (*n* = 100) working in different inpatient units of Abbasi Hospital in Miandoab, an Iranian hospital affiliated to Urmia University of Medical Sciences, were studied using a consensus method.\n\nRequired data were collected using a questionnaire consisting of two sections: the first section included nurses\\' demographic characteristics such as age, sex, marital status, education level, employment status, types of work shifts, experience, and their service unit. The second section, also, included 19 questions in 3 domains: fear of the consequences of reporting (with 11 items), managerial factors (with 5 items), and factors related to the process of reporting (with 3 items). The five-point Likert scale was used to determine the factors affecting the refusal of reporting on medication errors, whereby 1 refers to strongly disagree and 5 as strongly agree. Though the validity and reliability of this questionnaire had been confirmed in Tol and colleagues\\' study \\[[@B16]\\], in the present study, the reliability of this questionnaire was approved again (*\u03b1* = 0.82). For collecting data, one of the researchers attended the mentioned hospital in three work shifts---morning, evening, and night---and distributed the questionnaire among studied nurses after obtaining oral consent from them for participating in this study and explaining the objectives of the study and assuring the confidentiality of collected data.\n\nFinally, collected data were analyzed using SPSS 19.0 through some statistical tests including independent *t*-test, ANOVA, and chi-square. *P* \\< 0.05 was considered statistically significant.\n\n3. Results {#sec3}\n==========\n\nThe response rate was 83%. The results showed that among 83 nurses who participated in this study, most of them (*n* = 65, 78.3%) were female, 71% were married (*n* = 59), 55.5% were in the 25--35 age group (*n* = 46), 86.8% had a bachelor\\'s degree (*n* = 72), 49.4% were employed officially (*n* = 41), 38.6% had lower than 5 years job experience (*n* = 32), 88% were working in rotating work shifts (*n* = 73), and 39.7% were working in the Intensive Care Units (*n* = 33) ([Table 1](#tab1){ref-type=\"table\"}).\n\nThe mean self-reported scores of studied domains for not reporting on medication errors were as follows.\n\nFear of the consequences of reporting (3.01 \u00b1 1.039), managerial factors (3.56 \u00b1 0.99), and factors related to the process of reporting (3.32 \u00b1 0.79). Therefore, managerial factors were the most important reason for not reporting on the medication errors.\n\nThe results showed that the highest mean score in the managerial factors domain was related to \"the heads\\' focus only on finding the culprits and blaming them, regardless of other factors involved in the occurrence of errors\" (3.674 \u00b1 1.21). The highest mean scores in the domain of fear of the consequences of reporting was related to \"fear of judicial affairs following reporting on medication errors\" (3.68 \u00b1 1.21), and in the domain of factors related to the process of reporting was associated with \"lack of a clear definition of medication errors\" (3.144 \u00b1 1.29).\n\nThe frequency, mean, and standard deviation of nurses\\' responses to the questions related to each studied domains (fear of the consequences of reporting, managerial factors, and factors related to the process of reporting) and their items have been shown in [Table 2](#tab2){ref-type=\"table\"} based on their importance in occurring medication errors from the nurses\\' viewpoints.\n\nAlso, the results showed that there were not any significant relationships between some nurses\\' demographic characteristics (such as marital status, types of work shifts, age, education level, employment status, and service units) and the studied domains of not reporting on medication errors (including fear of the consequences of reporting, managerial factors, and factors related to the process of reporting) (*P*-value \\>0.05). However, there was a significant relationship between employment status and fear of the consequences of reporting on medication errors (*P* \\< 0.008).\n\n4. Discussion {#sec4}\n=============\n\nIdentification and reduction of medication errors requires a system to be designed for finding the root causes of occurring them. Based on such a system, first the occurred errors should be reported, and then, the major causes of occurring them should be found in an intimate and scientific situation. In this regard, the present survey aimed to study the factors influencing not reporting on medication errors from the nurses\\' viewpoints in Abbasi Hospital, Miandoab.\n\nThe study results showed that managerial factor was the most one causing not reporting on medication errors, and other factors including factors related to the process of reporting and fear of the consequences of reporting had the next priorities for not reporting on medication errors from the viewpoint of nurses. Tol and colleagues in a similar study conducted among the nurses of Baharloo hospital in Tehran also found the same results \\[[@B16]\\]. The similarity of these two studies in using the same methods and population studies (viz., the nurses of public hospitals) is a reason for those similar results. Kouhestani and Baghcheghi in their study on the nursing students reported the fear of the consequences of reporting errors as the most important factors of not reporting on medication errors by them \\[[@B19]\\]. The different population study of mentioned studies (viz., nursing student and nurses) can be a reason for their different results. In another study carried out by Osborne and colleagues the most important reasons for not reporting on medication errors were fear of being blamed, fear of being labeled as incompetent nurses and inadequacy, fear of their future professional career, fear of judicial issues, and adverse reactions of their heads and colleagues \\[[@B20]\\].\n\nAccording to the results of the present study, among the studied managerial variables, \"the heads\\' focus only on finding the culprits and blaming them, regardless of other factors involved in the occurrence of errors,\" had the greatest impact on not reporting on medication errors from the viewpoints of nurses. The results of the Tol and colleagues\\' study \\[[@B16]\\] as well as Hosseinzadeh and colleagues\\' study \\[[@B21]\\] confirm the present study results.\n\nIn the current study, unlike other researches in this field \\[[@B3], [@B15], [@B17], [@B22]\\], the factor of fear of the consequences of reporting had approximately the same mean score as the factor of managerial. Among the studied variables in the factor of fear of the consequences of reporting, \"fear of judicial issues following reporting on medication errors\" was introduced as the most important variable influencing not reporting on medication errors.\n\nOne of the new trends that are emerging today and should be paid particular attention is patients\\' litigation against possible and unwanted medical errors and malpractices throughout the diagnosis and treatment processes \\[[@B23]\\]. Kouhestani and Baghcheghi in their study had reported \"fear of the impact of reporting of errors on the students\\' evaluation scores and their educational outcomes\" as the most important variable in the factor of fear of the consequences of reporting which does not confirm the present study results. The difference between the population studies of these two studies can be a reason for the results differences \\[[@B19]\\]. However, the results of Tol and colleagues\\' study confirm this result of the present study \\[[@B16]\\].\n\nAmong the studied variables related to the process of reporting, \"lack of a clear definition of medication errors\" was the most important variable influencing not reporting on medication errors from the studied nurses\\' viewpoints. Hosseinzadeh and colleagues showed similar results in their study \\[[@B21]\\]; however, the results of Tol and colleagues\\' study \\[[@B16]\\] as well as Kouhestani and Baghcheghi\\'s study \\[[@B19]\\], which had introduced \"not paying attention to the reporting on some medication errors\" as the most important reason for not reporting on medication errors, do not confirm the present study results in the mentioned domain.\n\nNurses are the second group protecting against medication errors because they play an important role in the distribution, preparation, and administration of medication orders \\[[@B24]\\].\n\nNurses\\' proper and appropriate function of reporting on medication errors will prevent doing possible harm to the patients and also is considered as a source of valuable information to avoid occurring similar errors in the future and finally can dramatically provide safety for patients \\[[@B15]\\]. In other words, it can be said that as a rule, making errors is inevitable; however, reporting them and following their subsequent harm can prevent recurrence of those errors for the other patients \\[[@B25]\\].\n\nThe present study had some limitations such as studying only on one hospital and a small sample as well as its limitation of results generalizability. Therefore, it is suggested that similar studies should be carried out on other public and also private hospitals using large samples and population studies, and then those results should be compared with the present study results. Also, future studies should be conducted using interventional designs to identify the major causes of occurring medication errors, other reasons for not reporting them, and strategies to prevent or reduce their occurrence.\n\n5. Conclusion {#sec5}\n=============\n\nThe current study results showed that managerial factors had the greatest role in the refusal of reporting on medication errors. Therefore, designing a system for reporting on medication errors properly and accurately, training nurses in the quality of reporting on medication errors, and above all, establishing a mechanism to improve quality rather than focus only on finding the culprits and blaming them can result in more reporting on medication errors, reducing their occurrence, and, finally, improving patient safety.\n\nThe authors would like to thank the hospital\\'s heads and nursing staff of Abbasi Hospital for their kind cooperation with the researchers in collecting and analyzing the data.\n\n###### \n\nThe demographic characteristics of studied nurses.\n\n  ---------------------------------------\n  Variables               Frequency (%)\n  ----------------------- ---------------\n  Age (year)              \u2009\n\n  \u2003Less than 25           10 (12%)\n\n  \u200325--35                 46 (55.5%)\n\n  \u200335--45                 15 (18.1%)\n\n  \u2003More than 45           6 (7.2%)\\\n\n  \u2003Missing                6 (7.2%)\\\n\n  Sex                     \u2009\n\n  \u2003Male                   17 (20.5%)\n\n  \u2003Female                 65 (78.3%)\n\n  \u2003Missing                1 (1.2%)\n\n  Marital status          \u2009\n\n  \u2003Single                 23 (27.8%)\n\n  \u2003Married                59 (71%)\n\n  \u2003Missing                1 (1.2%)\n\n  Education level         \u2009\n\n  \u2003Diploma                7 (8.4%)\n\n  \u2003Associate degree       4 (4.8%)\n\n  \u2003Bachelor\\'s degree     72 (86.8%)\n\n  Employment status       \u2009\n\n  \u2003Official employee      41 (49.4%)\n\n  \u2003Contractual employee   29 (34.9%)\n\n  \u2003Other                  12 (14.5%)\n\n  \u2003Missing                1 (1.2%)\n\n  Types of work shift     \u2009\n\n  \u2003Fixed shifts           10 (12%)\n\n  \u2003Rotating shifts        73 (88%)\n\n  Job experience (year)   \u2009\n\n  \u2003Less than 5            32 (38.6%)\n\n  \u20035--10                  27 (32.5%)\n\n  \u2003More than 10           17 (20.5%)\n\n  \u2003Missing                7 (8.4%)\n\n  Service unit            \u2009\n\n  \u2003General Care Units     23 (27.8 %)\n\n  \u2003Intensive Care Units   33 (39.7%)\n\n  \u2003Missing                27 (32.5%)\n  ---------------------------------------\n\n###### \n\nThe frequency (%), mean, and standard deviation of nurses\\' responses to the variables based on their importance in occurring medication errors from their viewpoints.\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Factors                                                                                                                             Variables                                                                                             Scale        Mean \u00b1 SD\\   Mean \u00b1 SD\\                                            \n                                                                                                                                                                                                                                                         *F* (%)      total                                                 \n  ----------------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------- ------------ ------------ ------------ ------------ ------------- ------------- --------------\n  Fear of the consequences\\                                                                                                           Fear of the impact of reporting of errors on the personnel\\'s annual evaluation                       13 (15.9%)   19 (23.2%)   8 (9.8%)     30 (36.6%)   12 (14.6%)    3.11 \u00b1 1.35   3.01 \u00b1 1.039\n  of reporting                                                                                                                                                                                                                                                                                                              \n\n  Fear of the impact of reporting of errors on their salary and benefits                                                              10 (12.2%)                                                                                            25 (30.5%)   11 (13.4%)   28 (34.1%)   8 (9.8%)     2.99 \u00b1 1.24                 \n\n  Fear of being blamed by nursing heads                                                                                               6 (7.2%)                                                                                              16 (19.3%)   7 (8.4%)     34 (41%)     20 (24.1%)   3.55 \u00b1 1.25                 \n\n  Fear of being blamed by doctors                                                                                                     9 (11%)                                                                                               18 (22%)     7 (8.5%)     33 (40.2%)   15 (18.3%)   3.33 \u00b1 1.30                 \n\n  Fear of being blamed by colleagues                                                                                                  16 (9.8%)                                                                                             31 (38.3%)   13 (16%)     16 (19.8%)   5 (6.2%)     3.54 \u00b1 1.19                 \n\n  Fear of producing side effects in patients                                                                                          11 (13.8%)                                                                                            14 (17.5%)   2 (2.5%)     29 (36.3%)   24 (30%)     3.51 \u00b1 1.43                 \n\n  Fear of being labeled as incompetent nurses and inadequacy                                                                          14 (17.1%)                                                                                            17 (20.7%)   8 (8.9%)     32 (39%)     11 (13.4%)   3.11 \u00b1 1.35                 \n\n  Fear of colleagues\\' behavior                                                                                                       13 (15.9%)                                                                                            27 (32.9%)   14 (17.1%)   19 (23.2%)   9 (11%)      2.8 \u00b1 1.27                  \n\n  Fear of expressing a negative attitude towards the nurse(s) making errors by the patient and his/her family                         5 (6.2%)                                                                                              17 (21%)     7 (8.6%)     39 (48.1%)   13 (16%)     3.47 \u00b1 1.17                 \n\n  Fear of judicial issues following reporting on medication errors                                                                    6 (7.4%)                                                                                              11 (13.6%)   7 (8.6%)     36 (44.4%)   21 (25.9%)   3.68 \u00b1 1.21                 \n\n  Fear of informing colleagues working in other units and other facilities about one\\'s medication error                              4 (4.9%)                                                                                              19 (23.5%)   8 (9.9%)     34 (42%)     16 (19.8%)   3.48 \u00b1 1.19                 \n\n                                                                                                                                                                                                                                                                                                                            \n\n  Managerial factors                                                                                                                  Lack of receiving positive feedback from the nursing heads following to report on medication errors   4 (4.8%)     17 (20.5%)   10 (12%)     28 (33.7%)   24 (28.9%)    3.61 \u00b1 1.24   3.56 \u00b1 0.99\n\n  False beliefs in nursing heads and managers                                                                                         5 (6%)                                                                                                19 (22.9%)   12 (14.5%)   27 (32.5%)   20 (24.1%)   3.46 \u00b1 1.25                 \n\n  The heads\\' focus only on finding the culprits and blaming them, regardless of other factors involved in the occurrence of errors   5 (6%)                                                                                                12 (14.5%)   12 (14.5%)   30 (36.1%)   24 (28.9%)   3.67 \u00b1 1.21                 \n\n  Disproportionate reactions of the heads to the error seriousness                                                                    4 (4.8%)                                                                                              20 (24.1%)   2 (2.4%)     35 (42.2%)   22 (26.5%)   3.24 \u00b1 1.61                 \n\n  Disproportionate reactions of the heads to the error importance                                                                     3 (3.7%)                                                                                              22 (27.2%)   9 (11.1)     25 (30.9)    22 (27.2%)   3.50 \u00b1 1.25                 \n\n                                                                                                                                                                                                                                                                                                                            \n\n  Factors related to the process\\                                                                                                     Not paying attention to the reporting on some medication errors                                       21 (25.6)    14 (17.1%)   11 (13.4%)   28 (34.1%)   8 (9.8%)      2.85 \u00b1 1.39   3.32 \u00b1 0.79\n  of reporting\\                                                                                                                                                                                                                                                                                                             \n\n  Lack of a clear definition of medication errors                                                                                     8 (9.6%)                                                                                              25 (30.1%)   11 (13.3%)   25 (30.1%)   14 (16.9%)   3.14 \u00b1 1.28                 \n\n  To forget reporting on the medication errors                                                                                        13 (15.7%)                                                                                            20 (24.1%)   11 (13.3%)   28 (33.7%)   11 (13.3%)   3.05 \u00b1 1.32                 \n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n[^1]: Academic Editors: R. Constantino and B. Roberts\n"}
+{"text": "Introduction {#Sec1}\n============\n\nAs the number of people in modern societies with obesity continues to rise, increasing attention is given to the commensal intestinal bacterial communities for their role in affecting energy balance^[@CR1]--[@CR4]^. Extensive evidence from observational cross-sectional studies in humans, as well as from controlled intervention trials in rodent models, suggests that the intestinal bacterial communities, which have co-evolved with mammalian hosts throughout their evolution, play an important regulatory role for metabolic functions, including weight regulation, fat storage and appetite^[@CR5]--[@CR9]^. A seminal study by Liou et al. demonstrates that the conserved shifts in microbiota induced by gastric bypass surgery, including an increase in relative abundance of *Escherichia* and *Akkermansia*, are sufficient to cause weight loss and decreased fat mass in recipient germ-free mice transplanted with this perturbed microbiota^[@CR10]^. This effect may be due to changes in appetite regulation, which are known in part to be controlled by bacterial metabolites, such as short-chain fatty acids and secondary bile acids, binding to specific G-protein-coupled receptors on enteroendocrine L cells, thereby triggering release of satiety hormones^[@CR11],[@CR12]^. The microbiota has a rapid and pronounced effect on the L cell transcriptome, and germ-free mice have increased levels of circulating GLP-1 which suppresses the intestinal transit rate^[@CR13]^. The current understanding of long-term metabolic programming is that the very early phases of life are of pivotal importance^[@CR4],[@CR14]^. During the same time period, an orchestrated successional development of the bacterial community takes place^[@CR15]^. Exposure of the host to microbiota disrupting compounds, such as antibiotics, interferes with this natural process, and can thereby have long-lasting metabolic consequences^[@CR16]^. Current knowledge does, however, not provide an unambiguous link between early-life antibiotic exposure and overweight in later life^[@CR17],[@CR18]^, probably because the causations at play are influenced by many different factors, including timing of the antibiotic exposure, diet of the offspring^[@CR16]^ and class of antibiotic^[@CR4],[@CR19]^, and additionally confounded by factors such as exercise, maternal obesity^[@CR20]^ and infections during pregnancy^[@CR21]^.\n\nResults from a series of well-controlled rodent intervention trials addressing the effect of antibiotic exposure on host weight gain have been published in recent years. These studies have mostly focused on early-life exposure to antibiotic compounds and reveal that subtherapeutic antibiotic treatment (STAT) of young mice increases adiposity and hormone levels related to metabolism^[@CR4]^, transient perturbation of the microbiota by STAT in very early life is sufficient to induce sustained effects on body composition related to obesity^[@CR16]^ and that lifelong STAT induces microbiota changes and enhances the adiposity and insulin resistance associated with high-fat diet^[@CR22]^. Antibiotic therapy at clinically relevant concentrations during early life has similarly been shown to have long-term effect on metabolic outcomes and weight gain^[@CR19]^.\n\nSince early-life disruption of the microbiota by antibiotics thus appears to have consequences for future energy balance regulation, it is important to understand the role of intergenerational transfer of microbiota. A recent study has shown that a diet, which has a lowered content of microbiota-accessible carbohydrates (MAC), reversibly alters microbial composition in mice harbouring a human microbiota within a generation. However, if this diet is sustained over several generations, the changes and loss of bacterial species become irreversible unless the missing microbes are administered with the diet^[@CR23]^.\n\nIn the present study, we investigate the consequences of antibiotic-induced microbial perturbation in Wistar rat dams during the peripartum period on appetite regulation and weight gain in the pups. The dams were orally treated with therapeutic doses of either amoxicillin, which is known to be partly absorbed in the small intestine and enter the mammary glands, or with vancomycin, which is not absorbed into circulation following oral administration. Both of these antibiotics, which are used extensively in human therapy, inhibit bacterial cell wall biosynthesis^[@CR24]^ and have previously been shown to have marked and similar effects on the microbial community structure in Wistar rats^[@CR25]^. We find that pups born to antibiotic-treated dams have altered intestinal bacterial community composition in early life compared to controls, which gradually recedes into adulthood. Concurrently, we observe decreased weight gain in pups born to antibiotic-treated dams, which display a sustained lower body weight at 14 weeks. At 4 weeks of age, intestinal gene expression of the satiety hormone PYY in antibiotic treatment groups is higher compared to controls, which is consistent with the observation of decreased body weight gain.\n\nResults {#Sec2}\n=======\n\nPeripartum antibiotics causes lower body weight in pups {#Sec3}\n-------------------------------------------------------\n\nIn order to investigate the metabolic consequences of exposure to a perturbed microbiota at birth and during early life, we compared groups of rat pups born to dams exposed to either amoxicillin (AMX) or vancomycin (VAN) to untreated controls (CTR) (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). No differences in feed intake or animal weight were observed between groups of dams (Fig.\u00a0[1b, c](#Fig1){ref-type=\"fig\"}). Both AMX and VAN pups appeared to have a slightly lower feed intake from week 6 until termination at week 14; however, the difference was only statistically significant (*p*\u2009*\\<*\u20090.05, Student's *t* test) for VAN pups during weeks 7--9 (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). At Day 2, pups (P2D) in the AMX group were found to have significantly lower (*p*\u2009*=*\u20090.025, Student's *t* test) average body weight than CTR pups (Fig.\u00a0[1e, f](#Fig1){ref-type=\"fig\"}); however, this difference, which could in part be explained by uneven litter sizes at birth until culling at Day 2, disappeared within the first week and no significant differences in weight between groups were observed at weaning of pups at 4 weeks (P4W) (Fig.\u00a0[1g](#Fig1){ref-type=\"fig\"}). From around week 9, the weight of AMX and VAN pups was consistently lower than CTR pups and remained so until termination (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}). Furthermore, the epididymal fat pads, a marker of total body fat^[@CR26]^, were reduced in AMX pups at termination at 14 weeks (P14W) compared to CTR pups and the same trend was observed for VAN pups (Fig.\u00a0[1h](#Fig1){ref-type=\"fig\"}). Administration of antibiotics resulted in significantly increased (*p*\u2009=\u20090.039, Student's *t* test) spleen weight at termination in AMX dams, but no differences in liver weight (Supplementary Fig.\u00a0[1](#MOESM1){ref-type=\"media\"}). We found no difference in serum levels of the acute-phase protein haptoglobin between groups of pups at week 4 (Supplementary Fig.\u00a0[2a](#MOESM1){ref-type=\"media\"}).Fig. 1Peripartum antibiotic treatment affects weight gain in pups. **a** Study design. Wistar rat dams (*n*\u2009*=*\u200933) were housed individually and randomized to receive either amoxicillin (AMX), vancomycin (VAN) or water (CTR) by oral gavage daily from 8 days before expected delivery (Day 0) until weaning of their pups at 4 weeks. Pups did not receive antibiotics directly. The litters were reduced to 6 pups per cage (predominantly male) on Day 2 and further at 2 weeks (*n*\u2009=\u20095/cage) and at 4 weeks (*n*\u2009=\u20092/cage) as indicated. Animal weight and feed intake was recorded during the whole study period. Faecal samples were collected from dams just before expected delivery and dissections (intestinal-, tissue- and blood samples) were performed on pups at 4 time points; 2 days (P2D), 2 weeks (P2W), 4 weeks (P4W) and 14 weeks (P14W) as well as the dams at weaning (D4W). Feed intake and animal weight for **b**,**c** dams and **d**,**e** pups was determined weekly and shown as mean\u2009\u00b1\u2009SEM with *n*\u2009*=*\u200910 (CTR), *n*\u2009=\u200912 (AMX) and *n*\u2009=\u200911 (VAN). Pup weights at **f** P2D and **g** P4W for each treatment group are also shown as box-plots with whiskers indicating total range. **h** Epididymal fat pad weight was determined at termination of pups (P14W) and depicted as box-plots with whiskers indicating total range. Significant differences compared to the CTR group are indicated in all panels: \\**p*\u2009\\<\u20090.05\n\nAntibiotic treatment changes the gut microbiota of dams {#Sec4}\n-------------------------------------------------------\n\nAs expected, both amoxicillin and vancomycin dramatically reduced the bacterial \u03b1-diversity in the treated dams (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). At the time of delivery (D0W), the faecal microbiota in dams of both treatment groups (AMX and VAN) clustered separately and distantly from the CTR group (Fig.\u00a0[2b, d](#Fig2){ref-type=\"fig\"}) and were generally highly enriched for members of the Proteobacteria phylum, including *Escherichia* spp., *Morganella* spp. and *Proteus* spp., and depleted for Firmicutes, including genera known to produce butyrate, such as *Clostridium XIVa* (Fig.\u00a0[2f, g](#Fig2){ref-type=\"fig\"}). Interestingly, the bacterial diversity in faecal samples from dams, collected \u223c1 week after onset of the treatment (D0W), was lower than the diversity in caecal samples taken 4 weeks later (D4W) (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}), which suggests a gradual adaptation to the sustained antibiotic pressure, which could in part be driven by antibiotic- resistant strains. In line with this, a principal coordinate analysis (PCoA) based on unweighted UniFrac distances (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}), Bray--Curtis dissimilarity (Supplementary Fig.\u00a0[3a](#MOESM1){ref-type=\"media\"}) as well as relative abundance distribution patterns at phylum level (Fig.\u00a0[2f](#Fig2){ref-type=\"fig\"}) showed that the microbiota of the AMX and VAN dams were more similar to the CTR after 4 weeks. We have previously shown that faecal and caecal bacterial diversity and composition in both untreated and antibiotic-treated rats are similar^[@CR25]^. Notably, the total bacterial load in faecal samples from dams \u223c2 days before delivery as determined by culturing was significantly higher (*p*\u2009\\<\u20090.0001, Student's *t* test) in AMX and VAN dams than in CTR dams (Supplementary Fig.\u00a0[4](#MOESM1){ref-type=\"media\"}).Fig. 2Antibiotic treatment affects microbiota in both treated dams and their pups. **a** Bacterial diversity in faecal/caecal content estimated as Shannon diversity index in dams *n*\u2009*=*\u200910 (CTR), *n*\u2009=\u200912 (AMX) and *n*\u2009=\u200911 (VAN) at the time of giving birth (D0W, faecal sample) and 4 weeks later (D4W, caecal sample) as well as pups at 2 days (P2D, intestinal sample), 2 weeks (P2W), 4 weeks (P4W) and 14 weeks (P14W) (all caecal samples) of age shown as box-plots with whisker indicating total range. **b** Boxplot of unweighted UniFrac distances calculated from caecal microbial composition within all groups (grey) and between groups (yellow) of dams (D0W and D4W) and pups (P2W, P4W and P14W) and **c** boxplot showing numbers of shared OTUs between animals within the CTR group and between the CTR and antibiotic treatment groups of pups at P2D, P4W and P14W. Whisker indicates total range. Principal coordinate analysis (PCoA) based on unweighted UniFrac distances of **d** faecal and caecal samples from dams (D0W and D4W) and **e** caecal samples from pups (P2W, P4W and P14W) coloured according to treatment group. **f** Bacterial composition in dams and pups shown as average relative abundance at the phylum level. **g** Heatmap depicting log~10~ relative abundance scores for faecal and caecal agglomerated genera present in \\>50% of samples. Results are shown for dams at delivery (D0W) and dams 4 weeks after delivery (D4W) as well as pups at different ages (P2D, P2W, P4W and P14W). In panels **a**, **b** and **c**: \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001; \\*\\*\\*\\**p*\u2009\\<\u20090.0001 and in panel **g** FDR-corrected *p* values compared to the CTR group are shown *\\*q*\u2009\\<\u20090.05; \\*\\**q*\u2009\\<\u20090.01. The left-hand colour bar shows the taxonomic classification at the phylum level\n\nPups show delayed intestinal microbiota development {#Sec5}\n---------------------------------------------------\n\nFor all groups of pups, we observed a gradual increase in bacterial \u03b1-diversity reaching approximately the same levels as the untreated dams after 14 weeks (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). The \u03b1-diversity was significantly lower (*p*\u2009=\u20090.013, Student's *t* test) in AMX pups than in CTR pups at 4 weeks, with the same tendency for the VAN pups (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). Analysis of unweighted UniFrac distances within and between groups at different time points as well as PCoA analysis revealed group-specific clustering and significant differences in microbial composition in pups at both 2 weeks of age (P2W) and 4 weeks of age (P4W) (Fig.\u00a0[2b, e](#Fig2){ref-type=\"fig\"}). Separation of treatment groups was significant (*p*\u2009\\<\u20090.001, Adonis test) at all time points with gradually declining separation (P2W; *R*^2^\u2009=\u20090.22, P4W; *R*^2^\u2009=\u20090.14 and P14W; *R*^2^\u2009=\u20090.088). Analysis based on Bray--Curtis dissimilarity showed similar clustering (Supplementary Fig.\u00a0[3b](#MOESM1){ref-type=\"media\"}). Analysis of the number of shared OTUs between pups within the control group and pups between the control group and antibiotic treatment groups showed that at 4 weeks, but not at 2 and 14 weeks, there were significantly fewer (*p*\u2009\\<\u20090.0001, Mann--Whitney test) shared OTUs between treatment and the control group, than within the control group (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). This indicated that some of the shared OTUs in the CTR P4W pups are lacking in both the AMX and VAN pups. To further investigate this, we focused on OTUs that were omnipresent in the CTR animals at 4 weeks and determined their prevalence in the antibiotic-exposed pups of the same age (Supplementary Fig.\u00a0[5a](#MOESM1){ref-type=\"media\"}). We noted an interesting pattern for OTU_114 (99% homology to *Alistipes ihumii* strain AP11), which was absent in nearly all animals in both the AMX and VAN groups at P4W groups (whilst present in all CTR animals). We further found that this particular OTU remained absent in all VAN (11/11) and most AMX (9/12) pups to the age of 14 weeks and was also not present in the antibiotic-treated dams D0W at the time of giving birth (Supplementary Fig.\u00a0[5b](#MOESM1){ref-type=\"media\"}). At the genus level, the composition of the microbiota of the 2-week-old pups (P2W) of the antibiotic-treated dams differed significantly from CTR pups in several genera. This included a higher relative abundance of *Clostridium XI*, which contains *Peptoclostridium difficile* (formerly *Clostridium difficile*) in both AMX and VAN pups (*P*~adj~\u2009=\u20090.019 and *P*~adj~\u2009=\u20090.022, respectively, permutation-based *t* test) (Fig.\u00a0[2g](#Fig2){ref-type=\"fig\"}). At 4 weeks (P4W), only a few genera differed significantly from the CTR, namely *Escherichia*, which was higher in both AMX and VAN pups (*P*~adj~\u2009=\u20090.042 and *P*~adj~\u2009=\u20090.049, respectively), and *Ruminococcus* (*P*~adj~\u2009=\u20090.049), which was less abundant in VAN pups compared to CTR pups. At the of age 14 weeks (P14W), the pup microbiotas of all groups clustered together (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"} and Supplementary Fig.\u00a0[3b](#MOESM1){ref-type=\"media\"}) and no differences in UniFrac distances within or between treatment groups (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}), nor within or between bacterial groups at genus level were found (Fig.\u00a0[2g](#Fig2){ref-type=\"fig\"}). So the initially reduced diversity in the AMX and VAN pups, indicating a delayed microbial development, was mostly normalized at 14 weeks.\n\nPeripartum antibiotics affects SCFA production in pups {#Sec6}\n------------------------------------------------------\n\nTo understand whether the observed differences in microbiota in 4-week-old pups caused differences in bacterial activity related to energy extraction and signalling, we investigated short-chain fatty acid composition in the caecum at this age. No significant differences in acetate levels between groups were found (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). Compared to CTR pups, the AMX pups had significantly lower (*p*\u2009*=*\u20090.035, Student's *t* test) concentrations of butyrate, with the same trend observed for VAN, while the VAN pups also had lower concentrations of propionate (*p*\u2009*=*\u20090.0087, Student's *t* test) (Fig.\u00a0[3b, c](#Fig3){ref-type=\"fig\"}). The observed reduction in butyrate in the antibiotic groups is consistent with delayed microbial development, but was not accompanied by enlarged caecum size as noted for the dams (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}). At 14 weeks of age, the caecum size was found to be slightly reduced in the VAN group compared to the control group; however, this was attributed to a strong correlation between caecum weight and total body weight (Spearman's rho\u2009=\u20090.62, *p*\u2009=\u20090.0001). The caecal pH of VAN dams was increased compared to CTR dams; however, no difference in pH between the three groups of pups was found (Fig.\u00a0[3e](#Fig3){ref-type=\"fig\"}). Finally, no differences in faecal gross energy were found between the different groups of pups at 4 weeks (Fig.\u00a0[3f](#Fig3){ref-type=\"fig\"}), suggesting that the differences in body weight observed between the groups of pups were not sufficiently explained by differences in microbial energy extraction capacity.Fig. 3Antibiotic treatment of dams affects short-chain fatty acid levels in pups. Median concentrations of **a** acetate, **b** propionate and **c** butyrate in the caecum of 4-week-old pups. **d** Caecum weight and **e** caecal pH of dams 4 weeks after delivery (D4W) and pups at 4 (P4W) and 14 (P14W) weeks of age. **f** Faecal gross energy in samples from pups at 4 weeks (P4W). All panels show box-plots with whiskers indicating total range. \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 and \\*\\*\\*\\**p*\u2009\\<\u20090.001\n\nAppetite regulation is affected by peripartum antibiotics {#Sec7}\n---------------------------------------------------------\n\nSince we observed a reduced body weight together with a reduced feed intake in the pups that received an antibiotic-perturbed low-diversity microbiota, we considered that expression of intestinal genes potentially affected by the microbiota and related to appetite regulation may be involved. Indeed, expression of the satiety hormone Peptide YY (PYY) gene was higher in both AMX and VAN pups (*p*\u2009=\u20090.0075 and *p*\u2009=\u20090.0012, respectively, Student's *t* test) at the age of 4 weeks compared to CTR pups, in accordance with the differences observed in the antibiotic-treated dams (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}). No differences in PYY expression were observed in pups at 2 weeks despite significant differences in microbial community composition. At the age of 14 weeks also, no differences between groups were observed in the pups, suggesting a transitory effect on colonic PYY gene expression in the pups of antibiotic-treated dams. Colonic expression of the incretin Glucagon-like peptide-1 (GLP-1) gene was also higher in AMX-treated dams (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}). However, no difference was observed between groups of pups at any of the sampling times, which was supported by no differences in serum GLP-1 concentrations in 4-week-old pups (Supplementary Fig.\u00a0[2b](#MOESM1){ref-type=\"media\"}). Expression of genes encoding G-protein-coupled receptors GPR41 and GPR43 specific for short-chain fatty acids (SCFAs) did not differ in dams nor pups at any of the sampling points (Fig.\u00a0[4c, d](#Fig4){ref-type=\"fig\"}) despite changed SCFA levels in antibiotic- treated dams and some groups of pups. Expression of the bile acid-specific TGR5 membrane receptor gene in VAN pups was notably lower (*p*\u2009=\u20090.0002, Mann--Whitney test) at 2 weeks, and tended to be lower also at 4 weeks. In AMX pups, the expression of this gene was lower (*p*\u2009=\u20090.0090, Student's *t* test) at 4 weeks (Fig.\u00a0[4e](#Fig4){ref-type=\"fig\"}). Finally, the expression of the bile acid-sensing nuclear farnesoid X receptor (FXR) was not different between groups of dams and pups, respectively (Fig.\u00a0[4f](#Fig4){ref-type=\"fig\"}). Together, this suggests that changes in colonic gene expression related to appetite regulation may contribute to the observed lower body weight and lower feed intake in the AMX and VAN pups.Fig. 4Antibiotic treatment of dams affects gene expression in pups. **a--f** Relative gene expression (compared to the CTR group) in colon tissue from dams 4 weeks after delivery (D4W) and pups at 2 weeks (P2W), 4 weeks (P4W) and 14 weeks (P14W) for appetite hormones (PYY and GLP-1), short-chain fatty acid receptors (GPR41 and GPR43) and bile acid receptors (TGR5 and FXR) determined by quantitative PCR analysis. The results are shown as box-plots with whisker indicating total range of twofold change relative to the average of controls in each age group. In all panels, \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001; \\*\\*\\*\\**p*\u2009\\<\u20090.0001\n\nPeripartum antibiotics do not affect bile acid levels in pups {#Sec8}\n-------------------------------------------------------------\n\nSince we observed reduced colonic TGR5 expression in pups of antibiotic-treated dams at the age of 4 weeks, we speculated that this could be linked to differences in bile acid profiles resulting from a delayed gut microbiota development in the pups. Antibiotics had a dramatic effect on the serum bile acid profile in dams (Fig.\u00a0[5a--c](#Fig5){ref-type=\"fig\"}), characterized by a reduction of secondary bile acids and an overall increase in bile acids in the vancomycin-treated dams. Principal component analysis (PCA) of serum bile acids in the pups revealed differences according to age, but independent of peripartum antibiotics treatment (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"} and supplementary Fig.\u00a0[6a](#MOESM1){ref-type=\"media\"}). Serum bile acid profiles were very similar among pups at 2 weeks of age, as shown by the tight clustering of P2W samples in the PCA plot (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}). After weaning, at 4 weeks of age, the P4W samples were more scattered (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}), probably driven by the microbiota differences observed at this time point (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}). Finally, at the age of 14 weeks, the serum bile acid profile of the pups were very similar to that of CTR dams (D4W), as observed by the clustering of P14W samples with the CTR dams (D4W) in the PCA plot (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}). The collective mean concentration of serum bile acids in pups differed with age, peaking at the age of 4 weeks, at which time point also secondary bile acids were first markedly observed, and reached a profile similar to that of CTR dams at the age of 14 weeks (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). The bile acid profiles of the 2-week-old pups were dominated by conjugated primary bile acids, while the bile acid profiles at 4 weeks were dominated by de-conjugated primary bile acids, and the bile acid profiles at 14 weeks appeared not to be dominated by any specific bile acid group. The individual serum bile acid concentrations in AMX and VAN pups were not significantly different from CTR pups at any of the time points, possibly due to large inter-individual variations, while several serum bile acids in the antibiotic-treated dams differed from CTR dams (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}). In colonic content from pups at 4 weeks, no differences between groups of pups were detected for any of the quantified bile acids\u00a0(Fig. 5d), although a PCA plot suggested a slightly higher inter-individual variation within the bile acid profile of AMX and VAN pups (Fig.\u00a0[5e](#Fig5){ref-type=\"fig\"} and supplementary Fig.\u00a0[6b](#MOESM1){ref-type=\"media\"}) compared to CTR. Thus, peripartum antibiotics did not significantly affect bile acid profiles in pups.Fig. 5Bile acid concentrations and composition in pups is dependent on age but not peripartum antibiotics. **a** Principal component analysis (PCA) of the serum bile acid profile and **b** total mean concentration of bile acids and distribution of conjugated and unconjugated primary and secondary bile acids, respectively in dams and pups. **c** Temporal and treatment-induced changes in serum bile acid concentrations are visualized in a heatmap based on average row *z*-scores across all groups of samples and shown for pups at 2 weeks (P2W), pups at 4 weeks (P4W), pups at 14 weeks (P14W) and dams 4 weeks after delivery (D4W). **d** Heatmap of the mean abundance (row *z*-score) of the individual bile acids in colon content from pups at 4 weeks. **e** PCA plot of the colon bile acid profiles (P4W). In panel **b** significant differences in total bile acid concentrations are indicated: \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001 and in panel **c**--**d**: FDR-adjusted *p* values compared to the CTR group are shown: \\**q*\u2009\\<\u20090.05; \\*\\**q*\u2009\\<\u20090.01. Serum and colonic bile acid concentrations were auto-scaled before PCA to accommodate for highly abundant bile acids and associated loading plots for the PCA plots are shown in supplementary Fig.\u00a0[6](#MOESM1){ref-type=\"media\"}\n\nAppetite regulation is associated with specific bacterial groups {#Sec9}\n----------------------------------------------------------------\n\nTo investigate associations between the gut microbiota and intestinal gene expression levels of satiety hormones (GLP-1 and PYY), SCFA receptors (GPR41 and GPR43), bile acid receptors (FXR and TGR5) and caecal SCFA levels in pups at 4 weeks, we conducted a correlation analysis at the genus level (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}). Bacterial \u03b1-diversity (Shannon and richness indices) in the caecum correlated positively with caecal butyrate levels. Additionally, \u03b1-diversity was negatively associated with *Escherichia* spp., which was overrepresented in both antibiotic-treated groups at 4 weeks (Fig.\u00a0[2g](#Fig2){ref-type=\"fig\"}). Notably, *Escherichia* spp. was found to be the only genus that correlated positively with both PYY (Spearman's rho\u2009=\u20090.41, *p*\u2009*=*\u20090.019) and GLP-1 (Spearman's rho\u2009=\u20090.52, *p*\u2009*=*\u20090.002) gene expression in colonic tissue and also correlated positively with expression of FXR (Spearman's rho\u2009=\u20090.50, *p*\u2009*=*\u20090.003). Gene expression of FXR, but not TGR5, was strongly correlated to the gene expression of satiety hormones PYY and GLP-1 (Spearman's rho\u2009=\u20090.57, *p*\u2009*=*\u20090.0005 and rho\u2009=\u20090.65, *p*\u2009*=*\u20090.00005, respectively). The strongest negative correlations with appetite hormone expression were found for *Staphylococcus* spp. (Spearman's rho\u2009=\u2009\u22120.51, *p*\u2009*=*\u20090.002) and *Hespellia* spp. (Spearman's rho\u2009=\u2009\u22120.47, *p=*\u20090.006) for PYY and GLP-1, respectively. Colonic expression of bile acid receptors TGR5 and to a lesser extent FXR correlated with several specific primary and secondary bile acids in both serum and colonic content (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}). Collectively, only few bacterial groups were found to correlate to expression of appetite hormones GLP-1 and PYY.Fig. 6Associations between bacterial genera and intestinal gene expression, short-chain fatty acids and alpha-diversity. **a** Heatmap of Spearman's rank correlation coefficients between the relative abundance of genera (with a minimum prevalence of 50% across all samples) at 4 weeks of age (P4W) and levels of SCFAs, gene expression and bacterial diversity. The colour bar shows the taxonomic classification of included genera at phylum level (blue: *Firmicutes*, red: *Bacteroidetes*, green: *Proteobacteria*, yellow: *Actinobacteria* and pink: *Tenericutes*). **b** Heatmap of Spearman's rank correlation coefficients between concentrations of specific bile acids in serum and colon and gene expression in colon of bile acid receptors FXR and TGR5. Colour bars indicate primary (green), secondary (orange), conjugated (black) and unconjugated (grey) bile acids. In both panels, \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01 and \\*\\*\\**p*\u2009\\<\u20090.001\n\nDiscussion {#Sec10}\n==========\n\nHere we report that exposure of conventional Wistar rat dams to either amoxicillin or vancomycin during the peripartum period caused a sustained lower body weight in their male pups, as compared to a control group not exposed to antibiotics. The reduced body weight was also reflected in significantly reduced epididymal fat pats at 14 weeks of age. The pups did not directly receive antibiotics at any time, although a low exposure through coprophagy may occur and additionally amoxicillin may potentially be transferred during lactation due to intestinal absorption in the dams, which is an intrinsic limitation of the experimental design. Previous studies in rodents have, however, indicated such transfer to be minimal^[@CR27]^. We therefore propose that effects on growth parameters were driven primarily by an antibiotic-induced altered microbial transmission from dams during birth and early life, but cannot rule out direct effects of low doses of antibiotics. This is in line with a recent study demonstrating that colonization of germ-free mice with antibiotic-altered microbial communities may induce changes in the metabolic phenotype^[@CR16]^. Colonization of germ-free IL10-deficient mouse dams with antibiotic-perturbed bacterial communities as well as peripartum antibiotic treatments of dams has also been shown to enhance colitis in pups^[@CR28]^. Importantly, these previous studies have indicated a risk of lasting metabolic consequences despite recovery of the intestinal microbial community. In the current study, we found a gradual recovery of the disturbed pup microbiota, which at 14 weeks of age was very similar to the CTR group at the genus level, although grouping was still indicated by the Adonis test based on the UniFrac distance matrix. The number of shared OTUs between animals within the CTR groups at 4 weeks (P4W) was significantly higher than the number of shared OTUs between treatment groups. This observation is a clear indication that some OTUs are absent in the AMX and VAN groups at this time point. An interesting example of this was observed for OTU_114, classified as *Alistipes ihumii*, which is omnipresent in all animals in the CTR group at 4 weeks, but reduced to below-detection level for almost all AMX and VAN pups at P4W and remained so until P14W. In general, *Alistipes* spp. are strict anaerobes and have previously been reported to be underrepresented in the caecum and colon of undernourished mice^[@CR29]^. At 14 weeks, the body weight of both AMX and VAN pups appeared to approach a plateau significantly lower than that observed for the control group. This observation of stunted growth is different from several recent reports of increased weight gain and adiposity as a consequence of early-life antibiotic exposure^[@CR4],[@CR22]^. However, it is important to recognize the intrinsic variable parameters associated with antibiotic treatment, including the mode-of-action of the antibiotic compound as well as the concentration, timing and duration of exposure. A recent review proposed a pathway for antibiotic-mediated weight modulation inferring that early-life perturbation of the microbial community by high-dose antibiotics can lead to stunted growth due to loss of microbiota-derived calories^[@CR30]^, which is in line with previous work elucidating the role of the microbiota in relation to energy harvest and fat storage^[@CR5],[@CR31]^. Consistent with our findings, it was recently reported that stunted growth of malnourished Malawian children is associated with delayed maturation of the gut microbiota^[@CR32]^. They demonstrated that introducing *Ruminococcus gnavus* and *Clostridium symbiosum* to this immature microbiota ameliorated growth and metabolic abnormalities in recipient germ-free mice. Interestingly, we found reduced *Ruminococcus* spp. in the VAN-treated animals at 4 weeks and a negative correlation between this genus and gene expression of the satiety hormone PYY. *Ruminococcus bromii* has also previously been linked to obesity and resistant starch degradation in humans^[@CR33],[@CR34]^. Although several studies thus link antibiotic-induced microbial pertubation to changes in metabolic programming, it has, however, also been shown that antibiotic treatment may elicit microbiome-independent changes in host metabolites^[@CR35]^. Microbiome-independent effects of antibiotics may therefore also contribute to the observed effects in the current study.\n\nThe observed differences in weight gain during the first 14 weeks of life may be caused by differences in feed intake, altered energy extraction from the feed or a combination of these. We generally observed a minor reduction in feed intake for antibiotic pups after weaning, especially for VAN pups, which is consistent with a previous study showing reduced feed intake for adult rats treated orally with amoxicillin^[@CR25]^. No difference in faecal gross energy at 4 weeks was found, although differences in nitrogen and fat absorption cannot be ruled out. Since there were also no differences in animal weight and only minor differences in microbial composition between groups of pups at 4 weeks, we find it unlikely that microbiota-driven differences in energy extraction were sufficient to cause the observed 6--7% lower body weight of AMX and VAN pups aged 14 weeks.\n\nWe thus considered that early-life programming of appetite regulation could be involved in later-life weight regulation. The bacterial diversity, but not the total bacterial load, in the faeces of the antibiotic-treated dams shortly before birth was significantly reduced as compared to control animals, which is consistent with previous studies^[@CR25],[@CR36]^. It has also previously been shown that the vaginal microbiota is altered by antibiotics during pregnancy^[@CR37]^. Collectively, it is likely that the initial bacterial seeding of the pups of antibiotic-treated dams was very different from CTR dams. Specifically, the pups were initially exposed to a faecal microbiota characterized by low bacterial alpha-diversity with a relatively high abundance of Enterobacteriales, including *Escherichia* spp., *Morganella* spp. and *Proteus* spp. The increased levels of these facultative anaerobes in the dams may be explained by depletion of butyrate-producing Clostridia causing decreased levels of butyrate^[@CR25]^, which in turn has been reported to fuel aerobic pathogen expansion due to increased epithelial oxygenation^[@CR38]^. The high relative abundance of Enterobacteriales in the dams at the time of giving birth was, however, not directly reflected in the pups microbiota before the age of 4 weeks, at which point the alpha-diversity was lower compared to CTR pups and the relative abundance of *Escherichia* spp. was higher in both the AMX and VAN pups. Interestingly, the lower alpha-diversity, the increased PYY expression and the reduced weight first became apparent after the pups had started on solid foods. Leptin ingested during lactation may have initially regulated food intake and growth in all pups similarly^[@CR39],[@CR40]^ so that differences in appetite induced by a low-diversity microbiota only became noticeable after weaning.\n\n*Escherichia coli* has previously been associated with body mass index and appetite regulation^[@CR41]^. The suggested mechanisms behind this association include expression of the ClpB protein, a mimetic of the host \u03b1-MSH hormone known to induce satiety^[@CR42]^, bacterial production of the tryptophan metabolite indole, which modulates GLP-1 release from enteroendocrine L cells^[@CR43]^ and bacterial lipopolysaccharide binding to TLR4 receptors in the vagal afferents signalling to the brain^[@CR44]^. In line with this, we identified significant positive correlations between intestinal *Escherichia* spp. relative abundance and colonic gene expression of satiety proteins GLP-1 and PYY in the pups across all treatment groups as well as significantly higher gene expression of PYY in pups of antibiotic-treated dams at 4 weeks. Antibiotics have previously been shown to increase postprandial secretion of PYY in humans^[@CR45]^. A link between *Escherichia* spp. and weight regulation is consistent with a previous report of weight loss in germ-free mice receiving caecal microbiota from Roux-en-Y gastric bypass-treated mice characterized by increased abundance of *Escherichia*, *Akkermansia* and *Alistipes*^[@CR10]^.\n\nWe have previously reported that antibiotic treatment causes a decrease in the levels of microbially produced SCFAs and increases caecal size and luminal pH^[@CR25]^. Also in the present study, antibiotic treatment of pregnant dams increased caecal size and pH, consistent with an observed reduction in butyrate-producing Firmicutes. In the 4-week-old pups of antibiotic-treated dams, we found a reduction in caecal butyrate and propionate levels in the AMX and VAN pups, respectively. Very few differences in microbiota at the genus level were seen at this age, but lower levels of the butyrate-producing *Ruminococcus* spp., were detected in the VAN pups. Propionate activates PYY and GLP-1 release from L cells in humans^[@CR46],[@CR47]^, and it has recently been shown that GPR43 deficiency impaired the release of GLP-1 and PYY after administration of propionate^[@CR11]^. In the current study, no differences in gene expression of SCFA receptors GPR41 and GPR43 were seen between treatment groups at any age in the pups, which suggests that changes in expression patterns of these receptors did not drive the observed metabolic differences in the pups, although differences at the protein level cannot be ruled out.\n\nAntibiotic treatment of dams had a major effect on serum bile acid profiles characterized by a reduction of secondary bile acids in both AMX and VAN groups as compared to CTR. Since the generation of secondary bile acids largely depends on microbial transformation processes, the observed decreased concentration is in accordance with antibiotic-driven alteration of the gut microbiota in the dams and is consistent with previous studies^[@CR48]--[@CR50]^. An increase in primary bile acids (both conjugated and de-conjugated) as well as in the total bile acid pool occurred in the VAN dams but not in the AMX dams as compared to the CTR group. The difference between the antibiotic groups is likely to be explained by differences within the lactobacilli, which are resistant to vancomycin and capable of performing de-conjugation via bile salt hydrolase activity^[@CR51]^.\n\nThe differences in bile acid profiles of dams were not transferred to the pups, as bile acid profiles measured in pup serum at ages of 2, 4 and 14 weeks and in colon content at 4 weeks did not show any differences between the treatment groups. We did, however, find a down-regulation of expression of the TGR5 receptor in both 2- and 4-week-old pups of antibiotic-treated dams, which is an important ligand of secondary bile acids. Bile acids binding to the TGR5 receptor on enteroendocrine L cells may stimulate the secretion of satiety hormones PYY and GLP-1^[@CR52],[@CR53]^ and bile acid infusions have been shown to increase plasma PYY in a dose-dependent manner^[@CR54]^.\n\nThe observed temporal development in serum bile acid profiles with age, irrespective of the treatment group, is consistent with the co-occurring successional development of the gut microbiota. The immature gut microbiota in early life is expected to have limited bile acid biotransformation capacity consistent with the observed low levels of secondary and de-conjugated primary bile acids in 2-week-old pups. This was succeeded by a dramatic increase in total serum concentration of bile acids at 4 weeks, driven primarily by secondary and unconjugated primary bile acids. The total amount of bile acids then decreased again in the 14-week-old pups to resemble the level of bile acids found in control dams. An observed, spike in total serum bile acid levels at 4 weeks has previously been recognized in Wistar rats^[@CR55]^ and could be related to the shift from breast milk to solid diet during weaning. Indeed, Massimi et al. tracked the levels of CYP7A1 mRNA, encoding a rate-limiting enzyme in the bile acid synthesis pathway in Sprague Dawley rats, and found a distinctive peak around day 22, at which point weaning took place^[@CR56]^. Coincident with weaning, the leptin levels in rats increase^[@CR39]^ and leptin has been suggested to regulate bile salt metabolism^[@CR57]^.\n\nIn conclusion, we found that peripartum antibiotic treatment of rat dams caused a delay in successional microbiota development and sustained lower adult body weight in their pups. Concurrent with the lower body weight, we found differences in appetite hormone PYY and bile acid receptor TGR5 expression levels in colonic tissue as well as decreased levels of caecal SCFA levels in 4-week-old pups. These results are consistent with recent studies substantiating a microbial impact on early-life metabolic programming and provide new knowledge concerning potential risks associated with antibiotic administration during pregnancy.\n\nMethods {#Sec11}\n=======\n\nAnimals and housing {#Sec12}\n-------------------\n\nTime-mated specific pathogen-free female Wistar Hannover rats (*n*\u2009=\u200933), with a weight of 200\u2009\u00b1\u200910\u2009g, were obtained from Taconic Biosciences (Lille Skensved, Denmark). Upon arrival on gestation day (GD) 8, the animals were housed individually under controlled environmental conditions (12-h reverse light/dark cycles, temperature 21.5\u2009\u00b1\u20090.3\u2009\u00b0C, relative humidity 51.3\u2009\u00b1\u20093.1% and 8--10 air changes per hour). Animals had access to ad libitum water and feed (Altromin 1314 until giving birth and then Altromin 1324, Altromin Spezialfutter GmbH, Germany) throughout the study. Animal experiments were carried out at the DTU National Food Institute (M\u00f8rkh\u00f8j, Denmark) facilities. Ethical approval for the study was given by the Danish Animal Experiments Inspectorate with authorization number 2012-15-2934-00089, C2. The experiments were overseen by the DTU National Food Institutes in-house Animal Welfare Committee for animal care and use. Animals were monitored twice a day.\n\nExperimental design {#Sec13}\n-------------------\n\nFollowing a 5-day acclimatization period, dams were equally allocated into three treatment groups (with 10--12 animals in each group) based on weight. Blinding was not implemented during the animal trial due to practical reasons. Antibiotic treatment of dams started on GD 14, which is 8 days before expected birth, and consisted of a daily dosage of 0.5\u2009mL of antibiotic solution containing 60\u2009mg\u2009mL^\u22121^ amoxicillin, Sigma-Aldrich A8523 (AMX, *n*\u2009=\u200912), 8\u2009mg\u2009mL^\u22121^ vancomycin, Sigma-Aldrich 861987 (VAN, *n*\u2009=\u200911) or water (CTR, *n*\u2009=\u200910) by oral gavage. Antibiotic treatment continued until pups were weaned at 4 weeks of age. The scale weight of pups as well as feed and water intake after weaning was determined weekly during the study until termination at week 14 (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). Litters were reduced to 6 pups per dam at 2 days of age (P2D), 5 pups per dam at 2 weeks of age (P2W) and further to 2 animals following weaning at 4 weeks of age (P4W) with a preference for male animals remaining. Small litters were supplemented with excess pups from other dams within the same treatment group at P2D. At P2D, P2W, P4W and P14W, the culled animals from each litter were dissected for further analysis, including collection of blood, intestinal tissue and content from caecum and colon and weighing spleen, liver and caecum as well as determining pH in the latter (Orion^TM^ 3-star benchtop pH meter, Thermo Fisher Scientific). Faecal samples collected from dams 2 days before expected delivery (D0W) as well as samples taken during dissection of dams at weaning (D4W) were also included for analysis.\n\nExtraction of DNA and amplicon library preparation {#Sec14}\n--------------------------------------------------\n\nIntestinal samples for DNA extraction were stored at \u221280\u2009\u00b0C until analysis. Community DNA was extracted from faecal samples collected from dams just prior to delivery (D0W), from total gut content of 2-day-old female pups (P2D) and from caecal content of dams and male pups (D4W, P2W, P4W and P14W) using the MoBio PowerLyzer^\u00ae^ Power Soil^\u00ae^ DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA) according to the manufacturer's recommendations with minor modifications. A maximum of 200-mg samples was used for extraction and the bead beating step was conducted twice at 30\u2009cycles\u2009s^\u22121^ for 5\u2009min with a 10-min break in-between (Retsch MM 300 mixer mill). DNA concentrations were measured fluorometrically with the Qubit dsDNA HS kit (Life Technologies). The bacterial community composition was determined by amplification and sequencing of the V3-region in the 16S rRNA gene using the Ion Torrent PGM platform (Life Technologies) as previously described^[@CR58]^. Briefly, the V3-region of the 16S rRNA gene was amplified using a universal forward primer (PBU 5\u2032-A-adaptor-TCAG-barcode-CCTACGGGAGGCAGCAG-3\u2032) with a unique 10--12-bp barcode for each bacterial community (IonXpress barcode as suggested by the supplier, Life Technologies) and a universal reverse primer (PBR 5\u2032-trP1-adapter-ATTACCGCGGCTGCTGG-3\u2032). The PCR products were purified using the MAGBIO HigPrep\u2122 PCR-96 well protocol according to the manufacturer's recommendations, and DNA concentrations were determined with Qubit HS assay. Three amplicon libraries were constructed containing material from P2D/D0W, P2W/P4W and D4W/P14W, respectively, by mixing equal amounts of PCR products from each original community. Sequencing of the three libraries was performed on 318-chips for Ion Torrent sequencing using the Ion OneTouch\u2122200 Template Kit v2 DL.\n\nBacterial community composition {#Sec15}\n-------------------------------\n\nSequence data in FASTQ format was initially processed in CLC Genomic Workbench (version 8.5, Qiagen, Aarhus, Denmark) in order to de-multiplex and remove sequencing primers, retaining reads only if both forward and reverse primers were correctly identified with 100% homology as previously described^[@CR58]^. Next, operational taxonomic units (OTUs) were picked de novo using UPARSE^[@CR59]^ and an OTU table was generated. Taxonomical classification of OTUs was assigned using the Ribosomal Database Project multiclassifier version 2.10.1 and the RDP database^[@CR60]^ with confidence threshold set to 0.5 as recommended for sequences shorter than 250\u2009bp^[@CR61]^. Further downstream processing was performed in QIIME^[@CR62]^. A phylogenetic tree was generated (make_phylogeny.py) and rooted to an archaeal species following alignment of all OTUs. The OTU table was filtered to exclude singletons and include only OTUs assigned as bacteria (excluding the Cyanobacteria/Chloroplast group), and present in at least two samples, resulting in a total of 1426 OTUs and a median read depth of 32,906 (range 13,684--124,303). The QIIME workflow core_diversity_analysis.py was used to generate alpha- and beta-diversity indices and relative abundance tables at different taxonomical levels with a rarefied depth of 13,600 reads, including all samples except P2D. To determine UniFrac distances within and between groups of animals, the make_distance_boxplots.py script was used. The script compare_categories.py was used to perform Adonis tests on bacterial community composition. To determine the number of shared OTUs, the generated OUT table was first rarefied to 10,000 reads per sample using single_rarefaction.py and the script shared_phylotypes.py was used. A heatmap was generated at the genus level, including only those genera observed in a minimum of 50% across all samples and coloured according to average log~10~ relative abundances within each group/time point. The samples from 2-day-old pups (P2D) contained only 180 OTUs in reads assigned to bacteria with a median read depth of 8639 (range 1484--42,149) and were analyzed only for phylum-level distribution.\n\nGene expression analysis {#Sec16}\n------------------------\n\nApproximately 1\u2009cm of colon tissue from dissected animals (D4W, P2W, P4W and P14W), stored in RNAlater^\u00ae^ (Life Technologies) at \u221280\u2009\u00b0C, was homogenized in lysis buffer (MagMAX-96 RNA Isolation Kit; Ambion, Austin, TX) with glass beads using a FastPrep instrument (FP120, Bio 101, Thermo Savant; Qbiogene). Total RNA from the homogenized samples was extracted by MagMAX Express (Applied Biosystems) by using the MagMAX-96 RNA Isolation Kit (AM 1830, Ambion) for tissues, following the supplier's recommendations. Next, cDNA was produced from \u223c500\u2009ng of total RNA by using the High-Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer\\'s recommendations. The resulting cDNA was used as template for the qPCR reaction. The TaqMan^\u00ae^ Fast Universal Master Mix (2\u00d7) (4366073, Applied Biosystems, Thermo Fisher Scientific, Waltham, USA) was used together with specific TaqMan primers to assay for GLP-1, PYY and GPR41, and GPR43, FXR and TGR5 (Supplementary Table\u00a0[1](#MOESM1){ref-type=\"media\"}). The cDNA was amplified in duplicate on a StepOnePlus instrument (Applied Biosystems). Gene expression was calculated by the comparative cycle threshold (CT) method. The geometrical mean of ACTB and HPRT1 gene expression was used for normalization of the qPCR results. The expression of target genes was normalized to the geometrical mean of the reference genes by calculating \u0394C~T~\u2009=\u2009C~T(target)~\u2009\u2212C~T(GEOMEAN\\ reference).~ Fold change in gene expression was calculated as 2^\u2009\u2212\u0394\u0394CT^, where \u0394\u0394C~T\u2009=~\u2009\u0394C~T(sample)~\u2009\u2212\u0394C~T(calibrator)~ and the mean \u0394C~T~ of samples from control rats was used as a calibrator. Log2(\u0394\u0394C~T~) was taken to illustrate the fold change of the treatment groups compared to the control group, centred at zero.\n\nShort-chain fatty acid analysis {#Sec17}\n-------------------------------\n\nCaecum samples from 4-week-old pups (P4W) were analyzed for acetic acid, propionic acid and butyric acid by gas chromatography--mass spectrometry (GC--MS) by MS-Omics, Copenhagen, based on an established method^[@CR63]^. Briefly, the samples were homogenized in Milli-Q water by ultrasonication, acidified with HCl, supernatants were transferred to GC-vials and internal standards were added. The samples were then analyzed in random order by GC--MS and processed with software based on the PARAFAC2 model. Quantified values are calculated assuming that 1\u2009mg of faeces corresponds to a volume of 1\u2009\u00b5L.\n\nFaecal gross energy analysis {#Sec18}\n----------------------------\n\nFaecal gross energy from 4-week-old pups (P4W) was assessed using a BOMB calorimeter C6000 isoperibol (IKA, Staufen, Germany) according to the manufacturer's recommendations. Briefly, the faecal samples were dried at 50\u2009\u00b0C for 2 days before burning them with a fixed amount of benzoic acid (one pill, C723, IKA) as an accelerant. Results were calculated in J\u2009g^\u22121^ according to the isoperibol mode of the device.\n\nBile acid profiling {#Sec19}\n-------------------\n\nLC--MS-grade acetonitrile, methanol and acetic acid were obtained from Sigma-Aldrich (St. Louis, MO, USA). All aqueous solutions were prepared using ultrapore water obtained from a Millipore Milli-Q Gradient A10 system (Millipore, Bedford, MA). Oasis HLB 1 cc cartridges (30\u2009mg\u2009mL^\u22121^) were purchased from Waters (Milford, MA, USA).\n\nAuthentic compounds of 12 unconjugated, 9 taurine conjugated and 5 glycine-conjugated bile acids were ordered from either Sigma-Aldrich (Sigma-Aldrich, Schnelldorf, Germany) or Steraloids (Newport, RI, USA). Of these, four bile acids, which were absent in the rats, were used as internal standards (IS), namely glycoursodeoxycholic acid (GUDCA), dehydrocholic acid (DHCA), 23-nordeoxycholic acid (23-NDCA) and glycolithocholic acid (GLCA). The other 22 bile acids, which were to be quantified, included tauro-\u03c9-muricholic acid (T\u03c9MCA), tauro-\u03b1-muricholic acid (T\u03b1MCA), tauro-\u03b2-muricholic acid (T\u03b2MCA), tauroursodeoxycholic acid (TUDCA), taurohyodeoxycholic acid (THDCA), taurocholic acid (TCA), glycocholic acid (GCA), glycohyodeoxycholic acid (GHDCA), taurochenodeoxycholic acid (TCDCA), taurodeoxycholic acid (TDCA), \u03c9-muricholic acid (\u03c9MCA), \u03b1-muricholic acid (\u03b1MCA), 7-ketodeoxycholic acid (7-KDCA), glycochenodeoxycholic acid (GCDCA), \u03b2-muricholic acid (\u03b2MCA), hyocholic acid (HCA), taurolithocholic acid (TLCA), cholic acid (CA), ursodeoxycholic acid (UDCA), hyodeoxycholic acid (HDCA), chenodeoxycholic acid (CDCA) and deoxycholic acid (DCA). Details on the authentic compounds are provided in Supplementary Table\u00a0[2](#MOESM1){ref-type=\"media\"}.\n\nStock solutions (1\u2009mg\u2009mL^\u22121^) of 26 bile acids were individually prepared from their authentic compounds. Conjugated bile acids were prepared in 50% methanol and unconjugated bile acids were prepared in methanol. An IS solution containing four bile acids, namely GUDCA, DHCA, 23-NDCA and GLCA (1\u2009mg\u2009mL^\u22121^ each), was prepared in 50% methanol. The other 22 bile acids were mixed (BA mix) and further diluted with 50% methanol to give final concentrations of 0.01\u2009\u00b5g\u2009mL^\u22121^, 0.1\u2009\u00b5g\u2009mL^\u22121^, 1\u2009\u00b5g\u2009mL^\u22121^ and 4\u2009\u00b5g\u2009mL^\u22121^. Serum samples from dams at termination and pups at week 2, 4 and 14 were prepared in Eppendorf tubes where an aliquot of the IS solution (50\u2009\u00b5L, 1\u2009\u00b5g\u2009mL^\u22121^) had been dried with nitrogen. To each tube, 40\u2009\u00b5L of serum and 100\u2009\u00b5L of acetonitrile were mixed, vortexed and left at \u221220\u2009\u00b0C for 30\u2009min to precipitate the proteins. Lumen content of colon (100--150\u2009mg) from pups at week 4 was diluted 1:2 with the IS solution (4\u2009\u00b5g\u2009mL^\u22121^), vortexed for 10\u2009s and centrifuged at 16,000\u2009\u00d7\u2009*g*, 4\u2009\u00b0C for 10\u2009min. Subsequently, 150\u2009\u00b5L of colonic evacuate was mixed with 350\u2009\u00b5L of acetonitrile, vortexed and then left at \u221220\u2009\u00b0C for 20\u2009min to precipitate the proteins. After precipitation of proteins, serum and colon samples were centrifuged at 16,000\u2009\u00d7\u2009*g*, at 4\u2009\u00b0C for 10\u2009min in a microcentrifuge and each supernatant was loaded to a HLB cartridge (each cartridge was preconditioned with 1\u2009mL of 70% acetonitrile prior to use) to deplete phospholipids. After sample loading, the flow-through fraction was collected in a new tube. Subsequently, each cartridge was washed with 300\u2009\u00b5L of 70% acetonitrile, and the flow-fractions were pooled and dried with nitrogen. Dried samples were stored at \u221220\u2009\u00b0C until use.\n\nSerum and colon samples were reconstituted in 50 and 600\u2009\u00b5L of 50% methanol, respectively. Serum and colon samples were analyzed in random order in two separate runs. We used a modified protocol from our previous study^[@CR58]^: for each sample, a volume of 2\u2009\u00b5L was injected into an ultra-performance liquid chromatography quadrupole time-of-flight mass spectrometry (UPLC-QTOF-MS) system consisting of a Dionex Ultimate 3000 RS liquid chromatograph (Thermo Scientific, CA, USA) coupled to a Bruker maXis time- of-flight mass spectrometer equipped with an electrospray interphase (Bruker Daltonics, Bremen, Germany) operating in negative mode. The analytes were separated on a Poroshell 120 SB-C18 column with a dimension of 2.1\u2009\u00d7\u2009100\u2009mm and 2.7-\u00b5m particle size (Agilent Technologies, CA, USA). The column was held at 30\u2009\u00b0C and the sampler at 4\u2009\u00b0C. The UPLC mobile phases consisted of 0.01% acetic acid in water (solution A) and acetonitrile (solution B). The analytes were eluted using the following gradient: Solvent programming started with a linear gradient from 10 to 25% B in 10\u2009min at a flow rate of 0.3\u2009mL\u2009min^\u22121^. The flow rate was then increased to 0.4\u2009mL\u2009min^\u22121^ as the gradient went from 25 to 60% B until 18\u2009min, where the gradient was increased to 98% B until 20\u2009min. The gradient was kept constant at 98% B until 23\u2009min, where the flow rate was decreased to 0.3\u2009mL\u2009min^\u22121^. The gradient was returned to initial conditions at 23.1\u2009min and re-equilibrated until 25\u2009min. Mass spectrometry data were collected in full scan mode at 2\u2009Hz with a scan range of 50--1000\u2009mass/charge (m/z). The following electrospray interphase settings were used: nebulizer pressure 2\u2009bar, drying gas 10\u2009L\u2009min^\u22121^, 200\u2009\u00b0C and capillary voltage 4500\u2009V. To improve the measurement accuracy, external and internal calibrations were done using sodium acetate clusters (Sigma-Aldrich, Schnelldorf, Germany) and in addition a lock-mass calibration was applied (hexakis(1H,1H,2H,perfluoroetoxy)phosphazene, Apollo Scientific, Manchester, UK).\n\nBile acids were detected by extraction of ion chromatograms using the exact mass of the bile acids\u2009\u00b1\u20090.002\u2009Da and quantified by use of standard curves and IS with similar retention times as listed in Table\u00a0[S1](#MOESM1){ref-type=\"media\"}. Data were processed using QuantAnalysis version 2.2 (Bruker Daltonics, Bremen, Germany) and calibration curves were established by plotting the peak area ratios of all of the analytes with respect to the IS against the concentrations of the calibration standards. The calibration curves were fitted to quadratic regression.\n\nFaecal bacterial load {#Sec20}\n---------------------\n\nSerial dilutions of fresh faecal samples collected from dams 2 days before expected delivery (D0W) were prepared in peptone saline diluent and plated on Wilkins-Chalgren agar (WCA, Oxoid) and plate count agar (PCA, Oxoid). Plates were incubated at 37\u2009\u00b0C for 2\u2009days under anaerobic and aerobic conditions, respectively, before enumeration.\n\nELISA {#Sec21}\n-----\n\nBlood serum levels of haptoglobin and GLP-1 in 4-week-old rat pups (P4W) were determined by use of commercial sandwich ELISA assays (EKR750 and EKR1992, respectively, Nordic BioSite ApS, Sweden) as recommended by the supplier. Serum samples for haptoglobin measurements were diluted 1:500 and for GLP-1 1:2. All samples were at least assayed in duplicates.\n\nData handling and statistics {#Sec22}\n----------------------------\n\nAll statistical analysis was conducted in GraphPad Prism 7 (GraphPad Software Inc., La Jolla, CA) unless otherwise stated. Generally, the unpaired *t* test or the non-parametric Mann--Whitney test, when appropriate, was used to determine differences between the control group and treatment groups. Animal weight of the pups was calculated by averaging the individual weights of all pups in the appropriate cage. Average food intake per animal per day was calculated by dividing the total food intake with the number of animals per cage and the number of days between measurements. Statistical analysis of the 16S rRNA gene sequencing data to identify differentially represented bacterial groups was performed by permutation-based *t* tests with 10,000 iterations in R, and correction for multiple testing by the Benjamini--Hochberg false discovery rate method^[@CR64]^. Differences in total concentration of serum bile acids in pups at the different time points and between the three groups of dams, respectively, were tested by one-way ANOVA with Tukey's multiple comparison test. For multivariate analysis of bile acids, the data were auto-scaled and imported into LatentiX (version 2.11) (Latent5) for PCA to assess the variation of the data. Spearman's rank correlation coefficients were determined to investigate associations between the relative abundance of bacterial genera, alpha-diversity, SCFAs and gene expression as well as between bile acids and FXR and TGR5 gene expression.\n\nElectronic supplementary material\n=================================\n\n {#Sec24}\n\nSupplementary Information\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nElectronic supplementary material\n=================================\n\n**Supplementary information** accompanies this paper at 10.1038/s42003-018-0140-5.\n\nThe present work was funded by a grant from the Danish Council for Independent Research, Technology and Production Sciences (DFF-1335-00092). We thank Anne \u00d8rngreen, and her group for handling of animals, Li Zhang and Louise K. Vigsn\u00e6s for help during the animal dissections, Bodil Madsen and Anni Mehlsen for excellent technical support and Marlene Danner Dalgaard at the DTU Multi-Assay Core facility for performing the 16S rRNA sequencing. Morten Danielsen and MS-Omics are thanked for performing the SCFA analysis.\n\nM.I.B., M.V.L.T., T.R.L., H.M.R. and H.F. conceived and designed the study. M.V.L.T., H.M.R., I.C.T. and H.L.F. performed the experimental work. M.V.L.T., H.M.R., I.C.T., H.L.F. and M.I.B. performed the analytical work. The manuscript was drafted by M.V.L.T. and M.I.B. and all authors contributed to interpretation of the results and approved the final manuscript.\n\nThe 16S rRNA gene sequence datasets generated during the current study are deposited in the NCBI Sequence Read Archive with the accession number SRP131746. All other relevant data are available from the authors.\n\nCompeting interests {#FPar1}\n===================\n\nThe authors declare no competing interests.\n"}
+{"text": "1. Introduction {#sec1-polymers-12-00287}\n===============\n\nThe adsorption of polymers to particulate surfaces continues to attract much attention due to its diverse industrial relevance, and depending on the specific application, the role of the polymer may be to promote stabilization of the dispersion, to induce flocculation or otherwise destabilize the dispersion, or a combination of both, dependent on the polymer concentration \\[[@B1-polymers-12-00287],[@B2-polymers-12-00287],[@B3-polymers-12-00287],[@B4-polymers-12-00287],[@B5-polymers-12-00287]\\]. The adsorption of neutral homopolymers and copolymers onto surfaces is driven by a complex and subtle interplay between molecular weight and concentration-dependent enthalpic and entropic factors in defining the amount of adsorbed polymer, and inter alia its conformation at the particle surface \\[[@B6-polymers-12-00287],[@B7-polymers-12-00287]\\]. The presence of charge on the polymer leads to additional contributions to the thermodynamics of adsorption when exposed to both oppositely and like-charged surfaces, and usually, the electrostatic force dominates \\[[@B8-polymers-12-00287],[@B9-polymers-12-00287],[@B10-polymers-12-00287]\\]. The structure of the resultant polyelectrolyte-particle complex is defined by the polyelectrolyte and surface charge densities, polymer molecular weight and architecture, salt concentration/ionic strength of the medium, nonelectrostatic polyelectrolyte-surface interactions and the quality of the solvent \\[[@B11-polymers-12-00287],[@B12-polymers-12-00287],[@B13-polymers-12-00287]\\].\n\nFlocculation-inducing adsorption of polyelectrolytes to particle surfaces has been studied in some detail (e.g., \\[[@B14-polymers-12-00287],[@B15-polymers-12-00287],[@B16-polymers-12-00287],[@B17-polymers-12-00287],[@B18-polymers-12-00287]\\]), and generally, there are three dominant mechanisms dependent on the molecular weight and/or charge density of the polyelectrolyte: (i) neutralization of the charge on the particle surface leading to attractive van der Waals-driven aggregation; (ii) interparticle bridging by low/medium charge density, high-molecular-weight polymers; and (iii) attraction between uncoated surfaces and oppositely charged patches on adjacent particles formed by the localised adsorption of high charge density, low-molecular-weight polyelectrolytes \\[[@B19-polymers-12-00287],[@B20-polymers-12-00287],[@B21-polymers-12-00287]\\].\n\nThe impact of added surfactant on polyelectrolyte-induced flocculation of particle dispersions has also been explored. For example, Barany and Skvarla \\[[@B22-polymers-12-00287]\\] studied the adsorption of both cationic and anionic high-molecular-weight, high-charge-density polyelectrolytes to carboxylated polystyrene particles in the presence of hexadecyl pyridinium bromide (HPB) or anionic surfactant sodium dodecyl sulfate (SDS). Taubaeva et al. \\[[@B23-polymers-12-00287]\\] extended that study to bentonite dispersions and showed that addition of mixtures of cationic surfactants and polyelectrolyte induces a substantial reduction in the negative charge on the particles, including a reversal of the sign of their charge, whereas, the addition of anionic surfactant/polyelectrolyte mixtures leads to a substantial increase in the particle charge. Besra et al. \\[[@B24-polymers-12-00287]\\] studied the flocculation of kaolin suspensions with high-molecular-weight polyacrylamide flocculants (cationic, anionic and nonionic) of varying ionic character in the presence of cetyl trimethyl ammonium bromide (CTAB), SDS, and Triton-X 100. Karlson et al. \\[[@B25-polymers-12-00287]\\] explored the interaction of two cellulosic polymers, ethyl hydroxyethyl cellulose (EHEC) and hydrophobically modified ethyl hydroxyethyl cellulose (HM-EHEC), with both polystyrene latex and silica particles, in the presence of SDS, CTAB, and octaethylene glycol mono n-dodecyl ether (C~12~E~8~). Lele and Tilton \\[[@B26-polymers-12-00287]\\] examined the depletion and structural forces operating between silica spheres and plates in solutions containing the anionic polyelectrolyte sodium polyacrylate (Na-PAA) and SDS, and observed situations where the depletion attraction was synergistically enhanced or antagonistically weakened and where one species was the dominant depletant and indistinguishable from its single component behavior. Matusiak and Grz\u0105dka \\[[@B27-polymers-12-00287]\\] studied the efficiency of cationic starch to flocculate silica dispersions in the presence of anionic surfactants, highlighting the switch from bridging flocculation at low-polymer concentrations to a depletion mechanism at higher concentrations, surfactant-induced enhancement in polymer adsorption at low-surfactant concentrations and polymer-surfactant complex adsorption at higher concentrations.\n\nCationic hydroxyethylcellulose (cat-HEC) polymers, produced by quaternization of hydroxyethylcellulose, are widely used in water-based formulations, since the cationic groups along the polymer backbone lead to significant increases in viscosity and offer control over colloidal stability \\[[@B28-polymers-12-00287]\\]. Extending the work of Shubin \\[[@B29-polymers-12-00287]\\], Terada et al. examined the effect of the anionic surfactant SDS on the adsorption of cationic hydroxyethyl celluloses to silica surfaces (in the presence of 10 mM NaCl) using null ellipsometry. They showed that the surfactant-concentration-dependent associative binding of the surfactant to the polymer defined the interfacial behaviour, that maximum adsorption was observed in the low-surfactant concentration one-phase region, whilst higher SDS concentrations led to desorption of the complex from the surface \\[[@B30-polymers-12-00287]\\].\n\nThis study focuses on the interaction of cat-HEC with anionic particles, in the presence of the anionic surfactant sodium dodecyl sulfate (SDS) and its mixture with the nonionic surfactant hexaethylene glycol monododecyl ether (C~12~E~6~), to explore how the degree of apparent cationic charge on the polymer defines its ability to adsorb to silica and thus, to understand how the surfactant(s) modulate(s) the polymer-particle interaction.\n\n2. Materials and Methods {#sec2-polymers-12-00287}\n========================\n\nFour cationic cellulosic polymers have been explored in this work, in concert with an anionic silica dispersion (Ludox TM-50, particle size 22 (+/--2) nm, surface area 140 m^2^\u00b7g^\u22121^) and a hydrophilic, anionic polystyrene-polybutadiene latex (particle size 100 (+/--5) nm, surface area 53 m^2^\u00b7g^\u22121^). The majority of samples have been prepared in deionized water, with the particle dispersions being extensively dialyzed prior to use. Sodium dodecyl sulfate (SDS) and hexaethylene glycol monododecyl ether (C~12~E~6~) were used as received (Sigma-Aldrich). For both surfactants, the surface tension data showed no minima characteristic of significant levels of impurity, and the mass spectrometry and NMR data were consistent with the accepted structures.\n\nThe cationic cellulosic polymers are commercial in nature, [Scheme 1](#polymers-12-00287-sch001){ref-type=\"scheme\"} (Dow Chemical Company), and will be referred to by their nominal degrees of modification, expressed in terms of the nitrogen content; N = 0.5% (\u00b10.1)%, N = 0.95% (\u00b10.15)% and N = 1.8% (\u00b10.2)%. The parent polymer has also been included for comparison, and is labelled N = 0. The manufacturer's literature suggests these polymers have molecular weights in the vicinity of 500,000 g\u00b7mol^\u22121^.\n\nOrder of mixing is known to be a significant factor in these systems, and therefore all samples were prepared by adding polymer stock solutions (or their dilutions) to the particulate dispersions, and where surfactant (SDS or C~12~E~6~) is additionally present, this was added to the polymer stock solution prior to adding to the silica. The importance of order of addition highlights the nonequilibrium effects often observed in these systems.\n\nThe solvent-relaxation measurements were determined on a Zigo Acorn Drop bench-top NMR spectrometer operating at 13.0 MHz. Samples were equilibrated at 25 \u00b0C. The instrument's in-built configuration macros and data-analysis routines were used to optimize and obtain the experimental relaxation rates. All results are the average of at least triplicate measurements and frequent duplicate samples.\n\n3. Results and Discussion {#sec3-polymers-12-00287}\n=========================\n\nMixing oppositely charged polymers and particles leads to complex phase separation, characterized by two discrete phases, one phase being opaque, and therefore very rich in large(r) structures, the second showing various degrees of opalescence due to the presence of dispersed nanoparticles. [Figure 1](#polymers-12-00287-f001){ref-type=\"fig\"}a--d presents photographs of a large matrix of a series of samples in which the concentration of silica has been increased from 0 to 10 wt % in the presence of fixed amounts of polymer (typically 1000 ppm) at selected values of SDS concentration (0 \u2264 \\[SDS\\] \u2264 1 mM). Our previous electrophoretic NMR characterization of the binary polymer/SDS systems indicates that the polymer/SDS solution phases separate (due to charge neutrality) around typically 4--8 mM for a polymer concentration C*~polymer~* = 1 wt %, suggesting that over the concentration window studied here (0--1 mM SDS), the polymer retains some cationic character, except perhaps at the higher SDS values studied \\[[@B32-polymers-12-00287]\\]. [Figure 2](#polymers-12-00287-f002){ref-type=\"fig\"} presents a subset of these systems additionally in the presence of the nonionic surfactant, hexaethylene glycol monododecyl ether (C~12~E~6~). It should be noted that the uncharged polymer showed no measurable interaction with SDS.\n\nTraversing the row of samples from left to right, i.e., the only variable increasing being the silica concentration, the samples become increasingly opaque, with the volume of the opaque phase largely increasing. A similar pattern is observed in the uncharged and charged cases. Where SDS has been added, the volume of the opaque phase also depends on \\[SDS\\]. In some cases, the opaque phase is the supernatant, and in some cases the subnatant. As one moves vertically through the series, the same gross pattern is observed, but it is clear that the switch from super- to subnatant is displaced, and the phase volumes vary.\n\nFurther interpretation of these visual phase diagrams is challenging, as the composition of each phase is unknown, and that of the opaque phase continues to change with time. Given the density of the silica, it is slightly surprising that the opaque phase creams at all. If one removes the dense phase, and allows it to stand, it will further phase separate, leading to additional volume of clear phase and more dense phases. Therefore, to accelerate this phase separation, the samples were gently centrifuged.\n\nIn order to understand the composition of the separated phases, the less dense (termed the equilibrium phase) has been characterized by both solvent relaxation and by dry weight analysis. Making the assumption that the bulk of the weight of the dried sample recorded in this manner arises from the particles, we therefore define this as an equilibrium particle concentration, and thus provides a convenient parameter (*x*-axis) on which to base solvent relaxation data, and thereby to enable in parallel the comparison with bare silica dispersions over the same initial concentration range. For the silica comparators, no differences between initial and equilibrium concentrations were observed in the dry weight analysis. The solvent relaxation in these no-polymer systems can therefore be compared with the with-polymer data.\n\nSolvent relaxation provides a convenient method to characterize any near-surface structures via the dynamics of the solvent molecules, in this case, water \\[[@B33-polymers-12-00287],[@B34-polymers-12-00287],[@B35-polymers-12-00287],[@B36-polymers-12-00287],[@B37-polymers-12-00287],[@B38-polymers-12-00287],[@B39-polymers-12-00287]\\]. The relaxation rate is usually enhanced if a polymer layer is present, compared with the same surface area, due to the presence of a volume of trapped water. [Figure 3](#polymers-12-00287-f003){ref-type=\"fig\"} presents the solvent-specific relaxation rate, i.e., the rate normalized to pure water, plotted in terms of the equilibrium silica concentration. All samples studied here have been reproduced on this single figure. As may be seen, there is a universality to this curve that relates the measured specific solvent relaxation rate to the equilibrium silica concentration, this universality holding for particles that have been exposed to polymer, to surfactant and to mixtures of polymer and surfactant, as well as those that have not been exposed.\n\nMaking the reasonable assumption that any added polymer or surfactant that adsorbs to the surface would perturb the solvent-specific relaxation rate \\[[@B33-polymers-12-00287],[@B34-polymers-12-00287],[@B35-polymers-12-00287],[@B36-polymers-12-00287],[@B37-polymers-12-00287],[@B38-polymers-12-00287],[@B39-polymers-12-00287]\\], this universality strongly suggests that in all cases the particles detected in the solvent relaxation experiment are uncoated. In other words, the opaque phase consists of a concentration of polymer and particle with a composition to saturate that interaction, and any excess polymer or particle is displaced to the less-dense phase. When there is excess polymer present, all of the silica is removed, and one observes a solvent-specific relaxation rate that is characteristic of a simple polymer solution which is identical to the pure solvent, R~2sp~ tends to 0. When there is excess silica present, the polymer removes a fraction of the silica, and the solvent-specific relaxation rate is consistent with a bare silica dispersion commensurate with the remaining equilibrium concentration.\n\nIn those samples that also contain SDS, it is clear that the same universality between mass fraction and surface area exists. In this case, the anionic surfactant binds to the cationic polymer, thereby reducing the effective charge on the polymer, such that for a given polymer concentration, the polymer-surfactant complex removes less silica from the dispersion compared to the polymer alone. Nonetheless, all the detected silica surface is bare. In other words, in terms of the photographs in [Figure 1](#polymers-12-00287-f001){ref-type=\"fig\"} and [Figure 2](#polymers-12-00287-f002){ref-type=\"fig\"}, addition of SDS moves the phase diagram to the left. Similarly, one can also estimate the concentration of SDS required to nullify the charge on the cationic polymer, and therefore \"turning off\" any adsorption to the anionic surface.\n\nFurther confirmation of this simple binary phase separation may be sought by considering the zeta-potentials of the equilibrium (phase separated) silica phases ([Table 1](#polymers-12-00287-t001){ref-type=\"table\"}), these values being consistent with the original bare silica dispersion.\n\nAs an aside, a similar observation may be drawn from analogous studies using polystyrene-butadiene latex rather than silica ([Figure 4](#polymers-12-00287-f004){ref-type=\"fig\"}).\n\nConsider the impact of addition of a nonionic surfactant to the polymer/particle/anionic system. The two surfactants would be expected to interact strongly and synergistically, and it is known that the presence of the nonionic often leads to an enhancement in the counterion dissociation of the anionic surfactant, effectively increasing the charge on the mixed surfactant complex.\n\n[Figure 5](#polymers-12-00287-f005){ref-type=\"fig\"} presents the solvent-specific relaxation rate for a select set of experiments in which the cationic polymer has been first exposed to mixed anionic-nonionic (SDS-C~12~E~6~) surfactant blends, before being added to the silica dispersion. Phase separation still occurs, with the visual pattern not too dissimilar to that observed in the absence of the nonionic surfactant, and reminiscent of that observed at lower anionic surfactant concentrations. As may be seen, the same universality exists, between the dry weight analysis and the nature of the silica surface, implying the nonionic surfactant is merely a spectator in the polymer/SDS/particle interaction.\n\nThe uncharged (parent) polymer induces flocculation of the silica dispersion via a bridging mechanism, most likely through the hydroxyl groups. The grossly similar behavior of the three charged polymers indicates that the presence of the charge on the polymer leads to an increased capacity to remove silica, with the electrostatic interaction promoting the polymer bridging. Adding surfactant to the polymer solution induces complex formation through an associative interaction, which stoichiometrically reduces the prevailing charge on the polymer, and inter alia, its capacity to flocculate the silica. The uniformity of all these datasets, and the fact that all the surfaces detected in this manner are bare, indicates that the polymer-surfactant interaction must dominate the polymer-particle interaction, as found by Terada et al. \\[[@B30-polymers-12-00287]\\], thus accounting for the lack of dependence of the phase diagram or solvent-specific relaxation rate on \\[SDS\\]. Two populations of particles exist in these slowly flocculating systems---bare (equilibrium phase) and fully/partially coated particles (dense phase), the balance of which is defined by the polyelectrolyte concentration, or more accurately, the concentration of oppositely charged polymer segments, the latter also defined by the presence of any added competing surfactant. Adding anionic surfactant that competes preferentially for the cationic charges on the polymer weakens the polymer-particle interaction. Accordingly, experimental protocols that focus on the particle behavior (e.g., electrophoretic mobility) might not adequately capture this bare/coated particle distribution.\n\n4. Conclusions {#sec4-polymers-12-00287}\n==============\n\nDispersions of anionic particles are often destabilized by the addition of cationic polymers. Here, it is shown that macroscopic phase separation may be induced through a bridging mechanism when cationic hydroxycellulose derivatives are added to silica (and latex) dispersions, with the high-molecular-weight polymer aggregating sufficient particles to neutralize its charge completely, leaving any excess particle surface uncoated, i.e., a coexisting dense phase of aggregated material and an equilibrium phase of bare particles. This charge-neutralization-driven adsorption leads to a simple universal relationship between the initial polymer concentration, the degree of cationic modification of the polymer and the mass (or more precisely, surface area) of silica removed from solution. The same universality exists when anionic surfactants are present, with the appropriate modulation of the phase behavior being understandable in terms of the stoichiometric binding of surfactant to polymer.\n\nUniversity of Greenwich and Unilever are gratefully acknowledged for the provision of a Ph.D. studentship to W.A.\n\nAll authors were involved in the conceptualization of this study, which was conducted by M.C., P.C.G. and W.A., who were responsible for compiling the original draft preparation, whilst all authors were responsible for the final version. All authors have read and agreed to the published version of the manuscript.\n\nThis research was partially funded by Unilever in the form of a Ph.D. Studentship.\n\nThe authors declare no conflict of interest. The funders had a partial role in the design of the study, but the data collection, analyses and interpretation were conducted by W.A.\n\nFigures, Scheme and Table\n=========================\n\n![Generic structure of quaternary ammonium salts of hydroxyethyl cellulose polymers. Adapted from ref \\[[@B31-polymers-12-00287]\\].](polymers-12-00287-sch001){#polymers-12-00287-sch001}\n\n###### \n\n(**a**) Visual appearance of silica/polymer mixtures at 25 \u00b0C. The (uncharged) polymer (N = 0%) concentration is 1000 ppm throughout, the silica concentration is increased from 0 to 10 wt %, viz., 0 (left), 0.2, 0.5, 1.0, 2.0, 4.0, 6.0, 8.0, 10.0 (right). Photographs were taken approximately 2 days after preparation. (**b**) Visual appearance of silica/polymer/sodium dodecyl sulfate (SDS) mixtures at 25 \u00b0C. The polymer (N = 0.5%) concentration is 1000 ppm throughout, the silica concentration is increased from 0 to 10 wt % and SDS is 0, 0.1 and 0.25 mM (left-hand column), 0.50, 0.75 and 1.0 mM (right-hand column). Photographs were taken approximately 2 days after preparation. (**c**) Visual appearance of silica/polymer/SDS mixtures at 25 \u00b0C. The polymer (N = 0.95%) concentration is 1000 ppm throughout, the silica concentration is increased from 0 to 10 wt % and SDS is 0, 0.1 and 0.25 mM (left-hand column), 0.50, 0.75 and 1.0 mM (right-hand column). Photographs were taken approximately 2 days after preparation. (**d**) Visual appearance of silica/polymer/SDS mixtures at 25 \u00b0C. The polymer (N = 1.8%) concentration is 1000 ppm throughout, the silica concentration is increased from 0 to 10 wt % and SDS is 0, 0.1 and 0.25 mM (left-hand column), 0.50, 0.75 and 1.0 mM (right-hand column). Photographs were taken approximately 2 days after preparation.\n\n![](polymers-12-00287-g001a)\n\n![](polymers-12-00287-g001b)\n\n![Visual appearance of silica/polymer/SDS/C~12~E~6~ mixtures at 25 \u00b0C. The polymer (N = 0.95%) concentration is 1000 ppm throughout, the silica concentration is increased from 0 to 10 wt %. The first row has SDS = 0 mM, the second 0.50 mM. In the third and fourth rows, the total surfactant concentration is 0.50 mM, but the composition of the surfactant mixtures is 75%:25% (anionic: nonionic). The last row has 0.5 mM C~12~E~6~. Photographs were taken approximately 2 days after preparation.](polymers-12-00287-g002){#polymers-12-00287-f002}\n\n![(**a**,**b**) Solvent-specific relaxation rate as a function of equilibrium particle concentration for all the cellulosic polymer/anionic surfactant/silica systems studied here; polymer concentration dependence when \\[SDS\\] = 0, (**a**); surfactant concentration dependence when \\[polymer\\] = 1000 ppm. The symbols refer to the series of experiments in [Figure 1](#polymers-12-00287-f001){ref-type=\"fig\"}a--d and [Figure 2](#polymers-12-00287-f002){ref-type=\"fig\"}, as per figure legend.](polymers-12-00287-g003){#polymers-12-00287-f003}\n\n![Solvent-specific relaxation rate as a function of equilibrium particle concentration for cationic polymer N = 0.95% at 1000 ppm for anionic silica (circles, open symbols no polymer, filled symbols with polymer) and anionic latex systems (squares, open symbols no polymer, filled symbols with polymer).](polymers-12-00287-g004){#polymers-12-00287-f004}\n\n![Solvent-specific relaxation rate as a function of equilibrium particle concentration for polymer N = 0.95% at 1000 ppm in the presence of anionic-nonionic surfactant blends; 0 (simple SDS) red diamonds, 0.75 blue triangles and 1.0 (simple nonionic) purple triangles. Additionally shown for comparison is the silica only dataset (black circles) and the SDS = 0 dataset. The total surfactant concentration is 0.5 mM.](polymers-12-00287-g005){#polymers-12-00287-f005}\n\npolymers-12-00287-t001_Table 1\n\n###### \n\nZeta-potential analysis of the equilibrium phases obtained from the centrifugation of the two-phase silica/polymer/surfactant systems studied here. All measurements were conducted in triplicate.\n\n  Analysis of the Equilibrium Phase Extracted from the Following Samples   Zeta Potential/mV\n  ------------------------------------------------------------------------ -------------------\n  Stock, as-supplied Ludox, 1-in-10 dilution                               \u221240.4 (\u00b11.8)\n  Dialysed Ludox                                                           \u221235.5 (\u00b11.5)\n  Dialysed Ludox plus polymer N = 0.5%                                     \u221230.9 (\u00b11.3)\n  Dialysed Ludox plus polymer N = 0.95%                                    \u221230.8 (\u00b11.7)\n  Dialysed Ludox plus polymer N = 1.8%                                     \u221231.4 (\u00b12.4)\n  Dialysed Ludox plus polymer N = 0.5% and SDS                             \u221231.4 (\u00b11.6)\n  Dialysed Ludox plus polymer N = 0.95% and SDS                            \u221229.9 (\u00b12.2)\n  Dialysed Ludox plus polymer N = 1.8% and SDS                             \u221232.6 (\u00b12.8)\n"}
+{"text": "1. Introduction\n===============\n\nNitrophenols are important and versatile organic compounds in industrial, agricultural and defense applications. They are frequently used as intermediates in the manufacture of explosives, pharmaceuticals, pesticides, pigments, dyes, rubber chemicals and so on \\[[@b1-ijms-9-3-316]\\]. They are produced by microbial hydrolysis of several organophosphorous pesticides, such as parathion \\[[@b2-ijms-9-3-316], [@b3-ijms-9-3-316]\\] or by photodegradation of pesticides that contain the nitrophenol moiety \\[[@b4-ijms-9-3-316], [@b5-ijms-9-3-316]\\]. Nitrophenols also result from natural processes in the biosphere and are now common pollutants in several ecosystems in developed countries \\[[@b6-ijms-9-3-316]\\]. The compounds *o*-nitrophenol, *p*-nitrophenol and 2,4-dinitrophenol are listed on the United State Environmental Protection Agency\\'s (USEPA\\'s) \"Priority Pollutants List\" \\[[@b7-ijms-9-3-316]\\]. Therefore, the monitoring of *o*-nitrophenol is essential for environmental pollution control.\n\n*o*-Nitrophenol poses significant health risks since it is highly toxic to mammals, microorganisms and anaerobic bacteria. Its toxicity is thought to be due to the nitro group being easily reduced by enzymes to a nitro anion radical, nitroso and hydroxylamine derivatives \\[[@b1-ijms-9-3-316], [@b8-ijms-9-3-316], [@b9-ijms-9-3-316]\\]. These derivatives are responsible for the cytotoxic, mutagenic and carcinogenic action of nitro compounds \\[[@b10-ijms-9-3-316], [@b11-ijms-9-3-316]\\]. Experiments demonstrated that the nitro group reduction products react with nucleic acids *in vitro* and it is believed that DNA is the main target *in vivo* \\[[@b12-ijms-9-3-316]\\]. It was reported that the nitro anion radical is the species principally responsible for DNA damage, oxidizing the DNA double helix and liberating thymidine phosphate \\[[@b13-ijms-9-3-316], [@b14-ijms-9-3-316]\\]. The damage to DNA depends on the stability of the intermediate product formed on nitro reduction. Electrochemical methods have been used to study the reduction and mechanism of action of nitro compounds as an antimicrobial agent \\[[@b15-ijms-9-3-316]\\].\n\nAt present, there is a wealth of quantitative information on nitrophenol toxicity in anaerobic treatment systems. More information is required to determine the mechanism of toxicity of these organic compounds \\[[@b4-ijms-9-3-316]\\]. Some heterocyclic nitro compounds, such as metronidazole and benznidazole have been studied by DNA-biosensor \\[[@b16-ijms-9-3-316], [@b17-ijms-9-3-316]\\]. So an electrochemical DNA-modified GCE, prepared by immobilizing dsDNA on a GCE surface, is used to investigate possible DNA damage caused by *o*-nitrophenol. The electrochemical DNA-biosensor enables us to evaluate and predict DNA interaction and damage by health hazardous compounds, based on their binding to nucleic acid, exploring the use of voltammetric techniques for generation of reactive intermediates, which react with DNA *in situ* \\[[@b18-ijms-9-3-316], [@b19-ijms-9-3-316]\\]. The occurrence of damage to dsDNA can be detected by monitoring the oxidation current of the purine bases in DNA molecules. The interpretation of electrochemical data probably makes for elucidation the mechanism of DNA damage \\[[@b20-ijms-9-3-316]\\].\n\nIn this paper, the electrochemical behavior of *o*-nitrophenol on GCE was explored using cyclic voltammetry and differential pulse voltammetry. The redox mechanism of *o*-nitrophenol on GCE was discussed. The interaction of *o*-nitrophenol with DNA was investigated with immobilizing DNA on glassy carbon electrode and with adding DNA to solution. The change of electrochemical parameters indicated the *o*-nitrophenol could interact with DNA. Moreover, the oxidative damage to DNA was detected by obtaining the oxidative signal of the purine base in DNA molecules. The possible DNA damage mechanism was proposed.\n\n2. Results and discussion\n=========================\n\n2.1. Electrochemical behavior of o-nitrophenol\n----------------------------------------------\n\nThe reduction of nitrobenzenes is a complex process, especially for those compunds bearing hydroxyl or amine groups in the *ortho*- or *para*- positions \\[[@b21-ijms-9-3-316]\\]. The electrochemical behavior of *o*-nitrophenol on GCE was explored with cyclic voltammetry. [Figure 1](#f1-ijms-9-3-316){ref-type=\"fig\"} showed the cyclic voltammograms of 5.0\u00d710^\u22125^ M *o*-nitrophenol in 0.10 M pH 4.5 acetate buffer under different electrochemical conditions. [Figure 1A](#f1-ijms-9-3-316){ref-type=\"fig\"} was obtained with cycling between \u2212 0.80 V and + 1.20 V, initial potential \u2212 0.30V. A very large irreversible cathodic peak (R~1~) and an irreversible anodic peak (Ox~5~) appeared at around E~pc~ = \u2212 0.602 V and E~pa~ = + 0.998 V in the first cycle, respectively. The R~1~ peak corresponded to a four electrons / four protons reaction, forming hydroxylamine \\[[@b17-ijms-9-3-316], [@b22-ijms-9-3-316]\\], typical of a nitro reduction in acidic aqueous solution, and decreased in subsequent cycles. Peak Ox~5~ at E~pa~= + 0.998 V, was related to the irreversible oxidation of the phenolic group and also decreased in following cycles due to the phenol polymerization at the electrode surface \\[[@b18-ijms-9-3-316]\\]. It was interesting to observe another irreversible anodic peak (Ox~4~) at E~pa~ = + 0.335 V in the second and third cycle, which probably corresponded to condensation reactions expected concerning the oxidation of the reduced intermediates formed, generating electroactive azoxy derivatives \\[[@b17-ijms-9-3-316], [@b23-ijms-9-3-316], [@b24-ijms-9-3-316]\\].\n\nHowever, two reversible waves could be distinguished if the potential ranged from \u2212 0.80 V to + 0.40 V, with initial potential at \u2212 0.30 V from negative to positive potentials scan (see [Figure 1B](#f1-ijms-9-3-316){ref-type=\"fig\"}). On the second and subsequent cycles, the voltammogram was characterized with the one reversible wave centered at \u2212 0.173 V (Ox~2~/R~2~) and the other reversible peak centered at + 0.041 V (Ox~3~/R~3~). The irreversible reduction wave R~1~, corresponding to a four electrons and four protons typical nitro reduction in weak acid aqueous solution, was present in the first scan, but the reversible wave Ox~2~/R~2~ corresponding to nitroso/hydroxylamine derivative, occurred and slowly increased in the second and the following scan, as predicted. It was similar to the behavior of this compound observed at a hanging mercury drop electrode \\[[@b15-ijms-9-3-316], [@b21-ijms-9-3-316]\\]. The latter reversible peak Ox~3~/R~3~ possibly corresponded to the pair azoxy/azo.\n\nRepetitive differential pulse voltammograms for reduction of *o*-nitrophenol is shown in [Figure 2](#f2-ijms-9-3-316){ref-type=\"fig\"}. The R~2~ and R~3~ wave only appeared in the second and following cycles after the initial *o*-nitrophenol reduction, which was the same with what observed in the cyclic voltammogram. The reduction peak R~1~ corresponding to the nitro group was decreased, while the reduction peak corresponding to the nitroso derivative R~2~ and azoxy derivative R~3~ were increased in subsequent scans. This behavior could be explained by reduction of *o*-nitrophenol leading to its consumption on the electrode surface and electrogenerating a hydroxylamine derivative that adsorbed onto the electrode surface and was oxidized to nitroso and other derivatives.\n\n[Figure 3](#f3-ijms-9-3-316){ref-type=\"fig\"} shows the differential pulse voltammogram for oxidation of *o*-nitrophenol starting at different potentials and deposition at a constant potential. It was confirmed that peak Ox~5~ was due to the irreversible oxidation of the phenolic group, as only this group is able to undergo oxidation in *o*-nitrophenol ([Figure 3](#f3-ijms-9-3-316){ref-type=\"fig\"}, curve a), and that Ox~2~, Ox~3~ and Ox~4~ were generated only after *o*-nitrophenol was reduced at \u2212 0.60 V ([Figure 3](#f3-ijms-9-3-316){ref-type=\"fig\"}, curve b and c). It caused an increase in the peak Ox~2~, Ox~3~ and Ox~4~ after applying a potential of \u2212 0.6 V for 2 minters ([Figure 3](#f3-ijms-9-3-316){ref-type=\"fig\"}, curve c), which produced abundant hydroxylamine in this process.\n\nIn a similar way to what was observed for nitroaromatic compounds \\[[@b16-ijms-9-3-316], [@b17-ijms-9-3-316]\\], *o*-nitrophenol reduction was also pH dependent in weak acidic medium. The potential of reduction peak shifted to negative values with increasing pH. The variation of E~pc~ for the reduction of *o*-nitrophenol with pH showed the slop of 52 mV/pH (data not shown). This behavior indicated the same number of electrons and protons participated in the reduction of *o*-nitrophenol.\n\n2.2. o-Nitrophenol-DNA interaction\n----------------------------------\n\n### 2.2.1. Interaction of o-nitrophenol with DNA in solution\n\nThe above results indicate that *o*-nitrophenol could exhibit fine electrochemical response, so the interaction of *o*-nitrophenol with DNA could be studied by electrochemical methods \\[[@b25-ijms-9-3-316]--[@b27-ijms-9-3-316]\\]. The typical cyclic voltammograms of 5.0\u00d710^\u22125^ M *o*-nitrophenol on glassy carbon electrode in absence and presence of dsDNA in 0.10 M pH 4.5 acetate buffer are shown in [Figure 4](#f4-ijms-9-3-316){ref-type=\"fig\"}. A decrease in peak current and shift of peak potentials to more negative values when dsDNA was added to a solution could be easily observed. The decrease of peak current was due to the diffusion of *o*-nitrophenol-DNA complex decrement, not due to the increased viscosity of the solution or the blockage of the electrode surface by an adsorbed layer of DNA. Special cyclic voltammograms were recorded for 0.5mM K~4~\\[Fe(CN)~6~\\], which cannot interact with DNA at GCE in the absence and presence of DNA. The peak current of K~4~\\[Fe(CN)~6~\\] was not nearly affected and the peak potential didn\\'t shift any after addition of DNA. Thus, there was hardly impacted on the diffusion from the changed viscosity of DNA addition, and no significant obstruction on the GCE surface from DNA adsorption. A great decrease in peak current and a shift in potential in above CV experiments was attributed to *o*-nitrophenol bound to the bulky, slowly diffusing DNA to form *o*-nitrophenol-DNA adduct, which resulted in a considerable decrease in the apparent diffusion coefficient.\n\n### 2.2.2 Interaction of o-nitrophenol with DNA immobilized at GCE surface\n\n[Figure 5](#f5-ijms-9-3-316){ref-type=\"fig\"} shows the repetitive differential pulse voltammogram of *o*-nitrophenol electrochemical reduction using a DNA-modified GCE. It could be found that its reduction behavior on a DNA-modified GCE was similar to what happened at the bare GCE ([Figure 2](#f2-ijms-9-3-316){ref-type=\"fig\"}), but the potential shifted to negative value and the peak current decreased about 12% compared to their first scan. It indicated that pre-concentration of hydroxylamine onto the DNA matrix, denoted by the increase of the reduction peak for the nitroso R~2~ and azoxy R~3~ derivative with successive scans. Thus the DNA gel on the electrode surface could accumulate the analyte on it \\[[@b18-ijms-9-3-316]\\].\n\n2.3 DNA damage\n--------------\n\nIn order to investigate that electrochemical reduction of *o*-nitrophenol generates short-lived radical that interacts with DNA causing damage, differential pulse voltammetry was applied to detect DNA damage using a DNA-modified GCE. The blank experiment, i.e. oxidation at DNA-modified GCE without *o*-nitrophenol ([Figure 6d](#f6-ijms-9-3-316){ref-type=\"fig\"}), was performed in buffer solution to test the above hypothesis. There were almost no oxidation peaks of the DNA base, because nucleobases within native DNA are hidden inside the helix and stiffness of the structure keeps them far from the electrode surface \\[[@b28-ijms-9-3-316]\\]. [Figure 6a](#f6-ijms-9-3-316){ref-type=\"fig\"} shows the differential pulse voltammetry of *o*-nitrophenol oxidation using DNA-modified GCE. There was only an oxidation peak at + 0.982 V, which corresponded to the *o*-nitrophenol oxidation ([Figure 6](#f6-ijms-9-3-316){ref-type=\"fig\"} curve a). However, the damage to DNA was detected if *o*-nitrophenol was reduced at a constant deposition potential E~d~ = \u2212 0.60 V for 2 min ([Figure 6](#f6-ijms-9-3-316){ref-type=\"fig\"}, curve c). The appearance of two oxidative peak was attributed to guanine, E~pa~ = + 0.785 V and adenosine, E~pa~ = + 1.271 V \\[[@b29-ijms-9-3-316]\\]. In addition, the peak current at E~pa~ = + 0.982 V increased. It was interesting to observe DNA damage if the reduction of *o*-nitrophenol was performed in repetitive differential pulse scanning from 0 to \u2212 0.80 V (10 scans) ([Figure 6](#f6-ijms-9-3-316){ref-type=\"fig\"}, curve b). This distinctly indicated that *o*-nitrophenol could damage to DNA after the *o*-nitro-phenol underwent reduction process. The radicals generated in reduction process of the *o*-nitrophenol interacted with DNA \\[[@b22-ijms-9-3-316], [@b30-ijms-9-3-316]\\]. The appearance of purine oxidation peak demonstrated clearly that the radicals caused distortion of DNA double helix and exposure of the bases.\n\n2.4 Mechanism of the o-nitrophenol redox and damage to dsDNA\n------------------------------------------------------------\n\nThe above experimental results provided some evidence that reduction of *o*-nitrophenol was a complex process. Its electrochemically reduction process underwent some intermediate product formation such as nitro radical anion, nitroso, hydroxylamine, etc. The radical formed by reduction of the nitro group could interact with dsDNA and cause dsDNA damage. After the *o*-nitrophenol was reduced, the detected oxidation current signal of the guanine and adenosine nucleobase supplied evidences of DNA damage. According to above experimental results and reference to \\[[@b31-ijms-9-3-316]\\], a possible mechanism of the *o*-nitrophenol redox and damage to dsDNA was proposed as follows:\n\n3. Experimental\n===============\n\n3.1 Chemicals and Apparatus\n---------------------------\n\nCalf thymus DNA (sodium salt, type I) was purchased from Sigma and *o*-nitrophenol was received from Shanghai Chemical Reagent Company, respectively. They were used without further purification. Acetate buffer solution of 0.10 M (pH = 4.5) was prepared using analytical grade reagents. Double-distilled water was used throughout. All experiments were performed at room temperature.\n\nStock solution of 5.0\u00d710^\u22123^ M *o*-nitrophenol was prepared in a 1:1 mixture of pH = 4.5 0.10 M acetate buffer and ethanol. The stock solution was diluted with 0.10 M acetate buffer according to the demand. The solution was saturated with N~2~ by bubbling high purity nitrogen gas into solution for 15 min before experiment. Nitrogen gas was maintained over the solution by continuing with a flow of the pure gas during the voltammetric experiment.\n\nAll voltammetric measurements were carried out using a CHI440A electrochemical analyzer (Shanghai Chenhua Apparatus Corporation, China). The experiments were performed using a 5 mL one-compartment electrochemical cell with a three electrodes system, which was consisted of a glassy carbon electrode or DNA-modified glassy carbon electrode as working electrode, a platinum coil auxiliary electrode and a saturated calomel electrode (SCE) as reference electrode.\n\n3.2 The process of preparing DNA-modified glassy carbon electrode\n-----------------------------------------------------------------\n\nThe DNA-modified glassy carbon electrode (\u03a6 = 3 mm) was prepared as follows: It was polished using a piece of 1500 diamond paper, and then polished to mirror smoothness with about 0.05 um alumina water slurry on silk. Afterward, it was washed with absolute alcohol and double-distilled water in an ultrasonic bath to remove the adsorbates on the electrode. After electrode was dried in air, 20 \u03bcL of DNA solution (2.0 mg/mL) was dropped onto the surface of the clean glassy carbon electrode and solvent was evaporated at 4 \u00b0C for 12 hours. The modified electrode was immersed the double-distilled water for 15 min to remove the unadsorbed DNA before using.\n\n4. Conclusions\n==============\n\nThis work showed that the electrochemical behavoiur of *o*-nitrophenol at a GCE was complex. The interaction of *o*-nitrophenol-DNA was studied by using the dsDNA film-modified GCE. The result suggested that *o*-nitrophenol caused DNA damage *in situ* after it was reduced. These results also contributed in understanding the possible mechanism of DNA damage.\n\nWe thank the National Natural Science Foundation of China (Grant No. 20775002) for financial support. This research is also supported by program for Innovative Research Team in Anhui Normal University.\n\n![Cyclic voltammograms of 5.0\u00d710^\u22125^ M o-nitrophenol in 0.10 M pH 4.5 acetate buffer at a bare GCE under N~2~, (\\-\\--) first cycle; (---) second cycle; (\u00b7\u00b7\u00b7) third cycle. Scan rate: 50 mV/s, potential rang from (**A**) \u2212 0.8 V to + 1.2 V, E~i~ = \u2212 0.30 V and (**B**) \u2212 0.8 V to + 0.4 V, E~i~ = \u2212 0.30 V.](IJMS-9-3-316-f1){#f1-ijms-9-3-316}\n\n![Repetitive differential pulse voltammograms of reduction of 5.0\u00d710^\u22125^ M o-nitrophenol in 0.10 M pH 4.5 acetate buffer at a bare GCE under N~2~: (a) the first scan, (b) the second scan, (c) the six scan and (d) the fifteenth scan](IJMS-9-3-316-f2){#f2-ijms-9-3-316}\n\n![Differential pulse voltammograms of 5.0\u00d710^\u22125^ M *o*-nitrophenol in 0.10 M pH 4.5 acetate buffer at a bare GCE under N~2~: (a) E~i~ = \u2212 0.15 V, (b) E~i~ = \u2212 0.6 V and (c) E~i~ = \u2212 0.60 V after E~d~ = \u2212 0.6 V for t~d~ = 2 minutes](IJMS-9-3-316-f3){#f3-ijms-9-3-316}\n\n![Cyclic voltammograms of 5.0\u00d710^\u22125^ M *o*-nitrophenol at GCE in 0.10 M pH 4.5 acetate buffer in absence (1) and presence of 28 ug/mL (2), 56 ug/mL (3), 84 ug/mL (4), 112 ug/mL (5) dsDNA. Scan rate: 50mV/s.](IJMS-9-3-316-f4){#f4-ijms-9-3-316}\n\n![Repetitive differential pulse voltammograms of reduction of 5.0\u00d710^\u22125^ M *o*-nitrophenol in 0.10 M pH 4.5 acetate buffer at the DNA-modified GCE under N~2~: (a) the first scan, (b) the second scan, (c) the six scan and (d) the fifteenth scan.](IJMS-9-3-316-f5){#f5-ijms-9-3-316}\n\n![Differential pulse voltammograms using DNA-modified GCE in 0.10 M pH 4.5 acetate buffer containing 1.0\u00d710^\u22125^ M *o*-nitrophenol before (a) and after reduction (b) by successive differential pulse voltammetry scans (10) from 0 to \u2212 0.8 V and (c) at constant potential E~d~ = \u2212 0.60 V for t~d~ = 2 minutes. (d) Differential pulse voltammogram obtained with the DNA-modified GCE in 0.10 M pH 4.5 acetate buffer.](IJMS-9-3-316-f6){#f6-ijms-9-3-316}\n\n![Reduvtive process of o-nitrophenol and feasible damage to DNA (R = phenol group). Adapted from the reference \\[[@b31-ijms-9-3-316]\\].](IJMS-9-3-316-f7){#f7-ijms-9-3-316}\n"}
+{"text": "Country Demographics {#sec1-1}\n====================\n\nIn 2012, the population was estimated to be 164 million with 47% under 15 years. There is no national population based cancer registry but using worldwide incidence rates of between 80 and 150/million children we would expect 6-9000 new cases/year\\[[@ref1]\\] in Bangladesh. Only about 25% of those numbers are actually currently diagnosed. Under-5 mortality is now 56/1000 live births due to considerable government focus on maternal and child health, and disease prevention. Poverty is a major issue but the national economy is showing positive growth. Children present late with cancer as a result of poor public and local health worker awareness of the meaning of signs and symptoms of cancer. Consequently, only about 80% of children reaching secondary/tertiary hospitals can be offered potentially curative therapy, and of those many families cannot afford to pay for full treatment.\n\nBangabandhu Sheikh Mujib Medical University {#sec2-1}\n-------------------------------------------\n\nThe largest specialist pediatric oncology center (created in the early 1990s) is at the multidisciplinary Bangabandhu Sheikh Mujib Medical University (BSMMU) in Dhaka funded directly by the Government. The Ministry of Health and Family Welfare send pediatricians there for specialist training, including 10 who have completed Doctorate level degrees since 2002. The Social welfare department provides support and help to patients/families. Nevertheless at present the cost of cytotoxic therapy falls to parents. This center has other essential pediatric specialties including nephrology, cardiology, nutrition and gastro-enterology, surgery, and general pediatrics. There are good basic diagnostic hematology and pathology services. Some limited immune histo-chemistry, using a limited number of antibodies has been initiated in the last 6 months. There is a flow cytometer in the unit, but it is not currently functioning. Radiology includes ultrasonography, computed tomography scanning and magnetic resonance Imaging.\n\nThe number of children with cancer presenting at BSMMU has been rising year on year (280 in 2010, 400 in 2011, 455 in 2012).\n\nThe relative incidence of malignancies seen at BSMMU in 2012 was: Acute lymphoblastic leukemia 58%, non-hodgkin lymphoma 11%, acute myeloblastic leukemia 10%, neuroblastoma 5%, Wilm\\'s tumor 2%, hepatoblastoma 3.5%, hodgkin lymphoma 3%, retinoblastoma 2%, germ cell tumors 2%, histocytosis 2%, central nervous system tumors 1%, osteosarcoma 1%. As part of a twinning project in June 2012, an online database (POND) [\\@www.POND4kids.org](@www.POND4kids.org) was established and can provide comprehensive, reliable incidence and outcome data for the unit. Most brain tumor patients referred to BSMMU are managed by neurosurgeons and radiotherapists (latter at Dhaka Medical College (DMC)) or the National Cancer Research institute (NCRI)).\n\nIn Bangladesh, there are now 6 trained pediatric hematologists/oncologists working as professor, associate professor, assistant professor and consultant level at BSMMU and 7 Government funded pediatric haemato-oncologists at other MC and at the NCRI. Of the Government funded colleges, 3 are in Dhaka (Dhaka MC, Siri Sollimulllah MC and Mymensingh) and 1 each in Sylhet, Chittagong, Rajshahi, Barrishal and Rangpure. A number of private facilities also provide some services. BSMMU and the DMC see more leukemia cases and the NCRI more solid tumors. NCRI is the national referral center for retinoblastoma. All those requiring radiotherapy go to Dhaka Medical College or the NCRI.\n\nTraining/Network {#sec1-2}\n================\n\nIn addition to the MD training program at BSMMU (currently 17 residents are enrolled at different phases of the MD course), the Unit is spearheading an extensive training program and network development of pediatric oncology services, in part funded by a 5 year World Child Cancer ([www.worldchildcancer.org](www.worldchildcancer.org)) twinning program linking BSMMU with University College London Hospitals and Vancouver Children\\'s Hospital.\n\nThis project is supporting the creation of regular training workshops, subsidies for drug costs and data management, installation of online tumor registries and family support. Dhaka MC and NCRI are also using the POND database, since 2012.\n\nASHIC Foundation created a shelter/home from home in Dhaka for patients/families from distant parts and Chittagong has a parent support group (CLASS). More family support will be developed as the network expands.\n\nChallenges {#sec1-3}\n==========\n\nFollowing awareness campaigns through the media and workshops, the number of children referred are increasing, creating capacity problems within referral unitsLate diagnosis and advanced disease at presentation persists (about 20% of cases are incurable at presentation)The early toxic death rate is about 10% with very ill children presenting43% of families refused treatment and/or stopped treatment prematurely in 2012 (cost, family disruption, doubts about curability and after initial good response)Most cytotoxics are imported -- The cost of all drugs, but especially L-asparaginase and an inconsistent supply of 6-Mercaptopurine are the major drug problems encountered.\n\nConclusion {#sec1-4}\n==========\n\nThose who complete treatment have a 50-60% cure rate. Concerted efforts are being made to raise public and professional awareness, reduce diagnostic delays and subsidise drug and travel costs.\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared.\n"}
+{"text": "This work was supported by the CAMS Innovation Fund for Medical Sciences (CIFMS) (Grant Nos. 2020\u2010I2M\u2010CoV19\u2010006, 2016\u2010I2M\u20103\u2010024, and 2017\u2010I2M\u20101\u2010009).\n\nA novel human coronavirus, identified as Severe Acute Respiratory Syndrome Coronavirus 2 (SARS\u2010CoV\u20102), has rapidly spread across the world, resulting in an international COVID\u201019 pandemic. The World Health Organization (WHO) has announced that the outbreak of the novel coronavirus constitutes a global public health emergency. The full spectrum of COVID\u201019 symptoms ranges from mild, self\u2010limiting respiratory tract illnesses to severe progressive pneumonia, multi\u2010organ failure, and death.[^1^](#trf15918-bib-0001){ref-type=\"ref\"} Unfortunately, no specific therapeutic agents or vaccines for COVID\u201019 are available, apart from Lopinavir\u2010Ritonavir, which may be a promising treatment.[^2^](#trf15918-bib-0002){ref-type=\"ref\"} Several other therapies such as remdesivir and favipiravir are under investigation, but the antiviral effects of these drugs are not fully known.[^3^](#trf15918-bib-0003){ref-type=\"ref\"}, [^4^](#trf15918-bib-0004){ref-type=\"ref\"}\n\nPassive immunization for the treatment of human infectious diseases can be traced back to the past century.[^5^](#trf15918-bib-0005){ref-type=\"ref\"} Convalescent plasma has been tried to fight pathogens including Severe Acute Respiratory Syndrome Coronavirus (SARS\u2010CoV),[^6^](#trf15918-bib-0006){ref-type=\"ref\"} Ebola virus,[^7^](#trf15918-bib-0007){ref-type=\"ref\"} Middle East Respiratory Syndrome Coronavirus (MERS\u2010CoV),[^8^](#trf15918-bib-0008){ref-type=\"ref\"} and avian influenza A (H5N1) virus.[^9^](#trf15918-bib-0009){ref-type=\"ref\"} Few studies, however, have been conducted on the selection of donors to ensure therapeutic potency in convalescent plasma.[^10^](#trf15918-bib-0010){ref-type=\"ref\"} In 2007, the WHO Blood Regulators Network (BRN) issued a guidance document on the selection of donors in a pandemic.[^11^](#trf15918-bib-0011){ref-type=\"ref\"} The recent COVID\u201019 outbreak worldwide has prompted the exploratory use of convalescent plasma in treating COVID\u201019; and case reports and case series have shown encouraging results.[^12^](#trf15918-bib-0012){ref-type=\"ref\"}, [^13^](#trf15918-bib-0013){ref-type=\"ref\"}\n\nSARS\u2010CoV\u20102 is an enveloped virus with four structural proteins: spike (S) protein, membrane (M) protein, enveloped (E) protein, and nucleocapsid (N) protein.[^14^](#trf15918-bib-0014){ref-type=\"ref\"} The S protein receptor\u2010binding domain (RBD) has been identified as a key target for therapeutic antibodies, as it plays essential roles in tropism and virus entry into host cells and can induce neutralizing antibodies and protective immunity.[^15^](#trf15918-bib-0015){ref-type=\"ref\"}, [^16^](#trf15918-bib-0016){ref-type=\"ref\"}, [^17^](#trf15918-bib-0017){ref-type=\"ref\"}, [^18^](#trf15918-bib-0018){ref-type=\"ref\"}\n\nTo collect high\u2010quality convalescent plasma, potential donors with appropriate characteristics should donate plasma at the proper time. The selection of convalescent plasma donors and the timing of their donations are important to ensure therapeutic potency. In this study, we collected demographic information and health history from COVID\u201019 convalescent plasma donors and investigated relevant viral serology, to identify the appropriate selection criteria for convalescent plasma donors.\n\nMETHODS {#trf15918-sec-0006}\n=======\n\nThis study was conducted at the Blood Center of Wuhan, China, from February 12, 2020, to March 2, 2020. All convalescent plasma donors provided written informed consent. The study was approved by the Ethics Review Committee of the Institute of Blood Transfusion, Chinese Academy of Medical Sciences.\n\nDonor recruitment and sample collection {#trf15918-sec-0007}\n---------------------------------------\n\nPatients who had recovered from COVID\u201019 were recruited as blood donors for this study. All patients aged 18\u201055 who had a confirmed COVID\u201019 diagnosis in Wuhan, China as of February 12, 2020 were identified by searching hospital records. Prospective study subjects were contacted by phone to assess interest and study eligibility. Study eligibility was determined using a COVID\u201019\u2010specific health history questionnaire based on the following criteria: 1) aged 18\u201055\u2009years; 2) eligible for blood donation; 3) diagnosed with COVID\u201019;[^19^](#trf15918-bib-0019){ref-type=\"ref\"} 4) had two subsequent negative COVID\u201019 nasopharyngeal swab tests based on PCR (at least 24\u2009hours apart); 5) had been discharged from the hospital for more than 2\u2009weeks; and 6) had no COVID\u201019 symptoms prior to convalescent plasma donation.\n\nDemographic information on the donors was collected, including sex, age, initial symptoms, duration of the disease, presence and duration of fever, and ABO blood type.\n\nDonor testing samples were collected on the day(s) of intended convalescent plasma donation, stored at 4\u00b0C, and processed within 4\u2009hours.\n\nRoutine donor testing {#trf15918-sec-0008}\n---------------------\n\nWe tested for infectious pathogens according to the Chinese national quality standards for blood establishment technical procedures (2019 edition).[^20^](#trf15918-bib-0020){ref-type=\"ref\"} We conducted ELISA testing for anti\u2010HIV1/2, anti\u2010HCV, syphilis antibody, and HBs Ag; we performed NAT testing for HIV, HBV, and HCV.\n\nTesting for SARS\u2010CoV\u20102 nucleic acid {#trf15918-sec-0009}\n-----------------------------------\n\nAll donor samples were tested for the SARS\u2010CoV\u20102 RNA by quantitative RT\u2010PCR using the PerkinElmer New Coronavirus Nucleic Acid Detection Kit (PerkinElmer Healthcare Diagnostics Co., Ltd.). The assays were performed according to the manufacturer\\'s instructions.\n\nDetection of N\u2010specific IgM and IgG antibodies {#trf15918-sec-0010}\n----------------------------------------------\n\nN\u2010specific IgM and IgG antibodies were tested for using the enzyme\u2010linked immunosorbent assay (ELISA) (Livzon Pharmaceutical Group) according to the manufacturers' instructions.\n\nDetection of S\u2010RBD\u2010specific IgG antibody {#trf15918-sec-0011}\n----------------------------------------\n\nThe S\u2010RBD\u2010specific IgG ELISA assay was developed in\u2010house. In brief, we coated 96 well plates (Thermo Scientific) with 100\u2009ng of recombinant RBD polypeptides (Sino Biological) per well. After overnight coating, the coating solution was decanted from the plates and non\u2010specific activity was blocked with PBS containing 1% skim milk at 37\u00b0C for an hour. Plasma samples were diluted 160, 320, 640, and 1280 folds with 0.5% Triton X\u2010100 phosphate\u2010buffered saline and 5% fetal calf serum (Gibco) and added to the plates. After washing, mouse\u2010to\u2010human secondary antibodies were added to the plates; we observed the plates for the horseradish peroxidase reaction. The OD values were calculated by measuring the change in the absorbance at 450\u2009nm and 630\u2009nm using an automatic microplate reader (Sunrise, Tecan GmbH). Results were reported as the S/CO value, calculated as the ratio of the OD value to the cutoff value. Titers were reported as the highest dilution when the ELISA assay was still positive and ranged between 1:160 and 1:1280 (ELISA endpoint dilution titers).\n\nCorrelation of the S\u2010RBD\u2010specific IgG ELISA and SARS\u2010CoV\u20102 viral neutralization assay {#trf15918-sec-0012}\n-------------------------------------------------------------------------------------\n\nWe isolated SARS\u2010CoV\u20102 virus from a COVID\u201019 patient. The virus was cultured on Vero cells (American Tissue Culture Collection \\[ATCC\\], CCL\u201081) and the viral neutralization activity was determined based on TCID50 (Median Tissue Culture Infectious Dose).[^21^](#trf15918-bib-0021){ref-type=\"ref\"} In brief, we incubated study subjects\\' sera at 56\u00b0C for 30\u2009minutes to deactivate the complements and then diluted serially from 1:10, 1:20, 1:40, 1:80, to 1:160. The diluted serum was mixed with equal volumes of SARS\u2010CoV\u20102 virus at a dose of 100 TCID50, then incubated at 37\u00b0C for one hour. The mixture was added in quadruplicates to Vero cells cultured in 96\u2010well microtiter plates, then removed after 1 hour and replaced with 200\u2009mL fresh growth medium. Serum samples from healthy volunteers were used as the control. The cytopathic effect was observed 5\u2009days after incubation. The neutralizing antibody titers were calculated using the Reed\u2010Muench method.[^21^](#trf15918-bib-0021){ref-type=\"ref\"}\n\nStatistical analysis {#trf15918-sec-0013}\n--------------------\n\nStatistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software Inc.). Pearson\\'s r or Spearman rank\u2010order correlation was used for correlations. Nonlinear regression (curve fit) analysis was performed on the changes of antibody levels over time. One\u2010way analysis of variance followed by the Newman--Keuls test was used for comparison of all pairs. A p value of less than 0.05 was considered statistically significant.\n\nRESULTS {#trf15918-sec-0014}\n=======\n\nDonor characteristics {#trf15918-sec-0015}\n---------------------\n\nIn this study, 49 convalescent plasma donors were recruited successfully, including 33 males and 16 females. The median age was 37\u2009years (IQR, 25\u201054\u2009years). The most common self\u2010reported symptoms at the onset of illness were fever (39, 80%) and cough (30, 61%). Other symptoms included shortness of breath, muscle aches, and diarrhea. The majority (43, 88%) of patients presented with more than one symptom, 27 (55%) had a fever with cough, 16 (33%) had a fever with muscle aches, and 10 (20%) had a fever and shortness of breath (Table [1](#trf15918-tbl-0001){ref-type=\"table\"}). The average time from the onset of symptoms to plasma donation was 37.8\u2009days (IQR, 23\u201058\u2009days). All plasma samples were negative for SARS\u2010CoV\u20102 nucleic acid and the transfusion transmitted pathogens tested.\n\n###### \n\nDemographics, baseline characteristics and clinical symptoms of 49 donors\n\n                                                                Donors (n = 49)\n  ------------------------------------------------------------- -----------------\n  Age, years                                                    \n  Mean\u2009\u00b1\u2009SD                                                     37\u2009\u00b1\u20097.7\n  Range                                                         25\u201054\n  \u226429                                                           6 (12%)\n  30\u201039                                                         27 (55%)\n  40\u201049                                                         11 (23%)\n  \u226550                                                           5 (10%)\n  Sex                                                           \n  Male                                                          33 (67%)\n  Female                                                        16 (33%)\n  Severity of disease[\\*](#trf15918-note-0003){ref-type=\"fn\"}   \n  Mild                                                          24 (49%)\n  Moderate                                                      25 (51%)\n  Severe                                                        0 (0%)\n  Signs and symptoms                                            \n  Fever                                                         39 (80%)\n  Cough                                                         30 (61%)\n  Shortness of breath                                           10 (20%)\n  Muscle ache                                                   19 (39%)\n  Diarrhea                                                      13 (27%)\n  More than one sign or symptom                                 43 (88%)\n  Fever and cough                                               27 (49%)\n  Fever and shortness of breath                                 10 (20%)\n  Fever and muscle ache                                         16 (33%)\n  Time of disease onset to donation, days                       \n  Mean\u2009\u00b1\u2009SD                                                     37.8\u2009\u00b1\u20098.3\n  Range                                                         23\u201058\n  \\<28                                                          5 (10%)\n  28\u201034                                                         14 (28%)\n  35\u201041                                                         15 (31%)\n  \u226542                                                           15 (31%)\n\nThe definition of disease classification was based on the Chinese diagnosis and treatment guidelines for the new coronavirus pneumonia (trial version 5).[^19^](#trf15918-bib-0019){ref-type=\"ref\"}\n\nThe trend of changes over recovery time {#trf15918-sec-0016}\n---------------------------------------\n\nChanges in N\u2010specific IgM levels showed a decreasing trend and had a negative correlation with time from the onset of symptoms to the plasma donation (r = \u22120.3591, p = 0.0056) (Fig. [1A](#trf15918-fig-0001){ref-type=\"fig\"}). Changes in N\u2010specific IgG\u2010levels showed an upward trend and had a positive correlation with time from the onset of symptoms to the plasma donation (r = 0.2635, p = 0.0352) (Fig. [1B](#trf15918-fig-0001){ref-type=\"fig\"}). S\u2010RBD\u2010specific IgG showed similar results (r = 0.4540, p = 0.0011) (Fig. [1C](#trf15918-fig-0001){ref-type=\"fig\"}). The heat map of these antibodies over time also showed that the levels of S\u2010RBD\u2010specific IgG antibody in donors gradually increased, while N\u2010specific IgM antibody gradually decreased and remained at a lower level (Fig. [1E](#trf15918-fig-0001){ref-type=\"fig\"}). To further confirm the trend of S\u2010RBD\u2010specific IgG antibody changes over time, we analyzed antibody levels according to different recovery times. The levels of S\u2010RBD\u2010specific IgG antibodies were lower in donors who donated within 28\u2009days from the onset of symptoms than in those who donated later than 28\u2009days from the onset of symptoms. The difference was statistically significant, with a p value of 0.012 (Fig. [1D](#trf15918-fig-0001){ref-type=\"fig\"}).\n\n![Nonlinear regression (curve fit) of SARS\u2010CoV\u20102 antibodies levels. (A) The correlation between the time from disease onset to plasma donation and N\u2010specific IgM antibody level. There is an inverse correlation (r = \u22120.3591, p = 0.0056). (B) The correlation between the time from disease onset to plasma donation and N\u2010specific IgG antibody level. There is a positive correlation (r = 0.2635, p = 0.0352). (C) The correlation between the time from disease onset to plasma donation and S\u2010RBD\u2010specific IgG antibody level. There is a positive correlation (r = 0.4540, p = 0.0011). (D) The S\u2010RBD\u2010specific IgG antibody level at less than 28\u2009days, 28\u201034\u2009days, 35\u201041\u2009days, and more than 42\u2009days after disease onset. (E) The level of N\u2010specific IgM, IgG, and S\u2010RBD\u2010specific IgG antibodies during recovery time. The values are presented as the mean\u2009\u00b1\u2009SD. \\*p\u2009\\<\u20090.05, \\*\\*p\u2009\\<\u20090.01; NS, nonsignificant. \\[Color figure can be viewed at [wileyonlinelibrary.com](http://wileyonlinelibrary.com)\\]](TRF-9999-na-g001){#trf15918-fig-0001}\n\nCorrelation of S\u2010RBD\u2010specific IgG antibody and SARS\u2010CoV\u20102 viral neutralization assay {#trf15918-sec-0017}\n------------------------------------------------------------------------------------\n\nWe performed a correlation study between the SARS\u2010CoV\u20102 viral neutralization titer and the S\u2010RBD\u2010specific IgG antibody titer. The results indicated a positive correlation (r = 0.6222, p = 0.0308) between the two assays. A serum neutralization titer of 1:80 is approximately equivalent to a titer of 1:1280 for S\u2010RBD\u2010specific IgG.\n\nCorrelation of donor characteristics and S\u2010RBD\u2010specific IgG antibody {#trf15918-sec-0018}\n--------------------------------------------------------------------\n\nA recent study has demonstrated that SARS\u2010CoV\u20102 viral neutralization activity is correlated with the S\u2010RBD\u2010specific IgG antibody.[^22^](#trf15918-bib-0022){ref-type=\"ref\"} We assessed the correlation of the level of S\u2010RBD\u2010specific IgG antibody to donor factors including age, sex, initial symptoms, body temperature and duration of fever, and ABO blood type. The results showed that donors with high fevers exceeding 38.5\u00b0C or with fever lasting longer than 3\u2009days had a high level of S\u2010RBD\u2010specific IgG (Fig. [2A, B](#trf15918-fig-0002){ref-type=\"fig\"}). Notable, changes in antibody levels were not correlated to age, sex, or blood type (Note: we collected only one case of type AB) (Fig. [2C, D, and E](#trf15918-fig-0002){ref-type=\"fig\"}).\n\n![Correlation of donor characteristics and S\u2010RBD\u2010specific IgG antibody. (A) The S\u2010RBD\u2010specific IgG antibody level in non\u2010febrile (\\<37.5\u00b0C), mild\u2010febrile (37.5\u201038.4\u00b0C), moderate\u2010febrile (38.5\u201039.4\u00b0C), and extreme\u2010febrile (\u226539.5\u00b0C) donors. (B) The S\u2010RBD\u2010specific IgG antibody level in donors with fevers lasting \\<3\u2009days, 3\u20107\u2009days, and\u2009\u22658\u2009days. (C) The S\u2010RBD\u2010specific IgG antibody level of donors \\<30\u2009years old, 30\u201039\u2009years old, 40\u201049\u2009years old, and\u2009\u226550\u2009years old. (D) The S\u2010RBD\u2010specific IgG antibody level in male and female donors. (E) The S\u2010RBD\u2010specific IgG antibody level of A, B, O, and AB blood type donors. The differences between the groups were analyzed using the Newman--Keuls test and values are presented as the mean\u2009\u00b1\u2009SD. \\*p\u2009\\<\u20090.05, \\*\\*p\u2009\\<\u20090.01; NS, nonsignificant.](TRF-9999-na-g002){#trf15918-fig-0002}\n\nOptimal selection of COVID\u201019 convalescent plasma donors {#trf15918-sec-0019}\n--------------------------------------------------------\n\nTo obtain optimal convalescent plasma donor selection criteria, we analyzed the titer distribution of the S\u2010RBD\u2010specific IgG antibody in convalescent plasma donors under different selection conditions. Among the 49 convalescent plasma donors, 90% had a titer of \u22651:160 and 78% had a titer of \u22651:640. Based on the selection criterion of donating plasma 4\u2009weeks after the onset of symptoms, there were 42 convalescent plasma donors, of whom 90% had a titer of \u22651:160 and 84% had a titer of \u22651:640. Considering the additional selection criterion of a fever lasting longer than 3\u2009days or a body temperature exceeding 38.5\u00b0C, there were 30 convalescent plasma donors, of whom 100% had a titer of \u22651:160 and 93% had a titer of \u22651:640 (Fig. [3](#trf15918-fig-0003){ref-type=\"fig\"}). Of these, 29 donors had fever lasting longer than 3\u2009days, and 27 donors had body temperatures exceeding 38.5\u00b0C during the fever (Fig. [3](#trf15918-fig-0003){ref-type=\"fig\"}). Figure [4](#trf15918-fig-0004){ref-type=\"fig\"} shows the antibody titer distribution before and after the selection.\n\n![S\u2010RBD\u2010specific IgG antibody titers based on different selection criteria. The green line represents a total of 49 donors. The red line represents the distribution of antibody titer in donors with more than 4\u2009weeks after symptom onset. The gray line represents the titer in donors with a fever lasting longer than 3\u2009days at disease presentation and more than 4\u2009weeks after symptom onset. The yellow line represents antibody titer in donors with body temperature exceeding 38.5\u00b0C at disease presentation and more than 4\u2009weeks after symptom onset. The blue line represents antibody titer in donors with a fever lasting longer than 3\u2009days or with body temperature exceeding 38.5\u00b0C at disease presentation and more than 4\u2009weeks after symptom onset. \\[Color figure can be viewed at [wileyonlinelibrary.com](http://wileyonlinelibrary.com)\\]](TRF-9999-na-g003){#trf15918-fig-0003}\n\n![The proportion of S\u2010RBD\u2010specific IgG antibody titers. (A) The proportion of S\u2010RBD\u2010specific IgG antibody titers in 49 donors. (B) The proportion of S\u2010RBD\u2010specific IgG antibody titers in 30 donors with a fever lasting longer than 3\u2009days or body temperatures exceeding 38.5\u00b0C at disease presentation and with more than 4\u2009weeks after symptoms. \\[Color figure can be viewed at [wileyonlinelibrary.com](http://wileyonlinelibrary.com)\\]](TRF-9999-na-g004){#trf15918-fig-0004}\n\nDISCUSSION {#trf15918-sec-0020}\n==========\n\nThe COVID\u201019 pandemic has become a major public health challenge around the world. The use of convalescent plasma may be an effective and safe treatment option to help control the COVID\u201019 pandemic.[^12^](#trf15918-bib-0012){ref-type=\"ref\"}, [^23^](#trf15918-bib-0023){ref-type=\"ref\"} There are now thousands of recovered COVID\u201019 patients who could donate convalescent plasma. Establishing optimal selection criteria for COVID\u201019 convalescent plasma donors is necessary to ensure convalescent plasma therapeutic potency and to optimize collection efficiency.\n\nHere we have proposed a donor selection strategy for COVID\u201019 convalescent plasma collection. A total of 49 convalescent plasma donors were analyzed, all of whom had mild or moderate COVID\u201019 symptoms. At least 14\u2009days after symptom resolution, we collected the convalescent plasma and analyzed SARS\u2010COV\u20102 antibodies.\n\nThis study compares the changes of N\u2010specific IgM, IgG, and S\u2010RBD\u2010specific IgG antibody levels over time. Our data show that N\u2010specific IgM antibody levels continued to decline after 3\u2009weeks of SARS\u2010COV\u20102 infection and reached low levels after 6\u2009weeks. At the same time, S\u2010RBD\u2010specific and N\u2010specific IgG antibodies exhibited an upward trend and continued to rise after 4\u2009weeks from the onset of symptoms, which was consistent with the study by Ling et al.[^24^](#trf15918-bib-0024){ref-type=\"ref\"} Interestingly, the trends in increase showed by S\u2010RBD\u2010specific and N\u2010specific IgG antibodies were not quite the same; the former were more dramatic while the latter were milder. A recent study by Walls et al found that SARS\u2010CoV S\u2010murine polyclonal antibodies potently inhibit SARS\u2010CoV S mediated entry into cells and induce the production of protective neutralizing antibodies targeting S epitopes.[^25^](#trf15918-bib-0025){ref-type=\"ref\"} The RBD of Middle East respiratory syndrome coronavirus (MERS\u2010CoV), as a vaccine antigen, has been shown to induce neutralizing antibodies and protect mice transduced with a viral vector.[^26^](#trf15918-bib-0026){ref-type=\"ref\"} Thus, we predict that S\u2010RBD\u2010specific IgG may be a protective antibody and that its level would significantly increase in recovered patients. Cao et al. studied SARS\u2010CoV neutralizing antibody titers in 56 patients who recovered from SARS\u2010CoV infections.[^27^](#trf15918-bib-0027){ref-type=\"ref\"} Their findings show that SARS\u2010CoV IgG and neutralizing antibodies peaked at 4\u2009months and then declined, reaching undetectable levels in 25.6% (IgG) and 16.1% (neutralizing antibodies) of study subjects at 36\u2009months. Our observed positive correlation between S\u2010RBD\u2010specific IgG titer and viral neutralization titer is assuring. Ideally, COVID\u201019 convalescent plasma products should be measured with viral neutralization antibody titer to ensure product potency. However, viral neutralization assay is time\u2010consuming to perform and requires the use of Level 3 bio\u2010safety cabinets, thus limiting its use in routine practice as a quality measurement for convalescent plasma. Using S\u2010RBD\u2010specific IgG titer as a surrogate marker for viral neutralization titer is far more practical and brings convalescent plasma product quality control for potency a step closer to reality.\n\nMoreover, we observed a correlation between donor characteristics and S\u2010RBD\u2010specific IgG antibody. We found that donors who experienced disease symptoms including a body temperature exceeding 38.5\u00b0C or a fever lasting longer than 3\u2009days had higher levels of S\u2010RBD\u2010specific IgG antibody titers in convalescent plasma. Although a recent study indicates that people with blood group A have a significantly higher risk of acquiring COVID\u201019, whereas blood group O has a significantly lower risk for the infection,[^28^](#trf15918-bib-0028){ref-type=\"ref\"} our study did not show a significant correlation between S\u2010RBD\u2010specific IgG antibody and age, sex, or donor blood type.\n\nBased on the results of this study, optimal COVID\u201019 convalescent plasma donors with high S\u2010RBD\u2010specific IgG antibody levels (and most likely higher virus neutralization activity) can be selected based on the following criteria: disease symptoms comprising a fever lasting longer than 3\u2009days or a body temperature exceeding 38.5\u00b0C. Providing additional selection criterion of donating plasma 4\u2009weeks after the onset of symptoms could help achieve even higher S\u2010RBD\u2010specific IgG antibody levels.\n\nKey elements to ensure an adequate supply of high\u2010quality COVID\u201019 convalescent plasma products include the following: donor recruitment and selection, convalescent plasma collection, and product evaluation and quality control for the presence of adequate protective antibodies and the absence of transfusion transmitted pathogens. More studies, guidance, and regulations are needed to refine operation and product requirements for COVID\u201019 convalescent plasma. This study also suggests that the antibody titers from asymptomatic individuals who are incidentally found to be positive for IgG antibodies to SARS\u2010CoV\u20102 may be lower than those from symptomatic individuals. Future studies should evaluate this possibility.\n\nLimitations {#trf15918-sec-0021}\n-----------\n\nThis study has several limitations. First, the sample size was small, so our selection criteria should be verified by larger studies. Second, the study design precludes a definitive conclusion regarding the potential therapeutic efficacy and requires clinical validation through clinical trials.\n\nCONFLICT OF INTEREST {#trf15918-sec-0023}\n====================\n\nThe authors have disclosed no conflicts of interest.\n\nAUTHOR CONTRIBUTIONS {#trf15918-sec-0024}\n====================\n\nLing Li, Hanwei Chen, Qilu Lu, Ru Yang, Lei Zhao, Jue Wang, Haixia Xu, Chuanqiao Liu, Guoan Chen, Sitian Chen, Chenyue Li, and Jiajia Qiao were responsible for donor recruitment, plasma collection and performing all assays; Xunliang Tong and Rui He were responsible for data analysis and manuscript\u2010writing; Juntao Yang, Yanyun Wu, and Zhong Liu contributed to study design, research analysis, and manuscript revision.\n\nWe would like to thank the following individuals for their guidance, expertise, and assistance with the study design: Peter W. Marks, MD PhD, Center for Biologics Evaluation and Research, FDA; Anne Eder, MD, PhD, Center for Biologics Evaluation and Research, FDA; Nicole Verdun, MD, Center for Biologics Evaluation and Research, FDA; Tim Uyeki, MD, MPH, MPP, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention (CDC), USA; Scott Koepsell, MD, PhD, University of Nebraska Medical Center; Annie Winkler, MD, Instrumentation Laboratory; Xuan Qin, PhD, D(ABMM), Seattle Children\\'s Hospital and University of Washington; Toby L. Simon, MD, CSL Behring; Richard J Benjamin, MD, PhD, CERUS Corporation; Jerry A Holmberg, Grifols S A; Vicente Blanquer, Grifols S A; Daniel Fleta, Grifols S A; Amarant Martinez, Grifols S A; and Liu Yang, Grifols S A.\n\n[^1]: Co--first authors.\n"}
+{"text": "Introduction\n============\n\nOver the past 20 years or so, there has been considerable and growing interest in crystallization-driven block copolymer self-assembly (CDSA). Manners and co-workers reported the first example of CDSA, which involved the formation of cylindrical micelles by polydimethylsiloxane--poly(ferrocenyldimethylsilane) (PDMS--PFDMS) diblock copolymers in *n*-hexane.[@cit1] It was shown that the crystallization of the core-forming PFDMS block is solely responsible for the self-assembly behavior.[@cit2] Since this seminal work, the same group have utilized both ferrocenyl[@cit3]--[@cit10] and thiophene-based[@cit11]--[@cit15] diblock copolymers to generate rod-like cylindrical micelles with remarkably well-defined dimensions. In related work, Dove and co-workers have explored the CDSA of block copolymers comprising crystalline poly([l]{.smallcaps}-lactide) (PLLA) cores.[@cit16]--[@cit20] More recently, judicious solvent selection enabled the formation of relatively uniform micrometer-sized diamond-shaped lamellae from poly(*N*,*N*-dimethylacrylamide)--poly([l]{.smallcaps}-lactide) (PDMAC--PLLA) diblock copolymers.[@cit21] Furthermore, PDMAC--PLLA--PDMAC triblock copolymers formed either diamond-shaped lamellae or cylindrical micelles of varying length in dilute methanolic solution.[@cit22] Similarly, Lecommandoux and co-workers[@cit23] reported the 1D fusion of spherical diblock copolymer micelles to form fibres on prolonged heating at 65 \u00b0C in water, with wide-angle X-ray scattering (WAXS) studies indicating the gradual formation of crystalline cores. There have also been several reports of CDSA formulations utilizing poly(\u03b5-caprolactone) core-forming blocks.[@cit24]--[@cit26]\n\nPrior to the development of CDSA, Richter and co-workers[@cit27] demonstrated that poly(ethylene-propylene)--polyethylene diblock copolymers in *n*-decane form crystalline lamellar structures on cooling from 70 \u00b0C to ambient temperature. Later, Xu *et al.*[@cit28],[@cit29] utilized small-angle X-ray scattering (SAXS) to demonstrate that diblock copolymer micelles with crystalline cores can form higher order lamellar structures. For example, cooling solutions of poly(ethylene oxide)--poly(butylene oxide) diblock copolymers in *n*-hexane caused spherical micelles to become highly anisotropic owing to crystallization of the core-forming poly(ethylene oxide) chains.[@cit28] For diblock copolymers with relatively short corona blocks, unfavorable interactions between the crystalline micelle cores and the solvent led to particle aggregation, with pastes being formed even at relatively low copolymer concentrations (\\<1.0% w/v). In contrast, longer corona blocks prevented aggregation and such dispersions remained free-flowing at copolymer concentrations of up to 10% w/v. Similarly, Schurtenberger and co-workers[@cit30] observed a reversible spherical micelle-to-lamellar transition for poly(ethylene oxide)--polybutadiene nanoparticles in ethanol on cooling from 60 \u00b0C to 20 \u00b0C; concomitant WAXS studies confirmed crystallization of poly(ethylene oxide) stabilizer (or corona) blocks.\n\nRecently, many research groups have shown that polymerization-induced self-assembly (PISA)[@cit31]--[@cit35] is a versatile and efficient method for preparing a wide range of functional diblock copolymer nano-objects (typically spheres, worms or vesicles) in water,[@cit36]--[@cit40] lower alcohols[@cit41]--[@cit45] or non-polar solvents,[@cit46]--[@cit52] as well as various solvent mixtures.[@cit53]--[@cit61] Such PISA syntheses provide convenient access to new red blood cell cryopreservation protocols,[@cit62] sterilizable hydrogels for 3D cell growth,[@cit63] novel stem cell storage media,[@cit64] model Pickering emulsifiers,[@cit65]--[@cit70] controlled release of encapsulated payloads[@cit71] and hitherto unknown high-temperature oil-thickening mechanisms.[@cit72] Of particular relevance to the present work, reversible addition--fragmentation chain transfer (RAFT) dispersion polymerization enables the PISA synthesis of well-defined spherical nanoparticles of tunable diameter.[@cit73]--[@cit77] This is typically achieved by systematically varying the target degree of polymerization (DP) of the solvophobic block while utilizing a sufficiently long solvophilic macromolecular chain transfer agent (macro-CTA).[@cit41],[@cit75],[@cit76]\n\nIn the present study, a series of colloidally-stable poly(behenyl methacrylate)--poly(benzyl methacrylate) (PBeMA--PBzMA) spherical nanoparticles are synthesized *via* RAFT dispersion polymerization of benzyl methacrylate in mineral oil at 90 \u00b0C (see [Scheme 1](#sch1){ref-type=\"fig\"}). This industrially-sourced solvent was utilized to highlight the versatility and relevance of PISA formulations. We show that the PBeMA stabilizer chains exhibit crystallization-driven aggregation on cooling to 20 \u00b0C, which leads to the formation of a macroscopic paste. This colloidal (in)-stability is demonstrated to be thermoreversible and has been characterized using turbidimetry, differential scanning calorimetry (DSC), SAXS, WAXS and rheology.\n\n![Synthesis of a near-monodisperse poly(behenyl methacrylate) (PBeMA) macro-CTA *via* RAFT solution polymerization of behenyl methacrylate (BeMA) in toluene using 4-cyano-4-(2-phenylethane sulfanylthiocarbonyl)sulfanylpentanoic acid (PETTC) at 70 \u00b0C, followed by the RAFT dispersion polymerization of benzyl methacrylate (BzMA) in mineral oil at 90 \u00b0C.](c8sc00762d-s1){#sch1}\n\nExperimental\n============\n\nMaterials\n---------\n\nA 4 cSt American Petroleum Institute (API) group III mineral oil and behenyl methacrylate (BeMA, \\>99%) were kindly provided by The Lubrizol Corporation Ltd (Hazelwood, Derbyshire, UK). 4-Cyano-4-(2-phenylethane sulfanylthiocarbonyl)sulfanylpentanoic acid (PETTC, \\>99%) was synthesized according to the literature.[@cit69]*Tert*-butyl peroxy-2-ethylhexanoate (T21s, \\>97%) initiator was purchased from AkzoNobel (The Netherlands), THF and toluene were purchased from Fisher Scientific (UK), CDCl~3~ was purchased from VWR International (UK), CD~2~Cl~2~ was purchased from Goss Scientific (UK) and all other materials were purchased from Sigma-Aldrich (UK) and were used as received.\n\nSynthesis of poly(behenyl methacrylate) (PBeMA) macromolecular chain transfer agent (macro-CTA) *via* RAFT solution polymerization\n----------------------------------------------------------------------------------------------------------------------------------\n\nA PBeMA~37~ macro-CTA was synthesized as follows: a 250 mL round-bottomed flask was charged with behenyl methacrylate (BeMA; 23.8 g, 60.3 mmol), PETTC (410 mg, 1.21 mmol; target PBeMA DP = 50), 2,2\u2032-azobisisobutyronitrile (AIBN, 98%; 39.6 mg, 241 \u03bcmol; \\[PETTC\\]/\\[AIBN\\] molar ratio = 5.0) and toluene (24.2 g). The reaction solution was purged with nitrogen and placed in a pre-heated oil bath at 70 \u00b0C for 3 h. The resulting PBeMA (BeMA conversion = 57%; *M*~n~ = 12\u2009400 g mol^--1^, *M*~w~/*M*~n~ = 1.18) was purified by twice precipitating into excess 2-propanol. The mean degree of polymerization (DP) of this macro-CTA was calculated to be 37 (CTA efficiency = \\[(target DP \u00d7 monomer conversion)/(actual DP) = 77%\\]) using ^1^H NMR spectroscopy by comparing the integrated signals corresponding to the five aromatic protons at 7.0--7.5 ppm with that assigned to the two oxymethylene protons of PBeMA at 3.4--4.2 ppm (see Fig. S1a[\u2020](#fn1){ref-type=\"fn\"}).\n\nSynthesis of poly(behenyl methacrylate)--poly(benzyl methacrylate) (PBeMA--PBzMA) diblock copolymer nanoparticles *via* RAFT dispersion polymerization\n------------------------------------------------------------------------------------------------------------------------------------------------------\n\nA typical RAFT dispersion polymerization synthesis of targeted PBeMA~37~--PBzMA~100~ diblock copolymer nanoparticles at 20% w/w solids was conducted as follows: benzyl methacrylate (BzMA; 0.27 g, 1.54 mmol), T21s initiator (666 \u03bcg; 3.08 \u03bcmol; dissolved at 10% v/v in mineral oil) and PBeMA~37~ macro-CTA (0.23 g; 15.4 \u03bcmol; \\[macro-CTA\\]/\\[initiator\\] molar ratio = 5.0) were co-dissolved in mineral oil (2.00 g). The reaction mixture was purged with nitrogen for 30 min and the deoxygenated solution was then placed in a pre-heated oil bath at 90 \u00b0C for 5 h (final BzMA conversion = 99%; *M*~n~ = 22\u2009700 g mol^--1^; *M*~w~/*M*~n~ = 1.15).\n\nGel permeation chromatography (GPC)\n-----------------------------------\n\nMolecular weight distributions were assessed by GPC using THF eluent. The THF GPC set-up comprised two 5 \u03bcm (30 cm) mixed C columns and a WellChrom K-2301 refractive index detector operating at a wavelength of 950 \u00b1 30 nm. The mobile phase contained 2.0% v/v triethylamine and 0.05% w/v butylhydroxytoluene, and the flow rate was 1.0 mL min^--1^. A series of ten near-monodisperse poly(methyl methacrylate) standards (*M*~p~ values ranging from 645 to 2\u2009480\u2009000 g mol^--1^) were used for calibration.\n\n^1^H nuclear magnetic resonance (NMR) spectroscopy\n--------------------------------------------------\n\nHomopolymer and copolymer spectra were recorded in either CD~2~Cl~2~ or CDCl~3~ using a Bruker AV1-400 MHz spectrometer. Typically 64 scans were averaged per spectrum.\n\nDynamic light scattering (DLS)\n------------------------------\n\nDLS studies were performed using a Zetasizer NanoZS instrument (Malvern Instruments, UK) at a fixed scattering angle of 173\u00b0. Copolymer dispersions were diluted to 0.10% w/w using *n*-dodecane prior to light scattering studies. The intensity-average diameter and polydispersity index \\[PDI = (standard deviation/intensity-average diameter)^2^\\] of the diblock copolymer nanoparticles were calculated by cumulant analysis of the experimental correlation function using Dispersion Technology Software version 6.20, accounting for the temperature-dependent viscosity of the *n*-dodecane. Data were averaged over thirty runs each of thirty seconds duration.\n\nTransmission electron microscopy (TEM)\n--------------------------------------\n\nTEM studies were conducted using a Philips CM 100 instrument operating at 100 kV and equipped with a Gatan 1 k CCD camera. Diluted diblock copolymer dispersions (0.10% w/w) were placed as droplets on carbon-coated copper grids, allowed to dry and then exposed to ruthenium([viii]{.smallcaps}) oxide vapor for 7 min at 20 \u00b0C prior to analysis. This heavy metal compound acted as a positive stain for the core-forming PBzMA block in order to improve electron contrast. The ruthenium([viii]{.smallcaps}) oxide was prepared as follows. Ruthenium([iv]{.smallcaps}) oxide (0.30 g) was added to water (50 g) to form a black slurry; addition of sodium periodate (2.0 g) with stirring produced a yellow solution of ruthenium([viii]{.smallcaps}) oxide within 1 min.[@cit78]\n\nTurbidimetry\n------------\n\nTurbidimetry measurements were performed on an unstirred 1.0% w/w solution (or dispersion) using a Shimadzu UV-1800 spectrophotometer equipped with a twin Peltier temperature controller (DBS Analytical instruments, Italy). Data were recorded at 650 nm during cooling (50 \u00b0C to 20 \u00b0C) and heating (20 \u00b0C to 50 \u00b0C) steps, allowing 5 min for thermal equilibration at each temperature. Data were averaged over three measurements.\n\nDifferential scanning calorimetry (DSC)\n---------------------------------------\n\nDSC studies were performed using a TA Instruments Discovery DSC instrument equipped with TZero low-mass aluminium pans. Samples were equilibrated at 70 \u00b0C for 5 min before two consecutive thermal cycles (70 \u00b0C -- 10 \u00b0C -- 70 \u00b0C) were performed at a cooling/heating rate of 2.0 \u00b0C min^--1^.\n\nOscillatory rheology\n--------------------\n\nThe measurements were performed using an Anton Paar MCR502 rheometer equipped with a Peltier temperature controller, cone-and-plate geometry (a truncated 50 mm 2\u00b0 stainless steel cone) and TruGap functionality for online monitoring of the geometry gap. The storage (*G*\u2032) and loss (*G*\u2032\u2032) moduli were measured as a function of temperature at a fixed strain amplitude of 1.0% and an angular frequency of 10 rad s^--1^. Thermal cycles between 20 \u00b0C and 50 \u00b0C were performed at 1.0 \u00b0C intervals with an equilibration time of 5 min being allowed prior to each measurement.\n\nSmall-angle and wide-angle X-ray scattering\n-------------------------------------------\n\nSAXS and WAXS data were collected simultaneously using a laboratory SAXS/WAXS instrument (Xeuss 2.0, Xenocs, France) equipped with a liquid gallium MetalJet X-ray source (Excillum, Sweden, wavelength *\u03bb* = 0.134 nm), two sets of motorized scatterless slits for beam collimation, a Dectris Pilatus 1M pixel SAXS detector (sample-to-detector distance 1.889 m) and a Dectris Pilatus 100K pixel WAXS detector (sample-to-detector distance 0.178 m, tilted 36\u00b0 relative to the incident X-ray beam). SAXS and WAXS patterns were recorded over a *q* range of 0.06 nm^--1^ \\< *q* \\< 4.0 nm^--1^ and 10.5 nm^--1^ \\< *q* \\< 25 nm^--1^ (12.9\u00b0 \\< 2*\u03b8* \\< 31.0\u00b0), respectively, where *q* = (4\u03c0sin\u2009*\u03b8*)/*\u03bb* is the length of the scattering vector and *\u03b8* is one-half of the scattering angle. Glass capillaries of 2 mm diameter were used as a sample holder and the temperature was controlled using a heating/cooling capillary holding stage (Linkam Scientific Instruments Ltd., Tadworth, UK), with 5 min equilibration being allowed prior to data collection over 30 min (for 1.0% w/w dispersions) or 5 min (for 20% w/w dispersions). Data were reduced (normalization and integration) using the Foxtrot software package supplied with the instrument and further analyzed (background subtraction and data modelling) using Irena SAS macros[@cit79] for Igor Pro.\n\nResults and discussion\n======================\n\nSynthesis of PBeMA~37~--PBzMA~*x*~ diblock copolymer spheres\n------------------------------------------------------------\n\nThe near-monodisperse PBeMA~37~ macro-CTA (*M*~w~/*M*~n~ = 1.18) obtained from the RAFT solution polymerization of BeMA in toluene at 70 \u00b0C was utilized for the RAFT dispersion polymerization of BzMA in mineral oil at 90 \u00b0C to produce a range of PBeMA~37~--PBzMA~*x*~ diblock copolymer spheres (see [Table 1](#tab1){ref-type=\"table\"}). High (\\>99%) BzMA conversions were achieved in all cases, which is consistent with previously reported PISA formulations utilizing PBzMA core-forming blocks.[@cit41],[@cit48],[@cit49],[@cit74]--[@cit77] Reasonably narrow molecular weight distributions (*M*~w~/*M*~n~ \u2264 1.38) were obtained when targeting PBzMA DPs up to 300. A clear shift in molecular weight was observed for these diblock copolymers relative to the corresponding PBeMA~37~ macro-CTA (see [Fig. 1a](#fig1){ref-type=\"fig\"}). Moreover, a linear evolution of *M*~n~ with target PBzMA DP was indicated by THF GPC analysis (see [Table 1](#tab1){ref-type=\"table\"} and [Fig. 1b](#fig1){ref-type=\"fig\"}). Interestingly, all dispersions formed turbid, free-standing gels/pastes immediately after cooling from 90 \u00b0C to 20 \u00b0C (see Fig. S2a[\u2020](#fn1){ref-type=\"fn\"}), whereas free-flowing fluids of varying turbidity were observed on reheating to 50 \u00b0C (see Fig. S2b[\u2020](#fn1){ref-type=\"fn\"}). This thermal transition proved to be fully reversible.\n\n###### Summary of targeted copolymer compositions, BzMA conversions (% BzMA) as judged by ^1^H NMR spectroscopy, GPC and DLS data (particle diameter and polydispersity index, PDI) obtained for a series of PBeMA~37~--PBzMA~*x*~ diblock copolymers prepared by RAFT dispersion polymerization of BzMA in mineral oil. Synthesis conditions: 90 \u00b0C, \\[PBeMA~37~ macro-CTA\\]/\\[T21s\\] molar ratio = 5.0, 20% w/w solids. Relevant data for the PBeMA~37~ macro-CTA are also shown for reference\n\n  Target composition      \\% BzMA   THF GPC (*vs.* PMMA)   DLS at 50 \u00b0C         \n  ----------------------- --------- ---------------------- -------------- ----- ------\n  PBeMA~37~               ---       12\u2009400                 1.18           ---   ---\n  PBeMA~37~--PBzMA~50~    \\>99      16\u2009200                 1.15           21    0.08\n  PBeMA~37~--PBzMA~100~   \\>99      22\u2009700                 1.15           32    0.01\n  PBeMA~37~--PBzMA~150~   \\>99      28\u2009100                 1.18           37    0.02\n  PBeMA~37~--PBzMA~200~   \\>99      33\u2009800                 1.24           55    0.01\n  PBeMA~37~--PBzMA~300~   \\>99      43\u2009900                 1.38           67    0.01\n\n![(a) Normalized THF gel permeation chromatograms \\[against poly(methyl methacrylate) standards\\] obtained for the series of PBeMA~37~--PBzMA~*x*~ diblock copolymers synthesized *via* RAFT dispersion polymerization of BzMA in mineral oil at 90 \u00b0C and 20% w/w solids. The precursor PBeMA~37~ macro-CTA (prepared in toluene at 70 \u00b0C and 50% w/w solids) is shown as a reference (black dashed curve). (b) *M*~n~*vs.* PBzMA DP plot for the same PBhMA~37~--PBzMA~*x*~ series, where the *y*-intercept represents the PBhMA~37~ macro-CTA. The solid line represents the theoretical evolution of *M*~n~ with PBzMA DP.](c8sc00762d-f1){#fig1}\n\n![(a) DLS particle size distributions recorded for 0.10% w/w dispersions of PBeMA~37~--PBzMA~*x*~ diblock copolymer spheres in *n*-dodecane at 50 \u00b0C. (b) The apparent hydrodynamic diameter (*D*) determined by DLS at 50 \u00b0C *vs.* PBzMA DP (*x*) plot indicates an *\u03b1* scaling exponent of 2/3. (c) TEM images for selected PBeMA~37~--PBzMA~*x*~ spheres obtained after drying 0.10% w/w dispersions in *n*-dodecane. See Fig. S3[\u2020](#fn1){ref-type=\"fn\"} for additional higher magnification TEM images.](c8sc00762d-f2){#fig2}\n\nDLS studies were performed at 50 \u00b0C to determine the mean particle diameter for 0.10% w/w PBeMA~37~--PBzMA~*x*~ dispersions in *n*-dodecane (see [Table 1](#tab1){ref-type=\"table\"}, and [Fig. 3a and b](#fig3){ref-type=\"fig\"}). All dispersions exhibited narrow size distributions (PDI \\< 0.10) and a monotonic increase in apparent hydrodynamic particle diameter (*D*) was observed when targeting a higher PBzMA DP (*x*), where *D* = *kx*^*\u03b1*^ (see [Fig. 2b](#fig2){ref-type=\"fig\"}). The scaling exponent (*\u03b1*) of 2/3 indicates that the PBeMA~37~--PBzMA~*x*~ diblock copolymer chains lie within the strong segregation regime.[@cit80] This scaling relationship enables reproducible targeting of spheres with a predetermined diameter.[@cit75],[@cit76] TEM images confirm that well-defined spherical morphologies were obtained in all cases (see [Fig. 2c](#fig2){ref-type=\"fig\"}). It is emphasized that well-defined PBeMA~37~--PBzMA~*x*~ spheres of controllable diameter can be prepared despite the significant loss of control over the molecular weight distribution that occurs when targeting higher PBzMA DPs.\n\n![% Transmittance *vs.* temperature plots for (a) a 1.0% w/w solution of the PBeMA~37~ macro-CTA and (b) a 1.0% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil. The optical transmittance was monitored at a fixed wavelength of 650 nm on cooling from 50 \u00b0C to 20 \u00b0C (blue squares) and on heating from 20 \u00b0C to 50 \u00b0C (red circles), with 5 min being allowed for equilibration at each temperature. Solid lines are shown as a guide to the eye.](c8sc00762d-f3){#fig3}\n\nThermoreversible crystallization-driven aggregation of PBeMA~37~--PBzMA~*x*~ diblock copolymer spheres\n------------------------------------------------------------------------------------------------------\n\nA series of experiments were conducted to determine precisely how the PBeMA~37~ stabilizer block dictates the colloidal stability of PBeMA~37~--PBzMA~*x*~ spheres in mineral oil. Firstly, turbidimetry measurements were performed on a 1.0% w/w solution of PBeMA~37~ macro-CTA in mineral oil ([Fig. 3a](#fig3){ref-type=\"fig\"}) and a 1.0% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil ([Fig. 3b](#fig3){ref-type=\"fig\"}). At 50 \u00b0C, approximately 100% transmittance was observed in both cases. This indicates that the PBeMA chains are fully soluble and thus the sterically-stabilized spheres should be well dispersed and non-interacting under the same conditions.[@cit81] On cooling to below 32 \u00b0C, the PBeMA~37~ chains precipitate from mineral oil and the initially clear solution becomes turbid (\u223c0% transmittance). As anticipated for a first-order phase transition, thermal hysteresis is observed on heating; the critical temperature for redissolution is approximately 45 \u00b0C (see [Fig. 3a](#fig3){ref-type=\"fig\"}). Similar behavior is also observed for a 1.0% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres. However, in this case cooling from 50 \u00b0C leads to nanoparticle flocculation rather than homopolymer precipitation. Moreover, the critical temperature for this phase transition is approximately 27 \u00b0C (see [Fig. 3b](#fig3){ref-type=\"fig\"}). In principle, this difference may be partly attributable to the lower concentration of PBeMA~37~ stabilizer chains in this dispersion (\u223c0.45% w/w) compared to that of the PBeMA~37~ solution (1.0% w/w). Alternatively, crystallization may be suppressed because one end of each crystallizing PBeMA~37~ stabilizer chain is bound to an interface (the PBzMA nanoparticle cores). The melting temperature required for reconstitution of a colloidally stable dispersion on heating is approximately 42 \u00b0C (see [Fig. 3b](#fig3){ref-type=\"fig\"}), which is comparable to the critical temperature required for redissolution of the free PBeMA~37~ stabilizer chains (see [Fig. 3a](#fig3){ref-type=\"fig\"}). We hypothesize that the strong hysteresis observed in the present study arises primarily because flocculation is driven by crystallization of the alkyl side chains on the PBeMA stabilizer, which involves a first-order phase transition as opposed to a second-order phase transition (*i.e.* liquid--liquid phase separation). Similar local ordering has been previously reported for closely-related poly(*n*-alkyl methacrylates) such as poly(stearyl methacrylate): DSC studies indicate that side-chain crystallization occurs in the solid state for both this homopolymer and also for several poly(stearyl methacrylate)-based diblock copolymers.[@cit82],[@cit83] To test our hypothesis of crystallization-driven flocculation, we performed DSC experiments on (i) the PBeMA~37~ precursor in the solid state ([Fig. 4a](#fig4){ref-type=\"fig\"}), (ii) a 20% w/w PBeMA~37~ solution in mineral oil ([Fig. 4b](#fig4){ref-type=\"fig\"}) and (iii) a 20% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil ([Fig. 4c](#fig4){ref-type=\"fig\"}). For the PBeMA~37~ homopolymer in the solid state, a melting enthalpy (\u0394*H*~m~) of 77.2 J g^--1^ corresponding to the crystalline behenyl side-chains indicates a mean degree of crystallinity of \u223c31% relative to the theoretical value reported for the polyethylene unit cell.[@cit84] The \u0394*H*~m~ for PBeMA~37~ in a 20% w/w mineral oil solution is reduced approximately five-fold to 15.8 J g^--1^, simply owing to dilution. The same transitions can be observed in both cases, although the exotherm/endotherm becomes much weaker and the transition temperatures are suppressed for lower concentrations of PBeMA~37~ chains, as expected. These transitions can be attributed to the crystallization/melting of the behenyl side-chains within the PBeMA~37~ block. Interestingly, multiple peaks were observed in the DSC traces recorded for a 20% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil (see [Fig. 4c](#fig4){ref-type=\"fig\"}). These features indicate subtly different crystallization events. More specifically, behenyl side-chains within individual (isolated) nanoparticles crystallize first (see peak at \u223c26 \u00b0C in [Fig. 4c](#fig4){ref-type=\"fig\"}), before then acting as nucleation sites for neighboring nanoparticles (see peak at \u223c21 \u00b0C in [Fig. 4c](#fig4){ref-type=\"fig\"}), thus leading to the formation of colloidal aggregates of strongly-interacting PBeMA~37~--PBzMA~100~ nanoparticles.\n\n![Differential scanning calorimetry (DSC) experiments conducted at a cooling/heating rate of 2 \u00b0C min^--1^ for (a) PBeMA~37~ macro-CTA in the solid state, (b) a 20% w/w solution of PBeMA~37~ in mineral oil and (c) a 20% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil. Two thermal cycles beginning at the maximum temperature were performed and are presented on the plot. Blue and red data represent cooling and heating ramps, respectively. \\* Indicates the behenyl side-chain crystallization within individual (isolated) nanoparticles. \\*\\* Indicates the secondary crystallization between PBeMA~37~--PBzMA~100~ nanoparticles.](c8sc00762d-f4){#fig4}\n\nOscillatory rheology experiments performed on the same 20% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spheres in mineral oil (see [Fig. 5](#fig5){ref-type=\"fig\"}) are consistent with the transition temperatures determined by DSC (26 \u00b0C on cooling and 41 \u00b0C on heating, [Fig. 4c](#fig4){ref-type=\"fig\"}). On cooling from 50 \u00b0C to 27 \u00b0C, the 20% w/w dispersion exhibited fluid-like properties, with the loss modulus (*G*\u2032\u2032) being much larger than the storage modulus (*G*\u2032). The liquid-to-solid transition is defined by the crossover in the *G*\u2032 and *G*\u2032\u2032 curves at 25 \u00b0C, which occurs on further cooling. Hysteresis is observed as the dispersion is heated from 20 \u00b0C to 50 \u00b0C: its solid-like properties are initially retained (*G*\u2032 \\> *G*\u2032\u2032) before reverting to its original fluid-like state (*G*\u2032\u2032 \u226b *G*\u2032) at 41 \u00b0C. Despite this hysteresis, a fully thermoreversible transition is observed. These results also support those depicted in the optical digital images shown in Fig. S2:[\u2020](#fn1){ref-type=\"fn\"} free-flowing dispersions are obtained at 50 \u00b0C, while solid paste-like dispersions are formed at 20 \u00b0C. Moreover, the rheology and DSC studies performed on the 20% w/w dispersion of PBeMA~37~--PBzMA~100~ nanoparticles correlate well with the optical transmittance data recorded for the corresponding 1.0% w/w dispersion ([Fig. 3b](#fig3){ref-type=\"fig\"}).\n\n![Temperature dependence of the storage modulus (*G*\u2032, blue squares) and loss modulus (*G*\u2032\u2032, red circles) for a 20% w/w dispersion of 32 nm PBeMA~37~--PBzMA~100~ spherical nanoparticles in mineral oil on cooling from 50 \u00b0C to 20 \u00b0C (filled symbols) and on heating from 20 \u00b0C to 50 \u00b0C (open symbols). Data were recorded at 1.0% strain amplitude using an angular frequency of 10 rad s^--1^, with 5 min being allowed for thermal equilibration between each measurement.](c8sc00762d-f5){#fig5}\n\nX-ray scattering studies\n------------------------\n\nVariable-temperature SAXS studies were conducted to further probe the aggregation behavior of the PBeMA~37~--PBzMA~100~ nanoparticles (see [Fig. 6](#fig6){ref-type=\"fig\"}). Initially, data were collected for a dilute 1.0% w/w dispersion; inter-particle interactions are negligible under such conditions so the particle form factor could be estimated. At 50 \u00b0C (see [Fig. 6a](#fig6){ref-type=\"fig\"}, red squares), the zero gradient observed at low *q* in the *I*(*q*) *vs. q* plot is consistent with non-interacting spherical particles. Moreover, this pattern can be well-fitted using a spherical micelle model (see SAXS model section in the ESI[\u2020](#fn1){ref-type=\"fn\"}),[@cit85] which indicates a spherical core radius (*R*~s~) of 8.1 \u00b1 1.0 nm and a radius of gyration (*R*~g~) for the PBeMA~37~ stabilizer chains of 1.49 nm (see [Fig. 6a](#fig6){ref-type=\"fig\"} and [Table 2](#tab2){ref-type=\"table\"}). The micelle core diameter (*D*~c~ = 2*R*~s~) of 16.2 \u00b1 2.0 nm correlates well with the TEM image shown in Fig. S3[\u2020](#fn1){ref-type=\"fn\"} which indicates a mean core diameter of 16.2 \u00b1 1.8 nm (estimated from analysis of 50 nanoparticles). The corresponding volume-average diameter *D*~v~ is 22.1 \u00b1 3.7 nm (where *D*~v~ = 2*R*~s~ + 4*R*~g~), which is somewhat smaller than the hydrodynamic *z*-average diameter of 32 nm reported by DLS (see [Table 1](#tab1){ref-type=\"table\"}). In addition, SAXS analysis indicates that the volume fraction of solvent within the nanoparticle cores (*x*~sol~) is essentially zero. Similar results have been recently reported for other diblock copolymer nanoparticles comprising PBzMA core-forming blocks in both *n*-dodecane[@cit48] and mineral oil.[@cit76]\n\n![Small-angle X-ray scattering (SAXS) patterns recorded for (a) a 1.0% w/w dispersion and (b) a 20% w/w dispersion of PBeMA~37~--PBzMA~100~ spheres in mineral oil at 50 \u00b0C before the thermal cycle (red squares), 20 \u00b0C after cooling (black circles) and 50 \u00b0C after reheating (blue triangles). Solid lines indicate relevant data fits to an established spherical micelle model (see ESI[\u2020](#fn1){ref-type=\"fn\"}),[@cit85] with a sticky hard-sphere[@cit87] or a Percus--Yevick hard-sphere[@cit88] structure factor being incorporated where appropriate. Patterns are offset by an arbitrary factor (indicated on the left side of the patterns) for clarity.](c8sc00762d-f6){#fig6}\n\n###### Summary of parameters after fitting SAXS patterns obtained for 1.0% w/w and 20% w/w dispersions of PBeMA~37~--PBzMA~100~ nanoparticles in mineral oil at 50 \u00b0C and 20 \u00b0C using a spherical micelle model (see ESI),[@cit85] incorporating a sticky hard-sphere[@cit87] or a Percus--Yevick hard-sphere[@cit88] structure factor, where appropriate. *\u03c6* is the volume fraction of copolymers obtained from the fitting, *R*~s~ is the spherical micelle core radius, *R*~g~ is the radius of gyration of the PBeMA stabilizer block and the volume-average nanoparticle diameter, *D*~v~, is subsequently calculated using *D*~v~ = 2*R*~s~ + 4*R*~g~. *N*~s~ is the average number of copolymer chains per micelle, equal to ![](c8sc00762d-t1.jpg){#ugt1}, where *x*~sol~ is the volume fraction of solvent within the core domain (equal to 0 in all cases), and *V*~s~ is the molecular volume of one core-forming PBzMA chain ![](c8sc00762d-t2.jpg){#ugt2}. *S*~agg~ is the number of copolymer chains per unit surface area at the micelle core, equal to ![](c8sc00762d-t3.jpg){#ugt3}, and *d*~int~ is the average distance between neighboring chains at the core-corona interface, equal to ![](c8sc00762d-t4.jpg){#ugt4}. *D*~SHS~ is the mean interaction distance and *f*~SHS~ is the effective volume fraction of interacting spheres calculated by fitting the sticky hard-sphere model with a stickiness parameter (*\u03c4*) of 0.10, while *D*~PY~ is the mean interaction distance and *f*~PY~ is the effective volume fraction of interacting spheres returned from fitting to the hard-sphere structure factor. The fractal dimension, *p*, describes the upturn in scattering intensity at low *q*\n\n  Copolymer concentration/% w/w   *T*/\u00b0C   *\u03c6*       *R* ~s~/nm   *R* ~g~/nm   *D* ~v~/nm   *N* ~s~   *S* ~agg~/nm^--2^   *d* ~int~/nm   *D* ~SHS~/nm   *f* ~SHS~   *D* ~PY~/nm   *f* ~PY~   *p*\n  ------------------------------- -------- --------- ------------ ------------ ------------ --------- ------------------- -------------- -------------- ----------- ------------- ---------- ------\n  1.0                             50       0.00824   8.1 \u00b1 1.0    1.49         22.1 \u00b1 3.7   86 \u00b1 18   0.11 \u00b1 0.03         3.1 \u00b1 0.8      ---            ---         ---           ---        ---\n  1.0                             20       0.00772   8.1 \u00b1 1.6    0.95\\*       19.9 \u00b1 5.4   86 \u00b1 29   0.11 \u00b1 0.05         3.1 \u00b1 1.3      27.9           0.194       ---           ---        1.67\n  20                              50       0.135     7.7 \u00b1 0.7    2.45         25.1 \u00b1 3.0   74 \u00b1 11   0.10 \u00b1 0.02         3.2 \u00b1 0.6      ---            ---         52.1          0.184      ---\n  20                              20       0.105     7.7 \u00b1 1.1    2.04\\*       23.6 \u00b1 4.9   76 \u00b1 19   0.10 \u00b1 0.03         3.2 \u00b1 1.0      22.9           0.374       ---           ---        1.67\n\nOn cooling to 20 \u00b0C, a local maximum in scattering intensity was observed at *q* \u223c 0.26 nm^--1^ (see [Fig. 6a](#fig6){ref-type=\"fig\"}, black circles). Moreover, there is an upturn in scattering intensity at low *q*, suggesting the formation of scattering objects that are significantly larger than the individual nanoparticles. These observations indicate that the nanoparticles form large aggregates on cooling even at a relatively low copolymer concentration (1.0% w/w). Furthermore, the slope of the scattered X-ray intensity at low *q* is approximately --2 (see [Fig. 6](#fig6){ref-type=\"fig\"}), which is consistent with the formation of mass fractals[@cit86] by a network of interconnected spherical nanoparticles. This is consistent with the solid-like properties of a 20% w/w PBeMA~37~--PBzMA~100~ dispersion at 20 \u00b0C indicated by rheology (see [Fig. 5](#fig5){ref-type=\"fig\"}), the digital images of the dispersions (see Fig. S2[\u2020](#fn1){ref-type=\"fn\"}), and the transmission electron micrographs obtained for a dilute dispersion at 20 \u00b0C (see Fig. S3c[\u2020](#fn1){ref-type=\"fn\"}). Dividing the scattering pattern obtained at 20 \u00b0C with that recorded at 50 \u00b0C reveals the structure factor \\[see *S*(*q*) curve in Fig. S4a[\u2020](#fn1){ref-type=\"fn\"}\\], which is attributed to partial crystallization of the PBeMA~37~ coronal blocks between neighboring nanoparticles. Fitting the latter data to a sticky hard-sphere (SHS) model[@cit87] indicated a mean centre-to-centre interaction distance (*D*~SHS~) of 26.7 nm and an interacting sphere volume fraction (*f*~SHS~) of 0.078 (see Fig. S4a[\u2020](#fn1){ref-type=\"fn\"}). This approach does not provide a satisfactory fit, but this is not unexpected given that the two data sets used in this analysis were collected at different temperatures. Thus the volumetric contraction of the nanoparticles that occurs on cooling introduces a systematic error. Using the fitting parameters obtained for the colloidally stable 1.0% w/w dispersion at 50 \u00b0C and these *D*~SHS~ and *f*~SHS~ values as a starting point, the SAXS pattern obtained for the same flocculated 1.0% w/w dispersion at 20 \u00b0C was fitted. The primary nanoparticle core dimensions remain more or less unaffected, with an *R*~s~ of 8.1 \u00b1 1.6 nm being observed at this lower temperature (see [Fig. 6a](#fig6){ref-type=\"fig\"} and [Table 2](#tab2){ref-type=\"table\"}). However, the radius of gyration, *R*~g~, returned for the PBeMA~37~ stabilizer block when fitting the SAXS pattern recorded at 20 \u00b0C is lower than that indicated by the corresponding 50 \u00b0C pattern. In reality, these partially crystalline PBeMA chains are not expected to exhibit Gaussian behavior at 20 \u00b0C since their mobility is confined by interactions with neighboring PBeMA chains within the crystal. In addition, such crystallization should lead to an increase in mass density (and hence scattering length density) compared to the non-crystalline stabilizer chains at 50 \u00b0C. However, this rather subtle effect is not included in the spherical micelle model. This fitting strategy returned a *D*~SHS~ of 27.9 nm and an *f*~SHS~ of 0.194 (see [Table 2](#tab2){ref-type=\"table\"}). Importantly, the scattering pattern obtained on reheating to 50 \u00b0C (see Fig. S4a,[\u2020](#fn1){ref-type=\"fn\"} blue triangles) overlaps almost perfectly with that observed for the original spheres at 50 \u00b0C (see Fig. S4a,[\u2020](#fn1){ref-type=\"fn\"} red squares), which confirms the fully reversible nature of this thermal transition.\n\nFurther SAXS analysis was conducted to confirm that the temperature-dependent behavior observed for the 1.0% w/w copolymer dispersion in mineral oil is consistent with that found for the 20% w/w dispersion of PBeMA~37~--PBzMA~100~ spheres. Firstly, data were collected at 50 \u00b0C (see [Fig. 6b](#fig6){ref-type=\"fig\"}, red squares) and fitted using the same spherical micelle model[@cit85] (see [Table 2](#tab2){ref-type=\"table\"}). The sticky hard-sphere model used at 20 \u00b0C cannot be employed to represent the inter-particle interactions at 50 \u00b0C, since this is above the melting temperature of the partially crystalline PBeMA stabilizer chains, hence an amorphous free-flowing dispersion is obtained. Thus, a simple hard-sphere interaction based on the Percus--Yevick (PY) approximation suffices under such conditions.[@cit88] The numerical values for *R*~s~ (7.7 \u00b1 0.7 nm) and *D*~v~ (25.1 \u00b1 3.0 nm) returned from this fit are in reasonably good agreement with those obtained for the 1.0% w/w dispersion (see [Table 2](#tab2){ref-type=\"table\"}). The PY approximation indicates that a significant fraction of these PBeMA~37~--PBzMA~100~ spheres (*f*~PY~ = 0.184) are strongly interacting, with a mean centre-to-centre interaction distance (*D*~PY~) of 52.1 nm. Moreover, the latter parameter is larger than *D*~v~, which indicates that sphere--sphere interactions arise purely from their proximity in space. On cooling to 20 \u00b0C (see [Fig. 6b](#fig6){ref-type=\"fig\"}, black circles), *R*~s~ remained relatively unchanged (7.7 \u00b1 1.1 nm), which corresponds to a *D*~v~ of 23.6 \u00b1 4.9 nm. On cooling from 50 \u00b0C to 20 \u00b0C, the mean interaction distance is reduced from *D*~PY~ = 52.1 nm to *D*~SHS~ = 22.9 nm. The latter value is less than *D*~v~ at 20 \u00b0C, which provides compelling (albeit indirect) evidence that the interparticle interactions involve crystallization between PBeMA stabilizer chains located on neighboring spherical nanoparticles. The upturn in scattering intensity at low *q* is similar to that observed for the 1.0% w/w dispersion and again indicates the formation of loose mass fractals.[@cit86]\n\nDirect evidence for behenyl side-chain crystallization between neighboring PBeMA chains is provided by the subtle feature at *q* = 1.80 nm^--1^ in the SAXS pattern recorded for a 20% w/w dispersion at 20 \u00b0C ([Fig. 6b](#fig6){ref-type=\"fig\"}), which indicates a mean *d*-spacing of 3.49 nm (*d* = 2\u03c0/*q*). This characteristic length scale corresponds to a periodic lamellar structure formed by co-crystallizing methacrylic backbones and is similar to that reported for poly(stearyl methacrylate).[@cit82] Assuming every monomer repeat unit comprises two C--C bonds in an all-trans conformation, the theoretical maximum length of one fully-stretched behenyl side-chain is 2.81 nm. Hence the distance between non-interacting side-chains should be greater than 5.62 nm. Thus, a mean spacing of 3.49 nm implies significant interdigitation between PBeMA stabilizer chains on neighboring nanoparticles (see [Fig. 7](#fig7){ref-type=\"fig\"}). This interaction distance is physically reasonable when compared to the mean distance between adjacent diblock copolymer chains located at the nanoparticle surface (*d*~int~), which is calculated to be 3.2 nm for this 20% w/w dispersion at 20 \u00b0C (see [Table 2](#tab2){ref-type=\"table\"}).\n\n![Schematic representation of the characteristic length scales for the structural parameters associated with the formation of loose mass fractal aggregates *via* crystallization-driven self-assembly (CDSA) of a 20% w/w dispersion of PBeMA~37~--PBzMA~100~ nanoparticles in mineral oil at 20 \u00b0C. The mean *d*-spacing of 0.416 nm between co-crystallizing behenyl side-chains is calculated from WAXS analysis, whereas all other parameters are derived from SAXS analysis (see main text for further details).](c8sc00762d-f7){#fig7}\n\nWAXS patterns (recorded simultaneously with the SAXS data) were used to examine crystallization between the behenyl side-chains at shorter length scales. Comparing WAXS data recorded for the 20% w/w dispersion of PBeMA~37~--PBzMA~100~ nanoparticles at 20 \u00b0C to that obtained for mineral oil alone revealed a discernible broad peak at *q* = 15.1 nm^--1^ for the former system (see Fig. S5[\u2020](#fn1){ref-type=\"fn\"}). This corresponds to a mean *d*-spacing of 0.416 nm (see [Fig. 7](#fig7){ref-type=\"fig\"}), which is in good agreement with the 100 reflection reported for the hexagonal rotator phase of *n*-alkanes,[@cit89] and also with that reported for the closely-related poly(stearyl methacrylate).[@cit82] A schematic representation of the characteristic length scales associated with the formation of loose mass fractal aggregates *via* crystallization-driven self-assembly of sterically-stabilized nanoparticles is shown in [Fig. 7](#fig7){ref-type=\"fig\"}.\n\nIt is proposed that the onset of crystallization involves homogeneous nucleation within *isolated* PBeMA~37~--PBzMA~100~ nanoparticles, hence substantial undercooling is required.[@cit90] These crystalline domains decorate the nanoparticles, hence any contact with neighboring nanoparticles *via* Brownian motion initiates secondary nucleation, which is rapid compared to nanoparticle diffusion. Thus nanoparticles impinging on such crystalline domains rapidly crystallize and stick, quickly leading to the formation of loose fractal aggregates *via* reaction-limited aggregation.[@cit91] This mechanism is consistent with the DSC and rheology measurements, whereby crystallization and gelation are observed under similar conditions for the same 20% w/w dispersion of PBeMA~37~--PBzMA~100~ spheres. Importantly, when this relatively concentrated dispersion is returned to 50 \u00b0C, its new scattering pattern overlaps perfectly with that initially obtained at 50 \u00b0C (see Fig. S4[\u2020](#fn1){ref-type=\"fn\"}), confirming full reversibility for this crystallization-driven thermal transition. It is perhaps worth emphasizing that such CDSA-mediated nanoparticle aggregation behavior distinguishes the current study from the various literature reports of upper critical solution temperature (UCST)-driven aggregation of sterically-stabilized nanoparticles in which the amorphous stabilizer block becomes insoluble on lowering the temperature.[@cit92]\n\nConclusions\n===========\n\nIn summary, a series of PBeMA~37~--PBzMA~*x*~ spherical nanoparticles are prepared *via* RAFT dispersion polymerization of benzyl methacrylate at 90 \u00b0C in mineral oil at 20% w/w solids. On cooling to 20 \u00b0C, turbid pastes are formed for all PBeMA~37~--PBzMA~*x*~ spheres in mineral oil as a result of the crystallization of the insoluble PBeMA stabilizer block. However, heating to 50 \u00b0C leads to free-flowing dispersions, with DLS studies performed at this temperature indicating narrow size distributions and a strong correlation between the mean degree of polymerization of the core-forming PBzMA block and the *z*-average nanoparticle diameter. On cooling from 50 \u00b0C to 20 \u00b0C, turbidimetry studies conducted on a 1.0% w/w solution of the PBeMA~37~ macro-CTA in mineral oil indicated that precipitation occurred at around 32 \u00b0C, with redissolution occurring at around 45 \u00b0C on reheating. Thus this control experiment demonstrates that the thermosensitive nature of the PBeMA~37~ stabilizer chains determines the colloidal stability of the PBeMA~37~--PBzMA~*x*~ spheres. Oscillatory rheology experiments were performed to study the sol--gel behavior of 20% w/w PBeMA~37~--PBzMA~100~ nanoparticles in mineral oil: a free-flowing fluid of non-interacting spheres was observed at 50 \u00b0C, whereas a solid-like paste of strongly-interacting spheres was obtained on cooling to 20 \u00b0C. Notably, critical transition temperatures for the same 20% w/w dispersion of PBeMA~37~--PBzMA~100~ spheres in mineral oil determined *via* DSC and rheology were in excellent agreement. SAXS studies confirmed that PBeMA~37~--PBzMA~100~ spheres in mineral oil became strongly interacting and formed loose mass fractals at 20 \u00b0C. Importantly, the mean interaction distance for spherical nanoparticles at 20 \u00b0C was less than the volume-average diameter of the spheres, thus providing indirect evidence for crystallization between PBeMA chains on neighboring nanoparticles. Furthermore, WAXS studies indicated a crystalline hexagonal rotator phase between PBeMA chains at 20 \u00b0C, with a mean *d*-spacing of 4.16 \u00c5 between 100 planes formed by co-crystallizing behenyl side-chains. In contrast, well-dispersed nanoparticles are obtained at 50 \u00b0C and this thermal transition is fully reversible. However, pronounced hysteresis can be observed under certain conditions owing to the crystallization/melting behavior of the PBeMA stabilizer chains.\n\nConflicts of interest\n=====================\n\nThere are no conflicts to declare.\n\nSupplementary Material\n======================\n\nSupplementary information\n\n###### \n\nClick here for additional data file.\n\nS. P. A. acknowledges an ERC Advanced Investigator grant (PISA 320372) and an EPSRC Platform grant (EP/J007846). S. P. A. and O. O. M. are grateful to EPSRC for a capital equipment grant to purchase the Xenocs/Excillum SAXS laboratory instrument (EP/M028437/1). The Leverhulme Trust is also thanked for post-doctoral funding of M. J. D. (RPG-2016-330). Dr N. J. W. Penfold is thanked for the synthesis of the PETTC RAFT agent.\n\n[^1]: \u2020Electronic supplementary information (ESI) available: Assigned ^1^H NMR spectra; digital images of 20% w/w dispersions; TEM images obtained at 20\u00a0\u00b0C; overlaid SAXS patterns; WAXS patterns; SAXS models used. See DOI: [10.1039/c8sc00762d](10.1039/c8sc00762d)\n"}
+{"text": "INTRODUCTION {#cesec10}\n============\n\nOsteoarthritis (OA) is a disease of inflammatory and degenerative nature that causes destruction of the joint cartilage and can lead to acute and chronic painful conditions and joint deformities[@bib1]. Its prevalence ranges from around 4% to 30%, and its incidence is especially high among the elderly population, although this disease is not an inevitable consequence of advancing age[@bib2]. The knee is the joint most affected by OA, and its functionality status is strongly associated with changes to individuals' activities of daily living and autonomy[@bib3]. The World Health Organization (WHO) classifies OA in the knee as the fourth biggest cause of incapacity among women and the eighth among men[@bib4].\n\nDifferent clinical examinations can be used to delineate the diagnosis of OA, such as radiography[@bib2], magnetic resonance[@bib5] and, recently, bioimpedance[@bib6].\n\nOne characteristic common to these examinations is the fact that they all require interpretation and evaluation by clinicians, both to establish the diagnosis and to estimate the prognosis for the impairment. This introduces an important measurement bias into the process[@bib7].\n\nOver the last two decades, patients' perceptions of their state of health have been given value as a fundamental variable in making trustworthy clinical assessments and, consequently, in the therapeutic strategy[@bib8]. Information on patients perceptions of their state of heath is usually gathered through applying questionnaires. With regard to OA, the ones most used are the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Lequesne Index[@bib9], [@bib10], since these were developed specifically for use in assessing this disease. However, other instruments such as the subjective assessment of the \"International Knee Documentation Committee\" and the Lysholm scale also provide satisfactory results when applied to this population[@bib11]. The level of scientific authenticity of these instruments may vary depending on the sociocultural contexts and specific characteristics of populations, thus influencing the quality of the results from interpreting the questions[@bib12]. It is therefore necessary to identify the degree of adequacy of these instruments for assessing issues relating to pain, physical limitations and functionality, among Brazilian patients affected by OA. Furthermore, because this is a disease that predominantly affects one age group, an estimate of the influence of age on the final score may signify a better fit with reality for the results.\n\nThe aim of this study was to measure the validity and reliability of the WOMAC, IKDC and Lysholm questionnaires among patients with OA in the knee, and to determine the influence of age on the scores from these questionnaires.\n\nMETHODS {#cesec20}\n=======\n\nThe patients agreed to participate in the study by signing a commitment statement, thereby authorizing their participation. This study was approved by the institution's Ethics Committee.\n\nFifty-seven patients with a diagnosis of primary OA of the knee answered the Brazilian versions of the SF-36, WOMAC, Lysholm and IKDC questionnaires. Forty-one patients were female (71.9%) and sixteen were male (28.1%). The mean age of the sample was 61.7 years, with a range from 35 to 84 years.\n\nAccording to de Vet et al[@bib13], validity refers to the capacity of an instrument to measure what it is designed to measure. To measure the construct validity of the specific questionnaires for the knee, we correlated the values from these questionnaires with the physical domains of the SF-36, which deal with the factors of functional capacity, presence of pain and limitations due to physical problems. The convergent and divergent validities were assessed based on comparison between the results obtained through the WOMAC, IKDC and Lysholm questionnaires and the results relating to the eight domains of the SF-36, taking the hypothesis that the knee questionnaires would correlate better with the physical domains of the SF-36 than with its mental domains. The content validity was assessed through the distribution and occurrence of floor and ceiling effects, which occur respectively when a response attains a score of zero (the minimum) and when a response attains a score of 100 (i.e. the maximum possible).\n\nThe internal consistency and concordance between the questionnaires for examining the homogeneity between the items of a scale were calculated[@bib14]. Concordance refers to how close the values of two or more instruments are and, according to de Vet et al[@bib13], it denotes the absence of measurement errors.\n\nStatistical analysis {#cesec30}\n--------------------\n\nThe convergent and divergent construct validities were tested by means of Pearson's correlation coefficient (r). The internal consistency was calculated by means of Cronbach's \u03b1 coefficient[@bib14]. We examined the concordance between the questionnaires using the graphical representations of Altman-Bland[@bib15], [@bib16] and Concordance-Survival[@bib17]. Altman and Bland[@bib15], [@bib16] proposed that the concordance limits should be calculated from the differences observed, using the mean and standard deviation of the differences as the basis for calculating the limits. On the other hand, the concordance-survival curve[@bib17] expresses the degree of discordance as a function of the diverse tolerance limits, similar to the Kaplan-Myer survival analysis. However, instead of absolute differences observed between the measurements, this technique locates the modulus of the differences on the X axis and the proportion of discordant cases on the Y axis.\n\nTo determine the influence of age on the results, we examined the association between age and the scores from the WOMAC, IKDC and Lysholm questionnaires by means of Pearson's coefficient of determination (r^2^). The statistical analyses were performed using the Statistical Package for the Social Sciences software for Windows (SPSS Science Inc, version 13.00, Chicago, IL, USA), taking a confidence level of 5%.\n\nRESULTS {#cesec40}\n=======\n\nThe means, standard deviations and confidence intervals are presented in [Table 1](#tbl1){ref-type=\"table\"}.\n\nValidity {#cesec50}\n--------\n\nIKDC and WOMAC presented moderate to strong correlations in relation to the summary of physical capacity of SF-36, while the Lysholm questionnaire presented moderate correlations. All three questionnaires presented weak correlations with the mental components of SF-36, thus confirming the convergent and divergent validities ([Table 2](#tbl2){ref-type=\"table\"}). The three questionnaires presented good content validity, given that no floor or ceiling effects were observed.\n\nInternal consistency {#cesec60}\n--------------------\n\nCronbach's \u03b1 demonstrated values of 0.811 for IKDC, 0.959 for WOMAC and 0.734 for Lysholm. When one item at a time was excluded, Cronbach's \u03b1 was no larger than the original in any of the questionnaires, which negated any need to exclude any items from the scales when applied to patients with OA.\n\nConcordance {#cesec70}\n-----------\n\nDespite the strong correlations between WOMAC and IKDC (0.843), WOMAC and Lysholm (0.759) and IKDC and Lysholm (0.858), the graphical representations of Altman-Bland and Concordance-Survival showed that the concordance between the three questionnaires was low. The linear regression curve represented in the Altman-Bland graph demonstrated the presence of proportional bias between IKDC and WOMAC ([Figure 1](#fig1){ref-type=\"fig\"}) and IKDC and Lysholm ([Figure 2](#fig2){ref-type=\"fig\"}), while between WOMAC and Lysholm it seemed that there was a fixed bias, such that Lysholm was on average ten points lower than WOMAC ([Figure 3](#fig3){ref-type=\"fig\"}). The representation of concordance-survival ([Figure 4](#fig4){ref-type=\"fig\"}) confirmed the findings from the Altman--Bland graphs, regarding the lack of concordance between the three questionnaires.\n\nInfluence of age {#cesec80}\n----------------\n\nThe IKDC, Lysholm and WOMAC questionnaires presented Pearson's coefficients of determination (r^2^) of 0.004, 0.010 and 0.043 with age, respectively.\n\nDISCUSSION {#cesec90}\n==========\n\nMany studies have indicated the importance of using questionnaires for functional evaluations on patients affected by musculoskeletal diseases. Some of them have already been developed, translated and validated for the Brazilian population in relation to evaluations on patients with OA in the knee, including WOMAC[@bib18], IKDC[@bib12] and Lysholm[@bib19]. However, selection of the ideal questionnaire for assessing the limitations caused by specific diseases like OA is a matter that has been little discussed.\n\nAlthough all of these questionnaires present excellent psychometric properties for populations with joint diseases, it has not been determined which one would have the greatest capacity for ascertaining the physical limitations caused by OA. Since the incidence of OA increases with age, there is a need to estimate whether the responses obtained from the questionnaires would be influenced by this variable, independent of the severity of the clinical condition. In the present study, we chose to compare the criteria of scientific authenticity, validity, internal consistency and concordance between the WOMAC, IKDC and Lysholm questionnaires, in relation to patients with OA in the knee, and to identify the influence of the variable of age in determining the values of these parameters. No test-retest reproducibility verification was performed, given that the original studies of the Brazilian versions of the three questionnaires already presented excellent results. Consequently, it could be presumed that these properties would remain unchanged[@bib18].\n\nRegarding the validity of the questionnaires, it was found that WOMAC and IKDC presented stronger correlations with the physical components of SF-36 than did the Lysholm questionnaire. Nevertheless, none of the questionnaires presented very high correlations with these components ([Table 2](#tbl2){ref-type=\"table\"}). This suggests that WOMAC is more suitable for assessing the limitations relating to physical factors, while IKDC and Lysholm are more suited for assessing pain-related factors among these patients. Both WOMAC and IKDC were shown to be suitable for analyzing functional capacity. Despite these differences, all three questionnaires presented results that were more strongly associated with the physical capacities of SF-36 than with its mental capacities, thus ratifying the convergent and divergent validities of these questionnaires.\n\nThe internal consistency of the three questionnaires was shown to be adequate, although WOMAC presented the highest values, followed by IKDC. When one item was excluded from the analysis of the Cronbach index, the values remained smaller than with the complete scale, which does away with the need to exclude any item, for these questionnaires to be used with these patients. The higher values presented by WOMAC were expected, given that the original selection of the items[@bib9] was developed specifically for patients with osteoarthritis, while IKDC and Lysholm are questionnaires for overall assessments on knees.\n\nDespite the better correlation between IKDC and WOMAC in relation to Lysholm and WOMAC, the Altman-Bland graph ([Figure 1](#fig1){ref-type=\"fig\"}, [Figure 2](#fig2){ref-type=\"fig\"}, [Figure 3](#fig3){ref-type=\"fig\"}) showed low concordance between the questionnaires, which was confirmed in the concordance-survival representation ([Figure 4](#fig4){ref-type=\"fig\"}). According to de Vet et al[@bib13], concordance (agreement) and reliability are two psychometric properties that are included in the concept of reproducibility. They signify how close the values of two or more scales are to each other. De Vet et al[@bib13] defined concordance as the absence of measurement errors. In our study, the concordance-survival graph demonstrated that the best agreement was between the IKDC and Lysholm questionnaires ([Figure 4](#fig4){ref-type=\"fig\"}).\n\nTo achieve 70% concordance between individuals, the mean difference needed between these two questionnaires was approximately 10 points, while between IKDC and WOMAC and between WOMAC and Lysholm, it was 20 points. The presence of proportional bias between IKDC and WOMAC ([Figure 1](#fig1){ref-type=\"fig\"}) and between IKDC and Lysholm ([Figure 2](#fig2){ref-type=\"fig\"}) suggests that the greater the mean from the responses to the two questionnaires was, the greater the difference between them would be. On the other hand, between WOMAC and Lysholm ([Figure 3](#fig3){ref-type=\"fig\"}), we found that there was a relatively fixed bias, with a mean difference of approximately 10 points between the two questionnaires, such that WOMAC tended to have higher values that shown by Lysholm.\n\nNone of the questionnaires was influenced by the variable of age, since the low values of the coefficient of determination demonstrated that it would, at most, interfere by 5% in the final value from the questionnaires. If there had been a large negative correlation, we might have affirmed that age was influencing the values of the questionnaires, independent of the severity of the OA, but this did not happen. Hence, age was not shown to be a limiting factor for the use of any of the questionnaires applied in this study.\n\nWe take the view that the absence of the Lequesne Index in this assessment, an index that is specific for patients with OA, as is WOMAC, may be considered to be a limiting factor in this study. However, it is well known that extensive tests discourage correct responses from patients. Moreover, it is believed that between WOMAC and the Lequesne Index, the response would be specific, given that the construct validity is similar between these two questionnaires[@bib20]. The lack of tests commonly used to grade the severity of OA, such as the classifications of Kellgreen or Ahlback, for comparisons with the values obtained from the questionnaires may also be considered to be a limiting factor, but this was not an objective of the present study.\n\nCONCLUSION {#cesec100}\n==========\n\nThe concordance tests and the correlations with the physical components of the SF-36 suggest that none of the three questionnaires alone is capable of assessing all the matters relating to the physical limitations of patients with osteoarthritis. WOMAC in association with IKDC presented the best correlation values with the summary of the physical components of SF-36. While WOMAC assesses functional capacities and limitations relating to physical factors, IKDC seems to be better suited for assessing the functional limitations relating to pain. Age was not shown to be a factor limiting the use of any of the questionnaires applied in this study.\n\nWork performed at the Brazilian Institute of Health Technology (IBTS), Rio de Janeiro, Brazil.\n\nThe authors declare that there was no conflict of interest in conducting this work\n\n![Altman-Bland representation of the comparison between IKDC and WOMAC.](gr1){#fig1}\n\n![Altman-Bland representation of the comparison between IKDC and Lysholm.](gr2){#fig2}\n\n![Altman-Bland representation of the comparison between WOMAC and Lysholm.](gr3){#fig3}\n\n![concordance-survival test on the three comparisons (IKDC *versus* WOMAC, IKDC *versus* Lysholm and WOMAC versus Lysholm).](gr4){#fig4}\n\n###### \n\nMeans, standard deviations (SD) and 95% confidence intervals for the IKDC, SF-36, Lysholm and WOMAC questionnaires in the study sample.\n\n                      Mean   SD     95% CI\n  ------------------- ------ ------ -----------\n  **IKDC**            29.0   12.0   25.8-32.2\n  **Lysholm Score**   31.6   19.2   25.7-37.4\n  **WOMAC (Total)**   42.9   19.1   37.7-48.0\n  **SF-36 SPC**       34.1   19.6   28.9-32.2\n  **SF-36 SMC**       65.1   20.6   59.7-70.6\n  **SF-36 FC**        34.3   21.0   28.7-39.9\n  **SF-36 LP**        30.3   39.7   19.7-40.8\n  **SF-36 Pain**      35.1   21.9   29.3-40.9\n  **SF-36 GH**        58.9   25.6   52.1-65.7\n  **SF-36 VT**        57.9   21.4   52.2-63.6\n  **SF-36 SF**        68.9   28.0   61.4-76.3\n  **SF-36 LE**        43.3   43.6   31.7-54.9\n  **SF-36 MH**        66.5   21.8   60.7-72.3\n\n###### \n\nCorrelations between the questionnaires applied (IKDC, Lysholm and WOMAC) and the components of SF-36.\n\n                   Questionnaire   R           p-value\n  ---------------- --------------- ----------- -----------\n  **SF-36 SPC**    IKDC            0.62        \\< 0.0001\n  Lysholm          0.555           0.001       \n  WOMAC            0.642           \\< 0.0001   \n  **SF-36 SMC**    IKDC            0.251       0.0592\n  Lysholm          0.199           0.1945      \n  WOMAC            0.167           0.2217      \n  **SF-36 FC**     IKDC            0.485       0.0001\n  Lysholm          0.414           0.0052      \n  WOMAC            0.522           \\< 0.0001   \n  **SF-36 Lp**     IKDC            0.419       0.0012\n  Lysholm          0.374           0.0125      \n  WOMAC            0.524           \\< 0.0001   \n  **SF-36 pain**   IKDC            0.651       \\< 0.0001\n  Lysholm          0.652           \\< 0.0001   \n  WOMAC            0.580           \\< 0.0001   \n  **SF-36 GH**     IKDC            0.337       0.0103\n  Lysholm          0.205           0.182       \n  WOMAC            0.261           0.0546      \n  **SF-36 VT**     IKDC            0.275       0.0384\n  Lysholm          0.232           0.13        \n  WOMAC            0.244           0.073       \n  **SF-36 SF**     IKDC            0.302       0.022\n  Lysholm          0.256           0.093       \n  WOMAC            0.203           0.136       \n  **SF-36 LE**     IKDC            0.286       0.0308\n  Lysholm          0.282           0.0637      \n  WOMAC            0.350           0.0087      \n  **SF-36 MH**     IKDC            0.148       0.2713\n  Lysholm          0.099           0.5233      \n  WOMAC            0.061           0.6592      \n\nSPC: Summary of physical capacities; SMC: Summary of mental capacities; FC: Functional capacity; LP: Limitation due to physical factors; GH: General state of health; VT: Vitality; SF: Social factors; LE: Limitation due to emotional factors; MH: Mental health.\n"}
+{"text": "Dedication\n==========\n\nWe dedicate this paper to the memory of our dear friend and colleague Professor Colin Ingram who died December 15th 2013. Colin was the lead investigator on the CARMEN project, from which this study arose.\n\nBackground\n==========\n\nThe retina is the neural circuit within the eye responsible for converting light signals into neural activity. During at least the first postnatal week of life in the mouse, retinal ganglion cells (RGCs) are spontaneously active, generating waves of activity that propagate across the retina. These spontaneous activity patterns are thought to help refine the development of neuronal connections, since blocking or perturbing the activity leads to altered connectivity patterns. For reviews on the nature and role of spontaneous activity in the nervous system, see \\[[@B1],[@B2]\\].\n\nRetinal waves can be studied using both imaging methods and multielectrode arrays (MEAs). We have collected and annotated these recordings to allow researchers to compare the spatiotemporal properties of recordings obtained from different research groups. We have focused on curating recordings collected by MEAs, because although there are several types of array recording platforms available on the market, the underlying data after spike detection and sorting is simply a set of event times denoting when an action potential was detected on a particular electrode.\n\nThis paper describes the repository we have curated from many key papers investigating the nature of retinal spontaneous activity. We have converted these data files into a common format so that it can be easily shared with others, and provide some scripts to analyse these recordings. We created this repository for several reasons:\n\n1\\. By building a repository from many sources we are able to effectively compare findings from laboratories acquired under different experimental conditions, and from different transgenic mice.\n\n2\\. There are few public datasets of MEA recordings, although some are available accompanying research papers \\[[@B3]\\].\n\n3\\. We hope this platform will encourage future researchers to contribute their data. Many funding agencies now require data to be archived and shared for several years, and we hope this will serve as an example of how to share this type of data.\n\n4\\. Converting data to a standard open format, such as HDF5, should ensure that they can be read for many years to come. By contrast, keeping old datasets in proprietary formats may mean that the data are effectively unreadable in a few years.\n\n5\\. We have used these data as a demonstration for the workflow system in the CARMEN virtual laboratory.\n\nThis article is written as an example of \"reproducible research\", in that the results should be reproducible by others in a straightforward manner, given the same software and data \\[[@B4]\\]. The notion of reproducible research is beginning to be practised quite widely in some areas, such as Computational Biology \\[[@B5]\\], but is not yet that common within most fields, including Neuroscience \\[[@B6],[@B7]\\]. Figures and tables marked (Dynamic) in the legends of this article are regenerated dynamically, involving recomputation as needed. (Figures and tables marked (Static) are those that required no computation.) The source file for this article contains LATE X and R code, from which the paper is generated (see section \"Availability of supporting data\"). Links to all files required to regenerate the paper are provided on the accompanying website \\[[@B8]\\].\n\nData description\n================\n\nThe project web page \\[[@B8]\\] contains links to the data and code, and will list any updates to the repository. The data are freely available on the CARMEN portal \\[[@B9]\\]. Free registration to the CARMEN system is required to access the data.\n\nThe data provided to us from different laboratories arrived in several text and binary formats. We converted them to one common format to promote their reuse. We chose the open format HDF5 \\[[@B10]\\] because it provides an efficient and portable framework for storing large datasets. It is supported by many popular computational environments, such as R, Python, Mathematica, Matlab and Julia, and is freely available on all major operating systems. HDF5 is used across many scientific disciplines and is well-tested.\n\nHDF5 stores objects in a hierarchical tree that can be fully specified by the user. Our approach is to store the principal data items (such as spike times and electrode positions) for a recording in the top level of the tree. Relevant metadata about recordings (such as the age of the retina, and the species) are stored in objects under the `/meta/` group of the tree.\n\nData format for storage of MEA recordings\n-----------------------------------------\n\nThe following objects are stored in the root of the HDF5 data files:\n\n1\\. `epos` : an *N*\u00d7 2 matrix, where *N* is the number of spike trains in the recording. Row *i* stores the (*x*,*y*) location assigned to spike train *i* in this recording. The values *x* and *y* are specified in *\u03bc*m.\n\n2\\. `sCount `: a vector of length *N*. *sCount*\\[ *i*\\] stores the number of spikes included in spike train *i*.\n\n3\\. `spikes`: a vector of length S, where *S*= *sum*(*sCount*). These are the spike times (in seconds). The spike trains for each electrode are concatenated into one long vector, so that the spikes for electrode *j*\u2208 \\[ 1,*N*\\] are stored in elements *a* to *b*, where $a = \\sum\\limits_{i = 1}^{j - 1}\\textit{sCount}\\left\\lbrack \\; i \\right\\rbrack$ and *b*= *a*+ *sCount*\\[ *j*\\] - 1. Within each spike train, the spike times are sorted, smallest first.\n\n4\\. `array`: a string describing the MEA used to record the activity. Table\u00a0[1](#T1){ref-type=\"table\"} lists the values used to date.\n\n5\\. `names`: an optional vector of strings of length *N*; *names*\\[*i*\\] stores the name assigned to spike train *i*.\n\n###### \n\nDescriptions of the MEA layouts in the repository, along with their name and number (n) of recordings (Dynamic)\n\n  **Array**            **n**                                           **Description**\n  ------------------- ------- --------------------------------------------------------------------------------------------------\n  MCS_8 \u00d7 8_100um       154    60 electrodes in square lattice (corners missing) with 100 *\u03bc*m spacing (Multi Channel Systems).\n  MCS_8 \u00d7 8_200um       70     60 electrodes in square lattice (corners missing) with 200 *\u03bc*m spacing (Multi Channel Systems).\n  APS_64 \u00d7 64_42um      48               4096 electrodes in square lattice with 42 *\u03bc*m spacing (L Berdondini, IIT).\n  litke_hex_60um        29                512 electrodes on hexagonal lattice with 60 *\u03bc*m spacing (AM Litke, UCSC).\n  EJC1_hex_60um         54             61 electrodes on hexagonal lattice with 60 *\u03bc*m spacing (EJ Chichilnisky, Salk).\n  stanford_hex_60um     11                   61 electrodes on hexagonal lattice with 60 *\u03bc*m spacing (Stanford).\n\nTo help summarize each recording, we have also created a `/summary` group containing information which can be readily computed from the spike trains. These summary points can be read from HDF5 files on their own (rather than reading the entire file) and so provides an efficient cache of this information. The following fields are provided.\n\n1\\. `/summary/N`: the number of spike trains.\n\n2\\. `/summary/duration`: the duration of the recording in seconds, rounded up to the nearest second.\n\n3\\. `/summary/frate`: a vector of length *N*. Element *i* stores the firing rate in Hz of spike train *i*.\n\n4\\. `/summary/totalspikes`: the total number of spike trains in the file.\n\nData sources\n============\n\nTable\u00a0[2](#T2){ref-type=\"table\"} lists the main studies included in the repository, and the number of files in each collection. A key challenge in creating the repository was writing functions to parse the various formats of source data from the different research groups. This has now been done for each of the major formats. When each data set was converted, tests were performed to check that our results matched those presented in the original publications; some of these checks are discussed later. First we describe each of the key studies included in the repository: **Blankenship2011**\u2009This study investigated the impact of knocking out two connexin isoforms (Cx36 or Cx45) upon spontaneous activity \\[[@B11]\\]. Recordings of wild-type, single or double (Cx36 and Cx45) knock-outs were performed at postnatal day 11/12 (P11/12). **Demas2003**\u2009Recordings over an extended period (P9--P42) were made in wild-type mice, along with mice reared in the dark \\[[@B12]\\]. This study also served as a baseline for subsequent recordings in transgenic mice \\[[@B13]\\]. **Demas2006**\u2009This set contains data recorded from *nob*\u2009mutant mouse, where retinal waves persist at late developmental stages \\[[@B13]\\]. **Hennig2011**\u2009A set of eight recordings investigating the effects of chronic bicuculline application at two critical ages when the effect of GABA upon RGCs switches from excitatory to inhibitory \\[[@B14]\\]. **Kirkby2013**\u2009Recent recordings from wild-type and *\u03b2*2 knock-out (KO) mice recorded in the first postnatal week \\[[@B15]\\]. **Maccione2014**\u2009Wild-type recordings in developing mouse retina recorded from P2 to P13 using a high density (4096 electrode) MEA \\[[@B16]\\]. This study presents recordings with high spatial resolution obtained at pan retinal scale. It reports that waves become localized hotspots shortly before eye opening and that cellular recruitment within waves increases significantly during the second postnatal week. **Stacy2005**\u2009A set of further control recordings in wild-type mice are provided \\[[@B17]\\]. In this study, retinal waves were also recorded in transgenic mice where cholinergic neurotransmission was inhibited in most of the retina. However, recordings from the transgenic animal could not be found post-publication. **Stafford2009**\u2009A detailed set of recordings at one age (P6) showing that *\u03b2*2 KO mice still generate correlated waves \\[[@B18]\\]; see also \\[[@B19]\\]. A directional bias in control waves was also reported for the first time. **Sun2008** Two different versions of the *\u03b2*2 KO transgenic mice were studied and, in comparison to earlier studies \\[[@B20]\\], were shown to have correlated activity extending over larger distances than wild type. Although the key paper \\[[@B19]\\] focuses on postnatal days 4 and 5, recordings from a range of days are provided, and were analysed separately \\[[@B21]\\]. **Torborg2004** This key summarises data that appeared in several publications \\[[@B20],[@B22]-[@B24]\\]. These were the first MEA recordings of spontaneous activity in the *\u03b2*2 KO mouse, and showed that although individual RGCs were spontaneously active, the correlations in firing of neighbouring RGCs were strongly reduced. There are also recordings combining the *\u03b2*2 KO line with a gap junction knock-out (CX36), and recordings under different pharmacological conditions. **Wong1993**\u2009These are the first MEA recordings of retinal waves, in ferret at different developmental ages. (Some data from here were presented also in \\[[@B25]\\].) This was the first paper to introduce the key measure of correlated activity, the correlation index, that has been subsequently used in most studies. Although the recordings are relatively short, they highlighted strong distance-dependent correlations gradually decaying with age. (Conversion of these files was complicated as they were binary files, so we converted them through a custom macintosh program written for these data by Markus Meister.) **Xu2011** To further investigate the effect of cholinergic neurotransmission, a transgenic line (where *\u03b2*2-nAChR was expressed in only RGCs of *\u03b2*2 KO mice) was generated \\[[@B26]\\]. In this *\u03b2*2(TG) line, waves were restored, although the spatial extent of correlations was reduced.\n\n###### \n\nKeys and citations of the data sources in the repository\n\n  **Key**            **n**   **Citation**\n  ----------------- ------- --------------\n  Blankenship2011     17      \\[[@B11]\\]\n  Demas2003           37      \\[[@B12]\\]\n  Demas2006           41      \\[[@B13]\\]\n  Hennig2011           8      \\[[@B14]\\]\n  Kirkby2013          22      \\[[@B15]\\]\n  Maccione2014        48      \\[[@B16]\\]\n  Stacy2005            7      \\[[@B17]\\]\n  Stafford2009        29      \\[[@B18]\\]\n  Sun2008             62      \\[[@B19]\\]\n  Torborg2004         54      \\[[@B22]\\]\n  Wong1993            11      \\[[@B29]\\]\n  Xu2011              30      \\[[@B26]\\]\n\nn is the number of files associated with each key (Dynamic).\n\nCiting the data\n---------------\n\nWe are grateful to our colleagues for sharing their recordings. If you use any of these data sets, we request that you acknowledge the relevant authors by citing the corresponding papers (listed in Table\u00a0[2](#T2){ref-type=\"table\"}).\n\nMinimal metadata\n----------------\n\nOur approach to metadata is deliberately minimal, simply describing what we think are some of the essential features of the recordings, such as developmental age, and genotype. This of course means that many details are missing, but in most cases we hope that they can be extracted (manually) from the corresponding publications. We store the metadata as named items in the HDF5 file, under the `/meta/` group. For example, the developmental age of the recording is stored in `/meta/age`. Tables\u00a0[1](#T1){ref-type=\"table\"}, [3](#T3){ref-type=\"table\"}, [4](#T4){ref-type=\"table\"}, [5](#T5){ref-type=\"table\"}, [6](#T6){ref-type=\"table\"} and [7](#T7){ref-type=\"table\"} describe the metadata that are included in the HDF5 files.\n\n###### \n\nMetadata items stored in the repository\n\n  **Name**      **Type**   **Default**                                                                                                                        **Description**                                                                                                                                   **Table**\n  ------------ ---------- ------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ----------------------------------\n  `key`          String        ---                                            A primary key to indicate the study that the file is associated with. This is typically of the form *Wong1993*, i.e. surname of first author and year of main paper \\[[@B29]\\].                                        Table [2](#T2){ref-type=\"table\"}\n  `species`      String        ---           Species name; most recordings are from mouse, but we also have ferret recordings \\[[@B29]\\]. Although ages for both animals are given in postnatal days, ferret and mice have different gestational periods, and so cannot be directly compared.        Table [4](#T4){ref-type=\"table\"}\n  `age`         Integer        ---                 Postnatal age of the animal. For one study \\[[@B29]\\], adult recordings were represented as age 500 days. (There are currently no embryonic recordings, but these could be represented by negative values; zero is the day of birth.)             Table [5](#T5){ref-type=\"table\"}\n  `genotype`     String       \"wt\"                                                                                             Genotype (of the mouse), with the default \"wt\" indicating a wild-type mouse.                                                                                          Table [6](#T6){ref-type=\"table\"}\n  `cond`         String       \"ctl\"      A brief description of the conditions under which the spontaneous activity were made. e.g. some recordings were made in dark reared (\"dr\") mice, and sometimes pharmacological agents were used. If no condition is supplied, control (\"ctl\") is assumed.   Table [7](#T7){ref-type=\"table\"}\n\nValues with no default are compulsory. The final column refers to subsequent tables for more details on each name (Static).\n\n###### \n\nNumbers of recordings for each species (Dynamic)\n\n  **Species**    **n**\n  ------------- -------\n  Ferret          11\n  Mouse           355\n\n###### \n\nNumbers of files from each study (rows) for each postnatal age (columns) (Dynamic)\n\n  \u00a0                   **0**   **1**   **2**   **3**   **4**   **5**   **6**   **7**   **8**   **9**   **10**   **11**   **12**   **13**   **15**   **21**   **30**   **42**   **500**\n  ----------------- ------- ------- ------- ------- ------- ------- ------- ------- ------- ------- -------- -------- -------- -------- -------- -------- -------- -------- ---------\n  Blankenship2011         0       0       0       0       0       0       0       0       0       0        0       17        0        0        0        0        0        0         0\n  Demas2003               0       0       0       0       0       0       0       0       0       2        0        5        0        3        9        8        0       10         0\n  Demas2006               0       0       0       0       0       0       0      10       0       0        0        0       10        0        8       13        0        0         0\n  Hennig2011              0       0       0       0       0       2       2       0       0       2        2        0        0        0        0        0        0        0         0\n  Kirkby2013              0       0       0       0       2       7      10       3       0       0        0        0        0        0        0        0        0        0         0\n  Maccione2014            0       0       1       2       4       4       6       2       4       7        5        7        3        3        0        0        0        0         0\n  Stacy2005               0       0       0       0       0       7       0       0       0       0        0        0        0        0        0        0        0        0         0\n  Stafford2009            0       0       0       0       0       0      29       0       0       0        0        0        0        0        0        0        0        0         0\n  Sun2008                 0       1       3       2      13      17       3       4       4       4        2        6        1        2        0        0        0        0         0\n  Torborg2004             0       0       0       2      12       9       0       1       3       0       16       11        0        0        0        0        0        0         0\n  Wong1993                1       1       0       0       1       1       0       0       0       0        0        0        1        0        2        2        1        0         1\n  Xu2011                  0       0       0       0      30       0       0       0       0       0        0        0        0        0        0        0        0        0         0\n\n###### \n\nNumbers of recordings of each genotype included in the repository\n\n  \u00a0                 **Beta2 KO**   **Beta2**   **Beta2**   **Beta2(TG)**   **Beta2 Xu**   **Cx36 KO**   **Cx36**   **Cx45 KO**   **nob**   **wt**\n  ----------------- -------------- ----------- ----------- --------------- -------------- ------------- ---------- ------------- --------- --------\n  Blankenship2011   0              0           0           0               0              2             6          4             0         5\n  Demas2003         0              0           0           0               0              0             0          0             0         37\n  Demas2006         0              0           0           0               0              0             0          0             25        16\n  Hennig2011        0              0           0           0               0              0             0          0             0         8\n  Kirkby2013        11             0           0           0               0              0             0          0             0         11\n  Maccione2014      0              0           0           0               0              0             0          0             0         48\n  Stacy2005         0              0           0           0               0              0             0          0             0         7\n  Stafford2009      0              0           0           0               0              0             0          0             0         29\n  Sun2008           0              0           21          0               18             0             0          0             0         23\n  Torborg2004       11             7           0           0               0              19            0          0             0         17\n  Wong1993          0              0           0           0               0              0             0          0             0         11\n  Xu2011            0              0           0           17              0              0             0          0             0         13\n\nNote that Wong1993 data are from ferret (Dynamic).\n\n###### \n\nPharmacological and environmental conditions under which spontaneous activity were recorded\n\n  **Cond**                            **n**\n  ---------------------------------- -------\n  10 microM forskolin                   1\n  25 microM APV + 25 microM CNQX;       1\n  wash then 50 microM Mecamylamine      \u00a0\n  25 microM APV + 50 microM CNQX        1\n  2.88 microM FPL                       2\n  2 microM forskolin                    2\n  2 microM strychnine                   1\n  50 microM Carbenoxolone               1\n  bic                                   4\n  ctl                                  332\n  dr                                   19\n  wash                                  2\n\nWhere possible, the descriptions of the conditions follow those provided with the original data (Dynamic).\n\nWe attempted to include metadata regarding spike detection and sorting. However, many different methods for spike detection and sorting have been used in the last 25 years. These vary from manual, semi-automated to fully automated. Furthermore, the level of details included in publications varied significantly. Rather than encode these in the metadata, we instead provide a brief textual summary of the methods in Table\u00a0[8](#T8){ref-type=\"table\"}.\n\n###### \n\nSummary of spike sorting methods used to create spike trains (Static)\n\n  **Key**                  **Spike sorting method**\n  ----------------- ---------------------------------------\n  Blankenship2011            Plexon offline sorter\n  Demas2003                  Plexon offline sorter\n  Demas2006                  Plexon offline sorter\n  Hennig2011                 Wave_clus \\[[@B44]\\]\n  Kirkby2013                 Plexon offline sorter\n  Maccione2014                       None\n  Stacy2005                          None\n  Stafford2009       Mixture of Gaussians model \\[[@B45]\\]\n  Sun2008                    Plexon offline sorter\n  Torborg2004                  Manual clustering\n  Wong1993                     Manual clustering\n  Xu2011                     Plexon offline sorter\n\nAnalyses\n========\n\nWe now provide some examples of reproducible research using the repository. The aim is to summarise the main features of the repository, rather than to provide novel analyses of these data.\n\nR package\n---------\n\nWe have used the R programming environment to develop a package of tools for the analysis of spontaneous activity. R, however, is not required to use these data files. This R package (called `sjemea`) was created in 2001 (to support work subsequently published \\[[@B12]\\]) and is still under development. The package primarily focuses on the batch analysis of data; other open tools more suitable towards interactive analysis are available \\[[@B27],[@B28]\\]. We now give several examples of analysing the repository using code from this R package.\n\nOverview of the repository\n--------------------------\n\nFigure\u00a0[1](#F1){ref-type=\"fig\"} provides an overview of the repository, showing that recordings range from just a few minutes to many hours. In particular, we see that two different groups recorded waves continuously for up to eight hours \\[[@B14],[@B17]\\].\n\n![**Basic features of recordings in the repository.**\u2009Each recording is summarised by its number of spike trains and its duration. The colour of each point indicates which collection the recording comes from. Both axes are plotted on a log scale. We currently have 366 recordings in the repository, occupying 298 MB on disc (Dynamic).](2047-217X-3-3-1){#F1}\n\nAs all recordings were from extracellular electrodes, each electrode can detect activity from multiple neurons. Most, but not all, recordings have been spike-sorted to discriminate activity from multiple on each electrode. We therefore refer to spike trains coming from \"units\" throughout this paper to avoid confusion between multi-unit activity from several neurons and inferred activity from a single neuron. Most recordings contain fewer than 100 spike trains because they were made on MEAs consisting of 60 or 64 electrodes. There are also two clear groups of recordings with around 500--1300 units which were recorded from the two higher-density arrays \\[[@B16],[@B18]\\].\n\nFourplot\n--------\n\nThe \"fourplot\" is our one page summary plot of a recording which we use as an initial screen to check its quality. Figure\u00a0[2](#F2){ref-type=\"fig\"} shows one such example. This plot allows us to quickly evaluate the recording using the features described in the figure legend. The accompanying website has a gallery section showing the fourplot for each datafile in the repository.\n\n![**Ten minutes of spontaneous activity from P1 ferret retina recorded using a MEA.** The name of the data file is given at the top of the plot. **A**: the estimated position of each unit is plotted. Each spike train is given a unique number; overlapping numbers indicate that more than one unit was assigned the same position. **B**: the firing rate estimated in one second bins, averaged across the entire array. Periodic elevations in firing rate, followed by long periods of relative silence, are characteristic of retinal waves. **C**: the raster showing the spike times of all units (starting with unit one at the bottom). **D**: the correlation index plot \\[[@B29]\\], described in detail in Figure\u00a0[3](#F3){ref-type=\"fig\"} (Dynamic).](2047-217X-3-3-2){#F2}\n\nCorrelation indices in neonatal ferret retina\n---------------------------------------------\n\nRetinal waves induce correlations in the firing patterns of neighbouring RGCs. This was first demonstrated in the analysis of ferret retinal waves using the correlation index measure \\[[@B29]\\]. The correlation index measures the degree that two units spike together within some small time window (typically 50 ms). Figure\u00a0[3](#F3){ref-type=\"fig\"} shows the correlation index as a function of the distance separating any given pair of units. This figure almost exactly replicates Figure eight of the prior study \\[[@B29]\\], with the only exception that we also show correlation indices for pairs of units with zero distance separating them.\n\n![**Example correlation index plot.**\u2009For each pair of units we plot the correlation index computed between the two spike trains against the distance on separating the two units. This plot denotes the correlation index curve for the file Wong1993_P0.h5 (ferret P0), and matches the correlation index plot shown in Figure eight of \\[[@B29]\\]. (The y axis is plotted on a logarithmic axis to match the original figure.) The only observable difference is that in our plot we have included correlations for pairs of units that share the same array location (inter-unit distance = 0 *\u03bc*m) (Dynamic).](2047-217X-3-3-3){#F3}\n\nCorrelation indices in wild-type and transgenic mice\n----------------------------------------------------\n\nCholinergic neurotransmission is required for the generation of retinal waves in early development \\[[@B30],[@B31]\\]. One key transgenic line has been global knock-out of the *\u03b2*2 subunit of the nicotonic acetylcholine receptor (nAChR), termed *\u03b2*2 KO here. Initial reports suggested that *\u03b2*2 KO mice lack retinal waves \\[[@B20],[@B32]\\], but subsequent studies reported retinal waves in these mice \\[[@B18],[@B19]\\]. These differences might have occurred because of different recording conditions, notably temperature and bath medium \\[[@B18]\\]. Given the importance of these results, we have collected and curated most of the key recordings published to date that quantify spontaneous activity in *\u03b2*2 KO mice.\n\nThe effects of transgenically rescuing *\u03b2*2-nAChR into just RGCs, leaving it knocked out in the rest of the nervous system, were recently investigated \\[[@B26]\\]. In this genetic manipulation, termed *\u03b2*2(TG), retinal waves were correlated over shorter distances than waves in wild-type mice (Figure one(G) of \\[[@B26]\\]). We have recreated that result by recalculating the correlation indices. Figure\u00a0[4](#F4){ref-type=\"fig\"} has the same key properties as previously reported with the notable exception that the correlation indices in both wild-type and *\u03b2*2(TG) mice are about half the magnitude compared to those originally reported. However, the shape of the two groups, and the overall conclusions, are unaffected. The discrepancy between the two results is likely to be an artifact of the method previously used \\[[@B26]\\], although their code is no longer available to confirm this.\n\n![**Family of correlation index plots to summarise retinal wave data recorded at P4**\\[[@B26]\\]**.** Each line represents the mean correlation index, where the inter-unit distance has been grouped into approximately 100 *\u03bc*m bins, from one recording. The colour of the line denotes whether the recording was taken from a wild-type or *\u03b2*2(TG) mouse (Dynamic).](2047-217X-3-3-4){#F4}\n\nCARMEN application: burst analysis\n----------------------------------\n\nWe have made our data freely available in the CARMEN system \\[[@B33]\\]. The CARMEN Virtual Laboratory is a collaborative online facility for neuroscientists. Data can be uploaded to a repository, and shared with other neuroscientists. Extensive metadata \\[[@B34]\\] can be attached to the data, and a search facility allows data to be located in the repository. The system is currently targeted towards electrophysiology data, and predominantly MEA and electroencephalography (EEG) data.\n\nUseful neuroscience analysis routines can be converted into CARMEN services \\[[@B35]\\]. Service code can be written in a range of programming languages (including Matlab, Python, R, C/C++, Java), and can be easily wrapped into a service using the CARMEN Service Builder tool. Metadata attached to each service provides information for the user and the system. Once a service has been registered with the CARMEN system, users can run them via the portal, using any available data on the repository. The service is executed within the CARMEN system's execution environment, which is a private cloud of heterogeneous servers. The execution environment can support multiple execution server environments; currently these are Windows Server, Centos5 Linux and Scientific Linux 4 platforms. The execution details are hidden from the user.\n\nThe CARMEN portal also deploys a workflow tool within the browser, to allow users to tie services together into a processing pipeline. In order to support the interaction of services, a common data standard for all data types is used; the Neural Data Translation Format (NDF). NDF provides a standard format for neural time series data, segment data, and event data \\[[@B36]\\]. In addition it has a rich metadata header which provides a detailed description of the data contents, and which supports the addition of annotations and other relevant attachments (e.g., visual or audio files). A workflow editor provides a graphical means to construct and edit workflows. A workflow enactment engine allows the workflow to be run over the execution servers. The heterogeneous nature of the service infrastructure means that a workflow can be constructed from services that are written using different programming languages and for differing platforms.\n\nTo demonstrate the virtual laboratory workflows on these data, three services were built in Matlab and compiled into a standalone executable for inclusion into the service framework:\n\n1\\. HDF5 to NDF converter --- This reads in the HDF5 file and converts it into an NDF neural event data file.\n\n2\\. A burst detection service --- This finds bursts independently within each spike train of a recording \\[[@B14]\\]. The service takes input data in the form of an NDF neural event file.\n\n3\\. A graphing service to plot burst durations of multiple input files.\n\nThe CARMEN workflow facility (Figure\u00a0[5](#F5){ref-type=\"fig\"}) chains these services together so that given an input file, it is first converted into NDF and then the burst times are computed. The output from the independent burst analysis services are then compared to generate a plot such as Figure\u00a0[6](#F6){ref-type=\"fig\"}. This figure demonstrates that median burst duration is around 0.1 s and there is good agreement between recordings from different laboratories, albeit with one recording showing a few bursts longer than 2 s.\n\n![**Example of the CARMEN workflow facility.**\u2009Ellipses denote files and boxes denote CARMEN services. Four HDF5 files are independently converted into NDF for burst analysis. The outputs (in this case, the duration of each burst detected on each electrode) are then collated into a report generation service to produce a graphical summary, shown in Figure\u00a0[6](#F6){ref-type=\"fig\"} (Static).](2047-217X-3-3-5){#F5}\n\n![**Burst durations for spontaneous activity recorded from four laboratories at postnatal day 5 or 6.**\u2009This file is the output from the CARMEN workflow illustrated in Figure\u00a0[5](#F5){ref-type=\"fig\"}. One file was selected from four different datasets. Each curve shows the cumulative distribution (denoted CDF) of burst durations. The number of bursts detected from each recording were Stacy2005: 2903 bursts (from 60 units); Stafford2009: 3338 (571); Hennig2011: 16896 (58); Torborg2004: 4 (13) (Static).](2047-217X-3-3-6){#F6}\n\nDiscussion\n==========\n\nRole of the repository\n----------------------\n\nGiven the ongoing debate about whether neuronal activity instructs the development of neuronal circuits \\[[@B37],[@B38]\\], we believe it is important to understand the spatiotemporal properties of waves in different recordings. This repository provides a framework for systematic studies of spontaneous activity, looking for example at the effects of particular mutations, such as the *\u03b2*2 KO. Furthermore, we can now begin to study the variability between laboratories when recording spontaneous activity in the retina, as has already been reported for cortical cultures \\[[@B39]\\]. We hope this repository will also lead to increased data reuse.\n\nCollecting new data\n-------------------\n\nWe hope that the repository will prove useful and we encourage investigators to provide recordings of spontaneous activity. We typically require only spike times (rather than voltage traces) and a description of how to map units to positions on the array. The minimal metadata is also required, typically in a spreadsheet. It is preferable if the data have already been presented in an article, so that we can refer to that article. Unpublished data can also be accepted as long as the investigator is aware that the data are made freely available.\n\nThe current focus of the repository is on collecting spontaneous activity from developing retina. Since spontaneous activity is present in other systems, we also anticipate extending the repository to include data from, for example, cortical and hippocampal cultures \\[[@B40]\\].\n\nGenerating new standards\n------------------------\n\nTo date, there are no established standards for storing spike trains recorded from MEAs. However, one key aim of the datasharing program of the International Neuroinformatics Coordinating Facility (INCF) is to establish such standards. We anticipate that the data provided here can be a useful test case for evaluating any proposed standards \\[[@B41]\\]. Given the relatively simple format in which our data are currently provided, we imagine that changing the data files to accommodate any new standards should be straightforward.\n\nMethods\n=======\n\nAll data reported in this paper have been previously published, see Table\u00a0[2](#T2){ref-type=\"table\"}. Details of the experimental procedures are available in those articles. In all cases, recordings of spike times (rather than voltage traces) were provided. Files were then converted into the common HDF5 framework described in this paper. During this conversion, data were checked where possible with previous reports, for example, descriptions of mean firing rates. Metadata were validated using a separate script. The fourplot (Figure\u00a0[2](#F2){ref-type=\"fig\"}) for each recording was also checked. In a few instances this led to discussions with the original authors, or some data being excluded.\n\nAvailability of supporting data\n===============================\n\nThe HDF5 files are available as a zip file \\[[@B9]\\], and accompanying code is linked to from the project web page \\[[@B8]\\]. This article is an example of a literate programming document. It has been created in R using the `knitr` package \\[[@B42]\\]. Figures and tables in this paper are generated dynamically as the document is compiled. Several R packages are required to run the analysis. Materials are archived in the Gigascience database \\[[@B43]\\], and full details are also given on the \"Code\" section of the accompanying website.\n\nAbbreviations\n=============\n\nCARMEN: Code analysis, repository and modelling for E-Neuroscience; CDF: Cumulative distribution function; CRAN: Comprehensive R archive network; EEG: Electroencephalography; HDF5: Hierarchical data format, version 5; INCF: International neuroinformatics coordinating facility; KO: Knock-Out; MEA: Multielectrode array; NDF: Neural data translation format; Pn: Postnatal day *n*, e.g. P5 for postnatal day 5; RGC: Retinal ganglion cell; TG: Transgenic.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nSJE and ES conceived and designed the project. SJE and JDS collected and pre-processed data from groups. SJE curated and processed data, as well as contributed code. MJ, TJ and MW carried out CARMEN workflow analysis. SJE, TJ, ES and MW wrote the manuscript. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nWe thank all the investigators who have contributed data to the repository. Thanks to Matthias Hennig and Matthew Down for providing a burst analysis algorithm for use in the CARMEN workflow. Andrew Morton provided a Neuroexplorer script for converting data files. This work was supported by grants from EPSRC (EP/E002331/1), BBSRC (BB/H023577/1 and BB/I000984/1) and Wellcome Trust (083205/B/07/Z).\n"}
+{"text": "Introduction\n============\n\nAlthough some prophylactic medications have been proposed to treat chronic migraine (CM) there are still many refractory patients and other treatments are warranted. Peripheral nerve stimulation (PNS) of the occipital nerves is a potentially promising therapy for CM patients.\n\nAim\n===\n\nThe aim of this study is to evaluate the efficacy and tolerability of PNS of the occipital nerves for the treatment of refractory CM.\n\nMethods\n=======\n\nTwenty one patients (8 men, 13 women, average age 52.8\u00b112.2) meeting the IHS criteria for refractory CM were enrolled in this study and implanted with a neurostimulation device near the occipital nerves. The primary endpoint was the reduction in Analogical Visual Scale (AVS). Patient satisfaction, migraine frequency, side effects and reasons for discontinuation were also studied. Significance level was set at P\\<0.05.\n\nResults\n=======\n\nHeadache severity according to the AVS was reduced from 8.9\u00b10.7 before PNS to 3.3\u00b12.8 after treatment initiation. There was also a significant difference in reduction of number of headache days and 80% of the patients were satisfied or very satisfied with the procedure. The most common adverse event was persistent implant site pain and only one patient required to be explanted due to inefficacy.\n\nConclusions\n===========\n\nPNS has been explored as a possible treatment option in selective drug-resistant primary headache disorders and, according to our results, this technique may be effective, safe and well tolerated in treating refractory CM. An increasing experience and a more routine use of these techniques can be forecasted in the near future.\n\nNo conflict of interest.\n"}
+{"text": "Petersson LK, Dey DC, Felton AM, Gardiner ES, L\u00f6f M. Influence of canopy openness, ungulate exclosure, and low\u2010intensity fire for improved oak regeneration in temperate Europe. Ecol Evol. 2020;10:2626--2637. 10.1002/ece3.6092\n\n1. INTRODUCTION {#ece36092-sec-0001}\n===============\n\nPresent forest composition is the result of historical disturbance regimes, anthropogenic land\u2010use, climate, and local site conditions. For centuries, natural disturbance and traditional land\u2010use practices maintained relatively open canopies in many northern temperate forests (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Kirby & Watkins, [2015](#ece36092-bib-0035){ref-type=\"ref\"}; Vera, [2000](#ece36092-bib-0066){ref-type=\"ref\"}). During the 20th century, however, secondary succession following abandonment of past land\u2010use practices or introduction of modern high\u2010production forestry changed forest composition and structure, often creating dense forests (Kirby & Watkins, [2015](#ece36092-bib-0035){ref-type=\"ref\"}; Nowacki & Abrams, [2008](#ece36092-bib-0052){ref-type=\"ref\"}).\n\nOaks (*Quercus* L.) are typical components of northern temperate ecosystems in which they provide multiple ecosystem services (Johnson, Shifley, Rogers, Dey, & Kabrick, [2018](#ece36092-bib-0032){ref-type=\"ref\"}; L\u00f6f et al., [2016](#ece36092-bib-0041){ref-type=\"ref\"}; M\u00f6lder, Meyer, & Nagel, [2019](#ece36092-bib-0047){ref-type=\"ref\"}). Due to their disproportionate importance for biodiversity they are foundational species (Dayton, [1972](#ece36092-bib-0020){ref-type=\"ref\"}; Ellison et al., [2005](#ece36092-bib-0024){ref-type=\"ref\"}), creating crucial habitats for saproxylic invertebrates (Jonsell, Weslien, & Ehnstr\u00f6m, [1998](#ece36092-bib-0033){ref-type=\"ref\"}), lichens, fungi (Ranius, Eliasson, & Johansson, [2008](#ece36092-bib-0056){ref-type=\"ref\"}), and birds (Rodewald & Abrams, [2002](#ece36092-bib-0059){ref-type=\"ref\"}). In many areas, however, a combination of changing climate and human activities has favored other woody species over oaks, and thereby, lessened oak abundance relative to previous centuries (Lindbladh & Foster, [2010](#ece36092-bib-0040){ref-type=\"ref\"}; McEwan, Dyer, & Pederson, [2011](#ece36092-bib-0045){ref-type=\"ref\"}). Furthermore, oak forests face natural regeneration challenges throughout much of their range, which has been linked in part to intensification of forest\u2010use, increasing ungulate populations, efficient fire suppression, and a lack of sustained management in oak habitats of cultural origin (Bergmeier, Petermann, & Schr\u00f6der, [2010](#ece36092-bib-0007){ref-type=\"ref\"}; Crow, [1988](#ece36092-bib-0019){ref-type=\"ref\"}; Dey et al., [2019](#ece36092-bib-0022){ref-type=\"ref\"}; Petersson, Milberg, et al., [2019](#ece36092-bib-0055){ref-type=\"ref\"}; Shaw, [1968](#ece36092-bib-0063){ref-type=\"ref\"}; Watt, [1919](#ece36092-bib-0068){ref-type=\"ref\"}).\n\nThough their large acorns allow for seedling establishment in relatively dark understory conditions, oaks are often considered light demanding as they require ample irradiance for survival and growth once energy reserves of the cotyledons are exhausted (Annigh\u00f6fer, Becksch\u00e4fer, Vor, & Ammer, [2015](#ece36092-bib-0004){ref-type=\"ref\"}; Johnson et al., [2018](#ece36092-bib-0032){ref-type=\"ref\"}). In addition to sufficient light, successful oak regeneration hinges on a variety of other factors that often influence each other, including competing vegetation, ungulate browsing, and site conditions (Annigh\u00f6fer et al., [2015](#ece36092-bib-0004){ref-type=\"ref\"}; Harmer, [2001](#ece36092-bib-0027){ref-type=\"ref\"}; Jensen & L\u00f6f, [2017](#ece36092-bib-0031){ref-type=\"ref\"}; Lorimer, Chapman, & Lambert, [1994](#ece36092-bib-0043){ref-type=\"ref\"}). Though knowledge is lacking regarding the combined influence of different disturbance types on suitable environmental conditions, traditional land\u2010use systems that included grazing regimes, utilization of low\u2010intensity fire, and cutting of shade\u2010tolerant trees, likely promoted oak regeneration (Bobiec et al., [2019](#ece36092-bib-0009){ref-type=\"ref\"}; Bobiec, Reif, & \u00d6llerer, [2018](#ece36092-bib-0010){ref-type=\"ref\"}; Vera, [2000](#ece36092-bib-0066){ref-type=\"ref\"}).\n\nIn eastern North America, failed oak regeneration has been linked in part to changed disturbance regimes such as fire suppression during the last century (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Brose, Schuler, Lear, & Berst, [2001](#ece36092-bib-0014){ref-type=\"ref\"}; Dey et al., [2019](#ece36092-bib-0022){ref-type=\"ref\"}). This led to development of the fire--oak hypothesis, which states that fire has been an integral part of temperate oak ecosystems and that oaks have ecological traits (e.g., high sprouting capacity and thick bark) that make them better adapted to survive and benefit from periodic fires compared to other hardwood species (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Arthur, Alexander, Dey, Schweitzer, & Loftis, [2012](#ece36092-bib-0005){ref-type=\"ref\"}; Brose et al., [2001](#ece36092-bib-0014){ref-type=\"ref\"}; Nowacki & Abrams, [2008](#ece36092-bib-0052){ref-type=\"ref\"}). Thus, prescribed burns along with stand thinning to increase light availability are used extensively in the United States to restore and regenerate oak ecosystems (Brose, Dey, & Waldrop, [2014](#ece36092-bib-0013){ref-type=\"ref\"}; Dey & Hartman, [2005](#ece36092-bib-0021){ref-type=\"ref\"}). Depending on stage of life and fire regimes, oaks can be classified as fire resisters and endurers (Rowe, [1983](#ece36092-bib-0062){ref-type=\"ref\"}). Because of the complexities of fire regime effects on survival and mode of regeneration that complicate classification by Rowe ([1983](#ece36092-bib-0062){ref-type=\"ref\"}), we, hereafter, refer to oaks as being fire adapted.\n\nIn Europe, modern\u2010day forest fires have mainly been associated with the boreal and Mediterranean zones, and studies of fire effects on oak regeneration have almost entirely been conducted in Mediterranean forests with varying results (e.g., Catry, Moreira, Cardillo, & Pausas, [2012](#ece36092-bib-0015){ref-type=\"ref\"}; Monteiro\u2010Henriques & Fernandes, [2018](#ece36092-bib-0048){ref-type=\"ref\"}). However, reconstructions of past forest conditions suggest that fire may have had an important role in the dynamics of certain temperate European forest types as well (Bradshaw & Lindbladh, [2005](#ece36092-bib-0011){ref-type=\"ref\"}; Niklasson, Lindbladh, & Bj\u00f6rkman, [2002](#ece36092-bib-0050){ref-type=\"ref\"}; Niklasson et al., [2010](#ece36092-bib-0051){ref-type=\"ref\"}; Tinner, Conedera, Ammann, & Lotter, [2005](#ece36092-bib-0065){ref-type=\"ref\"}). In western Ukraine, woodland fires may effectively prevent understory development of hazel (*Corylus avellana* L.), hornbeam (*Carpinus betulus* L.), and beech (*Fagus sylvatica* L.), thereby favoring pedunculate oak (*Quercus robur* L.) reproduction (Bobiec et al., [2019](#ece36092-bib-0009){ref-type=\"ref\"}; Ziobro et al., [2016](#ece36092-bib-0070){ref-type=\"ref\"}). Moreover, Ad\u00e1mek, Hadincov\u00e1, and Wild ([2016](#ece36092-bib-0002){ref-type=\"ref\"}) found that survival rates of sessile oak \\[*Quercus petraea* (Matt.) Liebl.\\] trees following fire in central Europe was notably high and almost unaffected by fire intensity. A recent greenhouse experiment revealed that seedlings of *Q.\u00a0robur* responded similarly to the fire adapted North American white oak (*Q.\u00a0alba* L.) when they experienced shoot destruction typically sustained during fire (Petersson, L\u00f6f, Jensen, Chastain, & Gardiner, [2019](#ece36092-bib-0054){ref-type=\"ref\"}). Together, these findings suggest that European temperate oak species (*Q.\u00a0robur* and *Q.\u00a0petraea)* may respond to fire in a similar fashion as their North American congeners.\n\nAgainst this background, we investigated how the combined effects of different disturbances influence natural oak regeneration, and if the North American fire--oak hypothesis could be applicable during the regeneration phase of temperate European oak forests. A replicated field experiment using naturally regenerated oaks and a randomized block design with split--split plots including contrasting canopy openness, protection against browsing ungulates, and low\u2010intensity surface fire was established in southern Sweden. We hypothesized that: (a) greater light availability would increase tolerance of natural oak reproduction to disturbance; (b) protection against browsing would improve oak growth and survival; and, (c) oaks would survive a low\u2010intensity surface fire by producing new sprouts. Further, we evaluated the response of competing woody vegetation following the low\u2010intensity fire.\n\n2. MATERIAL AND METHODS {#ece36092-sec-0002}\n=======================\n\n2.1. Study sites and species {#ece36092-sec-0003}\n----------------------------\n\nThe experiment was conducted in five oak\u2010dominated forests, 45--95\u00a0m above sea level, in southern Sweden (Figure [1](#ece36092-fig-0001){ref-type=\"fig\"}; Table [1](#ece36092-tbl-0001){ref-type=\"table\"}). Southern Sweden consists of a mosaic landscape of forests, farmlands, and lakes. Due to active forest management favoring conifers, the dominant tree species in the region are Norway spruce (*Picea abies* L. Karst) and Scots pine (*Pinus sylvestris* L.). Two ecologically and morphologically similar oak species are native to the region, *Q.\u00a0robur* and *Q.\u00a0petraea*. The species naturally hybridize and are sometimes taxonomically treated as two subspecies under *Q.\u00a0robur* (Roloff, B\u00e4rtels, & Schulz, [2008](#ece36092-bib-0060){ref-type=\"ref\"}). In this study, they were treated as one species. At the start of the experiment, all study sites had naturally regenerated oak seedlings and saplings in the understory (Table [S1](#ece36092-sup-0001){ref-type=\"supplementary-material\"}).\n\n![Experimental design including location of the five sites (blocks) in southern Sweden: (1) Abbetorp; (2) Barnebo; (3) Horns\u00f6; (4) S\u00f6sdala; and, (5) Sperlingsholm. The enlarged figure shows one block with eight split--split plots, that is, treatment combinations: closed canopy (C), closed canopy and fence (CF), closed canopy, fence, and burn (CFB), closed canopy and burn (CB), canopy gap (G), canopy gap and fence (GF), canopy gap, fence, and burn (GFB) and canopy gap and burn (GB)](ECE3-10-2626-g001){#ece36092-fig-0001}\n\n###### \n\nCharacteristics of the five study sites\n\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n| Sites             | Species composition                                        | Height[a](#ece36092-note-0002){ref-type=\"fn\"} (m) | Basal area[a](#ece36092-note-0002){ref-type=\"fn\"} (m^2^/ha) | PACL (%) | Moose density[b](#ece36092-note-0003){ref-type=\"fn\"} (animals/km^2^) | Deer density[b](#ece36092-note-0003){ref-type=\"fn\"} (animals/km^2^) |          |           |\n+===================+============================================================+===================================================+=============================================================+==========+======================================================================+=====================================================================+==========+===========+\n| 1\\. Abbetorp      | *Quercus robur/petraea*,*Corylus avellana*,*Tilia cordata* | 20                                                | 22                                                          | 5        | 22\u00a0\u00b1\u00a01                                                               | 35\u00a0\u00b1\u00a02                                                              | 0.5--1.2 | 1.0--5.3  |\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n| 2\\. Barnebo       | *Q.\u00a0robur/petraea*,*Picea abies*,*Pinus sylvestris*        | 21                                                | 27                                                          | 16       | 23\u00a0\u00b1\u00a01                                                               | 39\u00a0\u00b1\u00a01                                                              | 0.4--1.5 | 3.2--4.2  |\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n| 3\\. Horns\u00f6        | *Q.\u00a0robur/petraea*,*Acer platanoides*,*P.\u00a0abies*           | 17                                                | 29                                                          | 21       | 20\u00a0\u00b1\u00a01                                                               | 38\u00a0\u00b1\u00a02                                                              | 0.5--0.7 | 5.2--8.3  |\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n| 4\\. S\u00f6sdala       | *Q.\u00a0robur/petraea*,*Sorbus aucuparia*,*Betula pendula*     | 19                                                | 28                                                          | 6        | 19\u00a0\u00b1\u00a01                                                               | 39\u00a0\u00b1\u00a02                                                              | 0.0--0.9 | 5.5--15.0 |\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n| 5\\. Sperlingsholm | *Q.\u00a0robur/petraea*,                                        | 21                                                | 24                                                          | 9        | 19\u00a0\u00b1\u00a01                                                               | 45\u00a0\u00b1\u00a01                                                              | 0.0--0.1 | 2.5--4.5  |\n|                   |                                                            |                                                   |                                                             |          |                                                                      |                                                                     |          |           |\n|                   | *P.\u00a0sylvestris*,*P.\u00a0abies*                                 |                                                   |                                                             |          |                                                                      |                                                                     |          |           |\n+-------------------+------------------------------------------------------------+---------------------------------------------------+-------------------------------------------------------------+----------+----------------------------------------------------------------------+---------------------------------------------------------------------+----------+-----------+\n\nThe three most frequent tree species (listed with respect to their abundance), mean height and basal area for all trees \\>10\u00a0cm dbh in August 2018; percentage of the above canopy light (PACL) in June averaged across the three study years; densities of moose and three deer species during the winter of 2016/2017 and 2017/2018.\n\nMeasurements conducted in a circular plot with a 10\u00a0m radius.\n\nDensities are estimated based on pellet counts, see Appendix [S1](#ece36092-sup-0001){ref-type=\"supplementary-material\"}.\n\nJohn Wiley & Sons, Ltd\n\nFour browsing ungulate species of the family Cervidae occur in southern Sweden: moose (*Alces alces* L.), roe deer (*Capreolus capreolus* L.), red deer (*Cervus elaphus* L.), and fallow deer (*Dama dama* L.). Moose and roe deer are common and likely responsible for most browsing damage in this study. Also, two species of hare (*Lepus timidus* L. and *L.\u00a0europaeus* Pallas) may cause browsing damage.\n\n2.2. Experimental design and treatments {#ece36092-sec-0004}\n---------------------------------------\n\nWe used a randomized block design with split--split plots and five blocks, that is, sites (Figure [1](#ece36092-fig-0001){ref-type=\"fig\"}). The main treatment was canopy manipulation to create different light levels (closed canopy or canopy gap), with protection against browsing (no fence or fence) nested within the canopy treatment, and a low\u2010intensity surface fire (no burn or burn) nested within the fence treatment. This created eight treatment combinations: closed canopy (C), closed canopy and fence (CF), closed canopy, fence, and burn (CFB), closed canopy and burn (CB), canopy gap (G), canopy gap and fence (GF), canopy gap, fence, and burn (GFB), and canopy gap and burn (GB). The size of a split--split treatment plot was 25\u00a0m^2^, with a distance of 25--188\u00a0m and 7--42\u00a0m between canopy\u2010 and fence\u2010treatments, respectively.\n\nIn April 2016, cutting to create canopy gaps of about 400\u00a0m^2^ was performed at each site, and a 2\u2010m steel wire fence was erected around two adjacent, randomly selected plots in each canopy treatment. Cutting was conducted with a chainsaw to remove all large canopy trees in the canopy gaps, but saplings, seedlings, and shrubs were not cut. The fence excluded all ungulates and hares, but provided free access to rodents (mesh size 5\u00a0\u00d7\u00a05 cm from 0--0.8\u00a0m, 16\u00a0\u00d7\u00a020\u00a0cm from 0.8--2.0\u00a0m). Two plots in each canopy treatment were thereafter randomly selected for the burn treatment, one inside and one outside the fence. All oak recruits, defined as seedlings, saplings, or sprouts, within browsing height\u00a0\u2264\u00a0300\u00a0cm tall (Nichols, Cromsigt, & Spong, [2015](#ece36092-bib-0049){ref-type=\"ref\"}) in treatment plots were marked with a uniquely numbered aluminum tag. In plots with\u00a0\\>\u00a0100 oak recruits, a random subset of at least 50 were chosen for measurements. In total, 2,357 oak recruits were measured through the duration of the experiment (Table [S1](#ece36092-sup-0001){ref-type=\"supplementary-material\"}).\n\nBurning was performed between 29 September and 7 October 2016, following an unusually warm and dry late summer and autumn (Table [S2](#ece36092-sup-0001){ref-type=\"supplementary-material\"}). Delaying the burn to these dates allowed understory vegetation time to acclimate to canopy manipulation and fencing for one growing season prior to burning. To ensure similar burn treatment across sites and plots, we used a propane fired blowtorch (121960L, Kemper) to simulate a low\u2010intensity surface fire. Plots were systematically burned from one side to the other so that herbaceous vegetation and the top of the fine litter layer was burned (Table [S3](#ece36092-sup-0001){ref-type=\"supplementary-material\"}), but larger pieces of litter and shrubs were charred. Flame heights were visually estimated to be \\<0.4\u00a0m at all sites. It took about one hour to treat one plot and all four plots per site were treated during the same day. Environmental conditions and litter depth were recorded before and after each burn (Table [S3](#ece36092-sup-0001){ref-type=\"supplementary-material\"}).\n\n2.3. Measurements {#ece36092-sec-0005}\n-----------------\n\nHeight (\u00b11\u00a0cm) and basal diameter (\u00b11\u00a0mm) of all oak recruits were recorded at the start of the experiment in April 2016 (Table [S1](#ece36092-sup-0001){ref-type=\"supplementary-material\"}), and in late August 2016, 2017, and 2018. There were no significant difference in height nor basal diameter between treatments at the start of the experiment (Table [S4](#ece36092-sup-0001){ref-type=\"supplementary-material\"}). In each measurement period, we noted if an oak was alive and the number of stems per individual. For oaks with multiple stems, height and basal diameter were recorded for the tallest stem.\n\nBrowsing damage, that is, herbivore removal of twigs, shoots, or buds, was recorded in each year in both April and August 2016--2018. In April 2017, that is, the spring following burning, browsing damage could not be recorded in burn plots as most oaks were top\u2010killed by the fire and, therefore, had no shoots available for browsers. When possible, the herbivore (hare or ungulate) responsible for damage was determined following Kullberg and Bergstr\u00f6m ([2001](#ece36092-bib-0036){ref-type=\"ref\"}). We defined browsing frequency as the proportion of browsed oaks per treatment plot. Fences remained intact throughout the experiment, and browsing damage did not occur inside fences. Densities of moose and the three deer species were estimated using pellet counts (Table [1](#ece36092-tbl-0001){ref-type=\"table\"}, Appendix [S1](#ece36092-sup-0001){ref-type=\"supplementary-material\"}) (Eberhardt & Van Etten, [1956](#ece36092-bib-0023){ref-type=\"ref\"}; M\u00e5nsson, Andr\u00e9n, & Sand, [2011](#ece36092-bib-0044){ref-type=\"ref\"}).\n\nIn August of each year in 2016--2018, competing woody vegetation \u2264300\u00a0cm tall was recorded in four circular 2\u00a0m^2^ subplots randomly placed in each treatment plot. Height (\u00b11\u00a0cm) and number of individuals per species were recorded, as well as the number of oaks per subplot. These measurements included seedlings that established during the course of the experiment. As oaks dominated the understory at most sites, woody vegetation was combined into species groups: conifers, broadleaves excluding oaks, and oaks.\n\nIn late June for each year of 2016--2018, light availability at 160\u00a0cm above ground level was estimated using hemispherical imagery (Table [1](#ece36092-tbl-0001){ref-type=\"table\"}). One photograph was taken in the middle of each treatment plot on an overcast day (Nikon Coolpix 8800VR, fisheye lens LC\u2010ER2). The camera lens was oriented perpendicular to the forest floor, and magnetic north was referenced in the image. Images were thresholded in the blue color plane, and percentage of above canopy light reaching the camera was calculated using Gap Light Analyzer software (Frazer, Canham, & Lertzman, [1999](#ece36092-bib-0025){ref-type=\"ref\"}).\n\n2.4. Calculations and statistical analysis {#ece36092-sec-0006}\n------------------------------------------\n\nTo account for height variation in oak recruits at the start of the experiment and any size\u2010related differences in growth rates, we calculated relative height growth rate per year (RGR~H~) for each oak following Hunt ([1982](#ece36092-bib-0029){ref-type=\"ref\"}):$$\\text{RGR}_{H} = \\frac{\\ln{(H_{2})} - \\ln{(H_{1})}}{t_{2} - t_{1}}$$where *H* ~2~ and *H* ~1~ were oak height the year of interest and the previous year, respectively, *t* ~2~ and *t* ~1~ were the year of interest and the previous year (in this work *t* ~2\u2212~ *t* ~1~ was always one year). We then calculated average RGR~H~ per treatment plot. Data from the last year of the experiment, that is, August 2018, were used in the statistical analyses. We analyzed RGR~H~ using a linear mixed\u2010effects model with the package \"lme4\" (Bates, Maechler, Bolker, & Walker, [2015](#ece36092-bib-0006){ref-type=\"ref\"}), with the following factors (treatments) and their interactions as fixed effects: canopy openness, fence, and burn (all binary). To account for the hierarchical design of the experiment, we included nested (site/light/fence) random effects in the model. Residual and random effect distributions were examined graphically. A possible outlier was identified, however, when excluded it did not affect model results.\n\nSurvival was analyzed using a generalized linear mixed\u2010effects model with binomial error distribution, using a similar model structure (i.e., fixed and random effects) as described for the RGR~H~ model. To investigate whether initial plant size affected survival after burning, an additional survival model using the fixed effects canopy openness, fence (both binary), and initial basal diameter (continuous), and only including the burn treatment, was analyzed. We tested for overdispersion and residual distribution using \"testDispersion\" and \"testUniformity\" functions from the DHARMa package (Hartig, [2019](#ece36092-bib-0028){ref-type=\"ref\"}).\n\nWe analyzed the change of competing woody vegetation, that is, number of plants, using a generalized linear mixed\u2010effects model with Poisson error distribution, using the following factors as fixed effects: canopy openness, fence, burn (all binary), and species group (three levels). We included nested (site/light/fence/burn) random effects and the initial number of plants per species group as an offset variable. A postanalysis of deviance (Wald *\u03c7* ^2^ Type III) revealed no significance of light or fence as predictor variables, hence we simplified the model using only burn and species group as predictors.\u00a0This was supported by Akaike\\'s Information Criterion (AIC). Overdispersion and residual distribution were tested as described above. Relative density of oaks, that is, number of oaks divided by total number of plants, and height of competing vegetation (log\u2010transformed) were analyzed using linear mixed\u2010effects models as described for the RGR~H~ model.\n\nAs ungulate browsing pressure may vary over time, due to, for example, fluctuating ungulate population densities and season, data from the three measurement occasions after burning (August 2017, April 2018, August 2018) were modeled separately to analyze browsing frequency on oak. We used generalized linear mixed\u2010effects models with binomial error distribution, as described for the survival model, only including the treatment without fences and only ungulate browsing (not hare).\n\nAll analyses were performed using R version 3.5.0 (R Core Team, [2017](#ece36092-bib-0018){ref-type=\"ref\"}). For all models, AIC was used to identify whether interactions between fixed effects were justified. Significance was determined at an alpha level of 0.05.\n\n3. RESULTS {#ece36092-sec-0007}\n==========\n\nFor oak survival and growth, as well as the development of competing woody vegetation, the statistical analyses presented here are based on data from the last year of the experiment (August 2018).\n\n3.1. Survival {#ece36092-sec-0008}\n-------------\n\nSurvival of naturally regenerated oak recruits declined in all treatment combinations over the course of the experiment, ranging between 54% (CB treatment) and 92% (GF treatment) at the end of the experiment (Figure [2](#ece36092-fig-0002){ref-type=\"fig\"}a). No interactions between effects were identified through AIC testing. Though not significant at *\u03b1*\u00a0\\<\u00a00.05, canopy openness (*\u03c7* ^2^\u00a0=\u00a02.83, *df*\u00a0=\u00a01, *p*\u00a0=\u00a0.092) and protection against browsing (*\u03c7* ^2^\u00a0=\u00a03.57, *df*\u00a0=\u00a01, *p*\u00a0=\u00a0.059) tended to increase survival. Burning, however, had a significant negative effect on survival (*\u03c7* ^2^\u00a0=\u00a0141.43, *df*\u00a0=\u00a01, *p*\u00a0\\<\u00a0.001). The majority of oak recruits affected by fire were top\u2010killed and survived by producing new sprouts the following year; 88%--96% of surviving oaks affected by burning were resprouts at the end of the experiment. Furthermore, analysis of survival among oaks receiving the burn treatment revealed greater survival for recruits with larger initial stem basal diameter (*\u03c7* ^2^\u00a0=\u00a040.02, *df*\u00a0=\u00a01, *p*\u00a0\\<\u00a0.001), and for recruits growing in a canopy gap (*\u03c7* ^2^\u00a0=\u00a04.56, *df*\u00a0=\u00a01, *p*\u00a0=\u00a0.033), while protection against browsing had no significant effect (*\u03c7* ^2^\u00a0=\u00a02.51, *df*\u00a0=\u00a01, *p*\u00a0=\u00a0.114).\n\n![Development of naturally regenerated oaks as (a) survival, (b) relative height growth (RGR~H~), and (c) absolute height under eight treatment combinations \\[closed canopy (C), closed canopy and fence (CF), closed canopy, fence, and burn (CFB), closed canopy and burn (CB), canopy gap (G), canopy gap and fence (GF), canopy gap, fence, and burn (GFB), and canopy gap and burn (GB)\\]. Time of canopy gap creation and fencing is indicated with a gray arrow, and time of burn with a black arrow](ECE3-10-2626-g002){#ece36092-fig-0002}\n\n3.2. Growth {#ece36092-sec-0009}\n-----------\n\nAs most oaks recruits that survived burning sprouted after top\u2010kill, their relatively small shoots resulted in a strongly negative RGR~H~ one year after burning (Figure [2](#ece36092-fig-0002){ref-type=\"fig\"}b). On average, oaks that sprouted produced 2.6\u00a0\u00b1\u00a00.1 shoots. RGR~H~ was greatest under canopy gaps (*F* ~1,4~\u00a0=\u00a010.10, *p*\u00a0=\u00a0.034), and, though not statistically significant, there was a trend that protection from browsing had a positive effect (*F* ~1,9~\u00a0=\u00a04.71, *p*\u00a0=\u00a0.058). Furthermore, RGR~H~ was greater for oaks receiving the burn treatment than those that did not receive burning (*F* ~1,19~\u00a0=\u00a013.65, *p*\u00a0=\u00a0.002). AIC testing supported no interactions between treatments. Despite their greater RGR~H~, height of oak recruits subjected to burning remained lower through the experiment than those that were not burned (Figure [2](#ece36092-fig-0002){ref-type=\"fig\"}c). By the end of the study, oaks had not grown above browsing height regardless of treatment.\n\n3.3. Competing woody vegetation {#ece36092-sec-0010}\n-------------------------------\n\nCompeting woody vegetation (number) was reduced by the burn treatment (*\u03c7* ^2^\u00a0=\u00a021.98, *df*\u00a0=\u00a01, *p*\u00a0\\<\u00a0.001). Moreover, differences were observed between species groups (*\u03c7* ^2^\u00a0=\u00a032.00, *df*\u00a0=\u00a02, *p*\u00a0\\<\u00a0.001). Burning reduced conifers from 750 to 125\u00a0plants/ha after two years, broadleaves (excluding oaks) from 9,812 to 5,500 plants/ha, and oaks from 51,652 to 38,250 plants/ha (Table [S5](#ece36092-sup-0001){ref-type=\"supplementary-material\"}). However, the relative density of oaks was not affected by treatment factors (Table [S6](#ece36092-sup-0001){ref-type=\"supplementary-material\"}).\n\nHeight of competing vegetation (broadleaves and conifers combined) was not affected by canopy openness (*F* ~1,4~\u00a0=\u00a00.20, *p*\u00a0=\u00a0.681; Table [2](#ece36092-tbl-0002){ref-type=\"table\"}). There was a trend, though not significant at *\u03b1*\u00a0\\<\u00a00.05, suggesting protection against browsing increased competitor height (*F* ~1,9~\u00a0=\u00a04.71, *p*\u00a0=\u00a0.058). Concurrently, competitor height was lowest where plots were burned (*F* ~1,19~\u00a0=\u00a023.36, *p*\u00a0\\<\u00a0.001).\n\n###### \n\nHeight of competing woody vegetation in August 2016 and 2018 (mean\u00a0\u00b1\u00a0*SE*, *n*\u00a0=\u00a05). Heights in 2018 also include individuals that were established after measurements in 2016\n\n  Treatment                              Height (cm)   \n  -------------------------------------- ------------- ----------\n  Closed canopy (C)                      95\u00a0\u00b1\u00a022       105\u00a0\u00b1\u00a026\n  Closed canopy and fence (CF)           51\u00a0\u00b1\u00a012       71\u00a0\u00b1\u00a013\n  Closed canopy, fence, and burn (CFB)   76\u00a0\u00b1\u00a017       45\u00a0\u00b1\u00a03\n  Closed canopy and burn (CB)            94\u00a0\u00b1\u00a019       37\u00a0\u00b1\u00a05\n  Canopy gap (G)                         40\u00a0\u00b1\u00a04        55\u00a0\u00b1\u00a09\n  Canopy gap and fence (GF)              73\u00a0\u00b1\u00a012       117\u00a0\u00b1\u00a010\n  Canopy gap, fence, and burn (GFB)      71\u00a0\u00b1\u00a015       60\u00a0\u00b1\u00a07\n  Canopy gap and burn (GB)               58\u00a0\u00b1\u00a019       39\u00a0\u00b1\u00a05\n\nJohn Wiley & Sons, Ltd\n\n3.4. Browsing frequency {#ece36092-sec-0011}\n-----------------------\n\nThe majority of browsing damage was caused by ungulates; hares were responsible for\u00a0\\<\u00a05% of all observed browsing damage. Overall, ungulate browsing frequency on oak recruits was greater during winter than summer (Figure [3](#ece36092-fig-0003){ref-type=\"fig\"}). Burning, which was applied in the autumn of 2016, temporarily influenced the pattern of browsing frequency. In August 2017 (one year after burning), browsing frequency was greatest in the burned plots (Figure [3](#ece36092-fig-0003){ref-type=\"fig\"}, Table [3](#ece36092-tbl-0003){ref-type=\"table\"}), presumably a response to the new sprouts that emerged in the summer of 2017. Thereafter, browsing frequency was greatest in nonburned plots (Figure [3](#ece36092-fig-0003){ref-type=\"fig\"}, Table [3](#ece36092-tbl-0003){ref-type=\"table\"}). Except for the winter 2017/2018, browsing frequency was unaffected by canopy openness. In that winter, an interaction between the canopy and burn treatments suggested that browsing frequency was lowest in the CB treatment.\n\n![Ungulate browsing frequency (proportion of browsed oaks per plot) on naturally regenerated oaks in four treatment combinations \\[closed canopy (C), closed canopy and burn (CB), canopy gap (G), and canopy gap and burn (GB)\\]. Measurements in April correspond to preceding winter browsing, measurements in August correspond to preceding summer browsing. Time of canopy gap creation is indicated with a gray arrow, and time of burn with a black arrow](ECE3-10-2626-g003){#ece36092-fig-0003}\n\n###### \n\nAnalysis of deviance table based on a generalized linear mixed\u2010effects model (binomial) explaining ungulate browsing frequency on naturally regenerated oak recruits among treatments and their interactions (excluding fence treatment). Browsing frequency measured in August corresponds to preceding summer browsing, measurements in April corresponds to preceding winter browsing\n\n  Factor                  *\u03c7* ^2^   *df*   *p*\n  ----------------------- --------- ------ --------\n  Aug 2017                                 \n  Canopy openness         0.97      1      .325\n  Burn                    37.02     1      \\<.001\n  Canopy openness: burn   3.19      1      .074\n  Apr 2018                                 \n  Canopy openness         8.07      1      .005\n  Burn                    30.14     1      \\<.001\n  Canopy openness: burn   10.32     1      .001\n  Aug 2018                                 \n  Canopy openness         3.67      1      .056\n  Burn                    5.63      1      .018\n  Canopy openness: burn   3.58      1      .058\n\nType III Wald *\u03c7* ^2^ tests.\n\nJohn Wiley & Sons, Ltd\n\n4. DISCUSSION {#ece36092-sec-0012}\n=============\n\nWe investigated if the fire--oak hypothesis (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Arthur et al., [2012](#ece36092-bib-0005){ref-type=\"ref\"}) is applicable to European temperate oaks by establishing a field experiment with contrasting canopy openness, protection against wild ungulate browsers, and a low\u2010intensity fire. The experimental design allowed us to examine potential combined effects of these disturbance\u2010related factors. At the end of the experiment, oak survival was high across treatment combinations, and relative height growth rate (RGR~H~) was greatest where burning was applied. Furthermore, the relatively higher light availability within canopy gaps increased RGR~H~ and oak survival where fire was prescribed. Burning also clearly effected competing woody vegetation by reducing the number of individuals, though it did not change relative density of oaks. These results, which are similar to reports from North American studies (e.g., Brose, Dey, Phillips, & Waldrop, [2013](#ece36092-bib-0012){ref-type=\"ref\"}; McEwan et al., [2011](#ece36092-bib-0045){ref-type=\"ref\"}), could indicate a common ecological mechanism impactful to the dynamics of oak regeneration in temperate forests across the Atlantic.\n\nThough fire reduced survival of naturally regenerated oaks, survival rates were still relatively high across treatments. This high survival was largely dependent on new shoot production following top\u2010kill. Top\u2010kill initially decreased RGR~H~, but by the end of the experiment sprouting resulted in a vigorous RGR~H~. One reason many North American oaks are considered fire adapted is the relatively greater sprouting capacity they exhibit over their competitors (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Brose et al., [2014](#ece36092-bib-0013){ref-type=\"ref\"}). The high sprouting capacity and increased growth rate of *Q.\u00a0robur/petraea* observed in this field experiment supports the hypothesis that natural regeneration of temperate European oaks could be promoted by low\u2010intensity fire. However, long\u2010term observations are needed to determine if the positive fire effects noted through this research will persist over time.\n\nAs expected, oak RGR~H~ was greater in canopy gaps than under closed canopies. However, oak survival was not significantly affected by canopy openness, except where burning was applied. *Q.\u00a0robur* and *Q.\u00a0petraea* require a minimum of 15%--20% of full light for sustained growth (L\u00f6f, Karlsson, Sonesson, Welander, & Collet, [2007](#ece36092-bib-0042){ref-type=\"ref\"}; von L\u00fcpke, [1998](#ece36092-bib-0067){ref-type=\"ref\"}). This corresponds to the available light beneath closed canopies in this study and can probably explain the high survival rates of the oaks growing there. Nevertheless, the observation that survival after prescribed burning was improved in canopy gaps suggests that oak resilience to disturbance (which is conferred by sprouting) is enhanced by light. North American studies have shown that prescribed burning offers the greatest to oak regeneration when conducted in combination with overstory treatments that increase light availability (e.g., Brose et al., [2013](#ece36092-bib-0012){ref-type=\"ref\"}; Hutchinson, Long, Rebbeck, Sutherland, & Yaussy, [2012](#ece36092-bib-0030){ref-type=\"ref\"}). Improving light availability the year prior to burning in this experiment likely favored development of belowground carbohydrate reserves by oaks, which supported sprouting and rapid growth following the prescribed fire (Kabeya & Sakai, [2005](#ece36092-bib-0034){ref-type=\"ref\"}; Welander & Ottosson, [1998](#ece36092-bib-0069){ref-type=\"ref\"}). Furthermore, the observation of survival being greatest for oaks with the largest stem diameters is consistent with previous research (Dey & Hartman, [2005](#ece36092-bib-0021){ref-type=\"ref\"}).\n\nThough oaks are preferentially browsed by ungulates (Bergqvist, Wallgren, Jernelid, & Bergstr\u00f6m, [2018](#ece36092-bib-0008){ref-type=\"ref\"}), they are considered browsing tolerant because they can survive moderate browsing for extended periods (Harmer, [2001](#ece36092-bib-0027){ref-type=\"ref\"}). However, browsing can severely limit height growth and high browsing pressure can prevent oak and other palatable tree species from advancing to the overstory (Churski, Bubnicki, Jedrzejewska, Kuijper, & Cromsigt, [2016](#ece36092-bib-0016){ref-type=\"ref\"}; Rooney & Waller, [2003](#ece36092-bib-0061){ref-type=\"ref\"}). We observed a trend indicating that protection from wild ungulate browsers positively affected oak recruit survival and RGR~H~. Considering the relatively short duration of our experiment, it was not surprising that protection against browsing had a somewhat limited impact. Additionally, there was considerable variation in browsing animal density between study sites and years, so it is possible that this variation masked positive effects of protection against browsers in this study.\n\nAccording to the fire--oak hypothesis, many oaks have traits that confer a competitive advantage over other tree species following periodic fires (Abrams, [1992](#ece36092-bib-0001){ref-type=\"ref\"}; Arthur et al., [2012](#ece36092-bib-0005){ref-type=\"ref\"}). Testing the fire--oak hypothesis on temperate European oaks should therefore also include an assessment of fire effects on competing vegetation. In this study, woody vegetation was assessed as species groups since oak seedlings and saplings dominated the understory on most sites. This limits our ability to distinguish species\u2010specific responses of oak competitors. Furthermore, it was not possible to distinguish sprouts from newly established seedlings. Nevertheless, we observed clear fire effects on competing woody vegetation as prescribed burning reduced the number of individuals, especially the number of conifers (*P.\u00a0abies* and *P.\u00a0sylvestris*). This was expected because these conifers lack the ability to sprout when top\u2010killed (Leonardsson & G\u00f6tmark, [2015](#ece36092-bib-0038){ref-type=\"ref\"}). It is, however, worth noting that there were fewer conifers present on the study sites as compared to broadleaves (Table [S5](#ece36092-sup-0001){ref-type=\"supplementary-material\"}). The height of competing vegetation was also lowered in burned plots, but competing vegetation remained roughly the same height or slightly taller than oak reproduction at the end of the experiment.\n\nOur review of literature from the eastern United States indicated that a single, dormant season prescribed fire is often not sufficient to favor oak regeneration (McEwan et al., [2011](#ece36092-bib-0045){ref-type=\"ref\"}). Rather multiple fires have been more successful (Dey & Hartman, [2005](#ece36092-bib-0021){ref-type=\"ref\"}; Hutchinson et al., [2012](#ece36092-bib-0030){ref-type=\"ref\"}). This corresponds to Ziobro et al. ([2016](#ece36092-bib-0070){ref-type=\"ref\"}) and Bobiec et al. ([2019](#ece36092-bib-0009){ref-type=\"ref\"}), who reported that reoccurring grass burning promoted *Q.\u00a0robur* regeneration by reducing competing shade\u2010tolerant species in Ukrainian woodlands. We, therefore, suggest that additional experiments involving multiple burning events are needed to determine whether reoccurring fire provides *Q.\u00a0robur* and *Q.\u00a0petraea* competitive advantage. The relatively small burn treatments in this study allowed us to experimentally examine potential combined effects of multiple disturbance\u2010related factors. Nevertheless, natural wildfires or prescribed burns generally impact larger areas, and the varying fire intensity of fire over larger areas can often result in pockets of unburnt vegetation (e.g., Lampainen, Kuuluvainen, Wallenius, Karjalainen, & Vanha\u2010Majamaa, [2004](#ece36092-bib-0037){ref-type=\"ref\"}). Future studies investigating the effects of larger fire events on oak regeneration would, therefore, be of interest. Furthermore, the timing of a fire is known to affect how vegetation responds after fire. North American studies have demonstrated that growing season fires have greater impact than do dormant season fires because vegetation is physiologically active (Brose et al., [2014](#ece36092-bib-0013){ref-type=\"ref\"}). Though burning was conducted at the very end of the growing season in this study, it is likely that the timing of the burns we applied impacted plants that were physiologically active.\n\nFire changed the pattern of ungulate browsing frequency. While the peak in browsing frequency was during the winter, summer browsing was also prominent where plots were recently burned. A similar observation has been reported for the United States (e.g., Andruk, Schwope, & Fowler, [2014](#ece36092-bib-0003){ref-type=\"ref\"}; Collins & Carson, [2003](#ece36092-bib-0017){ref-type=\"ref\"}). Considering the relatively small treatment plots of this study, it seems unlikely the increased browsing frequency was caused by a decrease in availability of alternative forage. Rather it may have resulted from changed foliar nutrient concentrations and/or foliar defensive chemistry that increased palatability of postfire--oak sprouts (Reich, Abrams, Ellsworth, Kruger, & Tabone, [1990](#ece36092-bib-0057){ref-type=\"ref\"}; Rieske, [2002](#ece36092-bib-0058){ref-type=\"ref\"}). This could also explain why browsing frequency was lower in the burn treatments compared to the nonburned treatments two years after the fire, as both Reich et al. ([1990](#ece36092-bib-0057){ref-type=\"ref\"}) and Rieske ([2002](#ece36092-bib-0058){ref-type=\"ref\"}) found increases in foliar nutrients resulting from prescribed burning diminished over time.\n\nConsidering most north temperate ecosystems have experienced dramatic increases in ungulate populations over the last few decades (Milner et al., [2006](#ece36092-bib-0046){ref-type=\"ref\"}; Rooney & Waller, [2003](#ece36092-bib-0061){ref-type=\"ref\"}) the interaction between fire and browsing frequency, we observed is particularly notable. It suggests that positive effects of prescribed fire on oak regeneration could be diminished in forests subject to high browsing pressure. In other words, oaks may readily survive moderate browsing pressure (Harmer, [2001](#ece36092-bib-0027){ref-type=\"ref\"}), but browsing in combination with fire is likely to be more detrimental to survival of oak recruits. Indeed, Nuttle, Royo, Adams, and Carson ([2013](#ece36092-bib-0053){ref-type=\"ref\"}) found that high browsing pressure reduced the benefits of fire and canopy gap creation on tree diversity in forest understories.\n\n4.1. Conclusion and management implications {#ece36092-sec-0013}\n-------------------------------------------\n\nOur results indicate that it is vital to increase light availability to promote naturally regenerated oak recruit survival and growth, especially when disturbances such as browsing and/or prescribed fire damage oak shoots. In areas of high conservation concern, this could be accomplished by removing canopy trees of less importance for biodiversity than oaks (cf. Leonardsson, L\u00f6f, & G\u00f6tmark, [2015](#ece36092-bib-0039){ref-type=\"ref\"}). Protection from ungulate browsing had a limited impact in this experiment; however, considering ungulate preference for oak browse, it is likely their impact will be maintained or even increase over time. This work also demonstrated relatively high survival and invigorated growth of *Q.\u00a0robur/petraea* sprouts following a low\u2010intensity surface fire.\n\nAltogether, our results indicate that a low\u2010intensity surface fire combined with some form of canopy cutting to increase light availability may serve as a first step for promoting natural regeneration of *Q.\u00a0robur* and *Q.\u00a0petraea* in temperate European forests. In particular, broadleaved forests of conservation concern in north temperate Europe, such as those examined in this study, often develop a significant regeneration pool of the shade\u2010tolerant *P.\u00a0abies* (G\u00f6tmark, Berglund, & Wiklander, [2005](#ece36092-bib-0026){ref-type=\"ref\"}). The reduction of conifer reproduction that we observed through burning supports the application of low\u2010intensity fire as an efficient management tool in such situations. Fire would likely also be efficient in situations where shade\u2010tolerant *F.\u00a0sylvatica* and *C.\u00a0betulus* constitute the main competitors, which is often the case in central temperate Europe (Bobiec et al., [2019](#ece36092-bib-0009){ref-type=\"ref\"}; Ziobro et al., [2016](#ece36092-bib-0070){ref-type=\"ref\"}). However, further research is needed in several areas before we can recommend prescribed burns in oak\u2010dominated forests. Long\u2010term vitality of oak regeneration and overstory oaks following fire needs to be evaluated, preferably within different light environments and in areas with varying browsing pressure to further elucidate combined effects. Also, an understanding of the possible benefits of multiple fires, as well as evaluation of fire frequency and seasonality, is needed. Furthermore, the species\u2010specific response of competing woody vegetation following fire needs further investigation to determine whether oaks gain a competitive advantage. Finally, as oak\u2010dominated forests are of great importance for conservation of biodiversity, the effects of fire on species requiring oak habitats needs thorough examination.\n\nAUTHOR CONTRIBUTIONS {#ece36092-sec-0015}\n====================\n\nAll authors contributed to conceive the ideas and designed methodology; LP collected and analyzed the data; LP and ML led writing of the manuscript. All authors contributed critically to drafts and gave final approval for publication.\n\nCONFLICT OF INTETEST {#ece36092-sec-0016}\n====================\n\nNone declared.\n\nSupporting information\n======================\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\nThe Foundation Oscar and Lili Lamms Memory as well as Erik and Ebba Larssons and Thure Rignells Foundation supported this research. We thank private landowners and Sveaskog for providing field sites, Max Jensen, Per Nordin, and Alessandra Salvalaggio for field assistance. The authors declare no conflict of interest.\n\nDATA AVAILABILITY STATEMENT {#ece36092-sec-0018}\n===========================\n\nData are available at Dryad Digital Repository: <https://doi.org/10.5061/dryad.cz8w9gj0h>.\n"}
+{"text": "Sir,\n\nAntimicrobial resistance refers to the phenomenon under which an antimicrobial drug is no longer effective against the microorganism.\\[[@ref1]\\] This is a major public health concern, as because the standard treatment becomes ineffective, the infection will persist and can be transmitted to others, eventually resulting in a number of complications, disabilities, and mortality.\\[[@ref1][@ref2]\\] In addition, it will result in an enormous financial burden for the individual, the community, and the health system.\\[[@ref1][@ref2]\\] Furthermore, amidst the globalization and increasing trends of trade and travel, no section of the world is immune to the public health menace of antimicrobial resistance.\\[[@ref3]\\] As it is very well-known that not many newer drugs are going to come for treating medical illnesses in the near future, it is very much crucial that the health sector should realize the gravity of the problem.\\[[@ref1][@ref3]\\]\n\nEven though, emergence of antimicrobial resistance is a natural evolutionary phenomenon, resulting because of the survival of resistant organisms (and death of susceptible organisms), on exposure to an antimicrobial agent, many human-related factors have aggravated the problem to an enormous extent.\\[[@ref2]\\] These factors include limited commitment from the policy makers and minimal orientation of the treating physicians about the importance of prescription of drugs in right dose for appropriate duration, and to only those patients in which it is indicated.\\[[@ref1][@ref2][@ref3]\\] In addition, factor like inappropriate usage (overuse or underuse or misuse) of medicines; exposure of patients to poor quality medicines and hence suboptimal drug dose; expansion of the counterfeit medicine market; and administration of subtherapeutic dosage of medicines to animals at times of their rearing, have also contributed to the global emergence of the problem.\\[[@ref3][@ref4]\\] However, at no stage the importance of parameters like existence of multiple lacunae\\'s in the infection prevention and control measures; infrastructure constraints (viz., no uniform availability and access to the laboratories where resistant microorganisms can be detected); and weaker surveillance mechanism because of which no comprehensive information is available regarding drug resistance, can be undermined.\\[[@ref1][@ref3][@ref5]\\]\n\nFindings of a recently released survey conducted across 133 nations of the world revealed that even though there is a positive intent from the policy makers in some of the nations, multiple gaps such as nonexistence of a holistic action plan to fight antimicrobial resistance; poor infrastructure support with limited laboratory capacity; and poor quality of monitoring and supervision, have been identified.\\[[@ref3][@ref6]\\] In fact, it has been even highlighted that issues such as absence of standardized guidelines for treatment of an illness; over-the-counter sale of drugs without prescription; poor public awareness; and limited number of strategies to combat nosocomial infections have also contributed to a great extent.\\[[@ref3][@ref6]\\] Thus, it is very important to understand that antibiotic resistance has spread to the entire world (viz., resistance to third generation cephalosporins used for treating gonorrhea, drugs used for treating tuberculosis/HIV/malaria/influenza, etc.), and is jeopardizing the act of treating common infections in both community and healthcare establishment settings.\\[[@ref2][@ref6]\\]\n\nAcknowledging the role of multiple factors and the complex nature of the problem, the need of the hour is to look for a prompt and coordinated response, in which roles and responsibilities of each stakeholder, namely:\n\nCommunity --- by adhering to practice of hand washing, getting immunized, discouraging self-medication, completing the full course of therapy, etc.;Health professionals and pharmacists --- implementing measures such as appropriate infection prevention and control measures in hospitals, prescribing antibiotics incorrect dosage and that too only when indicated, supporting public health sector by not selling antibiotics over-the-counter without prescription, etc.;Program managers --- by creating awareness among the general population, facilitating epidemiological studies to identify the extent and causes of resistance, strengthening infection control and monitoring practices, regulating and promoting appropriate use of antibiotics;International agencies --- by assisting nations to strengthen their ability to tackle antimicrobial resistance, conducting research and trials for development of new drugs or vaccines, development of tools for prompt diagnosis of infection, etc., is clearly specified.\\[[@ref3][@ref4][@ref5][@ref7][@ref8]\\]\n\nHowever, there is an immense need to simultaneously develop linkages with other sectors such as animal husbandry, and food and agriculture, to promote optimal practices so that antibiotics are used only in optimal dosages in both humans and animals.\\[[@ref1][@ref4]\\]\n\nTo conclude, in order to arrest the progress of the world toward postantibiotic era, it is high time that all the stakeholders should sit together and work out a global action plan to avoid the emergence and spread of antimicrobial resistance.\n\nFinancial support and sponsorship {#sec2-1}\n=================================\n\nNil.\n\nConflicts of interest {#sec2-2}\n=====================\n\nThere are no conflicts of interest.\n"}
+{"text": "The unprocessed reads have been deposited in the NCBI Sequence Read Archive (SRA) under the accession number SRR4934939 (study SRP092603). The Transcriptome Shotgun Assembly project has been deposited at DDBJ/EMBL/GenBank under the accession number GFPI00000000. The version described in this paper is the first version, GFPI01000000.\n\nIntroduction {#sec001}\n============\n\nThe interactions between host plants and their phytophagous insects are intrinsically complex and subject to remarkable evolution, where both have adapted strategies to avoid each other's defense systems. Plants have developed an extraordinary array of physical barriers, constitutive chemical mechanisms, and direct and indirect inducible defenses intended to counter/offset the effects of insect attack \\[[@pone.0192003.ref001]--[@pone.0192003.ref005]\\]. In parallel, insects have adapted tactics to combat the diverse arsenal of plant defenses, allowing them to feed, grow, and reproduce on their host plants \\[[@pone.0192003.ref006]\\]. A wide range of phytophagous insects possess highly modified piercing-sucking mouthparts, enabling them to use phloem sap as their exclusive food source. During the feeding process, saliva is injected into plant tissues to aid in penetration, ingestion of nutrients, and modulate plant responses \\[[@pone.0192003.ref007]\\]. Moreover, a subset of secreted proteins with structural, chelating, or enzymatic properties is thought to serve as the 'effectors' of these processes \\[[@pone.0192003.ref008],[@pone.0192003.ref009]\\].\n\nLeaf-feeding beetles (Order Coleoptera) possess \"chewing\" mouthparts consisting of two opposing mandibles, which are used to remove leaf sections or entire leaves. Lacking the salivary glands of their sap-sucking counterparts, these beetles are thought to regurgitate onto the surface of leaves while feeding to begin the digestion process. The bulk of the regurgitant likely originates from the foregut, though the gnathal glands in the cephalic regions may also contribute \\[[@pone.0192003.ref010],[@pone.0192003.ref011]\\]. Importantly, the deposited regurgitant is believed to be a major source of effector proteins and/or small molecules \\[[@pone.0192003.ref008]\\] that can evoke changes in host plant defenses, thereby making the plant more vulnerable to the herbivore attack \\[[@pone.0192003.ref009]\\]. Such effectors have been well documented in the saliva of sap-sucking insects \\[[@pone.0192003.ref012]--[@pone.0192003.ref021]\\] but little information is currently available for the regurgitant of chewing insects. In addition to effectors, recent studies have indicated these oral secretions contain diverse microbial communities that may alter plant-insect interactions \\[[@pone.0192003.ref022],[@pone.0192003.ref023]\\]. Taken collectively, the regurgitant of leaf-feeding beetles appears intricate and multifaceted but is only beginning to be explored.\n\nBeetles vector at least six groups of plant viruses: *Machlomovirus*, *Bromovirus*, *Carmovirus*, *Comovirus*, *Sobemovirus*, and *Tymovirus*. A unique and specific relationship exists between herbivore beetles and the plant viruses they vector: viruses that are transmissible by beetles are solely transmitted by beetles. Inoculative beetles deposit the active virus in regurgitant on the surface of the wounded leaf during feeding \\[[@pone.0192003.ref024]\\]. Unlike most other plant viruses, beetle-borne viruses can be inoculated into a chewing wound. This is because the virus particles are rapidly translocated in the xylem elements away from the inoculation site and infect unwounded cells at a distance from the feeding site \\[[@pone.0192003.ref025],[@pone.0192003.ref026]\\]. Previous work has revealed that factor(s) in the regurgitant also play a role in the virus-vector specificity \\[[@pone.0192003.ref011],[@pone.0192003.ref027],[@pone.0192003.ref028]\\]. Indeed, mixing purified virus with regurgitant prevents host plant infection by non-beetle-borne viruses but has no effect on beetle-transmissible viruses. Though this specificity is well established, little is known regarding the specific factor(s) in the regurgitant that govern the selective inhibition.\n\nOne of the most prevalent and destructive leaf-feeding beetles in North America is the Mexican bean beetle, *Epilachna varivestis* Mulsant \\[[@pone.0192003.ref029]\\]. Native to the plateau region of southern Mexico, *E*. *varivestis*' can now be found from Guatemala to southern Canada \\[[@pone.0192003.ref030],[@pone.0192003.ref031]\\]. Since its establishment in the United States in 1942, the beetle has become a major economic pest of *Phaseolus* spp., including soybean (Glycine max (L.) Merr.) \\[[@pone.0192003.ref031],[@pone.0192003.ref032]\\]. *E*. *varivestis* is also an efficient vector of several plant viruses, including *Cowpea severe mosaic comovirus* \\[[@pone.0192003.ref029]\\], *Southern bean mosaic sobemovirus* \\[[@pone.0192003.ref033]\\], and *Black gram mottle carmovirus* \\[[@pone.0192003.ref034]\\]. Perhaps the most important virus vectored by the beetle is *Bean pod mottle virus* (BPMV). This is a positive-sense single-stranded RNA comovirus, and one of the most ubiquitous viruses of soybean in North America.\n\nTo date, little is known about the molecular, chemical, or biological composition of beetle regurgitant or how the factor(s) within may modulate plant-insect interactions. To begin to characterize this, we carried out RNA sequencing on the oral secretions of *E*. *varivestis* to assemble the first ever leaf-feeding beetle \"regurgitome\". We also implemented 16 rDNA sequencing to characterize the bacterial communities in the regurgitant. The regurgitant consisted of a rich molecular assortment of genes encoding putative extracellular proteins involved in digestion, molting, immune defense, and detoxification. By carrying out plant inoculation assays that combined purified virus and regurgitant, we reinforced the importance of regurgitant in the unique and specific relationships between leaf-feeding coleopterans and the viruses they transmit. Ultimately, these studies begin to characterize the critical role of regurgitant in virus transmission and in the interactions between leaf-feeding beetles and their host plants.\n\nMaterials and methods {#sec002}\n=====================\n\nBeetle colony maintenance {#sec003}\n-------------------------\n\nA laboratory colony of *Epilachna varivestis* was established in 2014 from multiple, ongoing field collections in Ohio and reared in growth chambers under controlled conditions of 25 \u00b1 3\u00b0C, 65% RH with a 14-h:10 h light-dark cycle that included 1.5 h dawn and dusk transitions. Beetles were maintained on 'Sloan' seedlings placed in 47.5 cm x 47.5 cm x 47.5 cm cages.\n\nThe *E*. *varivestis* regurgitome {#sec004}\n---------------------------------\n\n### Regurgitant collection {#sec005}\n\nRegurgitant was isolated from adult feeding beetles (between 1--3 weeks old) at leaf wounding sites using capillary glass tubes, and immediately placed into 0.5 mL microcentrifuge tubes containing a 5:1 ratio of extraction buffer XB (Thermo Fisher Scientific, Waltman, MA) and 2-Mercaptoethanol (Sigma-Aldrich, St. Louis, MO). Samples were then snap frozen in liquid nitrogen and stored at \u221280\u00b0C until nucleic acid isolation.\n\n### Nucleic acid isolation {#sec006}\n\nTotal RNA and genomic DNA (gDNA) was extracted from approximately 200 *E*. *varivestis* by implementing a multifaceted approach. First, gDNA was isolated by following the DNA purification protocol of the AllPrep DNA/RNA Mini Kit (QIAGEN, Germantown, MD). The initial flow-through from the DNA spin column (step 5) was then collected for RNA extraction using the Arcturus PicoPure RNA Isolation Kit (Thermo Fisher Scientific, Waltman, MA), following the manufacturer's protocol but with the omission of LCM Caps. RNA and gDNA quality was evaluated using the Nanophotometer NP80 (Implen Inc., Westlake Village, CA), and quantity was calculated on the Qubit 3.0 fluorometer using the RNA HS or dsDNA HS assay kits (Thermo Fisher Scientific, Waltman, MA).\n\n### cDNA library preparation {#sec007}\n\nRNA (500 ng/sample) was used to generate one cDNA library for RNA sequencing using the TruSeq Sample Prep Kit V1 (Illumina, San Diego, CA) following the company recommended protocols. Quantity and quality of the cDNA library was assessed using the BioAnalyzer 2100 (Agilent Technologies, Santa Clara, CA) and then diluted to 100 fmoles.\n\n### Illumina sequencing {#sec008}\n\nThe cDNA and 16S rRNA libraries were sequenced in 300-bp paired-end fashion on one run of the llumina MiSeq System at the G\u00e9nome Qu\u00e9bec Innovation Centre. Illumina Analysis Package CASAVA 1.8.2 was used to perform bcl conversion and demultiplexing. Image deconvolution and quality value calculations were carried out using the Illumina GA pipeline v1.6.\n\n### Regurgitome assembly {#sec009}\n\nRaw reads of the cDNA library were imported into CLC Genomics Workbench (v6.5.1, CLC Bio) and trimmed for quality, adapter indexes and poly(A) tails using the default settings (Ambiguous limit = 2, quality limit = 0.05). Processed reads were assembled *de novo* into contigs using two independent approaches. First, using the CLC Bio algorithm based on de Bruijn graphs and the optimized parameters: Word Size = 64, Bubble Size = 900, Length Fraction = 0.65 and Similarity Fraction = 0.85. Second, using Oases v0.2.08 \\[[@pone.0192003.ref035]\\] with Kmer sizes of 53, 59, 65, 71, 77, 83, and 89. To obtain the set of non-redundant transcripts for each assembly, transcripts \u226580% sequence similarity were collapsed into clusters and the longest read retrieved using CD-HIT-EST \\[[@pone.0192003.ref036]\\]. The assemblies were then merged into a final assembly using Minimus2 \\[[@pone.0192003.ref037]\\]. Only contigs of \u2265300 nt in length with average coverage \u22655 were included in the final assembly.\n\nMicrobial contamination was identified and removed from the *E*. *varivestis* transcriptome using desktop-downloaded BLASTn against the NCBI bacteria non-redundant database (*E*-value \\<1 x 10^\u221250^) and a GC content threshold of 45%. Contigs of soybean origin were identified and removed using BLASTn against the most recent *Glycine max* \\[Glyma2.0\\] reference genome including scaffold sequences (*E*-value \\<1 x 10^\u221250^). The remaining contigs were assigned hierarchical gene ontologies (GO terms) on the basis of biological processes, molecular functions and cellular components using the platform-independent Java\u2122 6 implementation of the BLAST2GO software \\[[@pone.0192003.ref038]\\]. The top five BLASTx hits to the nr database with a cut-off *E*-value of 10^\u22123^ were considered for GO annotation.\n\n### Pair-wise comparisons among arthropods {#sec010}\n\nPair-wise comparisons of *E*. *varivestis* regurgitome contigs to the gut transcriptome assemblies of seven arthropod species across five orders were carried out using desktop downloaded tBLASTx software with a set *E*-value threshold of 10^\u221210^. The complete list of species included *Anopheles gambiae* (malaria mosquito, order Diptera, 22,889 sequences) \\[[@pone.0192003.ref039]\\], *Pectinophora gossypiella* (pink bollworm, order Lepidoptera, 11,746 sequences) \\[[@pone.0192003.ref040]\\], *Haemaphysalis flava* (hard tick, order Ixodida, 76,556 sequences) \\[[@pone.0192003.ref041]\\], *Periplaneta americana* (American cockroach, order Blattodea, 78,837 sequences) \\[[@pone.0192003.ref042]\\], *Leptinotarsa decemlineata* (Colorado potato beetle, order Coleoptera, 21,622 sequences) \\[[@pone.0192003.ref043]\\], *Chrysomela tremulae* (poplar leaf beetle, order Coleoptera, 10,876 sequences) \\[[@pone.0192003.ref044]\\], *Gastrophysa viridula* (green dock leaf beetle, order Coleoptera, 20,791 sequences) \\[[@pone.0192003.ref045]\\].\n\n### Identification of putative secreted proteins {#sec011}\n\nThe six open reading frame (ORF) amino acid sequences were predicted from the *E*. *varivestis* contig sequences using ORF-Predictor \\[[@pone.0192003.ref046]\\]. Only the subset of predicted sequences \u226550 amino acids was used in subsequent analyses. The SignalP 4.1 neural networks algorithm \\[[@pone.0192003.ref047]\\] was implemented to detect putative transmembrane proteins with signal peptide secretion and cleavage site signatures in their amino acid sequences using the default settings for D-score.\n\n### 16S rDNA library preparation and analysis {#sec012}\n\nBacterial 16S rRNA genes present within the beetle regurgitant were amplified using universal bacterial primers with the appending of Illumina adapter sequences to construct an amplicon library from the V3-V4 region of the 16S rDNA genes. By using the Illumina polymerase-binding regions, samples can be sequenced in lieu of sequencing primers thereby eliminating the need for an additional ligation step. The primer pairs, retrieved from \\[[@pone.0192003.ref048]\\], were: (A): S-D-Bact-0341-b-S-17, `5\u2019-CCTACGGGNGGCWGCAG-3\u2019` and S-D-Bact-0785-a-A-21, `5\u2019-GACTACHVGGGTATCTAATCC-3\u2019` \\[[@pone.0192003.ref049]\\]; and (B): S-D-Bact-0008-a-S-16, `5\u2019-AGAGTTTGATCMTGGC-30` \\[[@pone.0192003.ref050]\\] and S-D-Bact-0907-a-A-20, `5\u2019-CCGTCAATTCMTTTGAGTTT-3\u2019` \\[[@pone.0192003.ref051]\\]. PCR was performed at an initial denaturation temperature of 96\u00b0C for 3 min, followed by 25 cycles of 96\u00b0C for 30 s, 55\u00b0C for 30s and 72\u00b0C for 30 s. A final elongation step at 72\u00b0C was run for 5 min. PCR products were purified using 20 \u03bcL of AMPure beads (Beckman Coulter, Takeley, UK) and quantified on the Qubit 3.0 fluorometer.\n\nTaxonomy was assigned to amplicon sequences *in* the 16S rDNA library using a high performance version of the RDP Na\u00efve Bayes taxonomic classification algorithm via the BaseSpace 16S Metagenomics pipeline.\n\nBPMV-*E*. *varivestis* specificity assays {#sec013}\n-----------------------------------------\n\nThe experiments described below were largely developed based on the previously established protocols (see \\[[@pone.0192003.ref011],[@pone.0192003.ref028]\\]). One beetle transmissible virus, *Bean pod mottle virus* (BPMV) and one beetle non-transmissible virus, *Soybean mosaic virus* (SMV) were used in these experiments.\n\n### Virus extraction and isolation {#sec014}\n\nSoybean leaves infected with BPMV or SMV were homogenized in 10mM KHPO~4~ buffer (pH 7). The homogenate was clarified overnight by low speed centrifugation with 8% butanol at 4\u00b0C. After a centrifuge at 13,000 \u00d7 *g* for 20 min, virus particles were precipitated from the supernatant by using an equal volume of 16% PEG (Sigma Chemical Co., St. Louis, Mo.) with 0.5 M NaCl. After precipitation, the pellet was resuspended in 5 ml of KHPO~4~ buffer and incubated at room temp for 10 min. The resuspended pellet was transferred to a 15 ml polypropylene centrifuge tube containing 1.2 ml of chloroform, and centrifuged at 12,000 \u00d7 *g* for 5 min. The top aqueous layer (containing the RNA) was precipitated by addition of 0.5 volume of isopropanol for 5 min at room temp, followed by centrifuging at 5,000 \u00d7 *g* for 5 min. The resultant pellet washed with 80% ethanol and resuspended in washing buffer (10 mM Tris-HCl \\[pH 7.5\\], 0.15 M LiCl, 1 mM EDTA). This process was repeated 3 times with the final resuspension in 1 mL of KHPO~4~ buffer.\n\n### Regurgitant collection {#sec015}\n\nBeetles were induced to regurgitate by holding the individual between the thumb and forefinger and gently prodding the mouthparts with a capillary glass tube, which collected the regurgitant. Only freshly collected regurgitant were used for the experiments.\n\n### Inocula {#sec016}\n\nThe 'Sloan' cultivator (susceptible to BPMV and SMV) was used to assess the impact of *E*. *varivestis* regurgitant on virus infectivity using two different inoculation approaches: mechanical leaf-rub inoculations and gross wounding (described below). Inoculations were carried out for each virus independently using a predetermined dose of virus that gave \\>90% infection of the positive control for the mechanical and gross wounding assays, respectively. The inocula deposited at the leaf wounding sites consisted of the purified virus in 0.01 M KHPO~4~ mixed with five-fold dilutions of regurgitant (4 in total). After 14 d of visual assessment of symptom development, virus infection was confirmed and titer estimated in using enzyme-linked immunosorbent assays \\[[@pone.0192003.ref052]\\]. A total of 30 replicates were carried out for each combination of virus/regurgitant mixture.\n\n### Mechanical leaf-rub inoculations {#sec017}\n\nFollowing the protocols outlined in \\[[@pone.0192003.ref053]\\], isolated virus was mixed with and mechanically inoculated onto the youngest trifoliate of 1.5 wk old test plants lightly dusted with corundum. Rub-inoculation of plants with isolated virus and KHPO~4~ buffer served as a positive control.\n\n### Gross wounding assays {#sec018}\n\nA single hole 8 mm in diameter was bored into a leaf of the test plant using the fractured edge of a glass cylinder. Immediately prior to cutting the hole, the cylinder was dipped into an equal volume mixture of isolated virus and regurgitant. Control plants underwent identical treatment, but with the inoculum consisting of an equal volume mixture of isolated virus and KHPO~4~ buffer.\n\n### Statistical analysis {#sec019}\n\nTo test the statistical significance of differences in the relative proportions of infected plants for each virus and ratios of regurgitant to purified virus mixture, we subjected the data to a chi-square analysis followed by a 2 x 24 Marascuilo procedure \\[[@pone.0192003.ref054]\\], with a threshold of significance at *P* \\< 0.05.\n\nResults and discussion {#sec020}\n======================\n\nAssembly of the *E*. *varivestis* regurgitome {#sec021}\n---------------------------------------------\n\nA cDNA library derived from the regurgitant of roughly 200 *E*. *varivestis* individuals was sequenced, which produced 56,289,018 paired end reads of 300 nt. After trimming (quality, adapters, and poly (A) sequences) 40,118,669 reads were obtained. The unprocessed reads have been deposited in the NCBI Sequence Read Archive (SRA) under the accession SRR4934939 (study SRP092603). *De novo* assembly of the trimmed reads generated 104,977 non-redundant contigs of \\>300 nt with a mean length of 625 nt. Subsequent removal of bacterial (n = 38,352) and soybean (n = 21,077) contamination resulted in 45,548 contigs considered to be beetle in origin, and this subset represents the *E*. *varivestis \"*regurgitome\". These contigs are thought to be derived mostly from the beetle's gastrointestinal cell products. While the regurgitome is presumably similar in composition to the gut transcriptome, our approach has one important distinction: the regurgitome is likely produced by enterocytes lining the lumen side of the gastrointestinal tract, which play important roles in secretion and likely expresses the transcripts encoding for host plant modulating effector proteins. In contrast, assembly of the gut transcriptome would have yielded a mixed transcriptomic profile, the bulk of which is comprised of structural, muscular, and absorptive cells of little significance for the regurgitant.\n\nComparisons to gut transcriptomes of diverse arthropods {#sec022}\n-------------------------------------------------------\n\nWe surmised that the *E*. *varivestis* regurgitome would show extensive similarity to the gut transcriptomes of other arthropod species (see above), particularly phytophagous Coleopterans. To investigate this, we first wanted to compare the Mexican bean beetle regurgitome to the gut transcriptomes of arthropods displaying diverse feeding strategies. This included blood-feeding mosquitoes and ticks (*Anopheles gambiae* and *Haemaphysalis flava*), a herbivore caterpillar (*Pectinophora gossypiella*), and a generalist cockroach (*Periplaneta americana*). The pair-wise comparisons revealed that less than 20% (n = 8,909) of *E*. *varivestis* contigs had a significant match to at least one of the four arthropod transcriptomes (*E*-value \\<10^\u221210^), with only 1,773 contigs common to all species ([Fig 1A](#pone.0192003.g001){ref-type=\"fig\"}). Next, we sought to determine if the *E*. *varivestis* regurgitome carried considerably more commonality to the gut transcriptomes of herbivore Coleopteran beetles. This included two leaf-feeding beetles (*Leptinotarsa decemlineata* and *Chrysomela tremulae*) and a grass-feeding (*Gastrophysa viridula*) species. Similar to the previous analysis, \\<20% of *E*. *varivestis* contigs (n = 8,430) matched one or more of the beetle transcriptomes, with 3,057 contigs shared by all species ([Fig 1B](#pone.0192003.g001){ref-type=\"fig\"}).\n\n![Venn diagram \\[[@pone.0192003.ref084]\\] showing tBLASTx (E-value \\<10^\u221210^) pair-wise ortholog matches of the *E*. *varivestis* to the gut transcriptomes to the characterized gut transcriptomes.\\\n(A) four arthropod species with diverse feeding strategies (*Pectinophora gossypiella*, *Haemaphysalis flava*, *Periplaneta Americana*, *Periplaneta americana*); and (B) three herbivore Coleopteran beetles (*Leptinotarsa decemlineata*, *Chrysomela tremulae*, *Gastrophysa viridula*).](pone.0192003.g001){#pone.0192003.g001}\n\nOur pairwise comparisons indicated that only a small proportion of the *E*. *varivestis* regurgitome had significant sequence homology to the gut transcriptomes of a diverse grouping of arthropod species. Most of the 8,430 *E*. *varivestis* contigs matching other Coleopteran beetle(s) also matched at least one of the other arthropod species (\\~80%). This result was quite surprising given the greater evolutionary relatedness and similarity in feeding strategies among the beetle species. While it is probable that the *E*. *varivestis* regurgitome contains many genes encoding proteins that carry out unique digestive and defensive functions for *E*. *varivestis* and perhaps other Coccinellidae species, this result is likely partially attributed to sequencing coverage of our contrasting species. The Coleopteran gut transcriptomes were constructed using 454-mediated pyrosequencing and 0.28 to 1.24 million reads per species, whereas the gut transcriptomes of the other arthropods were generated using Illumina sequencing and between 18.6 to 223 million reads per species. Thus, the evolutionary relatedness among Coleopterans was probably offset by the deeper sequencing coverage of the other species. Constructed from nearly 57 million paired-end Illumina reads, the *E*. *varivestis* regurgitome thus represents the deepest coverage and most targeted effort to catalogue potential beetle effectors.\n\nIdentification of genes encoding putative extracellular proteins {#sec023}\n----------------------------------------------------------------\n\nBeetle effectors are proteins secreted into the cells of host plants during feeding. Therefore we initiated our search for putative *E*. *varivestis* effectors by translating the 45,548 contigs into their putative amino acid sequences. ORF analysis predicted 34,835 (76.5%) encoded peptide sequences of \u2265 50 AA. We then carried out *in silico* analysis on the peptide sequences using the SignalP server. A total of 1,555 sequences were predicted to have a secretion signal ([S1 Table](#pone.0192003.s001){ref-type=\"supplementary-material\"}). This Transcriptome Shotgun Assembly project has been deposited at DDBJ/EMBL/GenBank under the accession GFPI00000000. The version described in this paper is the first version, GFPI01000000.\n\nFor the subset of 1,555 contigs encoding putative secreted proteins, 992 (64%) had a significant BLASTx matches to the nr database. [Fig 2](#pone.0192003.g002){ref-type=\"fig\"} shows the top ortholog matches to organisms within the nr database. Not surprisingly, the largest number of significant matches were to a Coleopteran model organism, *Tribolium castaneum*, with \\>40% (n = 401) of all top matches. Other top matches include two other Coleopterans, *Oryctes borbonicus* (n = 16) and *Dendroctonus ponderosae* (n = 41) as well as two aphid species, *Acyrthosiphon pisum* (n = 101) and *Diuraphis noxia* (n = 67). The relatively large number of matches to aphids may be indicative of some commonality in the composition of secreted proteins in the salivary glands of sap-sucking insects and the regurgitant of beetles.\n\n![BLASTx top ortholog matches to organisms within the NCBI non-redundant (nr) database for the subset of 992 *Epilachna varivestis* contigs encoding putative secreted proteins with a significant BLASTx match (E-value \\<10^\u22123^).\\\nOnly organisms with \u226510 matches are shown.](pone.0192003.g002){#pone.0192003.g002}\n\nFunctional characterization of putative effector proteins {#sec024}\n---------------------------------------------------------\n\nBased on the gene ontologies (biological processes, molecular functions, and cellular components) assigned by BLASTx, contigs were manually placed into various categories and sub-categories based on consensus function. This was not possible for 23% of genes (n = 229), as they could not be assigned any putative function. The proportion of differentially expressed genes in twelve functional categories containing of a minimum of 20 genes is given in [Fig 3](#pone.0192003.g003){ref-type=\"fig\"} (other categories not shown). A complete description of the annotation for each gene can be found in [S2 Table](#pone.0192003.s002){ref-type=\"supplementary-material\"}. The most populous categories was \"Metabolism\" containing 40% of genes (n = 303) and \"Defense/Immune Response\" (n = 99).\n\n![Distribution of *Epilachna varivestis* contigs encoding putative secreted proteins among functional categories.\\\nBars indicate the proportion of genes in each category: Number of genes in each category is given beside each bar. Percentages do not total to 100 as not all categories are shown.](pone.0192003.g003){#pone.0192003.g003}\n\nOver the past few years, a large body of research has accumulated describing the molecular and chemical makeup of salivary glands and saliva of various sap-sucking insects \\[[@pone.0192003.ref016],[@pone.0192003.ref055]--[@pone.0192003.ref057]\\]. On the other side of the spectrum, virtually no studies have been undertaken to describe the composition of regurgitant in leaf-feeding beetles. The only data available are enzymatic assays indicating that the beetle regurgitant is rich in proteases \\[[@pone.0192003.ref058]\\] and ribonucleases \\[[@pone.0192003.ref059]\\]. Our study largely supports this at the molecular level by identifying a considerable number of genes encoding these enzymes in our dataset. However, as described below, *E*. *varivestis* regurgitant appears to also contain an extensive suite of other proteins involved in gastrointestinal, immunological, and developmental processes.\n\nRegurgitant genes are involved in digestion {#sec025}\n-------------------------------------------\n\nLacking the salivary glands of sap-sucking insects, it has been speculated that beetles regurgitate oral secretions onto the leaves to begin their digestive processes. Our findings of a large number of regurgitant genes encoding putative secreted proteins involved in digestion reinforces this idea. Nearly half of metabolic genes encode proteins with proteolytic activity (47%), of which the vast majority function as lysosomal proteases (e.g., cathespins) ([Fig 4](#pone.0192003.g004){ref-type=\"fig\"}). Lysosomes are intracellular organelles that play key roles in the digestive breakdown and recycling of diverse biological materials \\[[@pone.0192003.ref060]\\]. For herbivorous beetles, this includes the enzymatic break down of proteins in leaves into smaller peptides and amino acids that can be readily absorbed and utilized by the organism. The next most populous sub-categories within Metabolism were fatty acid/lipid and carbohydrate related. The former included an array of lipases, which are one of the main digestive enzymes involved in the insect digestion process \\[[@pone.0192003.ref061]\\]. Carbohydrate metabolism consisted of a variety of enzymes involved in degradation complex carbohydrates, such as amylases, maltases, and glycosyl hydrolases, some of which may also play an important role in breaking down plant cell walls \\[[@pone.0192003.ref062]--[@pone.0192003.ref064]\\]. The transport of molecules across cell membranes and between subcellular compartments is also an essential component of both digestion and normal cellular functions. Interestingly, our dataset produced several gene products associated with the transport of simple molecules, such as sugars and vitamins, and with vesicular trafficking which are essential for normal digestive function \\[[@pone.0192003.ref065],[@pone.0192003.ref066]\\].\n\n![Distribution of *Epilachna varivestis* contigs encoding putative secreted proteins involved in metabolism.](pone.0192003.g004){#pone.0192003.g004}\n\nRegurgitant genes function in defense responses and detoxification {#sec026}\n------------------------------------------------------------------\n\nFor nearly 400 million years phytophagous insects and their host plants have been entwined in an evolutionary arms race \\[[@pone.0192003.ref009]\\]. Plants have developed different mechanisms to fend off or deter insect attack, whereas insects employ a multitude of strategies to overcome these plant barriers \\[[@pone.0192003.ref006]\\]. Evidence has quickly accumulated that indicates sap-sucking insects secrete proteins and/or small molecules in their saliva to suppress host plant defenses \\[[@pone.0192003.ref012]--[@pone.0192003.ref017],[@pone.0192003.ref019]--[@pone.0192003.ref021]\\]. Lacking salivary glands, it is likely that herbivore beetles secrete these proteins in their regurgitant. Supporting this, we found approximately 13% of the genes encoding putative extracellular proteins function in defense/immune. This includes an array of attacins, defensins, toll-pathway genes, C-type lectins, glutathione peroxidase, 1,3-beta-D glucan binding proteins, and autophagy genes \\[[@pone.0192003.ref067]\\]. Several other genes are involved in detoxification processes, such as cytochrome p450s and a variety of esterases. Many of these genes likely aid in the detoxification and even sequestration of plant chemical defenses \\[[@pone.0192003.ref068]\\]. Overall this subset serves as strong candidates for effectors that play active roles in combatting the anti-herbivory defenses of soybean. Functional assays targeting our most promising candidates will better implicate specific defense and detoxification genes in modulating plant-insect interactions.\n\nRegurgitant contains genes involved in exocuticle molting {#sec027}\n---------------------------------------------------------\n\nOne of the larger categories contained genes involved in exocuticle-related functions, namely cuticle proteins and chitinases. In insects, these proteins belong to family 18 glycosyl hydrolases, and have been detected in gut tissues \\[[@pone.0192003.ref069]\\]. They are predicted to mediate the digestion of chitin present in the exoskeleton chitooligosaccharides \\[[@pone.0192003.ref070],[@pone.0192003.ref071]\\]. The *E*. *varivestis* life cycle consists of an egg stage followed by four instar stages and a pupal stage over a 30 to 70 d period before emerging as adults. Since it is unlikely that these the regurgitant plays a role in molting, these genes are probably highly expressed throughout the organism and are therefore also present in the regurgitome.\n\nMicrobiome of the regurgitant {#sec028}\n-----------------------------\n\nA large body of literature has amassed showing many insect groups are colonized by communities of diverse microbes, some of which act as symbionts \\[[@pone.0192003.ref072]--[@pone.0192003.ref076]\\]. There is growing evidence indicating these symbioses in the saliva of sap-sucking insects play an important role in host plant interactions \\[[@pone.0192003.ref077]--[@pone.0192003.ref079]\\]. Recent studies have demonstrated some chewing insects orally secrete symbiotic gut bacteria onto the surface of wounded leaves sites during feeding. Remarkably, these microbes can manipulate plant physiology to the benefit of their insect host in terms of nutrient acquisition \\[[@pone.0192003.ref022],[@pone.0192003.ref023]\\]. This prompted us to carry out 16S rDNA amplicon sequencing in an effort to catalogue the bacterial communities in the regurgitant of *E*. *varivestis*. We identified a total of 1,230 bacterial species representing 577 genera ([S3 Table](#pone.0192003.s003){ref-type=\"supplementary-material\"}), suggesting the regurgitant is comprised of a diverse microbiome.\n\nThe *E*. *varivestis* regurgitant contained a multitude of microbes that can be considered candidates for modulating plant physiology. By applying bacteria isolated from larval oral secretions to wounded plants, Chung and coauthors \\[[@pone.0192003.ref022]\\] demonstrated that microbial symbionts belonging to the genera *Stenotrophomonas*, *Pseudomonas*, and *Enterobacter* were responsible for host plant defense suppression in Colorado potato beetles (*Leptinotarsa decemlineata*). We identified several representatives of all three genera in *E*. *varivestis* regurgitant: 6 *Stenotrophomonas*, 55 *Pseudomonas*, and 12 *Enterobacter*. Jasmonic acid defense-suppressing Enterobacteriaceae-1 (*Serratia*) identified in fall armyworm oral secretions \\[[@pone.0192003.ref023]\\], were also found in *E*. *varivestis* regurgitant. Overall, the four phyla most represented in our dataset (Firmicutes, Actinobacteris, Bacterioidetes, and Proteobacteria) are also the most commonly associated with insect species \\[[@pone.0192003.ref080]\\]. Many of the other microbes identified are commonly associated with the soybean phyllosphere (e.g., Actinobacteria, Bacteroidetes, and Proteobacteria) \\[[@pone.0192003.ref081]\\]. Future studies are needed to explore the impact of these microbes on soybean-*E*. *varivestis* interactions, and to better disentangle the relative contributions of the regurgitant bacterial communities on overcoming plant defense strategies.\n\nRole of regurgitant of beetle-borne viruses specificity {#sec029}\n-------------------------------------------------------\n\nBeetle-bornes viruses can only be transmitted plant to plant by beetles, and it is thought that the beetle's regurgitant plays a significant role in this specificity. Previous research demonstrated that non-beetle-transmissible plant viruses lost their infectivity when mixed with the regurgitant of leaf-feeding beetles, whereas beetle-transmissible remained infectious \\[[@pone.0192003.ref011],[@pone.0192003.ref027],[@pone.0192003.ref028]\\]. To investigate this further, we assayed virus-vector specificity of *E*. *varivestis* and two prevalent soybean viruses: an aphid vectored virus not known to be transmissible by beetles, *Soybean mosaic virus* (SMV), and a virus naturally vectored by the beetle, *Bean pod mottle virus* (BPMV). Moreover, two different leaf wounding approaches were implemented: mechanical inoculation via an abrasive, as well as gross wounding technique that more naturally simulated wounding induced by beetle feeding. [Table 1](#pone.0192003.t001){ref-type=\"table\"} shows the effect of beetle regurgitant on virus transmission for the different wounding techniques and virus:regurgitant combinations deposited at the leaf wounding sites.\n\n10.1371/journal.pone.0192003.t001\n\n###### Impact of regurgitant on transmission of a beetle-borne virus (*Bean pod mottle virus*, BPMV) and a non-beetle transmissible virus (*Soybean mosaic virus*, SMV) using two different leaf inoculation techniques (mechanical, M and gross wounding, G).\n\nFor each inoculum/virus combination, 30 experimental plants were assayed. The percentage of virus-infected plants is shown and the number of infected plants is indicated in parenthesis. Statistical significance between treatments was tested using a Marascuilo procedure (see [methods](#sec002){ref-type=\"sec\"}), different superscript letters denote statistically different treatments (*P* \\< 0.05).\n\n![](pone.0192003.t001){#pone.0192003.t001g}\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Virus\\     Leaf\\      Ratio of beetle regurgitant to purified virus in inoculum mixture                                                               \n             Wounding                                                                                                                                   \n  ---------- ---------- ------------------------------------------------------------------- ------------ ------------ ---------------- ---------------- ------------\n  **BPMV**   G          93.3% ^a^ (28)                                                      90% ^a^\\     90% ^a^\\     86.7% ^a^\\       93.3% ^a^ (28)   90% ^a^\\\n                                                                                            (27)         (27)         (26)                              (29)\n\n  M          90% ^a^\\   96.7% ^a^ (29)                                                      100% ^a^\\    90% ^a^\\     93.3% ^a^ (28)   90% ^a^\\         \n             (27)                                                                           (30)         (27)                          (29)             \n\n  **SMV**    G          0% ^b^\\                                                             3.33% ^b^\\   0% ^b^\\      6.67% ^b^\\       0% ^b^\\          86.7% ^a^\\\n                        (0)                                                                 (1)          (0)          (2)              (0)              (28)\n\n  M          0% ^b^\\    0% ^b^\\                                                             0% ^b^\\      6.67% ^b^\\   0% ^b^\\          90% ^a^\\         \n             (0)        (0)                                                                 (0)          (2)          (0)              (29)             \n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nOur results indicated that *E*. *varivestis* regurgitant effectively suppresses infection of the non-beetle-borne SMV when the inoculum is diluted \u22641:20 (*P* \\< 0.05 across all dilutions and wounding approaches). At 1:500 dilution, the inhibitory effects of the regurgitant were negligible. On the other hand, the regurgitant did not impact the infectivity of the beetle transmissible BPMV, regardless of concentration. The results were consistent across virus/inocula combinations irrespective of leaf wounding technique implemented. Overall, \\<2% of plants inoculated with a mixture of SMV and *E*. *varivestis* regurgitant diluted \u226420-fold became infected, whereas 94% of plants of the BPMV-inoculated plants became infected.\n\nOur findings show some disagreement with that of \\[[@pone.0192003.ref011]\\]. Indeed, the authors found only gross wounding selectively inhibited infection of the non-beetle-borne *Tobacco ringspot virus* (TRSV). Inoculation by mechanical inoculation completely suppressed infection by both TRSV and the beetle-borne *Southern bean mosaic virus* at dilutions less than 1:320. The discrepancies could be attributed to deviations among studies in leaf-rub inoculation techniques, buffer composition, or could allude to some variability in specificity among Coleopteran viruses.\n\nIt appears very likely that the beetle regurgitant contains factor(s) that selectively prevent infection of plants by non-beetle-transmissible viruses. Previous work has shown the inability of virus particles to infect hosts is not due to inactivation since virus particles regained infectivity when purified away from the regurgitant \\[[@pone.0192003.ref026]\\]. This suggests that the inhibitor(s) in the regurgitant directly impact the host or alters the interactions between the virus and host in some capacity. Gergerich and coauthors \\[[@pone.0192003.ref059]\\] provided evidence that the high RNase activity in beetle regurgitant plays a role in the selective inhibition. This is substantiated by a study on *E*. *varivestis*, indicating that ribonucleases in the beetle's regurgitant may boost plant defenses and ultimately virus infection \\[[@pone.0192003.ref082]\\]. Still, we have shown beetle regurgitant is a highly complex molecular and biological substance that contains thousands of factors that could potentially influence vector-virus specificity. Moreover, some of these factors could even contribute to differences in vector competence found among beetle species and among individuals of the same species \\[[@pone.0192003.ref083]\\]. Studies have been far more numerous characterizing the salivary transcriptomes/proteomes of sap-sucking insects, and point to an exhaustive list of potential effectors \\[[@pone.0192003.ref012]--[@pone.0192003.ref021]\\]. Future studies are needed to disentangle the functional roles of the regurgitant components and how they relate to virus transmission.\n\nConclusions {#sec030}\n-----------\n\nThis study presents the first comprehensive high-throughput regurgitome of a beetle species. Leaf-feeding beetles, such as *E*. *varivestis*, deposit regurgitant onto wounded leaves during feeding. Analogous to the saliva of sap-sucking insects, it has been speculated that these oral secretions perform vital roles in the feeding process by initiating digestion and suppressing anti-herbivory host defenses. Moreover, the regurgitant is also thought to play a unique role in the remarkable specificity of beetle-transmissible viruses. Our study demonstrates that the regurgitant of *E*. *varivestis* is surprisingly complex, comprised of an impressive arsenal of putative extracellular proteins and microbes. Further, we show that the regurgitant is fundamental to the specificity of beetle-transmissible viruses. Ultimately, this study provides an exhaustive list of candidates, some of which could play important roles in plant-insect interactions and virus transmission.\n\nSupporting information {#sec031}\n======================\n\n###### SignalP information for the 1,555 *Epilachna varivestis* peptide sequences predicted to have a secretion signal.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### Functional annotation of the 992 *Epilachna varivestis* contigs predicted to have a secretion signal.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### The 1,230 bacterial species representing 577 genera found in *Epilachna varivestis* regurgitant.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n[^1]: **Competing Interests:**The authors declare no conflicts of interest.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nScience and technology are powerful engines of change, and generate heated controversies as well as high expectations. In response to this, sensitivity to societal issues, particularly within the life sciences, has increased over the years and has become a standard feature of funding programs. No less controversial than the normative and/or societal issues themselves, however, are the ways in which they are framed and the extent to which they become enlisted or delisted from various agenda's.\n\nIn Europe, two decades ago, in the context of the 4^th^ EU Framework Programme (launched in 1994), a new label was introduced to frame societal issues and to finance research, stakeholder dialogues, education and other activities to address them, namely ELSA: an acronym that stands for *ethical, legal and social aspects* of emerging sciences and technologies. The key signifier is the letter A, inserted at the end in order to distinguish it from ELSI, which refers to research on ethical, legal and social *implications* of emerging life sciences, notably human genomics, - a program funded by NIH/NHGRI in the U.S. (as part of the human genome sequencing initiative). Since then, the ELSA label has been adopted by other funding initiatives as well, notably in European partner countries, for instance the ELSA program of the Research Council of Norway and ELSAGEN (a collaboration of research funding agencies in Germany, Austria and Finland) (Chadwick and Zwart [@CR3]). Although the acronym ELSA was coined as a top-down funding mechanism, by funding agencies rather than by the research communities themselves, it nonetheless managed to evolve into a recognisable approach (Zwart and Nelis [@CR31]). In other words, rather than being an 'empty signifier', ELSA actually came to *signify* something, namely a particular research practice, with researchers using the opportunities ELSA programs offered (for instance in terms of relative proximity to large-scale life sciences research consortia) to strengthen the visibility and impact of their work. As Stegmaier ([@CR18]) phrases it, ELSA programs tend to fund activities that are geographically and organizationally close to the life sciences research programs with which they interact, serving as public and academic forums for addressing urgent societal issues arising in this context. The question whether and to what extent the research groups involved had really 'internalised' the label (seeing themselves as representatives of the ELSA-approach) has been addressed at various occasions, for instance by the *Centre for Society and Genomics* (funded by the *Netherlands Genomics Initiative*), via a workshop in September 2008 involving ELSA researchers from various countries (United Kingdom, Netherlands, Germany and Switzerland). In his report of this event, Peter Stegmaier ([@CR18]) stresses the diverse and often experimental nature of the various activities carried out under the ELSA heading. Nonetheless, a certain amount of \"convergence\" can be discerned as well, although \"the question of what is bottom-up---initiated by genomics or social researchers---and what is top-down---initiated by management or government---remains hard to answer. Initiatives come from all sides and the top-down programmes, in particular, are brought to life by researchers from the bottom-up in ways that governance could never have foreseen\" (p. 115).\n\nCurrently, and particularly in the context of recent EU funding initiatives (notably Horizon 2020), we witness a new initiative in the labelling arena, namely a shift from ELSA towards *Responsible Research and Innovation*, also known as RRI (Von Schomberg [@CR22], [@CR23], [@CR24]; Owen et al. [@CR13]). It is important to notice that, once again, this neologism is not introduced by the research field itself ('bottom-up'), but rather by science policy makers and funding agencies (notably within the European Commission) in a top-down manner, although the term seems to strike a cord and concurs with concomitant funding stratagems on the national level as well, for instance in the Netherlands, where RRI is known as MVI ('Maatschappelijk Verantwoord Innoveren'). Moreover, quite recently, a new journal, the *Journal for Responsible Innovation* has been launched by Taylor and Francis: an indicator of the readiness within the field (or at least among certain fractions within the field) to embrace/applaud (or at least adapt to) this relabeling move.\n\nIn other words, we are faced with a series of terminological shifts known as *metonymy*, to use the linguistic term for this phenomenon. A particular concept (in this case: ELSA) is no longer referred to by its usual name, but suddenly figures under a slightly different heading, one more or less associated it. Whereas ELSA always (to some extent at least) has been about safeguarding socially robust forms of fair and responsible innovation, some specific features are now more explicitly placed up front to re-frame and re-focus the approach as such. Notably, economic valorisation is given more prominence. There clearly is congruency, but a shift of focus can be discerned as well. Similar patterns of re-labelling tend to arise within life science domains themselves, of course, where labels such as 'genetics' came to be refashioned as 'genomics' (and its various derivatives), in order to reset the general focus of the field, while currently genomics is giving way to systems biology, synthetic biology, personalised medicine and other 'post-genomics' headings (not completely unlike genomics, but with slightly different emphases).\n\nAs Freud ([@CR7]) and Lacan ([@CR10]) have pointed out, however, metonymy (or 'displacement', *Verschiebung* in German) hardly ever represents a 'neutral' gesture. Quite the contrary, from a psychoanalytic viewpoint, metonymy is closely connected with (efforts to adapt to) strategies of censorship. In other words, it is by adopting this new label that the field in question hopes to be recognised as 'eligible for funding' by authoritative bodies such as funding agencies. The 'latent' research question (researcher-driven, as it were) has to be reframed in such a manner that it may fall under the heading of RRI, so as to ensure that obstacles connected with criticism of (or with biases towards) previous approaches (such as 'bioethics' or 'ELSA') may be circumvented. This type of semantic flexibility (perhaps even masquerade) comes with a price, however. It often means that some types of questions will become more important (or more difficult to ask) than others.\n\nPhrased in Lacanian terms, the question arises what this game of signifiers (notably the recent shift from ELSA to RRI) implies on the level of the 'subjects', i.e. the researchers and research groups involved. On the one hand, they depend on key signifiers to develop a more or less stable identity. On the other hand, with every new signifier appearing on the horizon, something will be gained and something will be lost. A recognisable label furnishes the subject with a discursive 'life-line', as it were. Without a credible label, the subject is left to his/her own devices and remains fragile and 'divided' (*\\$*), so that access to academic discourse and funding will be questionable and limited. The previous shift (on the level of signifiers) from ELSI (S~1~) to ELSA (S~2~) allowed the researchers involved to cleanse themselves of certain undesirable connotations connected with the ELSI label (such as the reproach of being too supportive of genomics research as such) and to develop a new identity (involving more opportunities for taking a critical stance, for instance). The question now is what is gained and what is lost with the current shift from ELSA to RRI.\n\nAs we will argue in this paper, whereas ELSA (S~2~) managed to develop a more or less recognisable profile over the years, allowing researchers to acquire a fairly established identity under this heading, the conceptualisation of RRI (S~3~) is still rather open-ended, for the time being at least. And yet it easily seems to brush the ELSA approach aside, putting the subject in a position that is both highly vulnerable and dense with opportunities. We will also argue, moreover, that this metonymy/displacement from ELSA to RRI must be regarded as a *symptom* reflecting a more fundamental shift/tension within the discourse and practice of research concerning the societal dimensions of life sciences innovations as such. Notably, within the RRI format, much more emphasis is now given to collaboration with industry and to potential socio-economic benefits of scientific and technological change. This not only pushes researchers into close proximity to their private-public 'objects' of research, but may also infect them with the aims and ideologies involved, such as: innovation, creating jobs and similar tangible socio-economic impacts. In other words, under the heading of RRI, ELSA may have lost (what was still left of its) innocence.\n\nRather than merely signalling these shifts, from a safe distance as it were, we would like to reflect on them from a position of proximity and involvement, as 'engaged scholars' so to speak. How to address the challenges involved, how to position ourselves vis-\u00e0-vis the developments outlined above? Basically, we will argue that a shift of signifiers is not an innocent game, and that losses may be involved. In other words, ELSA should not be brushed aside too quickly. Rather, (the conceptualisation of) RRI should build on the strengths of and past experiences of ELSA practices. Nonetheless, we recognise the importance of an open mind when it comes to acknowledging ELSA's weaknesses as well. We should be willing to at least consider new pathways for governance of research and innovation in the current era. At the same time, a reflection on the latent tensions reflected by this metonymy will make us more alert to some of the ideological threats and risks involved in working along the lines of the RRI concept. Thus, we will develop a preliminary assessment of RRI against the backdrop of a retrospect on ELSA, while fleshing out a number of theoretical and practical implications in this shift in grammar.\n\nAs to methodology, the focus of our paper is on (ELSA and RRI) *discourse*, rather than on *practice*, on conceptual analysis, rather than on empirical research. In other words, rather than discussing how ELSA/RRI is actually done, our paper addresses the way in which these basic approaches are fleshed out in primary and secondary sources. The focus is on the concepts, objectives and ideals, rather than on the work actually conducted under these headings. We will assess ELSA, RRI and their predecessors in chronological order. The composition of the paper, therefore, is as follows:\n\nPart I: ELSA - and its challenges: a retrospect {#Sec2}\n===============================================\n\nPredecessors of ELSA {#Sec3}\n--------------------\n\nThe assessment of societal dimensions of science and technology often involves a variety of 'third' parties besides scientists. In the past decades, these have included legal and ethical experts, NGO's, policy makers and a variety of 'publics'. In recent years, collaborative interactions with relevant stakeholder representatives were increasingly seen as a promising alternative to the 'expert' model. The various approaches emerging in this context all have their strengths and weaknesses, so it seems. Notably, finding a proper balance between embedded activity ('proximity') and autonomy (which is more easily maintained when working from a distance) is not self-evident. It is important, however, to see ELSI, ELSA and RRI (and the various discussions they raise) as efforts to respond to criticisms of previous approaches.\n\nAround 1990, existing philosophical, bioethical and Technology Assessment (TA) approaches to science and technology were increasingly seen as insufficient. Whereas philosophy was often regarded as a marginal, ivory-tower phenomenon, bio-ethics was seen by many as a mere service provider for scientists, while TA was considered to be a fairly technical field. Public engagement on the other hand seemed too open to gut-feelings, untested emotions and single issue politics. Against this backdrop, ELSA developed, using combinations of these approaches (Zwart [@CR29]). Before turning to ELSA proper, we will therefore briefly assess a number of complaints regarding these more 'traditional', pre-ELSA forms of research that (notably after 2000) gradually become 'infected' with the new model. These brief historical portrayals are not meant to be exhaustive (more comprehensive reviews are available elsewhere), but merely help to set the stage.\n\n### Philosophy {#Sec4}\n\nAs was already mentioned, one key issue in the reflection on science and technology is the issue of 'distance versus proximity'. Philosophers are often seen as preferring a position of detachment and distance towards science/technology as targets of research. Nonetheless, the plea for proximity to science has had various advocates in philosophy long before ELSA came about. One such voice has been the influential French philosopher Michel Serres ([@CR17]), who already in the 1970s argued that we live in an era of disruptive change, in science as well as in politics. As an important symptom of the transition, he mentioned the emergence of large-scale, high-tech, trans-disciplinary research fields such as molecular biology. Serres pointed out that philosophy runs the risk of losing track of these high-pace developments, thereby becoming outdated and irrelevant. Yet, according to Serres, philosophers could still play a pivotal role, provided that they acquaint themselves with these newly emerging worlds and 'oceans' of knowledge from a position of close proximity. To do so, they would have to enter the capillaries and tissues of techno-science and really become embedded scholars, addressing the various philosophical issues raised by these developments in close interaction with the scientists who are active within these fields. One important objective for philosophy is what Serres refers to as \"conceptual epidemiology\": analysing and assessing how techniques, vocabularies, research practices, forms of information and metaphors spread through research fields worldwide, infecting and inflaming the tissues of the broader societal life-world as well. This task, we believe, is as relevant now as it was in the 1970s.\n\n### Bioethics: the professional and the participatory turn {#Sec5}\n\nOne could argue that professional bioethics, as it emerged during the 1980s, tried to steer a kind of middle-course between proximity and distance (as well as between criticism and support) of (predominantly biomedical) research practices. In a well-known article entitled \"How medicine saved the life of ethics\", Stephen Toulmin ([@CR20]) argued that the involvement of professional ethicists in addressing moral dilemmas, emerging in contemporary medicine and health care, had allowed the field to revivify itself. Ethics suddenly evolved from a dull, self-centred and fairly marginal academic sub-discipline into a thriving, interdisciplinary research area. Whereas in the 1970s critical monographs written by single authors played a dominant role (with Ivan Illich's *Medical Nemesis* (1977) as the classical example), the 1980s and 1990s witnessed a process of professionalisation, exemplified not only by the emergence of ethics committees in academic hospitals, but also for instance by the launch of the journal *Bioethics* in 1986 and, more generally, by the rise of the ethical expert model. Rather than fleshing out a broad and comprehensive view of the role of medicine in contemporary culture, as Illich and others had done, professional ethicists opted for a consensus model, which entailed a focus on case studies, concrete dilemmas and principles. On the negative side, it resulted in a rather drastic delisting of the more substantial issues from the agenda of bioethical discourse. Virtually all items pertaining to 'metaphysical issues' (What is nature? What is life? What is embodiment?), 'world-views' (either secular or religiously inspired), or to the interwoveness of science, technology and society at large, tended to be regarded as being beyond the scope of ethical deliberation. Bioethics increasingly retreated to procedural aspects of ethical deliberation and to the process of weighing 'self-determination' versus 'harm to others', at the expense of broader content (Zwart [@CR28]).\n\nDuring the 1990s, ethics expertise was generally seen as an appropriate basis for advice to policy makers on ethical issues in science and technology, on national and European levels. In this capacity, it came to be institutionalised as a policy instrument rather than as an academic discipline (Tallacchini [@CR19]). A good example of the prominence of ethics as a governance tool have been the evaluation processes for EU Framework Programme research applications. Over the years, ethical review became increasingly prominent. Proposals for research, once they were evaluated positively in terms of scientific quality, were assessed in terms of their ethical quality as well, by panels of ethics experts drawn from across the EU^a^. Bioethics came to be seen as a major source for assessment, regulation and justification. As such, bioethics evolved into a normative instrument for policy and legislation (Tallacchini [@CR19]). Yet, it was especially this success of bioethics in terms of its institutionalised role which triggered a broad variety of concerns. Eckenwiler and Cohn ([@CR6]) notably voiced concerns with regard to the tension between bioethics as acting on behalf of morality, and bioethics as an institutionalised element *within* the field it purported to assess, conveying a 'technocratic style' of analysis. This gave rise to the criticism that bioethicists had become advocates and facilitators of science and technology rather than critical assessors of their societal impact.\n\nDiscontent with the professional bioethics model resulted in new developments in at least two directions, namely on the one hand the rise (during the 1990s) of the public participation paradigm, and on the other hand the coming into being of ELSA research, where bioethics was combined with empirical and participatory approaches. While public participation focussed on involving (future) stakeholders and various 'publics' (plural) in reflections on new technologies (Broerse [@CR2]), ELSA research took a slightly different course. The focus was on increased collaboration and interaction between experts from various fields: notably between researchers from the humanities and social sciences on the one hand with life scientists and life science research consortia on the other, although in practice both approaches (public participation and ELSA) could be combined in various ways, notably in the context of research that involved focus-groups with scientists, social sciences and humanities experts and (future) stakeholders (Broerse [@CR2]).\n\n### Technology assessment: the co-constructive turn {#Sec6}\n\nTechnology assessment (TA) was the dominant approach in science governance before the 1990s. Traditional TA, however, was seen by many as a rather technical form of research, conducted by science and technology experts, often with a background in science or engineering. It pretended to be neutral, but nonetheless seemed to convey an instrumentalist vision on both society and technology, seeing technologies as tools that could enable us to solve certain problems, depending on the circumstances. Gradually, the idea emerged that the concept of expertise should be broadened and that societal expertise (represented by future users) should be involved in this process. An important exemplification of the participatory turn within TA is *Constructive Technology Assessment*, developed by Arie Rip and others (Robinson [@CR15], Krabbenborg [@CR11]). In the context of CTA, the social scientists, instead of being 'experts' themselves, rather play the role of mediators, 'bridging' separate worlds by organising dedicated 'bridging events'^b^.\n\nCTA builds on the concept of 'co-evolution' between science and society that was put forward by 'science and technology studies' (STS) scholars, a branch of sociology focusing on the social shaping of emerging technoscience. Nowotny et al. for instance distinguished between 'mode-1' (disciplinary, predictive, linear) and 'mode-2' (context-driven, problem-focused, interdisciplinary) research, giving priority to the latter, which is oriented on integration of science and technology with society and vice versa. Similarly, Moor and Weckert ([@CR11]) state that ethics should already be integrated in science and technology processes in an early stage: the \"ethics first\"-approach. By this, they mean an approach to innovation in which ethical considerations (concerning both positive and negative societal impacts) are already taken into account in the design-phase of innovation trajectories (i.e. 'upstream innovation'), rather than defining moral constraints for innovation in hindsight. Many of these approaches exemplify a move from *analysing* how science and technology are shaped by societal factors to actively *supporting* ways to integrate science and technology in society.\n\nThe launch of ELSI {#Sec7}\n------------------\n\nAgainst the backdrop of discontent with existing approaches, and building on concepts such as 'mode 2', 'ethics first' and 'upstream innovation', ELSI (in the U.S.) and ELSA (in Europe) came about. ELSI came first (in 1990), while ELSA (introduced in 1994) was partly modelled on ELSI, but at the same time aimed to move beyond some of the restrictions of the ELSI format.\n\nELSI was conceived in 1988 when James Watson, at the press conference announcing his appointment as director of the Human Genome Project (HGP), suddenly and unexpectedly declared that the ethical and social implications of genomics warranted a special effort and should be directly funded by NIH (Cook-Deegan [@CR5]). Like the HGP itself, the US ELSI program was formally established in 1990. Its mission was to anticipate and address the ethical, legal, and social implications of genetic and genomic research. From 3 up to 5% of the NHGRI Research budget was devoted to this type of work. Thus, NIH all of a sudden became the largest public funder of bioethics research world-wide. Over the years, various ELSI and/or ELSA programs have been developed, in Canada,^c^ Europe and the Far East (Chadwick and Zwart [@CR3]).\n\nIn an issue of *Nature* published on November 9, 1989 (the day of the collapse of the Berlin Wall) an article entitled 'Dealing with the data' describes how HGP researchers were setting up a joint database where scientists could deposit their sequencing data^d^. Interestingly, this article was flanked by a second one entitled 'Ethical matters', containing a report of an international meeting outlining the societal hazards involved in human genome sequencing. This was a nice early example of the ELSI approach: genomics scientists on the one hand, experts from the social sciences and the humanities on the other, working side by side, as flanking endeavours, to create optimal conditions for an adequate embedding of genomics, notably of its flagship project, the HGP.\n\nFrom the very outset, like some of its predecessors (such as bioethics), ELSI has triggered scepticism and criticism from various corners. In a paper entitled *What's ELSI got to do with it?* Michael Yesley ([@CR27]) argues that, despite its unprecedented size, the ELSI program (\"the richest bioethics program in history\", p. 4) failed to have noticeable impact. Moreover, it was dedicated to \"advancing\" rather than to critically assessing genomics/genetics research. Also, although ELSI \"produced a large portfolio of academic literature primarily intended for, and communicated to, other academicians\" (p. 2, p. 5), it had little or no influence on policy. According to Yesley, an advisory committee would have done a much better job.\n\nFrom our perspective, the fairness and validity of Yesley's analysis may be questioned. The juxtaposition of 'academic research' *versus* 'policy relevance', for instance, seems unconvincing. Quite the contrary, we would argue, for how can policy be informed/improved in the absence of solid evidence-based research? Moreover, his claim that ELSI (in order to \"advance\" the scientific research it studied) has *avoided* to address more critical issues such as \"genetic reductionism\", \"biological warfare\" or the \"priority\" accorded to genomics/genetics in budget spending, is refuted by the actual ELSI output, which includes highly critical scholarly works, such as (to mention just one example) Lily Kay's (highly influential) analysis of the entanglement of molecular genetics and the militarisation of post-war American science (Kay [@CR10]). Often, such criticism has been brought forward against ELSI and ELSA in one and the same breath. Browsing through the tables of contents of typical ELSI/ELSA journals such as *New Genetics and Society* or *Life Sciences, Society & Policy*, shows that the accusation that ELSI (and/or ELSA) is only about streamlining implementation of new technologies does not hold. Nonetheless, ELSA programs tend to have a slightly different approach path (more bent on 'academic criticism') than the prototypical ELSI project, as will be more fully explained below.\n\nBirth of ELSA (or: the ELSI\u2009\u2192\u2009ELSA shift) {#Sec8}\n-----------------------------------------\n\nAs we have seen, the acronym ELSI (in the US) or ELSA (in Europe) refers to research and interaction activities that aim to anticipate and address ethical, legal and social *implications* (ELSI) or *aspects* (ELSA) of emerging life sciences, notably genomics. The shift in signifiers, from I (implications) to A (aspects) was generally seen as an effort to broaden the scope of the research (notably: to avoid the flawed linearity implied by 'implications' (Wickson et al. [@CR26])) and to launch a European alternative to the American counterpart. In the United Kingdom, this lead to a network of ELSA centres funded by the *Economic and Social Research Council* (the ESRC Genomics Network: EGN). In the Netherlands, the *Netherlands Genomics Initiative* (NGI) was established. 5% of its budget was spent on ELSA activities, on the one hand in the form of a research program issuing calls for researcher-driven, stand-alone projects (entitled: \"The societal component of genomics research\") and, on the other hand, in the form of a *Centre for Society and Genomics*, established in 2004. CSG evolved into a large-scale centre for interactive research and communication, with approximately 50 research projects designed and conducted in collaboration with the other 15 centres of the NGI genomics network. A budget of 25 million \u20ac was available for research, communication and education. Besides the substantial amount of funding involved, ELSA Genomics research in the Netherlands provided challenging opportunities for other reasons as well. It was regarded as a test-bed for a new style of doing research, as well as for setting up a new type or organisation. As a national Research Centre, CSG developed a relatively open network of trans-university collaborations for developing and conducting an interactive research programme of substantial size. Indeed, in purely quantitative terms, ELSA could be regarded as the Social Sciences and Humanities version of 'Big Science'. A key feature of the program consisted in proximity to (and collaboration with) prominent large-scale life sciences research programs. Virtually all research projects of the CSG program entailed collaborations or at least interactions with genomics research centres at various stages of the research trajectory (Zwart [@CR30]). Through these and similar initiatives, an ELSA research community, described by Stegmaier ([@CR18]) and others (for instance: Wickson et al [@CR26]), unfolded.\n\nChallenges of the ELSA approach {#Sec9}\n-------------------------------\n\nELSA was forged to amplify the strengths and forego the weaknesses of the previous approaches discussed above, in response to the on-going quest for ways to institutionalise normative assessment of science and technology in European policy processes. Point of departure was the view that scientific expertise as such cannot sufficiently guarantee the societal justification of developing and introducing new technologies. Specific strategies need to be developed, first of all in order to forego future public resistance (such as in the case of the GMO debacle), but also because a large portion of scientific research is funded by public means. Therefore, society should have a say in how the money is invested, i.e. trans-mutated into knowledge (which, hopefully, and eventually, can be trans-mutated back into money once again, through a process which is now often referred to as 'valorisation'). Moreover, society will also function as the future *consumer* of the scientific knowledge thus produced, and as the future *market* for new technological devices. Thus, it is highly relevant to be able to determine at a relatively early stage whether certain applications will be considered acceptable or usable or not, as public controversy will hinder innovation. Also, it is acknowledged that science and technology practices call for a broadening of the concept of responsibility. The internal moral codices of science should not merely involve issues of risk, safety and security, but should also include issues of social responsibility in a broader sense (such as: bringing about a more sustainable future). ELSA thus became the concept under which normative critique of science and technology was to be institutionalised. It held the two-way objective of critical assessment of on-going research on the one hand and of facilitating the future embedding of science (via co-constructive agenda-setting and policy advice, for instance) on the other.\n\nOver the years, much like its predecessors, ELSA research has been confronted with a number of internal and external challenges, or even \"pathologies\", as Wickson et al. ([@CR26]) have phrased it. To begin with, a problem of ELSA agenda-setting has been that research agendas of ELSA research were often pre-formatted by the scientific research programs they intended to study. The scientists had a say in the type of problems ELSA researchers were expected to address. This concern is of course reminiscent of the earlier discontent with regard to bio-ethics. Similar issues of pre-formatting and framing had played a role in embedded bio-ethics activities as well. Ethics advice committees, for example, were often seen as functioning within pre-defined procedures and aims: coming to consensus, finding solutions to tricky issues. This made them vulnerable to the accusation that they de-listed problematic items or 'closed' controversies over 'deeper' or 'broader' issues, as we have seen. Due to its proximity to science, and to large-scale funding initiatives for large-scale research programs, ELSA now became the target of similar concerns.\n\nAlso, and in relationship with a (real or perceived) dominance of ethics (notably medical ethics), ELSA (like its predecessor ELSI) has been accused of focussing on a limited set of issues, notably individual autonomy (and related items such as privacy) and harm or risk. Yet as science and technology, notably in the life sciences, included research in areas such as synthetic biology and nanotechnology as well, it tended to move far beyond the areas of health applications, involving issues such as bio-fuels, biomaterials, waste management and industrial bio-tech. Thus, whereas ELSA-expertise tended to remain associated with medical ethics, stressing issues of health risks and patient autonomy, other concerns and issues (naturalness, identity, global justice etc.) tended to be neglected. Overall, ELSA was often seen as an endeavour that focuses on the health domain rather than on agriculture, industry or the environment. In response to this, the ELSA community began to broaden its scope and to seek new partnerships with scientists.\n\nIn a reflection on the CSG program mentioned above, the following physiognomy of ELSA research was presented (Zwart and Nelis [@CR31]). First of all, the authors emphasise that ELSA must *not* be seen as a new 'discipline' or area of research. Rather, they argue that the ELSA acronym basically refers to a particular *style* of doing research, to a basic methodological *attitude*, one that may apply to all the disciplines involved (bio-ethics, philosophy of science, STS, TA, science communication, etc.). According to the authors, ELSA research involves:Proximity to life science researchAn anticipatory, forward-looking approach; a focus on the agenda-setting and design stages of innovation trajectories, rather than on the product stageInteraction with a broad range of societal stakeholders (media, policy, NGO, industry) as integral part of the researchInterdisciplinarity: ELSA research as a converging field involving a broad range of disciplines (philosophy of science, bioethics, social science, TA, STS, innovation studies, science communication etc.)A focus on micro-analysis (\"case studies\") rather than on macro analysis (socio-economic studies)A tendency to draw on a wide variety of sources: from academic philosophy via policy reports up to media coverage of public debates and genres of the imagination (novels, plays, movies and the like)\n\nAll these characteristics have their challenges/pathologies as well. A well-known challenge for anticipatory research, for instance, is the so-called Collingridge dilemma (Collingridge [@CR4]): although it is most effective to shape innovative technologies in a societally desirable direction at an early stage of development, it is difficult during this early stage to assess what the societal effects of the technology will be. In more advanced phases societal effects will become clearer, but there is less room for change. In other words, although impacts cannot be easily predicted until the technology is extensively developed and widely used, steering becomes more difficult when the technology has already become entrenched.\n\nAnother intricacy involves the interdisciplinarity of ELSA, notably the collaboration between philosophers/bioethicists on the one hand, and STS scholars on the other. The complexities of these collaborations were sometimes underestimated, notably in terms of the various forms of rivalry between these strands of ELSA research, not only over methods, but also over chairs, journals, program committees, funding opportunities and the like. The complexities of academic power play, also *within* ELSA, proved quite fierce on some occasions. But by far the most controversial feature of ELSA, as outlined above, has been 'proximity'.\n\nSocial sciences and humanities experts who were open to the ELSA approach have increasingly found themselves collaborating closely with scientific partners from the life sciences and from technical fields, even in the sense that they were institutionally bound to them. And this has made their work more relevant and well-informed. The question which ELSA communities continuously need to address basically is this: how to combine proximity with intellectual autonomy and societal credibility? The obvious answer is: through dialogue. Rather than undermining independence, proximity to and interaction with life scientists can make ELSA reflections, observations and criticisms more precise, more up-to-date, more targeted, and more relevant. Ideally, the scientists involved do not merely function as objects of research, as is the case in more traditional types of research, but are invited and allowed to comment on (preliminary versions of) ELSA analyses and assessments. Thus, interactive research entails a kind of empirical cycle of its own, where preliminary assessments are developed and tested in the context of critical dialogue and mutual learning. This is not only a benefit in terms of research ethics (granting the 'other side' the right to respond), but also makes ELSA analyses and assessments more robust. In the midst of efforts and disputes to address these issues, however, a new approach was suddenly heralded: RRI.\n\nPart II: the emergence of RRI {#Sec10}\n=============================\n\nWhat is RRI? And where does it come from? {#Sec11}\n-----------------------------------------\n\nThe first decade of the 21^st^ century (notably the period 2002 -- 2012) were the golden years of the ELSA approach. Fields such as technology assessment, philosophy of science, and bioethics were pushed towards a process of \"elsification\", adopting the ELSA profile outlined above (either voluntarily or in order to live up to funding prerequisites). But now, within the EU at least, ELSA as a label seems to have gone out of fashion. The term *Responsible Research and Innovation* is suddenly in vogue. In the Netherlands, for instance, funding for CSG is about to expire while the program \"Maatschappelijk Verantwoord Innoveren\" (MVI, the Dutch version of RRI) has been launched to fund collaborative projects with the private sector. In call texts issued by this program, run by the Dutch Science Council (NWO),^e^ it is explicitly pointed out that the emphasis should be on furthering socio-economic goals through partnerships with industry and private companies. The program is strongly linked to the Dutch innovation policy, aimed at strengthening 'top-sectors' of the national economy through research and innovation. Responsible innovation is *better* innovation, is the general adage, and innovation is expected to strengthen the competitiveness of core Dutch industry. Besides the fact that a rather prominent role is now given to private companies and public-private-partnerships,^f^ this also has consequences for the question which stakeholders are to be seen as the most relevant partners in societal dialogue. The most recent MVI-call (NWO [@CR12]) not only contained a rather restrictive listing of issues that were eligible for funding, but also (for every issue) a pre-fixed list of issues that research proposals were expected to address. Thus the profile of the type of research to be funded was quite explicitly pre-determined, and potential research proposals were steered in fairly specific directions: either acquiesce or perish, as it were.\n\nThe question now is: what *is* RRI and in what way does it differ from ELSA? We will argue, first of all, that at first glance RRI is *not* a radical departure from ELSA, and that, in the process of further developing the RRI approach, the inclusion of ELSA's heritage may well prove essential. Yet, we do see a new emphasis emerging in RRI in comparison with ELSA, namely the focus on socio-economic benefits and collaboration with private and industrial partners.\n\nELSA versus RRI? {#Sec12}\n----------------\n\nWhen it comes to defining (or at least profiling) RRI, it is important to keep in mind first of all that, very much like ELSA, RRI is *not* a (new) discipline or research field. Rather, as we will see, it is a basic strategy to *change* the way in which research and innovation is usually done. And this is one of the aspects which RRI shares with ELSA. Through interactive involvement, RRI initiatives propose to move agenda-setting processes in the direction of societal responsibility. Let us look into this in more detail.\n\nAn important voice in advocating RRI is Ren\u00e9 von Schomberg, a former academic with expertise in philosophy and science and technology studies (see for instance Wheale, von Schomberg, Glasner [@CR25]) who became a key figure at the Governance and Ethics Unit of the European Commission. In the opening lines of *A vision of Responsible Research and Innovation* (2013) he argues that \"RRI has become an increasingly important phrase within policy narratives, in particular in Europe, where it will be a cross-cutting issue under the prospective EU Framework Programme for Research and Innovation *Horizon 2020*\". And yet, he emphasises that \"there is no agreed definition of the concept, and approaches how it should be implemented may vary\". Indeed, the field is explicitly invited to join the debate as to what RRI (as a top-down signifier) exactly signifies.\n\nIn two recent publications on RRI (Schomberg [@CR22], [@CR24]), the following definition is proposed: *Responsible Research and Innovation is a transparent, interactive process by which societal actors and innovators become mutually responsive to each other with a view to the (ethical) acceptability, sustainability and societal desirability of the innovation process and its marketable products (in order to allow a proper embedding of scientific and technological advances in our society).* As we have seen, interaction between societal actors and innovators has been a key element within the ELSA approach as well. So, what's new? In a recent interview (2012), pressed to explain the difference between ELSA and RRI more clearly, Von Schomberg argues that RRI sees ethics \"as a stimulus, not as an obstacle\". Right at the start of the interview, he is asked whether RRI is really a new approach, or mainly a new label for ELSA? In his reply, Von Schomberg explains the specificity of RRI as follows: \"What's new about RRI is that we no longer see the ethical aspects of new technologies as constraints, as restrictions. Instead, we look at the aims of technology development. Which positive contributions do you wish to obtain from research and innovation? This positive basic attitude is an important difference in comparison with the ELSA approach.\" (p. 16) Further on, he argues that RRI rebels against the more traditional approach (which was focussing on the question whether a development has undesired effects), but rather uses the possible positive contributions of (the development of) a technology as an assessment criterion. Moreover, \"unlike ELSA, RRI considers the entire innovation process, from research and development to production and distribution\" (p. 16).\n\nThis downgrading of ELSA to a set of ethical constraints meant to curb negative outcomes and undesired effects of innovation, seems difficult to reconcile with the type of activities that have actually evolved under the ELSA label, as explained above. And similar to RRI, ELSA already had the ambition to become involved in innovation trajectories at a relatively early stage ('anticipatory' research). In another paper (Von Schomberg [@CR22]), however, some additional indications concerning the specificity of RRI compared to ELSA are given. Among the typical features of RRI are now mentioned the use of ethics as a *design principle* for technology (for example: privacy through design) as well the ensuring of market accountability through standards, certification, accreditation and labels as a new form of governance to manage the floods of products coming to the market.\n\nFrom our perspective, the idea that ELSA focusses solely on 'negative' consequences and functions basically as a constraint, notably at the 'end' stage of the innovation process, is a caricature, clearly at odds with the way ELSA-type research has actually evolved / has functioned in practice. In such cursory comparisons, important key aspects of ELSA, such as anticipation and proximity (see above), are conveniently left out of the picture. Yet, it is clear that some new elements or at least emphases are introduced as well. These are conveyed by the use of terms that are not absent, but relatively rare in ELSA discourse (which has a predilection for studying 'soft' rather than 'hard' impacts), such as 'design', 'markets' and 'accreditation'. It is in this direction, we believe, that the relative newness of the RRI-approach can be found. And that is why the term metonymy is used to specify this shift: there is a clear congruency between the ELSA and the RRI approach, but a number of items that used to be marginal in ELSA are now pushed to a much more prominent position in RRI.\n\nMeanwhile, Von Schomberg's phrasings reveal much about how science policy in the EU has evolved and is progressing (Blume [@CR1], Rodr\u00edgueza et al. [@CR16]). Today, science is expected to improve the functioning of society by creating innovations. These innovations are needed to maintain affluent living standards in the face of globalization, i.e. the (possible) erosion of the competitiveness of European business, as well as the need to enhance sustainability (not only in an environmental, but also in a socio-economic sense) and other 'grand challenges'. These innovations are not only to be implemented on the micro-level, so that companies can make a profit and create jobs, but also at the macro-level, leading to system-wide transformations of important industrial sectors and, ultimately, of the European economy as such.\n\nIn this sense, it comes as no surprise that Von Schomberg ([@CR24]) explicitly links research, and particularly research funding for universities, to tackling the grand challenges facing European societies. These challenges require citizens, researchers and entrepreneurs to channel the change-engine of science towards the common good. For Von Schomberg, the EU Treaty provides the 'normative anchors' for this process: sustainable development, competitive social market economy, full employment and social progress, protection and improvement of our environment, and no social exclusion (social justice added to quality of life). Yet, these normative anchors need to be taken as positive triggers for innovation, not as negative constraints. For instance, in the case of possible implementations of these elements one needs to ask what would be the 'right impacts', as well as what would be the 'right processes' to realise such impacts.\n\nThus, although some important elements of the RRI approach may sound very much like ELSA, the overall framing is different and linked to innovation and to addressing grand (socio-economic) challenges. In turn, the framing gives much more weight and urgency to the matter of channelling science to the common good. Indeed, the type of innovation needed, and the way these innovations are produced, have (according to Von Schomberg) changed. Nineteenth-century inventor-entrepreneurs created Frankensteinian monsters by themselves, but today, networks of university researchers, companies, pressure groups and individuals may produce innovations (responsible or otherwise) by diffusing accountability, while the consequences can be far more substantial and far-reaching. From a macro-perspective, moreover, society may need forms of innovation which private companies may not bring about by themselves because it is risky and unprofitable (or otherwise makes no sense for them to pursue). Getting a grip on the societal dimensions of innovation - i.e. not just market needs but also the ethical, legal and policy issues involved - could make innovation much more effective. Hence: we need RRI.\n\nSource 2: options for strengthening RRI {#Sec13}\n---------------------------------------\n\nAnother source for addressing the question *What is RRI?* is a recent report edited by Jeroen van den Hoven (Chair of the Dutch MVI programme committee) and other prominent voices in the RRI discourse, and published by the European Commission. Again, as we will show, although a close reading reveals a remarkable inability of the authors to flesh out the RRI concept in a precise and convincing manner (notably in contrast to existing approaches), the shift towards innovation and socio-economic development is nonetheless clear.\n\nInitiatives falling under the heading of RRI are bent on (Hoven et al. [@CR21], p. 12):considering societal needs and ethical aspects in research funding programs, e.g. through public and stakeholder dialogue;developing criteria for the early appraisal of research and innovation, e.g. technology assessments;establishing processes to better integrate societal needs in research and innovation, e.g. trans-disciplinary approaches in sustainability science;setting up advisory bodies such as councils on ethical aspects of new technologies.\n\nFor anyone familiar with the field, it is obvious that this is more or less exactly what ELSA envisioned to achieve as well. The same applies to the tactics the authors suggest for promoting this type of research through funding strategies:\n\n\"One way to promote RRI is to mainstream RRI in the existing funding programmes. In this case no new funding opportunities on RRI would be allocated, but criteria for RRI would have to be applied across all EU funding programmes. This would not only raise awareness for RRI but also create greater transparency with regard to the provisions for taking into account societal needs and ethical aspects in the proposed research.\" (p. 38/39)\n\nAnd this is how RRI is operationalised by the authors of the report: RRI is anticipatory, inclusive, reflexive and responsive:Anticipatory: Anticipation asks researchers and innovators to include new perspectives in the research and innovation process and to think through various possibilities to be able to design socially robust agendas for risk research and risk management.Inclusive: Inclusiveness asks researchers and innovators to involve diverse stakeholders (such as users, NGOs, etc.) in the process to broaden and diversify the sources of expertise and perspectives.Reflexive: Reflexivity asks researchers and innovators to think about their own ethical, political or social assumptions to enable them to consider their own roles and responsibilities in research and innovation as well as in public dialogue. Reflexivity should raise awareness for the importance of framing issues, problems and the suggested solutions.Responsive: If research and innovation claim to be responsible, if it has the capacity to change its direction or shape when it becomes apparent that the current developments do not match societal needs or are ethically contested. Hence, responsiveness refers to the flexibility and capacity to change research and innovation processes according to public values. (p. 58)\n\nAgain, this is very much how ELSA has been developed and defined during the past decade.\n\nAnd yet, the difference is there. In order to see it, however, we must shift the focus of attention from the description of the general methodological profile of RRI (which sounds very much like ELSA) to the broader, socio-economic context. Much like Von Schomberg ([@CR24]), the report frames the need for societal involvement and assessment in terms of innovation and competitiveness. Notably, reference is made to \"...the ambition of the European Union to ensure that research and innovative ideas can be turned into products and services that create jobs and prosperity, as well as help preserve the environment and meet the societal needs of Europe and the world (p. 11).\" The point of RRI is to help achieve this ambition: \"RRI has the potential to make research and innovation investments more efficient, while at the same time focusing on global societal challenges.\" (p. 16) Inclusion of ethics beforehand, it seems, will lead to less contestation of innovations afterwards.\n\nSo, what is really new? {#Sec14}\n-----------------------\n\nWe end up with a remarkable 'symptom': RRI is presented as something new, as a break with the ELSA past, and yet, upon closer inspection, RRI is defined in ways that tend to resemble the ELSA stratagem quite closely. Key proponents of RRI apparently struggle to articulate their own innovation in a convincing way.\n\nThis symptomatic failure (*Fehlleistung*) may be seen as a result of the disregard towards ('repression' of) the ELSA heritage, followed by the subsequent inevitable 'return of the repressed', namely the resurge of recognisable ELSA-speak in the definitions given of RRI. At the same time, however, it points to the fact that the difference is not in the methods or approaches used as such (as is sometimes suggested, notably by stressing the need for anticipation and interaction), but rather in the *finality* of RRI, the overall aim to be achieved, namely: ensuring that the EU economy remains internationally competitive and robust.\n\nVery much like ELSA, RRI builds on existing tendencies in academia and policy to integrate societal research and interaction into science and technology practices. The confusion concerning the question what RRI 'is' can only be overcome when, in elaborating the RRI concept, both the continuity (with ELSA) and the discontinuity (the newness of RRI compared to ELSA) is fleshed out with more precision. And the one is a prerequisite for the other, because it is only insofar as the value of ELSA's heritage is duly acknowledged, that the 'added value' of RRI can be articulated with sufficient precision.\n\nIn other words, rather than presenting RRI as a complete *novum*, while debunking/ignoring the ELSA experience, we argue that, in terms of proximity, co-creation, stakeholders participation and the like, RRI should build on, rather than break with, the ELSA approach. Proximity and interaction as such are not new (Rodr\u00edgueza et al. [@CR16]) and much can be learned from the ELSA heritage on whose accomplishments RRI could be engrafted. Only then can it be pointed out more clearly in what way RRI differs from its predecessor.\n\nIt is clear that RRI departs from ELSA most clearly in its focus on economic innovation. RRI is supposed to help research to move from bench to market, in order to create jobs, wealth and well-being. In turn, this focus on innovation has a number of important consequences. Micro-level case studies of knowledge production and innovation processes (so dominant in ELSA) will not be eliminated we expect, but will be complemented by mid- and marco-level studies of transformations and transitions, to bring the broader socio-economic context into view. Substantive normative questions also come into play at this level, particularly the question what kind of society and economy we want (which, of course, includes issues of sustainability and fairness). Private companies will be involved in RRI (being less prominent in ELSA). This will entail a shift of focus from analysing knowledge production to processes of co-design in innovation and public-private interaction.\n\nAs a consequence of this, RRI also entails a shift in terms of disciplines. Whereas ELSA was a collaboration between disciplines such as philosophy of science and/or technology, bioethics, STS, CTA, interactive science communication, and the like, RRI requires intense collaboration with a 'third' strand of research, namely management and innovation studies, fields that, broadly speaking, investigate how innovations come about, how they are managed and how policy affects them. RRI could profit from including innovation studies as a body of knowledge and/or as a scholarly community, to link the micro- with the macro-level and to make the results of research more relevant for policy makers.\n\nThus, we expect ELSA to partly merge into RRI, so that *in practice*, RRI will take the viable parts of the ELSA heritage more seriously than is done in printed form. For the ELSA community, RRI may increase the relevance of ELSA, as this type of research may now become part of broader transitory processes. New types of partnerships, notably with industry will evolve, urging the introduction of new research techniques (for instance adopted from innovation studies). However, the ELSA\u2009\u2192\u2009RRI metonymy also carries a risk: the envisioned merger of the ELSA-agenda (bent on promoting morally justified research) with innovation and industrial agendas may lead to silencing of critical, 'rogue' voices and outsiders in the debate, due to increased dependency on private sector parties and policy agendas. In that way, RRI may be faced with an unexpected result: it may be exactly the outsiders, i.e. small firms, small groups of researchers, interdisciplinary projects at the cutting edge of various disciplinary silos, that create the radical innovation we need to transform the EU economy and address the challenges it faces.\n\nPotential problems with the integration of science and technology under RRI {#Sec15}\n---------------------------------------------------------------------------\n\nOne important question is *how* an RRI approach could enhance innovation, particularly the beneficial effects of innovation (wealth, jobs, competitiveness, etc.). The RRI label suggests that some innovations will be selected over others, namely the 'responsible' ones. This is hard work in any kind of setting, particularly so when it is unclear as to exactly *what* a 'responsible innovation' is and *who* will be the one to determine this. It poses tough questions at the micro-level (what to do in the context of a particular RRI project?), at the meso-level (what kind of projects to select and reward?) and at the macro-level (what types of research should we prioritise?). EU funding institutions, as well as the values they represent, are currently targets of dispute (to put it mildly). Which interpretation of responsibility are we to pursue? Relying on 'process' and 'dialogue' runs the risks of missing the mark and delivering responsible innovations that fail to address the grand challenges on a macro-scale. In the end, RRI revolves around a crucial issue: what kind of innovations and what kind of economy do we want? And who will be the 'we'?\n\nELSA researchers (who already ran the risk of *going native* due to their proximity to and involvement with science programs) will (upon entering the promised land of RRI) become increasingly *co-responsible* for the innovations they help to develop. One important aspect of co-responsibility is *promise management*. Promises and expectations have always played an important role in science and technology development. They are currently facing a credibility deficit, a credibility crisis even, due to years of overpromising, i.e. the practice of recycling stereotypical lists of promises time and again for various purposes, until the public sphere becomes saturated with faltered promises and extended time-schedules. A closer involvement of societal stakeholders and future users in the innovation pathways implies that promisory discourse will be opened up to them as well, so that promises become co-constructed as well. At least implicitly, but often explicitly, 'promise management' becomes part the projects and activities developed. What promises have been made, to what extent have they been met; has the promised relevance been achieved? This is the handiwork of \"promisomics\" (Chadwick and Zwart [@CR3]), i.e. the co-constructive framing, monitoring and management of promises, counteracting fabulation and making promisory discourse more convincing, reliable and relevant. How can we frame promises that are reliable, plausible and socially robust? How can we assure that they concord with the needs and expectations of society?\n\nInvolving societal input in science and technology innovation implies that RRI- researchers become part of the very processes they study, immersing themselves in research consortia whose work they claim to critically assess. The tension between 'going native' and giving voice to critical concerns is there to stay. For the ELSA community, the liaison (or even 'marriage' with) RRI will bring up an issue that is uncannily reminiscent of the one that haunted Elsa of Brabant, the female lead in Wagner's *Lohengrin*: what will be the questions we are no longer supposed to (but will find impossible not to) ask?\n\nEndnotes {#Sec16}\n========\n\n^a^The prominence of the Ethics Review process can be seen from the FP7 web-portal pages on Ethics: <http://cordis.europa.eu/fp7/ethics_en.html> (last visited 14 May, 2012).\n\n^b^As Krabbenborg (2013) argues, a CTA researcher can move around in different worlds, make observations, ask questions, and point out problematic issues that are more difficult to notice for actors who operate from a particular disciplinary perspective. Notably, 'bridging events' can provide an opportunity for actors to freely explore emerging patterns, articulate problematic issues and develop strategies for how to deal with it. Thus, in the context of CTA, the focus shifted from expert reports to early stage preparatory and anticipatory interaction with stakeholders.\n\n^c^The Canadian acronym (GE^3^LS) added the G (for genomics) but dropped the postfix A/I.\n\n^d^Nature, 342 (6246), 9 November 1989, p. 108; doi:10.1038/342108a0.\n\n^e^Ironically, NWO, established in 1950 as ZWO (*Netherlands Organisation of Pure Scientific Research*, with the Z referring to 'pure' (i.e. fundamental) research - but this signifier was dropped from the acronym in 1988 in order to indicate the broadening of the mission), was traditionally focussed on funding 'basic' rather than 'applied' research activities.\n\n^f^Collaboration with industry was (perhaps) less prominent, but far from absent in ELSA, of course, see for instance: Penders and Nelis ([@CR14]).\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n\nAll authors read and approved the final manuscript.\n\nThis paper is based on an invited lecture presented at a Conference entitled *The road ahead for ELSA in Norway: Issues of quality, influence and network cooperation* (ELSA Conference Oslo, 4-5 December 2012), organised by Ellen-Marie Forsberg, Oslo and Akershus University College of Applied Sciences / The Research Council of Norway. It presents the results of research activities funded by the *Centre for Society and the Life Sciences* (CSG, which is part of the research network of the *Netherlands Genomics Initiative*, NGI) and by the MVI program of the Netherlands Organisation for Scientific Research (NWO).\n"}
+{"text": "I read with great interest the article by Toyota and colleagues published in PLOS ONE entitled, \"Short versus prolonged dual antiplatelet therapy (DAPT) duration after coronary stent implantation: A comparison between the DAPT study and 9 other trials evaluating DAPT duration\".\\[[@pone.0184513.ref001]\\] The authors compare results of the Dual Antiplatelet Therapy Study (DAPT Study) to pooled results of 9 other randomized trials evaluating duration of DAPT. The analysis concludes that the beneficial effects of prolonged DAPT in reducing rates of myocardial infarction (MI) and stent thrombosis (ST) seen within the DAPT Study were greater than in the other pooled trials, yet the magnitude of excess risk associated with prolonged DAPT was consistent across studies. In part, the authors suggest that this discrepancy may be due to higher rates of MI and ST within the control arm of the DAPT Study compared to the control arms of the other studies.\n\nWhile the desire to further understand the risks and benefits of prolonged DAPT holds merit, several points need to be considered when interpreting the presented results. First and foremost, the DAPT Study is a large, international, multicenter, randomized, double-blinded, placebo-controlled trial powered specifically to detect the impact of prolonged DAPT on the co-primary efficacy endpoints of definite or probably stent thrombosis and of major cardiovascular and cerebrovascular events.\\[[@pone.0184513.ref002]\\] The DAPT study therefore represents the gold standard of clinical experimental evidence. The comparator, on the other hand, is a pooled analysis of 9 heterogeneous randomized trials, none of which was statistically powered to study MI or ST rates. Each of the studies, with the exception of ISAR-SAFE,\\[[@pone.0184513.ref003]\\] were open label trials without placebo control, raising the possibility of inherent bias. Moreover, only 2 of the 9 studies (DES LATE and ARCTIC-Interruption) evaluated the clinical impact of prolonged DAPT (18 to 36 months) relative to the standard 12 months of therapy,\\[[@pone.0184513.ref004], [@pone.0184513.ref005]\\] as did the DAPT Study.\\[[@pone.0184513.ref002]\\] The other 7 trials, in contrast, focused on the safety of early DAPT discontinuation (3 or 6 months) after percutaneous coronary intervention (PCI) compared to 12 months of therapy. The authors' analysis assumes that the magnitude of risk reduction remains constant from 3 to 36 months after PCI, an assumption that is not supported by the literature. Within a recent analysis from the DAPT study, rates of MI and ST and the relative risk reduction seen with prolonged DAPT 15--30 months after PCI were approximately half of those 12--15 months after PCI.\\[[@pone.0184513.ref006]\\]\n\nThe authors posit that the robust reduction in ST and MI risks observed with prolonged therapy within in the DAPT Study is due to higher baseline rates of MI and ST, therefore insinuating that the DAPT Study cohort is a higher risk cohort that may benefit of more prolonged antiplatelet therapy. This point is further supported by the fact that a greater proportion of subjects within the DAPT study had undergone prior coronary revascularization prior to study enrollment, although subgroup analyses of the DAPT Study actually demonstrate numerically greater risk reduction with prolonged DAPT in those without prior revascularization.\n\nRegardless of the aforementioned points, I agree with the authors that the DAPT duration should be personalized to target prolonged therapy in those at greatest risk for ischemic events and abbreviated therapy in those at high risk of bleeding. Short-duration DAPT appears to be acceptably safe after contemporary DES use. In patients with hemorrhagic complications, at high risk of bleeding, or needing chronic anticoagulation, a short-duration DAPT is prudent, assuming patients are not concurrently at high risk for ischemic complications. In the absence of bleeding events or risk, the default duration of DAPT should be 12 months. Beyond 12 months of follow-up, the balance of hemorrhagic and ischemic risks should be reconsidered. Based on the proven efficacy and safety of prolonged DAPT,\\[[@pone.0184513.ref002], [@pone.0184513.ref007]\\] DAPT should preferably continue to 30 months to minimize risk of ST and non-stent-related MI, especially in those with recurrent ischemic events or at high risk of subsequent events. The recently developed DAPT score may provide further guidance for patients that do not fall within the extremes of risk.\\[[@pone.0184513.ref008]\\] Using five clinical factors and three index procedural characteristics (score range, -2 to 10), the score estimates the balance between ischemic and bleeding risks in patients who have remained free of ischemic or bleeding events 1 year after DES. Patients with scores \\< 2 possess a risk of bleeding that outweighs ischemic risk; whereas, with scores \u2265 2, ischemic risk outweighs bleeding risk. The DAPT score therefore answers the call by Toyota and colleagues by providing a means to personalize DAPT to fit individual patients.\n\n[^1]: **Competing Interests:**SE is the first author for the following reference within the Toyota manuscript: Reference \\#16: Elmariah S, Mauri L, Doros G, Galper BZ, O\\'Neill KE, et al. (2014) Extended duration dual antiplatelet therapy and mortality: a systematic review and meta-analysis. The Lancet 28: 792-798.\n"}
+{"text": "Related literature {#sec1}\n==================\n\nFor background to the synthesis of *N*-(*p*-chloro\u00adphen\u00adyl)acetamidine and related *N*-aryl\u00adacetamidines used as reagents in the formation of anti-leishmanial compounds, see: Shearer *et al.* (1997[@bb10]); Rousselet *et al.* (1993[@bb9]); Patai (1975[@bb8]). For background to leismaniasis, see: Ouellette *et al.* (2004[@bb7]); Croft *et al.* (2006[@bb2]); Ferreira *et al.* (2007[@bb4]); World Health Organization (2010[@bb14]).\n\nExperimental {#sec2}\n============\n\n {#sec2.1}\n\n### Crystal data {#sec2.1.1}\n\nC~8~H~9~ClN~2~*M* *~r~* = 168.62Orthorhombic,*a* = 9.6460 (9) \u00c5*b* = 9.0192 (4) \u00c5*c* = 19.3281 (5) \u00c5*V* = 1681.53 (18) \u00c5^3^*Z* = 8Mo *K*\u03b1 radiation\u03bc = 0.39 mm^\u22121^*T* = 120 K0.35 \u00d7 0.20 \u00d7 0.10 mm\n\n### Data collection {#sec2.1.2}\n\nNonius KappaCCD area-detector diffractometerAbsorption correction: multi-scan (*SADABS*; Sheldrick, 2007[@bb11]) *T* ~min~ = 0.792, *T* ~max~ = 1.00014006 measured reflections1924 independent reflections1185 reflections with *I* \\> 2\u03c3(*I*)*R* ~int~ = 0.081\n\n### Refinement {#sec2.1.3}\n\n*R*\\[*F* ^2^ \\> 2\u03c3(*F* ^2^)\\] = 0.047*wR*(*F* ^2^) = 0.150*S* = 1.051924 reflections107 parametersH atoms treated by a mixture of independent and constrained refinement\u0394\u03c1~max~ = 0.34 e \u00c5^\u22123^\u0394\u03c1~min~ = \u22120.33 e \u00c5^\u22123^\n\n {#d5e476}\n\nData collection: *COLLECT* (Hooft, 1998[@bb5]); cell refinement: *DENZO* (Otwinowski & Minor, 1997[@bb6]) and *COLLECT*; data reduction: *DENZO* and *COLLECT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb12]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb12]); molecular graphics: *ORTEP-3* (Farrugia, 1997[@bb3]) and *DIAMOND* (Brandenburg, 2006[@bb1]); software used to prepare material for publication: *publCIF* (Westrip, 2010[@bb13]).\n\nSupplementary Material\n======================\n\nCrystal structure: contains datablocks global, I. DOI: [10.1107/S1600536810011013/hg2664sup1.cif](http://dx.doi.org/10.1107/S1600536810011013/hg2664sup1.cif)\n\nStructure factors: contains datablocks I. DOI: [10.1107/S1600536810011013/hg2664Isup2.hkl](http://dx.doi.org/10.1107/S1600536810011013/hg2664Isup2.hkl)\n\nAdditional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?hg2664&file=hg2664sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?hg2664sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?hg2664&checkcif=yes)\n\nSupplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [HG2664](http://scripts.iucr.org/cgi-bin/sendsup?hg2664)).\n\nThe use of the EPSRC X-ray crystallographic service at the University of Southampton, England, and the valuable assistance of the staff there is gratefully acknowledged. JLW acknowledges support from CAPES and FAPEMIG (Brazil).\n\nComment\n=======\n\n*N*-(*p*-Chlorophenyl)acetamidine and related *N*-arylacetamidines (Shearer *et al.* 1997; Rousselet *et al.* 1993; Patai, 1975) were synthesized for use as reagents in the formation of 5-(difluoromethyl)-2-methyl-1-(substituted-phenyl)-1*H*-imidazoles, which are active anti-leishmanial compounds (Ferreira *et al.*, 2007). Leishmaniasis is caused by several species of protozoan parasites transmitted by the bite of the female phlebotomine sand fly. This neglected disease is currently prevalent in four continents, being endemic in 88 countries, 72 of which are developing countries, threatening 350 millions worldwide (World Health Organization, 2010). The treatment of Leishmaniasis, currently, is dependent on old and highly toxic drugs (Croft *et al.*, 2006). In addition, the development of clinical resistance and the increase of co-infections leishmaniasis AIDS, in some countries is causing further worries. Thus, the development of new, efficient, and safe drugs for the treatment of this disease is imperative (Ouellette *et al.*, 2004; Croft *et al.*, 2006; Ferreira *et al.*, 2007). This contribution describes the synthesis and crystallographic characterisation of an *N*-(*p*-chlorophenyl)acetamidine derivative, (I).\n\nThe molecular structure of (I), Fig. 1, is twisted about the C1--N1 bond as seen in the value of the C2--C1--N1--C7 torsion angle of -118.6 (2) \u00b0; the dihedral angle formed between the benzene ring and ethanimidamide residue is 66.54 (14) \u00b0. The molecule has approximate mirror symmetry with the non-hydrogen atoms of the ethanimidamide lying on the putative plane and the benzene ring being bisected by the plane. The conformation about the C7\u2550 N1 double bond \\[1.299 (3) \u00c5\\] is *Z*.\n\nThe crystal packing is dominated by N--H\u00b7\u00b7\u00b7N and N--H\u00b7\u00b7\u00b7Cl hydrogen bonding interactions, Table 1. These lead to the formation of 22-membered {\u00b7\u00b7\u00b7HNH\u00b7\u00b7\u00b7ClC~4~NCNH\u00b7\u00b7\u00b7ClC~4~N\u00b7\u00b7\u00b7HNCN}~2~ synthons that are connected into supramolecular arrays in the *ac* plane, Fig. 2; these have a zig-zag topology.\n\nExperimental {#experimental}\n============\n\nTo a stirred solution of *p*-chloroaniline (10.75 mmol) in acetonitrile (40 ml) was bubbled hydrogen chloride. A precipitate was formed immediately. The resulting suspension was refluxed and became homogeneous. Upon complete reaction, as shown by TLC, the mixture was rotary evaporated and the residue partitioned between CH~2~Cl~2~ and saturated aqueous NaHCO~3~. The aqueous layer was washed (3 times) with CH~2~Cl~2~, and the combined organic layers were dried over sodium sulfate, filtered, and the filtrate concentrated under reduced pressure to yield a white solid; yield 96%, m.p. 389--390 K. The sample used in the X-ray study was slowly grown from an ethanol solution of (I). IR (KBr, cm^-^1): 3451, 3295, 3079, 1640, 1586, 1482. ^1^H NMR (500 MHz, CDCl~3~): \u03b4 1.99 (*s*, 3H, CH~3~); 4.53 (*br s*, 2H, 2); 6.77 (*d*, 2H, J = 8.0 Hz); 7.24 (*d*, 2H, J = 8.0 Hz) p.p.m. ^13^C NMR (125 MHz, CDCl~3~): \u03b4 21.59 (CH~3~); 122.5 121.1; 128.6; 144.6; 155.3 (H~2~N---C=N) p.p.m. EI---MS (m/z): 168 (M^+^, 68%); 153 (M^+^-15, 38%); 127 (M^+^-41, 100%); 111 (M^+^ -57, 54%); 75 (M^+^-93, 42%).\n\nRefinement {#refinement}\n==========\n\nThe C-bound H atoms were geometrically placed (C--H = 0.95--0.98 \u00c5) and refined as riding with *U~iso~*(H) = 1.2-1.5*U~eq~*(C). The positions of the N--H atoms were refined with *U~iso~*(H) = 1.2*U~eq~*(N).\n\nFigures\n=======\n\n![The molecular structure of (I) showing the atom-labelling scheme and displacement ellipsoids at the 50% probability level.](e-66-0o958-fig1){#Fap1}\n\n![A view of a supramolecular array in (I) in the ac plane. The N--H\u00b7\u00b7\u00b7N and N--H\u00b7\u00b7\u00b7Cl hydrogen bonding interactions are shown as orange dashed lines. Colour code: Cl, cyan; N, blue; C, grey; and H, green.](e-66-0o958-fig2){#Fap2}\n\nCrystal data {#tablewrapcrystaldatalong}\n============\n\n  ------------------------- ---------------------------------------\n  C~8~H~9~ClN~2~            *F*(000) = 704\n  *M~r~* = 168.62           *D*~x~ = 1.332 Mg m^\u22123^\n  Orthorhombic, *Pbca*      Mo *K*\u03b1 radiation, \u03bb = 0.71073 \u00c5\n  Hall symbol: -P 2ac 2ab   Cell parameters from 2182 reflections\n  *a* = 9.6460 (9) \u00c5        \u03b8 = 2.9--27.5\u00b0\n  *b* = 9.0192 (4) \u00c5        \u00b5 = 0.39 mm^\u22121^\n  *c* = 19.3281 (5) \u00c5       *T* = 120 K\n  *V* = 1681.53 (18) \u00c5^3^   Block, colourless\n  *Z* = 8                   0.35 \u00d7 0.20 \u00d7 0.10 mm\n  ------------------------- ---------------------------------------\n\nData collection {#tablewrapdatacollectionlong}\n===============\n\n  --------------------------------------------------------------- --------------------------------------\n  Nonius KappaCCD area-detector diffractometer                    1924 independent reflections\n  Radiation source: Enraf Nonius FR591 rotating anode             1185 reflections with *I* \\> 2\u03c3(*I*)\n  10 cm confocal mirrors                                          *R*~int~ = 0.081\n  Detector resolution: 9.091 pixels mm^-1^                        \u03b8~max~ = 27.5\u00b0, \u03b8~min~ = 3.0\u00b0\n  \u03c6 and \u03c9 scans                                                   *h* = \u221211\u219212\n  Absorption correction: multi-scan (*SADABS*; Sheldrick, 2007)   *k* = \u221211\u21929\n  *T*~min~ = 0.792, *T*~max~ = 1.000                              *l* = \u221225\u219221\n  14006 measured reflections                                      \n  --------------------------------------------------------------- --------------------------------------\n\nRefinement {#tablewraprefinementdatalong}\n==========\n\n  ------------------------------------- ------------------------------------------------------------------------------------\n  Refinement on *F*^2^                  Primary atom site location: structure-invariant direct methods\n  Least-squares matrix: full            Secondary atom site location: difference Fourier map\n  *R*\\[*F*^2^ \\> 2\u03c3(*F*^2^)\\] = 0.047   Hydrogen site location: inferred from neighbouring sites\n  *wR*(*F*^2^) = 0.150                  H atoms treated by a mixture of independent and constrained refinement\n  *S* = 1.05                            *w* = 1/\\[\u03c3^2^(*F*~o~^2^) + (0.081*P*)^2^\\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3\n  1924 reflections                      (\u0394/\u03c3)~max~ = 0.001\n  107 parameters                        \u0394\u03c1~max~ = 0.34 e \u00c5^\u22123^\n  0 restraints                          \u0394\u03c1~min~ = \u22120.33 e \u00c5^\u22123^\n  ------------------------------------- ------------------------------------------------------------------------------------\n\nSpecial details {#specialdetails}\n===============\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Geometry. All s.u.\\'s (except the s.u. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell s.u.\\'s are taken into account individually in the estimation of s.u.\\'s in distances, angles and torsion angles; correlations between s.u.\\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell s.u.\\'s is used for estimating s.u.\\'s involving l.s. planes.\n  Refinement. Refinement of *F*^2^ against ALL reflections. The weighted *R*-factor wR and goodness of fit *S* are based on *F*^2^, conventional *R*-factors *R* are based on *F*, with *F* set to zero for negative *F*^2^. The threshold expression of *F*^2^ \\> 2\u03c3(*F*^2^) is used only for calculating *R*-factors(gt) etc. and is not relevant to the choice of reflections for refinement. *R*-factors based on *F*^2^ are statistically about twice as large as those based on *F*, and *R*- factors based on ALL data will be even larger.\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nFractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (\u00c5^2^) {#tablewrapcoords}\n==================================================================================================\n\n  ----- -------------- ------------- -------------- -------------------- --\n        *x*            *y*           *z*            *U*~iso~\\*/*U*~eq~   \n  Cl1   0.43525 (8)    0.20450 (7)   0.43084 (3)    0.0430 (3)           \n  N1    0.36389 (19)   0.4659 (2)    0.71058 (10)   0.0308 (5)           \n  N2    0.1234 (2)     0.4244 (2)    0.70948 (11)   0.0305 (5)           \n  H1N   0.044 (3)      0.438 (3)     0.7313 (13)    0.037\\*              \n  H2N   0.124 (3)      0.379 (3)     0.6740 (14)    0.037\\*              \n  C1    0.3774 (2)     0.4032 (3)    0.64327 (12)   0.0274 (6)           \n  C2    0.4543 (2)     0.2746 (3)    0.63454 (14)   0.0316 (6)           \n  H2    0.4938         0.2272        0.6738         0.038\\*              \n  C3    0.4741 (3)     0.2144 (3)    0.56933 (13)   0.0317 (6)           \n  H3    0.5266         0.1260        0.5638         0.038\\*              \n  C4    0.4169 (2)     0.2837 (3)    0.51261 (13)   0.0286 (6)           \n  C5    0.3428 (3)     0.4144 (3)    0.51965 (12)   0.0342 (6)           \n  H5    0.3054         0.4627        0.4802         0.041\\*              \n  C6    0.3243 (3)     0.4736 (3)    0.58491 (13)   0.0348 (6)           \n  H6    0.2745         0.5638        0.5901         0.042\\*              \n  C7    0.2408 (2)     0.4770 (3)    0.73731 (13)   0.0269 (6)           \n  C8    0.2239 (3)     0.5538 (3)    0.80540 (14)   0.0367 (6)           \n  H8A   0.3153         0.5706        0.8261         0.055\\*              \n  H8B   0.1680         0.4919        0.8364         0.055\\*              \n  H8C   0.1774         0.6492        0.7983         0.055\\*              \n  ----- -------------- ------------- -------------- -------------------- --\n\nAtomic displacement parameters (\u00c5^2^) {#tablewrapadps}\n=====================================\n\n  ----- ------------- ------------- ------------- -------------- -------------- --------------\n        *U*^11^       *U*^22^       *U*^33^       *U*^12^        *U*^13^        *U*^23^\n  Cl1   0.0537 (5)    0.0446 (5)    0.0307 (4)    0.0051 (3)     0.0060 (3)     \u22120.0034 (3)\n  N1    0.0210 (11)   0.0449 (13)   0.0264 (12)   \u22120.0010 (9)    \u22120.0007 (9)    \u22120.0032 (9)\n  N2    0.0205 (11)   0.0461 (14)   0.0248 (12)   \u22120.0033 (9)    0.0019 (9)     \u22120.0061 (10)\n  C1    0.0171 (11)   0.0372 (14)   0.0280 (14)   \u22120.0036 (10)   \u22120.0002 (10)   0.0003 (11)\n  C2    0.0286 (13)   0.0340 (14)   0.0321 (15)   \u22120.0010 (11)   \u22120.0040 (11)   0.0052 (11)\n  C3    0.0297 (13)   0.0288 (13)   0.0367 (16)   0.0031 (10)    0.0014 (11)    0.0001 (11)\n  C4    0.0284 (13)   0.0306 (15)   0.0266 (14)   \u22120.0016 (10)   0.0072 (10)    0.0012 (10)\n  C5    0.0327 (14)   0.0437 (16)   0.0263 (14)   0.0074 (11)    0.0018 (11)    0.0068 (11)\n  C6    0.0296 (14)   0.0403 (15)   0.0345 (15)   0.0119 (11)    0.0047 (12)    0.0019 (11)\n  C7    0.0233 (12)   0.0329 (13)   0.0246 (14)   \u22120.0026 (10)   \u22120.0006 (10)   0.0029 (10)\n  C8    0.0281 (13)   0.0512 (16)   0.0307 (14)   \u22120.0059 (12)   0.0014 (11)    \u22120.0066 (12)\n  ----- ------------- ------------- ------------- -------------- -------------- --------------\n\nGeometric parameters (\u00c5, \u00b0) {#tablewrapgeomlong}\n===========================\n\n  ---------------- ------------- ---------------- -----------\n  Cl1---C4         1.743\u00a0(3)     C3---C4          1.377\u00a0(4)\n  N1---C7          1.299\u00a0(3)     C3---H3          0.9500\n  N1---C1          1.425\u00a0(3)     C4---C5          1.385\u00a0(3)\n  N2---C7          1.340\u00a0(3)     C5---C6          1.381\u00a0(3)\n  N2---H1N         0.89\u00a0(3)      C5---H5          0.9500\n  N2---H2N         0.80\u00a0(3)      C6---H6          0.9500\n  C1---C2          1.387\u00a0(3)     C7---C8          1.496\u00a0(4)\n  C1---C6          1.393\u00a0(3)     C8---H8A         0.9800\n  C2---C3          1.386\u00a0(4)     C8---H8B         0.9800\n  C2---H2          0.9500        C8---H8C         0.9800\n                                                  \n  C7---N1---C1     118.52\u00a0(19)   C6---C5---C4     119.1\u00a0(2)\n  C7---N2---H1N    119.4\u00a0(17)    C6---C5---H5     120.5\n  C7---N2---H2N    121\u00a0(2)       C4---C5---H5     120.5\n  H1N---N2---H2N   119\u00a0(3)       C5---C6---C1     121.0\u00a0(2)\n  C2---C1---C6     118.7\u00a0(2)     C5---C6---H6     119.5\n  C2---C1---N1     119.5\u00a0(2)     C1---C6---H6     119.5\n  C6---C1---N1     121.6\u00a0(2)     N1---C7---N2     125.8\u00a0(2)\n  C3---C2---C1     120.8\u00a0(2)     N1---C7---C8     119.0\u00a0(2)\n  C3---C2---H2     119.6         N2---C7---C8     115.2\u00a0(2)\n  C1---C2---H2     119.6         C7---C8---H8A    109.5\n  C4---C3---C2     119.4\u00a0(2)     C7---C8---H8B    109.5\n  C4---C3---H3     120.3         H8A---C8---H8B   109.5\n  C2---C3---H3     120.3         C7---C8---H8C    109.5\n  C3---C4---C5     121.0\u00a0(2)     H8A---C8---H8C   109.5\n  C3---C4---Cl1    119.65\u00a0(19)   H8B---C8---H8C   109.5\n  C5---C4---Cl1    119.4\u00a0(2)                      \n  ---------------- ------------- ---------------- -----------\n\nHydrogen-bond geometry (\u00c5, \u00b0) {#tablewraphbondslong}\n=============================\n\n  -------------------- ---------- ---------- ----------- ---------------\n  *D*---H\u00b7\u00b7\u00b7*A*        *D*---H    H\u00b7\u00b7\u00b7*A*    *D*\u00b7\u00b7\u00b7*A*   *D*---H\u00b7\u00b7\u00b7*A*\n  N2---H1n\u00b7\u00b7\u00b7N1^i^     0.88\u00a0(3)   2.08\u00a0(3)   2.965\u00a0(3)   176\u00a0(3)\n  N2---H2n\u00b7\u00b7\u00b7Cl1^ii^   0.80\u00a0(3)   2.83\u00a0(3)   3.464\u00a0(2)   138\u00a0(3)\n  -------------------- ---------- ---------- ----------- ---------------\n\nSymmetry codes: (i) *x*\u22121/2, *y*, \u2212*z*+3/2; (ii) *x*\u22121/2, \u2212*y*+1/2, \u2212*z*+1.\n\n###### Hydrogen-bond geometry (\u00c5, \u00b0)\n\n  *D*---H\u22ef*A*        *D*---H    H\u22ef*A*      *D*\u22ef*A*     *D*---H\u22ef*A*\n  ------------------ ---------- ---------- ----------- -------------\n  N2---H1n\u22efN1^i^     0.88 (3)   2.08 (3)   2.965 (3)   176 (3)\n  N2---H2n\u22efCl1^ii^   0.80 (3)   2.83 (3)   3.464 (2)   138 (3)\n\nSymmetry codes: (i) ; (ii) .\n\n[^1]: Additional correspondence author, e-mail: j.wardell\\@abdn.ac.uk.\n"}
+{"text": "1. INTRODUCTION {#sec1}\n===============\n\nThe true incidence of popliteal artery entrapment syndrome (PAES) is unknown. In a review of 20,000 asymptomatic young people and in a study of autopsy specimens, the incidence was reported to range between 0.17 and 3.5%, leading the authors to conclude that only a small proportion of cases are symptomatic \\[[@r1]-[@r3]\\].\n\nPAES is mostly seen in young people with well-developed muscles. Because this pathology is related to the PA and the surrounding tissue of the popliteal musculotendinous structure, its diagnosis can be delayed. Reports of this pathology in the literature are limited to a number of small series \\[[@r4]-[@r7]\\]. The concomitant entrapment of the popliteal vein with the artery has been reported in 7.6% of cases. PAES is a congenital anomaly of muscle or tendon insertion, in relation to the PA, that causes functional occlusion of the artery (7). This entity results from a developmental defect in which the PA passes medial to and beneath the medial head of the gastrocnemius muscle or a slip of that muscle, with consequent compression of the artery. In rare cases, an anomalous fibrous band or the popliteus muscle deep to the medial head of the gastrocnemius is the compressing structure. Radiologic diagnosis and management of patients involved the surgical repair of anomalous anatomical relationship between the PA and the musculotendinous structures responsible for causing functional occlusion of the PA in young people with no risk factors. The aim of surgical management is to release the musculocutaneous structures, and if necessary, to perform a PA embolectomy or bypass procedures to prevent limb loss. The purpose of our review is to report our diagnostic experiences involving 31 patients with PAES, who over a 14-year period were treated surgically at 2 institutions.\n\n2. PATIENTS AND METHODS {#sec2}\n=======================\n\nOver a period of 14 years, from 2001 to 2015, 31 healthy young patients (26 males) underwent surgery for PAES at 2 institutions. PAES was unilateral in 27 patients and bilateral in 4 patients. The ages of the patients ranged from 23 to 49 years, with a mean of 32\u00b17.9 years. The symptoms related to PAES included swelling, pain, paraesthesia, claudication, and rest pain. Preoperative diagnosis of PAES was based on various combinations of radiologic investigations, including duplex ultrasonography (US), magnetic resonance angiography (MRA), and conventional angiography. The demographics and clinical properties of the patients are summarized in (Table **[1](#T1){ref-type=\"table\"}**).\n\nPatients were examined to determine whether they had any of the following preoperative associated risk factors: smoking history, presence of concomitant disease, hypertension, hyperlipidaemia, diabetes mellitus, chronic renal failure, or ischemic heart disease. Besides a history of smoking, which 22 patients reported to have, none of the patients had a clinical history of these diseases. The main complaints were progressive intermittent claudication and foot coldness during walking or exercise, with symptom duration lasting from 8 months to 2 years.\n\nVarious radiologic investigations, including Doppler ultrasonography US with positional stress test, MRA, and traditional angiography were used to diagnose PAES. Doppler ultrasonography US with positional stress test is an important clinical test that involves determining whether or not there are pedal pulses when the patient's foot is in plantar hyperflexion. Both limbs of all the patients were examined, with the results showing that 25 of the 35 limbs \\[[@r8]\\] (71.4%) had a positive result. Because the MRA clearly showed the abnormal anatomic relationship between the PA and the head of the musculotendinous anatomy, a repeat MRA was performed to confirm the definitive diagnosis of PAES in all the patients. Angiographic findings in the neutral position consisted of medial deviation of the proximal PA in 18 limbs (51.4%), poststenotic dilatation of the distal PA in 10 limbs (28.5%), and segmental or longitudinal occlusion of the PA in 3 limbs (8.5%). PAES were described by Rich **et al.** \\[[@r8]\\] Type II, III, and IV in 35 limbs. In our study, Type I was detected in 4 limbs, Type II in 12 limbs, Type III in 13 limbs, and Type IV in 6 limbs. None of the patients were Type V or Type VI. Surgery was recommended to the patients.\n\n3. SURGICAL PROCEDURES {#sec3}\n======================\n\nAll operations were performed under general anesthesia. We preferred a posterior approach to establish clear visualization of the relationships between the artery and musculotendinous structures. After dissection of the neurovascular bundle in the popliteal fossa, we incised through any anomalous insertions or attachments involving the medial head of the gastrocnemius muscle or other bands causing compression of the PA and veins. Musculotendinous section and PA release were performed. In the 4 patients with Type I PAES, simple release of the PA by division of the medial head of gastrocnemius and popliteus muscles and tendons were performed because the vessel remained undamaged. We detected intraluminal PA thrombus in 12 limbs. Therefore, PA embolectomy and simple release of the muscles and tendons was performed.\n\nBecause PA was severely fibrotic in the remaining 15 patients who had Type III and Type IV PAES, it was decided that an arterial reconstruction was the best option. We used the radial artery (RA) for arterial reconstruction in all operations. The RA was harvested from the non-dominant hand after applying the Allen Test. A patch angioplasty with thromboendarterectomy was performed in 8 limbs using the RA. The RA interposition was performed in the remaining 7 patients after division of adjacent muscles and tendons because there were irreversible fibrous thickening of the popliteal arterial wall. In patients who were referred to us due to arterial embolic event and for whom an arterial patch plasty or arterial graft interposition was performed, we administered a low molecular weight heparin in the early postoperative period.\n\nThe mean size of PA was 3.6 \u00b1 0.9 mm (2.9-5.1 mm) in 12 patients who underwent the RA graft interposition. Because of the RA size was smaller than PA in some patients, we used a mixed vasodilator solution prior to anastomosis. We incubated the RA in vasodilation solution (including a calcium channel inhibitor, papaverine, and nitroglycerine). We previously described the efficasy of this mixed solution \\[[@r9]\\]. The RAs' diameter increased significantly after incubation. The mean diameter of the RA was 2.2\u00b1 0.4 mm prior to incubation. After incubation, the mean diameter of the RA was measured as 4.06\u00b11.1 mm. Therefore, there was no mismatch between the size of popliteal and the RA.\n\nPostoperatively, antiplatelet and antiaggregant drugs, clopidogrel (daily dose of 75 mg), and acetylsalicylic acid (ASA) (daily dose of 100 mg) were prescribed for each patient after discharge.\n\n4. CASE PRESENTATIONS {#sec4}\n=====================\n\n4.1. Case 4 {#sec4.1}\n-----------\n\nA 23-year-old young woman presented with a 3-year history of experiencing numbness and pain in the right lower leg during exercise. Over the past 2 years, the patient had experienced worsening leg pain when running long distances. The patient had no cardiovascular risk factors. She had been hospitalized in the orthopaedic and physiotherapy clinic due to leg pain prior to admission to our clinic. However, her symptoms continued to persist. Examination of the right leg showed that there was swelling and no pulse below the femoral artery. Pulses were normal in left lower limb. Magnetic resonance imaging (MRI) revealed normal aortic and iliac arteries. Anterior and posterior views of MRA showed an occlusion of the right popliteal and peroneal arteries Figs. (**[1](#F1){ref-type=\"fig\"}** and **[2](#F2){ref-type=\"fig\"}**). In this case, intra-arterial thrombolysis was performed. Control angiography showed that there was a thrombus in the artery and a stricture at the mid-PA, with an irregularity of the wall. At the end of two 2 days, the angiogram demonstrated that the thrombus had largely resolved; however, the stricture of PA was present. MRI scanning was subsequently performed, which showed that the medial head of the gastrocnemius muscle, the PA and vein where it inserted in a more lateral position Fig. (**[3](#F3){ref-type=\"fig\"}**). Further evidence from the MRI revealed that the PA crossed below the popliteal muscle in this patient with Type IV PAES. Contralateral limb MRI was normal. Posterior popliteal approach was preferred. An S-shaped skin incision was performed. The PA crossed under the popliteus muscle and there was a hypertrophic band, which was the cause of occlusive pressure on the PA. The hypertrophic bands and the popliteus muscle were resected. PA was opened and it was partially occluded. The tibioperoneal artery was also occluded. A thromboendarterectomy (TEA) was performed. The artery was repaired using a RA. No rest pain was detected after surgery. US showed that the distal arterial system and the RA were patent. A low molecular weight heparin and acetylsalicylic acid were prescribed in the postoperative period. The patient was discharged home in good clinical condition. We prescribed a daily dose of 100 mg of acetylsalicylic acid. The patient was able to run without any symptoms in the lower limb.\n\n4.2. Case 8 {#sec4.2}\n-----------\n\nA 25-year-old man presented with a 3-year history of pain in the left lower leg during walking. Claudication was the main symptom. The patient had no other risk factors for atherosclerosis or Buerger's disease. Systemic and cardiac system examinations were normal. All pulses including popliteal and distal arteries were not palpable in the neutral position of left leg. Leg swelling, and ischemic symptoms were detected. There were palpable femoral pulses. Motor and sensory functions of both lower limbs was intact. The patient refused the conventional angiography. Therefore, an MRI scan was performed, which showed the aorta and iliac arteries to be normal. There was an occluded left PA. MRI demonstrated that the PA crossed below the popliteal muscle Fig. (**[4](#F4){ref-type=\"fig\"}**). We suggested an operation due to ischemia of the lower limb, but the patient declined a surgical approach. Therefore medical therapy, which included administration of antiaggregant and anticoagulant drugs, was prescribed. Six months later, this patient presented to our emergency clinic with the complaint of ischemic leg symptoms. Conventional angiography showed that there was a total occlusion of PA because of an acute thrombus. We operated in an emergent situation. The medial head of the gastrocnemius muscle was totally divided, and fibrotic bands were cut. Embolectomy was performed using a Fogarty catheter. PA blood flow was provided using a saphenous vein interposition. Dual antiplatelet treatment, clopidogrel (daily dose of 75 mgr), and acetylsalicylic acid (daily dose of 100 mgr) were prescribed after surgery. The patient was discharged to home in a good clinical condition 5 days after surgery.\n\n4.3. Case 11 {#sec4.3}\n------------\n\nA healthy 27-year-old male presented to our hospital with claudication in both legs. He was a non-smoker. No paraesthesia, weakness, or other ischemic symptoms were apparent. On physical examination, there was no pulse of the dorsalis pedis. A MRA showed an occluded PA. MRI of the right leg demonstrated that there was severe compression of the PA by an accessory slip of gastrocnemius muscle (Type III PAES) of the right leg Fig. (**[5A](#F5){ref-type=\"fig\"}**). Lateral view of MR imaging of the popliteal region showed the gastrocnemius muscle attached more laterally Fig. (**[5B](#F5){ref-type=\"fig\"}**). Surgery for relief of compressed PA was carried out. A myotomy was performed and fibrotic bands were cut. PA blood flow was provided using a RA interposition. Postoperative angiography showed the complete resolution of the PA compression. The RA graft was patent. The patient was discharged home on postoperative day 6. We prescribed clopidogrel for a year. The patient has no ischemic symptoms.\n\n4.4. Case 18 {#sec4.4}\n------------\n\nA 27-year-old professional female athlete was referred to our clinic because of sudden embolic events of the left lower leg. She described severe pain in the left leg during an exertion. The patient was a non-smoker. The bilateral femoral arteries were patent. However, the left PA and the pedal pulse were not palpable. Doppler US showed no colour filling with hyperechoic luminal thrombus. Motor and sensory functions of the legs were normal. A computed tomography (CT) showed an occlusion of the left PA due to intra-arterial thrombus Fig. (**[6A](#F6){ref-type=\"fig\"}**). We performed an intra-arterial thrombolysis using a glycoprotein IIb/IIIa inhibitor (tirofiban) followed by intra-arterial urokinase. Twenty-four hours later, ischaemic symptoms, including pain and pallor, started again on in the right limb. The distal pulses were not palpable. Because of the suspicion of PAES, a MRA was performed. The MRA results showed that total occlusion of the left PA Fig. (**[6B](#F6){ref-type=\"fig\"}**). In this case, the medial head of the gastrocnemius muscle was attached laterally. The PA was crossed under the gastrocnemius muscle insertion which described Type III PAES with the total occlusion of the right PA. Therefore, surgery was recommended. The medial head of the gastrocnemius muscle and tendon were carefully divided, and an embolectomy was performed. The distal pulses were palpable after surgery. The patient had no ischemic symptoms. Clopidogrel was prescribed as an antiaggregant. The patient was discharged from the hospital in good clinical condition. A CT angiography was performed 6 months later demonstrated that the popliteal, anterior tibial, and the peroneal arteries were patent (Fig. **[7](#F7){ref-type=\"fig\"}**).\n\n5. DISCUSSION {#sec5}\n=============\n\nWe presented our diagnostic modalities and treatment strategy for 31 young patients with PAES who were admitted to 2 cardiovascular surgery clinics because of calf pain, cramping during walking, ischemia, and swelling of the affected lower limb. There were 4 female patients (12.9%). For the diagnosis of PAES, we routinely used Doppler US for the first 12 patients with passive plantar flexion and traditional angiography was performed. However, because conventional angiography is not able to show the anatomical relationship between the PA and popliteal venous system (except slight medial deviation of the PA), MRI was the first choice of diagnostic modality after Doppler US. Symptoms were limited to intermittent claudication due to the intermittent compression of the artery during plantar hyper dorsiflexion.\n\nChronic extrinsic arterial compression leads to vascular endothelial trauma, early arterial wall fibrosis, and subsequent thrombosis. Thrombus formation may cause complete obstruction of the compressed PA, which in turn can lead to leg ischemia due to insufficient collateral circulation of the leg. We therefore performed a thromboembolectomy or TEA and patch plasty in 12 patients. In 15 patients, the RA interposition was performed after TEA. For the first time such a sequence of procedures has been conducted. In the literature, it has been reported that either the great saphenous vein or a 7 or 8 mm polytetrafluoroethylene graft can be used for the construction of the PA.\n\nThe the Popliteal Vascular Entrapment Forum in 1998 was an attempt to gain some consensus on the anatomic classification of the different types of PAES. Type I is characterized by an atypical course of the PA, Type II occurs as a result of an abnormal muscular insertion. In Type III, an accessory slip of muscle from gastrocnemius slings around the artery. In patients of Type IV PAES, an artery lies deep in popliteal fossa entrapped by popliteus or fibrous band. In Types V and VI, both PA and vein are entrapped.\n\nTraditional invasive angiography is the standard diagnostic tool for patients with ischemic symptoms. The use of this technique, however, is unable to demonstrate compression of the PA in some patients. Gourgiotis **et al.**, in addition to other authors, have noted that irregularity of the wall of the PA should raise suspicion of PAES \\[[@r8], [@r10]\\]. MRI is useful for evaluating aberrant muscular anatomy of the popliteal fossa and also for showing compression and deviation due to peripheral tissue effects on the PA and anomalous insertion of gastrocnemius or popliteal popliteal muscle and tendons \\[[@r11]-[@r15]\\]. Our experiences showed that MRA was useful for detecting the anatomic relationship between the artery and the calf muscles in young males with acute limb ischaemia.\n\nThe most common and successful treatment option is to perform a myotomy of popliteal muscle to release the entrapment, an effective treatment for all types of PAES. A resection of the hypertrophic structures should be done. If early diagnosis is not made, fibrosis or total occlusion of the PA could occur. Distal emboli due to poststenotic aneurysm or acute thrombus formation can be seen in patients with PAES \\[[@r16]\\]. Halici **et al.** reported on 12 patients (9 males, 3 females) with PAES in 2016 \\[[@r17]\\]. In a study by Meier **et al.**, patients with PAES were successfully treated by intravascular thrombolysis and thromboembolectomy, followed by musculotendinous dissection \\[[@r18]\\]. The surgical options available to these authors included TEA with patch angioplasty and an autogenous vein graft interposition after segmental arterial resection \\[[@r17]\\]. Interposition of vein grafting is superior to TEA and vein patching for stenosis, as it has a complication rate of 16.7% compared with 45.5% in the others \\[[@r18]\\]. In our patients who underwent a RA patch plasty, we did not detect any stenosis during the follow-up. The patency rate of bypass surgery using vein grafts has been reported to be 57-65% over a period of 8-10 years \\[[@r19], [@r20]\\]. Levien **et al.** reported a case series of 66 limbs that underwent musculotendinous section and 16 limbs that underwent segmental replacement of the occluded PA with a reversed saphenous vein \\[[@r21]\\]. The limbs treated using vein grafts demonstrated no graft occlusion. On the other hand, di Marzo **et al.** reported that 15 limbs that underwent revascularization procedures exhibited a patency rate of 65% \\[[@r22]\\]. Kim **et al.** also reported that bypass surgery was associated with a worse patency rate than interposition procedures \\[[@r23]\\]. Igari **et al.** suggested that the treatment of PAES with myotomy and selective revascularization achieved good long-term outcomes \\[[@r24]\\].\n\nGiven the seriousness of PAES, particularly as it threatens loss of limbs, clinicians should consider early diagnosis of PAES in young male patients who present with ischemic symptoms. PAES and thromboangiitis obliterans (Buerger's disease) are both characterized by distal arteriopathy in young males, but the treatment options are different. Patients with Buerger disease are generally heavy smokers. This pathology may sometimes be misdiagnosed as atherosclerosis or Buerger's disease. Igari **et al.** reported a case with PAES which had been misdiagnosed as Buerger's disease \\[[@r25]\\]. There are of course cases of male smokers who present with complaints of bilateral intermittent claudication and are referred from orthopaedic clinics for suspicion of Buerger's disease. We believe that an MRA and a Doppler US, accompanied by a plantar hyperflexion test, should be done for differential diagnosis and early treatment.\n\nWe performed MRA and CT as the main diagnostic procedures for diagnosis of PAES and analysed the relationship between the artery and musculocutaneous tissue in all patients. These modalities can demonstrate the vessel lumen as well as the surrounding anatomy. Traditional angiography was performed in the neutral limbs position, as well as with the foot in either dorsiflexion or plantar flexion to elicit compression and to confirm the diagnosis prior to surgery. Imaging commonly shows a normal arterial lumen when the foot is in the relaxed position, and a narrowing of the arterial lumen during stress manueuvres \\[[@r26]\\] for differentiation of susceptible arterial thrombosis \\[[@r27]\\].\n\nPAES is not an indication for angioplasty or stent placement; however, interventional thrombolysis would be appropriate treatment for patients who present with occlusion due to PAES. Thrombolysis of the distal popliteal and interventional thrombolysis would be appropriate therapy for patients who present with acute occlusion of popliteal or distal arterial trees due to PAES \\[[@r21]\\]. Thrombolysis of the distal popliteal and runoff vessels can be very important prior to surgical correction. In addition, MRI may not provide an adequate information of collateral arterial system of limb arteries. Therefore, to provide collateral circulation system during an angiography, we used conventional angiography to clearly visualise the run off of the limb arteries.\n\nAnkle-brachial index (ABI) has been suggested as a sign of arterial impairment in patients with PAES \\[[@r28]\\]. In previous publications, the results of post-exercise test with a doppler US/ABI have been reported as a main part of the investigations for PAES prior to treatment \\[[@r29]\\] \\[[@r30]\\] \\[[@r31]\\]. ABI is a simple, noninvasive, effective, and inexpensive method. Collins **et al.**, and Ruppert **et al.** proposed the provocation manoeuvres as a pathway for diagnosis of PAES \\[[@r29]\\] \\[[@r30]\\]. In this method, the ABI measurement should be obtained at rest, and just after treadmill using an initial speed of 6 km/h \\[[@r31]\\]. To produce symptoms, the speed of treadmill may be increased during the exercise. The pressures of exercise compared with the baseline brachial pressures. If there is a reduction in indices, a failure to increase the index with exercise with a normal increase in brachial pressure \\>5 mm Hg above resting systolic, the examination is accepted as a positive test \\[[@r30]\\] \\[[@r31]\\].\n\nPAES is a rare cause of a lower limb -threatening condition in healthy young people. Patients with PAES are reported as small case series in the English literature. As part of this study, we presented our 14-year experience with 31 patients (35 affected limbs). For the first time, we also described the use of the RA for PA reconstruction. If a young patient presents with unexplained lower limb pain experienced during walking, a diagnosis of PAES should be considered. Acute PA embolus can be seen as a first symptom in these patients who have no cardiovascular risk factors. In our opinion, Doppler US, using passive or active dorsiflexion of foot, and MRI should be performed as non-invasive methods for diagnosis of PAES. A posterior approach using an S-shaped incision to establish clear visualization of abnormal insertion of muscles and hypertrophic bands is the best option. If there is arterial fibrosis due to chronic arterial trauma, the fibrotic segment should be resected. The RA may be used as an arterial conduit to provide PA continuity. An antiaggregant such as clopidogrel, and a calcium channel blocker to inhibit RA spasm, can be used for up to 1 year in the post-surgery period.\n\nDeclared none.\n\nETHICS APPROVAL AND CONSENT TO PARTICIPATE\n==========================================\n\nNot applicable.\n\nDECLARATION OF PATIENT CONSENT\n==============================\n\nThe authors declared that the forms of patients' consents have been provided prior to study. The authors certify that they have obtained all appropriate patient consent forms. In the form the patients gave their consent for the images and other clinical information to be reported in the journal.\n\nHUMAN AND ANIMAL RIGHTS\n=======================\n\nNo animals were used in this research. All research procedures followed were in accordance with the ethical standards of the committee responsible for human experimentation (institutional and national), and with the *Helsinki Declaration* of 1975, as revised in 2008 (<http://www.wma.net/en/20activities/10ethics/10helsinki/>)\n\nCONFLICTS OF INTEREST\n=====================\n\nThe authors declare no conflict of interest, financial or otherwise.\n\n![](TOCMJ-12-18_F1){#F1}\n\n![](TOCMJ-12-18_F2){#F2}\n\n![](TOCMJ-12-18_F3){#F3}\n\n![](TOCMJ-12-18_F4){#F4}\n\n![](TOCMJ-12-18_F5){#F5}\n\n![](TOCMJ-12-18_F6){#F6}\n\n![](TOCMJ-12-18_F7){#F7}\n\n###### \n\nClinical charecteristics of the patients with popliteal artery entrapment syndrome.\n\n  Variable                    Number\n  --------------------------- --------\n  **Sex**                     \n  Male                        27\n  Female                      4\n  **Smoking**                 \n  No                          9\n  Yes                         22\n  **Laterality**              \n  Right                       15\n  Left                        12\n  Bilateral                   4\n  **Symptoms**                \n  Intermittent claudication   26\n  Foot Coldness               11\n  Calf Pain                   9\n  Swelling                    8\n"}
+{"text": "Background {#Sec1}\n==========\n\nSepsis is a leading cause of morbidity and mortality, during and after hospitalization in intensive care units (ICU) \\[[@CR1], [@CR2]\\]. Carbapenem-resistant (CR) *Acinetobacter baumannii*, *Pseudomonas aeruginosa* and Enterobacteriaceae have become major pathogens, especially in ICUs, implicated in healthcare-associated sepsis, causing prolonged hospitalization, high mortality, and increased costs \\[[@CR3]--[@CR6]\\].\n\nSeveral studies in the literature have investigated risk factors for infections by specific species of CR-Gram-negative bacilli (GNB) \\[[@CR7]--[@CR17]\\], but few considered all detected GNB species \\[[@CR18], [@CR19]\\], and most of them were in the context of nosocomial infection \\[[@CR9], [@CR10], [@CR12], [@CR15]--[@CR17]\\] or bacteremia \\[[@CR8], [@CR11], [@CR18], [@CR19]\\].\n\nThe empirical therapy of sepsis should be started within the first hour of presumed diagnosis, at a time when the clinic-epidemiological characteristics remain as the only determinants of a patient at greater risk.\n\nGiven these facts, we performed a case-case-control study to investigate predictive factors for sepsis by CR- and carbapenem-susceptible (CS) GNB in adult patients from a Brazilian public ICU. Our goal is to develop and validate a predictive score to identify patients at higher risk for CR-GNB sepsis in future studies.\n\nMethods {#Sec2}\n=======\n\nPatients, setting and study design {#Sec3}\n----------------------------------\n\nThe study followed a case-control design from a prospective cohort of patients with SIRS, sepsis-2 or sepsis-3 criteria in which blood and other samples' cultures were collected and antimicrobial therapy was instituted, for two or more days, in an adult clinical-surgical ICU, at a tertiary public hospital in Rio de Janeiro, Brazil, from August 2015 through March 2017. This study was approved by the institutional ethics committee and followed the Declaration of Helsinki and its later amendments. The study followed the STROBE recommendations for observational cohort studies (STROBE list in Additional file [1](#MOESM1){ref-type=\"media\"}: Appendix [S4](#MOESM1){ref-type=\"media\"}) \\[[@CR20]\\].\n\nWe evaluated all patients\\` clinical and surveillance samples cultured during the episodes, throughout the ICU stay, and the follow-up period of 30\u2009days following the end of sepsis treatment. Surveillance cultures were not used to assess patient inclusion or exclusion in the study. Patients were not matched by any variable, considering the homogeneity of this population.\n\nWe followed all detected sepsis episodes during the ICU stay and follow-up period. Patients with more than one episode were aleatory selected in each group in such a manner that patients who presented CR-GNB sepsis were selected as case 1, patients who had CS-GNB sepsis episode were selected as case 2 and those with unknown sepsis or due to other etiologies than GNB were selected as control group. Cases and controls entered the study once and were monitored closely during the follow-up period.\n\nWe excluded patients younger than 18\u2009years old, those with sepsis acquired in the community or associated with another healthcare institution, those who refused to sign the consent form and those suffering from polymicrobial sepsis by GNB and non GNB agents. In addition, we excluded patients initially enrolled in the control or CS-GNB case group that evolved respectively with CS-GNB or CR-GNB infection after discharge from ICU and during the follow-up period.\n\nThe variables investigated as predictive factors were studied during the period of hospitalization prior to sepsis episode for both cases and controls. We investigated demographics and comorbidities, length of hospital stays, prior ICU hospitalization, reasons for ICU admission, simplified acute physiology *score* (SAPS-3) and sequential organ failure assessment (SOFA) at baseline, previous use of invasive devices and antimicrobials. The variables and their definitions are described in Tables\u00a0[1](#Tab1){ref-type=\"table\"} and [2](#Tab2){ref-type=\"table\"}. The information was collected from multiple sources including hospital records, hospital laboratory system, radiological records, hospital infection control committee daily surveillance, ICU staff daily clinical round records, and data entered daily into the ICU Epimed System. We collected and managed study data using REDCap electronic data capture tools hosted at Instituto Oswaldo Cruz, Funda\u00e7\u00e3o Oswaldo Cruz (IOC/FIOCRUZ). The study formularies with the investigated variables are in Additional file [1](#MOESM1){ref-type=\"media\"}: Appendices [S1](#MOESM1){ref-type=\"media\"}, [S2](#MOESM1){ref-type=\"media\"} and [S3](#MOESM1){ref-type=\"media\"}. To avoid potential bias, all data collected were standardized and monitored throughout the study. Table 1Clinical characteristic investigated for association with CR-GNB sepsis (Case group 1) in an adult clinical-surgical ICUPredictive FactorsUnivariate analysisOdds ratio95% CI*P*^t^ valorCR-GNB (*n*\u2009=\u200960)Controls (*n*\u2009=\u200994)Demographic data\u2003Age in years, median (range)62 (23--91)62 (19--92)1.000.98--1.020.95\u2003Male sex, n (%)36 (60)46 (49)1.570.81--3.020.18Comorbidities^a^, n (%)\u2003Diabetes mellitus^b^21 (35)35 (37)0.910.46--1.780.78\u2003Renal failure^c^30 (50)38 (40)1.470.77--2.830.25\u2003Hemodialysis^d^22 (37)18 (19)2.441.17--5.100.02\u2003Chronic liver disease^e^2 (3)9 (10)0.330.07--1.560.16\u2003Immunossupressive condition^f^9 (15)11 (12)1.330.52--3.430.55\u2003Gastrointestinal disease16 (27)18 (19)1.540.71--3.310.27\u2003Geniturinary disease5 (8)9 (10)0.860.27--2.700.79\u2003Pulmonary disease10 (17)14 (15)1.140.47--2.770.77\u2003AIDS or chronic infectious disease3 (5)3 (3)1.600.31--8.180.56\u2003Surgery^g^26 (43)46 (49)0.800.42--1.530.50\u2003Infection/Colonization by CR-GNB19 (32)7 (7)5.762.24--14.79\\<\u20090.001\u2003Nosocomial diarrhea^h^17 (28)6 (6)5,.802.13--15.750.001\u2003Neutropenia^i^5 (8)4 (4)2.050.53--7.940.30\u2003Neoplasm23 (38)28 (30)1.470.74--2.900.27\u2003Infection46 (77)31 (33)6.683.20--13.95\\<\u20090.001Prior ICU hospitalization, n (%)7 (12)5 (5)2.350.71--7.780.26Length of hospital stay (in days)\u2003Median (range)26.5 (1--375)10 (0--143)1.041.02--1.06\\<\u20090.001ICU hospitalization reason, n (%)\u2003Elective or emergency surgery10 (17)28 (30)0.470.21--1.060.07\u2003Respiratory tract disease16 (27)14 (15)2.080.93--4.650.08\u2003Cardiovascular disease6 (10)8 (9)1.190.39--3.630.75\u2003Neurological disease2 (3)11 (12)0.260.06--1.220.09\u2003Gastrointestinal disease3 (5)5 (5)0.940.22--4.070.93\u2003Renal pathology5 (8)11 (12)0.690.23--2.080.51\u2003Sepsis31 (52)30 (32)2.401.23--4.680.01\u2003Sepsis shock23 (38)27 (39)1.630.82--3.240.17Total SOFA score^j^ at ICU admission,\u2003Median (range)6 (1--17)6 (0--17)1.060.98--1.150.16SAPs 3 score^k^, median (range)65 (30--103)64 (29--105)1.010.99--1.030.39Invasive devices, n (%)\u2003Mechanical ventilation56 (93)58 (62)8.692.90--26.01\\<\u20090.001\u2003Central vascular catheter60 (100)81 (86)\\...\\...0.003\u2003Urinary catheter55 (92)72 (76)3.161.12--8.920.03Previous use of antimicrobials, n (%)\u2003Aminoglycosides^l^18 (30)7 (7)5.332.07--13.740.001\u2003Cephalosporins, 3rd and 4rd generations^m^8 (13)16 (17)0.750.30--1.880.54\u2003Carbapenems^n^48 (80)32 (34)7.753.61--16.62\\<\u20090.001\u2003Glycopeptides^o^, linezolid and tigecycline45 (75)35 (37)5.062.47--10.37\\<\u20090.001\u2003Fluoroquinolones^p^14 (23)15 (16)1.600.71--3.620.26\u2003Metronidazole11 (18)13 (14)1.400.58--3.370.46\u2003Piperacilin-tazobactam30 (50)21 (22)3.481.72--7.01\\<\u20090.001\u2003Polymyxins^q^22 (37)9 (10)5.472.30--12.98\\<\u20090.001\u2003ATB with action for anaerobes^r^56 (93)59 (63)8.312.77--24.88\\<\u20090.001\u2003Antifungal agents^s^26 (43)13 (14)4.772.19--10.36\\<\u20090.001*Abbreviations*: *AIDS* Acquired Immunodeficiency Syndrome, *ATB* antibiotic agents, *CI* Confidence interval, *CR-GNB* Carbapenem resistant Gram-negative bacilli, *ICU* Intensive Care Unit^a^Prior comorbidities or conditions to investigated sepsis episode^b^Diagnosis of diabetes mellitus requiring oral or injectable hypoglycemic drug^c^Creatinine clearance \\<\u200930\u2009cc/min^d^Required in the last 90\u2009days^e^Laboratory clinical evidence^f^ Prednisone \\>\u200910\u2009mg for more than 50\u2009days, corticosteroid for \\>\u20097\u2009days or immunomodulatory agents (examples: monoclonal agents, methotrexate)^g^In the last 30\u2009days^h^Nosocomial diarrhea (3 or more daily episodes of stool for 2 or more days)^i^Granulocytes\u2009\\<\u2009500 cells/mm^3j^Sequential Organ Failure Assessment score^k^Simplified Acute Physiology Score III^l^Amikacin and gentamicin^m^Ceftriaxone, ceftazidime and cefepime^n^Ertapenem, imipenem-cilastatin and meropenem^o^Daptomycin, teicoplanin and vancomycin^p^Ciprofloxacin, levofloxacin, moxifloxacin^q^Polymyxin B and colistin^r^Antibacterial agents with action for anaerobes -- Amoxicilin-clavulanate, ampicillin-sulbactam, piperacilin-tazobactam, clindamycin, ertapenem, imipenem, meropenem and metronidazole^s^Amphotericin B family (standard, lipid complex or liposomal Amphotericin), echinocandins and azoles^t^Pearson's chi-square test or Fisher's exact test or Mann-Whitney-Wilcoxon U test, as required, and considering statistically significant *p*\u2009\\<\u20090.05Table 2Clinical characteristic investigated for association with CS-GNB sepsis (Case group 2) in an adult clinical-surgical ICUPredictive FactorsUnivariate analysisOdds ratio95% CI*p*^t^ valorCS-GNB (*n*\u2009=\u200930)Controls (*n*\u2009=\u200994)Demographic data\u2003Age in years, median (range)66 (27--82)62 (19--92)1.010.98--1.040.46\u2003Male sex, n (%)9 (30)46 (49)2.240.93--5.390.07Comorbidities^a^, n (%)\u2003Diabetes mellitus^b^15 (50)35 (37)1.690.74--3.860.22\u2003Renal failure^c^8 (27)38 (40)0.540.22--1.330.18\u2003Hemodialysis^d^3 (10)18 (19)0.470.13--1.720.25\u2003Chronic liver disease^e^2 (7)9 (10)0.670.14--3.310.63\u2003Immunossupressive condition^f^6 (20)11 (12)1.890.63--5.630.26\u2003Gastrointestinal disease9 (30)18 (19)1.810.71--4.610.21\u2003Geniturinary disease3 (10)9 (10)1.050.27--4.160.95\u2003Pulmonary disease7 (23)14 (15)1.740.63--4.820.29\u2003AIDS or chronic infectious disease2 (7)3 (3)2.170.35--13.620.41\u2003Surgery^g^19 (63)46 (49)1.800.77--4.200.17\u2003Infection/Colonization by CR-GNB5 (17)7 (7)2.490.73--8.510.15\u2003Nosocomial diarrhea^h^5 (17)6 (6)2.930.83--10.420.10\u2003Neutropenia^i^1 (3)4 (4)0.780.08--7.220.82\u2003Neoplasm17 (57)28 (30)3.081.32--7.190.009\u2003Infection10 (33)31 (33)0.920.39--2.210.86Prior ICU hospitalization, n (%)9 (30)5 (5)7.632.32--25.130.001Length of hospital stay (in days)\u2003Median (range)15 (0--142)10 (0--143)1.010.99--1.030.16ICU hospitalization reason, n (%)\u2003Elective or emergency surgery10 (33)28 (30)1.180.49--2.840.71\u2003Respiratory tract disease5 (17)14 (15)1.140.38--3.490.82\u2003Cardiovascular disease1 (3)8 (9)0.370.04--3.090.36\u2003Neurological disease3 (10)11 (12)0.840.22--3.230.80\u2003Gastrointestinal disease1(3)5 (5)0.610.07--5.470.66\u2003Renal pathology1 (3)11 (12)0.260.03--2.100.21\u2003Sepsis10 (33)30 (32)1.070.45--2.560.89\u2003Sepsis shock10 (33)27 (39)1.240.51--2.990.63Total SOFA score^j^ at ICU admission,\u2003Median (range)6 (1--18)6 (0--17)1.030.93--1.140.58SAPs 3 score^k^, median (range)66 (27--97)64 (29--105)1.000.98--1.020.95Invasive devices, n (%)\u2003Mechanical ventilation18 (60)58 (62)0.970.42--2.250.95\u2003Central vascular catheter25 (83)81 (86)0.800.26--2.470.70\u2003Urinary catheter23 (77)72 (76)1.000.38--2.650.99Previous use of antimicrobials, n (%)\u2003Aminoglycosides^l^2 (7)7 (7)0.890.17--4.520.89\u2003Cephalosporins, 3rd and 4rd generations^m^5 (17)16 (17)0.980.32--2.930.96\u2003Carbapenems^n^5 (17)32 (34)0.390.14--1.110.08\u2003Glycopeptides^o^, linezolid and tigecycline7 (23)35 (37)0.510.19--1.300.17\u2003Fluoroquinolones^p^2 (7)15 (16)0.380.08--1.750.21\u2003Metronidazole7 (23)13 (14)1.900.68--5.310.22\u2003Piperacilin-tazobactam8 (27)21 (22)1.260.49--3.250.63\u2003Polymyxins^q^3 (10)9 (10)1.050.27--4.160.95\u2003ATB with action for anaerobes^r^14 (47)59 (63)0.520.23--1.190.12\u2003Antifungal agents^s^3 (10)13 (14)0.690.18--2.610.59*Abbreviations*: *AIDS* Acquired Immunodeficiency Syndrome, *ATB* antibiotic agents, *CI* Confidence interval, *CR-GNB* Carbapenem resistant Gram-negative bacilli, *ICU* Intensive Care Unit^a^Prior comorbidities or conditions to investigated sepsis episode^b^Diagnosis of diabetes mellitus requiring oral or injectable hypoglycemic drug^c^Creatinine clearance \\<\u200930\u2009cc/min^d^Required in the last 90\u2009days^e^Laboratory clinical evidence^f^Prednisone\u2009\\>\u200910\u2009mg for more than 50\u2009days, corticosteroid for \\>\u20097\u2009days or immunomodulatory agents (examples: monoclonal agents, methotrexate)^g^In the last 30\u2009days^h^Nosocomial diarrhea (3 or more daily episodes of stool for 2 or more days)^i^Granulocytes\u2009\\<\u2009500 cells/mm^3j^Sequential Organ Failure Assessment score^k^Simplified Acute Physiology Score III^l^Amikacin and gentamicin^m^Ceftriaxone, ceftazidime and cefepime^n^Ertapenem, imipenem-cilastatin and meropenem^o^Daptomycin, teicoplanin and vancomycin^p^Ciprofloxacin, levofloxacin, moxifloxacin^q^Polymyxin B and colistin^r^Antibacterial agents with action for anaerobes -- Amoxicilin-clavulanate, ampicillin-sulbactam, piperacilin-tazobactam, clindamycin, ertapenem, imipenem, meropenem and metronidazole^s^Amphotericin B family (standard, lipid complex or liposomal Amphotericin), echinocandins and azoles^t^Pearson's chi-square test or Fisher's exact test or Mann-Whitney-Wilcoxon U test, as required, and considering statistically significant *p*\u2009\\<\u20090.05\n\nDefinitions {#Sec4}\n-----------\n\nRecurrent sepsis was defined as a new episode of sepsis developing after resolution of clinical and laboratory parameters of sepsis, or the recrudescence of sepsis with the evidence of new etiology by cultures during ICU stay. An adaptation of Singer et al. (2016) \\[[@CR21]\\] sepsis-3 criteria was used retrospectively as follows: delta SOFA \u22652 between SOFA scores measured on two calendar days between the period of 72\u2009h that preceded to 24\u2009h that succeeded the date of initial blood culture, and on the ICU admission date; and qSOFA applied in patients without mechanical ventilation or sedation within 72\u2009h before and 24\u2009h after the date of blood culture.\n\nPost-hoc analysis was performed by two research infectious disease physician investigators to review all clinical, radiological, and microbiological data. We reviewed the evidence of sepsis-2 \\[[@CR22]\\] and\u2009\u2212\u20093 \\[[@CR21]\\], and of the infectious source. We classified the plausibility of infectious source as definitive, probable, possible or undetermined, according to Klouwenberg et al. (2013) criteria \\[[@CR23]\\], adapted to include the updated Centers for Disease Control and Prevention definitions \\[[@CR24]\\].\n\nMicrobiological methods {#Sec5}\n-----------------------\n\nBlood cultures (aerobic and anaerobic) were processed using the BD BACTEC\u2122 system (Becton Dickinson, Sparks, MD, EUA), according to the routine of hospital microbiology laboratory. Identification and antibiotic susceptibility testing of any culture isolate were performed by VITEK\u00ae2 (BioM\u00e9rieux, Hazelwood, MO, USA) system, and confirmed by disk diffusion or E-test, according to the updated recommendations of Clinical and Laboratory Standards Institute \\[[@CR25]\\] and European Committee on Antimicrobial Susceptibility Testing \\[[@CR26]\\], including the use of meropenem, imipenem and ertapenem for all Gram-negative bacterial species, except *Stenotrophomonas maltophilia* which is naturally resistant to carbapenems. Carbapenemase production was investigated by phenotypic tests with phenylboronic acid and ethylene diamine tetra acetic acid.\n\nGram-negative bacterial isolates detected in blood and other cultures were referred to Laborat\u00f3rio de Pesquisa em Infec\u00e7\u00e3o Hospitalar for microbiological confirmation by using classical and molecular biochemical methods as described in previous publications \\[[@CR27], [@CR28]\\]. The search for the following genes of carbapenemases of Amber class A (*bla*~KPC-2~), B (*bla*~SPM-1~, *bla*~NDM-1,~ *bla*~VIM~) and D (*bla*~OXA-23-like~, *bla*~OXA-48-like~ and *bla*~OXA-51-like~) was performed by using in-house multiplex polymerase chain reaction (PCR) test.\n\nWe classified the antimicrobial susceptibility profile of the strains in multi-drug (MDR), extensively-drug (XDR) and pan-drug resistant (PDR) and described as \"possible\" profiles whenever not all antimicrobials of all selective classes for each bacterial group or species were tested, according to Magiorakos et al. (2012) \\[[@CR29]\\]. *S. maltophilia* and *Burkholderia cepacia* were considered MDR.\n\nSample size and statistical analysis {#Sec6}\n------------------------------------\n\nConsidering an alpha error of 5%, a power of 80%, a control to case ratio of 1:1, and respectively 40 and 18% exposure to carbapenem among cases and controls \\[[@CR19]\\], the sample size estimated was 152 patients.\n\nThe findings were used to build a model with clinical-epidemiological factors that can be easily identified by physicians during the first moment of patient evaluation, at a time when only clinical-epidemiological parameters can guide empirical antimicrobial therapy. All variables were analyzed using SPSS\u00ae statistics v22.0 software. Categorical variables were compared using Chi-Square or Fisher's exact test and for continuous variables, the Mann-Whitney-Wilcoxon test was used. Collinearity was investigated initially using Pearson correlation matrix and cross-tabulations between two or more variables \\[[@CR30]\\]. All variables investigated as predictors were explored in univariate and multivariate logistic regression analyses using the complete data set to identify independent risk factors for CR-GNB and CS-GNB sepsis \\[[@CR31]\\].\n\nTo optimize the model, we used our best knowledge not to include variables with collinearity together and give chances to those clinically meaningful. Our database has many variables that may be related to our outcome and correlated with each other. Although all variables were considered, those with small frequency and those with collinearity had to be excluded from multivariate analysis to improve the fit of the model. Possible interactions were also investigated. Using the approach described above, different models were evaluated but the best-fit model came with backward selection procedure. We also have used robust fit criteria for model comparisons (AIC and BIC) \\[[@CR32]\\]. Both sides of the curve and significance level of 5% were considered in all tests.\n\nResults {#Sec7}\n=======\n\nStudy population and clinical characteristics {#Sec8}\n---------------------------------------------\n\nAmong the total of 629 ICU admissions followed by 7797 patient-days, we evaluated 342 episodes of SIRS/sepsis detected (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). After applying the exclusion criteria, we enrolled a total of 184 patients: 60 patients who acquired CR-GNB sepsis, 30 patients with CS-GNB sepsis, and each group was compared with 94 patients with undetermined (*n*\u2009=\u200978) or non GNB (*n*\u2009=\u200916) sepsis. Demographic and clinical characteristics of included patients are shown in Tables [1](#Tab1){ref-type=\"table\"} and [2](#Tab2){ref-type=\"table\"}. Ninety-seven percent (29/30) of CS-GNB cases had severe sepsis-2, as well as 85% (51/60) of CR-GNB and 84% (79/94) of the controls. While 94% (51/54) of CR-GNB sepsis, 100% of CS-GNB cases (27/27) and 87% (74/85) of control group that could have their episode evaluable (166/184, 90%) fulfilled sepsis-3 criteria. Eighteen patients (10%, 18/184) were considered not evaluable for the modified sepsis-3 criteria for being on mechanical ventilation and sedation. Although these patients did not have delta SOFA \u22652, they had high median SOFA score at ICU admission (5.5, range 0--15) and on the date of initial blood culture collection (4, range 0--15). Septic shock was detected in 72% (43/60) of CR-GNB sepsis, 70% (21/30) of CS-GNB and 64% (60/94) of control patients with 30-day all-cause mortality rate of 50% (30/60), 40% (12/30) and 45% (42/94), respectively. Fig. 1Flowchart of patients included in the study. CR-GNB carbapenem-resistant Gram-negative bacilli; CS-GNB carbapenem-susceptible Gram-negative bacilli\n\nThe plausibility of infectious source was categorized as definitive or probable in 89% (25/28) and 84% (54/64) of CS-GNB and CR-GNB sepsis, respectively, while the sources of infection were determined in 64% (60/94) of controls. Bacteremia was identified in 25% (46/184) of all studied patients, and 43% (20/46) of them were diagnosed as probable catheter related bloodstream infections (BSI).\n\nCR-GNB cases were followed for the median of 47\u2009days (range 1--351\u2009days), while CS-GNB cases and controls for 40.5 and 37\u2009days (range 1--71\u2009days in both groups), respectively, during ICU stay until the end of follow-up period. The median length of ICU stays prior to the sepsis episode investigated was 10 times higher among CR-GNB cases (10\u2009days, range 0--291) than CS-GNB cases (1-day, range 0--34) (*p*\u2009=\u20090.006). Half of sepsis episodes were detected at ICU admission**.** Sepsis before 48\u2009h of ICU hospitalization occurred in 67% (20/30) of CS-GNB cases group, similar to the controls (57%, 54/94), and was statistically different among CR-GNB cases (20%, 12/60) (OR\u2009=\u20098.0; 95% CI 3.0--21.5; *p*\u2009\\<\u20090.001). Therefore, the majority of later sepsis was caused by CR-GNB.\n\nNon-fermenting bacilli corresponded to 73% (49/67) of the etiology of CR-GNB sepsis, and *A. baumannii* accounted for 43% of the occurrences (26/60) (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a). CR-GNB isolates were detected in tracheal aspirate or bronchoalveolar lavage (BAL) (59%, 37/63), blood (35%, 22/63), urine (3%, 2/63) and operative wound samples (3%, 2/63). The majority of isolates (83%, 45/54) had MIC \u226516\u2009mg/ml for meropenem and/or imipenem. Two isolates of *Klebsiella pneumoniae* had MIC \u226516\u2009mg/ml for polymyxins and 8% (4/52) of strains belonged to polymyxins naturally resistant species. CR-GNB strains causing sepsis were 89% MDR (48/54), 69% (37/54) possible extensively drug-resistant and 2% (1/54) had possible pandrug-resistant profile. The production of carbapenemases was estimated in 77% (39/51) of the isolates tested by phenotypic tests (92%, 12/13) and/or by PCR technique (74%, 35/47), since we included tested (*n*\u2009=\u20092) but also non-tested *S. maltophilia* isolates (*n*\u2009=\u20095). Fig. 2**a** Distribution of the etiological agents of sepsis by CR-GNB (*n*\u2009=\u200967) and CS-GNB (*n*\u2009=\u200936); **b** CR-GNB isolates tested (*n*\u2009=\u200942) and respective carbapenemase-producing genes detected (*n*\u2009=\u200944). CR-GNB, carbapenem resistant Gram-negative bacilli; CS-GNB, carbapenem-susceptible Gram-negative bacilli\n\nFigure\u00a0[2](#Fig2){ref-type=\"fig\"}b shows the CR-GNB isolates investigated (*n*\u2009=\u200942) and respective carbapenemase-producing genes detected (*n*\u2009=\u200944). Species of the Enterobacteriaceae family (69%; 25/36) outweighed among the etiologies of CS-GNB sepsis (Fig. [2](#Fig2){ref-type=\"fig\"}a). The isolates were recovered from blood (33%, 11/33), tracheal aspirate or BAL (30%, 10/33), operative wound samples (21%, 7/33) and urine (15%, 5/33). The susceptibility profiles of the isolates were 79% (27/34) non-MDR and 21% (7/34) classified as MDR. Polymyxin intrinsic resistance was detected in 13% (4/31) of CS-GNB isolates.\n\nVentilator-associated pneumonia (VAP) was the most frequent infectious source among CR-GNB sepsis (48%; 32/66) with the predominance of *A. baumannii* in 56% (18/32) of etiological agents. Among CS-GNB sepsis, respiratory tract infections (VAP, hospital-acquired pneumonia and tracheobronchitis) occurred in 37% (11/30). Intra-abdominal surgical site (SSI) plus urinary-tract (UTI) infections prevailed in CS-GNB sepsis (43%; 13/30) than CR-GNB sepsis (11%; 7/66) (*p*\u2009\\<\u20090.001). While in the control patients, excluding cases with undetermined focus (36%, 34/94), hospital-acquired pneumonia (32%, 21/66) and surgical wound (32%, 21/66) predominated as source of sepsis.\n\nCR-GNB septic patients versus control patients {#Sec9}\n----------------------------------------------\n\nThe results of comparative univariate analysis between cases of CR-GNB sepsis versus controls are shown in Table\u00a0[1](#Tab1){ref-type=\"table\"}. Previous infection was an important factor associated with CR-GNB sepsis (77%; 46/60) when compared to control group, with high percentage of healthcare-associated infection (57%; 26/46) and 48% (22/46) of previous sepsis at the same hospitalization period. Previous infection was predominantly of bacterial origin among CR-GNB cases (50%; 30/60) than controls (27%; 25/94) (OR\u2009=\u20092.74; 95% CI 1.39--5.48*; p*\u2009=\u20090.004). Sepsis as a reason for ICU admission occurred in 52% (31/60) of the CR-GNB cases, contrasting with only 32% (30/94) of controls (OR\u2009=\u20092.27; 95% CI 1.16--4.46; *p*\u2009=\u20090.016). Sepsis recurred during ICU stay in 65% (39/60) of CR-GNB cases and only 19% (18/94) of controls (OR\u2009=\u20097.71; 95% CI 3.72--16.51; *p*\u2009\\<\u20090.0001).\n\nThere was no difference between cases (77%; 46/60) versus controls (79%; 74/94) in respect to the use of antibiotics during the 3 days prior to the collection of blood culture (*p*\u2009=\u20090.76), including antibiotics active on GNB (70% vs. 73%; *p*\u2009=\u20090.64) or polymyxins (28% vs. 18%; *p*\u2009=\u20090.13) were used. However, there was a difference in carbapenem consumption among these groups (36/60, 60% vs. 39/94, 41%; OR\u2009=\u20092.11; 95% CI 1.09--4.12; *p*\u2009=\u20090.026).\n\nThe independent risk factors for CR-GNB sepsis are previous infection (mostly hospital-acquired bacterial infection or sepsis) (OR\u2009=\u20094.28; 95% CI 1.77--10.35; *p*\u2009=\u20090.001), previous use of mechanical ventilation (OR\u2009=\u20094.21; 95% CI 1.17--15.18; *p*\u2009=\u20090.028) and carbapenems (OR\u2009=\u20093.42; 95% CI 1.37--8.52; *p*\u2009=\u20090.008), and length of hospital stay (OR\u2009=\u20091.03; 95% CI 1.01--1.05; *p*\u2009=\u20090.007) (Table\u00a0[3](#Tab3){ref-type=\"table\"}). Table 3Independent predictive factors associated with CR-GNB sepsis; multivariate logistic regressionRisk factors for sepsis by CR-GNBCR-GNB cases (*n*\u00a0=\u200960)Control group (*n*\u00a0=\u200994)Odds ratio95% CI*p*^a^ valuePrevious infection, n (%)46 (77)31 (33)4.281.77--10.350.001Mechanical ventilation, n (%)56 (93)58 (62)4.211.17--15.180.028Use of carbapenem, n (%)48 (80)32 (34)3.421.37--8.520.008Length of hospital stay (days), median (range)26 (1--375)10 (0--143)1.031.01--1.050.007*Abbreviations*: *CR-GNB* carbapenem-resistant Gram-negative bacilli, *CI* confidence interval^a^Wald test for logistic regression, *p* significant \\<\u20090.05, accuracy of 80% for CR-GNB model\n\nCS-GNB septic versus control patients {#Sec10}\n-------------------------------------\n\nTable\u00a0[2](#Tab2){ref-type=\"table\"} shows the comparative univariate analysis between cases of CS-GNB sepsis versus controls. Repeated episodes of sepsis occurred during ICU stay in 23% (7/30) of CS-GNB cases with no significance when compared to controls (*p*\u2009=\u20090.619). Regarding the use of antimicrobial agents during the 3 days prior to the initial blood culture that could interfere with its result, there was statistically significant difference between CS-GNB sepsis (53%; 16/30) and controls (79%; 74/94) (OR\u2009=\u20090.31; 95% CI 0.13, 0.75; *p*\u2009=\u20090.009). The same occurred with the use of carbapenems (3/30, 10% vs. 39/94, 41%; OR\u2009=\u20090.16; 95% CI 0.03--0.57; *p*\u2009=\u20090.0009). However, patients with CS-GNB sepsis did not receive polymyxins previously, but control patients received them (18%; 17/94).\n\nIn multivariate analysis, readmission to the ICU (OR\u2009=\u20096.92; 95% CI 1.72--27.78; *p*\u2009=\u20090.006) and prior nosocomial diarrhea (OR\u2009=\u20095.32; 95% CI 1.07--26.45; *p*\u2009=\u20090.041) were detected as independent risk factors for developing CS-GNB sepsis in the study population.\n\nDiscussion {#Sec11}\n==========\n\nIn this study, patients with longer length of hospital stay, previous infection, mostly hospital acquired bacterial infection, who had been previously treated with mechanical ventilation and carbapenems presented a higher risk of sepsis due to CR-GNB than control group. Whereas readmission to the ICU and prior nosocomial diarrhea were factors associated with CS-GNB sepsis in the study population.\n\nPrevious infection has been rarely reported as risk for infection/colonization or bacteremia by CR *A. baumannii* \\[[@CR7], [@CR8]\\]. In our cohort, previous infection was the most important factor for CR-GNB sepsis, mainly of bacterial origin, mostly nosocomial infection and 48% previous sepsis. Recurrent pattern of sepsis was a striking feature of our studied population. Therefore, more attention for the prevention and control of nosocomial sepsis is required for the prevention of subsequent hospital-acquired CR-GNB sepsis, especially caused by *A. baumannii*, at the same hospitalization period in ICU patients.\n\nThe studies that have evaluated risk factors for CR-GNB or CR *A. baumannii* bacteremia, SIRS or nosocomial infection have found mechanical ventilation, respiratory failure, but also VAP as risk factors in critically ill patients \\[[@CR7], [@CR8], [@CR18], [@CR19]\\]. In our study we did not investigate types of infection as predictive factors because we sought to investigate variables that would readily discriminate patients with increased risk, in order to be useful to guide empirical therapy in future.\n\nProlonged hospital stay is a classically recognized risk for hospital and ICU infection \\[[@CR10], [@CR11], [@CR14], [@CR19], [@CR33]\\]. We found a 2% increase in the chances of developing sepsis by CR-GNB for every day of hospitalization. ICU readmission is also a documented risk factor for the acquisition of CS-GNB infection \\[[@CR10], [@CR16]\\]. A few studies have found diarrhea to be associated with GNB bacteremia \\[[@CR34]--[@CR36]\\].\n\nAs far as we know, this is the first risk factor study in CR-GNB sepsis that has considered repetitive episodes and not only bacteremia, but the broad variety of infection sources commonly observed in ICU sepsis. Studies that have investigated risk factors for infections by CR-GNB species are still rare \\[[@CR18], [@CR19]\\]. In general, the studies focus on risk factors for infections by specific species such as *K. pneumoniae* \\[[@CR12], [@CR13], [@CR16]\\] or other members of the Enterobacteriaceae \\[[@CR15], [@CR17]\\], *P. aeruginosa* \\[[@CR10], [@CR11]\\] and *A. baumannii* \\[[@CR7]--[@CR9]\\], and most of them in hospital infection \\[[@CR9], [@CR10], [@CR12], [@CR15]--[@CR17]\\] or bacteremia \\[[@CR8], [@CR11], [@CR18], [@CR19]\\]. However, bacteremia has been detected in less than 30% of septic cases in ICU \\[[@CR37], [@CR38]\\]. The extensive use of vascular catheter has been recognized as the most important factor contributing to BSI \\[[@CR4], [@CR39], [@CR40]\\], while VAP has predominated as source of sepsis in ICU \\[[@CR38]\\]. Therefore, although bacteremia can be considered the gold standard it represents only part of the population who were diagnosed with sepsis in ICU and is frequently associated with vascular catheter infection. In addition, several risk factor studies select the first episode of infection or bacteremia only \\[[@CR8]--[@CR10], [@CR13], [@CR16]--[@CR19]\\]. The successful longitudinal follow-up of this cohort study allowed identifying patients who presented recurrent sepsis, whose diagnoses were also essential to better select cases and controls. Giving the chance of inclusion of all episodes and the variety of sepsis infections may be more reliable to better discriminate patients at higher risk in ICU.\n\nThe case-case-control design is more effective for the identification of risk factors for antimicrobial resistant pathogens, avoiding bias of exposure to the antimicrobial of interest, for not using CS-GNB sepsis as control group \\[[@CR41], [@CR42]\\]. Our study aimed to confirm exposure to carbapenems as a risk for the development of CR-GNB sepsis \\[[@CR11], [@CR12], [@CR18], [@CR19], [@CR43]\\], since our control group was composed of patients with sepsis and antimicrobial treatment. This methodology allows the application of the factors found in the management of empirical therapy, which would not be possible if the control was a patient without sepsis and antimicrobial treatment of the episode investigated \\[[@CR42]\\].\n\nThe design resulted in the selection of cases so efficiently that the study detected a remarkable difference in the etiologies of sepsis by CR-GNB and CS-GNB. Indeed, the CR-GNB sepsis presented a predominance of non-fermenting bacteria, mainly *A. baumannii*, which was likely related to the higher prevalence of respiratory tract infections as source of sepsis \\[[@CR18], [@CR19]\\]. Whereas Enterobacteriaceae species predominated among CS-GNB sepsis, mostly associated with respiratory tract infections, but also intra-abdominal SSI and UTI.\n\nEvaluated sepsis episodes in cases and control patients had a good agreement when considering the adapted parameters for the diagnosis of sepsis-3 \\[[@CR21]\\]. The plausibility of infectious source as definitive and probable occurred in a high proportion in all groups of cases while the source of infection was determined in the majority of controls.\n\nOur data characterize sepsis by CS-GNB mostly as hospital-acquired infection outside the ICU, whereas sepsis by CR-GNB was mainly ICU-acquired infection. This issue corroborates with a well-known fact that ICU patients have more risk of infection by resistant bacteria \\[[@CR33]\\]. In fact, CR-GNB species typify a bacterial population with a high level of resistance and few treatment options, due to the often high carbapenems MIC and some combined resistance to polymyxins. Molecular investigation of carbapenemase production shows that we have determined the risk factors for CR-GNB sepsis predominantly with this mechanism of resistance, which is commonly described and disseminated worldwide \\[[@CR43]--[@CR47]\\].\n\nMain limitation of this study may be described as a single center study, which indicates caution to any generalization of our findings. Adapting management strategies to the local epidemiological data is a general recommendation for the prevention and control of hospital-acquired infection, which indicates the performance of cyclic evaluations locally. The methodology of case-control selection was one of the study's strengths. The performance of microbiological methods and the use of antimicrobials, which may have inhibited microbial growth in clinical cultures, could have influenced the risk factors results. Other non-investigated elements may also have interfered. Phylogenetic analysis of GNB isolates would have improved our knowledge about the epidemiological context of GNB hospital-acquired sepsis.\n\nConclusions {#Sec12}\n===========\n\nProlonged hospitalization with the development of healthcare-associated infection, requiring mechanical ventilation and treatment with carbapenems seem to be the natural history for subsequent sepsis by carbapenemase-producing GNB in this population. The concordance with the background knowledge suggests that these factors should be evaluated further for developing and validating a risk score to identify patients at higher risk for CR-GNB sepsis in ICU. Little is known about the influence of recurrent sepsis during the same hospitalization period. Consequently, investigation of repetitive episodes of sepsis in ICU patients is warranted.\n\nSupplementary information\n=========================\n\n {#Sec13}\n\n**Additional file 1 Appendix S1**. Study Form 1, with variables investigated in this study. **Appendix S2**. Study Form 2, with variables investigated in this study. **Appendix S3**. Study Form 2a, with variables investigated in this study. **Appendix S4**. STROBE Statement---Checklist of items included in this study.\n\nBAL\n\n:   Broncho alveolar lavage\n\nBSI\n\n:   Bloodstream infection\n\nCS\n\n:   Carbapenem-susceptible\n\nCR\n\n:   Carbapenem-resistant\n\nGNB\n\n:   Gram-negative bacilli\n\nICU\n\n:   Intensive care units\n\nMIC\n\n:   Minimum Inhibitory Concentration\n\nMDR\n\n:   Multidrug resistant\n\nPCR\n\n:   Polymerase chain reaction\n\nSAPS-3\n\n:   Simplified Acute Physiology Score III\n\nSIRS\n\n:   Systemic Inflammatory Response Syndrome\n\nSSI\n\n:   Surgical site infection\n\nSOFA\n\n:   Sequential Organ Failure Assessment\n\nUTI\n\n:   Urinary-tract infection\n\nVAP\n\n:   Ventilator-Associated Pneumonia\n\nXDR\n\n:   Extensively drug resistant\n\nPDR\n\n:   Pan-drug resistant\n\nPartly presented in the 13th World Congress of Intensive and Critical Care Medicine, Rio de Janeiro, Brazil, 2017. Abstract Number EP-099.\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary information\n=========================\n\n**Supplementary information** accompanies this paper at 10.1186/s13756-020-00791-w.\n\nNucleus of Hospital Research collaborators: Gl\u00f3ria Regina A. A. Magalh\u00e3es^2^, Priscila P. Silva^3^, Scyla Maria S. A. S. Reis Di Chiara^2^, Amanda Aparecida S. Machado^3^, Thaisa M. Tozo^2^, Lucas Lameir\u00e3o P. A. Rosas^3^, Paulo C. M. Barros^2^, Wania V. de Freitas^2^, Ana Paula D. C. Assef^1^**.** The research leading to these results is supported by the Technical Cooperation Agreement (number 51/2017) between FIOCRUZ and HFSE to develop research.\n\nEmail address of study collaborators: gramagal\\@hotmail.com, priscila.pinho\\@ioc.fiocruz.br, scylareiss\\@gmail.com, amandacidamachado\\@gmail.com, thaisa.mdrs\\@gmail.com, lucas.rosas40\\@gmail.com, paulocesar.unirio\\@outlook.com, freitas.wania\\@gmail.com, anapdca\\@ioc.fiocruz.br\n\nSupervised the study: MZRG Designed the analytic plan: MZRG, SCSG Obtained scholarships: MZRG Enrolled patients: MZRG, EML Collected data: EML, PAC, DSB, LHZP, JPST, NDL, CSB Created and managed the database: MZOA, MZRG Analyzed and/or interpreted the data: SCSG, MZRG, EML, CASR, MDA, RQS, JAM, APDCA Acted as supervisor for data analysis and/or interpretation: MZRG. Drafted the initial manuscript: EM L Reviewed and/or revised the manuscript: MZRG, SCSG. All authors read and approved the final manuscript.\n\nResearch Support Foundation of the State of Rio de Janeiro (FAPERJ) and National Council for Scientific and Technological Development (CNPq) for providing scholarships for scientific initiation for undergraduate medical students.\n\nThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\n\nThis study was approved by FIOCRUZ and HFSE Ethics Committees (CAAE: 39978114.5.0000.5248 and CAAE 39978114.5.3002.5252).\n\nNot applicable.\n\nThe authors declare that they have no competing interests\n"}
+{"text": "Introduction\n============\n\nEarly gastric cancer refers to gastric cancer confined to the mucosa or submucosa, regardless of local lymph node metastasis. The postoperative survival rate after 5 years of early gastric cancer is \\>90% ([@b1-ol-0-0-6376]). With the use of confocal endoscopy, narrow band imaging, and magnification chromoendoscopy, an increasing number of gastric cancers are found in the early stage ([@b2-ol-0-0-6376]).\n\nThe preferred treatment for early gastric cancer is traditional radical surgery with lymph node dissection, but the rate of postoperative complications is up to 43%. Radical resection alters the normal anatomical structure and can cause postoperative dysphagia, regurgitation, loss of appetite, weight loss and other symptoms, severely reducing the quality of life of the patient ([@b3-ol-0-0-6376]). Endoscopic mucosal resection (EMR) can simultaneously complete the diagnosis and treatment of gastric cancer in one operation. With its resultant reduced risk of trauma, rapid recovery and high complete resection rate, its application gradually increases in the clinical setting ([@b4-ol-0-0-6376]). Macrophage inhibitory cytokine (MIC)-1 was recently found to have a close relationship with the occurrence of gastric cancer, and its high expression often indicates a poor long-term prognosis ([@b5-ol-0-0-6376]).\n\nThe findings from research on EMR treatment of early gastric cancer at our center is summarized as follows.\n\nPatients and methods\n====================\n\n### Patient data\n\nOne hundred forty-seven cases of gastric cancer were diagnosed via biopsy with pathological confirmation at the Xuzhou Hospital Affiliated to Medical College of Southeast University (Jiangsu, China), from January 2010 to January 2013, and were selected continuously for participation in the study. Inclusion criteria were: i) Age \u226518 years; ii) the tumor diameter was \\<20 mm and not associated with erosion or ulcer; and iii) infiltration depth and scope of the mucosa had enough safety margins for resection. Exclusion criteria were: i) Patients with coagulopathy; ii) patients with other malignant tumors; iii) patients with comorbidities of the heart, liver and/or kidney; and iv) patients not willing to participate in the study.\n\nThis study was approved by the Ethics Committee of Xuzhou Hospital Affiliated to Medical College of Southeast University. Signed written informed consents were obtained from all participants before the study. Patients were informed of the detailed surgical conditions and the possible risks of adverse events and complications before surgery and informed written consent of the patients or their families was obtained. The patients were then divided into the observation group (75 cases) and the control group (72 cases), according to the chosen treatment method. There were 41 male and 34 female cases in the observation group. Ages ranged from 37 to 75, with an average of 51.9\u00b16.2 years. There were 46 cases with a differentiated type of morphology and 29 cases with an undifferentiated type. Thirty-five cases were intra-mucosal cancer, whereas 40 cases were submucosal cancer. The control group consisted of 44 male and 28 female cases. Ages ranged from 35 to 76, with an average of 52.1\u00b17.2 years. There were 40 cases with a differentiated type of morphology and 32 cases with an undifferentiated type. Thirty-one cases were intra-mucosal cancer, whereas 41 cases were submucosal cancer. The distribution of these general factors among patients between the two groups showed no statistical differences (P\\>0.05).\n\n### Research methods\n\nThe two groups of patients were treated using the same surgical and healthcare team, according to the standard medical process. The control group underwent radical resection with conventional laparotomy as follows: Administration of general anesthesia, access by laparotomy into the lateral rectus side, cutting off most of the gastric tissue along the transverse colon and freeing the greater omentum behind, and then reconstructing the digestive tract, with placement of a conventional drainage tube.\n\nPatients of the observation group were treated using EMR, as follows: Endoscopy was performed for observation of the lesion size and shape, and the depth of invasion. After the lesion was determined, the junction between the lesion and normal mucosa was stained with indigo carmine, and electric coagulation markers were made every 0.5 cm, starting at 0.5 cm away from the junction. Using mixed liquid containing glycerol fructose and methylene blue to inject at multiple points, each point was injected with \\~2 ml. The injection was repeated until the lesion became notably elevated relative to the normal mucosa. After the lesion was ridged upwards, the disposable snare was stretched into the lesion from the biopsy channel until the lesion was completely set into the snare. Once the snare was incised into the muscularis propria, it was activated and the lesion was resected. After the excision, electrocautery was used for coagulation and hemostasis of blood vessels. Lesion tissue was fixed with neutral formalin and sent for pathological examination. Following the procedure, we carried out routine postoperative measures to ensure infection prevention, fluid infusion and hemostasis. In cases where perforation occurred during the surgery requiring suturing of the wound, delayed postoperative feeding time was necessary.\n\n### Observation index\n\nA follow-up period was used to determine and compare progression-free survival and recurrence rates between the two groups. The deadline for follow-up of patients was January 2016 and the average follow-up period was 58.5 months. The amount of bleeding, operation time and postoperative complication rate were also compared between the two groups. The serum MIC-1 levels in the preoperative period and 3 months postoperative were compared. Detection of MIC-1 was as follows: Collection of \\~5 ml of early-morning fasting peripheral venous blood was centrifuged at 3,000 rpm for 20 min, and then stored at \u221220\u00b0C. Concentration of MIC-1 was then tested, using the standard double antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) detection methods with strict adherence to protocol (DAS-ELISA kit was purchased from the Beijing Zhongshan Jinqiao Biological Co., Ltd., Beijing, China). A routine abdominal CT scan was performed after the surgery at one, six, and then at six-month intervals with enhanced scanning used as necessary to evaluate any progression of the tumor.\n\n### Statistical analysis\n\nSPSS 22.0 (Chicago, IL, USA) was used to process the data and for statistical analysis. The data were expressed as standard deviation (mean \u00b1 SD) and enumeration data were expressed as cases or (%). Comparisons between groups were made using independent sample t-tests, and paired t-tests were used within groups for repeated measures. Further analyses of factors between groups were done using revised Chi-square tests and log-rank Chi-square tests. Kaplan-Meier survival analysis was used for assessment of progression-free survival. For all tests, P\\<0.05 was considered statistically significant.\n\nResults\n=======\n\n### Comparison of blood loss, operative time and postoperative complication rate\n\nThe lesion total resection rate was 100% and the resection margins were negative for both the observation and the control groups. The amount of bleeding, operation time and postoperative complication rate in the observation group were significantly lower than those in the control group ([Table I](#tI-ol-0-0-6376){ref-type=\"table\"}), with the difference being statistically significant (P\\<0.05).\n\n### Comparison of serum MIC-1 levels between the pre-operative period and 3 months post-operative\n\nComparison of preoperative serum MIC-1 levels between the two groups showed no statistical difference (P\\>0.05). Postoperative serum MIC-1 levels in the two groups decreased compared with the preoperative levels, and the serum MIC-1 levels in the observation group were significantly lower than those in the control group ([Table II](#tII-ol-0-0-6376){ref-type=\"table\"}), with the difference statistically significant (P\\<0.05).\n\n### Comparison of progression-free survival and recurrence in follow-up\n\nThe progression-free survival was more prolonged and the recurrence rate was lower in the observation group compared to the control group ([Table III](#tIII-ol-0-0-6376){ref-type=\"table\"} and [Fig. 1](#f1-ol-0-0-6376){ref-type=\"fig\"}), with the differences both being statistically significant (P\\<0.05).\n\nDiscussion\n==========\n\nGastrointestinal cancer, especially gastric cancer, has high incidence and mortality rates in China. Primary reasons for the high rates are insufficient patient screening and public health surveillance, limiting the capacity for early diagnosis. Most gastric cancers are in middle or advanced stages when diagnosed clinically ([@b6-ol-0-0-6376]). Early gastric cancers are confined to the mucosa and submucosa tissue, making the risk of lymph node metastasis relatively low. Endoscopic examination can enable early detection, diagnosis and treatment of gastric cancer before it reaches advanced stages. Minimally invasive endoscopic treatment for early gastric cancer, compared with conventional surgery, also has the advantages of reduced risk of trauma, less postoperative complications, less intensive operative procedures, and can effectively improve the life quality of patients. While clear clinical efficacy is evident for treatment of early gastric cancer by radical resection, the disadvantages are an increased risk of trauma, slow postoperative recovery, delayed transfer time, prolonged hospitalization and greater healthcare costs, resulting in greater potential for physical and mental health issues and economic burden for the patient ([@b7-ol-0-0-6376]). Therefore, EMR has become one of the most effective methods for the diagnosis and treatment of early gastric cancer.\n\nCurrently, there are four kinds of EMR: Peeling off biopsy resection, standard EMR, cap-EMR and EMR with ligation. Standard EMR is applied most often, the diseased tissue is cut off after being lifted completely with a snare ([@b8-ol-0-0-6376]). It can be difficult to cut the whole lesion using EMR when the diameter exceeds 20 mm. Therefore, it can become necessary to use endoscopic piece mucosal resection (EPMR); however, studies show ([@b9-ol-0-0-6376]) postoperative recurrence rate following EPMR is relatively higher than EMR. Broken specimens from EPMR are also difficult to use for accurate determination of surgical margins and depth of invasion.\n\nThrough this study, the complete resection rate was 100% for both the observation and the control groups. The amount of bleeding, operation time and postoperative complication rate in the observation group were significantly lower than those in the control group, which suggests that EMR has better safety and efficacy. The serum MIC-1 levels of the two groups of patients were both reduced, with the observation group significantly lower than the control group.\n\nMIC-1, as a member of the TGF-\u03b2 family, is widely involved in cell apoptosis, metastasis and invasion, and other biological mechanisms ([@b10-ol-0-0-6376]). In normal serum, mature MIC-1 protein have low levels of expression, whereas in pathological conditions, such as malignancy, inflammation and acute injury, MIC-1 expression levels can be significantly increased, and the duration of its expression significantly longer ([@b11-ol-0-0-6376]). Kim *et al* ([@b12-ol-0-0-6376]) found that the MIC-1 gene was overexpressed in gastric cancer tissue, and its role varied in different stages of the development of gastric cancer. Lee *et al* ([@b13-ol-0-0-6376]) found that the expression of MIC-1 mRNA increased significantly with chronic inflammatory reaction, precancerous lesions and gastric cancer development, as well as invasion and metastasis of gastric carcinoma, indicating that MIC-1 expression accompanies tumor progression, and is closely related with tumor growth, invasion and metastasis. MIC-1 mRNA expression of gastric cancer with lymph node metastasis or distant metastasis was significantly higher than that without metastasis, indicating that expression is related to tumor metastasis. MIC-1 acts as a tumor suppressor in early gastric cancer by several mechanisms, including activating the P53 pathway ([@b14-ol-0-0-6376]), inhibiting COX-2 expression ([@b15-ol-0-0-6376]), inducing caspase-8 pathway ([@b16-ol-0-0-6376]), and activating the phosphatidylinositol 3-kinase/serine/threonine protein kinase and glycogen synthase 3\u03b2 signaling pathway ([@b17-ol-0-0-6376]). Conversely, it is also involved in advanced gastric cancer with promoting effects by inhibiting catenin \u03b41 gene expression ([@b18-ol-0-0-6376]), upregulating uPA system to enhance invasiveness of gastric cancer cells ([@b19-ol-0-0-6376]) and inducing the overexpression of ErbB2 receptor tyrosine kinase in gastric cancer cells ([@b12-ol-0-0-6376]). These and other mechanisms illustrate its role as an indicator of prognosis for early gastric cancer cases. Therefore, MIC-1 expression may be a manifestation of poor prognosis.\n\nOur study showed that treatment of early gastric cancer with EMR resulted in significantly prolonged progression-free survival and significantly reduced rates of operative complications and disease recurrences compared to traditional surgery, suggesting that EMR in the treatment of early gastric cancer is both safe and effective. EMR can also reduce the postoperative expression of serum MIC-1, an important index for predicting the prognosis of the disease.\n\n![Survival rate of recurrence-free patients in the two groups.](ol-14-02-1967-g00){#f1-ol-0-0-6376}\n\n###### \n\nComparison of blood loss, operative time and postoperative complication rate.\n\n  Groups          Cases   Blood loss in operation (ml)   Operative time (min)   Blood loss after operation (%)   Infection (%)   Perforation (%)   Others (%)   Complication incidence (%)\n  --------------- ------- ------------------------------ ---------------------- -------------------------------- --------------- ----------------- ------------ ----------------------------\n  Observation     75      151.2\u00b138.4                     58.3\u00b114.5              3 (4.0)                          1 (1.3)         1 (1.3)           1 (1.3)      6 (8.0)\n  Control         72      258.3\u00b141.2                     86.4\u00b123.2              6 (8.3)                          2 (2.8)         3 (4.2)           3 (4.2)      14 (19.4)\n  t (\u03c7^2^) test           6.495                          7.147                                                                                                  4.093\n  P-value                 0.033                          0.026                                                                                                  0.043\n\n###### \n\nComparison of serum MIC-1 levels between the pre-operation and 3 months post-operation (pg/ml).\n\n  Groups        Pre-operation   Post-operation   t-test   P-value\n  ------------- --------------- ---------------- -------- ---------\n  Observation   237.5\u00b142.3      75.6\u00b114.5        13.630   \\<0.001\n  Control       226.8\u00b157.8      124.3\u00b138.2       6.134    \u00a0\u00a00.034\n  t-test        0.821           6.963                     \n  P-value       0.234           0.028                     \n\nMIC-1, macrophage inhibitory cytokine-1.\n\n###### \n\nComparison of progression-free survival and recurrence in follow-up.\n\n  Groups        Cases   Progression-free survival (months)   Recurrence in follow-up of 1 year   Recurrence in follow-up of 3 years   Recurrence in follow-up of 5 years\n  ------------- ------- ------------------------------------ ----------------------------------- ------------------------------------ ------------------------------------\n  Observation   75      46.8                                 2 (2.7)                             4 (5.3)                              6 (8.0)\n  Control       72      37.9                                 3 (4.2)                             11 (15.3)                            14 (19.4)\n  \u03c7^2^                  15.632                               0.002                               3.965                                4.093\n  P-value               \\<0.001                              0.963                               0.046                                0.043\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nPeriodontal diseases are prevalent globally, with approximately 60% of the adult population suffering from mild, moderate, or aggressive periodontitis \\[[@B1], [@B2]\\]. As a chronic inflammatory disease, periodontitis is characterized by the destruction of the gingiva, periodontal ligament, cementum, and alveolar bone and is initiated by the invasion of specific oral pathogens that colonize dental plaque biofilms on the tooth surface \\[[@B3]--[@B5]\\]. The pathogenesis of periodontitis is complex, and excessive tissue destruction occurs as a result of interactions between pathogenic bacteria and the host immune inflammatory response \\[[@B3]\\]. Previous studies have confirmed that periodontitis is commonly caused by gram-negative bacterial infection such as *Porphyromonas gingivalis* (*P. gingivalis*), *Tannerella forsythia*, and *Aggregatibacter actinomycetemcomitans* \\[[@B6]--[@B9]\\]. However, *Fusobacterium nucleatum* (*F. nucleatum*), which is also a fundamental periodontal pathogen, has not been extensively studied in the field of periodontics, and the pathogenicity of *F. nucleatum* has not been extensively studied in oral tissues and cells.\n\n*F. nucleatum*, as a bridging bacterium, transfers critical periodontal pathogens to periodontal infectious sites and recruits and activates local immune cells, which results in tooth-supporting tissue destruction \\[[@B10], [@B11]\\]. *F. nucleatum* has been identified as a high-frequency pathogen in periodontal disease \\[[@B12]\\] and many other infectious diseases, such as ventriculitis and brain abscesses \\[[@B13]\\], liver abscesses \\[[@B14], [@B15]\\], lung abscesses \\[[@B16]\\], septicemia-related infections \\[[@B17]\\], pelvic inflammatory disease \\[[@B18]\\], and intrauterine infections \\[[@B19]--[@B21]\\]. *F. nucleatum* attacks host tissues and obstructs the healing of damaged oral tissues by secreting large amounts of ammonia and butyrate \\[[@B22], [@B23]\\] and accelerates the initiation and progression of colorectal cancer by promoting tumor cell proliferation \\[[@B24]\\]. However, the effects of *F. nucleatum* on gingival fibroblast (GF) proliferation and apoptosis have not been reported.\n\nThe host response to pathogenic invasion is the determining factor of human health. As the first line of physical and chemical defense against infection, gingival epithelial cells release antimicrobial peptides such as human *\u03b2*-defensins to repress plaque biofilm formation, and activate innate and adaptive immune responses to protect the host from injury \\[[@B25], [@B26]\\]. Pathogens penetrate the connective tissue, directly damage the periodontal supporting tissues, and then induce the destruction of the integrity of the gingival epithelium \\[[@B5], [@B27]\\]. Human GFs, the predominant cell type of the periodontal connective tissue, are activated and recruited to sites of infection to regulate the occurrence and development of diseases \\[[@B28]--[@B30]\\]. It has been reported that GFs directly act on oral pathogens to defend against the progression of the infection by secreting various types of cytokines and chemokines, such as interleukin- (IL-) 1*\u03b2*, IL-6, IL-8, and tumor necrosis factor- (TNF-) *\u03b1* \\[[@B31], [@B32]\\]. Cytokine production is a crucial component for host defense against pathogenic invasion \\[[@B31], [@B32]\\]. A previous study confirmed that GFs showed no tolerance to bacterial stimulation and could continuously respond to exogenous stimuli and produce high levels of inflammatory cytokines \\[[@B33]\\]. However, the understanding of the pathogenic effects of *F. nucleatum* on inflammatory cytokine production by GFs and the potential mechanism has not been completely elucidated.\n\nIt has been reported that bacterial infection induces the mitochondrial electron transport for aerobic respiration and then dramatically elevates the intracellular reactive oxygen species (ROS) levels, which play critical roles in regulating cellular proliferation, apoptosis, and inflammatory response \\[[@B34]--[@B36]\\]. *F. nucleatum* has been shown to induce bladder cancer cell apoptosis through mediating ROS generation and mitochondrial dysfunction \\[[@B34]\\]. *F. nucleatum* enhances proinflammatory cytokine production by causing the impairment of autophagic flux in Caco-2 cells \\[[@B35]\\]. However, the effects of *F. nucleatum* on GF ROS generation remain unclear.\n\nIn this study, we aimed to explore the comprehensive gene expression profile of GFs over an *F. nucleatum* stimulation time course. We validated the effects of *F. nucleatum* on GF proliferation, apoptosis, intracellular ROS generation, and inflammatory cytokine production by biological experiments. Furthermore, we investigated the potential mechanism by which *F. nucleatum* regulated the biological properties of GFs according to the significantly enriched signaling pathways.\n\n2. Materials and Methods {#sec2}\n========================\n\n2.1. Human Subjects and Ethical Statements {#sec2.1}\n------------------------------------------\n\nThis study was approved by the Medical Ethical Committee of the Stomatology School, Shandong University (Protocol Number: 20170101). Five healthy individuals aged 30-35 who underwent impacted tooth extraction at the Department of Oral and Maxillofacial Surgery, Stomatology Hospital of Shandong Province, were recruited. All individuals were informed about the research project and signed the informed consent form according to the Helsinki Declaration of 1975.\n\n2.2. Cell Isolation and Culture {#sec2.2}\n-------------------------------\n\nThe excised gingival tissues were immediately immersed in Dulbecco\\'s modified Eagle\\'s medium (DMEM, HyClone, Logan, UT, USA) with 5% antibiotics (100\u2009U/ml penicillin and 100\u2009mg/ml streptomycin, Sigma-Aldrich, St Louis, MO, USA) and quickly transferred to the laboratory. Then, the free gingival tissue samples were washed, and the epithelium was removed. The samples were minced into small fragments of approximately 1-3\u2009mm^2^ and digested for 2\u2009h at 37\u00b0C by 3\u2009mg/ml collagenase I (Sigma Aldrich) and 4\u2009mg/ml Dispase II (Invitrogen, Carlsbad, CA, USA). The primary GFs were cultured with DMEM containing 20% fetal bovine serum (FBS, Biological Industries, Kibbutz, Israel) at 37\u00b0C in a humidified atmosphere of 5% CO~2~. Cells were fed fresh medium every three days until the cell monolayer reached 80-90% confluence. GFs were trypsinized and passaged at a dilution ratio of 1\u2009:\u20093 to expand the culture in 10% FBS medium. The fourth passage cells were used for the following experiments, and cells were cultured in 10% FBS DMEM in all experiments.\n\n2.3. RNA-Sequencing (RNA-Seq) Analysis {#sec2.3}\n--------------------------------------\n\nA total of 54 samples from 5 individuals were sequenced to analyze the gene expression at the whole genome level at BGI (Beijing Genomics Institute, Shenzhen, China) by RNA-seq (each individual sample included 1 sample at time point 0\u2009h; 5 control samples at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h with no *F. nucleatum* treatment; and 5 experimental samples with *F. nucleatum* stimulation at multiplicity of infection (MOI) of 100 at the same time points; sufficient RNA could not be extracted from one treated sample at the 12\u2009h time point). Total RNA was isolated, evaluated for quality, reverse-transcribed to cDNA, and sequenced on the BGISEQ-500 platform. After a quality control (QC) step (Agilent 2100 Bioanalyzer, Santa Clara, California, USA), the clean reads were mapped to the reference genome (GRCh38) via hierarchical indexing for spliced alignment of transcripts (HISAT) (v2.0.4) \\[[@B37]\\]. Bowtie 2 (v2.2.5) \\[[@B38]\\] was used to map the clean reads to the reference transcripts, and the gene expression level for each sample was calculated by RSEM \\[[@B39]\\]. The correlation among all samples was detected by principal component analysis (PCA). Based on the gene expression level, DESeq2 and PoissonDis algorithms were used to detect the differentially expressed genes (DEGs) between the control groups and the *F. nucleatum*-treated groups by an absolute\u2009of\u2009log\u2009base\u20092\u2009of\u2009fold\u2009change\u2009in\u2009DEGs \u2265 1 and an adjusted *P* value (%) \\< 5%. Gene ontology (GO) was used for screening and annotating DEGs. WEGO software \\[[@B40]\\] was used to generate the GO functional classification file. Pathway enrichment analysis of DEGs was performed based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database \\[[@B41]\\]. A Venn diagram was used to show the overlapping genes among the DEGs of each group. The protein-protein interaction network was generated by the STRING \\[[@B42]\\] website with default parameters. The Pathview \\[[@B43]\\] website was used to analyze the signaling pathway activation after *F. nucleatum* stimulation. In addition, Venn diagrams and heat maps were used to present the expression signatures of the DEGs involved in the GO biological processes, including cell proliferation (GO: 0008283), apoptotic process (GO: 0006915), response to reactive oxygen species (GO: 0000302), and defense response (GO: 0006952). The RNA sequence data have been deposited in the NCBI Gene Expression Omnibus (GEO, <http://www.ncbi.nlm.nih.gov/geo/>) and are accessible through GEO series accession number GSE118691.\n\n2.4. Cell Proliferation Assay {#sec2.4}\n-----------------------------\n\nGFs were seeded in 6-well (1 \u00d7 10^5^\u2009cells/well) plates in growth medium with 10% FBS. Cells were left untreated or incubated with *F. nucleatum* at multiplicities of infection (MOIs) (*F. nucleatum*\u2009:\u2009cell of 10\u2009:\u20091, 50\u2009:\u20091, 100\u2009:\u20091, 200\u2009:\u20091, and 400\u2009:\u20091). Cells were counted every other day using a hemocytometer (Corning, Corning, NY, USA) or an automated cell counter (Countstar, Shanghai, China). To assess the cell proliferation rate, cells were inoculated in 24-well plates (5 \u00d7 10^4^\u2009cells/well) and treated with the indicated concentration of *F. nucleatum*. The 5-ethynyl-2\u2032-deoxyuridine labeling assay was used to evaluate the cell proliferation rate according to the instructions of an EdU Apollo DNA in vitro kit (RiboBio, Guangzhou, China). The experiment was performed in sextuplicate and repeated three times.\n\n2.5. Cell Apoptosis and Viability Analysis {#sec2.5}\n------------------------------------------\n\nCell apoptosis was analyzed according to the instructions of an Annexin V-FITC/PI kit (Dojindo, Kumamoto, Japan). GFs were seeded in 6-well plates (2 \u00d7 10^5^cells/well) and stimulated with different densities of *F. nucleatum* for 2, 6, 12, 24, 36, and 48\u2009h. Afterward, cells were trypsinized, washed with phosphate-buffered saline (PBS), stained with a PI-conjugated anti-annexin V antibody for 15\u2009min at room temperature in the dark, and then subjected to flow cytometry; the data were analyzed by Accuri C6 Plus software (Becton Dickinson, Franklin Lakes, NJ, USA). Cell viability was detected by trypan blue (Solarbio) staining. GFs treated with 4% formaldehyde were used as a positive control, and untreated cells were used as a negative control. Trypan blue was added to the monolayers at a 1\u2009:\u200910 dilution and incubated for 10\u2009min. The cells were observed immediately under the microscope and photographed (Olympus, Tokyo, Japan). The experiment was performed in triplicate and repeated three times.\n\n2.6. Measurement of Intracellular ROS {#sec2.6}\n-------------------------------------\n\nIntracellular ROS levels were detected by a 2\u2032,7\u2032-dichlorofluorescein diacetate (DCFH-DA) assay. GFs were seeded in 6-well plates and treated with *F. nucleatum* at the above-mentioned MOIs for 2, 6, 12, 24, 36, and 48\u2009h. Cells were incubated with DCFH-DA (1\u2009:\u20091000) for 20\u2009min at 37\u00b0C in a cell incubator in the dark. Flow cytometry (Becton Dickinson) was used to determine the intracellular ROS production. The flow cytometry data were analyzed by Accuri C6 Plus software. Each experiment was performed in triplicate and repeated three times.\n\n2.7. RNA Isolation and Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) {#sec2.7}\n---------------------------------------------------------------------------------\n\nGFs were seeded in 6-well plates and cultured with *F. nucleatum*. GFs were collected at 6 different time points (0\u2009h, 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h), and total RNA was extracted with TRIzol\u00ae (CWBIO, Beijing, China). The mRNA concentration was determined using an ultramicro spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). One microgram of mRNA was reverse-transcribed to cDNA using a HiFiScript cDNA Synthesis kit (CWBIO). qRT-PCR was performed with an UltraSYBR Mixture (CWBIO) on a LightCycler 96 Real-Time PCR System (Roche, Basel, Switzerland) in triplicate. Briefly, the hot start enzyme was activated (95\u00b0C for 10\u2009min), and the cDNA was then amplified for 45 cycles of denaturation at 95\u00b0C for 10\u2009s, annealing at 60\u00b0C for 30\u2009s and extension at 72\u00b0C for 32\u2009s. Data were analyzed using the 2^(-*\u0394\u0394*Ct)^ method. The sequences of the primers for amplification are shown in Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}. The experiment was performed in triplicate and repeated three times.\n\n2.8. Enzyme-Linked Immunosorbent Assay (ELISA) {#sec2.8}\n----------------------------------------------\n\nGFs were seeded in 6-well plates and treated with *F. nucleatum* at an MOI of 100\u2009:\u20091. Cell culture supernatants were collected and centrifuged at 12,000\u2009rpm for 5\u2009min at 4\u00b0C. The levels of secreted IL-6, IL-8, IL-1*\u03b2*, and TNF-*\u03b1* protein were measured by ELISA (BioLegend, San Diego, CA, USA) according to the manufacturer\\'s instructions. The optical density values were measured by a microplate reader at 450\u2009nm and 570\u2009nm, and the 570\u2009nm values were subtracted from the absorbance at 450\u2009nm in the subsequent data analysis. The experiment was performed in triplicate and repeated three times.\n\n2.9. Western Blot Assay {#sec2.9}\n-----------------------\n\nGFs with or without *F. nucleatum* stimulation were harvested with a RIPA lysis buffer containing 1% protease inhibitors and 1% phosphatase inhibitors (Solarbio, Beijing, China). Protein concentrations were measured according to bicinchoninic acid (BCA) assays, and proteins (20\u00a0*\u03bc*g/lane) were separated to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). Membranes were blocked with 5% nonfat milk for 1\u2009h, incubated with primary antibodies overnight at 4\u00b0C, and then incubated with horseradish peroxidase-conjugated secondary antibodies (1\u2009:\u200910 000; Proteintech, Chicago, IN, USA) for 1\u2009h at room temperature. The protein bands were visualized with enhanced chemiluminescence reagents (Millipore) and scanned using an extra-sensitive imager (Amersham Imager 600; GE Healthcare Life Sciences, Pittsburgh, PA, USA). ImageJ 1.44 software (NIH, Bethesda, Maryland, USA) was used to quantify the protein expression levels. The primary antibodies and dilution ratio were as follows: rabbit anti-NF-*\u03ba*B p65 (1\u2009:\u200910000; Abcam, Cambridge, UK), rabbit anti-phospho-NF-*\u03ba*B p65 (1\u2009:\u20091000; Abcam), rabbit anti-I*\u03ba*B*\u03b1* (1\u2009:\u20095000; Abcam), rabbit anti-phospho-I*\u03ba*B*\u03b1* (1\u2009:\u20091000;Abcam), rabbit anti-p38 (1\u2009:\u20091000; Cell Signaling Technology, Danvers, MA, USA), rabbit anti-phospho-p38 (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-JNK (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-phospho-JNK (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-ERK1/2 (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-phospho-ERK1/2 (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-AKT (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-phospho-AKT (1\u2009:\u20091000; Cell Signaling Technology), rabbit anti-p53 (1\u2009:\u20091000; Abcam), and rabbit anti-phospho-p53 (1\u2009:\u20091000;Abcam). The experiment was performed in triplicate and repeated three times.\n\n2.10. Cell Immunocytochemistry Assay {#sec2.10}\n------------------------------------\n\nTo detect *F. nucleatum*-induced nuclear translocation of NF-*\u03ba*B p65 and p-p65, GFs were seeded in coverslip containing 24-well plates at a density of 2 \u00d7 l0^4^\u2009cells/well and treated with or without *F. nucleatum* for 5\u2009min to 120\u2009min. Cells were fixed with 4% paraformaldehyde and blocked with 10% normal goat serum in PBS for 1\u2009h. Then, cells were incubated with primary antibodies against p65 and p-p65 overnight at 4\u00b0C. The next day, the cells were incubated with an Alexa Fluor 594-conjugated goat anti-rabbit IgG secondary antibody (1\u2009:\u2009500) in the dark for 1\u2009h, and the nuclei were visualized using 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride (DAPI) for 5\u2009min. The images were captured by fluorescence microscopy (Olympus BX53, Tokyo, Japan). The experiment was performed in triplicate and repeated three times.\n\n2.11. Statistical Analysis {#sec2.11}\n--------------------------\n\nAll data were expressed as the mean \u00b1 standard\u2009deviation (SD). Tests were analyzed using GraphPad Prism software (version 6, Software MacKiev, Boston, MA, USA), and differences among more than two groups were analyzed by one-way or two-way ANOVA followed by Tukey\\'s honestly significant difference (HSD) comparison test. Variance between two groups was compared by a multiple *t*-test. *P* \\< 0.05 was considered indicative of statistical significance.\n\n3. Results {#sec3}\n==========\n\n3.1. RNA-Seq Analysis of *F. nucleatum*-Stimulated GFs over Time {#sec3.1}\n----------------------------------------------------------------\n\nTo better understand the global responses of GFs to *F. nucleatum* infection over time, we performed a genome-wide transcriptome analysis by RNA-seq to determine the global changes in gene expression. A total of 100\u2009Gb of sequence data was generated from 54 samples, and 23.90 million reads for each sample were generated on average (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). After QC, the clean reads were mapped to the reference genomes and transcripts with mapping percentages of 94.40% and 87.67%, respectively (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). A total of 18691 genes were detected, and the expression levels were calculated with RSEM (data can be obtained through GSE118691). The correlation analysis of the PCA on the whole-genome gene expression levels showed that the normal cells were relatively stable, while the *F. nucleatum-*stimulated GFs gradually deviated from normal conditions as the stimulus time increased (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). These results suggest that the effect of *F. nucleatum* stimulation on GFs may accumulate over time.\n\nTo characterize the DEGs influenced by *F. nucleatum*, the gene expression profiles of normal GFs versus infected cells at 2, 6, 12, 24, and 48\u2009h were compared at each time point. Approximately 228, 374, 616, 1208, and 1334 genes were found to be upregulated, and 28, 147, 448, 1302, and 2145 genes were downregulated at 2, 6, 12, 24, and 48\u2009h, respectively ([Figure 1(a)](#fig1){ref-type=\"fig\"}). The overlap analysis of the DEGs among the 5 time points identified that 62 genes were differentially expressed between nonstimulated and *F. nucleatum*-stimulated GFs across the entire time course, and the number of DEGs continuously increased as the stimulation time increased ([Figure 1(b)](#fig1){ref-type=\"fig\"}). These results suggest that the response of GFs to *F. nucleatum* infection increases over time, and a longer duration *F. nucleatum* stimulation might induce more side-effects than a short duration infection.\n\nThe heat map showed that the gene expression levels of 62 DEGs were upregulated and that some of these genes, such as IL-6, IL-1*\u03b2*, CCL2, CXCL2, CXCL8, and CSF3, are closely associated with host defense responses to bacterial infection ([Figure 1(c)](#fig1){ref-type=\"fig\"}). The KEGG annotation showed that the top 5 enriched pathways of these DEGs are immune-associated, namely, the TNF-signaling pathway, IL-17 signaling pathway, rheumatoid arthritis, NF-*\u03ba*B signaling pathway, and NOD-like receptor signaling pathway ([Figure 1(d)](#fig1){ref-type=\"fig\"}). The protein-protein interaction network of 62 DEGs showed the potentially complex interaction relationship among these genes ([Figure 1(e)](#fig1){ref-type=\"fig\"}).\n\nThe GO analysis confirmed that the 62 DEGs are involved in host defense to bacterial infection and cell growth pathways, such as cellular process, response to stimulus, regulation of biological process, immune system process, response to bacterium, and cell proliferation (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}), and the top five enriched biological process GO terms for the 62 DEGs are shown in Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}.\n\n3.2. *F. nucleatum* Inhibits the Proliferation of GFs {#sec3.2}\n-----------------------------------------------------\n\nTo study the effect of *F. nucleatum* on the proliferation of GFs, we applied a cell counting assay and EdU-labeling assay to quantify the cell proliferation rate over 5 days of *F. nucleatum* stimulation at MOIs of 10 and 100. The cell counting assay showed that *F. nucleatum* significantly inhibited the proliferation of GFs in a time- and dose-dependent manner (Figures [2(a)](#fig2){ref-type=\"fig\"} and [2(b)](#fig2){ref-type=\"fig\"}). To further test the inhibitory effect of different *F. nucleatum* concentrations, GFs stimulated with *F. nucleatum* at MOIs of 10, 50, 100, 200, and 400 were quantified at 3 time points over 5 days. The results showed that the inhibitory effect aggrandized with the increase of *F. nucleatum* concentration, and the high concentration of *F. nucleatum* (MOI of 400) directly blocked the cell growth ([Figure 2(c)](#fig2){ref-type=\"fig\"}). The EdU-labeling assay demonstrated that *F. nucleatum* significantly decreased the number of GFs and reduced the cell proliferation rate after 24\u2009h of stimulation ([Figure 2(d)](#fig2){ref-type=\"fig\"}--[2(f)](#fig2){ref-type=\"fig\"}). The DEGs of the 54 samples enriched in the cell proliferation pathway (GO: 0008283) were compared using an overlap analysis, and the results indicated that the number of cell proliferation-related DEGs continuously increased as the *F. nucleatum* stimulation time increased; 40, 62, 93, 152, and 183 genes were detected at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h, respectively, and 18 DEGs, including CCL2, SOD2, and NFKBIA, were consistently upregulated across all the time points ([Figure 2(g)](#fig2){ref-type=\"fig\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). The above results suggest that *F. nucleatum* dose- and time-dependently regulates cell proliferation-related genes and inhibits the proliferation of GFs.\n\n3.3. *F. nucleatum* Promotes the Apoptosis of GFs {#sec3.3}\n-------------------------------------------------\n\nTo further study the reason for the proliferation inhibition of GFs, the apoptosis rates of *F. nucleatum*-stimulated GFs were detected by flow cytometry. The results indicated that the low concentration of *F. nucleatum* (MOIs of 10) showed no influence on GF apoptosis, while high concentration of *F. nucleatum* (MOIs of 50, 100, 200, and 400) significantly decreased the number of normal cells and increased the number of apoptosis cells at 2\u2009h stimulation and maintained the similar apoptosis trend to 36\u2009h. At MOIs of 50 and 100, the number of early apoptotic cells was greater than the number of late apoptotic cells at the 2, 6, 12, 24, and 36\u2009h *F. nucleatum* stimulation, while at the MOIs of 200 and 400, the number of late apoptotic GFs was larger than the early apoptotic cells ([Figure 3(a)](#fig3){ref-type=\"fig\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). The DEGs of the 54 samples enriched in the apoptotic process (GO: 0006915) were compared using an overlap analysis, and the results suggested that 24, 31, 50, 84, and 97 apoptosis-related genes were differentially expressed at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h, respectively ([Figure 3(b)](#fig3){ref-type=\"fig\"}). Three apoptosis-related genes, CSF2, PTGS2, and SOD2, were continuously upregulated across the 5 time points ([Figure 3(b)](#fig3){ref-type=\"fig\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). A trypan blue staining assay indicated that *F. nucleatum* dose-dependently increased the number of nonviable cells, which were stained blue, and the cell morphology changed from long spindle to round, and fewer live cells were detected after *F. nucleatum* stimulation at MOIs of 200 and 400 ([Figure 3(c)](#fig3){ref-type=\"fig\"}). These results suggest that *F. nucleatum* might inhibit the proliferation of GFs by dose-dependently promoting GF apoptosis and time-dependently upregulating a group of apoptosis process-related genes.\n\n3.4. *F. nucleatum* Promotes Intracellular ROS Generation in GFs {#sec3.4}\n----------------------------------------------------------------\n\nTo investigate whether the alteration in GF biological properties is associated with intracellular ROS production, GFs were treated with *F. nucleatum* (MOIs of 10, 50, 100, 200, and 400) for 2, 6, 12, 24, 36, and 48\u2009h, and the level of ROS was detected at the different stimulation times and doses. The flow cytometry analysis results showed that the level of intracellular ROS was significantly elevated in a dose- and time-dependent manner (Figures [4(a)](#fig4){ref-type=\"fig\"} and [4(c)](#fig4){ref-type=\"fig\"}). The DEGs of the 54 samples enriched in the response to reactive oxygen species (GO: 0000302) were compared using an overlap analysis, and the results indicated that the number of differentially expressed ROS production-related genes also increased in a time-dependent manner, and 2, 4, 6, 16, and 22 DEGs were detected after *F. nucleatum* stimulation (MOI of 100) at 2, 6, 12, 24, 36, and 48\u2009h ([Figure 4(b)](#fig4){ref-type=\"fig\"}). SOD2 is a unique DEG that was continuously upregulated across the 5 time points. The relative gene expression level of SOD2 is shown in [Figure 4(d)](#fig4){ref-type=\"fig\"}. In addition, SOD2 also acts as the main contributor to the regulation of the proliferation and apoptosis of GFs (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}).\n\n3.5. *F. nucleatum* Promotes GF Inflammatory Cytokine Production {#sec3.5}\n----------------------------------------------------------------\n\nThe expression of Toll-like receptor- (TLR-) 2 and TLR4, which are closely related to inflammatory cytokine production, was analyzed by qRT-PCR. The expression of the TLR4 gene was elevated after *F. nucleatum* stimulation at 2\u2009h, 12\u2009h, 24\u2009h, and 36\u2009h, while showing no significant differences compared with the TLR4 expression in normal cells at 48\u2009h ([Figure 5(a)](#fig5){ref-type=\"fig\"}). The TLR2 gene expression level was increased in *F. nucleatum*-stimulated GFs at 2\u2009h, 6\u2009h, and 12\u2009h, while showing no significant differences at 24\u2009h and 48\u2009h compared with that in nonstimulated GFs ([Figure 5(d)](#fig5){ref-type=\"fig\"}). Compared with nonstimulated GFs at each time point, *F. nucleatum*-stimulated GFs showed significantly elevated gene expression levels of IL-6, IL-8, IL-1*\u03b2*, and TNF-*\u03b1* across the whole time course (Figures [5(b)](#fig5){ref-type=\"fig\"}, [5(e)](#fig5){ref-type=\"fig\"}, [5(h)](#fig5){ref-type=\"fig\"}, and [5(j)](#fig5){ref-type=\"fig\"}); however, at the protein level, IL-6 and IL-8 were significantly increased after *F. nucleatum* stimulation, while IL-1*\u03b2* and TNF-*\u03b1* were not (Figures [5(c)](#fig5){ref-type=\"fig\"}, [5(f)](#fig5){ref-type=\"fig\"}, [5(i)](#fig5){ref-type=\"fig\"}, and [5(k)](#fig5){ref-type=\"fig\"}). The DEGs of the 54 samples enriched in the defense response (GO: 0006952) were compared using an overlap analysis, which indicated that 31, 43, 47, 73, and 91 DEGs related to defense response were time-dependently increased after *F. nucleatum* stimulation; 14 overlapping DEGs were significantly upregulated among the 5 time points in the *F. nucleatum* stimulation groups ([Figure 5(g)](#fig5){ref-type=\"fig\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}).\n\n3.6. *F. nucleatum* Activates GF Nuclear Factor-*\u03ba*B (NF-*\u03ba*B), Mitogen-Activated Protein Kinase (MAPK), and Protein Kinase B (AKT) Signaling Pathway {#sec3.6}\n-----------------------------------------------------------------------------------------------------------------------------------------------------\n\nTo further explore the mechanisms of *F. nucleatum*-induced biological process alteration in GFs, the cell proliferation, apoptosis, and immune response-related NF-*\u03ba*B, MAPK, and AKT signaling pathways were analyzed by western blotting and immunofluorescence. The results showed that *F. nucleatum* increased the proportion of p-p65 to p65 and p-I*\u03ba*B*\u03b1* to I*\u03ba*B*\u03b1* after 5\u2009min of stimulation and that this increase was maintained for 120\u2009min ([Figure 6(a)](#fig6){ref-type=\"fig\"}--[6(c)](#fig6){ref-type=\"fig\"}). The immunofluorescence analysis of cells showed that p65 and p-p65 translocation from the cytoplasm to the nucleus also increased continuously after *F. nucleatum* stimulation from 5\u2009min to 120\u2009min (Figures [6(h)](#fig6){ref-type=\"fig\"} and [6(i)](#fig6){ref-type=\"fig\"}). These results indicate that *F. nucleatum* activates the NF-*\u03ba*B signaling pathway by increasing the phosphorylation levels of p65 and I*\u03ba*B*\u03b1* and increasing the nuclear-to-cytoplasmic ratio of NF-*\u03ba*B p65 and NF-*\u03ba*B p-p65 levels. Furthermore, we confirmed that *F. nucleatum* significantly activated AKT/MAPK signaling pathways by increasing the phosphorylation levels of AKT, ERK, JNK, and p38 (Figures [6(a)](#fig6){ref-type=\"fig\"} and [6(d)](#fig6){ref-type=\"fig\"}--[6(g)](#fig6){ref-type=\"fig\"}). The protein expression ratio of p-AKT/AKT and p-ERK/ERK was elevated after *F. nucleatum* stimulation at 5\u2009min and maintained at a high level until 120\u2009min (Figures [6(f)](#fig6){ref-type=\"fig\"} and [6(g)](#fig6){ref-type=\"fig\"}). In addition, the relative value of p-JNK/JNK was increased after 30\u2009min of *F. nucleatum* stimulation ([Figure 6(d)](#fig6){ref-type=\"fig\"}), and the phosphorylation level of p38 was significantly increased after 5\u2009min of stimulation with *F. nucleatum* ([Figure 6(e)](#fig6){ref-type=\"fig\"}). Furthermore, the protein level of p53 and phosphorylated p53 was detected, and the results indicated that *F. nucleatum* could not increase the protein expression ratio of p-p53/p53 after stimulation for 5\u2009min to 120\u2009min (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}).\n\n3.7. Whole-Transcriptome Analysis of NF-*\u03ba*B, MAPK, and PI3K-AKT Signaling Pathways in *F. nucleatum*-Stimulated GFs {#sec3.7}\n--------------------------------------------------------------------------------------------------------------------\n\nTo further clarify the correlation among cell proliferation, apoptosis, oxidative stress, cytokine production and AKT/MAPK, and NF-*\u03ba*B signaling pathways, the DEGs induced by *F. nucleatum* stimulation from the 54 samples at each time point were integrated and visualized by Pathview. The results indicated that the DEGs were significantly enriched in the NF-*\u03ba*B, PI3K-AKT, and MAPK signaling pathways, which suggest that these pathways are significantly activated by *F. nucleatum* stimulation at the gene expression level and are consistent with the results of our biological experiments ([Figure 7](#fig7){ref-type=\"fig\"}, Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). For the NF-*\u03ba*B signaling pathway, we confirmed that most DEGs downstream of the NF-*\u03ba*B signaling pathway were significantly upregulated by *F. nucleatum* stimulation, which is critical in regulating cell proliferation, apoptosis, ROS generation, and cytokine production ([Figure 7](#fig7){ref-type=\"fig\"}). For example, cIAP1/2 and TRAF1/2 were elevated after *F. nucleatum* stimulation at 2\u2009h, and Bcl-2 was upregulated at 12\u2009h after *F. nucleatum* stimulation; these genes play key roles in cell proliferation and apoptosis. The gene expression levels of IL-8, cyclooxygenase 2 (COX2), A20, I*\u03ba*B*\u03b1*, and MIP2 were significantly upregulated throughout the whole *F. nucleatum* stimulation (from 2 to 48\u2009h); these genes are closely associated with GF intracellular ROS generation and inflammatory cytokine production ([Figure 7](#fig7){ref-type=\"fig\"} and Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). In addition, we found that a large number of *F. nucleatum*-stimulated DEGs were enriched in the PI3K-AKT signaling pathway (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}) and MAPK signaling pathways (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}), which are mainly involved in cell proliferation and apoptosis. Furthermore, we found that the PI3K-AKT and MAPK signaling pathways were upstream of the NF-*\u03ba*B signaling pathway, which plays a synergetic role in regulating cell proliferation, apoptosis, ROS generation, and the inflammatory response.\n\n4. Discussion {#sec4}\n=============\n\nCharacterized by gingival inflammation and periodontal supporting tissue destruction, periodontitis is prevalent among adults over 30 years old \\[[@B44], [@B45]\\]. Understanding the pathogenic mechanisms of oral pathogens in periodontitis and identifying risk factors are important for maintaining human periodontal health. Previous studies have provided in-depth insights into the pathogenic effects of *P. gingivalis* \\[[@B46]\\]. *F. nucleatum* often appears simultaneously with *P. gingivalis* at periodontal infection sites \\[[@B47]\\], while its pathogenic role in periodontal disease has not been comprehensively reported. This study is the first to illuminate the comprehensive effects of *F. nucleatum* on cell proliferation, apoptosis, ROS generation, and inflammatory cytokine production in GFs.\n\nOur study confirms that *F. nucleatum* inhibits the proliferation and promotes the apoptosis of GFs, which is in contrast to the role *F. nucleatum* plays in colorectal cancer cell (CRC) proliferation. *F. nucleatum* has been reported to promote E-cadherin-expressing CRC proliferation by modulating the E-cadherin/*\u03b2*-catenin signaling pathway or activating TLR4 signaling to NF-*\u03ba*B and upregulating the expression of microRNA-21 \\[[@B24], [@B48]\\]. Our transcriptomic analysis demonstrates that *F. nucleatum* alters the expression of proapoptotic genes in GFs by upregulating TRAILR2, NOXA, and PUMA expression and downregulating the substrates *\u03b1*-tubulin, actin, fodrin, and PARP (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}), which are crucial for the regulation of cell proliferation and apoptosis \\[[@B49]--[@B55]\\]. We corroborated the gene expression levels of TRAILR2, NOXA, PUMA, *\u03b1*-tubulin, actin, fodrin, and PARP by qRT-PCR, and the results indicated that *F. nucleatum* significantly upregulated TRAILR2 and NOXA at 2 to 48\u2009h, while increased PUMA gene expression at 2, 24, and 48\u2009h; *F. nucleatum* downregulated *\u03b1*-tubulin, actin, fodrin, and PARP gene expression at the long time stimulation at 12\u2009h, 24\u2009h, and 48\u2009h, which were consistent with those of the RNA-seq analysis. However, at the 2\u2009h and 6\u2009h *F. nucleatum* stimulation, some disagreements between qRT-PCR and RNA-seq analysis were discovered in evaluating the gene level of actin and fodrin (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). Moreover, *F. nucleatum* elevated the expression of Mcl-1, which ultimately induced apoptosis through bim-mediated mitochondrial apoptotic events (Supplementary [](#supplementary-material-1){ref-type=\"supplementary-material\"}). *F. nucleatum* differentially regulates cell proliferation in GFs and CRCs, which indicates that the alteration of biological properties by *F. nucleatum* in normal human cells is quite different from that in tumor cell lines.\n\nIt has been reported that *F. nucleatum* stimulates ROS production and induces the tissue inflammatory response in TLR4-competent macrophages \\[[@B56]\\]. *F. nucleatum* induces IL-8, IL-1*\u03b2*, TNF-*\u03b1*, and ROS generation in Caco-2 colorectal adenocarcinoma cells by impairing autophagic flux \\[[@B35]\\]. In addition, recent studies have reported that in GFs, two NADPH oxidase isoforms, NOX1 and NOX2, are activated in response to *F. nucleatum* infection \\[[@B57], [@B58]\\]. However, the current study lacks the mechanism of ROS and inflammatory cytokine production in GFs after *F. nucleatum* stimulation. In this study, we confirm that *F. nucleatum* increases the production of ROS, activates TLR2 and TLR4 gene expression at early stimulation time points, and promotes the release of proinflammatory cytokines, such as IL-6 and IL-8 in GFs, which are concordant with the previous study. In addition, we found that *F. nucleatum* elevated the gene expression levels of IL-1*\u03b2* and TNF-*\u03b1*, but not the protein level. In gingival epithelial cells (GECs), *F. nucleatum* has been shown to activate the NLRP inflammasome and promote IL-1*\u03b2* secretion, which is different from our results in GFs \\[[@B59], [@B60]\\]. *F. nucleatum* differentially regulates IL-1*\u03b2* production in GFs and GECs, which indicates that the pathogenic mechanism of *F. nucleatum* in oral cells is quite complex, and it is necessary to clarify the mechanism of *F. nucleatum* in GFs.\n\nOur genome-wide transcriptome analysis showed that *F. nucleatum* activates the NF-*\u03ba*B, MAPK, and PI3K-AKT signaling pathways, which is consistent with our biological validation results through western blotting assays and immunofluorescence. In addition, we demonstrated that a large number of DEGs are enriched in the MAPK signaling pathway and that the JNK and p38 MAPK signaling pathways are upstream of the NF-*\u03ba*B signaling pathway, which plays a key role in regulating cell proliferation, apoptosis, and inflammatory response \\[[@B61]--[@B63]\\]. The ERK MAPK and PI3K-AKT signaling pathways are significantly activated in GFs, and the expression levels of proliferation-related genes, such as Myc and Bcl-2, are dramatically varied after *F. nucleatum* stimulation, which are crucial in deciding the cell survival \\[[@B64]--[@B67]\\].\n\nAccording to the Pathview analysis, we found that a larger number of DEGs that were induced by *F. nucleatum* stimulation downstream of NF-*\u03ba*B and AKT/MAPK signaling pathways ([Figure 7](#fig7){ref-type=\"fig\"} and Supplementary Figures [](#supplementary-material-1){ref-type=\"supplementary-material\"}, [](#supplementary-material-1){ref-type=\"supplementary-material\"}) are associated with cell proliferation, apoptosis, ROS generation, and production of inflammatory cytokines, such as bcl2, IL-8, COX2, JNK, and Mcl-1. We validated DEGs including COX2, IL-8, and bcl2 by qRT-PCR, and the results were consistent with those of the RNA-seq analysis. IL-8 is an important chemokine and plays key roles in the acute inflammatory response and various inflammatory diseases \\[[@B68], [@B69]\\]. COX2 is critical for the formation of prostaglandins and maintains ROS homeostasis under inflammatory conditions, which are pivotal in the pathogenesis of periodontitis \\[[@B70], [@B71]\\]. In the future, we will construct some gene knockout strains of *F. nucleatum* to further explore the mechanism of *F. nucleatum* in GFs.\n\n5. Conclusions {#sec5}\n==============\n\nOur study indicates that *F. nucleatum* alters the gene expression profiles of GFs in a time-dependent manner and plays multidimensional roles in regulating GF biological properties. *F. nucleatum* inhibits cell proliferation and facilitates cell apoptosis, ROS generation, and inflammatory cytokine production partly through the activation of AKT/MAPK and the NF-*\u03ba*B signaling pathways in GFs. Our study opens a new window for understanding the pathogenic effects of periodontal pathogens on the host oral system. A schematic diagram of the effects of *F. nucleatum* on GFs and the potential pathogenic mechanism is shown in [Figure 8](#fig8){ref-type=\"fig\"}.\n\nWe appreciate the volunteers who provided the samples. We thank Professor Shili Liu and other colleagues in our laboratory for all of the kind advice and support. We thank the Natural Science Foundation of Shandong Province (Nos. ZR201702190185 and 2018GSF118231), the National Natural Science Foundation of China (No. 81630072), the Construction Engineering Special Foundation of \"Taishan Scholars\" of Shandong Province (No. tsqn20161068), and the Program of Qilu Young Scholars of Shandong University for the financial support.\n\n*F. nucleatum*:\n\n:   *Fusobacterium nucleatum*\n\nGFs:\n\n:   Gingival fibroblasts\n\nROS:\n\n:   Reactive oxygen species\n\nIL:\n\n:   Interleukin\n\nTNF:\n\n:   Tumor necrosis factor\n\nDMEM:\n\n:   Dulbecco\\'s modified Eagle\\'s medium\n\nFBS:\n\n:   Fetal bovine serum\n\nRNA-seq:\n\n:   RNA sequencing\n\nQC:\n\n:   Quality control\n\nDEGs:\n\n:   Differentially expressed genes\n\nGO:\n\n:   Gene ontology\n\nKEGG:\n\n:   Kyoto Encyclopedia of Genes and Genomes\n\nPCA:\n\n:   Principal component analysis\n\nPBS:\n\n:   Phosphate-buffered saline\n\nMOI:\n\n:   Multiplicity of infection\n\nEdU:\n\n:   5-Ethynyl-2\u2032-deoxyuridine\n\nDCFH-DA:\n\n:   2\u2032,7\u2032-dichlorofluorescein diacetate\n\nqRT-PCR:\n\n:   Quantitative real-time polymerase chain reaction\n\nELISA:\n\n:   Enzyme-linked immunosorbent assay\n\nBCA:\n\n:   Bicinchoninic acid\n\nSDS-PAGE:\n\n:   Sulfate-polyacrylamide gel electrophoresis\n\nPVDF:\n\n:   Polyvinylidene fluoride\n\nDAPI:\n\n:   2-(4-Amidinophenyl)-6-indolecarbamidine dihydrochloride\n\nSD:\n\n:   Standard deviation\n\nANOVA:\n\n:   Analysis of variance\n\nTLRs:\n\n:   Toll-like receptors\n\nNF-*\u03ba*B:\n\n:   Nuclear factor-*\u03ba*B\n\nMAPK:\n\n:   Mitogen-activated protein kinase\n\nAKT:\n\n:   Protein kinase B\n\nCRCs:\n\n:   Colorectal cancer cells\n\nGECs:\n\n:   Gingival epithelial cells\n\nHSD:\n\n:   Honestly significant difference.\n\nData Availability\n=================\n\nAll raw RNA-seq data are accessible through GEO series accession number, GSE118691. Other data from Materials and Methods used to support the findings of this study are included in the article. If any other data are needed, please contact the corresponding author.\n\nConflicts of Interest\n=====================\n\nThe authors indicate no potential conflicts of interest.\n\nAuthors\\' Contributions\n=======================\n\nFQ and HKL conceived, designed, and supervised this study, and KWY and ZZW collected the samples. KWY, TD, and GH performed the experiments. KWY and JZL analyzed and interpreted the data. KWY and FQ wrote the paper. All authors commented on the manuscript.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\nTable 1: primer sequences for quantitative real-time PCR (qRT-PCR).\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable 2: the clean reads of 54 samples after RNA-seq.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable 3: genome mapping ratio summary of 54 samples.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable 4: gene mapping ratio summary of 54 samples.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTable 5: top five enriched GO terms in the biological process for the 62 DEGs.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 1: PCA of 54 samples of RNA-seq analyses.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 2: GO analysis of the 62 overlapped DEGs after *F. nucleatum* stimulation at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 3: heat map summarizing the 18 overlapped DEGs from the GO biological process analysis after the five paired comparisons of cell proliferation.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 4: (a) Flow cytometry analysis of cell apoptosis. (b) Heat map of the 3 overlapped DEGs from the GO biological process analysis after the five paired comparisons of cell apoptosis.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 5: heat map of the 14 overlapping DEGs from the GO biological process analysis after the five paired comparisons of cell defense response.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 6: (a) The protein levels of p53, phosphorylated p53 (p-p53). (b) The ratio of p-p53 and p53.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 7: the Pathview analysis of the PI3K-AKT signaling pathway.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 8: the Pathview analysis of MAPK signaling pathway.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 9: the relative gene expression level of Bcl-2 (a), COX2(b), A20 (c), I*\u03ba*B*\u03b1* (d), and MIP2 (e) by qRT-PCR.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 10: the Pathview analysis of the apoptosis signaling pathway.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nFigure 11: the relative gene expression level of TRAILR2 (a), NOXA (b), PUMA (c), tubulin (d), actin (e), fodrin (f), and PARP (g) by RT-PCR.\n\n###### \n\nClick here for additional data file.\n\n![RNA-sequencing analysis of GFs stimulated with *F. nucleatum* (MOI of 100). (a) The number of DEGs in 54 samples after *F. nucleatum* stimulation at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h. (b) Venn diagram summarizing the overlapping DEGs among the five time points. (c) Heat map of the 62 overlapping DEGs. (d) KEGG enrichment analysis of the 62 overlapping DEGs. (e) Network of 62 overlapping DEGs after *F. nucleatum* stimulation at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h.](OMCL2019-1681972.001){#fig1}\n\n![Effects of *F. nucleatum* on cell proliferation. Cell numbers of GFs detected by an electronic cell counter (a) and hemocytometer (b) with or without *F. nucleatum* stimulation at MOIs of 10 and 100 (*n* = 6). (c) Number of GFs detected by electronic cell counter after *F. nucleatum* stimulation (MOIs of 0, 10, 50, 100, 200, and 400, *n* = 6). (d) Cell proliferation rate of GFs detected by EdU assay (*n* = 6). (e) EdU assay of GFs after *F. nucleatum* stimulation (MOIs of 0, 10, 50, 100, 200, and 400) at 24\u2009h. Scale bar: 20\u00a0*\u03bc*m. NC: negative control; PC: positive control. (f) Statistical results of total and new cell growth after EdU labeling (*n* = 6). (g) Venn diagram of cell proliferation-related DEGs from the GO biological process analysis among the five time points. All data are shown as the mean \u00b1 SD. Statistical analyses were performed by one-way (d) and two-way (a, b, c, f) ANOVA with Tukey\\'s multiple-comparison test. ^\u2217^*P* \\< 0.05, ^\u2217\u2217^*P* \\< 0.01, and ^\u2217\u2217\u2217^*P* \\< 0.001 compared with the Control.](OMCL2019-1681972.002){#fig2}\n\n![Effect of *F. nucleatum* on GF apoptosis. (a) Statistical analysis of GF apoptosis with *F. nucleatum* stimulation (MOIs of 0, 10, 50, 100, 200, and 400) at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 36\u2009h. The histogram represents the mean \u00b1 SD (*n* = 3). (b) Venn diagram of apoptosis-related DEGs from the GO biological process analysis of the five paired comparisons. (c) Trypan blue staining results of GFs without or with *F. nucleatum* stimulation at MOIs of 10, 50, 100, 200, and 400 at 24\u2009h. Scale bar: 50\u2009*\u03bc*m. NC: negative control; PC: positive control. Statistical analyses were performed by two-way ANOVA with Turkey\\'s multiple-comparison test. ^\u2217\u2217\u2217^*P* \\< 0.001 compared with the control at each time point.](OMCL2019-1681972.003){#fig3}\n\n![Effects of *F. nucleatum* on ROS generation in GFs. (a) Flow cytometry analysis of ROS production with *F. nucleatum* stimulation (MOIs of 0, 10, 50, 100, 200, and 400) at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 36\u2009h. (b) Venn diagram of ROS production-related DEGs from the GO biological process analysis after the five paired comparisons. (c) Statistical results of ROS generation by flow cytometry (*n* = 3). (d) The relative gene expression level of SOD2 by qRT-PCR. The histogram represents the mean \u00b1 SD. Statistical analyses were performed by two-way ANOVA with Turkey\\'s multiple-comparison test (c) and multiple *t*-test (d). ^\u2217^*P* \\< 0.05 and ^\u2217\u2217\u2217^*P* \\< 0.001 compared with the control at each time point.](OMCL2019-1681972.004){#fig4}\n\n![Effects of *F. nucleatum* on inflammatory cytokine production in GFs. TLR4 (a), TLR2 (d), IL-6 (b), IL-8 (e), IL-1*\u03b2* (h), and TNF-*\u03b1* (g) gene expression with *F. nucleatum* stimulation (MOIs of 0 and 100) from 0 to 48\u2009h (*n* = 3). IL-6 (c), IL-8 (f), IL-1*\u03b2* (i), and TNF-*\u03b1* (k) protein expression with *F. nucleatum* stimulation (MOIs of 0 and 100) from 0 to 48\u2009h (*n* = 3). (g) Venn diagram of the defense response-related DEGs among the five time points. The histogram represents the mean \u00b1 SD. Statistical analyses were performed by multiple *t*-test. ^\u2217^*P* \\< 0.05, ^\u2217\u2217^*P* \\< 0.01, and ^\u2217\u2217\u2217^*P* \\< 0.001.](OMCL2019-1681972.005){#fig5}\n\n![Effects of *F. nucleatum* on the activation of the NF-*\u03ba*B, MAPK, and PI3K-AKT pathways in GFs. (a) The protein levels of p38, phosphorylated p38, JNK, phosphorylated JNK, ERK, phosphorylated ERK, NF-*\u03ba*B p65, phosphorylated NF-*\u03ba*B p65, I*\u03ba*B*\u03b1*, phosphorylated I*\u03ba*B*\u03b1*, AKT, and phosphorylated AKT were detected by western blotting. The relative level of phosphorylated NF-*\u03ba*B p65/NF-*\u03ba*Bp65 (b), phosphorylated I*\u03ba*B*\u03b1*/I*\u03ba*B*\u03b1* (c), phosphorylated JNK/JNK (d), phosphorylated p38/p38 (e), phosphorylated ERK/ERK (f), and phosphorylated AKT/AKT (g) was detected (*n* = 3). Immunofluorescence images of NF-*\u03ba*B p65 (h) and phosphorylated NF-*\u03ba*B p65 (i) in cells. The histograms represent means \u00b1 SD. Statistical analyses were performed using one-way ANOVA with Tukey\\'s multiple-comparison test. ^\u2217^*P* \\< 0.05, ^\u2217\u2217^*P* \\< 0.01, and ^\u2217\u2217\u2217^*P* \\< 0.001 compared with the Control.](OMCL2019-1681972.006){#fig6}\n\n![The Pathview analysis of the NF-*\u03ba*B signaling pathway. Every box is divided into five parts on average, and each part represents the relative gene expression level of *F. nucleatum*-stimulated GFs at 2\u2009h, 6\u2009h, 12\u2009h, 24\u2009h, and 48\u2009h from left to right. The red color indicates that the gene expression level is upregulated after *F. nucleatum* stimulation. The green color indicates that the gene level is downregulated after *F. nucleatum* stimulation. The white color indicates that the gene expression level is not influenced by *F. nucleatum* stimulation. The gene expression level is calculated by the log fold change in the *F. nucleatum* stimulation group relative to the control group at each time point.](OMCL2019-1681972.007){#fig7}\n\n![Putative mechanism for the effects of *F. nucleatum* on the biological processes in GFs. Schematic diagram depicting how *F. nucleatum* activates NF-*\u03ba*B, MAPK, and AKT signaling pathways and induces cell proliferation, apoptosis, and inflammatory response. TLR4: Toll-like receptor 4; TLR2: Toll-like receptor 2; I*\u03ba*B*\u03b1*: inhibitors of NF-*\u03ba*B; AKT/PKB: protein kinase B; NF-*\u03ba*B: nuclear factor kappa B; MAPK: mitogen-activated protein kinase; JNK: c-Jun N-terminal kinase; ERK: extracellular signal-regulated kinase.](OMCL2019-1681972.008){#fig8}\n\n[^1]: Academic Editor: Jos\u00e9 Lu\u00eds Garc\u00eda-Gim\u00e9nez\n"}
+{"text": "In this manual, the World Health Organization (WHO) updates its guidance for governments in preparing for a possible terrorist attack with biological or chemical weapons. The book has something for virtually everyone who may have an interest in this topic, from government officials to clinicians, including information about the history of biological and chemical warfare, applicable international treaties, procedures for requesting WHO technical consultation, fundamentals of public health emergency response, basics of infectious disease treatment, treatment of patients with specific infectious or toxic exposures, physical properties of various agents, utility of reconnaissance satellites for detecting weapons development, management of food and water safety programs, etc.\n\nAnd it is this ubiquity and ambitiousness that underlie the manual's limitations and strengths. At times the guidance is so general that is almost an inventory of truisms (e.g., \"If it is found that the \\[emergency\\] control measures are not effective, they must be changed or modified.\"); elsewhere, the manual is a detailed resource. Its utility for different users will depend on their backgrounds and information needs. The core chapter, Public Health Preparedness and Response, may disappoint those seeking more than general principles. Yet these principles merit articulation as the foundation for prevention and response.\n\nDescriptions of the sarin attack in Tokyo in 1995 and the anthrax attacks in the United States in 2001 illustrate lessons from governments' recent experiences with chemical and biological terrorism. Both episodes demonstrate that relatively small attacks can have a profound impact and expose weaknesses in public health systems. The anthrax case study lauds the success of laboratory preparations but does not sufficiently address three essential questions: Why did clinicians caring for the initial patients with cutaneous anthrax not establish and report the diagnosis sooner? Why did the Centers for Disease Control and Prevention not recognize earlier that anthrax spores could escape from sealed letters? Why did the federal government stumble initially in its efforts to communicate with the public? For each question, an assessment of systemic hurdles would benefit readers seeking to improve the functioning of the public health system.\n\nThe manual generally, but not consistently, avoids bureaucratic lingo. While clearly organized, the book lacks an index, complicating efforts to find information quickly. The appendices on chemical and biological agents offer concise, formatted summaries similar to those available through other resources, but ironically provide relatively little information about the agents' potential as weapons.\n\nThis manual will find a home on bookshelves worldwide among government officials and others concerned about the threat of biological and chemical terrorism. For those in countries most in need of this guidance, its scope may be overwhelming. But the book's underlying theme---that public health preparedness for biological or chemical terrorism depends on fundamental capacities to respond to more common health threats---is its most salient message, no matter where the user resides.\n\n*Suggested citation for this article*: Buehler JW. Public health response to biological and chemical weapons: WHO guidance \\[book review\\]. Emerg Infect Dis \\[serial on the Internet\\] 2005 Jan \\[*date cited*\\]. <http://dx.doi.org/10.3201/eid1101.040927>\n"}
+{"text": "Introduction\n============\n\nDacarbazine ([Figure 1](#F1){ref-type=\"fig\"}) is an antitumor prodrugused for the treatment of malignant metastatic melanoma and Hodgkin's disease ([@B1]- [@B3]). Ithas an imidazole ring in its structure and is structurally a 1-aryl-3,3-dimethyltriazene thatundergoes *in-vivo*metabolic *N*-demethylation ([@B4], [@B5]) to yield ultimately 5-aminoimidazole-4- carboxamide(AIC) ([Figure 1](#F1){ref-type=\"fig\"}). The metabolic precursor of 5-aminoimidazole-4-carboxamide (AIC) is 5-(3-methyltriazene-1-yl)imidazole- 4-carboxamide (MTIC)([@B3], [@B4], [@B6]-[@B8]) thatis a short-lived species. MTIC is believed to be the metabolitethroughwhich dacarbazine exerts its antineoplastic alkylation effect. MTIC undergoes tautomerization to a methylating species (compound I), thatreacts with nucleophiles such as DNA guanine bases and concomitantly generates 5-aminoimidazole-4-carboxamide (AIC) ([@B9]).\n\n![*In-vivo*metabolism of dacarbazine and alkylation of nucleophiles (Nu) by dacarbazine metabolite I](ijpr-12-255-g001){#F1}\n\nThe role of metabolism in the mode of cytotoxic action of dacarbazine has been deduced from the experiments on dimethyltriazenesthatpossess the substituted phenyl group instead of the imidazole moiety on N1 ([Figure 2](#F2){ref-type=\"fig\"}).\n\n![Title missing](ijpr-12-255-g002){#F2}\n\nMetabolism of such 1-phenyl-3,3-dimethyltriazenes generates cytotoxic monomethyltriazenes which act just like intermediate I ([Figure 1](#F1){ref-type=\"fig\"}) in the alkylation of nucleophile groups in the tissue.The immediate metabolic precursor of MTIC, 5-\\[3-(hydroxymethyl)-3-methyltriazene-1-yl\\]imidazole-4-carboxamide (HMTIC) ([@B10]) has been characterized as a urinary metabolite of dacarbazine and this compound is believed to act as a transported form of MTIC, which is the suggested antineoplastic species derived from dacarbazine.\n\nBased onthe rather selective methylation of guanine bases by dacarbazine, Lowe *et al*. ([@B11]) postulated that the carboxamide group in dacarbazine plays a recognition role in finding guanine rich moieties in DNA strands. Carboxamide group in one of its conformationsforms hydrogen bonds with cytosine similar to that of guanine. The alkylating methyltriazene group of Compound I is therefore placed in the proximity of guanine and alkylation will take place at guanine O-6 and *N*-7 moiety ([Figure 3](#F3){ref-type=\"fig\"}) ([@B3], [@B6], [@B12]-[@B15]).\n\n![Title missing](ijpr-12-255-g003){#F3}\n\nThe imidazole ring could exist in two tautomeric forms and X-ray crystal structure studies of dacarbazine shows that this compound exists in two tautomeric forms a and b as shown in [Figure 4](#F4){ref-type=\"fig\"}.\n\n![Two tautomeric forms of dacarbazine. a: NaNO~2~ , HCl, b: Suspension of Na~2~SO~4~ and 40% solution of dimethylamine in H~2~O](ijpr-12-255-g004){#F4}\n\nTautomeric form b is more stable due to the two intramolecular hydrogen bonds. The role of dynamic tautomerization of imidazole ring in dacarbazine alkylating activity has not been well documented yet ([@B16]). In this study a dacarbazine analog wassynthesized in which the imidazole ringwasreplaced by pyridine ring which could be considered as the non-tautomerizable form of imidazole ring. Evaluation and comparison of cytotoxic effect of pyridine congener and dacarbazinerevealed the role of imidazole tautomerization in dacarbazinecytotoxic activity.\n\n3-\\[3,3-Dimethyl-1-triazenyl\\]pyridine (compound III) was synthesized as shown in [Figure 5](#F5){ref-type=\"fig\"}.\n\n![Synthesis of compound III](ijpr-12-255-g005){#F5}\n\nExperimental\n============\n\n*Chemicals*\n\n5-(3,3-dimethyl-1-triazeno)-imidazole- 4-carboxamide (Dacarbazine, DTIC) was purchased from Fehlandtstrabe 3.D-20354 (Hamburg, Germany). All the other chemicals were ofsynthesis grade and purchased from Merck (Darmstadt, Germany). Collagenase (from *Clostridium histolyticum*), bovine serum albumin (BSA), Hepes, trypan blue, d\u00a0mannitol, dimethyl sulfoxide, catalase, superoxide dismutase (SOD), cyclosporine, butylatedhydroxyltoluene (BHT), chloroquinediphosphate, methylamine HCl, ethylene glycol\u00a0bis (p\u00a0aminoethyl ether) N,NN',N'\u00a0tetra acetic acid (EGTA) and heparin were obtained from Sigma (Taufkirchen, Germany). Carnitine was obtained from ICN Biomedicals (St. Thuringen, Eschwege, Germany). Acridine orange and dichlorofluorescindiacetate were purchased from Molecular Probes (Eugene, OR, USA). Rhodamine 123 was obtained from Aldrich Chemical Company (Milwaukee, WI, USA). Desferoxamine was a gift from CibaGeigy Canada Ltd. (Toronto, ON, Canada). All chemicals were of the highest commercial grade available. \u2033Sigma's caspase3 assay kit (CASP-3-C)\u2033 was purchased from Sigma-Aldrich (Taufkirchen, Germany).\n\n*Chemistry*\n\nA solution of sodium nitrite (0.02 mol, 1.38 g) in 3.5 mL of water was added dropwise to a solution of 3-aminopyridine (0.02 mol, 1.88 g) in concentrated HCl (8 mL) and water (5 mL) in an ice bath while the temperature was kept at 0 \u00b0C. The mixture was stirred for further15 minutes after completion of addition and then a solution of urea in water (0.1 g in 0.3 mL H~2~O) was added and the mixture stirred for 20 min. The mixture was then added to a mixture of sodium carbonate (7.42 g in 25 mL water) and 40% dimethylamin solution (6 mL), forming a red mixture. The mixture was then filtered and extracted by ethylactate (60 mL). The organic phase thus obtained was evaporated under vacuum after drying over anhydrous sodium sulfate. The crude product was further purified on a silicagel plate (60 GF 254) using ethylacetate:chloroform (1:1) as the eluent solvent. Yield = 61%. IR (KBr): \u03bb 2900, 1435, 1410, 1390, 1375, 1330, 1200, 1080, 800, 1H-NMR(CDCl3) \u03b4: 8.7 (1H;d,4J=2.1HZ;H-2), 8.3 (1H;d,3J=4.7HZ;H-6), 7.7 (1H;dt,3J=8.1HZ, 4J=1.7HZ;H-4), 7.25 (1H;dd,3J=8.1HZ, 3J=4.7HZ;H-5), 3.5 & 3.25 (6H;two broad singlets;CH~3~), Mass (EI), M/Z (%): 150 (82), 121 (40), 106 (86), 92 (28), 78 (100).\n\n*Animals*\n\nMale Sprague-Dawley rats (280-300 g) fed a standard chow diet and given water ad libitum, were used in all the experiments. All the experiments were conducted according to the ethical standards and protocols approved by the Committee of Animal Experimentation of ShahidBeheshti University of Medical Sciences, Tehran, IR Iran.\n\n*Isolation and incubation of hepatocytes*\n\nHepatocytes were obtained by collagenase perfusion of the liver as described by Pourahmad and O'Brien ([@B17]). Approximately 85-90% of the hepatocytes excluded trypan blue. Cells were suspended at a density of 106cells/mL in round bottomed flasks rotating in a water bath maintained at 37C in Krebs-Henseleit buffer (pH 7.4), supplemented with 12.5 mMHepes under an atmosphere of 10% O~2~, 85% N~2~, 5% CO~2~. Each flask contained 10 mL of hepatocyte suspension. Hepatocytes were preincubated for 30 min prior to the addition of chemicals. Stock solutions of all chemicals (\u00d7100 concentrated for the water solutions or \u00d71000 concentrated for the methanolic solutions) were prepared fresh prior to use. To avoid either nontoxic or very toxic conditions in this study we used EC50~2h~ concentration for DTIC and Compound III in the isolated hepatocytes (56 \u03bcM and 33 \u03bcM respectively). The EC50 of a chemical inACMS technique (ACMS: Accelerated Cytotoxicity Mechanism Screening) (with the total 3 h incubation period), is defined as the concentration which decreases the hepatocyte viability down to 50% following the 2 h of incubation ([@B18]). In order to determine this value for the investigated compound, dose-response curves were plotted and then EC50 was determined based on a regression plot of three different concentrations (data and curves not shown) ([@B19]). For the chemicals soluble in water, we added 100 \u03bcL sample of its concentrated stock solution (\u00d7100 concentrated) to one rotating flask containing 10 mL hepatocyte suspension. For the chemicals soluble in methanol we prepared methanolic stock solutions (\u00d71000 concentrated), and to achieve the required concentration in the hepatocytes, we added 10 \u03bcL samples of the stock solution to the 10 mL cell suspension. Ten microlitres of methanol did not affect the hepatocyte viability after 4 h incubation (data not shown). All the inhibitors were preincubated 30 min prior to DTIC and Compound III addition.\n\n*Cell viability*\n\nThe viability of isolated hepatocytes was assessed from the intactness of the plasma membrane as determined by the trypan blue (0.2% w/v) exclusion test ([@B20]). Aliquots of the hepatocyte incubate were taken at different time points during the 3 h incubation period. At least 80-90% of the control cells were still viable after 3 h.\n\n*Determination of reactive oxygen species (ROS)*\n\nTo determine the rate of hepatocyte ROS generation, dichlorofluorescindiacetate (DCFH-DA) was added to the incubatedhepatocytesas it penetrates hepatocytes and becomes hydrolyzed to non-fluorescent dichlorofluorescin (DCFH). The latter then reacts with ROS to form the highly fluorescent dichlorofluorescein (DCF), which effluxes the cell. Hepatocytes (1 106 cells/mL) were suspended in 10ml modified Hank's balanced salt solution (HBS), adjusted to pH 7.4 with 10 mM HEPES (2-hydroxyethyl)-1-piperazine --ethansulfonic acid-HBSH) and were incubated with DTIC and Compound III at 37\u00b0C for 30 min. After centrifugation (50 \u00d7 g. 1 min), the cells were re-suspended in HBS adjusted to pH 7.4 with 50 mM Tris-HCl and loaded with dichlorofluorescin by incubating with 1.6\u03bcL dichlorofluorescindiacetate for 2 min at 37\u00b0C. The fluorescence intensity of the ROS product was measured using a Shimadzu RF5000U fluorescence spectrophotometer. Excitation and emission wavelengths were 500 nm and 520 nm, respectively. The results were expressed as fluorescent intensity per 106 cells ([@B20], [@B21]).\n\n*Lysosomal membrane stability assay*\n\nHepatocyte lysosomal membrane stability was determined from the redistribution of the fluorescent dye, acridine orange ([@B22]). Aliquots of the cell suspension (0.5 mL) that were previously stained with acridine orange 5 \u03bcM, were separated from the incubation medium by 1min centrifugation at 1000 rpm (rotations per min). The cell pellet was then re-suspended in 2 mL of fresh incubation medium. This washing process was carried out for two times to remove the fluorescent dye from the media. Acridine orange redistribution in the cell suspension was then measured fluorimetrically using theShimadzu RF5000U fluorescence spectrophotometer set at 470 nm excitation and 540 nm emission wavelengths.\n\n*Mitochondrial membrane potential assay*\n\nMitochondrial uptake of the cationic fluorescent dye, Rhodamine 123 (1.5 \u03bcM), has been used for the estimation of mitochondrial membrane potential ([@B23]). Aliquots of the cell suspension (0.5 mL) were separated from the incubation medium by centrifugation at 1000 rpm (rotations per minute) for 1 min. The cell pellet was then re-suspended in 2 mL of fresh incubation medium containing 1.5 \u03bcM Rhodamine 123, and incubated at 37 \u00b0C in a thermostatic bath for 10 min with gentle shaking. Hepatocytes were then separated by centrifugation and the amount of Rhodamine 123 remaining in the incubation medium was measured fluorimeterically using a Shimadzu RF5000U fluorescence spectrophotometer set at 490 nm excitation and 530 nm emission wavelengths. The capacity of mitochondria to take up the Rhodamine 123 was calculated as the difference (between control and treated cells) in Rhodamine 123 fluorescence. Our data were shown as the percentage of mitochondrial membrane potential collapse (%\u0394\u03a8m) in all treated (test) hepatocyte groups ([@B23], [@B24]).\n\n*Determination of caspase-3 activity*\n\nCaspase-3 activity was determined by using the \u2033Sigma's caspase-3 assay kit (CASP-3-C)\u2033 (Sigma-Aldrich, Taufkirchen, Germany). This measurement was performed based on the hydrolysis of Ac- DEVD-pNA peptide substrate by caspase-3. The released moiety (p-nitroaniline) has a high absorbance at 405 nm ([@B25]).\n\n${AC} - {DEVD} - p{NA}\\overset{\\rightarrow}{{Caspase} - 3}{AC} - {DEVD} + p{NA}$\n\nThe concentration of the p-nitroaniline released from the substrate wascalculated from the absorbance values at 405 nm or from a calibration curve prepared with defined p-nitroaniline solutions. The activity of caspase-3 was obtained by pNA concentration (\u03bcM) using the following equation:\n\n${Caspase} - 3{activity},{\\mu M\\ pNA}/\\min/{mL} = \\frac{{\\mu M\\ pNA} \\times d}{t \\times v}$\n\n(t: time, v: volume of solution, d: dilution factor)\n\n*Statistical analysis*\n\nLevene\u00b4s test was used to check the homogeneity of variances. Data were analyzed using one-way analysis of variance (ANOVA) followed by Tukey Post-test. Results represent the mean \u00b1 standard deviation of the mean (S.D.) of triplicate samples. The minimal level of significance chosen was p \u2264 0.05.\n\nResults and Discussion\n======================\n\nUsing accelerated cytotoxicity mechanism screening (ACMS) technique, EC~50~ values were calculated as 56 \u03bcm for dacarbazine and 33 \u03bcm for compound III. These values indicate that pyridine derivative of dacarbazine (compound III) is almost two times more potent than dacarbazine. In order to investigate the molecular-cellular mechanism of cytotoxicity for compound III and dacarbazine, the effect of these compounds on hepatocyte cell death was evaluated in the presence of a wide variation of antioxidants (catalase, superoxide dismutase, *etc*.), ROS scavengers (mannitol, dimethylsulfoxide, *etc*.), a ferric chelator (desferoxamine), a CYP2E1 inhibitor (phenylimidazole), P~450~ reductase inhibitor (diphenyliodonium chloride - DPI), endocytosis inhibitors (chloroquineand methylamine) and mitochondrial permeability transitionpore inhibitors (cyclosporin and carnitine).In order to further investigate the mechanistic similarities between the cytotoxic activity of compound III and dacarbazine, the effect of these compounds on reactive oxygen species (ROS) formation, liposomal membrane leakiness and decrease in mitochondrial membrane potential were determinedby the measurement of the intensity of absorbance of fluorescence dyes with fluorescence spectrophotometer.\n\nWhen hepatocytes were incubated with 56 \u03bcm of dacarbazine and 33 \u03bcm of compound III, ROS formation increased very rapidly (peak in about 30 min, curve not shown) ([Table 1](#T1){ref-type=\"table\"}). The antioxidants: catalase, superoxide dismutase (SOD), butylatedhydroxytoluene (BHT) and ROS scavengers ([@B26]) mannitol and dimethylsulfoxide (DMSO) and the ferric chelator (desferoxamine) protected the hepatocytes against both DTIC and compound III induced cytotoxicity as well as ROS generation ([Table 1](#T1){ref-type=\"table\"}). All of these agents did not show any toxic effect on hepatocytes at the concentrations used (data not shown). However, the CYP2E1 inhibitor phenylimidazole ([@B26]-[@B30]) and P~450~ reductase inhibitor diphenyliodonium chloride (DPI) (26-30) showed significant effect on both DTIC and compound III induced cell lysis and ROS formation and protected the hepatocytes against dacarbazine and compound IIItoxicity ([Table 1](#T1){ref-type=\"table\"}). Endocytosis inhibitors including lysosomotropic agents (chloroquine ([@B31]) and methylamine ([@B32])) also protected the hepatocytes against DTIC and compound III induced cell lysis and ROS formation ([Table1](#T1){ref-type=\"table\"}). All of these agents did not show any toxic effect on hepatocytes at the concentrations used (data not shown). Cytotoxicity and ROS generationwere prevented by mitochondrial MPT pore sealing agents (carnitine and cyclosporine) ([Table1](#T1){ref-type=\"table\"}).\n\n###### \n\nEffect of antioxidant, ROS scavengers, ferric chelator, MPT pore sealing agents, lysosomotropic agents, and P~450~ reductase inhibitor on DTIC and Compound III -induced hepatocyte cytotoxicity and ROS formation\n\n  **Addition**                        **Cytotoxicity % (3h)**   **ROS (30min)**\n  ----------------------------------- ------------------------- -----------------\n  None                                20\u00b1 2                     79 \u00b1 4\n  Dacarbazine (56 \u03bcM )                76 \u00b1 4(1)                 230 \u00b1 4(1)\n  +Catalase (200 U/mL)                46 \u00b1 2(2)                 116 \u00b1 5(2)\n  +Superoxide dismutase (100 U/mL)    45 \u00b1 3(2)                 122 \u00b1 2(2)\n  +BHT (50 \u03bcM)                        42 \u00b1 3(2)                 118 \u00b1 4(2)\n  +Mannitol (50 mM)                   48 \u00b1 3(2)                 136 \u00b1 3(2)\n  +Dimethyl sulfoxide (150 \u03bcM)        44 \u00b1 3(2)                 121 \u00b1 2(2)\n  +Phenylimidazole (300 \u03bcM)           52 \u00b1 3(2)                 161 \u00b1 3(2)\n  +Diphenyliodoniumchloride (50 \u03bcM)   48 \u00b1 5(2)                 166 \u00b1 3(2)\n  +Methylamine (30 mM)                36 \u00b1 4(2)                 117 \u00b1 3(2)\n  +Chloroquine (100 \u03bcM)               40 \u00b1 3(2)                 128 \u00b1 2(2)\n  +Desferoxamine (200 \u03bcM)             36 \u00b1 2(2)                 121 \u00b1 3(2)\n  +Cyclosporine (2 \u03bcM)                34 \u00b1 3(2)                 138 \u00b1 3(2)\n  +Carnitine (2 mM)                   37 \u00b1 4(2)                 152 \u00b1 3(2)\n  **Compound III (33**\u03bc**M)**         73 \u00b1 2(1)                 256 \u00b1 5(1)\n  +Catalase (200 U/mL)                38 \u00b1 2(3)                 126 \u00b1 3(3)\n  +Superoxide dismutase (100 U/mL)    41 \u00b1 4(3)                 132 \u00b1 2(3)\n  +BHT(50 \u03bcM)                         37 \u00b1 4(3)                 128 \u00b1 2(3)\n  +Mannitol (50 mM)                   38 \u00b1 4(3)                 141 \u00b1 3(3)\n  +Dimethyl sulfoxide (150 \u03bcM)        36 \u00b1 3(3)                 145 \u00b1 2(3)\n  +Phenylimidazole (300 \u03bcM)           48 \u00b1 5(3)                 162 \u00b1 3(3)\n  +Diphenyliodoniumchloride (50 \u03bcM)   48 \u00b1 5(3)                 167 \u00b1 4(3)\n  +Methylamine (30 mM)                31 \u00b1 2(3)                 141 \u00b1 2(3)\n  +Chloroquine (100 \u03bcM)               46 \u00b1 3(3)                 155 \u00b1 3(3)\n  +Desferoxamine (200 \u03bcM)             35 \u00b1 3(3)                 136 \u00b1 3(3)\n  +Cyclosporine (2 \u03bcM)                28 \u00b1 2(3)                 141 \u00b1 2(3)\n  +Carnitine (2 mM)                   31 \u00b1 3(3)                 161 \u00b1 3(3)\n\nHepatocytes (106 cells/mL) were incubated in Krebs-Henseleit buffer pH 7.4 at 37 \u02daC for 3 h following the addition of DTIC and Compound III. Cytotoxicity was determined as the percentage of cells that take up trypan blue (19, 33). DCF formation was expressed as fluorescent intensity units (34). Values are expressed as means of three separate experiments (SD).\n\n\\(1\\) Significant difference in comparison with control hepatocytes (p \\< 0.05).\n\n\\(2\\) Significant difference in comparison with DTIC treated hepatocytes (p \\< 0.05).\n\n\\(3\\) Significant difference in comparison with compound III treated hepatocytes (p \\< 0.05).\n\nWhen hepatocyte lysosomes were preloaded with acridine orange, release of acridine orange into the cytosolic fraction ensued within 30min after treating the loaded hepatocytes with 56 \u03bcM of DTIC and 33 \u03bcm of compound III ([Table 2](#T2){ref-type=\"table\"}). The DTIC- and compound III-induced acridine orange release, which is a marker of lysosomal membrane damage, was prevented by ROS scavengers including dimethylsulfoxide, mannitol and antioxidants such as catalase,butylatedhydroxytoluene (BHT), superoxide dismutase (SOD) or the ferric chelatordesferoxamine ([Table2](#T2){ref-type=\"table\"}). Phenylimidazole and diphenyliodonium chloride (DPI) also inhibited dacarbazine and compound III acridine orange release ([Table 2](#T2){ref-type=\"table\"}). Dacarbazine- and compound III- induced acridine orange redistribution was prevented by chloroquine and methylamine ([Table 2](#T2){ref-type=\"table\"}). None of these agents alone at the concentrations used showedany significant effect on acridine orange release in acridine orange-loaded hepatocytes (data not shown).\n\n###### \n\nPreventing DTIC and Compound III induced hepatocyte lysosomal membrane damage by antioxidants, ROS scavengers, ferric chelator, CYP2E1 inhibitor, P~450~ reductase inhibitor, lysosomotropic agents\n\n  -----------------------------------------------------------------------------------------------------\n  **Addition**                        **% Acridine orange redistribution**\\                \n                                      **Incubation Time**                                  \n  ----------------------------------- --------------------------------------- ------------ ------------\n  None                                2 \u00b1 1                                   4 \u00b1 2        4 \u00b1 3\n\n  Dacarbazine (56 \u03bcM )                183 \u00b1 5(1)                              237 \u00b1 5(1)   250 \u00b1 4(1)\n\n  +Catalase (200 U/mL)                11 \u00b1 1(2)                               14 \u00b1 2(2)    18 \u00b1 2(2)\n\n  +SOD (100 U/mL)                     10 \u00b1 2(2)                               16 \u00b1 2(2)    20 \u00b1 2(2)\n\n  +BHT (50 \u03bcM)                        14 \u00b1 1(2)                               19 \u00b1 2(2)    25 \u00b1 3(2)\n\n  +Mannitol (50 mM)                   8 \u00b1 2(2)                                11 \u00b1 1(2)    13 \u00b1 1(2)\n\n  +Dimethyl sulfoxide (150 \u03bcM)        8 \u00b1 3(2)                                10 \u00b1 1(2)    12 \u00b1 1(2)\n\n  +Phenylimidazole (300 \u03bcM)           16 \u00b1 2(2)                               22 \u00b1 1(2)    30 \u00b1 3(2)\n\n  +Diphenyliodoniumchloride (50 \u03bcM)   18 \u00b1 3(2)                               26 \u00b1 3(2)    33 \u00b1 3(2)\n\n  +Methylamine (30 mM)                8 \u00b1 2(2)                                11 \u00b1 1(2)    14 \u00b1 1(2)\n\n  +Chloroquine (100 \u03bcM)               12 \u00b1 1(2)                               15 \u00b1 2(2)    20 \u00b1 2(2)\n\n  +Desferoxamine (200 \u03bcM)             8 \u00b1 2(2)                                10 \u00b1 2(2)    11 \u00b1 1(2)\n\n  Compound III (33 \u03bcM)                194 \u00b1 5(1)                              240 \u00b1 5(1)   264 \u00b1 5(1)\n\n  +Catalase (200 U/mL)                12 \u00b1 1(3)                               16 \u00b1 2(3)    18 \u00b1 2(3)\n\n  +SOD (100 U/mL)                     10 \u00b1 2(3)                               15 \u00b1 2(3)    21 \u00b1 2(3)\n\n  +BHT (50 \u03bcM)                        16 \u00b1 2(3)                               20 \u00b1 3(3)    24 \u00b1 3(3)\n\n  +Mannitol (50 mM)                   12 \u00b1 1(3)                               16 \u00b1 1(3)    19 \u00b1 2(3)\n\n  +Dimethyl sulfoxide (150 \u03bcM)        14 \u00b1 2(3)                               18 \u00b1 2(3)    21 \u00b1 2(3)\n\n  +Phenylimidazole (300 \u03bcM)           15 \u00b1 2(3)                               17 \u00b1 2(3)    24 \u00b1 2(3)\n\n  +Diphenyliodoniumchloride (50 \u03bcM)   18 \u00b1 3(3)                               22 \u00b1 2(3)    28 \u00b1 3(3)\n\n  +Methylamine (30 mM)                12 \u00b1 1(3)                               14 \u00b1 1(3)    17 \u00b1 2(3)\n\n  +Chloroquine (100 \u03bcM)               16 \u00b1 2(3)                               18 \u00b1 2(3)    25 \u00b1 2(3)\n\n  +Desferoxamine (200 \u03bcM)             10 \u00b1 1(3)                               12 \u00b1 1(3)    16 \u00b1 2(3)\n  -----------------------------------------------------------------------------------------------------\n\nHepatocytes (106 cells/mL) were incubated in Krebs-Henseleit buffer pH 7.4 at 37 \u02daC for 30 min. Lysosomal membrane damage was determined as intensity unit of diffuse cytosolic green fluorescence induced by acridine orange following the release from lysosome (16).\n\nValues are expressed as means of three separate experiments (SD).\n\n\\(1\\) Significant difference in comparison with control hepatocytes (p \\< 0.05).\n\n\\(2\\) Significant difference in comparison with DTIC treated hepatocytes (p \\< 0.05).\n\n\\(3\\) Significant difference in comparison with compound III treated hepatocytes (p \\< 0.05).\n\nAs shown in [Table 3](#T3){ref-type=\"table\"}, DTIC(56 \u03bcM) and Compound III(33 \u03bcM) induced a rapid decline of mitochondrial membrane potential (42% and 55% respectively) immediately after the incubation with hepatocytes and 77% and 90% after 30min of incubation at 37 \u00b0C which was prevented by reactive oxygen species scavengers (mannitol, DMSO), antioxidants (butylatedhydroxytoluene, catalase), suggesting that the observed decrease in mitochondrial membrane potential which was induced by DTIC and Compound III, was due to reactiveoxygen species formation. In addition, the NADPH P~450~ reductaseinhibitor, diphenyliodonium chloride and reduced CYP2E1 inhibitor, phenylimidazole, inhibited the decline of mitochondrialmembrane potential. Mitochondrial membrane potential collapse was prevented by mitochondrial MPT pore sealing agents (carnitine and cyclosporine) ([Table 3](#T3){ref-type=\"table\"}). All of these reagents including radicalscavengers, antioxidants, MPT pore sealing agents, NADPH P~450~ reductase inhibitor and reduced CYP2E1 inhibitor did not show any significant effect on hepatocyte mitochondrial membrane potential at the concentrations used while incubated alone (data not shown).\n\n###### \n\nMitochondrial membrane potential changes during DTIC and Compound III induced hepatocyte injury by antioxidants, ROS scavengers, CYP2E1 inhibitor, P450 reductase inhibitor and mitochondrial MPT pore sealing agents\n\n  **\u0394\u03a8m%Incubation Time**   Addition                \n  ------------------------- ----------- ----------- -----------------------------------\n  4 \u00b1 2                     3 \u00b1 1       2 \u00b1 1       None\n  77 \u00b1 3(1)                 56 \u00b1 2(1)   42 \u00b1 3(1)   **Dacarbazine (56**\u03bc**M )**\n  16 \u00b1 2(2)                 10 \u00b1 3(2)   6 \u00b1 2(2)    +Catalase (200 U/mL)\n  20 \u00b1 3(2)                 14 \u00b1 2(2)   6 \u00b1 3(2)    +BHT(50 \u03bcM)\n  21 \u00b1 2(2)                 16 \u00b1 2(2)   9 \u00b1 3(2)    +Mannitol (50 mM)\n  18 \u00b1 2(2)                 14 \u00b1 3(2)   6 \u00b1 2(2)    +Dimethyl sulfoxide (150 \u03bcM)\n  15 \u00b1 1(2)                 9 \u00b1 3(2)    6 \u00b1 3(2)    +Phenylimidazole (300 \u03bcM)\n  19 \u00b1 2(2)                 12 \u00b1 1(2)   8 \u00b1 2(2)    +Diphenyliodoniumchloride (50 \u03bcM)\n  16 \u00b1 2(2)                 10 \u00b1 1(2)   8 \u00b1 3(2)    +Cyclosporine (2 \u03bcM)\n  19 \u00b1 2(2)                 12 \u00b1 2(2)   8 \u00b1 2(2)    +Carnitine (2 mM)\n  90 \u00b1 4(1)                 66 \u00b1 6(1)   55 \u00b1 1(1)   **Compound III(33**\u03bc**M)**\n  18 \u00b1 2(3)                 12 \u00b1 2(3)   7 \u00b1 2(3)    +Catalase (200 U/mL)\n  18 \u00b1 3(3)                 15 \u00b1 3(3)   10 \u00b1 3(3)   +BHT(50 \u03bcM)\n  22 \u00b1 2(3)                 17 \u00b1 2(3)   9 \u00b1 2(3)    +Mannitol (50 mM)\n  18 \u00b1 2(3)                 15 \u00b1 2(3)   8 \u00b1 3(3)    +Dimethyl sulfoxide (150 \u03bcM)\n  22 \u00b1 2(3)                 16 \u00b1 3(3)   10 \u00b1 1(3)   +Phenylimidazole (300 \u03bcM)\n  26 \u00b1 3(3)                 18 \u00b1 2(3)   11 \u00b1 1(3)   +Diphenyliodoniumchloride (50 \u03bcM)\n  16 \u00b1 2(3)                 11 \u00b1 3(3)   5 \u00b1 2(3)    +Cyclosporine (2 \u03bcM)\n  20 \u00b1 2(3)                 14 \u00b1 2(3)   10 \u00b1 1(3)   +Carnitine (2 mM)\n\nHepatocytes (106 cells/mL) were incubated in Krebs-Henseleit buffer pH 7.4 at 37 \u02daC for 30 min. Mitochondrial membrane potential was determined as the difference in mitochondrial uptake of the rhodamine 123 between control and treated cells. Our data were shown as the percentage of mitochondrial membrane potential collapse (%\u0394\u03a8m) in all treated (test) hepatocyte groups (22, 35).\n\nValues are expressed as means of three separate experiments (SD).\n\n\\(1\\) Significant difference in comparison with control hepatocytes (p \\< 0.05).\n\n\\(2\\) Significant difference in comparison with DTIC treated hepatocytes (p \\< 0.05).\n\n\\(3\\) Significant difference in comparison with compound III treated hepatocytes (p \\< 0.05).\n\nAs shown in [Table 4](#T4){ref-type=\"table\"}, DTIC and Compound III induced caspase-3 activity in hepatocytes. Antioxidants (butylatedhydroxytoluene, catal ase), reactive oxygen species scavengers (mannitol, DMSO), NADPH P~450~ reductase inhibitor (diphenyliodonium chloride) and reduced CYP2E1 inhibitor (phenylimidazole), significantly decreased the activity of caspase-3. Mitochondrial MPT pore sealing agents (carnitine and cyclosporine) significantly decreased the activity of caspase-3 in comparison with hepatocytes incubated by DTIC and Compound III ([Table 4](#T4){ref-type=\"table\"}).\n\n###### \n\nBlockade ofDTIC and Compound III induced Caspase-3 activation by antioxidants, ROS scavengers, CYP2E1 inhibitor, p~450~ reductase inhibitor and mitochondrial MPT pore sealing agents\n\n  -------------------------------------------------------------\n  **Addition**                        **Caspase-3 Activity**\\\n                                      **2 h**\n  ----------------------------------- -------------------------\n  None                                297.65 \u00b1 5\n\n  Dacarbazine (56 \u03bcM )                659.52 \u00b1 7(1)\n\n  +Catalase (200 U/mL)                258.74 \u00b1 5(2)\n\n  +BHT (50 \u03bcM)                        260.52 \u00b1 3(2)\n\n  +Mannitol (50 mM)                   285.46 \u00b1 4(2)\n\n  +Dimethyl sulfoxide (150 \u03bcM)        281.33 \u00b1 5(2)\n\n  +Phenylimidazole (300 \u03bcM)           223.39 \u00b1 3(2)\n\n  +Diphenyliodoniumchloride (50 \u03bcM)   244.65 \u00b1 4(2)\n\n  +Cyclosporine (2 \u03bcM)                116.32 \u00b1 2(2)\n\n  +Carnitine (2 mM)                   124.38 \u00b1 3(2)\n\n  Compound III(33 \u03bcM)                 634.78 \u00b1 5(1)\n\n  +Catalase (200 U/mL)                138.73 \u00b1 4(3)\n\n  +BHT (50 \u03bcM)                        149.13 \u00b1 3(3)\n\n  +Mannitol (50 mM)                   281.22 \u00b1 3(3)\n\n  +Dimethyl sulfoxide (150 \u03bcM)        251.18 \u00b1 2(3)\n\n  +Phenylimidazole (300 \u03bcM)           173.89 \u00b1 3(3)\n\n  +Diphenyliodoniumchloride (50 \u03bcM)   188.36 \u00b1 3(3)\n\n  +Cyclosporine (2 \u03bcM)                96.18 \u00b1 2(3)\n\n  +Carnitine (2 mM)                   99.63 \u00b1 2(3)\n  -------------------------------------------------------------\n\nHepatocytes (106 cells/mL) were incubated in Krebs-Henseleit buffer pH 7.4 at 37 \u02daC for 30 min. The activating of caspase-3 (\u03bcM pNA/min/mL)) was determined based on hydrolysis of pNA labeled substrate (36). Values are expressed as means of three separate experiments (SD)\n\n\\(1\\) Significant difference in comparison with control hepatocytes (p \\< 0.05).\n\n\\(2\\) Significant difference in comparison with DTIC treated hepatocytes (p \\< 0.05).\n\n\\(3\\) Significant difference in comparison with compound III treated hepatocytes (p \\< 0.05).\n\nConclusion\n==========\n\nThe preliminarystudies for determination of EC~50~ of dacarbazine andcompound IIIreveals that the cytotoxicity of compound III is comparable with that of dacarbazine.Comparison ofthe EC~50~values for the two compounds even shows higher potency for compound III.\n\nThe results of cellular-molecular mechanistic studies for compound III and dacarbazine indicate that the two compounds show the same pattern of cytotoxicityform the mechanistic point of view.\n\nIn light of the fact that in compound III, an un-substituted pyridine ring hadreplaced the carbamoyl-imidazole ring in dacarbazine, it could be concluded that the imidazole ring and its dynamic tautomerization do not have significant role in dacarbazineinduced cytotoxic activity. Complementary detailed mechanistic studies such as investigation ofthe DNA-methylating properties for compound III will be further helpful to verify the accuracy of the above suggestion.\n"}
+{"text": "Glaucoma is the second most common cause of irreversible blindness, impairing 80 million people worldwide.[@bib1] Elevated intraocular pressure (IOP) is the most common indicator for diagnosis and progression of glaucoma, with reduction of IOP the only proven glaucoma intervention.[@bib2]^--^[@bib4] IOP is controlled by the balance between production and outflow of aqueous humor. In the conventional outflow pathway, the aqueous humor travels through the trabecular meshwork and juxtacanalicular connective tissue (JCT), into Schlemm\\'s canal (SC), through collector channels (CCs) passing through the sclera, and finally emptying into the episcleral veins.[@bib5]^,^[@bib6] The increased IOP characteristic of primary open-angle glaucoma is caused by increased resistance to aqueous humor outflow.[@bib7]^,^[@bib8] Most investigations of outflow resistance have focused on the JCT and inner wall of SC, which are widely regarded as the principal sites of outflow resistance.[@bib9] However, flow resistance is also generated in SC and CCs, particularly in the former as IOP is increased.[@bib8]^,^[@bib9]\n\nIn primate, humans, and mice, SC is essentially a flat tube traveling circumferentially around the limbus of the eye, whose opening thickness, the distance from the inner wall to outer wall, decreases as IOP increases.[@bib9] Greater lengths of SC are collapsed in glaucomatous eyes than healthy eyes, even when perfused at the same IOP,[@bib10] and such collapse has been speculated to contribute to the elevated IOP characteristic of glaucoma.[@bib11]^--^[@bib13]\n\nA clinical practice to improve visualization of SC is to increase pressure of the downstream episcleral veins so that blood flows backward from the normal path and into SC. Recent studies have shown that the quality of such blood reflux into SC immediately before or during surgery (both selective laser trabeculoplasty and canaloplasty) correlates with the success of the surgery and postsurgery IOP levels.[@bib14]^,^[@bib15] As a result, a method that can noninvasively image the outflow pathway will bring significant benefits to both clinical and preclinical investigations. We seek to image SC and associated limbal microvascular network in the mouse using optical coherence tomography (OCT).\n\nOCT is widely used in noninvasive ophthalmic imaging,[@bib16] which provides in vivo anatomical and functional information of both the anterior and posterior eyes at an axial resolution of several micrometers.[@bib17]^,^[@bib18] Researchers have used OCT to investigate properties of SC in vivo in mice and humans.[@bib17]^,^[@bib19]^--^[@bib26] These studies were limited in axial resolution (2--5 \u03bcm) and field of view (0.5--3 mm). Imaging the detailed structure of SC in mice requires overcoming several hurdles. First, the relatively small size of SC, especially in mice, requires a higher resolution to reveal the details of SC and to investigate the variation of SC in different conditions. Second, SC is a circular channel located around the limbus, with optimal image quality occurring when the OCT probing light is perpendicular to the curvature of the eye at the location of SC. However, existing OCT systems are usually designed such that the incident light is along the optical axis of the eye, requiring the mouse to be tilted significantly to optimally image SC. Third, the extreme variations in SC size and shape within a single mouse eye suggest that it is important to image the entire SC instead of a certain section of SC. If the entire SC is imaged with a traditional OCT system, the mouse must be rotated. When rotating the mouse, the IOP and SC size become unstable, where mouse demonstrates higher IOP and smaller SC when the head is lower than the body.[@bib27] Additionally, it has previously been shown that episcleral venous pressure (EVP) can also vary with the angle of the mouse body, affecting the natural outflow of aqueous humor.[@bib27]\n\nTo overcome these challenges, we developed an anterior segment visible-light OCT (vis-OCT) to conduct a circumlimbal scan and image the whole SC without rotating the mouse. Vis-OCT is a newly developed OCT technology, which uses visible light as its light source rather than the commonly used near-infrared light in commercial OCT systems.[@bib28]^--^[@bib30] Using shorter wavelengths allows for higher axial resolution for a given bandwidth and higher backscattering sensitivity in tissue.[@bib31]^,^[@bib32] The higher axial resolution allows more precise measurement of SC anatomy as well as SC changes. Additionally, taking advantage of improved sensitivity to slow blood flow, vis-OCT angiography (vis-OCTA) has the potential to better visualize the superficial limbal microvasculature and the outflow veins downstream from SC.\n\nIn this study, we first validated our vis-OCT\\'s capability to measure the morphology of SC at an axial resolution of approximately 1 \u00b5m in a three-dimensional (3D) printed phantom eyeball and in wild-type mice. Then, we artificially adjusted the IOPs in wild-type mice using a manometer and investigated SC cross-sectional size change with IOP. In addition, we introduced blood reflux into SC and showed that the backscattering contrast provided by blood reflux into SC allows for direct visualization of SC using vis-OCTA. Finally, we visualized the entire SC and its surrounding limbal microvascular network in 3D using a compound circumlimbal scan.\n\nMethods and Materials {#sec2}\n=====================\n\nAnimal Preparation {#sec2-1}\n------------------\n\nOur experimental protocol was approved by the Northwestern University Institutional Animal Care and Use Committee and conformed to the Association for Research in Vision and Ophthalmology Statement on Animal Research. We used adult wild-type C57BL/6 mice approximately 3 months old in our experiments. The mice were kept under normal lighting conditions with 12-hour-on, 12-hour-off cycles in the Center for Comparative Medicine at Northwestern University. Prior to any imaging or procedure, the mice were anesthetized by intraperitoneal injection (10 mL/kg body weight) of a ketamine/xylazine cocktail (ketamine: 11.45 mg/mL; xylazine: 1.7 mg/mL, in saline). During imaging, the mouse\\'s body temperature was maintained by a heating lamp, and a pulse oximeter was attached to the left rear paw to monitor peripheral arterial sO~2~ and heart rate. After imaging, the animals were placed in a warm cage for recovery, and toe/tail pinch withdrawal, heart rate, breath rate, and sO~2~ were monitored every 20 minutes until fully recovered.\n\nExperimental System Setup {#sec2-2}\n-------------------------\n\n[Figure\u00a01](#fig1){ref-type=\"fig\"}a shows the schematic of the experimental vis-OCT system. Briefly, light from a supercontinuum laser (SuperK EXTREME; NKT Photonics, Birker\u00f8d, Denmark) was filtered and sent to a 30:70 fiber coupler (Gould Fiber Optics, Millersville, MD). A pair of galvanometer mirrors (Nutfield Technology, Londonderry, NH) scanned the beam through an objective lens (LSM03, ThorLabs, Newton, NJ; SL in\u00a0[Fig.\u00a01](#fig1){ref-type=\"fig\"}a), which focused the light onto the sample. Prior to the objective lens, the beam diameter was 2 mm. A visible light spectrometer (Blizzard SR; Opticent Health, Evanston, IL) operating from 510 to 610 nm detected the interferogram signals for image reconstruction. The theoretical axial and lateral resolutions of the system were 1.3 and 6.8 \u00b5m, respectively.\n\n![(**a**) Schematic of the vis-OCT system. DC, dispersion compensation; FC, fiber coupler; GM, galvanometer mirrors; M, mirror; MRM, motorized rotational mount; SCL, supercontinuum light source; SL, scan lens; SM, spectrometer; OA, optical axis of the eye, which is also the axis of rotation of the motorized rotational mount. (**b**) Detailed illustration of the components being highlighted by the *red dashed box* in panel (**a**). M, mirror; OA, optical axis; SC, Schlemm\\'s canal. The incident light reflects off both the mirrors after the SL, such that it is angled perpendicularly to the surface of the eye at SC, as shown in the *blue* inset. (**c**) Illustration of the cannulation of the anterior chamber with an open syringe filled with physiologic saline, whose height can be adjusted to set IOP levels in the mouse eye. (**d**) Illustration of sutures to cause blood reflux into SC.](iovs-61-2-23-f001){#fig1}\n\nDue to the anatomical location of SC in the anterior segment, either the mouse eye or the vis-OCT light beam needs to tilt approximately 60\u00b0 from the optical axis of the eye to best image SC and the limbus. Because tilting the mouse by such a large angle has been shown to affect EVP,[@bib27] we chose to tilt the light beam 60\u00b0 using a two-mirror assembly ([Fig.\u00a01](#fig1){ref-type=\"fig\"}b) to avoid unwanted EVP elevation-induced SC perturbation. We mounted this assembly and the objective lens on a motorized rotational stage (PRM1Z8; ThorLabs). We rotated the two-mirror assemble and objective lens eight times at a step size of 45\u00b0 about the ocular axis ([Fig.\u00a01](#fig1){ref-type=\"fig\"}a, dashed line) and performed a raster scan at each rotational step. As a result, we created a compound circumlimbal scan to image the full SC and limbal microvascular network while the mouse stayed stationary.\n\nIOP-Level Control {#sec2-3}\n-----------------\n\nWe cannulated the anterior chamber and manometrically set the IOP level in the mouse ([Fig.\u00a01](#fig1){ref-type=\"fig\"}c). We first used a 33-gauge lancet to create a hole near the limbus, then placed a blunt 34-gauge needle through that hole. This 34-gauge needle was stabilized by attaching it to an external support. As shown in\u00a0[Figure\u00a01](#fig1){ref-type=\"fig\"}c, we connected the needle to an open syringe filled with physiologic saline. The height of the syringe manometrically controlled the IOP. We used a tonometer (TonoLab Tonometer, Icare, Finland) to measure the IOP values of the mouse eyes. We confirmed the height to set the saline column for baseline (\u03940) IOP position by using the tonometer. In our study, we adjusted the IOP to be \u00b110 mm Hg and \u00b15 mm Hg with respect to the baseline IOP, which varied from 7 to 10 mm Hg and agreed with a previously published average baseline IOP in wild-type C57BL/6 mice anesthetized with ketamine and xylazine.[@bib33] Using our real-time vis-OCT preview, we found that the shape and size of SC stabilized approximately 1 minute after changing the IOP. We allowed the eye to adjust to the each new IOP level for 3 minutes before acquiring a vis-OCT or vis-OCTA image. We also note that because of the needle in the anterior chamber, we will be unable to image the entire SC in manometry studies.\n\nEpiscleral Vein Blockage to Cause Blood Reflux {#sec2-4}\n----------------------------------------------\n\nWe cut approximately 1 to 2 mm at the outer canthus and added a suture (7-0 thread, using a round-tip needle) approximately 1 mm nasal of the original outer canthus ([Fig.\u00a01](#fig1){ref-type=\"fig\"}d). We then used forceps to gently squeeze the skin around the eyelid, causing the eye to slightly pop out of the eyelid. We then tightened the suture so that the eye could not fall back into the socket but remained slightly exposed.\n\nPhantom Eyeball and Compound Circumlimbal Scan Calibration {#sec2-5}\n----------------------------------------------------------\n\nTo calibrate geometries of the eight raster scans to form a montage circumlimbal scan, we made a 3D-printed phantom eyeball, which had comparable size and geometry to a mouse eyeball (sphere, diameter: 3 mm). The phantom eyeball was also 3D printed to include an annular groove around the sphere at the approximate position of SC in the anterior segment. We heated a polyethylene (PE-50) tube and pulled it over a flame to reduce the inner diameter from approximately 530 \u00b5m to approximately 150 \u00b5m. Then we laid the tube into the groove to mimic SC and connected it via a blunt 23-gauge needle to a syringe placed in a micro pump (Harvard Apparatus, Holliston, MA). The mechanical drawing of the phantom eyeball is shown in\u00a0[Figure\u00a02](#fig2){ref-type=\"fig\"}a, where the tube is colored yellow and flow direction is highlighted by the blue dashed arrows. We coated the phantom eyeball with ultraviolet-cured optical adhesive (Norland Optical Adhesive 81; Norland Products, Inc.), which contained ceramic scattering particles (200 nm) to mimic the highly scattering sclera.\n\n![(**a**) Schematic of the 3D-printed phantom eyeball, with tube inset into groove shown in *yellow*. Direction of the Intralipid flow is shown with the *blue dashed line*. (**b**) Vis-OCT B-scan image of the phantom eyeball without flowing Intralipid. (**c**) Vis-OCT B-scan image of the phantom eyeball with flow Intralipid. (**d**) Vis-OCTA B-scan image of the phantom eyeball with flowing Intralipid. (**e**) Geometry of how the eight raster scans are montaged onto a sphere. (**f**) 3D visualization of the montaged image of the phantom eyeball.](iovs-61-2-23-f002){#fig2}\n\n[Figure\u00a02](#fig2){ref-type=\"fig\"}b shows a cross-sectional vis-OCT B-scan of the phantom eyeball, with highlighted hollow tube and coating layer. We diluted Intralipid (Millipore Sigma, Burlington, MA) to 2% with phosphate-buffered saline and flowed it through the tube ([Fig.\u00a02](#fig2){ref-type=\"fig\"}c) to create a positive contrast to image SC using vis-OCTA ([Fig.\u00a02](#fig2){ref-type=\"fig\"}d). We acquired eight raster scans, each after rotating the motorized mount 45\u00b0. In each raster scan, the two-mirror assembly tilted the beam 60\u00b0 from the optical axis. To match this, each OCTA en face image was projected from a plane angled 60\u00b0 from the optical axis and radially spaced 45\u00b0 apart ([Fig.\u00a02](#fig2){ref-type=\"fig\"}e). Each projected image was manually fine adjusted to produce the final montaged image ([Fig.\u00a02](#fig2){ref-type=\"fig\"}f).\n\nFinally, we quantified the accuracy of the montaged image. The defining characteristic of the phantom eyeball is the angle of the inlet and outlet grooves, which were designed to be precisely 45\u00b0. Using ImageJ (National Institutes of Health, Bethesda, MD, USA), we measured the angle of both the inlet and outlet tubing five times each in the final montaged image. The inlet tubing has a measured mean path angle of 45.1\u00b0 \u00b1 0.5\u00b0 and the outlet tubing has a measured mean path angle of 44.4\u00b0 \u00b1 12\u00b0. To validate the accuracy of vis-OCT measurement of tube cross-sectional areas, we imaged a glass capillary tube (\\#1-000-800; Drummond Scientific Co., Broomall, PA) filled with water. Our measured inner diameter was 392 \u00b5m, which agreed well with the manufacturer\\'s specified diameter of 400 \u00b5m.\n\nVis-OCT Imaging {#sec2-6}\n---------------\n\nPrior to vis-OCT imaging, we immobilized the mouse on a homemade holder. We first aligned the vis-OCT system without the two-mirror assembly so that the optical axis of the mouse eye aligned with the vis-OCT probing beam and the pupil was centralized. After we centered the eye, we mounted the two-mirror assembly for SC imaging, and if needed, we readjusted the vertical position of the mouse for better optical focusing. The oblique-incident vis-OCT beam was approximately perpendicular to the scleral plane. In the compound circumlimbal scan, each raster scan area was 1.8 \u00d7 1.8 mm^2^, which covered roughly 20% of the perimeter of SC. For anatomical imaging, each raster scan contained 512 B-scans, and each B-scan contained 512 A-lines. For vis-OCTA imaging, we repeated each B-scan five times. The spectrometer integration time was 40 \u00b5s, and the vis-OCT illumination power was approximately 1.0 mW.\n\nImage Processing and Montage {#sec2-7}\n----------------------------\n\nWe developed a semiautomatic method to segment SC in vis-OCT B-scan using a custom MATLAB (Mathworks, Natick, MA, USA) program. Briefly, we upsampled the vis-OCT B-scan image 1580 \u00d7 2000 pixels and applied a mean filter (10 \u00d7 10 pixels). We then binarized the B-scan image based on a user-specified intensity threshold and automatically removed small structures that were less than 25 pixels in area. The user then clicks inside SC, and an automatic filling algorithm, based on morphologic reconstruction, extends from the pixel the user clicked outward until boundaries are reached on all sides. A change in intensity between a pixel and any of its four vertically or horizontally adjacent pixels creates a boundary.[@bib34] If necessary, the user can manually modify the segmented boundary. For 3D visualization and volume measurement, we segmented SC from every other vis-OCT B-scan image. The spacing between segmented B-scan images is 7.4 \u00b5m, matching well with the theoretical lateral resolution (6.8 \u00b5m, Rayleigh criteria under Gaussian beam profile) of the vis-OCT system. As a result, the segmented B-scans are very close to both nonoverlapping and without gaps. After SC was segmented from all B-scans, we used open-source software (3D Slicer) to visualize it in 3D. To calculate cross-sectional area of SC, we multiplied the vis-OCT pixel size within a B-scan image with the total number of pixels within the segmented SC. To calculate SC volume, we summed all the cross-sectional areas within a selected SC segment and multiplied the total area with the corresponding B-scan thickness.\n\nWe validated the repeatability of our segmentation using a test-retest method, where we repeated the IOP-level control sequence twice back to back: first decreasing from \u0394+10 mm Hg to \u0394--10 mm Hg in steps of 5 mm Hg, then increasing from \u0394--10 mm Hg to \u0394+10 mm Hg in steps of 5 mm Hg. Before comparing the two data sets, we corrected lateral displacement through cross-correlation across 400 B-scans. We calculated the Pearson correlation coefficient between the two \u0394--10-mm Hg data sets and between the two \u0394+5-mm Hg data sets. The two \u0394--10-mm Hg data sets were acquired successively without changing IOP values between two acquisitions. Their Pearson correlation coefficient was 0.81, representing a good match.[@bib35] The two \u0394+5-mm Hg data sets were acquired with a 24-minute delay and the six distinct IOP-level changes. The Pearson correlation coefficient between the two \u0394+5 data sets was 0.66, representing a moderate match.[@bib35] Please refer to [Supplementary Figure S1](#IOVS-61-2-23_s003){ref-type=\"supplementary-material\"} for more details.\n\nFor processing vis-OCTA data, we followed the method described by Chen et al.[@bib36] To create a full-view SC image, we montaged and mapped the individual raster scans vis-OCTA en face images taken at different positions onto a sphere using Solidworks (Dassault Systems, V\u00e9lizy-Villacoublay, France) as illustrated in\u00a0[Figure\u00a02](#fig2){ref-type=\"fig\"}e. Specifically, en face vis-OCTA images were projected onto the surface of a 3-mm-diameter sphere. To match the scanning geometry, each en face image was projected from a plane angled 60\u02da from the vertical axis with each subsequent plane radially spaced 45\u02da from the other. If needed, we manually readjust each projected image for better alignment as shown in\u00a0[Figure\u00a02](#fig2){ref-type=\"fig\"}f. We compared the measured SC from both segmentation and vis-OCTA by calculating the Dice coefficient[@bib37] in the respectively projected images.\n\nResults {#sec3}\n=======\n\n[Figure\u00a03](#fig3){ref-type=\"fig\"} shows that the ultrahigh resolution vis-OCT imaging of the anterior segment reveals the large variation in the shape and size of SC, even within a 1-mm section. In\u00a0[Figure\u00a03](#fig3){ref-type=\"fig\"}a, the red box highlights the approximate location of the imaged SC section and the three blue arrows highlight the scanning direction and locations of the B-scan images.\u00a0[Figure\u00a03](#fig3){ref-type=\"fig\"}b shows the final result after surface rendering of a segmented 1-mm SC section using 3D Slicer, an open-source software.[@bib38] The *x:y:z* aspect ratio was stretched to 1:2:3 for visualization purposes. Additionally, the model was smoothed along the *y*-axis (B-scan direction) across five B-scans. The two ends were not smoothed.\u00a0[Figures 3](#fig3){ref-type=\"fig\"}c--[3](#fig3){ref-type=\"fig\"}e show three representative vis-OCT B-scan images (from the positions highlighted in\u00a0[Fig.\u00a03](#fig3){ref-type=\"fig\"}a) used for segmentation and 3D visualization. We refer to these B-scan images as perpendicular views of SC. The segmented SC is outlined in red in each B-scan image. Due to lack of optical scattering contrast, the area inside SC shows noticeably lower vis-OCT intensity than the surrounding regions. We highlight the (i) sclera, (ii) limbus, (iii) cornea, (iv) ciliary body, and (v) iris in\u00a0[Figure\u00a03](#fig3){ref-type=\"fig\"}c, which are landmarks to facilitate locating SC. The magnified views of the areas highlighted by the blue dashed boxes in\u00a0[Figures 3](#fig3){ref-type=\"fig\"}c--[3](#fig3){ref-type=\"fig\"}e are respectively shown in\u00a0[Figures 3](#fig3){ref-type=\"fig\"}f--[3](#fig3){ref-type=\"fig\"}h, where segmented SC sizes are 8743 \u00b5m^2^, 8168 \u00b5m^2^, and 4047 \u00b5m^2^, respectively. The pixel size is approximately 1.1 \u00b5m along both the axial and lateral directions.\n\n![(**a**) Illustration of a top view of the mouse eye with positions and directions of three B-scans (*blue arrow*) shown within a raster scan (*red box*). (**b**) 3D visualization of the segmented SC from 270 adjacent B-scans, equivalent to 1 mm in length, with *x:y:z* aspect ratios of 1:2:3 (see [Supplementary Video S1](#IOVS-61-2-23_s001){ref-type=\"supplementary-material\"}). Panels (**c**), (**d**), and (**e**) are vis-OCT B-scan images from the three positions highlighted in panel (**a**); SC is outlined in *red*. (i) sclera, (ii) limbus, (iii) cornea, (iv) ciliary body, and (v) iris. Panels (**f**), (**g**), and (**h**) are magnified views of SC corresponding to the areas highlighted in panels (**c**), (**d**), and (**e**), respectively.](iovs-61-2-23-f003){#fig3}\n\n[Figure\u00a04](#fig4){ref-type=\"fig\"} shows the results of imaging SC size variation under different monometrically set IOP levels. We visualized SC size variations along the canal (parallel view) and perpendicular to the canal (perpendicular view) from the approximate positions illustrated in\u00a0[Figures 4](#fig4){ref-type=\"fig\"}a and\u00a0[4](#fig4){ref-type=\"fig\"}b, respectively.\u00a0[Figures 4](#fig4){ref-type=\"fig\"}c--[4](#fig4){ref-type=\"fig\"}g show the parallel-view B-scan images of SC from the same position when we changed the IOP from --10 mm Hg to +10 mm Hg with respect to the baseline IOP at a step size of 5 mm Hg. In these parallel-view images, we can see the elongated cross section of the imaged SC (dark area highlighted by red arrowheads).\u00a0[Figures 4](#fig4){ref-type=\"fig\"}h--[4](#fig4){ref-type=\"fig\"}l show perpendicular-view B-scan images taken from the position highlighted in\u00a0[Figure\u00a04](#fig4){ref-type=\"fig\"}b under the same five different IOP levels. In both the parallel and perpendicular views, SC cross-sectional areas were obviously larger at a lower IOP and decreased in size as IOP increased. When the IOP was 10 mm Hg above the baseline IOP, SC nearly collapsed completely ([Figs.\u00a04](#fig4){ref-type=\"fig\"}g and\u00a0[4](#fig4){ref-type=\"fig\"}l). At higher IOP levels ([Figs.\u00a04](#fig4){ref-type=\"fig\"}f--g,\u00a0[4](#fig4){ref-type=\"fig\"}k--l), entire sections of SC appeared collapsed, for up to tens of microns in length.\n\n![(**a**, **b**) Illustrations of a top view of the mouse eye with positions and directions of the parallel-view and perpendicular-view B-scan images, respectively. (**c--g**) Parallel-view vis-OCT images under different IOP levels from the position highlighted in panel (**a**). (**h--l**) Parallel-view vis-OCT images under different IOP levels from the position highlighted in panel (**b**). The *red arrowheads* highlight SC in the B-scan images. (**m**) Examples of SC cross-sectional area changes at different positions along the canal under different IOP levels. (**n**) Mean\u00a0\u00b1 standard deviation volumes of 1.5-mm SC segments at IOP levels of --10 mm Hg, --5 mm Hg, 0 mm Hg, +5 mm Hg, and +10 mm Hg with respect to the baseline IOP.](iovs-61-2-23-f004){#fig4}\n\nAt each of the five IOP levels, we segmented an approximately 1.5-mm length of SC acquired in the perpendicular view, with a spacing of 7.4 \u00b5m between segmented B-scans. We calculated the cross-sectional area of SC (in \u00b5m^2^) for each B-scan analyzed. The results of this analysis in one representative mouse are shown in\u00a0[Figure\u00a04](#fig4){ref-type=\"fig\"}m, with SC areas plotted against B-scan positions. There appears to be a periodic pattern to the area increasing and decreasing, which remains true across all the IOP levels. This observation is consistent among all the mice imaged (*n* = 8). In addition, the change in SC volume against IOP level is further illustrated in\u00a0[Figure\u00a04](#fig4){ref-type=\"fig\"}n.\n\n[Figure\u00a05](#fig5){ref-type=\"fig\"} shows the comparison between directly visualized SC based on blood reflux and results from segmenting the canal. We introduced blood reflux into SC through episcleral vein blockage to create optical scattering contrast inside SC as described in the Methods and Materials section and illustrated in\u00a0[Figure\u00a01](#fig1){ref-type=\"fig\"}d. Taking advantage of the motion of refluxed red blood cells in SC, vis-OCTA can directly visualize SC.\u00a0[Figure\u00a05](#fig5){ref-type=\"fig\"}a shows a normal mouse eye, with SC nearly invisible. After episcleral vein blockage, SC became bright red and clearly visible, confirming that blood was pushed back into SC ([Fig.\u00a05](#fig5){ref-type=\"fig\"}b).\u00a0[Figures 5](#fig5){ref-type=\"fig\"}c and\u00a0[5](#fig5){ref-type=\"fig\"}d show the parallel-view vis-OCT B-scan images of SC from the same location before and after suturing. SC (dark region highlighted by the red arrowheads) is enlarged after EVP increase, indicating that SC is no longer in its natural physiologic state after blood reflux. With blood reflux, the size and shape of SC are similar to what we observed when IOP was 10 mm Hg below the baseline IOP level. A 3D visualization of vis-OCTA images covering a 1.8 mm \u00d7 1.8 mm area is shown in\u00a0[Figure\u00a05](#fig5){ref-type=\"fig\"}e. SC is clearly visualized by vis-OCTA, which relies on motion contrast, confirming that blood reflux enhanced the imaging contrast of SC as a result of blood cells flowing inside SC. We also segmented SC from half of the vis-OCT volume and overlaid the segmented SC (pseudo-colored in yellow) onto the vis-OCTA volume (semitransparent gray). Both the segmented results and vis-OCTA results of the blood-refluxed SC showed near-perfect overlap, confirming that our semiautomatic segmentation faithfully detected SC from vis-OCT images. Indeed, after segmenting the projected vis-OCTA volume to remove most of the vessels surrounding SC, the overlap between the projected vis-OCTA SC and the projected segmented SC, as quantified by the Dice coefficient, was 0.87, implying good agreement.\n\n![(**a**, **b**) Photos of mouse eyes without and with the episcleral vein blocked. The *red arrows* in (**b**) highlight SC after it is filled with blood. (**c**, **d**) Vis-OCT parallel-view B-scan images before and after blood reflux. The *red arrows* in both the photos and vis-OCT parallel-view B-scan images point to SC. (**e**) 3D visualization of SC with blood reflux imaged by vis-OCTA directly (*gray*) and overlaid segmentation result (*yellow*).](iovs-61-2-23-f005){#fig5}\n\n[Figure\u00a06](#fig6){ref-type=\"fig\"} shows the results of imaging the full SC under elevated EVP and the associated limbal microvascular network using our compound circumlimbal scan as illustrated in\u00a0[Figure\u00a02](#fig2){ref-type=\"fig\"}e. We acquired eight raster scans using vis-OCTA, rotating the attachment 45\u00b0 between each raster scan. The final montaged images projected onto a sphere, forming the circumlimbal view, are shown in\u00a0[Figure\u00a06](#fig6){ref-type=\"fig\"}a.\u00a0[Figure\u00a06](#fig6){ref-type=\"fig\"}b is a projection view of the four quadrants of the entire SC and limbal microvascular network, mimicking the commonly used flat-mount microscopic images.[@bib39] The superficial limbus vasculature is color-coded in the left quadrant, by matching anatomical features with previously published histologic studies.[@bib40]\n\n![(**a**) 3D visualization of montaged vis-OCTA circumlimbal image of SC under elevated EVP and the surrounding limbal microvascular network (see [Supplementary Video S2](#IOVS-61-2-23_s002){ref-type=\"supplementary-material\"}). (**b**) Projection view of the imaged entire SC and limbal microvascular network in the flat-mount style four quadrants. The left quadrant is pseudo-colored to highlight different anatomical features of the limbal microvessels. PLV, perilimbal vein; CLA, circular limbal artery; CA, corneal arcades; SC, Schlemm\\'s canal.](iovs-61-2-23-f006){#fig6}\n\nDiscussion {#sec4}\n==========\n\nWe demonstrate, for the first time, in vivo imaging of the entire SC and its surrounding limbal microvascular network in mice. We developed an oblique-incident vis-OCT system to accommodate the curvature of the eyeball for the best SC imaging quality. We empirically identified that the best imaging angle between the vis-OCT incident light beam and the optical axis of the eye is approximately 60\u00b0 from the optical axis. To avoid rotating the mouse and the associated unwanted perturbations in both IOP and EVP, we added a two-mirror assembly that was rotated about the optical axis of the eye. We used a 3D-printed phantom eyeball to calibrate the geometries of this rotational scanning setup for full-field SC structure reconstruction. We found that each raster scan of 1.8 \u00d7 1.8 mm^2^ can cover approximately 20% of the total SC perimeter, and eight rotations of the raster scan will ensure adjacent images to have sufficient overlap for circumlimbal 3D reconstruction and montage.\n\nIn our study, we used two methods, manometry to change IOP and blood reflux, to validate the capability of our vis-OCT system on noninvasive imaging of SC and its size alterations. Several studies on both mouse and human support the observation that there is a large variation in SC cross-sectional area in a single subject, which is further dependent upon IOP, physical exercises, glaucomatous status, blood reflux, and congenital abnormalities.[@bib41]^--^[@bib45] We visualized such rapid changes in SC cross-sectional area within a 1-mm section of the mouse SC in the 3D surface rendering ([Fig.\u00a03](#fig3){ref-type=\"fig\"}b).\n\nUsing manometry, we confirmed that cross-sectional area can vary significantly even within a short 1.5-mm section of SC ([Fig.\u00a04](#fig4){ref-type=\"fig\"}m). Indeed, periodic peaks are visible roughly every 400 \u03bcm, and the locations of these peaks remain consistent throughout the IOP-level changes. As it has been reported in humans that aqueous outflow is increased near the collector channels, and SC is known to have a larger cross-sectional area near the collector channels,[@bib46] these peaks may coincide with the locations of collector channels in mouse. Thus, cautions should be taken when considering tracking SC alterations according to only a few B-scans. Instead, the volume of a given section of SC should be more reliable for quantitative analysis of SC changes. Our results show that SC volume significantly decreased with the IOP elevation ([Fig.\u00a04](#fig4){ref-type=\"fig\"}n) and demonstrate that even in wild-type mice, a 10-mm Hg increase in IOP leads to nearly total collapse of SC according to vis-OCT image ([Figs.\u00a04](#fig4){ref-type=\"fig\"}g and\u00a0[4](#fig4){ref-type=\"fig\"}l).\u00a0[Figure\u00a04](#fig4){ref-type=\"fig\"}n shows an average reduction of 72% at \u0394+10 mm Hg and 26% reduction at \u0394+5 mm Hg. As reported by Li et al.,[@bib25] the SC size reduces roughly 80% at 20 mm Hg and 37% at 15 mm Hg. Li et al.^25^ did not set the IOP levels relative to baseline IOP but rather at absolute values. Regardless, the values obtained are comparable to the degree of collapse that we observed. We recognize that there exists room for error in measuring SC cross-sectional area due to segmentation inaccuracy, limitations of vis-OCT resolution, and influence from signal-to-noise ratio and speckles. The inverse relationship between IOP and SC cross-sectional size in wild-type mice matches well with previously published human studies.[@bib47]^,^[@bib48] Several clinical studies have shown that SC is smaller in glaucomatous eyes as compared with normal eyes, and it is thought that decreased SC lumen size contributes to the elevated pressure characteristic of primary open-angle glaucoma.[@bib13]^,^[@bib44]^,^[@bib48]^,^[@bib49]\n\nSome researchers consider visualization of the blood in SC after sclera compression as a positive prognosis factor of successful antiglaucoma surgery.[@bib14]^,^[@bib15] Blood refluxed into SC may reflect the communication between the episcleral veins and SC. The mouse blood reflux test was performed by increasing the pressure of the episcleral vein, causing the blood to flow in the reverse direction of its usual flow and enter SC. Through vis-OCT imaging, blood reflux directly increased SC size after episcleral vein blockage. Additionally, the refluxed blood in SC also shows positive signal in vis-OCTA, thus allowing us to directly visualize SC without segmentation or introducing any exogenous contrast agent.[@bib50] Although several publications have used OCTA to image the SC in mouse, only B-scan images were shown and the angiography information was used as a tool in segmenting the SC rather than enabling direct visualization of SC.[@bib24]^--^[@bib26] Although not in its natural physiologic state, the entire structure of SC can be noninvasively and fully reconstructed. Therefore, the enlargement of SC and its uniformity under external perturbation imaged by OCT may be a potential measure to predict the success of certain antiglaucoma surgeries such as laser trabeculoplasty and canaloplasty.[@bib14]^,^[@bib15]\n\nThe perilimbal vein (PLV), pseudo-colored in blue in\u00a0[Figure\u00a06](#fig6){ref-type=\"fig\"}b, has been described as a plexus forming a continuous ring around the limbus.[@bib40] These features of the PLV are evident in the circumlimbal scan. The circular limbal artery (CLA), pseudo-colored in red in\u00a0[Figure\u00a06](#fig6){ref-type=\"fig\"}a, is described as a single narrow artery, which can often be found with the PLV.[@bib40] The corneal arcades (CAs), pseudo-colored in green, are looped vessels that anastomose with both CLA and PLV.[@bib40] Therefore, the CAs can be considered capillary equivalents in the limbal vasculature. Finally, SC (pseudo-colored in orange) often but not always runs directly parallel to the superficial limbal vasculature.\n\nIn summary, our findings demonstrate that we can image the entirety of SC and its surrounding limbal microvascular network in mice using vis-OCT and vis-OCTA. In agreement with other investigations,[@bib51]^,^[@bib52] we found that at normal IOP, SC is a continuous structure that extends around the limbus. In the future, we will investigate the size variation in normal SC and validate vis-OCT measurement by histologic analysis as well as electron microscopy. We will further investigate the hemodynamics of the limbal microvascular network under changing IOP levels. We believe that vis-OCT has the potential to shed new light on investigating the conventional outflow pathway in both healthy and glaucomatous eyes.\n\nSupplementary Material\n======================\n\n###### Supplement 1\n\n###### Supplement 2\n\n###### Supplement 3\n\nThe authors thank Pieter Norden and Benjamin Thomson for helpful discussions.\n\nSupported in part by National Institutes of Health grants R01EY026078, DP3DK108248, R01EY029121, R01EY028304, R44EY026466, P30 EY016665, and T32EY25202; a Cornew Innovation Award from the Chemistry of Life Processes Institute at Northwestern University; the National Science Foundation Graduate Research Fellowship 1000260620 (LB); and RPB Stein Innovation Award, an award from RPB to the Department of Ophthalmology, University of Wisconsin-Madison (NS).\n\nDisclosure: **X. Zhang**, None; **L. Beckmann**, None; **D.A. Miller**, None; **G.** **Shao**, None; **Z.** **Cai**, None; **C.** **Sun**, Opticent, Inc. (F); **N.** **Sheibani**, None; **X.** **Liu**, None; **J.** **Schuman**, Opticent, Inc. (F); **M.** **Johnson**, None; **T.** **Kume**, None; **H.F. Zhang**, Opticent, Inc. (F)\n\n[[**Supplementary Video S1**]{.smallcaps}](#IOVS-61-2-23_s001){ref-type=\"supplementary-material\"}.\u2003 Video showing the 3D visualization of the 1mm length of segmented SC from [Figure 3](#fig3){ref-type=\"fig\"}b.\n\n[[**Supplementary Video S2**]{.smallcaps}](#IOVS-61-2-23_s002){ref-type=\"supplementary-material\"}.\u2003 Video showing the 3D visualization of the montaged vis-OCTA circumlimbal image of SC under elevated EVP and the surrounding limbal microvascular network from [Figure 6](#fig6){ref-type=\"fig\"}a.\n\n[^1]: XZ and LB contributed equally to this work.\n"}
+{"text": "Introduction\n============\n\nThe incidence of colorectal carcinoma in the United States is among the highest in the world, affecting \\~52/100,000 individuals, and the incidence of colorectal cancer in India is among the lowest, affecting \\~7/100,000 individuals, suggesting that lifestyle factors may contribute to the development of the disease ([@b1-mmr-14-05-4559]). Previous epidemiological and dietary intervention studies have suggested that diet-derived flavonoids may have a beneficial contribution to cancer therapy, primarily due to their pro-apoptotic or anti-angiogenic activities ([@b2-mmr-14-05-4559]--[@b4-mmr-14-05-4559]). Quercetin (also termed 3,3\u2032, 4\u2032, 5,7-pentahydroxyflavone) is a ubiquitous flavonoid found in various fruits, vegetables, nuts and red wine. Its antitumor effects have been confirmed in various cancer cells, including leukemia, breast, ovarian, colon, cervical, prostate and lymphoma ([@b5-mmr-14-05-4559]--[@b8-mmr-14-05-4559]). Different molecular mechanisms underlying the antitumor activity of quercetin have been identified, including upregulation of cell cycle inhibitors, downregulation of oncogene expression and the inhibition of glycolysis ([@b5-mmr-14-05-4559],[@b9-mmr-14-05-4559]--[@b12-mmr-14-05-4559]). However, the precise target for quercetin and its mechanisms of action remain to be elucidated.\n\nThe constitutive photomorphogenesis 9 (COP9) signalosome (CSN), is an evolutionarily conserved multiprotein complex that is present in all eukaryotes. It consists of eight subunits termed CSN1-CSN8 ([@b13-mmr-14-05-4559]). Previous studies have identified that CSN6 of the COP9 complex is crucial for proteasome-mediated protein degradation, as it regulates E3 ligases, including MDM2 proto-oncogene and COP1 ([@b14-mmr-14-05-4559],[@b15-mmr-14-05-4559]). Notably, CSN6 overexpression has been identified in various types of cancer, including glioblastoma, breast cancer, myeloma and leukemia ([@b16-mmr-14-05-4559]). It has been determined that the CSN6-MDM2-p53 signaling axis is important for cell proliferation and has antiapoptotic effects ([@b14-mmr-14-05-4559]). Previous studies have demonstrated that CSN6 prevents MDM2 autoubiquitination at lysine 364, which results in the stabilization of MDM2 and the degradation of p53 ([@b14-mmr-14-05-4559],[@b17-mmr-14-05-4559],[@b18-mmr-14-05-4559]). A previous study revealed that the HER2-Akt signaling axis is associated with CSN6 regulation and that Akt acts as a positive regulator of CSN6 ([@b19-mmr-14-05-4559]). A recent study determined that CSN6 promotes carcinogenesis by positively regulating v-myc avian myelocytomatosis viral oncogene homolog (Myc) stability. Additionally, CSN6 overexpression was positively correlated with Myc protein expression. Additionally, the gene expression signature of Myc target genes have been identified in human breast and pancreatic cancer ([@b20-mmr-14-05-4559]). A previous study also determined that CSN6 overexpression may be is as high as 40% in colon adenocarcinoma ([@b20-mmr-14-05-4559]). The overexpression of Myc may be 70--80% in colorectal cancer ([@b21-mmr-14-05-4559]); however, the function of the Akt-CSN6-Myc signaling axis remains to be elucidated in colorectal cancer.\n\nCSN6 has been identified as a potential novel therapeutic agent in cancer treatment and has thus been widely previously investigated ([@b16-mmr-14-05-4559]). The present study determined that quercetin may reduce the protein expression levels of CSN6 in HT-29 colon cancer cells and that it may be one of the important targets for quercetin-induced apoptosis in HT-29 cells. The present study determined that quercetin may reduce cell viability and induce apoptosis of HT-29 cells by mediating the phosphorylation of Akt and increasing CSN6 protein degradation, which also affected the expression levels of Myc, p53, B-cell lymphoma 2 (Bcl-2) and Bcl-2 associated X protein (Bax), indicating that quercetin-induced apoptosis of HT-29 cells may involve the Akt-CSN6-Myc signaling axis.\n\nMaterials and methods\n=====================\n\n### Chemicals, reagents and growth media\n\nRPMI-1640 and 10% fetal bovine serum (FBS) were purchased from Gibco; Thermo Fisher Scientific, Inc. (Waltham, MA, USA). Antibodies against p53 (cat. no. 9282), Bax (cat. no. 2772), Bcl-2 (cat. no. 2872), caspase-3 (cat. no. 9665) and \u03b2-actin (cat. no. 4970s) were purchased from Cell Signaling Technology, Inc. (Beverly, MA, USA). The antibody against Myc (cat. no. NB600-336) was purchased from Novus Biologicals (Littleton, CO, USA) and antibody against CSN6 (cat. no. LS-C174568) was acquired from LifeSpan BioSciences Inc. (Seattle, WA, USA). The enhanced chemiluminescence kits used for the visualization of the proteins were purchased from GE Healthcare Life Sciences (Chalfont, UK). Quercetin, DMSO and MTT were purchased from Sigma-Aldrich; Merck Millipore (Darmstadt, Germany). All other reagents were purchased from Beyotime Institute of Biotechnology, Inc. (Jiangsu, China). Quercetin was dissolved in DMSO to a concentration of 100 \u00b5M. Further dilutions were performed in cell culture media.\n\n### Cell line and culture\n\nThe HT-29 human colorectal cancer cell line was purchased from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). HT-29 cells were cultured in RPMI-1640 containing 10% FBS and were maintained at 37\u00b0C in a humidified incubator in an atmosphere of 5% CO2.\n\n### Cell viability assay\n\nCell viability was determined using an MTT assay. HT-29 cells were cultured until the log-phase and were subsequently seeded into a 96-well plate at a density of 1.0\u00d7104 cells/well overnight prior to treatment with different concentrations of quercetin (12.5, 25, 50, 100 and 200 \u00b5M) or DMSO. Following an incubation of 24, 48 or 72 h, the cells were then incubated with medium containing MTT for 4 h and the formazan crystals were dissolved with 150 \u00b5l DMSO. The plates were incubated on a shaker for 15 min at room temperature. The absorbance was measured at 490 nm using a microplate reader. The drug dose at which the cell viability was reduced by 50% (IC50) at 48 h of treatment was quantified. The experiments were repeated in triplicate.\n\n### Ultrastructures observed by transmission electron microscopy (TEM)\n\nFollowing treatment with quercetin, the cells were washed with PBS, collected by centrifugation (1,500 \u00d7 *g*, 4\u00b0C, 5 min) and fixed in 2.5% electron microscopy-grade glutaraldehyde. Next, they were rinsed with 0.1 M PBS, fixed in 1% osmium tetroxide, dehydrated through a graded series of ethanol and processed for Epon epoxy embedding. Ultra-thin sections (60 nm) stained with uranyl acetate and lead citrate and were observed using a JEM-1230 electron microscope.\n\n### Apoptosis and cell cycle analysis by flow cytometry\n\nControl and quercetin-treated cells were collected, washed twice with ice-cold PBS and resuspended in 500 \u00b5l binding buffer (Nanjing KeyGen Biotech Co., Ltd., Nanjing, China). Next, 5 \u00b5l annexin V-fluorescein isothiocyanate (FITC) and 5 \u00b5l propidium iodide (PI) were added and the cells were incubated for 15 min at room temperature in the dark. FITC and PI staining was analyzed to determine the apoptotic rate. The percentage of total apoptotic cells was calculated by adding the percentages of early apoptotic gated cells (annexin-V^+^) and late apoptotic gated cells (annexin-V^+^/PI^+^)\n\nFor cell cycle analysis, the cells were fixed in 70% ethanol at 4\u00b0C for a minimum of 4 h and washed twice with ice-cold PBS. Subsequently, 100 \u00b5l RNase A was added and the cells were incubated in a 37\u00b0C water bath for 30 min. Following incubation, the cells were stained with 400 \u00b5l PI for 30 min in dark conditions. The assays were performed in triplicate using a FACSort flow cytometer and quantified using BD CellQuest\u2122 Pro software (BD Biosciences, Franklin Lakes, NJ, USA).\n\n### Western blot analysis\n\nThe protein expression levels of Akt, phosphorylated (p-)Akt, CSN6, Myc, Bax, Bcl-2, caspase-3 cleaved caspase-3 and p53 in HT-29 cells were determined using western blotting. Briefly, a cell lysis solution was prepared using an extraction reagents kit (Fermentas; Thermo Fisher Scientific, Inc.), according to the manufacturer\\'s protocol. A 50 \u00b5g sample of protein was separated by 10% SDS-PAGE and was transferred onto nitrocellulose membranes (Merck Millipore). The membranes were blocked with 5% not-fat dry milk for 2 h at room temperature and were then incubated with the appropriate primary antibodies in a shaker overnight at 4\u00b0C. Subsequently, the membranes were washed 3 times at room temperature with washing buffer (1X TBS T: 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20) for 10 min and then incubated with secondary antibodies (1:1,000; cat. no. G130321; Hangzhou HuaAn Biotechnology Co., Ltd., Hangzhou, China) for 2 h at room temperature. \u03b2-actin was used as a loading control. Enhanced chemiluminescence was used to visualize the proteins with SuperSignal West Pico Chemiluminescent substrate (Thermo Fisher Scientific, Inc.) on a Molecular Imager ChemiDoc XRS system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Densitometry was performed using 170--9600 Quantity One^\u00ae^ 1-D software (Bio-Rad Laboratories, Inc.).\n\n### Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)\n\nFollowing treatment with quercetin, the total RNA was extracted from the cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) under RNase-free conditions and reverse transcription was performed in a 20 \u00b5l reaction with 200 ng total RNA using a two-step reverse-transcription reaction kit (Takara Biotechnology Co., Ltd., Dalian, China). The RT-qPCR was performed on an Applied Biosystems 7500 Real-time PCR system using a SYBR Premix Ex Taq kit (Takara Biotechnology Co., Ltd.) in Axygen 96-well reaction plates.\n\nThe primers used in the present study were obtained from Sangon Biotech Co., Ltd. (Shanghai, China) and their sequences were as follows: CSN6 (NM_486571), forward (F) 5\u2032-AGAGGCCACAATGCTGTTTG-3\u2032 and reverse (R) 5\u2032-CGTGGTCTACACCAATGCGTT-3\u2032; GAPDH (NM_002046), F: 5\u2032-TGGCACCCAGCACAATGAA-3\u2032 and R: 5\u2032-CTAAGTCATAGTCCGCCTAGA-3\u2032. GAPDH was used as a housekeeping gene and internal control. The data was analyzed using the 2^\u2212\u2206\u2206Cq^ method ([@b22-mmr-14-05-4559]).\n\n### Retroviral constructs and transfection\n\nThe complete codon sequence of CSN6 (NM_474971) was amplified using Platinum Taq DNA Polymerase high fidelity (Invitrogen; Thermo Fisher Scientific, Inc.) and the following primers: F 5\u2032-GACTCGAGATGGCGGCGGCGGCGGCGGCGGCTGCAGCTA-3\u2032 and R 5\u2032-GAGAATTCTCAGAAAAAGAGCCCGCGCATTCTCCTGCCGA-3\u2032. The PCR product was cloned into *Xho*I and *Eco*RI sites on the retroviral vector MSCV MIGR1 (provided by Professor Duonan Yu, University of Pennsylvania, Philadelphia, PA, USA). Sequence fidelity was confirmed using DNA sequencing by Sangon Biotech Co., Ltd. HT-29 cells were seeded into 6-well plates (1.0\u00d7105 cells/well) overnight and transfected with the recombinant retroviral expression plasmid using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer\\'s protocol. The cells were visualized under a fluorescence microscope (Nikon Corporation, Tokyo, Japan) to detect transfection efficiency and were then treated with quercetin for an additional 48 h.\n\n### Statistical analysis\n\nThe data are expressed as the mean \u00b1 standard deviation. Each experiment was repeated at least three times. Statistical comparisons of \\>2 groups were performed using a one-way analysis of variance, followed by a Bonferroni post-hoc test. All statistical analyses were performed using SPSS version 18.0 statistical software (SPSS, Inc., Chicago, IL, USA). P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\n### Quercetin reduces cell viability of HT-29 cells\n\nThe MTT assay revealed that HT-29 cell viability was decreased in a dose-dependent manner with increasing concentration of quercetin. Dose-dependent inhibition of cell viability was also observed in the HT-29 cells. The IC50 value for treatment for 48 h was determined as 81.65\u00b10.49 \u00b5M quercetin ([Fig. 1](#f1-mmr-14-05-4559){ref-type=\"fig\"}). Therefore, it was determined that quercetin exerted an negative activity against viability of colorectal cancer cells.\n\n### Quercetin induces apoptosis in HT-29 cells\n\nAs presented in [Fig. 2A](#f2-mmr-14-05-4559){ref-type=\"fig\"}, ultrastructures in HT-29 cells were observed by TEM 48 h after quercetin treatment. The cells in the control group had intact organelles, with normal nuclei and nucleolus chromatin. However, upon treatment with 50, 100 or 200 \u00b5M quercetin, cell shrinkage, chromatin condensation and nuclear collapse were observed ([Fig. 2A](#f2-mmr-14-05-4559){ref-type=\"fig\"}).\n\nFlow cytometry analysis revealed an increase in the apoptotic rate in treatment groups with higher quercetin concentration compared with the control group ([Fig. 2B and C](#f2-mmr-14-05-4559){ref-type=\"fig\"}).\n\nCleaved-caspase-3 is a key factor required for apoptosis and is the active form of pro-caspase-3. Immunoblotting analysis determined that cleaved-caspase-3 was significantly higher in the 50, 100 and 200 \u00b5M quercetin treatment groups compared with the control group (0 \u00b5M quercetin; P\\<0.01; [Fig. 2D and E](#f2-mmr-14-05-4559){ref-type=\"fig\"}).\n\nMembers of the Bcl-2 family are crucial for the regulation of apoptosis. Therefore, the present study used western blot analysis to determine the protein expression levels of Bax and Bcl-2. Bcl-2 expression decreased and Bax expression increased with increasing concentrations of quercetin compared with the control group ([Fig. 2D and E](#f2-mmr-14-05-4559){ref-type=\"fig\"}). The protein expression of p53 was significantly increased compared with the control group (P\\<0.01; [Fig. 2D](#f2-mmr-14-05-4559){ref-type=\"fig\"})\n\n### Effect of quercetin on the cycle progression of HT-29 cells\n\nIn order to determine whether the proliferation-inhibiting effect of quercetin on HT-29 cells was a result of cell-cycle arrest, cell-cycle analysis was performed using flow cytometry ([Fig. 3A](#f3-mmr-14-05-4559){ref-type=\"fig\"}). The proportion of cells in the G0/G1 phase of the cell cycle was significantly increased in the treatment groups exposed to quercetin compared with the control group (P\\<0.01; [Fig 3B](#f3-mmr-14-05-4559){ref-type=\"fig\"}). The number of cells in the S and G2/M phases in the quercetin treatment groups were significantly decreased (P\\<0.05; [Fig. 2B](#f2-mmr-14-05-4559){ref-type=\"fig\"}). Therefore, quercetin inhibited the proliferation of HT-29 cells via G0/G1 phase arrest ([Fig. 3A and B](#f3-mmr-14-05-4559){ref-type=\"fig\"}).\n\n### Akt-CSN6-Myc signaling axis mediates quercetin cytotoxicity\n\nTo determine the effect of CSN6 on quercetin-induced apoptosis of HT-29 cells, the recombinant retrovirus plasmid MIGR1-CSN6, containing the full-length human CSN6 gene, was constructed. HT-29 cells were transfected with MIGR1-CSN6 and an empty plasmid (MIGR1, used as a control), and were separately selected by flow cytometry for GFP+ cells. A transduction of \\~100% was achieved ([Fig. 4A and B](#f4-mmr-14-05-4559){ref-type=\"fig\"}). To determine whether the Akt-CSN6-Myc signaling axis was involved in quercetin-induced apoptosis of HT-29 cells, the levels of Akt, p-Akt, CSN6, Myc and p53 were examined using western blotting. The protein expression levels of p-Akt-Thr308, CSN6 and Myc were significantly reduced in the quercetin treatment groups compared with the control group (P\\<0.01; [Fig. 4C and D](#f4-mmr-14-05-4559){ref-type=\"fig\"}). Additionally, the mRNA expression of CSN6 was determined using RT-qPCR. No significant difference was identified in terms of CSN6 mRNA expression when quercetin treatment groups were compared with the control group ([Fig. 4E](#f4-mmr-14-05-4559){ref-type=\"fig\"}). The western blot analysis demonstrated that the protein expression levels of cleaved-caspase 3, p53 and Bax were also downregulated, while Myc and Bcl-2 were upregulated compared with cells treated with 50 \u00b5M quercetin alone and empty plasmid controls ([Fig. 4F and G](#f4-mmr-14-05-4559){ref-type=\"fig\"}). The MTT assay revealed that the overexpression of CSN6 reduced the effect of quercetin on cell viability compared with the empty plasmid MIGR1 ([Fig. 4H](#f4-mmr-14-05-4559){ref-type=\"fig\"}).\n\nDiscussion\n==========\n\nEpidemiological evidence has revealed that cancer incidence may be significantly modulated by an increased dietary intake of flavonoids through increased consumption of fruits and vegetables ([@b23-mmr-14-05-4559]). Flavonoids are one of the largest groups of naturally occurring phenols, including flavones, flavanols, isoflavones, flavonols, flavanones and flavanonols ([@b24-mmr-14-05-4559]). These dietary antioxidants have been identified to exert significant antitumor effects and have been extensively investigated ([@b4-mmr-14-05-4559],[@b25-mmr-14-05-4559],[@b26-mmr-14-05-4559]).\n\nThe present study investigated the effect of quercetin, which is one of the frequently researched flavonoids ([@b27-mmr-14-05-4559]), due to its positive effect on the growth inhibition and the induction of apoptosis in HT-29 cells. The anti-proliferative effects of quercetin were initially assessed following incubation with different concentrations of quercetin for 24, 48 and 72 h. The cell viability of HT-29 cells was significantly inhibited in a time-and dose-dependent manner. Following a 24 h incubation, the inhibitory effect of quercetin on HT-29 cell viability was not evident at low concentrations of quercetin. However, following a 48 h incubation, significant inhibition of cell growth was observed at 50, 100 and 200 \u00b5M quercetin ([Fig. 1](#f1-mmr-14-05-4559){ref-type=\"fig\"}).\n\nQuercetin induces pro-apoptotic signaling pathways, which lead to cell death ([@b5-mmr-14-05-4559],[@b8-mmr-14-05-4559],[@b9-mmr-14-05-4559]). In the present study, TEM observation of the quercetin-treated HT-29 cells revealed chromatin condensation, nuclear collapse and apoptotic body formation in cells treated with 50, 100 and 200 \u00b5M quercetin ([Fig. 2A](#f2-mmr-14-05-4559){ref-type=\"fig\"}). Apoptotic cell death was also quantified by determining the percentage of early apoptotic gated cells (Annexin-V^+^) and late apoptotic gated cells (Annexin-V^+^/PI^+^). The results revealed an increase in the percentage of apoptotic cells in quercetin treatment groups in a dose-dependent manner ([Fig. 2B and C](#f2-mmr-14-05-4559){ref-type=\"fig\"}). Caspase-3 is a key factor in apoptosis execution and cleaved caspase-3 is an activated form of caspase-3. Both the mitochondria-initiated intrinsic apoptotic pathway and the death receptor-triggered extrinsic apoptotic pathway may lead to caspase-3 activation. Therefore, cleaved caspase-3 protein expression levels were evaluated using western blotting, the result revealed that the expression levels of cleaved caspase-3 were significantly increased following quercetin treatment ([Fig. 2D and E](#f2-mmr-14-05-4559){ref-type=\"fig\"}). Bcl-2 and Bax have also been identified as key proteins for controlling the release of cytochrome *c* and other pro-apoptotic factors from the mitochondria, which leads to subsequent caspase activation and apoptotic cell death ([@b28-mmr-14-05-4559]). The present study determined that quercetin treatment significantly decreased the expression of Bcl-2, while increasing the expression of Bax ([Fig. 2D and E](#f2-mmr-14-05-4559){ref-type=\"fig\"}). This indicated that quercetin induced apoptosis via the regulation of the expression levels of Bcl-2 family proteins.\n\nIn order to determine whether this quercetin-induced inhibition of cell viability was due to cell cycle arrest, PI staining was performed and revealed that quercetin treatment significantly increased cell cycle arrest in the G0/G1 phase of the cell cycle, and that the number of cells in the S and the G2/M phase was reduced ([Fig. 3A and B](#f3-mmr-14-05-4559){ref-type=\"fig\"}). This result was consistent with the findings of Kim *et al* ([@b29-mmr-14-05-4559]). The immunoblot analysis revealed that the expression levels of p53 increased and those of Myc decreased following treatment with quercetin for 48 h ([Figs. 2D](#f2-mmr-14-05-4559){ref-type=\"fig\"} and [4C](#f4-mmr-14-05-4559){ref-type=\"fig\"}). The upregulation of p53 proteins led to an inhibition of growth and proliferation, involved with the G1 and G2/M phase arrest in cancer cells ([@b30-mmr-14-05-4559]--[@b32-mmr-14-05-4559]). A previous study reported that the downregulation of Myc-associated genes was involved in cell cycle arrest in acute myeloid leukemia ([@b33-mmr-14-05-4559]). The cell cycle arrest of nasopharyngeal carcinoma cells also involved the inhibition of the c-Myc signaling pathway ([@b34-mmr-14-05-4559]).\n\nAdditionally, quercetin has been considered a powerful modulator of several cellular signaling pathways, including the phosphatidylinositol-3-kinase (PI3K)-mediated signaling pathway, which important for quercetin-repressed tumors ([@b35-mmr-14-05-4559]). Akt is a downstream target of PI3K and regulates cell survival through the phosphorylation of downstream substrates that control apoptosis either directly or indirectly. Previous studies have revealed that oncogenic activation through Akt may act as an antiapoptotic signal via the rapid destabilization of p53 ([@b36-mmr-14-05-4559],[@b37-mmr-14-05-4559]). A previous study revealed that quercetin inhibited lymphoma by downregulating the PI3K-Akt-p53 signaling pathway ([@b38-mmr-14-05-4559]); however, the mechanism by which Akt regulates p53 remains to be elucidated. Previous studies determined that the MDM2-p53 signaling axis may be regulated by CSN6 ([@b14-mmr-14-05-4559],[@b39-mmr-14-05-4559]), and a subsequent study revealed that the HER2-Akt axis was associated with CSN6 regulation and that Akt is a positive regulator of CSN6 ([@b19-mmr-14-05-4559]). A recent study also demonstrated that CSN6 contributed to carcinogenesis by positive regulation of Myc stability ([@b20-mmr-14-05-4559]). The present study aimed to determine the importance of the Akt-CSN6-Myc signaling axis in quercetin-induced apoptosis of HT-29 cells. The immunoblot analysis revealed that the expression of p-Akt-Thr308 and CSN6 decreased in quercetin treatment groups. The expression of direct or indirect CSN6 target genes, including Myc and Bcl-2 decreased, whereas p53 and Bax increased in HT-29 cells treated with quercetin ([Figs. 2D](#f2-mmr-14-05-4559){ref-type=\"fig\"} and [4C](#f4-mmr-14-05-4559){ref-type=\"fig\"}). In order to determine the effect of CSN6 on quercetin-induced apoptosis, HT-29 cells were transfected with plasmid MIGR1-CSN6 or an empty MIGR1 plasmid and then treated with 50 \u00b5M quercetin for 48 h. The MTT assay revealed that the overexpression of CSN6 reduced the effect of quercetin on cell viability compared with the empty MIGR1 plasmid ([Fig. 4H](#f4-mmr-14-05-4559){ref-type=\"fig\"}). Additionally, the western blot analysis determined that the protein expression levels of cleaved-caspase 3, p53 and Bax were downregulated, whereas the levels of Myc and Bcl-2 were upregulated in the CSN6 overexpression group compared with the control group where cells were treated with quercetin and transfected with an empty plasmid ([Fig. 4F and G](#f4-mmr-14-05-4559){ref-type=\"fig\"}), indicating that quercetin-induced apoptosis involves the Akt-CSN6-Myc signaling axis in HT-29 cells.\n\nIn conclusion, the present study demonstrated that quercetin inhibited cell viability, induced apoptosis and led to cell-cycle arrest in HT-29 cells. The protein expression levels of p-Akt-Thr308 and CSN6 were significantly downregulated following quercetin treatment. Additionally, the expression levels of genes downstream of CSN6, including Myc and Bcl-2 were reduced and the levels of p53 and Bax were increased following treatment with quercetin. The overexpression of CSN6; however, reduced the effect of quercetin treatment on HT-29 cells. Therefore, it is possible that the Akt-CSN6-Myc signaling axis may be a potential target for novel treatment strategies of colorectal cancer.\n\nThe present study was supported by grants from the National Natural Science Foundation of China (grant nos. 81403232, 81274141 and 81450051), the Natural Science Foundation of Jiangsu Province of China (grant nos. BK2012686 and SBK2014021480) and the Bureau of Traditional Chinese Medicine, Science and Technology Project of Jiangsu Province (grant no. YB2015183).\n\n![Cell viability of HT-29 cells following treatment with different concentrations of quercetin for varying durations. The data are presented as the mean \u00b1 standard deviation (n=3; ^\\#^P\\<0.05 and ^\\#\\#^P\\<0.01 vs. 0 \u00b5M quercetin treated group for 24, 48 and 72 h).](MMR-14-05-4559-g00){#f1-mmr-14-05-4559}\n\n![Quercetin-triggered apoptosis in HT-29 cells following 48 h treatment. (A) Transmission electron microscopy images of cellular structures; black arrows show chromatin condensation and nuclear collapse in the different treatment groups. (B) Flow cytometry analysis revealed that quercetin induced apoptosis in a dose-dependent manner. (C) The number apoptotic cells were quantified following flow cytometry. (D) The protein expression levels of Bax, Bcl-2, p53 and c-caspase 3 in HT-29 cells exposed to different concentrations of quercetin were determined by western blot analysis. (E) Quantification of western blotting was normalized against the expression of \u03b2-actin and is expressed as the mean \u00b1 standard deviation (n=3; \\*\\*P\\<0.01 vs. 0 \u00b5M quercetin group). Bcl-2, B-cell leukemia/lymphoma 2; Bax, Bcl-2 associated-X protein; c-, cleaved-; FITC, fluorescein isothiocyanate; PI, propidium iodide.](MMR-14-05-4559-g01){#f2-mmr-14-05-4559}\n\n![Quercetin altered the cell cycle of HT-29 cells. (A) HT-29 cells were treated with different concentrations of quercetin for 48 h and the cell cycle distribution was examined using propidium iodide staining and flow cytometry. (B) Quantification of the cells in G0/G1, S and G2/M phases. The data are expressed as the mean \u00b1 standard deviation (n=3; \\*P\\<0.05, \\*\\*P\\<0.01 vs. 0 \u00b5M quercetin group).](MMR-14-05-4559-g02){#f3-mmr-14-05-4559}\n\n![Quercetin-induced apoptosis involves the Akt-CSN6-Myc signaling pathway in HT-29 cells. (A) Overexpression of CSN6 was performed by transfection with the recombinant retroviral expression plasmid and the cells were selected by flow cytometry for GFP+ cells; \\~100% transduction efficiency was achieved. (B) Expression levels of CSN6 were analyzed by western blotting in HT-29 cells transfected with control (empty) or CSN6 vectors. The data are expressed as the mean \u00b1 standard deviation (n=3; \\*P\\<0.05 vs. control). (C and D) Western blot analysis of the protein expression levels of p-Akt, CSN6 and Myc in HT-29 cells treated with different concentrations of quercetin for 48 h. The data are expressed as the mean \u00b1 standard deviation (n=3; \\*P\\<0.05, \\*\\*P\\<0.01 vs. control). (E) The mRNA expression levels of CSN6 were examined by reverse transcription-quantitative polymerase chain reaction using total mRNA of HT-29 cells and compared with the control. (F and G) Western blotting was used to determine protein expression levels of CSN6, Myc, Bax, Bcl-2, p53 and caspase 3 in the CSN6 overexpression cell line and control (empty) cell line, which remained untreated or exposed to 50 \u00b5M quercetin for 48 h. The data are expressed as the mean \u00b1 standard deviation (n=3; \\*P\\<0.05, \\*\\*P\\<0.01 vs. untreated control). (H) Cell viability was examined in a cell line overexpressing CSN6 and control (empty) cell line treated with or without 50 \u00b5M quercetin for 48 h, \\*P\\<0.05 and \\*\\*P\\<0.01 vs. untreated control. MIGR1, plasmid; CSN6, COP9 signalosome subunit 6; Akt, Akt serine/threonine kinase 1; p-Akt, phosphorylated-Akt; Myc, v-myc avian myelocytomatosis viral oncogene homolog; Bcl2, B cell leukemia/lymphoma 2; Bax, Bcl2 associated X; c-caspase 3, cleaved-caspase 3.](MMR-14-05-4559-g03){#f4-mmr-14-05-4559}\n"}
+{"text": "1. Introduction\n===============\n\nWine production is of great importance in agro economic activities. The world grape production in 2012 exceeded 69 million tons and Europe was the largest producer of wine, with 66% of the total world production \\[[@B1-molecules-20-09686]\\]. The solid wastes generated by the wine industry represents between 25%--30% of the material used and it consists mainly of grape pomace (containing seeds, pulp, stem and skin) \\[[@B2-molecules-20-09686],[@B3-molecules-20-09686]\\]. It is well known that high quantities of valuable compounds like dietary fiber, oils from the seeds, anthocyanins and phenolics compounds still remain within the grape pomace after processing \\[[@B3-molecules-20-09686],[@B4-molecules-20-09686]\\]. The phenolics, such as resveratrol, have great potential due to their antioxidant capacity and health benefits against coronary diseases by the inhibition of LDL (low-density lipoproteins) and other chronic diseases, like cancer, diabetes and neurodegenerative disorders \\[[@B3-molecules-20-09686],[@B4-molecules-20-09686]\\].\n\nIn addition, from the economic point of view, the market of these compounds have been increased in the recent years by the increasing consumer demand for the use of more natural antioxidant compounds, achieving the value of US\\$30 billions, based on 2008 grape wine production data \\[[@B3-molecules-20-09686]\\]. In this sense, the valorization and reuse of these wastes from the wine-making industry would have a significant environmental and economic impact, and this possibility has been studied by several authors \\[[@B3-molecules-20-09686],[@B4-molecules-20-09686],[@B5-molecules-20-09686],[@B6-molecules-20-09686],[@B7-molecules-20-09686]\\].\n\nIn recent years, numerous methodologies for the extraction of compounds of relatively high polarity have been developed in an effort to displace conventional solvent extraction techniques. These novel alternative techniques significantly reduce solvent consumption and increase the speed of the extraction by simplifying the process.\n\nSupercritical fluid extraction (SFE) is an efficient technique that is widely applied in the separation of active compounds from herbs and other plants \\[[@B8-molecules-20-09686]\\]. This technique is appreciated due to the very high solvent power and the distinctive physicochemical properties of supercritical fluids (SCFs). The relatively low viscosity (near to the gas) and the high diffusivity of SCFs help to penetrate the porous solid materials more efficiently than liquid solvents, thus resulting in faster and more efficient extractions. For example, conventional solid-liquid extractions lasting several hours or even days can be achieved in ten minutes on using supercritical fluids \\[[@B9-molecules-20-09686]\\].\n\nThe first scientific work on SFE from grape residues concerned the use of carbon dioxide modified with 5% methanol at 350 bar and 50 \u00b0C for the extraction of phenolic compounds \\[[@B10-molecules-20-09686]\\]. The recovery of resveratrol from grape skin was also optimized by using CO~2~ + ethanol as the solvent system \\[[@B11-molecules-20-09686]\\]. Chafer *et al.* evaluated the SFE of polyphenols from five grape skin varieties and reported that the most suitable conditions were 60 \u00b0C, 250 bar and 20% ethanol as a CO~2~ modifier \\[[@B6-molecules-20-09686]\\].\n\nIt has been shown in different studies that SFE is selective for phenolics, such as gallic acid, catechin, epicatechin and quercetin and have showed high recovery of this polyphenols from grape pomace \\[[@B6-molecules-20-09686],[@B12-molecules-20-09686],[@B13-molecules-20-09686]\\]. However, this technique cannot be used to extract high molecular weight polyphenols, such as proanthocyanidins, which were more easily extracted by conventional extraction \\[[@B12-molecules-20-09686]\\].\n\nPLE is based on the use of conventional liquid solvents at subcritical conditions with controlled temperature and pressure. With respect to conventional extraction techniques, PLE has the advantage of using less solvent and the extraction is carried out in a shorter time. PLE is widely used for the extraction of antioxidant compounds from winery residues and other natural products \\[[@B14-molecules-20-09686],[@B15-molecules-20-09686],[@B16-molecules-20-09686],[@B17-molecules-20-09686]\\]. Pi\u00f1eiro *et al.* compared catechin and epicatechin extraction from tea leaves and grape seeds using ultrasound-assisted extraction (UAE) and PLE \\[[@B15-molecules-20-09686]\\]. In addition, different solvents, such as water, ethanol and methanol, have been evaluated as hot pressurized solvents for the extraction of anthocyanins and phenolic compounds from grape skin \\[[@B16-molecules-20-09686],[@B17-molecules-20-09686],[@B18-molecules-20-09686]\\].\n\nSeveral other methods have been applied for extraction of antioxidants from plant matrices, one such novel process being microwave-assisted extraction (MAE). The advantage of PLE and SFE over MAE is the applicability at different scales. PLE and SFE can be applied to systems on various scales, from the laboratory scale (a few grams) to the pilot plant scale (several hundred grams of samples), through to the industrial scale (tons of raw material) \\[[@B19-molecules-20-09686],[@B20-molecules-20-09686]\\]. In addition, for winery residues, comparative studies have shown that PLE was more efficient than conventional solvent extraction, MAE and UAE for the recovery of high levels of phenolics from grape pomace and grape skin \\[[@B5-molecules-20-09686],[@B16-molecules-20-09686]\\]. On the other hand, several studies have demonstrated the economic viability of SFE and PLE for the extraction of winery residues \\[[@B19-molecules-20-09686]\\] and other raw materials \\[[@B20-molecules-20-09686]\\].\n\nA great variety of phenolic compounds have been extracted under superheated pressurized conditions from white and red grape skins. Phenolic acids (caffeic acid, gallic acid and protocatechuic acid) and flavonols (catechin, epicatechin and gallate derivatives) were detected, but pyroanthocyanin was also tentatively identified \\[[@B5-molecules-20-09686]\\]. However, there has been a marked increase in the number of crops for the production of red wines in southern Europe. Many different varieties are grown and these include Tempranillo, Syrah, Cabernet Sauvignon and others. The by-products have already been used for the production of antioxidants using high pressure techniques. Campos *et al.* \\[[@B21-molecules-20-09686]\\], T\u00fcnde *et al.* \\[[@B22-molecules-20-09686]\\], and previous studies by Mantell *et al.* \\[[@B23-molecules-20-09686]\\] and Casas *et al.* \\[[@B24-molecules-20-09686]\\] are some examples.\n\nAs have been aforementioned, several studies have showed that both techniques, SFE \\[[@B6-molecules-20-09686],[@B12-molecules-20-09686],[@B13-molecules-20-09686]\\] and PLE \\[[@B14-molecules-20-09686],[@B15-molecules-20-09686],[@B16-molecules-20-09686],[@B17-molecules-20-09686],[@B18-molecules-20-09686]\\], are successfully used to recovery phenolic compounds from grape pomaces, but there are no comparative studies of these techniques to evaluated the efficiency to extract anthocyanins and phenolics from this raw material. Therefore, the aim of the work described here was to evaluate the effect of different experimental parameters, such as pressure, temperature and extraction solvent on the SFE and PLE from different varieties of red grape pomace (Petit Verdot, Tintilla, Syrah, Cabernet Sauvignon, Merlot and Tempranillo). The extracts were analyzed according to the global extraction yield, total phenolic content, total anthocyanin content and antioxidant activity. In addition, the effects of the extraction time and flow rate for PLE were evaluated at the best extraction conditions.\n\n2. Results and Discussion\n=========================\n\n2.1. Variety Selection\n----------------------\n\nIn order to analyze the influence of red grape varieties for the preparation of a product with a high antioxidant capacity, a series of experiments were designed in which the raw material was varied. Two extraction methods, namely Supercritical Fluid Extraction (SFE) with carbon dioxide with 20% of ethanol and PLE with ethanol, were tested and both the extraction yield and antioxidant capacity were analyzed. For SFE, several authors have reported that 20% of co-solvents is efficient enough to the high extraction yields of anthocyanins and polyphenols from red grape pomace and other raw materials \\[[@B6-molecules-20-09686],[@B23-molecules-20-09686]\\]. Raising the percent of ethanol as modifier to CO~2~ up to 15%--20% have not showed a significant increase in the yield or the phenolic extraction from grape skins \\[[@B6-molecules-20-09686],[@B21-molecules-20-09686]\\].\n\nThe results obtained in the SFE at 100 bar, 55 \u00b0C and flow rate of 25 g/min are shown in [Figure 1](#molecules-20-09686-f001){ref-type=\"fig\"} and [Table 1](#molecules-20-09686-t001){ref-type=\"table\"}.\n\n![Extraction yield and antioxidant activity obtained by supercritical fluid extraction (SFE) with CO~2~ + 20% ethanol as co-solvent (100 bar, 55 \u00b0C, 20 g/min CO~2~, 5 g/min ethanol, 3 h) from different grape pomace varieties.](molecules-20-09686-g001){#molecules-20-09686-f001}\n\nmolecules-20-09686-t001_Table 1\n\n###### \n\nExtraction yields of anthocyanins and phenolic compounds expressed as mg/g dry grape pomace.\n\n  Variety        SFE CO~2~ + 20% EtOH ^a^   PLE Ethanol ^b^                \n  -------------- -------------------------- ----------------- ------------ ------------\n  Petit Verdot   0.3 \u00b1 0.1                  2.5 \u00b1 0.1         16.0 \u00b1 1.0   28.9 \u00b1 1.3\n  Tintilla       3.8 \u00b1 0.1                  4.5 \u00b1 0.1         49.7 \u00b1 2.8   15.5 \u00b1 0.2\n  Syrah          3.2 \u00b1 0.3                  3.6 \u00b1 0.2         38.3 \u00b1 0.6   24.4 \u00b1 0.2\n  Cabernet       0.1 \u00b1 0.1                  2.3 \u00b1 0.3         11.1 \u00b1 1.2   23.8 \u00b1 0.1\n  Merlot         0.2 \u00b1 0.1                  2.1 \u00b1 0.1         10.1 \u00b1 0.1   22.4 \u00b1 0.1\n  Tempranillo    2.0 \u00b1 0.2                  2.2 \u00b1 0.1         30.9 \u00b1 1.0   19.6 \u00b1 0.1\n\n^a^ SFE at 100 bar and 55 \u00b0C for 3 h; ^b^ PLE at 120 bar and 100 \u00b0C for 3 h; ^c^ Yiel of anthocyanins expressed as mg malvin chloride/g dry grape pomace; ^d^ Yield of phenolics expressed as mg gallic acid equivalent/g dry grape pomace.\n\nIt can be seen from this figure that the best extraction yields were obtained with Petit Verdot, followed by Syrah and Merlot varieties. The lowest yields were obtained with Tintilla, Tempranillo and Cabernet Sauvignon, with values of just over half those of the aforementioned varieties. In contrast to the abovementioned, [Table 1](#molecules-20-09686-t001){ref-type=\"table\"} shows the SFE from Tintilla and Syrah presented higher recovery of anthocyanins and phenolics than the other grape varieties. The addition of higher concentrations of ethanol as CO~2~ modifier (20%) favored the extraction yield of phenolic compounds in comparison with previous studies using only 4.5% of ethanol (0.2--0.3 mg gallic acid/g dry grape skin) \\[[@B25-molecules-20-09686]\\].\n\nThe antioxidant activities of the extracts were analyzed and lower EC~50~ values (associated with a higher antioxidant activity) were obtained for the varieties Petit Verdot followed by Syrah ([Figure 1](#molecules-20-09686-f001){ref-type=\"fig\"}). In this case, the antioxidant capacity of the Cabernet Sauvignon samples was the lowest. Comparison of the antioxidant capacities of the extracts obtained in this work with data published previously by other authors shows that the antioxidant capacity of the extracts obtained in this work are lower and are still far from that obtained for the standard (+)-\u03b1-tocopherol (EC~50~ = 6.17 ug/mL) \\[[@B26-molecules-20-09686]\\]. As a consequence, PLE was explored, as it is considered to be an efficient technique for the extraction of polar or slightly polar compounds from different natural materials \\[[@B16-molecules-20-09686],[@B26-molecules-20-09686],[@B27-molecules-20-09686]\\].The results obtained in PLE experiments using ethanol for the different varieties of grapes analyzed are presented in [Figure 2](#molecules-20-09686-f002){ref-type=\"fig\"}. Once again, the highest global yields were obtained when Petit Verdot was used, followed by Syrah and Tempranillo. In this case, the lowest global yield was obtained when the extraction was carried out on the Tintilla variety. However, according to the anthocyanins and phenolic recovery, Tintilla, Syrah and Tempranillo showed higher yields of anthocyanins whereas Petit Verdot, Syrah and Cabernet presented the higest phenolic yields ([Table 1](#molecules-20-09686-t001){ref-type=\"table\"}).\n\n![Extraction yield and antioxidant activity obtained by pressurized liquid extraction (PLE) whit ethanol as extraction solvent (120 bar, 100 \u00b0C, 10 g/min ethanol, 3 h) from different grape pomace varieties.](molecules-20-09686-g002){#molecules-20-09686-f002}\n\nAs far as the antioxidant capacity is concerned, the highest antioxidant capacities were obtained in the extraction on the Petit Verdot variety. The results obtained with Merlot, Tempranillo, Cabernet Sauvignon and Syrah were very similar and presented values of EC~50~ close to (+)-\u03b1-tocopherol between 7.5 and 9.2 \u03bcg/mL. Nevertheless, there are significant differences with the Tintilla grape, which gave the lowest results (higher EC~50~).\n\nComparison of the yields and antioxidant capacities of the extracts obtained using SFE with carbon dioxide and 20% of ethanol as co-solvent and PLE with ethanol shows that the best results were achieved when the extraction was conducted with PLE. In some cases, double the global yield was obtained in comparison to the SFE method and around 10 fold was increased the yield of anthocyanins and phenolic compounds. The yield of anthocyanins and phenolics was in agreement as reported previously for different grape pomaces' varieties; however, it depends on the origin of the varieties, the genetic variation and the cultivation conditions \\[[@B7-molecules-20-09686],[@B16-molecules-20-09686]\\]. Similarly, the antioxidant activity results are significantly higher than those shown in [Figure 1](#molecules-20-09686-f001){ref-type=\"fig\"} and they are comparable to those described by other authors, such asBozan *et al.* \\[[@B28-molecules-20-09686]\\], who reported EC~50~ values of approximately 3 \u00b5g/mL for Merlot and Cabernet samples, and Tounsi *et al.* \\[[@B29-molecules-20-09686]\\], who reported values of 6.8 \u00b5g/mL for Carignan and 30 \u00b5g/mL for Syrah seeds, amongst others.\n\n[Table 2](#molecules-20-09686-t002){ref-type=\"table\"} shows the concentrarion of anthocyanins and phenolic in the extracts obtained by SFE and PLE from the different varieties of grape pomace. In all the cases, the contents of anthocyanins and polyphenols of PLE extracts were higher than those reported for the extracts obtained by SFE. The PLE extracts obtained from Merlot gave the highest concentration of phenolics (254.6 \u00b1 2.6 mg GAE/g extract) followed by Tintilla, Carbernet and Petit Verdot, which also presented high enough phenolic content. However, Tintilla extracts obtained by PLE presented a significant concentration of anthocyanins that was superior to the others grape varieties.\n\nmolecules-20-09686-t002_Table 2\n\n###### \n\nConcentration of anthocyanins and phenolics in extracts obtained by SFE and PLE from different varieties of red grape pomace.\n\n  Variety        SFE CO~2~ + 20% EtOH ^a^   PLE Ethanol ^b^                  \n  -------------- -------------------------- ----------------- -------------- -------------\n  Petit Verdot   4.9 \u00b1 0.3                  34.5 \u00b1 0.4        113.8 \u00b1 7.1    204.9 \u00b1 9.4\n  Tintilla       116.1 \u00b1 2.0                135.7 \u00b1 0.8       741.9 \u00b1 41.7   231.6 \u00b1 3.5\n  Syrah          55.1 \u00b1 4.6                 62.7 \u00b1 0.7        292.5 \u00b1 4.3    186.3 \u00b1 1.3\n  Cabernet       3.5 \u00b1 0.1                  59.7 \u00b1 1.8        107.0 \u00b1 11.3   228.4 \u00b1 0.9\n  Merlot         4.5 \u00b1 0.2                  52.2 \u00b1 1.1        114.6 \u00b1 0.2    254.6 \u00b1 1.0\n  Tempranillo    56.8 \u00b1 1.3                 59.0 \u00b1 0.9        278.3 \u00b1 9.3    177.1 \u00b1 1.3\n\n^a^ SFE at 100 bar and 55 \u00b0C for 3 h; ^b^ PLE at 120 bar and 100 \u00b0C for 3 h; ^c^ Total anthocyanins content in extracts expressed as mg malvin chloride/g dry extract; ^d^ Total content of phenolics in extracts expressed as mg gallic acid equivalent/g dry extract.\n\nHowever, a direct relationship between the total phenolics and the antioxidant activity of the extracts was not observed. Different authors have also reported similar results considering that the antioxidant activity depends on the quality of phenolic compounds. Besides, the antioxidant capacity of the extracts could also be affected by the synergistic effects caused by the interactions of antioxidant compounds and the presence of other non-phenolic compounds \\[[@B30-molecules-20-09686],[@B31-molecules-20-09686]\\].\n\nTherefore, considering that Petit Verdot grape pomace showed the highest extraction yield and extracts with high antioxidant capacity as well as adequate total phenolic content, this variety and PLE as extraction technique were selected for the subsequent experiments aimed at identifying the best process operating conditions.\n\n2.2. Extraction Conditions\n--------------------------\n\n### 2.2.1. Solvent System\n\nOn the basis of the results obtained in the previous analysis, the highest extraction yield and the best antioxidant activity were accomplished with the same method (PLE) and with the same grape pomace variety (Petit Verdot).\n\nPLE is increasingly being used as an alternative to carry out environmentally friendly green extractions since it avoids the use of large amounts of solvents, which also provides significant advantages in process automation and sample preparation. In regard to the solvent employed in the extraction, ethanol (EtOH) has been studied as one of the more environmentally friendly solvents (\"green\" solvent) and it is recognized as safe according to the European Food Safety Authority (EFSA) and FAO/WHO Expert Committee on Food Additives \\[[@B32-molecules-20-09686],[@B33-molecules-20-09686]\\]. In an effort to reduce the consumption of organic solvents, reports by \u0160tav\u00edkov\u00e1 *et al.* \\[[@B15-molecules-20-09686]\\] and Aliakbarian *et al.* \\[[@B34-molecules-20-09686]\\] have shown that PLE is favorable for the extraction process on grape skin. Moreover, it has been reported that ethanol/water mixtures are environmentally favorable compared to pure alcohol \\[[@B35-molecules-20-09686]\\].\n\nIn order to select the most suitable solvent system in this work, a study was carried out to evaluate PLE procedures with different solvent composition, namely pure ethanol, ethanol/water (50:50) and water, and the results for the global yield and antioxidant activity are represented in [Figure 3](#molecules-20-09686-f003){ref-type=\"fig\"}.\n\nIt was found out that the system with the ethanol/water mixture gave the highest extraction yield and the extract with the highest antioxidant activity; for this reason, this solvent was chosen to optimize the extraction method. The performance of this system can be explained because with a dual mixture, particularly a mixture of an organic solvent and water, the extraction efficiency is improved because the organic solvent enhances the solubility of the analyte and water increases the analyte desorption \\[[@B35-molecules-20-09686]\\].\n\n![Extraction yields and antioxidant activities, obtained by PLE from Petit Verdot grape pomace using different solvent systems (120 bar, 100 \u00b0C, 10 g/min solvent, 3 h).](molecules-20-09686-g003){#molecules-20-09686-f003}\n\nFurthermore, the analysis of the total phenolic content for the hydroalcoholic mixture (497.32 \u00b1 4.93 mg GAE/mg extract) was superior to that obtained using pure ethanol (204.92 \u00b1 9.43 mg GAE/mg extract) and water (334.61 \u00b1 9.16 mg GAE/mg extract).\n\nOther studies have also shown that for PLE, an increase in the percentage of ethanol has a negative effect on the extraction from grape skin or grape pomace and hydroalcoholic mixtures are more favorable for anthocyanins and polyphenols extraction \\[[@B23-molecules-20-09686],[@B27-molecules-20-09686],[@B36-molecules-20-09686]\\].\n\n### 2.2.2. Influence of Extraction Parameters\n\nTemperature and pressure both play a significant role in the extraction process and, as a result, these parameters were studied in order to select the best operating conditions.\n\nThe effects of temperature were studied considering that high temperatures favor extraction efficiency \\[[@B37-molecules-20-09686]\\]; although this might cause degradation of thermo-labile compounds \\[[@B38-molecules-20-09686],[@B39-molecules-20-09686]\\]. Three extraction temperatures were tested: 80 \u00b0C, 100 \u00b0C and 120 \u00b0C. It can be observed in [Figure 4](#molecules-20-09686-f004){ref-type=\"fig\"} that higher temperatures gave higher yields and, therefore, the highest extraction yield was obtained at 120 \u00b0C, although the best antioxidant activity was achieved with the extract obtained at 100 \u00b0C.\n\nThe use of higher pressure led to an improvement in the extraction because the solvent makes contact with the analyte more easily. Three values were analyzed: 90 bar, 120 bar and 150 bar. It can be deduced that the effect of pressure depends on the temperature; hence at a fixed pressure, both extraction yield and antioxidant activity are determined by the temperature. Nevertheless, there are very few differences between the EC~50~ values, and extraction efficiency is, consequently, the decisive result.\n\n![Extraction yield and antioxidant activity obtained by PLE with ethanol/water (50:50) from Petit Verdot grape pomace under different extraction conditions at a flow rate of 10 g/min during 3 h of extraction.](molecules-20-09686-g004){#molecules-20-09686-f004}\n\nIt can be seen in [Figure 4](#molecules-20-09686-f004){ref-type=\"fig\"} that a considerably higher extraction yield is obtained at 120 \u00b0C and, at this temperature, the highest antioxidant activity is achieved at 120 bar.\n\nThese results are consistent with those of Hawthorne and Miller \\[[@B40-molecules-20-09686]\\], who first studied PLE with water and showed that temperature has a predominant effect over pressure.\n\nThe extraction yields and antioxidant activities obtained for different temperatures and pressures were statistically analyzed. Regression analysis was performed on the experimental data and the coefficients of the model were evaluated for significance. It was observed in the Pareto diagram ([Figure 5](#molecules-20-09686-f005){ref-type=\"fig\"}) that temperature is the only factor that had a significant influence on the extraction yield (*p* \u2264 0.05).\n\nThe relationship between temperature and pressure for the global extraction yield is represented by Equation (1). where Y represents global extraction yield; P the pressure (bar), and T the temperature (\u00b0C).\n\n![Pareto diagram (**a**) for the global extraction yield (**b**) for the antioxidant activity, obtained by PLE with ethanol/water (50:50) from Petit Verdot grape pomace.](molecules-20-09686-g005){#molecules-20-09686-f005}\n\n### 2.2.3. Extraction Kinetics\n\nThe flow rate is an important variable in the design of extraction processes. Excessive solvent flow rates lead to insufficient loading of the solvent and, in severe cases, compacting of the bed, which results in preferential pathways through the bed and causes inefficient extraction. On the other hand, low flow rates lead to unprofitable processes due to the long extractor residence times.\n\nThe global extraction yields obtained with PLE from Petit Verdot variety at 90 bar of pressure and 120 \u00b0C of temperature with the mixture ethanol/water (50:50) as solvent at flow rates of 5, 10, 15 and 20 g/min are plotted against extraction time and solvent mass in [Figure 6](#molecules-20-09686-f006){ref-type=\"fig\"}.\n\nIt is remarkable that, in the initial stage of the process, the extraction rate is dependent on the flow. At 15 min, a higher extraction yield was obtained in the process with a solvent flow rate of 20 g/min than that obtained at 5 g/min. This behavior is as one would expect; an increase in the flow rate leads to higher yields of extracts. This effect is due to the presence of a large amount of solvent in the operation, a factor that enhances the extraction yield. For extraction times above 50 min, an increase in the flow did not lead to an improvement in the extraction process, despite the fact that solvent consumption was high. In this second part, the slope of the curve decreased and the extraction rate was reduced until a limiting yield was reached, which depended on the total amount of extractable solutes under the fixed operating conditions. As a result, it is advisable to work at a flow rate of 5 g/min for 90 min. Under these conditions a yield of 87% was obtained after 180 min. This behavior is commonly seen using different raw materials \\[[@B41-molecules-20-09686]\\].\n\n![Kinetic extraction in terms of extractions time (**a**) and solvent mass (**b**) for PLE from Petit Verdot grape pomace at 90 bar of pressure and 120 \u00b0C of temperature with ethanol/water (50:50) at different flow rates](molecules-20-09686-g006){#molecules-20-09686-f006}\n\n3. Experimental Section\n=======================\n\n3.1. Samples and Chemicals\n--------------------------\n\nThe materials used in this study were the wastes from the vinification of red wine from different varieties. The varieties analyzed were Cabernet Sauvignon, Merlot, Petit Verdot, Syrah, Tempranillo and Tintilla. All of the raw materials were provided by \"Bodegas Luis P\u00e9rez\" (Jerez de la Frontera, Spain). The grape pomaces were obtained immediately after the vinification process and were dried in an oven at 60 \u00b0C to constant weight. Prior to the extraction process, the samples were milled in order to reduce the particle size.\n\nCarbon dioxide (99.995%) was supplied by Abello-Linde S.A. (Barcelona, Spain). 2,2-Diphenyl-1-picrylhydrazyl free radical (DPPH). Methanol (HPLC grade), ethanol (HPLC grade) and the other reagents were provided by Panreac (Barcelona, Spain). All experiments were carried out using Milli-Q water (Millipore S.A.S., Molsheim, France).\n\n3.2. Extraction Methods\n-----------------------\n\nHigh pressure extraction tests were carried out in a system supplied by Thar Technology (model SF100, Pittsburgh, PA, USA) provided with an extraction vessel (capacity of 100 mL) and two pumps with a maximum flow rate of 50 g/min. In SFE, two pumps were used (one for carbon dioxide and the other for co-solvent) while in PLE only one pump was employed. Pressure was adjusted at the back pressure regulator and solvent pumps. A thermostatized jacket allowed control of the extraction temperature. The cyclonic separator permitted periodic discharge of the extracted material during the process.\n\nThe operating methodology involved loading the extraction vessel with approximately 35 g of the sample, which had previously been homogenized in order to maintain a constant apparent density in all experiments. The extracts were collected in a cyclonic separator and transferred to glass bottles, which were stored at 4 \u00b0C with the exclusion of light.\n\nFor SFE, the experiments were carried out at a flow rate of 25 g/min for 3 h at 100 bar and 55 \u00b0C using CO~2~ + 20% ethanol as solvent. Previous studies have reported that raising the percent of ethanol as modifier to CO~2~ up to 15%--20% have not showed a significant increase in the yield or the phenolic extraction from grape skins \\[[@B6-molecules-20-09686],[@B21-molecules-20-09686]\\].\n\nFor PLE, all experiments were carried out at a flow rate of 10 g/min for 3 h. The influence of pressure (P), temperature (T) and ethanol concentration (C) on the extraction process was analyzed. Pressures of 90, 120 and 150 bar, temperatures of 80, 100 and 120 \u00b0C, and different concentrations of ethanol in water between 0% and 100% were evaluated. In the case of hydro-alcoholic mixutres, previos studies have showed that 50% ethanol was successful to extract anthocyanins from red grape pomace with similar yields than conventional extraction \\[[@B35-molecules-20-09686]\\]. The experiments on each sample were carried out in duplicate in order to evaluate the variability of the measurements.\n\nThe global extraction yields obtained by SFE and PLE were calculated as the ratio of dry extract to dry raw material and the results are expressed as g extract/g raw material.\n\n3.3. Total Phenolic Content\n---------------------------\n\nThe determination of total phenolic compounds present in the extracts was carried out using an HPLC system 1100 series supplied by Agilent Technologies (Waldbronn, Germany) comprising a degasser, a quaternary pump, an autosampler and a UV/vis detector.\n\nThe extracts were previously filtered and subsequently 20 \u00b5L aliquots were injected and analyzed by HPLC using a Synergi Hydro--RP C18 column (150 mm \u00d7 3 mm i.d., 4 \u03bcm particle size) (Phenomenex, Torrance, CA, USA) with a C18 ODS guard column (4.0 mm \u00d7 2.0 mm i.d.). The mobile phase was acidified water containing 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B). The gradient profile was as follows: 0 min, 0% B; 0.2 min, 0% B; 0.3 min, 7% B; 14.7 min, 8.5% B; 40 min, 19% B; 45 min, 33% B; 48 min, 50% B; 50 min, 95% B; 57 min, 0% B; and 63 min, 0% B. The flow rate was 1 mL/min and the wavelength of detection was set at 278 nm.\n\nTotal phenol content was determined as the sum of the peak area of all the phenolic compounds and expressed as mg of gallic acid equivalent (GAE)/mg dry extract based on a calibration curve with gallic acid. Analyses were carried out in triplicate and SD was estimated. Other authors have also used chromatographic methods as an accurate approach to obtain the total phenolic content \\[[@B31-molecules-20-09686],[@B42-molecules-20-09686]\\].\n\n3.4. Anthocyanins Analysis\n--------------------------\n\nThe analysis of the anthocyanins present in the extracts was performed in and HPLC system provided by Agilent Technologies (Palo Alto, CA, USA) 1100 Series chromatograph. 5% Formic acid in water (v/v) (A) and methanol (B) were used as solvents in this HPLC method. The HPLC gradient program was executed as follows: 98% A to 40% A in 60 min, 40% A to 98% A in 5 min. The entire HPLC run time was 70 min using a flow rate of 1 mL/min.\n\nThe resultant extracts were filtered before HPLC assay using a 0.45 \u00b5m PTFE filter (Varian Inc., Palo Alto, CA, USA) and 100 \u00b5L of the filtered extract was injected into the column (250 mm \u00d7 4.6 mm) C18 Hypersil ODS (5 \u00b5m particle size) (Supelco). The total content of anthocyanins was calculated as the sum of the peak area of all the compounds detected at a wavelength of 510 nm. The results were reported based on the calibration curve of malvin chloride and expressed as mg of malvin chloride/mg dry extract. The experiments for each extraction were carried out in triplicate in order to evaluate the variability of the measurements. The method above described has also been used by other authors \\[[@B43-molecules-20-09686],[@B44-molecules-20-09686]\\].\n\n3.5. Antioxidant Assay with DPPH\n--------------------------------\n\nThe method used to measure the antioxidant activity of the extracts obtained from the grape pomace was based on the use of DPPH as a free radical. The technique proposed by Brand-Williams *et al.* \\[[@B45-molecules-20-09686]\\], and modified by Scherer *et al.* \\[[@B46-molecules-20-09686]\\], is based on the use of free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH\u2022), which absorbs at 515 nm. Reduction of this radical with an antioxidant leads to the disappearance of the absorption at this wavelength. Thus, the decrease in the absorbance allows an assessment of the ability of the compound to scavenge free radicals.\n\nThe 3.9 mL of 6 \u00d7 10^\u22125^ mol/L methanol DPPH solution were added to 0.1 mL of extract methanolic solutions at different concentrations. The blank sample consisted of 0.1 mL of methanol added to 3.9 mL of DPPH solution.\n\nThe absorbance of DPPH was monitored spectrophotometrically at 515 nm at 0 min and every 2 min until a steady state was reached. The DPPH concentration (C~DPPH~) in the reaction medium was calculated from a calibration curve determined by linear regression with the following equation: $$\\text{Abs} = 12709 \\times \\text{C}_{\\text{DPPH}} + 0.002$$\n\nThe percentage of DPPH remaining was calculated as described in Equation (3). $$\\text{\\%DPPH~remaining} = \\text{C}_{\\text{DPPH}_{\\text{t}}}/\\text{C}_{\\text{DPPH}_{0}} \\times 100$$\n\nThe EC~50~ (efficient concentration providing 50% inhibition) was calculated graphically using a non-linear calibration curve by plotting the extract concentration *vs.* %DPPH remaining on the steady state. The experiments were carried out in duplicate in order to evaluate the variability of the measurements.\n\n3.6. Experimental Design\n------------------------\n\nInitially, a preliminary study was conducted to select the most appropriate technique and the best variety of grape pomace. All 6 grape varieties were tested in the two systems studied: supercritical fluid extraction (SFE) with CO~2~ + 20% ethanol as co-solvent and pressurized liquid extraction (PLE) with ethanol. The use of ethanol as a modifier in the supercritical fluid extraction (SFE) is justified since it increases significantly the solubility of the polyphenols in the supercritical phase and improves their extraction. Moreover, ethanol is the most suitable polar modifier due to its volatility and non-toxicity, in addition to its regular use in the pharmaceutical, medical, cosmetic and food industries \\[[@B47-molecules-20-09686]\\]. The SFE process operating conditions were: 100 bar, 55 \u00b0C, flow rate of 25 g/min and 20% (v/v) of co-solvent. In PLE the extraction tests were performed at 120 bar, 100 \u00b0C, flow rate of 10 g/min. In each case, the response variables analyzed were the extraction yield (expressed as g extract/100 g of dry matter) and antioxidant activity of the extracts (AA) (expressed as \u00b5g/mL of extract).\n\nAfter selecting the most suitable variety of grape pomace, the PLE process was carried out with different solvents, namely ethanol, water and an ethanol/water mixture (50:50), with the operating conditions listed above kept constant for this technique.\n\nIn addition, with the variety of grape pomace and the solvent chosen for the PLE, an experimental factorial design 2^2^ was performed to study the extraction process and to identify the best operating conditions (according to the objective of the present article). The variables selected for the experimental design were pressure (P), with values of 90 and 150 bar, and temperature (T), with values of 80 and 120 \u00b0C. Finally, flow rates of 5, 10, 15 and 20 g/min were tested for PLE under the best operating conditions. All experiments were carried out with an extraction time of 3 h.\n\nThe results were analyzed using the Statgraphics Plus 5.1^\u00ae^ (1994--2001, Statistical Graphics Corp., Princeton, NJ, USA) program to evaluate the influence of the factors on the extraction process and to determine significant differences in the samples for each variable. Empirical correlations were developed in order to predict the influence of extraction conditions on both the extraction yield and the antioxidant activity. Significance levels of factors were defined using *p* = 0.05, so factors and their combinations with a *p*-value \\< 0.05 have a significant influence on the extraction process with a confidence level of 0.95. The sign associated with each factor indicates positive or negative effect caused by the variable.\n\n4. Conclusions\n==============\n\nTwo high pressure extraction techniques, SFE and PLE, were evaluated from wine industry wastes. The comparative study using different variaties of grape pomace showed PLE using hydro-alcoholic mixture as solvent was more efficient than SFE using CO~2~ + 20% ethanol in terms of both phenolic ontent and antioxidant activity. The global yield and the yield of anthocyanins and phenolic compounds was higher for PLE than SFE.\n\nThe comparison of six different grape varieties showed that Petit Verdot grape pomace provided the highest global and phenolic extraction yield, as weel as strong antioxidant activity using pressurized ethanol as extraction solvent. However, in terms of anthocyanins Tintilla showed the highest yield (49.7 mg/g dry pomace) and total content in the extracts (741.9 mg/g extract).\n\nIn addion the experimental design and the kinetic study showed that the highest extraction yield for PLE was obtained at 120 \u00b0C and 90 bar using hydro-alcoholic mixtures at a flow rate of 5 g/min for 90 min. Furthermore, from the statistical analysis of variance it was found that temperature was the only experimental parameter that has a statistical influence on the extraction yield from grape pomace by PLE. It was demosntrated that this technique is succesfull to extract antioxidant phenolic compounds from grape pomace in order to increase the value of this winery by-product with potential applications in differents industrial sectors like cosmetics or nutraceutics.\n\nThe authors would like express their gratitude to the Spanish Ministry of Science and Innovation for financial support through Project CTQ2011-22974v and the \"Bodegas Luis P\u00e9rez\" (Jerez de la Frontera, Spain) for supplying the raw material used in the present work.\n\n*Sample Availability*: Samples of the compounds extracted are available from the authors.\n\nL. Casas and C. Mantell have conceived and designed the experiments, and wrote the paper; M.J. Otero-Pareja has performed the experiments; M.T. Fern\u00e1ndez-Ponce and M.J. Otero-Pareja have analyzed the data; E.J. Mart\u00ednez de la Ossa has directed the whole research.\n\nThe authors declare no conflict of interest.\n"}
+{"text": "Introduction {#S1}\n============\n\nSociability is a widely studied trait that can be found in all personality theories and inventories, in one form or another. High sociability (i.e., preference for a company instead of solitude) is typically associated with favorable outcomes and low sociability with harmful outcomes ([@B37]; [@B25], [@B24]; [@B18]). Our understanding of the development of sociability traits and their early life etiologic factors remains, however, limited. Although prior work has established associations between early child care and the development of personality and social competence ([@B26]; [@B21]), there is a lack of longitudinal studies on the possible effects of child care arrangements on sociability in adulthood. As the majority of children in industrialized countries are cared for outside the home at least for some period of time, childcare arrangements offer a natural setting to study the role of different early care environments in the development of sociability. The unique characteristics of different care arrangements, such as opportunities to interact with peers or exposure to pedagogically planned early education, might play a role in the development of sociability. This study investigates early childcare as an exposure that could be linked to adulthood individual differences in overall sociability and its finer aspects.\n\nSociability {#S1.SS1}\n-----------\n\nUsing a standard definition in the field, sociability generally refers to a preference for others' company instead of solitude ([@B10]) that is found to be adaptive for many social species ([@B54]; [@B17]; [@B55]), including humans. In humans, even newborn babies have an inborn need to seek others' company and to bond with them ([@B1]). Depending on the theoretical framework surrounding sociability, willingness to form relationships with others express a different kind of social need ([@B61]). For example, some schools of sociology view close social relationships as important because through them the society organizes the individual's thinking and acting (e.g., moral values, goals, even the sense of self). The more psychological view recognizes that people have several needs that only social relationships can satisfy, such as the need for recognition, for attention, for care, for belonging, for intimacy, and many more ([@B61]). In other words, these social needs may partly explain the motivation for socially active behavior or higher sociability. The clearest form of sociability, a tendency to approach instead of withdrawing from others, can be measured rather easily by observation. In practice, however, people may have equal satisfaction with a few intense relationships or with a large number of relationships of lesser intensity -- depending on social needs. The more complex constructs of human sociability, with elements related to both quantity and quality of social relationships ([@B44]), are most often measured by self-evaluations due to their practicality and because they reflect the person's self-concept ([@B49]).\n\nThe theoretical frameworks behind personality inventories differ -- while some are developed by a consideration of the underlying biological and social determinants of individual differences (e.g., the biopsychological model by [@B12]), others are derived from a careful analysis of the personality assessment literature (e.g., the so-called Five Factor Model of personality; [@B14]), yet, they all include a dimension for sociability. In some inventories, the trait has a different name (e.g., Reward Dependence or Extraversion), but its core content corresponds to the standard definition of sociability.\n\nIn most personality theories, sociability consists of correlated subcomponents related to different kinds of social needs. In other words, depending on the theoretic framework, the subcomponents emphasize different reasons *why* an individual prefers others' company instead of solitude. For example, some subcomponents may focus more on dependency on others' approval or tendency to feel emotionally attached to them (i.e., the quality of social relationships), whereas some emphasize the willingness to be surrounded by other people or the need to actively seek for as wide a social network as possible (i.e., the quantity of social relationships). These finer aspects usually strongly correlate with each other ([@B43]), but may have differed developmental paths and associate with different outcomes.\n\nAs with any other personality trait, both genetic and environmental factors are implicated in the early development of sociability. Heritability studies estimate that approximately 20 to 65% of the within-population variation in sociability is of genetic origin ([@B7]; [@B45]; [@B12]). Thus far, attempts at understanding genetic etiology of sociability have not resulted in significant breakthroughs and previous genetic findings may be confounded and intertwined with environmental influences.\n\nAs a whole, social circumstances in childhood, including the home and care environment, have been associated with a variety of adulthood outcomes from environmental and socioeconomic status (SES) to psychological and physiological functioning ([@B46]; [@B35]). As for the environmental factors potentially associated with sociability, characteristics of the early environment have been shown to modify children's later personality and their social competence ([@B42]; [@B30]). Previous research has also established associations between early childcare and the development of personality traits, including prosociality ([@B5]) and dispositional compassion ([@B21]). In the current study, we examine if early childcare environments are also associated with adulthood sociability and its specific aspects. Investigating such early environmental factors may help shed light on the developmental mechanisms behind the different aspects of sociability.\n\nEarly Childcare Environments {#S1.SS2}\n----------------------------\n\nEarly care environments can be roughly divided into those which take place at home or outside the home. In home care, the child is cared for at home either by a parent, relative, or a nanny, together with possible siblings. Home care is the most familiar care environment for a child and where most of the children are typically cared for at least for the first months of their life. However, home care can be challenging to arrange for parents after their financially supported parental leave has ended. Thus, the majority of children in industrialized countries are cared for outside of the home at least for some period ([@B59]; [@B29]).\n\nThe most common forms of outside-home care are family daycare care and center-based daycare. In family daycare, the child is cared for at a caretaker's home, usually with a small group of peers and in some cases, including the caretaker's own children. Center-based daycare, in contrast, is pedagogically planned, goal-oriented early education with trained kindergarten teachers and with well-defined guidelines regarding, for example, peer group size and adult-to-child ratio.\n\nHome care, family daycare, and center-based daycare can be distinguished by peer-group size, the level of caregiver training and the presence of pedagogical curriculum that includes developmental aims for the children, as well as by the presence of close attachment figures. Center-based daycare is the most strictly regulated early childcare environment with the most exposure to peers, whereas home care presents the most familiar, the least formal and the most individual form of care. Family daycare is something between the other two forms of care: the children are exposed to group-based care, but the group-size is usually smaller than in center-based daycare. Furthermore, in the 1980s, there were no official educational requirements for family daycare providers in Finland, even though some degree of training was strongly recommended. We acknowledge that these three early childcare environments do not, by all means, cover all the forms of childcare, but they present a rough and widely used distinction between the most common forms of care arrangements.\n\nPrevious research has shown that children's early experiences in outside-home care can promote the development of both disadvantageous and favorable developmental outcomes, depending on the cumulative amount, the quality, and the timing of the childcare exposure ([@B34]; [@B3]; [@B58]; [@B29]; [@B5]). Notably, conclusions from the previous evidence on early childhood education and care tend to vary between children who are under and over 3 years during the care environment exposure. For children aged 0--3 years the research evidence is more mixed, with some studies indicating benefits for outside-home care, some negative effects, and some studies reporting no effects at all ([@B39]). For example, several studies have shown that a high level of cumulative time spent in outside-home childcare, especially before the age of 4.5 years, is related to elevated levels of aggression, assertiveness, and disobedience in adolescence ([@B3]; [@B58]; [@B29]). By contrast, some studies have found that center-based daycare already in the infant and toddler years may contribute to the development of social competence and prosocial behavior ([@B5]), or lower levels of later emotional or behavioral difficulties compared to informal childcare ([@B22]). For children from 3 years upwards, more consistent results about the benefits of center-based daycare, preschool and other forms of group-based care have been explored in numerous studies ([@B39]). For example, a recent study found that at age 6 (but not yet at age 3), center-based daycare may increase dispositional compassion in adulthood ([@B21]).\n\nOverall, previous findings suggest that the effects of early childcare depend both on the children's developmental preparedness for outside-home care (e.g., metacognition and 'theory of mind' in the early years; [@B62]; [@B9]) and on the characteristics of the care environment, such as quality of care. Potential explanations for these findings involve the developmental unpreparedness of 0--3-year-old children to cope with social demands in care environments that are characterized by large peer groups. Still, at age 3, children typically have not developed sufficient knowledge or skills to engage in social interaction without constant adult guidance ([@B51]; [@B29]). Good quality and other structural features in an early care environment (e.g., group size and/or child-to-staff ratio, and training, permanence, sensitivity and responsivity of caretakers) have been raised as one major explanation for findings where center-based daycare may be beneficial even when the child is less than 3 years old ([@B5]). However, the 'good quality' of center-based childcare can be challenging to achieve and maintain in practice. That, in turn, places more pressure on the child's developmental preparedness to cope in the care environment. After the age of 3 to 4 years, playing tends to be more interactive and less dependent on guidance ([@B27]), and interacting with peers has a more central role in a child's development ([@B23]). Thus, especially when the child has developed a sufficient level of cognitive preparedness, center-based care is more likely to be beneficial for a child's later social development and may promote children's positive behavior in early social interactions ([@B29]; [@B39]).\n\nAlthough outside-home care has become the norm in many modern societies and despite the vast interest in how early childcare shapes a child's development, findings on the longitudinal effects of different childcare environments are still limited. Namely, most of them have a follow-up period that covers only adolescence. As the largest changes in personality tend to occur in young adulthood (ages 20--40; [@B48]), studies spanning over this age period would increase knowledge whether early environmental effects can be detectable despite these normative changes. Furthermore, studies of this kind often focus on early antisocial (criminal) tendencies or (lack of) social competence rather than on normative personality development. In the current paper, we have two aims: (1) to investigate the extent to which different childcare environments at age 3 or 6 are associated with self-evaluated sociability at ages 20 to 35 years, and (2) with what kind of aspects of sociability are these associations found. Early childcare is a natural setting in which to study not only the effects of exposure to different types of care environments but also the effects of the timing of given childcare exposure on a child's personality development later.\n\nMaterials and Methods {#S2}\n=====================\n\nParticipants {#S2.SS1}\n------------\n\nThe participants were derived from the ongoing population-based Cardiovascular Risk in Young Finns Study, or Young Finns Study for short, which is one of the largest follow-up studies on cardiovascular risk from childhood to adulthood ([@B47]). The main aim of the Young Finns Study is to determine the contributions made in childhood lifestyle, biological, and psychological measures to the risk of cardiovascular diseases in adulthood. In addition to cardiovascular health, the data enables the study of personality in the participants exceptionally widely, both longitudinally and using several different personality assessments. The original sample (*N* = 3,596) were healthy children and adolescents randomly selected from six age-based cohorts (ages 3--18 years at the baseline in 1980) which have now been followed for 32 years. The study was approved by the local ethics committee, the Finnish National Advisory Board on Research Integrity (TENK) appointed by the Ministry of Education and Culture, and it was conducted in accordance with the Helsinki declaration. Informed written consent was provided from the parents when participants were still under-aged (i.e., for measurement waves 1980 and 1983 in the present study) and after reaching adulthood, from the participants themselves (i.e., in 1997--2012).\n\nThe participants of the present study consist of the youngest age cohort, born in 1977 (*n* = 577). Only this age cohort had information about their childcare arrangements both in 1980 and 1983 when the participants were, on average, 3 and 6 years old. In practice, participant age ranged from 2 years 9 months to 3 years 11 months in the first wave and from 5 years 9 months to 6 years 11 months in the second wave. First, we excluded participants who did not exclusively attend home care, family daycare, or center-based care (the most common forms of care), and those with a non-specified form of care (e.g., *\"\\[The child is\\] at home without a caregiver*\" or \"*Other daycare arrangement \\[not specified\\],\" n* = 174 in total). A total of 360 participants (62%) had full information for all the study variables and covariates derived from childhood (in 1980 and 1983), and at least one measurement for each aspect of sociability measured in 1997--2012. Males were more likely to have missing observations; otherwise, data were missing at random. All the analyses were performed both with and without missing data modeling ([@B58]) and the results were similar using both methods. All descriptive statistics and analyses reported are based on estimates employing missing data modeling to minimize the possible effects caused by missing information.\n\nMeasures {#S2.SS2}\n--------\n\n### Early Childcare Environment {#S2.SS2.SSS1}\n\nIn the present study, we decided to focus on the three most common forms of care: home care, family daycare, and center-based daycare ([Table 1](#T1){ref-type=\"table\"}). We did this by creating a categorical variable consisting of three non-overlapping forms of care based on the childcare arrangements reported by a parent (\"*How is child care arranged?*\") in 1980 (at age 3) and 1983 (at age 6). In home care, the participants had been cared for at home by a parent, a relative, or a nanny. Family daycare refers to a caring environment where the participants had been cared for in another family, typically at a caretaker's home, and where they were accompanied by two to four other children, not including the caretaker's own children. In the 1980s, there were no official educational requirements for family daycare providers, although some degree of training was strongly recommended. Center-based daycare as a care environment, by contrast, is characterized by more strictly regulated standards regarding, for example, caregivers per child ratio, educational level of kindergarten teachers, pedagogical goals, and the size of peer-groups (per guidelines, a maximum of 12 children per group with 3-year-olds and 16 with older children).\n\n###### \n\nDescriptive statistics of the study variables (*N* = 464).\n\n                                         ***M* (*SD*)**          **Data collection intervals (*M* \\[*SD*\\] or *N* \\[%\\])**                                                         \n  -------------------------------------- ---------------- ------ ----------------------------------------------------------- ----------- ------------- ------------- ------------- -------------\n  Gender (0 = women, 1 = men)            52%              464    --                                                          --          --            --            --            --\n  **Early childcare environment**                                                                                                                                                  \n  Home care                              --               --     264 (57%)                                                   234 (50%)   --            --            --            --\n  Family daycare                         --               --     118 (25%)                                                   65 (14%)    --            --            --            --\n  Center-based daycare                   --               --     82 (18%)                                                    165 (36%)   --            --            --            --\n  **Adulthood sociability (1**--**5)**                                                                                                                                             \n  EAS: Sociability                       3.38 (0.85)      1856   --                                                          --          3.53 (0.81)   3.46 (0.82)   3.27 (0.85)   3.26 (0.87)\n  NEO-FFI Extraversion                   3.37 (0.62)      928    --                                                          --          --            --            3.38 (0.63)   3.36 (0.62)\n  TCI RD1: Sentimentality                3.05 (0.67)      1856   --                                                          --          3.11 (0.64)   3.12 (0.63)   3.00 (0.69)   2.96 (0.70)\n  TCI RD3: Social attachment             3.59 (0.86)      1856   --                                                          --          3.63 (0.83)   3.66 (0.84)   3.56 (0.86)   3.50 (0.89)\n  TCI RD4: Dependence                    3.30 (0.62)      1856   --                                                          --          3.18 (0.60)   3.33 (0.61)   3.34 (0.60)   3.35 (0.65)\n\nEAS, Emotionality-Activity-Sociability Temperament Survey; NEO-FFI, The Neuroticism-Extraversion-Openness Five-Factor Inventory; TCI, Temperament and Character Inventory. At the baseline in 1980, the participants (born in 1977) were 3-year-olds and during the latest follow-up in 2012, 35 years old. Home care: care at home by a parent, a relative, or a nanny. Family care: care at the care provider's home with a maximum of four children. Center-based care: care in kindergarten with a maximum group size of 12 (at age 3) and 16 (at age 6). All the values presented are based on estimates employing missing data modeling.\n\nTypically, children in Finland attend preschool at age 6 and the official age for school entry is 7 years. However, in the 1980s, preschools were still only partially introduced in Finland, and those 6-year-olds who were in outside-home care typically attended either family care or center-based care. For some of our participants, center-based daycare may have corresponded to pre-school, but because \"preschool\" was not a specified care arrangement option in the questionnaire that was used in the Young Finns Study in 1980 and 1983, it is not possible to separate these participants from those who still attended center-based daycare. In practice, these two forms of care are close to each other both physically and in content: pre-schools are typically situated in the same or nearby building as center-based daycare, and the children who attended pre-school in the morning typically changed to center-based daycare during the afternoon.\n\n### Adulthood Sociability {#S2.SS2.SSS2}\n\nIn the Young Finns Study, adulthood personality has been self-evaluated by three different personality inventories which provides an extraordinary opportunity to compare them using the same individuals. A total of five sociability indicators were derived from these three commonly used personality inventories: Sociability, Extraversion, Sentimentality, Social attachment, and Dependence. Sociability was measured with five items (e.g., '*I like to be with people'*; Cronbach's alpha varied between \u03b1 = 0.77--0.82 over the measurement occasions) using Buss and Plomin's Emotionality-Activity-Sociability Temperament Survey (EAS; [@B6], [@B7]). The scale assesses a tendency to prefer and enjoy the presence of others over being alone, and how comfortable a person feels in a group. Extraversion was measured with 12 items ('*I really like to discuss with people'*; \u03b1 = 0.81--0.82) using Neuroticism-Extraversion-Openness Five-Factor Inventory (NEO-FFI; [@B38]; [@B14]). The trait refers to warmth, gregariousness, assertiveness, activity, excitement seeking, and positive emotions. Sentimentality (10 items, '*I like to please other people as much as I can*'; \u03b1 = 0.70--0.76), Social Attachment (eight items, '*I would like to have warm and close friends with me most of the time'*; \u03b1 = 0.82--86), and Dependence (six items, '*I don't care very much whether other people like me or the way I do things* (reverse scored)'; \u03b1 = 0.46--0.63) were measured using subscales of the Reward Dependence scale in Temperament and Character Inventory (TCI; [@B11]; [@B12]). Sentimentality refers to a tendency to be deeply moved by emotional appeals and to an inclination to show, share, and adapt emotions easily in the presence of others; Social Attachment to a person's tendency to prefer company and intimacy over solitude and privacy; and Dependence to a person's need for emotional support and approval from others, combined with a tendency to please and be preoccupied with fears of being abandoned. These five aspects do not cover all the variability in sociability but present an example of both the quality and quantitative side of sociability and demonstrate similarities and differences between three commonly used personality inventories in regard to sociability.\n\nIn the Young Finns Study, NEO-FFI scales have been measured twice ([Table 1](#T1){ref-type=\"table\"}; age range of 30--35 years), and TCI and EAS scale four times (age range of 20--35 years). With all these inventories, a five-point precision ranging from 1 (*Definitely false*) to 5 (*Definitely true*) was used from 1997 to 2001. From 2007 onwards, the response options were slightly modified to have a range from 1 \\[(*The definition fits me) poorly or not at all*\\] to 5 *\\[(The definition fits me) very well\\]*. A mean score for each aspect of the indicator was calculated for those participants who did not have more than one missing item. All indicators of adulthood sociability correlated with each other (*r* = 0.10--0.59, *p* \\< 0.003). Responses to all the sociability indicators were combined to represent 'overall adulthood sociability' (see section \"Statistical Analyses\"). After analyzing overall adulthood sociability, the five indicators of sociability were analyzed separately because they represent different aspects of sociability.\n\n### Covariates {#S2.SS2.SSS3}\n\nAll the constructed multi-level models (see below) were adjusted for nine covariates in total: for gender, for childhood home environment (parental SES, maternal age, the number of children in the family), for disruptive behavior in childhood (aggression, hyperactivity, lack of social adjustment), and parent--child relationship quality (emotional warmth and acceptance toward the child). Adjustment of these factors removes effects of some potential causes for non-random selection to daycare groups, thereby helping causal interpretations based on this naturally occurring experiment. All possible confounding factors cannot be removed, but these nine present covariates that have previously been widely studied in association with later personality development and social behavior (e.g., [@B42]; [@B30]; [@B52]; [@B31]; [@B16]; [@B21]).\n\nFor all the covariates, the scores from 1980 and 1983 were averaged. SES was measured by the total family income, and by parents' years of education and parental occupational status. These variables were standardized, summed, and standardized again in order to form the variable for parental SES, thereby giving equal weight to income and education variation (see Text-supplement S1 in [@B50], for further details). Participant's disruptive behavior in childhood, a scale derived from the Health Examination Survey ([@B63]), presents a form of the child's unpreparedness and challenges faced in being in peer-groups and group-based care environment. The scale contains three domains: aggression, hyperactivity, and lack of social adjustment. A child's aggression (six items) includes the aggressive behavior perceived by peers (e.g., \"Other children frequently accuse him or her of fighting\"), by other adults (e.g., \"Parents of other children have complained about his or her behavior\"), and by the child's own parents (e.g., \"'Accidentally' hits, trips or shoves other children\"). In 1980 the aggression item's scale was yes/no; in 1983 a five-point Likert-scale was used, from 1 (*totally disagree*) to 5 (*totally agree*). For better comparability, both were scaled to have a range of 0 to 1 where a higher value indicates greater aggression (\u03b1 = 0.62). A child's activity was evaluated on a four-point scale ranging from \"1: *He or she stays calm even after most other children have become restless*\" to \"4: *He or she is always on the move, talks non-stop, and his or her activity is striking*.\" A child's lack of social adjustment was reported with one item on a three-point scale (\"1: *He or she survives well in everyday life*; 2: *His or her behavior does not worry you*; 3: *Occasionally, his or her behavior worries you. You think of him or her as a problem child or are afraid that he or she may become one*\"). The parent--child relationship quality scale was developed based on the Operation Family Study ([@B36]). It contains two child-rearing components: emotional warmth and acceptance toward the child. Emotional warmth was measured using four items \\[e.g., \"The child is significant to me,\" from 1 (*not significant*) to 5 (*very significant)*; \u03b1 = 0.74\\], and acceptance toward the child was measured with three items \\[e.g., \"In difficult situations, the child is a burden,\" from 1 (*totally disagree*) to 5 (*totally agree*); \u03b1 = 0.71\\].\n\nStatistical Analyses {#S2.SS3}\n--------------------\n\nWe used multilevel regression modeling as the method that recognizes the hierarchical structure of the data, such as dependencies between repeated outcome measurements in adulthood ([@B20]; [@B15]). Furthermore, this method reveals whether exposure to childcare is differentially associated with different dependency structures, or random-effects, of the data. We did this by partitioning the population variance in adulthood sociability to trait (between-individual) variance, differences among used inventories (sociability indicator variance), the time-variant part of the overlapping variance of inventories in overall sociability (within-individual variance), and measurement error or idiosyncratic differences that cannot be attributed to an individual, to follow-up or to a sociability indicator. The effect of the childcare environment was analyzed separately at the age of 3 and 6 years. All the analyses were based on pooled multiple imputation estimates using chained equations ([@B57]) and the variables were standardized before being entered into the model. Variance Inflation Factors indicated no multicollinearity problems in any of the models, with all the factors being less than 10 ([@B41]).\n\nFirst, we predicted the overall adulthood sociability by exposure to different childcare environments while simultaneously controlling for all the covariates. Second, we predicted each of the five sociability indicators separately. In regression models, home care was established as the reference group, corresponding to an intercept. We also tested if the results would hold even after adjusting for daycare history (i.e., care arrangement at age 3 and 6). This resulted in nine different combinations of daycare history. However, in some of the groups (namely, in those whose child were in outside-home care at age 3 and in in-home care at age 6) the number of observations was less than 5%, which easily leads to high variance estimates and low reliability of the results. Therefore, we concentrated on the main effects of the daycare environment (i.e., either at age 3 or at age 6). The results of daycare history, which should be regarded as indicative, are presented as [Supplementary Material](#SM1){ref-type=\"supplementary-material\"}. All statistical analyses were done in R software version 3.3.2., supplemented with a MICE package, version 2.25 for imputation analysis ([@B57]) and a lme4 package version 1.1--12 for multilevel regression analyses ([@B2]).\n\nResults {#S3}\n=======\n\nOverall Adulthood Sociability {#S3.SS1}\n-----------------------------\n\nThe descriptive statistics of sociability indicators and the distribution of participants across different forms of care are presented in [Table 1](#T1){ref-type=\"table\"}. [Table 2](#T2){ref-type=\"table\"} shows the association between early care arrangements and adulthood sociability. Family daycare and center-based care at age 3 were independently associated with overall adulthood sociability. Relative to home care, exposure to family daycare (\u03b2 = 0.19, 95% CI 0.05 to 0.29, *p* = 0.007) or center-based daycare (\u03b2 = 0.21, 95% CI 0.06 to 0.34, *p* = 0.014) at age 3 predicted a higher degree of overall adulthood sociability later in life. In contrast, at age 6, only those who were cared for in a center-based daycare had higher overall adulthood sociability compared to home care (\u03b2 = 0.21, 95% CI 0.07 to 0.30, *p* = 0.004). When the childcare at age 6 was adjusted for child care status at age 3, daycare history of outside-home care (family daycare or center-based daycare) at age 3 combined with center-based daycare at age 6 associated with higher overall sociability in comparison to home care at age 3 and 6 (see [Supplementary Table S1](#SM1){ref-type=\"supplementary-material\"}). Male gender was the only adjusting covariate that was independently associated with overall adulthood sociability. Men had on average 0.44 standard deviations lower overall adulthood sociability than women (*p* \\< 0.001).\n\n###### \n\nMultilevel regression analyses of early childcare environment at age 3 and 6 predicting mean levels of standardized sociability indicators and overall adulthood sociability.\n\n                      **Early childcare environment**   **Childcare environment at age 3**   **Childcare environment at age 6**                                          \n  ------------------- --------------------------------- ------------------------------------ ------------------------------------ ------- -------------- --------------- -------\n  EAS: Sociability    Home care                         (Ref)                                                                             (Ref)                          \n                      Family care                       **0.32^\u2217\u2217^**                         0.07 to 0.46                         0.005   0.16           \u22120.01 to 0.46   0.280\n                      Center-based care                 **0.29^\u2217^**                          0.03 to 0.50                         0.037   **0.28^\u2217^**    0.11 to 0.49    0.017\n  NEO-FFI             Home care                         (Ref)                                                                             (Ref)                          \n  Extraversion        Family care                       0.17                                 \u22120.06 to 0.29                        0.094   0.11           \u22120.08 to 0.32   0.406\n                      Center-based care                 0.20                                 \u22120.00 to 0.41                        0.108   0.20           \u22120.02 to 0.32   0.073\n  TCI: RD1            Home care                         (Ref)                                                                             (Ref)                          \n  Sentimentality      Family care                       0.08                                 \u22120.04 to 0.22                        0.372   0.07           \u22120.08 to 0.23   0.452\n                      Center-based care                 0.14                                 \u22120.02 to 0.30                        0.130   0.14           \u22120.04 to 0.22   0.084\n  TCI: RD3            Home care                         (Ref)                                                                             (Ref)                          \n  Social attachment   Family care                       **0.33^\u2217\u2217^**                         0.09 to 0.47                         0.005   0.23           \u22120.07 to 0.39   0.130\n                      Center-based care                 **0.37^\u2217^**                          0.16 to 0.62                         0.011   **0.31^\u2217^**    0.06 to 0.44    0.014\n  TCI: RD4            Home care                         (Ref)                                                                             (Ref)                          \n  Dependence          Family care                       0.03                                 \u22120.05 to 0.21                        0.705   0.06           \u22120.05 to 0.26   0.535\n                      Center-based care                 0.05                                 \u22120.13 to 0.18                        0.612   0.12           \u22120.02 to 0.23   0.133\n  Overall adulthood   Home care                         (Ref)                                                                             (Ref)                          \n  sociability         Family care                       **0.19^\u2217\u2217^**                         0.05 to 0.29                         0.007   0.13           0.00 to 0.28    0.157\n                      Center-based care                 **0.21^\u2217^**                          0.06 to 0.34                         0.014   **0.21^\u2217\u2217^**   0.07 to 0.30    0.004\n\nThe statistically significant values are bolded,\n\n\u2217\n\n=\n\np\n\n\\< 0.05,\n\n\u2217\u2217\n\n=\n\np\n\n\\< 0.01. EAS, Emotionality-Activity-Sociability Temperament Survey; NEO-FFI, The Neuroticism-Extraversion-Openness Five-Factor Inventory; TCI, Temperament and Character Inventory. NEO-FFI was measured twice (2007, 2012; age range 30--35 years;\n\nN\n\ny\\[i\\]\n\n= 928), EAS and TCI four times (1997, 2001, 2007, 2012; age range 20--35 years;\n\nN\n\ny\\[i\\]\n\n= 1,856 for each trait). Overall adulthood sociability consists of all the five sociability indicators (\n\nN\n\ny\\[i\\]\n\n= 8,352 in total). The\n\np\n\n-value indicates the difference from home care which was set as a reference group. Models were done separately for each adulthood outcome, and they all were adjusted for gender, disruptive behavior in childhood, parental socio-economic status, parent-child relationship quality, maternal age, and the number of children in the family. All the values presented are based on estimates employing missing data modeling.\n\n[Table 3](#T3){ref-type=\"table\"} presents the associations between early care and adulthood sociability separately for men and women. Namely, at age 3, family care increased overall adulthood sociability in men (\u03b2 = 0.22, 95% CI \u22120.01 to 0.34, *p* = 0.034) and center-based care in women (\u03b2 = 0.27, 95% CI 0.10 to 0.50, *p* = 0.021) in comparison to home care. At age 6, center-based care associated with higher adulthood sociability only in women (\u03b2 = 0.24, 95% CI 0.09 to 0.41, *p* = 0.018). Men had a similar but weaker trend, possibly due to a smaller number of observations (*N~*y\\[i\\]*~* = 3,978 for men and 4,374 for women).\n\n###### \n\nMultilevel regression analyses of early childcare environment at age 3 and 6 predicting mean levels of standardized sociability indicators and overall adulthood sociability separately for men and women participants.\n\n                      **Early childcare environment**   **Men**       **Women**                                              \n  ------------------- --------------------------------- ------------- --------------- ------- -------------- --------------- -------\n  EAS: Sociability    **At age 3**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       **0.41^\u2217^**   0.00 to 0.57    0.016   0.29           0.01 to 0.57    0.063\n                      Center-based care                 0.19          \u22120.13 to 0.52   0.296   0.39           \u22120.01 to 0.67   0.052\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.15          \u22120.10 to 0.56   0.477   0.18           \u22120.07 to 0.60   0.382\n                      Center-based care                 0.22          \u22120.09 to 0.47   0.202   **0.35^\u2217^**    0.14 to 0.69    0.037\n  NEO-FFI             **At age 3**                                                                                           \n  Extraversion        Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.21          \u22120.09 to 0.38   0.159   0.16           \u22120.12 to 0.38   0.302\n                      Center-based care                 0.18          \u22120.11 to 0.44   0.282   0.20           \u22120.06 to 0.55   0.255\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.12          \u22120.09 to 0.46   0.477   0.10           \u22120.16 to 0.44   0.595\n                      Center-based care                 0.19          \u22120.08 to 0.38   0.201   0.20           \u22120.05 to 0.44   0.242\n  TCI: RD1            **At age 3**                                                                                           \n  Sentimentality      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.18          \u22120.10 to 0.31   0.135   --0.02         \u22120.10 to 0.24   0.869\n                      Center-based care                 0.22          \u22120.11 to 0.37   0.110   0.06           \u22120.06 to 0.36   0.625\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.02          \u22120.28 to 0.20   0.873   0.08           \u22120.04 to 0.38   0.517\n                      Center-based care                 0.19          \u22120.12 to 0.28   0.099   0.10           \u22120.07 to 0.27   0.345\n  TCI: RD3            **At age 3**                                                                                           \n  Social attachment   Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.27          \u22120.08 to 0.47   0.098   **0.41^\u2217^**    0.11 to 0.66    0.015\n                      Center-based care                 0.15          \u22120.18 to 0.44   0.402   **0.58^\u2217\u2217^**   0.29 to 0.96    0.005\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.08          \u22120.21 to 0.42   0.692   0.37           \u22120.08 to 0.60   0.102\n                      Center-based care                 0.23          \u22120.07 to 0.46   0.158   **0.40^\u2217^**    0.07 to 0.62    0.027\n  TCI: RD4            **At age 3**                                                                                           \n  Dependence          Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.02          \u22120.12 to 0.27   0.873   0.05           \u22120.09 to 0.27   0.659\n                      Center-based care                 0.02          \u22120.26 to 0.18   0.868   0.10           \u22120.12 to 0.33   0.494\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.06          \u22120.08 to 0.36   0.660   0.08           \u22120.11 to 0.32   0.552\n                      Center-based care                 0.10          \u22120.14 to 0.24   0.404   0.15           \u22120.01 to 0.34   0.162\n  Overall adulthood   **At age 3**                                                                                           \n  sociability         Home care                         (Ref)                                                                \n                      Family care                       **0.22^\u2217^**   \u22120.01 to 0.34   0.034   0.18           0.04 to 0.36    0.064\n                      Center-based care                 0.15          \u22120.09 to 0.31   0.187   **0.27^\u2217^**    0.10 to 0.50    0.021\n                      **At age 6**                                                                                           \n                      Home care                         (Ref)                                 (Ref)                          \n                      Family care                       0.08          \u22120.08 to 0.32   0.515   0.17           \u22120.00 to 0.39   0.183\n                      Center-based care                 0.18          \u22120.04 to 0.30   0.069   **0.24^\u2217^**    0.09 to 0.41    0.018\n\nThe statistically significant values are bolded,\n\n\u2217\n\n=\n\np\n\n\\< 0.05,\n\n\u2217\u2217\n\n=\n\np\n\n\\< 0.01. NEO-FFI, The Neuroticism-Extraversion-Openness Five-Factor Inventory; TCI, Temperament and Character Inventory; EAS, Emotionality-Activity-Sociability Temperament Survey. NEO-FFI was measured twice (2007, 2012; age range 30--35 years;\n\nN\n\ny\\[i\\]\n\n= 442 for men and 486 for women), TCI and EAS four times (1997, 2001, 2007, 2012; age range 20--35 years;\n\nN\n\ny\\[i\\]\n\n= 884 for men and 972 for women for each trait). Overall adulthood sociability consists of all the five sociability indicators (\n\nN\n\ny\\[i\\]\n\n= 3978 for men and 4374 for women in total). The\n\np\n\n-value indicates the difference from home care which was set as a reference group. Models were done separately for each adulthood outcome, and they were all adjusted for disruptive behavior in childhood, parental socio-economic status, parent-child relationship quality, maternal age, and the number of children in the family. All the values presented are based on estimates employing missing data modeling.\n\nRegarding the random effects, the between-individual variance with 3-year-olds was 0.24 (95% CI 0.20 to 0.27), repeated-measurements (i.e., within-individual) variance in overall sociability was 0.01 (CI 0.00 to 0.03), the sociability indicator variance was 0.08 (CI 0.03 to 0.30), and residual/error variance that cannot be attributed to an individual, to follow-up, or to a sociability indicator was 0.62 (CI 0.60 to 0.64). In other words, between-individual differences, within-individual changes, temporally stable differences between the sociability indicators, and measurement errors accounted for 25, 3, 8, and 65% of the variance in overall adulthood sociability, respectively. The error variance includes both measurement errors in individual indicators as well as within-individual changes that are not consistent across the indicators (i.e., do not reflect overall sociability nor stable indicator-specific differences). These results were similar for the model with the care environment at age 6 as predictors.\n\nDifferent Aspects of Adulthood Sociability {#S3.SS2}\n------------------------------------------\n\nWhen the five indicators of sociability were analyzed separately, family daycare and center-based care at age 3 predicted higher adulthood Sociability (derived from EAS) and Social Attachment (TCI) in comparison to those who have been cared for at home ([Table 2](#T2){ref-type=\"table\"}). At age 6, only center-based daycare predicted higher Sociability and Social Attachment. Regarding the daycare history, outside-home care at age 3 together with center-based care at age 6 predicted higher adulthood outcome only with these two aspects of sociability (see [Supplementary Table S2](#SM1){ref-type=\"supplementary-material\"}). A similar but statistically weaker trend was present for Extraversion (NEO-FFI) and Sentimentality (TCI; *p* = 0.073 and 0.084, respectively) both with the main effects and with the supplementary analysis of daycare history. Dependence (TCI) was not predicted by daycare. From covariates, male gender was associated with lower adulthood outcome with all the other sociability indicators except Extraversion. With Extraversion (NEO-FFI), the child's lack of social adjustment both at age 3 and 6 (\u03b2 = \u22120.11, 95% CI \u22120.16 to \u22120.01, *p* = 0.027 in both cases) was the only predictor that associated with the outcome.\n\nIn the analyses performed separately for men and women ([Table 3](#T3){ref-type=\"table\"}), an early care environment at age 3 and 6 associated with higher adulthood Social attachment (TCI) only in the women's subsample. Regarding Sociability (EAS), at age 3, family care predicted a higher adulthood outcome only in men whereas in women, both family care and center-based care had a positive, yet statistically weaker, trend on adulthood outcome.\n\nDiscussion {#S4}\n==========\n\nThe purpose of this study spanning over 32-years was to examine if the childcare environment at age 3 or age 6 is associated with self-reported adulthood sociability, and if these associations depended on the specific indicator of sociability. We focused on three forms of care that present the most common forms of care: home care, family daycare, center-based daycare. These three childcare arrangements can be intrinsically distinguished by peer-group size, the familiarity of the environment, the level of caregiver training and the presence of pedagogical curriculum that includes developmental aims for the children. Furthermore, The Young Finns Study sample gave us a special opportunity to investigate the association between these early daycare environments with several aspects of sociability derived from three commonly used personality inventories with the same participants. Whereas sociability as a whole is defined as a willingness to be with others instead of in solitude, the finer aspects of sociability provide more insights of the motivation for seeking others' company, such as quality or quantity of social interactions. Our results showed that group-based outside-home care associated with higher overall adulthood sociability. With center-based daycare, we found this association both for 3- and 6-year-olds. The association was strongest for aspects of sociability that emphasize the willingness to be surrounded by other people and to be attached to them.\n\nPrevious studies have shown that exposure to center-based daycare may predict several developmental outcomes, both beneficial and harmful ([@B34]; [@B3]; [@B58]; [@B29]). Most often, center-based care from age 0 to 3 or even to 4.5 years has been seen as a risk for child's later social development ([@B3]; [@B58]). In contrast, some studies have shown that if center-based daycare is of sufficient quality, it may do no harm or even be beneficial for the socio-emotional development of children under the age of 3 ([@B22]; [@B5]). It might be that, due the limited self-control, theory of mind, or language capabilities of a child, the quality of childcare (e.g., peer group sizes and adult-to-child ratio) tends to matter more in the toddler- than in the preschool years, regardless of the form of care. That, in turn, may partly explain why previous findings with children aged 0--3 years vary more than with older children for which the evidence of the benefits of group-based outside-home care to the child's later development is more coherent ([@B39]). In other words, children closer to the preschool age can cope in groups more independently than toddlers, and thus they are likely to be less dependent on the quality of care. However, none of these studies have focused on the possible effects of the childcare environment on the development of adulthood sociability.\n\nBased on our results, 3-year-olds in both family daycare (consisting of a group of 2 to 4 children who are taken care of at the caretaker's home, not including the caretaker's own children) and center-based daycare (typically consisting of a group of 12 children) associated with higher adulthood sociability in comparison to home care. In other words, despite the difference in peer-group sizes, home care and center-based daycare did not differ in their association with sociability when participants were 3 years old. In general, Nordic countries have an early childcare system which is considered to be of high quality on an international level ([@B39]). In Finland, the municipal daycare is generally homogeneous and of high quality in regard to, for example, caregivers' educational level, child-to-adult ratios, and other daycare conditions that are based on prevailing regulations that early childcare providers are bound to follow. Thus, our result with 3-year-olds are in line with previous findings where group-based care with sufficient quality may do no harm or, more interestingly, may even be beneficial for later development.\n\nIn contrast, 6-year-olds benefited relatively more only from the exposure to center-based daycare (typically consisting of a group of 16 children) in comparison to home care in regard to their later development of sociability. The results are logical not only because of better developmental preparedness of 6-year-olds compared to 3-year-olds but also because center-based daycare, which in the case of most of the 6-year-olds corresponds to preschool, is a caring environment which focuses on preparing children to the transition to school.\n\nAs previously discussed, sociability has a general definition as a preference for others' company instead solitude, which is widely used in both an animal and human context. However, with humans, sociability is seen as a more complex construct and by using different methods besides observation, like self-evaluations, it is possible to separate finer aspects of human sociability. Namely, some aspects of sociability emphasize the quantity of social relationships and activity to seek others' company (in this study, EAS Sociability and NEO-FFI Extraversion), whereas other aspects emphasize the quality of social relationships, such as dependence on others' company and warm social attachment (such as TCI Reward Dependence's subscales Sentimentality, Social attachment, and Dependence used in our study). Even though the aspects of sociability used in the present study cover only some examples of these differences and the diversity between them, they have been derived from some of the most widely used personality inventories.\n\nPreviously, the diversity of the same self-evaluated aspects of sociability used in the present study were acknowledged ([@B43]). In the present study, variation between the sociability indicators was almost three times more than the within-individual change from age 20 to 35 (8 vs. 3%). Thus, in addition to examining overall sociability, it was justified to study the different indicators of sociability separately. With 3-year-olds, childcare arrangements that took place outside the home (i.e., family daycare and center-based daycare) associated especially with the kind of adulthood sociability that emphasizes preference to be surrounded by other people (i.e., EAS Sociability) and willingness to be socially attached to them (i.e., TCI RD3: Social Attachment). At the age of 6, only center-based daycare increased EAS Sociability and RD3: Social Attachment in adulthood in comparison to those who were cared for at home. A similar, though statistically non-significant, positive trend of center-based daycare at the age of 6 was also present for NEO-FFI Extraversion (i.e., preference to actively seek for others' company) and TCI RD1: Sentimentality (i.e., tendency to share mental states and emotions with others). In other words, based on our study, the early care environment associated both with quantitative and qualitative elements of later sociability.\n\nThere can be several reasons why participants who, in childhood, were cared for outside the home are more sociable in adulthood than individuals who were cared for at home. More interestingly, there can be several reasons why any type of outside-home care at age 3 and only center-based daycare at age 6 associated with higher adulthood sociability. One possibility is that at age 3, exposure to any kind of group care environment encourages the development of later sociability (i.e., being in a group increases preference for others' company), whereas at age 6, only center-based daycare (or, preschool) promotes the development of higher sociability. In addition to providing exposure to peer-groups, center-based daycare involves an early education and pedagogical environment that encourages and models behaviors that are probably relevant for the development of sociability.\n\nMany other studies have previously acknowledged how women as a whole tend to have higher sociability than men ([@B19]; [@B4]; [@B13]; [@B40]; [@B32]; [@B60]; [@B43]). This was also noted in the present study. Furthermore, we noticed some gender-specific trends regarding the early care environment and different aspects of adulthood sociability. Namely, associations between outside-home care and adulthood TCI Social attachment was clearer in women, and for men, family care associated with adulthood outcome (overall adulthood and EAS Sociability) more often than center-based care. The results could be explained by different evolutionary and sociocultural roles that men and women have. From an evolutionary perspective, both animal and human research identifies that the peptide hormone oxytocin, which is strongly modulated by primary female sex hormone estrogen, is related to social bonding, attachment, and affiliation ([@B8]). In other words, women on average have a better biological preparedness for this kind of social behavior which is typically needed in a nurturing context compared to men. However, nurturing-relevant social behavior represents only a small part of sociability, defined as a willingness to be with others instead of solitude. [@B33] theory of the development of gender-typed behaviors present more sociocultural perspectives to the gender differences. She argues that gender differences in individual characteristics, such as sociability, are minimal when children are observed individually. Rather, gender differences in social behavior develop from social interaction, particularly from same gender peer groups. In such groups, gender-specific interaction styles and roles start to emerge. Namely, higher sociability of women may emerge because women are expected to be more nurturing, tender-minded or more \"warm\" and orientated toward other people than men are ([@B13]; [@B32]; [@B60]). Center-based daycare may offer more opportunities for children to find same-gender peer-groups than home or family care environments do, thus also exposing children to gender-specific expectations both from peers and caregivers. This, in turn, might encourage the development of gender-specific roles that boost the development of sociability especially in girls.\n\nStrengths and Limitations {#S4.SS1}\n-------------------------\n\nThe present study is particularly strong in examining the effects of the early childcare environment on adulthood sociability and its different aspects in that we were able to use a population-based sample with a 32-year follow-up time, with multiple widely used and thoroughly validated sociability indicators, and with several adjusted covariates. Furthermore, we avoided recall bias and some of the problems of common method variance by obtaining information about childhood factors directly from the participants' parents at the beginning of the study. The extensive study of temporal precedence, potential confounding covariates, and attrition effects alleviates the risk of confounded causality in correlational designs.\n\nOur childcare data dates back to the 1980s when possibilities of a Finnish child to attend municipal daycare were dependent on parental income level. As a result, the majority of 3-year-olds were cared for at home ([@B53]). In the current study, we adjusted for parental SES, but it should nevertheless be noted that children placed in daycare during the 1980s were likely to have a lower parental income compared to children cared for at home. Since 1997, Finnish parents have had a \"subjective right to childcare\" for children under the age of 3 regardless of family income or parental employment. Nevertheless, the ratio of 3-year-olds placed in outside-home care has changed relatively little in 30 years: in the present study, 43% of 3-year-olds were in outside-home care in 1980 whereas by 2009 the number has increased to 46% ([@B56]). For 6-year-olds, we were able to have a more balanced distribution of participants across different forms of childcare than what would be attained in studies of more recent birth cohorts. Nowadays, almost all 6-year-olds attend preschool ([@B28]) whereas in our sample the ratio was 36% (in 1983). Therefore, the lack of variation especially among 6-year-olds might preclude replication of the present comparison between early childcare arrangements in contemporary Finland. Data from other countries could be valuable in these respects.\n\nThere are, however, limitations in our study that should be considered. The main limitation is that we lacked detailed information about the characteristics of daycare, such as the exact peer-group size or the cumulative hours spent in care outside the home. Moreover, we did not have information about the age of entry into outside-home care or possible changes in care arrangements before the age of 3 or between the ages of 3 to 6. We presented a supplementary analysis of daycare history where we combined daycare at age 3 with care at age 6. However, due to the unbalanced distribution of observations in different daycare history combinations, these results should be cautiously read and would require further studies. We were also unable to separate those 6-year-olds who might have been in preschool from those who were actually cared for in center-based daycares. The blurred lines between these factors may attenuate, but should not change, the general trends we observed, as in 1980s Finland the difference between center-based daycare and preschool was still somewhat inconsequential. Although we controlled for various factors in childhood, child temperament was not directly assessed, and the child's disruptive behavior can only be regarded as a crude measure for this purpose. More knowledge on childhood temperament would enable investigation in more detail whether those attending childcare differ from those who stay at home until the preschool entry, or, for example, whether sociability in childhood is related to the association between outside-home care and a higher level of adulthood sociability. Although our previous findings in partly the same sample indicate that self- or parent-selection based on the most studied genes would not explain our present results ([@B43]), we cannot rule out a selection with respect to other genes. Additionally, from the indicators of sociability used in our study, the low Cronbach's alpha of Dependence should be noted, as it indicates somewhat poor internal consistency and reliability of the scale in our study sample. This may partly explain why this indicator did not associate with the predictors, but this would need more research.\n\nConclusion {#S5}\n==========\n\nFirst, we found an association between early childcare and higher overall adulthood sociability assessed 32 years later for 3- and 6-year-olds. At age 3, both being in family daycare and being in center-based daycare associated with higher adulthood sociability than being in home care, whereas at age 6, only center-based daycare/preschool associated statistically significantly with differential adulthood sociability. Second, among the indicators of sociability, exposure to outside-home care associated especially with a person's preference for other people's company over solitude and his or her tendency to be socially attached to them. Previously, daycare has been shown to predict a wide array of social behavior (e.g., prosocial behavior and compassion), but its role in the development of adulthood sociability is poorly understood. The effect sizes in our study were small, but considering the complexity and a lifespan of humans, it is noteworthy that individual differences in sociability based on the early childcare environment were found in adulthood, after the adjustment of several potential confounding covariates. The findings warrant future studies exploring the mechanisms through which early childcare environments become associated with later sociability.\n\nData Availability {#S6}\n=================\n\nThese data were used under a data processing agreement in compliance with the GDPR. The raw data for this study is not publicly available, but it may be obtained by contacting the corresponding author. R code for data processing and analyses is available upon a request from the first author (EO). For other requests related to the data please contact Professor Marko Elovainio (<marko.elovainio@helsinki.fi>) or Adjunct Professor LP-R (<laura.pulkki-raback@helsinki.fi>).\n\nEthics Statement {#S7}\n================\n\nThe study was approved by the local ethics committee, conducted in accordance with the Declaration of Helsinki, and the participants or their parents have provided informed written consent.\n\nAuthor Contributions {#S8}\n====================\n\nJV, OR, and LP-R provided the Young Finns data. EO and TR conceived the study and performed the data analysis. EO, TR, and LK-J wrote the manuscript. KG, AS, and MH took part in the writing of the manuscript and interpreting of the findings. All authors discussed the results and reviewed the final manuscript.\n\nConflict of Interest Statement {#conf1}\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This study was supported by grants from the Mannerheim League for Child Welfare's Research Foundation and Finnish Culture Foundation awarded to EO, Academy of Finland grant number 265869 awarded to LK-J, Academy of Finland project number 258578 awarded to MH, and Jenny and Antti Wihuri Foundation awarded to LP-R. The Young Finns Study has been financially supported by the Academy of Finland: Grants 286284, 134309 (Eye), 126925, 121584, 124282, 129378 (Salve), 117797 (Gendi), and 41071 (Skidi), the Social Insurance Institution of Finland, Competitive State Research Financing of the Expert Responsibility area of Kuopio, Tampere, and Turku University Hospitals (grant X51001), the Juho Vainio Foundation, the Sigrid Juselius Foundation, the Yrj\u00f6 Jahnsson Foundation, the Paavo Nurmi Foundation, the Finnish Foundation of Cardiovascular Research and Finnish Cultural Foundation, the Tampere Tuberculosis Foundation, the Emil Aaltonen Foundation, and Diabetes Research Foundation of Finnish Diabetes Association. Open access publication fee was provided by the Helsinki University Library.\n\nSupplementary Material {#S10}\n======================\n\nThe Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fpsyg.2019.02060/full#supplementary-material>\n\n###### \n\nClick here for additional data file.\n\n[^1]: Edited by: Kai S. Cortina, University of Michigan, United States\n\n[^2]: Reviewed by: Kristel Thomassin, University of Guelph, Canada; Xiuyun Lin, Beijing Normal University, China\n\n[^3]: This article was submitted to Developmental Psychology, a section of the journal Frontiers in Psychology\n"}
+{"text": "Introduction {#s1}\n============\n\nMost adults need, on average, about 8 h of sleep per night, and sleeping \\<6 h nightly is associated with decreased daytime functioning, poorer general health, increased risk of cardiovascular and metabolic diseases and an increased likelihood of accidents ([@B1]--[@B5]). However, many people do not get enough sleep daily because of conflicting obligations, personal choices, health conditions, and other factors. Both acute and chronic sleep restriction are especially common among college-aged students \\[e.g. ([@B6])\\].\n\nAttention is a fundamental mechanism underlying cognitive abilities that is affected by sleep debt. Experimentally shortened sleep impairs attention in children ([@B7], [@B8]), and impaired attention related to reduced sleep in members of the military causes performance deficits in reaction times and accuracy, as measured by the Attention Network Test (ANT), and in continuous visual tracking ([@B9]). Thus, the impacts of sleep loss on attention, and the consequences for cognitive performance, are far-reaching and may be applicable to a wide range of both clinical and otherwise healthy populations.\n\nThe model of attention proposed by ([@B10]) includes three functional systems that depend on separate, but interacting neural networks: vigilance/alerting; orienting/selection; and executive control (or regulation of attentional resources). The relative independence of these systems is substantiated by evidence that performance in these domains can differ significantly within individuals (as assessed using tests such as the ANT), and that there are low correlations among network scores ([@B11], [@B12]).\n\nThe proposed vigilance and alerting network is involved in preparing and maintaining attention to attend to high-priority stimuli and signals, and is strongly associated with the brain\\'s norepinephrine neurotransmitter system. There are two types of alertness: tonic, which involves lengthy and sustained vigilance, and phasic, which involves shifting into an alert state in response to an acute internal or external event. Alertness does not affect the rate of information accumulation, which underlies accuracy in choice situations, but it affects response speed. Thus, increased alertness is associated with faster responses to stimuli, but at the risk of a higher rate of errors, as participants choose a response more quickly but with potentially less information upon which to base their decisions ([@B4], [@B10], [@B12]).\n\nThe second proposed network, orienting, relates to searching and selecting stimuli for further processing. In the original model, one orienting network involving parietal, frontal, and subcortical areas of the brain was proposed to be responsible for prioritizing and orienting to sensory events by location and modality ([@B10]). More recently, two (sub)networks have been proposed to be involved in these processes ([@B12]). For the purposes of this study, we will refer to a single orienting network, given the prevalence of this model in the relevant literature and its incorporation into many attention tasks, and the fact that the visual search tasks used in this study incorporate aspects of both orienting subnetworks.\n\nThe third proposed network is executive control, which is responsible for allocating attentional resources. While originally one executive control network was proposed to be responsible for detecting signals from all sources ([@B10]), more recently, ([@B12]) proposed two separate subnetworks of executive control that stem from a common developmental origin. While cognitive theories differ with respect to whether executive control operates as a single system or not, the functions served by executive control are very similar in both views ([@B12]).\n\nA variety of different measurement tools, including the Psychomotor Vigilance Task (PVT), the ANT, the Sternberg item memory test and go/no-go tests have been used in the past to measure the impact of sleep loss on attentional performance. Various amounts of sleep loss have also been studied, including total sleep deprivation (remaining awake for one or more full nights); acute partial sleep restriction (receiving less than habitual amounts of sleep for one night); and chronic partial sleep restriction (receiving less than habitual amounts of sleep over a period of several days). See ([@B13]) for a review. Despite the large amount of previous research, there remain gaps in our understanding of how sleep loss affects different functional aspects of attention.\n\nIt is widely accepted that vigilance is reliably lowered by total sleep deprivation ([@B2], [@B3], [@B14]--[@B16]) and chronic ([@B3], [@B14]) and acute ([@B2]) sleep restriction, although there have been fewer studies on the effects of acute partial sleep loss. It is less well understood how sleep loss affects orienting and executive functions, and studies examining broader aspects of attention (e.g., using the ANT) have produced inconsistent results \\[e.g., ([@B4], [@B9], [@B17])\\]. The reasons for this lack of clarity likely include the fact that the performance of separate attentional networks is often compared between, rather than within, studies and the impact of the \"task impurity\" problem. Task impurity refers to the fact that many tasks used to measure executive functioning assess a complex combination of executive skills, including working memory, inhibition and task switching. It is therefore difficult to assign performance declines after sleep loss to effects on one or more individual component processes ([@B13]). Given that tests across studies also differ in task demands, it is difficult to determine whether these attentional processes are differentially sensitive to sleep loss.\n\nThe purpose of this study was to investigate the differential impact of acute, partial sleep restriction on all three attentional networks. Given the limitations of available measures of attention, the Dalhousie Computerized Attention Battery (DalCAB) was developed, based on the Posner and Peterson model ([@B10]), to simultaneously assess the vigilance, orienting, and executive control networks of attention. The DalCAB includes eight psychometrically stable computerized tasks that assess these networks, and has been shown to have good test-retest reliability ([@B18]) as well as construct validity as a measure of all three networks ([@B19]). In the present study, healthy female participants were assigned to either sleep restriction or control groups. For both groups, the DalCAB was administered on two separate occasions, once after a 9 h sleep opportunity, and again after either a 3 h sleep opportunity or after another 9 h sleep opportunity (to control for practice effects).\n\nBased on consistent findings in the literature related to the effects of sleep loss on vigilance ([@B2]--[@B4], [@B14], [@B15]), we predicted that vigilance performance would be significantly impaired following sleep restriction after controlling for possible practice effects. The less consistent literature related to effects of sleep loss on orienting and executive control ([@B4], [@B9], [@B16], [@B17]) did not allow a prediction as to how the function of these networks would be affected.\n\nMethods {#s2}\n=======\n\nParticipants\n------------\n\nBecause resource limitations did not allow studying an adequate number of participants of both sexes, and women are less frequently studied, and often without adequate consideration of potential effects of menstrual cycle phase, only female participants were included. Sleep characteristics and physiological responses to sleep loss have been shown to be affected by menstrual cycle phase ([@B20], [@B21]); thus, all participants were studied during the mid-follicular phase of their menstrual cycle, as determined by the self-reported timing of two immediately preceding menstrual cycles.\n\nThirty-nine healthy women (aged 19--25 years, mean age 21 years) were recruited from the community in Halifax, Nova Scotia, Canada. Inclusion criteria were: a self-reported habitual sleep duration of 7--9 h, confirmed by sleep diaries 7 days prior to study participation; no regular daytime naps; no history of sleep disorders; no intake of caffeinated beverages exceeding the equivalent of two cups of coffee (\\~300 mg caffeine) daily; no recent history of shift work or recent (past 6 weeks) transmeridian flights crossing more than two time zones; no history of regular recreational drug use; no history of neurological disorders or current psychiatric disorders; non-extreme chronotype, as determined by a Morningness-Eveningness Questionnaire \\[MEQ; ([@B22])\\]; no diagnosis of head injury with a loss of consciousness in the past 5 years; currently taking oral contraceptives regularly; normal or corrected-to-normal vision and hearing; body mass index between 18.5 and 25.0; and no experience of general anesthesia during the last 4 months.\n\nTests\n-----\n\nThe DalCAB comprises eight reaction time (RT) tasks (Table [1](#T1){ref-type=\"table\"}), each designed to measure the processes underlying different components of attention by collecting RTs and accuracy. These include: vigilance, orienting, and executive control components ([@B19]). The DalCAB tasks are as follows:\n\n###### \n\nDalhousie Computerized Attention Battery (DalCAB) tasks, task functions and task-specific variables.\n\n  **Proposed network**   **Task[^*a*^](#TN1){ref-type=\"table-fn\"}**   **Function**                   **Task-specific variables**\n  ---------------------- -------------------------------------------- ------------------------------ ----------------------------------\n  Vigilance              Simple RT                                    Response speed                 Response Stimulus Interval (RSI)\n                         Choice RT                                    Decision response speed        Response switch\n                         Feature visual search                        Search and select              Distractor set size\n  Orienting/selection    Conjunction visual search                    Search and select              Distractor set size\n  Executive              Go/no-go                                     Inhibition                     Go frequency\n                         Dual task                                    Dual task                      Response switch\n                         Flanker                                      Filtering, response conflict   Congruency of flankers\n                         Item memory                                  Verbal working memory          Set size, trial type\n                         Location memory                              Spatial working memory         Set size, trial type\n\n*Network designations determined by factor analysis in ([@B19]). Tasks are sorted into the different attentional networks proposed by ([@B10])*.\n\n### Simple reaction time (SRT; vigilance)\n\nThe reaction time from stimulus onset to button press was used to measure response speed. Response-stimulus intervals (i.e., the length of time between participant response and subsequent stimulus presentation) varied between 500 and 1500 ms.\n\n### Choice reaction time (CRT; vigilance)\n\nA two-choice reaction time task measured decision response speed. Response switching effects were also calculated on consecutive trials that required a different response choice as compared to consecutive trials requiring the same response.\n\n### Visual search (orienting; vigilance)\n\nParticipants searched and made an orientation judgment to colored stimuli among different set sizes of distractors ranging from 6-18 items. Targets were defined by a single feature (e.g., color, termed feature search) or by the combination of 2 features (e.g., color and shape; termed conjunction search).\n\n### Dual task (executive)\n\nControl of attentional resources was measured by combining the choice reaction time (CRT, described above) task with a counting task that involved responding to each trial as above, while simultaneously keeping track of how many times each of the 2 stimulus choices were presented.\n\n### Flanker task (executive)\n\nA flanker task assessed response conflict resolution and filtering (an aspect of executive control). Choice reaction time to identify the shape of a central stimulus with the same (congruent) or different (incongruent) vertical flanking stimulus shapes was measured. The congruency effect was calculated by subtracting the RTs or accuracy in the congruent condition from those in the incongruent condition.\n\n### Go/No-Go (GNG) task (executive)\n\nThe go/no-go task was used to measure initiation and inhibition. Participants were required to respond to certain stimuli (\"go\"), and to inhibit responses to others (\"no-go\"). The frequency with which \"go\" stimuli appeared varied from 20 to 80% across blocks of trials.\n\n### Item working memory task (executive)\n\nThe item working memory task presented a series of non-repeating stimuli (2--6 items), followed by a probe target stimulus after a delay. Participants were required to identify whether the target stimulus was present in the preceding series of stimuli.\n\n### Location working memory task (executive)\n\nThe location working memory task presented a stimulus at different sequential locations (2--6 locations), followed by a probe target stimulus. Participants were required to identify whether the target stimulus location was included or not included in the previous sequence.\n\nMood and sleepiness scales\n--------------------------\n\n### Morningness-eveningness questionnaire\n\nAll participants completed the Morningness-Eveningness Questionnaire ([@B22]) prior to participation in the study to ensure that they were not extreme morning or evening types (usually referred to as \"chronotypes\"). The questionnaire included 13 questions, each with four or five answer options, and was scored such that a higher score on a question indicated a morning chronotype, and a lower score indicated a more evening chronotype. The possible range of scores was from 13 to 55; a score between 23 and 43 was required for inclusion, as scores outside this range indicate extreme morning or evening chronotypes ([@B22]).\n\n### Profile of mood states (POMS)\n\nParticipants completed the Profile of Mood States ([@B23]) each evening and morning of the study, in order to assess self-reported changes in Tension-Anxiety, Depression-Dejection, Anger-Hostility, Vigor-Activity, Fatigue-Inertia, and Confusion-Bewilderment in conjunction with sleep restriction. Each of these six mood factors was scored individually, with a higher score indicating a stronger mood state. Factors were then summed (or subtracted, in the case of the Vigor-Activity score) to produce a Total Mood Disturbance Score. In the case of missing items, the mean of all completed items contributing to the same factor was used as an imputed item score ([@B23]).\n\n### Stanford sleepiness scale (SSS)\n\nParticipants completed the Stanford Sleepiness Scale each morning of the study ([@B24], [@B25]) to quantify changes in subjective sleepiness. The scale consists of seven ranked statements, with higher scores indicating more sleepiness. Participants were asked to record the rank of the statement that most accurately represented their current degree of sleepiness.\n\nApparatus\n---------\n\nDalCAB stimuli were presented on a 17\u2033 monitor of an Apple iMac G5 computer. Participant responses were collected using a Razer DeathAdder Gaming Mouse (sensitive to 1 ms response times), using the left and right mouse buttons.\n\nProcedure\n---------\n\nIndividuals interested in participating were first screened with respect to preliminary inclusion criteria during a telephone call. Eligible participants completed a screening questionnaire and the MEQ to confirm eligibility. After written informed consent was obtained, participants were assigned to either the control (C; *n* = 19) or sleep restriction (SR; *n* = 20) group. The first ten SR participants were recruited prior to the addition of C participants, after which group assignment was randomized. Each participant was scheduled to participate in the laboratory portion of the study near the midpoint of the follicular phase of her menstrual cycle. Participants completed a sleep diary every morning during the week prior to this point.\n\nParticipants spent three consecutive nights (Figure [1](#F1){ref-type=\"fig\"}) in the Chronobiology Laboratory at the QE II Health Sciences Centre (Halifax, Nova Scotia). The first night served to adapt participants to sleeping in the laboratory. They arrived at 21:30, completed the Profile of Mood States (POMS) and were allowed to sleep in a darkened bedroom from 22:00 to 07:00. The following morning, they completed the POMS and Stanford Sleepiness Scale (SSS) within an hour of awakening, were provided with breakfast (without a caffeinated beverage), and were instructed to go about their normal daily schedule without daytime napping or caffeine intake. The same procedures were followed on the second night. The following morning (Day 1), participants completed the POMS, SSS and started the DalCAB within an hour of waking. After completing the DalCAB (\\~1 h), participants were provided with breakfast (without a caffeinated beverage) and resumed their normal daily schedule, without daytime naps or caffeine intake.\n\n![Gray rectangles represent participant sleep opportunities. Blue arrows represent Profile of Mood States (POMS) completion. Yellow arrows represent Stanford Sleepiness Scale (SSS) completion. Red arrows represent Dalhousie Computerized Attention Battery (DalCAB) completion. DalCAB was always started within 1 h of waking in the mornings.](fpsyt-09-00499-g0001){#F1}\n\nDuring the third night, procedures for C participants were identical to those on the preceding night, with 9 h time in bed (TIB; 22:00--07:00). SR participants also arrived by 21:30 on the third night and completed the POMS. They then stayed awake until 04:00 and were allowed 3 h TIB until they were awakened at 07:00. The following morning (Day 2), both C and SR participants completed the POMS, SSS, and DalCAB, as on Day 1. During the time that SR participants stayed awake overnight, they participated in non-strenuous activities, such as reading, playing video or board games, listening to music, watching videos, or working on a computer, while continuously in the company of a research assistant. Participants in the SR group were not permitted to drive after sleep restriction, so they were sent home either by an arranged ride or taxi, if transportation was needed.\n\nParticipants received CAD \\$125 as compensation for their time and effort in participating in the study. The study protocol was approved by the Capital District (now Nova Scotia) Health Authority Research Ethics Board, in conformity with the Canadian Tri-Council Policy Statement 2: Ethical Conduct of Research involving Humans (2014).\n\nData analysis\n-------------\n\nAll statistical analyses were performed in R version 3.1.1 (R, Retrieved from <http://www.r-project.org/>). *P* values \\< 0.05 were considered statistically significant. Participant demographics were compared between groups using unpaired *t*-tests, with the exception of handedness which was compared using a chi squared test for independence. POMS and SSS scores in the morning were compared across days and treatment groups, using mixed design analysis of variance (ANOVA). POMS scores were assessed as individual mood dimensions.\n\nReaction times (RTs), measured in ms, were collected for each DalCAB task. Only correct responses \\>100 ms which fell within the trial length \\[varied by task, see ([@B18])\\] were included in analyses of RTs for all tasks. Accuracy was also calculated for each task, with accuracy defined specifically for each task (as described in Results). Anticipations were defined as RTs \u2264 100 ms. RTs, accuracy (% correct), and anticipations (as applicable) were compared across mornings (Day 1 vs. Day 2) and between groups (C vs. SR), as well as across the different levels of the task-specific variables, using mixed design ANOVA. Statistically significant results from the ANOVAs were followed by *post-hoc* tests for both accuracy and RTs using Tukey\\'s HSD for multiple comparison correction. With the exception of demographic data, only significant results are reported in the text. All findings and details of the RT ANOVA results are included in the [Supplementary Material](#SM1){ref-type=\"supplementary-material\"}.\n\nResults {#s3}\n=======\n\nDemographics\n------------\n\nControl and sleep restriction groups did not differ significantly with respect to age, years of education, body mass index, handedness, MEQ score or baseline sleep diary total sleep time; see Table [2](#T2){ref-type=\"table\"}.\n\n###### \n\nParticipant demographic information.\n\n  **Attribute**                                                             **Control group**   **Sleep restriction group**   ***p*-value**\n  ------------------------------------------------------------------------- ------------------- ----------------------------- ---------------\n  *n*                                                                       19                  20                            --\n  Age (years)                                                               20.7 \u00b1 0.38         21.3 \u00b1 0.42                   0.33\n  Education (years)                                                         14.5 \u00b1 0.39         15.3 \u00b1 0.39                   0.18\n  BMI                                                                       22.5 \u00b1 0.51         22.7 \u00b1 0.44                   0.70\n  Handedness (Right)                                                        84%                 95%                           0.56\n  MEQ score                                                                 34.4 \u00b1 1.56         35.5 \u00b1 0.94                   0.58\n  Sleep diary TST (avg. hours)                                              7.8 \u00b1 0.16          7.9 \u00b1 0.18                    0.58\n  Sleep diary sleep efficiency[^*a*^](#TN2){ref-type=\"table-fn\"} (avg. %)   90.6 \u00b1 0.97         91.2 \u00b1 1.02                   0.68\n\n*Proportion of time in bed spent asleep. Values shown are the mean \u00b1 SEM*.\n\nMood and sleepiness scores\n--------------------------\n\nFigures [2](#F2){ref-type=\"fig\"}, [3](#F3){ref-type=\"fig\"} present self-reported sleepiness (SSS) and fatigue and confusion scores (subscales of the POMS), respectively, across days for both C and SR groups. An interaction between day and group for SSS scores (*p* = 0.001; Figure [2](#F2){ref-type=\"fig\"}) revealed a significant change in sleepiness from Day 1 to Day 2 for SR participants (*p* = 0.002), and a significant difference between C and SR participants on Day 2 (*p* \\< 0.001). Similarly, significant interactions between day and group were found for fatigue (*p* \\< 0.001) and confusion (*p* = 0.02) (Figure [3](#F3){ref-type=\"fig\"}). Fatigue scores increased significantly from Day 1 to Day 2 for the SR group (*p* = 0.001) and differed between the C and SR groups on Day 2 (*p* \\< 0.001). No *post-hoc* tests of POMS confusion scores were significant.\n\n![Stanford Sleepiness Scale (SSS) scores on Day 1 and Day 2. In all figures, error bars indicate the standard error of the mean. In all figures, ^\\#^*p* \\< 0.05; ^\\*^*p* \\< 0.01; ^\\*\\*^*p* \\< 0.001.](fpsyt-09-00499-g0002){#F2}\n\n![Profile of Mood States (POMS) subscale scores for Fatigue and Confusion on Day 1 and Day 2.](fpsyt-09-00499-g0003){#F3}\n\nDalCAB performance\n------------------\n\nFor each DalCAB task, only significant Group X Day interactions for accuracy and reaction time (RT) outcomes are reported (see [Supplementary Material](#SM1){ref-type=\"supplementary-material\"} for all RT ANOVA results).\n\nVigilance performance (simple and choice reaction time, feature visual search)\n------------------------------------------------------------------------------\n\nFigure [4](#F4){ref-type=\"fig\"} plots RTs (in ms) in the Choice Reaction Time task across days and groups. A significant interaction between group and day (*p* = 0.02) revealed that RTs in both groups were faster on Day 2 than Day 1 (SR: *p* = 0.004; C: *p* \\< 0.001), but that the SR group was significantly slower than the C group on Day 2 (*p* \\< 0.001).\n\n![Reaction times (ms) by Day and Group for the Choice Reaction Time task.](fpsyt-09-00499-g0004){#F4}\n\nThe group by day interaction was also significant for anticipations on the Simple Reaction Time Task (*p* = 0.02), but no further comparisons were significant (Tukey\\'s *post-hoc* tests).\n\nFigure [5](#F5){ref-type=\"fig\"} plots RTs in the Feature Visual Search task across days and groups. While RTs were significantly faster for both groups on Day 2 than on Day 1 (SR: *p* \\< 0.001; C: *p* \\< 0.001), RTs for the SR group were faster on Day 1 (*p* \\< 0.001), but slower on Day 2 (*p* = 0.003), when compared to the C group.\n\n![Reaction times (ms) by Day and Group for the Feature Visual Search task.](fpsyt-09-00499-g0005){#F5}\n\nExecutive performance (Go/No-Go, item memory, vertical flanker)\n---------------------------------------------------------------\n\nAnalyses of the Go/No-Go task revealed slower RTs after sleep restriction (*p* \\< 0.001; Figure [6](#F6){ref-type=\"fig\"}). Participants in the C group showed significantly faster RTs on both days when compared to the SR group (Day 1: *p* \\< 0.001; Day 2: *p* \\< 0.001). The SR group, however, showed slower RTs on Day 2 than on Day 1 (*p* \\< 0.001), while the C group showed faster responses on Day 2 than on Day 1 (*p* = 0.001).\n\n![Reaction times (ms) by Day and Group for the Go/No-Go task.](fpsyt-09-00499-g0006){#F6}\n\nA significant Group X Day interaction on the Item Memory task (*p* = 0.039) revealed significantly faster RTs in the SR group than the C group on both Day 1 (*p* \\< 0.001) and Day 2 (*p* = 0.012), and faster RTs in both groups on Day 2 (C: *p* \\< 0.001; SR: *p* = 0.033). There was, however, less improvement in the SR than in the C group between Days (Figure [7](#F7){ref-type=\"fig\"}).\n\n![Reaction times (ms) by Day and Group for the Item Memory task.](fpsyt-09-00499-g0007){#F7}\n\nFigure [8](#F8){ref-type=\"fig\"} shows the significant interaction between Day and Group on Flanker Congruency effects on the Vertical Flanker task (*p* = 0.03). No further comparisons were significant (Tukey\\'s *post-hoc* tests).\n\n![Measure of the Congruency Effect (Incongruent -- Congruent RTs \\[ms\\]) by Group and Day for the Vertical Flanker Task.](fpsyt-09-00499-g0008){#F8}\n\nDiscussion {#s4}\n==========\n\nThis study was conducted to investigate the impact of acute, partial sleep restriction simultaneously on all three theoretical attentional networks: vigilance, orienting, and executive control ([@B10]). The DalCAB allows for a direct comparison of performance changes related to sleep loss on the three networks using a common methodology during a single 1 h test session.\n\nThe SSS scores and POMS Fatigue and Confusion scores on Day 2 showed that restriction for a single night to 3 h TIB effectively increased subjective sleepiness and associated mood changes, as expected ([@B2], [@B14]), while those with 9 h TIB showed no significant changes in these measures on Day 2.\n\nThe vigilance network, responsible for preparing and maintaining readiness to respond ([@B10]), was primarily measured in the DalCAB by the Simple and Choice RT tasks. For the Choice RT task, while both groups showed significant practice improvements on Day 2, there was less improvement in the SR group and a significant RT difference between the two groups on Day 2 that was not present on Day 1. These results are consistent with the known effects of sleep restriction in reducing vigilance ([@B4], [@B26], [@B27]), slowing reaction times ([@B3], [@B4], [@B26]--[@B29]), and increasing lapses and other errors ([@B2], [@B4], [@B14], [@B15], [@B30]) on RT tasks.\n\nA factor analysis of DalCAB variables showed that mean RTs from the Feature Visual Search task also assessed the function of the vigilance network of attention ([@B19]). SR participants showed significantly slower RTs on this task than the C group on Day 2, even though they initially had faster RTs on Day 1. Thus, the results on the tasks that measured vigilance indicate that SR reduced the function of this network as expected.\n\nThe literature related to sleep loss effects on the orienting network of attention is inconsistent, in part because of the use of different methodologies and study populations ([@B4], [@B7], [@B9], [@B17], [@B28]). In this study, the orienting component of attention was assessed with the Conjunction Visual Search task. There was no significant group by session interaction on outcomes of this task, in either reaction time or accuracy data, thereby indicating no effect of sleep loss on orienting performance, which is consistent with some ([@B4], [@B9], [@B17]) but not other ([@B27]) previous findings.\n\nOne possible contributor to the inconsistency in the published literature is variation in what is being measured using different orienting tasks. Various orienting tasks use different paradigms to assess the deployment of spatial attention, often by using a combination of invalid, valid, and neutral visual cues preceding a target stimulus. In the DalCAB Conjunction Visual Search task, participants were not cued to their target stimulus, and relied on internal control. This method is comparable to the neutral cues often used in control trials in similar studies. In these cases, reaction times to neutral cues are often the same before and after sleep loss ([@B26], [@B31]), even when there is impairment due to sleep loss on RTs for invalid cues ([@B31]). Thus, it is possible that internal control of spatial orienting is less affected by sleep loss than is external control (i.e., by cueing) of spatial attention.\n\nThis result may be specific to the costs of invalid cues; some studies have found that RTs on validly cued trials are not affected by sleep loss ([@B26], [@B31]). Additionally, orienting to a new sensory stimulus has been modeled as consisting of three component steps: disengaging attention from a previous target; looking (moving the gaze); and seeing (discriminating a new target) ([@B26], [@B31]--[@B33]). It has been suggested that it is the disengagement step, required for invalid cue conditions, which is most impacted by sleep loss ([@B32]). On the other hand, another study found that the RT costs of invalid cues actually go down after sleep loss ([@B4]). Alternatively, the degree of sleep loss in this study may not have been sufficient to impair the orienting network to a degree that could be measured by the conjunction visual search task.\n\nExecutive control of attention comprises several different processes that have been assessed using a variety of tasks. Some studies have reported impairment of executive function after sleep loss ([@B9], [@B17], [@B27], [@B29], [@B34]) while others have not ([@B4], [@B7], [@B28], [@B34], [@B35]). This variability probably reflects the use of different tasks and the fact that some tasks may target different aspects of executive control and/or the use of tasks with multiple components that do not target executive function specifically. Three of the five tasks that tap into executive mechanisms in this study showed impairment after sleep loss (i.e., Go/No-Go, Flanker, and Item Memory tasks). On the Go/No-Go task, C participants showed faster responses on Day 2, while SR participants showed slower responses. On the Flanker task, only the C group improved from Day 1 to Day 2. In addition, the performance benefit of stimulus congruity was lost on Day 2 for the SR group, but not the C group. Similarly, performance on the Item Memory task improved for both groups on Day 2, but the improvement was significantly greater for the C group. In contrast to these findings, sleep restriction did not impair performance on the Location Memory or Dual tasks.\n\nBased on these results, we can conclude that some components of executive function---inhibition (Go/No-Go), filtering (Flanker), and verbal working memory (Item Memory), were impaired by sleep loss. It is not clear, however, whether task switching was affected, since performance on two tasks that involve task switching were affected differently: Choice Reaction Time performance was impaired but Dual task performance was not.\n\nThe fact that C participants showed practice-related improvement on most executive control tasks on Day 2 raises a question of interpretation. There is a large literature demonstrating that performance on a wide array of learning tasks may improve after overnight sleep, presumably because of the benefits of sleep (or one or more sleep stages) on the process of consolidating newly acquired skills or information \\[for review, see ([@B36])\\]. These effects are seen for sensory and motor learning and for a variety of tasks that involve explicit learning \\[e.g., ([@B37], [@B38])\\]. It is unclear whether the practice effects observed in this study were related to changes in sensory processing, motor responses or strategic approaches to these tasks, but, in principle, any of these could reflect a form of incidental learning during the first experience of the tasks. If this learning were subject to sleep-related consolidation, relatively poorer performance by SR participants on Day 2 might reflect both direct effects of sleep loss on performance and indirect effects mediated by a lack of sufficient consolidation of skills or strategies acquired on Day 1. However, it is also possible that retesting participants later on Day 1 would have yielded similar improvements; i.e., that the improvement was not sleep-related. Different experimental approaches would be needed to evaluate these possibilities.\n\nStrengths and limitations {#s5}\n=========================\n\nAmong the strengths of this study are the use of one night of habituation to a novel environment to reduce any \"first-night\" effects on sleep quality ([@B39]--[@B41]), and testing baseline performance only after a second night of sleeping in the laboratory. We also used sleep diaries to ensure that participants did not experience unusual sleep timing or durations immediately before entering the study. Finally, the use of the DalCAB allowed assessment of a variety of cognitive mechanisms in a single study using an integrated testing approach (i.e., consistent stimuli, tasks purported to measure specific components of attention and all tests administered in a single session).\n\nOne limitation of this study is that, in the absence of polysomnography, we can only refer to time in bed (TIB) as an indication of sleep duration, rather than to actual sleep durations, nor do we have information on stages of sleep obtained in each condition. It is unlikely that participants slept a full 9 h when allowed to do so, but this duration was intended to allow them a typical, full night of sleep (reported as 7--9 h by participants). Control participants also reported sleeping typically during the nights with 9 h TIB. SR participants were accompanied by a researcher throughout their wake period, so they could not have slept more than 3 h on the SR night (04:00--07:00). The sleep manipulation was successful in inducing increased subjective sleepiness and fatigue (as measured by the SSS and POMS) in the SR but not the C group.\n\nBecause the night of sleep restriction followed immediately after the first test day, we cannot discriminate effects of sleep loss on next-day performance from effects of a reduced opportunity for sleep-related consolidation during the SR night. Although the tasks involved did not explicitly involve learning, improvement among C participants on Day 2 raises the possibility that some form of memory consolidation or skill improvement occurred, but there is no information as to whether it occurred during Day 1 testing, during subsequent waking or during the following night. Studies of nap effects on memory consolidation indicate that even short naps can provide an opportunity for memory consolidation that can affect next-day performance ([@B42]--[@B44]). Since we do not know whether improvement on Day 2 in controls was related to sleep, we also cannot speculate whether 3 h of sleep would provide an adequate consolidation opportunity.\n\nFinally, the conclusions about effects of sleep loss on attention in this study are applicable only to the current sleep restriction paradigm (i.e., a single night with 3 h TIB). Future studies would have to evaluate whether protocols involving more extreme or repeated episodes of sleep loss would have different effects on the processes measured using these tests. In addition, there were baseline differences between groups in performance on a few tasks (feature visual search, Go/No-Go, item memory). It is unknown why these otherwise well-matched groups differed in this way. These baseline differences did not affect achieving the goal of the study, which was to examine how performance changed from baseline after sleep restriction or adequate sleep.\n\nConclusions {#s6}\n===========\n\nA single night of sleep restriction to 3 h TIB increased daytime sleepiness and fatigue the next day. Associated with these changes were impairments in performance on tasks assessing vigilance and some executive processes related to visual attention (i.e., filtering, inhibition, working memory) but not others (task switching); no effects were observed on spatial orienting processes related to attention. These results imply that the relatively common experience of reduced sleep for a single night can have significant effects on several mechanisms related to attention, which can in turn affect many aspects of cognition. A secondary implication is that relatively modest, unintended sleep loss in study participants could affect performance on cognitive tasks, which may therefore not accurately reflect cognitive capacity that would be shown under sleep satiation conditions. This consideration may be especially relevant when testing populations with a variety of clinical conditions that are commonly associated with sleep disruption.\n\nAuthor contributions {#s7}\n====================\n\nJC: Study design, data collection, data analysis, and manuscript writing; SJ: Data analysis, manuscript writing, mentorship, and assistance; GE: Study design, data analysis, manuscript writing, and overall supervision; BR: Study design, data collection, data analysis, manuscript writing, and overall supervision.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe are grateful to Antonya Hall and Joshua Salmon for research assistance.\n\nThe datasets for this manuscript are not publicly available due to intellectual property protection. Requests to access the datasets should be directed to Gail Eskes, [gail.eskes\\@dal.ca](http://www.gail.eskes@dal.ca).\n\n**Funding.** Atlantic Canada Opportunities Agency (Atlantic Innovation Fund Project No. 195320); Dalhousie Psychiatry Summer Studentship; Natural Science and Engineering Research Council of Canada (RPGIN 305); NSERC Rehabilitative and Diagnostic Innovation in Applied Neurotechnologies Summer Undergraduate Research Fellowship.\n\nSupplementary material {#s8}\n======================\n\nThe Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fpsyt.2018.00499/full#supplementary-material>\n\n###### \n\nClick here for additional data file.\n\nANT\n\n:   Attention Network Test\n\nC\n\n:   Control (group)\n\nCRT\n\n:   Choice reaction time\n\nDalCAB\n\n:   Dalhousie Computerized Attention Battery\n\nGNG\n\n:   Go/no-go\n\nMEQ\n\n:   Morningness-Eveningness Questionnaire\n\nOSO\n\n:   Overnight sleep opportunity\n\nPOMS\n\n:   Profile of Mood States\n\nPVT\n\n:   Psychomotor Vigilance Task\n\nRT\n\n:   Reaction time\n\nSR\n\n:   Sleep restriction (group)\n\nSRT\n\n:   Simple reaction time\n\nSSS\n\n:   Stanford Sleepiness Scale\n\nTIB\n\n:   Time in bed.\n\n[^1]: Edited by: Lino Nobili, University of Genova, Italy\n\n[^2]: Reviewed by: Axel Steiger, Max-Planck-Institut f\u00fcr Psychiatrie, Germany; Karen Spruyt, Institut National de la Sant\u00e9 et de la Recherche M\u00e9dicale (INSERM), France\n\n[^3]: This article was submitted to Sleep and Chronobiology, a section of the journal Frontiers in Psychiatry\n"}
+{"text": "1. Introduction {#S1}\n===============\n\n1.1. History of ATRT {#S1-1}\n--------------------\n\nATRT, a cancer of the CNS, was christened by Rorke et al. in 1996, following a review of 52 pediatric cases ([@B1]). We may trace the first appearance of the term \"atypical teratoma\" to four decades earlier, where it was recognized to occur in the pineal gland ([@B2]). The \"atypical\" refers descriptively to the \"teratoid\" part of the tumor. This appears to show elements of two germ cell layers (ectoderm, i.e., primitive neuroepithelial, epithelial, and mesoderm, i.e., \"rhabdoid\") but is otherwise quite different from classical teratoid tumors. \"Rhabdoid\" (Greek, \"rod-like\") refers to the similarity, on microscopy, to tumors showing skeletal muscle differentiation.\n\nThe biology of and treatment strategies for ATRT (focusing on children) have been the subject of a recent, thorough review. Although the prognosis in children has traditionally been dismal, with few long-term survivors, developments over the last 10\u2009years have, at least for children, been encouraging ([@B3]). Recently, mutations in the chromatic remodeling complex *SMARCB1* (and, as a rare alternative, of *SMARCA4*) have been identified in the great majority of cases of ATRT ([@B4]). Such mutations appear necessary for the development of ATRT; no other consistent genetic abnormalities have been identified.\n\n1.2. ATRT in Adults {#S1-2}\n-------------------\n\nWhile ATRT is the most common malignant CNS tumor in children aged \\<1, cases in adults (i.e., age \\>18) are rare---we estimate \\<1/1,000,000 lifetime risk ([@B5]). As such, these patients are presented in detail as case reports or in small case series. There are no molecular markers recognized as unique to adults rather than children; in both cases, loss of INI1 staining via immunohistochemistry (IHC) is considered sufficient for diagnosis in the appropriate context.\n\n1.3. Rhabdoid Tumors {#S1-3}\n--------------------\n\nATRT is often described as analogous to \"malignant rhabdoid tumor of the kidney,\" a tumor recognized since at least the early 1980s ([@B6]). Other CNS tumors with rhabdoid histology include the \"epithelioid glioblastoma with rhabdoid component\" as well as the \"rhabdoid\" variants of more common tumors: meningioma, carcinoma, chordoma, and even sarcoma ([@B7]).\n\n1.4. EMT {#S1-4}\n--------\n\nEpithelial--mesenchymal transition (EMT) was initially identified as a phenomenon in developmental biology. The occurrence of EMT in certain cancers has been suggested since 1978, although it only began to receive widespread attention 10\u2009years later ([@B8]). The role of EMT in carcinogenesis has been the subject of an extensive review ([@B9]). EMT appears to be crucial to the development and maintenance of a pool of slowly dividing, chemoresistant, \"cancer stem cells.\" We hypothesize that the \"rhabdoid\" appearance of the tumors above reflects EMT.\n\n2. Case Reports {#S2}\n===============\n\n2.1. A Case in Pregnancy {#S2-1}\n------------------------\n\n### 2.1.1. Presentation {#S2-1-1}\n\nOur first patient was a 29-year-old woman who developed hoarseness during week 20 of her first pregnancy, which deteriorated over the subsequent 5\u2009weeks. Over the course of 1\u2009day, she developed weakness of the right arm and leg. A CT performed elsewhere showed the presence of an extra-axial left cerebello-pontine mass. This was interpreted as a probable meningioma, so further treatment was deferred.\n\nAt 35\u2009weeks gestation, she developed status migrainosus and worsening right leg weakness. Her voice had become weaker and she coughed intermittently when swallowing. Paralysis of her left vocal cord was demonstrated via laryngoscopy. Her MRIs at this time are shown in Figures [1](#F1){ref-type=\"fig\"} and [2](#F2){ref-type=\"fig\"}. A cesarean section was performed in anticipation of neurosurgery and a healthy baby boy was delivered.\n\n![First patient. MRI brain at time of diagnosis, 1.5\u2009T. Images are T1-weighted and Gd-enhanced. **(A)** Axial: heterogeneous extra-axial enhancement is seen in the left cerebello-pontine angle, with some local compression. **(B)** Coronal: enhancement is seen along the tentorium cerebelli on the right, which is characteristic of leptomeningeal spread. The left cerebellar lesion appears intraparenchymal.](fneur-08-00247-g001){#F1}\n\n### 2.1.2. Initial Treatment {#S2-1-2}\n\nA left retro-sigmoid craniotomy was performed. Pathology, showing loss of immunohistochemical staining for INI1, confirmed ATRT. Her spinal MRI showed leptomeningeal metastases in the cervical and thoracic cord. She rapidly developed hydrocephalus; a shunt and Ommaya reservoir were placed 1\u2009week after surgery.\n\nMultimodal treatment was instituted. Systemic chemotherapy involved two cycles (of 14\u2009days) of ICE (ifosfamide, carboplatin, etoposide). Following the second cycle, she developed pancytopenia and sepsis, and thereafter no further systemic treatment was pursued.\n\nIntrathecal chemotherapy was with liposomal cytarabine 50\u2009mg via Ommaya every 2\u2009weeks; she received 6 doses in total. Intensity-modulated radiotherapy to the cerebellum was 54\u2009Gy in 30 fractions; conformal radiotherapy to the cord at C5--T6 vertebral levels was 30\u2009Gy in 10 fractions. Disease in the brain and upper spinal cord improved overall with treatment, except for one focus in the pre-pontine cistern which grew in a plaque-like fashion.\n\n### 2.1.3. Treatment at progression {#S2-1-3}\n\nIn view of progressive weakness of the legs, she received an additional 30\u2009Gy in 10 fractions to the spinal cord at T12--L4. Within a month of completing this treatment, she developed a flaccid paraplegia with loss of bladder and bowel control and a sensory level at T10. Her MRI confirmed progression (Figure [2](#F2){ref-type=\"fig\"}). Thereafter, she was cared for by Hospice and died 5\u2009months following her initial surgery.\n\n![First patient. MRIs of spine, 1.5\u2009T, sagittal. **(A)** Gd-enhanced: diffuse, nodular enhancement is seen throughout. This is striking dorsally and at the conus medullaris. **(B)** T2-weighted: the thoracic cord is swollen with intra-medullary edema. This appears to be a reaction to the surrounding leptomeningeal tumor; there may also be an element of cord compression in the thoracic cord. **(C)** Gd-enhanced: disease in the upper thoracic cord is much improved following radiotherapy. Progression has occurred around the untreated cord caudally. **(D)** T2-weighted: similarly, intra-medullary edema has improved in the treated areas but worsened elsewhere.](fneur-08-00247-g002){#F2}\n\n2.2. A Case with a Durable Response to Treatment {#S2-2}\n------------------------------------------------\n\n### 2.2.1. Presentation {#S2-2-1}\n\nOur second patient was a 35-year-old man who presented with blurred vision that had been present for 3\u2009months. His MRI at this time, showing a suprasellar mass, is shown in Figure [3](#F3){ref-type=\"fig\"}.\n\n![Second patient. MRIs of brain. Upper row **(A--C)**: T1-weighted, Gd-enhanced (magnetization-prepared rapid gradient-echo (MPRAGE)). A suprasellar and interpeduncular lesion is seen with mixed solid and cystic (lobulated) elements. There is heterogeneous enhancement of intermediate intensity. Lower row **(A,B)**: T1-weighted, Gd-enhanced. There is no evidence of tumor recurrence. Smooth linear enhancement is present at the resection site, which is consistent with the long-term changes seen following surgery.](fneur-08-00247-g003){#F3}\n\nHe underwent craniotomy, via a pterional approach, for resection of the tumor. The pathology is shown in Figure [4](#F4){ref-type=\"fig\"}. This shows how ATRT may be suspected on the basis of hematoxylin and eosin (H&E) staining alone. The diagnosis was confirmed by IHC for INI1.\n\n![Second patient. Immunohistochemistry. **(A)** Hematoxylin and eosin (H&E) stain, 200\u00d7 magnification. The tumor comprises two main elements: sheets of large cells, many with a rhabdoid appearance (top) and small, more primitive-appearing cells (bottom). **(B)** (H&E) stain, 400\u00d7. This shows large cells, many with rhabdoid features and central mitoses. **(C)** INI1 stain, 400\u00d7. There is no nuclear staining in the tumor cells. A positive control is provided by the endothelial cells in the vessel at the center, and in other scattered glial and inflammatory cells.](fneur-08-00247-g004){#F4}\n\nA homozygous mutation in *SMARCB1* was confirmed on Sanger sequencing. This was performed on the fresh-frozen, paraffin-embedded tissue. The sequencing covered the coding region and 5\u2032 and 3\u2032 splice sites from nucleotide 37 to 47,120 in the reference sequence `NT_011520.gbk` ([@B10]). The mutation was described as `NM_003073.3: c.1148delC, NP_003064: p.P383fs*96`. Occurring in exon 9, this causes a *SMARCB1* protein frame-shift and the insertion of additional amino acids at the 3\u2032 end of the protein. This mutation is predicted to lead to inactivation of this protein in the tumor.\n\n### 2.2.2. Initial Treatment {#S2-2-2}\n\nMRIs performed following surgery demonstrated a nodule in the third ventricle that was believed to be persistent malignancy, as well as a nodule in the upper lumbar spinal cord, most likely a sign of leptomeningeal metastasis. Hence, cranio-spinal radiotherapy was given at a dose of 36\u2009Gy in 20 fractions. Localized boosts were given to the primary site (20\u2009Gy in 12 fractions) and the lumbar spine (8\u2009Gy in 4 fractions). This treatment was completed 3\u2009months following his initial surgery.\n\nHe proceeded to receive chemotherapy using the St. Judes ATRT protocol. This was developed for children aged \\>3 and comprises 4 cycles of cisplatin with high-dose cyclophosphamide and vincristine, with autologous peripheral blood stem-cell transplantation ([@B11]). This took 5\u2009months to complete.\n\nSince completing treatment there has been no sign of residual/recurrent disease at 2.5\u2009years follow-up, clinically or on MRI, as seen in Figure [3](#F3){ref-type=\"fig\"}.\n\nSince his initial resection, he has suffered from panhypopituitarism, a left homonymous hemianopia and short-term memory impairment. He also developed a peripheral neuropathy following chemotherapy, most likely due to vincristine.\n\n3. Methods {#S3}\n==========\n\n3.1. Review of Adult Cases {#S3-1}\n--------------------------\n\nA literature search was performed using PubMed and Google Scholar to search for all reported cases of ATRT in adults. All 50 cases were tabulated; this is given as Supplementary Material (worksheet d1 in `data sheet 1.xlsx`). Many of these reports include similar tables; however, we noted a number of inconsistencies in these tables and ours is based on our reading of the original cases. While many prior authors have reported the 'classical' features of ATRT via IHC (see below), we are the first to tabulate *all* pathological variables for each case.\n\n### 3.1.1. Clinical Variables {#S3-1-1}\n\nIn each case, we recorded age, gender, and location as well as presence of leptomeningeal metastases (LM) and details of treatment at initial diagnosis and time of first progression. As above, a meaningful response to treatment was rare if the disease progressed again after \"second-line\" approaches. 'Third-line' treatments of any kind were exceptional and were not recorded.\n\nThe diagnosis of LM was based on clinical details and/or imaging features, e.g., in the case report of Wang et al. recurrence was spinal (distal from the original tumor) and appears to have originated from the leptomeninges on MRI, with subsequent intra-medullary spread ([@B12]).\n\nA number of nominal (categorical) variables were 'collapsed' as this was such a small data set. Location, for example, initially had 15 values; we also analyzed this as a 5- and 3-category variable in order to look for more general patterns. We also assessed whether the location was 'central' vs. clearly lateralized and whether the tumor was \"next to CSF\" based on imaging findings. Similarly, the extent of resection was considered as an ordinal variable (gross total\u2009\\>\u2009sub-total resection\u2009\\>\u2009biopsy) as well as binary (surgery or not). We analyzed age as binary (i.e., \u226540\u2009years when diagnosed) as well as a continuous numeric variable.\n\n### 3.1.2. Pathological Variables {#S3-1-2}\n\nWe consider here only the pathology at the time of diagnosis of ATRT. We note that only 2/40 (5%) of patients in our series had an autopsy performed (where we could infer this information). A summary table of all pathologic results is given as Supplementary Material (worksheet ihc1 in `data sheet 1.xlsx`). Most of these variables relate to the presence of certain antigens via IHC.\n\nFor IHC, a variety of antibodies and names were used to assess the presence of various elements. We have standardized this by recording the antigen being tested. Keratin staining, in particular, used a range of antibodies and was variously reported by antibody, by weight of the stained keratin, by keratin type or at times as \"keratin present.\"\n\n3.2. Classification of IHC {#S3-2}\n--------------------------\n\n### 3.2.1. Diagnosis of ATRT {#S3-2-1}\n\nLoss of INI1 protein expression (via IHC) is considered a defining feature of ATRT. Partial deletion of chromosome 22 (detected via FISH), where the gene (*SMARCB1*) is located, is considered as an equivalent test. Both of these tests began to be used widely after the year 2000. While both tests are considered to have a specificity of 100% for ATRT (in the appropriate setting), a negative result does not exclude the diagnosis. Such testing was reported in 23/50 (46%) of the cases we reviewed.\n\nPrior to the advent of INI1 testing, the diagnosis was suspected on the basis of the appearance on standard H&E staining. Staining for the \"classical\" antigens vimentin, epithelial membrane antigen (EMA), and smooth muscle actin (SMA) was typically used as confirmatory. Both vimentin and SMA are classically associated with *mesenchymal* cells, although this does not imply a *mesodermal* origin for such cells. For example, constitutive expression of vimentin is recognized in arguably the most-studied *glial* cell line, U-251 ([@B13]). Again, a negative test result for one of these antigens does not exclude the diagnosis.\n\n### 3.2.2. IHC and Differential Diagnosis {#S3-2-2}\n\nWhile the diagnosis of ATRT is often quickly suspected in an infant, the initial differential diagnosis in adults is typically far wider. Thus, the range of IHC stains employed in adult cases is far greater and more variable. In total, 81 tests (using IHC) were performed in our case series. In 19 (23%) cases, the antigen of interest was tested for in just one specimen.\n\nIHC often included tests for metastasases, germ cell tumors, neuro-endocrine tumors, and other primary tumors of the nervous system (glial, medulloblastoma, neuroblastoma). Testing for a wider range of antigens (in adults vs. children) gives us a greater insight into the pathophysiology of ATRT.\n\nOther possible differential diagnoses in adults may include primitive neuroectodermal tumor (PNET), choroid plexus carcinoma, rhabdoid meningioma, and germinoma. However, these entities generally have more characteristic immunohistochemical staining patterns with specific markers than the poly-phenotypic patterns seen in ATRT; also INI1 staining is retained.\n\n### 3.2.3. Differential: Rhabdoid Glioblastoma {#S3-2-3}\n\nTumors resembling ATRT, staining with glial fibrillary acidic protein (GFAP) as well as vimentin, SMA and EMA have been suggested to represent rhabdoid glioblastoma (GB) rather than ATRT ([@B7]). EMT in the setting of GB appears plausible as an entity distinct from ATRT. GB with EMT *may* have been misdiagnosed as adult ATRT (prior to the advent of testing for INI1), although this appears unlikely given the clinical and histological findings in our cases.\n\n### 3.2.4. Germ Cell Layers {#S3-2-4}\n\nWe used IHC to determine the putative germ cell layer of origin of each specimen. In so doing, we were guided by the tissues that are known to stain for an antigen under \"normal\" conditions. The website accompanying the textbook \"Pathology Outlines\" was particularly helpful in this regard ([@B14]).\n\nSome stains are not restricted to one germ cell layer. An example is keratin 8, which predominantly occurs in ectodermal tissue but is also recognized in endodermal tissue ([@B15]).\n\n### 3.2.5. Classification {#S3-2-5}\n\nWe thus classified pathological findings into the following groups: 'ATRT specific''Defining' features (loss of INI1 or del(22))'Classical' features (vimentin, EMA, SMA)EctodermNeural crest (neuronal, glial, neuro-endocrine, melanoma markers)Other ectodermalEndodermKeratinsMesodermMesenchymalLeukemia/lymphoma markersGerm cellOthers (found in a variety of tissues).\n\n3.3. Quantitative Analysis {#S3-3}\n--------------------------\n\nData analysis was performed using R (RRID:SCR_001905) ([@B16]--[@B18]). The complete analysis is available as Supplementary Material (`data sheet 2.pdf`). Nominal variables are given as fractions (percentage) and continuous variables as median (range).\n\nWe did not attempt to correct for multiple hypothesis testing, as the present work is exploratory ([@B19]). As the sample size is small, *a priori* we considered p values of \\<0.1 to be potentially worth reporting.\n\n### 3.3.1. Associations {#S3-3-1}\n\nWe began by looking at the significance of *all* two-variable associations, using standard measures of strength and significance of association as follows:\n\n  ------------ ------------ ---------------- -------------\n  Variable 1   Variable 2   Correlation      p Value\n  Nominal      Nominal      Cramer's V       Chi-squared\n  Numeric      Numeric      Pearson's r^2^   t-Test\n  Nominal      Numeric      Pearson's r^2^   F-Test\n  ------------ ------------ ---------------- -------------\n\n### 3.3.2. Survival analysis {#S3-3-2}\n\nTime-to-event data were analyzed used the proportional-hazards model (PHM). We looked at the following outcomes: time to progressionoverall survivaltime from progression to death (to assess the effects of second-line treatments).\n\nGiven the small number of observations and deaths, over-fitting was a problem with multivariate PHMs.\n\nSurvival times are reported as a median with 95% confidence intervals (CI). Effect size is given as hazard ratio (HR, relative to control) and the p value is from the score test (or Wald test for multivariate models).\n\n### 3.3.3. Missing Data {#S3-3-3}\n\nGiven such a large quantity of missing data, we tried to compensate in a number of ways: Recursive partitioning 'looks ahead' at survival to find the best split for the given data. It is particularly well suited to data sets where there is near-perfect prediction of outcome by certain variables. It also allows for a clear ordering of variables in terms of predictive importance ([@B20]).\"Multivariate imputation by chained equations\" uses repeated regression on the existing variables to estimate missing values. The technique does have a tendency to strengthen existing associations rather than developing new insights ([@B21]).The \"nested cohort\" design allows for more 'accurate' estimates from a PHM where a covariate is observed for only some subjects (the cohort). Sampling is stratified by a variable that is available on all cohort members. This allows for frequency matching on confounders, or oversampling on the extremes of surrogate \\[markers\\] for exposures, to improve efficiency ([@B22]).That is, this technique allows one to control for major confounders in order to arrive at a more meaningful estimate of effect size and probability. In our case, we sought to better determine the effects of various treatments using a sampling scheme based on:the presence of LM or GFAP staining---as these were the major confounders of treatmentthe occurrence of death when the last recorded follow-up was performed---to try to control for censoring, i.e., the fact that certain patients were not followed up until death had occurred.\n\n4. Results {#S4}\n==========\n\n4.1. Demographics {#S4-1}\n-----------------\n\nThe 50 patients reviewed (including the present cases) had a median age of 32 (18--65); 27 (54%) were females. Of those of childbearing age, 3/20 (15%) were pregnant when the diagnosis was made.\n\n4.2. IHC {#S4-2}\n--------\n\nSome of the more significant findings on IHC are given in Table [1](#T1){ref-type=\"table\"}. As above, the complete table, with explanations, is given in (`data sheet 1.xlsx`).\n\n###### \n\nImmunohistochemistry (IHC) of adult ATRT.\n\n  Stain                             Long name                                     +ve   n    \\%\n  --------------------------------- --------------------------------------------- ----- ---- -----\n  **'Defining' features of ATRT**                                                            \n  del(22)(q)                        22q11.2 distal deletion                       13    15   87\n  INI1                              Integrase interactor 1/SMARCB1/hSNF5          1     23   4\n  **'Classical' ATRT markers**                                                               \n  vim                               Vimentin                                      33    33   100\n  EMA                               Epithelial membrane antigen                   29    35   83\n  SMA                               Smooth muscle actin                           18    32   56\n  **Neuronal**                                                                               \n  S100                              100% soluble in (NH4)2-SO~4~                  13    27   48\n  nfp                               Neurofilament protein                         10    15   67\n  nse                               Neuron-specific enolase                       2     6    33\n  Nestin                                                                          2     2    100\n  **Glial**                                                                                  \n  GFAP                              Glial fibrillary acidic protein               12    30   40\n  **Neuro-endocrine**                                                                        \n  Synaptophysin                     Major synaptic vesicle protein p38            6     22   27\n  **Keratin**                                                                                \n  ker                               Keratin (any, including type not specified)   20    49   41\n  ck8                               Keratin 8; 52\u2009Da                              12    16   75\n  **Mesenchymal**                                                                            \n  Desmin                                                                          0     19   0\n  Myoglobin                                                                       0     4    0\n  **Others**                                                                                 \n  CD34                              Hematopoietic progenitor cell antigen CD34    3     11   27\n  CD99                              Single-chain type-1 glycoprotein              2     7    29\n  p53                               Cellular tumor antigen p53                    4     5    80\n\n*+ve, positive result, i.e., staining for antigen present; n, number of cases tested; %, percentage positive/number tested*.\n\nDefining features of ATRT were found in almost all cases where these were assessed: loss of INI1 in 23/24 (96%), del22q in 13/15 (87%). It is tempting to consider those cases where these were absent as \"false negatives\"; we will return to this in the discussion.\n\nOf the \"classical\" features, only vimentin was universally positive (33/33) in our series. Germ cell markers, where assessed, were uniformly negative. Keratins were variably expressed; keratin 8 was found in the highest proportion of samples assessed, 12/16 (75%). As above, we cannot use this help attribute a germ cell layer of origin to a specimen ([@B15]).\n\n### 4.2.1. Nestin {#S4-2-1}\n\nMarkers of neuronal, glial, and neuro-endocrine differentiation were relatively common (*c*. 25--75%). Of particular interest is the presence of nestin in 2/2 cases studied. This is considered a marker of 'neuroepithelial stem cells,' is active in embryogenesis, and is downregulated in maturity, when expression of neurofilament protein and GFAP occur.\n\nInterestingly, nestin is also expressed in developing muscle (presomitic mesoderm), to be replaced by desmin in maturity. Yet, desmin was universally absent in our sample (0/19). Other markers of mature muscle differentiation were also lacking, as well as the myogenic regulatory factor MYOD1 (0/2 cases), which is involved in myogenesis.\n\n### 4.2.2. CD34 {#S4-2-2}\n\nThe hematopoietic system is thought to be derived from mesoderm; as such the presence of markers more typically used in the diagnosis of leukemia and lymphoma in our sample may at first appear anomalous. However, CD34 is recognized to occur in tumors of neuroepithelial origin, particularly those of childhood ([@B23]). It is also expressed on endothelial cells.\n\n### 4.2.3. CD99 {#S4-2-3}\n\nCD99 has been less well studied in CNS tumors. It is recognized to occur in \"primary neuroepithelial tumors of the kidney,\" an entity we believe to be closely related to ATRT ([@B24]). It is also present in ependymoma, peripheral primitive neuroectodermal tumor, and in the rare chordoid glioma, all of which are of neuroepithelial origin ([@B25]--[@B27]).\n\n4.3. Associations {#S4-3}\n-----------------\n\nA number of striking associations are given in Table [2](#T2){ref-type=\"table\"}. Additional relationships of significance statistically but of debatable importance biologically are given as in the Supplementary Material as `data sheet 2.pdf`.\n\n###### \n\nSignificant (*p*\u2009\\<\u20090.1) two-variable associations.\n\n                                                                                                                               Chi-square p   \n  --------------------------------------------- ------------------------------- ----------------------- ----------------- ---- -------------- -------\n  Male                                          4                               13                      2                 0    3              0.046\n  Female                                        3                               7                       4                 10   3              \n                                                                                                                                              \n  **Surgery, radiation, and chemotherapy**                                                                                                    \n  No                                            3                               14                      0                 6    4              0.007\n  Yes                                           3                               3                       5                 4    0              \n                                                                                                                                              \n  **Gender**                                    **Lateral**                     **Mid-line**                                                  \n                                                                                                                                              \n  Male                                          15                              7                                                             0.043\n  Female                                        9                               17                                                            \n                                                                                                                                              \n  **Age**                                       **Brainstem**                   **Lateral ventricle**   **Spinal cord**                       \n                                                                                                                                              \n  18--40                                        12                              22                      1                                     0.004\n  \\>40                                          6                               4                       5                                     \n                                                                                                                                              \n  **Leptomeningeal metastases at recurrence**                                                                                                 \n  No                                            2                               16                      2                                     0.039\n  Yes                                           6                               5                       1                                     \n                                                                                                                                              \n  **OTHER ASSOCIATIONS**                                                                                                                      \n                                                                                                                                              \n                                                **Leptomeningeal metastases**                                                                 \n  **GFAP**                                      **No**                          **Yes**                                                       \n                                                                                                                                              \n  \u2212ve                                           16                              2                                                             0.018\n  +ve                                           5                               7                                                             \n                                                                                                                                              \n                                                **Synaptophysin**                                                                             \n  **SMA**                                       **\u2212ve**                         **+ve**                                                       \n                                                                                                                                              \n  \u2212ve                                           9                               0                                                             0.031\n  +ve                                           5                               6                                                             \n                                                                                                                                              \n                                                **Neurofilament protein**                                                                     \n  **Gender**                                    **\u2212ve**                         **+ve**                                                       \n  Male                                          5                               4                                                             0.094\n  Female                                        0                               6                                                             \n\n*GFAP, glial fibrillary acidic protein; SMA, smooth muscle actin*.\n\n### 4.3.1. Demographics and Location {#S4-3-1}\n\nWhere the tumor was in or next to the pituitary, 10/11 cases were in females (91%). The predilection of adult ATRT for this location has been recognized, although the sex-specificity seen here is novel ([@B28]).\n\nTumors of the spinal cord were associated with increasing age (Pearson's r\u2009=\u20090.43, F test p\u2009=\u20090.002).\n\n### 4.3.2. IHC {#S4-3-2}\n\nTumors with LM when diagnosed were much more likely to show GFAP staining (n\u2009=\u200930, chi-squared p\u2009=\u20090.02), implying that glial differentiation is a major risk factor for this complication.\n\nAll tumors that showed synaptophysin staining also stained for SMA (n\u2009=\u200920, chi-squared, p\u2009=\u20090.03). Synaptophysin was the only neuroendocrine marker commonly assessed in the sample; this finding suggests that differentiation along such lines is associated with a greater tendency to manifest a mesenchymal phenotype.\n\nFindings regarding neural differentiation were difficult to interpret. Neurofilament protein (NFP) staining was present in all female patients (6/6) vs. 4/9 males (chi-square p\u2009\\<\u20090.1). Strangely, staining for NFP and neurospecific enolase (NSE) appeared mutually exclusive, i.e., where one was positive the other was negative, although there were only 4 subjects where both were checked. No cases where NFP staining was absent stained for GFAP (n\u2009=\u200915).\n\n4.4. Treatment {#S4-4}\n--------------\n\n### 4.4.1. Surgery {#S4-4-1}\n\nSurgery (SX) at diagnosis was a cornerstone of treatment; in only one case was this not undertaken, where imaging initially suggested vestibular schwannoma and radiotherapy was used instead ([@B12]). Gross total resection (GTR) was reported in 10/43 (23%) of cases. This was most common for those tumors located next to a lateral ventricle. Subtotal resection was most common for tumors of the pituitary and pineal gland.\n\n### 4.4.2. Adjuvant Treatment {#S4-4-2}\n\nRadiotherapy (RT) was undertaken in 33/42 (79%) and systemic chemotherapy (CT) in 17/42 (40%). Cranio-spinal irradiation (CSI) was performed in 5/34 (15%). All three approaches (SX, RT, and CT) were used in 15/42 cases (36%). This multimodality approach was most commonly used for those tumors next to a lateral ventricle (Table [2](#T2){ref-type=\"table\"}).\n\nIntrathecal CT was used in just two cases other than our first patient. In contrast to our use of liposomal cytarabine, these two patients were treated with small quantities of methotrexate: three and one dose(s), respectively ([@B29], [@B30]).\n\n### 4.4.3. Treatment at recurrence {#S4-4-3}\n\nWhen disease re-occurred/progressed, lower proportions of patients received treatment. Additional SX was attempted in 12/26 (46%), RT in 13/26 (50%) and CT in 11/26 (42%).\n\n4.5. Survival {#S4-5}\n-------------\n\nThe median time to progression (TTP) was 5\u2009months (95% CI 3--18). Overall survival (OS) was 23\u2009months (14--56). Time from progression until last observation or death (time from progression to death, TPD), in the 27 patients where this was applicable and available, was 8\u2009months (5--28). The fractions of patients without progression, and surviving, at various points in time from diagnosis are shown in Table [3](#T3){ref-type=\"table\"}.\n\n###### \n\nPercentages of patients without progression, and surviving, at timepoints following diagnosis.\n\n  Time from diagnosis   Progression-free (%)   Surviving (%)\n  --------------------- ---------------------- ---------------\n  6\u2009months              46                     76\n  12\u2009months             33                     64\n  2\u2009years               20                     39\n\n### 4.5.1. Long-term Survivors {#S4-5-1}\n\nApart from our second patient, two others are reported to have survived for more than 3\u2009years ([@B31], [@B32]). Unlike our patient, both re-occurred on multiple occasions following their initial treatment. The first was a frontal lesion that required complete excision 6 times (and once in the contralateral frontal lobe) as well as receiving radiotherapy, Gamma Knife radiosurgery and systemic chemotherapy; after 7\u2009years of relapsing disease, she remained disease free for another 10\u2009years. The other was a parietal lesion which re-occurred 3 times over the course of 9\u2009years and which appeared stable 1\u2009year after most recent treatment. Interestingly, this patients' tumor was diagnosed and treated as glioma, at presentation and at first re-occurrence. The diagnosis of ATRT was subsequently made in retrospect due to the absence of INI1 by IHC in all specimens.\n\n### 4.5.2. Clinical Variables and IHC {#S4-5-2}\n\nSignificant predictors of survival in PHMs are shown in Table [4](#T4){ref-type=\"table\"}. Of note, LM and GFAP staining appear as significant predictors in almost all of these models. While LM and GFAP are correlated, LM appears to be the more important predictor, particularly in multivariable PHMs. The effect of LM on OS is shown in Figure [5](#F5){ref-type=\"fig\"}.\n\n###### \n\nLeptomeningeal metastases and GFAP staining are significant predictors in most proportional-hazards models.\n\n                      Variable    n    HR        p\n  ------------------- ----------- ---- --------- --------\n  **Uni-variable**    **Score**                  \n                                                 \n  TTP                 RT          33   0.13      \\<0.01\n                      SX          33   Ordinal   0.02\n                      GFAP        22   1.9       0.19\n                      LM          33   2.4       0.04\n                      GTR         33   0.86      0.75\n  OS                  LM          41   4.6       \\<0.01\n                      GFAP        27   3.1       0.04\n                      SX          40   Ordinal   \\<0.01\n                      SRC         40   0.25      0.02\n                      Pregnant    18   4.5       0.05\n                      GTR         39   0.55      0.3\n  TPL                 LM          27   6.7       \\<0.01\n                      GFAP        18   7.5       \\<0.01\n                                                 \n  **Multivariable**   **Wald**                   \n                                                 \n  TTP                 LM               4.4       0.01\n  n\u2009=\u200921, e\u2009=\u200914      RT          21   0.32      0.13\n                      GTR              0.41      0.20\n                      GFAP             2.3       0.18\n  OS                  LM          27   5.5       0.02\n  n\u2009=\u200927, e\u2009=\u200915      GFAP             2.4       0.25\n                      GTR              0.27      0.07\n                      CT               0.24      0.03\n                      RT               1.5       0.6\n  TPL                 LM          18   3.7       0.4\n  n\u2009=\u200918, e\u2009=\u200914      GFAP             2.4       0.6\n\n*p values: for uni-variable models\u2009=\u2009score test; for multivariate models\u2009=\u2009Wald test*.\n\n*HR, hazard ratio; TTP, time to progression; OS, overall survival; TPL, time from progression to last follow-up/death; n, number of observations; e, number of events (progression or death); RT, radiotherapy; SX, surgery (ordinal; HRs not shown); GFAP, glial fibrillary acidic protein; LM, leptomeningeal metastases; GTR, gross total resection (vs. other value for surgery); SRC, surgery, radiation, and chemotherapy; CT, chemotherapy*.\n\n![Overall survival is significantly worse in patients with leptomeningeal metastases (LM) at the time of diagnosis. Time scale is shown in months. Hazard ratio 4.4, score test p\u2009=\u20090.0005.](fneur-08-00247-g005){#F5}\n\n### 4.5.3. Treatments {#S4-5-3}\n\n\"Over-fitting\" was a problem when using surgery as an ordinal-scale variable in multivariate models. Instead, we give this as \"gross total resection?\" (GTR, yes/no). Regarding OS, only SX and multimodality treatment (SX, RT, and CT, HR\u2009=\u20090.3) were significant predictors. Both GTR and CT were significant predictors in the multivariable PMH for OS.\n\nIn assessing the effects of treatment, we acknowledge that we are not \"comparing like with like,\" i.e., patients for whom a certain treatment is not possible will be expected to have a worse outcome. Regarding surgery, GTR is generally the goal, where possible; those for whom this is not feasible will tend to have a worse prognosis, no matter the approach to surgery. Similarly, those for whom CT is impossible will tend to have poorer survival outcomes. However, significant comorbidities or disabilities due to disease (thus limiting the use of CT) were exceptional at the time when CT was commenced.\n\n4.6. Recursive Partitioning {#S4-6}\n---------------------------\n\nThe most important predictor of OS was clearly LM; this is shown in Figure [6](#F6){ref-type=\"fig\"}. When LM was excluded from the analysis, multimodality therapy (SX, RT, and CT) became the most significant variable (Supplementary Material, `data sheet 2.pdf`).\n\n![Recursive partitioning analysis shows LM to be the most important predictor of overall survival. Abbreviations: LM, leptomeningeal metastases; SRC, surgery, radiation and chemotherapy (\u22650.5 means all modalities employed); Sx, Surgery; gtr, gross total resection; sx, surgery (extent unspecified); str, subtotal resection; bx, biopsy; none, no surgery. Key: circles, branching nodes; squares, terminal nodes; green, better outcomes; pink, worse outcomes; upper, no. predicted event rate; lower, no.\u2009no. events/no. at risk.](fneur-08-00247-g006){#F6}\n\n4.7. Multiple Imputation and \"Nested\" PHM {#S4-7}\n-----------------------------------------\n\nApplying these techniques to the potentially important predictors already identified tended to confirm the results above (`data sheet 2.pdf`).\n\nThe effects of the most common initial treatments on OS, when controlling for LM and GFAP (and death, in both cases), are shown in Table [5](#T5){ref-type=\"table\"}.\n\n###### \n\nChemotherapy is the most important predictor of overall survival, when controlling for leptomeningeal metastases and GFAP staining (using 'nested cohort' proportional hazards models).\n\n                   Variable   HR     p\n  ---------------- ---------- ------ --------\n  LM               CT         0.45   0.09\n  n\u2009=\u200939, e\u2009=\u200924   RT         0.56   0.3\n                   GTR        0.43   0.3\n  GFAP             CT         0.31   \\<0.01\n  n\u2009=\u200927, e\u2009=\u200915   RT         0.64   0.5\n                   GTR        0.31   0.1\n\n*HR, hazard ratio; GFAP, glial fibrillary acidic protein; LM, leptomeningeal metastases; n, number of observations; e, number of events (deaths); CT, chemotherapy; RT, radiotherapy; GTR, gross total resection (vs. other value for surgery)*.\n\n5. Discussion {#S5}\n=============\n\n5.1. Pathogenesis {#S5-1}\n-----------------\n\n### 5.1.1. Location at Diagnosis and Implications for Anatomical Origin {#S5-1-1}\n\nATRT in adults has a predilection for mid-line structures, particularly the pineal and pituitary glands. These are circumventricular organs, which are now recognized as a source of neural stem cells in adults ([@B33]).\n\nATRT may be said to be the prototypical tumor of infancy, in having the earliest onset and highest perinatal incidence of such tumors. Tumors of infancy in general are thought to arise from cells that have not completed the process of terminal differentiation and continue to undergo hyperplasia from the latter half of pregnancy to age 3 or so ([@B34]).\n\nOur case series suggests that there remains a pool of slowly dividing 'ectodermal stem cells' in at least some of these circumventricular organs that remains capable of acquiring carcinogenic mutations throughout most of adulthood ([@B35]). The proximity of these organs to CSF may explain the close association of these tumors with such locations and their tendency to develop LM. Of note, when ATRT occurs in a cerebral lobe, it typically appears in communication with a lateral ventricle.\n\nWe suspect that in many cases the tumor has already spread via CSF prior to diagnosis. This would help to explain the proposal of such entities as \"primary diffuse cerebral leptomeningeal ATRT\" and \"ATRT arising from the acoustic nerve\"; these structures appear unlikely to be points of origin for ATRT ([@B12], [@B36]).\n\nThe strong association between pituitary involvement and female gender is consistent with the greater mitotic activity of the pituitary in females throughout the lifespan ([@B37]). We acknowledge there remains much work to be done in this area.\n\n### 5.1.2. IHC and Implications for Cell Type of Origin {#S5-1-2}\n\nAs summarized in Table [1](#T1){ref-type=\"table\"}, findings on IHC suggest a tumor of ectoderm origin with EMT. For example, the combination of EMA (epithelial, 83%), vimentin (mesenchymal, 100%), and complete absence of markers of mature mesenchyme (desmin, myoglobin, 0%) are characteristic of such a phenotype.\n\nIn this regard, the name \"ATRT\" is somewhat misleading in that it is clearly not a teratoma and does not contain elements originating from mesoderm or showing skeletal muscle differentiation.\n\nClassification of IHC by cell type and germ layer in adult ATRT has also been undertaken by Raisanen et al. ([@B38]). Their table is similar to our own, albeit with a smaller sample size. The authors do not attempt to infer any implications for pathogenesis from their table.\n\n### 5.1.3. Germline Mutations and del22q {#S5-1-3}\n\nLoss of INI1 is not sufficient to cause ATRT. This is shown by the lack of complete penetrance of ATRT in germline mutations (distal 22q11 microdeletion syndromes) ([@B39]). None of the cases of ATRT in this series, including our own, had germline analysis performed to look for a mutations or deletion of chromosome 22q. The only case we reviewed where this seemed likely was our case with a prolonged response to treatment. The patent had macrocephaly and a history of spinal schwannoma as a teenager. None of the other reported a personal or family history of dysmorphism or of tumors seen in the rhabdoid-predisposition syndrome.\n\nNonetheless, it has been recognized that germline mutations in *SMARCB1* germline mutations are not always inherited and unaffected adult carriers are recognized ([@B40]). In this series of 100 patients, those with germline mutations tended to present at a much younger age, at a median of 5\u2009months (range birth---5\u2009years) vs. 18\u2009months (birth---17\u2009years) for those with acquired mutations. However, in 4 of 7 cases of inherited mutations, the parents appeared completely unaffected. The authors also recognize germline/gonadal mosaicism in their sample, i.e., the mutation must have been present within parental gametes but not somatic cells.\n\nSuch germline mutations result in a phenotype known as \"rhabdoid predisposition syndrome.\" In addition to ATRT, carriers appear predisposed to rhabdoid tumors (renal and extra-renal), PNET, medulloblastoma, choroid plexus carcinoma and schwannoma ([@B41]). These authors also suggest the mutation as a cause for cases of Li--Fraumeni-like syndrome (patients with a childhood cancer or \"sarcoma\" aged \\<45 and a family history of cancer at a young age, in those lacking germline *TP53* mutations).\n\nThus, we recommend somatic analysis in addition to analysis of the tumor, even in adult patients ([@B42]). Analysis of parental somatic and germline DNA also appears desirable.\n\n### 5.1.4. INI1, del22q, and Epigenetic Subgroups {#S5-1-4}\n\nThe only consistent genetic change in these tumors appears to be loss of INI1 due to *SMARCB1* mutations (which are variable) ([@B4], [@B43]). Mutation of *SMARCA4* is recognized as an alternative inciting event in rare cases, perhaps accounting for the single patient in our series with preserved INI1 expression. The lack of other consistent candidate oncogenic mutations has led to the recognition that epigenetic changes are crucial to pathogenesis. Three such epigenetic subgroups are now recognized ([@B44]). There were no adults in this recent work characterizing subgroups (age range: birth to 9.5\u2009years). The subgroups are named based on the pathways most commonly upregulated, here shown with some of the genes whose expression is characteristically increased:\n\n  ---------- ----------------------- -----------------------------------------------------------------\n  ATRT-TYR   Melanogenesis pathway   *EZH2*, DNA methyltransferases (*DNMT*s), *CCND1, VEGFA, ERBB2*\n  ATRT-SHH   Hedgehog pathway        *EZH2, DNMT*s, *CDK6*\n  ATRT-MYC   *MYC* pathway           *MYC, HOX* genes, *EZH2, DNMT*s, *ERBB2*\n  ---------- ----------------------- -----------------------------------------------------------------\n\nPhenotypically, the ATRT-TYR subgroup is strongly associated with a supratentional location vs. ATRT-MYC; ATRT-SHH appears in both locations.\n\nFurther characterization of phenotype by subgroup may be facilitated by preserving the methylation status of DNA in the tumor specimen. We suggest that fresh-frozen material be obtained for this purpose where practical. Technical advances continue to make methylation profiling more practical in formalin fixed, paraffin-embedded samples.\n\n### 5.1.5. Chromatin Remodeling and EMT {#S5-1-5}\n\nEMT is known to result in a more 'open' chromatic structure ([@B9]). This is necessary to facilitate de-differentiation, i.e., the transcription of mRNA, which is typically unavailable to the differentiated epithelial cell. This includes a shift in energy (ATP) production from oxidative phosphorylation to glycolysis. This is necessary for the supply of energy of biosynthetic precursors, of a balanced redox status and appears necessary to maintain the undifferentiated state.\n\nChromatin remodeling complexes such as INI1 have been implicated in various cancers ([@B45]). INI1 is now recognized as a tumor suppressor, due to its mutation in tumors including choroid plexus carcinoma, medulloblastoma, primitive neuroectodermal tumor, and chronic myeloid leukemia. INI1 is a member of the SWI2/SNF2 remodeling complex, which is ATP dependent.\n\nLoss of function of these chromatin remodeling complexes enables the development of the 'open' structure required for EMT. It remains unclear why loss of INI1 leads to diverging epigenetic phenotypes.\n\n5.2. Clinical Trials {#S5-2}\n--------------------\n\nOngoing and recently completed trials for this condition are tabulated in the recent review by Fruhwald et al. ([@B3]). The dearth of clinical trials enrolling adults \\>21\u2009years old, just seven at the time of writing, is shown in Table [6](#T6){ref-type=\"table\"}.\n\n###### \n\nOngoing clinical trials enrolling patients aged \\>21.\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  NCT ID                                         Agent                                          MOA            CA                    MC    Age     Phase   n     Cohorts               \n  ---------------------------------------------- ---------------------------------------------- -------------- --------------------- ----- ------- ------- ----- --------------------- -------------\n  <http://clinicaltrials.gov/show/NCT02601950>   Tazemetostat                                   EZH2 I         No                    Yes   \\>16    2       150   rhabdoid:             ATRT\\\n                                                                                                                                                                                       MRT\\\n                                                                                                                                                                                       RTK\n\n  Refractory synovial sarcoma\\                                                                                                                                                         \n  + SS18-SSX rearrangement                                                                                                                                                             \n\n  INI1 \u2212ve:                                      EMPNST\\                                                                                                                               \n                                                 EMC\\                                                                                                                                  \n                                                 Myoepithelial Ca\\                                                                                                                     \n                                                 Chordoma                                                                                                                              \n\n  RMC\\                                                                                                                                                                                 \n  Epithelial sarcoma                                                                                                                                                                   \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT01505569>   1--3 cycles:\\                                  Autologous\\    Yes                   No    \\<70    SOC     20    r/r solid tumor       \n                                                 Ifosfamide\\                                    peripheral\\                                                                            \n                                                 Etoposide\\                                     blood\\                                                                                 \n                                                 \u2192Carboplatin\\                                  stem cell\\                                                                             \n                                                 \u2192Thiotepa                                      transplant                                                                             \n\n  CNS tumor:                                     ATRT\\                                                                                                                                 \n                                                 PNET age \\<3\\                                                                                                                         \n                                                 High risk MB age \\<3\\                                                                                                                 \n                                                 MB age \\<8\u2009months\\                                                                                                                    \n                                                 Anaplastic MB                                                                                                                         \n\n  Germ cell tumor                                                                                                                                                                      \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT00445965>   ^131^I-3F8 Ab                                  Ganglioside\\   No                    No    Any     2       131   CNS/LM tumor\\         \n                                                                                                GD2                                                              GD2 +ve               \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT00983398>   Intra-arterial:                                               Yes                   No    1--30   1/2     55    r/r CNS\\              ATRT\\\n                                                                                                                                                                 embryonal             PNET\n\n  Melphalan\\                                     Alkylating\\                                    tumor:         MB\\                                                                     \n  Carboplatin\\                                   DNA repair\\                                                   Medulloepithelioma\\                                                     \n  Mannitol\\                                      Osmotic\\                                                      Pineoblastoma\\                                                          \n  Thiosulfate                                    Adsorbent                                                     Ependymoblastoma                                                        \n\n  Germ cell tumor                                                                                                                                                                      \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT02114229>   Alisertib                                      Aurora A I     No                    Yes   \\<22    2       180   ATRT (with chemoTx)   \n\n  r/r:                                           ATRT\\                                                                                                                                 \n                                                 MRT                                                                                                                                   \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT02684071>   Intraventricular methotrexate with systemic:                  Yes                   No    \\<22    2       10    r/r:                  ATRT\\\n                                                                                                                                                                                       PNET\n\n  Topotecan\\                                     Topoisomerase I\\                               MB\\                                                                                    \n  Cyclophosphamide                               Alkylating                                     Ependymoma                                                                             \n\n                                                                                                                                                                                       \n\n  <http://clinicaltrials.gov/show/NCT02458339>   Methotrexate via 4th ventricle                 Anti-folate    Yes                   No    1--21   1/2     18    r/r:                  ATRT\\\n                                                                                                                                                                                       PNET\\\n                                                                                                                                                                                       MB\\\n                                                                                                                                                                                       Ependymoma\\\n                                                                                                                                                                                       CPCa\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n*NCT ID, National clinical trials identifier; CA, commercially available; MOA, mode of action; MC, multi-center (i.e., \u22653 sites); n, number of patients (estimated accrual); Ab, antibody; I, inhibitor; EZH2, enhancer of zeste homolog 2 (enzyme); SOC, standard of care/protocol; r/r, recurrent/refractory (progressive); MRT, malignant rhabdoid tumor; RTK, rhabdoid tumor of the kidney; EMPNST, epithelioid malignant peripheral nerve sheath tumor; EMC, extra-skeletal myxoid chondrosarcoma; RMC, renal medullary carcinoma; PNET, primitive neuroectodermal tumor; MB, medulloblastoma; CPCa, choroid plexus carcinoma*.\n\nAs a rare condition, even in children, there is some justification for 'lumping' ATRT with similar tumors. This is particularly rational when the intervention is relatively non-specific, e.g., radiotherapy, traditional 'cytotoxic' chemotherapy (and soon, perhaps, immunotherapy).\n\nIn the case of more 'targeted' treatments, the pool of analogous diseases will be much smaller. Two of the seven trials may be said to be target aspects of the biology of ATRT. It is striking that the trial of alisertib aims to inhibit aurora kinase A, an enzyme that has been implicated in EMT. This trial includes malignant rhabdoid tumors (MRT), which show EMT but do not lose INI1 protein ([@B46]). MRTs, as well as rhabdoid tumors of the kidney (also showing EMT), are included in the trial of tamezostat (an EZH2 inhibitor). While beyond the scope of the current work, we suggest that the term \"rhabdoid\" is synonymous with EMT. Interestingly, the trial of tamezostat considers other tumors lacking INI1 to be sufficiently similar to be worth including.\n\nThe other five trials that include ATRT are relatively 'non-specific' and compensate for its low incidence by combining it with other 'serious' tumors, typically also of childhood. For example, four of five include PNET and medulloblastoma.\n\n5.3. Treatment {#S5-3}\n--------------\n\nAs ATRT is primarily a disease of infancy, co-ordination of treatment between pediatric and adult oncology services appears appropriate.\n\n### 5.3.1. Chemotherapy {#S5-3-1}\n\nATRT usually occurs before CSI can be given safely, i.e., in those aged \\<3. Thus, CT has traditionally been a cornerstone of treatment. We suggest that a primary guiding factor be penetration into CSF (given the high prevalence of LM) as well as into the intraparenchymal tumor. As these tumors tend to disrupt the blood--brain barrier (BBB) and are often next to parts of the brain with no BBB, penetration of CT into normal brain parenchyma does not appear to be of prime concern.\n\nIn the rare cases when ATRT is localized and can be completely excised, local RT appears appropriate; arguably, a protocol similar to that used for glioblastoma can be adopted in such cases.\n\nThe most common approach to CT typically involves cytotoxic and relatively non-specific agents. Ifosfamide, carboplatin, and etoposide (ICE) was the most common regimen in our series, a protocol typically used for sarcoma and lymphoma. These agents have reasonable brain and CSF penetration ([@B47]--[@B49]). However, given the grave prognosis of this condition, a more aggressive approach appears warranted in adults, as in our second case.\n\n### 5.3.2. Radiotherapy {#S5-3-2}\n\nSeveral reports demonstrated the beneficial role of adjuvant RT after surgical resection in patients with ATRT ([@B11], [@B50], [@B51]). However, RT is a significant challenge in those younger than 3\u2009years old ([@B52]).\n\nThe benefit of combining RT with SX in ATRT has been shown by Lau et al. ([@B53]). In a sample of 171 pediatric and 3 adult cases, they report an increase in median survival from 1.9\u2009\u00b1\u20090.4 to 5.9\u2009\u00b1\u20090.7\u2009years in those who had RT in addition to SX. Buscariollo et al. confirmed the value of RT in another retrospective series, this with 144 patients, where median survival improved from 6 to 34\u2009months (p\u2009\\<\u20090.001) ([@B54]).\n\nOnce a decision on the use of RT has been made for a patient with ATRT, factors such as timing, dosing, and technique need to be considered.\n\nChen et al. reported improved disease-free survival in patients who had a total dose more than 50\u2009Gy ([@B55]). Significant improvement in median survival was reported in patients who had a shorter interval between SX and RT. RT delivered early vs. latein the course of treatment has also been associated with improved survival ([@B11]).\n\nProton therapy may be especially attractive in younger patients, in order to minimize the risk of late complications. Encouraging results with local conformal proton therapy have been reported by Bernstein et al. ([@B56]). Nine of ten patients with ATRT (with a median age of 1.8\u2009years) had no evidence of disease, with a median follow-up of 27.3\u2009months.\n\nAs distal relapse is associated with higher mortality and as the majority of distant relapses occur within the CNS, CSI is recommended is those older than 3\u2009years old: 23.4\u2009Gy to the neuraxis (with 54\u2009Gy to the tumor bed) is recommended inpatients older than 3\u2009years old. Younger patients have been treated using either no adjuvant RT or local-only RT, with a trend toward improved survival with the addition of RT.\n\n### 5.3.3. A Combined Approach {#S5-3-3}\n\nPromising results have been already been shown in children \\>3 with PNET undergoing CSI followed by high-dose CT and in the setting of recurrent CNS tumors ([@B57], [@B58]). Stem-cell rescue was required with these more aggressive regiments, either from peripheral blood or bone marrow. Given its known role as a radiosensitizer and excellent brain and CSF penetration, the use of temozolomide during radiotherapy also appears reasonable ([@B59]). Early use of intrathecal chemotherapy also appears rational. As it is well tolerated in long-term use, it may also be an acceptable approach in those patients who would not tolerate more aggressive forms of CT.\n\n6. Conclusion {#S6}\n=============\n\nATRT in adults carries a grave prognosis, particularly in those cases with leptomeningeal spread. We postulate that ATRT is a tumor of neuroectodermal origin that demonstrates mesenchymal transition.\n\nOur cases indicate the possibility of improved outcome with more aggressive therapy. CSI should always be considered in adults. The use of stem-cell rescue allows for the use of more aggressive chemotherapeutic regimens than has hitherto been the case.\n\nConsent {#S7}\n=======\n\nWritten, informed consent was obtained from both patients specifically for the publication of these case reports.\n\nAuthor Contributions {#S8}\n====================\n\nCD drafted the manuscript and performed the statistical analysis. CD and KM wrote the first case report and performed the literature review. JY and EH wrote the second case report. EA performed the Sanger sequencing. All the authors contributed to data interpretation and to the final draft of the manuscript.\n\nConflict of Interest Statement {#S9}\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nSupplementary Material {#S10}\n======================\n\nThe Supplementary Material for this article can be found online at <http://journal.frontiersin.org/article/10.3389/fneur.2017.00247/full#supplementary-material>.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\nAdditional information is provided in supplemental files: `data sheet 1.xlsx` A spreadsheet with the following worksheets: `d1` Data set used for analysis.`k1` Key to data set.`ihc` Immunohistochemistry.`data sheet 2.pdf` Statistical analysis.\n\nCase reports and case series reviewed, but not specifically cited above ([@B60]--[@B89]).\n\n[^1]: Edited by: Gordon Li, Stanford University, United States\n\n[^2]: Reviewed by: Michael C. Fruehwald, Children's Hospital Augsburg, Germany; Ofelia Cruz, Hospital Sant Joan de D\u00e9u Barcelona, Spain\n\n[^3]: Specialty section: This article was submitted to Neuro-Oncology and Neurosurgical Oncology, a section of the journal Frontiers in Neurology\n"}
+{"text": "1. Introduction {#sec0005}\n===============\n\nTraumatic brain injury (TBI) is a serious and complex public health issue. In the USA alone, there are over 1.6 million TBI's reported every year ([@bib0105]) with many more unreported. A complex and debilitating neurological disorder, TBI is defined as an alteration in brain function, or other evidence of brain pathology, caused by an external force ([@bib0135]). Classified by severity, TBI ranges from mild to severe, based on the length of time of loss of consciousness, post-injury amnesia, Glasgow Coma Score ([@bib0190]) and trauma related findings on neuroimaging ([@bib0125]). The hallmark of TBI is diffuse axonal changes combined with continually evolving secondary changes. This feature combined with the heterogeneity of external factors such as location of injury, severity of injury, previous history and individual response to brain injury, can result in unique and dramatic changes in brain structure and brain function at many levels ranging from microscopic tears in white matter to global changes in functional brain networks. Headaches, dizziness, sleep disturbances, and fatigue often occurs immediately post injury, and may be severe. Attention, memory, processing speed and executive functions (i.e. working memory, cognitive flexibility, etc.) are also compromised early ([@bib0160]) and recover inconsistently ([@bib0020]; [@bib0130]). A significant proportion of individuals with TBI continue to suffer from persistent cognitive and behavioural complaints and disability ([@bib0090]; [@bib0130]), resulting in prolonged or lifelong disability and dependence on the health care system. It is estimated that 2% of the population live with permanent disability related to TBI, with an estimated economic impact of \\$60 Billion/year in the USA([@bib0055]). Developing effective rehabilitation strategies for this population is therefore of great importance.\n\nAt a behavioural level there is considerable evidence that targeted cognitive interventions can improve attention, executive function, and memory ([@bib0035]; [@bib0075]; [@bib0100]). In addition, cognitive rehabilitation is reportedly effective in helping patients learn and apply compensatory mechanisms for remaining deficits ([@bib0040]). Improvements in specific tasks related to executive function and memory have also been shown to generalize to functional improvement in everyday activities ([@bib0080]). Given that cognition is dependent on the activity of widely distributed functional brain networks ([@bib0015]), evaluating the characteristics of the underlying networks in response to behavioral change stimulated by cognitive intervention can provide important insights into the underlying neuroplastic process.\n\nA powerful and elegant method of evaluating the characteristics of brain networks is through the use of graph theory. In this framework, the brain is viewed as a network with interactions and communication occurring over multiple levels between local and distant areas ([@bib0170]). One approach to understanding the nature of these interactions is via functional connectivity. Functional connectivity refers to the statistical interdependencies between physiological time series recorded from the brain ([@bib0060]). Graph theory has emerged as a promising tool in recent years for characterizing brain connectivity at both global and regional levels ([@bib0025]; [@bib0170]). Graph theory allows for the quantitative analysis of network organization, characterizing the brain as a set of networks, with each network being composed of distinct brain regions or \"nodes\". The various nodes are functionally connected via edges. The relationship between nodes and edges provides information about the organization and efficiency of the network. Networks with an ordered structure have a high clustering coefficient (a measure that depicts the connectedness of immediate neighbors around individual vertices), long characteristic path length (an index reflecting the overall integration of the network), low global efficiency (defined as the average inverse shortest path) and low density or cost (the fraction of present connections to possible connections) ([@bib0205]). In contrast, randomly organized networks are characterized by a low clustering coefficient, short characteristic path lengths, and a high global efficiency and density.\n\nThe healthy brain has been shown to be a combination of ordered networks with a certain fraction of randomly rewired links which results in \"small-world\" networks, with tightly connected neighborhoods, short characteristic path lengths and a high local efficiency ([@bib0205]). These small-world networks balance between local specialization and global integration, which are optimal for information processing ([@bib0200]).\n\nGraph theoretical approaches have recently been applied to understand the effect of brain injury from a network perspective. In adults with TBI there appears to be a shift towards sub-optimal network organization ([@bib0030]; [@bib0140]; [@bib0150]). We have previously shown that mild TBI in adolescents does not alter resting state global network efficiency but does cause change in the local networks within the prefrontal cortex ([@bib0195]). This provides evidence for local changes in the frontal regions of the brain that are likely to affect the efficient processing of cognitive functions following mTBI. Arnemann et al. (2014) applied graph theory to predict individual responses to cognitive training in individuals with brain injury ([@bib0005]). They found that modularity (a measure of the density of links within a community relative to ones between communities) was able to predict improvements in attention and executive function following cognitive training. This suggests that functional brain imaging and graph theory has the potential to provide valuable information for understanding the mechanisms that influence recovery from TBI.\n\nThe purpose of this study was to pilot measures using graph theory analysis to evaluate the feasibility of these measures in capturing change following intervention. The specific objectives of this study were: (1) To establish the baseline global and local functional connectivity in individuals with TBI and evaluate the association with neuropsychological functioning, (2) To evaluate the changes in global and local functional connectivity in individuals with TBI following a 3-month cognitive intervention program and (3) To determine if network reorganization as measured by functional connectivity is associated with changes in cognitive recovery following the 3-month cognitive intervention program.\n\n2. Methods {#sec0010}\n==========\n\n2.1. Participants {#sec0015}\n-----------------\n\nParticipants with TBI were recruited from brain injury associations across the Greater Vancouver (British Columbia, Canada) area. All TBI participants were chronic with injuries occurring a minimum of 1 year prior to the start of the intervention. All healthy controls were recruited from the lower mainland of Vancouver in close proximity to the university. All controls were screened to ensure that they had no history of head trauma, neuropsychiatric disorders, substance abuse or any other neurological conditions. The participants with TBI were first interviewed to determine eligibility and to evaluate the severity of TBI. Participants were excluded if they were involved in litigation, had a history of current or recent substance abuse or if they had other severe medical conditions affecting brain function. They were also excluded if they had a diagnosis of psychiatric illness based on self report or when assessed with the Mini-international neuropsychiatric interview (MINI) ([@bib0175]). TBI severity was classified mild or moderate/severe, as per ACRM criteria (Mild Traumatic Brain Injury Committee, Head Injury Interdisciplinary Special Interest Group, American Congress of Rehabilitation Medicine. 1993). using retrospective self report of the duration of loss of consciousness and length of post-traumatic amnesia.\n\nThe experiments described in this study were approved by the Human Ethics Review Board at the University of British Columbia. All participants provided written consent according to the guidelines set forth by the Clinical Research Ethics Board at the University of British Columbia and this study complies with all regulations.\n\n2.2. Demographic characteristics {#sec0020}\n--------------------------------\n\n[Table 1](#tbl0005){ref-type=\"table\"} shows the demographic features and clinical features for all the participants in this study. The mean age of the TBI participants was 39.6 years (18, 54), with a mean of 15.3 years of formal education. The gender distribution was 50% male. Time since initial injury for the TBI group ranged from 2 to 29 years with an average of 11.7 years. The healthy controls had a mean age of 37.4 years (21, 53) with a mean of 16.6 years of formal education and a gender distribution of 73% male. There were no significant differences in age \\[t(19) = 0.429, p = 0.673, d = 0.19\\] or education \\[t(19) = -1.055, p = 0.305, d = -0.46\\] between groups. All participants were right handed. All participants with TBI were able to complete the 3-month program.Table 1Demographic and clinical profile of participants with TBI and Controls. M = male; F = female; TBI = Traumatic Brain Injury; MVA = Motor Vehicle Accident.Table 1Participant IDSexAgeYears of educationTime since injury (years)EtiologyMechanism of injurySeverityS001M511422TBIMVASevereS002F501829TBIMVASevereS003M361624TBIMultiple ConcussionMildS004M41128TBIMVASevereS005M18132TBIMultiple ConcussionMildS007F48204TBIMVAMildS008M26136TBIFallSevereS009F35138TBIMVASevereS011F38187TBIMVASevereS012F50177TBIMVAMildC001F5320NAHealthyNANAC002M3620NAHealthyNANAC003M3018NAHealthyNANAC004M5220NAHealthyNANAC005M2116NAHealthyNANAC006M3814NAHealthyNANAC008F5014NAHealthyNANAC009M2416NAHealthyNANAC010F2712NAHealthyNANAC011M4514NAHealthyNANAC012M2116NAHealthyNANA\n\n2.3. Neuropsychological testing {#sec0025}\n-------------------------------\n\nAll subjects were administered several neuropsychological tests assessing processing speed, memory, and executive function, as deficits in these cognitive skills have been documented across a range of TBI ([@bib0155]). The specific tests used were as follows:\n\n1\\. Rey Auditory Verbal Learning Test (RAVLT)([@bib0115]): On this verbal learning and memory test, the primary variables employed were total recall during learning trials and delayed free recall.\n\n2\\. Trailmaking Test Parts A and B ([@bib0165]): The Trails A test measures visual attention and processing speed, and Trails B assesses more complex attentional shifting. The time to completion for each task was employed.\n\n3\\. Verbal Fluency ([@bib0115]): On the phonemic verbal fluency task, subjects were asked to generate as many words beginning with the letters F, A, S. The main outcome was number of correct words produced.\n\nFor all tasks, demographics corrected z-scores were used as the primary measures. For the RAVLT and Verbal fluency tasks, alternate versions of the tests were used during follow-up testing (see below) to minimize practice effects. A global composite score was derived by averaging z-scores for the 5 primary measures described above.\n\n2.4. EEG recording and analysis {#sec0030}\n-------------------------------\n\nEEG was recorded using a 64-channel HydroCel Geodesic SensorNet (EGI, Eugene, OR). The EEG cap was placed on each participant's head and 5 minutes of resting state data was recorded with their eyes closed. EEG was recorded and amplified using Net Amps 300 amplifier, at a sampling rate of 250 Hz. Scalp electrode impedances were generally under 50 k\u2126. The signal was referenced to the vertex (Cz) and filtered from 4 to 40 Hz. A notch filtered at 60 Hz was applied. The EEG signals were analyzed offline using Brain Electrical Source Analysis (BESA) (MEGIS Software GmbH). An automated artifact scan was performed for extracting motion and excessive eye movement artifacts. BESA brain source montage was used to convert the EEG activity obtained from all the 64 scalp channels into predicted contributions of a set of 15 different brain source activity. The advantage of using a brain source montage is that the volume conduction effects are reduced in comparison with the surface electrodes and provides a better model of the underlying brain source activity.\n\n2.5. Graph theoretical analysis {#sec0035}\n-------------------------------\n\nBased on the learned connectivity networks, graph theoretical analysis was used to extract the structural features from learned networks ([@bib0025]). Traditional graph theoretical measures were used to characterize the network features in terms of density, global efficiency, clustering coefficient, and modularity. Density is defined as the fraction of present connections to all possible connections. Global efficiency describes the communication ability of the entire graph ([@bib0110]), and is defined as the average of the inverse shortest path. Clustering coefficient describes the degree to which nodes in a graph tend to cluster together. Modularity of the network is used to measure how well the network can be divided into the sub-modules ([@bib0145]). A higher value of modularity demonstrates that the graph is better divided with tighter connections within modules. We used the Brain Connectivity Toolbox ([@bib0170]) running Matlab (Natick, MA) to perform the graph theoretical analysis.\n\n2.6. Construction of connectivity matrix {#sec0040}\n----------------------------------------\n\nIn this paper, we constructed the brain functional connectivity networks using the preselected EEG signals and an error-rate controlled network learning algorithm. Based on the learned connectivity networks, the graph measures were further calculated to extract the functional network features. EEG signals were interpolated at 27 locations (FP2, FPZ, FP1, F10, F8, F4, FZ, F3, F7, F9, A2, T8, C4, CZ, C3, T7, A1, P10, P8, P4, PZ, P3, P7, P9, O2, OA, O1) on the scalp using BESA's Virtual Standard 10--10 Average montage. EEG time series from these 27 locations were used to construct the brain connectivity networks with each channel representing one brain region in the network. The connectivity network graphs were then computed for each individual subject and for each emotional expression using false discovery rate controlled PC (PCFDR) algorithm, which is a statistical model that tests the conditional dependence/independence between any two regions based on all other brain regions ([@bib0120]).\n\nWe used partial correlation to evaluate the conditional independence, which estimates the directed interactions between any two brain regions after removing the effects of all other brain areas. The PC algorithm starts from a complete graph and tests for conditional independence in an efficient way. The PCFDR algorithm is designed to control the false discovery rate (FDR), which evaluates the proportion between the connections that are falsely detected to all those detected, below a specified predefined level. Compared to the traditional Type-1 and Type-2 error rates, FDR has more conservative error rate criteria for modeling brain connectivity due to its direct relation to the uncertainty of the networks of interest. The PCFDR algorithm and pseudo-code are described in details in ([@bib0120]). F FDR threshold was set at the 5% level.\n\n2.7. Baseline and Post-intervention testing {#sec0045}\n-------------------------------------------\n\nPrior to the start of the program all participants took part in a comprehensive baseline assessment, which included resting state EEG and neuropsychological testing. All testing was completed at the University of British Columbia. The neuropsychological testing and clinical scales took approximately 90--120 min to complete. All participants took part in the EEG testing. Only 6 TBI patients completed the initial neuropsychological testing at the first time point. 4 additional TBI patients received their baseline neuropsychological testing at the 3-month time point and thus we elected to exclude the additional 4 patients from analyses involving cognitive data.\n\nFollowing the 3-month intervention program EEG was conducted on all 10 TBI patients. Symptoms of generalized anxiety, depression, and general psychological distress were assessed using the Generalized Anxiety Disorder \u2212 7 (GAD-7; ([@bib0180]), Patient-Health Questionnaire (PHQ-9; ([@bib0095]), and Brief Symptom Inventory \u2212 18 (BSI-18; ([@bib0050])).\n\n2.8. Cognitive intervention {#sec0050}\n---------------------------\n\nThe Arrowsmith Program is a suite of cognitive exercises. The goal of these exercises is to improve cognitive functioning across a broad domain, for example, executive functions, reasoning and memory. Each individual's program was based on an assessment of that individual's learning profile to identify his or her specific areas of difficulty. Each individual had his or her specific schedule of tasks and exercises to be completed during the course of a day. The exercises for each individual student were uniquely based on their individually identified learning profiles. These include written, visual, auditory and computer exercises. Each cognitive program had a series of intensive and graduated tasks. Performance criteria of automaticity, consistency and accuracy are built in at all levels and an individual was required to meet these criteria before mastering to the next level of complexity. Goals were set daily, weekly and monthly and each month the individual's attained levels in each program was entered into a database and analyzed to see if progress met benchmark expectations. In order to maximize evaluation of transfer effects, the neuropsychological battery that was used to assess outcome was assembled independently of the cognitive exercise tasks that were used within the cognitive intervention program.\n\nThe cognitive intervention program took place four days a week for 4 \u2212 5 hours a day for a 3-month period. Each day was composed of 6--8 blocks of different activities depending on individual's specific areas of weakness.\n\n2.9. Analysis {#sec0055}\n-------------\n\nDemographic data including age, gender and years of education were evaluated for between group differences with t-tests. T-tests were also performed to test differences in the EEG measures and the neuropsychological measures between groups at baseline. Two a priori EEG clusters of interest were selected for analysis based on our previous work examining changes in functional connectivity in adolescents with mTBI ([@bib0195]). EEG clusters of interest included F7 (L) inferior frontal gyrus (IFG) and F10 (R) inferior frontal gyrus. Paired t-tests were conducted to compare baseline and post-intervention scores in the TBI group. Pearson correlations were conducted to evaluate specific correlations between graph theory metrics and the individual and global composite neuropsychological scores.\n\n3. Results {#sec0060}\n==========\n\n3.1. Differences between groups at baseline {#sec0065}\n-------------------------------------------\n\n[Table 2](#tbl0010){ref-type=\"table\"} shows the mean scores (SD) on the global composite score based on the mean of all test scores. As expected, at baseline the TBI group performed worse on the overall composite score although the difference did not reach statistical significance as a likely result of diminished power. Consistent with this, the Cohen's d effect size for the difference was .78, which indicates a moderate to large difference.Table 2Global composite scores for the controls, TBI participants at baseline and TBI participants at 3 months post intervention.Table 2Cognititon Composite (Z score)NMeanStd. DeviationSignificance (2 tailed)Control vs. TBIControl12.1166.62026.102TBI6-.58231.10851Baseline vs. 3 monthBaseline6-.58231.1085.0003 Month6.01031.3002\n\nAt baseline there were no significant differences between groups for any of the four graph metrics associated with global connectivity: density (t(17) = 1.03, p = 0.32, d = 0.47), global efficiency (t(17) = -0.278, p = 0.78, d = -0.13), modularity (t(17) = 1.05, p = 0.31, d = 0.48), clustering coefficient, t(17) = 1.612, p = 0.125, d = 0.74.\n\nIn contrast with the global metrics we found significant differences in local metrics at the F10 and F7 electrode clusters corresponding to the right and left inferior frontal gyrus, respectively between groups ([Fig. 1](#fig0005){ref-type=\"fig\"}). Specifically, the F10 hub value was significantly higher compared to the healthy control group, t(17) = 2.24, p = 0.039, d = 1.01. A number of graph metrics for F7, showed significant decreases in the TBI group in comparison with the healthy control group. These included F7 degree, t(17) = -2.28, p = 0.036, d = -1.05, F7 betweenness, t(17) = -2.41, p = 0.028, d = -1.1 and F7 hub value, t(17) = -2.28, p = 0.035, d = -1.05. These results suggest that network connections are denser in the right inferior frontal regions and less dense in the left inferior frontal regions in the TBI group.Fig. 1Baseline functional connectivity measures showing differences between TBI \u2212 Baseline and controls in local connectivity measures. (A. F10 Hub Value; B. F7 Degree; C. F7 Betweenness; D. F7Hub Value).Fig. 1\n\n3.2. Cognitive training related changes {#sec0070}\n---------------------------------------\n\n[Table 2](#tbl0010){ref-type=\"table\"} also shows the scores following the three-month cognitive intervention in the participants that completed testing at baseline and post-intervention. A paired-test showed that there was a statistically significant increase in the composite score between baseline and post-intervention in the TBI group. In addition, there were no significant changes in generalized anxiety, t(5) = 0.08, p = 0.94, depressive symptoms, t(5) = 0.86, p = 0.43, or general psychological distress, t(5) = 0.58, p = 0.59.\n\nIn terms of functional connectivity, no changes were seen in the global connectivity measures. For the local connectivity measures, we noted changes in specific graph theory metrics at the F10 and F7 electrode clusters. [Fig. 2](#fig0010){ref-type=\"fig\"} shows that at F10, degree, \\[t(9) = 3.35, p = 0.008\\] ([Fig. 2](#fig0010){ref-type=\"fig\"}A), betweenness \\[t(9) = 3.52, p = 0.007\\] ([Fig. 2](#fig0010){ref-type=\"fig\"}B), and hub value \\[t(9) = 3.53, p = 0.006\\] ([Fig. 2](#fig0010){ref-type=\"fig\"}C) showed a statistically significant decrease following the intervention. At F7, measures increased, however none were statistically significant: degree, t(9) = -0.896, p = 0.394, betweenness, t(9) = -1.05, p = 0.323 and hub value, t(9) = -1.01, p = 0.338.Fig. 2Post-intervention connectivity measures for TBI \u2212 Baseline, TBI \u2212 3 Months Post Intervention compared to Controls, showing significant changes after 3 months in F10 Degree (A.), F10 Betweenness (B.) and F10 F10Hub Value (C.).Fig. 2\n\n[Fig. 3](#fig0015){ref-type=\"fig\"} shows the relationship between global density and the overall composite score. Note that there is a positive correlation in the controls with higher scores associated with higher global density (P=.63). In the TBI group there was a negative correlation between these variables at baseline (P = -.39). Following intervention, there was a trend towards a positive slope (P=.28). [Fig. 4](#fig0020){ref-type=\"fig\"} shows the relationship between the RAVLT and global density (4A) and the RAVLT and F10 degree (4B). Both correlations show a change from a negative correlations at baseline (P = -.38; P = -.33) to a positive correlation at 3 months post intervention (P=.69; P=.77). No other correlations were found to be significant.Fig. 3Scatter plot showing relationship between Global Density and Global composite score in controls, TBI participants at baseline and TBI participants at 3 months post intervention.Fig. 3Fig. 4(A). Scatter plot showing the relationship between RAVALT and Density in controls, TBI participants at baseline and TBI participants at 3 months post intervention. (B). Scatter plot showing the relationship between RAVALT and F10 Degree in controls, TBI participants at baseline and TBI participants at 3 months post intervention.Fig. 4\n\n4. Discussion {#sec0075}\n=============\n\nThe purpose of this study was to pilot measures using graph theory analysis to evaluate the feasibility of these measures to capture change following a 3 month program of cognitive intervention and to determine if these changes correlated with measures of cognition. The data from our pilot study provides preliminary evidence for improved performance in cognition and corresponding changes in local functional network connectivity in bilateral frontal regions following a 3 month cognitive intervention program in a small sample of adults with chronic TBI.\n\nGraph theoretical analysis at baseline revealed that there were no significant differences in global measures of functional connectivity between groups suggesting that in this cohort, there was no change in resting state global network connectivity, although there was evidence of disruption in the relationship between global metrics and measures of cognition. We did find significant changes in local networks in this group. Specifically, we observed increases in graph metrics in F10 and decreases in F7. F10 corresponds to the (R) inferior frontal gyrus (IFG) and F7 to the (L) IFG ([@bib0085]). The IFG is a key region involved in three processes of cognitive control: working memory, task switching and inhibitory control ([@bib0185]). The increase in hub value of IFG on the right and decrease on the left may seem paradoxical, but it suggests a disruption of network organization that is centered around the frontal regions. The hub value of one node is related to the structure of its neighbors, where a high value means that there are other key nodes that are connected with that hub. The increased value of the hub on the right indicates hyper-connectivity and the decrease on the left suggests hypo-connectivity. Overall, the consequences of disruption in the physical network due to TBI are reflected in disruptions in functional connectivity. These types of disruptions in functional connectivity have previously been reported in the literature. Hillary et al. ([@bib0065]) found increased functional connectivity specifically in the brain regions that are most highly connected or \"rich hubs\". We have also previously found increased functional connectivity in adolescents with mTBI in frontal regions ([@bib0010]). Palacios and colleagues recently reported increased functional connectivity within the frontal lobe in patients with chronic TBI. In contrast, Pandit et al., reported decreased functional connectivity in patients with chronic TBI ([@bib0150]).\n\nAlthough initial hypotheses in the literature related to increased connectivity have focused on compensatory strategies as a mechanism to account for loss of structural connectivity, more recently, Hilary et al. ([@bib0070]) have shown that hyper-connectivity is a common finding across a range of neurological conditions including TBI, multiple sclerosis, mild cognitive impairment and Alzheimer's disease. They hypothesize that increased connectivity may allow the brain to continue to meet task demands in the face of network disruption. Importantly, they suggest that this increased connectivity comes at the cost of slowed processing speed and cognitive fatigue. Our data shows that increased connectivity in the frontal regions is in fact correlated with lower cognitive scores suggesting that greater resources are being used that may lead to lower information processing efficiency.\n\nAlthough hyper-connectivity is a common finding across a range of neurological disorders, from a network perspective it is unlikely that this is the only response to structural changes in the brain. Across all brain networks it is more likely that there is a combination of hyper and hypo connectivity that reflects the changes in different brain networks ([@bib0070]). Indeed, our data supports this by showing that hyper-connectivity is not uniform across brain regions following TBI. Furthermore, the pattern of connectivity that we observed in each hemisphere provides data for the hypothesis that detailed analysis of network organization across the whole brain is necessary to understand and map the patterns of brain connectivity that underlie the cognitive deficits in individuals with TBI.\n\n4.1. Training related changes in cognition {#sec0080}\n------------------------------------------\n\nWithin several important constraints of the present study (i.e. small sample size and a lack of a TBI control group who did not receive cognitive intervention which is necessary to rule out practice effects for cognitive improvements), the analysis of cognitive performance from baseline to post-treatment in patients suggested trends in improvements in overall cognition. Importantly, these improvements were observed in a sample of chronic patients. There are few published studies that have examined the effectiveness of cognitive intervention in patients who are in the chronic stages of recovery([@bib0045]).\n\nIn a recent re-analysis of meta-analytic studies assessing the effects of cognitive rehabilitation on acquired brain injury, it was concluded that patients with traumatic brain injuries were more likely to receive benefit from cognitive retraining targeting attention based tasks ([@bib0040]). Because the cognitive measures included in the neuropsychological battery within this study were highly attention-based, this likely maximized the likelihood of observing cognitive gains as a result of the cognitive intervention.\n\n4.2. Training related changes in functional connectivity {#sec0085}\n--------------------------------------------------------\n\nInterestingly, our data show that cognitive intervention lead to changes in the organization of brain networks such that connectivity within hub areas are reorganized, and this occurred with a parallel improvement in cognition. In particular, we found a small shift in the correlation between overall Global density and the global composite score in the TBI group which was negatively correlated at baseline and moved to a slight positive trend following the intervention period. We also observed shifts in correlations between auditory verbal learning and global connectivity as well as measures related to brain regions in the right inferior frontal gyrus. These changes suggest a subtle reorganization between brain and behavior relationships that were disrupted at baseline.\n\nOur results provide preliminary evidence that participating in an intensive cognitive intervention program was associated with neuroplastic changes in adults with chronic TBI that occurred in parallel with improvements in cognition. Overall, we observed a shift from a baseline pattern of network organization that may be characterized by neural inefficiency and decreased cognition to a reorganization that reflected improved efficiency with possible improvements in fluid cognition. Importantly this data suggests that brain network organization is capable of reorganization even in chronic patients with intense intervention. Further work with a larger sample is clearly needed to understand the nuances of how brain organization impacts on cognitive ability and performance.\n\nDeclarations {#sec0090}\n============\n\nAuthor contribution statement {#sec0095}\n-----------------------------\n\nShaun Porter: Performed the experiments; Analyzed and interpreted the data; Wrote the paper.\n\nIvan Torres: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.\n\nWilliam Panenka: Conceived and designed the experiments.\n\nZahra Rajwani: Performed the experiments; Analyzed and interpreted the data.\n\nDelrae Fawcett, Amna Hyder: Analyzed and interpreted the data.\n\nNaznin Virji-Babul: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.\n\nFunding statement {#sec0100}\n-----------------\n\nThis work was supported by Mitacs-Accelerate Graduate Research Internship Program in collaboration with the Eaton Educational Group.\n\nCompeting interest statement {#sec0105}\n----------------------------\n\nThe authors declare no conflict of interest.\n\nAdditional information {#sec0110}\n----------------------\n\nNo additional information is available for this paper.\n\nWe would like to thank all the participants who took part in this study.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nLung cancer is the leading cause of cancer-related death in humans^[@CR1],[@CR2]^. Approximately 1,590,000 people died from lung cancer in 2012^[@CR3]^. Five-year survival rates vary from 4--17%, depending on stage and regional differences^[@CR4]^. Late diagnosis makes treatment options challenging, but early detection could significantly reduce mortality in lung cancer patients^[@CR5]--[@CR7]^.\n\nDifferent imaging modalities can be used for lung cancer detection, such as computed tomography (CT), magnetic resonance imaging (MRI), and ^18^F-fluorodeoxyglucose positron emission tomography and computed tomography (FDG-PET/CT). However, they lack sufficient specificity for the early detection of tumors^[@CR8]^. Molecularly targeted tumor imaging with a specific monoclonal antibody is superior to conventional nonspecific imaging technologies^[@CR9],[@CR10]^. Nevertheless, there are few reports on molecularly targeted imaging for the detection of lung cancer, because of the lack of molecular targets with high specificity for lung cancer.\n\nSP70 protein is a novel tumor marker of non-small-cell lung cancer (NSCLC), especially lung adenocarcinoma^[@CR11]^. SP70's monoclonal antibody NJ001 could specifically recognize and react to lung adenocarcinoma cells^[@CR12]^. In this study, we aimed to demonstrate the feasibility of SP70-targeted imaging with NJ001-conjugated nanomagnetic beads (immuno-nanomagnetic beads) in lung adenocarcinoma and the potential for its use in real-time monitoring and early detection.\n\nResults {#Sec2}\n=======\n\nBioluminescence intensity correlated with cell number {#Sec3}\n-----------------------------------------------------\n\nBoth SPC-A1-luc and U87-luc cells were diluted by a serial 2-fold dilution method, and then luciferase activity was measured. Bioluminescence intensity was highly correlated with the total number of cells (*R*^2^\u2009\\>\u20090.9900) for both cell lines (Fig.\u00a0[S1](#MOESM1){ref-type=\"media\"}). Furthermore, the average luciferase activity value of SPC-A1-luc *in vitro* was 1,223 photons/sec/cell.\n\nSP70-targeted fluorescence imaging in subcutaneous xenograft mouse models {#Sec4}\n-------------------------------------------------------------------------\n\nDIF showed that SP70 was located on the SPC-A1-luc cell membrane and in the cytoplasm but was not expressed in U87-luc cells (Fig.\u00a0[1A](#Fig1){ref-type=\"fig\"}). After subcutaneous implantation, three mice bearing subcutaneous SPC-A1-luc cell tumors were monitored weekly by both BLI and SP70-targeted fluorescence imaging in parallel. Three mice injected with U87-luc cells were used as controls. For SP70-targeted fluorescence imaging, fluorescent signals could be detected in the subcutaneous SPC-A1-luc mice starting in the third week. However, there were no fluorescent signals in the mice bearing glioma U87-luc cells (Fig.\u00a0[1B](#Fig1){ref-type=\"fig\"}). BLI could detect SPC-A1-luc xenograft tumor development by day 7, and bioluminescence intensity increased in parallel with tumor volume measured by calipers.Figure 1SP70-targeted fluorescence imaging in subcutaneous xenograft mouse models. **(A**) SP70 was located on the SPC-A1-luc cell membrane and in the cytoplasm, but was not expressed in U87-luc cells. (**B**) SPC-A1-luc subcutaneous xenograft tumor could be detected by fluorescence imaging using NIR fluorescence CF750 (red)-labeled NJ001 3 weeks after inoculation. In contrast, U87-luc xenograft tumors could not be detected by fluorescence imaging (black arrow).\n\nSP70-targeted fluorescence imaging in orthotopic lung tumor models {#Sec5}\n------------------------------------------------------------------\n\nThree mice with SPC-A1-luc orthotopic xenograft tumors were imaged with both BLI and fluorescence imaging. As shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}, SP70-targeted fluorescence imaging could detect the lesions at the third week. Meanwhile, the total lung BLI photon count was 7.2\u2009\u00d7\u200910^5^ photons/sec. Since *in vitro* BLI analysis showed that the luciferase activity value of a single SPC-A1-luc cell was 1,223 photons/sec, it was inferred that the tumor cell number reached approximately 600 on the third week.Figure 2Tumor monitoring with SP70-targeted fluorescence imaging in orthotopic lung tumor implantation models. **(A**) Three mice with SPC-A1-luc orthotopic xenograft tumors were imaged by both BLI (left) and fluorescence imaging (right) at the 3^rd^, 6^th^ and 9^th^ weeks. (**B**) FLI photon counts were correlated with BLI photon counts.\n\nImmuno-nanomagnetic bead characterization {#Sec6}\n-----------------------------------------\n\nNJ001 antibody-coated and noncoated nanomagnetic beads were characterized using transmission electron micrograph (TEM) with phosphotungstic acid staining (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}).Figure 3Immuno-nanomagnetic beads characterization with TEM. (**A**) TEM image of the nanomagnetic beads, showing a diameter of approximately 180\u2009nm; (**B**) TEM image of NJ001-conjugated nanomagnetic beads, using phosphotungstic acid for background staining.\n\nSignal enhancement of SP70-targeted micro-CT scan in orthotopic lung tumor models {#Sec7}\n---------------------------------------------------------------------------------\n\nAfter injection of the immuno-nanomagnetic beads, the micro-CT signal intensity of the orthotopic lung tumors increased significantly. The image density peaked at 4\u2009h after injection of the immuno-nanomagnetic beads. The tumor grayscale index in the immuno-nanomagnetic beads group was the highest among the three groups (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}).Figure 4Signal enhancement in SP70-targeted micro-CT scan in orthotopic lung tumor implantation models. Micro-CT scan at 0, 2, 4, 6 and 24\u2009h after NJ001 conjugated nanomagnetic beads or control injection in the sixth week. The image density increased and peaked at 4\u2009h postinjection with NJ001-conjugated nanomagnetic beads.\n\nTumor lesions could be detected at the sixth week in the mice injected with normal saline or bare nanomagnetic beads but were visible by the fourth week in the mice receiving immuno-nanomagnetic beads (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). In other words, orthotopic tumor lesions could be found 2 weeks earlier by SP70-targeted micro-CT scan compared with routine micro-CT scan.Figure 5Earlier tumor detection with SP70-targeted micro-CT imaging. SPC-A1-luc orthotopic xenograft tumors were detected by micro-CT scan weekly. Each group contained three mice. One image of each group is shown.\n\nDiscussion {#Sec8}\n==========\n\nThe insidious onset of lung adenocarcinoma usually results in metastases at the time of initial diagnosis^[@CR13],[@CR14]^. Early diagnosis of lung adenocarcinoma is challenging. Traditional imaging examinations such as CT and MRI are based on changes in tissue morphology^[@CR15]--[@CR18]^, and often lack sufficient specificity to fully predict a tumor's behavior^[@CR19]--[@CR21]^. Therefore, it is essential to find tumor-specific molecules and establish reliable imaging methods for accurate diagnosis.\n\nThe SP70 protein is expressed in lung adenocarcinoma but not in normal tissue, blood cells or other cell lines, such as U87 (a glioma cell line). And NJ001 also exhibits anti-tumor activity against NSCLC both *in vitro* and *in vivo*^[@CR12]^. SP70 is detectable in the serum of patients with lung adenocarcinoma. We also found that SP70 is a key protein that can regulate the expression of numerous genes (GEO accession number: GSE59655), promoting cancer cell proliferation and metastasis. Molecularly targeted imaging will allow clinicians to not only to see where a microscopic tumor is located in the body at an early stage, but also to specifically visualize the biological characteristics of tumors^[@CR22]--[@CR24]^. However, biomarker discovery becomes challenging because only a few biomarkers can be both diagnostic and prognostic^[@CR25]^. Thyroid transcription factor-1 (TTF-1) and napsin A have been well known as lung adenocarcinoma biomarkers for immunohistochemistry^[@CR26]^. However, they are not suitable for the application of targeted imaging technology due to their nononcogenic origins. Recently, Predina *et al*. reported that folate receptor alpha (FR\u03b1)-targeted intraoperative molecular imaging might be used during surgery^[@CR27]^.\n\nThree kinds of *in vivo* molecular imaging techniques were developed in this study: BLI based on the activity of luciferase that catalyzes the substrate luciferin in transfected cells, targeted fluorescence imaging with CF750-NJ001, and targeted micro-CT with NJ001-conjugated nanomagnetic beads. Since bioluminescence signal intensity positively correlates with tumor volume, we could perform a quantitative assessment of tumor growth *in vivo*. It was reported that molecular fluorescence imaging with near-infrared fluorescence could be used for detecting tumors and guiding surgery^[@CR28]--[@CR30]^. We chose the near infrared fluorescent dye CF750 to label NJ001 due to the wavelength's enhanced tissue penetrability and the dye's favorable pharmacokinetic properties. The fluorescent signal perfectly matched the bioluminescence signal. Conversely, there were no fluorescent signals in mice bearing glioma U87-luc cells, which were negative for the SP70 antigen. These results demonstrated that CF750-NJ001 had high specificity for lung subcutaneous xenografts.\n\nFurthermore, lesions in orthotopic lung models could be detected by SP70-targeted fluorescence imaging at the third week, while positive signals were found at the sixth week by conventional micro-CT. Considering photon attenuation when penetrating the body cavity, lesions may contain more than 600 tumor cells. Routine CT scans can only detect lesions of more than 1.5\u2009mm lesions which contain approximately10^6^ tumor cells. This result indicates that SP70-targeted fluorescence imaging is much more sensitive than traditional imaging techniques. Therefore, SP70 acts as an oncogenic protein and could be a target of molecular imaging, even for micrometastasis. Because of the weak penetration ability in the human body, SP70-targeted fluorescence imaging might only be used to locate lung adenocarcinomas and identify positive margins during surgery.\n\nCT is the first-line imaging tool for lung cancer diagnosis. Though contrast agents can enhance the resolution of CT, they cannot improve its ability to differentiate malignant from benign tumors. Current contrast agents are mainly (approximately 90%) excreted by kidneys. A larger dose of contrast agent often leads to the occurrence of nephropathy^[@CR31]^. Nanotechnology could be used in molecularly targeted imaging. Nanoparticles easily penetrate cancerous tissue capillaries^[@CR32]^, and can be deposited preferentially in cancer tissue. Recently, Patrick *et al*. used CNA35-conjugated gold nanoparticles as a CT contrast agent for the molecular imaging of myocardial scars^[@CR33]^. In our study, the diameter of the nanomagnetic beads was 180\u2009nm, so they could easily be distributed into cancer tissue. NJ001-conjugated nanomagnetic beads could specifically identified lung adenocarcinoma cells, and molecularly targeted enhancement significantly improved the sensitivity of micro-CT imaging for monitoring tumor growth. Lesions were detected at the fourth week in the SP70-targeted micro-CT group, 2 weeks (one third of the amount of time) earlier than in the groups injected with normal saline and bare nanomagnetic beads controls. It could be inferred that SP70-targeted imaging could greatly shorten the period from tumor onset to diagnosis, and therefore accomplish the aim of detecting the early stage of lung adenocarcinoma. Although the antigenicity of the murine monoclonal antibody NJ001 was minimized with conjugation to the nanomagnetic beads, its biosafety and metabolic properties remain to be determined.\n\nIn conclusion, SP70-targeted imaging can markedly improve the detection of lung adenocarcinoma detection. Molecularly targeted imaging with NJ001-labeled probes may have precision medical applications for the early diagnosis of lung adenocarcinoma.\n\nMaterials and Methods {#Sec9}\n=====================\n\nMaterials {#Sec10}\n---------\n\nSuperparamagnetic polymer nanospheres were purchased from Shanghai Allrun Nano New Science & Technology Ltd. (Shanghai, China). Bovine serum albumin (BSA), ethylcarbodiimide (EDC) and N-hydroxysuccinimide (NHS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). The CF488 Dye Antibody Labeling Kit and the CF750 Dye Antibody Labeling Kit, which are both fluorescent succinyl ester dye kits, were purchased from Biotium (Hayward, CA, USA). D-luciferin was purchased from Fanbo Biochemicals (Beijing, China).\n\nCell culture {#Sec11}\n------------\n\nSPC-A1-luc cells (SP70-positive) and U87-luc cells (SP70-negative), cell lines derived from human lung adenocarcinoma and human glioma, respectively, that express luciferase by stable transfection with the firefly luciferase gene, were purchased (Shanghai Baidaian Company, Shanghai, China). Both cell lines were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium containing 50\u2009U/mL penicillin, 50\u2009U/mL streptomycin, and 10% fetal bovine serum (Invitrogen, Carlsbad CA, USA). Cells were grown in an incubator with 5% CO~2~ at 37\u2009\u00b0C.\n\nAnimal models {#Sec12}\n-------------\n\nFour-week-old male BALB/c nude mice (Shanghai SLAC Laboratory Animal Co. Ltd. Shanghai, China) were maintained in a pathogen-free environment. All procedures were conducted in accordance with the Animal Care and Use Committee guidelines of Nanjing Medical University. To establish subcutaneous xenograft mouse models, SPC-A1-luc and control U87-luc cells were harvested in the logarithmic growth phase, counted, and resuspended in phosphate-buffered saline (PBS) to a final density of 5\u2009\u00d7\u200910^7^ cells/mL. Next, 5\u2009\u00d7\u200910^6^ SPC-A1-luc cells or U87-luc cells suspended in 0.1\u2009mL of sterile PBS were implanted subcutaneously into the right flank of each mouse. For an orthotopic lung tumor implantation mouse model, 5\u2009\u00d7\u200910^6^ SPC-A1-luc cells suspended in 0.1\u2009mL of sterile PBS were injected into mice intravenously via the tail vein. All animal studies were performed with the approval of the Animal Care and Use Committee of Nanjing Medical University.\n\nNJ001-conjugated nanomagnetic beads (immuno-nanomagnetic beads) {#Sec13}\n---------------------------------------------------------------\n\nSuperparamagnetic polymer nanospheres with diameters of 180\u2009nm were conjugated with NJ001 antibody. First, 2\u2009mg nanomagnetic beads were activated with 200\u2009\u00b5L of 5\u2009mg/mL EDC and NHS in a 37\u2009\u00b0C water bath for 30\u2009min, followed by three washes with phosphate-buffered saline with tween 20 (PBST). Second, the beads were mixed with 50\u2009\u00b5L of 5\u2009mg/mL NJ001 antibody and again incubated in a 37\u2009\u00b0C water bath for 4\u2009h. Third, after three more washes with PBST, the beads were resuspended in 1\u2009mL of PBST containing 5% BSA and gently shaken at 4\u2009\u00b0C overnight. Finally, the supernatant was removed, and the beads were resuspended in 200\u2009\u00b5L of PBST and kept at 4\u2009\u00b0C.\n\nImmuno-nanomagnetic bead characterization {#Sec14}\n-----------------------------------------\n\nTEM were taken by using a JEM-2100 HR transmission electron microscope (JEOL, Tokyo, Japan). The nanomagnetic bead sample (5\u2009\u00b5L) was dropped onto copper 200-mesh grids that were pretreated with UV light to reduce static electricity. After 30\u2009min, the solvent was drained with filter paper, and a phosphotungstic acid stain solution (1% by weight, adjusted to pH 6.0) was applied for 30\u2009s. After drying, TEM imaging was performed.\n\nDirect immunofluorescence (DIF) analysis {#Sec15}\n----------------------------------------\n\nThe Dye Antibody Labeling Kit was used to develop the fluorescently dyed probe CF488-NJ001 for *in vitro* DIF. SPC-A1-luc and U87-luc cells were grown on coverslips, fixed in 95% ethanol for 15\u2009min at room temperature and blocked with 5% BSA for 30\u2009min at 37\u2009\u00b0C. After washing three times with PBS, the coverslips were incubated with CF488-NJ001 (1:200 dilution) for 30\u2009min at 37\u2009\u00b0C. After three washes, a fluorescent dye, 4\u2032,6-diamidino-2-phenylindole (DAPI) was subsequently added for nuclear staining. Samples were examined using an Olympus IX71 inverted fluorescence microscope (Olympus Optical Co. Ltd, Tokyo, Japan) coupled with a charge-coupled device (CCD) camera (ProgRes, Jenoptik/Jena, Germany).\n\nBioluminescence imaging (BLI) {#Sec16}\n-----------------------------\n\nMice were injected intraperitoneally with 150\u2009mg/kg of D-luciferin in PBS, and imaged with IVIS 2000 imaging system (Caliper Life Sciences, Hopkinton MA, USA). Analyses of the BLI images were performed using Living Image software from Caliper Life Sciences by drawing regions of interest over the tumor region and obtaining maximum values in photons per second per cm^2^ per steradian or total flux as photons per second. BLI was repeated weekly.\n\nFluorescence imaging (FLI) {#Sec17}\n--------------------------\n\nFluorescence imaging was performed using the CF750 Dye Antibody Labeling Kit, a fluorescent succinyl ester dye kit, to label NJ001 for *in vivo* fluorescence imaging analysis. A dose of 5\u2009\u00b5g of CF750-NJ001 in 100\u2009\u00b5L PBS was injected via the tail vein one week after the cell inoculation and weekly thereafter. The fluorescence intensity in each mouse was assessed 48\u2009h after injection using the same IVIS 2000 imaging system with an excitation wavelength of 755\u2009nm and an emission wavelength of 800\u2009nm. Fluorescence from the ROI was defined manually, and fluorescence efficiency was expressed as (photons/s)\u2009\u00f7\u2009(\u00b5W/cm^2^)^[@CR34]^. Images were analyzed using Living Image software.\n\nMicro-computed tomography (micro-CT) {#Sec18}\n------------------------------------\n\nOrthotopic mouse models were imaged with BLI to monitor tumor growth one week after SPC-A1-luc cell inoculation. Nine mice with no significant difference in tumor size were selected for the following micro-CT scans. Three orthotopic mice were injected intravenously with 100\u2009\u00b5L NJ001-conjugated nanomagnetic beads, three mice were injected with 100\u2009\u00b5L normal saline and three mice were injected with bare nanomagnetic beads as controls. Micro-CT scans were acquired at 0, 2, 4, 6 and 24\u2009h after injection to determine tumor lesion and size. Micro-CT was performed on a SkyScan 1176 (SkyScan NV, Kontich, Belgium), a small animal imager, at 50\u2009kV and 490\u2009\u00b5A. A total of 360\u00b0 views were acquired at 1\u00b0 angle increments, each for an exposure time of 120\u2009ms, to give a resolution of 35\u2009\u00b5m using external respiratory gating. SkyScan software was used for multiplanar and 3D image reconstruction.\n\nStatistical analysis {#Sec19}\n--------------------\n\nStatistical analysis was conducted using SPSS Statistics v20.0 (SPSS Inc., Chicago IL, USA) and GraphPad Prism v6.0 (GraphPad Software, San Diego CA, USA). Dunnett's test was used to compare the experimental group to the control groups undergoing micro-CT scans. The correlation was analyzed by the Pearson correlation coefficient. All statistical assessments were performed in a two-sided setting, using a significance cut-off of 0.05.\n\nSupplementary information\n=========================\n\n {#Sec20}\n\nSupplementary information\n\n**Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThese authors contributed equally: Jian Xu and Shichang Zhang.\n\nSupplementary information\n=========================\n\nis available for this paper at 10.1038/s41598-020-59439-9.\n\nThis work was supported by grants from the National Natural Science Foundation of China (Grant ID 81672100), and the Key Laboratory for Laboratory Medicine of Jiangsu Province of China (Grant ID ZDXKB2016005) and was funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions. The authors are grateful to the Animal Core Facility of Nanjing Medical University.\n\nS.P. conceived the study and wrote the manuscript. J.X., S.Z. W.Z., E.X., M.G., Y.W., L.Y., B.Z. and D.L. performed the experiments. J.Z. performed the statistical analysis. C.G., T.X. and F.W. contributed methodological support. P.H. critically revised the manuscript.\n\nAll data generated or analysed during this study are included in this published article (and its Supplementary Information Files).\n\nThe authors declare no competing interests.\n"}
+{"text": "In the 19th century, Rudolf Virchow first proposed a potential link between inflammation and cancer based on his observations on the presence of leukocytes in tumors.[@bib1] Inflammation is a physiological mechanism evolved for wound healing and therefore is counter-intuitive to consider it to be oncogenic. Nevertheless, inflammation is a 'double-edged sword' with a pathologic edge that can promote various aspects of tumorigenesis deregulated such as cell proliferation, migration, angiogenesis, and apoptosis.[@bib1] Within the last decade, a multitude of studies demonstrating the a) abundance of inflammatory cells such as macrophages and fibroblasts in cancer biopsies, b) the role of proinflammatory molecules such as cyclooxygenase-2 (COX-2), prostaglandin E2, leukotrienes, transforming growth factor beta (TGF-\u03b2), hypoxia inducible factor-1 alpha, vascular endothelial growth factor (VEGF), nitric oxide synthase, nitric oxide, reactive oxygen species (ROS), cytokines and chemokines in the pathogenesis of several cancers, and the tumorigenic nurturing properties of the proinflammatory tumor microenvironment strongly indicates that inflammation plays a pathogenic role in several cancers.[@bib1], [@bib2], [@bib3], [@bib4], [@bib5], [@bib6], [@bib7], [@bib8] Chronic persistent inflammation is believed to play an important role in the pathogenesis of 15% of all malignancies.[@bib1], [@bib2], [@bib3], [@bib4], [@bib5] Depending on the type and stage of cancer, the physiological to pathologic switch of inflammation is triggered by various factors such as genomic instability, epigenetic changes, somatic mutations, tumor suppressor and oncogene mediated carcinogenesis, chronic persistent infections, and environmental stressors such as pollutants.[@bib1], [@bib7], [@bib8]\n\nThe role of tumor viruses in chronic persistent inflammation associated carcinogenesis is demonstrated in several malignancies such as Kaposi's sarcoma associated-herpes virus (KSHV/HHV-8) in Kaposi's sarcoma (KS) and primary effusion lymphoma (PEL), Epstein-Barr virus (EBV) in Burkitt's lymphoma and nasopharyngeal carcinoma, human papillomavirus (HPV) in cervical cancer, hepatitis B (HBV) and hepatitis C viruses (HCV) in hepatocellular cancer, and human T-lymphotropic virus (HTLV) in T-cell leukemia.[@bib6], [@bib9], [@bib10], [@bib11] Viruses are obligate intracellular parasites and use host proteins for genome replication and production of progeny.[@bib12] Piracy of inflammatory mechanisms is a recurring theme in the story of infections by KSHV, EBV, HCV, HPV, HBV, and HTLV because of the proliferative, angiogenic, immune-suppressive, and antiapoptotic niche that persistent inflammation provides.[@bib11] The purpose of this review is to highlight the salient findings demonstrating how KSHV uses the pivotal COX-2/PGE2/EP receptor mediated inflammatory axis for its survival and pathogenesis and, therefore, plays a crucial role in KSHV-associated malignancies.\n\nCOX-2 and Cancers {#sec1}\n=================\n\nCOX or prostaglandin-endoperoxide synthase catalyzes the conversion of arachidonic acid (AA) into prostaglandin H2, which is further converted into the proinflammatory lipid metabolites such as PGE2, PGI2, PGF2, and thromboxane-2 by specific enzymes and play crucial roles in diverse physiological functions such as platelet aggregation, inhibition of gastrointestinal (GI) acid secretion, regulation of glomerular function, and labor.[@bib13] The COX-1 isoform has a constitutively active promoter whereas COX-2 has an inducible promoter activated by stress, growth factors, cytokines, and infections.[@bib13] Numerous studies have demonstrated the induction of COX-2 and associated inflammatory pathways in the pathogenesis of several cancers such as colorectal, prostate, lung and breast cancers, as well as several hematological malignancies including chronic lymphocytic leukemia, Hodgkin's and non-Hodgkin's lymphomas (NHLs), and multiple myeloma.[@bib5], [@bib14], [@bib15], [@bib16], [@bib17], [@bib18] In recent years, COX-2 has been investigated as a potent chemotherapeutic target due to the well-studied anticancer properties of nonsteroid anti-inflammatory drugs (NSAIDs).[@bib5], [@bib13]\n\nThe major lipid metabolite of COX-2 implicated in tumorigenesis is PGE2.[@bib19] PGE2 is an autocrine and paracrine lipid signal inducer with a circulating half-life of approximately 30 seconds and normal plasma levels varying from 3-15 pg/mL.[@bib20] PGE2 exerts its effects through the 7-transmembrane rhodopsin family of G protein coupled (GPCR) eicosanoid (EP) receptors, namely, EP1, EP2, EP3, and EP4 (EP 1-4) that initiate signal transduction through Calcium (Ca)^2+^, Cyclic adenosine monophosphate (cAMP), protein kinase A (PKA), and Phosphatidylinositide 3-kinase (PI3K).[@bib5], [@bib21] EP receptor induction has been associated with several oncogenic pathways including Src, PI3K, PKC, NF\u03baB, Ras/Raf, ERK, VEGF, AKT/PI3K, PPAR, and interleukin (IL)-10/DAF and, therefore, forms the mechanistic context underlying the diverse aspects of COX-2 mediated tumorigenesis.[@bib22], [@bib23], [@bib24] In recent years, the link between EP receptors and tumorigenesis had also revealed the possibility of using highly specific EP receptor antagonists such as SC-51322 (EP1 antagonist), AH6809 (EP2 antagonist), and AH23848 (EP4 antagonist) as anticancer drugs.[@bib25], [@bib26], [@bib27]\n\nViral Infections and COX-2 {#sec2}\n==========================\n\nInfections by several viruses have been shown to regulate COX-2 expression and PGE2 production such as HBV in hepatocytes,[@bib28], [@bib29] HCV in Huh-7 cells,[@bib30] human herpesvirus 6 (HHV-6) in monocytes,[@bib31] human cytomegalovirus (CMV) in Peripheral blood mononuclear cells (PBMCs), smooth muscle cells, and fibroblasts,[@bib32], [@bib33], [@bib34], [@bib35] murine gammaherpesvirus 68 (MHV-68) in NIH 3T3 cells,[@bib36] HIV in monocytes,[@bib37], [@bib38] HTLV-1 in PBMCs,[@bib39] influenza virus in PBMCs,[@bib40] enterovirus 71 in human neuroblastoma cells,[@bib41] dengue virus in dendritic cells,[@bib42] *Severe acute respiratory syndrome* (SARS)-associated coronavirus in 293T cells,[@bib43] Theiler's murine encephalomyelitis virus in astrocytes,[@bib44] encephalomyocarditis virus in macrophages,[@bib45], [@bib46] coxsackie virus B3 in monocytes,[@bib47] respiratory syncytial virus in macrophages and dendritic cells,[@bib48] and canine distemper virus in monocytes.[@bib49] COX-2/PGE2 has been implicated in a multitude of viral mechanisms such as genome replication (HBV), (CMV, HTLV), gene expression (MHV-68), transmission (HTLV), cell tropism (rhesus CMV), cell invasion (CMV), T cell regulation (HIV), and even has identified a viral homologue of COX-2 in rhesus CMV revealing the significance of COX-2 in the evolution of inflammation mediated viral pathogenesis.[@bib28], [@bib29], [@bib30], [@bib31], [@bib32], [@bib33], [@bib34], [@bib35], [@bib36], [@bib37], [@bib38], [@bib39], [@bib40], [@bib41], [@bib42], [@bib43], [@bib44], [@bib45], [@bib47], [@bib48], [@bib49], [@bib50], [@bib51] Among the herpes viruses, studies using COX inhibitors have shown the role of COX-2/PGE2 pathways for replication and successful lytic cycle in HSV, CMV, HHV-6, and MHV-68.[@bib31], [@bib34], [@bib35], [@bib36], [@bib51], [@bib52], [@bib53], [@bib54], [@bib55], [@bib56], [@bib57], [@bib58] However, the role of the extensive molecular framework underlying the COX-2/PGE2/EP receptor inflammatory axis in herpes viral latency is described only in KSHV associated malignancies such as KS and PEL.[@bib59], [@bib60], [@bib61], [@bib62], [@bib63], [@bib64], [@bib65]\n\nKSHV Associated Diseases {#sec3}\n========================\n\nKSHV/HHV-8 is grouped in the \u03b3-2 herpes virus family and is the etiologic agent underlying KS, PEL, and multicentric Castleman's disease.[@bib66], [@bib67], [@bib68], [@bib69], [@bib70] Like other herpes viruses, the KSHV life cycle is characterized by 2 phases, the latent and the lytic cycles.[@bib70] After infection, KSHV enters the latency phase, where the virus remains evasive by transforming the infected cell into a stable reservoir.[@bib66], [@bib67], [@bib68], [@bib69], [@bib70] The lytic cycle results in the replication of the viral genome and production of new viral progeny.[@bib70] Both life cycles are associated with distinct viral proteins.[@bib70] Gene expression profiles of KS, PEL, and multicentric Castleman's disease biopsies have shown that the majority of tumor cells express latency transcripts with 1%-3% of tumor cells undergoing the lytic cycle at a given time point and both stages of the life cycle are implicated in the pathogenesis of KSHV associated diseases.[@bib70] Although, there are no specific treatments targeting KSHV associated diseases, highly active antiretroviral therapy (HAART) and consequent immune reconstitution is demonstrated to be beneficial in treating AIDS-KS.[@bib70], [@bib71], [@bib72]\n\nKS {#sec3.1}\n--\n\nEpidemiologically, KS is classified into 4 subgroups: (1) classical KS as described by Moritz Kaposi in elderly men of Mediterranean origin in 1872,[@bib73] (2) endemic KS in sub-Saharan Africa, (3) epidemic KS in AIDS patients, where KS forms the most common AIDS associated malignancy, and (4) transplant defining KS.[@bib74], [@bib75], [@bib76] Pathologically, KS is a multifocal angioproliferative tumor of vascular nature characterized by extravascular erythrocytes, spindle shaped cells of endothelial origin, inflammatory cells such as monocytes, fibroblasts, neutrophils, and lymphocytes interspersed between narrow, irregular angulated slits within a proinflammatory and angiogenic microenvironment.[@bib70] Fatality by KS is often due to systemic spread into the respiratory system, gastrointestinal tract, lymph nodes, and other organs.[@bib70]\n\nPEL {#sec3.2}\n---\n\nPEL is a rare, yet aggressive form of B cell lymphoma that accounts for 2%-4% of all AIDS associated NHLs with a prognosis of less than 6 months.[@bib66], [@bib67], [@bib69], [@bib71], [@bib77] PEL is characterized by primary lymphomatous aggregations within the major body cavities such\u00a0as the pleura, pericardium, and the peritoneum.[@bib66], [@bib67], [@bib69], [@bib71], [@bib77] Pathologically, PEL cells show\u00a0varying phenotypes, such as immunoblastic, plasmablastic, and anaplastic, and are proposed to lie\u00a0between the pro-B cell and plasma cell lineage.[@bib66], [@bib67], [@bib69], [@bib71], [@bib77] PEL cells are characterized by B cells transformed by persistent KSHV infection and consists of multiple copies (in the order of 50-150 copies/cell) of episomal KSHV genomes with the latent viral gene expression pattern involving latency associated nuclear antigen (LANA)-1, viral homologues of host proteins cyclin (vCyclin) and FLICE-inhibitory protein (Viral-FLICE-inhibitory protein (vFLIP)), a pre-microRNA transcript encoding viral microRNAs, as well as vIRF3/K10.5/LANA-2 and a homologue of IL-6 (vIL-6) is also expressed in some PEL cells.[@bib66], [@bib67], [@bib69], [@bib71], [@bib77] PEL cells express a variety of cell surface markers from different stages of B cell development such as the activation markers CD30, CD38, and CD71 and several plasma cell markers including CD138, VS38c, and MUM-1/IRF4 but are devoid of the B cell markers CD19 and CD20.[@bib66], [@bib67], [@bib69], [@bib71], [@bib77]\n\nKSHV Latency, Inflammation, and COX-2 {#sec4}\n=====================================\n\nKSHV latency is proposed to be a symphony of well-orchestrated interactions between viral and host proteins leading to the transformation of infected cells for viral survival through successful genome replication and immune evasion.[@bib70], [@bib78], [@bib79], [@bib80], [@bib81] The host and viral protein interactions initially established by KSHV infection for survival through the establishment and maintenance of latency progress pathologically as KS\u00a0and PEL under conditions of persistent selective pressures such as AIDS related or transplant associated immune suppression.[@bib70], [@bib79], [@bib80], [@bib81] The decrease in the incidence of KS post-HAART therapy in AIDS patients is suggestive of this scenario.[@bib70], [@bib71], [@bib72] The host mechanisms underlying the establishment and maintenance of KSHV infection and KSHV associated malignancies include cell signaling, anti-apoptosis, angiogenesis, immune modulation, and cell proliferation mediated by cytokines, growth factors, and inflammatory molecules.[@bib70], [@bib79], [@bib80], [@bib81] Thus, identification of molecules used by the KSHV latency program will enable us to delineate the pathogenesis of KS and PEL as well.\n\nStudies by Naranatt et\u00a0al (2004)[@bib82] and Sharma-Walia et\u00a0al (2006)[@bib61] first indicated the induction of COX-2 during *de novo* KSHV infection of human microvascular dermal endothelial (HMVEC-d) cells Studies by introduced a novel idea regarding the functional significance of COX-2/PGE2 within the context of the KSHV latency program. KSHV infection induced COX-2/PGE2 and PGE2 supplementation reversed the COX-1/COX-2 inhibitor mediated downregulation of latency gene LANA-1. Therefore, the studies for the first time generated a hypothesis that KSHV infection induced COX-2/PGE2 is crucial for establishment and maintenance of latency.[@bib61] Considering the oncogenic potency of COX-2 through the activation of inflammatory mechanisms, the proinflammatory mechanisms underlying KS and PEL pathogenesis, and the well characterized roles of COX-2 in other viral tumors such as Burkitt's lymphoma and cervical cancer, the study by Sharma-Walia et\u00a0al (2006)[@bib61] also raised several important questions as follows. (1) What are the different biological mechanisms regulated by COX-2 in KS? (2) What are the mechanisms underlying sustained COX-2 activation in the KSHV latency program? (3) How does COX-2/PGE2 regulate the KSHV latency program? (4) What is the role of COX-2 in PEL? (5) Do NSAIDs and EP receptor antagonists hold chemotherapeutic potential in treating KS and PEL? (6) Does simultaneous blockade of COX-2 and EP receptors provide synergistic anticancer effects?\n\nInduction of COX-2 and EP Receptors by KSHV {#sec5}\n===========================================\n\nSeveral gene array studies have demonstrated the induction of COX-2 in a multitude of malignant and premalignant human cancer lesions with progressive increase in expression as the stage of the cancer advances.[@bib83] We demonstrated CD31-COX-2 double stained spindle shaped cells in tissue microarray of human KS sections (eye orbit, tonsil, mouth, and small bowel) (Sharma-Walia et al (2010).[@bib59] Similarly, abundant expression of mPGES, PGE2, and EP1-4 was observed in human KS biopsies (George Paul et al (2010).[@bib62] Collectively, these findings corroborate with earlier in vitro observations[@bib61], [@bib82] and is the first detailed investigation of COX-2 and EP receptors in human KS and is the first detailed investigation of COX-2 and EP receptors in human KS biopsies. There are several possible mechanisms underlying COX-2/PGE2/EP receptor induction in KS lesions are several such as persistent KSHV infection, persistent chronic inflammation, and pathologic stress from chronic persistent infection and inflammation in KS patients.\n\nThe association of COX-2 and cancer is attributed to its inducible promoter activated by stress, infection, and inflammation.[@bib13], [@bib24] The role of infection mediated signaling in COX-2 induction is demonstrated by several viruses such as CMV (ROS, cAMP/nuclear factor of activated T cells (NFAT)), HBV\u00a0(Ca^2+^/ROS, cAMP/NFAT), HCV (Ca^2+^/ROS), encephalomyocarditis virus (EMCV) (NF\u03baB/MAPK/c-Jun N-terminal *Kinases*/p38), Enterovirus 71 (NF-\u03baB/AP-1/PKA/cAMP/Src/EGFR/p300/cAMP response element-binding (CREB)), SARS (NF\u03baB/CEBP), and Dengue virus (NF\u03baB/AP-1).[@bib28], [@bib29], [@bib30], [@bib31], [@bib41], [@bib42], [@bib43], [@bib45], [@bib46] We examined whether similar molecular mechanisms were at work in the induction of COX-2 in KS and demonstrated that *de novo* KSHV infection and exogenous PGE2 upregulates COX-2 promoter activity through Src, PI3K, PKC, focal adhesion kinase, JNK, p38, cAMP, PKA, and NF\u03baB.[@bib60] Promoter analysis using COX-2 promoter deletion constructs and mutation reporter constructs identified transcription factors CREB and NFAT cells downstream of the signal cascades synergistically modulating COX-2 promoter activity.[@bib60] A multitude of similar pathways are also activated by early KSHV binding and infection such as PI3K-PKC-\u03b6-MEK-ERK,[@bib84], [@bib85] Src-PI3K-RhoGTPase,[@bib86] RhoA-GTP-Diaphenous-2 microtubules,[@bib87], [@bib88] NF-\u03baB,[@bib89] FAK,[@bib90] VEGF,[@bib91] and lipid rafts[@bib92] indicating that KSHV entry associated signal transduction events and infection-induced PGE2 secretion work in concert to activate the COX-2 promoter.[@bib60] However, the presence of strong COX-2 expression in KS lesions strongly suggests the existence of viral mechanisms that sustain COX-2 expression post establishment of KSHV latency.\n\nCOX-2 induction has been demonstrated by other viral proteins as well such as Tax protein (HTLV-1), gp120 (HIV), HBx (HBV), and CoV N-protein (SARS virus).[@bib29], [@bib37], [@bib43], [@bib93] KSHV latency protein vFLIP and lytic\u00a0proteins KSHV G protein-coupled receptor, a constitutively active lytic phase protein with significant homology to the human IL-8 receptor, and K15 are the viral proteins proposed to be capable of inducing COX-2. v-FLIP has been shown to induce COX2 in other studies,[@bib94], [@bib95] Recently, we[@bib64] delineated the detailed mechanistic aspects of vFLIP mediated COX-2 expression that is mediated through NF\u03baB, p38, RSK, and transcription factor CREB. In addition, vFLIP activated COX-2 expression and PGE2 secretion was demonstrated to be part of a signaling loop where COX-2/PGE2 was required for vFLIP-induced NF-\u03baB activation.[@bib64] The induction of COX-2 by lytic proteins KSHV G protein-coupled receptor and K15 also raises the question of whether COX-2 plays a role in the lytic cycle and is still being investigated.[@bib96], [@bib97]\n\nThe induction of EP receptors by viral infections is largely an unexplored arena. Studies by George Paul et\u00a0al (2010)[@bib62] demonstrated that EP1, EP3, and EP4 protein levels are significantly upregulated in long-term-KSHV-infected endothelial cells. We observed upregulation of EP1-4 receptors in de novo KSHV infected HMVEC-d cells[@bib65] too. EP receptors are present in endothelial cells because of their general homeostatic functions such as GI mucosal protection.[@bib13] However, their pathologic upregulation is a characteristic of many malignancies such as colorectal cancer[@bib19] and, therefore, the work by George Paul et\u00a0al (2010)[@bib62] and George Paul et\u00a0al (2013)[@bib65] is strongly suggestive of their role in KS pathogenesis. Further work is required to\u00a0characterize the signaling and transcriptional mechanisms underlying the induction of EP receptor expression.\n\nBiological Mechanisms Regulated by COX-2 in KS {#sec6}\n==============================================\n\nThe pathogenesis of KS lesions consisting of spindle shaped endothelial cells, neovascular structures, and inflammatory cells is profoundly influenced by growth factors (GFs), proinflammatory cytokines (ICs), angiogenic factors (AFs) such as basic and acidic fibroblast growth factor (bFGF, aFGF), IL-1\u03b1 and IL-1\u03b2, granulocyte-monocyte colony stimulating factor (GM-CSF), platelet derived growth factor \u03b2 (PDGF-\u03b2), VEGF, interferon-\u03b3 (IFN-\u03b3), IL-6, tumor necrosis factor-\u03b1 (TNF-\u03b1), angiopoietin-2, angiogenin, heme oxygenase-1, TGF-\u03b2, adhesion molecules like intercellular/vascular-cell adhesion molecules (ICAM-1 and VCAM-1), and matrix metalloproteinases (MMPs) like MMP-1, -2, -3, -9, and -19.[@bib68], [@bib79], [@bib80], [@bib81] The combined effect of these molecules contributes to the\u00a0different aspects of KS pathogenesis such as neovascularization, angiogenesis, maintenance of KSHV latency, and metastasis whereas the mechanisms underlying the sustained activation of these molecules is still an active area of investigation. Considering the well-established roles of COX-2 in the progression of several cancers, Sharma-Walia et\u00a0al (2010)[@bib59] characterized the role of COX-2 in KSHV pathogenesis related processes such as secretion of GFs, ICs, AFs, MMPs, and ICAMs in endothelial cells. In *de novo* infected HMVEC-d cells, key molecules proposed to be important for KS pathogenesis, such as immune modulators (TNF-\u03b1, IFN-\u03b3, Stromal cell-derived factor-1, growth regulated oncogene, Regulated on activation, normal T cell expressed and secreted), cytokines (IL-8, IL-1\u03b2, IL-1 \u03b1,/\u03b2, ILs-2/-3/-8/-P40/-16), chemokines (MCP-2, MCP-3, TARC, MIP-1\u03b4, ENA-78, I-309, MIF, GCP-2, MIP-3-\u03b1, eotaxins -2/-3, IP-10, NAP-2, CK-\u03b28-1), growth and angiogenic factors (VEGF-A/C, PDGF-BB, MCSF, G-CSF, GM-CSF, angiogenin, oncostatin M, thrombopoeitin, SCF, insulin-like growth factor-binding protein (-2, -3, and -4), BDNF, PIGF, HGF, osteoprotegerin, NT-3, NT-4), and anti-inflammatory cytokines (IL-4, IL-13, and IL-15) and MMP-1, -9, and -10, were downregulated by the pharmacologic (NS-398, indomethacin) and small interfering RNA based inhibition of COX-2.[@bib59] COX-2 induction by KSHV infection thus play key roles in various aspects of KS pathogenesis such as angiogenesis and lymphangiogenesis (VEGF--A/C), regulation of T cell response (IFN-\u03b3, SDF-1, GRO, RANTES), chemotaxis of immune cells (SDF-1 and IL-8), inflammasome activation (IL-1\u03b2), and cell migration and metastasis (MMP-1, -9, 10, and SDF-1).[@bib59] Further characterization of the biological significance of COX-2 in KSHV infected cells has demonstrated that pharmacological and small interfering RNA based inhibition of COX-2 downregulated the (1) formation of intricate *in\u00a0vitro* capillary tubes and (2) cell adhesion and cell invasion capability of HMVEC-d cells.[@bib59] We[@bib64] have added an additional dimension to these observations by demonstrating that COX-2 inhibition could downregulate vFLIP mediated pathogenic mechanisms such as (1) expression of immune cell regulators and recruiters such as chemokines (CXCL-5 and CXCL-6), cytokines (IL-6 and IL-8), C-C ligand related molecules (CCL-2, CCL-5, CCL-20, MCP-1, RANTES-2, and MIP-3\u03b1), (2) expression of cell adhesion molecules ICAM-1, VCAM-1, and E-selectin and metalloproteinase MMP-10, (3) induction of actin cytoskeleton modulators FAK, Src, AKT, and Rac1GTPase, (4) ROS regulation through mitochondrial antioxidant enzyme manganese superoxide dismutase (MnSOD2), (5) endothelial-mesenchymal transitions (EMT) by downregulating EMT specific master regulator genes (snail, twist, slug, and laminin-\u03b31) and upregulating E-cadherin, and (6) anoikis resistance and anchorage resistant colony formation by upregulating proapoptotic proteins BIM and DR5.[@bib64]\n\nRegulation of KSHV Latency bY EP Receptors {#sec7}\n==========================================\n\nPGE2 and EP receptors are proposed to be the tumorigenic workhorses of COX-2.[@bib5] EP receptors are GPCRs and have been well-characterized in the pathogenesis of a multitude of cancers such as melanoma, breast cancer, and colon cancer by contributing to proliferation, immunosuppression, angiogenesis, invasion and blocking apoptosis through the activation of Src kinase, cAMP/CREB, PI3K/Akt, Ras/Raf, ERK-1/2, NF\u03baB, EGFR, PPAR\u03b4/\u03b2, and GSK-3\u03b2/\u03b2-catenin pathways.[@bib22], [@bib23], [@bib24] The role of Src kinase, PI3K, Akt, ERK, and NF\u03baB during the early events of KSHV infection and establishment of latency is well-characterized.[@bib84], [@bib85], [@bib86], [@bib89], [@bib98] Collectively, studies by Sharma-Walia et\u00a0al (2010; 2012)[@bib59], [@bib64] demonstrated that pathways downstream to COX-2 when activated by viral and nonviral mechanisms or both participate in enriching the tumor microenvironment and consequently various pathologic processes underlying KS such as endothelial transformation, neovascularization, and metastasis.[@bib59], [@bib64] However, pathways downstream of COX-2 resulting in the activation of Src, cAMP, PI3K, Akt, Ras/Raf, ERK, NF\u03baB, EGFR, and GSK-3\u03b2/\u03b2-catenin pathways, which form the first line of signal transducers in a molecular avalanche eventually resulting in the induction of ICs, GFs, AFs, and MMPs, is still an active area of investigation. We[@bib62] identified the involvement of EP receptors in the induction of various signaling molecules downstream of COX-2 in KSHV latency program. Specifically, the EP1 receptor was implicated in the activation of Ca^2+^, PI3K, and NF-\u03baB, the EP2 receptor in PI3K, PKC\u03b6/\u03bb, and NF\u03baB activation and the EP4 receptor in PI3K, PKC\u03b6/\u03bb, ERK 1, ERK 2, and NF-\u03baB activation in long-term-infected cells.[@bib62]\n\nEP1, EP2, and EP4 antagonists could also downregulate the expression of major KSHV latency gene LANA-1 by inhibiting the induction of Ca^2+^, phosphorylation of Src, PI3K, PKC\u03b6/\u03bb, and NF\u03baB signaling.[@bib62] The signal molecules that regulate the COX-2 promoter and PGE2 induced LANA-1 promoter activity were found to be similar to the EP receptor mediated signal transduction pathways in latently infected endothelial cells and COX-2 gene expression and PGE2 secretion was also significantly downregulated by the pharmacologic inhibition of EP2 and EP4 receptors.[@bib62] These observations implicate for the first time the role of EP receptors in any form of herpesvirus latency and, thus, substantiating the earlier observations by Sharma-Walia et\u00a0al (2006)[@bib61] and elucidated signal transduction network through EP receptors, initiated by KSHV infection mediated COX-2 activation and PGE2 secretion.[@bib62] PGE2 in the tumor microenvironment activates EP receptor mediated signal cascades in a paracrine and autocrine fashion that exert its effects on LANA-1 and COX-2 expression.[@bib62] Consequently, a self-sustained positive feedback loop networking the KSHV protein LANA-1 to the proinflammatory pathways regulated by COX-2/PGE2/EP receptors is created by viral infection. Recent work by Dupuy et\u00a0al (2012) further substantiates the role of EP receptors in KS pathogenesis by reporting the use of PGE2 inhibitors as an attractive approach to treat aggressive KS, as they could restore activation and survival of tumoricidal NK cells.[@bib99] These studies provided strong evidence that down-modulation of NKG2D is mediated by inflammatory PGE2, known to be released by KS cells, and also showed that PGE2 acts by preventing IL-15-mediated activation of NK cells.[@bib99] The role of EP receptors in the induction of several KSHV associated signal networks and consequently various pathogenic mechanisms is indicative of how KSHV subverts the COX-2/PGE2/EP receptor mediated protumorigenic signal pathways to sustain viral and host gene expression.[@bib62] However, these studies also demonstrated that neither chemical inhibitors (NS-398 and indomethacin) nor si-COX-2 could completely abolish the induction of ICs, GFs, MMPs, and AFs indicating the presence of a multitude of host molecules like COX-2 subverted by KSHV infection.[@bib59], [@bib60], [@bib62], [@bib63], [@bib64]\n\nRole of COX-2 in KSHV Associated B Cell Neoplasia (PEL) {#sec8}\n=======================================================\n\nPEL is comprised of B cells transformed by KSHV latent infection.[@bib77], [@bib100], [@bib101] Studies have proposed the cumulative interdependent vitality of the expression of\u00a0KSHV latency genes, the proinflammatory environment, and the manipulation of canonical anticancer host defense machinery, such as p53 and\u00a0p21, in the metamorphosis of PEL neoplasia.[@bib63], [@bib77], [@bib100], [@bib101] The mechanistic role of COX-2 in hematological malignancies[@bib18] and KSHV latency program in endothelial model systems is well established.[@bib59], [@bib60], [@bib61], [@bib62], [@bib64] The study by Paul et\u00a0al (2011)[@bib63] for the first time delineated the role of COX-2 in PEL pathogenesis using the COX-2 inhibitor nimesulide. Nimesulide downregulated KSHV latency genes vFLIP and LANA-1 and induced G1 cell cycle arrest and apoptosis through the activation of the p53/p21 tumor suppressor pathway and downregulation of cell survival kinases p-Akt1/2 and p-GSK-3\u03b2, and angiogenic factor VEGF-C in PEL cells.[@bib63] LANA-1 is a multifunctional protein and a major marker for KSHV latency.[@bib70], [@bib102] The diverse roles of LANA-1 in KSHV latency include maintenance of viral episomes, host gene manipulation through the recruitment of chromatin binding proteins, cell cycle regulation and blockade of apoptosis by downregulating p53 and Rb.[@bib70], [@bib102] vFLIP is one of the key KSHV latent proteins; it performs multiple functions such as IL-8 and IL-6 upregulation, induction of NF\u03baB, spindling of infected endothelial cells, and modulation of cell proliferation, and immune evasion.[@bib64], [@bib95], [@bib103], [@bib104], [@bib105], [@bib106] PEL consists of transformed B\u00a0cells with *in\u00a0vitro* clonogenic properties attributed to a multitude of molecules.[@bib77] A key observation by Paul et\u00a0al (2011) is the inhibition of the colony formation capacity of PEL cells by nimesulide because it encapsulates the pathologic consequence of COX-2 inhibition mediated latency blockade, G1 arrest, and apoptosis induction in PEL cells.[@bib63] Nimesulide mediated proliferation arrest, alteration in cell cycle profile, and apoptosis in PEL cells could be related to the downregulation of KSHV latency proteins LANA-1 and vFLIP resulting in the blockade of virus induced prosurvival mechanisms in PEL.[@bib107], [@bib108], [@bib109], [@bib110], [@bib111], [@bib112], [@bib113], [@bib114] However, considering the oncogenic potential of COX-2/PGE2/EP receptors in other cancer systems that are also important for PEL pathogenesis, the antigrowth effects of nimesulide could also be due to the drug's effects on these pathways as well as independent of viral proteins.[@bib5], [@bib15], [@bib115], [@bib116], [@bib117], [@bib118], [@bib119], [@bib120], [@bib121], [@bib122], [@bib123]\n\nChemotherapeutic Potential of NSAIDs in Treating PEL {#sec9}\n====================================================\n\nNSAIDs consist of COX-1/COX-2 inhibitors such as aspirin, indomethacin, and diclofenac and COX-2 specific inhibitors such as nimesulide and the COXIB (celecoxib, rofecoxib, valdecoxib, and lumiracoxib) family.[@bib124], [@bib125] COX-2 specific drugs such as COXIBs have gained popularity and notoriety in the last 2 decades because of their potent antipyretic and analgesic effects and numerous trials strongly suggesting an increase in cardiovascular events from the chronic use of rofecoxib and celecoxib, respectively.[@bib124], [@bib125] From a chemotherapeutic perspective after considering the severe side effects of existing anti-PEL drug regimens, which provide no specific cure for PEL, the goal should be to identify a drug with potent anti-KSHV and anti-cancer activity with the least side effects. Several lines of work are currently underway to develop anti-PEL therapies based on PEL pathogenesis such as the proapoptotic agents bortezomib and azidothymidine, antiproliferative antibiotic rapamycin, p53 activator nutlin-3a, antiviral compounds cidofovir and IFN-\u03b1, reactive oxygen species hydrogen peroxide, activation of unfolded protein response, and KSHV latency gene blocking agents glycyrrhizic acid, and small RNA transcripts.[@bib77], [@bib113], [@bib114], [@bib126], [@bib127], [@bib128], [@bib129], [@bib130], [@bib131], [@bib132], [@bib133], [@bib134], [@bib135], [@bib136], [@bib137], [@bib138] The well-established tumorigenic potential of COX-2/PGE2/EP receptor pathway,[@bib18], [@bib24] the availability of well-characterized EP receptor antagonists and Food and Drug Administration-approved COX-2 inhibitors with known\u00a0anticancer effects,[@bib18] the demonstration of COX-2/PGE2/EP receptors in KSHV latency[@bib59], [@bib61], [@bib62] and the correlation between COX-2 expression and poor NHL prognosis[@bib18] provided an excellent context to examine the chemotherapeutic potential of NSAIDs in treating PEL by Paul et\u00a0al (2011).[@bib63] The study by Paul et\u00a0al (2010) and work by George Paul et\u00a0al (2013)[@bib65] examined the chemotherapeutic potential of nimesulide and celecoxib against PEL and several NHL cell lines, respectively. Nimesulide is a well-characterized COX-2 inhibitor with known anticancer properties and is already prescribed to approximately 500 million people in 50 different countries since its introduction in 1985.[@bib14], [@bib121], [@bib139] Celecoxib was introduced in 1998, and several lines of work have strongly suggested the anticancer effects of both nimesulide and celecoxib.[@bib140], [@bib141] Celecoxib's anticancer effect is proposed to be due to COX-2 inhibition and non-COX dependent antigrowth effects.[@bib142] Nimesulide could induce significant proliferation arrest on a multitude of KSHV+/EBV- (BC-3, KSHV-BJAB), KSHV-/EBV+ (Akata/EBV+, LCL, Raji), KSHV+/EBV+ (JSC-1), and KSHV-/EBV- (Loukes, Ramos, Akata/EBV-) NHL cell lines with selective potency against KSHV+/EBV- cell lines suggesting that the proliferation arrest induced by nimesulide on all NHL cell lines tested is not due to the generalized antiproliferative effects of NSAIDs on tumor cell lines.[@bib63] In the work by Paul\u00a0et\u00a0al (2013),[@bib65] celecoxib had significant antiproliferative effects on KSHV+/EBV- (BCBL-1 and BC-3), KSHV-/EBV+ (Akata/EBV+), KSHV+/EBV+ (JSC-1), and KSHV-/EBV- (BJAB) cell lines.\n\nThe chemotherapeutic potential of EP receptor antagonists in any NHLs is still unexamined. The study by Paul et\u00a0al (2013)[@bib65] investigated the anticancer effects of EP receptor blockade on various NHL cell lines. The study demonstrates for the first time the anticancer effects of SC-51322 (EP1 antagonist), AH6809 (EP2 antagonist), and GW 627368X (EP4 antagonist) on NHL cell lines. EP1 and EP4 receptor antagonist had significant antiproliferative effects on KSHV+/EBV- (BCBL-1 and BC-3), KSHV-/EBV+ (Akata/EBV+), KSHV+/EBV+ (JSC-1), and KSHV-/EBV- (BJAB) cell lines. EP1 and EP4 antagonists also induced apoptosis in BCBL-1, Akata/EBV+, and Akata/EBV- cells but not in BJAB cells.\n\nAn intriguing idea proposed by the studies by Paul et\u00a0al (2010) and current work (Paul et\u00a0al 2013) is the multimodal anticancer and anti-KSHV effects of COX-2 inhibition on the PEL cell lines.[@bib63] The specificity of the 'knock-out' punch of nimesulide on PEL cells might be due to the additive antiviral (blockade of KSHV latency genes LANA-1 and vFLIP), anti-inflammatory/antisurvival properties (downregulation of VEGF-C, AKT 1/2, and GSK-3\u03b2), anti-PEL specific properties (downregulation of syndecan-1, VDR, and aquaporin-3) as well as anticancer (induction of G1 arrest and apoptosis) properties (Paul et\u00a0al 2011).[@bib63]\n\nCombinational Blockade of COX-2 and EP Receptors {#sec10}\n================================================\n\nPrevious work examining the role of COX-2 mediated tumorigenesis has proposed the combination of lower doses of COX-2 inhibitors with EP receptor antagonists to reduce the adverse effects of NSAIDs such as GI and cardiovascular toxicities.[@bib5], [@bib27] However, potential synergistic anticancer effects of combinational blockade of COX-2 and EP receptors are still unexamined. The current study by Paul et\u00a0al (2013)[@bib65] demonstrates that combining 1.0 \u03bcM each of celecoxib, SC-51322 (EP1 antagonist) and GW 627368X (EP4 antagonist) can potentiate the proapoptotic effect of celecoxib on KSHV+/EBV- and KSHV-/EBV+ cells. This strongly suggests the chemotherapeutic potential of a novel paradigm based on concurrent inhibition of COX-2 and EP receptors to obtain potent additive anticancer effects. The study shows that simultaneous inhibition of COX-2 and EP1/EP4 receptors modulates several classes of genes proposed to be important for KSHV and EBV associated lymphoma pathogenesis including tumor suppressors, and genes belonging to lymphoma survival, cell cycle arrest, cell adhesion, apoptosis, PI3K/Akt signaling, and epigenetic regulation.[@bib135], [@bib143], [@bib144], [@bib145], [@bib146], [@bib147], [@bib148], [@bib149], [@bib150], [@bib151], [@bib152], [@bib153], [@bib154], [@bib155], [@bib156] Furthermore, the study also shows the activation of several host genes proposed to be downregulated by either KSHV/EBV infection or respective viral proteins such as chemokine (C-X-C motif) receptor 4 (CXCR4), LIM domain only 2 (LMO2), v-myc myelocytomatosis viral oncogene homolog (MYC), Toll-like receptor 5 (TLR5), inhibitor of DNA binding 4 (ID4), TGF-\u03b21), antigen identified by monoclonal antibody Ki-67 (MKI67), ATM, lymphoid-restricted membrane protein (LRMP), TP53, membrane metallo-endopeptidase (MME), TIMP3, MLH1, CDH1, DLC1, CDKN1C, glutathione S-transferase pi 1 (GSTP1), HIC1, and RASSF1.[@bib157], [@bib158], [@bib159], [@bib160], [@bib161], [@bib162], [@bib163], [@bib164], [@bib165], [@bib166], [@bib167], [@bib168], [@bib169], [@bib170], [@bib171], [@bib172], [@bib173], [@bib174], [@bib175], [@bib176], [@bib177], [@bib178], [@bib179], [@bib180] Overall, the study by Paul et\u00a0al (2010)[@bib62] and Paul et\u00a0al (2013)[@bib65] propose that the anticancer effects of concurrent targeting of COX-2 and EP1/EP4 receptor is due to the simultaneous inhibition of viral and nonviral mediated tumorigenic mechanisms acting at multiple levels.\n\nConclusions and Future Studies {#sec11}\n==============================\n\nA key aspect of chronic inflammation and oncogenesis attributable to inflammation is the sustenance of the driving factors such as COX-2 activity, PGE2 secretion, and PGE2 mediated functional autocrine and paracrine signaling.[@bib5], [@bib27] An interesting finding in the study by George Paul et\u00a0al (2010)[@bib62] is the downregulation of COX-2 gene expression and PGE2 secretion by EP2 and EP4 antagonists indicating a positive feedback loop mediated through EP2 and EP4 receptor signaling that simultaneously regulates LANA-1 and COX-2 expression.[@bib62] Mechanistically, the stability of the COX-2 messenger RNA (mRNA) transcript has been shown to be mediated by p38/MK2 dependent signaling acting on the ARE sequences in the 3\u2032 UTR region of the COX-2 mRNA.[@bib181] Interestingly, the KSHV protein kaposin B is also shown to stabilize mRNA transcripts with 3\u2032UTR ARE sequences through p38/MK2 signaling.[@bib182] Further studies are critical to fully understand this pathway, such as examining the effect of EP receptor antagonism on the gene expression of Kaposin B, cytokine and p38/MK2 activation, and COX-2 protein levels. Multiple promoters (Lti, Ltc, Ltd) have been identified in the KSHV latency locus and account for the transcripts of LANA-1, vFLIP, vCyclin, viral microRNAs, and Kaposins.[@bib183], [@bib184] Therefore, the induction of the LANA-1 promoter by PGE2 and EP receptor agonists also raises the question whether the PGE2 and EP receptors could activate other latency promoters as well. Elucidating such mechanisms, if any, would provide a comprehensive perspective on how KSHV utilizes PGE2 and EP receptors for regulating latency.\n\nStudies with LANA-1 promoter deletion constructs identified a PGE2 response region in the KSHV latency locus.[@bib62] The KSHV latency locus is known to be regulated by Sp1, CTCF, and several other unidentified transcription factors (TFs).[@bib185], [@bib186] Among the TFs identified within the minimal region of the LANA-1 promoter required for PGE2 mediated LANA-1 promoter activity, there are several transcription factors that could be potentially stimulated by PGE2 and the EP receptor such as Sp1, C/EBP, c-Jun, Oct-1, and Oct-6.[@bib22], [@bib187] The functional significance of these TFs in inducing the LANA-1 promoter, their specific binding sites, and influence of PGE2 and EP receptors over these TFs remains to be determined. Key concept introduced by George Paul et\u00a0al (2010)[@bib62] is the role of the EP1 receptor in inducing Ca^2+^ signaling in the KSHV latency program. The study had identified a specific type of calcium signal induced by EP1 receptor in long-term-infected cells leading to several questions such as the effector molecules and the transcription factors activated by calcium signaling.\n\nOne of the most intriguing findings of the study by Paul et\u00a0al (2011)[@bib63] was the downregulation of syndecan-1, VDR, and AQP3 expression by nimesulide in PEL cells. Syndecan-1/CD138, VDR, and AQP3 are uniquely overexpressed in PEL cell lines unlike other NHLs.[@bib69], [@bib188] The role of transmembrane proteoglycan syndecan-1 in cell migration through Rac-1/PKC\u03b1 signaling and the significance of syndecan secretion in proteoglycan signaling are key aspects of oncogenesis.[@bib189] VDR is the natural receptor for 1\u03b125-dihydroxyvitamin D3.[@bib190] Induction of VDR is associated with chromatin remodeling and is also proposed to increase the risk of esophageal squamous, prostate, and pancreatic cancers by the activation of osteopontin and Ran-GTPase.[@bib190] AQP3 is a channel protein involved with the transportation of water and glycerol, and ATP generation.[@bib191] In lung adenocarcinoma, colorectal cancer, and squamous cell carcinoma, AQP3 has been proposed to play a role in promoting cell migration through actin depolymerization and ATP generation.[@bib191] The link between COX-2 and the expression of syndecan-1, AQP3, and VDR within the context of PEL raises several important questions such the role of proteoglycan mediated signaling, chromatin remodeling, and ATP metabolism in PEL and how COX-2 might be contributing to PEL pathogenesis through such a novel signal network.\n\nOverall, the studies reviewed here provide a glimpse of the molecular framework underlying the angiogenic stress response proinflammatory protein COX-2, its infamous lipid metabolite PGE2, and EP receptors in the establishment and maintenance of KSHV latency and, therefore, implicated COX-2 inhibitors and EP receptor antagonists as potent chemotherapeutic modalities in treating KSHV related lymphomas ([Fig 1](#fig1){ref-type=\"fig\"} ). Thus, the studies add a novel paradigm in the pathogenesis of KSHV associated diseases and raise several questions that could expand our understanding of the role of chronic persistent inflammation in KS and PEL.Fig 1Model summarizing the role of the cyclooxygenase-2 (COX-2)/Prostaglandin E2/Eicosanoid receptor pathway in Kaposi's sarcoma (KS) and primary effusion lymphoma (PEL) pathogenesis. During the early stages of Kaposi's sarcoma associated-herpes virus (KSHV) infection of target cells, KSHV binds to the cell surface receptors via its envelope glycoproteins, and by using multiple overlapping pathways, the virus enters the host cell.[@bib98], [@bib193] KSHV interactions with receptors, while binding and entering the target cell, induces a variety of overlapping cell signaling cascades (Extracellular signal-regulated kinase, Phosphatidylinositide 3-kinase, *Rho family of GTPases*, Focal adhesion kinase, Src, nuclear factor kappa-light-chain-enhancer of activated B cells, and protein kinase C) and transcription factors (c-Fos, c-Jun, c-Myc, and Signal transducer and activator of transcription 1-alpha) early during infection.[@bib59], [@bib84], [@bib85], [@bib86], [@bib87], [@bib88], [@bib89], [@bib90], [@bib92], [@bib193], [@bib194], [@bib195], [@bib196], [@bib197], [@bib198] KSHV infection via the induction of signal pathways also reprograms and modulates various host cell genes,[@bib82] and one of these molecules is the angiogenic stress response gene COX-2.[@bib61], [@bib82] KSHV infection induced COX-2 led to the secretion of its inflammatory metabolite PGE2.[@bib61] A variety of transcription factors (NF-\u03baB, NFAT, NF-IL-6/cEBP, AP-1, and CRE) can stimulate COX-2 expression. KSHV entry associated signal cascades involving FAK, Src, JNK, and p38 activate transcription factors NFAT and Cyclic adenosine monophosphate response element-binding CREB, which stimulate COX-2 gene expression and PGE2 secretion.[@bib60] PGE2 exerts its effect through the family of 7-transmembrane G-protein-coupled rhodopsin-type EP (1-4) receptors, which along with COX-2 and PGE2 were detected in human KS lesions.[@bib59], [@bib62] Besides manipulating host genes, KSHV establishes latency in the host cell as observed by increased expression of its viral latent genes latency associated nuclear antigen (LANA)-1 and vFLIP. PGE2 in the microenvironment of the infected cell functions in paracrine and autocrine fashion to augment its goal to establish and maintain the expression of viral latency protein LANA-1 through Ca^2+^, Src, PI3K, NK-\u03baB, and ERK1/2 mediated signal cascades.[@bib62] EP receptor antagonists downregulate LANA-1 expression through inhibition of Ca^2+^, p-Src, p-PI3K, p-PKC\u03b6/\u03bb, and p-NF-\u03baB while exogenous PGE2 and EP receptor agonists induced the LANA-1 promoter by activating transcription factors (yin-yang1, Specificity Protein 1, octamer transcription factor-1, octamer transcription factor-1, CCAAT-enhancer-binding proteins, and c-Jun).[@bib62] Collectively, our studies demonstrate that KSHV has pirated the proinflammatory PGE2 and its receptors for maintaining its latency in the host cell. Conversely, viral latency protein vFLIP mediated signaling sustains COX-2 expression and PGE2 secretion.[@bib64] KSHV oncogenic protein vFLIP induces COX-2/PGE2 to enhance its transforming ability (anchorage independent colony formation), metastatic potential (matrix metalloproteinase (MMP)-10), and inflammatory phenotype (inflammatory cytokines: monocyte chemotactic protein-1, RANTES, GRO-\u03b1/\u03b2, interleukin 8, and interleukin\u00a06; inflammation-related adhesion molecules: ICAM-1, VCAM-1; and chemokines: CXCL-6 and CXCL-5), and to promote anoikis resistance and prolong infected cell survival (cell survival genes: Cellular inhibitor of apoptosis protein-1, Cellular inhibitor of apoptosis protein-2, X-linked inhibitor of apoptosis protein, Superoxide dismutase 2, B-cell lymphoma 2, immediate early response gene X-1; antiapoptotic proteins: B-cell lymphoma 2, myeloid leukemia cell differentiation protein, B-cell lymphoma-extra large, Bcl-2 interacting mediator of cell death, and BAX translocation to the cytoplasm; and cell survival kinases; NF-\u03baB, PI3\u2009K, and AKT).[@bib64] In addition KSHV- induced COX-2/PGE2 regulated multiple events involved in KS pathogenesis such as secretion of proinflammatory cytokines and growth factors (Interleukin-1 alpha, Interleukin-1 beta, Subunit beta of interleukin 12/cytotoxic lymphocyte maturation factor 2, Tumor necrosis factor alpha, Interferon gamma-induced protein 10, neutrophil-activating protein-2, Oncostatin M, thrombopoeitin, fibroblast growth factors, Flt3-ligand, Fractalkine, Insulin-like growth factor-binding protein and Osteoprotegerin), angiogenic factors (vascular endothelial growth factor \\[VEGF\\]-A/-C), and invasive factors (MMP-2/-9).[@bib59] COX-2 blockade reduced latently infected endothelial cell adhesion/invasion, survival and proliferation (shortened S phase, arrested infected cells at G1/S phase).[@bib59] Similar to COX-2/PGE2 downstream effects in KS pathogenesis, we established that COX-2 contributes to PEL pathogenesis via viral gene independent and dependent pathways. COX-2 blockade reduced KSHV latent (LANA-1 and vFLIP) gene expression, disrupted p53-LANA-1 protein complexes, and activated the p53/p21 tumor-suppressor pathway in PEL cells.[@bib63] COX-2/PGE2 contributed to prosurvival mechanisms in PEL cells via regulating cell survival (p-Akt and p-GSK-3\u03b2), cell cycle and apoptosis blockade (cyclins E/A and cdc25C), angiogenesis (VEGF-C), transforming potential (colony forming capacity of PEL cells), and modulation of PEL defining genes (syndecan-1, aquaporin-3, and vitamin-D3 receptor).[@bib63] Collectively, these observations provide a comprehensive molecular framework linking COX-2/PGE2 with KS and PEL pathogenesis and identify the chemotherapeutic potential of targeting COX-2-PGE2-EP axis in treating KS and PEL.\n\nCurrently, NHLs are the fifth most common cancer in the United States and account for 5% of all cancers with an annual incidence increasing by 1%-2%.[@bib126], [@bib192] Keeping in mind the ultimate aim of cancer treatment is to inhibit the growth of precancerous and cancerous cells without affecting the normal cells, could the studies reviewed here suggest the antiproliferative effects of COX-2 inhibitors and EP receptor antagonists against various NHL cell lines? The data emanating from our *in\u00a0vitro* studies is valuable, informative, and requires further examination (ongoing studies) using an *in\u00a0vitro* angiogenic model and an *in\u00a0vivo* nude mice model to further validate COX-2, PGE2 inhibitors, and EP receptor antagonists as novel therapeutics to target latent KSHV infection, viral pathogenesis, and associated diseases; KS and PEL.\n\n**Arun George Paul***received his MD/PhD from the Department of Microbiology and Immunology at Rosalind Franklin University of Medicine and Science. The article is based on a presentation given at the Combined Annual Meeting of the Central Society for Clinical Research and Midwestern Section American Federation for Medical Research held in Chicago, Ill, on April 2011*.\n\nThe authors thank Keith Philibert for critically reading the review.\n\nThis study was supported in part by Public Health Service grant NIHCA128560 and RFUMS startup fund to NSW, NIHCA 075911, CA 168472, AI 091761 and RFUMS H. M. Bligh Cancer Research Fund to BC, Sigma Xi Grant-in-aid grantG2009101428 to AGP. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nBreast cancer has become a global health problem and caused more and more death in females in recent years. Invasive ductal carcinoma (IDC) is the most common breast cancer of women^[@CR1]^. Axillary lymph node is the most common site of metastasis for breast cancer and it is an extremely important factor when making treatment decision^[@CR2]^. Therefore, how to detect axillary lymph node metastasis (LNM) of breast cancer at an early stage is becoming a crucial issue. Traditionally, patients are subject to axillary lymph node dissection (ALND) to evaluate LNM. However, most axillary lymph nodes are normal and the value of ALND operation is still uncertain. In the past few years, several studies have reported that survival is not affected by clearance of regional lymph nodes^[@CR3],\\ [@CR4]^. In many centers, sentinel node biopsy (SLNB) as a minimally invasive modality has largely replaced ALND for evaluation of the axilla^[@CR5]--[@CR7]^. However, SLNB is dependent on the operator' s experience and may have false negative results^[@CR2]--[@CR4]^. In cases of several drawbacks and ineligibility of surgical method, a non-surgical method to predict axillary LNM is needed for drawing up a plan of systemic therapy, thus clinical management of patients could be improved.\n\nConventional high frequency ultrasound (US) is commonly used to detect breast lesion. Various studies have showed that conventional US may be helpful for predicting axillary LNM by evaluating the suspicious US features such as intralesional calcification, internal flow, and nodal size^[@CR8],\\ [@CR9]^. On the other hand, many studies have showed that breast cancers are stiffer than normal and benign tissues^[@CR10]--[@CR12]^, which can be evaluated by US elastography. As new US-based elastography techniques, acoustic radiation force impulse (ARFI) imaging of virtual touch tissue imaging (VTI; Siemens Medical Solutions, Mountain View, CA, USA) and shear wave imaging of virtual touch tissue imaging & quantification (VTIQ; Siemens Medical Solutions, Mountain View, CA, USA) have recently been used to diagnose breast cancers by measurement of the tissue stiffness^[@CR10],\\ [@CR11]^. For both VTI and VTIQ, a short-duration acoustic pulse from the transducer is used to generate vertical pressure and lateral local displacement. VTI is a qualitative method whereas VTIQ is a quantitative method. The latter reflects the tissue stiffness through the measurement of shear wave speed (SWS)^[@CR13]^. The maximum depth of generating sufficient shear waves (SWs) is approximately 4.5\u2009cm because of the attenuation of ARFI^[@CR14]^.\n\nAlthough shear wave imaging such as Supersonic imagine (SSI, Aix en Provence, France) has been used to evaluate axillary LNM in patients with breast IDCs^[@CR15]^, until present no study has been carried out using VTI and VTIQ for predicting axillary LNM. We hypothesized that VTI and VTIQ might be useful for predicting axillary LNM. What is more, histological grade was also used as a prognostic factor to assess patient survival and disease recurrence^[@CR16]^. On the other hand, estrogen receptor (ER), progesterone receptor (PR), and C-erbB-2 (i.e. human epidermal growth factor receptor 2 \\[HER 2\\]) expression are the most important prognostic factors for management of breast cancer, particularly in metastatic disease settings and adjuvant treatment. Thus, it was considered that the effect of the prognostic parameters including histological grade, ER, PR, and C-erbB-2 expression on axillary LNM of IDC patient should be taken into account. The purpose of the present study was to identify the risk factors including conventional US, VTI, VTIQ, and the histologic parameters and then to propose a new risk model for predicting axillary LNM in patients with breast IDCs.\n\nMaterials and Methods {#Sec2}\n=====================\n\nThe retrospective study was approved by the Ethical Committee of the Shanghai Tenth People's Hospital of Tongji University School of Medicine and informed consent from the patients were waived. The study was performed in accordance with relevant guidelines and regulations.\n\nPatients {#Sec3}\n--------\n\nFrom June 2014 to September 2015, a total of 174 consecutive women were enrolled in the study. All the patients were referred to examination by conventional US, VTI, and VTIQ for the evaluation of breast lesions that were detected incidentally during a prior imaging study or had been discovered with clinical palpation. The enrollment criteria for the patients were as follows: (1) breast IDC confirmed by histopathology after surgery; (2) solid or almost solid (\\<25% cystic) breast lesion with size larger than 5\u2009mm in diameter and depth\u2009\u2264\u20094.5\u2009cm on conventional US; and (3) patient without any previous treatment. Twelve patients with incomplete preoperative US, VTI and VTIQ data were excluded. For patients with multiple lesions (nine patients had two lesions and seven patients had three or even more lesions), only the largest one in diameter on US was included. Eventually, 162 patients (mean age, 57.2 years\u2009\u00b1\u200911.4; age range, 26--87 years) with 162 lesions comprised the final study cohort. Among them, 37 of 162 IDCs had been reported in a previous study with an aim to evaluate the value of VTIQ for US Breast Imaging Reporting and Data System (BI-RADS) category 4 lesions^[@CR13]^.\n\nConventional US, VTI, and VTIQ {#Sec4}\n------------------------------\n\nConventional US, VTI, and VTIQ were performed with the same Siemens S3000 US scanner (Siemens Medical Solutions, Mountain View, CA, USA). Conventional US, VTI, and VTIQ were performed by one of three skillful radiologists who had 2 years' experience in breast US imaging and 2 years' experience in VTI and VTIQ imaging. Each lesion was scanned with the patient holding their breath for a few seconds in the supine position. For conventional US, a 18L6 linear array transducer (frequency range, 6--18\u2009MHz) was used based on the American Institute of Ultrasound in Medicine practice guideline^[@CR17]^.\n\nVTI and VTIQ were performed following conventional US using the 9L4 linear transducer (frequency range, 4.0--9.0\u2009MHz). The scanning was performed in the longitudinal plane of the lesion and pre-compression were avoided. The VTI imaging reflects the tissue elasticity with two grayscale values (i.e. black and white) in the field of view (FOV), which includes the lesion and some surrounding breast tissues. The gray scale value of black indicates hard tissue and the white indicates soft tissue. VTIQ measurement was then performed to measure SWS within the breast lesion. The measurement result of SWS is expressed as \"m/s\" and ranges from 0.5 to 10\u2009m/s^[@CR18]^. VTIQ imaging involves the target lesion and sufficient surrounding tissue as well. The SW-quality map was firstly applied, which displayed quality in different colors from high (green), moderate (yellow), to low (red) quality^[@CR19]^. SWS measurement was then performed on SW-velocity map by placing the region of interest (ROI) on the areas with high SW quality in the breast lesion. In SW-velocity map, the speed distribution is shown in different colors from high (red), intermediate (yellow or green), to low (blue). The scale of the SW-velocity map is set at default value of 10\u2009m/s and is not adjusted thereafter. Seven SWS measurements were performed, and the cystic or calcified areas were avoided. The smallest size of ROI box is around 2\u2009\u00d7\u20092\u2009mm. The ROIs were placed at random when the distribution of lesion stiffness was homogeneous. Otherwise two ROIs were placed on the highest stiffness area and the lowest stiffness area respectively, and the remaining five ROIs were placed at random. The mean SWS values were computed. All dates from conventional US, VTI, and VTIQ were recorded for further analysis.\n\nImage Analysis {#Sec5}\n--------------\n\nTwo independent investigators were asked to analyze all the US, VTI, and VTIQ images in a blind manner. Both investigators had more than 3 years of experience in breast US. Disagreement was solved by consensus. Each investigator had been trained to review the images before the study. The conventional US features as following were evaluated: lesion size (\u2265\u200920\u2009mm/\\<\u200920\u2009mm), shape (regular/irregular), margin (circumscribed/not circumscribed), inhomogeneous echo, posterior features (absent/present), microcalcifications, taller than wide or not, and internal flow on color Doppler US (poor/rich). The calcification was judged to be microcalcification if the calcification diameter \u22640.5\u2009mm^[@CR20]^. In addition, the vascularity of the lesions were determined as described by Xu *et al*.^[@CR21]^. The internal flow was judged to be \"rich\" if more than 3 linear or tree-like signals existed on color Doppler US, otherwise the flow was judged to be \"poor\" (i.e. dotted signals or without signals on color Doppler US). The VTI of the breast lesions was scored from 1 (soft) to 6 (hard) as described by Li *et al*.^[@CR13]^: score of 1, predominantly white; score of 2, predominantly white with few black portions; score of 3, black and white portion equally; score of 4, predominantly black with a few white spots; score of 5, almost completely black; and score of 6, completely black.\n\nHistopathological Analysis {#Sec6}\n--------------------------\n\nPathological reports after surgeries were reviewed to determine histological grade of the tumor, presence of axillary LNM or not, and the number of the metastatic lymph node status. The histologic grade for each lesion was recorded as grade I, grade II, or grade III (Elston and Ellis method)^[@CR16]^. We regarded the grade I, grade II as \"low\" and grade III as \"high\". The routine paraffin sections was used as the reference standard and the immunohistochemical staining were then used to get the results such as ER, PR, and C-erbB-2 expression. Briefly, four micrometer paraffin sections were cut and air dried overnight at room temperature (temperature: \\<25\u2009\u00b0C). After deparaffinization with xylene and hydration in decreasing grades of ethanol, the activity of endogenous peroxidase was eliminated by treating the slides with a solution of hydrogen peroxide and methanol. The pressure cooker method in 0.1\u2009mol/l sodium citrate buffer (pH 6) was used to retrieve heat-induced epitope. Then the slides were treated with the diluted primary antibody, biotinylated antirabbit IgG immunoglobulin, and streptavidinperoxidase complex. Diaminobenzidine was used as a chromogen in the presence of hydrogen peroxide (View Universal DAB Detection Kit, Ventana\u00ae, Roche, Switzerland). Slides were sequentially counterstained with Mayer's haematoxylin, dehydrated in alcohol, cleared in xylene and mounted. The ER and PR were judged to be positive if the expression was \u226510%. The C-erbB-2 expression was scored as negative, 1+, 2+, and 3+. We judged the C-erbB-2 expression to be positive if the expression was 2+ or 3+^[@CR22]^, otherwise C-erbB-2 expression was judged to be negative.\n\nStatistical Analysis {#Sec7}\n--------------------\n\nStatistical analyses were performed with SPSS software (version 20.0 for Windows; SPSS, Chicago, IL, USA). Quantitative datas (i.e. patient age, lesion size, and SWS on VTIQ) were expressed as mean\u2009\u00b1\u2009standard deviation (SD) if normal distribution was achieved. Continuous variables were compared by independent t test. Categorical variables were analyzed by \u03c7^2^ test or Fisher's exact probability test. The best cut-off value for VTI was obtained to differentiate between patients with axillary LNM and no axillary LNM. The correlation between the predicting factors and axillary LNM was analyzed by univariate analysis. Multivariate logistic regression analysis was used to determine the statistically significant factors. Odd ratios (ORs) with 95% confidence interval (CI) were recorded and the equation for predicting axillary LNM was created. Receiver operating characteristic (ROC) curves were plotted to assess the diagnostic performance of the statistically significant factors and the predictive equation. The diagnostic performances for the statistically significant factors and the predictive equation were expressed as the area under the ROC curve (AUROC). The diagnostic performance was shown in different AUROC from low (AUROC\u2009=\u20090.5--0.7), moderate (AUROC\u2009=\u20090.7--0.9), to high (AUROC\u2009\\>\u20090.9). ROC analysis was used to assess the sensitivity and specificity. The best cut-off values for the predictive equation and VTI were obtained when Youden indexs (YIs) were maximum (sensitivity\u2009+\u2009specificity\u2009\u2212\u20091). *P*\u2009\\<\u20090.05 was considered as statistical significance.\n\nResults {#Sec8}\n=======\n\nOf the 162 breast lesions, the final pathological diagnoses were all IDCs. Among them, 20 (12.3%) were histologic grade I, 84 (51.9%) were histologic grade II, and 58 (35.8%) were histologic grade III. Axillary LNMs were present in 64 (39.5%) patients whereas absent in 98 (60.5%) patients (Figs\u00a0[1](#Fig1){ref-type=\"fig\"} and [2](#Fig2){ref-type=\"fig\"}). Number of metastatic lymph nodes were three or even more in 24 (37.5%) patients and one or two in 40 (62.5%) patients.Figure 1Images in a 60-year-old patient with breast invasive ductal carcinoma, no axillary lymph node metastasis (LNM), histologic grade II, positive estrogen receptor (ER), negative progesterone receptor (PR), and negative C-erbB-2. (**a**) A solid, heterogeneous hypoechogenicity, and poorly defined margin lesion (arrows) is shown on ultrasound (US). (**b**) Poor internal flow (i.e. dotted signals on color Doppler US) is found on color Doppler flow image (arrows). (**c**) Virtual touch tissue imaging (VTI) score of the lesion (arrows) is 3. (**d**) On virtual touch tissue imaging & quantification image, the lesion (arrows) is heterogeneous with a mean SWS value of 8.24\u2009m/s. (**e**) Pathological examination confirms the diagnosis of invasive ductal carcinoma (Hematoxylin-eosin stain, x200). (**f**) Pathological examination confirms the diagnosis of normal lymph node (Hematoxylin-eosin stain, x200). Figure 2Images in a 58-year-old patient with breast invasive ductal carcinoma, axillary lymph node metastasis (LNM), histologic grade III, negative estrogen receptor (ER), negative progesterone receptor (PR), and positive C-erbB-2. (**a**) A solid, marked hypoechogenicity, well defined margin, irregular, and taller than wide shape lesion (arrows) is shown on US. (**b**) Rich internal flow (i.e. 3 linear or tree-like signals) is found on color Doppler flow image (arrows) of the breast invasive ductal carcinoma. (**c**) Virtual touch tissue imaging (VTI) score of the lesion (arrows) is 4. (**d**) On virtual touch tissue imaging & quantification image, the lesion (arrows) is heterogeneous with a mean SWS value of 4.75\u2009m/s. (**e**) Pathological examination confirms the diagnosis of invasive ductal carcinoma (Hematoxylin-eosin stain, \u00d7200). (**f**) Pathological examination confirms the diagnosis of axillary LNM (Hematoxylin-eosin stain, \u00d7200).\n\nThe mean size of the breast lesions was 26.6\u2009mm\u2009\u00b1\u200915.4 (range, 6.0--86.0\u2009mm). The breast lesions with metastatic lymph nodes tended to be larger than those without metastatic lymph nodes (31.2\u2009\u00b1\u200916.0\u2009mm; range, 9.0--84.0\u2009mm vs. 23.4\u2009\u00b1\u200915.3\u2009mm; range, 6.0--86.0\u2009mm) (*P*\u2009=\u20090.002). The basic characteristics and US features of the patients and lesions are listed in Table\u00a0[1](#Tab1){ref-type=\"table\"}.Table 1Basic characteristics and ultrasound features of the patients.CharacteristicsAxillary LNM*P* valueYesNoMean age (y, range)56.16\u2009\u00b1\u20099.43 (34--81)57.87\u2009\u00b1\u200912.52 (26--87)0.352Lesion size (\u2009mm, range)31.16\u2009\u00b1\u200916.02 (9--84)23.43\u2009\u00b1\u200915.30 (6--86)0.002\\*Lesion position0.599\u00a0Right/left, n (%)28 (43.8)/36 (56.2)47 (48.0)/51 (52.0)Lesion shape0.368\u00a0Regular/irregular, n (%)5 (7.8)/59 (92.2)12 (12.2)/86 (87.8)Margin0.616\u00a0Circumscribed/not circumscribed, n (%)22 (34.4)/42 (65.6)30 (30.6)/68 (69.4)Homogeneous echo0.536\u00a0Yes/no, n (%)13 (20.3)/51 (79.7)24 (24.5)/74 (75.5)Posterior acoustic features0.273\u00a0Absent/present, n (%)37 (57.8)/27 (42.2)65 (66.3)/33 (33.7)Microcalcifications in a lesion0.241\u00a0Absent/present, n (%)38 (38.8)/26 (26.5)67 (68.4)/31 (31.6)Taller than wide shape\\<0.001\\*\u00a0No/yes, n (%)14 (21.9)/50 (78.1)65 (66.3)/33 (33.7)Internal flow\\<0.001\\*\u00a0Poor/rich, n (%)22 (34.4)/42 (65.6)73 (74.5)/25 (25.5)LNM lymph node metastasis. \\*Indicates statistically significant difference.\n\nComparisons between patients with axillary LNM and without axillary LNM in terms of VTI, VTIQ, and histologic parameters are listed in Table\u00a0[2](#Tab2){ref-type=\"table\"}. For VTI, tissue stiffness of the breast lesions (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}) was scored from 2 to 6 and the best cut-off value was VTI score \u22655 (YI\u2009=\u20090.470). No axillary LNM was found for those patients with VTI score of 1 and 2 lesions. were found Axillary LNM was found more frequently in patients with VTI scores of 5 and 6 lesions than in those with VTI scores of 3 and 4 lesions (*P*\u2009\\<\u20090.001). The sensitivity and specificity for VTI were 76.6% (49/64) and 70.4% (69/98) respectively.Table 2Comparisons with VTI, VTIQ, and histologic parameters between breast lesions with axillary LNM and those without axillary LNM.ParametersOverall lesions (n\u2009=\u2009162)Axillary lymph node metastasis*P* valueYes (n\u2009=\u200964)No (n\u2009=\u200998)VTI\\<0.001\\*\u00a0Score 10 (0.0%)0 (0.0%)0 (0.0%)\u00a0Score 23 (1.9%)0 (0.0%)3 (3.1%)\u00a0Score 315 (9.3%)3 (4.7%)12 (12.2%)\u00a0Score 466 (40.7%)12 (18.8%)54 (55.1%)\u00a0Score 572 (44.4%)45 (70.3%)27 (27.6%)\u00a0Score 66 (3.7%)4 (6.3%)2 (2.0%)VTIQ0.651\u00a0SWS (m/s)5.60\u2009\u00b1\u20091.625.53\u2009\u00b1\u20091.525.64\u2009\u00b1\u20091.68\u00a0Range (m/s)3.01--9.683.01--9.223.12--9.68Histologic grade\\<0.001\\*\u00a0I20191\u00a0II846915\u00a0III581048ER status0.716\u00a0Negative482028\u00a0Positive1144470PR status0.306\u00a0Negative993663\u00a0Positive632835C-erbB-2 expression\\<0.001\\*\u00a0Negative791663\u00a0Positive834835VTI virtual touch tissue imaging, VTIQ virtual touch tissue imaging & quantification, LNM lymph node metastasis, SWS shear wave speed, ER estrogen receptor, PR progesterone receptor. \\*Indicates statistically significant difference Figure 3Virtual Touch tissue imaging elasticity scores of the lesions (arrows): (**a**) score 2; (**b**) score 3; (**c**) score 4; (**d**) score 5; and (**e**) score 6.\n\nIn addition, the immunohistochemical results such as ER, PR, and C-erbB-2 expression (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}) were obtained. Metastatic lymph nodes were found in 38.6% (44/114) patients with positive ER (expression \u226510%) and 41.7% (20/48) patients with negative ER (expression \\<10%). On the other hand, metastatic lymph nodes were found in 44.4% (28/63) patients with positive PR (expression \u226510%) and 36.4% (36/99) patients with negative PR (expression \\<10%). However, there was no statistical differences between patients with positive and negative ER in terms of axillary LNM, as well as those with positive and negative PR (both *P*\u2009\\>\u20090.05). For C-erbB-2 expression, metastatic lymph nodes was found in 57.8% (48/83) patients with positive expression (2+ or 3+) and 20.3% (16/79) patients with negative expression (negative or 1+). Axillary LNMs were found more frequently in the patients with positive expression of C-erbB-2 (*P*\u2009\\<\u20090.001).Figure 4Representative immunohistochemical staining of estrogen receptor (ER), progesterone receptor (PR), and C-erbB-2. (**a**) ER status (negative); (**b**) ER status (positive); (**c**) PR status (negative); (**d**) PR status (positive); (**e**) C-erbB-2 expression (negative); and (**f**) C-erbB-2 expression (positive).\n\nUnivariate analysis showed that lesion size \u226520\u2009mm, taller than wide shape, rich internal flow on color Doppler US, VTI score \u22655, histologic grade III, and positive C-erbB-2 were significantly associated with axillary LNM (all *P*\u2009\\<\u20090.05) (Table\u00a0[3](#Tab3){ref-type=\"table\"}). None of the factors such as age, margin, shape, microcalcification, homogeneous echo, posterior features, SWS value on VTIQ, ER, and PR were found to be statistically significant in predicting axillary LNM in patients with breast IDCs (all *P*\u2009\\>\u20090.05).Table 3Univariate analysis in predicting axillary lymph node metastasis.ParametersOdds Ratios95% CI*P*Lesion size \u226520\u2009mm0.3342.321--9.115\\<0.001\\*Taller than wide shape0.2153.404--14.536\\<0.001\\*Rich internal flow5.5752.804--11.084\\<0.001\\*VTI score \u226557.7723.772--16.015\\<0.001\\*Histologic grade III0.18211.118--62.698\\<0.001\\*Positive C-erbB-25.4002.680--10.881\\<0.001\\*CI confidence interval. \\*Indicates statistically significant difference.\n\nOn multivariate logistic regression analysis, all independent predictors were determined as following: lesion size \u226520\u2009mm (OR: 3.526, *P*\u2009=\u20090.033), taller than wide shape (OR: 14.102, *P*\u2009\\<\u20090.001), VTI score \u22655 (OR: 9.003, *P*\u2009\\<\u20090.001), histological grade III (OR: 49.127, *P*\u2009\\<\u20090.001), and positive C-erbB-2 (OR: 6.869, *P*\u2009=\u20090.002) (Table\u00a0[4](#Tab4){ref-type=\"table\"}). Among them, histological grade III was the most significant risk factor.Table 4Multivariate analysis in predicting axillary lymph node metastasis.ParametersBSEOdds Ratios95% CI*P*Lesion size \u226520\u2009mm1.2600.5923.5261.104--11.2590.033\\*Taller than wide shape2.6460.71514.1023.471--57.292\\<0.001\\*VTI score \u226552.1980.6209.0032.671--30.345\\<0.001\\*Histologic grade III3.8940.75949.12711.100--217.418\\<0.001\\*Positive C-erbB-21.9270.6086.8692.088--22.6000.002\\*VTI virtual touch tissue imaging, SE standard error, CI confidence interval. \\*Indicates statistically significant difference.\n\nA multivariate logistic regression equation was then established with the significant predictors as follows: P\u2009=\u20091/1\u2009+\u2009Exp\u2211\\[\u22126.317\u2009+\u20091.260\u2009\u00d7\u2009(if lesion size \u226520\u2009mm)\u2009+\u20092.646\u2009\u00d7\u2009(if taller than wide shape)\u2009+\u20092.198\u2009\u00d7\u2009(if VTI \u22655)\u2009+\u20093.894\u2009\u00d7\u2009(if histological grade III)\u2009+\u20091.927\u2009\u00d7\u2009(if positive C-erbB-2)\\]. Then the ROC curves were plotted to assess the diagnostic performances of the predictive equation, lesion size \u226520\u2009mm, taller than wide shape, VTI score \u22655, histological grade III, and positive C-erbB-2. The diagnostic values of the predictive equation, lesion size \u226520\u2009mm, taller than wide shape, VTI score \u22655, histological grade III, and positive C-erbB-2 were 0.958, 0.681, 0.722, 0.735, 0.824, and 0.696, respectively (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). In terms of AUROC, the predictive equation achieved the highest diagnostic performance. The best cut-off value for the predictive equation was \u22120.22 (YI\u2009=\u20090.835). The sensitivity and specificity were 90.6% and 92.9% respectively.Figure 5Receiver operating characteristic (ROC) curves of (**a**) the lesion size \u226520\u2009mm (AUROC\u2009=\u20090.681), (**b**) taller than wide shape (AUROC\u2009=\u20090.722), (**c**) VTI score \u22655 (AUROC\u2009=\u20090.735), (**d**) histological grade III (AUROC\u2009=\u20090.824), (**e**) C-erbB-2 positive (AUROC\u2009=\u20090.696), and (**f**) the predictive equation (AUROC\u2009=\u20090.958) for prediction of axillary lymph node metastasis (LNM).\n\nA final predicting model was established based on the 5 risk factors derived from the multivariate logistic regression analysis. Risk score (RS) for each lesion was defined as follows: RS\u2009=\u20091.3\u2009\u00d7\u2009(if lesion size \u226520\u2009mm)\u2009+\u20092.6\u2009\u00d7\u2009(if taller than wide shape)\u2009+\u20092.2\u2009\u00d7\u2009(if VTI score \u22655)\u2009+\u20093.9\u2009\u00d7\u2009(if histological grade III)\u2009+\u20091.9\u2009\u00d7\u2009(if positive C-erbB-2). The rating system was divided into 6 stages as following: Stage I, RS was \\<1.3 and none of 5 risk factors was enrolled, including 19 (11.7%) patients; Stage II, RS was 1.4 to 3.9 and any 1 of 5 risk factors was enrolled, including 33 (20.4%) patients; Stage III, RS was 4.0 to 6.5 and any 2 of 5 risk factors were enrolled, including 39 (24.1%) patients; Stage IV, RS was 6.6 to 8.7 and any 3 of 5 risk factors were enrolled, included 29 (17.9%) patients; Stage V, RS was 8.8 to 10.6 and any 4 of 5 risk factors were enrolled, including 26 (16.0%) patients; and Stage VI, RS was 10.7 to 11.9 and all of 5 risk factors were enrolled, including 16 (9.9%) patients. The risk rates of axillary LNM were 0% (0/19) in Stage I, 6.1% (2/33) in Stage II, 7.7% (3/39) in Stage III, 65.5% (19/29) in Stage IV, 92.3% (24/26) in Stage V, and 100% (16/16) in Stage VI. According to the results of the rating system, as the number of risk factors increased, the probability of axillary LNM increased. With these findings, we regarded Stage I (none risk factors) as no axillary LNM probably, Stage II and Stage III (one or two risk factors) as low suspicion, Stage IV (three risk factors) as mediate suspicion, and Stage V and Stage VI (four or five factors) as highly suggestive of axillary LNM.\n\nDiscussion {#Sec9}\n==========\n\nTraditionally, ALND can remove the lymph node and reduce the concern for LNM. However, ALND had some drawbacks such as lymphedema, pain, or shoulder weakness^[@CR4],\\ [@CR23]^. Therefore, SLNB has become a safe method to replace ALND. Susan *et al*. found that the sensitivity of US-guided FNA for predicting positive results was 71--75%^[@CR24]^. However, the obvious disadvantage was that the sensitivity increased with increasing size of the primary lesion. Several studies showed that the false negative results would never reach zero^[@CR2],\\ [@CR4],\\ [@CR25]^. What is more, it is generally known that conventional US is wildly used to find abnormal lymph node by examination of axilla. However, not all metastatic lymph nodes could be seen by conventional US. Several studies have reported that metastatic lymph nodes were seen in only 35--65% of the patients by conventional US^[@CR2],\\ [@CR4],\\ [@CR26]^. This is probably a consequence of the latency metastasis, which may not cause the abnormity of lymph node on conventional US^[@CR27]^. As shown in the present study, the predicting model based on US, US elastography, and histologic parameters, as a noninvasive method, achieved the high diagnostic performance (AUROC\u2009=\u20090.958). Therefore, it may be a useful method to predict axillary LNM.\n\nIn the present study, seventeen clinical characteristics, US features, VTI features, VTIQ features, and histologic parameters were included as potential predictors for axillary LNM. In the past few years, several studies showed that some clinical characteristics and conventional US features were associated with axillary LNM^[@CR28],\\ [@CR29]^. In the present study, age, shape, margin, homogeneous echo, posterior acoustic features, microcalcifications, SWS, ER, and PR were not identified to be predictors for axillary LNM. In univariate analyses, lesion size \u226520\u2009mm was associated with axillary LNM, which was in consistent with previous studies^[@CR27],\\ [@CR30]^. In addition, rich internal flow (OR: 5.575, *P*\u2009\\<\u20090.001) was significantly associated with axillary LNM in univariate analysis. A study by Mehta *et al*. also showed that flow of breast cancer correlated strongly with detection of lymph node involvement, with an associated sensitivity of 93%^[@CR29]^. As we all know, the lymphatic drainage of breast is rich. The tumor cells may spread through the drainages. However, in multivariate logistic analysis, rich internal flow was not a independent risk factor after adjustment for other factors. What is more, the result of our study indicated that taller than wide shape (OR: 14.102, *P*\u2009\\<\u20090.001) was also associated with axillary LNM. Although Stavors *et al*. reported that the taller than wide shape was a critical factor to distinguish between benign and malignant breast lesions^[@CR31]^, it has not been reported before as an independent predictive factor of axillary LNM. The morphological change of IDCs reflects the enhancement of the tumor invasion ability to some extent, which may lead to cell growth, motility, and differentiation.\n\nRecently, tissue stiffness reflected by conventional US strain elastography (SE), strain ratio on elastography, or virtual touch tissue quantification (VTQ) imaging has been found to be associated with LNM^[@CR32]--[@CR35]^. However, these techniques have several limitations. SE and strain ratio on elastography have poor reproducibility and lack of quantitative information. For VTQ, the ROI for interrogation is fixed (i.e. 6\u2009\u00d7\u20095\u2009mm) and the measurement range (i.e. 0.5--8.4\u2009m/s) is limited.\n\nEvans *et al*. reported that metastatic lymph nodal involvement rates increased from 7% to 41% with the tumor stiffness increased from mean stiffness \\<50\u2009KPa to \\>150\u2009KPa by measuring tumor stiffness quantitatively^[@CR15]^. VTI and VTIQ, as the latest development of elastography techniques, have been used to differentiate between benign and malignant breast lesions^[@CR13],\\ [@CR19]^, whereas their role in predicting axillary LNM is unknown. According to our results, VTI score \u22655 had an approximately nine times (OR: 9.003) higher potential of axillary LNM than lesions with VTI score \\<5. Several *in vitro* studies reported that pathogenesis of tumor invasion and metastasis increased stiffness of matrix, which was an important microenvironment cue to regulate cell growth, motility, and differentiation^[@CR36]--[@CR38]^. On the other hand, VTIQ was not an independent risk factor for predicting axillary LNM, which was in contrast to the study by Evans *et al*.^[@CR15]^. The reason may be that different sampling methods were used. In the present study, mean SWS value of the 7 measurements on each lesion was computed, while the ROI covered the entire lesion in the study by Evans *et al*. with Supersonic imagine (SSI, Aix en Provence, France). Maybe the latter measuring method reflects the whole tissue stiffness and is a better choice.\n\nOur result demonstrated that ER and PR were not associated with axillary LNM, which was similar with the study by Evans *et al*.^[@CR15]^. ER and PR might play an important role in determining whether a patient needed adjuvant endocrine therapy^[@CR39]^. Histologic grade III had an approximately forty nine times (OR: 49.127) higher potential of axillary LNM than lesions with grade I or grade II. The enhanced tumor invasion ability may be the reason. What is more, in the present study, positive C-erbB-2 was an independent risk factor for predicting axillary LNM, which has been confirmed in several studies^[@CR40],\\ [@CR41]^. The phenomenon could be correlated with an enhanced tumor aggressiveness.\n\nA multivariate logistic regression equation was established in the present study. The result of the study suggested that the presence of axillary LNM was depending on the features such as lesion size \u226520\u2009mm, taller than wide shape, VTI score \u22655, histological grade III, and positive C-erbB-2. The reliability of the final predicting model was confirmed by AUROC of 0.958, which was higher than that of lesion size \u226520\u2009mm (AUROC\u2009=\u20090.681), taller than wide shape (AUROC\u2009=\u20090.722), VTI score \u22655 (AUROC\u2009=\u20090.735), histological grade III (AUROC\u2009=\u20090.824), and positive C-erbB-2 (AUROC\u2009=\u20090.696) individually. A risk model with six stages (i.e. Stage I, Stage II, Stage III, Stage IV, Stage V, and Stage VI) was established according to the five independent risk factors, and the corresponding risks of axillary LNM were 0%, 6.1%, 7.7%, 65.5%, 92.9%, and 100%, respectively. The results indicated that none of five predictive factors in patients at Stage I was not expected to undergo ALND. For Stage II and Stage III patients, axillary LNM was probably low, and combining other risk factors such as rich internal flow, poorly defined margin reported by previous studies might be helpful to determine the strategy of ALND^[@CR28],\\ [@CR29]^. For Stage IV, Stage V, and Stage VI patients, we would advocate the patients to determine the strategy of ALND. Thus, based on the model, ALND as a aspiring strategy was suggested in patients at Stage IV, Stage V, and Stage VI.\n\nThe current study had several limitations. First, this was a retrospective single-centre study, which might cause selection bias. Second, breast lesions with size \\<5\u2009mm in diameter were not included in the present study because the smallest size of VTIQ ROI box is around 2\u2009\u00d7\u20092\u2009mm, which might lead to inappropriate ROI placement on the lesion. Third, for VTIQ imaging, it was hard to completely avoid the calcification areas when the calcifications were diffused in the lesion, which might lead to bias since calcification in a lesion might increase tissue stiffness. Final, to confirm the factors associated with axillary LNM, a prospective, multicenter study with larger sample size is required in the future.\n\nConclusion {#Sec10}\n==========\n\nIn conclusion, lesion size \u226520\u2009mm, taller than wide shape, VTI score \u22655, histological grade III, and positive C-erbB-2 are risk factors for axillary LNM. The risk model developed in the study could predict the risk of axillary LNM in patients with breast IDCs and thus might be helpful to avoid unnecessary ALND.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupported in part by Grant SHDC12014229 from Shanghai Hospital Development Center, Grants 14441900900 and 15411969000 from Science and Technology Commission of Shanghai Municipality.\n\nConcept and design: H.X. and X.L. Performance of conventional US, VTI, and VTIQ: W.Y. Collection and assemble of data: X.L., Y.H., B.L. and L.F. Data analysis and interpretation: D.L., J.X. and L.S. Manuscript written: H.X. and X.L. Reviewed the manuscript: All authors.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "1. Introduction {#sec38574}\n===============\n\nMalaria is a protozoan infection that is transmitted to humans through the bite of the Anopheles mosquito. This is the most important parasitic disease which is responsible for about one to three million deaths each year. Malaria today, as in the past, is a very big threat for tropical communities and nonendemic countries as well as travellers ([@A8752R1], [@A8752R2]). Annually, more than 2,700,000 people are at risk in malarious areas of Iran and about 3,000 malaria cases are reported ([@A8752R3]). Annually, about 1,000 cases of malaria are reported in the United States, most of which are passengers or immigrants ([@A8752R4]). Four species of *Plasmodium* including *P. falciparum*, *P. malariae*, *P. ovale*, and *P. vivax,* cause nearly all malaria infections in humans. However, almost all malaria-attributable mortalities are due to *P. falciparum*.\n\nThe human infection starts when the bite of an infected female Anopheles mosquito injects the *Plasmodium* sporozoites from the salivary gland into the human blood circulation. These microscopic malaria parasites rapidly travel through the bloodstream to the liver and begin the period of asexual reproduction. During this replication process, known as intrahepatic or preerythrocytic merogony step, 10,000 to more than 30,000 daughter merozoites are produced from sporozoites. Afterwards, the liver cell containing these merozoites ruptures and releases them into the bloodstream. At this stage, the infection is symptomatic. The merozoites, immediately after entering the bloodstream, invade the red blood cells (RBCs) and become trophozoite. Since *P. vivax*, *P. ovale*, and *P. malariae* have tendency to invade mature RBCs, parasitemia rarely exceeds 2%; whereas in *Falciparum* malaria, severe parasitemia occurs due to invasion to RBCs of all ages ([@A8752R2]).\n\nIn this paper we introduce an Iranian case of severe malaria rapidly responding to treatment with intravenous quinine sulfate compared to other cases of severe malaria in the world.\n\n2. Case Presentation {#sec38575}\n====================\n\nIn January 2001, a 32-year-old man referred to Montaserieh hospital of Mashhad, Iran, with a history of fevers and vomiting 10 days after returning from Pakistan. He also complained of dark urine and hematuria. He had no history of receiving prophylactic antimalarial drugs. He had previously been hospitalized in a small hospital with a diagnosis *P. vivax* infection and had received oral chloroquine therapy. Then the patient was referred to a specialized hospital in Mashhad city due to worsening symptoms. At admission, he had a temperature of 38.5\u00b0C and his other vital signs and consciousness level were normal. On physical examination, the patient had a tender abdomen and his spleen was palpable 5 cm below the left costal margin.\n\nIn laboratory studies, hemoglobin was 11.5 g/dL, hematocrit 34%, white blood cell count 16,000/mm^3^, and platelet count was 45,000/mm ^3^. His creatinine level was 1.9 mg/dL. Serum was assayed for total bilirubin (Bil), lactate dehydrogenase (LDH), aspartate aminotransferase (AST), and alanine aminotransferase (ALT), using commercially available kits. The serum level of Bil was 2.3 mg/dL, LDH 1680 U/L (normal level: 80 - 100 U/L), AST 106 U/L, and ALT 29 U/L. Peripheral blood smear examination showed severe infection with *P. falciparum*and 75 to 85% of RBCs contained one to five parasites per cell ([Figure 1](#fig8096){ref-type=\"fig\"}).\n\n![Peripheral Blood Smear of a 32-Year-Old Iranian Patient With a Severe Infection of *P. falciparum*\\\nAmong RBCs, 75 to 85% contained one to five parasites per cell (Giemsa stain, X1000)](jjm-07-8752-g001){#fig8096}\n\nPatient probably had a mixed infection with *P. vivax* and *P. falciparum* that the *P. vivax* infection had been treated with oral chloroquine, but chloroquine-resistant *P. falciparum* remained and parasitemia had increased.\n\nThe patient was treated with intravenous quinine sulfate and oral primaquine. Then, he received primaquine (30 mg a day for 14 days) after the symptoms improvement for treating the probable mixed malaria infection. The improvement was considerable, so that parasitemia decreased by 40% after 24 hours and 1% after 4 days of posttreatment. During the treatment process, the patient developed dark urine and elevated creatine phosphokinase; thus, progression to acute renal failure was considered for him. Acute respiratory distress syndrome (ARDS) was exhibited in bilateral pulmonary infiltrates on the chest X-ray; also, he presented confusion due to cerebral malaria or superimposed a Gram-negative infection, so ceftriaxone was prescribed. However, patient was discharged after 10 days with no requirement for intubation, hemodialysis, or ICU admission. He had a good general condition at the last follow-up visit.\n\n3. Discussion {#sec38576}\n=============\n\nSevere *falciparum* malaria is a medical emergency that requires intensive care. Parenteral quinine or quinidine should be administered if any doubt exists about drug resistance. In very ill patients, blood exchange transfusion should be considered and it is indicated in high parasitemia (more than 30%) ([@A8752R5]). The patient presented in this paper was important in two aspects: one was because of high parasitemia which was certainly a very rare condition and the only similar reported case in the world with high parasitemia as our patients was a Swiss case (with a parasitemia of 70 - 80%), and second, the rapid response to treatment with intravenous quinine without the necessity of blood exchange or other supportive care proceedings (such as hemodialysis, ventilator, or ICU admission) and improvement after 10 days hospitalization.\n\nOutcomes of sever malaria differ in immune and nonimmune patients ([@A8752R6]). We found some reports of immune and nonimmune patients with severe malaria. The first case was a 34-year-old Swiss patient with a history of travel to Madagascar presenting severe *P. falciparum*malaria in which 70 - 80% of RBCs in the peripheral blood smear were infected ([Figure 2](#fig8097){ref-type=\"fig\"}) ([@A8752R7]). The treatment course took about 3 months and he was hospitalized during this process. In addition to administration of intravenous quinine sulfate, he underwent blood exchange transfusion as well as laparotomy and splenectomy due to splenic rupture. Also, he was placed on mechanical ventilation, and underwent hemodialysis. Although the rate of parasitemia was similar to that of our patient, response to the treatment was completely different ([@A8752R7]).\n\n![Peripheral Blood Smear in a 34-Year-old Swiss Patients With Severe Malaria\\\nAmong RBCs, 70 to 80% contained up to five *P. falciparum* cells ([@A8752R7]).](jjm-07-8752-g002){#fig8097}\n\nThe other case was a 54-year-old patient, progressed to severe malaria (*P. falciparum*) after returning from Kenya with a parasitemia rate of 10.4% ([@A8752R8]). Treatment with intravenous quinine and doxycycline were initiated for him. Despite the elimination of parasites from the blood, patient died from pulmonary edema and heart failure due to water and electrolyte abnormalities. Although the percentage of parasitemia was lower and even the treatment contained broad-spectrum drugs, he was a nonimmune case from a nonendemic region and so had a poorer prognosis than the two previously mentioned cases.\n\nThe other patient was a 17-year-old UK-resident woman who was born in Nigeria. She presented a maximum parasitemia of 30% (*P. falciparum*) after returning from her native country. Although she was admitted to ICU and undergone ventilation and hemodialysis, she was discharged with good general conditions and no need for blood exchange transfusion on the day 10 after the admission. In this case, although the disease progress was associated with more complications than ours, no need for blood exchange as well as the short length of hospital stay (10 days) indicated the relative immunity of patient which was due to her birthplace and history of living in an endemic region ([@A8752R9]).\n\nOther case was a 29-year-old Canadian man who had non-specific symptoms of malaria after returning from Zambia. The peripheral blood smear was negative for malarial parasite. The patient's consciousness was impaired 48 hours later and he died in the next few hours. *P. falciparum* was reported as the cause of death in autopsy. Although this patient received no medical treatment, the rapid and fatal course of malaria infection even with negative parasitemia in a nonimmune patient living in a nonendemic region, draws our attention to the importance of the relative immunity of people living in endemic regions ([@A8752R10]).\n\nFinally, in another article, four patients with *P. falciparum* were introduced (two 50-year-old Canadian patients, male and female, with parasitemia of 1%, a 28-year-old Norwegian patient with parasitemia of 2%, and a 64-year-old Indian patient with parasitemia of 6%). The Indian patient, compared to the three others (who lived in nonendemic regions) despite a higher rate of parasitemia, had lower symptoms and a faster response to treatment ([@A8752R11]). Presumably, the theory of a relative immunity existence among people living in the malaria endemic regions, leads to a faster response to treatments and better prognosis. This is quite reasonable considering the collected data and cases from around the world. Therefore, the indications of blood exchange, ICU admission, laparotomy, and splenectomy, as well as the disease prognosis in severe malaria, can be different based on a person's location and immunity.\n\nWe wish to thank staffs of Department of Infectious Disease in Mashhad University of Medical Sciences.\n\n**Implication for health policy/practice/research/medical education:**In this article, we introduce an Iranian patient suffering from *P. falciparum*. Peripheral blood smear for malaria parasites showed a severe infection of P. falciparum. The findings revealed that generally, patients from endemic regions significantly show a greater response to treatment and also a better prognosis in comparison to the patients from non-endemic regions.\n\n**Authors' Contribution:**Masood Ziaee, first edition and drafting of the manuscript. Farshid Abedi, corresponding author and contribution to abstracting the data and preparing the manuscript.\n\n**Financial Disclosure:**Authors have no financial interests related to the manuscript.\n\n**Funding/Support:**The authors have nothing to disclose.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nPulmonary arterial hypertension (PAH) is a severe condition that frequently leads to right ventricular failure and death \\[[@B1], [@B2]\\]. Remodeling of small pulmonary arteries represents the main pathologic finding related to PAH. The remodeling process includes marked proliferation of pulmonary artery smooth muscle cells (PASMCs) that contribute to vascular wall hypertrophy \\[[@B1], [@B3]\\].\n\nStudies in both animal models and in human patients have shown that cigarette smoke exerts a direct effect on pulmonary vascular structure, resulting in pulmonary vascular remodeling and pulmonary hypertension \\[[@B4], [@B5]\\]. However, the molecular mechanisms underlying this process remain unclear.\n\nPlatelet-derived growth factor (PDGF) is a potent mitogen and chemoattractant for pulmonary vascular smooth muscle cells \\[[@B6]\\]. Although there are several PDGF family members (A, B, C, and D), PDGFB is a prime candidate to study, because the PDGFB isoform is a potent inducer of VSMC proliferation and only acts through the beta-subunits of the PDGF receptor (PDGFR) \\[[@B6]--[@B8]\\]. The upregulated expression of PDGF and PDGFR has been correlated with PAH in various experimental animal models \\[[@B9], [@B10]\\] and in humans \\[[@B11]\\]. However, whether PDGF is also activated in rPASMCs by cigarette smoke extract (CSE) has not been explored.\n\nProtein kinase C (PKC) is a crucial family of the serine-threonine kinases in the intracellular signal transduction pathway. It has been implicated in a variety of cellular functions in vascular smooth muscle cells, including proliferation \\[[@B12]\\], apoptosis \\[[@B13]\\], and differentiation \\[[@B14]\\]. Recent studies have shown that CSE activates PKC in monocytes \\[[@B15]\\], bronchial epithelial cells \\[[@B16]\\], and pulmonary artery endothelial cells \\[[@B17]\\]. It has also been reported that PKC*\u03b4* plays an important role in hypoxia-induced proliferation of PASMCs \\[[@B18]\\]. Based on this background, we hypothesize that CSE may activate PKC*\u03b4* to stimulate PDGFB and activate a cellular response in rPASMCs.\n\nThis study was designed to examine the impact of CSE on PKC*\u03b4* expression to promote rPASMCs proliferation by upregulating the expression of PDGFB and PDGFR*\u03b2*, which could contribute to pulmonary vascular remodeling.\n\n2. Materials and Methods {#sec2}\n========================\n\n2.1. Medium and Reagents {#sec2.1}\n------------------------\n\nThe rPASMCs were purchased from ATCC (Jing-Ke-Hong-Da Company, Beijing, China), and fetal bovine serum (FBS) and penicillin/streptomycin were obtained from Gibco Life Technologies Inc. (Rockville, MD, USA). Dulbecco\\'s modified Eagle\\'s medium (DMEM) was obtained from Clontech. The mouse polyclonal antibody against *\u03b2*-actin and the rabbit polyclonal antibodies against PKC*\u03b4*, piPKC*\u03b4*, and Rottlerin were all purchased from Sigma-Aldrich (St. Louis, MO, USA). Mouse monoclonal antibodies against PDGFB and PDGFR*\u03b2* were purchased from Bioworld Technology Inc. (St. Louis Park, MN, USA). Fluorescein isothiocyanate- (FITC-) conjugated goat anti-rabbit IgG and horseradish peroxidase- (HRP-) conjugated goat anti-rabbit IgG were purchased from Pierce Biotechnology Co. Ltd (Rockford, IL, USA). Trizol reagent was obtained from Invitrogen (Carlsbad, CA, USA). All primers were synthesized by Aoke Bio-Engineering Co. (Beijing, China).\n\n2.2. Cell Culture and Treatment {#sec2.2}\n-------------------------------\n\nCells were cultured at 37\u00b0C in DMEM supplemented with 10% FBS and antibiotics (penicillin and streptomycin) in a 5% CO~2~ incubator. Cells were treated with varying concentrations of CSE (0, 2.5%, 5%, 10%, and 20%) for 24\u2009h, with or without a PKC*\u03b4* inhibitor. For all experiments, cells were made quiescent in serum-free media for 24\u2009h before exposure to CSE. Control untreated cells were maintained in serum-free media.\n\n2.3. Preparation of CSE Solution {#sec2.3}\n--------------------------------\n\nCSE was prepared as previously described by Oltmanns et al. with a few modifications \\[[@B19]\\]. Commercial filtered cigarettes (Daqianmen, Shanghai Cigarette Factory, China), which contain 13\u2009mg of tar and 1.0\u2009mg of nicotine/cigarette, were used. Briefly, cigarette smoke derived from one cigarette was drawn slowly into a 50\u2009mL syringe and bubbled through 30\u2009mL of DMEM at room temperature. A cigarette yielded 5 draws with a 50\u2009mL syringe, with each individual draw taking approximately 10 seconds to complete. The resulting solution, which was considered 100% strength, was adjusted to pH 7.4 with concentrated NaOH and filtered through a 0.22 *\u03bc*m pore filter before diluting it into DMEM as a percentage of the total volume. Final concentrations of the solution are expressed as percent values ((CSE solution volume/total volume) \u00d7 100%). Solution concentrations ranging from 0 to 20% (0%, 2.5%, 5%, 10%, and 20%) were used to evaluate the effect of CSE on rPASMCs proliferation. The experiment was repeated at least twice for each biological sample.\n\n2.4. Cell Proliferation Analysis (MTT Assay) {#sec2.4}\n--------------------------------------------\n\nThe effect of CSE treatment on cell viability and proliferation was examined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. Cells (1 \u00d7 10^4^ per well) were seeded in 96-well plates in replicates of five. After 24\u2009h, the absorbance at 490\u2009nm was recorded.\n\n2.5. Cell Cycle Analysis {#sec2.5}\n------------------------\n\nFlow cytometric analyses of cells were conducted according to the manual provided with the propidium iodide (PI) flow kit (KeyGEN, Nanjing, China). Briefly, the cells were harvested and stained with PI solution for 15\u2009min. Fluorescence was measured using a FACScalibur apparatus (Becton Dickinson, Franklin Lakes, NJ, USA). Data collection and analysis were performed using CellQuest software. The percentages of cells in different phases of the cell cycle were determined using the software program ModFit.\n\n2.6. RT-PCR Assay {#sec2.6}\n-----------------\n\nA total of 4 \u00d7 10^6^ cells were lysed to isolate total RNA using Trizol reagent (Invitrogen), according to the manufacturer\\'s instructions. Reverse transcription (RT) was performed as described by the company. Total RNA (5\u2009*\u03bc*g) was RT to synthesize cDNA in a volume of 20\u2009*\u03bc*L (Reverse Transcriptase M-MLV, Takara). For every 25\u2009*\u03bc*L of PCR reaction, 1\u2009*\u03bc*L of cDNA was used for 20--25 cycles. PCR products (10\u2009*\u03bc*L) were loaded onto a 2% agarose gel, stained with ethidium bromide and photographed. The primer sequences used for the RT-PCR are as follows: PDGFB forward 5\u2032-AAGTGAGCAGTGCCTCGGCG-3\u2032, reverse 5\u2032-TGGCCTCGGGGACCGTCAAA-3\u2032; PDGFR*\u03b2* forward 5\u2032-GCCGCCAGCGCCCATTTTTC-3\u2032, reverse 5\u2032-AGGCGTTTTCCTCTGCCCGC-3\u2032, and PKC*\u03b4* forward 5\u2032-AGCAGCGGGAGCCAGGACTA-3\u2032; reverse 5\u2032-GCGGTGCGGCATGTCGATGT-3\u2032.\n\n2.7. Immunoblot Analysis {#sec2.7}\n------------------------\n\nCells were lysed in ATM lysis buffer (containing 100\u2009mM Tris-Cl, pH 7.5, 150\u2009mM NaCl, 0.2\u2009mM EDTA, 20% glycerol, 0.4% NP-40, 2% Tween-20, and 0.2\u2009mM PMSF). The protein concentration of the supernatant was measured with a BCA Assay Kit (Calbiochem), and total protein (50\u2009*\u03bc*g) was loaded onto 15% SDS-PAGE gels. After electrophoresis, proteins were transferred onto PVDF membranes (Amersham) and hybridized with primary antibodies at the following dilutions: PDGFR*\u03b2* (Bio world; Cat No: BS1764; 1\u2009:\u20092,000), PDGFB (Bio world; Cat No: BS1290 1\u2009:\u20092,000), PKC*\u03b4* (Sigma; 1\u2009:\u20092,000), and phosphorylated PKC*\u03b4* (Sigma; 1\u2009:\u20092,000). HRP-labeled secondary antibodies (Zhongshan Golden Bridge) were all used at a dilution of 1\u2009:\u20092,000. The ECL detection system (Amersham) was used to detect the signals on the membranes.\n\n2.8. Statistical Analysis {#sec2.8}\n-------------------------\n\nAll data were presented as mean \u00b1 standard deviation. Statistical analyses were carried out using one-way ANOVA (for multiple-group comparison) followed by the least significant difference (LSD) test with the computer software SPSS 12.0 (Chicago, USA). For all tests, groups were considered statistically significant when *P* \\< 0.05.\n\n3. Results {#sec3}\n==========\n\n3.1. Effect of CSE Treatment on rPASMCs Proliferation {#sec3.1}\n-----------------------------------------------------\n\nTo investigate the effect of CSE treatment on cell proliferation, rPASMCs were stimulated with 0%, 2.5%, 5%, 10%, and 20% CSE for 24\u2009h. Cell proliferation was evaluated by MTT colorimetric assay. As shown in [Figure 1(a)](#fig1){ref-type=\"fig\"}, CSE treatment caused a significant increase in cell growth in a dose-dependent manner compared to the control group. The cell response reached its peak at the concentration of 20% CSE.\n\n3.2. Effect of CSE Treatment on rPASMCs Cell Cycle Progression {#sec3.2}\n--------------------------------------------------------------\n\nBecause cell proliferation is critically related to cell cycle changes, we further investigated the effect of CSE treatment on cell cycle progression. The rPASMCs were treated with 0%, 2.5%, 5%, 10%, and 20% CSE for 24\u2009h, cell cycle was examined using the FACS assay with PI staining. According to the flow cytometry analysis, we found that there was an obvious increase in the S and G2 phases in CSE-treated cells compared to the nontreated control group. The data also indicated that there was a slight reduction in G2 phase cells when cells were treated with CSE at concentration of 10% ([Figure 1(b)](#fig1){ref-type=\"fig\"}).\n\n3.3. Effect of CSE on rPASMCs PDGFB and PDGFR*\u03b2* Expression {#sec3.3}\n-----------------------------------------------------------\n\nWe first investigated the effect of CSE treatment on the expression of PDGFB and PDGFR*\u03b2* mRNA and protein levels. As shown in [Figure 1(c)](#fig1){ref-type=\"fig\"}, CSE significantly increased the mRNA and protein levels of PDGFB and PDGFR*\u03b2* compared to the nontreated control rPASMCs.\n\nAccording to the cell proliferation and cell cycle results, we selected the 10% CSE treatment to analyze the molecular mechanisms of CSE-induced cell proliferation. We further investigated the effect of PKC*\u03b4* inhibition with Rottlerin (a PKC*\u03b4*-specific inhibitor) on upregulated cell proliferation induced by CSE treatment. As shown in [Figure 2(a)](#fig2){ref-type=\"fig\"}, pre-incubation of rPASMCs with Rottlerin before CSE exposure significantly decreased the CSE-induced increased cell proliferation.\n\nSince cell proliferation was reduced with Rottlerin administration, we determined whether Rottlerin affected the cell cycle alterations with CSE treatment. As illustrated in [Figure 2(b)](#fig2){ref-type=\"fig\"}, preincubation of rPASMCs with Rottlerin before CSE exposure attenuated the increase of S and G2 phase populations induced by CSE treatment.\n\nWe also investigated the changes of mRNA and protein levels of PKC*\u03b4* after CSE treatment. We found that 10% CSE treatment had no obvious effect on total PKC*\u03b4* mRNA and protein levels. PKC*\u03b4* has many different modifications that affect its activity, including phosphorylation to generate the active form. Therefore, we explored whether CSE treatment affected the phosphorylation of PKC*\u03b4*. We measured the phosphorylation level of PKC*\u03b4* and found that CSE treatment indeed promoted the phosphorylation of PKC*\u03b4* ([Figure 2(c)](#fig2){ref-type=\"fig\"}).\n\nWe further investigated the effect of Rottlerin on the level of PDGFB, PDGFR*\u03b2*, and phosphorylated PKC*\u03b4*. Cells were incubated with different concentrations of Rottlerin prior to 10% CSE treatment, harvested, lysed, and subjected to RT-PCR or immunoblot assay. As shown in [Figure 2(c)](#fig2){ref-type=\"fig\"}, Rottlerin treatment significantly reduced the increase in PDGFB, PDGFR*\u03b2*, and phosphorylated PKC*\u03b4* levels induced by CSE treatment. These changes occurred at both the mRNA and protein level.\n\n4. Discussion {#sec4}\n=============\n\nCigarette smoking is a major public health problem associated with a multitude of diseases, including vascular diseases such as atherosclerosis and pulmonary hypertension. Proliferation of vascular smooth muscle cells plays an important role in the development of these vascular diseases. The relationship between cigarette smoking and cell function has been explored by stimulating isolated cells with CSE to determine direct causes \\[[@B20]\\]. In this study, we demonstrated that CSE induced rPASMCs proliferation by upregulating phosphorylating PKC*\u03b4* and PDGFB. This stimulation was suppressed with a PKC*\u03b4* inhibitor Rottlerin, which implicate PKC*\u03b4* and PDGFB as therapeutic targets to block abnormal smooth muscle cell proliferation induced by CSE.\n\nAbnormal vascular smooth muscle cell proliferation leads to medial hypertrophy of the vessel wall, vascular remodeling, and vascular lumen narrowing, all of which contribute to the development of pulmonary hypertension \\[[@B21], [@B22]\\]. In our *in vitro* study, CSE significantly promoted the proliferation of rPASMCs. Because cell proliferation is critically related to cell cycle changes, we further analyzed the effect of CSE on cell cycle distribution. Our study showed that CSE promoted cell cycle progression from the G1 phase to the S phase, indicating that CSE promoted cell proliferation. It is important to note that the results of the current study are consistent with those obtained by Luppi et al. \\[[@B23]\\], but contradictory to other investigations. In particular, a report by Ambalavanan et al. found that 20% CSE did not stimulate the proliferation of neonatal porcine vascular smooth muscle cells. Instead, CSE induced cell necrosis in the study \\[[@B24]\\]. A major difference between that study and ours is that neonatal porcine cells were used in the former while our study used adult rat cells. The different CSE preparation methods between the two studies could be another possible explanation for these discrepancies.\n\nMany different cell types synthesize PDGF, including smooth muscle cells, endothelial cells, and macrophages \\[[@B25]\\]. PDGF has the ability to induce the proliferation and migration of smooth muscle cells and fibroblasts and has been proposed as a key mediator in the progression of several fibroproliferative disorders such as atherosclerosis, lung fibrosis, and PAH \\[[@B11], [@B25], [@B26]\\]. There is much evidence to support the hypothesis that PDGF pathways play an important role in the pulmonary vascular remodeling process responsible in the progression of PAH \\[[@B10]\\]. Indeed, PDGF is known to induce proliferation of smooth muscle cells of different origins such as the aorta and carotid \\[[@B27], [@B28]\\]. Our results showed that after rPASMCs were exposed to CSE, the expression of PDGFB and PDGFR*\u03b2* mRNA and protein levels were markedly increased. In the current study, we extended the previous findings and demonstrated that PDGF was involved in the proliferation of rPASMCs induced by CSE.\n\nMoreover, we investigated the molecular mechanisms by which PDGF stimulated proliferation of rPASMCs. PKC is a critical regulator in the intracellular signal transduction pathway and plays an important role in PASMCs proliferation \\[[@B29]\\]. Currently, 10 members in the PKC family have been identified, and each isozyme has distinct cellular functions \\[[@B30]\\]. There are at least four isozymes in smooth muscle cells, including PKC*\u03b4*, PKC*\u03b1*, PKC*\u03b5*, and PKC*\u03b6* \\[[@B31]\\]. This study focused on the role of the PKC*\u03b4* isoform in CSE-induced rPASMCs proliferation. Although it has been shown that PKC*\u03b4* activation can induce smooth muscle cell proliferation \\[[@B32]\\], Sasaguri and colleagues found that activation of PKC*\u03b1* and PKC*\u03b5* blunted smooth muscle cell proliferation \\[[@B33]\\]. In our study, preincubation with Rottlerin, a specific PKC*\u03b4* inhibitor, inhibited cell proliferation, caused cell cycle arrest at the G1 phase, and downregulated PDGFB and PDGFR*\u03b2* mRNA and protein levels in rPASMCs exposed to CSE. These results suggest that CSE induces rPASMCs proliferation by upregulating PDGFB and PDGFR*\u03b2* expression mediated by the PKC*\u03b4* signaling pathway.\n\nAlthough the specific PKC*\u03b4* inhibitor Rottlerin significantly suppressed PDGFB and PDGFR*\u03b2* expression and CSE-induced rPASMCs proliferation, the upregulation was not completely inhibited. This result indicates that other mediators might also be involved in the process, such as EAK signaling, which plays an important role in mediating growth factor functions. The EAK signaling pathway has been shown to mediate cell proliferation in other cell types \\[[@B34]\\]. Future studies evaluating the roles of additional signaling cascades on cell proliferation in rPASMCs are warranted.\n\nIn conclusion, we determined that the effect of CSE on rPASMCs proliferation occurs through PKC*\u03b4* and PDGFB activation. This study provides a foundation on which to further examine the precise roles of PDGF in pulmonary vascular remodeling that leads to PAH. Understanding the intracellular mechanisms responsible for rPASMCs growth provides therapeutic targets to prevent or limit rPASMCs proliferation and pulmonary vascular remodeling.\n\n5. Conflict of Interests {#sec5}\n========================\n\nThe authors affirm that they have no conflict of interests.\n\nThis work was supported by Shanxi Scholarship Council of China (no. 2011-094).\n\n![(a) CSE promoted cell proliferation of rPASMCs. The rPASMCs were seeded in a 96-well plate and treated with increasing concentrations of CSE for 24\u2009h. Cell proliferation was analyzed by MTT assay. The data are shown as mean \u00b1 standard deviation from 6 experiments and were analyzed by ANOVA. \\**P* \\< 0.05 compared with control. (b) CSE promoted cell cycle progression of rPASMCs. The rPASMCs were treated with increasing concentrations of CSE for 24\u2009h. Cells were harvested and stained with PI solution for cell cycle assay. S phase and G2 phase cells were averaged. The data are shown as mean \u00b1 standard deviation from 6 experiments and were analyzed by ANOVA. \\**P* \\< 0.05 compared with control. (c) The PDGF signaling pathway was activated after CSE treatment. The rPASMCs were treated with increasing concentrations of CSE for 24\u2009h. Cells were lysed and submitted to RT-PCR or immunoblot assays, with the indicated primers or antibodies.](JBB2012-534384.001){#fig1}\n\n![(a) The PKC*\u03b4*-specific inhibitor Rottlerin inhibited 10% CSE-induced cell proliferation of rPASMCs. The rPASMCs were treated with different concentrations of Rottlerin for 1\u2009h before exposure to 10% CSE. Cell proliferation was analyzed by MTT assay. The data are shown as mean \u00b1 standard deviation from 6 experiments and were analyzed by ANOVA. \\**P* \\< 0.05 compared with control and ^\\#^*P* \\< 0.05 compared with 10% CSE. (b) The PKC*\u03b4*-specific inhibitor Rottlerin inhibited the CSE-induced cell cycle progression of rPASMCs. The rPASMCs were treated with different concentrations of Rottlerin for 1\u2009h before exposure to 10% CSE. Cells were prepared and stained with propidium iodide solution for cell cycle assay. S phase and G2 phase cells were averaged. The data are shown as mean \u00b1 standard deviation from 6 experiments and were analyzed by ANOVA. \\**P* \\< 0.05 compared with control and ^\\#^*P* \\< 0.05 as compared with 10% CSE. (c) The PKC*\u03b4*-specific inhibitor Rottlerin inhibited the CSE-induced upregulation of the mRNA and protein levels of PDGF signaling pathway. The rPASMCs were treated with different concentrations of Rottlerin for 1\u2009h before exposure to 10% CSE. Cell extracts were prepared and submitted to RT-PCR or immunoblot assay with primers or antibodies as indicated.](JBB2012-534384.002){#fig2}\n\n[^1]: Academic Editor: Leon Spicer\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\n*Musa paradisiaca* (plantain) is a crop in the genus Musa and all members of the genus are indigenous to the tropical and subtropical countries.[@ref1][@ref2] It has been cultivated for more than 4000 years, and its several varieties are staple food in the tropical regions of the world.[@ref3] *Musa paradisiaca* promotes healthy digestion, improves affective state, helps in the retention of and serves as good sources of potassium, calcium, phosphorus and nitrogen, which build and regenerate tissues in the body, and is also a rich source of iron and vitamins, especially Vitamins C and E.[@ref1]--[@ref3] Green plantain is also high in total dietary fibre content, especially hemicelluloses, which is higher than in most fruits and vegetables.[@ref4] The high fibre contents, particularly insoluble fibre, can lower glycaemic response by forming a physical barrier to enzymatic hydrolysis of starch. The leaves, roots and fruits of plantain have been a revolutionary breakthrough in the management of male sexual inadequacies.[@ref5][@ref6]\n\nInfertility is a public health concern in many developing nations of the world due to its high prevalence and especially because of its serious social implications.[@ref7] Infertility has social, economic and personal effects, which go beyond childlessness, and women bear the major brunt of the burden.[@ref8] It is a major cause of marital disharmony and separation, and personal misery in some other population groups.[@ref9] The highest prevalence is in low resource countries, particularly in sub-Saharan Africa.[@ref8] Male infertility can be caused by poor penile erection, abnormal sperm quality and volume, abnormal ejaculation, among other causes. Researches into natural diets like plantain showed that its consumption by men could enhance some reproductive functions, and also alleviate certain reproductive dysfunctions.[@ref6][@ref10] The aim of the current study, therefore, is to determine the effects of consumption of mature green plantain fruits on semen quality in normal adult Wistar rats.\n\nMATERIALS AND METHODS {#sec1-2}\n=====================\n\nThe study was carried out in conformity with the rules and guidelines of the Animal Ethics Committee of the University of Ilorin. Eighteen (18) adult male Wistar rats of average weight of 200 g were obtained and maintained in the Animal House of the College of Health Sciences, University of Ilorin. The animals were kept at normal room temperature and food and water made available *ad libitum*. They were grouped into three, each group having six Wistar rats.\n\nPreparation of the plantain flour {#sec2-1}\n---------------------------------\n\nGreen plantain fruits were obtained from a market in Ilorin, Nigeria. The fruits were cut longitudinally into chips of about 5 mm thickness and air-dried for 4 days after which they were grinded and made into flour. Two doses of the plantain flour were prepared: 500 mg/kg/day and 1000 mg/kg/day. The flour was dissolved in 2 ml of double distilled water, for easy administration.\n\nAnimal grouping {#sec2-2}\n---------------\n\nThe animals were divided into three groups:\n\nGroup A: Control (2 ml double distilled water, p.o.)\n\nGroup B: Given low dose *Musa paradisiaca* (500 mg/kg/day, p.o.)\n\nGroup C: Given high dose *Musa paradisiaca* (1000 mg/kg/day, p.o.).\n\nThe treatment lasted for a period of 28 days.[@ref10][@ref11]\n\nAnimal sacrifice {#sec2-3}\n----------------\n\nTwenty-four (24) hours after the 28^th^ day of treatment, the rats were sacrificed by cervical dislocation, and the testes and epididymis excised using a midline abdominal incision. The testes were immediately weighed and the left caudal epididymis transferred into sterile bottles containing 2 ml of normal saline for semen analysis.\n\nSemen analysis: Sperm concentration {#sec2-4}\n-----------------------------------\n\nA modified method of Yokoi and Mayi[@ref12] was adopted in counting the spermatozoa, with the aid of the new improved Neuber\\'s Counting Chamber (haemocytometer). About 10 \u03bcl of the diluted sperm suspension was transferred to each counting chamber of the haemocytometer and was allowed to stand for 5 min, and thereafter observed under a binocular light microscope.[@ref12][@ref13]\n\n### Sperm motility {#sec3-1}\n\nThe fluid from the caudal epididymis was diluted with Tris buffer solution[@ref14] to 0.5 ml. An aliquot of this solution was observed under the light microscope. The mean motility estimation was reported as the final motility score for each sample, and presented in percentages.\n\n### Sperm morphology {#sec3-2}\n\nThe morphology of the spermatozoa was determined using the original dilution for motility, diluted 1:20 with 10% neutral buffered formalin. The sperm cells were categorized based on the presence of one or more abnormal features, such as tail defects (short, irregular coiled or multiple tails); neck and middle piece defects (distended, irregular, bent middle piece, abnormally thin middle piece); and head defects (round head, small or large size, double or detached head). Findings were expressed as percentage of morphologically normal sperm.[@ref15]\n\nLife--death ratio {#sec2-5}\n-----------------\n\nLife-Death ratio was taken as, number of spermatozoa alive divided by total number of spermatozoa multiplied by 100. (No. of spermatozoa alive \u00f7 Total No. of spermatozoa) \u00d7 100\n\nStatistical analysis {#sec2-6}\n--------------------\n\nData were analyzed statistically by application of Student\\'s *t*-test, using the SPSS version 15.0 software and presented as mean and standard error mean (SEM). Values of *P* \\< 0.05 were considered to be statistically significant.\n\nRESULTS {#sec1-3}\n=======\n\nThe plantain flour was well tolerated, with no signs suggestive of regurgitations after oral administration. There was a relative gain in weight in the treated groups, and the weight gain was more in the group given the low dose *Musa paradisiaca* (Group B) with statistically significant difference (*P* \\< 0.05), while the weight gain recorded in the high dose group (Group C) was lower than in Group B (*P* \\> 0.05) \\[[Table 1](#T1){ref-type=\"table\"}\\].\n\n###### \n\nWeights of rats and testes following administration of *Musa paradisiaca*\n\n![](NMJ-54-92-g001)\n\nThe semen parameters obtained in the treated groups showed statistically significant differences between the two groups given *Musa paradisiaca* (*P* \\< 0.05) \\[[Table 2](#T2){ref-type=\"table\"} and [Figure 1](#F1){ref-type=\"fig\"}\\]. The animals administered with low dose of *Musa paradisiaca* had a slight increase in sperm count, which was not statistically significant (*P* \\> 0.05), whereas a high dose of *Musa paradisiaca* resulted in a significantly decreased sperm concentration (*P* \\< 0.05) compared with the control group.\n\nThe percentage motility increased in Wistar rats given low dose of *Musa paradisiaca*, but decreased slightly in the high dose group (*P* \\> 0.05) compared with the control \\[[Table 2](#T2){ref-type=\"table\"}\\]. Administration of *Musa paradisiaca* caused reduction in the number of morphologically normal sperm, with a very marked reduction noticed in the high dose group (*P* \\< 0.05) compared with the control animals \\[[Table 2](#T2){ref-type=\"table\"} and [Figure 1](#F1){ref-type=\"fig\"}\\].\n\n###### \n\nResults of semen analysis\n\n![](NMJ-54-92-g002)\n\n![Semen parameters following administration of *Musa paradisiaca* to Wistar rats. ^\u2020^Significant statistical difference between the low dose and high dose groups (*P* \\< 0.05) of all the parameters. ^\\*^Significant statistical difference between the high dose groups and the Control Group (*P* \\< 0.05). ^\\*\\*^Significant statistical difference between the low dose group and the Control Group (*P* \\< 0.05)](NMJ-54-92-g003){#F1}\n\nThere was no significant difference in the Life-Death ratio between the control animals and the high dose group (P \\> 0.05), but in the low dose group, whereas there was a significant increase compared with the control (*P* \\< 0.05) \\[[Table 2](#T2){ref-type=\"table\"} and [Figure 1](#F1){ref-type=\"fig\"}\\]. Sperm progressivity was reduced in the high dose group, but the low dose group had forward directional movement almost similar to those of the control group \\[[Table 2](#T2){ref-type=\"table\"}\\].\n\nDISCUSSION {#sec1-4}\n==========\n\n*Musa paradisiaca* has been documented to have many beneficial effects in the management of several diseased conditions, including atherosclerosis, diabetes mellitus, hypertension, hyperlipidaemia and thyroid dysfunctions[@ref16]--[@ref21] and also offers protective effects on organs of the body, such as the kidneys, in certain clinical conditions.[@ref22] Yakubu *et al*.[@ref6] reported that plantain fruits can be used in the management of sexual dysfunctions, though the exact mechanism of action or how it enhances this is not known. Similarly, studies conducted on methanolic extract of *Musa paradisiaca* (MEMP) fruit[@ref10] revealed that animal models with diabetic-induced testicular disorders can have the testicular damage reversed, when given MEMP. The precise mechanism of action by which MEMP does this is still unclear. *Musa paradisiaca* has hypoglycaemic effects in both normal and diabetic animal models, and this might be a reason for its ability to facilitate improvement in male sexual functions in diabetic animals.[@ref10][@ref20][@ref23][@ref24] This is true when the mature green fruit of plantain is consumed at a moderate dose, as revealed in the current study, where consumption of *Musa paradisiaca* at a low dose led to improvement in the quantity and quality of spermatozoa in adult Wistar rats. Specifically, sperm motility was improved, with a very significant increase in life/death ratio of spermatozoa, and a progressivity very similar to that of the control.\n\nIn animals given a higher (double) dose of *Musa paradisiaca*, all the semen parameters investigated were adversely affected. Both the sperm count and percentage of morphologically normal sperm cells were significantly reduced (*P* \\< 0.05) compared with the control animals. A good percentage of spermatozoa observed had some of the abnormal morphological features listed below: Tail defects (short, irregular coiled or multiple tails); neck and middle piece defects (distended, irregular, bent middle piece, abnormally thin middle piece); and head defects (round head, small or large size, double or detached head). These findings would definitely negatively impart the normal sexual functions of the animals, thereby reducing their ability to fertilise an ovum successfully.\n\nAt a higher dose of plantain fruit, the percentage motility and Life/Death ratio did not significantly alter, but the forward directional movement was reduced when compared with the control. Nevertheless, motile abnormal sperm cells are not expected to perform the functions of successful fertilization, as do motile and morphologically normal sperm cells.\n\nConclusively, the mature green fruit of *Musa paradisiaca* improves semen parameters of Wistar rats at a moderate dose, making it a possible remedy for male reproductive dysfunctions in which sperm cells have been adversely impaired.\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nIrritable bowel syndrome (IBS) is a common functional gastrointestinal disorder, appearing as complex symptoms such as abdominal pain/discomfort \\[[@R1]\\]. As a chronic biopsychological disorder, IBS is characterized by altered bowel habits excluding organic disorders. Except for the astrointestinal motility abnormality and distorted visceral perception of sensation, IBS is also connected with several gastrointestinal and extraintestinal manifestations \\[[@R2]\\]. Although IBS is one of the most common disease, the etiology of the disease remains unknown. Several physical and psychological factors, such as stress, anxiety, and abnormal attitudes towards illness, are known to contribute to IBS\\'s pathogenesis \\[[@R3]\\]. Nowadays the worldwide prevalence of IBS ranges from 3% to 22% of the population \\[[@R4]\\] and in America, IBS affects 15% of the population \\[[@R5]\\]. Current evidence figures out that the diagnosis of IBS is still according to Rome IV and the actual quantity of the presence of such underlying cause of IBS, however, IBS\\'s pathogenesis remains unknown. Therefore, the objective of our research was to find novel pathogenesis so as to contribute to the treatment of IBS.\n\nIn the human body, aquaporins (AQPs) are considered play important roles in the water transport system \\[[@R6]\\]. There are currently 13 types of AQP, AQP0 through AQP12, which are expressed in various organs. Taking intestinal tract as example, AQP1, AQP3, AQP4, AQP7, AQP8, AQP9, and AQP10 has been found expressed in the colon, which affects fecal water content \\[[@R7]--[@R9]\\]. It has been improved that the functions of AQPs in the stomach and intestine physiology may involve water transfer, gastric juice secretion, barrier function, as well as absorption and secretion of water and even small solutes through the epithelium \\[[@R7], [@R10]\\]. As the gastrointestinal tract is the major organ for water transport which is only secondary to kidney, IBS might have some change in water transport through AQPs. It has been found that AQP1,AQP3,AQP8 were abundantly expressed in rat intestinal epithelial cells, and were also expressed in the human colon, as a result of this, we chose to study these three AQPs \\[[@R11]--[@R12]\\].\n\nThe nuclear factor-kappa B (NF-B), which can be activated by a variety of stimuli such as virus infection, activation of kinases and oncogenes, overproduction of cytokines, and dysregulation of cell surface receptors, plays an important role in the process of immune responses and inflammation \\[[@R13]\\]. It has been reported that the expression of AQP5 mRNA in the parotid gland was down-regulated after LPS inducing, which is mediated via transcription factors NF-kB and p-c-Jun/c-Fos \\[[@R14]\\]. Another research reported that NF-kB activation is of great importance for the down-regulation of AQP2 channel and vasopressin receptor expression during sepsis \\[[@R15]\\]. As a result of the mechanism of IBS include immune responses and inflammation, we speculated that NF-kB pathway might play an important role in IBS. Our research was to detect the expression of AQPs in IBS rat models and to detect the expression of AQPs with the inhibitor of NF-kB so as to elucidate the mechanism of IBS.\n\nRESULTS {#s2}\n=======\n\nThe change of the model {#s2_1}\n-----------------------\n\nAfter completing the models, we found that there was no change of behavior. Rats of IBS model group (including model group and inhibitor group before injection) expressed excessive reaction when they were frightened or they were performed by intragastric administration. There was also no change of their stool.\n\nModel authentication {#s2_2}\n--------------------\n\nThe rectum effusion amount of the model rats(including model group and inhibitor group before injection) (0.88 \u00b1 0.16 ml) were lower than the control group (1.42 \u00b1 0.11 ml), the difference had statistical signifcance (*P \\<* 0.01), which suggested that model building method was successful (Table [1](#T1){ref-type=\"table\"}).\n\n###### The water injection rate when AWR score was 3\n\n                             Control group (ml)   IBS Model group (ml)\n  -------------------------- -------------------- ----------------------\n  The water injection rate   1.467 \u00b1 0.054        0.833 \u00b1 0.143\n\n*P \\<* 0.05 compare with control group.\n\nThe expression of AQP1 in the colon by immunohistochemical technique {#s2_3}\n--------------------------------------------------------------------\n\nAQP1 cell immunochemical staining {#s2_4}\n---------------------------------\n\nAQP1 positive reaction material presents brown, the cell membrane and cytoplasm were dyed, and the negative control was not dyed (Figure [1](#F1){ref-type=\"fig\"}).\n\n![The expression of AQP1 in the colon by immunohistochemical technique (\\*400) AQP1 positive reaction material presents brown, the cell membrane and cytoplasm were dyed](oncotarget-08-47175-g001){#F1}\n\nThe expression of AQP1 comparing among the different group {#s2_5}\n----------------------------------------------------------\n\nThe positive cell numbers of model group was less than the control group (*P \\<* 0.05), the positive cell numbers of inhibitor group was more than model group (*P \\<* 0.05), and the positive cell numbers of inhibitor group was more than control group(*P \\<* 0.05) (Table [2](#T2){ref-type=\"table\"}).\n\n###### The expressions of AQP1, AQP3, AQP8 and NF-kB p65 in the colon by immunohistochemical technique\n\n             Control group   Model group            Inhibitor group\n  ---------- --------------- ---------------------- ---------------------\n  **AQP1**   1.348 \u00b1 0.190   0.3031 \u00b1 0.048 *^a^*   0.675 \u00b1 0.077 *^b^*\n  **AQP3**   2.011 \u00b1 0.178   0.957 \u00b1 0.138 *^a^*    1.859 \u00b1 0.176 *^b^*\n  **AQP8**   2.401 \u00b1 0.216   0.609 \u00b1 0.120 *^a^*    1.847 \u00b1 0.190*^b^*\n  **P65**    0.204 \u00b1 0.022   2.129 \u00b1 0.263 *^a^*    0.425 \u00b1 0.054 *^b^*\n\n*^a^P \\<* 0.05 compare with control group; *^b^P \\<* 0.05 compared with model group.\n\nThe expression of AQP3 in the colon by immunohistochemical technique {#s2_6}\n--------------------------------------------------------------------\n\nAQP3 cell immunochemical staining {#s2_7}\n---------------------------------\n\nAQP3 positive reaction material presents brown, the cell membrane and cytoplasm were dyed, and the negative control was not dyed. (Figure [2](#F2){ref-type=\"fig\"}).\n\n![The expression of AQP3 in the colon by immunohistochemical technique (\\*400) AQP3 positive reaction material presents brown, the cell membrane and cytoplasm were dyed](oncotarget-08-47175-g002){#F2}\n\nThe expression of AQP3 comparing among the different group {#s2_8}\n----------------------------------------------------------\n\nThe positive cell numbers of model group was less than the control group (*P \\<* 0.05), and the positive cell numbers of inhibitor group was more than model group (*P \\<* 0.05), nevertheless, there was no significant divergence between control group and inhibitor group (*p \\>* 0.05) (Table [2](#T2){ref-type=\"table\"}).\n\nThe expression of AQP8 in the colon by immunohistochemical technique {#s2_9}\n--------------------------------------------------------------------\n\nAQP8 cell immunochemical staining {#s2_10}\n---------------------------------\n\nAQP8 positive reaction material presents brown, the cell membrane and cytoplasm were dyed, and the negative control was not dyed. (Figure [3](#F3){ref-type=\"fig\"})\n\n![The expression of AQP8 in the colon by immunohistochemical technique (\\*400) AQP8 positive reaction material presents brown, the cell membrane and cytoplasm were dyed](oncotarget-08-47175-g003){#F3}\n\nThe expression of AQP8 comparing among the different group {#s2_11}\n----------------------------------------------------------\n\nThe positive cell numbers of model group was less than the control group (*P \\<* 0.05), the positive cell numbers of inhibitor group was more than model group (*P \\<* 0.05), and the positive cell numbers of inhibitor group was more than control group(*P \\<* 0.05). (Table [2](#T2){ref-type=\"table\"})\n\nThe expression of NF-kB p65 in the colon by immunohistochemical technique {#s2_12}\n-------------------------------------------------------------------------\n\nNF-kB p65 cell immunochemical staining {#s2_13}\n--------------------------------------\n\nNF-kB p65 positive reaction material presents brown, the cell nucleus was dyed, the cell membrane and cytoplasm were not dyed, and the negative control was not dyed (Figure [4](#F4){ref-type=\"fig\"}).\n\n![The expression of NF-\u03baB p65 in the colon by immunohistochemical technique (\\*400) NF-kB p65 positive reaction material presents brown, the cell nucleus was dyed, the cell membrane and cytoplasm were not dyed](oncotarget-08-47175-g004){#F4}\n\nThe expression of NF-kB p65 comparing among the different group {#s2_14}\n---------------------------------------------------------------\n\nThe positive cell numbers of model group was more than the control group (*P \\<* 0.05), and the positive cell numbers of inhibitor group was less than model group (*P \\<* 0.05), nevertheless, there was no significant divergence between control group and inhibitor group (*p \\>* 0.05) (Table [2](#T2){ref-type=\"table\"}).\n\nThe expressions of AQP1, AQP3, AQP8 in the colon by PCR technique {#s2_15}\n-----------------------------------------------------------------\n\nThe expressions of AQP1, AQP3 and AQP8 mRNA of model group were less than the control group (*P \\<* 0.05), and the expressions of inhibitor group were more than model group (*P \\<* 0.05), however, there was no difference between control group and inhibitor group (*p \\>* 0.05) (Table [3](#T3){ref-type=\"table\"}).\n\n###### The expressions of AQP1, AQP3, AQP8 in the colon by RT-PCR technique\n\n         Control group    Model group           Inhibitor group\n  ------ ---------------- --------------------- ---------------------\n  AQP1   27.720 \u00b1 0.841   24.511 \u00b1 0.472*^a^*   26.776 \u00b1 0.695*^b^*\n  AQP3   24.163 \u00b1 1.027   20.814 \u00b1 0.569*^a^*   23.492 \u00b1 0.853*^b^*\n  AQP8   27.387 \u00b1 0.850   23.438 \u00b1 0.548*^a^*   26.007 \u00b1 0.781*^b^*\n\n^a^*P* \\< 0.05 compare with control group; ^b^*P* \\< 0.05 compared with model group.\n\nThe expressions of AQP1, AQP3, AQP8 and NF-kB p65 in the colon by Western blot technique {#s2_16}\n----------------------------------------------------------------------------------------\n\nThe expressions of AQP1, AQP3 and AQP8 protein of model group were less than the control group (*P \\<* 0.05), and the expressions of inhibitor group were more than model group (*P \\<* 0.05), nevertheless, there was no significant divergence between control group and inhibitor group (*p \\>* 0.05). The expressions of NF-kB p65 protein of model group were more than the control group (*P \\<* 0.05), the expressions of inhibitor group were less than model group (*P \\<* 0.05), and the expression of inhibitor group was more than control group(*P \\<* 0.05). (Figure [5](#F5){ref-type=\"fig\"}) (Table [4](#T4){ref-type=\"table\"})\n\n![The expressions of AQP1, AQP3, AQP8 and NF-\u03baB p65 in the colon by Western blot technique\\\nThe expressions of AQP1, AQP3 and AQP8 protein of model group were less than the control group (*p* \\< 0.05), and the expressions of inhibitor group were more than model group (*p* \\< 0.05), The expressions of NF-\u03baB p65 protein of model group were more than the control group (*p* \\< 0.05), and the expressions of inhibitor group were less than model group (*p* \\< 0.05).](oncotarget-08-47175-g005){#F5}\n\n###### The expressions of AQP1, AQP3, AQP8 and NF-\u03baB p65 in the colon by Western bolt technique\n\n             Control group    Model group           Inhibitor group\n  ---------- ---------------- --------------------- ---------------------\n  **AQP1**   29.325 \u00b1 0.305   19.708 \u00b1 0.744*^a^*   27.310 \u00b1 0.985*^b^*\n  **AQP3**   29.263 \u00b1 0.318   19.557 \u00b1 0.944*^a^*   27.633 \u00b1 0.683*^b^*\n  **AQP8**   28.647 \u00b1 0.355   21.105 \u00b1 0.621*^a^*   27.885 \u00b1 1.716*^b^*\n  **P65**    13.210 \u00b1 0.933   56.818 \u00b1 2.834*^a^*   21.888 \u00b1 1.237*^b^*\n\n^a^*P* \\< 0.05 compare with control group; ^b^*P* \\< 0.05 compared with model group\n\nDISCUSSION {#s3}\n==========\n\nAccording to Rome IV criteria, the diagnostic criteria for Irritable bowel syndrome is defined as recurrent abdominal pain on average at least 1 day/week in the last 3 months, associated with 2 or more of the following:related to defecation, associated with a change in frequency or form (appearance) of stool \\[[@R16]\\]. IBS is a chronic biopsychological disorder which has complex symptoms appearing as altered bowel habits without organic pathology\\[[@R17]\\]. As we know, IBS is so common that it affects nearly 10--20% of adolescents and adults in western societies and is also the frequentest cause of children with recurrent abdominal pain \\[[@R18], [@R19]\\]. Several hypotheses have been proposed that the pathogenesis of IBS includes altered gut microbiota, visceral hypersensitivity, dysmotility, gastrointestinal infection and infestation, dysregulation of brain-gut axis, psychological and genetic factors \\[[@R20]\\]. Despite the growing body of literature, IBS pathophysiology remains poorly understood.\n\nNowadays, at least 13 types of AQPs (AQP0--12) have been identified in mammals. AQP0, AQP1, AQP2, AQP4, AQP5, AQP6 and AQP8 are reported be selectively permeable to water; and AQP3, AQP7, AQP9, AQP10 are permeable to water, glycerol and urea \\[[@R21]\\]. In mammals, AQPs are widely present in the digestive tract, including salivary gland, esophagus, stomach, small and large intestines, liver, gallbladder, bile duct and pancreas\\[[@R22]\\]. Currently several researches reported that AQPs play important roles in gastrointestinal diseases, however, there is few research reported about IBS. It isreported that several AQP types are found in gastrointestinal epithelia, with AQP1, 3, 7, 10 and 11 being the most abundantly expressed in the whole digest tract, and AQP4 and 8 are expressed selectively in the stomach and colon \\[[@R23]\\]. Several studies revealed that AQP1, AQP3 and AQP8 are closely connected with water transportation in colon \\[[@R23], [@R24]\\]. Another research showed that AQP-1 increases osmotic water permeability and locally facilitates the rapid, trans-membrane flux of water in plasma membrane blebs \\[[@R25]\\]. It is also reported that in the mediation of AQP3, the fecal water content in the colon is controlled by the transport of water from the luminal side to the vascular side\\[[@R26]\\]. The knockdown of AQP3 was reported to be connected with a decreasing of the expression of Claudin-1 and Occludin and the increasing of bacteria translocation, which revealed that AQP3 was associated with intestinal permeability. In our research, we found that the expressions of AQP1, AQP3 and AQP8 of model group were down-regulated comparing with the control group which suggested that liquid water metabolic abnormalities might be one of the mechanisms of IBS connecting with AQPs. On the other hand, as the connection of AQPs and Occludin expression, it is also revealed that tight junction or intestine permeability might be another mechanism of IBS regulated by AQPs.\n\nNF-kB protein family include RelA (p65), Rel (c-Rel), RelB, NF-kB1 (p50), and NF-kB2 (p52), composed by a conserved Rel homology domain which is responsible for dimerization, nuclear localization and DNA-binding \\[[@R27]\\]. NF-kB pathway can be activated through two distinct signaling pathways: the classical pathway and the alternative pathway, and in the classical pathway, degradation of IkBa by IKKb activation triggers the translocation of various heterodimers, predominantly p65/p50, to the nucleus \\[[@R28]\\]. The heterodimer of NF-\u03baB p65 and p50 subunit is a common form of NF-kB. NF-kB was also reported as an important pathway in the chronic intestinal inflammation \\[[@R29]\\]. That periostin mediates intestinal inflammation via the activation of NF-\u03baB pathway, which suggests that periostin may be a potential therapeutic target for inflammatory bowel disease \\[[@R30]\\] has been reported. One latest study showed that the activation of TLR4 by NCI increase CBS expression, which is mediated by the NF-\u03baB pathway, final lead to visceral hypersensitivity \\[[@R31]\\]. In our research, we found that NF-\u03baB p65 of model group was up-regulated, and after the inhibitor of NF-\u03baB pathway, the expression of NF-\u03baB p65 was down-regulated again, which revealed NF-\u03baB pathway played a role in the mechanism of IBS. And as NF-kB is a main regulator of inducing several genes including inflammatory and immune response, we suspected that the mechanism of NF-\u03baB pathway for IBS might involve in the immune, inflammation and visceral hypersensitivity.\n\nUntil now we can\\'t search the study about IBS connected with AQP and NF-\u03baB pathway. There was a study found that once NF-\u03baB activated by a hypertonic medium, can decreased AQP-2 mRNA and protein expression through binding of NF-\u03baB complexes to specific kB elements of the AQP-2 promoter \\[[@R32]\\]. That IL-1\u03b2 induces expression of AQP4 through a NF-kB pathway without involvement of denove protein synthesis in rat astrocytes has been reported \\[[@R33]\\]. Another study reported that lack of AQP5 showed declined activation of mitogen-activated protein kinase and NF-\u03baB pathways in lungs before and after Pseudomonas aeruginosa(PA) infection \\[[@R34]\\]. In our research, the expression of AQP1, AQP3 and AQP8 was up-regulated after the injection of inhibitor of NF-\u03baB pathway. Thus, our research revealed that liquid water metabolic abnormalities and intestine permeability might be the mechanism of IBS by regulating AQP1, AQP3 and AQP8 via NF-\u03baB pathway. Worth attention, we found that the IBS model rats could not return to the normal state of the control group by inhibiting NF-\u03baB.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nSubjects {#s4_1}\n--------\n\nThere were 18 adult female SD big rats, the weight of every rat was about 200 g. Feeding environment was provided by experimental animal center of Zhejiang Chinese Medical University. 18 SD rats were divided randomly into three groups. The control group was 6, model group was 6, and inhibitor group was 6. They were put into the environment where the temperature was 22--24\u00b0C, the humidity \\< 60%, the noise \\< 50 db.\n\nExperimental procedure and methods {#s4_2}\n----------------------------------\n\nGroup {#s4_3}\n-----\n\nControl group {#s4_4}\n-------------\n\nThey were normal rats. The control group was injected by 1 ml saline for control. After 2 weeks of normal eating and drinking, we observed the condition of rats. Then the rats' visceral sensitivity was evaluated by abdominal withdraw reaction. The AWR scoring criteria are shown in Table [5](#T5){ref-type=\"table\"} \\[[@R35]\\].\n\n###### Abdominal withdrawal reflex (AWR) scoring criteria\n\n  Score 0       No behavioral response to colorectal distension (CRD)\n  ------------- --------------------------------------------------------------------------------------------------------------------------------------------\n  **Score 1**   Immobile during distension of colorectum (CR) and occasional appearance of brief head motion after a pause at the onset of the stimulation\n  **Score 2**   A mild contraction of abdominal muscles, but no lifting of abdomen off the plattorm\n  **Score 3**   A strong contraction of abdominal muscles and lifting of abdomen off the platform, no lifting of pelvic structure off the platform\n  **Score 4**   Arching body and lifting of pelvic structure and scrotum\n\nModel group {#s4_5}\n-----------\n\nThe conditioned stimulus was camphor ball special odor. The unconditioned stimulus was rectal distention pressure (\\> 60 mmHg (1 mmHg = 0.133 kPa)) combining with extremities constraint. Rats were put into the cage with camphor ball in it, we fixed the extremities and trunk of the rats for 45 min. At this time inserted catheter into rectum. The distance from air ballon distal end to anal was about 1 cm. The catheter was fixed at the root of the tail. The balloon volume was 1.6 ml (hydrostatic pressure in Ballon\\'s space \\> 60 mmHg) and lasted for 60 s, intermittent exhaust 3 mins and filled gas 10 times once. This was a stress process. There was one process in the first day, the same process was performed on the second day at the same time. The conditioned stimulus Without the unconditioned stimulus was done on the fourth day. The completed process was repeated once on the fifth day. The process same as the fourth day was performed on the sixth day. The conditioned stimulus was performed on the eighth day \\[[@R36], [@R37]\\]. Then the rats' visceral sensitivity was evaluated by abdominal withdraw reaction.\n\nInhibitor group {#s4_6}\n---------------\n\nThe IBS model rats were injected into the abdominal cavity with the inhibitor of NF-kB (PDTC, 50 mg/kg/d) once after completing the models.\n\nModel authentication {#s4_7}\n--------------------\n\nVisceral sensitivity was evaluated by abdominal withdraw reaction (AWR). 8 F urethral catheter which was lubricated by liquid paraffin was inserted per anum. The distance from air ballon distal end to anal was about 1 cm, and it was fixed at the root of the tail. The rats were put on the platform, after they accommodated the environment, we gradually affused water into sacculus, and recorded the water injection rate when the rats raised the abdomen and made the back like a bow when AWR score was 3. Rectal distention lasted for 30 s every time, and repeated 3 times. And then we took the mean number.\n\nExperimental sample {#s4_8}\n-------------------\n\nAfter a laparotomy incision, a portion of the colon was removed and placed in an oxygenated Tyrode\\'s solution. A segment of 2 cm length colon were mounted in a 10ml organ bath containing Tyrode\\'s solution that was bubbled with a 95% O2 and 5% CO2 mixture, and the temperature was held at 37\u00b0C.\n\nExperimental procedure (immunohistochemical technique) {#s4_9}\n------------------------------------------------------\n\nTwo footwork {#s4_10}\n------------\n\nWe dropwised 3% hydrogen dioxide on tissue away from light and incubated for 15min. And then the tissue was flushed by distilled water, and we put the chips into PBS balanced solution, and soaked for 5 min, for 3 times. AG dark was repaired. We dropped 50--100 ml antibody fluid on the tissue, and incubated for 30 min in the ambient temperature. Then we washed chips with PBS, and soaked the chips in PBS balanced solution for 4 min 3 times. Then we dropped appreciable proportion diluted biotin labeling antibody (1%BSA-PBS to dilute), and incubated for 30 min in the ambient temperature. After that we washed the chips with PBS, and soaked them in the PBS balanced solution for 5 min 3 times. Then we dropped 50--100 ml developer DAB fluid, and incubated for 5--20 min, after coloration completely, and then the tissue was washed by distilled water. The tissue was dewatered by 85%, 90%, 95%, 100%, 100%, 100% alcohol. Then we put the chips into xylene for 5 min, 3 times.\n\nNegative control {#s4_11}\n----------------\n\nReplaced one antibody with PBS, the consequence was negative.\n\nAnalytical method {#s4_12}\n-----------------\n\nWe used the computer image analysis software (the Carl Zeiss of the Imaging Systems of the Carl Zeiss company) to analyze images. The images were put under 40 times object glass. We found the typical places, and took 10 high power campus visualis (\u00d7400) successively. Then we analyzed the masculine expression by quantitative analysis and calculated photodensity.\n\nExperimental procedure (RT-PCR technique) {#s4_13}\n-----------------------------------------\n\nTotal RNA was extracted from duodenal tissues with Trizol (Invitrogen, Gaithersburg, MD, USA) using the one-step method. After purification, RNA concentration was analyzed using Nanodrop (Nanodrop Technologies, Wilmington, DE) and quality testing was conducted using BioAnalyzer (Agilent Technologies, Palo Alto, CA). After preparing total RNA, we did primer test and sample assay. Then we performed reverse transcription. After creating and setting up a plate document, we prepared the PCR reaction plate, after that we run the PCR reaction plate, then we analyzed the result. he main procedure was high-throughput sequencing on independent samples. Small RNA was purified from total RNA to enrich molecules in the range of 16--30 nt, and then 3\u2032 and 5\u2032 linker sequences were attached, and SuperScript II reverse transcriptase was used to synthesize cDNA. PCR amplification was conducted.\n\nExperimental procedure (Western blot technique) {#s4_14}\n-----------------------------------------------\n\nProtein measurements were performed on the prepared cell extracts and 7.5% SDS gel was also prepared. After the polymerization of the gel, the cell extracts were diluted in the sample buffer and boiled at 95\u00b0C for 5 min. With the samples executed in the gel, proteins were transferred into membranes by the use of semidry transfer method. The transfer will be realized at 100 V within 1 h. It was held in PBS solution (blocking solution) with 5% milk powder and 0.1% tween 20. The membrane will be incubated with primer antibodies overnight. And the membrane was washed 3 times using PBS solution with 0.1% Tween 20 for 5 min and second antibody application was performed. After membrane incubation with rabbit second antibody for 1 h, it was again washed 3 times using PBS solution with 0.1% Tween 20 for 5 min. ECL solution was used with the purpose of viewing and filming proteins in the membrane. Horseradish peroxidase enzyme linked to the second antibody catalyzes Lumigan PS-3 substrate in ECL solution.\n\nStatistical analysis {#s4_15}\n--------------------\n\nThe data was demonstrated as\\`x \u00b1 SE. We used SPSS 17.0 software as statistical method. evene\\'s test for equality of variances, differences between three groups were compared using one-factor analysis of variance (ANOVA), tatistical significance was taken as *P \\<* 0.05.\n\nCONCLUSIONS {#s5}\n===========\n\nAs a chronic biopsychological disorder, IBS is characterized by altered bowel habits excluding organic disorders. However, the real mechanism of IBS remains unknown. Our research showed that liquid water metabolic abnormalities and intestine permeability Alteration might be the mechanism of IBS by down-regulating AQP1, AQP3 and AQP8 via NF-\u03baB pathway.\n\n**Authors' contributions**\n\nGuanqun Chao did the research and wrote the article. Shuo Zhang guided the experiment.\n\nThis work was partially supported by the Fund of Medicine and health science and technology plan projects in Zhejiang province (2017179719); Traditional Chinese medicine science and technology plan of Zhejiang province (2017ZA089); Traditional Chinese medicine science and technology plan projects of Zhejiang province (2016ZB071); and National Natural Science Foundation of China (81573760); and Traditional Chinese medicine science and technology plan projects of Zhejiang province (2015ZZ012); and medical health platform plan projects of Zhejiang province (2015RCA020); Zhejiang Provincial Natural Science Foundation of China (LY16H030010).\n\n**CONFLICTS OF INTEREST**\n\nNone.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nTarsal coalitions can be osseous (synostosis) or nonosseous (synchondrosis or syndesmosis) \\[[@B1]\\]. Coalitions are most often congenital and are believed to result from a failure of differentiation and segmentation, probably of autosomal dominant inheritance \\[[@B2]\\]. The true incidence of tarsal coalitions is not known because the number of asymptomatic coalitions has never been studied extensively \\[[@B3]\\], but is estimated to be less than 1% of the population \\[[@B3], [@B4]\\]. Coalitions are bilateral in more than 50% of cases \\[[@B4]\\], even 80% according to Leonard \\[[@B2]\\]. The most common types are talocalcaneal and calcaneonavicular coalitions \\[[@B3]\\], which represent more than 90% of all tarsal coalitions \\[[@B4]\\]. Multiple coalitions have also been described \\[[@B5]--[@B7]\\]. Cubonavicular coalition represents an infrequent type, with only 38 cases described in the literature (Tables [1](#tab1){ref-type=\"table\"} and [2](#tab2){ref-type=\"table\"}), among which only 10 cases were osseous coalitions ([Table 1](#tab1){ref-type=\"table\"}).\n\nWe report a case of cubonavicular complete coalition, associated with other incomplete midtarsal coalitions, and with degenerative changes of the tarsometatarsal joint. This is the first case in the literature which associates a complete osseous cubonavicular coalition with tarsometatarsal arthritis.\n\n2. Case Presentation {#sec2}\n====================\n\nA 26-years-old man presented to the senior author for evaluation of bilateral heel pain, more severe on the right foot, for five months. He played hockey and tennis as leisure sports and worked as a dairy-cheese maker.\n\nThe physical examination demonstrated normal alignment of the lower extremities, straight hindfoot with painless limited motion of the subtalar joint bilaterally. Elective pain was present on both sides at the insertion of the Achilles tendon. The first tarsometatarsal (TMT1) joint was hypermobile and painless bilaterally.\n\nStandard X-rays showed insertional Achilles calcific spurs, an incidental finding of osseous cubonavicular coalition on the right foot (Figures [1(a)](#fig1){ref-type=\"fig\"} and [1(b)](#fig1){ref-type=\"fig\"}), and irregularity of the cubonavicular and calcaneonavicular articular surfaces on the left foot ([Figure 2](#fig2){ref-type=\"fig\"}). Arthritic changes were visible on the dorsal aspect of the first tarsometatarsal joint on both feet ([Figure 1(c)](#fig1){ref-type=\"fig\"}). Medial foot arch was normal with a Djian-Annonier angle of 123\u00b0 on both sides (angle defined between a line tangent to the inferior surface of the calcaneus and a line between the inferior point of talonavicular joint and the inferior point of the medial sesamoid, with normal value defined between 120 and 128\u00b0 \\[[@B8]\\]).\n\nMRI of the right foot showed insertional Achilles tendinosis and an osseous coalition between the navicular and the cuboid ([Figure 3](#fig3){ref-type=\"fig\"}), as well as arthritic changes of the TMT1 ([Figure 4(b)](#fig4){ref-type=\"fig\"}). Computed tomography of both feet was also performed and confirmed the osseous cubonavicular coalition on the right foot ([Figure 5](#fig5){ref-type=\"fig\"}), associated with a partial osseous coalition between the cuboid and the third cuneiform and bone irregularities between the first and second cuneiforms suggesting a nonosseous coalition ([Figure 6](#fig6){ref-type=\"fig\"}), as well as arthritis in the tarsometatarsal joint ([Figure 4(a)](#fig4){ref-type=\"fig\"}). On the left foot, a nonosseous calcaneonavicular associated with a cubonavicular coalitions was diagnosed ([Figure 7](#fig7){ref-type=\"fig\"}).\n\nAs the patient was not symptomatic of this midtarsal arthritis and of his rigid hindfoot, no surgical treatment was proposed. Insertional Achilles tendinosis was treated with physical therapy and foot orthoses with success.\n\n3. Discussion {#sec3}\n=============\n\nClassical tarsal coalitions can become symptomatic, usually during childhood or adolescence, causing stiffness and pain \\[[@B3], [@B9]\\]. Patients present typically with a history of ankle injury that is slow to resolve or diffuse pain exacerbated with activity, associated with difficulty to accommodate to uneven ground and repetitive ankle sprains \\[[@B3], [@B10]\\]. Clinical examination usually shows a diminished range of motion of the subtalar joint, rigid flat foot, tenderness over the site of the coalition, and possible peroneal spasm \\[[@B3], [@B11]\\]. Cubonavicular coalition can produce various clinical presentations, as summarized in Tables [1](#tab1){ref-type=\"table\"} and [2](#tab2){ref-type=\"table\"}. Nonoperative treatment represents the first line therapy \\[[@B1], [@B3], [@B11]\\]. Surgical treatment is considered in case of failure of conservative measures, with coalition resection and interposition, or arthrodesis \\[[@B1], [@B3], [@B10], [@B11]\\]. Degenerative changes related to tarsal coalitions were described regarding arthritis involving the joint affected by the coalition, especially for talocalcaneal coalitions \\[[@B12], [@B13]\\]. In the case of symptomatic calcaneonavicular or talocalcaneal coalition with severe degenerative changes of these joints, triple arthrodesis is recommended \\[[@B1], [@B3], [@B11], [@B14]\\].\n\nCohen et al. studied an adult population with incomplete calcaneonavicular coalitions and reported approximately 75% of degenerative changes involving adjacent joints (naviculocuneiform, subtalar and talonavicular joints) on preoperative x-rays \\[[@B15]\\]. Regarding cubonavicular coalitions, only two authors, to our knowledge, mentioned degenerative changes of adjacent joints, both associated with a nonosseous coalition: talonavicular arthritis was reported by Ehredt et al. \\[[@B16]\\], and arthritis of the tarsometatarsal joint was mentioned by Sarage et al. \\[[@B17]\\]. However, in this latter case, the authors report neither x-rays nor other details regarding these arthritic lesions. Among the few described cases of osseous coalition ([Table 1](#tab1){ref-type=\"table\"}), none was associated with arthritic changes. We described therefore the first case of a cubonavicular osseous coalition associated with other incomplete coalitions and midfoot arthritis.\n\nConsidering the relationship between the subtalar and the midtarsal joints during the stance phase, the axis of the talonavicular and the calcaneocuboid joints become parallel with pronation \\[[@B9], [@B18]\\]; thereby, the cuboid and navicular are independent from one another \\[[@B18]\\]. According to Cavallaro and Hadden \\[[@B18]\\], restriction of the normal movement between the cuboid and navicular could have an irritative effect by interfering with the physiological mobility of the talonavicular, calcaneocuboid, and subtalar joints and would result in stabilization of the subtalar joint and overloading of midtarsal joints. We can postulate that, if the subtalar become more rigid because of a cubonavicular coalition, accommodation on the uneven floor is less effective and that mechanical stress could be transmitted to tarsometatarsal joint and causing osteoarthritis.\n\nPain associated with a tarsal coalition is suspected to be related to an incomplete coalition, where the union is particularly prone to motion and strain \\[[@B15], [@B19]\\]. However, a complete osseous coalition can also be symptomatic \\[[@B20]\\]. In our case, the patient was asymptomatic, despite early degenerative changes of the tarsometatarsal joint, and the cubonavicular coalition was an incidental finding. The review of the literature shows three cases of asymptomatic cubonavicular coalitions, found out by radiological assessment following other injuries: violence to both feet in one case \\[[@B21]\\], lateral ankle sprain in another \\[[@B22]\\], and tibial anterior tendon rupture for the third \\[[@B23]\\]. A fourth case of incidental finding was described by Chu \\[[@B24]\\], after a pilon fracture, but it is unclear whether or not the patient was symptomatic from the coalition. Interestingly, these three asymptomatic patients \\[[@B21]--[@B23]\\] were in their forties (mean and median 45 years old), while symptomatic patients \\[[@B16]--[@B20], [@B25]--[@B35]\\] were younger (mean age 23 years old, median 17 years old). The analysis of the cases of the literature for cubonavicular coalition revealed that symptoms were more frequent for nonosseous (14/16, 88%) than osseous (7/9 cases, 78%) union but the small number of cases does not allow any assertion.\n\nThis case of a complete osseous cubonavicular coalition with early osteoarthritic changes shows that clinical and radiological presentation of such coalition could be very different among patients. Our patient was asymptomatic, despite multiple tarsal coalitions and degenerative signs of the tarsometatarsal joint, while other patients report daily disabling pain, whether the coalition is bony or fibrous. In any case, surgical treatment should only be proposed if the coalition is symptomatic and after failure of conservative treatment.\n\nDisclosure\n==========\n\nThe research was performed as part of the employment of the authors at the H\u00f4pital Intercantonal de la Broye (HIB, Payerne, Switzerland).\n\nConflicts of Interest\n=====================\n\nThe authors declare that there is no conflict of interest regarding the publication of this article.\n\n![Frontal (a) and oblique (b) x-ray image of the right foot show an osseous cubonavicular coalition (arrowheads) with an absence of joint space. Lateral (c) x-ray image of the right foot shows degenerative changes with a spiky osteophytic deformation on the dorsal aspect of the first tarsometatarsal joint and a joint line irregularity (side shot).](CRIOR2020-8850768.001){#fig1}\n\n![Oblique x-ray image of the left foot shows irregularity of the cubonavicular and calcaneonavicular articular surfaces with sclerotic changes.](CRIOR2020-8850768.002){#fig2}\n\n![MRI of the right foot. Transversal T1-weighted (a) and PD-Fat Sat-weighted (b) images show a coalition between the cuboid and the navicular with a perfect continuity of the bone marrow and joint surfaces, without any bone oedema.](CRIOR2020-8850768.003){#fig3}\n\n![(a) Volume rendered CT image of the medial aspect of the right foot shows the same degenerative disorders as those shown in x-ray images, with a spiky osteophytic deformation on the dorsal aspect of the first tarsometatarsal joint. (b) Sagittal T1-weighted MRI image of the right foot shows a dorsal joint space narrowing (arrowheads), hypo T1 subchondral sclerosis and osteophyte formation (arrow).](CRIOR2020-8850768.004){#fig4}\n\n![Osseous cubonavicular coalition of the right foot. Dorsolateral (a) and plantar (b) volume rendered CT images show the cuboidonavicular complex and its close relationship with the peripheral bones of the hindfoot and forefoot. Oblique multiplanar reconstruction CT image (c) shows as well as MRI the perfect continuity of the bone marrow and joint surfaces.](CRIOR2020-8850768.005){#fig5}\n\n![Oblique multiplanar reconstruction CT image of the right foot shows osseous cubonavicular coalition, associated with a partial osseous coalition between the cuboid and the third cuneiform with a focal lack of subchondral bone plate (black arrowheads) and bone irregularities between the first and second cuneiforms suggesting a nonosseous coalition (white arrowheads).](CRIOR2020-8850768.006){#fig6}\n\n![Dorsolateral volume rendered CT image of the left foot (a) shows hypertrophic arthritis changes between calcaneus and navicular due to nonosseous calcaneonavicular coalition. Oblique multiplanar reformatted CT images of the left foot (b-c) show a nonosseous calcaneonavicular coalition associated with a cuboidonavicular coalition (arrowheads), with degenerative changes between these tarsal bones.](CRIOR2020-8850768.007){#fig7}\n\n###### \n\nReview of the published cases of osseous cubonavicular coalition.\n\n  Authors                  Year   No. of cases   Sex   Age   Side   Nature of coalition   Symptoms                                          Physical examination\n  ------------------------ ------ -------------- ----- ----- ------ --------------------- ------------------------------------------------- ------------------------------------------------------------------------------\n  Waugh \\[[@B19]\\]         1957   1              M     15    L+R    O (L), NO (R)         Left foot pain with activity                      Left peroneal spasm\n  Del Sel \\[[@B21]\\]       1959   1              M     43    L+R    O (L), NO (R)         None                                              Cavus deformity, slight inversion/eversion\n  Cavallaro \\[[@B18]\\]     1978   1              F     12    L+R    O (L+R)               Bilateral ankle pain with activity                Pain and stiffness of subtalar joint, painful sinus tarsi\n  O\\'Neill \\[[@B36]\\]      1989   1              F     15    R      O                     n.a.                                              Limitation of inversion/eversion\n  Williamson \\[[@B20]\\]    1992   1              M     14    L+R    O (L+R)               Bilateral foot pain with activity                 Bilateral hindfoot valgus, decreased subtalar motion, peroneal spasm\n  Talkhani \\[[@B25]\\]      1999   1              M     42    L      O                     Midfoot pain with walking and morning stiffness   Diminished midfoot movements\n  Piqueres \\[[@B26]\\]      2002   1              M     14    L+R    O (L), NO (R)         Left midtarsal pain with weight bearing           Restriction of plantar flexion and eversion, valgus rearfoot, flat left foot\n  Johnson \\[[@B27]\\]       2005   1              M     15    L      O                     Midfoot and ankle pain with activity              Decreased subtalar and transverse tarsal motion, fixed pes planus\n  Prado \\[[@B28]\\]         2010   1              F     9     R      O                     Foot pain with exercises                          Limitation of mobility of midtarsal joint\n  Garc\u00eda-Mata \\[[@B22]\\]   2011   1              M     45    R      O                     None                                              Pain over cuboid\n\nM: male; F: female; L: left; R: right; O: osseous; NO: nonosseous; n.a.: nonavailable.\n\n###### \n\nReview of the published cases of nonosseous or undefined cubonavicular coalition.\n\n  Authors                   Year    No. of cases   Sex    Age    Side       Nature of coalition                             Symptoms                                              Physical examination\n  ------------------------- ------- -------------- ------ ------ ---------- ----------------------------------------------- ----------------------------------------------------- -----------------------------------------------------------------\n  Cowell \\[[@B37]\\]         1982    1              n.a.   n.a.   L          NO                                              Pain with sport activities                            Limited motion without details\n  Feliu \\[[@B29]\\]          1991    1              M      24     L          NO                                              Spontaneous pain dorsum of the foot                   Normal\n  Palladino \\[[@B30]\\]      1991    1              M      13     L+R        NO (L+R)                                        Rearfoot pain with activity (right more than left)    Bilateral peroneal spasm, rigid subtalar joint, pes planovalgus\n  Newman \\[[@B38]\\]         2000    1              F      10     R          NO                                              n.a.                                                  n.a.\n  Hounshell \\[[@B31]\\]      2011    1              F      37     L          NO                                              Persistent foot pain 8 months after a sprain          Decreased subtalar motion, painful sinus tarsi\n  Sarage \\[[@B17]\\]         2012    4              F      15     L          NO                                              Foot pain for 4 months with activity                  Cuboid and navicular pain, decreased subtalar motion\n                                    M              16     L+R    NO (L+R)   Bilateral dorsolateral midfoot pain             n.a.                                                  \n                                    M              35     L      NO         Dorsolateral midfoot pain for 1 year            n.a.                                                  \n                                    F              18     R      NO         Foot and ankle pain for 4 years with activity   Painful sinus tarsi, normal range of motion           \n  De Keyzer \\[[@B32]\\]      2013    1              F      40     R          NO                                              Mechanical pain for some duration                     Decreased subtalar motion, pes planovalgus\n  Lawrence \\[[@B33]\\]       2014    2              M      40     R          NO                                              Chronic midfoot pain                                  n.a.\n                                                   M      44     R          NO                                              Vague ankle pain                                      n.a.\n  Awan \\[[@B34]\\]           2015    1              M      17     R          NO                                              Foot pain for 6 months with activity                  Tenderness over tarsonavicular region\n  Kamiya \\[[@B35]\\]         2015    1              F      14     R          NO                                              Midfoot pain exacerbated with activity                Normal\n  Chu \\[[@B24]\\]            2017    1              F      34     R          NO                                              Not clear (x-rays for distal tibia fracture)          Symptoms of post-traumatic arthritis (localization n.a.)\n  Berger-Groch \\[[@B23]\\]   2018    1              M      47     L          NO                                              None                                                  Pes valgus and abductus\n  Ehredt \\[[@B16]\\]         2020    1              M      34     L          NO                                              Dorsolateral midfoot pain for 2 years with activity   Pes planovalgus, decreased subtalar motion\n  Harris \\[[@B39]\\]         1965    1              n.a.   n.a.   n.a.       n.a.                                            n.a.                                                  n.a.\n  Rankin \\[[@B40]\\]         1974    1              n.a.   n.a.   n.a.       n.a.                                            n.a.                                                  n.a.\n  Stormont \\[[@B4]\\]        1983    1              M      26     L          n.a.\u2020                                           n.a.                                                  n.a.\n  Sarrafian \\[[@B41]\\]      2011\u2021   8              n.a.   n.a.   n.a.       n.a.                                            n.a.                                                  n.a.\n\n^\u2020^naviculo-cubo-third cuneiform (x-rays n.a.). ^\u2021^4 cases reported by Gruber in 1871, 3 cases reported by Pfitzner in 1896, 1 case reported by Cruveilhier (1829-1835). M: male; F: female; L: left; R: right; NO: nonosseous; n.a.: nonavailable.\n\n[^1]: Academic Editor: Kaan Erler\n"}
+{"text": "Background\n==========\n\nThe amount of research utilizing health information has increased dramatically over the last ten years. Single, time-limited studies with tightly-defined research questions are giving way to programs of research that rely upon the systematic prospective collection of data in registries and biobanks for subsequent use in multiple projects to answer as yet unknown research questions. Many institutions have extensive biobank holdings collected over a number of years for clinical and teaching purposes, but are uncertain as to the proper circumstances in which to permit research uses of these samples. Research Ethics Boards (REBs) in Canada and elsewhere in the world are grappling with these issues, but have not received clear guidance regarding their role in the creation of and access to registries and biobanks. Historically, REBs have played an active role regarding specific project-by-project requests exclusively, thus not engaging in some of the larger issues concerning the creation and research uses of registries and biobanks. Indeed, some may not even be aware of the extent of registry and biobank holdings within their institutions. REBs we interviewed expressed concern and confusion both as to the handling of specific projects emanating from registries and biobanks, and to the broader issues surrounding them.\n\nIn this paper we outline variation in the responses of a wide sampling of REBs across Canada to a series of questions regarding the creation and use of registry and biobank information for health research purposes, and we note the heightened tension surrounding biobanks. We then discuss the implications of our findings for the development of policy and legislation.\n\nMethods\n=======\n\nDesign & Sample\n---------------\n\nWe approached the Chairs of 34 REBs affiliated with Faculties of Medicine in Canadian universities and requested interviews with them and/or with their REB Administrators. They were also invited to include other REB members in the interview. The interview was to be 90-minutes face-to-face. Ethics approval was obtained from Research Ethics Boards at the universities of McMaster, Dalhousie, and Montreal and at St. Joseph Healthcare, Hamilton, Ontario.\n\nProcedure\n---------\n\nInterviews consisted of structured questions dealing with diabetes-related scenarios, with open-ended responses and probing for rationales. The two scenarios discussed in this paper involved the development of a diabetes registry using clinical encounter data across several physicians\\' practices, and the addition of biological samples to the registry to create a biobank \\[[@B1]\\]. All interviews but one were audio-recorded.\n\nScenarios\n---------\n\nThe registry scenario involved the construction of a multi-centre multi-jurisdictional diabetes registry to serve as an ongoing resource for conducting epidemiologic and process-outcome studies. No specific research questions were identified. Instead, this was intended to provide a general research platform for future epidemiologic studies. The plan was to collect data through physician practices. At a regular patient visit, the physician would complete a duplicate encounter form. One copy would go in the patient\\'s file; the second would be supplied to the research associate at the principal investigator\\'s office, who was then to remove any direct identifiers and forward the data to the central registry. This registry was intended to be updated during routine patient care visits, and was to continue indefinitely. The research associate would hold the identification key in order to link newly received information with that already received for the particular patient.\n\n\\[See additional file [1](#S1){ref-type=\"supplementary-material\"}\\]\n\nIn the biobank scenario, blood samples were to be taken from patients during, but in addition to, routine patient care. The samples were to be retained indefinitely. The information garnered from the blood samples would be linked with the diabetes registry using a common study ID. The combined biobank and registry were intended to serve as an ongoing resource for studying biological markers of diabetes and conducting pedigree studies.\n\n\\[See additional file [2](#S2){ref-type=\"supplementary-material\"}\\]\n\nMain Outcome Measures\n---------------------\n\nMajor questions asked were as follows:\n\n\u2022 In terms of the *creation of registries*, we asked whether or not patient consent is required for inclusion in a registry, and the rationale. We also queried the duration of consent -- that is, whether it should last for the duration of the registry or if periodic renewal would be required -- and the reasoning behind their views on duration.\n\n\u2022 As to the *operation of registries*, we inquired into the need (or lack thereof) for ongoing monitoring of the registry by the REB, and the types of information that would need to be reported. We also probed whether the REB would review the individual research projects utilizing the registry, and the factors that contributed to this decision.\n\n\u2022 We asked whether the REB viewed *biobanks*as qualitatively different from registries, and the reasons behind their views. Further, we queried the need (or not) for consent and, if needed, its duration.\n\n\u2022 We also inquired into any additional reporting requirements surrounding *biobanks*.\n\nEmbedded Issues\n---------------\n\nA number of issues were built into the scenario. These included the implications of requesting REB review without a specific set of research questions attached to the creation of the registry/biobank, but rather only a general research agenda. In the case of the registry, the data were to be sent offsite to the principal investigator\\'s office for coding and removal of identifiers. For the biobank, a common study ID would be used for both the biological samples and the clinical data in order to facilitate linkage.\n\nAnalysis\n--------\n\nInterviews were transcribed, checked for accuracy against the original audio recordings, and forwarded to interviewees to review for accuracy and for clarification where the initial response may have been unclear. Transcript review moved through several iterations that can be summarized into two stages. In stage 1, all co-investigators reviewed the first 11 interviews and, based on these, identified themes and sub-themes to pursue in the analysis and response categories. In stage 2, the interviewers and a graduate student reviewed all transcripts (including those that had been reviewed in stage 1), coded responses according to the themes identified, and summarized respondents\\' rationales. In some cases, additional themes emerged or additional nuances were identified for the original themes. When responses were difficult to categorize, the P.I. independently coded these sections. Answers and rationales were then discussed as a group to reach consensus. In a few remaining instances, answers were not classifiable due to a lack of clarity; this is noted where applicable in the results section.\n\nTo support the interpretations drawn by the researchers, short examples or typical statements have been included in the text. Quotations are presented in italics. Minimal editing has been done to preserve authenticity while ensuring readability.\n\nResults\n=======\n\nThirty 90-minute face-to-face interviews were conducted with Chairs and/or Administrators (response rate 88%). In some cases, one or more other REB members also attended, to a maximum of seven in attendance. The median number of attendees was two.\n\nRegistry\n--------\n\nOf the thirty sites, one refused to entertain the scenario regarding the creation of the registry, indicating that its creation was not connected with any specific research question and therefore falls outside its mandate. Their concern was a blurring of the concept of creation of research infrastructure with that of review of research protocols, and that approving the infrastructure would open the door to unapproved data uses by the researcher. The remainder of questions in this section were skipped for this site. The other twenty-nine sites responded.\n\nIn response to the question as to whether patient consent is required for inclusion of her/his data in the registry, twenty-three of twenty-nine sites answered affirmatively. Reasons included the planned collection of identifiable data; the intention to utilize the data for research in future; the fact that identifiable data would be going offsite; and the plan to collect data prospectively, meaning that there would be ongoing contact with the patients and therefore seeking consent would not be onerous. Sixteen of these sites indicated they would not be sympathetic to an argument by the researcher that seeking individual consent is impracticable.\n\nSix sites indicated that consent would not be required for creation of the registry. Three of these did not consider this to be research; one indicated \\\"this sounds more like ongoing chart review\\\", and another that \\\"it\\'s an exploratory study on a large volume of data.\\\" Two of the sites would not require consent because the data would be stripped of direct identifiers prior to its entry into the registry. Two sites not requiring consent would place conditions on the creator of the database -- i.e., either an information letter to patients or notification with opt-out.\n\nOf the twenty-three sites that would require consent, there was a high degree of variation as to limits on its duration. Twelve agreed, but for differing reasons, that the patient\\'s consent would run for the duration of the registry in the absence of significant change. Of these, two sites saw no reason for periodic renewal of consent; five indicated providing an option to withdraw would obviate the need to require consent renewal; and three sites were motivated by the fact that they would require consent for specific research studies utilizing the registry. Another reason given was that the registry would lose scientific validity over time if periodic consent were required. Five sites would require periodic renewal, with the periods ranging from every subsequent patient visit to once every five years. This was viewed as feasible given that the patients are to be followed in the course of routine clinical care. Four sites were undecided, indicating that the duration of consent would be decided on a case-by-case basis. One site\\'s answer was indecipherable and we were unable to get clarification on follow-up.\n\nTwenty-four of twenty-nine sites would require periodic reporting to the REB by the registry custodians. Along with standard information for progress reports, the content of such reports would include registry-specific information such as how the registry is being managed, who has access, the evolution of the population (i.e., enrolments and withdrawals), and the source(s) of funding of the registry. In the case of the five remaining sites, one was undecided, in one case the answer was unclear, in one case the question was skipped, one indicated it would only require reporting in case of amendments to the registry, and the final site would not require periodic reporting due to a lack of resources for follow-up.\n\nThe twenty-nine sites were also asked whether specific research projects utilizing the registry would require REB review. Twenty answered in the affirmative, although three of these twenty indicated that such review would likely be expedited (i.e., not reviewed by the full REB). One site said \\'no\\', while six provided responses conditional on the circumstances; for three of the six, review would not be required if the data was de-identified; for two, it would depend on whether there were substantial changes to the protocol consented to upon establishment of the registry; and for one, review would only be required if dramatically different uses were to be made of the data (e.g., linkage to blood samples). Two sites were undecided.\n\nRegistry Combined with Biological Samples\n-----------------------------------------\n\nOne REB viewed the creation of a biobank with linkage to the registry information as outside the scope of REB scrutiny (the same REB that had indicated that registry creation was outside its scope). Of the remaining twenty-nine sites, twenty-three viewed the biobank information as qualitatively different from other types of personal health information, while six indicated the difference was at most a question of degree. One site stated that there is no difference; all information requires sensitive handling, whether or not it has genetic markers. Those that stated the difference is one solely of degree generally regarded information from the biological sample as being more sensitive, replicable, commercializable, and predictive. Of the sites that viewed it as qualitatively different, reasons given were its intra-familial and inter-generational nature; the implications for insurability and employability; the potential uses in deciding on paternity; and its regional or group implications, including discrimination on the basis of race.\n\nThe twenty-nine sites were unanimous with regard to the need for patient consent for blood samples to be placed in the biobank. All six of the sites that had not required consent for participation in the registry would now do so.\n\nIn terms of duration of consent for the biobank samples, six stated that its duration was time-limited, eleven indicated there would be no time limit for retention, five were undecided, in three cases the answer was unclear, and in four cases the question was skipped. Note that two of the sites that would require periodic re-consent for the registry alone would not require it when the biological samples were combined with the registry. These findings demonstrate indeterminacy on the part of REBs, as revealed by one site\\'s statement: \\\"We don\\'t have to provide answers to all these questions. They\\'re not all answerable.\\\"\n\nOf the six sites who said consent would be time-limited, one stated that \\\"blood is different\\\" and another that its potential uses are endless, unlike registries. For some of these sites, samples or linkages would be destroyed after a set time period, ranging from five to twenty-five years. Those indicating no time limit to consent sometimes included one or more qualifiers, such as a withdrawal option, and notification should there be significant changes to the biobank. Some additional requirements identified were: full REB review in the first year of operation; bio-collection treated as a separate protocol; periodic report on the activities and outcomes of research using the biobank; and scrutiny of physical security measures. As with the diabetes registry, one site would not require periodic reporting due to lack of REB resources.\n\nStudy Limitations\n-----------------\n\nInterviewees had been told that the interviews would take at maximum ninety minutes. Since questions were open-ended, it was necessary at times that the interviewer skip some questions in order to complete in the promised time frame. This led to some incompleteness in results.\n\nNote that this study was constructed around hypothetical situations. An REB faced with a real-life application for approval would have the opportunity to request further details and to deliberate at length. Further, the outcomes measure was what sites *said*they would do, based on these hypothetical facts. Not all of the sites had handled all of the types of requests included in our scenarios. Thus, some of the answers may have reflected their understanding of current guidelines, rather than reflecting past practice. In addition, responses may have been shaped in accordance with what the interviewee expected the interviewer wished to hear.\n\nDiscussion and Conclusion\n=========================\n\nWe found that participants were more attuned to issues surrounding biobanks as compared to registries, despite similarities regarding their creation and long-term research potential. This is not surprising given that the Tri-Council Policy Statement (TCPS), a statement agreed to by the major federal research funding agencies in Canada which aims to ensure the ethical conduct of research, is silent as to registries. Also, there is a dearth of literature, both in Canada and internationally, concerning the role of REBs vis-\u00e0-vis registries. There is also a significant degree of variation in how the sites in our survey indicated they would handle research proposals for creation and use of these entities. For example, six of the twenty-nine sites entertaining the scenario would not require patient consent for the entry of personal information into a registry, whereas all twenty-nine would require consent for entry of blood samples into a biobank. At least two factors are at play in creating the consensus as to biobanking. First, participants saw the scope of potential research activities to be much more broad for biobanks in comparison to registry information. Indeed, the limits for biobanks were identified as unknowable. Second, the TCPS does contain guidance as to human tissue, including the explicit requirement of informed consent to its collection and use \\[[@B2]\\].\n\nAccompanying this greater familiarity is a dramatically higher level of concern on the part of sites regarding biobanks. One referred to biobank information as a \\'gray box\\' in that its potential future uses are at present unknowable. Others referred to such information as providing a \\\"total picture of the person\\\" or \\\"a window into one\\'s soul\\\", and that \\\"the sum is greater than the parts\\\". These vivid and dramatic descriptors are indicative of trepidation on the part of participants regarding genetic information. There is a significant degree of ambivalence in the literature on biobanks as to whether or not genetic information is inherently different from other types of health information. Some argue that all personal health information is potentially sensitive \\[[@B3]\\]. Others lean to \\\"genetic exceptionalism\\\"\\[[@B4],[@B5]\\] despite the fact that other types of information may also implicate family or community as well as the individual, and may be highly sensitive (e.g. HIV status or psychiatric record). The majority of sites in our study (23/29) viewed genetic information as qualitatively different, thus weighing in on the \\'exceptionalism\\' side of the debate.\n\nGiven this acutely higher level of concern regarding biobanks, it is surprising that an equal number of sites would not require periodic renewal of consent for registries and for biobanks. Specifically, sixteen sites would permit the entry of information into a registry to run indefinitely or were undecided, and sixteen sites would either permit consent to banking of a blood sample to continue indefinitely into the future or were undecided. One of the sites that would not require periodic renewal of consent for the biobank in contrast to the registry provided this explanation: \\\"No, because once the sample is given, it\\'s for life, you don\\'t go a second time\\...up to now it\\'s once and for all.\\\" These findings give rise to serious concern about consent practices regarding biobanks, especially since the samples are often retained long-term. After describing the general standards for informed consent for research involving human subjects in the U.S., Natalie Ram notes that \\\" \\[a\\]gencies and courts have been hesitant to impose similar consent requirements on researchers obtaining human tissue for use in research, and human tissue research has therefore become a particularly thorny problem for traditional formulations of informed consent.\\\"\\[[@B6]\\]\n\nOne of the fascinating differences between sites with regard to their concerns or lack thereof with identifiability of registry data revolves around at what point in the process they were focussing on. Sites with concerns looked at an earlier period of time than entry into the registry -- i.e., the fact that identifiable data were to go offsite to the principal investigator\\'s office prior to being de-identified. One site indicated they would simply not allow release of personal information out-of-house, as had been proposed. The sites that were not concerned indicated that the data was de-identified upon entry into the registry. At least one REB member expressed trust that researchers would safeguard the personal information and not attempt to re-identify individuals.\n\nExpansion continues apace for registries and biobanks. This results in a need for critical analysis of suitable roles for REBs and subsequent guidance on these topics. A first step is to establish a dialogue on these issues, especially regarding registries; it is hoped that this project facilitates such discussion. Registries are of burgeoning importance in response to demands for evidence-based decision-making and the growth in numbers of epidemiological studies. They give rise to a number of privacy and consent issues that outstrip current guidance and yet will need to be dealt with by REBs.\n\nA second step will be the provision of urgently needed guidance regarding appropriate uses of information in biobanks and registries. One site referred to the rapidly changing context of genetics, and indicated that \\\"\\...we\\'re still disoriented.\\\" Sections of the TCPS on biobanking have not been updated since 1998 despite significant changes in practice combined with a huge expansion in their importance and significance. The Canadian Institutes of Health Research has developed a voluntary Best Practices code for the handling of personal information in health research \\[[@B7]\\]. While registries are addressed in the document, recommendations are currently very broad; more specific guidance as to both registries and biobanks would be in order for future editions. Further, we call on the Interagency Panel on Research Ethics to undertake a review and redrafting of parts of the TCPS of direct relevance to registries and biobanks. Accompanying this should be an education programme covering these topics aimed at researchers, REB members, and privacy commissioners.\n\nThird, we urge that REBs adopt an active role in guiding the creation of registries and biobanks. This holistic approach responds to the development of multi-project research platforms as opposed to simply individual projects. Several sites were concerned about jurisdiction and lack of specificity in being asked to review infrastructure; one site indicated that \\\"we shouldn\\'t be collecting data until we know what the future use might be\\...this is just a little too wide open, it\\'s a fishing expedition.\\\" However, it is our position that since registries and biobanks are indeed being created, it makes sense that any obvious potential problems be addressed up-front, prior to the infrastructure being developed. This will result in greater efficiency and less work later for both researchers and REBs, and the avoidance of future problems. The concern about a \\'fishing expedition\\' can be allayed by the fact that the individual research projects relying on the platform should still be subjected to REB review.\n\nFourth, the development of one or more specialized REBs with expertise in the area of registries and biobanks is well worth considering. Models that could be adapted exist in the form of committees that specialize in screening access to databases and in the governance of biobanks \\[[@B8]\\].\n\nAnd finally, in the longer term, we suggest the development of governing legislation. This would provide a superior form of guidance and control, given the sensitivity of personal health information generally and of genetic information in particular. This is particularly pertinent if use and/or disclosure of personal health information, including genetic material, without consent is under consideration. Caulfield et al. suggest that an authorization model for genetic databases may be superior to the present consent regime, but that legislation would be needed prior to adopting such a model \\[[@B9],[@B10]\\].\n\nThe development of legislation would not be free of complexities. For example, health and information are both primarily within provincial jurisdiction. All provinces have legislation governing information in the public sector, and most now also have legislation that covers aspects of information-handling in the private sector. There is also federal private sector legislation \\[[@B11]\\]. There is a lack of consistency as to the impact of these various statutes on the conduct of research, and often a lack of clarity. Thus, drafters of legislation would need to take into account these issues and contingencies. However, the difficulties are not insurmountable, and there is an obvious sense of need. To paraphrase one of the sites in our study:\n\nSome of the decisions should be taken by authorities above local REBs. Rules should be clarified so that each local REB does not have to take decisions. Such fundamental decisions \\[should\\] not rest upon the shoulders of local REBs. The consequences of certain decisions can impact on people or populations, which makes even more \\[persuasive\\] the case for the need for a regulatory framework on banks.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nEG was primary author of this paper. DW participated substantially in its revisions and was the primary investigator for the larger project of which the study of registries and biobanks formed a part. Other authors participated in the project and provided feedback on drafts.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1472-6939/9/17/prepub>\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\n**Creating a Diabetes Registry.** Scenario for Registry.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**Prospective Collection of Biological Samples for Diabetes Biomarkers and Pedigree Studies.** Scenario for Biobank.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis research was funded by a grant from the Canadian Institutes of Health\n\nResearch (grant number MOP-577484).\n"}
+{"text": "Introduction {#S0001}\n============\n\nMyocardial infarction is the major cause of morbidity and mortality in industrialized countries ([@CIT0001]). Acute coronary syndrome (ACS) is a set of signs and symptoms caused by rupture of an arterial plaque which provokes platelet--rich coronary thrombus formation. The thrombus leads to partial or complete coronary artery occlusion, which, in turn, results in myocardial ischemia and various clinical manifestations ranging from unstable angina (UA) to acute myocardial infarction (AMI) ([@CIT0002]). Platelets are blood cells with diverse sizes and densities. In addition to aggregation, platelets modulate important pathophysiological processes, including inflammation and coagulation. Larger platelets generally contain more dense granules, express more glycoprotein llb/llla receptors, have a higher thromboxane A2 level and aggregate more rapidly with collagen, and thus are considered to be more active. It has been shown that platelet size, measured as mean platelet volume (MVP), is a marker of platelet function and is positively associated with platelet activity. An increased MPV is an indicator of larger and more active platelet, associated with myocardial damage in ACS, and is an unfavorable predictive factor in AMI survivors ([@CIT0003]--[@CIT0004]).This study aimed to investigate whether there is an association between MPV and platelet count measured at admission with cardiac troponin T (cTnT) elevation in a large sample of patients with acute chest pain who were admitted to the emergency departments.\n\nMethods {#S0002}\n=======\n\n851 patients were selected with acute chest pain who were admitted to the emergency department of Rasoul-e-Akram Hospital during the year 2010. Two blood samples were taken from each patient in 4 hours of their arrival: one for routine hematologic tests and the other for cTnT. Peripheral venous blood for hematologic testing and cTnT measurements was collected in tubes containing K2--ethylenediaminetetraacetic acid and lithium heparin respectively. The cTnT samples were centrifuged at 1500g for 10 minutes at room temperature and immediately were analyzed by Roche cardiac reader. The samples for hematologic test, including platelet count and MPV, were assayed on cell counter of Sysmex kx 21. The cTnT results were stratified into two groups: troponin positive (troponin \u22650.1\u00b5g/L) and negative (troponin \\< 0.1 \u00b5g/L). All patients underwent a standard 12-lead electrocardiogram (ECG), which was interpreted by the cardiologists of the hospital according to conventional criteria. Patients with ACS did not undergo coronary angiography. All the patients with acute chest pain who were admitted to the emergency department of our hospital during the year 2010 were included in the study. Patients without cTnT, platelet count or MPV in their laboratory results, or ECG due to any reasons were excluded from the study.\n\nStatistical Analysis {#S20003}\n--------------------\n\nSignificance of differences and frequency distribution of values between the groups were assessed by the Mann-Whitney U-test. Statistical analyses were performed using statistical software SPSS. P\\<0.05 was considered statistically significant.\n\nResults {#S0004}\n=======\n\nThe baseline characteristic of our study population and the main results of this investigation are shown in [Table 1](#T0001){ref-type=\"table\"}. A total of 851 were enrolled in the study. 169 (19.9%) of admitted patients had positive troponin and 682 (80.1%) had negative results.\n\n###### \n\nMPV and platelet counts in troponin positive and negative groups.\n\n  Troponin               Patients with negative troponin (n = 682)   Patients with positive troponin (n = 169)   p                                    \n  ---------------------- ------------------------------------------- ------------------------------------------- ----------------------------- ------ ---------\n  Age (yr)               62.62 (range:18-100)                        0.62                                        68.59 (range:22-92)           1.07   \\<0.005\n  Platelet count (/\u00b5l)   221683 (range:26000-939000)                 3198                                        198814 (range:26000-637000)   6776   \\<0.001\n  MPV (fl)               9.57 (range:5.7-13.9)                       0.04                                        9.92 (range:7.5-13.4)         0.09   \\<0.001\n\nMPV and platelet count are significantly higher and lower in troponin positive group than troponin negative group respectively.\n\nIn troponin negative group, the mean of MPV was 9.57\u00b12.054 fl with max: 13.9 and min: 5.7. The mean age of the subjects was 62.62\u00b1 32.14 years with max: 100 and min: 18.The mean of platelet count was 221683\u00b1167032/\u00b5l with max: 939000 and min: 26000.\n\nIn troponin positive group, the mean of MPV was 9.92\u00b1 2.37 fl with max: 13.4 and min: 7.5.The mean age of the subjects was 68.59\u00b1 27.92 years with max: 92 and min: 22.The mean of platelet count was 198814\u00b1 176166/\u00b5l with max: 637000 and min: 26000.\n\nAs it is shown in [Fig. 1](#F0001){ref-type=\"fig\"} and [Table 1](#T0001){ref-type=\"table\"}, in troponin positive group, MPV values are significantly higher than that those in troponin negative group (9.92 vs. 9.57 fl with p\\<0.001), ([Fig. 2](#F0002){ref-type=\"fig\"} and [Table 1](#T0001){ref-type=\"table\"}), and platelet count is significantly higher than that in troponin positive group (221683 vs 198814/\u00b5l with p\\< 0.001). Also the age of patients in troponin negative group was significantly lower than troponin positive group (62.62 vs. 68.59 with p value \\<0.005) ([Table 1](#T0001){ref-type=\"table\"}).\n\n![Comparison of MPV in troponin positive and negative groups. MPV values were significantly higher in troponin positive group than troponin negative group.](MJIRI-26-17-g001){#F0001}\n\n![Comparison of platelet count in troponin positive and negative groups. Platelet counts were significantly lower in troponin positive group than troponin negative group.](MJIRI-26-17-g002){#F0002}\n\nDiscussion {#S0005}\n==========\n\nChest pain is a commonly presented symptom in hospitals. However, a minor group of patients admitted to the emergency department with chest pain are proved to have a cardiac etiology using electrocardiography and cTnT measurement ([@CIT0005]). Current cardiac markers are not sensitive enough for early diagnosis of ACS ([@CIT0006]), which is a leading cause of mortality. Physicians always search for rapid and independent markers for early and accurate diagnosis of ACS. Now the question need to be addressed whether these required markers are novel and expensive or they are ignored markers that could contribute to cells known to pathogenesis of thromboemboli.\n\nPlatelets play a pivotal role in atherogenesis ([@CIT0007]). Platelet activation ultimately leads to the formation of thromboxane A~2~ a potent vasoconstrictor and platelet aggregating substance, or leukotrienes, strong mediators of acute inflammatory response ([@CIT0008]). Larger hyperactive platelets play an important role in intracoronary thrombus formation and acute thrombotic events ([@CIT0009]). Decrease in platelet count can be due to participation of platelets due to thrombus process ([@CIT0010]).\n\nRecently, there has been some focus on platelet count and MPV for aiding the diagnosis of the ACS. The reason is these simple, inexpensive tests available in routine laboratories and used for nearly every patient admitted to emergency room by a routine CBC test. If these parameters are proven to be valuable in the diagnosis of ACS, then medical cost would be far less for patients.\n\nPrevious studies have shown for diagnosis of various illnesses that increased MPV or a decrease in platelet count might be useful. They can be conjunct with other conventional biochemical cardiac markers understanding etiology for patients admitted with chest pain. ([@CIT0008], [@CIT0010]--[@CIT0015]). It has also been reported that MPV measurement at admission might be valuable in the prediction of the infarct-related artery patency ([@CIT0016]) and is also a useful hematological marker for early and easy identification of patients with stable CAD who are at a higher risk of post percutaneous coronary intervention low-reflow ([@CIT0017]). Finally, elevated MPV and resistance to aspirin have proven to be prognostic factors, for death, MI and the composite endpoint ([@CIT0018]).\n\nTo the best of our knowledge, MPV and PLT count had not been previously studied in Iranian population presenting with chest pain. Our study was performed on the patients who were admitted to the emergency department of Rasoul-e-Akram Hospital for chest pain. The findings indicated that MPV level at admission is significantly higher in patients who have elevated troponin levels than those with normal levels. Also, PLT count at admission was significantly lower in patients who have elevated troponin levels than those with normal troponin levels.\n\nIn our study, the electrocardiography of the patients was also recorded and interpreted by the cardiologists according to conventional criteria for ACS. The patients were diagnosed with ACS according to the presence of suggestive cardiac symptoms, ischemic electrocardiographic changes, and elevation of cTnT. Based on the obtained results, elevated MPV and low PLT count are noted in ACS.\n\nThe obtained data have confirmed the results of previous studies ([@CIT0008], [@CIT0010]--[@CIT0015]). Our findings provide further evidence that platelet activation, measured by elevated MPV and low PLT count, may contribute to the pathogenesis of thrombosis-related complications in coronary artery disease ([@CIT0019]).\n\nThese valuable tests may also have some other advantages. High MPV may show adverse prognosis ([@CIT0018], [@CIT0020]--[@CIT0021]) and therapeutic needs ([@CIT0020]). Patients with ACS having high pre-procedural MPV values might benefit from an intensified antiplatelet therapy after coronary interventions ([@CIT0020]). MPV is also considered as a cost-effective tool that may be used even for predicting the possibility of acute coronary events ([@CIT0014]).\n\nThese simple parameters which are already available in routine laboratories can help with accurate diagnosis of ACS without adding an economic burden. Also, it can give us some insight about prognosis and therapeutic management of patients and predicting the possibility of acute coronary events in patients who do not fulfill the criteria of ACS on arrival.\n\nConclusion {#S0006}\n==========\n\nMPV and PLT count at admission are significantly higher and lower respectively in patients diagnosed with ACS than in those with chest pain of non-cardiac origin. Therefore, these simple and inexpensive laboratory measurements can be used in conjunction with other laboratory tests for patients admitted to the emergency department with chest pain. This could help to lower hospitalization and also misdiagnosis and having complications in patients with ACS. Nevertheless, further randomized clinical trials are needed to confirm the clinical usefulness of MPV and PLT count in risk stratification scoring systems that can predict myocardial infarction in admitted patients with ACS.\n"}
+{"text": "INTRODUCTION\n============\n\nPTEN (phosphatase and tensin homologue deleted on chromosome 10) is a tumour suppressor, with inactivating mutations that are among the most common in solid tumours \\[[@B1]--[@B3]\\]. Although initially identified as a protein tyrosine phosphatase (PTP) \\[[@B4]\\], the preferred substrate of PTEN is the lipid second messenger phosphatidylinositol 3,4,5-trisphosphate (PIP~3~) \\[[@B5]\\]. PTEN activity directly opposes the activity of the phosphoinositide 3-kinases (PI3Ks), preventing activation of the master protein kinase Akt and consequently inhibiting cell growth and proliferation \\[[@B6],[@B7]\\]. PTEN can also antagonize the PI3K pathway as a protein phosphatase through dephosphorylation of insulin receptor substrate 1 (IRS1) \\[[@B1]--[@B3],[@B8]\\]. Homozygous PTEN knockouts are lethal in mice, and heterozygous mutations result in a predisposition to tumour formation \\[[@B4],[@B9],[@B10]\\]. Characterization of PTEN mutations in tumours has led to a continuum model for tumour suppression in which loss of a single PTEN allele is sufficient to drive tumorigenesis \\[[@B5],[@B11]\\]. More recently, catalytically compromised PTEN mutants have been proposed to act in a dominant-negative mechanism due to dimerization \\[[@B6],[@B7],[@B12]\\], causing dramatic decreases in PTEN activity even when a single allele is mutated.\n\nPTEN is composed of three domains flanked by two motifs ([Figure 1](#F1){ref-type=\"fig\"}A): an N-terminal phosphatidylinositol 4,5-bisphosphate (PIP~2~)-binding motif (PBM, residues 1--7), a dual-specificity phosphatase domain, a C2 domain, a 50-amino-acid unstructured C-terminal tail and a C-terminal PDZ-binding motif \\[[@B13],[@B14]\\]. The phosphatase domain has an active site that is formed by the conjunction of three loops: the P loop, the WPD loop and the TI loop. The WPD loop (residues 88--98) is a flexible loop that carries Asp^92^, a residue critical for substrate catalysis \\[[@B13],[@B15]--[@B17]\\]. The TI loop (residues 160--171) forms a uniquely deep pocket to accommodate the bulky PIP~3~ substrate \\[[@B13],[@B16]\\]. The highly conserved P loop (residues 121--131) contains the HCXXGXXR phosphatase signature motif, including the catalytically crucial cysteine, Cys^124^, which forms a cysteinyl phosphoenzyme intermediate that is subsequently hydrolysed \\[[@B18]\\].\n\n![Phosphorylation of PTEN C-terminal tail regulates its activity and substrate specificity\\\n(**A**) Schematic representation of PTEN, highlighting various phosphorylation sites. The PBM (residues 1--7) is followed by a phosphatase domain (14--185), the C2 domain (190--351), the 50-amino-acid tail (353--400) containing multiple phosphorylation sites and the PDZ-binding motif (401--403). Differentially phosphorylated constructs of PTEN were produced, referred to as Phos-wtPTEN, Phos-PTEN-4A and Phos-PTEN-2A. (**B**) Phosphatase activity of PTEN against soluble diC~8~-PIP~3~ substrates. All activity assays and lipid-binding assays have error bars showing S.D. from the mean, derived from experiments carried out in triplicate. Some error bars are smaller than the marker. All experiments were conducted at least three times. (**C**) Phosphatase activity against IP~4~, a soluble head group substrate. (**D**) Phosphatase activity against a phosphotyrosine-containing polypeptide. (**E**) Phosphatase activity against lipid vesicles containing a long-chain PIP~3~ substrate. PIP~2~-containing vesicles (5%) were incubated with PI3K to generate PIP~3~, which was subsequently dephosphorylated with various concentrations of PTEN. (**F**) Protein/lipid FRET curves of variously phosphorylated PTEN constructs binding to 5% PIP~2~-containing membrane-mimicking vesicles.](bj4730135fig1){#F1}\n\nPTEN activity is controlled epigenetically \\[[@B19]\\], post-transcriptionally and post-translationally \\[[@B20]\\]. The 47-kDa PTEN protein is subject to phosphorylation \\[[@B21],[@B22]\\], acetylation \\[[@B23]\\], oxidation \\[[@B24]\\], ubiquitination \\[[@B25]\\] and SUMOylation \\[[@B26]\\]. Furthermore, recent discoveries have shown that PTEN can be transported from one cell to another, via ubiquitin-mediated exosomal sorting \\[[@B27]\\] and through secretion of PTEN-long (PTEN-L) \\[[@B28],[@B29]\\].\n\nThe N-terminal extension of PTEN-L (a longer variant of PTEN which results from an alternative translation initiation site) encompasses a poly-alanine secretion signal and a poly-arginine re-entry sequence that enable the protein to be secreted and enter neighbouring cells respectively \\[[@B29]\\]. PTEN-L also localizes to the mitochondria, where it regulates cytochrome *c* oxidase function \\[[@B30]\\]. The N-terminal extension is predicted to be highly disordered and the target of numerous post-translational modifications \\[[@B31]\\]. PTEN-L was suggested to be in a constitutively active state with regard to PIP~2~-mediated activation of PTEN, possibly due to an altered structure surrounding the PBM \\[[@B32]\\].\n\nPTEN activity is closely regulated by phosphorylation at multiple sites by serine/threonine kinases. Six inhibitory phosphorylation sites are located within the PTEN tail on residues Thr^366^, Ser^370^, Ser^380^, Thr^382^, Thr^383^ and Ser^385^ ([Figure 1](#F1){ref-type=\"fig\"}A). These sites are targeted by the CK2 and GSK3\u03b2 \\[[@B22],[@B33]--[@B35]\\], preventing proteasomal degradation of PTEN \\[[@B36]\\]. Two models of auto-inhibition have been proposed. In one model, the phosphorylated tail interacts with the calcium-binding region (CBR)3 loop of the C2 domain, either electrostatically \\[[@B35]\\] or non-electrostatically \\[[@B37]\\], preventing membrane binding and PIP~3~ dephosphorylation. The second model argues that the phosphorylated tail interacts with the phosphatase domain and the PIP~2~-binding domain (PBD) \\[[@B22]\\]. Auto-dephosphorylation of pThr^366^ and pSer^370^ can selectively remove these phosphates \\[[@B38]\\].\n\nIn the present study, we determine how phosphorylation of PTEN and the addition of the N-terminal extension of PTEN-L affect enzymatic properties such as substrate specificity and interfacial catalysis. Furthermore, we use HDX--MS to provide structural insights into how the phosphorylated C-terminal tail of PTEN and a putative \u03b1-helix in the N-terminal extension of PTEN-L regulate these catalytic properties.\n\nMATERIALS AND METHODS\n=====================\n\nProtein expression and purification\n-----------------------------------\n\nFor expression of recombinant PTEN proteins, 1--8\u00a0litres of *Spodoptera frugiperda* Sf9 cells at a density of 1\u00d710^6^ cells/ml were infected with an optimized ratio of virus. Expression of PTEN in *Escherichia coli* was found to produce unsatisfactory material that showed pronounced aggregation. All PTEN constructs contain an N-terminal His~6~-tag followed by a tobacco etch virus (TEV) cleavage site. After 48\u00a0h of infection at 27\u00b0C, the cells were harvested and washed with ice-cold PBS. The pellets were lysed by sonicating for 5\u00a0min in Buffer A \\[20\u00a0mM Tris/HCl, pH\u00a08.0, 300\u00a0mM NaCl, 10\u00a0mM imidazole, 5% glycerol, 2\u00a0mM 2-mercaptoethanol and 0.5% Triton X-100, with one Complete EDTA-free protease inhibitor tablet added (Roche) per 50\u00a0ml of buffer\\]. Lysate was centrifuged for 45\u00a0min at 140000\u00a0***g***. The supernatant was then passed through a 0.45\u00a0\u03bcm filter (Sartorius Biotech) and then loaded on to a 5\u00a0ml of HisTrap FF column (GE Healthcare). The column was washed with up to 30\u00a0mM imidazole before being eluted with a 0--100% gradient of Buffer B (20\u00a0mM Tris/HCl, pH\u00a08.0, 100\u00a0mM NaCl, 5% glycerol, 200\u00a0mM imidazole and 2\u00a0mM 2-mercaptoethanol). The pooled fractions containing the PTEN were then placed along with an aliquot of His-tagged TEV protease into a 10 000 Da molecular-mass cut-off Snakeskin Dialysis membrane and dialysed in 4\u00a0litres of Buffer C (20\u00a0mM Tris/HCl, pH\u00a08.0, 200\u00a0mM NaCl, 5% glycerol, 2\u00a0mM tris-(2-carboxyethyl(phosphine (TCEP)) at 4\u00b0C for 12\u00a0h. The ratio of PTEN/TEV in the dialysis was 20:1 (w/w). To remove the His~6~--TEV protease, the dialysed solution was subsequently passed back over a 5\u00a0ml of HisTrap FF column. Flow-through from the HisTrap column was then diluted 1:1 with Buffer D (20\u00a0mM Tris/HCl, pH\u00a08.0, 10% glycerol and 1\u00a0mM DTT), before being loaded on to a 5\u00a0ml HiTrap Q column (GE Healthcare) pre-equilibrated in Buffer E (20\u00a0mM Tris/HCl, pH\u00a08.0, 50\u00a0mM NaCl, 10% glycerol and 1\u00a0mM DTT). Protein was eluted with a gradient of 0--100% Buffer F (20\u00a0mM Tris/HCl, pH\u00a08.0, 1\u00a0M NaCl, 10% glycerol and 1\u00a0mM DTT). Pooled fractions were concentrated with an Amicon 10\u00a0000 Da molecular-mass cut-off centrifugal filter (Millipore) and injected on to a Superdex 75 16/60 gel-filtration column equilibrated in Buffer G (20\u00a0mM HEPES, pH\u00a07.4, 200\u00a0mM NaCl and 2\u00a0mM TCEP). Fractions were collected and concentrated to at least 1.5\u00a0mg/ml using an Amicon 10\u00a0000 Da molecular-mass cut-off centrifugal filter (Millipore) then divided into aliquots, frozen in liquid nitrogen and stored at--80\u00b0C.\n\nPTEN-L was expressed with the six C-terminal phosphorylation sites mutated to alanine (PTEN-L-6A; T539A-L, S543A-L, S553A-L, T555A-L, T556A-L and S558A-L), as all dephosphorylation conditions tested were found to cause the protein to precipitate. PTEN-L-6A was expressed with an N-terminal TEV-cleavable Protein A tag in Sf9 cells at 27\u00b0C for 41\u00a0h, before being harvested as above. The pellets were lysed by sonicating 5\u00a0min in Buffer A-L \\[20\u00a0mM Tris/HCl, pH\u00a08.0, 300\u00a0mM NaCl, 5% glycerol, 2\u00a0mM 2-mercaptoethanol, 0.1% Triton X-100 and 0.5% CHAPS, with one Complete EDTA-free protease inhibitor tablet added (Roche) per 50\u00a0ml of buffer\\]. Lysate was centrifuged for 45\u00a0min at 140000\u00a0***g***. The supernatant was then passed through a 0.45\u00a0\u03bcm filter (Sartorius Biotech). IgG beads (GE Healthcare, catalogue number 17096902) were added to the supernatant and incubated for 2\u00a0h at 4\u00b0C. The beads were then washed with 150\u00a0ml of wash buffer (50\u00a0mM Tris/HCl, pH\u00a08.0, 300\u00a0mM NaCl, 2\u00a0mM TCEP, 0.1% Triton X-100 and 0.5% CHAPS), followed by 150\u00a0ml of TEV buffer (50\u00a0mM Tris/HCl, pH\u00a08.0, 300\u00a0mM NaCl and 2\u00a0mM TCEP). PTEN-L-6A was then removed from the beads by incubation with TEV protease overnight, followed by elution using three 15-ml washes with TEV buffer. PTEN-L-6A was then loaded on to a HiTrap SP HP column equilibrated in Buffer SA (20\u00a0mM HEPES, pH\u00a07.5, 50\u00a0mM NaCl and 1\u00a0mM DTT) and eluted using a gradient of 0--100% Buffer SB (20\u00a0mM HEPES, pH\u00a07.5, 1\u00a0M NaCl and 1\u00a0mM DTT).\n\nProtein phosphorylation\n-----------------------\n\nProtein was incubated at room temperature for 24\u00a0h with glycogen synthase kinase 3\u03b2 (GSK3\u03b2) (New England Biolabs) and protein kinase CK2 as per the manufacturer\\'s instructions (approximately 10000\u00a0units of each kinase per 1000\u00a0\u03bcl of 40\u00a0\u03bcM PTEN). The reaction mixture was injected on to a Superdex 75 10/300 gel-filtration column (or a Superdex 75 16/600 for larger volumes) pre-equilibrated in Buffer G. Fractions containing protein were concentrated and frozen in liquid nitrogen for subsequent analysis. Phosphorylation status was checked with both intact and peptide fragmentation MS.\n\nProtein dephosphorylation\n-------------------------\n\nProtein was incubated at 30\u00b0C for 90\u00a0min with \u03bb-phosphatase as per the manufacturer\\'s instructions (approximately 50000\u00a0units per reaction). The reaction mixture was then injected on to a Superdex 75 10/300 gel filtration column pre-equilibrated in Buffer G. Fractions containing protein were concentrated and frozen in liquid nitrogen for subsequent analysis. Phosphorylation status was checked with both intact and peptide fragmentation MS.\n\nIntact MS\n---------\n\nPTEN samples (60\u00a0\u03bcl at 0.4\u00a0\u03bcM) were injected on to an ultra-performance liquid chromatography (UPLC) system and passed over a C~4~ Van-Guard pre-column (Waters) in 0.1% formic acid for 3\u00a0min to remove excess salt before elution using a 3\u03bb-p--100% B gradient of Buffer A (0.1% formic acid) and Buffer B (100% acetonitrile) over 13\u00a0min. This eluent was injected on to a Xevo QTOF (Waters) acquiring a mass range from 350 to 1500 *m*/*z*, with an ESI source operated at a temperature of 225\u00b0C and a spray voltage of 2.5 kV. Data were deconvoluted using MaxEnt Software (MicroMass).\n\nSoluble lipid phosphatase activity assay\n----------------------------------------\n\nRecombinant PTEN was incubated with 50\u00a0\u03bcM dioctanoyl-PIP~3~ (diC~8~-PIP~3~) (Echelon) in 50\u00a0mM Tris/HCl, pH\u00a08.0, and 2\u00a0mM DTT for 10\u00a0min at 37\u00b0C in a final volume of 25\u00a0\u03bcl. The reaction was stopped by the addition of 100\u00a0\u03bcl of Malachite Green Reagent (Millipore). The reaction was then incubated at room temperature, with agitation at 400 rev./min for 15\u00a0min to allow colour to develop. The absorbance at 620\u00a0nm was measured and the amount of phosphate released was determined by using a KH~2~PO~4~ (Millipore) standard curve. The specific activity of the PTEN was obtained by dividing the amount of phosphate released by the concentration of the enzyme.\n\nLipid vesicle preparation\n-------------------------\n\nLipid components were mixed together while in organic solvent (chloroform or chloroform/methanol). The solvent was then evaporated under a stream of nitrogen. Vesicles were composed of 5% brain PIP~2~, 20% brain phosphatidylserine (PS), 45% brain phosphatidylethanolamine (PE), 15% brain phosphatidylcholine (PC), 10% cholesterol and 5% sphingomyelin (w/v; Avanti Polar Lipids). The lipid film was allowed to dry for 1\u00a0h under vacuum and then resuspended in 20\u00a0mM HEPES, pH\u00a07.4, 100\u00a0mM KCl and 1\u00a0mM EGTA. The lipids were bath sonicated for 10\u00a0min then subjected to ten freeze--thaw cycles between liquid nitrogen and a 43\u00b0C water bath. The liposomes were finally extruded 11\u00a0times through a 100-nm filter. Vesicles were frozen in liquid nitrogen and stored at--80\u00b0C, before being defrosted at room temperature before use.\n\nProtein/lipid FRET assays\n-------------------------\n\nVesicles were prepared exactly as for the lipid kinase assays, except that 10% dansyl-PS was substituted for 10% of the brain phosphatidylserine. Vesicles were diluted with 20\u00a0mM HEPES, pH\u00a07.4, 100\u00a0mM KCl and 1\u00a0mM EGTA to give a final concentration of 50\u00a0\u03bcg/ml. PTEN was thawed and centrifuged at 13000 rev./min at 4\u00b0C to remove any precipitate. PTEN was diluted in 20\u00a0mM HEPES, 200\u00a0mM NaCl and 2\u00a0mM TCEP to a final concentration of 4\u00a0\u03bcM. Protein was serially diluted in a 2:3 ratio in the above buffer. The lipid solution (5\u00a0\u03bcl) was then mixed with 5\u00a0\u03bcl of the protein solution at various concentrations. The protein/lipid mixture was allowed to equilibrate for 10\u00a0min, while being agitated at 450 rev./min at room temperature. Reactions were measured with a PHERAStar HTS microplate reader (BMG Labtech) using a 280\u00a0nm excitation filter with 350- and 520-nm emission filters to measure tryptophan and dansyl-PS FRET emissions respectively. The FRET signal is *I--I*~o~, where *I* is the intensity at 520\u00a0nm and where *I*~o~ is the intensity of the solution containing only lipid (i.e. without protein). Binding curves were fitted with a one site-specific-binding model (GraphPad Prism version 5.00).\n\nPIP~3~ incorporated lipid vesicle phosphatase assay\n---------------------------------------------------\n\nPIP~2~-containing vesicles with the composition described above were resuspended at a concentration of 1\u00a0mg/ml and phosphorylated by incubation for 2\u00a0h with PI3K, purified as detailed previously \\[[@B42]\\]. The reaction was carried out in 20\u00a0mM Tris/HCl, pH\u00a07.4, 50\u00a0mM NaCl, 50\u00a0mM KCl, 1\u00a0mM EGTA, 3\u00a0mM MgCl~2~ and 0.1\u00a0mM ATP, before being quenched with 10\u00a0mM EDTA. The resulting PIP~3~-containing vesicles were then incubated with PTEN for 3 min at 25\u00b0C and quenched with 22\u00a0mM H~2~O~2~ (Sigma). The quenched reaction was agitated for 10\u00a0min at 500 rev./min at 25\u00b0C, before being spotted on to a nitrocellulose membrane. After the droplet had dried, the membrane was washed four times with TBS/Tween 20 solution (TBST) (50\u00a0mM Tris/HCl, pH\u00a07.4, 150\u00a0mM NaCl and 0.1% Tween 20) over the course of 1\u00a0h, before being blocked with TBST+2\u00a0mg/ml BSA for 1\u00a0h at 4\u00b0C. The membrane was then incubated with 0.5\u00a0\u03bcg/ml Alexa Fluor 488-labelled GRP1-PH domain in TBST+2\u00a0mg/ml BSA overnight at 4\u00b0C as described previously \\[[@B58]\\]. The membrane was washed six times in TBST and dried at 25\u00b0C before being imaged using a Typhoon imager (GE Healthcare). The intensities were quantified using ImageQuant (GE Healthcare).\n\nSurface dilution assay\n----------------------\n\nThe assay was conducted as described above, but with two lipid stock solutions, a 1\u00a0mg/ml 5% PIP~2~ plasma-membrane-mimicking composition and a 2\u00a0mg/ml 2.5% PIP~2~ plasma-membrane-mimicking composition. PIP~3~ was generated as above. Reactions were conducted with 10\u00a0nM PTEN over a 40-min period. PIP~2~ levels were determined using a GRP1-PH domain as described above.\n\nHydrogen/deuterium exchange (HDX) MS\n------------------------------------\n\nHDX reactions were conducted with 10\u00a0\u03bcl of protein in dilution buffer (20\u00a0mM HEPES, pH\u00a07.5, 150\u00a0mM NaCl and 2\u00a0mM TCEP) and initiated by the addition of 40\u00a0\u03bcl of D~2~O buffer solution (10\u00a0mM HEPES, pH\u00a07.5, 50\u00a0mM NaCl, 2\u00a0mM TCEP and 92% D~2~O), to give a final concentration of 74% D~2~O. Final protein concentrations were 1\u00a0\u03bcM. For HDX--MS membrane interaction experiments, lipid vesicles were mixed with the D~2~O buffer. Four time points of exchange (3, 30, 300 and 3000\u00a0s) were terminated by the addition of a quench buffer (final concentration 0.6\u00a0M guanidinium chloride and 0.8% formic acid). Samples were rapidly frozen in liquid nitrogen and stored at--80\u00b0C until MS analysis.\n\nMeasurement of deuterium incorporation\n--------------------------------------\n\nProtein samples were rapidly thawed and injected on to a UPLC system immersed in ice as previously described \\[[@B42]\\]. The protein was run over an immobilized pepsin column (Applied Biosystems; Poroszyme, catalogue number 2-3131-00) at 130\u00a0\u03bcl/min for 3\u00a0min and the peptides were collected on to a Van Guard pre-column trap (Waters). The trap was subsequently eluted in line with an Acquity 1.7\u00a0\u03bcm particle, 100\u00a0mm\u00d71\u00a0mm C~18~ UPLC column (Waters), using a gradient of 5--36% B (buffer A 0.1% formic acid, buffer B 100% acetonitrile) over 20\u00a0min. Eluent from the column was injected on to a Xevo QTOF (Waters) acquiring over a mass range from 350 to 1500 *m*/*z* for 30\u00a0min, using an ESI source operated at a temperature of 225\u00b0C and a spray voltage of 2.5 kV.\n\nProtein digestion and peptide identification\n--------------------------------------------\n\nPeptide identification was done by running MS/MS experiments using a 5--36% B gradient over 120\u00a0min with a Xevo QTOF (Waters). This was supplemented with a 20-min gradient separation to identify and correct the retention time for all samples. The MS tolerance was set to 3 ppm with an MS/MS tolerance at 0.1 Da. The resulting MS/MS datasets were analysed with the Mascot search within Proteome Discoverer (Thermo Scientific). All peptides with a Mascot score \\>20 were analysed using HD-Examiner Software (Sierra Analytics). The full list of peptides was then manually validated by searching a non-deuterated protein sample\\'s MS scan to test for the correct *m*/*z* state and check for the presence of overlapping peptides. Ambiguously identified peptides were excluded from all subsequent analysis. The first round of analysis and identification were performed automatically by the HD-Examiner software, but all peptides (deuterated and non-deuterated) were manually verified at every state and time point for the correct charge state, *m*/*z* range, presence of overlapping peptides and any deviation from the expected retention time. All HDX experiments were carried out in triplicate.\n\nMass analysis of peptide centroids\n----------------------------------\n\nAll results are presented as relative levels of deuterium incorporation and no correction for back exchange is applied, because no fully deuterated protein sample could be obtained. All percentage differences quoted in the Results section are the maximal changes in HDX seen at any time point of the analysis, unless explicitly stated otherwise.\n\nSize-exclusion chromatography--multi-angle light scattering (SEC--MALS)\n-----------------------------------------------------------------------\n\nPTEN solution (100\u00a0\u03bcl of 20--50\u00a0\u03bcM protein) was injected on to a Superdex 200 GL 10/300 gel-filtration column (GE Healthcare) equilibrated in 20\u00a0mM HEPES, pH\u00a07.4, 200\u00a0mM NaCl and 2\u00a0mM TCEP. Samples were run at 0.5\u00a0ml/min and eluted into a Dawn Heleos (Wyatt) (for light-scattering measurements) and Optilab rEX (Wyatt; for refractive index measurements). The system was calibrated using a BSA standard prior to samples being run.\n\nRESULTS\n=======\n\nPhosphorylation of the C-terminal tail differentially inhibits PTEN towards distinct substrates\n-----------------------------------------------------------------------------------------------\n\nPTEN was expressed in *Spodoptera frugiperda* Sf9 cells and purified to determine how the protein phosphatase and lipid phosphatase activities of PTEN were regulated by post-translational modifications of the C-terminal tail. To produce homogeneously and selectively phosphorylated PTEN, several constructs were produced. First, a construct containing six mutations, T366A, S370A, S380A, T382A, T383A and S385A (PTEN-6A), was prepared to prevent background heterogeneous phosphorylation of the tail, as previously described \\[[@B22],[@B37],[@B39]\\] ([Figure 1](#F1){ref-type=\"fig\"}A). We also generated two constructs that preserved only a portion of the tail phosphorylation sites: PTEN-2A with mutations T366A and S370A (mimicking auto-dephosphorylated PTEN \\[[@B38]\\]) and PTEN-4A with mutations S380A, T382A, T383A and S385A. We also expressed both full-length PTEN as well as a truncated PTEN lacking the tail (PTEN-\u0394Tail). We used \u03bb-phosphatase to dephosphorylate PTEN or CK2 and GSK3\u03b2 kinases to phosphorylate it. PTEN constructs were further purified after \u03bb-phosphatase or kinase treatment before biochemical and structural analysis. Proteins produced in this manner were homogeneously dephosphorylated or phosphorylated, as verified by LC--MS of intact proteins (Supplementary Figure S1). Using SEC--MALS, both the phosphorylated and the non-phosphorylated PTEN were shown to be monodisperse, with masses in agreement with what would be expected for monomers (Supplementary Figure S2).\n\nTo compare the activities and specificities of differentially phosphorylated PTEN constructs, phosphatase assays were conducted against four substrates: a phosphotyrosine-containing acidic polypeptide, the soluble lipid diC~8~-PIP~3~, inositol tetrakis phosphate (IP~4~) and liposome-incorporated PIP~3~ ([Figures 1](#F1){ref-type=\"fig\"}B--[1](#F1){ref-type=\"fig\"}E). The phosphatase activity was examined for each PTEN construct in either a phosphorylated or a dephosphorylated form. To verify that the phosphatase activity present in the dephosphorylated PTEN samples was solely due to PTEN, an inactive C124S-PTEN mutant, treated with \u03bb-phosphatase, was used as a control to ensure that there was no measurable contaminating phosphatase activity.\n\nWe found that phosphorylation of PTEN caused a dramatic reduction in activity against all substrates. Intriguingly, when PTEN was phosphorylated on only four sites (pSer^380^/pThr^382^/pThr^383^/pSer^385^) by using the phosphorylated PTEN-2A mutant (Phos-PTEN-2A), activity against the phosphopeptide and soluble diC~8~-PIP~3~/IP~4~ substrates was restored ([Figures 1](#F1){ref-type=\"fig\"}B--[1](#F1){ref-type=\"fig\"}D), but PTEN remained inhibited against liposome-incorporated PIP~3~ substrate ([Figure 1](#F1){ref-type=\"fig\"}E; Supplementary Figure S3). This suggests that pThr^366^ and pSer^370^ may have a role in occluding the active site of PTEN, whereas pSer^380^, pThr^382^, pThr^383^ and pSer^385^ have a role in occluding the membrane-binding surface. Phos-PTEN-4A, phosphorylated on only two sites, pThr^366^ and pSer^370^, showed similar activities with all substrates to that of dephosphorylated PTEN-4A, with only a slight reduction in activity against liposome-incorporated PIP~3~ ([Figures 1](#F1){ref-type=\"fig\"}B--[1](#F1){ref-type=\"fig\"}E).\n\nPhosphorylation of the C-terminal tail inhibits membrane binding\n----------------------------------------------------------------\n\nTo determine whether the substrate specificities of differentially phosphorylated PTENs are due to altered affinities for liposomes, we conducted protein/lipid FRET binding experiments \\[[@B40]\\] using liposomes with a composition mimicking plasma membranes \\[[@B41]\\] ([Figure 1](#F1){ref-type=\"fig\"}F). Dephosphorylated PTEN (subsequently referred to as wtPTEN) binds to these liposomes, but Phos-wtPTEN had no measurable membrane binding. PTEN-\u0394Tail maintained membrane affinity, and all dephosphorylated constructs showed an affinity similar to that of wtPTEN, suggesting that neither the tail nor the six threonine/serine residues have crucial roles in contacting the membrane. Phos-PTEN-4A (phosphorylated on Thr^366^ and Ser^370^) showed a slight decrease in membrane binding ([Figure 1](#F1){ref-type=\"fig\"}F) which may be due to electrostatic repulsion. Consequently, a slight reduction in liposome-incorporated PIP~3~ phosphatase activity was observed ([Figure 1](#F1){ref-type=\"fig\"}E). Phos-PTEN-2A (phosphorylated on Ser^380^, Thr^382^, Thr^383^ and Ser^385^) showed dramatically decreased membrane binding, exhibiting an affinity indiscernible from phosphorylated wtPTEN, accompanied with a great reduction in activity against liposome-incorporated PIP~3~.\n\nHDX--MS shows that the phosphorylated C-terminal tail forms intramolecular interactions with both the C2 and the phosphatase domains\n------------------------------------------------------------------------------------------------------------------------------------\n\nTo understand the molecular details behind phosphorylation-mediated auto-inhibition, we carried out HDX--MS and compared the HDX levels of the dephosphorylated enzyme with those of the fully phosphorylated enzyme, as described previously \\[[@B42]\\]. Further details of the experiments can be found in the Supplementary Online Data (see Supplementary Figure S4 and the deuterium incorporation for all experiments are shown in Supplementary Figures S5--S8). The changes in exchange described in the text are the largest seen at any time point. Using HDX--MS it was possible to determine the location of conformational changes that occur upon tail phosphorylation ([Figure 2](#F2){ref-type=\"fig\"}A). Phosphorylation produced decreases in HDX (defined as changes greater than both 0.7 Da and 6% at any time point) in regions of phosphatase domain, as well as part of the C2 domain. In the phosphatase domain, five peptides (residues 4--21, 35--45, 82--99, 111--134 and 155--177) showed decreases in exchange. These peptides span the N-terminal PBM, a positively charged loop termed here the 'arginine loop' (the loop spanning p\u03b22-p\u03b11, residues 35--49 \\[[@B13]\\]), the WPD loop, the P loop and the TI loop respectively. In the C2 domain, the C\u03b12 loop (residues 321--342) also shows a reduced HDX (peptide 319--341) upon phosphorylation. Overall, these decreases in HDX suggest that the fully phosphorylated C-terminal tail makes contact at the C2/phosphatase interface and more extensive contacts within the phosphatase domain, interacting extensively with all three loops of the active site and the PBM.\n\n![HDX-MS characterization of PTEN upon phosphorylation and membrane binding\\\n(**A**) Changes in HDX observed between wtPTEN phosphorylated by CK2/GSK3 and \u03bb-phosphatase-treated PTEN. Only peptides with HDX changes greater than 0.7 Da and 6% are shown. All HDX experiments were carried out in triplicate. None of the loops missing from the crystal structure had any observed changes and are not shown. Areas with no peptide coverage are shown as white (N/A). (**B**) Two peptides, residues 4--21 (PBM) and residues 82--99 (WPD), showed an increased rate of HDX in S366A/T370A phosphorylated mutant compared with the wild-type phosphorylated protein, with the WPD loop reaching the same level of HDX as the dephosphorylated PTEN. (**C**) Changes in HDX observed with dephosphorylated wtPTEN in the presence of PIP~2~-containing vesicles. (**D**) Schematic representation of the phosphorylation-dependent mechanism of inactivation.](bj4730135fig2){#F2}\n\nDephosphorylation of Thr^366^ and Ser^370^ exposes parts of the active site\n---------------------------------------------------------------------------\n\nWe used HDX--MS to determine the molecular mechanism that allows selective activation of Phos-PTEN-2A (T366A, S370A) against soluble substrates but not lipid substrates residing within membranes by comparing the HDX rates of fully phosphorylated PTEN (Phos-wtPTEN) with Phos-PTEN-2A. We found that two peptides, residues 4--21 spanning the PBM, and residues 82--99, spanning the WPD loop, had increased rates of HDX ([Figure 2](#F2){ref-type=\"fig\"}B). The largest increase in exchange (\\>15%) occurs in the WPD loop peptide, raising the HDX rate for this peptide to that seen in wtPTEN.\n\nMembrane interaction involves exposed loops on both the phosphatase and the C2 domain\n-------------------------------------------------------------------------------------\n\nTo understand the molecular details of how active PTEN interacts with membranes, we carried out HDX--MS experiments on the active wtPTEN in the presence and absence of lipid vesicles. Upon interaction with lipid vesicles, large reductions in the HDX rate were observed around the N-terminus of PTEN ([Figure 2](#F2){ref-type=\"fig\"}C), with numerous peptides showing greater than 25% reduction in HDX. A peptide encompassing residues 2--7 within the PBM had a 42% reduction. Peptide 35--45, which spans the 'arginine loop', saw a 39% reduction in HDX, along with peptide 72--81 showing a 19% reduction. Peptide 82--99, in the WPD loop, had a 13% reduction in HDX upon membrane binding. In addition, peptide 155--177, the TI loop, showed a 7% reduction in HDX.\n\nThe C2 domain also shows reductions in HDX for two C2 loops (CBRs), peptide 201--215\u00a0in the CBR1 (8% reduction) and peptide 258--273\u00a0in the CBR3 (11% reduction). Finally, peptide 316--330, incorporating the C\u03b12 loop, shows an 11% reduction in HDX when PTEN is bound to membrane. Our results using HDX--MS on non-labelled intact PTEN indicate that both the phosphatase and the C2 domain play roles in binding and orienting PTEN on membranes.\n\nHDX--MS of PTEN-L indicates an ordered segment in the intrinsically disordered N-terminal extension\n---------------------------------------------------------------------------------------------------\n\nPTEN-L-6A was purified from Sf9 cells, as attempts to dephosphorylate PTEN-L were unsuccessful. HDX--MS analysis of PTEN-L-6A (Supplementary Figure S9A*)* indicated that the first 145 residues of PTEN-L were intrinsically disordered as previously predicted \\[[@B31]\\], exhibiting very high levels of HDX--MS (\\>50% exchange at 3\u00a0s of exchange at 0\u00b0C). However, one region of the N-terminal extension of PTEN-L, residues 145--176, was below this threshold, suggesting that this region may be folded and forming an \u03b1-helix, as suggested by secondary structure prediction programs. Interestingly, this region has sequence similarity to the 'S4' transmembrane helix of Ci-VSP (*Ciona intestinalis* Voltage Sensitive Phosphatase), a paralogue of PTEN, further indicating that this region might form an \u03b1-helix ([Figure 3](#F3){ref-type=\"fig\"}A). We refer to this region as the 'membrane-binding helix (MBH)' for reasons detailed below.\n\n![Characterization of PTEN-L\\\n(**A**) Schematic representation of PTEN-L. The 173-residue N-terminal extension contains the six alanine residue signal sequence. The predicted \u03b1-helix between residues 151-L and 174-L (highlighted) has sequence similarity with the PTEN-family members Ci-VSP and TPTE (transmembrane phosphatase with tensin homology), identified here as the MBH. (**B**) Phosphatase activity against lipid vesicles containing a long-chain PIP~3~ substrate. PIP~2~-containing vesicles (5%) were incubated with PI3K to generate PIP~3~, which was subsequently dephosphorylated with various concentrations of PTEN. Experiments shown in (**B**)--(**E**) were conducted at least three times in triplicate. (**C**) Protein/lipid FRET of PTEN-L-6A and wtPTEN. (**D**) Interfacial kinetics experiment, showing different behaviours of wtPTEN, PTEN-L and PTEN-L-\u0394MBH. The reactions were conducted with 10\u00a0nM protein for 30\u00a0min. At this point the reactions were supplemented with more PIP~3~-containing liposomes to a 4-fold molar excess of liposomes to protein. (**E**) Surface dilution of PIP~3~ also highlights the different interfacial kinetics of wtPTEN, PTEN-L and PTEN-L-\u0394MBH. The bulk concentration of PIP~3~ was kept constant by varying the concentration of the carrier lipids. The surface concentration of PIP~3~ was altered between 5% or 2.5%. Although wtPTEN and PTEN-L-\u0394MBH show only minor perturbations in phosphatase rate between the two surface concentrations of PIP~3~, PTEN-L sees a drastic reduction in phosphatase activity upon dilution of PIP~3~'s surface concentration.](bj4730135fig3){#F3}\n\nThe PBM and the 'arginine loop' of PTEN-L-6A showed decreases in HDX compared with wtPTEN (Supplementary Figure S9A). The PBM (residues 177--194 in PTEN-L, 4--21 in PTEN) had a 15% decrease, whereas the arginine loop, (residues 208--217 in PTEN-L, 35--44 in PTEN) had a 7% decrease, suggesting a change in conformation in these regions. Additionally, the C2/PTP interdomain region of PTEN-L exhibited increases in HDX when compared with wtPTEN, suggesting a breakdown of the contacts in this region. A variant of PTEN-L with a deleted MBH (residues Ala^141^-L to Met^174^-L deleted), PTEN-L-\u0394MBH, had a HDX profile almost identical with that of wtPTEN (Supplementary Figures S9C and S9D).\n\nThe MBH of PTEN-L causes PTEN-L to become a membrane scooter rather than a hopper\n---------------------------------------------------------------------------------\n\nPTEN-L-6A was as active as dephosphorylated wtPTEN against soluble diC~8~-PIP~3~ (Supplementary Figure S9B), but not as active against liposome-incorporated PIP~3~ ([Figure 3](#F3){ref-type=\"fig\"}B). Deleting the MBH (PTEN-L-\u0394MBH) increased activity against membrane incorporated PIP~3~ to a level similar to wtPTEN. Despite its apparent lower activity in hydrolysing membrane PIP~3~, protein/lipid FRET measurements indicated that PTEN-L-6A bound to liposomes more tightly than wtPTEN ([Figure 3](#F3){ref-type=\"fig\"}C). We postulated that the reduced activity might be because PTEN-L less readily dissociates from vesicles and cannot move to new PIP~3~-containing vesicles, i.e., PTEN-L works in a scooting rather than hopping mode of catalysis \\[[@B43]\\]. To determine the interfacial kinetic mode of PTEN-L, we conducted two modified PIP~3~ liposome assays. In the first assay, a second aliquot of substrate was added to produce a 4-fold excess of liposomes relative to PTEN after the initial reaction had completed ([Figure 3](#F3){ref-type=\"fig\"}D). Surprisingly, PTEN-L and wtPTEN exhibited different interfacial kinetic behaviour. The addition of fresh substrate caused a burst of activity with wtPTEN, whereas PTEN-L failed to dephosphorylate the additional liposomes, indicating that wtPTEN works in a 'hopping' mode of interfacial kinetics, in agreement with previous studies \\[[@B44]\\], whereas PTEN-L-6A works in a 'scooting' mode. In a second assay, we measured the phosphatase rate of wtPTEN and PTEN-L with the PIP~3~ surface concentration of the membrane at 5% and 2.5%, while maintaining a constant bulk PIP~3~ concentration. The reduction in surface concentration greatly affected the phosphatase rate of PTEN-L ([Figure 3](#F3){ref-type=\"fig\"}E), whereas wtPTEN was affected to a much smaller extent. The results indicate that PTEN-L is carrying out a scooting catalysis, whereas wtPTEN shows predominately hopping behaviour. In both experiments, PTEN-L-\u0394MBH exhibited the same behaviour as wtPTEN, suggesting that the MBH is responsible for this change in interfacial kinetics.\n\nPTEN-L has a vastly altered membrane-binding footprint\n------------------------------------------------------\n\nTo determine how the N-terminal extension of PTEN-L caused a change in interfacial kinetics, we compared HDX of PTEN-L-6A in the presence and absence of lipid vesicles ([Figure 4](#F4){ref-type=\"fig\"}; Supplementary Figure S8). The membrane-binding footprint was very different from that of wtPTEN, with the MBH having the largest changes in HDX (MBH peptides 150--162, 166--176 and 167--176 decreasing 12%, 20% and 22% respectively, PTEN-L numbering). Additionally, PTEN-L showed decreases in HDX in areas associated with membrane binding for wtPTEN, although most changes were of a smaller magnitude. For example, a PBM-containing peptide (residues 177--194\u00a0in PTEN-L, 4--21\u00a0in PTEN) decreased 12% in PTEN-L as compared with 28% for the same peptide in PTEN, a peptide covering the arginine loop (208--215\u00a0in PTEN-L, 35--42\u00a0in PTEN) decreased 11% when PTEN-L binds to liposomes, compared with 32% in PTEN and a peptide spanning the C\u03b12 loop (6% decrease for 492--514 in PTEN-L, 11% decrease for 319--341 in PTEN). Additionally, the CBR3 loop of PTEN-L no longer showed a decrease in exchange when bound to membranes.\n\n![Changes in HDX observed on PTEN-L-6A in the presence of PIP~2~-containing vesicles\\\nThe N-terminal extension is shown as a cartoon, with the MBH as a cylinder. Areas with no peptide coverage are shown as white (N/A).](bj4730135fig4){#F4}\n\nDISCUSSION\n==========\n\nPTEN plays important tumour-suppressing roles as both a lipid and a protein phosphatase \\[[@B8]\\], and the membrane permeability of PTEN-L presents an opportunity to commandeer these activities as a therapeutic agent. We have shown that phosphorylation of the C-terminal tail switches PTEN\\'s substrate specificity and that a helix within the N-terminal extension of PTEN-L markedly alters the enzyme kinetics on membranes. Development of PTEN-L as a therapeutic agent will require consideration of these two disordered segments and the effect they have on ability of PTEN to function as a tumour suppressor.\n\nWe find that fully phosphorylated PTEN is completely inactive and incapable of binding membranes. However, selective dephosphorylation of only two residues, Thr^366^ and Ser^370^, causes the active site of PTEN to become partially exposed, allowing cytosolic PTEN to dephosphorylate its growing list of soluble substrates (both phosphoserine/threonine and phosphotyrosine residues) \\[[@B4],[@B45],[@B46]\\], including CREB \\[[@B47]\\], IRS1 \\[[@B8]\\], Shc \\[[@B48]\\], components of the MAPK pathway \\[[@B49]\\] and Dishevelled \\[[@B50]\\]. Furthermore, this selective activity may have an impact on an emerging role of PTEN in dephosphorylating soluble protein-bound PIP~3~, such as PIP~3~ bound to SF-1 within the nucleus \\[[@B51]\\]. Our data indicate that phosphorylation of Thr^366^ and Ser^370^ alone is not sufficient to occlude the active site, as Phos-PTEN-4A was still capable of dephosphorylating soluble substrates. HDX--MS suggests that dephosphorylation of Thr^366^ and Ser^370^ results in localized conformational changes; however, these localized changes are apparently not great enough to have been detected by using access to alkaline phosphatase as a measure of conformational change \\[[@B35]\\].\n\nBy using HDX--MS to map both the membrane-binding interface and the interface between the phosphorylated C-terminal tail with the rest of PTEN, we determined the structural basis for the substrate switch of PTEN. The phosphorylated tail interacts predominantly with the N-terminal PBM, the active site of the phosphatase domain and the C\u03b12 loop of the C2 domain. Previous studies have determined the CBR3 loop of the C2 domain as the binding partner for the phosphorylated tail, through either electrostatic \\[[@B35],[@B52]\\] or non-electrostatic cryptic binding motifs \\[[@B14],[@B37]\\]. However, these studies are based on the mutagenesis or deletion of the CBR3 loop. Our HDX--MS experiments, carried out on the native protein, are not consistent with the phosphorylated tail interacting with CBR3. This is in agreement with a previous study that highlighted the need for an intact active site and PBM to allow for the intramolecular binding event \\[[@B22]\\]. Here, HDX shows that the mechanism for inhibition involves occlusion of the active site by the phosphorylated tail, preventing substrate catalysis, rather than the prevention of PTEN membrane binding through occlusion of the CBR3 loop. The proximity of pThr^366^ and pSer^370^ to the active site may explain their propensity for auto-dephosphorylation \\[[@B38]\\].\n\nOur results also provide a comprehensive picture of how PTEN binds to membranes. For membrane binding and PTEN activity assays, we used a complex mixture of lipids that mimics the composition of the plasma membrane \\[[@B41]\\], as studies have shown that altering the relative proportions of various lipid moieties can effect both PTEN\\'s ability to interact with the membrane and its subsequent lipid phosphatase activity \\[[@B35],[@B44]\\]. Attention has focused on the CBR3 loop as the dictating element of membrane interaction \\[[@B9],[@B10],[@B15],[@B35],[@B37],[@B52]\\], but we observe a more complex picture, with very large reductions in HDX throughout the phosphatase domain and similar reductions in HDX in two of the three CBR loops, as well as in the C\u03b12 loop. Mutations in the C\u03b12 loop decrease activity *in\u00a0vitro* and membrane localization in cells \\[[@B13],[@B53]\\]. Within the phosphatase domain, our data highlight the membrane-binding contribution of the 'arginine loop', which forms a positively charged patch, composed of Arg^41^, Arg^47^, Arg^74^ and Lys^80^. Finally, although both the WPD and the TI loops are protected by membrane binding, the P loop, which contains the catalytically crucial cysteine, is not. These disparate regions of PTEN, i.e. the arginine loop, the TI loop, the CBR3 loop, the C\u03b12 loop and the PBM, have all been implicated in membrane interaction as modelled by molecular dynamic simulations \\[[@B54]\\]. Together, our results indicate that a large surface of PTEN interacts with the membrane.\n\nFrom our results, we propose that PTEN can exist in three distinct activation states ([Figure 2](#F2){ref-type=\"fig\"}D) depending on the phosphorylation status of the tail: a fully phosphorylated state that is inactive against both lipid and protein substrates, a partially dephosphorylated state (lacking phosphates on Ser^366^ and Thr^370^), active only against soluble substrates including phosphoproteins, and a fully dephosphorylated state that is active against both soluble and liposome-incorporated substrates. The activation states are mediated by intramolecular inhibitory interactions between the tail and the C2 and phosphatase domains that occlude both the membrane-binding region as well as the active site.\n\nWhether these three activation states are applicable to PTEN-L is not yet clear. PTEN-L is an N-terminally extended translational variant of PTEN, which was shown to be capable of exiting the cell and entering a neighbouring cell, down-regulating their p-Akt levels \\[[@B29]\\] and possibly localizing to mitochondria to regulate cytochrome *c* oxidase activity \\[[@B30]\\]. We have found that the N-terminal elongation of PTEN-L is not entirely devoid of secondary structure and that an ordered region, most likely an \u03b1-helix, exists between residues 151-L and 174-L. This region resembles the transmembrane S4-helix of voltage-sensitive phosphatase Ci-VSP, a PTEN family member. Upon membrane depolarization, the S4-helix changes conformation, causing the rest of the enzyme (which shares strong similarity with PTEN) to adopt an active conformation on the membrane \\[[@B55]\\]. The exact mechanism of how this conformational change is communicated to the rest of the enzyme is not well understood, although part of the linker, corresponding in sequence to the PBM (residues Arg^13^/Lys^14^ in PTEN), may interact with the TI loop of the phosphatase domain, thereby acting as a gate for activity \\[[@B56]\\].\n\nUsing HDX--MS, we have shown that this 151-L--174-L helix (MBH) in PTEN-L is strongly protected by liposomes, suggesting an interaction with the membrane. The MBH alters both the interfacial kinetics of the enzyme and the protein/membrane interface ([Figures 3](#F3){ref-type=\"fig\"}B--[3](#F3){ref-type=\"fig\"}E and 4). The presence of the MBH causes PTEN-L to function on membranes in a 'scooting' mode rather than a 'hopping' mode that is characteristic of PTEN, i.e. PTEN-L seems to remain attached to the membrane, diffusing two-dimensionally, rather than readily diffusing between separate membranes like PTEN. Deletion of the MBH caused PTEN-L to revert to the same hopping behaviour as PTEN. PTEN-L\\'s scooting mechanism may have important cellular implications, such as causing an enrichment of PTEN-L on certain membranes. Whether the increase in affinity of PTEN-L for the membrane imparted by MBH is sufficient to counter the decrease in membrane affinity observed on phosphorylation of the C-terminus of PTEN is not known. Whether the MBH is a transmembrane helix or lies against the membrane surface, as recently proposed for a helix found in the PBM \\[[@B57]\\], has yet to be determined ([Figure 5](#F5){ref-type=\"fig\"}). Our data also show that the PBM is structurally altered and stabilized in PTEN-L compared with PTEN, even in the absence of membranes, suggesting that there may be a network of interactions that link the MBH to the PBM and reduces the role of the PBM in the membrane binding of PTEN-L. Furthermore, the CBR3 and CBR1 loops of the PTEN-L C2 domain do not engage membranes, suggesting that the catalytic domain may be sitting on the membrane differently than PTEN.\n\n![Possible membrane-binding mode of PTEN-L\\\nThe MBH of PTEN-L may interact with the membrane surface or may be a transmembrane helix. This interaction changes the interfacial catalytic mechanism from a 'hopping' mode, characteristic of PTEN to a 'scooting' mode, characteristic of PTEN-L.](bj4730135fig5){#F5}\n\nPTEN-L and PTEN exhibit complex mechanisms of interdependent changes in conformation, membrane binding and substrate specificity that are mediated by two intrinsically disordered tails. A combination of differential phosphorylations of the C-terminal tail and extension of the N-terminus provides a large repertoire of active states with different membrane-binding properties. The use of recombinant PTEN-L has been proposed as a potentially novel therapeutic strategy to introduce PTEN activity in PTEN-deficient cancers and the present study underscores the complexity of regulation of PTEN and PTEN-L. Further work on the mechanistic aspects that underpin PTEN and PTEN-L function is essential to enable their use as a therapeutic agent.\n\nWe thank Mark Skehel, Sarah Maslen, Sew-Yeu Peak-Chew and Farida Bergum for help with the HDX--MS set-up. We also thank Chris Johnson for his expertise and help with the SEC--MALS experiments.\n\nAUTHOR CONTRIBUTION\n===================\n\nGlenn Masson designed and performed experiments and wrote the manuscript. Olga Perisic also performed experiments, constructed plasmids and helped to write the manuscript. John Burke aided in experiment design. Roger Williams designed experiments and helped to write the manuscript.\n\nFUNDING\n=======\n\nThis work was supported by the European Molecular Biology Organization \\[grant number ALTF268-2009 (to J.E.B.)\\]; the British Heart Foundation \\[grant number PG11/109/29247 (to R.L.W)\\] and the Medical Research Council \\[file reference U105184208\\].\n\nCBR\n\n:   calcium-binding region\n\nCi-VSP\n\n:   *Ciona intestinalis* Voltage Sensitive Phosphatase\n\nGSK3\u03b2\n\n:   glycogen synthase kinase 3\u03b2\n\nHDX\n\n:   hydrogen/deuterium exchange\n\nIP~4~\n\n:   inositol tetrakisphosphate\n\nIRS1\n\n:   insulin receptor substrate 1\n\nLC-MS\n\n:   liquid chromatography-mass spectrometry\n\nMBH\n\n:   membrane-binding helix\n\nPBM\n\n:   PIP~2~-binding motif\n\nPhospho-wtPTEN\n\n:   phosphorylated wild-type PTEN\n\nPI3K\n\n:   phosphoinositide 3-kinase\n\nPIP~2~\n\n:   phosphatidylinositol 4,5-bisphosphate\n\nPIP~3~\n\n:   phosphatidylinositol 3,4,5-trisphosphate\n\nPhos-PTEN-2A/4A\n\n:   phosphorylated PTEN-2A/4A mutant\n\nPS\n\n:   phosphatidylserine\n\nPTEN-\u0394Tail\n\n:   truncated PTEN lacking the tail\n\nPTP\n\n:   protein tyrosine phosphatase\n\nPTEN\n\n:   phosphatase and tensin homologue deleted on chromosome 10\n\nPTEN-L\n\n:   PTEN-long\n\nSEC--MALS\n\n:   size-exclusion chromatography--multi-angle light scattering\n\nTBST\n\n:   TBS/Tween 20 solution\n\nTCEP\n\n:   tris-(2-carboxyethyl)-phosphine\n\nTEV\n\n:   tobacco etch virus\n\nTPTE\n\n:   transmembrane phosphatase with tensin homology\n\nUPLC\n\n:   ultra-performance liquid chromatography\n\nwtPTEN\n\n:   dephosphorylated wild-type PTEN\n\n[^1]: Present address: Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, Canada, V8P 5C2.\n"}
+{"text": "INTRODUCTION\n============\n\nParallel-polarized light (PPL) photography is a method that can objectively evaluate reflections from the skin surface[@B1]. A preliminary study with various skin diseases indicated that PPL photography images taken with green light emitting diodes (LEDs) might be useful for analyzing specific diseases such as atopic dermatitis, rosacea, and xerotic dermatitis[@B2]. Based on our previous study, this study was designed to statistically correlate the CIELAB coordinates with rosacea severity. The L^\\*^ coordinate represents brightness, and the a^\\*^ and b^\\*^ coordinates represent the red to green axis and the blue to yellow axis, respectively[@B3]. In the future, clinical research on rosacea could apply this method as an objective assessment of disease severity.\n\nMATERIALS AND METHODS\n=====================\n\nPatients\n--------\n\nThis study was approved by the Institutional Review Board (IRB) of Korea University Anam Hospital (IRB ED13197). A total of 49 patients who visited the clinic between June and July 2015 were enrolled. Twenty-five were men and 24 were women. The mean age was 50.7\u00b114.3 years (range, 17\\~78 years). Patients with severe medical conditions such as malnutrition, thyroid disorders, or malignancy were excluded. Patients who were using a medication that might affect skin measurements such as diuretics, steroids, retinoids, or H2 antihistamines were excluded. We also excluded pregnant women.\n\nMaterials\n---------\n\nA high intensity power LED was used as a light source. A circuit was assembled with a radiator and constant current LED driver using white (PP00W-8L61-ESBI; Photron Co. Ltd., Anseong, Korea) and green (PP525-8L61-ESBI; Photron Co. Ltd.) LEDs. The circuit board assembly was operated by a 5 V power supply. A rechargeable battery with a power switch was installed to allow easy access and free movement. A rotatable polarizing filter (PL filter; Kenko Co. Ltd., Tokyo, Japan) was attached at the exit slit of the LED lamp, which enabled control of the polarization direction. A digital single-lens reflex camera (EOS-500D; Canon Inc., Tokyo, Japan) equipped with a macro lens (SP MF 90 mm F/2.8 Di Macro 1:1, Tamron Co. Ltd., Saitama, Japan) was used, and a rotatable polarizing filter was placed over the camera lens. The distance between the camera and the subject was maintained constant by using a manual body focusing technique.\n\nPPL photographic images using LED light\n---------------------------------------\n\nThe polarizing filters of both the LED lamp and camera were aligned in the same direction. The white and green LED illuminators were attached on each side of the camera at a 45 degree angle as shown in [Fig. 1](#F1){ref-type=\"fig\"}. The white balance of the camera was set to daylight on manual mode with F-number 2.5, shutter speed 1/60 s, and ISO 1600. The subject was placed about 9 cm from the camera to maintain focused images. Only the green or white LED illuminator was turned on in a darkroom to minimize environmental impacts on the images. The PPL images of the skin lesion and adjacent normal appearing skin were taken at the same time.\n\nThe analysis of PPL photographic images\n---------------------------------------\n\nThe PPL digital images were represented using 8-bit integer numbers of sRGB coordinates using a graphic program (Adobe Photoshop Element; Adobe System Inc., San Jose, CA, USA). These code values were converted to CIELAB coordinates using a spreadsheet tool (Excel 2010; Microsoft Corp., Redmond, WA, USA). The CIELAB coordinates of L^\\*^, a^\\*^, and b^\\*^ allowed quantitative characterizations of changes in color.\n\nClinical assessment\n-------------------\n\nThe clinical severity was assessed with a 5 point investigator\\'s global assessment (IGA) scale. All patients were also evaluated with the xerosis severity scale (from 0 to 6)[@B4]. We also used the standard grading system proposed by the National Rosacea Society Expert Committee to evaluate rosacea severity[@B5]. Relationships between the disease severity index, xerotic severity scale, age, and CIELAB coordinates were examined ([Table 1](#T1){ref-type=\"table\"}). We also analyzed the data according to sex and rosacea type ([Table 2](#T2){ref-type=\"table\"}, [3](#T3){ref-type=\"table\"}). Patients were classified according to clinical appearance into the erythematotelangiectatic rosacea (ETR) and papulopustular rosacea (PPR) subtypes.\n\nStatistical analysis\n--------------------\n\nThe data were analyzed using the statistical software IBM SPSS Statistics ver. 20.0 for Windows (IBM Co., Armonk, NY, USA). All values were evaluated using both Pearson correlation and Spearman correlation analyses as parametric and non-parametric tests, respectively. The ordinal scales such as the xerosis severity scale were compared between groups using the Mann-Whitney U-test. A *p*-value \\<0.05 was considered statistically significant.\n\nRESULTS\n=======\n\nOverall results\n---------------\n\n[Table 1](#T1){ref-type=\"table\"} shows the relationships between the disease severity index, xerotic severity scale, age, and CIELAB coordinates. The severity of rosacea using the IGA score had a significant positive correlation with the standard grading system proposed by the National Rosacea Society Expert Committee (r=0.79791, *p*\\<0.001). For the rosacea skin lesions, the standard grading system showed negative correlations with the L^\\*^ (r=\u22120.33348, *p*=0.0192) and b^\\*^ coordinates (r=\u22120.36298, *p*=0.0104), and a positive correlation with the a^\\*^ coordinate (r=0.34004, *p*=0.0168) using green PPL images. In the same environment, there was no correlation for the normal skin area for L^\\*^ (r=0.10488, *p*=0.4733), a^\\*^ (r=\u22120.11061, *p*=0.4493), or b^\\*^ (r=0.06473, *p*=0.6586).\n\nFor white PPL images, the xerosis severity scale showed a positive correlation with L^\\*^ (r=0.40175, *p*=0.0042) and a negative correlation with b^\\*^ (r=\u22120.29714, *p*=0.0381) for rosacea lesions. The same correlation was found for normal skin using white PPL images with L^\\*^ (r=0.52382, *p*=0.0001), b^\\*^ (r=\u22120.33021, *p*=0.0205).\n\nAge showed a negative correlation with the L^\\*^ (r=\u22120.49428, *p*=0.0003) and b^\\*^ coordinates (r=\u22120.46273, *p*=0.0008), and a positive correlation with the a^\\*^ coordinate (r=0.49321, *p*=0.0003) for the rosacea lesion using green PPL images. A similar correlation was found for age and normal skin area using green PPL images, with L^\\*^ (r=\u22120.33514, *p*=0.0186), a^\\*^ (r=0.34678, *p*=0.0146), and b^\\*^ (r=\u22120.30251, *p*=0.0346).\n\nAnalysis by sex\n---------------\n\nThe analysis based on sex is shown in [Table 2](#T2){ref-type=\"table\"}. When analyzed separately by sex, the severity of rosacea using the IGA score had significant positive correlations with the standard grading system in both male (r=0.83617, *p*\\<0.0001) and female patients (r=0.75593, *p*\\<0.001). The standard grading system for rosacea among the male patients had negative correlations with the L^\\*^ (r=\u22120.50907, *p*=0.0094) and b^\\*^ coordinates (r=\u22120.54441, *p*=0.0049), and a positive correlation with the a^\\*^ coordinate (r=0.52085, *p*=0.0076) using green PPL images. However, the standard grading system for the female patients did not show any significant correlation with the L^\\*^ (r=\u22120.12598, *p*=0.5575), a^\\*^ (r=0.14081, *p*=0.5117), or b^\\*^ coordinate (r=\u22120.14081, *p*=0.5117).\n\nThe age of the male patients was negatively correlated with the L^\\*^ (r=\u22120.54783, *p*=0.0046) and b^\\*^ coordinates (r=\u22120.50202, *p*=0.0106) and positively correlated with the a^\\*^ coordinate (r=0.53898, *p*=0.0054) using green PPL images. In the same environment, no significant relationship was found between the age of the female patients and the CIELAB coordinates, with L^\\*^ (r=\u22120.33660, *p*=0.1078), a^\\*^ (r=0.33355, *p*=0.1112), and b^\\*^ (r=\u22120.32920, *p*=0.1162).\n\nAnalysis of the ETR and PPR subtypes\n------------------------------------\n\nThe analysis based on the clinical subtypes is shown in [Table 3](#T3){ref-type=\"table\"}. The IGA score of the PPR type had negative correlations with the L^\\*^ (r=\u22120.55594, *p*=0.0485) and the b^\\*^ coordinate (r=\u22120.55594, *p*=0.0485) for the rosacea lesions using green PPL images. The standard grading system showed negative correlations with the L^\\*^ (r=\u22120.67862, *p*=0.0108) and the b^\\*^ coordinate (r=\u22120.67862, *p*=0.0108) and a positive correlation with the a^\\*^ coordinate (r=0.64194, *p*=0.0180) for the green LED only with the PPR type. Age also had negative correlations with the L^\\*^ (r=\u22120.66391, *p*=0.0133) and b^\\*^ coordinates (r=\u22120.68871, *p*=0.0092) and a positive correlation with the a^\\*^ coordinate (r=0.72452, *p*=0.0051) for the green LED with the PPR type.\n\nThe xerotic severity scale showed a positive correlation with L^\\*^ (r=0.36709, *p*=0.0276) and a negative correlation with the b^\\*^ coordinate (r=\u22120.33068, *p*=0.0489) in the white LED only for the ETR type. The xerotic severity scale had the same correlations with L^\\*^ (r=0.49009, *p*=0.0024) and b^\\*^ (r=\u22120.43852, *p*=0.0075) for the normal skin area using white LEDs with the ETR type. White LEDs showed a higher L^\\*^ value for ETR than for PPR, which means more brightness, for both the lesional (*p*=0.0358) and normal skin (*p*=0.0092). It was notable that the xerosis severity scale score was higher, indicating drier, in the ETR type than the PPR type (*p*=0.0287).\n\nDISCUSSION\n==========\n\nMany previous publications evaluated skin lesions based on direct visualization and standard flash photography. This kind of assessment often relies on some degree of subjective interpretation and different parameters affect any objective measure. Direct visualization and clinical photography can have different evaluations depending on the observer. Conventional photography may also be taken differently due to inconsistent framing, varying angles between the subject and the camera, and changes in exposure settings.\n\nIn order to improve diagnostic evaluation, various non-invasive devices such as ultraviolet light photography, polarized light photography, reflectance spectroscopy, dermoscopy, and confocal scanning laser microscopy have been developed and are widely used clinically[@B6]. Among these, polarized light photography is a well-known technique that uses the polarization of the light reflected from skin tissues. It can be applied to evaluate skin diseases such as acne vulgaris, photoaging, subclinical actinic keratosis, nonmelanoma skin cancer, subclinical levels of skin irritant reactions, and psoriasis[@B7][@B8][@B9][@B10].\n\nLight reflected from the skin can be divided into regular reflectance related to the skin surface and \"back-scattered\" light from structures within the skin tissue. The regular reflectance is a reflection at the stratum corneum-air interface. The \"back-scattered\" light originates from the papillary dermis. Light particles reflected from the stratum corneum-air interface have the same polarization direction of the incident light, whereas particles reflected from the papillary dermis have the scattering direction of polarization. Therefore, an observer can use the polarizing light source to selectively examine the surface or subsurface components of the skin[@B11][@B12]. When a polarizing filter attached to the camera is aligned parallel to the direction of polarization of the LED, information about the skin surface except the \"back-scattered\" light component can be obtained. Information on the subsurface component can be obtained by using perpendicular polarized photography.\n\nTherefore, parallel polarized photography can reduce intracutaneous and subsurface details such as pigmentation, vascularity, and color, while emphasizing skin surface detail such as texture and elevation scaling[@B6]. Despite these advantages, PPL images are not widely used in clinical practice due to challenges in quantitative assessment. Previous studies indicated that the CIELAB values from the PPL photographic images can serve as a quantitative indicator of skin surface reflection[@B13]. In our previous study, the degree of severity and dryness of skin had significant correlations with CIELAB values from the images taken from PPL photography using green LEDs, particularly those with atopic dermatitis, rosacea, and xerotic dermatitis[@B2].\n\nUnlike other noninvasive equipment like ultraviolet photography or dermoscopy, the CIELAB coordinates obtained from PPL photography are relatively consistent parameters regardless of the technique used or the physician\\'s experience. This study was conducted on patients with rosacea. Erythematous lesions appear black, while normal skin appears white using light emitting from a green LED[@B10]. The scale of whiteness is especially clearly visualized by PPL photography[@B10].\n\nSkin color is a result of complex interactions of skin microstructures and chromophores with the incipient light. One of the most important chromophores of human skin is hemoglobin, which is located in the lumen of vessels. If the skin color measuring device is pushed against the skin, then the amount of hemoglobin changes and skin color as well. Therefore, most contact type colorimetric devices have drawbacks that cannot be overcome easily. In this study, we used a technique modified from colorimetric photography, a non-contact type measurement that does not affect the vessels and, hence, does not affect erythema.\n\nWe observed significant relationships between the standard grading system for rosacea and the CIELAB coordinates, especially using the green PPL images ([Table 1](#T1){ref-type=\"table\"}). However, male and female patients also displayed important differences. Only the male rosacea patients had significant results on the green LED ([Table 2](#T2){ref-type=\"table\"}). Also, in the ETR and PPR subject groups, the coordinates of the L^\\*^, a^\\*^, and b^\\*^ had significant relationships with the standard grading system on the green LED only for the PPR type. No significant correlation with the ETR type was observed. The ETR type showed significant associations with both the xerosis severity scale and the L^\\*^ and b^\\*^ coordinates using the white LED ([Table 3](#T3){ref-type=\"table\"}).\n\nOnly male subjects had a significant association with the severity score of the disease and the L^\\*^, a^\\*^, and b^\\*^ coordinates, which may be due to a more severe form of the disease in male patients. This may be related to clinical evidence that rosacea occurs more frequently in middle aged women, but severe symptoms appear more frequently in men. In addition, most patients enrolled in this study had mild to moderate disease. If additional patients or patients with more severe complications are enrolled, the standard grading system in female patients may become significantly correlated with the CIELAB coordinates. The significant relationship between the standard grading system for only the PPR patients and the CIELAB coordinates may also be due to more severe forms of the PPR type enrolled in the study. The standard grading system of the ETR subjects could have a significant correlation if more severe cases of the ETR form were recruited.\n\nThe clinical differences between the PPR and ETR subtypes appeared to be relevant in the pathophysiology. As shown on [Table 3](#T3){ref-type=\"table\"}, dryness of the skin could be an important factor in the pathophysiology of the ETR type while age is more involved with the PPR type. The results in [Table 3](#T3){ref-type=\"table\"} indicate that ETR subtypes are observed more precisely using white LEDs and that of PPR subtypes were better evaluated using green LEDs. ETR subtypes were more associated with dryness and brighter (a higher L^\\*^ value) reflections than PPR subtypes. The results of this study could be used in clinical differentiation of the two subtypes, resulting in more active and rapid treatment. Further studies should examine the mixed type of the PPR and ETR subtypes.\n\nSun exposure, temperature changes, cytokines and chemokines, various microbial agents, neuroimmune and immune defense connections, angiogenesis, lymphangiogenesis, and other known factors contribute to the pathophysiology of rosacea[@B14]. Other factors involved in the pathophysiology of rosacea should be studied further. The results of this study will help us to detect pathophysiological differences between certain subtypes of rosacea.\n\nPPL photography is a method that can objectively evaluate reflections from the skin surface. The CIELAB coordinates from green PPL photography showed significant correlations with the severity index of rosacea. It can be used in clinical settings and may contribute to revealing the pathophysiology of rosacea. In addition, it seems that subtypes of rosacea, e.g., ETR and PPR, are distinct entities visually and optically.\n\nSupported by the Korea Research Foundation Grant funded by the Korean Government (NRF-2012R1A1A2044700).\n\n**CONFLICTS OF INTEREST:** The authors have nothing to disclose.\n\n![A schematic diagram showing the arrangement of the equipment. LED: light emitting diode, PL: polarized light.](ad-29-167-g001){#F1}\n\n###### Overall results\n\n![](ad-29-167-i001)\n\n  Parameter                        Normal skin area                 Lesion                                                           \n  -------------------------------- -------------------------------- -------------------------------- ------------------------------- --------------------------------\n  Standard grading system                                                                                                            L^\\*^ (r=\u22120.33348, *p*=0.0192)\n  a^\\*^ (r=0.36625, *p*=0.0096)    a^\\*^ (r=0.34004, *p*=0.0168)                                                                     \n  b^\\*^ (r=\u22120.36298,*p*=0.0104)                                                                                                      \n  Xerosis severity scale           L^\\*^ (r=0.52382, *p*=0.0001)                                     L^\\*^ (r=0.40175, *p*=0.0042)   \n  b^\\*^ (r=\u22120.33021, *p*=0.0205)   b^\\*^ (r=\u22120.29714, *p*=0.0381)                                                                    \n  Age                                                               L^\\*^ (r=\u22120.33514, *p*=0.0186)                                   L^\\*^ (r=\u22120.49428, *p*=0.0003)\n  a^\\*^ (r=0.34678, *p*=0.0146)    a^\\*^ (r=0.49321, *p*=0.0003)                                                                     \n  b^\\*^ (r=0.32456, *p*=0.0229)    b^\\*^ (r=\u22120.30251, *p*=0.0346)   b^\\*^ (r=\u22120.46273, *p*=0.0008)                                   \n\nLED: light emitting diode.\n\n###### Analysis by men and women\n\n![](ad-29-167-i002)\n\n  Parameter                        Normal skin area                 Lesion            \n  -------------------------------- -------------------------------- -------- -------- ------\n  Standard grading system          Male                             Male     Male     Male\n  L^\\*^ (r=\u22120.50907, *p*=0.0094)                                                      \n  a^\\*^ (r=0.47667, *p*=0.0160)    a^\\*^ (r=0.52085, *p*=0.0076)                      \n  b^\\*^ (r=\u22120.54441, *p*=0.0049)                                                      \n  Female                           Female                           Female   Female   \n  b\\* (r=\u22120.51134, *p*=0.0107)     b^\\*^ (r=\u22120.49652, *p*=0.0136)                     \n  Age                              Male                             Male     Male     Male\n  L^\\*^ (r=\u22120.54783, *p*=0.0046)                                                      \n  a^\\*^ (r=0.53898, *p*=0.0054)                                                       \n  b^\\*^ (r=\u22120.50202, *p*=0.0106)                                                      \n  Female                           Female                           Female   Female   \n  b^\\*^ (r=0.47594, *p*=0.0187)                                                       \n\nLED: light emitting diode.\n\n###### Analysis by ETR and PPR subtypes\n\n![](ad-29-167-i003)\n\n  Parameter                        Normal skin area                 Lesion         \n  -------------------------------- -------------------------------- -------- ----- -----\n  Standard grading system                                                          PPR\n  L^\\*^ (r=\u22120.67862, *p*=0.0108)                                                   \n  a^\\*^ (r=0.64194, *p*=0.0180)                                                    \n  b^\\*^ (r=\u22120.67862, *p*=0.0108)                                                   \n  IGA                                                                              PPR\n  L^\\*^ (r=\u22120.55594, *p*=0.0485)                                                   \n  b^\\*^ (r=\u22120.55594, *p*=0.0485)                                                   \n  Age                                                                              PPR\n  L^\\*^ (r=\u22120.66391, *p*=0.0133)                                                   \n  a^\\*^ (r=0.72452, *p*=0.0051)                                                    \n  b^\\*^ (r=\u22120.68871, *p*=0.0092)                                                   \n  Xerosis severity scale           ETR                                       ETR   \n  L^\\*^ (r=0.49009, *p*=0.0024)    L^\\*^ (r=0.36709, *p*=0.0276)                   \n  a^\\*^ (r=\u22120.15031, *p*=0.3816)   a^\\*^ (r=0.15445, *p*=0.3684)                   \n  b^\\*^ (r=\u22120.43852, *p*=0.0075)   b^\\*^ (r=\u22120.33068, *p*=0.0489)                  \n  PPR                              PPR                                             \n  L^\\*^ (r=0.24398, *p*=0.4218)    L^\\*^ (r=0.19518, *p*=0.5228)                   \n  a^\\*^ (r=-0.53675, *p*=0.0586)   a^\\*^ (r=\u22120.09759, *p*=0.7511)                  \n  b^\\*^ (r=\u22120.39036, *p*=0.1873)   b^\\*^ (r=\u22120.43916, *p*=0.1333)                  \n\nETR: erythematotelangiectatic rosacea, PPR: papulopustular rosacea, LED: light emitting diode, IGA: investigator\\'s global assessment.\n"}
+{"text": "INTRODUCTION\n============\n\nNon-LTR retrotransposons (NLRs) \\[retrotransposons lacking long terminal repeats (LTRs)\\], are a class of transposable elements that are mobilized by reverse transcription of an RNA intermediate ([@b1]). Multiple copies of NLRs inhabit variable fractions of the genome of most eukaryotes. Whereas most copies are defective, some are functional and their retrotransposition can result in several types of genetic effects including mutations associated with diseases in humans and mice ([@b2]). The *Drosophila* NLR *I factor* and the mammalian NLR L1 ([@b3]) provide well-known examples of the impact of NLR mobilization.\n\nMost NLRs contain two open reading frames, ORF1 and ORF2, which are translated from a bicistronic messenger RNA that presumably also serves as a template for retrotransposition ([@b4]--[@b6]). The sequence of the ORF2 product (ORF2p) is relatively well conserved among NLRs. ORF2p is involved in reverse transcription and integration, and its endonuclease and reverse-transcriptase activities have been demonstrated *in vitro* for different elements ([@b7]--[@b12]). The sequence of the ORF1 product (ORF1p) is, in contrast, poorly conserved in evolution. ORF1p of human and mouse L1 are expressed in some carcinoma cells and can be isolated as a 40 kDa full-size protein (p40) which forms, in association with L1 RNA, high molecular weight cytoplasmic ribonucleoprotein (RNP) complexes ([@b13],[@b14]). *I factor* ORF1p, expressed in bacteria or insect cells, shows both DNA and RNA binding properties and accelerates the annealing of complementary single-strand oligonucleotides without sequence specificity ([@b15]), similar to mouse L1 ORF1p when expressed in those heterologous systems ([@b16]--[@b18]). These *in vitro* experiments suggested that ORF1p from NLRs could act as a chaperone.\n\nIt is now known that the integrity of both ORF1 and ORF2 is required for retrotransposition ([@b19]--[@b21]), but our knowledge is limited on how and when the products of these ORFs interact with the RNA transposition intermediate during the mobilization process. The model NLR *I factor* allows the study of retrotransposition *in vivo*, making possible the tracking of specific transposition products in *Drosophila*. *I factor* transposes specifically in the oocytes of females (called SF) issued from mating males carrying active *I factor*s (inducer or I strains) with permissive females lacking active *I factor*s (reactive or R strains). A proportion of the eggs laid by the SF females fail to hatch. The severity of this sterility correlates with the frequency of transposition of *I factor*s \\[reviewed in ([@b22])\\].\n\nWe have previously described the localization pattern of ORF1p produced by active *I factors* in the germ cells of SF females, showing that it is correlated to the element retrotransposition ([@b23]). We show here that *I factor* RNA and ORF1p generated by active *I factor*s during mobilization co-localize with a specific cytoplasmic pattern (Loc+), exclusively in permissive oocytes. By analyzing the expression patterns of mutated/deleted *I factor*s in transgenic females, we show that ORF2p is not involved in the Loc+ patterning and that localization of ORF1p/RNA requires an RNA *cis*-acting signal located in the sequence of ORF2. Finally, the expression and complementation analyses of an element deleted in ORF1 indicate that ORF1p is required (either in *cis* or in *trans*) for efficient Loc+ patterning of the RNA. This result suggests that some chaperone-like properties of ORF1p described *in vitro* ([@b15],[@b17]), may apply *in vivo* to force proper folding of *I factor* RNA and enhance its targeting to the nuclear vicinity, favoring the efficiency of downstream steps of the retrotransposition process.\n\nMATERIALS AND METHODS\n=====================\n\nFly stocks\n----------\n\nThe JA (*white* and *yellow*) stock is a reactive (R) strain, permissive for *I factor* transposition. The strain *w~1118~* is a standard inducer (I) stock. The transgenic lines HT1 and HT2 carrying several copies of the marked I-HAO1 element ([@b23]) were also used as I stocks. Stocks HH16 and Is37 are transgenic lines for constructs *hs*ORF1 and I-\u0394Asu, respectively, and were established after P-element mediated transformation of the reactive strains *w*K and Cha ([@b24]). All strains are M in the P--M system of hybrid dysgenesis, thus transposition events cannot result from P-element activity.\n\nPlasmid constructs\n------------------\n\nAll constructs with internal deletions of the ORF2 region were derived from the element I-HAO1, a functional I element marked with the HA tag at the beginning of ORF1 in construct pCaSpeR/I-HAO1 ([@b23]), by performing restriction digests according to the sites indicated in [Figure 2](#fig2){ref-type=\"fig\"}. Nucleotide positions and the corresponding sequence corrections were described previously ([@b25],[@b26]). The integrity of the junction points was checked by sequencing. To obtain construct A, we cloned the thymidine kinase terminator, contained within an SmaI--EcoRI fragment from plasmid pBTK2, at the end of ORF1 in construct pCaSpeR/I-HAO1 cut with HpaI--EcoRI ([Figure 2](#fig2){ref-type=\"fig\"}). To obtain construct pCaspeR/I-HAO1-fsO2, an 8 nt oligomer was ligated to EcoRV-digested pCAspeR/IHAO1. The I-\u0394Asu construct contains an in-frame AsuII deletion within ORF1 and was derived from pI954, a plasmid containing a complete *I factor* cloned into vector pUC*hsneo*. Construct *hs*ORF1 was described previously ([@b24]). To obtain construct DF313/loc, a *lacZ*/*I factor* fusion, we generated the 552 nt localization fragment by PCR using primers 5\u2032-GTATCTAGAACTTAGCTCAGCAC-sense and 5\u2032-GACTAGTGGCTTGATGTATGCGG-antisense, digested it with XbaI/SpeI and cloned it at the 3\u2032 end of a \u03b2*-gal* gene in plasmid DF313 cut with SpeI. In plasmid DF313, the promoter of the maternal alpha4-tubulin gene 67C drives the expression of the \u03b2*-gal* gene (D. Ferrandon, personal communication).\n\nTransformations\n---------------\n\nAll constructs derived from I-HA-O1 were cloned into the pCaSpeR4 transformation vector containing the mini-*white*+ gene as a selection marker ([@b27]). They were introduced into reactive JA flies (*y*,*w*) by P-element mediated transformation ([@b28]). Transposase activity was provided by the PUC*hs*\u03942-3 helper plasmid (Flybase ID: FBmc-0000938). Several independent transformed lines were generated and three were analyzed for each construct, with the exception of CI21, for which only one line showing a full expression of the transgene was obtained.\n\nComplementation experiments\n---------------------------\n\nWe performed reciprocal crosses between flies from stocks HH16 and Is37. The F1 female progeny underwent two heat-shock treatments of 1 h at 37\u00b0C, 1 day before and 1 h before dissection of the flies. The expression of I-\u0394Asu transcripts in the ovaries of Is37/HH16 hybrids (with and without heat treatment) was specifically detected using a probe corresponding to *I factor* 5\u2032UTR, absent in the *hs*ORF1 construct.\n\nPCR detection of transposed copies\n----------------------------------\n\nThe method for obtaining flies in which the presence of transposed copies was investigated and the method for extracting their DNA was described previously ([@b23]). Briefly, the transgenic element integrated in w flies is linked to the w+ selection marker, while transposed copies are not. Therefore, after allowing a single cycle of transposition in heterozygote transgenic females, we selected w progeny and used PCR to detect transposed copies of the marked element in their DNA. The primers used were 5\u2032-TTACCATACGACGTCCCAGA (sense) which overlaps the HA tag and 5\u2032-GATCAGATCTGATCCTTTTAGA (antisense) which is specific to the frameshift mutation introduced in ORF2. The size of the expected product was 1354 nt.\n\n*In situ* hybridization\n-----------------------\n\nOvaries were dissected in ice cold phosphate-buffered saline and processed for *in situ* hybridization as described previously ([@b29]). The probe consisted of a PCR amplification product of *I factor* 5\u2032UTR purified on a Qiagen column and labeled with digoxygenin using the Boehringer nick-translation kit. Occasionally, we used a similarly prepared probe representing a short segment of ORF1 or ORF2.\n\nImmunofluorescence\n------------------\n\nAntibody staining was performed as described previously ([@b23]). The antibodies used were mouse anti-HA12CA5 (1:500 dilution) (Roche) and FITC-conjugated anti-mouse IgG (1:200 dilution) (Vector) or rat anti-HA high affinity 3F10 (1:500 dilution) (Roche) and FITC-conjugated anti-rat antibodies (1:500 dilution) (Vector).\n\nRESULTS\n=======\n\nORF1p and the *I factor* transcript co-localize temporally and spatially during oogenesis\n-----------------------------------------------------------------------------------------\n\nWe have analyzed the expression pattern of the transcripts produced by active *I factor*s during retrotransposition and compared it with the pattern previously identified for ORF1p \\[([@b23]), and [Figure 1a](#fig1){ref-type=\"fig\"}, panels B and D\\]. Briefly, ORF1p produced in nurse cells is transported to the oocyte cytoplasm from the early stages of oogenesis, where it accumulates at the posterior pole until mid-oogenesis. At this stage ORF1p re-localizes to the oocyte anterior pole. This movement coincides temporally and spatially with the change of microtubule polarity and the migration of the nucleus from the posterior to the anterior-dorsal pole ([@b30],[@b31]). In late developmental stages, coincident with nuclear membrane disassembly and resumption of meiosis I, ORF1p is no longer detected in the oocyte. We synthesized probes representing the *I factor* 5\u2032UTR and performed *in situ* hybridization on the ovaries of SF females resulting from crosses between R females and I males containing several active copies of the *I factor* (either from natural I strains or from transgenic strains containing the functional tagged I-HA-O1 element). As controls, we tested the ovaries of R and I females ([Figure 1a](#fig1){ref-type=\"fig\"}, panels E and F). We detected a strong expression of *I factor* RNA exclusively in SF females ([Figure 1a](#fig1){ref-type=\"fig\"}, panels A and C). This expression pattern was identical to that of ORF1p, i.e. accumulation in the cytoplasm of early oocytes, re-localization from the posterior to the anterior pole at mid-oogenesis and extinction of the signal after stage 10b ([Figure 1a](#fig1){ref-type=\"fig\"}, panels A--D). The co-localization of ORF1p and *I factor* transcripts during oogenesis strongly suggests that both molecular species interact and form RNP complexes during retrotransposition. We defined this dynamic subcellular localization pattern of ORF1p and RNA as the 'Loc+' phenotype.\n\nTo investigate how the different components of the *I factor* (ORF1p, ORF2p and RNA) interact and localize, we generated a frameshift mutation in ORF2 and several deletions downstream of the ORF1 sequence of the I-HA-O1 element ([Figure 2](#fig2){ref-type=\"fig\"}). Using standard procedures, we introduced the mutated elements in the JA reactive strain and established several independent homozygous transgenic lines for each construct. In the following paragraphs we describe the analysis of the expression patterns of ORF1p (by histo-immunofluorescence) and of the transcripts (by *in situ* hybridization) produced by these *I factor*s in a standard permissive background; i.e. in females resulting from crosses of homozygous transgenic males to JA reactive females. We tested several independent transgenic lines per mutant (see Materials and Methods).\n\nThe ORF2 protein is not required for the Loc+ phenotype\n-------------------------------------------------------\n\nWe disabled the translation capacity of ORF2 from its first AUG codon by creating a frameshift mutation near its 5\u2032 end (element I-HAO1-fsO2) ([Figure 1b](#fig1){ref-type=\"fig\"}, panels G and H). Interestingly, for each independent line analyzed, ORF1p and the transcripts presented a Loc+ phenotype. We verified that the frameshift in ORF2 actually abolished the transposition capacity of the element I-HAO1-fsO2. For this, we adapted a PCR-based method ([@b23]) to detect specifically the HA tag from transposed copies in the male progeny of SF females (see Materials and Methods). Whereas 20--70% of the progeny contained transposed copies of the parental I-HA-O1 element, no transposition events were detected in a sample of 300 flies pooled from three independent transgenic lines carrying the I-HAO1-fsO2 element (100 flies/line). These results indicate that: (i) I-HAO1-fsO2 is unable to transpose at a detectable frequency; (ii) functional ORF2 products were not produced from AUG codons located downstream of the first AUG of ORF2, identified as the initiation codon of this ORF ([@b5]); (iii) in the absence of ORF2p, *I factor* RNA and ORF1p are transported and co-localize in oocytes as efficiently as when they are produced by transposing elements.\n\nThe Loc+ phenotype requires *cis*-acting sequences within ORF2\n--------------------------------------------------------------\n\nWe examined the expression pattern of a construct B (see [Figure 2](#fig2){ref-type=\"fig\"}), containing only the 5\u2032UTR, ORF1 and the 539 nt of the 3\u2032end of the element, which includes a putative nucleic acid binding domain at the end of ORF2 ([@b26]), the 3\u2032UTR and four TAA repeats. To assess the effect of the 3\u2032end region in the localization process, we replaced it by the herpes thymidine kinase terminator (*tk*) in construct A ([Figure 2](#fig2){ref-type=\"fig\"}). For both the constructs and the three independent transgenic lines analyzed per construct, ORF1p was detected in the cytoplasm of nurse cells and oocytes. However, the intensity of the signals was weaker than that observed in females exhibiting a Loc+ phenotype. This signal was detected only at early stages of oogenesis (never later than stages 5--8). Interestingly, at stage 8, the weak fluorescent signal appeared spread throughout the cytoplasm of some oocytes, and was never accumulated at the anterior cortex of the cell. The *in situ* hybridization signal produced by the transcripts was also weak and diffuse throughout the cytoplasm of the nurse cells. This type of localization-defective phenotype will be referred to as a 'Loc\u2212' phenotype ([Figure 3A, B, and D](#fig3){ref-type=\"fig\"}). These results indicate that ORF1p and *I factor* sequences present in construct B were not sufficient to promote the Loc+ phenotype. Therefore, we analyzed several other deletion mutants, downstream of ORF1 to determine which sequences are required in *cis* for the patterning process ([Figure 2](#fig2){ref-type=\"fig\"}). According to the expression patterns observed for ORF1p and the I transcripts, we defined two categories of mutants: (I) those displaying a Loc+ phenotype ([Figures 3H--K](#fig3){ref-type=\"fig\"} and [4E](#fig4){ref-type=\"fig\"}), and (II) those displaying a Loc\u2212 phenotype similar to that presented by elements A and B ([Figures 3C, E, F, and G](#fig3){ref-type=\"fig\"} and [4C](#fig4){ref-type=\"fig\"}). We did not observe another category of mutants that would exhibit a Loc+ phenotype for the transcripts and Loc\u2212 for ORF1p (or the reverse). Sequence comparisons of the diverse constructs and their corresponding phenotypes revealed that a 552 nt segment, located in a central region of ORF2, is absent in all the mutants exhibiting a Loc\u2212phenotype ([Figure 2](#fig2){ref-type=\"fig\"}).\n\nThe 552 nt segment drives the localization of a heterologous RNA\n----------------------------------------------------------------\n\nTo further assess the significance of the *cis*-acting sequence identified above in the localization of the *I factor* RNA, we tested its capacity to localize a heterologous RNA expressed in the female germ line. We added the 552 nt fragment at the 3\u2032 end of a construct (DF313) containing the bacterial *lacZ* gene under the control of the promoter of the maternal alpha4-tubulin gene 67C that drives the expression in oocytes. The chimera *lacZ*/*I factor* construct was called DF313/loc. We transformed reactive flies either with DF313/loc, creating transgenic lines 313/loc, or with the unmodified DF313 construct, creating transgenic lines 313. In control lines 313, *lacZ* transcripts were not transported to the anterior pole of oocytes ([Figure 5B](#fig5){ref-type=\"fig\"}). In contrast, the chimera *lacZ*/loc transcripts localized to the oocyte anterior pole in lines 313/loc, exactly like *I factor* transcripts ([Figure 5A](#fig5){ref-type=\"fig\"}). This suggests that the 552 nt fragment of ORF2 contains all the necessary sequences to interact with cellular motor proteins and drive the localization of *I factor* products.\n\nA deletion in *I factor* ORF1 prevents an efficient expression of the Loc+ phenotype\n------------------------------------------------------------------------------------\n\nTo determine whether other *cis*-acting sequences contribute to *I factor* RNA localization, we analyzed the expression of the element I-\u0394Asu ([@b24]), which lacked 723 nt between the two AsuII restriction sites of ORF1. I-\u0394Asu produces a half size ORF1p, in which the zinc-knuckle region is removed ([Figure 6](#fig6){ref-type=\"fig\"}). Males of the Is37 transgenic line homozygous for the I-\u0394Asu element were crossed with JA females, and we analyzed the expression pattern of the transcripts and of the mutated ORF1 protein in the ovaries of their F1 progeny. This truncated protein appeared accumulated in the nurse cell cytoplasm. It appeared scarcely in oocytes and was not localized to the anterior pole (data not shown). The distribution of the transcripts was similar but, exceptionally, some very faint figures of localization at the anterior pole of stage 9 oocytes could be observed ([Figure 6A](#fig6){ref-type=\"fig\"}, arrows). These observations suggest that in the *I factor* context, the 552 nt region within ORF2 is not sufficient to promote a precise distribution of *I factor* RNAs, and that sequences of ORF1 missing in I-\u0394Asu and/or a functional ORF1p are necessary for their correct and efficient localization.\n\nORF1p can act in *trans* to restore the Loc+ phenotype\n------------------------------------------------------\n\nTo test whether ORF1p and/or other *cis*-acting sequences in ORF1 are required for the localization of *I factor* RNA, we conducted complementation experiments introducing I-\u0394Asu and full-length ORF1p into the same oocyte. For this, we performed reciprocal crosses between line Is37, containing the I-\u0394Asu element and a line containing the construct *hs*ORF1 (HH16 line) ([@b24]). In the construct *hs*ORF1 the expression of ORF1p is driven by the heat-shock promoter *hsp*70. An abundant and ubiquitous expression of ORF1p can be induced in line HH16 by heat treatment at 37\u00b0C (data not shown). A control heat shock on Is37 females had no effect on the localization of the I-\u0394Asu RNA ([Figure 6B](#fig6){ref-type=\"fig\"}). In contrast, when Is37/HH16 females were subjected to heat treatment, a large proportion of I-\u0394Asu transcripts was localized in the oocytes with a full Loc+ phenotype, exactly like complete *I factor* transcripts ([Figure 6C](#fig6){ref-type=\"fig\"}). Such hybrid females also expressed the typical sterility syndrome correlated with *I factor* retrotransposition. These results indicate that (i) in addition to the localization signal in ORF2, ORF1p is required to promote an efficient localization of *I factor* RNA; (ii) the central region of ORF1 (723 nt, absent in the I-\u0394Asu construct) does not contain the signals required in *cis* for correct localization; (iii) the Loc+ phenotype of an *I factor* RNA deficient in ORF1 can be rescued upon the supply of sufficient amounts of ORF1p provided in *trans*.\n\nDISCUSSION\n==========\n\nResults obtained from actively transposing *I factors* showed that ORF1p ([@b23]) and full length *I factor* RNA co-localize according to a dynamic pattern (Loc+) exclusively in permissive oocytes ([Figure 1a](#fig1){ref-type=\"fig\"}, panels A--D). The spatial and temporal co-localization of ORF1p and *I factor* RNA coincide with the expected place and time where the *I factor* is known to transpose (i.e. during the development of permissive young female germ cells), and suggests that during this period both molecules visualized here are a part of the transposition-related RNP complexes, expected early intermediates in the NLR retrotransposition process. Transcriptional activity is repressed in the oocyte and most maternal products synthesized during oogenesis are produced by the nurse cells and transported to the oocyte ([@b33]). Full-length *I factor* transcripts and ORF1p are both detected in the cytoplasm of very early oocytes, indicating that the RNA is transported to this cell immediately after synthesis in nurse cells, and that at least the first ORF of the bicistronic transcript is translated and transported at that time. The dynamic localization of these products coincides with nuclear migration to the anterior pole of the oocyte, indicating that both the nucleus and a fraction of *I factor* products remain in close contact during oogenesis. Some *Drosophila* morphogenes (proteins and/or RNAs), such as the *bicoid* or *gurken* mRNAs, follow similar re-localization patterns via motor proteins associated with the minus end of microtubules (dyneins) at mid-oogenesis ([@b34]--[@b37]). This similarity suggests that *I factor* products also localize using minus-end microtubule associated motor proteins. Through an ORF2p loss-of-function mutant (element I-HAO1-fsO2), we observed that ORF2p is not required for localization. Since localization of mRNAs is often coupled to translational control ([@b38],[@b39]), one can hypothesize that ORF2p is synthesized as a late product, only after ORF1p and the RNA have reached the anterior pole of the oocyte where they meet cellular factors required for the differential translation of ORF2p from the bicistronic transcript. ORF1p and the RNA become undetectable after stage 10b of oogenesis when nuclear breakdown is known to occur. It is possible that the translation process of ORF2 results in ORF1p release and degradation as the newly synthesized ORF2p is activated to proceed with the subsequent steps of reverse transcription and integration in the genome.\n\nUsing deletion derivatives of a tagged *I factor* we have identified a specific region of 552 nt within ORF2 containing a signal required for proper localization of both ORF1p and *I factor* RNA ([Figures 2](#fig2){ref-type=\"fig\"}--[4](#fig4){ref-type=\"fig\"}). This fragment fused to sequences of the *lacZ* bacterial gene promotes a similar localization of the *lacZ* transcripts, suggesting that cellular cargos or cargo-adaptor proteins can directly recognize it. Transport signals are often carried by stable stem-loop structures ([@b34],[@b40]). Several robust stem loops predicted in the 552 nt region of the *I factor* RNA are targets to test for interactions with ORF1p and for their involvement in localization.\n\nThe evidence that both ORF1p and the 552 nt signal in the RNA are implicated in the localization process of *I factor* products was obtained through the observation that an element deficient in ORF1 (I-\u0394Asu) localizes very poorly in oocytes but recovers a full Loc+ pattern when a full-size ORF1p is provided in large amounts in *trans* ([Figure 6](#fig6){ref-type=\"fig\"}). The extremely low efficiency of the *cis*-acting localization signal in the ORF1-mutant element ([Figure 6A](#fig6){ref-type=\"fig\"}), compared with its high capacity to promote anterior pole localization when fused to *lac*Z RNA ([Figure 5A](#fig5){ref-type=\"fig\"}), could be either due to the presence of the truncated ORF1p and/or due to a difference in the secondary structure resulting from a different sequence environment. The rescue of a full Loc+ phenotype when I-\u0394Asu was complemented by ORF1p suggests that a possible role of ORF1p is to act as a molecular chaperone directing RNA folding into a localization-competent structure, probably via site-specific binding, either inside and/or outside of the 552 nt stretch. Such a property is evocative of the chaperone Ncp7 protein which controls the conformational state of HIV RNA ([@b41]). The sequence-specific binding of the ORF1p of NLRs is a matter of debate. Assays performed with the human L1 p40 produced in heterologous systems have also led to the conclusion that binding does not require a specific sequence ([@b18]). However, p40 extracted from teratocarcinoma cells (in which L1 presumably transposes) specifically binds to two sites in the ORF2 RNA sequence ([@b42]), suggesting that some host dependent post-translational maturations may be essential for the full expression of ORF1p properties.\n\nOur complementation data as well as previous experiments ([@b20],[@b24]) indicate the importance of the localization process (Loc+) for efficient *I factor* transposition. Defective human L1 elements are poorly complemented in *trans* ([@b43],[@b44]) and a mechanism called *cis*-preference has been proposed to account for low complementation efficiency. This mechanism refers to functional proteins that preferentially remain associated with the RNA molecule from which they are synthesized, favoring the specific mobilization of the progenitor element ([@b45]). Early experimental data suggested that the *cis*-preference mechanism might also apply to the *I factor* since previous attempts to mobilize defective *I factor*s led to very low complementation efficiencies, even under conditions of protein overexpression ([@b20],[@b24]). Noticeably the elements used in those experiments did not contain the localization signal (ORF2 552 nt region). We have observed that the efficiency of complementation can be significantly improved by the use of deleted elements endowed with the localization signal and the appropriate supply of interacting proteins (ORF1p in our experiments). Another case of successful complementation when these conditions were met resulted in the rescue of the total transposition capacity of the non-functional element I-\u0394nuc (Loc+, deleted in the endonuclease domain) by a construct overexpressing a full-size ORF2p ([@b46]). Hence, at least part of the complementation efficiency to mobilize defective *I factor*s seems to depend on the localization signal interacting with one or more molecular partners. Under such conditions, genetic complementation analyses of *I factor* remain a powerful tool to continue dissecting the molecular components required for retrotransposition.\n\nAnalysis of the *Drosophila* genome has recently revealed the presence of 27 NLR families with two ORFs ([@b47]--[@b49]), but very little is known of the *in vivo* expression of any of these *Drosophila* NLRs. The expression in *Drosophila* cultured cells of ORF1p--GFP fusions from five of these *Drosophila* NLRs has shown a diversity of intracellular localizations: some of them enter the nucleus, but Doc or *I factor* ORF1p-GFP do not ([@b50],[@b51]). The *in vivo* expression pattern of *Drosophila* NLRs other than *I factor* is still undocumented except for the transcripts of the Doc and HetA elements in *Drosophila* oocytes ([@b52],[@b53]), but probably because the conditions of their retrotransposition are unknown, this pattern was not correlated to transposition. However, further investigations may reveal that the localization process described here for *I factor* RNPs reflects a more general situation.\n\nWe thank Dominique Ferrandon for the gift of the DF313 plasmid and Isabelle Busseau for providing transgenic lines HH16 and Is37. We thank Ned Lamb for a critical reading of the manuscript and students S\u00e8verine Guillemin, Corinne Nauroy and Camille Auziol for their help during their laboratory rotations. We thank Patrick Atger for assistance with the figures and Nicole Lautredou for assistance in confocal microscopy. This work was supported by the Centre National de la Recherche Scientifique and by grants from the Association pour la Recherche sur le Cancer (ARC). M.C.S. was the recipient of a fellowship from the Fondation Pour la Recherche M\u00e9dicale (FRM), and S.R. from the ARC. Funding to pay the Open Access publication charges for this article was provided by the ARC.\n\nFigures and Tables\n==================\n\n![The Loc+ phenotype. (**a**) The top diagram represents I-HA-O1, a functional *I factor* carrying the HA tag at the N-terminus of ORF1p ([@b23]). The arrow indicates the transcription start site of the *I factor* driven by an internal promoter contained within its 5\u2032UTR. The black bar underneath indicates the sequence used as the probe in *in situ* hybridizations. Below, the left panel shows a schematic diagram of an ovariole (the developmental unit of a *Drosophila* ovary where oogenesis takes place). At the anterior tip of the ovariole (ANT), in the germarium, stem germ cells divide to produce 16 cell cysts; one of these cells becomes the oocyte and the other 15 become nurse cells. Each cyst, surrounded by a layer of somatic follicle cells, constitutes an egg chamber. Egg chambers grow and mature as they progress from the germarium to the posterior part of the ovariole (POST), which ends in the oviduct. The different stages of maturation can be identified from the relative volume of the oocyte ([@b32]). (A, C, E, F and H) *In situ* hybridizations designed to detect *I factor* transcripts on whole-mount ovaries. (B, D and G) Immunostaining for the detection of HA-tagged ORF1p. (A--D) The Loc+ phenotype can be observed in the ovaries of SF females generated by crossing I males carrying the active tagged I-HA-O1 element with R females. *I factor* transcripts (A) and ORF1p (B) accumulate in the cytoplasm of the oocytes from the very early stages of oogenesis, as soon as the pro-oocyte is determined in the germarium (arrows). At mid-oogenesis (stages 8 and 9, and thereafter), when the oocyte nucleus migrates from the posterior to the antero-dorsal pole of the cell, both the *I factor* RNA (C) and ORF1p (D) are also re-localized and concentrate at the anterior cortex of the oocyte. *I factor* transcripts are not detected in the ovaries of R flies, which are devoid of active *I factor*s (E), nor in the ovaries of I flies, where active *I factor*s are silenced (F). (**b**) Loc+ pattern of I-HAO1-fsO2. The insertion of 8 nt, in red in the diagram, in I-HAO1 near the 5\u2032 end of ORF2 creates a frameshift in ORF2 and abolishes the transposing capacity of this element. The dynamic expression pattern of ORF1p (G) and RNA (H) is unchanged, indicating that the localization process does not require the participation of the ORF2 protein. \\[The concentration of ORF1p at the periphery of the oocyte anterior poles appears like green crescents in the oblique views of stages 8 and 9 egg chambers in (G) (arrows)\\]. Scale bars in (A) and (B) are 10 \u03bcm and in C--H are 50 \u03bcm.](gki221f1){#fig1}\n\n![Schematic representation of the I-HA-O1 element and of its deletion derivatives. The arrow at the beginning of the 5\u2032UTR indicates the transcriptional start site of the element. The relative locations of the endonuclease (EN), reverse transcriptase (RT) and RNAse H (RNH) domains within ORF2 are indicated. Restriction enzymes used to obtain the different constructs and their sites within the *I factor* sequence are shown. Thin lines represent the deleted sequences in each construct. The localization phenotypes with regard to ORF1p and the RNA are indicated at the right of each construct (Loc). tk represents the thymidine kinase terminator added to the 3\u2032 end of construct A. The 552 nt stretch missing in all constructs displaying a Loc\u2212 phenotype is indicated.](gki221f2){#fig2}\n\n![ORF1p expression produced by different *I factor* deletion derivatives. Immunostaining on ovaries of flies generated by crossing R JA females with transgenic males carrying either construct A (**A** and **B**), construct E (**C**), construct B (**D**), construct CI18 (**E**); construct CI3 (**F** and **G**), construct I-\u0394nuc (**H**), construct D (**I**), construct CI21 (**J**) or construct I-\u0394Hind (**K**). The constant and regular ORF1p expression pattern of Loc+ mutants shown in (H--K) complies exactly with the localization features of full-size elements (see [Figure 1](#fig1){ref-type=\"fig\"}). In contrast, Loc\u2212 mutants (A--G) present a large variability in the ORF1p expression patterns within each transgenic line. (A and B) and (F and G) illustrate this variability in the same genetic background. However, two features are constant: ORF1p is not preferentially accumulated in oocytes at the early stages of oogenesis but may occasionally label the oocytes in stage 6 and 7 cysts (arrows in A, C and E). ORF1p is generally missing in stages 8 and 9 oocytes, but when present, it appears scattered all over that cell (arrowheads in E and G) and does not migrate to the anterior pole.](gki221f3){#fig3}\n\n![RNA expression pattern produced by different *I factor* deletion derivatives. *I factor* transcripts are detected as a blue precipitate following *in situ* hybridization on ovaries (A--E) (see Materials and Methods). The probe used does not reveal any *I factor* RNA in control R JA flies (**F**). Transcripts of all Loc\u2212 mutants appear scattered in all the 16 cells of young cysts and are accumulated in nurse cells after stage 10 as shown in (A--D). Exceptionally, the RNA appears accumulated in some oocytes at stages 6 and 7 (arrow in A), like ORF1p in some Loc\u2212 phenotypes (arrows in [Figure 3A and C](#fig3){ref-type=\"fig\"}). In contrast, the phenotype of Loc+ mutants (E) is regular and identical to that of the full-size element. Construct A (**A**); construct B (**B**); construct CI18 (**C**); construct E (**D**) or construct I-\u0394nuc (**E**).](gki221f4){#fig4}\n\n![The *cis*-acting signal located within ORF2 drives the localization of a heterologous RNA. *In situ* hybridizations, using a *lacZ* probe, show that *lacZ* transcripts are not detected in the ovaries of control JA flies (**C**). In the ovaries of 313 flies, *lacZ* transcripts are scattered in the cytoplasm of the 16 cells of cysts (**B**). In 313/loc ovaries, the chimera *lacZ*/loc transcripts display a typical Loc+ phenotype like *I factor* transcripts (**A**).](gki221f5){#fig5}\n\n![ORF1p can act in *trans*. The transcripts of the I-\u0394Asu element, deleted in ORF1, do not accumulate in oocytes of flies issued from crosses of Is37 (line containing I-\u0394Asu) with JA reactive flies, but some weak signals may be observed occasionally at the anterior pole of stage 10 oocytes (arrows in **A**). A heat-shock treatment does not affect the Loc\u2212 phenotype of I-\u0394Asu (**B**). In the ovaries of females resulting from crosses between Is37 and HH16 (carrying the *hs*ORF1 construct), when overproduction of ORF1p is induced by a heat-shock treatment, a typical Loc+ localization pattern is observed (**C**).](gki221f6){#fig6}\n\n[^1]: Present addresses: Maria del Carmen Seleme, Department of Genetics, The University of Pennsylvania Medical School, Philadelphia, PA 19104, USA\n\n[^2]: Olivier Disson, Institut de G\u00e9n\u00e9tique Mol\u00e9culaire de Montpellier, C.N.R.S., 1919 Route de Mende 34090 Montpellier, France\n"}
+{"text": "1. Introduction {#sec1-sensors-20-02157}\n===============\n\nApproximately 30 percent of the United States' total energy consumption comes from the residential sector, and the amount of the residential energy consumption is expected to grow owing to increased use of home appliances (e.g., air conditioners (ACs) and washing machines (WMs)) and modern electronic devices \\[[@B1-sensors-20-02157]\\]. Thus, an efficient and economical methodology for residential energy management is required to reduce the electricity bill of consumers and keep the efficiency of their appliances. Furthermore, as distributed energy resources (DERs) (e.g., a rooftop solar photovoltaic (PV), a residential energy storage system (ESS), and an electric vehicle (EV)) become integrated into an individual residential house via an advanced metering infrastructure with smart meters for reliable residential grid operations, the complexity of the residential energy management increases. The aforementioned challenges require more intelligent systems, i.e., home energy management systems (HEMSs), through which an electric utility or a third party provides consumers with an efficient and economical control of home appliances.\n\nA primary goal of HEMS is to reduce the electricity bill of consumers while satisfying their comforts and preferences. To achieve this goal, HEMSs perform the following two functions: (1) real-time monitoring of the energy usage of consumers using smart meters; (2) scheduling of the optimal energy consumption of home appliances. To implement this second function, an HEMS algorithm is generally formulated as a model-based optimization problem. Recently, numerous studies have been published on the development of HEMS optimization algorithms \\[[@B2-sensors-20-02157],[@B3-sensors-20-02157],[@B4-sensors-20-02157],[@B5-sensors-20-02157],[@B6-sensors-20-02157],[@B7-sensors-20-02157],[@B8-sensors-20-02157],[@B9-sensors-20-02157],[@B10-sensors-20-02157],[@B11-sensors-20-02157],[@B12-sensors-20-02157]\\]. These studies address the scheduling of the energy consumption for home appliances and DERs, while maintaining the consumer's comfort level using mixed-integer nonlinear programming (MINLP) \\[[@B2-sensors-20-02157]\\], the load scheduling using mixed-integer linear programming (MILP) for single and multiple households \\[[@B3-sensors-20-02157],[@B4-sensors-20-02157]\\], robust optimization for scheduling of home appliances to resolve the uncertainty of consumer behavior \\[[@B5-sensors-20-02157]\\], and distributed HEMS architectures consisting of local and global HEMSs \\[[@B6-sensors-20-02157]\\]. A quickly distributed HEMS algorithm was developed for a large number of households using the MINLP approach with a nonconvex relaxation \\[[@B7-sensors-20-02157]\\]. Based on an emerging technology for ESSs and EVs, a home energy consumption model under the control of the ESS was presented in \\[[@B8-sensors-20-02157]\\]. A model predictive control-based HEMS algorithm was proposed using the prediction of the EV state \\[[@B9-sensors-20-02157]\\]. An HEMS optimization model considering both ESS and EV was formulated for a single household \\[[@B10-sensors-20-02157],[@B11-sensors-20-02157]\\] based on their bi-directional operation and multiple households with a renewable energy facility \\[[@B12-sensors-20-02157]\\]. In addition, many studies proposed the methods to evaluate and preserve the consumer comfort during the HEMS process. In \\[[@B13-sensors-20-02157]\\], a quality of experience (QoE)-aware HEMS was developed where the QoE-aware cost saving appliance scheduling and the QoE-aware renewable source power allocation are conducted to schedule the operation of the controllable loads based on the consumer preferences and the available renewable energy sources. A new demand management scheme based on the operational comfort level (OCL) of consumer was proposed to minimize the peak-to-average ratio while maximizing the OCL of consumers \\[[@B14-sensors-20-02157]\\]. A score-based HEMS method was presented to maintain the total household power consumption below a certain limit by scheduling various household loads based on the consumer comfort level setting \\[[@B15-sensors-20-02157]\\]. A QoE-aware HEMS algorithm was presented to reduce the peak load and electricity cost while satisfying the consumer comfort and QoE with a fixed threshold \\[[@B16-sensors-20-02157]\\]. More recently, an advanced QoE-aware HEMS method considering renewable energy sources and EVs was developed for adaptively varying the QoE threshold \\[[@B17-sensors-20-02157]\\]. Compared to the method with the fixed OoE threshold in \\[[@B16-sensors-20-02157]\\], a fuzzy logic controller was designed to dynamically adjust the QoE threshold for optimizing the consumer QoE in \\[[@B17-sensors-20-02157]\\]. A recent work on HEMS was summarized in \\[[@B18-sensors-20-02157]\\].\n\nHowever, the aforementioned optimization-based HEMS methods were executed according to deterministic equations to illustrate the operation characteristics of home appliances and DERs (e.g., consecutive operation time intervals of WMs and state of energy (SOE) dynamics of ESS) as well as consumer's comfort level (e.g., preferred indoor temperature using indoor temperature dynamics). Consequently, the model-based HEMS optimization approach suffers from two limitations. First, the characteristics for the operation of appliances/DERs and consumer preference are expressed through approximated unrealistic equations with fixed parameters, thereby leading to an inaccurate energy consumption schedule. Second, an optimization method including a large number of decision variables could significantly increase the computation complexity and could not scale well with a greater number of houses. Furthermore, the solution resulting from model-based optimization may not always be guaranteed and often diverges owing to a smaller feasible region with a large number of operational constraints for the HEMS optimization problem. To address the aforementioned limitations, we propose a data-driven approach that leverages model-free reinforcement learning (RL) to calculate the optimal schedule of home energy consumption.\n\nRecently, data-driven approaches based on various machine learning (ML) methods have gained popularity owing to their more efficient residential energy management. In \\[[@B19-sensors-20-02157],[@B20-sensors-20-02157]\\], methods to predict the generation output of a PV system accurately during the day for efficient energy management of buildings were presented. These methods employed an artificial neural network (ANN) and a deep neural network (DNN), respectively. DNN methods were also used in \\[[@B21-sensors-20-02157],[@B22-sensors-20-02157],[@B23-sensors-20-02157]\\] for load forecasting to minimize the energy usage of buildings and households. More recently, RL has received attention as a promising ML method for the energy management of buildings and homes. A pioneering study on RL-based energy management is Google DeepMind, which was developed using RL and proved to decrease the electricity bill by cooling the data center by approximately 40%. Another RL-based method, referred to as Q-learning, was applied to HEMS problems. It was integrated with the ANN module for estimating the consumer's comfort level, maintaining the energy efficiency of household appliances \\[[@B24-sensors-20-02157],[@B25-sensors-20-02157]\\] and predicting the pricing in real time \\[[@B26-sensors-20-02157]\\]. A new demand response strategy for HEMS was proposed through the combination of Q-learning and fuzzy reasoning, which reduces the number of state-action pairs and fuzzy logic for reward functions \\[[@B27-sensors-20-02157]\\]. Furthermore, methods based on deep reinforcement learning (DRL) such as Deep Q-Network, and policy gradient, were applied for energy management of building \\[[@B28-sensors-20-02157],[@B29-sensors-20-02157]\\] based on both discrete and continuous action spaces. A holistic DRL method for the energy management of commercial buildings was presented in \\[[@B30-sensors-20-02157]\\] where Heating, Ventilation, and Air conditioning (HVAC) system, lighting, blind, and window systems are controlled to achieve energy savings within the buildings' occupants comfort in terms of thermal, air quality, and illumination conditions. To resolve the limit of model-free DRL methods such as low sample efficiency, a model-based RL method was developed for building HVAC control that trains the system dynamics using neural networks \\[[@B31-sensors-20-02157]\\]. Based on the trained system dynamics, the operation of the HVAC system was managed by model predictive control to minimize both the energy cost and the indoor temperature constraints violation. The DRL approach was also applied to data centers with servers, which aim to minimize the energy used for moving air and on-demand cooling in the data centers through the control of the temperature and relative humidity of air supplied to the server \\[[@B32-sensors-20-02157]\\].\n\nRecent studies addressed on energy management systems for buildings and households using an RL-based method. However, to the best of the authors' knowledge, no study has presented an DRL-based algorithm that considers the continuous operations of heterogeneous home appliances and DERs according to the consumer's comfort and preferences. In prior studies \\[[@B24-sensors-20-02157],[@B25-sensors-20-02157],[@B26-sensors-20-02157]\\], the operations of home appliances and DERs were scheduled using a simple Q-learning method based on an unrealistic discrete action space. Furthermore, prior studies \\[[@B28-sensors-20-02157],[@B29-sensors-20-02157]\\] focused on the scheduling of the energy consumption of buildings without considering the operation characteristics of home appliances in detail.\n\nIn this study, we propose a two-level DRL framework that employs an actor--critic method where the controllable home appliances (WM and AC) are scheduled at the first level according to the consumer's preferred appliance scheduling and comfort level. The ESS and EV are scheduled at the second level to cover the aggregated WM and AC loads that are calculated at the first level along with the fixed load of the uncontrollable appliances. The proposed two-level scheme is motivated by the interdependent operation between the home appliances at the first level and the ESS/EV at the second level. If the proposed algorithm is executed in a single-level framework, the optimal policy for charging and discharging actions of the ESS and EV is independently determined without considering the energy consumption schedule of aggregated home appliances, thereby degrading the performance of the algorithm. [Figure 1](#sensors-20-02157-f001){ref-type=\"fig\"} presents the conceptual system model for the proposed two-level DRL-based HEMS that employs an actor--critic method, along with the data classification associated with the utility company, weather station, and consumer. The main contributions of this study are summarized as follows:We present a two-level distributed DRL model for optimal energy management of a smart home consisting of a first level for WM and AC, and a second level for ESS and EV. In such a model, the energy consumption scheduling at the second level is based on the aggregated energy consumption scheduled at the first level to determine the better policy of charging and discharging actions for the ESS and EV.Compared to the existing method using Q-learning in a discrete action space, we propose a hierarchical DRL in a continuous action space with the following two scheduling steps: (i) the controllable appliances including WM and AC are scheduled at the first level according to the consumer's preferred appliance scheduling and comfort level; (ii) ESS and EV are scheduled at the second level, thereby resulting in optimal cost of electricity for a household.\n\nThe simulation results confirmed that the proposed HEMS algorithm can successfully schedule the energy consumptions of multiple home appliances and DERs using a single DRL structure under various consumer preferences. In addition, through various case studies and comparative analysis, we evaluated the impact of different weather and driving patterns of the EV with different initial SOEs on the proposed algorithm. Furthermore, we verified that charging and discharging of the ESS and EV significantly contribute to the reduction of the cost of electricity.\n\nThe remainder of this paper is organized as follows. [Section 2](#sec2-sensors-20-02157){ref-type=\"sec\"} introduces the various types of smart home appliances and the traditional HEMS optimization approach; it also provides an overview of the RL methodology. [Section 3](#sec3-sensors-20-02157){ref-type=\"sec\"} presents the formulation of the proposed DRL-based HEMS algorithm based on the actor--critic method. The numerical examples for the proposed HEMS algorithm are reported in [Section 4](#sec4-sensors-20-02157){ref-type=\"sec\"}, and the conclusions are given in [Section 5](#sec5-sensors-20-02157){ref-type=\"sec\"}.\n\n2. Background {#sec2-sensors-20-02157}\n=============\n\n2.1. Types of Smart Home Appliances {#sec2dot1-sensors-20-02157}\n-----------------------------------\n\nWe consider the following types of smart home appliances in a single household where an HEMS automatically schedules their energy consumption under the time-of-use (TOU) pricing:(1)Uncontrollable appliance ($\\mathcal{A}^{uc}$): An HEMS cannot manage the energy consumption scheduling of uncontrollable appliances such as televisions, personal computers, and lighting. Thus, an uncontrollable appliance is assumed to follow fixed energy consumption scheduling.(2)Controllable appliance ($\\mathcal{A}^{c}$): It is an appliance for which the energy consumption scheduling is calculated by the HEMS. According to its operation characteristics, the controllable appliance is categorized into a reducible appliance ($\\mathcal{A}_{r}^{c}$) and shiftable appliance ($\\mathcal{A}_{s}^{c}$). A representative example of a reducible appliance is an air conditioner whose energy consumption can be curtailed to reduce the cost of electricity. By the contrast, under TOU pricing, the energy consumption scheduling of a shiftable appliance can be moved from one time slot to another to minimize the cost of electricity. A shiftable appliance has two types of load: (i) a non-interruptible load ($\\mathcal{A}_{s}^{c,NI}$), and (ii) an interruptible load ($\\mathcal{A}_{s}^{c,I}$). A shiftable appliance with an interruptible load can be interrupted at any time. For example, the HEMS must stop the discharging process and start the charging process of the ESS instantly when the PV power generation is greater than the load demand. However, the operation period of a shiftable appliance with a non-interruptible load must not be terminated by the HEMS. For example, a washing machine must finish a washing cycle prior to drying.\n\n2.2. Traditional HEMS Optimization Approach {#sec2dot2-sensors-20-02157}\n-------------------------------------------\n\nA conventional HEMS method that calculates the optimal operating scheduling of home appliances and DERs is formulated in terms of the following constrained multi-objective optimization problem described in the following section.\n\n### 2.2.1. Objective Function {#sec2dot2dot1-sensors-20-02157}\n\nThe objective function ([1](#FD1-sensors-20-02157){ref-type=\"disp-formula\"}) for the HEMS optimization problem comprises two terms, each of which includes different decision variables ($E_{t}^{net},T_{t}^{in}$) \\[[@B25-sensors-20-02157]\\]:$$\\min\\limits_{E_{t}^{net},T_{t}^{in}}\\underset{J_{1}{(E_{t}^{net})}}{\\underset{\ufe38}{\\underset{t\\in\\mathcal{T}}{\\sum}\\pi_{t}E_{t}^{net}}} + \\underset{J_{2}{(T_{t}^{in})}}{\\underset{\ufe38}{\\epsilon\\underset{t\\in\\mathcal{T}}{\\sum}\\left| {T_{t}^{in}-T^{set}} \\right|}}.$$\n\nThe first term $J_{1}\\left( E_{t}^{net} \\right)$ represents the total cost of electricity that is calculated under TOU pricing $\\pi_{t}$ and $E_{t}^{net}$, which is the net energy consumption accounting for the energy consumption of the controllable/uncontrollable appliances and the predicted PV generation output. The second term $J_{2}\\left( T_{t}^{in} \\right)$ represents the total penalty related to the cost of the consumer's discomfort. Here, discomfort is defined as a deviation of the consumer's preferred temperature $T^{set}$ from the indoor temperature $T_{t}^{in}$. $\\epsilon$ is a penalty for the cost of the consumer's discomfort. A larger $\\epsilon$ yields a smaller $J_{2}\\left( T_{t}^{in} \\right)$, thereby offering decreasing discomfort to the consumer at the expense of less energy saving. The value of $\\epsilon$ can be tuned by the HEMS operator to maintain the consumer's preferred comfort level at the cost of a higher electricity bill. The equality and inequality constraints for the HEMS optimization problem are illustrated in the following subsections.\n\n### 2.2.2. Net Energy Consumption {#sec2dot2dot2-sensors-20-02157}\n\nEquation ([2](#FD2-sensors-20-02157){ref-type=\"disp-formula\"}) expresses the constraint on the net energy consumption that represents the difference between the total consumption of all home appliances ($\\mathcal{A} = \\mathcal{A}_{r}^{c}\\bigcup\\mathcal{A}_{s}^{c,NI}\\bigcup\\mathcal{A}_{s}^{c,I}\\bigcup\\mathcal{A}^{uc}$) and the predicted PV generation output $E_{t}^{PV}$ at time *t*. In Equation (3), the total energy consumption of all appliances in Equation ([2](#FD2-sensors-20-02157){ref-type=\"disp-formula\"}) is decomposed into four different types of consumptions corresponding to (i) reducible appliances ($a \\in \\mathcal{A}_{r}^{c}$), (ii) shiftable appliances with a non-interruptible load ($a \\in \\mathcal{A}_{s}^{c,NI}$), (iii) shiftable appliances with an interruptible load ($a \\in \\mathcal{A}_{s}^{c,I}$), and (iv) uncontrollable appliances ($a \\in \\mathcal{A}^{uc}$) \\[[@B25-sensors-20-02157]\\]:$$E_{t}^{net}{= \\sum\\limits_{a \\in \\mathcal{A}}E_{a,t} - {\\hat{E}}_{t}^{PV}}$$$$\\begin{aligned}\n{\\sum\\limits_{a \\in \\mathcal{A}}E_{a,t}} & {= \\sum\\limits_{a \\in \\mathcal{A}_{r}^{c}}E_{a,t} + \\sum\\limits_{a \\in \\mathcal{A}_{s}^{c,NI}}E_{a,t}} \\\\\n & {+ \\sum\\limits_{a \\in \\mathcal{A}_{s}^{c,I}}\\left( {E_{a,t}^{ch} - E_{a,t}^{dch}} \\right) + \\sum\\limits_{a \\in \\mathcal{A}^{uc}}E_{a,t}.} \\\\\n\\end{aligned}$$\n\n### 2.2.3. Operation Characteristics of Controllable Appliances {#sec2dot2dot3-sensors-20-02157}\n\nFor a reducible appliance $a \\in \\mathcal{A}_{r}^{c}$, ([4](#FD4-sensors-20-02157){ref-type=\"disp-formula\"}) expresses the constraint for the indoor temperature dynamics of a reducible appliance (e.g., an AC) at time *t* ($T_{t}^{in}$), which is expressed in terms of $T_{t - 1}^{in}$ at time $t - 1$, the predicted outdoor temperature at time $t - 1$ (${\\hat{T}}_{t - 1}^{out}$), the energy consumption of the reducible appliances ($E_{a,t}$), and the environmental parameters ($\\alpha,\\beta$) characterizing the indoor thermal condition \\[[@B7-sensors-20-02157]\\]. Equation (5) illustrates the range of consumer's preferred indoor temperatures. The energy consumption capacity for the reducible appliances is limited according to (6): $$\\begin{aligned}\nT_{t}^{in} & {= T_{t - 1}^{in} + \\alpha\\left( {\\hat{T}}_{t - 1}^{out} - T_{t - 1}^{in} \\right) + \\beta E_{a,t}} \\\\\n\\end{aligned}$$$$\\begin{aligned}\nT^{\\min} & {\\leq T_{t}^{in} \\leq T^{\\max}} \\\\\n\\end{aligned}$$$$\\begin{aligned}\nE_{a}^{\\min} & {\\leq E_{a,t} \\leq E_{a}^{\\max}.} \\\\\n\\end{aligned}$$\n\nEquations (7)--(9) guarantee the consumer's preferred operation of shiftable appliances with a non-interruptible load $a \\in \\mathcal{A}_{s}^{c,NI}$ (e.g., a WM) with the binary decision variable $b_{a,t}^{c,NI}$ in different situations: (i) for a stopping period in which $\\omega_{s}^{pref}$ and $\\omega_{f}^{pref}$ are the consumer's preferred starting and finishing time ([7](#FD7-sensors-20-02157){ref-type=\"disp-formula\"}), respectively; (ii) for an operation period of $L_{a}$ hours during a day in (8); and (iii) for a consecutive operation period of $L_{a}$ hours in (9). The energy consumption capacity for the shiftable appliances with a non-interruptible load is calculated using (10):$$\\begin{aligned}\nb_{a,t}^{c,NI} & {= 0,\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu} t \\in \\left\\lbrack 1,\\omega_{s}^{pref} \\right) \\cup \\left( \\omega_{f}^{pref},T \\right\\rbrack} \\\\\n\\end{aligned}$$$$\\begin{aligned}\n{\\sum\\limits_{t = \\omega_{s}^{pref}}^{\\omega_{f}^{pref}}b_{a,t}^{c,NI}} & {= L_{a}} \\\\\n\\end{aligned}$$ $${\\sum\\limits_{t = p}^{p + L_{a} - 1}b_{a,t}^{c,NI}}{\\geq \\left( b_{p}^{c,NI} - b_{p - 1}^{c,NI} \\right)L_{a},}{\\forall p \\in \\left( \\omega_{s}^{pref},\\omega_{f}^{pref} - L_{a} + 1 \\right)}$$$$\\begin{aligned}\nE_{a,t} & {= b_{a,t}^{c,NI}E_{a}^{\\max}.} \\\\\n\\end{aligned}$$\n\nEquation ([11](#FD11-sensors-20-02157){ref-type=\"disp-formula\"}) presents the operational dynamics of the SOE for the ESS and EV ($a \\in \\mathcal{A}_{s}^{c,I}$) at current time instant *t* in terms of the SOE at previous time instant *t*-1, the charging and discharging efficiency, i.e., $\\eta_{a}^{ch}$ and $\\eta_{a}^{dch}$, and the charging and discharging energy, i.e., $E_{a,t}^{ch}$ and $E_{a,t}^{dch}$, respectively \\[[@B11-sensors-20-02157]\\]. Equation (12) represents the SOE capacity constraint for the ESS and EV. Equations (13) and (14) present the limits of the charging ($E_{a,t}^{ch}$) and discharging ($E_{a,t}^{dch}$) energies of the ESS and EV, respectively, where $b_{a,t}^{c,I}$ represents the binary decision variable that determines the charging and discharging status of the ESS and EV:$$\\begin{aligned}\n{SOE_{a,t} = SOE_{a,t - 1}} & {+ \\eta_{a}^{ch}E_{a,t}^{ch} - \\frac{E_{a,t}^{dch}}{\\eta_{a}^{dch}}} \\\\\n\\end{aligned}$$$$\\begin{aligned}\n{SOE_{a}^{\\min} \\leq} & {SOE_{a,t} \\leq SOE_{a}^{\\max}} \\\\\n\\end{aligned}$$$$\\begin{aligned}\n{E_{a}^{{ch},\\min}b_{a,t}^{c,I} \\leq} & {E_{a,t}^{ch} \\leq E_{a}^{{ch},\\max}b_{a,t}^{c,I}} \\\\\n\\end{aligned}$$$$\\begin{aligned}\n{E_{a}^{{dch},\\min}\\left( 1 - b_{a,t}^{c,I} \\right) \\leq} & {E_{a,t}^{dch} \\leq E_{a}^{{dch},\\max}\\left( 1 - b_{a,t}^{c,I} \\right).} \\\\\n\\end{aligned}$$\n\nNotably, the constraints ([11](#FD11-sensors-20-02157){ref-type=\"disp-formula\"})--(14) for the EV remain true in $t \\in \\left\\lbrack \\omega^{arr},\\omega^{dep} \\right\\rbrack$, whereas $E_{a,t}^{ch}$ and $E_{a,t}^{dch}$ become zero in $t \\notin \\left\\lbrack \\omega^{arr},\\omega^{dep} \\right\\rbrack$. In addition, when the EV departs from home at $t = \\omega^{dep}$, the SOE of the EV must be larger than the consumer preferred SOE $SOE_{a}^{pref}$ $$SOE_{a,t} \\geq SOE_{a}^{pref}.$$\n\n2.3. Reinforcement Learning Methodology {#sec2dot3-sensors-20-02157}\n---------------------------------------\n\n### 2.3.1. Reinforcement Learning {#sec2dot3dot1-sensors-20-02157}\n\nRL is an ML method that addresses a problem in a specific environment with the objective of maximizing a numerical reward. This learning process is applied to various types of general and special engineering problems. In the RL framework, while an agent interacts with an environment, it learns a particular type of action depending on the state of the environment and conveys the learned action to the environment. The environment then returns a reward along with its new state to the agent. This learning process continues until the agent maximizes the total cumulative rewards received from the environment.\n\nA policy is defined in terms of the procedure through which the agent acts from a specific state. The main objective of the agent is to find an optimal policy that maximizes the agent's cumulative reward in the environment. In our study, we consider that the environment is characterized by a Markov decision process, in which the change of the agent's next state depends only on the current state, along with the action chosen in the current state ignoring all previous states and actions.\n\nIn this study, the value function is selected as $Q\\left( s_{t},a_{t} \\right)$, namely the Q-value, which is written in terms of a pair of state $s_{t}$ and action $a_{t}$ at a discrete time *t*. By using the Q-value, the agent's main objective is to achieve the maximum Q-value at every time step *t*. Q-learning is one of the basic RL methods to find the optimal policy $\\nu^{*}$ in decision-making problems. The general *Q*-learning process computes and updates the Q-value $Q\\left( s_{t},a_{t} \\right)$ to achieve the maximum total rewards using the following Bellman equation:$$Q_{\\nu^{*}}^{*}\\left( s_{t},a_{t} \\right) = r\\left( s_{t},a_{t} \\right) + \\gamma\\max Q\\left( s_{t + 1},a_{t + 1} \\right)$$ where, based on the optimal policy $\\nu^{*}$, the optimal Q-value $Q_{\\nu^{*}}^{*}\\left( s_{t},a_{t} \\right)$ is obtained by the summation of the present reward $r\\left( s_{t},a_{t} \\right)$ and the maximum discounted future reward $\\gamma\\max Q\\left( s_{t + 1},a_{t + 1} \\right)$.\n\nIn general, a discounting factor $\\gamma \\in \\left\\lbrack 0,1 \\right\\rbrack$ is used to explain the relative importance of the current and future rewards. As the discounting factor $\\gamma$ decreases, the agent becomes short-sighted because it increasingly focuses on the current reward. However, a larger $\\gamma$ enables the agent to focus increasingly on the future reward and thus becomes far-sighted. The value of $\\gamma$ can be tuned by the system operator to balance the current and future rewards.\n\n### 2.3.2. Actor--Critic Method {#sec2dot3dot2-sensors-20-02157}\n\nThe actor--critic method is an extension of the policy gradient method that can improve the stability and reduce the variance of the gradient when the optimal solution of the algorithm converges \\[[@B33-sensors-20-02157]\\]. If the value of a certain state of the agent is known, the corresponding Q-value can be calculated and applied to the REINFORCE method (Algorithm 1) to calculate the gradient of policy network parameters and renew the agent's policy network, thereby leading to a better cumulative result by increasing the probability for the agent's action. In the actor--critic method, the agent can use an additional network to judge the goodness of the action the agent selects in a certain state. The policy network that returns the probability of the agent's action is called actor network, whereas the network that returns the evaluation value of the agent's action is called a critic network. The policy gradient method is suitable for handling the problem with a continuous action space; however, this method may have a poor convergence performance. An additional critic network in the actor--critic method can resolve the convergence issue in the policy gradient method. In our study, the actor and critic networks share a common body network owing to an effective convergence consideration.\n\nThe actor network returns the probability of the action that the agent selects in a particular state and the critic network returns the numerical future value that the agent would obtain in the terminal state. The critic network updates the function that distinguishes between the action and value, whereas the policy network updates its parameters in the direction suggested by the critic network. In this study, the parameters in the actor network are updated by the REINFORCE method (Algorithm 1) and the parameters in the critic network are updated by a linear temporal difference (TD) method \\[[@B34-sensors-20-02157]\\]. The TD method directly learns from episodes of experience, and the present and guessed values. This is known to be a useful method to solve the Markov decision process (MDP) problem. In our study, we select the linear value function approximation for applying the TD method to the critic network. The actor--critic method updates the parameters of the actor and critic networks to minimize the TD error, which encodes the difference between the value function of the present state and target value function. Algorithm 2 illustrates the actor--critic approach with the TD method. In Algorithm 2, *\u2202*, $\\theta$, and $\\omega$ represent the TD error from the Q-value, the parameter of the actor network, and the parameter of the critic network, respectively. $\\xi_{\\theta}$ and $\\xi_{\\omega}$ are the learning rates of the actor network and critic networks in the algorithm, respectively. **Algorithm 1:** REINFORCE method![](sensors-20-02157-i001.jpg)\n\nAlgorithm 2:\n\nActor-critic method\n\n3. Proposed Method for DRL-Based Home Energy Management {#sec3-sensors-20-02157}\n=======================================================\n\nIn this section, we propose a hierarchical two-level DRL framework that employs the actor--critic method to schedule the day-ahead optimal energy consumption of a single household with smart home appliances and DERs within the consumer's preferred appliance scheduling and comfort range. In the proposed framework, the energy consumption scheduling problem is decomposed into a two-level DRL problem corresponding to the energy consumption scheduling of i) WM and AC at the first level and ii) ESS and EV at the second level. A detailed illustration of the state, action, and reward for each level of the proposed actor--critic method is provided in the following two subsections. In addition, the superscript of the variables for the state/action spaces and the reward function represents the first and second level, respectively.\n\n3.1. Energy Management Model for WM and AC: Level 1 {#sec3dot1-sensors-20-02157}\n---------------------------------------------------\n\n### 3.1.1. State Space {#sec3dot1dot1-sensors-20-02157}\n\nWe consider the situation in which the proposed DRL-based HEMS algorithm performs optimal day-ahead scheduling of appliances with a 1-h scheduling resolution. For $\\forall t = 1,\\ldots,24$, the state space of the agent for WM and AC at the first level is defined as follows:$$\\mathcal{S}^{(1)} = \\left\\{ t,\\pi_{t},{\\hat{T}}_{t}^{out},T_{t - 1}^{in} \\right\\}$$ where the states *t*, $\\pi_{t}$, ${\\hat{T}}_{t}^{out}$, and $T_{t - 1}^{in}$ denote the scheduling time of the WM and AC, TOU price, and outdoor and indoor temperatures, respectively, at time *t*.\n\n### 3.1.2. Action Space {#sec3dot1dot2-sensors-20-02157}\n\nThe optimal action for the first level relies on the environment of the agent including the present state, as defined in [Section 3.1.1](#sec3dot1dot1-sensors-20-02157){ref-type=\"sec\"}. The action space at the first level is defined as follows:$$\\mathcal{A}^{(1)} = \\left\\{ E_{t}^{WM},E_{t}^{AC} \\right\\}$$ where $E_{t}^{WM}$ and $E_{t}^{AC}$ represent the energy consumption of the WM and AC at time *t*, respectively. In this study, $E_{t}^{AC}$ has a continuous value, whereas $E_{t}^{WM}$ has a discrete value; $E_{t}^{WM} = E^{{WM},\\max}$ when the WM turns on, otherwise, $E_{t}^{WM} = 0$.\n\n### 3.1.3. Reward {#sec3dot1dot3-sensors-20-02157}\n\nThe reward function for the first level is formulated as the sum of the negative cost of electricity and dissatisfaction of WM and AC related to the consumer's preferred comfort and appliance's operation characteristics. The total reward at the first level is expressed as:$$\\mathcal{R}_{t}^{(1)} = - \\left( c_{t}^{WM} + c_{t}^{AC} \\right)$$ where $c_{t}^{WM}$ and $c_{t}^{AC}$ are the cost functions for the WM and AC, respectively. Each cost function includes the cost of electricity for the appliance along with the cost of the consumer's dissatisfaction for the undesired operation of the WM and the indoor thermal discomfort.\n\nFirst, the cost function of the WM is defined as $$\\begin{array}{r}\n{c_{t}^{WM} = \\left\\{ \\begin{array}{l}\n{\\pi_{t}E_{t}^{WM} + \\overline{\\delta}\\left( \\omega_{s}^{pref} - t \\right),\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} t < \\omega_{s}^{pref}} \\\\\n{\\pi_{t}E_{t}^{WM} + \\underline{\\delta}\\left( t - \\omega_{f}^{pref} \\right),\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} t > \\omega_{f}^{pref}} \\\\\n{\\pi_{t}E_{t}^{WM},\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{otherwise}} \\\\\n\\end{array} \\right.} \\\\\n\\end{array}$$ where $\\omega_{s}^{pref}$ and $\\omega_{f}^{pref}$ are the consumer's preferred starting and finishing times of the WM, respectively, while $\\overline{\\delta}$ and $\\underline{\\delta}$ are the penalties for early and late operations, respectively, compared to the consumer's preferred operation interval. The cost of dissatisfaction is added to the cost function if the WM schedules the WM energy consumption earlier than $\\omega_{s}^{pref}$ or later than $\\omega_{f}^{pref}$; otherwise, the cost function includes only the cost of electricity.\n\nThe cost function of the AC is expressed as:$$\\begin{array}{r}\n{c_{t}^{AC} = \\left\\{ \\begin{array}{l}\n{\\pi_{t}E_{t}^{AC} + \\overline{\\kappa}\\left( T^{\\min} - T_{t}^{in} \\right),\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} T_{t}^{in} < T^{\\min}} \\\\\n{\\pi_{t}E_{t}^{AC} + \\underline{\\kappa}\\left( T_{t}^{in} - T^{\\max} \\right),\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} T_{t}^{in} > T^{\\max}} \\\\\n{\\pi_{t}E_{t}^{AC},\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{otherwise}} \\\\\n\\end{array} \\right.} \\\\\n\\end{array}$$ where $\\overline{\\kappa}$ and $\\underline{\\kappa}$ are the penalties for the consumer's thermal discomfort. The cost of dissatisfaction is defined as the deviation of the consumer's preferred temperature $T_{t}^{in}$ from $T^{\\min}$ and $T^{\\max}$.\n\nIt is noted that two terms in ([20](#FD20-sensors-20-02157){ref-type=\"disp-formula\"}) and ([21](#FD21-sensors-20-02157){ref-type=\"disp-formula\"}) have a trade-off relationship between the saving of the electricity cost and the reduction of the consumer's dissatisfaction cost in terms of the penalties $\\left\\{ \\overline{\\delta},\\underline{\\delta} \\right\\}$ and $\\left\\{ \\overline{\\kappa},\\underline{\\kappa} \\right\\}$, respectively. On the trade-off relationship, HEMS operators using our proposed DRL algorithm can adaptively adjust and tune the penalty to the situations where the consumer aims to save the electricity cost more or maintain the consumer's desired comfort and preference. The selection of the values of these penalties would become different depending on the consumer's desired comfort level and environment.\n\n3.2. Energy Management Model for ESS and EV: Level 2 {#sec3dot2-sensors-20-02157}\n----------------------------------------------------\n\nThe optimal schedules of the energy consumption of the WM and AC from the first level along with the fixed load of the uncontrollable appliances are embedded into the actor--critic module at the second level. In the second level, the agent for the ESS and EV initiates the learning process to determine the optimal charging and discharging schedules of the ESS and EV to minimize the cost of electricity. During the learning process in this second level, the energy generated by the PV system is assumed to be charged first to the ESS; then, the ESS will select an appropriate action.\n\n### 3.2.1. State Space {#sec3dot2dot1-sensors-20-02157}\n\nThe state space of the agent at the second level, which manages the operations of the ESS and EV, is defined as $$\\mathcal{S}^{(2)} = \\left\\{ t,\\pi_{t},SOE_{t}^{ESS},SOE_{t}^{EV},{\\hat{E}}_{t}^{PV},E_{t}^{(1)} \\right\\}$$ where the states *t*, $\\pi_{t}$, $SOE_{t}^{ESS}$, $SOE_{t}^{EV}$, ${\\hat{E}}_{t}^{PV}$, and $E_{t}^{(1)}$ are the scheduling time of the ESS and EV, the TOU price, SOE of the ESS and EV, the predicted PV generation output, and the aggregated energy consumption schedule calculated at the first level, respectively, at time *t*.\n\n### 3.2.2. Action Space {#sec3dot2dot2-sensors-20-02157}\n\nSimilar to the action space of the WM and AC in [Section 3.1.2](#sec3dot1dot2-sensors-20-02157){ref-type=\"sec\"}, the action space of the ESS and EV at the second level is expressed as $$\\mathcal{A}^{(2)} = \\left\\{ E_{t}^{ESS},E_{t}^{EV} \\right\\}$$ where $E_{t}^{ESS}$ and $E_{t}^{EV}$ represent the continuous energy charging and discharging of the ESS and EV, respectively, at time *t*.\n\nNote that, in the proposed two-level DRL architecture, the agent for the ESS and EV selects their optimal charging and discharging action using $E_{t}^{(1)}$ ([22](#FD22-sensors-20-02157){ref-type=\"disp-formula\"}) that includes the action of the agent for the WM and AC along with the fixed load of the uncontrollable appliances at the first level. If the DRL-based HEMS algorithm is modelled as a single-level framework (i.e., the state spaces ([17](#FD17-sensors-20-02157){ref-type=\"disp-formula\"}), ([22](#FD22-sensors-20-02157){ref-type=\"disp-formula\"}) and the action spaces ([18](#FD18-sensors-20-02157){ref-type=\"disp-formula\"}), ([23](#FD23-sensors-20-02157){ref-type=\"disp-formula\"}) at the first and second levels are combined, respectively), the agent for the ESS and EV may not find its optimal policy because the ESS and EV have no consumption data of other appliances in their state space. This is verified in [Section 4.2](#sec4dot2-sensors-20-02157){ref-type=\"sec\"}.\n\n### 3.2.3. Reward {#sec3dot2dot3-sensors-20-02157}\n\nThe reward for the second level is formulated as the sum of the negative cost of electricity and dissatisfaction of ESS and EV associated with the consumer's preferred comfort and appliance's operation characteristics. The total reward at the second level is defined as $$\\mathcal{R}_{t}^{(2)} = - \\left( c_{t}^{ESS} + c_{t}^{EV} \\right).$$\n\nIn ([24](#FD24-sensors-20-02157){ref-type=\"disp-formula\"}), $c_{t}^{ESS}$ and $c_{t}^{EV}$ represent the cost functions for the ESS and EV, respectively. Each cost function includes the cost of electricity of the appliance along with the cost of dissatisfaction for underdischarging and overcharging of the ESS and EV. Notably, these cost functions include the discharging energy from the ESS and EV, which supports the uncovered energy consumption of aggregated load for WM, AC, and uncontrollable appliances.\n\nFirst, the cost function of the ESS is expressed as follows:$$\\begin{array}{r}\n{c_{t}^{ESS} = \\left\\{ \\begin{array}{l}\n{\\pi_{t}E_{t}^{ESS} + \\overline{\\tau}\\left( SOE_{t}^{ESS} - SOE^{{ESS},\\max} \\right),} \\\\\n{\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} SOE_{t}^{ESS} > SOE^{{ESS},\\max}} \\\\\n{\\pi_{t}E_{t}^{ESS} + \\underline{\\tau}\\left( SOE^{{ESS},\\min} - SOE_{t}^{ESS} \\right),} \\\\\n{\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} SOE_{t}^{ESS} < SOE^{{ESS},\\min}} \\\\\n{\\pi_{t}E_{t}^{ESS},\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{otherwise},} \\\\\n\\end{array} \\right.} \\\\\n\\end{array}$$ where $\\overline{\\tau}$ and $\\underline{\\tau}$ are the penalties for ESS overcharging and undercharging, respectively. In this case, energy underutilization and dissipation of the ESS occur if the SOE becomes lower than $SOE^{\\min}$ (undercharging) or greater than $SOE^{\\max}$ (overcharging).\n\nNext, the cost function of the EV is expressed as $$\\begin{array}{r}\n{c_{t}^{EV} = \\left\\{ \\begin{array}{l}\n{\\pi_{t}E_{t}^{EV} + \\overline{\\nu}\\left( SOE_{t}^{EV} - SOE^{{EV},\\max} \\right),} \\\\\n{\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} SOE_{t}^{EV} > SOE^{{EV},\\max},\\mspace{600mu}\\mspace{600mu} t \\in \\left\\lbrack \\omega^{arr},\\omega^{dep} \\right\\rbrack} \\\\\n{\\pi_{t}E_{t}^{EV} + \\underline{\\nu}\\left( SOE^{{EV},\\min} - SOE_{t}^{EV} \\right),} \\\\\n{\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} SOE_{t}^{EV} < SOE^{{EV},\\min},\\mspace{600mu}\\mspace{600mu} t \\in \\left\\lbrack \\omega^{arr},\\omega^{dep} \\right\\rbrack} \\\\\n{\\pi_{t}E_{t}^{EV} + \\underline{\\eta}\\left( SOE^{pref} - SOE_{t}^{EV} \\right),} \\\\\n{\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{if}\\mspace{600mu} t = \\omega^{dep}\\mspace{600mu}{and}\\mspace{600mu} SOE_{t}^{EV} < SOE^{pref}} \\\\\n{\\pi_{t}E_{t}^{EV},\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}\\mspace{600mu}{otherwise},} \\\\\n\\end{array} \\right.} \\\\\n\\end{array}$$ where $\\overline{\\nu}$ and $\\underline{\\nu}$ are the penalties for overcharging and undercharging of the EV, respectively. Similar to the operation of the ESS, energy underutilization and dissipation occur if the SOE of the EV becomes lower than $SOE^{{EV},\\min}$ or higher than $SOE^{{EV},\\max}$. Unlike the reward function of the ESS, the reward function of the EV includes the parameter $\\underline{\\eta}$, which denote the consumer's preference penalty of the EV, corresponding to the deviation of the SOE of the EV from the consumer's preferred SOE when the EV departs. If $SOE_{t}^{EV}$ is lower than $SOE^{pref}$ at departure time $\\omega^{dep}$, the cost of dissatisfaction increases owing to insufficient SOE.\n\n3.3. Proposed Actor--Critic-Based HEMS Algorithm {#sec3dot3-sensors-20-02157}\n------------------------------------------------\n\nIn this subsection, we illustrate the proposed DRL method based on the actor--critic method that determines the optimal policy to minimize the electricity bill within the consumer's preferred comfort level and the appliance operation characteristics. Compared to value-based RL methods, the policy gradient approach is appropriate for engineering problems with continuous action spaces. In general, the continuous policy gradient network obtains state information from the agent and returns the appropriate action using a normal distribution. The network yields the mean and variance to achieve a normal distribution, and the agent samples the action randomly based on the resulting distribution. In the actor--critic approach, the additory method of criticizing the Q-value for its efficiency and convergence is added. Therefore, the network provides the mean, variance, and Q-values to find the optimal actions. As shown in [Figure 2](#sensors-20-02157-f002){ref-type=\"fig\"}, the proposed actor--critic network model for each level consists of one input layer for state elements, four hidden layers for a common body network with 512 neurons, one hidden layer for each actor and critic networks with 256 neurons, and one output layer with means and variances of the operation schedules of appliances and Q-values. In this study, a hyperbolic tangent function was used as a transfer function. In addition, the adaptive moment estimation (ADAM) optimization algorithm \\[[@B35-sensors-20-02157]\\] was used for training the proposed DRL model with a learning rate of 0.00004. Finally, Algorithm 3 illustrates the procedure of the actor--critic-based HEMS algorithm that learns the energy management policies, which optimize the cost of electricity and consumer's comfort level for level 1 (WM and AC) and level 2 (ESS and EV). **Algorithm 3:** Proposed actor--critic-based energy management of smart home at level 1 (or level 2).![](sensors-20-02157-i003.jpg)\n\n4. Numerical Examples {#sec4-sensors-20-02157}\n=====================\n\n4.1. Simulation Setup {#sec4dot1-sensors-20-02157}\n---------------------\n\nUnder the TOU pricing shown in [Figure 3](#sensors-20-02157-f003){ref-type=\"fig\"}a, we considered a household where the proposed DRL algorithm schedules the operation of two tasks: (i) two major controllable home appliances (WM and AC) at the first level and (ii) controllable DERs (PV-integrated ESS and EV) at the second level. The simulations were executed for 24 h with a 1-h scheduling resolution.\n\nWe assumed that the predicted PV generation energy ${\\hat{E}}_{t}^{PV}$ in [Figure 3](#sensors-20-02157-f003){ref-type=\"fig\"}b and predicted outdoor temperature ${\\hat{T}}_{t}^{out}$ in [Figure 3](#sensors-20-02157-f003){ref-type=\"fig\"}c could be calculated accurately. The maximum energy consumptions of the WM and AC were set to 300 and 1000 Wh, respectively. For the ESS, the battery capacity is 4000 Wh, and the minimum ($SOE^{{ESS},\\min}$), maximum ($SOE^{{ESS},\\max}$) and initial SOE ($SOE_{ini}^{ESS}$) of the ESS were set to 400 Wh (10% $SOE^{{ESS},\\max}$), 4000 Wh (100% $SOE^{{ESS},\\max}$), and 2000 Wh (50% $SOE^{{ESS},\\max}$), respectively. The maximum charging and discharging energy of the ESS were both 1200 Wh. For the EV, the battery capacity was 17,000 Wh, and the minimum ($SOE^{{EV},\\min}$), maximum ($SOE^{{EV},\\max}$) and initial SOE ($SOE_{ini}^{EV}$) of the EV were set to 1700 Wh (10% $SOE^{{EV},\\max}$), 17,000 Wh (100% $SOE^{{EV},\\max}$), and 9350 Wh (55% $SOE^{{EV},\\max}$), respectively. The maximum charging and discharging energy of the EV were both 10,000 Wh. For the reward function, the consumer's preferred operating period $\\left\\lbrack \\omega_{s}^{pref},\\omega_{f}^{pref} \\right\\rbrack$ for the WM was set to \\[9:00 a.m., 10:00 p.m.\\] along with 2 h of consecutive operation time. The range of consumer's comfortable indoor temperature $\\left\\lbrack T^{\\min},T^{\\max} \\right\\rbrack$ controlled by the AC was set to \\[22.5 \u00b0C, 25.5 \u00b0C\\]. For the EV, the preferred SOE ($SOE^{pref}$) and the departure time ($\\omega^{dep}$) of the EV were set to 12,750 Wh (75% $SOE^{{EV},\\max}$) and 8:00 a.m., respectively. The pairs of penalties for the cost of dissatisfaction of the WM, AC, ESS, and EV were {$\\left( \\overline{\\delta} = 50,\\underline{\\delta} = 50 \\right)$, $\\left( \\overline{\\kappa} = 200,\\underline{\\kappa} = 200 \\right)$, $\\left( \\overline{\\tau} = 100,\\underline{\\tau} = 100 \\right)$, $\\left( \\overline{\\nu} = 100,\\underline{\\nu} = 100 \\right)$, and $\\left( \\underline{\\eta} = 50 \\right)$}, respectively.\n\nThe performance of the proposed approach was tested for the following four cases according to different weather, weekday/weekend, and initial SOE $SOE_{ini}^{EV}$ of the EV:Case 1: Sunny, weekday, $SOE_{ini}^{EV} = 0.55 \\times SOE^{{EV},\\max},$Case 2: Rainy, weekday, $SOE_{ini}^{EV} = 0.55 \\times SOE^{{EV},\\max},$Case 3: Sunny, weekend, $SOE_{ini}^{EV} = 0.55 \\times SOE^{{EV},\\max},$Case 4: Sunny, weekday, $SOE_{ini}^{EV} = 0.15 \\times SOE^{{EV},\\max}.$\n\nOn a sunny day, the predicted PV generation output at time *t* (${\\hat{E}}_{t}^{PV}$) follows the profile in [Figure 3](#sensors-20-02157-f003){ref-type=\"fig\"}b, and on a rainy day, it is set to zero. On a weekday, the EV is assumed to arrive at the household at 6:00 p.m. and then the charging and discharging processes are conducted until the EV departs from the household at 8:00 a.m. During the weekend, the EV charges or discharges energy during 24 h. $SOE_{ini}^{EV}$ denotes the SOE when the EV arrives at home, and different values of $SOE_{ini}^{EV}$ represent different driving distances of the EV. All the cases were tested using Python 3.7.0 with the machine learning package pytorch 1.1.0.\n\n4.2. Simulation Results at Level 1 {#sec4dot2-sensors-20-02157}\n----------------------------------\n\nIn this subsection, we report the simulation results of the proposed approach associated with level 1 and verify the optimal energy consumption schedule of the WM and AC along with the consumer's comfort level. [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}a shows the energy consumption schedule of the WM. We observe from [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}a that the operation period is selected as \\[7:00 p.m., 8:00 p.m.\\] with two consecutive operation hours. This scheduling policy is optimal because the WM operates at the lowest TOU pricing during the consumer's preferred operation period \\[9:00 a.m., 10:00 p.m.\\], which in turn reduces the electricity bill while satisfying the consumer's preference and operation characteristics of the WM. [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}b illustrates the energy consumption schedule of the AC. Unlike the observation in [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}a, the AC energy consumption is scheduled at an even higher TOU pricing during the period \\[12:00 p.m., 4:00 p.m.\\]. This was expected because the agent at the first level considers the consumer's thermal comfort as well as saving on electricity bills in the reward function. As shown in [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}b, the AC turns off at midnight when the outdoor temperature is within the range of consumer's preferred temperatures. When an indoor temperature violation occurs at 7:00 a.m. owing to a sharp increase of the outdoor temperature, the AC turns on and its energy consumption increases to maintain the consumer's preferred indoor temperature. The maximum energy consumption schedule is verified during the interval \\[12:00 p.m., 4:00 p.m.\\], which presents the highest TOU pricing. In this interval, the AC aims to satisfy the consumer's comfort at the expense of increased electricity bills. [Figure 4](#sensors-20-02157-f004){ref-type=\"fig\"}c shows the total energy consumption schedule of the WM, AC, and uncontrollable appliances with fixed loads at level 1. Note that the sum of energy consumption schedules of WM, AC, and uncontrollable appliances at each period is used by the actor--critic module at the second level, which in turn determines the optimal policy of charging and discharging for the ESS and EV.\n\n4.3. Simulation Results at Level 2 {#sec4dot3-sensors-20-02157}\n----------------------------------\n\nIn this subsection, we present the simulation results of the proposed approach at Level 2. These results are divided into three comparison tests using four cases described in [Section 4.1](#sec4dot1-sensors-20-02157){ref-type=\"sec\"}: {Case 1, Case 2}, {Case 1, Case 3}, and {Case 1, Case 4}. The discharging ratio of the EV and ESS to cover the energy consumption at the first level simultaneously was set to 0.8 and 0.2, respectively. This ratio was determined as the battery capacity of the EV divided by the battery capacity of the ESS.\n\n### 4.3.1. Case 1 vs. Case 2 {#sec4dot3dot1-sensors-20-02157}\n\nIn this simulation, we investigate and compare the performance between Case 1 (with the PV generation output) and Case 2 (without the PV generation output). [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}a,b show the charging/discharging and SOE schedules of the ESS for Case 1 and Case 2, respectively. We observe from [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}a that, in general, more discharging (negative energy consumption) of the ESS for both cases occurs at high TOU pricing to support the household energy demand, thereby leading to consumer's energy savings. In addition, it can be observed through the comparison of [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}a,b that the SOE of the ESS increases (or decreases) as the ESS charges (or discharges) energy. However, Case 1 shows an unexpected phenomenon where the SOE of the ESS in the scheduling period between 10:00 a.m and 4:00 p.m is higher than in other periods even though the ESS conducts a significant energy discharging in this period. This phenomenon occurs because the PV generation energy is injected into the ESS further than the ESS discharging energy in this period.\n\nHowever, we observe from [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}a,b that the ESS charges more power at high TOU pricing for Case 2 than for Case 1. This observation justifies that the different weather conditions influence the charing and discharging schedules of the ESS significantly.\n\n[Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}c,d show the charging/discharging and SOE schedules for the EV. We observe from these figures that the EVs for Cases 1 and 2 does not perform neither charging nor discharging processes after 7:00 a.m. and that the SOE level remains unchanged, respectively. This observation is consistent with the EVs departure time setting ($\\omega^{dep}$=8:00 a.m.). We also observe from [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}c that the EVs charge significantly large amounts of energy from the grid at 10:00 p.m. and 5:00 a.m., whereas it discharges energy to support the household energy demand in the other time periods. This charging process derives from the fact that the EVs charge sufficient energy in advance to satisfy the consumer's preferred SOE condition ($SOE^{pref}$=12,750 Wh). It can be verified from [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}d that the SOE at 8:00 a.m. exceeds 12,750 Wh for both cases. However, no unexpected phenomenon observed in [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}a,b is identified in [Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}c,d. This is because the PV generation output affects only the ESS charging and discharging.\n\n[Figure 5](#sensors-20-02157-f005){ref-type=\"fig\"}e shows the net energy consumptions of the household for Cases 1 and 2, which is the difference between the energy consumption of the controllable/uncontrollable appliances and the predicted PV generation output. As shown in this figure, the values of the net energy consumption of both cases are much larger at 10:00 p.m. and 5:00 a.m. than at the other time slots owing to the large EV charging at these two time slots. However, in the period between 8:00 a.m. and 5:00 p.m., we can identify the large amount of energy consumption in Case 2. This is because the ESS needs to charge more power in Case 2 owing to no PV generation than in Case 1.\n\n### 4.3.2. Case 1 vs. Case 3 {#sec4dot3dot2-sensors-20-02157}\n\nIn this case study, we investigate and compare the performance between Case 1 (on a weekday) and Case 3 (on a weekend). We considered that the EV stays at home during 24 h in the weekend. Given that the EV does not depart, the consumer's preferred SOE of the EV is not considered in the proposed algorithm. First, we can observe from [Figure 6](#sensors-20-02157-f006){ref-type=\"fig\"}a that the SOE of the ESS in Case 3 is generally lower than that in Case1 even if the weather is sunny. We can interpret this observation as follows.\n\nIn Case 3, the ESS does not need to allocate much stored energy to satisfy the consumer's preferred SOE of the EV because the EV stays at home all day. Thus, the ESS increasingly supports the household energy demand through its charging process along with the EV charging. Unlike the results for Cases 1 and 2, we can verify from Case 3 in [Figure 6](#sensors-20-02157-f006){ref-type=\"fig\"}b that the consumer's preferred SOE of EV condition is ignored during the scheduling process of EV energy consumption, in which the SOE of the EV at 8:00 a.m. is much lower than $SOE^{pref}$ = 12,750 Wh. After 8:00 a.m., the EV keeps charging and discharging the energy in the same way as the ESS. We also observe from [Figure 6](#sensors-20-02157-f006){ref-type=\"fig\"}c that in Case 3 a high net energy consumption occurs at 10:00 p.m. and 1:00 p.m. owing to EV charging for discharging plan in future scheduling. Moreover, the net energy consumption at 5:00 a.m. in Case 3 is much smaller than in Case 1 because the consumer's preferred SOE of EV is ignored. Note from Case 3 in [Figure 6](#sensors-20-02157-f006){ref-type=\"fig\"}c that zero-energy consumption is verified in the period between 8:00 a.m. and 5:00 p.m. except 10:00 a.m. and 1:00p.m. This result shows that more energy saving can be obtained during the EV charging and/or discharging during the whole day.\n\n### 4.3.3. Case 1 vs. Case 4 {#sec4dot3dot3-sensors-20-02157}\n\nIn this simulation, we investigate and compare the performance between Case 1 (with a high initial SOE) and Case 4 (with a low initial SOE). [Figure 7](#sensors-20-02157-f007){ref-type=\"fig\"}a,b illustrate the SOE schedules of the ESS and EV for Case 1 and Case 4, respectively. First, we observe from [Figure 7](#sensors-20-02157-f007){ref-type=\"fig\"}b that the SOE of the EV at 6:00 p.m. is larger in Case 4 than in Case 1. This is because the low initial SOE of the EV enable the EV to require more charging power to support the upcoming household load demands. This observation is also verified for the ESS as shown [Figure 7](#sensors-20-02157-f007){ref-type=\"fig\"}a where the ESS charges more power in Case 4 than in Case 1 at 6:00 p.m. with the same reason. We also observe from Case 4 in [Figure 7](#sensors-20-02157-f007){ref-type=\"fig\"}c that the highest net energy consumption occurs at 6:00 p.m., owing to a significant charging of the ESS and EV. It is noted that this large amount of the net energy consumption is not observed in previous cases. This is because in these cases the EV has a high initial SOE so that it does not have to charge energy from grid in advance. Except the time slot at 6:00 p.m., the schedules of the net energy consumption at other time slots are similar to the schedules in Case 1.\n\n[Figure 8](#sensors-20-02157-f008){ref-type=\"fig\"} shows a relative increase of the total electricity bill for the aforementioned Cases 1, 2, and 4 with respect to Case 3 with the minimum total electricity bill using the following metric:$$\\frac{X_{n}^{bill} - X_{3}^{bill}}{X_{3}^{bill}} \\times 100\\left( \\% \\right),$$ where $X_{3}^{bill}$ is the total electricity bill for Case 3 and $X_{n}^{bill}$ is the total electricity bill for Case *n* where $n =$1, 2, and 4.\n\nNote in this figure that the relative total electricity bills for the three cases are listed in decreasing order of their bills as follows: Case 2 \\> Case 4 \\> Case 1. Note in turn from this list that the relative increase of the electricity bill in Case 1 (a sunny weekday with high $SOE_{ini}^{EV}$) is smaller than the other two cases (a sunny or rainy weekday with high or low $SOE_{ini}^{EV}$). Thus, a fraction of the charging and discharging periods of the EV constitutes the most influential aspect for saving on electricity bills.\n\n[Figure 9](#sensors-20-02157-f009){ref-type=\"fig\"}a,b show training curves that present the convergence of the total cost for the first and second levels, respectively. Each figure compares the three training curves using policy gradient method, actor--critic method with separate actor and critic neural networks (NNs), and proposed actor--critic method with a common NN. We observe from [Figure 9](#sensors-20-02157-f009){ref-type=\"fig\"}a,b that policy gradient and actor--critic with separate NNs show a poor performance of the convergence. By contrast, the proposed actor--critic shows that the training curves steadily decrease and then converge to an optimal policy within a moderate training period.\n\n[Figure 10](#sensors-20-02157-f010){ref-type=\"fig\"} compares the training curves for the level 1 and level 2 in the proposed DRL method and the single-level DRL method. We can observe from this figure that each training curve for the level 1 and level 2 in the proposed approach steadily decreases and converges to an optimal policy in the training periods of \\[1, 1,000\\] and \\[1,001, 4,000\\], respectively. By contrast, the training curve for the single-level approach shows a large fluctuation during the training process and a poor value of the result compared to the proposed two-level DRL method. The poor convergence performance of the single-level approach derives from the fact that the complexity of the HEMS problem increases dramatically as the state-action space dimension becomes larger in the single level. Furthermore, since the agent for the EV and ESS is not able to obtain energy consumption data of other appliances in the single level, the single-level approach may not calculate optimal charging and discharging actions of the ESS and EV.\n\n[Figure 11](#sensors-20-02157-f011){ref-type=\"fig\"} shows a relative increase of the total electricity bill in Cases 1--4 using a Building Energy Optimization Tool (BEopt) \\[[@B36-sensors-20-02157]\\] and MILP optimization method with respect to the proposed approach. BEopt is widely used as an energy simulation program for the residential building. To fairly compare the performance of the proposed method to that of the MILP method, the MILP method was executed for 24 h with a 1-h scheduling resolution. We can verify from this figure that our proposed method is the most economical for all four cases compared to two existing approaches using BEopt and MILP methods. In particular, it is observed that Case 3 (a sunny weekend with high $SOE_{ini}^{EV}$) shows the largest cost increase. This observation implies that the proposed DRL approach schedules the charging and discharging energy of EV in a much more cost-effective way.\n\nThe meaningful observations of the proposed HEMS approach in the numerical examples can be summarized as follows:Through a comparison between {Case 1, Case 3, Case 4} (with PV system) and Case 2 (without PV system), we conclude that PV generation has a significant impact on the reduction of the total cost of electricity. For example, the total cost of electricity in Case 2 is 11% higher than in Case 1.Given that the battery capacity of the EV is approximately four times larger than that of the ESS, the EV can discharge more power than the ESS to cover the total cost of electricity. This can be verified through a comparison between {Case 1, Case 2} (in weekday) and Case 3 on weekends. In contrast, different driving patterns associated with the initial SOE of the EV significantly influence the total cost of electricity. We conclude from a comparison between Case 1 (with high SOE) and Case 4 (with low SOE) that the total cost of electricity in Case 4 is 7% higher than in Case 1. This is because the EV with low SOE needs to charge more power than with high SOE to satisfy the consumer's preferred SOE at departure time.\n\n5. Conclusions {#sec5-sensors-20-02157}\n==============\n\nIn this study, we propose a two-level distributed deep reinforcement learning algorithm to minimize the cost of electricity through the energy consumption scheduling of two controllable home appliances (an air conditioner and a washing machine) and the charging and discharging of an energy storage system and an electric vehicle while maintaining the consumer's comfort level and appliance operation characteristics. In the proposed deep reinforcement learning method, two agents interact with each other to schedule the optimal home energy consumption efficiently. One agent for a washing machine and an air conditioner determines their continuous actions in the first level to schedule optimal energy consumption within the consumer's preferred indoor temperature and operation period, respectively. Based on the optimal energy consumption schedules from the first level, the other agent for an energy storage system and an electric vehicle conducts their continuous charging and discharging actions in the second level to support the aggregated load for controllable and uncontrollable appliances. The comparative case studies under different weather and driving patterns of the electric vehicle with different initial state of energy confirm that the proposed approach can successfully minimize the cost of electricity within the consumer's preference.\n\nIn future work, we plan to develop a multi-agent reinforcement learning algorithm based on a continuous action space that schedules the energy consumption of multiple smart homes with home appliances and distributed energy resources. A key challenge is to design an information exchange scheme between households to minimize the cost of electricity and maintain each consumer's comfort level. This future work can be implemented using advanced deep reinforcement learning methods such as deep deterministic policy gradient and asynchronous advantage actor--critic methods.\n\nS.L. proposed the DRL-based home energy system model and conducted the simulation study. D.-H.C. coordinated the approach that is proposed in this paper. All authors have read and agreed to the published version of the manuscript.\n\nThis work was supported in part by the National Research Foundation of Korea (NRF) Grant through the Korea Government (MSIP) under Grant 2018R1C1B6000965, and in part by the Competency Development Program for Industry Specialists of the Korean Ministry of Trade, Industry and Energy (MOTIE), operated by Korea Institute for Advancement of Technology (KIAT) under Grant P0002397 HRD program for Industrial Convergence of Wearable Smart Devices.\n\nThe authors declare no conflict of interest.\n\n![Conceptual architecture of the proposed deep reinforcement learning (DRL)-based home energy management system (HEMS) algorithm.](sensors-20-02157-g001){#sensors-20-02157-f001}\n\n![Architecture of the neural network model for the proposed actor--critic method.](sensors-20-02157-g002){#sensors-20-02157-f002}\n\n![Profiles of electricity price and weather. (**a**) time-of-use (TOU) price; (**b**) photovoltaic (PV) generation; (**c**) outdoor temperature.](sensors-20-02157-g003){#sensors-20-02157-f003}\n\n![DRL-based energy consumption schedule. (**a**) washing machine (WM); (**b**) air conditioner (AC); (**c**) WM+AC+Uncontrollable appliances at Level 1.](sensors-20-02157-g004){#sensors-20-02157-f004}\n\n![Performance comparison between Case 1 and Case 2. (**a**) energy consumption of the energy storage system (ESS); (**b**) state of energy (SOE) of the ESS; (**c**) energy consumption of the electric vehicle (EV); (**d**) SOE of the EV; (**e**) net consumption of household.](sensors-20-02157-g005){#sensors-20-02157-f005}\n\n![Performance comparison between Case 1 and Case 3. (**a**) SOE of the ESS; (**b**) SOE of the EV; (**c**) net consumption of household.](sensors-20-02157-g006){#sensors-20-02157-f006}\n\n![Performance comparison between Case 1 and Case 4. (**a**) SOE of the ESS; (**b**) SOE of the EV; (**c**) net consumption of household.](sensors-20-02157-g007){#sensors-20-02157-f007}\n\n![Comparison of a relative increase of the total electricity bill in Case 3 among the considered three cases.](sensors-20-02157-g008){#sensors-20-02157-f008}\n\n![Convergence of the total cost. (**a**) Level 1; (**b**) Level 2.](sensors-20-02157-g009){#sensors-20-02157-f009}\n\n![Comparison of the total cost convergence between the proposed two-level DRL approach and the single-level DRL approach.](sensors-20-02157-g010){#sensors-20-02157-f010}\n\n![Comparison of a relative increase of the total electricity bill in the proposed DRL method using building energy optimization tool (BEopt) and mixed-integer linear programming (MILP) programs for four cases.](sensors-20-02157-g011){#sensors-20-02157-f011}\n"}
+{"text": "![](envhper00396-0015-color.jpg \"scanned-page\"){.733}\n\n![](envhper00396-0016-color.jpg \"scanned-page\"){.734}\n"}
+{"text": "An 18-year-old male presented with painful papulo-nodular lesions on the upper limbs and back along with high-grade fever since 1 week. Most lesions had a lichenoid center with surrounding blisters ([Figure 1A](#f1){ref-type=\"fig\"}). There was associated tenderness and thickening of bilateral ulnar nerves. Slit skin smear was positive for *Mycobacterium leprae* (with a bacteriological index of 6+ and a morphological index of 0%) ([Figure 1B](#f1){ref-type=\"fig\"}). Skin biopsy revealed features of lepromatous leprosy (LL) with erythema nodosum leprosum (ENL) including granulomas, nerve destruction, vasculitis, and panniculitis, and lepra stain was positive ([Figure 1C](#f1){ref-type=\"fig\"}). Over the next 2 weeks, most lesions developed ulceration, suggesting necrotic ENL ([Figure 1D](#f1){ref-type=\"fig\"}). The patient is currently receiving WHO-recommended multidrug therapy (MDT) with oral corticosteroids.\n\n![(**A**) Patient with multiple papulo-nodules, having a lichenoid center with surrounding blisters, on upper limb and back. (**B**) Slit skin smear of the same patient showing acid fast bacilli. (**C**) Histopathological examination of skin biopsy showing granulomatous inflammation and vasculitis. (**D**) Same patient with ulcerative changes in lesions within 2 weeks, suggesting necrotic ENL. This figure appears in color at [www.ajtmh.org](http://www.ajtmh.org).](tpmd200177f1){#f1}\n\nErythema nodosum leprosum is a type III hypersensitivity reaction presenting as erythematous, tender papulo-nodules, commonly on extremities, with histology showing granulomatous inflammation with neutrophil infiltration, and variable degrees of panniculitis and vasculitis.^[@b1],[@b2]^ Reported incidence of ENL in multibacillary leprosy (especially LL) is as high as 11.8%.^[@b3]^ The lesions usually appear after starting MDT, although it may sometimes be the presenting feature.^[@b4]^ Rarely, ENL may be associated with ulceration, producing necrotic ENL, also called erythema necroticans. Necrotic ENL is a severe form often associated with systemic complications and heals with scarring.^[@b5]^ The case illustrates that lichenoid skin nodules may be an unusual initial presentation of necrotic ENL.\n\nWe thank Bishan Das Radotra for his inputs regarding histopathological diagnosis.\n\n[^1]: Authors' addresses: Akanksha Kaushik, Sunil Dogra, and Tarun Narang, Department of Dermatology, Venereology and Leprology, Postgraduate Institute of Medical Education and Research, Chandigarh, India, E-mails: <akskaushikk@gmail.com>, <sundogra@hotmail.com>, and <narangtarun@yahoo.co.in>.\n"}
+{"text": "Background\n==========\n\nOne major drawback in measuring ground-reaction forces during running is that it is time consuming to get representative ground-reaction force (GRF) values with a traditional force platform. A single force platform is only capable of measuring GRFs of one single stance phase per trial \\[[@B1],[@B2]\\]. Therefore, multiple force platforms are necessary for measuring consecutive steps which is space consuming and expensive. The limited length of a runway, also makes it difficult to simulate natural running at a constant speed in a laboratory situation \\[[@B3]\\]. For detection of small differences in GRFs during running, however, it is important to record sufficient steps during a stable running pattern \\[[@B4]\\]. An instrumented treadmill capable of measuring GRFs can overcome the limitations inherent to overground GRF testing during running at a short runway. With an instrumented treadmill it is possible to measure GRFs of multiple steps during one trial without interruptions in running speed, resulting in a more stable running pattern during the measurements \\[[@B3]\\].\n\nIn running, most runners make first ground contact with the posterior part of the foot, this is called heelstrike running. This running style results in a typical vertical GRF force-time curve that is characterized by two peaks, the impact peak and the active peak, as depicted in Figure\u2009[1](#F1){ref-type=\"fig\"}. Magnitude of the impact peak is speed dependent and occurs during the first 10% of stance (10-30ms) \\[[@B5]\\]. The active peak is reached approximately during mid-stance and can last up to 200ms. The absence of a separate impact peak in the force-time curve is typical for a non-heelstrike runner, as depicted in Figure\u2009[1](#F1){ref-type=\"fig\"}\\[[@B6]\\]. Besides a vertical component, GRFs also have an anterior-posterior and medio-lateral component. During running, the anterior-posterior force component shows a typical braking and propulsive phase while the medio-lateral force component is characterized by more variability \\[[@B7]\\]. Compared to the vertical GRF component, anterior-posterior and medio-lateral forces are small \\[[@B7]\\].\n\n![**Outcome measures in a typical vertical ground-reaction force (GRF) curve for a heelstrike runner and a non-heelstrike runner.** Figure is created from personal data.](1471-2474-13-235-1){#F1}\n\nAn underlying assumption when using a treadmill for running analysis is that running on a treadmill is similar to overground running. A comparison of spatio-temporal variables during overground and treadmill running was made in several studies. During treadmill running, runners tend to run with a shortened stride length and an increased stride rate \\[[@B3],[@B8],[@B9]\\]. Despite of these spatio-temporal differences, only small differences in knee flexion and a more flattened landing style during treadmill running were observed \\[[@B3],[@B10]\\]. An overground-treadmill comparison with respect to GRFs was made in only two studies \\[[@B1],[@B3]\\]. No systematic errors or extraordinary differences in vertical GRFs were found. Impact peaks and loading rates, however, have not been studied in these previous studies.\n\nThe purpose of this study was to determine the validity of a custom made instrumented force measuring treadmill to measure vertical GRF parameters during running. Validation of the treadmill was performed by comparing overground and treadmill measured vertical GRF parameters during running.\n\nMethods\n=======\n\nParticipants\n------------\n\nTwenty-four experienced runners (12 male, 12 female) between 18 and 35 years old participated in this study. The runners were voluntarily recruited by contacting two local track and field clubs. The criteria for inclusion in this study included a minimal training frequency of two times a week for at least a period of one year. Runners who reported an injury at time of measurement were excluded. Both heelstrike and non-heelstrike runners were allowed to participate in this study. All participants signed an informed consent before measurements started. The study was approved by the Medical Ethical Committee at the University Medical Center Groningen, The Netherlands; M12.112668.\n\nOverground measurements\n-----------------------\n\nDuring the overground measurements, GRFs were measured at three different individual speed conditions. Participants were instructed to run at their preferred speed (running speed for a normal endurance run), slower speed (running speed during a warming-up), and a faster speed (10km race speed) respectively. GRFs were collected with a force platform (0.60m x 0.40m) which was mounted in the middle of a 17.5m long runway. The sample frequency of the force platform was set at 1000Hz. Running speed was monitored with two pairs of photocells placed 2.5m before and after the force platform.\n\nBefore the actual overground measurements started, the participants performed several accommodation runs. During these accommodation runs, the exact start position for the measurements was determined. The start position was based on the position of foot placement at the force platform. Foot strike of the right foot should be completely at the force platform without an alteration in running pattern. An alteration can indicate aiming for the force platform, which modifies the GRF pattern \\[[@B11]\\]. Position of foot placement and running pattern were evaluated on sight. When participants were able to run several trials at the same speed, while landing with the right foot completely placed at the force platform, without visible alterations in running pattern, the actual measurements started. Since the participants were tested at three different speed conditions, accommodation runs were performed for each speed condition (preferred, slow and fast). The accommodation runs for the preferred speed were combined with a short warming-up and took longer (approximately 10 min), where the subsequent accommodation runs took approximately 5 minutes.\n\nDuring the actual measurements, GRF data were captured until five clean strikes of the right leg within a 5% speed range were recorded for all speed conditions. Trials with visible alterations in running pattern were not included in these five clean strikes. Afterwards, the mean running speed of the five steps was calculated for each speed condition.\n\nTreadmill measurements\n----------------------\n\nIn this study, an instrumented treadmill (Entred, Forcelink, Culemborg, The Netherlands) with a running surface of 1.60m by 0.60m that was driven by a 1.8 kW motor was used to measure vertical GRFs during running. The treadmill was equipped with three strain gage force transducers (ACB-500kg, Vishay Revere Transducers, Breda, The Netherlands) which were connected to bridge amplifiers. The force transducers were mounted on a stiff plate which was enforced with a non-deformable frame and were positioned as shown in Figure\u2009[2](#F2){ref-type=\"fig\"}. The signals from the amplifiers were sampled at 1000 Hz, digitized into a 16-bit signal by an AD converter (PCI-6220, National Instruments, Austin, TX, USA) and were connected to a computer.\n\n![Positioning of the three strain gage force transducers S1, S2 and S3.](1471-2474-13-235-2){#F2}\n\nBefore the treadmill measurements started, participants started with an accommodation run of 10 minutes at 10 km\u00b7h^-1^. After this accommodation period, participants were tested at three different individual speed conditions (slow, preferred and fast). Treadmill speed was matched to the average overground running speed for each speed condition because GRF parameters are speed dependent \\[[@B7]\\]. The three speed conditions lasted three minutes and were offered in random order. GRFs were recorded during the last 30 seconds of each speed condition. When treadmill measurements were finished, participants were given the opportunity for cooling-down at the treadmill. All measurements were conducted while participants were running in their personal running shoes.\n\nData analysis\n-------------\n\nVertical force data from both the force platform and the treadmill were processed using custom programs written in MATLAB R2010a (The MathWorks, Inc, Natick, MA). All steps which were recorded during the treadmill measurement were entered into the analysis. A 13-point moving average low-pass filter with a cut-off frequency of 33.3Hz was used to filter the GRF data that was recorded during the overground and treadmill measurements. Foot strikes in the overground and treadmill data were detected with a threshold of 30 Newton for impact and toe-off phase. Outcome measures for all right foot steps were identified, as described in Table\u2009[1](#T1){ref-type=\"table\"}. For each speed condition outcome measures of each participant were averaged. A distinction between heelstrike and non-heelstrike landing patterns was made based on the existence of an impact peak Fz1. Peak values Fz1 and Fz2 and the loading-rate were normalized to bodyweight.\n\n###### \n\n**Definition of outcome measures, as displayed in Figure**[1](#F1){ref-type=\"fig\"}\n\n  **Outcome measure**   **Description**\n  --------------------- ---------------------------------------------------------------------------------------------\n  Fz1                   Local maximum in the vertical GRF data, normalized to body weight (BW).\n  Fz2                   Maximum value in the vertical GRF data, normalized to BW.\n  LR                    The steepest part of the vertical GRF curve, from stance to impact peak. Expressed in BW/s.\n  ALR                   Average loading rate, the slope of the line from 20% to 80% of Fz1. Expressed in BW/s.\n  tFz1                  Time from heelstrike to Fz1 in ms.\n  tFz2                  Time from heelstrike to Fz2 in ms.\n  CT                    Contact time, from heelstrike to toe-off in ms.\n\nOutcome measures for the overground and treadmill data were identified with the same routine. Foot strikes were detected with a threshold of 30 Newton for both heelstrike and toe-off.\n\nStatistical analysis\n--------------------\n\nA within-subject repeated measures design was used to determine the validity of the instrumented treadmill for measuring vertical GRF parameters during running. Therefore, a two-way mixed-effects, consistency, single measure intraclass correlation coefficient (ICC~(3,1)~) model was used to examine the agreement between overground and treadmill measured GRF-parameters. Interpretation of the intraclass coefficients were as follows: poor (0 -- 0.39), modest (0.4 -- 0.74), or excellent (0.75 -- 1) \\[[@B12]\\]. ICCs were calculated by using SPSS (SPSS inc. Version 18.0, Chicago, IL, U.S.A.). Besides the intraclass correlations, Bland-Altman plots were used to examine the agreement between overground and treadmill measurements \\[[@B13]\\]. These plots were made for each outcome measure and each speed condition. The limits of agreement (LOA) were calculated (mean difference +/\u2212 1.96 times the standard deviation of the difference). Also ratio limits of agreement (RLOA) were calculated to express the LOA as percentage of the mean overground-treadmill value. The upper and lower LOA and the RLOA provide insight into how much random variation may be influencing the measurements.\n\nResults\n=======\n\nGround-reaction force (GRF) parameters of a different landing strategy cannot be compared, therefore only GRF parameters of participants who showed a consistent landing strategy during overground and treadmill running within a speed condition were examined. During overground running at preferred speed, 19 participants showed a heelstrike (HS) landing, while 16 of these runners showed a HS landing during treadmill running. This shows that 82.4% of the runners used a similar landing strategy during treadmill running at preferred speed. Results for the two other speeds can be found in Table\u2009[2](#T2){ref-type=\"table\"}.\n\n###### \n\nOverground landing strategy compared to treadmill landing strategy, displayed as number of persons and corresponding percentages of runners who showed a consistent landing strategy\n\n  \u00a0            **Heelstrike landing**  **Non-heelstrike landing**                     \n  ----------- ------------------------ ---------------------------- -------- ---- --- --------\n  Slow                   17            14                           82.4%     7   5   71.4%\n  Preferred              19            16                           84.2%     5   5   100.0%\n  Fast                   12            12                           100.0%    12  6   50.0%\n\nQualitatively, the overground and treadmill GRF curves for both HS and NHS running at slow, preferred and fast running speeds, were very similar, as can be seen in Figure\u2009[3](#F3){ref-type=\"fig\"}. In Table\u2009[3](#T3){ref-type=\"table\"} a quantitative evaluation of the vertical GRF-parameters of both HS and NHS runners can be found. The levels of agreement between overground and treadmill running for the time related variables (tFz1, tFz2 and CT) were excellent (ICCs between 0.76 and 0.95 and RLOAs between 5.7% and 15.5%). Also the active peak (Fz2) measured with both devices showed excellent agreement (ICCs between 0.77 and 0.89, RLOAs between 7.8% and 9.9%). Modest agreement was found for the impact peak, Fz1 (ICCs between 0.71 and 0.76, RLOAs between 19.9% and 28.8%). Maximal loading rate (LR) and average loading rate (ALR) also showed modest to excellent intraclass correlations (ICCs between 0.70 and 0.89), however the ratio limits of agreement were higher (RLOA values between 34.3% and 45.4%).\n\n![**Average GRF plots from all runners for overground (mean, solid black line; \u00b1 SD, dotted black lines) and treadmill running (mean, solid grey line) at slow, preferred and fast running speed for heelstrike and non-heelstrike runners.** Forces are in body weight (BW).](1471-2474-13-235-3){#F3}\n\n###### \n\nOutcome measures for overground and treadmill running\n\n  \u00a0                   **OG mean \u00b1 SD**   **TM mean \u00b1 SD**   **ICC**~**(3,1)**~**(95%CI)**   **Mean diff (LOA) diff (lowLim, upLim)**       **RLOA (%)**      \n  ------------------ ------------------ ------------------ ------------------------------- ------------------------------------------ ---------------------- ------\n  *Fz1* (*BW*)               \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow           1.67 \u00b1 0.26               1.70 \u00b1 0.23                       0.74 (0.37, 0.91)                0.03 (\u22120.32, 0.38)    20.8\n  \u00a0                      Preferred         1.94 \u00b1 0.45               1.93 \u00b1 0.30                       0.71 (0.35, 0.89)               \u22120.01 (\u22120.57, 0.55)    28.8\n  \u00a0                         Fast           1.94 \u00b1 0.25               2.06 \u00b1 0.32                       0.76 (0.35, 0.92)               0.12 ( \u22120.28, 0.52)    19.9\n  *Fz2* (*BW*)               \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow           2.54 \u00b1 0.20               2.53 \u00b1 0.18                       0.77 (0.49, 0.91)               \u22120.02 (\u22120.27, 0.23)    9.9\n  \u00a0                      Preferred         2.70 \u00b1 0.26               2.65 \u00b1 0.25                       0.89 (0.76, 0.96)               \u22120.03 (\u22120.25, 0.17)    7.9\n  \u00a0                         Fast           2.77 \u00b1 0.24               2.70 \u00b1 0.22                       0.86 (0.67, 0.95)               \u22120.06 (\u22120.27, 0.15)    7.8\n  NHS                       Slow           2.56 \u00b1 0.17               2.55 \u00b1 0.20                               \u00a0                                \u00a0              \u00a0\n  \u00a0                      Preferred         2.61 \u00b1 0.15               2.58 \u00b1 0.13                               \u00a0                                \u00a0              \u00a0\n  \u00a0                         Fast           2.79 \u00b1 0.15               2.78 \u00b1 0.20                               \u00a0                                \u00a0              \u00a0\n  *LR* (*BW*/*s*)            \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow          81.11 \u00b1 25.62             87.28 \u00b1 23.39                      0.76 (0.47, 0.90)               3.25 (\u221228.62, 35.12)   39.9\n  \u00a0                      Preferred        95.34 \u00b1 26.67           105\\. 33 \u00b1 25.08                     0.80 (0.57, 0.91)               6.11 (\u221226.21, 38.42)   34.3\n  \u00a0                         Fast          104.40 \u00b1 29.29           118.08 \u00b1 33.73                      0.70 (0.36, 0.88)               7.17 (\u221237.69, 52.02)   42.7\n  NHS                       Slow          70.03 \u00b1 14.68             65.09 \u00b1 13.74                              \u00a0                                \u00a0              \u00a0\n  \u00a0                      Preferred        77.00 \u00b1 22.35             74.25 \u00b1 16.47                              \u00a0                                \u00a0              \u00a0\n  \u00a0                         Fast          95.81 \u00b1 26.02             87.41 \u00b1 18.74                              \u00a0                                \u00a0              \u00a0\n  *ALR* (*BW*/*s*)           \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow          68.89 \u00b1 20.26             73.92 \u00b1 20.22                      0.84 (0.63, 0.93)               2.98 (\u221224.74, 30.70)   45.3\n  \u00a0                      Preferred        82.14 \u00b1 21.38             88.70 \u00b1 20.75                      0.89 (0.74, 0.95)               3.60 (\u221223.01, 30.21)   36.4\n  \u00a0                         Fast          90.70 \u00b1 23.66            100.77 \u00b1 29.10                      0.86 (0.67, 0.95)               4.08 (\u221231.26, 39.42)   45.4\n  NHS                       Slow           33.96 \u00b1 6.07             31.21 \u00b1 5.01                               \u00a0                                \u00a0              \u00a0\n  \u00a0                      Preferred        47.09 \u00b1 22.92             33.78 \u00b1 4.20                               \u00a0                                \u00a0              \u00a0\n  \u00a0                         Fast          43.63 \u00b1 13.89             36.00 \u00b1 4.20                               \u00a0                                \u00a0              \u00a0\n  *tFz1* (*ms*)              \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow            35 \u00b1 4.08                 35 \u00b1 4.86                        0.76 (0.40, 0.92)                 0.0 ( \u22125.4, 5.4)     15.5\n  \u00a0                      Preferred          34 \u00b1 4.42                 34 \u00b1 3.35                        0.82 (0.56, 0.93)                 0.3 ( \u22124.4, 5.0)     13.8\n  \u00a0                         Fast            32 \u00b1 5.00                 33 \u00b1 4.88                        0.87 (0.61, 0.96)                 0.6 ( \u22123.7, 4.8)     13.0\n  *tFz2* (*ms*)              \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow           112 \u00b1 13.55               109 \u00b1 10.22                       0.84 (0.63, 0.94)                \u22121.8 (\u221215.4, 11.8)    12.6\n  \u00a0                      Preferred         102 \u00b1 13.28               100 \u00b1 12.19                       0.94 (0.85, 0.97)                \u22121.6 (\u221210.1, 6.8)     8.5\n  \u00a0                         Fast            99 \u00b1 10.00               96 \u00b1 11.47                        0.87 (0.68, 0.95)                \u22123.0 (\u221213.5, 7.5)     11.0\n  NHS                       Slow           102 \u00b1 13.00               103 \u00b1 15.00                               \u00a0                                \u00a0              \u00a0\n  \u00a0                      Preferred          99 \u00b1 12.00               98 \u00b1 11.00                                \u00a0                                \u00a0              \u00a0\n  \u00a0                         Fast            92 \u00b1 80 0                91 \u00b1 10.00                                \u00a0                                \u00a0              \u00a0\n  *CT* (*ms*)                \u00a0                  \u00a0                         \u00a0                                    \u00a0                                \u00a0            \n  HS                        Slow           258 \u00b1 22.00               254 \u00b1 21.13                       0.92 (0.80, 0.97)                \u22124.0 (\u221221.4, 13.4)    6.9\n  \u00a0                      Preferred         232 \u00b1 23.34               232 \u00b1 20.49                       0.92 (0.82, 0.97)                \u22122.0 (\u221217.2, 13.2)    6.6\n  \u00a0                         Fast           223 \u00b1 21.00               220 \u00b1 21.14                       0.95 (0.87, 0.98)                \u22123.3 (\u221215.6, 9.1)     5.7\n  NHS                       Slow           240 \u00b1 17.00               237 \u00b1 19.00                               \u00a0                                \u00a0              \u00a0\n  \u00a0                      Preferred         229 \u00b1 12.00               222 \u00b1 12.00                               \u00a0                                \u00a0              \u00a0\n  \u00a0                         Fast           213 \u00b1 12.00               206 \u00b1 12.00                               \u00a0                                \u00a0              \u00a0\n\nIntraclass correlations, mean-differences with limits of agreement (LOA), and ratio limits of agreement (RLOA) were reported. Both HS and NHS runners were taken into account in the statistical analysis. Number of participants: HS (slow: N=14, preferred: N=16, fast: N=12), NHS (slow: N=5, preferred: N=5, fast: N=6).\n\nHS: Heelstrike-runner, NHS: Non-Heelstrike-runner, CI: Confidence Interval, LOA: Limit of Agreement, RLOA: Ratio Limit of Agreement, OG: Overground, TM: Treadmill, BW: Body Weight.\n\nSlow running speed: HS runners: 11.0 \u00b1 1.3 km\u00b7h^-1^, NHS runners: 10.9 \u00b1 1.5 km\u00b7h^-1^.\n\nPreferred running speed: HS runners: 12.7 \u00b1 1.6 km\u00b7h^-1^, NHS runners: 11.8 \u00b1 1.5 km\u00b7h^-1^.\n\nFast running speed: HS runners: 14.1 \u00b1 2.0 km\u00b7h^-1^, NHS runners: 13.9 \u00b1 1.9 km\u00b7h^-1^.\n\nDiscussion\n==========\n\nThe instrumented treadmill is capable of measuring vertical ground-reaction forces (GRFs) during running and seems to be a usable tool for simulating overground running kinetics. The results of this study demonstrated that the instrumented treadmill is a highly valid tool for the assessment of the vertical GRF parameters: tFz1, tFz2, CT and Fz2 and moderately valid for the assessment of Fz1, LR and ALR for runners who showed a consistent landing strategy during overground and treadmill running. A qualitative evaluation of the overground and treadmill vertical GRF curves as shown in Figure\u2009[3](#F3){ref-type=\"fig\"}, demonstrated that the vertical GRFs for both the heelstrike (HS) runners and the non-heelstrike (NHS) runners were similar during overground and treadmill running. The excellent intraclass correlations and low limits of agreement for contact time (CT), time to impact peak force (tFz1) and time to the active peak (tFz2) reflect this qualitative similarity. After all, these parameters show that the timing of peak values in the vertical GRF curve is not different for overground and treadmill running. The qualitative similarity of these GRF curves was also observed in other studies \\[[@B1],[@B3]\\]. In the current study, the overground and treadmill measured active peak (Fz2) showed no noteworthy differences. This is in accordance with the results of Riley et al., who also compared overground and treadmill running kinetics in a group of 20 runners \\[[@B3]\\]. Overground and treadmill measured impact peaks (Fz1), maximal loading rates (LR) and average loading rates (ALR), showed less consistent results with modest to excellent intraclass correlations and wider limits of agreement. To our knowledge this study is the first to compare overground and treadmill measured impact peaks and loading rates during running, therefore it is not possible to evaluate these results with previous studies.\n\nFor an overground-treadmill comparison with respect to vertical GRF parameters, a consistent landing strategy during both running conditions (overground and treadmill) is required. While most runners showed a consistent landing strategy during overground and treadmill running, some runners switched to another landing strategy. During slow and preferred running speed, these inconsistent runners mostly switched from an overground HS landing to a NHS landing during treadmill running. Considering that this behavior is in line with the more flattened landing style as observed in a previous study \\[[@B14]\\], it is likely that these inconsistencies in landing strategy are the result of accommodation to treadmill running. At fast self selected speed, however, the inconsistent runners switched from a NHS to a HS landing during treadmill running. These differences in landing strategy may indicate overground and treadmill differences in anterior-posterior GRFs which were not compared in the current study. The results of this study demonstrated that the inconsistencies in landing strategy are smallest during running at preferred speed. Therefore, to maximize certainty, it can be recommended to determine landing strategy with a treadmill measurement at preferred running speed.\n\nThe use of a treadmill in a research setting has been subject of much debate. Several factors are mentioned which may cause biomechanical differences between overground and treadmill running \\[[@B9]\\]. First, non-mechanical factors as accommodation to the changed visual and auditory surroundings or fear during treadmill running may result in differences between overground and treadmill running biomechanics \\[[@B15]\\]. Second, differences in air resistance may have an effect on treadmill running form \\[[@B16]\\]. The effects of air resistance on running kinematics, however, will only be visible during running at high speeds \\[[@B17]\\]. Third, intra-stride belt speed variations, due to an energy exchange between the treadmill and the runner, can cause differences in running kinematics compared to overground running. In particular low powered treadmills are more sensitive for opposite forces acting on the belt during running, resulting in larger belt speed variations. These variations in belt speed may lead to biomechanical differences during treadmill running when compared to overground running \\[[@B15]\\]. Fourth, during running, leg stiffness is adjusted to the stiffness of the running surface \\[[@B18]\\]. Adjusting leg stiffness results in subtle changes in the kinematics of the lower extremity \\[[@B19]\\]. Therefore, differences in running surface may lead to biomechanical differences when comparing overground and treadmill running.\n\nSeveral studies compared overground and treadmill running biomechanics \\[[@B3],[@B8],[@B14]\\]. Even though runners tend to run with a shortened stride length and an increased stride rate during treadmill running \\[[@B3],[@B8],[@B9]\\], overground and treadmill running kinematics are remarkably similar \\[[@B3],[@B9],[@B14]\\]. Only small differences in knee and ankle joint kinematics were reported. Nigg et al. observed a more flattened landing style during treadmill running \\[[@B14]\\]. Riley et al. did not find differences in ankle joint kinematics, but did find differences in minimal and maximal knee flexion \\[[@B3]\\]. Maximal knee flexion was lower and minimal flexion was higher during treadmill running, which could be a result of the observed decrease in flight phase and higher stride rate \\[[@B3]\\]. Thus, despite the theoretical factors which may influence treadmill running biomechanics, only small differences in overground and treadmill kinematics were observed. In the current study, also no significant differences in GRF parameters between overground and treadmill running were found. These findings are in line with previous studies where overground and treadmill running kinetics were compared \\[[@B1],[@B3]\\]. The between person variance in Fz1, LR and ALR during both overground and treadmill running was high, as indicated by the high standard deviations for these parameters. Stride-to-stride variance for these parameters was also high, which demonstrates the importance of measuring sufficient steps for representative GRF values. This is especially important for detecting small differences between different conditions or persons \\[[@B20]\\]. Because a treadmill makes it possible to measure multiple steps during one test trial, it can be argued that a treadmill measurement is more suitable for detecting small differences in vertical GRFs during running. However, this assumption was not assessed in the current study.\n\nSince the treadmill used in the current study only is capable of measuring vertical GRFs it cannot be used to assess joint kinetics using the standard inverse dynamics methodology, because anterior-posterior and medio-lateral GRFs are also needed for these calculations. It should also be noted that the inconsistencies in landing strategy may indicate differences in anterior-posterior GRFs between overground and treadmill running. Furthermore, this instrumented treadmill would have limited usefulness for walking studies, because the double support phase in walking cannot be measured directly. For measuring GRFs during walking, an instrumented split-belt treadmill may be more convenient.\n\nA limitation of this study was that participants first performed the overground measurements after which the treadmill measurements started. Due to this fixed order of the measurements, fatigue may have influenced the later treadmill measurements \\[[@B21]\\]. Nevertheless, this influence is expected to be low, since all participants were experienced runners who did not have to deliver a maximal performance and participants did not show signs of exaggerated fatigue during the measurements.\n\nConclusions\n===========\n\nThe results of this study demonstrated the treadmill is a moderate to highly valid tool for the assessment of vertical GRFs during running for runners who showed a consistent landing strategy during overground and treadmill running. Therefore, an instrumented treadmill can be used to measure vertical GRF parameters which correspond to normal overground values during running.\n\nIn a future study, the treadmill can be used to measure vertical GRF parameters in a large group of runners, for instance to identify possible kinetic risk-factors for running related injuries prospectively.\n\nAbbreviations\n=============\n\nGRF: Ground-reaction force; HS: Heelstrike; NHS: Non-heelstrike; Fz1: Impact peak; Fz2: Active peak; LR: Loading Rate; ALR: Average Loading Rate; CT: Contact Time; tFz1: Time to impact peak; tFz2: Time to active peak; BW: Body Weight; ICC: Intraclass correlation; LOA: Limit of Agreement; RLOA: Ratio Limit of Agreement.\n\nCompeting interests\n===================\n\nIn this study, an instrumented treadmill was used. The research group had no financial or other interest in the treadmill product or distributor of the treadmill. The project was not dependent on external financial assistance and the authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nSB, SZ, WZ and IB provided advice on the study design. BK recruited the participants, was responsible for the data acquisition/analysis and wrote the article. WZ provided advice on the data analysis. SB, SZ, WZ and IB contributed to the content of the article. All authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2474/13/235/prepub>\n"}
+{"text": "In the article, \"Abnormal glucose regulation in Chinese patients with coronary artery disease: A cross-sectional study\",^\\[[@R1]\\]^ which appeared in Volume 96, Issue 52 of *Medicine*, the authors need to add the following funding information \\\"This study is supported by the Project (No. 2006BAI01A02) supported by the 11th 5-Year National Key Technologies R&D Program for Coronary Heart Disease (CHD) from the Ministry of Science and Technology of China.\"\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nFirst described in 1898 by Dieulafoy, DL was identified in two patients with fatal upper gastrointestinal hemorrhage without ulceration. The incidence of this disease is rare; nonetheless, it can result in serious gastrointestinal bleeding \\[[@B1]\\]. In most patients, it is presented as melena, hematemesis, and hematochezia, and it accounts for 1--5.8% cases of acute upper GI bleeding \\[[@B2]\\]. With the advances in endoscopic techniques, the mortality of DL patients has been reduced from 80% to 8%, and surgical intervention is only considered in cases of failed endotherapy \\[[@B3]\\].\n\nAt the beginning of this century, the first perspective randomized-controlled study demonstrated that hemoclip and band ligation are equally effective compared to injection therapy for DL bleeding \\[[@B4]\\]. The overall hemostasis rate in hemoclip therapy has been shown to be up to 95% \\[[@B5]\\], but the ratio of emergent surgery is still above 5%, according to a recent multicenter report \\[[@B6]\\]. Furthermore, currently, there are no cohort studies on the effectiveness of cyanoacrylate injection for the treatment of duodenal DLs, since most of DLs are located in a proximal stomach. The best technique for endoscopic intervention in DL is still not clear \\[[@B3]\\]. In the present study, we performed a retrospective analysis of cyanoacrylate injection versus hemoclip placement for treating bleeding DLs in the duodenum.\n\n2. Patients and Methods {#sec2}\n=======================\n\n2.1. Study Design {#sec2.1}\n-----------------\n\nThe present study was a retrospective cohort analysis. The study protocol was approved by the Ethics Board of Beijing Ditan Hospital, Capital Medical University. Between October 2008 and February 2016, a total of 18 DL patients from three medical centers were enrolled in the study. All the patients received endoscopy within 12 hours of hospitalization and were diagnosed with duodenal DLs. The last follow-up was 3 months after the endoscopy treatment.\n\n2.2. Patients {#sec2.2}\n-------------\n\nThe diagnosis of duodenal DL was based on recently reported criteria \\[[@B7], [@B8]\\] including (i) active arterial spurting or bleeding from minute defect (\\<3\u2009mm); (ii) a protruding vessel within nearly normal mucosa; and (iii) a fresh, densely adherent clot within the normal-appearing mucosa. The patients\\' age, sex, blood pressure, hemoglobin, prothrombin time (PT), concurrent disease, location, the type of bleeding stigmata, and final outcomes were analyzed. Primary outcomes, including the primary hemostasis rate and rebleeding rate, were compared between the two groups. The secondary outcomes such as the number of endoscopic sessions, need for emergent surgery or transcatheter arterial embolization, bleeding-related deaths, transfusion requirements, hospitalization period, and survival time information were also retrospectively compared between the two groups.\n\n2.3. Endoscopic Therapy {#sec2.3}\n-----------------------\n\nData were collected according to following criteria: (i) endoscopic treatments were performed by gastroenterologists with at least 3 years of endoscopic experience; (ii) patients received endoscopic examination within 12 hours of hospitalization; (iii) duodenal DL diagnosis was confirmed by endoscopy; (iv) after being diagnosed, the patient received either cyanoacrylate injection (CI group) or hemoclip placement (HP group); and (v) 3--10\u2009days after the endoscopic treatment, patients received the first endoscopic follow-up, and after 1--3 months, they received the second endoscopic follow-up. For evaluating the efficacy of endoscopy, primary hemostasis and recurrent bleeding were determined based on previous reports \\[[@B8]\\]. The following apparatus/materials were used for endotherapy: GIF-260J or GIF-260 (Olympus, Japan); Sclerotherapy Needle and Interject\u2122 Injection Therapy Needle Catheter (Boston Scientific, USA); Resolution\u2122 Clip Hemoclip (Boston Scientific, USA); and N-butyl-2-cyanoacrylate 0.5\u2009mL (Braun, German).\n\n2.4. Statistical Analysis {#sec2.4}\n-------------------------\n\nGraphPad Prism 5.01 software was used for the statistical analysis. For comparing the differences in qualitative data between the two groups, chi-squared test or Fisher\\'s exact test was performed. Student\\'s *t*-test was used for comparing the mean differences between the two groups. *P* \\< 0.05 was considered statistically significant.\n\n3. Results {#sec3}\n==========\n\n3.1. Demographic Data {#sec3.1}\n---------------------\n\nA total of 18 duodenal DL patients from three medical centers met the inclusion criteria. Patient\\'s demographic data are shown in [Table 1](#tab1){ref-type=\"table\"}. The major manifestation was melena (61.11%), hematochezia (22.22%), and hematemesis and melena (16.67%). Comorbidities of duodenal DL were observed in 12 patients (66.67%), including hypertension in 2 patients (11.11%), diabetes mellitus in 3 patients (16.67%), hepatic cirrhosis in 4 patients (22.22%), fat liver in 2 patients (11.11%), cholelithiasis in 1 patient (5.56%), arteriosclerosis in 3 patients (16.67%), and hyperlipoidemia in 2 patients (11.11%). The compensated shock was observed in three patients (16.67%). No significant difference was observed between the two groups with regard to age, sex, hemoglobin and PLT levels, and comorbidities ([Table 1](#tab1){ref-type=\"table\"}).\n\n3.2. Endotherapy {#sec3.2}\n----------------\n\nAll endotherapies at three medical centers were performed with standard upper endoscopes by experienced gastroenterologists. Based on the present data, the use of hemoclips or cyanoacrylate injection usually depends on the preference of endoscopists; two patients with active bleeding received cyanoacrylate injection in the CI group ([Table 1](#tab1){ref-type=\"table\"}, [Figure 1](#fig1){ref-type=\"fig\"}). In the HP group, there were 12 patients ([Figure 2](#fig2){ref-type=\"fig\"}), and totally, 16 hemoclips were used; 8 patients had active bleeding, 3 patients had a protruding vessel, and one patient had a clot ([Table 1](#tab1){ref-type=\"table\"}, [Figure 2](#fig2){ref-type=\"fig\"}). In addition to 6 patients from the CI group, 4.5\u2009mL (0.5\u2009mL\u00a0\u00d7\u00a09) was used in total. None of the patients needed surgery or any other additional treatment. There were no marked side effects after our endoscopic treatment, except that one patient in the CI group developed mild ulcer at the injection site and completely recovered after rabeprazole therapy.\n\n3.3. Hemostasis Rate and Rebleeding Rate {#sec3.3}\n----------------------------------------\n\nAll patients received first endoscopic examination and/or endotherapy within 12 hours after being admitted into these medical centers. To observe the effects of CI and HP to duodenal DL, we compared primary hemostasis and recurrent hemorrhage between the two groups. As shown in [Table 2](#tab2){ref-type=\"table\"}, no differences were observed between the CI and the HP groups in primary hemostasis or recurrent hemorrhage, except for 1 patient from the HP group, who had melena reoccur three days after the first endotherapy; thus, cyanoacrylate injection was performed once again. The clinical outcome was summarized in [Table 2](#tab2){ref-type=\"table\"}.\n\n4. Discussion {#sec4}\n=============\n\nBefore the 1980s, a surgical approach was the only definitive treatment for DL patients failed by drugs, and the effectiveness of angiographic embolization was usually disappointing \\[[@B9]\\]. During that time, a surgical operation was considered the only life-saving option \\[[@B10]\\]. In the late 1980s, endoscopic electrocoagulation was first used for controlling hemorrhage in DL patients, and a selective arterial embolization was the alternative treatment used for failed endoscopic therapy \\[[@B11]\\]. After that, other endoscopic therapies have been used in clinical DL treatment as well, including epinephrine injection, heater probe technique, bipolar electrocoagulation, and Nd\u2009:\u2009YAG laser photocoagulation \\[[@B12]\\]. During the late 1990s, clipping and ethanol injection or sclerotherapy was introduced into the treatment of DL bleeding \\[[@B13], [@B14]\\]. However, subsequent retrospective analysis has suggested that endoscopic thermal coagulation should be the first choice of initial treatment for DL hemorrhage \\[[@B15]\\]. Hemoclip \\[[@B16], [@B17]\\] and band ligation \\[[@B18]\\] have also gradually become commonly used therapies for DL treatment \\[[@B19], [@B20]\\].\n\nSince most of DLs are located in the stomach \\[[@B21], [@B22]\\], there are almost no reports on the methods of duodenal DL treatment before the 1990s. In 1990, Goldenberg et al. have reported that a patient with duodenal DL was successfully treated with endoscopic injection of epinephrine and electrocoagulation \\[[@B23]\\]. Several years later, Hokama et al. have combined endoscopic clipping and ethanol injection in treating duodenal DL \\[[@B24]\\]. At the beginning of this century, endoscopic band ligation also became widely used for treating duodenal DL \\[[@B25]\\]. Nevertheless, further controlled studies are needed to evaluate the different methods or combined therapies. Recently, the first clinical controlled study about endotherapy of duodenal DL treatment has been published \\[[@B26]\\], showing that endoscopic band ligation and hemoclip placement are equally effective in duodenal DL treatment.\n\nCyanoacrylate injection has been used to treat hemorrhage of esophageal varices since the late 1970s \\[[@B27]\\]. This technique is a standard method to control varix bleeding in cirrhotic patients worldwide \\[[@B28], [@B29]\\]. However, there has been no clinical evaluation of its effectiveness in treating duodenal DL, except for a recent case report \\[[@B30]\\]. In the present study, we retrospectively analyzed the effectiveness of cyanoacrylate injection versus hemoclip placement in duodenal DL treatment. Based on our observation, there was no difference in effectiveness between the CI and the HP groups. In all of our cases, hemorrhage was completely controlled after cyanoacrylate injection. Moreover, no noticeable side effects were observed in either group. It seems that 0.5\u2009mL cyanoacrylate (usual dosage per patient) was more expensive than one hemoclip.\n\nThe major limitation of the present retrospective study was patient grouping. Only six patients received cyanoacrylate injection, and eight patients with active bleeding received hemoclip placement. It seems that the patient grouping (with or without active bleeding) may lead to a selection bias. As a new method, cyanoacrylate injection was mainly used to control active hemorrhage, such as varix bleeding. More importantly, compared with other endoscopic hemostatic methods in a recent meta-analysis report, cyanoacrylate injection is associated with increased likelihood of hemostasis of active bleeding \\[[@B31]\\]. Moreover, according to a recent study, there is no difference in the hemostasis rate in acute nonvariceal upper gastrointestinal bleeding between cyanoacrylate injection and hemoclip placement \\[[@B32]\\]. In the present study, the data indicated that the patient grouping did not lead to significant clinical bias in the hemostasis rate.\n\nIn conclusion, both CI and HP are effective approaches for treating duodenal DL. CI has a higher hemostasis rate without significant side effects. Further large sample size and prospective randomized trials are necessary to evaluate the efficacy and safety of duodenal DL treatments.\n\nThis work was partly supported by the Science Foundation of Capital Medicine Development (2014-2-2171) and the Capital Foundation for Clinical Characteristic Applied Research projects (Z141107002514132) to Professor Hongshan Wei.\n\nDL:\n\n:   Dieulafoy\\'s lesion.\n\nConflicts of Interest\n=====================\n\nDrs. Yu Jiang, Julong Hu, Ping Li, Wen Jiang, Wenyan Liang, and Hongshan Wei have no conflicts of interest or financial ties to disclose.\n\n![A male (58 years old) duodenal DL patient received cyanoacrylate injection during upper endoscopy. (a) An active bleeding was observed on the descending part of the duodenum. (b) Cyanoacrylate injection. (c) One month after endotherapy.](GRP2018-3208690.001){#fig1}\n\n![A male (28 years old) duodenal DL patient received hemoclip placement during upper endoscopy. (a) A protruding vessel 1 month post endoscopy, without active hemorrhage on the superior part of the duodenum. (b) Hemoclip placement. (c) 10 days after endotherapy.](GRP2018-3208690.002){#fig2}\n\n###### \n\nClinical characteristics of patients in the cyanoacrylate injection and hemoclip placement groups.\n\n                                                 CI (*n* = 6)    HP (*n* = 12)   *P* value\n  ---------------------------------------------- --------------- --------------- -----------\n  Age (years)                                    51.50\u2009\u00b1\u20097.29    44.58\u2009\u00b1\u200912.14   0.221\n  Blood pressure (mmHg)                          109.8\u2009\u00b1\u200932.73   104.3\u2009\u00b1\u200914.37   0.616\n  Hemoglobin level (g/dL)                        86.43\u2009\u00b1\u200930.15   88.30\u2009\u00b1\u200937.21   0.917\n  Platelet                                       177.0\u2009\u00b1\u200952.21   213.8\u2009\u00b1\u2009104.9   0.434\n  Prothrombin time                               13.93\u2009\u00b1\u20092.34    13.25\u2009\u00b1\u20092.45    0.579\n  Concurrent disease                             4               8               1.00\n  Endoscopic characteristic                                                      1.00\n  \u2003Active hemorrhage                             2               8               \n  \u2003Protruding vessel without active hemorrhage   3               3               \n  \u2003Blood clot                                    1               1               \n  Location                                                                       0.344\n  \u2003Duodenum                                                                      \n  \u2003\u2003Duodenal bulb                                3               9               \n  \u2003\u2003Descendant duodenum                          3               3               \n\n###### \n\nClinical outcomes of cyanoacrylate injection and hemoclip placement on Dieulafoy\\'s lesions.\n\n                         CI (*n* = 6)          HP (*n* = 12)         *P* value\n  ---------------------- --------------------- --------------------- -----------\n  Primary hemostasis     6                     11                    1.00\n  Recurrent hemorrhage   0                     1                     \n  Transfusion (mL)       *N* = 2, total 4\u2009IU   *N* = 1, total 1\u2009IU   \n  Hospital stay (days)   9.833\u2009\u00b1\u20091.11          9.333\u2009\u00b1\u20091.45          0.824\n\n[^1]: Academic Editor: Mohit Girotra\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nScurvy is caused by prolonged dietary deficiency of ascorbic acid (Vitamin C). It is a rare compared to other nutritional deficiency in children. High index of suspicion, detailed history, and physical examination is required in diagnosing disease. Children with mental illness or abnormal dietary habits are more prone to develop disease. Although uncommon, it still exists in high-risk groups including economically disadvantaged populations with poor nutrition.\n\nCase Report {#sec1-2}\n===========\n\nA 5-year-old female child coming from low socioeconomic class with global developmental delay was admitted to the pediatric department of Sir T Hospital with the history of increase head size and protrusion of the left eyeball for last 15 days. The child also had excessive irritability, especially when touched on any body part. There was no history of fever, trauma, bleeding from any site, vomiting, or altered sensorium. She was only on liquid and semisolid diet.\n\nOn examination, the child had severe acute malnutrition (weight for height below the third percentile). He was severely pale, afebrile and had no lymphadenopathy or hepatosplenomegaly. Posture of the child was frog-like. The child was extremely irritable. Tenderness was present all over the body mostly over the limbs. He had huge swelling over the entire head which had gradually increased in size. Swelling was soft in consistency with fluctuating baggy feel and tender. There was proptosis of the left eye with exposure keratitis.\n\nVitals: T-N, heart rate-116/min, respiration rate-28/min, and blood pressure-86/68 mmHg.\n\nInitially X rays of the skull bones was done, and there was no fracture in the skull. Ultrasonography of the scalp showed soft tissue outside the skull vault which was thickened (2.3 cm) with vascularity inside between the temporal and frontal regions. Neuroimaging was suggestive of diffuse extensive soft-tissue swelling involving the entire scalp with large necrotic collections and with mild proptosis of the left eye.\n\nDiagnostic aspiration was suggestive of blood.\n\nThe laboratory results were as follows:\n\nHemoglobin: 5.2 g/dl; total count: 6800 cells/mm^3^; and differential count: N 59%, L 37%, E 0, M 3%, and B 0%. Mean corpuscular volume: 90.5 fl, mean corpuscular hemoglobin: 30.4 pg, mean corpuscular hemoglobin concentration: 33.6%, and platelets: 3.2 \u00d7 10^3^/mcL. Bleeding time 1.30 min (normal 2--8), clotting time 3.30 min (up to 11 normal), erythrocyte sedimentation rate: 15 mm/h, prothrombin time 13 s (normal), and activated partial thromboplastin time 27 s (normal). Electrolytes, liver, and renal function test were within normal limits.\n\nThe radiograph of the knee showed ground glass appearance of the shaft of the tibia, fibula, and femur. Irregular, thickened white line at the metaphysic and White line at the cortical ends were seen. All radiological features pointed toward scurvy \\[[Figure 1](#F1){ref-type=\"fig\"}\\].\n\n![X-ray of the knee joint](JPN-11-355-g001){#F1}\n\nMagnetic resonance imaging of the brain revealed diffuse extensive soft-tissue swelling involving entire scalp appearing heterogeneously hyperintense on T2-weighted (T2W) and FLAIR images and hypointense on T1-weighted (T1W) images with no significant diffusion abnormality. The swelling showed large necrotic collections on postcontrast T1W FS images, suggestive of infective etiology with abscess formation. Orbit showed T2W hyperintense collection appearing hypointense on T1W in the left orbit extraconal compartment along the roof, with no obvious postcontrast enhancement or necrotic areas within and resultant mild proptosis of the left globe, less likely to be an abscess formation \\[[Figure 2](#F2){ref-type=\"fig\"}\\].\n\n![Headband sign on magnetic resonance imaging](JPN-11-355-g002){#F2}\n\nAscorbic acid level was not done. The diagnosis of scurvy was made, and the child was treated with 500 mg oral Vitamin C daily for 1 week and then 100 mg of Vitamin C daily thereafter. Incision and drainage of the hematoma was done. Mother was educated about dietary modifications including Vitamin C.\n\nTenderness and pain reduced and disappeared with treatment.\n\nDiscussion {#sec1-3}\n==========\n\nScurvy is less common in the pediatric population, but case reports still appear.\\[[@ref1][@ref2][@ref3]\\] A review of the literature by Noble *et al*. reveals 23 case reports of scurvy in children with behaviorally restricted diets including children with autism, mental retardation, and cerebral palsy.\\[[@ref4]\\] Scurvy is common in children with cerebral palsy as they subsist on predominant milk-based diets (due to pseudobulbar palsy and difficulty swallowing solids), and boiled cow\\'s milk is a very poor source of Vitamin C. Deficiencies may be noted in preterm babies who are on prolonged total parenteral nutrition therapy, children with malnutrition, and those with acute illnesses.\\[[@ref2][@ref5]\\] Musculoskeletal manifestations such as subperiosteal hemorrhages leading to bone pain and symptoms such as limb pain, swelling over long bones, and progressive leg weakness and fractures are present in 80% of patients with scurvy and are prominent in pediatric population.\\[[@ref6]\\] Dermatological manifestations include petechiae, ecchymoses, hyperkeratosis, and perifollicular hemorrhage.\\[[@ref3][@ref7]\\] Oral symptoms include gingival disease characterized by swelling, bleeding gums, and loosening of teeth.\\[[@ref3][@ref6][@ref8]\\] Systemic symptoms of scurvy in children include lassitude and fatigue, failure to gain weight, loss of appetite, and irritability.\\[[@ref6]\\] In addition to these symptoms, deficiency of ascorbic acid may lead to a hypochromic microcytic anemia because of decreased absorption of iron, bleeding, and dietary deficiencies.\\[[@ref3][@ref6]\\] The diagnosis of scurvy is based on history of poor dietary intake of Vitamin C, classic clinical features and radiological findings, and response to treatment with Vitamin C.\\[[@ref3]\\] Weinstein *et al*.\\[[@ref3]\\] recommend oral doses of 100--300 mg of Vitamin C daily until body stores are replenished per serum levels. Daily fruit and vegetable intakes should include a good source of Vitamin C such as citrus fruits. After Vitamin C supplementation, metaphyseal abnormalities will completely resolve. The large shells of periosteal bone are common radiographic findings particularly during the healing phase of disease.\\[[@ref9]\\] Various factors contribute to nutritional deficiencies in nonambulant children with severe spastic cerebral palsy such as poor intake, oral motor dysfunction, feeding problems, and use of antiepileptic drugs.\\[[@ref10]\\]\n\nFinancial support and sponsorship {#sec2-1}\n---------------------------------\n\nNil.\n\nConflicts of interest {#sec2-2}\n---------------------\n\nThere are no conflicts of interest.\n"}
+{"text": "1. Introduction {#sec0005}\n===============\n\nLumbar-peritoneal (LP) and ventriculo-peritoneal (VP) shunt placement is the treatment of choice for diversion of cerebrospinal fluid (CSF) from the subarachnoid space into the peritoneal cavity. In conditions where there is obstruction of CSF absorption or elevated CSF pressure such as hydrocephalus, LP works as an alternative route for removal of excess cerebrospinal fluid. However this invasive procedure has been associated with several complications, most commonly infection and obstruction. Perforation of the bowel is an extremely rare complication occurring in less than 0.1% of cases. Although infrequent, this delayed complication can be fatal if it goes unrecognized. We report a case of a 72 old female patient with LP shunt for raised intracranial pressure, who presented with LP shunt catheter protruding from anus. This was due to bowel perforation in the recto-sigmoid junction by the distal tip of lumbar-peritoneal shunt. She was surgically treated with removal of the distal part of the shunt, external drainage of the proximal part and primary closure of the perforation. The case report has been the work has been reported in line with the SCARE criteria. \\[[@bib0080]\\] Clinical suspicion of abdominal complications by the LP should be raised when patient with hydrocephalus develops acute abdominal symptoms or infection with unusual positive CSF cultures.\n\n2. Case presentation {#sec0010}\n====================\n\nIn our case a 72\u202fyear old female patient with LP shunt was admitted in the Neurosurgical department with symptoms of fever (38,6) vomiting and dysphasia. She had her first surgery for extraparenchymal tumor of frontal lobe 8 months ago, which was managed with bilateral craniotomy and total excision of the tumor. The patient had multiple prior admissions in the Neurosurgical department due to bulging, fever and surgical site infection which was managed with opening the wound, evacuating the pus and cleansing the wound. Eventually lumbar drain had been placed because of persistent bulging, recurrent infection and raised intracranial pressure. The Surgical team on duty was called when the doctor in the Neurosurgical Unit noticed an extrusion of a tube through the anus. The patient had fever (38,6) vomiting and dysphasia. On examination, the patient exhibited signs of incipient meningitis with cervical stiffness, mild abdominal pain and GCS: 13. The CT scan showed ventriculomegaly with the lateral ventricles become dilated and filled with air. Rectal examination revealed the VP shunt catheter protruding for a length of 10\u202fcm from a normal-appearing anus ([Fig. 1](#fig0005){ref-type=\"fig\"}). An emergent exploratory laparotomy was decided by both the neurosurgery and general surgery team.Fig. 1Extrusion of the VP shunt through the anus.Fig. 1\n\nIn surgery, the catheter was identified in the lateral wall of recto-sigmoid junction. The shunt was removed from the abdomen through the anus whereas the central part was cut and exteriorized at the abdominal wall. The site of the perforation was primary closed with two-layer suture.\n\nThe postoperative course was uncomplicated, fever was reduced the first postoperative day, bowel peristalsis was re-established on day 3 and a normal diet was tolerated by the patient on day 4. CSF cultures showed Citrobacter Freundii, therefore the ascending infection of VP shunt was implicated for meningitis. After repeated negative CSF cultures the following weeks, the peritoneal part of the LP catheter was placed intra-abdominally, with no complications postoperatively for the following 12 months.\n\n3. Discussion {#sec0015}\n=============\n\nCerebrospinal fluid (CSF) diversion devices or shunts have been used successfully and have become the primary therapy for obstruction of CSF absorption or elevated CSF pressure. An implanted shunt creates an alternative route for removal of CSF which is constantly produced within the brain by diverting CSF from the ventricles within the brain or the subarachnoid spaces around the brain and spinal cord to another body region (e.g. peritoneal cavity) where it will be absorbed and restoring the physiological balance between CSF production, flow, and absorption when one or more of these functions has been impaired. Lumbar-peritoneal (LP) shunt is a widely used shunt, where the proximal catheter drains the excess CSF from the lumbar thecal sac and distal catheter is typically placed in the peritoneal cavity \\[[@bib0005],[@bib0010]\\].\n\nIn comparison with ventrico-peritoneal shunts (VP), LP shunts give the advantage of avoidance of brain penetration with the shunt catheter, access to a large CSF space in the thecal sac and the potential of a large amount of CSF drainage \\[[@bib0005]\\].\n\nLumbar-peritoneal shunts can also be placed for management of localized pathology that affects the spinal cord. They have been used to drain and prevent reformation of pseudomeningoceles, or unwanted collections of CSF \\[[@bib0005],[@bib0010]\\]. Lumboperitoneal shunts have also been used to help minimize CSF leaks, either natural or iatrogenic. The principle is to decrease the fluid pressure on the dural defect and thus prevent excessive CSF leakage \\[[@bib0015]\\].\n\nLP shunting however has been associated with several complications such as obstruction, migration, malfunction (over drainage, under drainage), subdural hematoma, seizures and infection (ventriculitis, arachnoiditis, meningitis, sepsis) \\[[@bib0020], [@bib0025], [@bib0030]\\]. Furthermore abdominal complications have been reported in the literature including peritoneal pseudocysts, lost distal catheters and intestinal volvulus. The clinical presentation of shunt malfunction can be challenging \\[[@bib0030],[@bib0035]\\]. Symptoms can be either absent or present commonly with difficulty walking or gait disturbances, cognitive challenges and dementia, fever or signs of sepsis and redness along the shunt tract \\[[@bib0035]\\].\n\nSpontaneous bowel perforation is an extremely rare and potentially fatal complication of LP and VP shunt, occurring anytime, few weeks to several years, after the placement of the shunt device in 0.01% to 0.07% of patients \\[[@bib0035],[@bib0040]\\]. The first case in the literature was reported on 1996 by Wilson \\[[@bib0045]\\]. 94 cases of bowel perforation have been reported in the literature where the majority of cases were pediatric patients (0--10 years). In a systematic review of literature composed by A. Hai \\[[@bib0050]\\] reviled that the most common primary diagnosis was congenital hydrocephalus (33 patients), infective cause (10 patients) and 5 cases of tumor, normal pressure hydrocephalus respectively.\n\nThe causes of this uncommon complication vary and have been suspected to be the sharp end of the VP/LP shunt, the formation of a local inflammatory reaction or fibrosis around the distal catheter resulting in pressure on an area of the bowel, chronic irritation by the shunt, previous adhesions or infections and finally silicone allergy as a foreign body-like reaction \\[[@bib0035],[@bib0055]\\].\n\nThe mortality after perforation is relatively high, approaching 15--18%, and it is further increased when infection is present up to 22% with central nervous system (CNS) infection such as meningitis, encephalitis, or brain abscesses \\[[@bib0060]\\] and 33% with intra-abdominal infection \\[[@bib0065],[@bib0050]\\].\n\nTreatment of bowel perforation by VP or LP shunt depends on the clinical presentation. Peritonitis, intraperitoneal abscess and sepsis should always been treated with exploratory laparotomy, lavage, primary closure of the bowel wall and removal of the shunt \\[[@bib0035],[@bib0065], [@bib0070], [@bib0075]\\]. External drainage of the proximal part is mandatory for at least 20\u202fdays together with antibiotic prophylaxis \\[[@bib0035],[@bib0065]\\]. Once the CSF cultures are repeatedly negative, a new peritoneal shunt catheter can be placed on the opposite site.\n\nIn the absent of infection percutaneous or endoscopic removal of the abdominal shunt catheter without surgery has been reported \\[[@bib0035],[@bib0075]\\]. The shunt can be disconnected at the abdominal wall and the distal end can be removed through the rectum by colonoscopy \\[[@bib0075]\\].\n\n4. Conclusion {#sec0020}\n=============\n\nThe rare case of shunt perforation in viscera should be consider in patients with hydrocephalus when Gram-negative meningitis or CSF cultures are positive of unusual bacteria such as Escherichia coli or when the patient develops abdominal symptoms \\[[@bib0035],[@bib0050]\\].\n\nConflicts of interest {#sec0025}\n=====================\n\nNo conflict of interest.\n\nSources of funding {#sec0030}\n==================\n\nNo sources of funding.\n\nEthical approval {#sec0035}\n================\n\nStudy is exempt from ethical approval in our institution.\n\nConsent {#sec0040}\n=======\n\nPatient consent given.\n\nWritten informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request\n\nAuthor contribution {#sec0045}\n===================\n\nStudy design: Makris M., Mitrousias A., Krikelis A., Modestou E., Kokkalas G., KatsogiannosD., Liapi A.\n\nDATA COLLECTIONS: KARATZIAS VASILEIOS\n\nWRITING: KORNAROPOULOS MICHAIL\n\nGuarantor {#sec0050}\n=========\n\nEVRIPIDIS GETTIMIS.\n"}
+{"text": "Introduction {#s1}\n============\n\nThe secondary motor cortex (MOs) is located in the anterior lateral area of the brain cortex in rodents. The MOs is also termed the medial agranular cortex (AGm), medial precentral cortex (PrCm), second frontal area (Fr2), and frontal orienting field (FOF) (Van De Werd et al., [@B41]; Brecht, [@B3]; Sul et al., [@B39]; Barthas and Kwan, [@B2]). Recent studies have provided evidence for potential roles of the MOs in cognitive and executive processes, such as decision making, goal-directed actions, skill learning, and spatial memory (Sul et al., [@B40], [@B39]; Gremel and Costa, [@B12]; Cao et al., [@B4]; Siniscalchi et al., [@B38]; Yamawaki et al., [@B44]). Impairment or inactivation of the MOs results in neglect of the contralateral space in movement and motor order learning impairment (Barthas and Kwan, [@B2]).\n\nTo data, most knowledge of the MOs has been based on anatomical studies of its neural circuit structures and physiological studies of its neuronal activities. In recent decades, researchers have used virus-assisted anterograde or retrograde tracing techniques to characterize the long-range inputs and outputs of neurons in the MOs. Optogenetics and electrophysiology have also been used to investigate the neuronal activities of the MOs in brain functions. By combining viral tracing and electrophysiology recording, Nelson et al. demonstrated that neurons of the MOs make direct excitatory synapses on the auditory cortex (AUD) and exert a suppressive effect on AUD neuronal activity (Nelson et al., [@B30]; Schneider et al., [@B36]; Nelson and Mooney, [@B29]). Using genetic and anatomical manipulations, many researchers have revealed direct pathways from the MOs to the striatum, which integrates inputs from multiple regions to direct motor control (Wall et al., [@B42]; Rothwell et al., [@B35]; Hintiryan et al., [@B18]; Melzer et al., [@B27]). Several studies have also shown that the MOs has direct pathways to multiple brain cortexes, such as the retrosplenial cortex (RSP), primary visual cortex (VISp), and somatosensory cortex (SS), and participates in different brain functions (Manita et al., [@B26]; Yamawaki et al., [@B44]; Zhang et al., [@B45]; Leinweber et al., [@B21]). However, usually only one direct pathway from the MOs to its connected regions has been described in each of these studies, resulting in a lack of knowledge of the overall projections of the MOs and severely hindering the comprehensive understanding of how the MOs is wired to process information at a global scale.\n\nMore recently, several mesoscale connectomes of adult mouse brains, such as the Mouse Brain Architecture project (<http://mouse.brainarchitecture.org/>), Mouse Connectome Project (Zingg et al., [@B46]; <http://mouseconnectome.org/>), and The Allen Mouse Brain Connectivity Atlas (Oh et al., [@B31]; <http://www.brain-map.org/>), have been developed (Mitra, [@B28]). These atlases have revealed connectivity matrixes of multiple brain regions, including the MOs, across the whole brain. However, these atlases show the overall projections of a cluster of neurons in the MOs and are unable to reveal the projections of individual neurons due to dense neuronal labeling. In the past few years, an increasing amount of research has focused on the reconstruction of individual neurons. The tracing of single neurons by Han et al. showed that layer 2/3 neurons of the visual cortex distribute information to multiple areas, rather than to a single area (Han et al., [@B15]). Li et al. reconstructed the overall structure of cholinergic neurons in the basal forebrain and found that individual neurons in the same brain region have many different projections (Li X. et al., [@B23]). In addition, reports of single neurons in other brain regions \\[e.g., barrel cortex and mediodorsal thalamic nucleus (MD)\\] have also shown the diversity of axon projections in the same region (Aransay et al., [@B1]; Economo et al., [@B6]; Guo C. et al., [@B13]; Kuramoto et al., [@B20]). Therefore, overall projections of a cluster of neurons may not represent the projections of individual neurons. The lack of knowledge of projections of the MOs at a single-neuron level severely limits the understanding of its precise wiring diagram.\n\nTo address this issue, we fluorescently labeled \\~80 pyramidal neurons nearby or in the MOs using a Cre-dependent adeno-associated virus (AAV) and acquired an uninterrupted whole-brain 3D dataset at a voxel resolution of 0.2 \u00d7 0.2 \u00d7 1 \u03bcm with a whole-brain fluorescence imaging system (fMOST) (Gong et al., [@B11], [@B10]). Based on the 3D dataset, we successfully reconstructed the complete morphologies of 36 brightly labeled neurons nearby or in the MOs. Furthermore, we analyzed the projection patterns and projection strengths of these neurons at a single-neuron level based on several parameters, including the projection areas, the fiber length, the total number of terminal tips, and the total number of branches. As far as we know, this study is the first to reveal the complete morphologies and projection patterns of single neurons in the MOs. Our results lay a solid foundation for exploring the relationship between neuronal morphologies and behavioral functions of the MOs.\n\nMaterials and methods {#s2}\n=====================\n\nVirus injection\n---------------\n\nP56-P60 male C57BL/6J mice were anesthetized by intraperitoneal injection of a solution containing 10% urethane and 2% chloralic hydras (0.8 ml per 100 g body weight). pAAV-EF1a-double floxed-EYFP was combined at a 20,000:1 ratio with pAAV-CMV-Cre and packaged into an AAV virus. This recombinant adeno-associated virus (final titer: 4.8 \u00d7 10^12^ VG/ml) was used as the anterograde tracer (unpublished). A total of 100 nl of virus was injected into the MOs, which was located \u22121.5 mm lateral to the sagittal suture, 2.96 mm to bregma, and \u22121.3 mm to the dural surface. The stereotaxic coordinates for the injection site were chosen by referring to the Allen Reference Atlas (Dong, [@B5]). The mice survived 21 days before they were sacrificed for brain specimen preparation. All experiments and animal care followed procedures approved by the Institutional Animal Ethics Committee of Huazhong University of Science and Technology or the Administrative Panel on Laboratory Animal Care (APLAC) at Stanford University.\n\nTissue preparation\n------------------\n\nAnesthetized mice were fixed on the operating floor and then intracardially perfused with 50 ml of 0.01 M PBS (Sigma-Aldrich Inc., St. Louis, US), followed by the same volume of 4% paraformaldehyde (PFA) and 2.5% sucrose in 0.01 M PBS. The infusion speeds were strictly controlled to avoid bubbles in the brain, which seriously affect imaging quality. The brains were removed from the skull and immersed in 4% PFA at 4\u00b0C for 24 h. For embedding resin tissue, each intact brain was dehydrated by immersion in a graded ethanol series and then impregnated with HM20 working solution series (Electron Microscopy Sciences, cat. no. 14340). The details of the sample processing procedures were performed as previously described (Gang et al., [@B8]; Guo W. et al., [@B14]).\n\nWhole-brain imaging\n-------------------\n\nThe embedded brain sample was fixed on a 3D translation stage in a water bath filled with 0.01 M Na~2~CO~3~ and propidium iodide (PI) solution to preserve the EGFP fluorescence and counterstain the cell bodies, respectively. The whole-brain imaging system, the imaging parameters of which were set manually, automatically performed the sectioning and imaging process to acquire the whole-brain 3D dataset (Li et al., [@B22]; Gong et al., [@B10]). Data were saved at 16-bit depth in an LZW-compression TIFF format.\n\nWe acquired the dataset sections with a 1 \u03bcm thickness and imaging at a 0.2 \u00d7 0.2 \u00d7 1 \u03bcm voxel size. To save the data acquisition time, we changed the acquisition scheme to a 0.2 \u00d7 0.2 \u00d7 5 \u03bcm voxel size in the regions where there was no GFP single. The entire imaging process took nearly 1 week to complete. After collection, the images were spliced into intact coronal plane images. Thus, the whole dataset included 7,691 coronal images from the 10,630-\u03bcm specimen, while each coronal plane image was 30,442 \u00d7 54,600 pixels in size. Each image was nearly 1 GB, and the total dataset for a single GFP channel was up to 7.45 TB.\n\nVisualization and reconstruction\n--------------------------------\n\nWe used Amira 5.4.0 software (v5.2.2, FEI, Me\\'rignac Cedex, France) to visualize and reconstruct the complete morphologies of labeled pyramidal neurons in the brain. The preprocessed dataset was imported into Amira software on a Dell graphical workstation. To process the massive TB-sized data with the workstation typically found in biological laboratories, we used an efficient platform named TDat that can reduce computer memory consumption and processing time during data access (Li Y. et al., [@B24]). After extracting the data of interest into Amira, we applied the filament editor module of Amira to a brain-wide tracing of long-range axons in 3D view by human-machine interaction. For each neuron, we defined the initial starting point and then continue to import the data block into Amira in the direction of the fiber extension. In each loaded block, we manually assigned the initial point and the end of the fiber, and then the software automatically calculated the path between these two points. This procedure was repeated until all the fibers of the neuron were reconstructed. Tracking results with original location information were stored in AM or SWC format.\n\nBrain region segmentation\n-------------------------\n\nAfter the neuron were traced and reconstructed, we obtained the complete morphologies of 36 IT neurons distributed in the MOs, PL, and ORBm. To define their locations, we used a series of 5\\~10-\u03bcm PI-channel projection images to manually map the boundaries between brain regions by comparing the difference in cell architecture between them. Subsequently, we merged the 100-\u03bcm GFP channel projection images with processed PI-channel projection images. In this way, we obtained the location information of neurons we desired.\n\nIn addition to the brain regions, we also mapped the boundary between layers. As mentioned above, we imported PI-channel data nearing the soma into Amira and used the segmentation editor module to draw the line between layer 1 and layer 2 and the upper boundary of layer 5.The forth layer is known to not exist in the prefrontal area of rodents, and the boundary between layer 2 and layer 3 is difficult to distinguish. Therefore, we assessed layer 2 and layer 3 together. Conversely, the boundary for layer 5 was easily recognized because of the relatively large soma and sparse cell architecture. According to the layer boundaries, we further located the precise positions of these neurons in the brain.\n\nIn addition to the segmentation of brain regions around the projection site, we also used PI-channel images to map the outline of the striatum, the basolateral amygdalar nucleus, the subregions of the cortex, and the whole brain to help locate the projection areas of neurons in the brain. Due to the difficulty of loading the TB-sized data into the random access memory of a commonly used graphical workstation in biological laboratories, we downsampled the raw image data from 0.2 \u00d7 0.2 \u00d7 1 \u03bcm to 4 \u00d7 4 \u00d7 40 \u03bcm or a voxel size of 10 \u00d7 10 \u00d7 30 \u03bcm. After resampling, the images were then loaded into memory.\n\nData analysis\n-------------\n\nIn this study, we calculated the length, the total number of terminal tips, and the total number of branches for every fiber we reconstructed using a web-accessible tool L-Measure (Scorcioni et al., [@B37]). All measurements are listed as the mean \u00b1 s.e.m. Statistical comparisons were performed using Student\\'s *t*-test.\n\nResults {#s3}\n=======\n\nA 3D dataset of a whole mouse brain\n-----------------------------------\n\nWe coadministered a cocktail of two AAV genomes encoding the Cre recombinase and a fluorescent protein whose expression depends on Cre recombinase expression to label neurons in the MOs. The dual-plasmid mixture was packaged into an AAV virus and injected into adult C57BL/6J mice (Figures [1A,B](#F1){ref-type=\"fig\"}). We successfully labeled \\~80 neurons nearby or in the MOs in a whole mouse brain. To acquire a whole-brain imaging dataset, we employed the fluorescence micro-optical sectioning tomography system (fMOST) to image our labeled brain at a 0.2 \u00d7 0.2 \u00d7 1 \u03bcm voxel resolution (Supplementary Figure [1](#SM3){ref-type=\"supplementary-material\"}). Using the fMOST with two imaging channels, we obtained a colocalized dataset of both GFP-positive neurons and cell bodies counterstained with propidium iodide (PI). We processed the dataset into 100-\u03bcm coronal sections and then manually cropped the coronal images around the MOs based on the PI-stained cytoarchitectonic information (Hezel et al., [@B17]) and Allen Reference Atlas to obtain a normative anatomy of region position for labeled neurons (Figure [1C](#F1){ref-type=\"fig\"}).\n\n![A 3D dataset of a whole mouse brain. **(A)** Schematic showing the experimental processes, including virus injection, whole-brain imaging, and data processing. **(B)** The schematic diagram illustrates the dual-plasmid system of the AAV virus. **(C)** Distributions of labeled neurons in the frontal area. 1, 2/3, and 5 depict cortical layers. The dashed lines indicate the boundary of two adjacent regions. The length of the z stack is 100 \u03bcm. Representative raw images of the region near the cell bodies **(C1--C3)** and in the striatum **(C6--C8)** and terminals of axons **(C9,C10)**. Representative image of axonal segments at crossing fibers **(C3)** and branch points **(C8)**. **(C4,C5)** Magnifications of the region indicated in **(C2)**. The arrows in different colors indicate the fibers corresponding to the fibers in **(C3)**. Images are maximum intensity projections through a depth of 100 \u03bcm **(C)**, 200 \u03bcm **(C1,C6)**, 10 \u03bcm **(C1--C3, C7,C8)**, and 5 \u03bcm **(C4,C5)**. **(D)** Representative PI-merged local maximum intensity projections of the coronal planes in a GFP-positive area. AI, Agranular insular area; ccb, corpus callosum, body; fa, corpus callosum, anterior forceps; CP, Caudoputamen; IB, Interbrain; LS, Lateral septal nucleus; BLA, Basolateral amygdalar nucleus; cst, corticospinal tract. Scale bar, 200 \u03bcm **(C)**; 10 \u03bcm **(C1--C10)**; 50 \u03bcm **(D)**.](fnana-12-00086-g0001){#F1}\n\nA majority of GFP-positive neurons were located in the MOs with a few neurons in adjacent areas, such as the prelimbic area (PL) and medial part of the orbital area (ORBm) (Supplementary Figure [2](#SM3){ref-type=\"supplementary-material\"}). We randomly selected 36 brightly labeled pyramidal neurons in one specimen for tracing and reconstructing to avoid the influence of individual differences. The 36 neurons included 6 neurons in layer 5 of the MOs, 17 neurons in layer 2/3 of the MOs, 6 neurons in layer 2/3 of the PL, and 7 neurons in layer 2/3 of the ORBm (Supplementary Video [1](#SM2){ref-type=\"supplementary-material\"}). In our dataset, we used Z stack for reconstruction of nerve fiber morphology. The data shown that fibers in 200-\u03bcm maximum intensity projections of dense regions near the cell bodies (Figure [1C1](#F1){ref-type=\"fig\"}) or in the striatum (Figure [1C6](#F1){ref-type=\"fig\"}) were interwoven, making the different fibers difficult to distinguish. However, fibers in maximum intensity projection images of the same area of \\< 10 \u03bcm were sparse, and crossing fibers (Figures [1C2--5](#F1){ref-type=\"fig\"}) and branch points (Figures [1C7,8](#F1){ref-type=\"fig\"}) were easily distinguished according to their orientation, guaranteeing the accuracy of the tracing process. During reconstruction, we found obvious axonal terminal boutons at the end of fibers, providing clear evidence for the ending of axons (Figures [1 C9,10](#F1){ref-type=\"fig\"}). The axonal morphology here included non-varicose and varicose axonal segments. However, we must acknowledge that, despite our best efforts, there may still be splicing and missing errors in our reconstruction data. As shown in Figure [1D](#F1){ref-type=\"fig\"}, dense GFP fluorescence was distributed in several brain regions, such as the motor cortex (MO), striatum, corpus callosum (cc), basolateral amygdalar nucleus (BLA), and midbrain. We also found rare GFP fluorescence signals in many regions, such as the lateral septal nucleus (LS) and corticospinal tract (cst) (Figure [1D](#F1){ref-type=\"fig\"}). All the above projection areas of the MOs have been revealed in previous mesoscale connectomes (Barthas and Kwan, [@B2]; Peters et al., [@B33]). However, our dataset reveals the complete morphology and projection areas of each individual neuron rather than the overall projections of a cluster of neurons.\n\nIT neurons in layer 5 of the MOs\n--------------------------------\n\nIT projection neurons, which exist in all layers except layer 1, can be divided into associative projection neurons (APNs) and commissural projection neurons (CoPNs) (Fame et al., [@B7]; Lodato and Arlotta, [@B25]). CoPNs, which project to the contralateral hemisphere through the corpus callosum, are also named callosal projection neurons (CPNs).\n\nWe selected 6 brightly labeled neurons in layer 5 of the MOs and manually reconstructed their complete morphologies (neurons 1--6). All 6 of these neurons were CPNs with extensive projections in the contralateral hemisphere (Figures [2A,B](#F2){ref-type=\"fig\"}). The total axonal length of these 6 neurons exceeded 1,600 mm (Table [1](#T1){ref-type=\"table\"}). Moreover, three of them (neurons 2, 5, and 6) had an axonal length over 300 mm, and neuron 6 had the longest axonal length among them at 318.43 mm, which is \\~2.6-fold longer than the longest axonal length in previous reports (\\~120 mm) (Economo et al., [@B6]; Guo C. et al., [@B13]) (Figure [2C](#F2){ref-type=\"fig\"}). Neurons 1--6 not only had different axonal lengths ranging from 180.04 to 318.43 mm (Figure [2C](#F2){ref-type=\"fig\"}, Supplementary Table [1](#SM1){ref-type=\"supplementary-material\"}), but their projection areas were also different (Figure [2B](#F2){ref-type=\"fig\"}). The projections of neuron 1 were concentrated in the striatum and MO areas. Neurons 2, 4, and 6 had other projections to the agranular insular area (AI) or somatosensory areas (SS) of the contralateral hemisphere in addition to the striatum and MO areas. Neurons 3 and 5 also had dense projections to the SS, AI, or visual areas (VIS) of bilateral hemispheres (Figure [2B](#F2){ref-type=\"fig\"}). Remarkably, axons of neuron 5 covered almost the entire SS and AI areas and had more than 2,500 terminal tips throughout the whole brain (Figures [2B,C](#F2){ref-type=\"fig\"}). The terminal tips examined here represent only project targets, excluding en passant boutons. Therefore, our data can only display the projection but not the connections of a single neuron.\n\n![Projections of 6 IT neurons in layer 5 of the MOs. **(A)** Overview of the IT neurons 1--6 in layer 5 of the MOs reconstructed from the whole-brain 3D dataset. Each color represents a single neuron. **(B)** Neurons 1--6 are displayed separately. The dendrites are shown in the top right corner. **(C)** The illustration shows the length, the total number of terminal tips, and the total number of branches of axons of neurons 1--6. **(D)** Projection strengths for each output area of the ipsilateral and contralateral hemispheres. **(E)** Statistical results show projection strengths in the ipsilateral and contralateral target areas for each neuron. **(F)** Projection strengths in the ipsilateral and contralateral striatum for each individual neuron. **(G)** Projection strengths in the ipsilateral and contralateral hemisphere for each individual neuron. Str, striatum; MO, motor cortex; AI, agranular insular area; SS, somatosensory areas; BLA, basolateral amygdalar nucleus; VIS, visceral area.](fnana-12-00086-g0002){#F2}\n\n###### \n\nData statistics of neurons 1--36.\n\n  **Soma location**                 **MOs \\_ Layer 5 (neurons 1--6)**                                                                                                                 **MOs \\_ Layer 2/3 (neurons 7--23)**                                                                                                                                                                                                        **PL \\_ Layer 2/3 (neurons 24--29)**                                                                                                                                                                              **ORBm \\_ Layer 2/3 (neurons 30--36)**\n  --------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Axonal length                     270.58 \u00b1 58.93                                                                                                                                    70.05 \u00b1 9.61                                                                                                                                                                                                                                65.47 \u00b1 12.24                                                                                                                                                                                                     66.94 \u00b1 9.17\n  Axonal branches                   2964.33 \u00b1 1650.80                                                                                                                                 1054.71 \u00b1 322.76                                                                                                                                                                                                                            820.17 \u00b1 170.44                                                                                                                                                                                                   954.71 \u00b1 156.79\n  Terminal tips of axon             1483.00 \u00b1 825.38                                                                                                                                  528.06 \u00b1 167.27                                                                                                                                                                                                                             410.83 \u00b1 85.14                                                                                                                                                                                                    478.14 \u00b1 78.42\n  Dendritic length                  7.40 \u00b1 0.76                                                                                                                                       6.07 \u00b1 0.69                                                                                                                                                                                                                                 5.77 \u00b1 1.09                                                                                                                                                                                                       5.49 \u00b1 0.58\n  Dendritic branches                143.67 \u00b1 22.75                                                                                                                                    147.76 \u00b1 23.66                                                                                                                                                                                                                              122.33 \u00b1 5.46                                                                                                                                                                                                     121.71 \u00b1 10.73\n  Terminal tips of dendrite         72.83 \u00b1 11.37                                                                                                                                     74.88 \u00b1 11.83                                                                                                                                                                                                                               62.17 \u00b1 7.73                                                                                                                                                                                                      61.86 \u00b1 5.37\n  Projection area (ipsilateral)     Somatomotor areas (MO) Agranular insular area (AI) Caudoputamen (CP)                                                                              Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Orbital area,medial part (ORBm) Somatosensory areas (SS) Agranular insular area (AI) Caudoputamen (CP) Basolateral amygdalar nucleus (BLA) Entorhinal area (ENT)   Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Orbital area,medial part (ORBm) Agranular insular area (AI) Caudoputamen (CP) Amygdalar Retrosplenial area (RSP) Entorhinal area (ENT)   Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Orbital area,medial part (ORBm) Caudoputamen (CP) Agranular insular area (AI) Midbrain Visual areas (VIS)\n  Projection area (contralateral)   Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Somatosensory areas (SS) Agranular insular area (AI) Caudoputamen (CP)   Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Orbital area, medial part (ORBm) Somatosensory areas (SS) Caudoputamen (CP) Agranular insular area (AI) Entorhinal area (ENT)                                      Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL)                                                                                                                                          Somatomotor areas (MO) Anterior cingulate area (ACA) Prelimbic area (PL) Orbital area, medial part (ORBm) Agranular insular area (AI) Entorhinal area (ENT)\n\n*All units of length are millimeters*.\n\nWe then analyzed the projection strengths of these 6 neurons (Figure [2D](#F2){ref-type=\"fig\"}). We quantified the projection strength by the axonal length in the target area because the axonal length per area in the cerebral cortex can be directly correlated with synapse numbers (Ohno et al., [@B32]; Rodriguez-Moreno et al., [@B34]). Although all 6 neurons had dense projections in the striatum, their projection strengths in the striatum varied significantly, ranging from 36.18% (neuron 5) to 89.27% (neuron 1) (Figure [2E](#F2){ref-type=\"fig\"}), suggesting that these neurons control the striatum at different intensities. Neuron 1 preferentially projected to the ipsilateral striatum rather than to the contralateral striatum. In contrast, neurons 2, 3, 4, 5, and 6 exhibited more projections in the contralateral striatum than in the ipsilateral striatum (Figure [2F](#F2){ref-type=\"fig\"}). These results suggest that these neurons have different projection preferences for the ipsilateral and contralateral striatum regions. Moreover, the projection strengths of these neurons in the ipsilateral and contralateral hemispheres also differed significantly (Figure [2G](#F2){ref-type=\"fig\"}). For example, 66.4% of the projections of neuron 1 were in contralateral hemisphere while only 33.6% were in the ipsilateral hemisphere (Figure [2G](#F2){ref-type=\"fig\"}). Conversely, neuron 6 devoted 31.4 and 66.8% of the projections to the contralateral and ipsilateral hemispheres, respectively (Figure [2G](#F2){ref-type=\"fig\"}). These results imply that these neurons have different degrees of control over the ipsilateral and contralateral hemispheres in behavior functions.\n\nIT neurons in layer 2/3 of the MOs\n----------------------------------\n\nPrevious studies have shown that all layer 2/3 pyramidal neurons are IT neurons and project to the neocortex, striatum, and corticoid structures, such as the amygdala and claustrum (Harris and Shepherd, [@B16]; Lodato and Arlotta, [@B25]; Gerfen et al., [@B9]).\n\nWe traced and reconstructed the complete morphologies of 17 IT neurons (neurons 7--23) in layer 2/3 of the MOs, including 10 CPNs (neurons 7--16) and 7 APNs (neurons 17--23) (Figure [3A](#F3){ref-type=\"fig\"}). According to the projection areas, these 10 CPNs and 7 APNs were classified into three major categories, respectively. For CPNs, the first category included neurons 7, 8, 11, and 12, whose axons extended caudally to the frontal area of the contralateral hemisphere after transiting the corpus callosum (Figure [3A1](#F3){ref-type=\"fig\"}). The second category (neurons 9, 10, 13, and 14), in contrast, went rostrally to the posterior area of the contralateral hemisphere (Figure [3A2](#F3){ref-type=\"fig\"}). Although the projection areas of the above 8 CPNs (neurons 7-14) were different, they all exhibited dense projections in the contralateral hemisphere. The third category included neurons 15 and 16 in the lateral part of the MOs, which have extremely simple output in opposite hemisphere (Figure [3A3](#F3){ref-type=\"fig\"}). For APNs, the first category includes neurons 17, 18, and 19 with projections in the anterior part of the brain, such as the striatum, prefrontal cortex, AI, and MO (Figure [3A4](#F3){ref-type=\"fig\"}). Neurons in the second category (neurons 20 and 21) only projected to the striatum and amygdala with almost no output to local cortices (Figure [3 A5](#F3){ref-type=\"fig\"}). The neurons in the third category (neurons 22 and 23) extended their axons to the posterior part of the brain (e.g., entorhinal area (ENT)) (Figure [3A6](#F3){ref-type=\"fig\"}). We then calculated the projection strengths of these neurons. The results showed that neurons 7--23 also displayed various projection strengths in their target areas (Figure [3B](#F3){ref-type=\"fig\"}). Even for the neurons belonging to the same category with similar or same projection areas, their projection strengths differed. For example, neurons 20 and 21, which both targeted the striatum and BLA, exhibited different projection strengths to their targets (Figure [3B](#F3){ref-type=\"fig\"}). These results indicate that neurons 7--23 in layer 2/3 of the MOs have different projection areas and projection strengths; therefore, these neurons might be involved in different brain circuits at different intensities.\n\n![Projections of 17 IT neurons in layer 2/3 of the MOs. **(A)** Overview of the IT neurons 7-23 reconstructed from the whole-brain 3D dataset. Each color represents a neuron. **(A1--A3)** Associative projection neurons 7-16 (CPNs) are displayed by category. **(A4--A6)** Callosal projection neurons 17--23 (APNs) are displayed by category. The color of the dendrite corresponds to the neurons in A. The dendrites are shown in the top right corner. **(B)** Projection strengths in the ipsilateral and contralateral target areas for each neuron. Comparisons of axonal length **(C)**, dendritic length **(D)**, total number of axonal branches **(E)**, and total number of dendritic branches **(F)** between CPNs 7--16 and APNs 17--23 in layer2/3. Comparisons of axonal length **(G)**, dendritic length **(H)**, total number of axonal branches **(I)**, and total number of dendritic branches **(J)** between CPNs 7-16 in layer 2/3 and CPNs 1-6 in layer 5. \\* *p* \\< 0.05, \\*\\**p* \\< 0.01, \\*\\*\\**p* \\< 0.001, \\*\\*\\*\\**p* \\< 0.0001, n.s. Represents no significant correlations.](fnana-12-00086-g0003){#F3}\n\nWe also calculated the length and total number of branches of axons and dendrites of CPNs and APNs. For both axons and dendrites, the length and total number of branches of these two groups of neurons were similar (Figures [3C--F](#F3){ref-type=\"fig\"}), implying that these neurons have similar numbers of inputs and outputs (Guo C. et al., [@B13]). In contrast, although neurons 1--16 were all CPNs in the MOs, the average axonal length and total number of axonal branches of neurons 7--16 in layer 2/3 were much less than those in neurons 1--6 in layer 5, at only approximately a quarter and one third, respectively, implying that CPNs in layer 5 might have more outputs than those in layer 2/3 of the MOs (Figures [3G,I](#F3){ref-type=\"fig\"} and Table [1](#T1){ref-type=\"table\"}). The average dendritic length of neurons 1--6 in layer 5 was longer than that of neurons 7--16 in layer 2/3, but the total number of dendritic branches was not significantly different (Figures [3H,J](#F3){ref-type=\"fig\"}).\n\nIT neurons in layer 2/3 of the PL or ORBm\n-----------------------------------------\n\nWe also traced and reconstructed 6 IT neurons (neurons 24--29) in layer 2/3 of the PL and 7 neurons (neurons 30--36) in layer 2/3 of the ORBm **(**Figures [4B,C](#F4){ref-type=\"fig\"} and Supplementary Figure [3](#SM3){ref-type=\"supplementary-material\"}). The PL and ORBm are two areas adjacent to the MOs. The projection areas of each neuron are shown in Figure [4D](#F4){ref-type=\"fig\"}. For the neurons in the PL, neurons 24--28 were APNs, and only neuron 29 was a CPN (Supplementary Figure [3A](#SM3){ref-type=\"supplementary-material\"}). Neurons 24, 27, and 28 projected to the anterior part of the cerebral cortex, including the PL, ORBm, and AI. Neurons 25 and 26 had projections in the lateral-posterior part of the cerebral cortex, such as the ENT or posterior part of the AI. In addition, neurons 24 and 26 had a single fiber projecting to the RSP. Neuron 29 was the only neuron among the 6 neurons that projected to the contralateral frontal area (Supplementary Figure [3A](#SM3){ref-type=\"supplementary-material\"}). Of the neurons in the ORBm (Supplementary Figure [3B](#SM3){ref-type=\"supplementary-material\"}), neurons 30 and 31 were APNs and only had projections to the anterior part of the ipsilateral hemisphere. Moreover, neuron 30 had projections to the AI, but neuron 31 did not. The other 5 neurons (neurons 32--36) were all CPNs and projected to different areas in the ipsilateral and contralateral hemispheres. The results suggest that IT neurons in the ORBm and PL have significantly different projection areas.\n\n![Projections of IT neurons in layer 2/3 of the MOs, PL, and ORBm. Overview of the 17 IT neurons 7-23 in the MOs **(A)**, 6 IT neurons 24-29 in the PL **(B)**, and 7 IT neurons 30-36 in the ORBm **(C)**. The yellow arrow indicates the injection site. **(D)** Projection strengths in the ipsilateral and contralateral target areas for each neuron in the PL and ORBm. **(E--H)** Statistical comparisons of neurons in the MOs, PL, and ORBm. Comparisons of axonal length **(E)**, dendritic length **(F)**, total number of axonal branches **(G)**, and total number of dendritic branches **(H)**. \\**p* \\< 0.05; n.s., Represents no significant correlations.](fnana-12-00086-g0004){#F4}\n\nAlthough neurons 24--36 in the PL and ORBm are in same layer as neurons 7--23 in the MOs, the projection areas of these neurons were different. In the contralateral frontal area, neurons 24--36, regardless of whether they were located in the PL (Figure [4B](#F4){ref-type=\"fig\"}) or ORBm (Figure [4C](#F4){ref-type=\"fig\"}), preferentially projected to the medial part of the frontal cortex. In contrast, axons of neurons 7--23 in the MOs covered almost the entire frontal area and sensory area (Figure [4A](#F4){ref-type=\"fig\"}). Neurons 24--36 also had projections to the mediodorsal part of the brain, such as the RSP or ACA, but neurons 7--23 did not project to these areas (Figures [4A--C](#F4){ref-type=\"fig\"}). These results suggest that the neurons in layer 2/3 of the PL and ORBm innervate the neurons in the medial part of the cerebral cortex, while the neurons in layer 2/3 of the MOs tend to control the brain functions of the lateral part.\n\nWe then calculated the projection strengths of neurons 24--36. These neurons also tended to control the same area with different intensities (Figure [4D](#F4){ref-type=\"fig\"}). We further compared the fiber length and the total number of branches of these neurons with those of the neurons in layer 2/3 of the MOs. The results showed that for axonal and dendritic lengths, there were no significant differences among the neurons in the PL, ORBm, or MOs. However, there were significantly more dendritic branches of the neurons in the MOs than in the PL or ORBm, suggesting that the neurons in the MOs may receive more inputs than those in the PL or ORBm (Figures [4E--H](#F4){ref-type=\"fig\"}).\n\nDiscussion {#s4}\n==========\n\nComprehensive knowledge of the integral morphologies of individual neurons across an entire brain is essential for the understanding of how the nervous system processes information (Aransay et al., [@B1]; Economo et al., [@B6]; Guo W. et al., [@B14]; Kuramoto et al., [@B20]; Li X. et al., [@B23]; Han et al., [@B15]). Here, we labeled \\~80 pyramidal neurons nearby or in the MOs and obtained a 3D whole-brain dataset at a voxel resolution of 0.2 \u00d7 0.2 \u00d7 1 \u03bcm. The imaging resolution of 1 \u03bcm in the Z direction of the MOST system makes the reconstruction of individual neurons more complete and reduces the splicing and missing errors in the reconstruction data. Based on our dataset, we reconstructed uninterrupted complete morphologies of 36 IT neurons, including 6 neurons in layer 5 of the MOs, 17 neurons in layer 2/3 of the MOs, 6 neurons in layer 2/3 of the PL, and 7 neurons in layer 2/3 of the ORBm. The total axonal length of the 6 neurons in layer 5 of the MOs exceeded 1,600 mm and the axonal length of neuron 6 was \\~318.43 mm, which is by far the longest reported axonal length. We also located the projection areas of these neurons based on propidium iodide (PI)-stained cytoarchitecture and the Allen Brain Atlas. Our results showed that the projections of these neurons, regardless of whether the neuron originated in layer 2/3 or layer 5, were concentrated in the cortices and striatum, but almost no neurons had the same projection areas. The results indicated that each of these neurons possess different projection patterns and might be involved in different circuits even though they are located adjacent to one another. As far as we know, all existing mesoscale connectomes of the MOs only show the overall projections of a cluster of neurons. Distinguishing the projection of each neuron in these mesoscale datasets is difficult (Mitra, [@B28]; Oh et al., [@B31]; Zingg et al., [@B46]). Due to the large-scale variation in projections of each neuron, visualizing the complete morphologies of neurons at a single-neuron level rather than obtaining projections of a cluster of neurons is critical for further understanding the wiring diagram of the nervous system. Through the use of single-neuron reconstruction, we also find soma projection patterns that are different from those in previous reports (for example, both neurons 15 and 35 have two fibers passing through the callosum), which cannot be observed in fluorescence images of neuron clusters.\n\nThe motor cortex has been reported to play two parallel roles in rodents: producing dexterous movements and directing certain types of motor learning (Rothwell et al., [@B35]; Peters et al., [@B33]). Neurons of the motor cortex in layer 2/3 are specialized in learning movements, and corticocortical/corticostriatal neurons in layer 5 are involved in both roles. Our results showed that layer 2/3 neurons 7--23 of the MOs had similarly complex axons and dendrites. However, the projection patterns of these neurons were extremely different, suggesting that these neurons might take on different roles in movement learning. Layer 5 neurons 1--6 of the MOs were corticostriatal neurons (CSN) with dense outputs (hundreds of tips per neuron on average) in the striatum. Although these neurons had similar projection areas, their total axonal length (ranging from 180.04 to 318.43 mm) and projection strength (ranging from 36.18 to 89.27%) in the striatum were quite different. In addition, each neuron had a different bias toward ipsilateral or contralateral hemisphere. Neurons 3--6 preferred the ipsilateral hemisphere, whether neurons 1 and 2 tended to control the contralateral hemisphere. These results suggest that even these corticostriatal neurons located adjacent to one another with similar projection areas tend to control the same areas with different strengths. The projections of neurons 1--6 in layer 5 were more complex than those of neurons 7--23 in layer 2/3, suggesting that the neurons in layer 5 of the MOs might receive more inputs and have more outputs than neurons in layer 2/3. Furthermore, all the IT neurons that we reconstructed showed no projections to the spinal cord, which is necessary for generating dexterous movements, implying that these neurons are involved in motor learning rather than dexterous movements (Rothwell et al., [@B35]; Jeong et al., [@B19]; Wang et al., [@B43]).\n\nIn summary, we obtained uninterrupted complete morphologies of 36 IT neurons nearby or in the MOs based on our whole-brain 3D dataset and analyzed the projections of these neurons. To the best of our knowledge, this study is the first to show completed morphologies of individual reconstructed neurons in the MOs. Our results reveal the diversity of the projection patterns for neurons in the same brain region and the complexity of the axonal projections in a single brain region. Our results will be helpful for further understanding the wiring diagram of the MOs at the level of a single neuron and lay a solid foundation for exploring the behavioral functions of the MOs.\n\nAuthor contributions {#s5}\n====================\n\nY-HZ conceived of the project. H-ML performed the majority of experiments and data analysis. PS, NL and XL contributed to the raw data acquisition. J-XK participated in data processing. Y-HZ and H-ML wrote the manuscript with discussion and improvements from all authors.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe thank the Optical Bioimaging Core Facility of WNLO-HUST and the Analytical and Testing Center of HUST for the support in data acquisition.\n\n**Funding.** This work was supported by the National Natural Science Foundation of China (Grant No. 31470056, 31770924), the Fundamental Research Funds for the Central Universities (HUST No. 2018KFYXKJC038) and the Academic Frontier Youth Team Project to Xiaochuan Wang from HUST.\n\nSupplementary material {#s6}\n======================\n\nThe Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fnana.2018.00086/full#supplementary-material>\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nClick here for additional data file.\n\n[^1]: Edited by: Javier DeFelipe, Instituto Cajal (IC), Spain\n\n[^2]: Reviewed by: Francisco Clasca, Universidad Aut\u00f3noma de Madrid, Spain; Akiya Watakabe, RIKEN Center for Brain Science (CBS), Japan\n"}
+{"text": "Introduction {#Sec1}\n============\n\nParkinson's disease (PD) is the second most common neurodegenerative disorder \\[[@CR2]\\] characterized by progressive degeneration of dopaminergic (DA) substantia nigra (SN) neurons and their striatal axon terminals \\[[@CR40], [@CR68]\\]. One characteristic neuropathological hallmark of PD are intracytoplasmic eosinophilic inclusions, the so-called Lewy bodies, which develop in specific brain regions in a spatio-temporal pattern and consist predominantly of misfolded \u03b1-synuclein (aSYN) \\[[@CR11], [@CR83]\\]. The finding that duplications, triplications or missense mutations (e.g. A53T, A30P or G46\u00a0L) of the aSYN gene (SNCA) cause familial forms of PD \\[[@CR46], [@CR67]\\] has justified the assumption that aSYN plays a crucial role in the pathogenesis of PD.\n\nOnly within the last 20\u00a0years it is accepted that PD cannot solely be understood as a disease associated with the degeneration of DA SN neurons, as the PD pathology involves the central, peripheral, autonomic and enteric nervous system \\[[@CR11], [@CR12], [@CR16], [@CR79], [@CR95]\\]. The degeneration of DA SN neurons and the onset of motor dysfunction are preceded by a latency of several years, if not decades, in which the PD pathology develops in brain regions outside the DA SN. This phase, termed prodromal PD, is clinically characterized by the occurrence of certain non-motor symptoms, e.g. hyposmia, constipation, depression and idiopathic REM sleep behavior disorder \\[[@CR31], [@CR53], [@CR84]\\]. Since the prodromal phase is seen as the ideal time window for applying disease-modifying therapy \\[[@CR60], [@CR92]\\], it is of high importance to establish animal models, which allow testing of new and future therapeutic approaches on brain structures that are affected during prodromal PD. The noradrenergic locus coeruleus (LC), a monoaminergic nucleus located in the pontine brainstem \\[[@CR3], [@CR8]\\], plays a crucial role during the prodromal phase of PD and represents therefore an ideal brain structure for such in-depth characterization in an experimental animal model \\[[@CR97]\\]. Dysfunction and degeneration of neurons in the LC region are associated with several of the above listed non-motor symptoms, including depression, signs of reduced arousal, anxiety and REM sleep behavior disorder (RBD) \\[[@CR17], [@CR27], [@CR93]\\]. Neuropathological analysis of human PD brain samples revealed up to 80% LC neuronal cell loss in PD patients, thereby exceeding the degree of SN neurodegeneration in the same individuals \\[[@CR58], [@CR99]\\]. Moreover, experimental evidence indicates that toxin-induced LC cell loss sensitized DA SN neurons for neurodegeneration \\[[@CR9], [@CR22]\\], whereas noradrenergic hyperinnervation resulted in neuroprotective effects \\[[@CR41]\\]. This data implies that LC neurodegeneration itself plays a double role by firstly being responsible for several non-motor symptoms and secondly for accelerating the progression of PD at the nigral level \\[[@CR27]\\]. LC cells exhibit a common at-risk phenotype compared to other neuronal populations such as the DA nigral neurons and the cholinergic neurons of the dorsal motor nucleus of the vagal nerve which undergo neurodegeneration in PD \\[[@CR6], [@CR78]\\]. LC neurons integrate information from a broad range of different brain regions and broadcast information with extensively branched and thinly myelinated axons throughout the complete neuroaxis \\[[@CR8], [@CR75]\\]. Furthermore, they exhibit an intrinsic pacemaking activity, generating action potentials continuously \\[[@CR55]\\] thereby raising their basal metabolic stress level \\[[@CR78]\\].\n\nIn this study, we have characterized the first model of \u03b1-synucleinopathy in the murine LC. We show that targeted viral vector-mediated overexpression of human mutant A53T-aSYN in vivo in LC neurons of wild-type mice resulted in progressive LC neurodegeneration over a time frame of 9\u00a0weeks. Observed LC cell loss was accompanied by prominent and over time increasing micro- and astrogliosis. In addition, our data revealed accumulation of phosphorylated aSYN, progressive aggregation of aSYN as demonstrated by proteinase K-resistant aSYN aggregates and Ubi-1- and p62-positive inclusions comparable with findings from human PD samples. Co-staining with different cellular markers revealed that the p62- and Ubi-1-positive aggregates were found exclusively in microglial cells, while being absent in neurons, astrocytes and oligodendrocytes. Beside this local LC pathology, we observed abundant aSYN-positive axons in a high number of LC output regions, indicating rapid anterograde axonal transport of the human aSYN. In conclusion, our new murine LC model replicated cardinal morphological features of human PD pathology.\n\nMaterials and methods {#Sec2}\n=====================\n\nAnimals {#Sec3}\n-------\n\nA total of 70 wild-type male C57BL/6\u00a0N mice (Charles River, Sulzfeld, Germany), 8\u00a0weeks old at the beginning of the experiment, were used. Mice were housed in individually ventilated cages with ad libitum access to food and water under a 12\u00a0h/12\u00a0h light-dark cycle. All procedures performed in studies involving animals were in accordance with the ethical standards of the institution at which the studies were conducted (Regierungspr\u00e4sidium Giessen, Germany V54--19 c 20 15\u00a0h 01 MR 20/15 Nr. 66/2015).\n\nRecombinant adeno-associated viral (rAAV) vectors and stereotactic injection {#Sec4}\n----------------------------------------------------------------------------\n\nTwo different recombinant adeno-associated viral (rAAV) vectors of a mixed 1/2 serotype were used to overexpress human mutant-A53T-aSYN (rAAV1/2-CMV/CBA-human-A53T-aSYN-WPRE-BGH-pA (rAAV1/2-A53T-aSYN); viral titer 5.1\u2009\u00d7\u200910^12^ gp/ml, purchased from GeneDetect) or luciferase (rAAV1/2-CMV/CBA-luciferase-WPRE-BGH-pA (rAAV1/2-Luc), viral titer 5.0\u2009\u00d7\u200910^12^ gp/ml, purchased from GeneDetect). Each of the two vectors was driven by a chicken beta actin (CBA) promoter combined with a cytomegalovirus (CMV) immediate early enhancer sequence and a woodchuck post-transcriptional regulatory element (WPRE) to assess a high transcription rate \\[[@CR38], [@CR44]\\]. For stereotactic delivery of the rAAV vectors, mice were anesthetized with 100\u00a0mg/kg ketamine and 5\u00a0mg/kg xylazine via intraperitoneal injection. A volume of 1.25\u00a0\u03bcl of rAAV1/2-A53T-aSYN or rAAV1/2-Luc was stereotactically injected in the right LC region using a microinjector (UltraMicro Pump UMP3, World Precision Instruments) with a velocity of 125\u00a0nl/min based on the following coordinates: ML \u2212\u20090.9\u00a0mm, AP -5.4\u00a0mm and DV -3.65\u00a0mm relative to Bregma \\[[@CR66]\\].\n\nTissue preparation {#Sec5}\n------------------\n\nMice were sacrificed through transcardial perfusion with 0.1\u00a0M phosphate-buffered saline (PBS) for 5\u00a0min followed by 4% ice-cold paraformaldehyde (PFA) in 0.1\u00a0M phosphate buffer (PB) (pH\u00a07.4) for 5\u00a0min using a supply pump at a rate of 10\u00a0ml/min. Brains were carefully removed and post-fixed in 4% PFA for 3\u00a0days and then transferred to 30% sucrose solution for 3\u00a0days for cryoprotection. Brains were cut into 20\u00a0\u03bcm thick coronal sections using a cryostat microtome (Leica CM3050 S, Nussloch, Germany). Sections were then stored at 4\u00a0\u00b0C in cryoprotect-solution (1:1:3 volume ratio of ethylenglycol, glycerol and 0.1\u00a0M\u00a0PB) until further processing.\n\nImmunohistochemistry with 3,3-diaminobenzidine (DAB) {#Sec6}\n----------------------------------------------------\n\nFree-floating sections containing the LC/SN region were washed in 0.1\u00a0M\u00a0PB and quenched with 3% H~2~O~2~ and 10% methanol for 15\u00a0min. After a second wash, sections were blocked in 5% normal donkey serum with 0.3% Triton X-100 in 0.1\u00a0M\u00a0PB for 1\u00a0h before incubating them overnight with primary antibodies against TH, p-aSYN, Ubi-1 or p62 (Table\u00a0[1](#Tab1){ref-type=\"table\"}) at 4\u00a0\u00b0C in the same blocking solution. On the second day, sections were washed in 0.1\u00a0M\u00a0PB for 20\u00a0min and then incubated with the appropriate biotinylated secondary antibody (Table\u00a0[1](#Tab1){ref-type=\"table\"}) for 1\u00a0h, followed by incubation in avidin-biotin-peroxidase solution (ABC Elite, Vector Laboratories) for 1\u00a0h before initiating the color reaction with 5% DAB (Serva), diluted in 0.1\u00a0M\u00a0PB with 0.02% H~2~O~2~. All DAB-stained sections were mounted, dried, counterstained with cresyl-violet and coverslipped with mounting gel (Corbit-Balsam, Eukitt). Brightfield images were acquired using an AxioImager M2 microscope (Zeiss) equipped with an Axiocam 506 color camera (Zeiss).Table 1Characteristics of the primary and secondary antibodiesAntigenHostCat. No.ManufacturerDilutionTyrosine HydroxylaseRabbitAB152Merck Millipore1:1000Tyrosine HydroxylaseSheepAB1542Merck Millipore1:1000AAV VP1/VP2/VP3Rabbit61,084Progen1:250Alpha-synuclein (p-S129)Rabbitab51253Abcam1:2000Alpha-synuclein (syn211)MouseAHB0261ThermoFisher1:1000LuciferaseGoatNB100--1677Novus Biologicals1:250GFAPChickenab4674Abcam1:2000IbA1Rabbit019--19,741Wako1:500Ubiquitin (Ubi-1)Mouseab7254Abcam1:2000SQSTM1/p62Mouseab56416Abcam1:2000Olig2Rabbitab109186Abcam1:500MAP2Chickenab5392Abcam1:2000Anti-rabbit AlexaFluor488DonkeyA-21206Invitrogen1:1000Anti-goat AlexaFluor488DonkeyA-11055Invitrogen1:1000Anti-mouse AlexaFluor488DonkeyA-2102Invitrogen1:1000Anti-chicken Cy3Donkey703--165-155Jackson ImmunoResearch1:1000Anti-mouse Cy3Donkey715--165-150Jackson ImmunoResearch1:1000Anti-goat Cy3Donkey705--165-147Jackson ImmunoResearch1:1000Biotinylated anti-rabbitDonkey711--065-152Jackson ImmunoResearch1:1000Biotinylated anti-mouseDonkey715--065-151Jackson ImmunoResearch1:1000Biotinylated anti-goatDonkey705--065-147Jackson ImmunoResearch1:1000Streptavidin AlexaFluor647Donkey016--600-084Jackson ImmunoResearch1:1000\n\nImmunofluorescence staining {#Sec7}\n---------------------------\n\nSections were washed in 0.1\u00a0M\u00a0PB, then blocked in 10% normal donkey serum with 0.3% Triton X-100 in 0.1\u00a0M\u00a0PB for 1\u00a0h before incubating them with primary antibodies (Table\u00a0[1](#Tab1){ref-type=\"table\"}) at 4\u00a0\u00b0C in the same blocking solution overnight. On the second day, sections were washed in 0.1\u00a0M\u00a0PB containing 0.3% Triton X-100 and then incubated with fluorophore-conjugated, species-specific secondary antibodies (Table\u00a0[1](#Tab1){ref-type=\"table\"}) for 2\u00a0h at room temperature in 0.1\u00a0M\u00a0PB containing 0.3% Triton X-100 and 10% normal donkey serum. Before mounting sections were washed for 25\u00a0min in 0.1\u00a0M\u00a0PB containing 0.3% Triton X-100. Exceptions from this general protocol were made for staining luciferase, p-aSYN, IbA1 and Olig2, where after primary antibody incubation a biotinylated species-specific secondary antibody was used to further improve signal to noise by conjugation with streptavidin. Images were acquired using an AxioImager M2 microscope (Zeiss) equipped with an ORCA-Flash4.0 LT CMOS camera (Hamamatsu C11440-42\u00a0U). For confocal images, a TCS SP8 microscope (Leica) was used. Images were processed with FIJI image software \\[[@CR81]\\] to enhance signal-to-noise. Image data for 3D reconstructions were obtained with a Zeiss Spinning Disc Microscope (Axio Observer Z1) equipped with an Axiocam MRm (Zeiss) and an Evolve 512 EMCCD Camera (Photometrics) and post-processed with ZEN 2012 software (Zeiss).\n\nProteinase K treatment {#Sec8}\n----------------------\n\nTo analyze the formation of insoluble aggregates, sections were digested with Proteinase K (PK) using a modified protocol described elsewhere \\[[@CR21], [@CR86]\\]. 20\u00a0\u03bcm thick sections with 120\u00a0\u03bcm interslice distance containing the complete LC region were washed in 0.1\u00a0M\u00a0PB and subsequently digested in 0.1\u00a0M\u00a0PB containing 0.3% Triton X-100 and 12\u00a0\u03bcg/ml PK (Cat. No. 4333793, Invitrogen) at 65\u00a0\u00b0C for 10\u00a0min. To visualize insoluble aggregates, digested sections were double stained against human aSYN, p62, Ubi-1 or luciferase in combination with TH (Table\u00a0[1](#Tab1){ref-type=\"table\"}), following the fluorescence staining protocol described above. Complete absence of TH immunoreactivity served as an indicator for successful PK digestion, thus sections in which TH immunoreactivity was still visible were excluded from analysis implicating an incomplete protein digestion. Images were acquired using an AxioImager M2 microscope (Zeiss) equipped with an ORCA-Flash4.0 LT CMOS camera (Hamamatsu C11440-42\u00a0U).\n\nStereology {#Sec9}\n----------\n\nTo quantify TH-positive LC and SN neurons, the optical fractionator workflow (StereoInvestigator version 9, MicroBrightField Biosciences) was used. Therefore, tissue sections were stained against TH with DAB and counterstained with cresyl-violet as described above. To quantify LC cell numbers, five sections per animal containing the complete rostro-caudal extent of the LC region, separated by 120\u00a0\u03bcm, were selected. Contours including all TH-positive neurons of the LC were drawn, excluding neurons of the SubLC region. For quantification of TH-positive SN neurons, seven sections separated by 240\u00a0\u03bcm covering the complete caudo-rostral extent of the SN were used. Contours were drawn based on the cytoarchitectonic distribution of SN neurons \\[[@CR24]\\] including SN pars compacta but excluding SN pars reticulata or ventral tegmental area neurons. Parameters used for counting were: grid size 100\u2009\u00d7\u2009100\u00a0\u03bcm, counting frame 85\u2009\u00d7\u200985\u00a0\u03bcm, and 2\u00a0\u03bcm guard zones.\n\nQuantification of reactive micro- and astrogliosis {#Sec10}\n--------------------------------------------------\n\nTriple immunofluorescence stainings were performed to visualize astro- and microgliosis using antibodies directed against GFAP for astroglia, IbA1 for microglia and TH to label LC neurons (Table\u00a0[1](#Tab1){ref-type=\"table\"}). To quantify signs of reactive gliosis, we evaluated 5 LC sections of 6 animals per time point by measuring the optical density (OD) of the injected versus the non-injected side using FIJI. First, greyscale images were converted to 8 bit and the LC region was outlined with a rectangular contour (1200px x 800px). Then, OD was measured and lastly a background correction was performed by subtracting the mean background signal for every section. The background corrected OD values of all 5 sections of the injected side were summed and compared to the summed value of the non-injected side.\n\nQuantification of S129-phosphorylated aSYN {#Sec11}\n------------------------------------------\n\nTo analyze the degree of p-aSYN, a triple immunofluorescence staining against p-aSYN, human aSYN and TH was performed (Table\u00a0[1](#Tab1){ref-type=\"table\"}). Five sections of 4 animals per time point, containing the complete rostro-caudal extent of the LC region, were selected for analysis. First, images were converted to 8 bit before making them binary. By using a preset intensity threshold, pixels were given either an intensity value of 255 (when positive for p-aSYN) or 0 (when negative for p-aSYN). The resulting p-aSYN signal intensity value was then divided by the area positive for non-phosphorylated aSYN. This ratio was calculated for all five sections and averaged per animal.\n\nQuantification of aSYN transport {#Sec12}\n--------------------------------\n\nSeven coronal sections (Bregma: +\u20094.28, +\u20092.86, +\u20091.18, +\u20090.38, \u2212\u20090.58, \u2212\u20093.16 and\u2009\u2212\u20097.56) covering the complete mouse brain were stained against human aSYN (Syn 211) or Luc (Table\u00a0[1](#Tab1){ref-type=\"table\"}) and the degree of aSYN accumulation was assessed by scoring human aSYN positive axons/cell bodies as follows: -- no positive axons; + sparse (few positive axons); ++ mild (more positive axons); +++ moderate (many positive axons, covering almost the complete brain region) and ++++ severe pathology (large number of positive axons densely covering the complete brain region). (+) describes an intermediate state. Six animals per time point were analyzed and the scores for each brain region were averaged.\n\nStatistical analyses {#Sec13}\n--------------------\n\nIn general, all data values are expressed as mean\u2009\u00b1\u2009SEM or mean\u2009\u00b1\u2009min/max. Differences were considered significant at *p*\u2009\\<\u20090.05. Multiple comparisons were made by one-way or two-way ANOVA analysis followed by Tukey's or Sidak's multiple comparisons test. To calculate correlations, Pearson's correlation coefficient with 95% confidence interval was used. All statistical analyses were performed using GraphPad Prism version 7.00 (GraphPad Software, La Jolla California USA). Figures were created with Adobe Illustrator version 21.1 (Adobe Systems).\n\nResults {#Sec14}\n=======\n\nrAAV vector-mediated overexpression of human A53T-aSYN in LC neurons {#Sec15}\n--------------------------------------------------------------------\n\nTo determine whether and in which time frame aSYN overexpression induces PD-like pathology in LC neurons we chose to overexpress human mutant A53T-aSYN by injecting rAAV1/2-A53T-aSYN \\[[@CR38], [@CR44]\\] unilaterally in the right LC region of wild-type mice (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}, [b](#Fig1){ref-type=\"fig\"}). To verify that the resulting cellular effects were attributable to the aSYN protein itself, luciferase (Luc) was used as a control protein. To investigate time-dependent effects, animals were consecutively sacrificed after 3\u00a0days, 1, 3, 6 and 9\u00a0weeks (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}).Fig. 1Locally induced protein overexpression via injection of rAAV vectors in the LC region. **a** rAAV1/2 vectors contain a chicken \u03b2-actin promoter hybridized with a CMV immediate early enhancer sequence (CMV/CBA) to drive expression of either A53T-aSYN or luciferase (control). ITR, inverted terminal repeat; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element; BGH-pA, bovine growth hormone polyadenylation sequence. **b** Experimental design and schematic illustration of the injection site. Animals were consecutively sacrificed after 3\u00a0days, 1, 3, 6 and 9\u00a0weeks for immunohistochemical evaluation. **c**-**f** Analysis of the infection or transduction rates via double immunofluorescence staining for TH (red) and viral coating proteins (VP, green) (**c**, **d**) or TH (red) and human A53T-aSYN (green) or luciferase (green) (**e**, **f**), respectively. Co-localization of TH and VP indicates successful entry of viral particles, whereas co-localization of TH and A53T-aSYN/luciferase indicates successful protein expression. Student's t-test revealed no significant difference between the transduction rates of the two vectors (*p*\u2009\\>\u20090.05, *n*\u2009=\u20093 animals per protein) (**d**, **f**). Values (mean\u2009\u00b1\u2009SEM) represent the percentage (%) of TH-positive neurons that were also positive for VP, aSYN or Luc. **g** Overview of the pontine brainstem (Bregma: \u2212\u20095.30\u00a0mm) stained against TH (red) and human aSYN (green) depicting the transduced area 3\u00a0days post-injection. Abbreviations: L, left; R, right; PB, parabrachial nucl.; SUV, superior vestibular nucl.; MV, medial vestibular nucl.; DTN, dorsal tegmental nucl.; LDT, laterodorsal tegmental nucl. **h** Higher magnification overview image of the TH-positive LC region (red) transduced with human A53T-aSYN (green). Scale bars 25\u00a0\u03bcm in **c**, **e**; 500\u00a0\u03bcm in **g** and 100\u00a0\u03bcm in **h**\n\nBy analyzing the first set of animals 3\u00a0days after viral injection, we confirmed that both vectors entered LC neurons equally (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}, [d](#Fig1){ref-type=\"fig\"}), resulting in infection rates of 85.17\u2009\u00b1\u20092.53% for A53T-aSYN and 83.87\u2009\u00b1\u20093.31% for Luc (unpaired t-test, *p*\u2009=\u20090.77) (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). Double immunofluorescence stainings against TH and human aSYN or TH and Luc (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}, [f](#Fig1){ref-type=\"fig\"}) revealed that both vectors induced protein expression already at this early time point with similar strength (A53T-aSYN 59.89\u2009\u00b1\u20092.95% and Luc 54.39\u2009\u00b1\u20093.57%, unpaired t-test, *p*\u2009=\u20090.30). Protein expression was mainly restricted to the LC covering the whole nucleus (Fig.\u00a0[1g](#Fig1){ref-type=\"fig\"}, [h](#Fig1){ref-type=\"fig\"}). In addition, a variable number of immuno-reactive cells were observed in the adjacent regions (ncl. Parabrachialis, Barrington's nucleus, mesencephalic trigeminal nucleus and vestibular nuclei) (Fig.\u00a0[1g](#Fig1){ref-type=\"fig\"}). In LC neurons, cell bodies, as well as axons and dendrites were robustly labeled, indicating strong protein expression. Similar findings were observed for rAAV1/2-Luc injected animals. Notably, there was no aSYN or Luc signal in LC cells on the non-injected side at any time point (Fig.\u00a0[1g](#Fig1){ref-type=\"fig\"}). This allowed us to use the non-injected (left) side as an internal control.\n\nA53T-aSYN overexpression causes LC neurodegeneration {#Sec16}\n----------------------------------------------------\n\nIn the first set of experiments, the extent of aSYN induced LC cell loss was assessed with unbiased stereological quantification of TH-positive LC cells 1, 3, 6 and 9\u00a0weeks after viral vector delivery. In the A53T-aSYN group, significant degeneration of TH-positive LC cells was measured already 3\u00a0weeks post-injection, with 15.86\u2009\u00b1\u20092.09% cell loss compared to control side. Neurodegeneration increased progressively reaching 34.84\u2009\u00b1\u20093.39% after 6\u00a0weeks and 56.25\u2009\u00b1\u20095.19% after 9\u00a0weeks (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}, [b](#Fig2){ref-type=\"fig\"}). Cell loss was homogenously distributed over the complete rostro-caudal extent of the LC. No cellular pathology was observed in the Luc control group at any investigated time point, confirming that neither the viral vector nor overexpression of a cytoplasmic protein was able to induce neurodegeneration in our model (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}, [b](#Fig2){ref-type=\"fig\"}).Fig. 2Progressive loss of TH-immunoreactive LC cells after rAAV1/2-A53T-aSYN injection. **a**, **b** Representative images (Bregma \u2212\u20095.40\u00a0mm) and unbiased stereology of TH-positive LC-neurons in A53T-aSYN (*red bars, right column*) or Luc (*black bars, left column*) overexpressing animals. Values (mean\u2009\u00b1\u2009SEM) are expressed as cell numbers on the injected side compared to non-injected side (%). *n*\u2009=\u20098 per time point and group, two-way ANOVA analysis followed by Tukey's post-hoc test, \\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.01, \\*\\*\\**p*\u2009\\<\u20090.001. **c** Representative confocal images of neuronal morphology after 9\u00a0weeks of protein overexpression. Pyknotic cell bodies and dystrophic axons were observed in A53T-aSYN, but not in Luc overexpressing animals. Scale bars 250\u00a0\u03bcm in **a**, 50\u00a0\u03bcm in **c**\n\nMoreover, immunofluorescent TH-stainings and subsequent confocal imaging revealed that A53T-aSYN, but not Luc overexpression was accompanied by qualitative changes of neuronal morphology, including dystrophic axons and pyknotic perikarya (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}).\n\nAccumulation of phosphorylated-aSYN in the LC region {#Sec17}\n----------------------------------------------------\n\nPhosphorylation of aSYN at amino acid serine 129 (p-aSYN) is a commonly observed phenomenon in human PD brain tissue and in animal models artificially overexpressing aSYN \\[[@CR1], [@CR25], [@CR52], [@CR72], [@CR98]\\]. In these models, the S129 phosphorylation is frequently used as an indicator for aSYN aggregation. In our current study, we measured the signal intensity of p-aSYN systematically via double immunofluorescence stainings for TH and p-aSYN. Our data revealed that A53T-aSYN overexpression led to strong and progressive phosphorylation of aSYN in LC neurons (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}, [b](#Fig3){ref-type=\"fig\"}). Accumulation of p-aSYN started early with positive cells being observable already 1\u00a0week post-injection, reaching highest levels at the latest time point. Generally, the p-aSYN signal was homogenously distributed in the cytoplasm of TH-positive LC cells. In addition, robust labeling of the nucleus was observed (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}, [e](#Fig3){ref-type=\"fig\"}). To exclude the possibility of non-specific antibody labeling we analyzed rAAV-Luc injected animals, which showed no signal for p-aSYN at any time point (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}, [e](#Fig3){ref-type=\"fig\"}). Next, we wanted to quantify if the degree of phosphorylation correlated with the degree of noradrenergic cell loss. Therefore, the p-aSYN signal intensity values were plotted and correlated with the percentage of LC cell loss (Fig.\u00a0[3c](#Fig3){ref-type=\"fig\"}). The strong correlation (*r*\u2009=\u20090.67, *p*\u2009\\<\u20090.05) indicates that the degree of aSYN phosphorylation can be used as a predictor of aSYN toxicity in our model.Fig. 3Progressive accumulation of phosphorylated aSYN (p-aSYN) in the LC region. **a**, **b** Representative images (Bregma \u2212\u20095.40\u00a0mm) and quantification of p-aSYN in the LC region via double immunofluorescence staining for TH (green) and p-aSYN (magenta). Values are presented as mean\u2009\u00b1\u2009SEM, *n*\u2009=\u20094 per time point and group, one-way ANOVA analysis followed by Tukey's post-hoc test, \\**p*\u2009\\<\u20090.05. **c** Robust correlation was observed between loss of TH-positive LC cells and accumulation of p-aSYN (*r*\u2009=\u20090.67, *p*\u2009\\<\u20090.05). **d** Confocal microscopy confirmed accumulation of p-aSYN (magenta) in TH-positive LC cells (green) 3\u00a0weeks post injection in A53T-aSYN (red) overexpressing animals (*lower row*). **e** p-aSYN immunoreactivity in the LC-region of A53T-aSYN overexpressing animals (*lower row*) 9\u00a0weeks post injection. No A53T-aSYN or p-aSYN immunoreactivity was observed in Luc overexpressing animals (*upper row*) at any time point. Scale bars 250\u00a0\u03bcm in **a**, 25\u00a0\u03bcm in **d**-**e**\n\nFormation of proteinase K (PK)-resistant, p62-, Ubi-1- and aSYN-positive aggregates {#Sec18}\n-----------------------------------------------------------------------------------\n\nLewy bodies in human PD brain tissue are characterized by immunoreactivity for insoluble (PK-resistant) aSYN, but also for a variety of other proteins, such as ubiquitin-1 (Ubi-1) and p62/SQSTM1/sequestosome-1 (p62) \\[[@CR33], [@CR47]\\]. Both of the latter proteins are implicated in the cellular clearance of aSYN. Occurrence of PK-resistant Ubi-1-positive aggregates indicates an overburdened proteasomal clearing system, while dysfunction of the lysosomal system can result in accumulation of p62-positive aggregates \\[[@CR73]\\]. To test whether proteasomal and/or lysosomal clearance might be impaired in our model, we systematically screened A53T-aSYN and Luc overexpressing animals for p62- and Ubi-1-immunoreactivity. A53T-aSYN, but not Luc injected mice showed abundant p62- and Ubi-1-positive aggregates starting 3\u00a0weeks after viral vector delivery reaching highest numbers at the latest time point (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}, [b](#Fig4){ref-type=\"fig\"}). Ubi-1-, as well as p62-positive inclusions appeared as small circular objects surrounding the nuclei of the cells (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}) and were restricted to the ipsilateral side of injection. As in the previous experiments most of the p-aSYN signal was seen in TH-positive neurons (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}, [d](#Fig3){ref-type=\"fig\"}), we expected a high rate of co-localization for p62 and Ubi-1 with the LC marker TH. However, the majority of p62 and Ubi-1 immunoreactivity was located next to TH-positive LC cells, suggesting that other cells are involved in this process (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}). To elucidate in which cell type the p62-positive aggregates were located, double immunofluorescence stainings for p62 with MAP2 (neuronal marker), Olig2 (oligodendroglial marker), GFAP (astrocytic marker) or IbA1 (microglial marker) were performed. While p62 did not co-localize with MAP2 (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}), Olig2 (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}) or GFAP (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}), we observed clear co-localization with IbA1 (Fig.\u00a0[5d](#Fig5){ref-type=\"fig\"}), indicating that the p62-positive inclusions were located in microglial cells. Moreover, we further confirmed that Ubi-1-positive aggregates were also located in microglia (Fig.\u00a0[5e](#Fig5){ref-type=\"fig\"}). Double immunofluorescence stainings for IbA1 and aSYN (Syn211) (Fig.\u00a0[5f](#Fig5){ref-type=\"fig\"}, arrow) and GFAP and aSYN (Syn211) (Fig.\u00a0[5g](#Fig5){ref-type=\"fig\"}, arrow) revealed that microglia, as well as astroglia exhibited human aSYN after 3\u00a0weeks of aSYN overexpression.Fig. 4Formation of insoluble protein aggregates in A53T-aSYN overexpressing animals. **a** Staining for p62 (red, *upper row*) and Ubi-1 (red, *lower row*) revealed abundant p62- and Ubi-1-positive aggregates in the LC-region. These aggregates were found in close proximity to, but did not co-localize with TH-positive (green) LC cells, which were positive for p-aSYN (magenta). **b** Representative images of p62- (*upper left*) and Ubi-1 (*lower left*) positive aggregates in A53T-aSYN animals. No aggregates were observed in Luc overexpressing animals at any time point (*right column*). **c** p62-, Ubi-1 and aSYN stainings after proteinase K (PK) digestion or without digestion (\u00d8 PK, *lower right*) confirmed that p62- (*upper left*), Ubi-1- (*upper right*) and human aSYN-positive (*lower left*) aggregates were insoluble. Scale bars 50\u00a0\u03bcm in **a** and **c**, 25\u00a0\u03bcm in **b**Fig. 5p62- and Ubi-1 positive aggregates co-localize with IbA1-positive microglia. **a**-**c** Representative confocal microscopy images of p62 (green, *second column*) and MAP2 (**a**, red, *first row*), Olig2 (**b**, gray, *second row*), or GFAP (**c**, red, *third row*) show no co-localization of p62 and the different cellular markers. In contrast, co-staining p62 (green) and IbA1 (**d**, gray, *fourth row*) or Ubi-1 (green) and IbA1 (**e**, gray, *fifth row*) revealed clear co-localization. IbA1 positive microglial cells (**f**, gray, *sixth row*) and GFAP-positive astroglia (**g**, red, *seventh row*) also co-stained for aSYN (*arrows* in **f**, **g**). Scale bar for **a**-**g** 25\u00a0\u03bcm\n\nNext, we aimed to investigate if the observed p62- and Ubi-1-positive inclusions indeed consisted of insoluble aggregated proteins. Since PK resistance is accepted as a valid marker for the formation of insoluble aggregates in human PD samples and animal models \\[[@CR5], [@CR38], [@CR86]\\], we digested tissue samples of A53T-aSYN and Luc injected mice of all time points with PK. As a result, numerous PK-resistant insoluble aggregates positive for p62, Ubi-1 and aSYN were found in A53T-aSYN injected mice (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}). Notably, PK-resistant aSYN aggregates had the same shape and size as Ubi-1- and p62-inclusions. Further, all three kinds of aggregates started to appear 3\u00a0weeks after initiation of A53T-aSYN overexpression and were restricted to the site of viral injection. PK digestion and subsequent analysis of rAAV-Luc injected animals revealed no signal for aSYN, p62, Ubi-1 or Luc in any analyzed section.\n\nTargeted \u03b1-synucleinopathy induces reactive micro- and astrogliosis in the LC region {#Sec19}\n------------------------------------------------------------------------------------\n\nMicroglia activation and reactive astrocytes have been observed by respective PET imaging in human prodromal and manifest PD patients \\[[@CR26], [@CR63]\\], post-mortem PD brain samples \\[[@CR56], [@CR62]\\] and aSYN animal models \\[[@CR4], [@CR87], [@CR88]\\]. Most of the studies using animal models focused on the impact of microglia activation following nigrostriatal degeneration. In the current study, we aimed to investigate whether a focally induced \u03b1-synucleinopathy in the LC region would lead to reactive micro- and astrogliosis. Therefore, a triple immunofluorescence staining for IbA1 (microglial marker), GFAP (astroglial marker) and TH was carried out and the intensity of fluorescence signal was quantified (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}-[c](#Fig6){ref-type=\"fig\"}). Already 3\u00a0weeks of A53T-aSYN overexpression were sufficient to induce a 3.5-fold increase of astroglial signal intensity in the injected LC region compared to Luc control. The astrogliosis further progressed up to a 6-fold increase after 9\u00a0weeks (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}). Simultaneously, a 3-fold signal increase for microglia was measured after 3\u00a0weeks of A53T-aSYN overexpression and a 5-fold increase after 9\u00a0weeks, compared to Luc (Fig. [6c](#Fig6){ref-type=\"fig\"}). 3D reconstructed high magnification confocal images revealed a dense glial network in A53T-aSYN overexpressing animals, in which the remaining TH-positive LC neurons were embedded already 3\u00a0weeks after viral vector delivery (Fig.\u00a0[6d](#Fig6){ref-type=\"fig\"}). Abundant direct physical contacts between TH-positive LC neurons and astro- and microglia could be resolved. In addition, numerous LC cells appeared to be nearly completely engulfed by microglial processes (Fig.\u00a0[6d](#Fig6){ref-type=\"fig\"}, arrows). In contrast, overexpression of Luc did not lead to any significant increase of astro- or microglia intensity values (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}-[d](#Fig6){ref-type=\"fig\"}). Besides the interaction of astro- and microglia with LC neurons, we also observed direct physical contacts between astrocytes and microglial cells (Fig.\u00a0[6e](#Fig6){ref-type=\"fig\"}, arrow).Fig. 6A53T-aSYN overexpression leads to a pronounced reactive micro- and astrogliosis in the LC-region. **a** Representative images of the LC region of Luc (*left column*) or A53T-aSYN (*right column*) injected animals stained for TH (green), IbA1 (gray) and GFAP (red) display a marked increase of micro- and astroglia over time in A53T-aSYN overexpressing mice. Quantification of GFAP (**b**) and IbA1 (**c**) signal intensity revealed a progressive increase of astro- and microglia signal in A53T-aSYN injected animals (*red boxes*) compared to Luc control (*black boxes*). Values (mean\u2009\u00b1\u2009min/max) are expressed as the signal intensity ratio of the injected side compared to the non-injected side. *n*\u2009=\u20096 animals per time point and group. Two-way ANOVA analysis followed by Tukey's post-hoc test, \\**p*\u2009\\<\u20090.05, \\*\\**p*\u2009\\<\u20090.01, \\*\\*\\**p*\u2009\\<\u20090.001, \\*\\*\\*\\**p*\u2009\\<\u20090.0001. **d** Reconstructed high magnification confocal images of the LC region showing physical contacts between TH-positive (green) LC cells and IbA1-positive micro- (gray) and GFAP-positive astroglia (red) after 3\u00a0weeks of A53T-aSYN overexpression (*lower right*). Engulfment (arrow) of TH-positive neurons by glial cells was only observed in A53T-aSYN expressing animals and not in Luc control mice (*upper row*). **e** Direct physical contacts were also observed between micro- and astroglia (arrow). **f**, **g** Correlating TH cell loss with the microglia intensity values indicates a strong association between increase of microglia and severity of TH cell loss in A53T-aSYN overexpressing animals (*r*\u2009=\u20090.80, *p*\u2009\\<\u20090.05), whereas there was no correlation in Luc expressing animals (*r*\u2009=\u20090.09, *p*\u2009\\>\u20090.05). Pearson's correlation coefficient with 95% confidence interval. Scale bars 100\u00a0\u03bcm in **a**, 25\u00a0\u03bcm in **d** and **e**\n\nTo underline our hypothesis that the degree of aSYN-induced pathology is closely associated with the degree of microgliosis, we correlated the microglial intensity values with the percentage of LC neurodegeneration (Fig.\u00a0[6f](#Fig6){ref-type=\"fig\"}, [g](#Fig6){ref-type=\"fig\"}). This revealed a correlation coefficient of *r*\u2009=\u20090.80 (*p*\u2009\\<\u20090.05) for A53T-aSYN, whereas for the Luc overexpressing animals no significant correlation was found (*r*\u2009=\u20090.09, *p*\u2009\\>\u20090.05).\n\nExtensive transport of human A53T-aSYN to efferent brain regions {#Sec20}\n----------------------------------------------------------------\n\nAfter investigating the local effects of A53T-aSYN overexpression, we addressed the question whether the aSYN pathology can propagate to anatomically connected brain regions. A variety of studies using overexpression of rAAV-aSYN or injection of preformed aSYN fibrils (PFF's) have described transport or spread of aSYN to anatomically connected brain regions \\[[@CR35], [@CR52], [@CR54], [@CR70], [@CR77], [@CR91]\\]. To investigate the propagation of human A53T-aSYN after inducing the \u03b1-synucleinopathy in LC neuronal somata, we stained predetermined brain sections against human aSYN (Syn211) or Luc and rated the occurrence of aSYN- or Luc-positive axons or cell bodies (Table\u00a0[2](#Tab2){ref-type=\"table\"}). While overexpression of Luc resulted in a staining pattern, which was limited to the injection site and absent in distant brain regions, we observed aSYN signal in a high number of brain regions in A53T-aSYN injected mice (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}). One week after injection of rAAV-A53T-aSYN in the right LC region, abundant aSYN-positive axons were observed in various brain regions which are known output regions of LC neurons \\[[@CR85]\\]. The human aSYN signal was solely axonal and no aSYN-positive cell bodies were detected. Regions showing the strongest aSYN signal included the main olfactory bulb, lateral septal nucleus, diagonal band nucleus, bed nuclei of the stria terminalis, central amygdalar nucleus, periaqueductal gray, midbrain reticular nucleus, substantia nigra (SN) pars compacta and the ventral tegmental area (Table\u00a0[2](#Tab2){ref-type=\"table\"}). We counted 36 brain regions, which contained human aSYN-positive axons after one week, indicating that human A53T-aSYN was transported rapidly along the axons towards the synaptic terminals in an anterograde direction. Despite the increase of axonal aSYN signal, no aSYN-positive cell bodies were detected outside of the LC region at any investigated time point, arguing against the hypothesis that human A53T-aSYN is released in LC output regions and taken up by synaptically connected cells in the short time frame of 9\u00a0weeks. This is highlighted by the finding that staining against p-aSYN revealed no signs of phosphorylation or aggregation of endogenous aSYN in distant brain regions after 9\u00a0weeks whereas the axons containing human (non-phosphorylated) A53T-aSYN stained positive for TH (Fig.\u00a0[8c](#Fig8){ref-type=\"fig\"}).Table 2Semiquantitative analysis of human aSYN-pathology in distant brain regionsInjected (right) hemisphere\\\nTime of A53T-aSYN overexpressionNon-injected (left) hemisphere\\\nTime of A53T-aSYN overexpression1 wk3 wks6 wks9 wks1 wk3 wks6 wks9 wksBregmaMain olf. Bulb, gr. layer+++++(+)++(+)++(+)+4.28\u00a0mmInner plexiform layer--+(+)+(+)+--(+)(+)(+)Outer plexiform layer(+)+++--(+)(+)(+)BregmaCortex+++++(+)++(+)++(+)+\u20092.96\u00a0mmMain olf. Bulb+++++(+)++(+)(+)+(+)Ant. olf. Nucleus++(+)+(+)+--(+)(+)(+)Lateral olf. Tract(+)+(+)++--------BregmaCortex+++(+)++(+)++(+)--++++\u20091.18\u00a0mmTaenia tecta, dors. Part++++++(+)--++(+)Corpus callosum(+)+(+)++--(+)(+)(+)Lateral septal nucleus+(+)++(+)+++++(+)(+)++(+)+(+)Diagonal band nucleus+(+)++(+)+++++(+)(+)++(+)+(+)Nucleus accumbens++++--(+)(+)(+)Caudoputamen(+)++(+)--(+)(+)(+)BregmaMedial septal nucleus++(+)+++(+)(+)++(+)++\u20090.38\u00a0mmCortex(+)++(+)++(+)++--+++Bed ncl. of stria term.++++++++(+)+++(+)+++++Magnocellular ncl.(+)+++++--(+)++Hypothalamus+++(+)++(+)+++--+(+)+(+)++Lateral septal nucleus(+)++++(+)++(+)--++(+)+(+)Caudoputamen(+)++(+)--(+)(+)(+)Subst. innominata+++(+)+(+)(+)(+)++BregmaCentral amygdalar ncl.+++++++++(+)(+)+(+)++++\u22120.58\u00a0mmSubst. innominata+(+)++++(+)+++++(+)++++Globus pallidus(+)+++--(+)(+)(+)Hypothalamus+++++(+)++(+)++(+)+(+)Lat. hypothalamic area+(+)+++++++(+)(+)+++++Fimbria(+)+(+)(+)(+)+(+)(+)Stria terminalis++(+)+(+)+(+)--(+)++Corpus callosum--+++----(+)(+)Cortex(+)++(+)++(+)++--+++BregmaCortex--+(+)+(+)+--(+)(+)(+)\u22123.16\u00a0mmSup. colliculus, sens.--+(+)+(+)+--+++Sup. colliculus, motor(+)+(+)++++(+)+(+)++++Periaqueductal gray+(+)++++(+)++(+)(+)+(+)++++Midbrain reticular ncl.+(+)++++++(+)++(+)++++Red nucleus(+)++(+)++++++(+)++Ventral tegmental area++(+)++++(+)+(+)++++(+)++(+)Subst. nigra, pars compacta+(+)+(+)++++(+)++(+)+(+)Subst. nigra, pars reticulata(+)++(+)(+)(+)(+)(+)Hippocampus--++(+)+--(+)++Interpeduncular ncl.(+)+(+)++++(+)(+)+(+)++(+)++(+)Thalamus(+)+(+)++++(+)++(+)+(+)BregmaDorsal motor ncl. of n. X--++(+)+(+)--(+)++\u22127.56\u00a0mmHypoglossal ncl.--+++(+)(+)+(+)Parvicell. reticular ncl.+(+)+(+)+(+)+(+)(+)(+)++Intermediate ret. ncl.+++++(+)--(+)++Medullary reticular ncl.++(+)+(+)+(+)--+++Lateral ret. ncl.--(+)+(+)--(+)(+)(+)Ncl. of the solitary tract--(+)++(+)--(+)++Spinal ncl. of the n. V--(+)(+)(+)----(+)(+)Area postrema(+)(+)++(+)(+)++Arbor vitae--+++--+++Gr. layer (cerebellum)++(+)+(+)+(+)(+)+++Mol. layer (cerebellum)--+++--+++Occurrence of human aSYN-positive axons was graded out of seven coronal brain sections as follows: -- no positive axons; + sparse (few positive axons); ++ mild (more positive axons); +++ moderate (many positive axons, covering almost the complete brain region) and ++++ severe pathology (large number of positive axons densely covering the complete brain region); (+) describes an intermediate state between two categories to allow a more accurate description. *n*\u2009=\u20096 per time point. Abundance of aSYN-positive axons increased over time and was more prominent in the injected (right) hemisphere. The signal for human aSYN was solely axonal and no aSYN-positive cell bodies were detectedFig. 7Widespread transport of human A53T-aSYN to interconnected brain regions. **a**-**d** Representative images of analyzed brain sections stained against human aSYN (Syn211) (Bregma +\u20090.50\u00a0mm **a**, Bregma \u2212\u20090.94\u00a0mm **b**, Bregma \u2212\u20093.16\u00a0mm **c**, Bregma \u2212\u20097.56\u00a0mm **d**). *Scale bars* 1\u00a0mm in **a**-**d**, 25\u00a0\u03bcm in all high magnification images. Abbreviations: CTX, cortex; CP, caudoputamen; LS, lateral septal nucleus; MS, medial septal nucleus; aco, anterior commissure; BST, bed nuclei of stria terminalis; HY, hypothalamus; SI, substantia innominata; OT, olfactory tubercle; HPF, hippocampal formation; DG, dentate gyrus; fi, fimbria hippocampi; int, internal capsule; TH, thalamus; GP, globus pallidus; sAMY, striatum-like amygdalar nuclei; LHA, lateral hypothalamic area; CEA, central amygdalar nucleus; SC, superior colliculus; APN, anterior pretectal nucleus; PAG, periaqueductal gray; MRN, midbrain reticular nucleus; VTA, ventral tegmental area; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; AP, area postrema; NTS, nucleus of the solitary tract; CU, cuneate nucleus; ECU, external cuneate nucleus; DMX, dorsal motor nucleus of the vagus nerve; XII, hypoglossal nucleus; SPV, spinal nucleus of the trigeminal; MDRNd, medullary reticular nucleus, dorsal part; MDRNv, medullary reticular nucleus, ventral part; IRN, intermediate reticular nucleus; IO, inferior olivary complex; py, pyramid; RA, raphe nuclei; mlf, medial longitudinal fascicle; LRN, lateral reticular nucleus; sptV, spinal tract of the trigeminal nerve; icp, inferior cerebellar peduncle; L, left (contralateral); R, right (ipsilateral)Fig. 8No SN cell loss after 9\u00a0weeks of A53T-aSYN overexpression in LC region. **a** Abundant human aSYN-positive (red, *upper right*) axons were observed in the TH-positive SN region (green, *upper left*) after 9\u00a0weeks of A53T-aSYN overexpression in the LC. In contrast, no aSYN-positive cell bodies could be detected. **b** Quantification of TH-positive SN neurons 9\u00a0weeks post injection revealed no significant difference between A53T-aSYN group (*red bars*) and Luc control group (*black bars*) for either side. Values are presented as mean\u2009\u00b1\u2009SEM, *n*\u2009=\u20098 per group and side. One-way ANOVA analysis (*p*\u2009\\>\u20090.05). **c** Representative image of TH and human aSYN axonal co-localization in distant brain regions, exemplified for midbrain reticular nucleus (MRN). The majority of aSYN positive axons co-stained for TH indicating that they origin from the LC. Scale bars 250\u00a0\u03bcm in **a**, 25\u00a0\u03bcm in **c**\n\nNo substantia nigra (SN) cell loss after 9\u00a0weeks of human A53T-aSYN overexpression in LC neurons {#Sec21}\n------------------------------------------------------------------------------------------------\n\nAlready after 1\u00a0week of A53T-aSYN overexpression in LC neurons, human aSYN positive axons passing by DA SN neurons could be detected. After 9\u00a0weeks, SN neurons were densely surrounded by aSYN containing axons (Fig.\u00a0[8a](#Fig8){ref-type=\"fig\"}) but no human aSYN signal was observed in the somata of SN cells. Stereological quantification of TH-positive SN neurons (Fig.\u00a0[8b](#Fig8){ref-type=\"fig\"}) revealed no significant difference of TH-immunoreactive neurons between A53T-aSYN compared to Luc overexpressing mice neither for the left nor for the right SN (One-way ANOVA; *p*\u2009\\>\u20090.05). This result points out that LC degeneration, in combination with profound local axonal aSYN accumulation was not sufficient to induce degeneration of DA SN neurons within the relatively short period of 9\u00a0weeks.\n\nDiscussion {#Sec22}\n==========\n\nDegeneration of the LC noradrenergic system is a key event during PD pathogenesis in the prodromal phase of the disease. In this study, we present the first targeted LC \u03b1-synucleinopathy mouse model, which replicated cardinal features of human PD pathology. We have designed our rAAV vector-based overexpression model to generate robust and rapid induction of aSYN pathology, including phosphorylation and aggregation of aSYN, noradrenergic neurodegeneration, development of dystrophic axon morphology, signs of proteasomal and lysosomal dysfunction and prominent neuron-glial interactions. Furthermore, the herein characterized aSYN transport pattern allows investigating the effects of aSYN-induced LC neurodegeneration on anatomically connected LC output structures.\n\nProgressive S129 phosphorylation and formation of PK-resistant aSYN-positive aggregates {#Sec23}\n---------------------------------------------------------------------------------------\n\nPhosphorylation of aSYN at amino acid S129 is a dominant pathological modification of aSYN \\[[@CR1]\\] since approximately 90% of aSYN in human Lewy bodies is phosphorylated at this position, whereas only 4% of soluble aSYN exhibits this posttranslational modification \\[[@CR25]\\]. In PD animal models, phosphorylation at S129 is used as a key marker to investigate an induced \u03b1-synucleinopathy and its occurrence has often been interpreted as formation of aSYN aggregates \\[[@CR10], [@CR52], [@CR65], [@CR72], [@CR87]\\]. In our current study, we show abundant and over time increasing S129-phosphorylation of aSYN in the cytoplasm and nucleus of LC cells (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}, [d](#Fig3){ref-type=\"fig\"}). Previous studies have pointed out that aSYN has different cellular localizations. Beside the presynaptic and cytoplasmic localization, a nuclear occurrence of aSYN is known \\[[@CR80]\\]. Nuclear p-aSYN has been observed in previous studies where aSYN was overexpressed \\[[@CR30], [@CR96]\\] and it could be shown that nuclear aSYN interacts with histone molecules. It was even suggested that the S129-phosphorylation may play an important role for the nuclear translocation of aSYN \\[[@CR64]\\]. To confirm that phosphorylation of aSYN was accompanied by formation of high molecular weight aSYN aggregates we performed PK digestion experiments that revealed small circular aSYN positive inclusion bodies restricted to the injection site (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}). Our model thereby reproduces a key feature of the LC pathology observed in human PD patients. Importantly, since PK digestion led to the destruction of all soluble proteins it did not allow us to investigate if the developing aSYN-positive inclusions are located in neurons or glial cells. The observed discrepancy between a high amount of p-aSYN-positive cells and a relatively limited number of aSYN-positive PK-resistant inclusions raises the question whether aSYN S129-phosphorylation can solely be used as a sufficient marker for aSYN aggregation. Our data indicate that S129-phosphorylation is an important indicator for aSYN pathology, but immunohistochemistry for other aggregation markers should be added to confirm the occurrence of aSYN aggregates \\[[@CR69], [@CR90]\\].\n\nFormation of p62- and Ubi-1-positive proteinaceous inclusions in microglia {#Sec24}\n--------------------------------------------------------------------------\n\nAdditional markers, which are also commonly accepted to investigate protein aggregation and simultaneously serve as indicators for dysfunction of the proteasomal or lysosomal protein degradation system include Ubi-1 and p62 \\[[@CR33], [@CR47]\\]. Based on previous reports, which showed close co-localization of p-aSYN and Ubi-1 or p62 \\[[@CR54], [@CR74]\\], we expected to find overlap of these markers in our model. But notably, all p62 and Ubi-1 aggregates were located next to p-aSYN positive LC cells (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}). Co-staining p62 and Ubi-1 with different glial and neuronal markers revealed that the p62 and Ubi-1 inclusions were located in IbA1-positive microglial cells (Fig.\u00a0[5d, e](#Fig5){ref-type=\"fig\"}). Further, we also show that microglia exhibited human aSYN, probably as a result of local aSYN uptake or phagocytosis of aSYN containing cellular debris. (Fig.\u00a0[5f](#Fig5){ref-type=\"fig\"}). The possibility that the microglial cells were transduced by rAAV1/2-A53T seems to be unlikely, since the used rAAV1/2 vector possesses a high neuronal tropism \\[[@CR44]\\] and triple stainings for IbA1 (microglia), GFAP (astroglia) and aSYN revealed no aSYN-immunoreactivity 3\u00a0days after viral vector delivery within micro- or astroglial cells. Despite this, we cannot completely exclude the possibility of microglia transduction by the initial injection of rAAV1/2-A53T. Deposition of internalized aSYN aggregates in microglia has already been observed in vitro \\[[@CR49]\\], but our study represents (to our knowledge) the first in vivo evidence of inclusion formation in microglial cells. We hypothesize that these p62- und Ubi-1-positive aggregates develop de novo in microglial cells possibly because of massive human aSYN uptake, which exceeds the lysosomal degradation capacities and leads to protein aggregation.\n\nCurrently we can only speculate why p62 and Ubi-1 reactivity was observed in microglia but absent in LC neurons. Previous studies have shown that p62-positive inclusions co-localize with Ubi-1 not only in neuronal but also in glial cells in neurodegenerative diseases including Alzheimer's disease, dementia with Lewy bodies and PD \\[[@CR47]\\]. Furthermore, it has been shown that microglia rapidly internalize aSYN thereby representing the most efficient scavengers of neuronal released aSYN \\[[@CR48], [@CR71]\\]. By clearing aSYN, microglia might actively delay accumulation of aSYN and maturation of aSYN aggregates in LC neurons. One could hypothesize that in our model aSYN is rapidly released from LC neurons and taken up by microglia, which in turn leads to microglial but not neuronal accumulation of p62 und Ubi-1 aggregates. A longer duration of the experiment may clarify the question, whether Ubi-1- and p62-positive inclusions might also develop in LC neurons.\n\nReactive astro- and microglia and their implication in aSYN-induced LC pathology {#Sec25}\n--------------------------------------------------------------------------------\n\nAnother key aspect in several animal models in which aSYN was injected or overexpressed \\[[@CR87], [@CR88], [@CR100]\\] is the profound involvement of reactive astro- and microgliosis during the development of the aSYN pathology. It has been shown that activated microglial cells, besides their implication in clearing aSYN, are able to trigger the release of inflammatory cytokines and accelerate the production of reactive oxygen species, thereby likely contributing to the process of neurodegeneration \\[[@CR32], [@CR42], [@CR100]\\]. In our model, LC cells were surrounded by a massive network of astro- and microglia already after 3\u00a0weeks of aSYN overexpression, with many microglial cells almost completely engulfing the surviving LC neurons (Fig.\u00a0[6d](#Fig6){ref-type=\"fig\"}, arrows). This early induction of microgliosis (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}) is in line with previous findings where microgliosis even preceded the onset of neurodegeneration \\[[@CR4], [@CR13]\\]. Furthermore, we observed a strong correlation between the increase of microglial signal and LC cell loss (Fig.\u00a0[6g](#Fig6){ref-type=\"fig\"}), implying the conclusion that reactive microglial cells are important modulators of aSYN-induced toxicity not only in the dopaminergic SN but also in the noradrenergic LC. Microgliosis was accompanied by severe and progressively increasing astrogliosis. Reactive astrocytes surrounded and partially engulfed LC neurons. Furthermore, they formed direct physical contacts with reactive microglia (Fig.\u00a0[6e](#Fig6){ref-type=\"fig\"}, arrow) and exhibited clear signal for human aSYN (Fig.\u00a0[5g](#Fig5){ref-type=\"fig\"}). Importantly, reactive astrocytes can also take part in clearing aSYN by endocytosis and degradation in their lysosomal system \\[[@CR50], [@CR76]\\]. Furthermore, they interact closely with microglia and release pro- and anti-inflammatory molecules \\[[@CR18], [@CR57]\\]. Our LC model exemplifies this close interdependency between neurons, micro- and astroglia. We show that glial cells are highly involved in the process of aSYN degradation and that glial dysfunction or failure could be a factor of PD progression. However, it should also be considered that the LC itself plays a central role in decreasing neuroinflammation \\[[@CR20]\\]. Noradrenaline is able to suppress the expression of pro-inflammatory cytokines in glial cells while simultaneously elevating the expression of anti-inflammatory markers \\[[@CR19], [@CR59]\\]. Hence, it is reasonable to assume that loss of LC neurons additionally increases the neuroinflammatory response and contributes to the progressive increase of micro- and astroglial activity seen in our model.\n\nAnterograde axonal transport of aSYN to LC output regions {#Sec26}\n---------------------------------------------------------\n\nTo slow or prevent the progression of PD, it is essential to investigate if and how the \u03b1-synucleinopathy propagates within the brain. Recent evidence \\[[@CR36], [@CR45], [@CR52], [@CR54], [@CR70]\\] suggests that toxic aSYN species formed in a small number of cells can spread trans-synaptically to distant but anatomically connected brain regions where they act as seeds to trigger the formation of insoluble aSYN aggregates \\[[@CR29], [@CR54]\\]. Furthermore, cell culture experiments have demonstrated that aSYN can be taken up by cells and transported in both the retrograde and anterograde direction \\[[@CR15], [@CR23], [@CR94]\\]. The noradrenergic LC has a broad input-output connectome \\[[@CR82], [@CR85]\\] making this brain region suitable to investigate trans-neuronal spread. Moreover, Iba and colleagues \\[[@CR37]\\] have demonstrated in a tauopathy model that injections of synthetic tau fibrils were able to induce tau pathology in LC neurons which then propagated to LC afferents and efferents. To investigate if this also translates into our LC aSYN overexpression model we systematically analyzed and scored the aSYN pathology after 1, 3, 6 and 9\u00a0weeks (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}, Table\u00a0[2](#Tab2){ref-type=\"table\"}). Our results indicate that the overexpressed human A53T-aSYN, once produced in the cytoplasm of LC neurons, is only transported in the anterograde direction towards the synaptic terminals, as abundant aSYN-positive axons and terminals in efferent LC regions co-stained for TH (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}, Fig.\u00a0[8c](#Fig8){ref-type=\"fig\"}, Table\u00a0[2](#Tab2){ref-type=\"table\"}). The broad LC output connectome \\[[@CR85]\\] likely explains this high amount of aSYN-positive axons in distant brain regions of the ipsilateral but also contralateral hemisphere. The mild aSYN pathology of the contralateral (non-injected) side can be explained by LC projections crossing the midline and innervating brain structures of the contralateral hemisphere \\[[@CR39]\\]. In contrast to the profound axonal aSYN immunoreactivity, we found no aSYN-positive cell bodies outside of the LC region, arguing against trans-neuronal spread of aSYN in the relatively short time frame of 9\u00a0weeks. The absence of Luc-positive axons in LC output regions might be explained on one hand by protein size (Luc 62\u00a0kDa vs. aSYN 15\u00a0kDa) and on the other hand by the naturally presynaptic localization of aSYN \\[[@CR7], [@CR89]\\]. We therefore conclude that the aggregation prone aSYN species created by overexpression of human A53T-aSYN in the LC region are not transferred to other neuronal populations within the investigated 9\u00a0weeks. A longer time period and subsequently higher aSYN burden in the LC system may be necessary to enable such a transfer at later time points. This is in line with the finding that despite the severe degree of axonal aSYN accumulation in the SN region after 9\u00a0weeks of A53T-aSYN overexpression in the LC region, no statistically significant SN neurodegeneration was observed (Fig.\u00a0[8](#Fig8){ref-type=\"fig\"}).\n\nOpen questions and limitations {#Sec27}\n------------------------------\n\nIn this study, we have decided to overexpress human mutant A53T-aSYN by injection of a previously well-established rAAV vector. The vector used in this study has proven effective in inducing progressive neurodegeneration of SN neurons by several groups in several PD animal models \\[[@CR28], [@CR34], [@CR38], [@CR43], [@CR44]\\]. In this context, it would be of relevance to investigate whether overexpression of wild-type aSYN in the LC would have led to a different histopathological phenotype. Considering the lower rate of \u03b2-sheet and fibril formation of wild-type aSYN compared to the A53T-aSYN variant \\[[@CR14], [@CR51]\\], one could hypothesize that overexpression of wild-type aSYN might lead to milder histopathological alterations. However, this has to be demonstrated in the LC model in a further study. For this initial study we have focused on a relatively short time frame of up to 9\u00a0weeks which allowed us to characterize the initial, local and time-dependent histopathological alterations of LC neurons caused by A53T-aSYN. Since 9\u00a0weeks is likely too short to observe the full neuropathology, a future study containing longer observation times of up to 52\u00a0weeks or even longer would be suggested. This would allow to further investigate whether trans-synaptic spread of aSYN and subsequent degeneration of dopaminergic SN neurons occur at a later time-point. As our study primarily aims to address the histopathological consequences of A53T-aSYN overexpression in LC neurons, we have not carried out a behavioral assessment. Nevertheless, the model would benefit from a thoroughly carried out behavioral characterization, including sleep recordings covering the possible occurrence of any non-motor or subtle motor symptoms.\n\nConclusions {#Sec28}\n===========\n\nIn a time, in which on one hand the clinical research focus shifts away from the neurodegeneration of the dopaminergic nigrostriatal pathway towards the prodromal stages of PD and on the other hand the first potentially disease modifying therapies enter clinical testing \\[[@CR61]\\], animal models mimicking the prodromal phase of PD are needed. In this study, we have reproduced cardinal histopathological features of the human LC PD-pathology, delineated the time-course of noradrenergic neurodegeneration and characterized robust histological markers, which are sufficient to assess the pathological changes in a quantitative and qualitative way. Taken together, this animal model may contribute to the research on the pathophysiology of the prodromal stage of PD. Further studies with longer observation times and additional characterization (e.g. behavioral assessment, biochemical analyses) are required to determine whether the herein presented model will prove helpful in the development and testing of disease-modifying therapy.\n\naSYN\n\n:   \u03b1-synuclein\n\nDA\n\n:   Dopaminergic\n\nLC\n\n:   Locus coeruleus\n\nLuc\n\n:   Luciferase\n\np-aSYN\n\n:   Phosphorylated \u03b1-synuclein\n\nPB\n\n:   Phosphate buffer\n\nPBS\n\n:   Phosphate buffered saline\n\nPD\n\n:   Parkinson's disease\n\nPFA\n\n:   Paraformaldehyde\n\nPK\n\n:   Proteinase K\n\nrAAV\n\n:   Recombinant adeno-associated viral\n\nRBD\n\n:   REM sleep behavior disorder\n\nSN\n\n:   Substantia nigra\n\nTH\n\n:   Tyrosine hydroxylase\n\nMartin Timo Henrich and Fanni Fruzsina Geibl contributed equally to this work.\n\nLina Anita Matschke and Wolfgang Hermann Oertel are shared senior authors.\n\nWe thank Sabine Anfimov and Christine H\u00f6ft for technical assistance regarding the experiments. We also gratefully acknowledge the Core Facility Cellular Imaging of Philipps University-Marburg for the use of the Leica TCS SP8 confocal microscope and Zeiss Axio Observer Z1 spinning disc microscope. We thank Katrin Roth for the assistance with microscopy. WHO is a Hertie-Senior-Research Professor supported by the Charitable Hertie Foundation, Frankfurt/Main, Germany. LAM and WHC have received a grant from the intramural research fund of the Rh\u00f6n-Klinikum Aktiengesellschaft. BL is a fellow of Deutscher Akademischer Austauschdienst.\n\nFunding {#FPar1}\n=======\n\nWHO is supported by the Charitable Hertie Foundation, Frankfurt/Main, Germany. LAM and WHC have received a grant from the Rh\u00f6n-Klinikum Aktiengesellschaft. BL is a fellow of DAAD.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nAll data generated or analyzed during this study are included in this published article.\n\nWHO, LAM and WHC designed the study. MTH, FFG and BL performed experiments, conducted immunohistochemical analysis and analyzed data. WHO, LAM and ND supervised the project. MTH, FFG, LAM and WHO wrote the manuscript. ND, LT, JBK and JMB critically revised the manuscript. All authors read and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar3}\n==========================================\n\nAll applicable international, national and/or institutional guidelines for the care and use of animals were followed. All animal experiments were approved by the local authorities (Regierungspr\u00e4sidium Giessen, Germany V54--19 c 20 15\u00a0h 01 MR 20/15 Nr. 66/2015).\n\nCompeting interests {#FPar4}\n===================\n\nThe authors declare no competing financial interests. JBK and JMB have equity stakes in, and have received consultancy fees from, Atuka Inc., outside of the submitted work. WHO received personal fees for educational talks and/or consultancy, outside of the submitted work, from Abbvie, Adamas, Bristol-Myer-Squibb, Desitin, Mundipharma, Neuropore, Novartis, Roche and UCB Pharma and grants from the Deutsche Forschungsgemeinschaft, the International Parkinson-Fonds The Netherlands, the Michael J. Fox Foundation, USA, the National Research Fond Luxembourg and from Novartis Pharma, Germany.\n\nPublisher's Note {#FPar5}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nThe last 10--15 years have witnessed an explosion of studies examining whether working memory (WM) performance can be enhanced through various training procedures. Different approaches have been employed in this work, ranging from training of specific components of WM such as updating^[@CR1]--[@CR3]^, shifting^[@CR3],\\ [@CR4]^, and inhibition^[@CR5]^ to composite programs that involve practicing several WM functions^[@CR6],\\ [@CR7]^. Typically, the training runs for 4--5 weeks, with 3 sessions per week, each lasting for 45--60\u2009min. Several recent reviews^[@CR8]--[@CR11]^ give rise to the following conclusions: (a) WM training results in gains on the trained tasks; (b) transfer effects to tasks tapping untrained cognitive domains (e.g., reasoning, intelligence, multitasking) are small or non-existent; and (c) although transfer to untrained WM tasks that share constituent processes with the trained WM tasks may be seen, generalizability to other WM tasks is also quite limited.\n\nSeveral fMRI studies examining neural correlates of training-related WM gains report increased blood-oxygen-level dependent (BOLD) activity in striatum^[@CR1]--[@CR3],\\ [@CR12]^. This is an interesting observation in view of the fact that striatal neurons are thought to serve a gating function in letting new information enter into WM^[@CR13]--[@CR15]^. The functional relevance of increased striatal BOLD activity following WM training was demonstrated by Brehmer and colleagues^[@CR16]^, who reported a strong relationship between the magnitude of training-related increases in striatal BOLD activity and the size of WM improvement post training.\n\nAnimal^[@CR17],\\ [@CR18]^ and human^[@CR19],\\ [@CR20]^ data reveal a link between BOLD activity and measures of dopamine (DA) release. These associations open up for the possibility that increased striatal BOLD activity after WM training is related to an increased release of DA. To infer DA release during task performance, typically quantified using positron emission tomography (PET) and specific radioligands, two conditions that vary in cognitive demands are contrasted. The idea is that binding of the radioligand to DA receptors should be reduced during the more challenging condition relative to the control condition. This is so because, in the former case, binding of the ligand to receptors competes with binding of endogenous DA to the same receptors to a greater extent than in the control condition. Thus, reduced binding is assumed to reflect an increased release of DA. This displacement principle was initially formulated in the context of pharmacological DA challenges^[@CR21]^, and evidence for displacement has been observed in both striatal and extrastriatal regions during verbal^[@CR22]^ and spatial^[@CR23],\\ [@CR24]^ WM tasks.\n\nUsing the same updating training procedure as Dahlin *et al*.^[@CR1]^, B\u00e4ckman *et al*.^[@CR19]^ conducted a PET study that investigated whether 5 weeks of WM training was associated with increased striatal DA release. In that study, the radioligand raclopride was used to measure DAD2 receptor binding. There were two main PET findings: (a) increased DA release was observed during a trained updating task relative to a low-level control task before training, and (b) WM training resulted in a further increase of DA release. The locus of the DA effects was in left caudate, close by the training-related BOLD changes reported by Dahlin *et al*.^[@CR1]^ However, as this is the only study documenting increased DA release following WM training, replication of this finding is warranted. Replication of intriguing data is a rare happening in many fields, including cognitive neuroscience. When replication is attempted, a considerable number of findings reported in journals like *Science* and *Nature* cannot be reproduced^[@CR25],\\ [@CR26]^. In the current work, a major goal was therefore to seek to replicate the B\u00e4ckman *et al*.^[@CR19]^ findings. The training procedure, the criterion task, as well as the PET assessment of DA release were identical to those employed by B\u00e4ckman *et al*.^[@CR19]^.\n\nAs noted, transfer effects from this type of WM training to untrained tasks are small or non-existent. However, we observed transfer from WM updating training to an untrained n-back task that shares a similar demand on updating operations^[@CR1],\\ [@CR19]^. An interesting finding in Dahlin *et al*.^[@CR1]^ was that transfer effects were also observed in terms of neural activation patterns, as measured by BOLD fMRI. Specifically, increased BOLD activity was seen for both the letter-memory and n-back tasks in an overlapping part of the caudate. That said, the magnitude of the increases in the n-back task was considerably smaller compared to the letter-memory criterion task, both behaviorally and neurally. An additional objective of the present research was to investigate whether updating training would result in increased DA release also in an untrained n-back task, thus providing evidence for a neurochemical correlate of transfer of learning.\n\nResults {#Sec2}\n=======\n\nLetter-memory criterion task {#Sec3}\n----------------------------\n\nTwo participants in the updating training group and one participant in the control group were not included in the letter-memory analysis due to technical problems leading to loss of post-training behavioral data. The main analysis of the letter-memory criterion task revealed a significant group x time interaction (*F*(1, 23)\u2009=\u200924.579, *p*\u2009\\<\u20090.001, \u03b7^2^partial\u2009=\u20090.52; *d*\u2009=\u20092.07), reflecting the fact that the training group showed larger performance gains after training than the control group (Fig.\u00a0[1A](#Fig1){ref-type=\"fig\"}). An identical analysis involving only those participants who were included in the PET analyses yielded the same significant interaction (one participant could not be included in this analysis due to a technical error). Thus, the present behavioral findings replicated exactly those from our two previous studies^[@CR1],\\ [@CR19]^.Figure 1Behavioral and dopamine effects for the letter-memory criterion task. (**A**) Percent pre-post changes in letter memory. (**B**) Lower raclopride binding to striatal D2 receptors during letter memory compared to the control task before training in right striatum, reflecting greater DA release in response to the cognitive challenge (x,y,z\u2009=\u200914,\u22123,18; 30,\u22123,\u22122). (**C**,**D**) A training-induced decrease of raclopride binding to D2 receptors was found in left striatum (\u221218, 6, \u221212). (**E**) Training-related changes in raclopride binding to striatal D2 receptors in the peak region observed by B\u00e4ckman *et al*.^[@CR18]^. Error bars are standard errors.\n\nIn all analyses of the PET data, we only report effects that exceed 5 voxels to reduce the risk of spurious findings. We first examined a task effect on D2DR BP comparing the letter-memory and baseline tasks before training across all participants with a paired *t* test. We observed decreased D2DR binding during letter memory in bilateral striatum (x,y,z\u2009=\u200914,\u22123,18; 30,\u22123,\u22122, *p*\u2009\\<\u20090.01; Fig.\u00a0[1B](#Fig1){ref-type=\"fig\"}; x,y,z\u2009=\u2009\u221217,21,0; \u221223,2,11, *p*\u2009\\<\u20090.05). Critically, an ANOVA on the D2DR BP data yielded a significant group x time interaction in bilateral striatum (x,y,z\u2009=\u200926,\u22125,8; \u221218,6,\u221212, *p*\u2009\\<\u20090.05). Figures\u00a0[1C and D](#Fig1){ref-type=\"fig\"} highlight this effect for left striatum. This interaction reflected the fact that the trained persons, but not the controls, showed decreased BP after training. In a peak-voxel analysis of a potential training-related change in the exact same region as originally observed (x,y,z\u2009=\u2009\u221217,\u22127,22;^[@CR18]^), a similar trend of greater reduction of BP in the trained group compared to the control group was observed, although this effect did not reach conventional significance (Fig.\u00a0[1E](#Fig1){ref-type=\"fig\"}; *p*\u2009\\>\u20090.05).\n\nTransfer tasks {#Sec4}\n--------------\n\nFor several transfer tasks (in-scanner digit n-back and the offline tasks spatial n-back, letter-number sequencing, digit symbol, word recall, and number-letter RT), there were main effects of time (*p*s\u2009\\<\u20090.05), reflecting general retest effects. However, for the in-scanner 3-back task (Fig.\u00a0[2A](#Fig2){ref-type=\"fig\"}) as well as for all off-line transfer tasks (Table\u00a0[1](#Tab1){ref-type=\"table\"}), there were no disproportionate time effects as a function of group (*p*\u2009\\>\u20090.05; *d*\u2009=\u20090.00). Thus, no behavioral transfer effects were observed.Figure 2Behavioral and dopamine effects for the n-back transfer task. (**A**) Percent pre-post changes in 3-back. (**B**) Lower raclopride binding to striatal D2 receptors during 3-back compared to 1-back before training in bilateral striatum, reflecting greater DA release in response to the cognitive challenge (x,y,z\u2009=\u200927,0,0; 11,18,6; \u221223,0,8; \u221212,15,11). (**C**) Effects in bilateral striatum (x,y,z\u2009=\u200911,14,\u22129; \u221217,15,14) showing a training-induced decrease of raclopride binding to D2 receptors during 3-back. (**D**) Bar graph showing a selective training-related reduction in right striatum for trained subjects. Error bars are standard errors. Table 1Group means (*SD*) on the pre- and post-training cognitive measures.WM training groupControl groupPre-trainingPost-trainingPre-trainingPost-trainingLetter-memory accuracy %*n*\u2009=\u200912*n*\u2009=\u20091351.54 (11.63)83.33 (31.67)47.57 (12.87)56.28 (17.01)Digit n-back*n*\u2009=\u200910*n*\u2009=\u200912\u00a01-back target accuracy %91.23 (7.78)91.58 (7.71)93.62 (3.05)89.14 (8.42)\u00a03-back target accuracy %74.53 (18.24)86.12 (14.51)79.44 (20.66)91.10 (7.30)Letter-number sequencing*n*\u2009=\u200914*n*\u2009=\u20091413.50 (2.79)14.79 (2.61)14.00 (3.21)14.64 (3.84)Digit span forward*n*\u2009=\u200913*n*\u2009=\u2009147.00 (1.23)7.38 (0.87)6.43 (1.22)6.36 (1.15)Digit span backward*n*\u2009=\u200914*n*\u2009=\u2009146.15 (1.21)6.08 (1.50)6.36 (1.50)6.36 (1.28)Digit symbol*n*\u2009=\u200914*n*\u2009=\u20091471.00 (12.27)77.29 (13.98)68.07 (11.53)75.43 (11.65)Word recall accuracy %*n*\u2009=\u200914*n*\u2009=\u20091475.13 (10.54)80.69 (11.84)80.16 (15.70)82.14 (13.33)Simon task*n*\u2009=\u200914*n*\u2009=\u200913\u00a0Congruent RT413 (45)381 (49)430 (58)400 (42)\u00a0Incongruent RT434 (38)398 (43)460 (48)420 (35)\u00a0Congruent accuracy %99.00 (1.52)98.86 (1.51)98.62 (1.89)98.31 (1.97)\u00a0Incongruent accuracy %96.71 (2.16)97.71 (2.33)96.92 (2.78)97.69 (2.29)Number-letter task*n*\u2009=\u200914*n*\u2009=\u200914\u00a0Mixed task: no-switch RT680 (104)599 (78)767 (116)642 (97)\u00a0Mixed task: switch RT955 (288)816 (234)1115 (310)866 (193)\u00a0Mixed task: no-switch accuracy %98.94 (2.00)99.70 (1.14)98.48 (2.27)98.78 (1.37)\u00a0Mixed task: switch Accuracy %97.99 (3.15)97.99 (3.15)96.21 (3.51)96.21 (3.50)Spatial n-back*n*\u2009=\u200914*n*\u2009=\u200913\u00a01-back accuracy in %96.80 (2.81)98.31 (2.09)96.96 (2.91)98.08 (1.67)\u00a02-back accuracy in %94.84 (5.15)97.82 (2.97)93.27 (5.67)98.40 (2.53)\n\nHowever, as with letter-memory, comparing the 3-back and 1-back tasks before training across all participants with a paired *t* test, we observed decreased D2DR binding during 3-back in bilateral striatum (Fig.\u00a0[2B](#Fig2){ref-type=\"fig\"}; x,y,z\u2009=\u200927,0,0; 11,18,6; \u221223,0,8; \u221212,15,11, *p*\u2009\\<\u20090.01). Importantly, a group x time ANOVA on the D2DR BP n-back data yielded a significant interaction in bilateral striatum (Fig.\u00a0[2C](#Fig2){ref-type=\"fig\"}; x,y,z\u2009=\u200911,14, \u22129; 9,14,8; \u221217,15,14, *p*\u2009\\<\u20090.05). The interaction in right striatum reflected the fact that the training group showed decreased BP after training, whereas the control group's level was stable (Fig.\u00a0[2D](#Fig2){ref-type=\"fig\"}).\n\nDiscussion {#Sec5}\n==========\n\nThe main goal of this study was to replicate the B\u00e4ckman *et al*.^[@CR19]^ finding of increased DA release following WM updating training. First, replicating and extending prior work^[@CR19],\\ [@CR22]--[@CR24]^, DAD2 BP was lower during both letter memory and n-back than in the control conditions before training. This pattern indicates increased DA release as a function of the cognitive challenge. A key observation was reduced striatal DAD2 BP during letter memory in the training group at post test, suggesting increased DA release after training. The intervention effect on neurotransmission was accompanied by pronounced training-related behavioral gains in the letter-memory criterion task^[@CR1],\\ [@CR19]^. Although the DA effect did not overlap with the peak effect obtained in our previous study, both were located in left striatum. Also, a similar trend was seen in the same peak voxel as originally reported^[@CR19]^.\n\nA unique feature of this study was the use of four PET examinations, allowing us to assess neurochemical effects during an untrained n-back task. Interestingly, our training was associated with increased striatal DA release at the post-training assessment also during the 3-back transfer task. To the best of our knowledge, this is the first demonstration of changes in neurotransmission in relation to transfer of learning. For n-back the peak response was in right striatum, whereas the corresponding effect for letter memory was in left striatum. The reason for the difference in laterality remains unclear, and was not expected from our previous observation of a training-related overlap in left caudate BOLD signal change for letter memory and n-back^[@CR1]^.\n\nIn agreement with several recent meta-analyses and qualitative reviews^[@CR8]--[@CR11]^, we found no evidence that updating training transferred to performance on the tasks included in the off-line battery tapping verbal and spatial working memory, motor speed, episodic memory, set shifting, and inhibitory control. Moreover, unlike the findings reported by Dahlin *et al*.^[@CR1]^ and B\u00e4ckman *et al*.^[@CR19]^, we did not observe behavioral generalization to the 3-back transfer task. The reason thereof may be that the present participants performed at a very high level for 3-back already at the pre-training assessment. The high level may reflect the length (more room for task practice) and the nature of the task (performed during scanning, which likely increases the perceived performance demands compared to an off-line testing situation). Note also that recent meta-analyses indicate that effect sizes for this form of near-transfer are quite small^[@CR9]--[@CR11]^. Thus, it is unsurprising that such transfer effects may or may not be observed in a specific study. This is especially true in PET research with relatively small sample sizes and limited statistical power. In view of this concern and even though the measurement scale might not have been sensitive enough to capture behavioral transfer effects, the PET data nevertheless demonstrated DA alterations during the untrained n-back task. The present pattern of results raises the possibility that increased DA release may be a necessary, but not sufficient, condition for behavioral transfer to occur.\n\nThe present research extends previous observations that the DA system is plastic. Such plasticity has been observed during both pharmacological and cognitive challenges^[@CR18],\\ [@CR20],\\ [@CR21],\\ [@CR27]--[@CR31]^. The current replication of increased striatal DA release during letter memory after WM updating training, and the novel result of a corresponding increase during an untrained 3-back task, provide additional evidence that the DA system is malleable.\n\nMethods {#Sec6}\n=======\n\nParticipants {#Sec7}\n------------\n\nThe effective sample included 28 right-handed healthy, non-smoking, and non-medicated Finnish male university students (19--26 years). The participants underwent structural MRI and medical screening. They were randomized into training and control groups (*n*\u2009=\u200914 for each). The two groups were comparable regarding years of education, age, and BDI-II scores, and on all pre-training neuropsychological measures, except for TMT B, *t*(25)\u2009=\u20092.35, *p*\u2009\\<\u20090.05, where the control group showed significantly worse performance (Table\u00a0[2](#Tab2){ref-type=\"table\"}).Table 2Descriptive data on the study groups (means and *SD*).WM training group (*n*\u2009=\u200914)Control group (*n*\u2009=\u200914/13)Age22.21 (1.72)22.79 (1.48)Years of education14.21 (1.19)15.15 (1.45)BDI-II2.64 (3.13)1.79 (1.85)WAIS-III Vocabulary52.50 (6.97)49.86 (7.82)TMT A21.64 (5.23)22.93 (4.57)TMT B42.64 (6.79)50.69 (10.74)Pattern Comparison19.61 (3.70)20.04 (3.20)Number Copying51.46 (5.76)51.04 (6.25)Note. BDI\u2009=\u2009Beck Depression Inventory, WAIS\u2009=\u2009Wechsler Adult Intelligence Scale, TMT\u2009=\u2009Trail Making Test. Age in years, BDI-II scores, WAIS-III Vocabulary raw scores, TMT completion time in sec, Pattern Comparison and Number Copying raw scores (*SD*s in parentheses). One participant in the control group had missing data on years of education, one participant in the control group had unreliable performance data on the Pattern- Comparison task, and one participant in the control group was an extreme outlier in TMT B (hence, *n\u2009=\u2009*13 in the control group for these variables).\n\nThe study was approved by the Ethics Review Board of the Turku University Hospital District, the methods were carried out in accordance with relevant guidelines and regulations, and written informed consent was obtained from all participants.\n\nProcedure {#Sec8}\n---------\n\nWe employed a pre-training--intervention--post-training control group design. All participants took part in the pre- and post-training assessments, whereas only participants in the training group received training between these assessments. The training group practiced three times per week (45\u2009min/session) for five weeks. The pre-training assessments included a structural MRI scan, neuropsychological testing, and two consecutive PET scans performed during the same day. The post-training assessments included neuropsychological testing and two PET scans, but no MRI. Time in-between the pre- and post-training assessments was 6--9 weeks.\n\nThe training program was a computerized in-house developed Visual Basic program and consisted of the letter memory criterion task and five other updating tasks^[@CR1],\\ [@CR19]^. Four of the additional training tasks were similar to the letter memory task and involved updating of single items, but to foster generality different kinds of stimuli (i.e. numbers, letters, colors, and spatial locations) were used. In these training tasks, five lists of items were randomly presented and the task was to recall the four last presented items. Across the 5-week training period, list length varied to manipulate level of difficulty and thereby ensure that the training was sufficiently demanding (low level\u2009=\u20094--7 items; medium level\u2009=\u20096--11 items; high level\u2009=\u20095--15 items). Performance was monitored and at the end of each training week, level of difficulty was raised when the participant scored 80% or higher in the letter memory task. All subjects reached the most difficult level by the end of training. The final training task was a keep-track task. This task, too, taxes updating, but is structurally different from the other updating tasks in the training battery. The inclusion of the keep-track task should contribute toward strengthening of a general updating skill. In each trial, 15 words from different semantic categories were presented serially in random order (2.0\u2009s per word) and participants were instructed to place the words into categories (animals, articles of clothing, countries, relatives, sports, professions), indicated by boxes at the bottom of the screen. They had to continuously update their working memory content and remember the last presented word in each category at the end of the presentation. Participants responded by typing the last presented word under each category box when the trial ended.\n\nPET scanning {#Sec9}\n------------\n\nThe pre- and post-training PET scans were identical: both involved two scans of which one was performed in the morning and the other in the afternoon. The letter-memory task was administered during the morning scan, whereas the digit n-back task was administered during the afternoon scan. The two assessments were separated by a 90-min lunch break.\n\nDuring the first PET scan, we administered a computerized letter-memory task that taps verbal WM updating. Participants were shown 7--15 letters in a sequence; when a sequence suddenly ended, they were asked to report the last four letters in correct order by pressing buttons corresponding to A, B, C, and D. This task was preceded by a structurally equivalent control task not taxing updating (all letters in a sequence were identical and participants reported that letter). Prior to PET scanning, participants acquainted themselves with the task during a practice session. The PET session started with the control task (5--10\u2009min prior to bolus injection continuing for 55\u2009min post-injection), followed by the letter-memory task (25\u2009min). The sequence began with instructions shown for 7000 ms. Stimulus duration was 2000 ms with a fixation cross in-between for 1000 ms. The participants were allowed 7000 ms to respond.\n\nDuring the second PET scan, the participants performed a digit n-back task that measures verbal WM updating. In this task, they were to determine if a currently visible digit was the same as the previous digit (1-back) or the digit that was presented three steps back (3-back). The dependent measure was target accuracy on the 3-back vs. 1-back condition. After participants had read the task instructions and completed one 1-back and one 3-back practice task, the PET session was initiated. During scanning, they completed about 55\u2009min of the 1-back (baseline) task, after which they immediately continued with a period of 25\u2009min performing the 3-back task. Task instructions, informing the participant of which n-back version to perform, were shown at the beginning of each task version for 5000 ms, after which a digit was shown for 1500 ms. The digit was replaced by a fixation point that was visible for 450 ms, and after this the fixation point was once again replaced by a digit. The alternation of fixation points and digits continued until the end of the task. Targets encompassed 45% of the trials (i.e., the current digit and the digit n steps back matched), whereas 55% comprised non-targets. The participants responded by pressing a match button for targets and a non-match button for non-targets.\n\nThe study protocol and procedure was virtually identical to those employed by B\u00e4ckman *et al*.^[@CR19]^. The only differences between the studies concerned the duration of the letter- memory and digit n-back tasks performed during the PET scans. Unlike B\u00e4ckman *et al*.^[@CR19]^, the current letter-memory task included a 55-minute baseline phase and a 25-minute test phase, with no return to baseline. In other respects, the letter-memory tasks were identical. The digit n-back task was included in the neuropsychological off-line battery in B\u00e4ckman *et al*.^[@CR19]^, whereas here it was altered (i.e., much longer) and performed during PET scanning to mirror the design of the letter-memory task.\n\nImaging methods {#Sec10}\n---------------\n\nWe used HRRT-PET (High Resolution Research Tomograph; Siemens Medical Solutions, Knoxville, TN, USA) and \\[^11^C\\]raclopride with bolus-plus-infusion^[@CR32]^ to measure striatal dopamine D2 receptor availability during WM vs. control task performance. Raclopride was prepared from \\[^11^C\\]methyl triflate following an established procedure^[@CR33]^. Its radiochemical purity and specific radioactivity were determined using high-performance liquid chromatography and ultraviolet detection at 214\u2009nm. The aimed magnitude of the bolus was 50% of the total tracer volume^[@CR34]^, and constant infusion continued until 80\u2009min after bolus injection (Kbol\u2009=\u200980\u2009min). At scan start, the actual bolus injection and infused doses did not differ between groups or scans (Table\u00a0[3](#Tab3){ref-type=\"table\"}). Emission list-mode data were histogrammed into 3D sinograms in 20 time frames of variable length (8\u2009\u00d7\u20092\u2009min, 4\u2009\u00d7\u20093\u2009min, 2\u2009\u00d7\u20094\u2009min, 1\u2009\u00d7\u20095\u2009min, 1\u2009\u00d7\u20096\u2009min, 1\u2009\u00d7\u20098\u2009min, 3\u2009\u00d7\u20098.3\u2009min), taking declining radioactivity and task timing into account. Before each emission scan, a transmission scan was performed using a ^137^Cs point source. Tissue-attenuation maps were reconstructed using the maximum a posteriori transmission data (MAP-TR) algorithm with segmentation. Scattered events were estimated using the single-scatter simulation algorithm, and randoms were estimated from the block singles with a variance-reduction algorithm. All corrections were applied within the statistical image-reconstruction algorithm^[@CR35]^.Table 3Administered radioactivities (MBq; mean\u2009\u00b1\u2009*SD*) in fast bolus and constant infusion formulations, and *p*-values in *t*-tests across groups and scans.BolusInfusionBolusInfusionBolusInfusionControl groupTraining group*t*-test (groups)Letter memoryPre Training262\u2009\u00b1\u200918 (*n*\u2009=\u20099)252\u2009\u00b1\u200914272\u2009\u00b1\u200931 (*n*\u2009=\u200913)255\u2009\u00b1\u2009280.430.77Post training267\u2009\u00b1\u200919 (*n*\u2009=\u200910)253\u2009\u00b1\u200915280\u2009\u00b1\u200925 (*n*\u2009=\u200911)265\u2009\u00b1\u2009250.520.22*t* test (scans)0.760.720.880.43n-backPre training267\u2009\u00b1\u200936 (*n*\u2009=\u200910)249\u2009\u00b1\u200932262\u2009\u00b1\u200929 (*n*\u2009=\u200913)264\u2009\u00b1\u2009250.740.22Post training265\u2009\u00b1\u200924 (*n*\u2009=\u200910)260\u2009\u00b1\u200917269\u2009\u00b1\u200917 (*n*\u2009=\u200913)264\u2009\u00b1\u2009160.440.51*t* test (scans)0.930.430.480.96\n\nSpecial attention was paid to minimization of head-motion artifacts based on earlier observations of its detrimental effects on the PET signal and its interpretation^[@CR36]^. Specifically, head motion was minimized using an individually molded thermoplastic mask, and image reconstructions were made using an in-house version of the multiple-acquisition frame (MAF)^[@CR36]^ based motion-compensated image reconstruction algorithm. The algorithm employs external motion tracking (MT) as given by Vicra (Northern Digital) infrared detection, to define motion-free (amplitude less than 2.5\u2009mm) framing of list-mode PET data. This procedure compensates for attenuation map misalignments using mutual information- based image registration and finally combines the motion-free subframes into original, desired framing. In a phantom study by Johansson *et al*. \\[in preparation\\], we showed good performance of the MAF-based algorithm, but a residual bias was seen in the presence of rapidly oscillating motion. Rapidly oscillating, high amplitude motion is not often observed in human PET scans but, if present, may introduce erroneous perturbations in the PET signal, which in turn could be incorrectly interpreted as activation effects^[@CR36],\\ [@CR37]^.\n\nMovements that are beyond the capabilities of the motion-compensation algorithm were sought through analysis of the number of subframes, which was identified as an important factor in our previous phantom experiment with regard to the algorithm's performance \\[Johansson *et al*., in preparation\\]. Analysis showed that in 84% (89/106) of the sessions, no additional acquisition frames were generated, indicating subthreshold (2.5 mm) amplitude of motion in all frames. The remaining cases were investigated further through detailed assessment of the time-motion graphs representing three translations and three rotations of the motion-recording target at the forehead at 1\u2009Hz frequency. On the basis of visual inspection of the motion data, five letter-memory scans from four individuals, and three n-back scans from two individuals were deemed unrecoverable using the motion-compensation algorithm, and were thus excluded from analysis. In addition to those excluded due to extensive motion, the letter-memory PET variables of four subjects were missing due to other technical reasons. In one case, the ^11^C-raclopride infusion was started approximately 20\u2009min overdue, in one case the PET data acquisition failed, and two participants were not PET scanned post training. Thus, a total of eight participants were missing from the PET letter-memory analysis, resulting in an *n* of 20. In turn, the n-back PET variables were missing from three subjects, two of whom were not scanned after training, and in one case the data acquisition failed. Hence, altogether five participants were missing from the n-back analysis, resulting in an *n* of 23.\n\nStructural MR imaging was performed for excluding anatomical abnormalities and for anatomical reference of the PET data. T1-weighted MRI data were acquired using a 3T scanner (Philips Ingenuity TF PET/MR) with a 1\u2009mm\u2009\u00d7\u20091\u2009mm\u2009\u00d7\u20091\u2009mm voxel size covering the whole brain. The T1-weighted MRI data were pre-processed using FreeSurfer (Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, US) to obtain automated region-of-interest (ROI) definitions^[@CR38]^. Subsequently, a PET sum image was registered to the skull-stripped MR image, using the mutual-information optimization algorithm in Statistical Parametric Mapping (version 8, SPM8; Wellcome Institute, London, UK). Spatial normalization parameters to the Montreal Neurological Institute (MNI) space coordinates were determined on the basis of MRI data using the unified-segmentation algorithm in SPM8^[@CR39]^ yielding a mapping for registered PET data normalization. The PET-image analysis was restricted to striatum, which serves as a central hub of dopaminergic activation. Striatal subvolumes were defined on the basis of FreeSurfer segmentations, combining segments of caudate, putamen, and nucleus accumbens into one composite striatal structure in each subject's individual space. Striatal masks were used in edge-preserving spatial smoothing of the PET data; outside-mask voxels were omitted in the Gaussian filter kernel (10\u2009mm FWHM) calculation, attenuating the impact of partial-volume effects near the edges. Extensive spatial smoothing was deemed necessary for successful calculation of binding potential (BP) from the PET activation data.\n\nModeling of the dynamic PET data was based on an extension of the conventional simplified reference tissue model (SRTM)^[@CR40]^, including the effect of activation as an additional parameter. A similar approach as described by Alpert and colleagues^[@CR41]^ was adopted, where the activation effect on the ^11^C-raclopride signal was modeled by the time-dependent activation function *h*(*t*), and the relevant linear equations are:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$[\\begin{array}{c}{C}_{T}({t}_{1})\\\\ \\vdots \\\\ {C}_{T}({t}_{m})\\end{array}]=[\\begin{array}{c}{C}_{R}({t}_{1})\\,\\\\ \\vdots \\\\ {C}_{R}({t}_{m})\\end{array}\\begin{array}{c}\\,{\\int }_{0}^{{t}_{1}}{C}_{R}(u)du\\,\\\\ \\vdots \\\\ \\,{\\int }_{0}^{{t}_{m}}{C}_{R}(u)du\\,\\end{array}\\begin{array}{c}\\,-{\\int }_{0}^{{t}_{1}}{C}_{T}(u)du\\,\\\\ \\vdots \\\\ \\,-{\\int }_{0}^{{t}_{m}}{C}_{T}(u)du\\end{array}\\begin{array}{c}\\,-{\\int }_{0}^{{t}_{1}}{C}_{T}(u)h(u)du\\\\ \\vdots \\\\ \\,-{\\int }_{0}^{{t}_{m}}{C}_{T}(u)h(u)du\\end{array}]\\,[\\begin{array}{c}\\begin{array}{c}{R}_{1}\\\\ {k}_{2}\\end{array}\\\\ {k}_{2a}\\\\ \\gamma \\end{array}]$$\\end{document}$$where the weighting factor $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\,\\gamma $$\\end{document}$ is depicted as the activation effect, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${C}_{T}({t}_{i})$$\\end{document}$ and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${C}_{R}({t}_{i})$$\\end{document}$ are the instantaneously measured radioactivity concentrations in target and reference tissue, respectively, and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${R}_{1}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2a}$$\\end{document}$ are the conventional SRTM rate constants. Compared to our previous work^[@CR19]^, we now employed an explicit BP calculation algorithm for the activation phase, to quantitatively assess the magnitude of this effect relative to baseline. First, baseline BP (BP~0~) was determined from the fit parameters $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2}$$\\end{document}$, and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2a}$$\\end{document}$, which in turn were estimated using a linear least-squared solver in Matlab (version R2011b, Mathworks, US). Second, the activation BP (BP~1~) was determined from the fit parameters $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2}$$\\end{document}$, $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$${k}_{2a}$$\\end{document}$, and $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\gamma $$\\end{document}$, using an integral of $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\gamma h(t)$$\\end{document}$ and averaged over the activation period. As the activation function, we employed a simplified gamma-variate function^[@CR42]^, with a fixed peak time of 8\u2009min and 20\u2009sec from active task initiation, \u03bb\u2009=\u20093, and onset of activation at 55\u2009min from the injection. Model calculation was restricted to the aforementioned striatal volume, and the resulting parametric images were normalized using the MRI-based deformation into MNI space coordinates. Voxel-wise statistical analysis was performed in SPM8 within a striatal search volume (2934 voxels, \\~10\u2009cm^3^), determined as the intersection of individual striatal subvolumes warped in MNI space. Because the images could not be assumed to be normally distributed, non-parametric analyses were conducted using SnPM (version SnPM12, <http://www2.warwick.ac.uk/fac/sci/statistics/staff/academic-research/nichols/software/snpm>.\n\nA 2 (group)\u2009\u00d7\u20092 (time) ANOVA was conducted on BP\u2009=\u2009BP~1~\u2212BP~0~. Variance smoothing was used with an FWHM of 10 mm. Five thousand permutations were used to determine a *p*\u2009\\<\u20090.05 threshold within the striatal search space. Paired *t*-tests with a *p*\u2009\\<\u20090.01 threshold were used to assess main effects of task (letter memory and n-back) on BP, combining both groups.\n\nOff-line transfer tasks {#Sec11}\n-----------------------\n\nThe pre- and post-training neuropsychological assessments included three WAIS-III subtests (digit span, letter-number sequencing, and digit symbol), a number-letter task, a Simon task, a visuospatial n-back task, and an episodic recall task. A pattern-comparison task, a number-copying task, Trail Making parts A and B, and the WAIS-III vocabulary subtest were only included before training. The pre-training assessment also included a background questionnaire, the Edinburgh Handedness Inventory, and the BDI-II.\n\n### Digit span {#Sec12}\n\nThe digit span test was used to assess auditory attention as well as passive (forward span) and active (backward span) WM. Participants were asked to recall sequences of digits in the same or reversed order in which they were presented. Test administration, items, and scoring followed standard WAIS-III procedures^[@CR43]^. Maximum forward and backward spans were used as dependent measures.\n\n### Letter-number sequencing {#Sec13}\n\nThe letter-number sequencing test was also used to assess auditory attention as well as passive and active working memory. Participants were instructed to recall and sort number-letter sequences. Test administration, items, and scoring followed standard WAIS-III procedures^[@CR43]^. The total score was used as the dependent measure.\n\n### Visuospatial n-back {#Sec14}\n\nThe visuospatial n-back task was used to assess visuospatial WM updating^[@CR44]^. In this task, participants were asked to determine if a currently visible box was in the same location as the previous box (1-back) or the box that was presented two steps back (2-back). A matrix with eight stimulus locations was used (a 3\u2009\u00d7\u20093 matrix without the middle square). Accuracy was used as the dependent measure.\n\n### Digit-symbol coding {#Sec15}\n\nThe digit-symbol coding test was used to assess perceptual speed. Participants were asked to draw, as quickly as possible, different symbols in empty spaces on the basis of digits that were located above each empty space. A time limit of 90\u2009sec was used here, but in all other respects test administration, items, and scoring followed standard WAIS-III procedures^[@CR43]^. The total score was used as the dependent measure.\n\n### Number-letter task {#Sec16}\n\nThe number-letter task was used to assess set shifting. Here, participants were asked to categorize either the number or the letter in a number-letter pair depending on in which of two vertically aligned boxes the number-letter pair appeared. The task included both switch and no-switch trials, from which mixing and switching costs were calculated^[@CR44]^. The switching cost reflects the temporary cognitive load that is related to a task shift, whereas the mixing cost reflects the cost of maintaining attentional control in a situation where two task sets are active. The mixing and switching costs for RTs and accuracy were used as dependent measures.\n\n### Simon task {#Sec17}\n\nThe Simon task was used to assess inhibitory control. In this task, participants were to determine the color of a square while disregarding where the square appeared^[@CR45]^. On congruent trials, the color and location of the square matched the positioning of the appropriate response key, whereas the stimulus and correct response key were crossed on incongruent trials. The Simon effect was calculated by comparing the performance on congruent and incongruent trials. Simon effects for accuracy and RT were used as dependent measures.\n\n### Episodic word recall {#Sec18}\n\nThe word recall task was used to assess verbal episodic memory^[@CR46]^. Participants were instructed to read and remember 18 concrete nouns presented on a computer screen. They were instructed to recall as many words as possible, and the test leader recited any words that had been omitted. After hearing the omitted words, participants again attempted to recall as many words as possible, and once again the test leader recited any omissions. Finally, participants were again asked to recall as many words as possible. The total number of correctly recalled words was used as the dependent measure.\n\n### Pattern comparison {#Sec19}\n\nThe pattern-comparison task was used to assess motor speed^[@CR47]^. In this task, participants were to determine, as quickly as possible, if two symbols were identical. The total number of correctly evaluated symbol-pairs was used as the dependent measure.\n\n### Number copying {#Sec20}\n\nThe number-copying task was also used to assess motor speed^[@CR47]^. Here, participants were asked to copy, as quickly as possible, a set of numbers. The total number of correctly copied numbers was used as the dependent measure.\n\n### Trail Making {#Sec21}\n\nThis task was used to assess visual search, hand-eye coordination, and set shifting^[@CR48]^. In part A, participants were to connect, as quickly as possible, a set of numbers in ascending order. In part B, they were to connect numbers and letters in both ascending and alphabetical order, constantly switching between the two types of stimuli (1-A-2-B, etc.). The dependent variable was the difference score obtained by subtracting the completion time of part A from part B.\n\n### Vocabulary {#Sec22}\n\nThe vocabulary test was used to assess word comprehension. Participants were asked to explain the meaning of words of increasing complexity. Test administration, items, and scoring followed standard WAIS-III procedures^[@CR43]^. The total raw score was used as the dependent measure.\n\nStatistical analyses {#Sec23}\n--------------------\n\nThose participants who were extreme outliers (i.e., showing a deviation of at least three times the interquartile range) before training were removed from the specific statistical analyses on which they deviated. Training effects were analyzed by mixed-model ANOVAs.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThe study was financially supported by grants from the Academy of Finland, the \u00c5bo Akademi University endowment (the BrainTrain project), and the Turku University Hospital. LB was supported by an Alexander von Humboldt Research Award and a donation from the af Jochnick Foundation. LN was supported by a Scholar grant from the Knut and Alice Wallenberg Foundation.\n\nA.S., K.A., and O.W. contributed to study design, acquisition of data, analyzed the neuropsychological data, interpretation of data, and revising the article. J.J. and M.A. analyzed and interpreted the PET data, and revised the article. L.B., M.L., J.O.R. and L.N. all contributed to study design, interpretation of data, and drafting of the article. All authors have given their approval of the final version.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nKLF4, a member of the kr\u00fcppel-like factor (KLF) transcription factor family, is a potential tumor suppressor in several types of human malignancies \\[[@R1]--[@R4]\\]. For instance, KLF4 was significantly down-regulated in prostate cancer cell lines compared with nontumorigenic prostate cells \\[[@R5]\\]. RNA activation-mediated overexpression of KLF4 inhibited prostate cancer cell proliferation and altered the expression of several downstream cell-cycle-related genes \\[[@R5]\\]. Besides, ablation of KLF4 in gastric progenitor cells promoted transformation of the gastric mucosa and tumorigenesis in the antrum in mice \\[[@R6]\\]. Moreover, KLF4 protein expression was decreased or lost in hepatocellular carcinoma (HCC) tissues and, in particular, lymph node metastases when compared with that in normal liver \\[[@R7]\\]. Deficiency of KLF4 expression was significantly associated with poor survival, and also a prognostic marker in HCC patients \\[[@R7]\\]. Indeed, KLF4 could inhibit cell proliferation, invasion and epithelial to mesenchymal transition (EMT) through up-regulation of vitamin D receptor (VDR) and repression of \u03b2-catenin and SLUG, respectively \\[[@R7]--[@R10]\\]. However, the molecular mechanisms for the down-regulation of KLF4 in HCC tissues remain poorly understood.\n\nIt has been shown that KLF4 protein could be regulated by acetylation, phosphorylation and sumoylation \\[[@R11]--[@R14]\\]. More importantly, Lim KH et al. reports that KLF4 undergoes proteasomal degradation and some lysine residues are critical for its ubiquitination \\[[@R15]\\]. However, the factors regulating the ubiquitination and degradation of KLF4 have not been fully understood.\n\nThe Skp1-Cul1-F box protein ubiquitin ligases (SCFs), consisting of Skp1, Cul1, and one of a member of F box proteins, play important roles in various biological events by their ability to bind and destabilize substrates \\[[@R16], [@R17]\\]. For instance, FBXW7 could inhibit tumor growth by targeting Cyclin E, c-myc and SRC-3, all of which are thought to be oncoproteins \\[[@R18]--[@R21]\\]. Indeed, FBXW7 mutations have been found in a variety of primary human tumor types, providing novel insights into cancer progression and development of therapeutic drugs \\[[@R22], [@R23]\\]. In the current study, we used an immunoprecipitation and tandem mass spectrometry (IP-MS) analyses to search for KLF4-interacting protein with potential ubiquitin ligase activity and identified FBXO22 that binds with KLF4.\n\nRESULTS {#s2}\n=======\n\nIdentifying FBXO22 as a KLF4-interacting protein {#s2_1}\n------------------------------------------------\n\nTo identify novel KLF4-interacting proteins, HepG2 cells were transfected with adenovirus expressing Flag-KLF4 or empty vector for 36 hr. Then proteins were isolated by Flag M2 beads, separated by SDS-PAGE, and identified by tandem mass spectrometry (MS/MS). As shown in the Figure [1A](#F1){ref-type=\"fig\"}, the examination of copurified endogenous proteins revealed the presence of unique peptides derived from SCF subunits, including Cullin1, SKP1 and F-box protein FBXO22 as major KLF4-associated proteins. FBXO22 is of interest since the F-Box protein family has been reported as the largest and most versatile class of E3 ubiquitin ligase. A FBXO22-KLF4 interaction was further confirmed by substantial portion of endogenous FBXO22 and KLF4 co-immunoprecipitated together (Figure [1B](#F1){ref-type=\"fig\"} and [1C](#F1){ref-type=\"fig\"}). This interaction was largely attenuated in KLF4-knockdown cells, suggesting the specificity of this interaction (Figure [1D](#F1){ref-type=\"fig\"}). To confirm the specific binding between KLF4 and FBXO22, we screened 11 human F box proteins. These FLAG tagged F-box proteins were transfected into 293T cells and then immunoprecipitated to evaluate their interaction with endogenous KLF4. We found that, although all F-box proteins could interact with SKP1, only FBXO22 interacts with endogenous KLF4 ([Supplementary Figure 1](#SD1){ref-type=\"supplementary-material\"}). In contrast, related F box proteins such as FBXO4, FBXO6, FBXO30, FBXO34, or FBXO36, which all belong to F-box only (FBXO) proteins, did not bind endogenous KLF4.\n\n![FBXO22 interacts with KLF4 *in vivo*\\\n**A.** HepG2Cells were transfected with adenovirus expressing Flag-KLF4 or empty vector for 36 hr. The pull-down products from Flag M2 beads affinity purification were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), visualized by Colloidal blue staining, and analyzed by mass spectrometry. The identified SCF^FBXO22^ components are listed. **B--C.** Endogenous association between FBXO22 and KLF4 in HepG2 cells was performed by co-immunoprecipitation (CO-IP) experiments. **D.** The interaction of FBXO22 and KLF4 in cells depleted with KLF4. HepG2 cells were administered with adenoviral shRNA targeting KLF4 or a negative control (NC)](oncotarget-06-22767-g001){#F1}\n\nFBXO22 promotes KLF4 ubiquitination and degradation {#s2_2}\n---------------------------------------------------\n\nMembers of F-Box proteins could recognize and mediate the degradation of target proteins. Thus, the fact that FBXO22 interacts with KLF4 prompted us to ask whether FBXO22 can promote the degradation of KLF4 as well. To test this hypothesis, adenovirus containing FBXO22 or GFP was introduced into HepG2 cells. Forced expression of FBXO22 led to a decreased protein levels of KLF4 (Figure [2A](#F2){ref-type=\"fig\"}). The effect on KLF4 stability is selective since other members of KLF family (KLF2 and KLF5) were unaffected by FBOXO22 overexpression (Figure [2A](#F2){ref-type=\"fig\"}). Besides, the reduction in KLF4 protein is not due to changes at the KLF4 mRNA levels, as shown by quantitative real-time PCR ([Supplementary Figure 2](#SD1){ref-type=\"supplementary-material\"}). The decrease of KLF4 was also observed when FBXO22 was overexpressed in other HCC cells (HuH7 and Hep3B) and normal hepatocytes (L02) ([Supplementary Figure 3A--3C](#SD1){ref-type=\"supplementary-material\"}). Ubiquitin-conjugation to KLF4 was enhanced by FBXO22 overexpression (Figure [2B](#F2){ref-type=\"fig\"}). Besides, p21^Cip1/WAF1^ and Cyclin B1, two cell-cycle regulators, were activated and repressed by KLF4 respectively \\[[@R24], [@R25]\\] and modulated by FBXO22 ([Supplementary Figure 3D--3E](#SD1){ref-type=\"supplementary-material\"}).\n\n![FBXO2 promotes KLF4 turnover in HCC cells\\\n**A.** The levels of KLF4, KLF2 and KLF5 in cell lysates were determined by Western blot. HepG2 cells were transfected with adenovirus expressing GFP or FBXO22 for 36 hr. **B.** Ubiquitination of KLF4 in whole-cell lysates was determined by Western blot. **C.** HepG2 cells were transfected with adenovirus expressing GFP or FBXO22. 36 hr after transfection, the cells were either lysed directly or incubated in the presence of cyclohexamide (50 \u03bcg/ml) for the indicated time period in order to determine KLF4 turnover. **D.** The levels of KLF4 in cell lysates were determined by Western blot. HepG2 cells were transfected with adenoviral shRNA targeting FBXO22 or a negative control (NC) for 36 hr. **E.** KLF4 turnover was determined by Western blot in HepG2 cells transfected with adenoviral shRNA targeting FBXO22 or a negative control (NC) for 36 hr.](oncotarget-06-22767-g002){#F2}\n\n![FBXO22 promotes HCC cell proliferation and invasion\\\nCell proliferation **A, D.**, colony formation **B, E.** and invasion **C, F.** assays were determined in HepG2 cells with FBXO22 overexpression or knockdown.](oncotarget-06-22767-g003){#F3}\n\nTo further analyze the FBXO22-mediated down-regulation of KLF4 in more detail, degradation kinetics of KLF4 was determined in HepG2 cells. As expected, expression of FBXO22 significantly reduced the half-life of KLF4 (Figure [2C](#F2){ref-type=\"fig\"}). On the other hand, we examined whether knockdown of FBXO22 could inhibit KLF4 turnover. FBXO22 deficiency resulted in an increased expression and half-life of KLF4 (Figure [2D--2E](#F2){ref-type=\"fig\"}, [Supplementary Figure 4A--4C](#SD1){ref-type=\"supplementary-material\"}), indicating that FBXO22 contributes to the turnover of KLF4.\n\nFBXO22 promotes HCC cell proliferation and invasion *in vitro* {#s2_3}\n--------------------------------------------------------------\n\nGiven that KLF4 appears to function as a tumor suppressor in HCC, we next determined what physiologically stimulates FBXO22 to interact with and stabilize KLF4. As a result, the interaction of KLF4 and FBXO22 was enhanced during the cell-cycle progression, leading to a reduced protein levels of KLF4 ([Supplementary Figure 5](#SD1){ref-type=\"supplementary-material\"}). Besides, abilities of cell proliferation, colony formation and invasion were significantly enhanced by FBXO22 overexpression (Figure [3A--3C](#F3){ref-type=\"fig\"}). In contrast, disruption of FBXO22 in HepG2 cells inhibited cell growth (Figure [3D--3F](#F3){ref-type=\"fig\"}). Additionally, KLF4 deficiency largely blocked the proliferative roles of FBXO22 in HepG2 cells (Figure [4A--4B](#F4){ref-type=\"fig\"}). In agreement, expression levels of p21^Cip1/WAF1^ and Cyclin B1 inhibited by FBXO22 were also attenuated by KLF4 silencing (Figure [4C](#F4){ref-type=\"fig\"}), suggesting the oncogenic effects of FBXO22, at least in part, depend on its regulation of KLF4.\n\n![The oncogenic roles of FBXO22 rely on its regulation of KLF4\\\n**A--B.** Cell proliferation (A) and invasion (B) assays in cells with FBXO22 overexpression and (or) KLF4 knockdown as indicated. HepG2 cells were transfected with adenovirus expressing GFP or FBXO22 for 24 hr, and then transfected with adenoviral shRNA targeting KLF4 or a negative control (NC) for another 24 hr. **C.** Relative mRNA levels of p21 and Cyclin B1 in HepG2 cells as in (A-B)](oncotarget-06-22767-g004){#F4}\n\nFBXO22 promotes HCC growth *in vivo* {#s2_4}\n------------------------------------\n\nTo demonstrate the function of FBXO22 *in vivo*, HepG2 cells stably overexpressing FBXO22 or GFP protein were injected subcutaneously into two bilateral sites of BALB/c nude mice. In agreement, the tumor volume and weight were markedly increased in FBXO22-overexpressed tumors compared to control tumors (Figure [5A--5B](#F5){ref-type=\"fig\"}). KLF4 protein contents were reduced in FBXO22 overexpressed tumor tissues (Figure [5C](#F5){ref-type=\"fig\"}), suggesting an oncogenic role of FBXO22 *in vivo*. Consistently, in FBXO22 overexpressed tumor tissues, the expression levels of p21^Cip1/WAF1^, and Cyclin B1 were also affected (Figure [5D--5E](#F5){ref-type=\"fig\"}).\n\n![FBXO22 promote HCC growth *in vivo*\\\n**A--B.** HepG2 cells stably transfected with FBXO22 or GFP were injected into nude mice and followed up for tumorigenesis. Growth curve of tumor volumes (A) and tumor weights (B) were taken 4 weeks after injection. (*n* = 6--8 for each group) **C.** Representative protein levels of FBXO22, GFP and KLF4 were determined in the two groups of tumors. **D--E.** mRNA and protein levels of p21 and Cyclin B1 in two groups of tumors were determined by real-time PCR and Western blot.](oncotarget-06-22767-g005){#F5}\n\nUp-regulation of FBXO22 in human hepatocarcinoma {#s2_5}\n------------------------------------------------\n\nFinally, we asked whether FBXO22-mediated KLF4 degradation is involved in human HCC development. As shown in the Figure [6A](#F6){ref-type=\"fig\"}, the expression of FBXO22 was significantly increased in tumor samples compared with adjacent normal tissue. In contrast, down-regulation of KLF4 was detected in tumor tissues (Figure [6A](#F6){ref-type=\"fig\"}). Association analysis of 30 HCC tissues indicated that the protein levels of FBXO22 and KLF4 was negatively correlated (Figure [6B](#F6){ref-type=\"fig\"}), supporting the connection between FBXO22 and KLF4.\n\n![FBXO22 expression was increased in human hepatocarcinoma, correlating with reduced KLF4 expression\\\n**A.** Protein levels of FBXO22 and KLF4 were determined by Western blot in human HCC tissues and adjacent normal tissues. **B.** The correlation between FBXO22 and KLF4 protein levels in paired HCC and non-tumor tissues.](oncotarget-06-22767-g006){#F6}\n\nIt has been well-established that inflammatory process is tightly associated with many different cancer types, including HCC. We speculate that the up-regulation of FBXO22 in HCC tissues might be attributed to aberrant activation of inflammatory signaling. Therefore, HepG2 and HuH7 cells were treated with TNF\u03b1 and IL-1\u03b2. As expected, treatment of these pro-inflammatory cytokines significantly increased the mRNA and protein levels of FBXO22 ([Supplementary Figure 6A--6D](#SD1){ref-type=\"supplementary-material\"}). These results suggest that the up-regulation of FBXO22 in HCC tissues, at least in part, due to persistent activation of inflammatory signaling.\n\nDISCUSSION {#s3}\n==========\n\nIn the present study, we have demonstrated that KLF4 is a biological target of FBXO22, which promotes the turnover of KLF4 and thereby promotes HCC cell proliferation and invasion. It has been shown that promoter methylation of KLF4 locus could contribute to its down-regulation *in vitro* and *in vivo* \\[[@R26], [@R27]\\]. KLF4 was also identified as a direct target of several microRNAs, a class of small and non-coding RNA molecules \\[[@R28], [@R29]\\]. Besides, TGF-\u03b2 signaling could promote KLF4 degradation through Cdh1-anaphase promoting complex \\[[@R30]\\]. In addition, Phosphorylation of KLF4 by ERK1 or ERK2 enhanced its interaction with the F-box proteins \u03b2TrCP1 or \u03b2TrCP2, which led to KLF4 ubiquitination and degradation \\[[@R13]\\]. Because KLF4 is a tumor suppressor clearly involved in the initiation and (or) progression of HCC, it would be reasonable that important regulatory checks on its expression may exist at multiple levels. Therefore, our data present a novel mechanism for the dysregulation of KLF4 in HCC. However, further studies are still needed to establish the connection between FBXO22 and KLF4 in other gastrointestinal cancer.\n\nTan MK et al. reports that FBXO22 could control the activity of KDM4A through targeting it for proteasomal turnover \\[[@R31]\\], by which FBXO22 regulates histone H3 lysine 9 and 36 methylation levels in cell cycle. Besides, a recent study indicates that FBXO22 is essential for optimal synthesis of the N-methyl-D-aspartate (NMDA) receptor coagonist D-serine \\[[@R32]\\]. Here, we provide evidence that FBXO22 is up-regulated and might be an oncogene in HCC. However, the expression and roles of FBXO22 in other types of human cancers remain to be determined.\n\nIn conclusion, our study now suggests that FBXO22 controls the turnover of KLF4 protein *in vivo* and *in vitro*. Inhibition or down-regulation of FBXO22 by gene therapy might be beneficial in patients with malignant tumors.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nHuman tissue samples {#s4_1}\n--------------------\n\n30 paired of HCC tissues and adjacent non-tumor normal tissues were collected from routine therapeutic surgery at our department. All samples were obtained with informed consent and approved by the hospital institutional review board.\n\nCell culture {#s4_2}\n------------\n\nHCC cell lines (HepG2, HuH7 and Hep3B cells) were obtained from The Cell Bank of Type Culture Collection of Chinese Academy of Sciences (CAS, Shanghai). Cells were grown in Dulbecco\\'s modified Eagle\\'s medium (DMEM, Gibco, Shanghai) supplemented with 10% fetal bovine serum (Gibco) and maintained at 37\u00b0C in a humidified atmosphere with 5% CO~2~.\n\nMouse experiments {#s4_3}\n-----------------\n\nMale BALB/c nude mice aged 4 weeks were purchased from Shanghai Laboratory Animal Company (SLAC, Shanghai). 4.5 \u00d7 10^6^ HepG2 cells stably expressing GFP or FBXO22 were injected subcutaneously to the skin under the front legs of the mouse. The mice were observed over 4 weeks for tumor formation. After the mice were sacrificed, the tumors were recovered and the wet weights of each tumor were determined.\n\nImmunoprecipitation {#s4_4}\n-------------------\n\nThe purification process has been reported previously. Briefly, 5 \u00d7 10^7^ cells infected with adenovirus expressing Flag-KLF4 or empty vector were lysed in 5 mL of NETN lysis buffer for 20 min on ice. Lysates were cleared using centrifugation at 13000 rpm for 20 min; the supernatant was then subjected to immunoprecipitation with 50 \u03bcL of anti-FLAG M2 affinity resin overnight at 4\u00b0C with gentle inversion. Resin containing immune complexes was washed with 1 mL ice cold lysis buffer 6 times and proteins were eluted with 150 \u03bcL 150 \u03bcg/mL 3 \u00d7 Flag-peptide (Sigma) in TBS for 10 min. Proteins were precipitated with cold acetone.\n\nIn gel tryptic digestion {#s4_5}\n------------------------\n\nImmunoprecipitation samples were separated by SDS-PAGE, and visualized with colloidal Coomassie blue. The interest bands lane were cut into 1 mm slices, and each slice was washed twice with 50 mM NH4HCO3, 50% ACN and dehydrated with ACN. Proteins were reduced and alkylated by treating them with 10 mM DTT and 55 mM iodoacetamide, respectively. After washing with 50 mM NH4HCO3 and ACN, proteins were digested in gel with trypsin and incubated overnight at 37\u00b0C. Tryptic peptides were extracted from the gel pieces with 60% ACN, 0.1% trifluoroacetic acid. The peptide extracts were vacuum centrifuged to dryness.\n\nSamples were desalted with 10ul U-C18 pipette tips.\n\nNano-HPLC-MS/MS Analysis {#s4_6}\n------------------------\n\nThe dried samples were dissolved in solvent A (0.1% formic acid, 2% acetonitrile, 98% H2O). Samples were then injected onto a manually packed reversed phase C18 column (150 mm \u00d7 79 \u03bcm, 3-\u03bcm particle size, Dikma, China) coupled to Easy nLC (Thermo Fisher Scientific, Waltham, MA). Peptides were eluted from 7% to 80% solvent B (0.1% formic acid in 90% acetonitrile and 10% H~2~O) in solvent A (0.1% formic acid in 2% acetonitrile and 98%H~2~O) with a 1 h gradient at a flow rate of 300 nl/min. The fractions were analyzed by using a Q Exactive mass spectrometer in a top16 data-dependent mode. For full MS spectra, the scan range was 350 to 1300 with a resolution of 70, 000. For MS/MS scan, the 16 most intense ions with charge state 2 and 3 in each full MS spectrum were sequentially fragmentated by higher energy collisional dissociation (HCD) with normalized collision energy of 28%. The resolution was 17, 500. The dynamic exclusion duration was set to be 60 s, and the isolation window was 1.5m/z. All MS raw files were analyzed by MaxQuant software (version 1.0.13.13) and Mascot software (version 2.1) against the database uniprot_Human to identify proteins.\n\nReal-time PCR analysis {#s4_7}\n----------------------\n\nTotal RNA from tissues and cells was extracted using the RNA Isolation Kit (Takara, Dalian, China) according to the manufacturer\\'s instructions. Quantitative real-time PCR was performed by using an Applied Biosystems 7300 Real-time PCR System and a TaqMan Universal PCR Master Mix. Expression of the KLF4, p21 and Cyclin B1 was normalized to that of the \u03b2-actin.\n\nWestern blot {#s4_8}\n------------\n\nCells were harvested and lysed with ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 100 mM 2-Mercaptoethanol, 2% w/v SDS, 10% glycerol). After centrifugation at 10000 \u00d7 g for 10 min at 4\u00b0C, proteins in the supernatants were quantified and separated by 10% SDS PAGE. Western blot assay was performed using anti-FBXO22, KLF4, p21, Cyclin B1, GFP and GAPDH antibodies (Abcam, USA). GAPDH was determined as a loading control\n\nBrdU and cell invasion assays {#s4_9}\n-----------------------------\n\nA cell proliferation enzyme-linked immunosorbent assay (BrdU kit; Beyotime) was used to analyze the incorporation of BrdU during DNA synthesis following the manufacturer\\'s protocols. Absorbance was measured at 450 nm in the Spectra Max 190 ELISA reader (Molecular Devices, Sunnyvale, CA). For cell invasion assays, cells were analyzed using extracellular matrix-coated invasion chambers (Millipore, CA, USA), and quantitated with a colorimetric microplate reader at 570 nm, according to the manufacturer\\'s instructions.\n\nColony formation assay {#s4_10}\n----------------------\n\nCells were seeded in a 6-well plate 48 hours posttransduction and cultured for 8 to 10 days at 37\u00b0C in 5% CO~2~. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS), washed twice with PBS, and stained with a crystal violet solution (1% crystal violet, 10% ethanol in water). Stained cells were washed thrice with water and counted by under an optical microscope.\n\nStatistical analysis {#s4_11}\n--------------------\n\nThe data shown represent the mean \u00b1 standard error (SE) values of three independent experiments. Significance was analyzed using Student\\'s *t*-test (\\**p* \\< 0.05, \\*\\**p* \\< 0.01, \\*\\*\\**p* \\< 0.001).\n\nSUPPLEMENTARY FIGURES {#s5}\n=====================\n\nThis work was supported by grants from the National Natural Science Foundation of China (No. 81202955, 81401881 and 81372287).\n\n**CONFLICTS OF INTEREST**\n\nThe authors declared no conflict of interest.\n"}
+{"text": "The DNA damage response (DDR) is a coordinated signalling network involving a large number of interacting proteins controlled by post-translational modifications and includes signalling, cell--cycle checkpoint enforcement and DNA repair[@b1][@b2]. Following DNA double-strand break (DSB) generation, the apical DDR kinase ataxia telangiectasia mutated (ATM) undergoes activation and phosphorylates the histone H2AX at serine 139 (named \u03b3H2AX); this event is necessary for the recruitment of DDR proteins to sites of DNA damage, including ATM phosphorylated at serine 1,981 (pATM) itself and p53-binding protein 1 (53BP1). Recently, a class of small non-coding RNAs (ncRNAs), termed DNA damage response RNAs (DDRNAs), has been shown to be generated on transcription of the damaged locus following DSB induction and then processed by the endoribonucleases DICER and DROSHA[@b3][@b4][@b5]. DDRNAs are dispensable for the direct recognition of DNA damage and \u03b3H2AX phosphorylation, but necessary for the secondary recruitment of DDR proteins to DSBs to form the so-called DDR foci[@b6]. A similar set of small ncRNAs, named damage-induced RNAs, has been shown to be involved in DNA repair by homologous recombination (HR) in plants and human cells[@b7][@b8], and by non-homologous end-joining (NHEJ) in plants[@b9].\n\nTelomeres are the nucleoprotein complexes located at the tips of eukaryotic chromosomes, composed of repetitive DNA (TTAGGG in vertebrates), and coated by a set of proteins collectively known as the shelterin complex[@b10]. Dysfunctional telomeres resemble DSBs and they have been observed during ageing, cancer and a number of medical conditions[@b11][@b12][@b13][@b14][@b15][@b16][@b17]. Apart from telomeric repeat-containing RNA (TERRA), a non-coding UUAGGG-rich transcript starting from promoters located in the subtelomeric region[@b18][@b19], in mammals no other transcripts at telomeres have been characterized so far.\n\nHere we provide evidence that both strands of deprotected telomeres are transcribed to generate telomeric DDRNAs (tDDRNAs) and their precursors, whose inhibition leads to a reduction in the DDR activation at dysfunctional telomeres in living cells and *in vivo*.\n\nResults\n=======\n\nTelomere deprotection triggers both strands' transcription\n----------------------------------------------------------\n\nTo explore the potential generation of DDRNAs and their role at dysfunctional telomeres, we used *Trf2* conditional knockout mouse embryonic fibroblasts (*Trf2*^*F/*+^ and *Trf2*^*F/F*^ mouse embryonic fibroblasts (MEFs)) carrying a Cre recombinase (*Rosa26-CreERT2*) inducible by 4-hydroxytamoxifen (4OHT)[@b20]. TRF2 is a component of the shelterin complex and prevents telomeric DNA from being recognized as a DSB[@b10]. Recombination by Cre promptly and robustly induced telomere deprotection and DDR activation at telomeres in the *Trf2*^*F/F*^ cell line only ([Supplementary Fig. 1a](#S1){ref-type=\"supplementary-material\"}; ref. [@b20]). DDRNAs are small double-stranded RNAs (dsRNAs) transcribed at DSBs and carry the sequence of the damaged genomic locus[@b21]. To study tDDRNAs we employed miScript, a quantitative PCR with reverse transcription (RT--qPCR)-based method[@b22] developed to detect and quantify small RNAs, and designed primers to selectively amplify either strands of potential tDDRNAs generated at deprotected telomeres: teloC (C-rich strand of tDDRNA) and teloG (G-rich strand of tDDRNA; [Supplementary Fig. 1b](#S1){ref-type=\"supplementary-material\"}). Using a gel-extracted fraction of RNAs shorter than 40 nucleotides---thus, minimizing the contribution of much longer telomeric transcripts, including TERRA[@b18][@b19][@b23]---we observed an induction of both teloC and teloG RNAs ([Fig. 1a](#f1){ref-type=\"fig\"}), thus with the potential to form *in vivo* double-stranded tDDRNA molecules. The signals detected were not due to genomic DNA contamination, since very low signal was detectable in the absence of reverse transcription ([Supplementary Fig. 1c](#S1){ref-type=\"supplementary-material\"}). To extend these observations beyond mouse cells or TRF2 genetic ablation only, we used T19 cells, a human HT1080 fibrosarcoma cell line derivative, expressing a doxycycline-induced FLAG-tagged dominant negative form of TRF2 (*TRF2 \u0394B\u0394M*), whose expression leads to telomere deprotection and DDR activation ([Supplementary Fig. 1d,e](#S1){ref-type=\"supplementary-material\"}; ref. [@b24]). We observed that also in this cell system, tDDRNA levels increased following telomere deprotection ([Supplementary Fig. 1f](#S1){ref-type=\"supplementary-material\"}). This indicates that tDDRNA induction mechanisms are conserved among species.\n\nTo further characterize the length and sequence of tDDRNAs, we devised and employed an innovative method for target enrichment of RNA, based on in solution capture of low-abundance small RNA species followed by next-generation sequencing, developed in our laboratories (Nguyen *et al*., in preparation). By this approach, we observed that telomere deprotection induced the accumulation of small RNA species generated from the transcription of both telomere strands, including the expected DDRNA size range products ([Fig. 1b](#f1){ref-type=\"fig\"}; [Supplementary Fig. 1g](#S1){ref-type=\"supplementary-material\"}). Interestingly, 20--23 nucleotide RNAs displayed a base bias at both 5\u2032- and 3\u2032- ends significantly different from the telomeric locus ([Supplementary Fig. 2a](#S1){ref-type=\"supplementary-material\"}), suggesting a regulated processing.\n\nWe next sought evidence for longer RNA species---we named them damage-induced long ncRNAs (dilncRNAs)---potential precursors of tDDRNAs. For this purpose, we performed a strand-specific fixed-length RT--qPCR[@b25] that allows the quantitative amplification of repetitive sequences, such as telomeric repeats, but not of RNA molecules shorter than 40 nucleotides such as DDRNAs ([Supplementary Fig. 2b](#S1){ref-type=\"supplementary-material\"}). In cells with deprotected telomeres, both teloC and teloG long transcripts were 10- to 20-fold more abundant than in control cells ([Fig. 1c](#f1){ref-type=\"fig\"}), indicating that telomere deprotection strongly induces telomere transcription on both strands. Although PCR amplification of UUAGGG repeat-containing RNA may also detect the well-characterized telomeric transcript TERRA[@b18][@b19], C-rich telomere repeat-containing RNA has not been characterized in mammals. To generate additional evidence of the telomere transcription induced on telomere deprotection by an independent approach, and to dissect the sub-cellular localization of telomeric dilncRNAs, we employed single-molecule fluorescence *in situ* hybridization (smFISH)[@b26]. To visualize telomeres in *Trf2*^*F/F*^ MEFs, we stably expressed a green fluorescent protein (GFP)-tagged version of the telomeric protein TRF1, which co-localized with \u03b3H2AX foci on telomere deprotection ([Supplementary Fig. 2c,d](#S1){ref-type=\"supplementary-material\"}). By using strand-specific DNA probes, nuclear focal signals of RNA transcripts with telomeric sequence were detected. Strikingly, on telomere deprotection, we observed a marked increase in the number and the intensity of such nuclear signals for both G- and C-rich telomeric RNA, which often co-localized with GFP-TRF1 signals, and much less with centromeres used as a negative control ([Fig. 1d,e](#f1){ref-type=\"fig\"}; [Supplementary Fig. 2e](#S1){ref-type=\"supplementary-material\"}). In control cells expressing TRF2, a weaker signal was detected only by the probe against teloG, which likely detects also TERRA transcripts, while none was detected by probing for the complementary strand. The observed signals were specific for RNA, since they were lost when RNase A was added before hybridization ([Supplementary Fig. 2f](#S1){ref-type=\"supplementary-material\"}). This result further strengthens the evidence of transcriptional induction of both DNA strands at deprotected telomeres.\n\nDDR at deprotected telomeres is DICER and DROSHA dependent\n----------------------------------------------------------\n\nDICER and DROSHA have been shown to be involved in DDRNA biogenesis on DSB generation[@b21]. We therefore investigated their role in the processing of telomeric transcripts, and consequently in DDR activation at dysfunctional telomeres, by individually knocking down each of them in *Trf2*^*F/F*^ MEFs ([Supplementary Fig. 2g](#S1){ref-type=\"supplementary-material\"}). We observed that silencing of DICER or DROSHA fully abolished tDDRNA induction following telomere deprotection ([Fig. 2a](#f2){ref-type=\"fig\"}). Conversely, in telomere-deprotected cells, telomeric dilncRNAs increased significantly in *Drosha* knocked-down cells ([Fig. 2b](#f2){ref-type=\"fig\"}). This accumulation is consistent with dilncRNAs being DDRNA precursors, which are then processed by DROSHA---an increase on *Dicer* knockdown could not be appreciated likely because *Dicer* knockdown leads to the accumulation of RNA intermediates too short to be detected by qPCR in this experimental setting.\n\nNext, we aimed to study the impact of the DDRNA biogenesis inhibition on DDR activation through *Dicer* or *Drosha* knockdown. In MEFs exposed to ionizing radiation (IR), as reported in other cell lines[@b21], *Dicer* and *Drosha* knockdown impaired 53BP1 and pATM foci formation, while \u03b3H2AX foci remained unchanged ([Supplementary Fig. 2h,i](#S1){ref-type=\"supplementary-material\"}). This was not due to altered 53BP1 or ATM protein levels ([Supplementary Fig. 3a,b](#S1){ref-type=\"supplementary-material\"}), consistent with the reported unaltered levels of chromosomal integrity genes and telomere maintenance factors observed in conditional knockout mouse models for *Dicer* or *Dgcr8*, a *Drosha* cofactor[@b27].\n\nWe then tested whether DICER and DROSHA have a role in DDR activation at dysfunctional telomeres. As expected, DICER or DROSHA loss did not affect \u03b3H2AX foci at deprotected telomeres ([Fig. 2c,d](#f2){ref-type=\"fig\"}). When we studied the activation of the DDR apical kinase ATM, we observed that the fraction of cells containing pATM foci, as well as foci of proteins phosphorylated by PI3-like kinases, including ATM on their consensus target sequence (pS/TQ), were strongly reduced on *Dicer* and *Drosha* knockdown ([Fig. 2c,d](#f2){ref-type=\"fig\"}). 53BP1 foci formation was not detectably affected on telomere deprotection in *Dicer* and *Drosha* knocked-down cells ([Fig. 2c,d](#f2){ref-type=\"fig\"}); this is consistent with the notion that 53BP1 recruitment to DDR foci is controlled by redundant mechanisms[@b28] and is most dependent on DICER and DROSHA only at early time points (10\u2009min) following DNA damage generation[@b21], while the genetic deletion of TRF2 does not allow such early time points studies. Unbiased identification and quantification of DDR foci by an automatic imaging software[@b29] confirmed that in both DICER and DROSHA-depleted cells, the number, as well as the intensity, of pATM and pS/TQ foci per cell were reduced, whereas \u03b3H2AX and 53BP1 foci were unaffected ([Supplementary Fig. 3c](#S1){ref-type=\"supplementary-material\"}). Furthermore, in human T19 cells, both *Dicer* and *Drosha* knockdown reduced the recruitment of pATM and ATM kinase targets pS/TQ, to deprotected telomeres, but not \u03b3H2AX foci formation ([Supplementary Fig. 3d,e](#S1){ref-type=\"supplementary-material\"}). Therefore, in both mouse and human cells the stable recruitment of activated ATM and its phosphorylated targets to deprotected telomeres requires DICER and DROSHA.\n\nDICER-dependent small ncRNAs generated at sites of DNA damage, damage-induced RNAs, have been proposed to be necessary for efficient repair by HR in *Arabidopsis thaliana* and human cells[@b7][@b8]. While DICER seems dispensable for NHEJ DNA repair in a human cell-based reporter system[@b8], Dicer-like 3 has been involved in NHEJ in plants[@b9]. More recently, error-free DSBs repair by NHEJ have been shown to involve RNA transcripts[@b30], while transient DNA--RNA hybrids seem to be necessary for HR in fission yeast[@b31]. Chromosomal fusions on TRF2 depletion are DNA repair events mediated by NHEJ[@b32]. We therefore tested whether DICER or DROSHA have a role in NHEJ events at dysfunctional telomeres. Following telomere deprotection, siControl-transfected MEFs showed massive chromosomal fusions, as expected ([Supplementary Fig. 4a](#S1){ref-type=\"supplementary-material\"}; ref. [@b33]). Instead, *Dicer* and *Drosha* knockdown reduced the number of telomere fusions, reaching statistical difference in the case of DROSHA inactivation ([Supplementary Fig. 4a,b](#S1){ref-type=\"supplementary-material\"}). These results indicate that DICER and DROSHA, and likely their DDRNAs products, are involved in NHEJ DNA repair at dysfunctional telomeres in mammalian cells.\n\nDDR at deprotected telomeres is mediated by tDDRNAs\n---------------------------------------------------\n\nThus far, the demonstrated effects of DICER and DROSHA on DDR activation at deprotected telomeres do not exclude the possibility that such effects may be indirect, meaning mediated by the role of DICER and DROSHA in microRNAs (miRNAs) biogenesis and gene expression. To address this issue, we used an experimental set-up, in which we previously demonstrated[@b21] that the impact on DDR activation of DICER- and DROSHA-generated DDRNAs can be studied independently from the canonical translational effects of miRNAs. We thus used a mild detergent to transiently permeabilize cell membranes of *Trf2*^*F/F*^ MEFs, in which DDR was triggered by IR or by *Trf2* genetic ablation and treated them with RNase A, or bovine serum albumin (BSA) as control. Both IR- and telomere deprotection-induced 53BP1 foci were sensitive to RNase A ([Fig. 3a,b](#f3){ref-type=\"fig\"}; [Supplementary Fig. 4c,d](#S1){ref-type=\"supplementary-material\"}), despite unaltered 53BP1 protein levels ([Supplementary Fig. 4e,f](#S1){ref-type=\"supplementary-material\"}), while \u03b3H2AX foci were not affected. Thus, also at deprotected telomeres RNA is necessary for DDR foci maintenance. To test if DDR foci can reform in an RNA-dependent manner, permeabilised RNase A-treated, and telomere-deprotected, MEFs were incubated with RNA extracted from MEFs with either normal (uninduced) or deprotected (induced) telomeres, or yeast transfer RNA (tRNA) as control. We observed that only RNA from induced MEFs allowed the robust reformation of 53BP1 foci in RNase A-treated cells ([Fig. 3c](#f3){ref-type=\"fig\"}). We then characterized the RNA species required for DDR foci reformation and observed that RNA extracted from induced cells knocked down for *Dicer* or *Drosha* was unable to restore 53BP1 foci ([Fig. 3d](#f3){ref-type=\"fig\"}). This indicates that the RNA species responsible for DDR foci maintenance at dysfunctional telomeres are DICER and DROSHA products. To further demonstrate that telomeric DICER products are indeed the RNA molecules necessary for DDR activation at telomeres and are sufficient to allow telomeric DDR foci reformation following RNase A treatment, we generated small double-stranded tDDRNAs by incubating a long double-stranded telomeric, or control, RNA substrate with purified recombinant DICER ([Supplementary Fig. 4g](#S1){ref-type=\"supplementary-material\"}). We then added these *in vitro*-generated telomeric and control DDRNAs to telomere-deprotected, RNase A-treated MEFs. We observed that only telomeric, but not control, DICER products could restore 53BP1 foci formation ([Fig. 3e](#f3){ref-type=\"fig\"}). Importantly, we can exclude that this effect was mediated by miRNA-like translational inhibition by DICER products, since messenger RNAs, the potential targets, were degraded by RNase A treatment and the effects were observed within minutes from RNA addition. To further prove the crucial dependency of DDR signalling at telomeres on tDDRNAs, we chemically synthetized 21-nucleotide long dsRNAs with a telomeric or a control sequence and tested their activity in RNase A-treated *Trf2*-deleted MEFs. We observed that the addition of synthetic tDDRNAs was sufficient to allow the reformation of robust DDR signalling events, as indicated by the reformation of 53BP1 foci, while DDRNAs with a control sequence had no effect ([Fig. 3f](#f3){ref-type=\"fig\"}). Consistent with a role of tDDRNAs in 53BP1 foci reformation at dysfunctional telomeres, tDDRNAs localized to dysfunctional telomeres, as measured by proximity ligation assay (PLA) between GFP-tagged TRF1 and Alexa 647-conjugated to tDDRNAs, or control DDRNAs that localized to telomeres much less ([Fig. 3g](#f3){ref-type=\"fig\"}; [Supplementary Fig. 4h](#S1){ref-type=\"supplementary-material\"}). In summary, these results demonstrate that tDDRNAs have a crucial role in DDR activation at dysfunctional telomeres.\n\nAntisense oligonucleotides allow DDR inhibition at telomeres\n------------------------------------------------------------\n\nThe results described so far consistently support the notion of a crucial role of tDDRNAs in DDR signalling at deprotected telomeres. On the grounds of the above conclusions, we next tested whether DDR at telomeres could be affected by the direct inhibition of tDDRNA functions. Inhibitory antisense oligonucleotides (ASOs) are widely used to block the functions of RNA molecules, including ncRNAs and small RNAs such as miRNAs[@b34] by steric hindrance. We therefore designed ASOs complementary to either tDDRNAs (anti-teloG and anti-teloC) or an ASO with an unrelated control sequence. We observed no obvious increase in \u03b3H2AX, 53BP1, pATM and pS/TQ foci in wild-type MEFs transfected with any of the ASOs used, ruling out the possibility that ASOs could compete with telomeres for the binding of telomeric proteins ([Supplementary Fig. 5a,b](#S1){ref-type=\"supplementary-material\"}). ASOs were then transfected in MEFs before telomere deprotection, or exposure to IR as a source of non sequence-specific DNA damage, and DDR activation was monitored. ASO treatment did not have any impact on IR-induced DDR foci ([Supplementary Fig. 6a](#S1){ref-type=\"supplementary-material\"}). Differently, both anti-teloG and anti-teloC, but not the control ASO, reduced both the number and the intensity of 53BP1 and pS/TQ foci at deprotected telomeres ([Fig. 4a,b](#f4){ref-type=\"fig\"}). Similar to *Dicer* and *Drosha* knockdown, and RNase A treatment, no effect on \u03b3H2AX foci was observed.\n\nTo test the direct role of tDDRNAs on chromosomal fusions following TRF2 depletion, we inhibited them by ASOs in cultured MEFs and monitored fusion events. We observed a significant reduction in the frequency of chromosomal fusions in cells treated with both anti-teloG and anti-teloC ASOs, but not with the control ASO ([Supplementary Fig. 6b](#S1){ref-type=\"supplementary-material\"}).\n\nThese results demonstrate, for the first time, that DDR activation and NHEJ events at dysfunctional telomeres can be modulated in living cells by ASOs.\n\nTelomeric DDR inhibition *in vivo*\n----------------------------------\n\nASOs have been successfully used to treat a number of different diseases; many of them are being tested in clinical trials and some have already been approved for use in the clinics[@b35][@b36]. To validate the efficacy of anti-teloG and anti-teloC ASOs in a living organism, we used an inducible *Trf2* knockout mouse model in which, following tamoxifen administration, TRF2 expression is lost, leading to telomere deprotection and a very robust telomeric DDR activation ([Fig. 5a,b](#f5){ref-type=\"fig\"}). After tamoxifen administration, but before detectable DDR induction, mice received a systemic dose of anti-teloG or anti-teloC ASOs, or phosphate-buffered saline (PBS) as control, by intraperitoneal (i.p.) injection ([Supplementary Fig. 6c](#S1){ref-type=\"supplementary-material\"}). Four days later, when telomeric DDR was expected to reach its maximal activation, organs were collected and tissue sections studied for DDR activation by immunofluorescence. We observed a marked reduction in foci intensity of both 53BP1 and the ATM target pKAP1 (ref. [@b37]) in mice treated with both ASOs against tDDRNAs, compared with the control-treated mice, while \u03b3H2AX foci were unaffected. This striking effect could be observed in both organs studied: liver ([Fig. 5a,b](#f5){ref-type=\"fig\"}) and kidney ([Fig. 5c,d](#f5){ref-type=\"fig\"}). In summary, these unprecedented results demonstrate that sequence-specific targeting of tDDRNAs by ASOs allows telomeric DDR inhibition in both cultured cells and living mammals.\n\nDiscussion\n==========\n\nOur results show that dysfunctional telomeres are actively transcribed to generate double-stranded telomeric dilncRNAs and DDRNAs. While G-rich transcripts (TERRA) have previously been widely studied, this is the first report on the generation, processing and function of C-rich telomeric transcripts in mammals to our knowledge. Very low-abundant signals were detected for C-rich telomeric RNAs also in ref. [@b18]. The observation that *Schizosaccharomyces pombe* strains deleted for telomeric proteins Taz1 or Rap1 (yeast homologues of the mammalian shelterin components TRF1/2 and RAP1, respectively) also show an induction of C-rich transcripts[@b38][@b39] suggests that the observed induction is evolutionarily conserved. These results are consistent with RNA polymerase II recruitment at DSBs sites in fission yeast[@b31], and with the reported increased RNA polymerase II occupancy at telomeres on TRF2 downregulation in human cells[@b40].\n\nThe mechanism by which the proteins involved in the DDR are recruited to dysfunctional telomeres in a tDDRNA-dependent manner remains to be further investigated, although the ability of 53BP1 to bind RNA through its Tudor domain[@b41] can be involved.\n\nThe evidence that only DDRNAs bearing a telomeric sequence are able to restore DDR activation at telomeres following RNase A treatment, and that only ASOs against tDDRNAs inhibit DDR at telomeres, but not at DSBs randomly generated at various genomic locations by IR, strongly suggests that tDDRNAs act in a locus- and sequence-specific manner.\n\nFinally, we propose that the ability to detect tDDRNAs and to inhibit telomeric transcripts *in vitro* and *in vivo* by ASOs may have important implications in human physiology and pathology. tDDRNAs may be important biomarkers of DDR activation on telomeric damage, an event associated with physiological ageing and cancer[@b11][@b12][@b13][@b14][@b15][@b16]. In addition, many medical conditions, collectively known as telomere syndromes, and some progeric syndromes, are caused by accelerated telomere shortening and dysfunction with consequent DDR activation[@b17]. It is therefore conceivable that tDDRNA detection and modulation, for instance by ASO inhibition, can be exploited to address pathological conditions caused by telomere dysfunction.\n\nMethods\n=======\n\nCell culture\n------------\n\nRosa26-CreERT2 *Trf2*^*F/F*^ and *Trf2*^*F/+*^ (ref. [@b20]) MEFs, a gift from Eros Lazzerini Denchi (The Scripps Research Institute, La Jolla, USA), were grown in DMEM supplemented with 10% fetal bovine serum and 1% glutamine; for CreER activation, cells were treated with of 4OHT (600\u2009nM) and analysed 48--72\u2009h later. MEFs *Rosa26-CreERT2 Trf2*^*F/F*^ eGFP-TRF1 were obtained by retroviral infection of *Rosa26-CreERT2 Trf2*^*F/F*^ using eGFP-TRF1 pWzl-Hygro plasmid, a gift from Titia de Lange (Addgene plasmid \\# 19834), and selected with Hygromycin (200\u2009\u03bcg\u2009ml^\u22121^). T19 fibrosarcoma cells[@b24], a gift from Titia de Lange (The Rockefeller University, New York, USA), were grown in DMEM supplemented with 10% fetal bovine serum, 1% glutamine and doxycycline (100\u2009ng ml^\u22121^). For induction, cells were cultured without doxycycline for 7--8 days. All cell lines used in this study were negative for mycoplasma contaminations.\n\nAnimals and treatments\n----------------------\n\nRosa26-CreERT mice (Jackson Laboratories) and *Trf2* conditional knockout mice[@b33] and mice carrying a p53 conditional allele (Jackson Laboratories) were crossed to generate Trf2/p53/Rosa26 mice. Mice were maintained in 129/c57Bl6 genetic background. All mice were bred and maintained under pathogen-free condition at the Scripps Research Institute, La Jolla, USA, and were handled according to Institutional Animal Care and Use Committee guidelines. To activate CreER, 8--10-week-old mice (both males and females) were injected i.p. with tamoxifen dissolved in sunflower oil or with vehicle at a final concentration of 75\u2009mg\u2009kg^\u22121^. After 24\u2009h, ASOs dissolved in PBS were administrated by i.p. at concentration of 15\u2009mg\u2009kg^\u22121^. Mice were sacrificed after 5 days post tamoxifen injection. Tissues were collected and frozen in dry ice and embedded in optimal cutting temperature (OCT) tissue TEC (Sakura).\n\nNo power analysis was done to choose the size of the samples. No specific method of randomization was used to assign groups. Animals were assigned to experimental groups so as to minimize the influence of other variables such as age or sex on the outcome. No blinding approach was used for the experiments.\n\nRNA isolation\n-------------\n\nTotal RNA from cultured cells was extracted with RNeasy Mini Kit (Qiagen) for messenger RNA detection, or with mirVana miRNA Isolation kit (Life Technologies) for DDRNA and dilncRNA detection, according to the manufacturer's instructions. RNA extracted with mirVana miRNA Isolation kit using the Enrichment Procedure for Small RNAs was used as starting material for Target Enrichment experiments.\n\nReal-time quantitative PCR\n--------------------------\n\nA measure of 1\u2009\u03bcg of total cell RNA was reverse transcribed using SuperScript VILO cDNA Synthesis Kit. A volume corresponding to 10\u2009ng of initial RNA was used for each qPCR reaction using GoTaq qPCR Master Mix (Promega) on a Roche LightCycler 480 sequence detection system. Each reaction was performed in triplicate. Human ribosomal protein large P0 (Rplp0) and mouse beta-2 microglobulin (B2M) were used as control transcripts for normalization. Primers sequences (5\u2032--3\u2032 orientation) were:\n\nmDicer Fw GCAAGGAATGGACTCTGAGC\n\nmDicer Rv GGGGACTTCGATATCCTCTTC\n\nmDrosha Fw CGTCTCTAGAAAGGTCCTACAAGAA\n\nmDrosha Rv GGCTCAGGAGCAACTGGTAA\n\nmB2M Fw CTGCAGAGTTAAGCATGCCAGTA\n\nmB2M Rv TCACATGTCTCGATCCCAGTAGA\n\nhDicer Fw GCAAAGCAGGGCTTTTCAT\n\nhDicer Rv AGCAACACAGAGATCTCAAACATT\n\nhDrosha Fw TGCACACGTCTAACTCTTCCAC\n\nhDrosha Rv GGCCCGAGAGCCTTTTATAG\n\nhRplp0 Fw TTCATTGTGGGAGCAGAC\n\nhRplp0 Rv CAGCAGTTTCTCCAGAGC\n\nReal-time quantitative PCR for small RNAs\n-----------------------------------------\n\nA measure of 5\u2009\u03bcg of total cellular RNA were fractionated on a 10% polyacrylamide, 7\u2009M Urea gel and RNA species shorter than 40 nucleotides were gel-extracted. A measure of 10\u2009pg of an RNA spike-in was added to all samples before cellular RNA was loaded on the gel, to normalize for the efficiency of RNA extraction from gel. complementary DNA (cDNA) was synthesized using the miScript II RT kit (Qiagen) with HiSpec buffer. PCR was performed using the miScript PCR system (Qiagen) accordingly to the manufacturer's instructions. Each reaction was performed in triplicate. mir29b was used as a control gene for normalization. Primer sequences (5\u2032--3\u2032 orientation):\n\nmir29b TAGCACCATTTGAAATCAGTGTT\n\nspike-In CGAATTCCACAAATTGTTATCC\n\nteloG TAGGGTTAGGGTTAGGGT\n\nteloC CCCTAACCCTAACCCTAA\n\nStrand-specific real-time quantitative PCR\n------------------------------------------\n\nSamples were treated with DNase I (Thermo Scientific) at 37\u2009\u00b0C for 1\u2009h to remove any potential residual genomic DNA contamination (one unit of DNase I per 1\u2009\u03bcg of RNA). A measure of 1\u2009\u03bcg of total RNA was reverse-transcribed using the Superscript First Strand cDNA synthesis kit (Invitrogen) with strand-specific primers. To amplify telomeric repeats, we adapted a technique described in ref. [@b25], which allows the generation of a fixed-length amplification product. For reverse transcription, we used the following primers: RPP0 Rv for the detection of Rplp0 messenger RNA; teloC Rv for the detection of G-rich telomeric precursor; and teloG Rv for the detection of C-rich telomeric precursor. qPCR was performed using SYBR green (Roche). For each reaction, 50\u2009ng of cDNA were used. Each reaction was performed in triplicate. Rplp0 was used as a control gene for normalization. Primer sequences (5\u2032--3\u2032 orientation) were:\n\nRPP0 Fw TTCATTGTGGGAGCAGAC\n\nRPP0 Rv CAGCAGTTTCTCCAGAGC\n\nteloC Rv CCCTAACCCTAACCCTAA\n\nteloG Rv TAGGGTTAGGGTTAGGG\n\ntelo Fw CGGTTTGTTTGGGTTTGGGTTTGGGTTTGGGTTTGGGTT\n\ntelo Rv GGCTTGCCTTACCCTTACCCTTACCC TTACCCTTACCCT\n\nTargeted sequencing of small RNA\n--------------------------------\n\nTwo linkers were ligated to the two ends of the RNA molecules in the sample to be analysed. The 3\u2032-end of the starting RNA was ligated to a monoadenylated DNA linker by a T4 RNA ligase 2-truncated enzyme (NEB) incubated for 1\u2009h at 25\u2009\u00b0C. The 5\u2032 RNA linker was then ligated by a T4 RNA ligase 1 (NEB) to the target RNA at 20\u2009\u00b0C for 1\u2009h, after removing the 5\u2032 cap structure by Tobacco Acid Pyrophosphatase (Epicentre), incubated at 37\u2009\u00b0C for 1\u2009h. Linkers enabled cDNA synthesis using PrimeScript RT--PCR Kit (Takara). The reverse transcription reaction was incubated at 44\u2009\u00b0C for 1\u2009h. Subsequent PCR amplification using Phusion High-Fidelity DNA Polymerase (NEB) was carried out as follows: 98\u2009\u00b0C 2\u2009min; 22 cycles of: 98\u2009\u00b0C for 30\u2009s, 55\u2009\u00b0C for 30\u2009s, 72\u2009\u00b0C for 30\u2009s; 72\u2009\u00b0C for 5\u2009min; hold at 4\u2009\u00b0C. To capture the amplified cDNA targets, complementary RNA baits containing biotin-labelled nucleotides were used. These RNA baits were produced by using AMbion MAXIscript T7 *In Vitro* Transcription kit (Life Technologies) and Biotin RNA labelling Mix (Roche). A T7-promoter-containing double-stranded DNA was incubated at 37\u2009\u00b0C for 1\u2009h to allow for *in vitro* transcription. RNA baits and cDNA targets were incubated at 37\u2009\u00b0C for 48\u2009h in the presence of SUPERase-inhibitor (Life Technologies) and of the following blocking agents: Human Cot-1 (Life Technologies), UltraPure Salmon Sperm DNA Solution (Thermo Scientific) and a 200\u2009\u03bcM Customized Block. The hybrid RNA--cDNA molecules were captured by Dynabeads MyOne Streptavidin C1 (Life Technologies) beads, while non-targeted cDNAs were washed away. Captured cDNAs were then barcoded by PCR with Script Index PCR primers (Illumina) and sequenced by a MiSeq (Illumina) sequencer. Oligonucleotide sequences (5\u2032--3\u2032 orientation) were:\n\n3\u2032-DNA linker AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC-Amine\n\n5\u2032-RNA linker ACACUCUUUCCCUACACGACG CUCUUCCGAUCU\n\nRT primer GTGACTGGAGTTCAGACGTGTG CTCTTCCGATCT\n\nPCR Fw AATGATACGGCGACCACCGAGA TCTACACTCTTTCCCTACACGACGCTCTTCCGATCT\n\nPCR Rv CAAGCAGAAGACGGCATACGAGATCGGTCTCGGCATTCCTGCTGAACCGCTCTTCCGATCT\n\nBlock Fw AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT\n\nBlock Rv CAAGCAGAAGACGGCATACGAGATCGTGATGTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT\n\nAnalysis of small RNA sequencing data\n-------------------------------------\n\nIllumina adapters and linkers used in the targeted sequencing of small RNA were trimmed using Trim Galore! (<http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/>), discarding reads that became shorter than 10 nucleotides of length after trimming. Reads were mapped to a telomeric contig made of 15 consecutive TTAGGG repeats and in parallel to the mouse genome (GRCm38/mm10) using Bowtie2 (ref. [@b42]) with a very sensitive local option (-D 20 -R 3 -N 0 -L 20 -i S,1,0.50). Strand information was preserved to divide the aligned reads according to the original strand. Reads were parsed using SAMtools[@b43] and *ad hoc* bash and perl scripts. To avoid spurious reads contamination, telomeric reads with mismatches or soft clipped portions were removed with *ad hoc* R scripts.\n\nRNA smFISH\n----------\n\nCells were fixed in 4% paraformaldehyde (PFA) for 10\u2009min at room temperature (RT) and permeablized overnight at 4\u2009\u00b0C using 70% ethanol. Cells were rehydrated in a solution containing 20% formamide and 2 \u00d7 SSC for 5\u2009min and then treated with 10\u2009nM FISH probes for 12\u2009h in 2 \u00d7 SSC containing 10% dextran sulfate, 2\u2009mM vanadyl--ribonucleoside complex, 0.02% RNase-free BSA, 1\u2009\u03bcg\u2009\u03bcl^\u22121^ *E. coli* tRNA and 20% formamide at 37\u2009\u00b0C. After hybridization, the cells were washed twice for 30\u2009min at 37\u2009\u00b0C using a wash buffer (20% formamide in 2 \u00d7 SSC). Cells were then mounted in solution containing 10\u2009mM Tris/HCl, 2 \u00d7 SSC, 2\u2009mM trolox, 5\u2009mM protocatechiuc acid and 50\u2009nM protocatechuate dehydrogenase. HILO imaging was done, as described[@b26]. FISH signals were detected using custom-written macros in ImageJ that were based on the analysis routine as described[@b44]. To calculate the relative intensity of FISH spots, a 21 \u00d7 21 pixel area around the FISH signal locus was normalized with the intensity of a 21 \u00d7 21 pixel area outside of the locus. Intensity of 21 \u00d7 21 pixel area outside the cell was used for background subtraction. Red and green channels were corrected for signal bleed-through--samples that were labelled only with Cy5 or GFP were imaged using all excitation lines (488 and 640\u2009nm) and signal arising from non-cognate excitation were used as a correction factor. All probes were tested for RNA specificity by visualizing the loss of FISH signal on RNase A treatment. For these control experiments cells were treated with RNase A (1\u2009\u03bcg\u2009\u03bcl^\u22121^) in DPBS at 37\u2009\u00b0C, before probe treatment. FISH probes were purchased as high-performance liquid chromatography purified DNA oligonucleotides from IDT. Probe sequences were as follows (5\u2032--3\u2032 orientation):\n\nTeloG: CCTAACCCTAACCCTAACCCTAAC-Cy5\n\nTeloC: GGTTAGGGTTAGGGTTAGGGTTAG-Cy5\n\nCombined IF/RNA smFISH\n----------------------\n\nMEFs *Rosa26-CreERT2 Trf2*^*F/F*^ eGFP-TRF1 were fixed in 4% PFA for 10\u2009min and treated with a solution containing 1:1 methanol:acetic acid to irreversibly photobleach eGFP-TRF1. Cells were then washed three times in PBS for 5\u2009min each and the standard immunofluorescence protocol was continued. In addition, all blocking, primary and secondary antibody incubations were performed in 1 \u00d7 PBBR (1% Ultrapure/RNase-free BSA+0.1\u2009U\u2009\u03bcl^\u22121^ rRNAsin, 1\u2009mM EDTA in PBS). Importantly, the Human-anti-centromere antibody was purified by passing through G25 spin column (Millipore), which was pre-equilibrated in 1 \u00d7 PBS with 5% glycerol and 0.1\u2009U\u2009\u03bcl^\u22121^ rRNAsin. Cells were fixed again in in 4% PFA (in 1 \u00d7 PBS) for 10\u2009min after immunofluorescence protocol to stably retain the bound primary and secondary antibodies and the standard RNA smFISH protocol was continued.\n\nIonizing radiation\n------------------\n\nIR (1\u2009Gy) was generated by a high-voltage X-ray-generator tube (Faxitron X-Ray Corporation).\n\nTransfection\n------------\n\nTransfections were carried out with Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions.\n\nsiRNA\n-----\n\nON-TARGETplus SMARTpool short interfering RNA (siRNA) oligonucleotides (Dharmacon) were used at a final concentration of 5--20\u2009nM. Sequences were as follows:\n\nsiControl (siGFP) GCAAGCUGACCCUGAAGUUCAU\n\nsiDicer human UAAAGUAGCUGGAAUGAUG;\n\nGGAAGAGGCUGACUAUGAA; GAAUAUCGAUCCUAUGUUC;\n\nGAUCCUAUGUUCAAUCUAA\n\nsiDrosha human CAACAUAGACUACACGAUU;\n\nCCAACUCCCUCGAGGAUUA; GGCCAACUGUUAUAGAAUA;\n\nGAGUAGGCUUCGUGACUUA\n\nsiDicer1 mouse GGUAGACUGUGGACCGUUU;\n\nGGAAAUACCUGUACAACCA; GCAAUUUGGUGGUUCGUUU;\n\nACAGGAAUCAGGAUAAUUA\n\nsiDrosha mouse UGGAAGGAGUUACGCUUUA;\n\nGGAAUCCGCCACAGCAUUU; GUGAUCACUUUCCCGAUUA;\n\nUAAUGCACCUGGACAAGUU\n\nASOs sequences\n--------------\n\nThe locked nucleic acid mixmer oligonucleotides with a fully phosphorothioate backbone were produced by Exiqon. They were used at a final concentration of 20\u2009nM for transfection of cultured cells, 15\u2009mg\u2009kg^\u22121^ for mouse injections. Sequences were as follows (5\u2032--3\u2032 orientation):\n\nControl ACTGATAGGGAGTGGTAAACT\n\nanti-teloG CCTAACCCTAACCCTAACCC\n\nanti-teloC GGGTTAGGGTTAGGGTTAGGG\n\nImmunoblot\n----------\n\nCells were lysed in Laemmli sample buffer (2% SDS, 10% glycerol, 60\u2009mM Tris-HCl pH 6.8). A measure of 50\u2009\u03bcg of whole cell extracts were resolved by SDS--polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose membrane, which was blocked in 5% milk in TBST (0.1% Tween in tris-buffered saline), and incubated with the primary antibody for 1\u2009h at RT, and with a horseradish peroxidase-conjugated secondary antibody for 1\u2009h at RT. Quantification of protein bands was done by ImageJ software, subtracting the background signal and normalizing for the housekeeper. Uncropped scans of the blots are provided in [Supplementary Fig. 7](#S1){ref-type=\"supplementary-material\"}.\n\nImmunofluorescence for cultured cells\n-------------------------------------\n\nCells were fixed with 4% PFA or 1:1 methanol/acetone solution. After incubation with blocking solution, cells were stained with primary antibody for 1\u2009h at RT, washed and incubated with secondary antibodies for 40\u2009min at RT. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI; 1\u2009\u03bcg\u2009ml^\u22121^). Samples were mounted in mowiol or in imaging solution as in RNA smFISH.\n\nImmunofluorescence for mouse tissues\n------------------------------------\n\nA measure of 4-\u03bcm tissue sections were fixed for 10\u2009min in 4% PFA, incubated in blocking solution (2% BSA, 0.1% Tween in PBS) for 1\u2009hour at RT. Then sections were incubated for 1\u2009h at RT with primary antibodies, washed in blocking solution and incubated for 1\u2009h at RT with secondary antibody. Nuclei were stained with DAPI (1\u2009\u03bcg\u2009ml^\u22121^). Samples were mounted with glycerol solution.\n\nMetaphase spreads\n-----------------\n\nMEFs were incubated for 2\u2009h with 0.2\u2009\u03bcg\u2009ml^\u22121^ colcemid (Gibco). The cells were collected by trypsinization, resuspended in 0.075\u2009M KCl at 37\u2009\u00b0C for 30\u2009min, and fixed overnight in methanol/acetic acid (3:1) at 4\u2009\u00b0C. The cells were dropped onto glass slides and the slides were dried overnight. The next day, the slides were rehydrated with PBS for 5\u2009min. Slides were incubated consecutively with 75, 95, and 100% ethanol and allowed to air dry before FISH.\n\nDNA FISH\n--------\n\nCultured cells after immunofluorescence were fixed with 4% PFA, 0.1% Triton X-100 for 10\u2009min at RT, then incubated with 10\u2009mM glycine for 30\u2009min at RT. Interphase or metaphase cells were incubated with a telomeric PNA probe (0.5\u2009\u03bcM, TelC-Cy3 from PANAGENE, catalogue number: F1002-5) in a buffer containing 70% formamide, 1\u2009mg\u2009ml^\u22121^ blocking reagent (Roche), 10\u2009mM Tris-HCl pH 7.4 and the coverslips were denatured on a heat block 5\u2009min at 80\u2009\u00b0C and incubated for 2\u2009h in the dark. The coverslips were washed twice with 70% formamide, 10\u2009mM Tris-HCl pH 7.4 for 15\u2009min each and three times with 100\u2009mM Tris-HCl pH 7.4, 0.15\u2009M NaCl, 0.08% Tween 20 for 5\u2009min each. DNA was stained with DAPI (1\u2009\u03bcg\u2009ml^\u22121^). Samples were mounted in mowiol.\n\n*In situ* PLA\n-------------\n\nCells were labelled according to the manufacturer\\'s instructions (Sigma). Briefly, MEFs *Rosa26-CreERT2 Trf2*^*F/F*^ eGFP-TRF1 were treated with 4OHT (600\u2009nM). Forty-eight hours later cells were permeabilized with 0.6% Tween 20 in PBS for 20\u2009min at RT and incubated with 7\u2009ng\u2009\u03bcl^\u22121^ of Alexa Fluor 647-conjugated synthetic DDRNAs, in the presence of the RNase inhibitor RNaseOUT (1\u2009U\u2009\u03bcl^\u22121^, Invitrogen) for 25\u2009min at RT. Cells were then fixed with 4% PFA. After incubation with primary antibodies, appropriate PLA probes (secondary antibodies conjugated with oligonucleotides) were added to the samples. Ligation of the oligonucleotide probes that were in proximity (\\<40\u2009nm) was then performed, after which fluorescently labelled oligonucleotides were added, together with a DNA polymerase to generate a signal detectable by a fluorescence microscope.\n\nImaging\n-------\n\nImages were acquired using a widefield Olympus Biosystems BX71 microscope or a Leica TCS SP2 AOBS confocal laser microscope. Number and intensity of foci per cell and number of dots per cell were analysed by the imaging software CellProfiler[@b29], using the same pipeline for each sample in the same experiment. Percentages of DDR-positive cells were scored manually. A cell was counted as positive if showing \\>3 foci.\n\nAntibodies\n----------\n\nAnti-\u03b3H2AX (Millipore, 05-636, 1:400; Cell Signalling, 9718, 1:1,000); anti-ATM pS1981 (Millipore, 05-740, 1:100); anti-pS/TQ (Cell Signalling, 2851, 1:200); anti-53BP1 (Novus, NB100-304, 1:200; Bethyl, A300-272A, 1:1,000; Santa Cruz, sc-22760, 1:1,000); anti-pKap1 (Bhetyl, A300-767A, 1:1,000); anti-Flag (Sigma, F3165, 1:500); anti-ATM (Sigma-Aldrich, A1106, 1:1,000); anti-H2AX (Abcam, ab11175, 1:1,000); anti-vinculin (Sigma-Aldrich, V9131, 1:10,000); anti-GFP (Abcam, ab290, 1:4,000); anti-Cy5 (Abcam, ab52061,1:500); anti-Centromere Protein (Antibodies Incorporated, 15-234-0001, 1:100).\n\nRNase A treatment on permeabilized living cells\n-----------------------------------------------\n\nCells were permeabilized with 0.6% Tween 20 in PBS for 20\u2009min at RT, and RNase A treatment was carried out in 1\u2009mg\u2009ml^\u22121^ of ribonuclease A from bovine pancreas (Sigma-Aldrich R5503), or acetylated albumin from bovine serum (Sigma-Aldrich B8894), in PBS for 30\u2009min at RT. For complementation experiments samples were washed with PBS, treated with RNase inhibitor RNaseOUT (1\u2009U\u2009\u03bcl^\u22121^, Invitrogen) and \u03b1-amanitin (20\u2009ng\u2009\u03bcl^\u22121^, Sigma-Aldrich A2263) for 15\u2009min. Cells were incubated with total cell RNA or yeast tRNA (3\u2009ng\u2009\u03bcl^\u22121^), rDICER products (1.5\u2009ng\u2009\u03bcl^\u22121^) or synthetic dsRNA oligos (1--100\u2009nM using yeast tRNA up to 100\u2009nM) for 25\u2009min at RT. Cells were then fixed with 4% PFA.\n\nTurbo DICER RNAs generation\n---------------------------\n\nDICER RNA products were generated as follows. pTH5 (ref. [@b45]), a pSP73-backboned plasmid containing 27 telomeric repeats flanked by T7 and SP6 promoter, was used for *in vitro* transcription. SP6 RNA Polymerase (Promega) and T7 RNA Polymerase (Promega) were used together for *in vitro* transcription allowing the generation of a telomeric dsRNA. The DNA template was later removed by using RQ1 RNase-Free DNase (Promega) for 15\u2009min at 37\u2009\u00b0C. Acid-phenol:chloroform, pH 4.5 (Ambion) and subsequent ethanol precipitation was used to isolate the *in vitro*-transcribed RNA. The dsRNA was then incubated for 8\u2009h at 37\u2009\u00b0C using a Recombinant Human Turbo Dicer Enzyme Kit (Genlantis). To remove salts and undigested templates, RNA Purification Column 1 and 2 (Genlantis) were used. RNA products were quantified and checked on a 3% agarose gel. As a control, the same procedure was followed with a 1\u2009kb construct containing the neomycin DNA sequence. Equal amounts of DICER RNA products generated in this way were used in complementation experiments with MEFs following RNase A treatment.\n\nSynthetic DDRNAs\n----------------\n\n3\u2032-labelled Alexa Fluor 647 RNA oligonucleotides were resuspended in 1 \u00d7 siRNA Buffer (Dharmacon). Equal amount of two oligonucleotides were mixed to generate double-stranded DDRNAs and treated as follows: 90\u2009\u00b0C for 1\u2009min, 70\u2009\u00b0C for 2\u2009min, 55\u2009\u00b0C for 2\u2009min, 37\u2009\u00b0C for 5\u2009min, RT for 5\u2009min. Sequences were as follows (5\u2032--3\u2032 orientation):\n\nControl pair UGUUAGCUGGAGUGAAAACUU\n\nGUUUUCACAAAGCUAACACA\n\nTelomeric pair AGGGUUAGGGUUAGGGUUAGG\n\nUAACCCUAACCCUAACCCUAA\n\nStatistical analysis\n--------------------\n\nResults are shown as mean\u00b1s.e.m. or s.d. or percentages\u00b195% confidence interval as indicated. *P* value was calculated by the indicated statistical tests, using Prism software. Statistical significance of enrichment of small RNA species on telomere deprotection was calculated using the Fisher's exact test. The differences in the fraction of 20--23 nucleotides versus total small RNAs were calculated by performing a one-tailed binomial test, where the null hypothesis is the fraction of 20--23 nucleotide reads in the control sample and the alternative hypothesis is the count of 20--23 nucleotide reads versus total small RNAs in the sample with telomere deprotection. The deviation from a theoretically expected distribution in a UUAGGG or CCCUAA context was calculated by performing a one-tailed binomial test, where the null hypothesis is the expected fraction of U, A, G (or A, U, C) nucleotides in the wild type (that is, 2/6, 1/6 and 3/6, respectively), and the alternative hypothesis is the observed proportion of each nucleotide.\n\nIn figure legends, *n* indicates the number of independent experiments.\n\nData availability\n-----------------\n\nThe authors declare that all data supporting the findings of this study are available within the article and its Supplementary files or from the authors on a reasonable request. The sequencing data refer to the targeted sequencing of small RNA ([Fig. 1b](#f1){ref-type=\"fig\"}; [Supplementary Figs 1g and 2a](#S1){ref-type=\"supplementary-material\"}). These data are accessible through GEO (Gene Expression Omnibus) Series accession number GSE86964.\n\nAdditional information\n======================\n\n**How to cite this article:** Rossiello, F. *et al*. DNA damage response inhibition at dysfunctional telomeres by modulation of telomeric DNA damage response RNAs. *Nat. Commun.* **8,** 13980 doi: 10.1038/ncomms13980 (2017).\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nSupplementary Figures\n\nWe thank Eros Lazzerini Denchi for sharing the mouse model and reagents, and for helpful discussions; Valentina Matti for providing technical help; Amanda Oldani for providing assistance in image analysis; Arul Chinnaiyan for support; Titia de Lange for sharing reagents; all F.d'A.d.F. laboratory members for discussions. F.R. is supported by Fondazione Italiana per la Ricerca sul Cancro (FIRC, application number 12476). J.A. is supported by Marie Curie Initial Training Networks (FP7 PEOPLE 2012 ITN (CodAge Project No: 316354)). S.S. is supported by SIPOD 2 (Structural International Post Doc Programm 2)---the People Programme (Marie Curie Actions) of the European Union\\'s Seventh Framework Programme FP7 under grant agreement n.600399. F.d'A.d.F.'s laboratory is supported by Associazione Italiana per la Ricerca sul Cancro, AIRC (application 12971), Human Frontier Science Program (contract RGP 0014/2012), Fondazione Telethon (GGP12059), and an European Research Council advanced grant (322726). P.C.'s lab is supported by a Research Grant from the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) to the RIKEN Center for Life Science Technologies. F.R. and F.d'A.d.F. are inventors on the following patent applications: WO2013/167744 A1 and PCT130108. J.A. is an inventor on the following patent application: PCT130108.\n\nThe authors declare no competing financial interests.\n\n**Author contributions** J.A. generated data in Figs 1a,c and 2a,b and Supplementary Figs 1c,f,2b,3a,b,d,e and 4g, and performed the targeted sequencing of small RNA experiments; S.S. generated data in Fig. 5 and Supplementary Fig. 6c; F.I. analysed the next-generation sequencing data shown in Fig. 1b, and Supplementary Figs 1g and 2a; Q.N. set-up the targeted small RNA sequencing protocol; S.P. generated data in Fig. 1d,e and Supplementary Fig. 2c--f; P.C. supervised Q.N.; F.R. generated data in all remaining figures and wrote the manuscript; F.d'A.d.F. planned and supervised the project and wrote the manuscript; all authors edited the manuscript.\n\n04/13/2017\n\nA correction has been published and is appended to both the HTML and PDF versions of this paper. The error has not been fixed in the paper.\n\n![Deprotection of telomeres leads to enhanced transcription of both telomere DNA strands.\\\n(**a**) Total cell RNA was isolated from MEFs of the indicated genotype and treated with 4-hydroxytamoxifen (4OHT). Gel-extracted small RNA fraction (\\<40 nucleotides) was used for miScript PCR amplification to specifically detect DDRNAs. Error bars represent the s.e.m. *n*=3 independent experiments. \\**P*\\<0.05, \\*\\**P*\\<0.01, Student's *t*-test. (**b**) Small RNA (\\<200 nucleotides) fractions were isolated from 4OHT-treated MEFs of the indicated genotype, enriched for species with telomeric sequences using a telomeric bait, and sequenced. Histograms show for each reported read length the number of telomeric reads, either G rich or C rich, normalized on mir29b reads. For both teloG and teloC, MEFs *Trf2*^*F/F*^ have a significantly higher proportion of small RNA with respect to mir29b reads than MEFs *Trf2*^*F/+*^. *P*\\<0.001, Fisher's exact test. (**c**) Total cell RNA was isolated from MEFs of the indicated genotype and used for strand-specific RT--qPCR to detect telomeric dilncRNAs. Error bars represent the s.e.m. *n*=3 independent experiments. \\**P*\\<0.05; \\*\\*\\**P*\\<0.001, Student's *t*-test. (**d**,**e**) MEFs *Trf2*^*F/F*^-expressing GFP-TRF1 were treated with vehicle or 4OHT and analysed 48\u2009h later. (**d**) Representative images of teloG or teloC dilncRNA transcripts by smFISH. Zoomed in view of the boxed regions (12.5 \u00d7 12.5\u2009\u03bcm^2^), with GFP-TRF1 and FISH signals, is shown on the right. The indicated numbers show the percentage of RNA signals co-localizing with GFP-TRF1\u00b1s.e.m. Scale bar, 5\u2009\u03bcm (**e**) Quantification of data presented in **d**, showing the number of spots per cell and the relative intensity (a.u.=arbitrary units). Lines depict the mean\u00b1s.e.m. *n*=2 independent experiments; at least 50 cells per sample have been analysed; \\*\\*\\**P*\\<0.001.](ncomms13980-f1){#f1}\n\n![DICER and DROSHA are involved in dilncRNA processing and DDRNA generation and necessary for full DDR activation at deprotected telomeres.\\\n(**a**--**b**) Total cell RNA was isolated from MEFs of the indicated genotype that were previously treated with 4-hydroxytamoxifen (4OHT) and transfected with the indicated siRNA. (**a**) Gel-extracted small RNA fraction (\\<40 nucleotides) was used for miScript PCR amplification to specifically detect DDRNAs. Error bars represent the s.e.m. *n*=3 independent experiments. \\**P*\\<0.05, Student's *t*-test. (**b**) Total RNA was used for strand-specific RT--qPCR. Error bars represent the s.e.m. *n*=3 independent experiments. \\**P*\\<0.05, Student's *t*-test. (**c**) MEFs *Trf2*^*F/F*^ were treated with 4OHT, transfected with the indicated siRNA and stained for the indicated DDR markers. Scale bar, 10\u2009\u03bcm. (**d**) Quantification of data presented in **c**. Bar graphs show the percentage of DDR-positive cells\u00b195% confidence interval *n*=3 independent experiments; at least 150 cells per sample have been analysed; \\*\\*\\**P*\\<0.001, *\u03c7*^2^-test.](ncomms13980-f2){#f2}\n\n![DDR foci formation at deprotected telomeres is DDRNA dependent.\\\n(**a**) MEFs *Trf2*^*F/F*^ were treated with 4-hydroxytamoxifen (4OHT), permeabilised, treated with RNase A, or BSA as control, and stained for the indicated DDR markers. Scale bar, 20\u2009\u03bcm. (**b**) Quantification of data presented in **a**. Bar graphs show the percentage of DDR-positive cells\u00b195% confidence interval (CI); *n*=3 independent experiments; at least 100 cells per sample have been analysed; \\*\\*\\**P*\\<0.001, *\u03c7*^2^-test. (**c**--**f**) After RNase A treatment, cells were incubated with: (**c**) cell RNA from MEFs *Trf2*^*F/F*^ with normal (uninduced) or deprotected (induced) telomeres, or tRNA as a control; (**d**) cell RNA from MEFs *Trf2*^*F/F*^ treated with 4OHT and transfected with the indicated siRNA; (**e**) small double-stranded RNAs generated by recombinant DICER; (**f**) synthetic double-stranded RNAs. Bar graphs show the percentage of 53BP1-positive cells normalized on \u03b3H2AX-positive cells\u00b195% CI *n*=3 independent experiments; at least 100 cells per sample have been analysed; \\**P*\\<0.05, \\*\\**P*\\<0.01, \\*\\*\\**P*\\<0.001, *\u03c7*^2^-test. (**g**) MEFs *Trf2*^*F/F*^-expressing GFP-TRF1 were permeabilized and incubated with the indicated Alexa Fluor 647-conjugated double-stranded RNAs. The graph shows the mean number of dots per cell\u00b1s.e.m., detected by proximity ligation assay (PLA), using an anti-GFP antibody, recognizing the telomeric protein TRF1 and an anti-Cy5 antibody, recognizing the Alexa Fluor 647 antigen; *n*=2 independent experiments; at least 80 cells per sample have been analysed; \\*\\*\\**P*\\<0.001, Student's *t*-test.](ncomms13980-f3){#f3}\n\n![Antisense oligonucleotides against tDDRNAs inhibit DDR activation at deprotected telomeres.\\\n(**a**) MEFs *Trf2*^*F/F*^ were treated, or not, with 4-hydroxytamoxifen (4OHT), transfected with the indicated ASO, and stained for the indicated DDR markers. Scale bar, 10\u2009\u03bcm (**b**) Quantification of data presented in **a**. Dot plots show the number or intensity of DDR foci per cell (a.u.=arbitrary units). Lines depict the mean\u00b1s.e.m. *n*=3 independent experiments; at least 50 cells per sample have been analysed; \\*\\**P*\\<0.01, \\*\\*\\**P*\\<0.001, one-way analysis of variance with Sidak corrections for multiple comparisons.](ncomms13980-f4){#f4}\n\n![Antisense oligonucleotides against tDDRNAs inhibit DDR activation *in vivo* in mice with deprotected telomeres.\\\n(**a**--**d**) Mice were injected with vehicle or tamoxifen (tam) to induce telomere uncapping, and 24\u2009h later, with the indicated ASO or PSB. (**a**) liver and (**c**) kidney sections were stained for the indicated DDR markers. Scale bars, 10\u2009\u03bcm. (**b**,**d**) Quantification of the data presented in **a** and **c**. Dot plots show the intensity of DDR foci per cell in (**b**) liver and (**d**) kidney (a.u.=arbitrary units). Lines depict the mean\u00b1s.e.m. *n*=4 mice per group; at least 200 cells per sample have been analysed; \\*\\*\\**P*\\<0.001, one-way analysis of variance with Sidak corrections for multiple comparisons.](ncomms13980-f5){#f5}\n\n[^1]: These authors contributed equally to this work\n\n[^2]: Present address: Institute of Molecular Biosciences, The University of Queensland, Queensland Bioscience Precinct, 306 Carmody Road, St Lucia, Queensland 4067, Australia\n"}
+{"text": "\n"}
+{"text": "\n"}
+{"text": "Adult Neurogenesis: Relevance in Physiology, Pathology, and Therapy\n===================================================================\n\nExtensive experimental work demonstrated that new neurons can be generated in adult mammalian brain ([@B6]). One adult neurogenic area is the dentate gyrus, in the hippocampal formation. Here, in a subregion called the SubGranular Zone (SGZ), resident neural stem/progenitor cells (NSC/NPC) are present, can self-renew and give rise to transiently amplifying progenitor cells which, in turn, generate neuroblasts capable of terminal neuronal differentiation in the granular cell layer ([@B44], [@B45]; [@B77]; [@B10]).\n\nRecently a comprehensive review summarized data on how hippocampal function and related behavior may be modified by newborn neurons ([@B33]). At present, adding new neurons in the hippocampal circuitry is suggested to result in encoding of temporal information into memories and in cognitive flexibility during new task learning. Pattern separation is the process that reduces overlap in the representation of similar memories and adult hippocampal neurogenesis (ahNG) has been associated with improved performance in pattern separation behavioral tasks ([@B3]; [@B68]). In several studies a reduction in newborn cells was found to correlate with specific cognitive deficits, and in particular with spatial memory impairment ([@B78]; [@B2]; [@B22]; [@B21]; [@B15]; [@B1]). Context-dependent memory, and specifically performance in contextual fear conditioning tasks, was also found to be neurogenesis-dependent ([@B33]).\n\nA vast array of research investigated factors that regulate ahNG. External stimuli, including learning, environmental enrichment, and exercise have profound effects on proliferation/ differentiation of NPC as well as survival of their progeny ([@B85]; [@B31]). Aging, chronic stress and social isolation result in negative impact on ahNG ([@B75]; [@B76]; [@B28]) and reduction of hippocampal neurogenesis has been hypothesized to contribute to cognitive decline or mood alterations associated with those conditions ([@B1]).\n\nExtensive research efforts suggest that ahNG may be deregulated in several neuropsychiatric disorders, including major depressive disorder ([@B47]; [@B59]; [@B22]; [@B46]; [@B91]).\n\nAn intringuing aspect of ahNG is its susceptibility to pharmacological modulation ([@B24]; [@B23]; [@B88]; [@B94]; [@B11]). Antidepressants increase hippocampal neurogenesis ([@B54]). Interestingly, such increase requires several weeks and occurs in parallel with the onset of the antidepressant effects ([@B56]). A vast array of experimental studies demonstrated that antidepressants counteract stress-reduced ahNG in rodent models of depressive-like disorder ([@B67]; [@B27]). An increased number of hippocampal NPC was reported in postmortem brain of depressed patients who were treated with antidepressants ([@B9], [@B8], [@B7]). Although still controversial, it has been suggested that antidepressants require ahNG to exert some behavioral effects in rodents ([@B70]; [@B18]) and that ahNG may contribute to the antidepressant activity of these drugs in the clinical setting ([@B42]). In line with this hypothesis, preclinical studies evaluating effects on ahNG have become part of the discovery process of recently approved antidepressant drugs ([@B81]; [@B37]) or of clinically relevant drugs which, in addition to their approved indications, may potentially exert antidepressant activity ([@B84]; [@B16]).\n\nOpiates as Negative Modulators of ahNG\n======================================\n\nEmerging evidence suggest that several psychoactive drugs result in molecular changes that may negatively affect different aspects of ahNG ([@B25]; [@B90]). These findings have important clinical implications since they raise the possibility that cognitive dysfunction and/or mood alteration in the setting of such drug use and/or abuse may, at least in part, be related to alterations in ahNG ([@B91]).\n\nOpiate drugs are powerful analgesics which are also among the most commonly abused addictive drugs. Clinical research suggested the occurrence of deficits in memory tasks, attention, verbal fluency and general cognitive performance in opiate addicts ([@B14]; [@B36]; [@B35]). Controversial, but of potential clinical relevance, is the possibility that cognitive impairment may also occur in non-addicted patients subjected to chronic opiate treatment ([@B79]; [@B30]; [@B48]; [@B50]; [@B39]). Despite opiate effects on cognition are difficult to decipher due to their central depressant effects, preclinical research supports the idea that chronic administration of opiates may interfere with cognition independently of performance ([@B82]; [@B58]). As an example, in rats, chronic morphine administration impaired hippocampal dependent long-term memory retention. Moreover, a selective impairment in spatial memory, which is considered to be neurogenesis-dependent, was observed in morphine- compared to vehicle-treated animals ([@B58]).\n\nDepression often co-occurs with chronic pain and an emerging line of inquiry is also the association between opiate use and the risk of depression ([@B74], [@B73], [@B72]; [@B80]). In particular, the onset of a new depressive state has been associated with longer drug administration ([@B72]).\n\nAltogether several experimental data suggest that long-term opiate may produce maladaptive changes in brain structures involved in cognition and mood regulation such as the hippocampus. In morphine-dependent animals, mechanisms responsible for the disruption of long-term memory retention have been suggested, including effects on dendritic spine stability ([@B51]) and long-term potentiation ([@B63]). ahNG is among the forms of neural plasticity which are regulated by opiates ([@B92]). Actually, opiates were the first drugs shown to negatively impact hippocampal neurogenesis ([@B25]). Since then, numerous rodent studies confirmed that finding ([@B55]; [@B43]; [@B26]; [@B4]; [@B93]). Based on these reports, it can be hypothesized that opiates may produce long-lasting effects on the neuronal circuitry involved in mood and cognition through, at least in part, disruption of ahNG.\n\nOpiate Effect on Distinct Stages and Cell Types of ahNG\n=======================================================\n\nAdult neurogenesis is a complex, multistage process which consists of a series of developmental events, namely proliferation, differentiation, migration, maturation and survival of NSC/NPC and their progeny. Based on their developmental stage, different cells participating in ahNG can be recognized and classified. Briefly, radial-glia-like stem cells, which are both glial fibrillary acidic protein (GFAP) and nestin positive are defined as Type-1 cells; Type-2 cells are neural progenitors which are highly proliferative and nestin^+^/GFAP^-^; neuroblasts which express doublecortin and the polysialylated form of neural cell adhesion molecule (PSA-NCAM) are referred to as Type-3 cells.\n\nOver the last few years, extensive progress has been made on how morphine can disrupt neurogenesis in the adult rodent hippocampus ([@B92]). The first report correlating opiate administration and adult neurogenesis was by the group of Amelia Eisch that, in 2000, reported that chronic morphine, administered via a subcutaneous pellet, decreased the number of cells which incorporated the thymidine analog BromodeoxyUridine (BrdU) in rodent SGZ. Importantly, no effect of acute morphine administration was reported ([@B25]). Since then, extensive evidence has been accumulated on the negative impact of morphine on NSC/NPC proliferation, using BrdU or endogenous markers of cell proliferation/cell cycle ([@B55]; [@B43]; [@B4]). Morphine effects appeared to be independent from the route of administration (subcutaneous pellet vs. intraperitoneal injection) and to occur during settings of both forced and self-administration ([@B55]; [@B43]; [@B29]).\n\nSurvival of newborn cells is also reduced by morphine. After being born in the dentate gyrus, a large portion of newly generated cells die within a few days ([@B19]). Chronic morphine treatment decreased granule cell survival *in vivo*, by largely reducing the number of 4-week-old BrdU-labeled cells in the dentate gyrus of drug-treated compared to control rats ([@B25]). Although adult hippocampal NPC can be isolated and maintained in culture with self renewal and multipotential properties ([@B84]; [@B16]), surprisingly few studies investigated the effects of opiates *in vitro*. In one report morphine exposure in cultured NSC/NPCs resulted not only in reduced proliferation but also in an increased caspase-3 activity in nestin^+^/GFAP^+^ cells and not in neuroblasts ([@B86]).\n\nThe detailed mechanisms of regulation of ahNG by morphine remain to be fully clarified. There is evidence of morphine acting directly on neural progenitors by engagement of opioid receptors on their cell surface, but also indirectly. Opiate analgesics exert their effects through receptor subtypes, referred to MOR, KOR and DOR, which interact with endogenous opioid peptides. The contribution of different receptor subtypes to the negative effects on ahNG has been investigated by both genetic and pharmacological approaches, pointing to a major role of MOR. We demonstrated that morphine adversely impacted on neuronal differentiation, neurite outgrowth and survival of adult hippocampal NPC and their progeny ([@B57]). MOR pharmacological blockade confirmed that the receptor subtype was responsible for morphine effects ([@B57]). Interestingly, knock-out of MOR enhanced adult-born granule cell survival *in vivo*, suggesting that endogenous opioids may have a negative effect on ahNG ([@B38]). These findings are not consistently confirmed *in vitro*. In isolated rat hippocampal NPC, incubation with \u03b2-endorphin increased total DNA content and the number of cells expressing proliferation markers such as Proliferating Cell Nuclear Antigen (PCNA) and phosphorylated histone H3 (pHisH3), and these effects were antagonized by naloxone ([@B64]). The same research group reported that MOR and DOR antagonists decrease proliferation of cultured NPC ([@B65]). Similarly, the long acting opioid antagonist naltrexone was shown to reduce cell proliferation in the hippocampus of adult rats ([@B40]).\n\nOpiates also interfere with NPC fate specification. Type-3 cell number was diminished by chronic morphine ([@B43]). Moreover, in rat, repeated morphine treatment altered the GABAergic phenotype of adult-born hippocampal granule cells by increasing the GABA synthesizing enzyme glutamate decarboxylase-67 ([@B43]). A more detailed analysis using BrdU and Ki67 proliferation markers concluded that morphine treatment increased Type-2b and decreased Type-3 cells in mouse SGZ ([@B4]). Also *in vitro* a remarkable decrease in neuronal differentiation of mouse hippocampal NPC by morphine has been demonstrated, an effect which is MOR mediated ([@B57]; [@B86]). Chronic treatment with MOR and DOR antagonists decreased adult NSC/NPC differentiation into astrocytes and oligodendrocytes, while favoring their neuronal differentiation. In the same experimental setting KOR antagonists had no effect ([@B65]). It cannot be excluded that opiates may also affect astrocytes (that express opioid receptors) which, in turn, can modulate ahNG with different mechanisms, including via secreted molecules ([@B17]). Immature adult generated neurons are excited by GABA ([@B32]) and they need excitatory signals from the preexisting circuit to complete their differentiation and maturation. Opiate agonists may interfere with this process by decreasing GABA release ([@B62]). Studies are needed to further understand the role of endogenous opioids and receptors in ahNG homeostasis ([@B52]).\n\nThe intracellular signaling pathways involved in the negative effects of morphine on neural progenitors were investigated only in a few studies. The basic helix-loop-helix transcription factor NeuroD1 is negatively regulated by morphine in NPC cultures ([@B95]). Under the conditioned place preference paradigm morphine, through a mechanism involving NeuroD1, impaired the differentiation of NSC/NPC into immature neurons ([@B93]). [@B87] demonstrated that, in presence of morphine, mouse NPC preferentially differentiated into astrocytes and not neurons. This effect was mediated by MOR and by miR-181a/Prox1/Notch1 pathway activation. Interestingly, the same group also demonstrated that miR-181a/Prox1/Notch1 pathway regulates NPC differentiation in a ligand-dependent manner ([@B87]), pointing to differences in the effect of distinct opiate molecules on mouse NPC differentiation. Morphine modulates the lineage-specific differentiation of NPC by PKC\ud835\udf00-dependent ERK activation with subsequent TAR RNA-binding protein (TRBP, a cofactor of Dicer) phosphorylation and miR-181a maturation. Conversely fentanyl activated ERK via the \u03b2-arrestin-dependent pathway, followed by nuclear translocation of phosphoERK.\n\nOverall, available data support the idea that morphine negatively affects neurogenesis acting on multiple cellular types and stages of the neuroplasticity process. Morphine properties on neurogenesis are also shared by other opiates. The partial agonist buprenorphine, when administered via subcutaneous injections over a 3-day period, reduced the number of actively proliferating cells in the hippocampus of adult mice ([@B66]). On the other hand, differences in the signaling pathways activated in NPC by different opiate drugs may underlie potential differences in their impact on ahNG.\n\nIn the future studies, should be specifically designed to correlate more stringently the disruptive cognitive effects of distinct opiates with specific alterations in ahNG and to discriminate those that are strictly dependent on neurogenesis from the ones that are neurogenesis-independent.\n\nNot All Opiates are Created Equal: Different Impact on Hippocampal Neurogenesis of Distinct Drugs\n=================================================================================================\n\nA recent *in vivo* study in rat found that chronically administered methadone does not alter parameters relevant to ahNG including the number of Ki67-, doublecortin-, or BrdU-immunoreactive cells ([@B69]). These results suggest that, unlike morphine, methadone may not alter hippocampal plasticity. Interestingly methadone is an atypical opiate, since it is a MOR agonist and a non-competitive NMDA antagonist ([@B34]). Incidentally, NMDA antagonists positively modulate hippocampal neurogenesis in rodents ([@B61]; [@B53]). Future investigation should address whether the lack of negative effects on ahNG by methadone could be ascribed to its NMDA receptor antagonism.\n\nTapentadol is a centrally acting analgesic drug which combines, in a single molecule, MOR agonism and blockade of norepinephrine (NE) reuptake ([@B83]). Since ahNG is positively modulated by NE ([@B41]; [@B49]), we compared the effects of morphine and tapentadol on differentiation, maturation and survival of neurons generated from adult hippocampal NPC *in vitro*. Morphine significantly hampered neuronal differentiation, neurite outgrowth and survival of adult NPC and their progeny ([@B57]). In presence of tapentadol cell survival was not affected. Conversely, tapentadol reduced the number of newly generated neurons and their neurite outgrowth but only at concentrations which activate MOR and not at higher ones which block NE reuptake. Specifically, tapentadol proneurogenic and antiapoptotic effects appeared to be mediated by activation of \u03b22 and \u03b12 adrenergic receptors, respectively ([@B57]). As a proof of concept, in the same experimental model, morphine antineurogenic and proapoptotic effects were counteracted by reboxetine, an antidepressant which selectively blocks noradrenaline reuptake ([@B57]). In line with these *in vitro* results, chronic administration of a clinically relevant dose of tapentadol did not negatively affect proliferation and differentiation toward the neuronal lineage of newly generated cells in adult mouse hippocampus ([@B57]). Altogether these data support the idea that the noradrenergic component in tapentadol has the potential to counteract the adverse MOR-mediated effects on hippocampal neurogenesis both *in vitro* and *in vivo*. In principle, this counter-balancing effect may result, like for the atypical opiate methadone, in less or no dysfunction in adult neurogenesis and, potentially, in neurogenesis-associated functions after long-term treatment *in vivo*.\n\nClinical Implications and Long-Term Perspectives\n================================================\n\nCognitive dysfunction has been often reported in opiate drug abusers ([@B26]; [@B12]). Moreover, major depression represents an important comorbidity in chronic pain states and recent work has suggested an association between prolonged opiate use and the risk of new onset depression ([@B73], [@B72]).\n\nPreclinical evidence summarized in this review support the idea that disruption of ahNG may, at least in part, contribute to the long-lasting effects produced by some opiates, like morphine, on cognition and mood regulation. Although preliminary, these preclinical data mark the importance of taking into account inhibition of ahNG as a potential long-term consequence of opiate use also in the clinical setting. This aspect appears even more relevant in consideration of the fact that chronic pain *per se* may induce profound changes in hippocampal plasticity, including deregulated hippocampal neurogenesis ([@B60]; [@B20]).\n\nWhile we wait for a definitive answer on any causal correlation between long term-opiate treatment and cognitive and emotional impairment in chronic pain patients, preclinical studies should be undertaken to increase our current understanding of the cellular and molecular effects of opiates on adult NPC and their progeny. Moreover, experimental work should be aimed at understanding differences between distinct opiates in their potential for disrupting ahNG. Ultimately these studies may help us to unravel different level of risk associated with distinct treatment options. In principle, drugs like methadone and tapentadol may result, after drug long-term treatment, in reduced or absent dysfunction in ahNG compared to morphine and other opiates. Alternatively, other analgesic drugs which do not reduce but rather increase ahNG in preclinical models like pregabalin/gabapentin ([@B84]) and acetyl-[L]{.smallcaps}-carnitine ([@B16]) could be proposed, at least in neuropathic pain where they are effective. A better understanding of whether different opiate molecules affect molecular and cellular aspects of ahNG may also help designing and developing novel analgesic drugs for chronic pain states. Several drugs acting at different targets are under development for chronic pain treatment ([@B71]; [@B5]; [@B13]; [@B89]). We propose that future drug design should take into consideration the development of powerful analgesics that, by preserving ahNG, may also protect cognitive functions and counteract mood alterations which often represent comorbidities in chronic pain states.\n\nAuthor Contributions\n====================\n\nVB and MG were involved in the concept, literature screening, and writing of the article, and approved it for publication.\n\nConflict of Interest Statement\n==============================\n\nIn the past, MG has received research grants from drug companies manufacturing analgesic drugs, including Angelini S.p.A, Grunenthal GmbH, Pfizer, Sigma Tau. The other author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe would like to apologize for excellent work that could not be quoted in this review for space limitations.\n\n[^1]: Edited by: *Maria Angela Sortino, University of Catania, Italy*\n\n[^2]: Reviewed by: *Cesare Patrone, Karolinska Institutet, Sweden; Anna Maria Pittaluga, University of Genoa, Italy*\n\n[^3]: This article was submitted to Experimental Pharmacology and Drug Discovery, a section of the journal Frontiers in Pharmacology\n"}
+{"text": "Introduction {#s1}\n============\n\nSuppose that a researcher wants to study individual differences in how children respond to a training program. The predictor is a continuous measure of some individual property P. The dependent variable is improvement in ability, measured by a pre-training test and a post-training test. However, test scores are not perfect measures of ability. The same improvement in test scores from different pretest scores may therefore reflect different changes in ability. To control for this possibility the researcher may include the pretest score as a covariate in a regression analysis, thereby investigating whether property P predicts test score improvement *among children with equal pretest scores*. The topic of our paper is how this statistical adjustment may lead to incorrect conclusions because of regression artifacts arising from biased regression to the mean.\n\nIn 1999, Campbell and Kenny [@pone.0095949-Campbell1] devoted an entire book to warning about the dangers of statistical adjustments in comparisons of treatment effects between non-randomized groups. The basic problem was pointed out already thirty years earlier in a classic paper by Campbell and Erlebacher [@pone.0095949-Campbell2]. The problem has continued to attract attention, see [@pone.0095949-vanBreukelen1] for a review of the literature and a novel analysis. One reason to write yet another paper on this topic is that the literature has focused on comparisons between groups. Although the logic is the same for analysis of a continuous individual property, this case deserves an explicit discussion. The direct motivation comes from a rather recent empirical study in numerical cognition [@pone.0095949-Booth1], which analyzed the influence of a continuous property on learning. That study is particularly interesting because, in addition to a pre-training test and an end-of-training test, it included a follow-up test two weeks after the end of training with no training between end-of-training and follow-up. This feature will prove useful in assessing the size of the regression artifact.\n\nFormalizing the Problem {#s1a}\n-----------------------\n\nLet us formalize the abovementioned setup: A researcher wants to study individual differences in how children respond to a training program. The predictor is a continuous measure of some individual property P. The dependent variable is improvement in ability, (imperfectly) measured by a pre-training test score and a post-training test score . The researcher includes the pretest score as a covariate in a regression analysis. Denoting the property P level of child *i* by , this means estimating the following regression model:\n\nHere, *K* is the coefficient of interest, measuring the influence of property P on test score change. *D* is the intercept, *L* is a coefficient measuring the influence of pretest score on test score change, and the residual.\n\nAs discussed by other authors [@pone.0095949-vanBreukelen1] it is irrelevant whether the dependent variable is test score change or simply the posttest score, because addition of the pretest score to both sides yields an equivalent model predicting posttest score instead of change:\n\nAlthough inclusion of the pretest score as a covariate may seem both innocuous and sensible, it will lead to biased results when two critical conditions hold. The first condition is that P is correlated with pretest ability. The second condition is that test scores are not fully reliable measures of ability but subject to random within-individual variation, commonly represented by a \"true score\" model in which the test score is the sum of the child's latent ability (true score) and a random error term of positive variance:\n\nImprovement of test results will then reflect not only actual arithmetic learning () but also two random errors ().\n\nWhy Regression Artifacts Arise {#s1b}\n------------------------------\n\nThe only novelty of our setup is that property P is continuous. A classic setup is retrieved in the special case of P taking only values 0 or 1, indicating membership in one of two groups. Our first critical condition then reduces to the presence of a group differences in pretest scores. The risk of a regression artifact in that case was pointed out more than forty years ago by Campbell and Erlebacher [@pone.0095949-Campbell1]. The logic of their argument carries over directly to the continuous setting and goes as follows:\n\nChildren who are higher on property P will tend to have higher pretest scores than children who are lower on property P. By selecting to compare children with equal pretest scores, the researcher will inadvertently make a biased selection of the random errors. Specifically, consider a higher-P child and a lower-P child who happened to have the same pretest score. Equal test scores will arise by chance when a child with higher ability has less luck on the test than a child with lower ability. Because of regression to the mean, the child with worse luck on the first test--i.e., the one with higher ability--will tend to do better than the other child on the second test. Because children with higher P tend to have higher ability, equality in test scores will most often reflect a situation where the higher-P child has had worse luck and will therefore tend to be luckier on the next test. The observation that regression to the mean has this consequence of divergence between members of different groups is sometimes referred to as Kelley's paradox [@pone.0095949-Wainer1]. The consequence is that, without any improvement in latent ability, the higher-P child is likely to score better on the next test than the lower-P child with the same pretest score. Any genuine changes in ability will be confounded by this bias in random errors.\n\nNote how the bias was caused by the combination of initial differences and regression to the mean. The confounding effect on the results of the regression analysis is called a regression artifact.\n\nOutline of Paper {#s1c}\n----------------\n\nThe paper consists of three studies. The first study is a mathematical analysis of the emergence of the regression artifact. We derive an unbiased estimator of the regression artifact under certain simplifying assumptions. The second study is a computer simulation to illustrate the regression artifact, leading up to a discussion of Lord's paradox in a continuous setting. The third study applies our theoretical framework to a reanalysis of the main finding of the aforementioned study in numerical cognition [@pone.0095949-Booth1].\n\nStudy 1 {#s2}\n=======\n\nAnalysis {#s2a}\n--------\n\nTo demonstrate how a regression artifact may arise in comparisons of groups, Campbell and Erlebacher [@pone.0095949-Campbell2] specified a model for test scores and abilities in two hypothetical groups. The model assumed abilities to be constant over time. Test scores could change, but only because of random errors. In the model it was therefore absolutely certain that finding any influence of group membership on change in test scores must be an artifact of the random errors.\n\nHere we consider the case where children vary on a continuous property P rather than belong to one of two groups. We will demonstrate how a regression artifact arises when the test score difference is regressed on property P if the pretest score is included as a covariate. Adapting the model of Campbell and Erlebacher [@pone.0095949-Campbell2] to the case of a continuous property, we let child *i*'s actual ability (equal at pretest and posttest) be generated by the following equation:\n\nThe first two terms specify a linear relationship between ability and property P, while the last term denotes unexplained between-individual variation in ability. We assume these random errors to be independently drawn from a normal distribution with mean 0 and standard deviation *s.*\n\nFollowing equations (2) and (3), pretest and posttest scores of child *i* are obtained by addition of random errors to the child's level of ability. We assume these random errors to be independently drawn from a normal distribution with mean 0 and standard deviation *\u03c3.*\n\nIt is then possible to mathematically derive the expected size of the regression artifact. Under the model assumptions (2--4), the linear regression model (1) translates into\n\nTo analyze the results obtained from least-square estimation of this regression model we use the standard approach of letting the sample size tend to infinity, such that stochastic effects can be ignored. We can then identify coefficients between the left-hand and right-hand expressions to obtainand mean-square residual\n\nthe law of large numbers. Least-square estimation entails minimizing this expression with respect to . The minimum is attained for\n\nand plugging this into the second identity of [equation (5)](#pone.0095949.e012){ref-type=\"disp-formula\"} we obtain\n\nResults and Discussion {#s2b}\n----------------------\n\nFor infinite samples, [equation (6](#pone.0095949.e016){ref-type=\"disp-formula\"}) gives the exact size of the regression artifact. For finite sized samples there will also be stochastic effects and (6) is then an unbiased estimate of the regression artifact. Note that this estimate is the product of the strength of the relation between property P and ability (*b*) and the proportion of the total variance in pretest scores that is accounted for by random within-individual variation. The regression artifact arises when these two entities are nonzero, which is equivalent to the two critical conditions stated earlier.\n\nStudy 2 {#s3}\n=======\n\nFollowing Campbell and Erlebacher [@pone.0095949-Campbell1], we will use computer simulations of our model (i.e., equations 2--4) to demonstrate the regression artifact arising from estimation of the regression model (1).\n\nMethod {#s3a}\n------\n\nIn order to simulate data we need to choose values for the model parameters *a, b, s,* and *\u03c3.* We must also create a number of simulated children by assigning them a level of property P. To demonstrate the size of regression artifact that might arise in real empirical studies, we have chosen parameter values to roughly match the data of [@pone.0095949-Booth1]. We will discuss that study in detail in a later section. For now it suffices to say that 105 children were measured on the linearity of their numerical magnitude estimations, corresponding to our property P. We used this dataset as our P levels, fixed throughout our simulations. Model parameter values *a\u200a=\u200a*0.5, *b\u200a=\u200a*0.4, *s\u200a=\u200a*0.1, and *\u03c3\u200a=\u200a*0.1, were chosen to give a rough fit to the empirical data on test scores in [@pone.0095949-Booth1]. Simulated abilities and test scores were then obtained from the model equations (2--4), with random terms obtained from a generator of random numbers built into Excel. One thousand simulated datasets were generated in this way. For each dataset we then estimated regression model (1).\n\nResults {#s3b}\n-------\n\nOur interest lies in , the estimated size of the influence of property P on test score change. [Figure 1](#pone-0095949-g001){ref-type=\"fig\"} shows how was distributed over the 1,000 simulations. It was greater than zero in more than 98 percent of simulations, with a mean value of 0.2004. Because of the large number of simulations, the mean value should be very close to the unbiased estimate given by [equation (6](#pone.0095949.e016){ref-type=\"disp-formula\"}). Indeed, plugging the parameter values (*b\u200a=\u200a*0.4, *s\u200a=\u200a*0.1, and *\u03c3\u200a=\u200a*0.1) into [equation (6](#pone.0095949.e016){ref-type=\"disp-formula\"}) yields \u200a=\u200a0.2.\n\n![Histogram over the distribution of the unstandardized coefficient for the influence of property P on test score change, obtained from estimating regression model (1) in 1,000 simulated datasets in which actual abilities did not change between tests.](pone.0095949.g001){#pone-0095949-g001}\n\nLet us emphasize what these simulations tell us. They assume a situation where no child's ability change between tests, so *a fortiori* there is no influence of property P on change in abilities. Despite this absence of a genuine effect, the simulations indicate that researchers using regression model (1) will almost always find a substantial influence of property P on change in test scores.\n\nNote that the artifact tended to be half as large as *b*, the coefficient describing how property P relates to ability. [Equation (6](#pone.0095949.e016){ref-type=\"disp-formula\"}) explains why: The parameter values we used satisfied *s*\u200a=\u200a*\u03c3*, which implies that random within-individual variation accounted for half the total variance in pretest scores, with the other half accounted for by between-individual variation in abilities.\n\nDiscussion {#s3c}\n----------\n\nWe shall close the theoretical part of this paper by a discussion of Lord's paradox. In its original form, Lord's paradox is about comparison of groups. Let us therefore consider hypothetical children whose P values are either 0 or 1, such that groups can be based on P values. Using our simulation model we generated abilities and test scores for a hypothetical set of 105 children, equally distributed over the two P values. The results are presented in a scatter plot of pretest score against test score change (i.e., posttest minus pretest), see [Figure 2](#pone-0095949-g002){ref-type=\"fig\"}. We shall discuss this plot in some detail.\n\n![Scatter plot of simulated pretest score and test score change for children having either value 0 or 1 on property P.\\\nLatent abilities depended on P and did not change between tests.](pone.0095949.g002){#pone-0095949-g002}\n\nFirst consider the solid line, indicating no change in test score. Because abilities did not change between tests in our model, the solid line is where all datapoints would have been if test scores had been perfect measures of ability. Because test scores included random errors in our model, the data points are instead distributed above and below the solid line to the same extent. Based on this observation an empirical researcher could draw the conclusion that *property P had no systematic influence on change in test scores*.\n\nNow consider the dashed lines, showing the results of regressing test score change on pretest score in each group. These lines demonstrate another observation about this dataset: For children with the same pretest score, test score change tends to differ substantially between the high P group and the low P group. Based on this observation, an empirical researcher could instead draw the conclusion that *property P has a substantial systematic influence on change in test scores*.\n\nThis phenomenon, that the same data on pretests and posttests in two groups can yield conflicting conclusions depending on what aspect of the data is observed, was first pointed out in a classic paper by Lord [@pone.0095949-Lord1]. It is commonly referred to as Lord's paradox. By considering a continuous property P instead of group membership, we have the same paradox in a continuous setting.\n\nSo, which conclusion is correct? Given our knowledge about the model that generated this dataset, the answer is unambiguous: The first conclusion is correct and the second conclusion is incorrect. The relation between property P and change in test scores is just an artifact of regression to the mean and reflects no causal influence. [Figure 2](#pone-0095949-g002){ref-type=\"fig\"} allows an intuitive way to think about the regression artifact: Because of regression to the mean, high pretest scorers are more likely to score worse next time and data will therefore tend to slope downwards to the right. The high P group tended to score higher, and therefore lie to the right of the low P group. For data that slope downwards to the right, a slope that lies to the right of another slope will also look as if it lies above it.\n\nThe crux of the matter is that the empirical researcher would not know which model generated the data. Specifically, our mathematical analysis implies that equivalent data are generated by the following model:\n\nIn words, this says that equivalent data would be observed if test scores were perfect measures of ability and if change in abilities were to some degree random but positively influenced by property P and negatively influenced by pretest ability. In such a world, the influence of P found by including pretest score as a covariate would be genuine and not a regression artifact. Instead it would be the analysis *without* the covariate that would point toward an incorrect conclusion, as it would not detect the positive influence of property P. Finally, in a world represented by a mixture of the two models both analyses would be biased.\n\nThe empirical researcher who wants to draw a conclusion about how much influence, if any, property P has on change in ability therefore needs additional knowledge about the underlying processes. Such knowledge may well exist. For instance, one may have an understanding about the mechanisms whereby ability changes. According to such understanding it might be implausible that ability would systematically decrease between tests among highly able children. This would support the first conclusion that the observed decrease in test scores among high pretest scorers is due to regression to the mean. It is also likely that a researcher has some knowledge about the extent of within-individual variation in test scores. Such knowledge can come from the nature of the test itself as well as from analysis of repeated tests with no treatment in-between. We shall later appeal to this kind of knowledge in our reanalysis of an empirical study.\n\nLord's own version of the paradox presented data on weight change among male and female college students over an academic year [@pone.0095949-Lord1]. The question was whether men and women tended to respond differently to the diet provided in the college dining halls. This is exactly the dichotomous version of the problem we study in this paper. Thus, individuals differ on a property P (gender in Lord's problem), and the question is how this property of individuals affects their response to the same treatment. Most of the literature on Lord's paradox focuses on another setup in which, depending on group membership, individuals either receive treatment or no treatment (control) and the question is if treatment has a different effect than no treatment. The statistical analyses look the same but interpretations will be different.\n\nFor a recent review and analysis of Lord's paradox in treatment vs. control we refer to van Breukelen [@pone.0095949-vanBreukelen1]. The focus of van Breukelen's analysis is the circumstances under which the two methods of analysis (i.e., whether or not to include the pretest score as a covariate) will give unbiased results. Assuming non-randomized group assignment and presence of random errors, the conclusions can very briefly be summarized as follows: First, inclusion of the covariate will lead to biased results if groups differ in initial ability. This echoes what Campbell and Erlebacher [@pone.0095949-Campbell2] pointed out 40 years ago and is the same that we have said here in a continuous setting. Second, the alternative of not including the covariate may suffer from another source of bias, namely, that pretest ability actually influences the response to treatment, as in model (7). Third, both methods are unable to account for inherent differences between groups in how they respond to treatment, as treatment was only given to one of the groups. The third point obviously does not apply to the problem we (and Lord [@pone.0095949-Lord1]) consider, in which everyone is given the same treatment.\n\nTo be able to test for the presence of bias in results, van Breukelen [@pone.0095949-vanBreukelen1] suggests that experimental designs include two pretests with sufficient time in-between. If the analysis is unbiased, no effect should then be found treating the second pretest as posttest. The same logic applies if there are two posttests instead. In our below reanalysis of data from [@pone.0095949-Booth1], we shall follow van Breukelen's suggestion.\n\nStudy 3 {#s4}\n=======\n\nSo far we have theoretically discussed why and when a certain statistical analysis method will produce a regression artifact. We now turn to an empirical study where this method of analysis was used. The background is an interesting and well-established research finding that children's proficiency in solving arithmetic problems correlates with the linearity of their estimations of numerical magnitudes. Specifically, arithmetic performance tends to be better the more the child estimates numerical magnitudes in a linear rather than logarithmic way. This fact has been demonstrated in many studies, as reviewed by Booth and Siegler [@pone.0095949-Booth1]. The objective of their 2008 study was to investigate whether arithmetic *learning* was also influenced by this property. The researchers' hypothesis was that if children are trained on arithmetic problems, their arithmetic learning will be influenced by the linearity of their numerical magnitude estimations. Specifically, they predicted that children who make more linear estimations of numerical magnitudes would benefit more from training on arithmetic problems.\n\nMethods {#s4a}\n-------\n\nBooth and Siegler [@pone.0095949-Booth1] studied 105 first graders over several sessions across a few weeks, collecting a number of measures. For the purposes of our paper, only the below-mentioned measures are relevant as they were the ones that entered the critical analysis.\n\n### Numerical magnitude estimations {#s4a1}\n\nThe first session included the task of estimating the positions of 26 different numbers (ranging between 2 and 98) on a number line between 0 and 100. The researchers then measured the linearity of a child's estimations by calculating the proportion of variance in estimations explained by a best-fitting linear expression of the numbers to be estimated. This measure will be referred to as *R* ^2^ ~Lin~.\n\n### Test of arithmetic performance {#s4a2}\n\nA set of four arithmetic problems (9+18, 26+27, 17+29, and 49+43) was used to test arithmetic performance. Children were asked to solve these problems in the first session. A child's performance on the problem set was measured as the average absolute error in answers divided by 100, referred to as \"percent absolute error\" (PAE). For instance, a child giving answers 28, 50, 50, and 80 to these problems would have made absolute errors 1, 3, 4, and 12, yielding an average absolute error of 5 and a PAE of 0.05.\n\nIn two subsequent sessions, children received training on the same problems. (Training occurred in four between-subject conditions using different instructional procedures. However, all conditions were pooled in the analysis of the main hypothesis. Because this is the analysis we are concerned with in the present paper, the fact that there were different conditions will not be relevant to our account.) At the end of training, children were again given the same problems to solve. In a follow-up session two weeks after the end of training, children solved the same set of problems for a third time. Thus, three performance measures were collected: pre-training (PAE~pre~), at end-of-training (PAE~end~), and at follow-up (PAE~followup~).\n\nAnalysis {#s4b}\n--------\n\nBooth and Siegler [@pone.0095949-Booth1] analyzed their data using regressions that included pre-training performance (PAE~pre~) as a covariate. An impressive 39 percent of the variance in the test score difference PAE~end~\u2212PAE~pre~ was explained by a multiple regression on *R* ^2^ ~Lin~ and PAE~pre~, with both factors coming out as highly significant predictors. A similar result was obtained for the difference in performance between the first session and the follow-up session: *R* ^2^ ~Lin~ and PAE~pre~ together explained 29 percent of the variance in PAE~followup~\u2212PAE~pre~, again with both factors coming out as highly significant predictors. The researchers concluded that arithmetic learning is influenced by the linearity of children's numerical magnitude estimations.\n\n### Assessing the risk of a regression artifact {#s4b1}\n\nNote that this study fits perfectly with our previous theoretical discussion. The researchers' aim was to study how arithmetic learning is influenced by a certain continuous property, linearity of numerical magnitude estimations, operationalized by the quantity *R* ^2^ ~Lin~. This property is known to be related to arithmetic ability. The first condition for a regression artifact was therefore likely to be satisfied. Further, learning was measured as the change in test scores. These test scores measure how far off children were from the correct answers to difficult arithmetical problems. As a measure of arithmetic ability, this must be expected to suffer from substantial random errors. Children are likely to use guessing when they don't know the right answer. They will then, by chance, sometimes come close to the right answer and sometimes not. Thus, the second critical condition for a regression artifact was also likely to be satisfied. Because the researchers used a method of statistical analysis that suffers from a regression artifact under the combination of these two critical conditions, we must expect their results to be biased. It might be that their finding was entirely due to the regression artifact. This calls for a reanalysis of their data.\n\n### Aim of reanalysis {#s4b2}\n\nOur aim is to estimate to what extent Booth and Siegler's results suffer from the regression artifact and to assess whether or not their research conclusion still holds when the regression artifact is accounted for. To estimate the size of the regression artifact, we have conducted some additional analyses. We thank Julie Booth for sharing the raw data for this reanalysis. The data are presented in three scatter plots.\n\n### Relation between initial test score and R^2^ ~Lin~ {#s4b3}\n\n[Figure 3](#pone-0095949-g003){ref-type=\"fig\"} plots children's pretest score (PAE~pre~) against their linearity of numerical magnitude estimations (*R* ^2^ ~Lin~). Recall that the test score measures percent absolute error in responses, so *lower values mean better performance*. It is clear from [Figure 3](#pone-0095949-g003){ref-type=\"fig\"} that children higher on *R* ^2^ ~Lin~ performed better on the pretest. This is the first of the two conditions that give rise to the regression artifact. A simple linear regression of PAE~pre~ on *R* ^2^ ~Lin~ yields an estimated value of \u22120.36 of the unstandardized coefficient. This corresponds to parameter *b* in our simulation earlier. However, it should be noted that the relation in [Figure 3](#pone-0095949-g003){ref-type=\"fig\"} seems to be non-linear, whereas the simulation model was linear.\n\n![Scatter plot of pre-training performance (PAE~pre~) plotted against linearity of numerical magnitude representation (*R* ^2^ ~Lin~).](pone.0095949.g003){#pone-0095949-g003}\n\n### Relation between test score change and R^2^ ~Lin~ {#s4b4}\n\n[Figure 4](#pone-0095949-g004){ref-type=\"fig\"} plots the test score difference to the end-of-training test (PAE~end~\u2212PAE~pre~) against *R* ^2^ ~Lin~. Similarly, [Figure 5](#pone-0095949-g005){ref-type=\"fig\"} plots the test score difference to the follow-up test (PAE~followup~\u2212PAE~pre~) against *R* ^2^ ~Lin~. No correlations are evident in these plots. Statistical tests confirm that there was *no* statistically significant link between linearity of numerical magnitude estimations (*R* ^2^ ~Lin~) and improvement in test scores from training, neither when measured at end of training (PAE~end~\u2212PAE~pre~), *r* ~S~ *\u200a=\u200a*.07, *p\u200a=\u200a*.47, nor when measured at follow-up (PAE~followup~\u2212PAE~pre~), *r* ~S~ *\u200a=\u200a*.03, *p\u200a=\u200a*.76. To avoid any misunderstanding of this result, let us point out that the meaning of the positive sign of the non-significant correlations is that linearity of numerical magnitude estimations *negatively* correlated with test score improvement. In other words, this analysis if anything suggests an effect in the opposite direction to Booth and Siegler's reported finding. Next we replicate Booth and Siegler's analysis. Including the initial test score (PAE~pre~) as a covariate in a linear regression of the test score differences on *R* ^2^ ~Lin~, we find estimated unstandardized coefficients of \u22120.27 at end of training and \u22120.26 at the follow-up session. These values correspond to the outcome variable in our simulations in Study 2 (although with the opposite sign because PAE test scores measure negative ability). However, note that the real data do not support that relations between variables are linear, which they were assumed to be in our simulations.\n\n![Scatter plot of change in performance from the first session to the end of training (PAE~end~\u2212PAE~pre~) plotted against linearity of numerical magnitude representation (*R* ^2^ ~Lin~).](pone.0095949.g004){#pone-0095949-g004}\n\n![Scatter plot of change in performance from the first session to follow-up (PAE~followup~\u2212PAE~pre~) plotted against linearity of numerical magnitude representation (*R* ^2^ ~Lin~).](pone.0095949.g005){#pone-0095949-g005}\n\n### Lord's paradox in a continuous setting {#s4b5}\n\nTaken together, the above analyses show that the dataset exhibits our continuous version of Lord's paradox. On the one hand, simple correlations indicate that linearity of numerical magnitude estimations does not have any positive influence on test score improvement. On the other hand, when the pretest score was included as covariate the results clearly indicate a positive influence on test score improvement. Which of these results best reflect the answer to the real research question -- whether there is any influence on *arithmetic learning*? Recall that we have very good reason to believe that test scores include substantial random variation due to guessing. When a child happens to make worse guesses on the second test than on the first test it is incorrect to infer that the child's arithmetic ability has deteriorated. Similarly, it is incorrect to infer that a child's arithmetic ability has improved if a child happens to make better guesses on the second test than on the first test. For this reason, it is crucial to estimate how much of children's changes in test scores are due to learning and how much is due to random variation.\n\n### Arithmetic learning {#s4b6}\n\nDid arithmetic learning take place at all? In [Figures 3](#pone-0095949-g003){ref-type=\"fig\"} and [4](#pone-0095949-g004){ref-type=\"fig\"} it is not obvious to the eye that later test scores showed any systematic improvement from the pretest. A statistical analysis reveals that the median change in test score from pre-training to end of training was small (0.022, less than a fifth of the standard deviation of 0.118) but statistically significant, *p\u200a=\u200a*.014, signed rank test. From pre-training to follow-up, however, there was no statistically significant change in scores (median\u200a=\u200a0.005), *p\u200a=\u200a*.72, signed rank test. This indicates that arithmetic learning was weak or even non-existent. Changes in test scores seem mainly to be driven by random variation. We must therefore expect biased regression to the mean to be a strong driver of the results of Booth and Siegler's analysis. [Figure 6](#pone-0095949-g006){ref-type=\"fig\"} provides another way of looking at the data. Here we plot test score change (to the end-of-training session) against the pretest score. To illustrate *R* ^2^ ~Lin~ in the same plot we have conducted a median split of children into \"more linear\" and \"less linear\" (i.e., above and below median on *R* ^2^ ~Lin~, respectively). [Figure 6](#pone-0095949-g006){ref-type=\"fig\"} is analogous to how we presented Lord's paradox in simulated data in [Figure 2](#pone-0095949-g002){ref-type=\"fig\"}. In the plot it is apparent that children who performed well on the pretest tended to exhibit a worsening of performance on the next test. To reconcile this pattern with the hypothesis that test scores differences reflect actual learning, we must accept that children with high ability tend to lose ability from arithmetic training. This seems implausible. Indeed, Booth and Siegler's theoretical argument clearly assumes that training increases ability.\n\n![Scatter plot of change in performance from the first session to end of training (PAE~end~\u2212PAE~pre~) plotted against initial test score.\\\nBlack and white dots signify \"more linear\" and \"less linear\" children according to a median split based on *R* ^2^ ~Lin~.](pone.0095949.g006){#pone-0095949-g006}\n\n### The regression artifact in a no-training period {#s4b7}\n\nOur final analysis capitalizes on Booth and Siegler's inclusion of a follow-up test. Between the tests at end-of-training and follow-up no child received training. This is as close as an empirical study can get to ascertain that no actual learning affects the results between two tests. As suggested by van Breukelen [@pone.0095949-vanBreukelen1], we obtain an estimate of the size of the regression artifact by running the same statistical method using the first posttest as a pretest to the second posttest. In other words, we regress the test score difference PAE~followup~\u2212PAE~end~ on *R* ^2^ ~Lin~, including PAE~end~ as a covariate. The result is an estimated unstandardized coefficient of \u22120.25. This result is essentially identical to the results of our replication of Booth and Siegler's analyses, which yielded unstandardized coefficients of \u22120.27 and \u22120.26. We conclude that there is no evidence of any influence of *R* ^2^ ~Lin~ on test score changes beyond the influence that stems from random variation in test scores.\n\nDiscussion {#s5}\n==========\n\nIn this paper we have discussed a pitfall in regression analysis of individual differences in change of test scores between a pretest and a posttest. Inclusion of the pretest score as a covariate may produce a regression artifact of a kind that has been discussed for a long time in the statistical literature [@pone.0095949-Campbell1]--[@pone.0095949-vanBreukelen1], [@pone.0095949-Wainer1]--[@pone.0095949-Lord1]. The problem arises when two critical conditions hold, namely, when the property used to predict learning is correlated with initial ability and when individuals' test scores show some random change between tests even if their actual ability does not change. Under the second condition, test scores on the second test will exhibit regression to the mean. Under the first condition, this will make individuals that happen to have the same pretest score tend to differ systematically on the second test. The effect is that inclusion of the pretest score as a covariate in the regression will make it look as if the property actually predicts learning, when it actually just predicts the pretest score. Whereas previous literature has focused on the case when the property is group membership, we have discussed how the same problem occurs in a continuous setting.\n\nIn an empirical study, Booth and Siegler [@pone.0095949-Booth1] included the pretest score as a covariate in their regression analyses of how arithmetic learning is influenced by the linearity of numerical magnitude estimations. Their study satisfied the conditions under which the regression artifact arises and their results were therefore likely to be biased. Reanalysis showed that their data are consistent with the null hypothesis that arithmetic learning is not influenced by the linearity of numerical magnitude estimations. In other words, the null hypothesis should not be rejected. Thus, the regression artifact made the researchers draw the wrong conclusion about their research question. We first pointed out the regression artifact to the journal that published the original study. They responded that they receive far more manuscripts than they can publish and therefore declined to publish a refutation. We would like to offer an alternative viewpoint: Readers have reason to believe that publication in a highly selective journal is a reliable sign of correctness. Journals would safeguard this reliability by making sure to inform their readers in all (hopefully few) cases where the conclusions of research they have published later turns out to be seriously flawed. Such a practice would also help to avoid prolonged popular belief in incorrect findings [@pone.0095949-Eriksson1].\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Analyzed the data: KE. Wrote the paper: KE OH.\n"}
+{"text": "Background {#Sec1}\n==========\n\nLaryngeal injuries are often undiagnosed in the initial evaluation of the trauma patient. They are rare, with an estimated incidence of one in every 30,000 emergency department admissions \\[[@CR1]\\]. Delayed recognition and intervention may prove fatal in the presence of upper airway obstruction \\[[@CR2]\\].\n\nBlunt laryngeal trauma may present with varying degrees of severity, from mild to life-threatening extremes. A tracheostomy may be required to gain airway access distal to the site of injury. A systematic classification and management approach of blunt laryngeal trauma is crucial to guide early decision-making and improve patient outcome in the emergency department.\n\nCurrent standard of care for laryngeal trauma is determined according to the Schaefer classification of laryngeal injury. Schaefer group 1 and group 2 with minor endolaryngeal injuries can be managed conservatively with observation, antibiotics, steroids, voice rest and anti-reflux medications. However, for more severe Schaefer type 3--5 injuries, open surgical repair will be required to secure a definitive airway \\[[@CR1], [@CR2]\\].\n\nGood history taking, detailed clinical examination and a high index of suspicion are critical in the diagnosis of laryngeal trauma. The diagnosis can be aided using flexible nasendoscopy by direct visualization of the airway. CT scanning of the neck is still considered the gold standard to grade the severity of the injury and to direct appropriate management. Obtaining a timely CT scan may often be challenging due to logistical problems, primarily availability of radiology service support especially in resource-limited area, and stability for patient transfer.\n\nOf late, studies have integrated the use of upper airway ultrasound into point-of-care ultrasound examination, a paradigm shift in upper airway assessment \\[[@CR3]--[@CR5]\\]. The incorporation of ultrasound into the diagnostic arm may expedite the intervention process by removing some logistics problem and provide rapid information to guide timely management.\n\nWe discuss the role of focused airway ultrasound in upper airway trauma performed by point-of-care ultrasound trained emergency physician and propose a focused airway ultrasound classification in relation to Schaefer classification of laryngeal injury.\n\nCase presentation {#Sec2}\n=================\n\nCase 1---endolaryngeal hematoma without detectable fracture (Schaefer group 1) {#Sec3}\n------------------------------------------------------------------------------\n\nA 24-year-old male presented with neck swelling without signs of respiratory distress after a traumatic blunt neck injury. There was swelling of the anterior neck without palpable crepitus. Airway ultrasound showed disruption of the air--mucosal interface suggesting endolaryngeal disruption (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b). CT scan confirmed the diagnosis of endolaryngeal disruption without cartilaginous fracture. The patient was conservatively managed and discharged well on the third day.Fig.\u00a01**a** Surface landmark of probe position (in longitudinal view) in relation to scanning the area for **b** and **c**. **b** Normal sonoanatomy of upper airway showing the relationship between thyroid cartilage, cricoid cartilage, air--mucosal interface and surrounding soft tissues. A continuous and undisrupted line of air--mucosal interface seen as hyperechoic line (arrowhead) **c** Airway ultrasound showed disruption air--mucosal interface without obvious detectable laryngeal fracture (circle). SM: sternocleidomastoid muscle; TC: thyroid cartilage; CC: cricoid cartilage\n\nCase 2---undisplaced thyroid cartilage fracture (Schaefer group 2) {#Sec4}\n------------------------------------------------------------------\n\nA 66-year-old motorcyclist, was injured in a collision with a van. He presented with mild neck pain, difficulty in breathing, hoarseness, dysphagia and odynophagia. He had stridor, and his neck was swollen and tender with subcutaneous emphysema.\n\nAirway ultrasound using a 15-MHz linear transducer found discontinuity of the anterior cortex of thyroid cartilage with minimal surrounding tissue edema, consistent with Schaefer group 2 (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}c). CT scan confirmed the ultrasound findings, showing a defect in the posterolateral wall of the trachea and the esophagus, with fracture of the right anterior lamina of the thyroid cartilage and superior cornu of the left thyroid cartilage (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}d).Fig.\u00a02**a**, **b** Surface landmark of the probe position and ultrasound image of the normal thyroid cartilage (in transverse view). **c** Airway ultrasound showing undisplaced fracture of the thyroid cartilage (box) and disruption of the anterior cortex of the thyroid cartilage (arrowhead**). d** A Computerized tomography (CT) scan image shown right thyroid lamina fracture with surrounding prevertebral edema and hematoma. SM: sternocleidomastoid muscle; TC: thyroid cartilage\n\nHe was immediately intubated and was started on intravenous dexamethasone to reduce inflammation and edema, a proton pump inhibitor to prevent reflux and laryngeal irritation, nebulized adrenaline and a prophylactic antibiotic in the emergency department. The patient was managed conservatively and was discharged well from intensive care unit on the fifth day post trauma.\n\nCase 3---displaced thyroid cartilage fracture (Schaefer group 3) {#Sec5}\n----------------------------------------------------------------\n\nA 28-year-old male martial art athlete was kicked by his opponent and sustained a blow to the anterior part of the neck. He complained of pain, dysphagia and hoarseness. There was an abrasion to the anterior part of his neck, which was tender to palpation with localized crepitus.\n\nBedside airway ultrasound revealed a displaced fracture of the thyroid cartilage, disruption of anterior cortex of thyroid cartilage with surrounding mixed echogenicity denoting endolaryngeal edema (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a and Additional file [1](#MOESM1){ref-type=\"media\"}: Video S1) and paralyzed right vocal cord (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}e), consistent with Schaefer group 3.Fig.\u00a03**a** Image of focused airway of displaced thyroid cartilage fracture and disruption of anterior cortex of the thyroid cartilage (box). **b** A computerized tomography (CT) scan image showing defect in posterolateral wall of trachea with fracture of right anterior lamina of thyroid cartilage and superior cornu of left thyroid cartilage. **c**, **d** Assessment of vocal cord mobility can be done by looking at the movement of vocal ligament (white line) during abduction and adduction. **e** Blue arrow indicates reduced movement of the right vocal ligament (white line) to the midline during adduction compare to the left vocal ligament. R: right; L: left; VL: vocal ligament, TC: thyroid cartilage\n\nDirect visualization using a flexible fibreoptic scope revealed an edematous and medially deviated right arytenoid with paralyzed and erythematous right vocal cord. He was intubated and intravenous dexamethasone, proton pump inhibitor, nebulized adrenaline and prophylactic antibiotic were initiated early in the emergency department. The patient was sent for CT neck after stabilization, which showed a defect in the posterolateral wall of the trachea with a displaced fracture of right anterior lamina of thyroid cartilage and superior cornu of left thyroid cartilage, consistent with Schaefer group 3 (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}b), and that found on bedside ultrasound. The patient was stable throughout his entire hospitalization after open surgical repair and was allowed home on day nine with outpatient follow up.\n\nCase 4---displaced thyroid cartilage fracture (Schaefer group 3) {#Sec6}\n----------------------------------------------------------------\n\nA 35-year-old male lorry driver hit his neck against the steering wheel when he thrown forwards during a head-on collision. He presented with neck pain, severe swelling over the whole anterior region of the neck, stridor, hypoxia and a compromised airway. The patient was immediately intubated and ventilated.\n\nFocused airway ultrasound showed disruption of the air--mucosal interface, a displaced thyroid cartilage fracture with formation of endolaryngeal hematoma and a cricoid cartilage fracture (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}b). He was treated as Schaefer group 3 and was started on intravenous dexamethasone, proton pump inhibitor, nebulized adrenaline, prophylactic antibiotic and open surgical repair was planned.Fig.\u00a04**a** Normal sonoanatomy in longitudinal scan showing continuous and intact air--mucosal interface (arrowhead) in relation to thyroid and cricoid cartilage. **b** Airway ultrasound image in longitudinal scan showing disruption of the air--mucosal interface (arrowhead) and formation of endolaryngeal hematoma and fracture cricoid cartilage (box). TC: thyroid cartilage; CC: cricoid cartilage\n\nCT scan revealed a defect in the posterolateral wall of the trachea with a displaced fracture of the left anterior lamina of thyroid cartilage and hematoma surrounding the thyroid and cricoid cartilages. This confirmed the classification and injury details categorized under Schaefer group 3, which correlated with that of ultrasound assessment. Hospital stay was uneventful, and he went home after 2\u00a0weeks, to be reviewed in the outpatient department.\n\nDiscussion {#Sec7}\n==========\n\nThe ABCs of trauma, a mantra that prioritizes the primary survey starts, with \"A\" for airway. In this case series, we describe four encounters in which ultrasound of the upper airway was performed for suspected laryngeal trauma and correlated with CT scan assessment of severity based on the Schaefer classification of laryngeal injury.\n\nUpper airway obstruction as a consequence of laryngeal injury may be catastrophic. Apart from blunt force trauma, iatrogenic injuries can occur after percutaneous dilatational tracheostomy, fiberoptic bronchoscopy, airway manipulations and procedures, even tracheal intubation \\[[@CR6], [@CR7], [@CR8]\\]. The lack of correlation between symptoms, physical findings and severity of laryngeal injury may result in delayed recognition of such injuries. Additionally, patients with laryngeal injury are at risk of false passages, transforming an incomplete fracture to total separation of the upper airway, converting mild upper airway obstruction to complete obstruction especially in undiagnosed laryngeal trauma \\[[@CR9]--[@CR11]\\]. For these reasons upper airway ultrasound may play an important role in the early assessment for laryngeal injury.\n\nFor the past three decades the internationally accepted Schaefer classification of laryngeal injury stratification system has been used to categorize laryngeal injury. This classification not only categorizes, but it also guides management. It divides the management plan into 2 categories; non-invasive or conservative airway management for group 1 and group 2 injuries, and invasive airway management for higher grade injuries (group 3 to group 5). Further study is needed to determine if focused upper airway ultrasound can reliably be used to determine injury grade \\[[@CR1]\\].\n\nWhile the authors could not find previous publication of ultrasound assessment for upper airway injury in trauma, prior research by Osman et al. and You-Ten et al. briefly illustrated the usage of airway ultrasound in a step-by step manner to delineate the normal sonoanatomy of the upper airway such as thyroid cartilage, epiglottis, cricoid cartilage, cricothyroid membrane, tracheal cartilages, esophagus and the surrounding soft tissues \\[[@CR3]--[@CR5]\\]. Schick et al. published promising evidence on the use of airway ultrasound in the emergency setting to identify airway edema and impending threats to the airway \\[[@CR12]\\]. Airway ultrasound can also be used to assess laryngeal edema in the post-extubation period. \\[[@CR13]--[@CR15]\\]. Kameda et al. \\[[@CR16]\\] identified airway edema as hypoechoic thickening of the tracheal wall on airway ultrasound in a patient with inhalational burns. The findings on the sonogram were later confirmed by CT scan, demonstrating good correlation between focused ultrasound and CT Scan.\n\nCheng et al. \\[[@CR17]\\] found good correlation between sonographic visualization of abnormal vocal cords movement and laryngoscopic examination. They demonstrated that clinician-performed airway ultrasound is an accurate screening tool for preoperative assessment of vocal cord movement.\n\nUpper airway ultrasound findings that correlate with the Schaefer Classification System may be especially relevant in hemodynamically unstable patients where CT imaging is not feasible. While larger trials are needed, we propose that focused airway ultrasound can be used to correlate with the Schaefer Classification System (Table\u00a0[1](#Tab1){ref-type=\"table\"}) and therefore propose it be assimilated into the work-up of laryngeal trauma.Table\u00a01Proposed focused airway ultrasound findings in correlation to the Schaefer Classification System and standard management of laryngeal injuryGroup\\\n-Based on Schaefer classification \\[[@CR1]\\]CT scan findings\\\n-Based on Schaefer classification \\[[@CR1]\\]Focused airway ultrasound findingsStandard management and intervention \\[[@CR1], [@CR2]\\]Group 1Minor endolaryngeal hematoma or laceration without detectable fractureEndolaryngeal hematoma without detectable fractureSupportive care including observation, antibiotics, humidified air, supplemental oxygen, anti-reflux medications, voice rest and early steroid administration.\\*Patients with Group 2 injuries should be serially examined, since the injuries may worsen or progress with time. Occasionally group 2 injuries may require a tracheotomyGroup 2Edema, hematoma, minor mucosal disruption without exposed cartilage, nondisplaced fracture noted on CTEdema, endolaryngeal hematoma, minor mucosal disruption without exposed cartilage, nondisplaced fractureMucosal hematoma/edemaNondisplaced fracture of cartilage frameworkGroup 3Massive edema, mucosal tear, exposed cartilage, cord immobility, displaced fractureEdema, cord immobility, displaced fractureVocal fold immobilityObvious displaced fractureDirect laryngoscopy, esophagoscopy and immediate open surgical repair is deemed necessary due to extension of injuriesGroup 4Addition of more than two fracture lines or massive trauma to laryngeal mucosaAddition of more than two fracture linesComminuted fracture of laryngeal cartilage frameworkGroup 5Complete laryngeal separation\n\nPossible advantages of focused upper airway ultrasound in the diagnostic classification of blunt laryngeal trauma\u00a0are:In centers without CT scan capabilities---focused airway ultrasound can complement emergency department triage protocol to enable early airway management planning in blunt laryngeal injury.In centers with CT scan facility---focused airway ultrasound can hasten airway management planning prior to airway catastrophe during an emergency situation when transfer to the radiology suite is deemed unsuitable.In resource-limited situation such as---scarce resources, remote area, and humanitarian medical mission in environmental disasters and war-torn regions---focused airway ultrasound can supplement disaster and transfer protocol facilitating decision for early airway intervention while preparing for emergency transfer.It has the added advantage of real-time visualization of dynamic vocal cords function.\n\nLimitations {#Sec8}\n-----------\n\nUltrasound evaluation of the airway may not be practical in every case. Subcutaneous emphysema, posterior laryngeal injury, cartilage calcification and foreign bodies may result in artifacts interfering with ultrasound images and interpretation. Furthermore, ultrasound techniques and interpretation are operator-dependent, and have a steep learning curve. Adequate competency training and reproducibility is important to standardize findings.\n\nFuture directions {#Sec9}\n-----------------\n\nFurther studies to identify optimal management strategies for patients with laryngeal injury are required. Areas of interest include:Validation studies comparing accuracy of focused airway ultrasound to CT scan findings across a range of injury types.Management outcome, cost effectiveness, accuracy and time of diagnosis of focused airway ultrasound compared to existing radiological modalities.Reproducibility of results by different operators at different stages of seniority and proficiency level, their abilities to accurately detect pathology and studies on learning curve of this procedure.\n\nConclusion {#Sec10}\n==========\n\nUltrasound assessment of the upper airway is a promising adjunct in the rapid evaluation of patients with suspected laryngeal trauma. Early diagnosis and injury classification stratification with point-of-care ultrasound may play an important role in trauma patient care, particularly those too unstable for CT imaging or when advanced imaging is unavailable.\n\nSupplementary information\n=========================\n\n {#Sec11}\n\n**Additional file 1: Video S1.** Bedside airway ultrasound revealed a displaced fracture of the thyroid cartilage, disruption of anterior cortex of thyroid cartilage with surrounding mixed echogenicity denoting endolaryngeal edema.\n\nPOCUS\n\n:   Point-of-care ultrasound\n\nCT\n\n:   Computed tomography\n\nSm\n\n:   Sternocleidomastoid muscle\n\nTc\n\n:   Thyroid cartilage\n\nVm\n\n:   Vocalis muscle\n\nVL\n\n:   Vocalis ligament\n\nAC\n\n:   Arytenoid cartilage\n\n**Publisher\\'s Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary information\n=========================\n\n**Supplementary information** accompanies this paper at 10.1186/s13089-020-00186-3.\n\nWe would like to thank Ms. Anusha Bala, Cheong Chee Keong, Tan Wan Chuan, Lai Si Qi and World Integrated Network for Focused Ultrasound (WINFOCUS) Malaysia, and Society of Critical Ipoh Emergency Medicine Society (IEMS) and Emergency Sonography (SUCCES) for their assistance.\n\nDeclaration {#FPar1}\n===========\n\nI declare that this manuscript which depicts the clinical management of patient with laryngeal trauma. Contributions from respective authors have been explicitly mentioned in the respective segment. This work has not been submitted to any other publication for publishing.\n\nOA and KMS: primary author, involved in the management of case, and drafting, reviewing, editing, preparing, and final approval of the manuscript. AHA, MAW and LN: involved in the management of case, reviewing, editing, preparing, and final approval of the manuscript. NP: involved indirectly in reviewing, editing, and preparing the manuscript. All authors read and approved the final manuscript.\n\nAuthors received no funding for the case report from any institution/individual.\n\nThe material during the current case series is available from the corresponding author on reasonable request\n\nOurs is a retrospective report of a clinical event; therefore, ethical approval and consent to participate are not relevant.\n\nA signed written informed consent was taken.\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "On page 929, Ashton et al. report that an intracellular parasite regulates its own growth and exploits the inflammatory environment within its host to continue its survival and mediate long-term damage.\n\nFigure 1*T. cruzi* (blue dots) thrives in cells lacking (bottom) TXA~2~ receptors.\n\n*Trypanosoma cruzi*, the bug that causes Chagas\\' disease, initiates a short-lived acute infection in humans. A third of those infected, however, develop chronic cardiac disease that sets in after a long asymptomatic period. The mechanism by which the parasite facilitates this long-term pathology is still unclear.\n\nAshton et al. now show that the parasite infects vascular endothelial cells and secretes a bioactive lipid called thromboxane (TXA~2~). When produced by human cells, TXA~2~ has pro-inflammatory effects and can cause cardiac injury by triggering platelet aggregation, clotting, and vasoconstriction.\n\nThe parasitic TXA~2~ also promotes cardiac injury, the team finds, yet it somehow mediates an anti-inflammatory immune response in the host. Mice that lack cell surface receptors for TXA~2~ had increased inflammatory damage.\n\nThese mice also had a higher parasite load than wild-type mice, suggesting that the parasite suppresses its own replication by a negative feedback through the TXA~2~ receptor. The group is now dissecting TXA~2~ receptor signaling within infected cells to determine how the parasite controls its own growth rate.\n\nThe team speculates that, within infected cells, the parasite uses TXA~2~ to limit its own growth and minimize cell damage, thus ensuring its survival. But parasite-secreted TXA2 also engages receptors on uninfected host cells, which might eventually cause the heart disease seen in chronic infections. ![](jem_iti_end.jpg)\n\n[^1]: <hbashyam@rockefeller.edu>\n"}
+{"text": "Introduction {#sec1}\n============\n\nThe detection of biologically and environmentally related heavy metal ions (Cd^2+^, Hg^2+^, Cu^2+^, etc.) is crucial due to their important roles in biological processes and hazard to the environment.^[@ref1]^ Among various heavy metals, copper is an essential trace element for living systems. The excess of unbound Cu^2+^, the natural form of free copper in biological media, is quite harmful and can even cause various neurodegenerative diseases. Meanwhile, Cu^2+^ may be a common pollutant of drinking water that affects water quality and leads to multifarious infections.^[@ref2],[@ref3]^ Until now, the approaches for Cu^2+^ ion detection reported are inductively coupled plasma atomic emission spectroscopy (ICP-AES), atomic absorption spectroscopy, inductively coupled plasma mass spectroscopy (ICP-MS), and voltammetry. However, these approaches are time consuming, expensive, and quantitative; thus, it is necessary to develop a new method for the detection of Cu^2+^ ions.\n\nFor the past few years, the development of a range of smart fluorescence materials including transition-metal complexes,^[@ref4],[@ref5]^ organic dyes,^[@ref6]^ inorganic semiconductor nanocrystals,^[@ref7],[@ref8]^ carbon dots,^[@ref9],[@ref10]^ and lanthanide-doped upconverting nanoparticles^[@ref11],[@ref12]^ have attracted particular attention. These fluorescent compounds have great demand in optoelectronics, lighting, and so forth;^[@ref13],[@ref14]^ especially, fluorescent probing or sensing is now becoming an effective detection modality. Fluorescence detection is a highly versatile spectroscopic method with a short response period, high sensitivity, technical simplicity, and extensive applicability, which can be widely used at real time and in situ.^[@ref15]^ Therefore, it is a vital challenge to explore a robust fluorescent material with excellent sensitivity, selectivity, and a rapid response time for Cu^2+^ ions.\n\nPolyoxometalates (POMs) are a well-known class of environment friendly and structurally well-defined nanosized metal-oxide clusters with significant applications in optics, catalysis, electronics, magnetics, and biomedicine.^[@ref16]\u2212[@ref23]^ Especially, lanthanide (Ln) group element-doped polyoxometalates are of great importance in the fluorescence field because of their excellent photoluminescence properties, such as narrow emission bands, long lifetime, large Stokes shift, tunable emission, and abundant valence electrons in f orbitals.^[@ref24],[@ref25]^ Thus, Ln^3+^-doped nanoscale POM composite materials are suitable as fluorescent sensors to detect heavy metal ions.^[@ref26]^\n\nAlthough lanthanide-containing POMs (such as Na~9~\\[EuW~10~O~36~\\]\u00b732H~2~O, abbreviated to EuW~10~) have excellent fluorescent properties,^[@ref27]^ the biggest problem is the quenched emission in water solution due to the ubiquitous O--H vibrations.^[@ref28],[@ref29]^ To solve this problem, using the concept of self-assembly to suppress such quenching is a powerful methodology. Self-assembly is one of the advanced nanotechnologies used to design novel nanostructures, which is driven by noncovalent interactions such as hydrogen bonding, electrostatic interaction, hydrophobic interaction, steric effect, van der Waals force, and \u03c0--\u03c0 stacking,^[@ref30]\u2212[@ref33]^ and can efficiently enhance the fluorescence property. Thus, it is attractive to improve the fluorescent properties and construct novel supramolecular nanostructures of lanthanide-containing POMs through the self-assembly of POMs with other components.\n\nWith these issues in mind, we reported a new aggregate that comprises Weakley-type lanthanide-containing POMs (EuW~10~) and tetra-*n*-alkylammonium (TA), which form well-defined hybrid nanoparticles with enhanced emission. Moreover, our results further demonstrate that the fluorescent nanoparticles showed a pH-responsive behavior and can selectively detect Cu^2+^ in water with the detection limit of 0.15 \u03bcM, suggesting the potential of the fluorescence materials for optoelectronic applications and sensing device fabrication.\n\nResults and Discussion {#sec2}\n======================\n\nPhase Behavior of the EuW~10~/TA System {#sec2.1}\n---------------------------------------\n\nFirst, the complexes of EuW~10~ with four different chain lengths of tetra-*n*-alkyl ammonium (TA) (*n* = 1, 2, 4, 6) were constructed. The concentration of EuW~10~ is maintained at 0.6 mM, whereas that of TA is gradually increased. [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}f displayed the phase behavior of EuW~10~/TA systems as a function of *c*~TA~. It can be observed that all of them share the same phase behaviors in that the solution phase gradually turned to the precipitate phase with increasing \u2212*c*~TA~, which may be induced by the electrostatic interaction of the oppositely charged EuW~10~ and TA. Nevertheless, the difference between them was that the concentration of TA, the inflection point from solution to precipitate, was going down with the increase of chain lengths. It could be explained by the reason that TA with a longer alkyl chain owns poor flexibility and higher hydrophobicity, which induced the faster appearance of the precipitate.\n\n![Structures of (a) EuW~10~, (b) tetramethylammonium bromide (TMAB), (c) tetraethylammonium bromide (TEAB), (d) tetrabutylammonium bromide (TBAB), and (e) tetrahexylammonium bromide (THAB). (f) Phase behaviors of EuW~10~/TA system as a function of *c*~TA~ at 0.6 mM EuW~10~ for different chain lengths of tetra-*n*-alkyl ammonium.](ao-2018-01636r_0002){#fig1}\n\nCharacterization and Analysis of EuW~10~/TMAB Nanostructures {#sec2.2}\n------------------------------------------------------------\n\nFor the pellucid EuW~10~/TMAB solution, there was a typical Tyndall phenomenon when a laser goes through the solution, indicating the existence of aggregates. Hence, taking the sample of 0.6 mM EuW~10~/30 mM TMAB as an example, we carried out electron microscopic research to study the morphology of the aggregates in the EuW~10~/TMAB system. Nanoparticles with diameters varying from 90 to 300 nm were clearly detected in the transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}a,b. The structure of the nanoparticles was clearer with atomic force microscopy (AFM) analyses ([Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}c), which implied that their corresponding diameters are larger than their vertical heights (20--150 nm) by three to five times, indicating an oblate sphere structure. Besides, the Fourier transform infrared (FT-IR) spectra of EuW~10~, EuW~10~/TMAB, and TMAB were analyzed to explain the atomic interaction vibration, and the results are shown in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}d. The bands at around 2960, 2914, and 1486 cm^--1^ are assigned to asymmetric stretching vibrations, symmetric stretching vibrations, and C--H bending vibrations of the methyl groups in TMAB, respectively.^[@ref35],[@ref36]^ The bands in the low-frequency region are attributed to the vibrations of EuW~10~ ([Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}e); to be specific, the characteristic vibration bands are listed as follows: \u03bd (W = O~d~, 943 cm^--1^), \u03bd (W--O~b~--W, 844 cm^--1^), and \u03bd (W--O~c~--W, 784/703 cm^--1^), where O~b~ represents the bridged oxygen of two octahedra with a shared corner, O~c~ is the bridged oxygen of two octahedra with an edge shared, whereas O~d~ is the terminal oxygen.^[@ref37]^ These peaks moved to 921, 833, 789, and 671 cm^--1^ for the complex after the assembly with TMAB, confirming that electrostatic interaction and hydrogen bonding may be the main driving forces for successful hybridation.^[@ref38]^\n\n![(a) TEM, (b) SEM, and (c) AFM images of 0.6 mM EuW~10~/30 mM TMAB nanoparticles; the scale bar of images inside (a) and (b) is 100 nm. (d) FT-IR spectra of TMAB, EuW~10~, and 0.6 mM EuW~10~/30 mM TMAB complex; (e) partially enlarged (500--1100 cm^--1^) spectra of (d). (f) Zeta potential (\u03b6) of the aqueous solution with different *c*~TMAB~, while *c*~EuW~10~~ = 0.6 mM.](ao-2018-01636r_0003){#fig2}\n\nTo further verify the existence of electrostatic interaction, the \u03b6-potential of the EuW~10~/TMAB system was measured to further study the evolution of the assembly behavior. The results in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}f revealed that the \u03b6-potential values of EuW~10~/TMAB system increased from \u221235.6 to \u221211.2 mV when the concentration of TMAB increased. The increasing trend of the \u03b6-potential confirmed that electrostatic interaction between EuW~10~ and TMAB was the dominant interaction of self-assembly. Moreover, the negative value of \u03b6 indicated that the interaction of POMs and TA is weaker than that with other surfactants, and thus the cationic TA did not displace all countercations of the EuW~10~, and the EuW~10~ was located in the shell of the nanoparticles' structure.^[@ref39]^ Moreover, as the UV--vis absorption spectra show ([Figure S1A](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)), the absorption peak value at 260 nm of EuW~10~/TMAB decreased compared with pure EuW~10~, which further proved the formation of small and dense aggregates.^[@ref40]^\n\nFluorescence Behavior of the EuW~10~/TMAB System {#sec2.3}\n------------------------------------------------\n\nPOMs with rare earth elements doped have more prominent photoluminescent advantages than other inorganic fluorescent nanomaterials.^[@ref41]^ In our system, although EuW~10~ powder has excellent fluorescence property in the solid state, the emission of EuW~10~ in aqueous solution is quite weak. Hence, it would be significant if the fluorescence of the EuW~10~/TMAB system can be improved compared with pure EuW~10~ aqueous solution. Then, taking the EuW~10~/TMAB system as an example, we conducted a series of measurements to study its fluorescence performance. The photographs of 0.6 mM EuW~10~ aqueous solution and EuW~10~/TMAB with varied *c*~TMAB~ under 254 nm UV light are shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}A; it can be found by the naked eye that the luminescence of EuW~10~/TMAB enhanced obviously as *c*~TMAB~ increased gradually, and the confocal laser scanning microscopy (CLSM) image ([Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}B) exhibiting red shining dots further verified the great fluorescence performance of the EuW~10~/TMAB system.\n\n![(A) Photographs of aqueous solutions of ratio-dependent complexes with EuW~10~ fixed at 0.6 mM under 254 nm UV irradiation (from a to g, *c*~TMAB~/*c*~EuW~10~~ = 0, 1, 9, 20, 30, 40, and 50). (B) CLSM image of 0.6 mM EuW~10~/30 mM TMAB. (C) Fluorescence spectra and (D) the corresponding intensity ratios of 594--621 nm (*I*~594~/*I*~621~) of EuW~10~ (0.6 mM) with different concentrations of TMAB. (E) Time-resolved fluorescence decay curves of EuW~10~ (0.6 mM) before and after binding with TMAB. The experiment was performed at 25 \u00b0C, and the samples were excited at 265 nm.](ao-2018-01636r_0004){#fig3}\n\n[Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}C presents the emission spectra of EuW~10~ and EuW~10~/TMAB in aqueous solution with excitation wavelengths at 265 nm. The emission spectrum of pure EuW~10~ aqueous solution displayed four main characteristic transitions of Eu^3+^ ions from 570 to 720 nm and were ascribed to ^5^D~0~ \u2192 ^7^F*~j~* (*j* = 1, 2, 3, 4) transitions at 594, 621, 651, and 699 nm, respectively, caused by intramolecular energy transfer from the ligand-to-metal charge transfer (LMCT) band of O \u2192 W to the photoluminescent Eu^3+^ core, and the emission originated from the ^5^D~0~ excited states relaxing to the ^7^F*~j~* ground state eventually.^[@ref42],[@ref43]^ As the TMAB was added to the EuW~10~ solution, the fluorescence intensity enhanced significantly, whereas the emission bands remained unchanged, corresponding to the phenomenon observed under UV lamp. It is worth noting that the maximal luminescence of the EuW~10~ composite was observed in the sample with 30 mM TMAB, and the intensity of the ^5^D~0~ \u2192 ^7^F~1~ transition band increased by about 14 times compared with the discrete EuW~10~ solution. Hence, the remarkably enhanced fluorescence demonstrated that for the EuW~10~/TMAB system, EuW~10~ was located in a comparatively hydrophobic microenvironment where the cationic TMAB exhibited powerful affinity to the anionic EuW~10~ to take the place of water ligands by electrostatic interaction,^[@ref44]^ and thus enhanced the energy of the LMCT to Eu^3+^.^[@ref45]^\n\nIn addition, to the best of our knowledge, the relative intensity ratio of ^5^D~0~ \u2192 ^7^F~1~ transition to ^5^D~0~ \u2192 ^7^F~2~ transition is often applied to assay the symmetry variation degree of the Eu^3+^ coordination microenvironment of different systems.^[@ref46]^ As shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}D, the ratio of *I*~594~/*I*~621~ for EuW~10~ alone was calculated to be 0.87, whereas for the EuW~10~/TMAB mixed system, it improved to 1.62 at 30 mM TMAB, suggesting that the Eu^3+^ in complexes is located at a higher symmetry microenvironment.^[@ref47]^ What's more, the time-resolved fluorescent spectra of EuW~10~ before and after binding with TMAB were analyzed to gain an in-depth insight into the mechanism of fluorescence enhancement. [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}E illustrates the luminescence decay curves of the fluorescence intensity monitored at 594 nm, which was a biexponential decay and showed two different lifetimes and proportions as summarized in [Table S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf).^[@ref48]^ It can be found that the measured lifetime was gradually longer in the wake of the increase of the TMAB proportion. Compared with pure EuW~10~ (\u03c4~ave~ = 2.051 ms), the longest average lifetime of EuW~10~/TMAB (1:50) reached 2.582 ms, and these observed results confirmed the formation of compact self-assemblies because of strong electrostatic interactions between TMAB and EuW~10~, and replacing the water ligands of Eu^3+^ by TMAB is the main contributor to the emission enhancement of EuW~10~.^[@ref49]^\n\nInfluencing Factors for the Fluorescence Behavior of the EuW~10~/TA System {#sec2.4}\n--------------------------------------------------------------------------\n\nThe factors of alkyl chain lengths, amino group, and inorganic salt have a certain effect on the fluorescence behavior. Replace TMAB to TEAB and TBAB with longer alkyl chain lengths, the emission intensity decreased gradually as shown in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}A,B. The lower luminescence enhancement for longer alkyl chains suggested weaker electrostatic interaction and looser aggregation of the EuW~10~/TA system. The corresponding ratio of *I*~594~/*I*~621~ for EuW~10~/TA also decreased and the measured lifetime became shorter as the alkyl chain lengths increased ([Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}C,D). In addition, the relevant electrical conductivity and \u03b6-potential values in [Figure S2](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf) decreased clearly. All of these demonstrated that the Eu^3+^ in a longer alkyl chain system was located in a lower-symmetry microenvironment. This may be because the steric hindrance effect as well as nonpolarity of alkyl chains increase as the cationic alkyl chain length increases, which hinders the electrostatic interaction between EuW~10~ and TA and induces less fluorescence enhancement consequently.^[@ref35],[@ref50]^ Above all, the formation mechanism of EuW~10~/TA nanoparticles in the presence of different alkyl chain lengths of TA are summarized in [Scheme [1](#sch1){ref-type=\"scheme\"}](#sch1){ref-type=\"scheme\"}.\n\n![(A) Photographs under 254 nm UV irradiation, (B) fluorescence spectra, (C) the corresponding intensity ratios of 594--621 nm (*I*~594~/*I*~621~), and (D) time-resolved fluorescence decay curves of EuW~10~ (0.6 mM) with 30 mM TA of different chain lengths. (E) Fluorescence spectra of 0.6 mM EuW~10~/30 mM TMAB and 0.6 mM EuW~10~/30 mM TMPB. (F) Fluorescence spectra of EuW~10~ (0.6 mM), 0.6 mM EuW~10~/30 mM TMAB before and after binding with NaBr.](ao-2018-01636r_0005){#fig4}\n\n![Formation Mechanism of EuW~10~/TA Nanoparticles in the Presence of Various Alkyl Chain Lengths of TA](ao-2018-01636r_0008){#sch1}\n\nWhat's more, the effects of the amino group and inorganic salt on the fluorescence behavior were also studied. As shown in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}E, the luminescence of EuW~10~/TMPB had increased in comparison with pure EuW~10~ solution, but the intensity was much lower than that of EuW~10~/TMAB. It illustrated the existence of the more ordered arrangement between N^+^ and EuW~10~ than P^+^ and EuW~10~ in the self-assembly atop the active layer.^[@ref51]^ As for the sensitiveness of EuW~10~/TMAB system to ionic strength, it can be seen that for the same ratio of EuW~10~/TMAB and EuW~10~/NaBr solution, the latter emission intensity is lower, but is higher than that of pure EuW~10~ solution ([Figure S3A,B](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)). However, when NaBr was added to the EuW~10~/TMAB system, the fluorescence intensity became weaker than that of EuW~10~/TMAB ([Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}F), which proved that the increase of ionic strength destroyed the original assemblies. This is because the added salt might compress or shield the electrical double layer and weaken the electrostatic interaction of EuW~10~ and TMAB.^[@ref24]^\n\nResponse to pH of the EuW~10~/TMAB System {#sec2.5}\n-----------------------------------------\n\nStimuli-responsive POM-based supramolecular materials with multiple functions like reversible photoluminescence and self-assembly behavior in aqueous solution will broaden the potential applications of these materials. In case of the 0.6 mM EuW~10~/30 mM TMAB system (pH = 6.3), the fluorescent property at different pH adjusted by HCl or NaOH was also analyzed. The fluorescence of the system could be completely quenched at pH 7.4 and then recovered at pH 6.3 ([Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}a); in the meantime, the phase behavior transfers from clear solution to precipitate with pH from 6.3 to 7.4 and this procedure can reverse over four cycles with only little loss of fluorescence intensity ([Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}b). The morphological evolution by alternating the pH was also investigated in detail. When the pH increased to 7.4, the nanoparticles disassembled into a disordered network structure ([Figure S4a,b](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)), whereas the nanoparticles rebuilt when the pH of the system recovered to 6.3. It has been proved previously that TMAB acts as the cation and can interact with the EuW~10~ cluster through electrostatic interaction; thus, upon treatment with NaOH, the deprotonated amino group weakened the electrostatic interaction between EuW~10~ and TMAB and induced the breakout of nanoparticles.^[@ref52]^\n\n![(a) Photographs of 0.6 mM EuW~10~/30 mM TMAB at different pH values under daylight (left) and 254 nm UV irradiation (right). (b) Plot of the fluorescence intensity at \u03bb~em~ = 594 nm of the sample upon four cycles of alternating pH between 6.3 and 7.4 at an excitation wavelength of 265 nm.](ao-2018-01636r_0006){#fig5}\n\nFluorescent Detection of Cu^2+^ {#sec2.6}\n-------------------------------\n\nLanthanide-containing polyoxometalates are sensitive to ambient chemical environments and widely used in luminescent sensing or probing.^[@ref53]^ In our work, the strong fluorescence intensity of the Eu^3+^ ions is hypersensitive to the microenvironment. Thereby, a series of metal ions (Ca^2+^, Mg^2+^, Ba^2+^, Al^3+^, Pb^2+^, Cd^2+^, Zn^2+^, Fe^3+^, Cu^2+^) were added to the EuW~10~/TMAB aqueous solution to explore the possibility of fluorescence detection. It is obvious from the fluorescence spectra ([Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"}A) that only Cu^2+^ led to drastic quenching of the fluorescence (95%), which suggested that the nanoparticles can selectively detect Cu^2+^ ions. Aiming to assess the performance of the sensor, the selectivity for Cu^2+^ over other metal ions was also studied by competition experiments in the presence of 30 \u03bcM Cu^2+^ mixed with other equimolar interfering metal ions ([Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"}B). The results indicated that Cu^2+^ still induced strong quenching of EuW~10~/TMAB in the presence of other tested metal ions.\n\n![(A) Fluorescence spectra of 0.6 mM EuW~10~/30 mM TMAB with different metal ions (Ca^2+^, Mg^2+^, Ba^2+^, Al^3+^, Pb^2+^, Cd^2+^, Zn^2+^, Fe^3+^, Cu^2+^, and *c*~metal~ = 30 \u03bcM). (B) Selectivity of 0.6 mM EuW~10~/30 mM TMAB toward Cu^2+^: gray bars denote the solutions with metal ions and pink bars denote the succedent mix of Cu^2+^ to the above solutions (\u03bb~ex~ = 265 nm, \u03bb~em~ = 594 nm, *c*~metal~ = 30 \u03bcM, *c*~Cu^2+^~ = 30 \u03bcM). (C, D) Fluorescence spectra of the sample with different contents of Cu^2+^.](ao-2018-01636r_0007){#fig6}\n\nTo further study the sensitivity of the fluorescence sensor for Cu^2+^ ions, the fluorescence spectra as a function of Cu^2+^ ion concentration were measured. In [Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"}C, the fluorescence intensity of Eu^3+^ ions gradually decreases with the increase of *c*~Cu^2+^~, which can be found by the naked eye under UV light (inset of [Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"}C). An approximately linear relationship between fluorescence intensity (\u03bb~em~ = 594 nm) and *c*~Cu^2+^~ can be obtained when *c*~Cu^2+^~ is in the range of 0.2--1.0 \u03bcM ([Figure [6](#fig6){ref-type=\"fig\"}](#fig6){ref-type=\"fig\"}D). Thus, the detection limit was calculated to be ca. 0.15 \u03bcM for Cu^2+^ through the 3\u03c3 method, which was much lower than the EPA-defined safety level of Cu^2+^ in drinking water (20 \u03bcM).^[@ref54]^ Moreover, the 1/(*I*~0~ -- *I*) value as a function of 1/\\[Cu^2+^\\] displayed a wonderful linear fit (*R* = 0.998) based on the emission intensity changes, and the binding constant (*K*) calculated according to the Benesi--Hildebrand equation was close to 4.65 \u00d7 10^3^ M^--1^ ([Figure S5](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)).^[@ref55]^\n\nTo explore the nature (static or dynamic) of the fluorescence quenching of Cu^2+^, the relationship of the plot of *I*~0~/*I* (*I*~0~ and *I*, fluorescence intensity of EuW~10~/TMAB with and without Cu^2+^) with Cu^2+^ concentration was studied and it showed a good linear relationship (*R*^2^ = 0.9815) ([Figure S6](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)). Based on the modified Stern--Volmer relationship (*I*~0~/*I* = *K*~SV~ \\[Cu^2+^\\] + *C*) in which *K*~SV~ is the Stern--Volmer's constant and \\[Cu^2+^\\] is the concentration of Cu^2+^, the dynamic nature of the fluorescence quenching in our system can be proved.^[@ref56]^ Hence, the mechanism of the fluorescence quenching was speculated to be energy transfer.^[@ref57]^ Then, the reversibility of the quenching upon addition of ethylenediaminetetraacetic acid (EDTA) solution (2 mM, 30 \u03bcL) to the Cu^2+^-quenched EuW~10~/TMAB solution was checked. The initial fluorescence can recover in the presence of EDTA solution, suggesting the formation of a strong complex of Cu^2+^ with EDTA. Besides, TEM and FT-IR of EuW~10~/TMAB before and after the addition of Cu^2+^ were further compared to investigate the mechanism of the fluorescence detection of Cu^2+^ ions ([Figure S7](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf)). The results showed that both the morphology and structure were barely affected by Cu^2+^, indicating that the energy release from the O \u2192 W LMCT excited state transferred to Cu^2+^ rather than the ^5^D~0~ emitting state of Eu^3+^ is the possible reason for the fluorescence quenching.^[@ref58]^ Specifically, as Cu^2+^ diffused into the solution and came close to the Eu^3+^ ion surface, the energy transfer was more efficient for the smaller sized nanoparticles and externally distributed EuW~10~.^[@ref59]^\n\nConclusions {#sec3}\n===========\n\nIn our work, the design of notably fluorescence-enhanced solutions by a cationic component (TMAB) and an anionic cluster POMs (EuW~10~) by the ionic self-assembly strategy was presented. The fluorescent emission of the EuW~10~/TA system revealed a downward trend as the alkyl chain length of TA increased, which can be attributed to electrostatic interaction and steric hindrance effect. What's more, the EuW~10~/TMAB complex showed excellent responsiveness to pH and, most importantly, EuW~10~/TMAB composites can act as sensitive fluorescence sensors for Cu^2+^ in water, providing an environment friendly and real-time rapid detection method. Our results further proved that the fluorescent functional materials have unlimited potential in optoelectronic applications and sensing.\n\nExperimental Section {#sec4}\n====================\n\nMaterials {#sec4.1}\n---------\n\nEuW~10~ was synthesized according to the description of Sugeta and Yamase.^[@ref34]^ Tetramethylammonium bromide (TMAB), tetraethylammonium bromide (TEAB), tetrabutylammonium bromide (TBAB), tetrahexylammonium bromide (THAB), and tetramethylphosphonium bromide (TMPB) were all obtained from Macklin Biochemical Co. Ltd (Shanghai). The structures of EuW~10~, TMAB, TEAB, TBAB, and THAB are shown in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}a--e. NaBr and nitrates were acquired from Sinopharm Chemical Reagent Co. and used directly. Ethylenediaminetetraacetic acid (EDTA) was obtained from Tianjin Baishi Chemical Industry Co. Ltd. In all experiments, the water was obtained from a UPH-IV ultrapure water purifier (China) and the resistivity was 18.25 M\u03a9 cm.\n\nInstruments {#sec4.2}\n-----------\n\nTransmission electron microscopy (TEM) was carried out on a JEM-1011 (JEOL) instrument. Field-emission scanning electron microscopy was carried out on Hitachi SU8010 at 5.0 kV. Confocal laser scanning microscopy (CLSM) used a Panasonic Super Dynamic II WV-CP460 instrument, and the sample was excited at 488 nm. Atomic force microscopy (AFM) results were acquired from a Dimension Icon (American) with a ScanAsyst. Fourier transform infrared (FT-IR) spectra were observed from an \u03b1-T spectrometer of Germany Bruker Optics. The values of \u03b6-potential were acquired from Malvern Zetasizer Nano ZS ZEN3600. UV--vis spectra were collected on a UV--vis spectrophotometer (Hitachi, U-4100). A Lumina Fluorescence Spectrometer (Thermo Fisher) of the model Thermo Scientific Lumina was applied to observe the fluorescence spectra. The fluorescence lifetimes were determined using a spectrofluorometer (FLSP920, Edinburgh Instruments Ltd) with a time-correlated single-photon-containing method.\n\nSample Preparation of the EuW~10~/TMAB Composite {#sec4.3}\n------------------------------------------------\n\nIn this experiment, 3 mM of the 0.2 mL EuW~10~ aqueous solution was added to 0.8 mL TMAB aqueous solution under stirring. The solution was incubated for 1 day in a thermostat at 20.0 \u00b1 0.1 \u00b0C, after which the sample was used for other characterizations.\n\nDetection of Cu^2+^ {#sec4.4}\n-------------------\n\nMetal salts (30 \u03bcL; *c*~metal~ = 2 mM) were put into 2 mL of the hybrid nanostructure system (*c*~EuW~10~~/*c*~TMAB~ = 0.6/30 mM) for the detection of metal ions. Cu^2+^ salts (30 \u03bcL; *c*~metal~ = 2 mM) were put into 2 mL of the hybrid nanostructure system (*c*~EuW~10~~/*c*~TMAB~ = 0.6/30 mM) including other kinds of metal ions (*V* = 30 \u03bcL, *c*~metal~ = 2 mM) to investigate the selectivity of nanostructures for Cu^2+^.\n\nThe Supporting Information is available free of charge on the [ACS Publications website](http://pubs.acs.org) at DOI: [10.1021/acsomega.8b01636](http://pubs.acs.org/doi/abs/10.1021/acsomega.8b01636).Lifetimes and ratios, UV--vis spectra, electrical conductivity of the EuW~10~/TMAB, electrical conductivity and \u03b6-potential of the EuW~10~/TA, fluorescence spectra of 0.6 mM EuW~10~ with different ratios of NaBr, TEM images of 0.6 mM EuW~10~/30 mM TMAB for different pH, the linear relationship of 1/(*I*~0~ -- *I*) vs 1/\\[Cu^2+^\\] as well as *I*~0~/*I* vs \\[Cu^2+^\\] for 0.6 mM EuW~10~/30 mM TMAB, TEM images and FT-IR spectra of 0.6 mM EuW~10~/30 mM TMAB before and after adding 30 \u03bcM Cu^2+^ ([PDF](http://pubs.acs.org/doi/suppl/10.1021/acsomega.8b01636/suppl_file/ao8b01636_si_001.pdf))\n\nSupplementary Material\n======================\n\n###### \n\nao8b01636_si_001.pdf\n\nThe authors declare no competing financial interest.\n\nWe gratefully acknowledge the financial support from the National Natural Science Foundation of China (21201110, 21571115), Young Scholars Program of Shandong University (2016WLJH20) and the Major Research of Science and Technology, China (Grant No. 2016ZX05025-003-002).\n"}
+{"text": "Full data cannot be made available because of ethical and legal restrictions safeguarding patients\\' privacy, as per Italian laws and regulations. Excerpts of data relevant to the study can be requested to the Committee for Research Use of Human Samples, Department of Medical Science, University of Turin, Italy. (email: <spoke.dsm@unito.it>).\n\nIntroduction {#sec005}\n============\n\nClassification of primary CNS solitary fibrous tumour/hemangiopericytoma (SFT/HPC) has been recently redefined by the latest World Health Organization (WHO) Classification of Tumours of the Central Nervous System (CNS)\\[[@pone.0203570.ref001]\\] by merging into a single category two previously separated diagnostic entities (solitary fibrous tumour and hemangiopericytoma). This change mirrors what was already accepted for SFT/HPC of other sites (e.g. soft tissues and pleura) and is supported by a shared molecular hallmark: the chromosomal inversion at the 12q13 locus, fusing the NGFI-A-binding protein 2 (*NAB2*) and the signal transducer and activator of transcription 6 (STAT6) genes *\\[[@pone.0203570.ref002], [@pone.0203570.ref003]\\]*. This alteration is indeed present in most SFT/HPC cases regardless of tumour site or grade. NAB2-STAT6 fusion protein promotes tumour growth by localizing to the nucleus and activating the *EGR* gene. Nuclear localization of the fusion protein can be detected by anti-STAT6 antibodies\\[[@pone.0203570.ref004]\\] and STAT6 immunohistochemistry (IHC) is currently a recommended diagnostic tool by the latest WHO classification, since no nuclear STAT6 staining is present in meningiomas or in other frequent meningeal tumours, thus helping in the differential diagnosis.\\[[@pone.0203570.ref001], [@pone.0203570.ref004]\\]\n\nCNS SFT/HPC are rare (less than 1% of all primary CNS tumours), usually dural-based intracranial tumours, thought to develop from CD34-positive fibroblasts present in the dura mater or in the intraparenchymal perivascular connective tissue\\[[@pone.0203570.ref005], [@pone.0203570.ref006]\\] and mostly occur in adults, with a slightly higher incidence in males.\\[[@pone.0203570.ref007]\\] Exact epidemiological data are difficult to ascertain, due to both their rarity and to the changes of classification criteria over time.\n\nPrognosis is often good, especially in cases with the solitary fibrous tumour morphology and when gross total resection is achieved; conversely, the hemangiopericytoma morphology is associated with a higher rate of local recurrences (\\>75% in patients with long term follow up) and distant metastases occurrence.\\[[@pone.0203570.ref008]--[@pone.0203570.ref012]\\] Based on these findings, the current WHO grading criteria provide that SFT/HPC with solitary fibrous tumour phenotype and mitotic count \\<5 mitoses/10 high-power fields (HPF) are classified as grade I, whereas tumours with hemangiopericytoma morphology are graded as grade II or III based on the mitotic count with a cut-off value of 5 mitoses/10 HPF.\\[[@pone.0203570.ref001]\\] Radiotherapy may be of benefit to patients with incomplete resection or at higher risk of recurrence.\\[[@pone.0203570.ref013], [@pone.0203570.ref014]\\]\n\nOverall, compared to extra-CNS SFT/HPC, prognostic markers are less defined and available data are limited by the rarity of this tumour; thus, new and reliable prognostic factors are warranted to help personalize patients' therapeutic management. Therefore, we wanted to retrospectively assess novel possible prognostic factors in a series of SFT/HPC diagnosed at our Institution, evaluating both histopathological and molecular features. In particular, we evaluated the prognostic relevance of two IHC markers: CD34 and B-cell lymphoma 2 (Bcl-2), which, although non-specific, can be useful diagnostic tools and were especially important in the pre-STAT6 era.\\[[@pone.0203570.ref015]\\] CD34 loss has been reported as a characteristic finding in dedifferentiated extra-CNS SFT.\\[[@pone.0203570.ref016]\\] Secondly, considered the prognostic relevance of mitotic count in this tumour entity, we wanted to evaluate if a count based on phosphorylated histone H3 (PHH3) IHC could improve its grading and prognostic capability compared to H&E-based mitotic count. PHH3 antibody specifically recognizes the phosphorylated (at serines 10 and 28) histone H3, which is present in cells at the end of G2 or in the M phase of cell cycle and it has already been validated as a useful IHC marker to improve mitotic count sensitivity in different types of neoplasms, including breast cancer and meningiomas.\\[[@pone.0203570.ref017], [@pone.0203570.ref018]\\] In terms of molecular profile, telomerase maintenance has been recently described as a hallmark of most human cancers, needed by neoplastic cells to escape replicative senescence. *TERT* gene codes for the catalytic subunit of the telomerase complex and recurrent point mutations of its promoter have been shown to create a novel binding site for E-twenty-six (ETS) family transcription factors, leading to increased telomerase activity in affected cells. *TERT* promoter mutations are common to different malignancies, with variable prognostic significance.\\[[@pone.0203570.ref019]--[@pone.0203570.ref022]\\] Since this molecular alteration seems to be associated with adverse outcome in extra-CNS SFT,\\[[@pone.0203570.ref023]\\] we wanted to explore its possible prognostic meaning in our series.\n\nMaterials and methods {#sec006}\n=====================\n\nCase series and tissue samples {#sec007}\n------------------------------\n\nFifteen consecutive cases of primary CNS solitary fibrous tumours or hemangiopericytomas diagnosed at the Pathology Unit of City of Health and Science University Hospital of Turin between 2006 and 2016 were retrospectively collected. All patients underwent surgical resection at the Neurosurgery Unit of the same Institution. Follow up data were retrieved from patients' charts. The study was conducted in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and within guidelines and regulations by the Research Ethics Committee of the University of Turin. Written informed consent was not obtained because patients were no longer undergoing regular follow up visits directly at our institution or were deceased; considered the retrospective nature of the present study and that it had no impact at all on patients' care, verbal informed consent was obtained from participants or their relatives and recorded in a specific log. This study, including the verbal consent procedure, was approved by the Research Ethics Committee of the University of Turin.\n\nMorphology and immunohistochemistry {#sec008}\n-----------------------------------\n\nAll diagnoses were confirmed by a senior neuropathologist (PC) and reassessed according to the 2016 WHO classification. Thus, tumours were graded based on their phenotype (solitary fibrous tumour versus hemangiopericytoma) and mitotic count on H&E slides. Mitotic count was repeated after PHH3 IHC (polyclonal, Ventana Medical Systems Inc., Tucson, AZ, US). IHC for STAT6 (polyclonal, Spring Bioscience Corporation, Pleasanton, CA, US), Ki67 (clone 30--9, Ventana Medical Systems Inc., Tucson, AZ, US), CD34 (clone QBEnd/10, Ventana Medical Systems Inc., Tucson, AZ, US) and Bcl-2 (clone SP66, Ventana Medical Systems Inc., Tucson, AZ, US) were also performed. Immunohistochemistry for Ki67, CD34 and Bcl-2 was performed on all tissue blocks of each case; Ki67 labelling index was evaluated by counting at least 1000 tumour cells while CD34 and Bcl-2 were assessed by visual estimation. Nuclear staining for STAT6 was considered consistent with the SFT/HPC diagnosis. BenchMark ULTRA platform (Ventana Medical Systems Inc., Tucson, AZ, USA) was used for all IHC.\n\n*TERT* promoter sequencing {#sec009}\n--------------------------\n\nDNA extraction from formalin-fixed and paraffin-embedded (FPPE) tumour samples was performed as previously described\\[[@pone.0203570.ref024]\\] and concentrations/purity were measured by a Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).\n\nMutational status of the telomerase reverse transcriptase (TERT) promoter region from position -27 to -286 from ATG start site, including the polymorphic site represented by rs2853669, were determined by PCR and Sanger sequencing using the following primer pair: promoter forward 5\u2032-CAGCGCTGCCTGAAACTC-3\u2032 and reverse 5\u2032-GTCCTGCCCCTTCACCTT-3\u2032, as described by Horn et al.\\[[@pone.0203570.ref025]\\] PCR products were purified and used as template for the sequencing reactions, which were performed with BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). After purification, the sequences were analyzed by Sanger direct sequencing using the ABIPRISM 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA).\n\nStatistical analyses {#sec010}\n--------------------\n\nStatistical analyses were performed using Stata/MP 15.0 Statistical Software (StataCorp, College Station, TX, USA). Categorical variables were compared using Pearson's \u03c7^2^ test. Kruskall-Wallis has been used to compare not normally distributed variables. The variables tested were: age at diagnosis, gender, tumour phenotype, grading based on H&E and on PHH3 mitotic count (\\<5/10 HPF *versus* \u22655/10 HPF), tumour size based on MRI report before surgery (\\<3 cm *versus* \u22653), Ki67 labelling index (\u22645% *versus* \\>5%), CD34 expression (\\<80% versus \u226580%), Bcl-2 expression (\\<30% versus \u226530%), TERT promoter status and tumour recurrence. Differences were considered significant when p \\< 0.05 for reported two-sided p-values. Survival curves between different groups were plotted using the Kaplan-Meier method and the statistical comparisons were performed with Log-rank test.\n\nResults {#sec011}\n=======\n\nClinical characteristics {#sec012}\n------------------------\n\nFifteen SFT/HPC cases were retrospectively collected: 9/15 (60%) were female, median age at diagnosis was 60 (range: 10--67). Median follow up time was 2.4 years (6 months-7.4 years). Gender, age and tumour site were not significantly correlated with tumour morphology/phenotype ([Table 1](#pone.0203570.t001){ref-type=\"table\"}). Tumour site was as follows: cranial base 8/15 (53%), cerebral hemispheres 4/15 (27%) and posterior fossa 3/15 (20%). Gross total resection was not achieved in 7/15 (47%) and in 4 of these cases (57%), tumour was located at cranial base. Five patients (33%) developed local disease recurrence, but none showed distant metastases; all patients with disease recurrence had an incomplete surgical resection. Ten patients out of 15% (66%) were alive at end of follow up, whereas 5 patients died (3/5 because of SFT/HPC progression, 2/5 due to other causes). Gross total resection was significantly associated with a more favourable disease-free interval (DFI) (log-rank test p = 0.0185) ([Fig 1A](#pone.0203570.g001){ref-type=\"fig\"}), while no correlation was found for age, gender or tumour site.\n\n![Kaplan-Meier analysis of disease free interval according to different clinico-pathological variables.\\\nGross total resection (log-rank test p = 0.0185) (A), WHO grade (log-rank test p = 0.038) (B), tumour phenotype (solitary fibrous tumour *versus* hemangiopericytoma) (log-rank test p = 0.062) (C), tumour grade according to PHH3-based mitotic count (log-rank test p = 0.094) (D), CD34 expression (log-rank test p = 0.038) (E) and Bcl-2 expression (log-rank test p = 0.010) (F).](pone.0203570.g001){#pone.0203570.g001}\n\n10.1371/journal.pone.0203570.t001\n\n###### Clinico-pathological characteristics according to tumour phenotype.\n\n![](pone.0203570.t001){#pone.0203570.t001g}\n\n  --------------------------------------------------------------------------------------------------------\n                                                                        Tumour phenotype   p       \n  ------------------------------------------ -------------------------- ------------------ --- --- -------\n  **Gender**                                 **F**                      9                  4   5   0.667\n\n  **M**                                      6                          2                  4       \n\n  **Age**                                    **\u226460 years**              8                  3   5   0.833\n\n  **\\>60 years**                             7                          3                  4       \n\n  **Tumour site**                            **Cerebral hemispheres**   4                  1   3   0.526\n\n  **Cranial base**                           8                          3                  5       \n\n  **Posterior fossa**                        3                          2                  1       \n\n  **Tumour size**                            **\\<3 cm**                 9                  4   5   0.667\n\n  **\u22653 cm**                                  6                          2                  4       \n\n  **WHO grade (H&E-based mitotic count)**    **I**                      6                  6   0   0.001\n\n  **II**                                     4                          0                  4       \n\n  **III**                                    5                          0                  5       \n\n  **WHO grade (PHH3-based mitotic count)**   **I**                      6                  6   0   0.001\n\n  **II**                                     3                          0                  3       \n\n  **III**                                    6                          0                  6       \n\n  **H&E-based mitotic count**                **\\<5**                    10                 6   4   0.044\n\n  **\u22655**                                     5                          0                  5       \n\n  **PHH3-based mitotic count**               **\\<5**\\                   9                  6   3   0.028\n                                             **\u22655**                                                \n\n  **\u22655**                                     6                          0                  6       \n\n  **KI67 labelling index**                   **\u22645%**                    8                  6   2   0.003\n\n  **\\>5%**                                   7                          0                  7       \n\n  **CD34%**                                  **\\<80%**                  5                  0   5   0.025\n\n  **\u226580%**                                   10                         6                  4       \n\n  **Bcl2%**                                  **\u226430%**                   5                  1   4   0.264\n\n  **\\>30%**                                  10                         5                  5       \n\n  ***TERT* promoter mutations**              **Wildtype**               10                 5   5   1.000\n\n  **Mutated**                                2                          1                  1       \n\n  ***TERT* SNP rs2853669**                   **No**                     6                  2   4   0.248\n\n  **Yes**                                    6                          4                  2       \n\n  **Tumour recurrence**\\                     **No**                     10                 5   5   0.580\n\n  **Yes**                                    5                          1                  4       \n  --------------------------------------------------------------------------------------------------------\n\nHistological features and grading {#sec013}\n---------------------------------\n\nSix cases showed a solitary fibrous tumour phenotype (6/15, 40%) ([Fig 2A](#pone.0203570.g002){ref-type=\"fig\"}, while 9/15 (60%) had a hemangiopericytoma morphology ([Fig 2B](#pone.0203570.g002){ref-type=\"fig\"}). Mean number of mitoses/10 HPF on H&E was 4.8 \\[standard deviation (SD): 5.51\\]. According to WHO criteria, 6 cases were grade I (40%), 4/15 (27%) grade II and the remaining 5 cases (33%) were grade III tumours. No necrosis was identified in any case. H&E-based mitotic count \u22655/10 HPF was significantly correlated with the hemangiopericytoma phenotype (p = 0.044). WHO grade resulted significantly associated with DFI (log-rank test p = 0.038) ([Fig 1B](#pone.0203570.g001){ref-type=\"fig\"}), while tumour phenotype did not (log-rank test p = 0.062) ([Fig 1C](#pone.0203570.g001){ref-type=\"fig\"}).\n\n![Example images showing the different morphological and immunohistochemical features of the case series.\\\nImages of the solitary fibrous tumour (A) and the hemangiopericytoma (B) phenotypes. Mitotic figures highlighted by PHH3 immunohistochemistry (C). STAT6 nuclear staining (D), a characteristic finding of SFT/HPC. Diffuse (E) and focal (F) CD34 positivity.](pone.0203570.g002){#pone.0203570.g002}\n\nPHH3-based grading {#sec014}\n==================\n\nAlthough the mean number of mitoses/10 HPF observed following PHH3 IHC staining was higher than in H&E (8.4 versus 4.8, SD: 10.69) ([Fig 2C](#pone.0203570.g002){ref-type=\"fig\"}), in one case only the PHH3-based mitotic count would have changed the tumour grade (from grade II to III) according to the WHO criteria. As observed for the H&E-based count, PHH3-based mitotic count \u22655/10 HPF significantly correlated with the hemangiopericytoma phenotype (p = 0.028). WHO grade using PHH3-based mitotic count was not significantly associated with DFI (log-rank test p = 0.094) ([Fig 1D](#pone.0203570.g001){ref-type=\"fig\"}).\n\nImmunohistochemistry {#sec015}\n====================\n\nNuclear staining for STAT6 was present in 14/15 (93.3%) ([Fig 2D](#pone.0203570.g002){ref-type=\"fig\"}). The only STAT6-negative case displayed the characteristic features of the hemangiopericytoma phenotype and was partially positive for CD34 (40%). Overall, mean Ki67 labelling index was 9% (SD: 7.79%): 3.5% in grade I, 7% in grade II and 16.8% in grade III tumours. Regarding CD34 expression ([Fig 2E and 2F](#pone.0203570.g002){ref-type=\"fig\"}), overall mean number of positive cells was 64% (SD: 35.6%): mean rates of 86%, 90% and 19% positive cells were observed in grade I, II and III tumours, respectively (p\\<0.001). In 3/5 (60%) recurrent cases CD34 expression was \\<10%. Bcl-2 overall mean expression was 63% (SD: 38.6%), 74%, 76% and 38% in grade I, II and III tumours, respectively (p = 0.2262). Ki67 labelling index (p = 0.003) and CD34 expression (p = 0.025) were significantly correlated with the hemangiopericytoma phenotype ([Table 1](#pone.0203570.t001){ref-type=\"table\"}). CD34 expression \u226480% was associated with poorer DFI (log-rank test p = 0.038) ([Fig 1E](#pone.0203570.g001){ref-type=\"fig\"}) as well as Bcl-2 expression \u226430% (log-rank test p = 0.010) ([Fig 1F](#pone.0203570.g001){ref-type=\"fig\"}).\n\n*TERT* promoter sequencing {#sec016}\n==========================\n\n*TERT* promoter was successfully analysed in twelve cases out of 15 (80%) (one case was excluded because of lacking adequate material for analysis and in the other two cases sequencing failed). *TERT* promoter resulted wildtype in 10/12 (83%) cases, while 2/12 (17%) were mutated: the recurrent 1,295,228 C\\>T (C228T) mutation and a 1,295,254 C\\>T (C254T) mutation positioned respectively at 124 and 150 base pairs upstream of the ATG translational start site of TERT were identified; these mutations were found in a grade I and a grade II SFT/HPC, respectively. The first patient had no disease recurrence (follow up: 73 months), while the second patient had no recurrence too, but died of unrelated causes 5 months after diagnosis. Six out of 12 (50%) cases harboured the rs2853669 single nucleotide polymorphism (SNP). Neither of these findings was associated with tumour phenotype (p = 1.000 and p = 0.248) or DFI (log-rank test p = 0.387 and p = 0.739).\n\nDiscussion {#sec017}\n==========\n\nThe identification of prognostic markers is a crucial issue for appropriate patients' management, even more when it applies to rare entities or after changes in the classification criteria. Since SFT/HPC of the CNS well resume both these issues, in the present study we analysed and identified putative prognosticators which could be used in the daily practice. Prognostic factors are especially needed in this setting considered that some of the parameters used for risk-stratification of extra-CNS SFT/HPC, like tumour size, are not applicable in the CNS.\\[[@pone.0203570.ref026]\\]\n\nBeginning with clinical variables, GTR was significantly associated with a better DFI, a finding consistent with literature data.\\[[@pone.0203570.ref008], [@pone.0203570.ref012], [@pone.0203570.ref027]\\] As previous studies have found, surgical resection alone is sufficient to cure patients in many cases. Although tumour site was not related to DFI in our series, 57% of cases in which GTR was not achieved were located at cranial base, a site which could possibly hamper a complete resection as reported for other CNS tumours (i.e. meningiomas).\\[[@pone.0203570.ref028]\\] Among clinical variables neither gender nor age were related to patients' outcome.\n\nFocusing on histopathological features, the two different tumour phenotypes (SFT and HPC) were not specifically associated with clinical variables (gender, age and tumour site): as a matter of fact, a distinction based on their clinico-radiological characteristics was not deemed possible even when they were considered different entities.\\[[@pone.0203570.ref029]\\] The HPC phenotype was associated with a higher mitotic count, but tumour phenotype alone did not reach statistical significance in terms of DFI; on the contrary, a higher WHO grade was associated with a worse outcome. This finding supports the present WHO grading criteria which take into account both tumour phenotype and mitotic count.\n\nIn our series, PHH3-based mitotic count was higher than the H&E-based (as expected), but it would have changed the WHO grade (from grade II to grade III) in one case, only; moreover, grading based on this tool was not associated with a different DFI, therefore suggesting its lack of utility in this specific subset of tumours. These results need to be validated in a larger cohort, but for the time being the use of PHH3-based mitotic count in daily practice is not supported.\n\nAmong other IHC markers, STAT6 nuclear staining was detected in all cases except for one (14/15, 93.3%), a finding consistent with literature data.\\[[@pone.0203570.ref015]\\] Rare STAT6-negative SFT/HPC are possible and in these cases diagnosis should be based on a careful revision of tumour morphology and using other markers like CD34, Bcl-2 and CD99. In terms of prognostic meaning, we found that a lower CD34-expressions was associated with a significantly poorer outcome in our series. Previous research studies found an association between CD34 loss and higher-grade tumours\\[[@pone.0203570.ref030], [@pone.0203570.ref031]\\] and decreased CD34 expression is also linked to recurrent tumours, a finding confirmed by our data: we observed a very low (\\<10%) CD34-expression in most of our recurrent cases (3/5, 60%). Bcl-2 expression (when lower than 30%) was associated with prognosis as well. In our series, only one of the 5 recurrent cases showed discordant CD34/Bcl-2 expression with low CD34 and high Bcl-2 expressions. This finding suggests a relationship between these two markers and is worth being validated in a larger series.\n\n*TERT* promoter mutations have been identified in CNS SFT/HPC, but specific data in this setting are limited.\\[[@pone.0203570.ref032], [@pone.0203570.ref033]\\] In our cases, we observed a lower overall mutation rate compared to previously reported data (16.7% *versus* 25.6%), a difference even greater when considering the HPC phenotype alone (16.7% *versus* 50%).\\[[@pone.0203570.ref032]\\] We did not observe an association between *TERT* promoter mutational status and outcome, but this could probably be explained by the limited sample size. Moreover, one of the two identified mutations (C254T) was still unreported in this tumour entity. We also investigated a common *TERT* promoter SNP (rs2853669): its frequency in our series (50%) is in line with its prevalence in the general population and we did not observe an association with outcome.\n\nIn conclusion, WHO grade is the main prognostic tool for CNS SFT/HPC, but our data suggest that it could be integrated in the daily practice by other markers, like CD34 and Bcl-2. Also molecular traits, like *TERT* promoter mutations, could provide valuable information, but their significance in this setting needs further evaluation in larger series.\n\nWe would like to thank Francesca Veneziano and Chiara Musuraca for their excellent technical assistance.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "1. Introduction {#sec1-molecules-25-03005}\n===============\n\n*Calea pinnatifida* (R. Br) Less. (Asteraceae), popularly known as aruca, cip\u00f3 cruz or quebra-tudo, occurs in Brazil, mainly in the \"cerrado\" biome \\[[@B1-molecules-25-03005]\\]. This species is used in folk medicine as infusions to treat stomachaches, giardiasis, amoebiasis and gastric disorders in general \\[[@B2-molecules-25-03005]\\]. According to the literature, phytochemical studies on the plant have revealed sesquiterpene lactones, fatty esters, steroids and a polyacetylene as the main components \\[[@B2-molecules-25-03005]\\].\n\nSesquiterpene lactones (STLs) constitute a large group of secondary metabolites found in Asteraceae \\[[@B3-molecules-25-03005],[@B4-molecules-25-03005]\\]. The presence of \u03b1,\u03b2-unsaturated carbonyl systems in STLs has been described as an important structural feature responsible for bioactivity, since they are capable of acting as Michael acceptors when in contact with biological nucleophiles \\[[@B5-molecules-25-03005],[@B6-molecules-25-03005]\\]. Furthermore, Michael acceptor compounds are considered important alkylating agents, capable of supporting alkylation reactions through adequate nucleophiles \\[[@B7-molecules-25-03005]\\].\n\nAccording to the literature \\[[@B3-molecules-25-03005]\\], germacranolides appear as one of the most frequent types of STL found in *Calea*, with more than 40 isolated compounds from *C. pinnatifida*, *C. ternifolia*, *C. urticifolia* and *C. zacatechichi* \\[[@B8-molecules-25-03005],[@B9-molecules-25-03005],[@B10-molecules-25-03005],[@B11-molecules-25-03005],[@B12-molecules-25-03005],[@B13-molecules-25-03005]\\].\n\nDespite the fact that sesquiterpene lactones from *Calea* have been isolated since the 1970s, important structural information about many STLs, such as their absolute configuration, is still unclear. This is due especially to the large number of possible substituents present in STLs, as well as the lack of specific techniques available in the past for solution-state assignments. Many of the experiments carried out on these molecules were based on X-ray crystallography and degradation reactions, as in the case of neurolenin B, a sesquiterpene lactone that had its relative configuration assigned based on X-ray crystallographic data \\[[@B9-molecules-25-03005]\\]. Nowadays, bidimensional NMR experiments, combined with vibrational circular dichroism (VCD) measurements and calculations, represent powerful complementary tools to unambiguously define both the relative and absolute configurations of a given molecule directly in solution \\[[@B14-molecules-25-03005],[@B15-molecules-25-03005]\\].\n\nIn terms of biological potential, literature data suggest that STLs might be responsible for the cytotoxic potential of many species, such as *C. pinnatifida* \\[[@B16-molecules-25-03005],[@B17-molecules-25-03005],[@B18-molecules-25-03005]\\] and *C. urticifolia* \\[[@B19-molecules-25-03005],[@B20-molecules-25-03005]\\]. Regarding antitumor activity, our research group showed that calein C, an STL isolated from *C. pinnatifida*, inhibits mitotic progression and induces apoptosis in MCF-7 cell lines by the inhibition of cell cycle progression at the M-phase \\[[@B18-molecules-25-03005]\\]. Furthermore, analogues of calein C, such as arucanolide, displayed cytotoxicity against melanoma and HL60 cells \\[[@B17-molecules-25-03005],[@B21-molecules-25-03005]\\].\n\nOver the last years, the development of antineoplastic drugs based on natural products has been mainly due to combinatory chemistry, advanced spectroscopic and spectrometric instruments, and molecular networking \\[[@B22-molecules-25-03005],[@B23-molecules-25-03005]\\]. However, many cancer tumors are resistant to some antineoplastic drugs administrated, not only the ones obtained from natural sources, but also the most common drugs used, such as cisplatin \\[[@B24-molecules-25-03005]\\]. This scenario highlights the urgent need for prospecting new drugs.\n\nThyroid carcinoma is the fifth most common type of cancer among women worldwide \\[[@B25-molecules-25-03005]\\]. The most common forms of thyroid cancer originate from thyroid follicular cells and can be divided into three major pathological forms: differentiated thyroid cancer (papillary thyroid carcinoma---PTC and follicular thyroid carcinoma---FTC) and undifferentiated thyroid cancer (anaplastic thyroid carcinoma---ATC).\n\nAccording to literature data, natural products and plant extracts have demonstrated cytotoxic potential against thyroid carcinomas. For example, *Pulsatilla koreana*, a plant used in traditional Chinese and Korean medicines as an anti-inflammatory agent, showed cytotoxicity against ATC cell lines and caused a reduction in cell growth, inducing apoptosis \\[[@B26-molecules-25-03005],[@B27-molecules-25-03005],[@B28-molecules-25-03005]\\].\n\nBased on these issues, and also on the expressive bioactivity of related sesquiterpene lactones, this work aims to evaluate the cytotoxic potential of STLs isolated from *C. pinnatifida* (calein C, calealactone B and a derivative obtained from calein C) against thyroid tumor cell lines, exploring the differences observed between structures in terms of the presence of \u03b1,\u03b2-unsaturated carbonyl systems. Additionally the absolute configurations of natural products **1** and **2** were confirmed by VCD spectroscopy.\n\n2. Results {#sec2-molecules-25-03005}\n==========\n\n2.1. Structural Elucidation of Compounds **1**--**3** {#sec2dot1-molecules-25-03005}\n-----------------------------------------------------\n\nThe chromatographic procedures carried out allowed the isolation of three compounds: calein C (**1**), calealactone B (**2**) and a calein C derivative (**3**), a product of a Michael addition reaction with methanol ([Figure 1](#molecules-25-03005-f001){ref-type=\"fig\"}).\n\nCompound **1** was obtained as a colourless crystal, mp. 170 \u00b0C. ESI-HRMS (positive mode) data displayed a *pseudo*-molecular ion peak at *m*/*z* 429.1542 \\[M + Na\\]^+^ corresponding to the molecular formula C~21~H~26~O~8~Na (calculated *m/z*: 429.1553). All signals in the ^1^H- and ^13^C-NMR spectra were attributed by the analysis of the ^1^H--^1^H COSY, HMQC, HMBC and NOESY ([Figure 2](#molecules-25-03005-f002){ref-type=\"fig\"}) spectra recorded in benzene-*d*~6~.\n\nOnce the relative configuration of Compound **1** was unambiguously assigned by NOESY analysis, vibrational circular dichroism (VCD) experiments and calculations were used to determine its absolute stereochemistry for the first time in the literature. VCD associated with density functional theory (DFT) calculations has been demonstrated as a powerful probe for chirality in natural products \\[[@B14-molecules-25-03005],[@B15-molecules-25-03005]\\].\n\nThe very good agreement between the experimental VCD spectra obtained in CDCl~3~ and simulated data at the B3PW91/6-311G(d,p) level led to the assignment of the 4*R*,6*R*,7*S*,8*S*,9*R*,10*R* configuration to Compound **1** ([Figure 3](#molecules-25-03005-f003){ref-type=\"fig\"}). The most representative vibrational modes used in this assignment were those at (\u2212)-1025 cm^\u22121^ (C-O stretch of C-6, C-8, C-9 and C-10), (\u2212)-1125 and (\u2212)-1145 cm^\u22121^ (C-H bending of whole molecular framework), and (\u2212)-1345 and (\u2212)-1365 cm^\u22121^ (out-of-plane and in-plane C-H bending of C-6, C-7, C-8 and C-9).\n\nCompound **1**, identified as calein C, was previously isolated from *C. urticifolia* \\[[@B8-molecules-25-03005]\\] and *C. zacatechichi* \\[[@B9-molecules-25-03005]\\]. Our group has also studied this compound previously in order to determine the correct relative position of the acetate and methacrylate groups, by means of HSQC and HMBC NMR in benzene-*d*~6~ ([Table 1](#molecules-25-03005-t001){ref-type=\"table\"}) \\[[@B18-molecules-25-03005],[@B29-molecules-25-03005]\\], which were assigned to C-9 and C-8, respectively.\n\nNMR and MS data analyses of calealactone B (**2**) demonstrated that this STL is an analogue of calein C, containing an epoxide ring at C-2 and C-3. The ^1^H-NMR spectrum was similar to that of **1**, but it was possible to observe slight differences in H-2 (\u03b4~H~ 4.23 *d*) and H-3 (\u03b4~H~ 3.34 *dd*) patterns that correspond to oxirane hydrogens. The relative configuration of the stereogenic centers C-2, C-3, C-4, C-6, C-7, C-8, C-9 and C-10 in the structure of Compound **2** was established using the NOESY spectrum, and its absolute configuration was also unambiguously determined by means of VCD and DFT calculations. Comparisons of experimental and calculated spectra allowed the assignment of (\u2212)-**2** as 2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R* ([Figure 4](#molecules-25-03005-f004){ref-type=\"fig\"}).\n\nAs expected, the VCD spectra obtained for (\u2212)-**2** were very similar to those of (\u2212)-**1**, as they share most of the structural and stereochemical features. The negative bands at around 1025, 1090 and 1140 cm^\u22121^, however, contain significant contributions from the C-H bending (out-of-plane) of C-2 and C-3, comprising the oxirane ring. The positive features between 1230 and 1260 cm^\u22121^ arise mainly from oxirane vibrations such as C-O stretch coupled to C-H bendings. Finally, the region between 1350 and 1400 cm^\u22121^ is influenced, among other vibrational modes, by in-plane C-H bendings of C-2 and C-3 coupled to C-C stretches to neighbouring carbons. Calealactone B was previously isolated from the genus *Calea* \\[[@B11-molecules-25-03005]\\]. However, this compound has not been described in *C. pinnatifida*.\n\nCompound (**3**), a calein C derivative obtained as an artefact by reaction with MeOH in Michael addition, was also evaluated. The HR-ESI-MS spectrum revealed the \\[M + H\\]^+^ peak at *m*/*z* 439.1980, corresponding to the molecular formula C~22~H~30~O~9~ (calculated for C~22~H~31~O~9~: *m*/*z* 439.1968). According to ^1^H-NMR data ([Table 1](#molecules-25-03005-t001){ref-type=\"table\"}), it was possible to observe signals at \u03b4~H~ 3.36 (*s*, 3H) that suggested the presence of a methoxy group replacing the exocyclic double bond in the lactone ring. In the ^13^C-NMR, the presence of the signal \u03b4~C~ 66.4 attributed to C-13, which appears shifted upfield compared to that observed for calein C (\u03b4~C~ 126.3), indicated the absence of the exocyclic double bond at C-11/C-13 as present in the lactone ring of Compounds **1** and **2**. Additionally, the ^13^C data revealed the presence of an extra carbon \u03b4~C~ 59.2, attributed to the methoxyl carbon. According to HMBC spectrum, it was possible to observe a correlation of H-13 (\u03b4~H~ 3.70/3.35) with C-22 (\u03b4~C~ 59.2) and C-12 (\u03b4~C~ 174.1), as well as the correlations between H-7 (\u03b4~H~ 3.08) and C-13 (\u03b4~C~ 66.4), confirming the position of the methoxyl group at C-13 (see [supplementary data](#app1-molecules-25-03005){ref-type=\"app\"}).\n\nIn order to define the importance of the conjugated system in the lactone ring to the biological activity, Compound (**3**) was also investigated in terms of cytotoxicity against anaplastic and papillary thyroid cancer cell lines.\n\n2.2. Biological Assays {#sec2dot2-molecules-25-03005}\n----------------------\n\nCytotoxic assays revealed that KTC-2 and TPC-1 tend to be more sensitive to calein C (**1**) than to the other substances tested ([Table 2](#molecules-25-03005-t002){ref-type=\"table\"}). This pattern is also observed in nontumor cells treated with (**1**) (NIH-3T3). Calealactone B (**2**) resulted in similar cytotoxic values against the cell lines tested when compared to calein C (**1**), probably due to their analogous chemical structures. The calein C derivative (**3**) was shown to be the least cytotoxic of the STLs tested in all cell lines, including fibroblasts.\n\nIn general, it is possible to observe that all substances were cytotoxic against all of the cell lines. However, the IC~50~ data show that calein C has a slight tendency to be more cytotoxic to the cells, mainly for papillary cells TPC-1 (IC~50~ 1.49 \u03bcM). In addition to that, anaplastic cells KTC-2 were also affected by the treatment with calein C (IC~50~ 1.67 \u03bcM). The literature also reports the cytotoxicity of calein C against the breast cancer cell lines MCF-7, MDA-MB-231 and Hs578T. The data suggest that calein C is able to reduce cell viability as well as inhibit cell cycle progression, consequently resulting in cell death \\[[@B18-molecules-25-03005]\\].\n\nSimilarly to what is observed with calein C, calealactone B was also shown to be cytotoxic against both of the cell lines, whereas the IC~50~ values were slightly higher than those obtained for calein C. This difference might be due to the presence of an epoxide in positions C-2 and C-3 instead of the double bond present in calein C.\n\nOn the other hand, the cytotoxic potential of the calein C derivative (**3**) was the least effective against both cell lines tested, for anaplastic (IC~50~ 25.3 \u03bcM) and papillary (IC~50~ 23.6 \u03bcM). It was possible to see that the addition of the methoxy group at C-13, replacing the double bond, drastically changes the biological response.\n\nThe potential of a compound to be cytotoxic against anaplastic thyroid carcinoma is an interesting achievement considering that this type of carcinoma is highly aggressive and that no accurate treatment is currently available \\[[@B30-molecules-25-03005]\\]. According to literature data, the cytotoxic potentials of isolated natural products against ATC are not widely studied. However, *Pulsatilla koreana* extract*,* used in traditional Chinese and Korean medicines, was shown to suppress the growth of anaplastic carcinoma in a dose-dependent manner, possibly also acting as an apoptosis inductor \\[[@B26-molecules-25-03005]\\]. In addition, a flavonoid fraction obtained from *Citrus reticulata* juice also led to the reduction of ATC cell migration and proliferation in a time-dependent manner, acting in the G2/M phase of the cycle, and consequently promoting cell death \\[[@B31-molecules-25-03005]\\].\n\nOther phytochemicals, such as resveratrol, a phytoalexin present in several plant species; curcumin, a phenolic compound obtained from the rhizomes of turmeric (*Curcuma longa*); genistein, an isoflavone naturally found in numerous plants; and epigallocatechin gallate (EGCG), a polyphenolic compound found in green tea, induced a strong and significant reduction of the viability of ATC cell lines \\[[@B32-molecules-25-03005]\\]. These results reinforce the importance of testing natural products on thyroid cancer cells with the objective of prospecting new drug prototypes.\n\nAccording to the literature, the cytotoxicity of STLs containing an \u03b1,\u03b2 unsaturated carbonyl system was demonstrated to be important when tested against *Trypanosoma brucei* \\[[@B27-molecules-25-03005]\\]. Moreover, Padilla-Gonzalez and collaborators \\[[@B5-molecules-25-03005]\\] described that \u03b1,\u03b2 unsaturated carbonyl systems in STLs can participate in Michael addition with biological nucleophiles, showing high cytotoxic potential in biological media. Based on that, we believe that the presence of the \u03b1,\u03b2-unsaturated lactone found in Compounds **1** and **2**, might be important for the antitumor activity of STLs, since its loss in Compound **3** extinguished this activity.\n\n3. Materials and Methods {#sec3-molecules-25-03005}\n========================\n\n3.1. General Experimental Procedures {#sec3dot1-molecules-25-03005}\n------------------------------------\n\n^1^H- and ^13^C-NMR spectra of Compounds **1**--**3** were recorded, respectively, at 300 and 75 MHz in a Bruker Ultrashield 300 Advance III spectrometer. CDCl~3~ (Aldrich, St. Louis, MO, USA) and benzene-*d*~6~ (Aldrich) were used as solvents and as the internal standards. Silica gel *flash* (Merck, 230--400 mesh, New Jersey, NJ, USA) and Sephadex LH-20 (Amersham Biosciences, Little Chalfont, UK) were used for column chromatographic separations, while silica gel 60 PF~254~ (Merck) was used for analytical thin-layer chromatography (TLC). HPLC analysis was performed in a Dionex Ultimate 3000 chromatography system (Dionex, Sunnyvale, CA, USA), using a Luna Phenomenex RP-18 column (3 \u03bcm, 150 \u00d7 5 mm) and UV-DAD detector. IR and VCD spectra of Compounds **1** and **2** were recorded with a Single-PEM Chiral*IR*-2X FT-VCD spectrometer (BioTools, Inc., Jupiter, FL, USA) using a resolution of 4 cm^\u22121^ and a collection time of 6 h. The optimum retardation of the ZnSe photoelastic modulator (PEM) was set at 1400 cm^\u22121^. Minor instrumental baseline offsets were eliminated from the final VCD spectra of **1** and **2** by subtracting their VCD data from those obtained for the solvent under identical conditions. The IR and VCD spectra were recorded in a BaF~2~ cell with a 100 \u00b5m path length, using CDCl~3~ as the solvent. The samples were prepared as follows: 5 mg of **1** were added to 120 \u00b5L of CDCl~3~ and 3 mg of **2** were added to 110 \u00b5L of CDCl~3~.\n\n3.2. Plant Material {#sec3dot2-molecules-25-03005}\n-------------------\n\nLeaves of *Calea pinnatifida* (R. Br) Less. (Asteraceae) were collected from the Atlantic Forest area of S\u00e3o Paulo City, SP, Brazil (coordinates 23 53'08.86''S, 46 40'10.45''W), in October 2012. A voucher specimen (C.R. Figueiredo 25) has been deposited in the SPF Herbarium of Botany Department from the Biosciences Institute of University of S\u00e3o Paulo.\n\n3.3. Extraction and Isolation of Compounds {#sec3dot3-molecules-25-03005}\n------------------------------------------\n\nFresh leaves of *C. pinnatifida* (300 g) were dried, ground and then exhaustively extracted using MeOH at room temperature. After evaporation of the solvent under reduced pressure, the obtained crude extract (10 g) was resuspended in MeOH/H~2~O (2:1) and then partitioned with dichloromethane. The CH~2~Cl~2~ phase (650 mg) was subjected to column chromatography (CC) over Sephadex LH-20 using MeOH as a mobile phase to give seven fractions (A--G). Fraction B (420 mg) was subjected to CC over silica *flash* using increasing amounts of MeOH in CH~2~Cl~2~ as the solvent, to afford three fractions (B1--B3). Part of Fraction B1 (100 mg) was purified over semi-prep RP-18 HPLC, eluted with ACN/H~2~O (4:6) (flow rate: 3.6 mL min^\u22121^, \u03bb: 218 nm), to afford calein C (**1**, 40.0 mg) and calealactone B (**2**, 9.0 mg). In the course of extraction and fractionation, a calein C derivative (**3,** 2.8 mg; 7.0% of **1**) was produced and isolated.\n\n3.4. Cell Lines Culture {#sec3dot4-molecules-25-03005}\n-----------------------\n\nIn order to assess the cytotoxic effects of the STLs, two thyroid cancer cell lines---KTC-2 (anaplastic thyroid cell line) and TPC-1 (papillary thyroid cancer cell line), donated by Professor Edna Kimura---and the murine fibroblast nontumor cell line NHT-3T3 were selected. KTC-2 and TPC-1 were maintained in RPMI-1640 (ThermoFisher Scientific, Waltham, MA, USA) supplemented with 5% and 10% fetal bovine serum (ThermoFisher Scientific), respectively, and 0.01 \u03bcg mL^\u22121^ of penicillin-streptomycin (ThermoFisher Scientific). NIH-3T3 was maintained in DMEM (ThermoFisher Scientific) supplemented with 10% fetal bovine serum (ThermoFisher Scientific). All cell lines were cultured at 37 \u00b0C with 5% CO~2~.\n\n3.5. Cell Viability Assay and Determination of IC~50~ Value {#sec3dot5-molecules-25-03005}\n-----------------------------------------------------------\n\nThe cell viability after drug treatments was determined with PrestoBlue Cell Viability Reagent (ThermoFisher Scientific), following the manufacturer's instructions. Briefly, the thyroid cancer cells (KTC2 and TPC1) and the nontumor cell line (NIH-3T3) were seeded into 96-well plates at an initial density of 1000 cells per well and incubated at 37 \u00b0C for 48 h. Cells were treated with three concentrations of calein C (73.9, 36.9 and 18.5 \u03bcM), calein C derivative (68.3, 34.2 and 15.9 \u03bcM) and calealactone B (70.9, 35.4 and 17.7\u03bcM) for 72 h. A concentration of 10 \u03bcM of cisplatin was used as the positive control in the KTC2 and TPC1 cell lines for cytotoxicity. After 48 h of treatment, 10 \u03bcL of PrestoBlue reagent was added to each well and incubated for 1 h 30 min. Fluorescence (540 nm excitation/590 nm emission) was measured using a microplate reader, M3 (Molecular Devices, SoftMax Pro 7 Software). Vehicle-treated control cells (DMSO) were used to normalize the relative luminescence units from treated wells, and they were expressed as percentages of viable cells. The concentration to achieve 50% of cell death (IC~50~) was estimated from the non-linear regression analysis of the dose response curve using Prism 5 (GraphPad Prism 5, version 5.01). Cells were briefly treated with different concentrations of each compound for 72 h at concentrations ranging from 12.3 to 0.096 \u03bcM for calein C, 60.3 to 0.532 \u03bcM for calein C derivative and 17.7 to 0.137 \u03bcM for calealactone B. After 48 h treatment, 10 \u00b5L of PrestoBlue reagent was added to each well and incubated for 1.5 h. Fluorescence was also measured at M3. All experiments were performed in triplicate and data were expressed as the mean \u00b1 SEM.\n\n3.6. Calculations {#sec3dot6-molecules-25-03005}\n-----------------\n\nCalculations were performed for the arbitrarily chosen (4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**1** and (2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**2**. Conformational searches were carried out at the molecular mechanics level of theory with the Monte Carlo algorithm, employing the MM+ force field incorporated in the HyperChem 8.0.10 software package. Initially, 34 conformers of (4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**1** and 29 conformers of (2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**2** with a relative energy (rel E.) within 10 kcal mol^\u22121^ of the lowest-energy conformer were selected and further geometry-optimised at the B3LYP/6-31G(d) and B3PW91/6-311G(d,p) levels. The two conformers with rel E. \\< 4.0 kcal mol^\u22121^ for each compound, which corresponded to more than 99% of the total Boltzmann distributions, were selected for IR and VCD spectral calculations. DFT calculations were carried out at 298 K in the gas phase using Gaussian 09 software \\[[@B33-molecules-25-03005]\\]. The IR and VCD spectral simulations were created using dipole and rotational strengths from Gaussian, which were calculated at the same level used during the geometry optimisation steps, and converted into molar absorptivities (M^\u22121^ cm^\u22121^). Vibrational analysis at the B3LYP/6-31G(d) and B3PW91/6-311G(d,p) levels resulted in no imaginary frequencies, confirming the considered conformers as real minima. Each spectrum was plotted as a sum of Lorentzian bands with half-widths at half-maximum of 6 cm^\u22121^. The calculated wavenumbers were multiplied by a scaling factor of 0.975. The final spectra were generated according to Boltzmann weighting of the lowest-energy conformers identified for **1** and **2** and plotted using the Origin 8 software.\n\n4. Conclusions {#sec4-molecules-25-03005}\n==============\n\nFrom the MeOH extract of the leaves of *Calea pinnatifida*, two sesquiterpene lactones (STLs) were isolated: calein C (**1**) and calealactone B (**2**), while Compound **3** was obtained as a product of the Michael addition of MeOH to calein C. The absolute configurations of **1** and **2** were unambiguously confirmed as 4*R*,6*R*,7*S*,8*S*,9*R*,10*R* and 2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R*, respectively, by means of VCD and DFT calculations. According to the biological data obtained, it was possible to observe that calein C, calealactone B and the calein C derivative were effectively cytotoxic against thyroid cancer cells, both papillary and anaplastic. However, differences in conjugated double bonds among the structures of the STLs were shown to influence their cytotoxic potential. KTC-2 and TPC-1 were shown to be more sensitive to calein C than to the other tested substances. Thus, the obtained results demonstrate how the presence of an \u03b1,\u03b2-unsaturated lactone ring might be important for the antitumor activity. To the best of our knowledge, this is the first report of the calein C derivative (**3**) and the structure--activity relationship among sesquiterpene lactone analogues from *C. pinnatifida*, suggesting the importance of conjugated systems in their biological potential.\n\nThe authors thank the resources supplied by the Centre for Scientific Computing (NCC/GridUNESP) of S\u00e3o Paulo State University (UNESP), and the \"Central de Equipamentos e Servi\u00e7os Multiusu\u00e1rios (CESM-UNIFESP)\" for technical support.\n\n**Sample Availability:** Samples of the compounds **1** and **2** are available from the authors.\n\nThe following are available online, Figure S1: ^1^H-NMR spectrum of (**1**), Figure S2: ^13^C-NMR spectrum of (**1**), Figure S3: HMBC NMR spectrum of (**1**), Figure S4: NOESY NMR spectrum of (**1**), Figure S5: ^1^H-NMR spectrum of (**2**), Figure S6: ^13^C-NMR spectrum of (**2**), Figure S7: NOESY NMR spectrum of (**2**), Figure S8: ^1^H-NMR spectrum of (**3**), Figure S9: ^13^C-NMR spectrum of (**3**).\n\n###### \n\nClick here for additional data file.\n\nConceptualization, P.S.; methodology, L.A.C., M.T.R., A.N.L.B.; software, J.M.B.J.; Experimental work, data collection and evaluation, P.S., L.A.C., M.T.R., A.N.L.B.; spectral analysis, P.S., J.H.G.L., M.J.P.F.; Literature search and manuscript preparation, L.A.C., J.M.B.J., I.G.S.R., P.S.; writing---review and editing, J.M.B.J., I.G.S.R., P.S.; project administration, P.S., J.M.B.J.; funding acquisition, P.S., J.M.B.J. All authors have read and agreed to the published version of the manuscript.\n\nThis research was funded by the Funda\u00e7\u00e3o de Amparo \u00e0 Pesquisa do Estado de S\u00e3o Paulo (FAPESP), Grant\\# 2020/05741-2, 2016/24985-4, and 2014/25222-9. L.A.C, M.T.R and A.N.L.B. were funded by PhD and Post-Doc scholarships from Coordena\u00e7\u00e3o de Aperfei\u00e7oamento de Pessoal de N\u00edvel Superior (CAPES Cod. 001, \\#88887.313278/2019-00). P.S., J.H.G.L. and M.J.P.F. received a scientific research award from CNPq.\n\nWe wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.\n\n![Structures of calein C (**1**), calealactone B (**2**) and the derivative of calein C (**3**).](molecules-25-03005-g001){#molecules-25-03005-f001}\n\n![NMR NOESY correlations of **1**.](molecules-25-03005-g002){#molecules-25-03005-f002}\n\n![(Left) Comparison of experimental IR and vibrational circular dichroism (VCD) spectra of (\u2212)-**1** recorded in CDCl~3~ (black trace) with the calculated \\[B3PW91/6-311G(d,p)\\] IR and VCD spectra for the Boltzmann average of the two lowest-energy conformers identified for (4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**1** (red trace). (Right) Optimised structures, relative energies and Boltzmann populations of the lowest-energy conformers identified for (4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**1**.](molecules-25-03005-g003){#molecules-25-03005-f003}\n\n![(Left) Comparison of the experimental IR and VCD spectra of (\u2212)-**2** recorded in CDCl~3~ (black trace) with the calculated \\[B3PW91/6-311G(d,p)\\] IR and VCD spectra for the Boltzmann average of the two lowest-energy conformers identified for (2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**2** (red trace). (Right) Optimised structures, relative energies and Boltzmann populations of the lowest-energy conformers identified for (2*S*,3*S*,4*R*,6*R*,7*S*,8*S*,9*R*,10*R*)-**2**.](molecules-25-03005-g004){#molecules-25-03005-f004}\n\nmolecules-25-03005-t001_Table 1\n\n###### \n\n^1^H- (300 MHz) and ^13^C-NMR (75 MHz) spectral data for Compounds **1**--**3**.\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                          1 ^a^                                                                  2 ^b^                         3 ^b^                 \n  -------- ------------------------------ ------- -------------------------------- ----------------------------- ------- --------------------- --------------------- -------\n  **1**    \\-                             204.3   \\-                               \\-                            205.9   \\-                    \\-                    204.7\n\n  **2**    5.77 *d* (*J* 11.7 Hz)         124.9   H-15, H-7, H-3                   4.24 *d* (*J* 4.2 Hz)         62.9    H-15, H-3, H-9        6.55 *d*\\             125.6\n                                                                                                                                               (*J* 12.0 Hz)         \n\n  **3**    4.72 *t* (*J* 11.7 Hz)         147.5   H-14, H-15, H-4, H-6, H-2, H-5   3.33 *dd* (*J* 9.3; 4.2 Hz)   55.6    H-2, H-14             6.01 *t*\\             147.1\n                                                                                                                                               (*J* 12.0 Hz)         \n\n  **4**    2.57 *m*                       27.8    H-14, H-15\\                      1.60 *sl*                     26.0    OH, H-7               3.11 *m*              28.7\n                                                  H-7, H-3                                                                                                           \n\n  **5**    0.71 *m*/0.99 *m*              40.0    H-9, H-2, H-6                    1.46 *m*/1.90 *m*             38.8    H-14, H-21            1.61 *s*/1.85 *brs*   38.2\n\n  **6**    4.13 *dd* (*J* 11.7; 4.8 Hz)   77.2    H-7, H-9, H-5                    4.82 *dd*\\                    79.7    H-5, H-6, H-8         4.62 *dd*\\            77.2\n                                                                                   (*J* 11.9; 4.3 Hz)                                          (*J* 12.0; 4.1 Hz)    \n\n  **7**    2.24 *brs*                     41.3    H-6, H-8, H-9,\\                  2.36 *brs*                    40.9    H-5, H-6, H-8         3.08 *m*              40.2\n                                                  H-13                                                                                                               \n\n  **8**    5.70 *dd* (*J* 9.8; 2.0 Hz)    74.6    H-7, H-13, H-15                  5.68 *dd*\\                    71.5    H-5, H-21, H-7, H-6   5.64 *m*              69.3\n                                                                                   (*J* 9.9; 1.2 Hz)                                                                 \n\n  **9**    5.43 *d* (*J* 9.8 Hz)          73.9    H-6, H-7, H-15                   5.77 *d* (*J* 9.9 Hz)         73.9    H-2, H-6, H-15        5.60 *d*\\             74.7\n                                                                                                                                               (*J* 4.5 Hz)          \n\n  **10**   \\-                             79.1    \\-                               \\-                            74.6    \\-                    \\-                    79.2\n\n  **11**   \\-                             131.5   \\-                               \\-                            134.3   \\-                    \\-                    37.5\n\n  **12**   \\-                             168.0   \\-                               \\-                            168.3   \\-                    \\-                    174.1\n\n  **13**   5.02 *brs*/5.92 *brs*          126.3   H-7, H-18                        5.83 *s*/6.33 *s*             126.8   H-7                   3.35 *s*/3.70 *m*     66.4\n\n  **14**   0.33 *d* (*J* 6.3 Hz)          19.0    H-4, H-3                         1.22 *d* (*J* 6.1 Hz)         18.6    H-3                   1.16 *d*\\             19.8\n                                                                                                                                               (*J* 6.0 Hz)          \n\n  **15**   0.84 *s*                       23.2    H-2, H-9                         1.46 *s*                      24.5    H-2                   1.32 *s*              23.5\n\n  **16**   \\-                             165.2   \\-                               \\-                            165.3   \\-                    \\-                    165.4\n\n  **17**   \\-                             135.7   \\-                               \\-                            134.8   \\-                    \\-                    131.5\n\n  **18**   4.93 *s*/5.96 *s*              125.2   H-7, H-13, H-19                  5.57 *brs*/6.04 *brs*         127.3   H-19                  5.57 *s*/6.05 *s*     127.5\n\n  **19**   1.58 *s*                       17.8    H-18                             1.85 *s*                      18.1    H-18                  1.86 *s*              18.0\n\n  **20**   \\-                             170.0   \\-                               \\-                            170.4   \\-                    \\-                    170.1\n\n  **21**   1.38 *s*                       19.5    \\-                               2.05 *s*                      20.3    H-9, H-19             2.02 *s*              20.4\n\n  **22**   \\-                             \\-      \\-                               \\-                            \\-      \\-                    3.36 *s*              59.2\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n^a^ Benzene *d*~6~; ^b^ CDCl~3.~\n\nmolecules-25-03005-t002_Table 2\n\n###### \n\nIC~50~ (\u00b5M) determined for anaplastic cancer cells (KTC-2), papillary cancer cells (TPC-1) and murine fibroblasts (NIH-3T3).\n\n                                KTC-2   TPC-1   NIH-3T3\n  ----------------------------- ------- ------- ---------\n  Calein C (**1**)              1.67    1.49    3.06\n  Calealactone B (**2**)        4.69    3.54    4.40\n  Calein C derivative (**3**)   25.32   23.25   18.54\n  Cisplatin                     2.21    3.05    10.38\n\nIC~50~ values were calculated using non-linear regression in GraphPad Prism 5.0.\n"}
+{"text": "\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nObstructive sleep apnea (OSA) is a major type of sleep-disordered breathing prevalent in 2--7% of adults globally \\[[@pone.0177940.ref001]\\]. Co-morbid depression is common (21--41%) in OSA patients \\[[@pone.0177940.ref002]--[@pone.0177940.ref004]\\]. Recent studies showed that symptoms of depression were alleviated in OSA patients treated with continuous positive airway pressure \\[[@pone.0177940.ref005], [@pone.0177940.ref006]\\]. Besides, depressive-like behavior was observed in experimental animals given the treatment of chronic intermittent hypoxic (CIH) \\[[@pone.0177940.ref007], [@pone.0177940.ref008]\\]. These studies suggest causality between OSA and depression, but there is a paucity of mechanistic delineation of the pathophysiological link of the comorbidity.\n\nBrain monoamine oxidase A (MAO-A) plays an important role in maintaining the availability of monoamine neurotransmitters \\[[@pone.0177940.ref009]\\]. Dysregulated MAO-A activities significantly alter the homeostatic balance of monoamines that underpin pathogenesis of depression. In fact, overactivation of MAO-A has been reported in the brain of clinically depressed patients and in the postmortem brain \\[[@pone.0177940.ref010], [@pone.0177940.ref011]\\]. Also, neurodegeneration induced by elevated MAO-A activities was associated with depressive behavior in rodents with chronic stress \\[[@pone.0177940.ref012]\\].\n\nAlthough the role of inflammation in depression is highly contested, inflammation was reportedly observed in the brain of clinically depressed patients \\[[@pone.0177940.ref013]\\]. Inflammatory cytokine-responsive indoleamine-2,3-dioxygenase-1 (IDO-1) activation plays an important pathogenic role in the development of depressive-like behavior in experimental animals \\[[@pone.0177940.ref014], [@pone.0177940.ref015]\\]. IDO-1 catalyzes the first, rate-limiting step, in the tryptophan catabolism pathway, generating kynurenine and resulting in reduced levels of serotonin. Additionally, it has been demonstrated that a metabolite of the kynurenine pathway, quinolinic acid, can be neurotoxic. In fact, neurotoxic metabolites upon IDO-1 activation were reportedly to induce neurodegeneration \\[[@pone.0177940.ref016], [@pone.0177940.ref017]\\]. Here we examined the hypothesis that MAO-A upregulation induced by chronic intermittent hypoxia causes inflammation and IDO-1 activation, which significantly contribute to the serotonin deficiency and neurodegeneration.\n\nBrain permeable M30, 5\\[-N-Methyl-N-propargylaminomethyl\\]-8- hydroxyquinoline), is a synthetic compound composed of propargyl moiety and prototype of iron-chelator VK28 \\[[@pone.0177940.ref018]\\]. Thus, M30 possesses chemical properties of brain-selective MAO inhibitors and iron-chelating free radical scavengers \\[[@pone.0177940.ref019]\\]. These properties have been shown to be central to the protective effect of M30 against the pathogenic processes of neurodegenerative disease in animal models of Alzheimer's or Parkinson disease \\[[@pone.0177940.ref020], [@pone.0177940.ref021]\\]. A recent study has also reported an anti-inflammatory property of M30 via a down-regulation of the expression of inflammatory cytokines in a genetic model of Alzheimer's disease \\[[@pone.0177940.ref022]\\]. Yet, there is a lack of evidence on the mechanistic effect of M30 against the oxidative stress, inflammation and neurodegeneration induced by chronic intermittent hypoxia. In this study, we hypothesized that M30 could prevent depressive behavior induced by chronic intermittent hypoxia via its antagonistic effects on the MAO-A activity and oxidative stress, resulting in inflammation, IDO-1 activation, serotonin deficiency and neurodegeneration in the rat hippocampus.\n\nMaterials and methods {#sec002}\n=====================\n\nAnimal grouping and cell culture {#sec003}\n--------------------------------\n\nAnimal care and experimental protocol were approved and conducted according to the Committee on the Use of Live Animals in Teaching and Research (CULATR \\#2522--11, 3545--15), The University of Hong Kong. The Laboratory Animal Unit of the University of Hong Kong is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC international). Adult male Sprague-Dawley rats (220-250g) were put under pathogen-free condition in an air-conditioned room at constant temperature (23\u00b11\u00b0C) provided with water and standard diet (LabDiet, 5053 (LabDiet; St. Louis, MO, USA)) ad libitum. All animals were monitored on a daily basis for body health throughout the study. The animals were divided into four experimental groups (n = 12 each), namely normoxic control (Nx), M30-treated normoxic group (Nx+M30), hypoxia-treated group (IH), M30-treated hypoxic group (IH+M30).\n\nThe SH-SY5Y cells were obtained from ATCC (Manassa, VA, USA). The cells were cultured in DMEM/F-12, supplemented with 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 \u03bcg/mL), which were kept in incubators with 95% air and 5% CO~2~ at 37\u00b0C.\n\nIntermittent hypoxic protocol and drug preparation {#sec004}\n--------------------------------------------------\n\nThe normoxic control rats were kept in room air while hypoxic rats were maintained in an acrylic chamber for normobaric hypoxia in the same room. Levels of oxygen in the chamber were cycling between 21 to 5 \u00b1 0.5% per minute (i.e. 60 hypoxic episodes per hour) for 8 hours per day diurnally for 7 consecutive days. The desired oxygen levels were established by a mixture of room air and nitrogen and monitored by an oxygen analyzer (Vacumetrics Inc., St. Ventura, CA, USA).\n\nM30, 5\\[-N-Methyl-N-propargylaminomethyl\\]-8-hydroxyquinoline) was chemically synthesized and kindly provided by Dr. Moussa Youdim and Dr. Lin Bin. Rats were administered with a daily intraperitoneal injection of M30 (5mg/kg) 2 hours prior to the hypoxic treatment. The animals were anesthetized with halothane and then decapitated to harvest the hippocampus for experiments.\n\nCells were placed in an acrylic chamber in the incubator for intermittent hypoxia (repeated episodes of hypoxia at 1.5% oxygen for 4 hours followed by 21% oxygen for 4 hours) for 24 to 48 hours. MAO-A inhibitor clorgyline (10\u03bcM) or M30 (1\u03bcM) was added to the culture medium 1 hour prior to the hypoxic treatment.\n\nDetermination of MAO-A and MAO-B activities {#sec005}\n-------------------------------------------\n\nAccording to manufacturers' instruction, hippocampi were homogenized in 50mM potassium phosphate buffer (pH = 7.4) followed by optimal dilution with the use of reaction buffer provided by Amplex Red Monoamine Oxidase Assay Kit (Invitrogen, CA, USA). Enzyme activities of MAO-A and MAO-B were determined and normalized to total protein content in each sample. The results were expressed in the percentage of control.\n\nWestern blot {#sec006}\n------------\n\nLevels of protein expression of hippocampal tissue (including whole tissue cell lysate, cytosolic or nuclear fractions) were carried out as previously described \\[[@pone.0177940.ref023]\\]. The optical density of the bands was measured and quantified by Image J (National Institute of Health, MD, USA). Primary antibodies of SOD-2, GPx-1, NF\u03baB p65 and p50, I\u03baB\u03b1, TNF\u03b1, IL-1\u03b2, IL-6 and COX-2 were purchased from Santa Cruz Biotechnology, CA, USA; Synapsin-1 and Synaptophysin were purchased from Novus Biologicals, USA; PSD95 and Cleaved Caspase 3 was purchased from Cell Signaling Technology; Cleaved PARP1 was purchased from Bioworld Technology; IDO-1 was purchased from antibodies-online (ABIN1714836). The data were expressed as percentage of the control.\n\nEnzyme-linked immunosorbent assay (ELISA) {#sec007}\n-----------------------------------------\n\nHippocampal serotonin (5-HT) (Enzo Life Sciences), 5-HT metabolite 5-hydroxyindoleacetic acid (5-HIAA) (Elabscience), tryptophan (TRP) (LDN Labor Diagnostika Nord GmbH & Co.KG), kynurenine (KYN) (MyBioSource) and quinolinic acid (QUIN) (Cloud Clone Corp.) were determined by ELISA quantification. The results were expressed as ratio or ng/gram wet tissue and nM respectively.\n\nMalondialdehyde (MDA) aAssay {#sec008}\n----------------------------\n\nHippocampal MDA levels were examined by Bioxytech LPO-586TM kit (OxisResearch, Portland, OR). Briefly, according to manufacturer's protocol, the reaction products were detected at 586 nm and the standard curves was constructed with 1,1,3,3-tetraethoxypropan. Protein amount of each samples were measured by Bio-Rad Protein Assay Kit (Bio-Rad, Hercules, CA). The results were normalized by protein amount. The data was expressed as \u03bcmol/mg and percentage of the control.\n\nGSH/GSSG ratio {#sec009}\n--------------\n\nHippocampi were first homogenized in 5% metaphosphoric acid. Then, the homogenized lysate was centrifuged at 14,000g for 15min at 4\u00b0C to obtain the supernatant for glutathione determination (Enzo Life Sciences). Total amount of glutathione (GSH) and oxidized glutathione (GSSG) were determined according to manufacturer's protocol. For reduced GSH, it was calculated as below: Reduced GSH = Total Glutathione--Oxidized GSSH.\n\nGolgi staining {#sec010}\n--------------\n\nHippocampal CA1 and CA3 pyramidal neurons were visualized with the use of Golgi staining kit (FD Rapid Golgistain Kit (FD Neurotechnologies, MD). Dendritic spine density, length, and soma area were analyzed with the use of Neurolucida software (MicroBright-Field, USA). Soma volume was calculated with the below mathematical equation: 4/3\u03c0r^3^. Neurons chosen for analysis should be relatively separated in order to avoid the interference with neighboring impregnated neurons. Five neurons from each 100\u03bcm-thick brain section were sampled and analyzed. For the dendritic length and spine density, three to five dendrites with at least one branch point were selected for counting. Visible spines along the branch segment were counted and the spine density was expressed as number/10\u03bcm. The dendritic length, area and volume of soma were expressed as \u03bcm, \u03bcm^2^ and \u03bcm^3^ respectively.\n\nBehavioral tests {#sec011}\n----------------\n\nIn forced swimming test (FST), rats were put in a cylinder (60cm height X 25 cm diameter) containing tap water for a 15-min training session to learn helplessness on the following day after the hypoxic treatment ([Fig 1](#pone.0177940.g001){ref-type=\"fig\"}). The animals were put into the cylinder for 5 minutes on the next day and were recorded on a video-tape for the post-analysis of the immobility time. The results were expressed in seconds. The immobility time serves as an indicator of behavioral despair, which is a main phenotype observed in rat models for depression and clinically depressed patients.\n\n![Hypoxia induced depressive-like behavior in the rat, which was significantly prevented by the M30 administration.\\\nPanels A and B represent the immobility time and percentage of sucrose consumption of the rats in normoxic (Nx), hypoxic (IH) groups; M30-treated hypoxic (IH+M30) or normoxic (Nx+M30) groups respectively. Data are mean \u00b1 SEM (n = 12). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g001){#pone.0177940.g001}\n\nIn sucrose preference test (SPT), rats were individually caged and provided the training session in which two bottles of 1% (wt/vol) sucrose solution for 24 hours on the following day after the intermittent hypoxic treatment to prevent the subtle stress when applying the sucrose consumption assessment. After that, the rats were provided one bottle of water and the other contained 1% (wt/vol) sucrose solution for 24 hours. The positions of bottles were swapped at the middle of the assessment to avoid the bias towards a particular side. No food and water deprivation was applied before the test to prevent the interference with the metabolic demands of the rats. The consumption of water and sucrose solution was recorded by weighing bottles before and after the test. Results were presented as the percentage of the sucrose solution over the total weight of liquid consumed and expressed as percentage of the control.\n\nStatistical analysis {#sec012}\n--------------------\n\nData from each group were expressed as mean \u00b1 SEM. Statistical comparisons among groups were performed using One way ANOVA followed by Tukey's post-hoc test for multiple comparisons with the use of Graphpad Prism software (Graphpad Software Version 5.01, Inc., San Diego, USA). A p\\<0.05 is considered as significant difference between groups and indicated by different letter (e.g. a and b).\n\nResults {#sec013}\n=======\n\nHypoxia-induced depressive behavior in rats {#sec014}\n-------------------------------------------\n\nBehavioral despair and hedonic status of the rats were assessed by the immobility time of forced swimming test (FST) and percentage of sucrose consumption of reward-based sucrose preference test (SPT) respectively. The immobility time of hypoxic group was doubled when compared with that of the normoxic control (n = 12, [Fig 1](#pone.0177940.g001){ref-type=\"fig\"}). The percentage of sucrose consumption of the hypoxic group was reduced by 35% of the control. These values were not significantly different between the M30-treated groups and the control, supporting a prophylactic effect of M30 against hypoxia-induced depressive behavior.\n\nHypoxia elevated the hippocampal MAO-A activity {#sec015}\n-----------------------------------------------\n\nLevels of the expression and activity of MAO-A in the hypoxic group were significantly elevated by two folds of the control (n = 8, [Fig 2](#pone.0177940.g002){ref-type=\"fig\"}), but there were no changes in the expression and activity of MAO-B. Also, the 5-HIAA/5-HT ratio was markedly increased in the hypoxic group (n = 8, [Fig 2](#pone.0177940.g002){ref-type=\"fig\"}). No significant differences were found between the M30-treated hypoxic group and the control. Of note, MAO-A and MAO-B activities were lowered by half in the M30-treated groups. Results suggest that the upregulation of MAO-A induced by hypoxia resulted in a deregulated serotonin metabolism, which was neutralized by M30.\n\n![Hypoxia increased the metabolic turnover of serotonin mediated by the elevated activity of MAO-A but not MAO-B in the hippocampus.\\\nSerotonin (5-HT) level was also lowered in the hypoxic group. These alterations were significantly attenuated by M30. Levels of protein expression of (A) MAO-A, (B) MAO-B, activity of (C) MAO-A, (D) MAO-B, (E) 5-HT, (F) 5-HIAA and (G) 5-HIAA/5-HT are summarized. \u03b2-actin was the internal control. Data are presented as Mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g002){#pone.0177940.g002}\n\nNeuroarchitectural alterations induced by hypoxia {#sec016}\n-------------------------------------------------\n\nGolgi staining was performed to evaluate dendritic structural changes of hippocampal pyramidal CA1 and CA3 neurons. We observed dendritic abnormalities of both apical and basal branches of neurons shown by reduced dendritic spine densities and length and also the area and volume of somata in the hypoxic group (n = 12, [Fig 3](#pone.0177940.g003){ref-type=\"fig\"}). Also, a significant increase in the dendritic varicosity formations was observed in the hypoxic group. Administration of M30 effectively ameliorated neurodegeneration induced by hypoxia and no observable aberrant alterations were found in the M30-treated groups.\n\n![Hypoxia induced decreases in dendritic spine density, length, soma size and volume of the apical and basal branches of CA1 and CA3 pyramidal neurons.\\\nLevels of the dendritic spine densities of (A) CA3 basal branch, (B) CA3 apical branch, (E) CA1 basal branch, (F) CA1 apical branch; Levels of the dendritic length of (C) CA3 basal branch, (D) CA3 apical branch, (G) CA1 basal branch, (H) CA1 apical branch; Levels of the cell body area (I) CA3 pyramidal neuron, (J) CA1 pyramidal neuron; Levels of the cell body volume (K) CA3 pyramidal neuron, (L) CA1 pyramidal neuron in the hippocampus of the normoxic (Nx) or hypoxic (IH) groups; M30-treated hypoxic (IH+M30) or normoxic (Nx+M30) groups are summarized. Dendritic varicosity formations responsible for reduction in dendritic spine density were indicated by the arrows. Panel M summarized representative photomicrographs of apical and basal dendrites of both CA3 and CA1 pyramidal neurons showing dendritic spines. Magnification:100X, Scale bar = 10\u03bcm. Panel N illustrated Camera Lucida drawing of representative CA3 and CA1 pyramidal neurons. Data are presented as Mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g003){#pone.0177940.g003}\n\nInhibition of MAO-A by M30 mitigated hypoxia-induced oxidative stress {#sec017}\n---------------------------------------------------------------------\n\nHydrogen peroxide is one of the products of deamination by MAO-A, which is a source of reactive oxygen species (ROS) leading to oxidative stress. The level of MDA, a lipid peroxidation marker, was significantly elevated by two folds in the hypoxic group (n = 8, [Fig 4](#pone.0177940.g004){ref-type=\"fig\"}). Also, there was a significant decrease in the GSH/GSSG ratio (n = 8, [Fig 4](#pone.0177940.g004){ref-type=\"fig\"}). In addition, levels of protein expressions of antioxidant enzymes SOD-2 and GPx-1 were lowered, respectively, by 50% and 60% in the hypoxic group when compared to the control (n = 8, [Fig 4](#pone.0177940.g004){ref-type=\"fig\"}). There were no significant differences between the M30-treated groups and the control.\n\n![Administration of M30 attenuated oxidative stress induced by hypoxia.\\\nLevels of (A) MDA content, (B) GSH/GSSG ratio and the protein expression of (C) SOD-2 and (D) GPx-1 and in the hippocampus of the normoxic (Nx) or hypoxic (IH) groups; M30-treated hypoxic (IH+M30) or normoxic (Nx+M30) groups are summarized. \u03b2-actin was an internal control. Data are mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g004){#pone.0177940.g004}\n\nHypoxia induced redox-sensitive NF\u03baB canonical pathway-dependent inflammation {#sec018}\n-----------------------------------------------------------------------------\n\nThe level of I\u03baB\u03b1 was significantly decreased by 80% in the hypoxic group when compared to the control. Also, levels of NF\u03baB p65 and p50 in the hypoxic group were significantly reduced in the cytosolic fraction and were markedly increased in the nuclear fraction (n = 8, [Fig 5](#pone.0177940.g005){ref-type=\"fig\"}). Besides, there were significant increases in the protein expression of inflammatory cytokines TNF\u03b1, IL-1\u03b2, IL-6 and COX-2 by about two folds in the hypoxic group when compared to the control (n = 8, [Fig 6](#pone.0177940.g006){ref-type=\"fig\"}). M30 treatment normalized the I\u03baB\u03b1 degradation, nuclear translocation of NF\u03baB and cytokine expression to the control level.\n\n![Hypoxia induced the degradation of I\u03baB\u03b1 in the redox-sensitive NF\u03baB canonical pathway and the nucleus translocation of NF\u03baB member p65 and p50.\\\nLevels of (A) I\u03baB\u03b1, nuclear protein expression of (B) p65, (C) p50, cytosolic protein expression of (D) p65, (E) in the hippocampus of the normoxic (Nx) or hypoxic (IH) groups; M30-treated hypoxic (IH+M30) or normoxic (Nx+M30) groups are summarized. Lamin B1 and \u03b2-actin were an internal control of the nuclear fraction and cytosolic fraction respectively. Data are mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g005){#pone.0177940.g005}\n\n![Hypoxia induced inflammation and activated cytokine-responsive IDO-1 in the hippocampus.\\\nLevels of the protein expression (A) TNF\u03b1, (B) IL-1\u03b2, (C) IL-6, (D) COX-2, (E) IDO-1, (F) the ratio of KYN/TRP (IDO-1 activity) and (G) the level of QUIN are summarized. Data are presented as Mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g006){#pone.0177940.g006}\n\nThe expression and activity of IDO-1 activated by hypoxia {#sec019}\n---------------------------------------------------------\n\nIDO-1 activity were evaluated by the ratio changes of its substrate tryptophan and enzymatic metabolite kynurenine. The expression level of IDO-1 and the ratio of kynurenine to tryptophan were significantly increased, respectively, by two folds and four folds in the hypoxic group (n = 8, [Fig 6](#pone.0177940.g006){ref-type=\"fig\"}). Consistently, the level of downstream IDO-1 metabolite QUIN was also increased by hypoxia. M30 treatment mitigated the elevated level of IDO-1 expression and activity, and the increased level of QUIN.\n\nHypoxia induced losses of synaptic proteins and neuronal apoptosis {#sec020}\n------------------------------------------------------------------\n\nThere were dramatic decreases in the expression of pre-synaptic proteins synapsin-1 and synaptophysin, and post-synaptic protein PSD95 in the hypoxic group, respectively, by 55%, 50% and 68% of the control (n = 8, [Fig 7](#pone.0177940.g007){ref-type=\"fig\"}). Furthermore, the expression of anti-apoptotic protein Bcl-2 was much lowered by 55% of the control and the expression of apoptotic markers cleaved caspase 3 and cleaved PARP-1 was doubled in the hypoxic group (n = 8, [Fig 7](#pone.0177940.g007){ref-type=\"fig\"}). Administration of M30 effectively ameliorated losses of synaptic proteins and neuronal apoptosis induced by hypoxia.\n\n![Hypoxia induced neuronal apoptosis and losses of pre-synaptic vesicle proteins and post-synaptic.\\\nLevels of protein expression of (A) Bcl-2, (B) Cleaved Caspase 3, (C) Cleaved PARP-1, (D) Synapsin-1, (E) Synaptophysin and (F) PSD-95 in the hippocampus of the normoxic (Nx) or hypoxic (IH) groups; M30-treated hypoxic (IH+M30) or normoxic (Nx+M30) groups are summarized. \u03b2-actin was an internal control. Data are presented as Mean \u00b1 SEM (n = 8). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g007){#pone.0177940.g007}\n\nHypoxia increased MAO-A expression in SH-SY5Y cells {#sec021}\n---------------------------------------------------\n\nIn human SH-SY5Y cells expressing MAO-A but not MAO-B, the level of protein expression of MAO-A was significantly increased by 50% following the hypoxic treatment ([Fig 8](#pone.0177940.g008){ref-type=\"fig\"}). The level of MAO-A expression was normalized by the pretreatment of the cells with M30 or clorgyline. Hypoxia also significantly reduced the GSH to GSSG ratio, which was partially normalized by clorgyline ([Fig 8](#pone.0177940.g008){ref-type=\"fig\"}).\n\n![The level of protein expression of MAO-A was increased by exposure to intermittent hypoxia (IH) in SH-SY5Y cells, which was antagonized by M30 at 1\u03bcM or clorgyline at 10\u03bcM (Panel A and B). Hypoxia significantly reduced the GSH/GSSG ratio in SH-SY5Y cells, which was partially attenuated by clorgyline at 10\u03bcM (Panel C). Data are mean \u00b1 SEM (n = 6). Statistical comparisons between groups were performed using the One way Anova followed by Tukey post hoc test to detect differences in all groups. A p \\< 0.05 was considered to be statistically significant. Different letters (e.g. a and b) mean a statistical significant change between each other.](pone.0177940.g008){#pone.0177940.g008}\n\nDiscussion {#sec022}\n==========\n\nThis is the first report to delineate a pathophysiological role of the anomalous MAO-A activity upregulated by chronic intermittent hypoxia in oxidative stress and inflammation, which significantly activates the IDO-1 activity and contributes to serotonin deficiency and neurodegeneration closely associated with the clinical manifestation of the brain of patients with depression. This may help to explain high rates of depression in both community (17%) and clinical (21--41%) populations with obstructive sleep apnea (OSA) \\[[@pone.0177940.ref002], [@pone.0177940.ref004], [@pone.0177940.ref024]--[@pone.0177940.ref026]\\]. Our results are clinically relevant in its proposed pathophysiological causality between OSA and depression. Yet, we are aware of the limitation of the rodent model with the hypoxic treatment in an attempt to link degenerative neuropathies to the arterial oxygen desaturation simulating the pathophysiological consequences in a severe OSA condition, which is an important parameter in the clinical assessment. Intriguingly, we demonstrated the prophylactic effect of the synthetic compound M30 against the aberrant changes in the hippocampus and depressive behavior, which is indicative of targeting the MAO-A activity, oxidative stress and inflammation to retard the neurodegeneration induced by hypoxia. The administrative dose of M30 has been reportedly safe for the experimental usage and did not cause any noticeable physiological or histopathological changes in the healthy animals \\[[@pone.0177940.ref027]\\].\n\nA major finding of this study is that MAO-A but not the MAO-B activity was up-regulated in the hippocampus of hypoxic rats, which is in consistent with the increased MAO-A expression. This could cause an increased catabolic deamination of 5-HT resulting in increases in the level of the metabolite 5-HIAA and its ratio to 5-HT, leading to a lowered serotonin availability that were reportedly shown in the brain of clinically depressed patients and animal models of depression \\[[@pone.0177940.ref028], [@pone.0177940.ref029]\\]. These findings are also in agreement with studies reporting that MAO-A inhibitors could significantly improve depressive symptoms in the patient and rodent models \\[[@pone.0177940.ref030]--[@pone.0177940.ref032]\\]. Importantly, our results provide novel evidence to link the upregulation of MAO-A to neurodegeneration as a pathophysiological consequence of chronic intermittent hypoxia. We also demonstrated evidence for a proof of concept that the daily M30 administration significantly antagonized the MAO-A upregulation and consequently prevented the depressive behavior induced by hypoxia. Blockade of MAO-A activity could be a therapeutic approach to preventing the neurodegeneration in patients with severe conditions of OSA that is known to increase the risk for depression.\n\nOxidative stress is a significant adverse consequence of chronic intermittent hypoxia due to an increased production of reactive oxygen species (ROS). In this context, the upregulated MAO-A activity could result in an increased enzymatic production of hydrogen peroxide as a by-product of the catalytic deamination of monoamines. In effect, elevated levels of ROS deplete endogenous GSH antioxidant capacity and cause the lipid peroxidation of membrane forming MDA. Indeed, our results showed an elevated level of oxidative stress in the hippocampus of hypoxic rats with significant decreases in the GSH to GSSG ratio and increases in the MDA level. Also, the protein expression of antioxidant enzymes SOD-2 and GPx-2 was significantly lowered. These changes were markedly neutralized by the M30 treatment, strongly supporting that the chemical properties of M30, namely for the MAO inhibition and scavenging ROS, were effective to antagonize the pathophysiological consequences induced by chronic intermittent hypoxia. M30 is also highly permeable to the blood brain barrier and selective to the brain MAO-A activity \\[[@pone.0177940.ref033]\\], which is potentially an advantage to limit the non-specific side effects of MAO inhibitors.\n\nInflammation has been proposed as a prominent factor in the induction of depressive symptoms, which could be mediated by the activation of cytokine-responsive IDO-1. Studies have also shown that neuroinflammation was closely linked to oxidative stress via redox sensitive NF\u03baB canonical pathway, forming a vicious cycle in the production of ROS and inflammatory cytokines, causing tissue damages \\[[@pone.0177940.ref023], [@pone.0177940.ref034]\\]. Indeed, we found significant increases in the degradation of I\u03baB\u03b1 and the translocation of NF\u03baB p65 and p50 from the cytosol to nucleus. Consequently, the activation of NF\u03baB canonical pathway results in the elevated expression of inflammatory cytokines TNF\u03b1, IL-1\u03b2, IL-6 and COX-2. Intriguingly, the M30 treatment remarkably mitigated the activation of redox-sensitive NF\u03baB cascade and the expression of cytokines, which is consistent with an anti-inflammatory property of M30 reported recently \\[[@pone.0177940.ref022]\\].\n\nThe augmented IDO-1 activity has been proposed to play a pathogenic role in the catabolic conversion of tryptophan to kynurenine, causing a decrease in tryptophan availability for the 5-HT biosynthesis leading to serotonin deficiency \\[[@pone.0177940.ref035], [@pone.0177940.ref036]\\]. Our results showed a significant increase in the IDO-1 expression and activity in parallel with a significantly lowered 5-HT level in the hippocampus of hypoxic rats. These were in consistent with findings of previous reports showing 5-HT depletion as a result of elevated IDO-1 activities could be restored by the administration of anti-inflammatory agents \\[[@pone.0177940.ref037], [@pone.0177940.ref038]\\]. In fact, we showed that M30 significantly decreased MAO-A activity and oxidative stress, which could lessen the activation of redox-sensitive NF\u03baB cascade and the expression of cytokines that elevate the cytokine responsive IDO-1 activity in the hippocampus of hypoxic rats.\n\nNeurodegeneration is an important neuropathic feature of depression observed in both patients and animal models \\[[@pone.0177940.ref039], [@pone.0177940.ref040]\\], which is closely associated with and attributed to apoptosis as a result of oxidative stress and inflammation leading to the activation of intrinsic and extrinsic cascades \\[[@pone.0177940.ref023], [@pone.0177940.ref041]\\]. These are illustrated in our results showing significant increases in the level of apoptotic markers cleaved caspase 3 and cleaved PARP 1, and marked decreases in the level of anti-apoptotic protein Bcl-2. By targeting MAO-A upregulation, oxidative stress and inflammation, M30 could abrogated neuronal apoptosis induced by hypoxia, which is in line with previous reports illustrating anti-apoptotic property of M30 in the animal models of Parkinson's diseases and stress-induced depression \\[[@pone.0177940.ref020], [@pone.0177940.ref042]\\].\n\nLosses of synaptic proteins were observed in the brains of patients and animals with depression and were proposed to be closely related to the onset of depression \\[[@pone.0177940.ref043], [@pone.0177940.ref044]\\]. We found that there were significant down-regulation of expressions of pre-synaptic vesicular proteins synapsin-1 and synaptophysin, and post-synaptic protein PSD95 in the hippocampus of hypoxic rats, which were remarkably ameliorated by the M30 treatment. This is in agreement with the observation that the degradation of synaptic proteins was mediated by apoptotic caspase enzymes \\[[@pone.0177940.ref045]\\]. These results give strong support to the contention that M30 confers protective effects against synaptic degeneration induced by hypoxia.\n\nGolgi staining vividly revealed adverse effects on the neuroarchitecture induced by hypoxia. There were remarkable reductions in the dendritic spine density and dendritic length on both apical and basal branches of the CA1 and CA3 pyramidal neurons in the hippocampus of hypoxic rats. It has been reported that hypoxia could induce NMDA overactivation that leads to excitotoxicity and causes ultrastructural damages in the hippocampus \\[[@pone.0177940.ref046]\\]. In fact, NMDA excitotoxicity is a major cause of varicosities formation in the dendritic spines and decreases the spine density \\[[@pone.0177940.ref047], [@pone.0177940.ref048]\\]. Particularly, the catabolic degradation of tryptophan by IDO-1 produces neurotoxins such as quinolinic acid that agonistically effect on the activation of NMDA receptors and potentially cause the formation of varicosities when IDO-1 activity is augmented under hypoxic conditions. The area and volume of the soma were markedly decreased in the hippocampus of hypoxic rats, which are in line with the pathogenic processes of DNA fragmentation and condensation in neuronal apoptosis \\[[@pone.0177940.ref049]\\]. These defective changes in the spines and somata were not found in the rat treated with M30, strongly supporting a pathogenic role of MAO-A in oxidative stress and inflammation leading to IDO-1 activation that causes the neurodegeneration. Results are also in consistent with the observation that MAO-A inhibitors could rescue the abnormal dendritic structure upon exposure to chronic stress \\[[@pone.0177940.ref050]\\].\n\nFinally, the effect of hypoxia on the MAO-A upregulation was confirmed by the findings in cultured human SH-SY5Y cells, which constitutively express MAO-A but not MAO-B. The MAO-A expression was significantly upregulated in the hypoxic group, providing evidence to argue a non-specific effect of hypoxia on the MAO-A upregulation. The MAO-A upregulation was blocked by M30 or clorgyline, suggesting a crucial role of MAO-A in the pathophysiological cascade initiated by chronic intermittent hypoxia, which drives the adverse consequences towards neurodegeneration. This is summarized in [S1 Fig](#pone.0177940.s001){ref-type=\"supplementary-material\"}. illustrating the pathophysiological mechanism underlying the onset of neurodegeneration and depressive behavior induced by chronic intermittent hypoxia, which also highlights the effect of M30 targeting the MAO-A that significantly contribute to oxidative stress and inflammation leading to serotonin deficiency and neurodegeneration.\n\nConclusions {#sec023}\n===========\n\nChronic intermittent hypoxia upregulates the expression and activity of MAO-A resulting in oxidative stress, inflammation, activation of IDO-1, causing serotonin deficiency and neurodegeneration, which could be potently antagonized by M30 against the upregulation of MAO-A activity and the production of free radicals. MAO-A is an indicative drug target to prevent the onset of depressive symptoms in OSA patients.\n\nSupporting information {#sec024}\n======================\n\n###### A schematic summary of the pathophysiological cascade induced by chronic intermittent hypoxia (CIH) leading to depression.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nWe are grateful for the technical support of Mr. Y.M. Lo. The authors declare no competing financial interests. This work was supported by internal funding (104004440) granted by The University of Hong Kong and the General Research Fund (17102316) of Research Grants Council, Hong Kong.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: **Conceptualization:** MLF GLT CSL.**Data curation:** MLF GLT CSL JJL.**Formal analysis:** CSL JJL.**Funding acquisition:** MLF.**Investigation:** MLF GLT CSL JJL.**Methodology:** MLF GLT CSL JJL.**Project administration:** MLF.**Resources:** MLF GLT MBHY.**Software:** MLF GLT.**Supervision:** MLF GLT.**Validation:** MLF GLT CSL JJL.**Visualization:** MLF GLT CSL JJL.**Writing -- original draft:** CSL.**Writing -- review & editing:** MLF.\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec005}\n============\n\nAmniotic fluid (AF) is maintained in a dynamic equilibrium and its volume derived from the sum of inflow (from fetal urine and lung fluid) and outflow (fetal swallowing and intramembranous absorption) of fluid from the amniotic space.\\[[@pone.0144334.ref001]\\] Amniotic fluid volume (AFV) is an important parameter in the assessment of fetal wellbeing since it provides a number of functions vital to fetal development such as a supportive environment for growth,, protection from trauma and infection and a medium which allows fetal movement thus promoting the development of the musculoskeletal system. AF also prevents a possible compression of the umbilical cord and placenta thereby protecting the fetus from vascular and nutritional compromise.\\[[@pone.0144334.ref002]\\]\n\nTo date several methods are used to assess AFV ranging from subjective assessment (where the volume is described as average, above average, below average or scant) to semi-quantitative estimations including measurement of the deepest vertical pocket and amniotic fluid index (AFI).\\[[@pone.0144334.ref003]\\]\n\nWhile debate continues regarding the best method to estimate AFV, it has become evident that there are numerous maternal and fetal risk factors associated with a reduction of this parameter. \\[[@pone.0144334.ref004]--[@pone.0144334.ref008]\\] Indeed, an increased fetal and neonatal morbidity and mortality is described when pregnancies are complicated by oligohydramnios.\\[[@pone.0144334.ref009],[@pone.0144334.ref010]\\]\n\nGenerally about 3--5% of pregnancies are complicated by oligohydramnios, and in less than half of the cases, the diagnosis is made in the absence of maternal-fetal risk factors and is therefore defined \"isolated oligohydramnios\" (IO).\\[[@pone.0144334.ref011]\\]Nevertheless, the lack of a comprehensive understanding of the physiology and dynamics of IO contribute to the unresolved dilemma.\\[[@pone.0144334.ref011]\\]\n\nEmerging evidences suggest that oligohydramnios is only a weak predictor of poor perinatal outcome,\\[[@pone.0144334.ref012]\\] and, while in term pregnancies an appropriate threshold for intervention seems to be good medical practice,\\[[@pone.0144334.ref013]\\] in preterm IO conservative management and strict follow up are justified and should be considered gold standard.\n\nOccasionally IO may be diagnosed in pregnant women with a personal history of insufficient fluid intake. This fact triggered speculation regarding the potential role of maternal dehydratation in contributing to the development of this \"borderline condition\".\\[[@pone.0144334.ref014]\\] Maternal hydration (MH) has been proposed as a possible effective treatment for the conservative management of IO during pregnancy and prior to labour commencing.\\[[@pone.0144334.ref015]\\]\n\nSeveral trials have been conducted \\[[@pone.0144334.ref014]--[@pone.0144334.ref029]\\] to evaluate the efficacy of MH but the heterogeneity in patient selection criteria, sonographic diagnostic criteria, implementation of different hydration protocols and outcomes measured generated considerable confusion in defining the utility of maternal hydration.\\[[@pone.0144334.ref030],[@pone.0144334.ref031]\\] Despite the feeling that simple MH may increase AFV and be beneficial in the management of oligohydramnios, the last of two meta-analysis conducted concluded that further controlled trials are needed to assess the clinical benefits and possible risks of MH for specific clinical purposes.\\[[@pone.0144334.ref032]\\]\n\nAuthors evaluated the effects of MH in patients with normal AFV which may be of interest in obstetric care since the bid to reduce the rate of unnecessary caesarean section compels us to search for new strategies to facilitate vaginal deliveries.\\[[@pone.0144334.ref033]--[@pone.0144334.ref036]\\] The amount of AF is associated with the success of external cephalic version \\[[@pone.0144334.ref037]\\] and a reduction in its volume at term may increase the risk of caesarean section\\[[@pone.0144334.ref038]\\]. It is therefore necessary to definitively clarify the efficacy, the most effective strategy with which to administer MH and its clinical significance in improving the AFV in both pregnancies complicated by IO and pregnancies with normal amniotic volume.\n\nThe aim of this systematic-literature-review and meta-analysis was to collect all data regarding proposed strategies and their efficacy in relation to each clinical condition for which MH-therapy was performed with the aim of increasing amniotic-fluid (AF) and improving perinatal outcomes. In detail, the primary outcome was to determine, in cases of IO, whether MH-therapy could significantly improve AVF as compared with no treatment, the most effective hydration strategy and most importantly whether such intervention is associated with improvements in perinatal outcomes. Secondary outcome was to evaluate if the AVF variation in IO pregnancies can be comparable to that of patients with normal AFV.\n\nMaterials and Methods {#sec006}\n=====================\n\nData Sources {#sec007}\n------------\n\nA systematic literature search (English literature only) was conducted in electronic databases MEDLINE, EMBASE, ScienceDirect and the Cochrane library in the time interval between January 1991 and December 2014.\n\nAll studies which evaluated the effects of MH (including oral, intravenous or a combination of both strategies) were collected and analyzed.\n\nSearch strategy {#sec008}\n---------------\n\nThe Key search terms included: Ultrasound/Ultrasonography/Transabdominal sonography \\[Mesh\\] and amniotic fluid volume, adding the subsequent subheadings: isolated oligohydramnios OR abnormality in amniotic fluid OR oral fluid intake OR maternal hydration OR intravenous fluid administration OR hypotonic fluid OR water OR improvement in amniotic fluid index OR normal amniotic fluid index OR idiopathic abnormalities in amniotic fluid OR gestational age at diagnosis OR gestational age at delivery OR type of delivery OR neonatal adverse events.\n\nAfter screening of titles, abstracts and full texts, the selection of included studies was based on the availability of information regarding AFV, type of maternal fluid intake (hypotonic or isotonic), route of fluid administration (oral, intravenous or both), length of treatment (hours, days, weeks), gestational age at diagnosis, gestational age at delivery, type of delivery, and neonatal adverse events.\n\nA second search was performed by using the same criteria but instead searched for AFV changes in patients with normal AFI.\n\nStudies were selected in a 2-stage process. Titles and abstracts from electronic searches were scrutinized by 2 reviewers independently (A.V; ADA) and full manuscripts and their citations list were analyzed by a third reviewer (M.N) to retrieve missing articles and to select eligible manuscripts according to inclusion and exclusion criteria.\n\nInclusion/exclusion criteria {#sec009}\n----------------------------\n\nWe considered eligible all clinical trials comparing the effect of MH between interventional and control groups (no hydration, different route or type of fluid administration, different time of intervention) using a random allocation strategy with adequate allocation concealment and without violations of allocated management or exclusions after allocation which were insufficient to materially affect outcomes. We considered trials which included in the treatment group pregnant women diagnosed with IO in the second half of pregnancy and as controls women with a diagnosis of IO or with normal AFV either receiving no hydration or a different treatment protocol from that of the intervention group (considering volume and type of fluid, route of administration and length of treatment). To be considered eligible, data must be reported clearly and allow for statistical comparison regarding variation in AFV after intervention/no intervention.\n\nWe excluded studies in which the diagnosis of oligohydramnios was made by methods different from AFI, when outcomes measured did not minimize observer bias, when missed data influenced conclusions, when no data was available for analysis according to original allocation, or data format was not suitable for analysis.\n\nData extraction and management {#sec010}\n------------------------------\n\nWe designed a form for data extraction. In the eligible studies, two Authors (S.G; T.S.P.) extracted the data using the agreed upon form. We resolve discrepancies through discussion or, if required, by consulting the Authors of the original reports. Data was entered into the pre-installed Data Sheet form of free Excel software extension for meta-analysis (software MetaEasy 2013 version 1.0.5) and checked for accuracy.\n\nRegarding the strategies of maternal hydration we clustered trials and cohorts of patients according to the following outcome measures:\n\n1.  Route of administration: oral versus intravenous versus combination;\n\n2.  Total daily dose of fluid administered: less than 2000 versus more than 2000\n\n3.  Length of hydration treatment: 1 day, more than 1 day but less than 1 week, more than 1 week\n\n4.  Type of fluid administered: isotonic versus hypotonic versus combination of both\n\nRisk of bias (quality of the included studies) {#sec011}\n----------------------------------------------\n\nThe methodological quality of each study was evaluated with QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies 2).\\[[@pone.0144334.ref039]\\]\n\nThe risk of bias in patient selection, index test, reference standard, flow and timing as well as the concerns for applicability related to the first three domains are shown in [S1 Fig](#pone.0144334.s001){ref-type=\"supplementary-material\"}. Prisma checklist 2009 version is shown in [S1 PRISMA Checklist](#pone.0144334.s003){ref-type=\"supplementary-material\"}.\n\nEndpoints {#sec012}\n---------\n\nPrimary endpoint was to compare different strategies of MH versus no treatment in terms of \u0394 variation in amniotic fluid index (AFI), considering only the cohort of patients with diagnosis of IO.\n\nSecondary endpoint was to compare protocols in terms of type of fluid (isotonic versus hypotonic), route of, (intravenous versus oral) and length of fluid administration (single day versus multiple days) in order to evaluate \u0394 variation in AFI in the cohorts of patients diagnosed with IO.\n\nTertiary endpoint was to compare different strategies of MH in terms of \u0394 variation in AFI considering the cohorts of patients with diagnosis of IO versus patients with normo-hydramnios.\n\nFinally we evaluated \u0394 variation in AFI after different strategies of MH in a cohort of women with normo-hydramnios.\n\nAdditionally, when possible, we correlated \u0394 variation in AFI with mean gestational age at diagnosis, mean gestational age at delivery, type of delivery and percentage of adverse neonatal outcomes considering only the cohort of patients with diagnosis of IO.\n\nStatistical analysis {#sec013}\n--------------------\n\nThe Meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines using Comprehensive Meta-Analysis software V.2.2 (Biostat, Englewood, New Jersey, USA).\n\nFor dichotomous data, we presented results as summary risk ratio with 95% confidence intervals. For continuous data, we used the mean difference if outcomes were measured similarly between trials. We used the standardized mean difference to combine trials that measured the same outcome, but by different methods.\n\nThe impact of including studies with high levels of missing data in the overall assessment of treatment was explored by sensitivity analysis. For all outcomes, analysis was carried out, when possible, on an intention-to-treat basis, i.e. we attempted to include all participants randomized to each group in the analysis. The denominator for each outcome in each trial was the number of patients randomized minus any participants whose outcomes were known to be missing. We used the I^2^ statistic to measure heterogeneity among the trials in each analysis. If we identified substantial heterogeneity (I^2^ \\>50%) we would explore it by pre-specified subgroup analysis.\n\nWhere we suspected a reporting bias we attempted to contact study authors requesting missing outcome data. Where not possible, and the missing data was thought to introduce serious bias, the impact of including such studies in the overall assessment of results was explored by a sensitivity analysis.\n\nWe used fixed-effect inverse variance meta-analysis for combining data where trials were examining the same intervention, and the trials' populations and methods are judged sufficiently similar. Where we suspected clinical or methodological heterogeneity between studies sufficient to suggest that treatment effects may differ between trials we used random-effects meta-analysis.\n\nIf substantial heterogeneity was identified in a fixed-effect meta-analysis this was noted and the analysis repeated using a random-effects method. We carried out sensitivity analysis to assess the effect of including trials with greater risk of bias, if there were sufficient trials.\n\nResults {#sec014}\n=======\n\nUsing the above mentioned key search strategy, we identified 82 potentially relevant papers. Only 16 of the above manuscripts, reporting data collected in a cohort of 1121 pregnant women, were included in this meta-analysis after application of the inclusion and exclusion criteria.\\[[@pone.0144334.ref014]--[@pone.0144334.ref029]\\]\n\nOf these 16 articles, 8 met the inclusion criteria for the first endpoint \\[[@pone.0144334.ref015],[@pone.0144334.ref016],[@pone.0144334.ref018],[@pone.0144334.ref020],[@pone.0144334.ref021],[@pone.0144334.ref025]--[@pone.0144334.ref027]\\], 4 for the second endpoint \\[[@pone.0144334.ref014],[@pone.0144334.ref018],[@pone.0144334.ref020],[@pone.0144334.ref022]\\], 2 for the third endpoint \\[[@pone.0144334.ref019],[@pone.0144334.ref024]\\] and 4 for the fourth endpoint \\[[@pone.0144334.ref017],[@pone.0144334.ref023],[@pone.0144334.ref028],[@pone.0144334.ref029]\\].\n\nA full report regarding authors, study design, sample size, epidemiological features of studied population, type of intervention and outcome measures is shown in detail in [Table 1](#pone.0144334.t001){ref-type=\"table\"} and in the flow diagram ([S2 Fig](#pone.0144334.s002){ref-type=\"supplementary-material\"}).\n\n10.1371/journal.pone.0144334.t001\n\n###### Descriptive analysis of trials included in the meta-analysis.\n\n![](pone.0144334.t001){#pone.0144334.t001g}\n\n  AUTHORS and YEAR                STUDY SETTING \\[sample size\\]   PATIENTS & METHODS                                                                                                                                                                                                                                                                                                                                                                              RESULTS                                                                                                                                                                              GW AT DIAGNOSIS                                             GW AT DELIVERY                 CS RATE        NEONATAL DISTRESS/ADVERSE EVENTS RATE   CONCLUSIONS\n  ------------------------------- ------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ----------------------------------------------------------- ------------------------------ -------------- --------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  ***Kilpatrick et al 1991***     RCT \\[**36**\\]                  *IO (AFI\\<6 cm)*. **Study Group**: (n = 19) oral hydration 2000 ml/2-4 h. **Control Group** (n = 17): routine hydration (n = 10), routine hydration plus 100 ml water (n = 7)                                                                                                                                                                                                                   The mean AFI increased significantly in Study Group (p\\<0.01).                                                                                                                       AT TERM 37\u00b14.8 VS 39\u00b12.4                                    NR                             NR             NR                                      *Oral MH increases AFI in women with decreased AFV*.\n  ***Kilpatrick et al 1993***     RCT \\[**40**\\]                  *Normal AFI index (7--24 cm)*. **Study Group**: (n = 20) oral hydration of 2000 ml/2 h **Control Group**: (n = 20) oral hydration of 100 ml/2 h                                                                                                                                                                                                                                                 The mean AFI increased significantly in Study Group (p\\<0.0001), while decreased in Control Group (p\\<0.02)                                                                          THIRD TRIMESTER \\[MORE THAN 28\\]                            NR                             NR             NR                                      *MH status plays a role in AVF regulation in women with either normal or decreased amniotic fluid*.\n  ***Doi et al 1998***            RCT \\[**84**\\]                  *IO (AFI\\<5 cm)*. **Group A**: (n = 21) intravenous hydration of isotonic fluid 2000 ml/2 h (Lactated Ringer solution) **Group B**: (n = 21) intravenous hydration of hypotonic fluid 2000 ml/2 h (diluted Ringer solution) **Group C**: (n = 21) oral hydration of 2000 ml/2 h. **Control Group**: (n = 21) no hydration                                                                       The mean AFI increased significantly Group B and C (p\\<0.001), but not in Group A.                                                                                                   AT TERM 39.5 VS 39.7 VS 37.3 VS 38.9                        NR                             NR             NR                                      *In women with IO*, *a significant increase in the AFI was achieved by both IV hypotonic fluid loading and oral hydration*, *but not by IV isotonic fluid loading*.\n  ***Deka et al 2000***           OBS-PCS \\[**50**\\]              *IO (AFI\\<8 cm)*. **Study Group**: (n = 25) 2000 ml oral hydration in 1 hour **Control Group**: (n = 25) no hydration                                                                                                                                                                                                                                                                           The mean AFI increased significantly in all patients after 3 hours (p\\<0.001)                                                                                                        EARLY PRETERM 28                                            NR                             NR             NR                                      *Simple oral MH may help to sustain a 'steady state' amniotic fluid volume*, *and in the prevention and management of IO during pregnancy and labor*.\n  ***Chandra et al 2000***        RTS \\[**41**\\]                  *IO (AFI\\<6 cm)*. **Group A** (n = 16) oral hydration with 10--12 glasses of water per day (61.9\u00b111.7 hours between pre and post-treatment AFI measurements). **Group B** (n = 25) intravenous hydration (Lactated Ringer solution, 45.1\u00b18.9 hours between pre and post-treatment AFI measurements). **Subgroup B1** (n = 15) \u22642000 ml hydration. **Subgroup B2** (n = 10) \u22652500 ml hydration   The mean AFI increased minimally after hydration (51.2% of patients) with no differences between the two Groups. -AFI increase was not related to entity of hydration                AT TERM 38.5\u00b10.39 VS 38.8\u00b10.5                               NR                             NR             NR                                      *IO may respond to oral and intravenous hydration*.\n  ***Fait et al 2003***           OBS-PCS \\[**60**\\]              *IO (AFI\\<6 cm)*. **Study Group**: (n = 30) at least 2000 ml oral daily hydration **Control Group**: women with physiological pregnancy (routine hydration).                                                                                                                                                                                                                                    The mean AFI increased significantly in the Study Group after 1 week (p\\<0.01).                                                                                                      EARLY PRETERM 29 \\[RANGE: 26--34\\] VS 28 \\[RANGE:26--35\\]   NR                             NR             NR                                      *Long term oral hydration increase AFI for at least a week*.\n  ***Lorzadeh et al 2008***       RCT \\[**80**\\]                  *IO*. *(AFI\\<5 cm)*. **Group A**: (n = 20) intravenous hydration of isotonic fluid 2000 ml/2 h **Group B**: (n = 20) intravenous hydration of hypotonic fluid 2000 ml/2 h **Group C**: (n = 20) oral hydration 2000 ml/2 h (water) **Control Group**: (n = 20) no hydration                                                                                                                     The mean AFI increased in Group B and C (p\\<0.0001) without significant change in Group A.B Delta AFI was greater in Group C, in comparison with A and B Groups (p\\<0.0001)          AT TERM 39.0 \u00b1 1.3 VS 38.9\u00b11.27                             NR                             NR             NR                                      *MH with oral water*, *hypotonic fluid and isotonic fluid increases AFI in I*.*O*. *MH with oral water was more effective than other groups*.\n  ***YanRosemberg et al 2008***   RCT \\[**44**\\]                  *IO (AFI\\<6 cm)*. **Group A** (n = 21): 2000 ml iv (1/2 normal saline solution) in 2 h **Group B** (n = 23): placebo (20 ml iv 1/2 normal saline solution in 2 h)                                                                                                                                                                                                                               The mean AFI increased significantly in both groups (p\\<0.05), but not if a comparison was made.                                                                                     A TERM 39.2\u00b11.2 VS 39.1\u00b11.3                                 A TERM 40.1+1 VS 40+1.1        10% VS 9%      NR                                      *Acute intravenous hydration with hypotonic solution did not increase AFI in IO*\n  ***Umber et al 2010***          RCT \\[**50**\\]                  *IO (AFI\\<5 cm)*. **Group A**: (n = 25) intravenous hydration of 2000 ml/2 h (5% W/D solution) **Group B**: (n = 25) oral hydration of 2000 ml/2 h                                                                                                                                                                                                                                              The mean AFI increased significantlyin both Group A and Group B (p\\<0.05), with no significant differences between the two groups.                                                   THIRD TRIMESTER \\[RANGE: 28--42\\]                           NR                             NR             NR                                      *Intravenous as well as oral MH increases AFI*, *but neither appears to be particularly advantageous over the other*.\n  ***Borges et al 2011***         RCT \\[**99**\\]                  *Normal AFI*. **Group A**: (n = 34) oral hydration of 1500 ml/2-4 h of isotonic solution **Group B**: (n = 30) oral hydration of 1500 ml/2-4 h of water **Control Group**: (n = 35) oral hydration of 200 ml/2-4 h of water                                                                                                                                                                     The mean AFI increased in Group A and B (p\\<0.001) and was reduced in controls, but with no significant difference among Groups.                                                     LATE PRETERM 35\u00b11.53 VS 35.4\u00b11.6 VS 34.4\u00b12.2                NR                             NR             NR                                      *In women with normal AFI*, *a significant increase in AFI was achieved by oral hydration with both isotonic solution and water*.\n  ***Patrelli et al 2012***       RCT \\[**137**\\]                 *IO (AFI\\<5 cm)*. **Group A**: 66 patients with IO **Subgroup A1** (n = 33): 1500 ml iv (Ringer solution) + 1500 ml oral daily for 6 days. **Subgroup A**2 (n = 33): 1500 ml iv (Ringer solution) + 2500 ml oral daily for 6 days. **Group B**: 71 women with physiological pregnancy (routine hydration).                                                                                      The mean AFI increased in group A after therapy (P\\<0.001) without differences between subgroups. The mean AFI at birth was greater in subgroup A2 in comparison to A1 (P\\<0.001).   EARLY PRETERM 31.5\u00b11.2 VS 31.4\u00b11.3                          A TERM 39.5\u00b11.1 VS 39.4\u00b1 1.3   30% VS 18%     0% VS 0%                                *In pregnancies complicated by IO and treated with intravenous hydration therapy for 6 days the quantity of AF is significantly improved compared to pregnancies not complicated at the same gestational age*.\n  ***Ghafarnejad et al 2012***    RCT \\[**37**\\]                  *IO (AFI\\<6 cm)*. **Group A** (n = 22): 2000 ml oral in 2 h + routine hydration for 24 h. **Group B** (n = 22): routine hydration for 24 h                                                                                                                                                                                                                                                      The mean AFI increased in group A after therapy (P\\<0.001).                                                                                                                          LATE PRETERM 35.1\u00b11.4 VS 36\u00b12                               NR                             10% VS 9%      NR                                      *Acute oral hydration is a noninvasive*, *easily accessible and cheap way of increasing AFI*, *which should be encouraged*.\n  ***Shahnazi et al 2012***       RCT \\[**20**\\]                  *IO (AFI\\<5 cm)* **Study Group**: (n = 10) intravenous hydration of 1000 ml/30 min (isotonic saline solution) **Control Group**: (n = 10) no hydration                                                                                                                                                                                                                                          The mean AFI increased significantly in Study Group in comparison to controls (p = 0.03).                                                                                            AT TERM 38.6\u00b11.28 VS 39.37\u00b10.78                             NR                             30% VS 45.5%   10% VS 27.3%                            *MH is recommended as a low-cost method with no complications for the fetus and the mother*.\n  Akter et al 2012                RCT \\[**64**\\]                  *IO*. *(AFI\\<5 cm)* **Study Group**: (n = no data) oral hydration (water) 2 l + routine hydration/day for 7 days **Control Group**: (= no data) routine hydration                                                                                                                                                                                                                               The mean AFI increased in Study Group in comparison to controls. (p\\<0.05)                                                                                                           PRETERM \\[RANGE: 32--35\\]                                   NR                             28% VS 78.2%   16.2% VS 71.8%                          *Oral MH therapy significantly increases the AFI*, *reduces the caesarean section rate and improves the fetal outcome*.\n  \u00dclker et al 2013                RCT \\[**79**\\]                  *Normal AFI*. **Study Group**: (n = 40) oral hydration of 500 ml before first AFI assessment and 1250 ml before last AFI measurement (left lateral decubitus position) **Control Group**: (n = 39) no hydration (left lateral decubitus position)                                                                                                                                               The mean AFI increased in both Groups (p\\<0.05), but did not show any significant difference between the two Groups.                                                                 LATE PRETERM 36.32 \u00b1 0.86 VS 36.77 \u00b1 1.46                   NR                             NR             NR                                      *Maternal rest in the left lateral decubitus position with hydration and maternal rest in the left lateral decubitus position alone cause similar increases in the estimated AFV*.\n  Burgos et al 2014               OBS-PCS \\[**200**\\]             *Normal AFI*. **Study Group: (n = 100) intravenous hydration** of 2000 ml/2h (hypotonic saline) before the version attempt **Control Group**: (n = 100) no hydration before the version attempt                                                                                                                                                                                                 The mean AFI increased in the Study Group in comparison to controls (p\\<0.01).                                                                                                       AT TERM \\[RANGE: 37--41\\]                                   NR                             37% VS 37%     0% VS 0.1%                              *The intravenous option is a safe and effective way to increase the amount of amniotic fluid before external cephalic version*, *with no associated risks for the pregnancy*.\n\n**AFI**: amniotic fluid index, **AVF**: amniotic fluid volume, **IO:** isolated oligohydramnios, **MH**: maternal hydration, **GW**: gestational week, **CS**: cesarean section **RCT**: randomized controlled trial; **OBS:** observational study; **PCS**: perspective controlled study, **NR**: not reported\n\nAs above mentioned, the reference standard for single papers and quality of the included studies is illustrated in supplementary [S1 Fig](#pone.0144334.s001){ref-type=\"supplementary-material\"}.\n\n\u0394 variation in AFI index: different strategies of maternal hydration versus no treatment (IO only) {#sec015}\n--------------------------------------------------------------------------------------------------\n\nSix studies reported data regarding oral hypotonic fluid administration (276 women), \\[[@pone.0144334.ref015],[@pone.0144334.ref016],[@pone.0144334.ref018],[@pone.0144334.ref020],[@pone.0144334.ref025],[@pone.0144334.ref027]\\], three studies \\[[@pone.0144334.ref018],[@pone.0144334.ref020],[@pone.0144334.ref021]\\] reported data regarding intravenous isotonic fluid administration \u22652000ml (126 women) and one study regarding fluid intake \\<2000ml (20 women)\\[[@pone.0144334.ref026]\\], two studies regarding intravenous hypotonic fluid administration (82 women)\\[[@pone.0144334.ref018],[@pone.0144334.ref020]\\]. The analysis of data regarding \u0394 variation in AFI index in a cohort of women with diagnosis of IO showed significant differences between the intervention group compared to the no treatment group considering both fixed and random effects.\\[p\\<0.0001\\]\n\nIn detail, better results were collected when the intervention group was treated by oral hypotonic fluids (\u22652000ml) \\[p\\<0.0001\\] while inferior AFI improvements were noted after intravenous isotonic fluid administration (\u22652000ml)\\[p:n.s\\]. Contrasting results (despite collected only from 2 studies) were reported after intravenous hypotonic fluid administration (\u22652000ml). \\[p\\<0.01 in fixed model; p:n.s. in random model\\] ([Fig 1](#pone.0144334.g001){ref-type=\"fig\"}).\n\n![\u0394 variation in AFI index: different strategies of maternal hydration versus no treatment \\[isolated oligohydramnios only\\].](pone.0144334.g001){#pone.0144334.g001}\n\n\u0394 variation in AFI index: different types of fluid intake (isotonic versus hypotonic), hydration strategies (intravenous versus oral) and length of fluid administration (single day versus multiple days) (IO only) {#sec016}\n--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nTwo studies compared effects of intravenous versus oral hypotonic fluid administration (2000ml) in one day (82 women);\\[[@pone.0144334.ref018],[@pone.0144334.ref020]\\] two studies compared effects of isotonic intravenous versus hypotonic oral fluid administration (2000ml) in one day (90 women); \\[[@pone.0144334.ref020],[@pone.0144334.ref022]\\] one study compared isotonic versus hypotonic intravenous fluid administration (2000ml) in one day (42 women); \\[[@pone.0144334.ref018]\\] and one study compared variation in AFI index after 5 days of isotonic intravenous (\\<2000ml versus \\>2500ml) plus 1500ml hypotonic oral fluid administration (57 women)\\[[@pone.0144334.ref014]\\].\n\nThe analysis of data showed no significant improvements after the increase in daily dose of isotonic intravenous hydration and no significant differences between treatments completed within one day compared to those which lasted longer than 24 hours.\\[p:n.s.\\] The best results were achieved by intravenous hypotonic hydration of 2000ml administered within a single day, despite the fact that data was collected only from a single study \\[[@pone.0144334.ref018]\\]. Discordant results were reported regarding the administration of 2000ml of isotonic intravenous fluid versus 2000ml of hypotonic oral administration with significant improvements reported by Lorzadeh et al. \\[p\\<0.0001\\] \\[[@pone.0144334.ref020]\\] while no significant improvements were reported by Umber et al\\[[@pone.0144334.ref022]\\]. ([Fig 2](#pone.0144334.g002){ref-type=\"fig\"}).\n\n![\u0394 variation in AFI index: different types of fluid intake (isotonic versus hypotonic), hydration strategies (intravenous versus oral) and length of fluid administration (single day versus multiple days) \\[isolated oligohydramnios only\\].](pone.0144334.g002){#pone.0144334.g002}\n\n\u0394 variation in AFI index: different strategies of maternal hydration versus no treatment (IO versus normo-hydramnios) {#sec017}\n---------------------------------------------------------------------------------------------------------------------\n\nTwo studies compared the effects of different strategies of maternal hydration for amniotic fluid improvement in IO versus untreated matched pregnancies with normo-hydramnios (95 cases versus 101 controls)\\[[@pone.0144334.ref019],[@pone.0144334.ref024]\\].\n\nPatrelli et al. compared a 6 day treatment protocol consisting of isotonic intravenous fluid (1500ml) plus a hypotonic oral fluid intake (1500ml versus 2500ml) to a cohort of untreated controls. Both treatments resulted significantly effective in improving the AFI index \\[p\\<0.0001\\] with no significant differences observed between the two hydration schemes. \\[[@pone.0144334.ref024]\\] Interestingly, similar effects were collected by Fait et al. in a cohort of cases treated by a 2000ml intake hypotonic fluid administered orally for 14 days. \\[[@pone.0144334.ref019]\\]. The same hydration protocol appeared to be less effective when administered for a period of 7 days ([Fig 3](#pone.0144334.g003){ref-type=\"fig\"}).\n\n![\u0394 variation in AFI index: different strategies of maternal hydration versus no treatment \\[isolated oligohydramnios versusnormo-hydramnios\\].](pone.0144334.g003){#pone.0144334.g003}\n\n\u0394 variation in AFI index: effects of different strategies of maternal hydration (normo-hydramnios only) {#sec018}\n-------------------------------------------------------------------------------------------------------\n\nFour studies compared effects of different strategies of MH for amniotic fluid improvement in pregnant women with normo-hydramnios (453 women), all focused on short-term therapy (within one day).\\[[@pone.0144334.ref017],[@pone.0144334.ref023],[@pone.0144334.ref028],[@pone.0144334.ref029]\\]. In detail two studies compared oral hypotonic fluid administration (\\<2000ml) versus no treatments reporting no differences in term of AFI index variation \\[p:n.s.\\].\\[[@pone.0144334.ref028],[@pone.0144334.ref029]\\]A small, though not significant improvement was reported by Borges et al who administered 2000ml of isotonic solution orally \\[[@pone.0144334.ref023]\\] while a significant improvement was collected by Burgos et al and Kilpatrik et al \\[[@pone.0144334.ref017],[@pone.0144334.ref029]\\]who administered 2000ml of hypotonic solution intravenously and orally, respectively\\[p\\< 0.0001 and p\\<0.01, respectively\\] ([Fig 4](#pone.0144334.g004){ref-type=\"fig\"}).\n\n![\u0394 variation in AFI index: effects of different strategies of maternal hydration \\[normo-hydramnios only\\].](pone.0144334.g004){#pone.0144334.g004}\n\nRegarding gestational age at diagnosis, as shown in [table 1](#pone.0144334.t001){ref-type=\"table\"}, the majority of manuscripts reported heterogeneous or incomplete data (absence of absolute value, standard deviation or confidence interval). Additionally, only two studies, comparing different outcomes, reported data regarding gestational age at delivery \\[[@pone.0144334.ref021],[@pone.0144334.ref024]\\]. All remaining studies lacked data regarding the time interval between gestational age at intervention and delivery, therefore making statistical correlation studies impossible. Likewise, only 6 manuscripts reported data regarding the mode of delivery and the cesarean section rate \\[[@pone.0144334.ref021],[@pone.0144334.ref024]--[@pone.0144334.ref027],[@pone.0144334.ref029]\\] but the heterogeneity between the studies made a statistical evaluation of this outcome impossible. The same conditions applied for the evaluation of neonatal outcomes.\n\nDiscussion {#sec019}\n==========\n\nThough the ultimate goal of our meta-analysis was to evaluate the actual e effects of MH in both pregnancies with IO and with normal AFV, the majority of eligible studies (10 of 16 trials) merely reported information regarding the variation in AFV following intervention.\\[[@pone.0144334.ref014]--[@pone.0144334.ref020],[@pone.0144334.ref022],[@pone.0144334.ref023],[@pone.0144334.ref028]\\] Unfortunately, studies reporting data regarding perinatal outcomes were affected by relevant heterogeneity in outcome measures, data collection, eligibility criteria, and hydration strategy which made it impossible to make comparisons without incurring in bias in statistical evaluation.\n\nThe descriptive data reported in [Table 1](#pone.0144334.t001){ref-type=\"table\"} clearly demonstrated that in all cases of MH neither complications nor adverse events were reported. Interestingly, when hydration was associated with improvements in AFI, it also associated with a significant reduction in cesarean section rates and with a percentage of term deliveries comparable to that of normal pregnancies.\n\nThough the last Cochrane \\[[@pone.0144334.ref032]\\], published in 2002, concluded that \"since the studies reviewed have not assessed clinically relevant outcomes or possible complications, there is no evidence to support the use of MH in routine practice, except in the framework of further clinical trials designed to address these issues\", the encouraging data from subsequent trials (6 studies evaluating the impact on type of delivery and 4 on neonatal outcomes (see [Table 1](#pone.0144334.t001){ref-type=\"table\"}) allow us to postulate that the possibility of increasing AFV with a simple, non-expensive and less-invasive/non-invasive method such as MH should be considered a useful tool in a modern conception of obstetrical care.\n\nWe are conscious that stronger evidences are mandatory in order to introduce guidelines regarding MH intervention even during the third trimester of pregnancy (particularly when an isolated finding) and at term with the scope of improving the outcome of delivery and peri-partum care (favoring spontaneous cephalic version or increasing the success of external rotation maneuvers).\n\nDespite the limitations of the available data (impossibility to evaluate pregnancy, delivery and neonatal outcome after interventions, large confidence interval for gestational age at diagnosis, different cut-offs for diagnosis of oligohydramnios, different interval time of outcome measure-delta AFI) do not allow us to define the real clinical utility of MH, our results may prove useful in helping clinicians and researchers better design future studies aimed at solving e the dilemma regarding the potential utility of MH.\n\nEven if all above mentioned limitations suggest caution in the interpretation of these results, the statistical analysis of data clearly demonstrated the efficacy of MH in pregnancies affected by IO, since all proposed strategies significantly increased the AFV. On the contrary, when considering the efficacy of hydration in term pregnancy with normal AFI, the strength of the evidence seems to fall and dependence on the therapeutic strategy increased. In this last cohort, it seems that both routes of administration (intravenous versus oral) are equivalent while the administration of hypotonic fluids seems more effective than isotonic fluids in increasing AFV. However these results may be related to the small number of trials conducted and to the heterogeneity of both population and protocols.\n\nWhen considering IO, evidences suggest that oral hydration is more effective than intravenous hydration in increasing amniotic fluid volume. The indirect explanation of this fact may be identified in the duration of treatment. In all trials in which MH lasted longer than one day, the improvements in AFV appeared less dependent on the total volume of fluid administered. This fact suggests that most likely duration is more important than dose. Perhaps intravenous administration of fluids causes a more transitory increase in maternal volume compared to oral hydration and this may explain the lower efficacy of this strategy. On the other hand, while the dose of fluid intake does not seem to significantly affect the overall AFV increase, the type of fluid administered may significantly influence the final outcome. In general, all clinical trials suggested that in IO, regardless of the route of administration (oral versus intravenous versus combined), hypotonic solutions seems to be more effective than isotonic fluids. This fact is probably due to the physiologic homeostasis existing between the maternal and fetal compartment in maintaining a correct fetal AFV. Indeed Flack et al in 1995 demonstrated, by an elegant case-control study, that maternal hydration affects amniotic fluid precisely via increased transplacental passage, rather than affecting fetal production.\\[[@pone.0144334.ref040]\\]\n\nBecause isotonic solutions have the same concentration of solutes as plasma, infused isotonic solutions do not flow into cells. Rather, they remain within the extracellular fluid compartment and are distributed between the intravascular and interstitial spaces, thus increasing intravascular volume. On the contrary, infusing a hypotonic solution into the vascular system creates a concentration gradient between the fluid compartments. The infusion of hypotonic crystalloid solutions lowers the serum osmolality within the vascular space, causing a fluid shift from the intravascular space towards the intracellular and interstitial spaces. These solutions hydrate cells, although their use may deplete fluid within the circulatory system.\n\nThis physiological mechanism may perhaps explain the importance of duration of treatment rather than total dose administered since the passage of hypotonic fluid into the fetal compartment may be facilitated by time, while an increase in dose may cause an increase in maternal kidney function thus decreasing the dose dependent effect and potentially depleting maternal volume. The lack of data regarding follow-up after intervention in pregnancies affected by IO does not allow us to report regarding the time interval in which the increase in amniotic fluid remains clinically useful. Certainly further trials considering this endpoint may be useful in understanding whether the improvements are time limited and possibly define the average interval time in which further intervention is needed to maintain a safe AFV until delivery. The acquisition of this information may improve perinatal care and \"demedicalize\" delivery thus reducing the rate of unnecessary/avoidable cesarean section and preterm delivery with their associated complications \\[[@pone.0144334.ref041]--[@pone.0144334.ref048]\\]. Interpretation of reported data deserves caution since due to the amount of trials included in data analysis it was not possible to extrapolate information without incurring in some bias. In fact, it was not possible to exclude or quantify potential confounders such as maternal rest during hydration, maternal position during AFI assessment, absolute value of AFI at diagnosis (for example: an increase in AFI by 2 cm may be enough to move a patient from the IO to the low-normal group if the original AFI was above 3, whereas it would not significantly affect the clinical impression for AFI \\<3 cm).\n\nConclusion {#sec020}\n==========\n\nThough further studies are mandatory to reach sound conclusions, available data suggests that MH may be a safe, well-tolerated and useful strategy to improve AFV especially in cases of IO. Oral hydration should be preferred since it seems te more effective and potentially more feasible, avoiding invasiveness and necessity of hospitalization and potentially increasing patients' compliance to treatment. Hypotonic solutions should be preferred to isotonic solutions and administered at low dose (about of 1500 ml per day) for long periods (ideally for 2 weeks). In view of many obstetric situations in which a reduced AFV may pose threats, particularly to the fetus, the possibility of increasing AVF with a simple and inexpensive method like MH may certainly have useful clinical applications in obstetric care.\n\nSupporting Information {#sec021}\n======================\n\n###### Graphical display of study characteristics according to QUADAS II recommendations to report the risk of bias for patient selection and the concerns for applicability of data collected in manuscripts eligible for the meta-analysis.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### PRISMA 2009 Flow Diagram.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### PRISMA 2009 Checklist.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\nThe Authors acknowledge all the researchers who helped in collecting and analysing data.\n\nAF\n\n:   amniotic fluid\n\nAFI\n\n:   amniotic fluid index\n\nAFV\n\n:   amniotic fluid volume\n\nIO\n\n:   isolated oligohydramnios\n\nMH\n\n:   maternal hydration\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: MQ TF TSP. Performed the experiments: SG AV MN. Analyzed the data: SG AV MN. Wrote the paper: SG AV MN ADA DD CJA.\n"}
+{"text": "The competently performed and thoughtfully reported postmortem examination remains a vital component of today's autopsy practice. The integral role which autopsy plays in documenting diseases and injuries that cause death is well recognized. Of equal importance, however, is that in revealing anatomic pathology (and in some cases microbiologic and biochemical alterations, as well), an autopsy enables morbid anatomy to be correlated with clinical signs and symptoms, thereby enhancing an understanding of the decedent's ailments. The artful practice of clinical-pathologic correlation, advanced by Giovanni Battista Morgagni in 1761,[@B001] takes time to master, but becoming skillful at assisting clinicians in understanding a decedent's medical history or in explaining the cause of a decedent's symptoms to family members remains one of the most rewarding aspects of autopsy practice.\n\nIn order to provide meaningful clinical-pathologic correlations, autopsy pathologists should be aware of relevant issues that exist prior to, or arise during, an autopsy, and they should effectively address those issues in a satisfactory manner with the postmortem examination. For example, when pulmonary thromboembolism is identified at autopsy, deep veins of the legs and, if necessary, the arms should be dissected. (N.B. In consented autopsies additional permission(s) for the procedure(s) from next of kin may need to be obtained). Sections of the occluded or empty vessels should be submitted for histology, in addition to sections of pulmonary emboli within vessels.[@B002]\n\nSuch a thorough evaluation of venous thromboembolism assists in 1) determining the underlying cause of death (i.e., emboli by definition travel from somewhere, and most but not all pulmonary thromboemboli originate in the legs); 2) identifying a possible etiology of the thrombosis (e.g., potential phlebitis, extrinsic venous compression by tumor); and 3) approximately aging the thrombus and embolus (i.e., \"acute\", \"subacute\", or \"chronic\" clots). Complete evaluation and documentation of all relevant issues in such a manner optimizes clinical-pathologic correlation, allows objective medicolegal assessment of relevant issues, and reaffirms the value of autopsy among involved stakeholders.\n\nObjectively documenting evidence that can elucidate injury, disease, and/or death, such as in the aforementioned case of venous thromboembolism, is a fundamental purpose of the autopsy and underlies its persistent importance in the medicolegal realm. Indeed, over the past half century hospital autopsy rates have declined around the world,[@B003] while forensic autopsy rates have remained relatively unchanged (at least in the USA).[@B004] The distinction between \"hospital\" autopsies, in which natural disease processes are evaluated, and \"forensic\" autopsies, in which injuries as well as natural and non-natural pathologies are investigated, understandably exists to support jurisprudence in various countries. Although forensic pathologists are expected to become adept in performing, reporting, and testifying about autopsies in civil and criminal court proceedings, all pathologists conducting autopsies should recognize that any death may have medicolegal implications and the findings from any autopsy can be drawn into medicolegal proceedings.\n\nIn fact, among the various proposed reasons for declining hospital autopsy rates is the misperception among some physicians that autopsies increase the likelihood of malpractice litigation and/or render them culpable in such litigation.[@B005] However, literature from the USA,[@B006] Germany,[@B007] and Italy[@B008] exploring the role of postmortem examinations in cases of medical malpractice revealed that a majority of physicians were exonerated of the charges brought against them in cases where an autopsy had been performed. Moreover, the USA study showed that defendant physicians were acquitted of malpractice in the majority of cases, not only when autopsy findings favored the physician but also when autopsy findings favored the plaintiff initiating the lawsuit. In essence, unfavorable court rulings against defendant physicians involved standard of care issues rather than accuracy of clinical diagnoses in the cases examined by those authors.\n\nA poorly performed and/or reported autopsy can be detrimental to physicians, can negatively influence perceptions of the autopsy among healthcare professionals and the public, and does not allow the autopsy's potential value to be realized. The same review of USA medical malpractice cases discussed above also demonstrated that suboptimal autopsy performance and reporting adversely affected the appeals process in nearly 20% of the examined cases.[@B006] While the legal fate of a particular autopsy may not be known prior to its being completed, pathologists can always control the quality of their autopsy performance and reporting.[@B009] With the privilege of conducting autopsies and advancing clinical-pathologic correlation comes the responsibility to be consistent, complete, and competent in all facets of autopsy practice.\n\n**How to cite:** Williamson AK. The conscientious autopsy. Autops Case Rep \\[Internet\\]. 2019;9(2):e2019098. <https://doi.org/10.4322/acr.2019.098>\n\nFinancial support: None\n\n[^1]: Conflict of interest: None\n"}
+{"text": "Introduction {#s1}\n============\n\nIn the past decades, an increasing number of genomes have been completely sequenced (Adams, [@CIT0001]; Hillier *et al*., [@CIT0064]; Howe *et al*., [@CIT0067]; Venter *et al*., [@CIT0139]; Wood *et al*., [@CIT0148]). With the increasing knowledge of the sequence composition of genomes, the next challenge has been to comprehensively analyze the function of encoded genes. Of particular interest have been genes that are essential for survival at the cellular or organismic level. Identifying the minimal set of genes necessary for sustaining life will allow better understanding of life itself and provide insight into the origin of diseases.\n\nThe search for essential genes has been extensively conducted in prokaryotic organisms, with the aim to identify novel targets for antibiotic therapy (Clatworthy *et al*., [@CIT0023]) and critical building blocks for synthetic biology (Khalil & Collins, [@CIT0078]). Due to the small size of prokaryotic genomes and their easy accessibility to genetic manipulation, essential genes have been identified for a broad panel of prokaryotic organisms (de Berardinis *et al*., [@CIT0028]; Gerdes *et al*., [@CIT0047]; Glass *et al*., [@CIT0051]; Kobayashi *et al*., [@CIT0081]). In eukaryotic organisms, multiple loss-of-function technologies have been developed to investigate gene functions, including chemical mutagenesis (Hrab\u00e9 de Angelis *et al*., [@CIT0068]), insertional mutagenesis (Bellen *et al*., [@CIT0010]), RNAi technologies (Dietzl *et al*., [@CIT0033]; Kamath *et al*., [@CIT0075]) and CRISPR/Cas9 genome editing (Shalem *et al*., [@CIT0122]; Wang *et al*., [@CIT0141]). The effectiveness and also the limits of those screening technologies have determined the scope by which essential genes have been recovered.\n\nIn this review, we describe the development of screening technologies and their impact on discovery of essential genes for common eukaryotic model organisms. We illustrate how knowledge of gene essentiality contributes to understanding of human diseases and can be employed for anticancer therapy.\n\nWhat is an essential gene? {#s2}\n==========================\n\nEssential genes are defined as genes that are required for sustaining life (Juhas *et al*., [@CIT0072]). The concept of gene essentiality and its limits was first discussed in 1963 (Gluecksohn-Waelsch, [@CIT0052]). Presently, the general understanding of an essential gene is that it is required for survival and proliferation of single cell organisms. In multicellular organisms, loss of essential genes results in lethality during development or inability for reproduction. The estimated proportion of essential genes varies considerably between different species, and also among different publications ([Table 1](#t0001){ref-type=\"table\"}). Reasons for this discrepancy can be diverse and include differences in methods used to achieve loss-of-function, inability to perform genome-wide knockouts in many organisms and incomplete recovery of all phenotypes associated with gene essentiality. While there is a core set of essential genes that shows a stringently lethal phenotype upon loss, there is a larger group of genes on which survival depends on specific environmental conditions, in particular developmental stages or tissues. The impact of the environment on gene essentiality was extensively described for *Saccharomyces cerevisiae*, showing that under conditions other than rich in nutrients, the percentage and composition of essential genes varies (Giaever *et al*., [@CIT0048]; Hillenmeyer *et al*., [@CIT0063]). For instance, while 4769 homozygous deletion strains were considered non-essential in rich medium, only 205 strains (3% of the genome) were non-essential when growth was tested under multiple environmental conditions (Hillenmeyer *et al*., [@CIT0063]). Furthermore, defects in genes related to the immune system can also lack any visible phenotypes under laboratory conditions, but quickly become essential upon challenge by infectious agents (Galiana-Arnoux *et al*., [@CIT0043]; Gazit *et al*., [@CIT0045]). Table 1.Estimated proportion of essential genes in model organism.*Saccharomyces cerevisiae*\u00a0\u200317--18.1%(Giaever *et al*., [@CIT0048]; Winzeler *et al*., [@CIT0147])*Caenorhabditis elegans*\u00a0\u200313.9%(Johnsen & Baillie, [@CIT0071])\u20031%(Clark *et al*., [@CIT0022])\u20038.5%(Kamath *et al*., [@CIT0075])*Drosophila melanogaster*\u00a0\u20038-16.3%(Bellen *et al*., [@CIT0010])\u200330%(Dietzl *et al*., [@CIT0033])*Danio rerio*\u00a0\u20035.4%(Amsterdam *et al*., [@CIT0002])\u20039.3%(Haffter *et al*., [@CIT0057])*Mus musculus*\u00a0\u200313.3%(Bradley *et al*., [@CIT0016])*Homo sapiens* (core essential genes in a cancer cell line panel)\u00a0\u20031.4%(Hart *et al*., [@CIT0060])[^1]\n\nIn multicellular organisms, gene essentiality can be restricted to specific tissues and developmental stages. For example, mice deficient in SLC2744/FATP4, a protein responsible for the cellular import of free fatty acids, die shortly after birth due to skin abnormalities (Herrmann *et al*., [@CIT0062]). This lethal phenotype can be rescued by re-expression of the protein in the skin (Shim *et al*., [@CIT0124]). In contrast, knockout of SLC2744/FATP4 in either adipose tissue or the intestine did not show any striking phenotype (Lenz *et al*., [@CIT0088]; Shim *et al*., [@CIT0124]). In another example, post-developmental knockdown of essential genes in *Caenorhabditis elegans* revealed a fraction of genes than can actually prolong life-span when their function is lost at a later developmental stage (Curran & Ruvkun, [@CIT0026]).\n\nAnother concept that is closely related to gene essentiality is fitness. Fitness and fitness defects were originally used to describe changes of allele frequencies in population studies (Otto & Lenormand, [@CIT0109]) ([Figure 1](#F0001){ref-type=\"fig\"}). As opposed to gene essentiality, it is not measured on a single cell level, but is a population-level phenotype. Among others, it describes the exponential growth rate of a given population relative to its wild-type counterpart (Giaever *et al*., [@CIT0048]; Hillenmeyer *et al*., [@CIT0063]). Compared to essential genes, loss of genes that are associated with fitness defects can show only mild or no phenotypes within one generation. However, in a heterogeneous and dynamic population, selective pressure against fitness defects will result in the disappearance of individuals carrying the unfavorable trait within consecutive generations, as shown for *C. elegans* (Ramani *et al*., [@CIT0113]). Thus, genes that cause fitness defects can also be considered as essential in the context of population dynamics. Two studies in *C. elegans* and *S. cerevisiae* have shown that genes previously considered to be dispensable are actually associated with fitness defects (Breslow *et al*., [@CIT0017]; Kamath *et al*., [@CIT0075]). In summary, while the definition of gene essentiality is seemingly straightforward, unambiguously classifying a gene as essential is difficult and remains highly dependent on the context by which its function is measured. Figure 1.Model of essential genes and genes causing fitness defects. Loss-of-function of an essential gene X leads to cell death (A). In contrast, loss of a gene Y that is associated with a fitness defect leads to the gradual disappearance of the affected individuals from the population (B).\n\nDiscovery of gene essentiality in single cell organisms {#s2A}\n-------------------------------------------------------\n\n*Saccharomyces cerevisiae*, or budding yeast, was one of the first eukaryotic organisms in which essential genes were studied by a systematic approach and on a genome-wide scale (Winzeler *et al*., [@CIT0147]). *Saccharomyces cerevisiae* genes have many orthologs in common with other eukaryotes and the high rate of homologous recombination in *S. cerevisiae* enables its rapid genetic modification (Baudin *et al*., [@CIT0009]). Thus, it has been a favorable model for studying gene function on a larger scale. In a first set of experiments by Winzeler *et al*. ([@CIT0147]), deletion strains for 2026 ORFs were created, of which 17% were found to be essential for growth and survival in rich medium. Using a competitive growth assay with a pool of homozygous deletions strains of non-essential genes, the authors additionally showed that 40% of the strains have fitness defects. The second large-scale deletion screen in *S. cerevisiae* already included 5916 genes (96% of all annotated ORFs), of which 18.7% turned out to be essential for growth in rich medium (Giaever *et al*., [@CIT0048]). It was also observed that essential genes have more homologs in other organisms than their non-essential counterparts and that only 1% of essential genes had duplicates in the genome, as opposed to 8.5% of non-essential genes. While Winzeler *et al.* assessed phenotypes under two nutritional conditions (rich and low nutrients), Giaever *et al.* used five different conditions to demonstrate that gene essentiality and fitness both vary depending on the given environment. This observation was supported by another study in which a collection of \u223c11\u2009000 homo- or heterozygous deletion strains were tested against 726 different drugs or environmental stresses (Hillenmeyer *et al*., [@CIT0063]). The authors observed that 97% of all mutants exhibited growth defects under at least one condition, and therefore suggest that nearly all genes are required under a specific environmental condition.\n\nA main challenge for analyzing the function of essential genes in *S. cerevisiae* is the difficulty to generate hypomorphic mutants. Several methods have been developed to address this issue: essential genes can be shut off by inducible transcriptional repression (Mnaimneh *et al*., [@CIT0101]), by heterozygous deletion (Deutschbauer *et al*., [@CIT0030]) or by mRNA perturbation (DamP) (Breslow *et al*., [@CIT0017]).\n\nRecently, a genome scale collection of deletion mutations was generated for *S. pombe* (Kim *et al*., [@CIT0079]). *Saccharomyces cervisiae* and *S. pombe* are distantly related and differ in many cellular functions (Wood *et al*., [@CIT0148]), thus allowing for comparison and identification of genetic functions that are common to eukaryotes in general. In *S. pombe*, 4836 genes could be deleted, corresponding to 98.4% of all ORFs. Of those, 1260 or 26.1% were found to be essential. Similar to budding yeast, the proportion of single copy genes or singletons was higher among essential than non-essential genes. Gene sets of essential genes in both organisms were enriched for specific cellular processes (synthesis of DNA, RNA, lipids and proteins, transcriptional initiation, ribosome assembly).\n\nStudies of essential genes in multicellular organisms {#s3}\n=====================================================\n\nCompared to *S. cerevisiae*, the comprehensive study of essential genes in multicellular eukaryotic organism presents a greater technical challenge. Historically, most efforts to obtain genotype--phenotype interactions in multicellular organisms relied on forward genetic screening strategies using chemical or insertional mutagenesis. A major advance in functional genomics was introduced by the complete genome sequencing of model organisms and the development of RNAi technologies. The combination of both enabled the targeted knockdown of genes, allowing for reverse arrayed screens. Here, we describe how gene essentiality was explored in the three common model organisms *Caenorhabditis elegans, Drosophila melanogaster* and *Danio rerio*.\n\nIn *C. elegans*, the first screens that aimed at determining the number of essential genes relied on chemical mutagenesis with ethyl methanesulfonate (EMS) (Clark *et al*., [@CIT0022]; Johnsen & Baillie, [@CIT0071]). Based on the analysis of mutants in specific chromosomal regions including LGV(left) and unc-22 region, the authors estimated that the total number of essential genes in the *C. elegans* genome should range between 2850 and 3500 (Clark *et al*., [@CIT0022]; Johnsen & Baillie, [@CIT0071]). The first studies using RNAi in multicellular organisms were performed in *C. elegans*, by feeding animals with bacteria containing double stranded RNA or soaking animals in RNAi solution to achieve knockdowns (Fraser *et al*., [@CIT0040]; G\u00f6nczy *et al*., [@CIT0054]) ([Figure 2A](#F0002){ref-type=\"fig\"}). Both screens started with the knockdown of genes on single chromosomes and steadily increased genomic coverage to genome scale (Kamath *et al*., [@CIT0075]; Maeda *et al*., [@CIT0097]; S\u00f6nnichsen *et al*., [@CIT0129]). Roughly 800 genes were found to be critical for early embryonic development under laboratory conditions (S\u00f6nnichsen *et al*., [@CIT0129]), which is only 4% of the *C. elegans* genome. The total number of essential genes was estimated to be around 1750 (Kamath *et al*., [@CIT0075]). However, later studies indicate that the vast majority of non-essential genes show a measurable degree of fitness defect if measured over several generations, indicating that the number of essential genes might be underestimated (Ramani *et al*., [@CIT0113]). Figure 2.Screening strategies in different model organisms. (A) Schematic overview of RNAi screening approaches in *C. elegans* by ingestion of E.coli and in *Drosophila* cells by bathing (modified from Boutros & Ahringer ([@CIT0015])). Long double-stranded RNAs are introduced into the respective organisms and diced intracellularly into small-interfering RNAs (siRNAs). This results in many different siRNAs targeting a single transcript. (B) Outline of insertional mutagenesis screen in zebrafish (adapted by permission from Macmillan Publishers Ltd.: Nature Reviews Genetics (Patton & Zon, [@CIT0111]), (c) 2002). Embryos are injected with a retrovirus at a 1000--2000-cell stage. These embryos are raised (=\u2009founder generation P), mated and the mutations transmitted to the F~1~ generation. Individual fish from the F~1~ generation with multiple insertions are selected and further crossed with each other to generate the F~2~ generation. Siblings of each F~2~ family are crossed with each other to generate homozygous mutations. (C) Overview of screening approaches using ENU induced mutagenesis in mouse. For dominant mutations, ENU mutagenized males carrying mutations in their germ lines are crossed with wild-type females (G~0~). Dominant mutations will be detected in the G~1~ generation. For recessive mutations, males of the G~1~ generation carrying mutations are crossed with wild type females. Then females of the resulting G~2~ generation are backcrossed with G~1~ males, thereby generating the G~3~ generation that potentially carries individuals with homozygous mutations. (D) Arrayed versus pooled loss-of-function screens in cancer cell lines. siRNAs are used for arrayed screens in a multi-well plates. Each well harbors a distinct gene knockdown. Candidate genes are detected by measuring signal levels (e.g. luminescence) of individual wells. In contrast, pooled loss-of-function screens rely on viral infection of cells with shRNA vectors. Each vector contains a barcode allowing identification of the specific shRNA. Pools of cells with different gene knockdowns are generated and cultured for several doubling times. Depletion of cells with specific gene knockdowns is detected by sequencing of barcodes and measuring their relative abundance at different time points.\n\nThe use of transposable elements for insertional mutagenesis has been a major tool for studying genotype--phenotype interactions in *D. melanogaster*. Using P-element transposons, essential genes were identified by screening mutants on individual chromosomes (Bourbon *et al*., [@CIT0014]; Deak *et al*., [@CIT0029]; Oh *et al*., [@CIT0108]; Peter *et al*., [@CIT0112]). The number of identified essential genes ranged from 130 to 850. The effort to generate and classify P-element insertions in every gene is systematically conducted by the Berkeley Drosophila Gene Project. However, achieving genomic saturation with P-elements is difficult and so far, only 40% of all drosophila genes have been successfully disrupted (Bellen *et al*., [@CIT0010]). The percentage of lethal genes found by the Berkeley Drosophila Gene Project ranged between 8 and 16.3%, depending on the study included (Bellen *et al*., [@CIT0010]). The first genome-scale knockdown screen with dsRNA in *D. melanogaster* was performed in cultured blood cells and identified\u2009\\>400 genes that show a strong reduction of viability upon knockdown, many of which lacked mutant alleles (Boutros *et al*., [@CIT0154]) ([Figure 2A](#F0002){ref-type=\"fig\"}). In 2007, a genome-wide transgenic RNAi fly library was published and found that roughly 30% of the fly lines showed a lethal phenotype (Dietzl *et al*., [@CIT0033]).\n\nIn *D. rerio*, chemical mutagenesis with N-ethyl-N-nitrosourea (ENU) is an effective tool to introduce germline point mutations. Hence, several studies used ENU to generate large collections of mutants (Driever *et al*., [@CIT0036]; Haffter *et al*., [@CIT0057]). However, major general drawbacks of this method included laborious positional cloning to retrieve the underlying mutation and low genomic saturation that can be achieved (Justice, [@CIT0073]). In the most extensive ENU screen in *D. rerio*, mutants could be assigned to 375 genes, covering only a fraction of the zebrafish genome (Haffter *et al*., [@CIT0057]). Based on the results, the authors estimated that the percentage of lethal genes is roughly 2400, which is approximately 10% of the complete genome. However, only mutants with specific organ dysfunctions were selected for genotyping while mutants with multiple, non-viable malformations were not considered. Thus, the number of essential genes is most likely underestimated. Another forward genetic screening approach in *D. rerio* relied on the use of insertional mutagenesis (Patton & Zon, [@CIT0111]) ([Figure 2B](#F0002){ref-type=\"fig\"}). Two large-scale retroviral insertion screens were conducted in zebrafish, but only few essential genes could be retrieved (Gaiano *et al*., [@CIT0042]; Golling *et al*., [@CIT0053]). An insertional screen for embryonic and early larval development identified 315 essential genes, but only achieved a genomic saturation of 25% (Amsterdam *et al*., [@CIT0002]). Based on these numbers, the authors estimated that \u223c1,400 genes would be essential for embryonic development (Amsterdam *et al*., [@CIT0002]). Of the genes identified, a high proportion had homologs in yeast (72%) and human (99%), indicating that essential genes are phylogenetically conserved.\n\nEssential genes in mouse {#s4}\n========================\n\nThe mouse is the best studied mammalian model organism and identification of essential genes is of particular interest due to its close phylogenetic relationship to humans. Chemical mutagenesis with ENU has been the predominant screening tool to generate mouse mutants with novel phenotypes. ENU is a very powerful mutagen and predominantly creates single base mutations with the highest mutation rate in male spermatogonial stem cells (Balling, [@CIT0007]; Russell *et al*., [@CIT0116]). Chemical mutagenesis with ENU can cause both loss and gain of function mutations, and specific crossing strategies are required to obtain the desired mutations ([Figure 2C](#F0002){ref-type=\"fig\"}). One of the first efforts to identify lethal genes in mouse used ENU to generate mutants and back-crossings to identify affected genomic loci, but without recovering the precise point mutation (Rinchik & Carpenter, [@CIT0114]). A major drawback of ENU-based screens has been the laborious positional cloning necessary to identify underlying point mutations, which limited the rate by which novel mutants could be genotyped. In spite of this drawback, several large-scale ENU screens were initiated in the past to systematically generate, characterize and genotype novel mouse mutants (Hrab\u00e9 de Angelis *et al*., [@CIT0068]; Nolan *et al*., [@CIT0106]). While a large panel of phenotypic traits was documented for every mutant including fertility, other phenotypes of essential genes such as embryonic lethality were missed. Thus, only few essential genes were found in ENU screens. With the development of technologies for targeted gene disruption, many essential genes were identified by studying single gene functions in murine knockout models (Matsui *et al*., [@CIT0100]; Varfolomeev *et al*., [@CIT0138]).\n\nSince the establishment of homologous recombination in embryonic stem cells as a tool for targeted gene deletion (Thomas & Capecchi, [@CIT0132]), this technology has been further developed to enable generation of knockout cells on a larger scale (Skarnes *et al*., [@CIT0128]). Consequently, large efforts aiming at systematically generating knockout mouse models for every gene were started (Bradley *et al*., [@CIT0016]; White *et al*., [@CIT0144]). A subset of mice mutants with knockouts of 472 secreted proteins have been screened for specific phenotypes, and 8% of those showed pre-weaning lethality (Tang *et al*., [@CIT0131]). So far, roughly 3000 genes were identified to be essential upon knockout, which accounts for \u223c13% of the murine genome (Georgi *et al*., [@CIT0046]).\n\nBioinformatic resources {#s5}\n=======================\n\nWith the wealth of data available from both large loss-of-functions screens and genome sequencing projects, web-based depositories for genotype--phenotype interactions have been developed. The Online Essential Gene Database OGEE integrates results from large-scale screens from 16 prokaryotic and 8 eukaryotic organisms (Chen *et al*., [@CIT0020]). The database offers annotations to each essential protein-coding gene, including corresponding expression profile, duplication status or involvement in embryonic development. Another repository is the Database of Essential Genes (DEG), which since its first publication in 2004 has been updated several times (Luo *et al*., [@CIT0094]). The most recent release, DEG 11, includes essential genomic elements beyond protein-coding genes, such as promotors or non-coding RNAs. For common model organism such as *M. musculus, C. elegans* or *D. melanogaster*, community databases exist that systematically collect available phenotypes for every gene of the respective organisms (Blake *et al*., [@CIT0011]; Dos Santos *et al*., [@CIT0035]; Harris et al., [@CIT0059]) and are therefore a useful resource for detailed information on essential genes. The growing amount of data also enables comparative genomics approaches to explore common features of essential genes across different species ([Figure 3](#F0003){ref-type=\"fig\"}). For example, the propensity of genes to be lost in evolution was studied for a set of eukaryotic organisms (Krylov *et al*., [@CIT0086]). Essential genes were found to be associated with a low propensity to be lost during evolution, to accumulate fewer substitutions in the protein sequence and to be highly expressed. Furthermore, essential/lethal genes are found to be highly connected in protein networks (Jeong *et al*., [@CIT0070]) and have a high degree of phylogenetic retention (Gustafson *et al*., [@CIT0056]). In contrast, non-essential genes were more frequently targeted by many transcriptional factors, indicating that they are more dynamically regulated (Yu *et al*., [@CIT0150]). Based on functional data from yeast, essential genes were initially thought to be predominantly singletons. However, later studies could show that both singletons and duplicates are equally represented among essential genes in other organisms (Liang & Li, [@CIT0089]; Liao & Zhang, [@CIT0090]). Comparative genomics also allows dissecting characteristics of potentially essential genes in humans, which are not directly amenable to experimental studies. Analysis of the evolutionary and population genetics property of 2472 human orthologs of murine essential genes showed that they have less variants and are more frequently haploinsufficient (Georgi *et al*., [@CIT0046]). Whether human orthologs of murine essential genes are more frequently associated with human diseases is under debate, as some studies found associations (Dickerson *et al*., [@CIT0031]; Georgi *et al*., [@CIT0046]) while others did not (Park *et al*., [@CIT0110]). In summary, comparative genomic approaches help to understand the global structural features of essential genes and thus allow sequence-based prediction of essential genes in organisms without data from functional genomics experiments. Figure 3.General characteristics of essential genes. General characteristics of essential genes found across different species by comparative genomics are presented.\n\nGenetic interactions in model organisms {#s6}\n=======================================\n\nTo systematically study the functions and interaction partners of genes on a larger scale, synthetic genetic arrays have been developed in *S. cerevisiae* as a powerful tool (Boone *et al*., [@CIT0013]). In these experiments, every gene from a panel of query genes is deleted in combination with a gene from a second panel, resulting in large set of mutants with the loss of two genes. The growth behavior of all mutants is then individually measured and compared to each other. Three possible outcomes can result from a combinatorial loss-of-function: if the observed phenotype is the same as the single knockout, no genetic interaction is assumed. If the knockout of one gene can compensate for the loss of the other gene, then a positive or alleviating genetic interaction is present. If the knockout of one gene aggravates the phenotype caused by the loss of the second gene, a negative genetic interaction is found. If two genes share the same genetic interaction pattern across a large panel of genes, it can be assumed that they are functionally related.\n\nSynthetic genetic arrays were first performed with deletion strains of non-essential genes and could uncover several novel genetic interactions (Tong *et al*., [@CIT0133],[@CIT0134]). Based on these results, large-scale screens were conducted in *S. cerevisiae* mutants that harbor a conditional repressed essential gene and the loss of a query gene (Davierwala *et al*., [@CIT0027]). The effect of the genetic interaction between 575 essential genes and 30 query genes on growth behavior of mutants was analyzed. Similar to non-essential genes, essential genes also tend to share similar interaction partners if they are functionally related. However, the density of genetic interaction, i.e. the number of interaction partners was five times higher in essential genes, underlining that they are central hubs within the cellular network. Furthermore, the function of previously unknown essential genes could be assigned due to similarity of interaction partners, e.g. PGA1, which is required for specific functions of the endoplasmic reticulum.\n\nSynthetic genetic arrays were also performed in cultured *drosophila* cells using combinatorial knockdown with siRNA and image-based analysis of morphological features of cells (Fischer *et al*., [@CIT0038]; Horn *et al*., [@CIT0066]). In the publication by Horn *et al.*, 93 genes involved in MAPK, JNK and p38 signaling were knocked down in pairwise combinations, resulting in positive and negative genetic interactions. For example, single knockdown of Ras85D resulted in a reduced cell number, indicating that this gene is essential. However, the effect of Ras85D knockout could be compensated by the knockdown of a second gene, CG13197 (Horn *et al*., [@CIT0066]). The same approach was performed on a larger scale by Fischer *et al.*, by combined knockdown of a panel of 1367 genes involved in key cellular process (chromatin biology, cell cycle regulation, protein homeostasis) against a panel of 72 query genes. By grouping genes according to the pattern of their genetic interaction, genes could be assigned to known functional groups and directionality of genetic interaction could be inferred. Using this approach, novel links between the ERK signaling and chromatin remodeling could be discovered.\n\nEssential genes in cancer cells {#s7}\n===============================\n\nThe identification of essential genes in tumor cells is of outstanding interest in cancer research, as they present potential targets for novel therapeutic interventions. The first evidence that cancer cells may depend on specific mutated genes for proliferation and survival was shown in Kras mutant colorectal cancer cells (Shirasawa *et al*., [@CIT0125]). It was shown that the targeted deletion of mutant Kras resulted in a significant growth defect of cancer cells in nude mice. This observation was generalized under the concept of oncogene addiction (Weinstein & Joe, [@CIT0142]), which proposes that cancer can become dependent on specific mutated genes (oncogenes). These oncogenes then take over an essential role within a specific pathway that is not found in normal cells.\n\nThus, in pursuit of those conditionally essential genes, many loss-of-function screens have been performed. The two main screening strategies that are used to identify candidate genes in cancer cells are arrayed screens using siRNA and pooled screens using shRNA ([Figure 2D](#F0002){ref-type=\"fig\"}). A pilot study tested the effect of siRNA-mediated knockdown of 21 genes in transformed and non-transformed mammalian cells, measuring cell viability as outcome (Harborth *et al*., [@CIT0058]). While the chosen siRNA library was small, the study highlighted that screening with siRNAs was feasible in mammalian cell lines. Three years later, a first large-scale RNAi screen using 5305 siRNAs was performed to identify genes that regulate cell division in HeLa cells (Kittler *et al*., [@CIT0080]). Although not primarily focusing on essential genes, this study showed that knockdown of previously known essential genes such as ribosomal proteins or proteasome core units result in a lethal phenotype.\n\nOne of the first shRNA screens was performed in diffuse B-cell lymphoma using a retroviral library targeting 2500 genes, uncovering an essential role of NFkappaB pathway members for cell survival (Ngo *et al*., [@CIT0104]).Two years later, several screens that used large pools of shRNAs to identify essential genes in cancer cells were published. The screen by Silva *et al.* used shRNA pools of different scales to knockdown genes in five breast cancer cell lines (Silva *et al*., [@CIT0126]). Among the identified essential genes were several cell cycle regulators and components of the protein translation machinery. The sensitivity of cells towards knockdown of essential genes varied between cell lines and this observation could be confirmed using drugs with the same targets. Similar findings were obtained from another study that used 8204 shRNAs targeting 2924 genes in four cell lines (Schlabach *et al*., [@CIT0117]). The number of depleted genes varied significantly between cell lines, from as low as 2.5% of the gene panel to 23.8%. Another shRNA screen used the TRC library developed by the RNAi consortium (Moffat *et al*., [@CIT0102]), which contains 170\u2009000 shRNAs targeting 17\u2009200 human genes, to screen for essential genes in 12 cancer cell lines (Luo *et al*., [@CIT0093]). Two sets of essential genes could be discriminated: one set of global essential genes that was found in all cell lines and enriched for cellular processes including mRNA processing, translation and proteasomal degradation. In addition, another set of essential genes was identified that was specific to selected cell lines and often included oncogenes. One example of such a lineage-specific essential gene is IRF4, which was identified to be essential in multiple myeloma (Shaffer *et al*., [@CIT0120]). IRF4 is not genetically altered itself, indicating that essential genes do not necessarily correlate with mutation status in cancer. Another example is Brd4, which was found to be essential in a genetically defined model of acute myeloid leukemia and is a regulator of Myc (Zuber *et al*., [@CIT0152]).\n\nDue to the promising findings of the first screens, larger efforts were conducted to identify global and lineage-specific essential genes across larger sets of cell lines. The largest of such efforts to date, termed Project Achilles, assessed gene essentiality for 11\u2009000 genes across 216 human cancer cell lines by pooled lentiviral shRNA screens (Cowley *et al*., [@CIT0025]). In ovarian cancer, 54 lineage-specific essential genes were identified (Cheung *et al*., [@CIT0021]), among which PAX8 was further validated and found to be also amplified in ovarian cancer. Another large-scale screen used pooled shRNAs to target 16\u2009000 genes, but focused on a large panel of cell lines for a few selected tumor entities (Marcotte *et al*., [@CIT0099]). A total of 72 different breast, pancreatic and ovarian cancer cell lines were screened and a large set of essential genes was detected which overlaps with previous shRNA screens, indicating that the method is robust (Koh *et al*., [@CIT0082]). In total, a core set of 291 genes were discovered to be essential across many cell lines (Hart *et al*., [@CIT0060]).\n\nSynthetic lethality in cancer {#s8}\n=============================\n\nA concept of genetic interaction that has major implication for cancer therapy is synthetic lethality, with the potential to target the selected loss of tumor suppressor genes or addiction to oncogenes in cancer cells (Chan & Giaccia, [@CIT0019]; Kaelin, [@CIT0074]). Briefly, synthetic lethality occurs in cells that survive with an altered gene function in either gene A or gene B, but does not survive if the function of gene A and B are both altered ([Figure 4](#F0004){ref-type=\"fig\"}). Within the framework of gene essentiality, synthetic lethality can also be understood as an essential functional relationship between two genes. Figure 4.Concept of synthetic lethality. Loss-of-function of the essential gene x leads to immediate cell death (A), while loss-of-function of either gene A or gene B does not have a phenotypic effect. In contrast, combined loss-of-function of gene A and B results in a synthetic lethal interaction.\n\nSince most oncogenes or tumor suppressor genes are not directly amenable to pharmacological therapy, there is an urge to identify genes that become essential due to their functional interaction with oncogenes or tumor suppressors (Garber, [@CIT0044]; Kaelin, [@CIT0074]). In addition, genes that gain essentiality when specific cellular pathways are blocked by anticancer drugs are also of particular interest for combinatorial drug treatment. Knowing those genes would considerably enlarge the repertoire of cancer therapy and allow more selective killing of cancer cells.\n\nA first proof-of-principle experiment in a mammalian cell model used a hypoxanthine--guanine phosphoribosyl transferase (HPRT1) deficient cell line that expresses HPRT1 and a GFP reporter on an episomal plasmid (Simons *et al*., [@CIT0127]). HPRT1 is non-essential under normal conditions and the episomal plasmid is consequently lost. However, when the biosynthetic pathway leading to guanine monophosphate production is perturbed by specific inhibitors, HPRT1 becomes essential and only cells that were able to retain the expression plasmid survived.\n\nSubsequently, several studies exploited the concept of synthetic lethality to identify genes with chemosensitizing potential. For example, an arrayed siRNA screen was performed by Whitehurst *et al.* with the small lung cancer cell line NCI-H1155, measuring viability of gene knockdowns in the presence of sub-lethal concentrations of paclitaxel (Whitehurst *et al*., [@CIT0145]). The authors identified 87 genes that render the lung cancer cells sensitive to treatment with this microtubule inhibitor. In another arrayed siRNA screen, breast cancer cells were treated with a PARP-inhibitor and a number of kinases were identified to act synthetically lethal with the inhibitor (Turner *et al*., [@CIT0135]). Using the same approach, synthetic lethal interactions were found for a multiplicity of drugs, including inhibitors of PLK1 (Liu-Sullivan *et al*., [@CIT0091]; van der Meer *et al*., [@CIT0137]), DNA-PK (Dietlein *et al*., [@CIT0032]), ATR (Mohni *et al*., [@CIT0103]) or EGFR (Astsaturov *et al*., [@CIT0004]). A different approach making use of synthetic lethality aims at identifying genes that are essential in a specific genetic background, i.e. the presence of a gain or loss of function mutation. Targeting a panel of genes, either with RNAi or selective drugs, would reveal candidates that act synthetically lethal with the respective mutation. One of the first studies based on this concept screened 23\u2009550 compounds to identify drugs that selectively kill cells transformed by different combinations of oncogenes, but not their isogenic non-transformed counterparts (Dolma *et al*., [@CIT0034]). The authors found that specific combinations of oncogenes increased topoisomerase expression, rendering cells sensitive to topoisomerase inhibitors. A further early study aimed at identifying kinases that are required by clear renal cancers lacking the von Hippel-Lindau tumor suppressor (Bommi-Reddy *et al*., [@CIT0012]). A small shRNA screen was performed and revealed several kinases that act synthetically lethal, of which some could be confirmed using drugs.\n\nLater, several genome-scale RNAi screens have been performed using isogenic cell lines, which either have a RAS mutation (Schlabach *et al*., [@CIT0117]) or loss of TP53 (Krastev *et al*., [@CIT0085]) background. In another screen, Vizeacoumar *et al.* used genome-scale shRNA libraries to systematically identify negative genetic interactions across five isogenic cell lines with loss of function of major tumor suppressors (Vizeacoumar *et al*., [@CIT0140]). An alternative approach is to compare genetic interactions in multiple cell lines harboring the desired mutational background to a panel of cells without this background. Two hallmark papers used this approach to identify synthetic lethal interactions with mutated KRAS using a large set of cell lines (Barbie *et al*., [@CIT0008]; Scholl *et al*., [@CIT0118]), hereby identifying TBK1 and STK33 as negative interactors. TBK1 activates anti-apoptotic signals via NfKappaB (Barbie *et al*., [@CIT0008]), while STK33 suppresses mitochondrial apoptosis through S6K1 and BAD (Scholl *et al*., [@CIT0118]). However, different results were obtained with pharmacological inhibitors, indicating that the interaction might be independent of the kinase activity of STK33 (Babij *et al*., [@CIT0006]; Luo *et al*., [@CIT0095]; We\u00efwer *et al*., [@CIT0143]).\n\nWith the wealth of data from functional screens and mutation data from sequencing projects at hands, integrated approaches to identify synthetic lethal interactions were also possible. For example, RNAi profiles of cancer cells were compared to genome-wide copy number abberations, hereby identifying 56 genes for which reduction in growth only occurred if cells also harbored a copy number loss of the respective gene (Nijhawan *et al*., [@CIT0105]). Another study combined results from a Wnt pathway activity readout with data on lethal phenotypes identified by pooled shRNA screens for a set of 85 cell lines (Rosenbluh *et al*., [@CIT0115]). The authors found out that cancer with high active Wnt levels rely on YAP1 that forms a complex with TBX5 and mediates expression of anti-apoptotic genes BCL2L1 and BIRC5.\n\nAlthough there is a constantly growing amount of data from structural and functional studies, only few identified genes have been exploited as drug targets in a clinical setting. Initial successes by targeting lineage specific essential genes (BCR-Abl fusion protein in chronic myeloid leukemia by imatinib mesylate (Kantarjian *et al*., [@CIT0076])) have raised high hopes for a more efficient cancer therapy. However, while the repertoire for targeted therapy is constantly enlarging and improving overall survival rates of many cancers, the initial success of BCR-Abl imatinib remains an exception rather than the rule. Potential reasons for this development include inherent limitations of cell lines as a cancer model (Wilding & Bodmer, [@CIT0146]), tumor heterogeneity (Gillies *et al*., [@CIT0050]) and the involvement of the tumor microenvironment (Straussman *et al*., [@CIT0130]).\n\nClinical translation of synthetic lethality has been (in part) successful in two cases: PARP inhibitors in BRCA mutated breast cancer and the combination of retinoid acid and arsenic trioxide for treating promyelocytic leukemia. In 2005, two parallel publications described that in breast cancer cells with deficient BRCA1 and BRCA2, the Poly(ADP-ribose) polymerase PARP1 takes over an essential function in repairing DNA lesions (Bryant *et al*., [@CIT0018]; Farmer *et al*., [@CIT0037]). Pharmacological inhibition of PARP1 was highly effective in eradicating BRCA1 and BRCA2 deficient cancer cells. With a strong biological rationale behind, PARP inhibitors quickly went to phase I (Fong *et al*., [@CIT0039]) and phase II (Audeh *et al*., [@CIT0005]; Tutt *et al*., [@CIT0136]) trials for treatment of BRCA1/2 mutated breast and ovarian cancer, with promising results regarding response rate and clinical benefit. Several consecutive trials, including phase 3 trials, have been started, but the results so far are mixed and recent findings suggest that the significance of BRCA germline mutations in determining therapy response needs to be reassessed (Scott *et al*., [@CIT0119]).\n\nThe combination of all-trans retinoic acid (ATRA) and arsenic trioxide in treatment of promyelocytic leukemia (PML) is another example for clinical translation of synthetic lethality. The standard treatment of PML has been a combination of retinoic acid and anthracyclines, with overall high success rates. However, several studies show that ATRA and arsenic trioxide can strongly synergize to eradicate PML in *in vitro* and *in vivo* models (Lallemand-Breitenbach *et al*., [@CIT0087]; Shao *et al*., [@CIT0123]). Both compounds bind at different moieties of the PML-RARA fusion protein, thereby synergistically accelerating its degradation. A phase 3 clinical trial showed that combination of both substances is most likely superior to standard therapy and associated with less hematological toxicity, but higher hepatic toxicity (Lo-Coco *et al*., [@CIT0092]). In summary, while systematic screens in cancer cell lines have yielded a wealth of data on essential genes and synthetic lethality, translating these findings into novel clinical therapy still remains a major challenge.\n\nNovel methods for discovery of essential genes {#s9}\n==============================================\n\nRNAi has been the main workhorse for targeted identification of novel gene functions for almost a decade, with constant refinement of design and application. However, there are also limits and disadvantages inherent to RNAi technology. These include off-target effects (Ma *et al*., [@CIT0096]), toxicity and incompleteness of generating knockdown for selected genes (Boutros & Ahringer, [@CIT0015]). Knockout efficiency is also dependent on biological sources of variability, such as AGO2 expression levels (Hart *et al*., [@CIT0060]). Additionally, RNAi exclusively targets the mRNA of transcribed genomic regions, creating only loss-of-function. This however precludes analysis of non-coding genomic regions and gain-of-function phenotypes.\n\nRecently, the development of the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/Cas9 system of *S. pyogenes* into a genome editing tool (Cong *et al*., [@CIT0024]; Mali *et al*., [@CIT0098]) has opened new avenues for functional genomics. In brief, a synthetic small guide RNA targets the modified Cas9 endonuclease to a complementary sequence in the genome, where it introduces a double strand break. In most cases, these breaks are repaired by non-homologs end joining, which is error-prone and frequently results in loss-of-function deletion or insertions. CRISPR/Cas9 can be used in a wide range of species with high efficiency (Friedland *et al*., [@CIT0041]; Gratz *et al*., [@CIT0055]; Hwang *et al*., [@CIT0069]), making it an universal tool.\n\nThe simplicity of the CRISPR/Cas9 technology allows high-throughput screening in a similar scale as pooled shRNAs. CRISPR/Cas9 screens were performed in cancer cells and murine embryonic stem cells to identify resistance mechanisms towards toxins or drugs (Koike-Yusa *et al*., [@CIT0083]; Shalem *et al*., [@CIT0121]; Zhou *et al*., [@CIT0151]). In addition, screens have been conducted to identify essential genes in haploid and diploid leukemia cells (Wang *et al*., [@CIT0141]) and colorectal cancer cells (Hart *et al*., [@CIT0061]). Interestingly, identified hits between shRNA and CRISPR screens were partly non-overlapping, indicating that complete loss of function may result in different phenotypes (Hart *et al*., [@CIT0061]; Shalem *et al*., [@CIT0121]). In addition, the number of essential genes identified by a CRISPR/Cas9 screen was higher than by a shRNA screen in HCT116 (Hart *et al*., [@CIT0061],[@CIT0060]). It is thus anticipated that large-scale loss of function screens will be repeated using CRISPR/Cas9 technology. In addition to loss-of-function of protein coding genes, CRISPR/Cas9 also allows targeting and functional characterization of long non-coding RNAs and non-transcribed regions (Ho *et al*., [@CIT0065]; Kearns *et al*., [@CIT0077]; Yin *et al*., [@CIT0149]). Moreover, modifications of the CRISPR/Cas system for transcriptional activation of genes will in the future enable identification of gene essentiality and genetic interaction with gain-of-function CRISPR/Cas9 libraries (Gilbert *et al*., [@CIT0049]; Konermann *et al*., [@CIT0084]). CRISPR/Cas9 also enables reconstruction of point mutations frequently found in cancer (Antal *et al*., [@CIT0003]), thereby allowing screening for synthetic lethality in very specific genetic backgrounds. In face of all those possibilities opened up by novel screening methods, technical standardizations of screening procedures and bioinformatics analysis pipelines are essential to obtain comparable results across different screens.\n\nWe thank Marco Breinig and Christian Scheeder for careful reading of the manuscript and critical comments.\n\nDeclaration of interest {#s10}\n=======================\n\nWork in the lab of M.B. is in part supported by an ERC Advanced Grant. T.Z. is supported by a fellowship from the DKFZ International Postdoc program.\n\n[^1]: The percentage of essential genes is obtained by dividing the number of essential genes given in the indicated literature by the total number of protein coding genes in the respective genomes (retrieved from ENSEMBL database).\n"}
+{"text": "1.. Introduction {#S0001}\n================\n\nPorous ceramics are used in a variety of applications where high permeability is required, such as filtration of gases and liquids or catalytic supports.\\[[@CIT0001]\\] The main processing methods are extrusion, ceramic replication of an organic template, and direct foaming of a ceramic suspension. The final properties of porous ceramics are dependent on the processing technique and the resultant microstructure.\n\nIn general, ceramics processed by the replica method exhibit accurate porosity control and the pore size is relatively large (200--3000 $\\mathit{\\mu}$m), although the hollow struts characteristic of the replica method have a deleterious effect on strength.\\[[@CIT0002]\\] In contrast, materials obtained by direct foaming have higher mechanical properties compared to the replica method but larger pore size distribution and consequently a lower permeability.\\[[@CIT0003]\\] The microstructural effects on permeability of porous ceramics have been extensively studied in different cellular structures. Innocentini et al. \\[[@CIT0003]\\] compared the effect of pore volume (50--90%) and size (100--900 $\\mathit{\\mu}$m) in materials processed by replica and gel casting. Biasetto et al. \\[[@CIT0004]\\] conducted a similar study on porous ceramics obtained by organic burn out in a porosity range of 80--90% and pore size 10--150 $\\mathit{\\mu}$m. In both cases, permeability is increased by increasing porosity and pore size compromising the mechanical strength. One possible solution to increase both permeability and strength simultaneously is to engineer a unidirectional porosity (parallel to flow) to minimize the flow resistance without degrading the strength.\\[[@CIT0005]\\]\n\nThe idea of using ice crystals to template porosity in materials and in ceramics in particular have been known for a very long time \\[[@CIT0006]--[@CIT0008]\\], and the underlying principles (the segregation of matter by growing crystals) are more than a century old.\\[[@CIT0009],[@CIT0010]\\] Ice-templatinghas successfully produced anisotropic porous materials in a simpler manner coupled with a high flexibility to control the pore volume, morphology, and size.\\[[@CIT0011],[@CIT0012]\\] The process is based on the phase separation that takes place when a colloidal suspension is frozen. After freezing, the solvent is sublimated leaving pores whose morphology is a replica of the ice crystals. Finally, the green body is sintered to consolidate the structure.\n\nIn ice-templated materials, Fukasawa et al. \\[[@CIT0013]\\] studied pressure drop in Si${}_{3}$N${}_{4}$. These structures were specifically designed as filters with fibrous grains protruding into the macropores created by the ice-templating process. The samples ranged in porosity from 62.8% to 69.5% and pore size from 10 to 50 $\\mathit{\\mu}$m. They reported pressure drop data for four samples and found that increasing pore size reduced pressure drop. The data were not compared to known models of permeability, and the effects of the macroporosity was not separated from the effects of the fibrous grains protruding into the pores. Therefore, it is difficult to extend this data and use it to make predictions about permeability in samples with other types of pore morphologies.\n\nFukushima et al. \\[[@CIT0014]\\] studied SiC structures with 85% porosity and pore size between 34 and 147 $\\mathit{\\mu}$m. They found that increasing pore size increased permeability. They compared the permeability data to the capillary model.\\[[@CIT0015]\\] Assuming that tortuosity equaled one, they found that the capillary model matched the data for pore sizes from 34 to 70 $\\mathit{\\mu}$m but overestimated the permeability for samples with an average pore diameter of 147 $\\mathit{\\mu}$m.\n\nPekor et al. \\[[@CIT0016]\\] studied the porosity range 43--70% and pore size in the range 6--17 $\\mathit{\\mu}$m in ice-templated alumina. This investigation was focused on the impact of the polyvinyl alcohol (PVA) content on the permeability and found that increasing PVA content increases the non-Darcian component of the flow, most likely because increasing PVA content increases the surface roughness of the ceramic walls. The measured permeabilities were not compared to models.\n\nWhile there are several studies that discuss permeability in ice-templated ceramics, there is no single study that systematically explores the effects of pore volume, pore size, and pore morphology on permeability. Further, no article connects the mechanical properties of the samples with their permeability, and establishing this link is critical to the industrial application of ice-templated materials.\n\nIn this work, we aim to provide a deep understanding of permeability in ice-templated materials across different pore volumes, sizes, and morphologies. We systematically explore the permeability of ice-templated samples with different pore morphologies over the 54--72% porosity range with controlled pore sizes ranging from 2.9 to 19.1 $\\mathit{\\mu}$m. Importantly, we show that the Ergun and the capillary models accurately predict the permeability in materials with unidirectional porosity across the different pore morphologies, porosities, and pore sizes examined here. Finally, we link the permeabilities discussed in this work with mechanical properties obtained in our previous work \\[[@CIT0017]\\] in order establish links between strength and permeability in unidirectionally porous materials that may help researchers target specific structures to specific application where strength and permeability are critical.\n\n2.. Experimental procedure {#S0002}\n==========================\n\n2.1.. Sample preparation {#S2001}\n------------------------\n\nPorous samples with unidirectional, aligned porosity were prepared by ice-templating. Suspensions were prepared by mixing distilled water with 0.75 wt% of dispersant (Prox B03, Synthron, Levallois-Paris, France), 3 wt% of PVA (PVA2810, Wacker, Burghausen, Germany) as a binder, and 3 mol% yttria-stabilized zirconia (TZ-3YS, Tosoh, Tokyo, Japan) at different solids loading ranging from 50 wt% (14.7 vol.%) to 65 wt% (24.2 vol.%). In some suspensions, zirconium acetate (20 g l${}^{- 1}$) was added to the slurry to modify the pore morphology.\\[[@CIT0018]\\] All suspensions were ball-milled with alumina-milling media for a minimum of 18 h to achieve a good dispersion.\n\nAfterwards, 10 ml of slurry was poured into a PTFE mold (20 mm diameter and 25 mm height) placed on top of a copper plate. The top of the mold was exposed to air and kept at room temperature. Samples were unidirectionally frozen from the bottom to the top circulating silicone oil by the freezing plate. Temperature of the oil was regulated by a cryothermostat (Model CC 905, Hubert, Offenburg, Germany). The cooling rate was set at 2\u00a0${}^{\\circ}$C min${}^{- 1}$. A faster cooling rate was achieved by dipping a copper rod with the mold on top in a liquid nitrogen bath. An isolating panel was placed in between the mold and the top of the bath to avoid the direct contact of the vaporized liquid nitrogen with the mold. The cooling rate was monitored with a thermocouple and it was determined to be 25\u00a0${}^{\\circ}$C min${}^{- 1}$ on average. After solidification, samples were removed from their molds and sublimated for at least 48 h in a commercial freeze-dryer (Free Zone 2.5 Plus, Labconco, Kansas City, MO, USA) until the solvent was completely removed. The sintering cycle involved removal of organic components, which was achieved by heating the samples from room temperature to 500\u00a0${}^{\\circ}$C at a rate of 3\u00a0${}^{\\circ}$C min${}^{- 1}$ and holding them at 500\u00a0${}^{\\circ}$C for 5 h. Samples were sintered at 1400\u00a0${}^{\\circ}$C for 3 h using a heating and cooling rate of 5\u00a0${}^{\\circ}$C min${}^{- 1}$. Finally, we cut 2 mm off the bottom of each sample to remove the area with randomly oriented porosity \\[[@CIT0019]\\] and further minimize the pore size gradient between the bottom and the top, thereby testing the part with continuous porosity and with an almost constant cross section. The top of the specimens was also removed to avoid possible artifacts that may appear during the demolding.\n\nTo evaluate the effect of pore directionality, samples with randomly oriented porosity were prepared using pore formers. The same zirconia powder was mixed with commercially available pore former polypropylene with a mean particle size ($d_{50}$) of 45--55 $\\mathit{\\mu}$m and a maximum particle size ($d_{100}$) of 105 $\\mathit{\\mu}$m (Propyltex 140S, Micro Powders Inc., Tarrytown, NY, USA) at 50 wt% and mixed with distilled water. The slurry was magnetically stirred and ball milled for a minimum of 24 h to break up the agglomerates. Afterwards, the slurry was frozen by dipping the container in liquid nitrogen and freeze-dried to obtain an homogeneous mix. Then, 8 g of the obtained powder was pressed at 8 MPa in a mold of 20 mm diameter. The sintering temperature and dwell time were the same as those used for the ice-templated samples. To ensure a proper burn-out and crack free samples an extra hold of 1 h was added at 900\u00a0${}^{\\circ}$C.\n\n2.2.. Sample characterization {#S2002}\n-----------------------------\n\nThe overall porosity P(%) was calculated based on the mass (*m*) and volume (*V*) of the samples with respect to that of fully dense TZ-3YS ($\\mathit{\\rho}_{ysz} = 5.8gcm^{- 3}$), as:$$\\begin{aligned}\n\\mathit{\\rho}_{rel} & {= \\frac{\\mathit{\\rho}}{\\mathit{\\rho}_{ysz}} = \\frac{mV^{- 1}}{\\mathit{\\rho}_{ysz}}} \\\\\n\\end{aligned}$$ $$\\begin{aligned}\n{P(\\%)} & {= (1 - \\mathit{\\rho}_{rel}) \\times 100\\%} \\\\\n\\end{aligned}$$\n\nPore size was evaluated as the width of the lamellae and quantified using two techniques: image analysis and mercury intrusion porosimetry. Image analysis was performed in at least five cross-sections perpendicular to the freezing direction (height was kept constant at 7 mm from the bottom of the sample). Some samples were sectioned parallel to the freezing direction to evaluate the directionality of the pores. All the images were obtained using a scanning electron microscope (Nova NanoSEM 230, FEI, Hillsboro, OR, USA). Furthermore, pore size results were confirmed by mercury intrusion porosimetry (AutoPore IV 9500, Micromeritics, Norcross, GA, USA) with an applied pressure up to 0.31 bar.\n\nGas permeability was measured using a custom built equipment and is shown in Figure [1](#F0001){ref-type=\"fig\"}. Synthetic air was passed through the sample which was held by a silicone ring. Two sensors were placed before and after to record the inlet ($P_{i}$) and outlet pressure ($P_{o}$) at room temperature. The tested samples had a thickness (*L*) of 12--15 mm and a diameter 15 mm. In the case of ice-templated specimens, pores were always aligned parallel to the flow during the test.\n\nPermeability was evaluated using the Forchheimer's equation adapted for compressible fluids, Equation ([3](#M0003){ref-type=\"disp-formula\"}):$$\\begin{array}{r}\n{\\frac{\\Delta P}{L} = \\frac{{P_{i}}^{2} - {P_{o}}^{2}}{2P_{o}L} = \\frac{\\mathit{\\mu}}{k_{1}}\\mathit{\\nu}_{s} + \\frac{\\mathit{\\rho}}{k_{2}}{\\mathit{\\nu}_{s}}^{2}} \\\\\n\\end{array}$$\n\nwhere L is the thickness of the sample (parallel to the flow), $\\mathit{\\mu}$ and $\\mathit{\\rho}$ are the viscosity and the density of the fluid (air at 25\u00a0${}^{\\circ}$C), $\\Delta P$ is the pressure drop, $k_{1}$ and $k_{2}$ are the Darcian and non-Darcian permeabilities, and $\\mathit{\\nu}_{s}$ is the fluid velocity (air flow divided by the cross-sectional area of the sample). The maximum air flow ($\\mathit{\\nu}_{s}$) generated was 2 m s${}^{- 1}$ and $\\Delta P$ did not exceed 350 MPa.\n\n![Setup used to measure pressure drop.](tsta_a_1197757_f0001_oc){#F0001}\n\n3.. Results and discussion {#S0003}\n==========================\n\n3.1.. Microstructure of ice-templated samples {#S2003}\n---------------------------------------------\n\nFigure [2](#F0002){ref-type=\"fig\"}(a)--(d) shows cross sections perpendicular to the solidification front for the samples S1, S4, S5, and S8 in Table [1](#T0001){ref-type=\"table\"}. All of them exhibit a lamellar pore morphology characteristic of ice-templated samples with water used as a solvent and without ice-shaping additives incorporated to the initial slurry. Total pore volume and pore size were almost independently controlled by the initial solids loading and cooling rate respectively. Figures [2](#F0002){ref-type=\"fig\"}(a) and (b) show the microstructures obtained in two samples frozen at identical cooling rate (2\u00a0${}^{\\circ}$C min${}^{- 1}$) but different solids loading (50 and 65 wt%). The modification of the solids loading causes a variation on the relative density of the samples, and the total pore volume changes accordingly. Increasing the solids loading from 50 wt% (14.7 vol.%) to 65 wt% (24.2 vol.%) causes a decrease in total pore volume from 72% to 54%. The same effect has also been observed at higher freezing rate (Figure [2](#F0002){ref-type=\"fig\"}(c) and (d), but the pore size obtained was smaller. For example, when freezing temperature increased from 2 to 25\u00a0${}^{\\circ}$C min${}^{- 1}$, mean pore size ($d_{c}$) decreased from 20 $\\pm$ 9 $\\mathit{\\mu}$m (S1) to 5 $\\pm$ 2 $\\mathit{\\mu}$m (S5); however, pore volume remained unchanged, at 72%.\n\n![Microstructures of ice-templated samples listed in Table [1](#T0001){ref-type=\"table\"}, with cross-sections perpendicular to the freezing direction: (a) S1, (b) S4, (c) S5, and (d) S8; and parallel to it: (e) S1, (f) S4, (g) S5, and (h) S8.](tsta_a_1197757_f0002_oc){#F0002}\n\nTable [1](#T0001){ref-type=\"table\"} also highlights the effect of solids loading on pore size. Specimens ice-templated with 50 wt% solids loading exhibit a pore size remarkably larger (20 $\\pm$ 9 $\\mathit{\\mu}$m) than those ice-templated with 65 wt% (14 $\\pm$ 5 $\\mathit{\\mu}$m). This effect is caused by the increasing number of ceramics particles at higher solids loading that constrain the crystal growth, and consequently decrease the mean pore size. Although the pore size reduction caused by the solids loading holds a second order of importance compared with the cooling rate, it is important to consider it because certainly might affect the permeability.\n\n###### \n\nProperties of ice-templated samples. ${d_{c}}_{Hg}$ and ${d_{c}}_{IA}$ refer to the mean pore size measured by mercury porosimetry and image analysis respectively. $R^{2}$ corresponds to the correlation coefficient of the trend lines shown in Figure [4](#F0004){ref-type=\"fig\"}.\n\n  Sample   Solids loading (wt%)   Solids loading (vol.%)   Cooling rate (${}^{\\circ}$C min${}^{- 1}$)   Porosity (%)   Mean ${d_{c}}_{Hg}$ ($\\mathit{\\mu}$m)   Mean ${d_{c}}_{IA}$ ($\\mathit{\\mu}$m)   $k_{1}$ (m${}^{2}$)            $R^{2}$\n  -------- ---------------------- ------------------------ -------------------------------------------- -------------- --------------------------------------- --------------------------------------- ------------------------------ ---------\n  S1       50%                    14.7%                    2                                            72%            19.1                                    20 $\\pm$ 9                              1.39$\\, \\times \\, 10^{- 11}$   0.999\n  S2       55%                    17.4%                    2                                            67%            17.5                                    18 $\\pm$ 7                              9.44$\\, \\times \\, 10^{- 12}$   0.999\n  S3       60%                    20.6%                    2                                            60%            14.2                                    14 $\\pm$ 6                              5.95$\\, \\times \\, 10^{- 12}$   0.999\n  S4       65%                    24.2%                    2                                            54%            8.3                                     14 $\\pm$ 5                              2.17$\\, \\times \\, 10^{- 12}$   0.998\n  S5       50%                    14.7%                    25                                           72%            5.4                                     5 $\\pm$ 2                               1.08$\\, \\times \\, 10^{- 13}$   0.999\n  S6       55%                    17.4%                    25                                           68%            4.3                                     4 $\\pm$ 2                               7.99$\\, \\times \\, 10^{- 14}$   0.999\n  S7       60%                    20.6%                    25                                           60%            3.9                                     4 $\\pm$ 2                               4.27$\\, \\times \\, 10^{- 14}$   0.999\n  S8       65%                    24.2%                    25                                           55%            2.9                                     3 $\\pm$ 1                               2.36$\\, \\times \\, 10^{- 14}$   0.999\n\nFigure [2](#F0002){ref-type=\"fig\"}(e)--(h) show the microstructures of the same ice-templated samples parallel to the solidification front. In all cases, pores are continuous with an almost constant cross-section, independently of solids loading (Figure [2](#F0002){ref-type=\"fig\"}(e) and (f)) and cooling rate (Figure [2](#F0002){ref-type=\"fig\"}(g) and (h). Moreover, all the samples exhibited the characteristic dendritic growth of ice-templated materials.\n\nSince air flow in porous materials is a volumetric phenomenon and is highly sensitive to pore size, a complementary evaluation to image analysis was performed by mercury porosimetry. This technique probes the entire volume of the sample in contrast to the local measurements obtained by image analysis.\n\nFigure [3](#F0003){ref-type=\"fig\"} shows the pore size distribution of two representative samples (S2 and S6) described in Table [1](#T0001){ref-type=\"table\"}. As extensively reported \\[[@CIT0020]--[@CIT0024]\\], porous ceramics processed by ice-templating usually exhibited a bimodal pore distribution. In both samples, larger pore size corresponds to the open unidirectional porosity created by the sublimation of the ice crystals and, thus mainly controlled by the solids loading. The double peak observed at around 20 $\\mathit{\\mu}$m for S2 and 4 $\\mathit{\\mu}$m for S6 can be related with the primary and secondary arms characteristic of the ice crystal during the dendritic growth. Alternatively, the smaller peak (around 0.1 $\\mathit{\\mu}$m) corresponds to porosity located in the ceramic walls and can be tailored by the sintering temperature and powder characteristics.\n\n![Pore size distribution obtained by mercury intrusion porosimetry. S2 and S6 refer to specimens described in Table [1](#T0001){ref-type=\"table\"}. In both cases P(%) = 68%.](tsta_a_1197757_f0003_oc){#F0003}\n\n![Effect of air velocity ($\\mathit{\\nu}_{s}$) on pressure drop ($\\Delta P$) for different pore volumes parallel to the flow. Samples were ice-templated samples at (a) 2\u00a0${}^{\\circ}$C min${}^{- 1}$ and (b) 25\u00a0${}^{\\circ}$C min${}^{- 1}$.](tsta_a_1197757_f0004_oc){#F0004}\n\nResults obtained by both techniques, image analysis and mercury intrusion porosimetry, are in excellent agreement, as shown Table [1](#T0001){ref-type=\"table\"}. For example, S1 exhibits a pore size of 19.1 $\\mathit{\\mu}$m obtained by mercury porosimetry, and remarkably close to 20 $\\pm$ 9 obtained by image analysis. In general, the pore size measured by mercury intrusion should show smaller size than those measured by image analysis, due to the effect of bottle-necks. However, in these ice-templated unidirectional porous ceramics, the cross-section of the pores is constant, so the presence of bottle-necks tends to be minimized. The only exception is the specimen S4 that gives a slightly lower pore size value, probably caused by the presence of volumetric defects originated by the high solids loading.\n\n3.2.. Pressure drop and permeability {#S2004}\n------------------------------------\n\nFigure [4](#F0004){ref-type=\"fig\"}(a) and (b) show the variation of pressure drop ($\\Delta P$) with air velocity ($\\mathit{\\nu}_{s}$) for the specimens S1--S8 described in Table [1](#T0001){ref-type=\"table\"}. Obviously, increasing $\\mathit{\\nu}_{s}$ leads to increase $\\Delta P$. However, the different slopes exhibitedby all groups highlight the importance of pore volume and pore size on the gas flow resistance parallel to pore orientation and hence pressure drop.\n\nAll samples followed a linear relationship between $\\Delta P$ and $\\mathit{\\nu}_{s}$. This behavior can be explained based on the Forchheimer's equation (Equation ([3](#M0003){ref-type=\"disp-formula\"})). The first term ($\\mathit{\\mu}\\mathit{\\nu}_{s}/k_{1}$) in Equation ([3](#M0003){ref-type=\"disp-formula\"}) considers the viscous energy losses caused by the friction between fluid layers, and the second term ($\\mathit{\\rho}\\mathit{\\nu}_{s}^{2}/k_{2}$) accounts for the inertial effects created by turbulence and variations in the direction and acceleration of the fluid.\\[[@CIT0025]\\] The excellent linear correlation coefficient ($R^{2}$ \\> 0.998 in Table [1](#T0001){ref-type=\"table\"}) supports the hypothesis that all the tests were performed in viscous flow regime and thus, the determination of permeability is only subjected to the reduced form of Equation ([3](#M0003){ref-type=\"disp-formula\"}), i.e. Darcy's law:$$\\begin{array}{r}\n{\\frac{\\Delta P}{L} = \\frac{{P_{i}}^{2} - {P_{o}}^{2}}{2P_{o}L} = \\frac{\\mathit{\\mu}}{k_{1}}\\mathit{\\nu}_{s}} \\\\\n\\end{array}$$\n\nNevertheless, careful considerations should be made if $k_{1}$ is intended to be applied in systems with higher air flow. In these conditions the inertial effects and turbulence become important and both constants ($k_{1}$ and $k_{2}$) are required to describe the behavior.\\[[@CIT0026]\\]\n\nFigure [5](#F0005){ref-type=\"fig\"} and Table [1](#T0001){ref-type=\"table\"} show the effect of initial solids loading (and hence porosity) and cooling rate on permeability. Decreasing the solids loading causes an increase in $k_{1}$ that can be correlated with an increase in total pore volume as well as a slight increase in pore size. The specimens frozen faster are remarkably less permeable. The main explanation for this behavior relies on the important reduction in pore size observed at these freezing conditions. The fact that the decrease in permeability is around two orders of magnitude in samples with the same total pore volume emphasizes the great importance of pore size to maximize the fluid flow.\n\n![Effect of solids loading and freezing rate on permeability. Experimental and morphological parameters are specified in Table [1](#T0001){ref-type=\"table\"}. Slow and fast cooling rate correspond to 2 and 25\u00a0${}^{\\circ}$C min${}^{- 1}$ respectively. Circles around the markers are proportional to pore size.](tsta_a_1197757_f0005_oc){#F0005}\n\nThe values obtained in this work are in agreement with those found in the literature for ice-templated samples with similar microstructure. Pekor et al. \\[[@CIT0016]\\] reported a permeability of 1.0$\\, \\times \\, 10^{- 11}$ m${}^{2}$ for ice-templated alumina with 70% porosity and pore size around 17 $\\mathit{\\mu}$m. Fukushima et al. \\[[@CIT0014]\\] obtained a slightly higher $k_{1}$ (2.27$\\, \\times \\, 10^{- 11}$ m${}^{2}$) for SiC prepared by gelation freezing method but with a remarkably higher pore volume and pore size (86% and 34 $\\mathit{\\mu}$m).\n\nComparing the permeability exhibited by ice-templated materials with other macroporous materials obtained by different techniques, we can see in Figure [6](#F0006){ref-type=\"fig\"} that the former falls in the wide range exhibited by gelcasted foams and much lower than those typically exhibited by reticulated foams. However, the total pore volume, and more importantly the pore size of this type of cellular materials, is much higher compared with those obtained by ice-templating. For example, Innocentini et al. \\[[@CIT0003]\\] reported a $k_{1}$ of 1.5$\\, \\times \\, 10^{- 10}$ m${}^{2}$ for a a porous alumina processed by replica method with 70% porosity and 200 $\\mathit{\\mu}$m pore size and a permeability of 5$\\, \\times \\, 10^{- 11}$ m${}^{2}$ for a porous alumina prepared by gelcasting with the same pore volume and 125 $\\mathit{\\mu}$m pore size.\n\n![Comparison of the permeabilities that can be achieved by the most common processing techniques \\[[@CIT0001]\\]. Description of specimens labeled as 'IT-Lamellar' is shown in Table [1](#T0001){ref-type=\"table\"}. Blue and red markers correspond to samples S1--S4 and S5--S8 respectively.](tsta_a_1197757_f0006_oc){#F0006}\n\n![Representative microstructures of specimens shown in Figure [6](#F0006){ref-type=\"fig\"}. (a) Ice-templated, lamellar morphology, and slow cooling rate; (b) ice-templated, lamellar morphology, and fast cooling rate; (c) ice-templated, honeycomb morphology, and slow cooling rate; (d) pore formers. All of them exhibited a similar pore volume (P $\\, \\approx \\,$ 54%).](tsta_a_1197757_f0007_b){#F0007}\n\nAdditionally, we evaluated the effect of pore directionality (i.e. 'Pore Formers' in Figure [6](#F0006){ref-type=\"fig\"}) and morphology (i.e. 'IT-Honeycomb' in Figure [6](#F0006){ref-type=\"fig\"}) on permeability. The first group consists of samples with isotropic porosity processed by burn out of organic pore formers (Figure [7](#F0007){ref-type=\"fig\"}(d)), and the second group comprises ice-templated samples with zirconium acetate (ZRA) added to the original slurry. The addition of ZRA turns the pore morphology into a honeycomb-like structure with smooth surfaces (Figure [7](#F0007){ref-type=\"fig\"}(c)). As Figure [8](#F0008){ref-type=\"fig\"} shows, this porous structure posses a larger mean pore size and a broader distribution than the typical lamellar ice-templated morphology (Figure [7](#F0007){ref-type=\"fig\"}(a)). This variation in pore size distribution causes an increase in permeability of almost one order of magnitude at comparable total pore volume (Figure [6](#F0006){ref-type=\"fig\"}). In contrast, the specimens processed by pore formers have a similar permeability as the ice-templated specimens with lamellar morphology and therefore an apparent no effect of pore tortuosity (Figure [6](#F0006){ref-type=\"fig\"}). However, although the mean pore size is similar in both samples, the distribution tends to be more spread in materials processed by pore formers (Figure [8](#F0008){ref-type=\"fig\"}). The presence of these big pore clusters might facilitate the air flow through the samples and consequently mask the effect of pore directionality. Although the presence of larger isolated pores certainly increases the permeability, they can become stress concentrators and therefore affect the mechanical integrity of the system.\n\n![Pore size distributions obtained by image analysis of different ice-templated structures compared with sacrificial method (pore formers). Microstructures are shown in Figure [7](#F0007){ref-type=\"fig\"}.](tsta_a_1197757_f0008_oc){#F0008}\n\nMost of the important industrial applications involving porous materials are subjected to a trade-off between permeability and strength \\[[@CIT0027]\\]. Since both properties are inversely related to total pore volume, other strategies different than increasing/decreasing this parameter are necessary to improve both properties simultaneously.\n\nFigure [9](#F0009){ref-type=\"fig\"} shows the effect of microstructural variations in permeability studied in this work along with their corresponding compressive strength reported in a previous article.\\[[@CIT0017]\\] As expected, the ice-templated specimens with smaller pore size have the highest strength. This behavior might be caused by the larger connectivity among the walls that prevents the structure from buckling and shearing stresses.\\[[@CIT0024]\\] Unfortunately, this pore structure also exhibits the lowest permeability. On the other hand, ice-templated samples frozen at slow cooling rate, either lamellar or honeycomb morphology, exhibited both a higher compressive strength and permeability than the samples prepared by pore formers. Therefore, the unidirectionally aligned porosity of ice-templated materials seems to be more optimized to maximize mechanical stability and permeability than the randomly oriented porosity exhibited by materials processed by pore formers.\n\nThe ice-templated specimens with honeycomb morphology reached permeability values comparable to ice-templated samples with P = 72% coupled with a compressive strength around one order of magnitude higher. This microstructure exhibits higher permeability than required for diesel particulate filters,\\[[@CIT0028]\\] and it could be of particular interest in applications that require a reasonably high mass flow and low pressure drop (e.g. membranes for water/air filtration and catalytic supports).\\[[@CIT0029]\\]\n\n![Compressive strength and permeability of ice-templated (IT) and pore formers samples processed in this work. $d_{p}$ and P refer to pore size and porosity respectively. Pore size distributions are shown in Figure [8](#F0008){ref-type=\"fig\"}.](tsta_a_1197757_f0009_oc){#F0009}\n\n3.3.. Prediction by models {#S2005}\n--------------------------\n\nMost of the models used to predict the permeability constant ($k_{1}$) are extracted from correlations originally developed for granular beds and are dependent on the volumetric void fraction ($\\mathit{\\epsilon}$), and the equivalent particle size ($d_{p}$). The most common expression was developed by Ergun \\[[@CIT0030]\\] (Equation ([5](#M0005){ref-type=\"disp-formula\"})) and it has been extensively applied to all types of porous media.\\[[@CIT0031]--[@CIT0033]\\]$$\\begin{array}{r}\n{k_{1} = \\frac{\\mathit{\\epsilon}^{3}{d_{p}}^{2}}{150{(1 - \\mathit{\\epsilon})}^{2}}} \\\\\n\\end{array}$$\n\n![Comparison between the experimental permeability data and the values predicted by the original (A) and modified (B) Ergun equations. Two different methods were used to evaluated the pore size, image analysis (IA) and mercury porosimetry (poro). Samples labeled as 'Slow cooling rate' and 'Fast cooling rate' correspond to S1--S4 and S5--S8 in Table [1](#T0001){ref-type=\"table\"} respectively.](tsta_a_1197757_f0010_oc){#F0010}\n\nHowever, this approach is hard to apply to cellular materials due to the difficulties in extracting the characteristic parameter that could replace the particle size ($d_{p}$). One solution consists on replacing $d_{p}$ by an equivalent pore size ($d_{c}$) \\[[@CIT0034]\\] and then, $k_{1}$ can be rewritten as:$$\\begin{array}{r}\n{k_{1} = \\frac{2.25}{150}\\mathit{\\epsilon}{d_{c}}^{2}} \\\\\n\\end{array}$$\n\nThe experimental values of $k_{1}$ for ice-templated samples frozen at 2\u00a0${}^{\\circ}$C min${}^{- 1}$ and 25\u00a0${}^{\\circ}$C min${}^{- 1}$ were compared with those predicted by Equation ([6](#M0006){ref-type=\"disp-formula\"}). Since the evaluation of permeability by Ergun's equations is highly sensitive to the equivalent pore size ($d_{c}$), the assessment of the model was conducted with the pore sizes measured by image analysis and mercury porosimetry. As Figure [10](#F0010){ref-type=\"fig\"}(a) shows, all the experimental points diverge from the prediction of the Ergun model (dashed line) independently of the technique used to characterize the pore size. The only exceptions are the specimens frozen at slow cooling rate and lower permeability which seemingly exhibited a better fitting when the pore size was obtained by image analysis. However, this result clearly deviates from the tendency displayed by the other samples and can be attributed to the increasing presence of closed pores found at high solids loading (S4 in Table [1](#T0001){ref-type=\"table\"}).\n\nThe main factor that explains the deviation between experimental and calculated values of $k_{1}$ relies on the semi-empirical nature of Ergun expression. The value of 150 in Equation ([6](#M0006){ref-type=\"disp-formula\"}) was originally determined on different experiments on highly tortuous media, and thus it might not be valid for radically different porous structures like unidirectional porous materials. Nevertheless, since it is an expression broadly used and for the sake of comparison, we corrected the original Ergun equation (Equation ([6](#M0006){ref-type=\"disp-formula\"})) to represent the experimental data obtained in this work (Figure [10](#F0010){ref-type=\"fig\"}(B)). Equation ([7](#M0007){ref-type=\"disp-formula\"}) describes the permeability of samples frozen at 2\u00a0${}^{\\circ}$C min${}^{- 1}$ and Equation ([8](#M0008){ref-type=\"disp-formula\"}) for samples frozen at 25\u00a0${}^{\\circ}$C min${}^{- 1}$.$$\\begin{aligned}\nk_{1} & {= \\frac{2.25}{42}\\mathit{\\epsilon}{d_{c}}^{2}} \\\\\n\\end{aligned}$$ $$\\begin{aligned}\nk_{1} & {= \\frac{2.25}{407}\\mathit{\\epsilon}{d_{c}}^{2}} \\\\\n\\end{aligned}$$\n\nThe microstructure of ice-templated materials better resemble a capillary system where pores are aligned parallel to the gas flow than a porous granular media. The maximum permeability in an ideal capillary model is described by Equation ([9](#M0009){ref-type=\"disp-formula\"}):\\[[@CIT0015]\\]$$\\begin{array}{r}\n{K_{capillary} = \\frac{\\mathit{\\phi}d_{c}^{2}}{\\mathit{\\tau}32}} \\\\\n\\end{array}$$\n\nwhere $\\mathit{\\phi}$, $d_{c}$, and $\\mathit{\\tau}$ are porosity, pore size, and tortuosity respectively.\n\nThe capillary permeability ($K_{capillary}$) of ice-templated samples was calculated based on Equation ([9](#M0009){ref-type=\"disp-formula\"}) and using the measurements of porosity and pore size obtained by mercury porosimetry in Table [1](#T0001){ref-type=\"table\"}. Figure [11](#F0011){ref-type=\"fig\"} shows the experimental values of $k_{1}$ compared with those predicted by the model. The permeability of ice-templated samples frozen at slow cooling rate (2\u00a0${}^{\\circ}$C min${}^{- 1}$) is close to the maximum obtained when $\\mathit{\\tau} = 1$. This result indicates that the porosity of these samples is almost continuous and the capillary model is thus a good descriptor of permeability in these samples. However, the permeability of ice-templated materials is not always predicted by the ideal case of Equation ([9](#M0009){ref-type=\"disp-formula\"}) where $\\mathit{\\tau} = 1$. For example, the best fit for high freezing rate specimens (25\u00a0${}^{\\circ}$C min${}^{- 1}$) is exhibited around $\\mathit{\\tau} = 5$. Usually, in isotropic porous materials, tortuosity tends to decrease when pore volume increases.\\[[@CIT0035]\\] Unlike in isotropic samples, tortuosity of ice-templated materials remains constant in a wide porosity range (50--70%) but is affected by pore size.\n\n![Comparison between the experimental permeability data and the values predicted by the capillary model.](tsta_a_1197757_f0011_oc){#F0011}\n\n4.. Conclusions {#S0004}\n===============\n\nThe air permeability of ice-templated materials was evaluated as a function of different morphological parameters such as pore volume, size, and morphology. The results showed that pore size is the main parameter controlling the pressure drop and is therefore of capital importance to increase the permeability of porous materials without a detrimental effect on the mechanical stability. Based on this knowledge we produced ice-templated samples with a 53% pore volume and a permeability equivalent to 72% pore volume through a pore shape modification. This reduction on total pore volume allowed samples to be obtained with a high resistance coupled with a high permeability, two properties seemingly inversely related.\n\nMicrostructure control using ice-templating allows us to tailor the mechanical and gas flow properties almost independently and consequently could be of potential application in products such as filters or catalytic supports. Using a pressure drop test we evaluated the tortuosity of ice-templated samples and we obtained a value remarkably close to the ideal case ($\\mathit{\\tau} = 1$). In this case, tortuosity remained constant for a pore volume increase and it was mainly controlled by pore size, unlike in isotropic porous structures. Finally, we also demonstrate that the permeability of unidirectional porous materials can be described by the capillary model in a wide porosity range (50--70%). However, this study only covers the laminar regime and a deeper understanding of the microstructural effects on a turbulent flow is still required in these types of materials.\n\nThe authors would like to thank J. Leloup for his contribution to the permeability measurements.\n\nNo potential conflict of interest was reported by the authors.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nObesity has reached pandemic proportions worldwide, significantly contributing to reduce life quality and lifespan at a global scale [@bib1]. This condition is characterized by abnormal and excessive fat accumulation and is influenced by both genetic and environmental determinants. While several genetic loci have been associated with obesity, they explain only a fraction of the total variance within populations; moreover, genes deemed obesity-predisposing interact with environmental factors to regulate, for instance, satiety and energy expenditure [@bib2]. Among the environmental determinants of obesity and its associated dysmetabolic conditions, dietary habits play a central role. Diet\u00a0also strongly influences our \"other genome\" (*i.e.*, the metagenome), modeling gut microbial community structure [@bib3] and impacting host metabolism and energy partitioning [@bib4]. Research conducted throughout the last decade has revealed a clear association between obesity and gut microbial dysbiosis, which is generally characterized by a reduction in bacterial richness and by major taxonomic and functional changes [@bib5].\n\nThe consumption of fiberless diets rich in simple sugars and saturated fat (often referred to as Western diets) generates well-known detrimental metabolic consequences, leading to insulin resistance and glucose intolerance in the early-term, which later evolves to overt obesity, type 2 diabetes, and cardiovascular complications. NAFLD and NASH are highly prevalent diseases occurring in the setting of obesity and type 2 diabetes; they may eventually progress to hepatocellular carcinoma and contribute to dysregulate glucose and lipid homeostasis [@bib6]. The gut exerts major influences on liver physiology as both organs are anatomically and functionally connected by means of the portal circulation, but also since bacteria and bacteria-derived molecules can translocate from the gut to the liver and can potentially contribute to diet-induced insulin resistance and liver disease [@bib7], [@bib8].\n\nPlant-rich diets are abundant in fruits and vegetables and strongly linked to lean and healthy phenotypes [@bib9], which prompts the search for bioactive phytonutrients to treat or prevent obesity and its related dysmetabolic conditions. The use of polyphenol-rich fruit extracts or isolated polyphenols as strategies to alleviate obesity-linked diseases have been demonstrated in humans [@bib10], [@bib11] and in animal models [@bib12], [@bib13], but the mechanisms of action are not yet fully elucidated. Several dietary polyphenols are generally poorly bioavailable and build up in the colon, where they are modified by gut microbial enzymes and, in turn, reshape gut microbial communities [@bib14]. We have previously demonstrated that a polyphenol-rich cranberry extract prevents diet-induced obesity in high fat high sucrose-fed mice, and these findings were linked to improved gut-liver homeostasis and expansion of *Akkermansia muciniphila* population in the gut microbiota [@bib12]. Similar effects were reported by others using a polyphenol-rich extract of concord grape [@bib15] and apple proanthocyanidins [@bib16], but no studies have yet tested whether polyphenols can reverse an already established obesity and more severe metabolic alterations, including hepatic steatosis and inflammation. In the present study, we investigated the potential of a polyphenol-rich cranberry extract to reverse an already established obesity, insulin resistance, and NAFLD, and whether such effects may be linked to the reshaping of the gut microbiota and blooming of *A*. *muciniphila*, a well-known target of food polyphenols.\n\n2. Material and methods {#sec2}\n=======================\n\n2.1. Animals {#sec2.1}\n------------\n\nAll animal experiments reported in this manuscript comply with the Animal Research: Reporting of *In\u00a0Vivo* experiments (ARRIVE) guidelines. Eight week-old C57Bl/6J male mice (Jackson, USA) were housed 2--3 animals per cage, kept on Sani-chips bedding and in controlled environment (12\u202fh daylight cycle, lights off at 18:00) with food and water *ad libitum* in the animal facility of the Qu\u00e9bec Heart and Lung Institute (Qu\u00e9bec, Canada). After two weeks of acclimatization, mice were pre-fed either a healthy Chow (Teklad 2018, Harlan) or a High-Fat/High-Sucrose (HFHS) diet for 21 weeks. Diet composition was previously published [@bib12] and, although the abbreviation HFHS particularly refers to the enriched presence of saturated fat and simple sugars, it is important to stress that the lack of soluble fibers is a major obesogenic component of this diet [@bib17]. During the last 8 weeks of the study (i.e. from the beginning of week 13 to the beginning of week 21), control groups (Chow, n\u202f=\u202f8 and HFHS, n\u202f=\u202f8) were orally administered the animal facility\\'s drinking water whereas the treated groups (Chow\u202f+\u202fCE, n\u202f=\u202f11 and HFHS\u202f+\u202fCE, n\u202f=\u202f10) received a cranberry extract (CE, 200\u202fmg/kg, Nutra Canada, Qu\u00e9bec, Canada). The polyphenolic profile of CE was published elsewhere [@bib12], [@bib18]. Body weight gain and food intake were assessed twice weekly. At week 21, animals were anesthetized in chambers saturated with isoflurane and then sacrificed by cardiac puncture. Organs and tissues were carefully collected and blood was drawn in tubes containing 2\u202fIU of heparin and immediately centrifuged in order to separate plasma from cells. All interventions were carried out during the animals\\' light cycle. All procedures strictly followed the National Institutes of Health (NIH)'s Guide for the Care and Use of Laboratory Animals and were previously approved by the Laval University Animal Ethics Committee.\n\n2.2. Glucose homeostasis {#sec2.2}\n------------------------\n\nAt week 17, mice were fasted for 6\u202fh and insulin tolerance tests (ITT) were performed after intraperitoneal injections of insulin (0.75 UI/kg body weight). Glycemia was measured with an Accu-Check glucometer (Bayer) before (0\u202fmin) and after (10, 20, 30, 60, and 90\u202fmin) insulin injection. At the end of week 19, mice were fasted overnight (12\u202fh) and oral glucose tolerance tests were carried out (OGTT, 1\u202fg of glucose/kg body weight). Blood was collected before (0\u202fmin) and after (15, 30, 60, 90, and 120\u202fmin) glucose challenge for glycemia determination. Blood samples (\u223c30\u202f\u03bcL) were collected at each time point during OGTT and insulinemia was determined using an ultra-sensitive ELISA kit (Alpco, USA). The homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated based on the following formula: fasting insulinemia (\u03bcUI/mL) x fasting glycemia (mM)/22.5.\n\n2.3. Oil red O staining {#sec2.3}\n-----------------------\n\nDuring necropsies, mouse livers were embedded in Tissue-Tek\u00ae OCT, immediately snap-frozen in liquid nitrogen, and stored at\u00a0\u221280\u202f\u00b0C. Staining of neutral lipids was based on the methods described by Mehlem et\u00a0al. with some adaptations [@bib19]. Briefly, 12\u202f\u03bcm liver sections were allowed to equilibrate at room temperature for 5\u202fmin and then post-fixed with a Formalin (10%)/Calcium (2%) solution for 15\u202fmin. The sections were then incubated with oil red O (ORO) working solution at room temperature for 5\u202fmin, followed by a 5- minute clearing in 60% isopropyl alcohol and a counterstaining of 15\u202fs with Mayer\\'s hematoxylin.\n\n2.4. Q-PCR and antioxidant enzymes {#sec2.4}\n----------------------------------\n\nGene mRNA expression analysis by q-PCR and quantification of antioxidant enzymes were carried out as previously described [@bib12]. Primer sequences used were: (5\u2032\u2192 3\u2032) PPAR\u03b1 F- CGACCTGAAAGATTCGGAAA, R- GGCCTTGACCTTGTTCATGT; PPAR\u03b3 F- CAGGCCTCATGAAGAACCTT, R- GCATCCTTCACAAGCATGAA; SREBP1c F- GACCCTACGAAGTGCACACA, R- TCATGCCCTCCATAGACACA; SREBP2 F- CGACCAGCTTTCAAGTCCTG, R- CCTGTACCGTCTGCACCTG; LXR\u03b1 F- GGAGTGTCGACTTCGCAAAT, R- CTTGCCGCTTCAGTTTCTTC; LXR\u03b2 F- AAACGATCTTTCTCCGACCA, R- ATGGCTAGCTCGGTGAAGTG; COX2 F- GCTGTACAAGCAGTGGCAAA, R- CCCCAAAGATAGCATCTGGA; TNF\u03b1 F- GAACTGGCAGAAGAGGCACT, R- AGGGTCTGGGCCATAGAACT; NF\u03baB F- AGCTTCACTCGGAGACTGGA, R- ACGATTTTCAGGTTGGATGC; I\u03baB F- TGGCCAGTGTAGCAGTCTTG, R- GACACGTGTGGCCATTGTAG.\n\n2.5. Cecal mucin determination {#sec2.5}\n------------------------------\n\nCecal contents were collected at week 21, snap-frozen in liquid nitrogen and stored at\u00a0\u221280\u202f\u00b0C. Cecal feces were freeze-powdered and the presence of mucins was determined using a fluorometric assay kit (Cosmo Bio, Japan) that discriminates O-linked glycoproteins (mucins) from N-linked glycoproteins.\n\n2.6. Fecal samples {#sec2.6}\n------------------\n\nFecal samples were freshly collected at baseline (week 13) and week 21 and immediately stored at\u00a0\u221280\u202f\u00b0C. Bacterial genomic DNA was extracted from approximately 50\u202fmg of fecal material collected from each cage. Samples were resuspended in lysis buffer containing 20\u202fmg/ml lysozyme and incubated for 30\u202fmin at 37\u202f\u00b0C. Further lysis was performed by adding 10% SDS and proteinase K to 350\u202f\u03bcg/ml followed by incubation for 30\u202fmin at 60\u202f\u00b0C. Samples were homogenized using a bead beater and 0.1\u202fmm zirconium beads and then processed using a DNA extraction kit (DNeasy, Qiagen). DNA yield was assessed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific). Extracted DNA was stored at\u00a0\u221220\u202f\u00b0C until use. Each DNA sample was subsequently used for 16S amplification of the V3-V4 region using the primers 341F (5\u2032-CCTACGGGNGGCWGCAG-3\u2032) and 805R (5\u2032-GACTACHVGGGTATCTAATCC-3\u2032) adapted to incorporate the transposon-based Illumina Nextera adapters (Illumina, USA) and a sample barcode sequence allowing multiplexed paired-end sequencing. Constructed 16S metagenomic libraries were purified using 35\u202f\u03bcL of magnetic beads (AxyPrep Mag PCR Clean up kit; Axygen Biosciences, USA) per 50\u202f\u03bcL PCR reaction. Library quality control was performed with a Bioanalyzer 2100 using DNA 7500 chips (Agilent Technologies, USA). An equimolar pool was obtained and checked for quality prior to further processing. The pool was quantified using picogreen (Life Technologies, USA) and loaded on a MiSeq platform using 2\u202f\u00d7\u202f300\u202fbp paired-end sequencing (Illumina, USA). High-throughput sequencing was performed at the IBIS (Institut de Biologie Int\u00e9grative et des Syst\u00e8mes - Universit\u00e9 Laval). All raw sequences were deposited in the public European Nucleotide Archive server under accession number PRJEB23031.\n\n2.7. 16S rRNA gene-based gut microbial analysis {#sec2.7}\n-----------------------------------------------\n\nGenerated and demultiplexed sequences were analyzed using the QIIME software package (version 1.9.1). Paired-end sequences were merged with at least a 50-bp overlap using fastq-join. Resulting sequences containing ambiguous or low quality reads (Phred score\u202f\u2264\u202f25) were removed from the dataset. Forward and reverse primers were trimmed from the filtered sequences; reads with at least one reverse primer mismatch or where the reverse primer was not found were discarded. Chimera checking and filtering was performed using UCHIME (4). OTU (Operational Taxonomic Units)-picking from post-filtering reads was performed using USEARCH 61 version 6.1.544 [@bib20] with an open-reference methodology, which consisted of clustering sequences *de novo* at 97% identity threshold if they did not hit the reference sequence collection. Representative OTU sequences were assigned taxonomy against the Greengenes reference database (August 2013 release) [@bib21] using the RDP-classifier [@bib22]. Singleton OTUs and OTUs with a number of sequences \\<0.005% of total number of sequences were discarded at this step [@bib23]. Unclassified OTUs at the genus level against Greengenes were further investigated with the RDP classifier against the RDP database (version September 30, 2016) [@bib24] using a minimum bootstrap cutoff of 50% [@bib25].\n\n2.8. Statistical analysis {#sec2.8}\n-------------------------\n\nTwo-way ANOVA with a Student-Newman-Keuls was used to assign significance to the comparisons between groups (Sigmaplot, USA). The significance of the differences between time points was calculated using two-way repeated measures ANOVA with a Student-Newman-Keuls post hoc test (Sigmaplot, USA). Data are expressed as mean\u202f\u00b1\u202fSEM. All results were considered statistically significant at *P*\u202f\\<\u202f0.05.\n\nIn order to illustrate \u03b2-diversity of metagenomic samples, weighted UniFrac distance matrix was calculated at the genus level based on taxa having at least 1% of total relative abundance. PCoA (Principal Coordinates Analysis) was performed on the resulting distance matrix using the \\'phyloseq\\' R package (version 1.16.2). The statistical significance of differentially abundant and biologically relevant taxonomical biomarkers between two distinct biological conditions was measured using a linear discriminant analysis (LDA) effect size (LEfSe) [@bib26]. Only taxa meeting an LDA significant threshold of 2.5 were considered. A *P*-value \\< 0.05 was considered to indicate statistical significance for the factorial Kruskal--Wallis rank-sum test.\n\n3. Results {#sec3}\n==========\n\nCE administration throughout 8 weeks did not reverse body weight gain in both Chow- and HFHS-fed mice ([Figure\u00a01](#fig1){ref-type=\"fig\"}A). Accordingly, we found similar energy intake and fat mass accumulation when comparing Chow *versus* Chow\u202f+\u202fCE and HFHS *versus* HFHS\u202f+\u202fCE ([Figure\u00a01](#fig1){ref-type=\"fig\"}B,C). Interestingly, the livers of CE-treated HFHS-fed mice tended to be lighter than those of untreated HFHS-fed mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}A) and, during necropsies, we noted that the livers of untreated HFHS-fed mice were pale and clearly steatotic whereas those of HFHS\u202f+\u202fCE mice displayed a reddish healthy aspect ([Figure\u00a02](#fig2){ref-type=\"fig\"}C, top panel). Quantification of liver triglycerides and ORO staining revealed massive triglyceride accumulation in the livers of HFHS mice, which was fully reversed in HFHS\u202f+\u202fCE mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}B--D). CE administration did not reduce triglyceride levels in the liver of healthy Chow-fed mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}B--D) and fasting plasma triglycerides were not affected by diet or treatment ([Figure\u00a02](#fig2){ref-type=\"fig\"}E).Figure\u00a01**Impact of CE on body features of Chow- and HFHS-fed mice.** Mice were pre-fed a standard Chow diet or a high fat/high sucrose diet throughout 13 weeks and treated either with a cranberry extract (CE) or the vehicle for 8 additional weeks. (A) Weight gain and final body weight; (B) energy intake; (C) weight of visceral and subcutaneous fat pads. Two-way repeated measures RM-ANOVA with a Student-Newman-Keuls post hoc test was used to assign significance to the differences between time points within different groups. Two-way ANOVA with a Student-Newman-Keuls post hoc test was applied to calculate the significance of the differences between groups. Data are expressed as the mean\u202f\u00b1\u202fSEM; n = 8--11; \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001.Figure\u00a01Figure\u00a02**CE reverses hepatic steatosis and alleviates liver inflammation.** (A) Liver weight, (B) hepatic triglyceride accumulation, (C) representative images of hepatic lipid accumulation by oil red O (ORO) staining, (D) quantification of ORO-positive area, and (E) plasma triglycerides. (F) Hepatic quantification of \\[MDA\\] malondialdehyde, \\[SOD\\] superoxide dismutase, \\[GPx\\] glutathione peroxidase, and catalase. (G) Liver mRNA expression of \\[COX2\\] cyclooxygenase 2, \\[TNF\u03b1\\] tumor necrosis factor \u03b1, \\[NF\u03baB\\], nuclear factor \u03ba-light-chain-enhancer of activated B cells, \\[I\u03baB\\] NF\u03baB inhibitor. (H) Liver mRNA expression of \\[PPAR\u03b1/\u03b3\\] peroxisome proliferator-activated receptor \u03b1 and \u03b3, \\[SREBP1c/2\\] sterol regulatory element-binding protein 1c and 2 and \\[LXR\u03b1/\u03b2\\] liver X receptor \u03b1 and \u03b2. Two-way ANOVA with a Student-Newman-Keuls post hoc test was applied to calculate the significance of the differences between groups. Data are expressed as the mean\u202f\u00b1\u202fSEM; n = 8--11; \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001.Figure\u00a02\n\nLiver MDA levels were reduced in HFHS\u202f+\u202fCE mice when compared with vehicle-treated HFHS mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}F), suggesting lower lipid peroxidation. Conversely, levels of superoxide dismutase (SOD, SOD2), glutathione peroxidase (GPx), and catalase, all important constituents of the cell\\'s anti-oxidant defenses, were not affected by diet, treatment, or the interaction of both ([Figure\u00a02](#fig2){ref-type=\"fig\"}F). Interestingly, CE administration reversed the HFHS-induced mRNA overexpression of cyclooxygenase-2 (COX2), tumor necrosis factor-\u03b1 (TNF\u03b1), nuclear factor \u03ba-light-chain-enhancer of activated B cells (NF\u03baB) and NF\u03baB inhibitor (I\u03baB), suggesting a broad resolution of diet-induced hepatic inflammation in HFHS-fed CE-treated mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}G). Consistent with improved hepatic steatosis we found higher mRNA levels of peroxisome proliferator-activated receptor \u03b1 (PPAR\u03b1) and lower levels of both sterol regulatory element-binding protein 1 and 2 (SREBP1/2) transcripts in the livers of HFHS\u202f+\u202fCE mice *versus* HFHS mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}H). Unexpectedly, the mRNA expression of PPAR\u03b3 and liver X receptor \u03b1 and \u03b2 (LXR\u03b1/\u03b2), all nuclear factors linked to lipid anabolism, was upregulated in HFHS\u202f+\u202fCE in comparison with vehicle-treated HFHS-fed mice ([Figure\u00a02](#fig2){ref-type=\"fig\"}H).\n\nWe then sought to investigate whether glucose homeostasis and insulin sensitivity were affected by CE administration in diet-induced obese mice. We found lower glycemia 10 and 15\u202fmin after insulin injection ([Figure\u00a03](#fig3){ref-type=\"fig\"}A) and lower area under the ipITT curves ([Figure\u00a03](#fig3){ref-type=\"fig\"}B) in CE-treated *versus* vehicle-treated HFHS-fed mice. Despite lower glycemia 5\u202fmin after insulin injection in Chow\u202f+\u202fCE as compared with Chow mice ([Figure\u00a03](#fig3){ref-type=\"fig\"}A), the overall insulin response was not different between these two groups as suggested by similar area under the ipITT curves ([Figure\u00a03](#fig3){ref-type=\"fig\"}B). We found lower glucose excursions 30, 90, and 120\u202fmin after oral glucose challenge ([Figure\u00a03](#fig3){ref-type=\"fig\"}C) and reduced area under the OGTT curves ([Figure\u00a03](#fig3){ref-type=\"fig\"}D) in HFHS\u202f+\u202fCE *versus* HFHS mice. Importantly, determination of insulinemia during OGTT revealed that improved glucose tolerance in HFHS\u202f+\u202fCE mice was achieved despite lower basal and 15\u202fmin insulin levels post-glucose challenge ([Figure\u00a03](#fig3){ref-type=\"fig\"}E), which are in agreement with markedly improved insulin sensitivity in these animals. Glucose tolerance and insulinemia during OGTT were not different between CE-treated and vehicle treated mice on the Chow diet ([Figure\u00a03](#fig3){ref-type=\"fig\"}C--E).Figure\u00a03**CE improves glucose homeostasis and insulin sensitivity in diet-induced obese mice.** At week 17, mice were fasted for 6\u202fh and (A, B) insulin tolerance tests (ipITT) were carried out after intraperitoneal insulin injections (ipITT, 0.65\u202fIU/kg). At week 19, mice were fasted overnight (12\u202fh) and submitted to (C, D) oral glucose tolerance tests (OGTT). (E) Blood was collected during OGTT and used to assess insulinemia after glucose challenge. (A, C, E) Two-way repeated measures ANOVA with a Student-Newman-Keuls post hoc test was used to assign significance to the differences between time points within groups. \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001 for Chow *vs* HFHS; ^\\#^*P*\u202f\\<\u202f0.05, ^\\#\\#^*P*\u202f\\<\u202f0.01; ^\\#\\#\\#^*P*\u202f\\<\u202f0.001 for HFHS *vs* HFHS\u202f+\u202fCE; ^&^*P*\u202f\\<\u202f0.05 for Chow *vs* Chow\u202f+\u202fCE. (B, D, F) Two-way ANOVA with a Student-Newman-Keuls post hoc test was applied to calculate the significance of the differences between groups; n = 8--11; \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001. Data are expressed as the mean\u202f\u00b1\u202fSEM.Figure\u00a03\n\nFecal DNA was extracted and 16S rRNA-based microbial profiling was performed in order to investigate whether the phenotypic traits of Chow, Chow\u202f+\u202fCE, HFHS, and HFHS\u202f+\u202fCE were associated with changes in gut microbial community structure. \u03b2-diversity was generally assessed by means of principal component analysis (PCoA) on weighted unifrac distances and revealed a clear diet-induced separation in the microbial composition of Chow- and HFHS-fed mice (PCo1, 57,1%) ([Figure\u00a04](#fig4){ref-type=\"fig\"}A). Treatment also importantly influenced the gut microbiota (PCo2, 21,3%) and separated vehicle- and CE-treated microbial communities of both Chow- and HFHS-fed mice ([Figure\u00a04](#fig4){ref-type=\"fig\"}A). \u03b2-diversity changes between HFHS and HFHS\u202f+\u202fCE were accompanied by a drop in the Firmicutes to Bacteroidetes ratio in HFHS\u202f+\u202fCE mice *versus* vehicle-treated HFHS-fed mice ([Figure\u00a04](#fig4){ref-type=\"fig\"}B). LEfSe analysis disclosed that obesity-driven dysbiosis was mostly explained by a reduction in the populations of *Barnesiella*, *Bifidobacterium*, *Turicibacter*, *Anaerostipes*, and *Clostridium* and an expansion of *Peptostreptococcaceae*, Clostridiales, *Oscillospira*, *Oscillibacter*, *Clostridiaceae*, and *Anaerotruncus* ([Figure\u00a04](#fig4){ref-type=\"fig\"}C). Administration of CE to Chow-fed mice was associated with an increase of Clostridiales, *Lachnospiraceae* and *A*. *muciniphila* ([Figure\u00a04](#fig4){ref-type=\"fig\"}D and [Supplemental Figure\u00a01](#appsec2){ref-type=\"sec\"}), whereas in HFHS-fed mice CE-treatment was related to expansion of *A.\u00a0muciniphila*, *Coprobacillus*, and *Barnesiella* ([Figure\u00a04](#fig4){ref-type=\"fig\"}E and [Supplemental Figure\u00a01](#appsec2){ref-type=\"sec\"}).Figure\u00a04**CE administration alters the taxonomic profile of Chow- and HFHS-fed mice.** Genomic DNA was extracted from feces collected at week 21 and subsequent 16S rRNA-based gut microbial profiling was performed. Feces from mice housed in the same cage were pooled and considered as one biological sample (Chow n\u202f=\u202f3; Chow\u202f+\u202fCE n\u202f=\u202f4; HFHS n\u202f=\u202f3 and HFHS\u202f+\u202fCE n\u202f=\u202f4). (A) \u03b2-diversity among groups was initially observed by means of principal component analysis (PCoA) on weighted unifrac distances, and the (B) Firmicutes to Bacteroidetes ratio was calculated as a general index of obesity-driven dysbiosis. Linear discriminant analysis (LDA) effect size (LEfSe) was calculated in order to explore the taxa that more strongly discriminate between the gut microbiota of (C) Chow *vs.* HFHS, (D) Chow *vs* Chow\u202f+\u202fCE and (E) HFHS *vs* HFHS\u202f+\u202fCE. (B) Two-way ANOVA with a Student-Newman-Keuls post hoc test was applied to calculate the significance of the differences between groups. \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001.Figure\u00a04\n\nBecause the presence of *A.\u00a0muciniphila* in the gut microbiota has been linked to improved intestinal barrier and mucus layer integrity [@bib27], [@bib28], we assessed fecal mucin as a readout of mucus layer thickness. We found a reduction in fecal mucin in HFHS-fed mice as compared to Chow-fed mice, which tended to be reversed in HFHS\u202f+\u202fCE (*P*\u202f=\u202f0.06, HFHS *vs* HFHS\u202f+\u202fCE; two-way ANOVA with Student-Newman-Keuls post hoc test) ([Figure\u00a05](#fig5){ref-type=\"fig\"}). Regression analysis revealed that the abundance of *A.\u00a0muciniphila* in the gut microbiota explains 91% of the increase in fecal mucin and 82% of the decrease in triglyceride deposition in the liver HFHS\u202f+\u202fCE mice ([Suppl.\u00a0Figure\u00a02](#appsec2){ref-type=\"sec\"}).Figure\u00a05**Fecal mucin quantification.** Cecal contents were collected at week 21, snap-frozen in liquid nitrogen and stored at\u00a0\u221280\u202f\u00b0C. Cecal feces were freeze-powdered and the presence of mucins was determined using a fluorometric assay kit that discriminates O-linked glycoproteins (mucins) from N-linked glycoproteins. Two-way ANOVA with a Student-Newman-Keuls post hoc test was applied to calculate the significance of the differences between groups. \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001.Figure\u00a05\n\n4. Discussion {#sec4}\n=============\n\nWe and others have previously shown that concomitant and sustained treatment of diet-induced obese mice with polyphenol-rich extracts protected against obesity and also resulted in reduced hepatic steatosis, alleviated intestinal inflammation and a drastic expansion of *A.\u00a0muciniphila* in the gut microbiota [@bib12], [@bib15]. In these studies, however, it was not possible to determine to what extent the preventive effects of such extracts were secondary to lower body weight gain and reduced fat mass accretion. In the present study, we now document that CE can protect against two major metabolic complications of obesity, insulin resistance, and NAFLD, independently from changes in body weight or adiposity.\n\nThe marked reduction of fat deposition in the livers of CE-treated mice was accompanied by reduced hepatic inflammation, as suggested by downregulation of COX2, TNF\u03b1, NF\u03baB and I\u03baB mRNA expression. Conversely, enzymatic components of the cellular antioxidant machinery (*i.e.* SOD1, SOD2, GPx, catalase) were unaltered in the livers of HFHS\u202f+\u202fCE mice, and similar results were found in *db*/*db* mice supplemented with a cranberry powder [@bib29]. Importantly, lower hepatic levels of MDA indicated that lipid peroxidation is decreased in HFHS\u202f+\u202fCE mice. This is possibly explained by the lower availability of triglycerides in the livers of HFHS-fed CE-treated mice. In addition, CE may counter ROS-induced lipid peroxidation in the liver by alleviating inflammation and thereby toning down ROS formation. Since CE did not alter the cells\\' antioxidant defense in the liver, it is also plausible that CE polyphenols directly neutralize ROS, which would further contribute to reduce lipid peroxidation.\n\nConsistent with augmented lipid catabolism in the liver, we found higher mRNA levels of PPAR\u03b1 and lower amount of mRNA transcripts of SREBP1 and SREBP2 in HFHS\u202f+\u202fCE mice when compared with untreated HFHS-fed mice. However, in apparent contradiction with lower hepatic steatosis, CE treatment was associated with higher mRNA expression of PPAR\u03b3 and upregulation of LXR\u03b1 and LXR\u03b2 mRNA in HFHS-fed mice. PPAR\u03b1 is a key nuclear receptor to steatogenesis, being highly expressed in the liver and the principal activator of PPAR-responsive elements (PPREs) in this organ [@bib30], [@bib31], [@bib32]. It is therefore conceivable that PPAR\u03b1-related activation of \u03b2-oxidation genes likely overcomes PPAR\u03b3-driven stimulation of lipogenesis in the liver of HFHS-fed CE treated mice. Moreover, PPAR\u03b3 and LXR are both highly expressed in Kupffer cells (liver resident macrophages), where their activation is linked to anti-inflammatory effects [@bib33], [@bib34]. Thus macrophage PPAR\u03b3 and LXR\u03b1/\u03b2 transcripts may account for the increased expression of these nuclear receptors in the livers of HFHS\u202f+\u202fCE mice and contribute to alleviate hepatic inflammation. Interestingly, since activation of hepatic PPAR\u03b3 and LXR are both linked to reduced hepatic glucose output [@bib30], [@bib35], [@bib36], our results also point to a role of these two nuclear receptors in the CE-related benefits to glucose homeostasis. Taken together, our findings suggest that CE alleviates steatogenesis by targeting PPAR\u03b1, but also by upregulating PPAR\u03b3 and LXR\u03b1/\u03b2. This mRNA expression profile is consistent with enhanced hepatic immune-metabolic status and increased lipid catabolism in the liver. It is noteworthy that the massive lipid deposition observed in the livers of HFHS-fed mice did not result in dyslipidemia in our model, as suggested by similar levels of fasting plasma triglycerides found among groups. While this is presumably linked to the fact that, as opposed to humans, mice have a low LDL/high HDL cholesterol profile, which might favor hepatic lipid storage and limit dyslipidemia in diet-induced obese mice [@bib37], this observation indicates that the marked effect of CE on hepatic triglyceride accretion is not a consequence of increased lipid mobilization to the plasma, which is in accordance with enhanced lipid oxidation being the main mechanism of reduced liver fat accumulation in HFHS\u202f+\u202fCE mice.\n\nBecause gut microbial factors are increasingly recognized as key drivers of hepatic metabolism [@bib7], [@bib8], [@bib38] and because CE has been previously linked to major gut microbial changes [@bib12], we sought to investigate intestinal bacterial profiles in CE- and vehicle-treated mice. Similarly to our previous report using a preventive approach [@bib12], we found that CE administration to already obese mice triggered a remarkable bloom of *A.\u00a0muciniphila* in the gut microbiota of HFHS-fed mice. This effect of CE treatment was also observed, albeit to a lesser extent, in Chow-fed mice. *A.\u00a0muciniphila* is a Gram-negative mucin degrading bacterium strongly correlated with healthy and lean phenotypes [@bib5], [@bib39], and its administration as a probiotic was shown to reverse diet-induced obesity in mice [@bib27], [@bib40]. Previous studies have demonstrated a particular association between higher *A.\u00a0muciniphila* and better glycemic control, which was independent of reduced visceral fat mass deposition [@bib41], [@bib42]. These findings suggest that the interaction between *A.\u00a0muciniphila* and host metabolism is context-specific, being likely dependent on factors such as gut community structure, the severity of the dysmetabolic condition and the host\\'s genetic background. Indeed, Shin et\u00a0al. showed that the metformin-like effects of *A.\u00a0muciniphila* on glucose tolerance were only achieved with doses higher than 4\u202f\u00d7\u202f10^7^\u202fCFU in diet-induced obese mice [@bib28]. Our results revealed that, upon a 21-week long HFHS regimen, the CE-related increase in *A.\u00a0muciniphila* population was not linked to lower fat accumulation, but it was associated with improved glucose metabolism and alleviated hepatic steatosis.\n\nDespite the fact that LEfSe analysis did not classify *A.\u00a0muciniphila* as a key-phylotype of Chow\u202f+\u202fCE mice, we found a small yet significant increase in this taxon in Chow-fed CE-treated mice *versus* vehicle-treated Chow-fed mice by applying a distinct statistical approach. Because Chow-fed mice are metabolically healthy, this finding was not associated with major changes in glucose homeostasis or weight gain. However, this result is of great relevance as it suggests that CE may still favor an expansion of *A.\u00a0muciniphila* in healthy individuals, which might be protective in the long-term. Moreover, *Lachnospiraceae* and Clostridiales were both ranked as important discriminative taxa between Chow\u202f+\u202fCE and vehicle-treated Chow-fed mice. This may be related to the capacity of certain species within the family *Lachnospiraceae* and the order Clostridiales to resist to the antimicrobial effect of CE and/or utilize CE-polyphenols as substrates. More studies are warranted to further explore the gut microbial-related protective role of CE polyphenols in healthy mice.\n\nCranberry polyphenols have been shown to improve mucus layer and villi morphology in mice receiving elemental enteral nutrition [@bib43]. Moreover, dietary polyphenols increased the amount of mucin in the feces of high fat-fed mice [@bib44]. We therefore hypothesized that CE may create a favorable environment for *A.\u00a0muciniphila* to thrive by boosting mucus secretion [@bib12], [@bib45]. In accordance with this hypothesis, we found a strong trend (*P*\u202f=\u202f0.06) towards higher mucin concentration in the cecum of CE-treated HFHS-fed mice and that *A.\u00a0muciniphila* is a strong predictor of fecal mucin and liver triglyceride accumulation. Consistently, enhanced mucus layer thickness was previously observed in DIO-mice treated with live *A.\u00a0muciniphila* [@bib27]. Our data suggest that the benefits of CE to gut barrier and hepatic homeostasis are tightly linked to expansion of *A.\u00a0muciniphila* in the gut microbiota.\n\nPolyphenols possess antimicrobial activity [@bib46] and, because Gram-negative bacteria are generally more resistant to this effect [@bib47], *A.\u00a0muciniphila* may find a competitive advantage in the gut environment of CE-treated mice. Similar mechanism might favor the presence of *Barnesiella spp* in the gut microbiota of HFHS\u202f+\u202fCE mice. While little is known about the relevance of *Barnesiella spp* to the host, our data stress its beneficial impact on host metabolism as they classified this taxon as the main discriminative feature of Chow-fed mice when compared with HFHS-fed mice. Moreover, it has been suggested that *Barnesiella spp*. confer resistance to intestinal growth and bloodstream infection with vancomycin-resistant *Enterococcus* [@bib48], which supports a beneficial interaction between host and *Barnesiella*. *Coprobacillus* was ranked a key-feature of HFHS\u202f+\u202fCE mice; however, its abundance is much lower than 1% and its relevance to host metabolism might be minor in this study.\n\nLower Firmicutes to Bacteroidetes ratio is often considered as a key feature of the \"obese gut microbiota\" [@bib49], [@bib50]. In our model, however, we found lower Firmicutes to Bacteroidetes ratio in association with improved glucose and insulin tolerance in HFHS\u202f+\u202fCE mice, traits that were unrelated with changes in body weight and fat mass accumulation. While suggesting that CE-driven gut microbial remodeling is primarily linked to hepatic homeostasis and the regulation of glucose metabolism/insulin sensitivity, our findings also indicate that lower Firmicutes to Bacteroidetes ratio seems to influence host glucose homeostasis prior to affect host fat mass accretion. Indeed, lower Firmicutes to Bacteroidetes ratio has been particularly associated with better glycemic control in humans [@bib41], and reconstitution of germ-free mice with the fecal slurry of mice fed on a high-fat diet for 10 weeks rendered them glucose intolerant but not obese [@bib51]. Future studies are warranted to determine whether CE administration for a longer period of time, therefore more chronically exposing the host to lower Firmicutes to Bacteroidetes ratio and higher *A.\u00a0muciniphila*, would affect host fat mass storage.\n\nThe dose of CE used in this study (200\u202fmg of extract/kg, 75\u202fmg of polyphenols/kg) is equivalent to the consumption of approximately 120\u202fg of fresh cranberries/day by a 60\u202fkg individual [@bib18]. The translation of this dose to humans, however, is probably not straightforward. Considering the US Food and Drug Administration\\'s guidelines to calculate the human equivalent dose based on body surface area [@bib52] we found that a 16\u202fmg of extract/kg (6\u202fmg of polyphenols/kg) dose would be the human equivalent of a 200\u202fmg/kg dose in mice. Based on this, we estimated that obtaining 360\u202fmg of polyphenols/60\u202fkg from a cranberry juice cocktail (a very popular form of consuming cranberries) would require the daily consumption of 206.8\u202fmL of this product (considering a cocktail containing 54% of cranberry juice and 1.74\u202fmg of polyphenols/mL [@bib53]). It is, however, important to stress that while the polyphenolic composition of cranberry juices or cranberry juice cocktails may be different from that of the extract, these preparations often contain added sugars and/or artificial sweeteners [@bib53], which may contribute to metabolic disease. We believe that other sources of cranberry polyphenols, such as capsuled cranberry extract or unsweetened dried cranberries, are likely better options in order to target metabolic diseases.\n\nAs a prospect, changes in the bile acid profile may be investigated as a mechanism behind the metabolic benefits of CE. Bile acids can regulate their own synthesis and ileal re-uptake via binding to the nuclear receptor farnesoid-X-receptor (FXR). Furthermore, the role of bile acids goes beyond aiding lipid digestion and extends to regulation of glucose/lipid homeostasis and energy metabolism [@bib54]. Interestingly, while changes in bile acid profile have been associated with NASH in clinical studies [@bib55], pre-clinical analyses using mouse models of obesity have revealed that FXR agonism is protective against liver steatosis and insulin resistance [@bib56], [@bib57]. Since dietary polyphenols have been shown to bind to bile acids and alter their re-uptake (and possibly their ability to signal through FXR) [@bib58], [@bib59], it is possible that CE treatment changes the bile acid profile, which may in turn contribute to improve liver steatosis and insulin resistance. Bile acids are modified by the gut microbiota, yielding secondary bile acids. Since CE importantly affects gut microbial populations, it may also affect the microbial capacity to transform bile acids, which would further contribute to modify the bile acid profile and potentially modulate metabolic status. It is noteworthy that the ability of *A.\u00a0muciniphila* and *Barnesiella*, the two best representative taxa of HFHS\u202f+\u202fCE, to modify or resist to bile acids is currently unknown.\n\n5. Conclusions {#sec5}\n==============\n\nOur findings shed light on novel mechanisms by which CE improves metabolic health. Using a reversal experimental design allowed us to demonstrate that CE polyphenols strongly target the liver, where it modulates key hepatic nuclear factors and genes involved in the modulation of steatogenesis and inflammation, thus markedly improving liver homeostasis. This was associated with improvements in glucose tolerance and full restoration of insulin sensitivity despite the maintenance of obesity in these animals. Our data also put forward *A.\u00a0muciniphila* and *Barnesiella spp*. as potential microbial contributors to the effect of CE and that the abundance of these bacteria in the gut microbiota is not phenotypically linked to changes in obesity. Finally, while supporting the regular consumption of cranberries to help counteracting obesity-related diseases, our data provide the fundamental bases for future human trials with CE or CE-related products as a strategy against diet-induced NAFLD/NASH and metabolic syndrome.\n\nFinancial support {#sec6}\n=================\n\nThis work was funded by a Canadian Institutes of Health Research (CIHR) foundation grant (FDN\\#143247) to AM, J.D deS\u00e8ve grants to EL and AM and a Leahy Orchards Inc. & AppleActiv grant to EL.\n\nConflict of interest {#appsec1}\n====================\n\nThe authors declare no conflict of interest.\n\nAppendix A. Supplementary data {#appsec2}\n==============================\n\nThe following are the supplementary data related to this article:Supplemental Figure\u00a01**Gut microbial profile at genus level**. In order to eliminate the putative influence of baseline (week 13) gut microbiota, the \u0394 ~relative\\ abundance\\ week21-\\ relative\\ abundance\\ week13~ was calculated. Significance was calculated using Mann--Whitney *U* test with a Monte Carlo permutation test, \\**P*\u202f\\<\u202f0.05, \\*\\**P*\u202f\\<\u202f0.01 and \\*\\*\\**P*\u202f\\<\u202f0.001.Supplemental Figure\u00a01**Linear regression analysis.** How the abundance of *A.\u00a0muciniphila* in each cage explains the variability in the amount of (A) fecal mucin and (B) liver triglyceride deposition was calculated by regression analysis. The phenotype per cage (n\u202f=\u202f4) considers the average of mice housed together. r^2^ -- coefficient of determination.\n\nWe are grateful to Val\u00e9rie Dumais, Christine Dion, Christine Dallaire, and Joannie Dupont-Morissette for their expert help with animal care.\n\nSupplementary data related to this article can be found at [https://doi.org/10.1016/j.molmet.2017.10.003](10.1016/j.molmet.2017.10.003){#intref0010}.\n"}
+{"text": "Vertebrate organ development is a complex process that begins in the early embryo and continues until the functional capacity of the organ meets adult requirements. Organ growth and maturation occupy a significant period of life and involve major anatomic changes, yet relatively little is known about the cellular dynamics that underlie these changes.\n\nNew technologies can illuminate cellular mechanisms that drive organ morphogenesis. Recently, Livet and colleagues introduced a system that allowed labeling designations of \\~90 colors to murine neurons^[@R1]^. With this technology, termed 'Brainbow', they could visualize adjacent neurons and their connections in the brain with high resolution. The ability to assign many colors to different cells in a population can also be applied to understand cell proliferation and lineage decisions.\n\nThe heart is a set of chambers comprised predominantly of the contractile units, cardiomyocytes. Genetic fate-mapping has been performed to determine how separate lineages contribute to developing cardiac structures in mice and zebrafish^[@R2]--[@R6]^. Additionally, single-marker clonal analysis has traced the activity of individual cells during embryonic heart patterning^[@R7]--[@R10]^. These studies have enhanced our understanding of cardiogenic mechanisms in early embryos. Nevertheless, a large gap remains in comprehending how the size, shape, and structure of an adult heart are finalized through the individual and population behaviors of many cardiac cells.\n\nHere, we employed a multicolor clonal analysis to map the proliferative histories of many individual cardiomyocytes as the zebrafish cardiac ventricle transitions from a simple tube of single-cardiomyocyte thickness into a complex adult structure. Our experiments yielded several unexpected discoveries relevant to the number, nature, and mechanisms of cardiomyocyte contributions during heart morphogenesis.\n\nMulticolor labeling of cardiomyocytes {#S1}\n=====================================\n\nTo study cell clones in zebrafish, we adapted the Brainbow 1.0L construct for combinatorial expression of 3 spectrally different fluorescent reporter proteins^[@R1]^. Multiple copy integration of this transgene at a single genetic locus is a common outcome of transgenesis. Thus, combinatorial Cre recombinase-mediated excision events at paired *loxP* recognition sites can generate many possible permanent colors. ([Fig. 1a, b](#F1){ref-type=\"fig\"}). We generated several transgenic lines containing a *\u03b2-actin2* promoter-driven multicolor construct, and assessed them in combination with a strain harboring a tamoxifen-inducible, cardiomyocyte-restricted Cre recombinase, *cmlc2:CreER*^[@R11]^. We identified one line, *Tg(\u03b2-act2:Brainbow1.0L)*^pd49^ (referred to subsequently as *priZm*), showing limited mosaic recombination only in the presence of 4-hydroxytamoxifen (4-HT; [Supplementary Fig. 1](#SD1){ref-type=\"supplementary-material\"}).\n\nThe size and structure of the zebrafish heart is conducive to clonal analysis of cardiomyocytes. The 2--3 dpf zebrafish heart is looped, has a wall of single-cardiomyocyte thickness, and consists of 250--300 muscle cells; by our measurements \\~115 of these are contained within the ventricular wall ([Supplementary Fig. 2](#SD1){ref-type=\"supplementary-material\"}). Cardiomyocyte proliferation is detectable at 2 dpf and thought to account for most or all subsequent cardiogenesis^[@R2],[@R12]--[@R15]^. To trace the fates of individual embryonic cardiomyocytes, we briefly incubated 2 dpf *cmlc2:CreER; priZm* embryos in 4-HT and raised them to different ages ([Fig. 1c](#F1){ref-type=\"fig\"}).\n\nWe first assessed cardiac fluorescence at 10 dpf, a stage comparable in organismal size to 2 dpf. Our optimized 4-HT regimen generated \\>20 unique colors in cardiomyocytes, as viewed at 10 dpf and subsequent stages. Red fluorescent protein (RFP) is the initial reporter cassette and default expression marker, and 4-HT treatment induced recombination (non-red colors) in \\~50% of ventricular cardiomyocytes. Importantly, different colors were consistently assigned to adjacent cardiomyocytes on the surface of the 10 dpf ventricular wall ([Fig. 1d--f](#F1){ref-type=\"fig\"}; [Supplementary Fig. 3](#SD1){ref-type=\"supplementary-material\"}), a prerequisite for multicolor clonal analysis.\n\nThe zebrafish ventricle is recognized to contain two types of cardiac muscle^[@R16]^. These include a peripheral wall of compact muscle, and inner myofibers organized into trabeculae that initiate formation at 3 dpf^[@R17]^. Histological examination of 10 dpf hearts revealed three notable observations. First, the ventricular wall remained at single-cardiomyocyte thickness, as at 2 dpf. Second, trabecular myocytes connected to the wall were most often clonally unrelated to adjoining wall myocytes (58 of 63 observations, n = 5 ventricles; [Fig. 1e, f](#F1){ref-type=\"fig\"}). This observation supported a mechanism for trabecularization proposed recently based on different lines of evidence, in which myocytes delaminate from the ventricular wall, seed elsewhere in the chamber, and initiate trabecular growth from a second site^[@R17]^. Third, we saw that the trabecular myofibers themselves were comprised of cardiomyocytes arising from different clonal origins within the 2 dpf wall ([Fig. 1f](#F1){ref-type=\"fig\"}).\n\nJuvenile ventricular wall formation {#S2}\n===================================\n\nWe next examined the ventricles of juvenile zebrafish at 30 dpf, at which time major organismal and cardiac growth have occurred since cardiomyocyte labeling with 4-HT. By whole-mount imaging of surface myocardium, we readily identified multicellular, single-color regions of myocardium, indicating that progeny of embryonic cardiomyocytes generally remained connected with one another after division ([Fig. 2a--c](#F2){ref-type=\"fig\"}; [Supplementary Fig. 4a--c](#SD1){ref-type=\"supplementary-material\"}). Although less common (9.8% of clones), we also observed instances suggesting complete separation of a cardiomyocyte from its clonal partners ([Supplementary Fig. 4d](#SD1){ref-type=\"supplementary-material\"}).\n\nWe expected some uniformity to the clonal patches comprising the ventricular wall, with cardiomyocyte clones of similar shape and size. By contrast, ventricles from different animals presented unique patterns of surface color clones, and the shapes of clones within each ventricle were highly diverse ([Fig. 2a](#F2){ref-type=\"fig\"}; [Supplementary Fig. 5](#SD1){ref-type=\"supplementary-material\"}). Clone size also varied. Some clones displayed evidence of many cell divisions, while others contained a single cardiomyocyte ([Fig. 2d](#F2){ref-type=\"fig\"}). Histological analysis showed retention of a wall of single-cardiomyocyte thickness, indicating that the expansion of wall clones was limited to lateral directions along the surface. They also revealed that substantial expansion of trabecular myocardium had occurred since 10 dpf ([Fig. 2e](#F2){ref-type=\"fig\"}).\n\nTo estimate the number of embryonic cardiomyocytes that create the juvenile ventricular wall, we calculated the surface areas occupied by individual clones and divided each by the total ventricular surface area. Most clones (99/146; n = 5) each occupied less than 2% of the ventricular surface area, while a small number (15/146) of larger clones each represented \\~4--8% of the surface ([Fig. 2f](#F2){ref-type=\"fig\"}). Extrapolating for unrecombined myocardium, we determined that the 30 dpf ventricular surface was represented by 55.4 \u00b1 1.9 color clones (n = 10; mean \u00b1 SEM; [Fig. 2g](#F2){ref-type=\"fig\"}). Thus, our data indicate that the juvenile zebrafish ventricular wall is built by lateral expansion of \\~55 embryonic cardiomyocytes. These cardiogenic events create a patchwork of diverse clonal shapes and sizes that varies from animal to animal, indicating that juvenile cardiac form can be acquired with significant developmental plasticity.\n\nEmergence of a new adult muscle lineage {#S3}\n=======================================\n\nZebrafish are typically recognized as adults at 3 months post-fertilization. We examined ventricles of 6, 8, and 10 week-old animals that had undergone cardiomyocyte labeling at 2 dpf. At 6 weeks post-fertilization (wpf), most clonal patches comprising the surface myocardium appeared similar to those of 30 dpf ventricles. Additionally, we detected a population of clonally related cardiomyocytes layered upon these patches near the chamber base ([Fig. 3a, b](#F3){ref-type=\"fig\"}). By 8 wpf, large single-color swaths containing several hundreds of such cardiomyocytes extended from the ventricular base, wrapping around both sides of the chamber and often reaching its midpoint ([Fig. 3d--f](#F3){ref-type=\"fig\"}). At 10 wpf, we began to see evidence of these surface clones converging with each other. These external cardiomyocytes were typically more rod-shaped with more distinct striation than underlying cardiomyocytes ([Supplementary Fig. 6](#SD1){ref-type=\"supplementary-material\"}).\n\nWe examined histological sections and confocal slices through ventricles from 6--10 wpf animals, which confirmed that a new layer of ventricular muscle had emerged externally to the wall of single-cardiomyocyte thickness present at earlier stages ([Fig. 3c, g](#F3){ref-type=\"fig\"}). As indicated by whole-mount imaging, this external layer typically displayed substantial regions of clonally related cardiomyocytes. We will refer subsequently to the inner wall muscle as the 'primordial layer', as it retains the same single-cardiomyocyte thickness and characteristics of the embryonic ventricle throughout subsequent life stages. We will refer to the late-emerging, outermost muscle of the ventricular wall as the 'cortical layer'. Thus, by multicolor clonal analysis, we revealed two developmentally distinct and previously unrecognized forms of ventricular wall myocardium.\n\nDominant clones build adult wall muscle {#S4}\n=======================================\n\nWe next examined ventricles from mature, 90 dpf animals that had undergone cardiomyocyte labeling at 2 dpf. At this stage, the entire surface of the ventricle in each animal was covered by cortical cardiomyocytes ([Fig. 3h](#F3){ref-type=\"fig\"}; [Supplementary Fig. 7](#SD1){ref-type=\"supplementary-material\"}). Similar to the clonal representation of the primordial layer at earlier stages, cortical clone patterns appeared different in each animal. However, a common theme for each ventricle was the presence of one or two large cortical clones at the base of the heart extending across the ventricular surface toward the apex. We infer from our timepoint analysis that these areas of cortex formed by a base-to-apex wave of expansion over the primordial layer. A number of smaller surface clones were apparent at other locations. Clones converged by weaving or fitting with some overlay into each other, and were penetrated by coronary vessels ([Supplementary Fig. 8](#SD1){ref-type=\"supplementary-material\"}).\n\nHistological analysis of adult ventricular tissue indicated that the primordial layer remained at single-cardiomyocyte thickness and was comprised of multiple color clones. This appearance contrasted with the overlying cortical myocardium, which was predominantly single-color and could be several cells thick ([Fig. 3i](#F3){ref-type=\"fig\"}). We calculated the areas of surface clones from whole-mount images of ventricles that displayed basal clones of a recombined color. Ten of 20 ventricles met this criterion, each of which had a large cortical clone covering 30--60% of the total ventricular surface ([Fig. 3j](#F3){ref-type=\"fig\"}). An average of 7.8 \u00b1 1.2 clones contributed the entire cortical muscle of each ventricle (n = 10; [Fig. 3k](#F3){ref-type=\"fig\"}), a number several times lower than that clonal surface representation of the much smaller juvenile ventricle. Thus, a rare group of \\~8 clonally dominant cardiomyocytes in the embryonic ventricle ultimately contribute to building the adult cortical myocardium.\n\nMuscle lineage regeneration after injury {#S5}\n========================================\n\nZebrafish possess a robust capacity for heart regeneration throughout life^[@R18]^, based on the ability to activate the proliferation of spared cardiomyocytes after injury^[@R11],[@R19],[@R20]^. We examined the regenerative potentials of primordial and cortical muscle by amputating ventricular apices from 90 dpf *cmlc2:CreER; priZm* animals that had been labeled at 2 dpf. At 14 days post-amputation (dpa), we detected growth of adjacent cortical muscle clones in lateral and radial directions into the wound area ([Fig. 4a, b](#F4){ref-type=\"fig\"}), but the primordial layer lagged behind the amputation site. By 30 dpa, as the wall was reconstructed with clonal patches of cortical cardiomyocytes, clones of single-cardiomyocyte thickness contiguous with the primordial layer first became detectable in the regenerate ([Fig. 4c](#F4){ref-type=\"fig\"}). By 60 dpa, the primordial layer was largely restored as a complete structure positioned between cortical and trabecular muscle ([Fig. 4d](#F4){ref-type=\"fig\"}). Thus, regenerating primordial muscle undergoes restricted lateral expansion as during morphogenesis, while cortical muscle regeneration is less constrained and assumes the primary component of the new wall. Interestingly, these events occur in a temporally reversed manner compared to initial morphogenesis, with the underlying primordial layer regenerating subsequent to cortical muscle.\n\nOrigins of dominant cardiomyocyte clones {#S6}\n========================================\n\nTo determine when cortical cardiomyocytes originate during heart morphogenesis, we initiated color labeling at 30 dpf, after formation of juvenile structure but 1--2 weeks prior to emergence of cortical muscle. Strikingly, 90 dpf ventricles that had been incubated briefly with 4-HT at 30 dpf displayed surface representation by large clonal patches ([Fig. 5a](#F5){ref-type=\"fig\"} and [Supplementary Fig. 9](#SD1){ref-type=\"supplementary-material\"}), with largest clones present at the chamber base. We quantified the size and number of clones in ventricles with basal clones of recombined colors (11 of 20 animals). Our data indicated that adult cortical myocardium arises in patches of diverse sizes from an average of 8.6 \u00b1 0.7 labeled 30 dpf cardiomyocytes ([Fig. 5b, c](#F5){ref-type=\"fig\"}), a number similar to that observed after labeling at 2 dpf.\n\nIn some ventricles assessed 4 weeks after labeling (8 wpf), we observed trabecular muscle of the same color near the emergent cortical clone ([Fig. 5e, g](#F5){ref-type=\"fig\"}). To confirm this association, we examined sections of 6--7 wpf ventricles (labeled at 2 dpf), and consistently found nearby trabecular muscle of the same color as the small, basal cortical clone (13 of 13 ventricles; see cyan clone in [Fig. 3c](#F3){ref-type=\"fig\"}). We also could identify cases from these examples in which trabecular and cortical muscle of one color connected through an apparent breach in the primordial layer ([Supplementary Fig. 10](#SD1){ref-type=\"supplementary-material\"}). These observations suggested a clonal relationship between trabecular and cortical cardiomyocytes in the maturing zebrafish ventricle.\n\nWe noticed lower color recombination in the primordial layer (\\~15%) than in the trabecular and cortical muscle lineages (\\~61%) at 8 wpf ([Fig. 5d--g](#F5){ref-type=\"fig\"}). Taking advantage of this differential labeling efficiency, we titrated 4-HT and identified a low dose that induced sparse recombination at 30 dpf in trabeculae and no obvious recombination in primordial muscle. Importantly, even with very limited labeling, cortical clones of a recombined color were still discernable in 3 of 24 ventricles at 8 wpf ([Fig. 5h, i](#F5){ref-type=\"fig\"}), a finding that reaffirmed a trabecular source for the cortical lineage. As with other samples mentioned above, images of confocal slices from these ventricles could identify single-color clones containing both cortical and trabecular muscle, connecting through an apparent breach in primordial muscle of unrecombined color ([Fig. 5i--k](#F5){ref-type=\"fig\"}).\n\nIn total, our findings indicate a dynamic mechanism that generates a final ventricular muscle lineage and completes adult cardiac structure. Trabecular cardiomyocytes penetrate the primordial layer in rare, spatially segregated events at the juvenile stage, seed the ventricular surface, and undergo expansion to create the cortical myocardium.\n\nDiscussion {#S7}\n==========\n\nHere, we applied a multicolor clonal analysis and discovered unsuspected cellular mechanisms guiding heart morphogenesis in zebrafish. We found that three ventricular myocardial lineages are present in the adult form: primordial, trabecular, and cortical muscle, created in order from each other. Our data indicate that the innermost trabecular lineage is initiated predominantly by delamination from the embryonic primordial layer, as proposed in an earlier study^[@R17]^. Next, the outermost cortical layer is created as juveniles mature to adults, emerging wholly or in part via an 'inside-out' mechanism from trabeculae. Here, one or more particularly expansive juvenile cardiomyocytes access the ventricular surface at each of only \\~8 sites per animal.\n\nZebrafish share several aspects of cardiac development with higher vertebrates, and the multicolor results we report here indicate some basic commonalities with results of single-marker clonal analysis performed previously in early mouse and chick embryos. For instance, cardiomyocytes were also observed in coherent clonal populations during cardiac growth in these systems, after an early phase of dispersion^[@R9],[@R10]^. Most unpredicted is that the zebrafish ventricle maintains a primordial layer of single cardiomyocyte thickness without noticeable cell division in the z-plane, and that wall thickening instead occurs after seeding of a separate cortical lineage by dominant clones. It will be interesting to determine the evolutionary distribution of aspects of this uncovered morphogenetic mechanism within other bony fish and among other classes of vertebrates like mammals.\n\nThe origins and diversity of clonal patterns of cortical muscle we report here are more consistent with a stochastic model of source cell selection than a hierarchical or predetermined model. Additionally, the observation of clonal dominance in cardiomyocytes is reminiscent of stem cell compartments that drift toward clonality during homeostatic tissue maintenance, events that have been explained by stochastic models^[@R21]--[@R24]^. We suspect that molecular and/or physiologic cues enable the emergence and proliferative capacity of cortical cardiomyocytes, ostensibly with preferential influence at the ventricular base. It is likely that clonal dominance behavior is a recurring mechanistic strategy to help shape vertebrate organs.\n\nMETHODS SUMMARY {#S8}\n===============\n\nWild-type or transgenic zebrafish of the EK/AB background were used for all experiments and maintained at 3 fish/liter. Ventricular resection surgeries were performed as described previously^[@R18]^. All published transgenic strains used here are listed in the separate Methods section and were analyzed as hemizygotes. *Tg(\u03b2-act2:Brainbow1.0L)^pd49^* was generated and 4-HT labeling experiments were performed as described in Methods. Hearts were extracted at the indicated timepoints and fixed in 4% paraformaldehyde. After rinsing, the atrium was removed and the ventricle was placed on a cover slip in Fluoromount G. Another cover slip was used to gently compress the ventricle, allowing imaging of both ventricular surfaces. For surface images of whole-mounted ventricles, the z-position was adjusted until only surface muscle was visible. For confocal slices through whole-mounted hearts, the z-position was adjusted through the ventricle until trabecular muscle could be visualized. For histological analysis, 50 \u03bcm cryosections were mounted with Fluoromount G. Images from all multicolor samples were acquired using a Leica SP5 AOBS microscope equipped with X20 (0.7 NA) and X40 (1.25 NA) objectives. Antibodies were not used to enhance fluorescence. 458 nm, 515 nm, and 561 nm lasers were used to excite CFP, YFP, and RFP, respectively. Each channel was acquired sequentially and imported into ImageJ, where channels were overlaid. Uniform contrast and brightness adjustments were made using Adobe Photoshop. To visualize the entire outer surface area, images were joined using Photoshop. To quantify clone area, the size of a given clone in \u03bcm^2^ was traced using ImageJ software. The percentage area occupied by a clone was calculated by dividing its measured area by the total surface area of both sides of the ventricle.\n\nMETHODS {#S9}\n=======\n\nTransgenic lines {#S10}\n----------------\n\nPublished transgenic strains used in this study were (*Tg(cmlc2:CreER)^pd10^*)^[@R11]^, (*Tg(gata5:loxP-mCherry-STOP-loxP-nucEGFP)^pd40^*)^[@R11]^, (*Tg(cmlc2:EGFP)^f1^*)^[@R25]^, (*Tg(cmlc2:nucDsRed2)^f2^*)^[@R25]^, and *(Tg(\u03b2-actin:HRAS-EGFP)^vu119^*)^[@R26]^. Experiments with zebrafish were approved by Institutional Animal Care and Use Committee at Duke University.\n\nConstruction of *priZm* {#S11}\n-----------------------\n\nThe Brainbow 1.0L plasmid was digested with NheI and partially digested with NotI to obtain the entire cassette, and subcloned downstream of the 9.8 kb zebrafish *\u03b2-actin2* promoter from the *\u03b2-act2:RSG* construct^[@R11]^. The resulting plasmid was linearized by I-*Sce*I digestion and injected into one-cell zebrafish embryos. Red fluorescent protein (RFP) is the initial reporter cassette and default expression state from this construct. Two paired sites, *lox2272* and *loxp*, exist in the construct that enable Cre recombinase-mediated switching to expression of either Cyan (CFP) or Yellow fluorescent proteins (YFP).\n\nThirty founder lines were isolated, and *Tg(\u03b2-act2:Brainbow1.0L)^pd49^* was used for this study. The zebrafish *\u03b2-actin2* promoter does not drive expression in epicardial or endocardial cells^[@R11]^, facilitating clear visualization of *priZm* cardiomyocytes.\n\n4-HT labeling {#S12}\n-------------\n\nFor 4-HT labeling of *cmlc2:CreER; priZm* embryos, 2 dpf embryos were placed in egg water with 4-HT added to a final concentration of 4 \u03bcM, from a 1 mM stock made in 100% ethanol. Embryos were treated for 6 hours, rinsed once, and placed in fresh egg water. This labeling protocol was highly reproducible and induced recombination and appearance of diverse, non-red colors on half of the ventricular surface, calculated by digital quantification of red and non-red surface areas from whole-mount images of 30 dpf ventricles (50.0%; n = 10). By contrast, increased or reduced presence of 4-HT led to higher and lower recombination frequencies, respectively, that each reduced color diversity.\n\nFor labeling of *cmlc2:CreER; priZm* juveniles at 30 dpf, animals were incubated in 1\u03bcM 4-HT in aquarium water for 3 hours. To specifically label trabecular muscle at 30 dpf, animals were incubated in 0.25 \u03bcM 4-HT for 1 hour.\n\nEmbryonic ventricular cardiomyocyte quantification and proliferation indices {#S13}\n----------------------------------------------------------------------------\n\n*cmlc2:nucDsRed2; cmlc2:EGFP* double transgenic animals were raised to 3 dpf, fixed, sectioned at 30 \u03bcm, and stained with an antibody to DsRed2. Confocal stacks of the entire heart from each embryo were taken, and the number of myocytes within the ventricular wall was determined from 3D projections generated using Imaris. To determine effects of 4-HT on cardiomyocyte proliferation, we treated 2 dpf embryos with 4 \u03bcM 4-HT for 6 hrs and fixed at 3 dpf. Ten \u03bcm sections were stained with Mef2 and PCNA as described^[@R27]^, and cardiomyocyte proliferation indices for each group were calculated from 3 ventricular sections. The numbers of Mef2^+^ and Mef2^+^/PCNA^+^ cells were manually counted with the aid of ImageJ software.\n\nExamination of effects of color combinations on cardiomyocyte clone size {#S14}\n------------------------------------------------------------------------\n\nTo examine the possibility that recombined color combinations had differential effects on cardiomyocyte proliferation, we examined color diversity in the largest and smallest cortical clones of 90 dpf ventricles that were labeled at 2 or 30 dpf. From 21 ventricles, 16 unique colors were represented in the 21 large basal clones. Seventeen unique colors were represented in 21 of the smallest clones. Individual colors were observed in both small and large clone groups, including two shades of gray (in which the clone is expressing all three fluorescent proteins). This distribution was consistent with the idea that color combinations do not have significant effects on proliferative ability.\n\nSupplementary Material {#S15}\n======================\n\nWe thank K. Kikuchi for generating *cmlc2:CreER* animals and for advice; J. Burris, A. Eastes, P. Williams, and N. Blake for zebrafish care; A. Dickson for artwork; B. Hogan and Poss lab members for comments on the manuscript; and S. Johnson and Y. Gao for imaging advice. V.G. was supported by a National Heart, Lung, and Blood Institute (NHLBI) Medical Scientist Training Program supplement. K.D.P. is an Early Career Scientist of the Howard Hughes Medical Institute. This work was supported by grants from NHLBI (HL081674) and American Heart Association to K.D.P.\n\n[Supplementary Information](#SD1){ref-type=\"supplementary-material\"} is linked to the online version of the paper at [www.nature.com/nature](www.nature.com/nature).\n\n**Author Contributions.** V.G. and K.D.P. designed experimental strategy, analyzed data, and prepared the manuscript. V.G. performed all of the experiments.\n\n**Author Information.** The authors declare no competing financial interests.\n\n![Multicolor clonal labeling of embryonic zebrafish cardiomyocytes\\\n**a,** Recombination at paired *lox2272* (black triangles) or *loxP* (white triangles) sites leads to expression of CFP or YFP, respectively. **b,** Limited Cre-mediated recombination of tandem cassette insertions results in combinatorial expression of fluorescent proteins. **c,** Cartoon of lineage-tracing experiments. **d,** 10 dpf ventricular surface myocardium. Single cardiomyocytes are predominantly labeled with unique colors (arrowheads). **e, f,** 10 dpf ventricular confocal slice, indicating trabecular cardiomyocytes connected with clonally unrelated cardiomyocytes at the wall (**e**, **f,** arrow) and within trabeculae (**f**, arrowheads). Com, compact muscle. Tr, trabecular muscle. Scale bars, 10 \u03bcm.](nihms364290f1){#F1}\n\n![Several dozen embryonic cardiomyocytes build the juvenile ventricular wall\\\n**a,** Surface myocardium of half of a 30 dpf ventricular side, displaying clonal patches of varied shapes and sizes. **b, c,** Cardiomyocyte clones near the apex or chamber midpoint forming wedge/stripe shapes (arrowheads). **d,** Single-cell clone (green, arrowhead) positioned near a large clone (yellow, arrow). **e,** 30 dpf ventricular confocal slice, depicting a wall of single-cardiomyocyte thickness (Com) surrounding trabecular muscle (Tr). **f,** %Surface area occupied by 30 dpf clones (146 clones, 5 ventricles). **g,** Surface clones per ventricle (n=10). Scale bars, 50 \u03bcm.](nihms364290f2){#F2}\n\n![Clonally dominant cortical cardiomyocytes\\\n**a--c,** 6 weeks post-fertilization (wpf) ventricular surface (**a, b**) and confocal slice (**c**), indicating cortical (Cor, arrow), primordial (Pr, arrowhead), and trabecular (Tr) muscle. **d--g,** 8 wpf ventricular surface indicating a large green basal clone (**d--f**), and section indicating 3 muscle types (**g**). **h,** 90 dpf ventricular surface, showing only a few large cortical clones. **i,** 90 dpf ventricular section. **j,** %Surface area occupied by 90 dpf clones (56 clones, 10 ventricles); basal clones representing each ventricle are circled. **k,** Surface clones per ventricle (n=10). Scale bars, 50 \u03bcm (**a--g**); 100 \u03bcm (**h, i**)](nihms364290f3){#F3}\n\n![Regeneration of cortical and primordial muscle after injury\\\n**a,** Section of an uninjured ventricular apex, indicating the primordial (arrowheads) and cortical (arrows) layers. **b,** Regenerating ventricular apex at 14 days after resection (dpa). Cortical muscle clones converge within the injury site, while the primordial layer lags behind. Dashed line indicates amputation plane. **c,** 30 dpa ventricular apex, indicating multiple cortical clones and an incomplete primordial layer. **d,** Regenerated ventricular apex at 60 dpa, containing cortical muscle overlying a mostly contiguous layer of primordial muscle. n=6 animals for each timepoint. Scale bars, 50 \u03bcm](nihms364290f4){#F4}\n\n![Origins of clonally dominant cardiomyocytes\\\n**a,** 90 dpf ventricular surface, after 30 dpf 4-HT labeling. **b,** %Surface area occupied by 90 dpf clones (58 clones, 11 ventricles). **c,** Surface clones per ventricle (n=11). **d--g,** 8 wpf ventricular surface (**d, f**) and confocal slice (**e, g**), after 30 dpf 4-HT labeling. Cyan cortical muscle (arrows, **f, g**) overlies primordial muscle (arrowheads) that is largely red/unlabeled, and cyan trabeculae (arrow, **g**). **h--k,** 8 wpf ventricular surface (**h**) and confocal slices (**i--k**) after limited 4-HT labeling, showing no obvious primordial muscle labeling (arrowheads, **i--k**). Arrows indicate green (**h, i** (left); **j**) and hazel (**h, i** (right); **k**) cortical/trabecular clones. All cortical muscle in **h** is outlined in white. Scale bars, 50 \u03bcm](nihms364290f5){#F5}\n"}
+{"text": "1. Introduction {#sec1-molecules-25-01841}\n===============\n\nProtein-protein interaction (PPI) plays a distinctly important role in understanding cellular biological activities \\[[@B1-molecules-25-01841]\\]. Its research contributes to understanding the protein function, mechanism of biological activity, disease diagnosis and prevention, and new drug development \\[[@B2-molecules-25-01841],[@B3-molecules-25-01841],[@B4-molecules-25-01841]\\]. The research methods of PPI can be divided into two types: computational and experimental methods. Over the past few decades, many innovative experimental technologies have been designed to attempt to validate PPIs, such as glutathione *S*-transferase \\[[@B5-molecules-25-01841]\\], protein chip \\[[@B6-molecules-25-01841]\\], yeast two-hybrid \\[[@B7-molecules-25-01841]\\], tandem affinity purification (TAP) tag \\[[@B8-molecules-25-01841]\\], and other high-throughput technologies. Some direct interactions data of different species have been discovered and validated \\[[@B9-molecules-25-01841]\\]. However, the traditional experiment-based methods are not only costly and time-consuming, but also have high rates of false-positive predictions and weak generalization ability. Therefore, developing new reliable computational approaches has great practical significance for PPI identification at low cost and high efficiency \\[[@B10-molecules-25-01841]\\].\n\nIn recent years, some computational approaches based on various types of information about proteins have been suggested to predict PPIs, such as genomic information, structure information, evolutionary knowledge, protein domains, and phylogenetic profiles \\[[@B11-molecules-25-01841],[@B12-molecules-25-01841],[@B13-molecules-25-01841],[@B14-molecules-25-01841]\\]. However, the above prior information that can be used to predict the PPIs is scarce compared with the rapid growth of amino acid sequences \\[[@B15-molecules-25-01841]\\]. The above methods cannot be applied without prior knowledge of the proteins. In contrast, it is more significant to only use the protein amino acid sequences for predicting PPIs.\n\nExtensive experiments show that using the protein sequence information alone is enough for identifying PPIs \\[[@B16-molecules-25-01841]\\]. Many sequence-based computational methods have been explored to address the problems, such as support vector machine (SVM) with multi-scale discontinuous and continuous \\[[@B16-molecules-25-01841]\\], rotation forest algorithm with position-specific scoring matrix (PSSM) \\[[@B17-molecules-25-01841]\\], SVM with auto covariance (AC) \\[[@B18-molecules-25-01841]\\], average blocks (AB) using relevance vector machine (RVM) \\[[@B19-molecules-25-01841]\\], discrete cosine transformation using weighted sparse representation \\[[@B20-molecules-25-01841]\\], and so on. In 2018, G\u00f6ktepe et al. \\[[@B21-molecules-25-01841]\\] presented a feature representation method named weighted skip-sequential conjoint triads using principal component analysis (PCA) and SVM to capture the information of protein sequences. In 2019, Chen et al. \\[[@B1-molecules-25-01841]\\] designed an end-to-end framework which incorporated a deep residual recurrent convolutional neural network for capturing the information of protein sequences. In the same year, Zhang et al. \\[[@B22-molecules-25-01841]\\] presented a neural network-based model which separately used different descriptors (auto covariance descriptor, local descriptors, and multi-scale continuous and discontinuous local descriptor) to explore and represent the patterns of interactions between amino acids. Although the researches of these approaches have achieved good progress and application prospects, new methods are needed to further improve the performances of PPI predictions.\n\nThe concept of the energy of graph G is due to Gutman \\[[@B23-molecules-25-01841]\\] and is meaningful for the analysis of graph theory \\[[@B24-molecules-25-01841],[@B25-molecules-25-01841],[@B26-molecules-25-01841]\\]. Nowadays, the energy of the graph has been used in chemistry, bioinformatics, and related fields \\[[@B27-molecules-25-01841],[@B28-molecules-25-01841]\\]. In the literature, increasing studies have shown that the physicochemical properties of amino acids can improve the prediction performances of PPIs \\[[@B16-molecules-25-01841],[@B29-molecules-25-01841]\\]. The contact information among amino acids is also significant for prediction of PPIs \\[[@B30-molecules-25-01841]\\]. The multi-information fusion of different feature extraction methods can fuse different feature information of interacting protein sequences, and it is an effective technique in improving the prediction performance of PPIs \\[[@B22-molecules-25-01841]\\].\n\nIn this paper, we present a computational model to predict PPIs using only protein sequences and graph energy. Inspired by the graph energy theory, we design two feature extraction methods for PPIs---physicochemical graph energy (PGE) and contact graph energy (CGE)---to capture the feature information of interactions. Physicochemical graph energy is graph energy based on physical and chemical properties, while contact graph energy is graph energy based on amino acid contact information. The dipeptide composition method was also used to extract and supplement effective information. PCA was implemented to effectively reduce the influence of noise after integrating three feature extraction methods. The weighted sparse representation-based classification (WSRC) was used as the classifier of the proposed method after different classifiers were compared. The PPI-GE has been tested on human, *H. pylori*, and yeast data sets, and these three data sets achieved prediction accuracies of 99.49%, 97.15%, and 99.56%, respectively. In addition, we verified the validity of the proposed method on five independent data sets and two significant PPI networks. The comparative experimental results indicate that our feature extraction methods have a significant effect on PPI prediction and our method is superior to other state-of-the-art prediction methods.\n\n2. Results and Discussion {#sec2-molecules-25-01841}\n=========================\n\n2.1. Evaluation Metrics {#sec2dot1-molecules-25-01841}\n-----------------------\n\nIn this study, to ensure the reliability of experimental results and avoid over fitting of data, we implemented five-fold cross-validation to evaluate the effectiveness of PPI-GE and other computational models. Specifically, the experimental data set was split into five parts; each of the five parts is regarded as an independent testing data set and the other four parts were selected as training data sets. Several widely-used evaluation metrics were used, including accuracy (ACC), sensitivity (SEN), precision (Pre), and Matthews correlation coefficient (MCC), expressed as follows:$$ACC = \\frac{TP + TN}{TP + TN + FP + FN}$$ $$SEN = \\frac{TP}{TP + FN}$$ $$Pre = \\frac{TP}{TP + FP}$$ $$MCC = \\frac{TP \\times TN - FP \\times FN}{\\sqrt{\\left( {TN + FP} \\right) \\times \\left( {TP + FP} \\right) \\times \\left( {TP + FN} \\right) \\times \\left( {TN + FN} \\right)}}$$ where *TP*, *TN*, *FP*, and *FN* denote the number of true positives, true negatives, false positives, and false negatives, respectively. In the protein--protein interaction data sets, the unknown protein-protein interactions are considered negative samples, while the known interactions are called positive samples. The average performance of all evaluation metrics is obtained during the experiment. As the value of metric is larger, the performance of the method will be better. Moreover, the area under the receiver-operating characteristics curve (AUC) was calculated to further evaluate the performance of the method. The AUC value of 1 indicates perfect prediction and AUC value of 0.5 means random prediction.\n\n2.2. The Performance Comparisons of Different Classifiers {#sec2dot2-molecules-25-01841}\n---------------------------------------------------------\n\nIt is well known that the same method using different classifiers may achieve different prediction results. To further evaluate the proposed method, the K-nearest neighbors (KNN), support vector machine (SVM), and WSRC classifiers were adopted to predict PPIs using the same feature extraction methods. To ensure the universality of different data set, we implemented five-fold cross-validation 10 times and obtained the average value of three benchmark data sets for the same evaluation metric using every classifier, respectively.\n\nThe average results of three benchmark data sets with different classifiers are presented in [Figure 1](#molecules-25-01841-f001){ref-type=\"fig\"}. From the figure, the results of the comparison prove that the performance of the WSRC classifier has better stability and higher accuracy than the SVM and KNN classifiers for predicting PPIs. In this work, we used the WSRC classifier as the classifier of our model.\n\n2.3. Prediction Performances of the Proposed Method {#sec2dot3-molecules-25-01841}\n---------------------------------------------------\n\nFor verifying the efficacy and stability and reducing deviations of PPI-GE based on the WSRC classifier, five-fold cross-validation was performed in the experiment. The cross-validation results of three benchmark data sets are shown in [Table 1](#molecules-25-01841-t001){ref-type=\"table\"}, [Table 2](#molecules-25-01841-t002){ref-type=\"table\"} and [Table 3](#molecules-25-01841-t003){ref-type=\"table\"}.\n\nWhen performing on the human data set, ACC, SEN, MCC, Pre, and AUC achieved the average performance of 99.49%, 99.21%, 98.97%, 99.72%, and 99.99%, respectively (see [Table 1](#molecules-25-01841-t001){ref-type=\"table\"}). Similarly, the average results of ACC, SEN, MCC, Pre, and AUC on the *H. pylori* data set were 97.15%, 98.23%, 94.35%, 96.17%, and 99.19%, respectively (see [Table 2](#molecules-25-01841-t002){ref-type=\"table\"}). At the same time, we also gained better average results of these metrics at 99.56%, 99.14%, 99.13%, 99.98%, and 100% on the yeast data set, respectively (see [Table 3](#molecules-25-01841-t003){ref-type=\"table\"}). The experimental results show that PPI-GE is robust and promising for predicting PPIs. Our method achieved better prediction results which may be attributed to the choice of the classifiers and novel feature extraction methods.\n\n2.4. Comparison with Other Methods {#sec2dot4-molecules-25-01841}\n----------------------------------\n\nCurrently, many kinds of computational models based on protein sequences have been presented for predicting PPIs. In this section, to further objectively validate the prediction performance of the proposed method, seven state-of-the-art methods, including Ensemble Deep Neural Networks (EnsDNN) \\[[@B22-molecules-25-01841]\\], 3-mers-based \\[[@B31-molecules-25-01841]\\], Bio2vec-based \\[[@B31-molecules-25-01841]\\], pseudo Substitution Matrix Representation (pseudo-SMR) \\[[@B32-molecules-25-01841]\\], WSRC with continuous wavelet and discrete wavelet transform (WSRC+CW and DW) \\[[@B33-molecules-25-01841]\\], feature weighted rotation forest algorithm (FWRF) \\[[@B17-molecules-25-01841]\\], and Global encoding \\[[@B34-molecules-25-01841]\\] were compared on the human, *H. pylori*, and yeast data sets. The comparison results of three benchmark data sets based on five-fold cross-validation of different models are plotted in [Figure 2](#molecules-25-01841-f002){ref-type=\"fig\"}, [Figure 3](#molecules-25-01841-f003){ref-type=\"fig\"} and [Figure 4](#molecules-25-01841-f004){ref-type=\"fig\"}, respectively.\n\nSome previous algorithms did not use all three benchmark datasets in their papers, therefore we first compared the proposed method with the other five methods on the human data set. [Figure 2](#molecules-25-01841-f002){ref-type=\"fig\"} shows that the proposed method obtained higher average accuracy (99.49%) out of these methods. Meanwhile, the results of SEN, MCC, and AUC are superior to others. On the *H. pylori* data set, our method and six other methods were used for the comparison. From [Figure 3](#molecules-25-01841-f003){ref-type=\"fig\"}, it can be noted that our method is significantly better than that of others. On the yeast data set, the results of comparison among six different methods are shown in [Figure 4](#molecules-25-01841-f004){ref-type=\"fig\"}. We obtained similar results. The comparison results show that our method obtained satisfactory performance relative to current existing approaches. This further demonstrates that the proposed method based on the novel feature extraction methods is robust and effective.\n\n2.5. Performance on PPI Networks {#sec2dot5-molecules-25-01841}\n--------------------------------\n\nSince the development of the disease may involve proteins and pathways in multiple biological processes, PPI networks may help to understand the deregulated molecular mechanisms of disease development and progression and the functional organization of proteins. The general PPI networks are crossover networks from a biological perspective \\[[@B35-molecules-25-01841]\\]. It is necessary to evaluate the performance of the proposed method by predicting the PPI network. In this section, the Wnt signaling pathway network is a significant crossover network, which was used for evaluating.\n\nTo keep the same experimental conditions during the comparison, the yeast data set was regarded as the training data set and the Wnt-related network was regarded as the testing data set. Since they are different species, in the encoding, the dimension of fused feature vector $E$ was reduced to 20 for eliminating the influence of more noise. The network and prediction results are shown in [Figure 5](#molecules-25-01841-f005){ref-type=\"fig\"}. The red line is false prediction. It can be seen from the figure that our method can predict 92 of the 96 PPIs. We also compared some previous methods with the proposed method, and the comparisons are listed in [Table 4](#molecules-25-01841-t004){ref-type=\"table\"}. From the table, it can be noted that the proposed method is significantly better than others. In addition to this, we also tested our method on the multi-core network (Ras-Raf-Mek-Erk-Elk-Srf pathway) for predicting. The network is shown in [Figure 6](#molecules-25-01841-f006){ref-type=\"fig\"}. The core protein is colored yellow. Our method correctly predicts all PPIs. The results suggest that PPI-GE can be applied to predict PPIs encoded in the network and obtain better prediction results.\n\n2.6. Performance on Independent Data Sets {#sec2dot6-molecules-25-01841}\n-----------------------------------------\n\nFinally, to further validate the efficacy and stability of our method, we also tested the proposed method and compared it with several state-of-the-art methods on five independent data sets (*H. pylori*, *H. sapien*, *C. elegans*, *M. musculus*, and *D. mela* data sets). In the encoding, the yeast data set was regarded as the training data set and the independent data set was regarded as the testing data set, and the same feature extraction methods were used during the experiment. The comparison results between different methods are summarized in [Table 5](#molecules-25-01841-t005){ref-type=\"table\"}. The accuracies of five independent data sets were 93.80%, 99.93%, 86.24%, 94.57%, and 99.87%, respectively. The proposed method has better performance for PPI prediction on four data sets (*H. pylori*, *H. sapien*, *M. musculus*, and *D. mela*). However, our accuracy on the *C. elegans* data set is lower than Du's work and Ding's work. Overall, it indicates that our method can perform across species for PPI predictions.\n\n3. Materials and Methods {#sec3-molecules-25-01841}\n========================\n\nIn this section, a novel method called PPI-GE is described, which depends mainly on three steps. The flowchart of PPI-GE is shown in [Figure 7](#molecules-25-01841-f007){ref-type=\"fig\"}. First, the method only uses the amino acid sequences through physicochemical graph energy, contact graph energy, and dipeptide composition for feature extraction and multi-information fusion. Then, the PCA method was implemented for the descending dimension and eliminating noise. Finally, the WSRC classifier was applied for sample classification and predicting PPIs after different classifiers were compared.\n\n3.1. Datasets {#sec3dot1-molecules-25-01841}\n-------------\n\nIn this work, three high-quality benchmark data sets were used to ensure generality and evaluate the performance of the proposed method. The first data set is the human data set constructed by Huang et al. \\[[@B20-molecules-25-01841]\\]. They collected 3899 experimentally verified PPIs as a positive sub-dataset and obtained 4262 non-PPI pairs from different subcellular compartments as a negative sub-dataset. The *H. pylori* data set is used as the second data set constructed by Martin et al. \\[[@B38-molecules-25-01841]\\], and the third data set is the yeast data set collected by Guo et al. \\[[@B18-molecules-25-01841]\\]. The summary of three benchmark data sets can be seen in [Table 6](#molecules-25-01841-t006){ref-type=\"table\"}.\n\nIn addition, we tested on two significant PPI networks to objectively validate the performance of the proposed method. The first network is the crossover network (Wnt-related network) \\[[@B39-molecules-25-01841]\\] which contains 96 PPIs. It is a significant signaling pathway and plays a distinctly important role in the understanding of tumor formation, processes of cytoskeletal organization, patterning, and organogenesis. The second network is the multi-core network (Ras-Raf-Mek-Erk-Elk-Srf pathway) \\[[@B40-molecules-25-01841]\\] which includes 189 PPIs. It is an important consensus network and implicates a variety of transcriptional regulations and cellular processes. To further verify the efficacy and stability of our method, we also tested it on five independent data sets, including *H. pylori*, *H. sapien*, *C. elegans*, *M. musculus*, and *D. mela* data sets \\[[@B30-molecules-25-01841],[@B34-molecules-25-01841]\\].\n\n3.2. Feature Extraction {#sec3dot2-molecules-25-01841}\n-----------------------\n\nIn this work, we designed two feature extraction methods to capture the feature information of sequences: the graph energy based on the physical and chemical properties named physicochemical graph energy (PGE) and the graph energy based on the amino acid contact information named contact graph energy (CGE). The concept of the energy of a graph is due to Gutman \\[[@B23-molecules-25-01841]\\]. If graph $G$ is a simple graph, the energy of the graph is defined as the sum of the absolute eigenvalues of the adjacency matrix of the graph $G$. If $E\\left( G \\right)$ represents the energy of a graph, we get $$E\\left( G \\right) = \\sum\\limits_{i = 1}^{n}|\\lambda_{i}|$$ where $\\lambda_{i}$ is the $i$th eigenvalue of the adjacency matrix \\[[@B25-molecules-25-01841],[@B26-molecules-25-01841]\\].\n\n### 3.2.1. Physicochemical Graph Energy {#sec3dot2dot1-molecules-25-01841}\n\nAmino acids are the basic units of protein sequences and have different physicochemical properties which have a great significance for the prediction of protein functions and structures \\[[@B41-molecules-25-01841],[@B42-molecules-25-01841]\\]. Conventionally, the location information of the amino acid sequence is important in the prediction of PPIs, because amino acids make up protein sequences and have specific positions which are closely related to the local interaction information of amino acid neighborhoods.\n\nInspired by previous work \\[[@B27-molecules-25-01841]\\], we obtained a descending order $\\left. (D\\rightarrow E\\rightarrow C\\rightarrow N\\rightarrow M\\rightarrow F\\rightarrow Q\\rightarrow Y\\rightarrow S\\rightarrow P\\rightarrow T\\rightarrow V\\rightarrow L\\rightarrow I\\rightarrow W\\rightarrow H\\rightarrow G\\rightarrow A\\rightarrow R\\rightarrow K \\right.$) based on the isoelectric point and ionization equilibrium constant typical physicochemical properties of amino acids. Then, a unit substitution matrix $A \\in R^{20 \\times 20}($as shown in [Figure 8](#molecules-25-01841-f008){ref-type=\"fig\"}) was constructed by using this ordering to describe the location information of amino acids. A protein sequence with the length of $n$ can be transformed into a $\\left( {0,1} \\right)$-adjacency matrix $A_{G} = \\left( g_{i,j} \\right)_{20 \\times n}$ based on the unit substitution matrix $A$. It is defined as follows: when the $j$th amino acid type of protein sequence is the same as the $t$th amino acid type of above amino acid order, let $A_{G}\\left( {i,j} \\right) = A\\left( {i,t} \\right)$, where $i,t = 1,2,\\cdots,20;\\ j = 1,2,\\cdots,n$.\n\nNext, we constructed a sliding window with the length of 20 to transform the matrix $A_{G}$ into $n - 19$ matrixes of $20 \\times 20$ dimensions. Let $A_{G}^{k} \\in R^{20 \\times 20}$ be a sliding window which starts from the first amino acid at the left end of the protein sequence and moves the position of one amino acid at a time. Suppose $A_{G}^{k}$ is the sparse sub-matrix obtained by sliding the window to the $k$th amino acid of the sequence, then $A_{G}^{k}\\left( {i,j} \\right) = A_{G}\\left( {i,k + j - 1} \\right)$, where $i = 1,2,\\cdots,20;j = 1,2,\\cdots,20;k = 1,2,\\cdots,n - 19$. As shown in [Figure 8](#molecules-25-01841-f008){ref-type=\"fig\"}, $A_{G}^{k}$ can correspond to a sparse bipartite graph $G_{Ak}$ one by one, and each amino acid corresponds to a point in the bipartite graph. Finally, the physicochemical graph energy $E\\left( G_{Ak} \\right)$ of each bipartite graph $G_{Ak}$ was calculated as Equation (5). In this way, we turned a protein sequence into a numerical vector $E_{1}^{*} = \\left\\{ E\\left( G_{A1} \\right),E\\left( G_{A2} \\right),\\cdots,E\\left( G_{{An} - 19} \\right) \\right.$}.\n\nMachine learning methods need to input feature vectors with the same lengths, but different proteins may have different sequence lengths. In the literature, the top 30-dimensional feature vector of a protein sequence can contain some information for the similarity analysis of proteins \\[[@B28-molecules-25-01841]\\]. Here, we set 200 windows to obtain enough information of proteins for PPIs. The algorithm implements zero-padding if the length of the protein is less than 219 residues. Thus, a protein sequence with the length of $n$ can be characterized by a 200-dimensional numerical vector $E_{1}^{*} = \\left\\{ E\\left( G_{A1} \\right),E\\left( G_{A2} \\right),\\cdots,E\\left( G_{A200} \\right) \\right.$}.\n\n### 3.2.2. Contact Graph Energy {#sec3dot2dot2-molecules-25-01841}\n\nThe contact information among different types of amino acids is significant for predicting PPIs as described by Ding et al. \\[[@B30-molecules-25-01841]\\]. They considered 20 kinds of amino acid, 8 types of secondary structures, disability contact solvents, and 6323 complexes \\[[@B43-molecules-25-01841],[@B44-molecules-25-01841]\\]. Then, the average number of pairwise contacts observed at the interface was calculated from unbound protein to binding structure. In this section, the second alternative matrix $B \\in R^{20 \\times 20}$ (as shown in [Figure 8](#molecules-25-01841-f008){ref-type=\"fig\"}) is the contact matrix of amino acids which is based on the effective contact energy among different types of amino acids. It is noteworthy that we used the same amino acid contact matrix as used in Ding's work.\n\nNext, as the physicochemical graph energy is based on the substitution matrix $B$, the protein sequence was transformed into an adjacency matrix $B_{G} \\in R^{20 \\times n}$. We set a sliding window to slide on the protein sequence and obtain sub-matrix $B_{G}^{k} \\in R^{20 \\times 20}$, where $k = 1,2,\\cdots,n - 19$. Each matrix $B_{G}^{k}$ corresponds to a complete bipartite graph one by one. Finally, the contact graph energy $E\\left( G_{Bk} \\right)$ of each bipartite $G_{Bk}$ was calculated as Equation (5).\n\nTherefore, according to the proposed contact graph energy, we can characterize and transform a protein sequence into a numerical vector $E_{2}^{*} = \\left\\{ E\\left( G_{B1} \\right),E\\left( G_{B2} \\right),\\cdots,E\\left( G_{B200} \\right) \\right.$}.\n\n### 3.2.3. *N*-peptide Composition Representation {#sec3dot2dot3-molecules-25-01841}\n\nIt is generally known that amino acids are the basic units that make up protein sequences. A protein sequence can be simply expressed as follows: $$P = A_{1}A_{2}\\cdots A_{i}\\cdots A_{L - 1}A_{L}$$ where $A_{i}$ represents the $i$th amino acid and each belongs to one of the 20 native amino acid types; $L$ denotes the number of amino acids in the sequence.\n\nThe pseudo amino acid composition is widely used to extract sequence information of proteins. The simplest one is called n-peptide composition. In this way, it can preserve the protein sequence order information. When $n = 2$, this method degenerates the amino acid composition into a dipeptide composition \\[[@B21-molecules-25-01841],[@B45-molecules-25-01841],[@B46-molecules-25-01841]\\]. Dipeptide composition treats every two contiguous amino acids as a combination. Therefore, there are $L - 1$ combinations in a protein sequence. If the protein sequence information is known, we can calculate the frequency of these combinations and represent them by a 400-dimensional vector $E_{3}^{*}$. We can calculate the frequency value as the following formula:$$f_{mn} = \\frac{N_{mn}}{L - 1}\\ ,~~~1 \\leq m \\leq 20,\\ 1 \\leq n \\leq 20$$ where $N_{mn}$ represents the number of combinations that consist of the $m$th and $n$th types of amino acids and appear in the protein sequence.\n\n3.3. Principal Component Analysis {#sec3dot3-molecules-25-01841}\n---------------------------------\n\nPCA is an effective data analysis technique which can reduce the dimension of the raw data, eliminate some noise for promoting data processing speed, and save time. It has been widely used to process data in bioinformatics and related fields \\[[@B19-molecules-25-01841],[@B47-molecules-25-01841]\\]. It can retain the main information of variable interactions when the high-dimensional sample data set is transformed into a low-dimensional space.\n\nIn this work, we obtained the fused 600-dimensional numerical vector $E = \\left( {E_{1}^{*};E_{2}^{*};E_{3}^{*}} \\right)$ by combining three numerical vectors based on the physicochemical graph energy, contact graph energy, and dipeptide composition. The multi-information fusion fused different feature information of interacting protein sequences, which may bring more noise information. Thus, on the fused feature vector, the PCA method was applied to eliminate the influence of noise and integrate useful information. Considering that some important information may be ignored if the dimension is too small, the dimension of fused feature vector $E$ was reduced from 600 to 80 through many experiments to obtain the new feature vector and improve the prediction accuracy. After using the PCA method, the most discriminative new feature set was obtained and used as input information to train the classifier for PPI prediction tasks.\n\n3.4. Weighted Sparse Representation Based Classification {#sec3dot4-molecules-25-01841}\n--------------------------------------------------------\n\nIn this paper, the WSRC classifier was used as the classifier for predicting PPIs, which was proposed by Lu et al. \\[[@B48-molecules-25-01841]\\] in 2012. It is based on the sparse representation-based classification (SRC), uses the Gauss kernel function to measure the similarity between samples, and overcomes the shortcomings of sparse coding \\[[@B20-molecules-25-01841]\\].\n\nConsidering sample data set $X \\in R^{m \\times n}$, it consists of n samples, and each of the samples is composed of an m-dimensional feature vector. Set $L$ denotes the number of all classes in the sample data set. The samples belonging to the $l{th}$ class can be represented by a sub-matrix $X_{l} = \\left\\lbrack {s_{l1},s_{l2},\\cdots,s_{ln_{l}}} \\right\\rbrack$, where $s_{li}$ means the label of the $i$th sample belonging to the $l$th class and $n_{l}$ refers to the sample size of the $l{th}$ class. Therefore, the sample matrix can be represented as $X = \\left\\lbrack {X_{1},X_{2},\\cdots,X_{L}} \\right\\rbrack$.\n\nAssuming test sample $y \\in R^{m}$ is a sample of the $l$th class, $y$ can be expressed as:$$y = b_{l,1}s_{l1} + b_{l,2}s_{l2} + \\cdots + b_{l,n_{l}}s_{ln_{l}}$$\n\nThis equation can be further expressed as:$$y = X\\beta$$ where $\\beta = \\left\\lbrack {0,\\cdots,0,\\beta_{l1},\\beta_{l2},\\cdots,\\beta_{ln_{l}},0,\\cdots,0} \\right\\rbrack$. As the number of samples grows larger, $\\alpha$ becomes sparser, since the non-zero entries in $\\beta$ are only related to the $l$th class. The key to the principle of WSRC is to calculate the vector $\\beta$, which not only needs to satisfy Equation (8), but also minimize the $L1$-norm of $\\beta$. This can be expressed as follows:$${\\hat{\\beta_{1}} = \\arg min\\left\\| W\\beta \\right\\|_{1}}{,\\ {subject}\\ {to}\\ \\left\\| y - X\\beta \\right\\|_{2} < \\varepsilon}$$ where $\\varepsilon > 0$ is a threshold, and *W* is a block-diagonal matrix:$$diag\\left( W \\right) = \\left\\lbrack {d_{G}\\left( {y,x_{1}^{1}} \\right),\\cdots,d_{G}\\left( {y,x_{n_{L}}^{L}} \\right)} \\right\\rbrack$$ where $x_{i}^{j}$ denotes the $i$th sample of the $j$th class and $d_{G}\\left( {\\cdot , \\cdot} \\right)$ represents the Gaussian distance function:$$d_{G}\\left( {y,x_{i}^{j}} \\right) = e^{- {\\| y - x_{i}^{j}\\|}^{2}/2\\sigma^{2}}$$ where $\\sigma$ is the Gaussian kernel width, and $i = 1,\\ldots,n_{L},j = 1,\\ldots,L$. Then, the type of test sample $y$ will be determined by the sparse representation classifier, and the formula can be described as follows:$$\\min\\limits_{c}r_{c}\\left( y \\right) = \\left\\| y - X{\\hat{\\beta}}_{1}^{c} \\right\\|_{2}$$ where $c = 1,\\ldots,L$. In this paper, the WSRC classifier was applied for sample classification.\n\n4. Conclusions {#sec4-molecules-25-01841}\n==============\n\nIn this paper, we introduce graph energy to encode protein sequences and present a novel prediction method called PPI-GE for predicting PPIs using amino acid sequences alone. In the process of feature extraction, we designed two new feature extraction methods: physicochemical graph energy based on the ionization equilibrium constant and isoelectric point of amino acids and contact graph energy based on the contact information of amino acids. In addition, the dipeptide composition method was used to extract and supplement effective order information. These feature extraction methods can comprehensively consider the physical and chemical properties as well as the contact and location information of amino acids. The WSRC classifier was used as the classifier of the prediction model. The proposed method was tested on three benchmark data sets (human, *H. pylori*, and yeast data sets), two important PPI networks (Wnt-related pathway and Ras-Raf-Mek-Erk-Elk-Srf pathway), and five independent data sets (*H. pylori*, *H. sapien*, *C. elegans*, *M. musculus*, and *D. mela* data sets); good prediction results were obtained. The experimental results indicate that our proposed method is robust and superior compared to previous methods.\n\nWe thank the reviewers for their time reading the paper and constructive comments.\n\n**Sample Availability:** Samples of the compounds are available from the authors.\n\nD.X. and Y.Z. conceived and designed the study. D.X., H.X., and Y.Z. obtained and processed the data, performed the experiments, and analyzed the results. D.X. and H.X. wrote the manuscript. W.C. and R.G. provided suggestions and supervised the research. All authors have read and agreed to the published version of the manuscript.\n\nThis research was funded by the National Natural Science Foundation of China, grant number 61877064, U1806202, and 61533011.\n\nThe authors declare no conflicts of interest.\n\n![The performance comparisons of different classifiers. WSRC: weighted sparse representation-based classification; KNN: K-nearest neighbors; SVM: support vector machine.](molecules-25-01841-g001){#molecules-25-01841-f001}\n\n![Comparison results of different methods on the human data set.](molecules-25-01841-g002){#molecules-25-01841-f002}\n\n![Comparison results of different methods on the *H. pylori* data set.](molecules-25-01841-g003){#molecules-25-01841-f003}\n\n![Comparison results of different methods on the yeast data set.](molecules-25-01841-g004){#molecules-25-01841-f004}\n\n![The prediction results of crossover network (Wnt-related network).](molecules-25-01841-g005){#molecules-25-01841-f005}\n\n![The prediction results of multi-core network.](molecules-25-01841-g006){#molecules-25-01841-f006}\n\n![The flowchart of the proposed method for predicting protein-protein interactions (PPIs).](molecules-25-01841-g007){#molecules-25-01841-f007}\n\n![The schematic diagram of protein sequence feature extraction.](molecules-25-01841-g008){#molecules-25-01841-f008}\n\nmolecules-25-01841-t001_Table 1\n\n###### \n\nfive-fold cross-validation results on the human data set.\n\n  Testing Set   ACC (%)   SEN (%)   MCC (%)   Pre (%)   AUC (%)\n  ------------- --------- --------- --------- --------- ---------\n  1             99.33     98.76     98.66     99.87     99.99\n  2             99.63     99.34     99.26     99.87     100\n  3             99.57     99.22     99.14     99.87     100\n  4             99.39     99.24     98.77     99.49     99.98\n  5             99.51     99.49     99.02     99.49     99.99\n  Average       99.49     99.21     98.97     99.72     99.99\n\nmolecules-25-01841-t002_Table 2\n\n###### \n\nfive-fold cross-validation results on the *H. pylori* data set.\n\n  Testing Set   ACC (%)   SEN (%)   MCC (%)   Pre (%)   AUC (%)\n  ------------- --------- --------- --------- --------- ---------\n  1             95.55     98.21     91.24     92.88     99.27\n  2             97.94     98.26     95.89     97.59     99.34\n  3             98.11     99.65     96.27     96.64     98.94\n  4             97.94     97.69     95.88     98.34     99.41\n  5             96.23     97.32     92.46     95.41     99.01\n  Average       97.15     98.23     94.35     96.17     99.19\n\nmolecules-25-01841-t003_Table 3\n\n###### \n\nfive-fold cross-validation results on the yeast data set.\n\n  Testing Set   ACC (%)   SEN (%)   MCC (%)   Pre (%)   AUC (%)\n  ------------- --------- --------- --------- --------- ---------\n  1             99.60     99.18     99.20     100       100\n  2             99.46     98.95     98.93     100       100\n  3             99.55     99.20     99.11     99.91     100\n  4             99.51     99.00     99.02     100       100\n  5             99.69     99.38     99.38     100       100\n  Average       99.56     99.14     99.13     99.98     100\n\nNote: ACC: accuracy; SEN: sensitivity; MCC: Matthews correlation coefficient; Pre: precision; AUC: area under the curve.\n\nmolecules-25-01841-t004_Table 4\n\n###### \n\nComparison of different methods on the Wnt-related network using yeast data set as the training data set.\n\n  Wnt-Related Network                         Proportion   Accuracy (%)\n  ------------------------------------------- ------------ --------------\n  Proposed method                             92/96        95.83\n  Ding's work \\[[@B30-molecules-25-01841]\\]   89/96        92.71\n  Shen's work \\[[@B35-molecules-25-01841]\\]   73/96        76.04\n  Zhou's work \\[[@B36-molecules-25-01841]\\]   87/96        90.63\n  Chen's work \\[[@B29-molecules-25-01841]\\]   89/96        92.71\n\nmolecules-25-01841-t005_Table 5\n\n###### \n\nComparison of the accuracy (%) between different methods on the independent data sets using yeast data set as the training data set.\n\n  Data Set        Testing Pairs   Proposed Method   Huang's Work \\[[@B34-molecules-25-01841]\\]   Du's Work \\[[@B37-molecules-25-01841]\\]   Ding's Work \\[[@B30-molecules-25-01841]\\]\n  --------------- --------------- ----------------- -------------------------------------------- ----------------------------------------- -------------------------------------------\n  *H. pylori*     1420            93.80             85.77                                        93.66                                     92.03\n  *H. sapien*     1412            99.93             88.81                                        93.77                                     94.58\n  *C. elegans*    4013            86.24             72.79                                        94.84                                     90.28\n  *M. musculus*   313             94.57             83.39                                        91.37                                     92.25\n  *D. mela*       21975           99.87             89.35                                        N/A                                       N/A\n\nNote: N/A means not available.\n\nmolecules-25-01841-t006_Table 6\n\n###### \n\nThe details of three benchmark data sets.\n\n  Datasets      Protein Pairs   Interaction Pairs   Non-Interaction Pairs   References\n  ------------- --------------- ------------------- ----------------------- -------------------------------\n  human         8161            3899                4262                    \\[[@B20-molecules-25-01841]\\]\n  *H. pylori*   2916            1458                1458                    \\[[@B38-molecules-25-01841]\\]\n  yeast         11,188          5594                5594                    \\[[@B18-molecules-25-01841]\\]\n"}
+{"text": "The repair capacity of articular cartilage is limited and large cartilage defects generally fail to heal spontaneously, making intervention necessary in order to avoid the progression of osteoarthritis[@b1]. Nevertheless, the therapeutic effect for large cartilage damage is not satisfactory, as autografts suffer from the inadequate tissue availability and the associated morbidity of the donor site, and allografts are limited by transplant rejection[@b2][@b3]. The tissue engineering for the treatment of articular cartilage defects presents a promising strategy[@b4], however, many problems remain. For instance, natural materials suffer from the limitation of immunogenicity, potential risks of transmitting animal-originated pathogens, and weak mechanical properties[@b5]; synthetic materials lack natural sites for cell adhesion and may cause a local reduction in pH and potential inflammation due to the degradation through hydrolysis[@b6]. Moreover, scaffolds combining natural and synthetic materials barely show good results in *in vivo* studies[@b7]. Much attention has been paid to the biological safety and efficacy of the scaffolds[@b8]. It is safe and reliable when the scaffolds are obtained from the patient's own tissue.\n\nPlatelet-rich plasma (PRP) is an autologous blood product containing concentrated platelets. After activation by thrombin or calcium chloride, the platelets in PRP release platelet-derived growth factor (PDGF), transforming growth factor (TGF), insulin-like growth factor (IGF), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), and many other growth factors through degranulation[@b9][@b10]. During cartilage formation or chondrocyte differentiation, TGF-\u03b2 induces chondrogenesis of bone marrow-derived stem cells (BMSC), while PDGF aids chondrocytes to maintain the hyaline-like chondral phenotype and promote proliferation and proteoglycan synthesis[@b11]. In a previous study, we demonstrated that PRP gel presents a porous bioactive scaffold for cartilage repair[@b12].\n\nDespite the increasing use of PRP for cartilage lesions[@b13][@b14], the optimal PRP formulation is still unknown, and over the past few years, much attention has been focused on the leukocyte concentrations in PRP. In a clinical study by Filardo[@b15], leukocyte-rich PRP (L-PRP) resulted in a higher incidence of side effects in the treatment of osteoarthritis compared with pure PRP (P-PRP), which had a lower leukocyte concentration, possibly due to the fact that leukocytes in PRP may deliver pro-inflammatory cytokines, such as interleukin-1\u03b2 (IL-1\u03b2) and tumor necrosis factor-\u03b1 (TNF-\u03b1), at the site of injection, resulting in the production of destructive proteases that inhibit the formation and promote the degradation of extracellular matrix[@b15][@b16][@b17]. IL-1\u03b2 and TNF-\u03b1 induce deleterious effects through the nuclear factor \u03baB (NF-\u03baB) signalling pathway[@b18][@b19]. NF-\u03baB proteins are typically present in an inactive form in the cytoplasm bound to I\u03baB (an inhibitory protein). Cell stimulation through IL-1\u03b2 or TNF-\u03b1 leads to the nuclear translocation of NF-\u03baB to trigger the expression of a wide range of regulatory genes involved in apoptosis, inflammation, and other immune responses[@b20]. Therefore, the concentrated leukocytes in L-PRP may activate the NF-\u03baB pathway through IL-1\u03b2 and TNF-\u03b1 to inhibit tissue healing. However, this has not yet been substantiated.\n\nBMSC are defined as self-renewal, multi-potent progenitor cells that differentiate into several mesenchymal lineages[@b21], and they could be obtained in a less invasive procedure from autologous bone marrow compared with chondrocytes harvested from cartilage extraction. As a primary cell candidate for cartilage tissue engineering, BMSC has been extensively studied to repair osteochondral defects and enhance cartilage regeneration[@b22][@b23].\n\nThe objective of the present study was to compare the effects of L-PRP and P-PRP (PRPs) on rabbit BMSC (rBMSC) *in vitro* and on cartilage repair *in vivo* and preliminarily explore the mechanism to improve the efficacy of PRP therapy.\n\nResults\n=======\n\nComponents of PRPs and whole blood\n----------------------------------\n\nThe components of PRPs and whole blood (WB) are shown in [Fig. 1](#f1){ref-type=\"fig\"}. The mean leukocyte concentration of L-PRP was significantly higher than that of whole blood, while P-PRP had a lower concentration than whole blood (p\u2009\\<\u20090.01, [Fig. 1A](#f1){ref-type=\"fig\"}). The levels of IL-1\u03b2 and TNF-\u03b1 were also higher in L-PRP and lower in P-PRP compared with whole blood (p\u2009\\<\u20090.01, [Fig. 1B and C](#f1){ref-type=\"fig\"}). L-PRP and P-PRP had a similar level of platelet concentration, and produced an almost 6-fold increase in platelet concentration over whole blood (p\u2009\\<\u20090.01, [Fig. 1D](#f1){ref-type=\"fig\"}). Moreover, the concentrations of PDGF-AB and TGF-\u03b21 were similar between L-PRP and P-PRP (p\u2009\\>\u20090.01, [Fig. 1E and F](#f1){ref-type=\"fig\"}), but significantly lower in whole blood (p\u2009\\<\u20090.01).\n\nThe correlation analysis showed that the leukocyte concentration was positively correlated with the concentrations of IL-1\u03b2 (r\u2009=\u20090.959, p\u2009\\<\u20090.001, [Fig. 2A](#f2){ref-type=\"fig\"}) and TNF-\u03b1 concentration (r\u2009=\u20090.958, p\u2009\\<\u20090.001, [Fig. 2B](#f2){ref-type=\"fig\"}), and the platelet concentration was significantly positively correlated with the concentrations of PDGF-AB (r\u2009=\u20090.908, p\u2009\\<\u20090.001, [Fig. 2C](#f2){ref-type=\"fig\"}) and TGF-\u03b21 in the whole blood, L-PRP and P-PRP (r\u2009=\u20090.948, p\u2009\\<\u20090.001, [Fig. 2D](#f2){ref-type=\"fig\"}).\n\nEffects of PRPs on rBMSC\n------------------------\n\nThe trilineage differentiation and surface marker identification of rBMSC are shown in [Figure S1](#S1){ref-type=\"supplementary-material\"}(See the [supplementary file](#S1){ref-type=\"supplementary-material\"}). The Cell Counting Kit-8 (CCK-8) assay demonstrated that all groups presented good proliferation with no significant difference on the first day of culture (p\u2009\\>\u20090.01), while PRPs had better effects on rBMSC proliferation after 3, 5 and 7 days of culture than the fetal bovine serum (FBS) group. But the P-PRP group showed better effects on rBMSC proliferation than the L-PRP group after culture for 5 and 7 days (p\u2009\\<\u20090.01, [Fig. 3A](#f3){ref-type=\"fig\"}).\n\nWestern blot analysis showed that PRP groups had increased protein levels of Col II, Aggrecan and Sox9 compared with the FBS group, although the level of up-regulation was not as high as that in the CDK group. The results also showed that the expression levels were higher in the P-PRP group than those in the L-PRP group. The protein level of Col I was higher in the FBS group compared with other groups, and the L-PRP group showed higher Col I protein expression than the P-PRP group, while the CDK group showed the lowest expression ([Fig. 3B](#f3){ref-type=\"fig\"}).\n\nThe chondrogenic-related marker genes, including collagen type II (Col II), Aggrecan and Sox9, mRNA expression was significantly increased expression in the PRP groups than that in the FBS group. In addition, mRNA expression was significantly higher in the P-PRP group than in the L-PRP group, with the highest expression observed in the commercially available chondrogenesis differentiation kit (CDK) group (p\u2009\\<\u20090.01, [Fig. 3D--F](#f3){ref-type=\"fig\"}). The mRNA expression of collagen type I (Col I) was inhibited in the CDK group but increased in the PRP groups (p\u2009\\<\u20090.01, [Fig. 3C](#f3){ref-type=\"fig\"}).\n\nEffects of PRPs on NF-\u03baB pathway\n--------------------------------\n\nAfter culture for 1\u2009hour, the immunofluorescence staining showed that NF-\u03baB p65 was primarily located in the cytoplasm of cells treated with culture medium containing 10% FBS or P-PRP, but NF-\u03baB p65 was located in the nuclei of cells treated with culture medium containing 10% L-PRP ([Fig. 4A](#f4){ref-type=\"fig\"}). Western blot analysis revealed that nuclear NF-\u03baB p65 protein expression increased in the L-PRP group ([Fig. 4B](#f4){ref-type=\"fig\"}).\n\nThe qRT-PCR analysis showed that the L-PRP group showed higher inducible nitric oxide synthase (iNOS) and Cyclooxygenase-2 (COX-2) mRNA expression than the P-PRP and FBS groups (p\u2009\\<\u20090.01, [Fig. 4C and D](#f4){ref-type=\"fig\"}). The results of the nitric oxide (NO) production test demonstrated that the L-PRP group produced more NO than the P-PRP and FBS groups, and enzyme-linked immunosorbent assay (ELISA) indicated that the level of Prostaglandin E2 (PGE2) was significantly higher in the L-PRP group than in the P-PRP and FBS groups (p\u2009\\<\u20090.01, [Fig. 4E and F](#f4){ref-type=\"fig\"}). Moreover, the results indicated no significant difference between the P-PRP group and the FBS group ([Fig. 4C--F](#f4){ref-type=\"fig\"}).\n\nMacroscopic observations of cartilage repair\n--------------------------------------------\n\nAt 6 weeks after surgery, macroscopic observation of the full-thickness cartilage defect sites appeared glossy white, and mostly well-integrated regenerated tissue in the P-PRP group, but the surface was not as smooth as the normal articular surface. Defects in the L-PRP group were filled with regenerated bone tissue halfway covered with whitish cartilage-like tissue. Defects in the control group were concave, filled with fibrous-like tissue, while the joints were slightly degenerated ([Fig. 5A--C](#f5){ref-type=\"fig\"}).\n\nAt 12 weeks postoperatively, the regenerated tissue seemed similar in the P-PRP group and the L-PRP group. Specifically, the regenerated tissue looks more opaque and closer to the adjacent cartilage tissue in the P-PRP group than that in the L-PRP group. In the control group, the defects were partially concave, filled with fibrous tissue; meanwhile the joints were moderately degenerated ([Fig. 5D--F](#f5){ref-type=\"fig\"}).\n\nMicro-computed tomography (Micro-CT) scanning analysis\n------------------------------------------------------\n\nMineralized bone formation was assessed using Micro-CT scanning. Non-treated rabbits were used as a positive control, and representative pictures are shown in [Fig. 6A](#f6){ref-type=\"fig\"}. The Micro-CT parameter analysis showed that the values of BV/TV (bone volume/tissue volume) ratio, trabecular number, and connectivity density were significantly higher in the PRP and positive control groups compared with the control group at week 12 (p\u2009\\<\u20090.05, [Fig. 6B](#f6){ref-type=\"fig\"}). The values of trabecular pattern factor and structure model index in the control group were significantly higher than those in the other three groups. However, these values in the positive control group were significantly lower than those in the PRP groups (p\u2009\\<\u20090.05, [Fig. 6B](#f6){ref-type=\"fig\"}). However, the value of all 5 parameters was not significantly different between the L-PRP and P-PRP groups (p\u2009\\<\u20090.05, [Fig. 6B](#f6){ref-type=\"fig\"}).\n\nHistological findings\n---------------------\n\nThe H&E staining showed that, at 6 weeks after surgery, the defects in the control group were primarily filled with fibrous tissue and remained concave ([Fig. 7A and B](#f7){ref-type=\"fig\"}); in the L-PRP group, the articular surface was partially filled with fibrous tissue, with moderate integration and splitting ([Fig. 7C and D](#f7){ref-type=\"fig\"}); in the P-PRP group, the joint surface was integrated relatively well with the regenerated tissue in the defects, with subchondral bone formation at the bottom and regenerated round cells on the top ([Fig. 7E and F](#f7){ref-type=\"fig\"}). At 12 weeks postoperativel, the joint surface was still concave in the defects covered with fibrous tissue in the control group, and little new bone formation was observed ([Fig. 7G and H](#f7){ref-type=\"fig\"}); in the L-PRP group, more cartilage-like tissue formation was observed, but the new regenerated tissue was thinner compared with the surrounding normal cartilage ([Fig. 7I and J](#f7){ref-type=\"fig\"}); the defects in the P-PRP group were fully filled with cartilage-like tissue, and well integrated with the normal cartilage, although the amount of regenerated cells was not as abundant as observed in normal cartilage ([Fig. 7K and L](#f7){ref-type=\"fig\"}).\n\nThe toluidine blue staining showed that, at 6 weeks after surgery, there was few cartilage matrix in the control group and much more in the PRP groups ([Fig. 8A--F](#f8){ref-type=\"fig\"}). However, a wide split was observed in the L-PRP group and a relatively narrow one was observed in the P-PRP group. Moreover, more cartilage matrix was observed in the P-PRP group compared with the L-PRP group ([Fig. 8C--F](#f8){ref-type=\"fig\"}). At 12 weeks, new bone formation was clearly better than that at 6 weeks in the control group, but few cartilage matrix was observed ([Fig. 8G and H](#f8){ref-type=\"fig\"}); a narrow split was still observed in the L-PRP group, but it was almost not seen in the P-PRP group ([Fig. 8I--L](#f8){ref-type=\"fig\"}); in addition, there was much more cartilage matrix in the P-PRP group compared with the L-PRP group.\n\nImmunohistochemical staining at 12 weeks showed that the content of Col I decreased in the regenerated tissue, while the content of Col II simultaneously increased in the following sequence: the control group, the L-PRP group and the P-PRP group ([Fig. 9](#f9){ref-type=\"fig\"}).\n\nInternational Cartilage Repair Society (ICRS) scoring for macroscopic evaluation of cartilage repair showed that the scores in the PRP groups were higher than that in the control group at week 6 and week 12. There was no significant difference between the L-PRP and P-PRP groups at week 6 or at week 12 (p\u2009\\>\u20090.05, [Fig. 10A](#f10){ref-type=\"fig\"}).\n\nOsteoarthritis Research Society International (OARSI) scoring system was used to quantitatively analyze the histological evaluation. The results showed that the scores in the PRP groups were lower than that in the control group, and the scores in the P-PRP were lower than that in the L-PRP group at week 6 and 12 (p\u2009\\<\u20090.05, [Fig. 10B](#f10){ref-type=\"fig\"}). A low score means favorable outcome according to the OARSI scoring system.\n\nDiscussion\n==========\n\nThe present study evaluated and compared the effects of L-PRP and P-PRP on cartilage regeneration and the NF-\u03baB pathway. The findings showed that both L-PRP and P-PRP had \\~6-fold platelet concentration and increased PDGF-AB and TGF-\u03b21 concentrations compared with the whole blood. P-PRP, which had a significantly lower concentrations of leukocytes, IL-1\u03b2 and TNF-\u03b1 than L-PRP, presented better effects on rBMSC proliferation and chondrogenesis than L-PRP. In addition, L-PRP induced the nuclear translocation of NF-\u03baB p65, increased the mRNA expression of iNOS and COX-2, and enhanced the production of NO and PGE2 in rBMSC, indicating that L-PRP strongly activated the NF-\u03baB pathway. Moreover, rBMSC-containing P-PRP-derived scaffolds presented better cartilage healing results than rBMSC-containing L-PRP-derived scaffolds *in vivo*.\n\nPRP treatment has become widespread in sports medicine due to the use of patient's own blood product, ease of preparation, and beneficial effects on tissue repair[@b24][@b25][@b26][@b27][@b28]. However, the optimal method for preparing PRP for the treatment in cartilage repair is unknown, and different methods may result in different platelet and leukocyte concentrations, potentially affecting cell responses[@b29]. PDGF induces articular cartilage anabolism, chondrocyte proliferation and stem cell migration to the joint's injured tissues[@b30], and TGF-\u03b21 increases the cartilage extracellular matrix (ECM) synthesis, diminishes joint inflammation and promotes the differentiation of synovial membrane stem cells into chondrocytes[@b31]. In the present study, PRPs released more growth factors, approximately 6-fold higher than the baseline level. From the third day, rBMSC cultured with PRPs presented better proliferation than the FBS group, reflecting the increased concentration of growth factors. Shapiro *et al*. reported that chondrocytes from the native cartilage did not participate in the repopulation of the defect, and that the cartilaginous repair appeared to be mediated by the proliferation and differentiation of mesenchymal cells from the bone marrow[@b32]. TNF-\u03b1 was selected as a pro-inflammatory cytokine, and this protein along with IL-1\u03b2 are key cytokines associated with the catabolic state observed in OA[@b33], and leukocytes are the primary source of inflammatory cytokines such as IL-1 and TNF-\u03b1[@b34][@b35][@b36][@b37][@b38]. L-PRP contained more leukocytes, TNF-\u03b1 and IL-1\u03b2 than P-PRP in the present study. The cell proliferation assay showed that the L-PRP group had less cells than the P-PRP group after 5 and 7 days of culture, likely reflecting the harmful effects of TNF-\u03b1 and IL-1\u03b2. This finding is consistent with the study of Anitua *et al*.[@b35], showing that fibroblasts respond differently to various PRP formulations.\n\nRecent *in vitro* evidence suggests that P-PRP treatment could be more suitable for tendon[@b17][@b39] and joint[@b40][@b41][@b42] injuries than L-PRP, as the lower leukocyte concentrations in P-PRP induce less tissue catabolism/inflammation than L-PRP treatment. In the present study, the difference of platelet concentrations between P-PRP and L-PRP was not significant, but rBMSC treated with L-PRP expressed less chondrogenesis-related genes, such as Col II, Aggrecan and Sox9, and their respective proteins compared with with P-PRP. Thus, the different effects of PRPs on the chondrogenesis of rBMSC may result from different leukocyte concentrations. In the present study, the nuclear translocation of NF-\u03baB p65 was observed when rBMSC were cultured with L-PRP, and increased mRNA expression levels of iNOS, COX-2 and production levels of PGE2 and NO were determined, showing that the NF-\u03baB pathway was highly activated and may impede the chondrogenesis and anabolism of rBMSC. It was reported that PRP releasate inhibited the inflammatory process in osteoarthritic chondrocytes through the reduction of IL-1\u03b2-induced NF-\u03baB activation[@b43]; thus, the NF-\u03baB pathway plays an important role in the cartilage repair. Pereira *et al*. reported that PRP lysate could contribute to the down-modulation of the NF-\u03baB pathway and the COX-2 expression, thereby triggering the resolution of inflammatory responses[@b44]. However, neither platelet lysate nor PRP clot releasate contains viable leukocytes or platelets, while the excess of leukocytes may overwhelm the ability of growth factors on modulating pro-inflammatory cytokines[@b45]. Therefore, these studies may not reflect the effects of PRP on the NF-\u03baB pathway. The findings of the present study demonstrated that L-PRP promoted the NF-\u03baB pathway activation, which we reasonably ascribe it to the leukocytes in L-PRP. Leukocytes were significantly associated with an increased concentration of catabolic cytokines in PRP, including collagen-degrading MMPs, reactive oxygen species, and other catabolic proteases[@b46]. MMP-9 has been demonstrated to be a component of non-healing or poorly healing wounds[@b47] and it was of high concentration in L-PRP, positively correlated with leukocytes[@b46].\n\nThe harmful effect of leukocytes in PRP on chondrogensis *in vitro* may not occur *in vivo*. However, there has been no *in vivo* study that compares the cartilage repair in rabbit osteochondral defects between L-PRP and P-PRP. In our study, the rBMSC-containing P-PRP-derived scaffolds yielded better histological results than the rBMSC-containing L-PRP-derived scaffolds at 6 and 12 weeks after surgery. In addition, the interface of the joint surface in the PRP groups showed better integration compared with the control group, and that in the P-PRP group presented great integration. Valderrabano *et al*. reported that that chondral damage and poor cartilage integration promotes synovial fluid inflow, leading to cystic change seen in clinical studies[@b48]. Therefore it is important of the interface integration for the cartilage repair. PRP may enhance cartilage repair at the graft interface through recruitment of mesenchymal stem cells[@b49]. The introduction of rBMSC in PRP scaffolds may thus facilitate the recruitment efficiency.\n\nHowever, the present study has certain limitations. Initially, we only preliminarily explored the mechanism of the different effects of L-PRP and P-PRP on rBMSC *in vitro* experiments, and the precise principle was not confirmed. Second, we evaluated the *in vivo* effects at 6 and 12 weeks, but did not observe the long-term effects. To solve the problems mentioned above, further studies that evaluate the effects of L-PRP and P-PRP on the NF-\u03baB pathway *in vivo* are needed, and the observation period of *in vivo* studies should be prolonged. Third, the rabbit model of cartilage defects was not the same as the human cartilage lesion. The defects were created in the non-weight-bearing sites in our study, while most cartilage lesions occurred in the weight-bearing sites in clinics. Thus, we did not investigate the mechanical properties in this study.\n\nIn conclusion, the results suggest that both P-PRP and L-PRP can exert a beneficial effect on the cartilage repair, but P-PRP promotes better rBMSC in chondrogenic differentiation *in vitro* and *in vivo* compared with L-PRP. These findings also revealed that the better behavior of P-PRP in cartilage regeneration reflects the removal of the leukocytes reduces the concentrations of IL-1\u03b2 and TNF-\u03b1 and inhibits the activation of the NF-\u03baB pathway, which affects the repair of cartilage defects. Thus, P-PRP treatment may present a better alternative for cartilage repair.\n\nMaterials and Methods\n=====================\n\nThe Independent Ethics Committee and the Animal Care and Use Committee of Shanghai Jiao Tong University Affiliated Sixth People's Hospital approved the protocols used in the present study. All methods were performed in accordance with the relevant guidelines and regulations.\n\nPreparations of rBMSC, L-PRP, P-PRP and transplanted constructs\n---------------------------------------------------------------\n\nrBMSC were harvested from 6-week-old New Zealand white rabbits as per the Cao *et al*. procedure[@b50]. In brief, bone marrow was aspirated from each side of the iliac crests, anticoagulated with 1000\u2009U/mL preservative-free heparin, and filtered with a 70-mm filter mesh. The bone marrow was subsequently suspended in the \u03b1-modification of minimum essential medium (\u03b1-MEM, Sigma-Aldrich, St Louis, MO, USA) containing 10% FBS (Gibco, Carlsbad, CA, USA), 100\u2009U/mL penicillin G (Gibco) and 0.1\u2009mg/mL streptomycin (Gibco) at 37\u2009\u00b0C in a humidified atmosphere with 5% CO~2~. The medium was changed after 48\u2009hours to remove non-adherent cells and thereafter once every three days. The multi-potency characterization and surface marker identification of rBMSC were performed according to a previous study[@b51][@b52]. Cells at the fifth passage were used in the present study.\n\nApproximately 18\u2009mL of autologous whole blood was collected from each New Zealand white rabbit through the central auricular artery and anticoagulated with 2\u2009mL acid-citrate dextrose solution A (ACD-A) to make 20\u2009mL of anticoagulated whole blood. Approximately 19\u2009mL of anticoagulated whole blood was used in the preparation of PRPs, and 1\u2009mL of whole blood was retained to quantify the platelet, leukocyte and cytokine concentrations in whole blood.\n\nL-PRP was prepared using a buffy coat-based double-spin centrifugation method[@b53]. Briefly, 8\u2009mL of autologous whole blood was transferred to a 15-mL centrifuge tube and centrifuged at 250\u2009*g* for 10\u2009minutes at room temperature to separate the blood into three phases: erythrocytes at the bottom, buffy coat in the middle (rich in platelets, leukocytes and fibrinogen), and platelet-containing plasma at the top. Platelet-containing plasma and buffy coat were subsequently transferred to a new tube and centrifuged again at 1000\u2009*g* for 10\u2009minutes, during which most of the platelets, leukocytes and fibrinogen precipitated. The supernatant plasma was discarded and the precipitated platelets were resuspended in the residual plasma to obtain a total of 1\u2009mL of L-PRP.\n\nP-PRP was prepared using a plasma-based method developed in the laboratory that concentrates platelets and eliminates leukocytes and erythrocytes. In brief, 11\u2009mL of anticoagulated whole blood was centrifuged at 160\u2009g for 10\u2009minutes at room temperature to separate platelet-containing plasma from erythrocytes and buffy coat (rich in leukocytes). Care was taken to prevent the buffy coat and erythrocytes contamination. Platelet-containing plasma was subsequently transferred to a new tube and centrifuged at 250\u2009*g* for 15\u2009minutes. The supernatant plasma was discarded and the precipitated platelets were resuspended in the residual plasma to obtain a total of 1\u2009mL of P-PRP.\n\nThe PRP scaffolds and transplanted constructs were prepared as per the Xie *et al*. method[@b12]. In brief, 0.5\u2009mL of cell culture medium was added into 0.5\u2009mL of the whole blood, L-PRP, or P-PRP to form blood clot or PRP scaffolds within minutes. To prepare the transplanted constructs, rBMSC of the third passage were detached, centrifuged and resuspended in cell culture medium at a concentration of 2.0\u2009\u00d7\u200910^6^ cells/mL. Subsequently, 0.5\u2009mL of the cell culture medium (containing rBMSC) was mixed with PRPs to prepare the PRP scaffolds, and the rBMSC-containing PRP scaffolds were used as transplanted constructs.\n\nQuantification of platelets, leukocytes, and released cytokines in whole blood and PRPs\n---------------------------------------------------------------------------------------\n\nThe whole blood analyses were performed using an automatic hematology analyzer (XS-800i, Sysmex, Kobe, Japan) to determine the platelet and leukocyte concentrations of whole blood and PRPs.\n\nThe PRP scaffolds and blood clots were prepared as described above, followed by incubation at 37\u2009\u00b0C in a humidified atmosphere with 5% CO~2~ for 7 days, and centrifugation at 2800\u2009*g* for 15\u2009minutes to collect 0.5\u2009mL of supernatant[@b54]. The collected supernatant was assayed to determine the concentrations of IL-1\u03b2, TNF-\u03b1, PDGF-AB, and TGF-\u03b21 using ELISA according to the manufacturer's instructions (Xitang, Shanghai, China).\n\nCell proliferation assay\n------------------------\n\nrBMSC were seeded onto 96-well plates at a density of 4,000 cells/well for 20\u2009hours in \u03b1-MEM without FBS to allow the cells to adhere, and subsequently the cells were transferred to a medium supplemented with 10% (volume/volume) of FBS, L-PRP, or P-PRP. The concentrations of PRP were determined based on previous studies[@b12]. The proliferation of rBMSC was assessed after 1, 3, 5, and 7 days using CCK-8 (Dojindo, Kumamoto, Japan). Briefly, 10\u2009\u03bcL of CCK-8 solution was added to each well containing 100\u2009\u03bcL of medium and incubated for 3\u2009hours; blank medium plus an equal amount of CCK-8 was set as the background. The absorbance value was measured using a microplate reader (Bio-Rad, Hercules, CA, USA) at 450\u2009nm.\n\n*In vitro* chondrogenesis induction assessment\n----------------------------------------------\n\nThe *in vitro* chondrogenesis induction assessment of rBMSC was conducted as per Xie *et al*. and Zhu *et al*. procedure[@b12][@b55]. In brief, approximately 500,000 rBMSC were seeded in 25-cm^2^ flasks and cultured in \u03b1-MEM supplemented with 10% (volume/volume) of FBS, L-PRP, or P-PRP, or cultured in a commercially-available chondrogenesis differentiation kit (CDK, Invitrogen, Rockford, IL, USA). The medium was changed once in every three days in all the cultures. Three weeks later, the cells were collected for quantitative real-time polymerase chain reaction (qRT-PCR) analysis and Western blotting analysis to detect the messenger RNA (mRNA) and protein expression, respectively, of Col I, Col II, Aggrecan and Sox9.\n\nTotal RNA was extracted from rBMSC using TRIzol reagent (Invitrogen, Rockford, IL, USA), and measured for quantity and purity using NanoDrop 2000 (Thermo Fisher Scientific, Rockford, IL, USA) according to the manufacturer's instructions. Reverse transcription was performed using the High Capacity Reverse Transcription kit (Invitrogen). qRT-PCR was subsequently performed using a TP800 system (Takara, Japan) with SYBR Premix Ex Taq (Takara). \u03b2-Actin was used as a housekeeping gene for normalizing the results of the study. All the data were analyzed using the \u0394\u0394Ct method[@b56]. The sequences of the primers used are listed in [Table 1](#t1){ref-type=\"table\"}.\n\nrBMSC were lysed using mammalian protein extraction reagent (Pierce, Rockford, IL, USA) supplemented with complete protease inhibitor. The total protein concentration was quantified using BCA Protein Assay Kit (Pierce) according to manufacturer's instructions. The total protein was denatured at 95\u2009\u00b0C for 5\u2009minutes. 100\u2009\u03bcg of total protein was assessed using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), and the proteins were subsequently transferred to a polyvinylidene fluoride membrane (PVDF, Millipore, Billerica, MA, USA), which was blocked with a low-fat milk protein solution. The membranes were incubated with anti-Sox9 antibody (Abcam Cambridge, MA, USA), anti-Aggrecan antibody (Abcam), anti-Col I (Merck, Whitehouse Station, NJ, USA), anti-Col II (Merck), or anti-GAPDH antibody (Cell Signaling Technologies, Danvers, MA, USA), followed by peroxidase-conjugated secondary antibodies. The blots were then subjected to chemiluminescence detection using ECL western blotting substrate (Pierce).\n\nEvaluation of the effects of PRPs on NF-\u03baB pathway *in vitro*\n-------------------------------------------------------------\n\nrBMSC were seeded onto cell culture plates at the appropriate density and cultured in \u03b1-MEM supplemented with 10% FBS until a confluent layer was achieved. rBMSC were subsequently serum-starved for 24\u2009hours and treated in medium supplemented with 10% of FBS, L-PRP, or P-PRP for either 1\u2009hour (immunofluorescence staining) or 48\u2009hours (all other experiments).\n\nrBMSC were seeded onto a chamber slide (ibidi, Martinsried, Germany) at a density of 30,000 cells/mL and treated as described above. Subsequently, rBMSC were fixed with 4% paraformaldehyde for ten minutes, treated with 3% Triton-X for 10\u2009minutes, blocked with 10% FBS for 2\u2009hours, incubated with primary antibody (anti-NF-\u03baB p65 antibody, Abcam) at 4\u2009\u00b0C overnight, and incubated with Alexa Fluor-conjugated secondary antibody (Abcam) for 2\u2009hours. The cell nuclei were counterstained with DAPI (Life Technologies, Rockford, IL, USA), and subsequently washed subsequently with PBS. The nuclear translocation of NF-\u03baB p65 was observed using a Leica DMI6000 B with Leica AF6000 software.\n\nrBMSC were lysed to extract total RNA and the mRNA expression was analyzed for iNOS and COX-2 using qRT-PCR according to the methods described above. The primer sequences for the target genes are listed in [Table 1](#t1){ref-type=\"table\"}.\n\nNuclear protein extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents Kit (Thermo Fisher Scientific, Rockford, IL, USA) according to the manufacturer's instructions. Western blotting was subsequently performed as described above using an anti-NF-\u03baB p65 antibody (Abcam). The amount of PGE2 released into the medium was quantified using an ELISA kit (Xitang). The effects of PRPs scaffolds on NO production were measured using a NO Assay Kit ((Thermo Fisher Scientific, Rockford, IL, USA).\n\nAnimal surgery\n--------------\n\nTwenty-seven male mature New Zealand white rabbits (weighing 2.5--3.0\u2009kg) were enrolled in the present study. After anesthetization achieved, a lateral para-patellar skin was incised and the knee joint was exposed when the joint capsule was sliced open and the patella was extracted laterally. A full-thickness cylindrical cartilage defect of 5\u2009mm in diameter and 3\u2009mm in depth was generated in the patellar groove using a stainless steel drill. The 48 defects in 24 rabbits were left unfilled (the control group), or randomly filled with transplanted constructs prepared with L-PRP (the L-PRP group) or P-PRP (the P-PRP group). The remaining 3 rabbits, which were used as a positive control for Micro-CT analysis and as a reference in histological scoring assessments, were also surgically treated but without any defects creation. All the rabbits were housed in separate cages and allowed to move freely immediately after surgery. Then 12 rabbits were euthanized after 6 weeks postoperatively and the remaining 15 rabbits were euthanized after 12 weeks postoperatively to harvest the distal part of the femur. The samples were fixed with 4% paraformaldehyde for 72\u2009h and subsequently photographed.\n\nGross morphology\n----------------\n\nThe rabbits were sacrificed after 6 and 12 weeks postoperatively. Femoral condyle samples were dissected and fixed with 4% paraformaldehyde for 72\u2009h, and then photographed and evaluated according to the ICRS macroscopic assessment scores for cartilage repair[@b57].\n\nMicro-CT scanning\n-----------------\n\nMicro-CT (Skyscan, 1076 scanner, Kontich, Belgium) scanning was performed to evaluate the subchondral bone regeneration in the defects of samples harvested after 12 weeks postoperatively. The samples were fixed for 72\u2009h and immobilized with the femoral axis perpendicular to the scanning plane. The image data was reconstructed using NRecon software (Version 1.5.1.4, Skyscan) to visualize the 3D representation of the regenerated bone. Based on the CT data, a cylindrical region of interest (ROI) was analyzed which corresponds to the original defect location. For comparison between groups, the extent of bone regeneration within the defect was presented as a BV/TV ratio. The 3 rabbits without osteochondral defects were scanned as a positive control.\n\nHistological analysis\n---------------------\n\nThe samples were decalcified 10% EDTA for 1 month at 37\u2009\u00b0C and then embedded in paraffin and cut into 5-\u03bcm sections that were then stained with hematoxylin and eosin (H&E) for general histological evaluation and with Toluidine blue for cartilaginous matrix distribution evaluation. The regenerated tissue was graded and evaluated by three trained observers according to the OARSI scale.\n\nThe expression of Col I and Col II was analyzed by immunohistochemical staining. The sections were dewaxed in xylene and hydrated through graded ethanol solutions, followed by blocking with 1% bovine serum albumin. Then the sections were incubated with the primary antibodies (mouse clone, 1:1000; Merck) at 4\u2009\u00b0C overnight. The secondary anti-mouse IgG was added to the sections after washing three times with PBS and incubated for 1\u2009h at 37\u2009\u00b0C. The staining was developed in diaminobenzidine solution, with hematoxylin counterstaining.\n\nStatistical analysis\n--------------------\n\nAll experiments *in vitro* were repeated 3 times. The data were analyzed using the Statistical Package for Social Sciences version 22.0 (SPSS, IL, USA). The data are presented as the means\u2009\u00b1\u2009standard deviation (SD). One-way analysis of variance (ANOVA) and Bonferroni's post hoc test, or independent-samples Student's t test was performed for statistical analysis. Statistical significance was indicated at *p*\u2009\\<\u20090.01 in *in vitro* tests or *p*\u2009\\<\u20090.05 in *in vivo* tests.\n\nAdditional Information\n======================\n\n**How to cite this article**: Xu, Z. *et al*. Comparative evaluation of leukocyte- and platelet-rich plasma and pure platelet-rich plasma for cartilage regeneration. *Sci. Rep.* **7**, 43301; doi: 10.1038/srep43301 (2017).\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nThis study was supported by the National Natural Science Foundation of China (Grant No. 81401799) and Shanghai Youth Science & Technology Start-up Grant (No. 14YF1412100).\n\nThe authors declare no competing financial interests.\n\n**Author Contributions** X.Z.L. and Y.W.J. was responsible for the conduct of all experiments data collection and analyses, and manuscript preparation; Q.X. and C.Y.X. contributed to animal experimentation and review of manuscript; Z.Y.L. contributed to data analysis; Z.C.Q. and X.X.T. were responsible for the design of experimentation, project supervision and coordination, data analyses, and manuscript preparation.\n\n![Cellular and cytokine concentrations in WB, L-PRP and P-PRP.\\\nConcentrations of leukocytes (**A**), IL-1\u03b2 (**B**), and TNF-\u03b1 (**C**) in P-PRP were significantly lower than those in L-PRP and WB; concentrations of platelets (**D**), PDGF-AB (**E**), and TGF-\u03b21 (**F**) in L-PRP and P-PRP were similar, but significantly higher than those in WB. \\*p\u2009\\<\u20090.01 compared with WB; ^\\$^p\u2009\\<\u20090.01 L-PRP versus P-PRP; n\u2009=\u200910.](srep43301-f1){#f1}\n\n![Correlations between cytokine concentrations and cellular components in WB, L-PRP and P-PRP.\\\nPositive correlations were observed between the concentrations of leukocytes and those of IL-1\u03b2 (**A**) and TNF-\u03b1 (**B**), and between the concentrations of platelets and those of PDGF-AB (**C**) and TGF-\u03b21 (**D**) in WB, L-PRP and P-PRP; n\u2009=\u200930.](srep43301-f2){#f2}\n\n![PRPs promoted proliferation and chondrogenetic differentiation of rBMSC.\\\n(**A**) CCK-8 assay showed that both L-PRP and P-PRP promoted rBMSC proliferation on day 3, 5 and 7 compared with FBS, but P-PRP showed greater effects on rBMSC proliferation on day 5 and 7 compared with L-PRP; (**B**) western blotting analysis revealed that P-PRP upregulated protein expression of SOX-9, Aggrecan, and Col II, and downregulated protein expression of Col I, compared with L-PRP and FBS; (**C--F**), qRT-PCR analysis demonstrated that both L-PRP and P-PRP upregulated mRNA expression of SOX-9, Aggrecan, Col II, and Col I compared with FBS, but P-PRP showed greater effects on mRNA expression of chondrogenic-related marker genes (SOX-9, Aggrecan, and Col II) and weaker effects on mRNA expression of osteogenic-related marker gene (Col I) compared with L-PRP. \\*p\u2009\\<\u20090.01 compared with WB; ^\\#^p\u2009\\<\u20090.01 compared with CDK; ^\\$^p\u2009\\<\u20090.01 L-PRP versus P-PRP; n\u2009=\u20093.](srep43301-f3){#f3}\n\n![L-PRP induced the activation of NF-\u03baB pathway.\\\n(**A**) Immunofluorescence staining revealed that NF-\u03baB p65 was observed in the nucleus in the L-PRP group, and in the cytoplasm in the P-PRP and FBS groups; (**B**) western blotting analysis showed that L-PRP upregulated protein expression of NF-\u03baB p65 in the nucleus compared with P-PRP and FBS; (**C** and **D**) qRT-PCR analysis showed that the mRNA expression levels of iNOS and COX-2 were significantly higher in the L-PRP group than those in the P-PRP and FBS groups; (**E** and **F**) Production of PGE2 and NO was significantly enhanced in the L-PRP group in comparison with that in the P-PRP and FBS groups. \\*p\u2009\\<\u20090.01 compared with WB; ^\\$^p\u2009\\<\u20090.01 L-PRP versus P-PRP; n\u2009=\u20093.](srep43301-f4){#f4}\n\n![Effects of PRPs on cartilage repair in rabbits --- gross observation.\\\nMacroscopic appearance of the cartilage healing in the control, L-PRP and P-PRP groups at 6 weeks (**A--C**), and 12 weeks (**D--F**) after implantation. n\u2009=\u20096.](srep43301-f5){#f5}\n\n![Effects of PRPs on osteochondral repair in rabbits --- quantitative analysis of regenerated bone tissue using a 3-D Micro-CT model.\\\n(**A**) Region of interest (ROI) corresponding to the original defects in rabbits. The green colour represents regenerated bone, and the yellow colour represents other tissues in the original defect (top: 3-D models *in situ*; bottom: 3-D models *ex situ*). (**B**) Comparison of the bone regeneration in the 4 groups using the ratio of BV/TV, trabecular number, trabecular pattern factor, structure model index and connectivity density at 12 weeks. Non-treated rabbits were used as a positive control. BV, bone volume; TV, tissue volume; \\*p\u2009\\<\u20090.05 compared with the control group; ^\\#^p\u2009\\<\u20090.05 compared with the positive control group; n\u2009=\u20096.](srep43301-f6){#f6}\n\n![Effects of PRPs on cartilage repair in rabbits --- H&E staining.\\\nRepresentative H&E stained sections in the control, L-PRP and P-PRP groups at 6 weeks (**A--F**) and 12 weeks (**G--L**). The rectangular squares demonstrated the conjunction between the regenerated tissue (right half) and the normal tissue (left half). n\u2009=\u20096.](srep43301-f7){#f7}\n\n![Effects of PRPs on cartilage repair in rabbits ---toluidine blue staining.\\\nRepresentative toluidine blue stained sections in the control, L-PRP and P-PRP groups at 6 weeks (**A--F**) and 12 weeks (**G--L**). The rectangular squares demonstrated the conjunction between the regenerated tissue (right half) and the normal tissue (left half). n\u2009=\u20096.](srep43301-f8){#f8}\n\n![Effects of PRPs on cartilage repair in rabbits --- immunohistochemical staining.\\\nRepresentative sections using Col I (**A--F**) and Col II (**G--L**) staining in the control, L-PRP and P-PRP groups at 12 weeks. The rectangular squares demonstrated the conjunction between the regenerated tissue (right half) and the normal tissue (left half). n\u2009=\u20096.](srep43301-f9){#f9}\n\n![Effects of PRPs on cartilage repair in rabbits --- macroscopic and histological scoring.\\\n(**A**) macroscopic ICRS scores; (**B**) histological OARSI scores. \\*p\u2009\\<\u20090.05 compared with the control group; ^\\$^p\u2009\\<\u20090.05 L-PRP versus P-PRP; n\u2009=\u20096.](srep43301-f10){#f10}\n\n###### The primers sequences used for qRT-PCR.\n\n  Genes       Forward primer sequence (5\u2032-3\u2032)   Reverse primer sequence (5\u2032-3\u2032)\n  ---------- --------------------------------- ---------------------------------\n  Sox9           AAGGGCTACGACTGGACGCTGGTG            AGGGCCGCTTCTCGCTCTCG\n  Aggrecan        GTCTACAGAACAGCGCCATCATT           GCGAAGCAGTACACGTCATAGGT\n  Col II           TCGGCCTCCCTGGTATTGACG            GGAGGGCCCTGAGCACCATTGTT\n  Col I           AGCGTGGCCTACCTGGATGAAGC           ATGGGCGCGATGTCGGTGATGG\n  COX-2           CTTCACGCATCAGTTTTTCAAG          TCACCGTAAATATGATTTAAGTCCAC\n  iNOS              GCTGCCAAGCTGAAATTGA              GATAGCGCTTCTGGCTCTTG\n  \u03b2-Actin         GCATGGGCCAGAAGGACTCGTA            TCGCGGTTGGCCTTGGGGTTCA\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "\n"}
+{"text": "![](jove-48-2138-thumb)\n\nProtocol\n========\n\n1. Pre-requisites for Biomechanical Locomotion Analyses\n-------------------------------------------------------\n\nPrior to embarking upon purchasing expensive locomotor analysis equipment, and planning experiments where kinematics and/or kinetic locomotor assessment will be performed, it is imperative that the experimenter be familiar with the technical and practical aspects of biomechanical analyses, sensorimotor behavior, operant conditioning of animals, and handling/storing/managing large amounts of digitized data. Though these pre-requisites seem obvious to many, it is only after embarking upon these types of experiments where trainees realize the technical and practical complexities of performing relatively detailed locomotor analysis. The authors recommend that experimenters enroll in a course dealing with introductory biomechanics, be familiar with or hire someone familiar with a programming language required for data management, and of equal importance, spend substantial time interacting, handling, and working with laboratory animals. For understanding locomotion analysis and sensorimotor behavioral analysis in the neurosciences, experimenters are referred to several important references (see ^1,\\ 4-7^).\n\n2. Kinematic and Kinetic Testing Apparatus\n------------------------------------------\n\nA kinematic and kinetic testing apparatus, useful for collecting bilateral data, is comprised of the following components (see \\\"Table of Specific Reagents and Equipment\\\" for more detail):\n\n-   Quiet and sufficiently-sized room (not necessarily sound-proofed, though located in a low-traffic area)\n\n-   Radio to supply background \\\"noise\\\" to limit distracting noises from outside the room to animals\n\n-   Clear plexiglass, enclosed, flat-surface runway (180 cm long X 20 cm wide X 60 cm high) with a square hole (11 cm X 11 cm) located within the centre of the runway equidistant from either end\n\n-   Piece of plexi-glass (10.5 cm X 10.5 cm)\n\n-   Double-sided adhesive scotch tape\n\n-   3M reflective adhesive tape for skin markers\n\n-   Hole punch for creating skin markers\n\n-   Isofluorane anesthesia machine required for anesthetizing animals for shaving prior to placement of skin markers\n\n-   Cheerios for encouraging movement of the animal along the length of the runway\n\n-   AMTI HE6X6 forceplate\n\n-   Vicon Motus customized system (software, computer, etc)\n\n-   4 video cameras (at least 60 Hz sampling) and tripods with appropriate connector cables\n\n-   4 camera-mounted halogen lights provides optimal lighting for illumination of the animal subject and the reflective skin markers\n\n-   Appropriate sized calibration frame or volume necessary for Vicon Motus system to calibrate x, y, z positions into meaningful distance measures\n\n3. Animal Training\n------------------\n\nPrior to collecting data, each animal must be trained to cross a flat-surface, enclosed runway. Upon receiving rats from an appropriate animal supplier, animals should be acclimated to their new home for 1 week. During this acclimation time, several cheerios are placed daily into the rat\\'s cage. Animals are food restricted to their maintenance energy requirements to prevent obesity and ensure motivation to perform this task. Thereafter, each animal is handled by the experimenter for 10-15 minutes daily for 1 week. During this same time period, each animal is placed into the runway with cheerios located at either end. Once the animal becomes familiar with their environment, they will begin eating the cheerios. Once the animal is comfortable and eating cheerios within the runway, the experimenter must then operantly condition the animal to run the length of the runway for a food reward. This is accomplished by tossing \u00bc cheerio to the opposite end of the runway where the rat is positioned. Once the rat eats this cheerio, another \u00bc cheerio is placed at the other end of the runway. This is done for 15-20 minutes daily until the rat consistently (\\>90% of tosses) moves along the runway at a constant velocity (i.e. without starting, stopping, exploring, or without changing gait) to eat the cheerio without galloping/bounding. The rat should only be employing a trotting gait. Over-conditioning of the animals to this task can lead to animals galloping and bounding these gaits are indicative of animals traveling \\>90 cm/s. Bounding and galloping gaits, biomechanically, are more difficult to interpret for a variety of reasons (e.g. leg lead inclusion criteria, etc). In our experience, once rats consistently employ galloping or bounding gaits, it is difficult, if not impossible to have them use a trotting gait while locomoting in the runway. Velocities \\>90 cm/s are rarely seen after an animal has suffered from peripheral or central nervous system injury. Time to reach successful training is variable between strains and sexes of rats. Wistar, Lewis, Long-Evans, and Sprague-Dawley strains are able to consistently traverse the runway within 2 weeks from the onset of training ^8^. In our experience, Fischer (F-344) rats tend to take upwards of 4-6 weeks to learn this task ^8^.\n\n4. Joint Position Marking\n-------------------------\n\nForelimb kinematic analysis is unreliable due to skin movement artifact imposed by placing skin markers on the forelimbs which is exacerbated in species, like rats, that have a crouched posture 9. Instead, kinematics of the forelimbs must be achieved using x-ray cinematography or fluoroscopy ^10-13^. As such, hind limb joint position marker placement is only described herein.\n\nPrior to data collection, each rat must be anesthetized at least 24 hours in advance of data collection using an appropriate inhalational anesthetic (e.g. isofluorane, 1.5-2% dialed on a precision vaporizor) and administered in oxygen via face mask, and key topographic anatomical landmarks must be marked. Given the brevity of the procedure, and because long-acting anesthetic agents are not used, use of an animal warming device need not be used to maintain the animal\\'s body temperature. Once the animal is anesthetized, the hind limbs and the dorsum, to the level of the iliac crests are shaven. The animal is then placed in sternal recumbency and its hindlimbs are placed in an approximate standing position using firm packing foam to support it. The skin overlying the cranial-most portions of the iliac crests, the greater trochanter of the femur, the lateral tibial tuberosity, the tarsal joint, and the distal and lateral aspect of the 5^th^ metatarsal is marked with a non-toxic permanent marker. The animal is recovered from anesthesia. For temporal studies, periodic anesthesia may be required to shave the hindlimbs thereby permitting subsequent reflective marker placement (see below). Also, daily highlighting of the previously marked anatomical landmarks (using the same non-toxic marker) will be required as rats will slowly remove the markers through natural grooming behavior.\n\n5. Data Recording\n-----------------\n\nAll camera views are examined to ensure that their position is appropriate and capturing the same field of view. Each camera should be placed at approximately 60-80 degrees to each other. The field of view should include the forceplate in the centre and a length of runway sufficient to capture two strides.\n\nThe calibration volume is placed within the pre-determined area of the runway. A single frame of the calibration volume within the runway, from each of the cameras, is captured. All calibrated marks along the length of each of the poles are digitized. Only once a satisfactory error in digitizing is accomplished, can the experimenter proceed to collecting locomotor data. This calibration step is critical prior to collecting data. If calibration is not performed accurately, or if calibration does not occur immediately prior to a recording session, all resulting data will be inaccurate and unusable. Importantly, if any of the cameras are touched or moved, it is safest to assume that calibration of the system needs to be repeated.\n\nImmediately prior to placing the animal in the runway, its weight is recorded and pre-made conical reflective skin markers (using 3M reflective tape, see table) are adhered to the pre-determined felt marks made on the hind limb topographical landmarks. Recording the animal\\'s weight will permit retrospective normalization of ground reaction forces to body weight - an important aspect if one wishes to make comparisons between groups. Additionally, body weight measurement facilitates monitoring of the animal\\'s overall health for the duration of the experiment. Marker placement only requires appropriate animal handling and does not require anesthesia of the animal. If the adhesive on the reflective tape is insufficient to adhere to the animal\\'s skin, a very small amount of non-toxic glue (e.g. 3M VetBond Tissue Adhesive) can be used to facilitate adherence of the marker on the animal\\'s body. Once the markers are placed on the hindlimbs, the experimenter should be positioned comfortably near the keyboard of the computer and have in-hand the event marker attached through the Vicon Motus system. Using the calibrated file as a template, several files are made in advance of recording. Typically, 25 to 30 files need to be saved. Each file should be named uniquely. Each file will represent one recorded run of the animal being recorded. Twenty-five to 30 files are required to ensure that sufficient numbers of runs meeting the inclusion criterion velocity (i.e. 60-90 cm/s) are collected. After sufficient numbers of files are created the experimenter can begin collecting data. The experimenter must prompt the rat to shuttle within the runway by throwing \u00bc pieces of cheerios at either end of the runway. With careful coordination and timing, the event marker is triggered at initiation and just prior to completion of the rat successfully completing a pass along the runway. After examining the crude ground reaction force tracing, and after recording whether the left or the right limbs hit the force plate, the file is saved and closed. Equal numbers of left and right limb forceplate hits should be recorded. The process of recording the data from a given run is repeated until sufficient numbers of runs have been recorded.\n\n6. Data Analysis\n----------------\n\nUpon completing data collection for kinematics and ground reaction forces, each run from every animal needs to be evaluated for speed. Using two relatively fixed markers (e.g. wings of the iliums) one can evaluate a virtual point between the markers (done prior when creating a Vicon Motus file template). Before calculating the speed of this \\\"virtual\\\" point, markers for the iliac markers must be digitized. Using Vicon motus software, velocity of this virtual point in the X-direction (horizontal direction of movement) is calculated. In so doing, only runs within a given range speeds (determined *a priori*) are used in the final analyses. We find that animals moving between 60-90 cm/s are using consistent trotting gaits. A minimum of 10 runs (5 runs where the left limb makes contact with the force platform and 5 runs where the right limb makes contact with the force platform) are required. Once the acceptable runs are identified for each animal, digitization of the remaining skin markers must be completed.\n\nTo compensate for skin movement artifact over the knee, estimation of the knee position is calculated using triangulation (intersection between two circles 2D kinematics; or intersection between 2 spheres 3D kinematics), as has been previously described. Hip, knee, and hock joint angles, velocities, and accelerations can now be determined. Stance and swing times can also be evaluated, though their accuracy is limited based upon the sampling speed of the cameras being employed. These and other calculations can be performed directly (i.e. without export) using Vicon Motus KineCalc software, or data can be exported as ASCII data and analyzed using customizable routines in software such as MatLab.\n\nGround reaction force data is measured and amplified by the AMTI force platform and collected at 1200 Hz by Vicon Motus. As such, once the ground reaction force data is collected, an appropriate digital filter is applied to the data using Vicon Motus. Given that the experimenter has already determined the speed of travel and identified acceptable runs after digitizing appropriate skin markers, ground reaction force data that was collected simultaneously as the kinematic data, can be analyzed using Vicon Motus KineCalc directly, or indirectly using some other customizable software routine. A variety of variables for forces, in each of the three orthogonal directions, can be calculated. Such variables include peak force, area und the curve (i.e. impulse), etc. Importantly, however, the experimenter must keep right and left limb data for each run of each animal separate. Data extracted from right or left limbs is averaged for each animal and used as the representative data for that animal. Data is then analyzed using appropriate statistical procedures.\n\n7. Representative Results\n-------------------------\n\nTo represent the utility of this form of locomotor analysis, kinematics and ground reaction forces were determined for young, middle-aged, and geriatric female Wistar rats. From this analysis, age-related differences were found for female Wistar rats. In particular, ground reaction force analysis demonstrates that geriatric rats have reduced forelimb braking ability and tend to use their hindlimbs more for lateral stabilization compared to the other groups of animals (Figure 1). Kinematic analysis did not reveal any statistical differences between each group, though demonstrates that kinematics can be readily recorded from virtually any age of rat (Figure 2).\n\n**Figure 1.** Ground reaction force tracing taken from the left limbs of young (4 month old; n=7), middle-aged (13-14 months old; n=7), and geriatric (24 months old; n=5) female Wistar rats. Right limbs were similar. It is readily apparent that geriatric rats use their forelimbs less for braking (\\* = p\\<0.05) compared to young and middle-aged rats, and geriatric rats tend to use their hindlimbs more for lateral stabilization compared to young rats (\\*\\*). Solid lines represent mean, dotted lines represent mean + SE; dashed lines represent mean SE.[Please click here to see a larger figure.](http://www.jove.com/files/ftp_upload/2138/2138fig1.jpg)\n\n**Figure 2.** Kinematic stick figures for the left limb of young, middle-aged, and geriatric rats. Stick figures represent the mean movement of the left limb for each of the groups. There were no significant differences between the groups of animals for joint angle during the step cycle.[Please click here to see a larger figure.](http://www.jove.com/files/ftp_upload/2138/2138fig2.jpg)\n\nDiscussion\n==========\n\nThe present paper provides methodology for evaluating locomotion using continuous quantitative kinematics and ground reaction force determination. Important for anyone interested in embarking upon this methodology is a strong background in biomechanics of locomotion, animal sensorimotor behavior, and data management and manipulation. Though kinematic and ground reaction force determination requires additional time and expertise, compared to some other forms of locomotor analysis (e.g. endpoint measures, ordinal rating scales), the data obtained is sensitive, objective and quantitative for a variety of orthopedic^14-22^ and neurologic^1-3,\\ 23-32^ models of disease, in a variety of species.\n\nWe have provided data that describes differences in locomotion between various ages of strain-matched rats information that could not be gleaned using simple and less sensitive measures. Further, kinematic and kinetic analysis of locomotion has been used to describe locomotor alterations in a variety of nervous system conditions where other forms of evaluation would be unsuccessful.^8,\\ 23-25,\\ 27^ The use of sensitive measures becomes especially important when evaluating potential therapeutants for various models of disease. If a test is not sensitive enough to discern an effect of a potential therapeutant the experimenter runs the risk of committing a type-II statistical error (i.e. concluding there is no effect of a treatment when in fact there was an effect). Further, because endpoint measures and more subjective tests that evaluate locomotion, there exists a potential for bias. Kinematic and kinetic evaluation is purely objective in that, provided appropriate inclusion/exclusion criterion are made *a priori*, the experimenter simply collects, examines and applies appropriate statistics to the data (i.e. there is no subjective component to data determination).\n\nKinematic and kinetic analysis also affords the ability to be used for a multitude of species. In fact, kinematics, ground reaction force determination, or both have been used in a variety of species such as elephants^33-35^, cattle^36^, horses^37-40^, dogs^4,\\ 41-45^, cats^21,\\ 46-49^, various rodents^3,\\ 8,\\ 50,\\ 51^, birds^4,\\ 52-55^, and fish^56,\\ 57^ (this list is by no means exhaustive). In the authors\\' experience, however, the use of mice is problematic given that mice are not easy to operantly condition to travel along a runway. Given this, mice will not travel at a relatively constant speed and instead speed-up and slow-down when traversing the runway. This behavior can likely, in part, be overcome by running mice on a treadmill and video-taping the animal locomoting on the treadmill.^58^ If the experimental apparatus for ground reaction force determination was to be modified for treadmill usage, ground reaction force determination would likely only be easily obtained for vertical ground reaction force as the treadmill belt would interfere with fore-aft and medio-lateral force determination.\n\nAltogether, kinematic and kinetic analysis of locomotion is a reliable, sensitive, and objective method that can be employed for various models of orthopedic and neurologic conditions. Furthermore, all of the equipment has become available for use in rodents, thus negating any related reason for not performing this form of sensorimotor behavioral analysis.\n\nDisclosures\n===========\n\nNo conflicts of interest declared.\n\nThis research was supported by a Discovery Grant to AAW from the Natural Sciences and Engineering Research Council of Canada.\n\n[^1]: Correspondence to: Aubrey A. Webb at <cullenwebb@me.com>\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nChronic low back pain (CLBP) that lasts for six months or longer is estimated to occur in 60--80% of the general population in their lifetime \\[[@B1]\\] and is associated with substantial healthcare costs. The sacroiliac joint (SIJ) complex is one of the major sources of chronic low back pain, accounting for around 10--33% of the total number of CLBP cases \\[[@B1]--[@B5]\\]. The SIJ complex consists of the joint capsule, various muscular and ligamentous structures overlying the joint, and neural structures that innervate the SIJ \\[[@B6]\\]. Current treatment options for SIJ complex-mediated CLBP include intra-articular and periarticular steroid injections, SIJ fusion, and radiofrequency ablation of the neural structures innervating the SIJ. Intra-articular injection of the joint using a mixture of steroids and local anesthetics is a simple procedure and provides quick pain relief, but the effect is short-lived \\[[@B7]\\]. In addition, SIJ fusion is an invasive surgical procedure that should be reserved for cases refractory to nonoperative measures \\[[@B8], [@B9]\\]. On the other hand, radiofrequency ablation (RFA) of the SIJ complex offers longer-lasting effects and has gained wide attention in the last decade \\[[@B10]\\], with increasing numbers of reports advocating for its efficacy \\[[@B7], [@B11]--[@B14]\\]. RFA is usually performed under fluoroscopic guidance. The target structures are the lateral branches of the sacral rami, the dorsal ramus of L5, and the ligamentous structures overlying the joint. However, variations in the pattern of innervation exist between individuals, which provides a challenge for surgeons \\[[@B15]\\]. Due to these variations, different RFA target locations and techniques have been proposed to overcome this difficulty \\[[@B16], [@B17]\\]. Endoscopic radiofrequency ablation has been utilized in the treatment of facetogenic CLBP in a number of clinical reports with favorable results \\[[@B18]--[@B20]\\], but to our knowledge, the efficacy of this technique when applied to SIJ-associated CLBP has not been reported. In this study, we utilized endoscopy for the precise ablation of potential pain generators associated with the SIJ and evaluated the clinical efficacy of this new technique.\n\n2. Materials and Methods {#sec2}\n========================\n\nThe institutional review board at our institution approved this study. The medical records of 17 consecutive patients who underwent RFA of the SIJ for CLBP between April 2011 and December 2015 were reviewed. The inclusion criteria for treatment were as follows: patients with a chief complaint of CLBP with signs and symptoms of SIJ involvement on physical examination and radiological tests such as computed tomography (CT), unresponsive to conservative therapy including oral analgesics and physical therapy, persistent CLBP despite previous lumbosacral operation or pain procedures, and minimum follow-up period of 6 months.\n\nSIJ complex pathology as the main cause of CLBP is difficult to diagnose due to overlapping patterns with other sources of CLBP and varying patterns of pain between individuals \\[[@B7], [@B21]\\]. CT findings of arthropathy or erosion of the SIJ, while not specific, may suggest SIJ based pathology in patients with clinical suspicion \\[[@B22]\\]. While numerous physical examination methods have been suggested, provocative test was reported to have more reliability in numerous reports \\[[@B23], [@B24]\\]. In order to confirm the SIJ pain as the main source of CLBP, two separate diagnostic intra-articular and multisite lateral sacral branch blocks of the SIJ complex were performed at least 2 weeks apart. If patient experienced 50% or higher improvement in pain from the baseline according to visual analogue scale (VAS) after each block, SIJ complex was considered to be the main pain generator, and endoscopic RFA was scheduled. Patients with tumors of the SIJ, concern for secondary gain, previous surgery of the SIJ such as SIJ fusion, or other severe comorbid medical conditions were excluded. All patients were followed for a minimum of six months after the procedure in outpatient clinics. Endoscopic RFA of the SIJ complex was performed in the operating room. Patients were discharged the day after the procedure. All patients were followed up at the outpatient clinic at 1, 3, and 6 months after the procedure and annually thereafter.\n\n2.1. Surgical Techniques {#sec2.1}\n------------------------\n\nAll subjects were placed in the prone position on chest rolls on a radiolucent Jackson table. Before beginning the procedure, patients were fully informed of the procedure details. Patients were monitored and maintained communication with the surgeon throughout the procedure.\n\nAfter sterile prepping and draping, an anteroposterior fluoroscopic view was obtained using a C-arm. A transducer was tilted cephalad approximately 10--15 degrees and was tilted oblique 10--15 degrees contralaterally to optimally visualize the posterior aspect of the SIJ. The skin entry point was positioned at the inferior aspect of the posterior SIJ, and local anesthetic was injected into the entry point. An 18-gauge needle was docked onto the interosseous ligament overlying the posterior SIJ. Then, a guide wire was advanced through the needle, the needle was removed, and a 0.5 cm skin incision was made at the entry site. A cannulated obturator was inserted along the guide wire through the skin incision, and a beveled or nonbeveled working cannula of 7.9\u2009mm diameter was advanced along the obturator until the cannula reached the posterior SIJ. After removing the obturator, the endoscope (6.9\u2009mm \u00d7 5.6\u2009mm) was introduced through the cannula. The final position of the cannula was confirmed with fluoroscopy.\n\nUnder endoscopic visualization, the posterior sacroiliac ligament and its overlying soft tissue were ablated using a Trigger-Flex bipolar probe (Elman International, Inc.) that was introduced through the working channel of the endoscope. First, we ablated perforating branches that innervate the posterior capsule of the SIJ. After visual confirmation of the long posterior sacroiliac ligament, we proceeded with RFA along the course of the ligament in the cranial direction to the level of the posterior superior iliac spine ([Figure 1](#fig1){ref-type=\"fig\"}). The fluoroscope was then adjusted to obtain an anteroposterior view. Next, using the wanding maneuver of the cannula, the cannula tip was moved along the subcutaneous plane toward the region lateral to the S1--S3 sacral foramina, and a linear multidepth lesion was made along the line connecting the lateral margins of the S1--3 sacral foramina ([Figure 2](#fig2){ref-type=\"fig\"}). When uncertain about the position of the RF probe tip, we checked the tip position with the fluoroscope. We attempted to visually confirm the lateral branches exiting the sacral foramina and the branches coursing toward the SIJ when possible to ensure accurate nerve lesioning. Throughout the procedure, we maintained constant communication with the patient to assess the level of pain associated with each stimulus and to identify which stimulus area caused the most pain. Continuous saline irrigation was maintained throughout the procedure to minimize thermal injury to the surrounding structures. After ablation of the target points, the endoscope and cannula were removed. One-point suture with Nylon was used, and sterile dressing was applied.\n\n2.2. Clinical Assessment {#sec2.2}\n------------------------\n\nPatients were instructed to visit the outpatient clinic at 1, 3, 6, and 12 months after the procedure. Pain intensity and functional disability were assessed via questionnaires with outcome measurements before the procedure, immediately after the procedure, and at each follow-up outpatient visit. All clinical assessments were performed by a single coresearcher. At each follow-up visit, back and leg pain intensity was assessed using the visual analogue scale (VAS) and the Oswestry disability index (ODI). Additionally, all patients were asked to express their degree of satisfaction with the procedure on a percentage scale.\n\n2.3. Statistical Analysis {#sec2.3}\n-------------------------\n\nMean VAS scores for back and ODI scores immediately after the procedure, three and six months after, and one year after the procedure were compared to the scores recorded before the procedure. Statistical significance was assessed using paired Student\\'s*t*-tests. *p* values less than 0.05 were considered to be statistically significant. All statistical analyses were performed using SPSS version 18 (SPSS Inc.).\n\n3. Results {#sec3}\n==========\n\nIn this study, there were 2 male and 15 female patients whose age ranged from 37 to 81 years (mean age 61.9 \u00b1 11.8 years). On preoperative lumbosacral MRI, herniated nucleus pulposus was identified in eight patients, spondylolisthesis in four patients, and spinal stenosis in five patients. On CT scan of the pelvis, arthropathy of the SIJ was observed in 11 patients (64.7%).\n\nBefore the RFA procedure, ten patients (58.8%) had undergone operation or pain procedure on the lumbosacral spine. Four patients (23.5%) underwent lumbar or lumbosacral interbody fusion and posterior fixation, four (23.5%) underwent discectomy or laminectomy, and two (11.8%) underwent medial branch RFA. Seven patients (41.2%) underwent nerve block procedures including caudal block, medial branch block, root block, and transforaminal epidural block. All of the ten patients experienced persistent CLBP despite these procedures. Patient demographic data are summarized in [Table 1](#tab1){ref-type=\"table\"}.\n\nRFA was performed on the right side only in eight patients (47.1%), on the left side only in five patients (29.4%), and on both sides in four patients (23.5%). The mean duration of operation from the time of local anesthetic injection to wound closure was 26.6 \u00b1 22.5 (20--50)\u2009mins per side. All patients were discharged the next day, without perioperative complications such as hematoma collection, wound discharge, or development of acute neurological deficit. The mean VAS scores for back pain decreased from 6.7 \u00b1 1.41 preoperatively to 3.6 \u00b1 1.28, 3.2 \u00b1 1.06, 2.8 \u00b1 1.14, and 3.1 \u00b1 1.78 immediately postoperatively, and at 1, 3, and 6 months\\' follow-up visits, respectively. All of the follow-up VAS scores were significantly lower than the baseline (*p* \\< 0.005). The mean ODI score preoperatively was 22.2 \u00b1 3.36 and decreased to 14.1 \u00b1 3.35, 13.1 \u00b1 4.05, 12.9 \u00b1 4.32, and 12.0 \u00b1 4.69 immediately postop and at the 3, 6, and 12-month follow-up visits, respectively. All of the follow-up ODI scores were significantly lower than baseline (*p* \\< 0.001). Mean patient satisfaction rate was 86.6% (70--100). These results are summarized in [Table 2](#tab2){ref-type=\"table\"}.\n\n3.1. Example Case {#sec3.1}\n-----------------\n\nA 69-year-old female presented with chronic low back pain and left buttock pain for the last five years. She had undergone L4-5 and L5-S1 fusion at another institution in 2008. She had remained asymptomatic for two years, but in 2010, she started feeling left-sided buttock pain. Conservative treatment with oral analgesics and four separate root blocks and epidural blocks at a local pain clinic were ineffective. CT scan of the pelvis revealed bilateral SIJ arthropathy. Her initial VAS for back pain and her ODI scores were 7 and 40, respectively, and downward pressure on the sacrum with the patient in the prone position elicited severe pain (VAS 8). After two diagnostic SIJ complex injections, her VAS score dropped to 2 points on both occasions. Endoscope-guided RFA was performed one month later.\n\nLocal analgesics were applied to the entry point of the skin, which was about 1\u2009cm above where the buttock pain was elicited. Buttock pain was elicited when the long posterior sacroiliac ligament overlying the posterior articular capsule was stimulated with the RF probe in short bursts (Figures [1(a)](#fig1){ref-type=\"fig\"}, [1(b)](#fig1){ref-type=\"fig\"}, and [1(c)](#fig1){ref-type=\"fig\"}). While maintaining continuous saline irrigation, the painful areas along the length of ligament were coagulated with the RF probe to the level of the posterior superior iliac spine. Next, using the wanding maneuver, the tip of the cannula was gently mobilized in the subcutaneous plane and repositioned next to the lateral margin of the S1 neural foramen. The S1 lateral sacral branches were identified and coagulated in the same manner (Figures [2(a)](#fig2){ref-type=\"fig\"}, [2(b)](#fig2){ref-type=\"fig\"}, and [2(c)](#fig2){ref-type=\"fig\"}). Ablation of a targeted area was stopped when no further pain was elicited upon additional stimulation. The S2 and S3 lateral branches were ablated in a similar manner ([Figure 3](#fig3){ref-type=\"fig\"}).\n\n4. Discussion {#sec4}\n=============\n\nThe SIJ complex is increasingly being recognized as a major source of CLBP, and, yet, the exact pain generating mechanisms and anatomical properties of the SIJ complex have not been fully established. The literature surrounding the treatment of SIJ complex-mediated pain is still quite sparse. Diagnosing SIJ complex pain remains largely a diagnosis of exclusion. In SIJ complex pain, patients tend to complain of buttock pain, but many experience lower leg pain on the involved side as well, which can be confused with radiculopathy or referred pain from other low back structures \\[[@B25]\\]. There is no single reliable test that can identify the SIJ complex as the main pain generator in CLBP, and, in the majority of cases, there are a number of possible pain generators that are equally likely to contribute to the CLBP. For these reasons, accurate diagnosis and successful treatment of SIJ-related pain provide a challenge for surgeons, often causing delay in recognition of SIJ complex as the main source of CLBP. Recognizing the SIJ as the main pain generator takes time and effort, since SIJ complex pain is often only diagnosed after all other modalities fail to treat the patient\\'s CLBP. In our case series, all 17 patients had competing pathologies of the lumbar spine, including herniated lumbar disks, spinal stenosis, spondylolisthesis, and facet joint arthropathy that may have contributed to their symptoms. All had received one or more pain procedures at the lumbar level, including root blocks, epidural blocks, epidural neurolysis, facet joint blocks, and medial branch blocks/neurotomy with minimal or unsatisfactory results. Of the 17 patients, 7 (41.1%) had previously undergone some form of surgical procedure at the lumbar level, including five cases of interbody fusion and two cases of discectomy, which did not satisfactorily alleviate their CLBP. The average duration of symptoms prior to the diagnosis of SIJ complex pain was 2.8 \u00b1 6.5 years (6 months--10 years).\n\nThe SIJ complex consists of an articular region, a posterior ligamentous region, and a dorsal ligamentous region, which support the joint \\[[@B26]\\]. Neural innervations are found in both the posterior capsule of the joint and in the posterior sacroiliac and interosseous ligaments \\[[@B6]\\]. Although many studies have used intra-articular injections both to provide relief and to identify the origin of the SIJ pain, results from recent studies indicate that periarticular ligamentous structures may contribute more to SIJ pain than the articular region itself \\[[@B27], [@B28]\\], suggesting that the periarticular ligamentous structures are better targets for treatment. Dreyfuss et al. demonstrated that multisite, multidepth lateral sacral branch blocks are more reliable than intra-articular injections for identifying the SIJ complex as the main pain generator, and a recent systemic review by King et al. also supports this view. When performing diagnostic blocks, we targeted the intra-articular joint itself as well as the lateral sacral branches of S1--3 and the L5 dorsal ramus, since these will be the targets for RFA. Additionally, compared to intra-articular injection alone, relief after injection of these structures would be a better predictor of favorable outcomes after RFA.\n\nThe lateral branches of L5-S4, especially the branches of S1--S3 innervating the sacroiliac joint, are quite variable between individual patients in their course, branches, and location. To ablate all pain generators, a large lesion area would be required \\[[@B16], [@B17]\\]. Current techniques to accomplish these lesions include a periforaminal approach and a lateral sacral crest approach. The periforaminal approach involves making a series of lesions around the lateral border of the S1--S3 sacral foramina using a monopolar or bipolar RF probe. However, as described by Roberts et al. \\[[@B17]\\], it is often difficult to achieve a clear view of the foramina under fluoroscopy. The lateral sacral technique involves creating a lesion strip along the lateral sacral crest at regular intervals, effectively ablating the fine plexus of lateral branches overlying this region. This technique also has disadvantages in that the lateral crest is not easy to identify under fluoroscopy alone.\n\nFor both the periforaminal approach and the lateral sacral crest approach, multiple skin punctures are made in order to place the RF probe, which can result in patient discomfort. Because the cannula for the endoscope is considerably thicker and more rigid than the RF probe itself, it was possible to reach the periforaminal regions of the S1--3 sacral foramina through a single incision using the endoscope wanding maneuver. Notably, the wanding maneuver did not cause the patients more discomfort. If the procedure was uncomfortable, we offered the patient light sedation with midazolam. With the aid of the endoscope, we achieved better visualization of bony landmarks as well as the lateral branches of S1--S3 when possible. The branches of the posterior rami of S1--S3 travel deep to the long posterior sacroiliac and sacrotuberous ligaments. According to a cadaveric study by Roberts et al. \\[[@B17]\\], the diameter of the S1--S3 branches ranges from 0.21 to 1.51\u2009mm, and in many cases, the lateral branches could not be identified with certainty. If gentle stimulation of the suspected lateral branch with the RF probe elicited pain, we ablated the branch. Except for one patient, all procedures were performed with the patient under local anesthesia or light sedation, and communication with the patient was maintained throughout the procedure.\n\nAnother advantage of direct visualization with the endoscope is that it affords us the ability to identify areas that have already been ablated. This allows us to avoid damaging the soft tissue with excessive lesioning of the same region. Avoiding repeated lesioning helped curb complications such as postprocedural pain and dysesthesia ([Figure 4](#fig4){ref-type=\"fig\"}).\n\nOur study has several limitations. First, this is a retrospective study with a limited number of cases. Second, no direct comparison was made with conventional fluoroscopic-guided RFA method in the clinical results and perioperative parameters. Third, patients were not categorized by their VAS and ODI scores before the procedure or previous operation or procedure, and clinical effects of endoscopic RFA procedure could be confounded by other variables. We plan to conduct a randomized clinical trial comparing conventional RFA and endoscopic RFA in the future.\n\n5. Conclusions {#sec5}\n==============\n\nOur preliminary results suggest that endoscope-guided RFA may be alternative option to treat CLBP secondary to SIJ complex pain with favorable clinical outcomes, including a long-term pain-free period and improved physical function with minimal complications.\n\nCompeting Interests\n===================\n\nThe authors declare that there is no conflict of interests regarding the publication of this paper.\n\n![((a) and (b)) Long posterior ligament (black arrowheads) overlying the posterior capsule of the SIJ. (c) Corresponding position of the cannula tip in the anteroposterior fluoroscopic image.](BMRI2016-2834259.001){#fig1}\n\n![(a) Lateral sacral branches of S1 as they exit from the S1 foramen (black arrowheads). (b) Corresponding position of the endoscope cannula. (c) Small arteries or veins can often be seen coursing along the nerve branches (white arrowheads), which can help with identification of thin nerve branches.](BMRI2016-2834259.002){#fig2}\n\n![Fluoroscopic view of the endoscopic cannula tip in various positions during the procedure. The cannula tip can be moved in the subcutaneous plane and can be repositioned without causing much discomfort. If patients did experience discomfort, an additional lidocaine injection was applied.](BMRI2016-2834259.003){#fig3}\n\n![Under endoscopic view, it is possible to clearly discern areas that have already been ablated (surrounded by arrows) and which areas have not. It is also possible to gauge the depth of the ablation.](BMRI2016-2834259.004){#fig4}\n\n###### \n\nPatient demographic data.\n\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Patient number   Age   Gender   Follow-up period (months)   Side of procedure   Other diagnoses                  Previous operations/procedures\n  ---------------- ----- -------- --------------------------- ------------------- -------------------------------- ------------------------------------------------------\n  1                56    F        49                          Right               HNP, L4-5, Lt.                   Root block, L4, 5, Lt.\n\n  2                70    F        37                          Left                Extraforaminal HNP, L5-S1, Lt.   Transforaminal epidural block, L5-S1, Lt.\n\n  3                74    F        36                          Both                HNP, L3-4, 4-5, 5-S1             Caudal block\\\n                                                                                                                   medial branch block, L3-4, 4-5, 5-S1, both\n\n  4                73    F        35                          Both                Spinal stenosis, L3-4, 4-5       Medial branch block, L3-4, 4-5, both\n\n  5                76    F        35                          Left                HNP, L4-5, Rt.                   \\(1\\) Decompressive hemilaminectomy, L4, Rt. (2005)\\\n                                                                                                                   (2) Medial branch RFA, L4, 5, Rt. (2006)\n\n  6                81    F        22                          Right               Spinal stenosis, L3-4, 4-5       PLIF, L3-4, 4-5 (2008)\n\n  7                37    F        22                          Right               Spinal stenosis, L4-5, L5-S1     Root block, L4, 5, Rt.\n\n  8                72    F        21                          Left                Spinal stenosis, L5-S1           PLIF, L5-S1 (2001)\n\n  9                61    F        15                          Left                Spondylolisthesis, L4-5          2014.10.8 MIS TLIF 45\\\n                                                                                                                   (post-procedure)\n\n  10               57    F        13                          Right               Spinal stenosis, L4-5            PLIF, L4-5 (2012)\n\n  11               48    F        13                          Right               HNP, L3-4, Rt.                   Transforaminal epidural block, L3-4, Rt.\n\n  12               58    F        12                          Right               Spondylolisthesis, L3-4          Caudal block, epidural block, L3-4\\\n                                                                                                                   Trigger point injection of paravertebral muscles\n\n  13               56    F        12                          Left                HNP, L5-S1, Rt.                  Discectomy, L5-S1, Rt. (2013)\n\n  14               58    F        10                          Both                Facet arthropathy, L4-5, Lt.     Medial branch RFA,\\\n                                                                                                                   L4, 5, Lt. (2013)\n\n  15               60    F        9                           Both                Spondylolisthesis, L4-5          PLIF, L4-5 (2013)\n\n  16               47    M        9                           Right               HNP, L5-S1, Lt.                  Discectomy, L5-S1, Lt. (2013)\n\n  17               69    F        8                           Right               Spondylolisthesis, L4-5, L5-S1   PLIF, L4-5, 5-S1 (2008)\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n###### \n\nPreoperative and postoperative clinical data.\n\n  \u2009     Mean preoperative scores   Mean immediate postoperative scores   Mean 1-month follow-up scores   Mean 3-month follow-up scores   Mean 6-month follow-up scores\n  ----- -------------------------- ------------------------------------- ------------------------------- ------------------------------- -------------------------------\n  VAS   6.7 \u00b1 1.41                 3.6 \u00b1 1.28                            3.2 \u00b1 1.06                      2.8 \u00b1 1.14                      3.1 \u00b1 1.78\n  ODI   22.2 \u00b1 3.36                14.1 \u00b1 3.35                           13.1 \u00b1 4.05                     12.9 \u00b1 4.32                     12.0 \u00b1 4.69\n\n[^1]: Academic Editor: Tsung-Jen Huang\n"}
+{"text": "Introduction {#s1}\n============\n\nFlowering is an important process essential for sexual reproduction, seed development and fruit production. Although flowering is composed of a series of typically irreversible sequential events, reversion from floral to vegetative growth is frequently observed in nature. Reversions can be divided into two categories: inflorescence reversion, in which vegetative growth is resumed after or intercalated within inflorescence development, and flower reversion, in which vegetative growth is resumed in an individual flower [@pgen.1000818-Tooke1],[@pgen.1000818-Battey1]. Reversion, which can serve a function in the life history strategy (perenniality) or reproductive habit (pseudovivipary), is essential for the life cycle of some plant species [@pgen.1000818-Tooke1],[@pgen.1000818-Battey1].\n\nVivipary in flowering plants is defined as the precocious and continuous growth of the offspring while still attached to the parent plant [@pgen.1000818-Elmqvist1],[@pgen.1000818-Goebel1]. Vivipary can be divided into two distinct types: true vivipary and pseudovivipary [@pgen.1000818-Elmqvist1]. True vivipary is a sexual reproduction process in which seeds germinate before they detach from maternal plant. On the other hand, pseudovivipary is a specific asexual reproductive strategy in which bulbils or plantlets replace sexual reproductive structures [@pgen.1000818-Elmqvist1],[@pgen.1000818-Coelho1]. Pseudovivipary has been widely recorded in monocots, in particular grasses that grow in extreme environments [@pgen.1000818-Tooke1], [@pgen.1000818-Elmqvist1], [@pgen.1000818-Coelho1]--[@pgen.1000818-Moore1]. Characteristics of the environments which favour pseudovivipary include climate changes, high precipitation and humidity, drought, fungal infection, high altitudes and latitudes, late-thawing habitats, or arid/semi-arid areas [@pgen.1000818-Tooke1],[@pgen.1000818-Elmqvist1],[@pgen.1000818-Coelho1]. Several authors have argued that pseudovivipary has evolved in response to a short growing season, enabling plants to rapidly complete the cycle of offspring production, germination and establishment during the brief periods favourable to growth and reproduction [@pgen.1000818-Elmqvist1]. In developmental terms pseudovivipary occurs in two principal ways. The first way to proliferate, as in *Festuca ovina*, *Poa alpina* and *Poa bulbosa*, is through the transformation of the spikelet axis into the leafy shoot. The second way is to form the first leaf of the plantlet by lemma elongation, as is the case in *Deschampsia caespitose* and *Poa robusta* [@pgen.1000818-Tooke1],[@pgen.1000818-Moore1]. In some cases, such as *Deschampsia alpine* and *Phleum pratense*, both modes of propagule development have been found in a single plant [@pgen.1000818-Moore1], indicating that the molecular difference between the two types of pseudovivipary might be rather small.\n\nPseudovivipary has fascinated biologists, as elucidation of its mechanism could lead to an understanding of flower evolution and sexual reproduction; hence, it might provide an opportunity to manipulate a plant\\'s reproductive strategy. As pseudovivipary is always closely associated with various environmental factors, the molecular basis of pseudovivipary is still unknown. Here we report mutations of two MADS-box transcription factors that are essential for sexual reproduction and mutations of which lead to stable pseudovivipary in rice.\n\nResults {#s2}\n=======\n\nCharacterization of *pho* mutant {#s2a}\n--------------------------------\n\nIn this study, a naturally occurring mutant showing inflorescence reversion was found in the offspring of an *Oryza sativa* spp. *indica* var. Zhongxian 3037. Instead of normal floral organs, this mutant generated new plantlets ([Figure 1A and 1B](#pgen-1000818-g001){ref-type=\"fig\"}). The plantlets, like normal juvenile plants, generated roots, produced tillers and showed normal vegetative growth when explanted in paddy fields ([Figure S1A and S1B](#pgen.1000818.s001){ref-type=\"supplementary-material\"}). In the subsequent life cycle, plants again displayed inflorescence reversion. Thus, this mutant could be considered to be a complete pseudovivipary mutant in which the reproductive mode has completely changed from sexual to asexual. In fact, this mutant has accomplished six life cycles via this asexual reproductive method. This type of mutation has not been reported before in rice. We named the mutant *phoenix* (*pho*) to reflect its stable \"never die and reborn anew\" phenotype. Two additional mutants were also found in this segregating population. The first mutant was named *degenerative palea* (*dep*), and was characterized by shrunken paleas. Paleas in severe *dep* florets degenerated to glume-like organs that were prone to splitting. The lemmas and glumes in *dep* florets were slightly elongated ([Figure 1D](#pgen-1000818-g001){ref-type=\"fig\"}). The second mutant *abnormal floral organs* (*afo*) displayed a phenotype opposite to *dep*, with abnormalities primarily in lemma and the inner three whorls ([Figure 1E](#pgen-1000818-g001){ref-type=\"fig\"}).\n\n![Phenotypic characterization and genetic analysis of *pho*, *dep*, and *afo* mutants.\\\n(A) The phenotype of wild-type (left) and *pho* (right) plants. (B) All flowers are replaced by young plantlets in *pho* panicle. (C) The spikelet of wild-type rice. (D) The spikelets of *dep* in the order of increasing severity showing the defects of paleas. (E) The spikelets of the *afo* mutant showing pleiotropic defects in lemmas and the inner three whorls. (F) Genetic analysis of *pho*, *dep*, and *pho* mutants indicates that *pho* might be a double mutant containing both a Mendelian mutation in *DEP* and a non-Mendelian mutation in *AFO*; \"n\" indicates the line number.](pgen.1000818.g001){#pgen-1000818-g001}\n\nIn order to examine the genetic basis of the three mutations, seeds of the 28 individual plants showing the normal phenotype from the above population were planted into lines by parent plants. We found that those genotypes self-segregated into two categories. The first category only produced *afo* and wild phenotype plants, while the second category produced *dep*, *afo*, and *pho*, as well as wild phenotype plants. As the segregation ratios in both categories seemed unclear, seeds of the wild phenotype plants from each category were planted in individual lines for two more generations. Subsequently, all plants in the final generation were counted and analyzed (summarized in [Figure 1F](#pgen-1000818-g001){ref-type=\"fig\"}). In the first category lines, 35.34% of plants displayed the *afo* phenotype, while 64.66% of plants exhibited the wild phenotype (n\u200a=\u200a232). As the segregation did not follow Mendelian patterns (3\u22361 ratio, *\u03c7* ^2^ (1)\u200a=\u200a13.24, *P*\\<0. 01), we proposed that *afo* might be a non-Mendelian mutant. In the second category lines, 28.44% plants showed the *afo* phenotype, 18.35% plants showed the *dep* phenotype and 7.34% plants showed the *pho* phenotype (n\u200a=\u200a218). We observed that *pho* only appeared in the line where *afo* and *dep* mutants coexisted. In addition, when we put the wild phenotype plants and *afo* mutants into one group and *dep* and *pho* into another group, the segregation ratio would fit a 3\u22361 ratio (162\u223656, *\u03c7* ^2^ (1)\u200a=\u200a0.06, *P*\\>0.50), indicating that *dep* might be a Mendelian mutant. Therefore, we further hypothesized that *pho* might be a double mutant containing both a Mendelian mutation in *DEP* and a non-Mendelian mutation in *AFO*.\n\nSingle amino acid mutation disrupts the transcriptional activation of OsMADS15 in *dep* {#s2b}\n---------------------------------------------------------------------------------------\n\nTo understand the molecular mechanism of pseudovivipary in *pho*, we began by isolating the *DEP* gene through map-based cloning. The *dep* mutants from the second category line were crossed to *O. sativa* spp *japonica* var. Zhonghua11 to generate a mapping population. In the F2 population, 71 of 302 plants showed the *dep* phenotype (3\u22361 ratio, *\u03c7* ^2^ (1)\u200a=\u200a0.36, *P*\\>0.50), confirming that the phenotype of the *dep* mutant is controlled by a single recessive gene. 2,292 F2 and F3 plants showing the *dep* phenotype were used to map *DEP* to a 50-kbp region on the short arm of chromosome 7. All genes within this region were amplified and sequenced. A single nucleotide G to C substitution at position 94 in coding region was found in the first exon of the *OsMADS15* in the *dep* mutant. This substitution results in a change from a MADS-box conserved alanine residue to proline ([Figure 2A](#pgen-1000818-g002){ref-type=\"fig\"} and [Figure S5](#pgen.1000818.s005){ref-type=\"supplementary-material\"}). The same nucleotide mutation was also found in all the *pho* mutants analyzed (n\u200a=\u200a20), further implying that the mutation of *OsMADS15* might be partly responsible for the *pho* phenotype. To confirm that the loss of function of OsMADS15 is responsible for *dep*, we utilized an RNA interference approach to down-regulate *OsMADS15*. Forty transgenic plants expressing an inverted repeat of 317 bases of *OsMADS15* were generated in Nipponbare. Among them, 35 plants also displayed the *dep* degenerative palea phenotype ([Figure S1C and S1D](#pgen.1000818.s001){ref-type=\"supplementary-material\"}). Therefore, we concluded that the phenotype of the *dep* mutant is indeed caused by mutation in *OsMADS15*.\n\n![Molecular mechanisms of *dep* mutant and *afo* mutant.\\\n(A) Amino acid mutation corresponding to the nucleotide change in *dep*. (B) OsMADS15-GFP fusion protein is localized in nucleus while Osmads15 (*dep*)-GFP fusion protein is localized in cytosol. (C) Transcriptional activation assay of pOsMADS15, pOsMADS15-*dep*, pOsMADS15\u25b3C180-267, pOsMADS15\u25b3N1-66, and pLexA. White clones indicate no activation of the reporter gene while blue clones indicate activation of the reporter gene. (D) *OsMADS1* expression analysis by quantitative real-time PCR analysis in WT, *dep*, *afo*, and *pho* panicles shows the silencing of *OsMADS1* in *afo* and *pho*. (E) 294-bp sequence in the promoter region of *OsMADS1* gene shows different cytosine methylation in WT and *afo*. The yellow-marked cytosines were found to be methylated in WT or *afo*. (F) Profiles of DNA methylation in 294-bp region in WT (red line) and *afo* (blue line) plants. The numbers on the X axis represent cytosine positions in the analyzed region, and the Y axis represents methylation ratios in WT and *afo*.](pgen.1000818.g002){#pgen-1000818-g002}\n\nWe found five *OsMADS15* transcripts with differing sequences in GeneBank. To identify the WT *DEP* sequence, we performed RT-PCR and found that our cDNA sequence was identical to GB accession AB003325. This cDNA was used for subsequent analysis. MADS-box proteins are transcription factors, so we conducted experiments to evaluate whether amino acid substitution impaired the transcriptional activation function of OsMADS15 in the *dep* mutant. OsMADS15 from both WT and *dep* were fused with GFP protein and transiently expressed in onion epidermal cells as well as rice protoplast cells. The OsMADS15 GFP signal was localized in the nucleus, whereas the *dep* mutant caused redistribution of OsMADS15 GFP to the cytosol ([Figure 2B](#pgen-1000818-g002){ref-type=\"fig\"} and [Figure S2](#pgen.1000818.s002){ref-type=\"supplementary-material\"}). Previous study has revealed that the KC region of OsMADS15 (Amino acids of AF058698) does not show any transcriptional activation function [@pgen.1000818-Lim1]. However, a single amino acid substitution, from leucine to histidine mutation, has occurred at position 117 of the amino acids of AF058698. In our study, we found that the OsMADS15 protein itself exhibited transcriptional activator activity. Furthermore, when the MADS domain of OsMADS15 was eliminated, the residual IKC region of OsMADS15 also displayed transcriptional activator activity. However, the mutated protein in *dep* lost its transcriptional activator activity completely, though the amino acid mutation only occurred in the MADS domain ([Figure 2C](#pgen-1000818-g002){ref-type=\"fig\"}). Taken together, it is very likely that the mutated OsMADS15 protein has lost its transcriptional activation function in *dep*.\n\n*afo* is an epigenetic mutant of *OsMADS1*, while *pho* is a spontaneous mutant containing both genetic mutation in *OsMADS15* and epigenetic mutation in *OsMADS1* {#s2c}\n-------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nFrom the above genetic analysis, it was deduced that *pho* and *afo* were non-Mendelian mutants, so we proposed that they might be epigenetic mutants. Epigenetic mutations are often marked by a reduction or elimination of an associated transcript. Microarray experiments were carried out to investigate whether there were any variations in transcript accumulation between *pho* and WT young panicles ([Table 1](#pgen-1000818-t001){ref-type=\"table\"}). These experiments showed that the transcript levels of multiple genes were altered. Of those altered genes, *OsMADS1* (also known as *LEAFY HULL STERILE1*, *LHS1* [@pgen.1000818-Jeon1]), was the most significantly altered transcript, with a 2,208-fold reduced expression in *pho* relative to WT. Real-time PCR was further performed using WT, *dep*, *afo* and *pho* panicle transcripts to confirm this result and to examine whether the *afo* mutant also showed a reduced expression of *OsMADS1* transcripts. As expected, the expression of *OsMADS1* was hardly detectable in *afo* as well as *pho* ([Figure 2D](#pgen-1000818-g002){ref-type=\"fig\"}). Additionally, no mutations were detected in the 12,879-bp genomic sequence of the *OsMADS1* locus, including the eight exons, seven introns, 2,507-bp upstream sequence and 1,870-bp downstream sequence. We hypothesized that the *afo* mutant might be caused by an epigenetic modification of *OsMADS1*. Interestingly, recent studies in hexaploid wheat (*Triticum aestivum*) revealed that *WLHS1-B*, one of the homologs of *OsMADS1*, was silenced by cytosine methylation [@pgen.1000818-Shitsukawa1]. To test if this was also the case in rice, we used bisulfate sequencing of exon 1 and the 5\u2032 upstream regions of *OsMADS1* in *afo* to characterize their methylation status. Compared with the WT plants, the promoter region of *OsMADS1* in *afo* was more heavily methylated (from 31.43% to 62.86%), which might contribute to the silencing of *OsMADS1* ([Figure 2E and 2F](#pgen-1000818-g002){ref-type=\"fig\"}).\n\n10.1371/journal.pgen.1000818.t001\n\n###### Expression analysis of MADS-box genes in *pho* mutant according to the microarray data.\n\n![](pgen.1000818.t001){#pgen-1000818-t001-1}\n\n  Gene            WT Signal   *pho* Signal   WT-vs-*pho* Signal Ratio\n  -------------- ----------- -------------- --------------------------\n  *OsMADS1*        10165.8        4.6                2209.96\n  *OsMADS8/24*     9202.9        156.7                58.73\n  *OsMADS7/45*     8547.8        194.6                43.92\n  *OsMADS29*        119.6         4.1                 29.17\n  *OsMADS13*        247.9          14                 17.71\n  *OsMADS17*       3034.6        318.6                 9.52\n  *OsMADS4*        1842.2        224.9                 8.19\n  *OsMADS3*         414.3        103.2                 4.01\n  *OsMADS5*        7152.3        1577.8                4.53\n  *OsMADS6*        5314.5        1423.5                3.73\n  *OsMADS2*         6275         2591.8                2.42\n  *OsMADS14*         713         2063.4                0.35\n  *OsMADS34*        569.7        2269.7                0.25\n\nTo ascertain whether *pho* was a *dep/afo* double mutant, We crossed *dep* with *naked seed rice* (*nsr*), a mutant of the *OsMADS1* gene [@pgen.1000818-Chen1], to generate *dep/nsr* double mutants. In the F~2~ and F~3~ population, all the *dep/nsr* double mutants analyzed (n\u200a=\u200a35) showed a similar pseudovivipary phenotype to that of the *pho* mutants ([Figure S3](#pgen.1000818.s003){ref-type=\"supplementary-material\"}). This double mutant has accomplished three life cycles via asexual reproductive method. So, this result confirmed that *pho* was a double mutant of *Osmads1* and *Osmads15*.\n\n*dep* displays pseudovivipary occasionally {#s2d}\n------------------------------------------\n\nThe spikelet development of each of the three mutants was further analyzed to explore functions of the two MADS-box genes during spikelet development. Previous studies have characterized *OsMADS1* as a *SEPALLATA* (*SEP*)-like gene and performed multiple investigations in rice. However, the function of *OsMADS1* is still not fully elucidated [@pgen.1000818-Jeon1], [@pgen.1000818-Chen1]--[@pgen.1000818-Agrawal1]. The *afo* mutant shared many similarities with those severely affected *Osmads1* (*lhs1*) mutants and *OsMADS1*RNAi plants ([Figure 1E](#pgen-1000818-g001){ref-type=\"fig\"}): all spikelets were sterile; lemmas were more severely affected than paleas; palea marginal tissues (PMTs) were absent while palea main structures (PMSs) were only slightly effected; lodicules were converted into glume-like organs; and ectopic florets that are indicative of partial reversion had frequently arisen from the parent florets [@pgen.1000818-Jeon1],[@pgen.1000818-Chen1],[@pgen.1000818-Prasad2],[@pgen.1000818-Agrawal1]. In summary, the phenotype of *afo* mutant suggests that *OsMADS1* is required for the specification of lemma, PMTs and the three inner whorls [@pgen.1000818-Jeon1],[@pgen.1000818-Chen1],[@pgen.1000818-Prasad2],[@pgen.1000818-Agrawal1]. Its pleiotropic defects indicate that *OsMADS1* is essential for flower meristem (FM) determinacy [@pgen.1000818-Jeon1], [@pgen.1000818-Chen1], [@pgen.1000818-Prasad2]--[@pgen.1000818-Kater1].\n\nPhylogenetic analyses have characterized *OsMADS15* as an *APETALA1* (*AP1*)/*FRUITFUL* (*FUL*)-like gene ([Figure S4](#pgen.1000818.s004){ref-type=\"supplementary-material\"} and [Figure S5](#pgen.1000818.s005){ref-type=\"supplementary-material\"}) [@pgen.1000818-Yamaguchi1]--[@pgen.1000818-Kyozuka1]. In addition, previous study has shown that *OsMADS15* (*RAP1A*) RNA was expressed in the incipient floral primordium and later mainly accumulated in empty glumes, lemma, palea and lodicules [@pgen.1000818-Kyozuka1]. However, the function of *OsMADS15* is still unclear [@pgen.1000818-Yamaguchi1],[@pgen.1000818-Kater1]. The effects of *OsMADS15* on cell specifications of all spikelet whorls were histologically examined. In a severely affected *dep* spikelet, the transformed palea was actually only composed of two PMTs while the PMS was completely lost ([Figure 3A and 3B](#pgen-1000818-g003){ref-type=\"fig\"}). This implied that the identity of palea was lost in the *dep* spikelet with the severe phenotype. The lemma in the *dep* spikelet was also slightly affected, but its identity was still maintained ([Figure 3A and 3B](#pgen-1000818-g003){ref-type=\"fig\"}, and [Figure S6](#pgen.1000818.s006){ref-type=\"supplementary-material\"}). The glumes of *dep* spikelets contained many more bundles than the WT glumes, suggesting a possible partial reversion of glumes to leaf-like organs. No obvious difference was found in the inner three whorls, hinting that they are not affected by the mutation of *OsMADS15*. Thus, *OsMADS15* is required for the specification of PMS and empty glumes, those floral organs are just opposite to the affecting whorls of *OsMADS1*.\n\n![Spikelet morphologies of WT, *dep*, *afo*, and *pho* plants.\\\n(A) Transverse section of the WT spikelet shows normal glumes (gl), lemma (le), palea main structure (PMS) and palea marginal tissue (PMT). (B) Transverse section of the severely affected *dep* spikelet shows the loss of PMS. (C) Occasional emergence of root at the base of *dep* rachilla on the lemma side (left) and occasional emergence of tiller between palea and upper empty glume in *dep* spikelets (right, see also [Figure S7](#pgen.1000818.s007){ref-type=\"supplementary-material\"}). (D,E) SEM of the floral primordium in WT shows that only two empty glumes, lemma (le) and palea (pa) are arranged in alternate phyllotaxis. (F--H) SEM of the floral primordium in *pho* shows that all lateral organs are arranged in alternate phyllotaxy. Bars in (A,B), 200 \u00b5m; bars in (D--H), 10 \u00b5m.](pgen.1000818.g003){#pgen-1000818-g003}\n\n*dep* showed a stable degenerative palea phenotype when grown in paddy fields with a normal climate. Unexpectedly, however, we found that, under a continuous rain for several days during its heading stage, roots occasionally emerged from the base of *dep* rachillas ([Figure 3C](#pgen-1000818-g003){ref-type=\"fig\"}). Only one root was formed in each spikelet and it merely located at the lemma side (n\u200a=\u200a22). These roots would soon degenerate if the spikelets were dried. Interestingly, if the continuous rain occurred after the heading stage, the inner floral organs or developing seeds of *dep* always got mildewed because of the lack of protection by paleas, but emergence of new shoots was occasionally visible in *dep* spikelets ([Figure 3C](#pgen-1000818-g003){ref-type=\"fig\"} and [Figure S7A, S7B](#pgen.1000818.s007){ref-type=\"supplementary-material\"}, and [S7C](#pgen.1000818.s007){ref-type=\"supplementary-material\"}). In contrast to the emerged roots that were only formed on the lemma side, these emerged shoots only appeared between paleas and upper empty glumes on the other side (n\u200a=\u200a24). Moreover, prophylls were found on these shoots, indicating that these emerged shoots are actually tillers. These tillers also generated roots, produced new tillers and showed normal vegetative growth when replanted in fields ([Figure S7D and S7E](#pgen.1000818.s007){ref-type=\"supplementary-material\"}). So, *dep* can also be considered to be an unstable pseudovivipary mutant that was closely associated with environmental factors. In the *dep* mutant, most floral organs develop normally, demonstrating that *OsMADS15* might only play a minor role in the FM determinacy. However, the occasional emergence of roots and tillers in *dep* implies that the shoot apical meristem (SAM) identity is restored and begins to grow under a suitable environment (continuous rain), so *OsMADS15* might also participate in inhibiting SAM formation in incipient floral primordium. However, pseudovivipary has not been observed in *DEP* RNAi plants that grow in paddy fields; it is probably that the residual transcripts in RNAi plants are sufficient to inhibit SAM formation in incipient floral primordium. Alternatively, pseudovivipary, which is mainly observed in natural plants, might be a *dep* allele--specific phenomenon.\n\nFinally, the primordium development of *pho* mutant was also analyzed. In WT, two empty glumes, lemma and palea were arranged in alternate phyllotaxis while stamens and carpel were not ([Figure 3D and 3E](#pgen-1000818-g003){ref-type=\"fig\"}). In contrast, in the *pho* mutant, no stamen or carpel was observed and all lateral organs were arranged in alternate phyllotaxis ([Figure 3F--3H](#pgen-1000818-g003){ref-type=\"fig\"}). As those lateral organs finally grew into true leaves but not simple leaf-like organs, it is obvious that FM at least partially transformed into functional SAM although some following floral genes still expressed at this stage ([Table 1](#pgen-1000818-t001){ref-type=\"table\"}).\n\nDiscussion {#s3}\n==========\n\nPseudovivipary of *dep* and *pho* occurs in two distinct ways {#s3a}\n-------------------------------------------------------------\n\nMorphological studies in other grasses have revealed that pseudovivipary occurs either by proliferation of the spikelet axis or by transformation of the lemma [@pgen.1000818-Tooke1],[@pgen.1000818-Moore1]. In most cases, pseudovivipary is achieved by the transformation of the spikelet axis.\n\nThe grass spikelet is a structure consisting of two glumes subtending one or more small florets. The rice spikelet is generally considered to have three florets, which are subtended by two tiny glumes (rudimentary glumes) [@pgen.1000818-Yamaguchi1],[@pgen.1000818-Bommert1]. The uppermost floret is fertile while the two lower florets are reduced and sterile. The two empty glumes (or sterile lemmas) are considered to be reduced lemmas of two lower florets [@pgen.1000818-Yamaguchi1],[@pgen.1000818-Bommert1]. So, theoretically, rice spikelet axis is located between the palea and upper empty glume ([Figure 4](#pgen-1000818-g004){ref-type=\"fig\"}). In this study, new shoots in the *dep* mutant are merely found between paleas and upper empty glumes. Thus, we conclude that pseudovivipary in the *dep* mutant is also achieved by the transformation of the spikelet axis.\n\n![Diagrammatic representation of the spikelets of typical grass with three florets (left) and rice (right).\\\nThe arrows indicate the spikelet axes, which are transformed to shoots in *dep* plants.](pgen.1000818.g004){#pgen-1000818-g004}\n\n*Poa alopecurus* and *Poa fuegiana*, which are non-pseudoviviparous and pseudoviviparous species, respectively, can also be recognized as the same species because of the close affinities between them [@pgen.1000818-Moore1]. The characters of *Poa fuegiana* have been well described [@pgen.1000818-Moore1]. A detailed comparison of rice *dep* plant with *Poa fuegiana* shows that there are many similarities between the two pseudoviviparous plants: the palea is reduced or rudimentary; the lemma is elongated; new shoots are only formed on the palea side; both are not stable pseudoviviparous plants; and pseudovivipary mainly happens under high rainfall conditions. Considering so many similarities, it is very likely that the occurrence of pseudovivipary in *Poa fuegiana* and rice *dep* mutant might share the same mechanism. However, the validity of this speculation remains to be verified by molecular investigations on *Poa fuegiana*.\n\nThe *pho* mutant should be classified into the second type of pseudoviviparous plant since the lemma in *pho* undergoes elongation to form the first leaf of the propagule. However, *pho*, which differs from those environment-dependent pseudoviviparous grasses, shows stable pseudovivipary phenotype and is not associated with environmental factors. Till now, to our knowledge, no similar stable pseudoviviparous plant has been reported in nature. If similar stable pseudoviviparous plants are found in nature, they are very likely to be recognized as new species, because of the extreme difference in morphology and reproductive method.\n\nRoles of *OsMADS1* and *OsMADS15* {#s3b}\n---------------------------------\n\nEarly studies have showed that both *OsMADS1* and *OsMADS15* are expressed in the incipient floral primordium [@pgen.1000818-Malcomber1]--[@pgen.1000818-Jeon2],[@pgen.1000818-Kyozuka1]. Furthermore, *OsMADS1* interacts with *OsMADS15* in yeast two-hybrid experiments [@pgen.1000818-Lim1]. The defects of their mutants indicate that *OsMADS1* might work cooperatively with *OsMADS15* to determine FM, but their individual roles are divergent: *OsMADS1* mainly works in promoting the determinacy of FM while *OsMADS15* mainly functions in inhibiting the formation of SAM in incipient floral primordium. Consistent with those indications, the mutations of both *OsMADS1* and *OsMADS15* in *pho* result in a stable inflorescence reversion. In addition, *OsMADS1* is required for the specification of lemma, PMTs and three inner whorls. On the contrary, *OsMADS15* is required for the specification of PMS and empty glumes. So, it is also probably that all floral organs in the double mutant, *pho*, lost their modifications and transformed into their basal state, namely, leaves.\n\nIt has been shown that both transcripts of *OsMADS1* and *OsMADS15* are eventually accumulated in lemma and palea, suggesting that *OsMADS1* and *OsMADS15* might also be involved in the development of lemma and palea [@pgen.1000818-Prasad1],[@pgen.1000818-Kyozuka1]. In severely affected *Osmads1* spikelets, both lemma and palea are affected, but the lemma is affected to a greater extent, suggesting that *OsMADS1* might function as a lemma identity gene [@pgen.1000818-Prasad2],[@pgen.1000818-Kater1]. Additionally, PMTs are lost in *Osmads1* spikelets, indicating that *OsMADS1* is also essential for the specification of PMTs. In contrast, in severely affected *Osmads15* spikelets, both lemma and palea are affected, but the palea is affected to a greater extent and PMS is completely lost, implying that *OsMADS15* might be mainly involved in the specification of PMS. Collectively, both *OsMADS1* and *OsMADS15* might control the differentiation of lemma and palea, but their different roles might contribute to the asymmetric development of the first whorl of rice spikelets.\n\n*OsMADS1* and *OsMADS15* are characterized as *SEP*-like gene and *AP1*/*FUL*-like gene, respectively [@pgen.1000818-Lim1], [@pgen.1000818-Jeon1], [@pgen.1000818-Chen1]--[@pgen.1000818-Kyozuka1]. *AP1*, *FUL* and *SEP1/2/3/4* genes in dicot model plant *Arabidopsis* are also involved in floral meristem identity determination [@pgen.1000818-Ferrandiz1]--[@pgen.1000818-Pelaz1]. In addition, previous studies in *Arabidopsis* have transformed floral organs into leaf-like organs [@pgen.1000818-Ditta1],[@pgen.1000818-Bowman2],[@pgen.1000818-Meyerowitz1]. However, transformation of flowers into true plantlets that is indicative of pseudovivipary has not been found in *Arabidopsis*, but has been reported in many grasses in nature [@pgen.1000818-Tooke1]. The difference might be caused by the distinction of floral development between grasses and dicot plants, as well as the diversification of those floral genes during evolution [@pgen.1000818-Malcomber1],[@pgen.1000818-Yamaguchi1],[@pgen.1000818-Ikeda1].\n\nIs grass flower a modified plantlet meant for reproduction? {#s3c}\n-----------------------------------------------------------\n\nMore than 200 years ago, Goethe proposed that the floral organs are modified leaves. This belief is supported by the observation that triple mutants lacking the ABC genes in *Arabidopsis* have a conversion of all floral organs into leaf-like organs [@pgen.1000818-Bowman2],[@pgen.1000818-Meyerowitz1]. In this study, we revealed that mutations in *OsMADS1* and *OsMADS15* lead to the transformation of all rice flowers into plantlets that can produce true leaves, thereby further confirming Goethe\\'s hypothesis. The complete transformation of flowers into juvenile plantlets in rice, as well as similar transformations in other grasses, leads us to hypothesize that in grasses a flower may be a modified juvenile plantlet meant for reproduction.\n\nIt is widely accepted that sexual reproduction evolves from asexual reproduction, so we speculate that *pho* might be an atavistic mutant, and plants with similar phenotype might play an important role in the evolution of reproductive strategy from asexual to sexual. The *dep* mutant, which can produce both flowers and plantlets, is more similar to most natural pseudoviviparous plants than the *pho* mutant. Thus, its analogous plants might play an intermediate role in this evolution, because such environment-dependent pseudoviviparous plant has the ability not only to reproduce via sexual way under favourable conditions, but also to reproduce via asexual way when the harsh conditions affect its sexual reproduction.\n\nIn conclusion, we have shown that *dep* is a genetic mutant in *OsMADS15* while *afo* is an epigenetic mutant in *OsMADS1*, and their combination led to stable pseudovivipary. These findings suggest that the two MADS-box genes might play important roles in plant adaptation to various reproductive strategies.\n\nMaterials and Methods {#s4}\n=====================\n\nPlant materials {#s4a}\n---------------\n\nAll plant materials were grown in individual lines in paddy fields to monitor climate-change triggered pseudovivipary. In summer, all materials were planted in Beijing and Yangzhou, while, in winter, all materials were grown in Hainan Island in South China.\n\nPrimers {#s4b}\n-------\n\nThe primers used in this study are listed in [Table S1](#pgen.1000818.s008){ref-type=\"supplementary-material\"}.\n\nMolecular cloning of *DEP* {#s4c}\n--------------------------\n\nTo fine map *DEP*, STS markers (P1--P8) were developed based on sequence differences between *indica* variety 9311 and *japonica* variety Nipponbare according to the data published in <http://www.ncbi.nlm.nih.gov>.\n\nConstruction of RNA interference and rice transformation {#s4d}\n--------------------------------------------------------\n\nA 317-bp fragment of *OsMADS15* was amplified by PCR with their specific primers; this fragment was cloned into the pGEM-T vector (Promega) and sequentially cloned into the *Bam*HI*/Sal*I and *Bgl*II*/Xho*I sites of the pUCRNAi vector. Subsequently, the stem-loop fragment was cloned into the pCAMBIA2300-Actin vector. The resulting RNAi construct was transformed into an *A. tumefaciens* strain and used for further rice transformation.\n\nSubcellular localization {#s4e}\n------------------------\n\nThe amplified coding region of *OsMADS15* of both wild-type and *dep* was fused with green fluorescent protein (GFP) and cloned into the *Hind*III/*Bam*HI sites of the vector pJIT163. Those plasmids were bombarded into onion epidermal cells using a PDS-1000/He particle gun (Bio-Rad). The expression constructs were also transfected into rice Nipponbare protoplasts. Twenty hours after transfection, protein expression was observed and images were captured with a Zeiss LSM 510 Meta confocal laser scanning microscope.\n\nTranscriptional activation assay {#s4f}\n--------------------------------\n\nWe carried out the transcriptional activation assay using a MATCHMAKER LexA Two-Hybrid system (Clontech). Different length sequences were amplified and fused in frame to the pLexA to construct pOsMADS15, pOsMADS15-*dep*, pOsMADS15\u25b3C180-267 and pOsMADS15\u25b3N1-66. All constructs were used to transform the recipient strain EGY48 with p8op-lacZ. Transformants were selected on *Ura*/*His* depleted plates at 30\u00b0C for 3 days. The activation ability was assayed on Gal/Raf (*Ura^\u2212^*/*His^\u2212^*)/X-gal to test the activation of the *LacZ* reporter gene for 3 days.\n\nAffymetrix GeneChip hybridization and data analysis {#s4g}\n---------------------------------------------------\n\nIn order to generate gene expression profiles of WT and the *pho* mutant, we conducted 57K Affymetrix rice whole genome array. The total RNA of rice panicle (5--8 cm) samples was isolated using TRizol reagent (Invitrogen) and purified using Qiagen RNeasy columns (Qiagen). All the processes for cDNA and cRNA synthesis, cRNA fragmentation, hybridization, staining, and further scanning, were conducted according to the GeneChip Standard Protocol (Eukaryotic Target Preparation, Affymetrix). 5 ug of total RNA was used for making biotin-labeled cRNA targets. 10 ug of cRNA was hybridized for 16 h at 45\u00b0C on GeneChip Rice Genome Array. GeneChips were washed and stained in the Affymetrix Fluidics Station 450. GeneChip were scanned using the Affymetrix GeneChip Scanner. The information about GeneChip Rice Genome Array (MAS 5.0) could be accessed from Affymetrix [website: http://www.affymetrix.com/products_services/arrays/specific/rice.affx](http://www.affymetrix.com/products_services/arrays/specific/rice.affx). GCOS software (Affymetrix GeneChip Operating Software) was used for data collection and normalization. The overall intensity of all probe sets of each array was scaled to 500 to guaranty that hybridization intensity of all arrays was equivalent, each probe set was assigned with present \"P\", absent \"A\" and marginal \"M\" and p-value from algorithm in GCOS. The microarray data has been deposited in the Gene Expression Omnibus (GEO) of NCBI under accession GSE17194.\n\nPhylogenetic analysis {#s4h}\n---------------------\n\nAll MADS-box proteins were retrieved by BLAST searches using the conserved M-, I-, K-domain regions (174 amino acids) of OsMADS15 protein (<http://www.ncbi.nlm.nih.gov>). Protein sequences were aligned using the CLUSTALX 1.83 [@pgen.1000818-Thompson1]. The phylogenetic tree was constructed using the Molecular Evolution and Genetic Analysis (MEGA) package version 3.1 [@pgen.1000818-Kumar1].\n\nMorphological analysis {#s4i}\n----------------------\n\nFor SEM, samples were fixed overnight at room temperature with 2.5% glutaraldehyde in a 0.1 M phosphate buffer (pH 7.4) and dehydrated through an ethanol series. Then the samples were replaced by isoamyl acetate, critical point dried, sputter coated with gold, and observed with a scanning electron microscope. For histology, samples were fixed in FAA (5% formaldehyde, 5% glacial acetic acid and 63% ethanol) overnight at 4\u00b0C, dehydrated in a graded ethanol series, embedded in Technovit 7100 resin (Hereaus Kulzer) and polymerized at room temperature. Transverse sections were performed using an Ultratome III ultramicrotome (LKB), stained with 0.25% toluidine blue (Chroma Gesellshaft Shaud) and photographed using an Olympus BX61 microscope.\n\nQuantitative real-time PCR {#s4j}\n--------------------------\n\nTotal RNA was extracted from rice young panicles (5--8 cm) using TRIZOL reagent (Invitrogen) as described by the supplier. 3 \u00b5g RNA was reverse-transcribed with Oligo-dT(18) primer using the superscript II RNaseH reverse transcriptase (Invitrogen). For quantitative real-time RT-PCR, first strand cDNAs were used as templates in real-time PCR reactions using the SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer\\'s instructions. The amplification of the target genes were analyzed using the ABI Prism 7000 Sequence Detection System and Software (PE Applied Biosystems). Ubiquitin was used as a control to normalize all data.\n\nBisulfite sequencing {#s4k}\n--------------------\n\nFive micrograms genomic DNA isolated from panicles (5--8 cm) was digested with *EcoR*I and *Pst*I. After centrifugation, pellets were dissolved in 50 \u00b5L of water, heated at 95\u00b0C for 15 min, and quenched on ice. Fifty microliters of NaOH (3 M) was added and incubated at 37\u00b0C for 30 min, followed by the addition of 565 \u00b5L bisulfite solution to the denatured DNA. Samples were treated at 55\u00b0C for 20 h. After being purified using a Wizard DNA clean-up system (Promega), 50 \u00b5L bisulfite-treated DNA was added with 5 \u00b5L NaOH (3 M) and incubated at 37\u00b0C for 15 min. The Bisulfite-treated DNA was precipitated with ammonium acetate and ethanol, and the pellets were dissolved in 50 \u00b5L of water. PCR analysis was performed at 50\u00b0C using four primer sets (BSP1-4). PCR products were cloned into PMD18-T vectors. Ten clones of each product were sequenced to determine the methylation ratio. Cytosine methylation was only found in the BSP1 region.\n\nSupporting Information {#s5}\n======================\n\n###### \n\nThe plantlets formed in *pho* panicle show normal vegetative growth when explanted in paddy fields. (A) Young plantlets formed in *pho* panicle. (B) The emergence of normal roots in those plantlets after being replanted in field for three days. (C) The spikelet of WT. (D) The spikelets of ACT::RNAi*MADS15* plants.\n\n(1.41 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nOsMADS15-GFP fusion protein and Osmads15 (*dep*)-GFP fusion protein in rice protoplast. Bars: 5 \u00b5m.\n\n(0.24 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n*dep/nsr* double mutant shows a similar pseudovivipary phenotype to that of the *pho* mutant. (A) The panicles of *dep* (left), *nsr* (center) and *dep/nsr* (right) plants. (B) Young plantlet formed in dep/nsr panicle. (C) The emergence of normal roots in this plantlet after being replanted in field for two days.\n\n(1.48 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nPhylogenetic tree of deduced amino acid sequences shows that *OsMADS15* is an *AP1/FUL*-like gene. Phylogenetic tree construction was performed using the M, I, and K domains of these proteins.\n\n(0.23 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nAlignment of full-length sequences of *OsMADS15* with *AP1/FUL*-like proteins in other grass species and *Arabidopsis*. Black boxes indicate identical amino acids, and gray boxes indicate similar amino acids. The red box indicates the position of the amino acid substitution in *dep* and *pho* mutant.\n\n(3.13 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nPalea is more severely affected than lemma in *dep* spikelet. (A--C), SEM of the lemma (A), palea (B) and glume (C) epidermis of WT spikelet; (D--F), SEM of the lemma (D), palea (E) and glume (F) epidermis of severely affected *dep* spikelet. Scale bar is 10 \u00b5m in all panels.\n\n(1.75 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nNew shoots occasionally emerge from dep spikelets. (A) *dep* plant with emerged shoots (white arrows) in some spikelets. (B) *dep* spikelet with an emerging tiller (white arrow) between palea and upper empty glume. (C) SEM of the emerging tiller (white arrow) in *dep* spikelet. The upper empty glume has been removed. Bar is 0.5 mm (D) Tillers formed in *dep* spikelets. (E) The emergence of normal roots in those tillers after replanting in field for two days.\n\n(2.45 MB TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\nThe primers used in this study.\n\n(0.04 MB DOC)\n\n###### \n\nClick here for additional data file.\n\nWe thank Edwards Allen in Monsanto for critical reading of the manuscript. We also thank P. Wu in Zhejiang University for providing *nsr* seeds.\n\nThe authors have declared that no competing interests exist.\n\nThis work was supported by grants from the Ministry of Sciences and Technology of China (2005CB120805 and 2006AA10A101) and the National Natural Science Foundation of China (30530070 and 30621001). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: KW MG YX ZC. Performed the experiments: KW DT LH WX JH ML. Analyzed the data: KW. Contributed reagents/materials/analysis tools: DT. Wrote the paper: KW ZC.\n"}
+{"text": "Background\n==========\n\nMalaysia is one of the top ten countries in the world with the greatest number of diabetes patients \\[[@B1]\\] and the prevalence has increased dramatically from 8.3% in 1996 to 14.9% in 2006, affecting 1.4 million adults aged\u2009\u2265\u200930 years old \\[[@B2]\\]. According to the results of DiabCare Malaysia 2008, Type 2 diabetes (T2DM) accounts for more than 90% of all cases in adults \\[[@B3]\\]. Hypertension is a common comorbidity in T2DM patients, with a prevalence of up to two-thirds of the population, and it may be present by the time T2DM is diagnosed or even before the onset of hyperglycemia \\[[@B4]\\]. Hypertension enhances the risk of cardiovascular disease in T2DM patients \\[[@B4]\\]. It also increases the risk of developing microvascular complications such as diabetic nephropathy and retinopathy \\[[@B5],[@B6]\\].\n\nTo minimize the risk of complications, many guidelines recommend a target blood pressure (BP) of\u2009\u2264\u2009130/80 mmHg in all T2DM patients with hypertension \\[[@B4],[@B7],[@B8]\\]. However, achieving this target BP remains a great challenge and the majority of the patients require one or more antihypertensive agents in order to achieve this optimal BP control \\[[@B9]\\].\n\nT2DM patients with hypertension often receive multiple medications and this can lead to the occurrence of drug-related problems (DRPs) \\[[@B9]\\]. A high prevalence of DRPs has been observed in T2DM patients \\[[@B10],[@B11]\\]. DRPs may lead to suboptimal blood pressure control \\[[@B12]\\] which can contribute to significant morbidity or mortality, prolonged hospitalization, and increased health care expenditure if left unresolved \\[[@B13]\\]. However, in most cases, these DRPs are preventable \\[[@B14]\\].\n\nThere are several factors influencing DRPs in T2DM patients with hypertension. Polypharmacy (\u2265 5 concurrent medications) is an inherent factor as high blood pressure and diabetic complications usually complicate the treatment \\[[@B9]\\]. Age status (\u2265 65 years old) is another factor, given increased association with multiple medical conditions, multiple drug therapies and age-related changes in the pharmacokinetics and pharmacodynamics of drugs \\[[@B9]\\]. Multiple medical conditions and renal impairment also have been shown to contribute to DRPs \\[[@B15],[@B16]\\].\n\nTo date, there has been a lack of studies conducted locally and globally to investigate and document DRPs in T2DM patients with hypertension. Our study's aim was to provide baseline data regarding DRPs to allow the implementation of more effective management and to reduce the mortality and morbidity associated with DRPs.\n\nObjectives\n----------\n\n1\\. To assess the drug-related problems in type 2 diabetes patients with hypertension.\n\n2\\. To identify the factors influencing drug-related problems in type 2 diabetes patients with hypertension.\n\nMethods\n=======\n\nStudy design and setting\n------------------------\n\nThis was a retrospective study conducted in Malaysia's premier teaching hospital with 1000-beds, the University of Malaya Medical Centre (UMMC).\n\nStudy population\n----------------\n\nA total of 200 patients were included in this study. The sample size was calculated using the Epi Info, Version 6 (Centers for Disease Control and Prevention, Atlanta GA), which provided a minimum sample size of 195 patients. The study population consisted of T2DM patients with hypertension who fulfilled the requirements of the International Classification of Diseases Tenth Revision (ICD-10) code E11.0-11.9 for T2DM and who were admitted to the UMMC from January 2009 to December 2011. The study complied with the Declaration of Helsinki and the ethics committee of UMMC granted its approval. An overview of the study procedure is shown in Figure [1](#F1){ref-type=\"fig\"}.\n\n![Overview of study procedure.](1472-6823-13-2-1){#F1}\n\n### Inclusion criteria\n\n1\\. Patients diagnosed with T2DM and hypertension\n\n2\\. Patients who received at least one antidiabetic drug (oral antidiabetic drug or insulin) and at least one antihypertensive agent in the ward\n\n3\\. Adult patients aged 18 years old and above\n\n### Exclusion criteria\n\n1\\. Patient with missing data\n\nData collection\n---------------\n\nData were collected by the authors who are pharmacists. Demographic characteristics such as age, gender, ethnicity, height, weight, and body mass index were recorded. Clinical characteristics such as duration of hospital stay, duration of T2DM, duration of hypertension, presence of diabetic complications (referred to diabetic retinopathy, neuropathy or diabetic foot ulcer), presence of comorbidities, laboratory results and concurrent medications were also collected.\n\n### Definition used in the study\n\n\u25cf\u2002Comorbidities were defined as chronic illnesses or diseases which require long term treatment \\[[@B17]\\].\n\n\u25cf\u2002Cardiovascular events referred to the presence of acute coronary syndromes, ischemic heart disease, heart failure, arrhythmias, cardiomyopathy or as stated in the medical records \\[[@B18]\\].\n\n\u25cf\u2002Renal impairment was defined as chronic kidney disease, chronic interstitial nephritis, chronic glomerulonephritis, creatinine clearance of less than 35 mL/min, diabetic nephropathy, nephrosclerosis or as stated in the medical records \\[[@B19]\\]. The assessment of the creatinine clearance with the use of drugs was based on BNF and Lexicomp.\n\n\u25cf\u2002Liver impairment referred to liver cirrhosis, chronic hepatitis, hepatocellular carcinoma, elevations of liver enzymes (more than 3 times the upper normal limits) or as stated in the medical records \\[[@B19],[@B20]\\].\n\n\u25cf\u2002Polypharmacy was defined as the use of five or more medications \\[[@B21]\\].\n\n\u25cf\u2002DRP is defined as \"an event or circumstance involving drug therapy that actually or potentially interferes with desired health outcomes\" \\[[@B22]\\].\n\n\u25cf\u2002Significant potential drug interactions are defined as interactions that potentially can cause harm to patient and are well documented or can cause moderate harm without well documented studies \\[[@B23]\\].\n\nNote: The same drug used with different strengths was counted as one item whereas the same drug with different routes of administration was counted as separate items. Combinations of drugs were counted as a single item. However, this did not apply to antidiabetic and antihypertensive drugs, where the number of drugs used was counted according to the number of classes.\n\nClassification, identification and assessment of DRPs\n-----------------------------------------------------\n\nThe Pharmaceutical Care Network Europe (PCNE) classification of DRPs version 5.01 \\[[@B22]\\] was used to categorize DRPs. It is an established system that has been revised several times and its validity and reproducibility have been tested \\[[@B11],[@B24]\\]. It has been used by many recent studies \\[[@B11],[@B21],[@B25]\\].\n\nIn this study, the six domains of problems of the PCNE classification were used. The DRPs and their possible causes were identified from the patients' medical records, with reference to the standard guidelines and established literatures \\[[@B4],[@B7],[@B8],[@B26],[@B27]\\]. Two main references were used to assess the appropriateness of drug indications, appropriateness of drug and dosage, possible drug interactions, adverse drug reactions and contraindications \\[[@B28],[@B29]\\]. The authors who are pharmacists were involved in the identification and classification of DRPs.\n\n### Modified Beers criteria\n\nThe modified Beers criteria \\[[@B30],[@B31]\\] were used in this study. This is a consensus-based drug list that includes a number of drugs which should be avoided or used very cautiously in the elderly. For this study, the criteria were used as a reference to assess and identify the potential drugs that were inappropriately prescribed in the T2DM patients with hypertension who were aged 65 and above. The listed drugs were generally divided into low and high risk. In this study, only inappropriate prescriptions of \"Beers criteria high severity\" drugs were identified as DRPs because these drugs might pose clinically significant adverse effects when used in the elderly.\n\nStatistical Techniques\n----------------------\n\nThe statistical software Statistical package for Social Science (SPSS) version 19 (SPSS Inc., Chicago, IL, USA) was used to analyze all the data collected and extracted in this study. Categorical data were expressed as percentages and continuous data were expressed as mean\u2009\u00b1\u2009standard deviation. Independent Student's *t-*tests were used to compare group means of continuous dependent variables while the Kolmogorov-Smirnov Test of Normality were used to test the distribution of a sample.\n\nIn addition, the association or correlation between two categorical variables was examined using the chi-square test of independence. Statistical significance was defined as *p*-value\u2009\\<\u20090.05. Refer to Figure [1](#F1){ref-type=\"fig\"}.\n\nResults\n=======\n\nA total of 200 patients were included in this study. The mean age of patient was 62.3\u2009\u00b1\u200912.7 years old and the non-elderly patients were 15.0% more numerous than the elderly patients. The minimum and maximum ages of the patients were 31 and 95 years old, respectively. Table [1](#T1){ref-type=\"table\"} demonstrates the demographic and clinical characteristics of this study population. There was an average of 4.6\u2009\u00b1\u20091.3 chronic illnesses per patient. The mean number of medications was 6.9\u2009\u00b1\u20092.8. The number of medications taken by patients ranged from 3 to 20. Polypharmacy was common as 76.5% of the patients were taking five or more medications, of whom 46.4% were elderly patients.\n\n###### \n\nDemographic and clinical characteristic of the patients (N\u2009=\u2009200)\n\n  **Characteristics**                         **Number of patients (Percentage,%)**\n  ------------------------------------------ ---------------------------------------\n  **Sex**                                    \n  Male                                                      103(51.5)\n  Female                                                    97 (48.5)\n  **Age**                                    \n  Non-elderly                                              115 (57.5)\n  Elderly                                                   85 (42.5)\n  **Duration of hospital stay**              \n  Not more than 7 days                                     143 (71.5)\n  8 to 14 days                                              38 (19.0)\n  More than 15 days                                         19 (9.5)\n  **Duration of type 2 diabetes mellitus**   \n  Not more than 10 years                                    72 (36.0)\n  11 to 20 years                                            47 (23.5)\n  21 to 30 years                                            31 (15.5)\n  Unknown duration                                          50 (25.0)\n  **Duration of hypertension**               \n  Not more than 10 years                                    82 (41.0)\n  11 to 20 years                                            43 (21.5)\n  21 to 30 years                                            22 (11.0)\n  Unknown duration                                          53 (26.5)\n  **A1c**                                    \n  Achieved target (\\< 6.5)                                  48 (24.0)\n  Did not achieve target (\u2265 6.5)                           146 (73.0)\n  Unknown                                                    6 (3.0)\n  **Diabetic complications\\***               \n  Diabetic retinopathy                                      43 (21.5)\n  Diabetic foot ulcer                                       21 (10.5)\n  Diabetic neuropathy                                       17 (8.5)\n  **Comorbidities**^**\u2020**^                   \n  Renal impairment                                         100 (50.0)\n  Cardiovascular disease                                    96 (48.0)\n  Dyslipidemia                                              64 (32.0)\n  Stroke                                                    40 (20.0)\n  Gastrointestinal disease                                  14 (7.0)\n  Liver impairment                                          11 (5.5)\n  Bronchial asthma                                           7 (3.5)\n  Benign prostatic hyperplasia                               7 (3.5)\n  Gouty arthritis                                            6 (3.0)\n  Osteoarthritis                                             5 (2.5)\n\n^\u2020^One patient may have more than one diabetic complication or comorbidity.\n\nPatients on combination therapy were 4.5% more numerous than patients on monotherapy, and the most frequently used drug classes were calcium channel blockers, ACE inhibitors and diuretics. Amlodipine, perindopril and frusemide were the most widely used agents representing these classes, respectively.\n\nThe dual therapy that had the highest frequency was ACE inhibitors plus calcium channel blockers (5%), followed by ACE inhibitors plus beta blockers (4.5%). The combination of ACE inhibitors, calcium channel blockers and beta blockers (4%) was the most commonly used triple therapy. Additionally, 2 out of the 19 patients received antihypertensive agents from 5 different classes.\n\nMore than two-thirds of the patients were on monotherapy with oral antidiabetic agents or insulin. Insulin was more widely used than oral agents, with a difference of 17.5%. For combination therapy, oral agents plus insulin (17.5%) appeared to be the most common combination. Of these, insulin plus a single antidiabetic agent had the highest frequency (14%).\n\nInsulin was the most common antidiabetic prescribed. The most frequently used oral agents were biguanides (metformin) and sulfonylureas. The use of acarbose (alpha-glucosidase inhibitors) and sitagliptin (dipeptidyl peptidase-4 inhibitors) were 3.2% and 0.6%, respectively.\n\nA total number of 387 DRPs were identified (Table [2](#T2){ref-type=\"table\"}). There was an average of 1.9\u2009\u00b1\u20091.2 problems and 1.7\u2009\u00b1\u20091.1 causes per patient. A total of 90.5% of the patients had at least one DRP. The most frequent categories of DRPs were \"others\", drug choice problems and drug interactions. The mean number of chronic illnesses was 2.65\u2009\u00b1\u20091.3 in the group of patients with DRPs and 2.1\u2009\u00b1\u20091.1 in the group of patient without DRPs. There was a significant difference (p\u2009=\u20090.039) between the two group when tested using the independent sample *t*-test. Under the \"others\" category, the most frequent problem encountered was the \"insufficient awareness of health and diseases,\" accounting for 22% of all the cases. On the other hand, most of the drug choice problems were the results of inappropriate drug selection and the use of contraindicated drugs.\n\n###### \n\nDrug related problems in type 2 diabetes patients with hypertension (n\u2009=\u2009387)\n\n  **Code**   **Problems\\***                                                                         **Number of problem (Percentage,%)**\n  ---------- ------------------------------------------------------------------------------------- --------------------------------------\n  **P1**     **Adverse reactions**                                                                              **25 (6.5)**\n  P1.1       Side effects suffered (non-allergic)                                                                 25 (6.5)\n  **P2**     **Drug choice problems**                                                                          **87 (22.5)**\n  P2.1       Inappropriate drug (not most appropriate for indication)                                             34 (8.8)\n  P2.2       Inappropriate drug form (not most appropriate for indication)                                        4 (1.0)\n  P2.3       Inappropriate duplication of therapeutic group or active ingredient                                  4 (1.0)\n  P2.4       Contraindication for drug                                                                            29 (7.5)\n  P2.5       No clear indication for drug use                                                                     1 (0.3)\n  P2.6       No drug but clear indication                                                                         15 (3.9)\n  **P3**     **Dosing problems**                                                                               **62 (16.0)**\n  P3.1       Drug dose too low or dosage regime not frequent enough                                               5 (1.3)\n  P3.2       Drug dose too high or dosage regime too frequent                                                    44 (11.3)\n  P3.3       Duration of treatment too short                                                                      8 (2.1)\n  P3.4       Duration of treatment too long                                                                       5 (1.3)\n  **P4**     **Drug use problems**                                                                             **50 (12.9)**\n  P4.1       Drug not taken/administered at all                                                                  50 (12.9)\n  **P5**     **Interactions**                                                                                  **63 (16.3)**\n  P5.1       Potential interaction                                                                               63 (16.3)\n  **P6**     **Others**                                                                                        **100 (25.8)**\n  P6.1       Patient dissatisfied with therapy despite taking drug(s) correctly                                   5 (1.3)\n  P6.2       Insufficient awareness of health and diseases (possibly leading to future problems)                 90 (23.2)\n  P6.4       Therapy failure (reason unknown)                                                                     5 (1.3)\n\n\\*Only problems that have a frequency of more than one were included.\n\nA total of 25 adverse reactions were reported (Table [2](#T2){ref-type=\"table\"}). Antidiabetic drugs were associated with about one-third of all the cases. Eight patients experienced hypoglycemia secondary to either oral antidiabetic drugs or insulin. Tremor secondary to insulin was also recorded. Antihypertensive agents that caused adverse reactions were calcium channel blockers, diuretics and ACE inhibitors. It was reported that amlopidine had caused increased heart rate and bilateral leg swelling. Electrolyte imbalances were reported as adverse reactions of perindopril (hyperkalemia) and indapamide (hyponatremia). Also, urinary hesitancy secondary to hydrochlorothiazide had been reported. One patient suffered from rhabdomyolysis secondary to a lipid lowering agent. Drugs that did not fall into the above three classes were classified as \"others\". For example, bleeding gums and hematuria secondary to warfarin had been reported. Also, steroids had been associated with two cases of adverse reactions. Among the 87 drug choice problems identified, most of the problems were related to inappropriate drug choice and contraindications. The distribution of the drug choice problem is displayed in Table [2](#T2){ref-type=\"table\"}. Antihypertensive drugs were not used appropriately. The inappropriate choice of antihypertensive drugs resulted in 19 drug choice problems. For example, ACE inhibitors were used in five patients with ESRD. Alpha blockers such as prazosin and doxazosin were used as second or third add-on therapies when other better alternatives were available and not contraindicated (7 patients). Short-acting nifedipine was used in two elderly patients and spironolactone was prescribed to two renal impairment patients with creatinine clearances of less than 30 mL/minute. Aspirin was also given to patients with severe renal impairment who had creatinine clearances of less than 10 mL/minute.\n\nIn addition to these, there were several drugs which were inappropriately used in the elderly and were classified as high risk under the modified Beers criteria, namely, ticlopidine (4 cases), amitriptyline (2 cases), diphenhydramine (2 cases), liquid paraffin (2 cases), chlorpheniramine, and others.\n\nTwenty-eight drug choice problems involving the use of contraindicated drugs were identified. The three most common drugs that were most prescribed in the presence of contraindications were metformin, aspirin and fondaparinux. For instance, metformin was prescribed in 18 patients for whom it was contraindicated (e.g., creatinine clearance less than 30 mL/minute, more than 80 years old, or recent myocardial infarction). Also, gliclazide was prescribed to five patients with severe liver impairment and aspirin was given to patients who had a previous history of allergy. Three patients received fondaparinux despite their poor renal function (creatinine clearances of less than 30 mL/minute).\n\nThere were 15 cases where no drugs were prescribed for a clear indication. For example, hematinics were indicated in ESRD patients with chronic anemia but no drugs were prescribed (10 cases). Also, aspirin or simvastatin were not given to some of the patients as secondary prophylaxis of CVD.\n\nOf the 62 cases of inappropriate dosing identified, most of the drugs were prescribed at a higher dose than required, particularly in patients with existing renal or liver impairments. The most common drugs that were involved in wrong dosages were H~2~-antogonists (20 cases), antibiotics (8 cases), antihypertensive agents (6 cases), antidiabetic drugs (3 cases) and others. Ranitidine was commonly prescribed at a higher dosage than required in patients with creatinine clearances of less than 50 mL/minute.\n\nIn addition to wrong dosages, inappropriate durations of treatment were also identified. Durations of treatment were too short in eight cases whereas longer than required in four cases. For example, oseltamivir (Tamiflu\u00ae) was only given for three days for the treatment of H1N1 and oral azithromycin 500 mg was administered once daily for six days.\n\nOne-quarter of the patients had at least one drug use problem and this made up approximately 13% of all the DRPs. It was recorded in the patients' medical records that drugs were not taken or administered by them prior to admission. Non-compliance with antihypertensive agents and antidiabetic drugs was frequent.\n\nThere were a total of 63 significant potential drug interactions identified. The use of aspirin and clopidogrel posed a significant potential drug interaction in 25 patients. In addition, there were 14 patients who were prescribed simvastatin at more than 20 mg while receiving amlodipine. In this study, the drugs most implicated in drug interactions were aspirin (32 cases), clopidogrel (31 cases), simvastatin (23 cases), amlodipine (15 cases), omeprazole (15 cases), and iron salts (9 cases).\n\nAbout one-quarter of the DRPs that could not be classified under any other category were regarded as \"Others\". The majority of patients had insufficient awareness of health and diseases which could possibly lead to future problems. For example, many patients had a lack of knowledge about T2DM. Also, they were unaware of the management of the disease and its complications.\n\nThe total number of causes identified was 336 (Table [3](#T3){ref-type=\"table\"}). The causes varied with the problems identified. Some problems might have more than one cause whereas some might not have any causes. For example, most of the drug interactions did not have a cause. Of all the causes identified, inappropriate drug selection was the cause with the highest frequency, followed by inappropriate dosage selection and burden of therapy.\n\n###### \n\nCauses of DRPs in T2DM patients with hypertension (n\u2009=\u2009336)\n\n  **Code**   **Causes\\***                                                      **Number of problem (Percentage,%)**\n  ---------- ---------------------------------------------------------------- --------------------------------------\n  **C1**     **Drug/Dose selection**                                                      **158 (47.0)**\n  C1.1       Inappropriate drug selection                                                   70 (20.8)\n  C1.2       Inappropriate dosage selection                                                 58 (17.2)\n  C1.5       Synergistic/preventive drug required and not given                              15 (4.5)\n  C1.8       Manifest side effect, no other cause                                            15 (4.5)\n  **C2**     **Drug use process**                                                         **36 (10.7)**\n  C2.1       Inappropriate timing of administration and/or dosing intervals                  2 (0.6)\n  C2.2       Drug underused/under-administered                                               33 (9.8)\n  C2.6       Patient unable to use drug/form as directed                                     1 (0.3)\n  **C3**     **Information**                                                               **13 (3.9)**\n  C3.1       Instructions for use/taking not known                                           1 (0.3)\n  C3.2       Patient unaware of reason for drug treatment                                    6 (1.8)\n  C3.4       Patient unable to understand local language                                     6 (1.8)\n  **C4**     **Patient/Psychological**                                                    **124 (36.9)**\n  C4.1       Patient forgets to use/take drug                                                14 (4.2)\n  C4.2       Patient has concerns with drugs                                                 4 (1.2)\n  C4.3       Patient suspects side-effect                                                    10 (3.0)\n  C4.5       Patient unwilling to bother physician                                           6 (1.8)\n  C4.7       Patient unwilling to adapt life-style                                           28 (8.3)\n  C4.8       Burden of therapy                                                              47 (14.0)\n  C4.9       Treatment not in line with health beliefs                                       15 (4.4)\n  **C5**     **Logistic**                                                                  **5 (1.5)**\n  C5.1       Prescribed drug not available (anymore)                                         3 (0.9)\n  C5.2       Prescribing error (only in case of slip of the pen)                             2.(0.6)\n\n\\*Only causes that have frequency of more than one were included.\n\nWhen matching the cause to the problem, inappropriate drug or dose selection was found to be the most common cause for drug choice problems and dosing problems. The main cause for adverse reaction was \"patient or psychological\", particularly \"patient has concerns with drugs\". Drug use problems were commonly caused by \"patient or psychological\", followed by \"drug use process\". Similarly, \"patient or psychological\" was the most common cause for \"other\" problems.\n\nThere was no statistical significance identified when comparing factors with DRPs (Tables [4](#T4){ref-type=\"table\"}, [5](#T5){ref-type=\"table\"}). Despite this, several factors were found to have a statistically significant association with the 6 domains of DRPs. The elderly had a positive statistical association with drug choice problems (*p*\u2009\\<\u20090.001). Also, there were significant associations between renal impairment and drug choice problems (*p*\u2009=\u20090.029) or dosing problems (*p*\u2009=\u20090.027).\n\n###### \n\nComparison between factors and occurrence of adverse reactions, drug choice problems, and dosing problems\n\n  **Factors**                        **Adverse reactions (n\u2009=\u200925) Frequency (Percentage, %)**   **Drug choice problem (n\u2009=\u200976) Frequency (Percentage, %)**   **Dosing problem (n\u2009=\u200955) Frequency (Percentage, %)**                                                                    \n  --------------------------------- ---------------------------------------------------------- ------------------------------------------------------------ ------------------------------------------------------- ------------ ----------- --------------- ----------- ------------ ---------------\n  \u00a0                                                          **Yes**                                                      **No**                                                 ***p*-value**                        **Yes**      **No**     ***p*-value**    **Yes**      **No**     ***p*-value**\n  **Elderly**                                                                                                                                                                                                                                                                         \n  Yes                                                       11 (44.0)                                                   74 (42.3)                                                 \\> 0.999^a^                        28 (36.8)    87 (70.2)   \\< 0.001^a^\\*   34 (61.8)   81 (70.4)      0.548^a^\n  No                                                        14 (56.0)                                                   101(57.7)                                                  48 (63.2)                         37 (29.8)    21 (38.2)     64 (29.6)                             \n  **Polypharmacy**                                                                                                                                                                                                                                                                    \n  Yes                                                       19 (76.0)                                                   134 (76.6)                                                \\> 0.999^a^                        60 (78.9)    93 (75.0)     0.640^a^      38 (69.1)   115 (79.3)     0.182^a^\n  No                                                         6 (26.0)                                                   41 (23.4)                                                  16 (21.1)                         31 (25.0)    17 (30.9)     30 (20.7)                             \n  **Duration of hospital stay**                                                                                                                                                                                                                                                       \n  \u2264 1 week                                                  17 (68.0)                                                   126 (72.0)                                                 0.859^a^                          45 (59.2)    98 (79.0)    0.004^a^\\*     40 (72.7)   103 (71.0)     0.951^a^\n  \\> 1 week                                                  8 (32.0)                                                   49 (28.0)                                                  31 (40.8)                         26 (21.0)    15 (27.3)     42 (29.0)                             \n  **Microvascular complications**                                                                                                                                                                                                                                                     \n  Yes                                                       17 (68.0)                                                   107 (61.1)                                                 0.660^a^                          53 (69.7)    71 (57.3)     0.106^a^      40 (72.7)   84 (57.9)      0.078^a^\n  No                                                         8 (32.0)                                                   68 (38.9)                                                  23 (30.3)                         53 (42.7)    15 (27.3)     61 (42.1)                             \n  **Cardiovascular events**                                                                                                                                                                                                                                                           \n  Yes                                                       12 (48.0)                                                   109 (62.3)                                                 0.193^b^                          47 (61.8)    74 (59.7)     0.877^a^      36 (65.5)   85 (58.6)      0.471^a^\n  No                                                        13 (52.0)                                                   66 (37.7)                                                  29 (38.2)                         50 (40.3)    19 (34.5)     60 (41.4)                             \n  **Renal impairment**                                                                                                                                                                                                                                                                \n  Yes                                                       12 (48.0)                                                   88 (50.3)                                                 \\> 0.999^a^                        46 (60.5)    54 (43.5)    0.029^a^\\*     35 (63.6)   65 (44.8)     0.027^a^\\*\n  No                                                        13 (52.0)                                                   87 (49.7)                                                  30 (39.5)                         70 (56.5)    20 (36.4)     80 (55.2)                             \n  **Liver impairment**                                                                                                                                                                                                                                                                \n  Yes                                                        1 (4.0)                                                     10 (5.7)                                                 \\> 0.999^b^                         6 (7.9)      5 (4.0)      0.399^a^       5 (9.1)     6 (4.1)       0.306^a^\n  No                                                        24 (96.0)                                                   165 (94.3)                                                 70 (92.1)                         119 (96.0)   50 (90.9)    139 (95.9)                             \n  **Hyperlipidemia**                                                                                                                                                                                                                                                                  \n  Yes                                                        7 (28.0)                                                   57 (32.6)                                                  0.819^a^                          17 (22.4)    47 (37.9)    0.033^a^\\*     15 (27.3)   49 (33.8)      0.476^a^\n  No                                                        18 (72.0)                                                   118 (67.4)                                                 59 (77.6)                         77 (62.1)    40 (72.7)     96 (66.2)                             \n\n^a^ Computed using Continuity Correction; ^b^ Computed using Fisher's Exact Test; \\* Statistically significant (*p*\u2009\\<\u20090.05).\n\n###### \n\nComparison between factors and occurrence of drug use problems, drug interactions, and other problems\n\n             **Factors**             **Drug use problems (n\u2009=\u200950) Frequency (Percentage, %)**   **Drug interactions (n\u2009=\u200963) Frequency (Percentage, %)**   **Other problems (n\u2009=\u200997) Frequency (Percentage, %)**                                                                   \n  --------------------------------- ---------------------------------------------------------- ---------------------------------------------------------- ------------------------------------------------------- ----------- ------------ --------------- ----------- ----------- ---------------\n                  \u00a0                                          **Yes**                                                     **No**                                                ***p*-value**                        **Yes**      **No**     *p***-value**    **Yes**     **No**     ***p*-value**\n             **Elderly**                                                                                                                                                                                                                                                           \n                 Yes                                        17 (34.0)                                                  68 (45.3)                                                 0.215^a^                          37 (58.7)   78 (56.9)      0.933^a^      57 (58.8)   58 (56.3)     0.836^a^\n                 No                                         33 (66.0)                                                  82 (54.7)                                                     \u00a0                             26 (41.3)   59 (43.1)          \u00a0         40 (41.2)   45 (43.7)  \n          **Polypharmacy**                                                                                                                                                                                                                                                         \n                 Yes                                        34 (68.0)                                                  119 (79.3)                                                0.149^a^                          58 (92.1)   95 (69.3)     0.001^a^\\*     74 (76.3)   79 (76.7)    \\> 0.999^a^\n                 No                                         16 (32.0)                                                  31 (20.7)                                                     \u00a0                              5 (7.9)    42 (30.7)          \u00a0         23 (23.7)   24 (23.3)  \n    **Duration of hospital stay**                                                                                                                                                                                                                                                  \n              \u2264 1 week                                      37 (74.0)                                                  106 (70.7)                                                0.786^a^                          44 (69.8)   99 (72.3)      0.854^a^      65 (67.0)   78 (75.7)     0.227^a^\n              \\> 1 week                                     13 (26.0)                                                  44 (29.3)                                                     \u00a0                             19 (30.2)   38 (27.7)          \u00a0         32 (33.0)   25 (24.3)  \n   **Microvascular complications**                                                                                                                                                                                                                                                 \n                 Yes                                        31 (62.0)                                                  93 (62.0)                                                \\> 0.999^a^                        41 (65.1)   83 (60.6)      0.652^a^      65 (67.0)   59 (57.3)     0.204^a^\n                 No                                         19 (38.0)                                                  57 (38.0)                                                     \u00a0                             22 (34.9)   54 (39.4)          \u00a0         32 (33.0)   44 (42.7)  \n     **Cardiovascular disease**                                                                                                                                                                                                                                                    \n                 Yes                                        27 (54.0)                                                  94 (62.7)                                                 0.358^a^                          53 (84.1)   68 (49.6)    \\< 0.001^a^\\*   63 (64.9)   58 (56.3)     0.270^a^\n                 No                                         23 (46.0)                                                  56 (37.3)                                                     \u00a0                             10 (15.9)   69 (50.4)          \u00a0         34 (35.1)   45 (43.7)  \n        **Renal impairment**                                                                                                                                                                                                                                                       \n                 Yes                                        22 (44.0)                                                  78 (52.0)                                                 0.414^a^                          33 (52.4)   67 (48.9)      0.761^a^      49 (50.5)   51 (49.5)    \\> 0.999^a^\n                 No                                         28 (56.0)                                                  72 (48.0)                                                     \u00a0                             30 (47.6)   70 (51.1)          \u00a0         48 (49.5)   52 (50.5)  \n        **Liver impairement**                                                                                                                                                                                                                                                      \n                 Yes                                         4 (8.0)                                                    7 (4.7)                                                  0.472^b^                            0 (0)      11 (8.0)     0.018^b^\\*      5 (5.2)     5 (5.8)     \\> 0.999^a^\n                 No                                         46 (92.0)                                                  143 (95.3)                                                    \u00a0                             63 (100)    126 (92.0)         \u00a0         92 (94.8)   97 (94.2)  \n         **Hyperlipidemia**                                                                                                                                                                                                                                                        \n                 Yes                                        17 (34.0)                                                  68 (45.3)                                                     \u00a0                             37 (58.7)   78 (56.9)          \u00a0         57 (58.8)   58 (56.3)     0.836^a^\n                 No                                         33 (66.0)                                                  82 (54.7)                                                 0.215^a^                          26 (41.3)   59 (43.1)      0.933^a^      40 (41.2)   45 (43.7)  \n\n^a^ Computed using Continuity Correction; ^b^ Computed using Fisher's Exact Test; \\* Statistically significant (*p*\u2009\\<\u20090.05).\n\nIn addition, polypharmacy (*p*\u2009=\u20090.001) and cardiovascular events (*p*\u2009\\<\u20090.001) were found to be associated with drug interactions. Patients with polypharmacy or cardiovascular disease were more susceptible to potential drug interactions than those without these two factors. Moreover, a positive association between length of hospital stay and drug choice problem was discovered (*p*\u2009=\u20090.004). Those who stayed not more than one week had a higher likelihood of experiencing a drug choice problem.\n\nDiscussion\n==========\n\nThe mean age of this study population was higher compared with published local data which evaluated the diabetes control of 1670 patients in Malaysia \\[[@B3]\\]. Both local and global data revealed that in developing countries like Malaysia, the prevalence of diabetes mellitus was highest in the age group between 45 and 64 years \\[[@B2]\\] but in this study only 36% of the patients fell into this age group. The higher mean age reported in this study might be because the sample was not representative of the whole population. It also suggests that more older hypertensive diabetic patients were admitted to the ward than were younger patients.\n\nDrug-related problems\n---------------------\n\nIn this study, there was an average of 1.9\u2009\u00b1\u20091.2 DRPs per patient. To date, there has been no comparable study done specifically on DRPs in T2DM patients with hypertension both locally and globally. The number of DRPs identified was only half the number detected by some other studies which were conducted in diabetes mellitus patients \\[[@B10],[@B11],[@B25]\\]. When compared with a recent study with an almost equivalent sample size (193 geriatric clinic patients in Taiwan), which also used the PCNE classification system, the average number of problems identified was 2.2\u2009\u00b1\u20091.6 per patient, slightly higher than in this study \\[[@B21]\\].\n\nAlthough a similar PCNE classification of DRPs was used, the discrepancy with the study by Van Roozendaal *et al.* (2009) \\[[@B11]\\] could be due to the different methods and references used to identify DRPs. For example, the concurrent use of an ACE inhibitor and a sulfonylurea or insulin was considered as a potential DRP in that study and accounted for 46 cases out of the 682 DRPs detected. However, this combination of drugs was not considered as a potential drug interaction in this study because there is a lack of strong evidence of interaction \\[[@B29]\\]. Also, Van Roozendaal *et al.* (2009) \\[[@B11]\\] could not identify any possible contraindications as no information on patients' renal and hepatic functions were successfully retrieved but this is not the case in this study as several contraindications were identified based on patients' medical records and laboratory results.\n\nApart from that, the discrepancy with other studies may be attributed to the differences in study method and setting, different classification systems of DRPs used, and different methods to assess DRPs. Both the studies by Haugbolle & Sorensen (2006) \\[[@B10]\\] and Eichenberger *et al.* (2011) \\[[@B25]\\] conducted home visits and interviews. Also, the Problem Intervention Documentation (PI-Doc) coding system was used in the study by Haugbolle & Sorensen (2006)\\[[@B10]\\] whereas the classification system of DRPs used by Eichenberger *et al.* (2011)\\[[@B25]\\] was unclear, as it was not mentioned in its Methods section. Also, the clinical knowledge of the investigator(s) might also influence the assessment and identification of DRPs.\n\nThis study revealed that 90.5% of the patients had at least one DRP, which was much greater than the 80.7% reported by Haugbolle & Sorensen (2006) \\[[@B10]\\]. However, in a study conducted on ambulatory hemodialysis patients, 97.7% of the patients were found to have at least one DRP \\[[@B32]\\]. This variation across the studies can be attributed to the different study populations and study methods.\n\nAdverse reactions\n-----------------\n\nIn this study, almost one-third of the adverse drug reactions implicated antidiabetic drugs. Similar to the finding by Van Roozendaal & Krass (2008) \\[[@B11]\\], there was a potential risk of hypoglycemia in patients receiving oral antidiabetic drugs or insulin. Antihypertensive agents were also commonly associated with adverse reactions \\[[@B33]\\] and this finding was clearly demonstrated in this study. Calcium channel blockers caused a higher incidence of adverse reactions than diuretics, consistent with a study in an outpatient setting by Goncalves *et al.* (2007) \\[[@B34]\\]. Therefore, all these potential adverse reactions should be taken into consideration, especially in the elderly who might suffer significant deleterious effects. However, since this study was retrospective in nature, only the ones that were important for the hospital admission were noted.\n\nDrug choice problems\n--------------------\n\nA drug choice problem was the second most common DRP in this study and this finding was comparable to other studies \\[[@B10],[@B11]\\]. In this study, most of the contraindications identified were related to the use of metformin. Approximately 24% of the patients who received metformin were found to have contraindications. This was much less compared with the study carried out by Sweileh (2007) \\[[@B35]\\], in which up to 60% of the patients receiving metformin had contraindications to it. This difference is probably due to the variations in defining metformin's contraindications. For instance, the decompensated stage, but no other stage of congestive heart failure, was defined as a contraindication in this study.\n\nApart from metformin, drugs that were categorized as high risk in the modified Beers criteria were frequently prescribed to the elderly, placing them at higher risk of developing drug toxicity \\[[@B4],[@B31]\\]. The high frequency of drug choice problems may highlight a need for the health care providers to pay more attention when prescribing these drugs to older hypertensive diabetic patients.\n\nDosing problems\n---------------\n\nIn this study, excessive dosage was the most frequent dosing problem. Also, there were cases of subtherapeutic dosages and inappropriate durations of treatment. The percentage of dosing problems in this study was higher than that reported by Van Roozendaal and Krass (2009) \\[[@B11]\\] and this was probably due to the lack of assessment of patients' renal and hepatic functions in the latter study.\n\nRanitidine was the most implicated drug for dosing problems. In clinical practice, dosage adjustments of ranitidine are not frequently applied although recommended by manufacturers and this is probably because its potential adverse effects are underestimated \\[[@B36],[@B37]\\]. In a study by Manlucu *et al.* (2005) \\[[@B37]\\], H~2~-receptor antagonists were demonstrated to significantly increase the area under the curve (AUC) and the elimination half-life (*t*~1/2~) of serum drug concentrations when the glomerular filtration rate (GFR) was decreasing. Dosage reduction of drugs in patients with impaired renal function and low GFR may prevent adverse effects and decrease unnecessary drug expenditures \\[[@B36],[@B37]\\]. Therefore, efforts should be made to minimize these dosing errors such as the involvement of a pharmacist in deciding the dosing of drugs or a computerized dosing program \\[[@B37]\\].\n\nDrug use problem\n----------------\n\nThe most frequent drug use problem encountered was \"drugs not taken by patients prior to admission\", which were mostly antihypertensive and antidiabetic drugs. In this study, forgetfulness might be one of the reasons for non-adherence. Also, complicated regimens for the treatment of diabetic complications may contribute to non-adherence. A systematic review confirmed the poor compliance in diabetic patients who were prescribed diabetic medications, whether oral agents or insulin \\[[@B12]\\]. Non-adherence has proven to be associated with negative outcomes such as higher A1c levels and blood lipid levels in diabetes patients \\[[@B12],[@B38]\\].\n\nDrug interactions\n-----------------\n\nIn this study, the drugs that were most implicated in drug interactions were aspirin, clopidogrel, simvastatin and amlodipine. By contrast, beta-blockers, non-steroidal anti-inflammatory agents (NSAIDs) and ACE inhibitors were the drug most involved in drug interactions in a study conducted in Singapore \\[[@B39]\\]. The differences in prescribing patterns and practice in different hospitals may explain this discrepancy. The drug interactions identified in this study were mostly based on established literature and evidence. In clinical practice, several drugs can still be used together, yet close monitoring is fundamental and any toxicity should be identified and immediately followed by corrective actions.\n\nOthers\n------\n\nMany of the patients in this study did not engage in regular physical activity, did not adhere to diabetic diets, did not perform any routine blood glucose monitoring, and defaulted on follow-up or medications, and all these problems were clearly stated by the health care providers in the medical records. These problems would lead to poor glycemic control and accelerate the development or worsening of diabetic complications \\[[@B40]\\].\n\nA local study conducted in an endocrine clinic in a teaching hospital in Kuala Lumpur identified barriers to optimal control of Malay type 2 diabetic patients by interviewing 18 patients and health care providers. It was not surprising to find that most of the patients had a lack of understanding of the disease itself and of its management, which would eventually contribute to non-adherence \\[[@B41]\\]. This is particularly true for the elderly as they tend to have decreased memory, health beliefs not in line with drug therapy, and often neglect the importance of adherence to medications and dietary control. Hence, counseling may be important to increase the awareness and knowledge of this patient population since they frequently encounter these problems.\n\nCauses of DRPs\n--------------\n\nThe results from our study revealed that among all the causes, \"drug or dose selection\" was the most frequently identified cause for DRPs such as drug choice problems, dosing problems or drug interactions. According to the PCNE classification of DRPs, this domain of causes is directly related to the drug or dosage selection while the other domains are concerned with patient-related causes \\[[@B22]\\].\n\nOn the other hand, patients usually had \"drug use problems\" caused by \"drug use process, lack of information, and physiological or patient factor\". Generally, the number of causes identified was lower than the causes identified in other studies such as Chan *et al.* (2011) \\[[@B21]\\]. This is because most of the problems identified were matched with the one most relevant cause rather than several causes, which might be seen in other studies.\n\nFactors found to be associated with DRPs\n----------------------------------------\n\nGenerally, this study did not identify any factors with a statistically significant association with DRP. This was in agreement with a study by Koh *et al.* (2005) \\[[@B39]\\] which found no statistical correlation between DRPs and age or gender. Research by Samoy *et al.* (2006) \\[[@B42]\\] also concluded that there were no risk factors for drug-related hospitalization in a tertiary care hospital in Canada. One of the possible explanations is the nature of the PCNE classification system (6 problem domains with 22 categories) which could possibly affect the results. Also, it remains unclear whether the result was affected by the sample size (Samoy *et al.*, 2006) \\[[@B42]\\].\n\nWhen examining each of the problem domains with several possible factors, statistically significant associations were observed. These associations should receive the attention of the health care providers in order to minimize preventable DRPs.\n\nElderly\n-------\n\nIn our study, the non-elderly were found to be associated with drug choice problems. From the literature reviewed, the findings on the association between age and DRPs are conflicting. In one study on the elderly in an ambulatory setting, age of 80 and above was found to be an independent risk factor for adverse drug events \\[[@B43]\\]. While a study by Chan *et al.* (2011) \\[[@B21]\\] on geriatrics also reported a significant association between age and DRPs, a few studies did not agree with this finding. A study on hospitalized patients from several internal medicine departments found that age was not a risk factor of DRPs \\[[@B15]\\]. Similarly, Koh *et al.* (2005) \\[[@B39]\\] did not report any statistically significant correlation between these two.\n\nPolypharmacy\n------------\n\nIt is a well-known fact that polypharmacy is strongly associated with DRPs and this has been shown by numerous studies \\[[@B12],[@B14],[@B15],[@B43]\\]. It has been reported that a one unit increase in the number of drugs can lead to an increase of 8.6% in the number of DRPs \\[[@B43]\\].\n\nThe results of our study revealed a significant statistical association between polypharmacy and drug interactions, which was consistent with the result from Moura *et al.* (2009) \\[[@B44]\\], a retrospective study on drug interactions in a public hospital in Brazil. The increasing number of drugs used can lead to an increased risk of potential drug interactions \\[[@B23],[@B45]\\]. Since polypharmacy is an inherent problem in T2DM patients with hypertension, the clinically important and significant drug interactions should be routinely checked and monitored \\[[@B9]\\].\n\nRenal impairment\n----------------\n\nRenal impairment was associated with both the drug choice problem (*p*\u2009=\u20090.029) and the dosing problem (*p*\u2009=\u20090.027). Drugs with doses that were higher than required were often prescribed to T2DM patients with hypertension and renal impairment in our study. Also, dosage adjustment was commonly ignored by physicians, suggesting that the severity of inappropriate drug and dosing selection might be underestimated \\[[@B36],[@B37]\\]. DRPs were common among patients with renal impairment due to co-existing medical conditions, as most of them were receiving multiple medications which require dosage adjustment and routine monitoring \\[[@B36]\\].\n\nThe study by Manley *et al.* (2003) \\[[@B32]\\] revealed that in ambulatory hemodialysis patients, the presence of diabetes mellitus is one of the factors associated with DRPs. In other words, diabetes patients on hemodialysis were more likely to experience DRPs. Similarly, Leendertse *el al*. (2008) \\[[@B14]\\] also found that impaired renal function was a risk factor for potentially preventable DRPs. Therefore, in patients with renal impairment, dosage adjustment and close monitoring of renal function are fundamental in order to minimize drug toxicity or subtherapeutic effect \\[[@B46]\\].\n\nCardiovascular events\n---------------------\n\nPatients with cardiovascular events had more potential drug interactions than patients without cardiovascular disease in our study. This can be explained by the wide use of cardiovascular drugs such as antihypertensive drugs, antiplatelet drugs, anticoagulants, and lipid lowering drugs in T2DM patients with hypertension. Many studies concluded that the most common drug category involved in DRPs was cardiovascular agents \\[[@B12],[@B47]\\]. Also, cardiovascular events often add an additional burden to patient conditions and complicate their therapies.\n\nDuration of stay\n----------------\n\nPatients who stayed for not more than one week in the hospital tended to experience the drug choice problem as compared with those who stayed for more than one week. This finding of our study was not in line with that reported by Moura *et al.* (2009) \\[[@B44]\\], which revealed an association between duration of hospital stay and potential drug interactions.\n\nAnother study on hospitalized cancer patients also found a correlation between duration of hospital stay (\u2265 6 days) and potential interactions \\[[@B48]\\]. The relationship between length of hospital stay and DRPs needs to be investigated in future studies since there is a lack of published literature investigating this association.\n\nStudy limitations\n-----------------\n\nBecause of the retrospective nature of our study, the identification and assessment of the DRPs were based on the data available from the medical records with reference to established literature and guidelines. The number of studied patients was less than 50% of potentially eligible patients.\n\nConclusions\n===========\n\nThe most common drugs that were used in T2DM patients with hypertension were amlodipine, insulin, aspirin and simvastatin. Polypharmacy and multiple comorbidities were common in this patient population.\n\nThe most common DRPs were \"others\" (i.e., insufficient awareness of health and diseases), drug choice problems, dosing problems, and drug interactions. The most implicated drugs were aspirin, clopidogrel, simvastatin, amlodipine and metformin.\n\nSeveral factors were found to have statistically significant associations with the six domains of DRPs, including renal impairment, polypharmacy, cardiovascular disease, elderly age, and duration of hospital stay.\n\nSpecial attention should be given to T2DM patients with hypertension and risk factors who are prescribed commonly implicated drugs.\n\nCompeting interest\n==================\n\nThe authors have no conflict of interest to report.\n\nAuthors\\' contributions\n=======================\n\nHZH has made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data and drafting the manuscript or revising it critically for important intellectual content. HFW has been involved in acquisition of data and analysis and interpretation of data. HZH and HFW have given final approval to the version to be published. Both authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1472-6823/13/2/prepub>\n\nAcknowledgments\n===============\n\nThe authors would like to thank the University of Malaya, Malaysia, for financial (RG428/12HTM) and technical support.\n"}
+{"text": "Background {#Sec1}\n==========\n\nTracheostomy is recommended in critical ill patients when prolonged mechanical ventilation is presumed \\[[@CR1]\\]. Optimal timing of tracheostomy is still under debate \\[[@CR1], [@CR2]\\]. However, a recently published meta-analysis suggests that performing tracheostomy within 7\u00a0days after intubation may reduce intensive care unit stay \\[[@CR2]\\]. Most studies comparing early to late tracheostomy defined late tracheostomy as performed in week two after intubation \\[[@CR1], [@CR2]\\]. Venovenous extracorporeal membrane oxygenation (vv-ECMO) is a treatment option for severe adult respiratory failure \\[[@CR3]\\]. The average time on mechanical ventilation of vv-ECMO patients is 23\u00a0days \\[[@CR4]\\]. Therefore most vv-ECMO patients will have an indication for tracheostomy.\n\nWe have to report 2 cases of fatal air embolism into the vv-ECMO in patients undergoing percutaneous dilatational tracheostomy while being on full ECMO support.\n\nCase 1 {#Sec2}\n======\n\nA 65\u00a0year old male without significant comorbidities presented at a non-ECMO hospital with H1N1 pneumonia. As a severe ARDS developed (Horovitz index 58) a vv-ECMO was implanted via the right jugular vein using a 31 Fr AVALON ELITE\u00ae bi-caval catheter and the patient was transferred to our intensive care unit. Six days after ECMO initiation, the patient was still ECMO dependent (blood flow 3.9\u00a0l/min, gas flow of 5.0\u00a0l/min) while on invasive mechanical ventilation (FiO2 45\u00a0%, PEEP 15\u00a0mbar). We therefore presumed a prolonged weaning and aimed for a percutaneous dilatational tracheostomy using the ULTRAperc system (Portex\u00ae, Smith medical, England) with bronchoscopic guidance. After puncture of the trachea and immediately after the first dilation step a significant air embolism into the ECMO system was observed. The ECMO system was halted but air could not be removed. Invasive mechanical ventilation failed to achieve sufficient oxygenation and hypoxic cardiac arrest occurred. After 20\u00a0min of cardiopulmonary resuscitation the patient could be stabilized after implantation of a new ECMO system. Even though primary resuscitation was successful the patient never recovered from the event and died on ECMO at day 27.\n\nCase 2 {#Sec3}\n======\n\nA 57-year old male was admitted to hospital for acute interstitial lung disease. He developed progressive combined hypoxic (Horovitz index 54) and hypercapnic respiratory failure. Venovenous ECMO therapy was established using a 27 Fr bi-caval catheter (AVALON ELITE\u00ae, Maquet, Germany). A combined immunosuppressive therapy with cyclophosphamide and prednisolone was initiated. After 10\u00a0days of mechanical ventilation (fiO2 45\u00a0%, PEEP 9) and 4\u00a0days on vv-ECMO (blood flow 4.5\u00a0l/min, gas flow 7.0\u00a0l/min) we performed dilatational tracheostomy using the 8.0 Fr Ciaglia Blue Rhino\u00ae percutaneous introducer set (Cook medicals, USA). Stab incision and sequential dilation caused minor venous bleeding. The dilator was reintroduced and external compression applied. Within seconds air embolism into the ECMO system occurred. Rapidly, the patient developed hypoxemia leading to cardiac arrest. Despite immediate resuscitation and implantation of a new ECMO system via the femoral veins return of spontaneous circulation could not be achieved.\n\nDiscussion {#Sec4}\n==========\n\nPercutaneous dilatational tracheostomy is a frequent intervention in critical ill patients on the intensive care unit. In large case series, mortality rates range from 0.17\u00a0% \\[[@CR5]\\] to 0.60\u00a0% \\[[@CR6]\\]. Most of these fatal complications however were not directly associated with tracheostomy. Case series of fatal complications directly related to percutaneous dilatational tracheostomy \\[[@CR7]\\] report damage of large arteries, uncontrollable bleeding or airway loss. A significant air embolization after surgical tracheostomy was reported earlier \\[[@CR8]\\] but did not reach significant levels in a meta-analysis including over 113.000 procedures \\[[@CR6]\\].\n\nIn a recent review, 30 out of 168 patients underwent tracheotomy while on ECMO therapy \\[[@CR4]\\] with a significant variation between different centers (ranging from 4 to 46\u00a0% of all ECMO patients). No data is available concerning ECMO cannulation (bi-femoral, femo-jugular or bi-jugular) \\[[@CR4]\\]. A case of a fatal complication directly related to tracheostomy in ECMO patients has not been reported so far. Bleeding, the most frequent complication of tracheostomy \\[[@CR4], [@CR6], [@CR7]\\], however might be even pronounced in ECMO patients due to activation and consumption of coagualiton enzymes, platelet depletion and anticoagulant medication \\[[@CR9]\\].\n\nIn the 2 patients presented here, the air embolization into the ECMO was detected seconds after dilation with a percutaneous tracheostomy dilator. Both patients were on full ECMO support via an AVALON ELITE\u00ae (Maquet, Germany) bi-caval dual lumen cannula draining blood directly from the superior vena cava (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). Especially with higher ECMO blood flow, this might result in significant negative pressure in the draining vessels. We therefore hypothesize that suction of air through the inferior thyroidal vein draining blood directly to the vena cava superior might be responsible for fatal air embolism in both patients.Fig. 1Air embolization into the ECMO system during tracheostomy. The AVALON ELITE bi-caval cannula drains venous blood from the superior as well as the inferior vena cava. Since the inferior thyroid vein directly drains into the vena cava superior, a venous damage during tracheostomy might result in severe air embolization into the ECMO\n\nDuring the time period between 2013 to 2015, we performed a total of 30 percutaneous dilatational tracheostomy in patients on ECMO therapy, of which 24 had an Avalon elite cannula. As a routine measure, an ultrasound study is performed in all our patients in order to rule out large vessels at the puncture site before proceeding to dilatational tracheostomy. Since bleeding incidence is comparable in surgical versus dilatational tracheostomy \\[[@CR1], [@CR10]\\] it appears unlikely that surgical tracheostomy could have prevented these complications.\n\nIn response to the incidents described above we implemented the following points before performing a tracheostomy in ECMO patients on AVALON ELITE\u00ae cannulas.In case of severe hypoxia without ECMO (using 100\u00a0% FiO~2~ at the respirator) delay tracheostomy.During tracheostomy reduce ECMO blood flow (and thus negative suction pressure) as far as possible.Perform tracheostomy in a head down position.Cover puncture site at all times with wet compresses.\n\nConclusion {#Sec5}\n==========\n\nThere is limited data on safety of percutaneous dilatational tracheostomy in ECMO patients. Especially in patients on full ECMO support using the AVALON ELITE\u00ae bi-caval dual lumen catheter dilatational tracheostomy might be associated with a significant risk.\n\nAbbreviations {#Sec6}\n=============\n\nECMO, extracorporeal membrane oxygenation\n\nAuthors' contributions {#FPar1}\n======================\n\nAL: data analyzation, manuscript drafting, primary patient care, artwork TW: data analyzation, manuscript drafting CB: data analyzation, manuscript drafting CB data analyzation, manuscript drafting DL: conception and design, data analyzation, manuscript drafting, literature research. All authors read and approved the final manuscript.\n\nCompeting interests {#FPar2}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Commentary\n==========\n\nThe domestic dog offers unique advantages to the study of complex and multilocus diseases. Breeds are closed populations; membership requires that all parents and grandparents be registered members of the same breed. The division of the population into over 350 breeds simplifies the locus heterogeneity that is typically associated with complex diseases \\[[@B1],[@B2]\\]. Finally, many breeds share recent common ancestors \\[[@B3]-[@B5]\\], meaning that they also likely share common disease alleles, offering an avenue to increase power and resolution for genetic studies.\n\nThe genome of the dog was sequenced to 7.5x in 2005 \\[[@B6]\\]. Among the most interesting features are the extensive within-breed linkage disequilibrium (LD) \\[[@B6],[@B7]\\] and the high degree of across-breed heterogeneity. As a result, genome wide association studies (GWAS) require chips with no more the 100,000 single nucleotide polymorphisms (SNPs) (see \\[[@B8]\\] for review). Finally, the relatedness of breeds means that transitioning from marker to mutation can often be accomplished by combining data from affected individuals from related breeds \\[[@B9],[@B10]\\].\n\nNevertheless, the dog system has some disadvantages. For instance, while dog families may be very large, thus allowing disease genes to be found by linkage analysis of a single large family \\[[@B11]\\], locating and sampling all the necessary dogs can be problematic. Second, while GWAS have been successful at identifying many loci of interest \\[[@B12]\\], the extensive LD means that getting from associated locus to mutation can be difficult \\[[@B8],[@B13]\\].\n\nThe recent success of Forman et al, \\[[@B14]\\] who used genome-wide mRNA sequencing (RNA-Seq) to find the variant associated with a form of neonatal cerebellar cortical degeneration, circumvents many of these problems. By way of background, the disorder is a neurodegenerative disease occurring in several breeds, including the beagle. Affected dogs suffer from a loss of balance, uncoordinated gate, and an inability to regulate movement. Loss of Purkinje cells with swollen dendritic processes is the pathologic hallmark of this recessive disorder \\[[@B15]\\].\n\nThe authors performed genome wide mRNA sequencing using cerebellum tissue from one affected pup. Importantly, they focused their resulting search exclusively on the 27 dog genes that were known orthologs of human ataxia genes. In this sense they got lucky---there are 41 such human loci and causal genes identified for only 28. Of those, dog orthologs are known for 27. Had the causative gene not been previously identified as an ataxia gene in human studies, the cause of the canine disease would not have been found using this approach. After comparison to common dog SNPs and orthologous sequences, variants were eliminated if they were non-coding, heterozygous or conserved. The remaining variant, located in the \u03b2-III spectrin gene (*SPTB2),* was an eight base pair (bp) coding deletion that is predicted to cause both an aberrant run of 27 extra amino acids and premature termination of mRNA. The mutation segregated perfectly as an autosomal recessive in the small family tested, was found in the heterozygous state in other unaffected but at-risk dogs, and was absent in 37 other breeds. As expected, cerebellum tissue from the proband, showed a near total loss of both \u03b2-III spectrin mRNA and protein when compared to an unspecified control.\n\n\u03b2-III spectrin is a superb candidate gene. The gene family encodes cytoskeletal proteins that are important structural components of the plasma membrane. \u03b2-III spectrin is found in the nervous system, with the highest levels of expression in the Purkinje cell soma and dendrites \\[[@B16]\\]. There is, thus, no doubt the correct gene has been found.\n\nWhile the authors are to be commended for synthesizing a wide range of knowledge and tools to construct a strong argument for their findings, the generalizability of the approach has to be questioned. The authors point out, correctly, that because mRNA-seq requires far fewer samples than does a GWAS, less time is needed to collect samples and complete projects. In this case, however, the authors were well-informed regarding the correct tissue to sequence and candidate genes to consider. That will not always be the case. Also, this approach is likely to miss weakly expressed transcripts, which could be important for some diseases. While not simple, a library normalization step could certainly be incorporated. A major argument for sequencing and pursuing genetic studies in the dog is the simplified genetic architecture: reduced locus heterogeneity improves the chances for identifying variants underlying complex disease traits and a shared genetic background makes disease presentation more uniform. Focusing solely on known human disease genes dismisses both of these advantages.\n\nDoes this lessen the impact of the paper? In light of the rapid advances in human genetics, particular with regard to rare genetic diseases, we would argue no. There are many cases where families of human genes are known, but their precise matchup with clinical features remains ambiguous. This practice of using candidate gene/mRNA sequencing will resolve many such cases. The argument for greater clarity in phenotype because of the similar genetic background of breed members can still apply, albeit in a different way. Finally, animal models that result from studies such of that of Forman \\[[@B14]\\] are clinically valuable.\n\nThe number of human genes associate with both common and rare diseases is increasing at an amazing rate. These authors offer a way to take advantage of such advances while making use of the extraordinary advantages of the canine system, resulting in gene discovery that is applicable to both human and companion animal health.\n\nConclusions\n===========\n\nApplying mRNAseq methods from carefully selected tissues to pre-selected candidate genes, while not a substitute for the GWAS or linkage studies, can reveal causative mutations for multilocus diseases.\n\nAbbreviations\n=============\n\nbp, Base pair; GWAS, Genome wide associations study; LD, Linkage disequilibrium; RNA-seq, mRNA sequencing; SNP, Single nucleotide polymorphism.\n\nAuthors' contributions\n======================\n\nEAO wrote the first draft of the manuscript. HB provided extensive discussion and editing. Both authors contributed to and approved the final version.\n\nCompeting interests\n===================\n\nThe authors declare no completing interests.\n\nAcknowledgements\n================\n\nThe authors gratefully acknowledge the Intramural Program of the National Human Genome Research Institute.\n"}
+{"text": "1. Introduction {#sec0005}\n===============\n\nAmong the common ambient air pollutants, epidemiological studies have identified that exposures to particulate matter (PM) correlate most consistently with pulmonary and cardiovascular morbidity and mortality [@bib0050], [@bib0070], [@bib0110], [@bib0195]. While ambient PM range from 0.001 to 100 \u03bcm in aerodynamic diameter (AD) [@bib0035], particles larger than 10 \u03bcm AD are generally trapped in the nasal passages [@bib0095]. Based on size, PM smaller than 10 \u03bcm AD are divided into three fractions: PM~10~ (AD \\< 10 \u03bcm), PM~2.5~ (AD \\< 2.5 \u03bcm) and ultrafine (UF) (AD \\< 0.1 \u03bcm) [@bib0255]. PM~10~ is further divided into coarse (AD 2.5--10 \u03bcm) and fine (AD \\< 2.5 \u03bcm) PM [@bib0255]. The fate of inhaled particles depends on their size, which determines their site of deposition in the lung, which in turn influences their rate of clearance. While larger particles are primarily deposited in the extrathoracic and thoracic regions, smaller particles (\\<1 \u03bcm) deposit deep in the lung, particularly in the alveolar region [@bib0095]. Of further concern, ultrafine particles also have the potential to cross the lung alveolar-capillary border and gain access to the circulation [@bib0030], [@bib0165], which can have serious implications in terms of systemic toxicity and cardiovascular disease. Several studies have reported translocation of ultrafine particles to secondary target organs, including the liver, spleen, heart, and brain in rats [@bib0185], [@bib0230]. However, to date, translocation of particles has only been observed in spleens and livers of coal mine workers after long-term exposure to very high doses [@bib0125]. While exposure standards for PM~10~ and PM~2.5~ have already been established in North America, an ambient standard for ultrafine particles has yet to be established.\n\nRecent studies have suggested that the presence of excessive reactive oxygen species (ROS), i.e., hydrogen peroxide (H~2~O~2~), superoxide (O~2~\u2022^\u2212^), and hydroxyl radical (OH^\u2212^) in PM, may lead to oxidative stress in the pulmonary system [@bib0130]. ROS can be generated from soluble transition metals (Cu, Cr, Fe, Zn, etc.) or organic compounds (OCs), i.e., polycyclic aromatic hydrocarbons (PAHs), that are present on the surface of particles [@bib0235], [@bib0260]. Li et al. [@bib0130], [@bib0135] proposed that oxidative stress may initiate a specific sequence of cellular responses. At lower levels of oxidative stress, antioxidant enzymes are activated in order to protect the lung. If this antioxidant response of the cell fails to provide protection against the generation of ROS, an inflammatory response may be induced to attract inflammatory cells to the site of \"injury\". Finally, at toxic levels, cell death occurs through both apoptosis and necrosis [@bib0130], [@bib0135], [@bib0145], [@bib0280]. Thus, the progressive increase in the severity of the response to PM suggests a hierarchical dose-effect relationship, whereby the activation of antioxidant mechanisms precedes inflammatory responses and in turn ultimately cell death. Long-term exposure to PM resulting in chronic pulmonary inflammation may also impair lung development, and increase the risk of developing pulmonary diseases, including asthma and chronic obstructive pulmonary diseases [@bib0070], [@bib0075], [@bib0220].\n\nAn intensive, collaborative, multidisciplinary field campaign 'Health Effects of Aerosols in Toronto (HEAT)' was conducted in Toronto in 2010. The main objectives of this campaign were to: (1) characterize the physicochemical properties of Toronto\\'s size-fractionated concentrated ambient PM [@bib0200]; (2) measure different cellular responses (i.e., antioxidant defense, proinflammatory changes, and cytotoxicity) of airway epithelial cells to PM; (3) identify relationships between the physicochemical characteristics of the PM and the cellular responses; and (4) assess the differential effects of different PM size fractions on both pulmonary and cardiovascular functions [@bib0015]. We hypothesized that particle exposure would induce dose-dependent differential toxicological effects (i.e., antioxidant defense, inflammatory changes, and ultimately cell death) on the cells and that these cellular responses would vary depending on the diverse physicochemical properties of the different size-fractionated PM.\n\n2. Materials and methods {#sec0010}\n========================\n\n2.1. Collection of ambient particles {#sec0015}\n------------------------------------\n\nConcentrated ambient PM was collected using the high volume ambient particulate concentrator system located at the Gage Occupational and Environmental Health Unit, University of Toronto, which is located on a busy downtown street. The concentrator facility consists of three Concentrated Ambient Particle Systems (CAPS) that fractionate and concentrate ambient PM in the coarse (2.5 \\< AD \\< 10 \u03bcm), fine (1.5 \\< AD \\< 2.5 \u03bcm), or quasi-ultrafine (AD \\< 0.2 \u03bcm) size ranges. On average, coarse, fine, and quasi-ultrafine particles were concentrated (by mass) by factors of \u223c103, 36, and 28, respectively. We have previously reported a detailed description of the effectiveness of these three systems [@bib0155], [@bib0200].\n\nSize-fractionated PM was collected on filters on weekdays during the winter of 2010 (February 19--March 19). Samples were simultaneously collected over 4--8 h on 47 mm quartz (Whatman, USA), 47 mm Teflon (Pall Gelman, USA), and 37 mm Teflon filters (Pall Gelman, USA) for chemical characterization, determination of redox activity, and *in vitro* analyses, respectively. The number of filters used for this study was 7 for coarse, 6 for fine, and 8 for quasi-ultrafine PM. All filters were weighed gravimetrically before and after sampling to measure the total collected PM mass. To prevent photo-degradation and evaporation loss, the filter samples were then sealed and kept in the dark at -20 \u00b0C until further analyses. Information regarding the PM collection times and collected mass is described in Supplemental Table S.1. The filters were extracted using cell culture medium, as described previously [@bib0010]. The extraction efficiency varied from 65% to 90% for the three different particle sizes. Extraction efficiency was the highest for fine particles with the highest water-soluble fraction.\n\nDuring the study period, the number concentrations of the size-distributed ultrafine particles (diameter: 10--800 nm) were also measured online using a Scanning Mobility Particle Sizer (SMPS) (Model 3080, TSI) (Supplemental Fig. S.1.A). The particle surface area was calculated from the number concentration data, assuming the particles were spherical (Supplemental Fig. S.1.B). Particle-bound polycyclic aromatic hydrocarbons (p-PAHs) concentrations were determined using a Photoelectric Aerosol Sensor (PAS, Model PAS2000CE, EcoChem Analytics) (Supplemental Fig. S.2).\n\n2.2. Chemicals and assay kits {#sec0020}\n-----------------------------\n\nDulbecco\\'s modified Eagle\\'s medium (DMEM), penicillin--streptomycin, [dl]{.smallcaps}-dithiothreitol (DTT, 98%) and 1, 2-naphthoquinone were obtained from Sigma--Aldrich (Oakville, ON, Canada). [l]{.smallcaps}-Glutamine, 0.05% trypsin, and superscript II reverse transcriptase kits were obtained from Invitrogen (Carlsbad, CA, USA). Fetal bovine serum (FBS) and phosphate buffer saline (PBS) were obtained from Wisent Inc. (St-Bruno, QC, Canada). Tris--HCl, tetrasodium EDTA and trichloroacetic acid were obtained from Fisher Scientific (Whitby, ON, Canada), and 5, 5\u2032-dithio-bis (2-nitrobenzoic acid) was obtained from Alfa Aesar (Ward Hill, MA, USA). The MTT assay was performed using a kit from the American Type Culture Collection (ATCC) (Manassas, VA, USA). RNA extraction kits and TaqMan master mix were obtained from Qiagen^\u00ae^ (Germantown, MD, USA) and Applied Biosystems^\u00ae^ (Life Technologies Inc., Burlington, ON), respectively. Plasmids and primers for HMOX-1 and IL-8 were obtained from OriGene (Rockville, MD, USA) and Integrated DNA Technologies Inc. (Coralville, IA, USA), respectively.\n\n2.3. Cell culture and particle exposure {#sec0025}\n---------------------------------------\n\nA549 cells (American Type Culture Collection, Rockville, MD, USA), a cell line of human adenocarcinomatous cells derived from lung cancer that exhibits characteristics of alveolar epithelial cells, were used in this study [@bib0080]. Cell culture and *in vitro* PM exposures were performed, as described previously [@bib0010] with several modifications. Cells were seeded at 2.5 \u00d7 10^5^ cells/ml in 96-well plates for the MTT assay and at 1 \u00d7 10^6^ cells/ml in 6-well plates for real-time polymerase chain reaction (RT-PCR). PM suspensions were prepared in cell culture medium containing 10% FBS and were added to the cells at final concentrations of 10, 50, and 100 \u03bcg/ml (mass/volume) for 4 h along with appropriate media-treated controls. The exposure duration was selected based on the expression of IL-8 (protein) in our previous study [@bib0010]. Cells were exposed to PM isolated from each filter in triplicate at each PM concentration.\n\n2.4. PM chemical characterization {#sec0030}\n---------------------------------\n\nFor chemical speciation, the quartz filters were cut into three pieces: two quarters and one half portions. One quarter portion was used for the analyses of acid-soluble elements. From the half portion, 1.7 cm diameter punches were taken for elemental carbon (EC) and organic carbon (OC) analyses using a Sunset EC-OC analyzer, according to the National Institute of Occupational Health and Safety (NIOSH) 5040 protocol [@bib0025]. The mass of the OC fractions was converted to the organic mass (OM) by multiplying by a factor of 1.4 to account for the mass of oxygen in addition to the OC mass. The fractions of OM measured at various temperatures (310, 475, 615, and 870 \u00b0C) were named as OM-1, OM-2, OM-3, and OM-4, respectively, representing their volatility order. The remaining filter portion was extracted with deionized water (Millipore DQ-3, resistivity \\> 18.3 M\u03a9) for analysis of water-soluble cations, including amines, anions, metals, and organic carbon (WSOC). Acid and water-soluble extracts were analyzed for 26 elements (Ag, Al, As, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, Pb, Sb, Se, Si, Ti, Tl, V, and Zn) using Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES, Perkin Elmer Optima 3700 DV) in the axial mode, at the Analytical Laboratory for Environmental Science Research and Training (ANALEST) facility of University of Toronto. Anions and cations in the size-fractionated PM sample water extracts were measured by Ion Chromatography (Dionex ICS-2000). WSOC was measured using a Total Carbon Analyzer system (Shimadzu TOC-V~CPH/CPN~) equipped with an 8-port sampler. The DTT assay was used to measure the redox activity of the PM samples, as previously described [@bib0010]. Redox activity was expressed as the rate of DTT consumption per minute normalized to the quantity of PM used.\n\n2.5. Analyses of biological endpoints {#sec0035}\n-------------------------------------\n\nIn this study, heme oxygenase (HMOX-1) was selected as the biomarker of induction of antioxidant enzymes. HMOX-1 is known to be highly responsive to oxidative stress and plays a protective role by producing the antioxidant bilirubin through the degradation of heme. To evaluate the proinflammatory response to PM, IL-8 expression was measured at both the mRNA and protein levels. The results for IL-8 (protein) are presented in Supplemental Materials (S.1). IL-8 acts as a chemoattractant factor and is a ubiquitous early response marker of inflammation in many cell types. Following 4 h-exposure of cells to PM, cell lysates and supernatants were collected from each well. Total RNA was extracted from cell lysates using a RNeasy mini kit according to the manufacturer\\'s instructions (Cat. No. 74104, Qiagen) and was quantified using a Nanodrop-1000 system (Thermo Scientific). First strand cDNA was synthesized from 1 \u03bcg of mRNA for each sample using SuperScript^\u00ae^ II Reverse Transcriptase (Invitrogen\u2122). Expression of HMOX-1 and IL-8 mRNA was quantified using an Applied Biosystems 7900HT Fast Real-Time PCR System. The copy number of the target gene was normalized to 18S RNA, as a housekeeping gene. The primers and probe sets for HMOX-1, IL-8, and 18S are reported in [Table 1](#tbl0005){ref-type=\"table\"}. MTT assays were performed as described previously [@bib0010]. All assays were conducted in triplicate.Table 1Primers and probes used for real-time RT-PCR.Gene targetForward primerReverse primerProbeHMOX-15\u2032-TCAGGCAGAGGGTGATAGAAG-3\u20325\u2032-TTGGTGTCATGGGTCAGC-3\u20325\u2032-TGGATGTTGAGCAGGAACGCAGT-3\u2032IL-85\u2032-ATACTCCAAACCTTTCCACCC-3\u20325\u2032-TCTGCACCCAGTTTTCCTTG-3\u20325\u2032-CCACACTGCGCCAACACAGAAA-3\u203218S5\u2032-GGACATCTAAGGGCATCACAG-3\u20325\u2032-GAGACTCTGGCATGCTAACTAG-3\u20325\u2032-TGCTCAATCTCGGGTGGCTGAA-3\u2032\n\n2.6. Statistical analyses {#sec0040}\n-------------------------\n\nStatistical analyses were performed using GraphPad Prism 4.0c (La Jolla, CA, USA). All biological response results are reported as the mean \u00b1 standard error of the mean (SEM) and relative to controls. Non-parametric statistical analyses were used because the data were not normally distributed. Kruskal--Wallis one-way analysis of variance with post hoc Dunn\\'s multiple comparison test were used to compare the redox-activities of different sized PM and to determine the difference in biological responses between controls and different PM concentrations. Two-way ANOVA was used to investigate the effects of size and concentration on the biological endpoints. Biological response data for all three size fractions were also plotted on the same graph with elemental concentration to test for the effect of size on the responses. Spearman correlations (*\u03c1*) were used to assess the strength of relationships between composition and biological responses. Correlation tests were also used to examine the relationship between quasi-ultrafine particle number concentration and surface area (4 h averaged data) and biological responses. Statistical differences were considered significant at *p* \\< 0.05.\n\n3. Results {#sec0045}\n==========\n\n3.1. PM characterization {#sec0050}\n------------------------\n\nEach of the PM size fractions exhibited distinct chemical compositions ([Fig. 1](#fig0005){ref-type=\"fig\"}). The metal concentrations (including water-soluble metals) were the highest in the coarse fraction (22%), whereas quasi-ultrafine PM was dominated by OM (53%). On average, water-soluble components (ions, amines, and OC) comprised 59% of the fine PM, compared with 41% and 28% in the ultrafine and coarse PM, respectively. Among the water-soluble compounds, a significant contribution of anions (i.e., NO~3~^\u2212^, SO~4~^2+^, C~2~O~4~^2\u2212^, PO~4~^3\u2212^, Cl^\u2212^) was observed for both fine and coarse PM (Supplemental Fig. S.3). While ultrafine PM was comprised of the highest proportion of amines (38%) and OC (24%), coarse PM contained the highest proportion of water-soluble metals (17%). In ultrafine PM, trimethyl amine (TMA) contributed about 41% of the total amine concentration. Detailed information regarding the chemical compositions are provided in Supplemental Tables S.2--S.4.Fig. 1Average chemical composition of concentrated ambient coarse, fine, and quasi-ultrafine PM collected from Toronto (EC = elemental carbon, OM = organic mass, WS = water-soluble elements (anions, cations, amines, organic carbon (OC))).\n\n3.2. Redox activity measurement {#sec0055}\n-------------------------------\n\nThe ROS generation potential of the collected coarse, fine, and ultrafine PM samples was analyzed using the DTT assay. This assay demonstrated that each of the PM size fractions exhibited measurable redox activity ([Fig. 2](#fig0010){ref-type=\"fig\"}). The activity was higher for fine and quasi-ultrafine particles compared with coarse PM (*p* \\< 0.05).Fig. 2Redox activity of coarse, fine, and quasi-ultrafine PM, as measured by the DTT assay. Data are expressed as the mean \u00b1 SEM (*n* = 7, 6, and 8 for coarse, fine, and quasi-ultrafine, respectively). Kruskal--Wallis ANOVA followed by Dunn\\'s multiple comparison test showed that redox activity for fine and quasi-ultrafine PM were significantly higher than coarse PM (\\* *p* \\< 0.05, \\*\\* *p* \\< 0.01).\n\n3.3. Effect of different size-fractionated PM on biological endpoints {#sec0060}\n---------------------------------------------------------------------\n\n### 3.3.1. Activation of antioxidant defense {#sec0065}\n\nHMOX-1 mRNA expression increased in a dose-dependent manner in response to all three size fractions ([Fig. 3](#fig0015){ref-type=\"fig\"}A). Significant induction of HMOX-1 expression was observed as compared with controls at 10 \u03bcg/ml for ultrafine PM and at 50 \u03bcg/ml for the other size fractions. At exposure levels of 100 \u03bcg/ml, the order of response was coarse \u2265 ultrafine \\> fine. HMOX-1 expression was significantly lower for fine PM compared with both coarse (*p* \\< 0.05 by two-way ANOVA) and ultrafine PM (*p* \\< 0.001), but was similar between the coarse and ultrafine PM.Fig. 3Effect of different size-fractionated ambient PM on different biological endpoints. A549 cells were exposed to coarse, fine, and ultrafine PM at mass concentrations of 10, 50, and 100 \u03bcg/ml (mass/volume) for 4 h. (A) HMOX-1 (mRNA), (B) IL-8 (mRNA), and (C) cytotoxic responses. HMOX-1 and IL-8 expressions are presented as fold change and cytotoxic responses are presented as the percentage of cell viability relative to controls. The data are presented as mean \u00b1 SEM (*n* = 21, 18, and 24 for coarse, fine, and ultrafine, respectively, for HMOX-1 and IL-8 expression, *n* = 7 for coarse, 6 for fine, and 8 for ultrafine for cytotoxicity). Statistically significant differences from control and in between concentrations \\# (*p* \\< 0.05) and between sizes of PM \\*, \\*\\*, \\*\\*\\* *p* \\< 0.05, 0.01, 0.001, respectively).\n\n### 3.3.2. Activation of the proinflammatory response {#sec0070}\n\nSignificant induction of IL-8 (mRNA) expression was observed as compared to controls at 50 \u03bcg/ml for ultrafine PM and at 100 \u03bcg/ml for the other size fractions ([Fig. 3](#fig0015){ref-type=\"fig\"}B). At exposure levels of 100 \u03bcg/ml, the order of response was ultrafine = fine \\> coarse. IL-8 expression was significantly lower for coarse PM compared with both fine (*p* \\< 0.01) and ultrafine PM (*p* \\< 0.001), but was similar between the fine and ultrafine PM.\n\n### 3.3.3. Cell viability {#sec0075}\n\nThe cytotoxicity profiles for coarse, fine, and ultrafine PM were determined using the MTT assay ([Fig. 3](#fig0015){ref-type=\"fig\"}C). Similar to the antioxidant and proinflammatory responses, a PM concentration-dependent decrease in cell viability was observed. For coarse and fine PM, a statistically significant decrease was observed starting at PM mass concentrations of 50 \u03bcg/ml, whereas significant cytotoxicity for ultrafine PM was observed at 100 \u03bcg/ml. At 100 \u03bcg/ml, cell viability decreased to 70%, 84%, and 87% for coarse, fine, and ultrafine particles, respectively. Two-way ANOVA analyses showed that the coarse PM were more cytotoxic than fine and ultrafine PM (*p* \\< 0.001).\n\n3.4. Linear regression and correlation analyses {#sec0080}\n-----------------------------------------------\n\nScatter plots of selected chemical constituents with biological endpoints were used to examine whether the differences in the responses were due to the mass concentrations of different elements vs. particle size ([Fig. 4](#fig0020){ref-type=\"fig\"}, [Fig. 5](#fig0025){ref-type=\"fig\"}, [Fig. 6](#fig0030){ref-type=\"fig\"}). If the responses were governed by mass concentration alone, all data for the three PM size fractions would be expected to be fit along the same regression line. Three distinct and independent linear regression lines were observed for most of the elements, indicating that for the same elemental mass concentration, the responses were different between the different PM sizes. The slopes of the biological response vs. total PM mass concentration and the different PM constituent masses were also determined for the different size fractions (Supplemental Tables S.5). Among the 17 elements, the slopes of the dose-response relationships were significantly different among the sizes for 4 (HMOX-1), 15 (IL-8 mRNA), and 12 (cell viability) elements. Greater slope values were generally observed for metals in ultrafine PM and for OM for coarse PM. Although metal and OM concentrations were lower in ultrafine and coarse PM, respectively, the high slope values indicated that small increments of these PM constituents could have greater influence on the biological responses. Slope values were higher for Cr, Ti, and OM for all size fractions and biological endpoints than for total PM mass, suggesting that these two metals and OM impose comparatively greater effects compared to other chemical elements.Fig. 4Scatter plots of selected chemical elements vs. HMOX-1 upregulation for coarse, fine, and quasi-ultrafine PM.Fig. 5Scatter plots of selected chemical elements with IL-8 (mRNA) upregulation for coarse, fine, and quasi-ultrafine PM.Fig. 6Scatter plots of selected chemical elements with cell viability for coarse, fine, and quasi-ultrafine PM.\n\nThe Spearman correlation coefficients (*\u03c1*) of the biological endpoints with different chemical constituents, redox activity, and between the endpoints are reported in Supplemental Tables S.6--S.9. All the correlation values mentioned hereafter were statistically significant (*p* \\< 0.05). In the coarse size range, all three biological endpoints (HMOX-1 and IL-8 (mRNA) positively, and cell viability negatively) correlated with total Al, Ba, Ca, Cr, Cu, Fe, Mg, Mn, Na, Si, Ti, Zn, and OC fractions. For fine PM, Ba, Cr, Cu, Fe, Mg, Mn, Ti, Zn, and OC fractions correlated positively with HMOX-1 and IL-8 (mRNA), and inversely with cell viability. For ultrafine PM, total Mn, Na, Si, Ti, Zn, and OC fractions correlated positively with the upregulation of HMOX-1 and IL-8 (mRNA), and inversely with cell viability. Redox activities correlated with all biological responses (inversely with cell viability) for all size fractions. For ultrafine particles, all the biological responses were also correlated with trimethyl amine (TMA) and dimethyl amine (DME) (Supplemental Table S.7). No significant correlations were found between number concentration, surface area, and biological responses for ultrafine PM; except for cell viability, which correlated positively with particle number concentration (Supplemental Table S.8). Among the biological responses, IL-8 expression consistently correlated inversely with cell viability in all size fractions (Supplemental Table S.9).\n\n4. Discussion {#sec0085}\n=============\n\nThis *in vitro* study investigated the antioxidant, proinflammatory, and cytotoxic effects of coarse, fine, and quasi-ultrafine PM collected in Toronto, Ontario, Canada. Size-dependant cellular responses were observed after exposure to PM. All the size fractions exhibited differential biological effects with respect to oxidative stress and inflammatory responses, and cell death. Coarse PM exhibited higher cytotoxic responses than fine and ultrafine PM, whereas antioxidant and inflammatory responses for quasi-ultrafine PM were as high as coarse and fine PM, respectively. Based on the correlations of the chemical elements with biological endpoints, and the inter-correlations among the elements, it appeared that the observed responses for all size fractions were associated with traffic-related sources (exhaust and non-exhaust). The sampling site was located next to a busy road with \u223c20,000 automobiles/weekday. At this site, 25--51% of the PM~2.5~ mass comes from local sources and the main local source is vehicular traffic [@bib0105].\n\n4.1. Redox activity and biological responses {#sec0090}\n--------------------------------------------\n\nSince urban ultrafine particles are associated with traffic emissions, they are generally high in organic matter and contain higher mass loadings of PAHs than coarse or fine particles [@bib0045], [@bib0065]. Quinone species are highly active redox-cycling catalysts [@bib0235] and are present in gasoline and diesel engine exhaust particles [@bib0100]. Quinones may be generated during the oxidation of PAHs associated with exhaust particles [@bib0215], [@bib0270]. The DTT assay used in this study is particularly reactive to organic species, specifically to quinones. This assay measures quinone-catalyzed production of O~2~\u2022^\u2212^ by the transfer of electrons from DTT to oxygen [@bib0115]. Semi-continuous measurement of p-PAHs during the HEAT campaign showed that ultrafine PM had higher concentrations of p-PAHs than coarse and fine PM, in both ambient and concentrated PM samples (Supplemental Fig. S.2). Therefore, it was expected that ultrafine particles enriched with PAHs would exhibit greater redox activity than fine and coarse particles [@bib0045], [@bib0170], [@bib0210]. However, in this study fine PM was found to be more redox active than ultrafine and coarse PM. HMOX-1 is a sensitive biomarker of oxidative stress and its expression would thus be expected to be related to the redox potential of the particles. The increased expression of HMOX-1 following exposure to coarse and ultrafine particles compared with fine PM also suggested that the DTT assay did not comprehensively measure the redox potential of the coarse and ultrafine particles. The lack of an association between HMOX-1 expression and the redox activity, as measured by the DTT assay, could be due to two reasons. Firstly, there is evidence suggesting that aqueous chemistry within the CAPS can promote adsorption of condensable vapors, which likely adds a large amount of OM to the concentrated ultrafine particles [@bib0155]. For example, the abundance of amine compounds in these samples was much higher than that typical of urban ultrafine PM; concentrations of amine containing particles are usually quite low, but short-term exposure spikes do occur in cities around the world. The addition of these amine compounds and other OM may have skewed the observed redox activities, since these are expressed per unit PM mass. Thus, if the condensable material deposited on the particles exhibited lower redox activity than the original particles it would decrease the DTT activity per unit mass observed for the ultrafine particles. Secondly, the contribution of transition metals in coarse particles, i.e., Fe and Cu, to the redox activity could not be determined by this assay, as these metals catalyze ROS production following a different reaction pathway (Fenton reaction, metals reduce H~2~O~2~ producing OH\u2022) [@bib0085]. Thus, this approach did not allow us to differentiate between metal and organic-based redox activity, and could limit the interpretation of these findings.\n\n4.2. Differential responses for different sized PM {#sec0095}\n--------------------------------------------------\n\nDose-dependent increases in HMOX-1 and IL-8 expression and decreases in cell viability were observed in this study for all three PM size fractions. However, only the ultrafine particle data supported the differential cellular responses proposed by Li et al. [@bib0135]. Significant activation of the antioxidant defense response, i.e., upregulation of HMOX-1, was observed at a lower ultrafine PM concentration (10 \u03bcg/ml), whereas proinflammatory response (IL-8) and cell death were induced at 50 and 100 \u03bcg/ml particle mass concentrations, respectively. This finding was consistent with other studies using diesel exhaust particles (size \\<1 \u03bcm) that have shown that cytoprotective pathways are induced at relatively low levels of oxidative stress and that cellular apoptosis or necrosis occurs at a higher level of oxidative stress [@bib0135], [@bib0280]. Li et al. [@bib0135] suggested that the hierarchical oxidative stress model was governed by the OC content of particles, especially PAHs. The ultrafine samples collected in Toronto were also highly enriched in PAHs compared with coarse and fine PM. By contrast, for coarse PM, while cytotoxic effects were observed at the lowest PM dose (10 \u03bcg/ml), significant upregulation of HMOX-1 was not observed until a higher dose was reached. Thus, the results for the coarse PM were inconsistent with the hierarchical oxidative stress model proposed by Li et al. [@bib0135] and hence this model is likely not applicable for all particle sizes.\n\nMore generally, it appears that the cells were responding differently when exposed to the coarse vs. quasi-ultrafine PM. The possibility exists that metals, different water-soluble or organic species, and endotoxins could trigger different or a combination of biological pathways leading to different responses. For example, several studies have also suggested that the presence of endotoxins on the particle surface may initiate inflammatory responses in the cells, particularly for larger particles [@bib0040], [@bib0150], [@bib0240], [@bib0245], [@bib0265]. A recent study from our group measured the endotoxin levels in concentrated coarse (293 EU/mg) PM at this site, and found that human exposure to these concentrations was associated with inflammatory responses [@bib0020]. The possibility that particles of differing size or composition may elicit alternate or additional biological pathways further complicates interpretation.\n\n4.3. Effects of different physicochemical properties {#sec0100}\n----------------------------------------------------\n\nExposure to specific components within the different size fractions showed different degrees of responses ([Fig. 4](#fig0020){ref-type=\"fig\"}, [Fig. 5](#fig0025){ref-type=\"fig\"}, [Fig. 6](#fig0030){ref-type=\"fig\"}). If size has no impact on the biological responses, then similar responses should have been observed following exposure to the equivalent mass concentrations of the components of different size fractions. Thus, the different linear regression lines for the different components indicated that the observed responses were not caused by total or component mass alone; initial particle size may also play a role in inducing the response.\n\nSimilarly, the variability in our observed responses cannot be explained based on size alone; otherwise the order of response for different sizes would be expected to always be the same across studies. *In vitro* studies conducting spatial and temporal PM toxicity studies have not reported any consistent relationships between biological responses and PM sizes [@bib0040], [@bib0060], [@bib0120], [@bib0190], [@bib0225]. It is likely that the differences in composition within different PM size-fractions played an important role on the observed responses. Higher slope values for transition metals (Cr and Ti) and OM compared with total PM mass and other chemical elements suggested their stronger influence on the observed responses. The relative abundance of transition metals in coarse PM might be related to the observed cytotoxic effects [@bib0160], whereas ultrafine PM with higher OCs could induce greater activation of inflammatory response [@bib0140].\n\nA third important physical property that can play an important role in the extent of PM toxicity is particle surface area, particularly in the case of the smaller particles. Several studies have suggested that the surface area of particles is the critical determinant of their biological effects and should be used as a dose metric when comparing particle-induced effects [@bib0055], [@bib0090], [@bib0175], [@bib0275]. Depending on the size of the particles, equal masses of PM could be comprised of quite different numbers of particles, which would lead to significant differences in surface area. For monodispersed particles with unit density, only one particle/cm^3^ with a diameter of 2.5 \u03bcm would constitute a mass concentration of 10 \u03bcg/m^3^ compared with the same mass concentration of particles 20 nm in diameter would comprise about 2.4 million particles/cm^3^ with 100 times more surface area [@bib0180]. In this study of size-fractionated particles, all experiments were conducted based on an equal mass exposure. Thus, the quasi-ultrafine particles were comprised of the greatest number of particles and total surface area for a given mass. Given the differential responses to the three size fractions, it would not be appropriate to assume that the observed biological effects could be fully described by a single property, i.e., size or surface area. It is most likely that the observed biological effects were the combined effect of mass, size, surface area, and chemical composition. Furthermore, correlation coefficients calculated between surface area and biological responses did not support this hypothesis; suggesting that the complexity of the cellular responses is much greater than originally hypothesized.\n\n4.4. Potential emission sources related to the biological responses {#sec0105}\n-------------------------------------------------------------------\n\nThe correlation analyses between chemical elements and biological responses helped to determine the toxicological contribution of different constituents toward the biological endpoints for the different particle sizes. For coarse and fine particles, HMOX-1 expression and cell viability correlated significantly with transitional metals (Cr, Cu, Fe, and Ti) and OC fractions, whereas for ultrafine particles, correlations were found mainly with the OC fractions. IL-8 mRNA upregulation was affected by both transition metals and OC for all size fractions. Correlation analyses among elements also helped to identify potential PM sources responsible for the observed effects (data not shown). For coarse and fine particles, Ba, Cr, Cu, Fe, Mn, Si, Ti, and Zn, were significantly inter-correlated, indicating that the elements might be coming from similar sources. As the study site is located next to a busy street, potential sources for these metals could be traffic exhaust and non-exhaust emission sources, i.e., brake and tire wear [@bib0250]. Contaminated road dust and roadside soils could also contribute to the metal concentrations [@bib0205]. Further, these metals have all been found to exhibit diurnal and weekday/weekend patterns at this site, similar to other traffic-related pollutants. OC, which correlated significantly with Fe, Mn, and Ti for both coarse and fine particles, could also be produced by the same traffic sources [@bib0250]. In the ultrafine size range, higher EC and OC concentrations were observed in the morning and afternoon, respectively, indicating that the primary emission (in the morning) and secondary particle formation (in the afternoon) could be the sources of ultrafine in that location [@bib0005]. The particle size distribution of the ambient air (maximum number concentrations for 10--30 nm particles) also suggested that primary emissions from vehicles were a major source of ultrafine particles [@bib0200].\n\n4.5. Limitations of the study {#sec0110}\n-----------------------------\n\nThere are several limitations to this study. Firstly, filter extracts were used for the cell exposure experiments. The filter extraction methodology used in this study emphasizes the mass/composition-related responses rather than those associated with the particle number or surface area, as the filter collection/extraction process was presumed to not conserve the original number concentration, size distribution or surface area properties of the particles. Secondly, particles were added to the cells as a suspension in culture medium, which differs substantially from the actual deposition of airborne particles onto the respiratory cell surface. The quantity (i.e., mass or number) of particles that actually interacts with the cells could, therefore, not be determined. Thus, it is not surprising that the relationships between the physicochemical properties (number concentration, surface area, and chemical composition) and biological responses were significant when considering composition alone. Thirdly, particles for the chemical analysis and biological tests were collected on different filters and were extracted in different media. Therefore, the particle extraction efficiency could be different among the filters used. Fourthly, the concentrating process for the ultrafine PM causes a shift in the size distribution for small particles (20--40 nm) [@bib0200]. Composition of the ultrafine particles was also altered by the addition of excess OM during the condensation process [@bib0155]. These factors could have led to a reduction in the redox activity of ultrafine particles. Furthermore, as noted above, the method of assessment of redox activity did not allow the differentiation between metal and organic-based redox activity. Finally, the biological responses observed in the A549 cell line limits the interpretation of these findings to whole organism exposures. Future studies using primary cells will be helpful to confirm and translate these findings into higher models (i.e., whole animal/human exposures).\n\n5. Conclusions {#sec0115}\n==============\n\nThis *in vitro* study investigated the comparative biological effects of size-fractionated ambient PM on human alveolar epithelial cells. This study demonstrated that exposure to ambient PM can initiate mass-dependent antioxidant, proinflammatory, and cytotoxic responses. The biological responses correlated strongly with transition metals and OC fractions for the coarse and fine particles, and with OC fractions for the quasi-ultrafine particles. For all three size fractions, traffic-related emissions appeared to trigger the biological responses. However, the biological responses did not correlate consistently with any specific particle size fraction. These observations indicate that these responses are dependent upon physicochemical properties, such as size and composition, and that these yield complex patterns that preclude making simple generalizations. It appears that the observed biological responses were caused by the combined effects of different physicochemical properties of the particles. Therefore, it is important to consider the combined influences of these physicochemical properties when explaining PM toxicity.\n\nAppendix B. \\[{(Appendix A)}\\]Supplementary data {#sec0125}\n================================================\n\nThe following are the supplementary data to this article:\n\nThe authors thank Drs. Jeffrey R. Brook and Frances Silverman for supporting our collaboration in the HEAT campaign. Thanks to Dr. Mike Fila for his assistance with the particle collection. Operational funding was provided by Natural Sciences and Engineering Research Council of Canada (NSERC) and Canadian Institutes of Health Research (CIHR) through the Collaborative Health Research Projects (CHRP) and the AllerGen Network of Centres of Excellence (07-A5). Infrastructure for the Southern Ontario Centre for Atmospheric Aerosol Research (SOCAAR) was funded by the Canada Foundation for Innovation (CFI) and the Ontario Research Fund (ORF). Umme Akhtar was funded by Ontario Graduate Scholarship (OGS), Queen Elizabeth II/Richard Quittenton Scholarship, McAllister Graduate Fellowship, and Edward Jarvis Tyrell Fellowship.\n\nAvailable online 16 May 2014\n\nSupplementary data associated with this article can be found, in the online version, at [doi:10.1016/j.toxrep.2014.05.002](10.1016/j.toxrep.2014.05.002){#intr0005}.\n\n[^1]: Current address: Physical Research Laboratory, Ahmedabad, India.\n\n[^2]: Contributed equally as senior authors of this manuscript.\n"}
+{"text": "\n"}
+{"text": "Q: Last month you paid tribute to your colleague Dr\u00a0Sheikh Humarr Khan from Sierra Leone, who ran the only medical unit in the world devoted exclusively to the care of patients with viral haemorrhagic fever and who died of Ebola virus disease in July. Why are so many health workers dying, even those wearing the recommended protective clothing?\n\nA: There is a lot of focus on personal protective equipment -- the suits, gowns and masks seen on so many photographs of the outbreak. But this is only one of many important aspects of safe patient care. In many places the demand for patient care has outstripped the available human resources. For example, a couple of months ago a WHO physician, David Brett-Major, and I made rounds at the Kenema treatment centre in Sierra Leone, where the nurses were on strike. The two of us were the only health-care workers on that ward of 60 patients with Ebola virus disease. Even if you wear the recommended gear, much more is needed, such as supervision and sanitation officers to decontaminate the area regularly. Asking a health-care worker to safely care for patients without support personnel is like asking a pilot to fly a plane without mechanics and flight controllers.\n\nQ: Thousands more cases are expected in the coming weeks, but with so many health workers who are sick or have died, the health workforce in the affected countries is being depleted. How can WHO and its partners recruit more staff quickly enough to support these countries on the front line of this outbreak?\n\nA: First you need to find the right people. Ideally they should have expertise in viral haemorrhagic disease, which -- before this outbreak -- was rare. Very few of these experts were clinicians doing patient care. Most were laboratory researchers and epidemiologists. Recruiting new people is not easy. Many of them have jobs they cannot leave and families that don't want them to go. It's just as difficult to recruit people in the affected countries, especially when the previous post holder died, and no one in western Africa is getting rich from doing this kind of patient care. We, the expatriate field workers, get a lot of media attention, but -- when you think of the risks they take -- the real heroes in this outbreak are the Guineans, Liberians and Sierra Leoneans who are living and working in their communities, and facing unimaginable fatigue and stress every day. We go for a month, then get a rest. For them it never stops.\n\n*Q: You and your colleagues published a call for action to promote more research on the prevention and treatment of Ebola virus and other filovirus diseases in*The Journal of Infectious Diseases*in 2007* (*196:S136--41). What came of this?*\n\nA: We called for the establishment of infrastructure for clinical research on filoviruses in Africa, with skilled staff and a legal--logistical framework in the affected countries and internationally. We proposed that research protocols, data collection forms and culturally-appropriate methods of gaining informed consent should be prepared and approved in advance of any outbreak through an international ethics review process. If these and other basic principles had been put in place back then -- I'm not saying we could have prevented the current outbreak -- we would have been in a better position to tackle it. We proposed a way forward. It was a missed opportunity. Now people are discussing this again.\n\nQ: In the 2007 article, you noted that, in the absence of vaccines and drugs, quarantining and contact tracing were the key public health approaches to stopping a filovirus outbreak. Why have these failed in some places during the current outbreak?\n\nA: It's not the first outbreak where local communities have resisted these measures, but it's the one where resistance has been the most intense and violent. It's easy to point a finger and say \"how stupid\" that people in the affected communities don't follow the measures we recommend. But we have to understand that this distrust of foreigners is in part the legacy of the unjust colonial era and less than ideal governance since.\n\nQ: WHO has faced a storm of criticism over its response to the outbreak. Could the crisis have been prevented and, if so, at what point?\n\nA: When I went to Guinea in April, shortly after the epidemic was declared, it seemed to me to be \"a routine Ebola virus disease outbreak\". The days started with the usual meeting at the WHO Representative's office to go over the epidemiological updates with the usual international partners: M\u00e9decins Sans Fronti\u00e8res (MSF), CDC and the Red Cross. At the beginning, people say: \"We know how to do this, let's dive in and get this mopped up.\" When I came home at the end of May, it looked as if the outbreak was coming to an end, with very few new cases. The WHO clinical team debated whether we needed to recruit more clinicians. But you can't monitor every village. You watch and wait. Sometimes it's over. Sometimes someone knocks on the door again saying: \"Six people recently died in a village.\" And that's what happened. When I went to Sierra Leone six weeks later, there was a distinctly different feel about the outbreak response. The first morning I went to the WHO Representative's office, there were very few members of staff from WHO and the other international organizations. Clearly they had limited human and financial resources, and had expended them.\n\nQ: Is that when you realized the outbreak was not getting under control?\n\nA: Yes. We clearly recognized the need for more clinicians, logisticians, epidemiologists, everything. MSF said they had the capacity to establish a ward in Kailahun but not in Kenema -- the two worst affected districts in Sierra Leone. It was the first time since MSF had started working on Ebola virus disease outbreak response that I heard them say they didn't have enough resources. After a few months everyone is exhausted and needs to go home. When clinicians get tired, their work becomes more dangerous, but it's expensive to replace them. The outbreak response had outstripped the available resources. WHO recognized this and sounded the alarm at a meeting in Accra, Ghana (2--3 July). Since then, various groups have pledged resources. For example, CDC has deployed more than 50 epidemiologists across the affected region and WHO has increased their personnel as well, but we are all late and it has gotten out of control. It's too simplistic to lay the blame on one group. There has been a lot of finger pointing at WHO, no one is immune to criticism, but WHO has suffered a loss of personnel and resources. So it's not only about what we should have done at any particular time, but the whole foundation for an international public health response that has been eroded by the global economic downturn.\n\nQ: Why was this Ebola virus disease outbreak different from previous ones?\n\nA: It started in the border area of the three countries: remote, impoverished areas with virtually no health or surveillance infrastructure. So it has been difficult to do contact tracing, organize isolation and provide treatment. The local populations regularly cross the borders to visit or trade with their ethnic kin. The borders are porous, but the public health authorities have no system for coordinating cross-border matters and, even within each country, health systems are weak at best. This has been exacerbated by language barriers: in Guinea the national language is French, in Liberia and Sierra Leone it is English.\n\n\"The whole foundation for an international public health response ... has been eroded by the global economic downturn.\"\n\nQ: Research is vital for learning about Ebola virus disease and developing disease control methods. What have been the challenges for research in this outbreak?\n\nA: It's not easy to find people to work in the outbreak let alone those who are able to do research. For example, during the war when the medical school was shut down, Liberia had only a handful of trained doctors in the whole country. Over a decade ago, WHO, Tulane, and other partners started the Mano River Union Lassa Fever Network to build capacity for combatting that disease (another viral haemorrhagic fever) in Sierra Leone, Guinea and Liberia. I remember trying to recruit a new laboratory technician looking at r\u00e9sum\u00e9s from people who had spent two or three years in refugee camps and who hadn't had the opportunity to develop skills in the sciences. You can give them the technical training, but that's not enough for them to become a principle investigator, who can frame a hypothesis, write a research proposal, seek funds, implement the project with a team, analyse the data and publish the results. This requires a culture of research and this is something we were trying to build. But so many health-care workers have died of Ebola virus disease, in addition to the personal tragedy, this loss is a major blow to these efforts to advance research in these countries.\n\nQ: You were at the WHO meeting last month to discuss the possible use of experimental vaccines and drugs, has this brought the end of the outbreak closer?\n\nA: WHO brought together a vast array of experts from regulatory agencies, pharmaceutical companies, public health agencies and scientific institutes. The meeting was productive but the devil is in the detail. We have a moral imperative to accelerate the pace of these experimental products through the pipeline, but the challenge before us is formidable and the path not 100% clear. At present, in the absence of vaccines and therapies, many who are sick don't want to be traced, but prefer to disappear into the forest or urban jungle. If we can start providing vaccines and drugs, hopefully in the not too distant future, the problem will change and people will start knocking on the door demanding prevention and treatment, so this is a public health strategy as well. But we won't be able to get experimental compounds to everyone in need. Stemming this outbreak will still depend primarily on the classic strategy of case identification, with isolation and treatment, and contact tracing.\n\nQ: Are there other glimmers of hope?\n\nA: We have a major public health crisis across this region that threatens to get larger. There are still measures we can take to prevent the crisis spreading and we must continue these efforts. We are working to make things happen faster, but this requires cooperation from many groups across the world. Everyone recognizes the need to take action to stop the outbreak, but it's not easy. I do think that everyone is heeding the call now and doing what they can. In the long-term, the scale and public profile of this outbreak means that potential vaccines and therapies that were stalled are now being pushed through clinical trials. I hope that we have something to offer people during this outbreak but, at the very least, we must emerge more prepared for the next. Another long-term glimmer of hope is that there are discussions about how to help these three long-suffering countries. They are going to need a Marshall Plan for reconstruction when this is over, with infrastructure and educational programmes to put some of the world's poorest nations on the road to health and prosperity.\n\nDaniel Bausch has devoted the last 20\u00a0years to the study and control of Ebola virus and other emerging tropical diseases. He is an associate professor in the Department of Tropical Medicine at Tulane University Health Sciences Center, New Orleans, the United States, and is seconded to the US Naval Medical Research Unit No.\u00a06 in Lima, Peru, where he heads the Virology and Emerging Infections Department. From 1996 to 2003, he worked at the Centers for Disease Control and Prevention (CDC) Special Pathogens Branch. Bausch often serves as a consultant for the World Health Organization (WHO) in the field and as an adviser. He trained in internal medicine at Boston University and in infectious diseases at Tulane, where he also earned his Masters of Public Health in Tropical Medicine in 1994.\n\nDaniel BauschWHO/F Fleck\n"}
+{"text": "We appreciate the comments by Dr. Fukui and colleagues about our paper ([@bib1]), and we fully agree with some of their views. However, there are some issues that need to be clarified and discussed.\n\nWe acknowledge and agree that right ventricular (RV) function is closely linked with the afterload. In our cohort study, 40 (34.2%) patients developed acute respiratory distress syndrome (ARDS), which contributed to increase RV afterload. Furthermore, both ARDS and right ventricular longitudinal strain (RVLS) were found to be associated with mortality, consistent with the previous study ([@bib2]).\n\nIt is difficult for us to evaluate the right ventricular to pulmonary artery (RV-PA) coupling by measuring a ratio between tricuspid annular plane systolic excursion and pulmonary artery systolic pressure (PASP), as suggested by Cavalcante ([@bib3]). Noninvasive PASP measurement depends on the presence of tricuspid regurgitation (TR), and there were only 61 patients (50.8%) had interpretable TR jet signal in our study. Among them, the majority was mild or trivial TR, which may affect the accuracy of PASP assessment. As the right heart catheterization was not available in our designated treatment hospital, PASP derived from echocardiography may be less accurate to evaluate the RV-PA coupling.\n\nCOVID-19 patients with underlying cardiovascular disease are more likely to have more severe clinical course and represent a higher proportion of the patients who are at greater risk of mortality. Although cardiovascular disease has a marginal significance in the univariate analysis (p\u00a0=\u00a00.058), it was not included in the multivariate analysis due to the smaller number of endpoints.\n\nSpeckle-tracking echocardiography is dependent on image quality, which may relatively preclude its wide application in clinical practice, especially in patients who had a poor echocardiographic window. Therefore, we had to exclude 24 patients due to suboptimal image quality. Except for RVLS, RV size and function parameters were measured in these excluded cases. Our study showed that there were no significant differences in RV structure and function parameters between patients included for final analysis and those who were excluded. Therefore, we think the effect of excluding cases on our findings may be negligible.\n\nWe agree with the comments of Dr. Fukui and colleagues regarding the need for exploring the incremental prognostic value of RVLS over comprehensive clinical assessment. However, the small number of observed events does not enable us to explore the additive value of RVLS over the known risk factors. Future studies with larger simple sizes are needed to determine the additional prognostic value of RVLS over the other clinical and echocardiographic parameters in COVID-19 patients.\n\nWe are also delighted to respond to Dr. Schiavon and colleagues, who commented on the relatively high prevalence of deep vein thrombosis (DVT) and no cases of pulmonary embolization (PE) in our study. Each patient in the present study received both bedside echocardiography and the lower-extremity venous ultrasound. The levels of DVT included the bilateral common femoral, deep, and superficial femoral, and the popliteal veins as well as the posterior tibial, peroneal, and calf muscle veins. Moreover, if there was any clinical suspicion of PE, computed tomographic angiography of the pulmonary arteries (CTPA) was considered ([@bib4]) and obtained, if possible. In the present study, CTPA was performed in 2 patients with DVT, but neither of them was diagnosed with PE. Although previous studies demonstrated an increased risk of DVT and potential PE in respiratory and other intensive care settings ([@bib5]), our data suggested that CTPA-confirmed PE was less frequent in this group of patients. Nevertheless, the finding was limited because we did not perform screening CTPA for every patient with DVT.\n\nWe agree with the comments by Dr. Schiavon and colleagues about the effect of the timing of echocardiographic study and ventilation on RV echo parameters. In our study, there were 105 patients (87.5%) with oxygen therapy, 6 patients (5%) with noninvasive mechanical ventilation, and 15 patients (12.5%) with invasive mechanical ventilation at the time of echocardiographic examinations. Indeed, two-thirds of patients with invasive mechanical ventilation had the worst RV longitudinal strain.\n\nWe acknowledge that the proposed cutoff values for RV function are largely within the normal range. Except for the thoughtful explanation proposed by Schiavon and colleagues, another important reason may be the small proportion of patients (22%) admitted to the intensive care unit as well as those (12.5%) receiving invasive mechanical ventilation.\n\nPlease note: \u2020Drs. Y. Li, H. Li, S. Zhu, and Y. Xie contributed equally to this work. \u2021Drs. Li Zhang and Mingxing Xie are contributed equally to this manuscript as co-corresponding authors. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.\n\nThe authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the *JACC: Cardiovascular Imaging*[author instructions page](http://imaging.onlinejacc.org/content/instructions-authors){#interref1}.\n"}
+{"text": "Background {#Sec1}\n==========\n\nHeterotopic ossification (HO) is a relatively common complication following hip surgery treated with open reduction and internal fixation (ORIF), total arthroplasty (THA) or hemiarthroplasty. For example, a recent study showed the prevalence of HO after minimally invasive short-stem THA using a modified anterolateral approach to be 7.8% (16 out of 216 cases) \\[[@CR1]\\]. Development of HO after hip surgery is an important clinical issue as it can affect functional status \\[[@CR2]\\]. Means to prevent or reduce HO have been explored and include pre-operative \\[[@CR3]\\] and post-operative irradiation \\[[@CR4]\\] and the use of non-steroidal anti-inflammatory drugs \\[[@CR5], [@CR6]\\]. Rate and degree of HO after THA were shown to be affected by patient gender \\[[@CR7]\\], surgical approach \\[[@CR8], [@CR9]\\] as well as type of surgery \\[[@CR10]\\]. Moreover, African-American ethnicity has been shown to be an independent risk factor for HO formation after THA \\[[@CR11]\\]. An increased occurrence of HO was also reported specifically in patients with ankylosing spondylitis, elevated preoperative serum inflammatory markers and prolonged duration of surgery \\[[@CR12]\\]. Previous hip HO formation and bilateral hypertrophic hip osteoarthritis are other known risk factors for HO \\[[@CR13]\\].\n\nHO shows progression over time in radiographic appearance. In early stage, it is typically a soft tissue mass without overt calcification and can often be missed. In the mineralization stage which can occur within 10\u2009days after causative insult, calcification usually starts peripherally. Lesions can also be poorly organized without recognizable mineralization pattern. In mature HO, cortical bone is formed. The degree of HO can be semi-quantitatively assessed using the Brooker classification (grade 1--4) \\[[@CR14]\\]. So far, the relationship between time of surgery and the severity of postoperative HO has not been well established in the literature. In a single center retrospective study, interval from injury to surgery was not statistically significantly associated with development and severity of HO in a cohort of 241 patients with acetabular fractures \\[[@CR15]\\]. Another retrospective study showed that patients who underwent THA for acetabular fracture early after injury had higher (4-fold) chance of developing HO \\[[@CR16]\\]. However, in other retrospective and prospective studies of HO after surgical repair of elbow fractures, longer time to surgery was an independent predictor of HO \\[[@CR17], [@CR18]\\]. Thus, the existing literature evidence on the relationship between time to surgery and the incidence/severity of HO remains controversial. We hypothesized that the longer the time to surgery after hip injury, the more severe the postoperative HO will become.\n\nThe aim of our study was to determine the association between severity of heterotopic ossification around the hip joint and the interval between the time of injury and surgery.\n\nMethods {#Sec2}\n=======\n\nSubjects {#Sec3}\n--------\n\nOur retrospective study received Institutional Review Board approval and the need for informed consent from the patients was waived. We retrospectively reviewed the medical records in our institution for patients who had hip bipolar hemiarthroplasty (CPT codes 27125 and 27236) performed by an orthopedic surgeon to treat femoral neck fractures between 01/01/2003 and 11/22/2013. For each patient, the date of surgery, laterality (left or right hip), surgical approach (lateral, posterior, anterolateral), patient age, date of injury and interval between injury and surgery (days) were recorded. We excluded patients who had postoperative complications such as re-fracture, and hardware related complications including loosening, fracture and infection.\n\nRadiographic evaluation of heterotopic ossification {#Sec4}\n---------------------------------------------------\n\nUsing post-operative radiographs of the pelvis/hip, the severity of post-operative heterotopic ossification was semi-quantitatively graded using the Brooker classification as follows: class 1\u2009=\u2009islands of bone within the soft tissues about the hip; class 2\u2009=\u2009bone spurs from the pelvis or proximal end of the femur, leaving at least one centimeter between opposing bone surfaces; class 3\u2009=\u2009bone spurs from the pelvis or proximal end of the femur, reducing the space between opposing bone surfaces to less than one centimeter; class 4\u2009=\u2009apparent bone ankyloses of the hip (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} and Table\u00a0[1](#Tab1){ref-type=\"table\"}) \\[[@CR14]\\]. Pelvis/hip radiographs were read in consensus by one attending musculoskeletal radiologist and one musculoskeletal radiology fellow blinded to clinical information. If a patient had more than one follow-up radiographs, the most recent radiograph was reviewed.Fig. 1Brooker classification of heterotopic ossification in the hip. **a** Class 0 -- no heterotopic ossification is noted about the hip; **b** class 1 -- small islands of soft tissue ossifications are noted about the hip (arrows); **c** class 2- a moderate spur (arrow) arising from the right acetabulum, leaving an osseous gap of greater than 1\u2009cm measured from the greater trochanter; **d** class 3 -- a very large spur (arrow) arising from the greater trochanter of the right proximal femur approaching the acetabulum, leaving an osseous gap of less than 1\u2009cm; **e** class 4 -- arrows show large spurs arising from the acetabulum and also greater trochanter of the right proximal femur, resulting in ankylosisTable 1Brooker classification of heterotopic ossification in the hipClassDefinition1Islands of bone within the soft tissues about the hip2Bone spurs from the pelvis or proximal end of the femur, leaving at least 1\u2009cm between opposing bone surfaces3Bone spurs from the pelvis or proximal end of the femur, reducing the space between opposing bone surfaces to less than 1\u2009cm4Apparent bone ankylosis of the hip\n\nStatistical analysis {#Sec5}\n--------------------\n\nSince the number of subjects with HO class \"3\" and \"4\" were low (*N*\u2009=\u200911 and 4, respectively), they were combined together as \"3+\" in the following analysis. Chi-square tests with exact *P*-values based on Monte Carlo simulation were utilized to examine the marginal association between categorical variables (*gender, approach, side*) and HO class. Kruskal-Wallis tests were used to examine the marginal association between continuous variables (*days to surgery, age*) and HO class. Proportional odds model was used to further quantify the relationship between HO class (treated as ordinal variable) and days to surgery. Score test was utilized to confirm the proportional odds assumption. Of note, these additional analyses were not performed for other variables, since the results of the aforementioned Chi-squared test and Kruskal-Wallis test showed statistically non-significant association with the severity of HO with *p*-values well above 0.05, and thus calculation of odds ratio would also be non-significant. Statistical analysis was performed using SAS 9.4 and significance level was set at 0.05 (SAS Institute Inc., Cary, NC).\n\nEvaluation of possible influence of time interval between the date of surgery and date of follow-up radiograph {#Sec6}\n--------------------------------------------------------------------------------------------------------------\n\nAs we reviewed the obtained dataset, it became clear that there was a large variation in time interval between the date of surgery and date of follow-up radiograph (range 76--3049\u2009days). One might therefore think patients who had longer interval for follow-up were more likely to have HO. We therefore assessed the distribution of follow-up interval (in days) among different HO categories (0,1,2,3+) and assessed if the duration of time interval between the date of surgery and follow-up radiograph affected the severity of HO.\n\nResults {#Sec7}\n=======\n\nTable\u00a0[2](#Tab2){ref-type=\"table\"} shows the descriptive table for patients' gender, age, days to surgery, surgery approach and side by HO class. Most of the patients were 65\u2009years or older, but 18 patients younger than 65\u2009years received bipolar hemiarthroplasty due to clinical indications such as delayed surgery secondary to compromised systemic status, poor general health that would prevent a second operation or displaced fracture which was several days old. Severe heterotopic ossification (HO 3+) was associated with the longer interval between the time of acute hip fracture and surgery (median 6\u2009days) vs. median 2\u2009days in all other groups (HO classes 0--2) (*p*\u2009=\u20090.0015). In other words, HO class 1 and HO class 2 had the same the interval between the time of acute hip fracture and surgery as patients without ossifications. Patient age did not significantly differ amongst different HO classes with a large range of overlap around the age 80 (*p*\u2009=\u20090.2812). Patient gender was also not associated with HO class (*p*\u2009=\u20090.0705) although higher proportion of male patients (compared to female patients) had HO class 2 (44.74% vs. 33.63%) and class 3+ (15.70% vs. 7.96%). Conversely, a higher proportion of female patients had HO class zero (30.09% vs. 10.53%) compared to men. Surgical approach and side of surgery showed essentially no association with HO class, with *p*-values much higher than 0.05 (*p*\u2009=\u20090.1882 for surgical approach, *p*\u2009=\u20090.7383 for side of surgery). Figure\u00a0[2](#Fig2){ref-type=\"fig\"} shows the distribution of subjects according to the number of days to surgery. A majority of patients (90 of 151, 59.6%) had a surgical intervention within 2\u2009days of presentation.Table 2Descriptive table for patients' characteristics and surgery information by HO classVariablesTotal (*N*\u2009=\u2009151)HO class 0 (*N*\u2009=\u200938)HO class 1 (*N*\u2009=\u200943)HO class 2 (*N*\u2009=\u200955)HO class 3+ (N\u2009=\u200915)*P*-valuesDays to surgery2\u2009\u00b1\u200932\u2009\u00b1\u200912\u2009\u00b1\u200932\u2009\u00b1\u200936\u2009\u00b1\u200960.0015Age81\u2009\u00b1\u20091182.5\u2009\u00b1\u20091683\u2009\u00b1\u20091180\u2009\u00b1\u20091177\u2009\u00b1\u2009100.2812GenderFemale113 (74.83%)34 (30.09%)32 (28.32%)38 (33.63%)9 (7.96%)0.0705Male38 (25.17%)4 (10.53%)11 (28.95%)17 (44.74%)6 (15.79%)Surgical ApproachAnterolateral4 (2.65%)0 (0.00%)3 (75.00%)1 (25.00%)0 (0.00%)0.1882Lateral33 (21.85%)5 (15.15%)8 (24.24%)15 (45.45%)5 (15.15%)Posterior114 (75.50%)33 (28.95%)32 (28.07%)39 (34.21%)10 (8.77%)Side of surgeryLeft82 (54.30%)20 (24.39%)24 (29.27%)28 (34.15%)10 (12.20%)0.7383Right69 (45.70%)18 (26.09%)19 (27.54%)27 (39.13%)5 (7.25%)\\*For categorical variables, *p*-value was based on Chi-squared test with exact *p*-value from Monte Carlo simulation; for continuous variables, median +/\u2212 interquartile range were reported and *p*-value was based on Kruskal-Wallis testFig. 2Distribution of subjects according to the number of days to surgery. A majority of patients (90 of 151, 59.6%) had a surgical intervention within 2\u2009days of injury. Subjects who had extended delays of surgery all had mitigating medical reason which prevented medical clearance for undergoing surgery\n\nProportional odds model was used to further analyze the relationship between HO class and time to surgery. The odds ratio and 95% CI for one level higher HO class was 1.296 (1.152, 1.459), which means that the odds of having HO class one level higher increases by about 29.6% for every one-day increase in the days to surgery. This meant that as days to surgery increased, patients were more likely to have higher class of heterotopic ossification.\n\nThe median number of days between the surgery and follow-up radiograph was 321, with a range of 76--3049. In our study sample, distribution of follow-up interval among different HO categories (0,1,2,3+) was similar (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}) and longer follow-up interval did not necessarily correspond to higher HO category. The patient with longest follow-up interval (3049\u2009days) had class 2 HO, and there was a patient who had no HO at 2746\u2009days. Conversely, a patient whose follow-up radiograph was taken at 96\u2009days had class 4 HO. When our sample was stratified according to HO class, the median follow-up interval for class 0 was 353.5\u2009days (range, 89--2746); class 1 was 306\u2009days (range, 76--2296); class 2 was 279\u2009days (range, 94--3049); and class 3+ was 321\u2009days (range, 95--2345). Thus, shortest median follow-up interval was observed with class 2 HO group, and the median follow-up interval was actually shorter for class 3+ group compared to class 0 group.Fig. 3Distribution of follow-up interval among different HO categories. The median follow-up interval for class 0 was 353.5\u2009days (range, 89--2746); class 1 was 306\u2009days (range, 76--2296); class 2 was 279\u2009days (range, 94--3049); and class 3+ was 321\u2009days (range, 95--2345). Thus, shortest median follow-up interval was observed with class 2 HO group, and the median follow-up interval was actually shorter for class 3+ group compared to class 0 group\n\nDiscussion {#Sec8}\n==========\n\nOur results noted a strong association between the severity of postoperative HO and time to surgery. Available literature evidence regarding risk of postoperative HO and its association with time interval between injury and surgery has been mixed. No significant association between development/severity of HO and time interval from injury to surgery was found in d'Heurle et al's study in which hip fracture patients were evaluated \\[[@CR15]\\], while Chemaly et al. reported patients who underwent 'early' surgery (total hip arthroplasty within 60\u2009days of injury) had higher incidence of HO compared to those who underwent 'late' surgery (more than 60\u2009days after injury) \\[[@CR16]\\]. Hong et al. showed that the risk of developing HO after elbow fracture and surgical fixation increased with time to surgery: using the \u226424\u2009h category as the reference, the 2--7\u2009days category had an OR of 3.78 (95%CI, 1.12--12.78; *P*\u2009=\u20090.033) and the \\>\u20097\u2009days category an OR of 10.62 (95%CI, 2.96--38.09; *P*\u2009=\u20090.001) \\[[@CR17]\\]. Also, Bauer et al. reported longer time to surgery was a risk factor for the development of HO, with subjects waiting 8 or more days for surgery having 12 times the odds of HO than subjects having surgery within a day of injury \\[[@CR18]\\].These last two studies are in line with the findings of our study. However, exact reason for this finding is yet to be determined, and given mixed literature evidence, it remains to be a controversial issue.\n\nPatient age can be an important factor for severity of HO, as one might consider younger patients to have more tendency for bone formation compared to elderly patients. However, our analysis showed that age is not a statistically significant factor associated with severity of HO. Each HO grade had very similar median age and IQR, as well as age range (grade 0, min 54\u2009years and max 94\u2009years; grade 1, min 52\u2009years and max 95\u2009years; grade 2, min 34\u2009years and max 94\u2009years; grade 39, min 33\u2009years and max 86\u2009years). Upon detailed review of individual patient data, there were three particularly young patients. The patient aged 33\u2009years had grade 4 HO, in whom interval between surgery and follow-up radiograph was 1894\u2009days. The patient aged 34\u2009years had grade 2 HO with follow-up interval of 2410\u2009days. The patient aged 37\u2009years had grade 4 HO with follow-up interval of 1159\u2009days. Then, the next youngest patient was aged 50\u2009years, who had grade 2 HO. Thus, there were only three 'outliers' in terms of age distribution of our study sample, and two of these three subjects had grade 4 HO. This could be related to speculation that young patients may be more likely to get severe HO. However, the number of patients is too small to derive statistically meaningful conclusion regarding these very young patients. For patients aged 50--69\u2009years, there were only grade 0, 1 and 2 HO's. The remainder of high grade HO's (class 3 and 4) were only found in patients aged 70\u2009years or older. The oldest age for grade 3 HO was 91\u2009years and that for grade 4 HO was 86\u2009years. Despite all these detailed observations, overall age does not seem to be significant confounding factor for HO grade severity, as demonstrated by our formal statistical analysis.\n\nWe did not adjust the analysis for any demographic factors which were previously reported as possible risk factors (such as gender, surgical approach, ethnicity, etc.), as our samples did not suggest that these were risk factors. Nonetheless, we fit a model that adjusted for gender and surgical approach. The results suggest that after further adjusting for gender and surgical approach, the odds of having HO class one level higher increases by about 30.2% for every one-day increase in the days to surgery (OR\u2009=\u20091.302, 95% CI: 1.158--1.463, *p*\u2009\\<\u20090.0001.) This is in line with our original findings.\n\nA limitation of our study is that there was a large variation in time interval between the date of surgery and date of follow-up radiograph (range 76--3049\u2009days). However, in our study sample, distribution of follow-up interval among different HO categories (0,1,2,3+) was similar (Fig. [3](#Fig3){ref-type=\"fig\"}). Of note, there were total 50 patients who had radiographic follow-up within 180\u2009days of surgery. Of these patients, 11 patients (22%) had class 0 HO, 17 patients (34%) had class 1 HO, 18 patients (36%) had class 2 HO, 4 patients (8%) had class 3+ HO. In contrast, 101 patients had follow-up x-ray longer than 180\u2009days after surgery. Of these patients, 27 patients (27%) had class 0 HO, 26 patients (26%) had class 1 HO, 37 patients (37%) had class 2 HO, 11 patients (10%) had class 3+ HO. Fisher's exact test shows the distribution of HO class is not significantly different between early and late follow-up groups (*p*-value\u2009=\u20090.7323). Moreover, when only using 101 patients who had follow-up x-ray longer than 180\u2009days after surgery, the estimated odds ratio for days to surgery was 1.354 with 95% CI: 1.159--1.583 (*p*\u2009=\u20090.0001) or was 1.374 with 95% CI: 1.176--1.606 (*p*\u2009\\<\u20090.0001) after further adjusting for gender and surgical approach, which suggested that the odds of having HO class one level higher increased by about 35.4% for every one-day increase in the days to surgery. The conclusions are consistent with our original analysis. It is generally thought that HO increases and to be more manifest during longer-term observation \\[[@CR19]\\]. Our finding does not agree with this common belief, and suggests longer follow-up interval does not necessarily lead to increased severity of HO.\n\nPotential confounders for our study included the use of NSAIDS and radiation therapy for reduction of risk of HO. However, review of medical record of all patients showed that no patients received any prophylactic or therapeutic NSAIDS or radiation during the study period, and thus our study was not affected by these factors. Another potential confounder is the mechanism of injury (high velocity injury vs. low velocity injury). However, in our study sample 148 of 151 patients had hip fractures following a mechanical fall (i.e. low velocity injury) making our sample mostly homogeneous. Three patients had hip fractures following \"motor vehicle accidents\" according to electronic medical record, but precise circumstance of injury (e.g. what type of accident, speed of collision, etc) was not fully described. One patient had class 2 HO and two patients had class 4 HO, but effects of a high velocity injury on HO severity need to be further evaluated with a larger sample size. Finally, we did not correlate for use of medication other than NSAIDS, severity of trauma or post-operative rehabilitation.\n\nThe etiologies for the association between time to surgery and increased severity of HO remain undetermined. Our analysis has ruled out some plausible risk factors but did not identify the actual causative factor, which needs to be explored in further studies.\n\nConclusions {#Sec9}\n===========\n\nOur study showed class 3 or greater HO was associated with longer time interval between time of acute hip fracture and surgery (median 6\u2009days) compared to all other groups (HO class 0--2), which had similar time interval between the fracture and surgery (median 2\u2009days). While it is not always possible, every possible effort should be made to minimize the delay in surgery to reduce the degree of HO.\n\nHO\n\n:   Heterotopic ossification\n\nORIF\n\n:   Open reduction and internal fixation\n\nTHA\n\n:   Total arthroplasty\n\n**Author statement:** Our work described in this manuscript was previously presented as an oral presentation at the Society of Skeletal Radiology 41^st^ Annual Meeting in 2018 in Texas, Austin, United States, March 25-28, 2018. Hayashi D, Ho C, Caruana D, Komatsu D, Nicholson J, Mufti M, Yang J, Zhu C, Gould E. Association between severity of heterotopic ossification around the hip joint and the interval between the time of injury and surgery. Abstract Book, p.93\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nAuthors would like to thank the following individuals for their contributions to this study but not fulfilling the criteria for authorship: Mathew Teng, MD, for radiographic analysis, and Brian Campfield, MD, for clinical data acquisition.\n\nConsent to publication {#FPar1}\n======================\n\nNot applicable.\n\nDH, ESG, CH, DEK, JN, JY designed this study. ESG performed radiographic analysis. MM performed review of electronic patient record. DLC collected clinical data related to the surgical procedures. JN performed surgeries. DH, ESG, CH, JY, CZ performed data analysis and interpretation. DH, ESG and CH prepared the manuscript, tables, and figures. All authors have read and approved the final manuscript.\n\nNo funding was received for this study.\n\nThe dataset supporting the conclusions of this article is proprietary to Stony Brook University Hospital and will not be shared, because the hospital restricts sharing of the raw data with concerned personnel only. For permission to access the data, contact Department of Radiology, Stony Brook Medicine, HSC 4--120, Stony Brook, NY 11794, USA.\n\nOur retrospective study received Institutional Review Board approval and the need for informed consent from the patients was waived.\n\nESG is a Consultant to Endo Pharmaceuticals Inc. DH is a Section Editor of BMC Musculoskeletal Disorders journal. All other authors report nothing to disclose.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nFor most women, pregnancy is an enjoyable and pleasant experience, but for many it can open a window to chronic stress due to pregnancy-related biopsychosocial and cultural stressors. Antenatal stress is a major concern in low and middle-income countries (LMIC), with a prevalence of 29--60%, owing to a higher risk of social, psychological, and physical hardships \\[[@CR1]--[@CR4]\\]. It is associated with many psychopathological\u00a0issues among women including perinatal anxiety and depression, suicidality, eating disorders, and increased fear of childbirth \\[[@CR5]--[@CR7]\\]. It is also associated with many birth and developmental abnormalities among the offspring such as growth retardation, emotional and behavioral problems, neurodevelopmental disorders, preterm births and poor APGAR scores \\[[@CR8]--[@CR11]\\].\n\nPakistan is\u00a0the sixth most populous country globally, with a population exceeding 200 million. It boasts a unique cultural identity in Asia \\[[@CR12]\\]. Pakistan has its own distinct culture, languages, food, gender disparities, and, most importantly, cultural traditions revolving around pregnancy that moderate antenatal care. Unfortunately, Pakistan is also struggling with several demographic and socio-economic stressors, frequent infectious outbreaks, terrorism, and traumatic events such as wars and internally displaced refugees that are among common reasons for the high burden of psychopathologies \\[[@CR13], [@CR14]\\]. Hence, it presents a very peculiar set of biopsychosocial factors that can precipitate high levels of perinatal stress \\[[@CR3]\\].\n\nDue to the detrimental impact of antenatal stress, it is important to explore the prevalence and factors related to it. Factors related to treatment seeking attitudes, access to healthcare, and preference of birth-related procedures have been shown to correlate with common mental health disorders in Pakistan as well as in other countries \\[[@CR4], [@CR12]\\]. Most of the world's population lives in LMIC, boasting the highest population growth rates. Unfortunately, there are only a few studies focused on exploring antenatal stress in the sociocultural context of the developing world. Therefore, this study was designed to explore the prevalence of prenatal stress and its association with socioeconomic background, cultural beliefs, disparities related to gender of offspring such as desire for a son and access to healthcare in the province of Punjab, Pakistan.\n\nMain text {#Sec2}\n=========\n\nMethods {#Sec3}\n-------\n\nFrom November 2016 to March 2017, a cross-sectional study was conducted at two locations: the District of Faisalabad and Charanwala Village in the District Mandi Bahauddin. Faisalabad is a major metropolitan city in Punjab, Pakistan. Study participants were recruited from two teaching hospitals in Faisalabad, namely: District headquarter hospital and Children Hospital. The third site was the Basic Health Unit in Charanwala village in the district Mandi Bahauddin. The former two sites are the only secondary and tertiary care and publicly funded hospitals, providing obstetric and gynecological care to urban and rural population residing within Faisalabad and adjoining rural areas. While the Basic Health Unit in Charanwala village caters to the primary health care needs of the rural families. These three sites thus, provided a homogeneous study sample, representative of both the urban and rural population in the province of Punjab.\n\nA systematic sampling procedure was used, where every fifth patient presenting at obstetrics and gynecology departments of the selected hospitals were interviewed by a six-member team consisting of physicians, nurses and medical students. Ethical approval was sought and received from FMH College of Medicine and Dentistry, Lahore, Pakistan (FMH-06-2017-IRB-261-M). Participation in this study was voluntary and all participants signed the informed consent form.\n\nAll respondents were interviewed using the preformed questionnaire consisting of four parts: (1) variables related to demographic characteristics and obstetric history; (2) a 4 item Perceived Stress Scale (PSS-4) and (3) an Urdu version of Pittsburgh Sleep Quality Index (PSQI-U). Results for sleep pattern based on PSQI scale have been reported in another study (Waqas et al. In press).\n\nThe section on demographic characteristics enquired information pertaining to age, background, religion, ethnicity, socioeconomic class and education. The second section of the questionnaire pertained to obstetric variables including parity, status of pregnancy, trimester, previous complications during past deliveries by hospital staff and untrained midwives, any comorbidities, and gender of previous offspring and gender of fetus as evident on ultrasonography. While the third section included variables related to family dynamics included household decision maker, preference for education for daughters and sons, gender preference for fetus among in-laws, experience of marital problems, harassment and domestic abuse. These items were asked using close ended question format.\n\nPerceived Stress Scale (PSS-4) scores are obtained by reverse coding the positive items, and then summing across all 4 items. It exhibited a good internal consistency in the present study sample (Cronbach's alpha\u2009=\u2009.793). As indicated in previous studies, stress levels were dichotomized into low and high, by combining upper and lower two quartiles \\[[@CR15], [@CR16]\\].\n\n### Statistical considerations {#Sec4}\n\nTotal sample size for this study was estimated to explore several research questions. To estimate prevalence of high stress levels during pregnancy, a minimum sample size of 385 was required based on following parameters: 95% confidence level, 5% margin of error and a population of 33,85,000 in Faisalabad. For logistic regression analyses, most conservative estimates recommend 20 participants per predictor as a rule of thumb, thus, indicating a sample size of 240 to 300 as appropriate for a model comprising of 12--14 predictors. All data were analyzed in SPSS v.20 (IBM, Chicago, IL). Categorical variables were presented as frequencies and continuous data as mean (SD). Logistic regression analysis (backward method) was run with covariates demonstrating significant univariate associations, to identify the significant predictors of antenatal stress in our study sample. Introduction of predictors in the regression model was guided by significance levels of the bivariate analysis between PSS-scores and different variables, biological plausibility and previous literature (4).\n\nResults {#Sec5}\n-------\n\nOut of 550 respondents, 516 consented to participate in the study yielding a response rate of 93.82%. The mean age of respondents was 29.82 (6.50) years. Among socioeconomic background, most of the respondents were middle class 364 (70.54%), followed by low 135 (26.2%), and upper class 17 (3.3%). Most of the respondents were educated from 5th to 10th grade 260 (50.39%), followed by illiterate 134 (26%), and undergraduates 122 (23.64%). A total of 260 respondents (50.4%) belonged to rural areas, followed by semi-urban (185, 35.9%) and urban areas (71, 13.8%). Most were house-wives (426, 82.6%), while the rest were either hobbyists or employed. A majority were natives of Punjab (395, 76.6%) while the rest had migrated from other provinces.\n\nThe mean score of the respondents on Perceived Stress Scale was 7.55 (3.43). A total of 218 (42.2%) respondents reported higher stress levels. The mean number of offspring borne by women were 2.32 (2.1). Many respondents reported unplanned pregnancies, a history of miscarriage (n\u2009=\u2009101, 19.6%) and self-induced termination of pregnancy (n\u2009=\u200950, 9.7%). A higher proportion of women had past delivery procedures performed at hospitals (n\u2009=\u2009317, 61.43%). However, the respondents reported a significant percentage of past delivery procedures performed by midwives (n\u2009=\u2009256, 44.6%) as well. A higher proportion of respondents reported complications during birth procedures performed by midwives (51.95%) than physicians (17.98%). A significant percentage of respondents reported having chronic illnesses (Table\u00a0[1](#Tab1){ref-type=\"table\"}).Table\u00a01Response distribution to variables related to obstetric historyVariableSubcategoryFrequencyPercentage (%)Mean (SD)\u03c7^2^ statisticP-valueTrimester1st12323.851.78\\<\u2009.0012nd24848.13rd14528.1PregnancyPlanned30358.715.70\\<\u2009.001Unplanned21341.3Complication during past deliveries in hospitalYes5711.1313.19\\<\u2009.001No45988.9Complication during past deliveries by midwivesYes13325.8121.12\\<\u2009.001No38374.2Any chronic illnessYes11121.5167.51\\<\u2009.001No40578.5Number of live births2.32 (2.1)Number of still births.295 (.84)Terminated pregnancies.122 (.41)Number of sons1.06 (1.21)Number of daughters1.24 (1.38)Age of the eldest child6.24 (5.51)Children delivered in hospitals1.31 (1.37)Children delivered by midwives1.38 (2.09)\n\nMost of the respondents reported that their husbands were the primary household decision makers followed by their in-laws and themselves. Most of the respondents reported no preference for the gender of their child, however, a high proportion preferred son over daughters by themselves and their in-laws. Most of the respondents did not report any preference of education based on gender of offspring. A high percentage of respondents reported having relationship problems with their husbands, experience of harassment as well as domestic abuse (Table\u00a0[2](#Tab2){ref-type=\"table\"}).Table\u00a02Response frequency to variables highlighting gender related disparities and preference for sonsFrequency (n)Percentage (%)\u03c7^2^ statisticP-valueHousehold decision maker\u00a0Self468.9228.50\\<\u2009.001\u00a0Husband32362.6\u00a0In-laws14728.5Ultrasound\u00a0Yes15329.7\u00a0No36370.3Gender on ultrasound\u00a0Male8455.6\u00a0Female6744.4Gender preference\u00a0Male17734.3213.71\\<\u2009.001\u00a0Female346.6\u00a0No preference30559.1Preference for education\u00a0Sons479.1717.11\\<\u2009.001\u00a0Daughters112.1\u00a0No preference45888.8In-laws prefer sons\u00a0Yes19437.631.75\\<\u2009.001\u00a0No32262.4Marital problems\u00a0Yes12724.6133.03\\<\u2009.001\u00a0No38975.4Harassment\u00a0Yes13926.9109.78\\<\u2009.001\u00a0No37773.1Domestic abuse\u00a0Yes9618.6203.44\\<\u2009.001\u00a0No42081.4\n\nExperience of harassment had significant association with an increasing number of daughters (r~pb~\u2009=\u2009.231, P\u2009\\<\u2009.001) self-induced abortions (r~pb~\u2009=\u2009.237, P\u2009\\<\u2009.001) and miscarriages (r~pb~\u2009=\u2009.151, P\u2009\\<\u2009.001). Experience of domestic abuse also exhibited significant association with an increasing number of daughters (r~pb~\u2009=\u2009.232, P\u2009\\<\u2009.001) self-induced abortions (r~pb~\u2009=\u2009.163, P\u2009\\<\u2009.001) and miscarriages (r~pb~\u2009=\u2009.182, P\u2009\\<\u2009.001).\n\nThe logistic regression model (backward method) predicted 71.5% of the model correctly. A non-significant Hosmer and Lemeshow test represented a good model fit, explaining 29% of variance in PSS-scores. According to it, high stress levels among pregnant women were associated with low-income status, non-involvement of the respondent in household decision making, desire for a son than a daughter or having no preference, experience of birth complications in past pregnancies by midwives, and experience of marital problems with husband and previous experiences of harassment (Table\u00a0[3](#Tab3){ref-type=\"table\"}).Table\u00a03Gender disparity and obstetric history as predictors of high stress levels among pregnant women (n\u2009=\u2009516)PredictorsSubcategoriesBStd. error in BP-valueOdds ratio95% C.I. for ORLowerUpperHousehold income.353.142.0131.4231.0771.879Household decision makerIn-laws.050Husband\u2212\u2009.952.434.028.386.165.904Self\u2212\u2009.444.241.066.641.4001.029PregnancyUnplanned/planned\u2212\u2009.797.246.001.450.278.730Number of children\u2212\u2009.129.061.033.879.780.989Gender preferenceNo preference.011Male.319.233.1721.375.8702.173Female\u2212\u20091.398.583.016.247.079.775Complications during birth procedure by mid-wivesYes/no.466.273.0881.594.9332.722Marital problemsYes/no.779.338.0212.1781.1244.223HarassmentNo/yes1.188.322.0003.2801.7456.164Constant\u2212\u2009.621.437.155.538Hosmer and Lemeshow test statistic\u2009=\u20093.54, P\u2009=\u2009.896, Cox and Snell R^2^\u2009=\u2009.22, NegelKerke R^2^\u2009=\u2009.29, Model Chi^2^\u2009=\u2009113.50, P\u2009\\<\u2009.001\n\nDiscussion {#Sec6}\n----------\n\nThe present study reports a very high prevalence of antenatal stress in Pakistan, which is substantiated by previous studies conducted in the region. A prevalence ranging from 29% to as high as 66% has been identified within Pakistan and neighboring countries, highlighting the high burden of antenatal stress in LMIC, moderated by region specific sociocultural stressors \\[[@CR1]--[@CR4]\\].\n\nOur study identifies potential household dynamics and sociocultural stressors that moderate levels of antenatal stress in Pakistan, including low household income, marital problems and experience of harassment. Our findings are corroborated by a number of previous studies conducted in the South Asian region \\[[@CR17], [@CR18]\\]. The lack of autonomy in household decision-making and healthcare was found to be a strong predictor of antenatal stress. Autonomy in health care is considered the right to make informed medical decisions about oneself. It becomes imperative during a pregnancy when the mother has to make decisions about herself and the fetus \\[[@CR19]\\]. Moreover, the lack of autonomy in running everyday\u00a0household affairs leads to a dependency on in-laws. Restrictions on autonomy, cultural\u00a0constraints,\u00a0and limitations in making household decisions are a pervasive problem in Pakistan, and hence a potential cause of antenatal stress \\[[@CR20]\\].\n\nIn our study sample, most women had undergone birth procedures in a hospital setting. However, almost fifty percent of our respondents had undergone birthing procedures by midwives, associated with a higher proportion of complications and antenatal stress than at hospitals. Most of these complications occurred at home because of birth procedures performed by low trained midwives called \"Dai\" in Urdu & Punjabi language. This finding further strengthens the evidence in the literature that an adverse outcome during previous pregnancy is a risk factor for prenatal stress \\[[@CR12]\\]. These traumatic memories can haunt a mother and can precipitate stress in subsequent pregnancies. Moreover, respondents undergoing abortions and miscarriages also reported episodes of harassment and domestic abuse, highlighting the cultural stigma revolving around genetic defects, abortions and poor health information seeking attitudes \\[[@CR21]\\]. This adds insult to the injury and further potentiates the psychological effects of these traumatic events \\[[@CR4]\\].\n\nOur results reveal that the number of pregnancies is a distinct indicator for antenatal stress. The exact reason for these findings is unknown but it could be related to less excitement and novelty after many earlier pregnancies. Moreover, Islam, the dominant religion\u00a0in the country, along with sociocultural beliefs discourage termination of pregnancies and contraception, translating to high birth rates in the region \\[[@CR4]\\]. With increasing number of children mothers may be anxious about problems related\u00a0to raising more children, poor maternal and child health outcomes and finances associated with childcare \\[[@CR22]\\]. Grand multigravidity (more than 5 children) is also a known independent predictor of abortion, causing excessive Antenatal stress among women \\[[@CR22]\\]. In addition, we found that unplanned pregnancies predicted high levels of stress among women. In general, poor knowledge and practice of methods of contraception among women and less autonomy to make decisions about family planning can lead to more unplanned pregnancies \\[[@CR22]\\]. An unintended pregnancy can cause an excessive psychological burden on expectant mothers as it demands access to better health care,\u00a0improved nutrition, physical rest and may cause financial stress on the family \\[[@CR23], [@CR24]\\].\n\nOur study also explored gender disparities and stigmas that revolve around pregnancy in our region. According to our analysis, preference for male children was associated with high stress levels in our study sample. This gender preference behavior has also been highlighted in previous studies conducted in Lahore and Karachi, Pakistan \\[[@CR4], [@CR25], [@CR26]\\]. While the expectation of in-laws for a male child is culturally accepted by a mother, it can put undue stress on her. In rural Pakistan, contrary to all scientific\u00a0knowledge, it is a commonly held belief that a woman's\u00a0egg, not the father's semen, determines the gender of the fetus \\[[@CR1]\\]. This notion sometimes leads to men practicing polygamy in the society. However, in our study, mother's equal interest in a child's education irrespective of the gender is a welcoming finding, in a country that is trying to close the gender parity gap in education. But discrimination still remains with women giving birth to more daughters, reporting harassment and domestic abuse at their homes \\[[@CR26], [@CR27]\\].\n\nLimitations {#Sec7}\n===========\n\nSelf-reporting nature of the questionnaires may introduce recall bias in the study. Moreover, due to sensitive nature of the information, respondents might have underreported information about domestic violence. Due to its cross-sectional design, causal relationships cannot be established between two variables. Despite a higher sample size and an inclusive study sample, the results of this investigation should not be generalized to the whole Pakistani population. In addition, experiences of domestic violence and harassment by expectant mothers were inquired using dichotomous questions, which may not be a sensitive measure of these constructs.\n\nLMIC\n\n:   Low and middle income countries\n\nPSQI-U\n\n:   Urdu version of Pittsburgh Sleep Quality Index\n\nPSS-4\n\n:   Perceived Stress Scale\n\n**Publisher\\'s Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nNot applicable.\n\nSN, MZ and AW conceived the study, collected data and performed statistical analysis. KKA, SZ and MHM drafted the manuscript and contributed to data analysis. AW and HM participated in the study design, helped to draft the manuscript and revised it critically. All authors read and approved the final manuscript.\n\nThere was no funding for this study.\n\nA confidentiality agreement with participants prevents us from sharing the data, therefore, dataset cannot be shared.\n\nAll participants gave written informed consent, and approval for the study was granted by the Research Ethics Review Committee of FMH College of Medicine and Dentistry, Lahore, Pakistan.\n\nNot applicable.\n\nThe authors report no declarations of interest. Dr. Ahmed Waqas serves BMC Research Notes as an associate editor, however, he was not involved in the peer review process for this manuscript.\n"}
+{"text": "Introduction {#s1}\n============\n\nHypoxia, commonly found in solid tumors ([@B1]), is considered to exert adverse effects on radiotherapy outcome, and shows an association with poor prognosis ([@B2]--[@B5]). Increased oxygen amounts, termed reoxygenation, is generally associated with radiation therapy. Reoxygenation is likely to play a positive role in increasing irradiation sensitivity, therefore improving tumor control ([@B6]). In order to achieve tumor reoxygenation, dose escalation in radiotherapy for hypoxic regions represents a common method. However, because radiation damages the surrounding non-cancerous tissues, the dose for a hypoxic tumor cannot exceed 200% of the standard one ([@B7]). Detection of reoxygenation and the degree of oxygen increase in tumors during fractionated radiotherapy is of great importance for dose adjustment in follow-up radiotherapy.\n\n^18^F-fluoromisonidazole positron emission tomography/computed tomography (^18^F-FMISO PET/CT) non-invasively evaluates tumor hypoxia. Interestingly, FMISO uptake correlates with the \"gold standard\" of tissue measurement using invasive polarographic oxygen electrodes ([@B8], [@B9]), as well as pimonidazole uptake in animals with glioma ([@B10]). Our previous and several other studies have reported decreased FMISO uptake in hypoxic tumors during radiotherapy and post-treatment, indicating that reoxygenation occurs ([@B11]--[@B16]). However, cancer cells killed by irradiation no longer show FMISO uptake; thus, it is hard to correctly determine whether reduced FMISO uptake in radiotherapy results from tumor reoxygenation or reflects tumoricidal effects. In a study by Okamoto et al. ([@B17]), tumoricidal effects during radiotherapy were evaluated by assessing tumor metabolism using ^18^F-fluorodeoxyglucose (FDG) PET. However, irradiation usually causes radioactive inflammation, which can also exhibit FDG uptake. Uptake of ^18^F -FDG PET, therefore, could represent either the residual tumor or inflammation.\n\nApart from tumor oxygenation, another most influential factor affecting fractionated radiotherapy is the potential changes of cell proliferation, which indicates tumor activity. Fortunately, cell proliferation can be assessed by ^18^F-fluorothymidine (FLT) PET during radiation treatment. Indeed, the development of FLT as a PET tracer has enabled *in vivo* demonstration of cell proliferation ([@B18]). *In vivo* analyses suggested that FLT has higher tumor specificity than FDG, and can distinguish tumor tissue from inflammation ([@B19], [@B20]). Tumor cells with low FLT and FMISO uptake levels are considered to be inactive and will undergo death. Meanwhile, those with high FLT and low FMISO levels are active with no hypoxia.\n\nIn the current study, using an experimental murine tumor model, we investigated tumor reoxygenation and tumor proliferation changes during radiotherapy with ^18^F-FMISO PET/CT and ^18^F-FLT PET/CT prior to, during, and following fractionated radiotherapy (FRT), with the aim to detect the relationship between tumor reoxygenation and tumoricidal effects during radiotherapy.\n\nMaterials and Methods {#s2}\n=====================\n\nEstablishment of Tumor Model\n----------------------------\n\nAll experimental studies were approved by the Institute of Anhui Medical University, and followed AAALAC and IACUC guidelines. The head and neck squamous carcinoma cell line (FaDu) was from the Anhui Medical University animal center. Four to five weeks old female BALB/c nude mice (20--25 g), underwent anesthesia with 1% isoflurane and received a subcutaneous injection of 5.0 \u00d7 10^6^ cells in 0.2 mL phosphate-buffered saline (PBS) into the right flank. Tumors of 6--7 mm in diameter (10 days after injection) were selected for experiments. Tumor diameters were measured every day, and gross tumor volume (GTV) was derived as: V (cm^3)^ = length (cm) \u00d7 width^2^ (cm^2^) \u00d7 0.5.\n\nIrradiation of Tumors\n---------------------\n\nTumor-bearing mice were divided into two groups: (i) control group (*n* = 5) did not receive any treatment; (ii) IR group (*n* = 16) was exposed to 3 Gy daily to a maximum dose of 40 Gy with a VARIAN 23 EX medical linear accelerator (Varian Medical Systems, USA). The tumor-bearing mice were lightly anesthetized with 1% isoflurane and placed in a circular irradiation jig. Then, the tumor-bearing legs were gently extended into the central part of the jig, taped, and the animals were covered with a 3-mm-thick lead sheet. The irradiation factors were 6 mV, 6/100 Varian linear accelerator at a dose rate of 200 mU/min.\n\nPET/CT Imaging\n--------------\n\nAll mice were scanned with both ^18^F-FMISO and ^18^F-FLT PET/CT prior to (Pre-FRT, 0 Gy), during (Inter-FRT, 21 Gy), and after FRT (Post-FRT, 40 Gy). When reached a dose of 21Gy, radiotherapy was break for 2 days for Inter-FRT imaging. ^18^F-FMISO and ^18^F-FLT PET/CT scan (Inveon, Siemens, Micro PET research center of shanghai Ruijin hospital) were conducted on 2 consecutive days. Both ^18^F-FMISO and ^18^F-FLT were provided by the molecular imaging center of Shanghai Xinhua hospital (China), with radiochemical purity above 95% to qualify for use in imaging. For PET/CT, one dose of 5.55 MBq (150 \u03bcCi) of ^18^F-FMISO or ^18^F-FLT was intravenously administered to each mouse via the tail vein. PET was performed 2 h following ^18^F-FMISO and 1 h upon ^18^F-FLT administration, and image reconstruction was carried out with the Inveon Acquisition Workplace software v2.0 (Siemens Preclinical Solutions). CT-guided percutaneous biopsy was performed for pathological study at Pre-FRT and Inter-FRT while all mice were sacrificed for pathological analysis at Post-FRT.\n\nImage Analysis\n--------------\n\nTo evaluate tumor hypoxia, maximum standardized uptake (SUVmax) and the tumor-to-normal muscle ratio (TNR) were used in this study for quantitative assessment of ^18^F-FMISO PET/CT data. SUVmax was determined by assessing the maximal radioactivity in a region of interest (ROI). The TNR was derived as the tumor\\'s SUVmax divided by the SUVmax of non-cancerous tissue. Hypoxia volume (HV) was the volume with TNR \u22651.25 ([@B21]). Tumor reoxygenation was analyzed according to TNR and HV changes. To assess tumor proliferation, we measured the tumor\\'s SUVmax and calculated the proliferation target volume (PTV) in the ^18^F-FLT PET/CT image. The PTV was the volume with SUVmax \\>1.4 as previously reported ([@B22]). In ^18^F-FMISO and ^18^F-FLT PET/CT images, reduction rates from Pre-FRT to Inter-FRT were derived as (uptake in Pre-FRT---uptake in Inter-FRT)/uptake in Pre-FRT.\n\nPathological Assays\n-------------------\n\nTumor specimens underwent formalin fixation, paraffin-embedding, and sectioning at 4 \u03bcm for immunohistochemical staining. Hypoxia (hypoxia-inducible factor- \\[HIF-\\] 1\u03b1 and carbonic anhydrase IX \\[CAIX\\]) and proliferation (the proliferation antigen Ki67 and proliferating cell nuclear antigen \\[PCNA\\]) markers were evaluated immunohistochemically. Antibodies raised against HIF-1\u03b1 (1/200; Novus Biologicals, USA), CAIX (1/400; Santa Cruz, USA), Ki67 (1/500; Zhongshan Jinqiao, China) and PCNA (1/400, Santa Cruz) were probed. HIF-1\u03b1, CAIX, Ki67 and PCNA levels were obtained by counting stained cells in 10 randomly selected high-power fields (\u00d7400), and graded as follows based on the percentages of cells stained: 0+, 0--10%; 1+, 10--25%; 2+, 25--50%; 3+, 50--75%; 4+, \\>75%.\n\nStatistical Analysis\n--------------------\n\nAll statistical analyses were performed with SPSS v17.0. Data are mean \u00b1 SD. Changes in SUVmax, TNR, HV, PTV and the expression levels of biomarkers Pre-FRT, Inter- FRT, and Post-FRT were analyzed by *t*-test. Associations of SUVmax, TNR and tumor volume with the above biomarkers were assessed by Pearson correlation analysis. *P* \\< 0.05 indicated statistical significance.\n\nResults {#s3}\n=======\n\nGross Tumor Volume (GTV) Change\n-------------------------------\n\nThe preliminary studies were performed to maintain animal welfare and tolerance to the tumor and irradiation. However, even under the daily 2 Gy dose, owing to the continual increase in tumor volume, many tumors grew to exceed 8 or 10 mm in diameter. Consequently, the tumor ulceration and infection became issues of concern, and the increase in the abundance of necrotic tissue in the tumor may affect the detection of hypoxic tissue. Therefore, we increased the daily dose as 3 Gy and started the tumor radiotherapy when the diameter of the lesion was 6--7 mm to prevent the tumors from reaching the 10-mm ulceration threshold.\n\nTen days following tumor cell administration, the 16 animal models in IR group assessed showed overt subcutaneous tumors on the right hind flank. The tumor average diameter was 6.8 \u00b1 0.4 mm pre-irradiation, corresponding to an average volume of 144 mm^3^. As shown in [Figure 1](#F1){ref-type=\"fig\"}, GTV increased during the first week of irradiation and began to decrease only after 18 Gy (6th day of radiotherapy). GTV increased on average by 24.3% from pre-irradiation (144 \u00b1 29 mm^3^) to a cumulative dose of 18 Gy (179 \u00b1 46 mm^3^), with a mean overall decrease of 45% by treatment end (50 Gy; 90 \u00b1 18 mm^3^). Conversely, owing to the continual increase of GTV in the control group, some tumors grew beyond 10 and 12 mm in diameter. Thus, some of the tumor mice in the control group were sacrificed early for animal welfare.\n\n![Changes in GTV during radiotherapy (maximum dose 40 Gy) (*n* = 16).](fonc-10-01046-g0001){#F1}\n\nPET/CT Data\n-----------\n\nAs shown in [Table 1](#T1){ref-type=\"table\"}, both ^18^F-FMISO-SUVmax and ^18^F-FMISO-TNR were markedly reduced from Pre-FRT to Post-FRT in the IR group. In addition, the reduced rates of SUVmax and TNR from Pre-FRT to Inter-FRT (30.3 and 27.9%, respectively) were significantly higher than those from Inter-FRT to Post-FRT (14.5 and 18.3%, respectively). For ^18^F-FLT PET/CT, ^18^F-FLT-SUVmax was significantly and continually decreased from Pre-FRT to Post-FRT; however, the reduction rate from Pre-FRT to Inter-FRT (36.8%) was similar to that from Inter-FRT to Post-FRT (51.0%; *p* = 0.087). Both HV and PTV were significantly decreased from Inter-FRT to Post-FRT. The mean reduction rate of HV from Pre-FRT to Inter-FRT (85%) was significantly elevated as compared to that from Inter-FRT to Post-FRT (71.4%), while the PTV reduction rate from Pre-FRT to Inter-FRT (21.2%) was remarkably lower as compared to that from Inter-FRT to Post-FRT (82.7%) ([Figures 2](#F2){ref-type=\"fig\"}, [3](#F3){ref-type=\"fig\"}). Conversely, all the above PET parameters in the control group showed a significant continual increase during the study, which was due to the increased oxygen depletion and rapid tumor proliferation.\n\n###### \n\nUptake changes detected by ^18^F-FMISO and ^18^F-FLT PET/CT at Pre-FRT, Inter-FRT, and post-FRT.\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                            **Pre-FRT**      **Inter-FRT (*p*-value)**   **Post-FRT (*p*-value)**   **Reduction rate from pre- to inter-FRT**   **Reduction rate from inter- to post-FRT (*p*-value)**\n  ----------------------------------------- ---------------- --------------------------- -------------------------- ------------------------------------------- --------------------------------------------------------\n  **IR group (*****n*** **=** **16)**                                                                                                                           \n\n  ^18^F-FMISO-SUVmax                        2.67 \u00b1 0.78      1.86 \u00b1 0.41\\                1.59 \u00b1 0.39\\               30.3 \u00b1 21.6%                                14.5 \u00b1 10.0%\\\n                                                             (*p* \\< 0.001)              (*p* = 0.008)                                                          (*p* = 0.012)\n\n  ^18^F-FMISO-TNR                           1.97 \u00b1 0.40      1.42 \u00b1 0.23\\                1.16 \u00b1 0.11\\               27.9 \u00b1 18.2%                                18.3 \u00b1 9.8%\\\n                                                             (*p* \\< 0.001)              (*p* = 0.004)                                                          (*p* = 0.032)\n\n  ^18^F-FLT-SUVmax                          7.01 \u00b1 3.10      4.43 \u00b1 2.45\\                1.80 \u00b1 1.37\\               36.8 \u00b1 28.22%                               51.0 \u00b1 38.2%\\\n                                                             (*p* = 0.032)               (*p* = 0.017)                                                          (*p* = 0.087)\n\n  ^18^F-FMISO-HV\\                           74.53 \u00b1 18.20    11.20 \u00b1 6.79\\               3.27 \u00b1 4.12\\               85.0 \u00b1 28.2%                                71.4 \u00b1 26.3%\\\n  (mm^3^)                                                    (*p* = 0.031)               (*p* = 0.025)                                                          (*p* = 0.047)\n\n  ^18^F-FLT-PTV\\                            132.00 \u00b1 27.00   104.00 \u00b1 21.00\\             18.40 \u00b1 9.70\\              21.2 \u00b1 14.2%                                82.7 \u00b1 48.4%\\\n  (mm^3^)                                                    (*p* = 0.041)               (*p* \\< 0.001)                                                         (*p* = 0.012)\n\n  **Control group (*****n*** **=** **5)**                                                                                                                       \n\n  ^18^F-FMISO-SUVmax                        2.31 \u00b1 0.45      2.66 \u00b1 0.78\\                2.98 \u00b1 0.56\\               \u221215.2 \u00b1 11.1%                               \u221212.0 \u00b1 7.4%\\\n                                                             (*p* \\< 0.001)              (*p* \\< 0.001)                                                         (*p* = 0.034)\n\n  ^18^F-FMISO-TNR                           1.62 \u00b1 0.31      1.89 \u00b1 0.43\\                2.12 \u00b1 0.46\\               \u221216.7 \u00b1 12.1%                               \u221212.2 \u00b1 6.7%\\\n                                                             (*p* \\< 0.001)              (*p* \\< 0.001)                                                         (*p* = 0.041)\n\n  ^18^F-FLT-SUVmax                          6.35 \u00b1 3.15      9.78 \u00b1 4.76\\                11.79 \u00b1 5.37\\              \u221254.0 \u00b1 21.43%                              \u221220.6 \u00b1 13.2%\\\n                                                             (*p* \\< 0.001)              (*p* \\< 0.001)                                                         (*p* \\< 0.001)\n\n  ^18^F-FMISO-HV\\                           68.21 \u00b1 15.40    81.12 \u00b1 18.77\\              99.33 \u00b1 24.52\\             \u221218.9 \u00b1 10.77%                              \u221222.4 \u00b1 14.3%\\\n  (mm^3^)                                                    (*p* \\< 0.001)              (*p* \\< 0.001)                                                         (*p* = 0.004)\n\n  ^18^F-FLT-PTV\\                            137.00 \u00b1 32.00   198.00 \u00b1 57.45\\             252.11 \u00b1 71.46\\            \u221244.5 \u00b1 25.1%                               \u221227.3 \u00b1 17.3%\\\n  (mm^3^)                                                    (*p* \\< 0.001)              (*p* \\< 0.001)                                                         (*p* \\< 0.001)\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n![^18^F-FMISO and ^18^F-FLT-based micro-positron emission tomography (PET) detecting FaDu xenografts. White arrows indicate tumor location (Pre, Inter and Post represent data before (0 Gy), during (21 Gy), and after (40 Gy) radiotherapy, respectively). In ^18^F-FMISO PET images, the TNR, SUVmax, and HV were significantly decreased from Pre-FRT to Inter-FRT (*p* \\< 0.001, *p* \\< 0.001, *p* = 0.031, respectively) and from Inter-FRT to Post-FRT (*p* = 0.004, *p* = 0.008, *p* = 0.025, respectively). In ^18^F-FLT PET images, SUVmax and PTV were significantly decreased from Pre-FRT to Inter-FRT (*p* = 0.032, *p* = 0.041, respectively) and from Inter-FRT to Post-FRT (*p* = 0.017, *p* \\< 0.001, respectively).](fonc-10-01046-g0002){#F2}\n\n![TNR, SUVmax, HV and PTV changes detected by ^18^F-FMISO PET/CT and ^18^F-FLT PET/CT, respectively. **(A)** Changes of TNR (FMISO) detected by ^18^F-FMISO PET/CT during radiotherapy. **(B)** Changes of SUVmax detected by ^18^F-FMISO PET/CT during radiotherapy. **(C)** Changes of SUVmax detected by ^18^F-FLT PET/CT during radiotherapy. **(D)** Changes of HV detected by ^18^F-FMISO PET/CT during radiotherapy. **(E)** Changes of PTV detected by ^18^F-FLT PET/CT during radiotherapy. The reduction in all parameters was significant. \\**P* \\< 0.05, \\*\\**P* \\< 0.01, \\*\\*\\**P* \\< 0.01 by Student\\'s *t*-test (*n* = 16). TNR, tumor-to-normal muscle ratio of maximum radiotracer uptake; HV, hypoxic volume; PTV, proliferative tumor volume; FRT, fractionated radiotherapy.](fonc-10-01046-g0003){#F3}\n\nOf all the tumor-bearing mice in the IR group, five tumors showed hypoxia during Inter-FRT FMISO PET, which persisted Post-FRT in one mouse. Interestingly, decreased HV could render these lesions highly resistant to radiotherapy. Furthermore, nine tumor-bearing mice showed strong proliferation at Inter-FRT FLT PET, and the positive uptake persisted Post-FRT in four mice.\n\nExpression Levels of Hypoxia and Proliferation Markers\n------------------------------------------------------\n\nPathological examination was performed Pre (at 0 Gy), Inter (at 21 Gy) and Post (at 40 Gy) radiotherapy ([Figure 4](#F4){ref-type=\"fig\"}). Immunohistochemical assays were carried out to evaluate tumor hypoxia (HIF-1\u03b1, CAIX) and proliferation (Ki67, PCNA) markers. There were 77.63 \u00b1 15.96, 32.76 \u00b1 8.47 and 11.33 \u00b1 5.71% cells with positive HIF-1\u03b1 staining at Pre-FRT, Inter-FRT and Post-FRT, respectively. Cells expressing CAIX at Pre-FRT, Inter-FRT and Post-FRT accounted for 82.56 \u00b1 10.33, 37.32 \u00b1 7.28, and 13.12 \u00b1 6.12%, respectively. There were 77.54 \u00b1 16.42, 48.76 \u00b1 7.24, and 8.31 \u00b1 6.12% cells expressing Ki67 at Pre-FRT, Inter-FRT and Post-FRT, respectively. PCNA at Pre-FRT, Inter-FRT and Post-FRT was expressed by 72.96 \u00b1 18.49, 43.11 \u00b1 6.77 and 10.12 \u00b1 5.82% cells, respectively. These tumor biomarkers showed significant differences between Pre-FRT and Inter-FRT (*p* \\< 0.05), and between Inter-FRT and Post-FRT (*p* \\< 0.05). These findings suggested hypoxia and proliferation markers were continuously down-regulated during radiotherapy, which was consistent with PET imaging data.\n\n![Immunohistochemical assessment of multiple tumor biomarkers before, during and after radiotherapy. All the indicated biomarkers in tumors were continuously downregulated during irradiation. *P* \\< 0.05 by Student\\'s *t*-test (*n* = 16). HIF-1\u03b1, hypoxia-inducible factor; CAIX, carbonic anhydrase 9; Ki67, proliferation antigen; PCNA, proliferating cell nuclear antigen.](fonc-10-01046-g0004){#F4}\n\nAssociations of Uptake of Radiotracers With Tumor Volume (HV, PTV and GTV)\n--------------------------------------------------------------------------\n\nBoth SUVmax (FMISO) and TNR (FMISO) were significantly associated with HV (*r* = 0.686, *p* = 0.024 and *r* = 0.763, *p* = 0.016, respectively), but showed no associations with GTV (*r* = 0.325, *p* = 0.178 and *r* = 0.145, *p* = 0.216, respectively). SUVmax (FLT) was significantly correlated with PTV (*r* = 0.842, *p* = 0.009), but showed no correlation with GTV (*r* = 0.211, *p* = 1.435) ([Table 2](#T2){ref-type=\"table\"}).\n\n###### \n\nCorrelation analysis between PET tracers and tumor biological parameters.\n\n                                         **HV**                                   **PTV**                                    **GTV**   **HIF-1\u03b1**                               **CAIX**                                   **Ki67**                                   **PCNA**\n  ---------------- --------------------- ---------------------------------------- ------------------------------------------ --------- ---------------------------------------- ------------------------------------------ ------------------------------------------ ------------------------------------------\n  SUVmax (FMISO)   Pearson correlation   0.686[^\\*^](#TN1){ref-type=\"table-fn\"}   0.215                                      0.325     0.418[^\\*^](#TN1){ref-type=\"table-fn\"}   0.389[^\\*^](#TN1){ref-type=\"table-fn\"}     0.237                                      0.112\n                   Sig(2-tailed)         0.024                                    0.468                                      0.178     0.041                                    0.037                                      0.745                                      0.253\n                   N                     16                                       16                                         16        16                                       16                                         16                                         16\n  TNR (FMISO)      Pearson correlation   0.763[^\\*^](#TN1){ref-type=\"table-fn\"}   0.341                                      0.145     0.692[^\\*^](#TN1){ref-type=\"table-fn\"}   0.801[^\\*\\*^](#TN2){ref-type=\"table-fn\"}   0.412                                      0.215\n                   Sig (2-tailed)        0.016                                    0.842                                      0.216     0.015                                    0.006                                      0.341                                      0.547\n                   N                     16                                       16                                         16        16                                       16                                         16                                         16\n  SUVmax (FLT)     Pearson correlation   0.308                                    0.842[^\\*\\*^](#TN2){ref-type=\"table-fn\"}   0.211     0.304                                    0.523                                      0.792[^\\*\\*^](#TN2){ref-type=\"table-fn\"}   0.837[^\\*\\*^](#TN2){ref-type=\"table-fn\"}\n                   Sig (2-tailed)        0.264                                    0.009                                      1.435     0.271                                    0.222                                      0.003                                      0.004\n                   N                     16                                       16                                         16        16                                       16                                         16                                         16\n\nP \\< 0.05;\n\n*P \\< 0.01*.\n\nAssociations of Uptake of Radiotracers With Tumor Markers\n---------------------------------------------------------\n\nTNR (FMISO) was significantly correlated with hypoxia markers, including HIF-1\u03b1 and CAIX (*r* = 0.692, *p* = 0.015 and *r* = 0.801, *p* = 0.006, respectively). Meanwhile, weak correlations were found between SUVmax (FMISO) and hypoxia markers (*r* = 0.418, *p* = 0.041 and *r* = 0.389, *p* = 0.037, respectively). No significant associations were detected of FMISO uptake with proliferation markers. SUVmax (FLT) was significantly correlated with Ki67 and PCNA (*r* = 0.792, *p* = 0.003 and *r* = 0.837, *p* = 0.004, respectively). SUVmax (FLT) was not associated with hypoxia markers ([Table 2](#T2){ref-type=\"table\"}).\n\nDiscussion {#s4}\n==========\n\nRadiotherapy has been used extensively to treat cancer patients because it damages most solid tumors without penetration limits. It is well-known that the \"four R\\'s\" (repair of DNA damage, redistribution of cells in the cell cycle, repopulation, and reoxygenation of hypoxic tumor areas) play a role in radiotherapy. In this study, we mainly explored reoxygenation and tumor proliferation during radiotherapy. Among the \"four R\\'s,\" reoxygenation may be the most important in fractionated radiotherapy ([@B23]). A hypoxic microenvironment in solid tumors leads to severe radioresistance, which results in poor efficacy of radiotherapy. The maximal benefit of fractionation against tumors is attributed to the reoxygenation of surviving hypoxic cells before the next fraction. The assessment of tumor hypoxia during treatment discriminates between radiosensitive and radioresistant tumors, and hence, the application of doses in different areas can be optimized to maximize the tumor damage and protect normal tissue.\n\nIn the current study with head and neck squamous carcinoma (FaDu) xenograft model, serial ^18^F-FMISO and ^18^F-FLT PET/CT scans showed tumor reoxygenation and gradual decline in tumor cell proliferation during radiotherapy. FMISO uptake was reduced earlier than FLT uptake during radiotherapy. The reduction rates of FMISO uptake and HV at the beginning of radiotherapy (from 0 to 21 Gy) were significantly higher than those observed in the late phase of treatment (from 21 to 40 Gy). However, the decrease of FLT uptake gradually with the course of radiotherapy. Compared with the significant reduction of hypoxia, SUVmax (FLT) and PTV were still at relatively high levels at Inter-FRT. These findings support our previous hypothesis that \"the tumor is still active, but not hypoxic.\" In case such assumption holds, the current data suggest that the amounts of hypoxic cells are reduced earlier than those of active cells during fractionated radiation treatment. In other words, reduction in FMISO uptake in the early phase indicates tumor reoxygenation rather than tumoricidal effects.\n\nIn this study, HV was significantly decreased in Inter-FRT ^18^F-FMISO PET/CT at 21 Gy; meanwhile, PTV reduction in the early phase of treatment was markedly less pronounced than that of HV. It is well-known that tumor cells with hypoxia are more radiation-resistant than oxygenated counterparts ([@B24], [@B25]). Thus, hypoxic cancer cells are not likely killed earlier by irradiation compared with non-hypoxic ones. More likely, cells with elevated FMISO uptake in the HV pre-radiotherapy turned into cells with reduced FMISO uptake. This corroborated a previous study of patients with primary head-and-neck cancer by Okamoto et al. ([@B17]).\n\nIn addition, the FMISO uptake is continued in five tumor-bearing mice (31.2%, 5/16) at Inter-FRT and one tumor-bearing mouse (6.2%, 1/16) at Post-FRT. FMISO PET during radiotherapy may contribute to the detection of such highly radioresistant tumors. Clinically, this phenomenon indicated that additional chemotherapy or molecular targeted therapy might be needed for these patients.\n\nReoxygenation occurred early after radiotherapy initiation as shown above; this is consistent with previous reports assessing other xenograft tumor cell lines or *in vivo* tumors ([@B5], [@B26]). However, Harriss et al. ([@B6]) reported that reoxygenation occurs until near the end of irradiation treatment, based on pO~2~ measurements using the OxyLab probe. There are many possibilities which could explain such discrepancy. Tumor oxygenation depends on multiple individual features such as the amounts of cells under hypoxia, the extent of oxygen elevation, and the timing/progression of such effect during treatment ([@B27], [@B28]). Tumor type, volume and vascularization, as well as the structure and pressure of neighboring tissues are influential factors for reoxygenation initiation ([@B6]). As a consequence, pre-treatment evaluation of tumors is critical if individualized radiotherapy is to be implemented.\n\nMice underwent exposure to 3 Gy daily in this study. As shown in [Figure 1](#F1){ref-type=\"fig\"}, GTV was increased during the first week of radiotherapy, and began to decrease only after a cumulative dose of 18 Gy. In contrast, HV and PTV decreased gradually with increasing radiotherapy time. These findings indicate that tumor volume has no sensitivity for early efficacy assessment. Gross tumor volume is influenced by various factors such as fibrosis components, infiltration of inflammatory cells and the rate at which apoptotic cancer cells are cleared ([@B29]). Thus, functional imaging should be used for early efficacy evaluation in clinic.\n\nIn order to detect changes of the microenvironment in tumor cells during radiotherapy, immunohistochemical assays were performed at Pre-FRT, Inter-FRT, and Post-FRT. Biomarkers specific to cancer are frequently employed for monitoring tumor occurrence and progression. It is admitted that the adaptation of tumor cells to the hypoxic microenvironment is regulated by HIF-1a, which is considered an independent biomarker of response to hypoxia ([@B30]). HIF-1a activation promotes the expression of CAIX, which controls intra- and extra-cellular pH homeostasis in hypoxic conditions ([@B31]). Overexpression of CAIX is correlated with poor survival/prognosis in solid tumors ([@B32], [@B33]). In a study by Chen et al. elevated CAIX amounts were shown to promote tumor progression ([@B34]). Furthermore, the proliferation antigen Ki67 is closely associated with cell mitosis, playing a vital role in the initiation of cell proliferation ([@B35]). Proliferating cell nuclear antigen (PCNA) is an important trigger of multiple events in DNA replication, repair and recombination in eukaryotes and archaea, thus regulating cell proliferation ([@B36]). In this study, all these biomarkers were progressively down-regulated during radiotherapy, corroborating PET/CT imaging data. Moreover, statistical analysis showed that PET/CT imaging parameters had close associations with hypoxia and proliferation markers, further supporting the feasibility and effectiveness of non-invasive PET/CT monitoring in tumor radiotherapy.\n\nNew therapeutic strategies, especially dose escalation to the hypoxic area, is expected to improve the prognosis of patients. Thus, the evaluation of dynamic changes in intratumoral hypoxic and proliferative states by a dual-tracer study illuminates the understanding of tumor biology after radiotherapy, thereby facilitating its planning, including dose escalation and altered fraction schedules. Dose escalation requires an accurate determination of the hypoxic area in order to target the hypoxia. New developed PET scanners with high spatial resolution and the new generation of hypoxic tracers with lower lipid solubility \\[for example, ^18^F-fluoroazomycin arabinoside (^18^F-FAZA)\\] would have the potential to accurately demarcate the hypoxic area with high contrast ([@B37]).\n\nThis study was limited by its small sample size. In addition, the survival rate and prognosis of reoxygenated tumors after radiotherapy were not assessed. Additional research is needed to confirm these findings, especially studies involving patients.\n\nConclusion {#s5}\n==========\n\nReoxygenation of hypoxic tumors occurs early after radiotherapy initiation. In contrast, cell proliferation inhibition associated with tumoricidal effects takes place gradually with the course of radiotherapy. Finally, ^18^F-FMISO and ^18^F-FLT PET/CT are sensitive and non-invasive tools for monitoring tumor reoxygenation and activity during fractionated radiotherapy.\n\nData Availability Statement {#s6}\n===========================\n\nAll datasets generated for this study are included in the article/supplementary material.\n\nEthics Statement {#s7}\n================\n\nThe animal studies were reviewed and approved by the Institute of Anhui Medical University, and followed AAALAC and IACUC guidelines.\n\nAuthor Contributions {#s8}\n====================\n\nWY and HX designed the experiment and wrote the paper. WY, XS, DZ, and FQ performed the experiment and statistical analysis of data. HW and JJ contribute technical help. All authors have read and approved this version of the paper.\n\nConflict of Interest {#s9}\n====================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This work was supported by the National Natural Science Foundation of China (Nos. 81901779 and 81971643).\n\n^18^F-FLT\n\n:   ^18^F-fluorothymidine\n\n^18^F-FMISO\n\n:   F-fluoromisonidazole\n\nPET/CT\n\n:   positron emission tomography/computed tomography.\n\n[^1]: Edited by: Ester Orlandi, Istituto Nazionale dei Tumori (IRCCS), Italy\n\n[^2]: Reviewed by: Tao Xu, First People\\'s Hospital of Foshan, China; Alessandro Cicchetti, Istituto Nazionale dei Tumori (IRCCS), Italy\n\n[^3]: This article was submitted to Radiation Oncology, a section of the journal Frontiers in Oncology\n"}
+{"text": "Background\n==========\n\nRandom-sequence peptide library screening approaches represent an increasingly popular and powerful tool for identifying binding partners for antibodies and other proteins as well as carbohydrates, pharmaceuticals, and other small molecules. Peptide library methods generally fall into two categories: molecular display approaches such as phage display, and immobilized arrays such as SPOT. Display approaches can typically accommodate much larger libraries, but information is typically obtained only on the clones that survive several rounds of panning, resulting in a population that is heavily biased in favor of clones whose sequences facilitate growth \\[[@B1]\\]. In contrast, array based approaches may be used to screen smaller libraries with higher throughput than display approaches and semi-quantitative binding information is obtained on all of the peptides in the library. New technologies both on the display side and the array approach promise to overcome these limitations \\[[@B2]-[@B4]\\]. The decreasing cost of both sequencing and peptide synthesis as well as applications such as profiling the humoral immune response \\[[@B5]\\] promise to increase interest in connecting random-sequence peptide mimotopes to protein sequences occurring in nature. Therefore, an increase in the demand for appropriate algorithms and software to facilitate the data analysis would also be expected.\n\nWhile the peptides discovered in these library screening experiments serve as useful ligands in and of themselves, comparison of these sequences to natural protein sequences can reveal novel biological insight. Peptides selected by panning phage display libraries against monoclonal antibodies often closely match the antibody epitope making the sequence comparison rather straightforward \\[[@B6]\\]. If a strong enough motif is uncovered among the peptide sequences, it may even be used to search a database to predict an antibody target \\[[@B7]\\]. Though current array technology does not allow sufficient coverage of sequence space to contain sequences closely resembling natural protein sequences by chance, we have shown that experiments of this type still have utility for predicting monoclonal epitopes \\[[@B8]\\]. Other groups have shown that peptides selected to bind to other types of proteins have utility in understanding and predicting binding to natural binding partners \\[[@B9]-[@B11]\\]. Even small molecule binding peptides provide insight on their binding to natural proteins \\[[@B3],[@B12]\\].\n\nAnalysis of the peptide sequences obtained from any selection experiment poses two key challenges. First, a set of peptides need to be compared against a protein database. Second, an appropriate scoring scheme is needed to search for structural similarity rather than evolutionary relationships. At first glance, the FASTS/FASTF programs appear to address the first challenge, as they are designed to take peptide sequences generated from protein sequencing techniques and identify homologous proteins \\[[@B13]\\]. However, the FASTS/FASTF programs search for cases where peptides align to non-overlapping regions of the protein sequence, while we would like to identify regions where the peptides align to the same region of the protein sequence. Another approach is to identify a motif among the selected peptide sequences and use the consensus sequence or a probabilistic representation of the motif to compare to the protein sequence(s) of interest \\[[@B14]\\]. We previously demonstrated that the glam2 motif finding program is suitable for analyzing random-sequence peptide data \\[[@B8],[@B15]\\]. While the motif approach may be powerful in many cases, the peptides of interest may not always have a common pattern because different amino acids may match in the same region of the sequence, or peptides may align to different parts of the protein sequence(s). Another approach would be to align each discovered peptide sequence to the protein sequence targets and sum the alignment scores at each position. The RELIC MATCH program (not currently available or supported) used this approach with some success \\[[@B3],[@B9],[@B10],[@B12],[@B16]\\]. This program also had several limitations with regards to transparency, flexibility, statistical analysis, and the ability to search multiple sequences. Here we present an open source application that gives the user access to all parameters, can empirically estimate the statistical significance of the results, and enables the analysis of many sequences at once.\n\nMethods\n=======\n\nAlgorithm overview\n------------------\n\nThe user inputs protein sequence(s) to search, a set of selected peptides, and (optionally) a representative or complete list of peptides from the library. A scoring matrix may be generated by the program as described below or entered by the user. The maximal local alignment between each selected peptide and protein sequence is found. If the alignment score is greater than the user defined score threshold, the score at each protein residue position is added to the protein residue scores. If the moving average window size is set to greater than one, after all peptides have been aligned to a given protein, the moving average across the protein residue positions is calculated and the residue scores provided correspond to the score at the start of the window. The same number of peptides as in the selected list are randomly selected from the library if a library set was entered, and these are aligned to the protein(s) in the same manner as for the selected peptides; this process is repeated for the specified number of sampling iterations. If the subtract library scores box is checked, the average scores at each residue position from the randomly selected peptides from the library are subtracted from the residue scores. The selected peptide scores across each protein sequence are graphed, as well as the maximum and average scores from the random sampling iterations. The user may use the sort button to order the proteins by their maximal residue scores. The text output tab may be used to view a summary table of the maximum alignment scores for each protein or a table of all of the alignments identified for the number of proteins specified.\n\nScoring matrix\n--------------\n\nGuiTope generates a log-odds-like scoring matrix based on a given measure of amino acid distances and amino acid frequencies. The distance matrix is taken to be inversely proportional to the frequencies of an amino acid pair appearing in a true alignment after a pseudocount of 10% of the average distance is added to the distance matrix to avoid dividing by zero. The rows and columns are iteratively scaled to sum to the expected amino acid frequencies. This matrix is then divided by the product of protein and peptide amino acid frequencies at each position and log~10~transformed.\n\nAlignment algorithm and inversion scoring\n-----------------------------------------\n\nThe maximal gapless local alignment of each peptide with each protein is calculated using the Smith-Waterman algorithm. If the inversion weight is set to greater than 0, the program will identify sequence positions where the protein residue at position i is the same as the peptide residue at position j +1 AND the protein residue at position i+1 is the same as the peptide residue at position j. The residue scores for these inversions will be the product of the inversion weight and the average of the identity scores for the amino acids at the protein positions i and i+1.\n\nStatistical analysis\n--------------------\n\nFor each sampling iteration and each protein sequence, a set of peptides, with the same number of peptides as the selected peptide list, is randomly selected from the library and the residue scores are calculated. From these, the maximum and average residue scores are calculated for each position. If the \\'subtract library scores\\' option is selected, the average library scores are subtracted from the residue scores from each iteration. The maximum scores from each protein iteration are ranked. For each protein, the maximum residue score from the selected peptides is compared to the ranked scores. The percentage of library scores that are higher than the selected peptide score is reported as the significance.\n\nEvaluation datasets\n-------------------\n\nA dataset was previously described containing lists of peptide sequences identified from random-sequence peptide microarray experiments as binding to monoclonal antibodies with known epitopes \\[[@B8]\\]. This dataset was used to optimize Guitope\\'s alignment parameters. A polyclonal anti-peptide dataset from the same publication was used to evaluate the algorithm. Additionally, another set of monoclonal antibodies with known epitopes was used to probe a completely different set of 10,000 random-sequence peptides on a microarray. The two anti-P53 antibodies from the first monoclonal antibody dataset were repeated on both the first and second version of the 10,000 peptide microarrays. Additionally, an anti-cMyc clone 9E10 (AbD SeroTec, Raleigh), anti-Leu-Enkaphalin clone 1193/220 (AbD SeroTec, Raleigh), anti-PBEF clone E10 (Santa Cruz Biotechnology), and anti-V5 (AbD SeroTec, Raleigh) were used to probe the array and generate lists of peptides to which the antibodies bound. Anti-cMyc, anti-Leu-Enkaphalin, and anti-V5 recognize epitope tags, while the anti-PBEF was epitope mapped using tiling peptides (current authors, manuscript in preparation). Phage display datasets that identified the greatest number of unique peptides were selected from those listed in the \\\"several binding sites\\\" category in Derda *et al*. \\[[@B1]\\] and these were downloaded from MimoDB <http://immunet.cn/mimodb/>. These phage display datasets include peptides selected against a diverse set of targets, including two human extracellular proteins, one bacterial protein, and immune sera to a virus and a bacterium.\n\nImplementation\n--------------\n\nGuiTope was implemented in Visual Basic, using the Microsoft .NET framework. It may be installed on any computer running Microsoft Windows XP or a newer Windows operating system. It has a memory footprint of 400 MB and will take anywhere between seconds to several minutes to run a set of hundreds of peptides against a single protein with 100 sampling iterations on a single Pentium 4 core, 3.2 GHz and 2 GB RAM machine running Windows XP. On the same hardware, searching a protein database of \\~20,000 proteins with a set of several hundred peptides with a single sampling iteration, will utilize \\< 3 GB of memory and use approximately 20 hours of direct CPU time.\n\nResults and discussion\n======================\n\nThe optimal combination of parameters for GuiTope was determined by testing on a previously described dataset of peptide sequences bound by monoclonal antibodies with known epitopes that had been used to probe a random-sequence peptide array. \\[[@B8]\\] Epitope predictions were evaluated using ROC analysis and the AUROC scores are reported in Figure [1](#F1){ref-type=\"fig\"}. The most critical parameter appears to be the scoring matrix, with the BLOSUM62 matrix having an AUROC 0.15 less than the GuiTope method which adjusts for altered amino acid frequencies. The di-peptide inversion method also had a substantial improvement in the AUROC score. The di-peptide inversion method is a novel alignment approach that we developed after observing such alignments in our data. We hypothesize that the flexibility of the peptides enables the inverted amino acids to have similar interactions with the paratope and we have found some preliminary experimental and modeling evidence supporting the di-peptide inversion (data not shown). Here we have included results from analysis with and without di-peptide inversions since the approach is unusual. The library subtraction method only yields a small improvement to the score and a large number sampling iterations are required to accurately estimate the average library score, so we only used library subtraction for evaluating individual proteins rather than for database searches in order to keep run times reasonable.\n\n![**Parameter Optimization**. The AUROC (Area Under the Receiver Operator Characteristics Curve) is shown for each parameter value tested on the 1^st^Known Epitope Monoclonal Dataset shown below. The best parameter value was highlighted and that value was used when each other value was varied.](1471-2105-13-1-1){#F1}\n\nGuiTope was tested on two independent datasets obtained by probing random-sequence peptide microarrays with antibodies. The first was obtained by probing an array of 10,000 random-sequence peptides, having completely different sequences than those used in the training set, with monoclonal antibodies having known linear epitopes. The monoclonal epitopes were predicted with an AUROC score of 0.75 using the inversion method and 0.78 without inversions (Figure [2A](#F2){ref-type=\"fig\"}). It appears that this dataset is more difficult to predict as the RELIC and glam2 methods also perform worse. The second peptide microarray evaluation dataset was generated from polyclonal anti-peptide sera. Here GuiTope performs similarly to previously tested methods with an AUROC of 0.68 using the inversion method and 0.56 without inversions, compared to an AUROC of 0.48 using RELIC method and 0.68 using Glam2 (Figure [2B](#F2){ref-type=\"fig\"}). These microarray datasets are likely considerably more difficult than phage display datasets because of sparse sampling of sequence space.\n\n![**Peptide Microarray Evaluation Datasets**. Peptides selected to bind known epitope monoclonals (A) or anti-peptide polyclonal sera (B) were used to predict the epitope (A) or immunizing peptide (B) in GuiTope, RELIC, or Glam2 within a database of decoy sequences. The significance scores of the true epitope or immunizing peptide sequences was compared to the decoy sequences using ROC plots, where the true positive rate is plotted against the false positive rate for all possible score thresholds. The results using the inversion weight as one are plotted in blue, the results without inversions are plotted in red, the results for RELIC are plotted in green, and the results for Glam2 are plotted in black. The AUROC value shown in the legend indicates the probability that a true sequence would score higher than a decoy sequence for that dataset. \\*Note that the RELIC analysis of the monoclonal set is only based on five monoclonal antibodies because the sixth antibody was run after the server was no longer available.](1471-2105-13-1-2){#F2}\n\nPhage display datasets evaluated in GuiTope were selected based on the summary of the MimoDB published in Derda *et al*. \\[[@B1]\\]. Two of these datasets consisted of peptides selected to polyclonal sera. The phage display peptides selected against the anti-Nipah virus were used to map three epitopes on the nucleoprotein, and GuiTope also identified these epitope regions (Figure [3A](#F3){ref-type=\"fig\"}). GuiTope also predicted an epitope on Glycoprotein G that was also predicted by DiscoTope \\[[@B17]\\], which uses the crystal structure to identify accessible regions (Figure [3B](#F3){ref-type=\"fig\"}). Yang *et al*. identified some regions of sequence similarity between the anti-*Mycoplasma hyopneumoniae*selected peptides and several *M. hyopneumoniae*protein sequences, but did not test whether their epitope predictions were correct \\[[@B18]\\]. The only experimentally determined *Mycoplasma hyopneumoniae*B-cell epitopes in the Immune Epitope Database \\[[@B19]\\] were determined by a peptide tiling study of predicted lipoproteins \\[[@B20]\\]. None of these epitopes were predicted by the Yang *et al*. or the GuiTope analysis. Most likely the phage display selected peptides correspond to epitopes on proteins other than the lipoproteins. There is no structural or experimental data to evaluate GuiTope\\'s predictions of the *Mycoplasma hyopneumoniae*epitopes.\n\n![**Analysis of Anti-Nipah Dataset**. A. Screen shot of GuiTope mapping of anti-Nipah Virus selected peptides to the Nipah Nucleoprotein. Epitopes previously predicted and validated from this phage display peptide set are indicated with arrows. B. Novel GuiTope predictions using the inversion method of Nipah Glycoprotein G epitopes. The GuiTope alignment detail is shown as well as the locations of these epitopes in the crystal structure. The underlined glutamic acid is part of the receptor binding site.](1471-2105-13-1-3){#F3}\n\nThree protein panning datasets were also evaluated. In the first example, White *et al*. did not identify any similarity between the peptides found to bind to the endothelial protein C receptor (EPCR) and Protein C or any other known EPCR binding partners. GuiTope likewise did not find any significant similarity between any known EPCR interactors (Figure [4C](#F4){ref-type=\"fig\"}). In the second case, the peptides selected to bind to integrin \u03b15\u03b26 were mapped by GuiTope to the known interactors TGF beta 1 and TGF beta 3 as two of the top three hits (Table [1](#T1){ref-type=\"table\"}) and Guitope correctly identified the important interacting amino acids (Figure [4B](#F4){ref-type=\"fig\"}). Since these interactions were discovered after the publication of the phage display study, one may suppose that they could have been predicted from the phage display data if the proper analysis tools had been available. In the third set, neither Carettoni *et al*. nor the Guitope analysis reveal a clear similarity between the FtsA binding peptides and a known FtsA interactor. Carettoni *et al*. identified a weak motif that matched a site on FtsA, and used that site to develop a model for the structure of the FtsA dimer \\[[@B21]\\]. While several lines of evidence suggest that *E. Coli*FtsA does form a dimer, it is not clear whether the model proposed based on this phage-display data is correct \\[[@B22]\\]. We are not aware of any experimental evidence to support or refute the interactions predicted by GuiTope.\n\n![**Protein Interaction Predictions**. A. Peptides selected to bind Integrin AlphaV Beta6 clearly aligned in GuiTope to the integrin binding site on TGF beta 1. B. Detailed alignments of the peptides to TGF beta 1, with those that align to the binding site highlighted in yellow and those that do not contain the RGB motif shown in italic. Below the WebLogo view of peptides aligning to the region illustrates the relative importance of amino acids C. Peptides selected to bind to EPCR do not align to a particular region on protein C.](1471-2105-13-1-4){#F4}\n\n###### \n\nPhage display database search\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  MimoDB/Reference         Target                                       Database (number of Proteins)                          Known Interactor       Rank, p-value (Inv/No Inv)\n  ------------------------ -------------------------------------------- ------------------------------------------------------ ---------------------- -----------------------------\n  288 \\[[@B24]\\]           Endothelial protein C receptor               Human Extracellular and Cell Surface Proteins (5074)   Protein C              NA\\*\n\n  148 \\[[@B25]\\]           Polyclonal Anti-Nipah Virus                  Nipah Proteome (9)                                     Nucleoprotein          2, \\< 0.1/\\\n                                                                                                                                                      2, \\< 0.1\n\n  753,754,755 \\[[@B26]\\]   Integerin \u03b15\u03b26                               Human Extracellular and Cell Surface Proteins (5074)   TGF beta 1\\            2, \\< 0.0002/\\\n                                                                                                                               TGF beta 3             1, \\< 0.0002\\\n                                                                                                                                                      3, \\< 0.0002/\\\n                                                                                                                                                      2, \\< 0.0002\n\n  204-205 \\[[@B18]\\]       Anti-*M. hyopneumoniae*polyclonal antibody   *Mycoplasma hyopneumoniae*Proteome (691)               Lipoproteins and p97   None matched correct region\n\n  1127 \\[[@B21]\\]          *Escherichia coli*FtsA                       *Escherichia coli*(4311)                               FtsA                   1106,0.25/\\\n                                                                                                                                                      2067,0.58\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n\\*Protein C was not listed as Extracellular or a Cell Surface Protein GO annotation\n\nThe peptides that bind to a given target do not always have sequences that are similar to biologically relevant proteins. This problem is confounded when peptide array approaches are used because peptides that are highly similar to a given protein are unlikely to be present in the library. GuiTope was able to take these loosely similar sequences and predict antibody epitopes with modest accuracy (AUROC 0.75-0.9) in line with previously tested methods \\[[@B8]\\]. Random-sequence peptide microarrays have shown great promise in profiling the humoral immune response \\[[@B5],[@B23]\\], and it would be of great utility to be able to use the peptide sequences to trace back to the antigen that elicited the immune response. However, the current prediction accuracy would not be sufficient for this task \\[[@B8]\\]. In contrast to the peptide array datasets, the phage display selected peptides can sometimes be used to predict interaction partners from a database very accurately. As less biased molecular display methods are developed and higher density peptide arrays become available, we expect that the information content of the peptide sequences will improve, making the type of analysis facilitated by GuiTope even more useful.\n\nAvailability and requirements\n-----------------------------\n\nThe executable is available on <http://www.immunosignature.com/software> and will install and run on any PC with Windows XP or later. The source code is written in Visual Basic and available on sourceforge.net. The Microsoft .NET framework is required.\n\nAuthors\\' contributions\n=======================\n\nSAJ conceived the project. JE and RFH designed and implemented the software. KAN suggested the di-peptide inversion method. KAN and PS tested the software. RFH, PS, and SAJ designed the evaluation strategy. RFH performed the evaluation and drafted the manuscript. PS, SAJ, and KAN critically revised the manuscript. All authors have read and approved the final version of the manuscript.\n\nAcknowledgements\n================\n\nThe authors gratefully acknowledge Dr. J. Bart Legutki for software testing and valuable discussion, and Kevin Brown for critical suggestions on implementation. This work was supported by grants from the DoD Breast Cancer Research Program and the Defense Threat Reduction Agency to SAJ.\n"}
+{"text": "Background\n==========\n\nIn order to understand the living nature Systems Biology develops computational models of biological systems. These models are computational in the sense, that the models are expressed in an appropriate formal language, like the Systems Biology Markup Language (SBML, \\[[@B1]\\]) and CellML \\[[@B2]\\], and can be used by computer programs in order to infer statements about its dynamical behaviour (either quantitative or qualitative). In contrast to \\[[@B3]\\] we also call differential equation models \"computational\".\n\nWe call a computational model of a biological system a *bio-model* if it allows for an explanation of the mechanism behind the observed behaviour of the biological system. Therefore the model not only has to imitate the behaviour of the system. In addition, the components of the model must possess a biological meaning with respect to the modelled system. Only if the model has both the same *performance* (the behaviour) and the same *competence* (the mechanism) as the biological system, we can understand the living system by means of the model \\[[@B4]\\].\n\nToday's high-quality and high-throughput experimentation techniques in molecular biology are the basis for an increasing number of bio-models with growing size and complexity. Understanding biological systems on the system-level requires the integration of bio-models from different abstraction levels and with different paradigms \\[[@B5]\\]. Obviously, modelling on a system-level will require the very assistance of computers. Although computational bio-models themselves are represented in some formal language their meaning often is only described in natural language. Computer-aided modelling in Systems Biology will be impossible until the meaning of the models is formally described. In this paper we introduce the *meaning facets* of bio-models which are views of a bio-model from different perspectives. The meaning facets provide a conceptual framework for a systematic specification of the meaning of a bio-model and consequently are the basis for rigorous semantics of the bio-model.\n\nFormal semantics of bio-models which go beyond the usual formal specification of the model structure and comprehends all meaning facets would be desirable to provide computer support in the following tasks:\n\nSemantics based search\n----------------------\n\nGiven certain desired model properties find models that exhibit these properties. For example, both example models discussed below should be retrievable by search queries of the types: \"Find models describing the cell cycle!\", \"Find models related to p34 protein kinase!\", or \"Find models that exhibit both steady state and oscillating behaviour!\".\n\nModel comparison\n----------------\n\nGiven two models, do they semantically overlap? Is one model a sub-model of the other? Or is one of them an abstraction of the other? In general, a method for model comparison is needed for many higher level tasks like model matching or model integration. The comparison should apply to all perspectives of the model's meaning (see below). A comparison of two models can have different kinds of results: e.g. identical, similar, competing, contradictory, or subsuming models.\n\nAnnotating models\n-----------------\n\nThe annotation of a model can be done in an interactive mode: Starting with some elementary facts about a model an interactive system (see below) infers more facts and asks for missing information. Thereby it suggests possible answers. Furthermore, the system complains about inconsistencies. The result is a complete and consistent annotation of the model.\n\nBeside these tasks related to the storage, retrieval and exchange of models in a collaborative setting formal semantics could be the basis for computer-aided modelling. By means of automatic reasoning it would allow for higher-level tasks like:\n\nModel integration\n-----------------\n\nGiven two models that semantically overlap, what would an integrated model look like? Again, the formal semantics of the model's components is needed in order to automate this task.\n\nModel use\n---------\n\nIn order to simulate and predict the behaviour of a biological system the bio-model has to be implemented in a computer code. This causes further problems: Without formal semantics a biologist must directly modify the code in order to change the model. If the extrinsic meaning of model components and their inter-dependencies with the intrinsic model structure were formalised, it would be possible to modify the model on a more abstract semantic level without the need to refer to the implementation.\n\nModel revision\n--------------\n\nGiven desired behaviours, is the actual dynamics of a model in accordance with them? The diagnostics of a potential discrepancy will suggest possible changes of the model. The corresponding improvement could be used iteratively to \"evolve\" models.\n\nA formal semantic description of bio-models would not only be useful in corresponding computer-assisted application scenarios, but also would support biologists to access models, their use and their behaviour as well as the underlying assumptions and decisions. A formal description of the involved knowledge would allow to present relevant information about a model to biologists in a familiar way.\n\nThe biological scientist does not have to cope with this rather complicated formalisation of the semantics. We envision an interactive system for computer-aided annotation of bio-models. Based on a knowledge representation system working in the background this system can guide the user in entering all the necessary information while constantly checking the consistency of the resulting information. Furthermore, the system will be able to ask for specific kinds of information depending on the information already entered and can provide candidate answers to the user.\n\nResults\n=======\n\nThe semantics of bio-models is a formal account of their meaning. In order to specify the semantics for the intended application scenarios we therefore have to know what a bio-model means and which aspects of its meaning are relevant. From a closer investigation of the human understanding of bio-models and the way how bio-models describe biological phenomena we derived a conceptional scheme of the meaning of bio-models \\[[@B4]\\]. This scheme resembles results from knowledge representation of complex systems (see below). The conceptional scheme consists of six meaning facets (Figure [1](#F1){ref-type=\"fig\"}). The meaning facets are views at the meaning of a bio-model from different perspectives. The starting point for the interpretation of a bio-model with respect to the different facets is a model specification, i.e. an expression in some formal language. We claim that the formal semantics of a bio-model has to incorporate all of these meaning facets and the relations between them in order to enable full computer support for modelling. The proposed conceptual framework is a systematic account of the semantics of bio-models. It can guide the development of formal representations for bio-modelling and provides a coverage criterion for such efforts.\n\n![**The six meaning facets of a bio-model.** Three pairs of intrinsic/extrinsic meaning facets from left to right: structure (system, entities, relation), function (intention, instantiation, setup), and behaviour (dynamics, data, outcome). The terms in parentheses are sub-facets explained in the text.](1752-0509-7-43-1){#F1}\n\nDual interpretation\n-------------------\n\nA Bio-model has a dual interpretation: The mathematical expression bears meaning by itself without referring to the biological reality. It can be interpreted, analysed, and used in computational simulations without knowing what it represents. We call this interpretation the *intrinsic meaning* of the bio-model. However, a bio-model is more than a pure syntactical formal expression: it describes a piece of biological reality and thereby also exhibits an *extrinsic meaning*. Often, the extrinsic interpretation is referred to by the word \"represents\": for example, we say that a variable *xrepresents* the concentration of a specific substance and that the oscillation shown in simulations *represents* variations in concentrations during the cell cycle. An explanatory bio-model establishes a mapping between the two conceptual sides, i.e. between the intrinsic and extrinsic meaning. Note that the biological interpretation has to be consistent with the usual conceptualisation made in biology. This ensures that modelling results represent biological phenomena in such a way that the (intrinsic interpreted) model can explain biological reality (cf. Figure [2](#F2){ref-type=\"fig\"}).\n\n![**Dual interpretation of bio-models.** A model can be mathematically interpreted as a text in a formal language resulting in \"formal semantics\". This intrinsic meaning is necessary for using the model in computations. In order to exploit the results of such computations for the explanation of biological phenomena the model needs also a biological interpretation: the model possesses an extrinsic meaning relating its structure, its functionality, and its behaviour to biological reality. Ultimately, modelling is about making appropriate computational representation of biological reality.](1752-0509-7-43-2){#F2}\n\nIn the SBML community (see, e.g., \\[[@B6]\\]) the two sides of the meaning are often called \"model meaning\" (all information necessary to simulate a SBML model) and \"biological meaning\" (annotations of what is meant by a particular SBML component). The term \"model meaning\", however, is too general and therefore misleading. Furthermore, \"biological meaning\" is very specific to bio-models. We therefore use the terms \"intrinsic\" and \"extrinsic\" in order to (1) avoid the ambiguity of \"model meaning\" and (2) allow our framework to be applicable to other kinds of models.\n\nThree perspectives of meaning\n-----------------------------\n\nFollowing research from teleological modelling in engineering (see, e.g., \\[[@B7]\\] for recent work on this topic) three pragmatic meaning perspectives can be identified: (1) The meaning regarding the components of the model and the relations between them accounts for its *structure*. (2) The meaning regarding the model in connection to its context and its intended use accounts for its *function*. (3) The meaning regarding the dynamics of the model accounts for its *behaviour*. The extrinsic/intrinsic sides of the three perspectives together form the six meaning facets illustrated in the \"meaning diamond\" (Figure [1](#F1){ref-type=\"fig\"}). In order to represent the complete meaning of a bio-model one has to specify the intrinsic and extrinsic side of each of the three perspectives and the connections between them.\n\nThe following sections describe the three meaning perspectives in more detail. In order to illustrate the meaning facets the meaning of two semantically related models of the cell cycle from Tyson \\[[@B8]\\] is sketched with reference to existing formal approaches. The contribution of all mentioned formal approaches to the meaning facets is summarised in Table [1](#T1){ref-type=\"table\"}. Obviously, the extrinsic side is considerably less covered than the intrinsic side. This is due to the very complexity of biological reality and our restricted knowledge about it (see also the discussion of \"Biological Meaning\" below). The equations of the models are shown in subsection Example models, SBML encoded versions of the models can be found in BioModels Database \\[[@B9]\\]. In \\[[@B10]\\] we published a complete reconstruction of the meaning of this models which was based on a preliminary version of the meaning facets framework.\n\n###### \n\nFormal approaches to the meaning facets of bio-models\n\n  **Formal**        **Language element/**               **Intrinsic meaning facet**                                                                                         \n  ----------------- ----------------------------------- ----------------------------- -------------- --------------- ---------- ---------- ---------- ---------- ---------- ----------\n  SBML              `species`                           \u00a0                             x              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `reaction`                          \u00a0                             \u00a0              x               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `kineticLaw`                        \u00a0                             \u00a0              x               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `initialAmount`                     \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `parameter`                         \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  SBO               \"modelling framework\"               x                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  SED-ML            `Simulation`                        \u00a0                             \u00a0              \u00a0               x          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `Change`                            \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  KiSAO             \"modeling and simulation            \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          x          \u00a0          \u00a0          \u00a0\n  \u00a0                 algorithm\"                          \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  DYML              feature                             \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          x          \u00a0          \u00a0\n  \u00a0                 constraints                         \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  TEDDY             \"temporal behaviour\"                \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          x          \u00a0          \u00a0\n  \u00a0                 \"behaviour diversification\"         \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          x          \u00a0          \u00a0\n  Temporal logics   \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          x          \u00a0          \u00a0\n  SBRML             `Result`                            \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          x          x\n  Fielded Text      \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          x          x\n  **Formal**        **Language element/**               **Extrinsic meaning facet**                                                                                         \n  **approach**      **Ontology branch**                 **Structure**                 **Function**   **Behaviour**                                                          \n  \u00a0                 \u00a0                                   **(S1)**                      **(S2)**       **(S3)**        **(F1)**   **(F2)**   **(F3)**   **(B1)**   **(B2)**   **(B3)**\n  UniProt           \u00a0                                   \u00a0                             x              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  NCBI Taxonomy     \u00a0                                   x                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  Database          \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  Gene ontology     \"Biological Process\"                x                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          x          \u00a0          \u00a0\n  SBO               \"physical entity representation\"    \u00a0                             x              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 \"systems description parameter\"     \u00a0                             x              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 \"occurring entity representation\"   \u00a0                             \u00a0              x               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 \"systems biology representation\"    \u00a0                             \u00a0              x               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  Reactome          \u00a0                                   \u00a0                             \u00a0              x               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  Cell type         \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  Ontology          \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          \u00a0          \u00a0\n  SABIO-RK          \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  FuGE              `Material`                          \u00a0                             \u00a0              \u00a0               \u00a0          x          \u00a0          \u00a0          \u00a0          \u00a0\n  \u00a0                 `Investigation`                     \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          x          \u00a0          \u00a0          \u00a0\n  \u00a0                 `Data`                              \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          x          x\n  SBRML             `Result`                            \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          x          x\n  Fielded Text      \u00a0                                   \u00a0                             \u00a0              \u00a0               \u00a0          \u00a0          \u00a0          \u00a0          x          x\n\nFormal approaches (languages, ontologies and biological resources) for computer representations of bio-models and their contribution to the formalisation of the meaning facets as shown for the example models by Tyson. An \"x\" means that the respective formal approach or some given aspect (language element or ontology branch) of it is able to represent the specific meaning facet or at least parts of it. The table is divided into intrinsic (top) and extrinsic (bottom) meaning facets.\n\nExample models\n--------------\n\nIn this section we introduce two models of the cell cycle by Tyson \\[[@B8]\\] which are used as an example in the following description of the meaning facets. Both models describe the formation and activation of the maturation promoting factor (MPF), a hetero dimer made of the two proteins cyclin and cdc2.\n\n*Model 1* consists of six ordinary differential equations (ODEs) where each equation models the temporal evolution of the concentrations of one of the involved substances with respect to the concentrations of the other substances:\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{C2} \\right\\rbrack/dt} & {= k_{6}\\left\\lbrack \\;\\textsf{M} \\right\\rbrack - k_{8}\\left\\lbrack \\sim \\textsf{P} \\right\\rbrack\\left\\lbrack \\;\\textsf{C2} \\right\\rbrack + k_{9}\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack/dt} & {= - k_{3}\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack\\left\\lbrack \\;\\textsf{Y} \\right\\rbrack + k_{8}\\left\\lbrack \\; \\sim \\textsf{P} \\right\\rbrack\\left\\lbrack \\;\\textsf{C2} \\right\\rbrack - k_{9}\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{pM} \\right\\rbrack/dt} & {= k_{3}\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack\\left\\lbrack \\;\\textsf{Y} \\right\\rbrack - \\left\\lbrack \\;\\textsf{pM} \\right\\rbrack F\\left( \\left\\lbrack \\;\\textsf{M} \\right\\rbrack \\right) + k_{5}\\left\\lbrack \\; \\sim \\textsf{P} \\right\\rbrack\\left\\lbrack \\;\\textsf{M} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{M} \\right\\rbrack/dt} & {= \\;\\left\\lbrack \\;\\textsf{pM} \\right\\rbrack F\\left( \\left\\lbrack \\;\\textsf{M} \\right\\rbrack \\right) - k_{5}\\left\\lbrack \\; \\sim \\textsf{P}\\; \\right\\rbrack\\left\\lbrack \\;\\textsf{M} \\right\\rbrack - k_{6}\\left\\lbrack \\;\\textsf{M} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{Y} \\right\\rbrack/dt} & {= k_{1}\\left\\lbrack \\;\\textsf{aa} \\right\\rbrack - k_{2}\\left\\lbrack \\;\\textsf{Y} \\right\\rbrack - k_{3}\\left\\lbrack \\;\\textsf{CP} \\right\\rbrack\\left\\lbrack \\;\\textsf{Y} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{\\; d\\left\\lbrack \\;\\textsf{YP} \\right\\rbrack/dt} & {= k_{6}\\left\\lbrack \\;\\textsf{M} \\right\\rbrack - k_{7}\\left\\lbrack \\;\\textsf{YP} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{F\\left( \\left\\lbrack \\;\\textsf{M} \\right\\rbrack \\right)} & {= k_{4}^{\\prime} + k_{4}\\left( \\left\\lbrack \\;\\textsf{M} \\right\\rbrack/\\left\\lbrack \\;\\textsf{CT} \\right\\rbrack \\right)^{2}\\qquad} \\\\\n\\end{array}$$\n\nInvolved substances are: cdc2 (C2), phosphorylated cdc2 (CP), inactive MPF (pM), active MPF (M), cyclin (Y), phosphorylated cyclin (YP), adenosine triphosphate (\u223cP), and amino acids (aa). CT means total cdc2, i.e. \\[ CT\\]=\\[ C2\\]+\\[ CP\\]+\\[ pM\\]+\\[ M\\]. The *k*~*i*~ are kinetic rate coefficients.\n\n*Model 2* is a mathematical abstraction of *Model 1* under certain additional biological assumptions:\n\n$$\\begin{array}{ll}\n{du/dt} & {= k_{4}\\left( v - u \\right)\\left( \\alpha + u^{2} \\right) - k_{6}u\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n{dv/dt} & {= \\left( k_{1}\\left\\lbrack \\textsf{aa} \\right\\rbrack/\\left\\lbrack \\textsf{CT} \\right\\rbrack \\right) - k_{6}u\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\nu & {= \\;\\;\\left\\lbrack \\;\\textsf{M} \\right\\rbrack/\\left\\lbrack \\;\\textsf{CT}\\; \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\nv & {= \\left( \\left\\lbrack \\;\\textsf{Y} \\right\\rbrack + \\left\\lbrack \\;\\textsf{pM} \\right\\rbrack + \\left\\lbrack \\;\\textsf{M} \\right\\rbrack \\right)/\\left\\lbrack \\;\\textsf{CT} \\right\\rbrack\\qquad} \\\\\n\\end{array}$$\n\n$$\\begin{array}{ll}\n\\alpha & {= k_{4}^{\\prime}/k_{4}\\qquad} \\\\\n\\end{array}$$\n\n*u* and *v* are relative concentrations following the given equations.\n\nStructural facets\n-----------------\n\nA bio-model describes state changes of a formal system. The notion of structure refers to the aspects of the system which do not change. In most general terms structure can be described by entities having attributes and relations between the entities: the attributes of the entities constitute the state of the system, the relations describe inter-dependencies between the attributes of related entities. The structural entities and relations have to be rather classes than individuals. Whereas a individual molecule can be formed, changed and destroyed, a molecule sort (a class of molecules) remains the same all over the time. Based on the relations a programme determines how the system state is changing. The intrinsic structural meaning is obtained by interpreting the given model specification with respect to the formalism used. An explanation of the behaviour of the modelled biological system requires to map this intrinsic structure to relevant biological objects characterised by quantities and interactions establishing mechanisms. Usually, what is called a \"biological systems\" in fact denotes a class of concrete systems in reality. The concrete systems are considered on a specific conceptual level, e.g. as gene regulatory networks, protein interaction networks, signal transduction pathways, or metabolic networks. In turn, this common view onto the concrete systems establishes itself an abstract system, the \"biological system\". The conceptual level must be reflected by the formalism used. In detail the structural meaning can be characterised as follows:\n\n### (S1) System\n\n#### intrinsic\n\nWhich *formal system* is specified by the encoded model (the model itself)? Which *formalism* is employed by the model (modelling framework, spatiality, stochasticity)?\n\n#### extrinsic\n\nWhich *biological system* corresponds to the formal system (species, cell type, biochemical system)? Which *conceptual level* is reflected by the used formalism (system type, granularity, spatial and temporal resolution)?\n\n### (S2) Entities\n\n#### intrinsic\n\nWhat are the *entities* of the formal system (individuals, collections, agents)? Which *attributes* of the entities describe the state of the system (variables, terms)?\n\n#### extrinsic\n\nWhat biological *objects* correspond to the model entities (molecules, substances, cells)? Which *quantities* (amounts, concentrations, units) correspond to the model attributes?\n\n### (S3) Relations\n\n#### intrinsic\n\nWhat are the *relations* between the entities (inter-dependencies, correlation, neighbourhood)? What is the *programme* describing changes of the attributes of related entities (operations, equations, update rules)?\n\n#### extrinsic\n\nWhat biological *interactions* correspond to the model relations (reactions, transformations, diffusion)? What biological *mechanisms* realising the interactions between objects correspond to the model programme (reaction steps, bonding, activity)?\n\n### Notes\n\nThe specification of the formalism (S1) will restrict the ways one can use the model. This information is essential for the interpretation of a model specification as a formal system. If, for example, a model specification does not provide information about the intended modelling framework (like discrete and continuous) the specification of the formal system is incomplete (see \\[[@B11]\\]). However, with this information it will also be possible to automatically convert models from one modelling frameworks in another \\[[@B11]\\]. Fages and Soliman \\[[@B12]\\] investigate the different interpretations of SBML models depending on the chosen modelling framework and relate the resulting different semantics by the theory of abstract interpretations.\n\nBiological systems are hierarchically organised. Often, this is reflected by a partonomy of entities in a model. This partonomy is described as relations between the entities (S3). Furthermore, it has to be described how attribute changes in a part influence attribute changes in the corresponding whole and vice versa. Thereby the whole system can be seen as the top-level entity in the partonomic hierarchy.\n\nIn general, the programme has formal parameters. The actual parameters (i.e. the parameter values) must be appropriately instantiated, see facet (F2) below.\n\nFor understanding a model it is useful to capture the relationships between biological objects, e.g. between a protein and its phosphorylated versions or between a dimer and its part (maybe modelled as partonomic relations, see above).\n\nBiological processes often happen in separated compartments. There are two ways to account for this compartmentalisation: In which compartment an object resides can be represented by an attribute of the corresponding entity. In contrast, objects of the same type residing in different compartments can be modelled by different classes of entities. A relation has to describe the exchange between the compartments.\n\n### Example\n\nIntrinsically, both Tyson models are encoded in SBML. The respective *formal systems* (S1) are given by the equations in subsection Example models. The used *formalism* (S1) can be characterised as a set of coupled ordinary differential equations of continuous state variables in the common independent variable *t*, which describes a deterministic non-spatial state evolution. The modelling framework can be specified by a term from the Systems Biology Ontology (SBO, \\[[@B13]\\]): \"non-spatial continuous framework\" (SBO:0000293). The intrinsic structural meaning of SBML models is formalised by the SBML specification (we use `typewriter font` for SBML keywords): The *entities* (S2) are given as `species` in the `listOfSpecies`. Each species has an unique `id` and a name, e.g. the species C2 is called \"cdc2k\". Each `id` is also used as dependent variable within the `kineticLaw`s (see below), representing the amount as an *attribute* (S2) of the according `species`. The *relations* (S3) are `reactions` in the `listOfReactions`. Each `reaction` has a `kineticLaw` describing the corresponding changes of the species amounts. There can be (formal) `parameter`s in the `kineticLaw`. For instance there is the following reaction (Reaction1 in the SBML encoding) in the Tyson model:\n\n$$M\\underset{}{\\overset{k_{6}}{\\rightarrow}}C2 + \\textit{YP}$$\n\nwith the kinetic law *k*~6~\\[ M\\], where *k*~6~ is a parameter determining the reaction rate. The *programme* (S3) of an SBML model is just the set of ODEs (cf. the equations in subsection Example models) reflecting the kinetic laws of the single reactions. For better legibility, we use the common style for kinetic equations with square brackets denoting the amount of a species. However, SBML has a specific syntax for `kineticLaw` based on MathML \\[[@B14]\\].\n\nExtrinsically, both Tyson models describe a *biological system* (S1) of MPF (maturation promoting factor) formation and activation which controls major events of the cell cycle in different organism: frog, sea urchin, and fission yeast. The extrinsic meaning of each SBML tag can be given by `annotations` pointing to an appropriate description of biological knowledge. E.g. C2 represents the biological *object* (S2) \"Cyclin-dependent kinase 1\" for which the UniProt \\[[@B15]\\] entry P04551 can be given. MIRIAM Registry \\[[@B16]\\] can be used for a unified way of referring to all external resources used in describing the meaning of bio-models. For example, the UniProt entry for the extrinsic meaning of C2 will become urn:miriam:uniprot:P04551. By means of identifier.org \\[[@B16]\\] one can also provide a persistent URL for this information: <http://identifiers.org/uniprot/P04551>z. The addressed organism could be assigned by the NCBI Taxonomy Database \\[[@B17]\\], e.g. sea-urchins have the Taxonomy ID: 7625. But one could also use more general entries, like the common parent term of sea urchins and frogs Deuterostomia (Taxonomy ID: 33511), or even more general Eumetazoa (Taxonomy ID: 6072). It would not be suitable to go further up in the taxonomy, because the facts about embryonic development in \\[[@B8]\\] do not apply in general for higher taxa. An annotation can specify the addressed biochemical system. The addressed system of the example models can be specified by a link to the \"mitotic cell cycle\" (GO:0000278) entry of the Gene Ontology (GO, \\[[@B18]\\]). The biological systems are regarded on the *conceptual level* (S1) of pools of molecular entities without consideration of spatial effects. The justification for this are high enough numbers of molecules and fast diffusion. *Model 1* is a network of protein-protein interactions, where catalytic reactions between proteins changes their concentrations in time (extrinsic interpretation of the independent variable *t*). *Model 2* is an abstraction of the actual protein-protein interactions. In addition to the SBML `annotations`, it is possible to refer to SBO directly within a SBML tag. SBO terms can be used for a top-level classification of molecules and reactions. For example, C2 can be classified as \"polypeptide chain\" (SBO:0000252), and Reaction1 as \"dissociation\" (SBO:0000180). Each involved substance has as a *quantity* (S2) its molar concentration (SBO:0000472) in mol/1. Reaction1 represents the *interaction* (S3) \"cyclin cdc2k dissociation\" for which the Reactome entry REACT_6308 can be given. Furthermore, the role which a molecule plays in a reaction can also be described by SBO (e.g. reactant, product, modifier). The *mechanism* (S3) underlying most of the reactions is mass-action kinetics, which could be annotated by \"mass action rate law\" (SBO:0000012). Only Reaction9 represents a special mechanism for a autocatalytic feedback \\[[@B8]\\].\n\nThe structural meaning of *Model 2* can be expressed in a similar way. However, the extrinsic meaning of u and v is not straightforward. It can only be derived using the defining equations *u*=\\[ M\\]/\\[ CT\\], *v*=(\\[ Y\\]+\\[ pM\\]+\\[ M\\])/\\[ CT\\] and assigning meanings to the contained entities like M (see above). The extrinsic interpretation of the reactions of *Model 2* is even harder. As interpreted in BioModels Database the reactions do not contain a feedback loop anymore but still showing oscillating behaviour!\n\nFunctional facets\n-----------------\n\nA bio-model is simulated in order to get data for answering biological questions. The function of a model describes how the model structure is intended to be used in simulations to generate dynamic behaviour. Before a model can be used in simulations it has to be fully instantiated, i.e. all parameters should be given actual values and the initial state of the model has to be set. The simulation setup describes the exact procedure applied to the model instance. In addition, the post-processing describes how to produce the final outcome. The instantiation and the setup of the simulations performed with the model have to reflect the specific boundary conditions and the experimental settings under which the biological system is observed. The functional meaning can be characterised by the following questions:\n\n### (F1) Intention\n\n#### intrinsic\n\nWhat is the *intended use* of the model (simulation type, combination of simulations, desired outcome)? Which *constraints* are imposed on the model (value restrictions, ratios, conservation rules)?\n\n#### extrinsic\n\nWhich biological *questions* are addressed to the model (explanation, hypothesis testing, exploration, dependency analysis)? Which *assumptions* provide the basis for the constraints (likelihoods, justification, evidence)?\n\n### (F2) Instantiation\n\n#### intrinsic\n\nWhich *instantiation* of the model is used for the simulation (parameter values, parameter ranges)? Which *initial values* are chosen for the entities attributes (value assignment to variables)?\n\n#### extrinsic\n\nWhich *boundary conditions* correspond to the model instantiation (environment, kinetic data, plausible ranges)? Which *initial state* of the biological system corresponds to the initial values set for the model (initial concentrations)?\n\n### (F3) Setup\n\n#### intrinsic\n\nWhich *setup* is used for simulation experiments (simulation algorithm, algorithm settings, perturbations)? Which *post-processing* of the raw simulation data generates the desired outcome (normalisation, conversations of units, calculations)?\n\n#### extrinsic\n\nWhich biological *experimental settings* correspond to the setup for simulations of the model (experimental protocol)? Which *result calculation* produces the requested results of the experiment (normalisation, conversations of units, calculations)?\n\n### Example\n\nIn the following we describe the functional aspects of using *Model 1* for producing the time series of Figure [3](#F3){ref-type=\"fig\"}(a) in \\[[@B8]\\].\n\n![**Different meanings of \"Cell\".** Three different examples of models for which the extrinsic meaning \"Cell\" can be viewed at from different perspectives: a proto-cell model \\[[@B30]\\], a Virtual Cell model \\[[@B31]\\], and an immune response model \\[[@B32]\\].](1752-0509-7-43-3){#F3}\n\nIntrinsically, the functional meaning of the example models is equivalent to a complete description of simulation experiments applied to the model comprising all the details of (F1-3). Most simulation tools use their own proprietary format to encode this information which hampers the reuse of functional information. In order to overcome this situation the Simulation Experiment Description Markup Language (SED-ML) is developed \\[[@B19]\\]. In a SED-ML description algorithm used for the simulation can be specified using KiSAO (Kinetic Simulation Algorithm Ontology, \\[[@B13]\\]). The *intended use* (F1) is the generation of a time series through numerical integration of the model. Thereby the time evolution of the amounts of M and YT are reported. In SED-ML this intended type of simulation can be set in the `listOfSimulations` as `uniformTimeCourse`. Tyson gives some *constraints* (F1), e.g. \\[ CT\\]=*const*. and *k*~2~\u226a*k*~3~\\[ CT\\]. Constraints often are only implicit in the chosen simulation paradigm and algorithm; making them explicit will be a future challenge. It is an open issue how to formalise such constraints and the corresponding assumptions (see below). A combination of Constraint-Logic-Programming \\[[@B20]\\] with languages from Systems Biology seems to be a promising research direction for explicitly incorporating constraints and assumptions into models. For the *instantiation* (F2) of the model Tyson gives parameter values. Parameters can be set in SBML via the corresponding `value` attribute. There are no *initial values* (F2) explicitly given in \\[[@B8]\\]. However, one can find appropriate initial values from the time series in Figure [3](#F3){ref-type=\"fig\"}(a). The SBML file from BioModels Database contains such an assignment in the `initialAmount` attributes. The parameters and initial values can also be set or modified in SED-ML. In SED-ML it is possible to describe the *setup* (F3) of the experiment and the *post-processing* (F3) of the data: For Figure [3](#F3){ref-type=\"fig\"}(a) of \\[[@B8]\\] a uniform time course from time 0 min to 100 min with a step size of 0.001 min is produced with a fourth-order Adams-Moulton integration routine (KISAO_0000280). Subsequently, the raw amount data of M and the sum of all \"cyclin\" entities (called \"total cyclin\", \\[ YT\\]) are normalised by dividing by the amount of CT.\n\nTyson explicitly states biological *questions* (F1) addressed to the model. The corresponding question for Figure [3](#F3){ref-type=\"fig\"}(a) is: \"Can the same model also account \\[...\\], for rapid cycles of DNA synthesis and cell division (without cell growth) during the embryonic cell cycle, \\[...\\]\" \\[[@B8]\\], p.7329. That is, Tyson tries to explain a specific biological phenomenon. At the moment there is no way to formalise this question. However, we could imagine a classification of modelling aims. For each modelling aim appropriate simulation types could be identified. Tyson indicates *assumptions* (F1) as the basis for the parameter constraints. For example, he assumes that cdc2 is constantly synthesised in growing cells which supports \\[ *CT*\\]=*const*. The concrete *boundary conditions* (F2) corresponding to the instantiated model for Figure [3](#F3){ref-type=\"fig\"}(a) are the conditions found in early embryonic cells \\[[@B8]\\]. The Cell Type Ontology \\[[@B21]\\] can be used to specify the cell type: \"early embryonic cell\" has the ID CL:0000007. Tyson states that there is no experimental kinetic data available for this situation. Also, there is no *initial state* (F2) of the biological system given by Tyson. If such data would exist the corresponding SBML `parameter` and `initialAmount` could be annotated with kinetic data entries from appropriate sources (e.g. SABIO-RK, \\[[@B22]\\]). As with kinetic data and the initial state of the biological system, Tyson does not provide any information about *experimental settings* (F3) and the following *result calculation* (F3). Nevertheless, there are some standards for describing experimental protocols, like FuGE \\[[@B23]\\] for functional genomics experiments.\n\nThe functional facets of *Model 2* can be described in a similar way. However, there are specific assumptions underlying the abstraction of *Model 2* from *Model 1*. This assumptions are mainly reflected by the structure of *Model 2* (S2,3) and have to be met in a corresponding use of *Model 2* (F1,2).\n\nBehavioural facets\n------------------\n\nA bio-model is used in simulation experiments in order to investigate its behaviour. A simulation experiment produces raw numerical data. Often, this data is post-processed into the final desired outcome of the experiment. The model dynamics is a qualitative description of this experimental outcome. From the behavioural perspective the model dynamics should correspond to observed biological phenomena. The observed biological phenomena are supported by experimental results which are obtained from measurements of the biological system in question. The behavioural meaning can be characterised by the following questions:\n\n### (B1) Dynamics\n\n#### intrinsic\n\nWhich types of *dynamics* does the model show in simulations (fixed points, periodic behaviours, chaotic behaviour)? Which *diversification* in the dynamics does the model possess (stability, bifurcations)?\n\n#### extrinsic\n\nWhich biological *phenomena* correspond to the model dynamics (cyclic behaviour, steady state)? Which *variability* of the biological phenomena correspond to the diversification in the dynamics of the model (switching behaviour, excitability)?\n\n### (B2) Data\n\n#### intrinsic\n\nWhich *raw data* does the model produce in simulations (series of values)? Which *index* is used for the raw data (modelling time, parameter value, initial value)?\n\n#### extrinsic\n\nWhich experimental *measurements* correspond to the yielded raw data (series of values)? Which *key* is used to identify the single measurements (time, conditions, initial states)?\n\n### (B3) Outcome\n\n#### intrinsic\n\nWhat is the *outcome* of the simulation (specific values, time courses, phase portraits, bifurcation diagram)? Which *characteristics* of the model dynamics can be identified (maximal and minimal values, periods, Lyapunov exponents)?\n\n#### extrinsic\n\nWhich experimental *results* correspond to the outcome of the simulation (specific values, time courses, phase portraits, bifurcation diagram). Which *observables* correspond to the characteristics of the models dynamics (maximal and minimal concentrations, cycle length, stability)?\n\n### Example\n\nFor both example models Tyson identified three different types of *dynamics* (B1) dependent on the parameters setting: stable steady state, spontaneous limit cycle oscillation and excitable switch \\[[@B8]\\]. The intrinsic meaning of this dynamics types can be formalised by terms from the Terminology for the Description of Dynamics (TEDDY, \\[[@B13]\\]), e.g. `TEDDY_0000113` \"Stable Fixed Point\" for the stable steady state. TEDDY also provides terms for *diversification* (B1) of dynamics: In the example model there is a supercritical Hopf bifurcation (`TEDDY_0000074`) between the steady state and the oscillation if parameters *k*~4~ and *k*~6~ are varied. TEDDY only provides the vocabulary for describing the dynamics of models. We also need a language for relating conditions and types of dynamics. In an envisioned Dynamic Markup Language (maybe called DYML) the dynamics of *Model 2* could be formalised as follows (simplified notation):\n\nNote that only the parameters *k*~4~ and *k*~6~ are allowed to vary. All other parameters are set as in \\[[@B8]\\], Table two, p.7329. The constraints are taken from the phase plane analysis in \\[[@B8]\\], p.7332.\n\nThere exist other approaches for qualitative descriptions of model behaviour like temporal logics \\[[@B24]\\]. In BIOCHAM \\[[@B25]\\] a temporal logic is used as a query language for properties of the dynamics of bio-models. Instead of numerical simulations model checking techniques (see, e.g., \\[[@B24]\\]) can be used to answer such queries. In \\[[@B26]\\] a temporal logic extended by constraints over real numbers is used to express quantitative properties of temporal behaviour and to optimise parameters. Such quantitative temporal logics are worth to be considered as possible candidates for the needed model dynamics language.\n\nThe identification of the dynamics of the example models is based on simulation experiments which produce as *raw data* (B2) series of amount values for each entity. The *index* (B2) for these values is the modelling time (*t*). The intrinsic meaning of the values are the values itself. However, there is an issue to relate -- in a formalised manner -- the single values of a result table to the model attributes and the corresponding condition. There are some approaches to establish this connection, like SBRML \\[[@B27]\\] and Fielded Text \\[[@B28]\\]. The *outcome* (B3) of the simulations are plots of \\[ M\\]/\\[ CT\\] and \\[ YT\\]/\\[ CT\\] (see (F3) in the example above) against modelling time under different settings. Tyson also reports some *characteristics* (B3) of the model dynamics, e.g. relative amounts of M in steady state and period of the oscillation. Beside the time series there is another outcome in \\[[@B8]\\]: in the space of the parameters *k*~4~ and *k*~6~ regions of different qualitative behaviour are identified. Each region represents classes of concrete times series with common properties. These classes are the different dynamics of the model mentioned above and the plot of the regions in parameter space visualises the diversification in the dynamics.\n\nConcerning experimental *measurements* (B2) there is the same issue of connecting measurement values and the corresponding *keys* (B2) with model attributes and conditions as for simulation results (see above). Tyson does not provide concrete experimental measurements or *results* (B3). Instead, he refers to conclusions drawn from such data. For instance, he characterises the phenomenological *variability* (B1) by the different modes of operation observed in different developmental stages and states typical *observables* (B3) of these modes like the period of division cycles. The three different *phenomena* (B1) are mapped to the types of dynamics of the model: metaphase arrest in unfertilised eggs is represented by the steady state, rapid division cycles in early embryos by the spontaneous oscillation, and the growth-controlled division cycles in non-embryonic cells by the excitable switch. The extrinsic meaning of the dynamics can be grounded in external resources, e.g. \"cell cycle arrest\" (GO:0007050) for the metaphase arrest. The variability can be represented by linking conditions (e.g. early embryo stage) with the specific phenomena observed under this conditions.\n\nThe behaviour of *Model 2* is the same as for *Model 1* except for the number of dimensions of the dynamical system. Indeed, the simplified *Model 2* is used in \\[[@B8]\\] in order to also analyse the dynamics of *Model 1*.\n\nGlobal meta-information\n-----------------------\n\nBeside the meaning of a model itself there exists additional information describing the role of the model in scientific research. We call meta-information of this type \"global meta-information\". Global meta-information accounts for the origin of the model, the access to the encoded model in some formal language, and the relation of the model to other models. We will not provide a detailed systematics of global meta-information here. Instead we describe just the global meta-information for the example models.\n\n### Example\n\nBoth example models are originally published in \\[[@B8]\\]. The corresponding meta-information for the *origin* of the models comprises the paper itself (PubMed ID: 1831270), its author (John J. Tyson) and its date of publication (August 1991).\n\nImportant meta-information for the *access* of an encoded model involves the place (file name, URL, database ID), the used format (e.g. SBML, CellML), and the date and author of the encoding. If the model is stored in a database then there also exists meta-information about the curation process (curators, date, last modification). The example models are available in BioModels Database encoded in different formats: *Model 1*<http://identifiers.org/biomodels.db/BIOMD0000000005>*Model 2*<http://identifiers.org/biomodels.db/BIOMD0000000006>\n\nBioModels Database also lists the mentioned meta-information about the encoding and curation process. For example, one can access the encoded models in SBML, Level 2, Version 4. The format is determined by the `xmlns:m` attribute in the `sbml` tag.\n\nA model can have different *relations* to other models: It can be evolved from preliminary versions, it can be abstracted or integrated from other models, and it can be compared to competing models. The derivation of *Model 2* from *Model 1* is the result of an abstraction relation between the two models. Tyson also mentioned some existing related model. For example, he states that *Model 2* is a modified version of the famous \"Brusselator\". There are formal approaches to relate models, e.g. based on graph theory \\[[@B29]\\].\n\nDiscussion\n==========\n\nOur analysis showed that formalising the meaning of bio-models requires a significant effort and is not trivial, since the meaning appears from several perspectives and in different facets (cf. Figure [1](#F1){ref-type=\"fig\"}). We have nevertheless demonstrated how, in principle, it is possible to specify the meaning in a form that is understandable by both, computers and humans.\n\nThe proposed meaning facets framework allows for a systematic classification of existing approaches for computer-readable representations of model meaning. The framework therefore can be used to evaluate the coverage of representations and to identify missing pieces. Interesting next steps involve the extension of *BioModels Database*\\[[@B9]\\] by introducing the behavioural meaning perspective and by considering the intrinsic mathematical structure in order to grasp the semantics of variables like *v* in *Model 2*.\n\nFor the envisioned intelligent computer-aided working environment, which semantically guides model design and use and fosters the development of sound and well annotated bio-models, we have to establish appropriate languages for the missing pieces, like a description language for the behavioural perspective. Furthermore, existing languages and resources have to be improved in order to enable the necessary reasoning capabilities. The proposed meaning facets framework can direct this developments.\n\nBiological meaning\n------------------\n\nThe following are some explanatory notes regarding the biological (extrinsic) meaning of bio-models and its formalisation:\n\n1\\. In general, the extrinsic meaning will only be partial, i.e. there may be aspects of the model without counterparts in the biological world. But at least there has to be some aspect of a model which has an extrinsic interpretation. Without representing a concrete biological system a model would be (biological) meaningless!\n\n2\\. Even if an extrinsic interpretation of some model aspect exists it doesn't have to be intuitive. The more intuitive a model represents our perception of reality, the better it explains the modelled system and consequently contributes to an understanding of the living nature.\n\n3\\. The extrinsic interpretation depends on the intention of the model. Therefore the same mathematical construct can have more than one biological meaning. For example, the exponential grow $\\overset{\u02d9}{x} = {\\alpha x},\\alpha > 0$ can be a model for different biological phenomena.\n\n4\\. It is tempting to assume that familiar biological objects, like \"cell\" are represented in the model from the structural perspective, i.e. that there is a structural entity interpreted as \"cell\". This often is not the case. Figure [3](#F3){ref-type=\"fig\"} illustrates that all three perspectives of meaning can refer to \"cell\". This shows, that biological objects can also have some behavioural and functional aspects. The three meaning perspectives should not be regarded as independent from each other, but rather as different views of an indivisible unity.\n\nRelated work\n------------\n\nThe insight in the dual interpretation of mathematical models are of course not new: The \"knowledge representation hypothesis\" \\[[@B33]\\] demands that any useful formal representation needs both: to play a formal role and to have an \"external semantical attribution\". Also, Simon's notion of artefacts (like models) as interfaces between an inner and an outer environment \\[[@B34]\\] resembles the dual interpretation of bio-models. However, for a systematic formal specification of the meaning of bio-models it is very useful to distinguish between the intrinsic and the extrinsic interpretation. In fact, Rosen's central \"Modeling Relation\" \\[[@B35]\\] is formulated as a congruence between a natural system and a formal system (a model). Thereby, biological \"percepts\" and \"linkages\" between them are encoded by formal entities and relations. Inferences in the formal systems can be decoded as predictions about the behaviour of the natural system. Thus, \\[[@B35]\\] already distinguish between the intrinsic/extrinsic sides on the structural and the behavioural perspective with a focus on the interplay of the two sides, not on the details of the structure and behaviour provided in this paper.\n\nThere is a similar distinction between perspectives in \\[[@B36]\\]: Their \"model description\" is more or less what we call structural perspective. Their \"simulation description\" is part of the functional perspective described above. Our behavioural perspective is called \"simulation results description\" in \\[[@B36]\\]. Our meaning facets however are more systematic and provide more details from each perspective. \\[[@B36]\\], nevertheless, gives a good overview of important standards, languages, and ontologies for the three perspectives.\n\nAnother systematic approach to models is Zeigler's \"framework for modeling and simulation\" \\[[@B37]\\]. The framework consists of four elements: the source system, the experimental frame, the model, and the simulator. Each element involves knowledge on specific \"system specification levels\" (for details cf. \\[[@B37]\\]). There are some connections between Zeigler's framework and the meaning facets: Zeigler's \"state transition\" level 3 \\[[@B37]\\], p.17f corresponds to the *programme* in (S3), the \"coupled component\" level 4 corresponds to *entities* (S2) and *relations* (S3). Both levels together are used to specify models, therefore a \"model\" in Zeigler's framework is what we call *structural facets*. The \"experimental frame\" formalises the conditions for simulating the model, thus it corresponds to the *instantiation* (F2) and the *setup* (F3). Zeigler claims that the experimental frame \"is a operational formulation of the objectives that motivate a modeling and simulation project\" \\[[@B37]\\], p.27, so it corresponds also to the *intention* (F1). The \"source system\" is regarded as a source of data of the \"I/O behaviour\" level 1, which corresponds to *raw data* (B2). Although there are parallels between the two frameworks, Zeigler's work is focused on the mathematical side of building models and using them in simulations. In contrast, the approach proposed here regards models as \"integrators of knowledge\" \\[[@B13]\\] in the centre between computations and biological reality (cf. Figure [2](#F2){ref-type=\"fig\"}). As a consequence our conceptual framework provides a detailed account of the extrinsic meaning from different perspectives on models. Klir \\[[@B38]\\] also classifies the knowledge about investigated systems. He establishes what he calls \"epistemological levels of systems\" which are very similar to Zeigler's system specification levels. In fact, Zeigler starts his presentation with a review of Klir's levels and shows the correspondence with his approach \\[[@B37]\\], p.11ff.\n\nThe SemSim (for \"semantic simulation\") project \\[[@B39]\\] aims to support integration of bio-models by means of their semantics. In SemSim models are annotated from the structural perspective with links to different biological ontologies \\[[@B40]\\]. Additional, they use the Ontology of Physics for Biology (OPB, \\[[@B41]\\]) to describe the physical quantity represented by a model variable.\n\nIn \\[[@B42]\\] there is a distinction between function as mediating between structure and behaviour and function as purpose. The first determines the \"structural behaviours\", i.e. all possible behaviours the model is able to show. The second restricts the possible behaviours to the \"expected behaviours\" which are intended by the modeller making function \"the bridge between human intention and physical behavior of artifacts\" \\[[@B43]\\], p.271. The distinction between structural and expected behaviours originates from \\[[@B44]\\]. In this paper function is seen as purpose. Thus, the behaviour perspective describes expected behaviours.\n\nConclusion\n==========\n\nIn this paper, we present a systematic in-depth account of the semantics of bio-models. We show, that the meaning of bio-models has intrinsic and extrinsic aspects which can be viewed at from three perspectives: the structure, the function, and the behaviour of the model. The resulting six meaning facets provide a conceptual framework for the formalisation of the knowledge involved in building and using bio-models.\n\nThe proposed conceptual framework is a suitable foundation for computer-aided annotation, integration, and retrieval of bio-models. Obviously, this is only a first step in solving the \"semantic puzzle\" of formalising the meaning of bio-models. The framework helps in identifying how do the missing pieces look like and how they are fit together.\n\nOur meaning facets are also a way for structuring and clarifying our understanding of bio-models. They can guide the model builder during the model building process and can assist the model user in comprehending models. In fact, the meaning facets framework establishes a new methodology for computer-aided collaborative modelling in Systems Biology.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nCK originally developed the conceptual framework in continual discussion with the other authors. CB helped to theoretically analyse the knowledge involved in modelling and how to formalise it. NLN brought in the demand and relevance of semantics for bio-models. PD helped in structuring the meaning facets and with the case study. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nWe acknowledge financial support by the German Rosa Luxemburg Foundation (PhD scholarship), by the Marie Curie BIOSTAR (MEST-CT-2004-513973) and by the German Research Foundation priority program InKoMBio (SPP 1395, Grant DI 852/10-1). This work is part of the Computer Supported Research (CoSRe) initiative funded by Th\u00fcringer Ministerium f\u00fcr Bildung, Wissenschaft und Kultur under grant 12038-514.\n"}
+{"text": "\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nBilateral symmetry, characteristic of vertebrates, is extremely important for describing the morphology of the masticatory organ. This feature is strongly marked in the craniofacial area and is an important determinant of its correct structure. Assessment of bilateral symmetry in the craniofacial area is a fundamental component of the examination and description of people with and without disorders. It must be stated, however, that slight facial asymmetry is acceptable, being a common and frequently observed morphological feature \\[[@B1], [@B2]\\]. Such a disruption of symmetry is not a significant exception but a commonly accepted structural deviation. Unfortunately, the extent of acceptable craniofacial asymmetry has not been clearly defined. The concept of the bilateral symmetry of the human body is also connected with functional symmetry. An assessment of this feature in the craniofacial area is primarily related to the function of the largest and strongest facial bone, namely, the mandible. The symmetrical function of this bone, which is the single and only movable bone of the skull, is determined by two morphologically coupled temporomandibular joints.\n\nIn this context, a harmonised relationship between the dental arches is essential for maintaining functional symmetry. Malocclusion, particularly of the transverse type where disrupted symmetry of dental arches can be clinically observed, is a potential cause of functional disorders \\[[@B3]--[@B6]\\]. A priori knowledge clearly indicates the impact of the relationship between dental arches on the function of the masticatory organ \\[[@B7]--[@B10]\\].\n\nThe aim of this study was to assess the influence of unilateral posterior crossbite on the electrical activity of the temporal and masseter muscles in patients with subjective symptoms of temporomandibular dysfunctions (TMD).\n\n2. Materials and Methods {#sec2}\n========================\n\nFifty patients (22 women and 28 men) aged between 18.4 and 26.3 years (mean 20.84, SD 1.14) with subjective symptoms of TMD (Ai II-III) and unilateral posterior crossbite malocclusion were selected from patients referred to the Pomeranian University in Szczecin, Poland. The control group consisted of 100 subjects (54 women and 46 men) aged between 19.5 and 28.7 years (mean 21.42, SD 1.06) with no malocclusion and subjective symptoms of TMD (Ai I). Patients who had already finished their orthodontic treatment and those who were undergoing treatment at the time of the study were excluded.\n\nThe anamnestic interviews included the patients\\' general medical history as well as detailed information about their masticatory motor system. They were conducted according to a three-point anamnestic index of temporomandibular dysfunction---Ai ([Table 1](#tab1){ref-type=\"table\"}) \\[[@B11], [@B12]\\].\n\nThe assessment of the function of the masticatory motor system included clinical examination and electromyographic procedures. Clinical examination consisted of visual and auscultatory assessment as well as palpation. This made it possible to accurately and precisely evaluate the function of the masticatory system. Data obtained from the clinical study was analysed using the clinical temporomandibular dysfunction index (Di).\n\nAll the patients gave their informed consent to all of the procedures performed.\n\nEMG recordings were performed using a DAB-Bluetooth Instrument (Zebris Medical GmbH, Germany). Each patient was sitting on a comfortable chair without head support and was requested to assume a natural head position.\n\nSurface EMG signals were detected by four silver/silver chloride (Ag/AgCl), disposable, self-adhesive, bipolar electrodes (Noraxon Dual Electrode, Noraxon, USA) with a fixed interelectrode distance of 20\u2009mm. The electrodes were accurately positioned on the anterior temporal muscle and the superficial masseter on both the left and the right sides parallel to the muscular fibres. Anterior temporal muscle is vertically along the anterior margin of the muscle; masseter muscle is parallel to the muscular fibres with the upper pole of the electrode at the intersection between the tragus-labial commissura and exocanthion-gonion lines. A reference electrode was applied inferior and posterior to the right ear \\[[@B13]\\].\n\nTo reduce skin impedance, the skin was cleaned with 70% ethyl alcohol and dried prior to the placement of the electrode. The recordings were performed 5 minutes later.\n\nEMG activity was then recorded during three different tests.Rest activity of the masticatory muscles was performed in the clinical rest position.Maximum voluntary clench (MVC) was performed in the intercuspal position and the subject was asked to clench as hard as possible for 5 seconds.Maximum voluntary clench (MVC) was performed with two 10\u2009mm thick cotton rolls positioned on the mandibular second premolars and molars and the subject was asked to clench as hard as possible for 5 seconds.\n\nTo avoid any effects of fatigue, a rest period of at least 5 minutes was allowed between each of the recordings.\n\nFor each muscle, the EMG potentials were expressed as a percentage of the MVC value using cotton rolls (unit *\u03bc*V/*\u03bc*V%). This kind of standardization should obviate any variability due to skin and electrode impedance, electrode positioning, and relative muscular hypo- or hypertrophy \\[[@B14]--[@B16]\\].\n\nIn the current study, muscular coordination and symmetry of the masticatory muscles were expressed through the use of indices.\n\nThe asymmetry between the activity of the left and right jaw muscles was quantified by the Asymmetry Index (As, unit %). This ranges from 0% (total symmetry) to 100% (total asymmetry):$$\\begin{matrix}\n{\\text{As} = \\frac{\\sum_{i = 1}^{N}\\left| {R_{i} - L_{i}} \\right|}{\\sum_{i = 1}^{N}\\left( {R_{i} + L_{i}} \\right)} \\cdot 100.} \\\\\n\\end{matrix}$$\n\nTo assess the presence of a possible laterodeviating effect on the mandible during the test caused by unbalanced TR and ML and TL and MR couples, the Torque coefficient (Tc, unit %) was calculated as follows:$$\\begin{matrix}\n{\\text{Tc} = \\frac{\\sum_{i = 1}^{N}\\left| {\\left| {TR + ML} \\right|_{i} - \\left| {TL + MR} \\right|_{i}} \\right|}{\\sum_{i = 1}^{N}\\left\\lbrack {\\left\\lbrack {TR + ML} \\right\\rbrack + \\left\\lbrack {TL + MR} \\right\\rbrack} \\right\\rbrack_{i}} \\cdot 100.} \\\\\n\\end{matrix}$$Tc ranges from 0%, no torque during the test, to 100%, a significant laterodeviating effect on the mandible \\[[@B17]--[@B19]\\].\n\nThe Kruskal-Wallis test and the Mann-Whitney *U* test were used to verify the hypotheses relating to the existence or absence of differences between the mean values of the independent variables. The level of significance was set at *P* = 0.05.\n\nThe research was approved by the Ethics Committee of the Pomeranian Medical University in Szczecin (number BN-001/45/07).\n\n3. Results {#sec3}\n==========\n\nThe analysis of the results of EMG recordings confirmed the influence of unilateral posterior crossbite on the variability of muscle activity in the mandibular rest position ([Table 2](#tab2){ref-type=\"table\"}, [Figure 1](#fig1){ref-type=\"fig\"}). The rest activity of the temporal muscles was higher in subjects with crossbite and subjective symptoms of TMD (7.11\u2009*\u03bc*V/*\u03bc*V%, *P* \\< 0.0249) compared with healthy subjects (4.07\u2009*\u03bc*V/*\u03bc*V%). There were no significant differences in the rest activity of the masseter muscles in either examined group (*P* \\< 0.5902).\n\nThe results showed a significant increase in the Asymmetry Index in relation to both the rest activity of the temporal (29.30%, *P* \\< 0.0001) and masseter muscles (38.07%, *P* \\< 0.0006) in patients with unilateral posterior crossbite ([Figure 2](#fig2){ref-type=\"fig\"}).\n\nAdditionally, a significant increase was observed in the torque for the pair of muscles responsible for the lateral functional shift of the mandible in the rest position in patients with crossbite (14.56%, *P* \\< 0.0002, [Figure 3](#fig3){ref-type=\"fig\"}).\n\nThe differences presented regarding asymmetry in spontaneous muscle activity between the two examined groups were confirmed only for women.\n\nAn analysis of the EMG recordings during MVC confirmed the influence of transversal malocclusion on the activity of the masticatory muscles ([Table 3](#tab3){ref-type=\"table\"}, [Figure 1](#fig1){ref-type=\"fig\"}). In patients with unilateral posterior crossbite a significant decrease in the activity of the temporal (91.59\u2009*\u03bc*V/*\u03bc*V%, *P* \\< 0.0000) as well as masseter muscles (97.08\u2009*\u03bc*V/*\u03bc*V%, *P* \\< 0.0000) in relation to subjects without malocclusion (temporal and masseter muscles 116.27\u2009*\u03bc*V/*\u03bc*V% and 131.99\u2009*\u03bc*V/*\u03bc*V%, resp.) was observed.\n\nThe importance of transverse defects was also reflected in significantly higher rates of muscle asymmetry for the temporal (17.96%, *P* \\< 0.0000) and masseter muscles (17.10%, *P* \\< 0.0000, [Figure 2](#fig2){ref-type=\"fig\"}) during maximum isometric contraction in patients with unilateral posterior crossbite. Analysis revealed a considerable imbalance in the torque for the pair of muscles responsible for the lateral functional shift of the mandible in patients with crossbite (4.33%, *P* \\< 0.0380, [Figure 3](#fig3){ref-type=\"fig\"}).\n\n4. Discussion {#sec4}\n=============\n\nIn our study, an analysis of the influence of the relationships between dental arches on the electrical activity of muscles was performed with respect to two functions: in the mandibular rest position and during maximum isometric contraction. The results confirmed the significant impact of transversal malocclusions on the electrical activity of the temporal and masseter muscles in patients with subjective symptoms of TMD. In the analysis of the results of this scientific experiment, in addition to measurements which were specific for each research method, quotient indicators such as the Asymmetry Index and the Torque coefficient were used. The use of these mathematical tools made it possible to significantly increase the possibilities of describing the biomedical reality. As a result, the unbalanced torque of the two pairs of muscles responsible for the functional lateral shift of the mandible in subjects with unilateral posterior crossbite was observed. This was consistent with the clinical observations. Increased asymmetry in spontaneous muscles activity was revealed only in women and may suggest a higher sensitivity of this examined group for asymmetry of dental arches.\n\nA review of the literature presented by McNamara et al. \\[[@B20]\\] indicates that there are relatively weak links between function and the alignment of dental arches. Only five occlusal features such as skeletal anterior open bite, overjet greater than 6 to 7\u2009mm, retruded cuspal position/intercuspal position slides greater than 4\u2009mm, unilateral lingual crossbite, and five or more missing posterior teeth have been associated with functional disorders of the masticatory motor system.\n\nMohlin et al. \\[[@B21]\\] in a methodical review of 58 studies on the correlations between symptoms of TMD and malocclusions found that there were small differences in terms of functional disorders between subjects with and without malocclusions and thus the authors critically assessed these studies. Moreover, they confirmed a lack of unanimity in the denotation of correlations between function and specific types of malocclusion.\n\nThe studies conducted by Egermark-Eriksson et al. \\[[@B22]\\] in a group of 238 subjects aged 7 to 15 years also showed weak association between functional disorders and malocclusions. The confirmation of this thesis was provided by a lack of variability in the frequency of occurrence for the symptoms of functional disorders in the group who had received orthodontic treatment compared with the group of subjects without such treatment. Nevertheless, with regard to some malocclusions such as crossbite, both uni- and bilateral, anterior open bite, and post- and prenormal occlusion, a higher risk of developing functional disorders was recorded.\n\nLater, a 20-year follow-up by Egermark et al. \\[[@B23]\\] and Magnusson et al. \\[[@B24]\\] showed weak correlations between malocclusions and both symptoms and signs of TMD in a group of 402 subjects. Only unilateral crossbite was correlated with symptoms of TMD (*r* = 0.34, *P* \\< 0.01). Subjects with malocclusion over a long period of time tended to report more subjective symptoms of dysfunctions and to show a higher dysfunction index, compared with subjects with no malocclusion.\n\nAn analysis of masticatory muscle activity in patients with altered occlusal relationships due to malocclusion was the subject of studies conducted by Ferrario et al. \\[[@B25]\\]. The examined group consisted of 10 subjects aged 16--18 years, with posterior unilateral crossbite, bilateral angle Class I, and an overjet and overbite between 2 and 5\u2009mm. The control group consisted of 20 subjects with healthy dentition and with no malocclusion. Electromyographic recordings of masticatory muscles during chewing were performed. The findings of the study showed a decisive influence of crossbite on the electrical activity of the temporal and masseter muscles, manifesting itself in their disturbed coordination. Moreover, the functional changes were more apparent when the side with the altered transversal relationships was directly involved. Similar results were obtained by Rilo et al. \\[[@B26], [@B27]\\].\n\nThe relationship between transversal malocclusions and the electrical activity of the muscles was described by Alarc\u00f3n et al. \\[[@B28]\\]. Electromyographic recordings of anterior and posterior temporal and masseter muscles as well as anterior digastric muscles in 30 subjects with unilateral posterior crossbite and in a control group of 30 normocclusive subjects were made at rest position, during swallowing and during mastication. The results of the study revealed that the posterior temporal muscle on the noncrossbite side was more active than that of the same side in subjects with crossbite at rest position and during swallowing. The activity of both anterior digastrics was higher in subjects with crossbite during swallowing. Moreover, during chewing the masseter muscle was less active in patients with crossbite than in the subjects in the control group. The similar findings were reported by Kecik et al. \\[[@B29]\\]. The influence of transversal malocclusions on function was confirmed in the prospective electromyographic studies conducted by Sohn et al. \\[[@B30]\\]. The authors obtained an improvement in masticatory efficiency after orthodontic treatment of anterior crossbite. The duration of muscle activity and the incidence of silent periods in the superficial part of the masseter muscle during chewing in fact decreased after treatment. There were no significant differences in the electrical activity of the anterior and posterior temporal muscles before and after treatment.\n\nSaifuddin et al. \\[[@B31]\\] assessed the electrical activity of the temporal and masseter muscles in patients with lateral deviations of the mandible (from 5 to 14\u2009mm), crossbite, crowding, and those with no subjective symptoms in the masticatory motor system. The control group consisted of subjects without significant craniofacial asymmetry (acceptable range from 0 to 3\u2009mm), malocclusions, or subjective symptoms of functional disorders. The electromyographic recordings included not just selected activities but also the full daily activity including speech, eating, drinking, and sleeping. The analysis of muscle activity was divided into three periods: ordinary daily activities, mealtimes, and sleeping. The results showed that muscle activity in patients with disorders was significantly lower during all three periods for the masseter muscle and during ordinary daytime activities for the temporal muscle, in comparison to the control group. The Asymmetry Index (AI) in patients with a lateral shift of the mandible was significantly greater during usual daytime activities and sleep for the temporal muscle and significantly smaller during sleep for the masseter muscle, in comparison to the control group. The results clearly revealed that the asymmetry in the electrical activity during ordinary daytime activities and sleep in patients with lateral deviations of the mandible to a greater extent affects temporal muscles (anterior part) than masseter muscles. According to the authors, a reduction in temporal and masseter muscle electrical activity, with the accompanying asymmetry in the electrical activity of the temporal muscles, is closely related to occlusal instability due to malocclusions and lateral mandibular deviation.\n\nA review of the literature presented does not indicate a clear association between malocclusions and TMD. However, the results of the aforementioned studies suggest a higher risk of the prevalence of TMD in patients with unilateral posterior crossbite.\n\n5. Conclusions {#sec5}\n==============\n\nThe use of sEMG in the assessment of the function of the masticatory motor system provided tangible evidence of the determining influence of unilateral posterior crossbite on the electrical activity of the temporal and masseter muscles in patients with subjective symptoms of TMD.\n\nConflict of Interests\n=====================\n\nThe authors declare that there is no conflict of interests regarding the publication of this paper.\n\n![Electrical activity of muscles at clinical mandibular rest position and at maximal voluntary contraction (MVC) in intercuspal position depending on transversal malocclusion.](BMRI2015-259372.001){#fig1}\n\n![Asymmetry Index (As) of muscles at clinical mandibular rest position and at maximal voluntary contraction (MVC) in intercuspal position depending on transversal malocclusion.](BMRI2015-259372.002){#fig2}\n\n![Torque coefficient (Tc) of muscles at clinical mandibular rest position and at maximal voluntary contraction (MVC) in intercuspal position depending on transversal malocclusion.](BMRI2015-259372.003){#fig3}\n\n###### \n\nAnamnestic index of temporomandibular dysfunction (Ai).\n\n  Ai    Symptoms\n  ----- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  I     No subjective symptoms of temporomandibular dysfunction: no symptoms reported by patient.\n        \n  II    Mild symptoms of temporomandibular dysfunction: temporomandibular joint noise, feeling of \"jaw fatigue\" (fatigue of masticatory muscles), and feeling of \"jaw rigidity\" (increased tone of masticatory muscles).\n        \n  III   Severe symptoms of temporomandibular dysfunction: restricted mouth opening, painful lower jaw movements, temporomandibular joint pain, masticatory muscle pain, temporomandibular joint luxation, and lockjaw.\n\n###### \n\nElectrical activity of muscles at clinical mandibular rest position depending on transversal malocclusion.\n\n  Region                      Variable                             Gender    Group                                   \n  --------------------------- ------------------------------------ --------- ------- ------- ------- ------- ------- -------\n  Temporal muscles            Electrical activity \\[*\u03bc*V/*\u03bc*V%\\]   Females   54      4.07    2.02    22      7.08    6.45\n  Males                       46                                   4.08      2.06    28      7.14    5.10            \n  Total                       100                                  4.07      2.03    50      7.11    5.67            \n  Asymmetry index \\[%\\]       Females                              54        21.20   9.48    22      36.35   12.71   \n  Males                       46                                   17.12     9.00    28      23.77   15.60           \n  Total                       100                                  19.33     9.44    50      29.30   15.59           \n                                                                                                                     \n  Masseter muscles            Electrical activity \\[*\u03bc*V/*\u03bc*V%\\]   Females   54      2.12    0.89    22      1.78    0.98\n  Males                       46                                   2.13      1.09    28      2.10    0.57            \n  Total                       100                                  2.12      0.98    50      1.96    0.79            \n  Asymmetry index \\[%\\]       Females                              54        23.03   14.97   22      45.78   19.99   \n  Males                       46                                   31.63     14.14   28      32.01   17.37           \n  Total                       100                                  26.99     15.15   50      38.07   19.63           \n                                                                                                                     \n  Temporal/masseter muscles   Torque coefficient \\[%\\]             Females   54      7.20    7.66    22      19.43   13.53\n  Males                       46                                   4.37      3.95    28      10.73   12.83           \n  Total                       100                                  5.90      6.36    50      14.56   13.72           \n\n###### \n\nElectrical activity of muscles at maximal voluntary contraction (MVC) in intercuspal position depending on transversal malocclusion.\n\n  Region                      Variable                             Gender    Group                                     \n  --------------------------- ------------------------------------ --------- ------- -------- -------- ------- ------- -------\n  Temporal muscles            Electrical activity \\[*\u03bc*V/*\u03bc*V%\\]   Females   54      117.86   28.11    22      82.88   19.33\n  Males                       46                                   114.40    27.06   28       98.43    34.86           \n  Total                       100                                  116.27    27.54   50       91.59    29.84           \n  Asymmetry index \\[%\\]       Females                              54        9.95    8.18     22       22.10   11.14   \n  Males                       46                                   8.48      5.73    28       14.71    9.81            \n  Total                       100                                  9.27      7.16    50       17.96    10.95           \n                                                                                                                       \n  Masseter muscles            Electrical activity \\[*\u03bc*V/*\u03bc*V%\\]   Females   54      131.06   26.23    22      76.72   37.22\n  Males                       46                                   133.07    33.94   28       113.08   42.73           \n  Total                       100                                  131.99    29.89   50       97.08    43.96           \n  Asymmetry index \\[%\\]       Females                              54        9.07    5.23     22       21.72   12.81   \n  Males                       46                                   10.71     6.18    28       13.48    5.01            \n  Total                       100                                  9.83      5.72    50       17.10    10.06           \n                                                                                                                       \n  Temporal/masseter muscles   Torque coefficient \\[%\\]             Females   54      2.77     1.88     22      6.62    4.44\n  Males                       46                                   3.16      2.35    28       2.54     1.49            \n  Total                       100                                  2.95      2.11    50       4.33     3.72            \n\n[^1]: Academic Editor: Mieszko Wieckiewicz\n"}
+{"text": "Background\n==========\n\nChitin, one of the essential fungal cell wall components, is a \u03b2-1,4 linked *N*-acetylglucosamine (GlcNAc) homopolymer. Since it is absent in plants and mammals and has an important structural role in the fungal cell wall, its biosynthesis is recognized as a valuable target for fungicides \\[[@B1]\\]. Chitin is synthesized by multiple membranous isoenzymes called chitin synthases (CHS) \\[[@B2]\\]. In *Saccharomyces cerevisiae*, the organism in which chitin biosynthesis has been most studied, three differentially expressed genes code for three different proteins belonging to the family of glycosyltransferases 2 \\[[@B3],[@B4]\\]. CHS1p, is responsible for only 10% of the *in vivo*chitin pool, but accounts for most of the chitin synthase activity determined *in vitro*\\[[@B4]\\]. This activity is enhanced by trypsin and GlcNAc \\[[@B2]\\].\n\nCHS are processive glycosyltransferases capable of successively transferring GlcNAc monomers from the UDP-GlcNAc donor to a growing polymer acceptor. Elongation takes place at the non reducing end of the polymer \\[[@B5],[@B6]\\]. CHS isoenzymes belong to a subfamily of processive enzymes involved in polysaccharide biosynthesis such as cellulose and hyaluronan synthases \\[[@B7],[@B8]\\]. The 3D-structure of a non-processive member of glycosyltransferases family 2 had shed some light on both their donor and acceptor sites \\[[@B9]\\]. In contrast, due to the lack of structures, donor and acceptor sites have not been delineated in the case of polysaccharide synthases. While the donor sites can be inferred from sequence comparison between processive and non-processives glycosyltransferases, the way polysaccharide synthases accommodate the growing polysaccharide is totally unknown.\n\nAs for any polymer, chitin formation is expected to require chain initiation, elongation and termination. Several mechanisms of polysaccharide priming by oligosaccharides, either in a free or conjugated form, have been described, specifically Carbohydrate polymerase GlfT2 from *Mycobacterium tuberculosis*acts on glycolipid acceptors for the galactan polymer synthesis \\[[@B10]\\]. Also, experiments on cellulose synthase from cotton fiber led to the proposal of a model in which cellulose synthesis would be initiated by a sitosterol-\u03b2-glucoside \\[[@B11]\\]. Important outcomes of this model were the fact that a certain number of discrete steps should precede the final elongation, and the existence of an oligoglucoside species as a discrete intermediate. Free glycoside utilisation, as primers, are described for other glycosyltransferases, such as \u03b2-D-xylosides used for glycosaminoglycans chain biosynthesis on proteoglycan core proteins \\[[@B12]\\]. Other polysaccharide synthases responsible for the addition of a sugar monomer to oligosaccharide acceptors have been described for hyaluronan \\[[@B13]\\], chondroitin sulfate \\[[@B14]\\], heparan sulfate \\[[@B15]\\], glycogen \\[[@B16]\\] and heparosan \\[[@B17]\\]. Also, high concentrations of GlcNAc were shown to prime *N*-acetylchitopentaose synthesis by the oligosaccharide synthase NodC \\[[@B6]\\] for the synthesis of rhizobial lipo-chitin oligosaccharides (Nod factors). Concerning chitin synthase, a priming by GlcNAc has never been observed and GlcNAc is thought to be an allosteric activator \\[[@B18]-[@B20]\\]. McMurrough and Bartnicky-Garcia tested a range of compounds structurally related to GlcNAc but none successfully substituted for GlcNAc in the allosteric activation of chitin synthesis, although *N*-acetylchitooses exerted a slight stimulation of substrate incorporation supposed to a different mechanism as GlcNAc \\[[@B21]\\].\n\nIn order to elucidate the role of oligosaccharides, a systematic study was carried out with preformed oligosaccharides on a chitin synthase assay on *S. cerevisiae*microsomal preparations. We used oligosaccharides of various degrees of polymerization (DP) and from different structures as well as *N*-acetylchitooses (\u03b2-(1,4) GlcNAc units) or \u03b2-(1,4), \u03b2-(1,3) or \u03b1-(1,6) glucooligosaccharides.\n\nMethods\n=======\n\nAll reagents were purchased from Sigma and Merck Biochemicals, unless indicated otherwise in the text. UDP-*N*-acetyl\\[^14^C\\]glucosamine was from NEN Life Sciences Products. *N*-acetylchitooses were either obtained from Dr R. Geremia (CERMAV Grenoble, France) or purchased from Sigma. Both batches were of identical purity (\\>95%).\n\nPreparation of membrane fractions\n---------------------------------\n\n*Saccharomyces cerevisiae*cells (baker\\'s yeast, 10 g) were washed and resuspended in 10 mL of 50 mM Tris-HCl (pH 7.4) - 2.5 mM MgCl~2~. After addition of glass beads (0.25-0.50 mm, 10 mL), cells were broken by shaking on a vortex for 7 cycles of 30 s alternated with 30 s cooling on ice. After decantation of the solution, the supernatant was recovered, centrifuged at 1000 \u00d7 g for 10 min and then at 12 000 \u00d7 g for 10 min to remove mitochondria. Membrane fractions were isolated from this extract by centrifugation at 100 000 \u00d7 g for 1 h. The membrane pellet was resuspended in approximately 500 \u03bcl of 50 mM Tris-HCl (pH 7.4), 30% glycerol. Membrane preparations were conserved at -80 \u00b0C at protein concentration around 60 mg/ml, as determined by Bradford protein assay (Biorad) \\[[@B22]\\].\n\nPreparation of laminarin and dextran oligosaccharides\n-----------------------------------------------------\n\nLaminarin and dextran (1 g of each) were dissolved in 10 ml 0.1 N HCl and acid hydrolysis performed at ebullition for 4 h. Mixtures were diluted with water, neutralized with Dowex 2 \u00d7 8-200 resin (hydroxyl charged) and lyophilized. Oligosaccharides with DP from 3 to 6 were obtained after Bio-gel P2 chromatography purification as previously described for *N*-acetylchitooses \\[[@B23]\\].\n\nChitin synthase assay\n---------------------\n\nUnder our assay conditions, the main enzymatic activity is CHS1p activated by trypsin \\[[@B4]\\]. At 150 \u03bcg of *S. cerevisiae*membrane fractions containing chitin synthase activity a common standard reaction mix (in a total volume of 50 \u03bcl: 5 \u03bcl digitonin 2%, 5 \u03bcl magnesium acetate 50 mM, 5 \u03bcl trypsin 0,2 mg/mL (10000 units/mg protein, Sigma), 2 \u03bcl *N*-acetylglucosamine 1 M, 2.5 \u03bcl UDP-*N*-acetylglucosamine 10 mM, 1 \u03bcl UDP-*N*-acetyl\\[^14^C\\]glucosamine (10 nCi, 288 mCi/mmol) in Tris-HCl buffer 25 mM pH = 7.4) was added. Incubation was performed at 30\u00b0C for 10 min, the reaction was stopped with 1 ml of 10% trichloroacetic acid. The radioactivity in the insoluble fraction was counted after filtration through glass-fiber filter (Whatman, GF/C). The soluble fraction containing unincorporated UDP-*N*-acetyl\\[^14^C\\]glucosamine was discarded. The filter was washed three times with 1 ml of 1 M acetic acid/ethanol: 30/70 and once with 1 ml of 95% ethanol. The filter was transferred to a scintillation vial containing 4 ml of Optiphase Hisafe (Wallac) and the samples were counted in a scintillation counter (LKB 1214 RackBeta).\n\nTo determine the effect of adding different oligosaccharides chain lengths to the chitin synthase assay, membrane fractions were preincubated with either \u03b2-(1,3), \u03b2-(1,4) or \u03b1-(1,6) glucooligosaccharides or *N*-acetylchitoses (from bioses to octaoses) in 25 mM Tris-HCl pH 7.4 for 15 min at 30\u00b0C before addition of the common standard reaction mix. Oligosaccharides were used at a concentration of 1 mM unless otherwise stated.\n\nChitin digestion by chitinase\n-----------------------------\n\n\\[^14^C\\]-chitin, prepared as described above, was digested by *Streptomyces Griseus*chitinase (5.10^-3^U, 0.7 U/mg, Sigma) at 37\u00b0C for 24 h in phosphate buffer 2 mM pH = 6. Samples were boiled for 5 min and 1/5 of the sample was applied to TLC plates. Samples were analysed on precoated TLC plates silica gel 60, Merck. The chromatogram was developed with 1-propanol/water/ammonium hydroxide 30% (70:30:1.5, vol/vol/vol). Radioactive spots were detected by means of an automatic TLC-linear analyzer (LB 2821 Berthold).\n\nEffect of GlcNAc and trypsin\n----------------------------\n\nThe *N*-acetylchitose activation effect was tested in absence, or at low (1 mM) GlcNAc concentration as described above. Alternatively, activation by trypsin was carried out before polymerization. In this case, membrane fractions were preincubated at 30\u00b0C with trypsin from bovine pancreas (1 \u03bcg, 10000 units/mg protein, Sigma) in 25 mM Tris-HCl pH 7.4. After 15 min, soybean trypsin inhibitor (2 \u03bcg, Sigma) was added and incubation followed for 15 min.\n\nResults and Discussion\n======================\n\nCatalytic activity of CHS\n-------------------------\n\nCHS activity assays were performed using *Saccharomyces cerevisiae*microsomal preparations by measuring the rate of formation of \\[^14^C\\]-chitin from UDP-*N*-acetyl\\[^14^C\\]glucosamine \\[[@B4]\\]. Oligosaccharides of different nature and length were preincubated with microsomes before the chitin synthase assay was performed. Unless otherwise stated, trypsin was routinely included as a required activator in the common reaction mix. The product of the reaction was unambiguously identified as chitin by specific digestion with commercial chitinase and chromatographic characterization of released GlcNAc units (Figure [1](#F1){ref-type=\"fig\"}).\n\n![**Thin layer chromatography analysis of \\[^14^C\\]-chitin, produced *in vitro*by *S. cerevisiae*membrane fractionation**. The CHS reaction product was digested by *S. Griseus*chitinase and analysed on silica 60 TLC plates as described in Materials and Methods. Samples were loaded onto the TLC plate before (panel A) or after 24 h incubation with commercial chitinase (panel B). In panel C the reference sample containing commercial \\[^14^C\\]-GlcNAc was applied.](1756-0500-4-454-1){#F1}\n\nThe activating effect of *N*-acetylchitoses on chitin synthesis was studied over time with 0.5 mM UDP-GlcNAc and 1 mM *N*-acetylchitose concentrations. These concentrations are comparable to the UDP-GlcNAc Km of 0.25 mM displayed by yeast CHS 1 \\[[@B4],[@B24]\\]. Initial linear portions of the kinetics obtained in presence and absence of *N*-acetylchitopentaoses were compared and initial velocity values were calculated (Figure [2](#F2){ref-type=\"fig\"}). We obtained, 2.3 \u00b1 0.1 nmol GlcNAc incorporated.min^-1^.mg protein^-1^in the presence, and 1.9 \u00b1 0.09 nmol-GlcNAc incorporated.min^-1^.mg protein^-1^in the absence of *N*-acetylchitopentaoses. Thus, the initial velocity increase is about 20% and similar activation was observed for *N*-acetylchito -tetra or octaoses (Table [1](#T1){ref-type=\"table\"}).\n\n![**Time course of chitin synthesis in presence or absence of *N*-acetylchitopentaoses**. Kinetics of two chitin synthase assays following preincubation with (open boxes, solid line) or without (black circles, dotted line) 1 mM *N*-acetylchitopentaoses. The linear part of the curve (up to 12 min) is shown.](1756-0500-4-454-2){#F2}\n\n###### \n\nPer cent activation of \\[^14^C\\]-GlcNAc incorporation into chitin polymer in presence of oligosaccharides from different lengths and structures.\n\n  --------------------------------------------------------------------------\n  DP   *N*-acetyl-chitooses\\   Laminari-oses\\   Dextro-oses\\   Cello-oses\\\n       \u03b2-(1,4)-Glc-NAc         \u03b2-(1,3)-Glc      \u03b1-(1,6)-Glc    \u03b2-(1,4)-Glc\n  ---- ----------------------- ---------------- -------------- -------------\n  2    13,3% \u00b1 2               2% \u00b1 1,1         nd             nd\n\n  3    10% \u00b1 1,5               1,6% \u00b1 1,2       2% \u00b1 1,3       nd\n\n  4    17,3% \u00b1 0,8             0,5% \u00b1 0,5       1,5% \u00b1 1,2     nd\n\n  5    23,3% \u00b1 1,1             1% \u00b1 0,8         2,1% \u00b1 1,5     3% \u00b1 2\n\n  8    24,4% \u00b1 1,2             nd               nd             nd\n  --------------------------------------------------------------------------\n\nChitin synthase activity of *S. cerevisiae*membrane fractions was measured as incorporation of \\[^14^C\\]-GlcNAc into a filter-retained insoluble material. Assays were performed upon preincubation of microsomes with *N*-acetylchitooses \u03b2-(1,4)-GlcNAc, laminarioses \u03b2-(1,3)-Glc, dextrooses \u03b1-(1,6)-Glc, or cellooses \u03b2-(1,4)-Glc. Length of oligosaccharides tested varied from bioses to octaoses (DP = degree of polymerization). Per cent activations were calculated compared to standard assay without added oligosaccharides. Results are the average of five independent experiments (\u00b1SD). Activation for some oligosaccharides was not determined (nd).\n\nEffect of oligosaccharide structure and length on chitin synthesis\n------------------------------------------------------------------\n\n*N*-acetylchito-bi and *-*trioses in the CHS test mixture had little effect on chitin synthesis (enhancement around 11%), and these results were poorly reproducible (Table [1](#T1){ref-type=\"table\"}). By contrast, *N*-acetylchito-tetra, *-*penta and *-*octaoses clearly stimulated GlcNAc incorporation from UDP-\\[^14^C\\]-GlcNAc into chitin, reaching up to 25% enhancement for the *N*-acetylchitooctaoses (Table [1](#T1){ref-type=\"table\"}). A clear-cut effect of the oligosaccharide size on the extent of activation was observed.\n\nOligosaccharides from other carbohydrate series such as laminari-(\u03b2-(1,3-gluco)), dextro-(\u03b1-(1,6-gluco))- or cello-(\u03b2-(1,4-gluco))-oligosaccharides did not stimulate \\[^14^C\\]-GlcNAc unit incorporation into \\[^14^C\\]-chitin (Table [1](#T1){ref-type=\"table\"}). The lack of effect of oligosaccharides from other series argues in favor of a specific effect of GlcNAc oligomers. Moreover, in order to achieve acceptable activation reproducibility, the *N*-acetylchitooses must be of at least 4 residues.\n\nEffect of chitooligosaccharides versus chitin synthase activators\n-----------------------------------------------------------------\n\nIt is known that both trypsinolysis and a high (40 mM) GlcNAc concentration enhance *in vitro*chitin synthesis \\[[@B25],[@B19]\\]. In order to test whether the activation on chitin synthesis by *N*-acetylchitooses could interfere with these two activators, we used these *N*-acetylchitooses as potent substitutes. The effect of 1 mM *N*-acetylchitopentaoses was measured in the absence of GlcNAC or at low (1 mM) and high (40 mM) GlcNAc concentration. Activation profiles, by *N*-acetylchitopentaoses, (\\~20%) were identical to that of standard assay (40 mM GlcNAc), but with a 3-fold reduction of UDP-GlcNAc incorporation (Figure [3A](#F3){ref-type=\"fig\"}). Similar results were obtained with other *N*-acetylchitoose lengths (data not shown). One can thus conclude that the *N*-acetylchitooses did not substitute for GlcNAc. Both activating effects seem to be independent, suggesting that the two species bind to two different sites in a random manner.\n\n![**Activation by *N*-acetylchitooses is independent on Glc-NAc and trypsin**. (A) \\[**^14^**C\\]-chitin was prepared and quantified as described, except that Glc-NAc was present at 40 mM (grey bars), 1 mM (black bars) or absent (hatched bars) in the incubation mixture. (B) \\[**^14^**C\\]-Glc-NAc incorporation into chitin measured after two different conditions of CHS activation by trypsin: during (standard conditions, grey bars), or before (black bars) chitin polymerization. Results in presence or absence of ***N***-acetylchitopentaoses are shown.](1756-0500-4-454-3){#F3}\n\nInterference between the effect of *N*-acetylchitooses and activation by trypsin was also studied. In standard test conditions, microsomal preparations were activated by trypsin at the beginning of the reaction. In a parallel protocol, a preactivation of the microsomal preparation by trypsin was performed, then terminated by addition of trypsin inhibitor. Initial velocity in the presence of *N*-acetylchitooses clearly reached same value, whether or not trypsinolysis occured before or during chitin polymerization (Figure [3B](#F3){ref-type=\"fig\"}). This result excluded a possible effect of chitin synthase protection by *N*-acetylchitooses against overdigestion by trypsin.\n\nEffect of oligosaccharide concentration on chitin synthase activity\n-------------------------------------------------------------------\n\nThe chitin synthesis activation effect was studied over a range of oligosaccharide concentrations from 0 to 2 mM with the *N*-acetylchito-tetra, -penta and -octaoses. It appeared that activation was a saturable phenomenon for the three *N*-acetylchitooses tested, but saturation was more clearly established with higher *N*-acetylchitoose DP (Figure [4](#F4){ref-type=\"fig\"}). In each case we determined the concentration necessary to reach half-maximum activation (AC~50~) and concluded that AC~50~values decreased as the oligosaccharide sizes increased. AC~50~were 0.9, 0.5 and 0.25 mM for *N*-acetylchito-tetra, *-*penta and -octaoses, respectively (Figure [5](#F5){ref-type=\"fig\"}). Although chitin synthesis is best characterized in *S. cerevisiae*, the mechanism of chitin biosynthesis activation or regulation is still unknown \\[[@B26],[@B25]\\]. Three CHS isoenzymes are responsible for chitin synthesis and additional CHS proteins were identified as regulators of CHS in *S. cerevisiae*\\[[@B26]\\]. It is plausible that triggering of chitin synthesis by *N*-acetylchitooses results from oligosaccharides binding a protein partner of the complex leading to CHS activation. However, the direct binding of *N*-acetylchitooses to CHS enzyme can not be excluded. Indeed, defined oligosaccharide structures and lengths are required to observe this activation, which is also concentration dependent. Chavaroche et al. proposed that the polymerization reaction is more efficient for the synthesis of heparosan chains in presence of oligosaccharide templates, because the initiation step does not take place and the elongation of the templates occurs directly \\[[@B17]\\]. All these observations reflect the filling of an increasing number of subsites constituting the acceptor binding site of CHS. This attractive hypothesis should be compared to the reaction model scheme of chitinase, for which substrate subsites were clearly identified \\[[@B27]\\]. Cloning CHS genes and expression of recombinant proteins, in the appropriate host, would be the next step to elucidate chitin synthase activation. Our preliminary results on *Botrytris cinerea*CHS, suggest that active CHS cannot be expressed in *E. coli*. The truncated recombinant CHS protein showed specific binding to UDP-GlcNAc, but was unable to exhibit chitin synthase activity \\[[@B28]\\]. The expression of full-length CHS bound to the plasmic membrane of *S. cerevisiae*seems to be more favorable or even necessary for maintenance of enzymatic activity of CHS. Use of *N*-acetylchitooses as primers for chitin chain elongation will be verified as soon as an active recombinant chitin synthase will be obtained.\n\n![**Effect of *N*-acetylchitoose concentration on GlcNAc incorporation**. Different concentrations of *N*-acetylchito-tetra (A), -penta (B), -octaoses (C) were preincubated with membrane fractions before initiating the CHS reaction. GlcNAc incorporation into chitin versus *N*-acetylchitoose concentration is shown. AC~50~corresponds to half-activating *N*-acetylchitoose concentration.](1756-0500-4-454-4){#F4}\n\n![***N*-acetylchitooses units number influence half-activation concentration**. Half-activating ***N***-acetylchitoose concentration was determined from Fig. 4 for ***N***-acetylchito-tetra, -penta and -octaoses. DP = degree of polymerization.](1756-0500-4-454-5){#F5}\n\nConclusion\n==========\n\nIn this paper, we have shown that the addition of *N*-acetylchitooses to a standard chitin synthase assay resulted in an increase of initial velocity of GlcNAc unit incorporation, specifically a 25% enhancement for *N*-acetylchitooctaoses. The effect induced by *N*-acetylchitooses is different from other known activators such as trypsin and exogenous GlcNAc. The activating effect of *N*-acetylchitoses on chitin synthesis was described as a saturable phenomenon dependant of *N*-acetylchitoose structure and length.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nHFB performed most of the experiments, designed and coordinated the experiments, wrote the manuscript and edited the final text. AP assisted with the enzymatic analysis, AT participated in the interpretation of data and drafted portions of the manuscript. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nThis paper is dedicated to the memory of our colleague, Dr Anne Vidal-Cros, who passed away in 2006. AVC developed the study of chitin synthases in the laboratory and she assisted on the first design of these experiments. We are grateful to Prof. M. Boccara, Universit\u00e9 Pierre et Marie Curie, Paris, for helpful discussions and Dr R. Geremia, Dr E. Samain, CERMAV Grenoble, for the kind gift of chitooligosaccharides.\n"}
+{"text": "1. Introduction {#s0005}\n===============\n\nLike all desert ecosystems, the desert fauna of Saudi Arabia has received much less scientific research attention than forest systems ([@b0035]). Despite the vast size and diverse topography of Saudi Arabia, studies on the ecology and distribution of its terrestrial fauna has not been explored well, especially in the northern province of the country ([@b0010]). There still remain significant gaps in our understanding of how biodiversity in these systems is changing with time due to a lack of systematic monitoring ([@b0030]).\n\nCompared to insects and reptiles, mammals are generally not well represented in desert environments -- including the Saudi Arabia. Unfortunately, published literature on the terrestrial mammals of the Saudi Arabia is still scant although a few contributions by various authors have increased the information for a number of species. [@b0070] published an updated version of the Mammals of Arabia, which today is possibly still the most comprehensive study of mammals from the Arabian Peninsula. Other earlier books on mammals include a publication by [@b0040] on the carnivores of Arabia. [@b0050], [@b0055], [@b0060], [@b0065] described in very comprehensive scientific texts the mammals of the Arabian Peninsula and the neighboring countries. His excellent zoological work gives the most valuable information regarding the mammals of the region. Fifteen carnivore's species of Arabia are discussed in great detail by [@b0045]. A total of 22 species of mammals including three domestic animals were reported within the reserve\\'s boundaries of Harrat al-Harrah ([@b0080]). [@b0025] published a survey report on Mammal diversity\u00a0in\u00a0the Ibex Reserve of Saudi Arabia in which he mentioned that the deserts of Saudi Arabia contain a number of mammalian species highly adapted for the variable conditions of the desert including the Nubian ibex (*Capra nubiana),* Arabian oryx (*Oryx leucoryx)*, Arabian grey wolf (*Canis* m*lupus arabs)* and the Blanford's fox (*Vulpes cana*).\n\nThere is a need for an updated information to estimate the species richness of mammals in Turaif as the first step to increase our knowledge of the regional mammalian fauna. The aim of this study was to fill the gap in the inventory of mammals in this important area by updating information on mammalian distribution. A survey can be an important addition to understanding the biodiversity in the region of Turaif and can provide important information on mammals of the area.\n\n2. Materials and methods {#s0010}\n========================\n\nMammal survey took place in Turaif (31\u00b040\u203239\u2033N 38\u00b039\u203211\u2033E), northern west border Province, Saudi Arabia, close to the border with Jordan, with a total area of 20,400\u202fkm^2^. It is characterized by its unique topography, geomorphology, and biodiversity, possessing several landscape types, such as Sandy habitat, Highland habitat, Mountain habitat, Sabkha (Qa'a) habitat and Wetland habitat ([@b0010]). In the study area the air temperature averages 45\u202f\u00b0C in summer to below freezing point in winter, whereas, mean annual precipitation ranges from 14.3\u202fmm (December) to 5.4\u202fmm (March) and to 0\u202fmm in June--July ([@b0015]).\n\nMammal surveys were undertaken during different seasons of the year (2014--2015) and the report was compiled. Occasional field surveys were conducted to cover the main habitats of Turaif region. The most favorable time for collection was between March and July. All mammal sightings and signs were recorded. Active ground searches were undertaken throughout the study site. The survey was principally opportunistic and based on observations using different techniques (track surveys, live traps, Binoculars Jeory 6\u202f\u00d7\u202f30 and camera captures). Traps with baits for small and average size mammals like rodents, fox, wild cats, and hedgehog (8\u202f\u00d7\u202f9\u202f\u00d7\u202f23\u202fcm) whereas for fox, wild cats and large mammal traps of dimensions (40\u202f\u00d7\u202f40\u202f\u00d7\u202f110\u202fcm) were used.\n\nThe collected specimens were deposited at College of Science, Department of Zoology, King Saud University. Date, locality, Co-ordinates: latitude, longitude and altitude, of the collected specimens were recorded by a GPS.\n\n3. Results {#s0015}\n==========\n\n[Table 1](#t0005){ref-type=\"table\"} shows the mammal species collected or observed in different habitats of Turaif province and their co-ordinates: latitude, longitude and altitude recorded by a GPS. A total of 140 specimens belonging to 15 mammal species were recorded from the study area during our survey. These 15 species belong to 9 families (Erinaceidae, Leporidae, Cricetidae, Dipodidae, Erethizontidae, Canidae, Mustelidae, Hyaenidae and Felidae). Among the 9 families Cricetidae was represented by 4 genera, Canidae by 3 genera, Dipodidae by 2 genera and remaining all families were represented by 1 genus each. From the standpoint of species richness within the described families of the surveyed region, family Cricetidae represented 26.67% (n\u202f=\u202f4) followed by Canidae 20% (n\u202f=\u202f3) of the total species (n\u202f=\u202f15) collected from the studied area. No data concerning reproduction were collected from these specimens.Table 1Co-ordinates: Latitude, longitude and altitude, of the collected specimens by a GPS.Scientific nameStatusCoordinatesJaculus jaculus vocatorHighly abundantN 31 45.469\\\nE 039 01.329N 31 46.420\\\nE 039 01.620N 31 46.118\\\nE 039 54.989N 31 46.001\\\nE 039 00.763N 31 46.072\\\nE 038 59.552N 31 44.408\\\nE 039 05.185N 31 49.363\\\nE 039 03.925N 31 43.018\\\nE 038 55.951N 31 42.749\\\nE 038 55.821N 31 47.200\\\nE 038 55.145N 31 41.182\\\nE 038 55.314N 31 44. 395\\\nE 039 00.177N 31 43.903\\\nE 038 57.466N 31 42.541\\\nE 038 58.103N 31 48.605\\\nE 038 58.557N 31 40.851\\\nE 038 58. 820N 31 40.953\\\nE 038 54.989N 31 41. 696\\\nE 039 06.472N 31 36.688\\\nE 039 06.472N 31 39.007\\\nE 038 58.433N 31 39.233\\\nE 038 58.198N 31 44.670\\\nE 039 05.186N 31 47.483\\\nE 038 52.662N 31 58.597\\\nE 039 00.968N 31 56.944\\\nE 039 01.130N 31 57.115\\\nE 038 59.867N 31 57.307\\\nE 039 01.073N 31 57.534\\\nE 039 01.698N 31 57.530\\\nE 038 58.986N 31 57.976\\\nE 038 57.140N 31 58.208\\\nE 038 58.938N 31 51.208\\\nE 038 01.938N 31 51.503\\\nE 038 59.938N 31 51.503\\\nE 038 58.976Allactaga auphraticaCommonN 31 56.944\\\nE 039 01.130N 31 57.530\\\nE 038 58.986N 31 39.007\\\nE 038 58.433N 31 44. 395\\\nE 039 00.177Gerbillus cheesmaniCommonN 31 46.519\\\nE 039 02.444N 31 46.072\\\nE 038 59.552N 31 46.200\\\nE 038 00.450N 31 44.408\\\nE 039 05.185N 31 47.119\\\nE 039 02.071N 31 46.347\\\nE 039 56.102N 31 44.551\\\nE 039 02.493N 31 43.505\\\nE 039 02.991Gerbillus nanusAbundantN 31 46.244\\\nE 039 00.336N 31 46.519\\\nE 039 02.444N 31 42.944\\\nE 038 56.102N 31 46.072\\\nE 038 59.552N 31 44.551\\\nE 039 02.493N 31 44.395\\\nE 039 06.177N 31 44.408\\\nE 039 05.185N 31 47.200\\\nE 083 55.145N 31 40.549\\\nE 038 58.820N 31 40.549\\\nE 039 06.662N 31 39.746\\\nE 039 06.734N 31 44.996\\\nE 039 04.915N 31 44.334\\\nE 039 05.062N 31 44.297\\\nE 038 57.549N 31 41.806\\\nE 039 06.684N 31 40.753\\\nE 039 06.560N 31 39.172\\\nE 038 59.963N 31 57.976\\\nE 038 59.140N 31 58.321\\\nE 039 00.714N 31 58.451\\\nE 039 00.426N 31 58.597\\\nE 039 00.968N 31 57.534\\\nE 039 01.698N 31 57.925\\\nE 038 59.902Meriones crassusCommonN 31 46.170\\\nE 038 59.784N 31 46.244\\\nE 039 00.336N 31 44.408\\\nE 039 05.185N 31 46.519\\\nE 039 02.444N 31 42.944\\\nE 038 56.102N 31 50.841\\\nE 039 04.182N 31 43.989\\\nE 039 06.109N 31 39.801\\\nE 038 59.336N 31 40.858\\\nE 39 06.537N 31 44.666\\\nE 038.58.202N 31 42.908\\\nE 038 57.856N 31 57.115\\\nE 038 59.867N 31 57.214\\\nE 039 00.315N 31 58.573\\\nE 039 00.507Meriones libycusHighly abundantN 31 44.408\\\nE 039 05.185N 31 46.519\\\nE 039 02.444N 31 46.104\\\nE 038 56.028N 31 45.849\\\nE 038 55.291N 31 43.204\\\nE 039 05.185N 31 51.438\\\nE 039 04.951N 31 44.351\\\nE 039 06.109N 31 39.801\\\nE 038 59.336N 31 40.858\\\nE 039 06.537N 31 44.888\\\nE 039 05.154N 31 47.419\\\nE 038 55.647N 31 58.458\\\nE 039 01.918N 31 48.762\\\nE 039 04.048N 31 57.564\\\nE 039 01.808N 31 58.265\\\nE 039 00.393N 31 58.321\\\nE 039 00.714N 31 58.451\\\nE 039 00.426N 31 57.214\\\nE 039 00.315N 31 58.597\\\nE 039 00.968Vulpes ruppelli sabaeaHighly abundantN 31 45.515\\\nE 039 04.931N 31 46.244\\\nE 039 00.336N 31 45.849\\\nE 038 55.291N 31 46.072\\\nE 038 59.552N 31 44.408\\\nE 039 05.185N 31 47.093\\\nE 038 54.648N 31 45.238\\\nE 039 02.513N 31 45.403\\\nE 039 02.444Vulpes vulpes arabicaHighly abundantN 31 41.696\\\nE 039 06.477N 31 38 399\\\nE 039 04.967N 31 36.754\\\nE 039 04.967N 31 40.944\\\nE 038 56.102N 31 40.851\\\nE 38 58.820N 31 44.855\\\nE 038 55.288N 31 41.057\\\nE 038 57.739N 31 44.706\\\nE 039 05.164N 31 49.697\\\nE 038 02.107N 31 47.443\\\nE 038 55.647N 31 46.954\\\nE 038 55.925Paraechinus aethiopicusCommonN 31 46.104\\\nE 039 00.400N 31 46.244\\\nE 039 00.336N 31 44.408\\\nE 039 05.185N 31 47.013\\\nE 038 54.648N 31 46.925\\\nE 038 54.393N 31 58.597\\\nE 039 00.968N 31 58.458\\\nE 039 01.918N 31 58.581\\\nE 039 01.088Lepus capensisRareN 31 58.423\\\nE 039 00.127N 31 43.839\\\nE 039 01.521N 31 46.232\\\nE 039 55.450Hystrix indicaRareN 31 58.663\\\nE 038 59.533N 31 52.553\\\nE 039 01.636Canis lupusRareN 31 46.954\\\nE 038 55.925N 31 43.244\\\nE 039 01.361Mellivera capensisRareN 31 46.925\\\nE 038 54.393Hyaena hyaenaRareN 31 46.519\\\nE 039 02.444N 31 46.954\\\nE 038 55.925Felis margaritaRareN 31 46.104\\\nE 039 00.400N 31 47.251\\\nE 038 51.925N 31 43.206\\\nE 039 02.490\n\nThe following is a checklist of the mammals collected during this survey.\n\n3.1. Family: Erinaceidae {#s0020}\n------------------------\n\nMembers of this family can easily be distinguished by the thorns spread over a large part of the skin surface except the lower sides of the body. Their snout is free of hair and constantly moving and blowing. The tail is short. The front and rear limbs ends with five clawed fingers, and the palm of hands and soles of the feet are free of hair. One species of this family was recorded in Turaif region.\n\n### 3.1.1. *Paraechinus aethiopicus* Ehrenberg, 1832 {#s0025}\n\nThis Ethiopian hedgehog is a nocturnal animal common across the Arabian Peninsula. Eight specimens of this hedgehog were collected from mountainous areas, open deserts and desert plains of the study area. Their body is covered by brown thorns with white edges giving the animal a gray color. The ventral side is covered with soft fur. It has black tapering snout and long legs. Thorns acts as a natural protection for the hedgehog against predators.\n\n3.2. Family: Leporidae {#s0030}\n----------------------\n\nLeporids, the family of rabbits and hares are small to moderately sized mammals, adapted for rapid movement. These animals have soft fur, long ears, short tails, and powerful rear legs that they use to hop and jump. The hind legs are long, with four toes on each foot, and fore legs are shorter, with five toes each. The upper lip is bifid. Females often have three to five pairs of nipples of breast glands. Only one species was recorded in Turiaf region.\n\n### 3.2.1. *Lepus capensis* Linnaeus, 1758 {#s0035}\n\nThree specimens of *L. capensis* were collected from the study area from the plateaus and mountainous habitats. These nocturnal animals prefer shrubs, rather than grasses, to shelter under in summer. Cape hare is a typical hare in appearance, with long, slender limbs, large hind feet, a short tail, large eyes and large ears. The fur is soft and straight, body color ranges from gray-brown to reddish-brown, with white ventral parts and a black and white tail. Wild rabbits live almost in all habitats and they are abundant in deserts and sandy plains.\n\n3.3. Family: Cricetidae {#s0040}\n-----------------------\n\nCricetidae is an extremely diverse and one of the largest family of muroid rodents. They have small elongated bodies, long tails, large eyes, prominent ears and are gray or brown in color. Four species of this family were recorded in Turaif region.\n\n### 3.3.1. *Gerbillus nanus* Blanford, 1875 {#s0045}\n\nTwenty-three specimens of *G. nanus* were collected from the study area in rocky areas and gravel plains. The body of *G. nanus* is covered by colored fur suitable for the environment in which they live, and characterized by naked ankles, dark brown back hairs and the tail ends with a flock of dark hair. It is found in many parts of the Arabian Peninsula, in the northern regions, along the western coast and the northern east region. It is active at sunset and night, moving to long distances in search of food.\n\n### 3.3.2. *Gerbillus cheesmani* Thomas, 1919 {#s0050}\n\nEight specimens of this species were recorded in the surveyed area. *G. cheesmani* is characterized by sandy-beige color, while the belly is white, long limbs, large black eyes and long tail. This nocturnal rodent is small in size and like many gerbil species, looks like a mouse to a large extent. It live in sandy environments, and were recorded mostly around *Calligonum comosum* plants which provide them shelter. Like other rodents, they are prey for many birds, mammals and reptiles.\n\n### 3.3.3. *Meriones libycus* Lichtenstein, 1823 {#s0055}\n\nNineteen specimens of *M. libycus* were recorded in the surveyed area. This animal was characterized by the presence of long black tail that ends in a black tuft, black claws and the body is covered with sandy colored soft and fine fur. This species is found all over large areas of the Arabian Peninsula. This rodent is diurnal living, used to live in colonies and feed on desert plant.\n\n### 3.3.4. *Merionaes crassus* Sundevall, 1842 {#s0060}\n\nFourteen specimens of *M. crassus* were recorded in the surveyed area. It is the most common rodents in the central region especially in sandy areas of Saudi Arabia. The body is covered with thick sandy colored hair, with the presence of white tufts behind the ear. The abdomin is white and the tail ends with a black tuft. It is nocturnal species and used to live in colonies, inhabits burrows around the roots of trees. As the sun sets, it goes out searching for food, which is grains and grass.\n\n3.4. Family: Dipodidae {#s0065}\n----------------------\n\nThe members of this family are distributed throughout the Arabian Peninsula. They are small sized mammals characterized by much longer hind limbs than the forelimbs, and get around by jumping. The last third of the skull is fully occupied by hearing parts, where it is evidence that these rodents are badly in need of good hearing sense. The end of limbs often covered with a hair tuft. These rodents have long tufted tails used as a fulcrum. Two species belonging to this family were recorded in Turaif region.\n\n### 3.4.1. *Jaculus jaculus vocator* Linnaeus, 1758 {#s0070}\n\nThirty-four specimens, most abundant among all mammals recorded, were collected in the surveyed area. This rodent is undoubtedly one of the commonest rodents throughout the desert areas of the Arabian Peninsula. It does not appear to be restricted to sandy areas and most of the individuals observed and collected occurred in stony desert at considerable distance from sand. It is a small rodent known by its hopping form of locomotion like the kangaroo. Due to its short front limbs and long hind ones, this rodent is able to jump high to escape from enemies. This small rodent has soft fur coats on sandy colored body with a paler underside and a very long tail, used for balance when jumping. The presence of small hairs between fingers enables the animal to walk on the sand, and the structure of the ear also enables it to hear low sounds. The species is exclusively nocturnal, spending the daylight in its burrows. These rodents are prey to many wild animals such as birds, mammals and reptiles. They are also hunted and eaten in the central region and northern areas of the Kingdom.\n\n### 3.4.2. *Allactaga euphratica* Thomas, 1881 {#s0075}\n\nFour specimens of *A. euphratica* were recorded in the surveyed area. They are characterized by the existence of five fingers on its hind long feet, but it uses only three middle fingers in motion while the other two are vestigial which are found high up on the hind foot. It has small forelimbs, tall ears and black and white tufts of fur on the ends of their tails. This species is nocturnal and lives in burrows with many entrances so as to escape from enemies.\n\n3.5. Family: Erethizontidae {#s0080}\n---------------------------\n\nThe Erethizontidae is a family of rodents commonly known as the New World porcupines. These rodents are medium-sized to large nocturnal animals. All erethizontids share certain characteristics, such as hairs modified into spines with overlapping barbs and are mostly arboreal animals. Only one species belonging to this family was recorded in Turaif region.\n\n### 3.5.1. *Hystrix indica* Kerr, 1792 {#s0085}\n\nTwo specimens of *H. indica* were recorded in the surveyed area. This Porcupine is considered as a largest rodent in the Kingdom of Saudi Arabia, and it reach up to a meter in length and weighs about 15\u202fkg. It is the only species of family Erethizontidae within the Order Rodentia. It is an animal characterized by the presence of sharp long thorns that cover the animal\\'s body, while the head and the front part of the body covered with short black hairs and there is a strong coarse hair on the neck.\n\n3.6. Family: Canidae {#s0090}\n--------------------\n\nMembers of this family are characterized by presence of relatively long limbs with five toes on the forefeet and four on the hind feet but the first finger (thumb) on the forefeet seems to be very short and can be described as vestigial. These animals have non-retractile claws. The front half of the skull seems to be relatively long with elongated facial region. The famous mammals of this family are dogs, wolves, foxes, and jackals where they are fairly similar in terms of the external features. Three species belonging to this family were recorded in Turaif region.\n\n### 3.6.1. *Canis lupus arabus* Linnaeus, 1758 {#s0095}\n\nOnly two specimens of Arabian wolf were recorded in the surveyed area. The Arabian wolf has long limbs and the tail reaches the heels with misty coarse hair ending with black ones. The wolf is known from many localities of the Arabian Peninsula. It is a nocturnal living animal. Despite of their wide spread, they are continuously being killed especially in deserts, leading to poor populations in many areas.\n\n### 3.6.2. Vulpus *vulpus arabica* Thomas, 1902 {#s0100}\n\nEleven specimens of Arabian Red fox were recorded in the surveyed area in desert areas, plains, valleys and near residential areas. This is the largest of the four species of foxes that exist in the Kingdom of Saudi Arabia. The body is covered with soft reddish fur and has long tail ends with a white tuft of hairs. It has black hairs behind the ears and dark brown or black nose.\n\n### 3.6.3. Vulpus *rueppellii sabaea* Schinz, 1825 {#s0105}\n\nEight specimens of sand fox were recorded in the surveyed area. This sand fox is smaller than the red fox and is characterized by long ears and the tail ends with thick white fur. The body has sandy pale color and there is no black hair behind the ear. They were observed in sandy open areas. The Sandy fox is nocturnal. The thin legs and soft, fur-covered foot pads suggest that, unlike some other foxes, this fox does not burrow.\n\n3.7. Family: Mustelidae {#s0110}\n-----------------------\n\nThese mammals are characterized by elongate bodies with short legs and a short rostrum. The limbs end with five fingers and the first one (Thumb) on the front end grow well and in contact with the ground. The fingers end with non-retractile claws. Only one species was recorded in Turaif region.\n\n### 3.7.1. *Mellivera capensis* Schreber, 1776 {#s0115}\n\nOnly one specimens of *M. capensis* was recorded in mountainous habitat of the surveyed area. This Skunk also called the Honey Badger, is a distinctive animal with black or dark brown color with a white or yellowish bold line on the middle of the back. The forefeet are strong and wide, with large claws capable of drilling on the ground. In contrast, the hind feet are small with short claws. It has a thick skin which provides protection from enemies. It is one of the worst enemies of traditional honey breeders, where it digs and burgles the honeycombs. The thick skin provides him with a protection against bee stings when attacking beehives.\n\n3.8. Family: Hyaenidae {#s0120}\n----------------------\n\nThese mammals are considered as relatively big and characterized by back limbs shorter than the front ones, bushy tail and rounded ears. The limbs end with four fingers ending with non-retractile claws. only one species was recorded in Turaif region.\n\n### 3.8.1. *Hyaena hyaena* Linnaeus, 1758 {#s0125}\n\nOnly two specimens of *H. hyaena* were recorded in the surveyed area. Hyaena is also known as Striped Hyaena for the existence of black lines on the sides of the animal\\'s body that distinguish it from the dotted hyena. This nocturnal species is characterized by strong structure with long legs and large front teeth capable of smashing bones and skulls. The body is covered with gray thick hair and the animal can erect the long hair on the neck which it does when it feels threatened. The black and white tail is long and bushy and the feet bear short, blunt claws.\n\n3.9. Family: Felidae {#s0130}\n--------------------\n\nAll felids bear a strong resemblance to one another. The head is semi-round, making the mouth short. The front limbs end with five fingers, while the rear ones end with 4, and the fingers end with retractile claws using them for defense, predation and climbing trees. Tail is considered as relatively long. Only one species was recorded in Turaif region.\n\n### 3.9.1. *Felis margarita* Loche, 1858 {#s0135}\n\nThree specimens of *F. margarita* were recorded in the arid desert habitat of the surveyed area. The Sand Cat looks like a wild cat in the overall appearance, but is the smallest cat species in Arabia. The Sand Cat has bigger broad and sensitive ears. Its foot pads are covered with thick hair. The body has a yellowish gray color and the front limbs with broad strikes. The tail ends with black tuft. It is nocturnal and therefore rare to be seen during the day. Sandy cat is considered as one of the most beautiful animals in sandy environments.\n\n4. Discussion {#s0140}\n=============\n\nThe mammalian fauna of Kingdom of Saudi Arabia is highly heterogeneous and diversified. The animals scattered in different regions of Saudi Arabia have been explored, yet not enough studies were made to provide basic information about the species richness and diversity of mammals, especially in Turaif region. Mammals occupy specific habitats within the different ecozones of Saudi Arabia, which suit their environmental requirements. For the implementation of this study, various field visits were made during the different seasons of the year to study the mammal fauna of the Turaif region of Saudi Arabia. The mammal fauna has been studied in different regions of Saudi Arabia ([@b0075], [@b0005], [@b0080], [@b0020], [@b0025]).\n\nThis study revealed fifteen species of mammals in the Turaif region of Saudi Arabia ([Table 1](#t0005){ref-type=\"table\"}). The spreading of these species is different in the different habitats, some of them are abundant such as members belonging to the family Cricetidae and dipodidae, while others were not common in the study area. The *J. jaculus vocator* was most frequently encountered species followed by *G. nannus* and *M. libycus*. This survey shed new light on the distribution and abundance of various mammal species. The two species *G. nanus* and *F. margarita* recorded first time in the study area, Turaif was the highlight of this report. The presence of *J. jaculus vocator* and *G. cheesmani* were also confirmed from west of Turaif province by [@b0075]. No data concerning reproduction were collected from the specimens collected in the surveyed area.\n\nThe level of survey undertaken does not permit conclusions to be drawn about the total species comprising the mammal assemblages in the study area. However, information presented here is sufficient to provide a basis from which to discuss the potential significance of the study area for mammals. However, recording the different animal species by our group like [@b0010] (lizard fauna), [@b0015] (snake fauna) and the present study on mammal fauna inhabiting Turaif region, there is need for additional research concerning to the fauna and flora of the region, their history, adaptation, distribution, characteristics, and how they are influenced by different environmental factors. There is also a need for maps illustrating the past and present distribution of large mammals and presenting their relationships to other environmental factors and human activities.\n\nThe authors would like to extend their sincere appreciation to the Deanship of Scientific Research at the King Saud University, Riyadh, Saudi Arabia for funding this Research Group project no RGP-289. Also, the authors extend their sincere thanks to Ma'aden for their help and support during the study period in Turaif region.\n\nPeer review under responsibility of King Saud University.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nThe question of whether endurance running poses an increased risk of knee osteoarthritis (OA) remains controversial^[@CR1]--[@CR3]^. A number of authors state that prolonged running is not associated with an increased prevalence of cartilage degeneration^[@CR4]--[@CR10]^, while others indicate a risk for OA^[@CR1],[@CR11]--[@CR14]^. Chakravarty *et al*.^[@CR15]^ found no acceleration of radiographic knee OA in healthy older runners (50--71 years). Due to the non-pathological, but highly repetitive strain on the legs, excessive running stress can be regarded as mild chronic trauma to the joints of the lower extremities. Prolonged running burden over time is thought to alter the collagen network and lubricating proteoglycans, slowly wear out the articular cartilage, and cause microfractures in the subchondral bone^[@CR4],[@CR16]^. However, this hypothesis has not been proven, and clinical and radiographic cohort studies of regular long-distance running subjects, up to marathons, have shown no significance regarding knee OA incidence and progression in hip and knee joints without knee injury, poor muscle tone, obesity, or proprioceptive deficit^[@CR9],[@CR17]--[@CR24]^.\n\nThe physiology of ultra-endurance exercise is very important because it is believed that the ability to walk and run long distances has played an important role in human evolution^[@CR25]^. The German Ultramarathon Association defines a multistage ultra-marathon (MSUM) as a marathon with more than 4 stages, each with a distance of more than 50\u2009km. However, although the amount of data pertaining to long-distance running has increased in recent decades, little is known about the consequences of ultra-long-distance running, such as MSUM, on knee joint cartilage^[@CR26]^.\n\nOver the last 2 decades, magnetic resonance imaging (MRI) has become a validated and reliable *in vivo* method for quantitative and qualitative assessment of human joint cartilage^[@CR27]--[@CR29]^. Quantification of changes in the composition of hyaline cartilage during short (5\u2009km) and long-distance runs (marathon) can be carried out noninvasively by mapping the T1rho- and T2-relaxation times^[@CR30]--[@CR40]^. As these values are elevated in patients with OA, these mapping methods have played increasingly important role in detecting the onset of cartilage degeneration^[@CR40]--[@CR43]^. T1rho mapping shows a high negative correlation with intrachondral proteoglycan concentration and appears to be more sensitive to concentration changes^[@CR44],[@CR45]^. T2 mapping, however, is sensitive to changes in the equilibrium of free water, which depend on the orientation and content of collagen and the content of proteoglycans in the extracellular cartilage matrix^[@CR31],[@CR41],[@CR43],[@CR46]--[@CR50]^.\n\nAccording to the anisotropic properties of the collagen fibres in articular cartilage, the T2 relaxation changes with different spatial orientations in the magnetic field of the MR scanner^[@CR51]^. In the femorotibial joint (FTJ), T2 value inequalities and variabilities between corresponding cartilage regions are correlated with the limited water mobility within an anisotropic rigid matrix^[@CR52]^. They mostly depend on the different degrees of stiffness and creep rates^[@CR53]^ needed in relation to degrees of biomechanical stress and loading (cartilage compression)^[@CR54]^, resulting in regionally different glycosaminoglycan (GAG) content and collagen fibril network orientation^[@CR55]^. It is widely accepted that load-induced intrachondral T2 increases result from a loss of structural anisotropy of type II collagen matrix (spatial orientation of collagen fibrils) and from an increase in free cartilage water^[@CR41],[@CR47],[@CR56],[@CR57]^.\n\nAs a correlate of the disturbance of the cartilage structure and composition, Mosher *et al*. showed that running induced an initially significant decrease in the superficial chondral T2 values in the FTJ^[@CR37]^. Subburaj *et al*.^[@CR42]^ also described a significant decrease in T1rho- and T2-relaxation times after only 30\u2009min of running, which indicates that an increased cartilage running load leads to a temporary dehydration of the chondral tissue. Luke *et al*. evaluated an increase in chondral T1rho and T2 values in the first days after a single marathon compared with a control group^[@CR58]^, and the highest signal changes were found in the medial FTJ in the marathon runners.\n\nWith shorter measurement times and possible 3D recordings, the T2\\* technique^[@CR59]--[@CR61]^, which is also related to the collagen network and the water content, has been shown to be a practical alternative to the time-consuming T2 technique with multiple echo times^[@CR59],[@CR62]^. In addition, T2\\* imaging offers an inherently higher signal-to-noise ratio (SNR) and greater robustness^[@CR63],[@CR64]^ than the other methods (T1rho/T2 mapping, delayed gadolinium-enhanced MRI of cartilage (dGEMRIC)^[@CR61]^). As contrast media is not required, T2\\* is considered to be the best and only viable option for mobile MRI field studies, such as the TransEurope-FootRace (TEFR) project^[@CR65]^.\n\nThe TEFR project, which evaluated participants in the transcontinental MSUM from the southernmost (Bari, Italy) to the northernmost border of Europe (North Cape), is the first study to show the feasibility of recording mobile MRI from ultra-marathon runners over 4,486\u2009km through 6 countries and a period of 64-days without any day of rest; the mean stage distance was 70.1\u2009km or 1.7 marathons (min 44\u2009km, max 95.1\u2009km)^[@CR65]^. The prospective approach applied in this long-term MRI field study led to unique data accumulation, revealing in detail the influence of an ultra-endurance burden on different tissues^[@CR66]--[@CR71]^.\n\nThe main purpose of this article was to characterize the effect of MSUM running on the intrachondral free water distribution and collagen fibre content and organization using serial MRI (T2\\* response) recordings over several weeks. The results may provide evidence that ultra-running alters the extracellular cartilage matrix behaviour of the FTJ towards cumulative structural chondral damage with an increased risk of OA. Furthermore, we evaluated whether any detectable running-induced T2\\* reactions show regional differences in FTJ cartilage.\n\nMethods {#Sec2}\n=======\n\nStudy population {#Sec3}\n----------------\n\nAfter approval by the local ethics committee (University Hospital of Ulm, No.: 270/08-UBB/se) and obtaining written informed consent, 22 TEFR participants (demographics see Table\u00a0[1](#Tab1){ref-type=\"table\"}) were included in the mobile MRI study of the knee joints. All methods were performed in accordance with the relevant guidelines and regulations. The inclusion criterion was official acceptance as a participant at the TEFR by the organizers^[@CR65]^, including age\\> 17 years, the presence of a medical certificate indicating physical health, and the successful completion of at least one MSUM. No subject had to be excluded due to contraindications against native MRI^[@CR72]^. The subjects had no relevant knee injuries (meniscus tears, cruciate ligament ruptures, etc.) or prior surgical interventions of the knee joints, which could change the intrinsic cartilage conditions and thus normal chondral T2 signalling^[@CR65]^. The baseline data (before the start of the marathon) of the 18 of the 22 subjects (81.8%) who remained in the race until the last measurement interval (MI) were as follows: mean age of 51.0 years (SD 10.4, range 26--65), 17 males (94.4%), body mass of 70.4\u2009kg (SD 7.5, range 58.9--82.2), body mass index (BMI) of 23.0\u2009kg/m\u00b2 (SD 1.8, range 20.3--26.4), and body height of 175\u2009cm (SD 6, range 163--184). Reasons for premature termination of the TEFR were overuse reactions of the soft tissues of the lower extremities (\"shin splints\", myofascial inflammation) in 2 male participants, stress fracture of the left ventral pelvic ring in one female participant, and compliance problems by one male participant.Table 1Subjects demographics of TEFR-project at baseline (t0).TEFR-No.m/fage (yrs)BMI (kg)leg preferencefinisher/non-finisher (finished stages)total run time (hrs)**1**01m43.825.4rightF415.9**2**09f45.622.4rightF692.7**3**10m53.122.4rightF626.6**4**13m52.423.7rightF515.3**5**16f46.421.2leftNF (49)(419.2)**6**19m30.824.6rightF560.7**7**20m57.220.2rightF615.2**8**22m50.721.8rightNF (28)(208.6)**9**26m56.624.3rightF651.5**10**32m26.220.6rightF407.0**11**33m65.426.4leftF645.3**12**34m63.222.2rightF669.3**13**40m53.122.4rightF484.9**14**50m65.121.3leftF650.0**15**51m46.122.2rightF548.5**16**55m53.021.1rightF545.3**17**57m43.520.8rightF (left study at stage 42)(642.2)**18**59m49.925.6rightF506.7**19**60m58.521.6rightF629.0**20**63m49.424.6rightF517.3**21**64m48.922.2rightF566.1**22**68m46.424.1leftNF (10)(38.2)\n\nData acquisition {#Sec4}\n----------------\n\nMRI data were acquired with a mobile 1.5\u2009T MR scanner (Magnetom Avanto^TM^ mobile MRI 02.05, software version: Syngo^TM^ MR B15, Siemens Ltd., Germany) mounted on an MRI trailer (Model Mob. MRI 02.05, SMIT Mobile Equipment B.V., Farnham, UK), which was pulled by a truck tractor (38 tonnes in total) and travelled with the runners throughout entire TEFR from stage to stage, day by day^[@CR65]^. MRI scanning was planned at baseline (t0; within the 4 days prior to the start of the TEFR), at 3 MIs during the TEFR (t1 planned after 1,200\u2009km, t2 after 2,400\u2009km, and t3 after 3,600\u2009km of the run; between 2:00\u2009pm and 9:00\u2009pm after the daily stage), and 7 months after the TEFR (tx). However, this plan was not able to be implemented consistently during the field study. Influencing factors such as daily changing weather conditions, local conditions, and the individual changing needs and sensitivities of the subjects due to the immense physical and mental stress within the framework of such an MSUM required daily modification and rescheduling of the scanning plan^[@CR65]^. The actual measurement intervals implemented during the TEFR project were as follows: T1 was between stages 14--20 after a mean distance run of 1,103\u2009km (SD 106\u2009km, range 926--1,325\u2009km), T2 was between stages 39--45 after a mean distance run of 2,779\u2009km (SD 218\u2009km, range 2,313--3,082\u2009km), and T3 was between stages 51--57 after a mean distance run of 3,673\u2009km (SD 234\u2009km, range 3,516--3,973\u2009km). To minimize variation in joint positioning, a table-fixed, 8-channel knee coil was used. In the finisher group, 17 subjects (77.3%) underwent MRI at all MIs (t0-t3), and 12 of them underwent MRI at tx. T2\\* mapping was performed with a sagittal fast low-angle shot T2\\* weighted gradient-echo sequence with a 60\u00b0 flip angle, echo times of 4.2/12.2/19.9/27.7/35.4\u2009ms, a repetition time of 1,120\u2009ms, a slice thickness of 3.0\u2009mm, a field of view of 289\u2009cm\u00b2, a pixel size of 0.11\u2009mm\u00b2, and an in-plane resolution of 0.332 iso. T2\\* relaxation times were calculated from online reconstructed T2\\* maps by using a pixelwise, mono-exponential non-negative least squares fit analysis (syngo^TM^ MapIt; Siemens Ltd.)^[@CR50],[@CR73]^. Sagittal turbo inversion recovery measurements and coronal fat saturated proton density weighted sequences were also performed.\n\nImage evaluation {#Sec5}\n----------------\n\nFor measurements of cartilage T2\\* profiles, 4 sagittal slices through the FTJ were determined, and the 2 most median slices in the medial and lateral femorotibial compartment were used. To determine if there were regional differences in the T2\\* response of cartilage to ultra-running burden, a subset quantitative analysis was performed in which cartilage regions of interest (ROIs) were separately generated from the medial and lateral knee compartments^[@CR50]^. These ROIs were manually drawn on the 4 sagittal slices such that the whole femoral and tibial cartilage area was covered between the meniscal bases (see Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). These areas were separated into a superficial and a deep layer and in an anterior (cartilage covered by the anterior meniscal horn), central (uncovered cartilage between the anterior and posterior meniscal horn) and posterior zone (cartilage covered by the posterior meniscal horn)^[@CR74]^. In total, the representative cartilage T2\\* map was separated into 12 ROIs on each slice (see Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). Care was taken to avoid the subchondral bone or joint fluid and to set the ROIs in the exact same positions at every examination. The ROIs were drawn twice by 2 independent investigators with a time gap of 6 months. They were supervised by 2 radiologists with 30 to 15 years of experience in musculoskeletal imaging. As a strong significant correlation (*p*\u2009\\<\u20090.01) was found between the 2 series (r\u2009\\>\u20090.8) for all cartilage areas and most of the ROIs (see Supplementary Table\u00a0[S1](#MOESM1){ref-type=\"media\"}), averaged data of the mean T2\\* values of the 2 series were taken for further analysis. To determine the mean T2\\* for each cartilage area (layers, zones, or segments), the mean T2\\* values of the specific ROIs were pooled and calculated with regard to the ROI sizes (see Supplementary Table\u00a0[S2](#MOESM1){ref-type=\"media\"}). Sufficient intra- and interobserver precision and reliability results of the T2\\* measurement method used herein have already been presented^[@CR29],[@CR34],[@CR41],[@CR50],[@CR62],[@CR70],[@CR71]^. The precision (reproducibility) of the drawn ROI sizes in the course of the TEFR was calculated with 3.2 CV% in mean (SD 0.24) between the MIs (t0-t3) and no significant changes in the drawn ROI areas were detected for all investigated ROIs.Figure 1MR-image postprocessing for quantification of cartilage T2: (**A**) 4 sagittal slices (2 median slices in lateral and medial FTJ); **B**\u2009+\u2009**C:** fused colored T2\\* maps of 2 median sagittal FLASH T2\\* GRE (**C**) lateral FTJ, (**D**) medial FTJ).\n\nStatistical analysis {#Sec6}\n--------------------\n\nFor data documentation, analysis and graphical presentation, Microsoft^TM^ Excel^TM^ (Office-Business 2016), SPSS^TM^ (IBM^TM^-Statistics, release 25.0, 2017) and SigmaPlot^TM^ (release 14.0, 2018, Systat-Software Inc.) were utilized. For all tests, a *p* value of 0.05 indicated significance.\n\nSignificant changes in cartilage T2\\* between baseline (t0) and MIs (t1-t3) in the course of the TEFR were calculated by *1-way analysis on variance (ANOVA) with a general linear model for repeated measurements and pairwise post hoc analysis* to calculate the significance between MIs (including the Bonferroni procedure for multiple measurements and a Mauchly test for data sphericity with the Greenhouse-Geisser correction). *Significance of ANOVA inner-subject contrasts* was calculated to detect (linear or polynomial) trends in the T2\\* values during the progression of the TEFR (accepted for power\\> 0.8). To determine significant value differences at the last MI (t3) to any maximal change (peak) during exercise (t1-t2) compared with baseline, a *paired (2-tailed) samples t-test with calculation of the effect size according to Cohen*^[@CR75]^ was performed.\n\nDifferences in cartilage T2\\* between ROIs and zones at the same MI and within the same cartilage segment were calculated by *1-way analysis on variance (ANOVA) without repeated measurements and pairwise post hoc analysis on significance*: the Scheff\u00e9 test was employed for homogeneous variance (Levene test *p*\u2009\\>\u20090.1); otherwise, the Bonferroni test was used. To test inter-segmental differences and differences in cartilage T2\\* between ROIs and zones at the same MI and within the same cartilage segment, a *t-test for independent variables* was performed.\n\nResults {#Sec7}\n=======\n\nRegional cartilage T2\\* differences {#Sec8}\n-----------------------------------\n\nIn the right knee, the mean T2\\* value of the lateral femoral segment was significantly higher than that of the medial segment over the course of the race (t1-t3), while no relevant lateral to medial T2\\* difference was detected at baseline on either side or at any MI on the left side (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"}). At every MI (t0-t3, tx), the femoral cartilage showed a significantly higher mean T2\\* signal than the tibial cartilage in both compartments of the left and in the lateral compartment of the right knee (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Supplementary Fig.\u00a0[S4](#MOESM1){ref-type=\"media\"}).Table 2Mean T2\\* differences \\[ms\\] between specific cartilage areas (n~F~\u2009=\u200917).right FTJleft FTJt0t1t2t3txt0t1t2t3tx**mean differences between lateral and medial knee compartments**^**b**^femoral segments3.1**5.36.16.02.2**\u22121.62.3 ^c^1.9 ^c^\u22120.21.8tibial segments\u22120.41.31.91.60.1\u22122.3\u22122.0\u22121.2\u22122.20.2**mean differences between femoral and tibial cartilage segments**^**b**^lateral knee compartment**3.15.04.85.32.44.38.07.37.05.3**medial knee compartment\u22120.40.90.60.90.4**3.53.74.24.93.3mean differences between superficial and deep cartilage layers within the same cartilage segment**^**b**^femoral-lateral**6.710.111.111.07.36.410.6**^**c**^**11.0**^**c**^**10.0**^**c**^**8.4**tibial-lateral**6.810.811.89.88.56.110.19.99.7**^**c**^**6.8**femoral-medial**5.58.89.19.36.97.611.211.912.4**^**c**^**8.9**tibial-medial**10.214.8**^**c**^**14.4**^**c**^**14.5**^**c**^**12.66.38.9**^**c**^**8.2**^**c**^**9.3**^**c**^**6.2mean differences between ventral, central and dorsal zones within the same cartilage segment**^**a**^femoral lateral segment:anterior vs. central zone1.20.20.10.2\u22120.1\u22120.8\u22122.3\u22121.0\u22121.6\u22122.5anterior vs. posterior zone\u22122.3**\u22126.7**\u22124.6**\u22125.5**\u22123.1\u22122.6**\u22125.2\u22124.3**\u22125.0**\u22124.8**central vs. posterior zone\u22123.53**\u22126.9\u22124.7\u22125.7**\u22123.0\u22121.8\u22122.8\u22123.4\u22123.4\u22122.3tibial lateral segment:anterior vs. central zone0.901.80.51.20.61.91.61.51.1anterior vs. posterior zone**\u22125.8\u22129.9\u22128.2\u22128.2\u22125.1**\u22121.7\u22121.7\u22123.7\u22122.6**\u22122.3**central vs. posterior zone**\u22126.67\u22129.97\u221210.0\u22128.8\u22126.2**\u22122.4\u22123.6**\u22125.4\u22124.1\u22123.4**femoral medial segment:anterior vs. central zone0.12.43.41.63.63.4**4.24.6**4.11.8anterior vs. posterior zone**\u22126.0\u22125.3**\u22123.6**\u22126.5**\u22121.70.9**\u22124.5**\u22121.6\u22121**\u22125.0**central vs. posterior zone**\u22126.0\u22127.7\u22127.0\u22128.1\u22125.3**\u22122.4**\u22128.7\u22126.2\u22125.1\u22126.8**tibial medial segment:anterior vs. central zone2.32.7**3.8**3.6**4.4**1.31.12.82.21.2anterior vs. posterior zone\u22122.1**\u22123.8**\u22122.3\u22122.30.50.4\u22122.0\u22121.10.3\u22121.1central vs. posterior zone**\u22124.4\u22126.5\u22126.1\u22125.9\u22123.9**\u22120.8\u22123.1**\u22123.9**\u22121.9\u22122.2^a^(one-way) univariate ANOVA, ^b^independent t-test, c: variance homogeneity not given (Levene test \\<0.05).Bold fonts show significant differences (*p-value*\u2009\\<\u20090.05).\n\nAt every MI and in all cartilage segments on both sides, the superficial layers had a significantly higher total T2\\* than the corresponding deep layers (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"} and [S4](#MOESM1){ref-type=\"media\"}). Depending on the segment, this mean T2\\* layer difference was approximately 48% (tibial medial) and 66% (femoral medial) up to 70% (lateral) higher over the course of the race compared with baseline (see Table\u00a0[2](#Tab2){ref-type=\"table\"}).\n\nIn the right knee in all cartilage segments, the mean T2\\* values showed a significant inter-zone difference at every MI, with the exception of the femoral-lateral segment at baseline (t0). In all segments of the right side, the T2\\* value was significantly higher in the posterior zone than in the central zone. Tibial lateral at every MI and sometimes in the other segments, it was also significantly higher in the posterior than in the anterior zone (see Table\u00a0[2](#Tab2){ref-type=\"table\"}). In the left knee, the mean T2\\* differences between zones within the same cartilage segment were not significant at baseline in any segment and were not significant at t3 in the femoral-lateral segment, at t2 in the tibial-lateral segment or at t1 and t3 in the tibial-medial segment. On both sides, the mean T2\\* differences between the anterior and central zones were generally not significant.\n\nOn both sides, at every MI, a significant difference in mean T2\\* values between ROIs within the same cartilage segment was detected in all segments (see Supplementary Table\u00a0[S5](#MOESM1){ref-type=\"media\"}). As all T2\\* values of ROIs were significant higher during the race (t1-t3) than at baseline (see Supplementary Table\u00a0[S5](#MOESM1){ref-type=\"media\"}), further analyses also showed significantly higher mean T2\\* differences between zones, layers and segments in the race than at baseline (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"} and [S4](#MOESM1){ref-type=\"media\"}).\n\nCartilage T2\\* changes {#Sec9}\n----------------------\n\nThe T2\\* values changed significantly in all ROIs during the course of the TEFR (see Table\u00a0[3](#Tab3){ref-type=\"table\"}). The value changes occurred mainly in the first MI with a significant T2\\* increase: 44.0% femoral-lateral, 42.9% tibial-lateral, 34.9% femoral-medial, and 25.1% tibial-medial (see Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"} and [S4](#MOESM1){ref-type=\"media\"}). Statistical evaluation showed a significant quadratic trend in the T2\\* curves over the course of the race for nearly all ROIs on both sides (see Table\u00a0[4](#Tab4){ref-type=\"table\"}). Consecutively, the mean T2\\* of all cartilage areas showed a significant quadratic trend on both sides (see Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"} and [S4](#MOESM1){ref-type=\"media\"}). With the exception of the anterior and central superficial ROIs of the left medial FTJ, a significant difference in the initial mean T2\\* increase relative to baseline was not evaluated between any area on both sides (see Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"}). A significant secondary T2\\* decrease at t3 was detected for the tibial and left femoral segments (mainly in the superficial layer) with a medium to high effect size (see Table\u00a0[5](#Tab5){ref-type=\"table\"}).Table 3Changes of intrachondral T2\\* values \\[ms\\] in the course of TEFR (n~F~\u2009=\u200917).segmentROIright kneeleft kneeMauchly-TestANOVA^a^Mauchly-TestANOVA^a^*p-valuep-value*test power*p-valuep-value*test powerlateral femorotibial jointfemoral lateraldeep-ant.0.811**\\<0.001**1.0000.531**\\<0.001**1.000deep-central0.497**\\<0.001**1.0000.169**\\<0.001**1.000deep-post.0.422**\\<0.001**1.0000.385**\\<0.001**1.000superf.-ant.0.609**\\<0.001**1.0000.292**\\<0.001**1.000superf.-central0.686**\\<0.001**1.0000.083**\\<0.001**1.000superf.-post.0.468**\\<0.001**1.0000.053**\\<0.001**1.000tibial lateraldeep-ant.0.506**\\<0.001**1.0000.211**\\<0.001**1.000deep-central**0.030\\<0.001**^**b**^1.0000.484**\\<0.001**1.000deep-post.0.078**\\<0.001**1.0000.504**\\<0.001**1.000superf.-ant.0.713**\\<0.001**1.000**0.040\\<0.001**^**b**^1.000superf.-central0.370**\\<0.001**1.000**0.015\\<0.001**^**b**^0.999superf.-post.0.467**\\<0.001**1.000**0.025\\<0.001**^**b**^0.995medial femorotibial jointfemoral medialdeep-ant.0.881**\\<0.001**1.0000.393**\\<0.001**1.000deep-central0.493**\\<0.001**0.9990.201**\\<0.001**0.996deep-post.0.393**\\<0.001**0.9990.240**\\<0.001**1.000superf.-ant.0.267**\\<0.001**1.0000.805**\\<0.001**0.999superf.-central**0.0490.001**^**b**^0.9920.487**\\<0.001**1.000superf.-post.0.527**\\<0.001**1.000**0.014\\<0.001**^**b**^1.000tibial medialdeep-ant.0.707**\\<0.001**0.9940.558**\\<0.001**1.000deep-central0.839**\\<0.001**0.992**0.047\\<0.001**^**b**^0.998deep-post.0.753**\\<0.001**1.0000.651**\\<0.001**0.998superf.-ant.0.778**0.006**0.8720.131**\\<0.001**0.999superf.-central0.499**0.005**0.8890.127**\\<0.001**1.000superf.-post.0.291**\\<0.001**1.0000.435**\\<0.001**1.000^a^(one-way) univariate ANOVA for repeated measurements, ^b^with \"Greenhouse-Geisser\" correction procedure.Bold fonts show significance (*p-value*\u2009\\<\u20090.05).Figure 2Relative changes of T2\\* mapping compared to baseline (n~F~\u2009=\u200917), right knee.Figure 3Relative changes of T2\\* mapping compared to baseline (n~F~\u2009=\u200917), left knee.Table 4Analysis on quadratic trends of repeated T2\\* relaxation time \\[ms\\] in the course of TEFR. (univariate ANOVA for repeated measurements, n~F~\u2009=\u200917).ROIright FTJleft FTJmediallateralmediallateralfemoraldeep-ant.**0.0010.0010.010\\<0.001**deep-central**0.0030.001**0.054**0.001**deep-post.**0.0010.0040.0050.001**superf.-ant.**\\<0.0010.003**0.509**0.004**superf.-central**0.0110.001**0.067**0.001**superf.-post.**0.0010.001\\<0.0010.001**deep**0.0010.001\\<0.001\\<0.001**superf.**0.0010.0010.0010.001total0.0020.001\\<0.001\\<0.001**tibialdeep-ant.0.024**\\<0.0010.003\\<0.001**deep-central**0.001\\<0.0010.0040.001**deep-post.**\\<0.0010.001\\<0.001\\<0.001**superf.-ant.0.035**\\<0.001**0.050**0.001**superf.-central**0.0060.0010.0060.001**superf.-post.**0.000\\<0.001\\<0.0010.001**deep**\\<0.0010.001\\<0.0010.001**superf.**0.0060.0020.0010.001total0.0010.001\\<0.0010.003**Femoral segments**\\<0.0010.001**Tibial segments**\\<0.001\\<0.001**Total**0.002\\<0.0010.0010.001**Total FTJ**\\<0.0010.001**Bold fonts show significant p-values with high test power (\\>0.8).Table 5Analysis (paired *t*-test) on secondary T2\\* decrease \\[ms\\] at last MI (t3). (n~F~\u2009=\u200917).ROIright kneeleft knee*p-valuedp-valued*lateral femorotibial jointfemoral lateral segmentdeep-ant.0.0420.450.0090.47deep-central**0.028**0.710.0060.47deep-post.**0.027**0.630.0090.43superf.-ant.0.0600.380.0150.44superf.-central0.0300.450.0100.48superf.-post.0.0320.440.0100.47deep**0.050**0.640.0120.44superf.0.0370.450.0110.48**total0.042**0.54**0.008**0.55tibial lateral segmentdeep-ant.**0.006**0.62**0.009**0.55deep-central**0.022**0.630.0100.49deep-post.**0.014**0.720.0060.44superf.-ant.**0.009**0.730.0390.47superf.-central**0.012**0.640.0150.42superf.-post.**0.006**0.780.0050.40deep**0.034**0.660.0160.44superf.**0.036**0.620.0160.32**total0.33**0.630.0100.42medial femorotibial jointfemoral medial segmentdeep-ant.**0.003**0.670.0040.27deep-central0.0160.490.0220.35deep-post.0.0090.460.0020.46superf.-ant.0.0080.420.0270.29superf.-central0.0120.380.0230.25superf.-post.0.0290.410.0100.41deep**0.012**0.50**0.007**0.50superf.0.0200.38**0.028**0.69**total**0.0160.410.0140.45tibial medial segmentdeep-ant.0.0130.450.0430.44deep-central**0.003**0.650.0240.35deep-post.**0.003**0.70**\\<0.001**0.89superf.-ant.**0.001**0.57**0.036**0.59superf.-central**0.001**0.67**0.029**0.53superf.-post.**0.004**0.62**0.010**0.66deep0.0100.490.0130.49superf.**0.014**0.520.0120.39**total**0.0110.45**0.020**0.63Femoral segments0.0350.46**0.027**0.51Tibial segments**0.025**0.52**0.029**0.55Lateral FTJ**0.043**0.590.0170.47Medial FTJ0.0140.430.0400.45Total FTJ0.0330.48**0.041**0.50Bold fonts show significant *p-value*s with middle (d\u2009\\>\u20090.5) to high (d\u2009\\>\u20090.8) effect size (Cohen's d).\n\nT2\\* reaction of nearby structural osteochondral lesions {#Sec10}\n--------------------------------------------------------\n\nIn 2 male subjects with femoral medial focal grade 3--4 OA at baseline, with no further structural changes throughout the TEFR, the corresponding regional cartilage showed significantly higher absolute T2\\* values at baseline and a comparatively higher initial mean T2\\* increase (at t1) and higher secondary mean T2\\* decrease (at t3) compared with all other non-lesion ROIs (see Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}).Figure 4Osteochondral lesions (\\*) in the FTJ at baseline (t0) and last MI (t3) with no relevant signal changes throughout TEFR (sagittal TIRM: (1) right side, (4) left side; coronal PDfs: (2) right side, (3) left side). (**A**) 58-year-old male finisher (t3: stage 57/3,971\u2009km run); bilateral medial meniscal lesions in the cornu posterior, grade 3 on the right and grade 2 on the left side (*arrows*) with adjacent femoral chondral lesions Outerbridge grade 3 on the right, and Outerbridge grade 4 on the left side with focal femoral subchondral bone edema on both sides (\\*): mean T2\\* increases more in lesional posterior ROIs, but shows secondary decrease at t3. Non lesional anterior and central ROIs show only slight continuous mean T2\\* increase. (**B**) 48-year-old male non-finisher (t3: stage 63 / 3,669\u2009km run); bilateral osteochondral lesions at baseline (Outerbridge grade 2--3) with a focal subchondral bone edema (\\*) located femoral medial: mean T2\\* increases more in lesional medial layers but shows secondary decrease at t3. Non lesional median layers show only slight continuous mean T2\\* increase.\n\nDiscussion {#Sec11}\n==========\n\nWe found a significant T2\\* increase in all cartilage areas of the FTJ in the first third of the TEFR. After a \"steady state\" in the middle of the race and at approximately 3,500\u2009km of the run, the T2\\* signal showed a detectable decreasing trend in many, mostly tibial cartilage regions (see Table\u00a0[5](#Tab5){ref-type=\"table\"}, Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"}). To the best of our knowledge, this is the first and only prospective MRI study that analysed the cartilage response of a large number of human knee joints during an ultra-run load of almost 4,500\u2009km; therefore, a comparative discussion is only possible considering previous studies of animal models with similar distances or studies of shorter running distances in humans.\n\nImmediate, primary T2^(^\\*^)^ decrease with rapid recovery {#Sec12}\n----------------------------------------------------------\n\nIn static and acute loading studies^[@CR38],[@CR76]^ and in most post-race studies of short- (5\u2009km) to long-distance runs (marathons) that employed MRI within 1\u2009hour^[@CR30]--[@CR33],[@CR35]--[@CR37],[@CR77]^, a pressure-dependent significant T2 decrease was observed in all cartilage layers. With cyclical loading, cartilage deformation produces an increase in superficial collagen anisotropy and a concomitant decrease in (free) cartilage water, leading to lower post-race T2 values. For example, Mosher *et al*.^[@CR36]^ found a relevant initial T2 decrease in superficial cartilage (tibial: 1--3\u2009ms, femoral: 2--4\u2009ms) and no T2 change in the deep layer in subjects after a 30-min run. This immediate T2 response recovers rapidly after 30\u2009min up to a few hours^[@CR78]--[@CR81]^. However, we could not detect such primary T2\\* decreases at any MI (t1-t3), since our first data were acquired at t1 after a distance run of 962--1,325\u2009km (stage 14--20), at which time the effects of long-distance running dominate (masked) the immediate T2-response.\n\nLong-distance, primary T2(\\*) increase {#Sec13}\n--------------------------------------\n\nAnimal studies have shown that a more strenuous long-distance running burden leads to GAG decrease in the main weightbearing cartilage areas^[@CR82]--[@CR84]^. Only 2 studies have focused on T2 mapping from 1\u2009hour up to less than 24\u2009hours after a single marathon race in humans^[@CR34],[@CR58]^. Both reported significantly increased T2 values in knee cartilage, in contrast to all published studies with the first follow-up performed after less than 1\u2009hour. Using high-field MRI, Luke *et al*.^[@CR58]^ found not only significantly higher T2 values but also equivalent higher intra-chondral T1rho values in the FTJ 48\u2009hours after a marathon, postulating that this signal elevation is mainly caused by a proportional GAG decrease^[@CR45]^. Animal long-term running studies (young dogs: 1 year, 40\u2009km/day) showed that if the interstitial environment of chondrocytes is disturbed, high fluid pressure leads to disorganization of the collagen and proteoglycan network and structural changes in chondral fibrillation^[@CR83],[@CR85]^. Therefore, the initial significant T2\\* increase in the cartilage areas of the FTJ after the 1,100\u2009km TEFR run may not only be explained by the loss of spatial collagen structure; rather, it may also be caused by increased free cartilage water due to GAG depletion. Since no relevant morphometric and structural changes were observed during the T2\\* increase in the first third of the TEFR^[@CR86]^, our data may be explained by changes in chondral permeability and respective ion flow, which reached a new equilibrium from stage to stage.\n\nDuring the TEFR a significant mean T2\\* difference between ROIs was detected in all segments (see Table\u00a0[2](#Tab2){ref-type=\"table\"}). Since all these regional T2\\* differences were higher throughout the race than at baseline (see Supplementary Table\u00a0[S5](#MOESM1){ref-type=\"media\"}), this finding applies to all zones, layers and segments (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"}). However, we could not determine significant difference in T2\\* increase between cartilage areas (see Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"}).\n\nThere is a typical spatial variation in T2 values in healthy cartilage. For example, lower T2 values are found in the deeper radial layer, where the collagen fibres are highly ordered, and elevated T2 values are found in the superficial layer with less collagen organization^[@CR41],[@CR87],[@CR88]^. Mosher *et al*.^[@CR36]^ observed that this spatial dependency of cartilage T2 is a function of the distance from bone, even in recreational runners. This finding is explained by cartilage biochemistry: quantitative microspectroscopy has shown a significant running-induced decrease in GAG content in the weight-bearing superficial layer of the lateral femoral (5--13%) and tibial cartilage (5--35%) of the canine FTJ^[@CR55],[@CR85],[@CR89]^. The deformation rate of cartilage increases due to disorganization or lesion of the superficial collagen network due to reduced GAG content^[@CR55]^, which contributes to instant cartilage stiffness through osmotic pressure^[@CR90]^. Since cartilage stiffness is directly related to GAG content^[@CR91]^, the cartilage equilibrium shear modulus also decreases significantly in the lateral FTJ (12--14%)^[@CR85]^. Consistent with the fact that cartilage areas with regularly high loads are stiffer than areas with lower loads^[@CR92]^, we found significant mean T2\\* differences between the superficial and deep layers throughout the TEFR in all cartilage areas of the FTJ (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Supplementary Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"} and [S4](#MOESM1){ref-type=\"media\"}).\n\nZonal dependency of cartilage T2 has also been observed in several studies, and we found higher T2\\* values in the posterior zone than in the central and anterior zones, being significant in the right FTJ. As Mosher *et al*.^[@CR36]^ observed in normal running, we also found significantly higher T2 values and greater T2\\* variation in femoral cartilage than in tibial cartilage at all MIs during the MSUM. In humans, Subburaj *et al*.^[@CR42]^ indicated a greater load on the medial compartment of the FTJ during running and, therefore, that the medial cartilage is stiffer and shows a higher isotropy of collagen matrix. With increasing load, the contact area and pressure seem to increase, especially in the lateral compartment of the FTJ^[@CR93]^. This might explain why the left medial segments showed a significantly lower average T2\\* increase than the lateral segments in our study group (see Table\u00a0[2](#Tab2){ref-type=\"table\"} and Figs.\u00a0[2](#Fig2){ref-type=\"fig\"} and [3](#Fig3){ref-type=\"fig\"}).\n\nUltra-long distance, secondary T2^(^\\*^)^-decrease (\"recovery\") {#Sec14}\n---------------------------------------------------------------\n\nIn the FTJ cartilage, the long-distance T2 increase within 48\u2009hours after a single marathon recovered to baseline after 3 months^[@CR58]^. As our subjects continued to run daily ultra-marathons up to 64 stages (more than 2 months) without sufficient recreation time, T2\\* recovery could not be expected until t3. At the final MI (t3), we measured a significant secondary T2\\* decrease in the tibial and left femoral segments (mainly in the superficial layer) with a medium to high effect size (see Table\u00a0[5](#Tab5){ref-type=\"table\"}). Our results on ankle and hindfoot cartilage T2\\* behaviour at TEFR^[@CR70],[@CR71]^ showed the same reaction in the axial loaded joints: a primary T2\\* increase as the first long-distance T2\\* response after 1,100--2,400\u2009km run^[@CR85]^ but an earlier and stronger secondary T2\\* decrease (\"recovery\") as an ultra-long distance T2\\* response after more than approximately 2,000--2,500\u2009km run.\n\nA reasonable explanation is reorganization of the articular cartilage collagen network^[@CR94]^ and reduction in free water because it gets the ability to go back in chemical bond due to recovery of GAG content^[@CR95]^. A positive effect of joint loading (moderate to ultra-long running) on chondrocyte function with increased GAG and collagen synthesis has been reported in multiple animal studies^[@CR94],[@CR96]--[@CR100]^, and increased GAG has been reported in a few studies on running^[@CR95]^ and weight-bearing^[@CR101]^ in human knee cartilage. Since increasing hydrostatic pressure up-regulates proteoglycans and type-II collagen mRNA expression^[@CR102]^, de novo syntheses of proteoglycans will be initiated, and their increased concentration has been shown to impede hydraulic fluid flow^[@CR103]^. This mechanism shields the collagen-proteoglycan matrix from high stress and protects against cartilage degeneration and OA^[@CR94]^. In addition to these MR findings, changes in the serum concentrations of cartilage biomarkers of TEFR subjects also showed that articular cartilage is able to adapt during an MSUM^[@CR104]^. The observation of a secondary T2\\* decrease at t3 in cartilage regions with preexisting higher-grade cartilage degeneration after the above-average signal increase in 2 of our subjects (see Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}) may also support the hypothesis of ongoing T2\\* recovery in the FTJ of ultra-runners with an ongoing running burden beyond a distance of 3,500\u2009km.\n\nAs animal experiments have shown that ongoing ultra-long load bearing exercises minimize the development of OA in rats^[@CR96]^ and dogs^[@CR83]^, and as our collective of highly adapted and experienced ultra-runners^[@CR105]^ did not show increased cartilage degeneration, we found significant recovery of T2 values after TEFR (tx), as reported in all three published post-marathon studies with a second follow-up^[@CR34],[@CR39],[@CR40]^. Thus, ongoing ultra-endurance running burden does not correspond to increased OA or overuse lesions in the human knee joint when relevant cartilage injuries or malalignments are absent^[@CR16],[@CR106]^. Therefore, despite the problems that have been discussed with overuse injuries in ultra-marathon running^[@CR107]--[@CR114]^, MSUMs do not seem to have a negative effect, but rather a favourable influence, on cartilage.\n\nDue to the small number of non-finishers (3) and subjects with dominance of the left leg (3), sufficient data for statistical comparison of finishers or dominance of the right leg were not available (Table\u00a0[1](#Tab1){ref-type=\"table\"}). Arokoski *et al*.^[@CR55]^ showed side dependence in their experimental animal studies and tried to explain this by a side-dependent adaptation of the knee cartilage to mechanical loadings. The different degrees and types of loading on articular surface areas in the canine FTJ correspond to our observation of right to left side differences of the T2\\* cartilage response to the ultra-long-distance TEFR: a significant inter-zone T2\\* difference in all segments was found on the right but not regularly on the left, and a significant T2\\* increase in the medial compared with the lateral segments was found on the right but not on the left. As we also found comparable chondral T2\\* side differences in the ankle and hindfoot joints of TEFR subjects^[@CR70],[@CR71]^, which may be due to the right leg experiencing higher load and ground repulsion forces during the stance phase. The asymmetry of locomotion, which is based on the tendency of humans to use one leg preferentially during motor movements^[@CR115]--[@CR117]^, could possibly also be present during MSUM running. However, this explanatory model remains speculative, especially since this study has some limitations that do not allow any further reliable statistical evaluations about relevant causes and influencing factors with regard to the detected T2\\* reactions during the TEFR.\n\nLimitations {#Sec15}\n-----------\n\nThis study is subject to possible selection bias, as participants had been running ultra-marathons for almost a decade prior to enrolment. Therefore, these results may not be generalizable to novice individuals who begin ultra-long-distance running. The age dependency of the articular cartilage matrix content^[@CR118]^ is also a source of possible bias on the T2\\* measurement in this study, but we do not consider its relevance to be very strong, since the age range among the subject groups was not large (Table\u00a0[1](#Tab1){ref-type=\"table\"}). In addition, the test subjects did not adhere to a strict resting phase before the baseline measurement but carried out light running training (only a few kilometres per day during the last week), which could result in a discrete bias: slightly higher T2\\* values measured at baseline compared with a collective with a sufficiently long resting phase before the start of the TEFR. It must be assumed that the described immediate T2 decrease is present in parallel with the mechanism of the primary long-distance T2 increase. The amount of this \"masking\" effect and the attenuation of the real T2\\* increase and the influence of the stage run burden on T2\\* mapping in general is difficult to estimate^[@CR79]^. Including a control group was not feasible due to race conditions and the study setting.\n\nThe results of this mobile MRI-based field study allow us to refute a relevant negative influence of ultra-long running distances on knee joint cartilage in ultra-marathoners^[@CR106]^. However, in addition to the factor of running distance, there are other external running-specific characteristics (e.g., competition runners, mountain running with inclines and descents, cross-country running with enhanced multiaxial stress on the knee joints., interval running with partially high running speed, etc.) that may increase the risk of developing knee OA^[@CR119],[@CR120]^. These factors were not included in our investigation and require further clarification in other studies^[@CR121]^.\n\nConclusions {#Sec16}\n===========\n\nThe measured chondral T2\\* values showed similar physiological regional distribution patterns and relative differences between cartilage areas of the FTJ as in normal, healthy non-endurance and other athletic collectives, although these values were higher throughout the TEFR than at baseline. Due to initial chondral matrix degradation within the first 1,100\u2009km of running, a significant, primarily non-regional-related T2\\* increase occurred in all cartilage areas of the FTJ, caused by collagen disorientation and decrease and GAG depletion. As a further increase in the second half of the TEFR was not observed, we do not expect further softening or degradation of the FTJ cartilage with ongoing running burden over ultra-lengthy periods extending to 4,500\u2009km. However, we expect the FTJ cartilage matrix to be able to reorganize itself, as the T2 \\* curve showed a decreasing trend (recovery) after 3,500\u2009km. Therefore, our results are not consistent with the hypothesis that ultra-long-distance running in specifically trained athletes may be associated with increased incidence and severity of knee joint OA. The observed side differences in regional T2\\* changes are in line with the hypothesis of the 'braking' limb and also indicate side dominance of one leg in MUSM runners, which must be verified by specific studies in the future. This study shows the feasibility of mobile functional MRI for precise quantitative non-invasive measurements to evaluate the biochemical long-term effects of MSUM running on knee joint cartilage.\n\nSupplementary information\n=========================\n\n {#Sec17}\n\nSupplementary information\n\n**Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary information\n=========================\n\nis available for this paper at 10.1038/s41598-020-64994-2.\n\nWe would like to thank all the athletes of the TEFR who took part at this project. Considering their immense physical and mental stresses they showed extraordinary compliance on every day of the race. Heike Wiedelbach, our radiographer, did an excellent and outstanding job throughout the whole TEFR. The TEFR-project was mainly supported by the German Research Association (DFG: 'Deutsche Forschungsgemeinschaft'), under Grants SCHU 2514/1--1 and SCHU 2514/1--2. Other non-public funds were received from Siemens Healthcare Ltd. and the Medical Faculty of the University of Ulm. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No additional external funding was received for this study.\n\nAll authors read and approved the final manuscript: Sch\u00fctz U conceived the study (conception and design), implemented the project (including administrative, technical and logistical support), obtained the funding, participated in the data collection (MRI), data assembly and data evaluation with statistical analysis and drafted the article. Ehrhardt M participated in the implementation of the project, the data collection (MRI) and evaluation. Billich C participated in the implementation of the project (including administrative, technical and logistical support) and the data collection (MRI). G\u00f6d S participated in the data evaluation. Beer M supervised the publication process and did final approval of the article. Trattnig S supervised cartilage T2\\* evaluation, gave critical revision of the article for important intellectual content and did final approval of the article.\n\nAdditional materials, data and associated protocols are promptly available to readers without undue qualifications in material transfer agreements.\n\nThe authors declare no competing interests.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nGreen fodder produced from legume/cereal mixtures is a valuable source of dairy cattle feed, providing it has a beneficial chemical composition, including a low concentration of heavy metals (Ociepa-Kubicka and Ociepa [@CR15]; Chandra Sekhar et al. [@CR7]). Thus, it is necessary to check its safety in terms of impurities present in feeds, including above-ground parts. Monitoring of toxic and potentially toxic elements is one of the most important aspects of maintaining feed quality. Transfer of heavy metals from soil to animal body occurs through plants which are the most important link in the food chain. Heavy metals contaminating the soil hamper the development of soil microorganisms. Negative correlations between the biomass of microorganisms and Pb, Zn, and Cu contents in soils contaminated with metals have been reported by Ociepa-Kubicka and Ociepa [@CR15] as well as Sharma et al. [@CR19]. Heavy metal uptake by plants occurs through the rooting system and leaf blades. Heavy metals display various harmful effects. Lead, cadmium, chromium, and nickel are perceived as toxic metals but zinc and copper are microelements which are harmful when present in excess whereas their low amounts are necessary for the appropriate functioning of the body (Ali and Al-Qahtani [@CR4]; Abdulmojeed and Abdulrahman [@CR1]; Sharma et al. [@CR19]). The following plants have the greatest capacity to accumulate heavy metals: lettuce, cabbage, beetroot, carrot, spinach, and potato, whereas the uptake by tomato, cucumber, leguminous, and cucurbitaceous vegetables as well as fruit is much lower. Cereal grain, and as a result green fodder, can be contaminated as well (Ali and Al-Qahtani [@CR4]; Sobucola et al. [@CR21]). Too high levels of metals in plants destined for animal feed may lead to animal-derived products being contaminated by the metals. Legumes are of particular importance as, when mixed with cereals, they reduce the uptake of heavy metals from the soil and limit their concentration in the soil (Dabney et al. [@CR8]; Snapp et al. [@CR20]). There is a perceptible lack of studies on the aforementioned subject. Thus, the need arises to undertake research in this area to monitor heavy metals in animal feed. The objective of the study reported here was to determine the effect of component share in the mixture and harvest date on the concentration of selected heavy metals in the green fodder of field pea, oat, and their mixtures.\n\nMaterials and methods {#Sec2}\n=====================\n\nField research was conducted in the years 2010 to 2012 at the Zawady Experimental Farm (52\u00b0 03\u2032 39\u2033 N, 22\u00b0 33\u2032 80\u2033 E) which belongs to Siedlce University of Natural Sciences and Humanities. The experimental soil was Albic Luvisol (Arenic). The concentration of elements in the soil was as follows: P 5.21 mg 100 g^\u22121^, K 11.3 mg 100 g^\u22121^, Cu 1.3 mg 100 g^\u22121^, Zn 7.8 mg 100 g^\u22121^, Pb \\< 15.1 mg 100 g^\u22121^, Cd 0.10 mg 100 g^\u22121^, Cr\\< 12.6 mg 100 g^\u22121^, Ni \\< 12.6 mg 100 g^\u22121^. Soil reaction was neutral and humus content amounted to 1.39%. The experimental design was a split-block arrangement with three replicates. The area of plots under winter wheat was 30 m^2^ (5 m \u00d7 6 m), and the harvested area was 20 m^2^ (4 m \u00d7 5 m). Two factors were tested in the study: I---component share in the mixture: field pea---pure stand 100%, oat---pure stand 100%, field pea 75% + oat 25%, field pea 50% + oat 50%, field pea 25% + oat 75%; II---harvest date: field pea flowering stage, field pea flat green pod stage. The species proportion in a mixture was established in relation to the number of seeds planted in pure stands, that is 560 grains of oat and 120 grains of field pea m^2^. The following sowing rates were used: field pea 170 kg \u05bcha^\u22121^, oat 180 kg \u05bcha^\u22121^, field pea 128 kg\u05bc ha^\u22121^ + oat 45 kg \u05bcha^\u22121^, field pea 85 kg \u05bcha^\u22121^ + oat 90 kg \u05bcha^\u22121^, field pea 43 kg \u05bcha^\u22121^ + oat 135 kg \u05bcha^\u22121^.\n\nIn all the study years, the mixtures were preceded by winter triticale. Phosphorus and potassium fertilisers were applied in autumn and their rates, 35.2 kg \u05bcha^\u22121^ P in the form of granular triple superphosphate and 99.6 kg \u05bcha^\u22121^ K in the form of 60% potassium salt, depended on soil chemical composition. In spring, nitrogen fertiliser, at the rate of 30 kg \u05bcha^\u22121^ N in the form of ammonium nitrate, was applied preplant to all the plots, excluding the units assigned to field pea grown in pure stand. At the stage of stem elongation, additional N was applied (50 kg \u05bcha^\u22121^ for oat and 30 kg \u05bcha^\u22121^ for field pea/oat mixtures). Field pea and oat seeds were planted in early April as described for the first experimental factor. Plants were harvested in late June and early July. During harvest of mixtures, fresh matter samples were taken from each plot to determine microelements. Cu, Zn, Cd, Pb, Cr, and Ni contents were determined by means of inductively coupled plasma optical emission spectrometry (ICP-OES) using the spectrometer Perkin Elmer Optima 8300.\n\nEach characteristic tested was subjected to analysis of variance suitable for the split-block design. When significant sources of variation were confirmed, their means were separated using Tukey test. Calculations were performed in MS Excel 12.0.\n\nThe years of the study were characterised by significantly changeable weather conditions (Table [1](#Tab1){ref-type=\"table\"}). In 2010, mean air temperatures during the growing period fluctuated around the mean long-term temperatures. The precipitation totals, except for April, were higher than the mean long-term total precipitation. This year should be regarded as favourable for the cultivation of mixtures of field peas with oat. In 2011, the mean monthly air temperatures slightly differed from the mean long-term temperatures. However, the rainfall totals were lower than the mean long-term totals, except for July, where the recorded precipitation was 120.2 mm. In 2012, the average air temperature and total precipitation fluctuated around the long-term average. The lowest content of heavy metals was recorded in the green mass of mixtures of pea and oats in 2010, with the highest rainfall in June and July, and the highest content of heavy metals in 2012 with the lowest rainfall in July and throughout the growing season.Table 1Pluvio-thermal conditions in the growing season of pea and oat mixtures in 2010--2012 according to the Meteorological Station in RSD ZawadyYearMonthMeanAprilMayJuneJulyTemperature \u00b0C20108.914.017.421.615.5201110.113.418.118.315.020128.914.616.320.715.1Long-term mean 1990--20058.214.217.619.714.9Precipitation, mm201010.793.262.677.0243.5201131.036.139.1120.2226.4201229.953.476.243.0202.5Long-term mean 1990--200537.447.148.165.5198.1\n\nResults and discussion {#Sec3}\n======================\n\nCopper content in the green matter of field pea/oat mixtures was significantly affected by the experimental factors and their interaction (Table [2](#Tab2){ref-type=\"table\"}). The highest copper content (4.660 mg \u05bckg^\u22121^ d.m.) was recorded in field pea grown in pure stand, it being the lowest in oat (3.811 mg \u05bckg^\u22121^ d.m.). Also, Wo\u017aniak and Soroka ([@CR25]), Jarecki and Bobrecka-Jamro ([@CR12]), and Ali and Al-Qahtani ([@CR4]) observed a lower copper content in cereal grain rather than soy bean seed. In the experiment reported here, copper content in field pea, although higher than in oat, did not exceed the WHO/FAO standards, such feed being safe for animals. It should be emphasised that more copper was accumulated by legumes, although its amounts were low and did not pose a threat to human and animal health (Adefarati et al. [@CR2]; Chandra Sekhar et al. [@CR7]; Zarcinas et al. [@CR27]). Plants take up small amounts of copper which, transferred to animals eating the plants, is necessary for the body because it participates in oxidation-reduction processes where it is a component of the coenzyme, and regulates metabolism, iron transportation, and collagen metabolism (Cabrera et al. [@CR6]). In the present study, copper content in field pea/oat mixtures was significantly lower than in field pea grown in pure stand. The lowest Cu concentration was recorded in the mixture composed of the following respective shares of field pea and oat: 50 + 50% and 25 + 75%. A similar relationship was observed by Tr\u0105ba and Wola\u0144ski ([@CR24]) who reported that copper content in the dry matter of a grass/legume mixture was about twice as low as in legumes grown in pure stand. In the study reported here, harvest date had a significant influence on copper content in the green fodder of field pea/oat mixtures. A higher concentration of copper was recorded in field pea/oat mixtures harvested at the stage of field pea flowering compared with the stage of flat green pod. Research by Ladipo and Doherty ([@CR13]) as well as Abdulmojeed and Abdulrahman ([@CR1]) demonstrated that copper uptake by leafy vegetables is affected by such factors as climate, precipitation, concentration of heavy metals in soil, soil type, and maturity of plants at harvest. Fytianos et al. ([@CR10]), Demirezen and Aksoy ([@CR9]), and Muchuweti et al. ([@CR14]) have claimed that plants harvested at earlier development stages contain more copper. In the present study, an interaction between the experimental factors was confirmed indicating that field pea grown in pure stand and harvested at flowering contained the highest amount of copper, it being the lowest in pure stand oat harvested at the stage of flat green pod. All the test field pea/oat mixtures harvested at this stage had a lower copper content compared with field pea cultivated in pure stand and harvested at the flat green pot stage.Table 2Copper content in the green fodder of field pea/oat mixtures (means across 2010--2012), mg kg^\u22121^ d.m.Component share in the mixture, %Harvest dateMeansField pea flowering stageField pea flat green pod stageField pea in pure stand 100%5.0724.2484.660Oat in pure stand 100%4.2453.3763.811Field pea 75% + oat 25%4.6933.9144.304Field pea 50% + oat 50%4.4583.7084.083Field pea 25% + oat 75%4.2073.5373.872Means4.5353.757-LSD~0.05~Component share in the mixture0.212Harvest0.130Interaction0.272\n\nStatistical analysis demonstrated a significant effect of the experimental factors and their interaction on zinc content in the green fodder of field pea/oat mixtures (Table [3](#Tab3){ref-type=\"table\"}). Zinc plays a major role in plant metabolism. Plant growth and development are hindered by both zinc shortage and excess (Ociepa-Kubicka and Ociepa [@CR15]). Although zinc is necessary in plant nutrition, plants growing in a polluted environment may accumulate high amounts of this element, which can pose a serious threat to the health of people and animals (Yu-Wei et al. [@CR26]; Sharma et al. [@CR19]; Srinivas et al. [@CR22]). In the present study, the highest zinc content was recorded in field pea grown in pure stand, it being the lowest in oat. Also Adefarati et al. ([@CR2]) recorded a similar zinc content in green peas. However, the values are lower than standards set by WHO/FAO and are not harmful to humans or animals. According to Fytianos et al. ([@CR10]), Demirezen and Aksoy ([@CR9]), and Ali and Al-Qahtani ([@CR4]), zinc content in cereals is lower compared with beans and peas. In the experiment reported here, field pea mixed with oat increased zinc content in the green fodder of mixtures, the values being lower than in field pea but higher than in oat. This finding corresponds to values reported by Goli\u0144ski et al. ([@CR11]) and Szpunar-Krok et al. ([@CR23]). Harvest date had a significant effect on zinc content in the green fodder of field pea/oat mixtures. A higher concentration of this element was recorded in mixtures harvested at the stage of field pea flat green pod compared with the flowering stage. A delay in harvest was followed by an increase in the plant content of zinc. As zinc is a microelement, its higher amount in green fodder is of great importance for human and animal health (P\u0142aza et al. [@CR18]; Adefarati et al. [@CR2]; Chandra Sekhar et al. [@CR7]; Zarcinas et al. [@CR27]). In the present study, an interaction between the experimental factors was confirmed. The highest zinc content was recorded in the green fodder of field pea grown in pure stand and harvested at the stage of flat green pot, it being the lowest in oat grown in pure stand and harvested at the stage of field pea flowering. Of the examined mixtures, the highest concentration of zinc was determined in field pea/oat mixtures whose component shares were 75 + 25% and 50 + 50%, and which were harvested at the stage of field pea flat green pot.Table 3Zinc content in the green fodder of field pea/oat mixtures (means across 2010--2012), mg kg^\u22121^ d.m.Component share in the mixture, %Harvest dateMeansField pea flowering stageField pea flat green pod stageField pea in pure stand 100%41.0856.3748.73Oat in pure stand 100%29.0838.2333.66Field pea 75% + oat 25%39.2350.3244.78Field pea 50% + oat 50%36.9748.7842.88Field pea 25% + oat 75%33.1240.5436.83Means35.9046.85-LSD~0.05~Component share in the mixture2.31Harvest date1.43Interaction3.12\n\nLead and cadmium contents were insignificantly affected by the experimental factors and their concentrations in the green fodder of field pea/oat mixtures were too low to be determined by means of the emission spectrometer Perkin Elmer Optima 8300. Thus, they did not exceed the standards set for green fodder in the Official Journal of the European Union ([@CR16]). It was due to the fact that fields where the research was conducted were well away from roads and the green matter of field pea/oat mixtures was a safe cattle feed. The air and soil are major sources of heavy metals for plants (Adefarati et al. [@CR2]; Brigide et al. [@CR5]). Despite low Cd and Pb contents in the green fodder of field pea/oat mixtures determined in the present study, it is necessary to continually monitor and check plant-derived feeds in terms of their heavy metal content.\n\nStatistical analysis confirmed a significant effect of the experimental factors and their interaction on chromium content in the green fodder of field pea/oat mixtures (Table [4](#Tab4){ref-type=\"table\"}). The lowest chromium content was recorded in oat cultivated in pure stand, it being the lowest in field pea cultivated in pure stand. Field pea mixed with oat significantly reduced chromium content in green fodder. Some studies demonstrated excessive Cr amounts in cereals and rice (Akinyele and Shokunbi [@CR3]; Pirsaheb et al. [@CR17]; Brigide et al. [@CR5]; Ali and Al-Qahtani [@CR4]; Abdulmojeed and Abdulrahman [@CR1]). In contrast, legumes are characterised by a lower concentration of chromium (Brigide et al. [@CR5]; Akinyele and Shokunbi [@CR3]). In the current study, there was confirmed a significant effect of harvest date on chromium content in the green fodder of field pea/oat mixtures. A higher concentration of this element was recorded in the field pea/oat mixture harvested at the stage of field pea flat green pod versus the flowering stage, which can be explained by the fact that plants harvested at later dates take up more chromium from the soil (Adefarati et al. [@CR2]). However, in the experiment discussed here, chromium content in the green fodder of field pea/oat mixtures harvested at the stage of flat green pod was not high and posed no threat to animals. Similar findings were reported by Adefarati et al. ([@CR2]), Ladipo and Doherty ([@CR13]), and Sobucola et al. ([@CR21]) who demonstrated that chromium content in the plants they tested (cereals and legumes) did not exceed WHO/FAO standards. Chromium levels in plants are affected by their maturity at harvest. Small amounts of chromium are essential for life as the element plays a significant role in the metabolic transformations of glucose, some proteins, and fats (Ociepa-Kubicka and Ociepa [@CR15]; Sobucola et al. [@CR21]). In the study presented here, an interaction between the experimental factors was confirmed. The highest chromium content was recorded in the green fodder of field pea grown in pure stand and harvested at the flowering stage or the flat green pod stage, and in field pea/oat mixtures whose component shares were as follows: 75 + 25% and 50 + 50%, and which were harvested at the stage of field pea flowering. The highest chromium content was determined in oat grown in pure stand and harvested at the stage of field pea flat green pod.Table 4Chromium content in the green fodder of field pea/oat mixtures (means across 2010--2012), mg kg^\u22121^ d.m.Component share in the mixture, %Harvest dateMeansField pea flowering stageField pea flat green pod stageField pea in pure stand 100%1.3531.6961.525Oat in pure stand 100%3.5606.9145.237Field pea 75% + oat 25%1.8743.4232.649Field pea 50% + oat 50%2.1574.0723.115Field pea 25% + oat 75%2.9815.3404.161Means2.3854.289-LSD~0.05~Component share in the mixture0.623Harvest date0.249Interaction0.872\n\nNickel content in the green fodder of field pea mixed with oat was significantly affected by the experimental factors and their interaction (Table [5](#Tab5){ref-type=\"table\"}). The lowest nickel content was recorded in field pea grown in pure stand, it being the highest in oat cultivated in pure stand. Ni levels determined in the present study did not exceed WHO/FAO standards (Adefarati et al. [@CR2]). Small amounts of nickel are necessary for plant growth and development (Akinyele and Shokunbi [@CR3]), the element being toxic when present in excessive amounts (Cabrera et al. [@CR6]). It should be stressed that, in the study reported here, field pea mixed with oat contributed to a decline in the green fodder content of nickel. Harvest date significantly affected Ni content in the green fodder of field pea/oat mixtures. Nickel content was lower in plants harvested at the stage of field pea flowering compared with the stage of flat green pod. Also, Ladipo and Doherty ([@CR13]) as well as Sobucola et al. ([@CR21]) reported that nickel content in plants was influenced by their maturity stage at harvest. In the present study, interaction between the experimental factors was confirmed. The lowest nickel content was determined in the green fodder of pure stand field pea harvested at the flowering stage and the stage of flat green pod, and in field pea/oat mixtures with the following shares of components: 75 + 25% and 50 + 50% harvested at the stage of field pea flowering. Nickel content was the highest in oat harvested at the stage of field pea flat green pod.Table 5Nickel content in the green fodder of field pea/oat mixtures (means across 2010--2012), mg kg^\u22121^ d.m.Component share in the mixture, %Harvest dateMeansField pea flowering stageField pea flat green pod stageField pea in pure stand 100%1.5571.9721.765Oat in pure stand 100%3.2825.7054.494Field pea 75% + oat 25%1.9262.7472.337Field pea 50% + oat 50%2.2153.5382.877Field pea 25% + oat 75%2.7434.2163.480Means2.3453.636-LSD~0.05~Component share in the mixture0.593Harvest date0.189Interaction0.732\n\nConclusions {#Sec4}\n===========\n\nField pea grown in pure stand had the highest copper and zinc contents, and the lowest chromium and nickel contents. Field pea mixed with oat contributed to a significant decline in the green fodder content of heavy metals.\n\nField pea/oat mixtures harvested at the stage of field pea flowering contained more copper but less zinc, chromium, and nickel compared with mixtures harvested at the stage of field pea flat green pod.\n\nCadmium and lead contents in the green fodder of field pea mixed with oat were insignificantly affected by the experimental factors, and were too low to be determined by means of the emission spectrometer Perkin Elmer Optima 8300.\n\nGreen fodder of field pea mixed with oat, whether harvested at the flowering stage or the stage of field pea flat green pod, was safe for cattle because it did not contain excessive amounts of heavy metals.\n\nDespite a low heavy metal content in the green fodder of field pea mixed with oat, continual monitoring of these elements is recommended.\n\n**Publisher's note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Introduction\n============\n\nPhiladelphia chromosome-negative myeloproliferative neoplasms (MPNs) are a diverse group of clonal stem cell disorders derived from hematopoietic myeloid progenitors and include polycythemia vera (PV), essential thrombocythemia (ET), and myelofibrosis (MF).[@b1-btt-7-189] Of the classical MPNs, MF has the worst prognosis, with a median survival of 69 months.[@b2-btt-7-189] MF can present as a primary (PMF) disorder or evolve from PV or ET to post-PV MF or post-ET M F, respectively. MF is characterized by bone marrow fibrosis, splenomegaly, leukoerythroblastosis, extramedullary hematopoiesis, and a collection of debilitating symptoms. These symptoms include cachexia, worsening of patient role functioning, and quality of life (QoL) -- which ultimately result in increased rates of leukemic transformation -- and decreased survival.[@b3-btt-7-189]\n\nUntil recently, therapeutic options for patients with MF consisted of allogeneic hematopoietic stem cell transplant (alloHSCT), the use of cytoreductive agents (ie, hydroxyurea), splenectomy and splenic irradiation for treatment of splenomegaly, and management of anemia with transfusions, erythropoiesis-stimulating agents (ESAs), androgens, and immunomodulatory agents (IMiDs). Fortunately, our understanding of MPNs and the molecular mechanisms of the disease has been rapidly expanding. In 2005, the Janus kinase (*JAK*)*2* V617F mutation was discovered and observed in approximately 50%--60% of patients with PMF or ET and 90%--95% of patients with PV.[@b4-btt-7-189]--[@b7-btt-7-189] This discovery, along with the observation of other mutations in patients with MPNs found to activate the JAK/STAT (signal transducers and activators of transcription) pathway (*JAK2* exon 12, *MPL*, and *LNK*),[@b8-btt-7-189]--[@b10-btt-7-189] has established dysregulation of the JAK signaling pathway as the major contributor to the pathogenesis of MPNs. It has also translated into the development of small-molecule JAK inhibitors, the first of which, ruxolitinib, has been approved for the treatment of MF.\n\nThe purpose of this article is to review the clinical features of MF, discuss the use and future of JAK inhibitors, and reassess when and how to use conventional MF treatments in the context of JAK inhibitors.\n\nClinical features and prognosis of myelofibrosis (MF)\n=====================================================\n\nMF has an incidence of 0.5--1.5 per 100,000 individuals, and the median age of MF diagnosis is 67 years, with an equal distribution between the sexes. The clinical features of MF are similar regardless of the subtype and include anemia, leucopenia or leukocytosis, thrombocytopenia or thrombocytosis, and multi-organ extramedullary hematopoiesis -- commonly causing hepatomegaly, symptomatic splenomegaly, or portal hypertension. The single most common cause of death for patients with MF is transformation to acute myeloid leukemia (20%); however, most patients die from other disease-related events, such as progression without transformation and thrombotic or cardiovascular events.[@b2-btt-7-189]\n\nPatients with MF often have a substantial symptom burden, including constitutional symptoms such as fatigue, cachexia, pruritus, bone pain, and fever, and symptoms related to the presence of massive splenomegaly such as pain, early satiety, splenic infarction, and dyspnea. The severity of symptoms has been found to be similar to that of advanced cancer, resulting in a diminished QoL.[@b11-btt-7-189] Moreover, with median survival ranging between 2 and 15 years, patients often suffer for a long period with this considerable symptom burden.\n\nSeveral prognostic scoring systems have been developed for patients with MF, including the International Prognostic Scoring System (IPSS) used at the time of diagnosis[@b2-btt-7-189] and the dynamic IPSS (DIPSS)[@b12-btt-7-189],[@b13-btt-7-189] and DIPSS plus,[@b14-btt-7-189] which can be used to assess patients during the course of their disease ([Table 1](#t1-btt-7-189){ref-type=\"table\"}). It should be noted that these scoring systems have been validated only for patients with PMF. Nevertheless, they are often used for patients with post-ET MF and post-PV MF, despite potential inaccuracies associated with extrapolation to these populations. The IPSS comprises five risk factors for estimating survival from the time of diagnosis: age \\>65 years, hemoglobin (Hb) level \\<10 g/dL, leukocyte count \\>25 \u00d7 10[@b9-btt-7-189]/L, circulating blasts \u22651%, and presence of constitutional symptoms (eg, fever, weight loss, night sweats). Low-, intermediate-1\u2212, intermediate-2\u2212, and high-risk disease are defined as the presence of 0, 1, 2, and \u22653 adverse factors, respectively, with median survivals of 11.3, 7.9, 4, and 2.3 years.[@b2-btt-7-189] The DIPSS is a modification of the IPSS scale, in which two points are assigned for Hb level \\<10 g/dL. Unlike the IPSS, which is valid only at diagnosis, the DIPSS can be used to evaluate a patient's risk category during the course of the disease.[@b12-btt-7-189] The DIPSS plus, a modification of the DIPSS, incorporates three additional independent risk factors: red blood cell transfusion requirements, platelet count \\<100 \u00d7 10[@b9-btt-7-189]/L, and unfavorable karyotype (complex karyotype or one or two abnormalities, including trisomy 8, monosomy 7/7q\u2212, isochromosome \\[17q\\], inversion \\[3\\], deletion 5/5q\u2212, 12p\u2212, or 11q23 rearrangement).[@b14-btt-7-189] The DIPSS plus permits the identification of very low- and high-risk patients compared with the IPSS or DIPSS. For this reason, where it might alter the management, we recommend performing a karyotype for patients at diagnosis and for selected patients to follow the course of their disease.\n\nOur improved molecular understanding of MF is illustrated by the recent identification of mutations in *ASXL1*, *EZH2*, *IDH1/2*, and *SRSF2* that were associated with worse survival outcomes. If these data are validated, screening for these mutations could be used to identify patients in the IPSS groups who may have a greater likelihood of transforming to acute leukemia and could benefit from more aggressive or experimental therapies.[@b15-btt-7-189] However, at present, screening for such mutations is not carried out in routine practice nor is it incorporated into prognostic scores.\n\nJanus kinase inhibitors for the treatment of MF\n===============================================\n\nRuxolitinib\n-----------\n\nAs mentioned previously, discovery of the *JAK2* V617F mutation and an understanding of dysregulated JAK-STAT signaling in the pathogenesis of MF have led to the development of small-molecule JAK inhibitors. Ruxolitinib (Jakavi, Novartis AG, Basel, Switzerland; Jakafi, Incyte Corporation, Wilmington, DE, USA) is the first JAK inhibitor to gain approval in the USA, Canada, and Europe.[@b16-btt-7-189] These approvals were based on data from two randomized Phase III trials: the COntrolled MyeloFibrosis Study With ORal JAK Inhibitor Treatment (COMFORT) trials, which were conducted in patients with primary, post-ET, or post-PV MF with intermediate-2- or high-risk disease as assessed by IPSS and platelet count \\>100 \u00d7 10[@b9-btt-7-189]/L.[@b17-btt-7-189],[@b18-btt-7-189] In COMFORT-I, patients (N = 309) were randomized 1:1 to ruxolitinib or placebo; in COMFORT-II, patients (N = 219) were randomized 2:1 to ruxolitinib or best available therapy (BAT). In both trials, patients received ruxolitinib 15 or 20 mg twice daily based on their baseline platelet count (100--200 or \\>200 \u00d7 10[@b9-btt-7-189]/L, respectively).\n\nThe primary endpoint of both trials was achieved with a proportion of patients in the ruxolitinib arms exhibiting a \u226535% reduction in spleen volume as measured by magnetic resonance imaging at 24 weeks in COMFORT-I (41.9% ruxolitinib versus \\[vs\\] 0.7% placebo; *P* \\< 0.0001) and at 48 weeks in COMFORT-II (28.5% ruxolitinib vs 0% BAT; *P* \\< 0.0001).[@b17-btt-7-189],[@b18-btt-7-189] The spleen responses in both studies were observed regardless of *JAK2* V617F mutation status. Furthermore, spleen responses were durable, with 67.0% and 79.9% of responding patients in COMFORT-I and -II, respectively, maintaining their response for \u226548 weeks. With longer follow-up in both COMFORT-I and -II (median 102 and 112 weeks, respectively), the median duration of response to ruxolitinib had not been reached.[@b19-btt-7-189],[@b20-btt-7-189]\n\nThe COMFORT trials also demonstrated that, in addition to the profound effects on splenomegaly, ruxolitinib provided statistically significant improvements in patients' symptoms and QoL.[@b17-btt-7-189],[@b18-btt-7-189] Improvements in MF symptoms were rapid, with the majority of responses occurring within the first 4 weeks of ruxolitinib treatment. In COMFORT-I, there was a \\>50% improvement in the Myelofibrosis Symptom Assessment Form Total Symptom Score at 24 weeks in 45.9% of ruxolitinib patients compared with 5.3% of placebo patients (*P* \\< 0.001). Long-term follow-up of COMFORT-I (median 102 weeks) demonstrated that ruxolitinib treatment was associated with durable clinically significant improvements in global health status/QoL and the other functional domains of the European Organisation for Research and Treatment of Cancer QoL Questionnaire--Core 30 Items.[@b18-btt-7-189]\n\nConsistent with ruxolitinib's known mechanism of action as a JAK pathway inhibitor, anemia and thrombocytopenia were the most frequently reported adverse events (AEs) overall and of grade \u22653 in the ruxolitinib arms of both studies ([Table 2](#t2-btt-7-189){ref-type=\"table\"}). In both studies, Hb levels reached a nadir at week 12 and then stabilized at an average reduction of about 1 g/dL below baseline at week 24. Anemia and thrombocytopenia rarely led to treatment discontinuation (\\<1% of patients in any treatment group) and were manageable with dose modi-fications and/or blood transfusions. Rates of grade 3/4 non-hematologic AEs were low in both COMFORT studies.\n\nA survival analysis from COMFORT-I indicated a significant survival advantage with ruxolitinib therapy compared with placebo with a median follow-up of 51 weeks (hazard ratio \\[HR\\] 0.50; 95% confidence interval \\[CI\\] 0.25--0.98; *P* = 0.04).[@b18-btt-7-189] In additional follow-up at the 2-year time point, 41 patients randomized to placebo and 27 patients randomized to ruxolitinib died, representing a continued overall survival advantage in favor of ruxolitinib (HR 0.58; 95% CI 0.36--0.95; *P* = 0.028; [Figure 1](#f1-btt-7-189){ref-type=\"fig\"}).[@b21-btt-7-189] Furthermore, with a median follow-up of 112 weeks in COMFORT-II, patients randomized to ruxolitinib had longer overall survival than those randomized to BAT (HR 0.51; 95% CI 0.26--0.99; *P* = 0.041; the *P*-value from a log-rank test is provided for descriptive purposes only and was not adjusted for multiple comparisons).[@b20-btt-7-189] Potential reasons for this survival advantage included improved performance status, reduction of proinflammatory cytokines, improved nutritional status, and better overall physical functioning. Recent data, though only provisional, suggest that ruxolitinib therapy may lead to reduction of marrow fibrosis in a proportion of patients, while in other patients, fibrosis has remained either stable or indeed progressed.[@b22-btt-7-189],[@b23-btt-7-189] Furthermore, a modest reduction of mutant allele burden has also been reported.[@b24-btt-7-189]\n\nResponses to ruxolitinib are typically observed within the first 3--6 months after therapy initiation. For patients who have not had a reduction in spleen size or improvement in symptoms after this period, alternative therapies should be considered. In patients with some symptom response, symptoms returned to baseline levels within 1 week of discontinuing ruxolitinib. Therefore, dose tapering of ruxolitinib should be considered if a patient needs to discontinue ruxolitinib therapy. After 2 years of follow-up, no consistent pattern of AEs has been observed that would suggest a severe inflammatory syndrome after ruxolitinib discontinuation.\n\nOther JAK inhibitors\n--------------------\n\nSeveral other JAK inhibitors are in various stages of development.\n\nSAR302503 (Sanofi; Paris, France) is a selective JAK2 inhibitor that has also shown some inhibitory activity against *FLT3* and *RET*.[@b25-btt-7-189] In a Phase II study, patients were randomized to SAR302503 at 300 mg, 400 mg, or 500 mg once daily (N = 31).[@b26-btt-7-189] Reductions in spleen volume \u226535% at the end of cycle 3 appeared to be dose dependent (30%, 50%, and 64% for patients in the 300 mg, 400 mg, and 500 mg arms, respectively), which notably correlated with inhibition of STAT3 phosphorylation. The proportion of patients who achieved \u226550% reduction in the Myeloproliferative Neoplasm Symptom Assessment Form score was 50%, 50%, and 39% across the dosing arms, respectively. An initial report suggested that SAR302503 was associated with reduction in allele burden and improvement in bone marrow fibrosis scores;[@b27-btt-7-189] however, no updated results of these findings have been reported. Rates of grade 3/4 anemia were 33%, 30%, and 55%, and rates of grade 3/4 thrombocytopenia were 20%, 0%, and 9% for patients in the 300 mg, 400 mg, and 500 mg dosing arms, respectively. The most common non-hematologic AEs were gastrointestinal (rates of all-grade diarrhea were 70%, 90%, and 55%, respectively), although the incidence of diarrhea declined during the course of therapy. Currently, a Phase III study, JAK2 inhibition in a single-arm trial of SAR302503 in MF patients (JAKARTA-1), is underway comparing SAR302503 400 mg and 500 mg once daily with placebo (N = 225); the primary endpoint is spleen response. Results from this trial are anticipated soon. SAR302503 is also being evaluated in another Phase II study (JAKARTA-2) in patients previously treated with ruxolitinib (target enrolled, N = 70).\n\nMomelotinib (CYT387; Gilead Foster City, CA, USA) is a JAK1/JAK2 inhibitor currently under evaluation in a Phase II trial of patients with MF (N = 166).[@b28-btt-7-189] Durable reductions in spleen length of \u226550% as assessed by palpation were observed in 37% of patients, and the median duration of spleen response was 744 days. Achievement of a complete resolution or marked improvement of constitutional symptoms was reported in the majority of patients. After a median follow-up of 16.9 months, a substantial decrease in the percentage of patients requiring transfusions during study was observed (44% at baseline vs \\<10% at week 40), and 13% of patients had an increase in Hb level of at least 2 g/dL. A Phase III study of CYT387 is underway, and it will be important to determine whether these encouraging results are confirmed in this setting.\n\nPacritinib (SB1518; Cell Technology, Inc, Mountain View, CA, USA) is a JAK2 and FLT3 inhibitor currently being evaluated at a dose of 400 mg daily in a Phase II study (N = 34) that included patients with low platelet counts (\\<50 \u00d7 10[@b9-btt-7-189]/L). Nearly one third of patients (32%) had a \u226535% reduction in spleen volume at 24 weeks.[@b29-btt-7-189] Pacritinib therapy was associated with minimal myelosuppression and no new onset of anemia or change in transfusion requirements. Furthermore, no dose reductions due to thrombocytopenia were required. A Phase III study of pacritinib in patients with low platelets and symptomatic splenomegaly is currently open (Oral Pacritinib Versus Best Available Therapy in Patients With Primary Myelofibrosis, Post-Polycythemia Vera Myelofibrosis, or Post-Essential Thrombocythemia Myelofibrosis; PERSIST).\n\nPerspective: we would recommend the use of ruxolitinib first line for patients with burdening symptoms and/or splenomegaly. As with any new therapy, the long-term benefits and safety profile of ruxolitinib and the other JAK inhibitors (in time) will need to be evaluated as more patients receive treatment over longer periods. There has been some evidence of an increased incidence of herpes zoster and tuberculosis reactivation in the ruxolitinib arms of the COMFORT studies,[@b17-btt-7-189],[@b18-btt-7-189] and it will be important to assess the risk of increased viral reactivation in patients treated with JAK inhibitors as well as other concerns of immunosuppression (eg, opportunistic infections and secondary malignancies). We also believe that it will be important to understand how disease progression may present in patients receiving JAK inhibitors because increases in spleen size and worsening of constitutional symptoms may be masked by treatment. Additionally, as more JAK inhibitors are approved for MF and become commercially available, it will be necessary to determine which patients may benefit the most from a particular agent. For example, momelotinib may be a better choice for patients with severe anemia, while ruxolitinib may be better for those who experience non-hematologic AEs with other JAK inhibitors such as SAR302503. At present, our understanding of the reasons for these apparent differences in anemia is unclear but has been related to differing binding affinities for the various members of the JAK family. It will also be important to determine if any biomarkers exist to help select the most appropriate inhibitors for a particular patient. Furthermore, we need to be able to assess long-term benefits in comparing these agents with each other and with future therapeutic strategies (eg, combination studies) that may be developed.\n\nNon--JAK inhibitor therapies for MF\n===================================\n\nAllogeneic hematopoietic stem cell transplant\n---------------------------------------------\n\nAlloHSCT is currently the only curative treatment for patients with M F. However, as a large proportion of patients are not in the transplant age group at the time of diagnosis, alloHSCT has a limited role in the overall disease management of patients with M F. Significant regimen-related toxicities, graft failure, and graft-versus-host disease are major barriers to the success of alloHSCT in MF. However, it remains a valid option for patients in the transplant age group with adequate performance status and without any prohibitive comorbidities; among these typically younger patients, suitable donors are found in approximately 40%--50% of cases.[@b30-btt-7-189] Data from the most recent studies suggest that the expected progression-free survival rate after alloHSCT is in the range of 40%--50% at 3 years.[@b30-btt-7-189]\n\nPerspective: the recommended indications for transplant in our clinic are expected survival \\<5 years, transfusion dependency, and/or an increased risk of leukemic transformation, perhaps using the DIPSS plus or novel molecular markers as discussed earlier. The availability of a fully matched sibling donor would lead us to consider conducting a transplant earlier (for patients with intermediate-2 or intermediate-1 risk with anemia, transfusion, or a rising blast count). While splenectomy is not routinely recommended prior to alloHSCT, it is reasonable to explore the safety and efficacy of novel drugs that can provide rapid spleen shrinkage and improvement of constitutional symptoms in the immediate pretransplant period.[@b30-btt-7-189] Along these lines, a clinical trial (study MPD-RC114) to explore the safety of ruxolitinib in the pretransplant setting is underway.[@b31-btt-7-189]\n\nSplenomegaly and extramedullary hematopoiesis\n---------------------------------------------\n\n### Cytoreductive agents\n\nCytoreductive agents have been the treatment of choice for most patients with symptomatic splenomegaly. Hydroxycarbamide (hydroxyurea, HC) is the most commonly used cytoreductive agent, which usually produces modest responses at higher doses (1--2 g). However, HC can often exacerbate cytopenias and therefore is often not well tolerated. Reductions in spleen size \\>25% and 50% have been reported in up to 35% and 17%, respectively, of the patients treated with HC.[@b32-btt-7-189] In patients who do not respond to HC, busulfan or melphalan can be used, especially in older patients, since there is evidence that these agents can increase the frequency of leukemic transformation. Spleen responses with low-dose thalidomide (50 mg daily) are infrequent (\\<20%). However, lenalidomide has been shown to result in a 33% response rate in a study that included some patients who had failed prior thalidomide therapy. In cases of massive refractory splenomegaly, monthly intravenous cladribine courses have produced responses up to 50%, with severe but reversible cytopenias being the primary toxicity.[@b33-btt-7-189] Interferonalfa (standard and PEGylated versions) has demonstrated minimal clinical effect in reducing splenomegaly, and therefore its use is not generally recommended.[@b34-btt-7-189] Ruxolitinib has proven superior to BAT in the COMFORT-II study and thus we would use this agent first line to control symptomatic or progressive splenomegaly.\n\n### Splenectomy and radiotherapy\n\nWhile the management of MF-associated splenomegaly with splenectomy is well established, the procedure is associated with morbidity and mortality rates of approximately 31% and 9%, respectively.[@b35-btt-7-189] Hepatic extramedullary hematopoiesis, which sometimes leads to rapid hepatic enlargement, is an unusual but well recognized complication following splenectomy, as is the increased thrombotic risk. As a result, splenectomy should be restricted to selected patients with refractory hemolysis or anemia, symptomatic splenomegaly, significant splenic infarction, severe portal hypertension, and/or severe hypercatabolic symptoms. Furthermore, patients undergoing splenectomy need to be made well aware of the risks and provided with meticulous preoperative assessment and postoperative follow-up care.\n\nRadiotherapy can be an alternative to splenectomy in patients with symptomatic splenomegaly and an adequate platelet count (\\>50 \u00d7 10[@b9-btt-7-189]/L). In a report from the Mayo Clinic, a median radiation dose of 277 cGy administered in a median of 7.5 fractions reduced spleen size in the majority of cases for a median of 6 months. However, 44% of patients experienced cytopenias, of which 13% were fatal.[@b36-btt-7-189] Our preference is to use even lower dosing fractions with caution. Low-dose radiotherapy remains a preferred treatment for nonsplenic extramedullary hematopoiesis, including involvement of the peritoneum and pleura with resultant ascites and pleural effusions.\n\nPerspective: cytoreductive agents and/or surgical intervention have been the main approaches to treat symptomatic splenomegaly. JAK inhibitors, primarily ruxolitinib at present, will radically alter the way this aspect of disease is managed. In the COMFORT-II study, none of the 73 patients in the BAT arm, of whom 60% received HC, achieved a sustained \\>35% reduction in spleen volume.[@b17-btt-7-189] Ruxolitinib is likely to surpass HC as first-line treatment of symptomatic splenomegaly and will also be a valuable option in the management of extramedullary hematopoiesis at other sites.\n\nAnemia management\n-----------------\n\nThe management of anemia can be one of the most challenging aspects of treating patients with MF. Blood transfusion is a standard therapy for symptomatically anemic patients, and the transfusion target should be assessed individually. Since regular transfusions will ultimately result in iron overload, iron chelation is frequently required; therefore, it is useful to also consider other treatment options.\n\n### Erythropoiesis-stimulating agents\n\nResponses to ESAs are more likely in transfusion-independent patients with higher baseline Hb. In an analysis of 20 anemic patients with MF treated with ESAs, responses were seen in 45% of cases but were maintained long-term in only 20%.[@b37-btt-7-189] A pooled analysis that combined these 20 patients with 31 patients from the literature suggested an overall response rate of 55% (31% complete responses), with a median response duration of 12 months.[@b37-btt-7-189]\n\nPerspective: the use of ESAs in combination with ruxolitinib therapy has also been reported.[@b38-btt-7-189] In an analysis of 13 patients enrolled in COMFORT-II, concomitant ESA and ruxolitinib therapy was well tolerated, with an observed safety profile similar to that of ruxolitinib therapy alone. Furthermore, the combination did not appear to affect the efficacy of ruxolitinib regarding reductions in spleen size. Further analysis is required to determine if ESAs provided any substantial alleviation of anemia in these patients. Unless a patient has chronic kidney disease or an endogenous erythropoietin level ,125 IU/L, we generally do not use ESAs to treat patients in our clinic, although a short therapeutic trial may be useful.\n\n### Androgens\n\nAndrogenic hormones have been shown to stimulate erythropoiesis in patients with refractory anemia, leading to increased Hb level, reticulocytosis, and decreased need for blood transfusions.[@b39-btt-7-189] Danazol, a synthetic attenuated androgen, has demonstrated efficacy in treating anemic patients with MF and even reduced spleen size in a proportion of patients.[@b40-btt-7-189] Dosages are dependent on body weight (600 mg daily for patients weighing up to 80 kg and 800 mg daily for those weighing \\>80 kg) and should be continued for a period of 6 months. Patients achieving favorable responses can be maintained on danazol at a reduced dose of 400 mg daily for 6 months and then titrated down to the minimum dose required to maintain a response (generally 200 mg daily).\n\nSide effects of androgen therapy include fluid retention, increased libido, hirsutism, abnormal liver function tests, and hepatic tumors. Therefore, all patients receiving danazol should be monitored using monthly liver function tests during initial therapy and a periodic liver ultrasound to detect any hepatic malignancy. Males should be screened for prostate cancer before and during treatment.\n\nPerspective: in our clinic, we have had limited but successful experiences of combining androgens with ruxolitinib. However, it should be noted that these patients require careful liver function monitoring during androgen therapy.\n\n### Immunomodulatory agents\n\nIMiDs, such as thalidomide, lenalidomide, and pomalidomide, are agents that inhibit neoangiogenesis by downregulating vascular endothelial growth factor, basic fibroblast growth factor, and tumor necrosis factor. IMiDs have shown some efficacy in managing anemia, with some responses in patients with thrombocytopenia and splenomegaly (reviewed in the study by Thapaliya et al[@b41-btt-7-189]). In our clinic, thalidomide is used in combination with prednisolone. However, due to the side-effect profile of thalidomide, it would not be selected for first-line management of anemia.[@b42-btt-7-189] Lenalidomide is the recommended first-line therapy in rare cases of del(5q31)-associated anemia, because significant improvement -- with resolution of anemia and evidence of occasional molecular remission -- has been described.[@b43-btt-7-189] There is great interest in the potential for pomalidomide to treat anemia, and the benefits of combining pomalidomide with prednisolone have recently been reported.[@b44-btt-7-189] A number of Phase II studies are underway, [@b45-btt-7-189]--[@b47-btt-7-189] and results from a Phase III study (Pomalidomide in Persons With Myeloproliferative- Neoplasm-Associated Myelofibrosis and RBC \\[red blood cell\\]-Transfusion-Depen-dence Myelofibrosis and RBC-Transfusion-Dependence; RESUME) are anticipated.\n\nPerspective: in our clinic, IMiDs are rarely used as monotherapy for treating anemia in patients with M F, so other agents, such as androgens, are often selected instead. It will be interesting to see if the combination of IMiDs with a JAK inhibitor, such as ruxolitinib, can ameliorate therapy-induced anemia. However, we are doubtful that the combination will yield a benefit in this challenging scenario and would rather focus efforts on deepening disease response. The results of the RESUME trial are likely to be critical here, although we would note that anemia response is notoriously difficult to assess.\n\nOther experimental strategies\n-----------------------------\n\n### Everolimus\n\nIn addition to JAK/STAT, other related pathways, such as the phosphatidylinositol 3-kinase/mammalian target of rapamycin (PI3K/mTOR) pathway, have been found to be dysregulated in M F.[@b48-btt-7-189] In studies in vitro, the proliferation of *JAK2* V617F--positive cells decreased when treated with the mTOR inhibitor everolimus.[@b49-btt-7-189]--[@b51-btt-7-189] Results from a Phase I/II study of 39 high- or intermediate-risk patients with PMF or post-PV/ET MF treated with everolimus have also been reported.[@b49-btt-7-189] Of 30 evaluable patients, 69% and 80% experienced complete resolution of systemic symptoms and pruritus, respectively. The response rate was 60% when European Myelofibrosis Network criteria were applied (eight major, seven moderate, and three minor responses) or 23% when International Working Group for Myelofibrosis Research and Treatment criteria were used (one partial response, six clinical improvements). These results provide proof of concept that targeting the mTOR pathway may be clinically relevant in patients with M F.\n\n### Panobinostat\n\nDeacetylases (DACs) are enzymes that modify the acetylation of both non-histone and histone proteins. The inhibition of DACs has been shown to influence a number of cellular events involved in cancer initiation and progression.[@b52-btt-7-189],[@b53-btt-7-189] Panobinostat (LBH589) is a novel pan-DAC inhibitor that has demonstrated clinical activity in Phase I/II studies in patients with MF. At 16 months, one patient achieved a near-complete response, with resolution of palpable splenomegaly.[@b54-btt-7-189] This patient also achieved elimination of peripheral blood dacrocytes and leukoerythroblastosis, a 4 g/dL increase in Hb from baseline, and improvement in overall marrow cellularity and megakaryocyte atypia, with an increase in erythroid precursors and a significant reduction in reticulin/collagen fibrosis. In a Phase I/II study of panobinostat, reversible thrombocytopenia was observed to be the dose-limiting toxicity.[@b54-btt-7-189],[@b55-btt-7-189] Overall, low doses of panobinostat administered for \\>6 months improved symptoms and clinical features and reversed pathologic marrow changes in patients with MF.\n\n### Combination approaches\n\nThere is significant interest in improving patient outcomes using the combination of JAK inhibitors and traditional and/or experimental agents.[@b56-btt-7-189] Studies assessing the JAK1/2 inhibitor ruxolitinib combined with IMiDs, androgens, and DAC, PI3K, and smoothened inhibitors, or administered prior to HSCT have already been initiated ([Table 3](#t3-btt-7-189){ref-type=\"table\"}). Preliminary results in a small number of patients have been reported for panobinostat combined with ruxolitinib.[@b57-btt-7-189] The combination has been well tolerated, and reductions in spleen size and improvement of MF-related symptoms have been observed, even at doses below the therapeutic dosage of each agent as monotherapy.\n\nPerspective: it is our practice to offer patients trials exploring combination therapy if transplant is not an option or if ruxolitinib has either failed or may not be suitable. Other patients for consideration in combination studies are those seeking to delay disease progression who do not yet require transplant (eg, a young patient with leukocytosis but no other risk factors). Such patients are lower risk than those in the former categories and need to be very carefully counseled and monitored in experimental trials. Combination studies offer patients the opportunity to move toward ultimately attaining a possible cure or at least deeper/more meaningful responses. However, challenges lie ahead in understanding how to assess the benefits of combination approaches since observation of a survival benefit or leukemia-free survival would require very large trials conducted over a long period. Spleen volume response has almost become the standard endpoint in MF and may be selected as a primary endpoint in these studies, but other measures may be more appropriate.\n\nConclusion\n==========\n\nIt is encouraging to witness the recent developments in the understanding and treatment of MF and observe the benefits that these new options can provide to patients. However, it will be important to assess the long-term safety and efficacy of new treatments, such as JAK inhibitors, and establish the role of combination therapies in MF treatment.\n\nA challenge in the introduction of novel therapies is determining both the current prevalence of disease and the true societal cost -- this is poorly understood for M F, and efforts should focus here. If we could better understand the events surrounding either progression in MF or transformation in post-ET/post-PV MF, we might unlock even more biologically significant targets. As such, it will be important to assess whether early JAK inhibitor intervention in patients with PV and ET has an impact on reducing the transformation rate to post-PV MF and post-ET M F. Furthermore, such efforts may help us to identify surrogate markers of response that could enable more rapid advancement in the selection of novel therapies or combinations: the presence of splenomegaly and degree of spleen response seem to be of utility in identifying patients who may have a survival benefit, but other markers could be more relevant (eg, allele burden, cluster of differentiation molecule \\[CD\\]34+, fibrosis grade, and/or genetic signature).\n\nFinancial support for medical editorial assistance was provided by Novartis Pharmaceuticals Corporation. The authors thank Daniel Hutta, PhD, for medical editorial assistance with this manuscript.\n\n**Disclosure**\n\nDr Clodagh Keohane has received research funding from Novartis. Dr Deepti Radia has received honoraria from Novartis, Shire, Pfizer, and Vifor Pharma. Dr Claire Harrison has received honoraria from Novartis, Sanofi-Aventis, Celgene, and Shire; research funding from Novartis and Shire; and acted as a consultant to YM BioSciences, S\\*BIO, Sanofi-Aventis, and Novartis.\n\n![Kaplan--Meier analysis of overall survival in (**A**) COMFORT-I[@b21-btt-7-189] and (**B**) COMFORT-II[@b20-btt-7-189] with 24 months of follow-up.\\\n**Note:** ^a^*P*-values and CIs are unadjusted for repeat analyses.\\\nRepublished with permission of American Society of Hematology; High Wire Press, from A comprehensive review and analysis of the effect of ruxolitinib therapy on the survival of patients with myelofibrosis. Mascarenhas J, Hoffman R. *Blood*. 121(24):2013; permission conveyed through Copyright Clearance Center, Inc.[@b59-btt-7-189]\\\n**Abbreviations:** CI, confidence interval; HR, hazard ratio; BAT, best available therapy; COMFORT, COntrolled MyeloFibrosis Study With ORal JAK Inhibitor Treatment.](btt-7-189Fig1){#f1-btt-7-189}\n\n###### \n\nPrognostic risk assessment\n\n  Variable                           IPSS   DIPSS   DIPSS plus[a](#tfn2-btt-7-189){ref-type=\"table-fn\"}\n  ---------------------------------- ------ ------- -----------------------------------------------------\n  Age \\>65 years                     \u2713      \u2713       \n  Constitutional symptoms            \u2713      \u2713       \n  Hemoglobin \\<10 g/dL               \u2713      \u2713\u2713      \n  Leukocyte count \\>25 \u00d7 10^9^/L     \u2713      \u2713       \n  Circulating blasts \u22651%             \u2713      \u2713       \n  Platelet count \\<100 \u00d7 10^9^/L                    \u2713\n  RBC transfusion need                              \u2713\n  Unfavorable karyotype:                            \u2713\n  +8, \u22127/7q\u2212, isochromosome (17q),                  \n  inversion (3), deletion \u22125/5q\u2212,                   \n  12p\u2212, 11q23 rearrangement                         \n\n**Notes:** \u2713 = 1 point each.\n\nFor DIPSS plus, the score is derived from the DIPSS score and additional points added as per the table.\n\n**Abbreviations:** IPSS, international Prognostic Scoring System; DIPSS, dynamic IPSS; RBC, red blood cell.\n\n###### \n\nHematologic laboratory abnormalities\n\n  Laboratory parameter   COMFORT-I[@b17-btt-7-189]   COMFORT-II[@b18-btt-7-189]                            \n  ---------------------- --------------------------- ---------------------------- ---- ---- ---- ---- ---- ----\n  Anemia                 83                          45                           44   16   82   40   49   21\n  Thrombocytopenia       71                          14                           21   2    69   9    29   7\n  Neutropenia            19                          7                            4    3    12   6    8    1\n\n**Note:** Reproduced with permission from Novartis Pharmaceuticals Canada Inc. Jakavi (ruxolitinib) \\[product monograph\\]. Dorval, QC: 2012.[@b58-btt-7-189]\n\n**Abbreviations:** BAT, best available therapy; COMFORT, COntrolled MyeloFibrosis Study With ORal JAK Inhibitor Treatment.\n\n###### \n\nRuxolitinib combination therapy -- clinical studies\n\n  Combination agent           Study title                                                                                                                                                     Trial identifier\n  --------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------\n  Lenalidomide                Ruxolitinib and Lenalidomide for Patients with Myelofibrosis                                                                                                    NCT01375140\n  Pomalidomide                Ruxolitinib and Pomalidomide Combination Therapy in Patients with Primary and Secondary MF (POMINC)                                                             NCT01644110\n  Danazol                     Ruxolitinib Phosphate and Danazol in Treating Anemia in Patients with Myelofibrosis                                                                             NCT01732445\n  Panobinostat                Panobinostat and Ruxolitinib in Myelofibrosis (PRIME Trial)                                                                                                     NCT01693601\n  Buparlisib (BKM120)         A Study to Find the Maximum Tolerated Dose of the Experimental Combination of the Drugs INC424 and BKM120 in Patients with Primary or Secondary Myelofibrosis   NCT01730248\n  LDE225                      A Phase Ib/II Dose-finding Study to Assess the Safety and Efficacy of LDE225 + INC424 in Patients with MF                                                       NCT01787552\n  Stem cell transplantation   Ruxolitinib Prior to Transplant in Patients with Myelofibrosis                                                                                                  NCT01790295\n\n**Abbreviation:** MF, myelofibrosis.\n"}
+{"text": "Introduction\n============\n\nThe combined actions of neural input, muscles, and the joint(s) about which those muscles act serve to produce sufficient endpoint force for physical function, allowing the performance of activities of daily living, as well as the spectrum of athletic endeavors. Due to the complexity of the neuromuscular and musculoskeletal systems, many factors can influence strength, including, but not limited to, muscle moment arm, muscle size, activation, muscle architecture, and normalized muscle force (or specific tension) ([@ref-74]). Muscle size is of particular interest, as (1) it is highly plastic ([@ref-33]) and (2) a clear positive relationship exists between baseline muscle cross-sectional area (CSA) and strength, with greater CSAs correlating with greater strength capacities ([@ref-55]; [@ref-56]; [@ref-68]). However, this relationship is not necessarily linear, as several additional factors interactively influence strength capacity ([@ref-74]); studying the role of and relationship between muscle size and strength is therefore less straightforward under longitudinal contexts.\n\nWhile the cross-sectional correlation between muscle mass and strength remains well-established, some researchers have recently challenged the belief that resistance training (RT)-induced hypertrophy significantly impacts the ability to produce force, claiming improvements in these outcomes are separate and unrelated adaptations ([@ref-17]). Indeed, data remain somewhat equivocal on the relationship between changes in size and changes in strength resulting from regimented RT: A considerable range of correlation coefficients have been observed, from \u223c0 to \u223c0.6 ([@ref-2]; [@ref-4]; [@ref-5]; [@ref-6]; [@ref-22]; [@ref-30]; [@ref-31]; [@ref-52]; [@ref-53]; [@ref-59]; [@ref-64]; [@ref-79]). The discrepancies in findings between studies may be related, in part, to the statistical measures employed to analyze relationships between muscle hypertrophy and strength gain. For instance, analyses in a majority of studies are based on between-subject data using only two time points, but within-subject analyses are more appropriate for the question at hand. Inferentially, drawing individual-level conclusions from group-level data is a statistical fallacy, known as the ecological fallacy ([@ref-67]). Pragmatically, this problem can be better understood by differentiating between the question that each analysis addresses. Between-subject analyses answer the question, \"Do those who grow more also get stronger than those who grow less?\" Conversely, within-subject analyses answer the question, \"Is the growth of one's muscle related to their increases in strength?\" Due to individual differences, the former (between-subject) may not necessarily map to the latter (within-subject). For example, if subject A has a 30% larger muscle moment arm than subject B, then one may expect subject A to have a 30% greater slope between increases in muscular strength (force) and externally-measured strength (moment), all else being equal. To address the ecological fallacy and answer the within-subject question, more sophisticated statistical approaches are needed ([@ref-40]; [@ref-45]; [@ref-67]).\n\nA hierarchical approach can assist in avoiding the pitfall of the ecological fallacy ([@ref-40]; [@ref-45]). Traditionally, each participant's change in strength and change in size, from pre- to post-intervention, are calculated and regressed among one another ([@ref-2]; [@ref-4]; [@ref-5]; [@ref-6]; [@ref-22]; [@ref-30]; [@ref-31]; [@ref-53]; [@ref-59]; [@ref-64]; [@ref-79]). However, a hierarchical modeling approach allows for one to look at time points nested within participants, such that each participant's points are kept \"separate\" from other participants ([@ref-39]; [@ref-40]; [@ref-65]). Within the hierarchical model, each participant can receive varying intercepts and/or varying slopes, which allows for inter-individual differences to be appropriately accounted for ([@ref-39]; [@ref-40]; [@ref-65]). To carry out hierarchical modeling with varying slopes and intercepts, multiple (\u22653) time points are required (i.e., to quantify model variance), so most training datasets cannot be used to answer this question, as a majority only collect data at two time points (pre- and post-intervention). To date, only one study has employed a within-subject analysis: [@ref-52] used analysis of covariance (ANCOVA) ([@ref-12]) and found appreciably greater coefficients of determination in within- relative to between-subject models for the same muscle and strength test (e.g., *R*^2^\u00a0=\u00a00.004 vs. 0.35). However, in contrast to hierarchical linear models, ANCOVA has an affine assumption; participants receive different intercepts, but all are constrained to the same slope ([@ref-12]). Therefore, further work is needed to understand how model choice affects the strength of the relationship between hypertrophy and changes in strength.\n\nThe purpose of this study was to investigate the relationship between changes in muscle size and strength in the elbow flexors using a variety of statistical and measurement approaches, while also employing both between- and within-subject analyses over multiple time-points during periods of both training and detraining. It was hypothesized that different statistical models would produce different outcomes, with between-subject correlations showing the weakest relationships and hierarchical linear modeling showing the strongest.\n\nMethods\n=======\n\nParticipants\n------------\n\nThe study reanalyzed data from a previously published study, the methods of which have been described ([@ref-71]). In brief, young, recreationally active individuals (mean\u00a0\u00b1\u00a0SD, age = 24\u00a0\u00b1\u00a03\u00a0years, BMI = 22\u00a0\u00b1\u00a02, *n*\u00a0=\u00a019) were recruited for participation in the study. Participants reported exercising at least three times per week via various sporting activities but did not perform resistance training for the elbow flexors. Informed consent was obtained for all participants. The original study was approved by the University of Queensland Medical Research Ethics Committee (no. 2014001416).\n\nMuscle size\n-----------\n\nMeasures of muscle thickness were obtained via B-mode ultrasound imaging (Mindray DP-50) using a 7.5 MHz linear transducer probe. Images were taken at baseline and after each week of training throughout the 16-week study period. Scanning was carried out by a trained sonographer on both the dominant and non-dominant elbow flexors at 30, 50, and 70% of total length of the biceps brachii whilst participants were seated with the antebrachium in a neutral position. After Weeks 4, 8, and 16, CSA scans were acquired for both upper limbs via panoramic B-mode ultrasound (S3000 Siemens/Acuson system) using a 4--9 MHz linear transducer operating at 9 MHz. Imaging for CSA was obtained via lateral acquisition at 50% width of the biceps brachii. Values for both muscle thickness and CSA were determined using ImageJ (version 1.48; National Institutes of Health, Bethesda, MD, USA). Muscle thickness was not assessed for Week 4 due to a conflict in scheduling with CSA ultrasounds. All ultrasound measures were completed by a paid qualified professional, and not by the researchers of the paper. If the probe lost contact at any point during the measurement, the measurement was retaken. Test-retest intraclass correlation coefficients (ICC; model 2,1) of 0.99 and 0.97 for CSA and muscle thickness, respectively, have been previously reported ([@ref-46]). Because an ICC(2,1) model was used, these results are generalizable to the experienced rater in this study ([@ref-50]).\n\nResistance training protocol\n----------------------------\n\nResistance training for the non-dominant brachium was carried out five days per week for the initial eight weeks of the study, followed by a subsequent eight-week detraining period. Training consisted of unilateral dumbbell elbow flexion performed with a supinated forearm. During each session, participants performed nine sets of 12 repetitions with a 90-second rest interval afforded between sets. Loads were based on maximal voluntary isometric contraction (MVIC) values that were obtained each week using a Sundoo SN Analogue Force Gauge (model number SN-500) at 90\u00b0 elbow flexion. Subjects began each workout using 70% of that week's MVIC recording. If the full number of target repetitions (i.e., 12) was not achieved on a given set, the load was lowered to the next level of load until completion---e.g., if a participant achieved 8 repetitions at 70%, the load was decreased to 50% so that all 12 repetitions could be performed. Loads were progressively lowered on successive sets to 50% and 30% of MVIC as needed so that subjects could complete the target repetition range with proper form. The dominant brachium of each subject served as the control for the study throughout the training and detraining periods. Subjects were instructed to refrain from exercise involving the elbow flexors, other than activities of daily living, throughout the 16-week study period.\n\nStatistical analysis\n--------------------\n\nSeveral statistical analyses were carried out to investigate how methods of both measurement and analysis may affect the conclusions drawn from a study investigating the relationship between strength and hypertrophy. All analyses were carried out in R (version 3.4.3) ([@ref-62]). First, standard bivariate linear regression analyses of pre- and post-measures were utilized to investigate the relationship between muscle size (thickness or CSA) and strength, using a between-subject model. This was done for two different conditions: training and detraining. For each condition, a data point (\u0394~size~, \u0394~strength~) was calculated for each participant, where, in the general case, \u0394\u00a0=\u00a0post\u00a0\u2212\u00a0pre, where pre and post are the values before and after a given condition (training or detraining), respectively, as has been done in a number of previous investigations ([@ref-2]; [@ref-30]; [@ref-52]). Second, an ANCOVA was utilized to replicate the method of analysis used by [@ref-52]. In this analysis, strength was treated as a dependent variable, participants were treated as a categorical factor (dummy-coded), and size was treated as a covariate. Variance accounted for (VAF) was calculated using the formula $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}$\\text{VAF}= \\frac{S{S}_{\\text{size}}}{S{S}_{\\text{size}}+S{S}_{\\text{residual}}} $\\end{document}$, where *SS* is type III sum of squares ([@ref-12]). This is equivalent to a partial *\u03b7*^2^ for the size covariate. Lastly, because the ANCOVA method has a number of assumptions and does not allow for varying slopes, a more robust hierarchical linear model was used for the final analysis ([@ref-61]). In this analysis, the outcome measure (*y*~*ij*~) was the net joint moment during MVIC, and muscle size was used as a level-one predictor variable (*x*~*ij*~), which were group-mean centered for analyses. Subject was treated as a level-two variable. Finally, varied slopes and intercepts were permitted, creating the final model: $$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}\\begin{eqnarray*}\\text{}& Level~1 \\end{eqnarray*}\\end{document}$$ $$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}\\begin{eqnarray*}\\text{}& {y}_{ij}={\\beta }_{0j}+{\\beta }_{1j}{x}_{ij}+{\\epsilon }_{ij} \\end{eqnarray*}\\end{document}$$ $$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}\\begin{eqnarray*}\\text{}& Level~2 \\end{eqnarray*}\\end{document}$$ $$\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}\\begin{eqnarray*}\\text{}& {\\beta }_{0j}={\\gamma }_{00}+{r}_{0j}\\nonumber\\\\\\displaystyle & {\\beta }_{1j}={\\gamma }_{10}+{r}_{1j} \\end{eqnarray*}\\end{document}$$\n\nThe model was fit using restricted maximum likelihood in the *lme4* package ([@ref-9]). Sample variance of the residuals (*s* ^2^) were used to calculate VAF (or *R*^2^) using the following formula: $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}$\\text{VAF}=1- \\frac{{s}^{2}}{{s}_{\\text{uncond}}^{2}} $\\end{document}$, where $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}${s}_{\\text{uncond}}^{2}$\\end{document}$ is the sample variance of the residuals in the unconditional model, which contained only varied intercepts and no fixed effects (i.e., the same model, but with *\u03b2*~1*j*~\u00a0=\u00a00). This approach is mathematically equivalent to the VAF found for the ANCOVA using type III sums of squares (see [Appendix A](#supp-3){ref-type=\"supplementary-material\"}). Intraclass correlation coefficients (ICC) were calculated on the unconditional models to estimate the proportion of original variance explained by subject. To estimate 95% confidence intervals (CI) of the VAFs, each model was bootstrapped 2,000 times with replacement. The 0.025 and 0.975 quantiles of the VAF estimates were calculated as the lower and upper bounds of each estimate's 95% CI.\n\nTo understand how the different measures of hypertrophy relate to one another, within- and between-subject correlation matrices were constructed using the different thickness measures and CSA. The between-subject analysis included all thickness and CSA measures, across all subjects, for any time point at which both CSA and thickness were measured. The within-subject correlation matrix was constructed in a similar manner: (1) a correlation coefficient was calculated for each participant (*r*~*i*~); (2) using a Fisher *z*-transformation, *r*~*i*~ was transformed to a *z*-score (*z*~*i*~); (3) a weighted average was obtained using the number of points (*n*~*i*~) from each participant ($\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}$\\bar {z}= \\frac{\\sum {z}_{i}({n}_{i}-3)}{\\sum ({n}_{i}-3)} $\\end{document}$, for *i* participants); and (4) $\\documentclass[12pt]{minimal}\n\\usepackage{amsmath}\n\\usepackage{wasysym} \n\\usepackage{amsfonts} \n\\usepackage{amssymb} \n\\usepackage{amsbsy}\n\\usepackage{upgreek}\n\\usepackage{mathrsfs}\n\\setlength{\\oddsidemargin}{-69pt}\n\\begin{document}\n}{}$\\bar {z}$\\end{document}$ was transformed back to Pearson's *r* ([@ref-16]; [@ref-20]; [@ref-21]; [@ref-43]). Because CSA measures were only taken with thickness at two time points, within-subject correlation coefficients could not be estimated between CSA and muscle thickness.\n\nFurther exploratory analyses were performed to investigate if those with stronger strength-hypertrophy relationships also got stronger. To do this, Pearson correlation coefficients were calculated for each individual across the entire study (i.e., including both training and detraining periods). The resulting correlation coefficients were then correlated with \u0394~strength~ from the training period using Spearman's rank-order correlations (*\u03c1*). Spearman's *\u03c1* was used due to the heteroscedastic nature of the residuals. Qualitative interpretations of correlation coefficients and VAFs can be found in [Table 1](#table-1){ref-type=\"table\"}, which are in accordance with [@ref-44]. R code for all procedures can be found in the [Supplemental Files](#supplemental-information){ref-type=\"supplementary-material\"}.\n\n10.7717/peerj.5071/table-1\n\n###### Correlation coefficient and variance accounted for interpretations.\n\n![](peerj-06-5071-g001)\n\n  Interpretation      Correlation coefficient (*r* or *\u03c1*)   Variance accounted for (%)\n  ------------------- -------------------------------------- ----------------------------\n  Trivial             \\[0, 0.1)                              \\[0, 1)\n  Small               \\[0.1, 0.3)                            \\[1, 9)\n  Moderate            \\[0.3, 0.5)                            \\[9, 25)\n  Large/strong        \\[0.5, 0.7)                            \\[25, 49)\n  Very large/strong   \\[0.7, 0.9)                            \\[49, 81)\n  Nearly perfect      \\[0.9, 1)                              \\[81, 100)\n  Perfect             1                                      100\n\n**Notes.**\n\nAdapted from [@ref-44]. Note that all intervals are of the form *x*~low~\u00a0\u2264\u00a0*x*~*o*~\u00a0\\<\u00a0*x*~high~.\n\nResults\n=======\n\nDifferences in VAFs ranged from zero to an order of magnitude ([Table 2](#table-2){ref-type=\"table\"}). Similar differences were also observed between different statistical models for a given measure ([Table 2](#table-2){ref-type=\"table\"}). Intraclass correlation coefficients from the hierarchical linear models suggest that most of the original variance could be accounted for by including a level for subject (ICC = 0.89--0.91). Heterogeneity in correlation coefficients was observed when comparing different measures of muscle thickness, which ranged from *r*\u00a0=\u00a00.503 to *r*\u00a0=\u00a00.945 for between-subject correlations and from *r*\u00a0=\u00a00.326 to *r*\u00a0=\u00a00.875 for weighted within-subject correlations ([Table 3](#table-3){ref-type=\"table\"}). Finally, Pearson's *r* of each individual's strength-hypertrophy relationship was a moderate to strong predictor of strength for all measurements (US~30%~\u00a0*\u03c1*\u00a0=\u00a00.644; US~50%~\u00a0*\u03c1*\u00a0=\u00a00.356; US~70%~\u00a0*\u03c1*\u00a0=\u00a00.413; US~avg~\u00a0*\u03c1*\u00a0=\u00a00.480; CSA *\u03c1*\u00a0=\u00a00.449).\n\n10.7717/peerj.5071/table-2\n\n###### Percent (%) variance accounted for (95% CI) using different types of models.\n\n![](peerj-06-5071-g002)\n\n  Measure                Between-subjects   Within-subjects                      \n  ---------------------- ------------------ ----------------- ------------------ ------------------\n  Thickness (30%)        3.6 (0--61.9)      1.0 (0--45.1)     0.2 (0--6.1)       7.4 (0.8--16.0)\n  Thickness (50%)        0.8 (0--21.6)      0.0 (0--23.7)     0.3 (0--9.7)       24.1 (6.7--42.0)\n  Thickness (70%)        1.4 (0--39.1)      1.6 (0--38.0)     2.2 (0--10.9)      7.5 (2.1--23.7)\n  Thickness (Average)    0.4 (0--21.1)      0.0 (0--26.4)     1.2 (0--12.9)      18.1 (6.6--30.4)\n  Cross-sectional area   0.4 (0--32.2)      1.2 (0--35.4)     11.7 (1.1--34.2)   12.1 (2.0--69.5)\n\n**Notes.**\n\n30%, 50%, and 70% represent the position of the ultrasound probe on the brachium. Average represents the average of all three of the measured thicknesses at a given time point. Cross-sectional area was measured at 50%.\n\nTITLEANCOVAanalysis of covarianceHLMhierarchical linear model\n\n10.7717/peerj.5071/table-3\n\n###### Correlation matrix of measures of muscle size.\n\n![](peerj-06-5071-g003)\n\n                         Thickness (30%)                          Thickness (50%)                          Thickness (70%)                          Thickness (Average)                      Cross-sectional area\n  ---------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- ----------------------------------------\n  Thickness (30%)                                                 0.503[^a^](#table-3fn2){ref-type=\"fn\"}   0.618[^a^](#table-3fn2){ref-type=\"fn\"}   0.778[^a^](#table-3fn2){ref-type=\"fn\"}   0.557[^a^](#table-3fn2){ref-type=\"fn\"}\n  Thickness (50%)        0.344[^b^](#table-3fn3){ref-type=\"fn\"}                                            0.869[^a^](#table-3fn2){ref-type=\"fn\"}   0.916[^a^](#table-3fn2){ref-type=\"fn\"}   0.742[^a^](#table-3fn2){ref-type=\"fn\"}\n  Thickness (70%)        0.326[^b^](#table-3fn3){ref-type=\"fn\"}   0.687[^b^](#table-3fn3){ref-type=\"fn\"}                                            0.945[^a^](#table-3fn2){ref-type=\"fn\"}   0.730[^a^](#table-3fn2){ref-type=\"fn\"}\n  Thickness (Average)    0.659[^b^](#table-3fn3){ref-type=\"fn\"}   0.875[^b^](#table-3fn3){ref-type=\"fn\"}   0.871[^b^](#table-3fn3){ref-type=\"fn\"}                                            0.773[^a^](#table-3fn2){ref-type=\"fn\"}\n  Cross-sectional area                                                                                                                                                                       \n\n**Notes.**\n\n30%, 50%, and 70% represent the position of the ultrasound probe on the brachium. Average represents the average of all three of the measured thicknesses at a given time point. Cross-sectional area was measured at 50%.\n\nBetween-subject correlation.\n\nWeighted within-subject correlation.\n\nDiscussion\n==========\n\nTo the authors' knowledge, this is the first study to investigate the relationship between hypertrophy and changes in muscle strength using hierarchical linear modeling, which allows for robust within-individual analysis, in addition to the use of multiple types of measures of muscle size. Our results demonstrate that not only does measurement approach substantially affect outcomes, but so does the type of statistical model employed. These findings have important methodological implications for improving our understanding of the associative relationship between hypertrophy and changes in strength.\n\nPrevious literature has approached the question of how changes in muscle size relate to changes in strength from a between-subject perspective. However, it can be argued that a repeated-measures design allows for a more direct evaluation of the strength-hypertrophy relationship. Individual differences in muscle moment arms (MA), normalized muscle force (NMF), pennation angles (*\u03b8*~*p*~), voluntary activation (*\u03b1*), *et cetera* will greatly confound the relative relationship between changes in strength and muscle size (in this case, physiological CSA(PCSA)). All of the aforementioned components are multipliers in the formula used to calculate a muscle's contribution to a joint moment (*M*\u00a0=\u00a0*\u03b1*\u22c5PCSA\u22c5NMF\u22c5cos*\u03b8*~*p*~\u22c5MA) ([@ref-74]). To date, only one previous investigation has utilized a quantitative within-subject approach to investigate the relationship between hypertrophy and changes in strength ([@ref-52]); although, qualitative within-subject changes are depicted in a classic study by [@ref-25]. Specifically, [@ref-52] employed an ANCOVA with subject as a factor and muscle size as a covariate; from the resulting sum of squares, VAF could be calculated ([@ref-12]). ANCOVA is limited, however, in that it, in its basic form, assumes parallelism between all relationships, has several assumptions that may confound results (e.g., sphericity, compound symmetry, and homoscedasticity), and is not robust to missing data points ([@ref-12]; [@ref-13]; [@ref-61]). The parallel or affine assumption is of particular interest because there are several heterogeneities that confound this assumption (i.e., *\u03b1*, MA, NMF, and *\u03b8*~*p*~). Repeated-measures hierarchical models are a robust way to investigate longitudinal relationships within a group or person ([@ref-39]; [@ref-65]). By comparing these statistical models, a clear difference is apparent ([Table 2](#table-2){ref-type=\"table\"}). For all measurements, the hierarchical linear model resulted in greater VAFs than the ANCOVA ([Table 2](#table-2){ref-type=\"table\"}). These differences may be due to the hierarchical linear model allowing for varying slopes or, alternatively, some of the inherent assumptions and limitations of ANCOVAs ([@ref-61]). Interestingly, the VAFs found in this present study are much lower than those found by [@ref-52]. It is unclear from where these differences arise; that is, if they are due to measurement technique, differences in mechanisms of strength gain, differences in upper vs. lower extremities, or some other factor. However, our data provide a methodological proof of principle by delineating how different statistical models may drastically affect the conclusions formed from a given dataset, even when performed on the same set of regressors. Due to the robustness of hierarchical linear models, it is recommended that such analyses are used over ANCOVAs for future investigations with similar methods.\n\nHow muscle size is assessed will likely affect the strength of the relationship between changes in muscle size and strength. The measurement techniques utilized by previous and present investigations ([@ref-2]; [@ref-4]; [@ref-5]; [@ref-6]; [@ref-17]; [@ref-22]; [@ref-30]; [@ref-52]; [@ref-59]) have been limited in that they do not account for changes in architectural characteristics ([@ref-51]). There are several ways to measure muscle size, including limb circumference ([@ref-25]), estimates of total and segmental muscle mass (dual-energy X-ray absorptiometry and bioelectrical impedance analysis) ([@ref-49]), muscle thickness ([@ref-71]), anatomical CSA ([@ref-30]; [@ref-72]), muscle volume ([@ref-6]; [@ref-30]; [@ref-31]), and PCSA ([@ref-31]). There are strong physiological and mechanical rationales with basic science evidence to suggest that not all of these measures are equal, even when accounting for measurement error ([@ref-51]; [@ref-60]). For example, although muscle volume appears to be a strong predictor of strength in some contexts (even greater than anatomical CSA) ([@ref-3]; [@ref-36]), it does not perform as well in others ([@ref-10]), perhaps at least partly due to inter- and intra-muscular variation in architecture ([@ref-15]; [@ref-51]; [@ref-78]) and adaptation ([@ref-26]; [@ref-27]; [@ref-35]; [@ref-57]; [@ref-76]; [@ref-77]). Muscle volume is not only sensitive to changes in sarcomeres in parallel (PCSA), but also sarcomeres in series (fiber length). Sarcomeres in parallel will contribute to the magnitude of force production, while sarcomeres in series will affect the shapes of the force-length and force-velocity curves. Functionally speaking, not all muscle volume is equal ([@ref-51]). Importantly, in series hypertrophy appears to be limited to the initial weeks of commencing resistance training, further reinforcing potential issues when extrapolating correlative findings from novice to trained individuals ([@ref-14]). Similarly, thickness and anatomical CSA, as measured in this study, are also limited, as they only represent one part of the muscle and do not account for the intricacies of muscle architecture. This is further evidenced by [@ref-35], who found that, cross-sectionally, muscle thickness, anatomical CSA, and muscle volume are related, but the relative changes between muscle thickness and muscle volume did not strongly correlate following a training period. This is important when considering the formula for PCSA, in that the volume of the entire muscle must be taken into account ([@ref-51]); not just thickness or anatomical CSA. Moreover, the variability in correlation coefficients between these measures may be a cause for concern ([Table 3](#table-3){ref-type=\"table\"}), in that it suggests not all measures of muscle size are necessarily capturing the same effects, which is elucidated further by the statistical models ([Table 2](#table-2){ref-type=\"table\"}).\n\nSince PCSA has been shown to be a strong predictor of force production both *in vivo* ([@ref-37]) and *in vitro* ([@ref-60]), it is considered the gold standard for relating muscle form (architecture) to function (force production) ([@ref-51]). PCSA is, in essence, the \"effective\" CSA, as it is the average CSA perpendicular to the fibers' line of action. Thus, PCSA controls for pennation and is representative of the number of sarcomeres in parallel, making it highly indicative of a muscle's potential to generate force through the tendon ([@ref-51]). It is imperative to consider these differences in measurement techniques in the context of this study and similar investigations ([@ref-2]; [@ref-30]; [@ref-31]; [@ref-52]). Although this study ([Table 2](#table-2){ref-type=\"table\"}) and others ([@ref-52]) have observed what is analogous to a strong correlation (*r*\u00a0\u2265\u00a00.5) ([@ref-44]) with repeated-measures designs, substandard measurements of muscle size were used in the present study. Therefore, it is likely that PCSA measurements would produce different results ([@ref-1]). While PCSA is expensive to obtain and typically relies on MRI, newer technologies, such as 3D ultrasound, show promise as valid, affordable alternatives to MRI for estimating muscle volume and PCSA ([@ref-7]; [@ref-8]; [@ref-41]). Moving forward, it seems prudent that investigators utilize PCSA rather than other measures of muscle size, as the theory that hypertrophy leads to strength gains is predicated on this measure rather than other measures of muscle size.\n\nThe question of how changes in strength and changes in muscle size are related is one with broad clinical implications, ranging from the treatment and prevention of sarcopenia and dynapenia to exercise prescription for strength athletes. Clinically, if changes in muscle size are not important for strength, then exercise programs need not focus on variables that are more important for hypertrophy than strength, such as volume ([@ref-63]; [@ref-69]). Changes in strength do indeed arise from non-hypertrophic factors ([@ref-34]), including a myriad of neural adaptations ([@ref-28]), in addition to changes in muscle moment arms ([@ref-70]; [@ref-74]) and normalized muscle force production ([@ref-31]), in which lateral force transmission has been suggested to play a role ([@ref-48]). This implies that changes in strength are interactive rather than linear. As such, how this relationship is investigated and modeled should reflect such complexities. First, with more reductionist strength testing (i.e., single-joint isometric testing), it can be argued that the \"skill\" component of strength is less relevant (as opposed to one-repetition maximum tests ([@ref-18])), since little coordination is necessary and even untrained individuals see little-to-no changes in voluntary activation and co-contraction ([@ref-11]; [@ref-30]; [@ref-31]; [@ref-58]). Moreover, neural measures, such as voluntary activation, can be more accurately assessed during isometric efforts than during dynamic efforts ([@ref-32]; [@ref-75]) and thus can more easily be incorporated into a final model. Second, measures of muscle size should reflect those in the model (i.e., using PCSA). While this is expensive and time consuming, it will provide more appropriate biomechanical insight ([@ref-51]). Third, moment arm measures should be subject-specific and occur over the duration of an experiment, as moment arms may change with training ([@ref-70]; [@ref-74]). Finally, longer duration studies may be more appropriate for several reasons: (1) individual response trajectories will vary, as evidenced by the high ICCs in this present investigation and the heterogeneous rank orders between time points in previous work ([@ref-19]); (2) edema can greatly confound gross imaging measures of muscle size, depending on when the measurements are performed ([@ref-23]); (3) the magnitude of the difference between measurement points will be greater, which in turn will decrease the relative role of measurement error in parameter and VAF estimates ([@ref-38]); and (4) to understand the extent to which contributions may or may not change over time. While this present study did not incorporate these recommendations, since it was based on previously collected data ([@ref-71]), future studies should do so to properly isolate the associative contribution of muscle size (PCSA) to strength increases.\n\nThus far, our discussion has primarily focused on the associative, rather than causal, relationship between hypertrophy and strength gain. A conducive discussion of the causal nature of this relationship requires an operational definition of causality. In formal logic, causality is often broken down into two conditions: (1) necessary conditions, which state that *B* will not occur without *A* (\"if not *A,* then not *B*\"); and (2) sufficient conditions, which state that *A* will result in *B* (\"if *A*, then *B*\") ([@ref-29]; [@ref-42]). However, a less formal concept of causality is also possible without these conditions having been met, in the form of contributory causality. A contributory cause is neither necessary nor sufficient ([@ref-42]; [@ref-66]). Those who experience an effect need not experience its putative cause, and those who experience the putative cause need not experience its effect ([@ref-66]). For instance, although smoking causes lung cancer, not all of those who smoke develop lung cancer (i.e., it is not sufficient), and not all of those who develop lung cancer are smokers (i.e., it is not necessary); therefore, smoking may be viewed as a contributory cause of lung cancer ([@ref-66]). The arguments put forth by [@ref-17], [@ref-24] and [@ref-54] do indeed rule out hypertrophy as being a necessary or sufficient cause for strength gain, but we suggest that the contributory nature of hypertrophy to strength should not be dismissed on this basis. In other words, changes in strength can occur without changes in muscle size and *vice versa*, but this does not preclude muscle size from contributing to strength. Experimentally, it is important to consider the emergent, nonlinear, and interactive properties of strength; there are many moving parts that should be accounted for when attempting to understand such a complex system, which may concurrently change in different directions (e.g., increase in size but decrease agonist activation). Indeed, a systems rather than reductionist approach may be most appropriate for understanding strength emergence. In studying this system, it is necessary to measure all factors (confounders) that may contribute to strength to *truly* understand the role of hypertrophy, especially because different protocols may elicit differential adaptations ([@ref-47]). Thus, longitudinal, within-subject studies that incorporate all of the measures included in the formula to determine strength (PCSA, MA, activation and co-contraction, synergist characteristics, and NMF) are likely needed to better understand the emergent properties of strength. Finally, because the problem is so complex, the contributory role of hypertrophy in strength gain may not be able to be fully established from one study or line of evidence. Instead, a body of literature consisting of many forms of evidence---ranging from animal and agent-based models to observational and experimental human studies---may be required to elucidate the contributory role of hypertrophy in strength gain.\n\nThis study and its discussion have focused primarily on single muscle group hypertrophy and single-joint isometric strength gain. The larger question of multi-joint and dynamic strength gain is perhaps more relevant, but unfortunately much more complex ([@ref-73]). Starting with relatively simpler systems and research questions may bear more fruit, while also providing a conceptual basis that can be used when studying more complex systems and research questions.\n\nThis is the first study to utilize repeated-measures hierarchical linear modeling to investigate the relationship between muscle size and strength. We herein demonstrate that repeated-measures hierarchical linear models produce different results than other within-subject models (ANCOVA), in addition to between-subject models, which is in line with previous work by [@ref-52]. Moreover, it was found that different measures of muscle size can produce vastly different results. As such, we have advocated for more rigorous and reductionist experimental designs to better understand the mechanistic origins of single-joint strength following exercise programs, by suggesting that researchers measure PCSA and single-joint isometric strength, in addition to potential confounding variables.[^1^](#fn-1){ref-type=\"fn\"} [^1]These findings are important for the interpretation of previous studies, in addition to the design of future studies, on this same topic.\n\nConclusions\n===========\n\nThe strength of the associational relationship between muscle hypertrophy and strength gain is highly dependent upon the statistical model employed. We have demonstrated that hierarchical linear modeling, which allows for varying slopes and intercepts, provides greater estimates of the strength of the relationship between muscle hypertrophy and strength gain. Moreover, different assessments of muscle size do not perfectly correlate, and therefore, different methods of assessment may lead to different conclusions. These findings should be taken into consideration when planning and interpreting studies on the relationship between muscle hypertrophy and strength gain.\n\nSupplemental Information\n========================\n\n10.7717/peerj.5071/supp-1\n\n###### Raw data\n\n###### \n\nClick here for additional data file.\n\n10.7717/peerj.5071/supp-2\n\n###### Analysis code\n\n###### \n\nClick here for additional data file.\n\n10.7717/peerj.5071/supp-3\n\n###### Appendix A\n\n###### \n\nClick here for additional data file.\n\nAdditional Information and Declarations\n=======================================\n\nThe authors declare there are no competing interests.\n\n[Andrew D. Vigotsky](#author-1){ref-type=\"contrib\"} conceived and designed the experiments, performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft.\n\n[Brad J. Schoenfeld](#author-2){ref-type=\"contrib\"} conceived and designed the experiments, authored or reviewed drafts of the paper, approved the final draft.\n\n[Christian Than](#author-3){ref-type=\"contrib\"} and [J. Mark Brown](#author-4){ref-type=\"contrib\"} contributed reagents/materials/analysis tools, authored or reviewed drafts of the paper, approved the final draft.\n\nThe following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers):\n\nThe original study was approved by the University of Queensland Medical Research Ethics Committee (no. 2014001416).\n\nThe following information was supplied regarding data availability:\n\nThe raw data and R analysis code are provided as [Supplemental Files](#supplemental-information){ref-type=\"supplementary-material\"}.\n\n[^1]: Note that these recommendations only apply to studies that are investigating the strength-hypertrophy relationship with a reductionist approach. We are in no way suggesting that PCSA and single-joint isometric measures be used for all resistance training studies.\n"}
+{"text": "Introduction\n============\n\nE2F family of transcriptional factors has many regulatory functions implicating cellular proliferation, differentiation, DNA repair, cell cycle and cell apoptosis [@B1], [@B2]. The study of the E2F family began in 1980s, of which most extensive researches involve E2F1 [@B3]. As yet, E2F family has been found to have eight members in mammals, including from E2F1 to E2F8 [@B4]. The alternative splicing of some E2F family members and generation of different isoforms result in that the regulation of E2F functions is more complicated than other members [@B5]. E2F3 and E2F7 have two highly relevant isoforms, respectively [@B6]. Indeed, studies have shown that E2F evolution in plants and animals is very conservative but not in yeast [@B1]. According to their structure and function, E2F family can be divided into two groups (canonical E2F1-6 and atypical E2F7-8) and four subgroups (E2F1-3a, E2F3b-5, E2F6 and E2F7-8). Nowadays we get the difference between canonical and atypical E2F molecular structures, but it is still undefined for its evolutionary mechanism [@B3].\n\nThe canonical E2Fs possess a leucine-zipper domain in the N terminus, which is necessary for dimerization with a partner protein DP (dimerization partner). Except E2F6, there are also a transactivation domain and a retinoblastoma-binding domain for E2F1 to E2F5 in the C terminus [@B5], [@B7]. Based on functional division, E2F1-3a are usually considered to be a transcriptional activator. When overexpressed, it can drive G0-phase cells into S-phase, while E2F3b-6 are considered to be a transcriptional repressor [@B6]. The overexpression of E2F1-3a can drive G0-phase cells into S-phase, while the cells subjected to E2F1-3 triple knock-out were demonstrated to have defect for S-phase entry [@B6], [@B8]. In contrast, E2F3b-6 attach to the E2Fs-related regulatory promoters at the G0/G1 phase [@B9], [@B10]. The classic E2F-dependent transcriptional activation is executed by a RB/E2F/DP approach. E2F1-5 are associated with one of two DP (DP1/DP2) proteins and one member of pRB (pRB/p107/p130) family from a tripolymer [@B11], [@B12]. In the G1 to S period of the cell cycle, RB is phosphorylated by the cyclin-dependent kinases (CDKs), resulting in the dissociation of RB from the RB-E2F complex and dysfunction of E2F inhibition [@B9], [@B13]. Due to the lack of the retinoblastoma-binding domain, E2F6 is considered to regulate E2F-targeted genes that are independent on pRB family proteins [@B14]. E2F6 does not be regulated by pRB family. It can repress transcription by recruitment of chromatin remodeling complexes that include polycomb group proteins to E2F-target promoters [@B12], [@B15], [@B16]. It is further clarified that E2F1, E2F2 and E2F3 predominantly connect with pRB, and E2F4 is associated with all of the pRB/p107/p130 proteins, whereas E2F5 tends to bind with p130 [@B3], [@B14].\n\nThe atypical E2F consists of E2F7 and E2F8. They have two distinct DNA-binding domains (DBD), but lack DP-dimerization, retinoblastoma-binding, and transcriptional activation domains contained in the canonical E2F. Thus, the atypical E2F binding to DNA is different from other E2F members, which does not rely on DP [@B7], [@B17]. E2F7 and E2F8 control cell cycle genes through homodimers or heterodimers from DBD, functioning predominantly as transcriptional repressors [@B5], [@B18] (Fig.[1](#F1){ref-type=\"fig\"}). E2F8 is a recently discovered member of the E2F family. Christensen *et al.* noticed that there were several murine and human expressed sequence tags showing high similarity to E2F7, using the E2F7 sequence as query searching in the GenBank. Thus, they have designated the cloned transcripts as E2F8 [@B19]. E2F8 shows high similarity to E2F7 in structure and has two special domains exhibiting a high degree of resemblance to the E2F family in DNA-binding domain [@B8]. E2F8 shares the unique structure of E2F7 and performs overlapping functions in many cell metabolisms [@B16].\n\nIn spite of that the DNA binding domain of E2Fs is conservative, there is a lot of accumulated evidence that they also bind and regulate disparate sets of target genes through interactions with specific cofactors [@B20]. In fact, E2F family works more sophisticatedly than it is simply divided into an activator and repressor. Sometimes, they can play as either an activator or a repressor of transcription that depends on target genes, cofactors and cellular context. More importantly, E2Fs usually work as a balanced network [@B3].\n\nStructural properties of E2F8\n=============================\n\nAccessions NM 024680 (human) and XM149937 (mouse) were early identified as E2F8. They encode the FLJ23311 protein of Homo sapiens and the RIKEN cDNA 4432406C08 gene of Mus musculus, respectively [@B19]. NM 024680 and XM149937 include highly homologous ORFs that have the ability to encode proteins consisting of 867 and 860 amino acids, the molecular weights of which are 94272 and 93382 Da, respectively [@B2].\n\nE2F8 shows a high resemblance to E2F7. E2F8 is also found to be the high conservation and unique repeat DNA-binding domains. The result of alignment test shows that the human E2F8 has a 31.9% homology with the E2F7 in overall and the identity area mainly focuses on the DNA binding domain [@B19]. The conserved RRXYD motif is necessary for DNA binding in the E2F family. E2F8 presents the conserved motif in domain 1 (RRIYD) and domain 2 (RRLYD), respectively. The latest data demonstrate that they are located in the residues 113-182 and 261-347 of the E2F8\\'s N-terminal. These two domains of E2F8 are highly similar to the DNA-binding domains of E2F1 to E2F7 [@B2]. In fact, the similarity of DNA binding domains in E2F8 and E2F7 is far more than the conserved RRXYD motif, which extends to about 120 amino acid residues. The sequence between the domain 1 and domain 2 in E2F8 and E2F7 contains about 160 residues, appearing highly similar [@B2]. Like E2F7, E2F8 lacks a marked box, DP-dimerization, pRb-binding and transcriptional activation domains required for other E2F family members [@B19]. E2F8 has duplicated DNA-binding domains and three putative nuclear localization signal sequence (NLS) domains. The integrity of DNA-binding domains is essential for combination with consensus E2F-binding DNA elements. The amino acid mutation of E2F8 in the conserved RRXYD motif will result in failure to bind to DNA [@B2]. Similar to E2F7, the increasing expression of E2F8 negatively influences E2F-target genes, and represses cellular proliferation [@B19]. E2F7 and E2F8 can form homodimers or heterodimers. They usually present in the same adult tissues. These results suggest that the both may have partial overlapping and perhaps synergistic functions [@B2]. The similarity of genomic structure and function of the E2F8 and E2F7 strongly supports that they probably originate from a same ancestral gene.\n\nBoth of human and murine *E2f8* genes are made up of 13 exons and 12 introns. The human and murine *E2f8* genes are located in the short arm of chromosome11p15.1 and chromosome 7, respectively [@B8]. Their initiation codon ATGs are located in the second exons, and the termination codon TGAs present in the exon 13 [@B19]. The human *E2f8* mRNA consists of 323bp long 5\\'-UTR, 2604bp long ORF, and 610bp long 3\\'-UTR, appearing in a variety of tissues. The mouse *E2f8* mRNA comprises 490 bp long 5\\'-UTR, 2583 bp long ORF and a 632 bp long 3\\'-UTR. They have a consensus polyadenylation signals ATTAAA that is located in upstream of the poly (A) tail, respectively [@B8]. The similarity of sequence and genomic structure of the human and murine E2F8 genes strongly supports that they are true orthologues. The entire E2F8 protein is necessary for fully functional activity, but the first 200 residues are more important in the majority of the signaling activity. Molecular weight of E2F8 has slight differences reported in literatures. Predicted molecular mass of the human and mouse E2F8 proteins is about 105KD and 95KD, respectively [@B8]. E2F8 is predominantly expressed in liver, thymus, skin and testis, but not in the brain, stomach, and muscle [@B8]. Studies have shown that the E2F8 is very prone to degradation so that the level of the endogenous protein E2F8 is relatively low [@B19].\n\nE2F8 regulates the cell cycle and cellular proliferation\n========================================================\n\nAccording to the current knowledge of E2F family, we know that E2Fs play a key role in the regulation of cell cycle [@B13]. The E2Fs control transcription of the cell cycle-related genes in the different phases of the cell cycle [@B21], [@B22]. E2F3b-5 bind to E2F-regulated promoters to recruit chromatin remodeling factors in the G0/G1 phase of the cell cycle by dissociating with the pRB family proteins [@B23]. When cells reenter into late G1 phases, the CDKs phosphorylate RB and the free activators E2F1-3a replace the repressor E2F complexes [@B10]. The activators E2F1-3 reach the peak in this G1/S phases. Then, they are attenuated in G2 [@B9]. E2F1-3 also possess the ability to overcome arrest by the cyclin-dependent kinase inhibitor and induce quiescent cells to enter an S-phase [@B23], [@B24]. On the other hand, the group of atypical repressors play a vital role in the cell cycle exit, promoting the process of cell differentiation [@B7].\n\nE2F8 is not only similar to E2F7 in structure, but also has a lot of overlap in many metabolic functions [@B8], [@B20]. The synchronization treatment of HeLaS3 cells and qPCR analysis show that the E2F8 reaches the peak at early S phase, then drops after through S phase ,which is consistent with E2F7 [@B19], [@B25]. Ectopic expression of E2F7 and E2F8 causes the delayed cell cycle progression and the increased proportion of G1 cells [@B2]. E2F7 and E2F8 have an overlapping mechanism in the control of cell cycle. They reciprocally compensate to repress a network of fluctuant cell cycle genes to modulate S-phase progression [@B2], [@B25]. E2F7 and E2F8 can compete with E2F1 to combine with the promoters of E2Fs-targeted genes, inhibiting E2F1-activated gene transcription. On the other hand, E2F7 and E2F8 can directly bind to E2F1 to repress the expression of E2F1 [@B26]. E2F8 is also one of the E2F-targeted genes, and several family members (E2F1, E2F3, E2F4 and E2F7) have been found to combine to the E2F8 promoter, suggesting that there is a feedback-loop in E2Fs [@B19], [@B27]. E2F8 plays as a repressive arm component in monitoring the E2F network, coordinating the cell cycle progression [@B2].\n\nE2Fs are to be critical players in orchestrating the control of cellular proliferation [@B28], [@B29]. The roles of each family member in controlling cellular proliferation, cell cycle transition and apoptosis are different. The expressions of atypical E2Fs during the cell cycle are different from canonical E2Fs, and the former peak levels are found in the S/G2 phase. Interestingly, the ectopic expression of E2F8 leads markedly to the inhibition of E2F-targeted gene expression and reduces the proliferation capacity of mouse embryonic fibroblasts, whereas over-expression of E2F8 in liver cancer cells could promote cell proliferation and tumor generation [@B8], [@B29]. The specific mechanism for the opposite phenomenon of cellular proliferation is still not clear, needing to be further studied. These results also suggest that E2f8 may be an oncogene in the progression of HCC.\n\nE2F8 with E2F7 is essential for the placenta architecture and embryonic development\n===================================================================================\n\nA large number of studies suggest that E2F functions are involved in regulating a variety of cellular metabolism, and playing a key role in cellular homeostasis [@B15]. In fact, E2F function is closely related to the cellular context, whereas *in vivo* environment is more complicated than *in vitro* one. In order to further explore E2F7 and E2F8 functions *in vivo*, scientists utilized a homologous recombination technique and a conditional knockout technology to interfere with E2F7 and E2F8 functions in mice [@B16]. Indeed, genetic ablation of E2F7 and E2F8 demonstrates that at least one allele of *E2f7* or *E2f8* is necessary for embryonic development and viability. The lack of *E2f7* or *E2f8* in mice showed no significant effect on embryo development. However, their combined ablation resulted in massive apoptosis, dilation of blood vessels and hemorrhaging so as to finally induce embryonic death at day E11.5 [@B16]. Interestingly, E2F7 and E2F8 did not offer an equal contribution towards mice postnatal development. Studies have shown that *E2f7^+/-^E2f8^-/-^* mice are normal in postnatal development, but most of *E2f7^-/-^E2f8^+/-^* ones usually appear to be runts, died within the first three months of life [@B16]. In fact, E2F7 and E2F8 could form homodimers or heterodimers to bind to target DNA, which is not equivalent. E2F7:E2F7 is the first priority, and E2F7:E2F8 precedes E2F8:E2F8 [@B1],[@B3]. These data may explain that E2F7 and E2F8 offer an unequal contribution towards mouse postnatal development, while their molecular basis is not yet clear [@B16].\n\nIt was confirmed that there were a large number of cell apoptosis in the *E2f7* and *E2f8* double knockout (DKO) embryos by immunohistochemistry with BrdU-specific antibodies and TUNEL at embryonic day 9.5 (E9.5) [@B16]. The sequential ChIP assays in HEK 293 cells showed that E2F7 and E2F8 could directly bind to *E2f1*. The PCR genotyping of genomic DNA assays in DKO mouse embryonic fibroblast (MEFs) demonstrated that ablated E2F7 and E2F8 would significantly up-regulate E2F1 with a concomitant increase of p53 protein [@B15]. *E2f7^+/-^E2f8^-/-^* mice were mated with either *E2f1^-/-^* or *p53^-/-^* mice in order to generate cohorts of triple knockout (TKO) embryos. The result of TUNEL assays of the TKO mice showed that the massive apoptosis caused by a deficiency in *E2f7* and *E2f8* could be rescued on E9.5 TKO embryos by the loss of either *E2f1* or *p53* [@B16], [@B30]. These results reveal that E2F7 and E2F8 can go through the E2F1-p53 axis to control cell apoptosis what is indispensable for the E2F network. Intriguingly, although cell apoptosis was rescued in both *E2f7*/*8*/*1* and E2f7/8/p53 triple knockout mice, the extent of dilated vessels and hemorrhaging was still similar to DKO embryos when embryos were harvested at E10.5. In fact, because of the vascular defect of TKO, no embryos could exceed E12.5. These results suggest that TKO-rescued apoptosis is independent on vascular defects and embryonic lethality [@B16], [@B31]. DKO-resulted E2F7 and E2F8 target gene deregulation is likely to be involved in angiogenesis. At the same time, deletion of E2F7 and E2F8 resulted in a spectrum of embryonic defects impacting the vasculature and cell survival [@B31]. Their synergistic function may be considered as a special arm of the E2F network implicated in repression of transcription during S-G2, and the E2F1/p53 axis represents a particularly critical target that if not appropriately repressed can cause widespread apoptosis in developing embryos. These findings show that E2F7 and E2F8 are indispensable for embryonic development. They represent a critical role in cell survival and embryonic development by acting as a unique repressive arm [@B16], [@B30], [@B31].\n\nFurther researches reveal that the deficiency of E2F7 and E2F8 in trophoblast progenitor cells of placental is sufficient to kill embryos in mid-gestation. Indeed, a wild type placenta possesses the ability to support the *E2f7^-/-^and E2f8^-/-^* fetuses to birth and beyond. These support that the functions of atypical E2Fs in the extro-embryos are necessary and sufficient for mice embryonic development. Studies showed the TKO of *E2f3a*with *E2f7/8* at the same time rescued placental defects and lethality of fetuses, suggesting that E2F3a resists E2F7 and E2F8 to mediate repression by activating the relevant transcriptional program, which is important for placental development [@B30], [@B32].\n\nE2F8 mediates the DNA damage-correlated cell-cycle\n==================================================\n\nThe structure of atypical E2Fs is similar to a plant E2F-like (E2L) protein, which has two separate DNA-binding domains [@B33]. The function of the plant E2L has been involved in regulation of endo-replication so that atypical E2Fs may also have similar functions [@B34]. With a DNA-damaging agent etoposide or bleomycin to treat different mammal cell lines, including U2OS, T98G, MCF7, HeLa, HCT15, HCT116 and MEF, there were results similar to fluctuation of E2F7 and E2F8 which are concomitant with E2F1 [@B4]. The latter can regulate not only cell cycle, but also cell apoptosis [@B35]-[@B37]. Co-immunoprecipitation (Co-IP) and chromatin immunoprecipitation (ChIP) showed that E2F7 and E2F8 can interact to form dimmer to locate at the promoter of *E2f1*. The DD is independent on DBD, and mutant residues within the DBD do not influence the interaction [@B4]. Besides, the integrity of DD and DBD is necessary for forming dimers and performing biological functions [@B38], [@B39].\n\nEither E2F7- or E2F8-specific siRNA into U2OS cells caused a significant acceleration of cell death, and their combined knockout led to worse results [@B4]. Consistently, the sensitivity of *E2f7^-/-^E2f8^-/-^* cells to DNA-damage-inducing drugs camptothecin and cisplatin showed that these drugs induced a marked increase in apoptosis of DKO MEFs, compared to wild type MEFs [@B16]. Because E2F1 has ability to control apoptosis, overexpression of E2F1 can lead to the apoptosis in the cells treated with DNA damage-inducing agents [@B40]-[@B42]. E2F1 siRNA transfected into DNA-damaged U2OS cells can partly rescue apoptosis, which is caused by defective E2F7 and E2F8 [@B4]. In summary, E2F7 and E2F8 mediate the DNA damage-correlated cell-cycle through modulating E2F1 activity [@B18].\n\nAtypical E2Fs promote angiogenesis and lymphangiogenesis\n========================================================\n\nWith a large number of target genes identified, we have a further understanding of their other functions. Weijts *et al* deleted *E2f7/8* at the same time in zebrafish and mice embryonic, respectively. The results showed an interesting phenomenon, *i.e*., embryonic death without proliferation defects, but displaying massive cell apoptosis and vascular defects [@B16], [@B30]. The specific mutant of*E2f7/8* in the extra-embryonic trophoblastic cells also can lead to a poor structure of the placental vascular network [@B30]. Besides, embryos alone or federatively injected with *E2f7/8* MOs (morpholino oligomers) and genetic mutant embryos of *E2f7/8* showed different degrees of defects in angiogenesis and apoptosis, indicating that the E2f7/8 are necessary for angiogenesis. Intriguingly, deletion of *E2f1* or *p53*can rescue apoptosis caused by deficiency *E2f7/8,*but not do the defect of the blood vessels [@B31]. From these phenomena they reasonably speculate that E2F7/8 can regulate angiogenesis by an alternative mechanism.\n\nAs we know that vascular endothelial growth factor A (VEGFA) is an essential factor for angiogenesis, VEGFA acts as a chemoattractant through forming a correct concentration gradient to guide the formation of an appropriate vascular system [@B43]-[@B47]. The hypoxia-induced factor 1 (HIF1) is a major modulator of VEGFA in the background of angiogenesis [@B48]-[@B50]. The scarce Hif1\u03b1, Hif2\u03b1 or Hif\u03b2 mice are similar to E2f7/8 knockout mice that also die due to vascular defects around E10.5 [@B51]. The injection of *E2f7/8* mRNA increases the endothelial cell numbers to concomitantly raise the mRNA levels of vascular endothelial growth factor A (VEGFA), suggesting that *E2f7/8* excite endothelial cell proliferation through up-regulation of the VEGFA expression. Chromatin immunoprecipitation (ChIP) further confirmed that *E2f7/8* could directly bind to stimulate the VEGFA promoter in a way independent on canonical E2F-binding elements [@B31], [@B48]. Indeed, *E2f7/8* stimulate the VEGFA promoter activity by binding with HIF1 to form a transcriptional complex independent on the canonical E2F-binding site. In sum, E2f7/8 cooperate with HIF1 to form a transcriptional complex to activate the VEGFA promoter, and then to raise the VEGFA level to promote angiogenesis.\n\nAfter finding that atypical E2Fs (E2F7 and E2F8) can control angiogenesis, Weijts *et al*. also have a study on the E2F role in lymphangiogenesis [@B52]. The genome-wide microarray analysis on E2f7/8 deficiency in mouse fetuses at E10.5 revealed that only Ccbe1 and Flt4 were deregulated among the necessary genes for lymphangiogenesis. Ccbe1 and Flt4 contain a canonical *E2f* binding domain near the promoter, and ChIP showed that E2F7 and E2F8 can bind to their promoter [@B52]. The vascular endothelial VegfC/Flt4 (vascular endothelial growth factor receptor 3) signaling pathway is one of main ways to modulate the formation of lymphatic vessels [@B53], [@B54]. Collagen and calcium-binding EGF domain-containing protein 1 (CCBE1) was considered to enhance lymphangiogenesis by increasing the biological effect of VEGFC [@B55]-[@B57]. Atypical E2Fs can direct targets FLT4 and CCBE1, which was identified by a genome-wide approach. Loss of *E2f7/8* resulted in deregulated expression of CCBE1 and FLT4, and impaired venous sprouting and lymphangiogenesis. Intriguingly, E2F7/8 modulated CCBE1 and FLT4 in opposite ways. Inactivation of *E2f7/8* down-regulated CCBE1 and up-regulated Flt4. The expressional analysis in different cell lines (mesenchymal cells and endothelial cells) shows that*E2f7/8* are almost equally expressed in these cell lines, while productions of CCBE1 and FLT4 are dependent on cell types [@B52], [@B58], [@B59]. These results suggest that E2F7/8 do not function as an ON/OFF switch, but fine-tune the expression levels of CCBE1 and FLT4 to ensure the proper signal strength during lymphangiogenesis [@B52]. The regulation of CCBE1 and FLT4 expression by E2F7/8 enhances the biological effect of VEGFC partial monitoring of lymphangiogenesis, but the elaborating mechanism of lymphangiogenesis needs to be further elucidated.\n\nE2F8 influences polyploidization\n================================\n\nPolyploidization occurs in all mammalian species, being the most significant in well-known genome alterations [@B60]-[@B63]. It often exists in some specific tissues, and is an important feature of hepatocytes following weaning and trophoblast giant cells (TGCs) during gestation [@B61], [@B64]. Oxidative damage and surgery regeneration of liver often lead to the enhancement of proportion of polyploid cells in liver, which is also associated with aging [@B65]. The biological functions of polyploidy possibly include the increasing ability to resist adverse conditions, weaken sensitivity to apoptosis, and buffer defence on detrimental mutations [@B64], [@B66].\n\nThe E2F transcriptional factors are closely related to cell-cycle progression, with E2F activators and atypical repressors oppositely to adjust a common set of target genes [@B67], [@B68]. Using a quantitative PCR assay, E2F expression was determined at RNA levels of TGCs in the placenta and postnatal liver development, respectively. The result shows that the endocycling nuclei in TGCs and hepatocytes are most active when total E2F1-3 activator levels are minimized*. E2f1-3* ablation of TGCs in the placenta and hepatocytes in postnatal development results in hyperploidy, accompanying with abnormally large nuclei and increased liver weights. On the contrary, *E2f7^-/-^* and *E2f8^-/-^* DKO reduce ploidy and increase the proportion of G2/M^-^and M phase-related events. Indeed, the ploidy of wild-type TGCs could go up to 1,000C, whereas the *E2f7^-/-^* and *E2f8^-/-^* TGCs with genomes never exceeded 64C [@B64]. The canonical E2F mRNAs showed relatively high levels in 1 week old livers and then precipitous decrease from 3 weeks after birth, whereas they are relatively low in the livers of *E2f7^-/-^* and *E2f8^-/-^* mice at p0 (postnatal day 0) and have a markedly increase in 3 to 5 weeks. Atypical E2F expressional levels are consistent with the polyploidization of hepatocytes. Separated or combined inactivation of *E2f7* (Alb-7ko) and *E2f8* (Alb-8ko) showed that Alb-8ko and Alb-78dko mice had apparently smaller hepatocytes than Alb-7ko, but did not significantly affect hepatic mass. Flow cytometry revealed that Alb-8ko and Alb-78dko prevented hepatocyte endocycles over the whole mouse life, whereas Alb-7ko only had a modest reduction in genome ploidy. Indeed, E2F8 was more enriched on the polyploidy-related gene promoter than E2F7. Compared with the wild type mice, the up-regulated genes in 3-week-old livers were more marked in the Alb-7ko, Alb-8ko and Alb-78dko mice than the down-regulated genes, consistent with their inhibitor properties. Further analysis for these genes raised in both Alb-8ko and Alb-78dko livers revealed that they tended to regulate cytokinetic G2/M progression and mitosis. These data suggest that E2F8 has a vital role for mediating the repression of a set of cell-cycle-related target genes during the process of cell polyploidization [@B69].\n\nIt is worth noting that knock-out of *E2f4* or *E2f5* in mice seems to be little affection of the endocycling in TGCs and hepatocytes. The atypical E2Fs binding to DNA are in the way independent on DP so that they can mediate repression when CDK activity is high in S/G2 phases without the disturbing of the cyclin-CDK axis. Other studies have showed that through activation of the same target genes, deficiency of E2F1 can partially rescue the defect in hepatocyte polyploidization caused by a deficiency in *E2f7* and *E2f8* [@B69]. E2F8 can transcriptionally suppress CDK1 to induce the polyploidization of decidual cells [@B44]. It is necessary to maintain appropriate ploidy of cells, E2F8 abnormal ploidy of hepatocytes might be one of the important causes of HCC. However, the precise molecular mechanism of how the canonical activators and atypical repressors coordinate to control endocycling remains to be further elucidated.\n\nE2F8 and diseases\n=================\n\nThe E2F family has long been expected as potential therapeutic targets of cancer. However, the molecular mechanism about E2F in the occurrence and development of tumor is still not clear, especially the new member E2F8 [@B7], [@B70], [@B71]. Recently, literatures show that E2F8 is obviously up-regulated in HCC, speculating that it is implicated in oncogenesis and progression [@B3], [@B6], [@B11]. Ectopic overexpression of E2F8 by transiently transfecting the recombinant pcDNA3.1-E2F8 into Huh-7, Hep3B, YY-8103 and PLC/PRF/5 HCC cells causes the enhanced cell proliferation, the promoted colony formation and tumorigenicity. On the contrary, knock down of endogenous E2F8 by introduction of siRNAs and shRNA in some HCC cell lines inhibits cellular proliferation and colony formation, and reduces tumor burden [@B29]. Mechanism analysis supports that E2F8 overexpression partially attenuates the E2F1 binding to the oncogenic cyclin D1 promoter in HCC cells in a dominant-negative manner. Cyclin D1 functions as an oncogene involved in many cancers by regulating the G1- to S-phase transition of cell cycle progression. These data support that E2F8 can contribute to cell proliferation of HCC partially *via* E2F8/E2F1/ Cyclin D1 promoting the entry of S phase in cell cycle [@B29].These results suggest that E2F8 promotes HCC occurrence partly through cyclin D1, possibly becoming a potential therapeutic target. E2F8 is also shown to overexpress in ovarian cancer, lung cancer, breast cancer and prostate cancer, and up-regulated in the clinical samples from patients with cancer [@B10], [@B72]-[@B74]. The knockdown of E2F8 significantly inhibits prostate cancer cell growth *via* inducing G~2~/M arrest, but its mechanism is not fully understood.\n\nAlthough most studies indicate that E2F8 functions as an oncogene, recently few reports suggest that it might act as a tumor suppressor. Under the *in vivo* stress of DMBA/TPA skin carcinogenesis model, inactivation of E2f7/8 gives rise to the acceleration of tumorigenesis and deteriorating progression [@B75]. Temporal-specific ablation of E2f8 in the liver at 1-2 weeks of age resulted in severe liver cancer, but not in older age, suggesting that E2F8 acts as a tumor suppressor in postnatal liver development [@B27]. These contradictory results might be due to studies performed in different cell lines and context. No matter how to say, E2F8 expression must be affiliated to the occurrence and development of HCC. It may be a potential therapeutic target of HCC.\n\nE2F8 is not only associated with cancer, but also involved in other diseases, including obesity and hepatic steatosis [@B76]-[@B78]. Loss of RecQ-mediated genome instability 1 (RMI1) inhibits E2F8 expression [@B78], Knockdown of *E2f8* suppresses this glucose inducing up-regulation of RMI1 expression [@B77]. RMI1 and E2F8 consist of a feedback machinery to regulate energy balance in adipose tissue through modulating preadipocyte proliferation. E2F8 is up-regulated in hepatic steatosis. Ectopic expression of E2F8 significantly promotes the fatty acid binding protein 3 (FABP3) expression to cause hepatic steatosis [@B76]. As we know, the advanced stage of nonalcoholic hepatic steatosis results in ultimately cirrhosis and liver cancer, named trilogy for liver lesions [@B79], [@B80]. These results further support the possibility of E2F8 as a therapeutic target for liver cancer.\n\nConcluding remarks\n==================\n\nTaken together, the mammalian transcriptional factor E2F8 plays an essential role in embryonic development and regulation of cellular functions, including cell cycle, cell proliferation, cell survival, DNA damage, angiogenesis, lymphangiogenesis and cell polyploidization (Fig.[2](#F2){ref-type=\"fig\"}). Actually, it plays not only a role in regulating cell functions, but also an important one in balancing the E2F network. E2F8 has a remarkable increase in HCC and ovarian cancer, but no enhancement of E2F7 occurs though the both have functional compensation in many respects. E2F8 can transcriptionally suppress CDK1 to induce hepatocyte polyploidization. Moreover, E2F8 mediates the DNA damage-correlated cell-cycle through modulating E2F1 activity. Both E2F8 and E2F7 cooperate with HIF1 to form a transcriptional complex to raise the VEGFA level to promote angiogenesis. The regulation of CCBE1 and FLT4 expressions by both E2F7 and E2F8 can enhance the biological effect of VEGFC monitoring lymphangiogenesis. Besides, E2F8 also acts as a nonreceptor activator of heterotrimeric G proteins [@B81] that is an important second messenger, involving multiple cell signaling pathways [@B82], [@B83]. These provide a strongly theoretical support that E2F8 may become a promising therapeutic target for some tumors, especially HCC, and hepatic steatosis. The detailed mechanism by which E2F8 regulates cell functions still merits to be elucidated in the future.\n\nThis study was supported by the National Natural Science Foundation of China (grant numbers 81172824, 30971465) and the Foundation of Pearl River Science and Technology New Star (grant number 201506010087).\n\nAuthors\\' contribution\n======================\n\nAll authors contribution to the writing of the review manuscript. F.X and Y.L initiated, conceptualized and wrote the manuscript. F.X, J.L and X.J revised the manuscript.\n\n![**Structures of the E2F family members:** E2F8 has the duplicated DNA-binding domain (DBD), but lacks the dimerization partner (DP) and retinoblastoma-binding (RB) domains.](jcav08p1205g001){#F1}\n\n![**Mechanisms by which E2F8 mediates the cell metabolic functions and histological development**: E2F8 regulates the functions related to cell cycle, proliferation, DNA repair, apoptosis, polyploidization, angiogenesis and lymphangiogenesis.](jcav08p1205g002){#F2}\n\n###### \n\nMain similarities and differences between E2F7 and E2F8\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Object                                     Similarity                                                                    Difference                                                                                            Reference\n  ------------------------------------------ ----------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------- -------------\n  Structure                                  a\\. With two distinct DBD domains\\                                            E2F7 has two highly relevant isoforms, but not E2F8.                                                  2,6,7,17,19\n                                             b. Lacking DP-dimerization and retinoblastoma-binding domains\\                                                                                                                      \n                                             c. Absence of transcriptional activation domains                                                                                                                                    \n\n  Methods and functions of regulating gene   a\\. Controlling cell cycle genes through homodimers or heterodimers by DBD\\   Efficiency for the both to form dimerization is not equivalent.                                       5,16,18\n                                             b. Functioning predominantly as transcriptional repressors                                                                                                                          \n\n  Embryo and postnatal development           Being necessary for embryonic development and viability                       E2F7 appears more important than E2F8 in mice postnatal development.                                  16\n\n  Angiogenesis and lymphangiogenesis         a\\. Being necessary for angiogenesis\\                                                                                                                                               16,30,51,51\n                                             b. Being essential for lymphangiogenesis                                                                                                                                            \n\n  DNA demage                                 Resisting DNA damage                                                                                                                                                                4,18\n\n  Cell polyploidization                                                                                                    E2F8 plays a vital role in the process of cell polyploidization, but E2F7 only has a modest effect.   69\n\n  HCC                                                                                                                      E2F8 shows overexpression in HCC, but not E2F7.                                                       3,6,11\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nDBD: DNA-binding domains; DP: dimerization partner; HCC: hepatocellular carcinoma\n\n[^1]: Competing Interests: The authors have declared that no competing interest exists.\n"}
+{"text": "Introduction {#sec1}\n============\n\nOur perception of time is a highly subjective experience. Indeed, a long tradition in research has shown that the perception of time can be influenced by many different psychological factors ([@ref6]; [@ref27]; [@ref34]; [@ref1]). For example, [@ref26] showed that when we act voluntarily, actions are perceived to occur later in time and action effects are perceived to occur earlier in time, making the time intervals between actions and action effects appear shorter for voluntary compared to involuntary actions. This phenomenon is referred to as intentional binding, or temporal binding, and has been repeatedly demonstrated in previous research (e.g., [@ref23]; [@ref56]; [@ref40],[@ref41]; [@ref20]; [@ref43]; [@ref33]).\n\nGiven the differences in temporal binding effects when contrasting voluntary and involuntary actions, temporal binding has become strongly linked to the sensation of volitionally performing actions and causing effects known as agency ([@ref24]). This link is what makes temporal binding more than just an interesting phenomenon in fundamental research: As an implicit marker of volition, it is strongly related to cognitions about agency, responsibility, and ultimately, moral judgment. Temporal perception therefore covaries with -- and can sometimes even inform -- moral judgment ([@ref42]; [@ref12]). For example, we consider ourselves more responsible for outcomes that are perceived in close temporal proximity to our actions (e.g., [@ref50]). In similar vein, we typically perceive actions and outcomes to occur closer together in time when we also consider ourselves responsible (e.g., [@ref17]).\n\nBinding occurs when agency is high, and actions and effects are perceived as volitionally produced (e.g., [@ref17]). In contrast, binding has been shown to be reduced or absent when actions and effects are externally produced (e.g., [@ref26]). In daily life there are, however, many situations in which there is not such a stark contrast between self- and externally produced actions. Although we *ourselves* may perform actions, we often do so in the context of others who at times subtly, and at other times very overtly, influence the way we act and the way we think about acting. Only a limited number of studies have however looked at the influence of social cues on temporal binding (e.g., [@ref44]; [@ref30]) and there are virtually no studies investigating binding in the context of social action cues ([@ref11]). In the present research, we therefore investigated whether the compatibility of social cues with ongoing actions would influence the perception of time for those self-produced actions.\n\nPremotor and Inferential Accounts of Binding and Agency {#sec2}\n-------------------------------------------------------\n\nBinding occurs when participants willfully act (e.g., [@ref22]), and does not occur in situations when movements are externally induced ([@ref26]; [@ref57]; see also [@ref23], [@ref22]). Although the precise mechanism causing intentional binding is still a matter of empirical debate (e.g., [@ref49]; [@ref28]), traditionally both temporal binding and agency have been argued to emerge by dedicated motor control mechanisms predicting the sensory outcomes of actions (e.g., [@ref8], [@ref9]; [@ref7]; [@ref59]; [@ref26]; [@ref21]). To a large degree, binding is believed to be caused by joint simulations of motor actions and their predicted outcomes, causing temporal perception of the action to be moved toward the effect, and temporal perception of the effect moved toward the action.\n\nOther approaches suggest binding and agency also emerge through inferential mechanisms ([@ref53]; [@ref50], [@ref51]; [@ref39]). Accordingly, binding and agency are influenced when we can easily relate our actions or the action effects we produce to any thoughts preceding it ([@ref16]), and when our agency beliefs ([@ref17]) or the contextual information ([@ref18]) make us appear as the most likely agent.\n\nThe optimal cue integration approach ([@ref38]) and other theoretical approaches (e.g., [@ref47]) consider the premotor and inferential accounts as complementary. Both premotor and inferential cues are considered to be continuously integrated and weighted depending on their availability and reliability in a given situation -- thereby influencing binding and agency.\n\nInferential Manipulations of Temporal Binding {#sec3}\n---------------------------------------------\n\nA number of studies have shown that temporal binding can be influenced by inferential manipulations. For example, [@ref4] and [@ref3] showed that a high number of action possibilities increased temporal binding compared to a low number of action possibilities. And [@ref17] revealed that when participants had high beliefs about personal agency, they showed increased binding -- of the effects toward the actions -- compared to when participants believed their actions were caused by someone else (see also [@ref35]). These findings suggest that binding is influenced by manipulations that are (on face-value) not directly related to premotor prediction.\n\nManipulations that occur before action initiation -- such as the manipulation described above -- may nevertheless indirectly influence premotor prediction. For example, [@ref45] showed a relation between control-beliefs and premotor neural activity: When individuals were led to believe they generally had no personal control over events in their lives these individuals also showed reduced readiness potentials prior to action performance. The research by [@ref45] is a strong example of the way in which high-level manipulations (e.g., beliefs) may have low-level consequences (e.g., premotor activity). One further implication of this research is that it suggests that manipulations that are considered inferential in nature can nevertheless influence premotor prediction -- thereby casting doubt on whether inferential mechanisms directly influence binding.\n\nA way in which one can more convincingly test the unique contribution of inferential mechanisms is by introducing manipulations when actions have already been initiated. For example, a study by [@ref39] showed that binding could occur retrospectively on trials in which no effect was predicted but an effect occurred nevertheless. Moore and Haggard suggest that the emergence of an unexpected effect triggers *post hoc* inferential sense-making ultimately leading to a binding effect. However, whether binding is also influenced by inherently inferential manipulations (e.g., cues suggesting other agents) that are introduced after action-initiation is still unexplored.\n\nAnother possible way to establish the role of inferential processes in binding would be to investigate binding in situations in which motor predictions are less accurate, thereby increasing one's reliance on inferential information ([@ref38]). As we argue in the following section, situations of continuous action performance may provide exactly such a context.\n\nContinuous Action Performance {#sec4}\n-----------------------------\n\nPeople are often required to perform ongoing actions for which they receive continuous feedback. For example, steering and pedaling your bike on the way to work can be perceived as one long action in which you are provided with a strong and rich flow of sensory information. However, the methodology of studies in the agency domain typically does not feature actions that take relatively long to resolve -- instead actions and action outcomes are quick and discrete. As was argued by [@ref54], it is important that agency paradigms include longer actions as the processes underlying agency for discrete versus prolonged actions may differ. Specifically, from a motor prediction perspective it may be difficult to generate accurate predictions about sensory feedback for ongoing actions. Keeping to the previous example, biking to work would involve rich and continuous feedback, and due to the increased sensory input and overall complexity in this situation one's predictions may become less specific -- especially compared to discrete one-action one-outcome settings. In line with the optimal cue integration approach ([@ref38]), such a reduced reliability from motor predictions can increase the degree to which people rely on inferential information in their judgment of agency. However, more studies that feature inferential manipulations and relatively longer actions are required to validate the notion that inferential information is important in such settings.\n\nAction Cues Influence Agency {#sec5}\n----------------------------\n\nIn the agency domain a number of studies have investigated the degree to which cues related to actions can influence agency. [@ref55] subliminally cued participants with arrow symbols pointing to the left or right just before participants were about to press a left or right button. Their results showed that when the direction of the arrows pointed toward the spatial location of the action, individuals reported increased agency compared to when the arrows did not point toward the action location. According to [@ref55], the subliminally presented compatible primes enhanced the ability to select the appropriate action, and this action fluency then increased the sense of agency. A recent paper by [@ref46] reported similar results. Specifically, they showed that supraliminally presented arrow symbols pointing toward the compatible response locations in a Flanker task were related to increases in explicit agency ratings compared to arrow symbols pointing toward incompatible response locations. In conclusion, previous literature shows that cues that affect the fluency of action selection can influence the sense of agency. The question then remains whether action cues also influence measures of binding.\n\nSocial Processes in Binding and Agency {#sec6}\n--------------------------------------\n\nAgency is influenced by the principle of exclusivity, which holds that it is important to our sense of agency to be able to perceive ourselves -- and not others -- as the most likely cause for actions and effects. Many of our actions are however performed in a social context: We imitate ([@ref37]; [@ref52]) and conform to others ([@ref2]); we can be susceptible to suggestions from others ([@ref5]); and sometimes we are subject to direct commands of others telling us what to do ([@ref32]; [@ref36]). Social cues can therefore \"threaten\" our principle of exclusivity in ways that do not apply to non-social cues: While seeing an arrow pointing left may help us to press a left button and increase agency ([@ref55]; [@ref46]), an agent specifically telling us to go left may \"threaten\" the principle of exclusivity, and lead to a reduced sense of agency.\n\nThe unique influence of social action cues was shown in a paper by [@ref15]. Specifically, in their paradigm they introduced another agent (\"the computer\"), and presented participants with action instructions (the words \"left\" and \"right\"). When the instructions were subliminally presented the participants reported increased agency when there was a match between instructions and actions (similar to the findings by [@ref55]; [@ref46]). However, when the instructions were presented supraliminally and could be consciously perceived, the effects reversed: The visible and compatible instructions lowered the sense of agency instead. [@ref15] argued that in all likelihood, the conscious awareness of the other agent and the presented instructions could have threatened the participants' sense of having been an independent agent, resulting in a reduced sense of agency when their actions matched the instructions.\n\nChallenges {#sec7}\n----------\n\nTo reiterate, a number of important challenges remain to be addressed in the agency domain. First, although the principle of exclusivity is considered vital to the sense of agency ([@ref50], [@ref51]), only a small number of studies on agency in general -- and on binding in particular -- involve social cues. This reflects a gap in the literature that needs to be addressed, especially given the fact that a previous study measuring explicit agency suggests that social action cues decrease agency when the cues are consciously perceived and cued actions are subsequently performed ([@ref15]). However, whether such compatibility effects emerge on implicit measures of agency (i.e., temporal binding) is unclear, thereby reflecting an important second challenge to address. Third, although inferential agency manipulations have been investigated using implicit measures ([@ref17]; [@ref3]), such manipulations are typically introduced before actions are performed. This makes it difficult to determine whether the processes underlying the reported effects are inferential in nature or actually emerge from motor prediction. Fourth, as agency methodologies typically involve short actions, there is a lack of knowledge on agency and binding when actions take long to resolve ([@ref54]). The present research was designed to address these challenges.\n\nThe Present Research {#sec8}\n--------------------\n\nIn the present research, we report three studies in which we aimed to investigate whether temporal binding for self-produced actions is influenced by the compatibility of action cues with those actions. To measure temporal binding, participants were required to act, and when they had concluded their action, they were asked to estimate and report the time it took from the moment they initiated their action until they caused an effect.\n\nParticipants were presented with verbal action cues. The verbalizations related to actions and not to action outcomes, making them less likely to influence processes of outcome prediction. Additionally, the verbalizations were presented after individuals had already initiated their actions, reducing the probability that predictive mechanisms play a role. Furthermore, verbalizations were introduced during continuous action performance, a context in which people may rely more on inferential cues ([@ref54]).\n\nTo test whether action cues influence temporal binding, we manipulated the content of the verbalizations. In Study 1, participants heard verbalizations that were either compatible with the performed action or were in opposition to the performed action. The primary aim of Study 1 was to test the theory proposed by [@ref15], that acting \"against\" instructions or cues would be likely to increase the sense of having been an independent agent, whereas \"following\" suggestions is likely to threaten this view, and reduce agency instead. As such we expected that time interval estimates would be higher (indicating a lower sense of agency) when participants heard a verbalization that was compatible to the action they were performing, and time estimates would be lower (indicating a higher sense of agency) when opposite verbalizations were presented.\n\nAs we expected, these manipulations only to have effects in the context of action performance; in Study 2, participants were required to observe actions and action effects by another agent (the \"computer\") while also being presented with compatible and opposite verbalizations. The primary aim of Study 2 was therefore to establish that compatibility effects only emerge in the context of action performance. As such, we expected no effects of the verbalizations considering that participants were not actively involved.\n\nFinally, Study 3 was largely a replication of Study 1 besides that participants additionally heard verbalizations that were unrelated to the performed task. These verbalizations served as a control condition against which the directionality of (in)compatibility was investigated. Compatible primes may reduce agency, incompatible primes may increase agency, or both may apply. The primary aim of this study was to explore the direction of the compatibility effect.\n\nMethods {#sec9}\n=======\n\nParticipants {#sec10}\n------------\n\nEach study featured 50 undergraduate students who participated in exchange for a small fee (Study 1: 43 females; *M*~age~ = 22.28, age range 18--45; Study 2: 30 females; *M*~age~ = 22.34, age range 18--31; Study 3: 34 females; *M*~age~ = 22.44, age range 18--43). Although the majority of participants was female, gender is typically not a meaningful predictor in research on agency and binding. A participant involved in one study was excluded from participation in the other studies. The studies were approved by the Research Ethics Board of Utrecht University's Faculty of Social Sciences and all participants provided written informed consent prior to beginning the study.\n\nMaterials and Procedure {#sec11}\n-----------------------\n\nParticipants were presented a red circle on their monitor, ostensibly representing a balloon. In Studies 1 and 3, participants were instructed to inflate the balloon until it popped. They could start inflating the balloon by clicking once on the red circle and holding down the mouse-button, thereby causing the red circle to increase in size. In Study 2, participants were not required to click and hold down the mouse-button. Instead of performing an action, participants saw the \"computer's cursor moving toward the balloon, then inflating the balloon until it popped.\n\nAfter a fixed period of time (2.5 vs. 3.0 vs. 3.5 s), the balloon popped, indicated by the recorded sound of a balloon burst and the presentation of the picture of a popped red balloon for 0.5 s. Participants were told to release the mouse-button as quickly as possible when the balloon popped (In Study 1, we logged how quickly participants released the button after the balloon popped. This was after 0.314 s on average. There were no differences in release-times between the instruction conditions). Participants then used the mouse to give an estimation of the time it took for the balloon to pop from the moment they/the computer had started inflating it. Participants were required to provide their answer within a range of 2.0--4.0 s. In case participants did not hold down the mouse-button in Studies 1 and 3, and accidentally released it prior to the balloon burst, an error message appeared and the trial restarted.\n\nParticipants were explained that during the inflation process, they would hear recorded voices delivered over a headset. Participants were informed that the voices were contextual and not task-relevant -- they should listen to them, but not be influenced by them. Participants in Studies 1 and 2 would hear the words \"Press\" (compatible action verbalization) or \"Stop\" (opposite action verbalization) evenly divided over 60 trials. Participants in Study 3 would hear \"Press\" (compatible action verbalization); \"Stop\" (opposite action verbalization); or \"Swim\" (unrelated action verbalization) evenly divided over 90 trials. The verbalizations were delivered 2 s before trial end, were equalized in volume level, and were recorded and configured to last exactly 0.5 s. The overall procedure is visualized in [Figure 1](#fig1){ref-type=\"fig\"}.\n\n![Experimental overview. The upper half depicts a 3.0 s trial from Study 1. The lower half depicts a 3.0 s trial from Study 2.](fpsyg-11-00160-g001){#fig1}\n\nTemporal Binding {#sec12}\n----------------\n\nIn the domains of temporal binding and agency, time perception is often measured either using the Wundt-clock paradigm ([@ref26]; [@ref57],[@ref58]; [@ref21]; [@ref39]; [@ref40],[@ref41]) or using the interval judgment task ([@ref23], [@ref22]; [@ref13]; [@ref25]; [@ref14]). In the Wundt-clock paradigm, conclusions about temporal binding are drawn based on the differences with the estimated and actual time events, using a virtual clock interface as an instrument to gauge time, and featuring baseline blocks to be able to correct for individual error. This allows not only for contrasts between experimental conditions but also for contrasts of subjective time compared to real-time. The interval judgment task in the present experiment does not feature comparable procedures to make the subjective experience more in line with objective reality. It is thereby well able to capture one's subjective experience of time intervals, but it is also likely to show deviation from real-time ([@ref10]; [@ref29]). We therefore do not make comparisons between estimated and real-time intervals, and reported total interval estimates -- not shifts from real-time. Temporal binding in the present study refers to relatively lower estimates in contrasts between experimental conditions, not to a reduction compared to the actual time intervals.\n\nResults {#sec13}\n=======\n\nMeans and SD's of the experimental conditions in Studies 1--3 are reported in [Table 1](#tab1){ref-type=\"table\"}.\n\n###### \n\nMeans and SD's (in brackets) of the experimental conditions in Studies 1--3.\n\n                                           Compatible cue   Opposite cue    Unrelated cue\n  ----------------------- ---------------- ---------------- --------------- ---------------\n  **Study 1** (act)       2.5 s interval   2.736 (0.229)    2.693 (0.225)   \n  3.0 s interval          3.023 (0.279)    2.996 (0.299)                    \n  3.5 s interval          3.418 (0.361)    3.349 (0.351)                    \n  Mean (SD)               3.059 (0.269)    3.013 (0.262)                    \n  **Study 2** (observe)   2.5 s interval   2.852 (0.249)    2.821 (0.244)   \n  3.0 s interval          3.106 (0.267)    3.099 (0.272)                    \n  3.5 s interval          3.399 (0.283)    3.414 (0.322)                    \n  Mean (SD)               3.119 (0.233)    3.111 (0.247)                    \n  **Study 3** (act)       2.5 s interval   2.740 (0.334)    2.692 (0.341)   2.634 (0.328)\n  3.0 s interval          3.031 (0.310)    2.986 (0.320)    2.913 (0.318)   \n  3.5 s interval          3.362 (0.344)    3.327 (0.330)    3.245 (0.358)   \n  Mean (SD)               3.066 (0.304)    2.909 (0.269)    2.990 (0.283)   \n\n*Numbers represent seconds. Actual time intervals were 3 s on average*.\n\nStudy 1 {#sec14}\n-------\n\nA 3 (Time interval: 2.5 vs. 3.0 vs. 3.5 s) \u00d7 2 (Cue compatibility: compatible vs. opposite) repeated measures Analysis of Variance (ANOVA) on the average time estimates in Study 1 revealed a significant main effect of Cue compatibility, *F*(1, 49) = 8.507, *p* = 0.005, $\\eta_{p}^{2}$ = 0.148. Participants' time estimates were higher on the trials in which they heard a compatible cue, compared to the trials in which they heard an incompatible cue. As expected, there was also a main effect of the Time interval condition, *F*(2, 98) = 262.673, *p* \\< 0.001, $\\eta_{p}^{2}$ = 0.843. Participants' time estimates mapped the different trial intervals; estimates were smaller on the trials in which the timer interval was shorter, and higher on the trials in which the time interval was longer (*M*~2.5s~ = 2.715, SD = 0.212 vs. *M*~3.0s~ = 3.010, SD = 0.283 vs. *M*~3.5s~ = 3.384, SD = 0.346; all planned contrasts were statistically significant, *p*'s \\< 0.001; in all studies, linear approaches to the data were superior to quadratic approaches). There was no interaction effect between the Time interval and Cue compatibility conditions, *F*(2, 98) = 1.200, *p* = 0.306, $\\eta_{P}^{2}$ = 0.024.\n\nThe compatibility effect was in line with our hypotheses and Study 1's results provide support to the notion that overt and compatible action cues reduce agency compared to incompatible action cues ([@ref15]).\n\nStudy 2 {#sec15}\n-------\n\nStudy 2 aimed to show that no compatibility effects emerge when individuals do not act. A 3 (Time interval: 2.5 vs. 3.0 vs. 3.5 s) \u00d7 2 (Cue compatibility: compatible vs. opposite) repeated measures ANOVA on the average time estimates in Study 2 indeed showed no main effect of Cue compatibility when individuals observed another agent and did not act themselves, *F*(1, 49) = 0.341, *p* = 0.562, $\\eta_{P}^{2}$ = 0.007. Bayesian model averaging techniques using JASP ([@ref31]; [@ref48]) revealed a BFInclusion = 0.103, indicating that these data (strongly) support null-results over actual interaction effects. There was a main effect of Time interval, *F*(2, 98) = 200.352, *p* \\< 0.001, $\\eta_{P}^{2}$ = 0.803. Shorter intervals led to lower time estimates compared to longer intervals (*M*~2.5s~ = 2.837, SD = 0.233 vs. *M*~3.0s~ = 3.102, SD = 0.262 vs. *M*~3.5s~ = 3.406, SD = 0.290; all planned contrasts were statistically significant, *p*'s \\< 0.001). There was no interaction between the Cue compatibility and Time interval conditions, *F*(2, 98) = 1.101, *p* = 0.337, $\\eta_{P}^{2}$ = 0.022.\n\nAlthough much caution must be taken when statistically contrasting effects from separate studies, we performed a 2 (Cue compatibility: compatible vs. incompatible within-subjects) \u00d7 2 (Action: acting vs. observing between-subjects) repeated measures ANOVA that included participants from both studies. This analysis showed a main effect of the Action condition, as participants who acted showed more binding compared to participants who only observed, (*M*~act~ = 3.036, SD = 0.255 vs. *M*~observe~ = 3.208, SD = 0.255; *F*(1, 98) = 11.173, *p* = 0.001, $\\eta_{P}^{2}$ = 0.102). But more importantly, there was a significant interaction effect between the Cue compatibility and Action conditions, *F*(1, 98) = 107.395, *p* \\< 0.001, $\\eta_{P}^{2}$ = 0.523. This interaction reflects the significant difference we observed of Cue compatibility when participants acted (S1), and the absence of a compatibility effect when participants did not act (S2).\n\nStudy 3 {#sec16}\n-------\n\nStudy 3 required participants to act and featured cues that were compatible, incompatible, or unrelated -- thereby allowing us to investigate the direction of the compatibility effect. A 3 (Time interval: 2.5 vs. 3.0 vs. 3.5 s) \u00d7 3 (Cue compatibility: compatible vs. incompatible vs. unrelated) repeated measures ANOVA on the average time estimates in Study 2 revealed a significant main effect of Cue compatibility, *F*(2, 98) = 20.416, *p* \\< 0.001, $\\eta_{P}^{2}$ = 0.294. Compatible cues led to higher time estimates than unrelated or opposite cues, and incompatible cues led to lower time estimates than unrelated and compatible cues (all contrasts were significant, *p's* \\< 0.010). The analysis also revealed a significant main effect of Time interval, *F*(2, 98) = 146.964, *p* \\< 0.001, $\\eta_{P}^{2}$ = 0.750. Shorter intervals led to lower time estimates compared to longer intervals (*M*~2.5s~ = 2.689, SD = 0.319 vs. *M*~3.0s~ = 2.977, SD = 0.297 vs. *M*~3.5s~ = 3.311, SD = 0.332; all planned contrasts were statistically significant*, p*'s \\< 0.001). No interaction effect between the Cue compatibility and Time interval conditions emerged, *F*(4, 196) = 0.197, *p* = 0.940, $\\eta_{P}^{2}$ = 0.004.\n\nGeneral Discussion {#sec17}\n==================\n\nSummary of Findings {#sec18}\n-------------------\n\nThe present line of studies showed that individuals' time estimates were influenced by verbal cues. Even though these cues were contextual and participants were instructed not to be influenced by them, trials on which participants heard cues that were compatible with their actions were experienced as relatively longer, and trials on which participants heard cues that were in opposition or unrelated to their actions were experienced as relatively shorter. This effect only occurred when participants were acting (Studies 1 and 3), not when they passively observed the same actions and effects being caused by an external entity (Study 2).\n\nTheoretical Implications {#sec19}\n------------------------\n\nA paper by [@ref15] showed that compatible supraliminal action primes presented before participants performed their actions led to lower explicit agency ratings than incompatible action primes. As binding in time perception has been linked to increases in the sense of agency, and separation in time perception has been linked to decreases in agency ([@ref26]; [@ref23], [@ref22]; [@ref40],[@ref41]), our results are in line with the results by [@ref15]. Additionally, the present research expands on that study by showing similar effects can occur when actions are already ongoing and when the contextual action cues are presented during action performance.\n\nThe single-word-instructions in the present research arguably represent social cues in their most limited sense ([@ref60]). However, previous literature has shown that completely non-social cues can have very different effects on measures of binding and agency. For example, when actions are primed and cued by arrows and targets on the Flanker task, compatible primes have been shown to increase agency (or alternatively, incompatible primes decrease agency; [@ref46]). However, cues that signal the presence and/or wishes of other agents seem to influence agency very differently: they reduce agency when we perceive our actions to be in line with them ([@ref15]; [@ref11]).\n\nTemporal binding and agency emergence have both been explained through processes relating both premotor prediction ([@ref26]; [@ref21]; [@ref49]) as well as post-motor inferences ([@ref50], [@ref51]; [@ref39]). However, many studies investigating the relation between post-motor inferential mechanisms and binding featured manipulations prior to action performance (e.g., [@ref17]; [@ref4]). Given that these premotor inferential manipulations may (also) influence premotor prediction (e.g., [@ref45]), it was important to further validate the unique contribution of inferential mechanisms to temporal binding. We attempted to do this by using inferential cues that related to actions -- not outcomes -- by using manipulations that only occurred after action initiation, and, by presenting those cues in a context of ongoing action performance where people may rely more strongly on inferential cues in general ([@ref54]). As such, we believe our results further support the notion that inferential mechanisms are indeed important to effects of temporal binding.\n\nThe Use of Continuous Actions in Binding and Agency Research {#sec20}\n------------------------------------------------------------\n\nPrevious studies in the domains of binding and agency regularly required their participants to perform a single *short* action (e.g., a button-press), to observe a single event (e.g., a tone), and subsequently to report agency scores or measures of time perception. As far as we know, the present research is the first to link *implicit* agency and time perception in a task featuring an ongoing action. This approach was useful for our purposes, as we were able to show that action cues influenced time estimation. However, whether the mechanisms of temporal binding are actually similar for ongoing isometric action performance compared to discrete action performance needs to be addressed by future research.\n\nLimitations and Suggestions {#sec21}\n---------------------------\n\nSince participants always pressed the button there was a slight asymmetry to the conditions, as the \"press\" action cue condition was always compatible with participants' actions. As such, it is possible that our results were caused by more than mere compatibility but were also driven by differences in the psychological processing of default versus non-default options. Investigating prime-action compatibility when presenting multiple action possibilities and comparing default and non-default options therefore seem intriguing and important avenues for future research to take.\n\nConclusion {#sec22}\n----------\n\nThe perception of time may be one of the most pervasive yet subjective experiences in human consciousness ([@ref6]; [@ref1]; [@ref19]). The present research adds to line of research on the subjective nature of our perception of time. It shows that our perception of time is influenced not only by our own actions but also by the compatibility of those actions with contextual cues.\n\nData Availability Statement {#sec23}\n===========================\n\nMethodology and Data files are accessible via the Surfdrive repository at the following location: <https://surfdrive.surf.nl/files/index.php/s/fEi1h0bbtnamWk2>.\n\nEthics Statement {#sec24}\n================\n\nThe studies involving human participants were reviewed and approved by the Research Ethics Board of Utrecht University's Faculty of Social Sciences. All participants provided their informed consent to participate in this study.\n\nAuthor Contributions {#sec25}\n====================\n\nAll authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.\n\nConflict of Interest {#sec26}\n--------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nWe would like to thank Mirjam Moerbeek and Caspar van Lissa from the Department of Methodology and Statistics of Utrecht University for valuable feedback.\n\n[^1]: Edited by: Francesca Marina Bosco, University of Turin, Italy\n\n[^2]: Reviewed by: Mark Allen Hamilton, University of Connecticut, United States; Thora Tenbrink, Bangor University, United Kingdom\n\n[^3]: This article was submitted to Cognitive Science, a section of the journal Frontiers in Psychology\n"}
+{"text": "![](medchirrev77258-0155){#sp1 .153}\n"}
+{"text": "![](envhper00345-0208.tif \"scanned-page\"){.381}\n\n![](envhper00345-0209.tif \"scanned-page\"){.382}\n\n![](envhper00345-0210.tif \"scanned-page\"){.383}\n\n![](envhper00345-0211.tif \"scanned-page\"){.384}\n\n![](envhper00345-0212.tif \"scanned-page\"){.385}\n\n![](envhper00345-0213.tif \"scanned-page\"){.386}\n\n![](envhper00345-0214.tif \"scanned-page\"){.387}\n\n![](envhper00345-0215.tif \"scanned-page\"){.388}\n\n![](envhper00345-0216.tif \"scanned-page\"){.389}\n\n![](envhper00345-0217.tif \"scanned-page\"){.390}\n"}
+{"text": "1. Introduction {#sec0001}\n===============\n\nCurrently, with the significant advancement of nanotechnology, a variety of nanomedicines have been developed and introduced into the field of cancer treatment, including inorganic nanocrystals [@bib0001], polymer nanoparticles [@bib0002], biomimetic nanomaterials [@bib0003], and nanohybrids [@bib0004]. Most of the common nanomedicines incorporate nanocarriers such as nanoMOFs [@bib0005], 2D nanosheets [@bib0006], and nanogels [@bib0007] as anticancer drug-delivery nanoplatforms for oncotherapy. Typically, most of the nanomedicines, especially drug delivery nanoplatforms, suffer from the bottlenecks of low efficiency, serious side effects, and limited target efficiency to tumor tissues, which usually limit clinic translation potential. To overcome these disadvantages, the stimuli-responsive nanomedicine has been developed to precisely release drugs at specific sites for noninvasive cancer therapy [@bib0008]. Prodrug based nanomedicines are one of the typical representatives of stimuli-responsive nanomedicines that controllably release drug under external stimulations.\n\nThe prodrug (drugs conjugated to the precarrier), an inactive compound that can be enzymatically degraded into the parent bioactive drug in vivo, has gained wide attention as a practical approach to reduce off-target toxicity in cancer treatment. Due to its unique structure, most prodrugs can self-assemble into nanoparticles through simple chemical modifications. The prodrug exhibits several significant advantages, including high drug loading efficiency, ameliorative drug availability, drug release controllability, and less tendency to agglomerate after encapsulation [@bib0009]. Considering the different structures, prodrug-based nanomedicines can be classified into three categories: prodrug-based polymer conjugated nanomedicines, small molecular prodrug-based self-assembled nanomedicines, and prodrug-encapsulated nanomedicines.\n\nCompared to other types of nanomedicines, the prodrug-based nanomedicines are able to cleverly and sensibly respond to certain stimuli and release the desired drugs. After decades of development, numerous prodrug-based nanomedicines with different chemical modifications have emerged as corresponding stimulus responsive, including the internal tumor microenvironment (TME) stimulus (e.g., pH, redox environment, enzymes and hypoxic, etc.) [@bib0010], [@bib0011], [@bib0012], [@bib0013], [@bib0014], [@bib0015] and the external environmental stimulus (e.g., light, thermo, ultrasound, and magnetism field, etc.) [@bib0016], [@bib0017], [@bib0018]. Most importantly, selecting the proper type of stimuli and chemical modifications in different situations are the key to design satisfying prodrug-based nanomedicines.\n\nAlthough many reviews discuss the progress of prodrugs for cancer therapy, very few articles focus on the stimuli-responsive prodrug nanomedicines. This review will summarize essential aspects of stimuli-responsive prodrug nanomedicines in cancer therapy; including various types of prodrug-based nanoparticles, different stimulus, challenges, perspective, and recent advancement in clinical applications ([Scheme 1](#fig0001){ref-type=\"fig\"}). We expect this timely review article to provide deep insights into the development of stimuli-responsive prodrug-based nanomedicines for cancer therapy.Scheme 1Schemes of different types of prodrug-based nanomedicine and their stimuli-responsive pathways and applications in the field.Scheme 1\n\n2. Types of prodrug-based nanomedicines {#sec0002}\n=======================================\n\nThe types of prodrug-based nanomedicines mainly include, prodrug based polymer conjugated nanomedicines, small molecular prodrug-based self-assembled nanomedicines, and prodrug-encapsulated nanomedicines.\n\nFor prodrug-based polymer conjugated nanomedicine, drug molecules are usually covalently conjugated to polymers via chemical modifications to configure prodrug-based nanoparticles (NPs). Currently, conjugating drug molecules to the polymer scaffold is the most widely used approach to synthesize prodrug-based polymer nanomedicines. Various polymers, including hydrophilic polymer, hydrophobic polymer, and amphiphilic polymer, have been employed to synthesize polymer-drug conjugates for cancer therapy. On the structural bases, the polymers are divided into three categories that are widely utilized to synthesize polymer-drug conjugates: (i) comb-like copolymers [@bib0019]; (ii) block copolymers [@bib0020]; and (iii) dendritic copolymers [@bib0021]. These copolymers have common advantageous characteristics such as: it contains numerous functional sites on the polymer skeleton, and anticarcinogen molecules are mainly cross-linked to the polymer as branched chains to achieve high drug loading efficiency. In addition, poly(ethylene glycol) (PEG) block copolymers, polypeptides [@bib0022], polysaccharides [@bib0022], and polyaminoacids [@bib0023] also contain a great deal of active functional groups such as carboxyl, amino, and hydroxyl which could be easily modified in chemical manner to form satisfying polymer-drug conjugates nanomedicines.\n\nAnother prodrug-based nanomedicine is the small molecular prodrug-based self-assembled nanomedicine. This kind of nanomedicine has the following characteristics: (i) the molecular weight of the nanocarrier is very low with a small nanostructure, (ii) the anticarcinogen regulates the hydrophilic and hydrophobic balance of conjugates. Different from polymer prodrugs, the small molecular prodrugs are usually obtained via cross-linking the molecule to another molecule. Subsequently, the small molecular prodrug would self-assemble into prodrug-based nanomedicines. Most of these nanomedicines mainly contain cyclodextrin--drug conjugates [@bib0024], pillararene-drug conjugates [@bib0025], amphiphilic peptidic prodrugs [@bib0026], and lipid drug conjugates (e.g., squalenoylations [@bib0027], cholesteryl [@bib0028], and unsaturated fatty acids [@bib0029] prodrugs) of anticancer agents. Additionally, though the small molecular prodrug-based self-assembled nanomedicines possess many advantages, including smaller size and higher drug loading efficiency, nevertheless, shortcomings such as short blood circulation life and poor structural stability can not be ignored. Therefore, an amphiphilic copolymer containing a long amphiphilic chain such as phospholipid polyethylene glycol (DSPE--PEG) is usually introduced and implemented to modify the small molecular prodrug self-assembled conjugates for prolonging the blood circulation time and improving the structural stability [@bib0030].\n\nProdrug-encapsulated nanomedicines is a special kind of nano-assembled prodrug, which mainly utilizes nanocarriers (e.g. inorganic NPs, liposomes, micelles or nanogels) to load or encapsulate anticancer prodrugs through a noncovalent binding. Benefiting from the merits of prodrug and nanoscience, the obtained prodrug-encapsulated nanomedicines have higher targeting efficiency, better drug utilization and less side effect [@bib0009]. Notably, for these kinds of nanomedicines, any suitable biological material can be selected as the carriers for prodrug delivery without additional restrictions. However, with no difference from the common drug delivery systems, the low drug loading efficiency still limits the clinical development of prodrug-encapsulated nanomedicines [@bib0031]. Although this prodrug strategy is not more significant as compared to the above mentioned two strategies, it has dramatically attracted more attention in recent years due to the unique advantage of no carrier restrictions.\n\n3. Stimuli-responsive drugs release {#sec0003}\n===================================\n\nThis section will introduce the general stimuli-responsive drugs release classifications/types of prodrug-based nanomedicines, including endogenous stimuli and exogenous stimuli.\n\n3.1. Endogenous stimuli {#sec0004}\n-----------------------\n\n### 3.1.1. pH-responsive {#sec0005}\n\nThe anaerobic glycolysis of the tumor cells results in the accumulation of the acidic metabolite, therefore the TME presents weak acidity feature. Rapid tumor growth leads to irregular angiogenesis, thereby causing a rapid depletion of both oxygen and nutrients. The glycolytic metabolism involved in this process results in the generation of acidic metabolites in the tumor stroma. Under this condition, numbers of prodrug-based nanomedicines with low pH-responsive polymer conjugates have been designed. The pH-sensitivity relies on the degradation of acid-cleavable linkers or the protonation of ionizable groups in low acidic conditions [@bib0008]. These prodrug-based nanomedicines retains their stability during the systemic circulation under normal physiological pH conditions (pH 7.4) while releasing the anti-tumor drugs in the tumor tissues (pH 6.5--7.0).\n\nFor instance, Zhang et al. combined doxorubicin (DOX) with the polymer backbone through acid-labile hydrazone bond to form polymeric DOX (PDOX) prodrug-based nanoparticles([Fig. 1](#fig0002){ref-type=\"fig\"}a) [@bib0012]. Meanwhile, the photothermal agent of near-infrared (NIR)-absorbing dye IR825 and a targarting ligand of folic acid were modified to the self-assembled prodrug nanoparticles, respectively. Because of the acid-labile hydrazone linkage, the PDOX/IR825 could be readily triggered to release the conjugated DOX in the acidic TME**(**[Fig. 1](#fig0002){ref-type=\"fig\"}b**)**. Notably, significant tumor suppression performance was observed in the zebrafish liver hyperplasia model when they were treated with combined photothermal-chemotherapy of PDOX/IR825 nanomedicine **(**[Fig. 1](#fig0002){ref-type=\"fig\"}c). Furthermore, Hu et al. also developed a pH-responsive bortezomib (BTZ) dendrimer prodrug-based nanomedicine, which displayed prominent pH-dependent prototype drugs release behavior and good tumor inhibition [@bib0032].Fig. 1(a) Schematic illustration of IR825 loaded PDOX prodrug NPs for targeted and combined PTT and chemotherapy. (b) In vitro DOX release from PDOX prodrug NPs in different pH at 37\u00a0C. (c) The relative tumor volume changes of transgenic zebrafish based liver tumor after different treatment. Reproduced with permission\\[12\\]. Copyright 2016, American Chemical Society. (d) Schematic illustration of GSH-sensitive HDMP (CPT-ss-HPPH) with high loading capacity for efficient drug accumulation in tumor and synergistic chemotherapy and PDT. (e) In vitro CPT release with or without GSH in PBS at 37 \u00b0C. (f) The tumor growth curves of tumor-bearing mice with different treatment (*n*\u00a0=\u00a05). Black arrows represent intravenous injection of drugs; red arrows represent laser irradiation. Reproduced with permission\\[36\\]. Copyright 2018, Wiley-VCH. (g) Schematic illustration of the polyMTO NP platform for targeted and deeply penetrating cancer therapy. (h) KO~2~ concentrations-dependent release of MTO from the polyMTO NPs in PBS. (i) Tumor growth curve of the LNCaP xenograft tumor-bearing mice after various treatments. Reproduced with permission\\[10\\]. Copyright 2017, Wiley-VCH.Fig 1\n\n### 3.1.2. GSH-responsive {#sec0006}\n\nThe intracellular cytoplasm of cells contains more reductive species than the extracellular matrix. The concentration of GSH in the intracellular cytoplasm is two to three orders of magnitude higher than that in the extracellular matrixes [@bib0033]. Furthermore, GSH has a remarkable influence on many cell functions, including cell cycle regulation, gene expression, immune responses, protein function, and activation of cell death [@bib0034]. Meanwhile, there is evidence that the GSH level in tumor tissues is much higher than those in normal tissues [@bib0035]. During the past decades, the chemical bonds including disulfide bond, diselenide bond and some functional groups of *cis,cis*,trans-diamminedichlorodisuccinato-platinum (DSP) or trimethyl-locked benzoquinone (TMBQ) have been utilized to bridge anticancer drugs and polymers to synthesize GSH-responsive prodrug-based nanomedicines [@bib0008].\n\nFor example, Zhang et al. designed an engineered polymeric nanomedicine that loaded 59\u2009% of the GSH\u2010responsive heterodimeric multifunctional prodrugs (HDMP) [@bib0036]. The CPT conjugated with a photosensitizer of 2\u2010(1\u2010hexyloxyethyl)\u20102\u2010devinyl pyropheophorbide\u2010\u03b1 (HPPH) through a GSH\u2010sensitive disulfide bond to form the HDMP. The obtained HDMP then co-assembly with degradable and biocompatible poly(ethylene glycol)\u2010*block*\u2010poly(d,l\u2010lactic acid) (PEG\u2010*b*\u2010PLA) to compose HDMP\u2010loaded NPs **(**[Fig. 1](#fig0002){ref-type=\"fig\"}d). The in vitro drugs release result demonstrated the CPT concentration released from CPT\u2010ss\u2010HPPH NPs increased over time **(**[Fig. 1](#fig0002){ref-type=\"fig\"}e). The in vivo results suggested that the HDMP NPs exhibited significant synergistic therapeutic efficiency **(**[Fig. 1](#fig0002){ref-type=\"fig\"}f**)**. In addition, Ling et al. fabricated the glutathione (GSH) sensitive self-assembled prodrug-based nanoparticles (NPs) that are composed of amphiphilic lipid-polyethylene glycol (PEG) and Pt(IV) prodrugs [@bib0011]. These prodrug NPs also exhibited significant glutathione responsiveness and good tumor inhibition rate.\n\n### 3.1.3. ROS-responsive {#sec0007}\n\nReactive oxygen species (ROS, e.g., hydroxyl radicals, H~2~O~2~, and superoxide), are a class of oxidative molecules produced in cells and involved in many biological processes. An appropriate concentration of ROS act as signaling molecules in many metabolic pathways, while excessive ROS are harmful to tissues and result in a series of diseases such as cancer. To be more specific, ROS act as signaling molecules that play crucial roles in protein translation, transcription, survival, tumorigenesis, and proliferation [@bib0037]. It is generally known that cancer cells in tumor tissues are immersed in intrinsic oxidative stress with a relatively higher level of H~2~O~2~ compared with healthy tissues, which plays a vital role in tumor cells proliferation and tumor progression [@bib0037]. Meanwhile, the oxidative stress inherent in the tumor tissue also promotes the increase of reactive oxygen species in the tumor. Due to the high ROS level in TME, many types of ROS-responsive nanomedicines have been developed.\n\nFor instance, Xu et al. developed a new ROS-triggered nanomedicine through copolymerizing clinical anticancer drug mitoxantrone (MTO, a NIR fluorescence drug) with a ROS-susceptible thioketal-containing linker **(**[Fig. 1](#fig0002){ref-type=\"fig\"}g**)** [@bib0010]. The prepared polyprodrugs were then self-assembled with lipid-PEG to form polyprodrug NPs. Internalizing RGD (iRGD) peptide was then employed to modify the polyprodrug NPs for targeting \u03b1v integrins on tumor endothelium. The in vitro results demonstrated that there was a dramatically improved and sustained drug release in the presence of ROS **(**[Fig. 1](#fig0002){ref-type=\"fig\"}h**)**. The in vivo experiment suggested the prominent therapeutic efficacy of polyMTO NPs, and the iRGD modified NPs had better curative effects than other groups **(**[Fig. 1](#fig0002){ref-type=\"fig\"}i**)**. Additionally, Li et al. utilized copolymerized monomers CPT and PEG as well as piperidine-modified methacrylate (P(CPTMA-co-PEMA)) to load glucose oxidase (GOD) to construct a prodrug-based polymersome nanoreactor **(**[Fig. 2](#fig0003){ref-type=\"fig\"}a**)**. The obtained nanoreactor displayed H~2~O~2~-responsive CPT release behavior and superior tumor inhibition efficiency [@bib0020].Fig. 2(a) Schematic for the self-assembly process of GOD\\@PCPT-NR and its oxidation/chemotherapy for anticancer through tumor acidity-responsive activation, in situ H~2~O~2~ production, and active CPT drug release. Reproduced with permission\\[20\\]. Copyright 2017, American Chemical Society. (b) Schematic illustration of the cholinesterase-responsive SCD/QA-Cbl drug delivery system. (d) Percentage of Cbl released from the SCD/QA-Cbl assembly over time, with/without BChE. (e) Percentage of A549 tumor cells death in SCD/QA-Cbl assembly, with/without BChE. Reproduced with permission\\[13\\]. Copyright 2019, Royal Society of Chemistry. (e) Schematic of molecular mechanism of SPNpd for hypoxia-activated synergistic PDT and chemotherapy. (f) HPLC profiles of NADH, IPM-Br and the final solution of SPNpd incubated with NADH under hyoxia condition for 6\u00a0h. (g) Tumor volume changes of mice with different treatment. Reproduced with permission\\[42\\]. Copyright 2019, Wiley-VCH. (h) Schematic diagram for the structure of PEGylated multistimuli-responsive dendritic prodrug nanomedicine and its enhanced antitumor activity. (i) SEM images of the PEGylated dendritic polymeric prodrug nanomedicine after stored in acidic solution (C-1), acidic solution with GSH (C-2), and acidic solution with GSH/cathepsin B (C-3), respectively, for 24\u00a0h. (j) In vitro DOX release from the PEGylated multistimuli-responsive dendritic copolymer\u2212DOX-based nanomedicine activated in different conditions. (k) Relative tumor volume of tumor-bearing mice after different treatments. Reproduced with permission\\[21\\]. Copyright 2018, American Chemical Society.Fig 2\n\n### 3.1.4. Enzyme-responsive {#sec0008}\n\nEnzymes are crucial targets for drug development due to their vital roles in cell regulation. The drugs can be delivered to the lesion site programmatically through enzyme-activatable carrier when the target site with higher concentrations of enzymes or the enzymatic activity is related to a specific tissue [@bib0038]. Owing to the importance of enzymes in tumor metabolism, tumor-associated enzyme dysregulations have recently aroused attention in therapeutics. Using enzymes as a trigger has many favorable features in designing nanomedicine that can lead to numerous biochemical reactions under mild conditions nanomedicine [@bib0038]. Also, the exclusive selectivity of enzymes, allows sophisticated, specific, and biologically inspired reactions to occur [@bib0039]. Hence, many enzyme-cleavable chemical bonds were introduced into nanomedicines as targeted responses, especially for prodrug-based nanomedicines.\n\nFor example, Li et al. designed a smart Janus PEGylated dendrimer-PTX prodrug-based NPs, which was composed of PTX conjugated PEGlated peptide dendrimer and an enzyme cleavable linker oligopeptide glycyl phenylalanyl acylglycine showed a quickly release of PTX and significant tumor therapeutic effect in presence of enzyme [@bib0040]. Besides, Guan et al. reported a smart enzyme-responsive supramolecular nanoplatform, composed of biocompatible sulfato-b-cyclodextrin (SCD) and a water-soluble prodrug of choline modified anticancer drug chlorambucil (QA-Cbl) *via* electrostatic interactions **(**[Fig. 2](#fig0003){ref-type=\"fig\"}b**)** [@bib0013]. In this system, the butyrylcholinesterase (BChE) can cleave QA-Cbl into anti-cancer drug chlorambucil (Cbl) and choline due to the presence of a cleavable ester bond of QA-Cbl prodrug **(**[Fig. 2](#fig0003){ref-type=\"fig\"}c**)**, resulting in well therapy performance against cancer cells **(**[Fig. 2](#fig0003){ref-type=\"fig\"}d**)**.\n\n### 3.1.5. Hypoxia-responsive {#sec0009}\n\nOwing to malformed tumor blood vessel progression and irregular tumor cells proliferation, hypoxia has been observed to emerge in solid tumors. As one of the main features of solid tumors, hypoxia is closely related to tumor metastasis, invasion, and drug resistance. Considering its critical roles in tumor angiogenesis, tumor progression, and cancer metastasis, hypoxia has been identified as a primary stimulus for cancer diagnosis and treatment [@bib0008], especially for prodrug-based nanomedicines. In addition, there are mainly three representative classes of hypoxia-responsive moieties, including nitrobenzoyl alcohols, nitroimidazoles, and azo linkers. In general, these hypoxia-responsive moieties can accept electrons in hypoxic conditions, which would generate hydrophilic functional groups and further alter their physicochemical properties such as particle size and hydrophilicity [@bib0041].\n\nRecently, Cui et al. constructed a semiconducting polymer nanoprodrug (SPNpd) for hypoxia triggered synergistic oncotherapy [@bib0042]. The photosentisizer SPN core was grafted with PEG to form an amphiphilic polymer brush and then conjugated with the chemodrug side chains (bromoisophosphoramide mustard intermediate, IPM-Br) *via* hypoxia-cleavable linkers and further self-assembly to form SPNpd **(**[Fig. 2](#fig0003){ref-type=\"fig\"}e**)**. The obtained SPNpd possessed the features of generating singlet oxygen (^1^O~2~) under NIR irradiation that specifically tumor hypoxia-activatable drug release **(**[Fig. 2](#fig0003){ref-type=\"fig\"}f**)**. Because of these characteristics, SPNpd could exert synergistic chemotherapy and PDT, and effectively inhibits tumor growth even in the hypoxic conditions **(**[Fig. 2](#fig0003){ref-type=\"fig\"}g**)**. Also, Hua et al. designed a novel hypoxia-responsive angiopep-2-lipid-poly(MIs)n (ALP-(MIs)n) polyprodrug NP with hypoxic radiosensitization effects for targeted glioma therapy [@bib0043]. The ALP-(MIs)n polyprodrug NP comprised of P-(MIs)n (*n*\u00a0=\u00a025, 48), DSPE-PEG2000, angiopep-2-DSPE-PEG2000, and lecithin was prepared *via* nanoprecipitation process. The DOX was co-loaded into ALP-(MIs)n polyprodrug NP to achieve chemotherapy and radiation synergistic therapy. The obtained ALP-(MIs)n/DOX was disassembled and disordered to release DOX in the hypoxic conditions, and showed significant inhibition of glioma tumor growth with the assitance of RT.\n\n### 3.1.6. Multi-stimuli-responsive {#sec0010}\n\nGiven the complexity of TME, a combination of two or multiple stimuli in one nanoplatform can provide additional opportunities to maximize the therapy efficiency. Various endogenous stimuli, including low pH, GSH, ROS, enzymes, and hypoxia, coexist in the TME, as mentioned above, which opens up the possibility of designing a complex prodrug-based nanomedicine. Meanwhile, given the variations in the physiological TME, the multi-stimuli responsive nanomedicines can fully exploit the characteristics of the TME to enhance therapeutic accuracy. Thus, some multi-stimuli responsive prodrug-based nanomedicines have been developed in recent years\n\nFor example, Duan et al. developed a prodrug-based nanomedicine of PEGylated multistimuli-responsive dendritic copolymer coupled with DOX [@bib0021]. The dendritic polymers mainly consisted of poly \\[N-2-hydroxypropyl\\] methacrylamide (polyHPMA) segments and enzyme-responsive linkers of GFLG (Gly-Phe-Leu-Gly-tetrapeptide). Then the PEGylated (*via* the disulfide bond) dendritic polymers were further linked to DOX through hydrazine bonds **(**[Fig. 2](#fig0003){ref-type=\"fig\"}h**)**. The proposed dendritic nanomedicines could respond to pH, GSH, and enzyme due to the characteristics of the bonds in the particles. For instance, the GFLG crosslinkers for enzymes-responsiveness, disulfide bonds for GSH-responsiveness, and hydrazine bones for acidity-responsiveness **(**[Fig. 2](#fig0003){ref-type=\"fig\"}i and j**)**. Meanwhile, the dendritic polymers nanomedicines showed superior anti-tumor activity and tumor accumulation performance **(**[Fig. 2](#fig0003){ref-type=\"fig\"}k**)**. Sun et al. reported ROS and GSH sensitive paclitaxel (PTX) prodrug-based nanoassemblies [@bib0044]. PTX was conjugated to citronellol (CIT) *via* disulfide bonds (SS) and further PEGylated. The obtained nanoassemblies showed redox-responsive drug-release behaviours and tumor remarkable growth inhibition performance.\n\n3.2. Exogenous stimuli {#sec0011}\n----------------------\n\n### 3.2.1. Light-responsive {#sec0012}\n\nBecause of non-invasiveness, inexpensiveness, and practicability, light has aroused tremendous interest as an exogenous stimulus for prodrug-based nanomedicines. The external photoactivatable prodrug-based nanomedicines have many advantages over other internal stimuli owing to ease of handling, precise control of the time and location of treatment [@bib0045]. Various light-responsive smart prodrug-based nanosystems utilizing ultraviolet (UV), visible light, and near-infrared (NIR) have been intensively applied for non-invasive and controlled on-demand drug release against cancer \\[[@bib0046], [@bib0047], [@bib0048]\\].\n\nSince UV light (300--380\u00a0nm) can cause phase transitions in some polymers with special structures (such as O-nitrobenzyl, pyrene, spiropyran, and azobenzene), it is widely used in prodrug release \\[[@bib0048],[@bib0049], [@bib0050], [@bib0051]\\]. For instance, Liu\\'s group used a photocaged linkage to form prodrugs-based shell cross-linked (SCL) which were composed of P(CL-g-CPT)-b-P(OEGMA-co-MAEBA)-CPT and PCL-b-P(OEGMA-co-MAEBA-co-FA) amphiphilic diblock copolymers **(**[Fig. 3](#fig0004){ref-type=\"fig\"}a**)** [@bib0052]. This type of amphiphilic diblock copolymers micelles possessed a hydrophobic core conjugated with photocaged CPT prodrug. The cores of SCL micelles contained a lot of CPT drugs, which can be effectively cleaved to release drugs under UV irradiations **(**[Fig. 3](#fig0004){ref-type=\"fig\"}b**)**. The in vitro results also revealed that SCL micelles exhibited at least \\~9.7-fold enhanced cytotoxicity when introducing the UV light **(**[Fig. 3](#fig0004){ref-type=\"fig\"}c**)**. Although many UV-responsive prodrug-based platforms have been developed, the limited permeability and potential hazards of UV light terribly inhibited the potential clinic translations.Fig. 3(a) Schematic illustration of the fabrication of SCL micelles from P(CL-g-CPT)-b-P(OEGMA-co-MAEBA)-CPT and PCL-b-P(OEGMA-co-MAEBA-co-FA) amphiphilic diblock copolymers. (b) In vitro CPT release curves of SCL micelles (\u25cb) with and (\u25a1) without UV irradiation; (c) Cell viability of HepG2 cells and (e) A549 cells administered with a) control, b) UV irradiation, c) FA-decorated SCL micelles, d) FA-free SCL micelles, e) FA-decorated SCL micelles\u00a0+\u00a0UV irradiation, f) FA-free SCL micelles\u00a0+\u00a0UV irradiation, g) CPT parent drug, and h) CPT parent drug\u00a0+\u00a0UV irradiation. Reproduced with permission\\[52\\]. Copyright 2013, American Chemical Society. (d) Schematic diagram of antitumor mechanism of photo-responsive Pt(IV)--azide complexes, and self-assembly as well as light-triggered dissociation of an SPCA prodrug-backboned BCP micelle. (e) Platinum release curves of BCP micelles with intermittent light irradiation. (f) Tumor growth inhibition curves of A549 bearing nude mice after different treatments. Reproduced with permission\\[53\\]. Copyright 2016, Wiley-VCH. (g) Chemical structure of the PEGylated CPE covalently linked with anticancer drug DOX via a UV-cleavable linker CPE--DOX and schematic diagram of NIR laser regulated initiation of the photosensitizer for photodynamic therapy and on-demand drug release for chemotherapy. Reproduced with permission\\[56\\]. Copyright 2014, Royal Society of Chemistry. (h) Schematic diagram of HNP preparation and possible mechanism for metastatic breast cancer therapy. (i) In vitro release curve at 42 or 37\u00a0\u00b0C. (j) Tumor volume changes of 4T1 tumor-bearing mice. Reproduced with permission\\[31\\]. Copyright 2018, American Chemical Society.Fig 3\n\nCompared to UV light, visible light has a longer wavelength, which makes it more permeable to tissues and safer to clinical applications. Therefore, visible light is often used as a stimulant to programmatically regulate drug release. Zhou et al. conjugated the trans- \\[Pt(N~3~)~2~(OH)~2~(py)~2~\\] complex with PEG and polyurethane to form a prodrug-backboned block copolymer (BCP) micelles **(**[Fig. 3](#fig0004){ref-type=\"fig\"}d**)** [@bib0053]. The light with a wide range of wavelengths (from 365 to 500\u00a0nm) can activate the BCP micelles to release drugs. Meanwhile, BCPs were inactivated and kept in the silenced Pt(IV) prodrug state without light irradiation **(**[Fig. 3](#fig0004){ref-type=\"fig\"}e**)**. When irradiated with light, the chain was cleaved, then BCP micelles disassociated, and the Pt(IV) prodrugs were activated. Furthermore, the BCP micelles with good biocompatibility displayed significant antitumor activity **(**[Fig. 3](#fig0004){ref-type=\"fig\"}f).\n\nThe limitations of UV or visible light-triggered prodrug-based nanomedicines mainly contains low penetration depth (\\~10\u00a0mm), which prevents them from activating nanomedicines inside the tumor, resulting in poor therapeutic efficacy, therefore an urgent need to introduce NIR (700--1000\u00a0nm) light \\[[@bib0054],[@bib0055]\\]. For instance, upconversion nanoparticles (UCNPs) were ideal nanoplatforms for NIR-responsive drug delivery. Yuan et al. combined the polyelectrolyte (CPE)--drug conjugate and UCNPs to form NIR-responsive prodrug-based nanomedicines **(**[Fig. 3](#fig0004){ref-type=\"fig\"}g**)** [@bib0056]. The main component of this nanomedicine was DOX conjugated PEGylated CPE photosensitizer, and a UV-cleavable ortho-nitrobenzyl (NB) linker as well as UCNPs. In the presence of a 980\u00a0nm laser, the UCNPs emitted UV and visible light to activate the polymer photosensitizer to generate ROS and release DOX for suppression of U87-MG cell growth.\n\n### 3.2.2. Thermo responsive {#sec0013}\n\nIn addition to the above stimuli, heat can act as an exogenous stimulus for controlled drugs release *via* thermo-sensitive polymer-based nanoparticles, liposomes or polymer micelles [@bib0057]. Thermo-sensitive nanocarriers are formulated based on heat-responsive polymers with drug conjugation and release drugs at a critical temperature point [@bib0058]. Commonly, thermo-sensitive polymeric prodrugs undergo a change in solubility behavior at the transition temperature, mainly include the lower and upper critical solution temperature (LCST and UCST), which are the temperature for more and less soluble transition states, respectively. In the light of this, a variety of thermoresponsive prodrugs systems has been devised, including liposomes, polymeric micelles, or nanogels \\[[@bib0059],[@bib0060]\\].\n\nLv et al. reported a nanoplatform consisting of marimastat (MATT)-loaded thermosensitive liposomes (LTSLs) (MATT-LTSLs) and hyaluronic acid (HA)-paclitaxel (PTX) (HA-PTX) prodrugs **(**[Fig. 3](#fig0004){ref-type=\"fig\"}h**)** [@bib0031]. At 42\u00a0\u00b0C, near the phase transition temperature of LTSLs, HA-PTX/MATT-LTSLs HNPs released their payloads at a dramatically faster rate than that at 37\u00a0\u00b0C **(**[Fig. 3](#fig0004){ref-type=\"fig\"}i**)**. Tumor growth, metastasis, and angiogenesis (10-fold) were significantly inhibited after the injection of HNPs with mild hyperthermia addition **(**[Fig. 3](#fig0004){ref-type=\"fig\"}j**)**. The thermosensitive gel is also an attractive carrier to deliver prodrugs. For example, Peng et al. synthesized thermal and redox dual responsive biodegradable prodrug nanogels, which were composed of N-vinylcaprolactam (VCL) and N-succinimidyl methacrylate (Suma) crosslinked with diallyl disulfide **(**[Fig. 4](#fig0005){ref-type=\"fig\"}a**)** [@bib0061]. Good thermally controlled drug release behavior and significant antitumor performance toward HeLa cells were observed for this nanogel.Fig. 4(a) Scheme of synthesis of prodrug nanogel and intracellular release mechanism of DOX. Reproduced with permission\\[61\\]. Copyright 2018, American Chemical Society. (b) Construction of PFP-TNDs and the mechanisms of US-triggered imaging as well as delivery of cis-aconityl doxorubicin (CAD). Reproduced with permission\\[63\\]. Copyright 2019, Elsevier. (c) Schematic diagram of US combined with DPMC to deliver DOX into cell nuclei and induce cytotoxicity. (d) CLSM images and bright field images of DPMC pre-destruction and post-destruction by US. (i) In vivo tumor growth suppression of DPMC with/without US. Reproduced with permission\\[64\\]. Copyright 2017, Lvyspring. (f) Schematic illustration of RAW264.7 cell delivered thermochemotherapy. Reproduced with permission\\[67\\]. Copyright 2012, Beilstein. (g) T2-weighted images of the tumors obtained at 2\u00a0h-postinjection of USPIO/SQgemNPs \u2460 without an external magnetic field (MF) (mouse 2) and \u2461 with an external MF (1.1 T) (mouse 3). \u2462 Percentage of the hypo-intensity tissues with T2\u00a0\\<\u00a036\u00a0ms (white column), and with T2\u00a0\\<\u00a020\u00a0ms (gray column). Mouse 2 was injected with nanocomposites without exposure to MF. Mice 1, 3, and 4 were injected with USPIO/SQgem NPs with exposure to 1.1 T MF for 2\u00a0h. (g) In vivo anticancer activity of USPIO/SQgem NPs (with 1.1 T MF) compared with placebo-treated group (drug unloaded USPIO/squalene nanocomposites), USPIO/SQgem NPs (without MF), with SQgem NPs, and with free gemcitabine in L1210 tumor bearing mice. Untreated (\u25cf), placebo USPIO/squalene NPs (\u25cb), gemcitabine (\u25ca), SQgem NPs (\u25b2), USPIO/SQgem composite NPs (no MF) (\u0394), USPIO/SQgem composite NPs (with 1.1 T MF) (\u25a0). Reproduced with permission\\[18\\]. Copyright 2011, American Chemical Society.Fig 4\n\n### 3.2.3. Ultrasound-responsive {#sec0014}\n\nSonication is considered as a non-invasive, generally available, relatively inexpensive and portable technique that can penetrate deeply and accurately into the tissues. Recently, it has been used extensively in the biomedical field, such as imaging-guided drugs and gene delivery. Ultrasound (US) can trigger the release of the drugs through the thermal and mechanical generated from cavitation phenomena [@bib0062]. For example, Gao et al. developed an ultrasound activated prodrug delivery nanodroplet of PFP/C9F17-PAsp(DET)/CAD/PGA-g-mPEG, which was composed of acid-cleavable DOX prodrug, cationic amphiphilic fluorinated polymer carrier, and US responsive phase-change contrast agent aiming to achieve optimized US imaging, cellular uptake and antitumor therapeutic effect **(**[Fig. 4](#fig0005){ref-type=\"fig\"}b**)** [@bib0063]. Furthermore, Luo et al. developed an US-triggered and pH-sensitive DOX prodrug-microbubble complex (DPMC) **(**[Fig. 4](#fig0005){ref-type=\"fig\"}c**)** [@bib0064]. The DOX prodrug (DP) was composed of succinylated-heparin conjugated with DOX through hydrazone linker and modified with dual targeting ligands cRGD peptide and folate, and further combined with microbubble (MB) through an avidin-biotin bridge. The obtained DPMC possessed enhanced therapeutic efficiency with the assistant of US cavitation and sonoporation. Notably, after US treatment, DPMC was destroyed, and the aggregated DP dispersed into small uniform nanoparticles, thereby releasing DOX **(**[Fig. 4](#fig0005){ref-type=\"fig\"}d**)**. In particular, the antitumor ability of DPMC with US irradiation in vivo was dramatically promoted **(**[Fig. 4](#fig0005){ref-type=\"fig\"}e**)**.\n\n### 3.2.4. Magnetism and electric field - responsive {#sec0015}\n\nMagnetic responsive prodrugs are either fabricated by entrapment of inorganic magnetic nanocrystals within polymeric scaffold or by covalent immobilization [@bib0065]. So far, two different strategies are using for polymer contraction/shrinkage: alternating magnetic (AMF) field on magnetic nanoparticles (MNPs) to generate thermal to release prodrugs and introducing the static magnetic field to generate mechanical force to release the prodrugs [@bib0066]. For example, Wang et al. developed a magnetic heat responsive nanoplatform that was composed of a core/shell Fe/Fe~3~O~4~ MNPs and the prodrugs of SN38 (a topoisomerase I blocker) **(**[Fig. 4](#fig0005){ref-type=\"fig\"}f**)** [@bib0067], showing accelerated SN38 release under action of the Tet-On Advanced system.\n\nAdditionally, utilizing a magnetic field for local area targeting would dramatically enhance the drug accumulation in tumors, and promoted the diagnosis (magnetic guided imaging) accuracy [@bib0068]. Hence, magnetically responsive systems provide a precise selection of prodrug delivery. Above all, the magnetic-guided targeting concept has proven great potential in cancer therapy. Recently, Arias et al. designed a novel theranostic nanomedicine with the ability to target the delivery of the gemcitabine prodrug under the assistant of the magnetic field, monitored by the MRI imaging method as well **(**[Fig. 4](#fig0005){ref-type=\"fig\"}g**)** [@bib0018]. This multifunctional nanomedicine was prepared by embedding magnetite nanoparticles into a lipophilic self-assembling bioconjugate of squalenoyl gemcitabine (SQgem). The obtained nanomedicine had high drug loading efficiency about 93%, prominent magnetic susceptibility, and low burst release. Furthermore, magnetite/SQgem nanocomposites with magnetic field assistance showed considerably greater antitumor ability than the other groups **(**[Fig. 4](#fig0005){ref-type=\"fig\"}h**)**.\n\nFurthermore, the electric field is another emerging method to trigger the release of prodrugs. For example, J. Norman et al. presented a new concept of electroceuticals, summarized as the electrochemical activation of metal-based prodrugs [@bib0069]. However, due to the limitations of the electric field applied in vivo, there is little research on this part. It is believed that with the development of electrochemical activation prodrugs, the research on electric field-activated prodrugs-based nanomedicines will gradually emerge.\n\n4. Applications in clinical translation {#sec0016}\n=======================================\n\nThe history of the prodrugs-based nanomedicines composed of drugs and polymers is ancient and promotes the development of advanced stimuli-responsive drug release systems \\[[@bib0070],[@bib0071]\\]. So far, some polymeric prodrugs have been approved as promising therapeutic nanoagents that already entered in clinical trial phase I/II as given in [Table 1](#tbl0001){ref-type=\"table\"} \\[[@bib0072],[@bib0073]\\]. All these promising clinical prodrug-based nanomedicines are very easy to be fabricated. Contradictory, most of the developed intelligent prodrug-based nanomedicines were designed with sophisticated structures and formulations in order to obtain smart properties, which is challenging to scale up for industrial productions. Therefore, simplifying the structures of prodrug-based nanomedicines is still one of the critical points for successful clinic translation.Table 1Current status of some polymeric prodrugs in cancer therapy.Table 1Coupled drugsPolymersClinical TrialsCompanyAspartic acidPEGINCICPTPEGIIEnzonCPTPoly (glutamate)IIICell TherapeuticsCPTPoly N-(2-Hydroxypropyl) methacrylamide (P-HPMA)IPharmaciaCisplatinPEG-b-p(Glu)IIINanoCarrier/Orient EuroPharmaDOXP-HPMAIIPharmaciaDOX micelle and PlatinateP-HPMAII/IIIAccess PharmaDOX-GalactosamineP-HPMAIIPharmaciaDocetaxelmPEG~5000~-b-p(HPMAm-Lac~n~)ICristal TherapeuticsDACH-PlatinumPEG-b-p(Glu)INanoCarrierEpirubicinPEG-b-p(Asp-hydrazone)INanoCarrierPTXP-HPMAIPharmaciaPTXPEGIPharmaciaPTXPEG-*b*-p(ASP-4-phenyl-1- butanol)IIINanoCarrier/Nippon Kayaku\n\nBesides, the numbers of optimizations and improvements are also needed before the clinic translation. Especially, endogenous triggers (such as pH, enzyme) are difficult to control because of considerable variation from one individual to another. While the exogenous stimuli-responsive systems are much easier to be controlled and more promising for the clinic translation, however, problems such as normal tissue damage and tissue-penetration depth should be solved [@bib0045]. Accordingly, more attention should be paid on precisely controlling the prodrug conjugation process, high batch-to-batch reproducibility, and industrial scale-up possibility. The standard production methods and stimuli dosage control may eventually speed up the translation of prodrugs-based nanomedicines from the bench to the clinic.\n\nAlthough prodrugs-based nanomedicines face many challenges in clinical application, one special prodrug of antibody-drug conjugates (ADCs) is attracting more and more attention in clinical practice. As a next-generation precision prodrug, ADCs has been envisioned as a viable approach towards a cure to various types of cancers to be used in clinic [@bib0074]. ADCs can be defined as a prodrug that utilizes the site-specific quality of a monoclonal antibody (mAb) to deliver a highly potent cytotoxic drug (warhead) via a chemically engineered biodegradable linker [@bib0074]. ADCs drugs are similar to other common prodrugs with the natural advantage of avoiding damage to normal tissue; its cytotoxic agent remains inactive during circulation in the body and releases the drug only upon entering tumor cells. Generally speaking, the mAb of ADCs can specifically bind to the targeted antigens that are typically highly expressed on the surface of cancer cells to achieve cancer therapy. Due to high stability, selectivity, and favorable pharmacokinetic characteristics of ADCs nanomedicines, many of these are applied in the clinic. Some clinic data is also presented for the researcher: for example, a famous clinical ADCs approved in 2000, gemtuzumab ozogamicin, composed of an anti-CD33 mAb and calicheamicin [@bib0075]. This drug was aimed to be used on patients older than 60 years old diagnosed with relapsed acute myeloid leukemia, whose response rate was about 30%. Furthermore, the approved dosage of this drug is very low, which is about 9\u00a0mg/m^2^ or 0.22\u00a0mg/kg for an adult (65\u00a0kg), reflecting the high efficacy of calicheamicin.\n\nAs another representative Food and Drug Administration (FDA) approved ADCs drug, trastuzumab emtansine (T-DM1); which was formed by conjugation of anti-HER2 mAb trastuzumab with maytansinoid DM1 via linker for metastatic breast cancer therapy [@bib0075]. A clinical trial demonstrated that about a 33% response rate was observed after 110 patients with third-line metastatic breast cancer received the treatment of T-DM1 with the dosed of 3.6\u00a0mg/kg every three weeks, confirming the activity of T-DM1 in treatment-resistant HER2 positive metastatic breast cancer.\n\nIn addition to these drugs, by 2019 five other ADCs drugs: brentuximab vedotin(response rate of 54%) [@bib0075], ado-trastuzumab emtansine (response rate of 43.6%) [@bib0076], inotuzumab ozogamicin (response rate of 57%) [@bib0077], polatuzumab vedotin-piiq (response rate of 89%) [@bib0078], enfortumab vedotin (response rate of 44%) [@bib0079], have been approved by the FDA and are currently widely applied to treat cancer patients. With the rapid development of ADCs, according to incomplete statistics, there are currently about six hundred types of ADCs in clinical trials, and over 80 ADCs expected to be approved over the next few years [@bib0080]. Predictably, ADCs will be the main force of prodrug in clinical application and will play an increasingly important role in treating cancer. Meanwhile, we firmly believe that nanomedicines based on ADCs prodrugs will also step onto the clinical stage one day.\n\n5. Conclusion and perspectives: challenges and outlook {#sec0017}\n======================================================\n\nDespite the above-mentioned stimuli-responsive prodrugs-based nanomedicines are available for clinical applications. The potential toxicity, accumulation in the body, and biodegradability of the nanomedicines should be considered. For instance, acrylic acid-derived polymers cannot degrade in the aqueous phase, thus leads to the possible neurotoxicity [@bib0081]. Furthermore, the insufficient excretion of macromolecular conjugates leads to the accumulation in the body and harms the kidney [@bib0082]. Though stimuli-responsive prodrugs-based nanomedicines exhibited the efficient drug release ability after internalization in tumor or responded to TME. However, the drug-conjugates are also susceptible to non-specific and slow drug release in non-cancerous cells \\[[@bib0083],[@bib0084]\\].\n\nSo far, basic research has been well-explored for achieving the \"clinical prospects,\" but still, many of them might never enter into the clinic practices. Various probable obstacles have been portrayed as: Firstly, nanocarriers should be smart enough to cross biological barriers. Unfortunately, numerous of the nanomedicines is injected intravenously into the body, which would face a series of complex biological barriers that significantly limit the site-specific targeting. Due to the presence of opsonization, the mononuclear phagocyte system (MPS), cellular internalization, enormous intratumoral pressure, escaping from endosomal and lysosomal sections, as well as drug efflux pumps are the main obstacles to cross the biological barriers \\[[@bib0085], [@bib0086], [@bib0087]\\]. Furthermore, administration routes, disease types, and progression are also substantial challenges while designing smart prodrug delivery systems. Secondly, pharmaceutical manufacturing procedures should be safe, effective, and quality controlled. Many questions need to be addressed, such as biosafety, biodegradability, and metabolites of the nanomaterials. Moreover, compared with conventional drug delivery systems, prodrug-based nanomedicine may be more susceptible to inter-subject variability, particularly in therapy performance, therefore raising concerns with reproducibility in efficiency and safety. Thirdly, batch-to-batch reproducibility, scale-up validation, and controllable physicochemical properties are critical concerns for clinical interpretations. Authentication of the procedures should be confirmed for the designing and practices of prodrugs nanocarriers. For instance, stability, specificity, precision, and repeatability. In short, entirely, these issues argue the clinical applications of stimulus-sensitive drug carriers. Finally, to have a sound effect for nanotechnology to be incorporated with the pharmaceutical industry, it is worthy of designing a clear regulatory agenda to support novel nanomedicine products.\n\nCombining the polymers with an active functional group (hydroxyl, carboxyl or halogen) and drug molecules to form the polymeric prodrugs is the most common strategy for constructing prodrugs-based nanomedicines. Meanwhile, the polymer with the appropriate stimuli-responsive linker is a critical factor in the construction of the required prodrug. Although the field of polymeric prodrug design principles is well defined, it is still difficult to select the appropriate linker to meet the design requirements with keeping the activity of the drug. Among them, the major challenge is to understand the relationship between efficient conjugates design and biological features such as tumor heterogeneity, microenvironment, and metastasis. The progress brings in the designing of more efficient bio-conjugates for combination therapy using polymeric prodrugs with degradable long-circulating conjugates. Moreover, the implementation of new targeting strategies is also vital by using different bio-ligands in combination with other imaging techniques. The new targeting strategies with the more advantageous features may include (1) novel polymeric conjugates design for directly stem cells targeting; (2) modulation of polymeric conjugates to therapeutic vaccines; (3) subcellular targeted polymeric conjugates; (4) designing of polymeric conjugates as lysosomotropic nanomedicines; (5) antibodies-polymeric prodrugs conjugation for cancer therapy; (6) polymeric conjugates as anti-angiogenic agents; (7) development of biochemical, optical and chemical sensors based on polymeric conjugates; (8) design anticancer nanomedicine using polymeric conjugates; (9) genetically and hereditary disorders therapies *via* synthetic genetic polymeric conjugates. Finally, a strong belief is that prodrug-based nanomedicines incorporated with the interdisciplinary approach of sciences have great potential in the clinical treatment of cancer.\n\n6. Outstanding questions {#sec0018}\n========================\n\nThis review presents the current development and recent progress in stimuli-responsive prodrug-based nanomedicines. Understanding various stimuli-responsive pathways of prodrug-based nanomedicines could promote the design optimization of related drugs and advance their clinical application prospects. Several prodrug based nanomedicines have entered the clinical stage and achieved promising clinical results, but more future studies about pharmaceutical and pharmacokinetic, solubility, stability, biosafety, biocompatibility, and biodegradability of prodrug-based nanomedicines still need to be explored, which would further improve their feasibility, thereby obtaining the optimization of prodrug-based nanomedicines in clinical application. This timely review article provides deep insights into the development and challenge of stimuli-responsive prodrug-based nanomedicines for cancer therapy, which will help the researcher to understand the mechanism of prodrug-based nanomedicines effectively and establish a solid foundation for their clinical applications.\n\n7. Search strategy and selection criteria {#sec0019}\n=========================================\n\nResearch data of this review were searched by Google Scholar, PubMed, and Web of Science using the search terms nanomedicine, prodrug, stimuli-responsive, cilinic application, and ADCs to obtain relevant articles. Information was included when related directly to the relationship between nanomedicine, prodrugs and stimuli-responsive. Only articles published in English between 1995 and 2019 were included, and mostly articles are between 2015 and 2019.\n\nDeclaration of competing Interests\n==================================\n\nThe authors declare no conflict of interest.\n\nAcknowledgements {#sec0020}\n================\n\nThis work is supported by the US METAvivor Early Career Investigator Award (W.T.), and Harvard Medical School/Brigham and Women\\'s Hospital Department of Anesthesiology-Basic Scientist Grant (W.T.). T.W. is also a recipient of the Khoury Innovation Award and American Heart Association (AHA) Collaborative Sciences Award.\n\nAuthor\\'s contributions {#sec0021}\n=======================\n\nJ.O., B.S., and W.T. conceived the review out from our recently published work. A.X., H.S., and O.J presented the outline of this review. A.X., H.S., O.J., and Z.T. collected literature as well as wrote the review. N.K., N.Y.K., B.Q., D.P., B.S. and W.T. revised the manuscript, provided some relevant insights, and made some edits. All authors read and approved the final version of the manuscript.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "![](glasgowmedj75267-0070){#sp1 .356}\n\n![](glasgowmedj75267-0071){#sp2 .357}\n\n![](glasgowmedj75267-0072){#sp3 .358}\n\n![](glasgowmedj75267-0073){#sp4 .359}\n\n![](glasgowmedj75267-0074){#sp5 .360}\n"}
+{"text": "1.. Introduction {#s1}\n================\n\nThe exquisite sensitivity of animal sensory organs has been noted many times \\[[@RSIF20190049C1]--[@RSIF20190049C6]\\]. However, little attention has been paid to the mechanical properties that shape a sensor\\'s response as much, if not more, as the neuronal filters \\[[@RSIF20190049C7]\\]. The mosquito antenna is a well-known example of a highly sensitive particle-velocity receptor \\[[@RSIF20190049C2],[@RSIF20190049C3],[@RSIF20190049C8],[@RSIF20190049C9]\\], and in many species, the key function of the antenna is to locate the flying conspecific mate \\[[@RSIF20190049C2],[@RSIF20190049C10]--[@RSIF20190049C12]\\]. Different models on how they achieve complex mechanical behaviour, for example, active amplification, have been proposed and are reviewed by Mhatre \\[[@RSIF20190049C13]\\]. While knowledge of mechanical behaviour, for some sensory organs, has increased in the last decade \\[[@RSIF20190049C7]\\], rather little is known about the material composition and properties underlying these complex behaviours in terms of the geometry-defining distribution of stresses and strains within the sensor \\[[@RSIF20190049C7]\\].\n\nThe mosquito antenna comprises three parts---scape, pedicel and flagellum \\[[@RSIF20190049C1]\\] ([figure\u00a01](#RSIF20190049F1){ref-type=\"fig\"}). Control of the antennal direction is done in part by the scape, but the scape is not relevant to the present study. The pedicel houses some 16 000 sensory neurons \\[[@RSIF20190049C14]\\], the majority of which are used for acoustic detection. These neurons connect to radially distributed prongs, attaching the neurons to the base of the flagellum. The flagellum itself is the physical sensor---it consists of 13 sequential flagellomeres that project distally. These act as viscosity sensors, undergoing oscillatory displacement in the presence of acoustic fluid flow. The flagellum is covered in hair-like structures known as fibrillae (or setae) in which case the antenna is considered *plumose---*these fibrillae serve to increase the viscosity of the sensor improving its performance \\[[@RSIF20190049C2]\\]. It has been shown that for biologically relevant sounds, the flagellum moves like a paddle \\[[@RSIF20190049C2]\\]. Taken together, the whole system is resonantly tuned to respond maximally at the wingbeat frequencies of flying conspecifics and is the key sensor in the animal\\'s phonotactic mating behaviour \\[[@RSIF20190049C11]\\]. Figure 1.Schematic of antenna morphology. (*a*) Head of *T. brevipalpis*. (*b*) Cross section of the pedicel scale bar 0.1 mm. (*c*) Schematic of the inside of the pedicel (*b,c*) from Yack \\[[@RSIF20190049C15]\\], simplified and modified. Image abbreviations: bp, basal plate; fb, fibrillae; fl, flagellum; pd, pedicel; pdw, pedicel wall; pr, prongs; s, scape.\n\nMost research into mosquito audition focuses on the biophysical basis for mechano-electrical transduction that endows these sensors with high sensitivity and often assumes that the flagellum is simply a stiff rod in which the material properties are of negligible importance. However, we consider material properties to be an overlooked and potentially important factor in the biophysics of acoustic perception. As the resonance of the whole system will undoubtedly be determined by some contribution of the flagellum and the effective stiffness of the pedicel attachment, it is possible that the way in which the flagellum is built may be significant for resonance tuning, and thus mating behaviour. Therefore, with no *a priori* expectation on the spatial distribution of material properties, we investigated the stiffness distribution in mosquito antennae with confocal laser scanning microscopy (CLSM).\n\nThe use of CLSM for obtaining information about general types of cuticle present is well established (e.g. \\[[@RSIF20190049C16]--[@RSIF20190049C24]\\]). It enables us to visualize the insect exoskeleton using autofluorescences and to estimate material properties of structures. Michels & Gorb formulated that it is capable of estimating material properties, since well-sclerotized, flexible and resilin-dominated regions are visualized differently \\[[@RSIF20190049C16]\\]. It has also been previously cross-validated using three-point bending tests, AFM nanoindention and compression tests \\[[@RSIF20190049C17],[@RSIF20190049C22],[@RSIF20190049C23]\\]. Merits of CLSM are that the technique allows one to take sharp images of narrow sample planes by restriction of light entering the camera. This is achieved by the appropriate choice of pinhole size depending on the wavelength of the light, creating optical sections of the object. This image stack in turn then can be combined to create maximum intensity projections---in one image---showing the structure without loss of depth resolution (electronic supplementary material, table S1). One can therefore image the whole intact surface structure in great detail. In addition, the varying excitability of different unknown cuticle compounds with different laser wavelengths provides an estimation of material properties of the cuticle.\n\nIn many mosquito species, acoustic communication is essential for mating \\[[@RSIF20190049C10]--[@RSIF20190049C12],[@RSIF20190049C15]\\] and many are vectors for animal and human diseases, such as malaria, yellow fever or the zika virus \\[[@RSIF20190049C12],[@RSIF20190049C25]\\]---indicating how important it is to understand their biology.\n\nAntennae of two mosquito species of different size and ecology---*Toxorhynchites brevipalpis* and *Anopheles arabiensis*---were examined here. Many *Anopheles* species are swarming mosquitoes in which acoustics is crucial during mating \\[[@RSIF20190049C10],[@RSIF20190049C11],[@RSIF20190049C26],[@RSIF20190049C27]\\]. However, *T. brevipalpis* is a solitary non-swarming species, in which their behaviour is scarcely documented \\[[@RSIF20190049C28],[@RSIF20190049C29]\\].\n\nThere is significant interest in the *Toxorhynchites* genus for mosquito population control as its larvae predate on other mosquito larvae, many of which are species that have global importance \\[[@RSIF20190049C28],[@RSIF20190049C29]\\]*.* As this mosquito is large, robust and non-biting, it has been previously used as an amenable model system in earlier studies of insect auditory systems \\[[@RSIF20190049C8],[@RSIF20190049C9],[@RSIF20190049C26],[@RSIF20190049C30]\\]. According to Gibson & Russell \\[[@RSIF20190049C30]\\], the wingbeat frequency of males and females synchronize during aerial mating \\[[@RSIF20190049C12],[@RSIF20190049C26]\\]---and this is termed distinct flight/wingbeat frequency-matching during the mating display. As antennal ears are part of detecting the mating partner, and beam-like structures are most sensitive around their resonance, it is beneficial to have the resonance of the sensory organ close to the acoustic stimulus (the wingbeat frequency) \\[[@RSIF20190049C31]\\]. Similar acoustic behaviour has been observed in *Anopheles*, which are responsible for the spread of diseases in cattle and humans, and therefore play a substantial social and economic role \\[[@RSIF20190049C27]\\].\n\nNeither of the two species mentioned above have, to our knowledge, been studied for the potentially varying stiffness along the length of the flagellum. As we will show, CLSM proves to be very useful in demonstrating changes in material stiffness along the flagellum. From the results of the CLSM work detailed later, we used finite-element modelling (FEM) to determine the effect of the measured stiffness profiles along the antenna on a compliantly clamped beam.\n\n2.. Material and methods {#s2}\n========================\n\n2.1.. Specimen preparation {#s2a}\n--------------------------\n\nAnimals were deeply anaesthetized with CO~2~, and dissected in PBS (Carl Roth GmbH & Co KG, Karlsruhe, Germany). The antennae were treated with the surfactant Triton X-100 (Sigma-Aldrich Chemie GmbH, Steinheim, Germany), to ensure wetting of the entire surface, a necessary step as the fibrillae of plumose antennae easily trap air bubbles. Triton X-100 was washed off in triple steps with PBS. Antennae or their parts were transferred to glycerine (Carl Roth GmbH & Co. KG, Karlsruhe, Germany), which is a suitable medium as it has a similar refractive index to glass \\[[@RSIF20190049C16]\\].\n\nIn the present study, only males are included. There are three main reasons for this: (1) due to strong structural sexual dimorphisms, morphological sex comparisons are difficult. (2) While acoustics is important to both sexes (e.g. \\[[@RSIF20190049C30],[@RSIF20190049C32]\\]), the male antenna is the most studied with respect to their acoustic response, lending itself to easier comparison. (3) CLSM as relative method benefits more from comparing two separate, but structurally similar objects. Four individuals of each species were used for the present study, not all of which were CLSM-imaged, but it was confirmed the structures imaged are typical, either under the CLSM Zeiss LSM 700 (Carl Zeiss Microscopy GmbH, Jena, Germany) or a fluorescence microscope (Zeiss Axioplan).\n\n2.2.. CLSM operation {#s2b}\n--------------------\n\nFollowing an established standard method \\[[@RSIF20190049C16]\\], which is applied for many insect exoskeleton studies (e.g. \\[[@RSIF20190049C33]--[@RSIF20190049C35]\\]), antennae were visualized using autofluorescences using the same excitation wavelengths and emission filters applied in the referenced studies. Michels & Gorb \\[[@RSIF20190049C16]\\] demonstrated that the resilin autofluorescence is excitable at the wavelength of around 405 nm. Moreover with a combination of different excitation wavelengths, the method can estimate differences of material properties within a properly adjusted image or images taken with same settings, because in general well-sclerotized, flexible and resilin-dominated regions are visualized differently \\[[@RSIF20190049C16]\\]. The stack construction was performed with ZEN 2009 (6.0 SP2) (Carl Zeiss MicroImaging GmbH), which automatically corrects for wavelength-dependent slice thickness and resulting in varying overlap between stacks taken with different wavelengths.\n\nAfter adjustment of excitation level according to existing standards \\[[@RSIF20190049C16]\\], the specimens were exposed sequentially to four different excitatory wavelengths (405, 488, 555 and 639 nm), and emitted lights were filtered using a band-pass emission filter of 420--480 nm and long-pass emissions filters transmitting light with wavelengths \u2265490, \u2265560 and \u2265640 nm, respectively.\n\nEach sample was imaged only once to avoid the photobleaching which could alter the resulting visualization of the material, as different compounds are not similarly susceptible to this effect. This is important for our case, since relatively high laser power (up to max. 30%) was used in order to allow weakly fluorescent structures to be imaged. High laser power can lead to pronounced bleaching if the sample were to be imaged multiple times with the same laser \\[[@RSIF20190049C16]\\].\n\n2.3.. Image colour coding {#s2c}\n-------------------------\n\nConfocal laser scanning micrographs were colour-coded according to Michels & Gorb \\[[@RSIF20190049C16]\\]. Blue, green and two red image colours were assigned for each micrograph corresponding to excitation wavelengths 405, 488, 555 and 639 nm, and filters 420--480, \u2265490, \u2265560 and \u2265640 nm, respectively. The two 'red' channels, which provide similar results, were each set to 50% saturation and combined into one red channel to compensate for the double acquisition, a process required as otherwise the red autofluorescence would unduly dominate the image making assessment difficult. According to \\[[@RSIF20190049C16]\\], material properties were interpreted based on resultant micrographs as follows---in superimposed images of insect exoskeletal parts: (1) well-sclerotized structures are usually red, (2) tough-flexible cuticular structures are typically yellow-green, (3) relatively flexible parts containing a relatively high proportion of resilin are light blue and (4) resilin-dominated regions appear deep-blue. Note that the image colours represent autofluorescence and code the intensity of light in specific channels.\n\nThis interpretation was confirmed by observation of corresponding samples under a stereomicroscope. Since assessment of material based on CLSM are relative, not absolute, the comparison is only valid if images are taken with the same settings (electronic supplementary material, figure S1 and table S1), or if two structures are imaged in one single scan ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*b,c*).\n\nTherefore, when necessary we scanned multiple samples simultaneously. Images are reproduced here with increased brightness and contrast, to improve clarity (unaltered images are available in the electronic supplementary material). The exact imaging settings can be seen for each image in electronic supplementary material, table S1.\n\n2.4.. Finite-element modelling {#s2d}\n------------------------------\n\nFEM with COMSOL 5.3a (Comsol Inc., Stockholm, Sweden) was conducted in the frequency domain to investigate the vibrational characteristics of the antennae using the Solid Mechanics module. Simulations were performed on both desktop computers and the ARCHIE-WeSt supercomputer.\n\nThe objective of these simulations was to ascertain the relevance of the observed banded structure of the flagellomeres on the overall frequency response of an idealized compliantly clamped beam. We have reduced the mosquito antenna to a simple system, and thus the simulations are solely to observe the frequency-response changes due to the presence of hard and soft rings in the flagellomeres.\n\nControl simulations were done on uniform cylinders whose expected resonant frequencies are well established from Euler--Bernouilli beam theory. Once established that control simulations yield appropriate results, a subdivided ringed structure seen in the CLSM images was simulated ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}).\n\nThe fibrillae were not included in the model---they are apparently stiffly coupled to the beam, so that for biologically relevant frequencies, the fibrillae and flagellum move as one \\[[@RSIF20190049C2]\\]. From a mechanical perspective, these fibrillae add damping, but little mass, and thus broaden the response of the whole antenna but do not shift the resonance frequency appreciably. Additionally, the densities of the materials modelled are identical, since remarkable density differences are not known despite the range of stiffnesses in chitinous structures \\[[@RSIF20190049C36]\\].\n\nThe geometry of the model antenna was informed by the images that indicated that the cuticle of the flagellum is a thin sheet compared to the absolute volume, hence the flagellum is represented by a 10%v wall hollow cylinder filled with tissue (90%v, Young\\'s modulus 1 kPa), which we will describe as a beam, to avoid confusing it with the biological structure. The tissue was assumed to have a fraction of the literature value for soft tissue as those values are given for pulling on tissue assemblies in the direction of maximum resistance.\n\nThis beam, illustrated in [figure\u00a02](#RSIF20190049F2){ref-type=\"fig\"}, was subdivided into 13 long elements of constant size separated by 12 triplets of small elements, to represent the segment joint areas of the antenna. Size of the individual elements was 1/13 of the antenna length (3.3 mm for *T. brevipalpis* and 1.7 mm for *An. arabiensis*) and the diameter was 120 \u00b5m for *T. brevipalpis* and 15 \u00b5m for *An. arabiensis*. Of importance for the present study is ensuring the spatial distribution of elements of different stiffnesses was matched to the results of the CLSM imaging. Figure 2.Three-dimensional representation of our antenna model. This figure correlates modelling terminology with morphological terms. The antenna comprises 13 segments with 12 joints. In a model for *T. brevipalpis*, the antennal structure is a sequence of relatively hard long elements, interspersed with a series of three small elements of one soft material sandwiched between two hard small elements ([figure 3](#RSIF20190049F3){ref-type=\"fig\"}*a,b*). In contrast in *Anopheles*, the antennal structure is a sequence of soft elements, separated by thin discs of hard material (figures [3](#RSIF20190049F3){ref-type=\"fig\"}*c,d* and [4](#RSIF20190049F4){ref-type=\"fig\"}). Points used for simulations in [figure 6](#RSIF20190049F6){ref-type=\"fig\"} are shown in red. To simulate an impinging sound field, a load was applied perpendicular to the beam axis in the +X direction on elements 2--13. Image abbreviations, with morphological terms in brackets if applicable: 'bd' basal disc (approximating the pedicellar articulation); fl, beam (flagellum); j, three small elements (joint); sg, long element (segment). Figure 3.CLSM-based maximum intensity projections of the male antennae of *T. brevipalpis* (*a,b*) and *An. arabiensis* (*c,d*). Higher resolution images available in the electronic supplementary material. (*a*) Antenna, with increased brightness, showing the 12 more-or-less regular subunits and a varying 13th one. Inset: Zoomed image of bands of a different individual. (*b,c*) Comparison image of (*b*) *T. brevipalpis.* (2nd segment) and (*c*) *An. arabiensis* (4th to 8th segment), with increased brightness. The antenna of *T. brevipalpis* is larger and thicker. It is made up of relatively stiff cuticle with small, relatively flexible (blue) and hard (red/orange) rings, while *An. arabiensis* antenna is made up of relatively soft (light blue) cuticle interspaced with hard rings. In both species, the area where the fibrillae emerge is hard (orange). In (*a,b*) (*T. brevipalpis*), the image shows two red-orange discs sandwiching a blue disc at the base of the segments (white arrow). (*d*) Overview of *An. arabiensis* with increased brightness. The overall anatomy is dominated by relatively soft and flexible areas, interspaced with comparatively hard bands where fibrillae insert. Image abbreviations: fb, fibrillae; fl, flagellum; j, joint; pd, pedicel (grey arrowhead in (*d*)); rs, ring structure; sg, segment. Figure 4.Maximum intensity projection of *An. arabiensis*, with increased brightness, showing detail of two flagellomeres and the insertion position of fibrillae. Males of this species can erect the fibrillae depending on diurnal cycle and activity, due to the presumably soft, sac-like structure that is outlined by the granulae (**gr**) below the hard ring structure (**rs**). The combination of these two structures can potentially provide the mechanical basis to inflate by hydraulic pressure and erect the fibrillae. Fibrillae sockets on the ring structure (white arrowheads) are where the fibrillae insert. Image abbreviations: fb, fibrillae; fl, flagellum; gr, granulae; j, joint; pd, pedicel; rs, ring structure; sg, segment. \\*Deep-blue fluorescent small band, \\*\\*yellow-orange fluorescent band.\n\nThe articulation in the pedicel, which unquestionably contributes to the mechanical behaviour of the antenna, is represented by the round disc ([figure\u00a02](#RSIF20190049F2){ref-type=\"fig\"}) for simplicity and is modelled with an effective stiffness that takes into account the diameter of the prongs and an approximation of their stiffness. This compound quality of the pedicel articulation is later on referred to as 'basal stiffness'. It is important to note that our 'basal disc' is not matched in size or any other way to the basal plate, but its allometric relation to the beam length and diameter ensures it stays constant in relative size between species.\n\nAs estimations of Young\\'s modulus of insect cuticle vary greatly between different reports (see \\[[@RSIF20190049C36],[@RSIF20190049C37]\\] as examples), only the ratio of Young\\'s moduli between the materials was used for simulations, while keeping values in the natural range. The large and small segments were allocated stiffnesses within the range of the material property estimation according to CLSM images and literature values \\[[@RSIF20190049C36]\\].\n\nTaken together, the model beam is as follows: for *T. brevipalpis* the large elements are of medium stiffness (0.5 GPa) and the triplet of small elements is a stack of hard (5 GPa), soft (1 MPa) and hard element (5 GPa), followed by the next 'large element'. For *An. arabiensis* the large elements are soft (1 MPa) and all three of the triplet of small elements are hard (5 GPa). Vibrational characteristics of the modelled antennae were simulated in 10 Hz steps ([figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}) over the frequency range of 10--2720 Hz, which includes the typical hearing range of these insects. Only the range 20--2000 Hz is shown.\n\n3.. Results {#s3}\n===========\n\nResults are presented separately for each species and highlight differences and similarities in antennal structure as well as its putative material properties. According to the colour scheme of \\[[@RSIF20190049C16]\\], areas shown in blue are likely to be resilin-dominated structures, relatively soft structures will appear light blue, tough structures in yellow-green and sclerotized structures in red.\n\n3.1.. *Toxorhynchites brevipalpis* male {#s3a}\n---------------------------------------\n\nThe flagellum consists of 13 segments, each of which, except the first that articulates in the pedicel and therefore cannot be observed directly, has a blue ring followed by a yellow-reddish arrowhead or coronal structure that appears to be more sclerotized. At the tip of this structure, fibrillae emerge in a pattern shown in [figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*a,b*. Distal to the coronal structure, a comparatively weakly fluorescent material is located. The thin final ring-like structure, observable in detail in red, is interpreted as being part of the lower segment ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*b*), and the next segment therefore begins with a blue ring structure. The segments continuously decrease in length and diameter from proximal to distal locations.\n\nBy contrast, the fibrillae have a relatively uniform length and autofluorescence along the flagellum. The long fibrillae are absent on the 13th segment, giving the antenna a somewhat rounded appearance ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*a*). The 13th segment is different from the others; it continues for five to six times the length of the 12th, is noticeably thinner, and shows little tapering.\n\nThe 13th segment only sprouts a limited amount of much-shorter fibrillae ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*a*). Externally, the pedicel autofluorescence is rather uniformly green, except the uppermost ridge, where the flagellum emerges ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*a*). Where the pedicel and flagellum join, the material appears orange, due to a higher contribution of red autofluorescence compared to the pedicel\\'s outer wall ([figure\u00a05](#RSIF20190049F5){ref-type=\"fig\"} red arrowhead and electronic supplementary material, figure S1a). Figure 5.CLSM images of a *T. brevipalpis* male that shows a maximum intensity projection of an opened pedicel. This attachment is comparatively hard (red arrowhead) and the prongs run over a ridge (white arrowhead), which appears as hard as the prongs themselves. Image abbreviations: pdw, pedicel outer wall; pr, prongs; rg, ridge (white arrowhead).\n\nThe inside of the pedicel is dominated by centrally attached prongs. Note that the prongs are uniform in their green autofluorescence (and thickness) with their neighbours, as far as can be assessed ([figure\u00a05](#RSIF20190049F5){ref-type=\"fig\"} and electronic supplementary material, figure S1), indicating that prongs are of uniform stiffness.\n\nAn optical section (electronic supplementary material, figure S1a) of the pedicel indicates that the prongs either attach to, or pass over, a cuticular ridge. The thickness of the whole articulation point is less than approximately 2 \u00b5m (electronic supplementary material, figure S1).\n\nHowever, it should be noted that across individuals and depending on experimental settings, the colour of autofluorescence varies---the difference in settings between experiments makes comparison across individuals difficult.\n\n3.2.. *Anopheles arabiensis* male {#s3b}\n---------------------------------\n\nSegments 2--12 are very similar, albeit gradually tapering to about half the initial diameter by segment 12. Each segment consists of a blue part (figures\u00a0[3](#RSIF20190049F3){ref-type=\"fig\"} and [4](#RSIF20190049F4){ref-type=\"fig\"}), followed by a broadened ring-like sclerotized structure in orange-brown-red circumventing the flagellum from which the fibrillae crest of each segment emerges ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*d* white arrowhead [figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"}). The first segment is of similar structure, but along the flagellum axis, the segment first tapers after insertion, then becomes medially swollen ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*d*). The 13th segment is filamentous ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*c*), showing overall low autofluorescence.\n\nDetails of the more strongly orange-red fluorescent ring structure include the socket that each individual fibrilla inserts into ([figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"} white arrowheads). At the site where the ring joins the flagellum, a deep-blue fluorescent small band of similar width as the ring is present ([figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"}\\*), followed by a more yellow-orange fluorescent band ([figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"}\\*\\*) before the segment continues to show a light blue material fluorescence. Note that the area proximal to the ring structure and surrounding the flagellum is lacking autofluorescence, with the exception of scattered orange fluorescent granulae which outline the cuticle sac area ([figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"}, gr below rs). The rather sclerotized articulation of the flagellum in the pedicel is approximately hemispherical and at the upper end of the distalmost part of the pedicel ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*d* grey arrowhead). With CLSM, further internal structure, except the crest of uniformly fluorescent prongs, cannot be visualized (electronic supplementary material, figure S1b).\n\n3.3.. *Toxorhynchites brevipalpis* and *Anopheles arabiensis* male comparison {#s3c}\n-----------------------------------------------------------------------------\n\nComparison of the second segment in *T. brevipalpis* ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*b*) to the sixth to 10th segments of the antenna in *An. arabiensis* ([figure\u00a03](#RSIF20190049F3){ref-type=\"fig\"}*c*) shows a noticeable difference in size. These segments in *T. brevipalpis* are 120--150 \u00b5m wide and 200 \u00b5m long. The antenna of *An. arabiensis* is much smaller with each segment about 100 \u00b5m long, and only 10--20 \u00b5m wide. The long fibrillae inserting on the orange structure circumventing the flagellum lack an autofluorescence gradient and are barely visible in both species as autofluorescence is overall low in the fibrillae.\n\nWhereas large parts of *T. brevipalpis* show nearly no signal in contrast autofluorescence in *An. arabiensis* is in general stronger, every part of the antenna is either likely flexible (blue) or putatively sclerotized (orange) with no areas of intermediate properties. As a further difference, it can be stated that the most red part in *T. brevipalpis* is a coronal structure, while it appears more annular in *An. arabiensis* ([figure\u00a04](#RSIF20190049F4){ref-type=\"fig\"}). Both antennae are almost mirror images in the respect that in *T. brevipalpis* a rather small band (*ca* 10 \u00b5m) of a slightly deeper shade of blue is in between each segment, while the whole 100 \u00b5m long segment in *An. arabiensis* shows light blue autofluorescence. An antennal cross section shows that a rather thin surface layer constitutes the highest proportion of the blue autofluorescence (electronic supplementary material, figure S2) in both species.\n\nFurthermore, the optical sections (electronic supplementary material, figure S1) show that the articulation structure remains uniformly thin (*ca* 2--4 \u00b5m) and comparable in size between the two species*,* despite the fact that the pedicel of *An. arabiensis* is roughly half as large as the pedicel of *T. brevipalpis*.\n\n3.4.. FEM of the male antenna {#s3d}\n-----------------------------\n\nFrequency-domain studies were performed in COMSOL to observe the effect of the various segmental structures seen in CLSM experiments, as well as other standard parameters of an antennal system such as 'base stiffness' and geometry. [Figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*a* demonstrates the effect of changing 'base stiffness' on the frequency response of a simulated beam. It is clear that increasing basal stiffness leads to an increase in the frequency of the resonant peaks. As expected, further increases in basal stiffness yield diminishing changes in the frequency response as predicted by equation (4.1). In the investigated stiffness range for the base of 90 Pa (70 Hz)--100 kPa (1520 Hz), a diminishing gain per decade stiffness increase is observed. Figure 6.Simulation results. For a modelled beam, (*a*) shows the change of frequency response over four orders of magnitude of base stiffness, and (*b*) shows the effect of adding hard or soft elements to a uniform beam with base stiffness of 10 kPa as in [figure 2](#RSIF20190049F2){ref-type=\"fig\"}. Overlaid are dashed curves indicating simulations where both hard and soft elements are added---in this case 10\u00d7 harder elements in combination with 10\u00d7, 100\u00d7 and 1000\u00d7 softer elements---colour-coded as for soft elements. (*c*) Comparison of a *Toxorhynchites*-type model with either uniform stiffness, or hard and/or soft elements added. Only the addition of soft elements appreciably affects the frequency response. Note that the resonance is consistent with wingbeat frequencies in this species. (*d*) Comparison of the *Anopheles*-type model with either uniform stiffness or the addition of hard elements. Hard elements increase the resonant frequency.\n\nIn [Figure 6](#RSIF20190049F6){ref-type=\"fig\"}*b*, elements of different elasticity were added to a beam, distributed in agreement with the CLSM results. It is clear that the stiffness modification of segmented structures in the antennal model influences resonant frequencies.\n\nFor the beam simulated in [Figure 6](#RSIF20190049F6){ref-type=\"fig\"}*b*, softer segments are shown to reduce the peak frequencies by 35, 240 and 525 Hz as we decrease the stiffness by powers of 10. By contrast, the addition of harder elements leads to an increase in the peak frequencies, although to a lesser extent. Without hard elements the resonant frequency is between 690 and 700 Hz and with hard elements the resonant frequency shifts up to 700 Hz.\n\nWhile [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*a*,*b* is already dimensionally matched to *T. brevipalpis,* [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*c*,*d* shows simulations of frequency responses in male *T. brevipalpis* and *An. arabiensis,* respectively*.* The graph in [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*c* shows four lines associated with four cases: (1) uniform beam, (2) uniform beam with added stiff elements, (3) uniform beam with added soft elements and (4) uniform with both stiff and soft elements. The latter case is the most morphologically accurate. Clearly, the addition of soft elements is the only case which allows for significant changes in resonant frequencies, from approximately 695 Hz down to 390 Hz. By contrast, the addition of hard elements is insufficient to affect the natural vibration characteristics substantially, resulting in under a 10 Hz shift from approximately 695 Hz. In [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*d*, the two lines represent: (1) a uniform small and soft beam of the same size as the *An. arabiensis* antenna and (2) the same beam but with added hard elements. The latter case is most faithful to the antenna of *An. arabiensis.* The frequency shift appears to be less than the 305 Hz shift shown in [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*c* for soft elements in a harder beam. The addition of hard elements increases the resonant frequency of the overall softer beam from 160--190 Hz. While [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*a* demonstrates the model representing the situation in a compliantly clamped beam, [figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*b--d* demonstrates that distribution of varying material properties can have a strong effect on the overall resonant frequency. By contrast, in a simple compliantly clamped beam, the resonant frequency is mostly determined by the beam dimensions and the stiffness of the clamp \\[[@RSIF20190049C38]\\].\n\n[Figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*c,d* shows simulations based on stiffness ranges obtained from CLSM autofluorescences in the two species, as well as appropriate geometry. The example of *T. brevipalpis* demonstrates how it is possible to reduce the resonant frequency of a beam to match the observed frequencies seen in the insects studied (i.e. in *T. brevipalpis* and *An. arabiensis* resonant frequencies are 420 \u00b1 5 Hz \\[[@RSIF20190049C39]\\] and 380 \u00b1 46 Hz (new experimental data not shown) for the males, respectively). It is worth noting that within biologically relevant parameter ranges for antenna stiffness *without* adding disc elements to the beam, it is difficult to obtain resonant frequencies in simulation that match those seen in nature. Taken as a whole, the addition of elements of different stiffness in comparison to the main beam can change beam resonant frequency to a similar order of magnitude as changing the basal stiffness ([figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*a*).\n\n4.. Discussion {#s4}\n==============\n\nThe goal of this work was to investigate the material properties of the mosquito antenna. We performed a CLSM study of mosquito antennae and found that autofluorescence is not homogeneous along the antenna, but instead these antennae comprise repeating bands of harder and softer elements. In general, the presence of harder and softer elements in the antenna is similar between the two mosquito species---however, their distribution is inverted: *An. arabiensis* has large rather flexible bands interspaced with harder ring elements, while *T. brevipalpis* is medium-hard overall and has short flexible ring elements wedged between two hard rings. Given the nature of the results, we also simulated these geometrical configurations to determine whether these material changes and spatial distribution has any effect on the overall resonant frequency of the antenna, a property of significant importance to the animal, as sensitivity of these ears is best around the resonant frequency and acoustic perception is essential for mating \\[[@RSIF20190049C10],[@RSIF20190049C11],[@RSIF20190049C26],[@RSIF20190049C27],[@RSIF20190049C30],[@RSIF20190049C31]\\]. It is possible that the different sizes of the animals studied may influence the material properties of the antennae, as the larger antennae perhaps have different mechanical constraints to ensure robustness and structural stability. It is not clear in general whether the observed geometrical differences are driven by behavioural or other constraints unique to each species. These questions could be answered with studies on similarly sized and more behaviourly similar species, and are not addressed here. Some interpretation on how the observed stiffness distribution along the modelled beam could play a role in the resonant characteristics, and the implications of this, might be inferred from FEM simulations discussed below.\n\n4.1.. Confocal laser scanning microscopy {#s4a}\n----------------------------------------\n\nIn *T. brevipalpis* males*,* the larger of the species investigated, the overall structure is relatively tough and there exist small flexible to well-sclerotized ring structures. In *An. arabiensis* males*,* the combination of a sclerotized ring structure, on which the fibrillae are present in each segment, and the area of the membrane sac proximal to it is believed to play a role in the behaviour of male *Anopheles* antenna, in which they collapse and extend their fibrillae at different times of the day using hydraulic pressure \\[[@RSIF20190049C40]\\]. Our images are taken after suspension in fluid and show only antennae with extended fibrillae. This position in nature is only assumed during the active swarming phase \\[[@RSIF20190049C40]\\] and therefore is directly related to the detection of conspecifics. Pedicels both internally and externally look similar in structure, but in *An. arabiensis* the intensity of autofluorescence is higher. In male *T. brevipalpis,* a hard area ([figure\u00a05](#RSIF20190049F5){ref-type=\"fig\"} red arrowhead) is visible, where the flagellum leaves the pedicel. The prongs on the inside appear similar in the two species regarding fluorescence and dimensions compared with prongs of the same animal in the same image. Comparison between images demonstrates that different CLSM settings are necessary for proper visualization of material differences this does not allow judgement of material properties between species imaged individually. However, we can say with some confidence that the prongs are neither particularly flexible nor stiff and are all consistent in their autofluorescence and dimensions within the animal.\n\nThis is in agreement with Avitabile *et al*. \\[[@RSIF20190049C41]\\], in that the prongs act more or less as rigid-body extensions of the flagellum. Possible variations inside the pedicel would likely be due to the scolopidia, which have recently been shown, by direct measurement using atomic force microscopy, to be motile \\[[@RSIF20190049C42]\\], having long been suspected as the source of stiffness gating. Further studies have shown the importance of the scolopidia for both power gain of the antenna, and the intra- and interspecifc variations seen in antennal mechanics \\[[@RSIF20190049C32]\\].\n\nHowever, the present study demonstrates that the flagellum itself cannot be approximated as a rigid beam of uniform stiffness, but that it consists of repeating units of stiff and soft elements. A limitation of the present study is the lack of direct correlation of CLSM-based autofluorescence analysis with mechanical measurements, which is to be tackled in follow-up investigations.\n\n4.2.. Significance of material property differences in antennae {#s4b}\n---------------------------------------------------------------\n\nSince Johnston (1855) \\[[@RSIF20190049C1]\\], the nerve and cuticle structure inside the pedicel has been investigated regarding its auditory and general function. In this study, we visualized the antenna of *An. arabiensis*, a species where males form swarms and females fly into the swarm and are acoustically located by the males. *Toxorhynchites brevipalpis* is a solitary mosquito, where acoustics also plays a role in mating \\[[@RSIF20190049C30]\\]. Within individuals, consistency of the uniform prongs inside the pedicel has been found across both species, and remarkable differences in material distributions were found in the flagellum.\n\nA sensory organ exposed to the environment such as an insect antenna (that can move autogenously without stimulus) is potentially under continuous mechanical stress. This may explain the presence of resilin, a protein known to be used to protect from 'wear and tear' in insects \\[[@RSIF20190049C43],[@RSIF20190049C44]\\].\n\nThe FEM results indicate a potential for resonant tuning by alteration of material distributions along the flagellum. This is perhaps not that surprising, as the influence of differently stiff elements partially can be expected as the resonant frequency (*\u03c9*/*2\u03c0*) of a beam \\[[@RSIF20190049C38]\\] is dependent on Young\\'s modulus through$$\\omega = k^{2}\\sqrt{\\frac{EI}{\\rho A}}\\,,$$where *k* is the wavenumber, *I* is the area moment of inertia, *\u03c1* is the density, *A* is the cross section of the beam, and *E* is Young\\'s modulus. In a composite beam, Young\\'s modulus would likely be an effective Young\\'s modulus of the whole beam, which will be different when elasticity is not uniform. Furthermore, the non-homogeneous material distribution as suggested by CLSM results along the antenna could affect the area moment of inertia. Given the densities are suspected to be similar, the likelihood of this being important is low. Regardless, by changing the distribution of stiffnesses, a further mechanism to control the resonant frequency of the antenna is possible. There are potential benefits to this mechanism in conserving structural integrity in comparison to changing the material.\n\nFEM models of material property distributions in the antennae of the studied insects have been compared to a uniform beam structure. This shows different mechanical behaviours, suggesting that a more rigid antenna, presumably like that of *T. brevipalpis* ([figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*a*), and a soft antenna, like that of *An. arabiensis* ([figure\u00a06](#RSIF20190049F6){ref-type=\"fig\"}*b*), can both be tuned significantly by basal stiffness and distribution of stiffness along the beam. However, the effect of hard elements in soft beams seems less than that of soft elements in hard beams. How this different tuning affects behaviour requires further research.\n\nInterestingly, in order to make our model show resonant frequencies that are found in these insects, the addition of the triplet rings was essential. We found it difficult to reproduce the resonant frequencies found in *T. brevipalpis* and *An. arabiensis* (420 \u00b1 5 Hz \\[[@RSIF20190049C39]\\] and 380 \u00b1 46 Hz (data not shown) for the males, respectively) using only the known geometry and typical biological values for material properties---it was somewhat surprising that only the addition of the triplet rings allowed one to bring down the resonant frequencies in *T. brevipalpis* to observed values. The deviation in *An. arabiensis* is due to the other contributing factor---basal stiffness.\n\nOur FEM model shows a very weak first bending mode at the main resonant peak and a pendulum mode at lower frequencies. In particular, in the case of a small angular displacement, the weak bending mode can easily be perceived as a pendulum mode, as while tip displacement is largest, displacement overall is fairly uniform in both of these mode shapes. This is in line with earlier experiments in different species reporting a pendulum mode \\[[@RSIF20190049C2],[@RSIF20190049C7]\\].\n\nGenerally speaking, there are many factors that may contribute to the antennal mechanical behaviour. An exhaustive list would include the stiffness of the base articulation, the cell attachments to the scolopidia, prongs and scolopidia, geometry, and viscous effects of the fibrillae, among other things. For example, it has been shown in stick insects that tapering of their non-plumose antennae has the largest tuning effect, at least in the static case \\[[@RSIF20190049C44]\\]. The current study suggests that the configuration of spatial distribution of flexibility along the antenna does influence the antenna\\'s mechanical behaviour within the frequency range of the animal\\'s hearing. More in-depth experimental and theoretical investigations of antenna bending are required, which will add to our understanding of mechanical properties of insect hearing systems.\n\n5.. Conclusion and outlook {#s5}\n==========================\n\nGenerally, the highly variable and interesting material properties of insect cuticle have not gone unnoticed (e.g. \\[[@RSIF20190049C21],[@RSIF20190049C45],[@RSIF20190049C46]\\]). However, in the field of insect auditory systems, the fine detail of material properties and distribution is overlooked. The primary goal of this work was to observe and describe morphological and material differences and similarities in antennal hearing organs of the investigated species. While it is possible to consider the antenna as a simple beam, we have shown that the actual material properties of these antenna are more complex. We also show that these material complexities have the potential to modify the frequency responses of the acoustic sensors, providing a different mechanism to the animal to evolve and direct---mating-critical---frequency selectivity. The addition of soft elements to a hard beam shifts the resonant frequencies to lower values, while adding stiff elements to a soft beam does the opposite and shifts the frequency of resonant upwards. The study of varying material distribution of insect hearing organs with CLSM has a high potential for improving our understanding of the evolution and development of acoustic sensors in nature, especially if combined with FEM and possibly mechanical tests of materials and laser vibrometry, to characterize native system behaviour. The very high variability of mechanical and therefore acoustic properties in these insects studied suggests a potential for many interesting future findings and biomimetic engineering exploitation.\n\nSupplementary Material\n======================\n\n###### Figure S1\n\nSupplementary Material\n======================\n\n###### Figure S1a\n\nSupplementary Material\n======================\n\n###### Figure S1b\n\nSupplementary Material\n======================\n\n###### Figure S2\n\nSupplementary Material\n======================\n\n###### Figure S3a\n\nSupplementary Material\n======================\n\n###### Figure S3a - high resolution\n\nSupplementary Material\n======================\n\n###### Figure S3b&c\n\nSupplementary Material\n======================\n\n###### Figure S3d\n\nSupplementary Material\n======================\n\n###### Figure S4\n\nSupplementary Material\n======================\n\n###### Figure S5\n\nSupplementary Material\n======================\n\n###### Table S1\n\nWe thank Dr Jan Michels (University of Kiel) for CLSM training, Dr Alexander Kovalev (University of Kiel) for discussion and advice, Dr Jeremy Gibson, members of the Centre for Ultrasonic Engineering (University of Strathclyde) for their support and Dr Francesco Baldini (Vector Biology and Disease Group, University of Glasgow) for providing *Anopheles* specimens. We thank Shahida Begum (London School of Hygiene and Tropical Medicine) for the provision of *T. brevipalpis*.\n\nData accessibility {#s6}\n==================\n\nAdditional images, all original images and image values (electronic supplementary material, table S1) are available on (<https://figshare.com>) via Interface. Additionally, data is freely available from the University of Strathclyde KnowledgeBase at <https://doi.org/10.15129/8ec4ac84-3040-4c4d-854e-07e50786a68a>.\n\nAuthors\\' contributions {#s7}\n=======================\n\nB.D.S. and Y.M. carried out preparations and CLSM imaging*.* A.R. and B.D.S. conducted FEM simulations. B.D.S., J.F.C.W., S.N.G. and J.C.J. designed the study. B.D.S., A.R., S.G.N., Y.M., and J.C.J. wrote the manuscript. All authors gave final approval for publication.\n\nCompeting interests {#s8}\n===================\n\nWe have no competing interests.\n\nFunding {#s9}\n=======\n\nThis work was partially supported by the European Research Council under the European Union\\'s Seventh Framework Programme FP/2007-2013/ERC under grant agreement no. 615030 to J.F.C.W. by the EPSRC (J.C.J., EP/H02848X/1) and by the German Research Foundation (Y.M., DFG grant no. MA 7400/1-1).\n\n[^1]: Electronic supplementary material is available online at <https://dx.doi.org/10.6084/m9.figshare.c.4492862>.\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\nAcute coronary syndromes (ACS) are a set of signs and symptoms due to the rupture of a plaque and are a consequence of platelet-rich coronary thrombus formation. The thrombus leads to partial or complete coronary artery occlusion, which, in turn, leads to myocardial ischemia and various clinical manifestations ranging from unstable angina (UA) to acute myocardial infarction (AMI).\n\nPlatelets are heterogeneous blood elements with diverse sizes and densities. Platelet activation is a hallmark of ACS. In addition to aggregation, platelets modulate important pathophysiological processes including inflammation and coagulation. It has been shown that platelet size, when measured as mean platelet volume (MPV), is a marker of platelet function and is positively associated with indicators of platelet activity. An increased MPV, an indicator of larger and more reactive platelets, has been associated with myocardial damage in ACS and has been found to be predictive of an unfavorable outcome among survivors of AMI.\\[[@ref1][@ref2]\\] Aggregation also depends on platelet count. There are reports that systemic inflammation plays a role in development and progression of coronary heart disease. Previous studies have documented ethnic differences in MPV level.\\[[@ref3]--[@ref6]\\] Association of higher MPV values with ACS has been mostly studied among Caucasian patients.\\[[@ref7]\\] A few reports have revealed larger MPV values in Indian patients with ACS compared with healthy controls.\\[[@ref8]\\] However, there are less reports in comparison with stable coronary artery diseases. We currently lack understanding of the predictive accuracy of MPV for spectrum of CAD. Our aim was to study MPV and other platelet volume indices (PVI) in AMI and stable CAD and compare them with age- and sex-matched controls and to find predictive value of MPV in spectrum of CAD.\n\nMATERIALS AND METHODS {#sec1-2}\n=====================\n\nThis hospital-based case control study was designed to assess whether MPV and other PVI show variation in the spectrum of CAD. The study protocol was approved by the Institutional Review Board of the hospital, and written inform consent was obtained from the patients. Sample size was calculated based on standard error obtained in pilot study. Sample size was calculated to allow detection of a 30% difference in MPV between different groups and with \u03b1 of .05 and power of .80. Total 128 subjects were recruited and studied in three groups. Group 1A and 1B had patients of coronary artery disease and Group 2 had healthy controls. Group 1A had patients with AMI on admission. Group 1B had patients of coronary artery disease who had AMI at least 5 weeks prior and admitted for angiography without chest pain. Following patients were excluded from the study: Patients with severe hepatic or renal impairment, patients taking oral anticoagulation medicine (but Group 1B patients were thrombolysed and all were on anti-platelet therapy), myeloproliferative disorders and malignancy. The enrolment period was between September 2010 and April 2011.\n\nIn Group 1A patients, before administration of anti-coagulants and anti-platelet drugs, we collected blood samples within 6 hours on arrival at casualty into tubes containing EDTA who were subsequently diagnosed having AMI. For measurement of platelet count (PLC), mean platelet volume (MPV), platelet distribution width (PDW), platelet large cell ratio (P-LCR) and plateletcrit (PCT), samples were analyzed within 30 minutes after collection with Sysmex KX21-N automated flow meter. Blood samples of Group B were collected on the day of admission and were analyzed. Group C subjects came for routine check-up and their blood samples were collected in the outpatient department.\n\nAMI was diagnosed based on the following criteria: detection of rise or fall in cardiac biomarker Trop I or CKMB with at least one value above 99^th^ percentile of upper limit together with evidence of myocardial ischemia based on at least one of the following. 1) symptoms of ischemia, 2) ECG changes indicative of new ischemia, 3) development of pathological Q wave in the ECG, 4) imaging evidence of new loss of viable myocardium or a new regional wall motion abnormality. Group 1B patients were diagnosed based on the following criteria: Evidence of AMI at least 5 weeks prior to admission. Their case reports showed 1) development of pathological Q wave in the ECG and 2) imaging evidence of new loss of viable myocardium or a new regional wall motion abnormality.\n\nDATA ANALYSIS {#sec1-3}\n=============\n\nResults were presented as mean \u00b1 SD or frequency (percentage) as appropriate. One-way analysis was used for statistical analysis of categorical variables with comparisons of *P* \\< 0.05 between Group 1A, 1B and 2. Group 1A and Group 1B were included in CAD group (Group 1) and compared with control Group 2 using independent t test. To determine the accuracy and respective best cut-off values of MPV for predicting AMI/SCAD, the receiver operating characteristic (ROC) curves and their corresponding areas under the curve (AUC) were used. A *P* value of \\<0.05 was considered statistically significant. A common statistical package (SPSS 16.0) was used to perform all statistical tests.\n\nRESULTS {#sec1-4}\n=======\n\nDuring the 8 months of study, 128 individuals were selected for the study (17 females, 111 males). AMI was diagnosed in 39 patients (Group 1A). Total 24 patients who had AMI at least 5 weeks prior to getting admitted for angiography were enrolled (Group 1B); 65 individuals were selected from the outpatient department who were attending for routine check-up (Group 2).\n\nThere was no significant difference between age and sex among the 3 groups. ([Table 1](#T1){ref-type=\"table\"}) Increased MPV, PLC, P-LCR and PCT were observed in AMI Group compared to SCAD and control group (P 0.025). ([Table 2](#T2){ref-type=\"table\"}) MPV levels were significantly raised where other platelet indices were not raised significantly. Increased MPV, PLC, P-LCR and PCT were observed in CAD group compared to control group. MPV levels were significantly raised (*P* = 0.018) whereas other platelet indices were not raised significantly. Significantly increased WBC count was observed in AMI compared to SCAD and control groups (*P* = 0.004). But, increase in platelet count was insignificant. ROC curve of MPV when predicting AMI in patients was constructed and AUC was found to be 0.620 (95% CI) statistically significant (*P* = 0.031). Additionally, AUC of the MPV in predicting SCAD in patients was 0.483 (95% CI) statistically insignificant (*P* = 0.800). The best cut-off values for MPV when predicting AMI and SCAD in patients were 9.25 fl (sensitivity 56.4%; specificity 45.9%) and 9.15 fl (sensitivity 54.2%; specificity 42.23%), respectively.\n\n###### \n\nComparison of platelet volume indices in all the cases\n\n![](JCDR-3-272-g001)\n\n###### \n\nComparison of platelet volume indices in CAD and controls\n\n![](JCDR-3-272-g002)\n\nDISCUSSION {#sec1-5}\n==========\n\nWe observed increased platelet volume indices among AMI group compared to SCAD and controls. Specifically, MPV levels were significantly higher among patients with AMI compared to controls. The MPV comparison between these groups showed borderline significance. Future research with a larger sample size is needed to clarify this issue.\n\nOur study has certain limitations. Sample size was small. The relatively low number of included subjects was due to the study design attempting to limit the influence of several co-variables. We could not measure Trop I and CKMB in all the patients, which can have predictive value in ACS so that we could have compared the predictive value of MPV. There are potential confounding factors of MPV. It has been shown that MPV values vary between different ethnicities; furthermore, medications and illness also influence this value. For example, obesity, smoking, aging and diabetes increase MPV values, but aspirin, clopidogrel and inflammatory bowel disease decrease MPV values.\\[[@ref9]--[@ref13]\\] Chu *et al*. observed stepwise decrease in MPV in subjects with chest pain in AMI, UA and non-cardiac chest pain.\\[[@ref14]\\] Yilmaz *et al* observed a stepwise decrease in MPV between MI, UA and stable coronary artery disease among patients in Turkey.\\[[@ref15]\\] Lippi *et al*. reported that Italian patients with ACS had significantly higher MPV values than patients without ACS.\\[[@ref16]\\] Our results, like those of previous studies, demonstrated that MPV can be predictive of AMI, though it was not significant statistically for SCAD. However, in those studies, blood sampling was done either within 24 hours of onset of chest pain or at a time not specified except for the study by Chu *et al*. A few reports published have revealed a larger MPV in Indian patients with ACS compared with healthy controls or patients with stable coronary heart disease.\\[[@ref8]\\] However, those studies are all retrospective reviews of laboratory data, and the time between blood sampling and ACS events is not specified. To our knowledge, we are the first in the literature to study this issue in Indian population that is, within 4 hours of onset of chest pain. Automated cell counters in modern hospital laboratories have made MPV measurement routinely available. Thus, this effortless laboratory test can be added value to diagnosis of spectrum of CAD.\n\nThe clinical implications of our findings are multiple. First, a multi-marker approach to the diagnosis of AMI, combining markers reflecting different pathophysiology, has been shown to be clinically helpful. Second, we have found that signs of platelet activation seem to occur as early as 6 hours after onset of chest pain in patients with AMI. MPV had been proven to be a prognostic factor for angiographic reperfusion and 6-month mortality in patients with AMI treated with primary percutaneous intervention.\\[[@ref17]\\] Martin *et al*. have shown that the MPV, when measured 6 months after AMI, is an independent risk factor for recurrent MI.\\[[@ref18]\\] The differences in MPV between those with MI and healthy controls have been found to persist after 6 weeks.\\[[@ref19]\\] To substantiate these findings, we have also observed persistently increased MPV in Group 1B, that is, even after 5 weeks of myocardial infarction MPV was persistently high in our subjects. Taking all these findings together, MPV reflects an atherosclerothrombic tendency in the human body. Future studies including the use of MPV in a risk stratification system to predict MI or ACS as well as in response to intervention are worthy of consideration.\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared.\n"}
+{"text": "Introduction {#S0001}\n============\n\nRemarkable advancements have occurred in the development of combination antiretroviral therapy (cART) for the treatment of HIV infection since the approval of the first antiretroviral agent, zidovudine, in 1987. Since then, more potent, effective and better tolerated treatment regimens which require less frequent dosing and/or that are co-formulated into single-tablet regimens (STRs) have revolutionized the treatment of HIV and led to significant gains in life expectancy among people living with HIV (PLWH).\n\nThe current treatment paradigm, supported by guidelines developed by reputable organizations including the US Department for Health and Human Services (DHHS) and the International AIDS Society (IAS), includes initiating HIV treatment with two nucleoside or nucleotide reverse transcriptase inhibitors (NRTIs) in combination with either an integrase strand transfer inhibitor (INSTI), a protease inhibitor (PI) or a non-nucleotide reverse transcriptase inhibitor (NNRTI) with INSTIs being the preferred drug class and others reserved as alternative agents for certain clinical situations. Sustained and durable suppression of viral replication is now possible with such treatment combinations when taken consistently and the repertoire now includes more than 10 STRs.[@CIT0001]^--^[@CIT0004]\n\nGiven the expanded treatment options, regimen selection by HIV care providers include considerations such as dosing frequency, food requirement, drug--drug interaction (DDI) potential, short- and long-term toxicities as well as costs of what is expected to be life-long treatment. These considerations and regimen characteristics are also important for drug development and inform what drugs and treatment strategies should be advanced for clinical use.\n\nOne approach to simplifying treatment and limiting its associated toxicities is the use of dual ARV (treatment) regimens (DTRs) that have been evaluated in multiple studies for both initial and maintenance therapy for PLWH.[@CIT0005]^--^[@CIT0008] Results of these studies have shown variable efficacy but some of these regimen constructs have been associated with an increased risk for treatment-emergent drug resistance as well as decreased efficacy in individuals with high viral loads and/or low CD4 counts.[@CIT0009]^--^[@CIT0011] One regimen that has shown satisfactory efficacy and subsequently received approval by the US Food and Drug Administration (FDA) is dolutegravir/rilpivirine but is only for use for maintenance of virologic suppression in treatment-experienced patients.[@CIT0012]\n\nStudies in treatment-naive individuals have been performed comparing 2-drug to 3-drug regimens. An open-label randomized study involving 757 treatment-naive patients comparing ritonavir(r)-boosted lopinavir (LPV/r) or efavirenz (EFV), each in combination with two NRTIs (standard of care \\[SOC\\]) to the NRTI-sparing combination -EFV/LPV/r, found that the efficacy of dual therapy was similar to the EFV/2 NRTIs arm but was associated with more treatment-emergent resistance.[@CIT0006],[@CIT0013]--[@CIT0016] Another study showed that the 2-drug combination of LPV and 3TC had similar efficacy but superior tolerability compared to LPV/2 NRTIs among ARV-naive patients.[@CIT0017]\n\nSimilarly, dolutegravir (DTG) in combination with 3TC (DTG/3TC) has been extensively studied as a 2-drug regimen with favorable characteristics such as excellent tolerability (both drugs) and high resistance barrier (DTG). It is the first 2-drug STR approved as initial therapy for PLWH and is marketed as DOVATO^\u00ae^ (GlaxoSmithKline, Research Triangle Park, NC, USA).[@CIT0018],[@CIT0019] This review will include an overview of the chemistry, pharmacodynamic and pharmacokinetic properties, clinical trials data and drug--drug interactions of DTG/3TC.\n\nOverview Of The Market {#S0002}\n======================\n\nOver the past two decades, the paradigm of cART has been the use of 3-drug regimens as initial therapy for the treatment of HIV-1 infection. DTG/3TC enters a competitive market of over 50 approved medications many of which are co-formulated products. However, DTG/3TC is only the second approved 2-drug STR but holds the distinction of being the first one approved for treatment-naive PLWH.[@CIT0020]\n\nIntroduction To Compound {#S0003}\n========================\n\nDTG/3TC is a two-drug combination of DTG, an INSTI and 3TC, an NRTI, and is approved as a complete regimen for the treatment of HIV-1 infection in ARV-naive adults with no genotypic resistance to the individual ARV components.[@CIT0021]\n\nChemistry {#S0004}\n=========\n\nDolutegravir/Lamivudine (DTG/3TC) STR {#S0004-S2001}\n-------------------------------------\n\nThe single tablet contains 300 mg of lamivudine and 50 mg of dolutegravir (equivalent to 52.6 mg dolutegravir sodium) as active ingredients. Inactive ingredients include povidone K29/32, sodium starch glycolate, sodium stearyl fumarate magnesium stearate, mannitol and microcrystalline cellulose, as well as hypromellose, polyethylene glycol, titanium dioxide in the tablet's coat. The tablet is formulated for oral administration.[@CIT0019]\n\nDolutegravir (DTG) {#S0004-S2002}\n------------------\n\nDTG sodium is monocarboxylic acid amide and an organic heterocyclic compound with a chemical formula --sodium(4R,12aS)-9-{\\[(2,4-difluorophenyl)methyl\\]carbamoyl}-4-methyl-6,8-dioxo-3,4,6,8,12,12a-hexahydro-2Hpyrido\\[1\u02b9,2\u02b9:4,5\\]pyrazino\\[2,1-b\\]\\[1,3\\]oxazin-7-olate ([Figure 1](#F0001){ref-type=\"fig\"}). It has a sodium moiety with the molecular formula -- C~20~H~18~F~2~N~3~NaO~5~ with an exact mass of 441.36 g/mol. It is light yellow to white powder in appearance and is slightly water soluble.[@CIT0021],[@CIT0022] Figure 1DTG molecular structure (adapted from PubChem).[@CIT0022]\n\n### Pharmacodynamics {#S0004-S2002-S3001}\n\nDTG is a potent integrase inhibitor that employs divalent cations (Mg^2+^) to couple with the enzymatic active site of the viral integrase. Its structure allows DTG to penetrate the active and recently vacated enzymatic pocket where it binds farther in than prior drugs in its class. This provides a more stable and lasting bond compared to other precursor integrase inhibitors such that its dissociation constant (mean dissociation constant 2.7 x 10^--6^ s^\u22121^) is slower compared to either raltegravir (RAL) or elvitegravir (EVG) -- 22 and 71 x 10^--6^ s^\u22121^, respectively. It also has a lower half-maximal inhibitory concentration (IC50) for HIV-1 of 1.6 nM compared to 3.3 and 6 nM for RAL and EVG, respectively.[@CIT0023],[@CIT0024]\n\n### Pharmacokinetics {#S0004-S2002-S3002}\n\nFollowing ingestion, DTG peaks in 2--3 hrs. Low-, medium- and high-fat meals increase DTG's AUC by 33%, 41% and 66%, respectively. The median time to maximal concentration (Tmax) is 2.5 hrs. It is tightly protein bound (99%) in plasma and has a medium volume of distribution of 17 L. Its elimination half-life is about 14 hrs which allows for once-daily dosing. DTG is excreted mainly via feces (64%) and urine (31%). DTG is metabolized by UGT1A1 (major pathway) and CYP 3A (minor pathway) but does not inhibit UGT1A1 or inhibit or induce any CYP enzymes, therefore possesses a limited drug--drug interaction profile. However, it is a substrate of BCRP, P-gp, UGT1A3 and UGT1A9.^25^\n\nLamivudine (3TC) {#S0004-S2003}\n----------------\n\nLamivudine is a cytidine analogue with a chemical formula (2R,cis)-4-amino-1-(2-hydroxymethyl-1,3-oxathiolan-5-yl)-(1H)-pyrimidin-2-one, formulated as a sulfur salt ([Figure 2](#F0002){ref-type=\"fig\"}). Its molecular formula is C~8~H~11~N~3~O~3~S with an exact mass of 229.3 g/mol. It appears as an off-white to white crystal that is water soluble with a melting point of 160--162\u00b0C.[@CIT0026] Figure 23TC molecular structure (adapted from PubChem).[@CIT0027]\n\n### Pharmacodynamics {#S0004-S2003-S3001}\n\n3TC competitively inhibits HIV-1 reverse transcriptase by incorporation of its active tri-phosphorylated form into the viral DNA producing chain termination. It has excellent activity against HIV-1 and HIV-2 virus with an EC50 value of 120 nM for HIV-1 and HIV-2 in cell lines including PBMCs and monocytes. 3TC has activity against hepatitis B virus, although monotherapy is associated with the rapid development of resistance.[@CIT0028],[@CIT0029]\n\n### Pharmacokinetics {#S0004-S2003-S3002}\n\nFollowing oral ingestion, 3TC is rapidly absorbed with excellent bioavailability (87% approximately) that is not significantly impacted by meal intake and with a T~max~ of 1 hr. It distributes mainly to the extravascular spaces with a peak time of approximately 3.2 hrs and a volume of distribution of 96 L. Its half-life of 13--19 hrs allows for once-daily dosing. 3TC does not undergo hepatic metabolism and is excreted primarily through the kidneys, hence dose adjustments are required in the setting of compromised renal function. 3TC exhibits linear pharmacokinetics, achieving steady state by day 15 of administration.[@CIT0029]--[@CIT0033]\n\nClinical Efficacy {#S0005}\n=================\n\nInitial Therapy {#S0005-S2001}\n---------------\n\nThe DTG/3TC combination was first evaluated for efficacy, safety and tolerability in the proof-of-concept PADDLE study conducted in Argentina (see [Table 1](#T0001){ref-type=\"table\"}). Twenty ARV-naive subjects with HIV-1 RNA \\< 100,000 copies/mL, CD4 cell counts \\> 200 cells/mm^3^, no known genotypic resistance to lamivudine and with negative hepatitis B surface antigen (HBsAg) at screening were enrolled. Median age was 34 years and 19 subjects were male. The primary endpoint was the proportion of subjects with HIV-1 RNA below 50 copies/mL at week 48. All enrolled participants completed 48 weeks on the study except 1 patient (who committed suicide).[@CIT0034]Table 1Summary Of Clinical Trials -- DTG/3TC For Treatment-Naive IndividualsAuthorsTitleStudy DesignSubjects on DTG/3TC vs3-Drug RegimenHIV RNA\\< 50 Copies Per mL,\\\nDTG/3TC vs 3-Drug Regimen (% FDA Snapshot)Mean CD4 Count Increase\\\n(cells/\u00b5L)Discontinuation Due To AEs\\\nDTG/3TC vs3-Drug Regimen (%)Virologic Failure (%)Treatment Emergent Resistance,\\\nDTG/3TC vs 3-Drug RegimenCahn et al[@CIT0034]PADDLEProof of concept (Phase I)20/NA90(48 week)26705NoneTaiwo et al[@CIT0035]ACTG A5353Phase II120/NA90(24 week)NA0\\<1%M184V and R263R/K in 1 subjectCahn et al[@CIT0036]GEMINI 1 and 2Phase III719/72291(48 week)NA1/22/3 (virologic non-response)None\n\nOn intention to treat analysis, 90% achieved the primary endpoint by week 48 (per protocol viral suppression rate was 95%) and this included 3/4 subjects who were later found to have a baseline viral load \\>100,000 copies/mL. Rapid decline in viral loads was observed with all subjects having a VL \\< 50copies/mL by week 8. The sole subject who later experienced virologic failure re-suppressed without any change in therapy off-study at week 48. Median CD4 cell count change was by 267 cells/mm[@CIT0003] from baseline. Treatment was well-tolerated overall, and adherence was reported to be 100% by self-report and confirmed by pill counts.\n\nACTG5353 was a multicenter Phase II single-arm study of DTG/3TC combination for ARV treatment-naive individuals with HIV VLs between 1000 and 500,000 copies/mL with no CD4 count restrictions and no major resistance mutations to NRTIs, NNRTIs or INSTIs. Enrolled subjects (120) had a median age of 30 years were predominantly male (87%) with racial/ethnic composition including blacks (40%), Hispanics (28%) and Asians (2%). Thirty-one per cent of subjects had HIV VLs \\>100,000 copies/mL. Median baseline CD4 count was 387 cells/mm^3^. The primary endpoint was the proportion of subjects with VL \\< 50 copies per mL at week 24 using the FDA snapshot algorithm. Secondary measures included safety and tolerability assessments, change in CD4 cell count, emergence of resistance and efficacy evaluation by baseline viral load (\\< or \\> 100,000 copies/mL). DTG levels were also measured to assess for its relationship to the study outcomes.\n\nThe primary study outcome was achieved in 90% (108/120) of subjects. Results were comparable regardless of baseline viral load. Five (4%) had virologic non-success and 7 (6%) had no virologic data in the analysis window. Virologic failure in 3 subjects occurred after initial viral suppression and was thought to have been the result of suboptimal adherence as evidenced by DTG serum levels (\\<5ng/mL at 1 or more time points). The M184V (NRTI resistance) and R263R/K (INSTI resistance) mutations developed in 1 subject. Overall, the DTG/3TC regimen was well-tolerated with no study discontinuation due to adverse events. Self-reported adherence was high with 90% of subjects reporting perfect adherence.[@CIT0035]\n\nTwo identical multinational randomized, double-blind, non-inferiority Phase III studies conducted in 192 centers in 21 countries (GEMINI-1 and GEMINI-2) evaluated DTG/3TC compared to DTG/tenofovir disoproxil fumarate (TDF)/Emtricitabine (FTC) for treatment of ARV-naive individuals (see [Table 1](#T0001){ref-type=\"table\"}). Enrolled participants had to have HIV VL \\<500,000 copies/mL, and no major mutations conferring resistance to NRTIs, NNRTIs or PIs Women of childbearing potential who were on approved contraception modalities were also enrolled. A total of 1441 participants were randomized 1:1 to either study arm (719 in 2-drug and 722 in 3-drug regimen groups but with 3 and 5 individuals in each group not receiving a single dose of drug, respectively). Demographics and baseline clinical characteristics were similar among the groups studied, median age was 33 years, participants were predominantly male (85%) and white (68%). Most subjects had baseline had viral loads \\<100,000 copies per mL (80%) and CD4 counts \\>200 cell/mm^3^ (92%).\n\nThe primary endpoint was the proportion of participants with VL \\< 50 copies per mL at week 48 using the FDA snapshot algorithm with the objective to show non-inferior virologic efficacy with a margin of \u221210%. Secondary endpoints included proportion of subjects achieving VL \\< 50 copies per mL at week 24, time in weeks to achieve VL \\< 50 copies per mL and change in CD4 cell count from baseline to week 48. Safety endpoints included incidence and severity of adverse events and proportion of participants discontinuing treatment due to safety events. Renal and bone biomarkers were assessed at baseline, week 24 and week 48.\n\nIn GEMINI-1, the primary endpoint in the-intention-to-treat-exposed subjects was 90% (320 of 356) in the 2-drug group and 93% (332 of 358) in the 3-drug group. Similarly, in GEMINI-2, 93% (335 of 360) in the 2-drug group and 94% (337--359) in the 3-drug group achieved the primary endpoint. Median time to viral suppression was impressive and observed to be 29 days. Both studies met pre-set non-inferiority criteria for virologic efficacy. On combined analysis, virologic non-response (defined by Snapshot analysis as \u226550 copies per mL at week 48) was seen in 3% (20 of 716) in the 2-drug group and 2% (13 of 717) in the 3-drug group.\n\nTen individuals (\\<1%) met criteria for virological withdrawal (6 in the 2-drug regimen group and 4 in the 3-drug regimen group). Virologic withdrawal was met if a second and consecutive VL met any of the following: decrease in baseline of less than 1 log[@CIT0010], unless level was \\<200 by week 12; VL \\> 200 copies/mL at or after week 24; or rebound (VL \\> 200 after level was \\<200 copies/mL). Genotypic and phenotypic resistance testing did not reveal any reverse transcriptase or integrase mutations in these subjects.\n\nOverall, the combined results in GEMINI 1 and 2 trials were similar among individuals with higher (\\>100,000 copies per mL) and lower viral loads (\u2264100,000 copies per mL). Similar virologic response was also seen for individuals with CD4 counts \\>200 cells/mm^3^ (93% in both groups). Less virologic response occurred in those with CD4\\< 200 cells/mm^3^ (79% in 2-drug and 93% in 3-drug regimens), but was thought to be unrelated to efficacy of the treatment regimens or treatment failure. Safety outcomes were similar across both treatment groups as were changes in health quality of life assessment (EQ-5D-5L health utility scale).[@CIT0036] Results of safety outcomes are detailed in the section below.\n\nSafety And Toxicity Of DTG/3TC In Phase III Trials {#S0006}\n==================================================\n\nIn the combined GEMINI studies, adverse events (AEs) were reported in 76% (543 of 716) of subjects in the two-drug regimen and 81% (579 of 717) of subjects in the 3-drug group (see [Table 2](#T0002){ref-type=\"table\"}); safety and tolerability were comparable between treatment groups. Drug-related AEs were numerically lower in the 2-drug group and was attributable to less drug-related grade 1 nausea. Most adverse effects were considered non-serious and commonly included headache, diarrhea and nasopharyngitis. Adverse effects leading to study discontinuation and any serious event were similar between the two study groups. Neuropsychiatric AEs related to suicidal ideation or behavior occurred in 2% of each group. Two deaths occurred, one in each study group, but investigators thought they were unrelated to study drugs.Table 2Select Adverse Effects Reported From Phase III Trials At Week 48Adverse EffectsDTG/3TC (%)3-Drug Regimen (%)**Any adverse effect**\u2003Headache1010\u2003Diarrhea911\u2003Nasopharyngitis811\u2003Upper respiratory tract infection86\u2003Pharyngitis54\u2003Back pain54\u2003Nausea47\u2003Insomnia46**Drug-related adverse effect**1824**Serious adverse events**\\\n**Drug related serious adverse events**7\\\n18\\\n1**Adverse events leading to discontinuation of treatment or withdrawal from study**22**Discontinuation due to drug-related adverse events**11\n\nRegarding secondary outcomes, changes in biomarkers related to renal and bone safety favored the 2-drug regimen group. On the other hand, significantly lower elevations in lipids were observed in the 3-drug arm (see [Tables 3](#T0003){ref-type=\"table\"} and [4](#T0004){ref-type=\"table\"}). All these findings are consistent with the addition of TDF in the 3-drug treatment arm with its known effects on the kidneys, bone and cholesterol. Health-related quality of life scores was similar and high at baseline for both groups and remained high at week 48.Table 3Changes In Renal Biomarkers In GEMINI TrialsParametersTwo-Drug RegimenThree-Drug RegimenP-valueMean change in serum creatinine (mmol/mL)+10.4+13.50.0001GFR from creatinine CKD-EPI (mL/min per 1.73 mm^3^)\u221212.1\u221215.50.0001Protein creatinine ratio of week 48 to baseline (g/mol)+0.87+1.030.0001 Table 4Changes In Lipid Profiles In GEMINI StudiesAdjusted Mean Change (mmol/L)Two-Drug RegimenThree-Drug RegimenP-valueTotal cholesterol+0.32\u22120.150.0001HDL cholesterol+0.15+0.020.0001LDL cholesterol+0.17\u22120.140.0001Triglycerides+0.03\u22120.080.0457Ratio of total cholesterol to HDL cholesterol\u22120.12\u22120.240.0182\n\nUnique Safety And Toxicity Considerations {#S0007}\n=========================================\n\nDTG {#S0007-S2001}\n---\n\n### Use In Pregnancy {#S0007-S2001-S3001}\n\nA National Institute of health-funded observational surveillance study of outcomes at birth among pregnant women on cART in Botswana identified a possible increased risk of neural tube defects when DTG is taken at the time of conception.[@CIT0037]--[@CIT0039] Treatment guidelines have since been updated to reflect this safety concern such that the risks and benefits of treatment with DTG should be shared between women of childbearing potential and their providers.[@CIT0040]\n\n### Changes In Serum Creatinine {#S0007-S2001-S3002}\n\nDTG may cause minimal increases in serum creatinine by inhibition of renal transporter OCT-2. The increase is typically seen within the first 4 weeks of use but does not reflect a true decline in glomerular filtration rate (GFR) or renal function.[@CIT0041]\n\n3TC {#S0007-S2002}\n---\n\n### Resistance Barrier And Mutation {#S0007-S2002-S3001}\n\n3TC has broad activity against HIV-1 and HIV-2 virus. Resistance develops when there is a substitution of methionine for valine at amino-acid position 184 (M184V or M184I) of the reverse transcriptase *pol* gene of HIV-1. The mutation is easily selected for and appears days after exposure to 3TC monotherapy or with suboptimal adherence to regimens containing the medication. However, the M184V/I mutation reduces HIV replication and overall viral fitness; hence good clinical outcomes may still be observed for patients with 3TC-resistant strains of HIV-1 that continue 3TC treatment in combination with other regimens particularly those that include tenofovir to which such strains may be hypersusceptible.[@CIT0038],[@CIT0042],[@CIT0043]\n\n### Hepatitis B {#S0007-S2002-S3002}\n\nAs 3TC has antiviral activity against hepatitis B virus (HBV), severe acute exacerbation of hepatitis may occur in HBV/HIV co-infected patients who have used 3TC-containing regimens and discontinue it. Therefore, close monitoring of such individuals is warranted.[@CIT0044]\n\nDrug Interactions {#S0008}\n=================\n\nDolutegravir (DTG) {#S0008-S2001}\n------------------\n\nCompared to other classes of ARV agents, INSTIs have a low potential for drug interactions. Nonetheless, DTG has several important drug interactions to consider.\n\nThe absorption of DTG is significantly impaired by co-administration with divalent or trivalent cations such as iron, calcium, aluminum, and magnesium. Such interactions can be minimized if DTG is taken 2 hrs before or 6 hrs after ingestion of medications or substances that contain these agents. Drugs that induce CYP3A4 and the transporters UGT1A1 and P-glycoprotein will reduce DTG levels. These include rifampin, carbamazepine and EFV so these should not be co-administered with DTG (dosed daily). In cases where co-administration with rifampicin and carbamazepine is necessary (see section on dosing), the dosing frequency of DTG should be increased to twice-a-day to overcome the interaction.\n\nOn the other hand, the pharmacokinetics of metformin and dofetilide are altered when given with DTG with resultant increased levels of both drugs. The recommendation is to limit the dosing of metformin to a total daily dose of 1 g only with concurrent use. The use of dofetilide with DTG is contraindicated.[@CIT0045]\n\nLamivudine (3TC) {#S0008-S2002}\n----------------\n\n3TC is associated with limited significant drug interactions. Sorbitol-containing products can result in dose-dependent decreases in Cmax and AUC of 3TC when co-administered. Such products should therefore be avoided for patients on 3TC.[@CIT0046]\n\nRegulatory Affairs {#S0009}\n==================\n\nDOVATO received approval by the US FDA on April 8, 2019 for the treatment of HIV-1 infection in ARV-naive adults with no known viral mutations associated with resistance to the individual components.\n\nOn July 3, 2019, the European Medicines Agency (EMA) granted marketing authorization to DOVATO for the treatment of HIV-1 infection in ARV-naive adults and adolescents above 12 years of age weighing at least 40kg, with no known or suspected resistance to the integrase inhibitor class or 3TC.\n\nFurther applications are pending worldwide.\n\nDosing Routes {#S0010}\n=============\n\nDTG/3TC is to be taken as a single tablet once daily by oral route with or without food. As a fixed-dose tablet, it is not recommended for use in patients with creatinine clearance less than 50mL/min. If given with carbamazepine or rifampin (see drug interaction section), an additional dose of 50mg of DTG (as a single agent) should be taken 12 hrs from the first dose of DTG/3TC.[@CIT0020]\n\nConclusion And Expert Opinion {#S0011}\n=============================\n\nDTG/3TC represents an effective NRTI drug-sparing option with demonstrated efficacy as initial therapy for the treatment of HIV-1 infection. Studies evaluating DTG/3TC for maintenance of virologic suppression for treatment-experienced patients are ongoing, and the expectation is that it will be just as successful as the combination of DTG/rilpivirine (RPV) that holds the distinction of being the first approved 2-drug regimen for maintenance therapy.\n\nBoth DTG and 3TC, as have been described in the earlier sections of this article, are potent antiretroviral compounds that both have favorable tolerability profiles and limited drug interaction potential. Reassuringly, clinical trials have shown that the regimen works well for individuals with high viral loads (100,000--500,000 copies/mL) and low CD4 counts (\\<200 cells/mm^3^) and that the concerns about treatment-emergent viral resistance for treatment failures are unwarranted as even where resistance to 3TC develops, resistance to DTG remains very rare due to its high resistance barrier. Treatment providers, however, may be wary of offering DTG/3TC to patients who are likely to be poorly or sub-optimally adherent to treatment.\n\nIt is important to note that certain categories of patients were either underrepresented or excluded in the late-phase clinical trials of DTG/3TC including women, individuals with very high viral loads (\\>500,000 copies/mL) -- an important consideration for \"test and treat\" strategies, children less than 12 years of age or weighing \\<40kg, individuals with hepatitis B co-infection, those with resistance mutations to NRTI and/or INSTI class of ARVs and those with renal insufficiency (GFR\\<50mL/min). This impacts the generalizability of the prelicensure studies and highlight some populations for which more data are needed to justify its use. In addition, with emerging data suggestive of birth (neural tube) defects occurring at a low but significantly higher rate among pregnant women exposed to DTG compared to other comparator antiretroviral agents and the rate among the general population, there is lingering concern about its use for women of childbearing potential.[@CIT0039],[@CIT0040],[@CIT0047] Similarly, neuropsychiatric side effects occurring as both short-term and long-term complications have also been reported with DTG exposure, including exacerbation of depressive disorders with suicides are a consideration for prescribing providers.[@CIT0048]\n\nBy excluding the NRTIs -- abacavir (ABC) and tenofovir (TDF or TAF), DTG/3TC excludes compounds that have been associated with long-term toxicities including cardiovascular disease, renal and bone toxicity. However, the exclusion of tenofovir (specifically TDF) also confers weaknesses on the DTG/3TC combination such that it is not a preferred regimen for HIV/HBV co-infected patients as 3TC is quite resistance prone and the favorable effect of TDF on lowering lipids (cholesterol) is lost.\n\nOf note, INSTI resistance was not tested in GEMINI studies though the product label of DTG/3TC states that its use is only indicated in those without known INSTI mutations. This is not likely to be relevant in clinical practice as the prevalence of transmitted (pre-treatment) INSTI resistance remains extremely low (1.2%).[@CIT0049]\n\nEconomic considerations are worth highlighting as well. The availability of generic 3TC enables DTG/3TC to enter the ARV market with a reduced price that is approximately 26% less than that of bictegravir/TAF/FTC and almost 35% less than the price of DRV/cobicistat, co-formulated with TAF/FTC. This may have a favorable impact on utilization and inclusion of this drug on treatment formularies.[@CIT0020]\n\nIn the near future, with the anticipated emergence of long-acting ARV formulations, the era of the daily dosed STR as the preferred treatment approach may be experiencing the beginning of its end. Till that occurs, however, the approval of an effective and well-tolerated 2-drug STR for ARV-naive PLWH is certainly a game-changer and welcome development.\n\nDisclosure {#S0012}\n==========\n\nThe authors report no conflicts of interest in this work.\n"}
+{"text": "INTRODUCTION {#s0}\n============\n\nLeishmaniasis is the general name for a group of diseases caused by different species of the parasitic protozoan *Leishmania* that produce distinct clinical outcomes, ranging from localized, self-limiting cutaneous lesions, to more chronic forms of cutaneous and mucocutaneous disease, to visceral disease, which is fatal if untreated. According to the WHO there are around 1.3 million new cases of leishmaniasis around the world, with 20,000 to 30,000 deaths each year (<http://www.who.int/mediacentre/factsheets/fs375/en/>).\n\n*Leishmania* parasites have a dimorphic life cycle, shifting between the alimentary tract of their sand fly vector as extracellular, flagellated promastigotes and the phagolysosomal vacuoles of their mammalian host mononuclear phagocytes as intracellular amastigotes. The ability of the parasite to adapt to these vastly different environments has been the focus of a number of studies comparing the transcriptomes, proteomes, and metabolomes of amastigotes and promastigotes during their transformation *in vitro* ([@B1][@B2][@B6]). More recent studies have addressed the differential gene expression levels associated with the maturation of promastigotes to the mammalian-infective, metacyclic stage, again using *in vitro* conditions ([@B7]). Finally, two recent studies have demonstrated differential gene expression between *L.\u00a0infantum* promastigtoes in the anterior midgut of a natural vector, *Phlebotomus perniciosus*, versus those in stationary-phase culture ([@B8]) or versus intracellular amastigotes *in vitro* ([@B9]). To date, no studies have defined the genetic reprogramming associated with amastigote-to-promastigote transformation in the vector, nor have studies focused on the more complex series of promastigote developmental changes that accompany the maturation of transmissible infections *in vivo*.\n\nPhlebotomine sand flies are the only known natural vectors of *Leishmania* spp. responsible for human disease (reviewed by Killick-Kendrick in 1999 \\[[@B10]\\]); they are essential to sustain the life cycle of this parasite. Female sand flies require a blood meal before they will lay eggs ([@B10]), and the gonotrophic cycle is defined as the period between blood meals during which the ova develop and are subsequently deposited. The sequence of physiological events that delineate a normal gonotrophic cycle, and that define the ecology of the *Leishmania* parasite in the vector, can be summarized as follows: when a sand fly takes a blood meal, the ingested blood passes through the food canal and esophagus into the posterior midgut via the stomodeal valve (SV), which regulates the flow of fluids into the gut. A peritrophic matrix (PM) is rapidly secreted by midgut cells to completely envelop the fresh blood meal ([@B11], [@B12]). The PM protects the midgut epithelial cells from damage by blood meal contents, but it remains permeable to the digestive enzymes induced by blood feeding (reviewed by Lehane in 1997 \\[[@B13]\\]); nutrients derived from the digesting blood meal are required for egg development, which become fully mature and deposited around the time that the blood meal remnants are excreted ([@B10], [@B14]). Depending on the sand fly species and ambient conditions, varied proportions of females survive oviposition and undergo multiple gonotrophic cycles, with each additional cycle requiring another blood meal, thus increasing the capacity of the vector to transmit *Leishmania* ([@B10]). During each gonotrophic cycle, females will continue to feed on sugar meals that serve as an energy source for the fly during the interval between blood meals. The sources of the sugars are typically plant sap, nectar, or aphid and coccid honeydew, which are all rich in sucrose ([@B15][@B16][@B17]). The sugar feeds are stored in the crop and diffuse into the thoracic midgut ([@B15], [@B18]).\n\nSuprapylarian *Leishmania* spp. include all members of the genus with the exception of the *Vianna* subgenus, in which parasite development is confined to the midgut and the foregut. Some general aspects regarding the development of Suprapylarian species appear to be consistent ([@B19]): the infective blood meal containing *Leishmania* amastigotes is passed into the abdominal midgut, where the blood is quickly retained inside the PM. The transformation of amastigotes to promastigotes occurs within 12 to 18\u00a0h. These early transformed promastigotes are termed procyclics and appear as short, ovoid, and only slightly motile forms. For the next 36 to 60\u00a0h, rapid multiplication of procyclic promastigotes within the digesting blood meal continues, followed by their transformation to a long, slender, more actively motile form termed nectomonads. By 60 to 72\u00a0h, coincident with the excretion of the digested blood meal, tremendous numbers of nectomonads are found packed in the anterior portion of the abdominal midgut, with many attached via their flagella to the epithelial cell microvilli ([@B19], [@B20]). By days 7 to 10, the anterior migration of promastigotes to the region of the thoracic midgut and stomodeal valve proceeds until a massive accumulation of parasites behind the valve is achieved. This migration is associated with the transformation of nectomonads into short, actively dividing forms called leptomonads that produce a mucin-like substance termed promastigote secretory gel (PSG) in which the mass of promastigotes in the thoracic midgut are imbedded ([@B21], [@B22]). Broad forms termed haptomonads can be found in attachment to the cuticular surface of stomodeal valve ([@B20]). The final development phase is termed metacyclogensis, which refers to the differentiation to metacyclic promastigotes, thought to originate from leptomonads ([@B21], [@B22]), though their transformation directly from nectomonads cannot be ruled out. Metacyclic promastigotes are short, slender, highly motile, and unattached forms with a flagellum at least twice the length of the cell body, and they are never seen in division. They are believed to be the only forms egested into the skin when an infected fly takes a blood meal ([@B23]).\n\nThere are thus a number of promastigote developmental stages, and distinct microenvironments in the fly to which they are likely adapted, that are absent when *Leishmania* are axenically grown *in vitro*. We used transcriptome sequencing (RNA-Seq) to profile the transcriptomes of *L.\u00a0major* during its transition from amastigotes to promastigotes in the vector, and also during its subsequent development as nectomonads and its maturation to the infective, metacyclic stage. Given that the vast majority of experimental studies are confined to use of cultured promastigotes, we also found it relevant to compare the transcriptomes of sand fly-derived and culture-derived metacyclics.\n\nIt should be noted that transcription profiles do not always reflect the true proteomic and metabolomic conditions, which is an inherent weakness in transcriptomic studies. Since proteomic and metabolomic studies are still technically challenging when working with such small amounts of starting material as can be obtained from sand fly midguts, RNA-Seq is currently the most useful approach to predict the genes that are associated with the development of the *Leishmania* parasite inside the vector.\n\nRESULTS AND DISCUSSION {#s1}\n======================\n\nExperimental design. {#s1.1}\n--------------------\n\n*L.\u00a0major* amastigotes were used to initiate infection in *P. duboscqi* sand flies, which are a natural vector for *L.\u00a0major*. The flies were infected via artificial feeding through a chick skin membrane on mouse blood seeded with tissue-derived amastigotes prepared from the footpad lesions of BALB/c mice. Flies were dissected and midguts harboring homogeneous populations of procyclic (PP), nectomonad (NP), and metacyclic promastigotes (MP) were collected and pooled on days 2, 4, and 15 post-blood meal (PBM), respectively ([Fig.\u00a01](#fig1){ref-type=\"fig\"}). RNA was extracted from these samples, as well as from lesion-derived amastigotes (AM) and from culture-derived metacyclic promastigotes (CMP). These procedures were repeated to obtain duplicate, independent samples of each developmental form. Of note, midgut-derived leptomonads or haptomonads were not analyzed, because it is so far not possible to obtain homogenous populations of these *in vivo* forms.\n\n![Morphological observations of the different *L.\u00a0major* developmental stages. Light microscope images are of promastigotes recovered from sand flies on days 2 (PP), 4 (NP), and 15 (MP) postinfection and stained with Hoechst fluorescent dye. Culture-derived metacyclics (CMP) were purified from stationary-phase cultures of promastigotes growing in CM199.](mbo0021732540001){#fig1}\n\nA total of 1.9 \u00d7 10^8^ reads were generated across the 10 experimental samples after trimming 23.3% to 26.6% of the reads due to poor-quality sequences (see [Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"} in the supplemental material). In samples from replicate A, 76.4% of the high-quality reads in the RNA extracted from footpad-derived amastigotes mapped to the *L.\u00a0major* genome. As expected, RNA extracted from the sand fly midguts contained a relatively low proportion of *L.\u00a0major* sequences: 21.2% of the total midgut RNA sequences on day 2, 14.1% on day 4, and 14.1% on day 15. The rest of the reads mapped to the sand fly or other eukaryotic genomes, such as fungi, that may be normal constituents of the gut microbiota. In RNA extracted from culture-derived metacyclics, 87.2% of the reads mapped to the *L.\u00a0major* genome. Similar numbers were observed for the RNA samples from replicate B ([Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.00029-17.9\n\nRNA sequence reads from the different samples. Download TABLE\u00a0S1, DOCX file, 0.1 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nRelationship between samples. {#s1.2}\n-----------------------------\n\nPrincipal component analysis (PCA) ([Fig.\u00a02A](#fig2){ref-type=\"fig\"}; see also [Fig.\u00a0S1B](#figS1){ref-type=\"supplementary-material\"} and [C](#figS1){ref-type=\"supplementary-material\"} in the supplemental material) and Pearson correlation heat maps ([Fig.\u00a02B](#fig2){ref-type=\"fig\"}) were used to study the relationship between the samples and to evaluate the reproducibility of the biological replicates. The PCA plot shows the top two principal components which explain most of the variance between samples in the data set, 23%, 21% and 15.49% for PC1, PC2, and PC3, respectively (see [Fig.\u00a0S1](#figS1){ref-type=\"supplementary-material\"} for PC3). The analysis revealed a high degree of similarity between the duplicate samples, indicating a low batch effect with no clustering based on biological replicate (shapes) but rather on the developmental stage (color) ([Fig.\u00a02A](#fig2){ref-type=\"fig\"}; see also [Fig.\u00a0S1B](#figS1){ref-type=\"supplementary-material\"} and [C](#figS1){ref-type=\"supplementary-material\"}). This was further supported by the scatterplots of normalized gene counts for each pair of biological replicates, showing high correlation values ranging from 0.79 to 0.95 ([Fig.\u00a0S1A](#figS1){ref-type=\"supplementary-material\"}). Therefore, no batch correction was applied in any further analysis. The data set could be divided into three distinct clusters: cluster I, grouping together procyclic and nectomonad promastigotes; cluster II, with both culture- and sand fly-derived metacyclics, and cluster III, with tissue-derived amastigotes. This observation suggested distinct variations in the transcriptome between the earlier stages of infection, namely, procyclics and nectomonads, and the mature metacyclic promastigote forms. This deviation was also observed with the Pearson correlation heat map ([Fig.\u00a02B](#fig2){ref-type=\"fig\"}), which separated MP from the earlier sand fly stages and clustered them more closely with tissue amastigotes. This suggests a transcriptomic preadaptation of the fully mature metacyclic forms to the intracellular environment that they will encounter following transmission by bite. Notably, both culture- and sand fly-derived metacyclics clustered together (cluster II), suggesting high similarity between the transcriptomes of both forms. However, slight separation between MP and CMP was observed when plotting PC1 versus PC3 and PC2 versus PC3 ([Fig.\u00a0S1B](#figS1){ref-type=\"supplementary-material\"} and [C](#figS1){ref-type=\"supplementary-material\"}), suggesting that the transcriptomes of metacyclics derived under these two conditions are close but not identical.\n\n10.1128/mBio.00029-17.7\n\nScatterplots indicating correlation of expression values across replicates and PCA plots of PC1 versus PC3 and PC2 versus PC3. (A) Correlation of expression values across replicates. For each of the five conditions considered, scatterplots show the Voom and quantile normalized log~2~ counts per million values for each pair of replicates. (B and C) PCA results, with plots of PC1 versus PC3 (B) and PC2 versus PC3 (C). The different colors indicate the different developmental stages in biological replicates A (triangles) and B (circles). Download FIG\u00a0S1, TIF file, 0.3 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n![Global transcriptomic profiles of *L.\u00a0major* parasites in the different developmental stages. RNA-Seq was performed on replicate samples of sand fly-derived procyclic (PP), nectomonad (NP), and metacyclic promastigotes (MP), purified culture-derived metacyclics (CMP), and footpad lesion-derived amastigotes (AM). (A) PCA results. The different colors indicate the different developmental stages in biological replicates A (triangles) and B (circles). (B) Heat map of hierarchical clustering analysis using the Pearson correlation, with correlation levels indicated by colors (inset). (C) Distribution of DEG between different sequential developmental stages *in vivo* are indicated in a bar graph. The box width depicts the number of DEG downregulated (purple) and upregulated (blue) at an adjusted *P*\u00a0value of 0.05, with the total number of down- and upregulated genes shown. The color shading indicates the proportion of genes with at least 4-fold (dark), between 2- and 4-fold (medium), or 2-fold differential expression (light). (D) Venn diagram showing the differentially and commonly expressed mRNAs between AM and each of the sand fly promastigote stages.](mbo0021732540002){#fig2}\n\nIdentification of differentially expressed genes between the four developmental stages. {#s1.3}\n---------------------------------------------------------------------------------------\n\nA list of differentially expressed genes (DEG) with a *P*\u00a0value of \\<0.05 was generated for each of the different pairwise comparisons of the developmental stages (see [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"} for the complete list of DEG). The total number of DEG are summarized in [Table\u00a0S2](#tabS2){ref-type=\"supplementary-material\"}; [Fig.\u00a02C](#fig2){ref-type=\"fig\"} shows DEG that are at least 2-fold or 4-fold different, and for only the comparison of stages that occur sequentially. A total of 1,212 DEG were associated with the transition from AM to PP, of which 691 and 521 were up- and downregulated in AM, respectively. Comparisons between the promastigote stages revealed smaller differences in mRNA abundance. We found 561 DEG between PP and NP, 470 between NP and MP, and 791 between MP and AM.\n\n10.1128/mBio.00029-17.2\n\nList of differentially expressed genes for each of the different pairwise comparisons of the developmental stages. Download DATA SET\u00a0S1, XLS file, 0.8 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.00029-17.10\n\nSummary of differentially expressed genes between the different samples. Download TABLE\u00a0S2, DOCX file, 0.04 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nVenn diagrams were generated to reveal the unique and common mRNA abundance patterns between the different developmental stages ([Fig.\u00a02D](#fig2){ref-type=\"fig\"}; see also [Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"}). AM had the highest number of unique genes in its expression profile, with 287 genes that were different from PP, NP, and MP ([Fig.\u00a02D](#fig2){ref-type=\"fig\"}). In contrast, 81 genes were uniquely up- or downregulated in PP, while 53 and 52 genes showed unique expression patterns in NP and MP, respectively ([Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.00029-17.8\n\nVenn diagrams for differentially and commonly expressed mRNAs between the developmental stages. Venn diagrams show differentially and commonly expressed mRNAs between *L.\u00a0major* developmental forms in relation to PP (A), NP (B), and MP (C). Download FIG\u00a0S2, TIF file, 0.2 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nOf the 1,212 genes that were significantly up- or downregulated during transformation of AM to PP, 581 genes showed expression patterns that were unique to PP, while the differences between AM and NP and between AM and MP were significantly fewer, with 264 and 249 DEG, respectively ([Fig.\u00a02D](#fig2){ref-type=\"fig\"}). These observations are in line with the PCA and Pearson correlation analyses, suggesting that *in vivo* promastigotes become transcriptionally closer to amastigotes as they mature to the mammalian-infective, metacyclic stage. Overall, the major shift in the transcriptome was, as expected, between AM and the different sand fly stages. However, a unique signature was associated with each of the promastigote forms, validating that the distinctive cellular morphologies upon which their developmental stage designations have been based do in fact reflect distinct gene developmental programs.\n\nBased on the DEG, we used Gene Ontology (GO) and KEGG enrichment analyses (see Materials and Methods) to identify the cellular processes associated with parasite development *in vivo* and culminating in their maturation into MP. Upregulation and downregulation of GO categories in pairwise comparisons between the different developmental stages are listed in [Data Set\u00a0S2](#dataS2){ref-type=\"supplementary-material\"}. Five GO categories were downregulated during the differentiation from AM to PP, among which were categories related to antioxidant activity (GO:0016209), cysteine peptidase (GO:0008234), and DNA catabolism (GO:0006308) ([Data Set\u00a0S2](#dataS2){ref-type=\"supplementary-material\"}). Eighteen GO categories were shown to be upregulated in PP compared to AM, including categories related to nucleosomes (GO:0000786), motility (GO:0031514), and synthesis of ATP (GO:0000275). Seven GO categories were downregulated in the transition from PP to NP, mainly related to the cell cycle (GO:0000786 and GO:0000776), and no significant upregulated GO categories were found. Only two groups were downregulated during the transition from NP to MP, and these were related to rRNA processing (GO:0006364 and GO:0032040). No significant upregulated GO categories were found in MP compared to NP. Fourteen GO categories were shown to be upregulated in MP compared to PP, with most being related to transport and homeostasis of nonessential amino acids, like proline, alanine, and glutamate (GO:0015193, GO:0022858, and GO:0002036), and to signal transduction (GO:0035556). Eight GO categories were downregulated in MP compared to PP, including categories related to the cell cycle (GO:0000786), ATP hydrolysis (GO:0015991), and protein heterodimerization (GO:0051258).\n\n10.1128/mBio.00029-17.3\n\nEnrichment of GO categories in pairwise comparisons between the different developmental stages. Download DATA SET\u00a0S2, PDF file, 0.1 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nWe also generated coexpression gene modules for mRNAs with similar patterns of abundance across the different developmental stages (see [Data Set\u00a0S4](#dataS4){ref-type=\"supplementary-material\"} for the list of modules and corresponding genes). Our analysis resulted in 207 gene modules with a median number of 31 genes per module (minimum of 10, maximum of 295). All of the 207 coexpression modules are presented in [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}. In addition, coexpression gene modules were enriched by using GO and KEGG tools (see [Data Set\u00a0S5](#dataS5){ref-type=\"supplementary-material\"} for the list of GO and KEGG categories in the modules). In the sections below, based on the top up- and downregulated mRNAs in each of the stages ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}), the most dynamic cellular processes are described. To highlight interesting coexpression modules, we generated median expression plots for modules enriched for specific functions of interest described below ([Fig.\u00a03](#fig3){ref-type=\"fig\"}). Each plot shown was selected because it included the greatest number of genes annotated with a particular structure or function. For example, while amastin genes are represented in a number of different coexpression modules, module 1 was chosen because it contains the largest number of amastin genes. The specific genes that are mentioned throughout our manuscript were significantly different (*P \\<* 0.05) between at least two stages (see [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.00029-17.4\n\nCoexpression modules for genes with similar mRNA abundance patterns across the different developmental stages. There were a total of 207 coexpression modules. Each line represents a gene, with genes of particular interest labeled with different colors and/or line patterns. The counts per million data represent the proportion of reads mapped to each *Leishmania* gene, multiplied by 10^6^. Download DATA SET\u00a0S3, PDF file, 0.5 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.00029-17.5\n\nList of modules and corresponding genes. Download DATA SET\u00a0S4, XLSX file, 0.6 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.00029-17.6\n\nList of GO and KEGG categories in the modules. Download DATA SET\u00a0S5, PDF file, 0.2 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n![Selected coexpression modules showing dynamics of cellular processes across *L.\u00a0major* developmental stages in the sand fly. Each line represents a plot of the median expression level of a selected gene or group of genes in coexpression modules relating to cell surface (A), metabolism (B), cell cycle and gene expression (C), and differentiation, stress, and autophagy (D). Values shown are the log~2~ counts per million and represent the proportion of reads mapped to each *Leishmania* gene, multiplied by 10^6^.](mbo0021732540003){#fig3}\n\nSurface genes. {#s1.4}\n--------------\n\nStrong stage regulation was observed in surface proteins between the different developmental forms. Ten out of the 15 most downregulated transcripts as the parasites differentiate into PP, 2\u00a0days PBM, were amastin, the amastigotes surface proteins ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, PP versus AM). We found 30 significantly regulated amastin-like mRNAs, most of these transcripts affiliated with modules that show a strong downregulation in PP and an increase in NP or MP ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 1, 2 and 28). Many amastin-like mRNAs returned to their AM expression levels already at the MP stage. The most strongly upregulated mRNAs in PP were the promastigote surface antigen proteins (PSA) on chromosome 12 ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, PP versus AM). The majority of the PSA transcripts (19 out of 22) showed higher expression in PP compared to the other promastigote forms, with 5 of these genes showing a sharp peak of abundance in PP ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 3). The coordinate upregulation of these genes in PP may suggest their role in the resistance of the transforming extracellular promastigotes to the proteolytic enzymes that are induced by blood feeding and that reach their peak 18 to 48\u00a0h post-blood meal ([@B24]). Transcripts for other surface proteins, such as those encoded by adjacent genes on chromosome 4 and a single gene on chromosome 21, increased significantly at the NP stage and remained elevated in MP ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}, module 5, and [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, NP versus PP). Another surface antigen-like protein on chromosome 5 was upregulated only in NP versus the AM stage ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 6).\n\nThe *Leishmania* surface proteinase of 63\u00a0kDa (GP63, or leishmanolysin) is encoded by 7 genes in *L.\u00a0major* and is thought to be a virulence factor involved in complement resistance and also interaction with and survival in macrophages. mRNAs of gp63-1, -2 and -4 on chromosome 10 (LmjF.10.0460, LmjF.10.0465, and LmjF.10.0480) were significantly upregulated in PP and continued to increase at the later stages ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}, module 4). In contrast, GP63 on chromosome 28 (LmjF.28.0570) showed a sharp peak at the MP stage ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 7). The abundant surface lipophosphoglycan (LPG) is another virulence factor for promastigote survival in both the sand fly and mammalian host. LPG3 is essential for the parasite survival due to its role in the biosynthesis and assembly of glycophospholipid (GPI)-anchored glycoconjugates, including GP63 and LPG ([@B25]). LPG3 transcripts were strongly upregulated in PP but significantly decreased in MP ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}, module 10), consistent with its upregulated expression in log-phase versus stationary-phase culture promastigotes ([@B25]). Surprisingly, other LPG biosynthetic genes did not show significant differences between the stages, apart from the galactosyltransferases that modify the phosphoglycan side chains of the LPG (SCGs), for which SCG1 was downregulated in PP and NP versus AM while SCG7 was specific to AM and strongly downregulated in all other stages ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}; see also [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 12 and 13). Two different mRNAs encoding SCG5 showed different patterns of abundance. LmjF.31.3090 was upregulated in PP and NP, and LmjF.31.3190 was downregulated in NP and MP compared to AM ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 11 and 154). SCAL (LmjF.34.0510), the arabinosyltransferase that modifies the LPG side chain oligosaccharides with arabinose, is abundant in amastigotes but remains significantly downregulated at the NP and MP stages ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 183). This was surprising, as it is known that LPG side chains are arabinosylated on MP and are thought to facilitate the detachment of the parasite from the midgut epithelium ([@B26]).\n\nMetabolism. {#s1.5}\n-----------\n\n*Leishmania* spp. utilize different carbon sources for the production of energy throughout their life cycle. In the insect, extracellular promastigotes utilize mainly sugar and amino acids, while intracellular amastigotes use amino acids and fatty acids ([@B27][@B28][@B31]). Combining coexpression module analysis with GO and KEGG pathway enrichment analyses, our studies confirmed that transport of different sugars and amino acids was higher in PP than AM, further increased from PP to NP, and reached their peak at MP ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}; see also [Data Sets\u00a0S5](#dataS5){ref-type=\"supplementary-material\"}, module 17, and S2, E and G). This pattern was shared between the glucose transporters GT1, GT2, and GT3 (LmjF.36.6300, LmjF.36.6290, and LmjF.36.6280, respectively) and the amino acid transporter AAP24 (LmjF.10.0720, LmjF.10.0715, and LmjF.10.0720, also known as AAT20 1 to 3). This transporter is specific for \"nonessential\" amino acids, such as proline and alanine, that are known to serve as carbon sources in *Leishmania* ([@B4], [@B30][@B31][@B33]). Moreover, the transporter has a role in maintaining the homeostasis of amino acids in the free amino acid pool, in which proline, alanine, and glutamate are the most abundant constituents ([@B32], [@B34]). The significant increase in acquisition of carbon sources throughout the development of promastigotes in the fly is consistent with recent studies showing that the differentiation of *L.\u00a0mexicana* to amastigotes is coupled with them entering into a \"metabolic stringent state,\" in which transport and utilization of glucose and amino acids are strongly suppressed, while metabolism of fatty acids via \u03b2-oxidation and production of glutamate via the tricarboxylic acid (TCA) cycle is increased ([@B4]). These observations suggest that promastigotes build a reservoir of carbon sources that can be utilized in the nutrient-poor intracellular environment.\n\nIn contrast to glucose transport, mRNAs for glucose metabolism showed a different pattern of abundance. Three out of four glyceraldehyde-3-phosphate dehydrogenase (GAPDH) variants were significantly more abundant in all promastigote stages versus AM, with two of them reaching their peak already at PP (LmjF.30.2980 and LmjF.35.4750) ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}; see also [Data Set\u00a0S5](#dataS5){ref-type=\"supplementary-material\"}, module 15) and a third (LmjF.30.2970) peaking at the NP stage. Conversely, the putative cytosolic GAPDH, LmjF.36.2350, was upregulated in AM and MP and significantly downregulated in PP and NP ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 18). Other glycolysis enzymes, such as the phosphoglucose isomerase (PGI) (LmjF.12.0530) and the putative hexokinase (LmjF.21.0240), were significantly upregulated in PP and NP compared to AM, with no significant difference between MP and AM (modules 15 and 145, respectively), suggesting that glycolysis is downregulated at the MP stage, possibly as a preadaptation to the intraphagosomal environment, where energy is produced by other carbon sources, such as fatty acids and amino acids. A main pathway of amino acid utilization in *Leishmania* is through the metabolism of glutamate ([@B32], [@B33]). Glutamate is metabolized by glutamate dehydrogenase (GDH), which converts glutamate to \u03b1-ketoglutarate to enter the TCA cycle and vice versa. Two mRNAs for GDH were significantly regulated in this study: LmjF.15.1010 was upregulated by 2.79-fold in PP compared to AM, remained low in NP, and was downregulated again in MP by 2.38-fold ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}). Another GDH (LmjF.28.2910) was strongly downregulated between AM and PP (6.29-fold), slightly increased in NP, and returned to its high AM abundance level at the MP stage ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 19). Saunders et al. in 2014 reported that metabolism of glutamate takes place in intracellular amastigotes via acetyl coenzyme A (CoA) in a compartmentalized and highly active TCA cycle ([@B4]). The strong upregulation of LmjF.28.2910 in MP to its high abundance level in AM suggests that this process already arises in the MP stage.\n\nWhile mitochondrial activity is suggested to increase in MP and AM, general metabolism of ATP seemed to be more pronounced in NP and PP. GO categories that are related to ATP metabolism (GO:0000275, GO:0015991, and GO:0046034) were upregulated in PP and NP but were decreased in AM and MP ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}; [Data Sets\u00a0S5](#dataS5){ref-type=\"supplementary-material\"}, module 16, and\u00a0S2, B and F).\n\nAlthough in *Leishmania* the use of fatty acids as a carbon source is predominantly associated with the intracellular amastigote stage, it was also evident in nondividing promastigotes in culture ([@B27], [@B35]). We observed that the biosynthesis of fatty acids as well as the catabolism of ketone bodies were strongly upregulated in NP. The mRNAs of two fatty acid elongase enzymes (LmjF.14.0700 and LmjF.14.0720) were significantly upregulated in NP compared to AM and PP, while a third, LmjF.14.0730, showed a significant peak in NP compared to all other forms ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 5 and 43). Likewise, two putative fatty acid desaturases (LmjF.36.6950 and LmjF.24.2250) and one putative acetyl-CoA carboxylase (LmjF.31.2970) were affiliated with modules 6, 92, and 43, respectively ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}), all of which are characterized by a significant peak at the NP stage. In addition, we observed a strong increase in the mRNAs of succinyl-CoA:3-ketoacid (LmjF.30.1930, LmjF.30.1940, and LmjF.33.2340) in the NP stage that remained elevated in MP ([Data Sets\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 17, 20, and 22, and\u00a0S2, G and I; GO:0046952). This enzyme participates in the catabolism of ketone bodies by converting acetoacetate to acetoacetyl-CoA under low-nutrient conditions.\n\nAdditional evidence that NP are experiencing starvation conditions was found in the abundance patterns of specific amino acid transporters. While mRNAs of many amino acid transporters accumulated throughout promastigote development, with a peak of abundance in MP (module 17), others, including some members of the aATP11 family and the arginine transporter AAP3 (LmjF.31.0870 and LmjF.31.0880), showed a sharp increase in the NP stage ([Fig.\u00a03B](#fig3){ref-type=\"fig\"}, modules 20 and 126; [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, NP versus PP). aATP11 and AAP3 were previously associated with a response to purine and amino acid starvations, respectively ([@B36][@B37][@B38]), reinforcing the suggestion that the NP stage is associated with one or a number of stress conditions.\n\nCell cycle and gene expression. {#s1.6}\n-------------------------------\n\nHistone proteins were among the most strongly upregulated transcripts during differentiation from AM to PP ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, PP versus AM). GO enrichment along with coexpression clustering analysis indicated that most mRNAs of histone proteins (categories GO:0000786 and GO:0006334, nucleosome and nucleosome assembly) were affiliated with modules 8, 23, and 24 ([Fig.\u00a03c](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"} and [S5](#dataS5){ref-type=\"supplementary-material\"}). The shared expression pattern in these modules is a significant increase in PP followed by a decrease in NP ([Data Set\u00a0S2](#dataS2){ref-type=\"supplementary-material\"}, B and C) and a more modest increase at the MP stage. The tight association between histone protein abundance and cell cycle in *Leishmania* has been previously described ([@B39]). The mRNAs of genes that are related to cell replication showed a similar pattern to histones: strongly upregulated in PP and downregulated in NP, followed by a modest upregulation in MP. These include cyclin6 (LmjF.32.3320), CDC20 (LmjF.24.1720), cell cycle sequence binding phosphoprotein (RBP33) (LmjF.35.0950), and cyclin-dependent kinase regulatory subunit (LmjF.32.3790) ([Fig.\u00a03C](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 25, 26, 28, and 163). CyclinA (LmjF.25.1470) was an exception, as it was elevated in AM and PP but significantly decreased in NP and MP ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 27). Consistently, all of the mRNAs for DNA polymerases were downregulated in the nectomonad stage, including the mitochondrial DNA polymerase beta (LmjF.08.0890) and the mitochondrial DNA polymerase I protein C (LmjF.14.0920) ([Fig.\u00a03C](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 28 and 163).\n\nTogether, the results suggest cell cycle arrest in NP, which is consistent with the absence of dividing nectomonads observed by light microscopic examination of stained cells ([@B21]). Surprisingly, there was an increase in most cell cycle-related mRNAs in the metacyclic stage. Gossage et al. ([@B21]) identified two separate growth phases during promastigote development *in vivo*, the procyclic and leptomonad forms, that appeared prior to or coincident with the appearance of MP. The increase of cell cycle-related mRNAs in MP may be due to incomplete degradation of mRNAs from a previous, cycling promastigote form (i.e., leptomonad) that was not isolated in this work. Another possible explanation for the increase in cell cycle mRNAs in MP is the occurrence of meiosis, which shares both DNA synthesis and DNA-packaging processes with mitosis. Sexual mating of *Leishmania* occurs in the sand fly vector ([@B40], [@B41]), and while the mating competent form(s) is yet to be directly identified, the recovery of hybrid progeny was greatest during the later stage of infection (days 10 to 14) ([@B41]). Meiosis homologue genes were previously found to be active in *Trypanosoma brucei* in the salivary glands of the tsetse fly ([@B42]). Three orthologs of these potential meiosis-specific genes (SPO11 \\[LmjF.16.0630\\], HOP1 \\[LmjF.36.1110\\], and DMC1 \\[LmjF.35.4890\\]) were found to be significantly downregulated in PP and most abundant in MP ([Fig.\u00a03C](#fig3){ref-type=\"fig\"}, modules 18 and 107; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 199).\n\nGO and KEGG enrichment analyses revealed that the majority of ribosomal constituents and other translation-associated genes (GO:0005840, GO:0006412, GO:0003735, GO:0006412, and MD:03010) are downregulated at the NP stage ([Fig.\u00a03C](#fig3){ref-type=\"fig\"}, module 29; [Data Sets\u00a0S3](#dataS3){ref-type=\"supplementary-material\"} [and S5](#dataS5){ref-type=\"supplementary-material\"}, modules, 32, 69, 113, 120, 148, 150, and 207). Conversely, a small number of translation-related mRNAs reached their peak of abundance at NP ([Data Sets\u00a0S3](#dataS3){ref-type=\"supplementary-material\"} [and S5](#dataS5){ref-type=\"supplementary-material\"}, modules 98 and 113). Generally, the results suggest that translation decreases as the parasites differentiate into PP, remains low at the NP stage, and begins to increase again in the MP stage, as a possible preadaptation to the intracellular conditions.\n\nDifferentiation, autophagy, and the stress response. {#s1.7}\n----------------------------------------------------\n\nThe loci on chromosome 23 that contain the hydrophilic acylated surface proteins (HASP) (LmjF.23.1040, LmjF.23.1060, LmjF.23.1070, LmjF.23.1082, and LmjF.23.1088) and the small hydrophilic endoplasmic reticulum-associated protein (SHERP) (LmjF.23.1050, LmjF.23.1080, and LmjF.23.1086) were previously shown to be important in metacyclogenesis and in macrophage invasion by MP ([@B43], [@B44]). The HASP abundance pattern was consistent with these findings, as it was specific to AM and MP, downregulated in PP, and reached its negative peak in NP ([Fig.\u00a03D](#fig3){ref-type=\"fig\"}, module 9; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 144). SHERP mRNAs were among the most abundant in the MP stage, but unlike HASP, they did not remain upregulated in AM ([Fig.\u00a03D](#fig3){ref-type=\"fig\"}, module 34; [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, MP versus AM, MP versus PP, and MP versus NP), suggesting that following the initial establishment of infection, only HASP continues to play a role in the infectious process in the mammalian host.\n\nBiopterin metabolism provides cofactors for lipid cleavage, hydroxylation of aromatic amino acids and synthesis of nitric oxide ([@B45][@B46][@B48]). Previous studies showed that in *Leishmania*, both folate and biopterin can be metabolized via the same enzyme, PTR1 ([@B49]). Strikingly, it was shown that mutants lacking PTR1, and thus deficient in tetrahydropterin (H~4~B), generated significantly more metacyclics in culture and were more virulent ([@B50]). Moreover, PTR1 mRNA and protein were shown to be stage regulated and decreased dramatically in stationary versus log phase promastigotes ([@B51], [@B52]), reinforcing the role of pteridine metabolism in metacyclogenesis. In the present study, two folate/biopterin transporters (LmjF.10.0380 and LmjF.06.0310) and one pteridine transporter (LmjF.06.1260) were significantly upregulated in NP and MP ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 5, 37, and 122; [Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, NP versus PP). A third folate/biopterin transporter (LmjF.10.0390) was AM specific, with low abundance in all promastigotes ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 2). While acquisition of biopterin and pteridine appeared strongly upregulated in NP and MP, the mRNA abundance for PTR1 (LmjF.23.0270) which reduces pteridine, was strongly downregulated in NP and only slightly increased in MP ([Fig.\u00a03D](#fig3){ref-type=\"fig\"}; [Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 116), suggesting that shortage of tetrahydropterin in NP may contribute to the induction of metacyclogenesis, consistent with what was previously reported ([@B50]).\n\nRecycling of proteins by autophagic mechanisms is associated with the metabolism in cells that are under stress conditions and/or undergoing a differentiation process ([@B53], [@B54]). In *L.\u00a0major*, autophagy was shown to be essential for metacyclogenesis and was significantly upregulated in culture metacyclics ([@B55], [@B56]). It was shown in a later study that recycling of the glycosome compartment via autophagy is an important aspect of this response ([@B57]). ATG8 is involved in completion and expansion of the autophagosome vesicles and was previously used as a marker for autophagy in *Leishmania* ([@B56][@B57][@B58]). In our study, putative ATG8 mRNAs on chromosomes 9 and 19 were among the most strongly upregulated in the NP stage ([Fig.\u00a03D](#fig3){ref-type=\"fig\"}; [Data Sets\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, modules 5 and 126, and\u00a0S1, NP versus PP). A ubiquitin-dependent, proteasome-related protein (LmjF.14.0310) was also specific to NP and significantly decreased in MP ([Data Set\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 39).\n\nAnother factor that may trigger protein recycling in *Leishmania* is purine starvation. *Leishmania* parasites do not synthesize purines *de novo* and must scavenge them from the environment ([@B59]). In response to purine starvation, the parasites undergo major metabolic and cellular changes, including cell elongation, cell cycle arrest, and upregulation of the nucleotide transport and purine salvage machinery ([@B36], [@B60]). Changes in the transcriptome of *L.\u00a0donovani* following purine starvation have been previously reported ([@B36]), and are reflected in the current studies. For example, one of the most upregulated genes in purine starved parasites was the membrane-bound acid phosphatase (MBAP2) which has a role in endosomal trafficking ([@B61]). The mRNA for one MBAP2, LmjF.28.2650, was among the most strongly upregulated in NP compared to all other stages. A second, LmjF.23.1170, was significantly upregulated only in NP versus AM ([Fig.\u00a03D](#fig3){ref-type=\"fig\"}, module 43; [Data Sets\u00a0S3](#dataS3){ref-type=\"supplementary-material\"}, module 41, and S1, NP versus PP), suggesting an increase in lysosome-related recycling processes in NP. Other similarities include the increase in amino acid transporters aATP11 and in autophagy-related genes in the NP and MP stages, as discussed in the previous sections. Together, these observations suggest that promastigotes differentiation to NP and to MP involves a response to one or a number of stress conditions (e.g., purine shortage), which trigger protein recycling via autophagy or proteasome pathways. The initiation and in some cases the confinement of the peak of these responses to NP indicate that these stress conditions may vary according to the time and place within the post-blood meal midgut anatomy.\n\nMetacyclogenesis: sand fly versus culture. {#s1.8}\n------------------------------------------\n\nWe compared the change in transcriptome between sand fly-derived PP and MP to the difference recently described between log-phase and metacyclic promastigotes in culture ([@B7]) ([Tables\u00a01](#tab1){ref-type=\"table\"} and [2](#tab1){ref-type=\"table\"}). Six of the top 10 upregulated mRNAs observed in culture with metacyclics versus log-phase promastigotes were not significantly upregulated when sand fly-derived MP and PP were compared ([Table\u00a01](#tab1){ref-type=\"table\"}). This included the Meta1 gene (LmjF.17.0890), previously associated with metacyclogenesis in culture ([@B62]), which was upregulated in all promastigote forms *in vivo* with a small peak in NP, indicating that this gene is not directly related to metacyclogenesis. Four genes were consistent between the two differentiation conditions, including the RNA-binding protein (LmjF.23.0730) and the phosphoinositide phosphatase (LmjF.22.0250). Only 3 out of the 10 genes that were strongly downregulated in culture metacyclics were also downregulated in sand fly-derived MP: two histone H4s (LmjF.36.0020 and LmjF.31.3180) and the ATPase subunit 9 (LmjF.21.0740). Most of the top 10 mRNAs that were upregulated in sand fly-derived MP versus PP were also significantly upregulated in culture metacyclics versus log-phase promastigotes, although to a lesser extent ([Table\u00a02](#tab2){ref-type=\"table\"}). This included the three SHERP mRNAs (LmjF.23.1050, LmjF.23.1080, and LmjF.23.1086) and an aATP11 transporter (LmjF.31.0350). The ABCA6 transporter (LmjF.11.1290), which was significantly increased in MP versus PP *in vivo*, was downregulated in metacyclic versus log-phase promastigotes from culture. Importantly, whereas 6 out of the 10 most downregulated mRNAs in MP versus PP were the promastigote surface proteins (PSAs) on chromosome 12, these genes were not downregulated in cultured metacyclics, and in some cases they were slightly increased. Thirty-three GO categories were found to be significantly downregulated in metacyclics versus log-phase promastigotes in culture (see [Table\u00a02](#tab2){ref-type=\"table\"} in Dillon et al. \\[[@B7]\\]). In contrast, only 8 GO categories were downregulated in MP versus PP *in vivo*, including those related to nucleosomes and ATP metabolism ([Data Set\u00a0S2](#dataS2){ref-type=\"supplementary-material\"}). Seven GO categories were found to be upregulated in metacyclics versus log-phase cells in culture, and most were related to signal transduction and protein phosphorylation ([@B7]). In contrast, 14 GO categories were shown to be significantly upregulated in MP versus PP *in vivo* ([Data Set\u00a0S2](#dataS2){ref-type=\"supplementary-material\"}, E), and the majority of these were related to nutrient uptake, while two were related to phosphorylation of nucleotides and signal transduction. Together, these discrepancies reflect unique factors influencing promastigote survival and differentiation *in vivo*, including the diet, physiology, and microbiota of the fly. Thus, the abundance of PSAs expressed by PP in the fly but not by log-phase promastigotes from culture suggests a role in protecting the parasites in the hydrolytic environment of the blood-fed midgut ([@B63]). In addition, the upregulation in mRNAs for amino acid transporters in sand fly-derived MP but not metacyclics from culture may reflect a response to a shortage in these amino acids in flies transitioning from blood to sugar meals, or to competition for nutrients by the gut microflora.\n\n###### \n\nThe top 10 up- or downregulated mRNAs between culture MP and log-phase promastigotes compared to the difference between sand fly MP and PP\n\n  Direction of regulation and gene ID    Product description                                  Fold change   \n  -------------------------------------- ---------------------------------------------------- ------------- -----------------------------------------\n  Upregulated in culture metacyclics                                                                        \n  \u2003\u2003\u2003\u2003LmjF.34.0070                       Ascorbate peroxidase (APX)                           3.61          NS[^a^](#ngtab1.1){ref-type=\"table-fn\"}\n  \u2003\u2003\u2003\u2003LmjF.17.0890                       META domain-containing protein (META1)               3.07          NS\n  \u2003\u2003\u2003\u2003LmjF.02.0460                       Voltage-dependent anion-selective channel putative   3.03          NS\n  \u2003\u2003\u2003\u2003LmjF.23.0730                       RNA-binding protein putative                         2.78          4.22\n  \u2003\u2003\u2003\u2003LmjF.12.0480                       Hypothetical protein unknown function                2.73          6.88\n  \u2003\u2003\u2003\u2003LmjF.16.0500                       Hypothetical protein unknown function                2.71          NS\n  \u2003\u2003\u2003\u2003LmjF.28.0980                       P27 protein putative                                 2.69          NS\n  \u2003\u2003\u2003\u2003LmjF.23.0780                       Hypothetical protein conserved                       2.68          8.46\n  \u2003\u2003\u2003\u2003LmjF.22.0250                       Phosphoinositide phosphatase                         2.63          5.92\n  \u2003\u2003\u2003\u2003LmjF.29.1350                       RNA-binding protein putative                         2.59          NS\n  Downregulated in culture metacyclics                                                                      \n  \u2003\u2003\u2003\u2003LmjF.31.3070                       Iron-zinc transporter protein-like protein (LIT1)    \u22123.13         NS\n  \u2003\u2003\u2003\u2003LmjF.33.1760                       Hypothetical protein unknown function                \u22122.91         NS\n  \u2003\u2003\u2003\u2003LmjF.35.1310                       Histone H4                                           \u22122.89         NS\n  \u2003\u2003\u2003\u2003LmjF.35.2130                       Hypothetical protein unknown function                \u22122.82         NS\n  \u2003\u2003\u2003\u2003LmjF.36.0020                       Histone H4                                           \u22122.72         \u22122.08\n  \u2003\u2003\u2003\u2003LmjF.35.2160                       Adenine aminohydrolase (AAH)                         \u22122.72         7.04\n  \u2003\u2003\u2003\u2003LmjF.14.0470                       Hypothetical protein conserved                       \u22122.69         NS\n  \u2003\u2003\u2003\u2003LmjF.31.3180                       Histone H4                                           \u22122.66         \u22122\n  \u2003\u2003\u2003\u2003LmjF.33.3240                       h1 histone-like protein                              \u22122.58         NS\n  \u2003\u2003\u2003\u2003LmjF.21.0740                       ATPase subunit 9 putative                            \u22122.58         \u22122.08\n\nNS, not significant; differences between growth stages were not statistically significant.\n\n###### \n\nThe top 10 up- and downregulated mRNAs between sand fly MP and PP compared to the change between MP and log-phase promastigotes from culture\n\n  Direction of regulation and gene ID   Product description                                                   Fold change   \n  ------------------------------------- --------------------------------------------------------------------- ------------- --------\n  Upregulated in MP                                                                                                         \n  \u2003\u2003\u2003\u2003LmjF.23.1084                      Hypothetical protein                                                  1.98          32.93\n  \u2003\u2003\u2003\u2003LmjF.23.1086                      Small hydrophilic endoplasmic reticulum-associated protein (SHERP2)   1.90          31.55\n  \u2003\u2003\u2003\u2003LmjF.23.1075                      Hypothetical protein                                                  1.93          29.00\n  \u2003\u2003\u2003\u2003LmjF.23.1050                      Small hydrophilic endoplasmic reticulum-associated protein (SHERP)    1.88          28.80\n  \u2003\u2003\u2003\u2003LmjF.23.1080                      Small hydrophilic endoplasmic reticulum-associated protein (SHERP1)   1.88          25.64\n  \u2003\u2003\u2003\u2003LmjF.33.1290                      Hypothetical protein conserved                                        1.72          24.29\n  \u2003\u2003\u2003\u2003LmjF.31.0350                      Amino acid transporter aATP11 putative (AAT1.4)                       1.29          24.12\n  \u2003\u2003\u2003\u2003LmjF.13.0190                      Hypothetical protein unknown function                                 1.68          19.81\n  \u2003\u2003\u2003\u2003LmjF.26.0160                      Nuclear *lim* interactor-interacting factor-like protein              2.18          10.42\n  \u2003\u2003\u2003\u2003LmjF.11.1290                      ATP-binding cassette protein subfamily A member 6 putative            \u22121.34         10.30\n  Downregulated in MP                                                                                                       \n  \u2003\u2003\u2003\u2003LmjF.12.1040                      Surface antigen protein putative                                      1.36          \u221216.60\n  \u2003\u2003\u2003\u2003LmjF.12.1060                      Surface antigen protein putative                                      1.31          \u221214.20\n  \u2003\u2003\u2003\u2003LmjF.12.1020                      Surface antigen protein putative                                      1.36          \u221211.77\n  \u2003\u2003\u2003\u2003LmjF.12.0910                      Promastigote surface antigen protein                                  1.32          \u221210.79\n  \u2003\u2003\u2003\u2003LmjF.12.0860                      Surface antigen protein putative                                      1.35          \u22129.75\n  \u2003\u2003\u2003\u2003LmjF.12.0920                      Promastigote surface antigen protein                                  1.35          \u22129.15\n  \u2003\u2003\u2003\u2003LmjF.14.0130                      Inosine-guanine nucleoside hydrolase putative                         1.31          \u22127.74\n  \u2003\u2003\u2003\u2003LmjF.15.1090                      Developmentally regulated protein putative                            1.09          \u22127.25\n  \u2003\u2003\u2003\u2003LmjF.06.0210                      Hypothetical protein conserved                                        \u22121.51         \u22125.94\n  \u2003\u2003\u2003\u2003LmjF.13.0870                      Mitochondrial processing peptidase alpha-subunit putative             \u22121.24         \u22125.86\n\nWhile the shifting transcriptional programs comparing metacyclics and propagating promastigotes displayed important differences depended on their *in vitro* or *in vivo* origins, a comparison between the metacyclics themselves revealed a remarkably similar profile of transcript abundance, with only 26 DEGs distinguishing the *in vitro* and *in vivo* forms ([Data Set\u00a0S1](#dataS1){ref-type=\"supplementary-material\"}, SFMP versus CMP). Such a high level of concordance was not observed in a recent microarray-based transcriptome comparison of *L.\u00a0infantum* promastigotes from stationary-phase culture and from the anterior midgut of *P.\u00a0perniciosus* ([@B8]), for which over 260 DEGs were reported. We suggest that, whereas relatively homogeneous populations of metacyclics from sand flies (\\>90%) and culture (\\>95%) were compared in the present studies, the *L.\u00a0infantum* promastigote stages compared previously were not identified and likely contained heterogeneous populations of promastigotes. Moreover, compared to RNA microarrays, RNA-Seq requires less input RNA without the need for an artificial amplification step or the avoidance of sand fly tissue. Among the consistencies found, the pteridine transporters LmjF.06.1260 and its ortholog in *L.\u00a0infantum*, LinJ.06.1320, were in each case significantly higher in sand fly metacyclics than in culture-derived forms. Pteridine transport and metabolism are crucial for *Leishmania* differentiation and virulence ([@B51], [@B52]).\n\nIn the current study, the most highly upregulated gene in sand fly MP compared to culture was LmjF.14.0440, which according to GO prediction encodes the enzyme flavin adenine dinucleotide-binding oxidoreductase, which participates in energy production via an oxidation-reduction process. The nucleoside transporter 1 (LmjF.15.1230) and the glucose transporters 1 and 2 (LmjF.36.6290 and LmjF.36.6300, respectively) were also increased in sand fly versus culture metacyclics, as was MBAP (LmjF.36.2590), which was previously associated with the response to purine starvation ([@B36], [@B60], [@B61]). All together, the transcriptional programs associated with metacyclogenesis in the anterior midgut of the fly and in stationary-phase culture show exceptional consistency, with *in vivo* forms likely driven to a more fully mature state by exposure to different and/or more severe stress conditions. These conditions may include the hyperosmolality associated with the high concentration of sugar diffusing from the crop into the anterior midgut and the exceptionally high density of promastigotes that accumulate behind the stomodeal valve, which along with the microbiota will compete for nutrients in the anterior gut.\n\nConclusions. {#s1.9}\n------------\n\nThis work presents the most thorough examination to date of the transcriptomes of *Leishmania* associated with the life cycle progression in the sand fly midgut from tissue amastigotes ingested with the blood meal to fully mature, infective-stage metacyclic promastigotes that are transmitted by bite back to the mammalian host. The unique transcriptomic profile of each of the different stages likely reflects their adaptation to distinct microenvironments in the vector. The transformation of AM to rapidly dividing extracellular PP in the blood-fed midgut was accompanied by the greatest degree of genetic reprogramming, with over 1,200 DEGs. A substantial part of this shift was devoted to altering the expression of genes encoding surface proteins, with the downregulation of amastins observed in parallel to a strong increase in promastigotes surface antigens located on chromosome 12. It is likely that these surface changes reflect at least in part the need for the extracellular parasite to protect itself from the digestive enzymes induced by blood feeding ([@B63]).\n\nOur studies are the first to provide information about the transcriptome of NP, a distinctive developmental form that has so far not been studied, and only rarely even identified, in populations of promastigotes obtained from culture. NP differentiate from PP and dominate the gut immediately following excretion of a digested blood meal. Their differentiation was associated with a strong reduction in cell cycle-related and ribosomal protein mRNAs, and by a strong upregulation of responses associated with nutrient stress, including amino acid transport, fatty acid biosynthesis, catabolism of ketone bodies, and protein recycling via autophagy. Some of these DEGs, like those resembling the response to purine starvation, were confined to NP and were not apparent in MP. Despite both stages arising in a post-blood meal nutritional environment and both receiving cues to enter cell cycle arrest, their distinctive morphologies and gene expression profiles nonetheless argue for their exposures to distinctive nutrient and stress conditions. Sugar meals are stored in the crop and passively diffuse into the thoracic midgut ([@B18]), such that the concentration of sugars is apt to be greater for promastigotes that have migrated anteriorly. Indeed, the availability of nutrients in the anterior gut is adequate for some NP to reenter the cell cycle as leptomonads prior to their differentiation to MP and their return to a resting state ([@B21]). The earlier differentiation of PP to nonreplicating NP that occurs in the posterior midgut following excretion of the blood meal remnants may be in response to more extreme starvation conditions.\n\nExtensive transcriptomic differences were found between the MP and the earlier promastigote stages *in vivo* that in most cases did not reflect the DEG in the comparison of log-phase and metacyclic promastigotes from culture ([@B7]). Since the transcriptomes of metacyclics obtained from flies and culture were themselves highly similar, the fact that the changes in gene expression that accompanied metacyclogenesis *in vivo* and *in vitro* were in such poor agreement suggests that the transcriptomes of the respective proliferative forms (PP and log-phase promastigotes) are apt to be highly dissimilar. Interestingly, Pearson correlation and principal component analysis suggested a greater similarity between MP and AM than to the other promastigote stages that would indicate a preadaptation process of the metacyclics to the mammalian host environment. For instance, many of the mRNAs that encode amastin, the main surface glycoprotein that is essential for amastigote intracellular survival ([@B64]), are strongly upregulated already in the MP stage. From the metabolic point of view, increased mitochondrial activity that is associated with amastigote survival and growth already occurs at the MP stage.\n\nMetacyclic promastigotes obtained from culture rather than from flies are overwhelmingly relied upon for experimental studies, and the conditions and methods that promote their *in vitro* differentiation and purification have been well-described ([@B25], [@B65]). It seems generally encouraging then that the transcriptomic profiles of the sand fly- and culture-derived metacyclics were highly similar, with only 26 mRNAs significantly differing in abundance. This is consistent with the generally high level of infectivity that culture-derived metacyclics display. For example, injection of as few as 100 purified, culture-derived *L.\u00a0major* metacyclics was able to reproducibly establish infection in the ear dermis of C57BL/6 mice ([@B66]). Nonetheless, the few genes that differ in abundance between these populations suggest that sand fly metacyclics are a more fully differentiated, infective stage and that direct infectivity comparisons using even lower doses or other mammalian hosts might reveal meaningful differences.\n\nMATERIALS AND METHODS {#s2}\n=====================\n\nSand flies and parasites. {#s2.1}\n-------------------------\n\nA laboratory-reared colony of *Phlebotomus duboscqi*, a natural vector for *L.\u00a0major*, was used in this study. The colony was initiated from field specimens collected in Mali. The *L.\u00a0major* Ryan strain used in this study was derived from a strain originally isolated from a lesion biopsy specimen of a laboratory worker accidentally exposed to sand flies that were experimentally infected with a strain of *L.\u00a0major* (WR2885) originating in Iraq ([@B67]). Promastigotes were grown at 26\u00b0C in complete medium 199 (CM199) supplemented with 20% heat-inactivated fetal calf serum (FCS), 100\u00a0U/ml penicillin, 100\u00a0\u00b5g/ml streptomycin, 2\u00a0mM [l]{.smallcaps}-glutamine, 40\u00a0mM HEPES, 0.1\u00a0mM adenine (in 50\u00a0mM HEPES), 5\u00a0mg/ml hemin (in 50% triethanolamine), and 1\u00a0mg/ml 6-biotin. Lesion amasigotes were obtained from BALB/c footpads that were initiated using metacyclic promastigotes, that were purified by centrifugation through Ficoll as described elsewhere ([@B68]). Mice were maintained in the National Institute of Allergy and Infectious Diseases (NIAID) animal care facility under specific-pathogen-free conditions and used under a study protocol approved by the NIAID Animal Care and Use Committee (protocol LPD 68E).\n\nSand fly infection and parasite recovery. {#s2.2}\n-----------------------------------------\n\nAmastigotes were recovered from the mouse footpads, and \\~3 \u00d7 10^7^ parasites were used for tissue amastigote RNA. A small portion of the amastigotes (\\~1 \u00d7 10^6^) was seeded in CM199 and metacyclics were purified from 6-day stationary cultures by using a Ficoll gradient. Four million footpad-derived amastigotes were mixed with 1 ml mouse blood for sand fly infection. Flies were fed through a chick skin membrane as described elsewhere ([@B41]). Midguts were dissected 2, 4, and 15\u00a0days PBM to obtain homogeneous populations of procyclic, nectomonad, and metacyclic promastigotes, respectively. Parasites were released by macerating midguts individually with a pestle (Kimble Chase, Vineland, NJ) in an Eppendorf tube containing 100\u00a0\u00b5l phosphate-buffered saline, and promastigote stages and numbers were determined by counting with a hemocytometer. A representative sample of DNA from each stage was stained using Hoechst fluorescent dye and observed under a Leica fluorescence microscope at 100\u00d7 magnification. Only midguts yielding the respective promastigote stages with \\>90% homogeneity were pooled for RNA. The samples were homogenized in 1\u00a0ml of Trizol reagent (Tri reagent; Molecular Research Center Inc.) and stored at \u221280\u00b0C until further processing. Two biological replicates originating from different footpad lesions or groups of infected flies were prepared for each developmental stage.\n\nRNA extractions. {#s2.3}\n----------------\n\nRNA was extracted from all samples by using a combination of two kits as follows: 0.2\u00a0ml of chloroform was added to the samples in Trizol and vigorously vortexed. Samples were incubated for 5\u00a0min at room temperature followed by 15\u00a0min of centrifugation at 15,000\u00a0\u00d7 *g*, 4\u00b0C. The upper aqueous phase was carefully collected and mixed with 70% ethanol--diethyl pyrocarbonate. The mix of RNA and ethanol was added to an RNeasy column (catalog number 74106; Qiagen), and RNA was purified according to the manufacturer's instructions, including a DNA digestion step. To confirm the lack of DNA in the samples, cDNA was produced with and without reverse transcriptase enzyme and used as the template for PCR amplification of G6PDH on *Leishmania* chromosome 34, as described in reference [@B41]. Amplifications were observed only on cDNAs produced with the reverse transcriptase.\n\nRNA-Seq*.* {#s2.4}\n----------\n\nDue to low RNA yields, the SMARTer Ultra Low Input RNA kit for sequencing (v. 3; Clontech Laboratories, Inc., Mountain View, CA) was used for cDNA synthesis with a template input amount of 5\u00a0ng. The Illumina TruSeq Stranded mRNA-Seq HT sample preparation kit (Illumina, San Diego, CA) and its workflow were used for preparation of dual-indexed transcriptome libraries. A normalized input amount of 20 ng total RNA was used for the library prep. Final library products were measured by quantitative PCR using the Kapa Illumina Library quantification kit (Kapa Biosystems, Boston, MA), pooled at equal molar amounts, and sequenced on two lanes of a HiSeq 2500 8-lane flow cell (Illumina) to produce paired 100-bp reads. Trimmomatic v. 0.32 ([@B69]) was used to remove any remaining Illumina adaptor sequence and low-quality bases from the reads, and FastQC v. 0.11.2 ([@B70]) was used to assess the quality of reads before and after trimming.\n\nRead mapping and quantification. {#s2.5}\n--------------------------------\n\nTopHat version 2.0.13 ([@B71]) was used to map trimmed RNA-Seq reads to the TriTrypDB v. 6.0 *L.\u00a0major* Friedlin reference genome ([@B72]). Mapped reads were then sorted with SAMtools v. 1.3 ([@B73]) and quantified using HTSeq v. 0.6.0 ([@B74]). For a complete description of the precise commands used for each of the above steps, refer to [Text S1](#textS1){ref-type=\"supplementary-material\"} in the supplemental material.\n\n10.1128/mBio.00029-17.1\n\nDescription of the precise commands used to generate the different steps in our study. Download TEXT\u00a0S1, PDF file, 0.9 MB.\n\nCopyright \u00a9 2017 Inbar et al.\n\n2017\n\nInbar et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nNext, counts for each sample were loaded into R/Bioconductor (R Core Team, 2016 \\[[@B75]\\]) and combined into a single-count matrix. Unexpressed genes and genes with low levels of expression, defined as having less than 1 read per million in at least two of the samples, were removed. Quantile normalization ([@B76]) was applied, followed by log~2~ transformation of the counts per million and mean-variance adjustment using the Voom program ([@B77]).\n\nSample quality assessment, differential expression, and GO analysis. {#s2.6}\n--------------------------------------------------------------------\n\nPCA and Pearson correlation-based heat maps created using the base R princomp and gplots heatmap.2 functions, respectively, were used to determine the similarity between samples and make sure no outliers or strong batch effects were present in the data. Samples clustered closely by condition with no evidence for a strong batch effect.\n\nThe generalized linear model (GLM)-based method of Limma ([@B78]) was used to detect differentially expressed genes for each pairwise contrast. *P*\u00a0values were adjusted for multiple testing using the Benjamini-Hochberg (BH) method. Genes with an adjusted *P*\u00a0value of \\<0.05 were considered differentially expressed. The goseq ([@B79]) package for R was used to detect over- or underrepresented GO terms among each set of differentially expressed genes, using the *L.\u00a0major* Friedlin GO annotations provided by TriTrypDB. In order to improve the statistical power to detect enriched functions, up- and downregulated DEG were analyzed separately ([@B80]). The BH multiple-testing correction was applied, and GO terms with an adjusted *P*\u00a0value of \\<0.05 were detected for each DEG contrast.\n\nCoexpression cluster analysis and enrichment analyses. {#s2.7}\n------------------------------------------------------\n\nA modified version of the weighted gene coexpression network analysis (WGCNA) pipeline ([@B81]) was used to detect groups of *L.\u00a0major* genes with similar expression patterns across the different developmental stages and growth conditions. Starting with the raw RNA-Seq counts for each sample, genes with low expression levels were removed, and read counts (counts per million) were log~2~ transformed, as described in the differential expression analysis section. For each pair of genes *i* and *j*, a similarity score *S*~*ij*~ was computed using a weighted combination of the Pearson correlation and Euclidean distance and was calculated as follows: *S* = SIGN (corr~*x*~) \u00d7 {\\|corr~*x*~\\| + \\[1 \u2212 log(dist~*x*~ + 1)\\]/max\\[log(dist~*x*~ +1)\\]/2}, where corr~*x*~ is the correlation matrix and dist~*x*~ is the matrix used in calculation of the Euclidean distance.\n\nThe Euclidean distance component of the similarity metric serves to penalize gene pairs with significantly different magnitudes of expression, thus helping to split up clusters with similar profiles but very different expression levels. The similarity matrix was then converted to an adjacency matrix by shifting the matrix to the range (0, 1) and raising the each similarity score to the power 4, helping to reduce the number of spurious correlations.\n\nTo detect coexpression clusters in the adjacency matrix, hierarchical clustering was applied to group genes by their similarity in expression, and a dynamic branch cut algorithm ([@B82]) was used to divide the resulting dendrogram into individual clusters. This resulted in a partitioning of the genes into 207 coexpression clusters (median size, 31 genes; minimum of 10, maximum of 295).\n\nEach coexpression cluster was tested for enrichment of GO ([@B83]) and KEGG ([@B84]) annotations using the goseq package ([@B79]). Coexpression cluster analysis was performed in R/Bioconductor (R Core Team, 2016 \\[[@B75]\\]). All of the codes used for the coexpression cluster analysis, including commands used to generate all figures for this section, were documented using Knitr ([@B85]) and RMarkdown ([@B86]) (see [Text S1](#textS1){ref-type=\"supplementary-material\"}).\n\nAccession number(s). {#s2.8}\n--------------------\n\nRaw sequence data are available at the NCBI Short Read Archive (SRA) under record SRP096578. Accession numbers for the individual samples are summarized in [Table\u00a0S1](#tabS1){ref-type=\"supplementary-material\"} in the supplemental material.\n\n**Citation** Inbar E, Hughitt VK, Dillon LAL, Ghosh K, El-Sayed NM, Sacks DL. 2017. The transcriptome of *Leishmania major* developmental stages in their natural sand fly vector. mBio 8:e00029-17. <https://doi.org/10.1128/mBio.00029-17>.\n\nThis work was supported in part by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Also, N.E.S., L.A.L.D., and V.K.H. were supported, in part, by a grant to N.E.S. (NIH/NIAID AI094773).\n"}
+{"text": "**Core tip:** This was a retrospective study aiming to evaluate the presence of the so called \"statin escape\" phenomenon among hyperlipidemic subjects attending a lipid clinic and elucidate any potential confounding factors. This study confirms the limited bibliography reporting on statin escape phenomenon and its quite high prevalence. However, due to the small number of eligible participants, we were not able to identify potential predictors for the statin-escape phenomenon or establish an association between statin escape and incidence of cardiovascular disease. In this context, further investigation on the underlying pathophysiology of this phenomenon and its potential clinical ramifications is required.\n\nINTRODUCTION\n============\n\nStatins remain the cornerstone therapy for primary and secondary cardiovascular prevention, mainly due to their ability to reduce low-density lipoprotein cholesterol (LDL-C)\\[[@B1]\\]. Nevertheless, a notable cardiovascular risk remains in statin-treated individuals, which has been attributed to other residual factors, such as hypertension, diet and adherence to therapy\\[[@B2]\\]. Recently, the so called \"statin escape\" phenomenon has been reported as an independent cardiovascular risk factor in patients with acute myocardial infarction on prolonged statin treatment\\[[@B3]\\]. This phenomenon was first described in small studies including patients with familial hypercholesterolemia\\[[@B4],[@B5]\\] and afterwards in the Expanded Clinical Evaluation of Lovastatin (EXCEL) study\\[[@B6]\\]. The latter reported an increase in LDL-C levels after the first year of statin treatment, despite a marked decrease in those levels 1 mo after treatment initiation\\[[@B6]\\]. So far there have been few reports on this phenomenon and its underlying mechanisms remain obscure\\[[@B5],[@B7]-[@B9]\\].\n\nThe aim of this study was to provide additional data on the possible statin escape phenomenon based on the experience of a lipid clinic and try to elucidate potential risk factors.\n\nMATERIALS AND METHODS\n=====================\n\nThis was a retrospective (from 1999 to 2013) observational study as previously described\\[[@B10]-[@B12]\\]. Briefly, dyslipidemic adults followed-up for \u2265 3 years in the Outpatient Lipid Clinic of the University Hospital of Ioannina in Greece were included. A complete assessment of serum lipid profile along with cardiovascular risk factors and concomitant treatment was available. The study protocol was approved by the Institutional Ethics Committee.\n\nDemographic characteristics as well as various clinical and laboratory data were recorded at the baseline visit, at 6 mo and the most recent visit. These included: (1) age, gender, and smoking status; (2) body mass index (BMI) and waist circumference; (3) fasting glucose levels and glycated hemoglobin (HbA1c); (4) blood pressure (BP); (5) estimated glomerular filtration rate (MDRD - eGFR); and (6) a complete fasting lipid profile, including total cholesterol (TCHOL), triglycerides (TGs), high-density lipoprotein cholesterol (HDL-C), LDL-C and non-high-density lipoprotein cholesterol (non-HDL-C). The methods of blood sample collection and biochemical assessments have been previously described\\[[@B10]\\].\n\nThe evaluation of adherence to medication was based on the Hellenic national e-prescription web database. Subjects were classified according to their compliance with treatment as good or poor compliers if they refill \u2265 or \\< 80% of their expected prescriptions over time, respectively. We excluded those individuals meeting one of the following criteria: Use of statin therapy at baseline visit, discontinuation of statin treatment at most recent visit, change in statin treatment during follow-up and poor compliance to treatment. Statin escape phenomenon was defined as an increase in subject LDL-C levels at the most recent visit by \\> 10% compared with the value at 6 mo following initiation of statin therapy\\[[@B8]\\].\n\nStatistical analysis\n--------------------\n\nContinuous variables were tested for normality by the Kolmogorov-Smirnov test and logarithmic transformations were performed if necessary. Data are presented as mean \u00b1 standard deviation (SD) and median \\[interquartile range (IQR)\\] for normal and non-normal distributed data, respectively. \u03c7^2^ tests were performed for categorical values. The difference of variables between \u2265 2 groups was assessed by analysis of variance (ANOVA) and *post hoc* least significant difference tests were used for the comparison of variables or ratios of interest between the groups. Paired sample *t* tests were performed to assess the change of variables within each study group. Analysis of covariance (ANCOVA) was performed to assess the difference of variables between 2 subject groups, after adjusting for their baseline values. Binary logistic regression was performed to elucidate potential predictors for statin escape phenomenon. Two tailed significance was defined as *P* \\< 0.05. Analyses were performed with the Statistical Package for Social Sciences (SPSS), v21.0 software (SPSS IBM Corporation, Armonk, New York, United States).\n\nRESULTS\n=======\n\nOf 1240 hyperlipidemic individuals, 181 were considered eligible for the present analysis (Figure [1](#F1){ref-type=\"fig\"}). Study participant baseline characteristics are shown in Table [1](#T1){ref-type=\"table\"}. Of 181 eligible subjects, 56 (31%) exhibited the statin escape phenomenon and 125 (69%) did not. There were no differences between these 2 groups apart from the higher baseline prevalence of coronary heart disease noticed in the escape group (7% *vs* 1%, *P* \\< 0.05). As shown in Table [1](#T1){ref-type=\"table\"}, there was no difference between the 2 groups regarding statin treatment. No participant received any non-statin lipid-lowering therapy (*i.e*., fibrate, ezetimibe). In addition, no difference was found regarding drugs interfering with cholesterol or statin metabolism (*i.e*., \u03b2-blockers, thiazides, pioglitazone, atypical antipsychotics, levothyroxine, clopidogrel or proton-pump inhibitors; Table [1](#T1){ref-type=\"table\"}).\n\n###### \n\nBaseline characteristics of study participants\n\n  **Variable**                                      **Escape group**   **Non-escape group**\n  ------------------------------------------------- ------------------ -----------------------------------\n  *n*                                               56                 125\n  Gender (male), %                                  43                 52\n  Current smoking, %                                9                  14\n  Age, yr                                           56 (51-63)         57 (49-65)\n  Waist, cm                                         97 (90-101)        98 (90-105)\n  SBP, mmHg                                         134 (127-146)      140 (129-150)\n  DBP, mmHg                                         83 (79-95)         87 (80-92)\n  Follow-up, yr                                     4 (3-6)            4 (4-7)\n  Metabolic syndrome, %                             39                 40\n  Hypertension, %                                   59                 57\n  Diabetes, %                                       11                 9\n  Stroke, %                                         5                  4\n  Coronary heart disease, %                         7                  1[a](#T1FN1){ref-type=\"table-fn\"}\n  Abdominal aortic aneurysm, %                      2                  0\n  Carotid stenosis \u2265 50%, %                         0                  2\n  Peripheral arterial disease, %                    0                  1\n  Statin therapy and median dose, % (median dose)                      \n  Atorvastatin                                      38 (20 mg)         34 (20 mg)\n  Rosuvastatin                                      29 (10 mg)         24 (10 mg)\n  Simvastatin                                       21 (40 mg)         26 (40 mg)\n  Fluvastatin                                       7 (80 mg)          6 (80 mg)\n  Pravastatin                                       0                  1 (40 mg)\n  \u03b2-blocker, %                                      9                  7\n  Thiazides, %                                      11                 19\n  Pioglitazone, %                                   4                  1\n  Antipsychotics, %                                 0                  1\n  Levothyroxine, %                                  4                  5\n  Clopidogrel, %                                    2                  2\n  Proton-pump inhibitors, %                         4                  4\n\nMedian follow-up duration = 4 years (IQR: 3-6 years). Values are expressed as median (IQR), unless percentages as shown.\n\n*P* \\< 0.05 for the comparison with the escape group. DBP: Diastolic blood pressure; IQR: Interquartile range; SBP: Systolic blood pressure.\n\n![Flow chart of subject eligibility.](WJEM-7-25-g001){#F1}\n\nBaseline lipid and metabolic profile did not differ between the 2 study groups (Table [2](#T2){ref-type=\"table\"}). Six months after the initiation of statin treatment, LDL-C levels were lower in the escape compared with the non-escape group \\[88 (78-97) mg/dL *vs* 109 (91-129) mg/dL, *P* \\< 0.01; Figure [2](#F2){ref-type=\"fig\"}\\]. On the contrary, LDL-C levels at the most recent visit were lower in the latter group \\[103 (96-118) mg/dL *vs* 94 (79-114) mg/dL, *P* \\< 0.01; Figure [2](#F2){ref-type=\"fig\"}\\]. Similarly, non-HDL-C levels were lower six months after the initiation of statin therapy in the escape compared with the non-escape group among non-diabetic individuals \\[107 (97-121) mg/dL *vs* 132 (115-153) mg/dL, *P* \\< 0.01; Table [2](#T2){ref-type=\"table\"}\\]. On the other hand, higher non-HDL-C levels were noticed in the former group at the most recent visit (Table [2](#T2){ref-type=\"table\"}). TRG significantly declined by 11% and 18% in the escape and non-escape group during follow-up, respectively (*P* \\< 0.01 respectively for the change within each group; Table [2](#T2){ref-type=\"table\"}). Despite the fact, that the non-escape group exhibited higher TRG levels than the escape group 6 mo after the initiation of statin therapy \\[104 (83-140) mg/dL *vs* 97 (69-117) mg/dL, *P* \\< 0.05\\], there was no difference between 2 groups regarding TRG levels at the most recent visit and the change of TRG levels during follow-up (*P* = NS for the comparison between 2 groups). On the other hand, HDL-C levels did not change during follow-up and were not different between 2 groups (Table [2](#T2){ref-type=\"table\"}).\n\n###### \n\nLipid and metabolic profile of study participants\n\n                                                                 **Baseline visit**   **Visit at 6 mo**                               **Most recent visit**\n  -------------------------------------------------------------- -------------------- ----------------------------------------------- -----------------------------------------------\n  TCHOL, mg/dL                                                                                                                        \n  Escape group                                                   258 (233-283)        162 (147-174)                                   182 (170-201)\n  Non-escape group                                               259 (235-295)        184 (162-206)[a](#T2FN3){ref-type=\"table-fn\"}   172 (154-193)[a](#T2FN3){ref-type=\"table-fn\"}\n  TG, mg/dL                                                                                                                           \n  Escape group                                                   117 (89-175)         97 (69-117)                                     104 (87-129)\n  Non-escape group                                               132 (99-181)         104 (83-140)[a](#T2FN3){ref-type=\"table-fn\"}    108 (79-130)\n  HDL-C, mg/dL                                                                                                                        \n  Escape group                                                   53 (47-68)           55 (43-64)                                      54 (48-68)\n  Non-escape group                                               53 (46-65)           52 (44-60)                                      56 (46-62)\n  LDL-C, mg/dL                                                                                                                        \n  Escape group                                                   174 (152-189)        88 (78-97)                                      103 (96-118)\n  Non-escape group                                               177 (152-205)        109 (91-129)[a](#T2FN3){ref-type=\"table-fn\"}    94 (79-114)[a](#T2FN3){ref-type=\"table-fn\"}\n  Non-HDL, mg/dL[1](#T2FN1){ref-type=\"table-fn\"}                                                                                      \n  Escape group                                                   204 (181-223)        107 (97-121)                                    127 (116-143)\n  Non-escape group                                               209 (182-241)        132 (115-153)[a](#T2FN3){ref-type=\"table-fn\"}   118 (102-137)[a](#T2FN3){ref-type=\"table-fn\"}\n  BMI, kg/m[2](#T2FN2){ref-type=\"table-fn\"}                                                                                           \n  Escape group                                                   27.3 (23.5-29.9)     27.2 (23.5-30.1)                                27.6 (24-30.2)\n  Non-escape group                                               27.9 (25.5-30.6)     28.3 (25.1-30.9)                                28.4 (25.5-31.5)\n  Fasting glucose, mg/dL                                                                                                              \n  Escape group                                                   95 (88-105)          95 (87-129)                                     95 (88-106)\n  Non-escape group                                               93 (87-103)          94 (88-104)                                     96 (89-106)\n  HbA1c, %[2](#T2FN2){ref-type=\"table-fn\"}                                                                                            \n  Escape group                                                   8.5 (6.7-8.6)        6.6 (5.6-5.9)                                   6.7 (6.6-7.1)\n  Non-escape group                                               8.4 (7.7-10.9)       6.7 (6.3-7.9)                                   6.9 (6.3-7.6)\n  MDRD-eGFR, mL/min per 1.73 m[2](#T2FN2){ref-type=\"table-fn\"}                                                                        \n  Escape group                                                   77 (69.6-86.7)       76.6 (67.9-84.8)                                76.5 (65.4-81)\n  Non-escape group                                               81 (70.7-91.4)       79.7 (69-89.7)                                  76.9 (65.5-85.7)\n\nValues are expressed as median (IQR). To convert from mg/dL to mmol/L multiply by 0.0555 for glucose, 0.02586 for TC, HDL-C, LDL-C, and 0.01129 for TG.\n\nNon-HDL-C levels refer to non-diabetic individuals (*n* = 164);\n\nHbA1c values refer to diabetic individuals (*n* = 17).\n\n*P* \\< 0.05 for the comparison with the escape group. BMI: Body mass index; MDRD-eGFR: Estimated glomerular filtration rate according to The Modification of Diet in Renal Disease (MDRD) Study equation; HbA1c: Glycated hemoglobin; HDL-C: High-density lipoprotein cholesterol; IQR: Interquartile range; LDL-C: Low-density lipoprotein cholesterol; non-HDL-C: Non-high-density lipoprotein cholesterol; TCHOL: Total cholesterol; TG: Triglycerides.\n\n![Change in low-density lipoprotein cholesterol levels during follow-up. ^a^*P* \\< 0.05 for the comparison between the 2 groups. LDL-C: Low-density lipoprotein cholesterol.](WJEM-7-25-g002){#F2}\n\nThere was no significant difference between the 2 groups regarding BMI change. As also shown in Table [2](#T2){ref-type=\"table\"}, glucose levels did not change during follow-up and were not different between the 2 groups. eGFR declined by 0.5 and 4.1 mL/min per 1.73 m^2^ in the escape and non-escape group, respectively (*P* \\< 0.05 respectively for the change within each group), but the difference between the 2 groups was not significant. The same was true for the change in diabetics' HbA1c levels (Table [2](#T2){ref-type=\"table\"}, *P* = NS for the comparison between the 2 groups).\n\nBinary logistic regression assessing baseline characteristics along with the changes in BMI, eGFR or HbA1c levels during follow-up did not reveal any significant predictor for the statin escape phenomenon.\n\nDuring a median follow-up of 4 years, 1 of 56 escape individuals and 6 of 125 non-escape subjects were diagnosed with incident cardiovascular disease (*P* = NS for the comparison between the 2 groups).\n\nDISCUSSION\n==========\n\nThe present report confirms the existence of statin escape phenomenon in clinical practice.\n\nTwo small studies including patients with familial hypercholesterolemia were the first to notice a paradox rebound cholesterol increase following statin dose increase\\[[@B4],[@B5]\\]. Since then, the EXCEL study along with others, have described this so called statin escape phenomenon\\[[@B3],[@B5]-[@B7]\\]. Our results showing an initial marked LDL-C reduction but followed by a \\> 10% LDL-C increase after prolonged statin treatment in subjects exhibiting the statin escape phenomenon are in line with the results of these studies\\[[@B3],[@B5],[@B7]\\]. Similar to previous studies, we did not find any predictors for this phenomenon\\[[@B3],[@B5],[@B7]\\]. A recent study showed that statin escape phenomenon not only exists, but also might be an independent predictor of cardiovascular disease\\[[@B3]\\]. The mechanisms attributing to the statin escape phenomenon have not yet been elucidated. The failure of statin therapy to decrease LDL-C levels on a long-term basis may be attributed to poor compliance with lipid-lowering treatment and diet. Particularly, an increased intake of cholesterol in the diet may contribute to intermittent variations in cholesterol levels. In addition, weight changes or a poor glycemic control in diabetic individuals could also cause a LDL-C increase, which could be wrongfully considered as statin escape phenomenon. After excluding subjects with these characteristics, one study concluded that only 1.2% of 161 study participants exhibited the statin escape phenomenon, although 28% of those were initially considered to meet the criteria of statin escape\\[[@B7]\\]. Despite the fact that no data regarding diet and exercise was available in our study, there was no significant difference between groups in terms of BMI change, glycemic control and kidney function.\n\nWe also assessed non-HDL-C levels in non-diabetic individuals considering that atherogenic dyslipidemia may alter LDL-C changes\\[[@B10]\\]. Statin escapers had higher non-HDL-C levels after prolonged statin therapy in comparison with non-escapers, although they had a higher non-HDL-C reduction 6 mo after treatment onset.\n\nAlthough we checked for adherence to therapy, our study might have included non-compliant individuals. It may be possible that the escapers adhered less to statin therapy and diet after seeing a large drop in their LDL-C levels. Another possible explanation for the statin escape phenomenon could be the concomitant therapy, since a variety of drugs could increase LDL-C lowering action of statins by inducting cytochromes CYP450-3A4 and 2C9\\[[@B13],[@B14]\\]. According to a few experimental studies, statin escape phenomenon could be attributed to a slow increase in the 5-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase activity or to an increase in proprotein convertase subtilisin kexin-like 9 (PCSK9) levels caused by prolonged statin therapy\\[[@B9],[@B15]-[@B19]\\].\n\nOur data suggest that statin escape phenomenon is indeed noticed in clinical practice, although its clinical significance remains uncertain. Patients with larger than anticipated initial LDL-C lowering should be carefully monitored.\n\nCOMMENTS\n========\n\nBackground\n----------\n\nA few studies have reported on the so called \"statin escape phenomenon\", which describes an increase in low-density lipoprotein cholesterol (LDL-C) levels after prolonged statin therapy despite an initial marked decrease. Statin escape phenomenon has been recently reported as an independent cardiovascular risk factor.\n\nResearch frontiers\n------------------\n\nVery few studies have reported on statin escape phenomenon and its underlying mechanisms remain obscure. The present study contributes to clarifying whether this phenomenon exists in clinical practice.\n\nInnovations and breakthroughs\n-----------------------------\n\nThis was a retrospective observational study with a small sample size. However, only the EXCEL study, which is the only randomized trial reporting on statin escape phenomenon and a retrospective cohort had larger samples. The small number of eligible participants did not allow any analysis to identify potential predictors for the statin-escape phenomenon. Additionally, due to small sample and low incidence of cardiovascular disease this study did not have the power to stablish an association between statin escape and incidence of cardiovascular disease. Nevertheless, this study confirms the limited bibliography reporting on statin escape phenomenon and its quite high prevalence (28%-31%).\n\nApplications\n------------\n\nThis study suggests that further investigation on the underlying pathophysiology of the statin escape phenomenon and its potential clinical ramifications is required.\n\nPeer-review\n-----------\n\nThis study is well written and the patients have been well selected although several variables could have influenced the results.\n\nInstitutional review board statement: This study was reviewed and approved by the Ethics Committee of the University Hospital of Ioannina, Greece.\n\nInformed consent statement: All study participants, or their legal guardian, provided informed written consent prior to study enrollment.\n\nConflict-of-interest statement: .Professor Elisaf M is an editorial member of World Journal of Experimental Medicine. The rest of authors have no conflict of interest relevant to this publication to declare\n\nData sharing statement: No additional data are available.\n\nManuscript source: Unsolicited manuscript\n\nSpecialty type: Medicine, research and experimental\n\nCountry of origin: Greece\n\nPeer-review report classification\n\nGrade A (Excellent): 0\n\nGrade B (Very good): 0\n\nGrade C (Good): C, C\n\nGrade D (Fair): 0\n\nGrade E (Poor): 0\n\nPeer-review started: November 10, 2016\n\nFirst decision: December 1, 2016\n\nArticle in press: January 3, 2017\n\nP- Reviewer: Corso G, Mihaila RG S- Editor: Ji FF L- Editor: A E- Editor: Lu YJ\n\n[^1]: Author contributions: Barkas F designed and performed the research and wrote the paper; Klouras E and Dimitriou T contributed to the analysis; Tentolouris N provided clinical advice; Elisaf M and Liberopoulos E supervised the report.\n\n    Correspondence to: Evangelos Liberopoulos, MD, PhD, FASA, FRSH, Assistant Professor of Internal Medicine, Department of Internal Medicine, School of Medicine, University of Ioannina, Stavrou Niarchou Avenue, 45110 Ioannina, Greece. <vaglimp@yahoo.com>\n\n    Telephone: +30-26-51099265 Fax: +30-26-51007016\n"}
+{"text": "1. Introduction {#sec1-sensors-19-00798}\n===============\n\nThe wearable device has been widely studied in recent years. In the related researches, it is essential to monitor the cardiac and physical activities in users with congestive heart issues at home. The electrocardiogram (ECG) is an important signal in monitoring the cardiac activity. A Holter ECG device is frequently applied to record long-term ECG signals and helps to find arrhythmic heartbeats within twenty-four hours \\[[@B1-sensors-19-00798]\\]. Another apparatus is called an Event Recorder that is able to record one-minute ECG signals when a user feels uncomfortable on the chest region \\[[@B2-sensors-19-00798]\\]. Currently, the commercial Holter ECG apparatus or Event Recorder only record the ECG signals, and is lacking in the real time analysis of the ECG signals. Moreover, the user has been encouraged to neither do any severe exercises nor take a shower, because these two kinds of devices do not have a high ingress protection rating.\n\nAn ECG patch is a wearable device, which not only records ECG signal, but also shows some cardiac information on the smart phone in real time \\[[@B3-sensors-19-00798]\\]. However, its function for the arrhythmic beat detection is not better than the ECG Holter analysis. The problem is that the ECG patch is used under a body-motion condition, which is easily coupled with some noise, like motion artifact and electromyogram (EMG). Therefore, how to cancel the noise in real time would help to develop the applications of ECG patch. In the measurement of ECG signals, it is vulnerable to various noises and interferences, such as the EMG, 50 or 60Hz power line noise (PLn), baseline drift and measurement noise and so on. In \\[[@B4-sensors-19-00798],[@B5-sensors-19-00798]\\], it was found that for complex QRS wave of a standard ECG signal, its power is within the frequencies lower than 30Hz, and its peak power is within the range between 4 Hz and 12 Hz. Generally, the high frequencies in the PLn and EMG noise are able to cover up the characteristic of an ECG signal such that the QRS wave group cannot be located accurately \\[[@B5-sensors-19-00798]\\]. Besides, it is known that ECG signals affected by noises will lead to erroneous diagnosis by physicians. Therefore, it is crucial to remove the PLn and EMG noise in the ECG monitoring and clinical diagnosis.\n\nTraditionally, ECG noise cancellation methods applied a low-pass filter \\[[@B6-sensors-19-00798]\\] to remove the high-frequency components in noise while a high-pass filter and adaptive filter are used to rid of low-frequency vibrations, such as baseline drift \\[[@B7-sensors-19-00798],[@B8-sensors-19-00798]\\] and respiratory interference \\[[@B9-sensors-19-00798]\\]. Since most of the noises in ECG signals are distributed at the high-frequency components, and the traditional low-pass filters cannot solve the problem that signal and noise co-exist within the same bandwidth. Augustyniak \\[[@B10-sensors-19-00798]\\] used the time-frequency transform to rescale the ECG coupled with different noises and then eliminated the noises to enhance the signal noise ratio. Rundo et al. \\[[@B11-sensors-19-00798]\\] applied the bio-inspired nonlinear system to cancel the noise in an ECG signal and then to align the ECG and pulse wave. Some ECG denoising methods employ the frequency decomposition technique, such as the wavelet transform \\[[@B12-sensors-19-00798]\\], the empirical mode decomposition (EMD) \\[[@B13-sensors-19-00798]\\] and the ensemble empirical mode decomposition (EEMD) \\[[@B14-sensors-19-00798]\\]. The EMD is a pre-processing algorithm of Hilbert-Huang Transform (HHT) which was introduced by Huang et al. \\[[@B15-sensors-19-00798]\\] with a capability to perform time-frequency transformation for non-linear and non-stationary signals. In recent years, the EMD has been widely applied in biomedical signal processing, such as ECG \\[[@B13-sensors-19-00798],[@B14-sensors-19-00798]\\], EMG \\[[@B16-sensors-19-00798]\\] and blood pressure waves \\[[@B17-sensors-19-00798]\\].\n\nThe basic idea for ECG noise cancellation consists of two main stages. First, an ECG signal is decomposed by the EMD into a set of IMFs through which the noisy IMFs are discriminated and discarded. Second, the ECG is reconstructed by the retained IMFs. By this doing, the noise in the ECG signal is reduced or cancelled. To be successful in the EMD-based noise cancellation, a fundamental problem is to estimate the noise energy appropriately in each IMF. In addition, the problem of mode mixing occurs during the EMD decomposition. The problem of mode mixing comes from either an IMF mingled with signals of different scales or other IMF appeared in the combination of other components. To overcome the mode mixing problem in the EMD, the EEMD \\[[@B18-sensors-19-00798]\\] was proposed with an additive white noise into the ECG signal. The EEMD is able to reduce the mode mixing effect on the next IMF scale. The EEMD has been widely applied in noise cancellation. For example, Chang et al. presented a scheme to cancel the white noise in ECG signals where noisy low-order IMFs were removed by a predefined threshold \\[[@B12-sensors-19-00798]\\]. Jenitta et al. used the zero-crossing ratio of adjacent IMFs to discriminate noisy IMFs by its noise energy \\[[@B19-sensors-19-00798]\\]. Yannis et al. proposed a scheme according to the energy of the first-order IMF through which noise cancellation was performed among IMFs for ECG signals \\[[@B20-sensors-19-00798]\\]. Kumaravel et al. presented a genetic algorithm to determine the noise energy threshold in the first-order IMF for the PLn cancellation in ECG signals \\[[@B21-sensors-19-00798]\\].\n\nThe methods described above generally deal with white noise and the PLn only. Besides, All of them estimates the noise energy in the first-order IMF. However, only considering the noise energy in the first-order IMF might not be appropriate especially when the signal energy of the related IMF is high. This will lead to over-cancellation and result in distortion in the reconstructed ECG signal. Consequently, this paper considers noise energies in every IMFs through the noise magnitude spectrum.\n\nAs we know, very few approaches to ECG noise cancellation are based on noise magnitude spectrum. In this paper, two-stage discrimination scheme is proposed to estimate and cancel the PLn and EMG noise in ECG signals according to IMFs' noise magnitude spectrum where the EMD, the EEMD and the grey spectral noise estimation (GSNE) are employed. In the first stage, the EMD decomposes the input ECG signal into IMFs. Then the GSNE is used to estimate noise in IMFs and calculate the related noise energies through its noise magnitude spectrum. By a user-defined threshold, noisy IMFs are identified and put into the second stage. In the second stage, the noisy IMFs are reconstructed and decomposed by the EEMD. Then the IMFs are rechecked in a similar manner to the first stage. If an IMF is considered as noisy, it is discarded. The procedure is repeated for each IMF. At last, the ECG signal is reconstructed with all retained IMF components. To evaluate the performance, a noise energy ratio in dB ($NER_{dB}$) is employed in this paper.\n\nThis paper is organized as follows. [Section 2](#sec2-sensors-19-00798){ref-type=\"sec\"} describes the ECG signals with additive noise and the ECG signal decomposition algorithms, the EMD and the EEMD. [Section 3](#sec3-sensors-19-00798){ref-type=\"sec\"} introduces the GSNE based on the first-order grey model of one variable, GM(1,1) model \\[[@B22-sensors-19-00798],[@B23-sensors-19-00798]\\], and then describes how the GSNE is applied to ECG noise cancellation. Next, the proposed GSNC scheme is introduced in [Section 4](#sec4-sensors-19-00798){ref-type=\"sec\"}. In [Section 5](#sec5-sensors-19-00798){ref-type=\"sec\"}, the proposed GSNC scheme is justified by forty-three datasets from the MIT-BIH database \\[[@B24-sensors-19-00798]\\]. Discussions about the EMD, the EEMD and the proposed GSNC are given in [Section 6](#sec6-sensors-19-00798){ref-type=\"sec\"}. Finally, the conclusion is given in [Section 7](#sec7-sensors-19-00798){ref-type=\"sec\"}.\n\n2. The ECG Signal, EMD and EEMD {#sec2-sensors-19-00798}\n===============================\n\nIn this section, the ECG signals with additive noise, the EMD and the EEMD for the ECG signals used in the proposed GSNC scheme are described in the following.\n\n2.1. ECG Signal {#sec2dot1-sensors-19-00798}\n---------------\n\nIn this paper, the real ECG signals are from the cardiac arrhythmia database in the MIT-BIH database \\[[@B24-sensors-19-00798]\\] from which datasets are selected. Each dataset has a continuous period of 30 minutes with the sampling frequency of 360 Hz. The ECG signals in the database were processed by the Butterworth filter with bandwidth from 0.3 Hz to 40 Hz. The processed ECG signal is considered as clean ECG signal, $s\\left( k \\right)$. The PLn and EMG noise are added into $s\\left( k \\right)$ where the PLn is generated from the sinusoidal wave with frequencies from 59.5 Hz to 60.5 Hz and the sampling frequency of 360 Hz. As for the EMG noise, its bandwidth is from 100 Hz to 500 Hz with the sampling frequency 1000 Hz. The additive noise, the PLn or EMG noise, is denoted as $n\\left( k \\right)$. And the noisy ECG signal is denoted as $x\\left( k \\right)$ which is expressed as $$x\\left( k \\right) = s\\left( k \\right) + n\\left( k \\right).$$\n\nIn this paper, the signal-to-noise ratio ($SNR$) from $- 5{\\ {dB}}$ to $20{\\ {dB}}$ is under consideration whose definition is given as $$SNR = \\frac{\\sum_{k = 1}^{N_{s}}s^{2}\\left( k \\right)}{\\sum_{k = 1}^{N_{s}}n^{2}\\left( k \\right)},$$ where $N_{s}$ is the length of $s\\left( k \\right)$.\n\n2.2. The EMD Algorithm {#sec2dot2-sensors-19-00798}\n----------------------\n\nThis subsection briefly reviews the EMD which will be applied to decompose an ECG signal in the proposed GSNC scheme. For details, one may consult \\[[@B15-sensors-19-00798]\\]. The EMD algorithm consists of the following steps.Step 1.Find the local maxima and minima in $x\\left( k \\right)$.Step 2.Obtain the upper envelope by the local maxima and the lower envelope by local minima, respectively.Step 3.Calculate the average of the upper and lower envelops, $m\\left( k \\right)$.Step 4.Find the difference signal $d\\left( k \\right) = x\\left( k \\right) - m\\left( k \\right)$.Step 5.Check if $d\\left( k \\right)$ is a zero-average process. If yes, then stop and treat $d\\left( k \\right)$ as the first-order IMF (IMF 1), denoted as $c_{1}\\left( k \\right)$; otherwise, replace $x\\left( k \\right)$ with $d\\left( k \\right)$ and go back to Step 1.Step 6.Calculate the residual signal $r\\left( k \\right) = x\\left( k \\right) - c_{1}\\left( k \\right)$.Step 7.Replace $x\\left( k \\right)$ with $r\\left( k \\right)$ and repeat Step 1 to Step 6 to find the second-order IMF (IMF 2), i.e., $c_{2}\\left( k \\right)$.Step 8.Repeat Step 1 to Step 7 till $c_{M}\\left( k \\right)$ is obtained where $M$ is the total number of IMFs.\n\nAfter the EMD, the original signal can be expressed as $$x\\left( k \\right) = \\sum\\limits_{i = 1}^{M}c_{i}\\left( k \\right) + r\\left( k \\right),$$ where $r\\left( k \\right)$ is generally considered as $c_{M + 1}\\left( k \\right)$.\n\n2.3. The EEMD Algorithm {#sec2dot3-sensors-19-00798}\n-----------------------\n\nHere, the EEMD algorithm is briefly reviewed. For details, one may consult \\[[@B18-sensors-19-00798]\\]. The EEMD algorithm is given in the following. Step 1.Add a white noise sequence $w\\left( k \\right)$ into the target signal $x\\left( k \\right)$, i.e., $x_{1}\\left( k \\right) = x\\left( k \\right) + w\\left( k \\right)$. In this study, noise with $SNR = 5{\\ {dB}}$ is used which will be verified in [Section 5.1](#sec5dot1-sensors-19-00798){ref-type=\"sec\"}.Step 2.Apply the EMD algorithm to decompose $x_{1}\\left( k \\right)$, as described in [Section 2.2](#sec2dot2-sensors-19-00798){ref-type=\"sec\"}.Step 3.Repeat Steps 1 and 2 until the predefined number of trials, $N_{T}$, is reached. Each trial uses the same noise power level. Then a set of IMF components $c_{ij}\\left( k \\right)$ is obtained where $i$ is the iteration number and $j$ is the order of IMFs.Step 4.Calculate the ensemble average of $c_{ij}\\left( k \\right)$ as follows $${EEMD}_{c_{j}{(k)}} = \\ \\frac{1}{N_{T}}\\sum\\limits_{i = 1}^{N_{T}}c_{ij}\\left( k \\right),$$ where $N_{T}$ is the total number of trials.\n\n3. Application of GSNE to ECG Noise Cancellation {#sec3-sensors-19-00798}\n================================================\n\nIn this section, the noise estimation based on a grey model is introduced. [Section 3.1](#sec3dot1-sensors-19-00798){ref-type=\"sec\"} briefly reviews the first-order grey model with one variable, denoted as GM(1,1). Next, the grey spectral noise estimation (GSNE) based on GM(1,1) model is proposed in [Section 3.2](#sec3dot2-sensors-19-00798){ref-type=\"sec\"}. Then the way to determine noisy IMFs by the GSNE is given in [Section 3.3](#sec3dot3-sensors-19-00798){ref-type=\"sec\"}.\n\n3.1. GM(1,1) Model {#sec3dot1-sensors-19-00798}\n------------------\n\nIn this subsection, GM(1,1) model is briefly reviewed. For details, one may refer \\[[@B22-sensors-19-00798],[@B23-sensors-19-00798]\\]. The GM(1,1) modeling is described in the following. Given a non-negative sequence $\\left\\{ {x\\left( k \\right),\\ 1 \\leq k \\leq K} \\right\\}$, then $x\\left( k \\right)$ is put into the first-order accumulated generating operation (1-AGO) to convert into a new data sequence $x^{(1)}\\left( k \\right)$ as $$x^{(1)}\\left( k \\right) = \\sum\\limits_{i = 1}^{k}x\\left( i \\right).$$\n\nBy $x\\left( k \\right)$ and $x^{(1)}\\left( k \\right)$, a grey difference equation is formed as $$x\\left( k \\right) + az^{(1)}\\left( k \\right) = b,$$ for $2 \\leq k \\leq K$, where parameters $a$ and $b$ are called the developing coefficient and the grey input, respectively; and $z^{(1)}\\left( k \\right)$ is the background value and defined as $$z^{(1)}\\left( k \\right) = 0.5\\left\\lbrack {x^{(1)}\\left( k \\right) + x^{(1)}\\left( {k - 1} \\right)} \\right\\rbrack.$$\n\nLet $$\\mathbf{y} = \\begin{bmatrix}\n{x\\left( 2 \\right)} \\\\\n{x\\left( 3 \\right)} \\\\\n \\vdots \\\\\n{x\\left( K \\right)} \\\\\n\\end{bmatrix}\\ $$ and $$\\mathbf{B} = \\begin{bmatrix}\n{- z^{(1)}\\left( 2 \\right)} & & 1 \\\\\n{- z^{(1)}\\left( 3 \\right)} & & 1 \\\\\n{\\vdots \\ } & \\vdots & \\\\\n{- z^{(1)}\\left( K \\right)} & & 1 \\\\\n\\end{bmatrix}\\ $$ Then (6) can be written as $$\\mathbf{y} = \\mathbf{B}\\begin{bmatrix}\na \\\\\nb \\\\\n\\end{bmatrix}$$ where parameters $a$ and $b$ are found by $$\\begin{bmatrix}\na \\\\\nb \\\\\n\\end{bmatrix} = \\left( {\\mathbf{B}^{T}\\mathbf{B}} \\right)^{- 1}\\mathbf{B}^{T}\\mathbf{y}.$$\n\nIt can be shown that the solution of $x^{(1)}(k$) is given as $$x^{(1)}\\left( k \\right) = \\left\\lbrack {x\\left( 1 \\right) - \\frac{b}{a}} \\right\\rbrack e^{- a{({k - 1})}} + \\frac{b}{a}\\ .$$ By the first-order inverse accumulated generating operation (1-IAGO), the estimate of $x\\left( k \\right)$, $\\hat{x}\\left( k \\right)$, is obtained as $$\\hat{x}\\left( k \\right) = x^{(1)}\\left( k \\right) - x^{(1)}\\left( {k - 1} \\right).$$ In the GM (1, 1) modeling, the minimum number of samples is 4, i.e., *K* = 4.\n\n3.2. Grey Spectral Noise Estimation {#sec3dot2-sensors-19-00798}\n-----------------------------------\n\nIn this subsection, the grey spectral noise estimation (GSNE) based on GM(1,1) is introduced where an additive signal model is assumed, that is, $x\\left( k \\right) = s\\left( k \\right) + n\\left( k \\right)$ for $1 \\leq k \\leq L$, where $s\\left( k \\right)$, $n\\left( k \\right)$, and $L$ stand for the signal component, noise component and the length of $x\\left( k \\right)$, respectively. The block diagram of GSNE is shown in [Figure 1](#sensors-19-00798-f001){ref-type=\"fig\"} whose implementation steps are given as follows.\n\n1.  Level up $x\\left( k \\right)$ by a constant $C$, i.e., $\\left. x\\left( k \\right)\\leftarrow x\\left( k \\right) + C \\right.$ such that the condition $x\\left( k \\right) > 0$ is met.\n\n2.  Divide $x\\left( k \\right)$ into $N_{ss\\ }$ subsets as $\\left\\{ {x_{i}\\left( k \\right),\\ 1 \\leq i \\leq N_{ss\\ }} \\right\\}$, for $1 + \\left( {K - 1} \\right)\\left( {i - 1} \\right) \\leq k \\leq K + \\left( {K - 1} \\right)\\left( {i - 1} \\right)\\ $, where *K* is the number of data used in GM(1,1) model. [Figure 2](#sensors-19-00798-f002){ref-type=\"fig\"} indicates how $x\\left( k \\right)$ is divided into $N_{ss\\ }$ subsets with *K* = 4 where the square refers to the overlapped sample.\n\n3.  For each subset $i$, obtain the estimate of $x_{i}\\left( k \\right)$, ${\\hat{x}}_{i}\\left( k \\right)$, by the GM(1,1) model described in [Section 3.1](#sec3dot1-sensors-19-00798){ref-type=\"sec\"}. Then, ${\\hat{x}}_{i}\\left( k \\right)$ is considered as the estimate of $s_{i}\\left( k \\right)$, ${\\hat{s}}_{i}\\left( k \\right)$, that is, ${\\hat{s}}_{i}\\left( k \\right) = {\\hat{x}}_{i}\\left( k \\right)$. Finally, calculate the estimation error of GM(1,1) model as $\\ e_{i}\\left( k \\right) = \\ x_{i}\\left( k \\right) - {\\hat{x}}_{i}\\left( k \\right)$.\n\n4.  Since the additive noise $n_{i}\\left( k \\right)$ is not equal but related to the estimation error $e_{i}\\left( k \\right)$. Consequently, ${\\hat{n}}_{i}\\left( k \\right)$ is estimated as $\\alpha e_{i}\\left( k \\right)$ where $\\alpha > 0$ is the user-defined scaling factor and determined by experiences.\n\n5.  Apply Fast Fourier Transform (FFT) on ${\\hat{n}}_{i}\\left( k \\right)$, i.e.,$\\ \\hat{N}\\left( f \\right) = {FFT}\\left\\{ {\\hat{n}\\left( k \\right)} \\right\\}$ and find the magnitude of $\\hat{N}\\left( f \\right)$, $\\left| {\\hat{N}\\left( f \\right)} \\right|$ where $\\hat{n}\\left( 1 \\right) = \\hat{n}\\left( 2 \\right)$ is used.\n\n6.  Calculate the standard deviation of $\\left| {\\hat{N}\\left( f \\right)} \\right|$, $\\ \\sigma_{|{\\hat{N}{(f)}}|}$, as an indicator of noise energy.\n\nNote that the energy preserving property in Parseval's theorem \\[[@B25-sensors-19-00798]\\] which proves the sum of the square of ${\\hat{n}}_{i}\\left( k \\right)$, i.e., energy of ${\\hat{n}}_{i}\\left( k \\right)$, is related to the sum of the square of $\\left| {\\hat{N}\\left( f \\right)} \\right|$ as $$\\ \\sum\\limits_{i = 1}^{L}\\left\\lbrack {\\ {\\hat{n}}_{i}\\left( k \\right)} \\right\\rbrack^{2} = \\frac{1}{L}\\ \\sum\\limits_{i = 1}^{L}\\left\\lbrack \\left| {\\hat{N}\\left( f \\right)} \\right| \\right\\rbrack^{2}.$$\n\nSince the standard deviation $\\sigma_{|{\\hat{N}{(f)}}|}$ is a statistics of $\\left| {\\hat{N}\\left( f \\right)} \\right|$ related to noise energy, it thus can be used as an indicator of noise energy in the IMF under consideration.\n\n3.3. Noisy IMF Determination by the GSNE {#sec3dot3-sensors-19-00798}\n----------------------------------------\n\nIn this section, the determination of noisy IMFs and the ECG noise cancellation based on the GSNE are described. In the experiment, we randomly select 100 ECG signal segments with the continuous duration of 10 seconds from the mitdb/100 dataset. Then $\\sigma_{|{\\hat{N}{(f)}}|}$ is calculated for each of the ECG signals with and without noises. With $SNR = 5{\\ {dB}}$, [Figure 3](#sensors-19-00798-f003){ref-type=\"fig\"} shows the differences of $\\sigma_{|{\\hat{N}{(f)}}|}$ in IMF 1, i.e., $c_{1}\\left( k \\right)$, where ${ECG}_{raw}$ denotes the original ECG signal,${\\ {ECG}}_{PLn}$ the ECG signal with the PLn and ${ECG}_{EMG}$ the ECG signal with EMG noise, respectively. As shown in [Figure 3](#sensors-19-00798-f003){ref-type=\"fig\"}, $\\sigma_{|{\\hat{N}{(f)}}|}$ with and without noise has apparently different levels. Besides, the fluctuation is very small both for cases with and without noise. Thus, it gives us a hope to discriminate noisy and non-noisy $c_{i}\\left( k \\right)$ by a predefined threshold $\\tau$. According to our experiments, $\\tau = 10^{- 4}$ works well for most cases. If $\\sigma_{|{\\hat{N}{(f)}}|} > \\tau$, the related $c_{i}\\left( k \\right)$ is considered as noisy and is rechecked further in the proposed GSNC scheme. With the idea, $c_{i}\\left( k \\right)$ is modified as ${\\overline{c}}_{i}\\left( k \\right)$ which is given as $${\\overline{c}}_{i}\\left( k \\right)\\  = \\left\\{ \\begin{matrix}\n{c_{i}\\left( k \\right)\\ } & {,\\ \\sigma_{{|{\\hat{N}{(f)}}|},\\ i} < \\tau} \\\\\n{\\ 0\\ } & {,\\sigma_{{|{\\hat{N}{(f)}}|},\\ i} > \\tau} \\\\\n\\end{matrix} \\right.,$$ for $i = 1,\\ 2,\\ \\cdots,M$.\n\nBy ${\\overline{c}}_{i}\\left( k \\right)$, the ECG signal after noise cancellation, denoted as $\\widetilde{y}\\left( k \\right)$, is reconstructed, as follows, $$\\widetilde{y}\\left( k \\right) = \\sum\\limits_{i = 1}^{M}{\\overline{c}}_{i}\\left( k \\right) + r\\left( k \\right)\\ .$$\n\nIn order to justify the fixed threshold $\\tau = 10^{- 4}$ works well for most cases, further experiments are conducted with six MIT-BIH cardiac arrhythmia datasets which are mitdb/100, mitdb/105, mitdb/108, mitdb/203, mitdb/223 and mitdb/228. In the experiments, ECG signal segments with the continuous duration of 10 seconds are randomly selected in the six datasets where $SNR = 5{\\ {dB}}$ both for PLn and EMG noise. The results are shown in [Table 1](#sensors-19-00798-t001){ref-type=\"table\"} for the PLn and [Table 2](#sensors-19-00798-t002){ref-type=\"table\"} for the EMG noise, respectively. As shown in [Table 1](#sensors-19-00798-t001){ref-type=\"table\"}, the $\\sigma_{|{\\hat{N}{(f)}}|}$ of $x\\left( k \\right)$ decreases as the order of IMF increases and most cases are determined correctly except for mistaken cases. The noisy IMFs considered as clean are shadowed according to the threshold $\\tau = 10^{- 4}$. The mistaken cases found in IMF 1 are in datasets mitdb/100 and mitdb/223. For those cases, clean IMFs are considered as noisy ones. These mistakes will be rechecked in the second stage of the proposed GSNE scheme described later in [Section 4](#sec4-sensors-19-00798){ref-type=\"sec\"}. The mistaken cases found in IMF 4 and IMF 5 consider noisy IMFs as clean ones. The cases are not a major concern since they have little impact on ECG noise cancellation. Thus, they are neglected in this paper.\n\nAs for the EMG noise, [Table 2](#sensors-19-00798-t002){ref-type=\"table\"} shows most of the clean and noisy cases are determined correctly except some cases. Two clean cases are mistaken as noisy happened in IMF 1 in mitdb/100 and in mitdb/223. For those noisy IMFs, the proposed GSNC scheme introduced later in [Section 4](#sec4-sensors-19-00798){ref-type=\"sec\"} will recheck them further. Similar to the PLn cases, most of the noisy cases are considered as clean found in IMF 4 and IMF 5. Since those cases have little effect on the performance for ECG noise cancellation, thus we do not deal with them further. By the results, shown in [Table 1](#sensors-19-00798-t001){ref-type=\"table\"} and [Table 2](#sensors-19-00798-t002){ref-type=\"table\"}, the fixed threshold $\\tau = 10^{- 4}$ is justified working well for most cases.\n\n4. The Proposed GSNC Scheme {#sec4-sensors-19-00798}\n===========================\n\n[Figure 4](#sensors-19-00798-f004){ref-type=\"fig\"} shows the flowchart of the proposed two-stage ECG noise cancellation scheme based on the GSNE. The proposed scheme is called grey spectral noise cancellation (GSNC). The working flow of the proposed GSNC scheme is described as follows. In the first stage, the input ECG signal is decomposed into a set of IMFs by the EMD. Then the noise spectral energy of IMFs are estimated by the GNSE and the magnitude standard deviation of each noise spectral energy, $\\sigma_{|{\\hat{N}{(f)}}|}$, is calculated. Next, $\\sigma_{|{\\hat{N}{(f)}}|}$ is compared with a user-defined threshold $\\tau$. If $\\sigma_{|{\\hat{N}{(f)}}|} > \\tau$, the IMF is considered as noisy. Due to the IMFs may mix each other, the second stage is performed to avoid mistakes. In the second stage, the noisy IMFs considered in the first stage are reconstructed and decomposed by the EEMD. The decomposed IMFs are then checked by the GNSE as in the first stage. If $\\sigma_{|{\\hat{N}{(f)}}|} > \\tau$, the corresponding IMF is discriminated as noisy and discarded. Finally, the ECG is reconstructed with all retained IMFs. This finishes the procedure.\n\nTo evaluate the performance of the proposed GSNC scheme, a performance index to evaluate the improvement based on the noise energy ratio in dB ($NER_{dB}$) is employed which is defined as $$NER_{dB} = 10\\log\\left\\lbrack \\frac{\\sum_{k = 1}^{N_{s}}\\left( {x\\left( k \\right) - s\\left( k \\right)} \\right)^{2}}{\\sum_{k = 1}^{N_{s}}\\left( {\\widetilde{y}\\left( k \\right) - s\\left( k \\right)} \\right)^{2}} \\right\\rbrack,\\ $$ where $s\\left( k \\right)$ is the true ECG signal; $x\\left( k \\right)$ and $\\widetilde{y}\\left( k \\right)$ are the input noisy ECG signal and the reconstructed ECG signal after noise cancellation; $N_{s}$ is the total number of samples in $x\\left( k \\right)$.\n\n5. Results {#sec5-sensors-19-00798}\n==========\n\nIn this section, the proposed GSNC scheme is justified by forty-three datasets from the MIT-BIH cardiac arrhythmia database. In [Section 5.1](#sec5dot1-sensors-19-00798){ref-type=\"sec\"}, the energy of white noise added in the EEMD is investigated. In [Section 5.2](#sec5dot2-sensors-19-00798){ref-type=\"sec\"}, two types of noise, the PLn and EMG noise, are considered in the experiments.\n\n5.1. Effect of White Noise in the EEMD {#sec5dot1-sensors-19-00798}\n--------------------------------------\n\nIn this subsection, the effect of white noise energy on the EEMD employed in the proposed GSNC is investigated. It is observed that the energy of white noise added in the EEMD affects the performance of noise cancellation. Consequently, several levels of white noise energy are added into the EEMD to find an appropriate one. In the experiments, three $SNR$ used in the EEMD are 2 dB, 5 dB and 10 dB where an ECG signal with EMG noise of $SNR = 5\\ {dB}$ is under consideration. Besides, the number of trials $N_{T} = 100$ is employed in the EEMD. [Figure 5](#sensors-19-00798-f005){ref-type=\"fig\"} shows parts of the original ECG signal and the ECG signals after noise cancellation by the proposed GSNC scheme. The corresponding $NER_{dB}$ are 7.84 dB, 7.92 dB and 5.26 dB for the cases with $SNR =$2 dB, 5 dB and 10 dB in the EEMD, respectively. By the $NER_{dB}$, it suggests that the white noise with $SNR = 5\\ {dB}$ results in the best performance of ECG noise cancellation. Consequently, it is employed in the EEMD which is applied in the second stage of the proposed GSNC for better performance.\n\n5.2. Results for the PLn and EMG Noise {#sec5dot2-sensors-19-00798}\n--------------------------------------\n\nThe PLn and EMG noise are generated artificially and added into the original ECG signals whose $SNR = 5{\\ {dB}}$ and all forty-three MIT-BIH datasets are involved. As for the GSNE, the parameters $K = 4$ and $\\alpha = 1$ are set. In the experiment, the proposed GSNC scheme is also compared with the traditional noise cancellation schemes based on the EMD and the EEMD. As with the first experiment, the PLn with $SNR = 5{\\ {dB}}$ is considered. [Table 3](#sensors-19-00798-t003){ref-type=\"table\"} shows the $NER_{dB}$ for the proposed GSNC and compared schemes. In [Table 3](#sensors-19-00798-t003){ref-type=\"table\"}, the EMD scheme has the $NER_{dB}$ ranged from\u221212.83 dB to 7.96 dB with the largest standard deviation of 5.21 dB among the compared schemes. For the EEMD scheme, its $NER_{dB}$ ranges from \u22125.67 dB to 7.90 dB in the given datasets with a smallest standard deviation of 3.21 dB. As for the proposed GSNC scheme, its $NER_{dB}$ ranges from \u22125.33 dB to 9.97 dB. On the average of $NER_{dB}$, the proposed GSNC scheme has the best result (3.34$\\  \\pm \\ $4.03 dB) which is followed by the EEMD (2.20$\\  \\pm \\ $3.21 dB) and EMD (0.10$\\  \\pm \\ $5.21 dB) schemes.\n\nIn the second experiment, the EMG noise of $SNR = 5{\\ {dB}}$ is under study. [Table 4](#sensors-19-00798-t004){ref-type=\"table\"} gives the $NER_{dB}$ for the proposed GSNC, EMD and EEMD schemes. In this experiment, a similar result to the first experiment is found. In the EMD scheme, the $NER_{dB}$ ranges from \u22128.03 dB to 12.19 dB with the largest standard deviation of 4.52 dB. The EEMD scheme has the range of $NER_{dB}$ from \u22126.66 dB to 12.47 dB with the smallest standard deviation of 4.27 dB. And the $NER_{dB}$ for the proposed GSNC scheme is within \u22125.14 dB to 13.21 dB. On the average of $NER_{dB}$, the proposed GSNC scheme has the highest value (6.14$\\  \\pm \\ $4.29 dB) followed by the EEMD scheme (3.85$\\  \\pm \\ $4.27 dB) and the EMD scheme (3.52$\\  \\pm \\ $4.52 dB).\n\nTo investigate the proposed GSNC scheme further, the ECG signal with various noise energies in the PLn and EMG noise are under consideration. The SNR for the PLn and EMG noise used in the experiments are \u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB. Six datasets from MIT-BIH database are selected, including mitdb/100, mitdb/105, mitdb/108, mitdb/203, mitdb/223 and mitdb/228. For the case of PLn, [Figure 6](#sensors-19-00798-f006){ref-type=\"fig\"} shows the overall average of $NER_{dB}$ for the proposed GSNC scheme and the compared EMD and EEMD schemes. On the average of $NER_{dB}$ for all cases with the corresponding standard deviation, the EMD scheme obtains 4.16$\\  \\pm \\ $1.79 dB, 5.24$\\  \\pm \\ $1.81 dB, 3.82$\\  \\pm \\ $4.32 dB, 4.53$\\  \\pm \\ $2.57 dB, 4.72$\\  \\pm \\ $2.88 dB and \u22120.22$\\  \\pm \\ $8.29 dB for $SNR =$ \u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. The results for the EEMD scheme are 5.07$\\  \\pm \\ $2.13 dB, 6.41$\\  \\pm \\ $0.99 dB, 4.46$\\  \\pm \\ $1.98 dB, 3.39$\\  \\pm \\ $3.17 dB, 1.67$\\  \\pm \\ $4.68 dB and \u22123.09$\\  \\pm \\ $3.36 dB for $SNR = \\ $\u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. And the proposed GSNC scheme has 4.5$\\  \\pm$1.69 dB, 7.28$\\  \\pm \\ $1.41 dB, 5.44$\\  \\pm \\ $1.14 dB, 5.88$\\  \\pm \\ $2.07 dB, 6.08$\\pm \\ $0.91 dB and 5.07$\\  \\pm \\ $3.02 dB for $SNR =$ \u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. As shown in [Figure 6](#sensors-19-00798-f006){ref-type=\"fig\"}, the proposed GSNC scheme always has better $NER_{dB}$ than the compared schemes except the case $SNR = - 5{\\ {dB}}$.\n\nIn the case of EMG noise, [Figure 7](#sensors-19-00798-f007){ref-type=\"fig\"} depicts the averages of $NER_{dB}$ for different SNR. The $NER_{dB}$ and standard deviations for the EMD scheme are 8.88$\\  \\pm \\ $2.59 dB, 7.29$\\  \\pm \\ $3.75 dB, 5.13$\\  \\pm \\ $4.20 dB, 4.94$\\  \\pm \\ $4.21 dB, 2.13$\\  \\pm \\ $1.97 dB and \u22123.81$\\  \\pm \\ $10.24 dB for SNR \u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. For the EEMD scheme, the average $NER_{dB}$ with corresponding standard deviations are 10.52$\\  \\pm \\ $7.50 dB, 7.50$\\  \\pm \\ $1.93 dB, 6.50$\\  \\pm \\ $2.48 dB, 2.46$\\  \\pm \\ $5.32 dB, \u22122.13$\\  \\pm \\ $4.66 dB and \u22123.98$\\  \\pm \\ $4.56 dB for $SNR = \\ $\u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. The results for the proposed GSNC scheme are 10.89$\\  \\pm \\ $2.68 dB, 7.40$\\  \\pm \\ $2.56 dB, 7.32$\\  \\pm \\ $1.97 dB, 6.93$\\  \\pm \\ $2.70 dB, 2.41$\\  \\pm \\ $0.84 dB and \u22120.72$\\  \\pm$ 9.0 dB for $SNR =$ \u22125 dB, 0 dB, 5 dB, 10 dB, 15 dB and 20 dB, respectively. For all cases, the proposed GSNC scheme is superior to the compared schemes by $NER_{dB}$ except slightly less than the EEMD scheme in the case of ${SNR} = 0{\\ {dB}}$. By [Figure 6](#sensors-19-00798-f006){ref-type=\"fig\"} and [Figure 7](#sensors-19-00798-f007){ref-type=\"fig\"}, it suggests that the proposed GSNC scheme generally shows better performance when compared with the EMD and EEMD schemes in terms of average $NER_{dB}$, both for the PLn and EMG noise.\n\n6. Discussion {#sec6-sensors-19-00798}\n=============\n\nThe EMD acts like a filter-bank and has no strict bandwidth restriction with the IMFs. The frequency range of each IMF is adaptive, depending on the original signal content. Generally, the bandwidths of the PLn and EMG noise are assumed in the ranges 59.5--60.5 Hz and 100--500 Hz, respectively. Thus the PLn is considered as a lower frequency noise while the EMG noise is considered as a medium and higher frequency noise. Note that the noise energy is mainly distributed in the low order IMF components after the EMD. Because of different bandwidths, the performance for the EMG noise is better than that for the PLn in the EMD scheme for ECG noise cancellation. As shown in [Table 3](#sensors-19-00798-t003){ref-type=\"table\"} and [Table 4](#sensors-19-00798-t004){ref-type=\"table\"}, the average $NER_{dB}$ for the PLn is 0.10 dB and 3.52 dB for the EMG noise.\n\nCompared with the EMD, the EEMD has more concentrated band-limit IMF components. With the iterative EMD computation, the average of IMF with the same order yielded a sharper band transition than a single EMD-derived IMF, that is, the transition band overlap between adjacent IMFs is narrower than the EMD result. In other words, with the same filter specification the EEMD acts like a higher-order filter while the EMD works like a lower-order filter. Consequently, the performance of the EEMD scheme is better than the EMD scheme, as shown in [Table 3](#sensors-19-00798-t003){ref-type=\"table\"} and [Table 4](#sensors-19-00798-t004){ref-type=\"table\"} where the average $NER_{dB}$ for PLn and the EMG noise by the EEMD scheme are 2.20 dB and 3.85 dB, respectively. They are higher than the corresponding $NER_{dB}$ by the EMD scheme. However, the EEMD pays the price of computational complexity, that is, it takes more time to cancel the ECG noise. This hinders the realization in an ECG patch if only the EEMD is applied in the ECG noise cancellation. On contrarily, the proposed GSNC scheme decomposes the input ECG signal by the EMD in the first stage. When $\\sigma_{{|{\\hat{N}{(f)}}|},\\ i} < \\tau$, the proposed GSNC scheme stops further decomposition. This makes the proposed GSNC scheme possible to be embedded in an ECG patch. Moreover, the proposed GSNC scheme employs two-stage discrimination for noisy IMFs while the conventional EMD and EEMD schemes use one-stage discrimination. Thus, the proposed GSNC scheme is expected to have better performance since suspicious IMFs in the first stage can be rechecked in the second stage to avoid mistakes while the one-stage EMD and EEMD scheme fails to.\n\nIn \\[[@B17-sensors-19-00798]\\], Liu et al. showed that the energy of a lower frequency in the EMG noise is found in the first-order IMF component. In other words, the EMD or the EEMD like a filter-bank is able to decompose different intrinsic components in an ECG signal. This is also true for the PLn case. The results. shown in [Table 1](#sensors-19-00798-t001){ref-type=\"table\"} and [Table 2](#sensors-19-00798-t002){ref-type=\"table\"} have justified the idea where the IMFs for clean and noisy ECG signals can be discriminated in most cases. Another evidence is in [Table 3](#sensors-19-00798-t003){ref-type=\"table\"} and [Table 4](#sensors-19-00798-t004){ref-type=\"table\"}. The results indicate that the average $NER_{dB}$ are positive. That is, intrinsic noise and additive noise can be dealt and cancelled in the EMD, the EEMD and the proposed GSNC schemes. Among the three schemes, the proposed GSNC scheme has the best performance with the average $NER_{dB}$ 3.34 dB for the PLn and 6.14 dB for the EMG noise, as shown in [Table 3](#sensors-19-00798-t003){ref-type=\"table\"} and [Table 4](#sensors-19-00798-t004){ref-type=\"table\"}, respectively. It implies that the proposed GSNC scheme is able to estimate the noise appropriately and the two-stage discrimination can relieve the over-cancellation problem which generally happens in the one-stage discrimination.\n\nIn order to show the proposed GSNC scheme does not affect the morphology of arrhythmic beat, the arrhythmic ECG dataset mitdb/210 is given as an example where two premature ventricular contraction (PVC) beats are within the duration. The denoised results for the PLn and EMG noise ($SNR = 5{\\ {dB}}$) by the proposed GSNC scheme are given in [Figure 8](#sensors-19-00798-f008){ref-type=\"fig\"}. As shown in [Figure 8](#sensors-19-00798-f008){ref-type=\"fig\"}b,c, the proposed GSNC scheme is able to retain subtle signs in the denoised ECG signal, when compared with [Figure 8](#sensors-19-00798-f008){ref-type=\"fig\"}a.\n\n7. Conclusions {#sec7-sensors-19-00798}\n==============\n\nThis paper has presented a grey spectral noise cancellation (GSNC) scheme for ECG signals. In the proposed GSNC scheme, two-stage discrimination for noisy IMFs was employed which included the EMD, the EEMD and grey spectral noise estimation (GSNE). In general condition, the lower order IMFs would be easily coupled with the noise. The GSNE was applied to estimate the noise energy through the standard deviation of noise magnitude spectrum $\\sigma_{|{\\hat{N}{(f)}}|}$. By $\\sigma_{|{\\hat{N}{(f)}}|}$, noisy IMFs were determined and discarded in the process of reconstruction. The proposed GSNC scheme has been verified by forty-three datasets from the MIT-BIH database where different SNR levels for the PLn and EMG noise were considered. The results indicated that the proposed GSNC scheme was generally superior to the compared EMD and EEMD schemes in terms of average $NER_{dB}$ in the given datasets. The proposed GNSC scheme provides a new approach, based on the noise magnitude spectrum, to estimate the noise energy in the IMFs. The proposed GNSC scheme could be implemented in an embedded system, like the ECG patch, to deal with noises in the ECG signals.\n\nConceptualization: S.-H.L. and C.-H.H.; methodology: S.-H.L., C.-H.H., and W.C.; software: T.-H.T.; writing---original draft preparation: S.-H.L. and C.-H.H.; writing---review and editing: S.-H.L.; C.-H.H., and W.C.\n\nThis research was funded by the grants of the Ministry of Science and Technology of the Republic of China MOST 106-2221-E-324-011 and MOST 106-2221-E-324-001.\n\nThe authors declare no conflict of interest.\n\n![The block diagram of the the grey spectral noise estimation (GSNE).](sensors-19-00798-g001){#sensors-19-00798-f001}\n\n![One sample overlapped subsets in the GNSE $\\left( {K = 4} \\right)$.](sensors-19-00798-g002){#sensors-19-00798-f002}\n\n![Differences of $\\sigma_{|{\\hat{N}{(f)}}|}$ (**a**) ${ECG}_{raw}$ vs ${ECG}_{PLn}$ (**b**) ${ECG}_{raw}$ vs ${ECG}_{EMG}$.](sensors-19-00798-g003){#sensors-19-00798-f003}\n\n![The flowchart of the proposed GSNC scheme.](sensors-19-00798-g004){#sensors-19-00798-f004}\n\n![The original ECG signal and ECG signals after the proposed GSNC scheme with different SNR of white noise in the ensemble empirical mode decomposition (EEMD) (EMG noise, ${SNR} = 5{\\ {dB}}$).](sensors-19-00798-g005){#sensors-19-00798-f005}\n\n![Comparison of overall average $NER_{dB}$ with various *SNR* for the proposed GSNC, EMD and EEMD schemes (PLn).](sensors-19-00798-g006){#sensors-19-00798-f006}\n\n![Comparison of overall average $NER_{dB}$ with various *SNR* for the proposed GSNC, EMD and EEMD schemes (EMG noise).](sensors-19-00798-g007){#sensors-19-00798-f007}\n\n![The arrhythmic ECG signals, (**a**) the original ECG signal with two PVC beats, (**b**) the ECG with PLn, (**c**) the denoised ECG with PLn, (**d**) the ECG with EMG noise (**e**) the denoised ECG with EMG noise.](sensors-19-00798-g008){#sensors-19-00798-f008}\n\nsensors-19-00798-t001_Table 1\n\n###### \n\nThe $\\sigma_{|{\\hat{N}{(f)}}|}$ for clean signal $s\\left( k \\right)$ and noisy signal $x\\left( k \\right)$ with PLn ($SNR = 5{\\ {dB}}$).\n\n                        mitdb/100             mitdb/105       mitdb/108       mitdb/203       mitdb/223       mitdb/228        \n  --------------------- --------------------- --------------- --------------- --------------- --------------- ---------------- ---------------\n  IMF 1                 $s\\left( k \\right)$   1.16 \u00d7 10^\u22124^   7.90 \u00d7 10^\u22125^   6.97 \u00d7 10^\u22125^   8.37 \u00d7 10^\u22125^   1.27 \u00d7 10^\u22124^    7.84 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   2.56 \u00d7 10^\u22123^         4.54 \u00d7 10^\u22123^   2.78 \u00d7 10^\u22123^   4.40 \u00d7 10^\u22123^   6.39 \u00d7 10^\u22123^   2.60 \u00d7 10^\u22123^    \n  IMF 2                 $s\\left( k \\right)$   5.18 \u00d7 10^\u22125^   3.91 \u00d7 10^\u22125^   2.93 \u00d7 10^\u22125^   4.23 \u00d7 10^\u22125^   5.065 \u00d7 10^\u22125^   3.24 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   6.58 \u00d7 10^\u22124^         1.22 \u00d7 10^\u22123^   6.52 \u00d7 10^\u22124^   1.13 \u00d7 10^\u22123^   1.61 \u00d7 10^\u22123^   6.81 \u00d7 10^\u22124^    \n  IMF 3                 $s\\left( k \\right)$   2.66 \u00d7 10^\u22125^   3.05 \u00d7 10^\u22125^   1.70 \u00d7 10^\u22125^   2.37 \u00d7 10^\u22125^   2.98 \u00d7 10^\u22125^    1.57 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   2.08 \u00d7 10^\u22124^         3.51 \u00d7 10^\u22124^   2.06 \u00d7 10^\u22124^   3.31 \u00d7 10^\u22124^   4.85 \u00d7 10^\u22124^   2.02 \u00d7 10^\u22124^    \n  IMF 4                 $s\\left( k \\right)$   1.43 \u00d7 10^\u22125^   1.36 \u00d7 10^\u22125^   9.01 \u00d7 10^\u22126^   1.25 \u00d7 10^\u22125^   2.51 \u00d7 10^\u22125^    7.02 \u00d7 10^\u22126^\n  $x\\left( k \\right)$   8.93 \u00d7 10^\u22125^         1.01 \u00d7 10^\u22124^   6.09 \u00d7 10^\u22125^   1.05 \u00d7 10^\u22124^   1.65 \u00d7 10^\u22124^   5.63 \u00d7 10^\u22125^    \n  IMF 5                 $s\\left( k \\right)$   6.58 \u00d7 10^\u22126^   6.85 \u00d7 10^\u22126^   5.22 \u00d7 10^\u22126^   7.67 \u00d7 10^\u22126^   1.12 \u00d7 10^\u22125^    3.80 \u00d7 10^\u22126^\n  $x\\left( k \\right)$   3.28 \u00d7 10^\u22125^         4.26 \u00d7 10^\u22125^   2.40 \u00d7 10^\u22125^   3.59 \u00d7 10^\u22125^   6.25 \u00d7 10^\u22125^   2.61 \u00d7 10^\u22125^    \n\nsensors-19-00798-t002_Table 2\n\n###### \n\nThe $\\sigma_{|{\\hat{N}{(f)}}|}$ for clean signal $s\\left( k \\right)$ and noisy signal $x\\left( k \\right)$ with EMG noise ($SNR = 5{\\ {dB}}$).\n\n                        mitdb/100             mitdb/105       mitdb/108       mitdb/203       mitdb/223       mitdb/228       \n  --------------------- --------------------- --------------- --------------- --------------- --------------- --------------- ---------------\n  IMF 1                 $s\\left( k \\right)$   1.16 \u00d7 10^\u22124^   7.90 \u00d7 10^\u22125^   6.97 \u00d7 10^\u22125^   8.37 \u00d7 10^\u22125^   1.27 \u00d7 10^\u22124^   7.84 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   1.36 \u00d7 10^\u22123^         2.51 \u00d7 10^\u22123^   1.43 \u00d7 10^\u22123^   2.37 \u00d7 10^\u22123^   3.42 \u00d7 10^\u22123^   1.37 \u00d7 10^\u22123^   \n  IMF 2                 $s\\left( k \\right)$   5.18 \u00d7 10^\u22125^   3.91 \u00d7 10^\u22125^   2.93 \u00d7 10^\u22125^   4.23 \u00d7 10^\u22125^   5.06 \u00d7 10^\u22125^   3.24 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   4.00 \u00d7 10^\u22124^         7.75 \u00d7 10^\u22124^   4.75 \u00d7 10^\u22124^   7.04 \u00d7 10^\u22124^   1.03 \u00d7 10^\u22123^   4.82 \u00d7 10^\u22124^   \n  IMF 3                 $s\\left( k \\right)$   2.66 \u00d7 10^\u22125^   3.05 \u00d7 10^\u22125^   1.70 \u00d7 10^\u22125^   2.37 \u00d7 10^\u22125^   2.98 \u00d7 10^\u22125^   1.57 \u00d7 10^\u22125^\n  $x\\left( k \\right)$   1.40 \u00d7 10^\u22124^         2.06 \u00d7 10^\u22124^   1.30 \u00d7 10^\u22124^   2.09 \u00d7 10^\u22124^   3.17 \u00d7 10^\u22124^   1.20 \u00d7 10^\u22124^   \n  IMF 4                 $s\\left( k \\right)$   1.43 \u00d7 10^\u22125^   1.36 \u00d7 10^\u22125^   9.01 \u00d7 10^\u22126^   1.25 \u00d7 10^\u22125^   2.5 \u00d7 10^\u22125^    7.02 \u00d7 10^\u22126^\n  $x\\left( k \\right)$   5.53 \u00d7 10^\u22125^         6.60 \u00d7 10^\u22125^   3.93 \u00d7 10^\u22125^   5.89 \u00d7 10^\u22125^   9.61 \u00d7 10^\u22125^   3.92 \u00d7 10^\u22125^   \n  IMF 5                 $s\\left( k \\right)$   6.58 \u00d7 10^\u22126^   6.85 \u00d7 10^\u22126^   5.22 \u00d7 10^\u22126^   7.67 \u00d7 10^\u22126^   1.12 \u00d7 10^\u22125^   3.80 \u00d7 10^\u22126^\n  $x\\left( k \\right)$   2.14 \u00d7 10^\u22125^         3.13 \u00d7 10^\u22125^   1.28 \u00d7 10^\u22125^   2.14 \u00d7 10^\u22125^   3.65 \u00d7 10^\u22125^   1.52 \u00d7 10^\u22125^   \n\nsensors-19-00798-t003_Table 3\n\n###### \n\nThe $NER_{dB}$ for the proposed GSNC and the compared schemes with the PLn $\\left( {SNR = 5\\ {dB}} \\right)$.\n\n  Dataset     EMD      EEMD    GSNC    Dataset     EMD      EEMD    GSNC\n  ----------- -------- ------- ------- ----------- -------- ------- -------\n  mitdb/100   \u22124.59    6.36    6.51    mitdb/205   1.59     \u22121.14   5.08\n  mitdb/101   1.01     0.32    2.83    mitdb/207   4.12     6.92    9.48\n  mitdb/103   \u22121.72    1.09    1.17    mitdb/208   \u22124.45    2.79    \u22122.07\n  mitdb/105   6.04     3.53    5.02    mitdb/209   \u22127.96    \u22120.60   \u22125.02\n  mitdb/106   \u22125.32    0.12    \u22121.48   mitdb/210   6.58     4.24    7.68\n  mitdb/108   3.29     6.52    3.50    mitdb/212   \u221211.33   \u22123.40   \u22125.33\n  mitdb/109   7.96     7.90    9.97    mitdb/213   3.97     \u22123.15   7.10\n  mitdb/111   4.19     1.53    5.48    mitdb/214   \u22122.72    5.24    \u22120.91\n  mitdb/112   6.75     5.24    7.52    mitdb/215   \u22122.11    0.08    1.22\n  mitdb/113   \u22122.12    \u22121.81   \u22120.72   mitdb/219   \u22128.31    5.23    2.91\n  mitdb/114   \u22120.75    2.78    2.99    mitdb/220   \u22121.13    \u22120.52   0.40\n  mitdb/115   \u22123.01    \u22121.50   0.10    mitdb/221   2.23     2.87    3.46\n  mitdb/116   \u22123.52    3.77    3.32    mitdb/222   0.97     \u22120.39   6.22\n  mitdb/117   3.11     0.31    5.65    mitdb/223   5.48     5.62    6.24\n  mitdb/118   \u22121.84    \u22120.53   \u22120.19   mitdb/228   7.31     1.67    6.27\n  mitdb/119   2.03     0.37    3.32    mitdb/230   \u22121.58    0.32    \u22120.26\n  mitdb/121   2.76     2.83    3.88    mitdb/231   \u22126.55    \u22125.67   \u22124.45\n  mitdb/122   5.46     4.18    9.73    mitdb/232   \u22122.65    6.47    5.64\n  mitdb/123   \u22120.52    \u22120.42   \u22122.01   mitdb/233   0.73     6.18    6.68\n  mitdb/124   5.82     5.61    6.74    mitdb/234   \u22125.21    0.66    \u22120.22\n  mitdb/201   \u221212.83   5.44    6.72                                 \n  mitdb/202   7.83     4.53    8.42    Mean        0.10     2.20    3.34\n  mitdb/203   5.40     3.07    5.15    Std.        5.21     3.21    4.03\n\nsensors-19-00798-t004_Table 4\n\n###### \n\nThe $NER_{dB}$ for the proposed GSNC and the compared schemes with the EMG noise $\\left( {SNR = 5\\ {dB}} \\right)$.\n\n  Dataset     EMD     EEMD    GSNC    Dataset     EMD     EEMD    GSNC\n  ----------- ------- ------- ------- ----------- ------- ------- -------\n  mitdb/100   \u22123.02   8.66    7.94    mitdb/205   1.71    1.91    10.87\n  mitdb/101   0.14    2.18    2.65    mitdb/207   11.32   12.47   13.21\n  mitdb/103   \u22122.79   1.70    0.77    mitdb/208   \u22122.11   6.44    2.81\n  mitdb/105   6.44    7.68    8.21    mitdb/209   0.31    \u22121.44   1.65\n  mitdb/106   5.03    2.21    5.42    mitdb/210   7.79    3.66    9.50\n  mitdb/108   6.61    7.39    3.87    mitdb/212   \u22121.29   \u22124.15   \u22125.14\n  mitdb/109   10.84   11.17   12.61   mitdb/213   2.55    7.52    5.96\n  mitdb/111   8.23    3.79    9.44    mitdb/214   7.64    4.59    10.04\n  mitdb/112   1.50    3.16    1.97    mitdb/215   2.13    0.14    3.66\n  mitdb/113   \u22120.92   1.76    1.43    mitdb/219   7.62    4.16    9.44\n  mitdb/114   4.96    9.50    12.16   mitdb/220   0.42    \u22121.53   1.11\n  mitdb/115   1.43    \u22122.90   \u22120.77   mitdb/221   4.54    5.95    6.93\n  mitdb/116   3.27    6.24    9.45    mitdb/222   5.32    3.54    7.06\n  mitdb/117   \u22120.53   \u22120.08   6.12    mitdb/223   6.30    8.59    7.63\n  mitdb/118   \u22121.63   \u22121.98   \u22120.57   mitdb/228   5.25    3.47    6.55\n  mitdb/119   7.59    1.12    8.68    mitdb/230   0.82    0.27    3.21\n  mitdb/121   12.19   9.67    9.94    mitdb/231   \u22122.91   \u22126.66   \u22120.49\n  mitdb/122   2.73    2.50    5.21    mitdb/232   \u22128.03   7.15    7.64\n  mitdb/123   2.86    \u22121.12   3.38    mitdb/233   8.98    7.05    11.14\n  mitdb/124   8.53    6.74    11.42   mitdb/234   \u22120.39   4.08    2.93\n  mitdb/201   6.67    7.81    9.34                                \n  mitdb/202   4.10    7.98    9.79    Mean        3.52    3.85    6.14\n  mitdb/203   9.21    3.26    9.72    Std.        4.52    4.27    4.29\n"}
+{"text": "Introduction {#sec1}\n============\n\nChanges in training load -- particularly in the frequency, duration, and intensity of training sessions -- are associated with the principle of training stimulus variability that seeks to optimize sports performance ([@ref16]). Training load monitoring can be categorized into two forms: external load and internal load ([@ref23]). External load is understood as the physical repercussions of training performed by an athlete, encompassing indicators such as distance, duration, and race intensity ([@ref21]). Internal load is associated with the biological response of the athlete to the external load imposed by training ([@ref5]).\n\nIn the training's prescription, it is essential that the external and internal loads be appropriate and that there is a balance between them, allowing for improvements in the performance of the athlete and for the reduction of overload or underload ([@ref2]). The correct planning of the training load through microcycles allows an approximation of the training regarding the requirements of races ([@ref26]), causing fundamental specific adaptations in the athlete ([@ref25]).\n\nThe monitoring of training loads requires an accurate and reliable evaluation of the determinants of the training process ([@ref28]). However, the use of different methods and/or techniques is dependent on the context, namely considering the applicability and the resources. As an example, the internal load can be more objectively measured by using heart rate sensors or collecting a blood sample to determine the blood lactate ([@ref32]). However, such methods are somehow invasive or not practical in some contexts. On the other hand, subjective scales of intensity (e.g., rated of perceived exertion) have been presenting very good levels of validity and reliability, are less invasive and more practical in realistic training scenarios ([@ref15]). Similarly, the external load quantification is also dependent on the context and will provide different information than internal load, mainly considering specific sports that require a great perception of the pace and intensity of running. For these cases, the global positioning systems are often used considering that may provide complementary information to coaches and athletes ([@ref16]).\n\nDespite the unquestionable importance of quantification of the load to regulate the training process, the monitoring cycle of athletes does not finish with a simple quantification of the load. Other parameters related with the impact of the training stimulus on athletes are also a part of the monitoring process, namely considering the well-being parameters that include, among others, the perception of delayed onset muscle soreness (DOMS), fatigue, stress, or sleep quality ([@ref19]). In this sense, the literature refers to well-being questionnaires as a good indicator of the evaluation of these variables, and the Hooper questionnaire ([@ref19]) as being pointed out as a good tool to estimate the impact and to manage the dose of training in athletes.\n\nDespite a great number of publications considering the training load quantification and well-being determination, the great majority of the studies are related to team sports ([@ref28]; [@ref24]) while just a few, to the best of our knowledge, are dedicated to individual sports ([@ref31]; [@ref18]). Among individual sports, the trail running practice has been increasing in the last few years and is a sport with an apparent necessity of load management considering the great distances covered by the athletes. This sport can be characterized as a mountain run ([@ref30]) with race distances that may vary according to the type of competition, ranging between 10 and 894 km ([@ref29]). Trail running races are competitions that can last for several hours or even days because of accumulated unevenness and terrain specificity, with times varying from athlete to athlete ([@ref8]).\n\nDue to the specificity of trail running, researches have been carried out to characterize the load and the physiological requirements derived from races ([@ref33]). Usually, maximal oxygen uptake between 60 and 85 ml\u00b7kg^\u22121^\u00b7min^\u22121^ can be found in this type of athletes ([@ref13]). Therefore, it seems reasonable to assume that the training process should be adjusted to the requirements of the race and must be properly varied during the weeks aiming to fit the load with the performance expectations of the athletes. Despite this necessity, there is a lack of evidence about how athletes manage and apply the load. This is particularly important because a great number of these athletes are non-professional (recreational) and for that reason, it is important to characterize how they manage the load during the training and identify the variations of well-being parameters during the week.\n\nTo the best of our knowledge, no information has been reported about the intra- and inter-week variations of training load and well-being of trail runners. Based on that, the first purpose of our study was to characterize the training load (internal and external) and well-being parameters of trail runner athletes during a mid-season and competitive mesocycle of 4 weeks. As the second purpose of this study, we tested possible relationships (correlations) between aerobic capacity of athletes (estimated by a field-based test), performance in races (pace), and the training load variables, aiming to determine if the training process (namely intensity and volume) can be associated with the aerobic capacity and performance of these recreational athletes.\n\nMaterials and Methods {#sec2}\n=====================\n\nParticipants {#sec3}\n------------\n\nForty-seven Portuguese male trail running athletes (average age = 34.85 \u00b1 8.88 years; height: 177.34 \u00b1 5.81 cm; weight: 65.89 \u00b1 3.17 kg; experience: 4.72 \u00b1 2.11 years) participated in this study. On average, the athletes covered 35,159 m and trained for over 206 min per week. The absence of injuries during the study and the accomplishment of at least one and at most three races of the national championship of trail running were the criteria of inclusion. All participants provided informed consent in accordance with the recommendations of the Declaration of Helsinki for human study. The study was also approved by the local ethical committee (Polytechnic Institute of Viana do Castelo, School of Sport and Leisure) with the code number IPVC-ESDL171003.\n\nDesign {#sec4}\n------\n\nThe external and internal loads and the well-being of trail running athletes were monitored throughout the month of November 2017, during which 26 athletes participated in races of the national trail running championships. Despite participating in national trail running championships, the category of these athletes is recreational based on the fact that they are not professional and perform three sessions/week or less. However, for inclusion in this study, those reported less than three sessions per week were excluded from the analysis. From the athletes that participated in races, 65.4% competed in one race, 23.1% in two races, and 11.5% in three races. The races varied from a minimum of 10 km to a maximum of 300 km. The pace (min/km) made by the athletes during each race was collected to further correlations between aerobic capacity (as indicator of fitness level) and performance in race. In the week before the training load and well-being monitoring started, the 12-min Cooper test was implemented. During the training mesocycle, athletes were required to fill out the Hooper questionnaire before training sessions and races and to fill out the Borg scale after the end of training sessions and races. Both questionnaires were completed using an online form. External load was monitored using GPS (Global Positioning System) devices.\n\nData Collection: Global Positioning System {#sec5}\n------------------------------------------\n\nDuring the training sessions, the athletes used watches with GPS technology, enabling the collection of information regarding horizontal movement. The Polar V800 (37 mm \u00d7 56 mm \u00d7 12.7 mm and weight: 79 g) ([@ref27]) was used based on its validity for the collection of positional information.\n\nHooper Index {#sec6}\n------------\n\nThe Hooper index (HI) questionnaire for assessing athletes' well-being was administered individually, 30 min before training sessions and races, for the variables of sleep quality, stress, fatigue, and muscle soreness. Answers were given using scales of 1--7. For the variables of fatigue, stress, and muscle soreness, 1 = very, very low, and 7 = very, very high. For sleep quality, 1 = very, very bad and 7 = very, very good ([@ref19]).\n\nRated of Perceived Exertion {#sec7}\n---------------------------\n\nThe rated of perceived exertion (RPE) quantified by using the CR-10 Borg's scale ([@ref4]) was used as a measure of exercise intensity. On the CR-10 Borg's scale, the 1 = very, very easy and 10 = extremely hard. The CR-10 Borg's scale was firstly introduced to the participants aiming to familiarize them with the scale. After that, they have used the scale for 2 weeks without including the data in the study just aiming to increase the familiarization and the accuracy of the athlete's answers.\n\nAfter such period, and during the data collection, the athletes scored the RPE 30 min after the end of training session in a dedicated online form built for the effect. Moreover, they reported the time of the session in minutes. Using both information (i.e., RPE score and time of the session), it was possible to determine the session-RPE (sRPE) that represents the overall internal load of the session by multiplying the RPE score for the time of the session in minutes ([@ref10]). The sRPE has been used as a valid and reliable measure of internal load ([@ref15]). The data were collected in all training sessions that occurred in the period of data collection, thus the sRPE was calculated on a daily basis.\n\nThe 12-min Cooper Test {#sec8}\n----------------------\n\nA Cooper test with a duration of 12 min was performed for the estimation of cardiorespiratory capacity and the maximal oxygen uptake (VO~2max~). The test was performed while athletes were running or walking without interruption, and the total distance covered in the 12 min was recorded. All the participants reported previous experience and familiarization in this specific test, considering their previous participation in performance assessments.\n\nA 5-min warm-up run was performed with a 10-min interval between the warm-up run and the test. All athletes performed the test in the same place between 10:00 a.m. and noon, with no precipitation at a temperature of 14\u00b0C and with a relative humidity of 45%. The test took place 72 h after the previous race or training session. The test was performed in an official athletic track and the distance covered by each athlete was collected immediately after the 12-min Cooper test. The distance covered (m) was one of the measures associated with the performance in the Cooper test. Moreover, using the distance on the Cooper test and the equation proposed in Bandyopadhyay ([@ref1]), the maximal oxygen uptake of the athletes was estimated.\n\nData Analyses and Statistics {#sec9}\n----------------------------\n\nDescriptive statistics were presented in the form of mean, standard deviation, and 95% confidence intervals (presented in the [Figures 1](#fig1){ref-type=\"fig\"}--[4](#fig4){ref-type=\"fig\"}). The weekly RPE and well-being variables were treated as the average of the RPE and well-being variables in each week for each athlete and then integrated into the mean of the participants ([Figures 3](#fig3){ref-type=\"fig\"}, [4](#fig4){ref-type=\"fig\"}). The weekly accumulated load and well-being \\[sum of the arbitrary units (A.U.) of all sessions of each week\\] were also calculated for each athlete and then integrated into the mean of the participants ([Figures 1](#fig1){ref-type=\"fig\"}, [2](#fig2){ref-type=\"fig\"}).\n\n![Weekly variations of accumulated (sum of the A.U. of all sessions of each week) RPE and session-RPE (sRPE).](fphys-10-01189-g001){#fig1}\n\n![Weekly variations of accumulated (sum of the scores of all sessions of each week) stress, fatigue, DOMS, and sleep quality.](fphys-10-01189-g002){#fig2}\n\n![Intra-week variations of distance, duration, RPE, and session-RPE (sRPE) (averages of 1st, 2nd, and 3rd training sessions of the week).](fphys-10-01189-g003){#fig3}\n\n![Intra-week variations of sleep quality, stress, fatigue, and DOMS (averages of 1st, 2nd, and 3rd training sessions of the week).](fphys-10-01189-g004){#fig4}\n\nInter-week (comparisons of the weekly average of each measure between the 4 weeks) and intra-week (comparisons of the 4 weeks' average of each measure in each training session) comparisons were tested with one-way repeated-measures ANOVA after confirmation of the assumptions of normality and homogeneity of the samples. The partial eta squared ($\\eta^{2}$) tested the effect size of the repeated-measures ANOVA. The magnitude inferences of $\\eta^{2}$ were defined as ([@ref9]): no effect ($\\eta^{2}$ \\< 0.04), small effect (0.05 \\< $\\eta^{2}$ \\< 0.25), moderate effect (0.26 \\< $\\eta^{2}$ \\< 0.64), or strong effect ($\\eta^{2}$ \\> 0.65). The Tukey HSD *post hoc* test and the Cohen's *d* tested the significances and the effect size of differences between factors. The following magnitude inferences were made for the Cohen's *d*: 0.0--0.2, trivial effect; 0.2--0.6, small effect; 0.6--1.2, moderate effect; and 1.2--2.0, large effect.\n\nFor the study of associations between well-being variables, training load; aerobic capacity; and the race performance, Pearson's *r* test was performed. In the particular case of correlations tested with the performance in the races, it was used a mean of the pace of each participant in the races in he has participated. According to the Hopkins classification followed for the study of the magnitude of correlations, correlation values were classified as follows ([@ref20]): \\[0.0,0.1\\], trivial; (0.1,0.3\\], small; (0.3,0.5\\], moderate; (0.5,0.7\\], large; (0.7,0.9\\], very large; and (0.9,1.0\\], nearly perfect. Statistical procedures were performed in the statistical software SPSS (IBM, USA, version 23.0) for a significance level of 5%.\n\nResults {#sec10}\n=======\n\nIntra- and Inter-Week Variations of Training Load and Well-Being {#sec11}\n----------------------------------------------------------------\n\nThe descriptive statistics of accumulated (sum of the sessions of each week) training load parameters and well-being variables can be found in [Figures 1](#fig1){ref-type=\"fig\"}, [2](#fig2){ref-type=\"fig\"}, respectively.\n\nInter-week (changes between weeks) differences for the weekly average of RPE (*p* = 0.011; $\\eta^{2}$ = 0.078, small effect) and sRPE (*p* = 0.025; $\\eta^{2}$ = 0.065, small effect) were found. Weekly RPE was greater in week 1 than in week 3 (*p* = 0.001; *d* = 0.563, small effect). Moreover, weekly sRPE was greater in week 1 than in week 2 (*p* = 0.001; *d* = 0.441, small effect). No significant changes in weekly duration (*p* = 0.12; $\\eta^{2}$ = 0.40, no effect) and distance (*p* = 0.062; $\\eta^{2}$ = 0.062, small effect) were found.\n\nNo significant changes were found in weekly sleep (*p* = 0.389; $\\eta^{2}$ = 0.030, no effect), weekly stress (*p* = 0.537; $\\eta^{2}$ = 0.022, no effect), weekly fatigue (*p* = 0.319; $\\eta^{2}$ = 0.035, no effect), and weekly muscle soreness (*p* = 0.562; $\\eta^{2}$ = 0.020, no effect).\n\nIntra-week (changes within week) analysis of training load and well-being parameters can be found in [Figures 3](#fig3){ref-type=\"fig\"}, [4](#fig4){ref-type=\"fig\"}, respectively. Repeated-measures ANOVA did not reveal significant changes between the training sessions of the week in the distance (*p* = 0.618; $\\eta^{2}$ = 0.002, no effect), duration (*p* = 0.303; $\\eta^{2}$ = 0.005, no effect), RPE (*p* = 0.751; $\\eta^{2}$ = 0.001, no effect), and internal load (*p* = 0.915; $\\eta^{2}$ = 0.001, no effect).\n\nIntra-week comparisons of well-being variables also revealed no significant changes between training session in sleep quality (*p* = 0.776; $\\eta^{2}$ = 0.001, no effect); stress (*p* = 0.233; $\\eta^{2}$ = 0.006, no effect); fatigue (*p* = 0.557; $\\eta^{2}$ = 0.002, no effect); muscle soreness (*p* = 0.852; $\\eta^{2}$ = 0.001, no effect); and Hooper index (*p* = 0.733; $\\eta^{2}$ = 0.001, no effect).\n\nAssociation Between Well-Being and Performance in Training {#sec12}\n----------------------------------------------------------\n\nThe mean values of the RPE, s-RPE, and well-being variables during 4 weeks of training of trail running athletes can be seen in [Table 1](#tab1){ref-type=\"table\"}.\n\n###### \n\nDescriptive statistics (*M* \u00b1 SD) of well-being variables and training load during the mesocycle (average of the training session).\n\n                        Mean     SD\n  --------------------- -------- --------\n  Sleep (A.U.)          2.55     1.15\n  Stress (A.U.)         2.68     1.16\n  Fatigue (A.U.)        2.95     1.16\n  DOMS (A.U.)           2.55     1.27\n  Hooper index (A.U.)   10.72    3.99\n  Distance (km)         11.28    5.31\n  Duration (min)        66.43    39.07\n  RPE (A.U.)            2.95     1.54\n  sRPE (A.U.)           213.79   223.95\n\n*DOMS, delayed onset muscle soreness; RPE, rate of perceived exertion in the CR-10 Borg's scale; sRPE, session-RPE representing the multiplication of RPE by the time in minutes; A.U., arbitrary units*.\n\nThe correlation between well-being variables and training load variables was identified in order to identify possible associations between these variables throughout the training process. The results of Pearson' correlation coefficients *r* can be found in [Table 2](#tab2){ref-type=\"table\"}.\n\n###### \n\nCorrelation values (*r*) between well-being variables and training load along the mesocycle.\n\n                        Distance (km)                            Duration (min)                           RPE (A.U.)                               sRPE (U.A.)\n  --------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- ----------------------------------------\n  Sleep (A.U.)          0.207[^b^](#tfn2){ref-type=\"table-fn\"}   0.153[^b^](#tfn2){ref-type=\"table-fn\"}   0.287[^b^](#tfn2){ref-type=\"table-fn\"}   0.249[^b^](#tfn2){ref-type=\"table-fn\"}\n  Stress (A.U.)         \u22120.007                                   \u22120.038                                   0.217[^b^](#tfn2){ref-type=\"table-fn\"}   0.079\n  Fatigue (A.U.)        0.068                                    0.012                                    0.191[^b^](#tfn2){ref-type=\"table-fn\"}   0.109[^a^](#tfn1){ref-type=\"table-fn\"}\n  DOMS (A.U.)           0.134[^b^](#tfn2){ref-type=\"table-fn\"}   0.098[^a^](#tfn1){ref-type=\"table-fn\"}   0.240[^b^](#tfn2){ref-type=\"table-fn\"}   0.193[^b^](#tfn2){ref-type=\"table-fn\"}\n  Hooper index (A.U.)   0.120[^b^](#tfn2){ref-type=\"table-fn\"}   0.068                                    0.279[^b^](#tfn2){ref-type=\"table-fn\"}   0.188[^b^](#tfn2){ref-type=\"table-fn\"}\n\n*RPE, rated of perceived exertion in the CR-10 Borg's scale; DOMS, delayed onset muscle soreness; sRPE, session-RPE representing the multiplication of RPE by the time in minutes*.\n\n*Significant correlation at p \\< 0.05*.\n\n*Significant correlation at p \\< 0.01*.\n\nPhysical Variables and Performance in Evidence {#sec13}\n----------------------------------------------\n\n[Table 3](#tab3){ref-type=\"table\"} presents the mean values of the participants who underwent the 12-min Cooper test and the races performed during the mesocycle.\n\n###### \n\nDescriptive statistics (*M* \u00b1 SD) of the 12-min Cooper test and race values during the mesocycle.\n\n                             Mean      SD\n  -------------------------- --------- --------\n  Cooper test 12-min (m)     3168.97   287.57\n  VO~2max~ (ml/kg/min^\u22121^)   59.56     6.43\n  RPE (A.U.)                 6.15      2.24\n  Pace (min/km)              7.38      2.04\n\n*RPE, rated of perceived exertion in the CR-10 Borg's scale*.\n\nThe Pearson's correlation analysis was performed between the performance variables in the Cooper test and the sports performance measured in the race ([Table 4](#tab4){ref-type=\"table\"}). Positive and significant mean values were found between Cooper 12-min (m) and RPE (*r* = 0.380, *p* = 0.017), as well as negative values with moderate magnitude between Cooper 12 -min and the pace (*r* = \u22120.395, *p* = 0.016). Similar results were observed between estimated VO~2max~ and RPE (*r* = 0.379, *p* = 0.017) and the pace (*r* = \u22120.396, *p* = 0.015).\n\n###### \n\nCorrelation values (*r*) between the performance variables in the 12-min Cooper test and load during the races.\n\n                             RPE (A.U.)                               Pace (min/km)\n  -------------------------- ---------------------------------------- -----------------------------------------\n  Cooper test 12-min (m)     0.380[^a^](#tfn3){ref-type=\"table-fn\"}   \u22120.395[^a^](#tfn3){ref-type=\"table-fn\"}\n  VO~2max~ (ml/kg/min^\u22121^)   0.379[^a^](#tfn3){ref-type=\"table-fn\"}   \u22120.396[^a^](#tfn3){ref-type=\"table-fn\"}\n\n*RPE, rated of perceived exertion in the CR-10 Borg's scale*.\n\n*Significant correlation at p \\< 0.05*.\n\nDiscussion {#sec14}\n==========\n\nSignificant changes of RPE and s-RPE were found between the first and the third weeks and the first and the second weeks, respectively. Despite that, no more significant changes were found between weeks, possibly suggesting that there is a lack of progression in the stimulus and variability inter-week that is crucial to optimize the performance and to reduce the exposure to injuries ([@ref11]). In fact a stabilization of the load may contribute to a performance plateau and, for that reason, it is interesting to identify that these athletes are not promoting (in a significative scale) the principles of variability and progression of the load based on the general absence of changes in the training load during the week and even the general comparisons between accumulated load over the weeks analyzed. One possible cause to observe such tendency can be the fact that during the period the athletes participated in races, and this may be constrained the variability within the mesocycle. As previously mentioned, the mesocycles occurred in mid-season during the competitive period; however, and depending on the goal of each athlete, some variability in terms of the training process may occur. This should be considered a limitation of the present study. Despite that, the descriptive statistics of the internal load revealed that the sRPE per training day was relatively similar to that of the previous reports in race athletes ([@ref6]). Another evidence was that no intra-week changes (differences between sessions within the week) were found in training load or in well-being variables, suggesting that the training within the week is almost the same between training sessions, again revealing a lack of variability in the training stimulus.\n\nThe analysis of well-being variables and training load ([Table 1](#tab1){ref-type=\"table\"}) allowed us to observe possible associations between these two factors during the mesocycle. By analyzing [Table 2](#tab2){ref-type=\"table\"} for correlation values, it can be seen that sleep quality has a significant small magnitude with all training load variables, indicating that sleep may have consequences related to athlete performance ([@ref17]). The stress variable only correlates in a significant way with RPE, although with a small magnitude, which may lead to an incorrect perception of the internal load resulting from the training sessions or races, and this may provide mismatched feedback for training monitoring ([@ref16]). Fatigue correlates significantly at a small magnitude with RPE and internal load variables, showing that high levels of RPE indicate the presence of fatigue ([@ref12]). DOMS had a small magnitude with distance, RPE, and internal load. This correlation suggests that an athlete's perception of muscle soreness is related to the impact of the race ([@ref14]). The Hooper index scores correlate significantly, albeit with a small magnitude, with the variables of distance, RPE, and internal load. This result allows us to affirm that athletes who train or race with high Hooper index values are likely to have low levels of well-being, resulting in a reduction in performance ([@ref12]).\n\nAerobic capacity of the athletes was tested to further correlations with performance in race. Our results in the 12-min Cooper test presented mean values of 3168.97 m. The values were similar with those reported by [@ref22]. Also, mean values of 59.56 \u00b1 6.43 ml\u00b7kg^--1^\u00b7min^--1^ (VO~2max~) were estimated in our recreational athletes. The correlation between the Cooper test and RPE showed positive and significant values, suggesting that a greater performance in Cooper may allow achieving higher intensities in training sessions. However, more interestingly, negative correlations were found between distance covered at Cooper test and estimated VO~2max~ with the pace in races, suggesting that greater aerobic capacity increases the intensity of running during official races. These results are in agreement with the literature regarding VO~2max~ as being the variable with the greatest effect on success in medium- and long-distance races ([@ref3]), being determinant to be succeeded.\n\nThe values obtained in the present recreational trail runners during races revealed a mean of perceived intensity of 6.15 \u00b1 2.24 on the Borg scale. The pace of the athletes during races presented a mean of 7.38 \u00b1 2.04 min/km. According to a study on running athletes by [@ref7], the pace was 4.05 min/km over a distance of 10 km, 4.21 min/km over the distance of a half marathon, and 4.48 min/km over the distance of a marathon. The differences between both values can be associated with the typology of long-running activities considering that trail running means to run in mountains with great variations in terms of terrain and accumulated unevenness involved.\n\nDespite its contributions, our study had some limitations. For future studies, we recommend that heart rate during training sessions and competitions can be considered. Caloric intake should also be considered in order to determine the influence it has on the performance level. Moreover, hydration levels resulting from the excess body temperature of the athletes during a race should be also controlled. This can be also associated with the internal load in race considering that dehydration in trail running athletes causes increases in heart rate, which results in increases in fatigue levels and in an erroneous perception of effort. Finally, in the case of professional athletes, it would be important to compute some robust parameters associated with training load analysis, namely, the acute: chronic workload ratio, training monotony, and training strain.\n\nThis competitive 1-month analysis of trail running athletes demonstrated that well-being variables had small correlations between RPE and sRPE. Moreover, a negative correlation was observed between aerobic capacity measured in the Cooper test and the estimated VO~2max~ with the pace in race, demonstrating that increases in maximal oxygen consumption translate into improvements in the pace and performance of athletes.\n\nConclusions {#sec15}\n===========\n\nIt was found that, generally, there are no significant changes of training load and well-being parameters within and between weeks. Small correlations were found between training load parameters and well-being variables. A third evidence was that moderate correlations between aerobic capacity and performance in race revealed that higher levels in maximum oxygen consumption (VO~2max~) reflect a decrease in pace (min/km) and, consequently, in performance improvements during races.\n\nData Availability {#sec16}\n=================\n\nThe datasets for this manuscript are not publicly available because upon request from the first author. Requests to access the datasets should be directed to S\u00e9rgio Matos: <sfcmatos@gmail.com>.\n\nEthics Statement {#sec17}\n================\n\nThe study was also approved by the local ethical committee (Polytechnic Institute of Viana do Castelo, School of Sport and Leisure) with the code number IPVC-ESDL171003.\n\nAuthor Contributions {#sec18}\n====================\n\nSM and FC conceived the study. FC and AB designed the study. SM collected data. FC analyzed and interpreted the data. SM, FC, and AB drafted the manuscript. SM, FC, AB, JP, PN, TR, and BK revised the manuscript and approved the final version.\n\nConflict of Interest Statement {#sec19}\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n[^1]: Edited by: Robert James Aughey, Victoria University, Australia\n\n[^2]: Reviewed by: Brian Hanley, Leeds Beckett University, United Kingdom; Laurent Mourot, Universit\u00e9 Bourgogne Franche-Comt\u00e9, France\n\n[^3]: This article was submitted to Exercise Physiology, a section of the journal Frontiers in Physiology\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nThe development of urban rail transit systems (URTS) has been proposed as a coping strategy to relieve traffic congestion around the world. A URTS not only has the properties of large capacity with less land occupation, but also has the advantages of energy conservation, environmental protection, high safety and reliability. It is a sustainable transportation mode.\n\nUrban traffic demand is known as the time-varying origin--destination (O-D) demand. For example, during the morning rush hour, passengers gather from places of residence to their work place; during the afternoon rush hour, passengers disperse from their work place to other places. The time-varying demand makes the URTS operation complex.\n\nThe core of the URTS operation is the train schedule. A high-quality schedule must fit the time-varying flow, i.e., it should have a high train frequency when the flows are large in some time intervals. Train schedule design of urban rail network is a rather complex problem with considering the time-varying demand. To simplify the schedule design problem, it is always divided into the schedule design for each single rail line \\[[@pone.0188874.ref001]--[@pone.0188874.ref002]\\]. Niu and Zhou \\[[@pone.0188874.ref001]\\] used the time-varying O-D demand for a single line and optimized its schedule. Shi *et al*. \\[[@pone.0188874.ref002]\\] optimized the frequency setting, timetabling and the rolling stock circulation for a single line based on the time-varying section flow. As the time-varying O-D demands or section flows for a single line include the passengers taking this rail line directly and transferring from other lines, and it is need to design a method for obtaining the reasonable time-varying O-D demands or section flows for each single line, it motivates us to study the dynamic assignment for urban rail transit network (DAURTN), which can approximately estimate the space-time flow distribution in the urban rail network and be used to evaluate and optimize the space-time resource allocation for a URTS.\n\nThe space-time flow distribution in the urban rail network can be obtained by the traditional method of investigation and statistics \\[[@pone.0188874.ref003]--[@pone.0188874.ref004]\\], on condition that the rail networks and time-varying O-D demands are stable, but the workload is very heavy and the results are restricted by the schedule and transportation capacity distribution. Another method for solving the problem of DAURTN is the schedule-based transit assignment models \\[[@pone.0188874.ref005]--[@pone.0188874.ref014]\\]. When considering the travel behaviors such as capacity constraints, space priority, first-come-first-serve (FCFS) and congestion effects together, the scales of networks studied in the existing literature are very small, and it is difficult to solve the realistic scale networks in the schedule-based transit assignment. Moreover, it is obvious that the results are also restricted by the schedule, which make the time-varying section flow be discrete.\n\nTo fill up the research gap, we proposed the concept of continuous transmission, which means that the urban rail line can serve passengers at any time and it is not restricted by the train's schedule times. The concept of continuous transmission is similar to the continuous network in continuum equilibrium traffic assignment models \\[[@pone.0188874.ref015]--[@pone.0188874.ref018]\\]. The difference between the two concepts is that, for the concept of continuous transmission, the line service is continuous in the time axis and there is no schedule, while the concept of continuous network means the road network is approximated as a continuum \\[[@pone.0188874.ref015]\\]. With the concept of continuous transmission and maximal rail line capacity constraints, the problem of DAURTN is a new challenging problem considering the travel behaviors such as the capacity constraints, space priority, FCFS and congestion effects together. This problem of DAURTN plays an important role in evaluating and optimizing the schedule for a URTS, as well as optimizing the frequency and the rolling stock circulation, and evaluating the transportation capacity for urban rail networks. For solving the above problem, we design a cell transmission model (CTM).\n\nThe dynamic assignment problem (DAP) derives from urban road traffic assignment research, which can reflect how vehicles make appropriate route choices at any time according to the current road flow state, and also can depict travel choice behaviors more precisely. For solving the DAP, many works \\[[@pone.0188874.ref019]--[@pone.0188874.ref030]\\] in the literature have studied how to establish and solve models, and there are two classes of models, i.e., instantaneous dynamic route choice model and idea dynamic route choice model \\[[@pone.0188874.ref022]\\]. To solve the DAP considering the capacity constraints, there are also abundant studies \\[[@pone.0188874.ref031]--[@pone.0188874.ref037]\\].\n\nThe CTM is a practical method for dynamic traffic assignment, and has a relatively high solving efficiency. Danganzo \\[[@pone.0188874.ref038]--[@pone.0188874.ref039]\\] started to use the concept of a cellular automaton, and proposed a CTM to study the dynamic traffic problem. Lo and Szeto \\[[@pone.0188874.ref040]\\] developed a cell-based dynamic traffic assignment formulation which follows the ideal dynamic user optimal principle. This formulation automatically satisfies the first-in-first-out (FIFO) conditions as a result of encapsulating CTM, and can capture dynamic traffic phenomena such as shockwaves, queue formation and dissipation. Szeto and Lo \\[[@pone.0188874.ref041]\\] further considered departure time choice and elastic demand. For dynamic traffic assignment, they generally assume that traffic flow behaviors follow the FIFO principle, so they apply the CTM to solve the problem. To follow FIFO, they divide the flow into small groups according to different routes, and make them outflow from current cells on the basis of their proportion of the gross, but neglect the time sequence when they flow into the cells. FIFO is also discussed in Daganzo \\[[@pone.0188874.ref039]\\], Lo and Szeto \\[[@pone.0188874.ref040]\\], Carey \\[[@pone.0188874.ref042]--[@pone.0188874.ref043]\\], Blumberg and Bar-Gera \\[[@pone.0188874.ref044]\\], Long *et al*. \\[[@pone.0188874.ref045]\\] and Carey *et al*. \\[[@pone.0188874.ref046]\\]. In particular, Carey *et al*. \\[[@pone.0188874.ref046]\\] indicated that the usual recommended method for preserving FIFO will ensure FIFO for each cell taken separately, but does not fully ensure FIFO in the transition between cells or for links or for routes, and the paper is concerned with how to implement FIFO in the CTM. However, the CTM is not used to solve the transit assignment problem.\n\nFor transit assignment problem, there are two categories: one is frequency-based, and the other is schedule-based. The latter considers time-varying O-D demand. Tong and Wong \\[[@pone.0188874.ref005]\\] proposed a schedule-based, stochastic, dynamic transit assignment model, and a stochastic minimum path is generated by a specially developed branch and bound algorithm. Nuzzolo *et al*. \\[[@pone.0188874.ref006]\\] developed a dynamic process assignment model, both within-day and day-to-day, and tested it on a realistically sized network to verify its applicability for operations planning. Nguyen *et al*. \\[[@pone.0188874.ref007]\\] presented a new graph theoretic framework for the passenger assignment problem that encompassed the departure time and the route choice. The implicit FIFO access to transit lines was taken into account by the concept of available capacity. Poon *et al*. \\[[@pone.0188874.ref008]\\] proposed a predictive transit dynamic user equilibrium model, and the generalised cost function encompassed four components: in-vehicle time, waiting time, walking time, and a line change penalty. Passengers queued at platforms under the single channel first-in-first-out discipline. By using time-increment simulation, the passenger demand was loaded onto the network and the available capacity of each vehicle was updated dynamically. Hamdouch and Lawphongpanich \\[[@pone.0188874.ref009]\\] and Hamdouch *et al*. \\[[@pone.0188874.ref010]--[@pone.0188874.ref011]\\] proposed a user equilibrium transit assignment model that took into account transit schedules and individual vehicle capacities explicitly. When loading a vehicle, on-board passengers continuing to the next stop had priority and waiting passengers could be loaded on a FCFS or in a random manner. Sumalee *et al*. \\[[@pone.0188874.ref012]\\] proposed a stochastic dynamic transit assignment model with an explicit seat allocation process. Two priority rules were assumed in the seat allocation simulation: passengers arriving earlier at a stop can access the available seats prior to those arriving later; standing passengers already on-board can access the available seats prior to those just boarding at the stop/station. Zhang *et al*. \\[[@pone.0188874.ref013]\\] proposed a new multi-class user reliability-based dynamic transit assignment model, and the in-vehicle capacity constraint for random passenger demand was handled by an in-vehicle congestion parameter. Nuzzolo *et al*. \\[[@pone.0188874.ref014]\\] presented a schedule-based dynamic assignment model for transit networks, which took into account congestion through explicit vehicle capacity constraints, and solved the queue formation and dispersion through FCFS rules, the failure-to-board experience, as well as experienced LoS attributes.\n\nIn order to obtain the space-time flow distribution in the urban rail network, without the restriction of schedules, we proposed the concept of continuous transmission for DAURTN. To solve the problem of DAURTN based on the continuous transmission and maximal line capacity constraints, we design the cell transmission mechanism, and develop the CTM. In the construction of the cell transmission rules, the model considers the priority principle, the queuing process, the capacity constraint, and congestion effect. We design an efficient algorithm for solving the shortest path in the urban rail network, which decreases the computation cost of the algorithm for the CTM to be implemented on a large-scale network. Using the method of successive average (MSA), the instantaneous dynamic user optimal state can be reached. Many important indexes are generated by the CTM, which provides the effective support for the optimization of train schedules and the capacity evaluation for the urban rail transit network. We take a small-scale network and the Beijing Metro network as two numerical examples to show the model and its potential application.\n\nNomenclatures {#sec002}\n=============\n\n1.  *n*: total number of stations;\n\n2.  *m*: total number of lines;\n\n3.  *s*^*u*^: the *u*th station;\n\n4.  *S*: set of stations;\n\n5.  *l*: the *l*th line;\n\n6.  *U*: the up line direction;\n\n7.  *D*: the down line direction;\n\n8.  *d*: the direction variable (*d* \u2208 {U,D});\n\n9.  \u03a9: set of directional lines;\n\n10. *L*~*ld*~: direction *d* of line *l* in \u03a9;\n\n11. $s_{ld}^{i}$: the *i*th platform in the direction *d* of line *l*;\n\n12. *n*(*l*): number of stations of line *l*;\n\n13. $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$: the *i*th line section in the direction *d* of line *l* (*i* \u2264 *n*(*l*) \u2212 1);\n\n14. *E*: set of sections;\n\n15. ${\\overline{s}}_{ld}^{i}$: corresponding station of platform $s_{ld}^{i}$;\n\n16. *S*(*s*^*u*^): set of corresponding platforms of station *s*^*u*^;\n\n17. $d\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$: mileage of the section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$;\n\n18. $t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$: travel time of the section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$;\n\n19. \\[*T*~1~,*T*~2~\\]: operation period of the urban rail network;\n\n20. *\u03c4*~*l*~: minimum headway of line *l*;\n\n21. *C*~*l*~: maximum capacity of a train for line *l*;\n\n22. *S*~*F*~: set of transfer stations;\n\n23. \u0394*T*: the interval time;\n\n24. *N*: total intervals of \\[*T*~1~,*T*~2~\\];\n\n25. *t*~*k*~: the *k*th interval (*k* = 1,2,\u22ef,*N*);\n\n26. *RS*: set of O-D pairs;\n\n27. *q*~*rs*~(*t*~*k*~): O-D demand of pair (*r*,*s*);\n\n28. *V*: node set of the urban rail transmission network (*V* = *S* \u222a *S*~*\u03a9*~);\n\n29. *S*~*\u03a9*~: set of platforms;\n\n30. *A*~qu~: set of queuing arcs which point from the station node to the platform node;\n\n31. *A*~ar~: set of arriving arcs which point from the platform node to the station node;\n\n32. *A*~tr~: set of transmission arcs;\n\n33. *A*: set of arcs (*A* = *A*~qu~ \u222a *A*~ar~ \u222a *A*~tr~);\n\n34. *a*: an arc in *A*;\n\n35. *x*~*a*~(*t*~*k*~): number of passengers in the arc *a* \u2208 *A* in *t*~*k*~;\n\n36. $x_{a}^{qu}\\left( t_{k} \\right)$: number of passengers along arc *a* \u2208 *A*~qu~ in *t*~*k*~;\n\n37. $x_{a}^{ar}\\left( t_{k} \\right)$: number of passengers along arc *a* \u2208 *A*~ar~ in *t*~*k*~;\n\n38. $x_{a}^{tr}\\left( t_{k} \\right)$: number of passengers along arc *a* \u2208 *A*~tr~ in *t*~*k*~;\n\n39. ${\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)$: number of passengers departing from arc *a* \u2208 *A*~qu~ in *t*~*k*~;\n\n40. *c*(*a*): cost of the arc *a* \u2208 *A*;\n\n41. *c*~qu~(*a*): cost of the arc *a* \u2208 *A*~qu~;\n\n42. *c*~ar~(*a*): cost of the arc *a* \u2208 *A*~ar~;\n\n43. *c*~tr~(*a*): cost of the arc *a* \u2208 *A*~tr~;\n\n44. *\u03bb*: a parameter (0 \\< *\u03bb* \\< 0.5);\n\n45. *\u03b8*: converted factor of time transforming to cost;\n\n46. *k*(*s*^*u*^): number of time intervals transferring at *s*^*u*^;\n\n47. *\u03b7*,*\u03b1*: cost parameters;\n\n48. $n_{ld}^{i}$: number of cells in section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$;\n\n49. Cell(*s*^*u*^): station cell of station *s*^*u*^;\n\n50. Cell(*L*~*ld*~,*i*,*j*): the *j*th transmission cell in the *i*th section of directional line *L*~*ld*~;\n\n51. *y*~*h*~(*s*^*u*^,*t*~*k*~): flow in station cell Cell(*s*^*u*^,*t*~*k*~);\n\n52. *y*~*h*~(*L*~*ld*~,*i*,*j*,*t*~*k*~): flow in transmission cell Cell(*L*~*ld*~,*i*,*j*,*t*~*k*~);\n\n53. *x*~*h*~(*s*^*u*^,*t*~*v*~), flow which arrives at station *s*^*u*^ in *t*~*v*~, and is detained at the station in *t*~*k*~;\n\n54. *f*~*h*~(*L*~*ld*~,*i*,*end*,*t*~*k*~): outflow volume of tail cell Cell(*L*~*ld*~,*i*,*end*) in *t*~*k*~;\n\n55. *m*(*L*~*ld*~,*i*): number of cells in the *i*th section of the directional line *L*~*ld*~;\n\n56. $H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right)$: set of destinations to which the shortest path from platform $s_{ld}^{i + 1}$ passes through $s_{ld}^{i + 2}$ in *t*~*k*~;\n\n57. $H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right)$: set of destinations to which the shortest path from station *s*^*u*^ passes through $s_{ld}^{i} \\in S\\left( s^{u} \\right)$ in *t*~*k*~;\n\n58. *C*~rem~(*L*~*ld*~,*i*,1,*t*~*k*~): surplus capacity of head cell Cell(*L*~*ld*~,*i*,1) in *t*~*k*~;\n\n59. $M\\left( {s_{ld}^{i},t_{k}} \\right)$: flow from Cell(*s*^*u*^) to $Cell\\left( {L_{ld},i,1} \\right),s_{ld}^{i} \\in S\\left( s^{u} \\right)$;\n\n60. $s_{F}^{0}$: the first transfer station of a shortest path;\n\n61. $s_{F}^{1}$: the last transfer station of a shortest path;\n\n62. $p\\left( {s,s_{F}^{0}} \\right)$: length of the first segment of the shortest path from the origin *s* to the first transfer station $s_{F}^{0} \\in S$ along a directional line *L*^0^ \u2208 *\u03a9*;\n\n63. $p\\left( {s_{F}^{1},s^{u}} \\right)$: length of the last segment of the shortest path from the last transfer station $s_{F}^{1} \\in S$ to the destination *s*^*u*^ along a directional line *L*^1^ \u2208 *\u03a9*;\n\n64. *G*(*L*~*lU*~,*L*~*lD*~): network composed of directional lines *L*~*lU*~ and *L*~*lD*~;\n\n65. $Z\\left( {s_{ld}^{i},t_{k}} \\right)$: detained flow on the platform $s_{ld}^{i}$.\n\nQuantitative description of relevant concepts in the urban rail network {#sec003}\n=======================================================================\n\nUrban rail network {#sec004}\n------------------\n\nAn urban rail network comprises a number of lines, and a line comprises a number of stations and sections. In urban rail system, trains are usually planned individually for each line. We assume that trains do not run across lines, and passengers can move across lines by transfer stations.\n\nGiven an urban rail network with *n* stations and *m* lines. The set of stations is denoted as *S* = {*s*^1^,*s*^2^,\u22ef,*s*^*n*^}. Urban rail lines are almost linear, and some complex urban rail lines, for example, annular lines or Y-style lines, can be decomposed into the linear lines, so we represent urban rail lines as linear. Each line includes two directional lines according to two opposite directions of operation, and the set of directional lines can be denoted as \u03a9 = {*L*~1U~,*L*~1D~,*L*~2U~,*L*~2D~,\u22ef,*L*~*m*U~,*L*~*m*D~}, where *L*~*l*U~, *L*~*l*D~ denotes two directional lines of line *l*. We denote the direction variable as *d* \u2208 {U,D}, and the directional line as *L*~*ld*~ \u2208 \u03a9. For describing the queuing of passengers at the station, a station can be extended into many platforms for each directional line, and each platform only serves for a unique direction line, so a directional line can be described as a sequence of the platforms. We denote $s_{ld}^{i}$ as the *i*th platform of the directional line *L*~*ld*~, and *n*(*l*) as the number of the stations serving by line *l*. Then directional line *L*~*ld*~ can be denoted as a sequence of platforms $\\left( s_{ld}^{1},s_{ld}^{2},\\cdots,s_{ld}^{n(l)} \\right)$. Denote the set of platforms as $S_{\\Omega} = \\left\\{ {s_{ld}^{i},i = 1,2,\\cdots,n\\left( l \\right),L_{ld} \\in \\Omega} \\right\\}$. A section can be denoted as a dual group $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$. Therefore, the directional line *L*~*ld*~ can also be denoted as a sequence of sections $\\left( (s_{ld}^{1},s_{ld}^{2}),(s_{ld}^{2},s_{ld}^{3}),\\cdots,(s_{ld}^{n(l) - 1},s_{ld}^{n(l)}) \\right)$, and $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right) \\in L_{ld}$. Let *E* denote the set of sections. We define the stations passed through by two or more lines as the transfer stations, and let *S*~*F*~ \u2282 *S* denote the set of transfer stations. For convenience, we let ${\\overline{s}}_{ld}^{i} \\in S$ denote the corresponding station of platform $s_{ld}^{i}$, and *S*(*s*^*u*^) denote the set of platforms corresponding to station *s*^*u*^. For Section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$, we denote $d\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$ as the mileage of the section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$, and $t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$ as the travel time of the section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$.\n\nWe denote the urban rail network as (*V*,*A*), where the node set *V* = *S* \u222a *S*~\u03a9~. Station nodes mainly describe the departure, arrival and transfer of passengers, while platform nodes mainly describe the passenger flow getting on/off trains, queuing and passing through stations. $A_{qu} = \\left\\{ \\left( {s^{u},s_{ld}^{i}} \\right),s_{ld}^{i} \\in S\\left( s^{u} \\right),s^{u} \\in S \\right\\}$ denotes the set of queuing arcs which point from station nodes to platform nodes, $A_{ar} = \\left\\{ {\\left( {s_{ld}^{i},{\\overline{s}}_{ld}^{i}} \\right),s_{ld}^{i} \\in S_{\\Omega}} \\right\\}$ denotes the set of arriving arcs which point from the platform nodes to their corresponding station nodes and $A_{tr} = \\left\\{ {\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right),s_{ld}^{i},s_{ld}^{i + 1} \\in S_{\\Omega}} \\right\\}$ denotes the set of transmission arcs. Then, *A* = *A*~qu~ \u222a *A*~ar~ \u222a *A*~tr~.\n\nThe urban rail network with a single line is illustrated in **[Fig 1](#pone.0188874.g001){ref-type=\"fig\"}**, where the dashed line represents an urban rail operating line and doesn't belong to the urban rail network. The hollow nodes represent the station nodes and the solid nodes represent the platform nodes. The upside part in **[Fig 1](#pone.0188874.g001){ref-type=\"fig\"}** shows one directional line, while the downside part shows the other directional line. **[Fig 2](#pone.0188874.g002){ref-type=\"fig\"}** shows an urban rail network with three lines. In **[Fig 2](#pone.0188874.g002){ref-type=\"fig\"}**, ${\\overline{s}}_{1U}^{2} = {\\overline{s}}_{1D}^{3} = {\\overline{s}}_{3U}^{2} = {\\overline{s}}_{3D}^{3} = s^{1}$, and $S\\left( s^{1} \\right) = \\left\\{ {s_{1U}^{2},s_{1D}^{3},s_{3U}^{2},s_{3D}^{3}} \\right\\}$.\n\n![Illustration of the urban rail network with a single line.](pone.0188874.g001){#pone.0188874.g001}\n\n![Illustration of the urban rail network with three lines.](pone.0188874.g002){#pone.0188874.g002}\n\nThe above described network structure does not include the annular lines or Y-style lines (a Y-style line is a structure where two separate lines merge into one at a station), but it is easy to transform them into linear lines. For example, if $s_{ld}^{1}$ and $s_{ld}^{n(l)}$ are viewed as the same platform, the above network can describe the annular lines. For simplicity, we do not give specialized descriptions for lines with those special structures. Moreover, the above network does not describe the train stopping process in order to decrease the scale of the urban rail network.\n\nConstrained continuous transmission {#sec005}\n-----------------------------------\n\nTo avoid the restrictions of the schedules, the concept of continuous transmission is introduced. Continuous transmission means that each rail line can serve passengers at any time and passenger's traveling is not restricted by the train's schedule times, just like the road transportation.\n\nTo account for the urban rail's capacity constraint, continuous transmission has a capacity constraint, which means that the passenger transmission intensity of directional line *L*~*ld*~ at any time cannot exceed the transmission capacity *C*~*l*~/*\u03c4*~*l*~, where *C*~*l*~ is the capacity of each trains in line *l*, and *\u03c4*~*l*~ is the minimum headway of line *l*. We denote \\[*T*~1~,*T*~2~\\] as the operation period of the urban rail network. Divide \\[*T*~1~,*T*~2~\\] into *N* equal intervals by the interval \u0394*T*, and denote *t*~*k*~, *k* = 1,2,\u22ef,*N*, as the *k*th interval. For each time interval, its transmission capacity is \u0394*T C*~*l*~/*\u03c4*~*l*~, *L*~*ld*~ \u2208 \u03a9.\n\nThe benefit of the introduction of continuous transmission is that the space-time flow distribution obtained by DAURTN is not restricted by urban rail schedule, and can be used to evaluate and optimize the space-time resource allocation, for example, the schedule, the rolling stock circulation, and so on.\n\nPriority principle {#sec006}\n------------------\n\nIn each time interval, passenger flow cannot exceed the transmission capacity. When the flow exceeds the transmission capacity, only part of passengers can be transmitted during the current time interval, and surplus passengers have to wait at the station. According to the travel behavior of urban rail transit, the passengers' choices for different O--D pairs with capacity constrains must obey the following priority principles:\n\nSpace priority principle: the flows of upstream stations along the directional line have priority over those of downstream stations to occupy capacities.\n\nThis principle is from the papers of \\[[@pone.0188874.ref009]--[@pone.0188874.ref011]\\] and it means when loading a vehicle, on-board passengers continuing to the next stop had priority. Under the space priority principle, passengers at upstream stations will not reserve capacity for the waiting passengers at downstream stations.\n\nFirst-come-first-serve (FCFS) principle: passengers arriving earlier have priority over those arriving later to obtain service at a station.\n\nUnder the first-come-first-serve principle, the limited transmission capacity will be provided for passengers in the order of the batches arriving at the stations, and for one batch of passengers with different destinations, the method of equal proportional competition is used to determine the flow of departing passengers \\[[@pone.0188874.ref010]\\].\n\nDemands and costs {#sec007}\n-----------------\n\nThe origins and destinations of all O--D pairs belong to the station node set *S*. We denote *RS* as the set of O--D pairs, and *q*~*rs*~(*t*~*k*~), (*r*,*s*) \u2208 *RS*, 1 \u2264 *k* \u2264 *N* as O--D demands at time interval *t*~*k*~. In this study, it is assumed that passengers follow the instantaneous dynamic route choice principle \\[[@pone.0188874.ref022]\\], i.e., passengers choose the minimal cost routes under the currently time interval. Passenger flow reaches the instantaneous dynamic user optimal state that for each O-D pair at each decision node at each time interval, the instantaneous travel costs for all routes that are being used equal the minimal instantaneous route travel time \\[[@pone.0188874.ref022]\\].\n\nIn urban rail network (*V*,*A*), for any interval *t*~*k*~, 1 \u2264 *k* \u2264 *N*, denote *x*~*a*~(*t*~*k*~) as the flow on arc *a* \u2208 *A* in *t*~*k*~. When *a* \u2208 *A*~qu~,*A*~ar~ or *A*~tr~, use $x_{a}^{qu}\\left( t_{k} \\right),x_{a}^{ar}\\left( t_{k} \\right),x_{a}^{tr}\\left( t_{k} \\right)$ to replace *x*~*a*~(*t*~*k*~) respectively. Especially for *a* \u2208 *A*~tr~, $x_{a}^{tr}\\left( t_{k} \\right)$ is equal to the cumulating flow of differences between inflow and outflow from time interval *t*~1~ to *t*~*k*~, with the concept of the continuous transmission. This method of calculation is similar to that for road link flow in dynamic route choice models \\[[@pone.0188874.ref022]\\]. We denote ${\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)$ as the flow departing and transmitted from arc *a* \u2208 *A*~qu~ in *t*~*k*~. The cost of arc *a* is denoted as *c*(*a*), *a* \u2208 *A*. When *a* \u2208 *A*~qu~,*A*~ar~ or *A*~tr~, we use *c*~qu~(*a*),*c*~ar~(*a*) or *c*~tr~(*a*) to replace *c*(*a*) respectively.\n\nBased on the continuous transmission and instantaneous dynamic route choice principle, we calculate the queuing time by the flow state in the current time interval, i.e., the total queuing flow $x_{a}^{qu}\\left( t_{k} \\right)$ and the transmitted passenger flow ${\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)$ at the platform of the queuing arc *a* in the current time interval. Thus, the number of time intervals queuing at platform is ${x_{a}^{qu}\\left( t_{k} \\right)}/{{\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)}$, and the queuing time is ${\\Delta Tx_{a}^{qu}\\left( t_{k} \\right)}/{{\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)}$. However, ${\\overline{x}}_{a}^{qu}\\left( t_{k} \\right)$ may tend to or be equal to zero, which will result in too large congestion cost, so we assume that denominator has a lower limit, which is set to be *\u03bbC*~*l*~, where *\u03bb* is a parameter. Therefore, the queuing time estimated by passengers is $${\\Delta Tx_{a}^{qu}\\left( t_{k} \\right)}/{\\max\\left\\{ {{\\overline{x}}_{a}^{qu}\\left( t_{k} \\right),\\lambda C_{l}} \\right\\}}$$ and then, the queuing cost can be expressed as $$c_{qu}\\left( a \\right) = \\theta{{\\Delta Tx_{a}^{qu}\\left( t_{k} \\right)}/{\\max\\left\\{ {{\\overline{x}}_{a}^{qu}\\left( t_{k} \\right),\\lambda C_{l}} \\right\\}}}$$ where *\u03b8* is the converted factor for transforming time to cost.\n\nFor an arriving arc *a* \u2208 *A*~ar~,*c*~ar~(*a*) is the cost of the average transfer walking time, and needed to be expressed as a multiple of \u0394*T*. We denote *s*~*a*~ as the arrival station of arc *a*. Assuming that the average transfer time at *s*~*a*~ is a constant, denote it as const(*s*~*a*~), then $$c_{ar}\\left( a \\right) = \\theta const\\left( s_{a} \\right)$$ Denote the number of time intervals for transferring at station *s*~*a*~ as *k*(*s*~*a*~), then $$k\\left( s_{a} \\right) = \\left\\lceil {{const\\left( s_{a} \\right)}/{\\Delta T}} \\right\\rceil$$ where \u2308*x*\u2309 is the function of minimum integer no less than *x*.\n\nFor a transmission arc *a* \u2208 *A*~tr~, the cost *c*~tr~(*a*) is the sum of section travel cost and congestion cost, namely $$c_{tr}\\left( a \\right) = \\theta t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right) + g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right)$$ where $g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right)$ is the congestion cost.\n\nFor a transmission arc *a* \u2208 *A*~tr~, the total flow on arc *a* at time interval *t*~*k*~ is $x_{a}^{tr}\\left( t_{k} \\right)$, and the total capacity of arc *a* is denoted as *C*~*a*~, so the congestion cost $g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right)$ of arc *a* can be expressed as follows: $$g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right) = \\eta t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)\\left\\lbrack \\frac{x_{a}^{tr}\\left( t_{k} \\right)}{C_{a}} \\right\\rbrack^{\\alpha}$$ which is similar to the power form used in BPR functions and the congestion functions in the papers of \\[[@pone.0188874.ref010]\\] and \\[[@pone.0188874.ref047]\\], and where *\u03b7* and *\u03b1* are cost parameters and *\u03b7*,*\u03b1* \\> 0. The congestion cost function is increasing with travel time and passenger flow volume. With the concept of constrained continuous transmission, the transmission arc *a* can be regarded as a train with the length $\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)$, of which the capacity per length unit is *C*~*l*~/*\u03c4*~*l*~, so the capacity of the transmission arc *a* \u2208 *A*~tr~ is calculated as $$C_{a} = t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)*{C_{l}/\\tau_{l}}$$ The congestion cost $g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right)$ can be obtained by substituting Eq ([7](#pone.0188874.e068){ref-type=\"disp-formula\"}) into Eq ([6](#pone.0188874.e066){ref-type=\"disp-formula\"}): $$\\begin{array}{l}\n{g\\left( {x_{a}^{tr}\\left( t_{k} \\right)} \\right) = \\eta t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)\\left\\lbrack \\frac{x_{a}^{tr}\\left( t_{k} \\right)}{t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)*{C_{l}/\\tau_{l}}} \\right\\rbrack^{\\alpha}} \\\\\n{\\mspace{81mu} = \\eta{{t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)}^{1 - \\alpha}\\left( {{\\tau_{l}x_{a}\\left( t_{k} \\right)}/C_{l}} \\right)}^{\\alpha}} \\\\\n\\end{array}$$ In Eq ([8](#pone.0188874.e070){ref-type=\"disp-formula\"}), the travel time $t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)$ is fixed and determined, and only $x_{a}^{tr}\\left( t_{k} \\right)$ is variable. Hence, when *\u03b7*,*\u03b1* \\> 0, the calculation of congestion is feasible and the congestion influences in the travel choice of passengers.\n\nCell transmission model for DAURTN {#sec008}\n==================================\n\nCell transmission mechanism {#sec009}\n---------------------------\n\nTo solve the DAURTN based on the continuous transmission, we build the CTM for the urban rail network (*V*,*A*). For describing the CTM, the cell transmission network is constructed based on cell from the network as follows.\n\nWe define each section as a cell chain, and each station as a station cell. For any section $\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$, as travel time $t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$ is fixed, we divide the section into several transmission cells by \u0394*T*. The transmission cells of one section compose a cell chain, and passengers flows can be transmitted forward between the transmission cells. Note that travel time $t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)$ may not be exactly divided by \u0394*T*, so the time length of the tail cell can be equal to or exceed \u0394*T*. The number of cells divided is $n_{ld}^{i} = \\left\\lfloor {{t\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right)}/{\\Delta T}} \\right\\rfloor$, where \u230a*x*\u230b is a function of the maximum integer no larger than *x*. Therefore, we denote Cell(*L*~*ld*~,*i*,*j*) as the *j*th cell in the cell chain of section $\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)$. Denote *m*(*L*~*ld*~,*i*) as the number of cells in the *i*th cell chain, and for simplifying the notation, denote Cell(*L*~*ld*~,*i*,*end*) the last cell in the cell chain of section $\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right)$. Denote *y*~*h*~(*L*~*ld*~,*i*,*j*,*t*~*k*~) as the flow of Cell(*L*~*ld*~,*i*,*j*) traveling to destination *s*^*h*^ in *t*~*k*~. We also denote Cell(*s*^*u*^) as the station cell of *s*^*u*^, and *y*~*h*~(*s*^*u*^,*t*~*k*~) as the flow of Cell(*s*^*u*^) traveling to destination *s*^*h*^ in *t*~*k*~.\n\nThe transmission relationship between cells is illustrated in **[Fig 3](#pone.0188874.g003){ref-type=\"fig\"}**, where a hollow node represents a station cell, a solid node represents a transmission cell, a hollow rectangle represents the corresponding cell chain of a section, and the arrows represent transmission directions. In each time interval, passengers follow the instantaneous dynamic route choice principle \\[[@pone.0188874.ref022]\\] and are transmitted between cells. Transmission mechanism between cells is designed and transmission processes of flows between the cells are classified into 4 groups: $$\\begin{array}{l}\n\\left. {Cell}\\left( {L_{ld},i,end} \\right)\\Longrightarrow{Cell}\\left( {L_{ld},i + 1,1} \\right),\\quad\\forall\\ i \\leq n\\left( l \\right) - 1,L_{ld}; \\right. \\\\\n\\left. {Cell}\\left( {L_{ld},i,end} \\right)\\Longrightarrow{Cell}\\left( {\\overline{s}}_{ld}^{i + 1} \\right),\\quad\\forall\\ i \\leq n\\left( l \\right) - 1,L_{ld}; \\right. \\\\\n\\left. {Cell}\\left( {L_{ld},i,j} \\right)\\Longrightarrow{Cell}\\left( {L_{ld},i,j + 1} \\right),\\quad\\forall j \\leq m\\left( {L_{ld},i} \\right) - 1,i,L_{ld}; \\right. \\\\\n\\left. {Cell}\\left( s^{u} \\right)\\Longrightarrow{Cell}\\left( {L_{ld},i,1} \\right),\\quad s_{ld}^{i} \\in S\\left( s^{u} \\right),\\forall u,i,L_{ld}. \\right. \\\\\n\\end{array}$$ where '\u27f9' means the transmission process of flow from the left cell to the right cell.\n\n![Illustration of the cell transmission network.](pone.0188874.g003){#pone.0188874.g003}\n\nThe flows of station cells and transmission cells at the initial interval *t*~0~ are 0. In *t*~*k*~(*k* \u2265 1), the O--D demand {*q*~*uh*~(*t*~*k*~)\\|*s*^*h*^ \u2208 *S*} inflows into station cell Cell(*s*^*u*^). Therefore, the initial value of the variables are *y*~*h*~(*s*^*u*^,*t*~0~) = 0, *y*~*h*~(*s*^*u*^,*t*~*k*~) = *q*~*uh*~(*t*~*k*~), *y*~*h*~(*L*~*ld*~,*i*,*j*,*t*~*k*~) = 0, \u2200*h*,*u*,*i*,*j*,*L*~*ld*~,*k* \u2265 1.\n\nNext, we analyze the transmission mechanism in 3 steps.\n\nStep 1: The transmission processes Cell(*L*~*ld*~,*i*,*end*) \u27f9 Cell(*L*~*ld*~,*i* + 1,1), and $\\left. Cell\\left( {L_{ld},i,end} \\right)\\Longrightarrow Cell\\left( {\\overline{s}}_{ld}^{i + 1} \\right) \\right.$\n\nAs the length of time in the tail cell Cell(*L*~*ld*~,*i*,*end*) is equals to or greater than \u0394*T*, only a certain proportion of flow can outflow, and the proportion is ${\\Delta T}/\\left\\lbrack {t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right) - \\left( {n_{ld}^{i} - 1} \\right)\\Delta T} \\right\\rbrack$. Thus, the outflow of Cell(*L*~*ld*~,*i*,*end*) in interval *t*~*k*~ is $$f_{h}\\left( {L_{ld},i,end,t_{k}} \\right) = {{y_{h}\\left( {L_{ld},i,j,t_{k - 1}} \\right)\\Delta T}/\\left\\lbrack {t\\left( {s_{ld}^{i},s_{ld}^{i + 1}} \\right) - \\left( {n_{ld}^{i} - 1} \\right)\\Delta T} \\right\\rbrack},\\forall h$$\n\nThen we can obtain the detained flow of the tail cell $$\\left. y_{h}\\left( {L_{ld},i,end,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,end,t_{k - 1}} \\right) - f_{h}\\left( {L_{ld},i,end,t_{k}} \\right),\\forall h \\right.$$ where '\u2190' denotes that the value of the right variable is assigned to the left variable.\n\nAccording to the space priority principle, the outflow *f*~*h*~(*L*~*ld*~,*i*,*end*,*t*~*k*~) of tail cell Cell(*L*~*ld*~,*i*,*end*) has only two choices, i.e., Cell(*L*~*ld*~,*i* + 1,1) or $Cell\\left( {\\overline{s}}_{ld}^{i + 1} \\right)$, and the transmission choice is determined by the instantaneous dynamic route choice principle \\[[@pone.0188874.ref022]\\], i.e., the shortest path from platform $s_{ld}^{i + 1}$ to destination station *s*^*h*^ at the current time interval in network (*V*,*A*). If the shortest path passes through $s_{ld}^{i + 2}$, then flow *f*~*h*~(*L*~*ld*~,*i*,*end*,*t*~*k*~) is transmitted into cell Cell(*L*~*ld*~,*i* + 1,1); otherwise, it is transmitted to station cell $Cell\\left( {\\overline{s}}_{ld}^{i + 1} \\right)$. The above transmission choice is similar to the all-or-nothing assignment, i.e., the flows follow the shortest path.\n\nWe denote the set of destinations to which the shortest path from platform $s_{ld}^{i + 1}$ passes through $s_{ld}^{i + 2}$ as $$H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right) = \\left\\{ {s^{h} \\in S} \\middle| {in\\ t_{k},\\ the\\ shortest\\ path\\ from\\ s_{ld}^{i + 1}\\ to\\ s^{h}\\ passes\\ through\\ s_{ld}^{i + 2}} \\right\\}$$\n\nWhen $s^{h} \\in H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right)$, the flow *f*~*h*~(*L*~*ld*~,*i*,*end*,*t*~*k*~) is transmitted from Cell(*L*~*ld*~,*i*,*end*) to Cell(*L*~*ld*~,*i* + 1,1). Thus, $$\\left. y_{h}\\left( {L_{ld},i + 1,1,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i + 1,1,t_{k}} \\right) + f_{h}\\left( {L_{ld},i,end,t_{k}} \\right),s^{h} \\in H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right) \\right.$$\n\nWhen $s^{h} \\notin H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right),s^{h} \\neq {\\overline{s}}_{ld}^{i + 1}$, the flow *f*~*h*~(*L*~*ld*~,*i*,*end*,*t*~*k*~) is transmitted from Cell(*L*~*ld*~,*i*,*end*) to $Cell\\left( {\\overline{s}}_{ld}^{i + 1} \\right)$. Note that the average transfer time at station ${\\overline{s}}_{ld}^{i + 1}$ is $k\\left( {\\overline{s}}_{ld}^{i + 1} \\right)$, so $$\\left. y_{h}\\left( {\\overline{s}}_{ld}^{i + 1},t_{k + k({\\overline{s}}_{ld}^{i + 1})} \\right)\\leftarrow y_{h}\\left( {\\overline{s}}_{ld}^{i + 1},t_{k + k({\\overline{s}}_{ld}^{i + 1})} \\right) + f_{h}\\left( {L_{ld},i,end,t_{k}} \\right),s^{h} \\notin H\\left( {s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}} \\right),s^{h} \\neq {\\overline{s}}_{ld}^{i + 1} \\right.$$\n\nIn order to realize the FCFS in transmission mechanism, we introduce a variable *x*~*h*~(*s*^*u*^,*t*~*v*~), 1 \u2264 *v* \u2264 *N*, which represents the flows arriving at station *s*^*u*^ in *t*~*v*~ and detained at the station in *t*~*k*~.\n\n![](pone.0188874.e098.jpg){#pone.0188874.e098g}\n\nx\n\nh\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n,\n\nt\n\nk\n\n\\+\n\nk\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n)\n\n)\n\n\u2190\n\nx\n\nh\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n,\n\nt\n\nk\n\n\\+\n\nk\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n)\n\n)\n\n\\+\n\ny\n\nh\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n,\n\nt\n\nk\n\n\\+\n\nk\n\n(\n\ns\n\n\u00af\n\nl\n\nd\n\ni\n\n\\+\n\n1\n\n)\n\n)\n\nStep 2: The transmission process Cell(*L*~*ld*~,*i*,*j*) \u27f9 Cell(*L*~*ld*~,*i*,*j* + 1)\n\nAfter Step 1, in the tail cell of the cell chain, there may be some detained flows, and then the flow of the tail cell equals to the detained flows plus the flows from Cell(*L*~*ld*~,*i*,*end* \u2212 1), so $$\\left. y_{h}\\left( {L_{ld},i,end,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,end,t_{k}} \\right) + y_{h}\\left( {L_{ld},i,end - 1,t_{k - 1}} \\right) \\right.$$\n\nFor other cells in the chain, it is only need to move flows from the forward cell to the backward cell in the chain, that is, $$\\left. y_{h}\\left( {L_{ld},i,j + 1,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,j,t_{k - 1}} \\right),\\ 1 \\leq j \\leq m\\left( {L_{ld},i} \\right) - 2 \\right.$$\n\nStep 3: The transmission process $\\left. Cell\\left( s^{u} \\right)\\Longrightarrow Cell\\left( {L_{ld},i,1} \\right),s_{ld}^{i} \\in S\\left( s^{u} \\right) \\right.$\n\nAccording to the principle of the space priority, flows from Cell(*L*~*ld*~,*i* \u2212 1,*end*) are transmitted to Cell(*L*~*ld*~,*i*,1) and occupy the capacity of Cell(*L*~*ld*~,*i*,1) with priority. Thus, the surplus capacity of Cell(*L*~*ld*~,*i*,1) in *t*~*k*~ is $$C_{rem}\\left( {L_{ld},i,1,t_{k}} \\right) = \\Delta T\\ {C_{l}/\\tau_{l}} - y_{h}\\left( {L_{ld},i,1,t_{k}} \\right)$$ The flows in *t*~*k*~, which are queuing at station *s*^*u*^ and head to Cell(*L*~*ld*~,*i*,1), have to compete for the surplus capacity with the FSFC principle.\n\nIn order to determine the queuing flow, passengers at station cell Cell(*s*^*u*^) first determine which platform to queue. Similar to the method in Step 1, passengers determined the platform by the shortest path from station *s*^*u*^ to destination station *s*^*h*^ at the current time interval in network (*V*,*A*). If the shortest path passes through $s_{ld}^{i} \\in S\\left( s^{u} \\right)$, then flows traveling to destination station *s*^*h*^ queue on platform $s_{ld}^{i}$. We denote the set of destinations to which the shortest path from station *s*^*u*^ passes through $s_{ld}^{i} \\in S\\left( s^{u} \\right)$ as $$\\left. H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right) = \\left\\{ s^{h} \\in S \\right. \\middle| {in\\ t_{k},\\ the\\ shortest\\ path\\ from\\ s^{u}\\ to\\ s^{h}\\ passes\\ through\\ s_{ld}^{i} \\in S\\left( s^{u} \\right)\\}} \\right.$$ Thus, the flow competing for the surplus capacity is $y_{h}\\left( {s^{u},t_{k}} \\right),\\mspace{9mu} s^{h} \\in H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right),s_{ld}^{i} \\in S\\left( s^{u} \\right)$.\n\nIf ${\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{y_{h}\\left( {s^{u},t_{k}} \\right)}} \\leq C_{rem}\\left( {L_{ld},i,1,t_{k}} \\right)$, then $$\\left\\{ \\begin{array}{l}\n\\left. y_{h}\\left( {L_{ld},i,1,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,1,t_{k}} \\right) + y_{h}\\left( {s^{u},t_{k}} \\right),\\mspace{9mu} s^{h} \\in H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right) \\right. \\\\\n\\left. x_{h}\\left( {s^{u},t_{v}} \\right)\\leftarrow 0,\\mspace{198mu} v < k,s^{h} \\in H\\left( s^{u},s_{ld}^{i},t_{k} \\right) \\right. \\\\\n\\end{array} \\right.$$\n\nIf ${\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{y_{h}\\left( {s^{u},t_{k}} \\right)}} > C_{rem}\\left( {L_{ld},i,1,t_{k}} \\right)$, which means the surplus capacity is insufficient, then there exists $\\overline{k} < k$, and it makes that $${\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{\\sum\\limits_{v < \\overline{k}}{x_{h}\\left( {s^{u},t_{v}} \\right)}}} \\leq C_{rem}\\left( {L_{ld},i,1,t_{k}} \\right) < {\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{\\sum\\limits_{v \\leq \\overline{k}}{x_{h}\\left( {s^{u},t_{v}} \\right)}}}$$ According to the FCFS principle, the flow $\\sum_{v < \\overline{k}}{x_{h}\\left( {s^{u},t_{v}} \\right)}$ traveling to each destination station $s^{h} \\in H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right)$ can be transmitted, i.e., $$\\left\\{ \\begin{array}{l}\n\\left. y_{h}\\left( {L_{ld},i,1,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,1,t_{k}} \\right) + {\\sum\\limits_{v < \\overline{k}}{x_{h}\\left( {s^{u},t_{v}} \\right)}},\\mspace{9mu} s^{h} \\in H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right) \\right. \\\\\n\\left. x_{h}\\left( {s^{u},t_{v}} \\right)\\leftarrow 0,\\mspace{234mu} v < \\overline{k},s^{h} \\in H\\left( s^{u},s_{ld}^{i},t_{k} \\right) \\right. \\\\\n\\end{array} \\right.$$ and a portion of $\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{x_{h}\\left( {s^{u},t_{\\overline{k}}} \\right)}$ can also be transmitted. According to the equal proportion principle, the proportion of flows transmitted can be calculated by $$\\alpha = {\\left\\lbrack {C_{rem}({L_{ld},i,1,t_{k}}) - {\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{\\sum_{v < \\overline{k}}{x_{h}\\left( {s^{u},t_{v}} \\right)}}}} \\right\\rbrack/{\\sum_{s^{h} \\in H(s^{u},s_{ld}^{i},t_{k})}{x_{h}\\left( {s^{u},t_{\\overline{k}}} \\right)}}}$$ Then $$\\left\\{ \\begin{array}{l}\n\\left. y_{h}\\left( {L_{ld},i,1,t_{k}} \\right)\\leftarrow y_{h}\\left( {L_{ld},i,1,t_{k}} \\right) + \\alpha x_{h}\\left( {s^{u},t_{\\overline{k}}} \\right) \\right. \\\\\n{x_{h}\\left( {s^{u},t_{\\overline{k}}} \\right)\\leftarrow\\left( {1 - \\alpha} \\right)x_{h}\\left( {s^{u},t_{\\overline{k}}} \\right)} \\\\\n\\end{array} \\right.,\\quad s^{h} \\in H\\left( {s^{u},s_{ld}^{i},t_{k}} \\right)$$\n\nAfter the above processes, the flow of each cell make a choice by the shortest paths and are all transmitted to the next cell. But the cost of each arc will be changed with the variable flow, so the method of successive average (MSA) is adopted to reach the instantaneous dynamic user optimal state in each time interval. The variables in the above model are updated in MSA.\n\nAn efficient method for solving the shortest path {#sec010}\n-------------------------------------------------\n\nIn the CTM, it is needed to solve the shortest path in *t*~*k*~ from s \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* in network (*V*,*A*). We design a fast method for solving the shortest path as follows.\n\nIf the shortest path from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* passes through several transfer stations, then the shortest path can be divided into three segments at most. The first segment of the shortest path is from origin *s* to the first transfer station $s_{F}^{0} \\in S$, and its length is denoted as $p\\left( {s,s_{F}^{0}} \\right)$. The last segment of the shortest path is from the last transfer station $s_{F}^{1} \\in S$ to destination *s*^*u*^, and its length is $p\\left( {s_{F}^{1},s^{u}} \\right)$. As long as we solve the length of the shortest path between any two transfer stations $p\\left( {s_{F}^{0},s_{F}^{1}} \\right)$, we can obtain the cost of the shortest paths in three cases as follows: $$p\\left( {s,s^{u}} \\right) = \\left\\{ \\begin{array}{l}\n{\\min\\left\\{ p\\left( {s,s_{F}^{0}} \\right) + p\\left( {s_{F}^{0},s_{F}^{1}} \\right) + p\\left( {s_{F}^{1},s^{u}} \\right) \\middle| s_{F}^{0},s_{F}^{1} \\in S_{F},\\exists L^{0},L^{1} \\in \\Omega:s,s_{F}^{0} \\in L^{0},s_{F}^{1},s^{u} \\in L^{1} \\right\\},\\mspace{9mu} s,s^{u} \\notin S_{F}} \\\\\n{\\min\\left\\{ p\\left( {s,s_{F}^{1}} \\right) + p\\left( {s_{F}^{1},s^{u}} \\right) \\middle| s_{F}^{1} \\in S_{F},\\exists L^{1} \\in \\Omega:s^{u},s_{F}^{1} \\in L^{1} \\right\\},\\mspace{189mu} s \\in S_{F},s^{u} \\notin S_{F}} \\\\\n{\\min\\left\\{ p\\left( {s,s_{F}^{0}} \\right) + p\\left( {s_{F}^{0},s^{u}} \\right) \\middle| s_{F}^{0} \\in S_{F},\\exists L^{0} \\in \\Omega:s,s_{F}^{0} \\in L^{0} \\right\\},\\mspace{198mu} s \\notin S_{F},s^{u} \\in S_{F}} \\\\\n\\end{array} \\right.$$\n\nIf the shortest path from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* does not pass through any transfer station, then it will only use one line and can be solved easily.\n\nThe above analysis indicates that the solving method for the shortest path from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* can be decomposed into 3 steps.\n\nStep 1: calculate the shortest path from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* in each network *G*(*L*~*l*U~,*L*~*l*D~), which composed of a pair of opposite directional lines *L*~*l*U~,*L*~*l*D~ shown in **[Fig 1](#pone.0188874.g001){ref-type=\"fig\"}**.\n\nStep 2: calculate the shortest path between each two transfer stations in the network.\n\nStep 3: calculate all the shortest paths from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S*.\n\nIn step 1, for any destination $s^{u} = {\\overline{s}}_{lU}^{v} \\in \\left\\{ {{\\overline{s}}_{lU}^{i},i = 1,2,\\cdots,n\\left( l \\right)} \\right\\}$, we can structure two subsets of nodes bounded by node $s_{lU}^{v}$, i.e. $\\left\\{ s_{lU}^{i},s_{lD}^{({n(l) - i + 1})},{{\\overline{s}}_{lU}^{i},i = 1,2,\\cdots,v} \\right\\}$ and $\\left\\{ s_{lU}^{i},s_{lD}^{({n(l) - i + 1})},{{\\overline{s}}_{lU}^{i},i = v,v + 1,\\cdots,n\\left( l \\right)} \\right\\}$, which can form two generated sub-networks of *G*(*L*~*l*U~,*L*~*l*D~). Obviously, the shortest paths from other nodes to *s*^*u*^ in the two sub-networks are equal to solving the shortest paths in *G*(*L*~*l*U~,*L*~*l*D~). In the former sub-network, there are three cases.\n\nCase 1: solve the shortest paths from nodes $s_{lU}^{i},i = 1,2,\\cdots,v$ to *s*^*u*^ along the directional line *L*~*l*U~;\n\nCase 2: solve the shortest paths from $s_{lD}^{n(l)}$ and ${\\overline{s}}_{lU}^{1}$ to *s*^*u*^ passing through $s_{lU}^{1}$;\n\nCase 3: solve the shortest paths from $s_{lD}^{n(l) - i}$ and ${\\overline{s}}_{lU}^{i}$ to *s*^*u*^ containing the shortest path from $s_{lD}^{n(l) - i + 1}$ to *s*^*u*^, or from $s_{lU}^{i}$ to *s*^*u*^.\n\nThus we can solve the shortest paths from $s_{lD}^{n(l) - i}$ and ${\\overline{s}}_{lU}^{i}$ to *s*^*u*^ in the order of *i* = 2,3,\u22ef,*v*. In the latter sub-network, the solving method is similar. Therefore, the amount of calculation for the shortest paths from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* in network *G*(*L*~*l*U~,*L*~*l*D~) is only *O*(*n*(*l*)^2^), and the sum of calculation for the shortest paths of the whole *m* lines is $O\\left( {\\sum\\limits_{l = 1}^{m}{n\\left( l \\right)}^{2}} \\right)$.\n\nThe method for solving the shortest paths for an annular line only needs to make some supplements based on the above method for linear lines. For any destination *s*^*u*^, we can divide the annular line into a linear line by *s*^*u*^, and there are only two ways of dividing. Then we can adopt the above method to solve the shortest paths from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ for each way of dividing, and the shorter one between them is the shortest path. For a Y-style line, the similar method for solving the shortest paths is feasible.\n\nWe have obtained the length of shortest path between each two transfer stations in each line in step 1. In step 2, we use the Floyd--Warshall algorithm to solve the shortest path between any two transfer stations, and the computational complexity is *O*(\\|*S*~*F*~\\|^3^).\n\nIn step 3, we can solve the shortest paths from *s* \u2208 *S* \u222a *S*~\u03a9~ to *s*^*u*^ \u2208 *S* by formula ([23](#pone.0188874.e124){ref-type=\"disp-formula\"}), and the computational complexity is $O\\left( \\left| S \\middle| \\bullet {\\sum\\limits_{l = 1}^{m}{n\\left( l \\right)}} \\right. \\right)$.\n\nTherefore, the computational complexity of the shortest path solving algorithm in the urban rail transmission network is $O\\left( \\left| S \\middle| \\bullet {\\sum\\limits_{l = 1}^{m}{n\\left( l \\right)}} + \\left| S_{F} \\right|^{3} \\right. \\right)$, which is less than $O\\left( \\left( {\\sum\\limits_{l = 1}^{m}{n\\left( l \\right)}} \\right)^{3} \\right)$ of the classical methods, i.e., Dijkstra algorithm and Floyd algorithm.\n\nThe urban rail network in Beijing composes 15 operating lines and 231 stations until July 2014. There are only 40 transfer stations, which are far less than other stations. Thus, the solving algorithm of the shortest path is effective for the real urban rail network. We test our method and Floyd algorithm for solving the shortest paths of Beijing urban rail network. Using the Matlab(R2010b) to program, the shortest path problem is calculated for 100 times, and we record the CPU times for the two methods. The average CPU time of Floyd algorithm is 10.31s, while the average CPU time of our method is 1.06s, with the computer (IntelCore 2.90GHz, 8GB RAM).\n\nEvaluation indexes of passenger flow {#sec011}\n------------------------------------\n\nThe CTM can generate many important evaluation indexes, including time-varying section flow, operating line circulation volume, the detained passenger volume on the platform, and the queuing length on the platform.\n\nAs the flow $x_{a}^{tr}\\left( t_{k} \\right)$ for *a* \u2208 *A*~tr~ is equal to the sum of the flows of all cells in this cell chain, we defined that the time-varying section flow means the outflow of the last cell in each time interval, which means passenger flow transmitted by the line section in each time interval and can reflect the capacity constraints in CTM. Thus, the section flow is $${\\sum_{h = 1}^{n}{f_{h}\\left( {L_{ld},i,end,t_{k}} \\right)}},\\mspace{9mu}\\forall i,L_{ld},k$$ It is obvious that the time-varying section flow varies with \u0394*T*, i.e., if \u0394*T* becomes longer, then the time-varying section flow for each time interval is larger.\n\nWe can obtain the circulation volume of each directional line, which means the sum of travel mileages of passengers on each directional line, that is, $${\\sum_{k = 1}^{N}{\\sum_{d \\in \\{ U,D\\}}{\\sum_{i = 1}^{n(l) - 1}\\left\\lbrack {d\\left( s_{ld}^{i},s_{ld}^{i + 1} \\right){\\sum_{h = 1}^{n}{f_{h}\\left( {L_{ld},i,end,t_{k}} \\right)}}} \\right\\rbrack}}},\\quad\\forall l$$ and its unit is person kilometer. The circulation volume of the whole network can be obtained by summing the circulation volumes of all directional lines.\n\nIn CTM, we can calculate the detained flow volume on the platform as $$Z\\left( {s_{ld}^{i},t_{k}} \\right) = {\\sum_{v \\leq k}{\\sum_{s^{h} \\in H({s^{u},s_{ld}^{i},t_{k}})}{x_{h}\\left( {s^{u},t_{v}} \\right)}}},\\quad\\forall s_{ld}^{i},k$$\n\nIt is known that the queuing length on the platform depends on the headway *\u03c4*~*l*~, which means the larger the headway is, the longer the queuing length will be. The queuing length on the platform can be used to evaluate the service level of urban railway network. In the CTM, it is noted that the queuing length for each time interval is longest at the beginning of the current time interval, i.e., before flow transmission of each cell at each time interval, while it is shortest when at the ending of time interval, for the reason that some passengers queuing at the platform are transmitted at the ending of the current time interval.\n\nTo eliminate the above influence and obtain the reasonable queuing length to evaluate the service level, we define that the queuing length of each platform at time interval *t*~*k*~ means the total queuing flow during the time interval \\[*t*~*k*~\u0394*T* \u2212 *\u03c4*~*l*~,*t*~*k*~\u0394*T*\\] with headway *\u03c4*~*l*~, and it is calculated at the beginning of the above time interval. Thus, the queuing length includes the flows of the platform transmitted during \\[*t*~*k*~\u0394*T* \u2212 *\u03c4*~*l*~,*t*~*k*~\u0394*T*\\] and the detained flow at the time interval *t*~*k*~. The length of time interval is \u0394*T* and \u0394*T* \\< *\u03c4*~*l*~, so the headway *\u03c4*~*l*~ may cover more than one time interval. It is known that the flow from Cell(*s*^*u*^) to $Cell\\left( {L_{ld},i,1} \\right),s_{ld}^{i} \\in S\\left( s^{u} \\right)$ at time interval *t*~*k*~ is that $$M\\left( {s_{ld}^{i},t_{k}} \\right) = {\\sum_{h = 1}^{n}{y_{h}\\left( {L_{ld},i,1,t_{k}} \\right)}} - {\\sum_{s^{h} \\in H({s_{ld}^{i + 1},s_{ld}^{i + 2},t_{k}})}{f_{h}\\left( {L_{ld},i,end,t_{k}} \\right)}},\\ \\forall s_{ld}^{i},k$$ The flow at the platform $s_{ld}^{i}$ transmitted during \\[*t*~*k*~\u0394*T* \u2212 *\u03c4*~*l*~,*t*~*k*~\u0394*T*\\] is calculated as $$\\sum_{v = k - r + 1}^{k}{M\\left( s_{ld}^{i},t_{v} \\right) + M\\left( s_{ld}^{i},t_{k - r} \\right){\\left( {\\tau_{l} - r\\Delta T} \\right)/{\\Delta T}}}$$ where *r* meets 0 \u2264 *\u03c4*~*l*~ \u2212 *r*\u0394*T* \\< \u0394*T*. As the detained flow on the platform in *t*~*k*~ is $Z\\left( {s_{ld}^{i},t_{k}} \\right)$, so the queuing length on the platform in *t*~*k*~ is $$Z\\left( {s_{ld}^{i},t_{k}} \\right) + {\\sum_{v = k - r + 1}^{k}{M\\left( s_{ld}^{i},t_{v} \\right)}} + M\\left( {s_{ld}^{i},t_{k - r}} \\right){\\left( {\\tau_{l} - r\\Delta T} \\right)/{\\Delta T}},\\mspace{9mu}\\forall s_{ld}^{i},k$$\n\nNumerical examples {#sec012}\n==================\n\nExample 1: a network with three operating lines {#sec013}\n-----------------------------------------------\n\n### Description {#sec014}\n\nTake the urban rail network in **[Fig 2](#pone.0188874.g002){ref-type=\"fig\"}** as an example. This network has three operating lines. Each line has four stations, and there are total nine stations in this urban rail network, including three transfer stations. The section mileage of each line is 1.6 km, the section travel times are all equal to 2.5 min, the minimum headways are all equal to 8 min, and the maximum passenger capacity of each train is 1000 passengers per train. The network operation period is from 6:00 to 23:00.\n\n**[Fig 4](#pone.0188874.g004){ref-type=\"fig\"}** shows that the network is divided into three areas, where area 1 is a Central Business District (CBD), while areas 2 and 3 are residential areas. The density distributions of O--D demands arriving in area 1 are shown in **[Fig 5(A)](#pone.0188874.g005){ref-type=\"fig\"}**; the density distributions of O--D demands departing from area 1 are shown in **[Fig 5(B)](#pone.0188874.g005){ref-type=\"fig\"}**. The demands within each area and between areas 2 and 3 are all equal to 0. The detailed O--D demands and density distributions are listed in **[Table 1](#pone.0188874.t001){ref-type=\"table\"}**. It can be seen from the density distributions of travel demands that all the morning peaks of O--D demands are the period from 7:00 to 9:00, while all the evening peaks of O--D demands are the period from 18:00 to 20:00. The O--D demands between areas 2 and 1 are larger than those between areas 3 and 1.\n\n![Three areas of the urban rail network.](pone.0188874.g004){#pone.0188874.g004}\n\n![Two classes of density distributions of demands.](pone.0188874.g005){#pone.0188874.g005}\n\n10.1371/journal.pone.0188874.t001\n\n###### O--D total demand and corresponding probability density distributions (unit: thousand persons).\n\n![](pone.0188874.t001){#pone.0188874.t001g}\n\n  -------------------------------------------------------------------------------------------------------------\n  O--D total demand   Area 1   Area 2   Area 3                                                         \n  ------------------- -------- -------- -------- -------- -------- -------- -------- -------- -------- --------\n  Area\\               1                                   6.6(b)   5.5(b)   5.5(b)   5.5(b)   4.4(b)   4.4(b)\n  1                                                                                                    \n\n  4                                              5.5(b)   6.6(b)   6.6(b)   4.4(b)   5.5(b)   5.5(b)   \n\n  5                                              5.5(b)   6.6(b)   6.6(b)   4.4(b)   5.5(b)   5.5(b)   \n\n  Area\\               2        6.6(a)   5.5(a)   5.5(a)                                                \n  2                                                                                                    \n\n  6                   5.5(a)   6.6(a)   6.6(a)                                                         \n\n  7                   5.5(a)   6.6(a)   6.6(a)                                                         \n\n  Area\\               3        5.5(a)   4.4(a)   4.4(a)                                                \n  3                                                                                                    \n\n  8                   4.4(a)   5.5(a)   5.5(a)                                                         \n\n  9                   4.4(a)   5.5(a)   5.5(a)                                                         \n  -------------------------------------------------------------------------------------------------------------\n\nSet \u0394*T* = 1.25 min and each section has two cells. Divide the operation period into 816 intervals by \u0394*T*. The line transmission capacity in \u0394*T* is \u0394*T C*~*l*~/*\u03c4*~*l*~ = 1.25 \u00d7 1000/8 = 156.25 passengers. Set parameters *\u03bb* = 0.01, *\u03b8* = 1, *\u03b7* = 0.8, *\u03b1* = 1, and the relative gap is 10^\u22123^.\n\nWith the computer (IntelCore 2.90GHz, 8GB RAM), we use Matlab (R2010b) to program and solve the model for this example, and it takes 19s CPU time to solve this model.\n\n### Analysis of indexes {#sec015}\n\nThe circulation volumes of all lines are listed in **[Table 2](#pone.0188874.t002){ref-type=\"table\"}**, and the circulation volume of the network is 7.92 10^5^ person kilometers.\n\n10.1371/journal.pone.0188874.t002\n\n###### Operating line circulation volumes.\n\n![](pone.0188874.t002){#pone.0188874.t002g}\n\n  -------------------------------------------------------------------------------------\n  Line                                     Line 1         Line 2         Line 3\n  ---------------------------------------- -------------- -------------- --------------\n  **Operating line circulation volume**\\   3.40 \u00d7 10^5^   1.44 \u00d7 10^5^   3.08 \u00d7 10^5^\n  **(person kilometer)**                                                 \n\n  -------------------------------------------------------------------------------------\n\nAs some indexes are time-varying during an operation day, and the data are too large to be listed, we only calculate the statistical indexes about directional line $L_{1U} = \\left\\{ {s_{1U}^{1},s_{1U}^{2},s_{1U}^{3},s_{1U}^{4}} \\right\\}$ to demonstrate and analyze the model.\n\nThe time-varying section flow can be obtained by the method of MSA. There are three curves to show the time-varying section flows of $\\left( {s_{1U}^{1},s_{1U}^{2}} \\right),\\left( {s_{1U}^{2},s_{1U}^{3}} \\right),\\left( {s_{1U}^{3},s_{1U}^{4}} \\right)$ in **[Fig 6](#pone.0188874.g006){ref-type=\"fig\"}**. Note that *L*~1U~ is a directional line in which passengers mainly depart from the CBD, and its travel demand distributions are shown in **[Fig 5(B)](#pone.0188874.g005){ref-type=\"fig\"}**. The time-varying section flows are similar to **[Fig 5(B)](#pone.0188874.g005){ref-type=\"fig\"}**. During evening peak hours, the section flow of $\\left( {s_{1U}^{2},s_{1U}^{3}} \\right)$ reaches transmission capacity 156.25 from 18:06:15 to 20:18:45. The section flow of $\\left( {s_{1U}^{3},s_{1U}^{4}} \\right)$ is the lowest among three curves, because there is no demand starting from ${\\overline{s}}_{1U}^{3}$ along *L*~1U~.\n\n![Section flows along *L*~1U~.](pone.0188874.g006){#pone.0188874.g006}\n\nThere are three curves to show the detained flows on $s_{1U}^{1},s_{1U}^{2},s_{1U}^{3}$ in **[Fig 7](#pone.0188874.g007){ref-type=\"fig\"}**. The detained flow on $s_{1U}^{2}$ begins from 18:00:00, reaches the maximal value 155.91 at 20:02:30, and disappears at 20:20:00. It is for the reason that the demands from the upstream station ${\\overline{s}}_{1U}^{1} = s^{4}$ occupies most capacity, which make the demands departing from ${\\overline{s}}_{1U}^{2}$ during evening peak hours not be met.\n\n![The detained flows of all platforms along *L*~1U~.](pone.0188874.g007){#pone.0188874.g007}\n\nThree curves are illustrated in **[Fig 8](#pone.0188874.g008){ref-type=\"fig\"}** to show the queuing lengths on $s_{1U}^{1},s_{1U}^{2},s_{1U}^{3}$. On $s_{1U}^{1}$, all passengers can be transmitted in time due to the sufficient capacity, then the queuing length distribution is similar to the demand distribution. On $s_{1U}^{2}$, the detained flow begins to appear from 18:00:00 due to the insufficient capacity. As the travel demand intensity will not change from 18:00:00, the queuing length continues to increasing. On $s_{1U}^{3}$, the queuing flow does not appear at any time bacause there is no departure and transfer passenger flow.\n\n![The queuing lengths of all platforms along *L*~1U~.](pone.0188874.g008){#pone.0188874.g008}\n\n### Analysis of passenger flow characteristics {#sec016}\n\nTo demonstrate the reasonability for dynamic assignment with the CTM, we show the path choice, the space priority principle and the flow moving among platforms of one station.\n\n\\(a\\) Path choice\n\nAs passengers follows the instantaneous dynamic route choice principle, we now analyze the path choice for the passengers from platform $s_{1U}^{2}$ to station *s*^9^. We adopt two paths, i.e., path 1: $\\left( {s_{1U}^{2},s^{1},s_{3U}^{2},s_{3U}^{3},s^{3},s_{2U}^{3},s_{2U}^{4},s^{9}} \\right)$ and path 2: $\\left( {s_{1U}^{2},s_{1U}^{3},s^{2},s_{2U}^{2},s_{2U}^{3},s_{2U}^{4},s^{9}} \\right)$. The flows and costs of these two paths are shown in **[Fig 9](#pone.0188874.g009){ref-type=\"fig\"}**, we can see that the costs of path 1 for all time intervals are larger than those of path 2, so all passengers choose the path 2 to travel to *s*^9^.\n\n![The flows and costs of two paths for passengers from platform $\\mathbf{s}_{1\\mathbf{U}}^{2}$ to station *s*^9^.](pone.0188874.g009){#pone.0188874.g009}\n\nNow we consider the path choice for the passengers from platform $s_{2D}^{2}$ to station *s*^4^. We also adopt two paths, i.e., path 1: $\\left( {s_{2D}^{2},s^{3},s_{3D}^{2},s_{3D}^{3},s^{1},s_{1D}^{3},s_{1D}^{4},s^{4}} \\right)$ and path 2: $\\left( {s_{2D}^{2},s_{2D}^{3},s^{2},s_{1D}^{2},s_{1D}^{3},s_{1D}^{4},s^{4}} \\right)$. **[Fig 10](#pone.0188874.g010){ref-type=\"fig\"}** shows the costs and the flows of these two paths. We can see that the costs of the two paths are equal during the time period \\[07:08:45, 09:06:15\\], and the flows of the two paths are larger than zero. In other time periods, the costs of path 1 are larger than those of path 2, so the passengers all choose the path 2.\n\n![The flows and costs of two paths for passengers from platform $\\mathbf{s}_{2\\mathbf{D}}^{2}$ to station *s*^4^.](pone.0188874.g010){#pone.0188874.g010}\n\nFrom **Figs [9](#pone.0188874.g009){ref-type=\"fig\"} and [10](#pone.0188874.g010){ref-type=\"fig\"}**, it is obvious that the dynamic flow of the network is in an instantaneous dynamic user equilibrium state.\n\n\\(b\\) The space priority principle\n\n**[Fig 11](#pone.0188874.g011){ref-type=\"fig\"}** shows the curves of the transmitted passenger flows on $s_{1U}^{1},s_{1U}^{2},s_{1U}^{3}$ along directional line *L*~1U~. The flow on $s_{1U}^{3}$ is 0, and the flow on $s_{1U}^{1}$ shows that the transmitted passenger flow is consistent with the departure demand from *s*^1^. The transmitted flow on $s_{1U}^{2}$ increases sharply from 20:02:30.\n\n![The transmitted flows of all platforms along *L*~1U~.](pone.0188874.g011){#pone.0188874.g011}\n\nTo explain this phenomenon, we compare the curve of the transmitted flow on $s_{1U}^{2}$ with that of the passing flow in **[Fig 12](#pone.0188874.g012){ref-type=\"fig\"}**. As shown in **[Fig 12](#pone.0188874.g012){ref-type=\"fig\"}**, the passing flow on $s_{1U}^{2}$ decreases from 20:02:30, while the transmitted flow on $s_{1U}^{2}$ increases. Moreover, it can be seen in **[Fig 7](#pone.0188874.g007){ref-type=\"fig\"}** that there is no detained flow at $s_{1U}^{1}$ at any time, while the detained flow on $s_{1U}^{2}$ begins to appear from 18:00:00 to 20:20:00. With the space priority principle, when the capacity for $s_{1U}^{1}$ is sufficient, the flow on $s_{1U}^{1}$ can all be transmitted; when the capacity for $s_{1U}^{2}$ is insufficient, the flow on $s_{1U}^{2}$ can be transmitted according to the remaining capacity, then there are some detained flow on $s_{1U}^{2}$. It can be seen in **[Fig 12](#pone.0188874.g012){ref-type=\"fig\"}** that the sum of the transmitted flows and the passing flows of $s_{1U}^{2}$ at each time interval from 18:00:00 to 20:20:00, are exactly equal to the line transmission capacity (156.25 passengers per the unit time \u0394*T*). Thus, the reason for the phenomenon is that the decrease of demand in the upstream platform $s_{1U}^{1}$ makes the passing volume of $s_{1U}^{2}$ decline, and the increased remaining capacity can be used to transmit the detained passengers on $s_{1U}^{2}$, which makes the transmitted flow on $s_{1U}^{2}$ increase sharply. This exactly reflects the space priority principle.\n\n![Transmitted and passing flows on $\\mathbf{s}_{1\\mathbf{U}}^{2}$ along *L*~1U~.](pone.0188874.g012){#pone.0188874.g012}\n\n\\(b\\) Flow moving among platforms of the station\n\nIn each time interval, there are differences between the non-detained flow and the transmitted flow on the platform. The non-detained flow includes O-D demand departing from the platform and the transfer flow arriving at the platform in the current time interval, and it does not include the detained passenger flow. If the non-detained flow is equal to the transmitted flows, the detained flow in the current time interval is equal to that at the previous time interval. If the non-detained flow is larger than the transmitted flow, the detained flow at the current time interval is larger than that at the previous time interval. If the non-detained demand is less than the transmitted flow, the detained flow at the current time interval is less than that at the previous time interval. The curves of the differences between the non-detained flow and the transmitted flow on four platforms in *s*^1^ are illustrated in **[Fig 13](#pone.0188874.g013){ref-type=\"fig\"}**.\n\n![Curves of differences between the non-detained and transmitted flows of all platforms in *s*^1^.](pone.0188874.g013){#pone.0188874.g013}\n\nOn $s_{1D}^{3},s_{3D}^{3} \\in S\\left( s^{1} \\right)$, the value are 0, which indicates that the non-detained flows at any time are not restricted by the capacity. On $s_{1U}^{2} \\in S\\left( s^{1} \\right)$, the value in the period from 17:58:45 to 20:03:45 is larger than 0, which indicates that the capacity in this period cannot satisfy the non-detained flow, and the detained flow appears. The value in the period from 20:05:00 to 20:20:00 is less than 0, which indicates that the capacity in this period exceeds the non-detained demand, and the detained flow is transmitted. Note that the values of $s_{1U}^{2}$ before 17:58:45 and after 20:20:00 are both equal to 0, i.e., the detained flows in the two periods are both equal to 0.\n\nIn view of the figure area, the detained flow from 17:58:45 to 20:03:45 is much larger than the transmitted flow from 20:05:00 to 20:20:00 on the platform $s_{1U}^{2}$, but the detained flow disappear after 20:20:00. To explain this phenomenon, we can see that for $s_{3U}^{2} \\in S\\left( s^{1} \\right)$, the values in the period from 18:00:00 to 20:20:00 are less than 0, that is, the non-detained flow in this period is less than the transmitted flows. At other times, the non-detained demand is equal to the transmitted flows. It is obvious that some detained passengers on $s_{1U}^{2}$ change their travel route, and move to $s_{3U}^{2}$. In view of the figure area, the sum of the two negative areas is equal to the positive area.\n\n### Sensitivity analysis {#sec017}\n\nWe set the parameter *\u03b1* = 1.0,3.0,4.0,4.5,5.0 respectively, and don't change other parameters. The detained flows of platform $s_{1U}^{2}$ are calculated by the CTM for different values of parameter *\u03b1*, shown in **[Fig 14](#pone.0188874.g014){ref-type=\"fig\"}**. From **[Fig 14](#pone.0188874.g014){ref-type=\"fig\"}**, It can be seen that the detained flows of platform $s_{1U}^{2}$ decrease with the increasing of the parameter *\u03b1*. From the demand distributions in [Table 1](#pone.0188874.t001){ref-type=\"table\"}, we know that the demands from Area 1 to Area 2 are larger than those from Area 1 to Area 3. It results in that the congestion in *L*~1U~ is more than that in *L*~3U~. With the increase of parameter *\u03b1*, the cost differences between the paths traveling from Area 1 to Area 2 and from Area 1 to Area 3 increase. It results in that the flows change their paths and the flows shift from paths traveling from Area 1 to Area 2 to those traveling from Area 1 to Area 3. Thus, the increase of parameter *\u03b1* makes the decrease of the detained flows of platform $s_{1U}^{2}$.\n\n![The detained flows of platform $\\mathbf{s}_{1\\mathbf{U}}^{2}$ for different values of parameter *\u03b1*.](pone.0188874.g014){#pone.0188874.g014}\n\nFor parameter *\u03b7*, we now do the sensitivity analysis. Set parameter *\u03b7* = 1.0,1.8,1.9,2.0 and don't change other parameters. We also calculate the detained flows of platform $s_{1U}^{2}$ for different values of parameter *\u03b7*. **[Fig 15](#pone.0188874.g015){ref-type=\"fig\"}** shows the curves of the detained flows of platform $s_{1U}^{2}$ and we can see that the detained flow of platform $s_{1U}^{2}$ decreases with the increasing of the parameter *\u03b1*. The reason is similar to that of parameter *\u03b1*. It results in that the cost differences between the paths traveling from Area 1 to Area 2 and from Area 1 to Area 3 increase with the increase of parameter *\u03b7*. Thus, the increase of parameter *\u03b7* makes the decrease of detained flows of platform $s_{1U}^{2}$.\n\n![The detained flows of platform $\\mathbf{s}_{1\\mathbf{U}}^{2}$ for different values of parameter *\u03b7*.](pone.0188874.g015){#pone.0188874.g015}\n\nExample 2: Beijing Metro network {#sec018}\n--------------------------------\n\nWe take the Beijing Metro network as an example to illustrate the CTM. The total mileage of the network is 414.503 km, the number of stations in the network is 231, including 40 transfer stations, and the network has 15 lines, i.e. Line 1, Line 2, Line 4, Line 5, Line 6, Line 8, Line 9, Line 10, Line 13, Line 14, Line 15, Batong Line, Fangshan Line, Changping Line and Yizhuang Line. Line 2 and Line 10 are annular lines, while the other 13 lines are linear. The minimum headways are all 2.5 min, and the maximum passenger capacity of each train is 1200 passengers per train. The network operation period is from 6:00 to 23:00. The total amount of all O--D demands is 5.496 \u00d7 10^6^.\n\nWe set \u0394*T* = 1 min, and the total number of time intervals is 1020. The transmission capacity of a line within \u0394*T* is that $\\Delta T\\ {C_{l}^{max}/\\tau_{l}} = 1 \\times \\frac{1200}{2.5} = 480$. The relative gap is 10^\u22123^, and the total CPU time is 5.16h.\n\nThe circulation volumes of all lines are illustrated in **[Table 3](#pone.0188874.t003){ref-type=\"table\"}**. The total circulation volume of the network is 8.69 \u00d7 10^7^ person kilometers.\n\n10.1371/journal.pone.0188874.t003\n\n###### Operating line circulation volumes of all lines in Beijing Metro network.\n\n![](pone.0188874.t003){#pone.0188874.t003g}\n\n  ---------------------------------------- ------------ -------------- ---------------- --------------- ---------------\n  **Line**                                 Line 1       Line 2         Line 4           Line 5          Line 6\n\n  **Operating line circulation volume**\\   1.08s10^7^   6.06 \u00d7 10^6^   7.71 \u00d7 10^6^     6.70 \u00d7 10^6^    4.93 \u00d7 10^6^\n  **(person kilometer)**                                                                                \n\n  **Line**                                 Line 8       Line 9         Line 10          Line 13         Line 14\n\n  **Operating line circulation volume**\\   2.40s10^6^   2.42 \u00d7 10^6^   2.40 \u00d7 10^7^     9.92 \u00d7 10^6^    2.97 \u00d7 10^5^\n  **(person kilometer)**                                                                                \n\n  **Line**                                 Line 15      Batong Line    Changping Line   Fangshan Line   Yizhuang Line\n\n  **Operating line circulation volume**\\   2.41s10^6^   3.74 \u00d7 10^6^   1.91 \u00d7 10^6^     1.54 \u00d7 10^6^    2.15 \u00d7 10^6^\n  **(person kilometer)**                                                                                \n  ---------------------------------------- ------------ -------------- ---------------- --------------- ---------------\n\nWe take Line 5 as an example to analyze the space-time flow distribution. There are 23 stations in Line 5 illustrated in **[Fig 16](#pone.0188874.g016){ref-type=\"fig\"}**, where solid dots represent transfer stations, while hollow dots represent the non-transfer stations. There are 7 transfer stations including Songjiazhuang, Ciqikou, Chongwenmen, Dongdan, Dongsi, Yonghegong and Huixinxijie Nankou in Line 5.\n\n![Illustration of Line 5.](pone.0188874.g016){#pone.0188874.g016}\n\n**[Fig 17](#pone.0188874.g017){ref-type=\"fig\"}** illustrates the space-time section flow distribution in the direction from Songjiazhuang to Tiantongyuan North, while **[Fig 18](#pone.0188874.g018){ref-type=\"fig\"}** illustrates that in the opposite direction. Comparing **[Fig 17](#pone.0188874.g017){ref-type=\"fig\"}** with **[Fig 18](#pone.0188874.g018){ref-type=\"fig\"}**, there are a morning peak in the sections from Songjiazhuang to Dongsi in **[Fig 17](#pone.0188874.g017){ref-type=\"fig\"}**, and an evening peak in the sections from Dongsi to Songjiazhuang in **[Fig 18](#pone.0188874.g018){ref-type=\"fig\"}**. There are an evening peak in the sections from Dongdan to Huixinxijie Nankou in **[Fig 17](#pone.0188874.g017){ref-type=\"fig\"}**, and a morning peak in the sections from Huixinxijie Nankou to Dongdan in **[Fig 18](#pone.0188874.g018){ref-type=\"fig\"}**. The space-time flow distributions in the two directions have the character of symmetry in travel time, and show the tidal traffic flow phenomenon. The passenger flow of the morning peak from Songjiazhuang to Dongsi are more intensive than those of the evening peak from Dongsi to Songjiazhuang. As the morning peak is the time period when passengers need to get to their workplaces at specified times, while the evening peak is the time period of going off duty, so the travel time period chosen by the passengers is relatively wider. As shown in **[Fig 17](#pone.0188874.g017){ref-type=\"fig\"}** and **[Fig 18](#pone.0188874.g018){ref-type=\"fig\"}**, the section flow at transfer stations will vary greatly, because the section flow can gather from other lines, or disperse to other lines via the transfer stations. In addition, the density of the passenger flow space-time distribution in each section is less than 480, so it satisfies the capacity constraints.\n\n![The space-time section flow distribution from Songjiazhuang to Tiantongyuan North.](pone.0188874.g017){#pone.0188874.g017}\n\n![The space-time section flow distribution from Tiantongyuan North to Songjiazhuang.](pone.0188874.g018){#pone.0188874.g018}\n\n**[Fig 19](#pone.0188874.g019){ref-type=\"fig\"}** shows the variation of CPU time with the relative gap. We can see that the convergence speed is fast with a low relative gap, then as the relative gap gets lower, the CPU time becomes longer, and the convergence speed gets slower. When the relative gap reaches 10^\u22124^, the CPU time reaches 18.5h, for the reason that the algorithm of MSA takes a long time to reach a high relative gap.\n\n![The variation of CPU time with relative gap.](pone.0188874.g019){#pone.0188874.g019}\n\nConclusions {#sec019}\n===========\n\nIn this paper, the concept of continuous transmission is introduced to model the dynamic assignment for the urban rail network without the restriction of train schedules. Based on the cell transmission mechanism, the proposed CTM considers the priority principle, queuing process, capacity constraints and congestion effect. Using the MSA, the instantaneous dynamic optimal state can be reached at each interval. A fast and effective method is designed for solving the shortest path for the urban rail network. This method decreases the computing cost for solving the CTM, and it is applied efficiently to the large-scale urban rail network.\n\nThe CTM can generate some important evaluation indexes, including the time-varying section flow, the circulation volume, the detained flow on the platform, and the queuing length on the platform. It provides effective supports for optimizing the space-time resource allocation for the urban rail network. Finally, the model and its potential application are demonstrated via two numerical experiments using a small-scale network and the Beijing Metro network.\n\nThe proposed method is assumed that the passengers follow the instantaneous dynamic route choice principle, and as a topic of further interest, we can explore the CTM based on the continuous transmission with predictive/ideal dynamic user equilibrium. In the future studies, this method of DAURTN can be employed for optimizing the schedule, frequency and rolling stock, and evaluating the transportation capacity of network.\n\nSupporting information {#sec020}\n======================\n\n###### No. of each station.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### The distance between two adjacent stations.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### No. of each line.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### Line style.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### Total distance and average running speed of each line.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\n###### The stations of each line passing.\n\n(XLSX)\n\n###### \n\nClick here for additional data file.\n\nThis study was supported by National Natural Science Foundation of China (Grant No. 71171200, No. U1334207, No. 71701216), Project funded by China Postdoctoral Science Foundation (Grant No. 2017M612593) and postdoctoral foundation of Central South University.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "Background {#Sec1}\n==========\n\nEven though cell fusion has a pivotal role in several physiological and pathophysiological conditions such as fertilisation, placentation, muscle development, osteoclastogenesis, wound healing, tissue regeneration, infection with enveloped viruses, and cancer (for review see: \\[[@CR1]--[@CR4]\\]), the conditions that favour and the detailed mechanisms of how the plasma membranes of two or more cells merge are not fully understood.\n\nDuring evolution different cell fusion strategies developed using different proteins and protein-protein interactions. Some cell fusion-mediating proteins, such as EFF-1 in *C. elegans* and members of the syncytin family (syncytin-1 and syncytin-2 in humans or *syncytin-A* and *syncytin--B* in mice) are of viral and retroviral origin (for review see: \\[[@CR1], [@CR5], [@CR6]\\]). Because of the high homology of these proteins to viral fusion proteins, the cell fusion mechanisms might conceivably be similar. Syncytin-1 and syncytin-2 are expressed in trophoblasts and mediate their fusion to multi-nucleated syncytiotrophoblasts \\[[@CR5]\\]. Trophoblast fusion depends on cyclic AMP (cAMP)/PKA- and PKC-dependent mechanisms \\[[@CR7]\\]. In contrast, the fusion of myoblasts to multinucleated myofibres chiefly depends on the reorganisation of the actin cytoskeleton and the interplay of several actin regulatory proteins (for review see: \\[[@CR8]\\]). In order to fuse with osteoclasts, the macrophages first have to adopt a fusion-competent status, which is induced through several cytokines, including IL-4, IL-13, M-CSF and RANKL, and surface molecules such as TREM-2 and DAP12 (for review see: \\[[@CR9]\\]). Fusion-competent macrophages are characterised by the expression of several markers, including E-cadherin, DC-STAMP/OC-STAMP, CD200, SIRP\u03b1, CD9 and CD81, CCL2, and MMP9 (for review see: \\[[@CR9]\\]). The fusion of mesenchymal stem cells with MDA-MB-231 breast cancer cells depends on S100A4 \\[[@CR10]\\].\n\nThe induction of a fusion-competent status indicates that the cells per se are not fusogenic, but rather, have to acquire the ability to fuse, which has been termed priming \\[[@CR4]\\]. In addition to priming, four additional steps have been proposed for the entire cell fusion process, namely, chemotaxis, adhesion, fusion and post-fusion (an excellent and detailed overview is given in \\[[@CR4]\\]). Fusion-competent cells have to migrate towards their cellular fusion partners by chemotaxis and close cell-cell contact is a prerequisite for the subsequent fusion step. As important as the induction of a fusion-competent status is, fused cells must finally acquire a fusion-incompetent status to prevent further cell fusion events.\n\nEven though factors have been identified that induce a fusion-competent status in macrophages, the regulation of the entire process but particularly basal cell fusion events that occur between different body cell types, is not understood. In addition to macrophages \\[[@CR11]\\], basal cell fusion events were also observed for bone marrow-derived stem cells (BMDCs) \\[[@CR12]\\], intestinal cells \\[[@CR13]\\], and haematopoietic cells \\[[@CR14]\\]. It is unclear how these spontaneous cell fusion events are regulated. However, the data from several studies indicated an increased cell fusion frequency under injurious/inflammatory conditions. For instance, the fusion frequency of bone marrow-derived stem cells (BMDCs) with epithelial intestinal cells and Purkinje neurons is markedly increased in response to chronic inflammation \\[[@CR13], [@CR15]\\]. Likewise, an increased fusion frequency for different cell types, including macrophages and tumour cells, was observed in the presence of the pro-inflammatory cytokine TNF-\u03b1 \\[[@CR16]--[@CR19]\\]. In this context, it is worth speculating whether the chronically inflamed tumour microenvironment \\[[@CR20], [@CR21]\\] would also provide a fusion-friendly milieu. It is well-known that tumour cells could fuse with tumour cells and normal cells, like macrophages and stem cells, thereby giving rise to hybrid cells that could exhibit novel properties, such as an enhanced metastatic capacity and an increased drug resistance (for review see: \\[[@CR22]--[@CR24]\\]).\n\nTo clarify whether the fusion of tumour cells and normal cells is triggered by cytokines, chemokines and/or growth factors, we developed a *Cre-LoxP*-based cell fusion assay \\[[@CR18]\\]. Human breast cancer cells were stably transfected with a fluorescence double reporter (FDR) plasmid containing a *loxP* flanked HcRed/DsRed cassette followed by an EGFP cassette and were co-cultured with *Cre* recombinase-expressing human breast epithelial cells \\[[@CR18]\\]. The Cre-mediated recombination in fused cells led to a switch from red fluorescence to green fluorescence that could be easily quantified by flow cytometry \\[[@CR18]\\]. Using this assay we were able to show that the fusion of human breast cancer cells and human breast epithelial cells was positively triggered by TNF-\u03b1 in a dose-dependent manner \\[[@CR18]\\].\n\nHere, we investigated the impact of TNF-\u03b1-induced protein expression on cell fusion. MMP9 was markedly up-regulated in TNF-\u03b1 treated cells and the inhibition of MMP9 activity was correlated with a significantly impaired TNF-\u03b1-induced fusion rate. Similar results were achieved using the tetracycline-based antibiotic minocycline, which also has anti-inflammatory properties (for review see: \\[[@CR25]\\]).\n\nMethods {#Sec2}\n=======\n\nCell culture {#Sec3}\n------------\n\nM13SV1-Cre-Puro human breast epithelial cells and MDA-MB-435-pFDR1 human cancer cells were generated as previously described \\[[@CR18]\\]. The pFDR1 vector \\[[@CR26]\\] was kindly provided by Frank Edenhofer (University of W\u00fcrzburg, W\u00fcrzburg, Germany). M13SV1-mCherry-Cre human breast epithelial cells were derived from M13SV1 human breast epithelial cells \\[[@CR27]\\] by stable transduction with the pcDNA-mCherry-P2A-Cre vector. This vector was derived by excising the mCherry-P2A-Cre sequence of pLM-CMV-R-Cre (a gift from Michel Sadelain (Addgene plasmid \\#27546)) with NheI and SalI (both restriction enzymes were purchased from ThermoFisher Scientific GmbH, Schwerte, Germany) and cloning it into the pcDNA3.1 vector digested with NheI and XhoI (vector and restrictions enzymes were obtained from ThermoFisher Scientific GmbH, Schwerte, Germany). XhoI and SalI have identical cohesive ends. M13SV1-Cre-Puro and M13SV1-mCherry-Cre human breast epithelial cells (M13SV1-Cre cells) were cultivated in MSU-1 basal media (Biochrom GmbH, Berlin, Germany) supplemented with 10% foetal calf serum (FCS; Biochrom GmbH, Berlin, Germany), 1% penicillin/streptomycin (100\u00a0U/ml penicillin, 0.1\u00a0mg/ml streptomycin; Sigma-Aldrich, Taufkirchen, Germany), 10\u00a0\u03bcg/ml human recombinant EGF, 5\u00a0\u03bcg/ml human recombinant insulin, 0.5\u00a0\u03bcg/ml hydrocortisone, 4\u00a0\u03bcg/ml human transferrin, 10\u00a0nM \u03b2-oestrogen (all chemicals were purchased from Sigma-Aldrich, Taufkirchen, Germany) and 1\u00a0\u03bcg/ml puromycin (InvivoGen, Toulouse, France). MDA-MB-435-pFDR1 cells were cultivated in DMEM media (Sigma-Aldrich, Taufkirchen, Germany) supplemented with 10% FCS (Biochrom GmbH, Berlin, Germany), 1% penicillin/ streptomycin (100\u00a0U/ml penicillin, 0.1\u00a0mg/ml streptomycin; Sigma-Aldrich, Taufkirchen, Germany), and 2\u00a0\u03bcg/ml puromycin (InvivoGen, Toulouse, France). All cells were maintained in a humidified atmosphere at 37\u00a0\u00b0C and 5% CO~2~.\n\nCell fusion assay {#Sec4}\n-----------------\n\nQuantification of fusion events between M13SV1-Cre cells and MDA-MB-435-pFDR1 cells was performed as previously described \\[[@CR18]\\]. Briefly, M13SV1-Cre human breast epithelial cells and MDA-MB-435-pFDR1 human breast cancer cells were co-cultured in a ratio of 1:3 per well of a 96-well plate for 72\u00a0h in a humidified atmosphere at 37\u00a0\u00b0C and 5% CO~2~. Depending on the experiment 100\u00a0ng/ml TNF-\u03b1 (Bio-Techne GmbH, Wiesbaden-Nordenstadt, Germany), up to 10\u00a0\u03bcg/ml minocycline (Sigma-Aldrich, Taufkirchen, Germany), up to 10\u00a0\u03bcM SB-3CT (Sigma Aldrich, Taufkirchen, Germany), 10\u00a0\u03bcg/ml anti-ICAM1/CD54 (mouse monoclonal, clone 11C81; Bio-Techne GmbH, Wiesbaden-Nordenstadt, Germany) and appropriate combinations of inhibitors and TNF-\u03b1 were added to the cell culture media. Hypoxia was induced by culturing the cells in stepwise oxygen-deprivation conditions (24\u00a0h with 10% O~2~, 24\u00a0h with 5% O~2~, 24\u00a0h with 1% O~2~) in hypoxia chambers (Billups-Rothenberg, Del Mar, CA, USA) with a constant flow of 5% CO~2~ and the remainder of nitrogen, as previously described \\[[@CR18]\\]. All co-cultured cells were harvested, washed once with PBS and the amount of EGFP-expressing cells was quantified by flow cytometry (FACSCalibur; Becton Dickenson, Heidelberg, Germany). Freshly harvested M13SV1-Cre cells and MDA-MB-435-pFDR1 cells mixed in a ratio of 1:3 served as a negative control to adjust the flow cytometer. The relative fold change was calculated in relation to untreated co-cultured cells, which was set to 1. Each condition was assayed in triplicate.\n\nTotal RNA preparation {#Sec5}\n---------------------\n\nTotal RNA from all the cell lines/cells used in this study was isolated using the NucleoSpin\u00ae RNA Kit (Macherey Nagel, D\u00fcren, Germany) according to the manufacturer's instructions. For PCR and qPCR, RNA concentration and purity was determined by UV spectrophotometric analysis. For subsequent cDNA microarray studies, the RNA integrity number (RIN) was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany). The samples were diluted to a final concentration of 500\u00a0ng/\u03bcl and applied to an RNA chip according to the manufacturer's instructions. RNA samples with a RIN of 8 to 10 were used for microarray analysis.\n\nSingle-colour microarray analysis {#Sec6}\n---------------------------------\n\nTo investigate differential gene expression after the TNF-\u03b1 stimulation of MDA-MB-435-pFDR1 and M13SV1-Cre-Puro cells under normoxia and hypoxia, a single colour microarray analysis (Agilent Human 4\u00d744K v2 Microarray (Agilent Technologies, Waldbronn, Germany)) was performed. The total RNA of the cells (three independent experiments were pooled) matching the RIN criteria of 8 to 10 were sent on dry ice to Source BioSciences (Nottingham, UK), which performed synthesis and Cy3 labelling of the cDNA and hybridisation of the microarrays. Microarray gene expression data were analysed using GeneSpring GX v14.8 software (Agilent Technologies, Waldbronn, Germany). Expression data were normalized based on quantiles with a threshold of 1 and no baseline transformation was performed. For further analysis, data were filtered by flags (not detected flags and compromised spots were removed) and fold changes (2-fold up and down). All microarray data were deposited in the ArrayExpress database at EMBL-EBI (<https://www.ebi.ac.uk/arrayexpress>) under accession number E-MTAB-6084.\n\nRT-PCR and qPCR {#Sec7}\n---------------\n\nSynthesis of cDNA from mRNA by reverse transcription was performed using the RevertAid First Strand cDNA Synthesis Kit (ThermoFisher Scientific GmbH, Schwerte, Germany) as referred to in the instruction manual. Conventional PCR (total reaction volume of 25\u00a0\u03bcl) was performed with ready to use 5\u00d7 Master Mix containing Bio&Sell Taq-Polymerase, dNTPs and MgCl~2~ (Bio&Sell GmbH Nuremberg, Germany) and 10\u00a0\u03bcM primers (ThermoFisher Scientific GmbH, Schwerte, Germany). Cycling conditions comprised of an initial denaturation for 5\u00a0min at 94\u00a0\u00b0C and 30\u00a0cycles of 30\u00a0s at 94\u00a0\u00b0C, 30\u00a0s at the appropriate annealing temperature and 30\u00a0s at 72\u00a0\u00b0C followed by final elongation for 7\u00a0min at 72\u00a0\u00b0C. Primer pairs used in this study are summarised in Table\u00a0[1](#Tab1){ref-type=\"table\"}. PCR products were separated on a 1.5% agarose gel and the bands were visualized with GelRed\u2122 stain (VWR International GmbH, Darmstadt, Germany) and the GFelDoc\u2122 EZ Imager system (Bio-Rad, Munich, Germany). For qPCR (total volume of 10\u00a0\u03bcl per reaction), the SYBR Green Super Mix with ROX (Quanta Bioscience, Beverly, CA, USA) and 10\u00a0\u03bcM primers (Table [1](#Tab1){ref-type=\"table\"}) were used according to the manufacturer's instructions. The StepOne Plus Real-Time PCR System (ThermoFisher Scientific GmbH, Schwerte, Germany) was used for qPCR. The relative target gene expression level was determined in relation to GAPDH using the 2^-\u0394CT^ method.Table 1Summary of PCR/qPCR primer pairsNamePCR/ qPCRMean product sizePrimerSequence (5\u2032 to 3\u2032)MMP9qPCR106bpforwardTTCCAAACCTTTGAGGGCGAreverseCAAAGGCGTCGTCAATCACCICAM1qPCR149bpforwardGGTAGCAGCCGCAGTCATAAreverseGATAGGTTCAGGGAGGCGTGCCL2PCR297bpforwardGCTCAGCCAGATGCAATCAATGreverseGTGTCTGGGGAAAGCTAGGGCLDN1qPCR118bpforwardCTGTCATTGGGGGTGCGATAreverseCTGGCATTGACTGGGGTCATADAMTS9PCR301bpforwardTTAATCTCACCGCCAATGCreverseGCGCTGCGCCTATAAATGATVEGFCPCR320bpforwardCATGTACGAACCGCCAGreverseTTGGCTGTTTGGTCATTGGCIL4I1PCR451bpforwardTCACCAAGAGCTGGAGACACCreverseAACTTGGTCAGGTTGAGCCCIL7RPCR281bpforwardTAATAGCTCAGGGGAGATGGATreverseCTTGCAGAAAACCTTCCACTTCATNFAIP3PCR629bpforwardCAACTGAAACGGGGCAAAGCreverseGCCGTCACCGTTCGTTTTC\u03b2-actinPCR733bpforwardCCTCGCCTTTGCCGATCCreverseGGCCATCTCTTGCTCGAAGTGAPDHqPCR87bpforwardTGCACCACCAACTGCTTAGCreverseGGCATGGACTGTGGTCATGAG\n\nWestern blot {#Sec8}\n------------\n\nM13SV1-Cre cells and MDA-MB-435-pFDR1 cells were cultivated for 72\u00a0h at 37\u00a0\u00b0C and 5% CO~2~ under normoxic and hypoxic conditions in the presence of TNF-\u03b1 (100\u00a0ng/ml), minocycline (10\u00a0\u03bcg/ml) and a combination of both. Subsequently, cells were harvested and were lysed in ice-cold RIPA buffer (50\u00a0mM Tris-HCl pH\u00a08.0; 150\u00a0mM NaCl, 1% (*v*/v) NP-40, 0.5% (*w*/*v*) sodium deoxycholate, 0.1% (w/v) sodium dodecyl sulphate) supplemented with cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail (Sigma Aldrich, Taufkirchen, Germany) and Pierce Phosphatase Inhibitor Mini Tablets (ThermoFisher Scientific GmbH, Schwerte, Germany). The samples were sonicated three times (10\u00a0s on and 30\u00a0s off) and the total protein concentration was determined using the Pierce\u2122 BCA Protein Assay Kit (ThermoFisher Scientific GmbH, Schwerte, Germany) according to the manufacturer's instructions. For Western blots, 40\u00a0\u03bcg of total protein lysate was mixed with 3\u00d7 Laemmli Sample Buffer and was incubated for 6\u00a0min at 95\u00a0\u00b0C. The samples were separated by 10% or 15% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to an Immobilon polyvinyl difluoride (PVDF) nitrocellulose membrane (Merck Millipore, Darmstadt, Germany) under semi-dry conditions. The membranes were blocked in 5% (*w*/*v*) non-fat milk powder in PBS-T (phosphate-buffered saline) with 0.1% (*v*/v) Tween 20 (PBS-T) for 1\u00a0h at room temperature. The following antibodies were used for Western blot analysis: anti-MMP9 (rabbit monoclonal; Abcam, Cambridge, UK), anti-CCL2 (MCP-1; rabbit polyclonal; Abcam, Cambridge, UK), anti-claudin-1 (mouse monoclonal; Abcam, Cambridge, UK), anti-ICAM1/CD54 (rabbit polyclonal; Cell Signaling, Leiden, Netherlands), \u03b2-actin (rabbit monoclonal; Cell Signaling, Leiden, Netherlands); anti-mouse-IgG-HRP-linked (Cell Signaling, Leiden, Netherlands), and anti-rabbit-IgG-HRP-linked (Cell Signaling, Leiden, Netherlands). The bands were visualized using the Pierce ECL Western blot substrate (Thermo Fisher Scientific, Bonn, Germany), according to the manufacturer's instructions, and the Aequoria Macroscopic Imaging System (Hamamatsu Photonics Germany, Herrsching am Ammersee, Germany).\n\nZymography {#Sec9}\n----------\n\nThe proteolytic activity of MMP9 was measured using a gelatine zymography assay. M13SV1-Cre cells and MDA-MB-435-pFDR1 cells were plated at a density of 5\u00d710^5^ cells/well in a 6-well plate. After a 72\u00a0h incubation with 100\u00a0ng/ml TNF-\u03b1, 10\u00a0\u03bcg/ml minocycline, and a combination of both, the cell culture supernatants were collected and were mixed with non-reducing Laemmli sample buffer (250\u00a0mM Tris-HCl (pH\u00a06.8), 10% (*w*/*v*) SDS, 25% (*v*/v) glycerol, 0.01% (w/v) bromophenol blue) without boiling. The samples were separated on a 10% sodium dodecyl sulphate polyacrylamide gel containing 0.1% gelatine (Sigma Aldrich, Taufkirchen, Germany). After separation by electrophoresis, the gels were washed four times in wash buffer (50\u00a0mM Tris-HCl (pH\u00a07.5), 10\u00a0mM CaCl~2~, 2.5% (*v*/v) Triton X-100, 0.02% NaN~3~) for 2\u00a0h at room temperature to remove the SDS. Then, the gel was incubated in incubation buffer (50\u00a0mM Tris-HCl (pH\u00a07.5), 150\u00a0mM NaCl, 10\u00a0mM CaCl~2~, 0.02% NaN~3~) overnight at 37\u00a0\u00b0C. Subsequently, the gel was stained with Coomassie brilliant blue R-250 (Sigma Aldrich, Taufkirchen, Germany) for 1\u00a0h at room temperature with gentle agitation. Finally, the gels were destained until clear bands appeared visible, which were indicative of the proteolytic activity of proteases, including MMP9.\n\nXTT assay {#Sec10}\n---------\n\nTo determine the appropriate minocycline concentration for the quantification of cell fusions, M13SV1-Cre and MDA-MB-435-pFDR1 cells were cultured in triplicate at a ratio of 1:3 for up to 3 days in the presence of different concentrations of minocycline. Non-treated cells served as a control. After 24\u00a0h, 48\u00a0h, and 72\u00a0h the media was removed and the plates were analysed using the XTT reagent (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. Absorption of the XTT-formazan derivative that formed was measured using a BioTek EL800 microplate reader (BioTek, Bad Friedrichshall, Germany).\n\nsiRNA experiments {#Sec11}\n-----------------\n\nExpression of claudin-1 was knocked-down using a specific claudin-1 siRNA (Santa Cruz Biotechnology, Heidelberg, Germany). For control purposes, a non-targeting negative control siRNA was used (QIAGEN GmbH, Hilden, Germany). M13SV1-Cre cells (2.5\u00d710^5^) were resuspended in Opti-Mem (ThermoFisher Scientific, Schwerte, Germany) and were subsequently transfected with claudin-1 siRNA or control siRNA to a final concentration of 100\u00a0nM by lipofection (Lipofectamine\u2122 2000; ThermoFisher Scientific, Schwerte, Germany) as described in the user's manual. The cells were seeded in serum-free MSU medium in 6-well plates. After 6\u00a0h, the serum-free media was replaced with complete MSU medium containing FCS, antibiotics and additional supplements. The knock-down of claudin-1 expression in M13SV1-Cre cells was confirmed by Western blot.\n\nCo-immunoprecipitation {#Sec12}\n----------------------\n\nM13SV1-Cre cells and MDA-MB-435-pFDR1 cells were stimulated with 100\u00a0ng/ml TNF-\u03b1, 10\u00a0\u03bcg/ml minocycline, or a combination of both, for 72\u00a0h at 37\u00a0\u00b0C and 5% CO~2~ in a humidified atmosphere. Subsequently, the cells were washed once with PBS and lysed in ice-cold lysis buffer (1% (*v*/*v*) NP-40, 50\u00a0mM Tris-HCl (pH\u00a07.5), 150\u00a0mM NaCl, 1\u00a0mM EDTA, 1\u00a0mM EGTA and proteinase inhibitors (see above)) for 30\u00a0min on ice. The samples were sonicated three times (5\u00a0s on/20\u00a0s off) on a low pulse frequency and the total protein concentration was determined using the Pierce\u2122 BCA Protein Assay Kit (ThermoFisher Scientific GmbH, Schwerte, Germany) according to the manufacturer's instructions. The lysates were cleared by centrifugation (10\u00a0min, 12,000\u00a0rpm) and the supernatants were transferred to a new tube. Prior to immunoprecipitation, the lysates were pre-treated with 25\u00a0\u03bcl of Protein A Magnetic Beads (Cell Signaling, Leiden, Netherlands) for 2\u00a0h at 4\u00a0\u00b0C to remove any proteins that might bind non-specifically to Protein A. The samples were incubated with anti-ICAM1/CD54 (5\u00a0\u03bcg/ml; mouse monoclonal, clone 11C81; Bio-Techne GmbH, Wiesbaden-Nordenstadt, Germany) or IgG1 (5\u00a0\u03bcg/ml; Beckman Coulter, Krefeld, Germany) overnight at 4\u00a0\u00b0C. Thereafter, 30\u00a0\u03bcl of Protein A Magnetic Beads were added and the samples were incubated for an additional 2\u00a0h at 4\u00a0\u00b0C. The precipitates were washed three times in lysis buffer, separated on a magnetic rack and resuspended in 3\u00d7 Laemmli Sample Buffer (with DTT, without \u03b2-mercaptoethanol). ICAM1/CD54 and the co-immunoprecipitated proteins were detected by Western blot.\n\nSoluble ICAM1 detection {#Sec13}\n-----------------------\n\nTo detect soluble ICAM1, the cell culture supernatants from TNF-\u03b1-treated, minocycline-treated, and TNF-\u03b1 and minocycline-treated M13SV1-Cre cells and MDA-MB-435-pFDR1 cells were collected and concentrated using Vivaspin-2 centrifugal concentrators (100,000 MWCO; Sartorius, G\u00f6ttingen, Germany). The total protein concentration of the concentrated supernatants was determined using the Pierce\u2122 BCA Protein Assay Kit (ThermoFisher Scientific GmbH, Schwerte, Germany) according to the manufacturer's instructions. An equal amount of supernatant from each sample was resuspended in Laemmli sample buffer. Soluble ICAM1 was detected by Western blot using an anti-ICAM1/CD54 antibody (clone 14C11, mouse monoclonal; ThermoFisher Scientific GmbH, Schwerte, Germany).\n\nStatistical analyses {#Sec14}\n--------------------\n\nThe statistical significance of the data presented in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"} was calculated using an unpaired, two-tailed Student's *t*-test. The mean differences of the data presented in Figs.\u00a0[4](#Fig4){ref-type=\"fig\"}, [5](#Fig5){ref-type=\"fig\"}, [6](#Fig6){ref-type=\"fig\"} and [7](#Fig7){ref-type=\"fig\"} were tested by ANOVA F-tests. Afterwards, multiple comparisons were performed using Scheff\u00e9 post-hoc tests. Statistical analyses were performed using SPSS Version 23.0.0.2 and *p*-values \\<\u20090.05 were considered significant.\n\nLanguage editing {#Sec15}\n----------------\n\nThe manuscript was language edited by American Journal Experts (Durham, NC, USA).\n\nResults {#Sec16}\n=======\n\nAnalysis of the gene expression pattern of M13SV1-Cre cells and MDA-MB-435-pFDR1 cells with TNF-\u03b1 treatment in normoxic and hypoxic conditions {#Sec17}\n----------------------------------------------------------------------------------------------------------------------------------------------\n\nWe recently demonstrated that the fusion of human M13SV1-Cre breast epithelial cells and human MDA-MB-435-pFDR1 breast cancer cells under normoxic and hypoxic conditions is positively triggered by TNF-\u03b1 \\[[@CR18]\\]. To analyse the changes in the gene expression profile of both cell lines induced by TNF-\u03b1 (100\u00a0ng/ml) under normoxic and hypoxic conditions, cDNA-microarray studies were performed. In total, 51 genes were found to be significantly up-regulated in TNF-\u03b1-treated cell lines under both normoxia and hypoxia (Fig.\u00a0[1a, b](#Fig1){ref-type=\"fig\"}; Table\u00a0[2](#Tab2){ref-type=\"table\"}), including adhesion molecules such as ICAM1, ICAM2, and ICAM4, components of the extracellular matrix, including COL27A1, LAMB3, LAMC2, and MUC4, chemokines, growth factors, and interleukins, such as CCL2, VEGFC and IL32, and proteases including ADAMTS9 and MMP9 (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}). In particular, CCL2, MMP9 and ICAM1 have already been associated with cell fusion \\[[@CR9], [@CR28]\\]. In contrast, 15 genes were down-regulated in TNF-\u03b1-treated cells under normoxic and hypoxic conditions, such as ABCC6 and FOS (Fig. [1c, d](#Fig1){ref-type=\"fig\"}; Table\u00a0[3](#Tab3){ref-type=\"table\"}).Fig. 1Differentially regulated genes in TNF-\u03b1-treated M13SV1-Cre and MDA-MB-435-pFDR1 cells under normoxic and hypoxic conditions. The up-regulated and down-regulated genes in TNF-\u03b1-treated cells compared with untreated cells that matched the filter criteria (flags: not detected flags and compromised spots were removed) and fold changes (2-fold up and down). **a** A heat map analysis of TNF-\u03b1 up-regulated genes, **b** a Venn diagram of TNF-\u03b1 up-regulated genes, **c** a heat map analysis of TNF-\u03b1 down-regulated genes, **d** a Venn diagram of TNF-\u03b1 down-regulated genesTable 2Up-regulated genes (FC\u2009\u2265\u20092) in TNF-\u03b1 treated cells under normoxic and hypoxic conditionsNormoxiaHypoxiaNormoxiaHypoxiaGenbank\\\nAccessionGeneSymbolGeneNameM13SV1-Cre\u2009+\u2009TNF-\u03b1 vs. M13SV1-CreM13SV1-Cre\u2009+\u2009TNF-\u03b1 vs. M13SV1-CreMDA-MB-435-pFDR1\u2009+\u2009TNF-\u03b1 vs. MDA-MB-435-pFDR1MDA-MB-435-pFDR1\u2009+\u2009TNF-\u03b1 vs. MDA-MB-435-pFDR1NM_182920ADAMTS9ADAM metallopeptidase with thrombospondin type 1 motif, 915.3425.453.0510.59NM_058172ANTXR2anthrax toxin receptor 22.644.433.615.63NM_001165BIRC3baculoviral IAP repeat containing 34.5513.1123.7497.65NM_001734C1Scomplement component 1, s subcomponent5.666.976.917.03NM_000064C3complement component 34.314.24182.13115.58NM_002982CCL2chemokine (C-C motif) ligand 225.3639.20179.7362.92NM_001252CD70CD70 molecule3.156.994.764.52NM_001710CFBcomplement factor B2.412.2811.215.74NM_032888COL27A1collagen, type XXVII, alpha 12.704.452.095.29NM_004405DLX2distal-less homeobox\u00a022.397.505.384.42NM_014888FAM3Cfamily with sequence similarity 3. member C2.082.503.274.06NM_001024071GCH1GTP cyclohydrolase 12.203.072.933.83NM_001300919HMGA2high mobility group AT-hook 22.262.353.273.66NM_000201ICAM1intercellular adhesion molecule 14.466.915.763.58NM_000873ICAM2intercellular adhesion molecule 22.534.044.7146.87NM_022377ICAM4intercellular adhesion molecule 4 (Landsteiner-Wiener blood group)4.4715.537.1511.04NM_003897IER3immediate early response 32.022.846.612.96NM_022168IFIH1interferon induced with helicase C domain 12.433.472.012.48NM_004843IL27RAinterleukin 27 receptor. Alpha2.652.313.472.56NM_001012633IL32interleukin 324.474.2345.0561.04NM_001012631IL32interleukin 323.535.65174.41103.38NM_152899IL4I1interleukin 4 induced 113.7646.6350.4474.89NM_002185IL7Rinterleukin 7 receptor2.813.6523.528.37NM_001570IRAK2interleukin-1 receptor-associated kinase 25.2813.6614.7614.90NM_004031IRF7interferon regulatory factor 72.302.314.202.17NM_138433KLHDC7Bkelch domain containing 7B5.7410.452.7713.77NM_001017402LAMB3laminin, beta 32.123.485.682.89NM_005562LAMC2laminin, gamma 22.734.652.054.42NM_015541LRIG1leucine-rich repeats and immunoglobulin-like domains 12.292.494.452.27NM_004994MMP9matrix metallopeptidase 9 (gelatinase B, 92\u00a0kDa gelatinase, 92\u00a0kDa type IV collagenase)11.4527.2513.4944.92NM_018406MUC4mucin 4, cell surface associated2.494.722.342.76NM_181782NCOA7nuclear receptor coactivator 72.082.892.143.31NM_004289NFE2L3nuclear factor. Erythroid 2-like 33.087.275.008.99NM_020529NFKBIAnuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha2.977.635.415.74NM_004556NFKBIEnuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, epsilon3.864.403.413.14NM_002526NT5E5\u2032-nucleotidase, ecto (CD73)4.379.647.123.26NM_002526NT5E5\u2032-nucleotidase, ecto (CD73)3.378.283.622.54NM_030952NUAK2NUAK family, SNF1-like kinase. 22.245.485.712.87NM_001005377PLAURplasminogen activator, urokinase receptor4.002.1214.942.74NM_002852PTX3pentraxin 3, long4.376.1016.5917.75NM_006509RELBv-rel avian reticuloendotheliosis viral oncogene homolog B5.028.994.614.41NM_002960S100A3S100 calcium binding protein A32.012.884.907.84NM_001002236SERPINA1serpin peptidase inhibitor, clade A (alpha-1 antiproteinase. antitrypsin), member 12.502.735.442.35NM_015482SLC22A23solute carrier family 22, member 232.082.233.217.09NM_001024465SOD2superoxide dismutase 2, mitochondrial2.485.2016.5219.28NM_014220TM4SF1transmembrane 4\u00a0L six family member 12.083.235.343.43NM_017870TMEM132Atransmembrane protein 132A2.633.353.743.53NM_006290TNFAIP3tumor necrosis factor, alpha-induced protein 314.8116.9419.459.77NM_145725TRAF3TNF receptor-associated factor 32.213.612.022.50NM_006398UBDubiquitin D19.5958.56100.44226.21NM_005429VEGFCvascular endothelial growth factor C2.764.1674.9133.90Table 3Down-regulated genes (FC\u2009\u2264\u2009\u22122) in TNF-\u03b1 treated cells under normoxic and hypoxic conditionsNormoxiaHypoxiaNormoxiaHypoxiaGenbank\\\nAccessionGeneSymbolGeneNameM13SV1-Cre\u2009+\u2009TNF-\u03b1 vs. M13SV1-CreM13SV1-Cre\u2009+\u2009TNF-\u03b1 vs. M13SV1-CreMDA-MB-435-pFDR1\u2009+\u2009TNF-\u03b1 vs. MDA-MB-435-pFDR1MDA-MB-435-pFDR1\u2009+\u2009TNF-\u03b1 vs. MDA-MB-435-pFDR1NM_001079528ABCC6ATP-binding cassette, sub-family C (CFTR/MRP), member 6\u22122.72\u22122.26\u22122.51\u22122.36NM_020764CASKIN1CASK interacting protein 1\u22122.03\u22122.26\u22122.52\u22127.21NR_024482CECR5-AS1CECR5 antisense RNA 1\u22122.10\u2212\u20092.94\u22123.66\u22126.97NM_001885CRYABcrystallin, alpha B\u22122.76\u22124.62\u22122.03\u22126.23NM_001352DBPD site of albumin promoter (albumin D-box) binding protein\u22122.36\u22124.44\u22122.20\u22124.92NM_005252FOSFBJ murine osteosarcoma viral oncogene homolog\u22122.45\u22122.72\u22123.24\u2212\u20092.60NM_001136007FXYD3FXYD domain containing ion transport regulator 3\u22122.32\u22122.06\u221219.31\u221214.74NM_001136008FXYD3FXYD domain containing ion transport regulator 3\u22122.44\u22124.64\u221210.82\u221219.67XM_006715716HOXA6homeobox A6\u22122.75\u22124.61\u22122.38\u2212\u20092.30NM_030915LBHlimb bud and heart development\u22122.03\u221231.03\u22122.93\u22123.35NR_024485LOC100130093uncharacterized LOC100130093\u22122.28\u22123.05\u22123.11\u2212\u20093.30NM_014033METTL7Amethyltransferase like 7A\u22124.74\u22124.34\u22122.93\u22126.66BC128044MYO15Bmyosin XVB pseudogene\u22122.42\u22122.81\u2212\u20092.15\u22122.70NM_030625TET1tet methylcytosine dioxygenase 1\u22122.90\u22123.02\u22122.01\u2212\u20092.70NM_003881WISP2WNT1 inducible signaling pathway protein 2\u22123.96\u22124.69\u22123.99\u22125.53\n\nValidation of microarray data by qPCR and conventional PCR {#Sec18}\n----------------------------------------------------------\n\nTo validate the microarray data, Claudin-1 (CLDN1), ICAM1, and MMP9 expression was analysed by qPCR and ADAMTS9, CCL2, IL4I1, IL7R, TNFAIP3 and VEGFC expression was analysed by conventional PCR. CLDN1 did not pass the filter criteria but the microarray data revealed a marked up-regulation of this protein in TNF-\u03b1-treated (100\u00a0ng/ml) MDA-MB-435-pFDR1 cells under normoxic and hypoxic conditions (Additional file [1](#MOESM1){ref-type=\"media\"}). The qPCR and conventional PCR data partially matched the microarray data (Fig. [2a, d](#Fig2){ref-type=\"fig\"}). The microarray data revealed an up-regulation of ICAM1 expression in TNF-\u03b1-stimulated (100\u00a0ng/ml) M13SV1-Cre cells under normoxic conditions (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}), which was opposite to the qPCR data showing similar ICAM1 mRNA levels in untreated and TNF-\u03b1-treated M13SV1-Cre cells under normoxic conditions (Fig. [2b](#Fig2){ref-type=\"fig\"}). In contrast, ICAM1 was up-regulated in TNF-\u03b1-stimulated M13SV1-Cre cells under hypoxia (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}), which could be validated by qPCR (Fig. [2b](#Fig2){ref-type=\"fig\"}). The TNF-\u03b1-induced increase in CLDN1 and MMP9 expression in M13SV1-Cre cells under normoxia and hypoxia (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}) were validated by qPCR, whereas MMP9 mRNA expression was markedly up-regulated in TNF-\u03b1-stimulated cells (Fig. [2c](#Fig2){ref-type=\"fig\"}). For MDA-MB-435-pFDR1 cells, only the microarray data for CLDN1 and ICAM1 could be validated by qPCR; the qPCR data for MMP9 did not match the microarray data (Figs. [1](#Fig1){ref-type=\"fig\"}, [2c](#Fig2){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}). Markedly increased MMP9 expression levels were determined for TNF-\u03b1-stimulated MDA-MB-435-pFDR1 cells (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}), which is opposite to qPCR rather showing comparable MMP9 mRNA expression levels in untreated and TNF-\u03b1 treated cells (Fig. [2c](#Fig2){ref-type=\"fig\"}).Fig. 2Validation of microarray data by qPCR and conventional PCR. **a** qPCR CLDN1, **b** qPCR ICAM1, **c** qPCR MMP9, **d** conventional PCR. MMP9 was significantly up-regulated in TNF-\u03b1-treated (100\u00a0ng/ml) M13SV1-Cre cells, whereas significantly elevated CLDN1 and ICAM1 expression levels were detected in TNF-\u03b1-treated (100\u00a0ng/ml) MDA-MB-435-pFDR1 cells. Likewise, increased CCL2 and VEGFC expression levels were detected in TNF-\u03b1-treated (100\u00a0ng/ml) MDA-MB-435-pFDR1 cells. Data are presented as the mean \u00b1 SD of at least three independent experiments (**a**) or the representative data of three independent experiments (**b**). Statistical analysis: unpaired, two-tailed Student's *t*-test: \\*\u2009=\u2009*p*\u2009\\<\u20090.05; \\*\\*\u2009=\u2009*p*\u2009\\<\u20090.01\n\nCCL2, IL7R1, and VEGFC expression levels were partially up-regulated in TNF-\u03b1-stimulated MDA-MB-435-pFDR1 cells, which agreed with the microarray data (Fig. [1](#Fig1){ref-type=\"fig\"}, [2b](#Fig2){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}). In contrast, the microarray data revealed an up-regulation of IL4I1 in TNF-\u03b1-treated MDA-MB-435-pFDR1 cells, which could not be validated by conventional PCR (Fig. [2d](#Fig2){ref-type=\"fig\"}). Instead, only a weak PCR product was observed in TNF-\u03b1-stimulated MDA-MB-435-pFDR1 cells under normoxic conditions, whereas under hypoxia no PCR product was detected (Fig. [2d](#Fig2){ref-type=\"fig\"}). TNF-\u03b1 induced an increase in IL4I1, IL7R and TNFAIP3 expression in M13SV1-Cre cells under normoxic conditions (Fig. [2d](#Fig2){ref-type=\"fig\"}) and in hypoxic conditions there was only a slight increase in VEGFC levels in response to TNF-\u03b1 treatment (Fig. [2d](#Fig2){ref-type=\"fig\"}). Hence, the M13SV1-Cre microarray data were only partially validated by conventional PCR (Figs.\u00a0[1](#Fig1){ref-type=\"fig\"}, [2d](#Fig2){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}).\n\nValidation of the microarray data by western blot {#Sec19}\n-------------------------------------------------\n\nTo further validate the microarray and PCR data, Western blot analyses were conducted for CCL2, CLDN1, ICAM1 and MMP9. TNF-\u03b1-induced (100\u00a0ng/ml) CCL2 expression was observed in MDA-MB-435-pFDR1 cells with higher CCL2 expression under normoxia compared with hypoxia (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). In contrast, no CCL2 expression was found in M13SV1-Cre human breast epithelial cells (Fig. [3a](#Fig3){ref-type=\"fig\"}), which agrees with conventional PCR data (Fig. [2b](#Fig2){ref-type=\"fig\"}). The expression of CLDN1 was only observed in M13SV1-Cre cells concomitant with slightly increased CLDN1 levels in the presence of TNF-\u03b1 (Fig. [3a](#Fig3){ref-type=\"fig\"}), which is similar to qPCR data (Fig. [2](#Fig2){ref-type=\"fig\"}). Interestingly, no CLDN1 protein expression was detected in MDA-MB-435-pFDR1 cells (Fig. [3a](#Fig3){ref-type=\"fig\"}) even though qPCR data revealed increased CLDN1 mRNA levels in TNF-\u03b1-treated cells (Fig. [2b](#Fig2){ref-type=\"fig\"}). In agreement with the microarray data (Fig. [1](#Fig1){ref-type=\"fig\"}; Table [2](#Tab2){ref-type=\"table\"}), but opposite to the qPCR data (Fig. [2a](#Fig2){ref-type=\"fig\"}), increased ICAM1 protein expression was identified in TNF-\u03b1-treated M13SV1-Cre cells (Fig. [3a](#Fig3){ref-type=\"fig\"}). Increased TNF-\u03b1-induced MMP9 levels were only observed in M13SV1-Cre cells (Fig. [3a](#Fig3){ref-type=\"fig\"}), which agreed with the qPCR data (Fig. [2c](#Fig2){ref-type=\"fig\"}). In addition to Western blot analysis, MMP9 expression was further analysed by zymography, which also showed increased MMP9 expression in TNF-\u03b1-treated M13SV1-Cre cells (Fig. [3b](#Fig3){ref-type=\"fig\"}). Zymography analysis revealed a slight up-regulation of MMP9 expression in TNF-\u03b1-stimulated MDA-MB-435-pFDR1 cells (Fig. [3b](#Fig3){ref-type=\"fig\"}), which confirmed the qPCR data showing a weak MMP9 PCR product in TNF-\u03b1-treated cells under normoxic conditions (Fig. [2c](#Fig2){ref-type=\"fig\"}).Fig. 3Validation of qPCR and conventional RT-PCR data by Western blot and zymography. **a** Western blot analysis: CCL2 protein expression was only detected in TNF-\u03b1-treated (100\u00a0ng/ml) MDA-MB-435-pFDR1 cells, whereas CLDN1 expression was absent from these cells but was clearly detectable in M13SV1-Cre cells. Increased ICAM1 expression was observed in TNF-\u03b1-treated cells, whereas increased MMP9 expression was only found in TNF-\u03b1-stimulated M13SV1-Cre cells. Two discrete bands were detected for MMP9 in Western blot analysis; the upper band represents inactive MMP9 and the lower band represents active MMP9. ICAM1 Western blot data were rearranged because ICAM1 samples were originally loaded in a different order. **b** Zymography: markedly increased MMP9 expression levels and activity were detected in TNF-\u03b1-treated M13SV1-Cre cells, whereas even in TNF-\u03b1-stimulated MDA-MB-435 cells slightly enhanced MMP9 expression levels could be identified. The data shown are representative of at least three independent experiments\n\nFusion of M13SV1-Cre cells and MDA-MB-435-pFDR1 cells is impaired by the inhibition of MMP9 {#Sec20}\n-------------------------------------------------------------------------------------------\n\nTo investigate whether CLDN1, ICAM1 and MMP9 might be involved in the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells, the expression of these proteins was blocked. The knockdown of CLDN1 with siRNA in M13SV1-Cre cells did not impair the TNF-\u03b1-induced fusion of the cells (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}) even though the siRNA-mediated down-regulation of CLDN1 expression in M13SV1-Cre was stable for at least 3 days (Fig. [4b](#Fig4){ref-type=\"fig\"}). Interestingly, compared with M13SV1-Cre cells that were transfected with scrambled siRNA, M13SV1-Cre cells transfected with CLDN1 siRNA had a slightly increased fusion rate (Fig. [4a](#Fig4){ref-type=\"fig\"}).Fig. 4TNF-\u03b1-induced cell fusion could be blocked by the inhibition of MMP9. **a** The knockdown of CLDN1 expression with specific siRNA was not correlated with a decreased TNF-\u03b1-induced (100\u00a0ng/ml) fusion rate. **b** siRNA-mediated CLDN1 expression levels of M13SV1-Cre cells were stably down-regulated over 72\u00a0h. **c** Inhibition of ICAM1 activity using a blocking antibody did not impair the TNF-\u03b1-induced fusion between M13SV1-Cre and MDA-MB-435-pFDR1 cells. **d** The TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells was impaired by the MMP9 inhibitor SB-3CT in a dose-dependent manner. The mean \u00b1 SD of at least three independent experiments is shown. Statistical analysis: ANOVA F-test and Scheff\u00e9 post-hoc test: \\*\u2009=\u00a0*p*\u00a0\\<\u20090.05; \\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.01; \\*\\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.001\n\nThe inhibition of ICAM1 function with a blocking ICAM1 antibody was not correlated with a decreased TNF-\u03b1-induced rate of the fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells (Fig. [4c](#Fig4){ref-type=\"fig\"}). In contrast, the inhibition of MMP9 function using the specific inhibitor SB-3CT was associated with a dose-dependent, impaired, TNF-\u03b1-induced cell fusion rate of M13SV1-Cre and MDA-MB-435-pFDR1 cells (Fig. [4d](#Fig4){ref-type=\"fig\"}).\n\nThe TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells is inhibited by minocycline {#Sec21}\n---------------------------------------------------------------------------------------------\n\nThe tetracycline based antibiotic minocycline was tested because of its capability to down-regulate MMP9 expression in cells \\[[@CR29], [@CR30]\\]. First, XTT proliferation studies were performed to determine the optimal minocycline concentration and a range between 5\u00a0\u03bcg/ml and 24\u00a0\u03bcg/ml was tested. Prolonged cultivation of M13SV1-Cre and MDA-MB-435-pFDR1 cells with increasing concentrations of minocycline was correlated to a decreased cell proliferation rate (Additional file [2](#MOESM2){ref-type=\"media\"}). Hence, for on-going studies, a minocycline concentration of 10\u00a0\u03bcg/ml was chosen and only the impact of minocycline under normoxic conditions was analysed. The cultivation of cells under hypoxia and in the presence of 10\u00a0\u03bcg/ml minocycline was associated with numbers of dead cells that were too high to perform further cell fusion quantification studies.\n\nThe incubation of cells with 5\u00a0\u03bcg/ml or 10\u00a0\u03bcg/ml minocycline was correlated to a dose-dependent decreased TNF-\u03b1-induced fusion rate (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}), whereas the spontaneous fusion rate of the cells remained unaffected in the presence of different minocycline concentrations (Fig. [5](#Fig5){ref-type=\"fig\"}).Fig. 5TNF-\u03b1-induced fusion was inhibited by the tetracycline-based antibiotic minocycline. TNF-\u03b1-induced (100\u00a0ng/ml) cell fusion was impaired by minocycline in a dose-dependent manner. The data shown are the mean \u00b1 SD of three independent experiments. Statistical analysis: ANOVA F-test and Scheff\u00e9 post-hoc test: \\*\u2009=\u2009*p*\u2009\\<\u20090.05; \\*\\*\u2009=\u2009*p*\u2009\\<\u20090.01; \\*\\*\\*\u2009=\u2009*p*\u2009\\<\u20090.001\n\nEffect of minocycline on CLDN1, ICAM1, and MMP9 expression {#Sec22}\n----------------------------------------------------------\n\nQuantitative PCR and Western blot data showed that CLDN1 and ICAM1 expression in TNF-\u03b1-treated and TNF-\u03b1 and minocycline-treated M13SV1-Cre cells were comparable (Fig.\u00a0[6a, b](#Fig6){ref-type=\"fig\"}). In contrast, the qPCR results showed significantly reduced MMP9 expression in M13SV1-Cre cells co-treated with TNF-\u03b1 and minocycline as compared with TNF-\u03b1-treated cells (Fig. [6c](#Fig6){ref-type=\"fig\"}), which was further validated by zymography (Fig. [6c](#Fig6){ref-type=\"fig\"}). MMP9 protein expression was slightly increased in TNF-\u03b1 + minocycline co-treated M13SV1-Cre cells compared with TNF-\u03b1-stimulated cells (Fig. [6c](#Fig6){ref-type=\"fig\"}). This (reproducible) result remains ambiguous as did the finding that significantly higher MMP9 mRNA levels were observed in minocycline treated M13SV1-Cre cells (Fig. [6c](#Fig6){ref-type=\"fig\"}) although increased MMP9 protein expression levels were not detected (Fig. [6c](#Fig6){ref-type=\"fig\"}).Fig. 6TNF-\u03b1-induced MMP9 expression in M13SV1-Cre cells was blocked by minocycline. M13SV1-Cre and MDA-MB-435-pFDR1 cells were treated for 3 days with 100\u00a0ng/ml TNF-\u03b1, 10\u00a0\u03bcg/ml minocycline and a combination of both. The expression levels of **a**) CLDN1, **b**) ICAM1 and **c**) MMP9 were determined by qPCR, Western blot analysis and zymography (only MMP9). Significantly decreased MMP9 mRNA levels were detected in TNF-\u03b1 and minocycline-treated M13SV1-Cre cells as compared with TNF-\u03b1-treated cells and were validated by zymography. **d**) Co-immunoprecipitation data show that MMP9 binds to ICAM1. ICAM1 expression and co-immunoprecipitated MMP9 were markedly decreased in the lysates of TNF-\u03b1 + minocycline-treated M13SV1-Cre cells compared with the lysates from TNF-\u03b1-stimulated cells. **e**) No differences in sICAM1 levels were detected in M13SV1-Cre and MDA-MB-435-pFDR1 cells that were treated with TNF-\u03b1, minocycline and a combination of both. Shown are the mean \u00b1 SD or representative Western blot data for at least three independent experiments. Statistical analysis: ANOVA F-test and Scheff\u00e9 post-hoc test: \\*\u2009=\u00a0*p*\u00a0\\<\u20090.05; \\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.01; \\*\\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.001\n\nIn contrast to M13SV1-Cre breast epithelial cells, neither CLDN1 nor MMP9 protein expression was detected in MDA-MB-435-pFDR1 cells (Fig. [6a, c](#Fig6){ref-type=\"fig\"}). Because very weak MMP9 bands were observed in MDA-MB-435-pFDR1 cells by zymography (Fig. [6c](#Fig6){ref-type=\"fig\"}), and MMP9 and CLDN1 were detected by qPCR (Fig. [6a, c](#Fig6){ref-type=\"fig\"}) we assume that MMP9 and CLDN1 protein expression levels were below the detection threshold of the antibodies that were used. Minocycline treatment increased MMP9 mRNA expression in both M13SV1-Cre cells and MDA-MB-435-pFDR1 cells (Fig. [6c](#Fig6){ref-type=\"fig\"}), which, however, did not get translated into increased expression of the subsequent proteins (Fig. [6c](#Fig6){ref-type=\"fig\"}).\u00a0Compared with M13SV1-Cre breast epithelial cells, increased ICAM1 expression levels were found in TNF-\u03b1- and TNF-\u03b1 + minocycline-treated MDA-MB-435-pFDR1 cells (Fig. [6b](#Fig6){ref-type=\"fig\"}). Interestingly, significantly higher ICAM1 mRNA and protein levels were detected in the TNF-\u03b1 + minocycline-treated MDA-MB-435-pFDR1 cells than in the TNF-\u03b1-stimulated cells (Fig. [6b](#Fig6){ref-type=\"fig\"}).\n\nCo-immunoprecipitation assays were performed to investigate whether MMP9 bound to ICAM1 and whether this would be correlated to increased soluble ICAM1 (sICAM1) levels because of MMP9-dependent proteolytic degradation \\[[@CR31]\\]. As shown in Fig. [6d](#Fig6){ref-type=\"fig\"}, MMP9 could be co-immunoprecipitated with ICAM1 in M13SV1-Cre cells but not in MDA-MB-435-pFDR1 cells. Higher ICAM1 and MMP9 levels were co-immunoprecipitated in TNF-\u03b1-stimulated M13SV1-Cre cells, whereas lower levels were detected in cells co-treated with TNF-\u03b1 and minocycline (Fig. [6d](#Fig6){ref-type=\"fig\"}),\u00a0which is in agreement with the Western blot data (Fig. [6c](#Fig6){ref-type=\"fig\"}). In contrast, because of rather low MMP9 expression levels in MDA-MB-435-pFDR1 cells, MMP9 was not co-immunoprecipitated. To prove, whether the bindings of ICAM1 were proteolytically degraded by MMP9, the supernatants from cell culture were collected. Western blot analysis revealed no differences in the relative amount of sICAM1 monomers and dimers between the control cells and those treated with TNF-\u03b1 or TNF-\u03b1 + minocycline.\n\nBoth minocycline and the MMP9 inhibitor SB-3CT impaired TNF-\u03b1-mediated cell fusion {#Sec23}\n----------------------------------------------------------------------------------\n\nThe fusion rate of cells co-treated with both minocycline and SB-3CT (and TNF-\u03b1) was comparable to that of cells treated with minocycline and minocycline + TNF-\u03b1, respectively, and no additive effect was observed (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}). Interestingly, the cultivation of cells in the presence of both minocycline and the blocking ICAM1 antibody yielded in\u00a0an increased cell fusion rate, which was more intense in the presence of TNF-\u03b1 (Fig. [7](#Fig7){ref-type=\"fig\"}). Even though the data were not significant, the findings were reproducible in independent experiments, indicating that the increased fusion rate of cells in the presence of minocycline and anti-ICAM1 (and TNF-\u03b1) is a true effect.Fig. 7The co-administration of minocycline, SB-3CT and the ICAM1 blocking antibody, yielded different results. No additive inhibitory effect of minocycline and SB-3CT regarding the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells was observed. In contrast, the incubation of cells in the presence of minocycline, ICAM1 blocking antibody and TNF-\u03b1 yielded a significantly increased fusion rate. The data shown are the mean \u00b1 SD of at least three independent experiments. Statistical analysis: ANOVA F-test and Scheff\u00e9 post-hoc test: \\*\u2009=\u00a0*p*\u00a0\\<\u20090.05; \\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.01; \\*\\*\\*\u2009=\u00a0*p*\u00a0\\<\u20090.001\n\nDiscussion {#Sec24}\n==========\n\nIn a previous study we demonstrated that the pro-inflammatory cytokine TNF-\u03b1 is a potent trigger of cell fusion between human M13SV1-Cre breast epithelial cells and human breast cancer cell lines under both normoxic and hypoxic conditions \\[[@CR18]\\]. Here, we investigated the impact of TNF-\u03b1 on the gene expression profiles of M13SV1-Cre and MDA-MB-435-pFDR1 cells to identify which of the differentially-regulated genes might be involved in cell fusion. Of the identified target genes, only the inhibition of MMP9 was correlated with a decreased TNF-\u03b1-induced cell fusion rate.\n\nHere, we used the MDA-MB-435 cancer cell line, whose origin is debated to be either breast cancer or melanoma. A systematic analysis of the gene expression profiles of 60 human cancer cell lines first revealed that MDA-MB-435 cells express genes that are highly expressed in most melanoma-derived cell lines, suggesting that this cell line originated from melanoma rather than breast cancer \\[[@CR32]\\]. This assumption was further supported by comparative genomic hybridisation (VGH) and microsatellite polymorphism analyses showing that MDA-MB-435 cells and M14 cells share similarities \\[[@CR33]\\]. In contrast, MDA-MB-435 and M14 cells markedly differed in their DNA methylation profiles; more hypermethylated CpG islands were detected in MDA-MB-435 cells than in M14 cells and MDA-MB-435 cells were not grouped with melanoma cells after a hierarchical cluster analysis \\[[@CR34]\\]. Likewise, the phenotypic and molecular characterisation of MDA-MB-435 cells further revealed similarities to other breast cancer cell lines, such as MDA-MB-231, SUM1315 or HBL100 cells, which belong to the claudin-low intrinsic subtype of breast cancer \\[[@CR35]\\].\n\nPCR data showed that the microphthalmia-associated transcription factor (MITF), which is a lineage survival oncogene amplified in malignant melanoma cell lines, was expressed in MDA-MB-435 cells \\[[@CR36], [@CR37]\\]. However, MITF protein was not expressed in MDA-MB-435 cells \\[[@CR38]\\] but mRNA expression was detected in other breast cancer cell lines, such as MCF-10A, MCF-7, SKBR3, and U87 and D54 glioblastoma cell lines \\[[@CR37]\\]. The co-expression of neuronal, epithelial and melanocytic markers in breast cancer, melanoma and glioblastoma cell lines might be related to lineage infidelity, which seems to be a common phenomenon in cancer cells lines \\[[@CR37]\\]. In agreement with MCF-7 breast cancer cells, but in contrast to melanoma cells, the expression of breast epithelium-specific and epithelial-specific markers, such as \u03b2-casein, \u03b1-lactalbumin, epithelial membrane antigen (EMA), and keratin-19 as well as the induction of \u03b2-casein expression and production of milk lipids by \u03b2-heregulin and vitamin E was clearly observed in MDA-MB-435 cells \\[[@CR38]\\]. Moreover, MDA-MB-435 xenograft studies revealed that the vast majority of the cells were positive for cytokeratin and EMA, suggesting that MDA-MB-435 is a breast epithelial cell line that has gained melanocyte lineage characteristics \\[[@CR38]\\].\n\nOur data indicate that the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells depends on MMP9. Blocking MMP9 with a specific inhibitor or the inhibition of MMP9 expression using minocycline was associated with a decreased cell fusion frequency. These data agree with that of several studies that have already demonstrated that TNF-\u03b1 and/or MMP9 play a role in the fusion of different cell types \\[[@CR16]--[@CR19], [@CR39]--[@CR41]\\]. For instance, Song et al. showed that TNF-\u03b1 enhanced the fusion of oral squamous carcinoma cells and endothelial cells via a VCAM-1/VLA-4-dependent pathway \\[[@CR19]\\]. MMP9 was not investigated in this particular study and it is unclear whether it is involved in the fusion of these cells. TNF-\u03b1-induced VCAM-1 expression in endothelial cells is crucial for close cell-cell contact and is a prerequisite for the subsequent plasma membrane fusion step. Here, TNF-\u03b1 induced ICAM1 expression in both cell types, but no reduced cell fusion rate was observed in the presence of an ICAM1 blocking antibody. It is likely indicating that the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 was independent of ICAM1. Nonetheless, ICAM1 was recently identified to augment myoblast adhesion and fusion through homophilic trans-interactions and Rac-mediated actin remodelling \\[[@CR28]\\]. Therefore, the fusogenic property of ICAM1-ICAM1 interactions was restricted to myogenic cells, as forced expression of ICAM1 by fibroblasts did not augment their fusion to ICAM1-positive myoblasts/myotubes \\[[@CR28]\\].\n\nMMP9 is involved in macrophage fusion. The IL-4 induced fusion of macrophages was reduced in the presence of MMP9 function-blocking antibodies and similar effects were also observed in MMP9 null-macrophages, clearly showing the impact of MMP9 in macrophage fusion \\[[@CR40]\\]. Likewise, MMP9, E-cadherin and DC-STAMP were upregulated by both IL-4 signalling and DAP12 signalling in macrophages and DAP12 overexpression induced macrophage fusion \\[[@CR39]\\]. Conversely, DAP12 deficiency was associated with an impaired frequency of macrophage fusion because of low MMP9 levels \\[[@CR39]\\]. Low MMP9 expression levels, concomitant with a decreased macrophage fusion rate, were also found in MCP-1/CCL2-null mice \\[[@CR16]\\]. Moreover, the impaired fusion capacity of MCP-1/CCL2-null macrophages is rescued by exogenous TNF-\u03b1 and TNF-\u03b1-induced MMP9 expression \\[[@CR16]\\], which agrees with our data. The finding that TNF-\u03b1 could induce fusion via an MMP9-dependent mechanism is further supported by data showing that the fusion of osteoclasts in bone explants, which were stimulated by human breast cancer cells through TNF-\u03b1 secretion, could be blocked by the inhibition of MMP9 \\[[@CR41]\\].\n\nEven though these findings indicate a correlation between the induction of MMP9 expression and cell fusion, the detailed mechanism of how MMP9 is involved in the merging of plasma membranes still remains unclear. Cell fusion is a multi-step process that can be subdivided into i) priming, ii) chemotaxis, iii) adhesion, iv) fusion, and v) post-fusion. IL-4 and DAP12 signalling program cells into a fusion-competent status \\[[@CR4], [@CR39]\\], whereas MCP-1/CCL2 signalling is associated with chemotaxis \\[[@CR4], [@CR16]\\]. Because TNF-\u03b1 induces cell fusion, it might be capable of programming cells into a fusion-competent status. Because of its proteolytic activity, MMP9 may degrade extra-cellular matrix components, promote interaction with cell membranes or may induce signal molecules necessary for cell fusion \\[[@CR4], [@CR40]\\]. We wondered whether MMP9 may degrade membrane-bound ICAM1, which may allow the plasma membranes of two cells to come into close contact. Tsai and colleagues recently demonstrated that TNF-\u03b1 induced an MMP9-dependent release of soluble ICAM1 release in osteoblast-like MC3T3-E1 cells \\[[@CR31]\\]. However, increased levels of soluble ICAM1 were not observed in TNF-\u03b1-treated cells nor was the TNF-\u03b1 induced fusion of the cells inhibited by an ICAM1 blocking antibody.\n\nWhether CCL2 might be involved in the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells remains unclear. Markedly increased CCL2 expression levels were only observed in TNF-\u03b1-stimulated MDA-MB-435-pFDR1 cells. CCL2 might conceivably be involved in the induction of chemotaxis in M13SV1-Cre cells.\n\nThe finding that MMP9 plays a role in the TNF-\u03b1-induced fusion of M13SV1-Cre and MDA-MB-435-pFDR1 cells was further supported by minocycline studies, which revealed markedly lower MMP9 expression levels and a reduced fusion rate in the presence of TNF-\u03b1. This finding agrees with in vitro and in vivo studies that show that both tumour growth and osseous metastasis of breast cancer cells was effectively reduced by minocycline because of its inhibition of MMP9 expression \\[[@CR42]--[@CR44]\\]. It is well-known that besides its bacteriostatic efficacy against both gram-positive and gram-negative bacteria, minocycline also possesses anti-inflammatory properties (for review see: \\[[@CR25]\\]). The mechanisms involved in the anti-inflammatory activity of minocycline include its inhibitory effects on the activity of enzymes such as iNOS, the MMPs or COX2, the inhibition of apoptosis and the inhibition of immune cell activation and proliferation \\[[@CR25]\\]. Several lines of evidence indicated that minocycline exerted its inhibitory effects mainly through the inhibition of the NF-\u03baB pathway \\[[@CR45]--[@CR48]\\]. Minocycline attenuates bone cancer pain in rats by significantly decreasing both the total and nuclear expression of NF-\u03baB and p-IKK\u03b1 in astrocytes \\[[@CR46]\\]. Consistent with these findings, minocycline significantly suppresses IKK\u03b1/\u03b2 phosphorylation in LPS-stimulated THP-1 monocytic cells, suggesting that it inhibits NF-\u03baB signalling at the level of IKK\u03b1/\u03b2 phosphorylation \\[[@CR48]\\]. This agrees with the data that show that minocycline suppresses constitutive NF-\u03baB activation in OVCAR-3 and SKOV-3 ovarian carcinoma cells, which is correlated with the attenuation of IKK activation \\[[@CR47]\\]. Interestingly, further data revealed that the minocycline-induced suppression of NF-\u03baB activity was mediated, in part, through the inhibition of TGF-\u03b21 \\[[@CR47]\\]. It is well-known that TNF-\u03b1 signals via NF-\u03baB \\[[@CR49]\\]. Hence, minocycline most likely impairs TNF-\u03b1-induced cell fusion by blocking MMP9 expression in M13SV1-Cre cells because of the inhibition of NF-\u03baB signalling.\n\nConclusions {#Sec25}\n===========\n\nOur data indicate that the matrix metalloproteinase MMP9 is involved in the TNF-\u03b1-induced fusion of M13SV1-Cre cells and MDA-MB-435-pFDR1 cells. Likewise, the tetracycline-based antibiotic minocycline effectively impaired the TNF-\u03b1-induced fusion of the cells because of the inhibition of MMP9 expression.\n\nAdditional files\n================\n\n {#Sec26}\n\nAdditional file 1:Differentially-regulated genes in TNF-\u03b1-treated M13SV1-Cre and MDA-MB-435-pFDR1 cells under normoxic and hypoxic conditions. Genes that were up-regulated and down-regulated by at least 2-fold in TNF-\u03b1-treated cells compared with untreated cells (not detected flags and compromised spots were not removed). Genes shown in Fig. [1](#Fig1){ref-type=\"fig\"} are marked in yellow. (XLSX 29 kb) Additional file 2:Proliferation of M13SV1-Cre and MDA-MB-435-oFDR1 cells was impaired in a dose-dependent manner by minocycline. Cells were cultivated for up to 72\u00a0h with different minocycline concentrations. The data shown are the mean\u2009\u00b1\u2009SD of three independent experiments. (TIFF 478 kb)\n\nABCC6\n\n:   ATP binding cassette subfamily C member 6\n\nADAMTS9\n\n:   A disintegrin and metalloproteinase with thrombospondin motifs 9\n\nBMDCs\n\n:   Bone marrow-derived stem cells\n\nCCL2\n\n:   CC-chemokine ligand 2\n\nCLDN1\n\n:   Claudin-1\n\nCOL27A1\n\n:   Collagen type XXVII alpha 1 chain\n\nDC-STAMP\n\n:   Dendritic cell-specific transmembrane antigen\n\nDMEM\n\n:   Dulbecco's Minimal Essential Medium\n\nEGF\n\n:   Epidermal growth factor\n\nEMA\n\n:   Epithelial membrane antigen\n\nESA\n\n:   Epithelial specific antigen\n\nFCS\n\n:   Foetal calf serum\n\nFDR\n\n:   Fluorescence double reporter\n\nICAM1\n\n:   Intercellular cell adhesion molecule 1\n\nICAM2\n\n:   Intercellular cell adhesion molecule 2\n\nICAM4\n\n:   Intercellular cell adhesion molecule 4\n\nIL-13\n\n:   Interleukin-13\n\nIL32\n\n:   Interleukin 32\n\nIL-4\n\n:   Interleukin-4\n\nIL4I1\n\n:   Interleukin 4 induced 1\n\nIL7R\n\n:   Interleukin 7 receptor\n\nLAMB3\n\n:   Laminin subunit beta 3\n\nLAMC2\n\n:   Laminin subunit gamma 2\n\nlncRNA\n\n:   Long non-coding RNA\n\nM-CSF\n\n:   Macrophage-colony stimulating factor\n\nmiRNA\n\n:   microRNA\n\nMMP9\n\n:   Matrix metalloproteinase 9\n\nMUC4\n\n:   Mucin 4\n\nNF-\u03baB\n\n:   Nuclear factor-kappa B\n\nOC-STAMP\n\n:   Osteoclast stimulatory transmembrane antigen\n\nPBS\n\n:   Phosphate-buffered saline\n\nPCR\n\n:   Polymerase chain reaction\n\nqPCR\n\n:   Quantitative polymerase chain reaction\n\nRANKL\n\n:   Receptor activator of nuclear factor kappa-B ligand\n\nRIN\n\n:   RNA integrity number\n\nSDS-PAGE\n\n:   Sodium dodecyl sulphate-polyacrylamide gel electrophoresis\n\nsICAM1\n\n:   Soluble intercellular cell adhesion molecule 1\n\nSIRP\u03b1\n\n:   Signal-regulatory protein alpha\n\nTNF-\u03b1\n\n:   Tumour necrosis factor-\u03b1\n\nTNFAIP3\n\n:   Tumour necrosis factor-alpha induced protein 3\n\nVEGFC\n\n:   Vascular endothelial growth factor C\n\nZBAF\n\n:   Centre for Biomedical Education and Research\n\n**Electronic supplementary material**\n\nThe online version of this article (10.1186/s12964-018-0226-1) contains supplementary material, which is available to authorized users.\n\nThe authors are grateful to Frank Edenhofer from the Institute of Anatomy and Cell Biology, Julian-Maximilians-University of W\u00fcrzburg, W\u00fcrzburg, Germany for providing us with the pFDR1 vector. Likewise, the authors would like to thank Jan Postberg and Patrick Weil from the HELIOS Medical Centre Wuppertal, Wuppertal, Germany for helping us with the Agilent 2100 Bioanalyzer. Furthermore, we are grateful to Thomas Ostermann and Sebastian Appelbaum from the Department of Psychology and Psychotherapy, Witten/Herdecke University, for helping us with the statistical analysis.\n\nFunding {#FPar1}\n=======\n\nThis work was supported by the Fritz-Bender Foundation, Munich, Germany.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nThe datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.\n\nJW performed the experiments, analysed the data, wrote and corrected the manuscript. MM prepared samples for cDNA microarray data analysis; analysed and validated the cDNA microarray data. KSZ read and corrected the manuscript. TD designed the experiments and wrote and corrected the manuscript. All authors read and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar3}\n==========================================\n\nNot applicable.\n\nConsent for publication {#FPar4}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar5}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar6}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Background\n==========\n\nThe goal of genomic selection (GS) \\[[@B1]\\] is to capture all quantitative trait loci (QTL) influencing a trait by tracing all chromosome segments defined by adjacent markers. With use of highly dense markers, GS is supposed to be able to overcome the problem of traditional maker assisted selection (MAS) that only a limited proportion of the total genetic variance is captured by the markers of QTL. GS has become feasible very recently with the high throughput genotyping technology and the availability of highly dense markers covering whole genome. Genomic breeding value estimation is the key step in GS. A number of approaches have been proposed for estimating genomic breeding values \\[[@B1]-[@B9]\\], among which BLUP methods and Bayesian methods are most commonly used. Here, we applied two BLUP methods (GBLUP \\[[@B3]\\], TABLUP \\[[@B4]\\]) and three Bayesian methods (BayesA, BayesB \\[[@B1]\\], BayesC\u03c0 \\[[@B5]\\]) to the common dataset provided by the 15^th^QTL-MAS Workshop to evaluate and compare their predictive performances.\n\nMethods\n=======\n\nDataset\n-------\n\nThe common dataset consisted of an outbred population, which had been simulated using the LDSO software \\[[@B10]\\], with 1000 generations of 1000 individuals, followed by 30 generations of 150 individuals. 9990 SNP markers were distributed on 5 chromosomes. Each chromosome had a size of 1 Morgan and carried 1998 evenly distributed SNPs (1 SNP every 0.05 cM).\n\nThe final dataset used for evaluating genomic selection consisted of 3220 individuals, including 20 sires, 200 dams (each sire mated with 10 dams) and 3000 progenies (15 per dam). All individuals were genotyped for the 9990 SNPs without missing or genotyping error. Of the 15 progenies of each dam, 10 were phenotyped for a continuous trait. The 2000 progenies with phenotypic records and the other 1000 individuals (which had simulated true breeding values) without phenotypic records were treated as reference and validation population, respectively.\n\nEstimation of variance components and EBVs\n------------------------------------------\n\nThe variance components and the traditional BLUP EBVs were estimated using phenotypes and pedigree and the software DMUv6 \\[[@B11]\\] based on the following model:\n\n$$\\mathbf{y} = \\mathbf{1}\\mu + \\mathbf{Z}\\mathbf{a} + \\mathbf{e}$$\n\nwhere **y**is the vector of phenotypes of individuals in the reference population, *\u03bc*is the overall mean, **a**is the vector of additive genetic effects of the phenotyped individuals and their parents, **Z**is the incidence matrix of **a**, and **e**is the vector of residual errors. The variance-covariance matrices of **a**and **e**are $\\mathbf{A}\\sigma_{a}^{2}$ and $\\mathbf{I}\\sigma_{e}^{2}$, respectively, where A is the additive genetic relationship matrix, $\\sigma_{a}^{2}$ is the additive genetic variance, and $\\sigma_{e}^{2}$ is the residual variance.\n\nThe reliabilities of the traditional EBVs were obtained from DMU directly and calculated as the square of the correlation between EBVs and the true unknown breeding values.\n\nEstimation of SNP effects\n-------------------------\n\nBayesA, BayesB and BayesC\u03c0 were used to estimate SNP effects in the reference population based on the following model:\n\n$$\\mathbf{y} = \\mathbf{1}\\mu + \\mathbf{X}\\mathbf{g} + \\mathbf{e}$$\n\nwhere **g**is the vector of random SNP effects, **X**is the matrix of genotype indicators (with values 0, 1, or 2 for genotypes 11, 12, and 22, respectively).\n\nThe differences between the three Bayesian methods lay in the assumptions for the prior distribution of SNP effects. BayesA assumes that all SNPs have an effect, but each has a different variance. BayesB and BayesC\u03c0 assume that each SNP has either an effect of zero or non-zero with probabilities \u03c0 and 1-\u03c0, respectively, and for those having non-zero effects it is assumed that each SNP has a different variance in BayesB and a common variance in BayesC\u03c0. In addition, in BayesB \u03c0 is treated as a known parameter, while in BayesC\u03c0 it is treated as an unknown parameter with a uniform (0, 1) prior distribution. In this study, we set \u03c0 = 0.99 for BayesB, and adopted the same prior distributions of **g**and **e**for the three Bayesian methods as those in \\[[@B1],[@B5]\\].\n\nThe Markov chain was run for 50,000 cycles of Gibbs sampling (for BayesB, 100 additional cycles of Metropolis-Hastings sampling were performed for the SNP effect variance in each Gibbs sampling cycle), and the first 5000 cycles were discarded as burn-in. All the samples of SNP effects after burn-in were averaged to obtain the SNP effect estimate.\n\nCalculation of GEBVs\n--------------------\n\nThe genomic estimated breeding values (GEBVs) of all genotyped individuals were obtained using five methods: BayesA, BayesB, BayesC\u03c0, GBLUP and TABLUP.\n\nFor BayesA, BayesB and BayesC\u03c0, the GEBV of a genotyped individual was calculated as the sum of all marker effects according to its marker genotypes \\[[@B1]\\].\n\nFor GBLUP and TABLUP, the GEBVs were estimated based on the following model:\n\n$$\\mathbf{y} = \\mathbf{1}\\mu + \\mathbf{Z}\\mathbf{u} + \\mathbf{e}$$\n\nwhere **u**is the vector of genomic breeding values of all genotyped individuals with the variance-covariance matrix equal to $\\mathbf{G}\\sigma_{u}^{2}$ for GBLUP or $\\mathbf{T}A\\sigma_{u}^{2}$ for TABLUP. $\\sigma_{u}^{2}$ is the additive genetic variance estimated from the reference population.\n\nThe **G**matrix (realized relationship matrix) was constructed by using genotypes of all markers \\[[@B3]\\]. The **TA**matrix (trait-specific marker-derived relationship matrix), was constructed by using genotypes of all markers with each marker being weighted with its estimated effect obtained from BayesB following the rules proposed by Zhang et al. \\[[@B4]\\].\n\nThe accuracies of GEBVs were calculated as the correlation between GEBVs and the simulated true breeding values.\n\nResults and discussion\n======================\n\nVariance components\n-------------------\n\nThe estimated additive genetic variance and residual variance were 24.82 and 58.65, respectively. Therefore, the estimated heritability was 0.30. These estimates were used for the subsequent estimation of SNP effects and GEBVs.\n\nEstimates of SNP effects\n------------------------\n\nFigure [1](#F1){ref-type=\"fig\"} includes the profiles of SNP effects estimated by BayesA (Figure [1A](#F1){ref-type=\"fig\"}), BayesB (Figure [1B](#F1){ref-type=\"fig\"}) and BayesC\u03c0 (Figure [1C](#F1){ref-type=\"fig\"}). These estimated effects, which are obviously not evenly distributed, reflect the underlying architecture of the trait. The estimated value of \u03c0 in BayesC\u03c0 is 0.9986. In general, the SNP effect profiles from the three Bayesian methods are quite similar. In particular, all of the three methods show a big peak on chromosome 1, two peaks on chromosome 2, and a peak on chromosome 3. In addition, BayesC\u03c0 shows another peak on chromosome 3 and a peak on chromosome 4. No peaks appear on chromosome 5 for all of the three methods. The peak positions and the corresponding SNP effect estimates are given in Table [1](#T1){ref-type=\"table\"}. For chromosomes 1, 2 and 3, where one, two and two additive QTL were simulated, respectively, these peak positions match all the simulated QTL positions quite well, except that BayesA and BayesB missed one QTL on chromosome 3. For chromosomes 4 and 5, where an imprinted QTL and two epistatic QTL were simulated, respectively, either no peak was detected or the detected peak is far away from the simulated position. From these results, it seems that these methods could also serve as tools for QTL mapping and BayesC\u03c0 performed better in this respect. The drawback of BayesA and BayesB regarding the impact of prior hyperparameters and treating the prior probability \u03c0 as known has been addressed by Gianola et al. \\[[@B12]\\] and Habier et al. \\[[@B5]\\]. Our results partially confirmed their arguments.\n\n![**Absolute values of estimated SNP effects by BayesA (A), BayesB (B) and BayesC\u03c0 (C)**.](1753-6561-6-S2-S13-1){#F1}\n\n###### \n\nPeak positions of profiles of the estimated SNP effects and the corresponding estimated SNP effects\n\n  Method          Chr. 1   Chr. 2      Chr. 3   Chr. 4                                \n  --------------- -------- ----------- -------- ----------- ------ ----------- ------ -----------\n  BayesA          59       5.19\u00b10.37   3660     1.01\u00b10.90   4094   2.25\u00b10.40          \n                                       3914     0.35\u00b10.73                             \n  BayesB          59       1.96\u00b12.13   3660     0.73\u00b10.82   4092   0.91\u00b11.17          \n                                       3873     0.56\u00b10.65                             \n  BayesC\u03c0         58       5.15\u00b10.42   3660     0.93\u00b10.96   4092   2.50\u00b10.76   7234   0.53\u00b11.51\n                                       3873     0.76\u00b10.75   4331   0.41\u00b10.67          \n  Simulated QTL   57                   3638                 4100               6644   \n                                       3875                 4300                      \n\nCorrelations between GEBVs by different methods and between EBVs and GEBVs for the 20 sires\n-------------------------------------------------------------------------------------------\n\nFor the 20 sires, the reliability of traditional EBVs was 0.95. Table [2](#T2){ref-type=\"table\"} shows the correlations between GEBVs by different methods and between EBVs and GEBVs of the 20 sires. The correlations between EBVs and GEBVs by different methods ranged from 0.933 to 0.966, and the highest correlation was given by GBLUP and the lowest by BayesC\u03c0. In general, the GEBVs by different methods were highly correlated with the correlation coefficients over 0.95, indicating that the GEBVs for the 20 sires by different methods were quite consistent.\n\n###### \n\nCorrelations between GEBVs by different methods (the first 4 columns) and between traditional EBVs and GEBVs (the last column) for the 20 sires\n\n            BayesB   BayesC\u03c0   TABLUP   GBLUP   Traditional EBV\n  --------- -------- --------- -------- ------- -----------------\n  BayesA    0.999    0.995     0.995    0.972   0.942\n  BayesB             0.992     0.998    0.978   0.947\n  BayesC\u03c0                      0.986    0.956   0.933\n  TABLUP                                0.986   0.952\n  GBLUP                                         0.966\n\nCorrelations between GEBVs by different methods for the 1000 progenies without phenotypic values\n------------------------------------------------------------------------------------------------\n\nTable [3](#T3){ref-type=\"table\"} shows the correlations between GEBVs by different methods for the 1000 progenies without phenotypic values. The correlations ranged from 0.812 to 0.997, and the highest correlation was between TABLUP and BayesB, and the lowest between GBLUP and BayesC\u03c0. The correlations among the three Bayesian methods and TABLUP are all very high (over 0.97), indicating high similarity in GEBVs from these methods, while the correlations between them and GBLUP are all less than 0.9, indicating some differences in GEBVs exist herein.\n\n###### \n\nCorrelations between GEBVs by different methods for the 1000 progenies without phenotypic values.\n\n            BayesB   BayesC\u03c0   TABLUP   GBLUP\n  --------- -------- --------- -------- -------\n  BayesA    0.991    0.985     0.983    0.841\n  BayesB             0.986     0.997    0.860\n  BayesC\u03c0                      0.976    0.812\n  TABLUP                                0.876\n\nAccuracies and biases of GEBVs\n------------------------------\n\nThe availability of true breeding values (TBVs) of the 1000 progenies without phenotypic values allowed a more efficient assessment for methods. Table [4](#T4){ref-type=\"table\"} shows the correlations of TBVs and GEBVs, which measure the accuracies of GEBVs, and regressions of TBVs on GEBVs, which measure the biases of GEBVs, by different methods. In terms of both accuracy and bias, the three Bayesian methods and TABLUP performed similarly with correlations over 0.92 and slightly downward bias. BayesB and BayesC\u03c0 were slightly more accurate than BayesA and TABLUP, while TABLUP yielded smallest bias. GBLUP gave the lowest accuracy and the highest downward bias.\n\n###### \n\nAccuracies and biases of GEBVs for the 1000 progenies without phenotypic values.\n\n  Method    *r*     *b*\n  --------- ------- -------\n  BayesA    0.924   1.063\n  BayesB    0.933   1.068\n  BayesC\u03c0   0.938   1.057\n  TABLUP    0.924   1.033\n  GBLUP     0.774   1.648\n\n*r*: correlation coefficient between GEBV and TBV; *b*: regression coefficient of TBV on GEBV.\n\nTABLUP is an improvement of GBLUP in the way that the **G**matrix is replaced with **TA**matrix. In construction of the **TA**matrix, not only the marker genotypes, but also the marker effects are taken into account. The advantage of the **TA**matrix over the **G**matrix is that it not only accounts for the Mendelian sampling term, but also puts greater weight on loci explaining more of genetic variance for the trait of interest. This makes TABLUP more accurate than GBLUP. On the other hand, although TABLUP and the Bayesian methods gave similar accuracies, TABLUP has two important features that Bayesian methods lack. The first is that the reliability of an individual\\'s GEBV can be calculated by TABLUP through the method outlined for GBLUP by VanRaden \\[[@B3]\\] and Strand\u00e9n et al. \\[[@B13]\\]. The second is that TABLUP can be extended to estimate GEBVs for individuals without genotypes by constructing a joint pedigree-genomic relationship matrix according to the rule proposed by Legarra et al. \\[[@B14]\\].\n\nConclusions\n===========\n\nBayesA, BayesB, BayesC\u03c0 and TABLUP performed similarly and satisfactorily and remarkably outperformed GBLUP for genomic breeding value estimation in this dataset. TABLUP is a promising method for genomic breeding value estimation because of its easy computation of reliabilities of GEBVs and its easy extension to real life conditions such as multiple traits and consideration of individuals without genotypes.\n\nList of abbreviations used\n==========================\n\nQTL: quantitative trait locus; MAS: marker assisted selection; GS: genomic selection; BLUP: best linear unbiased prediction; GBLUP: BLUP with a realized relationship matrix; TABLUP: BLUP with a trait specific relationship matrix; EBV(s): estimated breeding value(s); GEBV(s): genomic estimated breeding value(s); TBV(s): true breeding value(s); SNP: single nucleotide polymorphism.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nCLW, PPM and ZZ contributed the data analyses and the manuscript. XDD and JFL contributed the modification of manuscript. WXF and ZQW carried out the data analyses. QZ coordinated the analyses and revised the manuscript. All authors have read and contributed to the final text of the manuscript.\n\nAcknowledgements\n================\n\nThis work was supported by the State High-Tech Development Plan of China (Grant No. 2008AA101002, 2011AA100302), the National Natural Science Foundation of China (Grant No. 30800776, 30972092, 31171200), Beijing Municipal Natural Science Foundation (Grant No. 6102016), and the Modern Pig Industry Technology System Program of Anhui Province.\n\nThis article has been published as part of *BMC Proceedings*Volume 6 Supplement 2, 2012: Proceedings of the 15th European workshop on QTL mapping and marker assisted selection (QTL-MAS). The full contents of the supplement are available online at <http://www.biomedcentral.com/bmcproc/supplements/6/S2>.\n"}
+{"text": "1. Introduction {#sec1-sensors-18-03568}\n===============\n\nThe Internet of Things (IoT) currently makes it possible to have a world sensed by and connected to all kinds of devices. Wireless Sensor Network (WSN) technology is the key for connecting physical and virtual environments. This technology is growing up so rapidly that in 2011 Cisco-IBSG estimated that globally there would be 50 billion interconnected \"things\" by 2020 \\[[@B1-sensors-18-03568]\\]. The IoT paradigm leads to an extremely large number of new opportunities and technical challenges in several fields in general, and in the industrial field in particular. In industry, wired technologies continue to be prevalent \\[[@B2-sensors-18-03568],[@B3-sensors-18-03568]\\]. Digital technologies like ModBus, ProfiBus, CanBus, HART \\[[@B4-sensors-18-03568]\\], and even analogue technologies like 4--20 mA \\[[@B5-sensors-18-03568]\\], are used to monitor and control most processes. Despite the high reliability of such technologies, proven over many years, wired technologies are expensive, difficult to install, time consuming, and unable to cope with the requirements of Cyber Physical Systems (CPS) and Industry 4.0. In Industry 4.0, CPSs will confer micro intelligence (namely processing and networking capabilities) to industrial objects, reducing even further today's already short production cycles \\[[@B6-sensors-18-03568]\\]. Additionally, in the near future, new markets will drive CPSs to increase their level of adaptability, directly connecting customers to manufacturing facilities, using Cyber Physical Manufacturing Systems (CPMS) \\[[@B7-sensors-18-03568],[@B8-sensors-18-03568]\\]. Thus, WSNs or, more specifically, Industrial Wireless Sensor Networks (IWSN), are fundamental for meeting the requirements of Industry 4.0.\n\nIWSN characteristics like low operating costs, self-organization, self-configuration, flexibility, rapid-deployment, and easy upgrading, make them ideal to industrial scenarios. However, despite all of these favourable characteristics, the adoption of WSNs in industry requires standards, dependability, ease of use, network security, extended battery life, low cost, and IP connectivity \\[[@B9-sensors-18-03568]\\]. In recent years, considerable effort was made in order to design technologies that meet these requirements, and to standardize IWSN technology. Standards like IEEE 802.15.4 \\[[@B10-sensors-18-03568]\\] and IEEE 802.15.1 \\[[@B11-sensors-18-03568]\\] are the technology foundation of many industrial applications for process and factory automation. IEEE 802.15.4 is the base technology for standards such as ZigBeePRO \\[[@B12-sensors-18-03568]\\], WirelessHART \\[[@B13-sensors-18-03568]\\], ISA100.11a \\[[@B14-sensors-18-03568]\\], and WIA-PA \\[[@B15-sensors-18-03568]\\]. These are widely used in process automation applications in the areas of chemical manufacturing, pulp and paper, oil and gas, and glass and mineral \\[[@B16-sensors-18-03568]\\]. On the other hand, IEEE 802.15.1 is the base technology for standards such as WISA \\[[@B17-sensors-18-03568]\\] and WSAN-FA \\[[@B9-sensors-18-03568]\\], widely used in factory automation applications in the areas of assembly process for automotive, consumer products and electronics \\[[@B18-sensors-18-03568]\\].\n\nNevertheless, although standards compliance is necessary, it is not enough to guarantee ISWN reliability per se. Sensor node components, either at hardware or firmware levels, and the network itself, can be at the root of a variety of faults \\[[@B19-sensors-18-03568]\\]. Sensor nodes are inherently resource-constrained devices in terms of energy, processing power and memory capacity. In this respect, the Internet Engineering Task Force (IETF) recently defined three classes of devices \\[[@B20-sensors-18-03568]\\]: Class 0 for devices with less than 10 KB of RAM and 100 KB of flash memory; Class 1 for devices with around 10 KB of RAM and 100 KB of flash; and Class 2 for devices that have more resources but are still quite constrained when compared to high-end devices. In addition to the device constraints mentioned, sensor, network, and application heterogeneity lead to extremely complex IWSN solutions and, consequently, to fault-proneness. These characteristics impose adequate, carefully-designed strategies in the development of sensor nodes firmware, hardware architectures, and operating systems (OSs) kernel (e.g., choosing between exokernel, microkernel, monolithic approach or hybrid approach) \\[[@B21-sensors-18-03568]\\]. As a way of example, OEM manufacturers can build their products based on a single chip (comprising wireless communications and processing capabilities) \\[[@B22-sensors-18-03568]\\], or separate microcontroller and radio (connected by SPI or UART) \\[[@B23-sensors-18-03568]\\]. Furthermore, applications may be developed on \"bare metal\" (which makes them very hardware-specific), or using one of the available OSs (e.g., Contiki \\[[@B24-sensors-18-03568]\\], RIOT \\[[@B25-sensors-18-03568]\\], FreeRTOS \\[[@B26-sensors-18-03568]\\]). Another important component is the network. For instance, despite the inclusion of security mechanisms in all of the referred standards, there are known attacks on WirelessHART, ZigBeePRO, ISA100.11a, and WIA-PA \\[[@B27-sensors-18-03568],[@B28-sensors-18-03568],[@B29-sensors-18-03568],[@B30-sensors-18-03568]\\]. Additionally, some of these technologies, namely WirelessHART, ZigBeePRO, and WIA-PA, are not immune to interference from equipment complying with other standards, such as IEEE 802.11, when operating in the ISM 2.4 Ghz frequency band. Such problems may lead to early sensor node energy depletion, and subsequent replacement, increasing the costs of network operation. Because of this, post-deployment tools are needed in order to adequately monitor IWSNs, thus contributing to the global system reliability.\n\nIn the last decade, a wide range of WSN post-deployment tools \\[[@B31-sensors-18-03568]\\] were developed. Some of them can improve the reliability of WSN by detecting network, firmware, and/or hardware problems. These tools help developers in both deployment and post-deployment environments by making several firmware- and hardware-related metrics accessible, and by detecting problems, using, for instance, sniffers or sink nodes. However, despite the effort to build such tools, most of them were designed for specific applications, require specific or dedicated hardware, consume non-negligible amounts of energy, do not implement security mechanisms, are complex to configure and/or to use, and do not allow the centralized management of multiple industrial standards like ZigBeePRO, WirelessHART, ISA100.11a, and WIA-PA.\n\nIn this context, the contributions of this paper are the following: (i) a survey of ISWN standards (for process-automation) and their features per OSI layer; (ii) an overview of the reports and metrics made available by each of the IWSN standards; (iii) a survey of the main techniques for WSN monitoring at network, firmware, and hardware levels; (iv) a proposal for an industrial IoT monitoring architecture that builds on IEEE 802.15.4-based industrial standards (WirelessHART, ISA100.11a, ZigBeePRO and WIA-PA) to provide management functionality; and (v) a proof-of-concept implementation of the proposed architecture. It should be noted that the proposed architecture does not increase the cost of node manufacturing, has negligible impact on the main sensor node application, hardware components, and network bandwidth and, last but not least, does not significantly increase energy consumption.\n\nThe remainder of this paper is organized as follows. [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"} details the main industrial IoT technologies, and provides a revision of applicable management protocols. In [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, representative diagnostic tools for collecting hardware, firmware and network metrics are described. [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"} presents the proposed architecture in detail and, at the same time, explains how technologies, approaches, and protocols can be used to implement it. [Section 5](#sec5-sensors-18-03568){ref-type=\"sec\"} presents a proof-of-concept implementation in a testbed using the WirelessHART standard, with the aim of assessing the architecture's impact in terms of energy consumption, application latency, RAM and flash utilization, and network traffic. [Section 6](#sec6-sensors-18-03568){ref-type=\"sec\"} compares the proposed architecture with related work in this domain. Lastly, [Section 7](#sec7-sensors-18-03568){ref-type=\"sec\"} presents the conclusions and guidelines for future work.\n\n2. Technologies for Industrial IoT {#sec2-sensors-18-03568}\n==================================\n\nReliable, controlled operation, and use of standardized technologies are key to the adoption of WSNs in industrial applications \\[[@B32-sensors-18-03568]\\]. In best-effort type networks, all devices in the network obtain an unspecified data rate and delivery time, depending on the traffic load \\[[@B33-sensors-18-03568]\\]. As a result, data is delivered without any quality of service guarantees. However, in general, industrial process automation applications have stringent latency requirements. For instance, monitoring applications should guarantee an average latency of 100 ms; control applications should guarantee 10 ms to 100 ms latency; and, lastly, safety applications should guarantee a maximum latency of 10 ms \\[[@B14-sensors-18-03568]\\]. To meet these requirements, IWSNs may have to pre-allocate network bandwidth and physical resources, thus avoiding statistical effects that lead to insufficient bandwidth, uncontrolled jitter, and congestion losses. With these constraints in mind, this section surveys the main IWSN technologies and standards (namely, IEEE 802.15.4, ZigBeePRO WirelessHART, ISA100.11a and WIA-PA), highlighting their main features, monitoring functionality, and applicable management protocols.\n\n2.1. IWSN Standards {#sec2dot1-sensors-18-03568}\n-------------------\n\nThis subsection analyses the characteristic features of each of the main IWSN standards, namely ZigBeePRO, WirelessHART, ISA100.11a, and WIA-PA. The analysis is done on a per-OSI-layer basis, starting with the physical layer and working up to the application layer. [Table 1](#sensors-18-03568-t001){ref-type=\"table\"}, below, presents a summary of the referred features, and may be used as guidance for the reader.\n\nIEEE 802.15.4 \\[[@B10-sensors-18-03568]\\] is a standard for low-rate wireless personal area networks (LR-WPANs) that specifies the physical (PHY) and medium access control (MAC) layers. These layers' design was optimized for very low-power consumption, high reliability (using mesh networks), low-data rates, and low-cost. ZigBee, WirelessHART, ISA100.11a, and WIA-PA have all adopted IEEE 802.15.4 at the PHY layer. However, because WirelessHART, ISA100.11a, and WIA-PA target worldwide adoption, they've chosen to use the 2.4 GHz frequency band only, as opposed to ZigBee, which can use the 868 Mhz and 915 Mhz bands as well \\[[@B18-sensors-18-03568]\\].\n\nAt the MAC layer, there are significant differences in the adoption of the IEEE 802.15.4 by each of the four standards. At this layer, networks can work in beacon or non-beacon modes of operation. When using the beacon mode, sensor nodes receive a specific message that is used to synchronize the network devices and, at the same time, to identify the network, and to describe the structure of the frame. Beacon networks are capable of detecting other beacon networks, and, for this reason, beacon networks can coexist in the same geographical area. Additionally, IEEE 802.15.4 uses a superframe structure in order to manage nodes channel access. The superframe is formed by three different parts: Contention Access Period (CAP), Contention Free Period (CFP), and inactive period. This superframe structure is used in ZigBee and WIA-PA because these two standards operate in beacon mode. On the other hand, WirelessHART and ISA100.11a do not operate in beacon-mode, as their developers considered that this mode is not good enough for industrial applications (Note: WIA-PA also shares the same view; however, its authors opted for maintaining full compatibility with IEEE 802.15.4-based networks, and implement additional features at the Data Link (DL) layer). As a result, WirelessHART and ISA100.11a implemented their own superframes \\[[@B34-sensors-18-03568]\\]. WirelessHART's superframe is composed of 10 ms timeslots, while ISA100.11a can operate in any of three superframe modes (short, long, and hybrid). Finally, ZigBee can use a slotted or unslotted CSMA/CA mechanism for managing the access to the wireless medium, while the remaining standards (WirelessHART, ISA100.11a and WIA- PA) use TDMA and CSMA, providing a high and medium level of latency determinism, respectively.\n\nNeither WirelessHART or ISA100.11a implement the full IEEE 802.15.4 MAC layer, as they consider that the MAC layer of IEEE 802.15.4 is not capable of delivering the deterministic latency needed by industrial applications. As such, these standards extend and complement medium access mechanisms with functionality at the data link (DL) layer \\[[@B9-sensors-18-03568]\\], namely, Time Synchronized Channel Hopping (TSCH), which then evolved to the new IEEE 802.15.4.e standard. The TSCH mechanism offers two significant improvements: the possibility to have deterministic latency (communication resources are pre-allocated); and a mechanism of channel hopping that minimizes interference with nearby devices that operate in the same frequency, such as IEEE 802.11 devices. On the other hand, the WIA-PA follows the IEEE 802.15.4 standard at the DL layer, including functionalities like time synchronization and frequency hopping techniques. Lastly, ZigBee does not implement any mechanism at this layer. It is also worthwhile mentioning that, in addition to extending/complementing MAC layer functionality, the DL layer is also used by ISA100.11a for implementing some network-related functions, specifically in what concerns routing. In fact, this standard implements two types of routing: one at the DL layer, which handles all the IEEE 802.15.4 traffic, and another one at the Network (NWK) layer, responsible for handling the IPv6 backbone traffic, as we will see below.\n\nAt the NWK layer, the choice of supported functionality and routing protocols is often influenced by the network architectures \\[[@B18-sensors-18-03568]\\]. For instance, ZigBee offers the possibility of having star, tree, and mesh topologies, and defines several field devices: coordinator, routers, and end-devices. Tree routing and Z-AODV protocols are used when the network operates in tree or mesh topology, respectively. Despite the fact that mesh networks can be considered more reliable, in ZigBee, the utilization of this topology is not suitable for industrial applications, due to the overhead and non-deterministic latency of on-demand protocols like Z-AODV \\[[@B9-sensors-18-03568]\\]. In the case of WirelessHART, the basic network devices are: field devices, gateway, access points, and network and security manager. Typically, WirelessHART gateways support the role of security and network manager, and access point \\[[@B35-sensors-18-03568]\\].\n\nWirelessHART networks may operate in star or mesh topologies. However, the mesh topology is the most used one, due to its flexibility, inherent fault-tolerance, and ease of deployment and configuration. With this topology, routing can be done by using graph routing or source routing. When graph routing is used, the network manager needs to compute all graphs in the network and share them with sensor nodes. All communications using graph routing always consider two different paths to ensure reliability. In contrast, when source routing is used, network packets are forward between intermediate devices without the need for prior route information (the path of the packet is specified in the packet itself) \\[[@B3-sensors-18-03568]\\]. Differently from the other standards, ISA100.11a specifies two types of networks: IEEE 802.15.4 (i.e., WSN) network, and backbone network. In the WSN network, ISA100.11a defines three device types: routing devices, field devices, and handheld devices. The routing mechanisms available in this network are the same as the ones in WirelessHART. Additionally, in the infrastructure (i.e., backbone) side, ISA100.11a uses 6LowPAN, thus allowing for direct communication between external IP devices and ISA100.11a devices. Lastly, WIA-PA supports a hierarchical topology that uses star and mesh, or star-only topology. In the case of the mesh topology, the network operates using routers and gateways, while in the star topology the network is composed of routers and field/handheld devices. At this level, field devices are cluster members that acquire sensor information and send it to the cluster heads (routers). Then, the cluster-heads form the mesh network. Each routing device in the network shares its neighbour information with the network manager, and then the network manager computes and shares the static routes. For each pair of devices that want to communicate, at least two routing paths are assigned.\n\nThe transport (TP) layer is responsible for providing host-to-host communication services between applications. As can be seen in [Table 1](#sensors-18-03568-t001){ref-type=\"table\"}, only WirelessHART and ISA100.11a implement data transport functions at this layer, supporting different service level agreements. Optionally, ISA100.11a offers end-to-end security at this layer. In contrast to ISA100.11a and WirelessHART, WIA-PA provides different service level agreements at the application sub-layer (APS), and not at the transport layer. The service-level agreements available in each standard are presented in [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"} and [Section 4.2](#sec4dot2-sensors-18-03568){ref-type=\"sec\"}. Last but not least, ZigBee does not support any transport layer functionality nor does it support traffic differentiation at the APS sub-layer. Moreover, in ZigBee, fragmentation, reassembly, and device discovery are implemented at the APS sub-layer.\n\nThe application (APP) layer is the layer at which the connection between legacy systems and IEEE 802.15.4-based systems takes place \\[[@B9-sensors-18-03568],[@B18-sensors-18-03568]\\]. At this layer, there are two important options that can be identified. WirelessHART uses a command-oriented approach; alternatively, ISA100.11a, ZigBee, and WIA-PA use a more flexible object-oriented approach. Object-oriented approaches are more flexible than command-oriented approaches because they allow for protocol translation by mapping attributes from one protocol to the other. As for native applications, ZigBee supports the ZigBee profiles; WirelessHART supports the HART protocol; WIA-PA supports native protocols like Profibus, FF, and HART; last but not least, ISA100.11a supports the ISA100.11 application protocol.\n\n2.2. IWSN Reports {#sec2dot2-sensors-18-03568}\n-----------------\n\nIWSN standard technologies make device and network state information available to a variety of entities, e.g., network neighbours or a central management device. In general, this information is shared between nodes to compute routes, allocate bandwidth, calculate link costs between network devices, or generate alarms when critical events occur (e.g., link failure, route failure, low battery level, etc.). This subsection presents the network and data link layer reports available in each of the standards being considered in this paper \\[[@B12-sensors-18-03568],[@B13-sensors-18-03568],[@B14-sensors-18-03568],[@B15-sensors-18-03568]\\], and describes the context in which the reports are used. [Table 2](#sensors-18-03568-t002){ref-type=\"table\"} presents a summary of the referred reports, and may be used as guidance for the discussion.\n\nThe ZigBee standard comprises three report types that are used for sharing network- and node-related information: (1) link status report; (2) network status report, and (3) network report. Link status reports share sensor node's neighbours incoming and outgoing link costs. The report is broadcasted by ZigBee coordinators and routers in one-hop fashion. This report is useful during network discovery, to find neighbour devices, and in the operation phase for updating the devices' neighbour table. Network status reports are sent by devices in order to report errors and other events that arise at the network layer. The report can be sent in unicast or broadcast modes over the network, and can only pertain to an event at a time. The list of the possible reported events is presented in the [Table 2](#sensors-18-03568-t002){ref-type=\"table\"}. Last but not least, network reports allow a device to report network events, like PAN conflict, or the radio channel condition to the coordinator. As limitation, in the tree topology, some of these packets nay not be received by the ZigBee coordinator, due to the presence of routing devices, that make their own routing decisions. On the other hand, when using a star topology, all the packets will be received at the ZigBee coordinator.\n\nContrary to what happens in ZigBee, in WirelessHART, the network manager controls all communication in the network, and only authorizes new services if resources are available. Consequently, field devices (here with full network capabilities) must share node state and network based information with the network manager. In WirelessHART, all network reports arrive to the network manager by using the maintenance service. WirelessHART specifies four types of reports: (1) device health; (2) neighbour health list; (3) neighbour signal levels; and (4) alarm report. Device health reports summarize all the communication statistics of a unique field device and are periodically sent to the network manager. The statistics include generated packets by device, terminated packets by device, power status, and others. Neighbour health list reports include statistics about the communication with all neighbours linked to a field device. These reports include the total number of linked neighbours, the mean Received Signal Level (RSL) to the neighbour, and packets and errors statistics. Neighbour signal level reports include statistics of discovered but not linked neighbour devices detected by a field device. When a device wants to connect to the network, it usually sends a join request and a neighbour signal level report. Lastly, WirelessHART defines several alarm types: path down alarm; source route failed alarm; and graph route failed alarm.\n\nSimilarly to what happens in WirelessHART, in ISA100.11a networks, the system manager controls all communication in the network. However, in this standard, WSN routing takes place at the DL layer, due to the use of 6LoWPAN at the network layer. Network metrics in ISA100.11a are shared at the MAC layer, instead of being shared at the NWK layer. ISA100.11a defines two groups of network reports: (1) connectivity alert, and (2) neighbour discovery alert. Connectivity alerts comprise two types of reports: per-neighbour report and per-channel report. Per-neighbour reports contain neighbours' connection statistics. On the other hand, per-channel reports contain per-channel all-neighbours statistics, and convey them to the system manager. Finally, neighbour discovery alerts are sent periodically to the system manager with a list of overheard neighbours. The system manager makes new routing decisions based on these reports. Per-neighbour reports, per-channel reports, and neighbour discovery alert, are similar to WirelessHART's neighbour health list, device health, and neighbour signal level, respectively.\n\nLast but not least, in WIA-PA, route computation is also performed by the network gateway, similarly to WirelessHART and ISA100.11a. WIA-PA defines four types of reports: (1) device status report, (2) channel condition report, (3) neighbour report, and (4) path failure report. Device status reports include statistics about the condition of field devices and routers, such as the number of packets exchanged with neighbours, number of restarts, and uptime. Channel condition reports send statistics grouped by channel and neighbour to the network gateway. These statistics include link quality, packet loss rate, and number of retries. Neighbour reports, also received by the network gateway, group neighbour statistics and neighbour scheduling details, such as backoff counter and exponent, transmitted and received packets, and number of acknowledgments. Lastly, path failure reports are generated when a route path failure occurs. These reports are sent by a routing device to the network gateway, whenever the retransmission counter of a specific path exceeds a given threshold.\n\n2.3. Management Protocols {#sec2dot3-sensors-18-03568}\n-------------------------\n\nNetwork management emerged in traditional networks with the need to control and monitor networks as they grew up in size and complexity. ISO/IEC 7498-4 \\[[@B36-sensors-18-03568]\\] was one of the first initiatives to establish a management framework, by defining a set of functional areas, namely: fault, configuration, accounting, performance, and security management. Furthermore, ISO/IEC 7498-4 proposed the use of managed objects as abstractions for network resources, which, in turn, contain several attributes. In current protocols, managed objects can be defined using several syntaxes, like SMIv2 \\[[@B37-sensors-18-03568]\\], YANG \\[[@B38-sensors-18-03568]\\], and XML \\[[@B39-sensors-18-03568]\\]. Managed objects data are stored in management databases and accessed by management protocols. In this sub-section, we identify two types of management protocols: (1) protocols designed for traditional networks; and (2) protocols designed for networks of resource-constrained devices, that can also be used in some traditional networks.\n\nAs one of most used protocols for the management of IP devices, SNMP \\[[@B37-sensors-18-03568]\\] provides most of the basic functionality defined in ISO/IEC 7498-4. The standard specifies an architecture based on agents and managers. Devices being monitored, run SNMP agents that share the device state information with an SNMP manager, by using query/response interactions and trap notifications. Each SNMP agent has a collection of managed objects whose data are stored in a Management Information Base (MIB). Each object is identified using a specific object identifier (OID). Despite the success of SNMP for monitoring purposes, the protocol failed to provide an effective and reliable way to configure devices. Thus, IETF started a working group that wrote an informational RFC \\[[@B40-sensors-18-03568]\\] with several requirements that coming network management standards should implemented. This work was the basis for the NETCONF \\[[@B41-sensors-18-03568]\\] protocol. Differently from SNMP's manager-agent architecture, NETCONF uses a client-server architecture. The server runs on a management device (SNMP agent) and shares the monitoring information (also by query/response and notification messages) with the client (SNMP manager). Additionally, when devices need to be configured, the client may send several configuration commands in one or several transactions. The transactions supported in NETCONF give operators the capability of sending out-of-order commands, and of performing rollback and commit operations.\n\nThe management protocols mentioned up to now were designed for traditional networks. In \\[[@B42-sensors-18-03568]\\], the authors implemented traditional network management protocols in networks of constrained devices, and concluded that SNMP and NETCONF are not suitable for this type of networks. For instance, the use of TLS encryption in NETCONF adds significant overhead in terms of session time (i.e., in the order of seconds). In later years, with the growth of IoT, new solutions that address resource-constrained devices and 6LowPAN networks were proposed. These solutions also take advantage of transport protocols specially designed for resource-constrained devices, like CoAP \\[[@B43-sensors-18-03568]\\].\n\nOne of the first solutions developed with a focus on the management of 6LowPAN networks was the LoWPAN Network Management (LNMP) protocol \\[[@B44-sensors-18-03568]\\]. The LNMP protocol implements a solution based on SNMP that targets 6LowPAN networks. In this solution, 6LowPAN gateways convert the information from 6LowPAN networks to SNMP MIBs and make it available over IP. On the other hand, other management architectures were evaluated, and new research directions that take advantage of new technologies like HTTP appeared. In \\[[@B45-sensors-18-03568]\\], the authors evaluated the use of different architectures like SNMP, Resource-Oriented Architecture (ROA), and Service-Oriented Architecture (SOA), and concluded that ROA architectures are more suitable for resource-constrained devices in terms of response time and power consumption, and are less sensitive to changes in timeout. As a result, a RESTful version of NETCONF was created, named RESTCONF \\[[@B46-sensors-18-03568]\\]. Distinct from NETCONF, which uses SSH and TCP, RESTCONF allows the communication between server and client using HTTP operations. Using HTTP at the application layer enables clients to receive notifications without maintaining a permanent connection with the server, as it is the case of NETCONF (one of its major drawbacks). The syntax used in RESTCONF is YANG, the same syntax used in NETCONF. Also using an ROA-based architecture but with a different application protocol, the CoAP Management Interface \\[[@B47-sensors-18-03568]\\] (COMI) is an internet draft that intends to provide access to resources specified in YANG or SMIv2, using CoAP. The draft defines, as in the cases of NETCONF and RESTCONF, a separation between operational and configuration data store, the use of DTLS, and a conversion of YANG string identifiers to numeric identifiers that contributes to reducing the payload size.\n\nLast but not least, LwM2M \\[[@B39-sensors-18-03568]\\] is also a protocol designed for resource-constrained devices. This protocol provides device management and application management, an aspect that differs from the previously mentioned network management protocols. As COMI, LwM2M also supports the use of CoAP at the application layer. Being a REST-based protocol, LwM2M uses GET/PUT/POST operations to perform read/write/execute operations over the managed objects resources. In this protocol, the definition of managed objects is done using XML. An extensive list of managed objects is available in OMA \\[[@B48-sensors-18-03568]\\]. In addition to these, OMA allows the creation of specific objects by individuals, organizations, and companies.\n\nSumming up, in this sub-section, a set of network management protocols were presented (see [Table 3](#sensors-18-03568-t003){ref-type=\"table\"}). By analysing the protocols presented here, we identified protocols suitable to managing traditional networks (switches, routers, computers), and protocols designed for networks of resource-constrained devices (e.g., sensor nodes). An important trend is that new protocols like LwM2M use general-purpose languages to describe the managed objects (e.g., XML), instead of YANG and SMIv2. All of these management protocols can be incorporated into the different gateways identified in the industrial standards presented before (i.e., ZigBee coordinator, WirelessHART network manager, ISA100.11a system manager, and WIA-PA network gateway) because these roles are performed by unconstrained devices. Finally, protocols like SNMP, NETCONF, RESTCONF, COMI and LwM2M cannot be directly used in sensor nodes because current standard industrial technologies do not allow running application protocols other than their own. Thus, the management of these kinds of networks can only be done in a standardized way at gateway level. Only new standards like 6tisch \\[[@B49-sensors-18-03568]\\] support the management of sensor node devices using COMI or LwM2M because 6tisch uses the CoAP protocol at the application level. Nevertheless, due to the fact that no 6tisch-based products are available and, consequently, it is not yet used in industrial settings, 6tisch will not be addressed in this paper.\n\n3. WSN Diagnostic Tools {#sec3-sensors-18-03568}\n=======================\n\nThe design of a diagnostic tool applicable to low-end IoT devices, like WSN nodes, is a tough task due to their characteristics and diversity. WSN nodes have limited resources, support a variety of application architectures, and rely on complex network mechanisms. In addition, WSN applications can be developed for a specific operating system, or even almost from scratch. These characteristics make it difficult, or even impossible, to develop a common diagnostic tool for all kinds of scenarios, and, at the same time, compatible with several operating systems. Despite this, in the last decade, several diagnostic tools were proposed in order to provide inside views on WSN networks' and nodes' behaviour. In \\[[@B31-sensors-18-03568]\\], the authors analyse an extensive set of diagnostic tools. In this section, some of the tools described in \\[[@B31-sensors-18-03568]\\], as well as more recent tools like \\[[@B50-sensors-18-03568],[@B51-sensors-18-03568],[@B52-sensors-18-03568],[@B53-sensors-18-03568],[@B54-sensors-18-03568]\\], will be presented from a diagnostic target perspective, organizing their presentation into network-based, firmware-based, and hardware-based tools.\n\n3.1. Network Tools {#sec3dot1-sensors-18-03568}\n------------------\n\nThe nature of wireless communications makes this technology unreliable and failure-prone. In this context, diagnostic tools that are able to collect network metrics and evaluate the state of the network are essential. Three types of approaches can be used to collect network information: (1) passive, using extra hardware to collect network information without interference; (2) active, using on-node available resources; and (3) hybrid, using a mix of active and passive approaches.\n\nWhen passive approaches are used, the acquisition of network-related information is made by installing extra sniffer nodes or sniffer capabilities in the existing sink nodes. These diagnostic tools may differ in the type of sniffer nodes, in the storage approach, and in the way the gathered information is transmitted. Specifically, some of the existing solutions use sink sniffer nodes to collect traffic data from the network; others use extra networks of sniffers deployed with the main WSN; others use sniffer nodes that collect data and store it in memory; and, lastly, the most expensive in terms of network installation, send network-related information by wired technologies (e.g., SNIF \\[[@B55-sensors-18-03568]\\], SNTS \\[[@B56-sensors-18-03568]\\], PDA \\[[@B57-sensors-18-03568]\\], LiveNet \\[[@B58-sensors-18-03568]\\] and L-SNMS \\[[@B59-sensors-18-03568]\\]).\n\nDifferently from passive tools that rely on extra hardware to monitor the network, active approaches use on-node metrics already available in sensors nodes. Active tools are easy to install, do not need extra hardware and, consequently, are less expensive when compared to passive approaches. However, active tools may have impact on node resources, as memory, energy, processing, and network throughput are needed to collect, store, and transport network-related information. Some of the tools gather traffic metrics from the sensor nodes' operating system, others from the network layer, and others directly from sink nodes, based on the received network traffic. In addition, there is a specific set of tools that use software scripts deployed in sensor nodes to collect statistics about the network traffic seen by the node. Subsequently to data acquisition, the transport of network-related information can be made using the main application channel or a secondary channel. To minimize the impact of data transport, some tools use compression and aggregation techniques so as to reduce the traffic overhead. Examples of this type of tools are Marionete \\[[@B60-sensors-18-03568]\\], Megs \\[[@B61-sensors-18-03568]\\], Memento \\[[@B62-sensors-18-03568]\\], Wringer \\[[@B63-sensors-18-03568]\\], 6PANview \\[[@B50-sensors-18-03568]\\], and D2 \\[[@B53-sensors-18-03568]\\].\n\nLastly, hybrid approaches use a mix of methods described in the cases of the active and passive approaches. Examples of this type of tools are Sympathy \\[[@B64-sensors-18-03568]\\] and Dustminer \\[[@B65-sensors-18-03568]\\].\n\n3.2. Firmware Tools {#sec3dot2-sensors-18-03568}\n-------------------\n\nAnother source of faults commonly addressed by diagnostic tools is the sensor nodes' firmware. After firmware development, sensor nodes are deployed in the field and, in some cases, faults may stay in a dormant state until an input activates them. If the error is not properly handled, a firmware failure may occur and sensor node data may not be delivered or may be corrupted. In an extreme although not uncommon situation, a firmware fault may even prevent a sensor node from entering sleep mode and, eventually, lead to battery exhaustion. With the aim of promptly detecting firmware faults, there are diagnostic tools that help developers in either the development phase, or in the WSN deployment phase.\n\nTools that are used during the development phase usually require debugging interfaces, specific hardware implemented in the microcontrollers architecture, and Integrated Development Environments (IDEs) that allow for accessing the available functionality. Examples of these technologies are: the old JTAG interface used to program microcontrollers and to have access to special internal registers (e.g., hardware and software breakpoints); UART ports used for outputting log messages that are useful for debugging purposes; the EnergyTrace \\[[@B66-sensors-18-03568]\\] technology, from Texas, that allows developers to measure the impact of firmware on nodes energy consumption; and the CoreSight \\[[@B67-sensors-18-03568]\\] technology, implemented by ARM in their microcontrollers, which allows performance profiling, memory access, real-time tracing, and software debugging through a new type of interfaces, namely the Serial Wire Debug and the Serial Wire Output Pin (examples of variables that can be access using this type of technology are cycles per instructions, sleep cycles, and exceptions). On top of that, IDEs like Code Compose Studio (CCS) \\[[@B65-sensors-18-03568],[@B68-sensors-18-03568]\\], among others, enable access to these tools and help developers to detect code faults.\n\nNone of the development phase diagnostic technologies are suitable for post-deployment because they require a wired connection between nodes and the tool's hardware. Consequently, in the last decade, several deployment phase tools were proposed that are able to provide monitoring information, although with considerable limitations when compared to the functionality delivered by development phase tools. Some examples are \\[[@B52-sensors-18-03568]\\], Nucleos \\[[@B69-sensors-18-03568]\\], Enverilog \\[[@B70-sensors-18-03568]\\], Marionete \\[[@B60-sensors-18-03568]\\], Clairvoyant \\[[@B71-sensors-18-03568]\\], NodeMD \\[[@B72-sensors-18-03568]\\], L-SNMS \\[[@B73-sensors-18-03568]\\], LIS \\[[@B74-sensors-18-03568]\\], Memento \\[[@B62-sensors-18-03568]\\], Dustminer \\[[@B65-sensors-18-03568]\\], DT \\[[@B75-sensors-18-03568]\\], Tracealyzer \\[[@B76-sensors-18-03568]\\], and Dylog \\[[@B51-sensors-18-03568]\\]. Despite their limitations, these tools allow WSN operators/managers to collect operating systems variables (e.g., task queue state, number of reboots), main application variables, and application events and states (like function calls). This information is usually stored in flash, RAM, or sent via the application main channel using logs. In order to deliver all of these without requiring an extra effort during the main application development, code instrumentation techniques are used, which make it possible to automate the process of adding firmware metrics collection and transmission functionality. Firmware data is usually sent using one of two paradigms: event-driven or query-driven. Additionally, some of the tools implement extra functionality, such as: source-level debugging, offering commands similar to hardware debugging (break, stop, watch, backtrace, etc.); remote call-specific functions; remote reprogramming; and specification of trace log events triggering conditions. Tools that usually read and write large amounts of data from/to flash memory (e.g., Tracealyzer) are not, in general, appropriated for industrial WSN technologies due to the energy wasted in the process.\n\n3.3. Hardware Tools {#sec3dot3-sensors-18-03568}\n-------------------\n\nFinally, a source of faults that diagnostic tools also commonly address is hardware faults. Hardware faults may occur before and/or after WSN deployment, and are due to one or more of several reasons, such as: bad hardware design, external environment phenomena, aging, and extreme temperatures. Hardware diagnostic tools can be divided in two groups: (1) external-physical-tools that are used by developers and operators; and (2) on-node-tools that infer hardware faults using available hardware resources.\n\nThe first group of tools consists of physical tools/equipment that operators and developers use to check hardware condition and the occurrence of faults. Tools like oscilloscopes, multimeters, and logic analysers are connected to the target system in order to check the condition of the electronics and the existence of faults. These tools are frequently used during the development phase, where developers search for electronics defects and for communication problems between hardware modules (e.g., using logic analysers \\[[@B77-sensors-18-03568]\\]).\n\nOn-node-tools use a different approach. This type of tools is designed during the hardware and/or firmware development phase, with the specific purpose of collecting state information from the different modules using on-node components. In \\[[@B78-sensors-18-03568]\\], the authors use the ARM CoreSight technology to create a WDP (Watch Dog Processor), by polling the main processor memory variables and setting conditions on these variables. Also using the same technology, in \\[[@B79-sensors-18-03568]\\], the authors extend Hardware in Loop (HIL) tests, by incorporating some metrics provided by the CoreSight technology. Moreover, in \\[[@B80-sensors-18-03568]\\], the authors propose a technique to read the energy wasted in boards that use switching regulators. By collecting these metrics, operators and developers are able to detect hardware faults that occur during the deployment phase without using external tools.\n\n4. Proposed Architecture {#sec4-sensors-18-03568}\n========================\n\nCurrent WSN diagnostic tools have several drawbacks, as already pointed in our previous work \\[[@B31-sensors-18-03568]\\]. In our opinion, some of these drawbacks are hampering the use of WSNs in industry because, nowadays, to the best of our knowledge, there aren't multi-network, standard-compliant monitoring tools that support the IWSN technologies addressed in this proposal (ZigBee, ISA100.11a, WirelessHART, WIA-PA). In order to make this clear, in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"} and [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, a review of the state-of-the-art in what concerns the main Industrial IoT standards, network management, and diagnostic tools was made. As a result, some questions arose: how can multiple networks, possibly comprising equipment compliant with different standards, be monitored in an integrated way? How can management functionality be added at the gateway level? How can firmware and hardware be monitored without extra costs in hardware, firmware, and network?\n\nIn what concerns network monitoring, three important aspects were identified in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"} and [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}: the set of techniques used by current diagnostic tools to monitor the network; the available metrics, provided by almost all industrial standards, that can be collected and shared at gateway level; and the current management standards. In what concerns hardware and firmware monitoring, several approaches and technologies were also presented, as well as log techniques that can be used to convey hardware and firmware metrics inside the network. Thus, the review made in these sections identified several crucial points and research directions that led us to the proposal of this architecture and related properties. Implementations of the proposed architecture will lead to diagnostic tools that are totally compatible with industrial standards. In this respect, a proof-of-concept implementation will be presented in [Section 5](#sec5-sensors-18-03568){ref-type=\"sec\"}, which also discusses evaluation results.\n\nIt should be highlighted that the current architectural proposal goes well beyond identifying components and respective abstract relations. More than defining the requirements, interactions, and roles to be performed by each architecture component, this section defines technologies, approaches, and protocols to be used in realistic, practical industrial scenarios. For instance, this section addresses the techniques and technologies for collecting hardware, firmware, and network metrics; defines the services used in each standard for conveying the monitoring information; and identifies the management protocols and technologies for dealing with the monitoring information.\n\nThe remainder of this section is organized as follows: initially, the proposed monitoring architecture is presented, providing the reader with a global view of its components and their main roles; subsequently, each component is described in detail, by specifying their role and requirements. Moreover, the technologies applicable to each component will be identified.\n\n4.1. Architecture Overview {#sec4dot1-sensors-18-03568}\n--------------------------\n\nWith the emergence of several IWSN standards and the increase in the number of IWSN deployments, it is crucial to define an architecture able to monitor multiple-network, standard-compliant technologies. The proposed architecture, presented in [Figure 1](#sensors-18-03568-f001){ref-type=\"fig\"}, represents a flexible, scalable, energy-efficient, low-cost, and multi-standard solution that covers hardware, firmware, and network monitoring. In order to ease the adoption by IWSN vendors, OEM producers, and developers, the architecture was designed according to six main guidelines: (i) it should support the monitoring of multiple IWSNs; (ii) it should support multiple IWSN standards; (iii) it should not lead to a significant increase in energy expenditure; (iv) the collection of hardware metrics should not increase the cost of manufacturing; (v) the acquisition of metrics should not have a significant impact on the main application size and delay, nor should it lead to a large traffic overhead; (vi) the network metrics defined by each IWSN standard should be used; and, lastly, (vii) the collection of sensor-node metrics (hardware/firmware) and network metrics should be independent. The proposed architecture has five base modules: (1) sensor node monitoring agent; (2) gateway monitoring agent; (3) monitoring logger; (4) management agents; and (5) management system.\n\nThe sensor node monitoring agent is responsible for the collection of node monitoring data (hardware and firmware metrics), and the sending of this metrics to the network gateway. The latter will forward the metrics to the monitoring logger, where they will be parsed and stored. The monitoring agent will use the most appropriate service in each industrial standard for forwarding the metrics. These metrics are encapsulated in a specific application format. During its operation (either when collecting information or when sending it), the monitoring agent should minimize the impact on the available resources.\n\nEach industrial gateway has a pair of agents (one monitoring agent and one management agent). The gateway monitoring agent is the component that collects the network metrics and the gateway state (globally called management objects) and stores them in a local database (the datastore). These are then accessed in a standardized way by management systems, through the services delivered by the gateway management agent. In order to support the interoperability with management systems, the gateway also stores the representation of the management objects (i.e., IWSN standard metrics, and gateway state).\n\nOn the other hand, the handling of the collected sensor node monitoring data is carried out by a monitoring logger component, which parses the log messages and stores them in its datastore. The monitoring logger is a software component with two main sub-components: log parser and management agent. The log parser is the component that receives the log messages from the gateway, parses them, and stores them in the local datastore. Like the gateway, the monitoring logger locally stores a representation of the management objects. This representation enables the management agent to share the sensor node metrics with the management system in a standardized way (i.e., by using description languages such as SMIv2, XML e YANG). It should be noted that the monitoring logger is a logical component that can be deployed either on the gateway (if the manufacturer supports it) or on the management system.\n\nThe management system receives network monitoring data from the gateway management agent, and sensor node data (hardware and firmware metrics) from the monitoring logger management agent. Besides the traditional functions of configuring the monitoring capabilities of IWSN devices, the management system can include, for instance, a diagnostic tool that alerts operators or developers of critical events in the network, hardware, or firmware. Thus, the management system is capable of monitoring sensor nodes and network behaviour of multiple IWSNs using the standards addressed in this paper. This section outlined a proposal for a monitoring architecture for IWSNs that: does not require any modification to IWSN standardized technologies, benefits from the management information provided by each IWSN standard, and communicates with management systems in a standardized way. This addresses requirements (i), (ii), (vi) and (vii), identified in the beginning of this section. The next sub-sections detail each architecture component and show how the identified questions and requirements (especially, the ones more related with implementation strategies) are address by this proposal.\n\n4.2. Sensor Node Monitoring Agent Overview {#sec4dot2-sensors-18-03568}\n------------------------------------------\n\nSensor node metrics (hardware and firmware) are collected by the sensor node monitoring agent, as opposed to network metrics, which are collect by the gateway monitoring agent. Independence between the acquisition of network and sensor node metrics is one of the main features of the proposed architecture. This makes it possible to monitor the network independently from the acquisition of sensor node metrics. This separation also allows having closed modules in IWSN gateways (provided by IWSN vendors) and a more open monitoring logger module (that can be extended by developers if needed).\n\nThis section presents and discusses the main requirements for hardware and firmware metrics acquisition in the sensor node monitoring agent, and provides guidelines as to the technologies that can be used in this scope. Additionally, the approaches used in the transport of the logged data over the IWSN will be also presented.\n\n### 4.2.1. Hardware Metrics Collection {#sec4dot2dot1-sensors-18-03568}\n\nIn order to persuade IWSM vendors and OEM producers to incorporate real-time monitoring in sensor nodes' hardware, the monitoring techniques used by the sensor node monitoring agent should meet the following requirements: the added hardware components should not significantly increase the cost of sensor node manufacturing; the board space required by monitoring components should be minimal; the energy consumed by the hardware monitoring components should have a negligible impact on the battery life; processing and memory overhead should be minimum; and lastly, access to the hardware metrics should be seamless across all hardware platforms.\n\nRegarding the latter requirement, considering the techniques presented in [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, those who enable direct access to microcontroller metrics (by using internal hardware registers and hardware counters) have a higher level of portability when compared to external physical tools. From the sensor node monitoring agent perspective, seamless access to hardware metrics, regardless the underlying technology, is extremely important. By using different Board Support Packages (BSPs) developed by hardware manufactures (e.g., Drivelib), or by using interfaces available in the sensor nodes operating systems (e.g., Contiki, RIOT), it is possible for sensor node monitoring agents to gather hardware metrics directly from each hardware platform. Hardware monitoring techniques that collect vast quantities of data to infer the hardware's condition are not appropriate for industrial WSN application scenarios.\n\nThe work performed in \\[[@B78-sensors-18-03568],[@B79-sensors-18-03568],[@B80-sensors-18-03568],[@B81-sensors-18-03568]\\] presents a set of metrics that can be collected with little cost and extra value to assess the health of hardware components. Specifically, in \\[[@B80-sensors-18-03568]\\], the authors present a low-cost technique that, using an extra wire connected to an MCU hardware counter, enables the measurement of sensor nodes energy consumption. This technique was used in \\[[@B81-sensors-18-03568]\\], in an anomaly detection system, with good results, and proved its low footprint in terms of memory, cost, and processing load. Using the same MCU technology (the MSP430 family), in \\[[@B81-sensors-18-03568]\\], the authors present a technique that enables data acquisition on microcontroller cycles and execution time. In 16-bit architectures, the MSP430 family is one of the most used microcontrollers in WSN hardware due to its low power consumption. On the other hand, in 32-bit architectures, ARM microcontrollers are the most used ones. For these platforms, the work presented in \\[[@B78-sensors-18-03568],[@B79-sensors-18-03568]\\] describes a set of metrics that can be collected from the CoreSight technology available in the ARM architecture. The Data Watch Point DWT block can provide several measurement statistics, such as interrupt overhead, and sleep cycles. Another block, the Embedded Trace Macrocell ETM, can provide statistics concerning the number of executed or skipped instructions.\n\nSumming up, the monitoring of hardware condition can be performed at hardware level in compliance with the defined requirements. Furthermore, the proposed methods allow data acquisition for different types of microcontroller architectures and technologies. A set of techniques already addressed and proven by the scientific community were presented, and are recommended for use in the proposed architecture\n\n### 4.2.2. Firmware Metrics Collection {#sec4dot2dot2-sensors-18-03568}\n\nDevelopment of firmware for sensor nodes can be made using one of two types of approaches: the \"bare metal\" approach or the OS-based approach. When the \"bare metal\" approach is used, the application is usually developed in C or C++, and access to the hardware is made by the BSP supplied by the manufacturer, resulting in a very hardware-dependent development. This type of development is less portable than OS-based development, in which hardware operations are managed by the OS. Developers only have to call OS generic functions that will, in turn, be translated into hardware-specific calls.\n\nFirmware monitoring techniques used by the sensor node monitoring agent should support the monitoring of applications developed either using bare metal or OS-based approaches. Thus, as main requirements, firmware monitoring techniques should: be application independent; support, at least, C and C++ (according to \\[[@B21-sensors-18-03568]\\] most OSs use these languages); be OS-independent; when enabled, have a negligible impact on the execution of the main application (i.e., a minimal increase in the microcontroller's load); minimize the use of RAM, program memory, and external memory; be easy to integrate into the developers work flow; and not depend on physical access to the hardware. Similar to hardware monitoring techniques, firmware techniques that collect vast quantities of data and need physical access are not appropriate to this architecture.\n\nFrom the set of techniques used in the tools presented in [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, instrumentation techniques are the only ones that fulfil the mentioned requirements. Tools that use instrumentation techniques with scripting languages allow developers to easily integrate them in their daily workflow and, at the same time, provide code tracing, debugging, profiling, and performance counters. Supporting instrumentation in C and C++ languages allows the use of instrumentation techniques in almost all OSs. In the proposed architecture, the sensor node monitoring tool collects firmware metrics by using instrumentation techniques.\n\nIn \\[[@B51-sensors-18-03568],[@B52-sensors-18-03568],[@B53-sensors-18-03568],[@B74-sensors-18-03568]\\], the authors present some technologies, techniques, and metrics that use code instrumentation. For instance, in \\[[@B74-sensors-18-03568]\\], the authors use the C Intermediate Language (CIL) to instrument the C source code. The instrumentation code is inserted by a parser before building the binary file. Using a similar approach, in \\[[@B52-sensors-18-03568]\\], the authors use the PYCParser tool to instrument the code to support logging. On the other hand, in \\[[@B50-sensors-18-03568],[@B51-sensors-18-03568]\\] the authors use binary instrumentation to inject the monitoring code using the trampoline technique. Differently from the other instrumentation techniques, in binary instrumentation, developers do not need to have access to the source code to change the program flow. Using trampoline techniques, the displaced instructions are executed and then another jump is made back to the actual code. This method allows the use of instrumentation in running code. These techniques allow the collection of several firmware metrics like function header, control flow (if, else, switch), function footer, function calls, variables values, and number of variable assignments.\n\nIn conclusion, from all the techniques presented in [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, solutions that need physical access to hardware interfaces (e.g., JTAG) to collect firmware monitoring metrics cannot be used in real industrial deployments. For this reason, the sensor node monitoring agent proposed in the current architecture collects firmware metrics using techniques deployed in the firmware by code instrumentation.\n\n### 4.2.3. Transport of Collected Data {#sec4dot2dot3-sensors-18-03568}\n\nIWSN standards were engineered to optimize four important sensor node resources: node battery, reliability, latency, and network bandwidth. In this type of networks, sensor nodes should operate during several years, and network resources should be optimized for conveying the collected data with a variety of QoS requirements. Typically, the transport of sensor node monitoring information has lower priority than the main application traffic because the collection of this information cannot compromise the sensor node main application (nor its operating lifetime). Thus, the transport of sensor node monitoring data made by the sensor node monitoring agent should fulfil the following requirements: the monitoring information should be sent using appropriate network services and without compromising the main sensor application; the monitoring information storage should be managed without compromising the main application (for instance using a FIFO buffer in RAM, and dropping data when the buffer is full); when available, data should have a network timestamp to make it possible to correlate several events in the network; when security is enabled for the main application, the monitoring data should also be secured; when log packets exceed the maximum payload size and no fragmentation and reassembly mechanisms are available in the underlying layers, the sensor node monitoring agent should support fragmentation and reassembly mechanisms; lastly, the impact of such operations on battery life should be minimal.\n\nRegarding the timestamp requirement, from the set of the four analysed standards ZigBee is the only one that does not define a time standard in its specification. On the other hand, WirelessHART, ISA100.11a, and WIA-PA define their own time standard as Universal Time Coordinated (UTC), International Atomic Time (TAI), and UTC, respectively, and provide services to synchronize sensor nodes \\[[@B9-sensors-18-03568]\\]. Consequently, in ZigBee, sensor node monitoring data cannot be correlated without implementing additional mechanisms in the sensor nodes that enable for synchronizing the network time between nodes.\n\nAs already mentioned, the Maximum Transmission Unit (MTU) in IEEE802.15.4 is 127 bytes. Thus, all of the four standards use fragmentation and reassembly mechanisms to overcome this limitation. WIA-PA and ISA100.11a implement this mechanism at the network layer, unlike ZigBee that implements it at the application support sub-layer. WirelessHART also supports the transport of large packets, but, in this case, sensor nodes need to allocate a special type of transport service---the block data transfer. This service only allows the existence of a unique block transfer service for the whole network \\[[@B9-sensors-18-03568],[@B12-sensors-18-03568]\\]. Thus, in WirelessHART, the transport of blocks of data that exceed the maximum MTU size requires fragmentation and reassembly to be implemented at application level.\n\nAnother important aspect is the security of sensor node monitoring data. All of the industrial standards within the scope of this architecture offer hop-by-hop and end-to-end encryption. In the cases of ZigBee, WirelessHART, and ISA100.11a, security is managed using a centralized approach. On the other hand, WIA-PA uses a distributed approach, where the security manager together with the routing devices configure the security measures and forward the keys to field devices. All the standards under consideration provide encryption using symmetric keys, although ISA100.11a can also use asymmetric encryption in device joining processes. Last but not least, the highest level of IEEE 802.15.4 security policy is AES-128-CCM. Consequently, most IEEE802.15.4 radio chips support AES-128-CCM operations at the hardware level \\[[@B9-sensors-18-03568]\\].\n\nIn order to minimize the impact on the main application traffic, the approach to transporting sensor nodes monitoring data should be carefully selected for each standard, as presented below.\n\nIn WirelessHART, data transport can be done using one out of four types of services: block transfer service, maintenance service, periodic service, and event service. The block transfer service can only be used by one sensor node at a time. Thus, this service is not appropriate to transport sensor node monitoring data. The maintenance service provides the network with a minimal bandwidth for basic sensor nodes control and management operations. Because this is a low bandwidth service, it cannot handle the transport of sensor node monitoring data. The event service is used by applications to send data packets during unexpected events, such as alarms and warnings (when this service is requested, the radio needs to define the latency associated to the service) \\[[@B35-sensors-18-03568]\\]. Finally, the publish service is used to periodically send data, like sensor readings (when this service is requested, the radio needs to define the interval for sending the data) \\[[@B82-sensors-18-03568]\\]. Taking into consideration the characteristics of each service, we conclude that the most appropriate service to transport the data is the publish service, since the event service is a service used by applications with latency constraints, and the monitoring data does not have such requirement.\n\nZigBee does not specify services for data transport, leaving this to the application layer. According to the ZigBee specification \\[[@B12-sensors-18-03568]\\], in order to guarantee compatibility among different manufactures, ZigBee devices need to implement application profiles (also called endpoints). An endpoint is a ZigBee application for a specific domain with a set of clusters (or application messages) that can be mandatory or not. In turn, clusters are composed by attributes that represent the exchanged data. ZigBee sensor nodes that implement the same ZigBee profile (endpoint) are able to communicate with each other. In the context of this architecture, the sensor node monitoring agent is an endpoint and a set of cluster messages that can be sent to the ZigBee coordinator (that must also support the same endpoint).\n\nCompared to the other standards, ISA100.11a is by far the most complex and customizable standard (see [Table 4](#sensors-18-03568-t004){ref-type=\"table\"}). Instead of using the term services, used in WirelessHART, in ISA100.11a, contracts establish the resources allocated by the system manager to devices. Before a device can send data to another device, a contract needs to be created. Contracts are identified by an ID, unique within the scope of the device (but not necessarily so in the scope of the network), and are unidirectional. The system manager is the device that has the authority to assign, modify, and revoke contracts. Like WirelessHART services, there are several types of contracts and attributes that can be used for establishing service levels. Firstly, contracts may be of two types: periodic, that schedule network resources for the periodic sending of data; or, otherwise, non-periodic. Secondly, contracts can also be negotiable. The system manager may change or revoke the contract to make resources available to other high priority contracts. Lastly, contracts can have several levels of priorities: best effort queued, used in client-server communications; real time sequential, used in voice and video applications; real time buffer, used for periodic communications; and network control, used for managing network devices by the system manager. Message retransmission can, additionally, be enabled or disabled inside the contract. Thus, sensor node monitoring agents that run inside an ISA100.11a network should request a non-periodic contract type, which can be negotiated and revoked if needed. In this way, the system manager can revoke the contract of the sensor node monitoring agent, thus guaranteeing that the main sensor application can deliver sensor data without disruption. Additionally, the contract priority used by the sensor node monitoring agent should have the best effort queued type.\n\nIn the case of WIA-PA, networks may operate in two distinct modes: a hierarchical network topology that combines star and mesh, or a star-only topology. The star-only topology is a special case of the hierarchical network. For this reason, this topology uses the same services available in the hierarchical topology for data transport. In WIA-PA, the Virtual Communication Relationship (VCR) is the main standard block to access the objects specified in the User Application Objects (UAOs). VCRs distinguish the routing and communication resources allocated to each UAO. Each VCR has a VCR identifier, a source UAO ID, a destination UAO ID, address of source device/destination device and the VCR type. Similar to WirelessHART services and ISA100.11a contracts, in WIA-PA, VCRs can be classified according to the application scope: publish/subscriber (P/S) VCRs, used for publishing periodic data; report source/sink (R/S) VCRs, used for transferring aperiodic events and trend reports (alarms); and client/server (C/S) VCRs, used for transferring aperiodic and dynamic paired unicast messages (for getting and setting operations in UAO). Additionally, VCRs also provide aggregation methods. In this architecture, the sensor node monitoring agent data are UAOs capable of representing log messages. Thus, sensor node monitoring agents, available in each field device, as well as routing devices, need to select the appropriate VCR. P/S VCRs are appropriate for real-time operations, using exclusive timeslots in intra and inter-cluster communication. For this reason, this type of VCR should not be used for sending the monitoring data. On the other hand, R/S and C/S VCRs use the CAP period of the IEEE802.15.4 slot inside clusters, and use shared timeslots in inter-cluster operations. Thus, C/S and R/S VCRs are the most appropriate for transporting sensor node monitoring data in WIA-PA networks.\n\n4.3. Gateway Monitoring Agent {#sec4dot3-sensors-18-03568}\n-----------------------------\n\nThe gateway monitoring agent is the component in charge of representing management objects data in a standardised way, and of sending network monitoring data to the management systems. This agent also deals with data from multiple standards. The gateway monitoring agent is independent from the sensor node monitoring agent, thus allowing the separation of sensor node data collection from network monitoring data coleection. In this context, gateway monitoring agents must meet the follow requirements: network monitoring should not significantly increase the cost of equipment nor the cost of installation; collecting network metrics should add low overhead to the gateway; monitoring should be energy-efficient; the monitoring solution should be easy to install and should be extensible; and, lastly, gateway monitoring agents should support widely used industrial standards, namely the ones addressed in this paper.\n\nConsidering the three types of techniques identified in [Section 3.1](#sec3dot1-sensors-18-03568){ref-type=\"sec\"} (active, passive, and hybrid), active type techniques are the only ones that are easy to install and do not rely on extra hardware to perform monitoring tasks. However, these types of techniques usually consume sensor node resources, as pointed in [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}. On the other hand, in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"}, the revision made to the metrics shared between nodes showed us that the standards under consideration can support the sharing of critical information related to network and sensor node state, which can be used for resource allocation and routing tasks. Thus, by using these metrics, active tools will be capable of fulfilling the identified requirements without spending significant sensor node resources. Furthermore, some current industrial solutions available on the market already provide these metrics at the gateway, using proprietary libraries. The only disadvantage of active type techniques is the partial coverage in ZigBee networks because some metrics cannot be gathered by the coordinator. This arises from the fact that, in ZigBee, route computation is done in a distributed fashion for some topologies.\n\n4.4. Monitoring Logger {#sec4dot4-sensors-18-03568}\n----------------------\n\nAs shown in [Table 1](#sensors-18-03568-t001){ref-type=\"table\"}, the standards under consideration use different application approaches. Specifically, ZigBee, ISA100.11a, and WIA-PA use object-oriented representation, whereas WirelessHART uses a command-oriented representation. Sensor node monitoring data needs to comply with the representation defined in each standard, which leads to different representations for the same sensor node data. To solve this issue, the monitoring logger is the component that parses distinct standard representations into the same representation.\n\nThe monitoring logger should fulfil the following requirements: support the connection to the different standards \"gateways\"; support the parsing of the distinct application protocols using a sub-component (the log parser); support the representation of sensor node monitoring data in the languages supported by the management protocols; allow access to the monitoring data by management systems, using the management agent; and, lastly, store the sensor node monitoring data in a datastore. As a software component, the monitoring logger can be installed as an extra software module in industrial gateways or in another compatible equipment.\n\n4.5. Management Agents and Management System {#sec4dot5-sensors-18-03568}\n--------------------------------------------\n\nManagement systems are one of the building blocks of current IP-based networks. By monitoring the networks and their equipment, operators and manufactures can maximize the network uptime, improving the delivered quality of service. When applied to IWSNs, network and sensor node monitoring can also provide similar benefits. It is unthinkable to have a large IWSN that uses one or more standards without a central management system to perform predictive maintenance of sensor nodes. As presented in the previous sections, hardware, firmware and network monitoring data can be delivered by the sensor node monitoring agent and by the gateway monitoring agent, respectively. However, a key to the puzzle of this architecture is missing. To improve interoperability between these agents, different manufactures, and diagnostic tools, appropriate management protocols and syntax languages are needed. In this context, management agents should fulfill the following requirements: agents should allow the representation of sensor node and network data as managed objects; when needed, managed objects should be able to be extended, supporting additional monitoring metrics; management agents should be able to share management object information with management systems, updating the model used in management systems when needed; the management protocol should be able to support security mechanisms; management agents should run in IWSN gateways, for which the minimum hardware requirements should be set to IETF Class 2 \\[[@B20-sensors-18-03568]\\] (e.g., gateways that can support, for instance, a simplified version of Linux-based operating system); management protocols should support notification messages (query-based architectures can be heavy for IWSN gateways and are not scalable); lastly, the management agent must operate in IP networks (IWSN gateways must be able to be connected to an IP network).\n\nWhen network management technologies were previously analyzed in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"}, some protocols, syntaxes, and key features were identified. Current network management protocols can be divided into protocols designed for resource-constrained devices, and protocols for traditional networks. While traditional network management protocols are more mature and widespread, resource-constrained management protocols use more advanced, simpler, and modern technologies for representing management objects (e.g., YANG), and for data transport (e.g., CoAP, HTTP). Here, we highlight that protocols designed for resource-constrained devices can also be used in more powerful devices, like IWSN gateways. However, there are other aspects that need to be addressed regarding the requirements defined in this section. From the set of syntaxes for representing management models, XML and YANG are the most flexible and most simple languages. The learning curve to describe management objects using SMIv2 is longer than using languages like YANG and XML. Thus, SMIv2 is not recommended in the scope of this architecture. Consequently, SNMP is also not recommended for this architecture. Moreover, when working with data models, one important requirement is the sharing of the model with management systems. From the remaining protocols proposed in the scope of this architecture, NETCONF \\[[@B41-sensors-18-03568]\\], COMI \\[[@B47-sensors-18-03568]\\], RESTCONF \\[[@B46-sensors-18-03568]\\], and LwM2M \\[[@B39-sensors-18-03568]\\] allow the discovery of the models and of the resources available in each management node, by using discovery commands or specific URIs that enable to retrieve the used models. Additionally, other important requirement is the use of notification messages. By creating subscriptions to different management topics, diagnostic tools will be able to have the monitoring information in real time and without extra effort, contrary to what happens in query-based protocols that need to actively look for the monitoring information. Considering the set of analysed management protocols, the only one that does not support notifications is the NETCONF protocol. Before a client and a server can exchange management messages, NETCONF has to establish a permanent connection using SSH and TCP. Thus, it is impossible to notify management clients in the case of a connection loss. RESTCONF, LWM2M, and COMI have support for notification messages.\n\nLast but not least, one of the main characteristics of protocols for resource-constrained devices is the capability to use the protocols in the node itself (something they were developed for). However, in IWSNs, for which application protocols are statically specified by the standard, it is not possible to use other protocols in sensor nodes, such as the management protocols we've been addressing. The only option is to represent managed objects in the gateway of each standard. The authors of \\[[@B83-sensors-18-03568]\\] present a solution to monitor and manage legacy systems with LWM2M protocol, by using several LWM2M clients to represent each legacy device behind the gateway. Additionally, as presented in \\[[@B84-sensors-18-03568]\\], using the YANG language it is possible to represent the network metrics shared with a gateway in a WirelessHART network. In this way, COMI and RESTCONF can also be used in the gateway to represent legacy devices.\n\n5. Proof-of-Concept {#sec5-sensors-18-03568}\n===================\n\nIn the previous section, the monitoring architecture was presented as well as its components, requirements, and available solutions for each component in light of the existing standards and protocols. Being a general monitoring architecture that integrates several network standards, hardware technologies, and firmware architectures, a proof-of-concept implementation of such architecture showing each supporting technology and monitoring technique is unfeasible. Thus, in this section, a proof-of-concept scenario is presented using a small testbed with a WirelessHART network. Using some of the technologies and solutions already presented in the [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"} and [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"}, we set out to prove that the proposed architecture is able to monitor the network and the sensor nodes (hardware and firmware) operations, with low impact on IWSN resources.\n\nThis section starts by presenting the test scenario, comprising the industrial application, and the deployed hardware and firmware components. Secondly, the collected metrics and associated mechanisms are presented. Thirdly, data acquisition and processing of sensor node metrics are explained. Lastly, the impact on the network and sensor node resources is analysed and discussed.\n\n5.1. Test Scenario {#sec5dot1-sensors-18-03568}\n------------------\n\nThe proposed monitoring architecture was tested using a typical industrial application. The proof-of-concept scenario consisted of four sensor nodes and a WirelessHART gateway. On one hand, sensor nodes monitor the temperature of the industrial assets and send their readings to the gateway. On the other hand, the gateway controls all the traffic in the network, performing the role of a network and security manager, [Figure 2](#sensors-18-03568-f002){ref-type=\"fig\"}a. According to the proposed monitoring architecture, WirelessHART sensor nodes and the gateway are capable of sharing sensor node monitoring data and network metrics.\n\nThe firmware deployed in the sensor nodes can be divided into two distinct modules: the industrial application, and the sensor node monitoring agent. The industrial application is responsible for temperature sampling and for controlling the network operations (by sending control messages to the radio, such as network joining, network notifications, and service requests). On the other hand, the sensor node monitoring agent collects hardware and firmware metrics and sends them by using a specific network service.\n\nAs presented in the sensor node monitoring agent description, [Section 4.1](#sec4dot1-sensors-18-03568){ref-type=\"sec\"}, WirelessHART defines four types of service level agreements: block transfer service, maintenance service, periodic service, and event service. The proof-of-concept application running at the microcontroller requests the following services to the gateway: a publish service, for sending temperature data every minute; an event service, for sending alarms if the temperature rises above a certain threshold (the verification is made every 30 s); an additional publish service that is requested and deleted each time the sensor node needs to send monitoring data (every 15 min); and, lastly, a maintenance service, directly allocated by the radio that handles all the control messages between the radio and the network manager.\n\nIn terms of hardware, the sensor nodes are formed by three components, [Figure 2](#sensors-18-03568-f002){ref-type=\"fig\"}b: radio, microcontroller, and power supply. The radio handles the communication tasks with other network nodes and with the microcontroller (Linear DC9003A-C). The microcontroller runs the industrial application and the sensor node monitoring agent (Texas MSP430F5 launchpad, Texas Instruments, Dallas, TX, USA). Lastly, the power supply is formed by a pack of batteries and a DC/DC converter used to supply energy to the radio and the microcontroller. Two of the nodes use TPS62740EVM-186 buck converters (Texas Instruments), and the other two use TPS61291EVM-569 boost converters (Texas Instruments). By using these DC/DC converters, we can measure the energy expenditure of each sensor node almost for free and in real time.\n\nThe gateway (LTP5903CEN-WHR, Analog Devices, Norwood, MA, USA) controls all network operations, manages the network, performs security operations and, at the same time, connects the IWSN with the industrial network through an IP network. Additionally, an application running at the gateway allows the subscription of specific types of messages (e.g., application messages, sensor node monitoring data messages, and network report messages). This application performs the role of the gateway monitoring agent, collecting the reports identified in [Table 2](#sensors-18-03568-t002){ref-type=\"table\"}.\n\nFinally, the monitoring logger was developed as a python application that implements the log parser sub-component. When the monitoring logger application starts, a specific subscription is made to the network manager in order for it to receive the sensor node monitoring messages. After that, when the data arrives at the sub-component, the log parser converts the packets to JSON objects and stores them in a datastore.\n\nIn this proof-of-concept scenario, we only intend to measure the impact of the architecture on sensor node resources (memory, processing, energy) and on the network operation (overhead in relation to the sensor node traffic). Thus, we did not implement the management agents presented in the architecture (at the gateway and at the monitoring logger). Such agents were already presented and are freely available, as described in \\[[@B85-sensors-18-03568],[@B86-sensors-18-03568]\\], and, thus, their implementation in this proof-of-concept scenario would not provide relevant added-value. However, we should emphasize that the WirelessHART network metrics management models that would be required for such implementation were already created by us and are available for download in \\[[@B84-sensors-18-03568]\\].\n\n5.2. Collected Metrics {#sec5dot2-sensors-18-03568}\n----------------------\n\nThe architecture proposed in this paper allows the collection of hardware, firmware, and network metrics using agents deployed in sensor nodes and network gateways. In [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"}, some of the requirements were presented, as well as some monitoring techniques already explored by other researchers. Thus, to prove that this architecture has low impact on sensor nodes resources, our test scenario implements some of these techniques, enabling the collection of metrics from the hardware of the Texas MSP430F5 launchpad, from the industrial application, and from the WirelessHART network.\n\nCollecting hardware metrics was done using techniques that directly access the registers and counters of the Texas MSP430F5 launchpad. As this is a 16-bit microcontroller, the implemented techniques were the ones presented in \\[[@B80-sensors-18-03568],[@B81-sensors-18-03568]\\], comprising a processing metric, an energy metric, and a time metric. The processing metric gives the number of instructions executed by the microcontroller. Using switching regulators in combination with the technique presented in \\[[@B80-sensors-18-03568]\\], it was possible to have an energy metric that provides the energy spent by sensor nodes. Lastly, the time metric gives a time measurement in milliseconds.\n\nSpecifically, the metrics were implemented in the following way. Firstly, there are two possible approaches to implementing the processing metric in the Texas MSP430F5 launchpad: (1) by configuring pin P7.7 to output the MCLK clock and connecting it to a hardware counter; or (2) by using the same clock source and same frequency of the MCLK clock in SMCLK, and configuring a counter to count it. As the Texas MSP430F5 launchpad does not give us access to pin P7.7, the second approach was used, and TimerA was configured to be sourced by the SMCLK. Secondly, the energy metric was obtained by connecting the output of the inductor used in the switching regulator (used as a DC/DC converter, either the TPS62740EVM-186 or the TPS61291EVM-569) to the TimerA clock signal input (P1.6) of the microcontroller. Lastly, the time metric was obtained using the TimerB, configured to be sourced by the ACLK (being sourced by ACLK allows to count time even during sleep periods such as Low Power Mode 3 (LPM-3)).\n\nApart from hardware metrics, the architecture also enables firmware monitoring by using instrumentation techniques applied to the main application code. In order to prove that it is possible to collect some metrics with low impact on the sensor node resources, manual instrumentation of the code was performed, with the objective of obtaining a trace of function calls. Specifically, the function calls trace was obtained by assigning a specific ID to each function in the code. Additionally, the instrumentation code enabled each executed function to collect the function identifier, the function duration, the spent energy, and the processing load (by using the metrics collected from the hardware). The details of the data processing and transport over the network will be presented in the following sub-section. Lastly, after collecting the metrics concerning the sensor node state (hardware and firmware), the gateway monitoring agent collects the reports identified in [Table 2](#sensors-18-03568-t002){ref-type=\"table\"}, namely device health report, neighbour health list report, neighbour signal levels, and network alarms.\n\n5.3. Sensor Node Instrumentation and Monitoring Information Processing {#sec5dot3-sensors-18-03568}\n----------------------------------------------------------------------\n\nThe proof-of-concept implementation presented here aims at demonstrating the applicability of the proposed architecture to several software development scenarios. Specifically, the architecture should support OS-based applications as well as \"bare metal\" applications, as in the current case. With this in mind, the application was developed based on a function-queue-scheduling software architecture. Similar to what happens in OS-based approaches, the functions are added to a queue of function pointers, and called when appropriate. The scheduler is responsible for getting the next function in the queue, which will then be executed. On completion, the microcontroller is put in low power mode (LPM) to save energy.\n\nThe main topic of this section is, nevertheless, the sensor node monitoring agent and its interactions with the BSP, the main application, and the network ([Figure 3](#sensors-18-03568-f003){ref-type=\"fig\"}). The sensor node monitoring agent is a library developed in C/C++ that collects the sensor node state data, stores it, and triggers a periodic event to send the monitoring data over the network. By manually inserting a simple call to the sensor node monitoring agent at the beginning of each function to be monitored (see [Figure 3](#sensors-18-03568-f003){ref-type=\"fig\"}a (1)), it is possible to record the state of the sensor node's firmware and hardware. The monitoring information is then stored in a ring buffer ([Figure 3](#sensors-18-03568-f003){ref-type=\"fig\"}a (2)), until it is sent over the network. In addition, the hardware state data is collected directly by the sensor node monitoring agent library from the BSP ([Figure 3](#sensors-18-03568-f003){ref-type=\"fig\"}a (3)).\n\nLast but not least, as illustrated in the [Figure 3](#sensors-18-03568-f003){ref-type=\"fig\"}b, when there is sufficient data to fill an IEEE 802.15.4 data packet payload, the sensor node monitoring agent requests a publish service to send the data available in the ring buffer (4). If the network has enough resources to send the data at the requested rate, an authorization is received (5), and the sending process starts (6--7). In case the network does not have enough resources, the microcontroller will receive a service denial, and the sending of the log will be postponed. When no more data is available in the ring buffer, the sensor node monitoring agent requests the deletion of the service in order to free the network resources (8).\n\n5.4. Results {#sec5dot4-sensors-18-03568}\n------------\n\nIn order to show that the proposed architecture can be implemented in industrial scenarios with efficiency and low impact on resource consumption, some experiments were conducted, whose results are described in this sub-section.\n\nThe performed tests can be divided in two groups: tests that measure the impact on the sensor node resources, and tests that measure the network operation overhead introduced by the sensor node monitoring agent. It should be noted that, as presented in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"}, the industrial standards addressed in this paper already share network metrics to perform routing and other network related operations and, thus, these metrics do not add extra overhead to the network operation. Hence, the only network overhead is the one caused by sensor node monitoring agents.\n\nFor the first group of tests, the used hardware and software were, respectively, the Texas Instrument MSP-FET (with EnergyTrace technology support) and the Code Composer Studio (CCS, version 7, Texas Instruments) connected directly to the launchpad debug pins and bypassing the board debugger. Each test was performed during 30 s and repeated 10 times. In the second group of tests, the gateway monitoring agent was used to collect the impact of the traffic generated by the sensor node monitoring agent on the network, by collecting device health reports. In these tests, we collected health reports over a period of 10 h. The results presented in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"} show (a) the impact of the hardware techniques on the sensor node battery lifetime; (b) the overall impact of the sensor node monitoring agent on the sensor node lifetime; (c,d) the impact of the sensor node monitoring agent on the microcontroller processing and memory, respectively; and (e) the overhead caused by the transport of the sensor node monitoring data over the network.\n\nIn order to measure the impact of the hardware monitoring metrics on sensor node resources, a simplified version of the industrial application was developed with the objective of guaranteeing the independence of each added metric. In this test, the energy expenditure was measured by the MSP-FET. Additionally, the radio was disconnected in order to allow the easy detection of the comparatively small increase of energy expenditure caused by each hardware metric.\n\nAs shown in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}a, the energy spent by the MSP430 in LPM3 without any hardware metric enabled is around 4.1 $\\mathsf{\\mu}$A. Enabling the processing metric does not increase the energy spent by the microcontroller. In this LPM, the MCLK and the SMCLK are turned off and, consequently, the processing metric interrupt will not occur when the microcontroller is in LPM. On the other hand, the energy counter increases the microcontroller energy expenditure in LPM because, even in this LPM mode, the system consumes energy and, as a result, the interrupt associated with the energy counter mechanism will be executed (this interrupt is tied with an hardware counter that is sourced by an external clock). In this mode, when enabled, the microcontroller counter consumes 12.1 $\\mathsf{\\mu}$A ($\\sigma$ = 0.00 $\\mathsf{\\mu}$A). Lastly, the time counter, sourced by the ACLK consumes around 33.2 $\\mathsf{\\mu}$A ($\\sigma$ = 0.05 $\\mathsf{\\mu}$A).\n\nContrary to the significant increase in energy consumption in LPM associated with each metric, the energy spent by the complete solution in active mode with the radio on (acquisition plus transport of the monitoring data) only corresponds to a 3.5% increase ([Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}b) in relation to a non-monitoring situation. Using two AA NiMh batteries (1.2 v, 1900 mA), the lifetime of the sensor node is 658 days ($\\sigma$ = 6 days) without monitoring and 635 days ($\\sigma$ = 12 days) with the sensor node monitoring agent enabled.\n\nThe latency generated by adding the monitoring mechanisms to the sensor nodes firmware, as well as the microcontroller's memory taken by the monitoring mechanisms are also two relevant aspects in assessing a monitoring solution. The firmware graphs, in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}c,d, show the number of cycles executed by the microcontroller in three firmware functions, and the RAM and ROM requirements, respectively, with and without monitoring. In terms of processing overhead, the manual instrumentation performed in the code of each function inserts an average of 758 cycles ($\\sigma$ = 3 cycles), adding 93 $\\mathsf{\\mu}$s of latency at 8 Mhz clock. As can be seen in (c), the overhead is quite small, when compared to the cycles spent to execute the functions without monitoring, representing 5.1%, 1.2%, and 2.2%, for functions 1, 2, and 3, respectively. This test was performed manually, by measuring the difference of cycles at the beginning and at the end of the instrumentation functions, using the Code Composer and the MSP-FET. An analysis in terms of memory was also made, measuring the increase of flash memory and RAM. In what concerns flash memory, the main application without monitoring occupies 18.590 KB, as represented in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}d. Adding the sensor node monitoring agent and the instrumentation code, the application size increases to 34.659 KB. This represents an increase of 12.26% of the total capacity of the flash memory in this microcontroller version (131.072 KB). Lastly, as also presented in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}d, the main application without monitoring takes 1.631 KB of RAM. Adding the sensor node monitoring agent to the application increased RAM usage to 2.506 KB. When compared to the initial utilization, the sensor node monitoring agent only consumes an additional 8.54% of all the RAM available in the microcontroller (10.240 KB).\n\nThe network overhead caused by the sensor node monitoring agent was also measured. Using health reports collected every fifty minutes by the gateway monitoring agent, we were capable of measuring the traffic generated by a leaf node in normal operation and with the monitoring capabilities enabled. As can be seen in [Figure 4](#sensors-18-03568-f004){ref-type=\"fig\"}e, the main application sends 76 packets/hour ($\\sigma$ = 0 packets) on average. Most of this traffic is generated by the publish service that sends the temperate values every minute. Enabling the sensor node monitoring agent causes a traffic increase of 46%, sending an average of 165 packets per hour ($\\sigma$ = 9 packets), of which 89 packets are generated by the sensor node monitoring agent. In WirelessHART, the IEEE802.15.4 payload is limited to 94 bytes per packet. Thus, each packet sent by the sensor node monitoring agent is capable of carrying six log entries, each with log id, function id, function duration, processing metric, and energy metric.\n\nSumming up, the assessment of the proposed monitoring architecture, carried out through this proof-of-concept implementation, shows its low impact and high efficiency in what concerns sensor node resources and network. [Table 5](#sensors-18-03568-t005){ref-type=\"table\"} summarises the obtained results. By causing a mere 3.5% reduction in sensor nodes' battery lifetime, introducing a 93 $\\mathsf{\\mu}$s latency overhead in each function, occupying 12.26% of flash memory, and using 8.54% of RAM, the implementation enabled the monitoring of hardware and firmware in sensor nodes. Despite the increase of 46% in network traffic generated by sensor nodes, the proposed solution, here implemented for WirelessHART, used network resources in a smart way. The monitoring data was only sent when the network had resources to send it, requesting the service and deleting it each time the sensor node monitoring agent needed to send monitoring data. Using this approach, the sensor node monitoring agent only requested free network resources, not used by the main industrial application.\n\n6. Related Work {#sec6-sensors-18-03568}\n===============\n\nThe solution presented in this paper proposes a multi-domain architecture that is able to monitor the condition of IWSNs from different perspectives. To the best of our knowledge, the proposed architecture is the first of its kind. The related work reviewed in this section also addresses the multi-domain nature of the architecture, by analysing the contributions in several domains. In [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, part of this multi-domain analysis was already made, in a detailed way, when different techniques to monitor the network, firmware and hardware were presented. On the other hand, this related work section will provide a broader overview, by analysing solutions from the Industry 4.0 perspective.\n\nCurrent related work in the Industry 4.0 topic is extensive. In general, most of the existing monitoring architectures target the prevention of security attacks in ICS systems, although without taking into consideration the IWSN technologies analysed in this paper \\[[@B87-sensors-18-03568]\\]. Some of these solutions have been proposed because of the paradigm shift that occurs nowadays in the ICS core, with the introduction of IP networks. Additionally, more recent technologies such as Software Define Networks (SDN) \\[[@B88-sensors-18-03568]\\] and Cloud-based platforms \\[[@B89-sensors-18-03568]\\] are being considered in such architectures, due to the rich network metrics and flow control mechanisms that these technologies make available. However, this type of control and monitoring is specific for wired IP networks, that can work in a decentralised manner, and with different application requirements in mind (e.g., troughput). However, as presented before, applications in industrial proccess automation need deterministic latency, instead of pure throughput. This is, by way of example, the kind of requirement being considered by the Time-Sensitive Task Group (TSN). The group analyses the possibility of deterministic services in IEEE 802 based networks. As consequence, technologies like SDN cannot be applied to IWSN standards like WirelessHART, ISA100.11a, WIA-PA and ZigBee, at least inside the wireless network domain, at this stage, due to their centralized and deterministic nature. Solutions like \\[[@B88-sensors-18-03568]\\] can only be used at gateway level, by virtualizing some of the non-deterministic features and roles in another device (e.g., the security manager in WirelessHART). Thus, in this section, only solutions proposed for the industrial standards addressed will be considered.\n\nFrom the first perspective, the authors of \\[[@B90-sensors-18-03568],[@B91-sensors-18-03568]\\] propose techniques that improve the reliability of the network at routing schedule-level: by using a finite-state Markov model \\[[@B90-sensors-18-03568]\\], and by specifying scheduling emergency and recovery mechanisms when a path-down alarm occurs \\[[@B91-sensors-18-03568]\\]. On the other hand, the work presented in \\[[@B92-sensors-18-03568],[@B93-sensors-18-03568],[@B94-sensors-18-03568]\\] proposes several techniques and platforms that can be used to monitor some IWSN. In \\[[@B92-sensors-18-03568]\\], the authors develop a passive monitoring tool for evaluation of the deployed WirelessHART networks using passive sniffers deployed in the network area. Additionally, in \\[[@B93-sensors-18-03568]\\], the authors propose an hybrid monitoring technique to monitor wireless and wired industrial technologies in the scope of the project HiFlecs. Lastly, in \\[[@B94-sensors-18-03568]\\], the authors present a debugging tool that is able to collect in the SensorLab2 environment routing topology information and end-to-end performance, by debugging the information using UART ports and then converting it to TCP/IP messages. Finally, some authors also show some concerns in the lack of integration solutions between the industrial standards \\[[@B95-sensors-18-03568],[@B96-sensors-18-03568],[@B97-sensors-18-03568]\\]. In \\[[@B95-sensors-18-03568]\\] the authors presented how the MQTT protocol can be used together with ISA S5.1 and ISA95/88 standards to represent the monitoring and control in topics using the URI scheme. Lastly, in \\[[@B96-sensors-18-03568]\\] the authors presents how the WirelessHART standard and the ISA100.11a standard can be integrated using the Modbus protocol and a proprietary software called Wonderware InTouch.\n\nWhen comparing the reliability solutions already proposed by other researchers with the architecture proposed in this paper, some differences appear. Most of the work performed in this field is not taking into consideration the main components that may produce faults in an IWSN (firmware, hardware and network). Additionally, most part of the work focus in a specific problem like \\[[@B90-sensors-18-03568],[@B91-sensors-18-03568]\\], and do not address the needs of monitoring several process automation networks at the same point. Additionally, some of the works that present solutions to monitor these standards \\[[@B92-sensors-18-03568],[@B93-sensors-18-03568],[@B94-sensors-18-03568]\\] keep proposing passive monitoring tools that need extra networks to be installed, adding extra costs to a technology designed to be low-cost. Lastly, most of the work \\[[@B95-sensors-18-03568],[@B96-sensors-18-03568],[@B97-sensors-18-03568]\\] in the field of integration focus their effort in the representation of the sensor data, and any of them try to explore the use of management protocols like LWM2M and COMI (protocols specific for constrained devices) to represent and make available the monitoring data.\n\n7. Conclusions {#sec7-sensors-18-03568}\n==============\n\nThe need for more dynamic and capable solutions in the Industry 4.0 field has contributed to an increase in the use of IWSN technologies in industrial applications over the last few years. As is widely recognised, WSN technologies have a considerable number of advantages over traditional solutions, although they are not free from drawbacks, namely in what concerns reliability, due to their resource-constrained nature. Nevertheless, hardware and firmware of sensor nodes are becoming more mature and capable. On one hand, the hardware is becoming more efficient and, at the same time, hardware manufactures are introducing new features in their architectures that allow the collection of state metrics. On the other hand, the adoption of operating systems in WSNs is also growing, enabling the development of firmware architectures and tools capable of collecting firmware state metrics. In a broader context, at the same time, ICS networks are shifting from networks disconnected from public infrastructures to systems connected to the Internet. This transformation is also exposing these systems to a new set of security threats, and new types of IDS systems are needed. To keep up with all of these changes and, at the same time, improve IWSN reliability, new network monitoring architectures are needed, more specifically, architectures that are not only capable of monitoring IWSNs from a network perspective, but also from the hardware and firmware perspective.\n\nSuch a multi-domain IWSN monitoring architecture was proposed in this paper, considering the main process automation IWSN standards. In order to back up the proposal, a comprehensive revision of the state of the art was needed, in order to introduce the different concepts in each domain. This was done in [Section 2](#sec2-sensors-18-03568){ref-type=\"sec\"} and [Section 3](#sec3-sensors-18-03568){ref-type=\"sec\"}, by surveying WSN diagnostic tools and the relevant industrial technologies, and by presenting several network, hardware and firmware metrics as well as the techniques used to collect and transport them over wireless networks. Additionally, being a multi-network monitoring architecture, several management models were also presented. Lastly, the architecture was presented in [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"}, and validated in a real testbed scenario, in [Section 5](#sec5-sensors-18-03568){ref-type=\"sec\"}. In [Section 4](#sec4-sensors-18-03568){ref-type=\"sec\"}, when the architectural components were described, some possible solutions were presented in order to provide some guidelines for implementation. Nevertheless, as shown by our proof-of-concept implementation, the architecture can be used to perform the monitoring of a variety of sensor and network-wide parameters, in a way that does not increase the cost of node manufacturing, does not have a significant impact on sensor node resources, and does not steal network bandwidth from the main application.\n\nBy proposing such architecture and by analysing relevant techniques and technologies, more questions and future research directions arrose. Thus, as future work, we plan to further explore the proposed monitoring architecture by extending and refining the existing implementation. In the firmware domain, we aim at extending the architecture in order to provide support to more advanced instrumentation techniques, such as those proposed in \\[[@B52-sensors-18-03568],[@B53-sensors-18-03568],[@B74-sensors-18-03568]\\]. On the other hand, in what concerns hardware monitoring, we would like to extend the compatibility to the ARM architecture, by supporting some of the metrics available in the CoreSight technology. From the network perspective, we would like explore new ways of reducing the overhead created over the network. Lastly, we plan to extend the implementation to other technologies (e.g., ZigBee, WIA-PA and ISA100.11a). Finally, by using the proposed architecture, we also plan to explore the architecture's data collection potential for building a diagnostic tool, and to share the acquired datasets with the community.\n\nReferences\n\nWe would like to show our gratitude to Thomas Watteyne and Jonathan Simon, for the support given through the dustcloud forum in the development of this proof-of-concept scenario.\n\nAll authors contributed to the conceptualisation and discussed the paper; D.R. wrote the original draft, developed and performed the architecture tests; A.R. contributed to the literature, participated in the implementation and tests, and provided supervision; S.S. contributed to the management technologies section; J.S.S. and F.B. provided supervision.\n\nThe work presented in this paper was partially carried out in the scope of the SOCIALITE Project (PTDC/EEI-SCR/2072/2014) and the MOBIWISE Project (P2020 SAICTPAC/0011/2015), both co-financed by COMPETE 2020, Portugal 2020-Operational Program for Competitiveness and Internationalization (POCI), European Union's ERDF (European Regional Development Fund) and the Portuguese Foundation for Science and Technology (FCT). Partial support was also given by Escuela Polit\u00e9cnica Nacional de Ecuador, and SENESCYT-Secretar\u00eda Nacional de Educaci\u00f3n Superior, Ciencia, Tecnolog\u00eda e Innovac\u00edon de Ecuador.\n\nThe authors declare no conflict of interest.\n\n![Proposed monitoring architecture.](sensors-18-03568-g001){#sensors-18-03568-f001}\n\n![Validation of the monitoring architecture using a WirelessHART testbed. On the left (**a**) the Gateway and four sensor nodes, on the right (**b**) the sensor node components.](sensors-18-03568-g002){#sensors-18-03568-f002}\n\n![On the left, (**a**) the application architecture and the sensor node monitoring agent acquiring the state information. On the right, (**b**) the request of the WirelessHART publish service.](sensors-18-03568-g003){#sensors-18-03568-f003}\n\n![Obtained results.](sensors-18-03568-g004){#sensors-18-03568-f004}\n\nsensors-18-03568-t001_Table 1\n\n###### \n\nSummary of main features, adapted from \\[[@B9-sensors-18-03568]\\] and \\[[@B18-sensors-18-03568]\\].\n\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------\n  Layer   ZigBee                                 WirelessHART                      ISA100.11a                              WIA-PA\n  ------- -------------------------------------- --------------------------------- --------------------------------------- ----------------------------------\n  APP     Object-oriented;\\                      Command-oriented;\\                Object-oriented;\\                       Object-oriented;\\\n          Profile Defined Protocol               HART protocol                     Native Protocol;\\                       Profibus/FF/HART Protocol;\\\n                                                                                   Multi-wired field bus protocols         Virtual Device\n\n  APS     Discovery of New Device;\\              \\-                                Basic Tunnelling;\\                      Data Communication\\\n          Binding;\\                                                                Smart Tunnelling                        and Management Services\n          Fragmentation/Reassembly                                                                                         \n\n  TP      \\-                                     Block Data Transfer;\\             Optional Security Features;\\            \\-\n                                                 Reliable Stream Transport;\\       Connectionless Services                 \n                                                 Convergence                                                               \n\n  NWK     Tree Routing/AODV Routing;\\            Graph/Source/Superframe Routing   Addressing (6LowPAN);\\                  Addressing;\\\n          Address Assignment;\\                                                     Routing Address Translation;\\           Static Routing;\\\n          Network Joining/Disjoining                                               Fragmentation/Reassembly                Fragmentation/Reassembly\n\n  DL      \\-                                     Slot Timing Communication;\\       Grapgh/Source/Superframe Routing;\\      Frequency Hopping;\\\n                                                 Time Synched TDMA/CSMA;\\          Slot Timing Communication; Duocast\\     Aggregation and Disaggregation;\\\n                                                 Channel Hopping                   Transaction; Time Synched TDMA/CSMA;\\   Time Synchronization\n                                                                                   Channel Hopping                         \n\n  MAC     IEEE 802.15.4\\                         IEEE 802.15.4\\                    IEEE 802.15.4\\                          IEEE 802.15.4\\\n          Mac Layer                              Mac Layer\\                        Mac Layer\\                              Mac Layer\n                                                 (partially implemented)           (partially implemented)                 \n\n  PHY     IEEE 802.15.4 PHY\\                     IEEE 802.15.4 PHY\\                IEEE 802.15.4 PHY\\                      IEEE 802.15.4 PHY\\\n          868 M/915 M/2.4 GHz Radio\\             2.4 GHz Radio\\                    2.4 GHz Radio\\                          2.4 GHz Radio\\\n          Data rate: 20 Kb/; 40 Kb/s; 250 Kb/s   Data rate 250 Kb/s                Data rate 250 Kb/s                      Data rate 250 Kb/s\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------\n\nsensors-18-03568-t002_Table 2\n\n###### \n\nIWSN Network metrics.\n\n  Standard                           Network Reports                                                                                                                                                                                                                                                          Metrics                                                                                                                                                                                                                                                                                                                                                                                                              Report Type   Visibility\n  ---------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------- ------------------------\n  ZigBeePRO                          Network Status                                                                                                                                                                                                                                                           No route available; Tree link failure; Non-tree link failure; Low battery level; No routing capacity; No indirect capacity; Indirect transaction expiry; Target device unavailable; Target address unallocated; Parent link failure; Validate route; Source route failure; Many-to-one route failure; Address conflict; PAN identifier failure; Network address update; Bad frame counter; Bad key sequence number   Event         Coordinator or Routers\n  Link Status                        Neighbour network address; Incoming cost; Outgoing cost                                                                                                                                                                                                                  Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Network Report                     Radio channel condition; PAN ID conflict                                                                                                                                                                                                                                 Event /Periodic                                                                                                                                                                                                                                                                                                                                                                                                                    \n  WirelessHART                       Device Health                                                                                                                                                                                                                                                            Packets generated by device; Packets terminated by device; DL mic failures; NWK mic failures; Power status; CRC errors                                                                                                                                                                                                                                                                                               Periodic      Network Manager\n  Neighbour Health List              Total number of neighbours; Nickname of Neighbour; Mean RSL; Packets transmitted to the neighbour; Packets received from the neighbour; Failed transmissions                                                                                                             Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Neighbour Signal Levels            Total number of neighbours; Nickname of neighbour; RSL of neighbour in DB                                                                                                                                                                                                Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Alarms                             Path Down (Nickname of the neighbour); Source Route Failed (Nickname of the neighbour and NWK MIC from the NPDU failed routing); Graph Route Failed (Graph Id of the failed route)                                                                                       Event                                                                                                                                                                                                                                                                                                                                                                                                                              \n  ISA100.11a                         Connectivity Alert per Channel                                                                                                                                                                                                                                           Attempted unicast transitions for all channels; Percentage of time transmissions on channel x did not receive an ACK); Percentage of time transmissions on channel x aborted due to CCA                                                                                                                                                                                                                              Periodic      System Manager\n  Connectivity Alert per Neighbour   RSSI level; RSQI level; Valid packets received by the neighbour; Successful unicast transmissions to the neighbour; Unsuccessful unicast transmission; Number of unicast transmissions aborted (by CCA); Number of NACKS received; Standard deviation clock\u00a0correction   Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Neighbour Discovery Alert          Total number of neighbours; Neighbour address; Neighbour RSSI; Neighbour RSQI                                                                                                                                                                                            Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  WIA-PA                             Path Failure report                                                                                                                                                                                                                                                      Route ID                                                                                                                                                                                                                                                                                                                                                                                                             Periodic      Network Gateway\n  Device Status report               Number of sent packets; Number of received packets; Number of MAC layer mic failures detected; Battery level; Number of restarts; Uptime since last restart                                                                                                              Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Channel Status report              Channel Id; Neighbour device address; LQI; Channel packet loss rate; Channel retransmission count                                                                                                                                                                        Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n  Neighbour Reports                  Neighbour address; Backoff counter; BackoffExponent; Last time communicated; Average RSL; Packets transmitted; Ack packets; Packets received; Broadcast packets                                                                                                          Periodic                                                                                                                                                                                                                                                                                                                                                                                                                           \n\nsensors-18-03568-t003_Table 3\n\n###### \n\nNetwork management protocols revision.\n\n  Standard   Modeling Language   Supported Operations                                       Resource Constrained Devices Support   Notifications Support   Used Protocols\n  ---------- ------------------- ---------------------------------------------------------- -------------------------------------- ----------------------- -----------------------\n  SNMP       SMIv2               Monitoring/Configuration (not\u00a0used to configure devices)   No                                     Yes                     UDP\n  NETCONF    YANG                Monitoring/Configuration                                   No                                     No                      SSH/SSL/HTTP\n  RESTCONF   YANG                Monitoring/Configuration                                   Yes                                    Yes                     HTTP/TLS/TCP\n  COMI       YANG/SMIv2          Monitoring/Configuration                                   Yes                                    Yes                     CoAP/DTLS/UDP\n  LWM2M      XML/YANG            Monitoring/Configuration/ Application Management           Yes                                    Yes                     CoAP/DTLS/UDP and SMS\n\nsensors-18-03568-t004_Table 4\n\n###### \n\nNetwork transport revision.\n\n  Standard       Time Standard   Fragmentation and Reassembly   Security                                                                                                                                     Service Level Agreement (Available Options) Support          Service Level Agreement (Recommended) Protocols\n  -------------- --------------- ------------------------------ -------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------ -------------------------------------------------\n  ZigBee         \\-              Application Sub-Layer          Symmetric encryption available in all standards and supported by IEEE802.15.4 hardware                                                       \\-                                                           \\-\n  WirelessHART   UTC             Block transfer service         Publish, Event, Maintenance, and Block Transfer service                                                                                      Publish service                                              \n  ISA100.11a     TAI             Network Layer                  Contract type: periodic and non-periodic Contract priority: best-effort queued, real-time sequential, real-time buffer and network control   Contract type: Non-periodic Contract priority: Best-effort   \n  WIA-PA         UTC             Network Layer                  Publish/subscribe VCRs, source/sink VCRs, and client/server VCRs                                                                             Source/sink VCR or Client/server VCR                         \n\nsensors-18-03568-t005_Table 5\n\n###### \n\nEvaluation results overview.\n\n  Metric                                        Value\n  --------------------------------------------- --------------------\n  Node lifetime                                 \u22123.5%\n  Overhead latency                              93 $\\mathsf{\\mu}$s\n  Flash overhead (each instrumented function)   +13 bytes\n  RAM overhead (each instrumented function)     +14 bytes\n  Flash usage                                   \\* 12.26%\n  RAM usage                                     \\* 8.54%\n  Network traffic                               +46%\n\n\\* when compared with the max capacity available.\n"}
+{"text": "Introduction {#j_jvetres-2018-0028_s_001}\n============\n\nPyrrolizidine alkaloids (PAs) are synthesised by plants as their secondary metabolites and are considered to be one of the most widespread toxins of natural origin ([@j_jvetres-2018-0028_ref_001], [@j_jvetres-2018-0028_ref_023]). PAs have gained attention in the recent years due to their hepatotoxic, carcinogenic, genotoxic, and pneumotoxic properties ([@j_jvetres-2018-0028_ref_017]).\n\nMore than 6,000 plant species produce alkaloids and more than 660 PAs have been identified so far ([@j_jvetres-2018-0028_ref_005], [@j_jvetres-2018-0028_ref_015]). PA-producing plants are introduced or are native species which are considered invasive and noxious weeds. They often infest open ranges and fields leading to contamination of crops and feeds ([@j_jvetres-2018-0028_ref_012], [@j_jvetres-2018-0028_ref_030]).\n\nPAs are esters of amino alcohols, consisting of two basic structure elements: a pyrrolizidine-derived moiety necine and different nono- or dicarboxylic acids. The necine base consists of two fused fivemembered rings with a nitrogen atom at the bridgehead, which can be saturated or contain a double bond in the 1,2-position ([@j_jvetres-2018-0028_ref_015]). PAs occur as free base/tertiary forms and as *N*-oxides which are characterised by different reactivity and solubility leading to different toxicity ([@j_jvetres-2018-0028_ref_003]). Neither form is significantly toxic to humans or animals *per se*. Their hazard arises when they are converted into pyrrolic metabolites in the liver. These highly reactive electrophilic alkylating agents are capable of binding strongly to nucleophilic centres in tissues or cross-linking DNA, leading to hepatotoxicity or carcinogenicity ([@j_jvetres-2018-0028_ref_014]).\n\nCases of intoxication have been noted worldwide and toxicity effects of some PAs have been well documented in numerous publications ([@j_jvetres-2018-0028_ref_010], [@j_jvetres-2018-0028_ref_013], [@j_jvetres-2018-0028_ref_016], [@j_jvetres-2018-0028_ref_024]). Besides acute intoxications, pyrrolizidine alkaloids are considered to cause delayed progressive chronic effects, including hepatic cirrhosis following even short-term low-level exposure ([@j_jvetres-2018-0028_ref_014]). Chronically poisoned animals often do not show any clinical signs and their serum biochemistry may be normal for several months or even years after PAs ingestion. Nevertheless, the lasting hepatocellular damage leads to increased hepatocyte death with subsequent inflammation, fibrosis, and ultimately cirrhosis ([@j_jvetres-2018-0028_ref_030]). In an analogous way in humans, even single dietary exposure to dehydro PAs can lead to silently progressing chronic diseases which are unlikely to be attributed to dehydro PAs in food ([@j_jvetres-2018-0028_ref_012]).\n\nAnimals are exposed to PAs mainly through feed contaminated with plants containing these compounds, though due to the bitter taste, animals usually avoid direct consumption of PA-producing plants ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_015]). It has also been proved that PAs can be transferred to food of animal origin such as milk, eggs, and meat, if animals consume contaminated feed ([@j_jvetres-2018-0028_ref_008], [@j_jvetres-2018-0028_ref_019], [@j_jvetres-2018-0028_ref_027]).\n\nOnly a few studies on the contamination of feed materials with PAs have been published hitherto ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_018], [@j_jvetres-2018-0028_ref_020], [@j_jvetres-2018-0028_ref_025]). The problem of PA-contaminated feeds originating from Polish territory has not been well recognised. Only one paper on this subject has been published ([@j_jvetres-2018-0028_ref_021]). In the previous study, PAs were determined with the gas chromatography mass spectrometry method (GC-MS) using the sum parameter approach. This approach allows PAs determination without the necessity of having analytical standards of all the particular PAs. However, the drawback of this approach is that 1,2-unsaturated alkaloids are reduced to their common base structures retronecine and heliotridine, and all information concerning the whole structure is lost. Where that information retained, it could lead to indication which of the alkaloids are present in the analysed material. For this reason, a new liquid chromatography method for determination of selected alkaloids has been developed.\n\nTaking into consideration results reported by other authors ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_018], [@j_jvetres-2018-0028_ref_020], [@j_jvetres-2018-0028_ref_025]), the European Food Safety Authority (EFSA) recommendations, and PA-producing plants occurring in Poland, the following alkaloids were selected for the study: senecionine-type PAs including jacobine, retrorsine, senecionine, and seneciphylline; lycopsamine-type PAs including lycopsamine, intermedine, and echimidine; heliotrine-type PAs including heliotrine and lasiocarpine; and senkirkine ([@j_jvetres-2018-0028_ref_015]). The final determined concentration of PAs also includes the *N*-oxides contribution, as they are reduced with zinc dust and are measured as free base form.\n\nMaterial and Methods {#j_jvetres-2018-0028_s_002}\n====================\n\n**Chemicals and reagents** Sulphuric acid (95%) was from Chempur (Poland), 25% ammonia solution was from POCH (Poland), and formic acid and zinc dust were from Sigma-Aldrich (USA). Ethyl acetate and triethylamine were from Merck (Germany), and methanol and acetonitrile were from J.T. Baker (the Netherlands). Water was purified with the Milli-Q water purification system (Millipore, USA). Analytical standards of intermedine, lycopsamine, jacobine, retrorsine, heliotrine, seneciphylline, senecionine, echimidine, senkirkine, and lasiocarpine were purchased from PhytoLab (Germany). Solid phase extraction (SPE) Strata SCX and polymeric Strata XC cartridges were supplied by Phenomenex (USA). Oasis MCX mixed-mode polymeric cartridges were from Waters (USA) and Bond Elut Plexa PCX and HF Bond Elut-SCX cartridges from Agilent (USA). The cartridges were of 500 mg bed weight and 6 mL volume.\n\n**Standard solutions** Stock standard solutions of intermedine, lycopsamine, jacobine, retrorsine, heliotrine, seneciphylline, senecionine, echimidine, senkirkine, and lasiocarpine were prepared at a concentration of 1 mg mL^-1^ in methanol and stored at \u221218\u00b0C. A mixed standard solution of 1 \u03bcg mL^-1^ was prepared by mixing the appropriate volume of each stock standard solution and subsequent serial dilution. The mixed standard solution was stored between 2 and 4\u00b0C.\n\n**Feed samples** Analysed feed materials included grass and alfalfa silage (27 and 1 sample, respectively) and hay (4 samples). All analysed samples were available from earlier microbiological studies.\n\n**Sample preparation** Silage was dried in an oven and then ground and homogenised as well as hay. Homogenised feed material (5 g) was weighed into 50 mL polypropylene tubes and 40 mL of 0.05 M sulphuric acid was added. The samples were extracted for 2 h on a horizontal shaker. After extraction, the samples were centrifuged and filtered through cellulose filters. About 1 g of zinc dust was added to filtered extracts to reduce *N*-oxides and the samples were left overnight. On the next day, the samples were slowly shaken for 0.5 h, subsequently centrifuged (4,000 *g*, 10 min), and 15 mL of the extract was purified with SPE. Strata SCX cartridges were preconditioned with 9 mL of methanol and 9 mL of 0.05 M H^2^SO^4^ acid. After sample application, the cartridges were washed with 12 mL of water and 12 mL of methanol and vacuum dried for 2 min, then 6 mL of ethyl acetate was added. For the elution of PAs, 12 mL of solvent mixture consisting of ethyl acetate, methanol, acetonitrile, ammonia, and triethylamine (8:1:1:0.1:0.1 v/v) was used. After evaporation at 40\u00b0C in a nitrogen stream the residues were reconstituted in 0.2 mL of water and 0.2 mL of methanol and passed through 0.2 \u03bcm PVDF syringe filters into chromatographic vials.\n\n**Instrumental parameters** HP 1200 Series separation modules from Agilent Technologies (USA) were used for the analysis. These modules consisted of a degasser system, binary pump, automatic injector, and column thermostat. A single quadrupole mass spectrometry detector (6140 Agilent Technologies) was also employed. The separation of the alkaloids was carried out on a Gemini 3 \u03bcm NX-C18, 150 mm \u00d7 4.6 mm column (Phenomenex, USA) coupled with C18 guard column (Phenomenex). The column was thermostatted at 30\u00b0C. The mobile phase containing 0.2% formic acid in water (A) and a mixture of methanol and acetonitrile (1:1, v/v) (B) was used in a gradient mode as follows: 0--2 min, 5.5% B; 2--8 min, 12% B; 8--11 min, 20% B; 11--12 min, 30% B; 12--15 min, 40% B; 17--16 min, 70% B; 16--17 min, 85% B; and 17--23 min, 5.5% B. The flow rate was 0.6 mL min^-1^ and the injection volume was 5 \u03bcL. Electrospray ionisation (ESI) was set in a positive mode, the capillary voltage was set at 2,000 V, nebulizer pressure was 35 psi, and drying gas flow and temperature were 11.0 L min^-1^ and 300\u00b0C, respectively. Fragmentor voltage was set at 100 V for all monitored alkaloids. Selected ion monitoring was used for the detection, and the protonated molecular (M + H)^+^ ions (m/z) monitored are listed in [Table 1](#j_jvetres-2018-0028_tab_001){ref-type=\"table\"}.\n\n###### \n\nSelected (m/z) ions and retention times of the monitored alkaloids. Determination coefficients obtained for matrix calibration curves in a concentration range corresponding to 0 -- 100 \u03bcg kg^-1^ for all analysed alkaloids, and results of matrix effect evaluation. PA -- pyrrolizidine alkaloid, RT -- retention time, R^2^-- coefficient of determination, ME -- matrix effect\n\n  PA               m/z     RT      R^2^    ME (%)\n  ---------------- ------- ------- ------- --------\n  Intermedine      300.1   6.14    0.999   99.9\n  Lycopsamine      300.0   6.37    0.999   108.6\n  Jacobine         352.1   6.88    0.999   93.7\n  Retrorsine       352.1   8.29    0.996   98.3\n  Heliotrine       314.1   8.58    0.999   102.2\n  Seneciphylline   334.1   9.11    0.998   98.9\n  Senecionine      336.1   11.09   0.998   100.3\n  Echimidine       398.0   13.11   0.997   100.6\n  Senkirkine       366.1   13.37   0.997   89.9\n  Lasiocarpine     412.1   15.03   0.995   101.4\n\n**Identification and quantification** Identification was made by comparison with the relevant reference standard by the retention time and the protonated molecular ion (M + H)^+^. Quantification was achieved using calibration curves prepared by adding the mixed reference standard solution to blank matrix in appropriate amounts before the extraction procedure. Calibration curves were constructed by plotting the peak area *versus* the alkaloid concentrations.\n\n**Validation of the method** In-house validation was conducted according to SANTE/11945/2015 guidelines ([@j_jvetres-2018-0028_ref_028]). Parameters such as linearity, recovery, repeatability, reproducibility, specificity, limit of quantification (LOQ), matrix effect, robustness, and PA stability were evaluated. Tested feed with no determined content of the target compounds was used as a blank matrix.\n\n**Linearity and selectivity** The method's linearity was evaluated using matrix calibration curves. All curves were composed by blank samples fortified with the standards before the extraction procedure and analysed in triplicate. Feed samples were spiked with appropriate concentrations corresponding to 0, 5, 10, 20, 30, 50, and 100 \u03bcg kg^-1^.\n\nTo determine the selectivity of the method a set of blank feed samples was analysed in order to check the possible presence of endogenous interferences in retention times of the monitored alkaloids.\n\n**Recovery and precision** Recovery and precision were determined by the analysis of feed blanks spiked at three different concentrations, corresponding to 5, 20, and 100 \u03bcg kg^-1^ (a set of six samples for each level). The samples were analysed under the same conditions. The repeatability was expressed as relative standard deviation of the results, RSD (%). To determine the reproducibility, also expressed as RSD (%), other two sets of feed samples were spiked at 5, 20, and 100 \u03bcg kg^-1^, and analysed on different days with the same instrument.\n\n**Quantification limit and matrix effect** According to the SANTE document ([@j_jvetres-2018-0028_ref_028]) the lowest spike level meeting the method performance criteria for trueness and precision was assumed to be the limit of quantification. Matrix effects are generally recognised as a suppression or enhancement of the analytical signal due to co-eluting matrix components ([@j_jvetres-2018-0028_ref_009]). This effect was evaluated by comparison of the slopes of matrix matched calibration curves with slopes of calibration curves in a solvent.\n\n**Robustness and stability test** The Youden procedure was used to determine the robustness of the method. Blank feeds spiked at the concentration of 20 \u03bcg kg^-1^ were analysed in order to assess the influence of seven parameters which were slightly changed. The effects of change of percentage of sulphuric acid (0.05 M, 0.1 M), volume of elution mixture (12 mL, 10 mL), evaporation temperature (40\u00baC, 45\u00baC), percentage of formic acid in the mobile phase (0.2%, 0.18%), injection volume (5 \u03bcL, 4.5 \u03bcL), thermostat temperature (30\u00b0C, 27\u00b0C), and flow rate (0.6 mL min^-1^, 0.63 mL min^-1^) were evaluated. Student's *t-*test was used to determine the impact of changes in individual parameters on the results of the analysis.\n\nA stability test was performed on SPE purified extracts. Blank feeds were contaminated at the level of 20 \u03bcg kg^-1^ before the extraction procedure. Dry extracts were stored at \u221218\u00baC, 4\u00baC, and 20\u00baC. The extracts were subsequently analysed on days 1, 2, 4, 8, 14, and 30.\n\nResults {#j_jvetres-2018-0028_s_003}\n=======\n\n**Validation results** The developed method was validated in-house according to SANTE/11945/2015 guidelines ([@j_jvetres-2018-0028_ref_028]). Matrix calibration curves were linear in the concentration range corresponding to 0--100 \u03bcg kg^-1^ for all analysed alkaloids. The linearity was confirmed as coefficients of determination were higher than 0.99 for all monitored PAs ([Table 1](#j_jvetres-2018-0028_tab_001){ref-type=\"table\"}).\n\nAll validation parameters ([Table 2](#j_jvetres-2018-0028_tab_002){ref-type=\"table\"}) were determined on the basis of the analysis of sets of blank feed samples spiked at the levels corresponding to concentrations 5, 20, and 100 \u03bcg kg^-1^. The recovery was from 84.1% to 112.9%. The repeatability expressed as RSD (%) ranged from 3.0% to 13.6%, and reproducibility was from 4.8% to 18.9%.\n\n###### \n\nValidation parameters evaluated for ten alkaloids in feed matrix\n\n                   Concentration level (\u03bcg kg^-1^)   Concentration level (\u03bcg kg^-1^)   Concentration level (\u03bcg kg^-1^)                                      \n  ---------------- --------------------------------- --------------------------------- --------------------------------- ------ ------ ------ ------ ------ ------\n  Intermedine      99.5                              99.1                              89.9                              11.7   13.4   11.2   18.4   16.8   13.9\n  Lycopsamine      89.1                              97.9                              87.6                              9.6    11.6   12.8   18.3   10.6   16.8\n  Jacobine         99.8                              102.5                             92.3                              9.3    13.6   9.5    10.8   13.7   9.8\n  Retrorsine       112.9                             84.1                              97.1                              10.1   10.9   4.0    16.3   18.9   4.8\n  Heliotrine       100.5                             93.9                              92.9                              5.7    11.8   5.2    16.3   14.1   5.3\n  Seneciphylline   102.3                             100.2                             97.1                              9.5    12.2   4.3    18.8   12.7   5.2\n  Senecionine      100.6                             104.6                             94.7                              8.7    12.9   4.5    14.6   13.4   5.5\n  Echimidine       110.8                             109.5                             95.0                              7.4    8.9    7.0    13.9   11.8   5.5\n  Senkirkine       94.2                              106.4                             97.5                              3.0    5.6    5.9    18.0   9.9    8.0\n  Lasiocarpine     100.5                             99.4                              97.2                              6.0    7.9    3.8    11.6   15.1   6.3\n\nThe robustness test revealed that chosen variables do not affect the analysis as the calculated standard deviation was not significantly higher than the standard deviation of the method carried out under within-laboratory reproducibility conditions. The results were also compared with the use of Student's *t*-test to determine the impact of changes in individual parameters on the result of the analysis and no significant statistical differences were determined. Only a small retention time shift was observed when the flow rate was changed to 0.63 mL min^-1^, or the thermostat temperature was decreased to 27\u00b0C.\n\nOn the basis of the performed stability test, it can be stated that extracts of feed samples can be stored at \u221218\u00b0C and at 4\u00b0C for about a week without a significant change in PAs concentrations ([Fig. 1](#j_jvetres-2018-0028_fig_001){ref-type=\"fig\"}, data shown only for \u221218\u00b0C).\n\n![Stability of ten PAs in dry SPE-purified extracts of feed matrix, stored at \u221218\u00b0C](jvetres-62-183-g001){#j_jvetres-2018-0028_fig_001}\n\nThe method is selective as no interfering peaks were determined in the retention times of the monitored alkaloids ([Fig. 2c](#j_jvetres-2018-0028_fig_002){ref-type=\"fig\"}). The LOQ was established at 5 \u03bcg kg^-1^ for the individual alkaloid, as it was the lowest validation level meeting the criteria of SANTE document for trueness and precision.\n\n![SIM chromatograms obtained for: a) 3% ammonia in methanol SPE elution (PAs concentration 20 \u03bcg kg^-1^); b) new solvent mixture consisting of ethyl acetate, methanol, acetonitrile, ammonia, and triethylamine (8:1:1:0.1:0.1, v/v) SPE elution (PAs concentration 5 \u03bcg kg^-1^); c) blank feed sample. Intermedine (Int), lycopsamine (Lyc), jacobine (Jcb), retrorsine (Ret), heliotrine (Hel), seneciphylline (Snc), senecionine (Sen), echimidine (Ech), senkirkine (Skn), and lasiocarpine (Las)](jvetres-62-183-g002){#j_jvetres-2018-0028_fig_002}\n\nThe matrix effect was in the range \u00b120% ([Table 1](#j_jvetres-2018-0028_tab_001){ref-type=\"table\"}) for all compounds which complies with the SANTE document.\n\n**Real sample application** The developed method was applied to the analysis of 32 feed samples. Analysed feed materials included grass and alfalfa silage and hay. All analysed samples were available from earlier microbiological studies.\n\nAmong 32 analysed feed samples, 13 (40.6%) contained at least one of the monitored PAs (2 positive hay samples and 11 positive silage samples). In two samples PAs concentrations were below LOQ value ([Fig. 3](#j_jvetres-2018-0028_fig_001){ref-type=\"fig\"}).\n\n![The content of PAs in positive feed samples. Silage (s), hay (h)](jvetres-62-183-g003){#j_jvetres-2018-0028_fig_003}\n\nOnly one sample (silage) showed a high total PAs content of 62.4 \u03bcg kg^-1^. The highest determined PAs concentration in hay samples was 42.6 \u03bcg kg^-1^. The overall detected PAs concentrations ranged from 5.6 to 62.4 \u03bcg kg^-1^, and the average content and median for positive samples were 20.1 \u03bcg kg^-1^ and 14.9 \u03bcg kg^-1^, respectively. When all samples were included in the calculation, the average PAs concentration was 8.11 \u03bcg kg^-1^.\n\nSix samples contained more than one monitored alkaloid. The most abundant PAs were lycopsamine and intermedine. Lycopsamine was present in 61.5% of all positive samples, with the determined concentrations ranging from 5.5 to 31.8 \u03bcg kg^-1^. The concentration range of intermedine was \\<LOQ -- 55.1 \u03bcg kg^-1^ and the alkaloid was present in 46.2 % of all positive samples. Senecionine (\\<LOQ -- 14.9 \u03bcg kg^-1^), retrorsine (\\<LOQ -- 7.2 \u03bcg kg^-1^), and senkirkine (6.1 \u03bcg kg^-1^) were among other detected alkaloids.\n\nDiscussion {#j_jvetres-2018-0028_s_004}\n==========\n\nThe analysis of plant materials is still problematic even if state-of-the-art analytical equipment is used. PAs are of purely natural origin in contrast to antibiotics or pesticides ([@j_jvetres-2018-0028_ref_011]), and are coextracted with many other natural compounds. The variation in composition of feed materials and trace concentrations of PAs pose analytical challenges regarding sample preparation and instrumental analysis. Therefore, an appropriate cleaning strategy seems to be one of the most important steps in the sample preparation process.\n\nIn a previously conducted study GC-MS was used for the determination of PAs in feeds ([@j_jvetres-2018-0028_ref_021]). However, LC-MS technique has one advantage over GC-MS: it provides structural information and therefore alkaloids contributing to the overall contamination can be identified. Moreover, the procedure of sample preparation can be simplified as it does not require PAs reduction to the necine base backbone structures and a subsequent derivatisation step.\n\nPreviously hydrochloric acid was chosen as the most effective solvent for PAs extraction from feed matrix. However, in the case of liquid chromatography, better results, especially when it comes to the purity of the chromatograms, were obtained with sulphuric acid. Nevertheless, the recoveries, as well as the conversion of *N*-oxides into their free base forms were very comparable with the use of both acids, as it was also found before ([@j_jvetres-2018-0028_ref_021]).\n\nIn contrast to honey analysis, in the case of feed and other plant materials, there is no single dominant method of sample preparation and extract purification. Bolechov\u00e1 *et al* ([@j_jvetres-2018-0028_ref_002]) applied the QuEChERS method and extraction with acetonitrile and 0.1% formic acid in water. Huybrechts and Callebaut ([@j_jvetres-2018-0028_ref_020]) extracted the PAs with an aqueous solution of HCl with the addition of NaCl and analysed the extracts without further purification. Mulder *et al* ([@j_jvetres-2018-0028_ref_025]) used 2% formic acid for the extraction of PAs from feeds followed by purification on Strata X cartridges with methanol used for PAs elution. Gottschalk *et al* ([@j_jvetres-2018-0028_ref_018]) also used 2% formic acid for PAs isolation, however, for the purification cation exchange cartridges and PAs elution with 5% ammonia in methanol were applied. For extraction of PAs from herbal products and teas Crammer *et al* ([@j_jvetres-2018-0028_ref_006]), Bodi *et al* ([@j_jvetres-2018-0028_ref_001]), and Schulz *et al* ([@j_jvetres-2018-0028_ref_029]) utilised 0.05 M sulphuric acid and purification with C18 or cation exchange cartridges. For the PAs elution methanol or ammonia in methanol were selected. In all cases LC-MS/MS was used for the instrumental analysis. However, none of the approaches described above were useful for the LC-MS analysis. The obtained eluates were not properly purified and many other compounds interfered with the target PAs. In fact, the detection of PAs at the concentration of 20 \u03bcg kg^-1^ was impossible when 3% ammonia in methanol was used as an eluent in combination with an Strata SCX cartridge (Fig. 2a). For this reason a new protocol for sample purification had to be developed.\n\nIn the previously conducted studies, a new elution mixture consisting of ethyl acetate, methanol, ammonia and triethylamine (8:2:0.1:0.1 v/v) was developed to improve the sufficiency of the SPE clean-up of honey matrix ([@j_jvetres-2018-0028_ref_022]).\n\nThis elution mixture was also tested in the SPE purification of feed extracts, and it gave much better results than 3% of ammonia in methanol. However, to obtain the most efficient results, a different combination of solvents (ethyl acetate, methanol, ammonia, triethylamine, and acetonitrile) and different cation exchange cartridges (MCX, Strata SCX, PCX, Strata XC, and HF) were also evaluated.\n\nThe best purification effect was obtained for the Strata SCX cartridge; however, when the solvent combination was regarded, the best results were yielded by the mixture consisting of ethyl acetate, methanol, acetonitrile, ammonia, and triethylamine (8:1:1:0.1:0.1, v/v). The recovery rates of all target alkaloids were also at satisfactory levels. Application of this new elution mixture had crucial importance especially in the case of identification and quantification of the low concentration of monitored PAs in feed materials. In comparison to elution with 3% ammonia in methanol, the quantification of PAs at the level of 5 \u03bcg kg^-1^ could be carried out easily (Fig. 2b). Consequently, it positively influenced the sensitivity of the LC-MS method and could be compared with protocols based on the LC-MS/MS technique ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_025]).\n\nThe mobile phase has to be optimised to obtain the best separation of the analysed compounds. Formic acid of different concentrations, in several cases in combination with ammonium formate ([@j_jvetres-2018-0028_ref_001], [@j_jvetres-2018-0028_ref_006], [@j_jvetres-2018-0028_ref_018], [@j_jvetres-2018-0028_ref_029]), was used as the most popular modifier of water and organic mobile phases ([@j_jvetres-2018-0028_ref_004], [@j_jvetres-2018-0028_ref_011], [@j_jvetres-2018-0028_ref_017], [@j_jvetres-2018-0028_ref_023]). In all published methodologies acetonitrile or methanol was used as the organic phase. However, proposed mobile phases did not always provide good separation of all alkaloids, especially when it came to enantiomeric structures such as intermedine and lycopsamine that usually are co-eluted. In our method, the mixture of methanol and acetonitrile (1:1, v/v) was used as the organic phase. This combination of the organic solvents provided good separation of all analysed alkaloids, including the enantiomeric structures.\n\nSeveral reversed phase C18 columns were tested during optimisation of the separation of the analysed PAs. The best results as far as the separation and peak shape are concerned were obtained with a Gemini NX-C18, 150 mm \u00d7 4.6 mm, 3 \u03bcm column and a mobile phase consisting of 0.2% formic acid in water (A) and a mixture of methanol and acetonitrile (1:1, v/v) (B).\n\nPAs have easily ionisable nitrogen in the ring system and the protonation of PAs molecules can be achieved ([@j_jvetres-2018-0028_ref_007]), that is why ESI in a positive ionisation mode was chosen for the analysis. To optimise all parameters affecting ESI source operation, flow injection analysis (FIA) was used.\n\nThe method was successfully validated according to the SANTE/11945/2015 recommendations ([@j_jvetres-2018-0028_ref_028]). Quantification was based on matrix calibration curves. The method showed adequate linearity in a concentration range 0--100 \u03bcg kg^-1^ with determination coefficient R^2^ \\> 0.99 for all analysed alkaloids. All determined parameters met the SANTE document criteria as recovery was in the range of 70%--120% and relative standard deviations of repeatability and reproducibility were \u226420%. The obtained recovery values for the target PAs were in the range 84.1%--112.9% and were slightly higher than values reported by Bolechov\u00e1 *et al* ([@j_jvetres-2018-0028_ref_002]), who achieved recoveries in the range of 72%--98%, and Gottschalk *et al* ([@j_jvetres-2018-0028_ref_018]), who found recoveries in the range of 69%--104%. The method proved to be selective, as no interfering peaks were determined in the retention times assigned to the particular alkaloid. It can also be affirmed that the method is robust to slight changes in selected performance parameters.\n\nFeeds are considered difficult and complex materials containing a lot of compounds which might interfere and cause both signal suppression and signal enhancement at MS detection. A matrix effect could be noticed in the case of alkaloids such as jacobine or senkirkine; however, the values are still in the acceptable range. According to the PA stability results, the purified extracts can be stored for about a week, which is in line with previously obtained results for extracts prepared for GC-MS analysis ([@j_jvetres-2018-0028_ref_021]).\n\nPlant material is a very difficult matrix, and often methods based on LC-MS/MS have higher LOQ values for the same alkaloids when analysed in plant material than matrices such as honey, meat, or eggs ([@j_jvetres-2018-0028_ref_001], [@j_jvetres-2018-0028_ref_026]). For the developed method the LOQ was established at 5 \u03bcg kg^-1^ for the individual alkaloid, as it was the lowest validation level. This LOQ value is comparable with, or even lower than LOQs reported for methods based on LC-MS/MS methodology that were used for determination of PAs in feeds ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_025]) or in other plant materials such as herbal teas ([@j_jvetres-2018-0028_ref_029]).\n\nThe developed method was applied to the analysis of 32 feed samples. Among detected alkaloids, lycopsamine and intermedine were the most abundant compounds, followed by senecionine and retrorsine. This is in line with Huybrechts and Callebaut ([@j_jvetres-2018-0028_ref_020]), who also reported that lycopsamine and intermedine were the most frequently detected alkaloids in feed samples with a lower contribution of senecionine and retrorsine. Also, Gottschalk *et al* ([@j_jvetres-2018-0028_ref_018]) reported the presence of intermedine, lycopsamine and senecionine in analysed silage samples. Bolechov\u00e1 *et al* ([@j_jvetres-2018-0028_ref_002]) determined retrorsine and senecionine in feed for chickens and senkirkine in silage samples. However, in contrast to other authors reporting PAs contamination in feed materials ([@j_jvetres-2018-0028_ref_002], [@j_jvetres-2018-0028_ref_020], [@j_jvetres-2018-0028_ref_025]), seneciphylline was not detected in any of the analysed materials. This may be due to the difference in the plant species, which was a source of contamination, as well as the variability of the composition and concentration of alkaloids in the same plant species, which strongly depends on the environmental conditions prevailing in a particular year, growth stage, part of the plant, and geographical origin.\n\nDetermined concentrations can be regarded as low and are consistent with results reported in the previously conducted study where a GC-MS technique was used for PAs determination in feed materials ([@j_jvetres-2018-0028_ref_021]). The determined levels are also in agreement with the PAs contamination results of feeds of European origin reported by Mulder *et al* ([@j_jvetres-2018-0028_ref_025]), Bolechov\u00e1 *et al* ([@j_jvetres-2018-0028_ref_002]), Huybrechts and Callebaut ([@j_jvetres-2018-0028_ref_020]), and Gottschalk *et al* ([@j_jvetres-2018-0028_ref_018]). Mean values of PAs concentrations in silage reported by the authors were in the range of 4.8--25.7 \u03bcg kg^-1^. In our study the average content of PAs in positive samples was 20.1 \u03bcg kg^-1^. Mulder *et al* ([@j_jvetres-2018-0028_ref_025]) and Gottschalk *et al* ([@j_jvetres-2018-0028_ref_030]) reported 30 \u03bcg kg^-1^ as the highest PAs concentration detected in silage samples. In our study, the highest PAs content was determined at the level of 62.4 \u03bcg kg^-1^, which is higher but is still comparable with other authors' results.\n\nHoogenboom *et al* ([@j_jvetres-2018-0028_ref_019]) studied the transfer of PAs from feed to milk in dairy cows. The transfer rate was about 0.1% of the overall daily dose of PAs. The authors showed that the rate of PAs transferred to milk depended on the type of alkaloid. *N*-oxides were not detected in milk even though they made a large contribution to the content of PAs in contaminated feed. However, the highest transfer rate was observed for jacoline, even though it was not the main contributor to the overall content of PAs in feed ([@j_jvetres-2018-0028_ref_019], [@j_jvetres-2018-0028_ref_027]). This was also confirmed by Mulder *et al* ([@j_jvetres-2018-0028_ref_027]), who studied the transfer of PAs to eggs and meat in laying hens. The author found that the transfer rates were relatively low, ranging from 0.02% to 0.23% and also depended on the PAs composition of the plant containing the alkaloids ([@j_jvetres-2018-0028_ref_027]). Considering the transfer rates and the highest content of PAs determined in our study, it can be concluded that the detected concentrations should not pose the risk of alkaloid transfer into food such as milk or meat.\n\nRegarding acute poisoning, it can also be stated that the determined PAs concentrations should not pose a risk to the animals. The reported cumulative lethal dose for cows is 2.5 mg PAs kg^-1^ body weight per day (which translates to 1,500 mg for a cow of 600 kg) during exposure for 18 days ([@j_jvetres-2018-0028_ref_018], [@j_jvetres-2018-0028_ref_030]). In our study, the highest detected concentration was 62.4 \u03bcg kg^-1^, which would be only 0.06% of the lethal dose in the case of 15 kg of silage consumption. However, the detected concentrations, if consumed in the long term, especially in combination with other toxins, may adversely affect animals.\n\nIn conclusion, the EFSA has recommended the development of sensitive and selective methods suitable for determination of PAs in feeds, and the presented method fulfils the Agency's requirements of selectivity and sensitivity. Application of the new elution mixture consisting of ethyl acetate, methanol, acetonitrile, ammonia, and triethylamine significantly improved the purification effect and markedly improved the sensitivity of the method. Validation according to SANTE/11945/2015 proved the method's reliability. The developed method was applied to the analysis of 32 feed samples, of which almost 41% were positive for at least one of the PAs. Detected PAs concentrations were relatively low and should not be of concern as a cause of acute poisoning of animals or the transfer of PAs to food of animal origin; however, in the long term consumption they may affect the health of animals.\n\n**Conflict of Interests Statement** The authors declare that there is no conflict of interests regarding the publication of this article.\n\n**Financial Disclosure Statement** The research was financed from the Institute's statutory sources.\n\n**Animal Rights Statement** None required.\n"}
+{"text": "The Simbu serogroup viruses, one of the largest serogroups within the genus *Orthobunyavirus* of the family *Peribunyaviridae*, comprises at least 24 antigenically different, but serologically related viruses that are transmitted mainly by *Culicoides* biting midges (De Regge [2017](#vms3129-bib-0003){ref-type=\"ref\"}; Hirashima *et\u00a0al*. [2017](#vms3129-bib-0007){ref-type=\"ref\"}). Several Simbu serogroup viruses have been shown to cross the placenta of ruminants to the developing fetus and to cause outbreaks of abortion, stillbirth and malformations (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; Tsuda *et\u00a0al*. [2004](#vms3129-bib-0012){ref-type=\"ref\"}; Yanase *et\u00a0al*. [2010](#vms3129-bib-0013){ref-type=\"ref\"}; Hoffmann *et\u00a0al*. [2012](#vms3129-bib-0008){ref-type=\"ref\"}; Golender *et\u00a0al*. [2015](#vms3129-bib-0006){ref-type=\"ref\"}; Brenner *et\u00a0al*. [2016](#vms3129-bib-0001){ref-type=\"ref\"}). The congenital malformations seen at birth are known as arthrogryposis hydranencephaly syndrome and correlate with the stage of pregnancy at which the mother is infected (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; De Regge [2017](#vms3129-bib-0003){ref-type=\"ref\"}). Brain damage can range from microscopic to mild or severe malformations and includes hydranencephaly, microencephaly and polio\u2010encephalomyelitis (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; Brenner *et\u00a0al*. [2016](#vms3129-bib-0001){ref-type=\"ref\"}). In cattle, severe brain malformations may occur if the dam is infected between 76 and 106\u00a0days of pregnancy. Surprisingly, some calves with little more than an intact brain stem are born alive. They are often blind, poorly responsive to external stimuli and have a weak suckling reflex. Nevertheless, they may be capable of standing and getting around, and are able to develop physically if carefully tended and fed. The colloquial term for them is \"dummy calves\", and the veterinary term is weak calf syndrome (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; De Regge [2017](#vms3129-bib-0003){ref-type=\"ref\"}).\n\nTwo members of the Simbu serogroup, Akabane and Shuni viruses, have been isolated from the brains of congenitally malformed ruminants in Israel (Golender *et\u00a0al*. [2015](#vms3129-bib-0006){ref-type=\"ref\"}; Brenner *et\u00a0al*. [2016](#vms3129-bib-0001){ref-type=\"ref\"}). A recent retrospective serosurvey revealed that Israeli ruminants have been exposed to several additional Simbu viruses, including *Shamonda* and *Sathuperi* (Brenner *et\u00a0al*. [2018](#vms3129-bib-0002){ref-type=\"ref\"}). These viruses are considered potential teratogenic agents in ruminants (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; Yanase *et\u00a0al*. [2012](#vms3129-bib-0014){ref-type=\"ref\"}; Hirashima *et\u00a0al*. [2017](#vms3129-bib-0007){ref-type=\"ref\"}), and seem to be circulating in Israel (Brenner *et\u00a0al*. [2018](#vms3129-bib-0002){ref-type=\"ref\"}).\n\nIn late April 2017, an apparently healthy 1\u2010month\u2010old male calf was transferred to the Kimron Veterinary Institute (KVI). A few days later, the herdsman reported that the calf was slow to respond to its surroundings and was not able to feed on its own. Blindness was observed upon clinical examination. Epidemiological investigation revealed that the calf was born to a dairy heifer in a large dairy farm located in the central coastal plain of Israel, which is one of 11 farms that are monitored for arboviruses and their vectors by the KVI. The heifer was of the same age as a group of na\u00efve heifers serving as sentinels in 2016. Seroconversion to Simbu serogroup viruses was detected in the sentinels in July 2016 by ELISA (IDEXX Schmallenberg Ab Test Kit, Liebefeld\u2010Bern, Switzerland), indicating that the aforementioned dam may have been exposed to at least one Simbu serogroup virus during the first trimester of her pregnancy. This dam died from dystocia during the calf\\'s birth.\n\nDespite being carefully tended for 5\u00a0months, the calf remained weak and in poor body condition. It was humanely sacrificed in the first week of September 2017. Blood samples were taken from the calf upon arrival in April, in the last week of July, and just before euthanasia were negative for Simbu serogroup viruses by real\u2010time PCR (Fischer *et\u00a0al*. [2013](#vms3129-bib-0005){ref-type=\"ref\"}). While opening the calf\\'s skull, it was apparent that the skull cavity was narrower than usual. The cerebellum and brainstem were intact, the cerebrum was absent and the space contained a copious volume of cerebrospinal fluid (CSF) (Fig.\u00a0[1](#vms3129-fig-0001){ref-type=\"fig\"}). This was defined as hydranencephaly. No other gross pathological changes were seen in the internal organs or the skeleton. Total viral nucleic acids were extracted from internal organs (brain, thymus, spleen, liver and testis) and CSF using the Maxwell 16 Viral Total Nucleic Acid Purification Kit (Promega). Nested PCR targeting the large segments of Simbu serogroup viruses was performed as follows: Pan Simbu reverse transcription PCR targeting the large RNA segment was conducted according to Fischer *et\u00a0al*. [2013](#vms3129-bib-0005){ref-type=\"ref\"};. The PCR products served as the template for the nested PCR with primer pairs SimbunestF; 5\u2032\u2010TGATTAGTGAACCAGGGGACTC\u20103\u2032; and SimbunestR; 5\u2032\u2010 GCACTCCATTTTGACATATCAGC\u20103\u2032 (expected product size ca. 200\u00a0bp). Reverse transcription PCR targeting the small and medium RNA segments was conducted according to Hirashima *et\u00a0al*. [2017](#vms3129-bib-0007){ref-type=\"ref\"} (expected product size ca. 443\u00a0bp) and Otani *et\u00a0al*. [2013](#vms3129-bib-0010){ref-type=\"ref\"} (expected product size ca. 488\u00a0bp), respectively. Partial nucleotide sequences of the three segments were amplified only from two samples: CSF and testes. sequence data obtained in this study have been deposited in GenBank with accession no.: MH331909\u2010 MH331911. Phylogenetic analysis performed for each segment using maximum likelihood implemented in PhyML (Dereeper *et\u00a0al*. [2008](#vms3129-bib-0004){ref-type=\"ref\"}) showed that the virus sequences detected in the Israeli calf were virtually identical to those of Peaton virus (PEAV) (Fig.\u00a0[2](#vms3129-fig-0002){ref-type=\"fig\"}a--c). The presence of all three segments of this virus also confirms that this is indeed PEAV and not a reassortant containing one PEAV segment. Our attempts to isolate the virus from the CSF were unsuccessful.\n\n![Gross pathology: Narrow skull cavity with partially developed frontal cerebral lobes (mark with stars). Deformed and reduced cerebral mass (microencephaly), with enlarged internal cisterns (hydranencephaly) (Full arrows). The cerebellum appears normal (dotted arrows).](VMS3-5-87-g001){#vms3129-fig-0001}\n\n![Rooted maximum\u2010likelihood phylogenetic trees of Simbu serogroup viruses for the S (a), M (b) and L (c) segments, based on a general time\u2010reversible and gamma\u2010distributed rate heterogeneity (GTR_G) model of nucleotide substitution. S, M and L segments of the Israeli calf\\'s virus were compared with the respective sequences from the validated Simbu viruses. Whenever possible, we used Simbu serogroup viruses for which full segment sequences were available. Homologous sequences from Oropouche virus were used as the outgroup. Scale bar indicates estimated nucleotide substitutions. Only bootstrap values greater than 70% are shown.](VMS3-5-87-g002){#vms3129-fig-0002}\n\nThis is the first genomic detection of PEAV in a ruminant in Israel. A recent retrospective serosurvey suggested circulation of this virus in Israel between 2008 and 2014 (Brenner *et\u00a0al*. [2018](#vms3129-bib-0002){ref-type=\"ref\"}). Nevertheless, based on serological cross\u2010reactivity, neutralization tests can only differentiate between Simbu serogroup viruses up to a certain level. To specifically identify an individual virus, detection of its genome is necessary. Previous studies on Simbu serogroup viruses have shown that virus isolation from offspring with congenital malformations is difficult and viral RNA is often no longer detectable by real\u2010time PCR in malformed progeny (St George & Standfast [1989](#vms3129-bib-0011){ref-type=\"ref\"}; De Regge [2017](#vms3129-bib-0003){ref-type=\"ref\"}). Therefore, the effective genomic detection of PEAV from the tissues of a calf with hydranencephaly exhibiting weak calf syndrome has an epidemiological importance, as it shows that PEAV is present in Israel and affects cattle. Further support for the presence and circulation of PEAV in Israel was provided during the writing of this manuscript: PEAV genomic materials were detected for the first time in two pools of *Culicoides imicola* trapped during summer and autumn of 2017 at two different locations along the Israeli central coastal plain (data not shown). The presence of PEAV presented here might also indicate that other Simbu serogroup viruses such as *Shamonda* and *Sathuperi* may be circulating in Israel as was previously proposed (Zentis *et\u00a0al*.[2012](#vms3129-bib-0015){ref-type=\"ref\"}; Brenner *et\u00a0al*. [2018](#vms3129-bib-0002){ref-type=\"ref\"}), and it is only a matter of time until they will be gnomically detected in animals and/or vectors.\n\nFinally, a recent sero\u2010survey has indicated the presence of PEAV in Africa (Mathew *et\u00a0al*. [2015](#vms3129-bib-0009){ref-type=\"ref\"}), but this is the first detection of the virus genetic material outside Australia and Japan (Yanase *et\u00a0al*. [2010](#vms3129-bib-0013){ref-type=\"ref\"}). Thus, our results are of relevance to a potential spread of Simbu serogroup viruses into Europe.\n\nSource of funding {#vms3129-sec-0003}\n=================\n\nMinistry of Agriculture and Rural Development (MOAG) Chief Scientist\\'s grant \\# 33\u201003\u20100004.\n\nConflict of interest {#vms3129-sec-0004}\n====================\n\nThe authors declare they have no conflict of interest.\n\nEthical statement {#vms3129-sec-0005}\n=================\n\nThe authors confirm that the ethical policies of the journal, as noted on the journal\\'s author guidelines page, have been adhered to and the appropriate ethical review committee approval has been received. The US National Research Council\\'s guidelines for the Care and Use of Laboratory Animals were followed.\n\nContributions {#vms3129-sec-0006}\n=============\n\nDr. Leibovich and Dr. Brenner preformed all clinical examination and ensure the calf\\'s health. Dr. Edry preformed the post mortem examination and pathology. Dr. Yanase advised on the molecular analysis and assists in writing the manuscript. Dr. Behar is head of KVI\\'S monitoring system for arboviruses and their vectors, was responsible for the epidemiological investigation and conducted all molecular analysis and insect collections. Dr Behar and Dr. Brenner wrote the manuscript together.\n\nWe wish to thank KVI\\'s herdsman Ran Zarecki for his devotion in tending to the calf and hand\u2010feeding it for 5\u00a0months, and to Dr. Orly Friedgut for her helpful comments.\n"}
+{"text": "Introduction {#s1}\n============\n\nDirected cell migration is fundamental to diverse physiological processes, including developmental morphogenesis, cancer metastasis, and the immune response. To directionally migrate, cells must sense signals from the extracellular environment and transduce the signals to downstream effectors that polarize the cytoskeleton and regionally regulate its organization and dynamics [@pone.0041413-Horwitz1]. It is well established that the actin and microtubule (MT) cytoskeletal systems are critical to directed cell movement. The actin cytoskeleton is the primary force generator for cell migration: Rapid polymerization of F-actin drives lamellipodia/lamella protrusions in the cell leading edge, and contraction of acto-myosin bundles in the cell body causes retraction of the trailing cell rear [@pone.0041413-Pollard1].\n\nThe MT cytoskeleton is required for rapid, directed migration of tissue cells [@pone.0041413-Rodriguez1], however its role in cell motility is less clear than that of actin. MTs are thought to provide critical cell polarization cues that dictate spatial regulation of protrusive and contractile actin-based processes [@pone.0041413-Rodriguez1]. Indeed, the MT cytoskeleton displays a characteristic polarized organization and dynamics in migrating cells. MTs growing from the MT-organizing center (MTOC) and Golgi apparatus orient towards the leading edge lamella/lamellipodia region [@pone.0041413-Rodriguez1], [@pone.0041413-Miller1]. In addition, MT dynamic instability is polarized in migrating cells. MTs in the lamellipodia grow more slowly, persistently and parallel to the cell edge than those in the lamella [@pone.0041413-WatermanStorer1], [@pone.0041413-Wadsworth1], [@pone.0041413-Wittmann1]. These leading edge MTs have been termed \"pioneer MTs,\" and in PtK epithelial cells, exhibit a decreased catastrophe frequency (switch from growth to shrinkage), resulting in an increased growth time [@pone.0041413-Wittmann1]. Pioneer MTs are thought to guide directed migration by localizing signals that promote lamellipodial actin dynamics [@pone.0041413-WatermanStorer2], and/or facilitating turnover of leading edge focal adhesions [@pone.0041413-Kaverina1]. However, the mechanisms that establish pioneer MTs in the lamellipodia of migrating cells are not understood.\n\nThe Rho family of small GTPases act as molecular switches that locally control cytoskeleton dynamics through multiple effectors [@pone.0041413-EtienneManneville1]. In migrating cells, activated Rac1 accumulates in lamellipodia and stimulates local actin polymerization to drive leading edge protrusion [@pone.0041413-Gardiner1], [@pone.0041413-Wang1]. In addition, we previously showed that lamellipodial pioneer MTs can be induced by activation of Rac1 [@pone.0041413-Wittmann1]. We identified two distinct pathways downstream of Rac1 that promote pioneer MT growth. One depends on Rac1-mediated activation of p21-activated kinase1 (Pak1), which phosphorylates and inactivates the MT-destabilizing protein Op18/stathmin [@pone.0041413-Wittmann1], [@pone.0041413-Wittmann2]. Alternatively, Rac1-mediated inactivation of glycogen-synthase kinase 3\u03b2 (GSK3\u03b2) induces MT binding and growth stabilization by CLASP [@pone.0041413-Kumar1], [@pone.0041413-Wittmann3]. However, in those studies, we found that neither Pak1 activation nor GSK3\u03b2 inhibition were sufficient for inducing pioneer MTs [@pone.0041413-Wittmann1], [@pone.0041413-Wittmann2], [@pone.0041413-Wittmann3], indicating that additional factors must regulate pioneer MTs downstream of Rac1.\n\nTo find downstream targets of Rac1 signaling that promote pioneer MTs, we performed an RNAi-based screen of MT regulatory factors to identify proteins whose depletion suppresses MT growth and orientation induced by constitutively activated Rac1. For unbiased and statistically robust analysis of millions of MT growth excursions in many cells across a range of conditions ([Tables S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}, [S3](#pone.0041413.s004){ref-type=\"supplementary-material\"}, [S5](#pone.0041413.s006){ref-type=\"supplementary-material\"}, [S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}, [S8](#pone.0041413.s009){ref-type=\"supplementary-material\"}), we utilized plusTipTracker, an automated image analysis program that tracks motion of fluorescent-tagged EB3, a MT plus-end binding protein (+TIP) [@pone.0041413-Matov1], [@pone.0041413-Applegate1], [@pone.0041413-Myers1]. We identified six RNAi treatments that suppressed Rac1-induced MT assembly dynamics, and one of these, microtubule-affinity-regulating kinase2 (MARK2), that additionally suppressed Rac1-mediated MT orientation in lamellipodia. The MARK2 homologues are known to mediate organization of MTs and polarity establishment in epithelia, oocytes and neurons [@pone.0041413-Cohen1], [@pone.0041413-Doerflinger1], [@pone.0041413-Shulman1], [@pone.0041413-Cox1], [@pone.0041413-Matenia1], but how MARK2 controls MT dynamics in cell migration has not been demonstrated. Our results suggest that MARK2 is required for polarization in migrating cells through promotion of pioneer MTs downstream of Rac1.\n\nResults {#s2}\n=======\n\nRac1 Activity Promotes Pioneer MTs in Lamellipodia of U2-OS Osteosarcoma Cells {#s2a}\n------------------------------------------------------------------------------\n\nWe previously showed that Rac1 activation in PtK1 kidney epithelial cells promotes formation of pioneer MTs, characterized by persistent, slow plus-end growth in lamellipodia parallel to the leading cell edge [@pone.0041413-Wittmann1]. To identify downstream targets of Rac1 that promote pioneer MTs, we sought to utilize an RNAi screening approach with U2-OS osteosarcoma cells due to their human origin and amenability to transfection. To validate the efficacy of these cells for our purpose, we transfected them with a constitutively activated Rac1 mutant (CA-Rac1) and analyzed their MT organization and dynamics. Fluorescent localization of F-actin showed that expression of CA-Rac1 in U2-OS cells promoted lamellipodia formation and reduced stress fibers compared to mock-transfected controls or cells expressing dominant negative Rac1 (DN-Rac1) ([Fig. 1A](#pone-0041413-g001){ref-type=\"fig\"}) and similar to effects in other cell types [@pone.0041413-Hall1]. Immunolocalization of MTs showed that like previous studies in PtK1 cells, expression of CA-Rac1 had a strong effect on MT organization, inducing a bundle of MTs parallel to the cell edge in lamellipodia around most of the cell periphery. In contrast, in controls and cells expressing DN-Rac1, MTs were for the most part oriented perpendicular to the cell edge and absent from lamellipodia. Thus, Rac1 activity promotes formation of oriented lamellipodial MTs in U2OS cells, similar other cell types [@pone.0041413-Wittmann1].\n\n![Rac1 promotes pioneer MTs in U2-OS cells.\\\n(A) Immunolocalization of microtubules (MTs) and fluorescent phalloidin staining of F-actin in mock-transfected cells (control) or cells expressing CA-Rac1 or DN-Rac1. Contrast inverted, left, center. Zoom of boxed region, second column. Insets, BFP-Rac1. Bar, 10 \u00b5m. (B) Workflow of plusTipTracker software for detecting mKO-EB3 comets, tracking them, and classifying MT growth excursions. (C) Top: Proportion of MT growth excursions in each subpopulation in non-targeting control vector, CA-Rac1 or DN-Rac1 expressing cells. Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. Box-plots of speed (D) and lifetime (E) of MT growth excursions, conditions as in C. (\\*, p\\<0.001, Kolmogorov-Smirnov, \\*, p\\<0.05, Students). (F) Top: mKO-EB3 tracks from 2 min time-lapse movies (frame rate \u200a=\u200a3 s) colored according to the key in C overlaid on images of mKO-EB3 (inverted contrast), conditions described in A. Bottom: zoom of boxed region. Bars, 10 \u00b5m. (G) Percentage of MT growth tracks within 5 \u00b5m from the leading edge whose angle relative to the edge is between 0--45\u00b0 (blue) or between 45--90\u00b0 (green), conditions as in C.](pone.0041413.g001){#pone-0041413-g001}\n\nTo determine the effects of Rac1 activity on MT growth dynamics, we performed live-cell imaging of the +TIP protein EB3 fused to monomeric Kusabira Orange (mKO-EB3, Video S1) as a non-perturbing marker of growing MT plus ends [@pone.0041413-Stepanova1], [@pone.0041413-Piehl1] and used plusTipTracker software package [@pone.0041413-Matov1], [@pone.0041413-Applegate1], [@pone.0041413-Myers1] for high-throughput measurement of MT growth dynamics. Because +TIP proteins only associate with MT ends during active growth excursions, we were limited to measuring MT growth speed and growth persistence lifetime, i.e. the duration of a growth excursion before loss of fluorescent EB3 from the MT plus end due to transition to shortening (catastrophe) or to pause [@pone.0041413-Stepanova1]. Automated tracking is powerful because of the ability to measure \u223c10,000--30,000 MT growth excursions per condition ([Tables S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}, [S3](#pone.0041413.s004){ref-type=\"supplementary-material\"}, [S5](#pone.0041413.s006){ref-type=\"supplementary-material\"}, [S8](#pone.0041413.s009){ref-type=\"supplementary-material\"}), allowing robust statistical analysis that facilitates identification of treatments that may have subtle, yet reproducible effects on MT growth. The limitations and accuracy of this program are described in detail in [@pone.0041413-Matov1], [@pone.0041413-Applegate1].\n\nWe classified MT growth excursions as \"slow\" vs. \"fast\" or \"short-lived\" vs. \"long-lived\" based on thresholds of the mean value for each parameter from the population of all measurements from non-targeting vector-transfected control cells ([Fig. 1B](#pone-0041413-g001){ref-type=\"fig\"}, [@pone.0041413-Applegate1], [@pone.0041413-Myers1]). These classes of MT growth dynamics were color-coded and the percentage of growth excursions in each class in the measured population was depicted as a bar graph to allow easy visual assessment and comparison of the effects of perturbations on MT growth dynamics across a range of conditions [@pone.0041413-Applegate1], [@pone.0041413-Myers1]. Mean growth speeds and growth lifetimes, standard errors and the number of cells and MTs analyzed, for all conditions in all figures are presented in the Supplemental tables. We also generated image overlays of color-coded MT growth tracks to allow qualitative visualization of regional differences in MT growth dynamics throughout the cell ([Fig. 1B](#pone-0041413-g001){ref-type=\"fig\"}).\n\nAnalysis of MT growth with plusTipTracker software showed that compared to control, CA-Rac1 expression increased the proportion of slow, short-lived (red) MT growth excursions, whereas expression of DN-Rac1 increased the proportion of fast, short-lived (yellow) MT growth excursions ([Fig. 1C](#pone-0041413-g001){ref-type=\"fig\"}, [Table S2](#pone.0041413.s003){ref-type=\"supplementary-material\"}). Changes in proportions of MTs in each class were mirrored by changes in mean values, as CA-Rac1 decreased and DN-Rac1 increased MT growth speed, and either CA-Rac1 or DN-Rac1 reduced MT growth lifetime ([Fig. 1D, 1E](#pone-0041413-g001){ref-type=\"fig\"}, [Table S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}) compared to control. These results differ from previous results where CA-Rac1 was shown to inhibit catastrophe and thus promote longer-lived MT growth [@pone.0041413-Wittmann1]. This discrepancy could arise from the technical improvements afforded by our current methodology including automated tracking of EB3 comets compared to hand-tracking fluorescent tubulin in our previous studies [@pone.0041413-WatermanStorer1], [@pone.0041413-Wittmann1]. These improvements include a shorter image acquisition frequency, which would allow measurement of shorter-lived growth excursions than the previous study, and higher accuracy localization of the MT end, which would allow measurement of slower growth that may have been counted as pause in the previous study. Moreover, manual measurements of MT dynamics tend to focus on longer-lived growth phases. These subjective biases are reduced by automated tracking approaches. Alternatively, Rac1 activity may promote subtle differences in MT dynamic instability depending on cell type. Nonetheless, the overall phenotype of CA-Rac1 expression, with extended MTs running parallel to the cell edge in the lamellipodia are similar in both studies, indicating that Rac1 activity generates major effects on MTs in multiple cell types.\n\nWe next examined the spatial distribution of MT growth induced by CA-Rac1. Given that pioneer MTs grow parallel to the leading edge [@pone.0041413-WatermanStorer1], we classified the orientation of MT growth tracks within 5 \u00b5m of the cell edge as \"parallel\" (0--45\u00b0) or \"perpendicular\" (45--90\u00b0, [Fig. 1G](#pone-0041413-g001){ref-type=\"fig\"}, [Table S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}). This analysis and examination of image overlays of color-coded MT growth tracks ([Fig. 1F](#pone-0041413-g001){ref-type=\"fig\"}) showed that CA-Rac1 promoted slow MT growth (red and green) in the cell periphery parallel to the cell edge, and long-lived MT growth (blue and green) in the cell center. In contrast, controls and cells expressing DN-Rac1 displayed fast, long-lived MT growth (blue) in the cell center, and both fast and slow, short-lived MT growth (red and yellow) perpendicular to the leading edge at the cell periphery. Thus, our results show that activated Rac1 mediates pioneer MT formation in U2-OS cells at least in part by promoting slow, short-lived MT growth excursions in lamellipodia parallel to the cell edge, thus allowing for the basis of a phenotypic screen.\n\nAn RNAi Screen Identifies Proteins that Modulate MT Dynamics Downstream of Rac1 in U2-OS Osteosarcoma Cells {#s2b}\n-----------------------------------------------------------------------------------------------------------\n\nTo find downstream targets of Rac1 that promote pioneer MTs, we performed RNAi-based depletion of 23 MT-regulatory factors ([Figs. 2](#pone-0041413-g002){ref-type=\"fig\"} and [3](#pone-0041413-g003){ref-type=\"fig\"}) and screened for proteins whose depletion suppresses the slow, short-lived MT growth parallel to the leading edge induced by CA-Rac1 in U2-OS cells ([Fig. 3](#pone-0041413-g003){ref-type=\"fig\"}). We focused our screen on non-motor MT-binding target proteins that are known to regulate MT growth behavior, but it is unknown if they are regulated by Rac1. These included APC, APC2, ACF7, CLASP2, XMAP215, CLIP170/115, p150*^glued^*, p50/dynamitin, doublecortin, EB1, katanin p60 subunit, MARK1, MARK2, MARK3, MAP1A, MAP1B, MAP1S, MAP2, MAP4, STOP, spastin, or Op18/stathmin. shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, spastin and katanin p60 (see Methods). Op18/stathmin and CLASP that are both known to regulate MT dynamics downstream of Rac1 in other cell types [@pone.0041413-Wittmann2], [@pone.0041413-Wittmann3] were included as positive controls. We used plusTipTracker to analyze the speed, lifetime, and orientation of MT growth excursions in movies of cells expressing mKO-EB3 and treated with either siRNAs or shRNAs targeting (see Methods) one of these MT regulatory proteins together with ([Fig. 3A](#pone-0041413-g003){ref-type=\"fig\"}, [Tables S5](#pone.0041413.s006){ref-type=\"supplementary-material\"}, [S6](#pone.0041413.s007){ref-type=\"supplementary-material\"}) or without ([Fig. 2](#pone-0041413-g002){ref-type=\"fig\"}, [Tables S3](#pone.0041413.s004){ref-type=\"supplementary-material\"}, [S4](#pone.0041413.s005){ref-type=\"supplementary-material\"}) the additional expression of CA-Rac1. Expression of target proteins in U2OS cells was verified for as many target proteins as possible (see methods), however in many cases antibodies were not available. Thus, negative results in this screen should be interpreted with caution.\n\n![Effects of MT regulatory protein depletion on MT growth dynamics in U2-OS cells.\\\nTop: Results of analysis of time-lapse movies of mKO-EB3 with PlusTipTracker software. Proportion of MT growth excursions in each subpopulation in control shRNA vector-transfected cells (WT) or cells treated with RNAis targeting the protein noted (kd). shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60 (see Methods). Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for each subpopulation.](pone.0041413.g002){#pone-0041413-g002}\n\n![RNAi screen for proteins whose depletion blocks CA-Rac1 effects on MT growth and orientation.\\\n(A) Top: Proportion of MT growth excursions in each subpopulation in control shRNA vector-transfected cells (control), cells expressing CA-Rac1, CA-Rac1 and additionally treated with RNAis targeting the protein noted (kd), DN-Rac1. shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60 (see Methods). Dashed line 1: proteins which pass the first criteria (red tracks\\<50%); solid line 2: proteins which pass the second criteria (40%\\<red+green tracks \\<60%). Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. (B) mKO-EB3 tracks from 2 min time-lapse movies (frame rate\u200a=\u200a3 s) colored according to the key in A overlaid on images of mKO-EB3 (inverted contrast) in cells treated with RNAis targeting the noted proteins with (bottom) or without (top) CA-Rac1 expression. Bar, 10 \u00b5m. (C) Percentage of MT growth tracks within 5 \u00b5m from the leading edge whose angle relative to the cell edge is between 0--45\u00b0 (blue) or between 45--90\u00b0 (green), conditions as in B.](pone.0041413.g003){#pone-0041413-g003}\n\nWe first analyzed the effects of RNAi treatments targeting MT regulatory proteins in the absence of CA-Rac1. As expected, RNAi treatment targeting most of these MT regulatory proteins induced alterations of MT growth behavior compared to non-targeting RNAi controls ([Fig. 2](#pone-0041413-g002){ref-type=\"fig\"}, see Methods). RNAi treatments targeting either the +TIP rescue factor CLIP115/170 [@pone.0041413-Komarova1], the XMAP215 protein that promotes rapid MT assembly and disassembly [@pone.0041413-Vasquez1], the MT stabilizer and assembly factor DCX [@pone.0041413-Coquelle1], the plus end anchor and MT stabilizing protein APC2 [@pone.0041413-McCartney1], [@pone.0041413-Shintani1], the MT growth enhancer MAP1S [@pone.0041413-Halpain1] or the MT stabilizer STOP [@pone.0041413-Margolis1] resulted in a greater fraction of both slow short-lived (red) and slow long-lived (green) MT growth excursions compared to control. This suggests that these proteins promote fast MT growth in U2-OS cells, and confirms demonstrated (XMAP215 [@pone.0041413-Vasquez1], DCX [@pone.0041413-Coquelle1]) and suggested (STOP [@pone.0041413-Margolis1], MAP1S [@pone.0041413-Halpain1], APC2 [@pone.0041413-Kita1]) functions of several of these proteins from previous studies in other cell types. RNAi treatments targeting either the dynactin subunit p50/dynamitin, the MT stabilizer MAP2, the microtubule affinity regulating kinases (MARK1, MARK2, MARK3) that promote MT disassembly [@pone.0041413-Bright1], the catastrophe factor Op18/stathmin [@pone.0041413-Belmont1], the MT severing protein katanin p60 [@pone.0041413-McNally1], the MT-actin crosslinking and catastrophe-promoting protein ACF7 [@pone.0041413-Kodama1], or the MT growth stabilizer CLASP2 [@pone.0041413-MimoriKiyosue1] resulted in a greater fraction of both fast short-lived (yellow) and fast long-lived (blue) MT growth excursions compared to control. This suggests that these proteins slow or inhibit MT growth in U2-OS cells, and confirms demonstrated (Op18/stathmin [@pone.0041413-Belmont1], katanin [@pone.0041413-McNally1], ACF7 [@pone.0041413-Kodama1]) and suggested (MARKs [@pone.0041413-Bright1], CLASP [@pone.0041413-MimoriKiyosue1]) functions of these proteins from previous studies in other cell types. RNAi treatments targeting each of the remaining proteins (the +TIP EB1, the MT severing protein Spastin, the assembly promoting MAP1 family members (MAP1A and MAP1B), the dynactin subunit p150*^glued^*, and the growth-promoting +TIP APC had little effect on the rate or duration of MT growth excursions, suggesting possibly inefficient RNAi targeting, low protein expression in U2-OS cells, the presence of redundant proteins (i.e. APC2, EB3, MAP1s), or cell type-specific effects. Differences in expected effects of RNAi treatment on MT dynamics based on previous studies (i.e. CLIP115/170, MAP2) could be due to similar issues with RNAi targeting and/or cell type specificity. However, in general, these results demonstrate that depletion of known MT regulatory proteins produces expected effects on U2-OS MT growth dynamics.\n\nTo identify proteins whose depletion suppresses the slow, short-lived MT growth parallel to the leading edge induced by CA-Rac1, we then determined the effects of RNAi treatments targeting MT regulatory proteins in the presence of activated Rac1. We considered a \"hit\" based on three criteria. First, because the percentage of slow, short-lived (red) MT growth tracks was increased in cells expressing CA-Rac1 ([Fig. 1C](#pone-0041413-g001){ref-type=\"fig\"}, 51%) compared to controls (38%), we sought targets whose RNAi treatment reduced the percentage of slow, short-lived (red) MT growth tracks in cells expressing CA-Rac1 to less than 50% (black dashed line, [Fig. 3A](#pone-0041413-g003){ref-type=\"fig\"}). RNAi treatments meeting this criterion included those targeting dynactin subunits (p50/dynamitin, p150*^glued^*), MAPs (MAP4, MAP1B, MAP2), MT severing proteins (Spastin, katanin p60), MT plus end-binding proteins (EB1, APC2, CLASP2), the MT depolymerizing factor Op18/stathmin, the cytoskeletal linker ACF7, and MARK2 ([Table S6](#pone.0041413.s007){ref-type=\"supplementary-material\"}). Our identification of Op18 and CLASP, previously known targets of Rac1 regulation [@pone.0041413-Wittmann2], [@pone.0041413-Wittmann3], validates the efficacy of our automated MT analysis and selection criteria approach. Second, because CA-Rac1 increased the total proportion of slow-growing MTs (red+green, 65%) compared to control (red+green, 52%), our second criteria for a \"hit\" was an RNAi treatment that reduced the percentage of slow growth tracks (red+green) in cells expressing CA-Rac1 to between 40% and 60% (black lines, [Fig. 3A](#pone-0041413-g003){ref-type=\"fig\"}). RNAi treatments meeting both criteria included those targeting p150*^glued^*, APC2, Spastin, EB1, Op18, and MARK2 ([Table S6](#pone.0041413.s007){ref-type=\"supplementary-material\"}), suggesting that these proteins are required for the slow, short-lived MT growth induced by CA-Rac1.\n\nAs a third criterion, we next sought to identify target proteins whose RNAi targeting additionally suppressed the effects of CA-Rac1 on MT orientation in the lamellipodium. We analyzed image overlays of color-coded MT growth tracks in cells expressing CA-Rac1 and treated with siRNAs or shRNAs targeting p150*^glued^*, APC2, spastin, EB1, Op18 and MARK2 ([Fig. 3B, 3C](#pone-0041413-g003){ref-type=\"fig\"}, [Table S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}). Of these, only cells expressing CA-Rac1 and shRNA targeting MARK2 showed a reduction in MTs oriented parallel to the leading edge compared to cells expressing CA-Rac1 alone. Together, these results suggest that MARK2 may be a key Rac1 target promoting leading edge pioneer MTs by regulating both MT growth speed and lifetime as well as growth orientation in the cell periphery.\n\nMARK2 Promotes Short-lived MT Growth Parallel to the Cell Edge Downstream of Rac1 {#s2c}\n---------------------------------------------------------------------------------\n\nWe next sought to validate the results of our RNAi screen suggesting that Rac1's effects on MT dynamics and orientation are mediated by MARK2. Western blot of U2-OS cell lysates showed that either shRNA or siRNA effectively depleted MARK2 protein ([Fig. 4A](#pone-0041413-g004){ref-type=\"fig\"}, lanes 2 and 5) and expression of an shRNA-resistant GFP-tagged MARK2 in cells expressing MARK2-specific shRNA restored near wild-type protein level ([Fig. 4A](#pone-0041413-g004){ref-type=\"fig\"}, lane 3). In fixed cells, both the elongated MTs in lamellipodia parallel to the cell edge and the discoid cell morphology induced by CA-Rac1 were abolished by shRNA-mediated MARK2 depletion, and MT organization and cell morphology were rescued by re-expression of GFP-MARK2 ([Fig. 4B](#pone-0041413-g004){ref-type=\"fig\"}). Analysis of MT growth dynamics via tracking mKO-EB3 (Video S2) showed that shRNA-mediated MARK2 depletion in cells expressing CA-Rac1 increased the proportion of fast, long-lived (blue) MT growth and decreased slow, short-lived (red) growth ([Fig. 3A](#pone-0041413-g003){ref-type=\"fig\"}, [4C](#pone-0041413-g004){ref-type=\"fig\"}, [Table S2](#pone.0041413.s003){ref-type=\"supplementary-material\"}), increasing mean MT growth speed and lifetime compared to cells expressing CA-Rac1 alone, and similar to these values in control shRNA treated-cells ([Fig. 4D, 4E](#pone-0041413-g004){ref-type=\"fig\"}, [Table S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}). Expression of GFP-MARK2 in shRNA-mediated MARK2-depleted, CA-Rac1-expressing cells rescued the effects of MARK2 depletion on MT growth lifetime ([Fig. 4E](#pone-0041413-g004){ref-type=\"fig\"}) and growth orientation in lamellipodia ([Fig. 4F, 4G](#pone-0041413-g004){ref-type=\"fig\"}, [Table S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}). However, despite rescue of MT growth lifetime and orientation, expression of GFP-MARK-2 in MARK2 shRNA-treated cells further increased MT growth speed compared to cells expressing CA-Rac1 alone ([Fig. 4D](#pone-0041413-g004){ref-type=\"fig\"}). To determine if the increased MT growth speed was due to an off-target effect, we tested to see if an shRNA targeting a distinct sequence in MARK2 had similar effects on MT dynamics ([@pone.0041413-Suzuki1], Fig. S1, [Tables S1](#pone.0041413.s002){ref-type=\"supplementary-material\"} and [S8](#pone.0041413.s009){ref-type=\"supplementary-material\"}). This showed that both shRNAs had similar effects: increases in both MT growth speed and lifetime, and expression of GFP-MARK2 rescued the increased MT growth lifetime but showed further increase of MT growth speed ([Fig. 4D, 4E](#pone-0041413-g004){ref-type=\"fig\"}, Fig. S1, [Table S8](#pone.0041413.s009){ref-type=\"supplementary-material\"}). This suggests that MT growth speed may be tightly regulated by a specific MARK2 protein level [@pone.0041413-Matenia2]. However, these results show that MARK2 promotes pioneer MT formation in lamellipodia downstream of activated Rac1, at least in part by inducing short-lived MT growth excursions parallel to the cell edge.\n\n![MARK2 regulates MT growth dynamics and orientation downstream of Rac1.\\\n(A) Western blot of lysates of U2-OS cells transfected with control shRNA (lane 1), MARK2-shRNA (lane 2), MARK2-shRNA and shRNA-resistant GFP-MARK2, (lane 3), non-targeting siRNA pool (lane 4) or MARK2 siRNA (lane 5). GAPDH and DM1A were used as a loading control. (B) Immunostaining of MTs (inverted contrast) in cells expressing BFP-CA-Rac1 (Rac1) and GFP-shRNA targeting MARK2 (shRNA), or rescued with GFP-MARK2 expression. Center column: zoom of boxed region. (C) Top: Proportion of MT growth excursions in subpopulations, conditions as in B. Control represents control shRNA vector transfected cells. Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. Box-plots of speed (D) and lifetime (E) of MT growth excursions, conditions as in B. (\\*p\\<0.001, Kolmogorov-Smirnov, \\*p\\<0.05, Students). (F) Top: mKO-EB3 tracks from 2 min time-lapse movies (frame rate\u200a=\u200a3 s) colored according to the key in C overlaid on images of mKO-EB3 (inverted contrast), conditions as in B. Bottom: Zoom of boxed regions. (G) Percentage of MT growth tracks within 5 \u00b5m from the leading edge whose angle relative to the cell edge is between 0--45\u00b0 (blue) or between 45--90\u00b0 (green), conditions as in B. Bars, 10 \u00b5m.](pone.0041413.g004){#pone-0041413-g004}\n\nMARK2 is Required for Pioneer MTs in the Leading Edge and Polarized Cell Migration {#s2d}\n----------------------------------------------------------------------------------\n\nWe then characterized the function of MARK2 in U2OS cells, comparing its role in MT organization and dynamics in non-polarized and polarized migrating cells. Staining of MTs in fixed non-polarized cells showed that neither shRNA-mediated MARK2 depletion nor re-expression of GFP-MARK2 in MARK2-depleted cells substantially affected overall MT organization ([Fig. 5A](#pone-0041413-g005){ref-type=\"fig\"}). However, analysis of MT growth dynamics via tracking mKO-EB3 (Video S2) showed that MARK2 depletion did, in fact, affect MT dynamics in non-polarized cells. As noted, MARK2 shRNA treatment increased the proportion of fast, long-lived MT growth excursions (blue) and reduced the proportion of both slow short-lived and slow long-lived growth excursions (red and green, [Fig. 2](#pone-0041413-g002){ref-type=\"fig\"}, [Fig. 5B](#pone-0041413-g005){ref-type=\"fig\"}, [Table S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}), increasing both speed and lifetime of MT growth ([Fig. 5C, 5D](#pone-0041413-g005){ref-type=\"fig\"}, [Table S2](#pone.0041413.s003){ref-type=\"supplementary-material\"}). Expression of GFP-MARK2 in MARK2-depleted non-polarized cells increased the proportion of fast, short-lived MT growth excursions (yellow), rescuing the effects of MARK2 depletion on MT growth lifetime, but further increased MT growth speed compared to control ([Fig. 5B--D](#pone-0041413-g005){ref-type=\"fig\"}). This is similar to effects of rescuing MARK2 depletion with GFP-MARK2 in the presence of CA-Rac1 ([Fig. 4D, 4E](#pone-0041413-g004){ref-type=\"fig\"}, Video S2), supporting the notion that RNAi effects on MT lifetime were MARK2-dependent and effects on growth speed may be dependent on the specific level of MARK2 protein or were MARK2 independent. Analysis of the spatial organization of MT growth excursions in image overlays of color-coded MT growth tracks showed that in non-polarized cells, neither shRNA-mediated MARK2 depletion nor re-expression of GFP-MARK2 in MARK2 depleted cells substantially affected the spatial distribution or orientation relative to the cell edge of MT growth excursions ([Fig. 5E, 5F](#pone-0041413-g005){ref-type=\"fig\"}, [Table S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}). Thus, although MARK2 clearly regulates MT growth lifetime and orientation in lamellipodia downstream of activated Rac1, in non-polarized cells in the absence of exogenous Rac1 activation, MARK2 promotes short-lived MT growth excursions, but has little effect on MT organization or growth orientation.\n\n![MARK2 regulates MT growth lifetime.\\\n(A) Immunolocalization of MTs (inverted contrast) in cells expressing control shRNA-GFP (top), MARK2 shRNA-GFP (middle), or MARK2 shRNA-GFP together with RNAi-resistant GFP-MARK2 (bottom). Cells co-expressing MARK2 shRNA-GFP and GFP-MARK2 could be recognized by the targeting of GFP-MARK2 to the centrosome (arrowhead). (B) Top: Proportion of MT growth excursions in each subpopulation in cells under the conditions described in A. Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. Box-plots of speed (C) and lifetime (D) of MT growth excursions, conditions as in A. (\\*, p\\<0.001, Kolmogorov-Smirnov, \\*, p\\<0.05, Students) (E) Top: mKO-EB3 tracks from 2 min time-lapse movies (frame rate\u200a=\u200a3 s) colored according to the key in B overlaid on images of mKO-EB3 (inverted contrast), conditions as in A. Bottom: zoom of boxed regions. (F) Percentage of MT growth tracks within 5 \u00b5m from the leading edge whose angle relative to the cell edge is between 0--45\u00b0 (blue) or between 45--90\u00b0 (green), conditions as in A. Bars, 10 \u00b5m.](pone.0041413.g005){#pone-0041413-g005}\n\nTo resolve these apparently disparate results, we sought to determine if MARK2 specifically functions in spatial organization of MT growth in polarized cells undergoing directed migration in a scratch wound assay, where Rac1 activity has been shown to be polarized towards the leading edge in other cell types [@pone.0041413-Gardiner1]. We analyzed MT growth in polarized cells migrating at a wound edge, comparing MT growth parameters averaged from whole cells (W) to MT growth parameters from within 5 \u00b5m of the leading edge (LE). In control shRNA treated-wound-edge cells, MTs in the leading edge exhibited more slow and short-lived growth excursions that were more often parallel to the cell edge compared to MTs throughout the whole cell, typical of pioneer MTs and similar to the effects of CA-Rac1 on MT growth ([Fig. 6A--E](#pone-0041413-g006){ref-type=\"fig\"}, [Tables S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}, [S2](#pone.0041413.s003){ref-type=\"supplementary-material\"}, [S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}, Video S3). shRNA-mediated MARK2 depletion in wound-edge cells homogenized differences in MT growth speed between MTs at the leading edge and throughout the whole cell, increased MT growth lifetime in both the leading edge and whole cell, and reduced the proportion of MTs growing parallel to the cell edge compared to controls or cells expressing CA-Rac1 ([Fig. 6A--E](#pone-0041413-g006){ref-type=\"fig\"}, [Tables S1](#pone.0041413.s002){ref-type=\"supplementary-material\"}, [S2](#pone.0041413.s003){ref-type=\"supplementary-material\"}, [S7](#pone.0041413.s008){ref-type=\"supplementary-material\"}, Video S3). Thus, MARK2 is required for polarizing MT growth dynamics to form pioneer MTs that exhibit slow, short-lived growth parallel to leading lamellipodia in polarized, directionally migrating cells.\n\n![MARK2 is required for leading edge MT dynamics of directed migrating cells.\\\n(A) Above: Proportion of MT growth excursions in each subpopulation in control shRNA vector-transfected whole cells (W), expression of CA-Rac1, MARK2 shRNA treatment or from within 5 \u00b5m from the leading edge of wound edge cells with MARK2 shRNA (MARK2 RNAi (LE)) or control shRNA (control (LE)) treatment. Below: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. Box-plots of speed (B) and lifetime (C) of MT growth excursions, conditions as in A. (\\*, p\\<0.001, Kolmogorov-Smirnov, \\*, p\\<0.05, Students). (D) mKO-EB3 tracks from 2 min time-lapse movies (frame rate\u200a=\u200a3 s) of cells at the edge of a scratch-wound, colored according to the key in A overlaid on images of mKO-EB3 (inverted contrast) with MARK2 shRNA (MARK2 RNAi) or control shRNA (control RNAi) treatment. Only MT growth tracks that come within 5 \u00b5m of the leading edge are shown. Bar, 10 \u00b5m. (E) Percentage of MT growth tracks within 5 \u00b5m from the leading edge whose angle relative to the cell edge is between 0--45\u00b0 (blue) or between 45--90\u00b0 (green), conditions as in A.](pone.0041413.g006){#pone-0041413-g006}\n\nWe then sought to determine whether localization of MARK2 is polarized in migrating cells and/or controlled by Rac1 activity ([Fig. 7](#pone-0041413-g007){ref-type=\"fig\"}). Unfortunately, available antibodies to MARK2 were not effective for immunofluorescnce under conditions required to preserve MT structure in fixed U2-OS cells. Comparing mCherry fluorescence as a volume marker to expressed GFP-MARK2 showed that GFP-MARK2 accumulated in protrusive regions of non-polarized cells ([Fig. 7A](#pone-0041413-g007){ref-type=\"fig\"}). In migrating cells at a wound edge GFP-MARK2 localized in cell-cell junctions ([@pone.0041413-Cohen1], [Fig. 7B](#pone-0041413-g007){ref-type=\"fig\"}) and was concentrated in the single leading protrusion facing the direction of migration ([Fig. 7B](#pone-0041413-g007){ref-type=\"fig\"}). Co-expression of GFP-MARK2 and Rac1 mutants showed that CA-Rac1 induced concentration of GFP-MARK2 in lamellipodia around the entire cell periphery, and all peripheral localization was abolished in cells expressing DN-Rac1 ([Fig. 7C](#pone-0041413-g007){ref-type=\"fig\"}). Thus, MARK2 localizes to the leading edge of directionally migrating cells in a Rac1-activity dependent-manner, suggesting MARK2 regulates MT organization in these regions.\n\n![Accumulation of MARK2 in the leading edge depends on Rac1 activity.\\\n(A) Fluorescence images of GFP-MARK2 (upper left), mCherry-C1 (upper right) and phase image (lower left) in a non-polarized cell. Line scans (lower right) show the average normalized (to maximal) fluorescent intensity of three different cells expressing GFP-MARK2 (red) and mCherry-C1 (blue) along lines places similarly as those shown in the upper panel. Arrow represents cell protrusion. (B) Fluorescence image of GFP-MARK2 (upper left), and phase image (upper right) in migrating cell at a wound edge. Arrow: cell protrusions, Asterisks: cell-cell junctions. Line scans (lower panel) show average normalized (to maximal) fluorescent intensity of lines in three different cells expressing GFP-MARK2 along protrusion area (Lines 1) or non-protrusion area (Line 2). (C) Fluorescence images of GFP-MARK2 (upper left and upper right) in cell expressing BFP-CA-Rac1 (inset upper left) or BFP-DN-Rac1 (inset upper right). Insets show the BFP fluorescence images. Line scans indicate average normalized (to maximal) fluorescent intensity of three different cells expressing GFP-MARK2 along lines placed similarly to those in in CA-Rac1 (left) or DN-Rac1 (right). Bars, 10 \u00b5m.](pone.0041413.g007){#pone-0041413-g007}\n\nWe finally sought to determine the role of MARK2 in polarization and directional migration of U2-OS cells. To assay cell polarization, we examined effects of siRNA-mediated MARK2 depletion on centrosome orientation by immunolocalizing \u03b3-tubulin in cells in a scratch wound assay. We utilized siRNA-mediated targeting in this experiment because of its increased transfection efficiency compared to shRNA. This allowed better determination of effects of MARK2 depletion on monolayer wound healing, which is coordinated by all cells in the monolayer, and may not be apparent with the lower level transfection afforded by shRNA methods. This showed that two hours after wounding, MARK2-depleted cells were inhibited in forward orientation of centrosomes, and MTs were less extended towards the leading edge compared to non-targeting siRNA treated-control cells ([Fig. 8A, 8B](#pone-0041413-g008){ref-type=\"fig\"}). Analysis of time-lapse phase contrast movies showed that MARK2 depletion caused wound-edge cells to migrate significantly more slowly and with reduced directional persistence compared to control ([Fig. 8C--F](#pone-0041413-g008){ref-type=\"fig\"}, Video S4, [@pone.0041413-Simpson1]). Together, these results suggest that MARK2 promotes directed cell migration via establishing polarity of the MT cytoskeleton and enhancing pioneer MT growth in the leading edge.\n\n![MARK2 is required for cell polarization and directional migration.\\\n(A) Immunostaining of MTs (left and right) and \u03b3-tubulin (merge in center) in wound edge cells in non-targeting siRNA pool (top, control RNAi) and MARK2 siRNA treatment (bottom, MARK2 RNAi). Dotted circles, centrosome; bar, 10 \u00b5m. (B) Percentage of cells in the wound edge with centrosomes in front of the nucleus. In B, D, F \\*p\\<0.05, Student's t-test. (C) Wound-healing assay in non-targeting siRNA pool (control RNAi; left) or MARK2 siRNA (MARK2 RNAi; right), time after wounding shown. Dashed line, position of wound edge at t\u200a=\u200a0. Bar, 50 \u00b5m. (D) Average migration velocity of non-targeting siRNA pool (control) and MARK2 siRNA-treated (MARK2 RNAi) cells. (E) Rose plots of the position of nuclei over 5 hr (each cell track colored differently) for non-targeting siRNA pool-(above, control RNAi) or MARK2 siRNA-treated cells (below, MARK2 RNAi) at the edge of a wound. Arrows: open region of wound. (F) Distance from origin (position at time\u200a=\u200a0) divided by total distance travelled over 5 hr for non-targeting siRNA pool and MARK2 siRNA treated cells.](pone.0041413.g008){#pone-0041413-g008}\n\nDiscussion {#s3}\n==========\n\nWe utilized automated image analysis algorithms to screen for regulators of MT assembly dynamics induced downstream of Rac1 activation. Our results suggest that Op18/stathmin, the dynactin component p150*^glued^*, the +TIP proteins APC2 and EB1, the MT severing enzyme Spastin, and the kinase MARK2 may be potential downstream mediators of Rac1 signaling that regulate MT dynamics. Mechanistic links between Rac1 and APC or Op18 have previously been reported [@pone.0041413-Wittmann2], [@pone.0041413-Kawasaki1], however the other factors identified here represent novel potential Rac1 targets. Thus, our results suggest that Rac1 acts as a master regulator of leading edge MT dynamics through multiple molecular pathways.\n\nOur study represents the first time that parameters of MT growth dynamics have been used as the basis of a phenotypic screen, and was made possible by the development of automated image analysis programs [@pone.0041413-Matov1], [@pone.0041413-Applegate1], [@pone.0041413-Myers1]. This method allowed measurement of MT growth excursions throughout the cell, including short-lived, slow growth excursions that were not measurable by hand-tracking in previous studies [@pone.0041413-Wittmann1], suggesting that activated Rac1 primarily promotes less persistent and slower polymerization of MTs, especially in the leading edge ([Fig. 1C--E](#pone-0041413-g001){ref-type=\"fig\"}).\n\nBy further analyzing MT growth orientation in lamellipodia, we identified a single protein, MARK2, as required for Rac1 effects on both MT growth dynamics and orientation mediating pioneer MT formation. MARK2 is the mammalian homologue of the polarity protein PAR-1 which is thought to mediate polarity through effects on MTs [@pone.0041413-Goldstein1]. However, this is the first report of MARK2/PAR1's effects on polarity and MT dynamics being controlled by the Rac1 signaling cascade. We found that RNAis targeting any of the three MARK isoforms (1, 2, and 3) promoted fast MT assembly, suggesting that MARKs function to slow MT growth or destabilize MTs in situ, in agreement with previous studies [@pone.0041413-Bright1]. Interestingly, depletion of MARK 1, 2 or 3 suppressed the slow, short-lived growth induced by CA-Rac1, but only depletion of MARK2 suppressed the effects of Rac1 on MT orientation in lamellipodia. Taken together, this suggests that all three MARK isoforms function downstream of Rac1 to regulate MT growth, but MARK2 additionally regulates Rac-1 mediated MT orientation.\n\nThe molecular mechanism by which Rac1 promotes MARK2-mediated regulation of MT growth dynamics and orientation in the leading edge is of interest. MARK2 is known to be activated by phosphorylation [@pone.0041413-Matenia1], and active MARK2 in turn phosphorylates MT-associated proteins (MAPs) such as tau, MAP2 and MAP4, which causes them to dissociate from MTs [@pone.0041413-Matenia1], [@pone.0041413-Drewes1], [@pone.0041413-Ebneth1]. Notably, we found that RNAis targeting several MAPs including MAP4, MAP1B and MAP2, reduced the proportion of slow, short-lived MT growth excursions ([Fig. 3A](#pone-0041413-g003){ref-type=\"fig\"}), similar to the effects of MARK2 depletion. Further study is needed to determine if Rac1 mediates local downstream inactivation of MARK2 in lamellipodia to reduce MAP phosphorylation and promote MAP-MT binding to locally stabilize MTs and drive pioneer MT growth. In addition, although we showed that Rac1 activity controls MARK2 localization and Rac1's effects on MTs require MARK2, it would be of interest to additionally know if Rac1 controls MARK2's kinase activity. MARK2 has been shown to interact with the Rac1 targets, GSK3\u03b2 and a PAK kinase [@pone.0041413-Matenia2], [@pone.0041413-Kosuga1], [@pone.0041413-Timm1], and these in turn regulate Op18/stathmin and CLASP [@pone.0041413-Wittmann2], [@pone.0041413-Wittmann3]. This suggests that the MARK2, GSK3\u03b2, and PAK pathways of MT regulation could intersect downstream of Rac1.\n\nThe question of how MARK2 regulates MT orientation in lamellipodia is unclear. Since MT bending and growth parallel to the cell edge is thought to be driven by MT interactions with lamellipodial F-actin retrograde flow [@pone.0041413-WatermanStorer1], [@pone.0041413-Salmon1], it is possible that MARK2's effect on pioneer MT orientation are mediated by phosphoregulation of MAPs that bind both MTs and actin, as has been demonstrated for ACF7, APC and MAP2 [@pone.0041413-Kodama1], [@pone.0041413-Munemitsu1], [@pone.0041413-Okada1], [@pone.0041413-Kim1], [@pone.0041413-Sattilaro1], [@pone.0041413-Griffith1]. Alternatively, MARK2 may coordinately regulate MT and actin cytoskeletal dynamics through Pak kinases, which target both MT and actin binding proteins [@pone.0041413-Bokoch1].\n\nOur study builds on previous work showing that MARK kinases and their PAR1 homologues regulate cell polarity through downstream effects on MTs [@pone.0041413-Goldstein1]. In a recent study of dendritic spines, it was found that that knockdown of the MARK2 homologue PAR-1 decreased the distance and duration of MT growth, as analyzed by tracking fluorescent +TIP proteins [@pone.0041413-Hayashi1]. This contrasts our study in which MT growth persistence was increased in MARK2 knock-down cells ([Figs. 3](#pone-0041413-g003){ref-type=\"fig\"}, [4](#pone-0041413-g004){ref-type=\"fig\"} and [5](#pone-0041413-g005){ref-type=\"fig\"}). This suggests that MARK2 may utilize distinct effectors to control MT dynamics differently in each system. Notably, tracking of GFP-EB3 dynamics in *Drosophila* wing epidermis [@pone.0041413-Harumoto1] and oocytes [@pone.0041413-Parton1] revealed that the MARK2 homologue, Par-1, plays a crucial role in establishing a small bias in the orientation of growing MTs without a substantial change in overall MT distribution, resulting in asymmetric segregation of polarity proteins. In addition, similar to lamellipodia in migrating cells, MTs tend to bend and grow along the posterior cortex of *Drosophila* oocytes where Par-1 localizes, and Par-1 mutants fail to show these polarized MT growth excursions [@pone.0041413-Parton1]. These data indicate that MARK2/PAR-1 has a conserved role in organizing MTs and regulating their growth and orientation in many systems.\n\nConsistent with the role of MARK2 in MT reorganization, our results demonstrate that MARK2 is essential for cell polarization and directed cell migration. We found that MARK2-depleted cells in a wound edge have reduced pioneer MTs in the leading lamellipodia and show faster MT growth ([Fig. 6](#pone-0041413-g006){ref-type=\"fig\"}). Indeed, depletion of MARK2 caused a failure in centrosome polarization towards the wound edge, and inhibited migration velocity and directionality ([Fig. 8](#pone-0041413-g008){ref-type=\"fig\"}). MARK2/PAR-1 has also been shown to play a crucial role in directed migration in other cell types. For example, in the cerebral cortex, neurons depleted of MARK2 exhibit altered directed migration and reduced centrosome motility [@pone.0041413-Sapir1]. In addition, *par-1* null mutants in *Drosophila* border cells also fail in directed cell migration, showing abnormal protrusion in their front [@pone.0041413-McDonald1]. Therefore, MARK2/PAR-1 is a conserved key mediator for establishing cell polarity during directed migration, likely through regulating MT dynamics under the control of Rac1.\n\nTaken together, our study shows the power of high-resolution, quantitative live cell imaging assays for refined screening of protein function, and identifies MARK2 as essential for linking Rac1 activation to polarization of MTs and their assembly dynamics critical to directed cell migration.\n\nMaterials and Methods {#s4}\n=====================\n\nCDNA Expression Constructs {#s4a}\n--------------------------\n\nPCR products and products of restriction digests were purified by gel electrophoresis and extraction using the QIAquick\u2122 gel extraction kit (QIAGEN, Valencia, CA, USA). Plasmid DNA was purified from overnight cultures using the QIAprep\u2122 Spin Miniprep kit (QIAGEN). Restriction endonucleases were purchased from Life Technologies (Grand Island, NY, U.S.A.) or New England Biolabs (Ipswich, MA, U.S.A.). Sequencing was used to confirm the complete cDNA sequence (Florida State University Bioanalytical and Molecular Cloning DNA Sequencing Laboratory). mKusabira Orange (mKO) fused to EB3 was constructed using a N1 (Clontech-style) cloning vector. The fluorescent protein cDNA was PCR amplified (Phusion Flash; Finnzymes, Espoo, Finland) with a 5\u2032 primer encoding an AgeI site and a 3\u2032 primer encoding a NotI (N1) site. The purified and digested PCR products were ligated into a similarly digested EGFP-N1 cloning vector backbone (Clontech, Mountain View, CA, U.S.A.). Human EB3 cDNA (gift of Lynne Cassimeris, Lehigh University; NM_012326.2) was PCR amplified (Phusion Flash) with primers containing NheI and BamHI restriction enzyme sites and ligated into the mKO-N1 cloning vector to produce a fusion with EB3 attached to the N-terminus of mKO separated by a 6-amino acid linker. DNA for mammalian cell transfection was prepared using the Plasmid Maxi kit (QIAGEN). Proper localization was confirmed using widefield (Nikon 80i; TRITC filter set) and spinning disk microscopy (Olympus DSUIX81; TRITC filter set). RNAi-resistant MARK2 tagged with EGFP for rescue experiments was constructed by cloning a BglII-KpnI fragment of pENTR(tm)221-MARK2 (Invitrogen) into pEGFP-C1 (CLONTECH Laboratories, Inc., Mountain View, CA) and substituting sequence (5\u2032-AATTATCGATAAAACG-3\u2032) for sequence (5\u2032-GATCATTGACAAGACT-3\u2032, from GM Bioscience (Rockville, MD)). Human Rac1 G12V and T17N cDNA (Missouri S&T cDNA Resource Center, Rolla, MO) were PCR amplified with primers containing with a 5\u2032 primer encoding an KpnI and a 3\u2032 primer encoding a ApaI restriction enzyme sites and ligated into the BFP-tagC1 cloning vector to produce a fusion with Rac1 G12V or T17N attached to the C-terminus of BFP.\n\nCell Culture and Transfection {#s4b}\n-----------------------------\n\nU2-OS cells were obtained from American Type Culture Collection and maintained in McCoy's 5A with 10% FBS and 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA) at 37\u00b0C in 5% CO~2~. For live imaging, medium was supplemented with 25 mM HEPES pH\u200a=\u200a7.2 and 30 units/ml Oxyrase. Transfections were performed with an nucleofector (solution kit V (Lonza, Walkersville, MD)). For stable expression of fluorescent EB3, transfected cells were selected for 2 weeks in media supplemented with 0.8 mg/ml G-418 (Invitrogen), and low level-expressing cells collected by FACS.\n\nWestern Blot {#s4c}\n------------\n\nCells were lysed in 1X Laemmli sample buffer, separated on 4--12 or 4--20% SDS-PAGE gels (Invitrogen), transferred to immobilon-P membranes, and visualized by immunoblotting with enhanced chemiluminescence (GE Healthcare, Piscataway, NJ) with the following antibodies: Primary antibodies: Mouse anti-tubulin (1\u22363,000, DM1A, Sigma-Aldrich, St. Louis, MO); rabbit anti-Par1 (1\u2236200, obtained from A. Suzuki, Yokohama-City University, Yokohama, Japan); mouse anti-GAPDH (1\u22365,000, Cell Signaling, Danvers, MA); rabbit anti-Op18 (1\u22361000, Agrisera, V\u00e4nn\u00e4s, Sweden); rabbit anti-X-orbit antibody (1\u22361000, obtained from R. Heald, UC Berkley, Berkley, CA); and mouse anti-p150*^glued^* (1\u22361000, BD Biosciences, San Jose, CA). Secondary antibodies: HRP-conjugated anti-mouse or rabbit (1\u22363,000, Jackson ImmunoResearch Laboratories, Inc, West Grove, PA).\n\nImmunofluorescence {#s4d}\n------------------\n\nCoverslips with bound cells were fixed in 4% paraformaldehyde and 0.5% Glutaraldehyde (Electron Microscopy Science) in PHEM buffer (60 mM PIPES, 27 mM HEPES, 10 mM EGTA, 8 mM MgSO~4~, pH7.0) for 20 minutes at room temperature, permeabilized with 0.1% Triton X-100 in PHEM for 5 min, free aldehydes were reacted with 0.1 M glycine for 5 min, cells were washed 3x for 5 min in PBS, and blocked in blocking solution (2% BSA IgG free and protease free in PBS; Jackson ImmunoResearch Laboratories, Inc.). Coverslips were incubated with mouse anti-\u03b3 tubulin (1\u2236200, AKT-15, Sigma-Aldrich) and/or rat anti-tubulin \u03b1 (1\u2236300, YL 1/2, AbDSerotec, Oxford, UK) antibodies diluted in blocking solution overnight in a humid chamber at 4\u00b0C, washed 3X for 5 min in PBS, incubated with fluorophore-conjugated secondary antibodies (1\u2236250, Jackson ImmunoResearch Laboratories, Inc.) and Alexa Fluor 655 phalloidin (1\u2236400; Invitrogen) for 1 h, washed again, and mounted on a slide in mounting media (Dako).\n\nRNAi {#s4e}\n----\n\nU2-OS cells stably expressing mKO-EB3 were transfected by Amaxa nucleofector with SMARTpool (3 \u00b5g siRNA, Thermo Fisher Scientific, Lafayette, CO) or shRNA vector (1 \u00b5g DNA, Thermo Fisher Scientific), and cultured for 24 h at 37\u00b0C with 5% CO~2~. Control cells were either transfected with non-targeting pool siRNA (D-001810, Thermo Fisher Scientific) ([Fig. 8](#pone-0041413-g008){ref-type=\"fig\"}) or control shRNA vector (Thermo Scientific,) ([Figs. 1B--F](#pone-0041413-g001){ref-type=\"fig\"}, [2](#pone-0041413-g002){ref-type=\"fig\"}, [3](#pone-0041413-g003){ref-type=\"fig\"}, [4](#pone-0041413-g004){ref-type=\"fig\"}, [5](#pone-0041413-g005){ref-type=\"fig\"}, and [6](#pone-0041413-g006){ref-type=\"fig\"}). The next day, cells were transfected with BFP-tagC1-Rac1V12 and replated on coverslips, and experiments were performed 14--20 hours later to allow time for Rac1 expression. The following SMARTpools were used: APC (APC: L-003869), APC2 (APC2: L-009847), MACF1 (ACF7: L-013618), CKAP5 (XMAP215: L-006847), STIMN1 (Op18: L-005152), DCTN (p150*^glued^*: L-012874), CLIP1 (CLIP170: L-005294), CYLN2 (CLIP115: L-013511), MAP6 (STOP: L-026713), BPY2IP1 (MAP1S: L-016881), SPAST (Spastin: L-014070) and KATNA1 (katanin p60: L-005157, Thermo Scientific Open Biosystems). The following shRNA vectors were used: EB1 (pSuper-EB1, obtained from T. Wittmann, UCSF, San Francisco, CA), an shRNAmir-GFP library (Thermo Fisher Scientific Open Biosystems ) included shRNA vectors for CLASP2, DCTN2, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. For depletion of MARK2 in a scratch wound assay ([Fig. 8](#pone-0041413-g008){ref-type=\"fig\"}), SMARTpools (L-004260, Thermo Fisher Scientific) were used. For depletion of MARK2 in [Figure S1](#pone.0041413.s001){ref-type=\"supplementary-material\"} and [Tables S8](#pone.0041413.s009){ref-type=\"supplementary-material\"}, human par-1b shRNA vector (obtained from A. Suzuki, [@pone.0041413-Suzuki1]) was used. Western blot analysis was performed to verify protein depletion for CLASP2, p150*^glued^* dynactin, Op18/stathmin, and MARK2.\n\nMicroscopy {#s4f}\n----------\n\nmKO-EB3 dynamics were imaged on a Yokogawa spinning disk confocal microscope system described in [@pone.0041413-Myers1] using a 60\u00d71.4 NA Plan Apo oil immersion objective lens (Nikon). Images of EB3 dynamics were generated using illumination from the 561 nm laser and were captured using a 500-ms exposure at 3 sec image intervals for 3 min on a Coolsnap HQ2 cooled CCD (Photometrics) operated in the 14-bit mode. Images of immunofluorescence, GFP-MARK2 and mCherry volume marker were acquired using the same microscope system, using the 561 or 488 nm lasers as appropriate. Microscope system automation was controlled with Metamorph software (Molecular Devices).\n\nFor wound healing assays, cells were grown to confluency and wounded 48 h after transfection by generating a longitudinal scratch using a razor blade. After a 2 h recovery at 37\u00b0C in 5% CO~2~, phase-contrast images were acquired at 5-min intervals for 20 h on the microscope described above using a 20\u00d70.45 NA phase objective and an 0.52 NA LWD condenser using Metamorph's Multi-dimensional Acquisition (MDA) software module.\n\nImage Analysis {#s4g}\n--------------\n\nMT growth dynamics were analyzed from EB3 time-lapse movies using plusTipTracker [@pone.0041413-Matov1], [@pone.0041413-Applegate1] (<http://lccb.hms.harvard.edu/software.html>). For this study MT shrinkage or pause events were not estimated. Accordingly, we set the gap length to relatively short intervals to limit the gap closing mechanism in the software mostly to bridge short-term out-of-focus movements of the comets. The following parameter set was used for all movies in the dataset: maximum gap length, 5 frames; minimum track length, 3 frames; search radius range, 8--12 pixels; maximum forward angle, 30\u00b0; maximum backward angle, 10\u00b0; maximum shrinkage factor, 1.5; fluctuation radius, 1 pixel.\n\nTo categorize EB3 tracks based on MT growth speed and growth excursion lifetime, we utilized the tool within plusTipTracker called 'Quadrant Scatter Plots.' Briefly, the function generates a 2D scatter plot of speed versus lifetime with each point representing a single MT growth excursion defined by a single continuous EB3 track. Then, the points on the graph are divided into four subpopulations based on whether they were above or below the average growth speed (13 \u00b5m/min) and average growth lifetime (18 sec) of all EB3 tracks from all control cells analyzed in the study. These four subpopulations are coded by color, and a percentage bar showing the relative proportion of the subpopulations was generated. Colored tracks were overlaid on an inverted image of mKO-EB3 using the same color scheme to show how each subpopulation is distributed across the cell. To analyze the angle of MT growth excursions near the leading edge, EB3 tracks in these color overlays were analyzed within 5 \u00b5m from the edge of cell and their angle relative to the leading edge was categorized as parallel if the track was less than 45 degrees or perpendicular if more than 45 degrees.\n\nFor analysis of centrosome position in wound healing assays, images of fixed cells stained for MTs, \u03b3-tubulin and DNA were aligned with the wound parallel to the image x axis and lines were plotted parallel to the image x axis to bisect each nucleus of cells at the wound edge. Centrosomes in front of the line were considered \"oriented\".\n\nCell migration in wound healing assays was quantified by hand-tracking the nucleus in time-lapse phase-contrast image series using the \"track points\" function in MetaMorph software to determine instantaneous velocity and distance to origin divided by the total path length.\n\nStatistical analysis was performed using the Analyze-It plug-in (Analyze-It Software Ltd.) for Excel (Microsoft) or Matlab. Box plots were used to represent the distributions of MT growth speed and lifetime measurements under different conditions: red lines represent median and boxes around them represent 25^th^ and 75^th^ percentile of the dataset; 98% of the data points are inside the whisker area and the residual outliers are represented by the red points. For analysis of MT growth speed and lifetimes, Kolmogorov-Smirnov test was used to analyze potential differences between the distributions of all measured MT growth excursions pooled from all cells under each condition, and Student's t-test was used to analyze potential differences between the distributions of mean values for each cell under each condition. For centrosome position and cell migration velocity and directionality measurements, Student's t-test was used to analyze potential differences between the distributions of mean values for each condition.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**shRNA targeting of a distinct sequence of MARK2 has similar effects on MT dynamics.** An shRNA vector with sequence targeting MARK2 (human par-1b) was obtained from A. Suzuki [@pone.0041413-Suzuki1]. (A) Top: Proportion of MT growth excursions in subpopulations of CA-Rac1-transfected (CA-Rac1), CA-Rac1 and MARK2 shRNA-transfected (+CA-Rac1, MARK2 RNAi \\#2), or CA-Rac1, MARK2 shRNA and GFP-MARK2-transfected cells (+CA-Rac1, MARK2 RNAi \\#2+GFP-MARK2). Bottom: Color key showing MT growth speed and growth excursion lifetime ranges for subpopulations. Box-plots of speed (B) and lifetime (C) of MT growth excursions, conditions as in A. (\\*, p\\<0.001, Kolmogorov-Smirnov, \\*, p\\<0.05, Students).\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Mean MT growth speed and growth excursion lifetimes.** Control cells were transfected with control shRNA vector; MARK2 shRNA vector was used for MARK2 RNAi. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics. Data shown is depicted graphically in [Fig. 1D, 1E](#pone-0041413-g001){ref-type=\"fig\"}; [Fig. 4D, 4E](#pone-0041413-g004){ref-type=\"fig\"}; [Fig. 5C, 5D](#pone-0041413-g005){ref-type=\"fig\"}; [Fig. 6B, 6C](#pone-0041413-g006){ref-type=\"fig\"}.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Proportion of MT growth excursions in subpopulations grouped according to growth speed and growth excursion lifetime.** Control cells were transfected with control shRNA vector; MARK2 shRNA vector was used for MARK2 RNAi. Data shown is depicted graphically in [Fig. 1C](#pone-0041413-g001){ref-type=\"fig\"}, [Fig. 4C](#pone-0041413-g004){ref-type=\"fig\"}, [Fig. 5B](#pone-0041413-g005){ref-type=\"fig\"}, [Fig. 6A](#pone-0041413-g006){ref-type=\"fig\"}.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Mean MT growth speed and growth excursion lifetimes for cells treated with RNAis.** shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Proportion of MT growth excursions in subpopulations grouped according to growth speed and growth excursion lifetime for cells treated with RNAsi.** shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics. Data shown is depicted graphically in [Figure 2](#pone-0041413-g002){ref-type=\"fig\"}.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Mean MT growth speed and growth excursion lifetimes for cells expressing CA-Rac1 and treated with RNAis.** shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150glued, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Proportion of MT growth excursions in subpopulations grouped according to growth speed and growth excursion lifetime for cells expressing CA-Rac1 and treated with RNAis.** shRNA vectors were used for RNAi targeting of EB1, CLASP2, dynamitin, DCX, MAP1A, MAP1B, MAP2, MAP4, MARK1, MARK2 and MARK3. siRNA oligos were used for RNAi targeting of APC, APC2, ACF7, XMAP215, Op18, p150*^glued^*, CLIP115, CLIP170, STOP, MAP1S, Spastin and Katanin p60. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics. Data shown is depicted graphically in [Figure 3A](#pone-0041413-g003){ref-type=\"fig\"}.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Orientation of MT growth at the leading edge.** shRNA vectors were used for control, and RNAi targeting of EB1 and MARK2. siRNA oligos were used for RNAi targeting of APC2, Op18, p150*^glued^* and Spastin. mKO-EB3 tracks were overlaid on images of mKO-EB3 using PlusTipTracker software and the orientation of tracks within 5 \u00b5m from the leading edge was classified according to their angle relative to the cell edge. Data shown is depicted graphically in Figures. ([Fig. 1G](#pone-0041413-g001){ref-type=\"fig\"}, [Fig. 3C](#pone-0041413-g003){ref-type=\"fig\"}, [Fig. 4G](#pone-0041413-g004){ref-type=\"fig\"}, [Fig. 5F](#pone-0041413-g005){ref-type=\"fig\"}, [Fig. 6E](#pone-0041413-g006){ref-type=\"fig\"}).\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Mean MT growth speed and growth excursion lifetimes in [Figure S1](#pone.0041413.s001){ref-type=\"supplementary-material\"}.** MARK2 shRNA \\#2 vector was used for MARK2 RNAi. Results of analysis of mKO-EB3 time-lapse movies using PlusTipTracker software to measure MT growth dynamics. Data shown is depicted graphically in [Figure S1B and S1C](#pone.0041413.s001){ref-type=\"supplementary-material\"}.\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**EB3 dynamics in U2-OS cells expressing Rac1 mutants.** Spinning disk confocal fluorescence time-lapse image sequences of cells expressing mKO-EB3 to mark growing MT plus ends in cells expressing control non-targeting vector (WT, left), CA-Rac1 (middle), and DN-Rac1 (right). The same cells are shown in [Figure 1F](#pone-0041413-g001){ref-type=\"fig\"}. Bar \u200a=\u200a10 \u00b5m, elapsed time in min: sec shown.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**EB3 dynamics in U2-OS cells treated with shRNAs targeting MARK2.** Spinning disk confocal fluorescence time-lapse image sequences of cells expressing mKO-EB3 to mark growing MT plus ends in cells expressing shRNA targeting MARK2 (left), CA-Rac1 and shRNA targeting MARK2 (middle) or expressing CA-Rac1, shRNA targeting MARK2, and RNAi-resistant GFP-MARK2 (right). The same cells are shown in [Figure 4F](#pone-0041413-g004){ref-type=\"fig\"} and [Figure 5E](#pone-0041413-g005){ref-type=\"fig\"}. Bar \u200a=\u200a10 \u00b5m, elapsed time in min: sec.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**EB3 dynamics in control and MARK2 depleted cells in a monolayer wound edge.** Spinning disk confocal fluorescence time-lapse image sequences of cells expressing mKO-EB3 two hours after scratch-wounding a cell monolayer. Cells treated with control shRNA (left), or MARK2 shRNA (right) at the edge of a monolayer wound. The same cells are shown in [Figure 6D](#pone-0041413-g006){ref-type=\"fig\"}, bar \u200a=\u200a10 \u00b5m, elapsed time in min: sec.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Effects of MARK2 depletion on directed cell migration**. Phase contrast time-lapse image series of U2-OS cells treated with non-targeting siRNA pool (left) or MARK2 siRNA (right). The same cells are shown in [Figure 8C](#pone-0041413-g008){ref-type=\"fig\"}, bar \u200a=\u200a50 \u00b5m, elapsed time after scratch-wounding shown in hr: min.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\nWe thank K. Myers, L. Cassimeris, A. Suzuki, T. Wittmann, R. Heald and the NHLBI cell sorting facility.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This work supported by National Heart Lung and Blood Institute intramural research program National Institutes of Health U01 GM067230. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: CMW YN. Performed the experiments: YN. Analyzed the data: YN. Contributed reagents/materials/analysis tools: GD KA MWD. Wrote the paper: CMW YN.\n"}
+{"text": "INTRODUCTION\n============\n\nAcupuncture has been practiced for thousands of years in Eastern countries to manage pain, and it is now considered as one of the most important strategies in complementary and alternative medicine. Electroacupuncture (EA) is a modified acupuncture technique that utilizes electrical stimulation. Zusanli (ST36) is one of several the acupuncture point that are used for clinical pain control \\[[@B1], [@B2]\\]. Many studies have demonstrated that acupuncture and/or EA stimulation applied to the ST36 acupoint can increase the experimental pain threshold mediated by the release of endogenous opioid peptides \\[[@B3]\\] and/or by activation of the descending pain inhibitory system \\[[@B3]-[@B10]\\].\n\nHowever, there are marked individual variations in the sensitivity to EA, which often cause clinical and experimental difficulties. Previous studies reported that 20\\~40% of rats were insensitive to EA in an acute thermal pain test \\[tail flick latency (TFL) test\\] \\[[@B11]\\]. We previously identified and characterized the genes that cause these differences between the responder rats that were sensitive to EA and the non-responder rats that were insensitive to EA based on the TFL test, using the hypothalamus, a main center of the descending inhibitory system and \u03b2-endorphin secretion \\[[@B12]\\]. Among the genes that were more abundantly expressed in the hypothalamus of responder rats compared to the non-responder rats, dopamine beta-hydroxylase (DBH) gene was significantly higher in the responder rats \\[[@B12]\\]. DBH catalyzes the conversion of dopamine to norepinephrine. Norepinephrine is an essential neurotransmitter of descending pain inhibitory system. Many pharmacological studies indicate that norepinephrine is antinociceptive and potentiate the analgesic effect \\[[@B13]-[@B15]\\].\n\nBased on those observations, we hypothesized that the expression of DBH gene in the hypothalamus modulates EA analgesia in rats. To explore this hypothesis, we constructed a DBH-encoding adenovirus and a control virus expressing only green fluorescence protein (GFP) and evaluated the effects of microinjection of the viruses into the hypothalamus on EA analgesia in rats using the TFL test. Furthermore, we examined the effects of microinjection of the DBH virus into hypothalamus on the body core temperature, body weight, and learning and memory ability, all of which may be influenced by hypothalamic dysfunction.\n\nMETHODS\n=======\n\nAnimals\n-------\n\nMale Sprague-Dawley (SD) rats were housed in groups of four, with water and food available ad libitum in a controlled environment (12:12-hour light:dark cycle, temperature 23\u00b12\u2103, humidity 50\u00b110%). The SD rats were purchased from Orient Animal Corp. (Korea). The experiments began when the rats were seven weeks old, and all of the procedures involving animals were approved by the Institutional Animal Care and Use Committee of Kyung Hee University.\n\nProduction of DBH recombinant adenovirus\n----------------------------------------\n\nThe protocol for generation of recombinant adenovirus has been described elsewhere \\[[@B16]\\]. Briefly, human DBH genes were cloned to the multicloning site of a pShuttle-IRES-CMV vector with *Xho* I / *Sal* I through PCR. The constructs were bicistronic and contained a GFP gene controlled by a separate cytomegalovirus (CMV) promoter. After homologous recombination, the recombinants were selected with kanamycin and screened by restriction enzyme analysis. Recombinant adenoviral DNA was then extracted from the correctly recombined clones, digested with *Pac* I to expose its ITR (inverted terminal repeats) and transfected into packaging 293 cells using Lipofectmine 2000 reagent (Invitrogen, Rockville, USA) to produce viral particles. Viruses were harvested and purified using Vivapure AdenoPACK 500 (Vivagen, Seoul, Korea) according to the manufacturer\\'s instructions. The final titer of the adenovirus was about 2.5\u00d710^9^ plaque-forming units (pfu)/ml. Small-volume aliquots were kept in liquid nitrogen, and fresh aliquots were used for each experiment.\n\nMicroinjection of adenovirus into the hypothalamus\n--------------------------------------------------\n\nSD Rats were positioned in a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA) under isoflurane anesthesia. After making a longitudinal incision of the scalp and disinfecting an exposed dorsal cranium with betadine, a small hole approximately 2 mm in diameter was drilled into the skull over the hypothalamic arcuate nucleus (-3.8 anterior-posterior, 0.5 mediolateral, 9.8 dorsoventral), based on the rat brain atlas Paxinos and Watson (1998). All injections were made using a Hamilton syringe equipped with a 30S gauge beveled needle and attached to a syringe pump (KD Scientific, New Hope, PA). Infusions were made at a rate of 0.2 \u00b5l/min for PBS (4 \u00b5l of sterile PBS; Sigma, St. Louis, MO), for control GFP virus (1\u00d710^7^ pfu/4 \u00b5l) and for DBH virus (1\u00d710^7^ pfu/4 \u00b5l). Following injection, the syringe was left in place for an additional 10 min, after which the needle was removed from the brain very slowly.\n\nMeasurement of TFL and EA stimulation\n-------------------------------------\n\nWe allowed the rats to adapt to handling for at least three weeks to minimize stress so that they would remain keep clam during the TFL testing and EA stimulation \\[[@B17],[@B18]\\]. This acclimation enabled us to eliminate the need for anesthetics or cylindrical restrainers that cause the TFL values to be more variable (unpublished observations). Rats were individually placed on the palm of an experimenter\\'s hand with its tail between the middle and ring fingers, and their head and back were continuously rubbed and stroked so that they became accustomed to being handled. The antinociceptive response was assessed by the TFL test, which records tail-flick latency as the time from the onset of stimulation to the withdrawal of the tail from a heat stimulus produced by a light beam in the TFL test. Three successive assessments of TFL were recorded at time intervals of 1-min and these values were averaged. The cut-off time was set at 15 s to eliminate tissue damage. For EA stimulation, a pair of stainless steel acupuncture needles (0.25 mm in diameter and 3 cm long) was inserted into the Zusanli acupoint (ST36) on the right leg located in the tibialis cranials muscle, 5 mm below lateral to the anterior tubercle of the tibia, and into the point 5 mm distal from the first needle. The two acupuncture needles, which were vertically inserted to a depth of 5 mm, were connected to an electrical stimulator to generate electrical stimulation, and train pulses (2 Hz, 0.5 ms pulse duration) and an intensity of the threshold of muscle contraction (0.2\\~0.3 mA) were then applied for 20 minutes. To determine if adenoviral gene transfer of DBH alone changes the TFL, the baseline was recorded before starting the experiments. We then compared the TFL between the DBH virus-injected (n=16) and GFP control virus-injected rats (n=16) in the absence of EA stimulation at -1d, 3d, 7d and 14d following viral injection. We also compared the increase in the EA analgesia ratio between the DBH virus-injected (n=16) and the GFP control virus-injected (n=16) rats to determine if viral gene transfer of DBH increased EA-induced analgesic effects. Following EA stimulation at -1d, 3d, 7d and 14d after viral injection, the average of three successive TFL assessments was determined. The effects of viral gene transfer on EA analgeisa were then presented as percent changes from the pre-injection TFL (i.e. -1d TFL) according to the following formula:\n\nAssessment of possible side effects\n-----------------------------------\n\nWe evaluated body core temperature (rectal temperature) and body weight in the DBH virus-injected and GFP control virus-injected rats (n=7\\~8, per group) at -1d, 3d and 7d after viral injection to determine if microinjection of the DBH virus into the hypothalamus induced side effects. Learning and memory ability were also tested based on passive avoidance 7d after viral injection as previously described \\[[@B19]\\]. The apparatus consisted of equal-sized light and dark chambers (28\u00d723.5\u00d720 cm) separated by a guillotine door (6\u00d76.5 cm) leading to a runway (platform) at one end and to the grid floor of a dark chamber for electrical foot shock. The retention test session was measured when the rats entered the dark chamber. The hesitation time was measured, with the maximum time being set at 90 s.\n\nTissue preparation\n------------------\n\nAnimals were anesthetized with chloral hydrate (360 mg/kg, i.p.) at 15 days after injection and transcardially perfused with saline solution containing 0.5% sodium nitrate and heparin (10 U/ml), followed by fixation with 4% paraformaldehyde dissolved in 0.1M phosphate buffer (PB). Brains were removed and postfixed overnight at 4\u2103 in buffered 4% paraformaldehyde, and stored at 4\u2103 in 30% sucrose solution until they sank. Brains were frozen sectioned using a sliding microtome into 40 \u00b5m coronal sections and collected in twelve separate series. For Nissl staining, some of the Hypothalamus tissues were mounted on gelatin-coated slides, dried for 1 h at RT, stained with 0.5% cresyl violet (Sigma), dehydrated, coverslipped, and then analyzed under a bright-field microscope (Olympus Optical). The correct transfection of the adenovirus was verified by the presence of green fluorescence in the hyphothalamus under a confocal microscope (Zeiss, Germany).\n\nWestern immunoblot analysis\n---------------------------\n\nFor analyses of DBH expression, the ipsilateral hypothalamus were dissected 7 days after injection with PBS, control GFP or DBH virus in the absence of EA stimulation. Hypothalamic samples were homogenized with ice-cold lysis buffer containing 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 5 mM MgCl~2~, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride and protease inhibitor cocktail (Sigma) in a Dounce homogenizer (Wheaton, Millville, NJ, USA). Proteins (50 \u00b5g/lane) were separated by 10% sodium dodecyl sufate-polyacrylamide gel (SDS-PAGE), transferred to polyvinylidene difluoride membrane (Millipore, Bedford, MA USA) using an electrophoretic transfer system (Bio-Rad Laboratories, Hercules, CA, USA), and subjected to immunoblotting with rabbit anti-DBH (1:500; Santa Cruz Biotechnology, Santa Cruz, CA, USA). The blots were subsequently stripped and re-probed with rabbit anti-actin (1:2000; Cell Signaling Technology, Beverly, MA, USA).\n\nStatistical analysis\n--------------------\n\nThe data were expressed as the means\u00b1SEM. An unpaired t-test or one way ANOVA was used for statistical analyses. In all cases, p values of \\<0.05 were considered to be significant.\n\nRESULTS\n=======\n\nEffects of adenoviral gene transfer of DBH on the basal pain threshold and EA analgesia\n---------------------------------------------------------------------------------------\n\nIn our previous study, we showed that EA stimulated rats were divided into responders that were sensitive to EA and non-responders that were insensitive to EA based on the results of a TFL test. The ratio of responders to non-responders stimulated EA was approximately 3:2. In addition, the TFL increase ratio was 72.3% and 11.9% for responders and non-responders following EA stimulation, respectively \\[[@B12],[@B20]\\]. We previously found that the DBH gene (gene profiling) was more abundantly expressed in the hypothalamus of responders than non-responders based on the TFL test \\[[@B12]\\]. Based on these results, we evaluated to prioritize the effects of adenoviral gene transfer of DBH into the hypothalamus on EA analgesia in responders and non-responders. As a result, both the responder and the non-responder rats showed a significant increase in TFL to a similar extent following microinjection of the DBH virus (data not shown) (i.e. the initial responders became more responsive to EA after the viral expression of DBH and the initial non-responders rats became responders). Therefore, we did not classify the rats into responders or non-responders in the current study.\n\nTo determine if gene transfer of DBH affects the basal pain threshold and analgesic effect of EA stimulation, we injected adenovirus co-expressing the DBH and GFP genes into the hypothalamus. As a control, the adenovirus expressing the GFP gene alone was injected into a different group of rats. Next, the change in TFL or the increase in the ratio of EA analgesia in the DBH virus-injected rats was compared to GFP control virus-injected rats at -1, 3, 7 and 14 days following viral injection. As shown in [Fig. 1A](#F1){ref-type=\"fig\"}, there were no significant differences in the TFL values of the DBH virus-injected and the GFP control virus-injected rats during the 2-week experimental period (t-test; t=0.5, df=30, p=0.618 at -1 day; t=0.34, df=30, p=0.739 at 3 days; t=0.34, df=30, p=0.738 at 7 days, t=1.5, df=30, p=0.139 at 14 days). In addition, TFL values measured at 3, 7 and 14 days after the injection of DBH or control virus were not significantly different from the value at -1 day following the injection (ANOVA; F~3,45~=3.975, p\\>0.05, DBH; F~3,45~=1.230, p\\>0.05, control). [Fig. 1B](#F1){ref-type=\"fig\"} shows the ratio of EA analgesia following microinjection of the DBH or GFP control virus. The EA-induced analgesic effect gradually increased in the DBH virus-injected rats, but not in the GFP control virus-injected rats. A statistically significant increase in EA analgesia was observed from 7 days after injection of the DBH virus (t-test; t=2.4, df=30, p=0.025 at 3 days; t=3.8, df=30, p=0.001 at 7 days; t=2.4, df=30, p=0.024 at 14 days).\n\nTo confirm the expression of the DBH adenoviral gene, we performed confocal microscopic analysis at 15 days after viral injection. The results revealed the localization of GFP expression in the adenoviral injected hypothalamus ([Fig. 2A](#F2){ref-type=\"fig\"}). Moreover, Nissle staining showed the location of the needle tip in the hypothalamus, showing only a little tissue damage, which was primarily limited to the needle track ([Fig. 2B, D](#F2){ref-type=\"fig\"}). The hypothalamic DBH expression was analyzed by western blotting using anti-DBH antibody at 7 days after viral injection. DBH expression in the DBH virus-injected rats was also significantly higher than GFP control virus-injected rats ([Fig. 2C](#F2){ref-type=\"fig\"}).\n\nEffects of adenoviral gene transfer of DBH on body temperature, body weight gain, and learning and memory ability\n-----------------------------------------------------------------------------------------------------------------\n\nBecause the hypothalamus controls a variety of body functions in addition to pain modulation, we determined if microinjection of the DBH virus into the hypothalamus generates side-effects. There were no significant differences in body temperature (rectal temperature) ([Fig. 3A](#F3){ref-type=\"fig\"}) and body weight gain ([Fig. 3B](#F3){ref-type=\"fig\"}) between the DBH virus-injected rats and the GFP control virus-injected rats at -1, 3 and 7 days after viral injection (body temperature in the GFP group versus DBH group: 36.9\u00b10.1\u2103, 37.0\u00b10.2\u2103 and 37.0\u00b10.1\u2103 versus 36.9\u00b10.1\u2103, 37.0\u00b10.1\u2103 and 37.0\u00b10.1\u2103; body weight gain: 288.9\u00b14.1, 325.8\u00b19.1 and 357.1\u00b14.8 g versus 292.0\u00b12.8, 335.7\u00b16.9 and 358.3\u00b15.4 g). In addition, learning and memory ability were assessed by a passive avoidance test administered 7 days after viral injection. The test revealed no signs of memory impairment in the GFP virus-injected and the DBH virus-injected rats because the hesitation time of all rats in both groups was greater than the cut-off time (90 seconds) ([Fig. 3C](#F3){ref-type=\"fig\"}).\n\nDISCUSSION\n==========\n\nAlthough it has been difficult to establish the biological basis of acupuncture, it has long been a unique method used to treat various diseases and to relieve pain in East Asia, and it is a common complementary and alternative medicine in Western countries \\[[@B21]\\]. Previous studies have shown that acupuncture or EA increases the experimental pain threshold in animals as well as humans and that EA is more effective than acupuncture alone \\[[@B6],[@B9],[@B22]\\]. More recently, reasonable neurophysiological theories have been reported that signaling molecules, such as opioid peptides, glutamate, 5-hydroxytryptamine, adenosine, and cholecystokinin octapeptide, are involved in the acupuncture induced analgesia \\[[@B5],[@B8],[@B23],[@B24]\\]. Specifically, low frequency EA stimulation, such as the 2 Hz EA that was used in the present study, activates the arcuate nucleus of the hypothalamus leading to the release of beta-endorphin and subsequent activation of the downstream descending inhibitory system \\[[@B3]\\].\n\nThe differences in the response to pain and analgesics have been studied in many research fields and the results of these studies have indicated that it might be attributed to genetic factors \\[[@B25]-[@B27]\\]. However, many of these studies have been conducted to evaluate only a single factor. Therefore, we conducted a cDNA microarray experiment to assay the expression of thousands of genes in parallel and identified the genes that were differentially expressed in the hypothalamus of responder and non-responder rats that received EA analgesia. The microarray analysis revealed that the DBH gene was expressed in abundance in the responder group. In addition, brain factor-1, which is a transcription factor that is essential for the proliferation of progenitor cells in the cerebral cortex \\[[@B28]\\], and the acetylcholinesterase T subunit, which is the only catalytic subunit involved in all vertebrate cholinesterases \\[[@B29]\\], was also expressed abundantly in the responder group. Conversely, the neurodegeneration markers, tau and cathepsin B \\[[@B30],[@B31]\\], were abundantly expressed in the non-responder group. In a previous study, we showed that the microinjection of the acetylcholinesterase T (AChET) adenovirus into the hypothalamus significantly enhanced the EA analgesia of without side-effect on the body core temperature, body weight, or learning and memory ability \\[[@B20]\\].\n\nThe present study focused on the relationship of EA analgesia and DBH expression in the hypothalamus. DBH is the catalytic enzyme involved in the conversion of dopamine to norepinephrine, which is a key neurotransmitter of the endogenous pain inhibitor system. Recent report showed that norepinephrine played an important role in antinociceptive modulation of hypothalamic paraventricular nucleus (PVN), with the result that pain stimulation increased endogenous norepinephrine and microinjection of norepinephrine into the PVN elevated pain threshold \\[[@B13]\\]. Moreover DBH deficiency caused chronic thermal hyperalgesia associated with a lack of norepinephrine \\[[@B32]\\].\n\nTo explore the role that the expression of DBH in the hypothalamus plays on EA analgesia in vivo, we constructed an adenovirus that expressed DBH and GFP control virus. The data obtained in this study revealed that there was no difference in the TFL value of DBH virus-injected and control virus-injected rats, and that the post-injection TFL values did not differ from the pre-injection TFL values in either group ([Fig. 1A](#F1){ref-type=\"fig\"}). These findings indicate that the basal pain threshold was not changed by adenoviral gene transfer of DBH in the hypothalamus. However, the present study clearly demonstrates that microinjection of the DBH virus, but not the control virus, potentiates the analgesic effect of EA ([Fig. 1B](#F1){ref-type=\"fig\"}). Taken together, these findings suggest that adenoviral expression of DBH in the hypothalamus enhances EA analgesia without affecting the basal pain threshold. The influence of the hypothalamus on body physiology is diverse. Indeed, it controls a variety of body functions, such as the autonomic nervous system and the feeding and neuroendocrine systems. Therefore, we evaluated microinjection of the DBH virus into the hypothalamus to determine if it generated side-effects that might disturb the exact analysis of the TFL data ([Fig. 3](#F3){ref-type=\"fig\"}). The results showed that DBH expression in the hypothalamus does not affect the body core temperature, body weight, or learning and memory ability, which suggests that viral gene transfer of DBH into the hypothalamus produces no apparent side-effects.\n\nCurrently, we cannot sufficiently explain why the results of basal pain threshold after manipulations of the DBH and norepinephrine in the present study and the others are different. Contradictory to our data showing no changes in baseline pain threshold in the TFL test after microinjection of DBH virus into the hypothalamic arcuate nucleus, DBH knockout mice showed thermal hyperalgesia in the hotplate (50\u2103) and cold-plate (4\u2103) test, as compared to their littermate controls \\[[@B32]\\]. In addition, microinjection of norepinephrine into the rat hypothalamic PVN dose-dependently elevated basal pain threshold measured by the potassium iontophoresis-induced tail flick test \\[[@B13]\\]. The differences in the manipulation site of the hypothalamus (arcuate nucleus vs. PVN), in the manipulation method (viral gene transfer vs. direct acute injection or genetic deletion resulting in DBH KO of the whole nervous system) and in the strain used (WT rats vs. KO mice) may contribute to such contradictory results. Specifically, the role of the hypothalamic arcuate nucleus in mediating EA analgesia is more related to the endotenous opioids (i.e. \u03b2-endorphin) \\[[@B33],[@B34]\\]. Thus, it would be very interesting to identify a functional interrelationship between the noradrenergic system and opioid system in the hypothalamic arcuate nucleus during EA analgesia.\n\nEven though many reports have indicated that the supraspinal mechanisms are involved in the TFL test, it is mainly based on a spinal reflex \\[[@B35]\\]. In addition, it would be of high importance to investigate the analgesic effects of EA on chronic pathological pain, such as neuropathic pain \\[[@B36]\\], the mechanism of which is somewhat different from acute nociceptive pain like the TFL test. Thus, further studies using non-reflexive (operant) pain assays, spontaneous pain behavior tests and chronic pain models \\[[@B37]\\] might provide a better understanding of EA-induced analgesia.\n\nWe found that expression of DBH in the hypothalamus is closely associated with individual differences in response to EA analgesia. Although, the functional role that DBH in the hypothalamus plays in the analgesic effect of EA is unclear, our findings suggest that expression of the DBH gene in the hypothalamus promotes EA analgesia without obvious side-effects. Taken together, these data provide a new perspective regarding the role that peripheral somatic stimulation plays in pain control.\n\nThis work was supported by the SRC/ERC program of MOST/KOSEF (R11-2005-014).\n\nDBH\n\n:   dopamine beta-hydroxylase\n\nTFL\n\n:   tail flick latency\n\nEA\n\n:   electroacupuncture\n\n![Effects of adenoviral gene transfer of DBH on the basal pain threshold and EA analgesia. (A) Basal TFL values in the DBH virus-injected and GFP control virus-injected rats at -1, 3, 7 and 14 days after viral injection. Data are presented as the mean\u00b1SEM. p\\>0.05, DBH versus GFP group at each day as determined by the unpaired t-test; p\\>0.05, -1 day TFL versus 3, 7 and 14 days TFL in either group as determined by one-way ANOVA. (B) The increase ratio of EA-induced analgesic effect at 3, 7 and 14 days following viral injection, compared to the EA effect prior to viral injection. ^\\*^p\\<0.05 and ^\\*\\*\\*^p\\<0.001, between the two groups as determined by the unpaired t-test. The data are the means\u00b1S.E.M (n=16/group).](kjpp-17-505-g001){#F1}\n\n![Verification of the correct injection and infection of the adenovirus into the hypothalamus. Tissues were prepared 15 days after PBS or virus injection. (A) Representative confocal microphotographs of GFP fluorescence in the hypothalamus from DBH virus injected rats. 3v, 3rd ventricle. Scale bar, 200 \u00b5m. (B) Representative photographs of the Nissle staining following microinjection of the viral suspension. Arrow indicates needle track. (C) Western blot analysis showing significantly increase of DBH in the DBH virus-injected rat compared to PBS or GFP control virus injected rats at 7 days later virus injection. (D) High magnification of B.](kjpp-17-505-g002){#F2}\n\n![Effects of adenoviral gene transfer of DBH on body temperature, body weight, and learning and memory ability. The body temperature (A) and body weight (B) of the DBH virus-injected and the GFP control virus-injected rats at -1, 3 and 7 days following viral injection (n=7\\~8/group). (C) The learning and memory ability in the passive avoidance test at 7 days after viral injection. The data shown are the means\u00b1S.E.M (n=7\\~8/group).](kjpp-17-505-g003){#F3}\n"}
+{"text": "INTRODUCTION\n============\n\nSystemic lupus erythematosus (SLE) is an autoimmune disease involving multiple systems, including the nervous, circulatory, pulmonary, renal, and immune systems.^[@r01],[@r02]^ SLE is prevalent among younger individuals, primarily those aged in their 20s to 40s, and is more frequent in women than men.^[@r03]^ SLE is often complicated with seizures, renal insufficiency, and anemia, which substantially reduce quality of life.^[@r04],[@r05]^ In order to improve management of SLE, it is necessary to clearly understand the prevalence of the disease, including the clinical and socioeconomic characteristics of SLE patients.^[@r06]--[@r08]^ However, no population-based prevalence survey of SLE in South Korea has so far been conducted.\n\nCohort studies involving primary data collection have been used to estimate SLE prevalence, but such studies tend to be costly and time-consuming.^[@r09],[@r10]^ Alternatively, population-based administrative databases, containing physician billing and insurance information, have received increased attention for their potential to provide epidemiologic information, particularly for rare conditions like SLE.^[@r11]--[@r13]^ Along with reduced costs and time requirements, administrative data also offer the advantage of simplicity in establishing and maintaining a population-based surveillance system. However, optimal methods for extracting information from these databases have yet to be determined.^[@r14],[@r15]^ Researchers have therefore called for further studies regarding the usefulness of administrative sources such as physician claims databases and insurance databases.^[@r08],[@r16],[@r17]^ To this end, studies concerning the positive predictive value (PPV) of diagnoses based on administrative data are required, especially for complex conditions such as SLE.\n\nKorea has a medical insurance system, in which all citizens are required to join the national public insurance system and pay monthly medical insurance premiums. The information from medical practices is then collected by the National Health Insurance (NHI) system. This delicate process is strictly governed by the Health Insurance Review and Assessment Service (HIRAS). Therefore, national epidemiological data can be easily collected by analyzing the NHI payment request data from HIRAS. However, this database had a minor problem to overcome. In the past, the International Classification of Diseases, Tenth Revision (ICD-10) code for SLE diagnosis, M32, was often given tentatively to suspected SLE patients before the year of 2009. As such, the PPV of the M32 diagnosis was uncertain before 2009. We tried to estimate the prevalence of SLE using the national database from 2004 to 2006, which used the M32 code (probable SLE diagnosis). We calculated the \"raw\" prevalence of SLE from NHI data and then verified the SLE-coded diagnosis as follows: 1) verification of randomly selected SLE-coded samples; 2) verification of all registered SLE-coded patients from more than 56 hospitals nationwide; and 3) application of a PPV, obtained from subgroups of a sample institute, to the NHI database. Then, the calculated PPV was applied to adjust the raw prevalence, thereby estimating the prevalence of SLE in South Korea.\n\nRecently, the South Korean government enacted a law requiring physicians to make a definitive diagnosis to bestow better insurance benefits on only patients whose diagnosis is confirmed to be SLE. Since this measure was introduced, more accurate data on the prevalence of SLE diagnosis have become available.\n\nMATERIALS AND METHODS\n=====================\n\nData assessment and acquisition of HIRAS data\n---------------------------------------------\n\nSLE data were compiled from resources of HIRAS and provided to the research team for primary analysis. This process was officially permitted by the Korean Food and Drug Administration (KFDA) and HIRAS. The HIRAS statisticians compiled data on all patients from institutes nationwide who had been treated under the ICD-10 code of \"M32\" for 3 years from June 2004 to June 2007. Due to enforced privacy laws, detailed personal information was unable to be retrieved; instead, each patient was allocated an identification number. Data extraction and compilation processes are shown in Figure [1](#fig01){ref-type=\"fig\"}. The database contained information on service code (admission or outpatient clinic), year (billing year), age (year old), region (province), sex (male or female), hospital (private clinic, hospital, university hospital), department (internal medicine, dermatology, and so forth), admission (admission or not), and billing number.\n\n![Three algorithmic approaches were adopted for the calculation of SLE prevalence, and the outcomes of each method were compared. The first method was to randomly select M32-coded patients in nationwide hospitals (A. Random sampling). Diagnostic accuracy was then evaluated by analysis of clinical data with verification on 2000 patients' records. The second method was to register all possible patients obtained from rheumatologists in the Korean Rheumatism Association (B. All registered patients). A total of 3504 patients were registered, and their patient records were reviewed. Diagnostic accuracy was also delineated by analysis of clinical data in patient records. The third method was to exclude patients least likely to have SLE and include patients most likely SLE based on chronology of insurance claim issuing pattern (C. Interpreting HIRAS database).](je-24-295-g001){#fig01}\n\nAnalysis of HIRAS data\n----------------------\n\nThe HIRAS database is relatively credible and well-grounded for its use in managing the national statistical resources. Nevertheless, apparent limitations of the database stem from lack of accuracy of diagnosis, because the original purpose of the data was to manage national health insurance rather than for use in academic evaluation. To overcome this limitation, verification of the SLE-coded patients was necessary; if the PPV of the M32 code could be deduced, then the actual prevalence of SLE could be estimated. However, since the law restricts revelation of personal and institutional identifications, the diagnostic PPV of the M32 code could not be confirmed using this method.\n\nVerification processes of M32-coded group\n-----------------------------------------\n\nOur initial plan was to randomly select a relatively small number of patients who were coded M32 and to review their medical records in order to validate their diagnosis. However, legal restriction regarding the protection of personal information mandated other alternatives, as follows: 1) verification after random sampling, which would randomly sort regional distribution, sex ratio, hospital grades, and age patterns of the original database; 2) verification after recruiting as many M32 patients as possible with the help of the 150 rheumatologists in the Korean Rheumatology Association; and 3) sub-grouping of M32 patients from the original SLE data and matching diagnostic PPVs between subgroups and Korean reference groups. Detailed explanations about randomization, data gathering, and analysis are presented in Figure [1](#fig01){ref-type=\"fig\"}.\n\nVerification of the patients who were randomly sampled\n------------------------------------------------------\n\nA total of 2000 patients were randomly selected and the validity of their diagnostic accuracy was analyzed (Figure [1](#fig01){ref-type=\"fig\"}A). We sought to investigate 15% of all M32-coded patients from each institute as an optimal proportion. Random sampling was done according to the ratio of regional differences and types of hospitals, with 868 patients (43.4%) in Seoul, 270 patients (13.5%) in Kyounggi province, 212 patients (10.6%) in Pusan, 126 patients (6.3%) in Daegu, and so on. In terms of hospital types, 1308 patients (65.4%) were in university hospitals, 358 patients (17.9%) were in general hospitals, 18 patients (0.9%) were in local hospitals, and 316 patients (15.8%) were in private clinics. Identifications of clinics were provided by the Korean Rheumatology Association. Before experienced rheumatologists were sent to each institute, randomization of patients was verified by a statistician. In addition, patient lists were sent to the institutes in advance so that their charts could be prepared for the verification process. National prevalence of SLE per 100 000 was based on the population in 2006 (<http://kosis.nso.go.kr/Magazine/NEW/YD/VD00013.xls>).\n\nVerification of all registered patients\n---------------------------------------\n\nAfter verification of randomly selected M32-coded patients, we decided to recruit as many M32-coded patients as possible to compensate for the limitations of random sampling (Figure [1](#fig01){ref-type=\"fig\"}B). Since SLE is relatively rare and sometimes fatal, most patients are referred to experienced rheumatologists. We sent official letters to every rheumatologist who was affiliated with the Korean Rheumatology Association. As a result, more than 160 rheumatologists participated in the survey, and a total of 3504 M32-coded patients were recruited for verification.\n\nApplication of PPV of sample groups to the national database\n------------------------------------------------------------\n\nPatterns of insurance coverage, frequency of a patient's visits, type of healthcare provider, billing, and age are significant factors in the accurate diagnosis of SLE (Figure [2](#fig02){ref-type=\"fig\"}). Most importantly, an annual clinic visit is essential for SLE patients. We categorized 7 groups of M32-coded patients from the database according to number of annual visits to a clinic over three consecutive years. In the Group \"abc\", patients visited clinics for three consecutive years, indicating a high probability of a correct SLE diagnosis. Group \"bc\" and \"c\" meant that patients visited a clinic continuously till the time of this investigation for 2 years and 1 year, respectively. Diagnosis of SLE seemed less accurate in other groups such as group \"ab\" and \"a\", whose patient visits were remote, suggesting much lower probability of SLE. We calculated the diagnostic PPV of 7 groups in our institute that were applied to the national database (Figure [1](#fig01){ref-type=\"fig\"}C).\n\n![A. Database build-up process from government insurance data pool. Data were requested on M32-coded patients, including duration of disease, regional code, year of insurance claim, age, sex, hospital type (university hospital, general hospital, regional hospital, private clinic), department (internal medicine, neurology, pediatrics, etc), and type of treatment (admission, outpatient). Data were extracted from HIRAS and modified by statistical teams. Duplicate data were eliminated, and the finalized database was provided to our research team. B. The number of patients by year (2004, 2005, and 2006) when the medical records were issued in a single institute. \"a, b, c\" equals years 2004, 2005, and 2006, respectively. \"abc\" signifies the group of SLE patients with the diagnosis of SLE who were treated in a single institute in consecutive years of 2004, 2005, and 2006. \"ac\" signifies patients with the SLE diagnosis treated in 2004 and 2006. \"a\" signifies that the M32 code was used for only one year, 2004. C. The number of patients according to the year 2004, 2005, and 2006 when the bills were issued. \"a, b, c\" equals the years 2004, 2005, and 2006, respectively. \"abc\" signifies the group of SLE patients whose M32-coded insurance billings were issued in consecutive years of 2004, 2005, and 2006. \"ac\" signifies patients with the SLE diagnosis billed in 2004 and 2006. \"a\" signifies that the M32 code was used for only one year, 2004, for insurance billing. SLE, systemic lupus erythematosus; HIRAS, Health Insurance Review and Assessment Service.](je-24-295-g002){#fig02}\n\nApplication of estimated prevalence of SLE to isolated regional database\n------------------------------------------------------------------------\n\nWe finally figured out SLE prevalence via three different methods. To verify whether this estimate is realistic or not, we compared this estimate with actual local prevalence in an isolated area. We accessed the database of Jeju Island, where two hospitals usually take care of SLE patients. We analyzed the data of these two hospitals and calculated the PPV of patients with an M32 code actually having SLE. We then calculated local prevalence on the island by the number of diagnosed patients multiplied by PPV and divided by the entire island population. This local prevalence was compared with the estimate from the national database.\n\nThe pilot study\n---------------\n\nThe pilot study was conducted in 7 hospitals affiliated with our institute. An initial data collection form for analysis was prepared and verified in the pilot study; however, preparing the data collection form was arduous and time-consuming. Even after careful evaluation of the medical record, the retrieved data were insufficient for the survey. Hence, a simple data collection form to register minimal yet essential SLE information was created. In the process of generating the data collection form, patients' social security numbers were partially obscured to protect personal information and privacy.\n\nThe data collection form and definition of SLE\n----------------------------------------------\n\nThe modified, simple data collection form solicited 11 types of patient information, which were the 1982 revised criteria of SLE^[@r01]^: presence of malar rash, discoid rash, photosensitivity, oral ulcer, arthritis, serositis, renal disorder, hematologic disorder, neurologic disorder, immunologic disorder, and positive fluorescent anti-nuclear antibody (FANA) test. Diagnosis of SLE was considered to be confirmed when 4 out of 11 criteria were satisfied.\n\nStatistical analysis\n--------------------\n\nWe had problems calculating the confidence intervals (CIs) shown in [Table](#tbl01){ref-type=\"table\"}. The 95% CI should be obtained from prevalence data itself. However, our research environment prevented us from accessing full sets of data. During the estimation of SLE prevalence, we chose to calculate CIs of PPV and then multiplied by total population of M32-coded. $$\\begin{array}{l}\n{\\text{95}\\%{CI~of~fraction~is}} \\\\\n{\\lbrack p - 1.96*{sqrt}\\left\\{ p(1 - p)/n \\right\\},p + 1.96*{sqrt}\\left\\{ p(1 - p)/n \\right\\}\\rbrack} \\\\\n\\end{array}$$ where \"p\" is \"PPV,\" derived from the investigation of medical records of patients, and \"n\" is the number of patients whose medical records were reviewed.\n\n###### Estimated SLE prevalence according to three different ascertainment methods\n\n  ------------------------------------------------------------------\n  Algorithms            Random\\        All registered\\   Subgroup\\\n                        sampling       patients          analysis\n  --------------------- -------------- ----------------- -----------\n  Estimated patients\\   9167\\          9533\\             10 633\n  (95% CI)              (8886--9447)   (9326--9741)      \n\n  Prevalence per\\       18.8\\          19.5\\             21.7\n  100 000 people\\       (18.2--19.3)   (19.1--19.9)      \n  (95% CI)                                               \n  ------------------------------------------------------------------\n\nSLE, systemic lupus erythematosus; CI, confidence interval.\n\nWe thought that random sampling and the method of analyzing all registered M32 patients were more direct means of estimating PPV and could be compared with other potential methods of subgroup analysis. All statistical analyses were weighted to the Korean population to provide nationally representative estimates. We used the mid-year estimated population from 2006, provided by National Statistical Office website (<http://www.kosis.kr>). SAS (version 8.12; SAS Institute, Cary, NC, USA) was used for random sampling. Results were analyzed with SAS and Excel (version 2003; Microsoft Corporation, Redmond, WA, USA).\n\nRESULTS\n=======\n\nBasic characteristics of database subjects\n------------------------------------------\n\nAccording to the NHI database, the number of insurance claims with the M32 code incrementally increased from 2004 to 2006. Specifically, female patients increased by 1000 annually, while male patients increased by 100 annually (Figure [3](#fig03){ref-type=\"fig\"}A). The female-to-male ratio was 89.4:10.6. The 30- to 39-year-old age group was the most affected, followed by 40--49, 20--29, and 50--59 (Figure [3](#fig03){ref-type=\"fig\"}B). Korean SLE patients were managed in university hospitals (70%), general hospitals (19%), private clinics (18%), and regional hospitals (1%). Over 90% of SLE patients visited a rheumatology clinic, while approximately 3% of SLE patients went to either dermatology clinics or pediatric clinics.\n\n![A. Annual increase in insurance claims with the M32 code in South Korea. Male to female ratio is about 10 to 1. The number of male and female patients increased each year. B. The peak prevalence is located in the age group of 30- to 39-year-olds. The population demonstrates a normal distribution around the peak.](je-24-295-g003){#fig03}\n\nDiagnostic PPV calculated in patients with SLE-coded by randomization\n---------------------------------------------------------------------\n\nAs described in Figure [1](#fig01){ref-type=\"fig\"}A, we randomly selected 100 patients from private clinics, 400 patients from 13 general hospitals, and 1500 patients from 30 university hospitals in order to calculate the PPV of diagnosis in each of these settings. The PPV of diagnosis in private clinics was 0.65, while diagnostic PPVs in general hospitals and university hospitals were 0.70 and 0.76, respectively. Calculation of the prevalence was performed using the following equation: $$\\begin{array}{l}\n{Presumed\\, number\\, of\\, SLE\\, patients} \\\\\n{= Total\\, number\\, of\\, M32\\, insurance\\, claims*PPV} \\\\\n\\end{array}$$ where the total number of insurance claims concerning M32 was 13 652, and the PPV was the actual number patients who meet the criteria divided by 2000. Among the random sample of SLE patients, the calculated PPV was 0.6715 (95% CI 0.6509--0.6920), resulting in a prevalence estimate of 9167 (95% CI 8886--9447).\n\nDiagnostic PPV calculated among SLE-coded patients from all institutes which participated in nationwide survey\n--------------------------------------------------------------------------------------------------------------\n\nAs displayed in Figure [1](#fig01){ref-type=\"fig\"}B, every accessible M32-coded patient was investigated by rheumatologists contacted by the Korean Rheumatism Association. Twenty-two hospitals participated in this survey and registered 3504 SLE patients. With the previously calculated PPV of 69.8% in this setting, the prevalence was deduced using the following equation: $$\\begin{array}{l}\n{Presumed\\, number\\, of\\, SLE\\, patients} \\\\\n{= Total\\, number\\, of\\, M32\\, insurance\\, claims*PPV} \\\\\n\\end{array}$$ where the total number of insurance claims concerning M32 was 13 652, and the PPV was the actual number of patients who meet the criteria divided by 3504. Among all participants seen by study rheumatologists, the PPV was 0.6983 (95% CI 0.6831--0.7135), resulting in an estimated prevalence of 9533 (95% CI 9326--9741).\n\nDiagnostic PPV of a single institute applied to the national database after subgroup analysis\n---------------------------------------------------------------------------------------------\n\nAs displayed in Figure [1](#fig01){ref-type=\"fig\"}C, a single institute, in which patient data could be investigated thoroughly, was chosen for sub-grouping according to clinic visitation patterns for the past 3 years (Figure [2](#fig02){ref-type=\"fig\"}B). We then calculated PPVs for each subgroup. Sub-grouping was concurrently performed in the original HIRAS database (Figure [2](#fig02){ref-type=\"fig\"}C) so that the previously calculated subgroup PPVs could be applied to the nationwide database. The PPVs of subgroups abc, bc, ac, and c, which were considered most likely to be true SLE patients, were 0.90, 0.76, 0.36, and 0.57, respectively. The prevalence of this algorithm was estimated using the following equation: $$\\begin{array}{l}\n{Presumed\\, number\\, of\\, SLE\\, patients} \\\\\n{= 7842*\\alpha + 1672*\\beta + 260*\\gamma + 3878*\\delta} \\\\\n\\end{array}$$ where \u03b1 is PPV of group abc calculated from one institute, \u03b2 is PPV of group bc, \u03b3 is PPV of group ac, and \u03b4 is PPV of group c. With subgroup analysis, estimated number of SLE patients was 10 633. Estimated number of SLE patients is about 10% more than other two methods.\n\nAlthough prevalence was estimated by three different methods, the results were similar ([Table](#tbl01){ref-type=\"table\"}). The total number of SLE patients was between 9000 and 11 000, and the prevalence of SLE was estimated to be 18.8--21.7 per 100 000 people in South Korea. Furthermore, after estimating SLE prevalence via three different methods, these estimates were compared to local SLE prevalence. Prevalence of SLE on the isolated Jeju Island was also calculated as 20.2 per 100 000 people, a similar estimate to the three shown in [Table](#tbl01){ref-type=\"table\"} (Figure [4](#fig04){ref-type=\"fig\"}).\n\n![Calculation of the prevalence of SLE among the population of the isolated Jeju Island. Estimated SLE prevalence was again tested by the validation of M32-coded patients on the basis of the local island population. Local population based-estimation of SLE prevalence corresponded closely to the prevalence estimated from the nationwide database deduced by three different analyses.](je-24-295-g004){#fig04}\n\nDISCUSSION\n==========\n\nThis study is the first to estimate the prevalence of SLE in South Korea. Until now, most of the data regarding SLE prevalence have been based on surveys of western countries,^[@r03],[@r18]--[@r20]^ and relatively old ones at that, although there are a few North American and European studies that have published data on unselected populations in the last 10 years.^[@r06],[@r07]^ Results of a previous analysis estimated the prevalence of SLE to be about 14.6--50.8 per 100 000 people. In the recent US study of unselected populations, those cases were obtained from clinical records. Those studies estimated the prevalence to range from 78.5 (95% CI 59.0--98.0) per 100 000 people^[@r21]^ to 124 (95% CI 40.0--289.0) per 100 000 people.^[@r22]^\n\nOur estimated prevalence, which was calculated using a single source of government data, was about 20 per 100 000 people. Although there may be ethnic differences in SLE prevalence between Korea and western countries, estimation by administrative data alone may not be a very sensitive means of identifying cases. Our methods would have underestimated SLE prevalence in South Korea if, for example, patients did not seek SLE-related medical care over the three consecutive years on which our prevalence study was focused. Though rare, it is likely to result in the under-ascertainment of milder forms of SLE. Omission of SLE records is also possible, especially when the medical care is primarily focused on diseases other than SLE. In a study using US Medicare data, Katz and colleagues compared diagnosis of SLE appearing in physician claims data with those appearing in medical records, which estimated the sensitivity of claims data to be 85% (95% CI 73%--97%).^[@r14]^ While the present study focused only on the billing data of rheumatologists, a more recent study by Nightingale and colleagues discussed the prevalence of SLE according to study duration.^[@r23]^ Their report pointed out that the likelihood of detecting SLE cases increased in proportion to the length of a patient's contribution to the observational study, claiming that the incremental increase in SLE prevalence over time was, in fact, not a true outbreak of SLE, but an observational artifact.^[@r14]^\n\nRecently, we asked the HIRA official statistical department of the South Korean government to provide the prevalence of SLE-coded patients who were diagnosed after 2009. HIRAS received 16 585 insurance claims reported with M32-coded cases in 2010. There was an annual increase of about 600--800 new cases from 2004 to 2006, while our final prevalence was 10 615 in 2006. If this annual increase continued to 2010, 4000--5000 new cases could be added to the 2006 prevalence data. Extrapolating these increases from the 2006 data resulted in similar numbers of patient cases reported in 2010 (16 585) and patients estimated from the 2006 database (around 15 000).\n\nThe annual increase of M32-coded patients in Figure [3](#fig03){ref-type=\"fig\"}A has several possible explanations: 1) decreased mortality or improvements in diagnostic methods might have led to an increase in new SLE cases; 2) increased numbers of rheumatologists might have enabled the diagnosis of SLE in previously underserved rural areas; or 3) environmental factors may have increased the actual incidence. The annual increase was seen primarily in patients aged 20--60 years, but was not prominent in groups less than 19 years old or more than 70 years old.\n\nIn observational studies, particularly when using administrative databases to identify cohorts of SLE patients, careful contemplation of the data source and consideration of alternative algorithmic definitions of the disease are imperative. The point estimates of prevalence can differ considerably depending on which approach is adopted for SLE detection, hampering determination as to whether or not one approach has greater validity than another. For this reason, we adopted three different algorithms to minimize over- or under-estimation of the true prevalence and obtained similar values in all attempted approaches.\n\nAnother strength of our study is that the presented prevalence incorporates the most widespread data available in South Korea to date. The distribution of our survey was extended to include many institutes in rural areas, where previous statistical investigations have omitted collection of data due to the distance and lack of resources. The rationale for our extensive survey was based on the fact that: (1) relatively few medical institutes provide specialized SLE care; (2) relatively few rheumatologists are in charge of SLE patients; and (3) despite the low number of specialized health care providers, the population of an isolated area should be sufficiently large for a valid analysis.\n\nAfter calculating SLE prevalence via three different methods, we sought to verify whether or not these estimates are applicable to local prevalence. Our estimated prevalence in the isolated area of Jeju Island was 20.2 out of 100 000 people, which was similar to the results reported for national prevalence (Figure [4](#fig04){ref-type=\"fig\"}).\n\nIn summary, this study is the first attempt to estimate the SLE prevalence in South Korea, finding a prevalence of SLE in 2006 of 18.8--21.7 per 100 000 people. This study constructed three feasible algorithms to ascertain SLE cases in administrative data and validated these algorithms with a peer review process. Administrative databases may be a useful source of information in observational studies of patients with SLE. Since no method of case ascertainment can be absolutely accurate, there obviously is a risk of error in administrative data, which may have contributed to the variation of estimates in our study. Further research is needed to develop a disease model and an algorithm for the accurate estimation of the prevalence of rare autoimmune diseases.\n\nThis research was supported by a grant of the Korea Healthcare Technology R&D Project, Ministry of Health, Welfare & Family Affairs, Republic of Korea \\[A092258\\]. This surveillance was organized and supported by Korean Lupus Study Group.\n\nConflicts of interest: None declared.\n"}
+{"text": "Introduction\n============\n\nIn 1966, Worth and Wollin described a condition of hyperostosis corticalis generalisata which was dominantly inherited \\[[@B1]\\]. It is radiologically similar to the autosomal recessive condition called van Buchem syndrome, and some authors refer to the two conditions as endosteal hyperostosis \\[[@B2]\\]. Facial dysmorphism and diaphyseal radiographic changes are present by adolescence and consist of an elongation of the mandible and an increased gonial angle. The forehead becomes flattened and there is a slowly enlarging osseous prominence of the hard palate (torus palatinus). The early radiographic changes include thickening of the endosteum of the long bones and the skull. A progressive increase in the density of the posterior elements of the spine has been noted in some patients \\[[@B3]-[@B5]\\].\n\nMore than 13 kindreds with endosteal hyperostosis have been reported. Four families out of 13 had autosomal dominant inheritance, including male-to-male transmission. Only two patients have been reported in the literature with a history of fractures \\[[@B6]\\].\n\nWe present a father and daughter with endosteal hyperostosis. They were unusual in that both had fractures, odontoid hyperplasia (subclinical basilar invagination), and a simultaneous process of anterior longitudinal spinal sclerosis along the thoracic vertebrae, associated with progressive sclerosis of the posterior elements of the spine. We found no previous reports describing this constellation of spinal abnormalities in association with autosomal dominant, endosteal hyperostosis.\n\nCase presentation\n=================\n\nPatient 1\n---------\n\nA 6-month-old-girl was referred to the department of paediatric orthopaedics because of a fracture of the right humerus. The child was born at full term, the product of an uneventful gestation. At birth her weight, length and head circumference were around the 50^th^percentile. Her mother was a 25-year-old gravida 1, abortus 0, married to a 26-year-old-unrelated father.\n\nThe child\\'s development was normal and there was no history of serious illnesses. Clinical examination showed extensive flattening of the posterior aspect of the skull, which was brachycephalic. She had a flat face with mild frontal bossing, small and deeply set eyes, and low-set ears. Musculo-skeletal examination showed normal musculature and no associated anomalies. Blood biochemistry was normal. A lateral skull x-ray showed extensive flattening across the posterior skull and marked vault and convolutional sclerosis (figure [1](#F1){ref-type=\"fig\"}). An AP radiograph of the right humerus showed mid-diaphyseal endosteal hyperostosis (arrow) and fracture (figure [2](#F2){ref-type=\"fig\"}). The lateral spine radiograph showed no platyspondyly, but marked sclerosis of the entire vertebral rim circumference and unusual enlargement of the spinous processes (figure [3](#F3){ref-type=\"fig\"}).\n\n![**(patient1)**. A lateral radiogram showed extensive flattening across the posterior skull and marked vault and convolutional sclerosis.](1752-1947-1-142-1){#F1}\n\n![**(patient 1)**. An anteroposterior radiograph of the right humerus showed mid-diaphyseal fracture and unusual diaphyseal endosteal hyperostosis (arrow).](1752-1947-1-142-2){#F2}\n\n![**(patient 1)**. The lateral spine radiograph showed no platyspondyly, but marked sclerosis of the entire vertebral cortical rim (arrow) and unusual enlargement of the spinous processes.](1752-1947-1-142-3){#F3}\n\nPatient 2\n---------\n\nA 26-year-old patient, the father of patient I, had a total of six fractures from early childhood till the preadolescent period, involving the clavicles and humerus, but none of the lower limb bones. Thereafter no fractures were reported. Clinical examination showed a man of normal height and normal phenotype. He had normal sclera, teeth and hearing. He had ligamentous stiffness but no muscle wasting or myopathic features. Rigidity over the vertebral column, particularly over his kyphotic thoracic spine was notable. His limbs were not bowed.\n\nRadiographic examination of the femora showed thick, and dense, endosteal hyperostosis more marked at the mid-diaphyses. The medullary canals were narrow but patent, and the endosteal surface was irregular (figure [4](#F4){ref-type=\"fig\"}). Lateral radiogram of the skull shows thick and sclerosed skull vault-arrow- (fig [5](#F5){ref-type=\"fig\"}). Sagittal reformatted multiplanar computed tomography of the craniocervical junction showed a hyperplastic odontoid process. The tip of the dens projected 4.8 mm above a line joining the back of the hard palate to the lowest point of the occipital squama (McGregor line). Subclinical basilar invagination was therefore present. Note the increased bone density of the posterior vertebral elements (figure [6](#F6){ref-type=\"fig\"}).\n\n![**(patient 2)**. Radiographic examination of the femora showed thick, and dense, endosteal hyperostosis more marked at the mid-diaphyses. The medullary canals were narrow but patent, and the endosteal surface was irregular (arrow).](1752-1947-1-142-4){#F4}\n\n![(patient 2). Lateral radiogram of the skull showed thick and sclerosed skull vault (arrow).](1752-1947-1-142-5){#F5}\n\n![**(patient 2)**Sagittal reformatted multiplanar computed tomography of the craniocervical junction showed a hyperplastic odontoid process. The tip of the dens projected 4.8 mm above a line joining the back of the hard palate to the lowest point of the occipital squama (McGregor line). A mild basilar invagination was therefore present. Note the increased bone density of the posterior vertebral elements.](1752-1947-1-142-6){#F6}\n\nSagittal reformatted multiplanar computed tomography demonstrated irregular increased density in the mandibular bone and the increased density of cancellous bone at the inferior borders (figure [7](#F7){ref-type=\"fig\"}-arrow). Coronal reformatted multiplanar computed tomography of the cervical spine demonstrating the increased density of the transverse processed and the unduly long odontoid process-arrow- (fig [8](#F8){ref-type=\"fig\"}). Lateral radiograph of the thoracic spine showed progressive sclerosis of the anterior longitudinal ligament with subsequent development of bony ankylosis (arrow-fig [9](#F9){ref-type=\"fig\"}).\n\n![**(patient 2)**. Sagittal reformatted multiplanar computed tomography showed irregular increased density in the mandibular bone and the increased density of cancellous bone at the inferior borders.](1752-1947-1-142-7){#F7}\n\n![**(patient 2)**. Coronal reformatted multiplanar computed tomography of the cervical spine, demonstrating the increased density of the transverse processes and the unduly long odontoid process.](1752-1947-1-142-8){#F8}\n\n![**(patient 2)**. Lateral radiograph of the thoracic vertebrae showed progressive sclerosis of the anterior longitudinal ligaments with subsequent development of bony ankylosis (arrow).](1752-1947-1-142-9){#F9}\n\nWe measured the bone density in the hip and spine by means of central DEXA (Dual-energy x-ray absorptiometry). Lumbar spine showed osteoporosis (T-Score -2.9 SD), whereas the hip showed normal limits (T-Score -1.0SD).\n\nDiscussion\n==========\n\nHyperostosis corticalis generalisata refers to a group of sclerosing bone diseases, the principal feature of which is a generalised, symmetrical, endosteal sclerosis at different sites of intramembraneous ossification. Four types of endosteal hyperostosis (van Buchem disease, Worth disease, Nakamura disease, and Truswell-Hansen disease) have been reported.\n\nThese differ from each other on the basis of the mode of inheritance and the involvement of the cranial nerves. Van Buchem\\'s disease, for instance, is an autosomal recessive disease whose main features include involvement of the mandible, causing a misshapen jaw, unlike the relative normal facial phenotype seen in our patient \\[[@B1]-[@B9]\\].\n\nWorth and Wollin, first described an autosomal dominant endosteal hyperostosis \\[[@B1]\\], in which early radiographic changes included thickening of the endosteum of the long bones and of the skull. Classically, the affected bones in patients with endosteal hyperostosis are resistant to fracture and only two patients with a fracture have been reported \\[[@B6]\\]. Our patients were unusual in that both presented with multiple fractures.\n\nSpinal changes have rarely been reported in the Worth type of endosteal sclerosis. Perez-Vicente *et al*., \\[[@B10]\\] reported a father and daughter, with endosteal hyperostosis. Both showed severe sensorineural hearing loss, chronic intracranial hypertension, and mild corticospinal tract involvement. A cranial CT scan documented a reduction in size of the posterior fossa with encroachment at the foramen magnum. Neither fractures nor distinctive spinal changes were documented. Ades *et al*., \\[[@B5]\\] described the gradual increase in density of the posterior elements of the spine, which may be associated with arthritis and may lead to nerve entrapment.\n\nCamurati-Engelmann disease (CED) \\[[@B11]\\], was considered. But, in our patient, the absence of muscle weakness, fatigue, poor appetite, headache and pain in the limbs, important features encountered in CED, were absent. Moreover, in CED the long bones are thickened and bone prominences and tenderness occur. These were not present in our patient.\n\nMalformations of the craniocervical junction have been reported in patients with bone fragility. They have been found in association with osteomalacia, Paget\\'s disease, hyperparathyroidism, rheumatoid arthritis, and osteogenesis imperfecta. In osteogenesis imperfecta, basilar impression rather than basilar invagination, is the rule \\[[@B12]\\]. Our patient had hyperplasia of the odontoid. This is defined as hypertrophy of the apical portion of the odontoid process of the axis. It can occur in Marfan syndrome, where height acceleration is a prominent feature and dolichoodontoid can cause sudden death \\[[@B13]\\]. Neither the phenotype, nor the radiographic features in our patient, was in favour of Marfan syndrome.\n\nOur patient had a progressive anterior and posterior vertebral hyperostosis in connection with progressive sclerosis. The pattern was irrelevant to the diffuse idiopathic skeletal hyperostosis syndrome (DISH). The latter is observed in elderly patients with a history of disturbed metabolic parameters \\[[@B14]\\]. In our patient there is thoracic spine ankylosis involved the anterior and the posterior spine elements simultaneously; in a way it gave no chance to maintain skip areas of ankylosis. Spinal segments ankylosis might be a major predisposing factor for the development of marked spinal osteopenia and possibly osteoporosis and fracture \\[[@B15]\\]. Minor trauma can lead to serious fracture of the ankylosed spine.\n\nConclusion\n==========\n\nWe present an unusual father-daughter pair with endosteal hyperostosis not compatible with Worth syndrome, spinal changes and multiple fractures. Although we have not been able to elucidate the actual pathogenesis in our patients, it sounds the likelihood that the disorder results from a simultaneous uncoupling of formation and resorption.\n\nWe suggest that the constellation of spinal changes seen in our patient was possibly caused by progressive prenatal, axial and extra-axial sclerosis. The fact of decreased bone density in the father could possibly correspond with the disturbed postnatal osteoblast-mediated bone formation to defective osteoclastic bone resorption.\n\nOn the other hand, the craniocervical changes were important to detect, as subclinical basilar invagination has the potential for severe neurological consequences. Anterior bony ankylosis over the thoracic spine is an additional risk for the development of myelopathy and fractures and nerve entrapment might result from the progressive involvement of the sclerotic posterior elements of the spine. Finally we wish to stress, that in the light of incompatibility between our current family and other forms of well-documented endosteal hyperostosis, the overall clinical and radiographic phenotype were distinctive and possibly represent a new syndrome.\n\nCompeting interests\n===================\n\nThe author(s) declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nAll authors read and approved the final manuscript and all participate in this work\n\nAcknowledgements\n================\n\nWe wish to thank the patient for his cooperation and for providing written consent to publish the data.\n"}
+{"text": "Clinical translation of advances in understanding the genetics of common disease will require primary care practitioners to play an increasing role in providing genetic advice ([Qureshi *et al*, 2004](#bib21){ref-type=\"other\"}). Cancer genetics provides a model for the genetics of common disease and their clinical implications for primary care ([Emery *et al*, 2001](#bib10){ref-type=\"other\"}). The discovery of genes that place individuals at increased risk of breast, ovarian and colorectal cancer has had important and immediate clinical applications ([Wooster and Weber, 2003](#bib29){ref-type=\"other\"}). Key tasks for primary care practitioners include identifying individuals likely to be at increased genetic risk and advising those for whom genetic testing and increased disease surveillance offer little benefit. Previous audits of referrals to familial cancer clinics in the UK show that approximately 30% of referrals from general practice are for patients whose risk is not significantly raised on current evidence ([Wonderling *et al*, 2001](#bib28){ref-type=\"other\"}).\n\nPreviously we reported experimental and qualitative evaluations of a prototype computer decision support tool for the management of familial cancer risk in primary care ([Emery *et al*, 1999](#bib11){ref-type=\"other\"}, [2000](#bib12){ref-type=\"other\"}). These demonstrated the functionality and design of the software, and also demonstrated its potential to improve general practitioners\\' management decisions in simulated cases. This work underpinned the development of the GRAIDS software (Genetic Risk Assessment on the Internet with Decision Support) ([Emery, 2005](#bib9){ref-type=\"other\"}). We now report the results of a randomised controlled trial of an assessment strategy using the GRAIDS software in British general practice compared with best current practice.\n\nMATERIALS AND METHODS\n=====================\n\nDesign\n------\n\nThis was a pragmatic cluster randomised controlled trial with randomisation, at the level of the general practice, to GRAIDS or current 'best practice\\' ([Campbell *et al*, 2000](#bib4){ref-type=\"other\"}). Outcomes were measured at practice, practitioner and patient levels. Within the intervention group, we used an exploratory design with fixed and adaptive arms based on a threshold of software use.\n\nObjectives\n----------\n\nWe hypothesised that the new assessment strategy (GRAIDS) would result in a greater proportion of referrals to the Regional Genetics Clinic that were consistent with the risk assessment guidelines for familial breast/ovarian cancer and familial colorectal cancer, than current best practice. Secondary hypotheses included that patients from intervention practices would have greater knowledge about familial cancer without higher cancer worry, at the point of referral than from comparison practices. Within the GRAIDS practices, we predicted that patients who were not referred would have lower risk perception and lower cancer worry than those who had been referred, and that the intervention would increase practitioner confidence in management.\n\n### Participant practice teams\n\nWe invited 170 general practice teams in the Eastern Region of England, with a minimum of three full-time-equivalent doctors, to join the trial. Inclusion criteria were that the practice was connected to the health service intranet (NHSnet) and referred patients with a family history of cancer to the Eastern Regional Genetics Clinic at Addenbrookes Hospital NHS Trust, Cambridge. Forty-five practice teams agreed to participate and were randomised to GRAIDS (intervention) or best practice (comparison) strategies.\n\n### Interventions and recruitment of patients\n\n*Intervention; GRAIDS strategy*: All general practitioners and practice nurses attended a 45-min educational session on cancer genetics, delivered at their general practice. They were also introduced to the principles of the GRAIDS intervention. Each practice team selected a single clinician (general practitioner or practice nurse) to act as the 'lead clinician\\' to manage all patients who expressed concerns about their family history of breast or colorectal cancer. In larger practices, this role could be shared by two clinicians. The lead clinicians attended a further 90-min interactive training session to learn to use the GRAIDS software.\n\nThe GRAIDS software links a user-friendly pedigree-drawing tool to patient-specific management advice regarding a family history of breast/ovarian and colorectal cancer, and provides additional numerical risk information about breast cancer ([Supplementary Figure 1](#sup1){ref-type=\"other\"}). The software applies Cyrillic technology ([Benson, 2000](#bib1){ref-type=\"other\"}) to create pedigrees and assesses familial cancer risk using two parallel methods: the implementation of risk assessment guidelines and an epidemiological risk model. In the GRAIDS Trial, we implemented the regional guidelines for familial breast/ovarian cancer and familial colorectal cancer ([Table 1](#tbl1){ref-type=\"table\"}). In addition, the Claus model ([Claus *et al*, 1991](#bib5){ref-type=\"other\"}) was applied to provide breast cancer risk information in a range of verbal and graphical modes. The regional guidelines are principally designed to assess cancer risk and categorise people into increased risk or population risk; the familial breast/ovarian cancer guidelines additionally categorise women into moderate and high risk, the latter representing people who are also at clinically significant risk of carrying a BRCA1/2 mutation. The guidelines are used to inform referrals for those at increased risk of cancer to the Regional Genetics Clinic.\n\nThe GRAIDS software operates on a central server within a computer network. In this trial, the software was installed on a secure server at the Addenbrookes Hospital NHS Trust. Each lead clinician accessed the GRAIDS software via their NHSnet connection, using a practice-specific password.\n\nPatients were invited to participate if they expressed concerns about their family history of breast or colorectal cancer in a consultation. They were referred to the lead clinician and given a detailed explanation of project participation, and a family history questionnaire to complete before the next consultation. The family history questionnaire was designed as part of the GRAIDS strategy to improve the accuracy of the family history information provided by participants.\n\n*Fixed and adaptive sub-groups*: Within the intervention arm, practices were randomised to either a fixed or adaptive sub-group. In the adaptive group, the lead clinician was interviewed 3 months after training, if frequency of software usage was below a predefined level based on size of the practice population. The interview identified reasons for low usage and aimed to resolve any problems using the software. In the fixed group, practices received the intervention as described above, with no opportunity for additional input. The purpose of this was to include the option of additional clinician training or adjustment to the software in the adaptive arm during the trial, to increase software use.\n\n*Comparison; current 'best practice\\'*: All general practitioners and practice nurses attended a 45-min educational session on cancer genetics delivered at their general practice ([Watson *et al*, 2001](#bib24){ref-type=\"other\"}). Afterwards, they were each mailed a paper copy of the regional guidelines for familial breast/ovarian cancer and familial colorectal cancer. Patients were unaware that they would be invited to participate in the trial, until they were referred to the Regional Genetics Clinic.\n\nMeasures\n--------\n\n### Randomised comparisons\n\n*Practice level, principal outcome and appropriate referral rates*: We audited all referrals to the Regional Genetics Clinic from trial practices during the study period. 'Appropriateness\\' of referral was defined in two ways: (1) (principal outcome) consistency of the family history reported in the general practitioner\\'s referral letter with the regional guidelines for familial breast/ovarian and colorectal cancer; (2) (secondary outcome) the final expert risk assessment conducted by the Regional Genetics Clinic staff among attending patients, according to current best evidence, to determine if the patient was at significantly increased risk of familial cancer. A 'relevant referral\\' was for the individual person referred, rather than their relative, and about their family history of either breast, ovarian or colorectal cancer.\n\n*Patient level: (referred patients only):* Questionnaires were sent to patients when a referral was received at the Regional Genetics Clinic. The questionnaire measured risk perception, knowledge about familial cancer and cancer worry using disease-specific measures. Items were taken from established instruments identified in a systematic review of genetic counselling for familial cancer ([Braithwaite *et al*, 2004](#bib2){ref-type=\"other\"}). Risk perception was measured on a scale of 1--7, relative to the general population, considering 1 as 'much less likely...\\' and 7 as 'much more likely to develop breast/bowel cancer than other people of your age\\'. For one analysis, responses of 1--4 were classified as 'population risk\\' and 5--7 as 'increased risk\\'. This was compared to the risk assessment conducted by the Regional Genetics Clinic to classify patients as under-estimators, accurate-estimators and over-estimators ([Watson *et al*, 1999](#bib26){ref-type=\"other\"}).\n\n### Comparisons within GRAIDS arm only\n\n*Practitioner level*: We measured the frequency of use of the software remotely from server activity. A questionnaire examined the attitudes of lead clinicians towards using the GRAIDS software, their confidence in risk assessment and managing patients with a family history of cancer, problems using the software and intention to continue using it ([Braithwaite *et al*, 2002](#bib3){ref-type=\"other\"}). This questionnaire was provided before the lead clinician training and 2 weeks and 12 months after. Lower scores on these instruments reflected more positive attitudes or greater agreement with a statement. Data on the length of consultation with the lead clinician were obtained from electronic appointment systems in 20 intervention practices.\n\n*Patient level*: Those in the GRAIDS arm who were not referred, were sent questionnaires 2 weeks after their consultation with the lead clinician for comparison with those referred from GRAIDS practices.\n\nSample size\n-----------\n\nTwenty intervention and 20 comparison practices allowed an estimation of effect size on appropriateness of referrals for familial cancer with a precision of \u00b111% as measured by the 95% confidence interval (CI) width, equivalent to 80% power to detect a 15% difference between arms. Ten practices per intervention arm also provided 80% power to detect a 25% relative difference (adaptive *vs* fixed arms) in use of software at 1 year, and 33% relative difference at 3 months using practice-level *t*-tests at the 5% level of significance. Sample sizes were determined through simulation using S-plus 2000 software (MathSoft Inc., Seattle, WA, USA).\n\nRandomisation\n-------------\n\nPractices were randomised using a partial minimisation procedure that dynamically adjusted the randomisation probabilities in order to provide a balance between arms in the mean number of patients aged 20--50 years per practice and to achieve the planned allocation ratio of 1\u2009:\u20091\u2009:\u20092 for adaptive intervention, fixed intervention and comparison arms. Randomisation was conducted independently by a statistician (ATP) who had no contact with practices.\n\nStatistical methods\n-------------------\n\nRates of software use were calculated, for each practice, as the annual number of software uses per 10\u2009000 registered patients, and were compared between fixed and adaptive arms using a *t*-test. The same method was used to compare referral rates between GRAIDS and comparison arms. The binary outcomes of referral appropriateness were analysed using a generalised linear mixed-effects model, allowing for practice as a random effect in order to account for the cluster randomised design.\n\nPatient questionnaire outcomes were analysed using linear or generalised linear mixed-effects models as appropriate, allowing for practice as a random effect. Numerical knowledge and worry scales were analysed as continuous variables.\n\nWithin the GRAIDS arm, practitioner outcomes were analysed using a paired *t*-test to compare attitudes towards using the GRAIDS software 2 weeks and 12 months after training with pre-training levels. A significance level of 5% was used for all tests. All analyses were conducted on an intention-to-treat basis. Analysis was carried out using SPSS version 12.5 and R version 2.0.1.\n\nEthical and research governance approval for the study were received by the Eastern Multi-Centre Research Ethics Committee and the relevant primary care trusts, respectively. For the intervention arm, participants signed a consent form at the beginning of their consultation with the lead clinician. For the comparison arm, consent was assumed from return of the questionnaire.\n\nRESULTS\n=======\n\nPractice characteristics\n------------------------\n\nAll 45 practice teams were in the trial for a minimum of 12 months and none withdrew. Twenty-three practice teams were randomised to the intervention and 22 to the comparison group. [Table 2](#tbl2){ref-type=\"table\"} presents the main characteristics of practices and participants in the two trial arms. There were no statistically significant differences in practice characteristics at baseline. The flow of participants through the trial is shown in [Figure 1](#fig1){ref-type=\"fig\"}.\n\nOne hundred and sixty-two relevant referrals were registered at the Regional Genetics Clinic from GRAIDS practices (162/168 referred) and 84 from comparison practices during the trial. These denominators were used for analyses of referral rates. One hundred and sixty-nine participants had attended their Regional Genetics Clinic appointment by the end of the trial and contributed to the comparison of appropriateness of referral against final Regional Genetics Clinic risk assessment (GRAIDS arm: 117/162 participants, Comparison arm: 52/84 participants). Twenty-two participants were referred for a family history of both breast and colorectal cancer, and therefore contributed two risk assessments in the analyses of appropriateness of referral letters (GRAIDS arm 21; comparison arm 1). Sixteen of these attended the Regional Genetics Clinic within the trial period (GRAIDS arm 15; comparison arm 1). In total, 132 risk assessments in the GRAIDS arm and 53 in the comparison arm were obtained. Questionnaires were returned by 75% of referred participants from the GRAIDS arm and 64% from the comparison arm.\n\nRandomised comparisons\n----------------------\n\n### Practice level\n\nPrincipal outcome: appropriate referral rates: There were 162 relevant referrals made by the GRAIDS practices and 84 referrals by the comparison practices to the Regional Genetics Clinic for familial cancer risk assessment. GRAIDS practice teams referred a mean 6.2 (standard deviation (s.d.), 3.1) per 10\u2009000 registered patients per practice per year compared to 3.2 (s.d., 2.8) per 10\u2009000 registered patients per practice per year in comparison practices (mean difference 3.0; 95% CI 1.2--4.8, *P*=0.002).\n\nA significantly higher proportion of referral letters was consistent with the regional guidelines in the intervention arm than in the comparison arm (breast cancer alone and breast and colorectal cancer combined; [Table 3](#tbl3){ref-type=\"table\"}). There was no overall difference between groups in the final risk assessment conducted by staff at the Regional Genetics Clinic. This was due to the large proportion of referrals from GRAIDS practices for family history of colorectal cancer that, while consistent with regional guidelines, on final risk assessment by the Regional Genetics Clinic staff, were deemed to be at population risk. Participants referred from intervention practices about colorectal cancer risk were significantly more likely to be at population risk than those from comparison practices when their risk was determined at the Regional Genetics Clinic, even though the referral letter was consistent with increased risk defined by the familial colorectal cancer guideline. The referrals about colorectal cancer did not account for the significantly higher number of referrals overall from intervention practices.\n\n### Patient level\n\nThere were no significant differences in knowledge scores between patients referred from intervention or comparison practices. Cancer worry scores were significantly lower in patients from intervention practices than those from comparison practices. There was no difference in mean risk perception between patients referred from intervention or comparison practices ([Table 4](#tbl4){ref-type=\"table\"}). There was a non-significant trend towards more accurate risk perception at the point of referral in intervention patients with fewer overestimating risk (odds ratio (OR) 1.50, 95% CI 0.62--3.67; *P*=0.36) ([Table 5](#tbl5){ref-type=\"table\"}).\n\nComparisons within GRAIDS arm only\n----------------------------------\n\n### Practitioner level\n\nSoftware use: The software was used with patients 219 times during the trial; this equates to a mean use of 8.27 per 10\u2009000 registered patients per practice per year. There was no clear trend in frequency of use of the software during the first 12 months of the trial.\n\nFixed and adaptive arms: All lead clinicians from the 11 practices in the adaptive sub-group were interviewed at 3 months due to lower than predicted software use. No specific problems in using the software were identified. All lead clinicians felt low use reflected low patient demand. There was no difference in software use between fixed and adaptive practices at 12 months: fixed arm, mean 7.8 (s.d., 4.7) uses per practice per year per 10\u2009000 registered patients; adaptive arm, mean 8.8 (s.d., 4.1); mean difference 0.9; 95% CI \u22122.8--4.8; *P*=0.60). We therefore combined data from the fixed and adaptive sub-groups of the intervention for all subsequent analyses.\n\nPractitioner confidence and attitudes: Lead clinicians\\' confidence in managing people with a family history of cancer increased significantly after training and this increase was maintained at 12 months ([Figure 2](#fig2){ref-type=\"fig\"}). Their attitudes towards the software were generally positive, such that it was felt to be simple, easy, beneficial and cost-effective and these positive attitudes remained at 12 months ([Figure 2](#fig2){ref-type=\"fig\"}). However, there was some reduction over time, in agreement with the statement that the software enhanced consultations (mean score 2.1 (s.d., 0.8) post-training; 3.0 (s.d., 1.7) at 12 months; mean change 0.8 95% CI 0.1--1.6; *P*=0.04; *n*=26) and persistent agreement that it would prolong consultations (mean score 2.5 (s.d., 1.2) post training; mean score 2.3 (s.d., 1.2) at 12 months). All but one lead clinician intended to continue using the software if it remained available. Median consultation time with the lead clinician was 28\u2009min.\n\n### Patient level\n\nA total of 219 patients received the GRAIDS intervention, of whom, 141 were referred to the Regional Genetics Clinic. A further 27 patients were referred to the Regional Genetics Clinic from GRAIDS practices without a GRAIDS consultation. Of the 168 referrals, 162 were identified at the Regional Genetics Clinic and were for relevant cancers. The 78/219 patients not referred were identified by the consent procedures. Risk perception was significantly lower in patients not referred to the Regional Genetics Clinic than in those referred.\n\nDISCUSSION\n==========\n\nPractice team access to a family history assessment strategy using the GRAIDS software resulted in increased referral rates from primary care to a regional genetics clinic for familial breast and colorectal cancer, compared with current best practice. Referrals from GRAIDS practices were more appropriate than from comparison practices, when judged by their consistency with referral guidelines, the most relevant measure of general practitioners\\' clinical behaviour. Cancer worries were lower among patients referred from GRAIDS practices than from comparison practices. For the GRAIDS arm, patients who were not referred had lower cancer risk perception than those who were. Clinicians were generally positive about the software and intended to continue to use it if available.\n\nLimitations of design\n---------------------\n\nCluster randomisation at practice level to intervention and comparison strategies is the design of choice when the intervention is applied at that level ([Donner, 2000](#bib7){ref-type=\"other\"}). It avoids contamination between arms, which can reduce differences observed, but can pose problems in equivalence of recruitment and consent across arms. Thus we recruited patients in the GRAIDS arm as they consulted their GP. However, we wished in the comparison arm to mimic as closely as possible routine best practice ([Watson *et al*, 2002](#bib25){ref-type=\"other\"}). We therefore did not recruit patients who expressed concerns about their family history of cancer in this arm, to avoid an intervention effect by, for example, raising patient expectation and increasing clinician referrals independent of usual best practice. Consent to data collection in the comparison arm was seen only among those referred to the regional clinic. This meant that we were unable to collect data on patients in comparison practices, who presented but were not referred to the Regional Genetics Clinic.\n\nThe design precluded direct comparison of practice-based recruitment rates in the two arms; we therefore analysed referrals standardised by practice-registered population, where appropriate. Differences in principal referral outcomes between arms are thus an unbiased estimate of differences between the two overall service models, except for the timing of consent, the effect of which we believe would be small and not create any systematic bias.\n\nKey findings\n------------\n\nThis is the first report of a clinical trial demonstrating the value of family history assessment software designed for general practice. The trial design accounted for many of the recognised flaws in past trials of computer decision support systems ([Mitchell and Sullivan, 2001](#bib19){ref-type=\"other\"}). A previous study of mailing a CD Rom containing electronic guidelines for familial breast cancer to general practitioners in Scotland found minimal uptake of the software ([Wilson *et al*, 2005](#bib27){ref-type=\"other\"}). Our intervention differed in several critical ways: the software had greater utility by supporting collection of family history information and creating a pedigree; it made patient-specific recommendations about management at the point of decision; clinicians received interactive training in its use and the service model of training a single clinician in a practice led to more frequent use than if all practitioners had been trained ([Kawamoto *et al*, 2005](#bib15){ref-type=\"other\"}). The trial was testing the service model of providing a risk assessment service in the practice by supporting a single clinician in the practice regardless of background knowledge or interest. In some practices the lead clinician may have had a specific interest in familial disease or computer support, although to our knowledge this was only true in one practice. Over a quarter of practice teams approached were able to participate, a recruitment rate consistent with similar primary care trials ([Montgomery *et al*, 2000](#bib20){ref-type=\"other\"}). Practices recruited into the trial are likely to reflect the teams of the future who will be leading the incorporation of genetic medicine into practice ([Rogers, 1995](#bib22){ref-type=\"other\"}).\n\nThe GRAIDS intervention resulted in significantly more referrals that were consistent with referral guidelines, for breast cancer alone and combined with colorectal cancer. This difference was not apparent when appropriateness of referral was judged on the final Regional Genetics Clinic risk assessment; indeed the reverse was seen for colorectal cancer. This may reflect differences in the validity of self-reported family history of cancer across disease sites, which is usually confirmed by the genetics clinic ([Ziogas and Anton-Culver, 2003](#bib32){ref-type=\"other\"}). More importantly it reflects differences in the complexity and accuracy of the guidelines. The familial breast cancer guidelines are more complex than for colorectal cancer with several more criteria that define moderate and high-risk groups. Consequently they are more specific, but more difficult to implement in general practice. Computer implementation of familial breast cancer guidelines is more likely therefore to have a positive effect. The familial colorectal cancer guideline attempted to account for additional cancers associated with hereditary non-polyposis colon cancer (HNPCC), while still capturing individuals at moderate risk, and were therefore less specific. As a result, when applied rigidly by the software, practitioners made more referrals that were subsequently assessed as at population risk by the Regional Genetics Clinic, even though the referral would be considered 'appropriate\\', as it met the category for increased risk in the familial colorectal cancer guideline. This highlights the need for accurate guidelines for familial colorectal cancer to underpin decision support in primary care, similar to those published for breast cancer ([McIntosh *et al*, 2004](#bib17){ref-type=\"other\"}). Current criteria to identify individuals with potential HNPCC, such as the modified Bethesda criteria, aim to identify only high-risk individuals, and would fail to identify those at moderately increased risk, who may also benefit from referral ([Umar *et al*, 2004](#bib23){ref-type=\"other\"}).\n\nAccurately identifying patients who may be at increased risk of breast and colorectal cancer, and reassuring the majority who are unlikely to benefit from referral is an important outcome of this trial. Although it resulted in increased referrals, which in the United Kingdom might threaten overstretched regional genetics services, it is the role of primary care to identify patients who might benefit most from specific referral, in this case for genetic counseling ([Hayflick *et al*, 1998](#bib14){ref-type=\"other\"}). There is growing evidence of benefit for a range of surveillance and prophylactic measures for individuals at increased risk of breast and colorectal cancer ([Cuzick *et al*, 2003](#bib6){ref-type=\"other\"}; [Dove-Edwin *et al*, 2005](#bib8){ref-type=\"other\"}; [Leach *et al*, 2005](#bib16){ref-type=\"other\"}). Access to predictive genetic testing can inform patients\\' decisions about prophylactic surgery and other preventive strategies ([Meijers-Heijboer *et al*, 2000](#bib18){ref-type=\"other\"}).\n\nPatients referred from intervention practices reported significantly lower cancer worry than those from control practices. This was not mediated by risk perception or the knowledge items we measured. It is possible that the longer and more detailed assessment in primary care better prepared patients for referral with consequent reduction in anxiety. Alternatively, the GRAIDS intervention may have resulted in a different population referred to the Regional Genetics Clinic, who would not usually have been referred. Participants who were not referred from intervention practices showed lower cancer worry and mean risk perception than those who were referred. While acknowledging the absence of baseline data, this suggests that the intervention helps clinicians to reassure patients at population risk of familial cancer and manage them in primary care.\n\nThe intervention increased GPs\\' confidence in managing familial cancer, an area of medicine that is relatively new and complex from a primary care perspective ([Emery *et al*, 2001](#bib10){ref-type=\"other\"}). Clinicians were generally positive about the software in terms of simplicity and utility. However, concerns were evident about the time taken to conduct a consultation using the software. At the time of the trial, the majority of practices had limited bandwidth available via their NHSnet connection, making the software relatively slow to run. Recent expansion of broadband connections to practices in the UK would reduce the consultation time. However, this type of consultation cannot be conducted in a 'standard 10\u2009min\\', as is true for a growing number of complex conditions managed in primary care ([Freeman *et al*, 2002](#bib13){ref-type=\"other\"}). What may be required is a brief, sensitive triage tool that identifies people with a family history of cancer and other common familial conditions for a more detailed assessment using GRAIDS-based software.\n\nThere is growing interest in the broader application of the family history of common disease in preventive health ([Yoon *et al*, 2003](#bib31){ref-type=\"other\"}). The Center for Disease Control has recently developed a similar electronic family history tool for use in primary care that is undergoing evaluation ([Yoon and Scheuner, 2003](#bib30){ref-type=\"other\"}). This trial demonstrates the potential of the GRAIDS software to improve the management of familial cancer in primary care, assuming the accuracy of current risk assessment guidelines. The software is now being developed to implement validated risk-assessment guidelines for other common familial conditions, to support the broader management of the family history of common disease in primary care.\n\nExternal data objects {#sup1}\n=====================\n\n###### Supplementary Figure 1\n\nThis trial was funded by Cancer Research UK and NHS R&D Support for Science. Jon Emery received a Cancer Research UK Primary Care Clinician Scientist Award during the conduct of the trial. The General Practice and Primary Care Research Unit is an NHS R&D Academic Unit and a member of the English National School of Primary Care Research. Jon Emery is the guarantor of this paper.\n\nWe thank all the individual participants and the lead clinicians, general practitioners, practice managers and administration staff at the following participating practices: Acorn Community Health Centre, Huntingdon; Dr Bailey & Partners, Cambourne; Botesdale Health Centre, Diss; Bridge Street Surgery, Cambridge; Cornerstone Practice, March; Cottenham Surgery; Cromwell Place Surgery, St Ives; East Barnwell Health Centre, Cambridge; Eaton Socon Health Centre; Gayton Road Health Centre, King\\'s Lynn; George Clare Surgery, Chatteris; Guildhall Surgery, Bury St Edmunds; Harston Surgery; Dr Hassan & Partners, Hopton; Dr Hewlett & Partners, Willingham; Jenner Health Centre, Peterborough; Dr Lazar & Partners, Ipswich; Litcham Health Centre; Moat House Surgery, Warboys; Nene Valley Medical Centre, Peterborough; Newnham Walk Surgery, Cambridge; Nuffield Road Medical Centre, Cambridge; Orchard House Surgery, Newmarket; Orchard Surgery, St Ives; Oundle Medical Practice, Peterborough; Dr Oxley & partners, Attleborough; Park Medical Centre, Peterborough; Paston Health Centre, Peterborough; Queen Street Surgery, Whittlesey; Rainbow Surgery, Ramsey; Ramsey Health Centre; Dr Rawlinson & Partners, Kimbolton; Roysia Surgery, Royston; Sawston Medical Practice; South Street Surgery, Bishop\\'s Stortford; Stoke Park Medical Centre, Ipswich; Spinney Surgery, St Ives; Stowhealth, Stowmarket; Thistlemoor Medical Centre, Peterborough; Trinity Medical Practice, Norwich; Trumpington Street Medical Practice, Cambridge; Upwell Health Centre; Wickhambrook Surgery; Yaxley Group Practice; York St Medical Practice, Cambridge.\n\nWe also thank the following: David Spiegelhalter, MRC Biostatistics Unit, Cambridge, for advice on the exploratory design of the study; Theresa Marteau, Guys, Kings and St Thomas\\' School of Medicine, for advice on measurement issues and overall study design and Fiona Walter, General Practice and Primary Care Research Unit, Cambridge for helpful discussions throughout the study. We thank Joan Paterson, Julia Rankin, Sarah Downing, Anna Middleton, Vicki Wiles and Michaela Twyford of the Regional Genetics Clinic, Cambridge for support in the planning stages, as well as throughout the course of the study. We thank Craig Livingstone and other programmers at AP Benson for their support in developing and maintaining the GRAIDS software.\n\n**Competing interest**\n\nJon Emery has received a consultancy from AP Benson Ltd, which produces Cyrillic software. All other authors have no competing interests.\n\n[Supplementary Information](#sup1){ref-type=\"other\"} accompanies the paper on British Journal of Cancer website (<http://www.nature.com/bjc>)\n\n![Changes in lead clinicians\\' attitudes towards the GRAIDS software, and confidence in managing familial cancer over time (mean score with 95% CI shown; *P*-values refer to comparison of pre- and post-training responses).](6603897f1){#fig1}\n\n![CONSORT Flowchart.](6603897f2){#fig2}\n\n###### \n\nEastern Region Familial Breast/Ovarian and Colorectal Cancer Guidelines\n\n  -------------------------------------------------------------- -----------------------------------------------------------------------------------------------------------------------------------\n  *Breast/ovarian cancer*                                        \n  \u2003*High risk criteria:*                                         \n  \u2003\u20031.                                                           Two relatives who are FDR of each other with breast cancer where average age of diagnosis is under 40 years.\n  \u2003\u20032.                                                           Three or more relatives who are FDR of each other with ovarian or breast cancer, where average age of diagnosis is under 60 years\n  \u2003\u20033.                                                           Four or more relatives who are FDR of each other with breast or ovarian cancer at any age\n  \u2003\u20034.                                                           One individual in family with breast and ovarian cancer\n  \u00a0                                                              \u00a0\n  \u2003*Moderate risk:*                                              \n  \u2003\u20031.                                                           One female FDR with breast cancer \\<40 years\n  \u2003\u20032.                                                           One paternal female SDR with breast cancer \\<40 years\n  \u2003\u20033.                                                           One female FDR with bilateral breast cancer \\<60 years\n  \u2003\u20034.                                                           Two FDR/SDR with breast cancer \\<60 years or ovarian cancer any age\n  \u2003\u20035.                                                           Three FDR/SDR with breast or ovarian cancer any age\n  \u2003\u20036.                                                           One male FDR with breast cancer any age\n  \u00a0                                                              \u00a0\n  *Colorectal cancer*                                            \n  \u20031.                                                            One affected FDR \\<45 years\n  \u20032.                                                            One affected FDR and 1 affected SDR on same side of family\n  \u20033.                                                            Two FDR (inc both parents)\n  \u20034.                                                            Three affected relatives any age\n  \u00a0                                                              \u00a0\n  \u2003*'Affected\\' means diagnosed with either of the following:*   \n  \u2003\u2003a.                                                           CRC, colorectal cancer: 3\u2a7e adenomatous polyps, one adenomatous polyp \\<60 years.\n  \u2003\u2003b.                                                           HRC, HNPCC-related cancer: endometrium, ovary, gastric, small bowel, ureter, renal pelvis\n  There should be at least one CRC in the family.                \n  -------------------------------------------------------------- -----------------------------------------------------------------------------------------------------------------------------------\n\nAbbreviations: CRC, colorectal cancer; FDR, first-degree relative; HRC, HNPCC-related cancer; SDR=second-degree relative.\n\n###### \n\nCharacteristics of practices and participants in trial arms\n\n  \u00a0                                                    **Intervention practices**   **Comparison practices**\n  --------------------------------------------------- ---------------------------- --------------------------\n  *Practice factors at baseline*                                                   \n  \u2003Number                                                          23                          22\n  \u2003Mean list size (s.d.)                                      8787 (3840)                 8718 (4614)\n  \u2003Mean number of patients aged 20--50 years (s.d.)           3881 (1747)                 3843 (2136)\n  \u00a0                                                                \u00a0                           \u00a0\n  *Participant factors at baseline*                                                \n  \u2003No. in trial not referred to RGC                                78                  Unknown by design\n  \u2003No. relevant referrals detected at RGC                         162                          84\n  \u00a0                                                                \u00a0                           \u00a0\n  \u2003*No. (%) referred to RGC for family history of*                                 \n  \u2003\u2003Breast/ovarian cancer                                       86 (53)                     60 (71)\n  \u2003\u2003Colorectal cancer                                           55 (35)                     23 (27)\n  \u2003\u2003Both                                                        21 (13)                      1 (1)\n\n###### \n\nProportion of referrals meeting guidelines and number of referrals for increased risk, as determined by RGC, by randomised group\n\n  \u00a0                                              **Intervention**   **Control**   **Odds ratio (95% CI)**\n  --------------------------------------------- ------------------ ------------- -------------------------\n  *Proportions meeting referral guidelines*                                      \n  \u2003Breast                                          93% (99/107)     73% (44/60)       4.5 (1.6--13.1)\n  \u2003Bowel                                           99% (75/76)      92% (23/25)       6.5 (0.5--83.7)\n  \u2003Combined                                       95% (174/183)     79% (67/85)       5.2 (1.7--15.8)\n  \u00a0                                                     \u00a0                \u00a0               *P*=0.006\n  *Proportions with increased RGC risk level*                                    \n  \u2003Breast                                          77% (60/78)      70% (23/33)       1.4 (0.6--3.5)\n  \u2003Bowel                                           56% (30/54)      85% (17/20)       0.2 (0.1--0.8)\n  \u2003Combined                                        68% (90/132)     75% (40/53)       0.7 (0.3--1.5)\n  \u00a0                                                     \u00a0                \u00a0               *P*=0.35\n\nAbbreviations: CI, confidence interval.\n\nOdds ratios and 95% confidence intervals shown for intervention *vs* comparison, allowing for the cluster randomised design.\n\nNB: 21 participants were referred for a family history of breast and colorectal cancer, of whom 15 attended the RGC; these participants contribute two risk comparisons in the analyses.\n\n###### \n\nPatient knowledge, cancer worry and risk perception mean scores (standard deviations), and mean differences allowing for cluster randomised design\n\n  \u00a0                              **Intervention arm**        \u00a0                   \u00a0                           \n  ----------------------------- ---------------------- ------------- ------------------------- ------------- ----------------------------\n  Knowledge breast cancer                 NA            5.77 (2.90)             NA              5.66 (2.78)       0.11 (\u22121.05--1.27)\n  \u00a0                                       \u00a0               *n*=65                 \u00a0                *n*=38                  \u00a0\n  Knowledge colorectal cancer             NA            5.50 (2.46)             NA              4.86 (3.30)       0.64 (\u22121.01--2.29)\n  \u00a0                                       \u00a0               *n*=44                 \u00a0                *n*=14                  \u00a0\n  Cancer worry                       4.95 (2.99)        5.74 (3.04)     0.79 (\u22120.19--1.76)      7.18 (3.43)   \u22121.44 (\u22122.64 to \u22120.23)^\\*^\n  \u00a0                                     *n*=57            *n*=110                \u00a0                *n*=51                  \u00a0\n  Risk perception                    4.25 (0.80)        4.99 (1.14)   0.74 (0.38--1.09)^\\*\\*^   5.04 (0.88)     \u22120.09 (0.34 to \u22120.51)\n  \u00a0                                     *n*=51            *n*=104                \u00a0                *n*=47                  \u00a0\n\nAbbreviations: CI, confidence interval; NA, not analysed since no hypothesised difference. ^\\*^*P*=0.02; ^\\*\\*^*P*\\<0.0001.\n\n###### \n\nAccuracy of patients\\' risk perception compared with Regional Genetics Clinic assessment\n\n  \u00a0              **Under-estimator**   **Accurate assessor**   **Over-estimator**\n  -------------- --------------------- ----------------------- --------------------\n  Comparison     9 (23%)               22 (55%)                9 (23%)\n  Intervention   18 (21%)              59 (68%)                10 (11%)\n"}
+{"text": "Introduction\n============\n\nThe current standard therapy for patients with locally advanced rectal cancer (LARC) is neoadjuvant chemoradiation (CRT) followed by total mesorectal excision (TME) and postoperative adjuvant chemotherapy ([@ref-11]; [@ref-15]). This trimodal therapy provides excellent local tumor control and long-term survival ([@ref-9]; [@ref-14]). However, TME is associated with some fatality, morbidity and long-term sequelae, with negative effects on quality of life. The pathological complete response (pCR), defined as the complete absence of tumor cells in the resected specimen and lymph nodes (ypT0N0) is associated with improved local control, overall survival, and disease-free survival which on behalf of improved prognosis ([@ref-7]; [@ref-10]; [@ref-23]). Patients who have a complete response could be eligible for less invasive surgeries or even a watch-and-wait approach, questioning the added value of TME in these patients ([@ref-5]; [@ref-18]). However, the proportion of patients who achieve a pCR remains low with only 8--24% after neoadjuvant CRT ([@ref-10]). Increasing the rate of pCR, and therefore expanding the number of rectal cancer patients who could potentially benefit from a watch-and-wait strategy, has been a hot area of research for years ([@ref-17]). The intensified strategy includes escalating radiation dose, adding neoadjuvant treatment with induction or consolidation chemotherapy, and prolong the time of between neoadjuvant CRT and resection ([@ref-14]). However, the effects of these approaches on tumor response and long-term survival are still controversial. Two prospective phase II trials demonstrated that adding modified FOLFOX6 or XELOX after CRT and before TME increased pCR rate without increase the surgical difficulty in patients with LARC ([@ref-4]).\n\nThe Timing of Rectal Cancer Response to CRT trial was designed to investigate the effect of adding an increasing number of cycles mFOLFOX6 (folinic acid, fluorouracil and oxaliplatin) after CRT and lengthening the CRT-to-resection interval on the rate of pCR in patients with LARC. The results showed that adding consolidation chemotherapy after CRT and delaying surgery increased the pCR rate without increasing surgical complications ([@ref-4]; [@ref-12]). [@ref-8] found that pCR and DS rate could be marginally improved with two cycles XELOX consolidation chemotherapy after preoperative CRT before TME. Will the above phase II trials result be the same in the real-world? Therefore, we aimed to investigate the effect of adding consolidation chemotherapy between CRT and resection on the pCR as well as surgical outcomes.\n\nMaterials and Methods\n=====================\n\nPatients\n--------\n\nThis study was a single-center retrospective cohort study based on patients with clinical stage II (T3--4, N0) or III (any T, N1--2) LARC who underwent neoadjuvant CRT followed by surgery at Shandong Cancer Hospital, affiliated with Shandong University, from January 2014 to December 2019. Eligibility criteria included a diagnosis of LARC with confirmed by pathological diagnosis; age was \u226518 years; invasive rectal adenocarcinoma within 12 cm from the anal verge; no history of receiving chemotherapy or pelvic radiation for rectal cancer; Eastern Cooperative Oncology Group (ECOG) performance status (PS) \u22642; no distant metastasis. The exclusion criteria were as follows: existence of distant metastases, concurrently diagnosed colon cancer, history of other cancer, other serious life conditions. Patients with a history of pelvic radiation, recurrent rectal cancer, metastatic disease, other primary tumors within the previous 5 years were excluded. The clinical T and N staging was using MRI scans according to the American Joint Committee on Cancer's (AJCC) seventh staging system. Imaging studies including abdomen and pelvis CT, and rectal MRI were applied before treatments to assess primary tumor and exclude distant metastases. The study was approved by the Institutional Research Ethics Committee of Shandong Cancer Hospital in Jinan, Shandong (approval number \\#2019-23). The requirement for informed consent was waived because of the retrospective nature of the study. All methods were performed following the relevant guidelines and regulations.\n\nTreatment strategy\n------------------\n\nEach patient received initially radiotherapy (DT 50 Gy/25 fractions) concurrent with capecitabine (825 mg/m^2^ twice daily for 5 days/week during CRT). The choice of whether to consolidation chemotherapy was according to the physician's decision. For patients without consolidation chemotherapy (group neoadjuvant chemoradiation therapy (nCRT)), TME was treated 4--8 weeks after the completion of CRT. For patients with consolidation chemotherapy (group CCT), two cycles of consolidation chemotherapy in XELOX (Oxaliplatin 130 mg/m^2^ day 1, Capectiabine 1,000 mg/m^2^ twice daily days 1--14 every 3 weeks) was administered between CRT and TME. Repeat every 3 weeks to a total of two cycles.\n\nOutcomes\n--------\n\nWe used the pCR rate to evaluate the efficacy of adding consolidation chemotherapy. The pCR was defined as the complete absence of tumor cells in the resected specimen and lymph nodes (ypT0N0). The tumor downstaging (DS) was defined as the proportion of patients with ypT0-2N0M0 (ypStage 0 or I) for those who underwent TME and assessed using the AJCC 7th staging system. In this study, we reported pCR and DS rate for outcome results. Besides, adverse events (AEs) during and post treatment were also compared between the treatment groups to evaluate the safety of consolidation chemotherapy.\n\nStatistical methods\n-------------------\n\nAll categorical variables were assessed using the one-sided Fisher's exact test or Chi-squared test. Continuous variables were tested using the Student's *t-*test. To adjust the unbalanced variables for the primary endpoint, logistic regression analysis and stratified analysis were performed, and one-sided *P*-value was derived. *P* \\< 0.05 was regarded as statistically significant in two-tailed tests. IBM SPSS STATISTICS version 25.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis.\n\nResults\n=======\n\nPatient characteristics\n-----------------------\n\nTotally 144 eligible LARC patients with CRT were treated in our cancer institute from January 2014 to December 2019 were retrospectively reviewed. Eighty-three (58%) and 61 (42%) patients were categorized into with the consolidation chemotherapy group and without consolidation chemotherapy group respectively. The median age was 54 years (range 20--78). All of the patients were adenocarcinoma. A summary of the baseline for the 144 patients' characteristics like the patient demographics and tumor characteristics of the two groups is presented in [Table 1](#table-1){ref-type=\"table\"}. No differences were found in demographics, tumor characteristics, or clinical and pathological features between two groups, indicating balanced groups in our study.\n\n10.7717/peerj.9513/table-1\n\n###### Baseline for the 144 patients' characteristics in consolidation chemotherapy (CCT) and neoadjuvant chemoradiotherapy (nCRT).\n\n![](peerj-08-9513-g001)\n\n                            nCRT (*n* = 61)     CCT (*n* = 83)      *P* value\n  ------------------------- ------------------- ------------------- -----------\n  Age (years)                                                       \n  Median (range)            56 (24--78)         54 (20--72)         0.34\n  Gender                                                            \n  Male                      34 55.7%            57 68.6%            0.11\n  Female                    27 44.3%            26 31.4%            \n  BMI (kg/m^2^)                                                     \n  Median (range)            23.5 (17.3--30.4)   24.2 (18.4--33.9)   0.34\n  ECOG performance status                                           \n  0                         28 45.9%            37 44.5%            0.87\n  1                         33 54.1%            46 55.5%            \n  Clinical T stage                                                  \n  T3                        17 27.8%            33 39.7%            0.21\n  T4                        44 72.2%            50 60.3%            \n  Clinical N stage                                                  \n  N0                        12 19.6%            12 14.4%            0.48\n  N1                        38 62.2%            50 60.2%            \n  N2                        11 18.2%            21 25.4%            \n  Clinical stage                                                    \n  II                        12 19.7%            12 14.5%            0.41\n  III                       49 80.3%            71 85.5%            \n  CEA (ng/ml)                                                       \n  Median (range)            4.3(0.93-152.2)     6.23(0.24-247.6)    0.32\n  \\<5 ng/ml                 33 54.1%            38 45.7%            \n  \u22655 ng/ml                  28 45.9%            45 54.3%            \n\n**Note:**\n\nBMI, body mass index; ECOG, Eastern Cooperative Oncology Group; CEA, carcinoembryonic antigen.\n\nSurgical procedures and pathologic outcome\n------------------------------------------\n\nAll of the patients eventually underwent TME with R0 resection. As shown in [Table 2](#table-2){ref-type=\"table\"}, there was no significant difference in the estimated blood loss and operation time during surgery. In the postoperative pathological staging, group nCRT had a significantly higher percentage of ypT4 than group CCT (63.9% vs 28.9%). The pCR rate in the CCT group was significantly higher than in the nCRT (19.3% vs 4.9%, *P* = 0.01). Also, DS rate (45.8% vs 24.6%, *P* = 0.009) was significantly higher in the CCT group than in the nCRT group. When analyzing time interval as a categorical variable, the highest proportion of patients in the nCRT group underwent TME within 6 weeks (60.6%) or between 6 and 10 weeks (32.7%) while in the CCT group most patients had TME between 6 and 10 weeks (61.4%) or after 10 weeks (27.7%) (*P* \\< 0.001).\n\n10.7717/peerj.9513/table-2\n\n###### Differences in pathology outcomes between consolidation chemotherapy (CCT) and neoadjuvant chemoradiotherapy (nCRT).\n\n![](peerj-08-9513-g002)\n\n                                                     nCRT (*n* = 61)   CCT (*n* = 83)                            *P* value\n  -------------------------------------------------- ----------------- ----------------------------------------- -----------\n  Interval time (weeks)                                                                                          \n  Median (range)                                     5 (3--14)         9 (2--16)                                 \\<0.001\n  \\<6                                                37 60.6%          9[\\*](#table-2fn1){ref-type=\"fn\"} 10.9%   \n  6--10                                              20 32.7%          51 61.4%                                  \n  \\>10                                               4 6.7%            23 27.7%                                  \n  Approach                                                                                                       \n  Laparoscopic                                       18 29.6%          40 48.1%                                  0.02\n  Open                                               43 70.4%          43 51.9%                                  \n  Operation mode[\\*\\*](#table-2fn2){ref-type=\"fn\"}                                                               \n  APR                                                31 50.8%          34 40.9%                                  0.16\n  LAR                                                20 32.7%          40 48.1%                                  \n  Hartmann resection                                 10 16.5%          9 11%                                     \n  Operation time (min)                                                                                           \n  Median (range)                                     215 (90--420)     210 (120--480)                            0.53\n  Estimtated blood loss (g)                                                                                      \n  Median (range)                                     50 (20--100)      50 (20--500)                              0.23\n  ypStage                                                                                                        \n  0                                                  3 4.9%            16 19.4%                                  0.008\n  I                                                  12 19.7%          25 30.1%                                  \n  II                                                 25 41%            28 33.7%                                  \n  III                                                21 34.4%          14 16.8%                                  \n  ypTstage                                                                                                       \n  0                                                  3 4.9%            16 19.2%                                  0.001\n  1                                                  3 4.9%            3 11%                                     \n  2                                                  11 18%            29 27.7%                                  \n  3                                                  5 8.3%            11 13.2%                                  \n  4                                                  39 63.9%          24 28.9%                                  \n  ypNstage                                                                                                       \n  0                                                  40 65.5%          69 83.1%                                  0.11\n  1                                                  12 19.6%          8 9.6%                                    \n  2                                                  9 5.9%            6 7.3%                                    \n  pCR rate                                           3 4.9%            16 19.3%                                  0.01\n  DS rate (ypStage 0 + I)                            15 24.6%          38 45.8%                                  0.009\n\n**Notes:**\n\nNine patients had only one cycle of the XELOX because they refused the second cycle of XELOX.\n\nPatients with T4b tumors with adjacent organ invasion underwent combined resection.\n\nAPR, abdominoperineal resection; LAR, low anterior resection; pCR, pathologic complete response; DS, downstaging.\n\nIn stratified analysis, CCT was associated with a higher pCR rate in patients with a cT4, cN1 and 6--10 weeks of time interval ([Table 3](#table-3){ref-type=\"table\"}).\n\n10.7717/peerj.9513/table-3\n\n###### Stratified analysis of the association between consolidation chemotherapy (CCT) or neoadjuvant chemoradiotherapy (nCRT) and pathological complete response (pCR) by clinical tumor stage, clinical nodal stage and categories of time interval.\n\n![](peerj-08-9513-g003)\n\n  Strata                     pCR in CCT   pCR in nCRT   *P*-value\n  -------------------------- ------------ ------------- -----------\n  Clinical tumor stage                                  \n  cT3                        2/17         4/32          0.09\n  cT4                        1/44         12/50         0.002\n  Clinical nodal stage                                  \n  cN0                        2/12         3/12          0.61\n  cN1                        1/38         10/50         0.01\n  cN2                        0/11         3/21          0.18\n  Time interval to surgery                              \n  \\<6 weeks                  2/35         0/9           0.46\n  6--10 weeks                0/21         11/51         0.02\n  \\>10 weeks                 1/5          5/23          0.93\n\nIn multivariable analysis, the probability of pCR was significantly higher in the CCT group compared to the nCRT group (Odds Ratio = 4.91, 95% CI \\[1.01--23.79\\], *P* = 0.048 adjusted for gender, categories of CEA, cT-stage, cN-stage and categories of time interval) ([Table 4](#table-4){ref-type=\"table\"}).\n\n10.7717/peerj.9513/table-4\n\n###### Multivariable analysis of the association between consolidation chemotherapy (CCT) or neoadjuvant chemoradiotherapy (nCRT) and pathological complete response (pCR).\n\n![](peerj-08-9513-g004)\n\n                        Crude OR (95% CI)      Adjusted OR (95% CI)   *P* value\n  --------------------- ---------------------- ---------------------- -----------\n  Neoadjuvant therapy                                                 \n  CRT                   Ref.                   Ref.                   \n  CCRT                  4.6 \\[1.3--16.6\\]      4.91 \\[1.01--23.79\\]   0.048\n  Sex                                                                 \n  Male                  Ref.                   Ref.                   \n  Female                0.47 \\[0.18--1.25\\]    2.85 \\[0.93--8.75\\]    0.07\n  CEA                                                                 \n  \\<5 (ng/ml)           Ref.                   Ref.                   \n  \\>5 (ng/ml)           0.94 \\[0.35--2.52\\]    0.88 \\[0.28--2.67\\]    0.82\n  Clinical T-stage                                                    \n  cT3                   Ref.                   Ref.                   \n  cT4                   1.21 \\[0.44--3.31\\]    1.38 \\[0.44--4.33\\]    0.58\n  Clinical N-stage                                                    \n  cN0                   Ref.                   Ref.                   \n  cN1                   0.54 \\[0.17--1.75\\]    0.49 \\[0.12--2.01\\]    0.32\n  cN2                   0.39 \\[0.08--1.84\\]    0.23 \\[0.38--1.42\\]    0.11\n  Time interval                                                       \n  \\<6 weeks             Ref.                   Ref.                   \n  6--10 weeks           3.79 \\[0.79--17.97\\]   1.97 \\[0.30--12.9\\]    0.47\n  \\>10 weeks            5.72 \\[1.07--30.77\\]   4.41 \\[0.58--33.2\\]    0.14\n\n**Note:**\n\nCEA, carcinoembryonic antigen; OR, odds ratio.\n\nSafety\n------\n\n[Table 5](#table-5){ref-type=\"table\"} describes AEs during the two groups. Nausea, diarrhea, anal pain, leukopenia and neutropenia were more common in CCT group than in nCRT group, whereas thrombocytopenia and anorexia were more common in group nCRT. Any grade 1--2 AE occurred in 57 of 61 patients (93.4%) in group nCRT and 79 of 83 (95.1%) in group CCT. Any grade 3 AE occurred in 1 of 61 patients (1.6%) in group nCRT and 2 of 83 (2.4%) in group CCT. No grade 4 AE was reported during treatment in both groups.\n\n10.7717/peerj.9513/table-5\n\n###### Differences in adverse events (AEs) between consolidation chemotherapy (CCT) and neoadjuvant chemoradiotherapy (nCRT).\n\n![](peerj-08-9513-g005)\n\n                            nCRT (*n* = 61)   CCT (*n* = 83)              \n  ------------------------- ----------------- ---------------- ---------- --------\n  Any toxicities            57 93.4%          1 1.6%           79 95.1%   2 2.4%\n  Nausea                    6                 0                14         0\n  Fatigue                   4                 0                0          0\n  Diarrhea                  5                 0                21         0\n  Anal pain                 4                 0                17         1\n  Peripheral neuropathy     0                 0                2          0\n  Anorexia                  7                 0                5          0\n  Abdominal pain            5                 0                9          0\n  Constipation              2                 0                3          0\n  Urinary tract infection   1                 0                1          0\n  Myalgia                   1                 0                1          0\n  Sepsis                    0                 0                0          0\n  Leukopenia                8                 1                26         2\n  Hand foot syndrome        0                 0                2          0\n  Neutropenia               6                 0                10         0\n  Anemia                    2                 0                3          0\n  Thrombocytopenia          3                 0                0          0\n  Hyponatremia              0                 0                1          0\n\nAs for surgical complications, there was no significant difference between postoperative morbidities in the two groups ([Table 6](#table-6){ref-type=\"table\"}). Incisional pain was the most common morbidity in both groups (21.3% and 18,1%). Surgical-site infection was more common in group nCRT than group CCT (14.7% vs 7.2%).\n\n10.7717/peerj.9513/table-6\n\n###### The comparison of postoperative complications between the consolidation chemotherapy (CCT) and neoadjuvant chemoradiotherapy (nCRT).\n\n![](peerj-08-9513-g006)\n\n                                           nCRT (*n* = 61)   CCT (*n* = 83)\n  ---------------------------------------- ----------------- ----------------\n  Overall operative morbidity: *n* (%)     33 (54.1)         41 (49.3)\n  Intestinal obstruction: *n* (%)          0 (0)             0 (0)\n  Surgical-site infection: *n* (%)         9 (14.7)          6 (7.2)\n  Transient urinary dysfunction: *n* (%)   0 (0)             0 (0)\n  Pneumonitis: *n* (%)                     1 (1.6)           3 (3.6)\n  Perineal wound infection: *n* (%)        2 (3.2)           7 (8.4)\n  Intraabdominal abscess: *n* (%)          3 (4.8)           7 (8.4)\n  Enteritis: *n* (%)                       1 (1.6)           0 (0)\n  Catheter infection: *n* (%)              1 (1.6)           2 2.4\n  Anastomotic leakage: *n* (%)             3 (4.8)           1 (1.2)\n  Incisional pain: *n* (%)                 13 (21.3)         15 (18.1)\n\nDiscussion\n==========\n\nHow to increase pCR rate after neoadjuvant therapy and therefore to expand the number of LARC patients who could potentially benefit from a watch-and-wait strategy has been a hot research. In our study, we retrospectively used real-world data to demonstrate that the consolidation chemotherapy significantly increased the pCR in patients with LARC. Just as importantly, we showed that this approach appears to be safe from oncology and surgical perspective. It does not increase the risk of AEs or surgical complications before and after treatment. If these findings can be applied clinically, more patients with LARC will be eligible for organ preservation, which will avoid surgical sequelae and improve quality of life.\n\nCurrently, some prospective phase II trials had focused on the efficacy of consolidation chemotherapy between neoadjuvant CRT and TME in patients with LARC. In a nonrandomized phase II clinical trial from [@ref-4], patients received two, four, or six cycles of modified FOLFOX between neoadjuvant CRT and resection at different interval time of 12, 16, and 20 weeks, which resulted in improved pCR of 25%, 30% and 38% respectively compared with the pCR of 18% for patients without consolidation chemotherapy. In our study, the pCR rate was increased to 19.3% in the group CCT compared with the pCR of 4.9% for patients without consolidation chemotherapy. The pCR rate was 4.2% in our study for patients without consolidation chemotherapy, which is smaller than the data with 8--24% reported before ([@ref-20]). Although the pCR rate in this study was lower than reported in both groups (4.9% and 14.3%), consolidation chemotherapy was related to enhanced tumor regression. The pCR rates in our study are lower due to that the patients in our study are unusually adverse in that 65% are cT4 which normally is associated with a low chance of achieving a pCR ([@ref-8]). Improved pCR and downstaging rates could be the result of the study treatment, but we cannot exclude that it may also have been affected by a significant delay of surgery in group CCT compared with group nCRT. Results from several retrospective studies have confirmed that longer chemoradiotherapy to surgery intervals are associated with a higher percentage of patients achieving a pathologic complete response ([@ref-1]; [@ref-3]; [@ref-5]; [@ref-6]; [@ref-19]; [@ref-13]; [@ref-16]; [@ref-21]; [@ref-22]). In the Lyon R90-01 study ([@ref-3]), Francois and his colleagues found that 14% of patients who underwent surgery after 8 weeks of radiotherapy achieved a pCR, while only 7% of patients who underwent surgery after 4 weeks of radiotherapy achieved a pCR. However, considering the results of a large observational study that showed a 2--4% increase in pCR rates for each 1-week delay after the start of CRT within 17 weeks ([@ref-16]), it is possible to explain the 14% increase in pCR rates for a delay of 2 weeks. The purpose of our study was to consider whether the addition of consolidation chemotherapy could increase the proportion of pCR patients regardless of the time interval. Inevitably, in a subset of patients, the addition of consolidation chemotherapy between neoadjuvant chemoradiotherapy and TME resulted in an unintended increase in the time interval. Therefore, we used stratified analysis to exclude the time interval difference between the two groups. The stratified analysis results show that CCT was associated with a higher pCR rate in patients with 6--10 weeks of time interval ([Table 3](#table-3){ref-type=\"table\"}). Also, the multivariable analysis which used logistic regression also suggested that the probability of pCR patients was significantly higher in the group CCT compared to the standard treatment group (Odds Ratio = 4.91, 95% CI \\[1.01-- 23.79\\], *P* = 0.048 adjusted for cT-stage, cN-stage and categories of time interval) ([Table 4](#table-4){ref-type=\"table\"}). In a randomized phase II clinical trial, [@ref-8] implemented a randomized phase II study, which suggested pCR and DS rate could be marginally improved with two cycles XELOX consolidation chemotherapy after preoperative CRT before TME. The interval time between the two groups were both 6--10 weeks but included a difference of only 10 days. Unlike our results, in their study, there was a numerical difference in pCR rates between the two groups, but it was not statistically significant (13.6% vs 5.8%, *P* = 0.167). However, the authors did not conduct further analysis to explore whether the difference in PCR was caused by the addition of consolidation chemotherapy or the extension of interval time.\n\nAs time prolongs, the potential development of pelvic fibrosis that could increase postoperative complications ([@ref-2]). In the current study, patients' operation time, intraoperative blood loss and postoperative complications were also collected. As shown in the [Tables 2](#table-2){ref-type=\"table\"} and [3](#table-3){ref-type=\"table\"}, there was no significant difference in operation time (*P* = 0.53) or intraoperative blood loss (*P* = 0.23) between the two groups. For surgical complications, overall surgical complications in group nCRT were slightly higher than those in group CCT (54.1% vs 49.3%, *P* = 0.57), the most common postoperative complication in both groups was incisional pain, and the incidence of surgical-site infection in group nCRT was slightly higher than that in group CCT.\n\nAdverse events are an important part of evaluating experimental safety, so not only we reported surgical-related complications, we evaluated the incidence of AEs during chemoradiotherapy. In our study, the top three AEs were Leukopenia, Anorexia and Nausea in group nCRT, while Leukopenia, Diarrhea and Anal pain were in group CCT. The incidence of any grade 1--2 AE in our study was 93.4% in group nCRT and 95.1% in group CCT (*P* = 0.93), which was slightly higher than the previous randomized controlled trial (91% vs 92%). Considering the fact that the dose of oxaliplatin and capecitabine in our trial (100 mg/m^2^ and 850 mg/m^2^, respectively) was higher than in the previous trial (130 mg/m^2^ and 1,000 mg/m^2^, respectively), the incidence of AEs was similar to the previous trial which suggested that consolidation chemotherapy is a safe way to improve the proportion of pCR for patients with LARC.\n\nOur study has several limitations. First, the use of a retrospective database limited our capacity to investigate other sources of potential bias. Second, the number of cases is limited, and it needs expanding the number of patients to confirm our conclusions. Approximately 14% increment in the DS rate and 21% increment in the pCR rate could be significant, but the sample size was small eventually may permitted a room for type-I and type-II error.\n\nUp to now, modify and optimize the radiotherapy dose, chemotherapy regimen, treatment duration and interval time is still needed to seek the optimal treatment mode in patients with LARC. Nevertheless, the current study based real-world data demonstrated the role of consolidation chemotherapy may improve pCR rate after neoadjuvant CRT, adding a real-world data piece of evidence to support the effectiveness of consolidation chemotherapy in patients with LARC. Further randomized phase III trial is warranted to explore the role of consolidation chemotherapy.\n\nConclusions\n===========\n\nIn conclusion, our study demonstrated that adding neoadjuvant consolidation chemotherapy after CRT significantly increased the pCR rate and did not increase the AEs or surgical complications in patients with LARC.\n\nSupplemental Information\n========================\n\n10.7717/peerj.9513/supp-1\n\n###### Raw data.\n\n###### \n\nClick here for additional data file.\n\nAdditional Information and Declarations\n=======================================\n\nThe authors declare that they have no competing interests.\n\n[Jin Cui](#author-1){ref-type=\"contrib\"} analyzed the data, prepared figures and/or tables, and approved the final draft.\n\n[Xue Dou](#author-2){ref-type=\"contrib\"} performed the experiments, prepared figures and/or tables, and approved the final draft.\n\n[Yanlai Sun](#author-3){ref-type=\"contrib\"} conceived and designed the experiments, authored or reviewed drafts of the paper, and approved the final draft.\n\n[Jinbo Yue](#author-4){ref-type=\"contrib\"} conceived and designed the experiments, performed the experiments, authored or reviewed drafts of the paper, and approved the final draft.\n\nThe following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers):\n\nThe study was approved by the Institutional Research Ethics Committee of Shandong Cancer Hospital in Jinan, Shandong (\\#2019-23).\n\nThe following information was supplied regarding data availability:\n\nThe raw data are available as a [Supplemental File](#supplemental-information){ref-type=\"supplementary-material\"}.\n"}
+{"text": "Introduction\n============\n\nStroke is a leading cause of disability and death in most countries \\[[@b1-jos-17-3-302]\\]. Control of modifiable risk factors is the most effective approach to decreasing the burden of stroke \\[[@b2-jos-17-3-302]-[@b4-jos-17-3-302]\\]. According to an international case-controlled study (INTERSTROKE), ten risk factors account for about 90% of the risk of stroke, and our efforts at stroke prevention should be concentrated on reducing blood pressure (BP) and promoting smoking cessation, physical activity, and a healthy diet \\[[@b5-jos-17-3-302]\\].\n\nHowever, the impact of modifiable risk factors on stroke incidence is not equal between different ethnic groups, sexes, or age groups \\[[@b1-jos-17-3-302],[@b6-jos-17-3-302]-[@b9-jos-17-3-302]\\]. For example, in the INTERSTROKE study, hypertension was more strongly associated with strokes in women than in men, and in people aged 45 years or younger than in those aged 45 years or older \\[[@b5-jos-17-3-302]\\]. The recommended strategies for preventing strokes in women with hypertension are different from those in men with hypertension \\[[@b7-jos-17-3-302]\\], and the optimal target BP is higher in people aged 60 years or older than in those less than 60 years \\[[@b10-jos-17-3-302]\\]. Additionally, the risk of strokes in diabetic patients varies with age \\[[@b8-jos-17-3-302]\\] and sex \\[[@b9-jos-17-3-302]\\]. Therefore, stroke prevention strategies in a specific populations would be more effective if they were designed based on knowledge of the population attributable risks (PARs) of individual risk factors in each age- and sex-specific subpopulation because the PAR indicates the proportion of a disease that could be prevented by eliminating the risk factor of interest in a specific population \\[[@b11-jos-17-3-302]\\].\n\nThe burden of stroke, as compared to coronary heart disease (CHD), is disproportionately high in East Asian countries, including China, Japan, and South Korea \\[[@b12-jos-17-3-302]\\]. Furthermore, these countries are the world's fastest aging countries \\[[@b13-jos-17-3-302]\\]; their burden of stroke is therefore expected to increase substantially in the near future unless effective prevention strategies are launched. However, there is a paucity of information about age- and sexspecific PARs of the major risk factors in Asian populations \\[[@b14-jos-17-3-302]-[@b17-jos-17-3-302]\\].\n\nIn this study, we aimed to estimate the age- and sex-specific PARs of the major risk factors for strokes in the Korean population.\n\nMethods\n=======\n\nRecruitment of cases\n--------------------\n\nThe Clinical Research Center for Stroke (CRCS) project sponsored by the Korean Government was established in 2006 to facilitate a multicenter collaborative stroke research program in Korea and to develop and implement clinical practice guidelines for stroke treatment and prevention. The fifth section of CRCS (CRCS-5), which was dedicated to epidemiologic research, constructed a web-based, prospective stroke registry database and, in April 2008, began to collect clinical information on patients hospitalized for strokes. Using the CRCS-5 registry database, cases were selected from acute ischemic stroke patients admitted within 7 days of onset. Ischemic stroke was defined as acute focal neurological deficits and relevant ischemic brain lesions found on computed tomography or diffusion-weighted magnetic resonance imaging. Details of the CRCS-5 database have been reported elsewhere \\[[@b18-jos-17-3-302]\\].\n\nThis study focused on ischemic stroke because 1) consecutive registration of hemorrhagic stroke patients was not guaranteed at some centers and 2) acute and hemorrhagic strokes have fairly different risk factor profiles \\[[@b19-jos-17-3-302]\\].\n\nAmong the 6,023 patients hospitalized between December 1, 2008 and June 30, 2010, we excluded 274 who were admitted \u2265 7 days after stroke onset, 135 with hemorrhagic stroke, and 340 with a transient ischemic attack. Patients younger than 19 years or older than 90 years were also excluded because matched control cases of these age ranges were rare in the Korea National Health and Nutrition Examination Survey (KNHANES) database. Patients with missing height or weight information (n = 101) were also excluded. Ultimately, 5,107 cases were eligible for matching ([Figure 1A](#f1-jos-17-3-302){ref-type=\"fig\"}).\n\nRecruitment of controls\n-----------------------\n\nThe 4th (2007-2009) and 5th (2010-2012) KNHANES was conducted by the Korean Center for Disease Control to investigate the health status of the Korean population through annual nationwide representative sampling \\[[@b20-jos-17-3-302]\\]. The KNHANES employed a stratified, multi-stage probability sampling design; subjects were selected from sampling units predicated on geographical area, sex, and age, based on household registers. Among the subjects surveyed in 2008, 2009, and 2010, 19,435 subjects, aged 19-90 years with no missing information about risk factors of interest, were selected for matching ([Figure 1B](#f1-jos-17-3-302){ref-type=\"fig\"}).\n\nMatching process\n----------------\n\nEach case was matched to one control by age (\u00b1 3 years) and sex. Despite limiting the age range to 19-90 years, 364 elderly patients were excluded due to a lack of corresponding controls. Therefore, 4,743 case-control sets were used for the analysis.\n\nThe CRCS-5 registry got approval for collection and analysis of anonymized clinical data without patients' consent for the purpose of monitoring and improving the quality of stroke care from the institutional review boards (IRBs) of the participating centers. We obtained additional approval for the use of the registry database and the KNHANES database to conduct this study.\n\nDefinition and selection of vascular risk factors\n-------------------------------------------------\n\nRisk factors of interest were chosen based on their relative importance in the primary prevention guidelines \\[[@b4-jos-17-3-302]\\] and the comparability of their definitions between the CRCS-5 registry and the KNHANES. Hypertension, diabetes, hypercholesterolemia, smoking, prior stroke, CHD, and obesity were selected. Despite its importance as a stroke risk factor, atrial fibrillation was not included because the KNHANES did not gather information on it.\n\nHypertension was defined as a mean systolic BP \u2265 140 mmHg or a mean diastolic BP \u2265 90 mmHg in repeated measures, prior use of antihypertensive medications, or a self-reported physician diagnosis. BP measurements used for the diagnosis of hypertension in the cases were obtained after patients were medically and neurologically stabilized. Diabetes was defined as a fasting blood glucose \u2265 126 mg/dL, prior use or prescription-at-discharge of hypoglycemic medications, or a self-reported physician diagnosis. Hypercholesterolemia was defined as prior use of lipid lowering agents, fasting total serum cholesterol \\> 240 mg/dL, or a self-reported physician diagnosis. For the cases, smoking was defined as a current smoker who had smoked at least one cigarette in the last month, whereas for the controls it was defined as a current smoker with a lifetime history of smoking \u2265 100 cigarettes. Prior stroke was defined as a stroke history diagnosed by a physician. CHD was defined as a self-reported physician diagnosis for angina pectoris or myocardial infarction. A history of coronary intervention (angioplasty and/or stent insertion) and of coronary arterial bypass graft surgery was added to the definition of CHD only for the cases. Obesity was diagnosed according to the World Health Organization guidelines. Subjects whose body mass index (BMI) was \\< 18.5 kg/m^2^, 25 to 29.9 kg/m^2^, or \u2265 30 kg/m^2^ were classified as underweight, overweight, or obese, respectively.\n\nStatistical analysis\n--------------------\n\nContinuous variables are presented as means \u00b1 SD or medians (interquartile range, IQR), and categorical variables are reported as proportions. Before estimating the PAR of a risk factor, we determined the strength of the association between the risk factor and ischemic strokes in all subjects and within each age and sex subgroup by using odds ratios (ORs) and 95% confidence intervals (CIs). Conditional logistic regression was applied to the estimation of ORs and CIs for the matched sets and adjustments were made for risk factors of interest and age. PARs of individual risk factors were calculated using the following formula: *PAR=PR (OR-1)/\\[1+PR(OR-1)*\\] \\[[@b21-jos-17-3-302]\\], where PR represents the age- and sex-specific prevalence of a risk factor in the Korean population, estimated from the age- and sex-specific prevalence in the 4th KNHANES and standardized according to the age and sex structure of the general population in the 2010 Census of Korea ([Table 1](#t1-jos-17-3-302){ref-type=\"table\"}). OR represents the age- and sex-specific OR obtained as described above. CIs of the PARs were estimated using a substitution method \\[[@b22-jos-17-3-302]\\].\n\nTo estimate age-specific PARs, age was categorized into 3 groups; \u2264 45 years (young), 46 to 65 years (middle-aged), and \u2265 66 years (elderly). In men and women, age-specific PARs were estimated by using the age-standardized prevalence within each age group ([Tables 2](#t2-jos-17-3-302){ref-type=\"table\"}, [3](#t3-jos-17-3-302){ref-type=\"table\"}). A two-sided P value \\< 0.05 was considered statistically significant. All statistical analyses were performed with SAS version 9.3 (SAS, Cary, NC, USA).\n\nResults\n=======\n\nOf the 4,743 cases in this study, 60% were men, and the mean age was 63.7 \u00b1 11.6 years in men and 69.2 \u00b1 11.0 years in women. The proportion of cases \u2264 45 years, 46-65 years, and \u2265 66 years were 5.7%, 35.6%, and 58.7%, respectively. The median National Institute of Health Stroke Scale score was 4 (IQR, 2-8). Hypertension was the most frequent risk factor (63.6% in men, 70.1% in women), followed by diabetes and smoking ([Table 4](#t4-jos-17-3-302){ref-type=\"table\"}).\n\nWhen combining all age groups, hypertension, diabetes, stroke history, and CHD related to an increased risk of strokes in both sexes, whereas hypercholesterolemia and smoking were related to an increased risk in men but not in women ([Table 4](#t4-jos-17-3-302){ref-type=\"table\"}). Being underweight increased the risk of strokes in women while being overweight was protective in both sexes.\n\nAge- and sex-specific ORs and PARs of individual risk factors are presented in [Table 5](#t5-jos-17-3-302){ref-type=\"table\"}. The OR of hypertension was highest in the young and decreased with age in both sexes. However, its PAR was similar in most of age and sex subgroups because the prevalence of hypertension increased with age ([Tables 1](#t1-jos-17-3-302){ref-type=\"table\"}-[3](#t3-jos-17-3-302){ref-type=\"table\"}). Approximately one fourth of strokes were attributed to hypertension across all age groups except for young women.\n\nFor cases involving diabetes, the OR was highest in young men but the PAR was highest in elderly women. The OR decreased with age in men, but was similar between middle-aged and elderly women.\n\nStroke history was a strong risk factor in most age and sex subgroups, especially in the elderly, where the PAR was 19.7% in elderly men and 17.3% in women. The OR and PAR of CHD were highest in elderly women, which on average, were twice those in elderly men (2.99 vs. 1.67; 9.5% vs. 4.1%).\n\nThe increased risk of strokes due to smoking was highest in middle-aged women (OR, 2.66; 95% CI, 1.39-5.12). However, the PAR for strokes due to smoking was highest in young men. About half of strokes in young men and more than one third of strokes in middle-aged men were attributed to smoking. The magnitude of the increased risk and the contribution to stroke incidence by smoking were mitigated in elderly men.\n\nHypercholesterolemia increased the risk of strokes only in elderly men. The association of BMI with strokes was inconsistent throughout all age and sex subgroups. Being underweight did not increase the risk in any subgroup. Obesity increased the risk only in young women. Being overweight in middle-aged men and obesity in middle-aged women were both associated with a decreased risk of strokes.\n\nTo summarize the major contributions of the risk factors among age and sex subgroups, smoking (PAR, 45.1%) and hypertension (28.5%) were the major contributing risk factors in young men. There were no risk factors with a PAR greater than 10% in young women. In middle-aged men, smoking was still the greatest risk factor (37.4%), followed by hypertension (22.7%). In middle-aged women, hypertension was the greatest risk factor (22.7%). The contributions of diabetes and stroke history were larger in the middle-aged than in the young.\n\nIn the elderly, the contribution of hypertension was similar in both sexes and not different from that in the middle-aged. The contribution of stroke history and diabetes was also remarkable in the elderly.\n\nDiscussion\n==========\n\nOur study shows that the contributions of the major risk factors to ischemic strokes differ by age and sex. Therefore, tailored preventive strategies based on this knowledge may be an effective approach to stroke prevention at a population level.\n\nStroke incidence and mortality have declined over the past decades in developed countries. This could be explained by wide-spread implementation of public health programs for the control of hypertension, diabetes, and dyslipidemia, and the cessation of smoking \\[[@b23-jos-17-3-302],[@b24-jos-17-3-302]\\]. Stroke mortality declined by 34.2% in Korea from 2002 to 2011 \\[[@b25-jos-17-3-302]\\]. However, the declining trend in stroke mortality leveled out after 2011, with only a 0.8% decrease over the next 3 years, which could be at least partly attributed to the fast-aging population. Thus, it is time to develop new preventive strategies using the knowledge obtained from the current study.\n\nAlthough several studies have reported the PARs of individual risk factors for strokes in Korean population \\[[@b26-jos-17-3-302]-[@b28-jos-17-3-302]\\], only one study \\[[@b28-jos-17-3-302]\\] described the age- and sex-specific PARs. The relative risks used in that study were mostly obtained from non-Koreans, and therefore, might not reflect the properties of Korean populations. The estimated age- and sex-specific PARs in the current study indicate that there should be an emphasis on BP control in the entire population, on smoking cessation in young- and middle-aged men, on diabetes control in the middle-aged and elderly of both sexes, and on the secondary prevention of strokes in the elderly of both sexes.\n\nAlthough hypertension is the greatest contributing risk factor for ischemic strokes \\[[@b4-jos-17-3-302]\\], the proportion of stroke incidence attributable to hypertension varies depending on the demographic features (age, sex, and ethnicity) of the study population \\[[@b5-jos-17-3-302],[@b6-jos-17-3-302],[@b15-jos-17-3-302],[@b29-jos-17-3-302],[@b30-jos-17-3-302]\\]. With the exception of young women, the age- and sex-specific PARs of hypertension (23%-28%) in the current study were greater than the PARs among non-Hispanic white (3%) and black (16%) populations but were smaller than the PAR among Hispanics (51%), as reported in the recent Northern Manhattan Study \\[[@b6-jos-17-3-302]\\]. The contribution of hypertension to stroke incidence in Asian countries might have declined during the past decades. A recent cohort study in Japan reported that the PAR of hypertension decreased from 51% to 30% over 27 years \\[[@b30-jos-17-3-302]\\]. However, in order to sustain this declining trend we should pay attention to BP control in the elderly because of the rapid aging of populations in East Asian countries.\n\nGiven the high smoking rate in Asian men compared to other parts of the world \\[[@b16-jos-17-3-302]\\], it may be no surprise that about one-third of ischemic strokes in men were attributed to smoking. However, in contrast to the findings of a recent meta-analysis \\[[@b31-jos-17-3-302]\\], we could not find a significant association between smoking and strokes in young or elderly women. One possible explanation for the discrepancy is that this study included only ischemic stroke cases and smoking was associated with a greater risk for hemorrhagic strokes in women than in men \\[[@b31-jos-17-3-302]\\]. Additionally, our study might lack the statistical power to detect a significant association between smoking and strokes in young women due to the small sample size of this demographic subgroup. Only 21% of the general population in Korea is aware that smoking is a stroke risk factor \\[[@b32-jos-17-3-302]\\]. This underscores the urgent need to inform the public that smoking is an important risk factor for stroke, especially for young men in Korea, and to find a way to promote smoking cessation in this subpopulation.\n\nWe found a greater risk of stroke in relation to diabetes in young men (OR, 2.78) than in middle-aged (2.10) and elderly men (1.71). The increased risk of stroke among the younger diabetic men has been previously reported \\[[@b8-jos-17-3-302],[@b33-jos-17-3-302]\\]. However, we did not observe any sex differences in stroke risk related to diabetes, as reported in previous studies \\[[@b9-jos-17-3-302],[@b34-jos-17-3-302],[@b35-jos-17-3-302]\\]. A recent meta-analysis of 64 cohorts, including 775,385 individuals and 12,539 strokes, clearly showed an increased risk of stroke associated with diabetes in women; the pooled ratio of relative risks for women compared to men was 1.27 (95% CI, 1.10-1.46) \\[[@b9-jos-17-3-302]\\]. The above meta-analysis showed a lower, although statistically insignificant, relative risk of stroke for women compared to men in Asian than in non-Asian populations and that women have a lower risk of ischemic stroke than hemorrhagic stroke. This may explain, at least in part, the discrepancy between our study and the previous studies. The greatest PAR of diabetes was observed in elderly women, followed by the PARs in middle-aged and elderly men. These high PARs may be attributable to the higher prevalence in these age and sex subgroups ([Tables 2](#t2-jos-17-3-302){ref-type=\"table\"}, [3](#t3-jos-17-3-302){ref-type=\"table\"}). Thus, the importance of diabetes control should be emphasized to alleviate the burden of stroke in the elderly of both sexes and in middle-aged men.\n\nAn increased risk associated with stroke history found in our study is consistent with the findings from Western populations where the risk of a second stroke is six to nine times greater than the risk of a first stroke \\[[@b36-jos-17-3-302],[@b37-jos-17-3-302]\\]. The proportion of stroke incidence attributable to stroke history was distinctly high in the elderly and the most plausible explanation is the high prevalence in this age group. In order to alleviate the stroke burden in elderly populations, in addition to strictly controlling factors such as BP and diabetes, improving patient adherence to antithrombotic drugs may also be effective as they are known to reduce the risk of recurrent vascular events \\[[@b38-jos-17-3-302]\\].\n\nThe contributions of hypercholesterolemia and CHD were not noticeable in our study population. The weak association of hypercholesterolemia with ischemic stroke in Koreans is concordant with the findings from a meta-analysis of 29 Asia-Pacific cohorts \\[[@b17-jos-17-3-302]\\]. The PAR for CHD was measurable only in the elderly, and was comparable to that found in Western populations \\[[@b29-jos-17-3-302],[@b39-jos-17-3-302]\\]. A stronger association of CHD with strokes in elderly women than in elderly men might be attributable to either a higher mean age or a higher prevalence of atrial fibrillation in women than in men \\[[@b40-jos-17-3-302]\\]. The inconsistent association of BMI with stroke risk throughout the age and sex subgroups in this study probably resulted from the small sample size of each subgroup and the modest effect of BMI on stroke risk. The protective effect of being overweight or obese, which was seen in middle-aged men and elderly women in this study, may be accounted for by an inverse relationship between obesity and better survival rates observed in some studies of CHD and stroke \\[[@b41-jos-17-3-302],[@b42-jos-17-3-302]\\].\n\nThis study has several limitations. First, there may be a concern about the national representativeness of both of the cases and the controls. Although national representativeness of stroke cases in the CRCS-5 registry has been commented on in a previous paper \\[[@b43-jos-17-3-302]\\] and the age and sex distribution of our cases was almost identical to that of a report from the National Health Insurance Review Board in Korea ([Supplementary Table 1](#SD1){ref-type=\"supplementary-material\"}), the fact that most of the study sites were university hospitals could have introduced a selection bias. The CRCS-5 registry received a waiver of informed consent from the local IRBs and the completeness and consecutiveness of case registration was monitored aggressively through the monthly meetings of the steering committee \\[[@b18-jos-17-3-302]\\]. Therefore, a selection bias related to loss of cases in participating hospitals was less likely to occur. For controls, national representativeness was ensured by using the KNHANES data, the target population of which comprises nationally representative non-institutionalized civilians in Korea \\[[@b20-jos-17-3-302]\\]. Second, ORs observed in our study might be underestimated because our control group was not a stroke-free cohort, which limits the ability to measure the PAR of a prior stroke. However, the number of stroke survivors included was small (4% of the control population) and likely had little effect on ORs. Third, differential recall and differences in risk factor definitions between the cases and controls might have led to biased estimates of ORs. For example, the risk of stroke history or CHD might be overestimated by a recall bias in the cases. As for smoking, the strength of association might be overestimated due to a looser definition for the cases. Fourth, subjects taking statins, even though they do not have hypercholesterolemia, might be misclassified as having it because prior use of lipid lowering agents was defined as hypercholesterolemia in this study. However, this misclassification was not likely to influence ORs because the same definition was applied to both the cases and the controls. Finally, we did not analyze other modifiable risk factors, such as atrial fibrillation, physical activity, or alcohol intake, due to the unavailability of these data for the cases and the controls. In particular, given that the risk factors evaluated in this study are primarily related to atherosclerosis, the contribution of the unmeasured risk factors to the risk of stroke might be expected to be greater in the young than in the elderly ([Supplementary Table 2](#SD2){ref-type=\"supplementary-material\"}).\n\nConclusions\n===========\n\nMeasurements of the PARs of major risk factors for strokes by age and sex subgroups may be useful to help guide preventive strategies in public health programs. For example, we need to reinforce public campaigns and health education programs to reduce the smoking rate among young and middle aged populations. In the elderly, more medical surveillance to prevent stroke recurrence is required.\n\nPark TH, Ko Y, Lee SJ, Lee KB, Lee J, Han MK, Park JM, Cho YJ, Hong KS, Kim DH, Cha JK, Oh MS, Yu KH, Lee BC, Yoon BW, Lee JS, Lee J report no disclosures.\n\nBae HJ is involved as a principal investigator, a member of the steering committee, and/or a site investigator of multicenter clinical trials or clinical studies sponsored by Otsuka Korea, Bayer Korea, Boehringer Ingelheim Korea, Handok Pharmaceutical Company, SK Chemicals, ESAI-Korea, Daewoong Pharmaceutical Co. Ltd, Daichi Sankyo, AstraZeneca Korea, Dong-A Pharmaceutical, and Yuhan Corporation; served as the consultant or scientific advisory board member for Bayer Korea, Boehringer Ingelheim Korea, BMS Korea, and Pfizer Korea, and received lecture honoraria from AstraZeneca Korea, Bayer Korea, BMS Korea, Coviden Korea, and Daichi Sankyo Korea (modest).\n\nThis work was supported by a grant from the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (HI10C2020).\n\nThe authors have no financial conflicts of interest.\n\nSupplementary Material\n======================\n\n###### \n\nComparison of stroke population from the Clinical Research Center for Stroke-5 (CRCS-5) and from the Health Insurance Review and Assessment Service (HIRA)\n\n###### \n\nOdds ratios and population attributable risks according to age groups for atherosclerotic stroke and non-atherosclerotic stroke\n\n![Flow diagram for cases (A) and controls (B). ICH, intracerebral hemorrhage; SAH, subarachnoid hemorrhage; DWI, diffusion-weighted image; TIA, transient ischemic attack; KNHANES, Korea National Health and Nutrition Examination Survey.](jos-17-3-302f1){#f1-jos-17-3-302}\n\n###### \n\nPrevalence of individual risk factors in the general population according to age group\n\n                           Prevalence^[\\*](#tfn1-jos-17-3-302){ref-type=\"table-fn\"}^ %, (95% CI)                                               \n  ------------------------ ----------------------------------------------------------------------- --------------------- --------------------- ---------------------\n  Hypertension             8.09 (7.41-8.77)                                                        34.50 (33.15-35.84)   58.98 (57.05-60.91)   23.98 (23.13-24.83)\n  Diabetes                 2.78 (2.38-3.18)                                                        12.90 (11.98-13.81)   20.89 (19.21-22.56)   8.68 (8.20-9.15)\n  Smoking                  33.27 (32.11-34.44)                                                     26.62 (25.35-27.89)   16.54 (15.06-18.01)   28.73 (27.92-29.54)\n  Stroke history           0.11 (0.04-0.18)                                                        1.86 (1.49-2.24)      6.00 (5.20-6.79)      1.51 (1.35-1.68)\n  Coronary heart disease   0.33 (0.20-0.47)                                                        2.71 (2.29-3.14)      5.70 (4.91-6.48)      1.87 (1.68-2.07)\n  Hypercholesterolemia     5.45 (4.93-5.98)                                                        16.32(15.23-17.41)    17.31 (15.74-18.87)   10.74 (10.22-11.26)\n  Obesity, BMI (kg/m^2^)                                                                                                                       \n  \u200325.0-29.9               23.77 (22.75-24.79)                                                     33.44 (32.24-34.64)   28.96 (27.20-30.72)   27.73 (27.00-28.47)\n  \u2003\u2265 30                    4.23 (3.75-4.71)                                                        3.66 (3.15-4.16)      3.31 (2.62-4.00)      3.91 (3.58-4.25)\n  \u2003\\< 18.5                 6.42 (5.81-7.03)                                                        2.04 (1.61-2.46)      4.76 (4.04-5.48)      4.72 (4.35-5.09)\n  \u200318.5 \u2264 , \\< 25          65.58 (64.42-66.74)                                                     60.86 (59.59-62.14)   62.97 (62.21-64.74)   63.64 (62.81-64.47)\n\nAge- and sex-standardized prevalence using a direct standardization method.\n\nCI, confidence interval; BMI, body mass index.\n\n###### \n\nPrevalence of individual risk factors in men of the general population according to age group\n\n                           Prevalence^[\\*](#tfn2-jos-17-3-302){ref-type=\"table-fn\"}^ %, (95% CI)                                               \n  ------------------------ ----------------------------------------------------------------------- --------------------- --------------------- ---------------------\n  Hypertension             12.41 (11.20-13.61)                                                     37.33 (35.24-39.43)   51.59 (48.65-54.54)   25.30 (24.16-26.44)\n  Diabetes                 3.16 (2.55-3.77)                                                        15.49 (14.00-16.97)   20.20 (18.04-22.36)   9.27 (8.59-9.95)\n  Smoking                  57.26 (55.56-58.97)                                                     48.76 (46.54-50.98)   32.00 (29.04-34.96)   51.52 (50.21-52.82)\n  Stroke history           0.16 (0.04-0.28)                                                        2.21 (1.60-2.83)      7.69 (6.32-9.05)      1.71 (1.45-1.97)\n  Coronary heart disease   0.46 (0.23-0.69)                                                        2.85 (2.18-3.51)      6.35 (5.11-7.60)      1.94 (1.65-2.23)\n  Hypercholesterolemia     7.26 (6.36-8.16)                                                        13.59 (12.10-15.09)   9.75(7.86-11.63)      9.69 (8.93-10.44)\n  Obesity, BMI (kg/m^2^)                                                                                                                       \n  \u200325.0-29.9               31.83 (30.15-33.51)                                                     36.84 (34.97-38.70)   22.83 (20.47-25.18)   32.51 (31.36-33.65)\n  \u2003\u2265 30                    4.76 (3.99-5.53)                                                        2.77 (2.08-3.45)      1.23 (0.65-1.82)      3.69 (3.19-4.19)\n  \u2003\\< 18.5                 3.29 (2.63-3.95)                                                        1.97 (1.43-2.51)      6.55 (5.21-7.89)      3.21 (2.77-3.66)\n  \u200318.5 \u2264 , \\< 25          60.12 (58.37-61.87)                                                     58.43 (56.52-60.33)   69.39 (66.79-71.99)   60.60 (59.40-61.80)\n\nAge standardized prevalence using a direct standardization method.\n\nCI, confidence interval; BMI, body mass index.\n\n###### \n\nPrevalence of individual risk factors in women of the general population according to the age group\n\n                           Prevalence^[\\*](#tfn3-jos-17-3-302){ref-type=\"table-fn\"}^ %, (95% CI)                                               \n  ------------------------ ----------------------------------------------------------------------- --------------------- --------------------- ---------------------\n  Hypertension             3.62 (3.05-4.20)                                                        31.72 (29.98-33.47)   63.92 (61.50-66.34)   22.72 (21.72-23.72)\n  Diabetes                 2.38 (1.86-2.90)                                                        10.37 (9.25-11.49)    21.35 (19.06-23.63)   8.11 (7.48-8.73)\n  Smoking                  8.44 (7.39-9.50)                                                        4.98 (4.09-5.88)      6.19 (4.92-7.46)      6.93 (6.26-7.60)\n  Stroke history           0.07 (0.00-0.14)                                                        1.52 (1.13-1.90)      4.86 (3.92-5.81)      1.33 (1.13-1.53)\n  Coronary heart disease   0.20 (0.09-0.31)                                                        2.58 (2.04-3.13)      5.25 (4.17-6.34)      1.81 (1.54-2.08)\n  Hypercholesterolemia     3.59 (2.99-4.18)                                                        18.99 (17.53-20.45)   22.36 (20.07-24.66)   11.74 (11.04-12.44)\n  Obesity, BMI (kg/m^2^)                                                                                                                       \n  \u200325.0-29.9               15.44 (14.21-16.66)                                                     30.13 (28.53-31.72)   33.06 (30.79-35.33)   23.17 (22.21-24.13)\n  \u2003\u2265 30                    3.68 (3.06-4.58)                                                        4.53 (3.73-5.32)      4.70 (3.61-5.78)      4.13 (3.68-4.58)\n  \u2003\\< 18.5                 9.66 (8.60-10.71)                                                       2.10 (1.47-2.73)      3.56 (2.76-4.36)      6.16 (5.54-6.78)\n  \u200318.5 \u2264 , \\< 25          71.23 (69.75-72.70)                                                     63.24 (61.50-64.99)   58.68 (56.39-60.97)   66.54 (65.48-67.61)\n\nAge standardized prevalence using a direct standardization method.\n\nCI, confidence interval; BMI, body mass index.\n\n###### \n\nFrequency of individual risk factors in cases and controls\n\n                           Men, n (%)     Women, n (%)                                                    \n  ------------------------ -------------- -------------- ------------------ -------------- -------------- ------------------\n  Hypertension             1,811 (63.6)   1,349 (47.4)   1.76 (1.55-2.00)   1,329 (70.1)   1,067 (56.2)   1.64 (1.39-1.95)\n  Diabetes                 954 (33.5)     526 (18.5)     1.89 (1.63-2.19)   626 (33.0)     367 (19.3)     1.83 (1.53-2.19)\n  Hypercholesterolemia     675 (23.7)     422 (14.8)     1.38 (1.18-1.62)   541 (28.5)     455 (24.0)     1.07 (0.89-1.27)\n  Smoking                  1,286 (45.2)   911 (32.0)     2.03 (1.78-2.31)   111 (5.9)      108 (5.7)      1.26 (0.92-1.74)\n  Stroke history           641 (22.5)     140 (4.9)      4.96 (3.97-6.17)   406 (21.4)     75 (4.0)       6.21 (4.56-8.45)\n  Coronary heart disease   291 (10.2)     146 (5.1)      1.76 (1.38-2.25)   191 (10.1)     70 (3.7)       2.65 (1.89-3.72)\n  BMI (kg/m^2^)                                                                                           \n  \u2003\\< 18\u00b75                 110 (3.9)      120 (4.2)      0.87 (0.63-1.19)   120 (6.3)      66 (3.5)       1.87 (1.29-2.70)\n  \u200318.5-24.9               1,874 (65.8)   1,844 (64.8)   1                  1,188 (62.6)   1,142 (60.2)   1\n  \u200325.0-29.9               792 (27.8)     821 (28.8)     0.85 (0.74-0.98)   504 (26.6)     607 (32.0)     0.68 (0.57-0.80)\n  \u2003\u2265 30.0                  70 (2.5)       61 (2.1)       0.77 (0.51-1.17)   85 (4.5)       82 (4.3)       0.80 (0.55-1.15)\n\nOR was adjusted for age and other risk factors of interest.\n\nOR, odds ratio; CI, confidence interval; BMI, body mass index.\n\n###### \n\nOdds ratio and population attributable risk in age- and sex-specific subgroups\n\n  Risk factor              Age groups (N/n)        Men                 Women                                                                    \n  ------------------------ ----------------------- ------------------- ----------------------- ------------------------------------------------ ------------------------\n  Hypertension             \u2264 45 yr (410/132)       4.21 (2.26-7.83)    28.45 (13.51-45.86)     2.57 (0.78-8.53)                                 5.39 (-0.82-21.45)\n                           46-65 yr (2,430/946)    1.79 (1.47-2.17)    22.66 (14.93-30.38)     1.93 (1.41-2.62)                                 22.69 (11.61-33.95)\n                           \u2265 66 yr (2,852/2,716)   1.60 (1.33-1.92)    23.67 (14.70-32.21)     1.48 (1.20-1.82)                                 23.40 (11.44-34.31)\n  Diabetes                 \u2264 45 yr (410/132)       2.78 (1.14-6.80)    5.32 (0.43-15.49)       2.14 (0.38-11.98)                                2.65 (-1.49-20.72)\n                           46-65 yr (2,430/946)    2.10 (1.68-2.63)    14.56 (9.52-20.12)      1.81 (1.25-2.63)                                 7.76 (2.49-14.47)\n                           \u2265 66 yr (2,852/2,716)   1.71 (1.39-2.10)    12.48 (7.28-18.11)      1.83 (1.49-2.25)                                 15.07 (9.43-21.10)\n  Smoking                  \u2264 45 yr (410/132)       2.43 (1.46-4.07)    45.05 (20.74-63.72)     1.74 (0.36-8.48)                                 5.89 (-5.74-38.72)\n                           46-65 yr (2,430/946)    2.22 (1.83-2.69)    37.36 (28.91-45.24)     2.66 (1.39-5.12)                                 7.65 (1.88-17.04)\n                           \u2265 66 yr (2,852/2,716)   1.77 (1.45-2.15)    16.73 (12.64-26.95)     0.84 (0.57-1.26)                                 -0.98 (-2.75-1.55)\n  Hypercholesterolemia     \u2264 45 yr (410/132)       1.65 (0.84-3.22)    4.49 (-1.17-13.90)      1.23 (0.28-5.52)                                 0.83 (-2.66-13.94)\n                           46-65 yr (2,430/946)    1.19 (0.95-1.50)    2.54 (-0.70-6.32)       0.85 (0.61-1.19)                                 -2.89 (-7.95-3.41)\n                           \u2265 66 yr (2,852/2,716)   1.63 (1.27-2.08)    5.76 (2.58-9.54)        1.18 (0.95-1.46)                                 3.81 (-1.15-9.29)\n  Stroke history           \u2264 45 yr (410/132)       4.22 (0.68-26.08)   0.51 (-0.05-3.88)       4.37 (0.12-164.75)                               0.22 (-0.06-9.65)\n                           46-65 yr (2,430/946)    6.91 (4.54-10.52)   11.58 (7.27-17.41)      8.83 (4.18-18.63)                                10.61 (4.60-21.11)\n                           \u2265 66 yr (2,852/2,716)   4.18 (3.21-5.46)    19.66 (14.51-25.51)     5.30 (3.79-7.41)                                 17.30 (11.96-23.76)\n  Coronary heart disease   \u2264 45 yr (410/132)       1.66 (0.18-15.19)   0.30 (-0.38-6.13)       ^[a](#tfn6-jos-17-3-302){ref-type=\"table-fn\"}^   \n                           46-65 yr (2,430/946)    1.84 (1.21-2.80)    2.34 (0.59-4.87)        1.95 (0.97-3.93)                                 2.40 (-0.08-7.04)\n                           \u2265 66 yr (2,852/2,716)   1.67 (1.23-2.28)    4.10 (1.44-7.50)        2.99 (2.02-4.43)                                 9.45 (5.06-15.26)\n  BMI (kg/m^2^)                                                                                                                                 \n  \u2003\\< 18.5                 \u2264 45 yr (13/8)          0.81 (0.20-3.34)    -0.62 (-2.71-7.16)      1.47 (0.33-6.50)                                 4.33 (-6.89-34.68)\n                           46-65 yr (58/23)        1.32 (0.68-2.53)    0.62 (-0.63-2.93)       0.95 (0.39-2.33)                                 1.87 (-0.70-8.50)\n                           \u2265 66 yr (159/155)       0.74 (0.50-1.08)    -1.76 (-3.38-0.52)      1.26 (0.18-8.71)                                 3.60 (1.23-6.99)\n  \u200325.0-29.9               \u2264 45 yr (152/29)        1.48 (0.87-2.49)    -13.16 (-4.18-32.20)    1.91 (0.67-5.42)                                 -0.82 (-10.49-17.04)\n                           46-65 yr (826/311)      0.79 (0.65-0.97)    -8.38 (-14.95\\--1.27)   0.80 (0.58-1.10)                                 -6.45 (-14.41-2.90)\n                           \u2265 66 yr (635/771)       0.84 (0.68-1.04)    -3.84 (-8.01-0.95)      0.88 (0.40-1.91)                                 -14.41 (-19.66\\--8.54)\n  \u2003\u2265 30.0                  \u2264 45 yr (27/9)          1.31 (0.47-3.70)    1.47 (-2.61-11.40)      2.05 (1.35-3.11)                                 0.96 (-3.10-22.12)\n                           46-65 yr (57/36)        0.67 (0.36-1.25)    -0.92 (-1.81-0.70)      0.62 (0.50-0.76)                                 -0.57 (-2.79-3.96)\n                           \u2265 66 yr (47/122)        0.75 (0.37-1.54)    -0.31 (-0.79-0.66)      0.73 (0.47-1.13)                                 -1.29 (-2.55-0.60)\n\nOR was adjusted for age and other risk factors of interest;\n\nPAR was calculated using the following formula: PAR=Pg (OR\u22121)/\\[1+Pg (OR\u22121)\\], where Pg represents risk factor prevalence in the general population, estimated from the prevalence in the portion of the population over 19 years old in the 4th KNHANES, with standardization to the age structure of the general population from the 2010 Census;\n\nOR was not available due to the rarity of coronary heart disease in this age group.\n\nN represents the number of men in each age group, n represents the number of women; OR, odds ratio; PAR, population attributable risk; CI, confidence interval; BMI, body mass index.\n"}
+{"text": "1. Introduction {#sec1-ijms-17-00122}\n===============\n\nA thorough understanding of the molecular machineries involved in the underlying pathophysiology and pharmacodynamics are crucial to the treatment of metabolic disorders including fatty liver disease. We recently elucidated a novel lipophagic mechanism by which metformin, a kind of anti-diabetic drug that disturbs mitochondrial complex I of the electron transport chain, alleviates non-alcoholic fatty liver disease (NAFLD) by inducing the sirtuin class of histone/protein deacetylases sirtuin 1 (SIRT1)-mediated autophagy independent of 5\u2032 adenosine monophosphate-activated protein kinase (AMP-activated protein kinase) \\[[@B1-ijms-17-00122]\\]. Of the intracellular organelles involved in this underlying pathophysiologic mechanism, the degradation of dysfunctional mitochondria and the reciprocal biogenesis of mitochondria, *i.e.*, mitophagy, are areas of debate and investigation \\[[@B2-ijms-17-00122],[@B3-ijms-17-00122]\\]. Although controversies remain as to the distinct machinery utilized for mitophagy in response to cell stressors, dysfunctional mitophagy in response to glucose and lipid toxicities in hepatocytes might play an important role in the accumulation of intracellular or extracellular oxidative stress, which causes NAFLD, ultimately leading to hepatocellular carcinoma \\[[@B3-ijms-17-00122],[@B4-ijms-17-00122],[@B5-ijms-17-00122]\\]. In addition, recent studies have reported that cytosolic p53 interferes with mitochondrial integrity and mitophagy \\[[@B6-ijms-17-00122],[@B7-ijms-17-00122]\\] by inhibiting the removal of damaged mitochondria through an inhibitory interaction with Parkin, thereby inducing mitochondrial dysfunction \\[[@B8-ijms-17-00122],[@B9-ijms-17-00122]\\]. Based on these previous reports, we hypothesized and aimed to elucidate the mechanism by which metformin induces mitophagy through restoration of the Parkin-mediated mitophagy suppressed by increased cytosolic ER stress and p53.\n\n2. Results {#sec2-ijms-17-00122}\n==========\n\n2.1. Metformin Treatment Upregulates Mitophagy {#sec2dot1-ijms-17-00122}\n----------------------------------------------\n\nThe electron microscopic findings resulting from *ad libitum* feeding of chow diets ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}A-a), caloric restriction (CR) ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}A-b), and *ad libitum* feeding of chow diets and 300 mg/kg metformin treatment ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}A-c) in mice are shown. Electron microscopy readily detected spheroid mitochondria forming a clearly defined internal space surrounded by mitochondrial membranes in mice fed a standard chow diet ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}A-a). The compressed mitochondria (white arrow), the lost cristae, and some portion of the matrix might connect the extra-mitochondrial space through a small orifice (black arrow) or form a lumen containing subcellular contents (asterisk). The white arrowheads indicate either autophagic double membranes surrounding abnormal mitochondria or an autophagic vacuole containing mitochondria ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}A-c).\n\n![Metformin induced mitophagy. Electron microscopy was performed on hepatocytes from *ob*/*ob* mice under *ad libitum* feeding of chow diet (**A**-**a**), caloric restriction (**A**-**b**), and *ad libitum* feeding of chow diet with 300 mg/kg metformin treatment (**A**-**c**). In the *ad libitum* group, spheroid mitochondria were readily detected. However, in the metformin-treated group, autophagic double membranes were detected, indicating mitophagy; (**B**) co-localization between a marker of autophagosomes (LC3) and a mitochondrial marker (MitoTracker Red) was assessed in HepG2 cells stably transfected with green fluorescent protein (GFP)-LC3. Although there was little co-localization between mitochondria and autophagosomes in the 0.25 mM palmitate-treated HepG2 cells, LC3-labeled structures were seen surrounding the fragmented mitochondria in cells treated with 0.5 mM metformin and 0.25 mM palmitate. Scale bar = 20 \u03bcm.](ijms-17-00122-g001){#ijms-17-00122-f001}\n\nWe assessed co-localization between a marker of autophagosomes, LC3, and a mitochondrial marker (MitoTracker Red) in HepG2 cells stably transfected with GFP-LC3. HepG2 cells ectopically expressing GFP-LC3 exhibited an increase in the number of punctate GFP-LC3 structures upon exposure to 0.5 mM metformin. Furthermore, LC3-labeled structures were found surrounding the fragmented mitochondria in cells treated with metformin and palmitate, although there was little co-localization between mitochondria and autophagosomes in cells treated with 0.25 mM palmitate alone ([Figure 1](#ijms-17-00122-f001){ref-type=\"fig\"}B).\n\n2.2. Palmitate Increases but Metformin Decreases the Expression of p53 Protein {#sec2dot2-ijms-17-00122}\n------------------------------------------------------------------------------\n\nTo determine whether metformin down-regulated p53 expression, Western blotting was performed using the cytosolic fraction. Western blot analysis showed a dose-dependent increase in cytosolic p53 protein in response to palmitate in the presence of 35 mM glucose in HepG2 cells ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}A-a). In contrast, when metformin was added to the same conditions, p53 expression decreased in a dose-dependent manner ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}A-b). Treatment of HepG2 cells with both 0.25 mM palmitate and 35 mM glucose significantly activated cytosolic p53 (1.0 \u00b1 0 *vs.* 1.226 \u00b1 0.028, *p* \\< 0.0001). However, pretreatment with 0.5 mM metformin significantly attenuated this gluco-lipid induced expression of p53 in HepG2 cells (1.226 \u00b1 0.028 *vs.* 1.078 \u00b1 0.038, *p* \\< 0.05, [Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}B-a). To confirm the effects of metformin in an animal model, we performed Western blotting using hepatocytes from *ob*/*ob* mice. As shown in [Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}B-b, expression of p53 was significantly attenuated in both the CR- (0.639 \u00b1 0.072, *p* \\< 0.001) and metformin-treated groups (0.783 \u00b1 0.114, *p* \\< 0.05) compared to the *ad libitum*-fed group.\n\n2.3. Metformin Induces Parkin-Mediated Mitophagy through Inhibition of Glucolipotoxicity-Induced Cytosolic p53 {#sec2dot3-ijms-17-00122}\n--------------------------------------------------------------------------------------------------------------\n\nTo highlight the impact of metformin on mitophagy, we assessed co-localization between Parkin and a mitochondrial marker (MitoTracker Red) using immunofluorescence staining, and performed cytosolic and mitochondrial fractionation experiments in HepG2 cells. In the immunofluorescence study, there was little co-localization between Parkin and mitochondria in cells treated with 0.25 mM palmitate and 35 mM glucose, whereas metformin restored the co-localization (yellow dots) between Parkin and mitochondria. In the fractionation experiments, Parkin translocation from the cytosol to mitochondria was attenuated in cells treated with 0.25 mM palmitate and 35 mM glucose, whereas metformin restored the Parkin translocation suppressed by glucolipotoxicity ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}C). We also found a negative relationship between Parkin and mitofusin translocation in response to metformin treatment. To examine p53 and Parkin protein--protein interactions, the endogenous Parkin--p53 complex was observed in immunoprecipitates (IP) of Parkin and p53 ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}D). Treatment of HepG2 cells with both 0.25 mM palmitate and 35 mM glucose significantly upregulated the expression of Parkin. However, pretreatment of HepG2 cells with 0.5 mM metformin significantly attenuated the gluco-lipid induced expression of Parkin ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}D-a); similar findings were also found using *ad libitum*-fed *ob*/*ob* mice and *ob*/*ob* mice treated with CR or metformin ([Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}D-b). These results indicate the involvement of p53 in the significant decrease in Parkin-dependent mitophagy observed under diabetic conditions.\n\n2.4. Metformin Induces Parkin-Mediated Mitophagy through Inhibition of Glucolipotoxicity-Induced Endoplasmic Reticulum (ER) Stress {#sec2dot4-ijms-17-00122}\n----------------------------------------------------------------------------------------------------------------------------------\n\nTo elucidate possible ER stress-mediated glucolipotoxicity in a hepatocyte cell line, we investigated the effects of palmitate with glucose and metformin on ER stress markers using Western blotting. We demonstrated that HepG2 cells treated with both 0.25 mM palmitate and 35 mM glucose for 24 h showed significant activation of the expression of ATF6 and Xbp-1. However, pretreatment with 0.5 mM metformin attenuated the expression of ER stress markers in HepG2 cells ([Figure 3](#ijms-17-00122-f003){ref-type=\"fig\"}A-a). Similar findings were noted in both the CR- and metformin-treated groups compared to the *ad libitum*-fed group ([Figure 3](#ijms-17-00122-f003){ref-type=\"fig\"}A-b). In addition, we investigated this pathway using the traditional ER stressor thapsigargin. As shown in [Figure 3](#ijms-17-00122-f003){ref-type=\"fig\"}B, HepG2 cells exposed to different concentrations (0, 0.1, 0.2, and 0.5 \u03bcM) of thapsigargin showed upregulated expression of ER stress markers and of p53 protein. Based on this finding, we utilized 0.1 \u03bcM thapsigargin in subsequent experiments. Compared to the untreated controls, HepG2 cells exposed to 0.1 \u03bcM thapsigargin for 24 h showed significantly decreased levels of viability (100.0 \u00b1 2.95 *vs.* 94.0 \u00b1 1.24, *p* \\< 0.05) although there is no significant difference in viability in HepG2 cells exposed to 0.5 mM metformin for 24 h, as assessed by CCK-8. Pretreatment with 0.5 mM metformin for 3 h significantly restored the HepG2 cell viability reduced by 0.1 \u03bcM thapsigargin (94.0 \u00b1 1.24 *vs.* 101.1 \u00b1 2.18, *p* \\< 0.001, [Figure 3](#ijms-17-00122-f003){ref-type=\"fig\"}C-a). To elucidate possible ER stress-mediated mitophagic dysfunction, we performed cytosolic and mitochondrial fractionation experiments in HepG2 cells. Parkin translocation from the cytosol to mitochondria was attenuated in cells treated with 0.1 \u03bcM thapsigargin, whereas metformin restored the Parkin translocation suppressed by thapsigargin ([Figure 3](#ijms-17-00122-f003){ref-type=\"fig\"}C-b,c).\n\n###### \n\nMetformin restored the Parkin-mediated mitophagy inhibited by glucolipotoxicity-induced cytosolic p53. Western blotting analysis was performed to assess cytosolic p53 levels with the cytosolic fractions isolated from palmitate- (**A**-**a**) or metformin- (**A**-**b**) and 35 mM glucose-treated HepG2 cells; the expression of cytosolic p53 protein was assessed following 0.5 mM metformin and/or 0.25 mM palmitate with 35 mM glucose treatment of HepG2 cells (**B**-**a**) or in hepatocytes from *ob*/*ob* mice with *ad libitum* feeding of chow diet, caloric restriction, and *ad libitum* feeding of chow diet with 300 mg/kg metformin treatment (**B**-**b**); (**C**) Expression of Parkin (green) on mitochondria (MitoTracker Red) was observed with a confocal microscopy. The bottom panels show enlarged views of the boxed areas. The yellow dots indicate the mitochondria that co-localize with Parkin. The cytosolic and mitochondrial fractionation experiments demonstrated Parkin translocation from the cytosol to mitochondria depending on the conditions used in the HepG2 cells; a co-immunoprecipitation assay was conducted to investigate the interaction between p53 and Parkin, depending on the conditions used in the HepG2 cells (**D**-**a**) or in hepatocytes from mice (**D**-**b**). GAPDH or COX IV was used for normalization. Values displayed are mean \u00b1 SEM (*n* = 3 independent experiments, respectively). Scale bar: white = 20 \u03bcm, yellow = 10 \u03bcm. Asterisks (\\* *p* \\< 0.05, \\*\\* *p* \\< 0.01, \\*\\*\\* *p* \\< 0.001) indicate significant differences. Abbreviation: Mfn1, mitofusin1.\n\n![](ijms-17-00122-g002a)\n\n![](ijms-17-00122-g002b)\n\n###### \n\nMetformin restored Parkin-mediated mitophagy inhibited by glucolipotoxicity-induced ER stress. The expression of ER stress markers was assessed following 0.5 mM metformin and/or 0.25 mM palmitate with 35 mM glucose treatment of HepG2 cells (**A**-**a**) or in hepatocytes from *ob*/*ob* mice with *ad libitum* feeding of chow diet, caloric restriction, or *ad libitum* feeding of chow diet with 300 mg/kg metformin treatment (**A**-**b**) using Western blotting analysis. Values displayed are mean \u00b1 SEM (*n* = 3 independent experiments, respectively); (**B**) the expression of ER stress markers and p53 protein was assessed in response to thapsigargin in HepG2 cells; (**C**-**a**) cell viability was assessed in HepG2 cells exposed to 0.1 \u03bcM thapsigargin and/or 0.5 mM metformin using CCK-8. Values displayed are mean \u00b1 SEM (*n* = 5 independent experiments); (**C**-**b**) the expression of an ER stress marker was also examined under these conditions; (**C**-**c**) the cytosolic and mitochondrial fractionation experiments demonstrated Parkin translocation from the cytosol to mitochondria, depending on the conditions used. Values displayed are mean \u00b1 SEM (*n* = 3 independent experiments, respectively). GAPDH or COX IV was used for normalization. Asterisks (\\* *p* \\< 0.05, \\*\\* *p* \\< 0.01, \\*\\*\\* *p* \\< 0.001) indicate significant differences. Abbreviation: Mfn1, mitofusin1; CCK-8, cell counting kit-8.\n\n![](ijms-17-00122-g003a)\n\n![](ijms-17-00122-g003b)\n\n2.5. A p53 Inhibitor Restores Parkin-Mediated Mitophagy Suppressed by Glucolipotoxicity {#sec2dot5-ijms-17-00122}\n---------------------------------------------------------------------------------------\n\nTo determine whether p53 actively mediates mitophagy dysfunction, we treated cells with pifithrin-\u03b1 (PFT-\u03b1), a p53 inhibitor, to act as an ER stress reliever. Parkin translocation from the cytosol to mitochondria was attenuated in cells treated with 0.25 mM palmitate and 35 mM glucose, whereas PFT-\u03b1 restored the Parkin translocation suppressed by glucolipotoxicity ([Figure 4](#ijms-17-00122-f004){ref-type=\"fig\"}A). We additionally assessed co-localization between a marker of autophagosomes, LC3, and mitochondria in HepG2 cells stably transfected with GFP-LC3. LC3-labeled structures were observed surrounding the fragmented mitochondria in cells treated with PFT-\u03b1 and palmitate, although there was little co-localization between mitochondria and autophagosomes in cells treated with 0.25 mM palmitate alone. ([Figure 4](#ijms-17-00122-f004){ref-type=\"fig\"}B).\n\n![A p53 inhibitor restored the Parkin-mediated mitophagy inhibited by glucolipotoxicity. (**A**-**a**) the expression of cytosolic p53 was assessed in HepG2 cells treated with 50 \u03bcM PFT-\u03b1, a p53 inhibitor, and/or 0.25 mM palmitate with 35 mM glucose using Western blotting analysis; (**A**-**b**) the cytosolic and mitochondrial fractionation experiments demonstrated Parkin translocation from the cytosol to mitochondria, depending on the conditions used. Exposure time of Western blotting was long, compared to [Figure 2](#ijms-17-00122-f002){ref-type=\"fig\"}C; (**B**) co-localization was assessed between a marker of autophagosomes (LC3) and a mitochondrial marker (MitoTracker Red) in HepG2 cells stably transfected with GFP-LC3, depending on the conditions used. LC3-labeled structures were observed surrounding the fragmented mitochondria in cells treated with PFT-\u03b1 and palmitate, although there was little co-localization between mitochondria and autophagosomes in cells treated with 0.25 mM palmitate alone. Scale bar = 20 \u03bcm. *Abbreviation*: Mfn1, mitofusin1; PFT-\u03b1, pifithrin-\u03b1.](ijms-17-00122-g004){#ijms-17-00122-f004}\n\n3. Discussion {#sec3-ijms-17-00122}\n=============\n\nMitochondria not only produce cellular energy through the oxidation of nutrients, but also generate reactive oxygen species (ROS) under pathologic conditions that induce mitochondrial dysfunction \\[[@B10-ijms-17-00122]\\]. The generation of ROS in conjunction with endoplasmic reticulum (ER) stress induces cellular stress and, consequently, cell death \\[[@B11-ijms-17-00122],[@B12-ijms-17-00122]\\]. Because the existence of damaged and aged mitochondria threatens cells, the turnover processes of mitochondrial degradation and biogenesis are essential for maintenance of cellular integrity \\[[@B2-ijms-17-00122],[@B3-ijms-17-00122],[@B4-ijms-17-00122],[@B13-ijms-17-00122]\\]. With advances in the understanding of autophagy, mitophagy, the selective degradation of mitochondria by autophagy, has garnered attention in mitochondrial metabolism. Conventionally, metformin has been viewed as a type of anti-diabetic drug belonging to the biguanide class that functions by disturbing complex I of the electron transport chain, thereby resulting in activation of 5\u2032 adenosine monophosphate-activated protein kinase (AMP-activated protein kinase) and induction of the activation of NAD+-dependent deacetylase SIRT1 (sirtuin 1) \\[[@B14-ijms-17-00122],[@B15-ijms-17-00122]\\]. Recently, we demonstrated that metformin activates lipophagy, resulting in relief of hepatosteatosis due to induction of SIRT1-mediated autophagy independent of AMP-activated protein kinase \\[[@B1-ijms-17-00122]\\]. Based on these reports, we hypothesized that metformin plays additional roles in mitochondrial homeostasis through the activation of mitophagy, resulting in an improvement in hepatosteatosis.\n\nAmong the cell stress or apoptotic pathways, p53 is known to utilize many mechanisms in its anticancer function and to play roles in apoptosis and genomic stability. Recent studies have reported that cytosolic p53 contributes to mitochondrial integrity and mitophagy \\[[@B6-ijms-17-00122],[@B7-ijms-17-00122]\\]. Cytosolic p53 has been shown to inhibit the removal of damaged mitochondria through an inhibitory interaction with Parkin, thereby inducing mitochondrial dysfunction \\[[@B8-ijms-17-00122],[@B9-ijms-17-00122]\\]. Based on these reports, we studied the interplay of metformin in mitophagy and p53 abundance. We therefore investigated (1) whether metformin alleviates the glucolipotoxicity induced cell stress by assessing the expression of ER stress markers and p53 protein; (2) whether Parkin-mediated mitophagy is inhibited by glucolipotoxicity-induced cytosolic p53; and (3) whether metformin induces mitophagy through restoration of the suppressed Parkin-mediated mitophagy induced by increased cytosolic ER stress and p53.\n\nWith respect to glucolipotoxicity-induced hepatocellular toxicity and the role of metformin in alleviating glucolipotoxicity, we observed a dose-dependent increase in p53 protein expression in HepG2 cells treated with palmitate in the presence of 35 mM glucose, as well as a dose-dependent decrease in p53 expression in HepG2 cells treated with metformin. Additionally, we found that metformin alleviated glucolipotoxicity-induced p53 overexpression. Similar to these findings, metformin alleviated glucolipotoxicity-induced ER stress, as assessed by Western blotting using ATF6 and Xbp-1 antibodies. To further demonstrate ER stress pathway-mediated glucolipotoxicity, we treated HepG2 cells with the well-established ER stressor thapsigargin (TG). In these experiments, we observed a dose-dependent upregulation of cytosolic p53 and the ER stress markers ATF6 and Xbp-1 in TG-treated HepG2 cells. From a translational point of view \\[[@B16-ijms-17-00122]\\], nutritional deprivation through caloric restriction (CR) and metformin administration in *ob*/*ob* mice also showed significantly decreased levels of cytosolic p53 and the ER stress markers ATF6 and Xbp-1 compared to *ad libitum* fed *ob*/*ob* mice. Endoplasmic reticulum (ER) plays a role as a nutrient sensor in cells, and excess fuel can induce ER stress, which triggers the unfolded protein response (UPR). The UPR leads to both translational attenuation of new protein synthesis and transcriptional activation of stress-response genes to relieve ER stress. However, the UPR also initiates death signals, which function when the stress is pathologically prolonged \\[[@B17-ijms-17-00122]\\]. ER stress has been generally shown to be inhibited by metformin \\[[@B18-ijms-17-00122],[@B19-ijms-17-00122]\\], thereby preventing cell injury and apoptosis. Based on results in this study, we suggest that increased expression of p53 accompanies the ER stress pathway.\n\nWith respect to mitophagy in hepatic cells, *ob*/*ob* mice treated with metformin or CR showed increased mitophagy, as demonstrated by autophagic double membranes surrounding abnormal mitochondria or by the presence of an autophagic vacuole containing mitochondria. In contrast, *ob*/*ob* mice under metabolic stress caused by *ad libitum* feeding of chow diets showed numerous mitochondrial spheroids, as demonstrated by the compressed mitochondrion and lost cristae and matrix. To further validate mitophagy induction by metformin treatment, we additionally demonstrated increased co-localization of the autophagy marker GFP-LC3 and the selective mitochondrial probe MitoTracker Red in metformin-treated HepG2 cells. Mitochondrial spheroids are described as having a unique ring-like morphology with a squeezed mitochondrial matrix. Under various pathophysiologic stresses, mitochondria can undergo direct remodeling to form mitochondrial spheroids, which requires the presence of ROS and either mitofusin1 (Mfn1) or mitofusin2 (Mfn2). In this process, Parkin prevents mitochondrial spheroid formation by causing proteasomal degradation of Mfn1 and Mfn2, which are required for mitophagy \\[[@B20-ijms-17-00122]\\]. Mitochondrial spheroids can envelop the contents of the cytosol, including damaged mitochondria. Regarding fat accumulation in the liver, each hepatocyte contains about 800 mitochondria, and mitochondrial dysfunction can contribute to the development of NAFLD \\[[@B21-ijms-17-00122],[@B22-ijms-17-00122],[@B23-ijms-17-00122]\\]. In a recent study, Ibdah *et al.*, using heterozygous mice deficient in mitochondrial trifunctional protein, found that impairment in \u03b2-oxidation resulted in hepatic steatosis \\[[@B24-ijms-17-00122]\\]. These alterations were also demonstrated in Otsuka Long-Evans Tokushima Fatty (OLETF) rats, which is a model of type 2 diabetes and obesity \\[[@B25-ijms-17-00122]\\]. Based on the results of our study, we suggest that the metformin-induced autophagic machinery might play important roles in alleviating hepatosteatosis through the induction of both lipophagy and mitophagy.\n\nWith respect to the inverse relationship between metformin-induced mitophagy induction and p53 protein abundance \\[[@B26-ijms-17-00122]\\], the present study showed that glucolipotoxicity inhibited the translocation of Parkin from the cytosol to mitochondria. In contrast, we also showed that metformin or PFT-\u03b1, which is the p53 inhibitor, induced translocation of Parkin from the cytosol to mitochondria by decreasing the inhibitory interaction with cytosolic p53. These results were additionally validated by examining protein--protein interactions between p53 and Parkin using immunoprecipitates (IP) of endogenous Parkin and p53, as well as the co-localization of autophagy marker GFP-LC3 and the selective mitochondrial probe Mitotracker Red in PFT-\u03b1-treated HepG2 cells. Regarding the role of mitofusin in mitochondrial spheroid formation, ROS and either Mfn1 or Mfn2 are required. The inner membrane protein optic atrophy 1 (OPA1) might not be important for this process because it is rapidly degraded upon mitochondrial depolarization. Ding *et al.* reported that Parkin promotes mitophagy through the proteasomal degradation of mitofusins and, consequently, the inhibition of mitochondrial spheroid formation \\[[@B20-ijms-17-00122]\\]. We confirmed the negative relationship between Parkin and mitofusins in mitochondrial quality control in response to metformin treatment in this study. Based on these results, we suggest that metformin treatment facilitates Parkin-mediated mitophagy instead of mitochondrial spheroid formation by decreasing the inhibitory interaction with cytosolic p53 and increasing the degradation of mitofusins.\n\nThe limitation of this study is that we did not conduct *in vivo* experiments with genetically engineered animal models, such as Parkin knockout mice, to provide a more convincing mechanistic role of metformin in mitophagy. Consequently, these studies should be performed in the future. However, to the best of our knowledge, our study is the first to investigate the molecular interplay among glucolipotoxicity-induced ER stress, cytosolic p53, and Parkin-mediated mitophagy in response to metformin treatment. To summarize our findings, metformin treatment might facilitate Parkin-mediated mitophagy instead of mitochondrial spheroid formation by decreasing the inhibitory interaction with cytosolic p53 and increasing the degradation of mitofusins. This working thesis is summarized in [Figure 5](#ijms-17-00122-f005){ref-type=\"fig\"}.\n\n![Summary of the working thesis of this study. See the text for details. Black and red arrows indicate the effects of glucolipotoxicity and the drugs (metformin or PFT\u03b1), respectively.](ijms-17-00122-g005){#ijms-17-00122-f005}\n\n4. Experimental Section {#sec4-ijms-17-00122}\n=======================\n\n4.1. Animal and Experimental Procedures {#sec4dot1-ijms-17-00122}\n---------------------------------------\n\nEight-week-old *ob*/*ob* mice (C57bl/6j background, Jackson Laboratory, Bar Harbor, ME, USA) were maintained at 60% \u00b1 5% relative humidity and 22 \u00b1 2 \u00b0C with a 12-h light/dark cycle over a four-week course of the following treatment regimens: (1) *ad libitum* feeding of chow diet (Dyets Inc., Bethlehem, PA, USA) and vehicle injection of PBS (pH 7.8, *n* = 8); (2) *ad libitum* feeding of chow diet and intraperitoneal injections of 300 mg/kg metformin (Merck KGaA, Darmstadt, Germany, *n* = 8); and (3) caloric restriction (CR) group (3 g/day, *n* = 8). After each animal was sacrificed, fresh liver tissues were fixed with 2% glutaraldehyde-2% paraformaldehyde buffered with 0.1 M phosphate buffer (pH 7.2) overnight at 4 \u00b0C, post-fixed with 1% osmium tetroxide in a 0.1 M sodium cacodylate buffer (pH 7.4) for 1 h at room temperature, and dehydrated with a graded series of ethanol, as previously described \\[[@B5-ijms-17-00122]\\]. The dehydrated tissues were then embedded in Epon, using poly/Bed812 Embedding kit/ DMP-30 (Polysciences, Warrington, PA, USA). All experimental procedures performed in this study followed the ethical guidelines for animal studies and were approved by the Institutional Animal Care and Use Committee of Yonsei University College of Medicine (IACUC No. 2011-0302-1).\n\n4.2. Cell Culture and Preparation of Subcellular Fractions {#sec4dot2-ijms-17-00122}\n----------------------------------------------------------\n\nThe human hepatoma cell line HepG2 was cultured in Dulbecco's modified Eagle's medium (Welgene, Daegu, Korea) supplemented with 10% FBS and antibiotics. HepG2 cells were washed and incubated for 10 min in ice-cold PBS, scraped into PBS with 1 mM phenylmethylsulfonylfluoride (PMSF), and pelleted by centrifugation (500\u00d7 *g*) at 4 \u00b0C for 10 min. Cell pellets were resuspended in RIPA buffer (Life Science, Berkeley, CA, USA) with protease inhibitors (Thermo Scientific, Waltham, MA, USA), passed 10 times through a 30-gauge syringe to break open the cells, and incubated on ice for 10 min to generate a whole cell lysate. After centrifugation (10,000\u00d7 *g*, in a microcentrifuge) at 4 \u00b0C for 30 min, the supernatants were collected as whole extracts. Fractionation to separate the mitochondrial and cytosolic fractions was carried out as previously described \\[[@B27-ijms-17-00122]\\]. Briefly, HepG2 cells were harvested, washed in PBS, and resuspended in cytosolic extraction buffer (250 mM sucrose, 70 mM KCl, 137 mM NaCl, 4.3 mM Na~2~HPO~4~, 1.4 mM KH~2~PO~4~, pH 7.2, 200 \u03bcg/mL digitonin, 100 \u03bcM PMSF, protease inhibitor cocktail) (Sigma, St. Louis, MO, USA) for 5 min on ice. Digitonin-permeabilized cells were confirmed by staining with 0.2% trypan blue solution. Cells were centrifuged at 1000\u00d7 *g* for 5 min; the supernatant was saved as the cytosolic fraction, and the pellet was solubilized in two volumes of mitochondrial lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 0.2% (*v*/*v*) Triton X-100, 0.3% NP-40, PMSF, protease inhibitor cocktail) (Sigma) for 5 min on ice. Cells were then centrifuged at 10,000\u00d7 *g* for 10 min at 4 \u00b0C, and the supernatant was collected as the mitochondrial extract.\n\n4.3. Detection of Mitophagy {#sec4dot3-ijms-17-00122}\n---------------------------\n\nFor the detection of autophagosomes and mitochondrial spheroids using electron microscopy, sectioning for electron microscopic examination was accomplished with an ultramicrotome (Leica, EM UC7, Wetzlar, Germany), and electron microscopy was performed with a JEM-1011 transmission electron microscope (JEOL). Detection of mitophagy were also performed by transfecting HepG2 cells with a plasmid driving the expression of GFP-LC3 (excitation wavelength 488 nm, emission filter 530 nm) (OriGene Technologies, Inc., Rockville, MD, USA) for 16 h, followed by loading of cells with Mitotracker Red (500 nM) (excitation wavelength 488 nm, emission filter 585 nm) (Invitrogen, Carlsbad, CA, USA). Transfected cells were fixed with 4% paraformaldehyde and then observed with a confocal microscope.\n\n4.4. Immunofluorescence Staining {#sec4dot4-ijms-17-00122}\n--------------------------------\n\nTo detect Parkin expression on mitochondria, immunofluorescence staining was performed on cells using Parkin antibody and MitoTracker Red. The cells were stained with MitoTracker Red (500 nM) and fixed with 4% paraformaldehyde. The fixed cells were incubated with Parkin antibody for 2 h, followed by incubation with secondary antibody (Alexa 488, excitation wavelength 495 nm, emission filter 519 nm) (Invitrogen) for 2 h. Next, the cells were observed under the confocal microscope.\n\n4.5. Free Fatty Acid (FFA) Preparation {#sec4dot5-ijms-17-00122}\n--------------------------------------\n\nFFA solutions were prepared as previously described \\[[@B15-ijms-17-00122]\\]. Briefly, 100 mM palmitate (PA) (Sigma) stocks were prepared in 0.1 M NaOH at 70 \u00b0C and filtered. A one percent (weight/volume) palmitate-free BSA (Sigma) solution was prepared in serum-free DMEM. After the palmitate dissolved, the palmitate solutions were added to serum-free DMEM containing BSA. The 5 mM palmitate/1% BSA solution was prepared by complexing the appropriate amounts of palmitate to 1% BSA in a 55 \u00b0C water bath.\n\n4.6. Cell Viability Assay {#sec4dot6-ijms-17-00122}\n-------------------------\n\nHepG2 cells were dispensed in wells of 24-well plates at a density of 5 \u00d7 10^4^ cells/well. HepG2 cells were pretreated with 0.5 mM metformin for 3 h and incubated in the presence or absence of thapsigargin for 24 h. The cells were then treated with 10 \u03bcL of cell counting kit-8 (CCK-8) solution (Sigma-Aldrich, St. Louis, MO, USA) at 37 \u00b0C for the indicated times, according to the manufacturer's instructions. Absorbance was measured at 450 nm using a microplate reader (Molecular Devices, Sunnyvale, CA, USA).\n\n4.7. Western Blotting and Antibodies {#sec4dot7-ijms-17-00122}\n------------------------------------\n\nMouse livers and HepG2 cells were lysed in PRO-PREPTM protein extraction solution (iNtRON Biotechnology, Kyungki-Do, Korea), and the protein contents of the resulting lysates were measured using the Bradford assay (Bio-Rad, Berkeley, CA, USA). Equal amounts of protein were resolved using SDS-PAGE and were electroblotted onto a nitrocellulose membrane (Bio-Rad). Membranes were subsequently probed as indicated with the following primary antibodies: p53 (Cell Signaling, Danvers, MA, USA), ATF6, sXBP1, GAPDH, Parkin, mitofusin (Mfn1) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and COXVI (Life Science). Secondary antibodies (anti-rabbit and anti-mouse) were from Santa Cruz Biotechnology.\n\n4.8. Co-Immunoprecipitation Assay {#sec4dot8-ijms-17-00122}\n---------------------------------\n\nFor co-immunoprecipitation (IP) assays, HepG2 cells and liver tissue were harvested in lysis buffer (50 mM Tris (pH 7.4), 140 mM NaCl, 1% Triton X-100, 30 M MG132, and protease inhibitors). Cell extracts were centrifuged, and the supernatants were incubated with anti-p53 or anti-Parkin antibodies for 12 h at 4 \u00b0C. ProteinA/G PLUS-Agarose (Santa Cruz Biotechnology) was then added to each sample, and incubation was carried out overnight at 4 \u00b0C on a rotating device. Immunoprecipitates were collected by centrifugation at 1000\u00d7 *g* for 5 min at 4 \u00b0C and washed with lysis buffer containing 500 mM NaCl or PBS. The pellets were eluted by heating at 95 \u00b0C for 5 min in 1\u00d7 electrophoresis sample buffer.\n\n4.9. Statistical Analysis {#sec4dot9-ijms-17-00122}\n-------------------------\n\nStatistical analysis was performed using PRISM (GraphPad Software Inc., San Diego, CA, USA). Results are expressed as mean \u00b1 SE, and statistical significance was calculated using Student's *t*-test. Alternatively, for comparisons involving more than two groups, one-way analysis of variance (ANOVA) with a post hoc Bonferroni multiple comparison test was used to assess the differences. Statistical significance was defined as the conventional *p*-value of \\<0.05.\n\n5. Conclusions {#sec5-ijms-17-00122}\n==============\n\nIn conclusion, the present study suggests that metabolic stresses in *ob*/*ob* mice induce mitochondrial spheroid formation and ultimately result in fat droplet accumulation in hepatocytes. In contrast, metformin treatment of *ob*/*ob* mice induces Parkin-mediated mitophagy upregulation by inhibiting glucolipotoxicity-induced cytosolic p53.\n\nThis research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2014R1A1A2057906).\n\nYoung Mi Song and Woo Kyung Lee performed experiments, analyzed the data, and wrote and edited the manuscript; Byung-Wan Lee was responsible for experimental design and initiated, planned and supervised the study; Yong-ho Lee, Eun Seok Kang and Bong-Soo Cha have given much advice and technological support. All authors approved the final version of the article, including the authorship list.\n\nThe authors declare no conflict of interest.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Correction to: *Scientific Data* 10.1038/s41597-019-0229-9, published online 22 October 2019\n\nIn the original version of this Data Descriptor, the URL provided to access the accompanying machine-accessible metadata file was incorrect. This has now been corrected.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nLack of drug sensitivity of cancer cells and their ability to acquire resistance to anti-cancer drugs is one of the most challenging problems in modern oncology. Theoretically, we know the basis of this process (at least one of the mechanisms), which is mainly associated with a cross type of drug resistance that is driven by ABC family (ATP-binding cassette) members, e.g., PgP, BCRP, LRP, MRP, and ca 40 others \\[[@CR1]\\]. It was also reported that an increased DNA repair in cancer cells (mismatch repair pathway activation after the damage provoked by a drug) is another defense mechanism in those cells \\[[@CR2]\\].\n\nThe mechanism of drug resistance differs for various cancer types and depends on the category of a drug used in therapy. However, most anti-cancer agents contribute to the generation of reactive oxygen species (ROS) followed by target cells apoptosis. Nevertheless, continuous treatment with the same drug may result in less efficient ROS production that may lead to drug resistance. Importantly, even drugs that directly bind DNA and induce its damage usually require an increased ROS signaling that may be necessary for excessive DNA damage and apoptosis.\n\nLatest reports reveal that drug resistance may also result from an increased translocation of telomerase hTERT (telomerase reverse transcriptase) subunit to mitochondria or translocation of another telomerase-associated factor (telomerase-associated protein 1, TEP1) to vaults. As reported, mitochondrial enrichment with hTERT may be accompanied by ROS level modulation or an increased number of copies of mtDNA that may be followed by apoptosis repression and may play a protective role during therapy \\[[@CR3], [@CR4]\\]. We also know that one of the characteristic features of cancer cells is an increased telomerase expression or activity (circa 95% of cancers, including cancer stem cells) \\[[@CR5], [@CR6]\\].\n\nNumerous studies indicate a relation between telomerase and sensitivity of cancer cells to therapy. This observation is confirmed by the fact that an increased expression of telomerase in cancer cells correlates with their resistance to drugs \\[[@CR7]\\]. Simultaneously, inhibition of telomerase activity results in an increased sensitivity of cancer cells (e.g., breast cancer) to doxorubicin (DOX) \\[[@CR8]\\]. Thus, the correlation between drug resistance and telomerase seems to be evident especially since it is observed during the studies of sensitization process (via telomerase expression downregulation) in cancer cells exposed to radiomimetic compounds (i.e., DNA damaging) \\[[@CR9]\\]. This aspect is highly interesting, because telomerase is responsible for genome stabilization. Apparently, after therapy accompanied by telomerase blocking, we might observe a synergistic process based on two mechanisms---DNA damage (drugs) and mismatch repair/genome stabilization attenuation (telomerase inhibition). Consequently, it might lead to an increased efficacy in cancer cell elimination. Surprisingly, after chemotherapy, the activity of telomerase was reported to increase in LoVo (colon cancer) cells that were accompanied by an increase in telomere length and induction of telomerase subunit expression hTERT and hTR (human telomerase RNA or TERC) \\[[@CR10]\\]. This might implicate a protective role of telomerase in cancer cells exposed to drugs. The efficacy of telomerase and telomeres targeting in cancer therapy is perceived as a promising strategy and is being studied in a broad range of approaches (Table\u00a0[1](#Tab1){ref-type=\"table\"}).Table\u00a01Overview of the key telomerase therapeutic strategies directed against telomerase and telomeres \\[[@CR30]\\], modifiedStrategyFactorTargetMechanismReferencesASOGRN163LhTRExpression inhibition\\[[@CR11]\\]RT inhibitionAZT azidothymidineDNA elongationReplication termination\\[[@CR12]\\]BIBR 1532RTNon-competitive inhibition\\[[@CR13]\\]ImmunotherapyGV1001HLACD4+ stimulation\\[[@CR14]\\]Vx-001HLA-A\\[[@CR15]\\]GRNVAC1/2 (AST-VAC1/2)Dendritic cellsAntigen presentation\\[[@CR16], [@CR17]\\]G-quadruplex stabilizationBRACO-1910Telomeric DNAAccess blocking\\[[@CR13]\\]RHPS4\\[[@CR18]\\]TMPyP4\\[[@CR19]\\]Telomestatin\\[[@CR20]\\]Quarfloxin/CX-3543^a^Telomeric DNArRNA biogenesis inhibition^a^\\[[@CR21]\\]Uncapping mimickingT-oligoTelomeric overhangApoptosis, and autophagy induction\\[[@CR22]\\]RNAihTERT-siRNA\\\nhTERC-siRNAhTERT\\\nhTERCExpression repression\\[[@CR23], [@CR24]\\]Preventing binding of telomerase to telomereIWR1, IWR2, JW55, flavone, XAV939Tankyrase inhibitionTRF1 dissociation from telomeres\\[[@CR25]\\]Nitroreductase-based gene therapyCB1954DNA crosslinksSuicide gene therapy\\[[@CR26]\\]Carboxypeptidase-based gene therapyZD2767PDNA-damaging alkylation\\[[@CR27]\\]Oncolytic virusesTelomelysin (OBP-301)hTERT positive cellsCancer cell lysis\\[[@CR28]\\]Gene therapyCRISPR/CAS9hTERT\\\nhTERCExpression repression\\[[@CR29]\\]All strategies aim to stop the main telomerase subunit expression, complex activity, or prevent substrate access to provoke cancer cell senescence, death, or proliferation of target cells*ASO* antisense oligonucleotides, *RT* reverse transcriptase^a^Requires verification in further studies\n\nHowever, some questions still appear, e.g., if the efficacy is an effect specific for the inhibitor or for the way of telomerase inhibition/downregulation? Is this effect similar when the key telomerase subunit hTERT is downregulated at the level of expression? Or is the same effect observed when telomeres are shortening? How are signaling pathways modulated at the same time (also in the context of genome destabilization)? How are those pathways modulated when drug therapy is accompanied by telomerase downregulation?\n\nThe association of telomerase and mismatch repair pathway seems to be related if we consider the fact that MSH2 (protein element of the DNA mismatch repair complex) binds hTERT promoter leading to its activation, and most probably to an increased expression of the key telomerase subunit hTERT, that is especially important during carcinogenesis \\[[@CR31]\\]. Surprisingly, studies concerning telomerase inhibition with BIBR1532 show no cumulative effect of telomerase inhibition and DOX in drug-resistant cells \\[[@CR32]\\]. Thus, a question arises, what is the mechanism of the correlation between drug sensitivity, drug resistance, and telomerase expression/activity, but also with the resulting mechanism of DNA damage leading to cancer cell death.\n\nThe role of hTERT translocation in cancer resistance to drugs {#Sec2}\n=============================================================\n\nThe hTERT gene encodes telomerase reverse transcriptase, the catalytic subunit of telomerase. As demonstrated, this is one of the two key telomerase subunits that enable enzyme activity, detected mainly in cancer or stem cells \\[[@CR33]\\]. The main function of hTERT is a reverse transcriptase activity that adds a six-base DNA repeat onto chromosome ends and prevents their shortening during successive cell divisions. Telomerase is associated with cell immortality and cancer, which may be related to the ability of hTERT to prevent apoptosis by stabilizing telomeres. However, fundamental information concerning the antiapoptotic function of hTERT is lacking, including a crucial question---whether its activity and/or nuclear localization are required and where telomerase acts to suppress the cell death process. It was demonstrated that overexpression of wild-type human TERT in HeLa cells, and in cells lacking hTERT but containing the telomerase RNA template, increases their resistance to apoptosis induced by the DNA damaging agent etoposide or the bacterial alkaloid staurosporine \\[[@CR34]\\]. In contrast, hTERT mutants with disruptions of either the RT domain or a 14-3-3 binding domain fail to protect cells against apoptosis. Similarly, overexpression of hTERT in cells lacking the telomerase RNA template was also ineffective in preventing apoptosis. Another finding shows that hTERT suppresses apoptosis at an early step before the release of cytochrome c and apoptosis-inducing factor from mitochondria, suggesting that hTERT can suppress a nuclear signal(s) that is an essential component of apoptotic cascades triggered by diverse stimuli.\n\nhTERT in response to stress {#Sec3}\n===========================\n\nMitochondria are the major source of ROS, which are mainly produced through the respiratory electron transport chain. Normally, intracellular ROS are dynamically balanced. When cells are exposed to oxidative stress, the endogenous production of ROS exceeds the capacity of the cellular antioxidant defenses, resulting in chemical damage of mtDNA. Mitochondrial DNA contains 37 genes encoding 13 structural proteins that are subunits of various respiratory chain complexes, 22 tRNAs, and two rRNAs. ND1 and COXII, which are encoded by mtDNA, are important components of respiratory chain complexes I and IV. Mitochondrial TERT has been shown to act as a TERC-independent reverse transcriptase and to exhibit RNA-dependent DNA polymerase activity using mitochondrial tRNA as a template \\[[@CR35]\\]. Mitochondrial TERT can bind to the RNA module of mitochondrial RNA processing endoribonuclease (RMRP) and form a complex similar to RNA-dependent RNA polymerase, which affects gene silencing at the post-transcriptional level \\[[@CR36], [@CR37]\\]. Other results suggest that mitochondrial TERT is involved in the regulation of COXII, a subunit of respiratory chain complexes \\[[@CR38]\\].\n\nFor the first time, the association between telomerase key subunit and cell resistance to stress was shown in 2003. More precisely, it was an oxidative stress. Authors revealed that not only telomerase activity but specifically the C-terminus of hTERT plays a role in hTERT-mediated cellular resistance to oxidative stress. As demonstrated, a 27-kDa hTERT C-terminal polypeptide (hTERTC27) devoid of domains required for telomerase activity was capable of nuclear translocation/telomere-end targeting. It was reported that a low level of hTERTC27 renders hTERT positive HeLa cells sensitive to H2O2-induced oxidative stress and subsequent cell senescence \\[upregulation of the cyclin/cdk inhibitor p21 (Waf1)\\]. Most importantly, no alteration in the expression of endogenous hTERT, significant telomere shortening, or inhibition of telomerase activity was observed \\[[@CR39]\\]. Another studies indicated the role of mitochondrial translocation of hTERT in resistance to drugs of human hepatocellular carcinoma (HCC) cells. As demonstrated, using three HCC cell lines (with different resistance index to cisplatin), the apoptosis rates in drug-resistant cells were significantly reduced. Cell-cycle analysis revealed the ratio of drug-resistant cells in *G*2/*M* and S phases increased, while that in *G*1 phase decreased. This process was accompanied by shortening of telomere length in drug-resistant cells under the chemotherapeutic stress. A reduction of damaged mtDNA with the increase in resistance index was revealed indicating protective effect of the mitochondrial translocation of hTERT. Therefore, it may suggest that the mitochondrial translocation of hTERT increases in multidrug-resistant cells and exerts protective effect on mitochondrial function which suggests that drug-resistant tumor cells escape from apoptosis through hTERT-mediated mitochondrial protection. Thus, it might be concluded that mitochondrial translocation of hTERT may serve as an underlying mechanism of MDR (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}) \\[[@CR40]\\].Fig.\u00a01Role of hTERT translocation in cancer\n\nAs reported elsewhere, mitochondrial translocation of hTERT led to chemotherapeutic resistance in HEPG2 cells. As suggested, phenomenon might result from the contribution of hTERT to reducing ROS production as well as mtDNA damage \\[[@CR18]\\]. In the studies performed in human leukemia cell line K-562, it was shown that telomerase activity inhibition by curcumin (measured using the TRAP assay) is due to suppression of translocation of hTERT subunit, from cytosol to nucleus. More importantly, this process was accompanied by induction of apoptosis in studied cells \\[[@CR41]\\]. The molecular mechanisms by which telomerase demonstrated a pro-survival function resulting in an increased resistance against DNA damage and decreased apoptosis induction remains elusive. It is still unclear whether it is associated with telomere maintenance or is rather a non-telomeric function of the telomerase protein, hTERT. However, the fact is that the protein subunit of telomerase can shuttle from the nucleus to the mitochondria upon oxidative stress. Consequently, it protects mitochondrial function and decreases intracellular oxidative stress. In addition, a significant correlation between nuclear localization of telomerase and high DNA damage was found. At the same time, cells which excluded telomerase from the nucleus displayed no or very low DNA damage. It is known that nuclear DNA damage can be caused by mitochondrially generated reactive oxygen species (ROS). It was shown that the mitochondrial localization of telomerase specifically prevents nuclear DNA damage by decreasing levels of mitochondrial ROS. This decrease of oxidative stress might be a possible cause for high stress resistance of cancer cells and could be especially important for cancer stem cells \\[[@CR42]\\].\n\nThe role of vaults in drug resistance and association with telomerase {#Sec4}\n=====================================================================\n\nIt has been some time already, since vaults were reported to contribute to the resistance of cancer cells to drugs. The vault complex, a large-sized ribonucleoprotein, was first described by Kedersha and Rome \\[[@CR43]\\]. The barrel-shaped structures were identified in preparations of clathrin-coated vesicles from rat liver. Their name comes from the fact that they resemble a morphology similar in shape to the vaulted ceilings in cathedrals. Such structures with similar dimension, morphology, and composition are found in cells of diverse eukaryotic organisms including mammals \\[[@CR44], [@CR45]\\]. Since they reveal a high degree of evolutionary conservation, it is suggested that they play an important cellular function.\n\nVaults' components are almost ubiquitously expressed and have been implicated in the regulation of several cellular processes including intracellular transport, signal transduction, and immune response. However, most importantly, they also contribute to chemotherapy resistance \\[[@CR46]\\]. Consequently, they are considered to play a protective role against xenobiotics and other stressing factors. The mammalian vault complex consists of multiple copies of three proteins: *M* ~r~ 100,000 (MVP), 193,000 (VPARP), and 240,000 (TEP1) (*M* ~r~, relative molecular mass). It also contains small untranslated RNA molecules of 88--141 bases.\n\nComponents of the vault complex {#Sec5}\n===============================\n\nMVP {#Sec6}\n---\n\nThe major vault protein (MVP, 110\u00a0kDa) was found to be identical to the previously described lung resistance-related protein (LRP) \\[[@CR47]\\] and constitutes over 70% of the total mass of the complex. One of the mechanisms leading to drug resistance is based on the ability to pump drugs away from intracellular drug targets through exocytotic vesicles or pump molecules. In addition, MVP/LRP is one of those pumps. The LRP gene is located on chromosome 16, close to the genes coding for the multidrug resistance-associated protein, MRP1. Overexpression of LRP constitutes a prediction of a poor response to chemotherapy in acute myeloid leukemia and ovarian carcinoma. As demonstrated, human MVP promoter lacks a TATA box, as well as other core promoter elements, but harbors putative transcriptional factor binding sites for Sp1 but also p53, STAT1, MyoD, GATA, and YB-1. Consequently, diverse transcription factors are involved in the regulation of MVP in different cell types. Some reports show that MVP expression is induced by a variety of cellular stresses such as DNA damaging agents, UV irradiation, hypoxia, and hyperthermia. Consequently, its role in radiotherapy response was also postulated \\[[@CR48]\\]. Kowalski et al. \\[[@CR49]\\] reported that MVP was involved in host resistance to infection with *Pseudomonas aeruginosa*. Ryu et al. \\[[@CR50]\\] demonstrated that MVP enhanced the resistance of cells to apoptosis induced by H2O2 in senescent human diploid fibroblasts. These findings suggest that MVP plays an important role in cellular responses to stress.\n\nMVP protein and mRNA levels were found to be upregulated in SW-620 cells after treatment with doxorubicin (DOX), etoposide (VP-16), cis-platinum (II) diammine dichloride (CDDP), or SN-38 at their IC50 concentration. Moreover, DOX could directly induce the transcriptional activity of MVP promoter regions, but it did not affect the stability of MVP mRNA. Interestingly, analysis of *MVP* knockout animal derived tissues indicated that the absence of MVP resulted in dramatically lowered cellular levels of other vaults' component, VPARP \\[[@CR51]\\].\n\nVPARP {#Sec7}\n-----\n\nThe vault poly-(ADP-ribose) polymerase (VPARP, 193\u00a0kDa) contains a functional poly-(ADP-ribose) polymerase (PARP) domain \\[[@CR52]\\]. VPARP is capable of ADP-ribosylating itself and the MVP, but this activity was not yet shown to be of functional importance within the vault complex.\n\nVPARP was originally identified as a minor protein component of the vault ribonucleoprotein particle, which may be involved in molecular assembly or subcellular transport. In addition to the association of VPARP with the cytoplasmic vault particle, subpopulations of VPARP localize to the nucleus and the mitotic spindle, indicating that VPARP may have other cellular functions. VPARP was also found to be associated with telomerase activity and interacted with exogenously expressed telomerase-associated protein 1 (TEP1) in human cells. Experiments with mice deficient in mVparp showed that the animals were viable and fertile for up to five generations, with no apparent changes in telomerase activity or telomere length. Vaults purified from mVparp-deficient mouse liver appeared intact, and no defect in association with other vault components was observed. Mice deficient in mTep1, whose disruption alone does not affect telomere function but does affect the stability of vault RNA, showed no additional telomerase or telomere-related phenotypes when the mTep1 deficiency was combined with an mVparp deficiency. These data suggest that murine mTep1 and mVparp, alone or in combination, are not necessary for normal development, telomerase catalysis, telomere-length maintenance, and vault structure in vivo \\[[@CR53]\\]. There are no data concerning human cells deprived VPARP, and noteworthy, telomeres and telomerase metabolism in those two species differ substantially.\n\nTEP1 {#Sec8}\n----\n\nTelomerase-associated protein (TEP1, 240\u00a0kDa) was primarily found to be associated with the telomerase complex \\[[@CR54]\\], where its function remains still unknown. TEP1 is an RNA binding protein that is not restricted to the telomerase complex and plays a redundant role in the assembly or localization of the telomerase RNP in vivo \\[[@CR55]\\]. It was also shown to interact specifically with the vault RNAs (vRNAs) \\[[@CR52]\\]. Since only two components of the telomerase complex seem essential for its function in vitro, i.e., hTERT and TR, TEP1 was thought to be a structural component \\[[@CR56], [@CR57]\\]. Indeed, analysis of a TEP1-deficient mouse model showed that this protein is not essential for telomerase activity. Moreover, telomere length was also unaffected after disruption of *TEP1* \\[[@CR55]\\]. However, contribution of TEP1 to telomere elongation in human is still postulated \\[[@CR58]\\]. Although not much is known about the function of TEP1, it was shown that polymorphisms of that gene may correlate with some types of cancer risk, e.g., bladder cancer \\[[@CR59]\\].\n\nVaults: summary {#Sec9}\n===============\n\nIt seems that vaults are new players in the field of drug resistance of cancer cells. They are not new, but it looks that novel functions have been assigned to them recently. The MVP components were shown to be associated with the insulin-like growth factor-1, hypoxia-inducible factor-1 alpha, and DNA double-strand break repair systems including non-homologous end joining and homologous recombination. Furthermore, MVP has been proposed as a useful prognostic factor associated with radiotherapy resistance \\[[@CR48]\\].\n\nVaults are abundantly present in the cytoplasm of eukaryotic cells and they were found to be associated with cytoskeletal elements as well as occasionally with the nuclear envelope. Vaults and MVP have been connected with several cellular processes which are also involved in cancer development like cell motility and differentiation. Due to the overexpression of MVP in several *P*-glycoprotein-negative chemoresistant cancer cell lines (*P*-gp/neg), vaults have been linked to multidrug resistance (MDR). Accordingly, high levels of MVP were found in tissues chronically exposed to xenobiotics. In addition, the expression of MVP correlated with the degree of malignancy in certain cancer types, suggesting a direct involvement in tumor development and/or progression. Based on the finding that MVP binds several phosphatases and kinases including PTEN, SHP-2 as well as Erk, a hypothesis arises that MVP might be involved in the regulation of important cell signaling pathways including the PI3K/Akt and the MAPK pathways \\[[@CR60]\\]. It is not obvious that telomerase or any subunit of the complex plays a pivotal role in those processes, but TEP1 is for sure a common component of both, vaults and telomerase.\n\nTelomerase and ABC-related drug resistance {#Sec10}\n==========================================\n\nNumerous reports show that in most cancer cases that do not respond to therapy, an abundant expression of ABCB1, ABCC1, or ABCG2 appears as poor prognostic factor \\[[@CR32]\\]. These proteins function as membrane transporters thus protect cells from potentially toxic substances and participate in the distribution of drugs. Changes in the expression and activity of ABC proteins may lead to an increased pumping out of drugs from cells and hence a reduction of the effective concentration of the drug in cancer cell. ABC transporters are often found to be inherently expressed in a wide variety of stem cells, where they provide improved protection from toxins \\[[@CR61]\\].\n\nCells gain the ability to acquire drug resistance at early stage of tumorigenesis process. Yague et al. \\[[@CR62]\\] have shown that it is possible to obtain drug-resistant cells by altering only expression of telomerase, p53, and pRb. The state of the art does not confirm if there is a direct correlation between telomerase expression and ABC-related drug resistance. However, in the literature, there are some examples of attempts to associate telomerase with drug resistance related to ABC overexpression. The concept of such association is based on work of Wang et al. \\[[@CR63]\\], who showed a prognostic value and correlation of telomerase expression with expression of genes associated with multidrug resistance in lung cancer patients. On the other hand, Sakin et al. \\[[@CR64]\\] suggested a lack of such correlation in an in vitro study performed in breast cancer cells. A tenuous correlation was shown in melanoma cells revealing co-expression of ABC transporters, ABCB5 and ABCC2 and hTERT \\[[@CR65]\\]. Smith et al. \\[[@CR66]\\] assessed a panel of human tumor cell lines resistant to vindesine, gemcitabine, and cisplatin, and disclosed that all have displayed changes in telomerase activity and/or telomere length compared to their parental lines. Recent findings suggest that signal transducer and activator of transcription 5 (STAT5) connects these two mechanisms through binding to the promoter regions of both the human TERT gene and the MDR1 gene \\[[@CR67], [@CR68]\\].\n\nAnother study shows that doxorubicin treatment of imatinib resistant cell line with *P*-glycoprotein overexpression was accompanied by an increased phosphorylation of BCR-ABL and STAT5, as well as increased hTERT expression. Interestingly, silencing of STAT5 expression reduced both the expression of *P*-glycoprotein and telomerase activity and restore sensitivity to imatinib in drug-resistant CML cells \\[[@CR67], [@CR68]\\]. A direct regulation of the MDR pathways and telomerase by interpheron-alpha (IFN-alpha) (used in the therapy of advanced cutaneous melanoma) has also been hypothesized. Treatment of melanoma cells with IFN-alpha, the MDR genes regulator, revealed no correlation between hTERT and TEP1 mRNA expression, whereas significant positive correlations were found between TEP1 and MDR1 mRNA, and between TEP1 and LRP/MVP mRNA \\[[@CR69]\\]. Alternative results were reported by Gomez et al. \\[[@CR70]\\]. They showed that in A549 derived clones resistant to a DNA-alkylating agent (ethyl methanesulfonate), and selected for resistance to 12549 (telomerase inhibitor; a potent G-quadruplex ligand), hTERT expression was significantly increased (accompanied by an increased telomerase activity) but with no multidrug resistance phenotype alterations. As reported, no variations in multidrug-related protein 1 (MRP1) and breast cancer resistance protein (BCRP) transcripts were observed while multidrug resistance 1 (MDR1) transcript was undetectable. It was also suggested that resistance to 12549 (or other G-4 ligands, i.e., telomestatin or BRACO-19) was associated with both, upregulation of telomerase expression/activity and alteration of telomere capping functions that may participate directly or indirectly in the mechanism of resistance in some cell types. Thus, a direct correlation of telomerase modulation and resistant phenotype acquisition was revealed but in a non-ABC-related mechanism.\n\nResistance mechanism specific to G-quadruplex inhibitors {#Sec11}\n========================================================\n\nThere is area of telomerase/telomere-based cancer strategy that focuses on the use of G-quadruplex stabilizers that prevent telomerase access to its substrate, and it constitutes a promising expansion. Eventually, it appears that G-quadruplex stabilizers play some more complex functions. Gomez et al. reported a potential correlation of telomerase and cancer cells resistance to G-quadruplex ligands that might be associated with a subtle balance between telomerase inhibition and telomere uncapping. These studies provided an evidence that telomerase activity and telomere length are key cellular determinants of the resistance phenotype \\[[@CR70], [@CR71]\\]. Moreover, cancer cells resistance to G-quadruplex ligands seem to be selective to this group of compounds that was confirmed by a cross resistance (as mentioned above) with other telomerase inhibitors with no correlation to other anti-cancer agents showing alternative mechanisms of action. An interesting report shows that ROS activation of PPM1D/WIP1 (DNA damage signaling pathway responsible for phosphorylation status of Chk1 and \u03b3-H2AX) by 12459 in A549 cells may lead to an inhibition of DNA repair signaling that triggers telomeric dysfunction or other genomic DNA damages but that finally leads to a delayed cell death \\[[@CR72]\\]. In addition, the relation between the way of treatment and observed effect (including the initiated mechanism) was observed implicating an effect dependent on variability of factors including long-/short-term treatment, concentration as well as the chemical structure of the ligand or the nature of the studied cell line. Noteworthy, the biological effect of G-quadruplex stabilization that is associated with a specific degradation of the telomeric overhang seems to be independent of the Bcl-2 expression status. Altogether, it gives an impression that telomere/telomerase targeting drugs trigger effects that are probably more directly associated with DNA stability, damage, and repair pathways but also with the mechanism of action of the therapeutic compound, e.g., direct DNA targeting or stress provoking (including oxidative stress) \\[[@CR73]\\]. It is clear that although telomere length in humans is determined by genetic predispositions, it is also significantly affected by environmental factors that altogether results in different (individual) attrition rate. It is also evident that direct modulation of telomerase activity by inhibitors, quadruplex stabilizers, repressors or ROS, and other stress factors (including psychological and life stress) affect telomere length. Consequently, even smoking (an excellent example for higher ROS formation) leads to a progressive shortening of telomeres. Especially, the GGG triplet within the human telomere sequence (TTAGGG) is particularly vulnerable to chemical modifications. Thus, telomere or telomerase targeting may ultimately lead to genomic instability. Some authors claim that on one hand, it may be used in cancer therapy, but on the other hand, it may contribute to initiation of carcinogenesis \\[[@CR74]\\].\n\nTelomerase-positive and ABC genes expressing cancer stem cells (CSC): neoplasm maintenance {#Sec12}\n==========================================================================================\n\nYears of research work in the field of cancer biology confirmed highly heterogenic character of cancer cell populations \\[[@CR75]\\]. A variety of types of cells differing in morphology as well as in molecular status were observed within malignant tumor or among the leukemic cells in circulation of one patient. It is believed that most, if not all, tumors contain a small sub-population of cancer stem cells (CSC), called also cancer stem-like cells or tumor-initiating cells \\[[@CR76]\\]. CSCs share some features with normal adult stem cells. Both types of cells exhibit high levels of telomerase activity, increased level and/or activity of transmembrane ATP-binding cassette (ABC) transporter proteins, and possess the ability to self-renewal and differentiation, induction of antiapoptotic mechanisms, and migration potential \\[[@CR77], [@CR78]\\]. Cancer stem cells are able to reproduce themselves and their presence in the heterogenous neoplastic cells population enables sustaining the cancer. CSCs can \"differentiate\" into nontumorigenic mature cancer that drives the tumor growth and spread \\[[@CR79]\\].\n\nOne of the theories explaining the immortality of tumor cells proposes that cancer originates from normal stem cells, which show constitutive telomerase activity. According to this cancer stem cell model of carcinogenesis, the event that initiates neoplastic transformation occurs in normal stem cell \\[[@CR80]--[@CR82]\\]. There are scientific evidences proving that during aging, stem cells can accumulate so many mutations that the scenario where nontumorigenic normal stem cell becomes tumor-initiating cell is very likely \\[[@CR83]\\]. Undoubtedly, telomerase is crucial for cancer stem cell existence. Studies with the use of animal experimental model (Sca1-BCRABLp210) with the established telomerase deficiency revealed that telomerase presence is necessary for the generation and maintenance of cancer stem cells. Complete deletion of telomerase coding gene or even loss of one allele of the gene resulted in prevention of BCR-ABL-induced chronic myeloid leukemia (CML) development \\[[@CR82]\\]. Interestingly, one of the key antiapoptotic proteins---Bcl-2---is not required to sustain leukemic stem cells population in mice \\[[@CR84]\\]. Inhibition of telomerase activity with the use of Imetelstat caused a significant reduction of clonogenic growth of CD138^\u2212^ cancer stem cells isolated from patients with myeloma \\[[@CR85]\\]. There are reports showing that some cancer stem cells and mature cancer cells have short telomeres while exhibiting significant telomerase activity \\[[@CR86]\\]. These observations may indicate that telomerase in cancer stem cells may play some other function(s) than only telomere lengthening.\n\nAnother important feature of cancer stem cells is their intrinsic drug resistance related to overexpression of genes encoding ABC transporter proteins. It is a serious problem from the clinical and pharmacological point of view, because CSCs, that are rather rare in the tumor mass, survive the therapy (chemo- and radiotherapy) and propagate giving rise to new tumor. Cancer stem cells are characterized by high resistance to antimitotic drugs resulting from the upregulation of drug efflux pumps. ABCB1, ABCG2, and ABCB5 transmembrane proteins are listed among CSC's markers \\[[@CR87]\\]. In human melanoma cells, coexistence of ABCB1, ABCG2, and ABCC2 together with stem cell markers on the cell surface was observed in a subset of cells. Moreover, it was reported that in melanoma stem cells, concomitant occurrence of ABCG2 and CD133 can be a stemness marker \\[[@CR88]\\]. CSCs isolated from pancreatic cancer cell line PANC1 were verified as cells expressing both *ABCB1* and *ABCG2* \\[[@CR89]\\]. In addition, the ability of CSCs to evade chemo- and radiotherapy is enhanced by the fact that the cells are relatively quiescent and slow cycling. These two features protect them against drugs targeting rapidly proliferating cells \\[[@CR87]\\].\n\nConcluding, telomerase-positive cancer cells with stem-like properties are considered the most resistant subset of cells within the malignant tumor. One of the extratelomeric function of telomerase is an interplay with molecules regulating gene expression. TERT (telomerase subunit) is hypothesized to act as a transcription factor and play a key role in the modulation of many processes related to neoplastic transformation, induction of cancer stemness as well contribution to the drug resistance development \\[[@CR90]\\].\n\nSensitization of cancer cells through telomerase inhibition {#Sec13}\n===========================================================\n\nNumerous studies reported a correlation between telomerase inhibition and sensitization of cancer cells. As demonstrated, pharmacological inhibition of telomerase catalytic activity by BIBR1532 sensitized both drug-sensitive and drug-resistant leukemia and breast cancer cells; moreover, these effects were telomere-length-dependent \\[[@CR32]\\]. Another telomerase inhibitor, imetelstat, sensitized primary CLL lymphocytes to fludarabine in vitro \\[[@CR91]\\]. Treatment of flavopiridol-resistant cells with G-quadruplex stabilizer BRACO-19 alone also led to a rapid inhibition of cell growth that was not observed in the parental cell line. It is noteworthy that combination of BRACO-19 and flavopiridol caused a threefold reduction in cell proliferation compared to the use of flavopiridol alone. Interestingly, this effect was not observed in parental colon cancer cell line that was sensitive to the flavopiridol and showed lower telomerase activity \\[[@CR92]\\]. Treatment with doxorubicin and taxol combined with inhibition of telomerase increased senescence and apoptosis of breast cancer cells. Another study revealed that telomerase inhibition itself, not necessarily short telomeres, was crucial for sensitization in treated cells \\[[@CR9]\\]. Similarly, telomere-length independent cellular response was observed by Park et al. \\[[@CR93]\\]. Knockdown of hTERT in resistant bladder cancer cell line increased cisplatin-induced cell death and accelerated the translocation of Bax into the mitochondrial membrane and the release of cytochrome C. Thus, once more, telomerase was suggested to play an important role in mitochondrial metabolism during the process of intrinsic apoptosis.\n\nTelomerase and DNA damage/apoptosis resistance {#Sec14}\n==============================================\n\nMultiple studies show that telomeres and telomerase are centrally involved in the capability for an unlimited proliferation of cells \\[[@CR94]\\]. Since telomeres are essential for maintaining and protecting chromosomes from degradation, telomere dysfunction is considered to play a crucial role in initiating genomic instability. Abnormalities in telomere function lead to chromosome rearrangements, breakage--fusion--bridge cycles (BFB), and the generation of novel chromosomal variants. As revealed, in the presence of intact p53, critically short telomeres induce cell-cycle arrest, cellular senescence, or apoptosis what suppress tumorigenesis. Dysfunctional telomeres are also recognized as double-strand DNA breaks, activating the DNA damage response (DDR) checkpoints, including ataxia telangiectasia mutated (ATM) and checkpoint kinase 2 (CHK2) \\[[@CR95]\\]. Thereby, telomere shortening decreases lifespan and functions as a barrier to cancer development.\n\nAs suggested, telomerase itself has a pro-survival function resulting in an increased resistance against DNA damage and decreased apoptosis induction \\[[@CR2]\\]. Rubio et al. \\[[@CR96]\\] showed that telomerase confers resistance to ionizing radiation, bleomycin, hydrogen peroxide, and etoposide in human fibroblast cells with short, near-dysfunctional, telomeres. This resistance depends on the ability of telomerase to elongate the short telomeres. Moreover, the authors revealed that telomerase did not reduce genotoxic stress in cells with long telomeres. As mentioned already, telomerase exclusion from nucleus into mitochondria could be one of the possible non-telomeric mechanisms that decrease nuclear DNA damage and apoptosis caused by an anti-cancer treatment. Beside the decrease of mitochondrial ROS level, telomerase can also disrupt DNA repair processes through interference with DNA repair enzymes under DNA damage conditions \\[[@CR42]\\]. This might contribute to increased resistance of cancer cells to various anti-cancer treatments.\n\nSeveral reports show that hTERT might play an important role in direct inhibition of at least two cell death pathways. It was shown that hTERT depletion upregulates the induction of apoptotic cell death by cisplatin, etoposide, mitomycin C, and reactive oxygen species in cervical and colon cancer cell lines. Interestingly, inhibition of mitochondrial membrane permeabilization precluded the induction of cell death by the combination of hTERT downregulation and chemotherapeutic agents. These findings indicate that hTERT functions as an endogenous inhibitor of mitochondrial cell death pathway \\[[@CR97]\\].\n\nEctopic expression of hTERT is responsible for suppressing a p53-dependent cell death induced by DNA-damaging agents. The same effect was observed for both normal and mutant hTERT (without telomerase activity) and allowed cells to be resistant to DNA damage and to suppress activation of p53 \\[[@CR98]\\] (see Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). Similarly, bortezomib inhibition of hTERT expression in leukemic and gastric cancer cells contributed to telomere dysfunction and cellular apoptosis. In addition, hTERT overexpression protected cells against bortezomib-induced DNA damage \\[[@CR99]\\].Fig.\u00a02Potential interactions between telomerase and drug resistance in cancer cells\n\nOne of the telomere-binding proteins, TRF2, regulates both telomere protection and telomere length as well as interacts with several DNA repair proteins. It was found that TRF2 expression increases after DNA double-strand breaks induction by anti-cancer drugs or irradiation in gastric cancer cells. As demonstrated, TRF2 upregulation was more dramatic in drug-resistant cells, whereas its inhibition by RNAi in drug-resistant cells partially reversed its resistance phenotype \\[[@CR100], [@CR101]\\].\n\nConclusions {#Sec15}\n===========\n\nTelomerase regulation studies concern mainly the relationship between telomerase activity inhibition and cancer treatment. In addition, the impact on the cell biology and the non-telomeric functions of telomerase is also investigated. First observations of connection between telomerase and drug resistance in cancer concerned sensitization of resistant cells after telomerase inhibition. Surprisingly, only imatinib mesylate (telomerase inhibitor) is documented to reveal desired effects in treating CML patients. However, lot of potential telomerase inhibitors still is under research and clinical trials. Direct connection between telomerase (or its components) and ABC transporter expression/function seems to occur only in some types of cancer cells. This may be due to common telomerase but limited ABC genes expression regulation.\n\nWe assume that non-canonical functions of telomerase might play a more important role and contribute towards the higher resistance of cancer cells. That is supported by a fact that hTERT translocation between the nucleus and mitochondria can be considered as a negative prognostic factor in cancer cells. Finally, telomerase components are involved in death pathways at various levels, and thus, its regulation may be crucial for the development of death-resistance mechanisms (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). Unfortunately, some telomerase-based approaches provoke activation of alternative lengthening of telomerase (ALT), which may underlie the development of another resistant mechanism. Anyway, the correlation of telomerase and drug resistance is a fact. The question is how those processes are associated? However, another question appears---how to use this knowledge to overcome cancer cells resistance? How useful those data could be, partly was already presented in some pioneering studies performed in 3D culture systems. Liu et al. \\[[@CR102]\\] revealed that CaSki sphere-forming cells (exhibiting higher telomerase activity) were more resistant to chemotherapeutic drugs than the control CaSki cells. These studies, however, are in progress but without a doubt should be continued.\n\nNatalia Lipinska and Przemyslaw Kopczynski have equally contributed to the work.\n\nThe present review was supported by grants from the National Science Centre: 2016/21/B/NZ7/01079, 2011/03/B/NZ7/00512, and 2014/13/N/NZ7/00307 and Poznan University of Medical Sciences 2016/21/B/NZ7/01079.\n"}
+{"text": "Background\n==========\n\nColorectal cancer is one of the most common causes of cancer deaths worldwide \\[[@B1]\\]. The median overall survival of patients with metastatic colorectal cancer has increased from 12 months to approximately 24 months over the past decade as a result of improvements in systemic therapies, including new chemotherapeutic agents such as irinotecan and oxaliplatin and monoclonal antibodies against epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF). However, the 5-year survival of patients is still poor \\[[@B2]-[@B5]\\]. Thus, one of the main challenges for the treatment of colorectal cancer remains to identify new strategies beyond chemotherapy to inhibit disease progression.\n\nColorectal cancers are characterized by infiltration with multiple stromal cells, among which are tumor-infiltrating lymphocytes (TILs) that act as prognostic and predictive factors \\[[@B6]-[@B15]\\]. TILs include natural killer (NK) cells, CD8+ T cells, and CD4+ T cells, including Th1, Th2, Th17, and Treg cells. Although the role of TILs in tumor progression is controversial, CD45RO+ T cells have been identified as the main anti-tumoral effectors in early colorectal cancers. Hierarchical clustering has shown that markers of T-cell migration, activation, and differentiation are increased in tumors without signs of early metastatic invasion. These tumors have an increased number of CD8+ T cells, ranging from early memory (CD45RO+, CCR7-, CD28+, and CD27+) to effector memory (CD45RO+, CCR7-, CD28-, and CD27-) T cells. The presence of high levels of infiltrating memory CD45RO+ T cells is correlated with the absence of signs of early metastatic invasion, a less advanced pathologic stage, and increased survival, which has been confirmed in several series of patients \\[[@B16],[@B17]\\]. These correlations indicates that protective immune responses exist in a subset of colorectal cancer patients. Activated tissue-resident memory T cells, which have a potent lytic potential with the expression of perforin and granzyme B, are not only capable of providing immediate effector function at the site of cancer cells but can also generate an effective secondary immune response \\[[@B18]-[@B21]\\]. Radiotherapy and some chemotherapeutics have the potential to induce the immunogenic death of cancer cells and subsequently activate memory T cells. Furthermore, clinical trials have shown that the efficacy of chemotherapy against colon cancer can be improved when combined with cytokines \\[[@B22]-[@B27]\\]. These data indicate that the activation of memory T cells that have infiltrated into tumor tissues might be a promising therapeutic strategy.\n\nTissue-resident memory T cells result from the acute immune response, which may be induced by damage-associated molecular pattern molecules (DAMPs). High-mobility group box 1 (HMGB1), previously named HMG1, amphoterin, and sulfoglucuronyl carbohydrate binding proteinis a DAMP. Under physiologic conditions, HMGB1 localizes in the nucleus. HMGB1 is released into the extracellular matrix by spontaneous release or by dying cells. HMGB1 is a multi-functional protein that accelerates cell growth, invasion, and angiogenesis in cancer tissues, induces apoptosis in macrophages, and promotes immune responses \\[[@B28]-[@B31]\\]. With respect to modulation of the immune response, HMGB1 binds to surface receptors expressed on dendritic cells (DCs), leading to the maturation of DCs, antigen processing, and a subsequent protective effect. There is also evidence that HMGB1 induces immune tolerance by interfering with the antigen-presenting process and decreasing the number of DCs or by binding to the TLR9 agonist, which leads to human plasmacytoid DC suppression \\[[@B32]-[@B36]\\]. In addition, extracellular HMGB1 triggers the inflammatory cascade through multiple pathways, facilitating progression of the tumor.\n\nBecause HMGB1 has a dual effect on tumor progression, the feasibility of targeting the release of HMGB1 remains elusive. To this end, this study determined whether the expression of HMGB1 is related to the infiltration of T cells and patient prognosis in locally advanced colon cancers.\n\nMethods\n=======\n\nMaterials\n---------\n\nSeventy-two cases of pathologically-confirmed specimens matched with adjacent normal mucosa were obtained from patients with stage IIIB (T3N1M0; AJCC, 2002) colon cancer between January 1999 and May 2002 at the Cancer Center of Sun Yat-Sen University in Guangzhou, China (Table [1](#T1){ref-type=\"table\"}). All of the patients underwent radical resection and 5-FU-based adjuvant chemotherapy post-operatively for 6 months. Patients were evaluated every 3 months during the 1^st^year, every 6 months in the 2^nd^year, and by telephone or mail communication once every year thereafter for a total of 5 years. If recurrence or metastasis occurred, 5-FU-based chemotherapy was given according to the NCCN guidelines. Overall survival was defined as the time from surgery to death. Alternatively, censoring was done at the last known date the patient was alive. This study was approved by the institutional ethical review committee of Sun Yat-Sen University Cancer Center.\n\n###### \n\nPatient Characteristics (N = 72)\n\n  Characteristic              No. of patients (%)\n  --------------------------- ---------------------\n  **Age, years**              \n  \\< 60                       33(45.8)\n  \u226560                         39(54.2)\n  **Gender**                  \n  Male                        40(55.6)\n  Female                      32(44.4)\n  **Tumor sites**             \n  Left hemicolon              46(63.9)\n  Right hemicolon             26(36.1)\n  **Pathological grade**      \n  G1                          10(13.9)\n  G2                          54(75.0)\n  G3                          8(11.1)\n  **Survival time, months**   \n  \u226560                         52(72.2)\n  \\< 60                       20(27.8)\n\nImmunohistochemical assay and scoring systems\n---------------------------------------------\n\nFormalin-fixed, paraffin-embedded archived tissues were cut into 4-\u03bcm sections. The size of each tissue section was about 1.0 cm\u00d71.5 cm. Then, the sections were de-waxed, rehydrated, blocked with hydrogen peroxide, and the antigens were retrieved in a microwave in 10 mM citrate buffer (pH 6.0) for 10 minutes and cooled to room temperature. After blocking with sheep serum, the sections were incubated overnight at 4\u00b0C with either rabbit polyclonal antibody against human HMGB1 at a dilution of 1:1000 (Abcam, Cambridge, MA, USA) or mouse monoclonal antibody against human CD3, CD45RO, CD4, CD8, and CD56. All of these antibodies (Zymed, San Diego, CA, USA) were diluted 1:100. Subsequently, biotinylated secondary antibodies and streptavidin-biotinylated horseradish peroxidase complex were used. The sections were developed with diaminobenzidine tetrahydrochloride (DAB) and counterstained with hematoxylin. Negative controls were employed in which the primary antibody was replaced by phosphate-buffered solution (PBS).\n\nThe density of TILs within the tumors and the expression of HMGB1 in cancer cells were scored with two scoring systems. Hussein\\'s method was used to score the density of TILs as follows: 1) the cells were counted in at least 10 different fields of each section, and the size of each high-power field (\u00d7400) was about 300 \u03bcm\u00d7300 \u03bcm; 2) the cells were counted in the tumor stroma; 3) the areas of highest density were chosen; 4) necrotic areas were avoided; 5) two observers counted the cells at the same time and in the same field using a multiple-lens microscope; and 6) the results were expressed as the mean \u00b1 standard error of the mean \\[[@B37]\\]. The expression of HMGB1 was interpreted via immunoreactivity using the 0-4 semi-quantitative system derived from Soumaoro \\[[@B38]\\] for both the intensity of staining and the percentage of positive cells (labeling frequency percentage). The intensity of nuclear or cytoplasmic staining was grouped into the following four categories: no staining/background of negative controls (score = 0), weak staining detectable above background (score = 1), moderate staining (score = 2), and intense staining (score = 3). The labeling frequency was scored as 0 (\u22641%), 1 (1%-24%), 2 (25%-49%), 3 (50%-74%), and 4 (\u226575%). The sum index was obtained by totaling the intensity and percentage scores, as follows: (-), (+), (++), and (+++) indicated sum-indexes of 0-1, 2-3, 4-5, and 6-7, respectively; (-) and (+) were defined as no or modest expression, and (++) and (+++) were defined as strong expression. Each section was independently scored by two pathologists. If an inconsistency occurred, a third pathologist was consulted to achieve consensus.\n\nDouble immunohistochemistry for HMGB1 and CD45RO\n------------------------------------------------\n\nDouble immunohistochemical staining was performed with a Polymer kit (GBI, Seattle, WA, USA). The sections were simultaneously incubated with two primary antibodies derived from different species. After blocking nonspecific binding with goat serum, the sections were incubated overnight at 4\u00b0C with a rabbit anti-human HMGB1 polyclonal antibody (1:1000; Abcam, Cambridge, MA, USA) and a mouse anti-human CD45RO monoclonal antibody (1:100; Zymed, San Diego, CA, USA) simultaneously. After washing in PBS, the sections were incubated with both alkaline phosphatase-labeled goat anti-rabbit IgG antibodies and horse radish peroxidase-labeled goat anti-mouse IgG antibodies with a definite percentage for 30 minutes at room temperature. After three washes with PBS, sections were developed sequentially with DAB and AP-Red and counterstained with hematoxylin. Negative controls were employed in which both of the primary antibodies were replaced with PBS.\n\nFlow-cytometric analysis\n------------------------\n\nFresh tumor specimens were processed via sterile mechanical dissection. The tissue was cut into small pieces and stirred for 2 hours at 37\u00b0C in an enzymatic bath containing RPMI 1640 (Invitrogen, San Diego, CA, USA) and HEPES buffer (20 mmol/L) containing penicillin/streptomycin (120 \u03bcg/ml and 100 ug/ml, respectively), fungizone (0.25 mg/ml), ceftazidime (50 \u03bcg/ml), collagenase type III (200 U/ml), and DNase type I (100 U/ml). The suspension was then filtered through two wire grids (70 and 40 \u03bcm), and the cells were washed three times with Hank\\'s balanced salt solution (HBSS). Subsequently, cells were separated on a Percoll density gradient (Pharmacia Biotech AB, Uppsala, Sweden) for 30 min at 1500 \u00d7 *g*at room temperature. The dense layer, enriched for lymphocytes, was collected and washed. Then the cells were incubated for 30 minutes at 4\u00b0C with antibodies conjugated to APC against CD3 (Biolegend, San Diego, CA, USA) and phycoerythrin (PE) against CD45RO (Biolegend, San Diego, CA, USA). Analyses were performed with a five-color fluorescence-activated cell sorter (FC500; Beckman-Coulter, Fullerton, CA, USA) and Cxp cytometer software (Beckman-Coulter, Fullerton, CA, USA). Immune subpopulations were measured as a percentage of the total number of all cells and a percentage of the total number of CD3+ cells.\n\nStatistical analysis\n--------------------\n\nThe correlation between the density of TILs or HMGB1 expression with patient characteristics and the correlation between the density of TILs and levels or subcellular location of HMGB1 were analyzed with a chi-square test or Fisher\\'s exact test. The following factors were assessed with both univariate and multivariate analyses to determine their influence on overall survival: gender, age, pathologic grade, tumor site, the density of CD3+ cells, the density of CD45RO+ cells, and the level or subcellular location of HMGB1 expression within colon cancer tissues. Kaplan-Meier curves were used to estimate the distributions of those variables in relation to survival, which were compared using the log-rank test. The Cox regression model was used to correlate assigned variables with overall survival. All statistical analyses were carried out using SPSS 13.0 software (SPSS Inc., Chicago, IL, USA). Statistical significance was assumed for a two-tailed *P*\\< 0.05.\n\nResults\n=======\n\nPatterns of TIL infiltration and HMGB1 expression within stage IIIB colon cancer tissues\n----------------------------------------------------------------------------------------\n\nCD3+ and CD45RO+ cells were observed in all cases in the tumor stroma and the adjacent normal mucosa to different extents. Both antigens stained the cell membrane. In addition, we performed immunostaining against CD4, CD8, and CD56 in 10 of the tissue sections to determine which of these proteins might play an important role in this group of patients. Among the CD3+ cells, CD8+ cells outnumbered CD4+ cells, while CD56+ cells were rare (Fig. [1](#F1){ref-type=\"fig\"}, [2](#F2){ref-type=\"fig\"}. To confirm the phenotype of the CD45RO+ cells, co-expression of CD3 with CD45RO was examined in five fresh cancer tissues via flow cytometry, which indicated that 72% of all tumor-infiltrating CD45RO+ lymphocytes were also CD3-positive (Fig. [3](#F3){ref-type=\"fig\"}).\n\n![**Patterns of TILs within stage IIIB colon cancer tissues**. A high density of CD3+ cells infiltrated into the colon cancer stroma (\u00d7400, A), a low density of CD3+ cells infiltrated into the colon cancer stroma (\u00d7400, B), a high density of CD45RO+ cells infiltrated into the colon cancer stroma (\u00d7400, C), and a low density of CD45RO+ cells infiltrated into the colon cancer stroma (\u00d7400, D, \\\"\u2191\\\"refers to tumor tissue). All of the antigens were immunolocalized to the membrane.](1471-2407-10-496-1){#F1}\n\n![**Subtype of CD3+ cells within stage IIIB colon cancer tissues**. A high density of CD3+ cells infiltrated into the colon cancer stroma (\u00d7400, A); CD4+ cells infiltrated into the colon cancer stroma (\u00d7400, B), CD8+ cells infiltrated into the colon cancer stroma (\u00d7400, C), and CD56+ cells infiltrated into the colon cancer stroma (\u00d7400, D). All of the antigens were immunolocalized to the membrane.](1471-2407-10-496-2){#F2}\n\n![**Co-expression of CD3 and CD45RO antigens among TILs analyzed with flow cytometry**.](1471-2407-10-496-3){#F3}\n\nAlthough HMGB1 staining existed only in the nucleus in normal mucosa, co-expression of nuclear and cytoplasmic HMGB1 (co-expression pattern) existed in a subset of cancer tissues (12/65 \\[18.5%\\]), whereas most of the cancer tissues only showed nuclear staining of HMGB1 (nuclear expression pattern; 53/65 \\[81.5%\\]; Table [2](#T2){ref-type=\"table\"}, Fig. [4](#F4){ref-type=\"fig\"}).\n\n###### \n\nPatterns of HMGB1 expression in stage IIIB colon cancers (N = 72)\n\n  ------- ---------------------- ------------------------- -------------- --------------------------------------\n          **HMGB1 expression**                                            \n  **0**   **+**                  **++-+++**                               \n                                                                          \n          nuclear only           nuclear and cytoplasmic   nuclear only   nuclear only nuclear and cytoplasmic\n  7       20                     5                         33             7\n  ------- ---------------------- ------------------------- -------------- --------------------------------------\n\n![**Expression of HMGB1 in colon cancer and the adjacent normal mucosa**. Nuclear staining of HMGB1 (\u00d7100 in A and \u00d7400 in B) and nuclear and cytoplasmic staining of HMGB1 (\u00d7100 in C and \u00d7400 in D).](1471-2407-10-496-4){#F4}\n\nThe relationship between HMGB1 expression and the density of CD3+ or CD45RO+ cells\n----------------------------------------------------------------------------------\n\nThe density of TILs within the tumors was within the 75^th^percentile for all those specimens in which the survival difference between the \\\"high\\\" and \\\"low\\\" groups was the largest \\[[@B17]\\]. The cut-off values for the density of CD3+ and CD45RO+ cells were 16 and 24 cells per high-power field in the center of the tumor, respectively. Thus, the density of CD3+ cells was recorded as high if \\> 16 cells were observed per high-power field, and the same cutoff used for CD45RO+ cells. The log-rank test was used to compare the level of HMGB1 with the density of CD3+ and CD45RO+ cells within the tumor. A stronger expression (++-+++) of HMGB1 was associated with a higher density of both CD3+ and CD45RO+ cells (Table [3](#T3){ref-type=\"table\"}, Fig. [5](#F5){ref-type=\"fig\"}). Furthermore, the log-rank test was used to compare the subcellular localization of HMGB1 with the density of CD3+ and CD45RO+ cells within the tumor. The co-expression pattern of HMGB1 in the nucleus and cytoplasm was inversely associated with infiltration of both CD3+ and CD45RO+ cells, respectively (Table [4](#T4){ref-type=\"table\"}, Fig. [5](#F5){ref-type=\"fig\"}).\n\n###### \n\nRelationship between the expression of HMGB1 and the density of TILs as well as patient characteristics (N = 72)\n\n  ----------------------------------------------------------------------------------------------\n  Characteristic                                             HMGB1 expression\\     *P*   \n                                                             (-)- (+) (++)-(+++)         \n  ---------------------------------------------------------- --------------------- ----- -------\n  **Gender**                                                                             0.058\n\n  Male                                                       22                    18    \n\n  Female                                                     10                    22    \n\n  **Age, years**                                                                         0.113\n\n  \u226560                                                        14                    25    \n\n  \\<60                                                       18                    15    \n\n  **Tumor sites**                                                                        0.476\n\n  Left hemicolon                                             19                    27    \n\n  Right hemicolon                                            13                    13    \n\n  **Pathological grade**                                                                 0.182\n\n  G1                                                         4                     6     \n\n  G2                                                         22                    32    \n\n  G3                                                         6                     2     \n\n  **Density of CD3+cells (cells per high-power field)**                                  0.013\n\n  \u226516                                                                                    \n\n  \\<16                                                       20                    35    0.006\n\n  **Density of CD45RO+cells (cells per high-power field)**   12                    5     \n\n  \u226524                                                        19                    35    \n\n  \\<24                                                       13                    5     \n  ----------------------------------------------------------------------------------------------\n\n![**Double staining of HMGB1 and CD45RO+ cells in stage IIIB colon cancer (red staining of HMGB1 in cancer cells and brown staining of CD45RO+ cells in the stroma)**. Strong nuclear staining of HMGB1 was associated with a high density of CD45RO+ cells infiltrating into colon cancer tissues (\u00d7100 in A and \u00d7400 in B, \\\"\u2191\\\"refers to nuclear staining in cancer cells). Strong nuclear and cytoplasmic staining of HMGB1 was also associated with a low density of CD45RO+ cells infiltrating into colon cancer tissues (\u00d7100 in C and \u00d7400 in D, \\\"\u2191\\\"refers to nuclear or cytoplasmic staining in cancer cells).](1471-2407-10-496-5){#F5}\n\n###### \n\nRelationship between subcellular localization of HMGB1 and density of TILs (N = 65)\n\n  Subcellular localization of HMGB1                      *P*value       \n  ------------------------------------------------------ ---------- --- --------\n  Density of CD3+cells (cells per high-power field)                     \n  \u226516                                                    47         5   0.000\n  \\<16                                                   6          7   \n  Density of CD45RO+cells (cells per high-power field)                  \n  \u226524                                                    46         5   0.0001\n  \\< 24                                                  7          7   \n\nRelationship between the density of TILs, HMGB1, clinical characteristics, and 5-year survival rate as analyzed via univariate analysis\n---------------------------------------------------------------------------------------------------------------------------------------\n\nThe log-rank test was used to compare the density of CD3+ cells and CD45RO+ cells and the level of HMGB1 with patient characteristics. The density of CD3+ cells within the tumor stroma was associated with pathologic grades (*P*=0.023), i.e., a lower pathologic grade was associated with less CD3+ cell infiltration. However, neither the density of the CD45RO+ cells nor the levels of HMGB1 were associated with any of the clinical characteristics.\n\nBy the end of the 5-year follow-up period, 52 patients were alive, making the 5-year survival rate equal to 72% in this group of patients. Kaplan-Meier survival analysis indicated that a higher density of either CD3+ or CD45RO+ cells within the tumor was associated with a higher 5-year survival rate (Fig. [6](#F6){ref-type=\"fig\"}). However, the association between the patterns of HMGB1 expression and 5-year survival rate was complicated. The survival curves crossed at 24 months between patients with nuclear-only expression of HMGB1 and those with co-expression of nuclear and cytoplasmic HMGB1 (Fig.7A). Kaplan-Meier survival analysis revealed no statistical difference in survival between the patients with nuclear-only expression of HMGB1 and those with co-expression of nuclear and cytoplasmic HMGB1 within 24 months after surgery. After 24 months, the patients with the co-expression of nuclear and cytoplasmic HMGB1 showed a lower 5-year survival rate compared with the patients with nuclear-only expression of HMGB1 (Table [5](#T5){ref-type=\"table\"}, Fig. [7B](#F7){ref-type=\"fig\"}).\n\n![**The association between overall survival with the density of CD3+ and CD45RO+ cells as well as HMGB1 expression in stage IIIB colon cancers**. A high density of CD3+ cells within the tumor was associated with a longer 5-year survival rate (A); a high density of CD45RO+cells within the tumor was associated with a longer 5-year survival rate (B).](1471-2407-10-496-6){#F6}\n\n###### \n\nRelationship between the localization of HMGB1 and prognosis after 24 months of follow-up (N = 57)\n\n                           Subcellular location of HMGB1   *P*value   \n  ------------------------ ------------------------------- ---------- ----------\n  Survival time (months)                                              \\< 0.001\n  \u226560                      43                              5          \n  \u226524and \\< 60             3                               6          \n\n![**The association between overall survival and the subcellular localization of HMGB1 in stage IIIB colon cancers**. No significant correlation with 5-year survival for patients with nuclear or nuclear and cytoplasmic expression of HMGB1 in cancer cells was observed because the survival curves crossed at 24 months (A), nuclear and cytoplasmic expression of HMGB1 in cancer cells was inversely associated with prognosis after 24 months of follow-up (B).](1471-2407-10-496-7){#F7}\n\nRelationship between the density of TILs, the level or subcellular location of HMGB1, clinical characteristics, and 5-year survival rate as analyzed via multivariate survival analysis\n---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nThe Cox regression model revealed that only those patients with a higher density of CD45RO+ cells within the tumor stroma had a longer survival, whereas no relationship was observed between the survival and the density of CD3+ cells. There was also no relationship between the survival and the levels or the subcellular distribution (nucleus only or nucleus combined with cytoplasm) of HMGB1 in primary tumors. These results indicated that in this group of patients, the density of CD45RO+ cells within the tumor was independently prognostic (Table [6](#T6){ref-type=\"table\"}). The same result occurred in the time-dependent multivariate survival analysis (data not shown).\n\n###### \n\nMultivariate survival analysis (N = 72)\n\n                             B       SE     Wald    df   **Sig**.   Exp(B)   95%CI for Lower   Upper Exp(B)\n  -------------------------- ------- ------ ------- ---- ---------- -------- ----------------- --------------\n  Pathological grade         .375    .257   2.120   1    .145       1.454    .878              2.408\n                                                                                               \n  Gender                     -.111   .270   .168    1    .682       .895     .527              1.520\n                                                                                               \n  Age                        .019    .242   .006    1    .936       1.020    .634              1.638\n                                                                                               \n  Tumor sites                .015    .253   .004    1    .952       1.015    .619              1.666\n                                                                                               \n  Density of CD3+ cells      -.569   .351   2.632   1    .105       .566     .285              1.126\n                                                                                               \n  Density of CD45RO+ cells   -.883   .376   5.524   1    .019       .414     .198              .864\n                                                                                               \n  Level of HMGB1             .099    .300   .110    1    .740       1.105    .613              1.989\n                                                                                               \n  Localization of HMGB1      -.130   .229   .322    1    .570       .878     .560              1.376\n\nDiscussion\n==========\n\nTo determine the role of HMGB1 in the progression of colon cancer, this study examined the patterns of HMGB1 expression and their relationship with the infiltration of T cells and the 5-year survival rate in patients with stage IIIB colon cancer. The results showed that the density of CD45RO+ T cells within the cancer tissue was independently prognostic. More importantly, the co-expression pattern of nuclear and cytoplasmic HMGB1 in cancer cells was inversely related to the infiltration of CD3+ or CD45RO+ T cells and patient 5-year survival rate. These data indicate that the pattern of HMGB1 expression is related to the progression of colon cancer.\n\nThe role of TILs in colorectal cancer has been discussed extensively during the past 3 decades, but controversy continues. Recently, two studies have focused on the role of memory T cells in the progression of colorectal cancer. Pag\u00e8s et al. \\[[@B16]\\] concentrated on the role of TILs in the early metastatic invasion of 959 cases of colorectal cancer and observed increasing levels of TILs, especially CD45RO+ memory cells, in tumors without pathologic signs of early metastatic invasion. The increasing levels of TILs were correlated with a less advanced pathologic stage and increased survival, indicating the potential of CD45RO+ memory cells for limiting early metastasis of colorectal cancer. Galon et al. \\[[@B17]\\] explored the role of the adaptive immune response in controlling growth and recurrence in 454 cases of human colorectal cancer and found that patients with a higher density of CD3+ and CD45RO+ cells in the tumor center and the invasive margins of the tumor samples had prolonged disease-free survival, indicating that the type, density, and location of immune cells within the tumor samples were predictors of survival. Both of these studies revealed the protective role of CD45RO+ memory cells in early colorectal cancer. The present study revealed that the infiltration of CD45RO+ cells was also protective in locally advanced colon cancer. As 72% of CD45RO+ cells were also positive for CD3, the protective CD45RO+ cells in colon cancer should be predominantly CD3+ CD45RO+ memory T cells. Although a higher density of CD3+ T cells was also associated with a favorable prognosis in univariate analysis, this prognostic value was not apparent in multivariate analysis, an observation that is also consistent with a recent study \\[[@B39]\\].\n\nSince the memory T cells residing in colon cancers result from an acute immunologic response, this study further examined whether HMGB1 was involved in this process. A higher level of HMGB1 was associated with a poorer prognosis or more malignant phenotypes in multiple solid tumors, such as nasopharyngeal, lung, skin, hepatic, gastric, and prostate cancers, with the exception of breast cancer \\[[@B40]-[@B46]\\]. Previous studies have shown that the co-expression of RAGE and HMGB1 led to enhanced migration and invasion by colon cancer cell lines. Furthermore, increased RAGE expression in colon cancers has been associated with atypia, adenoma size, and metastasis to other organs. Additionally, stage I tumors have relatively lower expression levels of HMGB1, whereas stage IV tumors have near-universal expression of HMGB1 \\[[@B47]-[@B52]\\]. However, only a few studies have shown the association between HMGB1 expression and 5-year survival in colon cancer patients. Recent clinical evidence showed that the overexpression of HMGB1 in colorectal cancer was related to malignant phenotypes and shorter survival times \\[[@B53]-[@B55]\\]. To decrease the bias from factors like TNM stage and rectal cancers, this study analyzed specimens only from stage IIIB colon cancer patients. The results indicated that the co-expression of nuclear and cytoplasmic HMGB1 in colon cancer cells was associated with a poor prognosis. In normal colon mucosa, the expression of HMGB1 was observed only in the nucleus. In colon cancer cells, the co-expression pattern of HMGB1 in the nucleus and cytoplasm was observed in a subset of cancer cells, whereas nuclear-only expression of HMGB1 existed in most colon cancer cells. The survival curves of patients with the nuclear-only expression pattern and those with a pattern of co-expression crossed at 24 months after surgery, a result that might derive from the relatively small number of cases in this study, as no statistically significant difference was detected. However, Kaplan-Meier survival analysis showed that the patients with the pattern of co-expression of HMGB1 had a lower 5-year survival rate after 24 months of follow-up than those patients with HMGB1 expression in the nucleus only. This effect suggested that the pattern of co-expression of HMGB1 in cancer cells interfered with the long-term survival by partially modulating the immune response in a subset of colon cancer patients.\n\nBecause HMGB1 is a DAMP, the association between the expression of HMGB1 and the infiltration of TILs was further analyzed in this study. The results showed that co-expression of nuclear and cytoplasmic HMGB1 was also inversely related to the infiltration of CD45RO+ cells and CD3+ cells, again arguing that the effect of the pattern of co-expression of HMGB1 on survival was associated with the local modulation of the immunologic response by HMGB1. The mechanism underlying this inverse association is still elusive. One possibility is that HMGB1 secreted by colon cancers promotes the angiogenesis switch and increases cancer cell invasiveness \\[[@B56]-[@B59]\\]. Another potential mechanism is that extracellular HMGB1 triggers the inflammatory cascade, modulates the local immunologic microenvironment towards tolerance by polarizing the response of helper T cells and improperly activating macrophages and DCs \\[[@B60]-[@B64]\\]. There is evidence suggesting that both aforementioned mechanisms might contribute to the progression of colon cancer. For example, in colon cancer cell lines, E-selectin downregulated the cellular expression of HMGB1 but enhanced the release of HMGB1 into the culture medium. The released HMGB1, in turn, activated endothelial cells to express E-selectin \\[[@B65]\\]. Additionally, the higher level of HMGB1 detected in DCs existed in the metastatic lymph nodes of colon cancer patients \\[[@B66]\\]. Therefore, HMGB1-releasing colon cancer cells might promote effective angiogenesis and immune escape.\n\nSeveral lines of evidence indicate that chemotherapy or radiation therapy induce the release of HMGB1 from dying cells, and this damage-related, acute release of HMGB1 is involved in the protective immune response \\[[@B67]-[@B69]\\]. This study observed that the co-expression of HMGB1 in the nuclear and cytoplasm was associated with poorer prognoses and lower infiltration of CD45RO+ cells, suggesting that spontaneous release and damage-induced release of HMGB1 might act differently in the progression of colon cancer. However, This study was just a descriptive one, more functional work should be done to reveal the mechanisms underlying this process. Therefore, different strategies should be considered when targeting HMGB1 under the conditions of spontaneous chronic release and damage-related, acute release of HMGB1 \\[[@B70]-[@B72]\\].\n\nConclusion\n==========\n\nThis study observed that the co-expression of nuclear and cytoplasmic HMGB1 was inversely associated with the infiltration of CD45RO+ cells and the 5-year survival rate in patients with stage IIIB colon cancer. These observations indicated that the distribution patterns of HMGB1 contribute to the progression of colon cancer. The co-expression of nuclear and cytoplasmic HMGB1 might be used as a marker for poor survival in patients with local advanced colon cancer.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nWXJ, DY, ZX, PZZ, WDS, and ZLM carried out the case collection. PRQ, LCY, and YXJ carried out the immunohistochemical staining and flow cytometry analyses. ZXS and ZYX conceived the study, participated in its design and coordination, and helped draft the manuscript. All authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2407/10/496/prepub>\n\nAcknowledgements\n================\n\nThis study was supported by the National Nature Science Foundation (30872931) and the National Nature Science Foundation, China (30972882).\n"}
+{"text": "Introduction {#Sec1}\n============\n\nChromaffin cells are neuroendocrine cells that synthesize and secrete catecholamines in response to sympathetic nervous system stimulation and therefore participate in regulation of stress-modified parameters such as heart rate and blood pressure. A variety of additional agents are costored and released along with the catecholamines from the chromaffin granules, including neuropeptides such as the enkephalins and adenosine triphosphate (ATP) \\[[@CR1]\\]. In addition to being released into the circulation, these agents may regulate chromaffin cell activity in an autocrine or a paracrine manner, allowing the cells to adjust to varying levels of stimulation.\n\nThe role of ATP in chromaffin cell function has not been well defined, even though it is secreted in high concentrations from chromaffin cells \\[[@CR1]\\]. It has been suggested that ATP regulates chromaffin cell secretion, either positively \\[[@CR2]\\] or negatively \\[[@CR3], [@CR4]\\], and that ATP regulates the function of voltage-dependent calcium channels \\[[@CR5]\\]. ATP exerts its effects through either G-protein-coupled receptors, designated P2Y; or ion channels, designated P2X. These receptor types are further divided into subtypes, including P2Y~1~, P2Y~2~, P2Y~4~, P2Y~6~, and P2Y~11--14~ for the G-protein-coupled P2Y receptors and P2X~1--7~ for those that activate ion channels \\[[@CR6]\\]. Previous work with chromaffin cells indicated ATP stimulation results in increases in inositol phosphates \\[[@CR7]\\], cyclic adenosine monophosphate (cAMP) \\[[@CR8]\\], and \\[Ca^2+^\\]~i~ accumulation \\[[@CR7]\\], likely via activation of a P2Y~2~ or P2Y~4~ receptor, both of which are present in bovine chromaffin cells (unpublished observations). The downstream effects of P2Y receptor stimulation by ATP in chromaffin cells are not known. The observed increases in signaling messengers may bring about the activation of multiple protein kinases or tyrosine kinases.\n\nIn several cell types, ATP signaling has been shown to activate extracellular signal-regulated kinase 1 and 2 (ERK1/2). P2Y receptors have been shown to couple to ERK1/2 activation via activation of protein kinases such as phosphoinositide-3 kinase (PI3K) \\[[@CR11]\\] or protein kinase C (PKC) \\[[@CR10], [@CR11]\\]. P2Y receptor-mediated ERK1/2 has also been shown to be dependent on activation of tyrosine kinases such as Src or proline-rich tyrosine kinase (Pyk2) \\[[@CR9], [@CR10]\\] and/or on transactivation of the epidermal growth factor receptor (EGFR) \\[[@CR9]\\]. Additionally, increases in cAMP in response to ATP may result in activation of protein kinase A (PKA), which has been shown to activate ERK1/2 \\[[@CR12]\\]. Increases in \\[Ca^2+^\\]~i~ in response to ATP may result in activation of Ca^2+^/calmodulin-dependent protein kinase II (CaMKII), which has also been shown to phosphorylate ERK1/2 \\[[@CR13]\\]. Therefore, we elected to examine ATP-mediated ERK1/2 phosphorylation and the receptor subtype and signaling mechanism present in bovine chromaffin cells.\n\nMaterials and methods {#Sec2}\n=====================\n\n**Chromaffin cell isolation and cell culture** Bovine adrenal chromaffin cells (BACC) were isolated using a collagenase perfusion method as described previously \\[[@CR14], [@CR15]\\]. Cells were maintained on six-well plates at a density of 3\u2009\u00d7\u200910^6^ cells/well at 37\u00b0C with 5% CO~2~. Viability and purity were verified to be \\>95% by Trypan blue exclusion and neutral red staining, respectively.\n\n**Western blot analysis** Cells were incubated with agonist for 10 min, and inhibitor preincubations were 15 min, unless otherwise indicated. Cells were rinsed with phosphate buffered saline (PBS) and lysed with 200\u00a0\u03bcl of 50\u00a0mM N-2-hydroxyethylpiperazine-N\u2032-2-ethanesulfonic acid (HEPES) (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), pH 7.2, containing 1\u00a0mM ethylenediaminetetraacetate (EDTA), 1\u00a0mM ethyleneglycoltetraacetic acid (EGTA), 0.2% triton X-100, 10\u00a0mM \u03b2-glycerol 2-phosphate disodium salt, 1\u00a0mM sodium orthovanadate, 1\u00a0mM benzamidine, 4\u00a0\u03bcg/ml leupeptin, 1\u00a0\u03bcM microcystin-LR, and 0.5\u00a0mM DTT. Lysates were centrifuged at 13,000 g, and protein concentrations of supernatants were determined with the Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA). Loading buffer (2X) was added to cell lysates, consisting of 125\u00a0mM Tris(hydroxymethyl)aminomethane HCl (Research Organics, Cleveland, OH, USA), 4% sodium dodecylsulfate (SDS) (Research Organics, Cleveland, OH, USA), 20% glycerol, and 0.02% bromophenol blue. Samples were subsequently boiled and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) on 10% Tris-HCl Criterion gels (BioRad, Hercules, CA, USA).Protein was transferred to fluorescent-polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). Membranes were next blocked in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE, USA) for 1 h, then incubated in Odyssey blocking buffer containing 0.2% Tween 20 (polyoxyethylene-sorbitan monolaurate) and mouse-anti-phospho-ERK 1/2 (Thr^202^/Tyr^204^) and rabbit-anti-ERK 1/2 (Cell Signaling Technology, Boston, MA, USA) primary antibodies. Blots were then incubated in Odyssey blocking buffer with 0.2% Tween 20 and 0.02% SDS and with the secondary antibodies goat-anti-mouse immunoglobulin (Ig)G Alexa Fluor680 (Molecular Probes, Eugene, OR, USA) and goat-anti-rabbit IgG IR800 (Rockland Immunochemicals, Gilbertsville, PA, USA). Membranes were developed with the Odyssey Infrared Imaging System, which utilizes two infrared channels (700\u00a0nm and 800\u00a0nm), allowing for detection of two target proteins simultaneously, in this case phosphorylated and total ERK1/2.\n\n**Statistics** Band integrated intensities were determined with Odyssey Imaging software. Phospho-ERK1/2 intensities were divided by total ERK1/2 intensities and normalized to fold increases over control. Data were analyzed with GraphPad Prism software, and one-way analysis of variance (ANOVA) was utilized to determine statistical significance. The decision was made to utilize ERK2 for graphical representations, as band intensities for phosphorylated ERK2 were stronger than phosphorylated ERK1, though results obtained were quantitatively similar for both.\n\n**Chemicals** Nucleotides and analogs were obtained from Sigma-Aldrich (St. Louis, MO, USA). PD98059 (2\u2032-amino-3\u2032-methoxylflavone), PMA (phorbol 12-myristate 13-acetate), suramin, and reactive blue 2 (RB2) were also obtained from Sigma-Aldrich. NF279 \\[8,8\u2032-\\[Carbonylbis(imino-4,1-phenylenecarbonylimino-4,1-phenylenecarbonylimino)\\]bis-1,3,5-naphthalenetrisulfonic acid hexasodium salt\\] was purchased from Tocris (St. Louis, MO, USA). KT5720 \\[(9S,10S,12R)-2,3,9,10,11,12-Hexahydro-10-hydroxy-9-methyl-1-oxo-9,12-epoxy-1H-diindolo\\[1,2,3-fg:3\u2032,2\u2032,1\u2032-kl\\]pyrrolo\\[3,4-i\\]\\[1,6\\]benzodiazocine-10-carboxylic acid hexyl ester\\], Ro-31--8220 \\[2-\\[1-(3-(Amidinothio)propyl)-1H-indol-3-yl\\]-3-(1-methylindol-3-yl)maleimide methanesulfonate\\], LY294002 \\]2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one\\], PP1 \\[4-amino-5-(4-methylphenyl)-7-(*t*-butyl)pyrazolo-D-3,4-pyrimidine\\], PP2 \\[4-amino-5-(4-chlorophenyl)-7-(*t*-butyl)pyrazolo-D-3,4-pyrimidine\\], AG1478 \\[4-(3-Chloroanilino)-6,7-dimethoxyquinazoline\\], and GM6001 \\[N-\\[(2R)-2-(hydroxamidocarbonylmethyl)-4-methylpentanoyl\\]-L-tryptophan methylamide\\] were purchased from Biomol (Philadelphia, PA, USA). Bis-I (bisindolylmaleimide-I), KN-92 \\[2-\\[N-(4-methoxybenzenesulfonyl)\\]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine\\], KN-93 \\[2-\\[N-(2-hydroxyethyl)\\]-N-(4-methoxybenzenesulfonyl)\\]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine\\], and wortmannin were purchased from Merck-Calbiochem Biosciences (Darmstadt, Germany). H89 \\[N-\\[2-(p-Bromocinnamylamino)ethyl\\]-5-isoquinolinesulfonamide\\] was obtained from Upstate (Charlottesville, VA, USA).Other reagents were obtained from either Sigma-Aldrich or Fisher Scientific.\n\nResults {#Sec3}\n=======\n\nWestern blot analyses with antibodies specific to activated ERK1/2 phosphorylated at Thr^202^/Tyr^204^ were utilized to examine the time course of ERK1/2 phosphorylation in response to nucleotide stimulation. Both ATP and UTP potently increased ERK1/2 phosphorylation, with a peak between 5 and 15 min (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). ERK1/2 has been shown to be catalyzed only by MEK; therefore, the MEK inhibitor PD98059 was utilized to confirm the immediate upstream signaling event responsible for nucleotide-mediated ERK1/2 phosphorylation. PD98059 (10\u00a0\u03bcM) significantly decreased both ATP-mediated (\u223c90%) and UTP-mediated (\u223c70%, Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}, Table\u00a0[1](#Tab1){ref-type=\"table\"}) ERK1/2 phosphorylation. Fig.\u00a01Nucleotide-stimulated ERK1/2 phosphorylation is time dependent. BACCs were treated with ATP (100\u00a0\u03bcM) or UTP (100\u00a0\u03bcM) for 2 to 30 min. Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa, and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Points* on the graph represent mean \u00b1 standard error of the meanFig.\u00a02MEK inhibition decreases ATP- and UTP-mediated ERK1/2 phosphorylation. BACCs were treated with PD98059 (10\u00a0\u03bcM) or dimethylsulfoxide for 15 min, followed by a 10-min stimulation with ATP (100\u00a0\u03bcM) or UTP (100\u00a0\u03bcM). Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Bars* on the graph represent mean \u00b1 standard error of the mean. \\*\\*\\* *p*\u2009\\<\u20090.001 vs. stimulator aloneTable\u00a01Involvement of signaling pathways in ATP- and UTP-mediated ERK1/2 phosphorylationInhibitor/antagonistATPUTPERK1ERK2ERK1ERK2P2RSuramin (100\u00a0\u03bcM)61.8\u2009\u00b1\u20093.1\\*\\*\\*54.7\u2009\u00b1\u20093.8\\*\\*\\*62.0\u2009\u00b1\u20095.7\\*\\*\\*62.0\u2009\u00b1\u20095.7\\*\\*\\*AntagonistsRB2 (100\u00a0\u03bcM)32.3\u2009\u00b1\u20094.5\\*\\*\\*29.5\u2009\u00b1\u20094.5\\*\\*\\*44.9\u2009\u00b1\u200910.7\\*\\*\\*44.9\u2009\u00b1\u20099.1\\*\\*\\*MEK inhibitorPD98059 (10\u00a0\u03bcM)95.3\u2009\u00b1\u20095.3\\*\\*\\*94.4\u2009\u00b1\u20094.5\\*\\*\\*69.5\u2009\u00b1\u20093.4\\*\\*\\*64.3\u2009\u00b1\u20094.2\\*\\*\\*PKC inhibitorsBis-I (3.5\u00a0\u03bcM)12.7\u2009\u00b1\u200920.54.6\u2009\u00b1\u20097.95.2\u2009\u00b1\u200912.9\u22121.3\u2009\u00b1\u200910.4Bis-I-PMA76.1\u2009\u00b1\u20092.9\\*\\*\\*67.5\u2009\u00b1\u20093.2\\*\\*\\*----Ro-31-8220 (10\u00a0\u03bcM)\u221240.0\u2009\u00b1\u200928.8\u221213.5\u2009\u00b1\u200910.2\u221213.4\u2009\u00b1\u200914.9\u22128.7\u2009\u00b1\u200910.9Ro-31-8220-PMA66.6\u2009\u00b1\u20092.6\\*\\*\\*54.4\u2009\u00b1\u20092.3\\*\\*\\*----PKA inhibitorsH89 (10\u00a0\u03bcM)\u221267.1\u2009\u00b1\u200936.3\u221242.4\u2009\u00b1\u200914.4\u221227.5\u2009\u00b1\u200931.4\u221216.4\u2009\u00b1\u200916.5KT5720 (100\u00a0nM)\u22121.5\u2009\u00b1\u200913.3\u22121.9\u2009\u00b1\u200916.51.9\u2009\u00b1\u200913.96.5\u2009\u00b1\u200913.2CaMKII inhibitorKN93 (1\u00a0\u03bcM)22.5\u2009\u00b1\u200927.27.1\u2009\u00b1\u200912.7\u221220.1\u2009\u00b1\u200918.2\u221214.2\u2009\u00b1\u200915.4PI3K inhibitorsWortmannin(300\u00a0nM)\u221210.0\u2009\u00b1\u200934.2\u22123.9\u2009\u00b1\u200915.710.0\u2009\u00b1\u200925.912.5\u2009\u00b1\u200914.6LY294002 (20\u00a0\u03bcM)\u22123.9\u2009\u00b1\u200915.71.0\u2009\u00b1\u200917.418.6\u2009\u00b1\u200918.730.3\u2009\u00b1\u200912.1Src inhibitorPP2 (1\u00a0\u03bcM)40.9\u2009\u00b1\u200916.1\\*39.5\u2009\u00b1\u20099.1\\*\\*\\*46.0\u2009\u00b1\u20098.1\\*\\*\\*32.4\u2009\u00b1\u20099.6\\*\\*\\*EGFR inhibitorsAG1478 (2.6\u00a0\u03bcM)73.7\u2009\u00b1\u20092.0\\*\\*\\*71.1\u2009\u00b1\u20092.3\\*\\*\\*62.9\u2009\u00b1\u20097.2\\*\\*\\*69.6\u2009\u00b1\u20096.7\\*\\*\\*MMP inhibitorsGM6001 (2.5\u00a0\u03bcM)69.2\u2009\u00b1\u20095.8\\*\\*60.8\u2009\u00b1\u20097.1\\*\\*\\*66.7\u2009\u00b1\u200910.1\\*\\*\\*62.0\u2009\u00b1\u200910.4\\*\\*\\*Cells were pretreated with inhibitors for 15\u00a0min then stimulated with 100\u00a0\u03bcM UTP (or 1\u00a0\u03bcM PMA) for 10\u00a0min. Values are percent inhibition of UTP-mediated ERK1/2 phosphorylation \u00b1 standard error of the mean for three experiments in triplicate. Bis-I-PMA and Ro-31-8220-PMA refer to using PMA as the stimulator rather than UTP\\**p*\u2009\\<\u20090.05 vs. stimulator alone,\\*\\**p*\u2009\\<\u20090.01 vs. stimulator alone, \\*\\**p*\u2009\\<\u20090.001 vs. stimulator alone\n\nMost studies designed to determine purinergic receptor subtypes use ligand potency studies due to the lack of available highly selective receptor subtype antagonists \\[[@CR16]\\]. Correspondingly, we used a similar approach to characterize the receptor subtype involved in ERK1/2 phosphorylation. Examination of several purine analogs revealed a rank order of potency of UTP (EC~50~\u2009=\u20091.6\u00a0\u03bcM)\\>ATP\u03b3S (6.5\u00a0\u03bcM)\u2265ATP (13\u00a0\u03bcM)\\>uridine diphosphate (UDP) (120\u00a0\u03bcM) = adenosine diphosphate (ADP) (220\u00a0\u03bcM)\u2009= 2-methylthio (ATP)\u20092-MeSATP (320\u00a0\u03bcM)\\>\\>\u03b1,\u03b2-methylene ATP (\u03b1,\u03b2-meATP) (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}), consistent with the involvement of a P2Y~2~ or P2Y~4~ receptor, as both UTP and ATP exhibit strong agonist action. Fig.\u00a03P2Y~2~ or P2Y~4~ receptor activation increases ERK1/2 phosphorylation. BACCs were treated with increasing concentrations of nucleotides and analogs for 10 min. Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa, and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Points* on the graph represent mean \u00b1 standard error of the mean. *Con* (control) refers to results obtained with unstimulated cells\n\nThe involvement of a P2 receptor in ERK1/2 phosphorylation was further supported using the nonselective P2 receptor antagonists suramin and RB2. Suramin (100\u00a0\u03bcM) significantly decreased ATP- or UTP-mediated ERK1/2 phosphorylation (\u223c60%, Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, Table\u00a0[1](#Tab1){ref-type=\"table\"}). RB2 (100\u00a0\u03bcM) also decreased the effect of ATP- or UTP-stimulation on ERK1/2 phosphorylation (\u223c35%, Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, Table\u00a0[1](#Tab1){ref-type=\"table\"}). The P2X-specific receptor agonist \u03b1,\u03b2-meATP had no effect on ERK1/2 phosphorylation at concentrations up to 100\u00a0\u03bcM (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}), eliminating the involvement of several of the P2X receptor subtypes. Moreover, UTP is selective for P2Y receptors, precluding the involvement of a P2X receptor in nucleotide-mediated ERK1/2 phosphorylation. Fig.\u00a04P2 receptor antagonists partially block ATP- and UTP-mediated ERK1/2 phosphorylation. BACCs were pretreated with or without suramin (100\u00a0\u03bcM) or reactive blue 2 (RB2, 100\u00a0\u03bcM) for 15 min, followed by a 10-min stimulation with or without ATP (100\u00a0\u03bcM) or UTP (100\u00a0\u03bcM). Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa, and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Bars* on graph represent mean \u00b1 standard error of the mean. \\*\\*\\* *p*\u2009\\<\u20090.001 vs. stimulator alone\n\nATP stimulation of chromaffin cells has been shown to increase inositol phosphates \\[[@CR7]\\], cAMP \\[[@CR8]\\], and \\[Ca^2+^\\]~i~ accumulation \\[[@CR7]\\], which may lead to activation of several protein kinases, which may in turn be involved in ATP-mediated phosphorylation of ERK1/2. Therefore, a variety of inhibitors were used to examine which kinases are involved in ATP-mediated ERK1/2 phosphorylation. The PKA inhibitors H89 (10\u00a0\u03bcM) and KT5720 (100\u00a0nM) had no effect on ATP- or UTP-mediated ERK1/2 phosphorylation (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The broad-spectrum PKC inhibitors Ro-31--8220 (10\u00a0\u03bcM) and Bis-I (3.5\u00a0\u03bcM) also had no effect on ATP- or UTP-mediated ERK1/2 phosphorylation, although they were capable of blocking PMA-mediated ERK1/2 phosphorylation (\u223c60%, Table\u00a0[1](#Tab1){ref-type=\"table\"}). Moreover, the CaMKII inhibitor KN93 (1\u00a0\u03bcM) and the PI3K inhibitors wortmannin (300\u00a0nM) and LY294002 (20\u00a0\u03bcM) had no effect on ATP- or UTP-mediated ERK1/2 phosphorylation (Table\u00a0[1](#Tab1){ref-type=\"table\"}).\n\nIn addition to activation of PKs, P2Y receptors may utilize tyrosine kinases to activate ERK1/2 \\[[@CR10]\\]. Therefore, the tyrosine kinase inhibitor PP2 was used to examine the role of Src family members in ATP-mediated ERK1/2 phosphorylation. PP2 (1\u00a0\u03bcM) decreased ATP- or UTP-mediated ERK2 phosphorylation (\u223c40%, Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}, Table\u00a0[1](#Tab1){ref-type=\"table\"}). In addition to inhibiting Src family members, PP2 is also a weak inhibitor of the EGFR. Also, G-protein-coupled receptor activation of Src family members may result in activation of ERK1/2 via activation of Ras \\[[@CR17], [@CR18]\\] or via transactivation of the EGFR. Therefore, the involvement of EGFR in ATP-mediated ERK1/2 phosphorylation was determined by treating cells with the EGFR inhibitor AG1478 (2.6\u00a0\u03bcM), which decreased ATP- and UTP-mediated ERK2 phosphorylation by about 70% (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}a, Table\u00a0[1](#Tab1){ref-type=\"table\"}). EGF-mediated ERK1/2 phosphorylation was completely blocked by AG1478 (100%, *p*\u2009\\<\u20090.001, data not shown). Transactivation of the EGFR by G-protein-coupled receptors may be mediated by tyrosine kinases such as Src or via metalloproteinases, which release EGFR ligands such as heparin-binding EGF-like growth factor (HB-EGF) from the cell membrane \\[[@CR19]\\]. The metalloproteinase inhibitor GM6001 (2.5\u00a0\u03bcM) decreased ATP- and UTP-mediated ERK1/2 phosphorylation by about 65% (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}b, Table\u00a0[1](#Tab1){ref-type=\"table\"}). Fig.\u00a05Tyrosine kinase inhibition decreases ATP- and UTP-mediated ERK1/2 phosphorylation. BACCs were pretreated with PP2 (1\u00a0\u03bcM) or dimethylsulfoxide (DMSO) for 15 min, followed by a 10-min stimulation with or without ATP (100\u00a0\u03bcM) or UTP (100\u00a0\u03bcM). Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa, and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Bars* on the graphs represent mean \u00b1 standard error of the mean. \\*\\*\\* *p*\u2009\\<\u20090.001 vs. stimulator aloneFig.\u00a06Nucleotide-mediated ERK1/2 phosphorylation is dependent on EGFR transactivation. BACCs were pretreated with or without AG1478 (2.6\u00a0\u03bcM) (**a**), GM6001 (2.5\u00a0\u03bcM) (**b**) and dimethylsulfoxide (DMSO) for 15 min, followed by a 10-min stimulation with or without ATP (100\u00a0\u03bcM) or UTP (100\u00a0\u03bcM). Blots are representative of three independent experiments performed in triplicate (*n*\u2009=\u20093); the *upper band* (phosphorylated or nonphosphorylated) is ERK1\u2009=\u200944\u00a0kDa ,and the *lower band* is ERK2\u2009=\u200942\u00a0kDa. Blot intensities were measured with the Odyssey Imaging System; values are phosphorylated ERK2 intensity divided by total ERK2 intensity. *Bars* on the graphs represent mean \u00b1 standard error of the mean. \\*\\*\\* *p*\u2009\\<\u20090.001 vs. stimulator alone\n\nDiscussion {#Sec4}\n==========\n\nATP and UTP potently increase ERK1/2 phosphorylation, with a peak between 5 and 15 min. This rapid peak in ERK1/2 phosphorylation in response to ATP would allow the cells to respond quickly to varying levels of stimulation. Although the physiological effects of ERK1/2 phosphorylation in these cells are unknown, possible actions requiring a rapid response include either the acute activation of proteins involved in catecholamine secretion and/or stimulation of protein expression important for exocytosis.\n\nLigand potency and inhibitor studies suggest either the P2Y~2~ or P2Y~4~ receptor subtype is responsible for nucleotide-mediated ERK1/2 phosphorylation, similar to data obtained for increases in inositol phosphates (unpublished observations). Both of these receptor subtypes are present in chromaffin cells, based on reverse transcriptase real-time polymerase chain reaction (PCR) data for P2Y~2~ and P2Y~4~, and appear to be expressed in these cells according to Western blot analysis with specific antibodies (unpublished observations.) Several lines of evidence suggest P2X receptors are not involved in the increase in ERK1/2 in response to nucleotide stimulation. First, UTP does not activate P2X ion channels but potently increases ERK1/2 phosphorylation. Additionally, \u03b1,\u03b2meATP, an agonist selective for several P2X receptor subtypes, had no effect on ERK1/2 phosphorylation. The P2Y receptor involved is most likely either P2Y~2~ or P2Y~4~. UTP is highly selective for two P2Y receptor subtypes P2Y~2~ and P2Y~4~ and weakly effective on the P2Y~6~ receptor. The P2Y~6~ subtype can be ruled out because of the subtypes activated by UTP; only the P2Y~2~ and P2Y~4~ subtypes are also strongly activated by ATP. The weak effect of ADP, UDP and 2-MeSATP confirms this designation, as these agonists are specific for P2 receptor subtypes other than the P2Y~2~ or P2Y~4~ subtypes \\[[@CR6]\\]. There are no available agonists or antagonists to distinguish between the P2Y~2~ and P2Y~4~ receptors. Even so, suramin and RB2 are commonly used to characterize these receptors in a given cell type, and their partial effectiveness is not contradictory to results found in other cell types for P2Y~2~ or P2Y~4~ receptors \\[[@CR20]\\].\n\nNucleotides utilize multiple signaling pathways in different cell types to bring about increases in ERK1/2 phosphorylation. PKC and PI3K have been implicated in P2Y-mediated ERK1/2 phosphorylation \\[[@CR10], [@CR11]\\]. In PC12 cells, ERK1/2 phosphorylation in response to P2Y~2~ receptor activation has been shown to be both dependent \\[[@CR9], [@CR10]\\] and independent \\[[@CR21]\\] of the small tyrosine kinase Pyk2. P2Y~2~ receptors have also been shown to contain an integrin-binding domain, arginine-glycine-aspartic acid (RGD), which is necessary for ERK1/2 activation in astrocytes \\[[@CR22], [@CR23]\\]. Additionally, P2Y~2~ receptors contain SH3-binding sites that associate with Src in astrocytoma cells \\[[@CR24]\\] and astrocytes \\[[@CR23]\\]. Also, in PC12 cells P2Y~2~ receptors have been shown to require EGFR transactivation to increase ERK1/2 phosphorylation \\[[@CR9]\\].\n\nInitially, we examined whether signaling pathways mediated by protein kinases were involved in ATP-mediated ERK1/2 phosphorylation. Previous studies determined that ATP-mediated stimulation of bovine chromaffin cells results in increases in inositol phosphates \\[[@CR7]\\], cAMP \\[[@CR8]\\], and \\[Ca^2+^\\]~i~ accumulation \\[[@CR7]\\]. P2Y~2~ or P2Y~4~ receptors couple to G~q~ to increase activation of PKC. Additionally, the observed increases in cAMP may result in activation of PKA, whereas increased \\[Ca^2+^\\]~i~ may result in activation of CaMKII. However, inhibitors of each of these protein kinases had no effect on ATP- or UTP-mediated ERK1/2 phosphorylation, including the PKA inhibitors H89 and KT5720, PKC inhibitors Bis-I and Ro-81--3220, or the CaMKII inhibitor KN93. The PKC inhibitors Bis-I and Ro-81--3220 were capable of blocking PMA-mediated ERK1/2 phosphorylation, suggesting PKC can couple to ERK1/2 phosphorylation in these cells, and yet confirming the lack of involvement of this pathway in ATP-mediated ERK1/2 phosphorylation. We also examined the PI3K inhibitors wortmannin and LY294002, as this kinase has been implicated in P2Y-mediated ERK1/2 phosphorylation. These inhibitors also proved to be ineffective, suggesting the protein kinases examined were not responsible for ATP-mediated ERK1/2 phosphorylation.\n\nWe next examined the involvement of tyrosine kinases in ATP-mediated ERK1/2 phosphorylation, as these kinases have been shown to be involved in P2Y-mediated ERK1/2 phosphorylation. PP2 (1\u00a0\u03bcM) significantly decreased ATP- and UTP-mediated ERK1/2 phosphorylation. The reported IC~50~s for PP2-mediated inhibition of Src family members are in the nanomolar range (<http://www.biomol.com>), whereas the dose used in these studies is reported to cause weak inhibition of the EGFR \\[[@CR25]\\]. Lower doses of PP2 had no effect on ATP- or UTP-mediated ERK1/2 phosphorylation (data not shown.) Therefore, at the effective dose used, it is not possible to conclude whether Src or EGFR inhibition was responsible for the decrease in ATP-mediated ERK1/2 phosphorylation. Additionally, activation of Src family members by G-protein-coupled receptors may increase ERK1/2 phosphorylation via activation of the renin angiotensin system (Ras) \\[[@CR17], [@CR18]\\] or via transactivation of the EGFR \\[[@CR26]\\]. Thus, the involvement of EGFR in ATP-mediated ERK1/2 phosphorylation was investigated further. Inhibition of the EGFR with the specific inhibitor AG1478 decreased ATP- and UTP-mediated ERK1/2 phosphorylation, strongly suggesting EGFR transactivation is important for ATP-mediated ERK1/2 phosphorylation.\n\nG-protein-coupled receptors may transactivate the EGFR via activation of tyrosine kinases such as Src, or via activation of metalloproteinases to generate EGFR ligands such as HB-EGF \\[[@CR19]\\]. The role of metalloproteinases in ATP-mediated ERK1/2 phosphorylation was investigated with broad-spectrum inhibitor GM6001, which significantly decreased the response, suggesting that HB-EGF cleavage mediated by metalloproteinases in response to ATP may be responsible for transactivation of the EGFR and subsequently stimulation of ERK1/2 phosphorylation.\n\nAs expected, the MEK inhibitor PD98059 blocked ATP- and UTP-mediated ERK1/2 phosphorylation, confirming that MEK, the only known kinase upstream of ERK1/2, contributes to ERK1/2 phosphorylation. PD98059 blocked ATP-mediated ERK1/2 phosphorylation (\u223c90%) to a greater extent than UTP-mediated ERK1/2 phosphorylation (\u223c70%). This may be due to the fact that UTP is a more potent agonist and elicited a larger response than ATP for ERK1/2 phosphorylation. For the other inhibitors, the responses were very similar; however, none of the other inhibitors had such a pronounced effect on ATP-mediated ERK1/2 phosphorylation. Alternatively, ATP and UTP may activate multiple receptors with distinct signaling pathways that are variously more specific for ATP or UTP. Activation of ERK1/2 independent of MEK1 may involve inhibition of phosphatases.\n\nFurther studies are necessary to determine the consequence(s) of ERK1/2 phosphorylation in response to ATP stimulation of chromaffin cells. As chromaffin cells are nonproliferating, the stimulation of ERK1/2 phosphorylation by ATP and UTP may couple to regulation of gene transcription essential to exocytosis.\n\nTo our knowledge this is the first study demonstrating phosphorylation of ERK1/2 in response to ATP or UTP stimulation in bovine chromaffin cells. Our data show that the ERK1/2 phosphorylation response to ATP is mediated by either a P2Y~2~ or P2Y~4~ receptor. Protein kinases are not involved in nucleotide-mediated ERK1/2 phosphorylation, but rather, metalloproteinase-dependent transactivation of the EGFR is necessary for ATP-mediated ERK1/2 phosphorylation.\n\nWe thank Matthew Beaver for excellent technical assistance and Robin Taylor for expert assistance with graphics. This work was supported by a grant to Terry D. Hexum from the American Heart Association (\\#0550030Z).\n\nBACC\n\n:   bovine adrenal chromaffin cells\n\nERK1/2\n\n:   extracellular signal-regulated kinase 1 and 2\n\nEGF\n\n:   epidermal growth factor\n\nEGFR\n\n:   EGF receptor\n\nMEK\n\n:   mitogen-activated protein kinase/ERK kinase\n\n\\[Ca^2+^\\]~i~\n\n:   cytosolic free Ca^2+^ concentration\n\nPKC\n\n:   protein kinase C\n\nPKA\n\n:   protein kinase A\n\nCaMKII\n\n:   Ca^2+^/calmodulin-dependent protein kinase II\n\nPI3K\n\n:   phosphoinositide-3 kinase\n\n2-MeSATP\n\n:   2-methylthio ATP\n\n\u03b1,\u03b2-meATP\n\n:   \u03b1,\u03b2-methylene ATP\n\nHB-EGF\n\n:   heparin-binding EGF-like growth factor\n\nPyk2\n\n:   proline-rich tyrosine kinase\n\nSH3\n\n:   Src homology 3\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nCardiovascular diseases (CVD) constitute the major risk of morbidity and mortality in chronic kidney disease (CKD) patients \\[[@B1], [@B2]\\]. Uremic patients have both traditional cardiovascular (CV) risk factors (i.e., old age, hypertension, diabetes, smoking, dyslipidemia, heart failure, and physical inactivity) and nontraditional CV risk factors, including malnutrition, anemia, hyperhomocysteinemia, neuropathy, hyperparathyroidism, and chronic inflammation \\[[@B3]--[@B5]\\]. Patients with end-stage renal disease (ESRD) undergoing renal replacement therapy (RRT), either hemodialysis (HD) or peritoneal dialysis (PD), may face a partial loss of some low-molecular-weight plasma factors (i.e., vitamins A, C, and E) \\[[@B6], [@B7]\\] that normally contrast inflammation by neutralizing reactive oxygen species (ROS) \\[[@B8]\\]. Indeed, the latter are increased during the two therapies \\[[@B9]\\]. The imbalance in antioxidant and pro-oxidant factors generates an oxidative stress (OS) that increases the inflammatory state already present in these patients.\n\nIn recent years, OS has been postulated to be an important risk factor for CVD \\[[@B10]\\]. OS results from an imbalance between prooxidant and antioxidant defence and mechanisms with increased levels of prooxidants leading to tissue damage \\[[@B10]\\]. Antioxidants can be divided into intracellular and extracellular antioxidants. Intracellular enzymatic antioxidants are superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GSH-GPx), which convert substrates (superoxide anion radicals and hydrogen peroxide) to less reactive forms. Various extracellular antioxidants, such as reduced glutathione (GSH), bilirubin, uric acid, and iron (Fe), prevent free radical (FR) reaction by sequestering transition metal ions by chelation in plasma \\[[@B10]\\].\n\nMany studies have demonstrated increased OS in patients with CKD, including accumulation of reactive carbonyl compounds as markers of elevated protein peroxidation \\[[@B10]--[@B12]\\] and increased concentrations of thiobarbituric acid reactive substances (TBARS), malondialdehyde (MDA), and Hydroperoxides (LPO) as markers of high lipid peroxidation \\[[@B13], [@B14]\\].\n\nOS is particularly detrimental in patients receiving HD after each dialysis session, due to the contact of blood with dialysis membrane, facing a chronic deficit in antioxidant defense system \\[[@B12]\\].\n\nThus, this study was undertaken in order to determine the effect of different stages of CKD, hemodialysis, and peritoneal dialysis on lipid peroxidation, protein oxidation, and antioxidant defence.\n\n2. Subjects and Methods {#sec2}\n=======================\n\n2.1. Patients {#sec2.1}\n-------------\n\nThe study was carried out in 167 patients with CKD. Patients were divided into 6 groups according to the classification of the National Kidney Foundation (NKF) \\[[@B15]\\]. They included 28 patients with GFR equal to 86.28\u2009mL/min (CKD 1), 28 patients with GFR equal to 46\u2009mL/min (CKD 2), 28 patients with GFR equal to 20.31\u2009mL/min (CKD 3), 18 patients with GFR equal to 10.75\u2009mL/min (CKD 4), 40 hemodialysis (HD) patients, and 25 peritoneal dialysis (PD) patients ([Table 1](#tab1){ref-type=\"table\"}).\n\nWe excluded from our study patients with clinical signs of infection (hepatitis B and hepatitis C), malignancy, active immunological diseases, and immunosuppressive or immunomodulatory and anti-inflammatory drugs (i.e., conditions that conceal cytokine release). We also excluded patients with diabetes and nephrotic syndrome because diabetes could induce an inflammatory response and rise OS. None of the patients were taking lipid-lowering drugs or antioxidant supplements. Patients take antihypertensive drugs, calcium, vitamin D, and erythropoietin.\n\nThe etiology of CKD in our study included hypertension (66%), cystic kidney disease (14%), glomerulonephritis (6%), prostatic obstruction (3%), and pulmonary tuberculosis treatment (1%) and 10% were unknown.\n\nWe calculate the creatinine clearance from the serum creatinine and through the following estimation formula: the Cockroft and Gault \\[[@B16]\\] formula \\[GFR (1/4) (140 \u2212 age) BW \u00d7 1.23/creatinin\\]. In women, this value was multiplied by 0.85. HD patients were on standard bicarbonate using polysulfone membrane. Patients were dialyzed from 14 to 109 months, three times a week, and each session lasting 4\u2009h. PD patients were in dialysis from 05 to 49 months, using a standard procedure (four exchanges: three isotonic 1.36% glucose solutions, then a hypertonic one at 3.86% glucose). We note that the patients\\' nutrition contains small amounts of protein and phosphate.\n\nAll patients were treated at the nephrology ward of the University Hospital of Oran. The purpose of this study was explained to the subjects and the investigation was carried out with their consent. The experimental protocol was approved by the Committee for Research on Human Subjects of Oran.\n\n2.2. Assays {#sec2.2}\n-----------\n\nIn all patients, blood samples were drawn after 12-hour overnight fast from antecubital venipuncture in uremic and PD patients and by the dialysis fistula in HD patients. We used vacutainer tubes with different anticoagulants during the blood samples. Tubes containing the lithium heparin are used for biochemical experiments, and those containing ethylene diamine tetraacetic acid (EDTA) (8%) are used for the hematology analysis. We collected plasma according to the technique low speed centrifugation at 3000\u2009\u00d7g at 4\u00b0C for 15\u2009min. The fresh plasma was removed, aliquoted, and stored at \u221220\u00b0C.\n\n### 2.2.1. Lipid and Protein Peroxidation {#sec2.2.1}\n\nLipid peroxidation was estimated by measuring thiobarbituric acid reactive substances (TBARS) and hydroperoxides (LPO). TBARS concentrations were measured according to the method of Quintanilha et al. \\[[@B17]\\], using tetramethoxypropane (Prolabo) as precursor of malondialdehyde (MDA). One milliliter of diluted plasma (protein concentration about 2\u2009mg/mL) was added to 2\u2009mL of thiobarbituric acid (final concentration, 0.017\u2009mmol/L) and butylated hydroxytoluene (concentration, 3.36\u2009mmol/L) and incubated for 15\u2009min at 100\u00b0C. After cooling and centrifugation, the absorbance of supernatant was measured at 535\u2009nm. Data were expressed as mmol of TBARS produced/mL of sample.\n\nThe plasma was also assayed to determine LPO with an assay kit from Cayman Chemical (cat. no. 705003); the rate of increase in the absorbance at 500\u2009nm is directly proportional to the LPO produced.\n\nOxidized proteins were estimated by measuring carbonyls concentrations. The latter were analyzed in plasma with an assay kit from Cayman Chemical (cat. no. 10005020); the absorbance of the samples was measured between 360 and 385\u2009nm by a plate reader.\n\n### 2.2.2. Antioxidant Measurements {#sec2.2.2}\n\nActivity of antioxidant enzymes was measured in erythrocyte. Superoxide dismutase (SOD; EC 1.15.1.1) activity was determined with Sigma Chemical kits (cat. no. 19160) at 450\u2009nm by measuring the dismutation of superoxide radicals generated by xanthine oxidase and hypoxanthine. One unit was defined as the amount of enzymes necessary to produce 50% inhibition in the rate of p-iodonitrotetrazolium reduction.\n\nCatalase (CAT; EC 1.11.1.6; 2H~2~O~2~ oxidoreductase) activity was measured with an assay kit from Cayman Chemical (cat. no. 707002). CAT is involved in the detoxification of hydrogen peroxide (H~2~O~2~). CAT enzyme activity could be determined by using the peroxidatic function of CAT at 540\u2009nm.\n\nGlutathione peroxidase (GSH-Px; EC 1.11.1.9) enzyme activity was measured with an assay kit from Sigma Chemical (cat. no. CGP1). GSH-Px activity was measured indirectly by a coupled reaction with glutathione reductase (GSH-GR). Oxidized glutathione (GSSG), produced upon reduction of an organic hydroperoxide by GSH-Px, was recycled to its reduced state by GSH-GR and NADPH. The oxidation of NADPH to NADP+ is accompanied by a decrease in absorbance at 340\u2009nm and is indicative of GSH-GPx activity.\n\nGlutathione reductase (GSH-GR; EC 1.6.4.2) enzyme activity was determined with an assay kit from Sigma Chemical (cat. no. GRSA). GSH-GR catalyzed the reduction of GSSG to reduced glutathione (GSH). GSH-GR activity can be measured either by the decrease in absorbance caused by the oxidation of NADRH at 340\u2009nm or by the increase in absorbance caused by the reduction of DTNB at 412\u2009nm.\n\nFor the vitamin E (Vit E) we measured it in plasma with an assay kit from Cayman Chemical (cat. no. 10010621). The absorbance of samples was measured between 405 and 420\u2009nm by a plate reader. The plasma concentrations of the bilirubin and iron (Fe) were estimated with a HumaStar 600 automatic analyzer (Human, Germany).\n\n2.3. Statistical Analysis {#sec2.3}\n-------------------------\n\nStatistical analysis was performed using SPSS 20.0 (IBM SPSS statistics; USA). Data were expressed as the mean \u00b1 SD (standard deviation). The distribution of variables was compared by the *\u03c7* ^2^ analysis. Difference between the arithmetical averages was assessed by ANOVA, which was adjusted for multiple comparisons. Depending on the normality of distribution of variables, the comparisons between groups were performed using one-way of analysis of variance (ANOVA) or the Mann-Whitney *U*-test when results are nonparametrically distributed. All statistical tests were two tailed, and a *P* value below 0.05 was considered statistically significant.\n\n3. Results {#sec3}\n==========\n\n3.1. Oxidative Status {#sec3.1}\n---------------------\n\nResults showed that the levels of lipid oxidative product (TBARS and LPO) and protein carbonyls were significantly increased in HD and PD patients compared to the others\\' stage of CKD (*P* \\< 0.001) ([Table 2](#tab2){ref-type=\"table\"}). Indeed, TBARS concentrations were 2.4-fold higher in HD, 2-fold higher in PD, and 1.5-fold higher in CKD4 than in CKD1 patients (*P* \\< 0.001). LPO were also 5-fold higher in HD, 4.2-fold higher in PD, and 4.3-fold higher in CKD4 than in CKD1 patients (*P* \\< 0.001). Carbonyls were 3.3-fold higher in HD, 3.4-fold higher in PD and 2.4-fold higher in CKD4 than in CKD1 patients (*P* \\< 0.001). Our results showed a considerable increase on protein carbonyls in PD patients. Indeed, it is important to note that the levels of the latter products were already increased at stage 3 and stage 4 of CKD.\n\n3.2. Antioxidative Status {#sec3.2}\n-------------------------\n\nOur results on the antioxidants status were reported in [Table 3](#tab3){ref-type=\"table\"}. Our results showed that SOD activity was decreased by \u221260% in HD, by \u221264% in PD, and by \u221266% in CKD4 than in CKD1 patients (*P* \\< 0.001). CAT activity was also significantly decreased by \u221276% in HD, \u221282% in PD, and \u221281% in CKD4 than in CKD1 patients (*P* \\< 0.001). A decrease in GSH-Px activity was noted in HD by \u221252%, in PD by \u221264%, and in CKD4 by \u221265% compared to CKD1 (*P* \\< 0.001). Similar result was observed for GSH-GR activity, which diminished by \u221257% in HD, by \u221266% in PD, and by \u221262% in CKD4 compared to CKD1 patients (*P* \\< 0.001).\n\nVit E concentrations were lower in all groups, especially in HD patients, which lowered by \u221236% compared to CKD1 (*P* \\< 0.001). Similar results were obtained for iron and bilirubin concentrations in which both of them decreased in all study groups especially in HD patients. Iron concentration was diminished by \u221273% in HD patients compared to CKD1 (*P* \\< 0.001), whereas bilirubin concentration was diminished around \u221265% in HD and PD patients compared to CKD1 patients (*P* \\< 0.001). Indeed, patients with severe and advanced stage of CKD had marked decreases in the concentrations of the antioxidants enzymes. Patients also had decreased Vit E, iron, and bilirubin. Further, this decrease was more pronounced in HD and PD patients.\n\n4. Discussion {#sec4}\n=============\n\nA comparative study was conducted in CKD patients in order to assess lipid peroxidation, protein oxidation, and antioxidant defence during the progression of CKD and to evaluate the effect of dialysis treatment on redox status.\n\nThe case-control studies agree that CKD patients have increased OS produced by an imbalance between pro- and antioxidant capacities \\[[@B18], [@B19]\\]. This OS is responsible for the peroxidation of macromolecules such as lipids and proteins causing significant damage. Several pathophysiologic explanations have been claimed; some attribute it to malnutrition and hypoalbuminemia having in these cases low availability of \"thiol\"; others to \"uremic status\" itself with solute retention that may favor their pathogenicity; and others to the association of comorbid factors such as advanced age, diabetes, and inflammatory and infectious phenomena \\[[@B20], [@B21]\\]. Besides, when these uremic patients receive treatment with an extrarenal depurative technique, as in HD, OS is promoted by several reasons among which the usage of low biocompatible synthetic membranes and the lack of ultrapure dialysis water stand out \\[[@B22]\\].\n\nOS is highly present in CKD patients, and it may contribute directly to the endothelial activation which favors the precocious atherosclerosis seen in these patients. Further investigation is still required in order to estimate the imbalance between antioxidant and oxidant factors, and in particular it is essential to investigate whether different RRT achieve similar levels of OS \\[[@B18]\\], especially in patients with high risk to develop CVD. Moreover, the effect of different RRT on oxidative status compared to undialyzed patients with different stage of CKD has not been clearly described.\n\nIn the study, we found an increase in markers of lipid peroxidation, in which a production of TBARS and LPO was significantly higher already at severe stage of CKD. Furthermore, we noted a significant increase in the production of protein oxidation. Carbonyls were increased already at severe stage of CKD and its production was intensifying in both HD and PD patients. Our results were in agreement with previous findings which associated the increased levels of lipid peroxidation with the severity of coronary atherosclerosis in patients with CKD \\[[@B23]\\].\n\nFree radicals are the source of lipid peroxidation derived from oxygen, and the first line of defense against them is SOD. Its function is to catalyze the conversion of superoxide radicals to hydrogen peroxide (H~2~O~2~). Indeed, the decreased SOD activity in our study, especially in HD patients, suggests that accumulation of superoxide anion radical might be responsible for increased lipid peroxidation \\[[@B24]\\]. Glutathione is a tripeptidic thiol found in the inside of all animal cells and likely is the most important cellular antioxidant. Oxidized glutathione (GSSG) is highly toxic to cells so that the organism tends to reduce GSSG to GSH through glutathione reductase. Thus, determining GSH-GR is considered a reliable estimate of the degree of cellular OS \\[[@B25]\\]. Moreover, GSH-Px is responsible for the most decomposition of lipid peroxide and protects the cell from the deleterious effects of peroxides \\[[@B24]\\]. H~2~O~2~ in the presence of sufficient CAT activity will be converted to harmless H~2~O and O~2~. Our patients have decreased levels of the four enzymes; this decrease was more important in HD and PD patients. Hence our result explained the reason of the increase in TBARS and LPO levels.\n\nVitamin E has an antioxidant effect that protects tissue from the oxidative damage \\[[@B26]\\]. Our investigation confirms the previous studies which found a significant decreased level of vitamin E in advanced stages of CKD patients especially those on HD \\[[@B27]\\].\n\nPlasma concentrations of the essential metals as iron (Fe) were significantly lower in HD patients than in the other groups of the study population. We found that patients had a strong association between increased plasma OS and decreased concentration of Fe. Long-term HD patients have significantly lower blood concentrations of Fe, and the possible causes of the loss of these essential metals have been studied \\[[@B28], [@B29]\\]. Indeed, accumulating FR accompanied reduced tissue concentrations of Fe and induced oxidative damage to brain tissue \\[[@B30]\\].\n\nAdditionally, bilirubin concentration is the primary defense against OS in extracellular fluid results, generated during normal metabolism or introduced in the body by the consumption of dietary products rich in antioxidants \\[[@B29]\\]. Knowing that our HD patients had dietary restriction intakes that are a part of a renal care, this unbalanced diet contributes to a significant OS.\n\nMoreover, in the HD population, the interaction between dialysis membranes and blood can trigger the release of oxygen-FR and oxidizing agents, such as superoxide anion, hydrogen peroxide, and myeloperoxidase. In turn, these molecules contribute to the oxidation of lipid products, proteins and nucleic acids. This oxidation has several pathophysiological consequences, including enhanced atherogenicity of Ox-LDL, and accelerated demise of circulating erythrocytes, leading to a shorter life span \\[[@B31], [@B32]\\].\n\nOur patients were treated by a conventional PD considered as bioincompatible. Repeated and long-term exposure to conventional glucose-based PD fluids plays a central role in the pathogenesis of the functional and structural changes of peritoneal membrane. Low pH, high glucose concentration, high osmolarity and heat sterilization represent major factors of low biocompatibility \\[[@B33]\\]. Bioincompatibility of PD solutions seems to play a central role in the increase of ROS production \\[[@B34]\\].\n\nIn conclusion, our study showed that impaired renal function and duration of dialysis treatment are associated with increased OS. Consequently decreased antioxidants defence which leads to promote the malnutrition, inflammation, and atherosclerosis (MIA syndrome). Our data suggest that patients subjected to RRT need new methods aimed to reduce intradialytic OS, such as incorporating antioxidant therapy into the dialysis membrane, hemolipodialysis, using of electrolyte-reduced water for dialysate, or using an ultrapure dialysate system to reduce acute phase inflammation.\n\nThe authors declare that they have no conflict of interests regarding the publication of this paper.\n\n###### \n\nCharacteristics of the study populations.\n\n  -----------------------------------------------------------------------------------------------------------------------------\n  \u2009                            CKD 1\\          CKD 2\\          CKD 3\\          CKD 4\\          HD\\              PD\\\n                               *n* = 28        *n* = 28        *n* = 28        *n* = 18        *n* = 40         *n* = 25\n  ---------------------------- --------------- --------------- --------------- --------------- ---------------- ---------------\n  Age (years)                  37 \u00b1 13         55 \u00b1 11         45 \u00b1 15         46 \u00b1 14         42 \u00b1 11          39 \u00b1 15\n\n  Weight (kg)                  66 \u00b1 12.74      67 \u00b1 16.12      67 \u00b1 13.96      61 \u00b1 13.87      56.44 \u00b1 11.37    68.22 \u00b1 10.55\n\n  BMI (Kg/m^2^)                24.07 \u00b1 7.16    25.35 \u00b1 5.63    24.91 \u00b1 3.54    23.63 \u00b1 4.01    23.15 \u00b1 3.02     25.43 \u00b1 3.28\n\n  Sex ratio (M/F)              10/18           11/17           10/18           07/11           22/18            06/06\n\n  GFR (mL/min)                 86.25 \u00b1 23.87   46 \u00b1 6.95       20.31 \u00b1 5.10    10.75 \u00b1 2.56    ---              ---\n\n  Dialysis duration (months)   ---             ---             ---             ---             14--109          05--49\n\n  Urea (g/L)                   0.34 \u00b1 0.15     0.53 \u00b1 0.20     1.05 \u00b1 0.34     1.67 \u00b1 0.82     1.32 \u00b1 0.36      1.04 \u00b1 0.38\n\n  Creatinin (mg/L)             10.30 \u00b1 2.89    18.54 \u00b1 5.26    41.75 \u00b1 11.46   70.03 \u00b1 26.15   102.12 \u00b1 29.37   69.36 \u00b1 36.95\n\n  Uric acid (g/L)              52.63 \u00b1 11.28   64.10 \u00b1 13.55   80.85 \u00b1 14.11   86.67 \u00b1 20.01   65.55 \u00b1 13.44    68.10 \u00b1 15.62\n\n  Total proteins (g/L)         75.17 \u00b1 5.21    74.29 \u00b1 7.52    71.16 \u00b1 11.15   70.86 \u00b1 10.52   69.92 \u00b1 9.60     67.04 \u00b1 7.36\n\n  Cholesterol (g/L)            1.68 \u00b1 0.33     1.89 \u00b1 0.29     1.73 \u00b1 0.32     1.74 \u00b1 0.45     2.17 \u00b1 1.58      1.95 \u00b1 0.56\n\n  HDL-cholesterol (g/L)        0.53 \u00b1 0.10     0.48 \u00b1 0.09     0.50 \u00b1 0.10     0.44 \u00b1 0.11     0.36 \u00b1 0.11      0.42 \u00b1 0.07\n\n  LDL-cholesterol (g/L)        0.93 \u00b1 0.33     1.13 \u00b1 0.22     0.95 \u00b1 0.93     0.94 \u00b1 0.41     1.14 \u00b1 0.51      1.21 \u00b1 0.45\n\n  Triglycerides (g/L)          1.17 \u00b1 0.53     1.39 \u00b1 0.55     1.47 \u00b1 0.74     1.30 \u00b1 0.55     1.64 \u00b1 0.90      1.75 \u00b1 0.89\n\n  Erythrocytes, \u00d710^6^/*\u03bc*L    4.72 \u00b1 0.70     4.43 \u00b1 0.59     4.03 \u00b1 0.46     3.25 \u00b1 0.50     3.07 \u00b1 0.73      3.68 \u00b1 0.86\n\n  Hemoglobin (g/dL)            13.01 \u00b1 1.36    12.28 \u00b1 1.31    11.46 \u00b1 1.53    8.71 \u00b1 1.37     9.55 \u00b1 1.86      10.38 \u00b1 2.44\n  -----------------------------------------------------------------------------------------------------------------------------\n\nBMI: body mass index (weight\u2009kg/height\u2009m^2^); GFR: glomerular filtration rate. Data are spoken in mean \u00b1 SD.\n\n###### \n\nOxidative status in the study groups.\n\n  \u2009                     CKD 1         CKD 2         CKD 3         CKD 4         HD            PD            *P*\n  --------------------- ------------- ------------- ------------- ------------- ------------- ------------- ---------\n  TBARS (*\u03bc*mol/L)      0.55 \u00b1 0.19   0.62 \u00b1 0.11   0.71 \u00b1 0.08   0.82 \u00b1 0.08   1.36 \u00b1 0.12   1.11 \u00b1 0.24   \\<0.001\n  LPO (nmol/mL)         0.34 \u00b1 0.26   0.80 \u00b1 0.39   1.25 \u00b1 0.36   1.46 \u00b1 0.17   1.71 \u00b1 0.43   1.45 \u00b1 0.12   \\<0.001\n  Carbonyls (nmol/mg)   0.56 \u00b1 0.15   0.95 \u00b1 0.13   1.04 \u00b1 0.33   1.37 \u00b1 0.36   1.85 \u00b1 0.16   1.92 \u00b1 0.13   \\<0.001\n\nTBARS: thiobarbituric acid reactive substances; LPO: hydroperoxides. Data are presented as the mean \u00b1 SD. Statistically significant differences between all the groups (*P* \\< 0.001).\n\n###### \n\nAntioxidative status in the study groups.\n\n  \u2009                  CKD 1           CKD 2          CKD 3           CKD 4           HD              PD              *P*\n  ------------------ --------------- -------------- --------------- --------------- --------------- --------------- ---------\n  SOD (U/mL)         81.89 \u00b1 8.38    68.37 \u00b1 2.45   60.23 \u00b1 6.15    54.29 \u00b1 2.94    49.01 \u00b1 3.03    52.53 \u00b1 4.33    \\<0.001\n  CAT (U/mL)         62.75 \u00b1 5.42    57.76 \u00b1 7.25   53.52 \u00b1 8.48    51.22 \u00b1 3.66    47.99 \u00b1 5.90    51.92 \u00b1 5.90    \\<0.001\n  GSH-Px (U/mL)      7.71 \u00b1 1.01     6.62 \u00b1 0.50    5.23 \u00b1 0.71     5.01 \u00b1 0.82     3.99 \u00b1 1.39     4.94 \u00b1 1.12     \\<0.001\n  GSH-GR (U/mL)      1.72 \u00b1 0.06     1.45 \u00b1 0.11    1.34 \u00b1 0.12     1.07 \u00b1 0.19     0.98 \u00b1 0.19     1.13 \u00b1 0.21     \\<0.001\n  Vit E (ng/mL)      0.81 \u00b1 0.06     0.62 \u00b1 0.04    0.43 \u00b1 0.05     0.36 \u00b1 0.01     0.29 \u00b1 0.03     0.36 \u00b1 0.01     \\<0.001\n  Fe (*\u03bc*g/dL)       81.44 \u00b1 20.45   74.90 \u00b1 8.53   67.00 \u00b1 10.98   61.42 \u00b1 22.08   59.22 \u00b1 16.39   60.90 \u00b1 19.72   \\<0.001\n  Bilirubin (mg/L)   6.37 \u00b1 1.80     5.95 \u00b1 1.96    5.50 \u00b1 1.43     5.00 \u00b1 1.41     4.48 \u00b1 1.68     4.40 \u00b1 1.07     \\<0.001\n\nSOD: superoxide dismutase; CAT: catalase; GSH-Px: glutathione peroxidase; GSH-GR: glutathione reductase; Vit E: vitamin E; Fe: iron. Data are presented as the mean \u00b1 SD. Statistically significant differences between all the groups (*P* \\< 0.001).\n\n[^1]: Academic Editor: Paul W. Doetsch\n"}
+{"text": "Early recognition of invading bacteria by the innate immune system fundamentally contributes to antibacterial defence by triggering inflammatory responses that prevent the spread of infection and suppress bacterial growth[@b1][@b2][@b3]. A key step in the induction of the inflammatory response to gram-negative bacteria is the activation of Toll-like receptor 4 (TLR4) signalling by lipopolysaccharide (LPS), a major component of the outer membrane of all gram-negative bacteria[@b1][@b2][@b3]. Dendritic cells (DCs), defined by their dendritic morphology and unique phenotype, are involved in the initiation of inflammation in response to gram-negative bacteria[@b3][@b4][@b5]. Moreover, DCs are the most potent professional antigen-presenting cells and thus play a pivotal role in linking the innate and adaptive immune response[@b3][@b4][@b5]. In addition to their vital roles in antibacterial defence, DCs are also indispensable for the induction and maintenance of immunological tolerance. Recently, the identification and characterisation of DCs with regulatory properties (so-called regulatory or tolerogenic DCs) has attracted much attention[@b4][@b5][@b6][@b7][@b8][@b9][@b10]. Regulatory DCs usually produce large amounts of interleukin-10 (IL-10), thereby promoting the generation of IL-10-producing T cells[@b6][@b7][@b8][@b9][@b10]. However, whether regulatory DCs can modulate inflammatory T cell responses through other mechanisms remains unclear.\n\nSeveral reports have discussed the potential regulatory function of a DC subset characterised by its particular CD11c^low^CD45RB^+^ surface marker expression[@b6][@b7][@b8][@b9][@b10]. Naturally occurring CD11c^low^CD45RB^+^ DCs are present in the spleens and lymph nodes of normal mice and are present at an increased level in transgenic mice expressing high levels of IL-10 and in mice experiencing a parasitic infection[@b6][@b10]. Naturally occurring CD11c^low^CD45RB^+^ DCs and those induced by a parasitic infection have been demonstrated to induce IL-10-expressing CD4^+^ T cells[@b6][@b10]. A similar expansion of splenic CD11c^low^CD45RB^+^ DCs has also been reported in mice injected with sublethal doses of LPS[@b10]. Changes in the number and function of DCs have been reported to play an important role in endotoxin tolerance[@b4][@b5]. However, the function of endotoxic shock-expanded CD11c^low^CD45RB^+^ DCs has not been examined.\n\nIn this work, we show that intra-peritoneal (i.p.) *Escherichia coli* (*E. coli*) infection and i.p. administration of *E. coli*-derived LPS both induce the expansion of CD11c^low^CD45RB^+^ cells with dendritic morphology and the production of low levels of inflammatory cytokines and costimulatory molecules. The expanded CD11c^low^CD45RB^+^ DCs modulate inflammatory T cell responses through induction of regulatory T cells (Tregs), incapability to stimulate T cells, and induction of T cell apoptosis.\n\nResults\n=======\n\nEndotoxic shock promotes the expansion of CD11c^low^CD45RB^+^ cells\n-------------------------------------------------------------------\n\nWith antibodies against CD11c and CD45RB, we identified three subpopulations of CD11c-expressing cells, CD11c^hi^CD45RB^\u2212^, CD11c^low^CD45RB^+^ , and CD11c^low^CD45RB^\u2212^, in the spleens of C57BL/6 (B6) mice ([Fig. 1A](#f1){ref-type=\"fig\"}). To verify whether endotoxic shock leads to the expansion of CD11c^low^CD45RB^+^ cells, we used the laboratory *E. coli* strain K12. Four days after i.p. infection with *E. coli* K12, the percentage of CD11c^low^CD45RB^+^ cells, but not of the other subpopulations, increased ([Fig. 1A](#f1){ref-type=\"fig\"}). However, in a model of acute self-limiting sterile inflammation[@b11], the percentage of CD11c^low^CD45RB^+^ cells remained largely unchanged 4 days after i.p. injection of thioglycolate ([Fig. 1A](#f1){ref-type=\"fig\"}). These data suggest that the expansion of CD11c^low^CD45RB^+^ cells depends on the intensity of inflammation. Because of the splenomegaly induced by *E. coli* infection, the absolute number of CD11c^low^CD45RB^+^ cells increased over 5-fold, reaching its peak on day 5 after infection ([Fig. 1B](#f1){ref-type=\"fig\"}). This expansion was significantly reduced by simultaneous treatment with cholera toxin ([Fig. 1B](#f1){ref-type=\"fig\"}), which has been shown to suppress inflammation *in vivo*[@b12], further suggesting that the intensity of inflammation is indispensable for the expansion of CD11c^low^CD45RB^+^ cells. The expansion of these cells also occurred in BALB/c mice ([Fig. 1C](#f1){ref-type=\"fig\"}). A similar change in splenic DC subsets was observed when mice were injected with sublethal doses of purified *E. coli* LPS ([Fig. 1D](#f1){ref-type=\"fig\"}). Therefore, endotoxic shock promotes the expansion of CD11c^low^CD45RB^+^ cells.\n\nPhenotypic characterisation of the expanded CD11c^low^CD45RB^+^ cells\n---------------------------------------------------------------------\n\nNext, we attempted to identify the expanded CD11c^low^CD45RB^+^ cells. Regarding lineage markers, these cells showed weak expression of Gr1 and F4/80 and no expression of the T cell markers CD4 and CD8 or the B cell marker CD19 ([Fig. 2A](#f2){ref-type=\"fig\"}). Giemsa staining revealed the expanded cells exhibited immature dendritic morphology, as compared to the mature dendritic morphology of CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 2B](#f2){ref-type=\"fig\"}). As plasmacytoid DCs also express low levels of CD11c, we stained splenocytes from mice using the plasmacytoid DC marker CD317 (also called PDCA-1)[@b13]. The expanded cells did not express CD317, but CD11c^low^CD45RB^\u2212^ cells exhibited significant expression of that marker ([Fig. 2C](#f2){ref-type=\"fig\"}). The analysis of the cytokine profiles revealed that the expanded cells secreted slightly less tumor necrosis factor-\u03b1 (TNF-\u03b1) and IL-6, a similar level of transforming growth factor-\u03b2 (TGF-\u03b2), and a slightly higher level of IL-10 in response to LPS as compared to CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 2D](#f2){ref-type=\"fig\"}). Regarding functional markers, the expanded cells showed almost undetectable levels of major histocompatibility complex (MHC) molecule I-A and costimulatory molecules (CD40, CD80, CD86), in contrast to the significant expression of these molecules in CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 2E](#f2){ref-type=\"fig\"}). In addition, the expanded cells expressed the adherence molecules CD54 and CD11b, but to a lesser extent than CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 2E](#f2){ref-type=\"fig\"}). MHC molecule I-A and costimulatory molecule CD40 remained undetectable after the expanded cells underwent *in vitro* stimulation with LPS, indicating a stable phenotype for these cells ([Fig. 2F](#f2){ref-type=\"fig\"}). Taken together, these data suggest that endotoxic shock-expanded CD11c^low^CD45RB^+^ cells are less capable of stimulating T cells than CD11c^hi^CD45RB^\u2212^ conventional DCs. On the other hand, only 15% of CD11c^low^CD45RB^+^ cells in untreated mice showed low level of MHC molecule I-A expression ([Supplementary Figure 1](#S1){ref-type=\"supplementary-material\"}). Moreover, the majority of CD11c^low^CD45RB^+^ I-A^\u2212^ cells purified from untreated mice upregulated the expression of MHC molecule I-A after they underwent *in vitro* stimulation with LPS ([Supplementary Figure 2](#S1){ref-type=\"supplementary-material\"}). These data suggest naturally occurring CD11c^low^CD45RB^+^ cells are heterogeneous and only a small portion of them have regulatory effects. Therefore, it is more interesting to explore the functions of the expanded CD11c^low^CD45RB^+^ cells.\n\nThe expanded CD11c^low^CD45RB^+^ cells induce apoptosis of CD4^+^ T cells *in vitro*\n------------------------------------------------------------------------------------\n\nBecause we did not observe an inhibitory role of the expanded CD11c^low^CD45RB^+^ cells on cytokine production by macrophages *in vitro* (data not shown), we next explored how the expanded cells might affect T cell responsiveness. The potential mechanisms of inflammatory T cell responses include molecular mimicry, bystander activation, epitope spreading, and superantigen activation of T cells[@b14][@b15][@b16]. In this regard, a polyclonal T cell stimulation approach was employed. Allogeneic CD4^+^ CD25^\u2212^ T cells were co-cultured with the expanded CD11c^low^CD45RB^+^ cells or CD11c^hi^CD45RB^\u2212^ conventional DCs. Stimulation was performed using beads conjugated with antibodies against CD3 and CD28 in the presence of neutralisation antibodies against IFN-\u03b3 and IL-4. After 5 days, secretion of IFN-\u03b3, IL-4, IL-10, and IL-17A was evaluated by ELISA ([Fig. 3A](#f3){ref-type=\"fig\"}). Co-culture of allogeneic CD4^+^ CD25^\u2212^ T cells with conventional DCs significantly enhanced IFN-\u03b3 and IL-17A secretion and slightly enhanced IL-10 secretion without affecting IL-4 production ([Fig. 3A](#f3){ref-type=\"fig\"}). Co-culture of allogeneic CD4^+^ CD25^\u2212^ T cells with the expanded cells led to a partial decrease in IFN-\u03b3 secretion, similar IL-4 and IL-10 production, and a dramatic decrease in IL-17A secretion compared to co-culture with conventional DCs ([Fig. 3A](#f3){ref-type=\"fig\"}). Intracellular staining revealed that under the Th1 condition, co-culture with conventional DCs significantly augmented the proportion of CD4^+^ T cells that produced IFN-\u03b3 or IL-17A ([Fig. 3B](#f3){ref-type=\"fig\"}). Co-culture of allogeneic CD4^+^ CD25^\u2212^ T cells with the expanded cells resulted in a decreased proportion of CD4^+^ T cells producing IL-17A but a similar proportion of CD4^+^ T cells producing IFN-\u03b3 as compared to co-culture with conventional DCs ([Fig. 3B](#f3){ref-type=\"fig\"}). In addition, the expanded cells showed similar ability to promote Foxp3 expression compared with conventional DCs under both Th0 condition ([Supplementary Figure 3](#S1){ref-type=\"supplementary-material\"}) and the Treg condition ([Fig. 3C](#f3){ref-type=\"fig\"}).\n\nThe similar percentages of CD4^+^ T cells expressing IFN-\u03b3 and the different IFN-\u03b3 levels in the supernatants obtained from the two co-culture systems suggest that the expanded cells might affect T cell proliferation and/or survival. *In vitro* CFSE dilution analysis revealed that the antigen-nonspecific proliferation of CD4^+^ T cells was slightly enhanced by conventional DCs and slightly inhibited by the expanded cells ([Fig. 3D](#f3){ref-type=\"fig\"}). Apoptosis analysis demonstrated that conventional DCs slightly promoted apoptosis of CD4^+^ T cells, but the expanded cells exhibited a much stronger ability to promote apoptosis ([Fig. 3E](#f3){ref-type=\"fig\"}; [Supplementary Figure 4](#S1){ref-type=\"supplementary-material\"}). Taken together, these data suggest that the expanded cells partially affect T cell responsiveness *in vitro*.\n\nFas ligand (FasL) mediates the induction of CD4^+^ T cell apoptosis by expanded CD11c^low^CD45RB^+^ cells *in vitro*\n--------------------------------------------------------------------------------------------------------------------\n\nBecause the expanded CD11c^low^CD45RB^+^ cells did not show significantly higher production of IL-10 or TGF-\u03b2 as compared to CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 2D](#f2){ref-type=\"fig\"}), these expanded cells must employ other mechanism(s) to affect T cell responsiveness *in vitro*. Multiple molecules, including nitric oxide, reactive oxygen species (ROS), programmed death ligand 1 (PD-L1), PD-L2, and Fas ligand (FasL), have been reported to trigger apoptosis of activated CD4^+^ T cells[@b17][@b18][@b19][@b20]. The expanded CD11c^low^CD45RB^+^ cells did not exhibit higher levels of inducible nitric oxide synthase (iNOS), ROS, or PD-L1 than CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 4A--C](#f4){ref-type=\"fig\"}). Consistent with the proliferation data, the expanded cells also did not show enhanced production of arginase I ([Fig. 4A](#f4){ref-type=\"fig\"}), which has been reported to mediate the inhibition of T cell proliferation[@b16]. However, the expanded cells expressed higher levels of PD-L2 and FasL than conventional DCs ([Fig. 4C](#f4){ref-type=\"fig\"}). With neutralisation antibodies against PD-L2 and FasL, we tested the roles of these two molecules in the induction of apoptosis. As expected, the blockade of FasL significantly reversed the enhanced apoptosis ([Fig. 4D](#f4){ref-type=\"fig\"}) and the reduced cell number ([Fig. 4E](#f4){ref-type=\"fig\"}) of CD4^+^ T cells co-cultured with the expanded cells, whereas the blockade of PD-L2 exhibited only a marginal effect ([Fig. 4D,E](#f4){ref-type=\"fig\"}). Thus, FasL mediates the induction of CD4^+^ T cell apoptosis by expanded CD11c^low^CD45RB^+^ cells *in vitro*.\n\nExpanded CD11c^low^CD45RB^+^ DCs suppress colitis induction by CD4^+^ CD25^\u2212^ T cells *in vivo*\n-----------------------------------------------------------------------------------------------\n\nTo verify whether endotoxic shock-expanded CD11c^low^CD45RB^+^ cells partially suppress T cell responsiveness *in vivo*, colitis induction by CD4^+^ CD25^\u2212^ T cells was employed[@b21][@b22]. CB-17 SCID mice were divided into four groups, and CD4^+^ CD25^\u2212^ colitogenic effector T cells were transferred into three of the groups to induce colitis, as described previously[@b21][@b22]. At the time of T cell transfer, mice were treated with PBS containing CD11c^hi^CD45RB^\u2212^ conventional DCs or the expanded CD11c^low^CD45RB^+^ DCs (1\u2009\u00d7\u200910^6^ cells/mouse). Four weeks after T cell transfer, mice treated with CD4^+^ CD25^\u2212^ T cells/PBS showed significantly lower body weights than untreated mice ([Fig. 5A](#f5){ref-type=\"fig\"}). Treatment with CD11c^hi^CD45RB^\u2212^ conventional DCs was insufficient for protection against the loss of body weight ([Fig. 5A](#f5){ref-type=\"fig\"}). However, treatment with the expanded CD11c^low^CD45RB^+^ DCs significantly reversed the loss of body weight ([Fig. 5A](#f5){ref-type=\"fig\"}). Macroscopic examinations at week 4 after transfer revealed that treatment with CD11c^hi^CD45RB^\u2212^ conventional DCs tended to reverse the shortening of the colon but not at a statistically significant level ([Fig. 5B](#f5){ref-type=\"fig\"}). In contrast, treatment with the expanded cells led to significantly longer colons ([Fig. 5B](#f5){ref-type=\"fig\"}). Upon histological examination, colitis was characterised by severe epithelial hyper-proliferation, mucus depletion, massive infiltration of inflammatory cells, crypt abscesses, reduced numbers of goblet cells, and erosions. CB-17 SCID mice injected with CD4^+^ CD25^\u2212^ T cells and treated with PBS had severe colitis ([Fig. 5C](#f5){ref-type=\"fig\"}). Treatment with CD11c^hi^CD45RB^\u2212^ conventional DCs partially improved these histological signs ([Fig. 5C](#f5){ref-type=\"fig\"}). More importantly, treatment with the expanded cells almost completely abolished signs of colitis ([Fig. 5C](#f5){ref-type=\"fig\"}). There was a significant reduction of inflammatory cell infiltration, and goblet cell number and mucus were preserved at normal levels ([Fig. 5C](#f5){ref-type=\"fig\"}). To rule out the possibility that the DC preparations contained residual T cells that were responsible for the protective effects in the co-transfer experiment, rather than the DCs, SCID mice were used as recipients of the expanded cells. Four weeks later, flow cytometry revealed that no CD4+ or CD3+ cells could be detected in the spleens of SCID mice with or without transfer of the expanded cells ([Supplementary Figure 5](#S1){ref-type=\"supplementary-material\"}). Thus, the expanded CD11c^low^CD45RB^+^ DCs suppressed colitis induction by CD4^+^ CD25^\u2212^ T cells *in vivo*.\n\nAdministration of expanded CD11c^low^CD45RB^+^ DCs reduces proinflammatory cytokine responses *in vivo*\n-------------------------------------------------------------------------------------------------------\n\nTo elucidate the mechanism(s) underlying the protective role of the expanded cells against colitis, the expression levels of Foxp3 and proinflammatory cytokines were measured. Treatment with either CD11c^hi^CD45RB^\u2212^ conventional DCs or the expanded CD11c^low^CD45RB^+^ DCs significantly induced Foxp3 expression in splenic CD4^+^ T cells ([Fig. 6A](#f6){ref-type=\"fig\"}). Furthermore, treatment with either of the two cell types led to reduced Th1 differentiation in spleens, as revealed by intracellular IFN-\u03b3 staining ([Fig. 6B](#f6){ref-type=\"fig\"}). However, there was no significant difference in Foxp3 or IFN-\u03b3 expression in splenic CD4^+^ T cells between the groups of mice treated with conventional DCs and those treated with the expanded cells ([Fig. 6A,B](#f6){ref-type=\"fig\"}). To get a better idea about the proinflammatory cytokine responses *in vivo*, mesenteric lymph node (MLN) cells (1\u2009\u00d7\u200910^6^) of colitic mice were cultured for 24\u2009h in 24-well plates in the presence of 1\u2009\u03bcg/ml CD3 and 1\u2009\u03bcg/ml CD28. The culture supernatants were then harvested and subjected to ELISA. As shown in [Fig. 6C](#f6){ref-type=\"fig\"}, treatment with the expanded CD11c^low^CD45RB^+^ DCs, but not treatment with CD11c^hi^CD45RB^\u2212^ conventional DCs, led to significantly reduced IFN-\u03b3 and TNF-\u03b1 production, even though both conditions resulted in significantly reduced IL-17A production. In addition, neither treatment showed an effect on IL-10 expression by MLN cells ([Fig. 6C](#f6){ref-type=\"fig\"}). Since colitis is a colonic inflammatory disease, it would be more meaningful to determine proinflammatory cytokine responses in the colonic tissues. Indeed, immunohistochemistry revealed significantly reduced IFN-\u03b3 expression in the colonic tissues upon treatment with the expanded CD11c^low^CD45RB^+^ DCs, but not treatment with CD11c^hi^CD45RB^\u2212^ conventional DCs ([Fig. 6D](#f6){ref-type=\"fig\"}).\n\nAdministration of expanded CD11c^low^CD45RB^+^ DCs induces T cell apoptosis *in vivo*\n-------------------------------------------------------------------------------------\n\nSince the expanded CD11c^low^CD45RB^+^ cells significantly induce apoptosis of CD4^+^ T cells *in vitro* ([Fig. 3E](#f3){ref-type=\"fig\"}), the better control of proinflammatory cytokine responses by the expanded cells as compared to the conventional DCs might be attributed, at least partially, to their ability to induce T cell apoptosis. To test this notion, the colon samples of colitic mice were subjected to terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) staining. As expected, more apoptosis was observed in the Payer's patches ([Fig. 7A](#f7){ref-type=\"fig\"}) and in colonic tissue-infiltrating cells ([Fig. 7B](#f7){ref-type=\"fig\"}) of mice treated with the expanded cells than in those treated with conventional DCs. Another issue of interest is how long these expanded cells could provide protection. The percentage of this subpopulation in the spleen and MLN was examined until 2 weeks after LPS challenge. Flow cytometry analysis revealed that even though the expanded cells disappear in the mouse spleen after reaching their peak on day 4-5 of LPS treatment ([Fig. 1B](#f1){ref-type=\"fig\"} and [Supplementary Figure 6](#S1){ref-type=\"supplementary-material\"}), an elevated percentage of this subpopulation was still detected in mouse MLN cells 10-14 days after LPS treatment ([Supplementary Figure 6](#S1){ref-type=\"supplementary-material\"}). Thus, the expanded cells can last for a significant amount of time after endotoxic shock.\n\nDiscussion\n==========\n\nRecent work has provided evidence supporting the importance of DCs, especially regulatory DCs, in peripheral tolerance and thus as potential immunotherapies for autoimmune diseases. Regulatory DCs usually produce a large amount of IL-10, thereby promoting the generation of IL-10-producing T cells. However, in this study, endotoxic shock-expanded CD11c^low^CD45RB^+^ regulatory DCs exhibited low levels of basal IL-10 expression and LPS-induced IL-10 expression ([Fig. 2D](#f2){ref-type=\"fig\"}). Consequently, the expanded cells failed to potently induce IL-10 production by T cells *in vitro* ([Fig. 3A](#f3){ref-type=\"fig\"}) and *in vivo* ([Fig. 6C](#f6){ref-type=\"fig\"}). Thus, the expanded cells ameliorated colitis without expressing an abundance of IL-10.\n\nThese results led us to examine how these expanded CD11c^low^CD45RB^+^ DCs modulate inflammatory T cell responses, and our results suggest that at least three factors are involved. First, the expanded cells induce the generation of Foxp3+ Tregs. Administration of the expanded cells led to an increase in Tregs *in vivo* ([Fig. 6A](#f6){ref-type=\"fig\"}). *In vitro*, the expanded cells promoted Foxp3 expression under both the Th0 condition ([Supplementary Figure 3](#S1){ref-type=\"supplementary-material\"}) and Treg condition ([Fig. 3C](#f3){ref-type=\"fig\"}). Because Tregs play important roles in reducing colitis[@b23], it is reasonable to propose that the expanded cells inhibit inflammatory T cell responses through, at least partially, the promotion of the generation of Tregs. CD11c^hi^CD45RB^\u2212^ conventional DCs showed a similar ability to induce Foxp3 expression *in vitro* and *in vivo* ([Fig. 3C](#f3){ref-type=\"fig\"}, [Supplementary Figure 3](#S1){ref-type=\"supplementary-material\"}, and [Fig. 6A](#f6){ref-type=\"fig\"}). The induction of Tregs by splenic conventional DCs has been reported to contribute to peripheral tolerance[@b24]. Consistently, CD11c^hi^CD45RB^\u2212^ conventional DCs exhibited partial protection against the induction of colitis by CD4^+^ CD25^\u2212^ T cells ([Fig. 5](#f5){ref-type=\"fig\"}), even though these conventional DCs significantly augmented IFN-\u03b3 and IL-17A expression in activated CD4^+^ T cells ([Fig. 3A,B](#f3){ref-type=\"fig\"}) and slightly enhanced the antigen-nonspecific proliferation of CD4^+^ T cells ([Fig. 3D](#f3){ref-type=\"fig\"}) *in vitro*. *In vivo*, the induction of Tregs by conventional DCs would likely overcome their detrimental effects.\n\nSecond, the expanded cells were relatively incapable of stimulating T cells. Compared with conventional DCs, the expanded cells expressed reduced levels of proinflammatory cytokines ([Fig. 2D](#f2){ref-type=\"fig\"}) and adherence molecules ([Fig. 2E](#f2){ref-type=\"fig\"}). Furthermore, the expanded cells showed almost undetectable levels of MHC molecule I-A and costimulatory molecules ([Fig. 2E](#f2){ref-type=\"fig\"}), which remained stable even after *in vitro* exposure to LPS ([Fig. 2F](#f2){ref-type=\"fig\"}). Consistently, the expanded cells were less potent at inducing IL-17A-producing T cells than conventional DCs *in vitro* ([Fig. 3B](#f3){ref-type=\"fig\"}). While conventional DCs slightly enhanced the antigen-nonspecific proliferation of CD4^+^ T cells *in vitro*, the expanded cells exhibited the opposite effect ([Fig. 3D](#f3){ref-type=\"fig\"}). Thus, the relative incapability of the expanded cells to stimulate T cells should allow them to exert better protection against colitis induced by CD4^+^ CD25^\u2212^ T cells than conventional DCs.\n\nThird, and possibly most importantly, the expanded cells induce T cell apoptosis. More apoptosis was observed in the Payer's patches ([Fig. 7A](#f7){ref-type=\"fig\"}) and in colonic tissue-infiltrating cells ([Fig. 7B](#f7){ref-type=\"fig\"}) of mice treated with the expanded cells than in those treated with conventional DCs. Induction of T cell apoptosis by the expanded cells was also observed *in vitro* ([Fig. 3E](#f3){ref-type=\"fig\"}; [Supplementary Figure 4](#S1){ref-type=\"supplementary-material\"}), and this was reversed by treatment with a neutralisation antibody against FasL ([Fig. 4D,E](#f4){ref-type=\"fig\"}). Although some DC subsets have been reported to induce T cell apoptosis or death, rare studies have reported this ability in regulatory DCs. Splenic CD8a^+^ DCs expressing high levels of FasL can kill CD4^+^ T cells via Fas-mediated apoptosis[@b20]. Interestingly, the expanded CD11c^low^CD45RB^+^ regulatory DCs showed no CD8a expression ([Fig. 2A](#f2){ref-type=\"fig\"}), and this result is similar to splenic stroma-educated regulatory DCs, which also express FasL and induce apoptosis of activated T cells without expressing CD8a[@b18]. These findings suggest that induction of T cell apoptosis is one of the mechanisms by which regulatory DCs exert their immunoregulatory function. The induction of T cell apoptosis by the expanded cells is another mechanism that helps them to protect against colitis induced by CD4^+^ CD25^\u2212^ T cells better than conventional DCs.\n\nThe mechanism(s) by which endotoxic shock promotes the production of CD11c^low^CD45RB^+^ cells remain elusive. As we failed to detect the *in vitro* expansion of naturally occurring CD11c^low^CD45RB^+^ DCs in the presence of LPS (data not shown), the crosstalk of this cell type with other cell types must be essential for their *in vivo* expansion and for the functional switch after exposure to LPS. Moreover, our data suggest that the intensity of inflammation is indispensable for the expansion of these cells ([Fig. 1](#f1){ref-type=\"fig\"}). Another possible origin of the expanded CD11c^low^CD45RB^+^ cells is *de novo* induction of haematopoietic progenitor cells. Future studies are required to clarify the origin of CD11c^low^CD45RB^+^ cells.\n\nThe advantages and disadvantages of these expanded regulatory DCs in various inflammatory diseases are of interest. It is known that the incidences of Th1- and Th17-type inflammatory diseases---type I diabetes mellitus, multiple sclerosis and Crohn's disease---have increased considerably over the past half century, especially in Europe and North America[@b25]. It has been suggested that the lack of exposure of persons to sublethal doses of endotoxin due to improved sanitation in industrialised and urban areas increases the incidence of such diseases[@b25]. It is possible that exposure to sublethal doses of endotoxin leads to the generation or enhancement of certain regulatory mechanism(s) that protect against such diseases. In this regard, it is more important to determine the characteristics of the expanded CD11c^low^CD45RB^+^ cells than those of their naturally occurring counterparts. Our data show the expanded cells can last for a significant amount of time after endotoxic shock ([Supplementary Figure 6](#S1){ref-type=\"supplementary-material\"}) and consequently might provide some protection against Th1- and Th17-type inflammatory diseases. In contrast, during sepsis, CD11c^low^CD45RB^+^ cells may not be beneficial because of their extraordinary ability to induce CD4 T cell apoptosis, which could promote immune suppression in the host, making the host more vulnerable to secondary infection[@b26].\n\nMethods\n=======\n\nMice\n----\n\nC57BL/6, CB-17 SCID, and BALB/c (H-2d) female mice of 8-12 weeks of age were purchased from Beijing Vital River Laboratory Animal, Inc. (Beijing, China, <http://www.vitalriver.com.cn>). All mice were maintained under specific pathogen-free conditions. The care, use and treatment of mice in this study were in strict agreement with the guidelines set by the Institute of Basic Medical Sciences. Briefly, mice were anaesthetised by administrating pentobarbital i.p. at 50\u2009mg/kg and were subjected to euthanasia by cervical dislocation. All *in vivo* experiments were in accordance with the ARRIVE guidelines, and all efforts were made to minimise the suffering of the mice. The food and water provided to the mice were subjected to aseptic processing. The number of mice used in this study was approximately 150, and all mice were euthanised.\n\nInduction of peritonitis\n------------------------\n\nLog-phase *E. coli* K12 were washed and resuspended in 0.5\u2009ml of PBS to the desired concentration. Each mouse received an i.p. injection of 1\u2009\u00d7\u200910^8^ cfu *E. coli* K12 or 2\u2009ml of thioglycolate per 20\u2009g of body weight.\n\nFlow cytometry\n--------------\n\nCells from spleens and lymph nodes were depleted of erythrocytes by hypotonic lysis. The cells were washed with FACS washing buffer (2% FBS, 0.1% NaN~3~ in PBS) twice and were then incubated with fluorescence-conjugated antibodies against cell surface molecules for 30\u2009min on ice in the presence of 2.4G2 mAb to block Fc\u03b3R binding. Isotype antibodies were included as negative controls. For intracellular cytokine staining, single-cell suspensions were stimulated with 50\u2009ng/ml PMA and 1\u2009\u03bcM ionomycin in the presence of brefeldin A solution for 4\u2009h. After stimulation, cells were stained with fluorescence-conjugated antibodies against CD4 and CD25, fixed and permeabilised using a fixation/permeabilisation kit (eBioscience, San Diego, CA, USA) and stained with fluorescence-conjugated specific antibodies against IFN-\u03b3, IL-17A, and Foxp3 in accordance with the manufacturer's instructions. Flow cytometry was performed using a Becton Dickinson FACSCalibur machine.\n\nPurification of DC subsets\n--------------------------\n\nFirst, DCs were enriched using an OptiPrep density gradient (Axis-Shield PoC AS, Oslo, Norway) by centrifugation at 600\u2009\u00d7\u2009*g* for 30min at 24\u2009\u00b0C. This protocol generated more than 30% CD11c^+^ cells. The mixture was stained with fluorescence-conjugated antibodies against CD11c, CD45RB, and MHC molecule I-A. Different subsets of DCs were then sorted using a Becton Dickinson FACSVantage machine. Cell purity was verified to be at least 95%.\n\nGiemsa staining\n---------------\n\nCells were collected on precoated (poly-L-lysine) coverslips, fixed with methanol, and stained with Giemsa dye for 10\u2009min. After washing with water, the morphology of these cells was observed under a microscope.\n\nPurification of CD4^+^ CD25^\u2212^ T cells\n--------------------------------------\n\nFirst, CD4^+^ T cells were isolated by negative selection from single-cell suspensions of spleens using a CD4^+^ T cell isolation kit II (Miltenyi Biotech, Bergisch Gladbach, Germany). Then, the cells were incubated with CD25 microbeads, and all CD25^+^ cells were eliminated using an MS column. The purity of CD4^+^ CD25^\u2212^ T cells was \\>95% as confirmed by flow cytometry.\n\n*In vitro* assays for T cell differentiation\n--------------------------------------------\n\nPurified CD11c^low^CD45RB^+^ cells or CD11c^hi^CD45RB^\u2212^ cells were co-cultured with CD4^+^ CD25^\u2212^ splenic cells (2.5\u2009\u00d7\u200910^4^\u2009cells/well) in a 96-well plate at a ratio of 4:1. For the Th0 condition, the cells were stimulated with Dynabeads mouse CD3/CD28 T cell expanders (Invitrogen, Carlsbad, CA, USA) in the presence of neutralisation antibodies against IFN-\u03b3 and IL-4 (1\u2009\u03bcg/ml each) for 5 days. For the Th1 condition and the Treg condition, 10\u2009ng/ml recombinant murine IL-12 and 10\u2009ng/ml recombinant human TGF-\u03b2 were used in addition to the reagents used for the Th0 condition, respectively. The supernatants were harvested for ELISA. CD4^+^ T cell differentiation of cells was assessed by flow-cytometric analysis of IFN-\u03b3, IL-17A, and Foxp3.\n\n*In vitro* assays for T cell proliferation and apoptosis\n--------------------------------------------------------\n\nPurified DC subsets were co-cultured with CD4^+^ CD25^\u2212^ splenic cells as stated above. 72 or 96\u2009h later, cells were stained with anti-CD4-PECy5, PI, and Annexin V-FITC resuspended in 300\u2009\u03bcl binding buffer containing calcium ion. Apoptosis was assessed by flow-cytometric analysis of Annexin-V/PI staining in CD4^+^ cells. Cells were collected for the same time period and the number of CD4^+^ Annexin-V^\u2212^ cells was counted. For the analysis of T cell proliferation, CD4^+^ CD25^\u2212^ splenic cells at 10^6^/ml were labelled by incubation in RPMI 1640 medium with 0.1\u2009\u03bcM CFSE at 37\u2009\u00b0C for 20\u2009min before co-culturing them with DCs. Proliferation was assessed by flow-cytometric analysis of CFSE dilutions of CD4^+^ cells after 96\u2009h of co-culture.\n\nReal-time PCR\n-------------\n\nRNA was extracted by using TRIzol reagent. First-strand synthesis was performed with Oligo dT primers and reverse transcription was performed with M-MLV reverse transcriptase. Quantitative real-time PCR was performed using SYBR Green reagent (TOYOBO, Tokyo, Japan) in a real-time PCR machine Realplex 2 (Eppendorf, Hamberg, Germany). Reactions were performed three times independently, and GAPDH values were used to normalise gene expression. The following primers were used: murine iNOS, 5\u2032-gttctcagcccaacaatacaaga-3\u2032 (forward) and 5\u2032-gtggacgggtcgatgtcac-3\u2032 (reverse); murine arginase 1, 5\u2032-ctccaagccaaagtccttagag-3\u2032 (forward) and 5\u2032-aggagctgtcattagggacatc-3\u2032 (reverse); and GAPDH, 5\u2032-ggcaaattcaacggcacagt-3\u2032 (forward) and 5\u2032-agatggtgatgggcttccc-3\u2032 (reverse).\n\nROS production assays\n---------------------\n\nA LIVE Green Reactive Oxygen Species Detection Kit (Life Technologies, Eugene, OR, USA) was used for detecting the generation of ROS. Briefly, cells were incubated in serum-free RPMI medium containing 2\u2009\u03bcM carboxy-H~2~DCFDA, anti-CD11c-PECy5 and anti-CD45RB-PE at 37\u2009\u00b0C for 30\u2009min. Cells were washed with PBS and were immediately subjected to flow cytometry to analyse the intensity of green fluorescence at a 488\u2009nm excitation wavelength.\n\nInduction of colitis by transfer of CD4^+^ CD25^\u2212^ T cells\n----------------------------------------------------------\n\nMice within litters were randomly divided into different groups. Each litter (containing mice in different groups) was housed in the same cage. Induction of colitis by the transfer of CD4^+^ CD25^\u2212^ T cells was performed according to methods described in a previous report, with some modifications[@b18][@b19]. Briefly, CD4^+^ CD25^\u2212^ T cells (3\u2009\u00d7\u200910^5^\u2009cells/mouse) obtained from BALB/c mice were suspended in 0.2\u2009ml PBS and i.v. injected into SCID mice. SCID controls received 0.2\u2009ml PBS alone. The day of this transfer was designated as day 0. Then, also on day 0, DCs (1\u2009\u00d7\u200910^6^\u2009cells/mouse) generated from BALB/c mice were injected i.v. The body weights of all mice were measured weekly.\n\nHistological assessment of colitis induced by transfer of CD4^+^ CD25^\u2212^ T cells\n--------------------------------------------------------------------------------\n\nMice were euthanised, and colons were removed 4 weeks after cell transfer. The transverse colons were fixed for 24\u2009hours in 4% paraformaldehyde, dehydrated, infiltrated with paraffin and sectioned at 5\u2009\u03bcm. Slides were stained with haematoxylin/eosin. The grade of inflammation was scored as follows: 1) degree of cell proliferation: 0, none; 1, mild cell number increase and crypt length; 2, moderate cell number increase or focally marked increase; 3, marked increase in entire field of section. 2) severity of inflammation: 0, none; 1, mild lymphocyte infiltration; 2, massive lymphocyte infiltration or visible focal degeneration of crypts; 3, multifocal crypt degeneration and/or tissue structure destruction. 3) extent of inflammation: 0, none; 1, to the mucosal layer; 2, to the submucosal layer; 3, to the transmural layer. 4) amount of mucus: 0, normal; 1, slight decrease of mucus; 2, mild decrease or focal absence of mucus; 3, severe absence of mucus; 4, total absence of mucus. The cumulative histological score was calculated as the sum of the four individual parameters.\n\nELISA\n-----\n\nThe levels of cytokines in the supernatants were determined using ELISA kits according to the manufacturer's protocols. ELISA kits for IL-6, IL-17A, and TNF-\u03b1 were from eBioscience (San Diego, CA, USA). ELISA kits for IL-4, IL-10, IFN-\u03b3, and TGF-\u03b2 were from R&D Systems (Minneapolis, MN, USA).\n\nImmunohistochemistry\n--------------------\n\nColons were removed from euthanised mice 4 weeks after cell transfer. The transverse colons were fixed overnight in 4% paraformaldehyde, dehydrated, infiltrated with paraffin and sectioned at 5\u2009\u03bcm. Immunohistochemistry was performed using standard protocols with citrate buffer (pH 6.0) pretreatment. Briefly, the sections were incubated with a primary antibody against IFN-\u03b3 at 4\u2009\u00b0C overnight and then with a corresponding horseradish peroxidase-conjugated secondary antibody at 37\u2009\u00b0C for 30\u2009min. Immunohistochemical detection of apoptosis was performed using a TUNEL assay kit (Promega, Madison, WI, USA) according to the manufacturer's protocol. The sections were finally counterstained with haematoxylin for detection.\n\nStatistical analysis\n--------------------\n\nThe data were shown as mean\u2009\u00b1\u2009standard deviations (SD). Student's *t* test was employed to determine significance between two groups (paired or unpaired) and One Way Anova analysis was used to determine significance among several groups. Differences were considered statistically significant when *P*\u2009\\<\u20090.05.\n\nEthics statement\n----------------\n\nAll experimental protocols used in this work were approved by the institutional review board of the Institute of Basic Medical Sciences.\n\nAdditional Information\n======================\n\n**How to cite this article**: Wang, X. *et al.* Endotoxic shock-expanded murine CD11c^low^CD45RB^+^ regulatory dendritic cells modulate inflammatory T cell responses through multiple mechanisms. *Sci. Rep.* **5**, 10653; doi: 10.1038/srep10653 (2015).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Figures\n\nThis work was supported by grants from the National Natural Science Foundation of China (81472736, 31270960 to J.Z.).\n\n**Author Contributions** J.Z. conceived and designed the study, analyzed the data, and wrote the paper, X.W., Q.W., X.Z., Y.L., J.W., C.H. and J.C. performed experiments, B.S. and Y.S. helped conceive the study.\n\n![Endotoxic shock promotes the expansion of CD11c^low^CD45RB^+^ cells. *A*, B6 mice were intra-peritoneally (i.p.) injected with *Escherichia coli* (*E. coli*) K12 or thioglycolate. Four days later, mice were euthanised and the splenic cells were subjected to flow cytometry analysis of CD11c and CD45RB expression. ns, not significant. *B*, B6 mice were i.p. injected with cholera toxin CHX, (5\u2009\u03bcg/mouse) or PBS of equal volume, followed by i.p. injection of *E. coli* K12 one hour later. Mice were euthanised at the indicated time points, and the splenic cells were subjected to flow cytometry analysis of CD11c and CD45RB expression. *C* and *D*, BALB/c mice were i.p. injected with *E. coli* K12 or *E. coli*-derived LPS. The splenic cells were subjected to flow cytometry analysis of CD11c and CD45RB expression 5 days later.](srep10653-f1){#f1}\n\n![Phenotypic characterisation of the expanded CD11c^low^CD45RB^+^ cells. BALB/c mice were intra-peritoneally (i.p.) injected with *E. coli*-derived LPS. Five days later, mice were euthanised and the splenic cells were subjected to the following analyses. ***A***, ***C***, and ***E***, The expression of lineage markers and functional molecules on the surface of different subpopulations of CD11c^+^ cells was analysed by flow cytometry. The black lines represent cells stained with isotype antibodies. *B*, *D*, and *F*, CD11c^low^CD45RB^+^ cells or CD11c^hi^CD45RB^\u2212^ cells were purified, the morphology was analysed by Giemsa staining (***B***), cytokine profiles of the two subpopulations stimulated with or without 0.5\u2009\u03bcg/ml LPS for 24\u2009h were analysed by ELISA (***F***), and the expression of functional molecules on the surface of CD11c^low^CD45RB^+^ cells stimulated with 0.5\u2009\u03bcg/ml LPS for 24\u2009h was analysed by flow cytometry. The black lines represent cells stained with isotype antibodies (***F***).](srep10653-f2){#f2}\n\n![The expanded CD11c^low^CD45RB^+^ cells induce apoptosis of activated CD4^+^ T cells *in vitro*. Purified CD11c^low^CD45RB^+^ cells or CD11c^hi^CD45RB^\u2212^ cells were co-cultured with CD4^+^ CD25^\u2212^ splenic cells at the ratio of 4:1. Stimulation was affected by the Dynabeads mouse CD3/CD28 T cell expander. *A*-*C*, The co-culture was subjected to Th0 (***A***), Th1 (***B***), or Tregs (***C***) condition. Five days later, the supernatants were subjected to ELISA (***A***). IFN-\u03b3, IL-17, Foxp3, or CD25 expression in CD4^+^ cells was analysed by flow cytometry (***B*** and ***C***). *D*, CD4^+^ CD25^\u2212^ splenic cells were labelled with CFSE before co-culture. Proliferation was assessed by flow-cytometric analysis of CFSE dilution 96\u2009hours later. The grey line represents cells with CFSE labelling but without proliferation. *E*, Apoptosis was assessed by flow-cytometric analysis of the percentages of CD4^+^ Annexin-V^+^ cells 96\u2009hours later. *Left*, representative data; *right*, statistical data (n\u2009=\u20093).](srep10653-f3){#f3}\n\n![Fas ligand (FasL) mediates the expanded CD11c^low^CD45RB^+^ cell-induced apoptosis of CD4^+^ T cell *in vitro*. ***A*** Purified CD11c^low^CD45RB^+^ cells and CD11c^hi^CD45RB^\u2212^ cells were subjected to real-time PCR. *B* and *C*, The levels of ROS (***B***) and cell surface markers (***C***) in different subpopulations of CD11c^+^ cells were analysed by flow cytometry. The grey lines represent background fluorescence (***B***) or staining with isotype antibodies (***C***). *D* and *E*, Purified CD11c^low^CD45RB^+^ cells and CD11c^hi^CD45RB^\u2212^ cells were co-cultured with CD4^+^ CD25^\u2212^ splenic cells at the ratio of 4:1. Stimulation was affected by Dynabeads mouse CD3/CD28 T cell expander in the presence or absence of neutralisation antibodies against programmed death ligand 2 (PD-L2) and FasL. Ninety-six hours later, cells were stained with anti-CD4-PECy5, PI, and Annexin V-FITC resuspended in 300\u2009\u03bcl binding buffer containing calcium ion. Apoptosis was assessed by flow-cytometric analysis of the percentages of CD4^+^ Annexin-V^+^ cells. *Left*, representative data; *right*, statistical data (n\u2009=\u20093) (***D***). Cells were collected using a FACSCalibur flow cytometer at high speed for 60\u2009seconds, and the number of CD4^+^ Annexin-V^\u2212^ cells was counted. The results are expressed as the percentage of basal growth of CD4^+^ T cells (***E***).](srep10653-f4){#f4}\n\n![The expanded CD11c^low^CD45RB^+^ dendritic cells (DCs) suppress colitis induction by CD4^+^ CD25^\u2212^ T cells *in vivo*. *A*, Relative changes in percent body weight over time for no-transfer mice (\u25ca, n\u2009=\u20099) and for colitis-induced mice (transfer of CD4^+^ CD25^\u2212^ T cells) subsequently treated with PBS (o, n\u2009=\u20097), CD11c^hi^CD45RB^\u2212^ cells (\u25a1, n\u2009=\u20095), or CD11c^low^CD45RB^+^ cells (\u2206, n\u2009=\u20098). *B*, Left, macroscopic findings of the colon on day 28 after transfer of CD4^+^ CD25^\u2212^ T cells. Right, the colon lengths of colitic mice were measured on day 28. *C*, Top, histological analysis of the colon was carried out on day 28. Scale bar: 50\u2009\u03bcM. Bottom, histological scores of the colons from CD4^+^ CD25^\u2212^ T cell-transferred mice. The results were interpreted independently by two pathologists who were not given previous information regarding the treatment of the mice. Two independent experiments gave similar results.](srep10653-f5){#f5}\n\n![Administration of the expanded CD11c^low^CD45RB^+^ dendritic cells (DCs) reduces proinflammatory cytokine responses *in vivo*. Mice were treated as described in [Fig. 5A](#f5){ref-type=\"fig\"}. The percentages of Foxp3^+^ CD4^+^ (***A***) or IFN-\u03b3^+^ CD4^+^ (***B***) splenic T cells were analysed by flow cytometry on day 28. Mesenteric lymph node (MLN) cells (1\u2009\u00d7\u200910^6^) of colitic mice were cultured for 24\u2009h in 24-well plates in the presence of 1\u2009\u03bcg/ml CD3 and 1\u2009\u03bcg/ml CD28. The culture supernatants were then harvested and subjected to ELISA (***C***). The colon samples of colitic mice were subjected to immunohistochemical staining with a primary antibody against IFN-\u03b3. Scale bar: 50\u2009\u03bcM (***D***).](srep10653-f6){#f6}\n\n![Administration of expanded CD11c^low^CD45RB^+^ dendritic cells (DCs) induces T cell apoptosis *in vivo*. Mice were treated as described in [Fig. 5A](#f5){ref-type=\"fig\"}. The colon samples of colitic mice were subjected to terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) staining and the apoptosis in the Payer's patches (***A***) and the colonic tissues (***B***) was shown. Scale bar: 50\u2009\u03bcM.](srep10653-f7){#f7}\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Background\n==========\n\nTylosin, a member of the macrolide group of antibiotics, is widely used in the treatment of canine chronic diarrhea of unknown etiologies \\[[@B1]\\],\\[[@B2]\\]. Currently, oral tylosin treatment is not approved for the treatment of canine enteritis, but it may be prescribed as an *ex tempore* drug, dispensed by the pharmacy on request. Consequently, no official recommendations are available for the oral dosage regimen in dogs, and the literature shows that individual authors' recommended dosages differ.\n\nIn 1976, Van Kruiningen drew attention to the usefulness of tylosin in treating enterocolitis in dogs in a study about the clinical efficacy of tylosin in canine inflammatory bowel disease \\[[@B3]\\]. This study comprised 27 dogs receiving a daily oral dosage of tylosin between 11--200\u00a0mg/kg. The dosing interval was twice daily, and the duration of treatment varied from one week to five years. Two decades later, a case--control study involved seven pet dogs with secondary chronic diarrhea due to naturally occurring exocrine pancreatic insufficiency. The dogs received tylosin for 5--7 days at a dose of 15\u00a0mg/kg twice daily as an adjunctive medication to the pancreatic enzyme \\[[@B4]\\]. In 2005, Westermarck et al. introduced the term tylosin-responsive diarrhea (TRD) in dogs in a case series involving nine client-owned dogs suffering from recurrent diarrhea and responding to tylosin at a dose of 6--16\u00a0mg/kg once daily for two weeks \\[[@B5]\\]. Characteristic of TRD was that the dogs with recurrent diarrhea, for which no underlying cause could be found, usually responded to tylosin within a few days. The stool remained normal as long as treatment continued, but diarrhea recurred in many dogs within weeks of discontinuing the treatment. After initiating a repeated course, however, the dogs' response to tylosin was as good as the initial treatment, even after several treatments.\n\nA recent randomized controlled trial of the effect of tylosin on fecal consistency in dogs with recurrent diarrhea revealed that tylosin proved significantly more effective than placebo. At a dose of 25\u00a0mg/kg once daily for seven days, tylosin proved effective in controlling diarrhea in 20 client-owned dogs \\[[@B6]\\].\n\nHowever, despite the wide use and acceptance of tylosin in the treatment of chronic enteropathies in dogs, a clear consensus about the dosage regimen is lacking. In textbooks, the dose recommendations for tylosin in dogs with chronic diarrhea vary considerably, with daily dosages ranging from 25 to 80\u00a0mg/kg, and dosing intervals from one to three times daily \\[[@B2]\\],\\[[@B7]\\],\\[[@B8]\\]. These textbooks contain little information about the original source of this information. To date, no published dosage determination studies are available on the oral treatment of dogs with chronic diarrhea by tylosin.\n\nTherefore, the primary objective of the current study was to determine whether tylosin doses of 5\u00a0mg/kg or 15\u00a0mg/kg administered orally once daily for seven days have a similar effect on the fecal consistency in diarrhea relapses to that of 25\u00a0mg/kg dose of tylosin once daily for seven days, a dosage that has proved effective in controlling canine TRD. A further objective was to compare the efficacy of the 5\u00a0mg/kg and 15\u00a0mg/kg dosages.\n\nSecondary objectives were to assess the recurrence rate and the time to recurrence of diarrhea at a tylosin dosage of 25\u00a0mg/kg once daily over a seven-day treatment duration. Further objectives were to investigate the influence of the three different tylosin dosages on the time to the cessation of diarrhea.\n\nMethods\n=======\n\nTrial design, compliance and ethics approval\n--------------------------------------------\n\nThis trial was designed according the guidelines of evidence-based medicine \\[[@B9]\\] and reported utilizing the Consolidated Standards of Reporting Trials (CONSORT) \\[[@B10]\\]-\\[[@B13]\\]. The study design is a prospective, single-blinded, two-arm parallel clinical dosage determination field trial. The study complied with the European Economic Community Guideline on the demonstration of efficacy for veterinary medicinal products containing antimicrobial substances. Both the Ethics Committee for Animal Experiments of the University of Helsinki in Finland and the National Animal Experiment Board in Finland approved the study protocol. The owners provided their written informed consent in which they agreed to participate with their dog.\n\nStudy population, inclusion and exclusion criteria\n--------------------------------------------------\n\nThis study was conducted at the Small Animal Teaching Hospital, Faculty of Veterinary Medicine, University of Helsinki, Finland. The study population was recruited from the patient material of a placebo-controlled, randomized, double-blinded, prospective clinical trial on the effectiveness of oral tylosin in treating recurrent diarrhea in dogs, that was conducted just prior to this current dosage determination trial \\[[@B6]\\]. Dogs were considered eligible for the oral tylosin dosage determination trial if they responded to a tylosin treatment at a dosage of 25\u00a0mg/kg once daily for seven days in the aforementioned trial. Inclusion criteria were that the dogs were at least six months old and had been suffering from recurrent diarrhea responding to tylosin treatment. They were not allowed to have received systemic corticosteroids, nonsteroidal anti-inflammatory drugs, or antibiotics other than tylosin in the 90\u00a0days preceding the tylosin dosage determination trial. Lactating bitches were excluded, as were dogs with evidence of systemic or organ-related disease that could secondarily cause diarrhea.\n\nThe dogs underwent an initial examination prior to enrollment in the tylosin effectiveness trial to ensure that they met the selection criteria. In brief, the initial examination covered the dogs' clinical history and physical examination, blood samples for a complete blood count and serum biochemical profile, urinalysis, fecal analysis for endoparasites and enteropathogenic bacteria, and gastroduodenoscopy with mucosal biopsies from the duodenum. The protocol and the results of the initial examination have been published previously \\[[@B6]\\].\n\nOutcome measures\n----------------\n\nThe primary objective of the current trial was to assess whether daily doses of 5\u00a0mg/kg or 15\u00a0mg/kg tylosin at a once daily dosing interval for seven days would have similar effects on the fecal consistency in diarrhea relapses to that of an initial dose of 25\u00a0mg/kg tylosin once daily for seven days in these dogs, a dosage that has proved effective in controlling canine TRD \\[[@B6]\\]. Further to assess whether the tylosin dose 5\u00a0mg/kg was as effective as the 15\u00a0mg/kg dose. To evaluate this, the elimination of diarrhea served as the main criterion. The elimination of diarrhea was fulfilled when noting that on three consecutive days the fecal score for each of the dog's feces was and remained three or less. A mean fecal consistency score assigned during the last three days of the seven-day treatment period served as the primary outcome measure. To determine the mean fecal consistency score, the owners evaluated and recorded throughout the study the consistency of each of their dog's stool according to a standardized fecal scoring system previously established by Moxham \\[[@B14]\\]. This kind of fecal scoring system has proved useful in evaluating treatment success in diarrhea patients \\[[@B5]\\],\\[[@B6]\\],\\[[@B15]\\]-\\[[@B17]\\]. In brief, the fecal scoring system is based on a nine-point scale, consisting of scores from one to five, with half-point intervals. To ensure uniformity in the fecal consistency scoring, it comprises both a verbal description and a visual picture for each of the fecal consistency scores. Scores 1 and 1.5 represented feces of a hard and dry consistency. Scores 2 and 2.5 indicated well-formed feces that left no mark when picked up. A score of 3 represented feces of a slightly moist consistency, whereas score 3.5 represented feces of a moist consistency, but which still had some definite form. Scores 4 through 5 described feces of poor consistency.\n\nTo evaluate whether the tylosin dosage was effective, the investigator calculated a mean fecal consistency score of the scores assigned for each of the dog's feces during the last three days of the seven-day treatment period based on the owner's records. A responder was defined as having a mean fecal consistency score of three or less; a non-responder had a score of more than three. The owners were unaware of these definitions.\n\nSecondary outcome measures were defined in order to evaluate the recurrence rate and the time to recurrence of diarrhea at a 25\u00a0mg/kg dosage of tylosin. The recurrence of diarrhea was defined as when the dog's feces had a consistency score of at least four on the scale for at least two consecutive defecations. Further secondary outcome measures were defined to assess the influence of the three different tylosin dosages (5, 15 and 25\u00a0mg/kg) on the time of ceasing of diarrhea. The day on which the diarrhea ceased was defined as the day of the treatment period on which the fecal score for each of the dog's feces was and remained three or less.\n\nInterventions\n-------------\n\nWhen assessing the dogs' eligibility for this present dosage determination trial all dogs had diarrhea and began their tylosin treatment at a dose of 25\u00a0mg/kg once daily for seven days. The dog owners evaluated and recorded the consistency of each of their dogs' feces. On day seven of the treatment period each dog underwent a physical examination and its fecal consistency records were collected. The investigator recorded the day on which the dog's diarrhea ceased and calculated a mean fecal consistency score for each dog over the last three days of the treatment period at a tylosin dosage of 25\u00a0mg/kg; the dogs defined as responders were deemed eligible to participate in the dosage determination trial. Treatment at 25\u00a0mg/kg tylosin was then discontinued for the participating dogs and a follow-up period of up to two months served to determine whether the diarrhea recurred. No additional changes were permitted in the dog's feeding management or medication. Upon meeting the criteria for the recurrence of diarrhea (fecal score\u2009\u2265\u20094 for at least two consecutive defecations) each dog underwent a physical examination, during which the investigator also assessed and confirmed the fecal consistency by rectal palpation. The dogs then received tylosin (Tylosin tartrate 120\u00a0mg and 240\u00a0mg tablets, University Pharmacy, Helsinki, Finland; equivalent to 100\u00a0mg and 200\u00a0mg tylosin) once daily for seven days in doses of either 5\u00a0mg/kg or 15\u00a0mg/kg.\n\nDuring the seven-day treatment period, the owners continued evaluating each of their dog's stools and recorded its fecal consistency according to the guidelines. On day three, the investigator phoned the owner to assure whether the clinical condition of the dog met the criteria of the humane endpoint of the study and whether the dog had to be released from the study and treated accordingly. A humane endpoint was defined as the point at which the dog's clinical condition was poor and it suffered from excessive diarrhea. On day seven, the dog underwent a second physical examination, including rectal palpation and an evaluation of the dog's fecal consistency. The fecal consistency records were collected, and a mean fecal consistency score was calculated for each dog based on the last three days of the treatment period at the 5\u00a0mg/kg or 15\u00a0mg/kg dosages.\n\nAllocation to treatment groups\n------------------------------\n\nAfter the onset of diarrhea within the two-month follow-up period the dogs received an ascending study number denoting the order in which the diarrhea randomly recurred. The investigator then assigned the dogs in a 1:1 ratio to two different dosage groups - either 5\u00a0mg/kg or 15\u00a0mg/kg - receiving tylosin orally once daily for seven days. All dogs with odd study numbers were assigned to the 5\u00a0mg/kg dosage group, and those with even study numbers to the 15\u00a0mg/kg dosage group.\n\nBlinding\n--------\n\nThe study was single-blinded and thus open to the investigator but blinded for the dog owner. The blinding procedure was carried out by labeling the 120\u00a0mg tylosin tartrate tablet bottles as \"Tylosin tart. S\" and the 240\u00a0mg tylosin tartrate tablet bottles as \"Tylosin tart. V\". The tablets were visually identical.\n\nStatistical methods\n-------------------\n\nDescriptive statistics were used to summarize the data. The Shapiro-Wilk test served to determine whether the data were normally distributed; non-parametric data were expressed as median and range. The Wilcoxon signed rank test was used for non-parametric data to compare the mean fecal consistency scores of the last three days of the treatment period at the 25\u00a0mg/kg tylosin dosage to the scores at the 15\u00a0mg/kg and 5\u00a0mg/kg tylosin dosages. The Mann--Whitney *U* test was used to examine differences between the mean fecal consistency scores in the 5\u00a0mg/kg and 15\u00a0mg/kg tylosin dosage groups. To evaluate the effect of each dosage on the time until the diarrhea ceased, the Wilcoxon signed rank test and the Mann--Whitney *U* test were used. The level of statistical significance was set at *P*\u2009\\<\u20090.05; all statistical analyses were performed with a commercially available statistical software system (SPSS 18.0 for Windows).\n\nResults\n=======\n\nParticipant flow and follow-up\n------------------------------\n\nBetween October 2006 and May 2008, 20 client-owned dogs responding to 25\u00a0mg/kg tylosin once daily for the seven day treatment in the tylosin effectiveness trial \\[[@B6]\\] were deemed eligible to participate in the present tylosin dosage determination trial. One owner declined to participate with the dog for personal reasons, and two dogs were excluded due to other concurrent medications that violated the inclusion criteria. A total of 15 of 17 dogs relapsed and diarrhea recurred within the two-month follow up period. The results of the enrollment, allocation and response to the two different tylosin dosages appear in Figure\u00a0[1](#F1){ref-type=\"fig\"}. It is noteworthy that in treating their recurrent diarrhea, 14 of 15 dogs (93%) responding to a dose of 25\u00a0mg/kg tylosin once daily also responded to the lesser dosages of 5\u00a0mg/kg or 15\u00a0mg/kg once daily for seven days. No adverse events or side effects were recorded at any of the dosages in any of the dogs. No deviations from the study protocol occurred during the trial, and no dog had to be released from the study due to excessive diarrhea or poor clinical condition.\n\n![**Flowchart of the trial.** Flowchart of the results of the study enrollment, the allocation to treatment groups, and the response to the two different tylosin dosages.](s13028-014-0043-5-1){#F1}\n\nBaseline data\n-------------\n\nThe 15 dogs included in the prospective treatment trial consisted of 12 different breeds: Golden Retriever (n\u2009=\u20093), German Shepherd (n\u2009=\u20092), and one of each of the following breeds (n\u2009=\u20091): Pumi, Rough Collie, Miniature Schnauzer, Kromfohrl\u00e4nder, Doberman Pinscher, Border Collie, German Shorthaired Pointer, Nova Scotia Duck-Tolling Retriever, Old English Sheepdog, and mixed breed. The dogs' ages at the time of enrollment ranged from seven months to four years and three months (median three years and eleven months). The weight of the dogs at the time of enrollment ranged from 10\u00a0kg to 37.5\u00a0kg (median 25.2\u00a0kg). The study population included nine intact males and six females, three of which were intact.\n\nPrimary outcomes\n----------------\n\nFigure\u00a0[2](#F2){ref-type=\"fig\"} shows the mean fecal consistency scores assigned during the last three days of the treatment period for each dog participating in the trial at tylosin doses of 5\u00a0mg/kg (n\u2009=\u20098), 15\u00a0mg/kg (n\u2009=\u20097), and 25\u00a0mg/kg (n\u2009=\u200915). Figure\u00a0[3](#F3){ref-type=\"fig\"} shows the minimum and maximum, the upper and lower quartiles, interquartile range and the median of the mean fecal consistency scores at all three tylosin dosages. The median of the mean fecal consistency scores from the last three days of the treatment period in the dogs receiving tylosin at the 25\u00a0mg/kg dosage was 2.5 (95% confidence interval, 2.32; 2.65). In the dogs receiving tylosin at the 15\u00a0mg/kg dosage, the median of the mean fecal consistency scores from the last three days of the treatment period was likewise 2.5 (95% confidence interval, 2.10; 3.36). The median of the mean fecal consistency scores from the last three days of the treatment period in the dogs receiving 5\u00a0mg/kg tylosin was 2.63 (95% confidence interval, 2.36; 2.82). The mean fecal consistency scores for the 25\u00a0mg/kg dosage of tylosin showed no significant difference from scores for the 15\u00a0mg/kg or 5\u00a0mg/kg dosages of tylosin (Wilcoxon signed rank test, *P*\u2009=\u20090.345, *P*\u2009=\u20090.672). Nor was the difference in the mean fecal consistency scores between the dogs receiving 15\u00a0mg/kg tylosin and those receiving 5\u00a0mg/kg tylosin significant (Mann--Whitney *U* test, asymptotic *P*\u2009=\u20090.370 and exact, not corrected for ties *P*\u2009=\u20090.397).\n\n![**Mean fecal consistency scores per dog.** Mean fecal consistency scores at dosages of 25\u00a0mg/kg (n\u2009=\u200915), 15\u00a0mg/kg (n\u2009=\u20097), and 5\u00a0mg/kg (n\u2009=\u20098) tylosin per dog assessed during the last three days of the seven-day treatment period.](s13028-014-0043-5-2){#F2}\n\n![**Box and whisker plots of the mean fecal consistency scores.** Box and whisker plots of the mean fecal consistency scores assessed during the last three days of the seven-day treatment period at tylosin dosages of 25\u00a0mg/kg (n\u2009=\u200915), 15\u00a0mg/kg (n\u2009=\u20097) and 5\u00a0mg/kg (n\u2009=\u20098). The limits of the box represent the 25^th^ and 75^th^ percentile values, the line within the box represents the median values and the whiskers represent the range. The black line indicates the cut-off value for a responder, set at a score of 3 on the mean fecal consistency score scale.](s13028-014-0043-5-3){#F3}\n\nSecondary outcomes\n------------------\n\nAt a dosage of 25\u00a0mg/kg tylosin and a treatment duration of seven days, 15 of 17 dogs relapsed, and diarrhea recurred within the two-month follow-up period on a median of day 9 (range 1--33), resulting in a diarrhea recurrence rate of 88%.\n\nThe day on which the diarrhea ceased in the dogs when receiving the 25\u00a0mg/kg dosage of tylosin (n\u2009=\u200915) ranged from days 1 to 5 (median day 3; 95% confidence interval, 2.20; 3.40). At the 15\u00a0mg/kg dosage of tylosin (n\u2009=\u20097), the day on which diarrhea ceased ranged from days 2 to 5 (median day 2; 95% confidence interval, 1.68; 3.82) of the treatment period, and at the 5\u00a0mg/kg dosage of tylosin (n\u2009=\u20098) from days 1 to 5 (median day 2.5; 95% confidence interval, 1.40; 3.94). The number of days until the diarrhea ceased in the dogs receiving 5\u00a0mg/kg tylosin compared to those receiving 15\u00a0mg/kg tylosin was not statistically significantly different (Mann--Whitney *U* test, asymptotic *P*\u2009=\u20090.835).\n\nDiscussion\n==========\n\nThis study was conducted to evaluate the efficacy of two low dose oral tylosin treatment regimens to control the relapse of diarrhea in dogs suffering from TRD. Our results showed that 14 of the 15 dogs (93%) suffering from recurrent diarrhea that previously responded to oral tylosin therapy at a dose of 25\u00a0mg/kg once daily for seven days, responded to the lower doses of 5\u00a0mg/kg and 15\u00a0mg/kg once daily for seven days after diarrhea relapse. These data indicate that a suitable daily oral dose of tylosin for the treatment of diarrhea relapses in canine TRD may thus far be less than the current recommendations in textbooks, which vary from 25 to 80\u00a0mg/kg \\[[@B2]\\],\\[[@B7]\\],\\[[@B8]\\]. The limitations of this study were its rather small sample size, as well as lack of randomization and controls. Nevertheless, our results are in agreement with those of a previous TRD study in which dog owners reduced the tylosin dose for their dogs as low as possible (6--16\u00a0mg/kg per day) while maintaining their dogs' health \\[[@B5]\\]. In both the aforementioned and the present studies, however, all dogs had previously received several oral tylosin treatments, so whether the efficacy would remain the same had the dog initially received tylosin at low dosages remains uncertain. The groups receiving 5\u00a0mg/kg and 15\u00a0mg/kg tylosin were not independent of the 25\u00a0mg/kg tylosin group. Still a statistical comparison was carried out to investigate whether the mean fecal consistency scores after the low dose treatments for diarrhea relapse were similar to the scores at the initial treatment and to compare only the treatment outcomes, not its efficacy.\n\nIn the present study, one dog failed to respond to the 15\u00a0mg/kg once daily dosage, but did respond the 25\u00a0mg/kg once daily dosage, possibly because this time the cause of diarrhea in this dog may have differed. Since the owners declined further diagnostic investigations in this case, the reasons for the non-response remain uncertain.\n\nAlthough pharmacokinetic studies have explored the intramuscular and intravenous administration of tylosin in dogs \\[[@B18]\\], studies on per oral administration are lacking. Consequently, the dosing interval of per oral administration remains unknown. Individual authors' recommendations in textbooks for the dosing interval of tylosin differ, ranging from one to three times daily \\[[@B2]\\],\\[[@B7]\\],\\[[@B8]\\]. Our data show that a once daily dosing interval is sufficient to stop diarrhea in 14/15 (93%) of the participating dogs during the treatment period. This finding is consistent with those of earlier reports on TRD in which a once-daily dosing interval proved successful in controlling diarrhea in dogs with recurrent diarrhea \\[[@B5]\\],\\[[@B6]\\],\\[[@B15]\\], and shows that a more frequent dosing interval is unnecessary in the treatment of diarrhea relapses in TRD dogs. A once daily dosing makes the treatment cost effective and more convenient for both the dog and the owner.\n\nControversy also surrounds recommendations regarding the duration of tylosin therapy in chronic enteropathies in dogs. So far, recommendations for the duration of antibiotic treatment in antibiotic-responsive diarrhea can range up to six weeks, because premature cessation of treatment could lead to relapse \\[[@B1]\\]. However, to the best of our knowledge, no studies have yet investigated whether the treatment duration affects the recurrence rate. All of the dogs in this present trial received tylosin for seven days at a dosage of 25\u00a0mg/kg before it was terminated. During the two month follow-up 88% of the dogs relapsed. This is a higher recurrence rate than the one in our previous study, where diarrhea recurred in only 43.3% of the dogs investigated \\[[@B6]\\]. In the aforementioned study, the duration of tylosin treatment varied between the dogs. However, the high prevalence of recurrence is in accordance with that of an earlier report on TRD in which the low dosage tylosin treatment lasted for at least six weeks. When the treatment was terminated, the diarrhea recurred within 30\u00a0days in 85.7% of the dogs \\[[@B5]\\]. Since an at least six-week tylosin course in TRD dogs has not lead to smaller relapse rates than a seven-day course, we conclude that a seven-day treatment period may prove sufficient in the treatment of TRD in dogs. One limitation of this study was that we were unable to follow the dogs beyond the end of this trial to determine whether, after receiving the low dosages, they showed signs of diarrhea relapse and whether the time interval between the cessation of treatment and the relapse of clinical signs depended on the dose or duration of the treatment. Additional studies are warranted to further assess the effect of treatment duration on diarrhea relapse.\n\nAt a dosage of 25\u00a0mg/kg tylosin and treatment duration of seven days diarrhea recurred in the dogs of this trial on a median of day 9, which is in accordance with a former study on TRD where at tylosin dosages varying individually from 6 to 16\u00a0mg/kg, the diarrhea reappeared on a median day of 7 \\[[@B5]\\].\n\nUntil now, some have suggested that in dogs with antibiotic-responsive diarrhea who have shown a suboptimal response, the antibiotic should be changed after the first two weeks \\[[@B1]\\]. In the present study the diarrhea ceased in all dogs by day five of the tylosin treatment. The diarrhea ceased in the dogs receiving 5, 15 and 25\u00a0mg/kg tylosin on a median day of 2.5, 2 and 3, respectively. This result falls within the same time frame as those published in earlier studies on TRD \\[[@B5]\\],\\[[@B15]\\] and indicates that in the absence of a response after a few days of tylosin treatment, the treatment should be terminated and further diagnostics introduced as soon as possible. In small animal practice the uncritical use of antibiotics, including tylosin, contributes to the development of resistant intestinal bacterial flora which also poses a risk to public health. Thus, until alternative treatment options are found, one should aim to restrict the usage of tylosin in chronic enteropathies in dogs when systemic disorders have been ruled out and, despite thorough investigations, no other underlying primary intestinal disorder is evident.\n\nConclusions\n===========\n\nIn conclusion, our results indicate that a suitable oral dosage of tylosin in the treatment of diarrhea relapses in canine tylosin-responsive diarrhea could be as low as 5\u00a0mg/kg once daily for seven days.\n\nAbbreviation\n============\n\nTRD: Tylosin-responsive diarrhea\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nSK made the practical arrangements, carried out the clinical examinations and data collection and drafted the manuscript. TS made a substantial contribution to the design of the study and helped draft the manuscript. EW made a substantial contribution in the design of the study and assisted significantly in drafting the manuscript. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nWe acknowledge Dr. Katja Mustonen and Dr. Nina Menna of Vetcare for their cooperation during the study as well as Dr. Merja Rantala for improving the manuscript with her constructive comments.\n\nThe study was funded by Vetcare Ltd., M\u00e4nts\u00e4l\u00e4, Finland and supported by both the Finnish Veterinary Foundation and the Finnish Foundation of Veterinary Research.\n\nPart of the work was presented as an abstract at the 2010 ACVIM Forum, Comparative Gastroenterology Society Session, Anaheim, California, USA.\n"}
+{"text": "Facts\n=====\n\nNatural polyphenolic compounds that are present in our diet, such as rottlerin, genistein, quercetin, curcumin, and resveratrol can alter the effects of signaling pathways and induce cell death not only via apoptosis but also via autophagy. Thus, these compounds could be used as a co-therapy with standard therapies in cancer.These compounds can trigger type II PCD via various mechanisms through the canonical (Beclin-1 dependent) and non-canonical (Beclin-1 independent) routes of autophagy.Rottlerin or its related analogs may be used in the development of novel agents for the induction of autophagic cell death as it has been proven, pharmacodynamically in a mice xenograft model, to be efficiently absorbed in cells and tissues against pancreatic cancer.Genistein induced autophagy due to changes in apoptotic signaling, which is beneficial against chemoresistance usually seen in cancer cells.Quercetin induced extensive autophagy and subsequent death in cancer cells mediated by the inhibition of proteasomal activity and mTOR signaling.Curcumin induced G~2~/M arrest and autophagy in malignant glioma cells through the inhibition of the Akt/mTOR/p70S6K and activation of the extracellular signal-regulated kinase (ERK)1/2 pathways, which implied that cell death via autophagy, might be pathway specific.Resveratrol induced cell death through autophagy in five ovarian cancer cell lines, suggesting that it may be effective treatment in apoptosis-resistant ovarian cancer. Autophagy can be induced with acute exposure to resveratrol, whereas prolonged exposure activates a caspase-mediated cell death pathway.\n\nOpen Questions\n==============\n\nThe induction of cellular senescence was accompanied by autophagy in colon cancer cells with an increase in Beclin-1 and p62/SQSTM1 protein levels. Therefore, the functional link between senescence and autophagy in these curcumin-treated cancer cells should be further investigated.Autophagy inhibitors may have the potential to enhance resveratrol antitumor efficacy.The ability of natural polyphenolic compounds to induce autophagic cell death that enhances the effects of standard therapies should be taken into consideration for designing novel therapeutic strategies.Combining FDA (food and drug activation)-approved drugs with these polyphenolic compounds such as rottlerin, genistein, quercetin, curcumin, and resveratrol may provide novel therapeutic strategies in the treatment of cancer.\n\nNatural plant-derived polyphenols are chemical substances characterized by the presence of more than one phenol unit per molecule. They are present in some foods and have been shown to exert anticancer properties. Some important examples are rottlerin, genistein, quercetin, curcumin, and resveratrol, all of which have been shown to induce autophagy death in various cancer cells ([Figures 1](#fig1){ref-type=\"fig\"},[2](#fig2){ref-type=\"fig\"},[3](#fig3){ref-type=\"fig\"} and [Table 1](#tbl1){ref-type=\"table\"}).\n\nMacroautophagy (hereafter called autophagy) is a lysosomal catabolic process conserved through evolution in eukaryotes for degrading long-lived proteins, macromolecules, and organelles from the cytoplasm.^[@bib1],\\ [@bib2]^ Autophagy has different roles in normal and cancer cells, especially in the tumor microenvironment. Although genetic evidence indicates that autophagy functions as a tumor suppressor in normal cells, it can promote survival of established tumors in the presence of cellular stress factors, including nutrient deprivation, hypoxia, metabolic, and therapy-induced stress.^[@bib3]^ If excessively induced, autophagy can lead to a non-apoptotic form of programmed cell death (type II PCD) which is caspase independent. Autophagy may be seen in cells that have a high threshold for induction of apoptosis or defective apoptotic machinery, such as inappropriate regulation of pro- and antiapoptotic Bcl-2 family member proteins.^[@bib4]^\n\nAutophagy requires the sequestration of cytoplasmic content or organelles through the formation of double-membrane vesicles, controlled by autophagy-related genes (*ATG*) and their protein products. Beclin-1, one of the major Atg proteins (its counterpart in yeast is known as Atg6), controls vital steps in the autophagic pathway. It is interesting that autophagy can be regulated by various canonical and non-canonical pathways, some involving Beclin-1,^[@bib5],\\ [@bib6]^ others Beclin-1 dependent and Atg5/Atg7 independent,^[@bib7]^ Beclin-1/Vps34 independent,^[@bib8],\\ [@bib9]^ or Beclin-1 independent and Vps34 dependent.^[@bib10],\\ [@bib11]^ Studies characterizing metabolic cascades known to have a role in autophagy involved common pathways, such as PI3K/Akt/mTOR, NF-*\u03ba*B, MAPK/MEK/ERK, protein kinase C delta (PKC\u03b4)/ transglutaminase 2 (TG2), JNK/p62 /SQSTM1, and AMPK/TSC2, are reviewed here to gain insight into the regulation of autophagy ([Table 1](#tbl1){ref-type=\"table\"} and [Figure 3](#fig3){ref-type=\"fig\"}).\n\nIn this review, which is an update to previous reports by our group,^[@bib3],\\ [@bib12],\\ [@bib13]^ we look into the effects of five common dietary polyphenols -- rottlerin, genistein, quercetin, curcumin, and resveratrol -- on autophagy in several cancer types and their regulation of the various metabolic cascades that leads to the induction or inhibition of this alternative cell death mechanism.\n\nAutophagic pathways\n===================\n\nAutophagy is a multistep process involving induction, phagophore formation, sequestration, formation of autophagosome, and finally its fusion with lysosome to form autophagolysosome, which then induces either the death or survival pathway ([Figure 2](#fig2){ref-type=\"fig\"}). The autophagic pathway dependent on Beclin-1 defines the canonical pathway; the other, independent of Beclin-1, is accepted as non-canonical autophagy. The presence of non-canonical autophagy means that autophagy can be triggered in cells in which the expression of Beclin-1 is too low to induce canonical autophagy.\n\nCanonical/classic pathway\n-------------------------\n\n*Beclin-1*, an autophagy-promoting gene, was determined to function as a tumor suppressor after it was shown to be deleted in breast, ovarian, and prostate cancers.^[@bib5]^ Beclin-1 forms a core complex known as PtdIns3K class III with Vps34 and Vps15 regulatory subunits ([Figures 2](#fig2){ref-type=\"fig\"} and [3](#fig3){ref-type=\"fig\"}), and this complex has been shown to promote autophagy by mediating autophagosome nucleation.^[@bib14]^ Within the Beclin-1-dependent autophagy are two alternative downstream pathways, the more common, Atg5/Atg7 dependent and the other, Atg5/Atg7 independent.^[@bib7]^\n\nNon-canonical/non-classic pathway\n---------------------------------\n\nIn contrast to canonical, non-canonical autophagy is a process that does not require the entire set of Atg proteins to form the autophagosome. This alternative form of autophagy, which is independent of Beclin-1 and, typically Vps34, has been described elsewhere.^[@bib8],\\ [@bib15]^ It is important to note that Beclin-1-independent autophagy does not always imply independence from Vps34.^[@bib10],\\ [@bib11]^ A common hallmark of this Beclin-1 independent autophagy is its dependence on the activity of the UNC-51 like kinase 1/2 complex to induce autophagy and LC3 for phagophore formation ([Figures 2](#fig2){ref-type=\"fig\"} and [3](#fig3){ref-type=\"fig\"}). The stimuli that trigger these forms of autophagy and the various interacting proteins involved in the formation and maturation of autophagosomes are not completely understood.\n\nAutophagy and cancer\n====================\n\nDefects in autophagy alter cells\\' metabolic state and their capacity for protein degradation, and are associated with various disease conditions.^[@bib16],\\ [@bib17],\\ [@bib18]^ Malignant cells often display defective autophagic activities compared with their normal counterparts. However, the role of autophagy in cancer is paradoxical in normal and malignant cells, and available data suggest that autophagy can promote tumor growth by helping tumor cells survive and inhibit tumor growth through its tumor-suppressor function in normal cells.^[@bib19]^ Removal of damaged proteins and organelles may prevent cancer initiation while allowing established tumors to adapt to nutrient-deprived or hypoxic conditions during cancer progression.^[@bib20]^\n\nInitial genetic evidence in mice suggested that *Beclin-1* functions as a haplo-insufficient tumor suppressor and that its mono-allelic deletion leads to spontaneous tumors and upon re-expression, it restores autophagy and suppresses tumorigenesis.^[@bib21],\\ [@bib22]^ Although Beclin-1 and LC3 are important mediators of autophagy, other molecules and signaling pathways (for example, p53, PI3K/AKT/mTOR) require critical examination to determine their roles in cells\\' autophagic capacity toward cell death or survival following various triggers.\n\nOn the other hand, the antiapoptotic Bcl-2 family members from the endoplasmic reticulum and not the mitochondria, such as Bcl-2 and Bcl-xL, inhibit autophagy by binding to the BH3 domain of Beclin-1.^[@bib23]^ Furthermore, recent data suggest that the oncogenic effect of Bcl-2 arises from its ability to inhibit autophagy but not apoptosis. Inhibition of Bcl-2 leads to autophagic cell death in MCF7 breast cancer cells.^[@bib24]^ Therefore these evidences suggest that modulating autophagy may be important in designing anticancer therapies.^[@bib22]^ Finally, studies suggest that the genetic makeup of cells will determine its fate in terms of undergoing autophagy and response to standard or novel therapies.\n\nSignaling pathways regulating autophagy\n=======================================\n\nDisruption of the normal balance between pro- and anti-autophagic signaling pathways is linked to cancer and other diseases. Several signaling pathways kinases, such as mTOR, AMPK, PI3Ks, MAPKs (ERK and JNK), and PKC, respond to various external factors and are often dysregulated in cancer ([Figure 3](#fig3){ref-type=\"fig\"}). These kinases may be part of an energy-sensing mechanism and stress response. For example, the mTOR kinase is an important repressor of autophagy and controller of cell growth and proliferation. Therefore, it is not surprising that it is regulated by multiple signaling pathways; it is activated by the Akt/PKB proto-oncogenic pathway and often upregulated in various cancer types.^[@bib25]^\n\nThere are two mTOR complexes, mTORC1 and mTORC2, which are regulated differently even though they are both induced in response to nutrient starvation, stress, and reduced growth factor signaling. TSC 1/2, an inhibitor of mTOR, is the gateway by which other signaling pathways influence mTOR activity. Alexander *et al.*^[@bib26]^ provided evidence that the ATM protein signals to TSC2 via the AMPK metabolic pathway in the cytoplasm to repress mTORC1 in response to reactive oxygen species (ROS), thereby inducing autophagy. Alternatively, the mTOR kinase can be activated by Akt independently of TSC 1/2 and can be regulated by MAPKs. mTORC2 is rapamycin insensitive and activates Akt by phosphorylation, to contribute to another route in the regulation of autophagy.^[@bib25],\\ [@bib27]^\n\nWong *et al.*^[@bib28]^ observed the activation of ERK and JNK as upstream effectors controlling autophagy induced by ROS production. Wang *et al.*^28^ reported a non-canonical MEK signaling pathway positioned downstream of AMPK and upstream of TSC that mediates autophagy via regulation of Beclin-1.^[@bib6]^ Puissant and Auberger^[@bib29]^ observed the activation of AMPK/mTOR- and JNK-mediated p62/SQSTM1 in triggering autophagic cell death. AMPK, which promotes autophagy by activating TSC 1/2 and suppressing mTOR, has been implicated in an energy check point because of its roles in phosphorylating p53 and inducing cell cycle arrest following energy deprivation.^[@bib30]^ We previously reported that PKC\u03b4 positively regulates the expression of TG2,^[@bib13],\\ [@bib14]^ which leads to the suppression of autophagy through various downstream pathways, including NF-*\u03ba*B, PI3K/Akt/mTOR/p70S6K and Bcl-2 anti-autophagic/apoptotic protein.^[@bib3],\\ [@bib12]^ Hence, studies suggest that activated or inhibited signaling pathways that are involved in the regulation of autophagy will direct it in cancer cells.\n\nClinical significance of autophagy and its modulation for treatment of cancer\n=============================================================================\n\nThe observations in clinical trials showed that Beclin-1 expression is altered in some cancers, overexpressed in others (intrahepatic cholangiocarcinoma and gastric tumors), whereas underexpressed in several solid tumors (breast, ovarian, cervical, lung, brain, liver, esophageal, gastric, and pancreatic) and osteosarcomas.^[@bib22],\\ [@bib31],\\ [@bib32],\\ [@bib33],\\ [@bib34],\\ [@bib35],\\ [@bib36]^ *Beclin-1* gene is found to be deleted in ovarian, breast, and prostate cancers, suggesting that reduced autophagic capacity is usual in some commonly diagnosed cancers.^[@bib37],\\ [@bib38]^ Apparently, the expression of LC3 and Beclin-1 is associated with cancer stage, low expression levels being associated with poorly differentiated tumors and more advanced clinical stage of disease. Examples of this correlation of Beclin-1 level with clinical cancer stage include ovarian,^[@bib32]^ esophageal squamous cell carcinoma,^[@bib39]^ chondrosarcoma,^[@bib40]^ and lymphomas,^[@bib41],\\ [@bib42]^ whereas no such connection has been found in cervix,^[@bib43]^ nasopharynx[@bib44] or adenoid cystic carcinoma.^[@bib44]^ More importantly, many chemotherapeutic agents, such as cisplatin, plant alkaloids, antimetabolites,^[@bib45],\\ [@bib46]^ tyrosine kinase inhibitors,^[@bib47],\\ [@bib48]^ and radiotherapy^[@bib49]^ induce autophagy. However, it remains to be determined whether the effectiveness of these therapies is dependent on hyper or under functional autophagy.^[@bib20]^\n\nCurrently, \\>30 clinical trials are investigating the effects of autophagy in combination with cytotoxic chemotherapy or targeted agents in various human cancers.^[@bib50]^ For example, chloroquine and hydroxychloroquine, routinely used for the treatment of diverse diseases, inhibit lysosomal acidification and prevent autophagy. In cancer treatment, chloroquine is often used in combination with chemotherapeutic drugs, such as cisplatin or PI3K inhibitor, LY294002 or the mTOR inhibitor, rapamycin. With all these agents, however, sensitization occurs independently of autophagy inhibition, and was not mimicked by Atg12, Beclin-1 knockdown, or bafilomycin treatment, and occurred even in the absence of Atg12.^[@bib51]^ The results of these clinical trials are critical for better understanding the process and role of autophagy in tumor biology and to validate the strategy of targeting autophagy to enhance therapeutic benefits to patients.\n\nCell death by autophagy\n=======================\n\nApoptosis (type I PCD) and necrosis (type III PCD) are well-known mechanisms of cell death induced by anticancer therapies. Recent studies have shown a non-apoptotic form of programmed death called autophagy, which is termed type II PCD and is often caspase independent.^[@bib52]^ In both apoptosis and autophagy, the degraded cells are disposed by phagocytosis without an inflammatory response, in contrast to the extensive cellular disintegration and subsequent inflammation that occur in necrosis.\n\nStudies have shown that there is a complex interplay between the apoptotic and autophagic processes. Autophagy may precede apoptosis or be induced simultaneously depending on the genetic context and cellular background of the cells. Oligomerized caspase-8 binds autophagosome membrane, leading to its activation and creating a mechanism for transition from apoptosis to autophagy.^[@bib53]^ Experimental findings suggest that activated caspase-3 can cleave Beclin-1, producing a fragment that translocates to mitochondria and induces apoptosis.^[@bib54]^\n\nOther studies indicate that the Bcl-2 family of proteins not only regulates apoptosis but also controls cell death that depends on the autophagy genes.^[@bib24]^ Cytotoxic signals can induce autophagy in apoptosis resistant cells, such as, those expressing high levels of Bcl-2 or Bcl-xL, lacking Bax and Bak, or exposed to pan-caspase inhibitors, such as, zVAD-fmk, suggesting autophagy to act as a default mechanism leading to cell death.^[@bib4]^ For instance, although embryonic fibroblasts from Bax/Bak double-knockout mice are resistant to apoptosis, they can undergo autophagic death after stimulation.^[@bib4]^ This non-apoptotic cell death was suppressed by autophagy inhibitors, such as, 3-MA, bafilomycin, or hydroxyl-choloroquine and or genetic silencing of autophagic genes (for example, *ATG5, ATG7*, *or Beclin-1*).^[@bib4]^ This indicates that the autophagic process has a significant role in caspase-independent cell death. Further studies supported these findings when apoptosis-resistant Bax\u2212/\u2212 and Bak\u2212/\u2212 knockout fibroblast cells underwent autophagic death following induction, such as, starvation, growth factor withdrawal, chemotherapy (etoposide), or radiation.^[@bib55]^ In addition, the knockdown of ATG5 or Beclin-1 in cancer cells has shown marked reduction in cell death and autophagic effects in response to death stimuli, with no sign of apoptosis.^[@bib12]^\n\nAutophagy has important roles in both, maintenance of cellular homeostasis under regular growth conditions and protection of cell viability under stress. In the housekeeping pathway, autophagy removes sources of ROS such as damaged or aggregated proteins and organelles to prevent tumor initiation via suppression of oxidative and genotoxic stress.^[@bib56],\\ [@bib57]^ Under stress, however, autophagy supports tumor cell survival by providing substrates for mitochondrial metabolism.^[@bib58]^ *In situations* of defective mitophagy, autophagy-deficient cells would accumulate damaged mitochondria and deregulate ROS levels, which have been suggested to contribute to induction of tumors.^[@bib59]^ Kaminskyy *et al.*^[@bib60]^ showed that autophagy suppression led to the inhibition of proliferation of non-small cell lung carcinoma cells and sensitized them to cisplatin-induced caspase-dependent and -independent apoptosis by stimulation of ROS formation.\n\nAutophagy leading to cell death is induced in various cancer cells in response to treatment by several polyphenolic compounds,^[@bib61]^ including rottlerin ([Figure 4](#fig4){ref-type=\"fig\"}),^[@bib12]^ curcumin,^[@bib62]^ resveratrol,^[@bib63]^ genistein,^[@bib64],\\ [@bib65]^ and quercetin,^[@bib66]^ as well as some chemotherapeutic agents, such as cytosine arabinoside,^[@bib67]^ etoposide,^[@bib4]^ and staurosporine,^[@bib68]^ and growth factor deprivation.^[@bib67]^\n\nPolyphenolic compounds and autophagic cell death\n================================================\n\nPolyphenols are a structural class of natural organic chemicals characterized by the presence of large multiples of phenol structural units. These compounds found in foods possess anticancer activities in their ability to alter the effects of signaling pathways and induce cell death not only via apoptosis but also autophagy.^[@bib69],\\ [@bib70]^\n\nRottlerin\n=========\n\nRottlerin (5, 7-dihydroxy-2, 2-dimethyl-6-(2, 4, 6-trihydroxy-3-methyl-5-acetylbenzyl)-8-cinnamoyl-1, 2-chromine), also called mallotoxin, is isolated from *Mallotus phillippinensis* (the monkey-faced tree). Rottlerin induces autophagy via three distinct mechanisms, that is, through PKC\u03b4/TG2, PKC\u03b4-independent, and mTORC1 pathways, in pancreatic cancer, fibrosarcoma, and breast cancer, respectively ([Table 1](#tbl1){ref-type=\"table\"}and [Figure 3](#fig3){ref-type=\"fig\"}).\n\nRottlerin displayed antioxidant properties and inhibitory effect on NF-*\u03ba*B in breast and colon cancer cells.^[@bib71]^ In fact, in these cells, expressions of PKC\u03b4 and TG2 led to activation of NF-*\u03ba*B,^[@bib72],\\ [@bib73]^ whereas inhibition led to induction of autophagy death.^[@bib74]^ Importantly, rottlerin is accepted widely as a PKC\u03b4-selective inhibitor.\n\nThe first evidence of massive autophagy induction leading to death only in cells with PKC\u03b4 via induction of TG2 was reported in pancreatic cancer cells.^[@bib12],\\ [@bib13]^ However, recent studies suggested that rottlerin can also induce apoptosis through PKC\u03b4-independent mechanisms in fibrosarcoma cells.^[@bib75]^ It was suggested that the early autophagy might serve as a survival mechanism against late apoptosis in this cancer type.^[@bib76]^\n\nRottlerin was able to inhibit mTORC1 signaling, via its negative regulator, TSC2 to induce autophagosome accumulation in breast cancer cells in nutrient-rich conditions.^[@bib77]^\n\nRottlerin inhibition of NF-\u03baB was able to induce AMPK induction, which led to significantly reduced cellular ATP levels and induction of autophagy in cancer cells.^[@bib78]^ AMPK can also activate the cyclin-dependent kinase inhibitor, p27, by a mechanism involving the SIRT1/FOXO pathway to induce autophagy.^[@bib79]^\n\nOverall, rottlerin-induced autophagy may involve multiple signaling pathways and cellular mechanisms for induction of autophagy and eventual cell death. However, the most important factor determining the fate of cells is probably the cellular context, increased apoptotic threshold/resistance and activated/inhibited signaling pathways. Thus rottlerin or its related analogs may be used in the development of novel agents for induction of autophagic cell death as it has been proven pharmacodynamically in a mice xenograft model to be efficiently absorbed in cells and tissues against pancreatic cancer.^[@bib80]^\n\nGenistein\n=========\n\nGenistein (4\u2032, 5, 7-trihydroxyisoflavone), a naturally occurring isoflavonoid found in soy products, has been shown to have anticancer properties. It has the capacity to induce cell death through both apoptosis^[@bib81]^ and autophagy.^[@bib82]^ Thus, genistein may be beneficial against chemoresistance owing to changes in apoptotic signaling ([Table 1](#tbl1){ref-type=\"table\"}).\n\nIt has been shown to completely protect the cytokeratin network in stress, nutrient- and growth factor-deprived environments. Various studies showed its ability to overcome the disruptive effects of okadaic acid, a strong inhibitor of autophagy, on the organization of the cytoskeleton and cytokeratin in rat hepatocytes.^[@bib64],\\ [@bib65]^ This is of consequence, as the cytokeratin filaments are involved in the development of autophagy.\n\nGenistein has been shown to be cytotoxic in ovarian cancer cells with mechanism of death involving not only apoptosis but also autophagy.^[@bib82]^ Treatment markedly inhibited glucose uptake in these cells, and methyl pyruvate, the substrate for oxidative phosphorylation and fatty acid synthesis, could rescue cells from genistein-induced autophagy. Gossner *et al.*^82^ also showed that treatment reduced levels of phosphorylated Akt. This may have contributed to limiting glucose utilization, which suggested that a starvation-like signaling response would eventually lead to autophagy death. Christian *et al.*^83^ reported its ability to inhibit both PKC and ERK inhibitors via inhibition of PDE4A4 aggregate formation to activate autophagy in ovarian cancer cells. These authors suggested that, as PDE4A4 aggregates are neither autophagosomes nor aggresomes and constitutively co-immunoprecipitated with the p62 protein (SQSTM1). Therefore, inhibiting their formation would be beneficial in ensuring induction of autophagy, as p62 interacts with LC3, which is critical for membrane encapsulation in autophagosomes.^[@bib84]^\n\nAli *et al.*^85^ reported evidence for genistein inhibition of N-CoR misfolding, an important component in the activation of the oncogenic survival pathway in non-small cell lung carcinoma, which was found to be associated with Hsc70, a molecular chaperone in autophagy ([Figure 3](#fig3){ref-type=\"fig\"}).\n\nAlthough, genistein can induce autophagy and apoptotic death in cancer cells, ADME studies revealed that it has intrinsically low oral bioavailability because of metabolic enzyme and efflux transporter. This should be further investigated to improve its efficacy in the treatment of apoptotic resistant cancers.^[@bib86]^\n\nQuercetin\n=========\n\nQuercetin (3, 3\u2032, 4\u2032, 5, 7-pentahydroxyflavone), a natural flavonoid molecule found in fruits, vegetables, leaves, and grains, has anticancer effects linked to its capacity for targeting key molecules, organelles, and tumorigenic pathways ([Table 1](#tbl1){ref-type=\"table\"}).^[@bib87],\\ [@bib88],\\ [@bib89]^ To confirm the involvement of autophagy, Psahoulia *et al.*^66^ treated RAS-transformed colon cells with 3-MA at the early stages, which resulted in the inhibition of vacuolization. Moreover, zVAD-FMK also failed to inhibit vacuolization, showing that the autophagy induced was caspase independent. Treatment of gastric cancer cells induced the vital stages that initiated autophagy progression.^[@bib90]^ Administration of the inhibitor chloroquine or selective ablation of Atg5 or Beclin-1 using siRNA increased apoptotic cell death, suggesting that autophagy has a protective role against quercetin-induced apoptosis. Functional studies revealed that the activated autophagy is modulated via the Akt-mTOR and HIF-1*\u03b1* signaling. Therefore, these xenograft models provided *in vivo* evidence for quercetin-induced apoptosis and autophagy.^[@bib90]^\n\nMartinez-Outschoorn *et al.*^91^ presented evidence that treatment with quercetin promoted the removal of defective mitochondria from cancer-associated fibroblasts by autophagy /mitophagy that was induced by oxidative stress. As a consequence, the 'reverse Warburg effect,\\' whereby these fibroblasts provided nutrients to stimulate mitochondrial biogenesis and oxidative metabolism in adjacent cancer cells, was proposed.\n\nQuercetin treatment induced extensive intracellular vacuolization and phagolysosome formation with accumulation of autophagic biomarkers in epithelial cancer cells, which led to cell cycle arrest and induction of apoptosis.^[@bib92]^ Before the formation of autophagosomes, inhibition of mTOR activity was observed, accompanied by a marked reduction in the phosphorylation of its substrates, the ribosomal S6 subunit via p70S6 kinase and the eIF4 via its inhibitor 4E-BP1. Inhibition of proteasome activity by quercetin was also observed with accumulation of polyubiquitinated protein aggregates, suggesting that proteasome inhibition is a major cause of the cancer cell death. Therefore, quercetin induces extensive autophagy and subsequent death in cancer cells mediated by the inhibition of proteasomal activity and mTOR signaling.^[@bib92]^\n\nLi *et al.*^93^ investigated whether autophagy contributes to HSP72-mediated cytoprotection in lipopolysaccharide-induced peritonitis. This is because HSP72 is known to induce autophagy but provide protection against apoptosis. The initial exposure of cultured peritoneal mesothelial cells to lipopolysaccharide resulted first in cell death via autophagy, with subsequent death seen to occur via apoptosis. Therefore, the activation of autophagy acted as a prosurvival mechanism. When autophagy was inhibited by 3-MA or Beclin-1 siRNA, the cells were sensitized to apoptosis, and the antiapoptotic effect of HSP72 was abolished. Also, overexpression of HSP72 enhanced autophagy through JNK phosphorylation and Beclin-1 upregulation. The suppression of JNK activity reversed HSP72-mediated Beclin-1 upregulation and autophagy, which indicated that HSP72-mediated autophagy, is JNK dependent. In the *in vivo* rat model of lipopolysaccharide-induced peritonitis, autophagy was seen to occur prior to apoptosis. When HSP72 was upregulated by geranylacetone, autophagy was increased, whereas apoptosis was inhibited and peritoneal injury reduced. These effects were reversed by downregulation of HSP72 with quercetin. When Li *et al.*^93^ blocked autophagy by chloroquine, there was induction of apoptosis and increased peritoneal dysfunction. Thus, they concluded that HSP72 protects peritoneum from lipopolysaccharide-induced mesothelial cell injury by inducing JNK activation-dependent autophagy and inhibiting apoptosis.\n\nIn a study to characterize the bioavailability and metabolic pharmacokinetics of quercetin in rats, 93.8% of the dose was circulating as its sulfates and glucuronides in the bloodstream when administered intravenously and 53% when given orally. These metabolites were seen to be responsible for the *in vivo* effects of quercetin. In both instances, the parent form of quercetin was not detected.^[@bib94]^\n\nTherefore, treatment with quercetin has numerous anticancer effects, including not only the induction of cell cycle arrest and apoptosis but also of autophagy through modulation of important autophagy signaling pathways such as Akt-mTOR and HIF-1*\u03b1* ([Figure 3](#fig3){ref-type=\"fig\"}).\n\nCurcumin\n========\n\nCurcumin (diferuloylmethane), an active ingredient of the spice turmeric *Curcuma longa*, has a potent anticancer effect on cancer cells.^[@bib95]^ Curcumin targets mainly the PI3K/Akt/mTOR signaling pathway and NF-*\u03ba*B-regulated proteins ([Table 1](#tbl1){ref-type=\"table\"}).\n\nVarious studies provided evidence that curcumin induced G~2~/M arrest and autophagy in malignant glioma cells through inhibition of the Akt/mTOR/p70S6K and activation of the ERK1/2 pathways, which implied that cell death via autophagy, might be pathway specific.^[@bib62],\\ [@bib96],\\ [@bib97]^ In a xenograft glioma model, Aoki *et al.*^96^ observed that curcumin induced autophagy and inhibited tumor growth significantly. Shinojima *et al.*^62^ observed that inhibition of NF-*\u03ba*B, which is the main anticancer target of curcumin, does not have a major role in the death of malignant glioma cells. Therefore, autophagy and not NF-*\u03ba*B, is the causal factor of the anticancer effects seen in these cells. Chadalapaka *et al.*^98^ showed that curcumin decreased expression of Sp proteins, whose overexpression in gastric and pancreatic cancers correlates with poor survival and tumor aggressiveness. Downregulation of *EGFR* (a Sp-regulated gene that suppresses autophagy) and decreased phosphorylation of Akt led to the induction of LC3 and cell death in bladder cancer. Curcumin protected HUVECs cells from oxidative stress by inducing autophagy via pI3K/Akt/mTOR and FOXO pathways to interact with Atg7; the involvement of FOXO (a mediator of autophagy) was confirmed in a knockdown study in which autophagy was inhibited by an siRNA.^[@bib99]^\n\nCurcumin induced autophagy in mesothelioma cells as indicated by increased conversion of LC3-I to LC3-II and formation of autophagosomes, which were reduced by RNA silencing of Atg5.^[@bib100]^ Mosieniak *et al.*^[@bib101]^ showed that induction of cellular senescence was accompanied by autophagy in colon cancer cells with an increase in Beclin-1 and p62/SQSTM1 protein levels. Interestingly, the inhibition of autophagy due to diminished expression of Atg5 by RNA interference also decreased the number of cells induced into senescence by curcumin, but did not lead to increased cell death. This study revealed a possible functional link between senescence and autophagy in curcumin-treated cells, a finding that should be investigated further.\n\nJia *et al.*^102^ presented evidence that curcumin is able to induce both autophagy and apoptosis in a chronic myeloid leukemia cell line via downregulation of the Bcl-2 protein. These effects were confirmed when the potent inhibitor of autophagosome--lysosome fusion, bafilomycin A1 and pan-caspase inhibitor zVAD-FMK suppressed the cell death. In a study on prostate cancer cells, curcumin was shown to induce autophagy cell death through downregulation of another important Bcl-2 family member, Bcl-xL. Curcumin treatment did not induce the cleavage of procaspase-8, -9, -3, or -7 or PARP but led to appearance of the LC3B-II isoform and increased number of autophagosomes.^[@bib103]^\n\nTreatment of human colon cancer cells with curcumin induced the conversion of LC3-I to LC3-II and degradation of SQSTM1 and autophagosomes.^[@bib104]^ The autophagic changes induced by curcumin were almost completely blocked in the presence of the antioxidant NAC, indicating that treatment led to ROS production, autophagosome formation, and autolysosomal degradation. The reduction of SQSTM1 degradation by bafilomycin, further confirmed the activation of autophagy cell death.^[@bib104]^ Kim *et al.*^105^ reported anticancer activity involving ROS in oral squamous cell carcinoma via both apoptosis and autophagy, which was confirmed when NAC blocked autophagic vacuole formation.\n\nCurcumin has been proven to not only induce apoptosis on its own but to also have synergistic effects with various FDA-approved drugs via major inflammatory biomarkers and oncoproteins.^[@bib106]^ As curcumin promotes autophagy similarly via most of these proteins, therefore this compound is effective in inducing both apoptosis and autophagy in cancer therapy.\n\nHowever, a major restriction on the use of curcumin as an anticancer agent is its poor absorption, biodistribution, metabolism, and bioavailability. When 400\u2009mg of curcumin were fed to rats, about 60% were found to be absorbed. To address these problems, several formulations have been used, which include nanoparticles, liposomes, micelles, and phospholipid complexes, all with limited success. Also curcumin undergoes metabolism to form various metabolites, such as, glucuronide, sulfate, tetrahydrocurcumin, hexahydrocurcumin, octahydrocurcumin, and hexahydrocurcuminol after oral administration. However, all these metabolites displayed anticancer effects. ^[@bib107]^\n\nResveratrol\n===========\n\nResveratrol (3,5,4-trihydroxystilbene), a natural phytoalexin present in grapes, nuts, and red wine, has chemopreventive properties and multiple mechanisms of action, which may be activated depending on the specific cell type and cellular environment ([Table 1](#tbl1){ref-type=\"table\"}).\n\nOpipari *et al.*^63^ provided the first evidence that resveratrol induces cell death through autophagy in five ovarian cancer cell lines, which suggested it as an effective treatment in apoptosis-resistant cells. As autophagy is an adaptive response to nutrient starvation and resveratrol\\'s ability to induce a starvation-like signaling response, that is, reducing the levels of phosphorylated Akt and mTOR to initiate autophagy, is observed in ovarian cancer cells.^[@bib108]^\n\nResveratrol was shown to bind to a novel estrogen receptor coactivator, PELP1, and induce its accumulation in autophagosomes.^[@bib109]^ PELP1 was identified for the first time by a trafficking molecule, hepatocyte growth factor-regulated tyrosine kinase substrate, which binds to it.^[@bib110]^ The role of hepatocyte growth factor-regulated tyrosine kinase substrate in facilitating the transport of cytoplasmic proteins to autophagosomes for their selective degradation was confirmed in another study involving resveratrol treatment of lung cancer cells.^[@bib111]^\n\nPuissant *et al.*^29^ presented evidence of resveratrol triggering autophagic death in chronic myeloid leukemia cells via both JNK-mediated p62/SQSTM1 overexpression and AMPK/mTOR activation.^[@bib108]^ Also resveratrol enhanced the expression of several tubulin subunits, which is important for movement of autophagosomes inside the cell.\n\nTrincheri *et al.*^112^ reported that autophagy can be induced with acute exposure to resveratrol, whereas prolonged exposure activated a caspase-mediated cell death pathway. They observed that genetic inactivation of PI3K, Beclin-1, and Lamp2b inhibited resveratrol toxicity. Beclin-dependent autophagy was confirmed when supplementing the cells with asparagine or knocking down Beclin-1 by RNA interference, abrogated the effect. The effect of Lamp2b was confirmed when its silencing inhibited the fusion of autophagosomes with lysosomes and induced cell viability. Interestingly, zVAD-FMK inhibited cell death but not autophagy. This study uncovered a novel pathway of resveratrol cytotoxicity in which autophagy has two roles, that is, a prosurvival stress response that later in the process, changes to a caspase-dependent apoptotic response. In another study, resveratrol increased ROS level with induction of caspase-8 and caspase-3 cleavage and elevation of LC3-II expression in colon cancer cells; these effects were diminished by NAC.^[@bib113]^\n\nAn interesting non-canonical autophagic process characterized by a Beclin-1/Vps34-independent pathway was observed in MCF-7 cancer cells in response to resveratrol treatment. Overexpression of Bcl-2, which is known to block canonical starvation-induced autophagy by binding to Beclin-1, was unable to reverse the non-canonical autophagy triggered by resveratrol in these breast cancer cells.^[@bib8],\\ [@bib114]^\n\nSCCA 1, an endogenous cathepsin L inhibitor, is expressed widely in uterine cervical cells. Hsu *et al.*^115^ showed that the cathepsin L--SCCA 1 lysosomal pathway and autophagy were involved in resveratrol-induced cytotoxicity in cervical cancer cells. Inhibition of the autophagic response by wortmannin or asparagine resulted in decreased autophagic death.\n\nDihydroceramide, an immediate precursor of the apoptotic mediator ceramide in the *de novo* sphingolipid synthesis pathway, was accumulated when resveratrol-induced autophagy occurred as a result of inhibition of dihydroceramide desaturase activity in gastric cancer cells. These effects of resveratrol were mimicked by a dihydroceramide desaturase inhibitor.^[@bib116]^\n\nResveratrol-induced autophagy in human glioma cells has the ability to inhibit resveratrol-induced apoptosis.^[@bib117]^ Autophagy and apoptosis seemed to have different roles, apoptosis causing these cells\\' death, whereas autophagy delayed apoptosis and protected the cells from death. This suggests that autophagy inhibitors may have the potential to enhance resveratrol antitumor efficacy.^[@bib117]^\n\nIn a large-scale *in vitro* kinase screen, p70 S6 kinase (S6K1) was identified as a target of resveratrol. Blocking S6K1 activity by expression of a dominant-negative mutant or RNA interference was sufficient to disrupt autophagy to an extent similar to resveratrol. Also, co-administration of resveratrol with S6K1 knockdown did not produce an additive effect, which indicated that S6K1 is important for the full induction of autophagy and some of the beneficial effects of resveratrol are due to modulation of S6K1 activity.^[@bib118]^\n\nSIRT1, one of the best characterized targets of resveratrol, upon activation by this compound induced both autophagy and apoptosis.^[@bib119]^ However Armour *et al.*^118^ observed that resveratrol decreased autophagy in response to nutrient limitation in multiple cell lines through a pathway independent of SIRT1. Resveratrol is able to induce autophagy in lung cells with cigarette smoke-mediated oxidative stress via regulation of SIRT1 and PARP.^[@bib120],\\ [@bib121],\\ [@bib122]^ Resveratrol prevented the decline in ATP concentration and SIRT1 expression, as well as the increase in HIF-1*\u03b1* expression and autophagy, in the livers of endotoxin-challenged wild-type mice but not in the liver of SIRT1 knockout mice.^[@bib123]^ These findings provided an insight to the potential roles of SIRT1 and HIF-1*\u03b1* expression in systemic inflammation.\n\nPersistent human papillomavirus infection may stabilize ATAD3A (an anti-autophagy factor), inhibit cell autophagy and apoptosis and to increase drug resistance in uterine cervical cancer. Resveratrol\\'s anticancer effects were confirmed by its capacity for reducing ATAD3A expression, increasing abrasion of the mitochondrial outer membrane, and increasing the numbers of autophagosomes.^[@bib124]^\n\nResveratrol was shown to trigger autophagic cell death through increased expression of Atg5, 7, 9, and 12 proteins in a human hepatitis C-induced hepatoma cell line.^[@bib125]^ Also, Filippi-Chiela *et al.*^126^ observed that resveratrol induced the formation of autophagosomes through upregulation of Atg5, Beclin-1, and LC3-II in glioblastoma cells. The PtdIns(3)P effectors, WIPI-1, and WIPI-2 were shown to function downstream during initiation of autophagosome formation. Localization of WIPI-1 at the endoplasmic reticulum and the plasma membranes upon the induction of autophagy confirmed its involvement with the autophagosomal membrane.^[@bib127]^ In another study, WIPI-1 specifically bound PtdIns(3)P, accumulated at the phagophore, and become a membrane protein of the autophagosome generated. WIPI-1 was observed to function upstream of both Atg7 and 5 and stimulated an increase of LC3-II upon nutrient starvation. These findings constituted evidence that resveratrol-mediated autophagy was via the non-canonical pathway, Beclin-1 independent but Atg7 and 5 dependent.^[@bib128]^\n\nAs seen with curcumin, resveratrol can induce premature senescence that is associated with a blockade of autolysosome formation, as assessed by the absence of colocalization of important markers of autophagosomes and lysosomes, LC3 and Lamp-2, respectively. Resveratrol was also able to downregulate the level of Rictor, a vital component of the mTORC2 complex, which led to decreases in RhoA-GTPase activity, altered actin cytoskeleton network and increases in senescence-associated *\u03b2*-gal activity. Therefore, resveratrol has the ability to attenuate the autophagic process via downregulation of Rictor, which may be the mechanism of tumor suppression associated with premature senescence.^[@bib129]^\n\nResveratrol upon administration is readily absorbed and metabolized as glucuronides and sulfates and the oral bioavailability of resveratrol is almost zero owing to its rapid metabolism activity.^[@bib130]^ However, as resveratrol has shown versatility in inducing autophagy in numerous cancer types and targeting a wide array of autophagy-associated proteins, further clinical investigations are important to fully evaluate the efficacy and bioactivity of resveratrol in the killing of cancer cells via autophagy.\n\nConclusion\n==========\n\nOne of the most important unsolved problems in cancer therapy is increased tumor resistance to treatment, be it chemotherapy, radiotherapy, or any of the targeted therapies. This increased resistance is a direct effect of defects in apoptosis. An alternative form of cell death, namely, autophagy, may be the ultimate solution for this problem. The evidence presented here ([Table 1](#tbl1){ref-type=\"table\"} and [Figure 2](#fig2){ref-type=\"fig\"}) demonstrates that several phytochemical polyphenolic compounds found in foods are able to mediate both canonical and non-canonical autophagy via multiple pathways targeting important proteins in a number of cancer types. Understanding their mechanisms is of vital importance in ensuring cell death *versus* survival. These polyphenolic compounds have the ability to induce both apoptosis and autophagy, thereby maximizing death of cancer cells. As they are food compounds, furthermore, they may offer greater safety, both through their inherently lower toxicity and through allowing reduction of doses and side effects as compared with synthetic drugs. Combining FDA-approved drugs with known polyphenolic compounds such as rottlerin, genistein, quercetin, curcumin, and resveratrol may provide novel therapeutic strategies in the treatment of cancer to combat the substantial problem of drug resistance in cancer therapy.\n\nThis work was supported by the Centre for Research in Biotechnology for Agriculture (CEBAR) University Malaya RU Operation Grant (RU005C-2014). The funder had no role in decision to publish or preparation of the paper.\n\n3-MA\n\n:   3-methyl adenine\n\n4E-BP1\n\n:   eIF4E binding protein 1\n\nADME\n\n:   absorption, distribution, metabolism and excretion\n\nAMPK\n\n:   AMP-dependent protein kinase\n\nATAD3A\n\n:   ATPase family AAA domain containing 3A\n\nATG\n\n:   autophagy regulated genes\n\nATM\n\n:   ataxia telangiectasia mutated\n\nCML\n\n:   chronic myeloid leukemia\n\neIF4\n\n:   eukaryotic initiation factor of protein biosynthesis\n\nER\n\n:   endoplasmic reticulum\n\nERK\n\n:   extracellular signal-regulated kinase\n\nFDA\n\n:   Food and Drug Administration\n\nFOXO\n\n:   forkhead box class O\n\nHIF-1*\u03b1*\n\n:   hypoxia-induced factor 1*\u03b1*\n\nHRS\n\n:   hepatocyte growth factor--regulated tyrosine kinase substrate\n\nHSP72\n\n:   heat shock protein 72\n\nHUVECs\n\n:   human umbilical vein endothelial cells\n\nJNK c-Jun\n\n:   *N*-terminal kinase\n\nLamp2b\n\n:   lysosome-associated membrane proteins 2b\n\nMAPK\n\n:   mitogen-activated protein kinase\n\nMDC\n\n:   monodansylcadaverin\n\nmTOR\n\n:   mammalian target of rapamycin\n\nNAC\n\n:   N-acetylcysteine\n\nN-CoR\n\n:   nuclear receptor co-repressor\n\nNF-\u03baB\n\n:   nuclear factor kappa B\n\nNSCLC\n\n:   non-small cell lung carcinoma\n\nPARP\n\n:   poly(ADP ribose) polymerase\n\nPCD\n\n:   programmed cell death\n\nPELP1\n\n:   proline-, glutamic acid--, and leucine-rich protein-1\n\nPDE4A4\n\n:   cyclic AMP phosphodiesterase-4A4\n\nPI3K/Akt\n\n:   phosphatidylinositide 3-kinase\n\nPKB\n\n:   protein kinase B\n\nPKC\u03b4\n\n:   protein kinase C delta\n\nROS\n\n:   reactive oxygen species\n\nSCCA\n\n:   squamous cell carcinoma antigen\n\nsiRNA\n\n:   small interfering RNA\n\nSIRT1\n\n:   sirtuin 1\n\nSQSTM1\n\n:   p62/sequestome 1\n\nTG2\n\n:   transglutaminase 2\n\nTSC2\n\n:   tuberous sclerosis complex 2\n\nWIPI-1/2\n\n:   Wd-repeat proteins interacting with phosphoinositides\n\nzVAD-fmk\n\n:   benzyloxycarbonyl-Val--Ala--Asp fluoromethylketone\n\nEdited by M Diederich\n\nThe authors declare no conflict of interest.\n\n![Chemical structures of autophagy-inducing polyphenolic compounds](cddis2014467f1){#fig1}\n\n![The major signaling pathway regulation and core machinery of autophagy. Various pathways regulate autophagy either positively or negatively. Most of these pathways, including AMPK and PKA, merge at mTORC1. PKA directly activates mTORC1, inactivating both AMPK and autophagy. AMPK negatively regulates mTORC1 in two ways, either directly or by activating the TSC2 protein. The mTORC1 substrate p70S6K is a positive regulator of autophagy. Another important upstream factor is Akt/PKB, a negative regulator of the TSC1/2 complex. The core machinery includes the processes of induction, phagophore assembly (membrane nucleation), sequestration, autophagosome formation, and autophagolysosome formation. The process of induction involves initiation of the UNC-51-like kinase complex members ULK 1/2, Atg13, Atg101, and FIP200. Activation of the PtdIns3K complex (Beclin-1, Vps34, and Vps15) is an essential step in phagophore assembly (membrane nucleation). The E1-like enzyme Atg7 activates both Atg12 and LC3-I, and the E2-like enzyme activates Atg10 (for Atg12) and Atg3 (for LC3-I). Atg12 is conjugated to Atg5 and the Atg12-Atg5 complex acts as an E3 ubiquitin ligase to catalyze the conjugation of LC3-I to the lipid phosphatidylethanolamine (PE) in the process of sequestration. The subsequent autophagosome formation is dependent on the Atg12-Atg5-Atg16 complex. After autophagosome completion, the Atg12-Atg5-Atg16 complex dissociates from autophagosomes to allow Atg4 access to LC3-II-PE for deconjugation. When the autophagosome is completed, it fuses with lysosome to form an autophagolysosome in which the cytosolic macromolecules, proteins, and organelles will either be degraded by acid hydrolases in what is known as type II programmed cell death or involved in a survival mechanism to be released back into the cytosol with the help of permeases against cellular stress. Akt/PKB, protein kinase B; AMPK, adenosine monophosphate-activated protein kinase; CaMKK\u03b2, calmodulin kinase kinase \u03b2 Atg, autophagy-related genes; FIP200, focal adhesion kinase (FAK) family-interacting protein of 200kD; LC3, microtubule-associated protein 1 light chain 3; LKB1, liver kinase B1; mTORC1, mammalian target of rapamycin complex 1; PE, phosphatidylethanolamine; p70S6K, p70S6 kinase; PKA, protein kinase A; PtdIns3K, phosphatidylinositol 3-kinase; TAK1, TGF\u03b2-activated kinase 1; TSC1/2, tuberous sclerosis complex 1/2; ULK 1/2, UNC-51--like kinases 1/2; Vps 15/34, vacuolar protein sorting 15/34](cddis2014467f2){#fig2}\n\n![Polyphenols promote autophagic cell death in cancer cells. The polyphenolic compounds rottlerin, genistein, quercetin, curcumin, and resveratrol induce cell death via autophagy through inhibition or activation of multiple signaling pathways and cellular targets in cancer cells. The common pathways involved are PI3K/Akt/mTOR, NF-\u03baB, MEK/ERK, PKC\u03b4/TG2, JNK/p62/SQSTM1, and AMPK/TSC2. Akt/PKB, protein kinase B; AMPK, adenosine monophosphate-activated protein kinase; ATM, ataxia telangiectasia mutated; ATP, adenosine triphosphate; Bcl-2, B-cell lymphoma 2; Bcl-XL, B cell lymphoma-extra large; BNIP3, [B]{.ul}cl-2 and [n]{.ul}ineteen KD [i]{.ul}nteracting [p]{.ul}rotein-3; eIF4E-BP1, eukaryotic translation initiation factor 4E-binding protein 1; ERK1/2, extracellular signal-regulated kinase 1/2; FOXO, forkhead transcription factors; HIF-1*\u03b1*, Hypoxia-inducible factor 1-alpha; HRS, hepatocyte growth factor-regulated tyrosine kinase substrate; Hsc70, 70-kDa heat shock protein family expressed constitutively; HSP72, 70-kDa heat shock protein family stress- and heat shock-induced; JNK, c-Jun N-terminal kinase; Lamp2b, lysosome-associated membrane proteins 2b; MEK1/2, mitogen-activated protein kinase 1/2; mTORC1/2, mammalian target of rapamycin complex 1/2; N-CoR, putative corepressor; NF-*\u03ba*B, nuclear factor-kappa B; PARP1, poly(ADP-ribose) polymerase 1; PDE4A4, cyclic AMP phosphodiesterase-4A4; PELP1; proline-, glutamic acid-, and leucine-rich protein-1; PI3K, phosphatidylinositol 3-kinase; PKC\u03b4, protein kinase C-delta; PTEN, phosphatase and tensin homolog deleted on chromosome ten; p62/SQSTM1, p62 protein /sequestosome 1; p70S6K, p70 ribosomal protein S6 kinase; Raf-1, oncoprotein activated by Ras; Ras, oncoprotein; Rheb, Ras homolog enriched in brain; ROS, reactive oxygen species; SIRT1, sirtuin 1; TG2, transglutaminase 2; TSC1/2, tuberous sclerosis complex 1/2; Vps 15/34, vacuolar protein sorting 15/34](cddis2014467f3){#fig3}\n\n![Rottlerin induced massive autophagy and cell death in pancreatic cancer cells. (**a**) Morphological changes and formation of autophagic vacuoles in MDA-Panc28 pancreatic cancer cells that were treated with 4\u2009*\u03bc*M rottlerin. Microphotographs were taken using a phase-contrast microscope ( \u00d7 300 magnification). Left panel, untreated cells at 48\u2009h; center panel, cells treated for 4 days; and right panel, cells treated for 7 days. (**b**) MDA-Panc28 cells were treated with Rottlerin for 48\u2009h and then were stained with acridine orange for detection of acidic vesicular organelles by fluorescent microscopy (left panels) and localization of LC3-II at autophagosomes in cells after transfection with GFP-LC3 plasmid (middle panels).^[@bib12],\\ [@bib13]^ Green fluorescence indicates cytoplasm and nucleus, whereas red fluorescence shows acidic vesicular organelles. Right panels, electron micrographs showing the ultrastructure of rottlerin-treated cells (4\u2009M). Numerous autophagic vacuoles, indicated by arrows, were observed in the rottlerin-treated cells](cddis2014467f4){#fig4}\n\n###### Effect of polyphenols on induction of autophagy in cancer\n\n  **Treatment**   **Cancer cell/tissue**                     **Mechanism**                 **Dose**   **Reference**  \n  --------------- ------------------------------------------ ---------------------------- ---------- --------------- -----------------------\n  Rottlerin       Pancreatic                                 PKC\u03b4/TG2                        2--4          ---        ^[@bib12],\\ [@bib13]^\n  \u00a0               Fibrosarcoma                               PKC\u03b4-indep                    0.5--10         ---             ^[@bib131]^\n  \u00a0               Breast                                     mTORC1                           3            ---             ^[@bib77]^\n  Genistein       Rat hepatocytes                            Cytokeratin                     ---           100        ^[@bib64],\\ [@bib65]^\n  \u00a0               Ovarian                                    Akt                           25--100         ---             ^[@bib82]^\n  \u00a0               Ovarian                                    PDE4A4& p62/SQSTM1               10           ---             ^[@bib83]^\n  \u00a0               Lung                                       N-CoR/Hsc70                    25--50         ---             ^[@bib85]^\n  Quercetin       Colon                                      Ras                              20           ---             ^[@bib66]^\n  \u00a0               Fibroblast-breast                          ROS                              10           ---             ^[@bib91]^\n  \u00a0               Gastric                                    Akt/mTOR&HIF-1\u03b1               10--160         50              ^[@bib90]^\n  \u00a0               Rat mesothelial                            HSP72/JNK&Beclin-1              ---           100             ^[@bib93]^\n  \u00a0               Breast, Cervical, Ovarian                  mTOR/eIF4E-BP1/p70S6K          30--90         ---             ^[@bib92]^\n  Curcumin        Brain                                      Akt/mTOR/p70S6K                10--50         100        ^[@bib62],\\ [@bib96]^\n  \u00a0               \u00a0                                          ERK1/2                           \u00a0             \u00a0                   \u00a0\n  \u00a0               CML                                        Bcl-2                          5--20          ---             ^[@bib102]^\n  \u00a0               Bladder                                    Akt                              40           ---             ^[@bib98]^\n  \u00a0               Prostate                                   Bcl-xL                         10--50         ---             ^[@bib103]^\n  \u00a0               Colon                                      ROS                            10--40         ---             ^[@bib104]^\n  \u00a0               Brain                                      PI3K/Akt/mTOR                    2            300             ^[@bib97]^\n  \u00a0               Mesothelioma                               ND                             10--50         ---             ^[@bib100]^\n  \u00a0               Oral                                       ROS                              10           ---             ^[@bib105]^\n  \u00a0               Colon                                      Beclin-1&p62/SQSTM1              10           ---                ^01^\n  \u00a0               Endothelial                                PI3K/Akt/mTOR&FOXO1            1--10          ---             ^[@bib99]^\n  Resveratrol     Ovarian                                    ND                               50           ---              ^[@bib3]^\n  \u00a0               Salivary gland                             PELP1/HRS                     50--100         ---             ^[@bib110]^\n  \u00a0               Ovarian                                    Akt/mTOR/ p70S6K              25--100         ---             ^[@bib108]^\n  \u00a0               Lung                                       PELP1/HRS                     50--100         ---             ^[@bib111]^\n  \u00a0               Colorectal                                 PI3K/Beclin-1/Lamp2b            100           ---             ^[@bib112]^\n  \u00a0               Breast                                     Akt/PKB/mTOR/p70S6K              64           ---             ^[@bib132]^\n  \u00a0               Cervical                                   Cathepsin L                     100           ---             ^[@bib115]^\n  \u00a0               Gastric                                    Dihydroceramide desaturase       50           ---             ^[@bib116]^\n  \u00a0               Brain                                      Beclin-1                        150           ---             ^[@bib117]^\n  \u00a0               Fibroblast, Cervical                       p70S6K                        50--200         ---             ^[@bib118]^\n  \u00a0               CML                                        JNK/p62, AMPK/mTOR             10--50         ---             ^[@bib133]^\n  \u00a0               Lung                                       SIRT1/PARP-1                     10           ---             ^[@bib120]^\n  \u00a0               Hepatoma                                   ND                               20           ---             ^[@bib125]^\n  \u00a0               Liver                                      SIRT1, AMPK, HIF-1*\u03b1*           ---           20              ^[@bib123]^\n  \u00a0               Colon                                      SIRT1                           100           ---             ^[@bib122]^\n  \u00a0               Brain                                      ND                               30           ---             ^[@bib126]^\n  \u00a0               Cervical                                   ATAD3A                        30--100         ---             ^[@bib124]^\n  \u00a0               Osteosarcoma, melanoma, cervical, breast   WIPI-1                           64           ---             ^[@bib128]^\n  \u00a0               Colon                                      ROS                           25--150         ---             ^[@bib113]^\n  \u00a0               Skin                                       mTORC2 (Rictor)                  50           ---             ^[@bib129]^\n\nAbbreviations: Akt/PKB, protein kinase B; AMPK, adenosine monophosphate-activated protein kinase; ATAD3A, ATPase family AAA domain containing 3A; Bcl-2, B-cell lymphoma 2; eIF4E-BP1, eukaryotic initiation factor 4E binding protein 1; ERK 1 / 2, extracellular signal-regulated kinases 1 / 2; HIF-1\u03b1, hypoxia-inducible factors -1 alpha; HRS, hepatocyte growth factor-regulated tyrosine kinase substrate; HSP72, heat shock protein 72; JNK, c-Jun N-terminal kinases; Lamp 2b, lysosome associated membrane proteins 2b; mTOR, mammalian target of rapamycin; mTORC 1 / 2, mTOR complex 1 / 2; N-CoR, nuclear receptor corepressor; ND, not determined; p62/SQSTM1, the ubiquitin-binding protein p62 or sequestosome 1; p70S6K, p70S6 kinase; PARP-1, Poly \\[ADP-ribose\\] polymerase 1; PDE4A4, cyclic AMP phosphodiesterase-4A4; PELP1, proline-, glutamic acid-, and leucine-rich protein-1; PI3K, phosphatidylinositol 3-kinases; PKC\u03b4, protein kinase C delta; Ras, rat sarcoma protein; ROS, reactive oxygen species; SIRT1, Sirtuin 1; Tcf-4, transcription factor -4; TG2, tissue transglutaminase; TSC1/2, tuberous sclerosis 1 / 2; Vps 34/ Class III PI 3-kinase; WIPI, WD-repeat protein interacting with phosphoinosides; Wnt, wingless integration signaling pathway. Induction of autophagic death by the five dietary polyphenolic compounds rottlerin, genistein, quercetin, curcumin and resveratrol in various cancer types, the regulatory mechanisms involved and the effective doses for *in vitro* as well as *in vivo* analyses\n"}
+{"text": "![](medchirrev76516-0184){#sp1 .177}\n\n![](medchirrev76516-0185){#sp2 .178}\n\n![](medchirrev76516-0186){#sp3 .179}\n\n![](medchirrev76516-0187){#sp4 .180}\n\n![](medchirrev76516-0188){#sp5 .181}\n\n![](medchirrev76516-0189){#sp6 .182}\n\n![](medchirrev76516-0190){#sp7 .183}\n\n![](medchirrev76516-0191){#sp8 .184}\n\n![](medchirrev76516-0192){#sp9 .185}\n\n![](medchirrev76516-0193){#sp10 .186}\n\n![](medchirrev76516-0194){#sp11 .187}\n\n![](medchirrev76516-0195){#sp12 .188}\n\n![](medchirrev76516-0196){#sp13 .189}\n\n![](medchirrev76516-0197){#sp14 .190}\n\n![](medchirrev76516-0198){#sp15 .191}\n\n![](medchirrev76516-0199){#sp16 .192}\n\n![](medchirrev76516-0200){#sp17 .193}\n\n![](medchirrev76516-0201){#sp18 .194}\n\n![](medchirrev76516-0202){#sp19 .195}\n\n![](medchirrev76516-0203){#sp20 .196}\n\n![](medchirrev76516-0204){#sp21 .197}\n\n![](medchirrev76516-0205){#sp22 .198}\n\n![](medchirrev76516-0206){#sp23 .199}\n\n![](medchirrev76516-0207){#sp24 .200}\n\n![](medchirrev76516-0208){#sp25 .201}\n\n![](medchirrev76516-0209){#sp26 .202}\n\n![](medchirrev76516-0210){#sp27 .203}\n\n![](medchirrev76516-0211){#sp28 .204}\n\n![](medchirrev76516-0212){#sp29 .205}\n\n![](medchirrev76516-0213){#sp30 .206}\n\n![](medchirrev76516-0214){#sp31 .207}\n\n![](medchirrev76516-0215){#sp32 .208}\n\n![](medchirrev76516-0216){#sp33 .209}\n\n![](medchirrev76516-0217){#sp34 .210}\n\n![](medchirrev76516-0218){#sp35 .211}\n\n![](medchirrev76516-0219){#sp36 .212}\n\n![](medchirrev76516-0220){#sp37 .213}\n\n![](medchirrev76516-0221){#sp38 .214}\n\n![](medchirrev76516-0222){#sp39 .215}\n\n![](medchirrev76516-0223){#sp40 .216}\n\n![](medchirrev76516-0224){#sp41 .217}\n\n![](medchirrev76516-0225){#sp42 .218}\n\n![](medchirrev76516-0226){#sp43 .219}\n\n![](medchirrev76516-0227){#sp44 .220}\n\n![](medchirrev76516-0228){#sp45 .221}\n\n![](medchirrev76516-0229){#sp46 .222}\n\n![](medchirrev76516-0230){#sp47 .223}\n\n![](medchirrev76516-0231){#sp48 .224}\n"}
+{"text": "1. Introduction {#j_med-2019-0088_s_001}\n===============\n\nPain is a very complex physiopathological entity. It can take on different aspects both from a clinical and biological/ physiological point of view, and present itself as an epiphenomenon of a pathophysiological process, until it becomes a real pathological entity in its own right. Pain is defined as chronic when it is continuous for more than three tosix months \\[[@j_med-2019-0088_ref_001]\\]. It is a condition reported in 20% of people worldwide, in 15-20% of all physician examinations \\[[@j_med-2019-0088_ref_002]\\], and should receive more attention because a proper pain therapy is a human right \\[[@j_med-2019-0088_ref_002],[@j_med-2019-0088_ref_003]\\]. The trigger for the development of chronic pain may be different in different situations. Inflammation causes inflammatory pain, while nerve injuries as the result of mechanical trauma (iatrogenic or not), metabolic or autoimmune disorders, and cancer and chemotherapy may give neuropathic pain \\[[@j_med-2019-0088_ref_004], [@j_med-2019-0088_ref_005], [@j_med-2019-0088_ref_006]\\]. The excitation of the primary neurons due to prolonged inflammation induce a pathological response that persist beyond the period of recovery of the tissue, constantly stimulating the nociceptive pathways and thus generating chronic pain with changes in ion channels, receptors and nerve synapses. The distribution of neurotransmitters and neuromediators allows peripheral and central neurons to reach the depolarization threshold early to cause ectopic discharges to amplify and activate nearby cells, with chronic pain \\[[@j_med-2019-0088_ref_007]\\]. Neuronal pathophysiological mechanisms are integrated with immunological response, and neuropathic pain is considered a neuro-immune disorder \\[[@j_med-2019-0088_ref_008], [@j_med-2019-0088_ref_009], [@j_med-2019-0088_ref_010], [@j_med-2019-0088_ref_011], [@j_med-2019-0088_ref_012], [@j_med-2019-0088_ref_013]\\]. In fact, patients with complex regional pain, peripheral neuropathy and neuropathic pain associated with spinal cord injury syndrome have increased serum IL-4, IL-6 and TNF-\u03b1, as well as reduced serum IL-10 levels \\[[@j_med-2019-0088_ref_014], [@j_med-2019-0088_ref_015]\\]. The serum increases of IL-1\u03b2, IL-6, IL-2, TNF-\u03b1, and IFN-\u03b3 increase the intensity of chronic pain \\[[@j_med-2019-0088_ref_016], [@j_med-2019-0088_ref_017], [@j_med-2019-0088_ref_018], [@j_med-2019-0088_ref_019]\\]. Therefore, inflammatory and allogenic processes are supported by a complex balance between cells and cytokines (due to both pro and anti-inflammatory molecules) and the nervous system.\n\nIL-6, TNF-\u03b1 and IL1-\u03b2, produced by macrophages, are cytokines that amplify chronic pain. According to some studies, the administration of TNF-\u03b1 is associated with thermal hyperalgesia, mechanical allodynia and hypersensitivity associated with pain. In the electrophysiological field, TNF-\u03b1 can also increase the conductivity of K+ ions of the membrane in a non-voltage dependent manner, resulting in overall neuronal hyperexcitability and therefore neuropathic pain. TNF-\u03b1 induce the release of glial mediators TNF-\u03b1-induced glial mediators cause endocytosis of GABA receptors with consequent reduction of inhibitory modulation of the GABAergic system. Longterm potentiation (LTP) is a physiological mechanism for strengthening a neuronal circuit and is involved in numerous nerve functions. Intrathecal therapy (IT) is a good choice to improve therapeutic results in chronic pain.\n\nThe aim of this study was to evaluate in clinical practice the effectiveness of the intrathecal pump in 140 patients consecutively enrolled from January 2010 to July 2018 who underwent pain therapy at our Center.\n\n2. Materials and methods {#j_med-2019-0088_s_002}\n========================\n\nOne hundred and forty patients who underwent pain therapy at our Center were consecutively enrolled from January 2010 to July 2018. All procedures conducted in the study were in accordance with international guidelines, with the standards of human experimentation of the local Ethics Committees and with the Helsinki Declaration of 1975, revised in 1983.\n\nAt the baseline visit, each patient signed their informed consent for the use of their data in clinical research.\n\nAll 140 patients underwent complete physical examination. Their pain was evaluated with the McGill Pain Questionnaire (MPQ), recording the Numerical Rating Scale (NRS) of average pain, minimum pain, maximum pain and pain during exertion (scale 0--10: 0 = no pain, 10 = worst pain ever). All patients were evaluated for HBV, and HCV serum markers. Serum samples were tested for HBsAg, anti-HCV, total anti-HBc, and anti-hepatitis B surface antibody (HBs) using commercial immunoenzymatic assays (Abbott Laboratories, North Chicago, IL, USA: AxSYM\u00ae HBsAg (v2) M/S for HBsAg, AxSYM\u00ae HCV (v3) for anti-HCV, AxSYM\u00ae CORE\u2122 (v2) for total anti-HBc, and AxSYM\u00ae AUSAB\u00ae for anti-HBs), as described in previous studies \\[[@j_med-2019-0088_ref_020], [@j_med-2019-0088_ref_021], [@j_med-2019-0088_ref_022], [@j_med-2019-0088_ref_023], [@j_med-2019-0088_ref_024], [@j_med-2019-0088_ref_025], [@j_med-2019-0088_ref_026], [@j_med-2019-0088_ref_027], [@j_med-2019-0088_ref_028]\\].\n\nEpidemiological, clinical, and immunological data recorded at the baseline visit included patients' age and sex and risk factors for the acquisition of pain as stated by the patients.\n\nAll cases were followed-up in these 8 years with complete physical examinations, laboratory tests were carried out and pain was re-evaluated.\n\n3. Results {#j_med-2019-0088_s_003}\n==========\n\nThe demographic and clinical data obtained at enrolment are shown in [Table 1](#j_med-2019-0088_tab_001){ref-type=\"table\"}.\n\n###### \n\nDemographic and clinical characteristics of the 140 enrolled patients\n\n  ------------------------------------------------------- -------------\n  **N. of patients**                                      140\n  **Males, N. (%)**                                       98 (70.0%)\n  **Age, years (M + SD)**                                 51.2 \u00b1 8.2\n  **Caucasian, N (%)**                                    140 (100%)\n  **Etiology of the pain, N(%): *chronic cancer pain***   99 (71.0%)\n  ***chronic non-cancer pain***                           41 (29.0%)\n  **Therapy, N (%)**:                                     \n  ***Morphine in primary infusion***                      120 (85.6%)\n  ***Ziconotide in primary infusion***                    3 (0.2%)\n  ***Baclofen in primary infusion***                      11 (0.8%)\n  ***Ziconotide plus morphine***                          3 (0.2%)\n  ***Bupivacaine plus morphine***                         3 (0.2%)\n  ------------------------------------------------------- -------------\n\nThe 140 patients were predominantly male: 98 (70%), with a media age of 51.2 years old (range: 42-63). Of 140 patients, 99 cases had been subjected to a system for the treatment of chronic cancer pain; 41 patients were diagnosed with non-cancer pain. In non-cancer pain patients, the main diagnoses were neuropathic or mixed pain, chronic, not responsive to conventional drug therapy and / or burdened by the incidence of significantly elevated side effects with high doses. Some patients were not responders to spinal cord stimulation ([Table 1](#j_med-2019-0088_tab_001){ref-type=\"table\"}).\n\nMorphine was administered as the first intrathecal continuous infusion in 120 cases, in three patients ziconotide was used alone as the primary infusion, and 11 patients received primary Baclofen infusion alone. Ziconotide and Bupivacaine are analgesic adjuvants to morphine. In three patients, Ziconotide had been used in association with morphine as a secondary infusion, and three patients had a secondary infusion of Bupivacaine used in association with morphine. The patient was then given an intrathecal continuous infusion of one of the following: ziconotide 2.5 mL (2.5 \u03bcg at 1 \u03bcg per mL) daily, morphine 0.1-8 mg daily, baclofen 250-1000 \u03bcg daily or bupivacaine 3-8 mg daily.\n\nFrom the follow-up carried out over these 8 years regarding the infusion modalities and the titration of the described drugs, the following data concerning pain relief was obtained: 71 (50,7%) patients out of 140 experienced satisfactory pain control globally, settling on a percentage of pain relief variable from 50 to 70%.\n\nOnly one patient, followed by oncologist, showed pain relief less than 30%.\n\nIn patients on follow-up with baclofen infusion for a diagnosis other than chronic pain, a variable reduction of the painful physiological symptomatology in spasticity was observed.\n\n4. Discussion {#j_med-2019-0088_s_004}\n=============\n\nIntrathecal therapy is a good choice to improve therapeutic results in patients with chronic pain. In fact, the morbidity and mortality of IT opioids are lower than those of systemic administration of opioids. A trial on 6398 patients in over a period of ten years \\[[@j_med-2019-0088_ref_029]\\], showed that there weren't deaths correlated with the IT opioid infusion. In the US, death from systemically administered opioids are about 19,000 a year. Intrathecal drug delivery (IDD) is potentially a safer option for the patient \\[[@j_med-2019-0088_ref_030]\\].The Food and Drug Administration (FDA) has approved two intrathecal pain relief drugs: morphine free of preservatives and ziconotide peptide. Intrathecal opioid administration may cause side effects (vomiting, nausea, itching, constipation, urinary retention and neuroendocrine dysfunction) \\[[@j_med-2019-0088_ref_031]\\]. The most serious adverse reaction reported is respiratory depression \\[[@j_med-2019-0088_ref_032]\\], which occurs with the administration of opioid IT therapy or restarting of IT opioids after stopping \\[[@j_med-2019-0088_ref_033]\\]. The risk may be higher with hydrophilic opioids, such as morphine, and in cases where other depressants of the central nervous system (CNS), such as benzodiazepines, are administered \\[[@j_med-2019-0088_ref_034]\\]. The Polyanalgesic Consensus Conference (PACC) has recently recommended testing opioid therapy in an outpatient setting with a conservative dose, as a single dose of 0.15 mg morphine produces respiratory depression (defined as a PaCO2 above 40 or a lower respiratory rate at 10 breaths per minute) \\[[@j_med-2019-0088_ref_035], [@j_med-2019-0088_ref_036], [@j_med-2019-0088_ref_037]\\]. Ziconotide is a non-opioid drug approved for IT \\[[@j_med-2019-0088_ref_038], [@j_med-2019-0088_ref_039], [@j_med-2019-0088_ref_040], [@j_med-2019-0088_ref_041]\\], with adverse events of the CNS reported (nausea, nystagmus, vertigo, dysmetria, ataxia, agitation, hallucinations and coma) \\[[@j_med-2019-0088_ref_042]\\]. Neuropsychiatric adverse effects can occur after many months of asymptomatic infusion and are the main reason for its suspension \\[[@j_med-2019-0088_ref_043]\\]. Although an increase in serum creatinine kinase is present in 40% of patients, this significant increase typically occurs in the first two months after initiation of ziconotide therapy and only three cases of rhabdomyolysis have been described. There is no antidote for the overdose of ziconotide IT. The use of intrathecal bupivacaine (usually in combination with morphine) is off-label and it has occasionally been associated with numbness and rarely with weakness \\[[@j_med-2019-0088_ref_044]\\]. The PACC of 2016 presented evidence-based recommendations on patient survival, pathological process and use of IT drugs. Regarding cancer pain, IT is frequently thought to offer reliable, safe and effective treatment \\[[@j_med-2019-0088_ref_045], [@j_med-2019-0088_ref_046], [@j_med-2019-0088_ref_047], [@j_med-2019-0088_ref_048], [@j_med-2019-0088_ref_049], [@j_med-2019-0088_ref_050]\\].\n\nRegarding the algorithm for IT management in patients with cancer pain, the prognosis, the probable disease progression in anatomical regions, features of tumors and findings of peri-procedural imaging are useful considerations for the selection of the device. Patients with cancer pain deserve special attention, in terms of implementation of IT and selection of drugs, together with the sustainability of a regime. Therefore, patients are stratified into 3 main categories by stage of disease and life expectancy. In the last PACC update, IT is used to treat related/unrelated cancer pain, and chronic-severe refractory pain \\[[@j_med-2019-0088_ref_051]\\]. The focus was on the patient's age, although age contributions are reflected in the sustainability of the therapy, which is addressed elsewhere in the recommendations \\[[@j_med-2019-0088_ref_052],[@j_med-2019-0088_ref_053]\\]. The costs of this therapy and its safety are better than the use of systemic opioids \\[[@j_med-2019-0088_ref_054], [@j_med-2019-0088_ref_055]\\].\n\nThe need to resort to the continuous infusion of subarachnoid drugs occurs when pharmacological, neurostimulation or surgical therapies have been unsuccessful. The main indications are: chronic non-cancer pain with a strong neuropathic component; chronic cancer pain with more than 6 months of life expectancy, and spasticity of degenerative and traumatic origin.\n\nThe procedure is completely reversible, the catheter can be removed if the expectations in terms of reduction / disappearance of the symptomatology are not met; the greatest advantage is the possibility of treating the symptoms (pain, spasticity) with minimal doses of effective drug, avoiding the appearance of troublesome side effects. Fixed-flow pumps have an indefinite duration, while programmable flow pumps are constrained to when the battery runs out, so they must be replaced before the battery expires.\n\nThe implant procedure involves positioning the catheter in the subarachnoid space, connecting it to the pump through a second catheter (extension) that is tunneled to the abdomen where the subcutaneous pocket containing the pump is packaged. Both the first and second phases of the operation are performed under local anesthesia, possibly paired with a mild analgosedation. In addition to the risks associated with the completion of a surgical procedure, the implants and / or the use of IT devices may also entail the following risks and complications: the displacement or disconnection of the catheter may occur; insertion of a catheter into the subarachnoid space may result in epidural bleeding, headache, hematoma, infection, spinal cord compression and/ or paresis; loss of cerebrospinal fluid (fluid that bathes nerve structures), with possible headache; persistent pain at the site of implantation of the catheter or pump; serum collection at the pump site just under the skin, which can be evacuated with a simple puncture; catheter migration, which can cause changes in the analgesic effect; allergic reaction or rejection of implanted materials;and pain localized at the implant site.\n\nOur study demonstrates the effectiveness of intrathecal therapy in reducing severe chronic refractory pain compared to intravenous (i.v.) therapies. The importance of specialized staff dedicated to these procedures can drastically reduce adverse events and complications. It is a safe and manageable solution in pain management and can be considered for chronic refractory pain from all sources. Our data shows a longer follow up and case study than the others in the literature \\[[@j_med-2019-0088_ref_056], [@j_med-2019-0088_ref_057], [@j_med-2019-0088_ref_058], [@j_med-2019-0088_ref_059], [@j_med-2019-0088_ref_060], [@j_med-2019-0088_ref_061], [@j_med-2019-0088_ref_062]\\] and this reinforces our obtained results.\n\n5. Conclusion {#j_med-2019-0088_s_005}\n=============\n\nChronic pain is correlated with lower quality of life and depressive/anxiety disorders (15.5%) in patients without a history of these disorders. Depression and anxiety are associated with more severe pain.\n\nIntrathecal therapy is one of the best options in the treatment of chronic/severe refractory pain. The costs associated with therapy in terms of safety are significantly better than those of systemic opioids.\n\nThe greatest advantage of this therapy is due to the possibility of treating the pain with minimal dosages of the drug, avoiding the appearance of troublesome side effects. Non-responsive patients represent the ideal category to be addressed with \"niche\" techniques such as cordotomy.\n\nOur results confirm that intrathecal therapy is one of the best choices in the management and treatment of severe chronic refractory pain. This therapy, overall, is safer than systemic opioids, which often need higher dosages to be effective. With systemic opioid therapy, the higher dosages and possible interferences with other drugs used in this type of complex patient allow an increase in the possibility of serious adverse events. Intrathecal therapy has demonstrated superiority in the management of dosages of these drugs which can be reduced to the minimum effective with this technique.\n\nThe use of minimum effective dosages of opioids in pain therapy allows a reduction in serious adverse events, and for this reason also a considerable reduction in the overall costs of treatment of chronic patients.\n\nThe need to resort to the continuous infusion of subarachnoid drugs occurs when pharmacological, neurostimulation or surgical therapies have been unsuccessful. The catheter can be removed calmly, if the expectations in terms of symptomatic reduction are not met.\n\n**Conflict of interest**\n\nConflict-of-interest statement: All the authors of the manuscript declare that they have no conflicts of interest in connection with this paper.\n"}
+{"text": "INTRODUCTION\n============\n\nCystic hypersecretory carcinoma (CHC) and cystic hypersecretory hyperplasia (CHH) were first described in 1984 ([@B1]). The characteristic features of these lesions are dilated ducts and cysts containing an eosinophilic secretory product resembling thyroid colloid. After the initial report of CHC and CHH, the entire spectrum of cystic hypersecretory lesions of the breast was described by Guerry et al. in 1988 ([@B2]). These lesions range from benign CHH to the intermediate CHH with atypia and the frankly malignant CHC.\n\nThere have been only seven cases of invasive CHC reported in the literature ([@B1]-[@B5]). We describe an additional case of invasive CHC in a 45-yr-old female.\n\nCASE REPORT\n===========\n\nA 45-yr-old woman was admitted to the hospital for diagnosis and treatment of a palpable mass in the lower quadrant of the left breast. The mass was soft and focally hard. She had no past or family history of a breast disease. The sonograph showed a cystic and lobulated mass. An excisional biopsy of the left breast was performed. Gross pathologic examination of the excisied specimen revealed an ill-defined, mucoid mass. The cut surface of the mass, which was 4.7\u00d73.7\u00d73 cm, revealed multiple cystic spaces, and the cysts were filled with thick, gelatinous secretions. The individual cysts varied from 0.1 cm to 1.2 cm in dimension ([Fig. 1](#F1){ref-type=\"fig\"}). Intervening solid areas were also noted. Microscopically, many cystically dilated ducts contained thyroid colloid-like eosinophilic secretions. The homogeneous secretions were retracted from the surrounding epithelia, producing scalloped margins ([Fig. 2A](#F2){ref-type=\"fig\"}). The secretory material also showed linear cracking artifact. The cyst-lining epithelium exhibited variable patterns. The lining of the cysts in most areas was of flat or cuboidal epithelium and devoid of cellular atypia ([Fig. 2B](#F2){ref-type=\"fig\"}). The epithelium of some cysts showed proliferative change ranging from atypical hyperplasia to intraductal carcinoma, micropapillary type ([Fig. 2C](#F2){ref-type=\"fig\"}). The intraductal carcinoma component was accompanied by an invasive component with small solid nests ([Fig. 2D](#F2){ref-type=\"fig\"}). The invasive component was a high-grade carcinoma lacking cystic and papillary traits. Histochemical staining of the secretory material in the cysts was positive with periodic acid-Schiff (PAS) and alcian blue. Immunohistochemically, the cystic contents were positive for carcinoembryonic antigen (CEA) (Zymed, San Franciscoa, CA, U.S.A., predilute), but negative for thyroglobulin (Zymed, predilute). Immunostains for estrogen and progesterone receptors (DAKO, Glostrup, Denmark, predilute) and p53 protein (DAKO, dilution 1:100) were negative in the neoplastic epithelial cells.\n\nThe diagnosis was invasive CHC. Modified radical mastectomy with axillary lymph node dissection was performed. Axillary lymph nodes were free of tumor metastasis. Subsequent radiotherapy was performed. The seven-month follow-up period was uneventful.\n\nDISCUSSION\n==========\n\nCystic hypersecretory lesions of the breast have a spectrum of morphologic features ranging from the clearly benign (CHH), a combination of benign and atypical epithelium (CHH with atypia), to cases that combine benign, atypical, and frankly malignant epithelium (CHC) ([@B2]). The characteristic gross features are the formation of dilated ducts and cysts filled with a colloid-like secretion. Although cystic hypersecretory lesions have a distinctive gross appearance, it is usually not possible to distinguish CHC from CHH grossly. CHC is differentiated from CHH by a micropapillary cyst lining with cytologic atypia. If no cytologic atypia is present and the epithelium is flat or cuboidal, the lesion is characterized as CHH. Invasion is heralded by solid nests of invasive carcinoma and is usually poorly differentiated with no secretory characteristics. As a consequence, total excisional biopsy is required for definitive diagnosis of cystic hypersecretory lesions of the breast.\n\nAbout 50 cases of cystic hypersecretory breast lesions have been reported ([@B1]-[@B7]). Most cases of CHC have been intraductal and only seven cases of invasive CHC have been reported ([@B1]-[@B5]). Most invasive carcinomas have been poorly differentiated duct carcinomas with a solid growth pattern. Invasive CHC tends to have an aggressive behavior. Four cases were diagnosed with lymph node metastases ([@B2], [@B5]). Metastatic foci in the axillary lymph nodes had cystic foci that contained eosinophilic secretion. One patient developed invasive lobular carcinoma of the contralateral breast 10 yr after ipsilateral mastectomy for invasive CHC ([@B3]). The patient described herein had negative lymph nodes.\n\nThe present case is the eighth case of invasive CHC. In this case all features of CHH were identified. In addition, micropapillary intraductal CHC and focal high-grade invasive carcinoma were also observed. Our case supports the concept that cystic hypersecretory breast lesions encompass a spectrum of pathologic lesions including CHH, atypical CHH, CHC, and invasive CHC. The progression of these lesions from CHH, through intraductal CHC, to invasive CHC may be possible.\n\nThe differential diagnosis of invasive CHC includes secretory carcinoma, mucinous carcinoma, malignant mucocele-like tumor, and metastatic thyroid carcinoma. Secretory carcinoma contains vacuolated cytoplasm and more bubbly secretions, which are not typical features of CHC ([@B8]). Mucinous carcinoma and malignant mucocele-like tumor also show cystically dilated ducts ([@B9]). However, the secretions in these lesions are rather pale and basophilic and do not show linear cracking artifacts. Metastatic follicular thyroid carcinoma of the breast may mimic CHC. Histologic differentiation of these two conditions may require immunohistochemical stain for thyroglobulin. Negative reaction for thyroglobulin was observed in the cyst contents of our case.\n\n![The cut surface of the mass shows numerous cysts, measuring up to 1.2 cm in diameter, with a gelatinous secretion.](jkms-19-149-g001){#F1}\n\n![Microscopic findings. The lesion is composed of multiple cysts and ducts containing eosinophilic secretion (**A**, H&E, \u00d720). Most of the cysts are lined by flat epithelium. The secretion retracts from the surrounding epithelium (**B**, H&E, \u00d7200). The epithelium of some cysts grows as micropapillary intraductal carcinoma (**C**, H&E, \u00d7200). An invasive component (arrow) is found adjacent to the intraductal carcinoma component (**D**, H&E, \u00d7100).](jkms-19-149-g002){#F2}\n"}
+{"text": "![](edinbmedj75053-0042){#sp1 .192}\n\n![](edinbmedj75053-0043){#sp2 .193}\n\n![](edinbmedj75053-0044){#sp3 .194}\n\n![](edinbmedj75053-0045){#sp4 .195}\n\n![](edinbmedj75053-0046){#sp5 .196}\n\n![](edinbmedj75053-0047){#sp6 .197}\n\n![](edinbmedj75053-0048){#sp7 .198}\n\n![](edinbmedj75053-0049){#sp8 .199}\n\n![](edinbmedj75053-0050){#sp9 .200}\n\n![](edinbmedj75053-0051){#sp10 .201}\n\n![](edinbmedj75053-0052){#sp11 .202}\n\n![](edinbmedj75053-0053){#sp12 .203}\n\n![](edinbmedj75053-0043-a){#f1 .193}\n\n![](edinbmedj75053-0043-b){#f2 .193}\n\n![](edinbmedj75053-0044-a){#f3 .194}\n\n![](edinbmedj75053-0044-b){#f4 .194}\n\n![](edinbmedj75053-0045-a){#f5 .195}\n\n![](edinbmedj75053-0045-b){#f6 .195}\n\n![](edinbmedj75053-0046-a){#f7 .196}\n\n![](edinbmedj75053-0046-b){#f8 .196}\n\n![](edinbmedj75053-0047-a){#f9 .197}\n\n![](edinbmedj75053-0047-b){#f10 .197}\n"}
+{"text": "\n"}
+{"text": "Introduction {#s1}\n============\n\nThe *Ralstonia solanacearum* species complex (RSSC) (Gillings and Fahy, [@B7]) is responsible for bacterial wilt on a broad range of plant hosts comprising more than 200 species in at least 50 families (Hayward, [@B10]). RSSC is particularly destructive for vegetable crops, including potato, tomato, eggplant and pepper plants. RSSC strains are known for their unusually broad genetic basis and phenotypic diversity in tropical and subtropical areas (Hayward, [@B12]). Soil-borne RSSC strains invade the roots and colonize the xylem vessels (Vasse et al., [@B32]), leading to wilt symptoms and the death of their hosts. RSSC strains have been frequently reported to develop latent infections that are maintained at high concentrations in asymptomatic hosts (Grimault and Prior, [@B8]). Breeding for resistance remains the most effective and sustainable strategy to control bacterial wilt (Prior et al., [@B24]). Unfortunately, resistance to bacterial wilt often breaks down due to the genomic plasticity and the large genetic and phenotypic diversity within RSSC (Gillings and Fahy, [@B7]; Lebeau et al., [@B15]). The sustainability of host resistance is tried to a large-scale local management strategy that includes study of the genetic diversity of bacterial wilt-causing strains and their virulence patterns (Lebeau et al., [@B15]).\n\nHistorically, RSSC strains have been classified into races and biovars based on their host ranges and biochemical properties (Buddenhagen et al., [@B1]; Hayward, [@B11]), but these classifications are neither predictive nor phylogenetically meaningful. Subsequently, strains unifying the RSSC were distributed into four major phylotypes of different geographical origins named phylotypes based upon phylogenetic analyses of sequence data generated from the 16S-23S internal transcribed spacer (ITS) region as follows: I from Asia, II from the America, III from Africa and the Indian Ocean, and IV from Australia, Japan, and Indonesia (Fegan and Prior, [@B6]). More recently, the RSSC was taxonomically organized into three species that classified phylotypes I and III as *R. pseudosolanacearum*, phylotype II as *R. solanacearum* and phylotype IV as *R. syzygii* (Safni et al., [@B28]; Prior et al., [@B23]). The phylotypes are subdivided into sequevars based on sequence variation in the endoglucanase (*egl*) partial gene (Fegan and Prior, [@B6]).\n\nMayotte is a small island located in the southwest Indian Ocean, more precisely in the Comoros archipelago between Eastern Africa and Madagascar. Small-scale farming systems have been developed, especially with vegetable production, which strongly contributes to secure agricultural resources. From the panel of tropical plant diseases, the severity of bacterial wilt outbreaks is the major constraint to vegetable production, especially during the warm season. Tomatoes are largely consumed throughout the year; therefore, improvement of tomato production during the off-season (corresponding to the rainy and wet season) remains highly strategic for farmers income. In Mayotte, bacterial wilt has been recognized for a long time on tomato (*Solanum lycopersicum*), eggplant (*Solanum melongena*), sweet pepper (*Capsicum annuum*), hot pepper (*C. frutescens*) and European black nightshade (*Solanum nigrum*) plants. Various unpublished studies have also been undertaken during the last two decades to assess the resistance of tomato and eggplant varieties against local RSSC strains. As a secluded island in the Comoros archipelago, Mayotte is located in a strategically poor documented area that covers the spectrum of RSSC diversity.\n\nThe literature available on the distribution and economic importance of RSSC in Africa and the Indian Ocean remains discrete (Elphinstone, [@B5]), although broad diversity has been reported, with three of the four known phylotypes identified. In African countries where large samplings have been achieved, such as Cameroon, the Ivory Coast and Ethiopia, phylotypes I and IIA have been reported to be more prevalent in the lowlands, whereas phylotypes IIB-1 and III are mostly observed in the highlands (Lemessa and Zeller, [@B16]; Mahbou Somo Toukam et al., [@B19]; N\\'Guessan et al., [@B20]). Broad genetic diversity of phylotype III strains has been reported in a few countries in Sub-Saharan Africa (Angola, Burkina Faso, Cameroon, Guinea, the Ivory Coast, Kenya, and Zimbabwe) and the southwest Indian Ocean (Madagascar and Reunion) (Ravelomanantsoa et al., [@B26]).\n\nIn this study, we investigated an almost closed and uniform agronomical landscape in which we assumed that the RSSC population was well-established due to its insularity and limited international exchanges compared to continental environments. A 140-strain collection was sampled, which allowed us to unravel the genetic diversity of RSSC and assign the phylogenetic positions of the strains. As a prerequisite to understanding, extending and managing the success of commercially resistant tomato cultivars, the virulence of a subset of representative strains was assayed using resistant tomato breeding lines under field and controlled conditions.\n\nMaterials and methods {#s2}\n=====================\n\nBacterial strains\n-----------------\n\nAlthough Mayotte is a small island (20 \u00d7 40 km), the main vegetable production areas are distributed into five agro-ecological zones (Figure [1](#F1){ref-type=\"fig\"}). Bacterial wilt occurred in all areas surveyed in September 2012. Solanaceous plants (tomato, eggplant, hot pepper, sweet pepper, and black nightshade) showing typical bacterial wilt symptoms were sampled at 24 sites with a particular emphasis on the tomato, which is highly susceptible to bacterial wilt in Mayotte. For each plant, a large stem segment (5--10 cm in length) was cut and maintained at approximately 25\u00b0C prior to bacterial isolation and purification. The stems were surface-disinfected with 70% ethanol, and a sub-fragment (0.5 cm) was shredded and macerated in sterile distilled water to allow bacterial release. Then, the macerates (50 \u03bcL) were individually streaked onto tetrazolium chloride (TZC) agar medium (Kelman, [@B14]) for 48--72 hours at 28\u00b0C. For each sample, one typical RSSC colony was re-streaked onto new TZC medium for further purification and bacterial species validation. Finally, each strain was assigned a CIRAD Reunion (RUN) identification number (Table [S1](#SM1){ref-type=\"supplementary-material\"}) and maintained at \u221280\u00b0C on Cryobank\u00ae microbeads at Cirad (Saint-Pierre, Reunion).\n\n![Mayotte map showing the five main vegetable areas (C, Centre; N, North; NE, Northeast; S, South; SE, Southeast) and the 24 sites where phylotype I strain sequevars 31, 18, 46, and 15 were collected.](fpls-08-02209-g0001){#F1}\n\nPhylogeny based on the endoglucanase (*egl*) partial gene\n---------------------------------------------------------\n\nThe endoglucanase partial gene was sequenced from 140 strains to identify their phylotypes and sequevars and to create a phylogenetic tree. A 750-bp fragment of the *egl* gene was amplified using the Endo-F /Endo-R primer pair as previously reported (Cellier and Prior, [@B3]). The PCR products were dehydrated in a vacuum and sent to Beckman Coulter Genomics (Takeley, UK) for further purification and double-strand sequencing using the PCR primers as the sequencing primers. Raw sequences from both strands were edited, trimmed and aligned under the ARB software package (<http://www.arb-home.de/>) (Ludwig et al., [@B18]). The sequences were trimmed starting at 5\u2032-ACGGCGAT-3\u2032 and ending at 5\u2032-ACGGCGGC-3\u2032. A phylogenetic tree was constructed with the 140 isolated strains from Mayotte together with a set of international reference strains (Table [S2](#SM2){ref-type=\"supplementary-material\"}) using the neighbor-joining method (Saitou and Nei, [@B30]) with 5,000 bootstrap resampling runs to test the robustness of the tree topology. Assignment to *egl*-based sequevars was conducted using reference *egl* sequences (Prior and Fegan, unpublished data). The sequences of the newly described strains were deposited into the GenBank database under accession numbers [MF359095](MF359095) - [MF359234](MF359234) (Table [S1](#SM1){ref-type=\"supplementary-material\"}).\n\nEvaluation of tomato resistance to mayotte phylotype I\n------------------------------------------------------\n\nFirst, we tested the resistance properties of 10 tomato accessions (Core collection) that were internationally considered references for tomato resistance origins (Table [S3](#SM3){ref-type=\"supplementary-material\"}) (Lebeau et al., [@B15]). The field assay was conducted in a field located at Dembeni (Southeastern Mayotte), which is an area that is naturally infested with sequevar I-31 (the most prevalent sequevar in Mayotte; this study). This experiment was conducted under a full ground tunnel during the hot and wet season that was favorable for the development of bacterial wilt. This assay was performed using a random experimental plan of four repetitions corresponding to four blocks of 10 plants per modality defined as following:CRA66 (T1), Okitsu Sozai no.1 (T2), NC 72 TR 4-4 (T3), IRAT L3 (T4), Hawaii 7996 (T5), TML46 (T6), CLN1463 (T7), R3034 (T8), L285 (T9), and L390 (10). The plantlets were grown under nursery conditions and planted at the three to four fully expanded leaf stage in double lines with a density of 3 plants per m2. The water supply was provided by a drip system using a pump equipped with two sand filters, a disk filter and a pressure regulator. Disease progression was visually assessed weekly over a 2-month period by marking each plant as \"dead\" or \"not dead\" (all leaves wilted).\n\nA second virulence test was conducted at Cirad Reunion in growth chambers (Rotoplan) under a routine security norm level (NS2). Six tomato lines of the Core collection (Lebeau et al., [@B15]) were chosen based on the field assay results. Two commercial tomato hybrid F1 cultivars (Cobra and Platinum) were added to this test, since they were widely cropped in Mayotte (Table [S3](#SM3){ref-type=\"supplementary-material\"}). All of these accessions were challenged with a subset of 8 strains representative of the genetic diversity of RSSC from Mayotte (Table [1](#T1){ref-type=\"table\"}). The strains were chosen according to their phylotype-sequevar classification, geographical location and isolation host. Two strains were selected from each of the 4 sequevars identified. The sequevar I-31 strains RUN2108 and RUN2170 were selected since they were isolated at the \"Station Agronomique de Dembeni\" and the \"Lyc\u00e9e Agricole de Coconi,\" which are the main experimental sites for vegetable production in Mayotte. Additionally, the RUN2108 strain was isolated in the experimental full ground tunnel used for the tomato field assay. Virulence was tested on plantlets with 3--4 fully expanded leaves. Three repetitions of 10 pots with one healthy plantlet with three to four true leaves were selected for each plant x strain combination. The strains were grown at 30\u00b0C on Kelman solid medium with 0.5 g of yeast extract (Kelman, [@B14]). Bacterial suspensions calibrated to 5 \u00d7 10^8^ colony-forming units were prepared as previously described (Cellier and Prior, [@B3]). Inoculation was achieved by pouring 2 mL of inoculum on lateral roots previously wounded with a scalpel. All plants were placed in growth chambers with a 12-h photoperiod, 25 \u00b1 2\u00b0C (night) and 30 \u00b1 2\u00b0C (day) temperature, and 80% relative humidity. Bacterial wilt development was monitored for 15 days after inoculation by scoring the number of wilted plants as follows: asymptomatic (no symptoms), wilting plant (at least one wilted leaf) or dead (all wilted leaves) (Lebeau et al., [@B15]). After the final record, the remaining healthy plants were tested to determine whether they were carrying a latent infection. For each plant, a cutting 2 to 3 cm in length at the stem base was immersed in 5 mL of Tris buffer solution at room temperature for 2--3 h to allow bacterial release. A 50-\u03bcL aliquot of each extract was streaked onto semi-selective modified Granada and Sequeira medium (Poussier et al., [@B38]) and incubated for 72 h at 28\u00b0C. Latent infections were scored as positive or negative depending on the development of colonies on the plates. A colonization index was calculated for each plant x strain combination according to the formula CI = N~wp~ + (N~s~ \u00d7 R~s~), where N~wp~ was the percentage of wilted plants, N~s~ was the percentage of asymptomatic plants and Rs was the percentage of asymptomatic plants with a latent infection (Lebeau et al., [@B15]).\n\n###### \n\n*Ralstonia solanacearum* phylotype I strains from Mayotte selected for the virulence test against tomato accessions under controlled conditions.\n\n  **Strains**   **Location (area)**   **Isolation host**   **Species**              **Sequevar**\n  ------------- --------------------- -------------------- ------------------------ --------------\n  RUN2170       Coconi (C)            Eggplant             *Solanum melongena*      31\n  RUN2108       Dembeni (SE)          Tomato               *Solanum lycopersicum*   31\n  RUN2083       M\\'romouhou (SE)      Hot pepper           *Capsicum annuum*        18\n  RUN2150       Miangani (N)          Tomato               *S. lycopersicum*        18\n  RUN2127       Combani (C)           Sweet pepper         *C. annuum*              46\n  RUN2146       Miangani (N)          Tomato               *S. lycopersicum*        46\n  RUN2143       Mits\u00e9ni (N)           Sweet pepper         *C. annuum*              15\n  RUN2140       Mits\u00e9ni (N)           Eggplant             *S. melongena*           15\n\nData analysis\n-------------\n\nEffect of accessions on resistance phenotypes was tested by analysis of variance on the wilted rate means 120 days after plantation. The data was transformed with arcsine square root function to fit normal distribution of the variable. Then means differentiation between accessions was done using a pairwise comparison test by Tukey\\'s \"Honest Significant Differences\" method (package agricolae, R statistical freeware software, version 3.4.2).\n\nThe strain phenotypes were scored according to the pathotypes as defined previously by testing the CoreRs2 (RSSC strain collection) on the Core-TEP collection (Lebeau et al., [@B15]). A pathotype was defined as a group of strains with similar virulence profiles on reference accessions of a host species, such as the tomato (Lebeau et al., [@B15]). A clustering approach was used to determine the phenotype scores according to the final wilting incidence and the colonization index of each accession-strain combination used in the study. The phenotype scores of each accession-strain combination were assigned using the \"k-nearest neighbors\" algorithm of the package *class* of the R statistical freeware (version 3.0.2) (Venables and Ripley, [@B33]) to the reference phenotype score classification, which comprised five levels: 1 = highly resistant, 2 = moderately resistant, 3.1 = partially resistant, 3.2 = latently infected, 4 = moderately susceptible, and 5 = highly susceptible (Lebeau et al., [@B15]). An agglomerative hierarchical nesting classification using the package *agnes* confirmed by a fuzzy clustering analysis using the package *cluster* was undertaken independently of the Core-Rs2 to identify tomato pathotypes for the 8 tomato accessions used for the study (Lebeau et al., [@B15]).\n\nResults {#s3}\n=======\n\nBacteria isolation and identification\n-------------------------------------\n\nA total of 140 RSSC strains were sampled from the 24 surveyed sites throughout the 5 vegetable cropping areas in Mayotte. Strains were collected from five Solanaceous crops at rates of 57.1% for tomato (*n* = 80), 32.9% for eggplant (*n* = 46), 5.0% for sweet pepper (*n* = 7), 2.1% for hot pepper (*n* = 3), and 2.9% for black nightshade (*n* = 4).\n\nPhylotype and sequevar phylogeny\n--------------------------------\n\nOnly phylotype I strains were identified according to the most recent RSSC taxonomy (Safni et al., [@B28]; Prior et al., [@B23]). The sequevar assignation of the strains based on *egl* sequencing (140 strains) revealed the presence of four sequevars (I-15, I-18, I-31 and I-46) (Figure [2](#F2){ref-type=\"fig\"}). The prevalence and distribution of the four sequevars appeared to be variable throughout the five vegetable production areas in Mayotte (Table [S4](#SM4){ref-type=\"supplementary-material\"}). Sequevar I-31 (*n* = 120) was the most prevalent, representing 85.7% of the collection compared to sequevars I-18 (*n* = 7), I-15 (*n* = 8), and I-46 (*n* = 5), which represented 5.0, 5.7, and 3.6% of the collection, respectively. Sequevar I-31 was isolated from the northern area where 15.7% of the strains were collected (*n* = 22). Sequevars I-18 and I-15 were isolated from the central and southeastern areas, whereas I-46 was detected only in the central area. Only one sequevar (I-31, *n* = 17) was identified in the northeastern area. The central area is the most important vegetable production area in Mayotte. Sequevar diversity was the highest in this area. Of the 4 identified sequevars, sequevar I-31 (*n* = 48) represented 72.7% of the diversity, followed by the remaining strains, which belonged to sequevars I-18 (9.1%, *n* = 6), I-46 (7.6%, *n* = 5), and I-15 (10.6%, *n* = 7). Sequevar I-31 also predominated in the south and southeastern areas where it represented 100.0% (*n* = 14) and 90.5% (*n* = 19) of the collected strains, respectively. In the latter area, sequevars I-18 and I-15 were also identified and represented 4.8% (*n* = 1) of the diversity each. The prevalence of the four sequevars appeared to be variable depending on the host species in Mayotte (Table [S5](#SM5){ref-type=\"supplementary-material\"}). Sequevar I-31 was consistently isolated and was predominant on all sampled host species, followed by I-18 on four host species and I-46 and I-15 on 2 host species each. Sequevars I-31, I-18, and I-46 were isolated on tomatoes with frequencies of 91.3% (*n* = 73), 3.8% (*n* = 3), and 5.0% (*n* = 4), respectively. Additionally, three sequevars each were identified on sweet pepper \\[I-31 (71.4%, *n* = 5), I-46 (14.3%, *n* = 1), and I-15 (14.3%, *n* = 1)\\] and eggplant \\[I-31 (80.4%, *n* = 37), I-18 (4.3%, *n* = 2), and I-15 (15.2%, *n* = 7)\\]. Two sequevars each were identified on hot pepper and black nightshade, with I-31 and I-18 representing 66.7% (*n* = 2) and 75.0% (*n* = 3) of the identifications on the hot peppers and 33.3% (*n* = 1) and 25.0% (*n* = 1) of the identifications on the black nightshade, respectively. A neighbor-joining tree based on partial *egl* sequences allowed us to determine the phylogenetic positions of the 140 Mayotte strains compared with the worldwide reference strains of RSSC phylotypes I (18 strains), II (31 strains), III (12 strains), and IV (6 strains) (Figure [2](#F2){ref-type=\"fig\"}). The 120 sequevar I-31 strains isolated in Mayotte were phylogenetically identical to the reference sequevar I-31 strain JT519 from La R\u00e9union. The eight sequevar I-15 strains and the five sequevar I-46 strains isolated in Mayotte appeared to be phylogenetically identical to the reference sequevar I-15 and I-46 strains PSS358 and MAD17 from Taiwan and Madagascar, respectively. The seven sequevar I-18 strains isolated in Mayotte were phylogenetically closely related to the reference sequevar I-18 strain GMI1000 from French Guiana.\n\n![Phylogenetic neighbor-joining tree based on partial endoglucanase (*egl*) gene sequences from strains from Mayotte and reference RSSC strains. The number localized at each node is the bootstrap value (*n* = 5,000), with significance \\<100% indicated at each node. The scale bar represents 1/100 nucleotide substitutions.](fpls-08-02209-g0002){#F2}\n\nTomato resistance to phylotype I in mayotte\n-------------------------------------------\n\nBacterial wilt progression on 10 tomato accessions was followed for 2 months in a field naturally contaminated with sequevar I-31, which was the most prevalent sequevar in Mayotte. The statistical analysis (allowed us to separate the 10 tomato accessions into six groups (Figure [3](#F3){ref-type=\"fig\"}). The first goup included the most susceptible control T10 with 100.0% of wilted rate. The second group comprised susceptible accessions tomatoT2, T1, and T3, which showed 87.5, 80, and 72.5%, wilted plants, respectively, at the end of the experiment. The third group was composed of one tomato accession (T4) that showed 64.1% of wilted plants. The fourth group included also one accession, T9 that appeared to be moderately resistant, with 32.5% wilted plants. The fifth group was resistant with one accession T7 showing 15.0% wilted plants. The last group was composed of the three most resistant tomato accessions (T5, T6, and T8), resulting in 8.3, 2.9, and 2.5% wilted plants, respectively. Considering these results, the resistance properties of the six tomato accessions (T4, T5, T7, T8, T9, and T10) belonging to five groups were further evaluated under controlled conditions.\n\n![Bacterial wilt progression curves on 10 Core tomato accessions screened under field conditions at the Dembeni experimental station during the 2011 hot and wet season. Means and standard errors are shown. Homogeneous groups were created transforming wilted datas with arcsin square root function and performing variance analysis followed by a Tukey\\'s HSD test.](fpls-08-02209-g0003){#F3}\n\nThe virulence of 8 strains (2 strains each of sequevar I-15, I-18, I-31, and I-46) representing phylotype I sequevar diversity in Mayotte (YT) were tested on 6 tomato accessions and 2 tomato cultivars that were widely cropped in Mayotte (Table [2](#T2){ref-type=\"table\"}). Interestingly, the mean wilting rate of the six tomato accessions used in the field and the growth chamber assays appeared to be very similar (Table [S6](#SM6){ref-type=\"supplementary-material\"}). The susceptible control accession T10 scored as highly susceptible to all tested strains \\[wilting incidence (= W) between 100 and 76.7% and colonization index (= CI) between 100 and 75%\\], even though this accession was slightly more affected by the two sequevar I-18 strains (RUN2083 and RUN2150; W = 30.0 and 53.3% and CI = 35.3 and 53.3%, respectively). Higher levels of resistance were observed with T5 (W from 0 to 10% and CI from 0 to 25.3%) and T8 (W from 0 to 8.1% and CI from 0% to 26.7%). T5 was classified as highly resistant (phenotype score *P* = 1) to 6 strains and moderately resistant (*P* = 2) to 2 strains. T8 was classified as highly resistant to 5 strains and moderately resistant to 3 strains. The commercial hybrid cultivar Platinum appeared to be more resistant than the commercial hybrid cultivar Cobra. The Platinum cultivar was classified as highly resistant to 4 strains (I-15, I-18, and I-46) and moderately resistant to 4 strains (I-15, I-31, and I-46), whereas the Cobra cultivar was classified as highly resistant to two strains (I-18), moderately resistant to four strains (I-15 and I-46), partially resistant to one strain (I-31) and moderately susceptible to one strain (I-31).\n\n###### \n\nPhenotypic responses of six tomato accessions and two commercialized tomato cultivars inoculated with the Core-RsYT under controlled conditions in La Reunion and compared with reference data from the Core-Tep/Core-Rs2.\n\n  **Accessions**[^**a**^](#TN1){ref-type=\"table-fn\"}   ***R. pseudosolanacearum* strain*s* from Mayotte (phylotype-sequevar)**[^**b**^](#TN2){ref-type=\"table-fn\"}                                                                                                                                                                                                                                         \n  ---------------------------------------------------- ------------------------------------------------------------------------------------------------------------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- --------- --------- --- --- --------- --------- --- --- --------- ------- --- --- --------- ------ --- --- --------- ------ --- --- --------- ------ --- ---\n  C                                                    30.0                                                                                                          46.4       3.1        3          56.7       68.2       4          6          0.0       3.0       1   1   0.0       13.7      1   1   10.0      23.2    2   2   10.0      24.8   2   2   30.0      41.6   2   3   10.0      17.0   2   2\n  P                                                    20.0                                                                                                          36.0       2          3          23.3       39.0       2          3          0.0       13.7      1   1   0.0       5.3       1   1   **0.0**   13.7    1   1   13.3      26.2   2   2   6.7       24.4   2   2   3.3       11.7   1   1\n  T4                                                   58.5                                                                                                          51.2       4          5          56.0       70.3       4          6          5.6       18.3      2   1   6.7       12.9      1   1   3.3       17.3    1   1   10.0      26.4   2   2   30.0      41.6   2   3   21.1      29.3   2   3\n  T5                                                   **0.0**                                                                                                       15.3       1          1          10.0       25.3       2          2          **0.0**   **0.0**   1   1   **0.0**   **0.0**   1   1   **0.0**   8.0     1   1   **0.0**   8.3    1   1   6.7       14.5   1   1   10.0      16.6   2   2\n  T7                                                   16.7                                                                                                          38.5       2          3          16.7       27.3       2          2          **0.0**   **0.0**   1   1   10.0      18.6      2   2   3.3       19.8    2   1   13.3      20.7   2   2   **0.0**   11.0   1   1   6.7       16.3   2   1\n  T8                                                   8.1                                                                                                           26.7       2          2          3.3        22.4       2          1          **0.0**   **0.0**   1   1   **0.0**   8.3       1   1   **0.0**   9.0     1   1   **0.0**   11.0   1   1   **0.0**   20.3   2   1   **0.0**   8.0    1   1\n  T9                                                   26.7                                                                                                          39.5       2          3          40.0       53.3       3.1        4          6.7       13.0      1   1   3.3       6.3       1   1   16.7      40.6    2   3   **0.0**   14.0   1   1   20.0      30.0   2   3   26.7      37.1   2   3\n  T10                                                  93.3                                                                                                          90.0       5          8          100.0      100.0      5          8          30.0      35.3      2   3   53.3      53.3      4   5   100.0     100.0   5   8   76.7      88.6   4   7   96.7      90.0   5   8   83.3      75.0   4   7\n  Pathotype                                            type T-2                                                                                                      type T-3   type T-1   type T-1   type T-1   type T-1   type T-1   type T-1                                                                                                                                                            \n\n*Tomato accessions. T, pathotype on tomato; C, Cobra F1; P, Platinum F1; Path, Classification of Mayotte tomato pathotypes compared to those of the tomato Core-tep/CoreRs2*.\n\n*W, wilted plants (%); CI, colonization index (%); P, phenotype; C, Mayotte classification, identified using the k-nearest neighbors classification with reference to the Core-Tep for P and Prof, the tomato pathotype from the Core-tep/CoreRs2 for Path and by agglomerative hierarchical nesting classification and the average linkage method for C. P scale: 1, highly resistant; 2, moderately resistant; 3.1, partially resistant, 3.2, latent infection; 4, moderately susceptible; and 5, highly susceptible (Lebeau et al., [@B15])*.\n\nThe eight strains were separated into three clusters by referring to previously defined pathotype typing (Lebeau et al., [@B15]) (Table [2](#T2){ref-type=\"table\"}). The first cluster encoded T-1 and grouped the less virulent strains belonging to sequevars I-15, I-18, and I-46. Clusters T-2 and T-3 comprised the more virulent strains RUN2170 and RUN2108 that belonged to the sequevar I-31 strains. Strain RUN2108 isolated at the Dembeni site was highly virulent to the 6 tomato lines and the commercial cultivars used during this study. An agglomerative hierarchical nesting classification was used to distinguish the phenotypic variability of the strains on the tomato lines and cultivars used in this study. Two clusters ranked as types YT-1 and YT-2 were identified. Again, the two sequevar I-31 strains scored with the highest virulence and were typed as YT-2, whereas strains I-18 and I-46 were assigned to the weakest virulence type YT-1. The sequevar I-15 strains were distributed in YT-1 (RUN2140) and YT-2 (RUN2143).\n\nDiscussion {#s4}\n==========\n\nThe RSSC strains are distributed into four monophyletic clusters of strains termed phylotypes based on a hierarchical classification scheme. These phylotypes are further subdivided into sequevars based on polymorphisms of the endoglucanase gene (*egl*) (Fegan and Prior, [@B6]). RSSC was reported to be broadly diverse in most subtropical and tropical areas where wide sampling and genetic diversity studies were performed. For example, in Africa, three phylotypes (I, II, and III) were identified in Cameroon and the Ivory Coast (Mahbou Somo Toukam et al., [@B19]; N\\'Guessan et al., [@B20]), and two phylotypes (I and II) were identified in Ethiopia (Lemessa and Zeller, [@B16]). In Asia, three phylotypes (I, II, and IV) were reported in India (Sagar et al., [@B29]), and two phylotypes (I and IV) were identified in Japan (Horita et al., [@B13]) and (I and II) China (Xu et al., [@B35]).\n\nIn this study, we assessed the phylogenetic diversity of RSSC in Mayotte. We found that the phylogenetic diversity was highly homogeneous, since only phylotype I was identified (*R. pseudosolanacearum*) according to the most recent RSSC taxonomy (Safni et al., [@B28]; Prior et al., [@B23]). Phylotype I is the most prevalent phylotype in other southwest Indian Ocean islands (SWIO), such as Comoros, Mauritius, Reunion, Rodrigues, and Seychelles, where it accounts for 87% of the strain phylogenetic diversity and is mainly isolated from solanaceous crops (Yahiaoui et al., [@B36]). Phylotype I has also been reported in Madagascar and eastern African countries bordering the Indian Ocean, such as Kenya and South Africa (Wicker et al., [@B34]; Carstensen et al., [@B2]; Ravelomanantsoa, [@B25]). Phylotype I affects a wide range of crops that include both herbaceous and woody plants (Hayward, [@B10]), is distributed worldwide (Hayward, [@B12]) and is reported to be highly recombinogenic (Coupat et al., [@B4]; Wicker et al., [@B34]). Phylotype I is known for its broad infrasubspecific diversity and comprises 16 out of the 57 sequevars that are currently known. In Mayotte, we identified four sequevars (I-15, I-18, I-31, and I-46) of which sequevar I-31 had a high prevalence (85.7%). Interestingly, a similar situation was reported in another island in Taiwan, where only phylotype I was identified based on a 58-strain collection with the exception of two introduced *R. solanacearum* phylotype IIB-1 isolates (Lin et al., [@B17]). However, in Taiwan, the intraspecific diversity was different and was much higher than the diversity in Mayotte, since 10 sequevars were identified (Lin et al., [@B17]) and sequevar I-15 was the most prevalent (60.7%). Importantly, although the strains in Mayotte were isolated only from solanaceous plants, the strains in Taiwan were isolated from 22 host species ranging from annual herbaceous plants to perennial woody plants, with solanaceous and non-solanaceous species representing 64.2 and 35.8% of the collected strains, respectively. These results obtained in Mayotte and Taiwan on solanaceous and non-solanaceous crops emphasize the importance of host species as a motor driving RSSC genetic diversity. This finding strongly suggests that futures studies focusing on RSSC genetic diversity should take into account host diversity by collecting RSSC from both agro- and natural ecosystems.\n\nIn Mayotte, we found that the most prevalent sequevar was sequevar I-31, which represented 85.7% of sequevars collected. Our study showed that the I-31 strains from Mayotte were phylogenetically identical to the sequevar I-31 strains from La Reunion. I-31 has also been reported in other southwest Indian Ocean islands, such as Comoros, Mauritius, Reunion, Rodrigues, and Seychelles (Yahiaoui et al., [@B36],[@B37]), as well as in Brazil (Rodrigues et al., [@B27]), and some African countries, such as the Ivory Coast (N\\'Guessan et al., [@B20]), Democratic Republic of Congo, Uganda, South Africa (Carstensen et al., [@B2]), Benin (Sikirou et al., [@B31]), and Kenya (unpublished data). Interestingly, similar to the situation in Mayotte, sequevar I-31 was also the most widespread and prevalent sequevar in countries where extensive RSSC surveys were conducted, including the Ivory Coast (N\\'Guessan et al., [@B20]) and southwest Indian Ocean islands (Yahiaoui et al., [@B36]). The higher prevalence of I-31 strains may be explained by their higher virulence. I-31 strains have the capacity to infect a wide host range. In Mayotte, I-31 strains were isolated from the 5 solanaceous plants sampled (tomato, eggplant, hot pepper, sweet pepper, and black nightshade). In other countries, these strains have been isolated on both herbaceous plants (*Solanaceae, Geraniaceae, Begoniaceae*, and *Asteraceae*) (N\\'Guessan et al., [@B20]; Rodrigues et al., [@B27]; Yahiaoui et al., [@B36],[@B37]); unpublished data) and woody plants (Eucalyptus) (Carstensen et al., [@B2]). Moreover, in the Ivory Coast, I-31 strains were characterized by a broad virulence spectrum and high virulence level on all tomato and eggplant accessions tested (except two eggplant accessions) and were assigned to the two highest virulence pathoprofiles (e and f) defined by Lebeau et al. ([@B15]). In Mayotte, the sequevar I-31 strains were clustered into the most virulent pathotypes T-2 and T-3 referring to the classification of Lebeau et al. ([@B15]) when tested on tomatoes, whereas strains of sequevars I-18, I-46, and I-15 were clustered into the least virulent pathotype T-1. Under field and growth chamber conditions, Hawaii 7996 (T5) and R3034 (T8) showed the highest levels of resistance amongst all tomato breeding lines tested against the sequevar I-31 strains and the strains belonging to the three other sequevars (I-15, I-18, and I-46) uncovered in Mayotte. In the Ivory Coast, Hawaii 7996 (T5) and R3034 (T8) were also the best genitors against phylotype I and IIA strains even though their resistance was overcome by strains from two agroecological zones (N\\'Guessan et al., [@B20]). These results are consistent with the virulence tests conducted with the Core-Rs2 collection, which comprised global strains from phylotypes I, IIA, IIB, and III (Lebeau et al., [@B15]).\n\nInterestingly, in Mayotte, the most resistant genitors \\[Hawaii 7996 (T5), R3034 (T8), CLN1463 (T7), and TML46 (T6)\\] to the four sequevars (I-15, I-18, I-31, I-46) were selected by AVRDC in Taiwan. Sequevars I-15 and I-18 but not sequevars I-31 and I-46 were also reported in Taiwan (Lin et al., [@B17]), indicating that these genitors were effective against strains not found in Taiwan. In Mayotte, the commercial hybrid Platinum F1 was classified as highly or moderately resistant depending on the strain tested, suggesting that resistance in this hybrid likely came from the AVRDC breeding lines. The other commercial hybrid (Cobra F1) was not resistant, particularly against the I-31 strains, strongly suggesting that its deployment in Mayotte was considered riskier and required the implementation of additional control measures against bacterial wilt. Altogether, these results represent a real incentive for plant breeders to use resistant tomato breeding lines and the commercial hybrid Platinum as genetic resources to create and deploy adapted cultivars in Mayotte.\n\nThis study is the first report linking RSSC virulence to phylotype I genetic diversity in Mayotte. This study represents an important starting point for regional breeding programmes for bacterial wilt resistance and deployment strategies not only in Mayotte but also in neighboring southwest Indian Ocean islands and African countries, where the I-31 sequevar is widespread. To understand the epidemiological and evolutionary relationships associated with this genetic lineage, population structure studies must be conducted using methods such as multilocus sequence typing (MLST) and multilocus variable number tandem repeat analysis (MLVA), which have been developed to study RSSC populations on different scales (from cropping areas to continents) (Wicker et al., [@B34]; N\\'Guessan et al., [@B21]; Parkinson et al., [@B22]; Ravelomanantsoa et al., [@B26]; Guinard et al., [@B9]).\n\nConclusion {#s5}\n==========\n\nThis is the first study of RSSC epidemiology performed in a small isolated island. Low phylogenetic diversity was identified with only Phylotype I strains collected on solanaceous crop. The absence of cross interactions with phylotype II and III strains provide a unique environment to better understand phylotype I plant resistance properties from worlwide breeding projects.\n\nAuthor contributions {#s6}\n====================\n\nPP, SP, MR-C, and TC: designed the study; LV, TC, and PP: prospected the strains; CB, J-JC, and TC: performed the microbiology and molecular biology experiments; GM and TC: conducted the phenotyping experiments; TC: drafted the manuscript; PP, MR-C, and SP: contributed to writing, drafting and revising the manuscript. All authors approved the final manuscript.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This study was supported by grants from the Office de D\u00e9veloppement de l\\'Economie Agricole des D\u00e9partements d\\'Outre Mer (ODEADOM) as part of the R\u00e9seau d\\'Innovation et de Transfert Agricole (RITA). The authors thank all of the contributors, including the farmers and technical staff who helped isolate RSSC strains in Mayotte and screen them at the Plant Protection Platform in La R\u00e9union.\n\nSupplementary material {#s7}\n======================\n\nThe Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fpls.2017.02209/full#supplementary-material>\n\n###### \n\n*Ralstonia solanacearum* strains isolated from Mayotte Island. ^a^Mayotte location and area of strain isolations (C, Centre; N, North; NE, Northeast; S, South; SE, Southeast). ^\\*^Strain used for the virulence tests.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nRSSC reference strains used in this study to structure the phylogenetic tree.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nTomato accessions used for the virulence tests.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nSequevar distributions of the *Ralstonia solanacearum* phylotype I strains within the five vegetable production areas of Mayotte.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nSequevar distributions of the *Ralstonia solanacearum* phylotype I strains within the different host species sampled in Mayotte.\n\n###### \n\nClick here for additional data file.\n\n###### \n\nComparison table of the mean wilted rate of several tomato lines from the Core collection measured under field and controlled conditions against *Ralstonia solanacearum* phylotype I-31 strains from Mayotte. ^a^Tomato accessions from the tomato Core-Tep. ^b^Field experiment was assessed on the tomato Core-tep in Mayotte in 2011 and screened against strains characterized as *R. pseudosolanacearum* phylotype I sequevar 31. Notably, the RUN2108 (*Solanum lycopersicum*), RUN2116 (*Capsicum annuum*), and RUN2121 (*Solanum nigrum*) strains were collected in the same tunnel in 2012. ^c^Screening results obtained for strain RUN2108 (I-31), which was collected in the same tunnel in 2012 against several tomato Core-Tep lines under controlled conditions in La Reunion in 2013.\n\n###### \n\nClick here for additional data file.\n\n[^1]: Edited by: Brigitte Mauch-Mani, University of Neuch\u00e2tel, Switzerland\n\n[^2]: Reviewed by: Eduardo S. G. Mizubuti, Universidade Federal de Vi\u00e7osa, Brazil; Sudisha Jogaiah, Karnatak University, India\n\n[^3]: This article was submitted to Plant Microbe Interactions, a section of the journal Frontiers in Plant Science\n"}
+{"text": "![](indmedgaz71845-0051){#sp1 .336}\n\n![](indmedgaz71845-0052){#sp2 .337}\n"}
+{"text": "Background\n==========\n\nMultidrug-resistant tuberculosis (MDR-TB) refers to disease caused by *Mycobacterium tuberculosis* strains with *in vitro* resistance to both isoniazid (INH) and rifampin, two first line anti-tuberculosis drugs. The World Health Organization (WHO) estimates that there were 650,000 MDR-TB cases worldwide in 2008, corresponding to 3.6% of all tuberculosis (TB) cases \\[[@B1]\\]. MDR-TB is associated with high rates of patient default, treatment failure and death \\[[@B2]\\]. Moreover, MDR-TB treatment is more costly when compared with first line therapy and requires enhanced clinical and laboratory support \\[[@B1]\\]. For these reasons, MDR-TB is a threat to achieving successful TB control \\[[@B3]\\].\n\nTransmission of MDR-TB strains has been examined from several perspectives, including *in vitro* and animal studies, mathematical modeling and epidemiological investigation \\[[@B4]\\]. Results from epidemiological studies have been variable \\[[@B4],[@B5]\\]. In studies utilizing molecular typing, drug-resistant strains demonstrate both increased and decreased clustering \\[[@B6]-[@B12]\\], while data from traditional contact tracing studies note variable proportions of tuberculin skin test (TST) positivity and TB disease in MDR-TB contacts \\[[@B13]-[@B20]\\].\n\nContact tracing studies consistently report a significant proportion of contacts with TB disease that demonstrate a distinct resistance profile from their identified MDR-TB source, indicating that not all supposed transmission events involve MDR-TB strains \\[[@B12],[@B14]-[@B20]\\]. Despite this discrepancy, there is considerable evidence to support human-to-human MDR-TB strain transmission. Indeed over half of global MDR-TB cases are thought to result from primary transmission \\[[@B21]\\]. Yet appropriate preventative treatment of MDR-TB contacts remains unclear, as the relative infectivity MDR-TB strains is unknown and no high quality evidence exists to guide treatment of latent TB infection (LTBI) in contacts of drug resistant cases \\[[@B22]\\].\n\nWe examined local MDR-TB contact tracing outcomes in a low incidence setting. To better understand the transmissibility of MDR-TB strains, the proportion of close contacts of MDR-TB source that developed TB disease or were found to be TST positive in follow-up was compared to the proportion of close contacts of drug-susceptible TB (DS-TB) and isoniazid mono-resistant TB (HMR-TB).\n\nMethods\n=======\n\nSetting\n-------\n\nThe study was conducted in British Columbia (BC), a Canadian province with a population of 4.4 million and a TB case rate of 7.1 per 100 000 population per year \\[[@B23]\\]. The BC Centre for Disease Control (BCCDC) maintains a provincial population based TB registry that is informed of all TB cases through legal notification, as well as case notification through a centralized provincial mycobacteriology laboratory and pharmacy. The same agency dispenses all medications used to treat contacts. Contact tracing was performed according to Canadian guidelines and was recorded in the registry using standardized protocols \\[[@B24]\\].\n\nBacterial culture and susceptibility testing\n--------------------------------------------\n\n*M. tuberculosis* was isolated from clinical specimens using the BacT/Alert mycobacterial culture detection system (BioMerieux, Durham, NC). Phenotypic susceptibility testing was performed using the BACTEC 460 radiometric method and interpreted as per Clinical and Laboratory Standards Institute (CLSI) recommendations \\[[@B25],[@B26]\\].\n\nDefinitions\n-----------\n\nFor the purposes of this study, a *source case* was defined as the first household member to present with culture-positive pulmonary TB and drug susceptibility profile results. The source case date of diagnosis was defined as the date that the first culture positive sample was received by the laboratory. A *close contact* was defined as any individual identified as a \"household contact\" or \"Type 1 contact\" of a source case in the provincial TB registry. In this registry, both \"household contact\" and \"Type 1 contact\" represent a close contact and refer to household contacts or those sharing airspace with the source case for \\>4 hours per week. TSTs were performed according to Canadian guidelines \\[[@B24]\\]. A *positive TST* in a given contact was defined as any TST \u22655 mm measured \\<3 months before to \\<1 year after source diagnosis. *Prior positive* referred to a positive TST \u22653 months before source diagnosis. A close contact identified by the source case without a TST measurement or TB disease diagnosis was recorded as *no result*. Contacts with a positive TST or with signs and symptoms of TB disease were screened with a chest x-ray and in some cases, sputum examination. C*ontacts with TB disease* was defined as a contact with *M. tuberculosis* complex demonstrated on culture or an individual with radiological, pathological or therapeutic responses consistent with TB disease as per established guidelines. Contacts with TB disease were not classified by their TST result.\n\nData acquisition and analysis\n-----------------------------\n\nWe accessed the BCCDC registry for all cases of MDR-TB, HMR-TB and DS-TB diagnosed between 1990 and 2008. Close contacts of all MDR-TB cases were identified from the registry and their demographic profile, clinical features and TST results were recorded. For comparison, close contacts of DS-TB and HMR-TB were also identified with data recorded in a similar fashion. Only close contacts were recorded to limit bias introduced by enhanced contact tracing of drug-resistant sources. To further limit bias introduced by misclassification of close contacts, any sources with \\>15 close contacts were excluded from analysis.\n\nThe number of contacts with TB disease was recorded, and when available, contact resistance profiles were compared with their purported source case. TST positive or TST negative contacts with two sources listed within the same household in the same year were excluded from further TST analysis. BCG status was extracted for each contact, along with demographic variables. When BCG status was unavailable (in 36% of cases), BCG status was estimated from <http://www.bcgatlas.org> using data on country of origin, age and year of entry to BC \\[[@B25],[@B26]\\].\n\nStatistical analysis\n--------------------\n\nDescriptive statistics were computed using Stata version 11.1 (StataCorp, College Station, Texas). The Wilcoxon Rank Sum test, chi-square test and Ficher's exact test were used in univariate analysis to assess statistical difference between variables (alpha 0.05). A multivariable logistic regression model was constructed to test the association between source resistance profile and TST positivity. Factors known to predict TST positivity, including adult age (\u226518 years), source smear status (positive), male gender, foreign birth, and BCG status were included in the model. After analysis of contact data, the number of contacts per source was added to the model, as the number of contacts varied significantly between HMR contacts and DS-TB contacts. A second model was constructed which excluded BCG status given the limited data for this variable. Odds ratios with 95% confidence intervals were reported. Models were assessed with chi-square goodness-of-fit and ROC curve construction.\n\nResults\n=======\n\nMDR-TB population\n-----------------\n\nWe identified 35 MDR-TB cases from the BCCDC registry between 1990 and 2008 (Table [1](#T1){ref-type=\"table\"}). Contact tracing was not performed for 7 cases: 5 had extra-pulmonary TB, 1 case entered BC on MDR-TB therapy, and 1 case left BC before contact tracing was initiated. From the remaining 28 MDR-TB sources, 89 close contacts were identified, with a median of 3 total contacts per source (range 1-7) and median follow-up of 123 months (range 19-239). Of the 89 close contacts, 42 (47%) were TST positive, 33 (37%) were TST negative and 9 (10%) were prior positive or no result (Table [2](#T2){ref-type=\"table\"}). Latent TB therapy was completed in 12 MDR-TB contacts, including 11 contacts that initiated preventative therapy tailored to the source case susceptibility profile. Five close contacts developed TB disease during follow-up; fully susceptible *M. tuberculosis* was isolated from 4 cases and the fifth case was diagnosed based on clinical criteria. All 5 contacts with TB disease were diagnosed within 3 months of source diagnosis and had not received preventative therapy.\n\n###### \n\nCharacteristics of TST positive and negative contacts\n\n  \u00a0                              **DS-TB (%)**   **HMR-TB (%)**   **MDR-TB (%)**   **p-value vs DS-TB**  \n  ----------------------------- --------------- ---------------- ---------------- ---------------------- ---------\n  Source Cases                       2895              96               28                  \u00a0                \u00a0\n  Household contacts                 7309             249               89                  \u00a0                \u00a0\n  Median contacts/source               3               3                3               \\<0.001\\^         0.839\\^\n  **Source Characteristic**                                                                              \n  Source Smear Positive            4709 (64)        139 (56)         57 (64)              0.005            0.939\n  **Contact Characteristics**                                                                            \n  **Demographics**                                                                                       \n  Mean Age (Sd)                   32.9 (21.2)     28.9 (20.4)      26.7 (19.1)              \u00a0                \u00a0\n  Age Over 18                      5489 (75)        170 (68)         57 (64)              0.013            0.015\n  Male Gender                      3521 (48)        112 (45)         43 (48)               0.32            0.98\n  **Birth Country**                                                                                      \n  Canadian born                    3163 (43)        87 (35)          60 (67)                \u00a0                \u00a0\n  Foreign born                     3609 (49)        154 (62)         23 (26)                \u00a0                \u00a0\n  Unknown                           536 (7)          8 (3)            6 (7)                 \u00a0                \u00a0\n  **BCG status\\***                                                                                       \n  Positive                         3340 (46)        132 (53)         44 (49)                \u00a0                \u00a0\n  Negative                         2977 (41)        94 (38)          42 (47)                \u00a0                \u00a0\n  Unknown                          991 (14)          23 (9)           3 (3)                 \u00a0                \u00a0\n\n\\^Wilcoxon rank sum test.\n\n\\* 22% of values were calculated using age, country of origin and information from <http://www.bcgatlas.org>.\n\n###### \n\nResults of contact tracing\n\n  \u00a0                 **DS-TB (%)**   **HMR-TB (%)**   **MDR-TB (%)**   **p-value vs DS-TB**  \n  ---------------- --------------- ---------------- ---------------- ---------------------- -------\n  Source cases          2895              96               28                  \u00a0               \u00a0\n  Close contacts        7309             249               89                  \u00a0               \u00a0\n  TST positive        2321 (32)        121 (49)         42 (47)             \\<0.001          0.002\n  Secondary case       168 (2)          8 (3)            5 (6)               0.347           0.039\n  Prior positive       448 (6)          12 (5)           8 (9)               0.395           0.265\n  No result            475 (6)          9 (4)            1 (1)               0.067           0.045\n\nIsoniazid-resistant and drug-susceptible populations\n----------------------------------------------------\n\nBetween 1990 and 2008, contact tracing of 96 infectious HMR-TB source cases yielded 249 close contacts, of whom 121 (49%) were TST positive (Table [2](#T2){ref-type=\"table\"}) and 8 (3%) developed TB disease during follow-up. Of the six contacts with culture-confirmed TB, 3 developed HMR-TB, while 3 developed DS-TB (Table [3](#T3){ref-type=\"table\"}). DS-TB contact tracing over the same period yielded 7309 close contacts from 2895 sources (Table [2](#T2){ref-type=\"table\"}). There were 2321 contacts (32%) with a positive TST and 168 (2%) contacts with TB disease. Despite a common median of 3 close contacts per source case, the distribution of contacts varied significantly between DS-TB and HMR-TB (p\u2009\\<\u20090.001).\n\n###### \n\nDrug resistance patterns in source and secondary cases\n\n  **Related pairs**    **Resistance pattern**              \n  ------------------- ------------------------ ----- ----- ---\n  MDR-TB1                       0.1             2.0   4.0   S\n  S                              S               S     S   \n  MDR-TB2                       0.1             2.0   NR    S\n  S                              S               S     S   \n  MDR-TB3                       0.1             2.0   4.0   S\n  S                              S               S     S   \n  MDR-TB4                       0.4             2.0   2.5   S\n  S                              S               S     S   \n  MDR-TB5                       0.4             2.0   2.5   S\n  S                              S               S     S   \n  HMR-TB1                       0.1              S     S    S\n  0.1                            S               S     S   \n  HMR-TB2                       0.1              S     S    S\n  0.1                            S               S     S   \n  HMR-TB3                       0.1              S     S    S\n  S                              S               S     S   \n  HMR-TB4                       0.1              S     S    S\n  S                              S               S     S   \n  HMR-TB5                       0.1              S     S    S\n  0.1                            S               S     S   \n  HMR-TB6                       0.1              S     S    S\n  S                              S               S     S   \n\nThe first row in each pair represents source case sensitivity profile, while the second represents the secondary case susceptibility. *S* = sensitive, *NR* = no result. Resistance values reported in microG/microL.\n\nMultivariate analysis of TST results\n------------------------------------\n\nAll seven pre-specified variables predicted TST positivity in the logistic regression model, including adult age (OR 1.76; 95% CI 1.51-2.06), male gender (OR 1.18; 95% CI 1.05-1.32), BCG vaccination (OR 1.42; 95% CI 1.23-1.64), foreign birth (OR 5.37; 95% CI 4.55-6.33), source smear positivity (OR 1.23; 95% 1.09-1.39), source HMR-TB (OR 2.13; 95% CI 1.57-2.90), and source MDR-TB (OR 1.75; 95% CI 1.07-2.86) (Table [4](#T4){ref-type=\"table\"}). The number of close contacts attributed to the index case was included in the model because the distribution varied significantly between the HMR-TB and DS-TB groups, and because a lower threshold for classifying contacts as Type 1 may impact the rate of TST positivity. Each single increase in the number of close contacts was associated with a decrease in TST positivity (OR 0.96; 95% CI 0.94-0.98). Estimates did not change significantly after BCG status was excluded from the model. Chi-square goodness of fit was non-significant for models that included and excluded BCG status, indicating that the goodness-of-fit for this model appeared adequate (p\u2009=\u20090.32, p\u2009=\u20090.09 respectively).\n\n###### \n\nMultivariate analysis of factors associated with TST positivity\n\n  **Variable**                                 **Adjusted OR**   **p-value**   **95% CI**\n  ------------------------------------------- ----------------- ------------- -------------\n  Age over 18                                       1.77           \\<0.001     1.53 - 2.06\n  Male gender                                       1.19            0.002      1.07 - 1.33\n  Foreign birth                                     6.97           \\<0.001     6.11 - 7.95\n  HMR-source                                        1.99           \\<0.001     1.48 - 2.67\n  MDR-source                                        1.72            0.030      1.05 - 2.81\n  Smear positive source                             1.23            0.001      1.09 - 1.38\n  Increase in source contacts (per contact)         0.96           \\<0.001     0.95 - 0.98\n\nDiscussion\n==========\n\nGuidelines addressing latent TB infection (LTBI) treatment in MDR-TB contacts are vague and somewhat contradictory \\[[@B24],[@B27],[@B28]\\]. The WHO, citing the unknown efficacy of tailored preventative regimens, recommends observing contacts with careful clinical follow-up over two years \\[[@B28]\\]. The American Thoracic Society recommends observation or treatment with one of two regimens for 6-12 months \\[[@B27]\\]. These recommendations reflect the lack of high quality evidence required to direct decisions. Indeed, a recent Cochrane review failed to identify randomized control trials on MDR-TB contact tracing, while a 2006 systematic review found only two comparative studies suitable for analysis \\[[@B23],[@B29]\\].\n\nGiven our results, combined with the discouraging completion rates of susceptibility-profile tailored regimens \\[[@B30]-[@B32]\\], IPT could be seen as a potential option in close contacts of MDR-TB patients in low incidence settings. This strategy, however, is not without risk \\[[@B33]\\]. MDR-TB transmission does occur in low incidence settings \\[[@B12]\\]. Moreover, in high incidence settings, the majority of patients with TB disease in MDR-TB contact populations develop MDR-TB \\[[@B16],[@B18],[@B19]\\]. Such contacts will not likely benefit from IPT. More importantly, however, this strategy could propagate drug-resistant disease by selectively killing drug sensitive organisms \\[[@B34]-[@B37]\\]. Thus, the risk of IPT appears to outweigh its benefit in MDR-TB contacts.\n\nOur results demonstrate that MDR-TB contacts have higher rates of TST positivity compared with DS-TB contacts. This may reflect the increased transmissibility of MDR-TB strains. Alternatively, the high rates of TST positivity in MDR-TB and HMR-TB contacts may reflect differential distribution of unmeasured determinants for LTBI, such as source time-to-diagnosis, source cavitary disease or contact environmental and socioeconomic determinants. Indeed the development of DS-TB in all close contacts of MDR-TB with incident TB disease supports the notion that close contacts of drug-resistant source cases may be at higher risk for LTBI and TB disease independent of transmission from an identified MDR-TB source case.\n\nThere are several limitations in our study, the most significant being the lack of molecular typing data, which could help determine whether the discrepant drug susceptibility profiles were from strains with the same genotype. A second limitation is the small population of drug-resistant cases and contacts available for analysis. Previous studies examining traditional contact tracing in MDR-TB patients have demonstrated disease rates ranging from 0-8% in close contacts, consistent with our data \\[[@B10]-[@B19]\\]. However, in studies examining drug-susceptibility profiles, 62-92% of contacts that subsequently developed active disease had MDR-TB \\[[@B13]-[@B16],[@B18]-[@B20]\\]. These rates are consistent with data demonstrating that 70% of secondary cases have the same genotype as their source case \\[[@B38]\\].\n\nFinally, data on several determinants for TB infection are absent from analysis, including time-to-diagnosis, socioeconomic status and medical co-morbidities. Differential rates of TST positivity may be, in part, related to the differential distribution of these determinants. Nonetheless, our data suggests that risk for *M. tuberculosis* infection is higher in the contacts of drug resistant cases, and that thorough contact tracing should be performed in this population.\n\nConclusions\n===========\n\nIn conclusion, we described our experience with the contact tracing results of drug-resistant TB in a low prevalence region over nearly two decades. Our findings demonstrate that close contacts of MDR-TB at higher risk for LTBI and active TB. We suggest enhanced contact tracing in MDR-TB contacts, but caution against IPT in this population. Further research is urgently needed to determine the optimal management of LTBI in MDR-TB contacts.\n\nAbbreviations\n=============\n\nTB: Tuberculosis disease (also known as active tuberculosis); LTBI: Latent tuberculosis infection; MDR-TB: Multidrug resistant tuberculosis; DS-TB: Drug sensitive tuberculosis; HMR-TB: Isoniazid mono-resistant tuberculosis; INH: Isoniazid; IPT: Isoniazid preventative therapy; WHO: World Health Organization; TST: Tuberculin Skin Test.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nJJ: design, data acquisition, writing, statistical analysis. AA: statistical analysis, writing. AI: data acquisition. writing. PT: data acquisition, writing. VC: design, writing. KE: design, writing. MF: design, writing. All authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2334/12/266/prepub>\n"}
+{"text": "Introduction {#s1}\n============\n\n*Staphylococcus aureus* is one of the most prevalent and clinically significant pathogens worldwide, causing a variety of illnesses ranging from benign, superficial skin eruptions to life-threatening infections with bacteraemia, endocarditis, pneumonia and toxic shock syndrome [@pone.0027328-Lowy1]. Since methicillin-resistant *S. aureus* (MRSA) was first identified in 1961, it has become the most common cause of nosocomial and community infections worldwide [@pone.0027328-Deresinski1]. The recent emergence of highly virulent community-associated MRSA (CA-MRSA) isolates which cause disease in individuals with no apparent risk factors for hospital acquisition of MRSA has raised widespread concern over how CA-MRSA is transmitted in the community setting [@pone.0027328-Taiwo1], [@pone.0027328-Cohen1].\n\n*S. aureus* is a member of commensal microflora and readily colonizes the anterior nares in humans. Many infections caused by *S. aureus* occur in persons with prior nasal carriage [@pone.0027328-Wertheim1], and this carriage is an important risk factor for nosocomial *S. aureus* infection in patients undergoing surgery, hemodialysis, implantation of intravascular devices, and among HIV-infected patients [@pone.0027328-Kluytmans1]. One study revealed that nasal carriage of *S. aureus* is an important source of *S. aureus* bacteremia [@pone.0027328-vonEiff1], and eliminating nasal carriage of *S. aureus* may prevent systemic *S. aureus* infection [@pone.0027328-vonEiff1]. Determination of the prevalence of *S. aureus* nasal carriage, as well as antimicrobial resistance profiles and molecular typing for nasal *S. aureus* isolates, in healthy populations is beneficial for identifying risk factors associated with *S. aureus* infection [@pone.0027328-Bischoff1], [@pone.0027328-LamaroCardoso1], [@pone.0027328-Lowy2]. Molecular typing of *S. aureus* is also helpful for supporting infection control measures, investigating suspected outbreaks and preventing nosocomial transmission [@pone.0027328-Enright1]. Identification of (Panton-Valentine Leukocidin) PVL genes, pulsed-field gel electrophoresis (PFGE), staphylococcal cassette chromosome *mec* (SCC*mec*) typing, spa typing, and multilocus sequence typing (MLST) have been used to monitor the evolutionary process of pandemic clones [@pone.0027328-Diep1]. However, studies characterizing *S. aureus* isolates in China have largely focused on isolates recovered from clinical specimens in the context of clinical disease. One surveillance study performed in China showed that 63% of *S. aureus* isolates were MRSA, including 77% of nosocomial isolates and 43% of community isolates [@pone.0027328-Xiao1]. In addition, two major epidemic MRSA clones, ST239-MRSA-SCCmec type III and ST5-MRSA-SCCmec type II, are distributed across China in unique geographic patterns [@pone.0027328-Liu1]. These data suggest that dissemination of virulent MRSA clones among healthy persons in China may contribute to the presence of clinically significant MRSA infections in some locations. To our knowledge, there are limited published studies characterizing MRSA isolates recovered from nares of healthy individuals in China who work and reside in the health care setting. Therefore, we sought to determine the prevalence, antimicrobial resistance profiles, and molecular characteristics of nasal *S. aureus* isolates from students on a large medical college campus in China.\n\nMaterials and Methods {#s2}\n=====================\n\nIsolation and identification of nasal *S. aureus* isolates {#s2a}\n----------------------------------------------------------\n\n935 volunteers from the campus of Wenzhou Medical College in Wenzhou, Southeast China without symptoms or signs of clinical illness were enrolled in the study over a one-month time period. The healthy volunteers included had not used antibiotics in past two months. Samples were collected from both anterior nares of volunteers by rotating a sterile polyester fiber-tipped swab moistened with sterile saline. Swabs were placed in 3 mL of Luria-Bertani broth and transported to the Department of Clinical Microbiology of the first Affiliated Hospital of Wenzhou Medical College. Gram-positive, catalase-positive, coagulase-positive isolates were confirmed as *S. aureus* using a Vitek-60 microbiology analyzer (bioMe\\'rieu, Marcy l\\'Etoile, France). The Ethics Committee of the first Affiliated Hospital of Wenzhou Medical College exempted this study from review because the present study focused on bacteria.\n\nSusceptibility testing {#s2b}\n----------------------\n\n*S. aureus* susceptibility to penicillin (10 \u00b5g), ampicillin/sulbactam (20/10 \u00b5g), cefazolin (30 \u00b5g), vancomycin (30 \u00b5g), erythromycin (15 \u00b5g), clindamycin (2 \u00b5g), rifampicin (5 \u00b5g), linezolid (30 \u00b5g), mupirocin (5 \u00b5g), quinupristin/dalfopristin (15 \u00b5g), tetracycline (30 \u00b5g), trimethoprim/sulfamethoxazole (1.25/23.75 \u00b5g), gentamicin (10 \u00b5g), ciprofloxacin (5 \u00b5g), and levofloxacin (5 \u00b5g) were determined using the disk diffusion method in accordance with standards recommended by the Clinical and Laboratory Standards Institute (CLSI) [@pone.0027328-Institute1]. All disks were obtained from Oxoid Ltd., and *S. aureus* ATCC 25923 was used as a quality control strain. Susceptibility of *S. aureus* to mupirocin was also determined using disk diffusion, with a zone diameter \u226514 mm using a 5-\u00b5g disc indicating susceptibility as described previously [@pone.0027328-Fuchs1]. MICs of mupirocin for mupirocin-resistant isolates were further determined by the agar dilution method in accordance with the CLSI guidelines [@pone.0027328-Institute1]. *mupA* was detected among mupirocin-resistant isolates as described previously[@pone.0027328-Anthony1].\n\nDNA Extraction {#s2c}\n--------------\n\n*S. aureus* isolates were cultured on blood agar overnight. Three to four bacterial colonies were suspended in 150 \u00b5L sterile distilled water with lysostaphin (1 mg/mL) (Sangon, China) and incubated at 37\u00b0C for 30 min. DNA was extracted using the Genomic DNA Extraction kit according to the manufacturer\\'s instructions (Sangon, China). DNA was stored at \u221220\u00b0C and prepared for PCR amplification.\n\nDetection of MRSA and PVL genes {#s2d}\n-------------------------------\n\nA multiplex PCR protocol described previously was used for simultaneous amplification of *mecA*, 16S rRNA, and PVL genes [@pone.0027328-McClure1]. Isolates harboring *mecA* were confirmed as MRSA using MRSA N315 as a positive control strain.\n\nSCC*mec* typing {#s2e}\n---------------\n\nSCC*mec* typing of MRSA isolates was performed using eight unique and specific pairs of primers for SCC*mec* types and subtypes I, II, III, IVa, IVb, IVc, IVd, and V as described previously [@pone.0027328-Zhang1]. The MRSA isolates with unexpected fragments or lacking fragments by multiplex PCR were defined as non-typeable (NT). MRSA NCTC 10442 (SCC*mec*I), MRSA N315 (SCC*mec* II), MRSA 85/2082 (SCC*mec* III), MRSA JCSC 4744 (SCC*mec* IV) and MRSA WZ153 (SCC*mec* V) were used for positive controls.\n\n*spa* typing {#s2f}\n------------\n\nThe *spa* variable repeat region from each MRSA isolate was amplified by simplex PCR oligonucleotide primers as previously described [@pone.0027328-Koreen1], [@pone.0027328-Harmsen1]. Purified *spa* PCR products were sequenced, and *spa* types were assigned by using the *spa* database website (<http://www.ridom.de/spaserver>).\n\nMultilocus sequence typing (MLST) {#s2g}\n---------------------------------\n\nMLST of MRSA isolates was conducted through amplification of internal fragments of the seven housekeeping genes of *S. aureus* as described previously [@pone.0027328-Enright2]. Following purification and sequencing of these genes, allele quantification and sequence typing were assigned using a well-characterized online database (<http://saureus.mlst.net/>). The DNA sequences of novel STs were verified by Pro. Enright and deposited in MLST database.\n\nResults {#s3}\n=======\n\nA total of 144 *S. aureus* isolates were isolated from the nares of the 935 volunteers in the study. The nasal carriage rate of *S. aureus* was 15.4% (144/935). 19.4% (28/144) of *S. aureus* isolates were identified as MRSA. The prevalence of nasal MRSA carriage was 3.0% (28/935). The resistance profiles of MRSA, MSSA and *S. aureus* isolates to antimicrobials tested were listed in [Table 1](#pone-0027328-t001){ref-type=\"table\"}. All *S. aureus* isolates tested were susceptible to vancomycin, quinupristin/dalfopristin and linezolid. The majority of *S. aureus* isolates were resistant to penicillin (96.5%), ampicillin/sulbactam (83.3%) and trimethoprim/sulfamethoxazole (93.1%). Isolates displayed complete or intermediate levels of resistance to gentamicin (16.0% and 18.1%), tetracycline (28.5% and 16.7%), ciprofloxacin (23.6% and 22.2%), clindamycin (31.2% and 23.6%) and erythromycin (52.1% and 30.6%). The resistance rates of MRSA isolates to gentamicin, ciprofloxacin, levofloxacin, erythromycin and clindamycin were significantly higher than those among MSSA isolates ([Table 1](#pone-0027328-t001){ref-type=\"table\"}). However, the intermediate resistance rates of MSSA isolates to gentamicin, tetracycline, erythromycin, clindamycin and rifampicin were higher than for MRSA isolates ([Table 1](#pone-0027328-t001){ref-type=\"table\"}). 82.1% (23/28) of MRSA isolates and 66.4% (77/116) of MSSA isolates were resistant to at least 3 different antimicrobials tested. Three MRSA isolates with no zone of inhibition to mupirocin (2.1%, 3/144) were resistant to mupirocin. The mupirocin MICs for the 3 MRSA isolates were \\>512 \u00b5g/mL determined by agar dilution method. The 3 MRSA isolates were positive for *mup*A detected by PCR.\n\n10.1371/journal.pone.0027328.t001\n\n###### Antimicrobial susceptibility profiles of MRSA, MSSA and *S. aureus* nasal isolates.\n\n![](pone.0027328.t001){#pone-0027328-t001-1}\n\n                                    MRSA (n\u200a=\u200a28)[a](#nt101){ref-type=\"table-fn\"}   MSSA(n\u200a=\u200a116)   *S. aureus*(n\u200a=\u200a144)                \n  -------------------------------- ----------------------------------------------- --------------- ---------------------- ------ ------ ------\n  penicillin                                             100                                                95.7                  96.5    0\n  Ampicillin/Sulbactam                                  92.9                                                82.8                  83.3   1.4\n  cefazolin                                              7.1                             7.1                0.9             0     2.1    1.4\n  trimethoprim/sulfametho-xazole                        85.7                              0                 94.8            0     93.1    0\n  gentamycin                                            28.6                             7.1                12.9           20.7   16.0   18.1\n  tetracycline                                          28.6                              0                 28.4           20.7   28.5   16.7\n  vacomycin                                               0                               0                  0              0      0      0\n  ciprofloxacin                                         53.6                            17.9                16.4           23.3   23.6   22.2\n  levofloxacin                                          39.3                            17.9                6.9            6.0    13.2   8.3\n  erythromycin                                           75                             14.3                46.6           34.5   52.1   30.6\n  clindamycin                                           53.6                            10.7                25.9           27.6   31.2   23.6\n  rifampicin                                             3.6                              0                 6.9            8.6    6.3    6.9\n  Linezolid                                               0                               0                  0              0      0      0\n  dalfopristin/quinupristin                               0                               0                  0              0      0      0\n  mupirocin                                             10.7                              0                  0              0     2.1     0\n\n: R, resistance; I, intermediate susceptibility.\n\nMRSA isolates expressing SCC*mec* V (15 isolates), SCC*mec* II (5 isolates), SCC*mec* III (4 isolates) and SCC*mec* IVd (3 isolates) were identified ([Table 2](#pone-0027328-t002){ref-type=\"table\"}). The remaining isolate was non-typeable by multiplex PCR. MLST revealed 16 different sequence types (STs) for 22 of the MRSA isolates, including ST6, ST15, ST25, ST59, ST88, ST188, ST338, ST438,ST398, ST943, ST946, ST1295, ST1556, ST1623, ST1778 and ST1779 were identified ([Table 1](#pone-0027328-t001){ref-type=\"table\"}). Since at least one of the 7 housekeeping genes for MLST typing could not be amplified among the remaining 6 MRSA isolates, STs for these isolates were not available. The STs most commonly identified were ST59 (4 isolates), ST25 (3 isolates), ST188 (2 isolates) and ST438 (2 isolates). Two novel STs recently deposited in the MLST database (<http://saureus.mlst.net/>), ST1778 and ST1779, were also identified.\n\n10.1371/journal.pone.0027328.t002\n\n###### Molecular characterization of MRSA nasal isolates and their antibiotic resistance profiles.\n\n![](pone.0027328.t002){#pone-0027328-t002-2}\n\n          SCC*mec*   spa type                  STs                   Resistance profile[a](#nt102){ref-type=\"table-fn\"}   *pvl* [b](#nt103){ref-type=\"table-fn\"}\n  ------ ---------- ---------- ------------------------------------ ---------------------------------------------------- ----------------------------------------\n  N2         V         t437                    1556                                      P,AML,CIP                                          \\+\n  N18        V         t163                    338                                  P,AML,SXT,E,DA,TE,RD                                    \\+\n  N30       VId        t437                     59                                     P,AML,STX,E,KE                                       \u2212\n  N39        V          NT                     438                                 P,AML,SXT,E,DA,CIP,LEV                                   \u2212\n  N44        V         t289                    943                                      P,AML,SXT,E                                         \u2212\n  N54        V         new                      88                                P,AML,E,DA,CIP,LEV,CN,TE                                  \u2212\n  N55       VId        t437                     59                                       P,AML,E,DA                                         \u2212\n  N57        V         t177                    438                              P,AML,SXT,E,DA,CIP,LEV,CN,TE                                \u2212\n  N58        II        new      NT[c](#nt104){ref-type=\"table-fn\"}               P,AML,SXT,E, DA,CIP,CN,TE                                  \u2212\n  N77        II        new                      NT                              P,AML,SXT,E,DA,CIP,LEV,CN,TE                                \u2212\n  N86        V         t571                    398                                 P,AML,SXT,E,DA,CIP,CN                                    \u2212\n  N91       III        t304                     59                                       P,AML,SXT                                          \u2212\n  N92       III        t304                     NT                              P,AML,SXT,E,DA,CIP,LEV, MUP                                 \u2212\n  N97       VId        t437                     59                                   P,AML,SXT,E,DE,TE                                      \u2212\n  N98        V         new                     946                                     P,AML,SXT,E,DA                                       \u2212\n  N99       III         NT                      6                                         P,SXT,DA                                          \u2212\n  N108       V         t189                    188                                       P,AML,SXT                                          \u2212\n  N112       II        t081                     NT                                P,AML,SXT,E, DA, CIP,LEV                                  \u2212\n  N114       V         t081                     25                                   P,AML, SXT,E,CN,TE                                     \u2212\n  N121       V          NT                     1778                             P,AML,SXT,E,DA,CIP,LEV,CN,TE                                \u2212\n  N123       V          NT                     1623                                 P,AML,SXT,CIP,CN,KZ                                     \u2212\n  N128       NT        t078                     25                                      P,AML,SXT,E                                         \\+\n  N135      III        t062                    1779                                   P,AML,E,CIP,LEV                                       \u2212\n  N145       V        t1751                    188                                  P,AML,SXT,E,CIP,LEV                                     \u2212\n  N154       V         t078                     25                                         P,AML                                            \u2212\n  N156       V         new                     1285                                        P,SXT                                            \u2212\n  N167       II       t5196                     15                                  P,AML, SXT,E,CIP,LEV                                    \u2212\n  N168       II       t1751                     NT                               P,AML,SXT,E,DA,CIP,LEV,MUP                                 \u2212\n\nP, penicillin; AML, Ampicillin/Sulbactam; CIP, ciprofloxacin; LEV, levofloxacin; SXT, trimethoprim/sulfamethoxazole; E, erythromycin; DA, clindamycin; RD, rifampicin; CN, gentamycin; KZ, cefazolin; TE, tetracycline; MUP, mupirocin.\n\n\u2212, *pvl* negative; +, *pvl* positive.\n\nNT, not typeable.\n\n19 MRSA isolates were assigned 12 previously characterized *spa* types, and 5 other isolates were assigned 5 novel *spa* types ([Table 2](#pone-0027328-t002){ref-type=\"table\"}). *spa* typing could not be assigned for the remaining 5 MRSA isolates. Furthermore, MRSA isolates were distributed among different sporadic clones. ST59-MRSA-VId-*spa* t437 was identified for 3 MRSA isolates. Three isolates, representing ST1556-MRSA-V- t437, ST25-MRSA-NT- t078 and ST338-MRSA-V- t163 clones, harbored PVL genes. One isolate exhibiting multidrug resistance was characterized as an ST398-MRSA-V-*spa* t571 clone. Three MRSA isolates typed as ST25 exhibited different spa types and SCC*mec* types and belonged to different clones. Although 2 ST188 MRSA isolates harbored SCC*mec* (V), these isolates exhibited two different *spa* types, t189 and t1751.\n\nDiscussion {#s4}\n==========\n\nThe prevalence of *S. aureus* nasal carriage among healthy adults ranges from approximately 20% to 30%, with higher prevalences in overcrowded populations [@pone.0027328-Wertheim1], [@pone.0027328-Kluytmans1], [@pone.0027328-Bischoff1], [@pone.0027328-Choi1]. In the United States, a national population-based survey conducted over a 3-year period revealed nasal *S. aureus* colonization for 32% of study participants [@pone.0027328-Gorwitz1]. In contrast, a relatively low rate of nasal *S. aureus* carriage (6.3%) was found for children under the age of 5 in India [@pone.0027328-Pathak1]. The prevalence of nasal colonization with *S. aureus* in our study was 15.4%, a prevalence similar to that observed (16%) for a suburban military camp cohort in Beijing, China [@pone.0027328-Qu1]. However, the prevalence of nasal MRSA carriage in our study (3.0%) was higher than prevalences previously reported for cohorts from developed countries, including the United States [@pone.0027328-Gorwitz1], [@pone.0027328-Sdougkos1], [@pone.0027328-Graham1]. Furthermore, no nasal MRSA colonization was found among Chinese healthy military volunteers in another study [@pone.0027328-Qu1]. In contrast, a relatively high prevalence of MRSA colonization was found (11.6%) for a cohort of healthy children aged \u226414 years in community settings in Taiwan over a 5-year period [@pone.0027328-Lo1]. These data indicate that MRSA colonization varies significantly, even within similar populations in developed countries. The isolation of nasal *S. aureus* with a single sample may result in underestimating the number of intermittent carriers. In order to improving the isolation of nasal *S. aureus*, multiple nasal swabs should be taken from both anterior nares of volunteers. Additional studies are needed to clarify relationships between MRSA colonization, age, general antibiotic use, and other variables within overcrowded populations at higher risk for MRSA infections.\n\nOur study revealed that 93% of nasal *S. aureus* isolates were resistant to trimethoprim-sulfamethoxazole, with MSSA isolates exhibiting the highest rates of resistance. These data contradict several studies reporting low rates of resistance to trimethoprim-sulfamethoxazole for *S. aureus* isolates recovered from the nares [@pone.0027328-Lo1]--[@pone.0027328-Liu2]. For example, in one study of 371 nasal *S. aureus* isolates, only 7 (1.9%) were resistant to trimethoprim-sulfamethoxazole [@pone.0027328-Lo1]. Our data are not in agreement with another study reporting trimethoprim-sulfamethoxazole susceptibility rates of 78.6% and 95.3%, for MRSA and MSSA isolates, respectively, recovered from patients in 12 cities across China [@pone.0027328-Wang1]. Similar to other surveillance studies [@pone.0027328-Xiao1], [@pone.0027328-Wang1], we found high rates of resistance for nasal *S. aureus* isolates to penicillin, ampicillin/sulbactam and erythromycin, and no discernable resistance of these isolates to vancomycin, quinupristin/dalfopristin or linezolid. Mupirocin is used commonly for treating MRSA skin and soft-tissue infections and eliminating nasal MRSA colonization among patients and medical staff. Our study indicated that 10.7% of nasal MRSA isolates were resistant to mupirocin, in agreement with mupirocin resistance rates observed for MRSA isolates recovered in the context of clinical infection in another study conducted in Wenzhou, China [@pone.0027328-Liu2]. Therefore, different decolonization strategies including mupirocin may prove more successful for reducing nasal MRSA carriage among members of the Chinese medical community, and possibly Chinese hospital-associated MRSA infections.\n\nUnderstanding the molecular characteristics of MRSA isolates recovered from healthy individuals may facilitate a better understanding of mechanisms for geographic distribution of MRSA in certain populations, as well as the population-based risk for aggressive community-acquired MRSA infections. CA- MRSA isolates from different geographic areas have demonstrated significant genetic diversity [@pone.0027328-Diep1]. The vast majority of CA-MRSA isolates worldwide belong to 5 clones: pandemic clone USA300 (ST8), European clone (ST80), midwest clone (ST1), southwest pacific oceanic clone (ST30) and pacific clone (ST59) [@pone.0027328-Diep1]. Studies performed in Taiwan indicate that ST59 was the most common MLST type for MRSA isolates associated with community-acquired infections and nasal colonization, particularly pediatric infections [@pone.0027328-Lo1], [@pone.0027328-Wang2], [@pone.0027328-BoyleVavra1], [@pone.0027328-Huang1]. The majority (90%) of MRSA isolates recovered from healthy individuals over a 5-year period in Taiwan were clustered into minor variants of two clonal types, ST59 and ST338 [@pone.0027328-Lo1]. However, the nasal MRSA isolates recovered in our study showed considerable heterogeneity, with ST59 and ST338 accounting for only 14.3% and 3.6% of STs, respectively. One recently published study found that among MRSA isolates associated with pediatric infections on the Chinese mainland, ST59-MRSA-SCCmecVI-t437 was the most common clone identified (59%), followed by ST1 (8%) and ST338 (8%) [@pone.0027328-Geng1]. ST59-MRSA-SCCmecVI-t437 and ST338-MRSA-SCCmecV-t163 clones were identified from healthy volunteers in our study, indicating that these two clones may contribute to community infection and dissemination among hospital employees. HA-MRSA infections are caused by a relatively small number of epidemic MRSA clones [@pone.0027328-Gomes1], and ST239-MRSA-SCC*mec* type III and ST5-MRSA-SCC*mec* type II are two major epidemic MRSA clones identified in China [@pone.0027328-Liu1]. Interestingly, these two clones were not isolated in our study. In previous studies published by our group [@pone.0027328-Yu1], [@pone.0027328-Yao1], ST88 was the most common clone identified among PVL-positive isolates. In the present study, ST88-MRSA-SCCmecV was identified for only one PVL-negative nasal MRSA isolate. Whether CA-MRSA clones identified within the nares of healthy individuals are responsible for clinical infections caused by CA-MRSA in China remains to be determined.\n\nCA-MRSA pathogenesis has been linked to carriage of PVL encoding genes [@pone.0027328-BoyleVavra2]. Relatively high prevalences of PVL gene expression were also found among nasal *S. aureus* isolates in Chengdu, China (22%) [@pone.0027328-Lo2], and Taiwan (19.1%) [@pone.0027328-Fan1], [@pone.0027328-Lo2]. We noted a somewhat lower prevalence (11%) of PVL gene expression for nasal MRSA isolates in our study. Whether a higher prevalence of PVL expression for nasal *S. aureus* isolates portends a higher rate of clinical *S. aureus* infections, and more severe CA-MRSA infections, in Chinese populations remains to be determined. Epidemiologic studies tracking incident CA-MRSA infections in our institution, and determining whether infections caused by PVL-expressing CA-MRSA isolates arise in persons in contact with healthy volunteers in our study harboring these isolates, is ongoing.\n\nST398-MRSA is usually associated with animal infection, especially in pigs [@pone.0027328-vanDuijkeren1], and persons in close contact with animals are more likely to harbor ST398-MRSA isolates [@pone.0027328-Yu1], [@pone.0027328-Krziwanek1], [@pone.0027328-vanBelkum1], [@pone.0027328-Hallin1]. Recently published data suggest that human infections caused by MRSA-ST398 are increasing [@pone.0027328-Yu1], [@pone.0027328-Krziwanek1], [@pone.0027328-vanBelkum1], [@pone.0027328-Hallin1]. One nasal MRSA isolate belonging to ST398-MRSA-V-*spa* t571 was found in our study. It would be interesting to determine whether medical students performing cadaveric procedures using animals are more likely to harbor ST398-MRSA.\n\nIn conclusion, we found a relatively high prevalence of nasal MRSA carriage within healthy persons from a Chinese medical college campus. Among these nasal MRSA isolates, epidemic clones associated with community infection were found, supporting policies for reducing nasal carriage in order to prevent spread of MRSA to community members and hospitalized patients. Moreover, antibiotic resistance patterns for these isolates indicate that unique decolonization strategies may be needed.\n\nWe gratefully acknowledge the students from Wenzhou Medical College, Zhongyong Wang and Liqing Zhu, for participation in the collection of nasal specimens and Yuanyuan Lu for performing experiment. We are also grateful to T. Ito for the kind gift of MRSA NCTC 10442 (SCC*mec* I), MRSA N315 (SCC*mec* II), MRSA85/2082 (SCC*mec* III), and MRSA JCSC4744 (SCC*mec* IV).\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This study was supported by grants from Wenzhou Municipal Science and Technology Bureau, China (H2010064) and the 11th Five-Year Plan of the Ministry of Sciences and Technology (2010DFA32100, 2009ZX09303-005, 2008ZX10003-016). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: FY QB LZ JP LW. Performed the experiments: JD CC QC YS JT BD. Analyzed the data: ZQ CP CS. Contributed reagents/materials/analysis tools: BD. Wrote the paper: FY. Revised the manuscript critically for important intellectual content: ZQ CP CS.\n"}
+{"text": "\n"}
+{"text": "![](hosplond72630-0023){#sp1 .487}\n"}
+{"text": "Background\n==========\n\nInherited defects in DNA mismatch repair (MMR) genes, usually *hMLH1*or *hMSH2*, lead to microsatellite instability (MSI) and subsequent malignancy in hereditary nonpolyposis colon cancer (HNPCC) \\[[@B1]\\]. While typical of HNPCC cancers, MSI also occurs in a minority of sporadic cancer \\[[@B2]\\]. Recently, a distinct type of MSI has been described where microsatellite alterations are present at selected tetranucleotide repeat regions; this is referred to as \\'elevated microsatellite alterations at selected tetranucleotide repeats\\' (EMAST) \\[[@B3]-[@B5]\\]. Although the genetic basis for MSI associated with defective MMR is being increasingly clarified, the mechanism underlying microsatellite alterations for EMAST tumors remains to be unclear \\[[@B3]\\]. Recent work has suggested that EMAST may be associated with alterations in the tumor suppressor gene p53 \\[[@B4],[@B6]\\].\n\nThe incidence of MSI in endometrial carcinoma has recently been described in several articles. It has been shown to be present in 25% to 30% of type I endometrial carcinoma, which is related with estrogenic stimulation, and has been rarely demonstrated in type II endometrial carcinoma, which is unrelated with estrogen \\[[@B10]-[@B12]\\]. However, EMAST in endometrial carcinoma has not yet been studied. The aim of the current study was to evaluate the prevalence and clinicopathologic significance of MSI including EMAST in sporadic type I endometrial carcinoma. We also examined the association between *hMLH1*, *hMSH2*or p53 and EMAST using immunohistochemical staining techniques.\n\nMethods\n=======\n\nSample collection\n-----------------\n\nFormalin-fixed, paraffin-embedded tissue samples from sporadic type I endometrial carcinomas were selected from 39 patients who had undergone hysterectomy at Chonnam National University Hospital during the period from 1998 to 2002. None of these patients met the criteria for a diagnosis of HNPCC. Normal lymph node tissue from the same patient was used for control DNA samples. Hematoxylin and eosin (H&E) -stained slides from each case was reviewed previously for determination of depth of invasion and grade. Histological grade and stage were assigned according to the International-Federation of Obstetrics and Gynecology (FIGO) criteria. The depth of myometrial invasion was divided into no evidence of myometrial invasion, \u2264 50% and \\> 50%.\n\nMicrosatellite instability assay\n--------------------------------\n\n### 1) DNA extraction\n\nParaffin embedded normal and tumor tissues were cut into 10 \u00a7 sections. The tumor cells were microdissected under the microscope on the H-E stained slide. One tissue section from each case was transferred with an disposable razor blade to 1 ml of xylene in a 1.5 ml microcentrifuge tube. An equal volume of 100% ethanol was added and the samples were pelleted (10 minutes in microcentrifuge), dried under a vacuum, and digested overnight with TEN buffer (0.1 M NaCl, 10 mM Tris\u00b7Cl pH 8.0, 1 mM EDTA pH 8.0) 375 \u03bc*l*, 10% SDS 25 \u03bc*l*, and 20 \u03bcg/\u03bc*l*proteinase K 15 \u03bc*l*at 56\u00b0. The DNA was isolated by phenol-chloroform extraction and ethanol precipitation.\n\n### 2) Polymerase Chain Reaction\n\nEach matched pair of normal and tumor DNA was subject to PCR analysis with 12 microsatellite markers. The tetranucelotide repeat markers were MYCL1, D19S394, D9S242, UT5320, D8S321, and L17686. Dinucleotide repeat markers were D2S123, D17S250, and D8S554. Mononucleotide repeat markers were BAT25, BAT 26, and BAT40.\n\nPrimer sequences for PCR amplification of microsatellite loci were designed and are shown in Table [1](#T1){ref-type=\"table\"}. PCR was performed in a total volume of 20 \u03bc*l*with 10--20 ng DNA, 0.175 uM each primer, 2 mM dNTP, 15 mM MgCl~2~and 0.5 units of Taq polymerase (Supertaq, Republic of Korea). PCR conditions consisted of 28 to 35 cycles at 92\u00b0 for 1 min, 53\u00b0 \\~63\u00b0 for 1 min and 72\u00b0 for 1 min. The annealing temperature was determined empirically for each primer pair. For microsatellite analysis, one primer was end-labeled with T4 polynucleotide kinase (Takara, Japan) and ^32^P-\u03b3 \\[dATP\\]. PCR products were separated by electrophoresis in denaturing 7 M urea/6% polyacrylamide gels followed by autoradiography.\n\n###### \n\nMicrosatellite primer for MSI assay\n\n  Primer      Primer sequence                                 AT(\u00b0)   PCRcycle   Repeat type\n  ----------- ----------------------------------------------- ------- ---------- -------------\n  *BAT25*     Forward : 5\\'-TCGCCTCCA AGAATGTAA GT-3\\'        55      28         Mono\n              Reverse: 5\\'-TCTGGATTTTAACTATGGCTC-3\\'                             \n  *BAT26*     Forward : 5\\'-TGACTACTTTTGACTTCAGCC-3\\'         55      30         \n              Reverse: 5\\'-GTTTCTAACCATTCAACATTTTTATCCC-3\\'                      \n  *BAT40*     Forward : 5\\'-GTAGAGCAAGACCACCTTG-3\\'           55      30         \n              Reverse: 5\\'-GTTTCTACAACCCTGCTTTTGTTCCT-3\\'                        \n                                                                                 \n  *D2S123*    Forward : 5\\'-AAACAGGATGCCTGCCTTTA-3\\'          55      35         Di\n              Reverse: 5\\'-GGACTTTCCACCTATGGGAC-3\\'                              \n  *D17S250*   Forward : 5\\'-GGAAGAATCAAATAGACAAT-3\\'          55      35         \n              Reverse: 5\\'-GCTGGCCATATATATATTTAAACC-3\\'                          \n  *D8S554*    Forward : 5\\'-TTTCCAGACAGGGCCTA-3\\'             58      35         \n              Reverse: 5\\'-AATGCACAGGACATAATTT-3\\'                               \n                                                                                 \n  *MYCL1*     Forward : 5\\'-TGGCGAGACTCCATCAAAG-3\\'           58      35         Tetra\n              Reverse: 5\\'-CCTTTTAAGCTGCAACAATTTTC-3\\'                           \n  *D19S394*   Forward : 5\\'-AGACTACAGYGAGCTGTGG-3\\'           58      35         \n              Reverse: 5\\'-GTGTTCCTAACTACCAGGC-3\\'                               \n  *D9S242*    Forward : 5\\'-GTGAGAGTTCCTTCTGGC-3\\'            56      35         \n              Reverse: 5\\'-ACTCCAGTACAAGACTCTG-3\\'                               \n  *UT5320*    Forward : 5\\'-ACCGACAGACTCTTGCCTC-3\\'           58      35         \n              Reverse: 5\\'-TTGAGATGACCCTGAGACTG-3\\'                              \n  *D8S321*    Forward : 5\\'-GATGAAAGAATGATAGATTACAG-3\\'       58      35         \n              Reverse: 5\\'-ATTCTTCTCATGCCATATCTGC-3\\'                            \n  *L17686*    Forward : 5\\'-GCACCAATGCTCCAGAAATG-3\\'          63      35         \n              Reverse: 5\\'-TCATGGTGCCATGATAGGAG-3\\'                              \n\nAT : Annealing temperature.\n\nMono : Mononucleotide repeat.\n\nDi : Dinucleotide repeat.\n\nTetra : Tetranucleotide repeat\n\n### 3) Criteria for EMAST, MSI, and MSS\n\nEMAST was determined when the new peak, in the tumor samples, appeared in more than one of the tetranucleotide repeat markers, but was not present in the mono or dinucleotide repeat markers. A high frequency of MSI (MSI-H) was present when new peak were observed at two or more (\u2265 30--40%) in mono or dinucleotide repeat markers. A low frequency of MSI (MSI-L) was present when the new peak was at one mono or dinucleotide repeat marker. Microsatellite stable (MSS) was present when there was no new peak observed in any of the repeat markers studied.\n\nImmunohistochemical staining\n----------------------------\n\nThe immunohistochemical staining in formalin-fixed, paraffin-embedded sections was performed with the labeled avidin-biotin complex peroxidase-AEC (3-amino-9-ethylcarbazol) system. Tissue sections on glass slides were de-paraffinized with xylene, hydrated in serially diluted alcohol, and then immersed in 3% H~2~O~2~in order to quench endogenous peroxidase activity. Antigen retrieval was performed using citrate buffer (Antigen Retrieval Citra; Biogenex, San Ramon, CA, USA) with a pressure cooker. The sections were then incubated with anti-*hMLH1*(mouse monoclonal; 1:100 dilution; Zymed, San Francisco, USA), anti-*hMSH2*(mouse monoclonal; 1:100 dilution; Zymed, San Francisco, USA), and anti-*p53*(mouse monoclonal; 1:100 dilution; DAKO, Carpinteria, USA). The tissue sections were incubated with horseradish peroxidase (HRP) -conjugated streptavidin and chromogen were developed, and counter-stained with hematoxylin. The staining of tumor nuclei for *hMLH1*and *hMSH2*was evaluated and noted as absent or present. For *hMLH1*and *hMSH2*, loss of nuclear protein expression in tumor cells was considered as a defect in the *hMLH1*or *hMSH2*gene. The immunostaining for *p53*was evaluated within tumor nuclei and divided to absent or present.\n\nStatistical analysis\n--------------------\n\nWe performed statistical analysis using the Fisher\\'s exact test between clinicopathologic status or extent of positive immunoreactivity and the MSS, MSI and EMAST groups. All *p*values represent two-sided statistical tests with statistical significance at *p*\\< 0.05.\n\nResults and discussion\n======================\n\nResults\n-------\n\n### CLINICOPATHOLOGIC DATA\n\nThe patient age ranged from 30 to 68 (mean 53.4) years. Grade 1 was seen in 20 tumors; grade 2, in 11 tumors; and grade 3, in eight tumors. Most of the patients were FIGO stage I or II (33 of 39; 84.6%), and only six tumors were identified as stage III. In 15 (38.5%) tumors, the tumor was confined to endometrium. A more than 50% myometrial invasion was seen in six tumors (15.4%), and less than 50% in 18 tumors (46.1%). (Table [2](#T2){ref-type=\"table\"})\n\n###### \n\nClinicopathologic features, MSI status, and the expression of *hMSH2*, *hMLH1*, and p53 protein in type I endometrial carcinomas\n\n  No.                          Immunohistochemistry   Tetranucleotide repeat   Mononucleotide repeat   Dinucleotide repeat   Subtype of                                                               \n  ----- ---- ----- --- ------- ---------------------- ------------------------ ----------------------- --------------------- ------------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -------\n  1     50   III   2   \\>50%   \\+                     \\+                       \\+                      H                     MSI          H     H     H     H     H     H     H     H     H     H     EMAST\n  2     67   I     1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     H     H     H     MSI   MSI   H     H     MSI   H     MSI-H\n  3     56   III   1   \\<50%   \\+                     \\+                       \\+                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  4     43   I     3   Em      \\+                     \\+                       \\+                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  5     39   II    1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     MSI   H     H     H     H     H     H     H     H     EMAST\n  6     50   I     1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  7     54   I     2   \\<50%   \\-                     \\+                       \\-                      H                     H            H     H     H     H     MSI   H     MSI   H     MSI   MSI   MSI-H\n  8     40   I     2   \\<50%   \\-                     \\-                       \\-                      H                     H            H     H     H     H     MSI   H     MSI   H     H     H     MSI-H\n  9     51   I     1   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  10    55   I     1   \\<50%   \\-                     \\+                       \\-                      MSI                   H            H     H     MSI   H     H     H     H     H     H     H     EMAST\n  11    60   I     1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  12    68   III   2   \\>50%   \\-                     \\-                       \\-                      H                     H            H     H     H     H     H     MSI   MSI   H     MSI   H     MSI-H\n  13    55   I     1   Em      \\+                     \\+                       \\+                      H                     MSI          H     H     H     H     H     H     H     H     H     H     EMAST\n  14    66   I     2   \\<50%   \\-                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  15    60   I     2   Em      \\-                     \\+                       \\-                      H                     H            H     MSI   H     H     H     H     H     H     H     H     EMAST\n  16    42   I     1   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  17    53   I     2   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  18    63   I     3   \\>50%   \\-                     \\-                       \\+                      H                     H            H     H     H     H     MSI   H     MSI   H     MSI   MSI   MSI-H\n  19    58   I     1   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  20    36   I     1   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  21    40   III   1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     H     MSI   H     H     H     H     H     H     H     EMAST\n  22    68   I     1   \\<50%   \\+                     \\-                       \\-                      H                     H            H     H     H     H     MSI   MSI   MSI   H     H     H     MSI-H\n  23    41   I     2   Em      \\+                     \\+                       \\+                      H                     H            H     H     MSI   H     H     H     H     H     H     H     EMAST\n  24    50   I     2   \\<50%   \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  25    54   I     1   Em      \\+                     \\+                       \\-                      H                     MSI          H     H     MSI   H     H     H     H     H     H     H     EMAST\n  26    51   I     3   \\<50%   \\+                     \\-                       \\-                      MSI                   MSI          MSI   H     H     H     H     H     H     H     H     H     EMAST\n  27    63   I     3   Em      \\+                     \\+                       \\+                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  28    46   I     3   Em      \\+                     \\+                       \\+                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  29    64   III   3   \\>50%   \\+                     \\+                       \\+                      H                     H            H     H     MSI   H     H     H     H     H     H     H     EMAST\n  30    44   II    3   Em      \\+                     \\+                       \\+                      H                     H            H     MSI   H     H     H     H     H     H     H     H     EMAST\n  31    55   III   3   \\>50%   \\-                     \\+                       \\+                      H                     H            H     H     MSI   H     H     H     H     H     H     H     EMAST\n  32    44   I     1   \\<50%   \\-                     \\-                       \\-                      H                     H            H     H     H     H     MSI   MSI   MSI   MSI   H     MSI   MSI-H\n  33    43   I     1   \\<50%   \\+                     \\+                       \\-                      H                     H            H     MSI   H     H     H     H     H     H     H     H     EMAST\n  34    55   I     1   \\<50%   \\+                     \\+                       \\+                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  35    51   I     1   \\>50%   \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  36    33   I     2   Em      \\+                     \\+                       \\-                      H                     H            H     H     H     H     H     H     H     H     H     H     MSS\n  37    49   I     1   Em      \\+                     \\-                       \\-                      MSI                   H            H     MSI   H     H     H     H     H     H     H     H     EMAST\n  38    52   I     1   \\<50%   \\+                     \\+                       \\+                      H                     H            H     H     H     MSI   H     H     H     H     H     H     EMAST\n  39    30   I     2   \\<50%   \\+                     \\-                       \\-                      H                     H            H     H     H     H     H     MSI   MSI   MSI   MSI   MSI   MSI-H\n\nEm : Endometrium H: Heterozygosity.\n\nMSS : Microsatellite stable.\n\nMSI : Microsatellite instability.\n\nMSI-H : Microsatellite instability -- high frequency.\n\nEMAST : Elevated microsatellite alterations at selected tetranucleotide repeat.\n\n### Molecular data\n\nSuccessful PCR amplification of six mono- or dinucleotide repeat markers and six tetranucleotide markers was obtained for 39 tumors and each normal lymph node samples. The data is shown in Table [2](#T2){ref-type=\"table\"}. Microsatellite instability at mono- or dinucleotide repeat markers was noted in eight of 39 tumors (20.5%). All eight tumors displayed MSI at two or more loci for mono and dinucleotide repeat markers. Therefore, these eight tumors were categorized as the MSI-H group. There were no tumors classified as MSI-L. The MSI rates at mono or dinucleotide repeat markers were as follows: 15.4% (6/39) in BAT25, 12.8% (5/39) in BAT26, 17.9% (7/39) in BAT40, 5.1% (2/39) in D17S250, 12.8% (5/39) in D8S554, and 10.3% (4/39) in D2S123. (Figure [1](#F1){ref-type=\"fig\"}) Microsatellite instability at tetranucleotide repeat markers was observed in 15 of 39 tumors (38.5%). All 15 tumors did not exhibit any MSI at the mono- or dinucleotide repeat markers, and were categorized as EMAST. Among 15 EMAST tumors, 11 tumors displayed MSI at one locus for tetranucleotide repeat markers, three displayed at two loci, and one displayed at three loci. The highest MSI rates, for tetranucleotide repeat markers, were seen in *L17686*(7/39, 17.9%), followed by *D9S242*(4/39, 10.3%), *UT5320*(4/39, 10.3%), *MYCL1*(3/39, 7.7%), *D19S394*(1/39, 2.6%), and *D8S321*(1/39, 2.6%). (Figure [2](#F2){ref-type=\"fig\"}) MSI was not observed in 16 of 39 tumors (41.0%) for all markers studied. Therefore, these 16 tumors were categorized as MSS. There was no significant differences between MSI-H, EMAST and MSS group in respect to stage (*p*= 0.28), grade (*p*= 0.66) and depth of invasion (*p*= 0.06). (Table [3](#T3){ref-type=\"table\"})\n\n![**MSI at mono- and dinucleotide repeat markers in type I endometrial carcinioma**. Alterations in the electrophoretic mobility of PCR products from tumor compared with normal tissue DNA are seen (arrowheads). (N: normal, T: tumor).](1756-9966-27-88-1){#F1}\n\n![**MSI at tetranucleotide repeat markers (EMAST) in type I endometrial carcinioma**. Alterations in the electrophoretic mobility of PCR products from tumor compared with normal tissue DNA are seen (arrowheads). (N: normal, T: tumor).](1756-9966-27-88-2){#F2}\n\n###### \n\nMSI status in relation to clinicopathologic features.\n\n                                No.   MSS         MSI-H      EMAST       *p*-value\n  ------- --------------------- ----- ----------- ---------- ----------- -----------\n  Stage   I or II               33    15(93.8%)   7(78.5%)   11(73.3%)   0.28\n          III                   6     1(6.2%)     1(12.5%)   4(26.7%)    \n                                                                         \n  Grade   1                     20    9(56.3%)    3(37.5%)   8(53.3%)    0.66\n          2                     11    4(25.0%)    4(50.0%)   3(20.0%)    \n          3                     8     3(18.7%)    1(12.5%)   4(26.7%)    \n                                                                         \n  Depth   Endometrium           15    9(56.3%)    0(0.0%)    6(40.0%)    0.06\n          \\<50% of myometrium   18    6(37.5%)    6(75.0%)   6(40.0%)    \n          \\>50% of myometrium   6     1(6.2%)     2(25.0%)   3(20.0%)    \n                                                                         \n  Total                         39    16(41.0%)   8(20.5%)   15(38.5%)   \n\nMSS : Microsatellite stable.\n\nMSI : Microsatellite instability -- high frequency.\n\nEMAST : Elevated microsatellite alterations at selected tetranucleotide repeat.\n\n### Immunohistochemical staining data\n\n#### 1) hMSH2 and hMLH1\n\nLoss of nuclear *hMSH2*protein expression was seen in 23.0% (9 of 39), and loss of *hMLH1*in 20.5% of tumors (8 of 39). In 13 of total 39 tumors (33.3%), the protein expression loss was found at least one of either *hMSH2*or *hMLH1*. (Figure [3](#F3){ref-type=\"fig\"}) No correlation between the loss of *hMSH2*or *hMLH1*and clinicopathologic data was observed. Loss of *hMSH2*or *hMLH1*was mainly seen in MSI-H group (7 of 8), and was rarely seen in MSS group (1 of16). The MSI-H group was significantly correlated with loss of at least one of both proteins. (p \\< 0.01) In 33.3% (5 of 15) of EMAST group, the loss of *hMSH2*or *hMLH1*was observed. (Table [4](#T4){ref-type=\"table\"}) Of 11 tumors that displayed MSI at one locus for tetranucleotide markers, nine tumors showed positive immunoreactivity for both *hMSH2*and *hMLH1*. Of four tumors that displayed MSI at two or three loci for tetranucleotide repeat markers, three tumors showed the expression loss of at least one of either *hMSH2*or *hMLH1*.\n\n![**Examples of type I endometrial carcinoma with (A) loss of *hMSH2*protein expression, (B) preservation of *hMSH2*protein expression, (C) loss of *hMLH1*protein expression, and (D) preservation of *hMLH1*protein expression**. Positive expression for *hMSH2*or *hMSH1*is noted in non-tumor cells. (A-D, X200).](1756-9966-27-88-3){#F3}\n\n###### \n\nMSI status in relation to the expression of *hMSH2*, *hMLH1*, and p53 protein\n\n                       IHC                      No.   MSS          MSI-H      EMAST       *p*-value\n  -------------------- ------------------------ ----- ------------ ---------- ----------- -----------\n  *hMSH2*              Positive                 30    15(92.8%)    3(37.5%)   12(80.0%)   0.01\n                       Negative                 9     1(7.2%)      5(62.5%)   3(20.0%)    \n                                                                                          \n  *hMLH1*              Positive                 31    16(100.0%)   2(33.3%)   13(86.7%)   \\<0.01\n                       Negative                 8     0(0.0%)      6(66.7%)   2(13.3%)    \n                                                                                          \n  *hMSH2*and *hMLH1*   Positive, both           26    15(93.7%)    1(12.5%)   10(67.7%)   \\<0.01\n                       Negative, at least one   13    1(6.3%)      7(87.5%)   5(33.3%)    \n                                                                                          \n  p53                  \\+                       13    5(31.3%)     1(12.5%)   7(46.7%)    0.25\n                       \\-                       26    11(68.7%)    7(87.5%)   8(53.3%)    \n                                                                                          \n  Total                                         39    16           8          15          \n\nIHC : Immunohistochemistry.\n\nMSS : Microsatellite stable.\n\nMSI : Microsatellite instability -- high frequency.\n\nEMAST : Elevated microsatellite alterations at selected tetranucleotide repeat\n\n#### 2) p53\n\nThe p53 immunohistochemical expression was noted in 31.3% (5 of 16) of MSS, 12.5% (1 of 8) of MSI-H, and 46.7% (7/15) of EMAST. Although p53 expression was more common in the EMAST group than the other groups, the difference was not statistically significant. (p = 0.25) (Table IV). The p53 expression was noted in seven of 11 tumors that displayed MSI at one locus for tetranucleotide markers (63.6%). None of four tumors displayed MSI at two or three loci for tetranucleotide markers showed the p53 expression. For the clinicopathologic data, p53 expression was not correlated with stage and depth of invasion. But, observation of p53 expression was noted in 87.5% (7/8) of grade 3, 18.2% (2/11) of grade 2, and 20.0% (4/20) of grade 1 tumors. Thus, p53 expression was highly associated with FIGO grade 3 tumors. (p \\< 0.01)\n\nDiscussion\n==========\n\nIn our study, the incidence of EMAST in type I endomterial carcinoma was 38.5%. To date, those tumors showing the highest levels of EMAST are associated with known environmental carcinogens, for example, cigarette smoking or sunlight exposure, suggesting a causal relationship \\[[@B13]\\]. For example, EMAST is common in head and neck (56%), lung (51%) and bladder cancer (21--40%), but infrequently found in renal (12%) and prostate cancer (6%) \\[[@B5],[@B14]\\]. Danaee *et al*. found 75.4% and 43.3% of MSI at tetranucleotide repeat markers in non-melanomatous skin cancer and bladder tumors, respectively \\[[@B14]\\]. Wooster *et al*. reported that human cancers including breast (11%), brain (2%), ovarian (10%) cancers and soft tissue sarcoma (11%) exhibited EMAST \\[[@B15]\\]. The EMAST rate in type I endometrial cancer was relatively higher than that of kidney, prostate, brain and ovary. The estrogen which is widely known as a causal factor in type I endometrial carcinoma can be possible as a novel factor associated with EMAST. However, the tetranucleotide repeat makers are different between our study and others. In order to clarify whether the estrogen is associated with EMAST or not, a standard gene panel for EMAST study, like Bethesda panel used in mono- & dinucleotide repeat markers \\[[@B3]\\], should be requested.\n\nSeven of eight MSI-H tumors (87.5%) of our study showed a loss of at least one of both *hMLH1*and *hMSH2*. However, for the EMAST group, it was observed in five of 15 tumors (33.3%). Malfunction of the MMR genes including *hMLH1*and *hMSH2*can result from a somatic mutation, deletion, and hypermethylation of the coding region. In sporadic endometrial carcinoma cell lines, somatic mutations are infrequent, and hypermethylation of the *hMLH1*promoter region appears to be the main cause of gene silencing \\[[@B21],[@B22]\\]. Thus, our results said that a defect of *hMLH1*or *hMSH2*was less frequently detected in EMAST group than MSI-H group.\n\nThe majority of sporadic endometrial carcinomas (at least approximately 70%--80%), are designated as type I endometrial carcinomas and are estrogen related. Histologically, most type I endometrial carcinomas show an endometrioid differentiation and are of low grade. About 10--20% of endometrial carcinomas, designated as type II endometrial carcinomas, are estrogen-unrelated; these tumors usually present in older patients and are typically high-grade carcinomas, with non-endometrioid differentiation \\[[@B8]\\]. Molecular data from multiple studies support the hypothesis of different genetic pathways for the development of type I and type II endometrial carcinoma. The most frequent alteration in type 1 endometrial carcinoma is the *PTEN*inactivation by mutation, followed by MSI and mutations of *K-ras*and *\u03b2-catenin*. In type II endometrial carcinoma, alterations of p53 and loss of heterozygosity (LOH), at several chromosomal loci, are thought to drive the process of neoplastic transformation \\[[@B8],[@B9]\\]. In this study, the incidence of MSI, in type I endometrial carcinoma, is similar to that previously reported \\[[@B10],[@B12]\\]. Our study for type I endometrial carcinomas showed no significant differences between three groups with regard to stage, FIGO grade, and the extent of myometrial invasion. Peiro et al. reported that tumors in an advanced FIGO stage (III to IV) were more frequently MSI positive than those in a low stage (I to II) \\[[@B20]\\]. This difference can be interpretated as two problems : 1) Peiro\\'s study included type II endometrial carcinomas as well as type I endometrial carcinomas. 2) The samples included in our study are not evenly distributed in aspect of FIGO stage. Most of the samples were FIGO stage I or II (33 of 39; 84.6%). The unbalanced stage of our study\\' cases seems to be a factor affecting the result.\n\nAlteration of the p53 gene is a common event in many cancers, and is primarily associated with protein overexpression. Given its role in maintaining genomic stability, it is possible that altered p53 might play a mechanistic role in the genesis of EMAST \\[[@B14]\\]. In fact, a recent report by Ahrendt et al. found that most tumors with EMAST also had p53 gene mutations \\[[@B4]\\]. In this study, although the p53 gene was expressed more in EMAST (+) tumors than EMAST (-) tumors, this difference was not statistically significant. The p53 mutation was known as a genetic alteration in type II endometrial carcinomas \\[[@B24]\\]. However, p53 mutation was also found in a subset of approximately 10--20% of type I endometrial carcinomas, which were mostly grade 3 tumors \\[[@B25]\\]. Catasus et al. found that most grade 3 tumors in type I endometrial carcinomas showed the immunohistochemical expression for p53 \\[[@B10]\\], this finding is presented in our result. The p53 protein was highly expressed in grade 3 tumors (p \\< 0.001).\n\nAmong tetranucleotide repeat markers that were examined in our study, MSI was highly detected in *L17686*gene (7/39). The gene is located in long arm of chromosome 7, and was also highly detected marker in lung cancer and head & neck cancer \\[[@B17]\\]. *UT5320*and *MYCL1*that showed the higher detection rate in bladder cancer were observed in 10.3% (4/39) and 7.7% (3/39) in type I endometrial carcinomas of our study, respectively \\[[@B18],[@B19]\\]. The detection rate at each tetranucleotide repeat maker seems to be different according to organ or tumor types. In order to solve this question, EMAST study within more various organ and tumor types should be preceded.\n\nTo date, seven human MMR genes whose products function in MMR have been identified. These include homologs of the *Escherichia coli mutS*-- *hMSH2, hMSH3, hMSH6*-, and *mutL*homologs -- *hMLH1, hPMS2, hMLH3*and *hPMS1*\\[[@B26],[@B27]\\]. DNA strand slippage at tri- or tetranucleotide repeats during replication will result in formation of loops containing more than three extra nucleotides. A complex between the heterodimer MutL\u03b1 (*hMLH1/hPMS2*) and the heterodimer MutS\u03b2 (*hMSH2/hMSH3*) is responsible for repairing these loops \\[[@B28]\\]. Recently, yeast MLH3 was found to form a complex with y *MLH1*and to participate in repairing some groups of loops in collaboration with MutS\u03b2. This suggests that *hMLH3*, a homolog of yeast *MLH3*, might have a similar function. Therefore, the loss of the *hMSH3*and/or *hMLH3*function may result in loss of loop repair, and eventually may be responsible for EMAST. This should inspire further investigations to prove a correlation between *hMSH3*and/or *hMLH3*and EMAST\n\nIn this study, 11 tumors of 15 EMAST (73.3%) showed MSI at one locus for tetranucleotide repeats, and four tumors (26.7%) showed MSI at two or more loci. Of 11 tumors, a defect of *hMLH1*or *hMSH2*expression was found in 18.2% (2 of 11) and p53 overexpression was observed in 63.6% (7 of 11). In 4 tumors at two or more loci, 3 tumors of 4 showed the expression loss of *hMLH1*or *hMSH2*and none showed the p53 expression. It is possible to be a different genetic alteration between two groups. However, after the molecular events associated with EMAST become clear, the research about the genetic difference appears to be a subject.\n\nIn conclusion, the incidence of EMAST in endometrioid carcinoma was 38.5%, and an alteration of *hMSH2*or *hMLH1*was not frequent in EMAST.\n\nAbbreviations\n=============\n\nMMR: Mismatch repair; MSI: Microsatellite instability; HNPCC: Hereditary non-polyposis colon cancer; PCR: Polymerase chain reaction; EMAST: Elevated microsatellite alterations at selected tetranucleotide repeats; H&E: Hematoxylin and eosin; FIGO: International-Federation of Obstetrics and Gynecology; MSI-H: a high frequency of MSI; MSI-L: A low frequency of MSI; MSS: Microsatellite stable; HRP: Horseradish peroxidase.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nYDC, JSL, and JHN carried out the microsatellite instability assay and participated in the design of the study. JC and HSC collected the samples and drafted the manuscript. CSP and JHK carried out the immmunohistochemical staining. CC performed the statistical analysis. JHL, MCL and SWJ conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.\n"}
+{"text": "INTRODUCTION\n============\n\nBaseline whole-body ^18^F-fluoride PET/CT is ideal for staging and restaging prostate cancer and has been shown to be an independent prognostic imaging biomarker of patients undergoing radium-223 dichloride (^223^RaCl~2~) therapy^(^[@r1]^)^. However, although treatment with ^223^RaCl~2~ improves survival in prostate cancer patients^(^[@r2]^-^[@r4]^)^, not all patients respond to this therapy. It would be beneficial to identify nonresponders early in the course of treatment, thereby reducing morbidity and unnecessary costs.\n\nAfter successful treatment of osteoblastic bone metastases, an osteoblastic reaction (flare) can occur, which increases bone uptake even in responsive cases. That can be confused with the osteoblastic reaction and inflammation that occur in response to tumor-associated growth factors during progression. This phenomenon has been well described in conventional bone scintigraphy, and that method is therefore not recommended for use as the sole means of determining the response to treatment^(^[@r5]^,^[@r6]^)^.\n\nAlthough interim studies performed with ^18^F-FDG PET/CT can change the management of patients with a variety of cancer types^(^[@r7]^-^[@r10]^)^, the exact role of ^18^F-fluoride PET/CT in evaluating the early response to therapy (interim study) is not well established. The importance of ^18^F-fluoride PET/CT has extended beyond the diagnosis of metastases to the evaluation of optimal strategies for use in patients submitted to treatment with new therapeutic agents. Chemotherapy, hormone therapy, immunotherapy, and radionuclide therapies such as those involving ^223^RaCl~2~^(^[@r11]^)^ are costly approaches. Therefore, the ability to predict response, thereby avoiding overtreatment and reducing costs, will improve patient management. The purpose of this study was to determine whether an interim ^18^F-fluoride PET/CT study is able to evaluate treatment responses in prostate cancer patients submitted to ^223^RaCl~2~ therapy.\n\nMATERIALS AND METHODS\n=====================\n\nThe local institutional review board approved this retrospective analysis (reference no. PA14-0848). We retrospectively reviewed histologically confirmed cases of hormone-refractory prostate cancer with bone metastasis in patients receiving ^223^RaCl~2~ therapy and undergoing two ^18^F-fluoride PET/CT studies-a baseline study and an interim study (immediately prior to the fourth cycle of ^223^RaCl~2~)-at our facility. All patients completed at least four cycles of ^223^RaCl~2~ (Xofigo; Bayer Pharma AG, Berlin, Germany), receiving intravenous infusions of 50 kBq/kg (1.4 \u00b5Ci/kg) of ^223^RaCl~2~ at monthly intervals.\n\n^18^F-fluoride PET/CT acquisition\n---------------------------------\n\n^18^F-fluoride PET/CT images were acquired immediately prior to initiation of the ^223^RaCl~2~ therapy (baseline study) and immediately before the fourth cycle (interim study). True whole-body PET images were obtained 50-60 min after intravenous injection of 158-370 MBq of ^18^F-sodium fluoride in dedicated PET/CT scanners (Discovery STe, RX, or VCT; 16 or 64 channel; GE Healthcare, Milwaukee, WI, USA, or Siemens mCT Flow; 64 channel; Siemens Healthcare, Knoxville, TN, USA), and whole-body noncontrast CT scans were used for attenuation correction.\n\n^18^F-fluoride PET/CT interpretation and quantification\n-------------------------------------------------------\n\nTwo board-certified nuclear medicine physicians evaluated baseline and interim ^18^F-fluoride PET/CT images. Visual and quantitative analyses were performed.\n\n### Visual analysis\n\nIn the visual analysis, we compared the baseline and interim studies, classifying the responses as follows:\n\nComplete response - Osteoblastic bone metastases identified in the baseline study no longer being present in the interim study.Partial response - Interim study showing decreased uptake in pre-existing bone metastases.Stable disease - No difference between the interim and baseline scans in terms of the uptake in pre-existing bone metastases.Progressive disease - Interim study showing an increase in the uptake or volume of a pre-existing bone metastases or new osteoblastic metastases.\n\nThe patients were followed for confirmation of these response classifications. The follow-up reference standards used in order to determine if the response classification was correct (i.e., to identify true-positive, true-negative, false-positive, and false-negative responses) included clinical parameters-such as clinical worsening, disease progression, bone events, and death; biochemical parameters-such as the levels of alkaline phosphatase (ALP) and prostate-specific antigen (PSA); and imaging findings-such as those obtained with ^18^F-fluoride PET/CT, ^18^F-FDG PET/CT, bone scans, or CT scans. On the interim ^18^F-fluoride PET/CT study, images that demonstrated stable disease, a partial response, or a complete response were all considered to represent a true-positive response to therapy if the reference standards also indicated that the patient had responded to therapy (no clinical worsening, progression, or increase in the levels of the biochemical markers) or a false-positive response to therapy if those same standards demonstrated progressive disease (new areas of disease, clinical worsening, or death). In contrast, images that demonstrated progressive disease were considered to represent a true-negative response to therapy if the reference standards also indicated that the patient had not responded to therapy and it was confirmed during follow-up that there was no response to therapy or a false-negative response to therapy (flare response) if those same standards demonstrated a response (no clinical worsening, progression, or increase in the levels of the biochemical markers).\n\n### Quantitative analysis\n\nUsing quantitative analysis, we determined the whole-body skeletal tumor burden in the baseline and interim ^18^F-fluoride PET/CT images. The skeletal tumor burden was determined after establishing the maximum standardized uptake value (SUV~max~) threshold \u2265 10 to exclude normal bone^(^[@r12]^)^. To that end, we initially obtained the volume (in milliliters) of total fluoride activity, defined as the fluoride tumor volume above an SUV~max~ of 10 (FTV~10~), within the volume of interest (VOI). The FTV~10~ calculation is equivalent to the metabolic tumor volume calculation used in ^18^F-FDG PET/CT studies. The total fluoride lesion uptake above an SUV~max~ of 10 (TLF~10~) was then calculated as the product of mean SUV~max~ \u00d7 FTV~10~. The TLF~10~ is equivalent to the total lesion glycolysis used in ^18^F-FDG PET/CT studies. To evaluate the performance of interim ^18^F-fluoride PET/CT, the percent change in the skeletal tumor burden between the baseline and interim studies was calculated as follows:\n\n%TLF~10~ = interim TLF~10~ - baseline TLF~10~ / baseline TFL~10~\n\nStatistical analyses\n--------------------\n\nCategorical variables were expressed as absolute and relative frequencies, whereas continuous variables were expressed as mean \u00b1 standard deviation when presenting normal distribution and as median (minimum-maximum) when presenting non-normal distribution. All outcome measures were correlated with the %TFL~10~ values obtained. The primary end point was overall survival (OS), which was calculated from the first ^223^RaCl~2~ cycle to the date of death or last follow-up. Secondary end points were progression-free survival (PFS), time to a bone event (TTBE), and bone marrow failure (BMF). PFS was calculated from the first ^223^RaCl~2~ cycle to the date of progression, death, or last follow-up. The TTBE was calculated as the time from the date of the first ^223^RaCl~2~ cycle to the next bone event. Lastly, BMF was defined as the development of hematologic toxicity (World Health Organization grade 3 or 4), together with no recovery after six weeks or death due to BMF after the last ^223^RaCl~2~ cycle.\n\nKaplan-Meier survival curves were generated, and Cox proportional hazards regression was used in order to analyze predictors of survival. Backward stepwise selection was performed for multivariate Cox models. Logistic regression was used in order to model the odds of a bone event as a function of all of the PET variables. We used Spearman\\'s correlation coefficient to assess the level of agreement between the PET variables. For the statistical analyses, we used the Statistical Analysis System, version 9.3 for Windows (SAS Institute Inc., Cary, NC, USA).\n\nRESULTS\n=======\n\nWe analyzed the cases of 34 patients, with a mean age of 72.4 \u00b1 10.2 years (median, 72.5 years; range, 43.3-88.8 years) ([Table 1](#t1){ref-type=\"table\"}), who had had prostate cancer for a mean of 6 \u00b1 4 years (range, 2-20 years). The mean Gleason score was 7 \u00b1 3. Prior to the initiation of ^223^RaCl~2~ therapy, 26.9% of the patients had received chemotherapy, 5% had received radiotherapy, 59% had received hormone therapy, and 9% had received blood transfusion. At the first ^223^RaCl~2~ cycle, the mean ALP was 193.9 IU/L and the mean PSA was 103.2 ng/mL. The median time of follow-up after the interim study was 28.1 months (range, 11-52 months). The 34 patients were submitted to a collective total of 179 ^223^RaCl~2~ cycles: 55.9% of the patients received six cycles of ^223^RaCl~2~; 14.7% received five cycles; and 29.4% received four cycles. The principal causes of treatment interruption were progression (in 44.4%), hematologic toxicity (in 17.8%), a significant decline of the Eastern Cooperative Oncology Group performance status (in 13.3%), and a bone event (in 2.2%).\n\n###### \n\nDemographic and clinical characteristics of 34 patients prior to 2^23^RaCl~2~ therapy.\n\n  Characteristic                      Median   Range\n  ----------------------------------- -------- ------------\n  Age                                 72.4     43.3-88.8\n  Prostate specific antigen (ng/mL)   103.2    2.1-761.1\n  Alkaline phosphatase (IU/L)         193.9    48.0-913.0\n  Hemoglobin (g/dL)                   11.2     6.6-13.6\n  Platelets (K/\u00b5L)                    215.8    114-413\n  Absolute neutrophils (K/\u00b5L)         5.7      1.5-21.4\n\nVisual analysis of interim ^18^F-fluoride PET/CT\n------------------------------------------------\n\nA complete response was not perceived in any of the interim ^18^F-fluoride PET/CT studies or on the basis of the follow-up reference standards. A partial response was identified in 16 (47%) of the patients in the interim ^18^F-fluoride PET/CT studies ([Figure 1](#f1){ref-type=\"fig\"}), and the reference standards demonstrated that a partial response had indeed been achieved in eight of those patients (true-positive cases), whereas the other eight patients had progressed (false-positive cases), as shown in [Figure 2](#f2){ref-type=\"fig\"}. Stable disease was noted in five (15%) of the patients in the interim ^18^F-fluoride PET/CT studies, although only three of those patients were categorized as true-positive cases (showing stable disease or a partial response), whereas the two remaining patients progressed. Progressive disease was identified in 13 (38%) of the patients in the interim ^18^F-fluoride PET/CT studies, 12 (35.3%) of whom were categorized as true-negative cases ([Figure 3](#f3){ref-type=\"fig\"}), the remaining patient (3.0%) being categorized as a false-negative case because the increased uptake noted on the interim ^18^F-fluoride PET/CT (when compared with that observed in the baseline study) was actually due to a flare phenomenon ([Figure 4](#f4){ref-type=\"fig\"}). Therefore, the responses were categorized as true positive in 11 cases (32.4%), false positive in 10 (29.4%), true negative in 12 (35.3%), and false negative in 1 (2.9%).\n\nFigure 1A patient with hormone-refractory prostate cancer, accompanied by bone metastasis, who showed a partial response to ^223^RaCl~2~, and the interim ^18^F-fluoride PET/CT study demonstrating a true-positive response. **A:** The baseline ^18^F-fluoride PET/CT study revealing widespread osteoblastic metastases. **B:** The interim ^18^F-fluoride PET/CT study, performed after the third ^223^RaCl~2~ cycle, showing a reduction in osteoblastic metastases, especially in the rib cage, pelvis, and right femur, consistent with a partial response to ^223^RaCl~2~. There was a 70% reduction in the %TLF~10~. During follow-up, the ALP levels dropped and no new bone lesions appeared. After the last ^223^RaCl~2~ cycle, the patient resumed enzalutamide to control lymph node metastases that had been present prior to the first ^223^RaCl~2~ cycle.\n\nFigure 2A patient with hormonerefractory prostate cancer, accompanied by bone metastasis, who showed progression during ^223^RaCl~2~ therapy but was categorized as a false-positive case on the basis of the imaging findings. **A:** The baseline ^18^F-fluoride PET/ CT study showing osteoblastic metastases. **B:** The interim ^18^Ffluoride PET/CT study, performed after the third ^223^RaCl~2~ cycle, showing a slight reduction in uptake by the known osteoblastic metastases and no new lesions, consistent with a partial response. Although the %TLF~10~ decreased by 44%, the PSA and ALP levels continued to rise and there was rapid progression of the bone metastases. Therefore, the patient was started on cyclophosphamide and subsequently on dasatinib. C: A follow-up ^18^F-fluoride PET/ CT study, conducted after the sixth ^223^RaCl~2~ cycle, showed wide- A B C spread osteoblastic metastases.\n\nFigure 3A patient with hormone-refractory prostate cancer, accompanied by bone metastasis, who showed progression during ^223^RaCl~2~ therapy. **A:** The baseline ^18^F-fluoride PET/CT study showing widespread osteoblastic metastases. **B:** The interim ^18^F-fluoride PET/CT study, performed after the third ^223^RaCl~2~ c ycle, s howing i ncreased u ptake i n t he k nown o steoblastic metastases and new lesions, especially in the pelvis, consistent with progression. The %TLF~10~ increased by 104%; PSA and ALP levels continued to rise; new bone metastases developed; a liver metastasis developed; and there was further enlargement of previously enlarged lymph nodes. The patient started a new chemotherapy regimen but died eight months after the last ^223^RaCl~2~ cycle.\n\nFigure 4A patient with hormone-refractory prostate cancer, accompanied by bone metastasis, who responded to ^223^RaCl~2~ but was categorized as a falsenegative case on the basis of the imaging findings. **A:** The baseline ^18^F-fluoride PET/CT study showing osteoblastic metastases. **B:** The interim ^18^F-fluoride PET/CT study, performed after the third ^223^RaCl~2~ cycle, showing increased uptake in the known osteoblastic metastases but no new lesions. Although that pattern is consistent with progression (with a %TLF~10~ increase of 65%), the PSA and ALP dropped remarkably, after which the patient responded and was stable at 12 months after the last ^223^RaCl~2~ cycle. Therefore, the images were clearly due to a flare (false-negative) response.\n\nThe interim ^18^F-fluoride PET/CT study was found to have a sensitivity of 91.6%, a specificity of 54.5%, a positive predictive value of 52.4%, a negative predictive value of 92.3%, and an accuracy of 67.6% ([Figure 5](#f5){ref-type=\"fig\"}). For distinguishing between responders and nonresponders, a reduction in the ALP level had a sensitivity of 38% and a specificity of 85% when the follow-up parameters were taken as the reference.\n\nFigure 5Visual analysis and evolution of the 34 patients.\n\nQuantitative analysis of interim ^18^F-fluoride PET/CT\n------------------------------------------------------\n\n[Figure 6](#f6){ref-type=\"fig\"} illustrates the quantitative method employed to obtain the TLF~10~ and FTV~10~ values. Spearman\\'s correlation coefficient showed that the %TLF~10~ and %FTV~10~ values correlated strongly with each other (rho = 0.95). Therefore, only the %TLF~10~ values were applied to subsequent analyses. The median TLF~10~ was 7374 (range, 391-46,550) in the baseline ^18^F-fluoride PET/CT study and 5632 (range, 486-30,200) in the interim study.\n\nFigure 6Example of determination of TLF~10~ and FTV~10~. **A:** A semi-automatic VOI (orange rectangle) is placed within the whole-body maximum intensity projection image. A threshold SUV~max~ of 10 is then established as the cut-off to separate normal bone from abnormal bone. Consequently, the software will automatically delineate only SUV~max~ regions above the set threshold of 10, defining the VOI with an isocontour threshold set at 41% of the SUV~max~. After all regions have been defined, a careful inspection should be performed to exclude all non-tumor-related VOIs. The sum of all the VOIs outlined with the SUV~max~ of 10 provides the FTV~10~. To obtain the TLF~10~, the FTV~10~ is multiplied by the SUVmean~10~ (VOI~10~ \u00d7 mean~10~), which is also measured in milliliters. **B:** In this particular example, the patient had only one lesion with an SUV~max~ higher than 10, which corresponded to a rib metastasis with an SUV~max~ of 25. The TLF~10~ was 65.8, and the FTV~10~ was 4.2.\n\nAt the end of the follow-up period, 32 (94%) of the patients had progressed and 17 (53%) had died ([Table 2](#t2){ref-type=\"table\"}). The average time to progression was 4.7 \u00b1 2.9 months (median, 3.1 months; range, 0.9-12.1 months), and the most common type of progression was metastasis to the bone (in 39.1%), followed by nodal metastases (in 25.0%) and visceral metastases (in 21.9%).\n\n###### \n\n%TLF~10~ variation and outcome measures.\n\n  \u00a0    \u00a0        OS         \u00a0      PFS   \u00a0     TTBE   \u00a0   BMF                    \n  ---- -------- ---------- ------ ----- ----- ------ --- ----- ------ --- ----- ------\n  1    46.9%    Alive      17.4   \u00a0     Yes   3.0    \u00a0   No    17.4   \u00a0   No    9.9\n  2    -44.0%   Alive      11.8   \u00a0     Yes   2.8    \u00a0   No    11.8   \u00a0   No    8.8\n  3    -22.4%   Alive      18.7   \u00a0     Yes   11.9   \u00a0   No    18.7   \u00a0   No    13.2\n  4    -70.0%   Alive      11.4   \u00a0     Yes   0.9    \u00a0   No    11.4   \u00a0   No    8.0\n  5    -7.3%    Deceased   8.6    \u00a0     Yes   6.5    \u00a0   Yes   6.5    \u00a0   No    6.6\n  6    -33.0%   Deceased   8.7    \u00a0     Yes   3.4    \u00a0   No    8.7    \u00a0   Yes   3.4\n  7    -35.1%   Alive      9.0    \u00a0     Yes   5.2    \u00a0   No    9.0    \u00a0   Yes   7.7\n  8    -33.8%   Deceased   10.4   \u00a0     Yes   3.0    \u00a0   No    10.4   \u00a0   No    6.7\n  9    17.8%    Deceased   11.1   \u00a0     Yes   2.8    \u00a0   No    11.1   \u00a0   No    10.5\n  10   -51.3%   Deceased   9.0    \u00a0     Yes   2.2    \u00a0   Yes   2.2    \u00a0   No    4.5\n  11   -9.8%    Alive      13.9   \u00a0     Yes   2.8    \u00a0   Yes   6.2    \u00a0   No    9.5\n  12   39.6%    Alive      3.7    \u00a0     Yes   2.8    \u00a0   No    3.7    \u00a0   No    3.7\n  13   -24.5%   Deceased   6.0    \u00a0     Yes   3.1    \u00a0   No    6.0    \u00a0   Yes   4.5\n  14   -17.1%   Alive      10.7   \u00a0     Yes   7.7    \u00a0   No    10.7   \u00a0   No    7.6\n  15   34.6%    Alive      11.5   \u00a0     Yes   8.1    \u00a0   No    11.5   \u00a0   No    10.3\n  16   -22.4%   Alive      15.4   \u00a0     Yes   9.3    \u00a0   No    15.4   \u00a0   No    11.4\n  17   14.2%    Alive      4.8    \u00a0     Yes   2.8    \u00a0   Yes   2.8    \u00a0   Yes   4.3\n  18   -20.4%   Alive      11.9   \u00a0     Yes   7.1    \u00a0   No    11.9   \u00a0   No    6.1\n  19   -25.6%   Alive      9.5    \u00a0     Yes   4.2    \u00a0   No    9.5    \u00a0   No    7.9\n  20   -8.8%    Alive      13.9   \u00a0     Yes   8.1    \u00a0   No    13.9   \u00a0   No    11.7\n  21   16.1%    Deceased   8.7    \u00a0     Yes   2.8    \u00a0   No    8.7    \u00a0   No    3.7\n  22   65.5%    Alive      16.3   \u00a0     No    12.1   \u00a0   No    16.3   \u00a0   No    8.1\n  23   -26.0%   Deceased   4.7    \u00a0     Yes   1.8    \u00a0   No    4.7    \u00a0   No    3.5\n  24   -66.3%   Alive      17.7   \u00a0     Yes   10.0   \u00a0   Yes   4.5    \u00a0   No    12.1\n  25   -11.6%   Deceased   5.9    \u00a0     Yes   2.0    \u00a0   No    5.9    \u00a0   Yes   5.1\n  26   -18.9%   Deceased   5.0    \u00a0     Yes   2.8    \u00a0   No    5.0    \u00a0   Yes   2.8\n  27   104.0%   Deceased   12.1   \u00a0     Yes   3.2    \u00a0   No    12.1   \u00a0   Yes   11.9\n  28   -43.8%   Deceased   8.2    \u00a0     Yes   5.1    \u00a0   Yes   8.2    \u00a0   No    5.1\n  29   -28.8%   Alive      9.6    \u00a0     Yes   4.6    \u00a0   No    9.6    \u00a0   No    4.6\n  30   19%      Deceased   8.2    \u00a0     Yes   5.0    \u00a0   No    8.2    \u00a0   No    5.0\n  31   65.5%    Deceased   6.8    \u00a0     Yes   2.8    \u00a0   No    6.8    \u00a0   Yes   6.4\n  32   -84.5%   Deceased   5.8    \u00a0     Yes   2.8    \u00a0   No    5.8    \u00a0   Yes   5.6\n  33   -62.8%   Alive      9.3    \u00a0     No    4.9    \u00a0   No    9.3    \u00a0   No    4.8\n  34   8.4%     Deceased   5.4    \u00a0     Yes   3.1    \u00a0   No    5.4    \u00a0   No    4.2\n\nIn our study sample, the %TLF~10~ on the interim ^18^F-fluoride PET/CT was not able to predict OS (*p* = 0.6320; HR = 0.753; 95% CI: 0.236-2.401) or PFS (*p* = 0.5908; HR = 1.248; 95% CI: 0.557-2.797). Six patients had a bone event, and %TLF~10~ was also unable to predict the TTBE (*p* = 0.5114; HR = 1.588; 95% CI: 0.399-6.312). Nine patients developed BMF after ^223^RaCl~2~, and %TLF~10~ was also not a significant univariate predictor of the odds of developing that condition (*p* = 0.6071; HR = 1.401; 95% CI: 0.387-5.070). We found that OS did not correlate with the SUV~max~ (*p* = 0.7989), any nodal disease (*p* = 0.1342), or visceral disease (*p* = 0.1496).\n\nDISCUSSION\n==========\n\nWe have demonstrated that an interim ^18^F-fluoride PET/CT study is unable to predict outcomes after ^223^RaCl~2~ therapy. Novel therapies for osteoblastic metastases, including ^223^RaCl~2~ therapy, are costly, and it is therefore important to establish a diagnostic test to predict responses to these new, expensive treatments. In one study evaluating treatment responses after six cycles of ^223^RaCl~2~ in ten patients^(^[@r13]^)^, conventional bone scintigraphy demonstrated that increased areas of uptake were due not only to treatment response but also to reparative bone changes after therapy (a flare response).\n\nPrevious studies have shown that a baseline ^18^F-fluoride PET/CT study plays a prognostic role in patients with breast or prostate cancer treated with ^223^RaCl~2~^(^[@r1]^,^[@r14]^)^. However, ^18^F-fluoride PET/CT is not traditionally used in evaluating the response to any therapy, because the process of bone healing involves an osteoblastic reaction than can increase ^18^F-fluoride uptake, as in conventional bone scintigraphy^(^[@r15]^)^. Because of comparable pharmacokinetics between ^223^RaCl~2~^(^[@r2]^)^ and ^18^F-fluoride^(^[@r16]^)^, we hypothesized that ^18^F-fluoride would be able to evaluate osteoblastic metastases before, during, and after ^223^RaCl~2~ therapy.\n\nIn our study sample, the interim study demonstrated that a decrease in uptake was generally due to a response (partial or stable disease) to therapy. However, we find it interesting that, in six (17.6%) of the patients, the decreased uptake was caused by extensive tumor infiltration of the bone marrow, ultimately leading to BMF. To our knowledge, there have been no previous studies describing the latter imaging pattern (caused by BMF) in interim studies. In contrast, although the interim study was able to demonstrate that increased uptake was due to progression, that pattern of uptake was in fact a flare phenomenon in one case. This increased uptake most likely occurred because of the bone healing process after successful ^223^RaCl~2~ treatment, which involves an osteoblastic reaction. In the subset of patients in which the flare phenomenon occurred, the CT portion of the study revealed reparative changes with increased extent of the sclerotic lesions. However, even on CT, it was not possible to determine which patients were progressing and which were responding. Although we hypothesized that bone levels of ALP could help evaluate patient outcomes, it demonstrated higher specificity and lower sensitivity than did the interim ^18^F-fluoride PET/CT study.\n\nQuantitative analyses of ^18^F-fluoride PET/CT images have been conducted to assess its role in predicting outcomes, by determining the peak SUV~max~ values of bone metastases. Apolo et al.^(^[@r17]^)^ performed ^18^F-fluoride PET/CT after 6 and 12 months of standard therapy in prostate cancer patients, reporting that progression was associated with SUV increases of more than 57%, as well as that a greater increase in SUV was associated with worse survival. Yu et al.^(^[@r18]^)^ evaluated responses to therapy with dasatinib using SUV~max~ in five target lesions on ^18^F-fluoride PET/CT and detected only a borderline correlation with PFS; the changes also correlated with the ALP level. Another study, involving only five patients, showed a reduction in SUV~max~ at 6 and 12 weeks after the use of ^223^RaCl~2~^(^[@r19]^)^. In our population, the SUV~max~ did not correlate with OS. Although the above mentioned studies performed ^18^F-fluoride PET/CT for therapeutic evaluation, its precise role in determining the early response to therapy has yet to be extensively studied, especially in terms of assessing survival as an end point.\n\nThe reported frequency of the flare phenomenon in prostate cancer patients undergoing conventional bone scintigraphy ranges from 6% to 25%^(^[@r20]^,^[@r21]^)^. Although the flare phenomenon has also been described in patients undergoing ^18^F-fluoride PET/CT^(^[@r15]^)^, there have been no reports of its frequency in patients treated with ^223^RaCl~2~ and undergoing ^18^F-fluoride PET/CT. Although we identified the flare phenomenon on ^18^F-fluoride PET/CT in only a small proportion of our patient sample (3%), that proportion is probably higher than in conventional bone scintigraphy, given the greater sensitivity of PET/CT. The most likely explanation for the fact that the frequency of the flare phenomenon was not higher is that our study sample was composed of patients with extensive disease, in whom the likelihood of progression is greater than is that of a response to therapy. In addition, the number of patients might have been insufficient to detect this phenomenon.\n\nIn our patient sample, the interim ^18^F-fluoride PET/CT (%TLF~10~) after three cycles of ^223^RaCl~2~ was not able to predict OS, PFS, TTBE, or BMF. These findings are quite similar to those of a previous study, involving ten prostate cancer patients treated with ^223^RaCl~2~^(^[@r22]^)^, although the interim ^18^F-fluoride PET/CT studies were performed at different time points: at baseline; after one (or two) cycles of ^223^RaCl~2~; and at the end of treatment. A correlation with outcome was only noted between baseline and end-of-treatment ^18^F-fluoride PET/CT results were found to correlate with outcomes, as previously reported^(^[@r1]^)^.\n\nOne major limitation of our study was the relatively small number of patients. We believe that the interim ^18^F-fluoride PET/CT could have potential for the prediction of BMF, given that 6 of the 9 patients who evolved to BMF showed a reduction in uptake. However, due to the small sample size, those results were not significant. Another limitation was the fact that it was not possible to obtain histological confirmation in the patients who showed progression. Although the ^18^F-fluoride PET/CT images were acquired in different scanners, the same software was employed in all quantifications, guaranteeing uniformity in the metrics.\n\nTo our knowledge, this is the first study to evaluate the role of interim ^18^F-fluoride PET/CT in predicting the response to ^223^RaCl~2~ therapy, using quantitative methods to determine the skeletal tumor burden. It would be interesting to know whether these findings could be replicated in other populations, such as that of breast cancer patients treated with ^223^RaCl~2~.\n\nCONCLUSION\n==========\n\nIn prostate cancer patients undergoing ^223^RaCl~2~ therapy, interim ^18^F-fluoride PET/CT performed after three cycles of ^223^RaCl~2~ does not seem able to predict outcomes. It also appears to be unable to differentiate a flare response from progressive disease, and we therefore discourage the use of interim ^18^F-fluoride PET/CT to evaluate the response to ^223^RaCl~2~ therapy in prostate cancer patients. There is a need for studies involving a larger number of patients and patients with other types of cancer, in order to verify our findings.\n\n**Financial support**\n\nThis work received funding from the James E. Anderson Distinguished Professorship Endowment and from the Funda\u00e7\u00e3o de Amparo \u00e0 Pesquisa do Estado de S\u00e3o Paulo (FAPESP, S\u00e3o Paulo Research Foundation).\n\nWe would like to thank Patricia S. Fox for her advice regarding the statistical analysis.\n"}
+{"text": "Ehlers--Danlos Syndrome (EDS) affects collagen synthesis, resulting in pathologic connective tissue and joint hypermobility that clinically present as joint subluxations and dislocations.[@bib1] Recurrent instability of the hip limits the quality of life of patients due to constant apprehension and chronic pain, and it may predispose them to early joint degeneration.[@bib2] Conservative management is the initial treatment; however, in refractory cases, surgical intervention should be considered. Arthroscopic and open capsular plication techniques can be attempted; however, the native pathologic connective tissue increases risk of recurrence. Previously described by our institution, an open caspsuloligamentous reconstruction technique using Achilles tendon allograft has been successful for post-hip arthroscopy instability.[@bib3]\n\nWe present a surgical technique demonstrating an open anterior capsule reconstruction using an Achilles tendon allograft for native hip instability in EDS. We used an allograft to ensure normal collagen is used, as diseased connective tissue of EDS can subsequently stretch and instability can recur. This technique is unique as it is performed on a native hip and shows how the Achilles tendon allograft is fashioned to adopt the dynamic and static stabilizing features of the native iliofemoral ligament. Furthermore, this may be useful for hip arthroscopy surgeons, as it demonstrates a stabilization procedure for the hip that can be used in refractory instability cases post-hip arthroscopy.\n\nThe indication for this reconstructive procedure is native hip instability that has failed nonoperative treatment. A trial of activity modification, hip bracing, and gait aids is strongly recommended. The preoperative workup includes a functional history, hip examination focusing on instability, standard hip radiographs, and a magnetic resonance (MR) arthrogram. The MR arthrogram inflates the capsule and helps depict complete defects in the capsule, or patulous capsules, and is recommended over a nonarthrogram study.\n\nSurgical Technique {#sec1}\n==================\n\nRoom Setup {#sec1.1}\n----------\n\nThe room is set up with a flat top radiolucent table to allow for easy intraoperative fluoroscopic images, if needed. A graft table is required for graft preparation.\n\nSpecialized Equipment {#sec1.2}\n---------------------\n\nA direct anterior approach retractor set (Stryker, Kalamazoo, MI) is used to facilitate the capsular exposure ([Figs 1](#fig1){ref-type=\"fig\"} and [2](#fig2){ref-type=\"fig\"}). The reconstruction is performed with an Achilles tendon allograft with the calcaneous tuberosity still attached ([Fig 3](#fig3){ref-type=\"fig\"}). Suture anchor choice includes a 5.5-mm suture anchor (CrossFT; ConMed, Utica, NY), which is triple loaded with \\#2 HiFi sutures and is a biocomposite product. In areas of dense cortical bone such as the lesser trochanter (LT), a 5-mm titanium self-drilling triple-loaded anchor (ThRevo; ConMed) with \\#2 HiFi sutures can be used. HiFi suture (ConMed) is a high-strength braided nonabsorbable suture. The remaining equipment includes standard orthopedic surgery instrumentation.Fig 1Axial-cut magnetic resonance (MR) arthrogram of the left hip demonstrating a patulous capsule and deficient iliofemoral ligament anteriorly. The anterior capsule is heterogeneous in nature and not a homogenous black band as seen on a normal MR imaging study.Fig 2Stryker direct anterior approach retractor set. From left to right: Standard Cobra, Wide Cobra, Standard Homan, and Curved Homan.Fig 3Stryker retractors. Cobras are blunt (left), and Homans are sharp (right).\n\nPatient Position, Preparation, and Draping {#sec1.3}\n------------------------------------------\n\nThe patient is placed supine on a radiolucent flat top operative table. A dynamic examination of the hip is undertaken preoperatively to allow for a postoperative clinical comparison ([Video 1](#appsec1){ref-type=\"sec\"}). A bump is placed under the ipsilateral hip to allow extension for easier surgical exposure. All pressure points are padded. The operative leg, foot, groin, and as proximal as the umbilicus is prepped in standard fashion. The operative leg is then draped free to allow for intraoperative hip mobilization.\n\nApproach: Superficial Dissection {#sec1.4}\n--------------------------------\n\nWe use a direct anterior approach (Hueter\\'s approach) to the hip to provide the capsule exposure and working space. The anterior superior iliac spine is palpated and determines the location of the skin incision ([Fig 4](#fig4){ref-type=\"fig\"}). The straight leg is internally rotated, which accentuates the tensor fascia lata (TFL) muscle belly. The incision is based over the TFL muscle belly. The incision can be extended proximally for exposure ([Fig 4](#fig4){ref-type=\"fig\"}). The superficial fascia of TFL is identified and cleaned to aid with closure. The superficial fascia of the TFL is divided in line with the skin incision, taking care to protect the lateral femoral cutaneous nerve (LFCN), which lies medial to the fascial incision ([Fig 5](#fig5){ref-type=\"fig\"}).Fig 4Type of graft used. Achilles tendon allograft with the calcaneal tuberosity still attached. It originally comes with a large bone block and a single tendinous structure. The bone block is fashioned to a width and length of 2\u00a0cm each and a depth of 0.8\u00a0cm. The single tendinous structure is split longitudinally 50% of its length into 2 strands starting at the free end (non-bony end) to replicate the iliofemoral ligament (allowing for a Y-configuration). The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.Fig 5Surgical step 1. Location of the skin incision is 2 finger breadths lateral and 2 finger breadths distal to the anterior superior iliac spine. The tensor fascia lata muscle belly is accentuated with internal rotation of the leg, and the skin incision is centrally based on the muscle belly. Proximal extension above the flexor crease is marked out, which can help with exposure of the anterior inferior iliac spine (AIIS) and the gluteal pillar on the outer table of the pelvis where the graft will be docked. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.\n\nApproach: Deep Dissection {#sec1.5}\n-------------------------\n\nThe TFL muscle is dissected bluntly off the medial fascia. Retractor placement includes a wide Cobra (Stryker) at the lateral femoral neck, a standard Homan (Stryker) inferolateral to the greater trochanter, and a broad right-angle is used to retract the medial tissue. The ascending branch of the lateral femoral circumflex artery is now visualized and cauterized. The deep investing fascial layer of the TFL is divided. The precapsular fat pad of the hip is visualized and removed. The capsule is exposed with the help of the standard Cobra on the medial femoral neck and the curved Homan over the pelvic brim elevating the iliocapsularis, rectus femoris, and iliopsoas ([Fig 6](#fig6){ref-type=\"fig\"}). It is crucial to keep the curved Homan perpendicular to the inguinal ligament to prevent a femoral nerve palsy. Scar tissue on the anterior capsule is anticipated in patients with recurrent dislocations, and development of the appropriate planes with a Cobb elevator is helpful. This completes the anterior capsule exposure. The dissection is carried along the outer table of the pelvis, elevating the TFL off bone ([Fig 7](#fig7){ref-type=\"fig\"}). The 3 main anchor points for the allograft are now visualized; outer table of pelvis adjacent to the anterior inferior iliac spine (origin for the graft), adjacent to the lesser trochanter (medial limb of the graft), and the anterior prominence of the greater trochanter crest (lateral limb of the graft).Fig 6Surgical step 2. Incision of the tensor fascia lata (TFL) superficial fascia is in line with the skin incision. The Hueter\\'s approach through the TFL fascia helps to protect the lateral femoral cutaneous nerve as it lies medial to the superficial fascial incision. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.Fig 7Surgical step 3. Stryker direct anterior approach retractor placement to exposure the anterior hip capsule includes the wide Cobra at the lateral femoral neck, the standard Cobra at the medial femoral neck, and the curved Homan over the pelvic brim. The curved Homan needs to stay perpendicular to the inguinal ligament to prevent a femoral nerve palsy. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.\n\nCapsule Imbrication {#sec1.6}\n-------------------\n\nA new No. 15 scalpel is used to perform a capsule cut running from superolateral to inferomedial, avoiding the cartilage of the femoral head ([Fig 8](#fig8){ref-type=\"fig\"}). The capsule is then imbricated in a pants over vest technique using a \\#1 Vicryl stitch ([Fig 8](#fig8){ref-type=\"fig\"}).Fig 8Surgical step 4. Exposure of the anterior inferior iliac spine (AIIS) and outer table of the pelvis. Just adjacent to the AIIS is where the docking site of the graft is prepared. This 2-cm\u00a0\u00d7 2-cm area is decorticated with an oscillating saw until bleeding bone is visualized. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.\n\nGraft Preparation {#sec1.7}\n-----------------\n\nThe Achilles tendon allograft is prepared by drilling six 2.5-mm holes in the bone block to allow for horizontal mattress suture passage. The tendinous portion of the graft is divided into 2 equal width longitudinal limbs encompassing 50% of the overall length of the tendon, creating a Y-configuration ([Fig 4](#fig4){ref-type=\"fig\"}).\n\nGraft Deployment {#sec1.8}\n----------------\n\nThe docking sites for the graft were chosen to mimic the native iliofemoral ligament.[@bib4] The proximal docking area is located on the outer table of the pelvis just adjacent to the anterior inferior iliac spine abutting the anterolateral acetablulum.[@bib4] This 2\u00a0\u00d7 2-cm area is decorticated using an oscillating saw. A 5.5-mm suture anchor (CrossFT; ConMed) is then deployed in the decorticated docking site in a distally based direction to allow for better pullout strength, a theory borrowed from the shoulder deadman\\'s angle ([Fig 9](#fig9){ref-type=\"fig\"}). Each limb of suture in the triple loaded anchor is passed through the bone block to allow tying over the bone block in a horizontal mattress fashion ([Fig 10](#fig10){ref-type=\"fig\"}). The allograft limbs are now spun 180\u00b0 to create the corkscrew phenomenon and reciprocal tightening of the iliofemoral ligament. The medial limb is deployed at the anteroinferior femur just anterior to the LT,[@bib4] and finally the lateral limb to the anterior prominence of the greater trochanter crest[@bib4] in the same horizontal mattress fashion ([Fig 11](#fig11){ref-type=\"fig\"}). Occasionally, a ConMed ThRevo anchor, which is a 5-mm titanium self-drilling anchor, is required for the medial limb deployment as the cortical bone in this location can be very thick. We tension the allograft in 45\u00b0 of hip flexion and 15\u00b0 of internal rotation to allow for appropriate tension of the graft. Finally, each limb is oversewn to the capsule using \\#1 Vicryl stitch to augment the fixation ([Fig 12](#fig12){ref-type=\"fig\"}).Fig 9Surgical step 5. The capsule imbrication is an augmentation to the procedure. A superolateral to inferomedial cut in the capsule is made with a new No. 15 scalpel and the femoral head is exposed (left). When performing the capsulotomy, heavy pickups are used to lift the capsule off the surface of the femoral head to avoid iatrogenic damage to the underlying cartilage. The capsule is imbricated with a pants over vest technique using \\#1 Vicryl suture (right). The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.Fig 10Surgical step 6. Proximal anchor insertion is done at a 45\u00b0 distal angle at the docking site, just adjacent and lateral to the anterior inferior iliac spine (AIIS). The 45\u00b0 distal angle of anchor insertion helps with pullout strength. A triple-loaded anchor is used. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.Fig 11Surgical step 8. The Achilles tendon allograft shown fixed in place by use of proximal, medial, and lateral triple-loaded suture anchors. The Y-limbs are spun 180\u00b0 to allow for the corkscrew effect and reciprocal tightening of the iliofemoral ligament. The medial suture anchor is positioned just anterior to the lesser trochanter, and the lateral suture anchor is positioned on the anterior prominence of the greater trochanter crest. The limbs are tensioned in flexion and internal rotation of the hip. The wound is then irrigated, and closed in layers starting with the tensor fascia lata fascia, subcutaneous layer, and finally skin. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.Fig 12Surgical step 7. The bone block attachment of the Achilles tendon allograft is docked by passing all 6 strands of the triple loaded anchor to allow for horizontal mattress suture fixation over a bone bridge. The figures depicting the surgical procedure are in a supine position, on a left hip. Proximal, distal, medial, and lateral are the top, bottom, left, and right of the image, respectively.\n\nTesting {#sec1.9}\n-------\n\nThe graft undergoes clinical testing with range of motion, and the corkscrew phenomenon with reciprocal tightening of the reconstructed iliofemoral ligament is observed ([Video 1](#appsec1){ref-type=\"sec\"}).\n\nClosure {#sec1.10}\n-------\n\nThe wound is irrigated with normal saline to prevent infection and heterotopic ossification formation. We use 3\u00a0g of tranexamic acid topically to allow for added hemostasis. Pericapsular local anesthetic is infiltrated for postoperative pain control. The superficial fascia of the TFL is closed in a running fashion with intermittent locking stitches, paying close attention to not capture the LFCN. The wound is then closed as per the surgeon\\'s preference.\n\nPostoperative Examination and Postoperative Rehabilitation {#sec1.11}\n----------------------------------------------------------\n\nThe hip is taken through the same examination as done preoperatively to evaluate and compare the stability of the hip ([Video 1](#appsec1){ref-type=\"sec\"}). The patient remains touch weight bearing for 6\u00a0weeks. Progressive physiotherapy is then started for range of motion. Strengthening begins at 12\u00a0weeks.\n\nDiscussion {#sec2}\n==========\n\nThe presentation of native hip instability is rare; however, it can occur in patients with connective tissue disorders. In 2015, Larson et\u00a0al.[@bib5] reported outcomes on 16 patients with EDS-hypermobility type who underwent hip arthroscopy for femoroacetabular impingement and capsular laxity with capsular plication. After a mean follow-up of 45\u00a0months, there were significant improvements in all outcome measures, with 1 patient requiring revision arthroscopy for recurrent pain and subjective instability. In the present case, the authors felt that a capsular plication alone would be insufficient, given the recurrent history of symptomatic instability and evidence on MR imaging (MRI) of an attenuated anterior hip capsule. Furthermore, we believe there are limits to arthroscopic management in patients with global capsular insufficiency, especially those with collagen disorders. The use of native collagen for capsular plication in patients with EDS is at greater risk of recurrence. We advocate allograft reconstruction using an open direct anterior technique to reduce the risk of recurrence.\n\nRisks to an open anterior capsule reconstruction of the hip should be mentioned ([Table 1](#tbl1){ref-type=\"table\"}). Risks can be divided into immediate and late risks. Due to the open and invasive characteristics of this approach, immediate risks include nerve (lateral femoral cutaneous nerve, femoral nerve), artery and blood loss (ascending branch of the lateral circumflex femoral artery), infection (open nature of the procedure), iatrogenic cartilage injury (during the capsulotomy), scar formation (superficial and deep), and postoperative pain management. Late risks are recurrence of instability, and theoretical overconstraint of the hip leading to early arthritis, akin to the historic procedure of shoulder capsulorrhaphy leading to shoulder arthropathy.Table\u00a01Advantages and Disadvantages for Open Anterior Capsule Reconstruction of the Hip Using an Achilles Tendon AllograftAdvantagesDisadvantages1.Open approach allows for accurate and full exposure of the docking sites for the allograft.2.Open approach allows for augmentation with a capsular plication.3.Open approach allows to insert a large and robust graft acting as an internal brace for the hip.4.An open approach is not subject to traction time as seen in an arthroscopic approach.5.Allograft technique decreases the recurrence rate in the background of connective tissue disorders.6.Using an Achilles tendon allograft with the calcaneal tuberosity as the bone plug allows for bone to bone healing.1.Surgical expertise required in open surgery, specifically facile in the direct anterior approach.2.Increased risk to local structures, including nerves, arteries, and cartilage.3.Allograft use is of greater cost than autograft.4.Increase risk of infection with and open approach.5.Without the arthroscope, a full diagnostic evaluation of the intraarticular compartments of the hip is difficult.\n\nThis technique allows for a step-by-step guide in anterior capsule reconstruction using an Achilles tendon allograft with a calcaneal bone block in situ. The bone block is helpful, as it provides bone on bone healing at the proximal docking site. The Achilles tendon allograft is sturdy, and can be split distally into a Y-configuration to replicate the iliofemoral ligament.\n\nWe like to use biocomposite suture anchors to allow for postoperative MRI if necessary. Using metal suture anchors will result in a large metal artifact on MRI and limit the quality and interpretation of the images. However, metal suture anchors can be useful in dense cortical bone areas such as the medial limb anchor site anterior to the LT ([Table 2](#tbl2){ref-type=\"table\"}).Table\u00a02Pearls and Pitfalls for Open Anterior Capsule Reconstruction of the Hip Using an Achilles Tendon AllograftPearlsPitfalls1.Use a flat-top radiolucent table to prepare for possible intraoperative fluoroscopy.2.Place a bump under the ischiums (folded flannel blanket works well) to allow some extension to the hip, which allows for easier exposure.3.Prep the leg free to allow full control of the limb throughout the procedure.4.Bias the skin incision lateral, centered over the TFL to prevent unintentional dissection medial to the Sartorius.5.Ensure complete hemostasis of the ascending branch of the lateral circumflex femoral artery to prevent uncontrolled bleeding later on in the case.6.Use of DAA retractors is advised for proper and full exposure.7.Consider a metal suture anchor at the medial site anterior to the LT; this is an area of dense cortical bone.1.Rushing through the exposure. The exposure is the foundation of this reconstructive procedure.2.One must stay superficial to the hip capsule; do not stray into the anterior capsule remnant while reflecting the iliocapsularis and reflected head of the rectus femoris from the anterior capsule.3.Allowing the superior acetabular retractor to fall away from the line perpendicular to the inguinal ligament (this may cause a femoral nerve palsy).4.Fracture of the bone plug on the allograft while either drilling the suture holes or docking the graft.5.Inadequate tensioning of the graft in flexion and internal rotation of the hip, causing a lax graft and defeating the purpose of the reconstruction.[^1]\n\nHip instability cases invariably have had repetitive subluxations and resultant tissue trauma to the hip. Impressionable scar tissue forms anterior to the capsule, and the reflected head of rectus femoris as well as the iliocapsularis are significantly adhered at this location. Care must be taken to elevate these tissues completely to obtain the proper tissue plane and allow adequate exposure.\n\nMany capsuloligamentous reconstruction techniques have been previously described to address symptomatic instability in both native and non-native hip joints.[@bib6], [@bib7], [@bib8], [@bib9] However, the technique in this paper is unique in that using a direct anterior approach to a native hip, the anatomic configuration of the iliofemoral ligament is replicated with an Achilles tendon bone block allograft to turn the original insufficient capsule into a now enhanced reconstructed iliofemoral ligament.\n\nSupplementary Data {#appsec1}\n==================\n\nVideo 1Surgical video on the step-by-step technique for anterior capsule reconstruction using an Achilles tendon allograft performed through a direct anterior approach to a left hip.ICMJE author disclosure forms\n\nThe authors report the following potential conflicts of interest or sources of funding: D.W. is a consultant for Intellijoint. O.R.A. is on the Speakers Bureau for ConMed. Full ICMJE author disclosure forms are available for this article online, as [supplementary material](#appsec1){ref-type=\"sec\"}.\n\n[^1]: DAA, direct anterior approach; LT, lesser trochanter; TFL, tensor fascia lata.\n"}
+{"text": "Introduction {#s1}\n============\n\nThe pathogenesis of IgA nephropathy (IgAN) involves multiple 'hits' \\[[@SFV076C1], [@SFV076C2]\\]. It is thought that increased levels of circulating galactose-deficient IgA1 in association with the production of unique anti-glycan antibodies leads to the formation of pathogenic IgA1 containing circulating immune complexes that are deposited within the mesangium leading to activation of mesangial cells and thus glomerular damage. That complement components including C3, C4d, C5b-9, properdin, factor H, C4BP and MBL are deposited within the glomerulus in IgAN has been known for some time \\[[@SFV076C3]--[@SFV076C5]\\]. In addition it has been shown that there are genetic susceptibility factors for IgAN that involve complement. A common deletion in the regulators of complement activation cluster at 1q32, which incorporates the genes (*CFHR3* and *CFHR1*) encoding factor H-related protein 3 and 1, has been shown to protect against IgAN \\[[@SFV076C6]\\], possibly by reducing the ability of complement factor H-related proteins to inhibit the regulatory function of CFH \\[[@SFV076C7]\\]. It thus seems possible that complement inhibition might be beneficial in the treatment of IgAN and in particular the crescentic form with rapidly progressive glomerulonephritis (RPGN), which has a poor prognosis \\[[@SFV076C8]\\]. We report here the use of eculizumab in a young man with Henoch--Sch\u00f6nlein purpura and crescentic IgAN.\n\nCase history {#s2}\n============\n\nA 16-year-old man was admitted in September 2013 with a 2-week history of mild abdominal pain, arthralgia and purpura on the lower extremities. He had previously been treated with penicillin for an unconfirmed diagnosis of erysipelas. Blood pressure was 116/69 mmHg, serum creatinine was 59 \u00b5mol/L (1.08 g/dL) and creatinine clearance 159 mL/min. Urinalysis showed haematuria and proteinuria. Serum albumin was 30 g/L but there was no evidence of oedema. Since 2010 he had been treated with atomoxetine for attention deficit--hyperactive disorder. Otherwise there was no significant past medical history. He reported occasional use of marihuana, amphetamines and cocaine. He smoked 15--30 cigarettes daily. Serological tests to exclude ANCA-associated vasculitis and SLE were negative. C3 and C4 levels were normal. A diagnosis of Henoch--Sch\u00f6nlein purpura was made and he was started on an angiotensin-converting enzyme inhibitor. Two months later he was readmitted with increasing creatinine of 98 \u00b5mol/L (1.11 g/dL) and declining serum albumin of 19 g/L.\n\nA biopsy (Figure [1](#SFV076F1){ref-type=\"fig\"}A) revealed epithelial crescents in 6 of 14 glomeruli, mesangial hyperplasia, endocapillary hypercellularity and deposition of both IgA (++++/4+) and C3 (+++/4+) on immunofluorescence (M1 S0 E1 T0 according to Oxford classification). Fig. 1.(**A**) Glomerulus with a cellular crescent in a first renal biopsy; \u00d720, PAS. (**B**) Segmental glomerulosclerosis and atrophy of tubuli and fibrosis of interstitium in a second renal biopsy 11 months later, \u00d720, PAS (left) and IgA immunofluorescence (right).\n\nBecause of the rapid loss of renal function and the acute changes on biopsy with no evidence of chronic damage it was elected to treat him with prednisolone 40 mg daily and cyclophosphamide 100 mg (1.5 mg/kg) \\[[@SFV076C9]\\]. He was also anticoagulated with warfarin and low molecular weight heparin. Despite this, renal function deteriorated further. He was, therefore, given three daily pulses of methylprednisolone 500 mg followed by five plasma exchanges (40 mL/kg). This did not result in any improvement. As a last resort---and without any direct proof of complement involvement---we decided to treat with eculizumab. Four weekly doses of 900 mg were administered intravenously followed by a single dose of 1200 mg accompanied throughout by penicillin. Immediately after the first dose of eculizumab a remarkable improvement in renal function was evident. After 3 months cyclophosphamide was replaced with azathioprine and the dose of prednisolone was tapered. The course is shown in Figure [2](#SFV076F2){ref-type=\"fig\"}. Fig. 2.Course of the disease.\n\nA further renal biopsy (Figure [1](#SFV076F1){ref-type=\"fig\"}B) taken 11 months after presentation revealed chronicity with 2 of 19 glomeruli completely and 7 partially sclerosed, tubular and interstitial changes with 50% atrophy and fibrosis. There was evidence of residual activity with a single glomerulus showing a cellular crescent. According to Oxford classification the changes were considered as M1 S1 E1 T1. Accordingly, renal function slowly declined to end-stage in spring 2015 with unabated nephrotic range proteinuria but consistently well-controlled blood pressure.\n\nFurther complement analysis has been undertaken. C3 was 1.77 g/L, C4 0.30 g/L, factor H 0.79 g/L and factor I 81 mg/L. CD46 expression on peripheral blood mononuclear cells was normal and factor H autoantibody screening was negative. Mutation screening of *CFH*, *CFI*, *CD46*, *C3* and *CFB* showed no abnormalities. Multiplex ligation-dependent analysis (MLPA) showed no evidence of genomic disorders affecting *CFH* or *CFHR1-5*.\n\nDiscussion {#s3}\n==========\n\nKDIGO guidelines recommend that steroids and cyclophosphamide be used to treat crescentic (\\>50% crescents) IgAN \\[[@SFV076C10]--[@SFV076C12]\\] with RPGN. We report here a case of crescentic IgAN unresponsive to this regimen. Furthermore, the introduction of plasma exchange did not appear to be beneficial. Therefore, because of the possible involvement of complement in the pathogenesis of IgAN we elected to treat the patient with eculizumab as a last resort. With this there was a significant immediate improvement in renal function. Proteinuria, however, persisted and a repeat renal biopsy has shown evidence of chronicity. As with all anecdotal reports it is difficult to know whether the improvement in renal function can be ascribed to the intervention. It is indeed possible that the effect of the massive therapy provided incidentally materialized at exactly the moment eculizumab was started. There is, however, evidence emerging from other forms of glomerulonephritis that complement blockade may play a role in inducing remission in aggressive forms including crescentic dense deposit disease \\[[@SFV076C13]--[@SFV076C15]\\] and lupus nephritis \\[[@SFV076C16]\\]. This is perhaps not surprising given that complement activation results in the generation of powerful inflammatory anaphylatoxins such as C5a. This is supported by the temporal relationship seen in the time course. Macrohaematuria may cause acute kidney injury and persistent injury \\[[@SFV076C11]\\] but one would then expect tubular injury rather than crescentic disease on biopsy. Notwithstanding free haemoglobin could exacerbate the disease process by activating complement \\[[@SFV076C17]\\].\n\nUse of eculizumab in another patient with IgAN has been reported \\[[@SFV076C18]\\]. In that case eculizumab was started when deterioration in renal function occurred despite treatment with mycophenolate mofetil and steroids. The introduction of eculizumab was associated with temporary stabilization in renal function but the therapy was eventually withdrawn when a further renal biopsy showed severe chronicity. These two cases suggest that the mechanisms underlying chronic progression may not be amenable to complement inhibition as we provided it. Recently it was shown that inhibiting C5aR resulted in more rapid amelioration of proteinuria than steroid in patients with ANCA-positive vasculitis \\[[@SFV076C19]\\].\n\nConclusion {#s4}\n==========\n\nFollowing the comprehensive treatment in our patient, crescents improved, but nonetheless severe and progressive chronicity developed. If this deterioration is the result of the initial severe crescentic disease, perhaps earlier treatment with eculizumab could improve the final outcome. However, it is also possible that even if complement is involved in the causation of nephritis, either components proximal to the effect of eculizumab could be paramount or there is no direct relationship between initial inflammation and development of chronic scarring. Many more complement studies would be needed to assess these possibilities. We were not able to accomplish that during the acute treatment of our desperately ill patient.\n\nPatient consent {#s5}\n===============\n\nThe patients concerned in this case report consented to its publication after seeing the final version.\n\nConflict of interest statement {#s6}\n==============================\n\nT.R. has received lecture fees from Alexion Pharmaceuticals. T.H.J.G. has received honoraria for consultancy work from Alexion Pharmaceuticals.\n\n(See related articles by Rojas-Rivera *et al.* Rapidly progressive IgA nephropathy: a form of vasculitis or a complement-mediated disease? *Clin Kidney J* (2015) 8: 477--481 and by Yang *et al*. Clinical features of IgA nephropathy with serum ANCA positivity: a retrospective case--control study. *Clin Kidney J* (2015) 8: 482--488.)\n"}
+{"text": "Introduction {#S1}\n============\n\nHeart failure (HF) is a clinical syndrome characterized by structural and/or functional impairment of ventricular filling or ejection of blood resulting in insufficient perfusion to meet metabolic demands. There is no single diagnostic test for HF, it is a clinical diagnosis based on history, physical examination, laboratory and imaging parameters \\[[@R1]\\]. Thyroid hormones (TH) have numerous effects on body systems, especially the heart and cardiovascular system including effects on the relaxation and contractile properties of the heart and are critical in preserving cardiac structure \\[[@R2]\\]. In recent years, studies have shown that alterations in TH are associated with a wide spectrum of cardiovascular diseases - specifically, hypothyroidism and subclinical hypothyroidism have been reported to be associated with increased incidence and worsening of HF, with and without underlying heart disease \\[[@R3],[@R4]\\]. The aim of this review is to evaluate the effects of hypothyroidism and subclinical hypothyroidism. We will also discuss the postulated mechanisms that may induce and/or exacerbate HF and highlight the appropriate management strategies.\n\nPrevalence of Decrease T3/HF/Subclinical and Associations (Prevalence of Hypothyroidism/Subclinical Hypothyroidism in HF) {#S2}\n=========================================================================================================================\n\nNearly 10 million people (4.6%) in the United States have hypothyroidism. Most of them are asymptomatic, i.e. with subclinical hypothyroidism (4.3%). In iodine-replete communities, the prevalence of spontaneous hypothyroidism is between 1 and 2%, and it is 10 times more common in women than in men, and particularly prevalent among older women. Studies in Northern Europe, Japan and the USA have found the prevalence to range between 0.6 and 12 per 1000 women and between 1.3 and 4.0 per 1000 in men investigated \\[[@R5]\\].\n\nTunbridge et al. conducted a study in Whickham, England to determine the prevalence of thyroid disorders in the community and reported that 7.5% of women and 2.8% of men of all ages had thyroid stimulating hormone (TSH) levels greater than 6 mlU/L. After reviewing 12 studies across different cultures, the Whickham study concluded that primary thyroid gland failure (TSH\\>6 mlU/L) is 5% in multiple populations \\[[@R6]\\]. Moreover, in the Colorado Thyroid Disease Prevalence Study, 9.4% of the subjects had a high-serum TSH concentration, of whom 9.0% had subclinical hypothyroidism \\[[@R7]\\]. Among those with an elevated serum TSH concentration, 74% had a value between 5.1 and 10 mlU/l and 26% had a value \\> 10 mlU/l. Women had a higher percentage of high serum TSH concentration versus men in each decade of age, and ranged from 4 to 21% in women and 3 to 16% in men.\n\nThe National Health and Nutrition Examination Survey, composed of 4392 participants conducted between 1999--2002, noted a 3.7% prevalence of hypothyroidism in the general population. It also demonstrated that the serum TSH concentrations increased with age in both men and women and were higher in whites than in blacks, independent of serum anti-thyroid antibody concentrations \\[[@R8],[@R9]\\].\n\nHeart failure (HF) has been considered an epidemic and a global health problem, with a prevalence of over 5.8 million in the USA and over 23 million worldwide \\[[@R10]\\]. The estimates of HF prevalence in developed countries generally range from 1--2% of the adult population \\[[@R11]\\]. Although the age-adjusted incidence and prevalence of HF are decreasing, the absolute number of patients with HF has drastically increased, secondary to shifts in the global age distribution, increased life expectancy, medical advancement and general population growth \\[[@R12]\\]. HF incidence has shown signs of stabilization and possible reduction in developed countries based on community-based cohorts, such as Framingham and Olmstead county \\[[@R13],[@R14]\\]. However, the incidence of HF varies between ethnic groups in the USA. The Multi-Ethnic Study of Atherosclerosis reported the highest incident rates of HF among African-American individuals, intermediate rates among Whites and Hispanic individuals, and the lowest rates among Chinese-American individuals \\[[@R15]\\].\n\nNing et al. conducted a meta-analysis including 19,354 subjects with HF, 2173 with hypothyroidism, to clarify the association of hypothyroidism and all-cause mortality and morbidity in patients with HF. The analysis reported that hypothyroidism and subclinical hypothyroidism were associated with increased all-cause mortality even after correction with thyroid replacement therapy. Moreover, hypothyroidism and subclinical hypothyroidism were associated with increased risk of hospitalization in HF patients \\[[@R3]\\].\n\nPathophysiology of Decrease T3 and Effect on HF {#S3}\n===============================================\n\nThe American College of Cardiology defines HF as a complex clinical syndrome that impairs the ability of the ventricle to fill with or eject blood \\[[@R16]\\]. In recent years, studies have shown that untreated overt hypothyroidism is an important cause of HF \\[[@R3],[@R17]\\]. Moreover, persistent subclinical hypothyroidism has been associated with the development of HF in patients with and without underlying heart disease \\[[@R18]\\].\n\nThyroid gland produces and releases hormones mostly as the prohormone thyroxine (T4). triiodothyronine (T3), the biological active thyroid hormone, derives from the conversion of T4 by deiodinase enzymes. Three deiodinase enzymes regulate circulating and tissue concentration of THs: type 1 (D1); type 2 (D2), and type 3 (D3). D1 is considered the major peripheral source of circulating T3 and is commonly found in the liver and the kidney, whereas D2 plays a critical role in providing local conversion to T3in the heart, skeletal muscles, brain and pituitary tissues. D3 is involved in the inactivation and degradation of T3 \\[[@R19]\\]. Due to the necessity of T3 in preserving both cardiac morphology and performance in adult life, the heart is sensitive to reductions in local T3 levels.\n\nGenomic nuclear effects of thyroid hormones initiate most of physiologic effects in humans. T3 binds to specific nuclear thyroid hormone receptors (TRs), which subsequently bind as homodimers or heterodimers to thyroid hormone response elements in the promoter region of some genes \\[[@R20]--[@R22]\\]. There are two isoforms of receptors generated by two TR genes in the heart. TRa1 transcripts binds to T3 with high affinity and acts as a positive regulator, whereas TRa2 acts as a negative regulator as it binds TREs but does not bind to T3 \\[[@R20]--[@R23]\\].\n\nThere are many cardiac structures that are transcriptionally regulated by T3, such as sarcoplasmic reticulum calcium ATPase (SERCA2), alpha - myosin heavy chain (\u03b1MHC), B1 adrenergic receptors sodium/potassium ATPase, voltage-gated potassium channels, malic enzyme and atrial and brain natriuretic hormone \\[[@R20],[@R21]\\]. Other cardiac genes are negatively regulated by T3 namely \u03b2MHC, phospholamban, sodium/calcium exchanger, TRa1 and adenylyl cyclase type V and VI \\[[@R20]--[@R21]\\]. However, T3 also has non genomic effects on cardiac myocyte and peripheral vascular resistance. These effects involve the transport of ions across the plasma membrane, glucose and amino acid transport and mitochondrial function \\[[@R21],[@R23]\\].\n\nThe effects of T3 allow it to have control on the inotropic and lusitropic properties of the myocardium and have an influence on cardiac growth, myocardial activity and vascular function. As a result, thyroid hormone deficiency is likely responsible for an increased risk of HF events, by causing cardiac atrophy, chamber dilation and impaired myocardial blood flow. ([Figure 1](#F1){ref-type=\"fig\"})\n\nCases of hypothyroidism and reversible dilated cardiomyopathy have been reported. Changes in \u03b1MHC expression by measuring mRNA, extracted from endomyocardial biopsy, of a hypothyroid patient with dilated cardiomyopathy before and after T4 replacement therapy. The administration of thyroid hormone therapy and restoration of euthyroidism produced an increase in \u03b1MHC gene expression and reversed the cardiomyopathy \\[[@R24]\\].\n\nRelation of Decreased T3 and Other Heart Diseases {#S4}\n=================================================\n\nThe total thyroid hormones produced by the thyroid gland is 80% thyroxine (T4) and 20% triiodothyronine (T3), however the concentration of T3 in the plasma is 1/40 of T4. Thyroxine has an intrinsic effect whereas T3 is necessary for all metabolic activities and used by all organs. In general, the effects of low T3 are essentially the opposite of hyperthyroidism, with indirect and direct effects on the heart ([Table1](#T1){ref-type=\"table\"}) \\[[@R20],[@R1]\\]. In the previous section the effects of Hypothyroidism on heart failure were explained. In this section we will focus on the effects of low T3 on the cardiovascular system other than heart failure.\n\nThe major changes of low T3 on the cardiovascular system are related to decreased cardiac output, cardiac contractility, bradycardia, and increased peripheral vascular resistance. But many other atherosclerotic modifiable risk factors like high concentrations of cholesterol; accelerated atherosclerosis, coronary artery disease and impaired endothelial derived relaxation factor (nitric oxide) are also affected \\[[@R25],[@R26]\\]. Systemic vascular resistance can increase as much as 50% due to the decrease in arterial compliance \\[[@R26],[@R27]\\]. In hypothyroidism, as opposed to hyperthyroidism, atrial arrhythmias are rare and ventricular ectopy is common \\[[@R26]\\]. Low T3 levels by regulatory effects on the expression of numerous ion channels in the heart (B-adrenergic, Na/K ATPase, Voltage-gated K channels, Na/Ca exchanger etc.) \\[[@R28],[@R29]\\], tend to prolong the cardiac action potential and hence the QT interval. Consequently, attenuated activity on precordial examination if often appreciated predisposes the patient to ventricular irritability and, in rare cases, acquired torsade de pointes \\[[@R30]\\].\n\nEffect of Treating Hypothyroidism in Heart Failure {#S5}\n==================================================\n\nHypothyroidism is a reversible cause of HF. Consequently, thyroid function should be evaluated in patients with HF and non-ischemic dilated cardiomyopathy. The American College of Cardiology guidelines for HF recommend screening for serum thyroid hormones levels for all newly diagnosed HF patients \\[[@R16]\\]. Hypothyroidism has many effects on the heart's physiology and internal blood supply. Studies have shown the administration levothyroxine (LT4) can actually reverse myocyte apoptosis, improve cardiovascular performance and ventricular remodeling in hypothyroidism \\[[@R3],[@R31]\\]. Moreover, diastolic dysfunction, impaired ventricular filling and coronary flow improve when euthyroidism is restored. The Cardiovascular Health Study demonstrated that LT4 may decrease the risk of developing heart failure in patients with subclinical hypothyroidism (SHypo) and TSH\\>10 \\[[@R4]\\]. Consequently, it appears that replacement doses of LT4 should be considered in patients with SHypo and TSH\\>10 mU/l to prevent the risk of developing HF.\n\nAdditionally, thyroid hormone replacement has been shown to decrease total cholesterol, low-density lipoproteins (LDL) and triglycerides, reduce blood pressure, improve diastolic function, and control heart rate both at rest and during exercise. These documented effects all contribute to a theoretical risk reduction in developing atherosclerosis \\[[@R20],[@R24],[@R32],[@R33]\\].\n\nHowever, a problem that many physicians encounter is the fact that hormonal therapy may precipitate both fatal arrhythmias and/or myocardial ischemic events. Per Somwary et al., 20% of patients with hypothyroidism maybe over treated during replacement therapy leading to increased risk of atrial fibrillation, especially among the elderly \\[[@R34],[@R35]\\].\n\nHence, the usual approach (start low, go slow) is to start with the lowest dose of levothyroxine and titrate up until a euthyroid state has been achieved. Despite the improvement in both symptoms and cardiac contractility, close observation is needed when starting thyroid hormonal therapy, particularly in elderly and in those with known coronary ischemia \\[[@R36],[@R37]\\].\n\nEffect of Subclinical Hypothyroidism on Heart Failure and the need for Therapy? {#S6}\n===============================================================================\n\nSubclinical hypothyroidism is a biochemical diagnosis; it is defined as a normal T4 concentration while also having an elevated serum TSH concentration. Many patients usually do not have any symptoms; therefore, it is merely considered a laboratory diagnosis. Its prevalence ranges from 4 to 15 percent in the United States \\[[@R18]\\]. There have been prospective studies suggesting that the annual rate of progression to overt hypothyroidism ranges from 2 to 4 percent depending on the initial TSH level. Higher TSH levels have been associated with increased cardiovascular disease including heart failure and coronary heart disease. There is some discrepancy about the exact levels related to increased risk, but the general consensus is that levels greater than 10 mU/L are associated with more cardiovascular pathology. The Cardiovascular Health Study revealed that patients with a TSH equal or greater than 10 mU/L had an increased risk of HF and a greater baseline peak E velocity, which is a measurement of diastolic function associated with HF incidence, after adjustment for age, gender, and systolic blood pressure compared to euthyroid participants \\[[@R4]\\]. However, patients with TSH 4.5 to 9.9 mU/l had no increased HF risk. Increased LV mass, impairment of LV relaxation were also exclusively associated with subclinical hypothyroidism with TSH \\>10.0 \\[[@R4]\\]. The Health Aging and Body Composition population-based study showed that patients (aged 70--79 years) with TSH level 7 mU/l or greater had a higher rate of CHF incidence and recurrence compared with euthyroid patients. Among the 127 patients who had HF, 51 had recurrent HF events \\[[@R38]\\]. Furthermore, in the PROSPER study patients with persistent SHypo had an increased rate of HF hospitalization compared with euthyroid controls with an age- and sex-adjusted HR of 4.99 (95% CI, 1.59--15.67) \\[[@R39]\\].\n\nWhen treating a patient with heart failure, that is also found to have subclinical hypothyroidism, there are many discrepancies in regards to starting treatment with levothyroxine. Some studies showed beneficial outcomes when treating with hormone replacement, including decreased blood pressure, LDL, total cholesterol and carotid intima thickness \\[[@R40]\\]. But other studies have shown no significant reduction on CAD \\[[@R41]\\]. On the contrary, analysis from a population based cohort study has linked increased cardiovascular mortality in terms of ischemic heart disease and dysrhythmias in patients treated for subclinical hypothyroidism with levothyroxine. This trend has been seen mainly in elderly patients (\\>70 years old). After the European Thyroid Association proposed guidelines in regards to this matter, there has been a general consensus to treat with levothyroxine all those patients \\<70 years old who have a TSH serum concentration \\>10 mIU/L. For patients \\>70 years old with TSH levels \\<10 mIU/L the guidelines discourage from treating, given higher chances of exacerbating ischemic events and arrhythmias in CAD patients \\[[@R42]\\].\n\nStudies show that 7 out of 10 elderly women will have subclinical hypothyroidism with nearly the same changes in cardiovascular function, but less marked, than overt hypothyroidism \\[[@R43]\\]. A study performed in Netherlands comprised of 1149 postmenopausal women demonstrated women with subclinical hypothyroidism where more likely to have a myocardial infarction and a higher frequency of calcification to the aorta \\[[@R44]\\]. When patients with subclinical hypothyroidism are treated with thyroxine, systolic and diastolic cardiac function improves \\[[@R45]\\]. Nevertheless, screening and management strategies are still a subject of disagreement, but from the cardiac perspective, treating a patient with high serum thyrotropin and normal serum thyroid hormone concentration, seems to offer a benefit with minimal risks \\[[@R46],[@R47]\\].\n\nAt the present time, there is no randomized clinical trial evaluating the long-term cardiovascular outcomes in patients with subclinical hypothyroidism receiving hormonal replacement with levothyroxine. Further studies are needed to inform guidelines when encountering these patients, especially in the elderly population \\>65 years old.\n\nConclusion {#S7}\n==========\n\nIn this review, we outlined the epidemiology and the relationship between hypothyroidism and heart failure. We also discussed the pathogenetic mechanisms by which thyroid hormone, directly and indirectly influences cardiac function. Furthermore, we highlighted the rationale for therapy with T4, particularly in those with TSH\\>10mU/l\n\nIt is clear that both overt and subclinical hypothyroidism can have varied but profound effects on cardiac hemodynamics and physiology contributing to cardiac related morbidities. While the cardiovascular effects of subclinical hypothyroidism are clear, the current management guidelines are still debated. Currently, very high TSH levels in an asymptomatic patient are generally treated, but less definitive data is available regarding initiation of therapy. It is also important to note that close monitoring is recommended especially in the elderly. Overall, there is evidence to suggest that treatment of subclinical hypothyroidism can improve cardiovascular outcomes; however, randomized controlled clinical trials in this field are lacking and warranted.\n\nThis work is supported, in part, by the efforts of Dr. Moro O. Salifu M.D., M.P.H., M.B.A., M.A.C.P., Professor and Chairman of Medicine through NIH Grant number S21MD012474.\n\nCompeting Interests\n\nThe authors declare that they have no competing interests.\n\n![T3 has both genomic and nongenomic effects on the cardiac myocyte. Genomic effects involve binding of T3 to TRs, which in turn, positively or negatively regulates the transcription of specific cardiac genes. Nongenomic mechanisms include direct modulation of membrane ion channels. AC = adenylyl cyclase; \u03b2-AR = \u03b2-adrenergic receptor; Ca2+ = calcium ions; Ca2+ ATPase = sarcoplasmic reticulum calcium adenosine triphosphatase; cAMP = cyclic adenosine monophosphate; GS = stimulatory G (guanine nucleotide binding) protein; K+ = potassium ions; Kv = voltage-gated potassium ion channel; mRNA = messenger ribonucleic acid; Na-K ATPase = sodium-potassium adenosine triphosphatase; Na+ = sodium ions; NCX = sodium calcium exchanger; PLB = phospholamban; T3 = triiodothyronine; TR = thyroid hormone receptor; TRE = thyroid hormone response element.](nihms-1600817-f0001){#F1}\n\n###### \n\nEffects of Low Triiodothyronine (T3) and on the Cardiovascular system.\n\n  ---------------------------------------------------------------------\n  **Indirect Effect on Cardiovascular Function**\n  Mild hypertension\n  Narrowed pulse pressure\n  High serum concentrations of cholesterol\n  High serum concentrations of creatinine kinase (specifically CK-MM)\n  Pericardial effusions\n  Non-Pitting edema (myxedema)\n  Cardiac arrhythmias (Torsade de pointes)\n  Accelerated atherosclerosis and coronary artery disease\n  Increased carotid intimal thickening\n  Impaired endothelial-dependent vasodilation\n  **Direct Effect on Cardiovascular Function**\n  Increase systemic vascular resistance\n  Decrease heart rate\n  Decrease ejection fraction\n  Decrease cardiac output\n  Increase isovolumetric relaxation time\n  Decrease blood volume\n  ---------------------------------------------------------------------\n"}
+{"text": "1. Introduction {#sec1-insects-10-00319}\n===============\n\nThe cotton--melon aphid, *Aphis gossypii* Glover (Hemiptera: Aphididae), is a worldwide polyphagous insect species, with a large ecological and host range \\[[@B1-insects-10-00319]\\]. *A. gossypii* is an important pest of many agricultural plants, including cotton, melon, potato, chili pepper, sweet pepper, and eggplant \\[[@B2-insects-10-00319],[@B3-insects-10-00319]\\]. It causes significant damage to host plants by sapping, dripping honeydew, and transmitting viruses.\n\n*A. gossypii* exhibits anholocyclic or holocyclic forms of life cycle \\[[@B4-insects-10-00319]\\]. Anholocyclic *A. gossypii* overwinters as nymphs or adults, while holocyclic *A. gossypii* shows a heteroecious or autoecious lifecycle. Heteroecious *A. gossypii* migrates from buckthorn as its primary host to herbaceous plants as its secondary host, where it reproduces asexually to produce numerous offspring in the spring. It then returns to the primary host in the fall to lay overwintering eggs \\[[@B4-insects-10-00319]\\]. On the contrary, autoecious *A. gossypii* does not require any secondary host. Only holocylic *A. gossypii* was found in several countries including Korea where harsh winter is common \\[[@B5-insects-10-00319]\\]. In Korea, *A. gossypii* hatches from eggs on its primary host in the beginning of April and reproduces for two to three generations before adults (apterous viviparous females) migrate to secondary hosts from May to June \\[[@B6-insects-10-00319],[@B7-insects-10-00319]\\].\n\nUnderstanding population genetic structure of *A. gossypii* could help to manage aphid populations by providing more reliable estimates of population dynamics and the risk of resistance genes arising \\[[@B8-insects-10-00319]\\]. The genetic structure of *A. gossypii* populations is influenced by various factors such as host plants, geographical barriers, insecticides, and dispersal ability \\[[@B9-insects-10-00319],[@B10-insects-10-00319]\\]. Significant genetic differentiation and different population structure are observed for most aphid species because their gene flow is limited among populations due to their weak flying ability and reproductive characteristics \\[[@B11-insects-10-00319]\\]. Aphid is a parthenogenetic species with a high clonal diversity, i.e., a rapid change in time with respect to the genotypes \\[[@B12-insects-10-00319],[@B13-insects-10-00319]\\]. Moreover, the application of insecticides has potential to maintain clones with different levels of resistance in population, and evolution of resistance may have the possibility of dramatic shifts in clonal frequencies. However, information regarding the genetic diversity of *A. gossypii* in Korea is currently unavailable. In Korea, *A. gossypii* is an important insect pest of pepper in greenhouses because greenhouses provide sustained warm temperature. Since greenhouses are a relatively closed environment, pest populations in greenhouses are generally more affected by chemical control and host plant changes than populations in fields. These practices may lead to reductions in pest population size and selection for resistant genotypes in populations, resulting in increased homozygosity within populations and differentiation between pest populations \\[[@B14-insects-10-00319]\\]. In addition, population dynamics may account for genetic variation of pest populations in greenhouses. Rochat et al. \\[[@B15-insects-10-00319]\\] revealed that a population of *A. gossypii* infesting cucurbit in greenhouses experiences extreme demographic fluctuations and strong founder effects through a small number of winged individuals immigrating from the exterior. Founder effects are common in aphid populations because of their high rate of population increase due to high fecundity and overlapping generations \\[[@B16-insects-10-00319]\\]. They have the potential to rapidly colonize surrounding plants, leading to very high-density infestations and inter-clonal competition \\[[@B15-insects-10-00319]\\]. Aphid populations in greenhouses may also experience local extinctions and serious bottlenecks due to plant resource exhaustion or insecticide treatments. Despite the importance of information on genetic variation, details on the genetic structure of *A. gossypii* in Korea are scarce.\n\nThus, the purpose of this study was to determine genetic diversity and geneflow of *A. gossypii* on greenhouse peppers (and a few field peppers) in Korea using microsatellite markers. We then discuss population genetic structure for this species based on the results, which may provide a practical framework for developing appropriate management strategies against *A. gossypii*.\n\n2. Materials and Methods {#sec2-insects-10-00319}\n========================\n\n2.1. Insect Samples {#sec2dot1-insects-10-00319}\n-------------------\n\n*A. gossypii* was collected in 18 populations from 2016, 14 populations in 2017, and five populations in 2018 in Korea ([Table S1, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). Most samples were collected from greenhouse peppers in summer season (late May to early August). However, GJ, CJ, YC, and GwJ were collected from field peppers in 2017 due to a change of cultivated crop or no occurrence of *A. gossypii*. All samples were placed in vials containing 95% ethanol and stored at \u221220 \u00b0C until DNA extraction.\n\n2.2. Microsatellite Genotyping {#sec2dot2-insects-10-00319}\n------------------------------\n\nDNA was extracted from individuals using a Qiagen Gentra Puregen Tissue Kit (Qiagen, MD, USA). A total of 1420 aphids were genotyped for eight microsatellite loci as described by Vanlerberghe-Masutti et al. \\[[@B17-insects-10-00319]\\]. The forward primer of each microsatellite locus was labeled with a fluorescent dye (FAM, NED, PET, VIC). Different dyes were chosen for loci having the same allele size to be analyzed simultaneously (Ago24-FAM, Ago53-VIC, Ago59-NED, Ago66-VIC, Ago69-NED, Ago84-PET, Ago89-PET, and Ago126-FAM). Polymerase chain reaction (PCR) was performed in two separate multiplex groups \\[[@B12-insects-10-00319]\\]. The first PCR, using primers specific to seven loci (Ago53, Ago59, Ago66, Ago69, Ago84, Ago89, and Ago126), was performed in a final volume of 10 \u00b5L containing 3.1 \u00b5L of distilled water, 1.0 \u00b5L of 10\u00d7 PCR buffer, 1.0 \u00b5L of 10 mM dNTP mixture, 0.2 \u00b5L of each primer (final concentration, 10 pmol/\u00b5L), 0.1 \u00b5L of Taq polymerase (Takara Taq \u2122, Tokyo, Japan), and 2.0 \u00b5L of template DNA. Amplifications were performed in a thermocycler with the following parameters: initial denaturation at 95 \u00b0C for 15 min; 25 cycles of 30 s at 95 \u00b0C, 90 s at 56 \u00b0C, and 30 s at 72 \u00b0C; and a final extension for 30 min at 60 \u00b0C. The second PCR using primers specific to the eighth locus of Ago24 was performed in the same conditions except for distilled water volume (5.5 \u00b5L) with the following parameters: 5 min at 95 \u00b0C; 35 cycles of 30 s at 95 \u00b0C, 45 s at 62 \u00b0C, and 30 s at 72 \u00b0C; and a final elongation of 7 min at 72 \u00b0C. Then, 1 \u00b5L of each of these two PCR products was mixed with 8.5 \u00b5L of Hi-Di Formamide (Applied Biosystems, Foster City, USA) for denaturing and 0.5 \u00b5L of GeneScan\u2122 500 ROX\u2122 Size Standard (Applied Biosystems, Foster City, USA). PCR products were separated and detected by capillary electrophoresis with an ABI 3730xl automatic sequencer (Applied Biosystems, Foster, USA) using the GENESCAN-500 \\[Rox\\] size standard. The genotype data were analyzed using GeneMapper 3.7 (Applied Biosystems, Foster City, USA).\n\n2.3. Genetic Variation and Genetic Structure {#sec2dot3-insects-10-00319}\n--------------------------------------------\n\nMicrosatellite genotype data of *A. gossypii* were analyzed with Micro-Checker \\[[@B18-insects-10-00319]\\] for the existence of null alleles, scoring errors, and large allele dropout. The confidence interval for Monte Carlo simulations of homozygote frequencies was set to 95%.\n\nBasic parameters were analyzed to measure genetic diversity in *A. gossypii* population. An exact test for Hardy--Weinberg equilibrium (HWE) was conducted per locus and over all loci in each population using Genepop v. 4.2.1 \\[[@B19-insects-10-00319]\\]. Significance was tested using the Markov chain method (10,000 dememorizations, 100 batches, and 5000 iterations per batch). We used Poppr package \\[[@B20-insects-10-00319]\\] of R software version 3.5.0 to identify the multilocus genotypes \\[[@B21-insects-10-00319]\\]. Number of alleles (*N*~A~), observed heterozygosity (*H*~O~), expected heterozygosity (*H*~E~), and inbreeding coefficient (*F*~IS~) were calculated using GenAlEx version 6.5 \\[[@B22-insects-10-00319]\\]. Allelic richness (*A*~R~) was estimated using FSTAT version 2.9.3.2 \\[[@B23-insects-10-00319]\\]. *N*~A~, *H*~O~, *H*~E~, polymorphic information content (PIC), and measures of genetic diversity for each locus and averaged across loci were calculated with CERVUS software \\[[@B24-insects-10-00319]\\]. An *F*~ST~ value of zero implies a lack of divergence between populations, while an *F*~ST~ of one implies complete isolation of the population. Index of pairwise *F*~ST~ of Weir and Cockerham \\[[@B25-insects-10-00319]\\] between population and their associate 95% confidence intervals were estimated using FSTAT software. The ENA (excluding null alleles) method was also used to verify unbiased pairwise *F*~ST~ values (*F*~ST~^\\[ENA\\]^) using the FreeNA program \\[[@B26-insects-10-00319]\\].\n\nBottleneck events in *A. gossypii* populations were tested with BOTTLENECK program version 1.2 \\[[@B27-insects-10-00319]\\] using a two-phase model (TPM) and stepwise mutation model (SMM). We excluded the infinite alleles model (IAM) because IAM was not appropriate for microsatellites due to high microsatellite mutation rates and mutation processes that might retain memory of ancestral allelic states \\[[@B28-insects-10-00319],[@B29-insects-10-00319]\\]. The SMM model can predict all mutations corresponding to increment or decrement of a single base-pair repeat. The TPM model observes the occurrence of an occasional multiple base-pair repeat \\[[@B30-insects-10-00319]\\]. It is suggested that TPM can closely simulate microsatellite mutation \\[[@B31-insects-10-00319]\\]. Thus, we used both TPM and SMM models. They are widely adopted for use with microsatellite markers \\[[@B32-insects-10-00319],[@B33-insects-10-00319]\\]. Parameters chosen for TPM were as follows: variance = 30.00, probability = 70.00%, and estimations based on 10,000 iterations. Deviations from equilibrium were examined using Wilcoxon signed-rank test with significance level *p* \\< 0.05. The Wilcoxon signed-rank test is efficient and reliable when eight microsatellite loci are analyzed \\[[@B31-insects-10-00319]\\]. The genetic bottleneck test was reconfirmed through a mode shift indicator test based on a qualitative descriptor of allele frequency distribution.\n\nHierarchical analysis of molecular variance (AMOVA) was performed using GenAlEx that partitioned genetic variation among populations and individuals within populations. Significance of AMOVA analysis was estimated using 10,000 permutations. Isolation by distance (IBD) was analyzed by regressing pairwise population estimates of linearized *F*~ST~/ (1 \u2212 *F*~ST~) \\[[@B34-insects-10-00319]\\] on a natural log of the geographical distance between all pairs of sample location. Mantel's test was implemented with 9999 permutation using ADEGENET package \\[[@B35-insects-10-00319]\\] of R software version 3.5.0 \\[[@B21-insects-10-00319]\\].\n\nPopulation structure of *A. gossypii* was calculated by the Bayesian clustering procedure using STRUCTURE 2.3.3 \\[[@B36-insects-10-00319]\\] to explore different numbers of populations *K* to the population structure based on microsatellite data. Ten replicate runs with 600,000 MCMC (Markov chain Monte Carlo) iterations and a burn-in of 60,000 steps were performed for 1 \u2264 *K* \u2264 10 (*K* = number of clusters) to verify the consistency of estimates and determine the most likely number of genetic clusters. The optimal value of *K* was determined using STRUCTURE HARVESTER \\[[@B37-insects-10-00319]\\] to compute \u0394*K* \\[[@B38-insects-10-00319]\\]. Clustering pattern was applied to the CLUMPP program and visualized using DISTRUCT software version 1.1 \\[[@B39-insects-10-00319]\\]. Population structure was estimated using principal coordinate analysis (PCoA) applied in GenAlex, and then scatter diagram was plotted based on factor scores along the two PCoAs revealing the most variations. PCoA was based on the covariance of the genetic distance matrix, and analysis was implemented separately for each year. Discriminant analysis of principal component (DAPC) was carried out using ADEGENET package \\[[@B35-insects-10-00319]\\] of statistical package R software version 3.5.0 \\[[@B21-insects-10-00319]\\].\n\n3. Results {#sec3-insects-10-00319}\n==========\n\nIn this study, a total of 57 alleles were verified across the eight microsatellite loci for 1420 *A. gossypii* individuals from 37 locations in Korea (18 populations in 2016, 14 populations in 2017, and five populations in 2018). Fisher's exact tests showed that 291 of 1036 locus/population combinations deviated significantly from Hardy--Weinberg equilibrium (HWE). The number of alleles (*N*~A~), allelic richness (*A*~R~), observed heterozygosity (*H*~O~), expected heterozygosity (*H*~E~), and inbreeding coefficient (*F*~IS~) of populations are presented in [Table 1](#insects-10-00319-t001){ref-type=\"table\"}. *N*~A~ ranged from 2.750 (JiJ_16) to 6.375 (JE_16), with an average of 4.091 across populations. *A*~R~ ranged from 2.707 (JiJ_16) to 5.949 (BS_16), with an average of 3.947. The lowest *H*~O~ was detected in JJ_17 (0.466), while the highest was detected in BS_16 (0.878). The lowest *H*~E~ was observed in GwJ_16 (0.354), while BS_16 showed the highest *H*~E~ (0.719). Inbreeding coefficient (*F*~IS~) ranged from \u22120.810 (JiJ_16) to 0.128 (JJ_17), with an average of \u22120.256 across populations. The presence of potential null alleles was indicated by a general excess of homozygotes for most allele size classes for no or one loci within at each population. The most polymorphic marker with the highest *N*~A~ per locus was 10 (Ago66), and the mean of *N*~A~ per locus was 7.125. Polymorphic information content (PIC) per locus ranged from 0.425 (Ago53) to 0.760 (Ago66). PIC values showed that all markers were informative ([Table S2, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}).\n\nGenetic structures of 37 geographic populations in 2016 (18 populations), 2017 (14 populations), and 2018 (five populations) were analyzed using pairwise comparisons of multilocus *F*~ST~ with or without ENA correction (*F*~ST~ ^\\[ENA\\]^) ([Tables S3--S5, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). Pairwise *F*~ST~ values between the populations in 2016 ranged from 0.0047 for GJ_16 and GS_16 populations (*F*~ST~^\\[ENA\\]^ = 0.0093; GJ_16 and GS_16 populations) to 0.4396 for YC_16 and GwJ_16 populations (*F*~ST~^\\[ENA\\]^ = 0.4287; YC_16 and GwJ_16 populations) ([Table S3, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). Pairwise *F*~ST~ values between populations in 2017 ranged from 0.0080 for AD_17 and GwJ_17 populations (*F*~ST~^\\[ENA\\]^ = 0.0109; AD_17 and GwJ_17 populations) to 0.3819 for CY_17 and CJu_17 populations (*F*~ST~^\\[ENA\\]^ = 0.3718; CY_17 and CJu_17 populations) ([Table S4, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). Pairwise *F*~ST~ values between populations in 2018 ranged from 0.0033 for CJu_18 and BS_18 populations (*F*~ST~^\\[ENA\\]^ = 0.0043; CJu_18 and BS_18 populations) to 0.2022 for CJu_18 and JJ_18 populations (*F*~ST~^\\[ENA\\]^ = 0.2008; CJu_18 and JJ_18 populations) ([Table S5, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). *F*~ST~ adjusted for null alleles and *F~ST~* assuming no null allele results were similar to each other. Overall *F*~ST~ value (uncorrected *F~ST~* = 0.1888, 95% confidence interval (CI): 0.1222--0.2532; ENA corrected *F*~ST~ = 0.1794, 95% confidence interval (CI): 0.1166--0.2411) indicated a high level of genetic differentiation among geographic populations.\n\nBased on bottleneck analysis, significant heterozygote excess was shown in 10 of 37 populations under the two-phase model (TPM) and three of 37 populations using the stepwise mutation model (SMM). Accordingly, there was potential evidence for recent population reductions in a few populations. Most populations had a normal L-shaped distribution (except for YC_16, JiJ_16, and HS_18), indicating that *A. gossypii* expanded spatially without severe bottleneck in most regions of Korea. Population YC_16, JiJ_16, and HS_18 showed a significant bottleneck effect at *p* = 0.05 and a shifted shape (S) allele frequency distribution, providing evidence for recent population reduction in these populations ([Table 2](#insects-10-00319-t002){ref-type=\"table\"}).\n\nAnalysis of molecular variance (AMOVA) for *A. gossypii* populations revealed a high variance component within individuals (85%), followed by among populations (15%) ([Table 3](#insects-10-00319-t003){ref-type=\"table\"}). The component of variance among populations was significant (*F*~ST~). Based on the Mantel test for IBD, no significant correlation was found between genetic and geographical distances among populations (*r^2^* = 0.0004, *p* = 0.370), suggesting high gene flow among populations in Korea ([Figure 1](#insects-10-00319-f001){ref-type=\"fig\"}).\n\nResults of Bayesian analysis of population genetic structure indicated that the best dataset partitioning involved two genetic clusters since the value of \u0394*K* (Evanno method) occurred at *K* = 2 with a maximum value of 1251.3 ([Figure S1, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}). Using partition *K* = 2, graphics were drawn with DISTRUCT to visualize the clustering pattern of individuals and populations ([Figure 2](#insects-10-00319-f002){ref-type=\"fig\"}). In addition, results of Bayesian cluster analysis of multilocus microsatellite genotypes in 2016, 2017, and 2018 are displayed as pie graphs ([Figure 3](#insects-10-00319-f003){ref-type=\"fig\"}). Opposite genetic clusters appeared between 2016 and 2017, mostly in northwest (HS, CJu) and southeast (MY, BS, JiJ) parts of Korea, although samples were collected from the same pepper greenhouse. HS population showed an opposite pie graph compared to 2016 and 2017. However, the cluster was divided in half in 2018. JE and BS populations showed similar pie graphs in 2017, while CJu showed the process of changing to opposite cluster.\n\nPrincipal coordinate analysis (PCoA) showed that the pattern of genetic structure was similar to the genetic cluster analysis result (two distinct groups) ([Figure 4](#insects-10-00319-f004){ref-type=\"fig\"}). Total variance explained by the first and second axes in 2016, 2017, and 2018 was 56% (30.39% for axis 1 and 25.11% for axis 2). Discriminant analysis of principal component (DAPC) results were highly similar to PCoA based on the scatter plot of the population. However, most populations were at the gravity center because genetic variation between populations was not large enough to be divided into two distinct groups. ([Figure S2, Supplementary Materials](#app1-insects-10-00319){ref-type=\"app\"}).\n\n4. Discussion {#sec4-insects-10-00319}\n=============\n\nThis study is the first attempt to understand the pattern of genetic variability in *A. gossypii* in Korea. In the present study, we analyzed genetic diversity within and between *A. gossypii* populations collected from pepper plants (mostly in greenhouses). Genetic variability found in *A. gossypii* through the use of molecular markers revealed that its clonal diversity is structured by its host plants \\[[@B17-insects-10-00319]\\]. In the present study, *A. gossypii* had low to moderate genetic diversity based on microsatellite data ([Table 1](#insects-10-00319-t001){ref-type=\"table\"}), in which *H*~E~ varied from 0.354 (GwJ_16) to 0.719 (BS_16). Most *H*~E~ values were lower than the average (average = 0.524). A similar level of genetic variation was shown in other aphid species (*Sitobion avenae*), for which H*~E~* ranged from 0.409 to 0.873 on the basis of eight microsatellite loci \\[[@B40-insects-10-00319]\\]. Within a greenhouse population of *A. gossypii*, clonal diversity declined significantly as spring/summer season progressed \\[[@B16-insects-10-00319]\\]. Similarly, in *S. avenae*, genetic variation significantly decreased from spring to summer \\[[@B41-insects-10-00319]\\]. Br\u00e9vault et al. \\[[@B42-insects-10-00319]\\] revealed that *A. gossypii* populations, collected from cotton crops, vegetable crops, and weeds in northern Cameroon, show very low genetic diversity (only 11 multilocus genotypes identified). The final predominance of the clone may occur through a combination of genetic drift related to population foundation, demographic explosion, and/or clonal competition. The final fittest founder genotype will become dominant during the period of rapid population growth, while recessive genotypes will decrease in frequency or may extinct completely. Moreover, *Eriosoma lanigerum* and *Myzus persicae* show low levels of genetic diversity due to adaptation of aphids to heavy selection pressures, including distribution of host plants and the use of insecticides \\[[@B43-insects-10-00319],[@B44-insects-10-00319]\\]. Cyclical parthenogenesis may explain the low level of genetic variability detected in aphids compared to other insects. It could be an important factor that leads to local or temporal genetic differentiation of populations \\[[@B45-insects-10-00319]\\]. In our study, most *A. gossypii* samples were collected from greenhouse peppers in the summer, during which most *A. gossypii* were under cyclical parthenogenesis. Chemical control was common in all greenhouses. This apparently led to heavy selective pressure for *A. gossypii*. These points may explain such low levels of genetic variability of *A. gossypii* observed in Korea. Moreover, *A. gossypii* developed resistance to insecticides such as carbamates and organophosphates. Its exponential growth due to parthenogenesis favors rapid selection for insecticide resistance traits \\[[@B46-insects-10-00319],[@B47-insects-10-00319]\\]. Therefore, it would be important to understand the genetic structure and level of gene flow of *A. gossypii* populations, to estimate insecticide resistance over temporal and spatial scales and investigate distribution of resistance genes among *A. gossypii* populations because resistance genes are related to genetic factors.\n\nThe moderate flight and dispersal ability of aphids might allow short-distance movement. Usually, the IBD test is widely used to examine spatial patterns of gene flow and genetic relatedness between populations \\[[@B48-insects-10-00319]\\]. Results of IBD analysis revealed that geographic distance had no effect on *A. gossypii* population structure in Korea ([Figure 1](#insects-10-00319-f001){ref-type=\"fig\"}). In the present study, the lack of a significant IBD pattern may indicate unrestricted gene flow among *A. gossypii* populations in Korea. Genetic structure analysis based on STRUCTURE and PCoA revealed two genetic clusters in *A. gossypii* populations in Korea ([Figure 2](#insects-10-00319-f002){ref-type=\"fig\"} and [Figure 3](#insects-10-00319-f003){ref-type=\"fig\"}). Interestingly, different genetic clusters were found between 2016 and 2017, mostly in northwest (HS, CJu, IS) and southeast (MY, BS, JiJ) parts of Korea, although samples were collected from the same greenhouses. Interestingly, Jeju (JJ) population showed similar genetic clusters during three years. Such similar genetic structures might be due to limited immigration as growers raise pepper seedlings by themselves and Jeju is an island. Within a greenhouse, clonal composition does not persist over time. These changes might result from local extinction (due to insecticide, crop change, cold winter temperature, etc.), followed by recolonization by winged immigrants coming from neighboring greenhouses, from refuges, or from raising of seedlings. In addition, different clusters observed over the years can be attributed to a founder effect. A greenhouse is a comparatively enclosed space. Once it is founded, migration to the outside would be limited and vice versa. A few new immigrants that succeed in dispersing into a greenhouse after insecticide treatment for control might have a significant effect on the genetic structure of *A. gossypii*. Moreover, the population structure of A. *gossypii* might change dramatically due to bottlenecks and rapid gene flow, because several locations (HS, CJu, BS, JiJ) showed a significant *p*-value in the TPM model. Moreover, a change in genetic cluster in a period of one year might be caused by different fitness between two genetic clusters of *A. gossypii* in Korea. Two genetic clusters of *A. gossypii* might have coexisted in the same regions, and one genetic cluster could be a dominant genotype if there is a fitness difference between them.\n\n5. Conclusions {#sec5-insects-10-00319}\n==============\n\nGenetic diversity was revealed among *A. gossypii* populations in Korea based on eight microsatellite loci. *A. gossypii* populations in Korea appeared to be classified into two genetic clusters. However, its genetic structure rapidly changed into opposite clusters in several regions, although samples were collected from the same locations. Based on the results of bottleneck analysis (TPM, SMM model), most *A. gossypii* populations experienced a recent spatial expansion without any severe bottleneck in most regions of Korea. However, several locations (HS, CJu, BS, JiJ) were shown to be significant in TPM, which may affect the rapid turnover in genetic structure. These results provide important information for understanding its feasible local adaptation and dispersal patterns. *A. gossypii* populations in pepper-growing (field and greenhouse) areas in Korea showed low genetic diversity and high gene flow due to cyclical parthenogenesis and heavy insecticidal selection pressure. Thus, future work should elucidate any phenotypic fitness difference between the two genetic clusters of *A. gossypii* in Korea. Also, further studies should focus on the relationship between the genetic structure of *A. gossypii* populations in various crop-producing areas and the extent of insecticide selection pressure.\n\nThe following are available online at <https://www.mdpi.com/2075-4450/10/10/319/s1>: Figure S1: Bayesian inference to identify suitable cluster (K) and cluster proportion using STRUCUTRE for *Aphis gossypii* in Korea. Delta K are analyzed against number of genetic clusters (K); Figure S2: Scatter plot of DAPC analysis of the nine populations using \"adegent\" in R package ([Table 2](#insects-10-00319-t002){ref-type=\"table\"} indicates the population ID). Dots: individuals, ellipses: populations; Table S1: Sampling information of *A. gossypii* collected in Korea (2016, 2017, and 2018); Table S2: Microsatellite loci, number of alleles observed at each locus (NA), observed (HO) and expected (HE) heterozygosity at each locus, and mean PIC (polymorphic information content) per locus in *A. gossypii* population; Table S3: Pairwise FST\\[ENA\\] values (lower-left matrix), and pairwise FST values and significance (upper-right matrix) based on eight microsatellite loci between the populations *of A. gossypii* in Korea (2016); Table S4: Pairwise FST\\[ENA\\] values (lower-left matrix), and pairwise FST values and significance (upper-right matrix) based on eight microsatellite loci between the populations of *A. gossypii* in Korea (2017); Table S5: Pairwise FST\\[ENA\\] values (lower-left matrix), and pairwise FST values and significance (upper-right matrix) based on eight microsatellite loci between the populations of *A. gossypii* in Korea (2018).\n\n###### \n\nClick here for additional data file.\n\nConceptualization, H.Y.N. and J.-H.L.; methodology, H.Y.N.; software, H.Y.N.; validation, J.-H.L.; formal analysis, H.Y.N.; investigation, H.Y.N. and Y.P.; resources, H.Y.N. and Y.P.; data curation, H.Y.N.; writing---original draft preparation, H.Y.N. and J.-H.L.; writing---review and editing, H.Y.N. and J.-H.L.; visualization, H.Y.N. and J.-H.L.; supervision, J.-H.L.\n\nThis research was funded by a grant from the Rural Development Administration in Korea (PJ01194804) and was supported by the Brain Korea 21 Plus.\n\nThe authors declare no conflict of interest.\n\n![Geographical distance versus genetic distance (*F*~ST~/1 \u2212 *F*~ST~) for populations of *Aphis gossypii*, using pairwise *F*~ST~. Correlations and probabilities were estimated from a Mantel test with 9999 bootstrap repeats.](insects-10-00319-g001){#insects-10-00319-f001}\n\n![Structure profile under *K* = 2, permuted in CLUMPP, plotted with DISTRUCT on 37 *A. gossypii* populations, depicting classifications with the highest probability under the model that assumes independent allele frequencies and inbreeding coefficients among assumed clusters. Each individual is represented by a vertical bar, often partitioned into colored segments with the length of each segment representing the proportion of the individual's genome from *K* = 2 ancestral populations. On the bottom of the plot, the name of population localities is indicated and the year of sampling is shown in parentheses.](insects-10-00319-g002){#insects-10-00319-f002}\n\n![All populations partitioned in two clusters (*K* = 2) and the pie graphs revealing the results from a Bayesian cluster analysis of multilocus genotypes in 2016, 2017, and 2018. The population identifiers (IDs) are indicated in pie graphs (\\* = samples collected from field pepper).](insects-10-00319-g003){#insects-10-00319-f003}\n\n![Scatter diagram of factor scores from a principal coordinate analysis of genotype data for eight microsatellite loci in samples of *A. gossypii* collected from 37 locations in Korea (2016, 2017, and 2018). The percentage of total variation attributed to each axis is indicated.](insects-10-00319-g004){#insects-10-00319-f004}\n\ninsects-10-00319-t001_Table 1\n\n###### \n\nGenetic variation estimates of geographic population of *Aphis gossypii*. Number of alleles (*N*~A~), allelic richness (*A*~R~), observed heterozygosity (*H*~O~), expected heterozygosity (*H*~E~), inbreeding coefficient (*F*~IS~), probability (*p*-value) of being in Hardy--Weinberg equilibrium (HWE), and loci showing potential null alleles. ID---identifier.\n\n  Sampling Name   Population ID   Sample Size   N*~A~*   *A* ~R~   *H* ~O~   *H* ~E~   *p*-Value   *F* ~IS~ ^1^    Loci with Null Alleles\n  --------------- --------------- ------------- -------- --------- --------- --------- ----------- --------------- -----------------------------\n  PT              PT_16           40            4.250    4.124     0.759     0.575     0.00002     \u22120.257 \\*\\*\\*   Ago53\n  DJ              DJ_16           40            3.250    3.172     0.744     0.502     0.00002     \u22120.438 \\*\\*\\*   No\n  HS              HS_16           40            5.250    5.080     0.741     0.623     0.00002     \u22120.172 \\*\\*\\*   Ago69\n  GJ              GJ_16           40            3.000    2.735     0.616     0.404     0.00002     \u22120.322 \\*\\*\\*   Ago66, Ago126\n  CJu             CJu_16          40            3.500    3.361     0.828     0.510     0.00002     \u22120.548 \\*\\*\\*   No\n  GS              GS_16           40            3.250    3.014     0.594     0.417     0.00002     \u22120.297 \\*\\*\\*   Ago66\n  CJ              CJ_16           40            5.500    5.266     0.851     0.709     0.00002     \u22120.214 \\*\\*\\*   No\n  YC              YC_16           40            2.875    2.803     0.716     0.442     0.00002     \u22120.554 \\*\\*\\*   No\n  AD              AD_16           40            5.625    5.334     0.750     0.623     0.00002     \u22120.241 \\*\\*\\*   Ago53\n  MY              MY_16           40            4.625    4.429     0.694     0.570     0.00002     \u22120.178 \\*\\*\\*   Ago53, Ago69\n  JiJ             JiJ_16          40            2.750    2.707     0.850     0.472     0.00002     \u22120.810 \\*\\*\\*   No\n  KH              KH_16           40            4.625    4.580     0.766     0.635     0.00002     \u22120.168 \\*\\*\\*   Ago53\n  BS              BS_16           40            6.125    5.949     0.878     0.719     0.00002     \u22120.228 \\*\\*\\*   No\n  IS              IS_16           40            5.000    4.736     0.541     0.516     0.00002     \u22120.025 \\*\\*\\*   Ago53, Ago66, Ago126\n  JE              JE_16           40            6.375    5.919     0.859     0.656     0.00002     \u22120.315 \\*\\*\\*   No\n  GwJ             GwJ_16          40            3.375    3.161     0.544     0.354     0.00002     \u22120.373 \\*\\*\\*   Ago53\n  Bos             BoS_16          40            4.500    4.299     0.850     0.561     0.00002     \u22120.502 \\*\\*\\*   No\n  JJ              JJ_16           40            5.625    5.363     0.687     0.632     0.00002     \u22120.078 \\*\\*\\*   Ago59\n  HS              HS_17           40            3.375    3.271     0.494     0.422     0.00002     \u22120.110 \\*\\*\\*   Ago53, Ago59, Ago66\n  CY              CY_17           30            3.625    3.625     0.513     0.438     0.00002     \u22120.205 \\*\\*\\*   Ago53\n  GJ              GJ_17 \\*        40            3.000    2.883     0.559     0.433     0.00002     -0.166\\*\\*\\*    Ago53, Ago59, Ago66\n  CJu             CJu_17          40            3.000    2.904     0.631     0.469     0.00002     \u22120.277 \\*\\*\\*   Ago59, Ago69\n  CJ              CJ_17 \\*        40            4.875    4.566     0.719     0.574     0.00002     \u22120.217 \\*\\*\\*   Ago53\n  YC              YC_17 \\*        40            3.750    3.639     0.741     0.563     0.00002     \u22120.289 \\*\\*\\*   No\n  AD              AD_17           40            3.250    3.203     0.725     0.511     0.00002     \u22120.332 \\*\\*\\*   Ago53, Ago69\n  MY              MY_17           40            3.750    3.611     0.547     0.428     0.00002     \u22120.048 \\*\\*\\*   Ago53, Ago59, Ago69, Ago126\n  BS              BS_17           40            5.375    5.131     0.691     0.602     0.00002     \u22120.155 \\*\\*\\*   Ago53, Ago59, Ago126\n  JiJ             JiJ_17          40            3.625    3.461     0.656     0.549     0.00002     \u22120.235 \\*\\*\\*   Ago53, Ago66\n  IS              IS_17           40            3.875    3.641     0.616     0.473     0.00002     \u22120.301 \\*\\*\\*   Ago53\n  JE              JE_17           40            3.875    3.708     0.763     0.527     0.00002     \u22120.346 \\*\\*\\*   No\n  GwJ             GwJ_17 \\*       40            4.375    4.131     0.734     0.527     0.00002     \u22120.280 \\*\\*\\*   Ago53\n  JJ              JJ_17           40            4.875    4.620     0.466     0.477     0.00002     0.128 \\*\\*\\*    Ago24, Ago53, Ago59\n  HS              HS_18           30            2.875    2.875     0.608     0.513     0.00002     \u22120.138 \\*\\*\\*   Ago53, Ago59, Ago69\n  CJu             CJu_18          30            3.250    3.250     0.579     0.501     0.00002     \u22120.164 \\*\\*\\*   Ago53, Ago59, Ago69\n  BS              BS_18           30            3.500    3.500     0.550     0.485     0.00002     \u22120.107 \\*\\*\\*   Ago53, Ago59, Ago69\n  JE              JE_18           30            3.250    3.250     0.596     0.443     0.00002     \u22120.322 \\*\\*\\*   Ago59, Ago69\n  JJ              JJ_18           30            4.375    4.375     0.700     0.542     0.00002     \u22120.171 \\*\\*\\*   Ago53, Ago69\n\n\\* *A. gossypii* was collected from field pepper; HW test: Hardy--Weinberg exact test \\[[@B19-insects-10-00319]\\] with Bonferroni correction (*p* = 0.000017). ^1^ Significance *F*~IS~ value was obtained after a 1000-permutation test (\\*\\* *p* \\< 0.05; \\*\\*\\* *p* \\< 0.01).\n\ninsects-10-00319-t002_Table 2\n\n###### \n\nWilcoxon signed-rank test for mutation-drift equilibrium estimated based on eight microsatellite loci.\n\n  Population ID   TPM                  SMM              Mode Shift   Population ID   TPM              SMM              Mode Shift\n  --------------- -------------------- ---------------- ------------ --------------- ---------------- ---------------- ------------\n  PT_16           0.191                0.680            L            CY_17           0.281            0.578            L\n  DJ_16           **0.037 \\*\\*^,2^**   0.156            L            GJ_17 \\*        0.191            0.422            L\n  HS_16           0.371                0.629            L            CJu_17          **0.020 \\*\\***   0.273            L\n  GJ_16           0.422                0.473            L            CJ_17 \\*        0.371            0.680            L\n  CJu_16          0.098                0.273            L            YC_17 \\*        **0.004 \\*\\***   **0.006 \\*\\***   L\n  GS_16           0.289                0.813            L            AD_17           0.125            0.371            L\n  CJ_16           **0.002 \\*\\***       **0.037 \\*\\***   L            MY_17           0.727            0.844            L\n  YC_16           **0.027 \\*\\***       0.188            S            BS_17           0.527            0.809            L\n  AD_16           0.727                0.994            L            JiJ_17          **0.027 \\*\\***   0.098            L\n  MY_16           0.422                0.770            L            IS_17           0.371            0.809            L\n  JiJ_16          **0.004 \\*\\***       **0.020 \\*\\***   S            JE_17           0.098            0.191            L\n  KH_16           **0.027 \\*\\***       0.230            L            GwJ_17 \\*       0.473            0.875            L\n  BS_16           **0.010 \\*\\***       0.473            L            JJ_17           0.973            0.994            L\n  IS_16           0.809                0.990            L            HS_18           **0.014 \\*\\***   0.098            S\n  JE_16           0.629                0.963            L            CJu_18          0.098            0.273            L\n  GwJ_16          0.711                0.813            L            BS_18           0.191            0.527            L\n  BoS_16          0.422                0.727            L            JE_18           0.289            0.594            L\n  JJ_16           0.473                0.963            L            JJ_18           0.320            0.809            L\n  HS_17           0.469                0.469            L                                                              \n\n^2^*p* is test for heterozygosity excess, \\*\\* *p* \\< 0.05; TPM: two-phase model; SMM: stepwise mutation model; L: normal L-shaped distribution S: shifted mode.\n\ninsects-10-00319-t003_Table 3\n\n###### \n\nAnalysis of molecular variance (AMOVA) analysis on eight microsatellites in different populations of *A. gossypii* in Korea (\\*\\*\\* *p* \\< 0.01). df---degrees of freedom\n\n  Source of Variation                    df     Sum of Squares   Mean Sum of Squares   Estimated Variance   \\% of Variation   *F*-Statistics\n  -------------------------------------- ------ ---------------- --------------------- -------------------- ----------------- ------------------------\n  Among populations                      36     1434.821         39.856                0.500                15%               *F*~ST~ = 0.190 \\*\\*\\*\n  Among individuals within populations   1383   2099.471         1.518                 0.000                0%                *F*~IS~ = \u22120.286\n  Within individuals                     1420   3879.000         2.732                 2.732                85%               *F*~IT~ = \u22120.041\n  Total                                  2839   7413.292                               3.231                100%              \n"}
+{"text": "Introduction {#Sec1}\n============\n\nLand ecosystems presently absorb around 25% of the carbon annually emitted by anthropogenic activities as carbon dioxide (CO~2~) and thus beneficially slow down the trend of increasing atmospheric CO~2~ concentrations and the associated global warming (IPCC [@CR19]; Le Qu\u00e9r\u00e9 et al. [@CR26]; Raupach et al. [@CR33]; Schimel et al. [@CR39]). To project whether land ecosystems will continue to remove CO~2~ from the atmosphere at this rate (Ballantyne et al. [@CR4]), robust models of the carbon cycle are needed. Because the net ecosystem exchange of CO~2~ is the small difference of two large fluxes of opposing sign, gross primary production (GPP) and ecosystem respiration (ER), models typically simulate these two component fluxes separately (Cramer et al. [@CR15]). For model calibration and validation, ecosystem-scale observations of GPP and ER are critical; however, these are characterized by large uncertainties (Wohlfahrt and Gu [@CR50]).\n\nA promising new approach is the use of carbonyl sulfide (COS) flux measurements to estimate photosynthesis (Asaf et al. [@CR3]; Berry et al. [@CR8]; Wohlfahrt et al. [@CR51]). COS and CO~2~ have similar diffusion pathways into leaves (Seibt et al. [@CR40]; Stimler et al. [@CR41]) and are processed by the same enzyme---carbonic anhydrase. However, in contrast to CO~2~, COS is not emitted by plants and could, therefore, be used to estimate the gross CO~2~ uptake by the vegetation.\n\nCarbonyl sulfide is the most abundant sulfur-containing trace gas in the atmosphere, with a mean concentration of about 500 pptv in the troposphere. It is transported to the stratosphere, where it contributes to the formation of sulfur aerosols via photolysis and oxidation, leading to a stratospheric lifetime of 68\u00a0\u00b1\u00a020\u00a0years at polar latitudes and 58\u00a0\u00b1\u00a014\u00a0years at tropical latitudes (Krysztofiak et al. [@CR22]). COS sources are direct emission from oceans, oxidation of CS~2~ (Watts [@CR46]), biomass burning (Notholt et al. [@CR30]), anthropogenic activities (Campbell et al. [@CR13]), wetlands, and anoxic soils (Kettle et al. [@CR21]; Whelan et al. [@CR48]). Reaction with OH in the troposphere, photolysis and oxidation in the stratosphere, and uptake by vegetation and oxic soils are assumed to be the major sinks for COS (Berry et al. [@CR8]; Kettle et al. [@CR21]; Launois et al. [@CR23], [@CR24]).\n\nTo use COS as a proxy for the CO~2~ uptake by plants, a quantification of all other sink and source terms within an ecosystem is necessary (Wohlfahrt et al. [@CR51]). Knowledge about the contribution of soils, listed as both sources and sinks in current COS budgets, is especially scarce. Both COS producing and consuming processes in soils are partially known (Conrad [@CR14]). The hydrolysis catalyzed by carbonic anhydrase, which is present in microorganisms, was of particular interest in the past and was tested in experiments and found to be substantial (Kesselmeier et al. [@CR20]). Van Diest and Kesselmeier ([@CR44]) found in a lab study a dependency of the COS flux on the soil water content (SWC) in four different soils, with a higher uptake at a soil water content (SWC) between 9 and 11.5%, with a considerable joint influence of temperature. The COS fluxes in that study exhibited an optimum curve depending on SWC and temperature suggesting biotic processes to be dominant. Abiotic processes would more likely lead to a monotonic increase, especially with temperature. In a recent comparison of soil COS exchange from several contrasting biomes, Whelan et al. ([@CR49]) confirmed generally small COS uptake rates, except for an agricultural soil, which, in accordance with Billesbach et al. ([@CR9]) and Maseyk et al. ([@CR28]), showed large emissions.\n\nRecent lab experiments (Whelan and Rhew [@CR47]) pointed in a different direction focusing on abiotic processes by comparing dead versus living soil samples from an agricultural study site and came to the conclusion that the magnitude of abiotic reactions, driven by temperature and radiation, outweighs those of biotic ones. To date, it is not clear whether these observed differences are reflective of particular soil characteristics, e.g., associated with natural vs. agricultural soils, experimental setups, and methods or reflect the true continuum of soil COS exchange which ranges from significant uptake to emissions governed by an interplay between biotic and abiotic processes.\n\nThe overarching aim of this study was to identify the main drivers of in situ COS soil fluxes at temperate mountain grassland in support of a related ongoing study investigating on the potential of COS to be used as a tracer of canopy photosynthesis. Based on the previous laboratory (Kesselmeier et al. [@CR20]; Van Diest and Kesselmeier [@CR44]) and current modelling studies (Ogee et al. [@CR31]; Sun et al. [@CR42]), we hypothesized that soil water content would have a decisive influence on the soil COS exchange and we thus compared the soil COS exchange between a rain exclusion treatment and a well-watered control treatment.\n\nMaterials and methods {#Sec2}\n=====================\n\nThe study site (47\u00b07\u2032N, 11\u00b018\u2032E) is located near Neustift (Austria) in the Central Alps at an elevation of 990\u00a0m above sea level. The climate is temperate with alpine influences; the average annual temperature is 6.5\u00a0\u00b0C; the average annual rainfall amounts to 852\u00a0mm. The soil was classified as a Fluvisol with an estimated depth of 1\u00a0m; the bulk of the roots was located within the first 10\u00a0cm. The organic volume fraction of the A horizon is approx. 14%. Soil water content at field capacity (matrix potential of \u221210\u00a0kPa) was 0.48\u00a0m^3^m^\u22123^ and at the wilting point (matrix potential of \u22121500\u00a0kPa) 0.02\u00a0m^3^m^\u22123^ (Brilli et al. [@CR11]). H\u00f6rtnagl et al. ([@CR18]) described the vegetation as a Pastinaco-Arrhenatheretum dominated by *Dactylis glomerata, Festuca pratensis, Alopecurus pratensis, Trisetum flavescens, Ranunculus acris, Taraxacum officinalis, Trifolium repens, Trifolium pretense,* and *Carum carvi*. The site is managed as a hay meadow, with harvests typically occurring at the beginning of June, beginning of August and the end of September. Organic fertilizer is applied typically in the form of manure during late October.\n\nThe measurement period stretched from 10-Jun-2015 to 13-Aug-2015 with 6 measurement days. The first day was prior to the simulated drought and the last day after the rainout shelters had been removed. The duration of the first 3 measurement days was 24\u00a0h; afterwards, the duration of the daily measurements varied from 7 to 12\u00a0h during daytime. Two rainout shelters were constructed with visible light- and UV-permeable foil (Lumisol clear AF; folitec Agrarfolien Vertriebs GmbH, Westerburg, Germany; visible light permeability: approx. 88--90%, UV-A permeable, UV-B permeability: \\>70%). Both shelters covered an area of 5.28\u00a0m^2^; the foil was removed during the measurements. In one shelter, rain was excluded for the whole treatment period (11-Jun-2015--05-Aug-2015; 55\u00a0days), the other one was used as a control and watered according to the mean precipitation of the period 1971--2014 (178.22\u00a0mm across 55\u00a0days). In each shelter, three stainless steel (SAE grade: 316L) rings were inserted 5\u00a0cm into the soil, and remained there for the whole experiment. The aboveground biomass was removed within the rings one day prior to each measurement day; the surrounding vegetation within the shelters was allowed to grow and was not cut. Roots inside the rings were not removed and natural litter was left in place.\n\nCOS and CO~2~ concentrations were measured with a Quantum Cascade Laser Mini Monitor (Aerodyne Research, Billerica, MA, USA) at a wavenumber of ca. 2056\u00a0cm^\u22121^. The QCL was operated at a pressure of 20\u00a0Torr using a built-in pressure controller and temperature of the optical bench and housing controlled to 35\u00a0\u00b0C. Fitting of absorption spectra at 1\u00a0Hz, storing of calculated mole fractions, switching of zero/calibration valves, control of pressure lock, and other system controls were done by the TDLWintel software (Aerodyne Research, Billerica, MA, USA). The QCL was housed in an air-conditioned trailer next to the study site and regularly calibrated against working standards cross-referenced against a standard from NOAA (557 ppt COS in air). A hand-held sensor (WET-2, Delta-T Devices, Cambridge, England) was used to measure soil water content (SWC) and soil temperature at a soil depth of 5\u00a0cm simultaneously with the soil chamber measurements next to the rings. Incoming solar radiation was measured (CNR-1, Kipp & Zonen, Delft, The Netherlands) at an adjacent meteorological station.\n\nTo measure the soil flux, a fused silica bell was placed in a water-filled channel on top of a stainless steel ring (see Online Resource 1 for a sketch of the experimental setup). All tubing and fittings used were either made of PFA or stainless steel. Each measurement cycle started with air being drawn with approx. 1.5\u00a0l/min from a tube outside, but near the chamber to quantify the ambient concentration, whilst the chamber line was flushed at the same flow rate. After 5\u00a0min, the lines were switched and the air in the headspace of the chamber was measured until a steady state was reached (11.63\u00a0\u00b1\u00a04.6\u00a0min); meanwhile, the ambient line was flushed at the same flow rate. Afterwards, the ambient concentration was measured a second time and a linear interpolation was performed to estimate the ambient concentration at the time of the steady-state conditions within the chamber. The soil COS and CO~2~ flux was calculated according to the following equation:$$\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$F = \\frac{{q(C_{ 2} - C_{ 1} )}}{A}.$$\\end{document}$$Here, *F* denotes the soil flux in pmol\u00a0m^\u22122^\u00a0s^\u22121^ for COS and \u00b5mol\u00a0m^\u22122^s^\u22121^ for CO~2~ (air temperature and pressure were measured to calculate the molar density), *q* is the flow rate (l/min) while measuring the chamber concentrations, *C* ~1~ is the ambient concentration derived from the linear interpolation (in pmol\u00a0m^\u22122^ s^\u22121^ for COS and \u00b5mol\u00a0m^\u22122^s^\u22121^ for CO~2~), *C* ~2~ the concentration in the chamber at steady-state, and *A* the surface area (0.032\u00a0m^2^) covered by the chamber. For each of the 6 measurement days, we made between 12 and 42 flux measurements, yielding a total of 172 soil COS and CO~2~ measurements with associated environmental conditions. Data were processed with MATLAB (MATLAB 8.6.0.267246 (R2015b), The MathWorks Inc., Natick, MA, USA) and the function PropError by Ridder ([@CR100]) was used for error propagation.\n\nTo assess potential emissions of COS by materials used in the experimental setup, measurements without soil (by covering the bottom of the stainless steel rings with a Teflon foil) were conducted and no significant flux was measured under dark conditions. Under high light intensities, however, a flux from the empty chamber was measured. It led to a maximum difference of 40 ppt between chamber and ambient air translating to a flux of 2\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ (Online Resource 1 contains sample data for high light intensity blank measurements). We were not able to identify the source of the emission, as the air stream had only contact with stainless steel (ring and fittings), PFA (tubes and fittings), and fused silica (chamber). To further assess the influence of incoming radiation, one fused silica bell was covered in aluminum foil and plots were subsequently measured with the transparent and the covered bell on 4\u00a0days later in the season.\n\nDifferential pressure measurements (MKS Baratron Type 226A Differential Pressure Transducer, MKS Instruments Inc., Andover, MA, USA) were conducted prior to the experiment to find the best flow rate, avoiding under inflation or overpressure within the chamber (Rayment and Jarvis [@CR34]). A flow rate of 1.55 slm resulted in pressure differences being lower than 0.2\u00a0Pa, corresponding to the instrument resolution.\n\nStatistical analyses were carried out with R version 3.2.2---\"Fire Safety\" (R Core Team [@CR32]). To obtain normality and homogeneity, the response variable (COS flux) was log-transformed (to the base of 10). A single influential observation (high cook's distance and studentized residual \\<\u221212) was excluded. A multiple linear regression was performed, data from an 8\u00a0h interval for each day during the drought experiment were used, first with a mixed effect model (Bates et al. [@CR6]) to check for subject (plot) specific effects. A variance close to zero for the random effect indicated that the model is degenerate (Bates [@CR5]) and the random effect was discarded. In addition, the measurement days were dummy coded to test if there are significant differences between the measurement days, which was not the case, resulting in a model without random effects and differentiation between days. The relative importance of the predictors for the log-transformed response was calculated using the relaimpo package (Gr\u00f6mping [@CR17]) with the method \"lmg\" (Lindeman et al. [@CR27]). Kruskal--Wallis tests were performed to test for significant differences between dark/light, night/day, and daily measurements, night was defined as the period with a solar elevation angle equal to or below \u22126 degrees.\n\nResults {#Sec3}\n=======\n\nDuring the simulated drought, the soil water content dropped to a minimum of 5% (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}), starting from a minimum of 30.7% prior to the treatment. The SWC means of the control and the rain exclusion were significantly different on all days (all *p* values \\<0.01). Soil temperature was not significantly different between the treatments and reached its maximum of 31.7\u00a0\u00b0C on Day 4 (21-Jul-2015) (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). CO~2~ fluxes decreased slightly, compared to the control treatment, as the soil water content decreased, on the contrary COS fluxes remained unaffected (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}).Fig.\u00a01**a**,**b** *Boxplots of the COS and CO* ~2~ fluxes for an 8\u00a0h interval (10:00--18:00) for both treatments, a *star* (\u02df) connotes a significant (Kruskal--Wallis, *p*\u00a0\\<\u00a00.05) difference between the treatments. *Arrows* denote start and end of the treatment period. The *light grey* area denotes the estimated measurement error. On each day, the first box plot is the control and the second one the rainout. **c** Daily mean incoming shortwave radiation (*bars*) and the daily soil temperatures for the drought treatment and the control are shown (*symbols*). **d** *Bars* show the precipitation before and after the rain exclusion and the watering of the control plots during the manipulation. *Symbols* denote the volumetric soil water content (%). Both treatments were measured on the same dates, but are shifted in time to improve visibility. The figure is available in color in the online version\n\nA clear difference between day (12.5\u00a0\u00b1\u00a013.8\u00a0pmol\u00a0m^\u22122^ s^\u22121^) and night soil COS flux (0.4\u00a0\u00b1\u00a01.7\u00a0pmol\u00a0m^\u22122^ s^\u22121^) could be seen during the 24\u00a0h measurements (Figs.\u00a0[2](#Fig2){ref-type=\"fig\"}, [3](#Fig3){ref-type=\"fig\"}). Nighttime fluxes ranged from a minimum of \u22122.0\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ to a maximum of 5.6\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ and daytime fluxes from \u22122.0 to 76.0\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^. The maximum residual emission of empty chambers at high light intensities of 2.0\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ determined during extensive blank tests thus had a negligible effect on the observed range of fluxes. As shown in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}, the clear contrast between day and night conditions was due to the soil COS exchange closely following changes in incoming radiation. Soil CO~2~ fluxes exhibited, relative to their magnitude, a weaker daily pattern (data not shown) and a smaller difference between day (8.1\u00a0\u00b1\u00a04.3\u00a0\u00b5mol\u00a0m^\u22122^\u00a0s^\u22121^) and night (5.4\u00a0\u00b1\u00a01.1\u00a0\u00b5mol\u00a0m^\u22122^\u00a0s^\u22121^). Daytime measurements of the soil COS flux with darkened chambers were significantly different from light measurements (*p*\u00a0\\<\u00a00.001) and were similar in magnitude to nighttime fluxes (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}).Fig.\u00a02COS fluxes (*symbols*) measured during the experiment, with bars indicating the solar incoming radiation and error bars denoting the propagated errors for each chamber measurement. Plots show data from **a** 10-Jun-2015, **b** 26/27-Jun-2015, **c** 9/10-Jun-2015, **d** 21-Jul-2015, **e** 5-Aug-2015, and **f** 13-Aug-2015. The color coding of the*symbols* reflects the soil temperature measured next to the chambers. The figure is available in color in the online version Fig.\u00a03*Boxplots*showing the difference between soil COS fluxes with and without light (light excluded by covering the fused silica bell in aluminum foil) (sig. different, Kruskal--Wallis, *p*\u00a0\\<\u00a00.05), and between daytime and nighttime soil COS fluxes (sig. different, Kruskal--Wallis, *p*\u00a0\\<\u00a00.05) Fig.\u00a04Relationship between soil COS fluxes and **a** soil water content, **b** soil temperature in 5\u00a0cm, and **c** incoming solar radiation. An exponential function ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\widehat{COS}flux = \\, 3.196 \\, * \\, \\exp \\left( {0.002616*{\\text{Incoming radiation}}} \\right)$$\\end{document}$) was fitted to the data (*black line*) in panel (**c**)\n\nNo multicollinearity was detected (all variance inflation factors \\<4) for the multiple linear regression ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\log (\\widehat{\\cos }{\\text{flux}}) = \\hat{\\beta }_{0} + \\hat{\\beta }_{1} {\\text{Incoming}}\\; {\\text{radiation}} + \\hat{\\beta }_{2} {\\text{Tsoil}} + \\hat{\\beta }_{3} {\\text{SWC}}$$\\end{document}$) (see Table\u00a0[1](#Tab1){ref-type=\"table\"} for summary) and the whole model has predictive capability (F Statistic: *p*\u00a0\\<\u00a00.05). The coefficients for incoming radiation and soil temperature (Tsoil) were highly significant and the coefficient for SWC was significant. By plotting the response variable log(COS flux) against the model estimates, no over- or underestimation within the range of the model was detected; the intercept was not significantly different from 0 (*p* value\u00a0=\u00a01). The model explained 68.2% of the variation in log(COS flux). The relative importance of the predictors for *R* ^2^ normalized to 100%, for log(COS flux) in the model amounted to 70.4% for incoming radiation, 25.1% for soil temperature, and 4.5% for SWC. Note that the percentages refer to the variance of the common logarithm of the COS flux. Due to the monotonic property of the logarithm, the order of relative importance remains the same after back transformation. According to the model, an increase of 1\u00a0Wm^\u22122^ in incoming radiation leads to an increase of 0.3% in the COS flux and an increase of 1\u00a0\u00b0C in soil temperature leads to an increase of 12.4% in the COS flux.Table\u00a01Summary of the linear regression model $\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$\\log (\\widehat{\\cos }{\\text{flux}}) = \\hat{\\beta }_{0} + \\hat{\\beta }_{1} {\\text{Incoming radiation}} + \\hat{\\beta }_{2} {\\text{Soil temperature}} + \\hat{\\beta }_{3} {\\text{Soil water content}}$$\\end{document}$EstimateStd. error*t* valuePr(\\>\\|*t*\\|)(Intercept)\u22120.86090.2072\u22124.160.0001Incoming radiation0.00100.000111.040.0000Tsoil0.05060.00766.620.0000SWC0.00670.00203.280.0015Observations55*R* ^2^0.6824\n\nDiscussion {#Sec4}\n==========\n\nThe objective of this study was to examine the influence of soil water content, previously found significant in laboratory experiments (Kesselmeier et al. [@CR20]), on in situ COS soil fluxes. The SWC, which varied between 5 and 47% and thus covered almost the entire range of plant extractable soil water at this site, had no significant effect on the COS flux (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). This finding is in conflict with results from the literature (Kesselmeier et al. [@CR20]; Van Diest and Kesselmeier [@CR44]), especially considering that microbial communities are negatively impacted by droughts (Borken and Matzner [@CR10]; Lavigne et al. [@CR25]; Schimel et al. [@CR38]) and enzymatic activity within the soil is generally reduced under water stress (Davidson and Janssens [@CR16]; Sardans and Penuelas [@CR36]). However, in past studies, the activity of some enzymes was unaffected by drought (Sardans et al. [@CR37]) or was even stimulated (Sanaullah et al. [@CR35]), very much depending on microbial communities, their demand for nutrients, their interaction with plants, and the severity of the drought. The lack of knowledge concerning microorganisms involved in the production and consumption of COS makes it difficult to predict their reaction to water stress. In addition, the previous studies (Kesselmeier et al. [@CR20]; Van Diest and Kesselmeier [@CR44]) linked SWC primarily to COS uptake, whereas in our study, COS emission prevails.\n\nA proxy for biological activity within soils is soil respiration, which was lower in the drought treatment compared to the control, indicating differences in the activity of soil biota between the treatments (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The relatively small reduction in soil respiration can be explained by the relatively high groundwater level and the 'water-spending' strategy of the plant species present at the study site, which keep their activity up until very low levels of soil water content are reached (Brilli et al. [@CR11]). These aforementioned impacts on soil biota by a drought event should be visible in the soil COS fluxes, if COS consumption or production is mainly driven by biotic processes. The consistent lack of an SWC effect on soil COS flux in our model thus hints towards strong abiotic influences. Ogee et al. ([@CR31]) and Sun et al. ([@CR42]) demonstrated that changes in the soil COS flux caused by changes in the SWC are mainly due to changing diffusion rates in the soil column. Sinks and sources for COS on the soil surface would, therefore, be less affected.\n\nSoil temperature at a depth of 5\u00a0cm was the second most important predictor for the soil COS flux in the model. Within the measured soil temperature range (from 17.7 to 31.7\u00a0\u00b0C), the COS flux exhibited no optimum in contrast to Kesselmeier et al. ([@CR20]), but increased exponentially with rising temperature (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}), a finding shared with other studies (Maseyk et al. [@CR28]; Whelan et al. [@CR49]; Whelan and Rhew [@CR47]). An increase in temperature can stimulate and accelerate microbial and enzymatic activity in soils explained by the Arrhenius equation (Arrhenius [@CR2]), which links chemical reaction rates to temperature. However, temperature also interacts with other soil properties like SOM adsorption and desorption onto mineral surfaces and soil water thickness (important for diffusion processes of compounds and, therefore, substrate availability) (Davidson and Janssens [@CR16]). The ability of organisms and communities to adapt to changing environmental conditions, such as increasing temperature, can modify their reaction over time (Allison et al. [@CR1]).\n\nThe most important parameter in the model was incoming solar radiation. Even though data about photoproduction of COS in the ocean (von Hobe [@CR45]) and in precipitation (Mu et al. [@CR29]) are available, data concerning terrestrial production are scarce (Whelan and Rhew [@CR47]), especially the global magnitude is unknown. The previous experiments (Mu et al. [@CR29]; von Hobe [@CR45]) concluded that especially the UV fraction of sunlight is responsible for abiotic COS production, though reactions involved are largely unknown. Sun et al. ([@CR42]) compared the results from their soil diffusion--reaction model to observed data (Maseyk et al. [@CR28]) and realized that there is larger mismatch between observed data and modelled data at higher positive fluxes and an underestimation of the flux at midday. The authors considered photochemical production as one source not fully captured by their model, even though the chambers used in the field observation itself were opaque. Due to the use of fused silica chambers in our experiment, UV radiation was able to reach the soil surface largely unattenuated and thus presumably was able to trigger reactions involving precursor compounds on the soil surface and top-most soil layers. Past lab experiments already pointed out that radiation is a considerable driver for COS fluxes in soils and on the soil surface (Whelan et al. [@CR49]; Whelan and Rhew [@CR47]). Our study supports this notion, COS fluxes closely following the intensity of incoming solar radiation (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}), even though we did not disentangle pure photochemical processes from secondary effects like an increase of the surface temperature (thermoproduction).\n\nFurthermore, in this study, the vegetation was removed to be able to access the soil surface, which allowed a much higher fraction of the solar radiation to reach the ground as compared to natural conditions, when the soil surface is shaded by vegetation. This circumstance is important for interpreting our comparably high fluxes, which are more similar to fluxes measured on agricultural fields after harvest (thus with less shading by plants) (Berkelhammer et al. [@CR7]; Billesbach et al. [@CR9]). Maseyk et al. ([@CR28]) showed persistent COS emission in the late season, especially after harvest, and no direct correlation between COS fluxes and soil water content for an agricultural study site, similar to our study. In their study, the relationship between soil temperature and COS flux changed at an SWC threshold of ca. 20%, a pattern we did not observe. The authors hypothesized that COS production could be attributed to the remaining roots in the soil. Simulating the soil COS exchange under real-world conditions within a grassland ecosystem, i.e., with the plant canopy shading the soil surface, thus requires measuring or simulating the intensity of radiation reaching the ground. A simple way to approximate the radiation incident on the soil surface is to use the Beer--Lambert law (Campbell and Norman [@CR12]). If we assume an extinction coefficient of 0.4 (Zhang et al. [@CR52]), incoming solar radiation of 1000\u00a0Wm^\u22122^ (light intensity around noon on a sunny day) and an LAI/GAI of 3.5\u00a0m^2^m^\u22122^ (medium canopy height and density; GAI defined as half of the green surface area in m^2^ per m^2^ projected base area), 173.8\u00a0Wm^\u22122^ would reach the soil surface, after the cut (GAI of 1.8) 407 Wm^\u22122^. Using the regression formula presented in this paper, at a soil temperature of 25\u00a0\u00b0C, the COS fluxes would be 2.9 and 5\u00a0pmol\u00a0m^\u22122^ s^\u22121^, respectively. In comparison, the two equations presented in Maseyk et al. ([@CR28]) would yield, at a soil temperature of 25\u00a0\u00b0C and without considering radiation, fluxes of 6.9\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ (\\>20% SWC) and 3.2\u00a0pmol\u00a0m^\u22122^\u00a0s^\u22121^ (\\<20% SWC). Therefore, after factoring in shading by the plant canopy, the fluxes presented here are similar in their magnitude compared to the previous studies (Billesbach et al. [@CR9]; Maseyk et al. [@CR28]).\n\nOur results also highlight the need to systematically review and compare soil COS measurement approaches. Most recent studies used opaque chambers (Berkelhammer et al. [@CR7]; Maseyk et al. [@CR28]; Sun et al. [@CR43]) and thus excluded potential photoproduction of COS. The resulting data may thus not be reflective of the true soil COS exchange, as even in dense canopies, there are periods during the season (e.g., before leaf-out in deciduous forests or after harvesting in agricultural systems) or certain times of the day, when significant sunlight penetrates to the soil surface. Our study faced a different problem, as the plant canopy had to be removed to access the soil, resulting in much more radiation reaching the soil surface during chamber measurements and the need to simulate the true light availability at the soil surface to estimate the actual soil COS flux. A potential source of bias in our flux estimates results from chamber COS concentrations exceeding 1000 ppt (compared to ambient mole fractions of 500--550 ppt) in cases of high emissions during periods of high light intensity, which might affect the COS soil diffusion profile and thus the magnitude of the soil surface flux (Ogee et al. [@CR31]; Sun et al. [@CR42]). The fact that our measurement system, despite using the most inert materials (fused silica, stainless steel, and PFA), was characterized by a residual COS emission (which, however, was close to negligible compared to the magnitude of the measured fluxes), highlights the importance of carefully characterizing soil COS measurement systems, for which no standard reference exists.\n\nConclusions {#Sec5}\n===========\n\nContrary to our hypothesis, which motivated the simulated drought experiment, soil water content, even though important for soil biota (as evident from the reduction in soil respiration during the course of the drought experiment), had no impact on measured soil COS fluxes. To our surprise, but consistent with recent literature (Whelan and Rhew [@CR47]), the soil COS exchange was driven primarily by radiation, with a secondary influence by temperature, suggesting abiotic production to dominate the soil--atmosphere exchange of COS. The dominating influence of radiation can be explained by the high radiation levels incident on the soil surface with our experimental approach, which would be lower under in situ conditions, when the vegetation cover is present, but could occur in agricultural systems after harvesting.\n\nCurrent attempts (Ogee et al. [@CR31]; Sun et al. [@CR42]) to devise a universal theoretical framework explaining the roles of biological and abiotic processes in governing the direction and magnitude of soil COS exchange need to be extended to be able to explain the range of observed soil COS fluxes and in particular light-driven emissions. More empirical data from a larger diversity of soil types, both in situ and under laboratory conditions, will be essential to this end. Especially, the small number of transparent chamber measurements up to this date could be a reason for the dominance of studies reporting predominately soil COS uptake. To locate potential COS sinks and sources depth-resolved soil, COS measurements should be conducted; among them, less intrusive measurement techniques (e.g., using Radon-222 calibrated methods) could also help to avoid artifacts. With regard to the use of COS as a tracer for canopy photosynthesis, the strong light-dependence of the soil COS exchange suggest that accounting for the soil contribution may be more complicated than previously thought as it is likely to require measuring/simulating the radiation intensity reaching the soil surface.\n\nElectronic supplementary material\n=================================\n\n {#Sec6}\n\nBelow is the link to the electronic supplementary material. Supplementary material 1 (PDF 415 kb)\n\nOpen access funding provided by Austrian Science Fund (FWF). This study was financially supported by the Austrian National Science Fund (FWF) under contract P27176-B16 and the Tyrolean Science Fund (TWF) under contract UNI-0404/1801. Thanks to all who helped during our field campaign and to family Hofer (Neustift, Austria) for granting us access to the study site.\n\nAuthor contribution statement {#FPar1}\n=============================\n\nGW originally formulated the idea, AH, FS, and FK developed methodology, AH, TL, FS, and FK conducted fieldwork, FK processed the data, FK performed statistical analyses, and FK and GW wrote the manuscript.\n\n[^1]: Communicated by David R. Bowling.\n"}
+{"text": "**Core tip:** The time interval between neoadjuvant therapy and surgery is an underlying debate. This study mainly analyzed the effect of prolonging the time interval to surgery after neoadjuvant therapy on pathologic response and disease-free survival. Our study concluded that time interval \\> 9 wk will result in better prognosis and oncologic outcomes than time interval \u2264 9 wk. In practice, we should prolong the time interval to obtain favorable outcomes in advanced rectal cancer after neoadjuvant therapy.\n\nINTRODUCTION\n============\n\nAccording to the latest report, with the increase in morbidity and mortality, rectal cancer ranks as the third most common cancer in China and is a major public health problem\\[[@B1]\\]. In addition, according to the epidemiological characteristics of rectal cancer in China, locally advanced rectal cancer is most common. Advanced primary tumor stage negatively impacts prognosis with local recurrence and synchronous metastases\\[[@B2]\\]. Over the past few decades, neoadjuvant chemoradiotherapy has been shown to reduce the rate of local recurrence after surgery for advanced rectal cancer to some extent, but overall survival (OS) was not improved significantly in some large randomized trials\\[[@B3],[@B4]\\]. In recent years, with the increasing incidence of rectal cancer in China and the higher proportion of advanced rectal cancer, neoadjuvant chemoradiotherapy has gradually become the standard treatment scheme for locally advanced rectal cancer in China. In several prospective cohort and retrospective studies\\[[@B5],[@B6]\\], neoadjuvant chemoradiotherapy has been indicated to lead to longer disease-free survival (DFS) in patients with pathologic complete response (pCR) and tumor downstaging. Subsequently, related studies were carried out on how to improve the pCR and tumor downstaging rates. By studying different surgical procedure methods and preoperative chemoradiotherapy doses and protocols, researchers further improved the pCR rate and tumor downstaging rate\\[[@B7]-[@B9]\\]. In recent years, some centers have also begun to observe whether the pCR rate and DFS rate can be further improved by examining the time interval to surgery after preoperative chemoradiotherapy.\n\nThe timing of surgery after neoadjuvant therapy for rectal cancer is an unresolved subject. Therefore, the first prospective randomized controlled study on the issue was established as the Lyon Trial R90-01. Francois et al\\[[@B10]\\] enrolled 201 patients between 1991 and 1995 and published the results of the study in 1999. In the trial, the registered patients were divided into two groups according to the time of surgery after neoadjuvant therapy: A 2-wk group and a 6-8 wk group. They concluded that the longer time interval group (6-8 wk) had a significant advantage in clinical tumor response and pathologic downstaging, while there were no significant differences between the groups in postoperative complications, local recurrence, sphincter-preservation, or short-term survival. The second phase of the trial was the 17-year follow-up of the two groups of patients. Although the conclusion was that preoperative neoadjuvant therapy can effectively eliminate residual tumor cells, which is a sign of a good prognosis, the OS rates were similar in the groups with long follow-up, and the time interval to surgery after neoadjuvant chemoradiation of 6-8 wk as the standard treatment protocol was widely accepted in clinical practice\\[[@B11]\\]. More recent studies have described that a prolongation of the time interval achieved higher rates of downstaging and pCR and an especially favorable DFS\\[[@B12]\\]. The guidelines of the National Comprehensive Cancer Network recommend that surgery should be performed 5--12 wk following neoadjuvant chemoradiation\\[[@B13]\\]. In the GRECCAR-6 randomized trial\\[[@B14]\\], which divided patients into interval times of 7 and 11 wk, there was no significant difference in the effect of the time intervals on the pCR rate, whereas the quality of specimens was poorer and the rate of complications was higher in patients with an interval of 11 wk. Overall, the optimal time interval for operation is still a topic to be researched through clinical trials.\n\nThis study aimed to evaluate the optimal time interval to perform surgery following neoadjuvant chemoradiotherapy for increasing downstaging and pCR rates and improving DFS.\n\nMATERIALS AND METHODS\n=====================\n\nPatients\n--------\n\nOur retrospective cohort included 231 patients with locally advanced rectal cancer who underwent neoadjuvant therapy followed by radical surgery at the Colorectal Surgery Unit of the National Cancer Center/National Sciences Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College between November 2014 and August 2017 and had clinical, neoadjuvant, pathologic, and follow-up information available. This research was authorized by the ethics committee of our institution, and informed consent to collect clinical data was obtained from each patient following the principles of the World Medical Association Declaration of Helsinki. The patients enrolled in this study needed to meet the following criteria: (1) All patients had endoscopy biopsy-proven adenocarcinoma within 12 cm of the anal verge through endoscopy; (2) Tumors were scanned by pelvic magnetic resonance imaging and thoracic and abdominal-pelvic enhanced computed tomography (CT) and classified as locally advanced rectal cancer without metastases (stage cT3/T4 or cT any cN1/2, cM0) at the time of diagnosis; and (3) Laparoscopic surgery was performed following the completion of neoadjuvant therapy. The exclusion criteria were as follows: (1) Clinical tumor-node-metastasis (TNM) stages I and IV; (2) Patients diagnosed with other carcinomas and lesions; (3) Obstruction, perforation, bleeding, and palliative resection; and (4) Patients who refused to accept chemoradiotherapy.\n\nNeoadjuvant therapy\n-------------------\n\nBased on clinical stages II and III, all enrolled patients received the following neoadjuvant therapy regime. Preoperative radiotherapy was performed at a dose of 45 Gy to all pelvic tissue, 25 fractions/5 wk in a long-course regime, and then the linear accelerator was used to inject 5.4 Gy/3 fractions/3 d into the primary tumor with an energy of 10 MV. Radiation therapy technique used was intensity modulated radiotherapy. The radiating field involved the upper margin 1.5 cm above the sacral region and the lateral margin 1.5 cm outside the pelvic bone, including pelvic lymph nodes. One of the three following chemotherapy regimens was administered concurrent with radiation: (1) XELOX regimen and capecitabine 825 mg/m^2^ orally administered twice daily 5 d/wk and oxaliplatin 130 mg/m^2^ IV on the first day every 3 wk as a cycle; (2) Oral capecitabine 825 mg/m^2^ twice daily 5 d/wk alone; and (3) Other regimens of oxaliplatin combined with others.\n\nSurgical procedure\n------------------\n\nSurgery was scheduled 6-8 wk after the completion of chemoradiotherapy, but due to clinical factors such as logistics and scheduling, the actual interval varied from 2 to 60 wk. All surgical procedures were performed by colorectal surgeons specialized in oncology using the standardized laparoscopic total mesorectal excision (TME) technique defined by Heald et al\\[[@B15]\\] and a pelvic autonomic nerve preservation technique. The proximal and distal colon was anastomosed end to end by using the double stapler technique. To prevent severe complications from being caused by postoperative anastomotic leakage, a protective stoma was used as a diversion method based on the anastomotic tension and colonic blood supply during the operation, and 51 (22.1%) patients underwent protective stoma with the proximal ileum. The other key surgical procedures were left colic artery preservation (6.9%), multivisceral resection (5.2%), and lateral lymph node dissection (9.1%). All specimens were of high quality and completely measured by expert pathologists.\n\nPathologic analysis\n-------------------\n\nAll postoperative pathological specimens were fixed in paraffin within 24 h. The prepared specimens were analyzed by two pathologists specializing in colorectal cancer, and group information was kept confidential to the pathologists. The resected specimens were reviewed according to TNM staging (ypTNM), and the tumor tissues were analyzed by microscopy by the pathologists. The Dworak classification was used to determine tumor regression grades, with grades 0 to 4. pCR was defined as no residual cancer cells in the primary tumor tissue and no metastasis of tumor cells in the lymph nodes of the dissected region. Meanwhile, according to the WHO classification of digestive system tumors\\[[@B16]\\], the differentiation degree of tumor tissue was determined.\n\nFollow-up\n---------\n\nThe first follow-up occurred 3 mo after surgery, and follow-up visits were conducted every 3 mo for the first two years at our center. Clinical examination and enhanced CT examination of the chest, abdomen, and pelvic cavity were performed, and colonoscopy was performed every 6 mo during the first two years. Beyond two years, the patients were followed every 6 mo, and colonoscopy was performed once a year. DFS was defined as the time from the date of radical resection to the time when disease recurrence was determined by radiological examinations. OS was defined as the time from the date of radical resection to the time when death caused by cancer occurred. The last follow-up date of this study was March 22, 2020. Patients who did not reach the endpoint or who were lost to follow-up before that date were censored. The median follow-up was 41 mo.\n\nStatistical analysis\n--------------------\n\nFor statistical analysis, the chi-squared test or Fisher's exact test was used to compare binary and categorical variables. Continuous data were analyzed by *t*-test if they were normally distributed and are displayed as the mean \u00b1 SD. If continuous data were not normally distributed, they were compared using the Mann-Whitney *U* test and are displayed as medians (ranges). Comparisons of the effects of the following variables on the two groups were carried out: Clinical characteristics, neoadjuvant information, surgical procedure, postoperative complications, pCR status, and other factors on intervals. All variables that achieved *P* \\< 0.05 in univariate analysis were included in the multivariable analysis. Multivariable regression analysis was performed to further identify factors independently associated with the time interval to surgery. The Kaplan-Meier method and Cox proportional hazards regression analysis were used to extract independent factors, especially time interval, associated with DFS and OS. All statistical analyses were performed using the Statistical Package for the Social Sciences version 26.0 for Mac (IBM Corp., Armonk, NY, United States).\n\nRESULTS\n=======\n\nClinical characteristics and outcomes\n-------------------------------------\n\nOverall, 231 patients with locally advanced rectal cancer received neoadjuvant therapy followed by laparoscopic TME surgery from November 2014 to August 2017 at the Department of Colorectal Surgery, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. The enrolled patients had an average age of 58.7 \u00b1 11.2 years at the time of surgery, and the population consisted of 160 (69.3%) males and 71 (30.7%) females. The average body mass index (BMI) was 24.1 \u00b1 3.2 kg/m^2^. The time interval between neoadjuvant chemotherapy and surgery ranged from 2 to 60 wk, with a median of 9 wk. The patient and tumor characteristics of the 139 (60.2%) patients who underwent surgery \u2264 9 wk after chemoradiotherapy (group A) and the 92 (39.8%) patients who underwent surgery \\> 9 wk after chemoradiotherapy (group B) are presented in Table [1](#T1){ref-type=\"table\"}. The two groups were balanced by age, sex, American Society of Anesthesiologists (ASA) score, BMI, preoperative carcinoembryonic antigen level, clinical T stage, clinical N stage, and preoperative concurrent chemotherapy regimen. The surgery results and tumor outcomes are presented in Table [2](#T2){ref-type=\"table\"}. All surgical procedures were performed by laparoscopy, and the surgical procedures were not significantly different and were unaffected by the neoadjuvant surgery time in the groups (*P* = 0.241). More patients underwent sphincter preservation surgery in group A than in group B, although the difference was not statistically significant (group A, 52.5% *vs* group B, 43.5%; *P* = 0.179). In the comparison of complications between the two groups, there were no significant differences in the incidence or type of complications; the incidence of complications in group A was similar to that in group B (8.6% *vs* 7.6%; *P* = 0.894). Blood loss and operation time were significantly different between the two groups. More blood loss and longer operative times were found in group B than in group A (B, 100.9 \u00b1 133.9 mL *vs* A, 68.0 \u00b1 92.1 mL; *P* = 0.007 and B, 222.2 \u00b1 67.8 min *vs* A, 194.5 \u00b1 65.2 min; *P* = 0.001). Pathologic T stage (ypT) and pathologic N stage (ypN) were significantly different between the two groups (ypT, *P* = 0.022; ypN, *P* = 0.014). The rate of pCR was obviously higher in group B than in group A (group B, 27.2% *vs* group A, 10.8%; *P* = 0.001). Additionally, patients in group A had more lymph nodes retrieved (group A, 19.0 \u00b1 8.8 *vs* group B, 15.7 \u00b1 8.7; *P* = 0.001) and positive lymph nodes (group A, 1.6 \u00b1 3.2 *vs* group B, 0.8 \u00b1 1.7; *P* = 0.045) than those in group B. The exhaust time, defecation time, micturition time, and degree of differentiation were not affected by the interval time.\n\n###### \n\nPatient and tumor characteristics\n\n  **Variable**                                              **Interval \u2264 9 wk** (**group A, *n* = 139**)   **Interval \\> 9 wk (group B, *n* = 92)**   ***P* value**\n  --------------------------------------------------------- ---------------------------------------------- ------------------------------------------ ---------------\n  Sex, *n* (%)                                                                                                                                        0.279\n  Male                                                      100 (71.9)                                     60 (65.2)                                  \n  Female                                                    39 (28.1)                                      32 (34.8)                                  \n  Age (year, mean \u00b1 SD)                                     58.6 \u00b1 10.6                                    58.9 \u00b1 11.7                                0.645\n  ASA, *n* (%)                                                                                                                                        0.280\n  1                                                         4 (2.9)                                        3 (3.3)                                    \n  2                                                         111 (79.9)                                     67 (72.8)                                  \n  3                                                         24 (17.2)                                      22 (23.9)                                  \n  BMI (kg/m^2^ , mean \u00b1 SD)                                 23.8 \u00b1 3.4                                     24.4 \u00b1 2.9                                 0.063\n  cT, *n* (%)                                                                                                                                         0.318\n  2                                                         2 (1.4)                                        0 (0)                                      \n  3                                                         110 (79.1)                                     70 (76)                                    \n  4                                                         27 (19.4)                                      22 (24)                                    \n  cN, *n* (%)                                                                                                                                         0.355\n  0                                                         40 (28.8)                                      22 (23.9)                                  \n  1                                                         71 (51.1)                                      48 (52.2)                                  \n  2                                                         28 (20.1)                                      22 (23.9)                                  \n  Preoperative carcinoembryonic antigen (\u03bcg/L, mean \u00b1 SD)   5.4 \u00b1 10.8                                     5.8 \u00b1 13.7                                 0.182\n  Preoperative concurrent chemotherapy regimen, *n* (%)                                                                                               0.712\n  Capecitabine + oxaliplatin                                32 (23)                                        29 (31.5)                                  \n  Capecitabine oral                                         87 (62.6)                                      58 (63)                                    \n  Oxaliplatin union                                         21 (15.1)                                      5 (5.4)                                    \n\nBMI: Body mass index; ASA: American Society of Anesthesiologists.\n\n###### \n\nSurgical results and tumor outcomes\n\n  **Variable**                               **Interval \u2264 9 wk**   **Interval \\> 9 wk**   ***P* value**\n  ------------------------------------------ --------------------- ---------------------- ---------------\n  Surgical procedure                                                                      \n  Miles (*n*, %)                             55 (39.6)             45 (48.9)              0.161\n  Dixon (*n*, %)                             73 (52.5)             40 (43.5)              0.179\n  Hartmann (*n*, %)                          11 (7.9)              7 (7.6)                0.933\n  Operative time (min, mean \u00b1 SD)            194.5 \u00b1 65.2          222.2 \u00b1 67.8           0.001\n  Estimated blood loss (mL, mean \u00b1 SD)       68.0 \u00b1 92.1           100.9 \u00b1 133.9          0.007\n  Postoperative complications                7 (5)                 5 (5.4)                0.894\n  Hemorrhage, *n* (%)                        1 (0.7)               1 (1.1)                \n  Anastomotic leakage, *n* (%)               5 (3.6)               2 (2.2)                \n  Ileus, *n* (%)                             2 (1.4)               0 (0)                  \n  Reoperation, *n* (%)                       3 (2.2)               1 (1.1)                \n  Cardiovascular, *n* (%)                    0 (0)                 1 (1.1)                \n  Urinary retention, *n* (%)                 1 (0.7)               2 (2.2)                \n  Exhaust, d (mean \u00b1 SD)                     3.6 \u00b1 1.5             3.9 \u00b1 1.5              0.148\n  Defecation, d (mean \u00b1 SD)                  4.7 \u00b1 1.7             5.2 \u00b1 2.0              0.249\n  Micturition, d (mean \u00b1 SD)                 4.5 \u00b1 1.4             4.4 \u00b1 1.0              0.614\n  Degree of differentiation, *n* (%)                                                      0.475\n  Low and low-moderate grades                22 (15.8)             10 (10.9)              \n  Moderate, moderate-high, and high grades   114 (82)              79 (85.9)              \n  Signet-ring and mucinous adenocarcinoma    3 (2.2)               3 (3.3)                \n  ypT, *n* (%)                                                                            0.009\n  1                                          3 (2.2)               2 (2.2)                \n  2                                          29 (20.9)             19 (20.7)              \n  3                                          82 (59)               39 (42.4)              \n  4                                          10 (7.2)              7 (7.6)                \n  ypN, *n* (%)                                                                            0.014\n  0                                          68 (48.9)             60 (65.2)              \n  1                                          50 (36)               24 (26.1)              \n  2                                          21 (15.1)             8 (8.7)                \n  pCR                                        15 (10.8)             25 (27.2)              0.001\n  Total lymph nodes                          19.0 \u00b1 8.8            15.7 \u00b1 8.7             0.001\n  Positive lymph nodes                       1.6 \u00b1 3.2             0.8 \u00b1 1.7              0.045\n\npCR: Pathologic complete response.\n\nMultivariate analysis\n---------------------\n\nAll factors that achieved a *P* \\< 0.05 in univariate analysis were selected for binary logistic regression analysis since those factors have a possible association with being affected by interval time, including pathologic T stage (ypT), pathologic N stage (ypN), intraoperative blood loss (mL), operation time, total lymph nodes, positive lymph nodes, and rate of pCR. The binary logistic regression analysis showed that a shorter time interval (\u2264 9 wk) between neoadjuvant therapy and surgery had the effect of decreasing the rate of pCR \\[odds ratio (OR) = 2.668; 95%CI: 1.276-5.578; *P* = 0.009\\]. In addition, operative time (OR = 1.006; 95%CI: 1.001-1.01; *P* = 0.01) and the total number of lymph nodes retrieved (OR = 0.952; 95%CI: 0.918-0.986; *P* = 0.007) were independent factors affected by time interval. The results are displayed in Table [3](#T3){ref-type=\"table\"}.\n\n###### \n\nMultivariate logistic regression analysis identifying independent factors affected by interval time\n\n  **Variable**           ***P* value**   **OR**   **95%CI**\n  ---------------------- --------------- -------- -------------\n  ypT                                             \n  ypT1 *vs* ypT4         0.799           0.753    0.085-6.655\n  ypT2 *vs* ypT4         0.779           0.841    0.25-2.826\n  ypT3 *vs* ypT4         0.538           0.705    0.233-2.14\n  ypN                                             \n  ypN0 *vs* ypN2         0.511           0.492    0.059-4.082\n  ypN1 *vs* ypN2         0.431           0.469    0.071-3.082\n  Positive lymph nodes   0.234           0.922    0.807-1.054\n  Total lymph nodes      0.007           0.952    0.918-0.986\n  pCR                    0.009           2.668    1.276-5.578\n  Operative time         0.010           1.006    1.001-1.01\n  Estimated blood loss   0.093           1.002    1-1.005\n\npCR: Pathologic complete response; OR: Odds ratio.\n\nPathologic outcomes\n-------------------\n\nSeventy-three (31.6%) patients in the cohort experienced downstaging of disease, which was defined as clinical TNM \\> pathologic TNM without pCR. Forty (17.3%) patients achieved a pCR. Unfortunately, 118 (51.1%) patients did not reach the point of downstaging. There was a significant difference in T downstaging (group A, 40.3% *vs* group B, 43.5%; *P* = 0.016). Patients who underwent surgery after a longer interval had a significantly higher rate of N downstaging than those who underwent surgery after a shorter interval following neoadjuvant therapy (group A, 28.1% *vs* group B, 44.6%; *P* = 0.010, Table [4](#T4){ref-type=\"table\"}).\n\n###### \n\nPathologic response\n\n  **Variable**                   **Interval \u2264 9 wk**   **Interval \\> 9 wk**   ***P* value**\n  ------------------------------ --------------------- ---------------------- ---------------\n  No downstaging, TNM (*n*, %)   83 (59.7)             35 (38)                0.001\n  Downstaging, TNM (*n*, %)      41 (29.5)             32 (34.8)              0.129\n  T, *n* (%)                                                                  0.016\n  No downstaging                 83 (59.7)             52 (56.5)              \n  Downstaging                    56 (40.3)             40 (43.5)              \n  N, *n* (%)                                                                  0.010\n  No downstaging                 100 (71.9)            51 (55.4)              \n  Downstaging                    39 (28.1)             41 (44.6)              \n\nTNM: Tumor node metastasis.\n\nSurvival outcomes\n-----------------\n\nThe follow-up period ranged from 2 to 62 mo, and it was not different between the two groups. Twelve (8.7%) patients in group A and 8 (8.6%) patients in group B experienced local recurrence (*P* = 0.987). The total number of deaths in the cohort was 26 (18.7%) from group A and 12 (13%) from group B (*P* = 0.661). The patients in group A experienced more distant recurrence than those in group B (group A, 28% *vs* group B, 10.9%, *P* = 0.001). Table [5](#T5){ref-type=\"table\"} presents the tumor outcomes between the interval groups. The impacts of potential prognostic factors on DFS and OS (age, sex, BMI, ASA score, clinical T stage, clinical N stage, surgical procedure, pathologic T stage, pathologic N stage postoperative complications, differentiation, and preoperative concurrent chemotherapy regimen) were examined with Kaplan-Meier curves. There were significant differences in the DFS curves between the two groups for pathologic T stage, pathologic N stage, and time interval (Figure [1](#F1){ref-type=\"fig\"}). There was no significant difference between the two OS curves for pathologic N stage or time interval (Figure [1](#F1){ref-type=\"fig\"}), whereas there was a significant difference in the two OS curves for postoperative complications, as shown in Table [6](#T6){ref-type=\"table\"}. According to Table [7](#T7){ref-type=\"table\"}, the time interval was found to be independently associated with DFS but not with OS (\\> 9 wk *vs* \u2264 9 wk: OR: 0.570; 95%CI: 0.328-0.991; *P* = 0.046). Meanwhile, pathologic T stage was an independent factor for DFS.\n\n![Kaplan-Meier curves of disease-free survival and overall survival in the \u2264 9-wk group and \\> 9-wk group with 5-year follow-up. DFS: Disease-free survival; OS: Overall survival.](WJG-26-4624-g001){#F1}\n\n###### \n\nPatterns of disease recurrence according to study group\n\n  **Variable**                  **Interval \u2264 9 wk**   **Interval \\> 9 wk**   ***P* value**\n  ----------------------------- --------------------- ---------------------- ---------------\n  Local recurrence, *n* (%)     12 (8.6)              8 (8.7)                0.987\n  Distant recurrence, *n* (%)   40 (28.8)             10 (10.9)              0.001\n  Death                         26 (18.7)             12 (13)                0.661\n\n###### \n\nKaplan-Meier survival estimates\n\n  **Variable**                                            **No. of cases**   **DFS *P* value**   **OS *P* value**\n  ------------------------------------------------------- ------------------ ------------------- ------------------\n  Sex (*n*, %)                                                               0.668               0.448\n  Male                                                    160 (69.3)                             \n  Female                                                  71 (30.7)                              \n  Age (year)                                                                 0.518               0.688\n  \\< 60                                                   111 (48.1)                             \n  \u2265 60                                                    120 (51.9)                             \n  ASA (*n*, %)                                                               0.691               0.658\n  1                                                       7 (3)                                  \n  2                                                       178 (77)                               \n  3                                                       46 (19.9)                              \n  BMI (kg/m^2^)                                                              0.759               0.587\n  \\< 25                                                   149 (64.5)                             \n  \u2265 25                                                    82 (35.5)                              \n  Interval time, w (%)                                                       0.002               0.259\n  \u2264 9 wk                                                  139 (60.2)                             \n  \\> 9 wk                                                 92 (39.8)                              \n  cT, *n* (%)                                                                0.343               0.483\n  2                                                       2 (0.9)                                \n  3                                                       180 (77.9)                             \n  4                                                       49 (21.2)                              \n  cN, *n* (%)                                                                0.059               0.084\n  0                                                       62 (26.8)                              \n  1                                                       119 (51.5)                             \n  2                                                       50 (21.6)                              \n  ypT, *n* (%)                                                               \\< 0.001            0.001\n  0                                                       40 (17.3)                              \n  1                                                       5 (2.2)                                \n  2                                                       48 (20.8)                              \n  3                                                       121 (52.4)                             \n  4                                                       17 (7.4)                               \n  ypN, *n* (%)                                                               0.001               0.185\n  0                                                       128 (55.4)                             \n  1                                                       74 (32)                                \n  2                                                       29 (12.6)                              \n  Surgical procedure                                                         0.109               0.464\n  Miles (*n*, %)                                          100 (43.3)                             \n  Dixon (*n*, %)                                          113 (48.9)                             \n  Hartmann (*n*, %)                                       18 (7.8)                               \n  Postoperative complications, *n* (%)                                       0.265               0.043\n  Yes                                                     12 (5.2)                               \n  No                                                      219 (94.8)                             \n  Degree of differentiation, *n* (%)                                         0.094               0.183\n  Low and low-moderate grades                             32 (13.9)                              \n  Moderate, moderate-high, and high grades                193 (83.5)                             \n  Signet-ring and mucinous adenocarcinoma                 6 (2.6)                                \n  Preoperative concurrent chemotherapy regimen, *n* (%)                      0.357               0.533\n  Capecitabine + oxaliplatin                              71 (30.7)                              \n  Capecitabine oral                                       145 (62.8)                             \n  Oxaliplatin union                                       15 (6.5)                               \n\nBMI: Body mass index; ASA: American Society of Anesthesiologists; DFS: Disease-free survival; OS: Overall survival.\n\n###### \n\nCox regression analysis\n\n  **Variable**                  **Disease-free survival**   **Overall survival**                                   \n  ----------------------------- --------------------------- ---------------------- --------- ------- ------------- -------\n  Interval time                                                                                                    \n  \\> 9 wk *vs* \u2264 9 wk           0.570                       0.328-0.991            0.046     0.825   0.411-1.656   0.589\n  Postoperative complications   1.322                       0.527-3.368            0.544     2.187   0.764-6.255   0.145\n  ypT                                                                                                              \n  ypT0 *vs* ypT4                0.047                       0.01-0.214             \\< 0.01   0.065   0.014-0.307   0.001\n  ypT1 *vs* ypT4                0.173                       0.022-1.337            0.093     0.253   0.031-2.031   0.196\n  ypT2 *vs* ypT4                0.102                       0.038-0.275            \\< 0.01   0.076   0.02-0.288    0.001\n  ypT3 *vs* ypT4                0.387                       0.205-0.731            0.003     0.287   0.128-0.641   0.002\n  ypN                                                                                                              \n  ypN0 *vs* ypN2                0.705                       0.35-1.421             0.329     0.765   0.282-2.077   0.599\n  ypN1 *vs* ypN2                0.967                       0.504-1.853            0.919     0.939   0.363-2.426   0.897\n\nOR: Odds ratio.\n\nDISCUSSION\n==========\n\nThe standard treatment regimen for clinical TNM stage II-III rectal cancers is neoadjuvant chemoradiotherapy followed by surgery with TME in China\\[[@B17],[@B18]\\]. The treatment regimen can effectively decrease the rates of local recurrence and distant metastasis and may increase the rate of DFS\\[[@B19]\\]. Our results suggested that a longer time interval (\\> 9 wk) resulted in a better pathologic response and more favorable DFS. The results also demonstrated that a high pCR rate (*P* = 0.009; OR: 2.668; 95%CI: 1.276-5.578) and downstaging were favorable prognostic factors. Similarly, in the pathologic response study, although the TNM downstaging rate did not reach statistical significance in the longer interval group, the data analysis presented a trend towards an increased TNM downstaging rate in patients who underwent surgery \\> 9 wk after chemoradiotherapy (group A, 29.5% *vs* group B, 34.8%, *P* = 0.129). Recently, an increasing number of studies have suggested that a longer time interval to surgery is associated with a high incidence of pCR. Our findings are consistent with this conclusion regarding pathologic response. Garcia-Aguilar et al\\[[@B20]\\] investigated the time interval to surgery in a 6-wk group and a longer time interval group. They compared the tumor response and surgical complications of the two groups. Finally, they concluded that a longer time interval to surgery may increase the pCR rate without increasing complications. de Campos-Lobato et al\\[[@B21]\\] described that perioperative complications were not affected by chemoradiotherapy or interval time. The pCR rate of patients with a longer time interval (\u2265 8 wk) after chemoradiotherapy was significantly improved (*P* = 0.03), and the 3-year local recurrence rate decreased (*P* = 0.04). Probst et al\\[[@B22]\\] collected 17255 patients from a national database and stated that a longer time interval to surgery of 8 wk improved the pCR rate. Similarly, a systematic analysis\\[[@B23]\\] found that 4 of 7 studies reported that the rate of pCR was significantly higher with a longer interval time, and 3 of 8 studies indicated that prolonging the time interval to surgery increased the rate of tumor downstaging. In the first phase of the Lyon\\[[@B10]\\] trial, which was reviewed above, the conclusion was that a prolonged time interval to surgery can lead to better pathologic downstaging, but it does not contribute much to the control of local recurrence. We believe that the differences in the results of this study may be because the neoadjuvant radiotherapy dose was small (39 Gy) and that there was no concentrated radiation on the tumor. In contrast, in 2016, Sun et al\\[[@B24]\\] found that downstaging was increasing in less than 8 wk after radiotherapy. Stein et al\\[[@B35]\\] concluded that prolonging the time interval to surgery after neoadjuvant therapy did not lead to tumor downstaging. Our study described a favorable prognosis of DFS in patients who underwent surgery more than 9 wk after neoadjuvant chemoradiotherapy. However, there was a significant balance in OS between the two groups. To some extent, in the second phase of the Lyon trial\\[[@B11]\\], the DFS and OS rates were similar in the two groups at 17 years of follow-up. Regarding the pCR rate results, our study result was also in agreement with that of the randomized trial by Akgun et al\\[[@B12]\\]. They also concluded that prolonging the interval (\\> 8 wk) to surgery after chemoradiotherapy could increase the pCR rate, with the highest increase at 10-11 wk. However, there was no information on DFS or OS outcomes. In many studies, lower pathologic T stage or N stage was associated with improving DFS and decreasing recurrence and was even a vital predictor of survival\\[[@B25],[@B26]\\]. Kim et al\\[[@B27]\\] reported that pathologic N downstaging was an important prognostic factor. Our study demonstrated that pathologic T downstaging and time interval delayed beyond 9 wk significantly improved the oncological outcomes of DFS and OS, and we found that there was a lower recurrence rate associated with delaying surgery. Our conclusion coincided with that of Tulchinsky et al\\[[@B28]\\], who found that a longer time interval (\\> 7 wk) to surgery was associated with a higher rate of pCR and near pCR (17% *vs* 35%, *P* = 0.03), decreased recurrence, and improved DFS (*P* = 0.05).\n\nSphincter preservation was performed in more than half of the patients in our study (56.7%), and there was no significant difference between the two groups (*P* = 0.161). The longer interval of neoadjuvant surgery had no influence on the rate of sphincter-saving procedures, confirming the findings of other studies. A meta-analysis published in 2017 involving 13 studies that registered 19652 patients found that there were no significant differences (*P* = 0.743) in sphincter-preserving procedures\\[[@B29]\\]. Conversely, Campbell et al\\[[@B30]\\] reported that in the longer time interval group (8-16 wk), 74 (65%) patients underwent abdominoperineal resections. However, the most recently published studies show no beneficial effect of longer intervals on sphincter preservation\\[[@B31],[@B32]\\].\n\nThe rate of postoperative complications was 12% in our study. According to the Clavien-Dindo classification\\[[@B33]\\], all complications were graded as grades 1 to 3. Anastomotic leakage occurred more often in the shorter interval (\u2264 9 wk) group than in the longer interval group (3.6% *vs* 2.2%). Three patients underwent reoperation in the interval \\> 9 wk group, and one patient underwent reoperation in the interval \u2264 9 wk group. The other types and rates of complications were similar between the groups in our analysis. However, many surgeons hold the view that further prolongation of the time interval to surgery will increase the risk of the operation because pelvic fibrosis caused by radiotherapy may lead to intraoperative or postoperative bleeding. According to our previous research, intraperitoneal chemotherapy can lead to postoperative anastomotic leakage, especially in patients undergoing preoperative neoadjuvant therapy\\[[@B34]\\]. In fact, Stein et al\\[[@B35]\\] reported that in the shorter interval group, the incidence of complications was higher than that in the longer time interval group. Three patients suffered from anastomotic leakage, and two patients healed by reoperation. In addition to the blood supply factors, leakage of the anastomotic site may be caused by injury to the intestinal canal at the anastomotic site caused by radiation irradiation. The Lyon R90-01 trial\\[[@B10]\\] compared the postoperative complications of the two groups and found that the incidence of complications was 18% in the 2-wk group and 17% in the 6-8 wk group, and there was no significant difference between the two groups. A total of 265 patients from 24 medical centers were randomly enrolled in the GRECCAR-6\\[[@B36]\\] trial, and the patients were divided into a 7-wk interval group and an 11-wk interval group. The study described that complications were significantly increased in the longer interval group, but there was no significant difference compared with the 7-wk interval group (*P* = 0.404); regarding the rate of anastomotic leaks, there was balance in the two groups (*P* = 0.611). The study concluded that a longer time interval (11 wk) after surgery may improve the risk of complications. We analyzed the data of Moore et al\\[[@B37]\\] and found a high incidence of anastomotic complications in the longer time interval group undergoing surgery after chemoradiation (*P* = 0.05). In summary, most studies suggested that a prolonged time interval may lead to higher complications and lower sphincter preservation.\n\nOur study has several limitations. The first limitation of this study is its retrospective nature. Due to the lack of data from large, prospective, randomized controlled studies, we believe that the results of this study are of value. The second may be bias in the information collected, but the homogeneity in the cohort was well compared. A future multicenter randomized control or cohort trial with a larger sample size may be needed to identify the optimal time interval for increasing the rate of pCR and improving oncologic outcomes. Third, the complication rate in our center is relatively low for neoadjuvant therapy in advanced rectal cancer.\n\nIn conclusion, a time interval to surgery of \\> 9 wk after neoadjuvant therapy increases the pCR rate and has an impact on DFS, but it does not have an impact on OS.\n\nARTICLE HIGHLIGHTS\n==================\n\nResearch background\n-------------------\n\nSurgery with total mesorectal excision following neoadjuvant therapy is a standard regime for locally advanced rectal cancer. The optimal interval time between neoadjuvant therapy and surgery is still under debate.\n\nResearch motivation\n-------------------\n\nThere is a lack of consensus concerning the interval time between neoadjuvant therapy and surgery. Whether shorter or longer interval time is a controversial topic. And there are limited data regarding outcomes associated with different neoadjuvant therapy-surgery times.\n\nResearch objectives\n-------------------\n\nThe main aim of this study was to investigate whether different interval times affect the rate of pathologic complete response (pCR), preoperative outcomes, and survival status.\n\nResearch methods\n----------------\n\nWe performed a retrospective cohort study and enrolled locally advanced rectal cancer patients with neoadjuvant therapy. Information regarding the clinicopathological features, clinical outcomes, and follow-up was collected and analyzed. Multivariate logistic regression analysis was performed to evaluate the possible factor affected by interval time.\n\nResearch results\n----------------\n\nThe interval time between neoadjuvant therapy and surgery \\> 9 wk increased the incidence of pCR and had a better impact on disease-free survival (DFS).\n\nResearch conclusions\n--------------------\n\nProlonging the interval time between neoadjuvant therapy and surgery may be associated with improved rates of pCR, decreased disease recurrence, and improved DFS but has little impact on postoperative complications and sphincter preservation.\n\nResearch perspectives\n---------------------\n\nProspective randomized trials are required to evaluate the optimal time interval that is needed to achieve minimum morbidity, maximal tumor downstaging, and minimum disease recurrence.\n\nInstitutional review board statement: Our investigation received approval from the ethics committee of the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College.\n\nInformed consent statement: All study participants or their legal guardian provided informed written consent about personal and medical data collection prior to study enrolment.\n\nConflict-of-interest statement: The authors declare that there is no conflict of interest in regard to this research.\n\nSTROBE statement: The authors have carefully read the STROBE statement checklist of items and prepared the manuscript based on the requirements of the STROBE statement checklist of items.\n\nManuscript source: Unsolicited manuscript\n\nPeer-review started: May 26, 2020\n\nFirst decision: June 4, 2020\n\nArticle in press: August 1, 2020\n\nSpecialty type: Oncology\n\nCountry/Territory of origin: China\n\nPeer-review report's scientific quality classification\n\nGrade A (Excellent): 0\n\nGrade B (Very good): 0\n\nGrade C (Good): C\n\nGrade D (Fair): 0\n\nGrade E (Poor): 0\n\nP-Reviewer: Mohammadianpanah M S-Editor: Zhang H L-Editor: Wang TQ P-Editor: Ma YJ\n\nData sharing statement\n======================\n\nNo additional data are available.\n\n[^1]: Author contributions: Mei SW, Chen JN, and Wang ZJ designed the research; Wei FZ, Shen HY, Li J, and Zhao FQ collected the data; Pei W, Wang Z, Liu Z, and Wei FZ analyzed the data; Mei SW drafted the manuscript; Liu Q and Wang XS revised the paper.\n\n    Supported by the National Key Research and Development Plan \\\"Research on Prevention and Control of Major Chronic Non-Communicable Diseases\\\", No. 2019YFC1315705; and the Medicine and Health Technology Innovation Project of Chinese Academy of Medical Sciences, No. 2017-12M-1-006.\n\n    Corresponding author: Qian Liu, MD, Chief Doctor, Professor, Surgeon, Department of Colorectal Surgery, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, No. 17 Panjiayuan Nanli, Chaoyang District, Beijing 100021, China. <fcwpumch@163.com>\n"}
+{"text": "In a recent study Soon et al. ([@B16]) predicted abstract intentions from fMRI BOLD activities in localized areas of the brain. Activities in a spherical cluster of voxels served as input to a multivariate pattern classifier (linear SVM). The accuracy for predicting the intention to add or subtract two numbers was determined for clusters centered on different voxels. A prediction accuracy of 60% averaged across participants and based on 10-fold cross-validation was achieved for patterns of voxel activities in the *medial frontopolar cortex* and *precuneus* up to 4 s *before* participants reported being consciously aware of their decision. The prediction accuracy in this study was similar to studies on predicting spontaneous left or right motor decisions (Soon et al., [@B15]; Bode et al., [@B1]). Since the task demands placed on the participants create similar methodological issues as in previous studies (Lages and Jaworska, [@B10]), it seems possible that the multivariate classifier picked up sequential information processing between trials (Bode et al., [@B2]).\n\nAlthough the average prediction accuracy of 60% returned to chance level for patterns of voxel activity in the two brain areas shortly after the onset of a new trial and remained at 50% between trials, this observation is neither necessary nor sufficient for the *absence* of sequential information processing. In order to investigate sequential dependencies the outcome of at least one preceding trial and the current trial needs to be taken into account. Depending on task and response, sequential information processing between trials may emerge in distributed form within the default mode network (DMN) at variable time points (Guggisberg and Mottaz, [@B6]). In the following we illustrate the issue and suggest how the data may be analyzed.\n\nTo illustrate sequential effects consider the following inconspicuous sequence of ten responses (S, A, A, A, S, A, S, A, S, S) where A and S stand for freely choosing addition and subtraction, respectively. There are five As and five Ss suggesting a binomial process with rate parameter *p* = 0.5. However, if we consider the nine subsequent pairs of responses {(S,A), (A,A), ..., (S,S)} then we obtain unequal transition probabilities. A trained classifier that predicts the next response from the preceding response would be 3 out of 5 times or 60% correct if the preceding response is A and 3 out of 4 times or 75% correct if the response is S. Starting with a random guess in the first trial, this gives an average prediction accuracy of 65%. In two behavioral studies replicating two different choice tasks (Haynes et al., [@B7]; Soon et al., [@B15]) we used subsequent response pairs to train a linear classifier (SVM) and obtained an average prediction accuracy of 61.6 and 64.1%, respectively (Lages and Jaworska, [@B10]).\n\nWhen asked to generate a random sequence, people typically alternate between binary responses with a probability of about 0.6 (Lopes, [@B14]). This response pattern appears to be relevant in Soon et al.\\'s study ([@B16]) since the only behavioral evidence for memory-less choice in the 17 (out of 34) selected participants is a histogram plotting average frequencies for different lengths of response sequences fitted by an exponential distribution (Figure S1). The authors take the excellent fit as evidence for random performance. Recently Allefeld et al. ([@B1a]) released a detailed account of the behavioral data but there are no further details how the data were compiled and fit. Nevertheless, it is discernable from their Figure S1 that the fit represents an *exponential function* with two parameters rather than an *exponential distribution* with a single parameter and that observed frequencies do not add up to probability 1.0. In addition, an exponential distribution would only provide a meaningful approximation of the geometrically distributed phase lengths if choosing addition and subtraction were equally probable \\[*p* = (1\u2212p) = 0.5\\]. However, even with a best-fitting rate parameter of 1-exp(\u22120.826) \u22480.56 the exponential distribution underestimates the relative frequency of alternations (A,S) and (S,A) with phase length 1 as well as repetitions (A,A,S) and (S,S,A) with phase length 2 (see Figure [1](#F1){ref-type=\"fig\"}). Increased frequencies for short phase lengths are a hallmark of non-random human choice behavior (Wagenaar, [@B21]; Lopes, [@B14]; Treisman and Faulkner, [@B19]; Falk and Konold, [@B4]) and these characteristics are not only present in the behavioral data of Soon et al. ([@B16]) but also in Soon et al. ([@B15]); Bode et al. ([@B1]), and Haynes et al. ([@B7]) suggesting that free or spontaneous choice tasks result in non-random behavior.\n\n![**Histogram for length of response sequences (phase length or runs) re-plotted as relative frequencies (adapted from Figure S1 in Soon et al., [@B16]).** The data points are fitted by an exponential function with two parameters (red curve) and an exponential distribution with a single parameter (blue curve). The red curve provides an excellent fit (*R*^2^ = 0.99) whereas the blue curve underestimates phase length 1 and 2 (*R*^2^ = 0.77). See text for explanation.](fpsyg-04-00925-g0001){#F1}\n\nA related concern arises from the searchlight analyses. If patterns of voxel activities are analyzed within a moving spherical cluster to predict behavioral responses then pre-processing of the data and definition of the searchlight are important (Etzel et al., [@B3]; Todd et al., [@B18]). The implementation of regions of interest, temporal constraints (hemodynamic delay), pre-processing (covariates), and data selection can invalidate the results of a searchlight analysis (Kriegeskorte et al., [@B8]; Lindquist et al., [@B13]). In Soon et al. ([@B16]) trials were selected (undersampled) in order to balance the mean response rate. It is therefore possible that the searchlight found a cluster of voxels that was predictive of the next response in the context of the preceding response, simply because transitions between successive responses remained unbalanced. A repeated choice task with self-monitoring of the decision process invites sequential dependencies because the observer has to remember goals, constraints, and execution of the task. If, for example, participants shift a decision criterion following each response (Lages and Treisman, [@B12], [@B9]; Lages and Paul, [@B11]; Treisman and Lages, [@B20]) or engage in metacognition by recalling the last response before making the next then neural correlates of these response-dependent processes introduce a confound that would be picked up by a searchlight analysis as soon as transitions between response categories are unbalanced.\n\nWe recommend that rather than postulating a 50% chance level, prediction accuracy should be tested with a permutation test (Stelzer et al., [@B17]) and/or separate multivariate classification analyses conditional on the previous response. Only if individual prediction accuracies reliably exceed observable benchmarks such as response bias and transition probabilities can we rest assured that results are not confounded. The interested reader is invited to test predictability of their own free choice behavior by downloading the demo program in the Appendix.\n\nMartin Lages was supported by a research grant from The Leverhulme Trust F00-179/BG (UK). Stephanie Boyle and Katarzyna Jaworska were both funded by BBSRC WestBio DTP studentships.\n\n*After the game is before the game---*Josef \"Sepp\" Herberger (1897-1977, football player and manager)\n\nThe demo program can be downloaded from <http://www.psy.gla.ac.uk/~martinl/Assets/Software/Predictatron.txt>\n\nPlease save the text file as Predictatron.m before running it under MatLab (MathWorks Inc., Natick MA). The program illustrates the difficulty for a human decision maker to generate a truly random sequence in a free choice task. Different choice tasks may be simulated by freely choosing between addition/subtraction (Haynes et al., [@B7]; Soon et al., [@B16]), left/right (Soon et al., [@B15]; Bode et al., [@B1]), or delay/non-delay (Filevich et al., [@B5]) before pressing the corresponding left or right arrow key on the keyboard. The program records 100 binary responses before it computes the rate of left/right key presses, number of left/right alternations, and left/right repetitions. Then the program determines the prediction accuracy of a linear support vector machine (SVM) using one preceding response to predict the next response (1-back SVM) for 10-fold cross-validation. If the algorithm does not exceed 60% prediction accuracy then two preceding responses are used to predict the next response (2-back SVM). If both prediction accuracies stay below 60% then the participant is considered \"unpredictable\" and wins against the \"Predict-a-tron\". Results of the non-parametric Wald-Wolfowitz `runstest()` and parametric Lilliefors `lillietest()` as implemented in MatLab are also reported.\n\n[^1]: This article was submitted to Decision Neuroscience, a section of the journal Frontiers in Psychology.\n\n[^2]: Edited by: Bj\u00f6rn Brembs, University of Regensburg, Germany\n\n[^3]: Reviewed by: Allen Neuringer, Reed College, USA; Alexander Maye, University Medical Center Hamburg-Eppendorf, Germany\n"}
+{"text": "1. Background {#sec1-ijms-18-02618}\n=============\n\nCharacterizing the organization of an organism into autonomous genetic and phenotypic modules has gained strong theoretical and empirical interest \\[[@B1-ijms-18-02618],[@B2-ijms-18-02618],[@B3-ijms-18-02618],[@B4-ijms-18-02618]\\]. Such modular organization helps to understand the evolutionary dynamics of complex organisms given that the compartmentalization of a function (or of a set of functions) will permit this functional aspect to evolve independently of how selection acts on other functional aspects of the organisms (i.e., it alleviates the so-called \"cost of complexity\" or \"cost of pleiotropy\" \\[[@B5-ijms-18-02618],[@B6-ijms-18-02618],[@B7-ijms-18-02618]\\]). However, understanding the evolution of such functional modules requires a characterization of the interactions between the different elements within a module and how these interactions determine their joint evolution \\[[@B1-ijms-18-02618],[@B2-ijms-18-02618]\\]. Recently-developed computational tools can complement our understanding of molecular evolution within a genetic system or network by estimating the probability that a nucleotide site coevolved with a site in another gene across a phylogenetic tree \\[[@B8-ijms-18-02618],[@B9-ijms-18-02618]\\]. The use of such tools is a major difference in relation to previous studies, where only the coevolution within a protein or gene sequence (e.g., due to structural constraints) could be considered \\[[@B10-ijms-18-02618]\\].\n\nThe melanocortin system is a key hormonal pathway that exhibits the features of a module. It is composed of a set of G protein-coupled membrane receptors (MC1-5R) responsible for the regulation of very distinct functions in vertebrates: from pigmentation to metabolic homeostasis and sexual behavior ([Figure 1](#ijms-18-02618-f001){ref-type=\"fig\"}) \\[[@B11-ijms-18-02618],[@B12-ijms-18-02618],[@B13-ijms-18-02618],[@B14-ijms-18-02618],[@B15-ijms-18-02618],[@B16-ijms-18-02618],[@B17-ijms-18-02618]\\]. The activity of the different receptors and their associated functions is under the control of a shared set of agonist and antagonist ligands. The upregulation of the distinct receptors is achieved by binding one of the melanocortin hormones that are derived from the same prohormone: the pro-opiomelanocortin, POMC ([Figure 1](#ijms-18-02618-f001){ref-type=\"fig\"}) \\[[@B14-ijms-18-02618],[@B15-ijms-18-02618],[@B18-ijms-18-02618]\\]. The proprotein convertase, PC1/3, encoded by the gene *PCSK1*, cleaves the POMC prohormone into the adrenocorticotropic hormone, ACTH, which, in turn, can be cleaved by the convertase PC2 (*PCSK2*) to release the melanocortin \u03b1-MSH (Melanocyte Stimulating Hormone \\[[@B18-ijms-18-02618],[@B19-ijms-18-02618]\\]. \u03b2-MSH as well as other melanocortins (\u03b3-, and \u03b4-MSH) are present only in certain vertebrate taxa \\[[@B11-ijms-18-02618]\\] and are also derived from POMC by the cleaving action of PC1 and PC2 \\[[@B20-ijms-18-02618]\\]. Activation of a melanocortin receptor has several cascading effects by raising intracellular cAMP levels \\[[@B21-ijms-18-02618]\\]. In contrast, down-regulation of the MC receptors activity, is triggered by binding the antagonists and inverse agonists: the agouti-signaling protein (ASIP, or ASIP1 the fish orthologue of the mammalian ASIP \\[[@B17-ijms-18-02618]\\]) and the agouti-related protein (AgRP), in a tissue and melanocortin-receptor dependent manner \\[[@B22-ijms-18-02618]\\].\n\nThe evolution of the different genes in the melanocortin system has been previously studied \\[[@B11-ijms-18-02618],[@B23-ijms-18-02618],[@B24-ijms-18-02618],[@B25-ijms-18-02618]\\]. However, to what extent the genes in the melanocortin system have influenced each other's evolution has never been assessed. On the one hand, we can expect genes in the melanocortin system to have coevolved to a certain extent owing to selection for module conformation and/or maintenance \\[[@B1-ijms-18-02618],[@B7-ijms-18-02618]\\]. In particular, the evolution of the regulatory genes (e.g., *POMC* or *ASIP*) of the melanocortin system may have favored further changes in one or several receptor sequences to induce, for instance, compensatory or co-adaptive effects in relation to the changes in the regulatory gene sequence \\[[@B26-ijms-18-02618],[@B27-ijms-18-02618]\\]. Changes in a melanocortin receptor may have also favored evolutionary changes in other receptors owing to selection for co-adaptation through the functions controlled by several receptors \\[[@B26-ijms-18-02618]\\]. On the other hand, other factors, such as pleiotropy, are expected to limit coevolutionary changes \\[[@B6-ijms-18-02618],[@B28-ijms-18-02618],[@B29-ijms-18-02618]\\]. In this sense, we could expect that the regulatory genes of the melanocortin system, which have more pleiotropic effects by acting on all the different receptors ([Figure 1](#ijms-18-02618-f001){ref-type=\"fig\"}), might have been under higher evolutionary constraints. This might have limited their influence on the evolution of other melanocortin genes and reduce the extent with which coevolution occurred among genes in the system.\n\nHere, we estimated the selective and coevolutionary processes that occurred during the evolution of the melanocortin system. The widespread occurrence and preserved functions of these genes across vertebrates make it a highly suitable system to investigate how and to what extent genes may function within a module coevolve. We focused on the 10 major genes of the melanocortin system (*POMC*, *PCSK1*, *PCSK2*, MC1-5R, *ASIP*, and *AGRP*), whose sequences were obtained for a total of 138 species representing the main vertebrate lineages (16 birds, 2 snakes, 7 lizards, 1 turtle, 1 monotreme, 3 marsupials, 81 placental mammals, 3 amphibians, 1 coelacanth, 20 teleost fish, 2 sharks, and 1 lamprey species: [Supplementary Materials Figure S1](#app1-ijms-18-02618){ref-type=\"app\"}). We first assessed the evolutionary dynamics of the genes in the melanocortin system by comparing the level of sequence conservation and the proportion of codons evolving neutrally or under positive or purifying selection. This allowed us to understand the genes in the system that are evolutionarily more constrained and to associate this with the coevolutionary dynamics between genes. We then estimated to what extent the genes in the melanocortin system have coevolved by estimating the number of sites that show evidence of correlated evolution \\[[@B8-ijms-18-02618],[@B9-ijms-18-02618]\\]. We focused on sites within codons evolving under positive selection and inducing coevolutionary changes between genes to better understand the importance that selection had on the distinct genes of the melanocortin system (and possibly on the functional aspects these genes regulate) on determining coevolutionary dynamics within the system.\n\n2. Results and Discussion {#sec2-ijms-18-02618}\n=========================\n\n2.1. Molecular Evolution of the Genes in the Melanocortin System {#sec2dot1-ijms-18-02618}\n----------------------------------------------------------------\n\nGenes belonging to the melanocortin system have been traditionally suspected to be highly conserved across vertebrates, although empirical support for this idea has been based on the sequences of a few vertebrate species \\[[@B11-ijms-18-02618],[@B25-ijms-18-02618]\\]. Using a wider coverage of the vertebrate tree than in previous studies, our results showed that the genes of the melanocortin system have indeed a high level of sequence identity across vertebrates (mean \u00b1 SD: 77.8% \u00b1 3.6, [Figure 2](#ijms-18-02618-f002){ref-type=\"fig\"}a). Similar levels of identity have been observed for other genes such as those associated with the regulation of general vertebrate development \\[[@B30-ijms-18-02618],[@B31-ijms-18-02618]\\] and it is in line with the conserved functions of the melanocortin system across vertebrates \\[[@B11-ijms-18-02618],[@B13-ijms-18-02618]\\]. Although the differences between genes in sequence identity were not major (maximal difference observed was 12%, [Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}), among the most evolutionary conserved genes were the proopiomelanocortin gene (*POMC*) and its two associated convertases, *PCSK1* and *PCSK2* ([Figure 2](#ijms-18-02618-f002){ref-type=\"fig\"}a). This suggests that the genes involved in the production and post-processing of the receptor agonist ligands may be subjected to stronger evolutionary constraints than other genes in the system, which may be the consequence of their higher number of pleiotropic effects \\[[@B6-ijms-18-02618],[@B13-ijms-18-02618]\\]. However, this contrasts with the fact that the genes encoding the antagonistic ligands (*ASIP* and *AGRP*) were among the least conserved genes, even though they are also expected to mediate some pleiotropic effects by regulating the activity of the different receptors \\[[@B13-ijms-18-02618],[@B14-ijms-18-02618]\\].\n\nCongruent with the high level of sequence conservation, the MEME analysis indicated that a large proportion of the codons were evolving under purifying selection (mean \u00b1 SD 90.7% \u00b1 3.6), whereas a low proportion were under positive selection (2.6% \u00b1 1.6 SD) or neutral evolution (6.7 \u00b1 3.9 [Figure 2](#ijms-18-02618-f002){ref-type=\"fig\"}b, [Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}). The proportion of codons under purifying selection across the genes in the melanocortin system was strongly and negatively correlated with the proportion of codons under positive selection (Spearman's \u03c1 = \u22120.80, *N* = 10, *P* = 0.005) or evolving neutrally (Spearman's \u03c1 = \u22120.95, *N* = 10, *P* \\< 0.0001). This suggests that the evolution of some genes in the melanocortin system has been less constrained, although all the genes exhibited values within the same order of magnitude and, as suggested by the similar levels of sequence identity, major differences among the genes of the melanocortin system in their evolutionary dynamics do not seem to exist ([Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}). The genes exhibiting a higher proportion of codons under positive selection and neutral evolution were those coding for the antagonist receptor ligands: *ASIP*, and *AGRP*, and the *MC1R*. Both, *ASIP* and the *MC1R* are key regulators of melanin synthesis in vertebrates \\[[@B32-ijms-18-02618],[@B33-ijms-18-02618]\\] and have been observed to mediate adaptive color evolution in several vertebrate species \\[[@B34-ijms-18-02618],[@B35-ijms-18-02618],[@B36-ijms-18-02618]\\]. Although *ASIP* and, to a minor extent, *MC1R* have functions outside melanogenesis \\[[@B37-ijms-18-02618],[@B38-ijms-18-02618],[@B39-ijms-18-02618],[@B40-ijms-18-02618]\\], we can hypothesize that vertebrate coloration may be the functional component of the melanocortin system that can be freer to vary and/or have mediated adaptation more frequently than the other functional aspects controlled by the melanocortin system. The *POMC* gene exhibited a relatively high proportion of codons evolving neutrally and the lowest proportion of codons evolving under purifying selection, which contrasts with the relatively higher level of identity of its sequence compared to other genes ([Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}). However, it is important to consider that, because of their high degree of conservation among vertebrates, only one third of the *POMC* codons had sufficient statistical signal to be assigned by MEME into one of the three categories of selective pressures ([Figure 2](#ijms-18-02618-f002){ref-type=\"fig\"}b, [Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}).\n\n2.2. Molecular Coevolution among the Melanocortin System Genes {#sec2dot2-ijms-18-02618}\n--------------------------------------------------------------\n\nWe estimated coevolution as the joint substitutions between nucleotide sites in two different protein sequences following the simulation procedure described in Dib et al. \\[[@B9-ijms-18-02618]\\]. Using this method, we estimated the number of significant coevolving nucleotide pairs and, separately, the number of significant coevolving nucleotide pairs that were included in codons under positive selection ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"} and [Table 2](#ijms-18-02618-t002){ref-type=\"table\"}, [Supplementary Materials Tables S3 and S4](#app1-ijms-18-02618){ref-type=\"app\"}). These sites may be more indicative of adaptive coevolution between genes rather than of constraints due to, for instance, receptor-ligand specificity. Using a \u0394AIC threshold representing the 95th percentile (see Methods), the mean (\u00b1S.D.) number of coevolving nucleotide pairs was 1071.51 \u00b1 526.7, whereas the mean (\u00b1S.D.) number of coevolving pairs with at least one of the nucleotides in the pair being in a codon subjected to positive selection was 181.2 \u00b1 90.23 ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"} and [Table 2](#ijms-18-02618-t002){ref-type=\"table\"}). Using the 90th or the 97.5th percentiles as thresholds resulted in a larger and a smaller number of coevolving sites detected, respectively ([Supplementary Materials Tables S3 and S4](#app1-ijms-18-02618){ref-type=\"app\"}). Nevertheless, the number of coevolving nucleotide pairs detected for each gene, as well as the number of these coevolving pairs including at least one of the nucleotides within a codon under selection, were highly correlated among the different thresholds used (all Spearman's \u03c1 \\> 0.73, all *P* \\< 0.02).\n\nConsistently across the three thresholds used, the genes that had a significantly higher frequency of coevolving pairs (including nucleotides in codons under selection) compared to all the other genes in the system were the antagonist *ASIP*, *AGRP*, and the receptors *MC1R*, and *MC3R* ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}). The number of coevolving pairs can consider the same sites several times when these have induced evolutionary changes in different sites in another gene ([Supplementary Materials Figure S2](#app1-ijms-18-02618){ref-type=\"app\"}). When considering only the number of distinct sites in a gene sequence that belongs to a codon under selection and that induced an evolutionary response in other genes, the list of genes highlighted above was narrowed down to *ASIP* and *MC1R* ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}). Those genes with a significantly lower frequency of coevolving pairs (including nucleotides in codons under selection) across all the thresholds were the ligand-related genes *POMC*, *PCSK1*, *PCSK2* and the receptor *MC4R* ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}). When also considering the number of different sites, only *PCSK2* was highlighted using the three different thresholds ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}).\n\nThe coevolutionary effects of the genes in the melanocortin system are therefore in line with the differences among these genes at the level of genetic identity and selection. Genes producing agonist ligands (*POMC*, *PCSK1*, *PCSK2*) seem to be more conserved and also to induce less evolutionary changes in other melanocortin genes. Contrarily, genes, such as *ASIP* and *MC1R*, which are less conserved and may have evolved more frequently under selection or by neutral means seem to induce more evolutionary changes in the other melanocortin genes in response to positive selection. This is in part supported by a significant, negative association between sequence identity and the percentage of sites in a gene sequence that induced evolutionary changes in other genes (for all thresholds, Spearman's \u03c1 \\> \u22120.69, all *P* \\< 0.027). In contrast, the percentage of codons under positive selection, purifying selection, or evolving neutrally was not consistently correlated with any parameters of gene coevolution measured using the different thresholds (the correlation between the percentage of sites in a gene sequence that induced evolutionary changes in another gene and the percentage of sites under positive or purifying selection were significant only when using the less restrictive 90th percentile: Spearman's \u03c1 = 0.66, *P* = 0.038, \u03c1 = \u22120.68, *P* = 0.03, respectively). This finding suggests that the selection exerted on individual genes is not associated with the extent to which they influence other genes in the system. However, the relatively low number of genes in the melanocortin system limit the power to assess this confidently using this system. Thus, studies based systems or networks enclosing a larger number of genes may be more suitable in testing to what extent evolutionary patterns of individual genes are linked to their coevolutionary dynamic with other genes.\n\nIn addition to measuring the magnitude of coevolutionary effects that a given gene may have towards other genes in the system, it is also relevant to consider the frequency with which such coevolutionary processes have occurred across the vertebrate tree (i.e., the percentage of branches of the vertebrate tree where genes of the melanocortin system coevolved as a result of positive selection in one gene; [Supplementary Materials Table S5](#app1-ijms-18-02618){ref-type=\"app\"}). We observed that sites under positive selection in the *MC1R* gene have induced coevolutionary changes in other genes more frequently than any of the other genes in the melanocortin system (10.2%, mean across all of the genes of the melanocortin system: 6.3% \u00b1 2.9 S.D., [Supplementary Materials Table S5](#app1-ijms-18-02618){ref-type=\"app\"}). On the other hand, the lowest frequency (1.5%) was observed for the *POMC* gene. Surprisingly, *PCSK2*, which hosts a low number of coevolving pairs and sites (see paragraph above), has mediated coevolutionary effects with a frequency (9.9%) similar to that of the *MC1R*. This evidences that the coevolutionary action of a gene can be also based on a few sites having influenced the evolution of other genes repeatedly across evolutionary time.\n\nWe applied an ordination method (Principal Component Analysis, PCA) to assess how distant the genes in the melanocortin system are from each other when accounting for the different parameters that assessed their coevolutionary influence onto other genes. We measured this as the number of pairs containing a nucleotide in a codon under positive selection that induce coevolution in other genes, the number of different sites containing a nucleotide in a codon under positive selection that induce coevolution in other genes, and the percentage of branches in the vertebrate tree where positive selection induced coevolution in other genes. For all the thresholds, only the first axis was retained owing to its eigenvalue greater than 1 \\[[@B41-ijms-18-02618]\\] (explained variance: 58.5, 67.9, 79.6 for the 90th, 95th, and 97.5th percentile thresholds, respectively). For all the thresholds, the variables included in the PCA loaded positively, indicating that positive scores on the PC1 was associated with a higher coevolutionary influence in terms of the number of pairs and different sites inducing coevolution in other genes, and the percentage of branches where a coevolutionary effect of a gene has been observed. All the PC1 derived using the data estimated with the different thresholds ordered *POMC* and *MC1R* as the most negative and positive extremes, respectively ([Figure 3](#ijms-18-02618-f003){ref-type=\"fig\"}). This indicates that, when taking all the parameters into account, *POMC* and *MC1R* can be distinguished as the genes within the melanocortin having less and more general coevolutionary effects, respectively ([Figure 3](#ijms-18-02618-f003){ref-type=\"fig\"}).\n\n2.3. Coevolution of the Melanocortin System {#sec2dot3-ijms-18-02618}\n-------------------------------------------\n\nOur study provides unprecedented insights into the evolution of a genetic system across vertebrate evolution shedding some light on the evolution of modular organization from a molecular perspective. The analysis of the coevolutionary patterns among the genes of the vertebrate melanocortin system suggests that genes differ to a certain extent on their relative influence towards other genes. Considering the different aspects that determine the coevolutionary influence that sites on a gene subjected to positive selection have on other genes, our study highlights the genes *MC1R* and *POMC* as those having a larger and a smaller coevolutionary influence on other genes of the melanocortin system, respectively. *POMC* is an upstream gene that is key for the regulation of the multiple functions controlled by the melanocortin system \\[[@B13-ijms-18-02618],[@B25-ijms-18-02618]\\] and it is therefore expected to influence the different elements in the system at several levels. In this sense, the relatively lower coevolutionary influence of the sites at the *POMC* subjected to positive selection towards other genes is somehow surprising. This could be explained by the fact that *POMC* mainly interacts with the other elements of the system through the small and highly conserved melanocortin peptides ([Figure 1](#ijms-18-02618-f001){ref-type=\"fig\"}) \\[[@B25-ijms-18-02618]\\] and may offer few chances for inducing coevolution at the other genes.\n\nContrarily to *POMC*, *MC1R* mainly functions in a single context, melanogenesis \\[[@B42-ijms-18-02618]\\]. Although it is also involved in regulating the inflammatory response \\[[@B43-ijms-18-02618]\\] and mechanisms of analgesia \\[[@B39-ijms-18-02618]\\], its main function in regulating the synthesis of melanin pigments allows us to hypothesize that coloration is, among all the functional aspects controlled by the melanocortin system (e.g., anti-inflammatory activity, energy homeostasis, aggressive behaviour, sexual activity, and social behaviour; reviewed in \\[[@B13-ijms-18-02618]\\]), the one that has been more influential in the coevolution of all the genes in the system. A number of studies have identified different *MC1R* alleles associated with differences in coloration within or between populations of the same species \\[[@B44-ijms-18-02618]\\], and there is a clear interest in better understanding the evolutionary relevance of the sites harboring these mutations \\[[@B45-ijms-18-02618],[@B46-ijms-18-02618]\\]. However, it should be noted that our study focused on the long-term evolutionary processes that have led to the divergence of the melanocortin genes across vertebrates and, consequently, it does not account for sequence variation within species. It is therefore unlikely that the same sites would be involved at the two different scales (i.e., within and among species) because the mechanisms that lead to the fixation of specific alleles in species and their maintenance through time are certainly different from the selective forces acting within populations of a species. Indeed, we detected very little overlap between the sites identified among species in our study and the sites known to be involved at the intraspecific level \\[[@B46-ijms-18-02618]\\]. Bridging the two evolutionary scales would require a much denser sampling within and across species, in order to track the fixation process of specific alleles and their divergence between closely related species.\n\nAlthough we restricted our analysis to codons under positive selection so that the coevolutionary dynamics among genes in the melanocortin system can be more easily understood from the prism of adaptive and co-adaptive changes, it is difficult using our purely molecular approach to envision the rationale for the coevolutionary changes that the *MC1R* is inducing in other genes of the melanocortin system. Distinct hypotheses can explain why *MC1R* is inducing co-evolutionary changes in other genes of the melanocortin system although specific empirical studies will be needed to test how changes in the *MC1R* sequence influence other functional aspects of the melanocortin system. Regardless of the threshold used, we observed that *MC1R* induced coevolutionary changes mainly on the genes coding for the other melanocortin receptors ([Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}). The fact that positive selection on the *MC1R* gene induced evolutionary responses on the other receptors supports the idea of modularity underlying complex adaptations based on multiple phenotypic aspects \\[[@B47-ijms-18-02618]\\] Alternatively, it has been shown that the formation of MC1R-MC5R heterodimers is of functional relevance for the regulation of coloration in fish by altering MC1R sensitivity to its ligand, \u03b1-MSH \\[[@B48-ijms-18-02618]\\] and similar heterodimers may exist between MC1R and, at least, MC3R \\[[@B49-ijms-18-02618]\\]. The formation of such heterodimers may allow for coevolution among the melanocortin receptors, given that selection on the *MC1R* in relation, for instance, to adaptive coloration may require changes in the other receptors so that heterodimers can still be formed with the MC1R.\n\n3. Material and Methods {#sec3-ijms-18-02618}\n=======================\n\n3.1. Data Set {#sec3dot1-ijms-18-02618}\n-------------\n\nOur study comprises a total of 138 species representing the main vertebrate lineages (16 birds, 2 snakes, 7 lizards, 1 turtle, 1 monotreme, 3 marsupials, 81 placental mammals, 3 amphibians, 1 coelacanth, 20 teleost fish, 2 sharks, and 1 lamprey species: [Supplementary Materials Figure S1](#app1-ijms-18-02618){ref-type=\"app\"}). Complete coding sequences of the five melanocortin receptors (*MC1-5Rs*), the agonist ligand (*POMC*), antagonists (*ASIP* and *AGRP*), and convertases (*PCSK1* and *PCSK2*) were downloaded from Ensembl (<http://www.ensembl.org>) and NCBI (<http://www.ncbi.nlm.nih.gov>) databases on 02.04.2012 (see also [Supplementary Materials Figure S1 and Table S2](#app1-ijms-18-02618){ref-type=\"app\"}). The melanocortin receptors expanded early during vertebrate evolution, most likely as a consequence of two rounds of duplication, followed by a single gene duplication event that led to the MC5R \\[[@B50-ijms-18-02618]\\]. The most ancient melanocortin receptor is suspected to belong to the sea and river lampreys (*MCa* and *MCb*, respectively) \\[[@B50-ijms-18-02618],[@B51-ijms-18-02618]\\]. Some authors have proposed that *MCa* and *MCb* receptors represent two branches of the ancestral melanocortin receptors and that vertebrate *MC1R* and *MC2R* emerged from an ancestral MCa receptor, while *MCb* gave birth to the *MC3R* and a *MC4/5R* gene that later split into *MC4R* and *MC5R* \\[[@B24-ijms-18-02618],[@B50-ijms-18-02618]\\]. Based on this interpretation, the lamprey's *MCa* and *MCb* have been integrated as outgroups into the *MC1-2R* and *MC3-5R* data sets, respectively. Melanocortin receptors seem to have experienced duplication events followed by subfunctionalisation in certain fish species \\[[@B50-ijms-18-02618],[@B52-ijms-18-02618],[@B53-ijms-18-02618]\\]. In those species with paralogous sequences, we excluded the sequence that was non-orthologous to the mammalian gene copies.\n\nThe *POMC* gene has been reported in invertebrates, suggesting that POMC-derived peptides probably appeared before the split between protostomes and deuterostomes \\[[@B25-ijms-18-02618],[@B52-ijms-18-02618],[@B54-ijms-18-02618],[@B55-ijms-18-02618]\\]. Unfortunately, no sequences from invertebrates are available. Lampreys possess two copies of *POMC*, namely *POM* and *POC*, that we did not include into our data set because they likely experienced subfunctionalization \\[[@B25-ijms-18-02618]\\]. We kept the paralog sequences of *Danio rerio*, *Gasterosteus aculeatu*, *Oreochromis niloticus* and *Oryzias latipes* species because they are fully functional copies of *POMC* that resulted from a latter duplication. Out of the 62 sequences available in Ensembl, 40 *POMC* sequences were finally kept for the analysis.\n\nWe translated the nucleotide sequence for the melanocortin receptors, *ASIP* and *AGRP* genes into amino acids before aligning them using the software Muscle \\[[@B56-ijms-18-02618]\\]. Once aligned, we translated the sequences of amino acids back into nucleotide sequences for the analysis. The alignments of the coding sequences of the *POMC* and the convertases *PCSK1* and *PCSK2* were directly exported from Ensembl (ENSGT00390000016811, Euteleostomi.001). The alignments were filtered to exclude partial sequences and incomplete sequences. To perform the analyses of coevolution between pairs of genes, we concatenated the alignments of two genes from the same species ([Supplementary Materials Table S2](#app1-ijms-18-02618){ref-type=\"app\"}). Full sequences for all the 10 genes were available for only a few vertebrate species and, consequently, the number of genes varies from one species to another. *MC1R* is the most studied gene with a large number of sequences available across the vertebrates ([Supplementary Materials Figure S1](#app1-ijms-18-02618){ref-type=\"app\"}). However, the availability of *MC1R* sequences should not bias our results because the sampling level across vertebrates is consistent for each gene and all genes had sufficient numbers of sequences to have a correct power for the analyses performed \\[[@B57-ijms-18-02618]\\]. For the coevolution analyses, we further reduced the concatenated alignments to include only the species that had both genes present in Ensembl ([Supplementary Materials Table S2](#app1-ijms-18-02618){ref-type=\"app\"}), which removed any effects of the larger sampling available for *MC1R*.\n\n3.2. Phylogeny {#sec3dot2-ijms-18-02618}\n--------------\n\nWe performed all analyses on the species tree of the organisms sampled ([Supplementary Materials Figure S1](#app1-ijms-18-02618){ref-type=\"app\"}). We constrained the topologies of the phylogenetic tree to follow the species tree from the Interactive Tree of Life (<http://itol.embl.de/other_trees.shtml>) and estimated, for each gene of the melanocortin system, the branch lengths of the species tree using the HKY85 + Gamma model as implemented in PhyML \\[[@B58-ijms-18-02618]\\] from the alignments obtained from each gene. For our analyses, we used the topology of the species phylogenetic tree rather than the trees estimated from each gene. Our goal was to combine the estimation of selective pressure with the coevolution occurring between genes; thus, we needed to have a comparable topology for all our genes. Whenever possible (for the melanocortin receptors), we used the lamprey (*Lampetra fluviatilis*) as outgroup and a fish species when the sequence of a given gene was not available for the lamprey.\n\n3.3. Estimation of Evolutionary Constraint and Selective Pressure {#sec3dot3-ijms-18-02618}\n-----------------------------------------------------------------\n\nWe measured the percentage of identity to evaluate the level of evolutionary conservatism of each gene of the melanocortin system. The nucleotide positions with a score higher than 95% were considered highly conserved positions. The reported percentage of identity for the alignment is the mean of the percentage of identity of all alignment positions. Although the species sampled are not identical between the studied genes, we took particular care to sample evenly the main vertebrate lineages so that the conservation scores can be compared between genes.\n\nWe used the MEME model \\[[@B59-ijms-18-02618]\\] to classify each codon of the different melanocortin genes as either having evolved neutrally or under positive or purifying selection. The identification of selective pressures provided additional information that could not be measured by the percentage of identity. Purifying selection can be interpreted as an evolutionary force whose identification incorporates the phylogenetic relationship and the substitution process along the entire gene, whereas conservation is the description of a pattern observed in the alignment. We considered a maximum likelihood framework to fit a model allowing codons to be under selective pressure along branches of the phylogenetic tree. We compared this model based on the ten genes of the melanocortin system using likelihood ratio tests (LRT; *P*-value \\< 0.05) to a null model that did not allow the rate of non-synonymous substitutions to be larger than the rate of synonymous substitutions \\[[@B59-ijms-18-02618]\\]. We controlled for multiple testing through false discovery rate, FDR, using the qvalue package in R \\[[@B60-ijms-18-02618]\\]. We used the MEME model because we needed to determine the level of selection at codon sites assuming that episodic events of positive selection occurred throughout the evolution of vertebrates. We retrieved the LRT scores and their associated q-values for each codon position and determined whether they were under positive or purifying selection or whether they evolved neutrally. Mixed effects models of codon evolution \\[[@B59-ijms-18-02618]\\] detected a statistically significant signal of selection or neutral evolution in approximately 59.5% of the codon sites (31.1% and 75.9% being the lowest and highest values, respectively; [Figure 2](#ijms-18-02618-f002){ref-type=\"fig\"}b; [Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}). Within these unambiguous codon sites for each gene, we then calculated the proportion of codons that were under these different types of selective pressure for each gene.\n\nWe also counted the codons that were detected as evolving under positive selection with a probability higher than 95% and the proportion of branches of the vertebrate phylogeny where these codons were effectively inferred to be under positive selection. We used a Bayes factor greater than 1 to determine which branches should be counted in this procedure. Although a threshold of 1 for the Bayes factors might seem low, our goal here is not to identify the best model for each branch, but to compare the proportion of branches that show, for each gene, signs of positive selection. We did not consider the intensity of the selection (i.e., the exact level of the dN/dS ratio) in addition to the thresholds that are used to test whether the selection pressure is positive or purifying but counted only the codons that were significantly assigned to these types of selection. An accurate estimation of the parameter representing the dN/dS ratio can be difficult to obtain and it is better to assess this ratio using model comparisons (as describe above), which incorporated the uncertainty in the parameter value during the testing procedure.\n\n3.4. Estimation of Coevolution {#sec3dot4-ijms-18-02618}\n------------------------------\n\nWe used the maximum likelihood implementation of the model Coev \\[[@B9-ijms-18-02618]\\] to estimate the nucleotide positions that were coevolving between pairs of genes. The analyses were based on the concatenated gene sequences and the species phylogenetic tree ([Supplementary Materials Figure S1 and Table S2](#app1-ijms-18-02618){ref-type=\"app\"}). We measured the score of coevolution for 219,556 pairs of positions for 45 combinations of inter-molecular alignments by comparing the fit of the Coev model against a null model of independent evolution. We characterized pairs of nucleotide positions as coevolving based on the difference in Akaike information criterion (\u0394AIC) between the two models. However, assessing coevolution is difficult, and using the standard \u0394AIC threshold for model selection can lead to spurious detection of coevolution \\[[@B9-ijms-18-02618],[@B58-ijms-18-02618]\\]. We therefore used simulations to correct for bias in the evaluation of the \u0394AIC distribution \\[[@B9-ijms-18-02618]\\]. To confidently delimit the list of coevolving pairs for a given data set, we used the phylogenetic tree and branch lengths estimated from the concatenated alignments to simulate one alignment of 200 base pairs for each data set based on the independent JC69 model of substitution, as implemented in the package \"evolver\" from the PAML software \\[[@B61-ijms-18-02618]\\]. For each pair of positions from the simulated alignments, we calculated the \u0394AIC between the Coev model and the independent model that was used to simulate the data (i.e., JC69). We considered the 97.5th, 95th and 90th percentile values of these simulated \u0394AIC distributions (i.e., one per pair of genes) as the thresholds to accept that two sites were coevolving. We used these three thresholds to assess the robustness of the coevolution measures and investigate the strength of the effect in our analyses. In total, we evaluated the score of coevolution of 40,000 simulated pairs of nucleotide positions for each inter-molecular analysis, which amounts to 1,800,000 pairs of positions analyzed.\n\n3.5. Evolutionary Response in the Melanocortin System {#sec3dot5-ijms-18-02618}\n-----------------------------------------------------\n\nWe combined the predictions from the coevolution and selection analyses to estimate the extent to which positive selection exerted on a given site in a gene has resulted in coevolution of a site in a different gene. With this aim, we calculated for every gene the number of codons subject to positive selection that contained nucleotides inducing a coevolutionary response with another gene. For instance, among the 199 pairs of coevolving sites linking *MC2R* and *AGRP* (threshold 0.975, [Supplementary Materials Table S4a](#app1-ijms-18-02618){ref-type=\"app\"}), five were included in codons subject to positive selection in *AGRP* ([Supplementary Materials Table S3a](#app1-ijms-18-02618){ref-type=\"app\"}). Similarly, three nucleotides in *MC2R* induced an evolutionary response on *AGRP* and were contained in codons subject to positive selection ([Supplementary Materials Table S3a](#app1-ijms-18-02618){ref-type=\"app\"}). We used Principal Component Analysis (PCA) as implemented in the *prcomp* function in R \\[[@B62-ijms-18-02618]\\] to summarize the evolutionary responses of the melanocortin genes.\n\n4. Conclusions {#sec4-ijms-18-02618}\n==============\n\nOur study is a first step in the ultimate aim of considering all potential interactions between all genes within a genetic system. We show that the method developed by Dib et al. \\[[@B8-ijms-18-02618],[@B9-ijms-18-02618]\\] allows to assess to what extent genes within functional modules such as the melanocortin system jointly evolve. Genes in the melanocortin system differ in their relative coevolutionary influence suggesting that, from this point of view, selection on a few aspects within the system can be of more importance than others for the evolution of the entire system. Future studies should consider potential co-evolutionary dynamics among the genes in the melanocortin system and the genes that also contribute to regulate the functions controlled by the melanocortin system. Despite the importance of the melanocortin system in regulating functions such as melanogenesis, energy homeostasis, or steroidogenesis, several other genes outside the melanocortin system can also influence these phenotypic aspects (e.g., several dozens of genes have been described to influence pigmentation in vertebrates \\[[@B32-ijms-18-02618]\\]). The common participation of several genes in regulating a given function increases the likelihood for coevolution. Conducting these studies also offers the opportunity to test if co-evolution among genes that regulates the same function is, as intuition would suggest, stronger than among genes that have a common regulatory basis but code for different functions, such as those in the melanocortin system. In addition to these studies, replication in other known genetic networks or systems is required to assess the generality of our findings, particularly in systems with different evolutionary dynamics (for instance, with lower levels of sequence conservation). In contrast to other mechanisms such as regulation of gene expression, coevolution is not expected to be a major mechanism driving modular function \\[[@B1-ijms-18-02618]\\]. However, we show that coevolution between genes in a genetic system can still be of certain importance. Future studies will be needed to assess whether coevolution among the genes within a system is higher than with genes belonging to a different one and try to disentangle the major factors triggering coevolution between genes.\n\nWe thank Reto Burri, Julien Roux and Marc Robinson-Rechavi for providing useful comments on the manuscript, and Susana Figueiredo Pinto for help in preparing the data.\n\nSupplementary materials can be found at [www.mdpi.com/1422-0067/18/12/2618/s1](www.mdpi.com/1422-0067/18/12/2618/s1).\n\n###### \n\nClick here for additional data file.\n\nA.R., L.D. and N.S. conceived the study, L.D., N.S., and L.M.S.-J. performed the analyses, and A.-L.D., A.R., L.D., L.M.S.-J. and N.S. interpreted the analyses and wrote the manuscript.\n\nThe authors declare that they have no conflicts of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.\n\nThe authors would like to acknowledge funding support for this study from the the Swiss National Science Foundation: grant 31003A-120517 to A.R. and CR32I3_143768 to N.S.\n\nThe dataset supporting the conclusions of this article are available upon request.\n\n![Schematic presentation of the melanocortin system in human. The melanocortins (ACTH, \u03b1-, \u03b2- and \u03b3-MSHs) are peptide hormones derived from the proopiomelanocortin (POMC) prohormone through tissue-dependent post-translational modification by two convertases: PC1/3 (encoded by *PCSK1*) cleaves the prohormone to obtain ACTH and \u03b2-LPH, and cleavage of ACTH by PC2 (encoded by *PCSK2*) gives \u03b1-MSH. \u03b2- and \u03b3-MSH are also obtained by PC2 cleavage of POMC. Together with the antagonists, agouti-signalling protein (ASIP) and agouti-related protein (AgRP), the melanocortins bind to five melanocortin receptors MC1-5Rs with various affinities. MC1R mainly regulates melanogenesis, whereas the specific ACTH-receptor MC2R is essential for the regulation of glucocorticoidogenesis. MC3-5Rs are involved in the regulation of energy homeostasis (food intake, energy storage, lipolysis), autoimmune response, anti-inflammatory, cardiovascular and natriuretic processes, sexuality and social behavior. See references in the main text. In the figure, only the main tissues and functions in human were noted for simplification.](ijms-18-02618-g001){#ijms-18-02618-f001}\n\n![Sequence conservation (**a**) and selection patterns (**b**) among the main genes of the vertebrate melanocortin system. (**a**) Sequence conservation is given by the percentage of identity; (**b**) The percentages of nucleotide sites that experienced positive or purifying selection or that evolved neutrally have been standardized (\\[value--mean\\] divided by 1 standard deviation) to facilitate comparisons (see unstandardized data in [Supplementary Materials Table S1](#app1-ijms-18-02618){ref-type=\"app\"}). The proportion of nucleotide sites that could be unambiguously assigned by MEME to a codon that experienced positive or purifying selection or that evolved neutrally is indicated between parenthesis below the name of each gene.](ijms-18-02618-g002){#ijms-18-02618-f002}\n\n![Ordination of genes in the melanocortin system according to their coevolutionary influence onto other genes within the system. Shown are the scores derived from a principal component analysis on the number of pairs containing a nucleotide in a codon under positive selection that induce coevolution in other genes, the number of different sites containing a nucleotide in a codon under positive selection that induce coevolution in other genes, and the percentage of branches in the vertebrate tree where positive selection induced coevolution in other genes. For each gene, we showed the scores using different threshold to detect coevolution between sites.](ijms-18-02618-g003){#ijms-18-02618-f003}\n\nijms-18-02618-t001_Table 1\n\n###### \n\nCoevolutionary response of the melanocortin system to selection on target genes. The table shows the number of nucleotides of a target gene that belong to a codon under positive selection (in rows) and induced a coevolutionary response in the other nine genes of the melanocortin system (in columns). For example in the [Table 1](#ijms-18-02618-t001){ref-type=\"table\"}, 10 nucleotides, which are part of a codon under positive selection in *MC1R*, induced a change in 10 sites in *MC5R* during vertebrate evolution. For each target gene, we give the number of pairs of sites that coevolved (note that a given site within a codon can be implicated in more than one pair of sites). We also provide the number of sites that belong to a codon under positive selection of a target gene that induced an evolutionary response in at least one of the other genes of the melanocortin system (this number can be smaller than the number of pairs because the same site can be involved in several pairs). Finally, we give the length in nucleotides of each human target gene and the percentage (in relation to this sequence length) of the number of different sites that belong to a codon under positive selection of a target gene and that induced an evolutionary response, which indicates the mean percentage of sites that coevolved with other genes of the melanocortin system. The sequences used in the coevolution and selection analyses where trimmed to remove conserved sites and regions of the alignment containing ambiguities. The length of the nucleotide sequence is therefore shorter than the total length given in [Supplementary Materials Table S2](#app1-ijms-18-02618){ref-type=\"app\"}. To evaluate the robustness of our method in assessing coevolution, we applied the \u0394AIC threshold based on the 0.95 percentiles of the null distribution of \u0394AIC obtained by simulations (the percentiles 0.975 and 0.90 are reported in [Supplementary Materials Table S3](#app1-ijms-18-02618){ref-type=\"app\"}). Whether the frequency of number of pairs or the frequency of number of sites per gene was significantly different than the frequency estimated for the rest of the genes in the melanocortin system was tested using Pearson's \u03c7^2^ tests corrected for multiple testing using the Benjamini-Hochberg approach. Genes with frequencies in the number of pairs or the number of sites significantly above the frequency for the rest of the genes are denoted with an 'a' superscript and those with frequencies below are denoted with a 'b' superscript. No superscript denotes non-significant differences.\n\n                                                        Coevolutionary Response to Positive Selection on Target Genes                                                                               \n  ---------------------------------------- ------------ --------------------------------------------------------------- ---- ---- ---- ---- ---- ---- ---- ---- ------ --------- --------- -------- ------\n  **Positive selection on target genes**   ***ASIP***   \\-                                                              4    1    6    6    3    14   9    2    3      396       176 ^a^   22 ^a^   5.6%\n  ***AGRP***                               2            \\-                                                              4    4    4    0    7    2    4    6    396    141 ^a^   9         2.3%     \n  ***POMC***                               0            2                                                               \\-   2    0    0    3    0    0    1    801    33 ^b^    8         1.0%     \n  ***PCSK1***                              6            8                                                               9    \\-   10   4    13   5    8    6    2256   322 ^b^   25 ^b^    1.1%     \n  ***PCSK2***                              3            1                                                               0    10   \\-   3    6    2    4    3    1914   208 ^b^   13 ^b^    0.7%     \n  ***MC1R***                               5            3                                                               1    1    3    \\-   11   8    7    13   951    311 ^a^   34 ^a^    3.6%     \n  ***MC2R***                               3            4                                                               3    0    1    4    \\-   1    12   5    891    135       21        2.4%     \n  ***MC3R***                               0            3                                                               0    0    3    8    12   \\-   5    3    969    209 ^a^   17        1.8%     \n  ***MC4R***                               0            2                                                               0    0    2    5    3    1    \\-   0    996    85 ^b^    10        1.0%     \n  ***MC5R***                               6            8                                                               2    4    3    15   3    3    2    \\-   972    192       30 ^a^    3.1%     \n\nijms-18-02618-t002_Table 2\n\n###### \n\nNumber of coevolving nucleotide sites between pairs of genes of the melanocortin system using the \u0394AIC thresholds based on the 0.95 percentile of the null distribution of \u0394AIC obtained by simulation (see Methods). The percentiles 0.975 and 0.90 are given in [Supplementary Materials Table S4](#app1-ijms-18-02618){ref-type=\"app\"}.\n\n                Coevolving Pairs                                                                \n  ------------- ------------------ ----- ----- ------ ------ ------ ------ ------ ------ ------ ------\n  ***ASIP***    \\-                 181   101   519    170    643    913    631    105    323    398\n  ***AGRP***    \\-                 \\-    276   761    714    477    511    364    299    352    428\n  ***POMC***    \\-                 \\-    \\-    1348   \\-     606    682    104    340    192    405\n  ***PCSK1***   \\-                 \\-    \\-    \\-     4012   2672   3105   927    2704   1137   1909\n  ***PCSK2***   \\-                 \\-    \\-    \\-     \\-     2088   2207   1025   1410   1754   1487\n  ***MC1R***    \\-                 \\-    \\-    \\-     \\-     \\-     1253   1203   1636   1519   1344\n  ***MC2R***    \\-                 \\-    \\-    \\-     \\-     \\-     \\-     1336   1662   2011   1520\n  ***MC3R***    \\-                 \\-    \\-    \\-     \\-     \\-     \\-     \\-     1281   1454   925\n  ***MC4R***    \\-                 \\-    \\-    \\-     \\-     \\-     \\-     \\-     \\-     1250   1187\n  ***MC5R***    \\-                 \\-    \\-    \\-     \\-     \\-     \\-     \\-     \\-     \\-     1110\n"}
+{"text": "###### Strengths and limitations of this study\n\n-   Economic models of rotavirus vaccination simulate the absence from work due to rotavirus infection in children as well as the reduction in work absenteeism of mothers due to vaccine introduction, but nobody has evaluated these reductions in a real-life setting.\n\n-   The objective of this study was to evaluate the impact of introducing the rotavirus vaccine in Belgium on the reduction in work absenteeism over an observation period of 9\u2005years (from 2003 to 2012), using real-life data from a database of administrative personnel in the City of Antwerp. The vaccine was introduced with a high uptake in November 2006.\n\n-   The analysis suggests that rotavirus vaccination resulted in a reduction of absences from work among mothers with a first child during the first, second and third epidemic periods of rotavirus after birth, with a cumulative gain of 2.24\u2005days per woman.\n\n-   This translated into a net cost gain for the employer of \u20ac187 per working mother.\n\n-   The main limitation of the study is that the results are based on retrospective data analysis with no causal relationship between the introduction of the vaccine and the reduction in absenteeism specified. However, several indirect arguments have been brought forward to support the potential link.\n\nIntroduction {#s1}\n============\n\nThe rotavirus (RV) epidemic is an annual recurrent public health problem of severe diarrhoea in young children, with a peak incidence before the age of 2\u2005years.[@R1] [@R2] RV disease preferentially occurs during the winter months in temperate countries in the northern hemisphere. The virus spreads among young children but may manifest a higher rate of transmission at around 10\u2005months of age, because at this age a child in a day-care centre is a conducive virus transmitter to younger and older children.[@R3]\n\nRV vaccination was introduced in Belgium in November 2006 as a new management strategy against the illness.[@R4] Belgium was one of the first countries in Europe to integrate this vaccine into its routine childhood immunisation programme.[@R5] Vaccine uptake was high from the start (\\>85%) because it was recommended by the High Committee of Health Promotion. Moreover, the organisational structure for implementing immune protection in children and a good follow-up process are both well developed in the country.[@R6]\n\nSeveral economic evaluations of the RV vaccine have been conducted and most of these analyses have included indirect cost estimates.[@R7] An analysis of the financial burden of RV disease in four European countries indicated that the indirect costs could be substantial: half of the total cost of the disease per child at risk could be linked to these indirect costs.[@R11] However, until now, these estimates have always been simulated and nobody has been able to evaluate the reduction of work absenteeism using real-life data subsequent to the introduction of the RV vaccine.[@R12] Obtaining that type of evidence is not easy, as we need to have an environment where employment is stable among a large number of employees in order to follow enough working mothers with young children under the same working conditions and having the same exposure to the disease. In addition, we needed to obtain detailed information on each period of absenteeism with a start and end date linked to the employee\\'s family condition when a new child is born. The data should be available over a long enough period of time (at least 5\u2005years) and in electronic format with easy access, so that the time periods before and after vaccine introduction can be analysed and compared.\n\nIt was postulated that during epidemic periods of RV disease, working mothers with a first child would be absent from work for short durations (\u22645\u2005days) more often than during non-epidemic periods or after the introduction of the vaccine.[@R13] In addition, working women with no exposure of children to the rotavirus disease should not experience any reduction in work absenteeism as a result of the vaccine. We first constructed a simple model that could give us guidance in our search for parameters in real-life data sets.\n\nDesign and setting {#s2}\n==================\n\nSimple model construction {#s2a}\n-------------------------\n\nThe simple model calculates the expected difference in worker absenteeism when comparing exposure versus non-exposure to the RV vaccine. It is expressed as the estimated number of days per year and per working mother with a first child. The value estimated would then serve as a benchmark for analysing the observed data.\n\nObserved data {#s2b}\n-------------\n\nWe selected a database from the City of Antwerp which has a sufficient number of subjects from the target group (n\u224811\u2005600 women per year) over a long period of time (from 2003 to 2012). This database collects detailed information on absences from work for all its administrative personnel, including the general reason, duration and time period of the absence (start and end dates). It was unfortunate that electronic data before 2003 were not available in order to better capture overall trend results related to the disease over time. The data are reasonably accurate because personnel payment is linked to that system. Moreover, through a unique subject number, the database could be linked to other databases in the city which compile information on family composition and the birth dates of children born to each employee. The data were made available after decoding subjects to prevent identification of individuals and after approval of the project and its objectives by the administrative head of the city.\n\nWe performed the analysis in three steps. First, to increase the chances of observing a difference in absenteeism due to vaccination, we selected a target group of women with a first child approximately 10\u2005months of age during the typical RV epidemic season (January--May) of each year. These children are known transmitters of the virus. In a second step, we conducted an annual analysis of the same working mothers but with a first child born any time during the year prior to the next epidemic period. We expected a larger difference (ie, less absenteeism) from the first analysis than from the second one. In a third step, we selected from the same database women aged 30--35\u2005years with no first child, but working during the same observation period from 2004 to 2012, from January to May. It was hypothesised that these women, considered as a control group, should not benefit from the rotavirus vaccination, and therefore we would not observe any decrease in work absenteeism over time.\n\nThus, in the first step, we selected working women with a first child born during the months of April to July on a yearly basis from 2003 to 2011. That number was variable per year. We then recorded short absences from work (\u22645\u2005days) which were registered 10\u2005months after the birth of their child. A maximum of 5\u2005days of absence was considered the right limit to be linked to the disease event that is frequent enough to occur and should not interfere with planned holiday periods. A longer period of absenteeism, such as \u226410\u2005days, could be considered a holiday and therefore become a confounding factor in the analysis. The period of 10\u2005months postpartum was equal to the normal RV epidemic season. The sum of all work absences during that annual period of time in all years from 2004 to 2012 was then divided by the number of mothers considered during the previous year in order to obtain an average value per working mother with a first child during the following epidemic period.\n\nWe compared the data by year to observe any marked difference in work absenteeism after 2006, which was the year RV vaccination was introduced. We also analysed absenteeism data of short duration from the same working mothers outside the epidemic period, expecting a much lower rate and no major change over time. To be able to compare the same values by time period, we analysed the average value by month for each period (epidemic and non-epidemic).\n\nIf an important difference was observed in the first step, we then proceeded to the second step of evaluating work absences during each epidemic period among mothers with a first child born any time during the previous year (whole-year birth cohort). We again reported the sum of all days absent from work in the postpartum year during epidemic and non-epidemic periods. We hypothesised that if the difference in absenteeism was large enough in the first step, it should still be present in the second step, and that would facilitate the analysis of other time periods. In addition, we evaluated the same type of absences from work among mothers with a firstborn child in its second and third years of life (eg, absences of mothers with a first child born during 2003 were evaluated in the epidemic periods of 2005 and 2006, respectively). Finally, we compared these observed data with the estimates obtained from the simple model. The control group was analysed in the same way as the other groups. We report the same type of outcome measure over time, which is the average number of days of absence from work per woman during the epidemic period per year.\n\nOn the basis of the above results, we could calculate the net cost gain per working mother through the average reduction in absenteeism postvaccination. This was adjusted by the cost of the vaccine, which was considered at \u20ac60/dose.[@R14] The average gross salary for a working mother in the City of Antwerp was estimated at \u20ac135/day.[@R15]\n\nTo observe statistically significant differences between prevaccination and postvaccination absenteeism per working mother, we compared the data by ranking mothers into six categories according to the number of days absent from work (0, 1, 2, 3, 4 and 5\u2005days) during the epidemic period. We then applied a statistical ranking test (Mann-Whitney U-test, with a significance level of p\\<0.05, two-sided) for the pooled years (2004--2006) prevaccination with the pooled years (2010--2012) postvaccination. We also evaluated for [tables 1](#BMJOPEN2014007453TB1){ref-type=\"table\"} and [2](#BMJOPEN2014007453TB2){ref-type=\"table\"} a pooled analysis of the period prevaccination (2004--2006) versus postvaccination (2010--2012) regarding the average days of absenteeism per woman per period with a t test for a statistical significance level of p\\<0.05, two-sided. Finally, the comparison between the control group (women aged 30--35\u2005years) and the target group (mothers with a first child) was analysed using linear regression analysis over the period of 2004--2012 using year as the independent variable and the average number of days absent from work during the observation period as the dependent variable. The \u03b2 regression coefficients were compared for significant difference between the two groups using a t test comparison for a significance level of p\\<0.05, two sided. Statistical analyses were performed using IBM SPSS Statistics V.22.0 and GraphPad v6.\n\n###### \n\nAverage number of short work absences per targeted woman with a first child during the epidemic and non-epidemic seasons\n\n  Year   Women in target period   Epidemic period (January--May)   Non-epidemic period (June--December)                                                        \n  ------ ------------------------ -------------------------------- -------------------------------------- ------- ------- ------- ---- ------- ------- ------- -------\n  2003   56                                                                                                                                                    \n  2004   57                       98                               1.750                                  2.252   1.247   0.350   76   1.357   1.764   0.950   0.194\n  2005   62                       97                               1.702                                  2.234   1.168   0.340   27   0.474   0.742   0.204   0.068\n  2006   66                       98                               1.581                                  2.048   1.113   0.316   58   0.935   1.309   0.561   0.134\n  2007   80                       109                              1.652                                  2.116   1.186   0.330   54   0.818   1.151   0.485   0.117\n  2008   65                       148                              1.850                                  2.285   1.414   0.370   67   0.838   1.136   0.538   0.120\n  2009   62                       65                               1.000                                  1.346   0.653   0.200   39   0.600   0.900   0.299   0.086\n  2010   96                       63                               1.016                                  1.354   0.677   0.203   66   1.065   1.484   0.644   0.152\n  2011   114                      64                               0.667                                  0.907   0.426   0.133   68   0.708   1.049   0.491   0.101\n  2012                            98                               0.860                                  1.113   0.588   0.172   84   0.737   0.972   0.501   0.105\n\n###### \n\nFrequency distribution of days absent from work prevaccination and postvaccination among mothers with a firstborn child in the first year of life during the epidemic period and the non-epidemic period, and among women aged 30--35\u2005years old with no firstborn child during the epidemic period only\n\n  Days                                                             2004    2005    2006    2007    2008    2009    2010    2011    2012\n  ---------------------------------------------------------------- ------- ------- ------- ------- ------- ------- ------- ------- -------\n  Full cohort of women with firstborn child, epidemic period                                                                       \n  \u20030                                                               70      85      78      94      101     116     131     159     167\n  \u20031                                                               34      26      32      22      39      54      37      32      41\n  \u20032                                                               10      17      14      33      16      22      13      30      19\n  \u20033                                                               8       8       14      15      13      14      16      16      10\n  \u20034                                                               12      13      11      10      19      8       10      10      11\n  \u20035                                                               31      48      31      35      35      13      14      9       17\n  \u2003Total N                                                         165     197     180     209     223     227     221     256     265\n  \u2003Total N of absent days                                          281     376     301     348     361     237     221     225     238\n  \u2003Average                                                         1.703   1.909   1.672   1.665   1.619   1.044   1       0.879   0.898\n  \u200395% CI +                                                        2.005   2.201   1.955   1.924   1.873   1.264   1.236   1.09    1.116\n  \u200395% CI\u2212                                                         1.401   1.617   1.389   1.406   1.365   0.825   0.764   0.668   0.68\n  Cohort of women aged 30--35\u2005year, epidemic period                                                                                \n  \u20030                                                               248     303     365     399     435     242     440     500     464\n  \u20031                                                               179     176     168     191     257     372     167     206     230\n  \u20032                                                               106     104     123     109     128     162     109     126     195\n  \u20033                                                               63      60      76      79      79      72      101     139     111\n  \u20034                                                               61      77      93      81      97      77      65      71      89\n  \u20035                                                               157     155     160     215     203     81      144     144     200\n  \u2003Total N                                                         814     875     985     1074    1199    1006    1026    1186    1289\n  \u2003Total N of absent days                                          1609    1647    1814    2045    2153    1625    1668    1879    2309\n  \u2003Average                                                         1.977   1.882   1.842   1.904   1.796   1.615   1.626   1.584   1.791\n  \u200395% CI +                                                        2.106   2.008   1.96    2.021   1.903   1.711   1.74    1.688   1.892\n  \u200395% CI\u2212                                                         1.847   1.757   1.724   1.787   1.689   1.519   1.512   1.481   1.691\n  Full cohort of women with firstborn child, non-epidemic period                                                                   \n  \u20030                                                               96      142     116     129     142     158     142     163     176\n  \u20031                                                               25      27      24      32      27      20      29      34      31\n  \u20032                                                               22      13      16      19      22      22      13      14      20\n  \u20033                                                               5       3       9       16      14      12      13      17      14\n  \u20034                                                               8       7       6       9       10      9       6       12      12\n  \u20035                                                               9       5       9       4       8       6       18      16      12\n  \u2003Total N                                                         165     197     180     209     223     227     221     256     265\n  \u2003Total N of absent days                                          161     115     152     174     193     166     208     241     221\n  \u2003Average                                                         0.976   0.584   0.844   0.833   0.865   0.731   0.941   0.941   0.834\n  \u200395% CI +                                                        1.203   0.801   1.07    1.026   1.063   0.931   1.156   1.136   1.023\n  \u200395% CI\u2212                                                         0.749   0.367   0.618   0.639   0.668   0.532   0.727   0.746   0.645\n\nN, working mothers.\n\nResults {#s3}\n=======\n\nModelled data {#s3a}\n-------------\n\nAs shown in [table 3](#BMJOPEN2014007453TB3){ref-type=\"table\"}, the simple model indicated that the introduction of the RV vaccine produced a gain of 0.73--0.80 working days per mother, with a first child in the vaccinated cohort. Since there was no maximum vaccine coverage in the vaccinated cohort, we needed to include a normal rate of infection among the unvaccinated in that cohort (='Rest'). A sensitivity analysis around that value was performed, since high vaccination levels result in a herd effect in the 'Rest' group. Therefore, the difference between prevaccination and postvaccination absenteeism could be higher.\n\n###### \n\nModel estimates\n\n  Parameter                                                                                                 Value                Absolute numbers   Difference\n  --------------------------------------------------------------------------------------------------------- -------------------- ------------------ ----------------\n  No vaccination                                                                                                                                    \n  \u2003Working mothers with a 1st child                                                                         75                                      \n  \u2003% of mothers with a 1st child having diarrhoea 1st year                                                  20%                  75\u00d720%=15          \n  \u2003Average duration (days) for being absent for diarrhoea in a child                                        5                    15\u00d75=75            \n  \u2003Average number of days absent/woman                                                                                           75/75=1            1\n  Vaccination                                                                                                                                       \n  \u2003Working mothers with a 1st child                                                                         75                                      \n  \u2003% of mothers with a vaccinated child                                                                     85%                  75\u00d785%=64          \n  \u2003% of mothers with no vaccinated child                                                                    (1--85%)=15%         75\u00d715%=11          \n  \u2003Vaccine efficacy against diarrhoea                                                                       85%                                     \n  \u2003% of mothers with a vaccinated child still having diarrhoea                                              20%\u00d7(100%--85%)=3%   64\u00d73%=2            \n  \u2003% of mothers with an unvaccinated child still having diarrhoea (Rest)                                    20%                  11\u00d720%=2           \n  \u2003Average duration (days) for being absent for diarrhoea in a child                                        5                    4\u00d75=20             \n  \u2003Average number of days absent/woman                                                                                           20/75=0.27         0.27\n  \u2003Gain in working days avoided/woman after vaccination 1st year                                                                                    (1--0.27)=0.73\n  Sensitivity analysis                                                                                                                              \n  \u2003Proportion of children with diarrhoea is lower because of the vaccine\\'s herd effect in the Rest group   10% instead 20%      11\u00d710%=1           \n  \u2003Average duration (days) for being absent for diarrhoea in a child                                                             3\u00d75=15             \n  \u2003Average number of days absent/woman                                                                                           15/75=0.20         0.20\n  \u2003Gain in working days avoided/woman after vaccination 1st year                                                                                    (1--0.20)=0.80\n\nObserved data {#s3b}\n-------------\n\n### First-step analysis {#s3b1}\n\n[Table\u00a01](#BMJOPEN2014007453TB1){ref-type=\"table\"} summarises the annual number of days absent from work in each epidemic season among the target group of working mothers with a first child born between April and July during the previous year. It should be noted that a reduction in absenteeism is observed only after 2008 because the vaccine was introduced in November 2006. In the non-epidemic period, no large changes were seen in any given month. We observed a reduction in absenteeism (average days per woman (2003--2008)\u2212average days per woman (2009--2012) =0.821) noted from 2009 onwards, so that we could proceed with the second step. Applying t-testing with unequal variance on a pooled analysis of days being absent from work during the prevaccination period (2004--2006; n=175 observations, average value= 1.674\u2005days) versus the postvaccination period (2010--2012; n=272 observations, average value=0.827\u2005days), the test result of 5.0365 with df=283 was statistically very significant (p\\<0.0001, two sided) with a mean difference of 0.847\u2005days and 95% CI of 0.540 to 1.154\u2005days.\n\n### Second-step and third-step analyses {#s3b2}\n\n[Table\u00a02](#BMJOPEN2014007453TB2){ref-type=\"table\"} reports the number of absent workdays among mothers with a firstborn child during the epidemic period of the first postpartum year in six different categories (0, 1, 2, 3, 4 and 5\u2005days), together with average values and the 95% CI. From 2009 onwards, we observed a clear increase in the number of 0-day absences from work and a decline in the number of 5-day absences. In the control group of women aged 30--35\u2005years, no marked change in absenteeism was seen over time. There is not much difference noted in the average values for the full cohort of women ([table 2](#BMJOPEN2014007453TB2){ref-type=\"table\"}) compared with the targeted cohort ([table 1](#BMJOPEN2014007453TB1){ref-type=\"table\"}), except for a narrower CI in [table 2](#BMJOPEN2014007453TB2){ref-type=\"table\"} because of the higher number of persons enrolled in the analysis. During the non-epidemic period, there was no big variation in the numbers noted year after year. Rank testing using the Mann-Whitney U statistics revealed a statistically significant change in distribution of the different day categories for being absent from work (n=1284 (2004, 2005, 2006, 2010, 2011, 2012); z-statistics=7.8903; p\\<0.0001, two sided) linked to an observed average reduction in days of absence from work over time. [Figure\u00a01](#BMJOPEN2014007453F1){ref-type=\"fig\"} reports the percentage change per year of each of the 6-day categories for absence from work, showing a dramatic change in category 0 versus category 5\u2005days during the epidemic period for the target group ([figure 1](#BMJOPEN2014007453F1){ref-type=\"fig\"}A). This is not observed during the non-epidemic period ([figure 1](#BMJOPEN2014007453F1){ref-type=\"fig\"}B) or in the control group ([figure 1](#BMJOPEN2014007453F1){ref-type=\"fig\"}C).\n\n![Proportional distribution of days absent from work during the epidemic period reported per year for (A) working mothers with a first child and (B) the control group (women aged 30--35\u2005years); and (C) proportional distribution of days absent from work during the non-epidemic period reported per year for working mothers with a first child.](bmjopen2014007453f01){#BMJOPEN2014007453F1}\n\nApplying t-testing with unequal variance for the target group during the epidemic period on a pooled analysis of days absent from work during the prevaccination period (2004--2006; n=542 observations, average value=1.767\u2005days) versus the postvaccination period (2010--2012; n=742 observations, average value=0.921\u2005days), the test result of 8.343 with df=946 was statistically very significant (p\\<0.0001, two sided) with a mean difference of 0.846\u2005days and 95%CI of 0.655 to 1.035\u2005days. The same analysis, performed for the non-epidemic period and for the control group, revealed the following t test results. For the non-epidemic period, the 2004--2006 average absent day value per woman was 0.789 and 0.902 for the 2010--2012 period. T test statistic\\'s value was 1.40 for df 1228, p value two sided was 0.16, indicating a non-significant difference. For the control group during the epidemic period, the 2004--2006 average absent day value per woman was 1.896 and 1.330 for the 2010--2012 period. T test statistic\\'s value was 12.542 for df of 5315, with a p value two sided \\<0.0001, indicating a very significant difference. This can be explained by two facts, the high number of observations (n=6175) and the remarkable dip in 5\u2005days of absence from work in 2009 as shown in [figure 1](#BMJOPEN2014007453F1){ref-type=\"fig\"}C. Linear regression analyses of the average number of absent days by year (2004 to 2012) were undertaken for the two groups in [table 2](#BMJOPEN2014007453TB2){ref-type=\"table\"}, the target and the control group, during the epidemic period. \u03b2-Coefficients were, respectively, \u22120.138 (95%CI \u22120.19 to \u22120.08) and \u22120.039 (95%CI \u22120.07 to \u22120.01). The t test value of \u22123.887 with 16 df results in a p value two-sided \\<0.05. The null hypothesis, that the slopes are equal, can be rejected, and a significant difference between the two groups for any incremental year in the analysis is present. This may be caused by an external factor such as the introduction of the rotavirus vaccine.\n\n[Figure\u00a02](#BMJOPEN2014007453F2){ref-type=\"fig\"}A reports the results of [table 2](#BMJOPEN2014007453TB2){ref-type=\"table\"}, comparing the average value per month and per time period, because the epidemic and non-epidemic seasons have different durations (5 and 7\u2005months, respectively). It is important to note that the full cohort of women with vaccinated children does not reach the same absenteeism level as in the non-epidemic period. At the same time, the control group does not manifest a substantial decline in absenteeism during the same observation period.\n\n![(A) Average number of short work absences per woman per month for the control group (blue), the full cohort of mothers during the epidemic period (red), and during the non-epidemic period (green) and (B) Average number of short work absences per woman in the full cohort in the first (red), second (blue) and third (yellow) years postpartum during the epidemic period and the non-epidemic period (green).](bmjopen2014007453f02){#BMJOPEN2014007453F2}\n\n[Figure\u00a02](#BMJOPEN2014007453F2){ref-type=\"fig\"}B also shows the average number of absent workdays during the epidemic period when the firstborn child is in the first, second and third years of life. It is interesting that the same type of decline in absenteeism is observed in subsequent years as for the first year postpartum analysis. As already mentioned, reductions in absenteeism started after 2008 following the introduction of the RV vaccine by the end of 2006. A comparison of averages over the whole observation period shows that prior to the introduction of the vaccine, the average number of days absent from work during the epidemic period was an estimated 1.71\u2005days (average value from 2004 to 2008). The average number of days absent from work in the non-epidemic period was 0.83\u2005days (average value from 2004 to 2012). Thus, the estimated difference in absenteeism obtained from switching from no vaccination to vaccination is approximately 0.88\u2005days per working mother with a first child in the first year of life (1.71 to 0.83), 0.70\u2005days for 2-year-old children (1.53 to 0.83) and 0.67\u2005days for 3-year-old children (1.50 to 0.83).\n\nThe accumulated gain over a 3-year period per working mother with a first child during the epidemic period is 2.24\u2005days (0.88+0.7+0.67). The absolute gain is difficult to measure from the database, which reports fluctuating numbers and different lengths of duration each year. [Table\u00a04](#BMJOPEN2014007453TB4){ref-type=\"table\"} shows the calculated values, based on the average numbers obtained.\n\n###### \n\nEstimated cumulative gain per working mother with a first child over a 3-year period\n\n             Postpartum   Prevaccination: days absent from work\\*   Postvaccination: days absent from work\\*   Difference (days)   Days gained (216 women)\\*\n  ---------- ------------ ----------------------------------------- ------------------------------------------ ------------------- ---------------------------\n  Average    1st year     1.71                                      0.83                                       0.88                190\n  2nd year   1.53         0.83                                      0.70                                       150                 \n  3rd year   1.50         0.83                                      0.67                                       144                 \n  Sum        4.74         2.49                                      2.25                                       484                 \n\n\\*Per working woman with a first child.\n\nThe benefit of vaccination to an employer with an average annual workforce of 216 working mothers with a first child is an estimated gain of 484 working days over a 3-year period. At an average gross monthly salary of \u20ac3000 (\u20ac135 per workday), the gross gain is \u20ac67\u2005095 for the entire working mother cohort. The employer will spend a total of \u20ac26\u2005640 for the vaccine if the cost is an estimated \u20ac120 per mother. This results in an estimated net gain of \u20ac40\u2005455 for the cohort, or \u20ac187 per working mother.\n\nDiscussion {#s4}\n==========\n\nThe current study investigated whether we could use real-life data to obtain an estimate of the reduction in absenteeism among working mothers when rotavirus vaccination was introduced in Belgium in November 2006 with a high uptake at the start. We compared the observed results with estimates calculated through a simple modelling exercise. The analysis confirms a measurable reduction in work absenteeism observed in the real-life data setting. This reduction was observed among cases with a high number of absentee working days (eg, 5-day absences, see [table 2](#BMJOPEN2014007453TB2){ref-type=\"table\"} and [figure 1](#BMJOPEN2014007453F1){ref-type=\"fig\"}A).\n\nThe initial intention of the project was primarily to increase our chances of successfully selecting the right target group to show a difference in absenteeism which could be linked to the introduction of the rotavirus vaccine. Therefore, we opted for mothers with a first child, since in the Belgian culture they are the most likely to take time off work when the child is ill. A reduction in work absenteeism was observed, and the critical question is whether this observed reduction was linked to the introduction of rotavirus vaccination, as many other reasons for short absences from work could cause a fluctuation in this parameter. The analysis here indicates five reasons supporting a potential link between the observed reduction in absenteeism and the introduction of the rotavirus vaccine. First, the reduction happened after the introduction of the vaccine in 2006, at the time we would expect to observe a major reduction during the epidemic rotavirus season. Second, mothers with a first child in their second and third years of life also manifested a reduction in absenteeism starting during the same year (2008), which can be explained by the known herd effect postvaccination. If no herd effect was known for this vaccine, we would not expect to observe these additional reductions in the other age groups at the same period. Third, the observed reduction per working mother closely matches the modelled estimates (see [table 3](#BMJOPEN2014007453TB3){ref-type=\"table\"}). Fourth, no reduction in work absenteeism was seen during the non-epidemic period, making it unlikely that the reduction in absenteeism resulted from a factor such as a change in rules put in place by the employer to minimise short-term absences, as such a factor would have affected absenteeism during all time periods after 2008. Finally, women with no exposure of their children to any rotavirus vaccination did not show a marked reduction in absenteeism during the same observation period. It is unlikely that the reduction in absenteeism was linked to fluctuations in other childhood infections such as influenza, because there were no reports of high epidemic infectious diseases during the prevaccine periods in the literature or in local disease reports.\n\nSpecific conditions in retrospective data analyses must be fulfilled before it is possible to measure the changes observed. Those conditions are: (1) the demographic composition (gender and age) of the study population must remain stable in order to retain the same denominator; (2) the rules and conditions for taking time off work must be maintained; (3) the disease must be causing a serious public health problem over a certain period so that a change in working patterns (eg, absenteeism) can be observed; (4) the new intervention (eg, the vaccine) must have an immediate high uptake as well as a large and rapid impact on the disease; (5) the data registry must be adequate, of high quality, consistent over time and easily stored and accessible and (6) finally, the target population must be a well-defined group. We cannot work with cultural changes over time (eg, fathers instead of mothers suddenly becoming the main carers for young children who fall sick). All these different factors applied to the database of the administrative personnel of the City of Antwerp. For instance, the fact that the data collected on absenteeism was linked to the payment condition of an employee makes the quality of the data very rich. If one of the conditional elements mentioned above is of poor quality, it automatically decreases the value of the whole investigation and the analysis.\n\nHaving an initial estimate from the model was a helpful tool in understanding the potential gain to be observed in the real-life database. Looking at all the days of work absences and all mothers with children was not a viable option because the specific condition of interest was lost in the large numbers that were not related to disease in children necessitating short (\u22645\u2005days) absences from work. Surprisingly, it appeared that the benefit was a little higher in real life than in the model. This could potentially be related to a higher incidence or distribution of the disease than was predicted by the model, or to a herd effect in the vaccinated age group itself. It is clear from the data that a herd effect was realised in other age groups who were unvaccinated when the vaccine was introduced, as reported in [figure 2](#BMJOPEN2014007453F2){ref-type=\"fig\"}B. The data confirmed our observations in the RotaBIS study, in which the vaccine made a large impact among unvaccinated age groups as soon as it was introduced.[@R3]\n\nThis study has some limitations. First, the sample size of working mothers with a first child was small. However, the analysis was the best we could make in a country where the vaccine coverage rate for rotavirus vaccines was very high from the start. Second, the prevaccination period was short, from 2003 to 2006. This was because the data were only available in electronic format from 2003 onwards, so it was not possible to include earlier years in the analysis. Third, the database did not give a specific reason for absence from work, and therefore it was not possible to identify absences that resulted directly from rotavirus infection of a child. However, as discussed above, our findings are consistent with a link between the reduction in absenteeism and the effect of rotavirus vaccination.\n\nTo the best of our knowledge, this analysis is the first to demonstrate a potential effect of the RV vaccine on absenteeism using real-life data. The conditions of RV infection provide the opportunity for this to happen: the disease is very contagious, preferentially affects very young children who need care by an adult, and occurs mainly during a short epidemic season every year at the same time. These conditions allow the comparison of work absenteeism during epidemic and non-epidemic periods. It reinforces the circumstantial evidence showing a link between the observed reduction in absenteeism and the introduction of the vaccine, but we cannot claim or prove a clear causality from the data here. There are potentially other methods for collecting the same data prospectively, but it would be difficult to attain the same quality in the final results.[@R16] This is not a clinical trial but an analysis of an administrative database, in which the prospect of collecting that type of information in such a rigorous way is not obvious. It should also be clear that the overall benefit of the vaccine on a reduction in absenteeism in the workplace could be greater than measured here in a very specific subgroup of working mothers (those with a first child). It is likely that we could observe the same benefit among working mothers with a second or third child who was never previously exposed to RV.\n\nFinally, will the findings in this particular environment be easily transposable to other settings? Given the many conditions necessary to observe and measure the effect of a vaccine, it is likely that in other settings different amounts of benefit will be seen. For example, the facilities needed for an adult to be easily absent from work for childhood illness must be present before similar results could be observed.\n\nConclusions {#s5}\n===========\n\nWorking mothers with a first child benefit from RV vaccination through a reduction in work absenteeism. The modelled estimates and the observed data fit well for absences from work during the year following birth. The higher observed gain (0.88-day vs a 0.72-day gain) could be explained by a herd effect of the vaccine. There is possibly an underestimate of the total gain as only a select group (mothers with a first child) was investigated. In the case of the City of Antwerp, the benefit can be expressed as a cost gain per woman as a cost--benefit ratio of 1.85 (working days gained/vaccine cost). Confirmation of these results with data sets from other public organisations in Belgium is expected in the near future.\n\nThe authors wish to thank Carla Lefebvre (independent research and writing consultant) and Carole Nadin (Fleetwith Ltd) for assistance in editing this paper; Gregory Collet for his editorial support (Business & Decision Life Sciences on behalf of GSK Vaccines); and the City of Antwerp for making the data set of administrative personnel available for analysis.\n\n**Contributors:** BS developed the protocol, prepared the analysis, and the report and wrote the first draft of the manuscript. EVdM helped in getting access to the data and reviewed the statistical analysis plan, as well as giving input into the review of the manuscript. VN helped in designing the study and reviewed the different drafts of the manuscripts.\n\n**Funding:** This work was supported by GlaxoSmithKline Biologicals SA (GSK study identifier HO-12-12768). GlaxoSmithKline Biologicals SA was also responsible for all the costs associated with the analysis, as well as the development and publication of the present manuscript. BS received a discount on open access fee due to his second affiliation (University of Groningen).\n\n**Competing interests:** BS is an employee of the GSK group of companies and holds stock in the GSK group of companies. VN reports grants from the GSK group of companies while conducting the study.\n\n**Ethics approval:** The Head of the City Administration of Antwerp.\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n\n**Data sharing statement:** All data were anonymised before sharing among the researchers. The method of analysis and the results were evaluated by all the researchers. Persons who are interested can get access to the data and the method of analysis on request. The anonymised database with information on days and dates of being absent from work and duration, as well as a child\\'s date of birth (MM/YY), was made available per year from 2003 to 2012. All the authors and coauthors have had access to the data. We explained back to the administration personnel how we analysed the data and how we made the link to the data.\n"}
+{"text": "![](brjcancer00504-0227.tif \"scanned-page\"){.607}\n\n![](brjcancer00504-0228.tif \"scanned-page\"){.608}\n\n![](brjcancer00504-0229.tif \"scanned-page\"){.609}\n\n![](brjcancer00504-0230.tif \"scanned-page\"){.610}\n\n![](brjcancer00504-0231.tif \"scanned-page\"){.611}\n\n![](brjcancer00504-0232.tif \"scanned-page\"){.612}\n\n![](brjcancer00504-0233.tif \"scanned-page\"){.612-1}\n\n![](brjcancer00504-0234.tif \"scanned-page\"){.612-3}\n\n![](brjcancer00504-0235.tif \"scanned-page\"){.612-5}\n\n![](brjcancer00504-0236.tif \"scanned-page\"){.613}\n\n![](brjcancer00504-0237.tif \"scanned-page\"){.614}\n"}
+{"text": "INTRODUCTION {#sec1-1}\n============\n\nThe radiological examinations in dentistry can be classified into intraoral (with film or the sensor placed in the mouth) and extraoral imaging techniques. Extraoral imaging includes (1) panoramic X-ray showing a curved section of the whole-maxillo-facial block (more or less mandible shape), and (2) cephalometric X-ray showing a projection, as parallel as possible, of the whole skull. The cephalometric X-ray may be achieved from the lateral side or from the anterior--posterior view.\n\nThere is no doubt that having an exact and noiseless view of such images can offer an invaluable help to the dentist in better diagnosis and treatment. Most of prevalent techniques of dental imaging are going to be acquired, processed, and even evaluated in digital form; however, the low contrast and the noise pollution are inevitable problems. The proposed method tries in reduction of noise level in different kinds of the mentioned dental image types and provides a better visualization for the dentist and a good contrast to the noise ratio by numerical computation. Furthermore, the method is of benefit to real-time image processing due to its fast response.\n\nSeveral papers reported current methods in dental noise reduction, first of which proposed by Goebel\\[[@ref1]\\] in 2005 and was based on removal of a multiplying background using polynomial scaling and the A-Trous multiresolution transform. In 2006, Frosio\\[[@ref2]\\] proposed a mixture model made up of two Gaussian distributions and one inverted lognormal distribution to analyze the image histogram. A similar approach was also proposed by Sayadi\\[[@ref3]\\] for enhancing digital cephalic radiography using mixture models and local gamma correction. In 2008, Frosio introduced a filter based on the classical switching scheme in which the pulses were first detected and then corrected through a median filter.\\[[@ref4][@ref5]\\] One year later, Lucchese\\[[@ref6]\\] reported a principled method for setting the regularization parameter in total variation filtering, which was based on the analysis of the distribution of the gray levels on the noisy image. Along with this research, Frosio\\[[@ref7]\\] proposed a statistical based impulsive noise removal. The application of anisotropic nonlinear tensor-based diffusion\\[[@ref8]\\] in dental images was also reported in 2009 and 2010 by Kroon\\[[@ref9][@ref10]\\] for cone-beam CT images. Ju\\[[@ref11]\\] also proposed the application of Ridgelet transform for denoising of the root canal image in dental films. In 2010, Serafini\\[[@ref12]\\] used a gradient projection method for image denoising and deblurring on dental radiographies. More recently, in 2011, Bonettini\\[[@ref13][@ref14]\\] reported the application of the extra gradient method for total variation-based image restoration from Poisson data and interior point methods for edge-preserving removal of Poisson noise in dental radiographies.\n\nIn spite of the sophistication of the recently proposed methods, most algorithms have not yet attained a desirable level of applicability. All show an outstanding performance when the image model corresponds to the algorithm assumptions but fail in general and create artifacts or remove image fine structures. Among a great deal of denoising and deblurring methods introduced for digital images\\[[@ref15]--[@ref17]\\] the mentioned list seems too short for a wide variety of dental images, the quality of which is of fundamental importance in diagnosis and treatment of diseases pertaining to dentistry. Dental image denoising problem seems to be better solved if a powerful signal/noise-separating tool (e.g., wavelet analysis) is incorporated in the noise-reducing diffusion process. In simple words, multiresolution and sparsity properties of wavelet transform on the top of edge preservation ability of nonlinear diffusion can make a significant help to the noise reduction process.\\[[@ref18]\\]\n\nHaving this in mind, we proposed the application of wavelet diffusion in dental image denoising. Three different approaches to selecting the constant value of the nonlinear diffusion are proposed: the manual selection of \u03bb, semi-automatic calculation of \u03bb, and the automatic method. In the final approach, a classification of image to homogenous and nonhomogenous areas is required subject to modeling the distribution of wavelet modulus. Two novel plans in this application are proposed: Laplace-mixture model and circular symmetric Laplacian model.\n\nThe rest of this paper is organized as follows: In the \"Material\" section the material is illustrated and the proposed method is further elaborated in the \"Wavelet Diffusion,\" \"The Manual and Semiautomatic Method,\" and \"The Automatic Method\" sections. Quantitative results on natural noisy dental images are reported in the \"Experimental Results\" section. The method is discussed and conclusions are drawn in the \"Discussion and Conclusion\" section.\n\nMATERIALS AND METHODS {#sec1-2}\n=====================\n\nMaterial {#sec2-1}\n--------\n\nFor evaluation of our algorithm, we used 104 images of cephal-lateral, 38 images of anterior\\--posterior, 27 images of intraoral radiology, and 12 panoramic images. The intraoral and panoramic datasets were recorded in Dental Department of Isfahan University of Medical Sciences and the rest of the dataset is obtained in Dental Department of Shahid Beheshti University. The dental apparatus for cephal-lateral, anterior-posterior, and panoramic imaging was a Planmeca ProMax X-ray unit, and for intraoral images, a Planmeca intra X-ray unit was utilized along with two sizes of Planmeca ProSensor sets in 33.6\u00d723.4 mm and 39.7\u00d725.1 mm with a resolution of 17 lines per millimeter.\n\nWavelet Diffusion {#sec2-2}\n-----------------\n\nTo clear up our method, we should consider that the nonlinear diffusion\\[[@ref8]\\] technique relies on the gradient operator to distinguish signal from noise. Such a method often cannot achieve a precise separation of signal and noise. During last decades, discrete wavelet transform has also been introduced as a powerful denoising method;\\[[@ref19]\\] however, wavelet shrinkage suffers from ringing artifact. To eliminate this artifact several methods such as cycle spinning\\[[@ref20]\\] were suggested which usually increase the computational complexity.\n\nDental image denoising problem is better solved if a powerful signal/noise-separating tool (e.g., wavelet analysis) is incorporated in the noise-reducing diffusion process. In simple words, multiresolution and sparsity properties of wavelet transform on the top of edge preservation ability of nonlinear diffusion can make a significant help to the noise reduction process.\\[[@ref18]\\] Furthermore, the time complexity of this proposed method is considerably lower than conventional nonlinear diffusion. This combination was employed by Yue\\[[@ref21]\\] for speckle suppression in ultrasound images and it was proved that a single step of nonlinear diffusion can be considered equivalent to a single shrinkage iteration of coefficients of Mallat-Zhong dyadic wavelet transform (MZ-DWT).\\[[@ref22]\\]\n\nTwo years later in 2008, Rajpoot\\[[@ref23]\\] extended the diffusion wavelet idea by investigating the ability of conventionally used orthogonal and biorthogonal filters like Haar, Daubechies, and Coiflet (undecimated form) to be replaced by quadratic Mallat-Zhong filters. The steps of wavelet diffusion can therefore be classified in next steps:\\[[@ref23]\\]\n\nWavelet decomposition into low frequency subband (*A~j~*) and high frequency subbands (*W^i^~j~*)Regularization of high frequency coefficients (*W^i^~j~*) by multiplication with a function *p*Wavelet reconstruction from low frequency subband (*A~j~*) and regularized high frequency subbands ![](JMSS-2-103-g001.jpg)\n\nA two level structure of wavelet diffusion is shown in [Figure 1](#F1){ref-type=\"fig\"}\\[[@ref23]\\].\n\n![A two-level structure of wavelet diffusion\\[[@ref23]\\]](JMSS-2-103-g002){#F1}\n\nRegularization is the second step of the algorithm and the regularization function should be defined as *P~j~*(\\|*\u03b7~j~*(*x,y*)\\|) = 1--*g*(\\|*\u03b7~j~*(*x,y*)\\|) where *g*(\\|*\u03b7~j~*(*x,y*)\\|) can be of the form\n\n![](JMSS-2-103-g003.jpg)\n\nAnd \\|*\u03b7~j~*(*x,y*)\\| is an edge estimate for each pixel which can be approximated by the modulus of high.frequency wavelet subbands at each scale *j* by\n\n![](JMSS-2-103-g004.jpg)\n\nIn implementation of algorithms based on nonlinear diffusion (Perona-Malik diffusion, Anisotropic Diffusion, or new wavelet-based approaches), the selection of proper value for the constant \u03bb (in eq. 1) has a profound effect on performance of the denoising algorithm. We, therefore, proposed three different approaches in selection of this value. The first approach is manual selection of \u03bb and finding an acceptable result with trial and error. The second method (semiautomatic) tries in calculating of \u03bb based on the information obtained from a homogenous area in the image, the location of which should be determined by the user. And, finally, in the third approach, the homogenous area is found automatically using a likelihood classification and cross-scale edge consistence. The classification of the image to homogenous and nonhomogenous areas is subject to modeling the distribution of wavelet modulus which is proposed to be a Laplace-mixture model or a circular symmetric laplacian model.\n\nThe Manual and Semiautomatic Method {#sec2-3}\n-----------------------------------\n\nUsing a manual manner, we tried the mentioned method utilizing Haar wavelet with different values of \u03bb (in eq. 1) and for different levels of shrinkage, the outcomes of which are presented in the results.\n\nFurthermore, we examined semiautomatic and automatic methods for deciding on the value of the best \u03bb for each image. In the semiautomatic method, according to,\\[[@ref21]\\] the user choused a homogenous area of the image (X_homo), which could be simply located for dental images with a wide consistent region. Then, the wavelet transform was applied on X_homo and \\|*\u03b7~j~^homo^*(*x, y*)\\| was calculated as the modulus of high-frequency wavelet subbands, in a similar manner to \\|*\u03b7~j~*(*x, y*)\\|. Then a new scale was defined:\n\n![](JMSS-2-103-g005.jpg)\n\nNow, the best fitting \u03bb could be estimated for each level by\n\n![](JMSS-2-103-g006.jpg)\n\nThe Automatic Method {#sec2-4}\n--------------------\n\n### Rayleigh mixture model {#sec3-1}\n\nIn the automatic method for \u03bb value, Yue\\[[@ref21]\\] used a likelihood classification and cross-scale edge consistence to find the homogenous areas of an image, automatically. They tried to find a threshold for classification of image to edge-related class and the noise-related category. So, a Rayleigh model was proposed for the speckle-related modulus, since they assumed that both of speckle-related and edge-related wavelet coefficients are Gaussian distributed:\n\n![](JMSS-2-103-g007.jpg)\n\nSimilarly, *p*(*x*\\|*edge*) has the same form with \u03c3^2^~e~.\n\nTherefore the distribution of the wavelet modulus could be estimated by a Rayleigh-mixture model:\n\n*p*(*x*)=*\u03c9~n~p*(*x*\\|*noise*)+(1--\u03c9~n~)p(x\\|edge) \u2003\u2003\u2003\u2003 (6)\n\nparameters of which can be estimated by the well-known expectation-maximization (EM) method that is an iterative numerical algorithm. The requested threshold for each scale can then be estimated by\n\n![](JMSS-2-103-g008.jpg)\n\nA coarse-to-fine classification method\\[[@ref24]\\] can be used to determine the homogenous region *U~j~* :\n\n![](JMSS-2-103-g009.jpg)\n\n*K* may be changed and the higher amount of *K* make more coefficients close to edges contributed in calculation of the threshold.\n\nThe Laplace-mixture Model {#sec2-5}\n-------------------------\n\nIn this paper, we also proposed another aspect in modeling the speckle-related and edge-related modulus. In 2006, Rabbani\\[[@ref25]\\] suggested to design a maximum a posterior (MAP) estimator which relies on Laplace mixture distributions to better model the heavy-tailed property of wavelet coefficients. We use a similar idea to model the speckle-related modulus. Therefore, the wavelet modulus will be in the form of two Laplace pdfs:\n\n![](JMSS-2-103-g010.jpg)\n\nFor this mixture model, we use the EM algorithm to estimate its parameters. This iterative algorithm has two steps. **S(k,m)** denotes variable S at point k for iteration m and we start the algorithm with m = 0 (first iteration), assuming the observed data *\u03c9* (*k*), the E-step calculates the responsibility factors:\n\n![](JMSS-2-103-g011.jpg)\n\nr~2~(k,m) = 1--r~1~(k,m)\n\nThe M-step updates the parameters *a*(*k,m*), \u03c3~1~(*k,m*) and \u03c3~2~(*k,m*). *a*(*k,m*) is computed by\n\n![](JMSS-2-103-g012.jpg)\n\nwhere *M* is the number of coefficients and the parameters \u03c3~1~(*k,m*) and \u03c3~2~(*k,m*) are computed by\n\n![](JMSS-2-103-g013.jpg)\n\nThe Circular Symmetric Laplacian Model {#sec2-6}\n--------------------------------------\n\nThe most important novelty of this paper is based on modeling the speckle-related modulus with a circular symmetric Laplacian\\[[@ref26][@ref27]\\] model. It is well known that one of the most important properties of the wavelet transforms is the persistence, i.e., the large/small values of wavelet coefficients tend to propagate across scales. This property means that the bivariate pdfs, such as circular symmetric Laplacian pdf,\\[[@ref26]\\] which exploit the dependence between coefficients, better model the statistical properties of wavelet coefficients in comparison with univariate pdfs.\\[[@ref26]\\] In this paper, we use a mixture of two circular symmetric Laplacian pdfs to describe the above property. Then we define a parent-child schematics of the wavelet domain as illustrated in [Figure 2](#F2){ref-type=\"fig\"}.\n\n![Illustration of neighborhood\\[[@ref26]\\]](JMSS-2-103-g014){#F2}\n\nConsequently, *\u03c9* in (9) should now be described by ![](JMSS-2-103-g015.jpg):\n\n![](JMSS-2-103-g016.jpg)\n\nwhere *\u03c9*~1~(*k*) and *\u03c9*~2~(*k*) are the values in parent and child scales. The following circular symmetric Laplacian pdf with local variance is proposed in order to describe that *\u03c9*~1~(*k*) and *\u03c9*~2~(*k*) are uncorrelated while are dependent\\[[@ref26]\\]:\n\n![](JMSS-2-103-g017.jpg)\n\nThe bivarate mixture model then can be written as\n\n![](JMSS-2-103-g018.jpg)\n\nIn the EM algorithm, the formulas will change to\n\n![](JMSS-2-103-g019.jpg)\n\n*r*~2~(*k,m*) = 1--*r*~1~(*k,m*)\n\n![](JMSS-2-103-g020.jpg)\n\nEXPERIMENTAL RESULTS {#sec1-3}\n====================\n\nIn this section, we demonstrate the performance of the proposed algorithms on different dental images. According to a great deal of similar papers, the denoising performance may be evaluated in different ways\\[[@ref16]\\] like the method noise, the mean square error, PSNR, contrast-to-noise ratio (CNR), and the visual quality of the restored images. We chose CNR and the visual quality for this performance because this application is designed for natural noisy images (i.e., the noisy images are not constructed manually); therefore, methods like PSNR and mean-square error are not acceptable since we have no golden noise-free image to which we may compare the denoised results of the proposed algorithm.\n\nIn order to have a meaningful comparison between mentioned algorithms, we use contrast-to-noise ratio (CNR)\\[[@ref28]\\] defined as\n\n![](JMSS-2-103-g021.jpg)\n\nwhere *\u03bc~ROI~* is the mean signal value computed for a small region of interest (ROI). The desired ROI can be a homogeneous area of tissue with high signal intensity. The noise standard deviation (std) \u03c3 is computed from a large region outside the object, which represents the background noise. The CNR represents the contrast between two ROIs. [Figure 3](#F3){ref-type=\"fig\"} shows the images from which we selected ROI samples. The white circles show the noise ROIs and the black ones indicate the ROI representing a homogeneous area of tissue. [Table 1](#T1){ref-type=\"table\"} shows the CNR values for different dental images using discussed methods.\n\n![The selected ROI samples](JMSS-2-103-g022){#F3}\n\n###### \n\nThe CNR values of various automatic methods\n\n![](JMSS-2-103-g023)\n\nIn [Figure 4](#F4){ref-type=\"fig\"}, the results of the manual method described in the \"The Manual and Semiautomatic Method\" section are shown. Figures [5](#F5){ref-type=\"fig\"}--[8](#F8){ref-type=\"fig\"} are also demonstrating the performance of the automatic algorithm using the Rayleigh mixture model, Laplace mixture, and circular symmetric Laplacian.\n\n![The results of choosing different values of *\u03bb* for different levels in the manual method](JMSS-2-103-g024){#F4}\n\n![The comparison of different automatic methods in anterior-posterior images. Left image is the original one, the performance of the automatic algorithm with the Rayleigh mixture model (middle left), Laplace mixture (middle right), and circular symmetric Laplacian (most right)](JMSS-2-103-g025){#F5}\n\n![The comparison of different automatic methods in Cephal-lateral images. The left image is the original one, the performance of the automatic algorithm with the Rayleigh mixture model (middle left), Laplace mixture (middle right), and circular symmetric Laplacian (most right)](JMSS-2-103-g026){#F6}\n\n![The comparison of different automatic methods in intraoral images. The left image is the original one, the performance of the automatic algorithm with the Rayleigh mixture model (middle left), Laplace mixture (middle right), and circular symmetric Laplacian (most right)](JMSS-2-103-g027){#F7}\n\n![The comparison of different automatic methods in panoramic images. The left image is the original one, the performance of the automatic algorithm with the Rayleigh mixture model (middle left), Laplace mixture (middle right), and circular symmetric Laplacian (most right)](JMSS-2-103-g028){#F8}\n\nThe other important aspect considered in dental image denoising is the ability of the filtering method in retaining the possible cavities on dental images. For this purpose, we collected new series of intraoral images with natural cavities on them and examined the cavity preserving ability of the method. Furthermore, we made radiographic images from a healthy extracted dent and then applied an artificial cavity to determine the mentioned ability. Figure [9](#F9){ref-type=\"fig\"} and [10](#F10){ref-type=\"fig\"} demonstrate the results of the circular symmetric Laplacian mixture model method for cavity preserving. As it can be seen in [Figure 9](#F9){ref-type=\"fig\"}, the left column demonstrates two extracted dents with natural intrinsic cavities and the right column shows the filtered images with the proposed method, preserving the local cavities. Similarly, it is obvious in [Figure 10](#F10){ref-type=\"fig\"} that the manually applied cavity ([Fig. 10](#F10){ref-type=\"fig\"}, left) is filtered with the new method and the result ([Fig. 10](#F10){ref-type=\"fig\"}, right) has a good preservation on this cavity. The main advantage of this aspect is the reality that in clinical applications it is an important feature not to miss the cavities due to denoising and unfortunately this problem is prevalent in most of common denoising techniques which are usually found in software sold along with dental imaging equipment.\n\n![Cavity preserving in natural cavities (left column: original images, right column: filtered images)](JMSS-2-103-g029){#F9}\n\n![Cavity preserving in artificial cavities (left: original image, right: filtered image)](JMSS-2-103-g030){#F10}\n\nDISCUSSION AND CONCLUSION {#sec1-4}\n=========================\n\nIn this paper, a particular combination of wavelet shrinkage and nonlinear diffusion for dental image denoising was proposed. For this purpose, wavelet diffusion was used and its threshold is selected automatically using the Laplacian mixture model and circular symmetric Laplacian mixture models for speckle-related modulus. The circular symmetric Laplacian mixture model could make a better model of data because of its compatibility with heavy tailed structure of wavelet coefficients besides their interscale dependence and produced better CNR in comparison with other models. The method also has an acceptable speed even when implemented in MATLAB\u2122 without using mex files -- Math Works, Inc., Natick, MA, USA.\\[[@ref29]\\] This sounds quite promising for real-time application of this method in dental offices.\n\nBIOGRAPHIES {#sec1-5}\n===========\n\n![](JMSS-2-103-g031){#F11}\n\n**Raheleh Kafieh** received the BS degree from Sahand University of Technology, Iran, in 2005 and the MS degree from the Isfahan University of Medical Sciences, Iran, in 2008, both in biomedical engineering. She is currently working toward the PhD degree in the Department of Biomedical Engineering at Isfahan University of Medical Sciences. Her research interests are in biomedical image processing, computer vision, graph algorithms, and sparse transforms. She is a student member of the IEEE and the IEEE Signal Processing Society.\n\n**E-mail:** <r_kafieh@resident.mui.ac.ir>\n\n![](JMSS-2-103-g032){#F12}\n\n**Hossein Rabbani** is an Associate Professor at Isfahan University of Medical Sciences, in Biomedical Engineering Department also Medical Image & Signal Processing Research Center. Involved research topics include medical image/volume processing, noise reduction and estimation problem, image enhancement, blind deconvolution, video restoration, probability models of sparse domain\\'s coefficients especially complex wavelet coefficients. He is a member of IEEE, Signal Processing Society, Engineering in Medicine and Biology Society, and Circuits and Systems Society.\n\n**E-mail:** <h_rabbani@med.mui.ac.ir>\n\n![](JMSS-2-103-g033){#F13}\n\n**Mehrdad Foroohandeh** is a dental student since 2007 at dental Department, Isfahan University of Medical Sciences. His research interests are in dental image process and oral cancer.\n\n**E-mail:** <mehrdad_foroohande@yahoo.com>\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared\n"}
+{"text": "Key Points {#FPar1}\n==========\n\nAsenapine monotherapy is approved in the US for the acute treatment of manic and mixed episodes of bipolar I disorder in children and adolescents aged 10--17\u00a0years.Although asenapine has been studied in pediatric patients with schizophrenia, it is not approved for this indication.Asenapine is administered twice daily as a fast-dissolving tablet that is placed under the tongue; generally transient oral effects related to sublingual administration (e.g., oral hypoesthesia, dysgeusia, paresthesia) have been observed.Absolute bioavailability is markedly decreased if asenapine is swallowed, so it should not be ingested; exposure can also be reduced by food or drink, so eating and drinking should be avoided for at least 10\u00a0min after administration.Asenapine was generally safe and well tolerated in clinical trials of pediatric patients with bipolar I disorder or schizophrenia; sedation, somnolence, weight gain, and changes in some metabolic parameters have been observed and should be monitored.\n\nIntroduction {#Sec1}\n============\n\nPsychiatric disorders that are commonly treated with pharmacotherapy in adulthood often have onset in childhood or adolescence \\[[@CR1]\\]. Similar to treatment recommendations for adult patients with schizophrenia or bipolar mania, practice parameters addressing the early onset of these disorders recommend the use of antipsychotic agents as a first-line treatment \\[[@CR2], [@CR3]\\]. The relative scarcity of safety and efficacy information for antipsychotic use in juveniles means that prescribers may have to rely on data from studies in adults to treat their young patients; this lack of information is concerning since antipsychotics may have effects that are developmentally dependent and differ from what is observed in adult patients \\[[@CR4]\\].\n\nData indicate that 14--28% of adult patients with bipolar disorder experience onset of the illness before the age of 13\u00a0years, and 36--38% experience onset between the ages of 13 and 18\u00a0years \\[[@CR5], [@CR6]\\]. Early onset of bipolar disorder, which is generally associated with a delay in first treatment and a more pernicious course of illness than later onset illness, is characterized by greater comorbidity with other psychiatric disorders and substance abuse, more mood episodes, more days depressed, shorter periods of euthymia, and higher lifetime risk of suicide attempt \\[[@CR6], [@CR7]\\].\n\nThe onset of schizophrenia before the age of 13\u00a0years appears to be rare, with a prevalence of approximately 1 in 40,000 \\[[@CR8]\\]. Rates of illness increase during adolescence, with the peak age of onset for schizophrenia cited as ranging from 15 to 30\u00a0years \\[[@CR3]\\]. It has been reported that up to 20% of individuals with schizophrenia become ill before the age of 18\u00a0years \\[[@CR9]\\], and early onset is associated with poor long-term prognosis, social deficits, and high suicidality \\[[@CR3], [@CR10]\\].\n\nAsenapine is a novel second-generation antipsychotic that is administered twice daily (BID) as a sublingual tablet. As the only second-generation antipsychotic with this unique formulation, asenapine is available as 2.5-, 5-, and 10-mg black-cherry-flavored tablets that dissolve within seconds when placed under the tongue. Rapidly absorbed through the oral mucosa, in adult patients asenapine reaches peak plasma concentration in 30--90\u00a0min following a single 5-mg dose, and steady state is reached within 3 days of BID administration \\[[@CR11]\\]. Absolute bioavailability is markedly decreased if asenapine is swallowed, so it should not be ingested; exposure can also be reduced by food or drink, so eating and drinking should be avoided for at least 10\u00a0min after administration \\[[@CR11]\\].\n\nUS Food and Drug Administration (FDA) approval of asenapine monotherapy in adult patients was based on efficacy demonstrated in two 6-week, double-blind, placebo- and active-controlled studies in schizophrenia \\[[@CR12], [@CR13]\\] and two 3-week, double-blind, placebo- and active-controlled studies in manic or mixed episodes associated with bipolar I disorder \\[[@CR14], [@CR15]\\]. Maintenance efficacy was demonstrated in a 26-week relapse prevention study in schizophrenia \\[[@CR16]\\] and a 26-week relapse prevention study in bipolar I disorder \\[[@CR17]\\]; these studies served as the basis for subsequent indications for maintenance treatment in each disease state in adults \\[[@CR11]\\]. Additional short-term and longer-term studies support the safety and efficacy of asenapine in adult patients with schizophrenia \\[[@CR18], [@CR19]\\] and bipolar I disorder \\[[@CR20]--[@CR23]\\]. Asenapine is also approved for the treatment of adults in the European Union (moderate to severe manic episodes associated with bipolar disorder), in Canada (as monotherapy or in combination with lithium or divalproex for manic or mixed episodes associated with bipolar I disorder), in Japan (schizophrenia), and in Australia (schizophrenia, and as monotherapy or in combination with lithium or sodium valproate for acute mania associated with bipolar I disorder and prevention of relapse of manic or mixed episodes in bipolar I disorder).\n\nClinicians who treat pediatric and adolescent patients with psychiatric disorders should be aware that asenapine is approved in the US as monotherapy for the acute treatment of manic and mixed episodes of bipolar I disorder in children and adolescents aged 10--17\u00a0years. Although asenapine has been studied in pediatric patients with schizophrenia, it did not separate from placebo on the primary efficacy parameter in a double-blind randomized trial \\[[@CR24]\\] and is not approved for this indication. Asenapine is not approved for pediatric use in bipolar I disorder or schizophrenia in any other major market. A recent literature search using the terms asenapine, pediatric, adolescent, childhood, bipolar I disorder, and schizophrenia yielded scant results. Consequently, we reviewed and collated the available information on the neuropharmacology, pharmacokinetics, clinical trials, and clinical use of asenapine in pediatric patients to assist clinicians who treat young patients with these disorders.\n\nOverview of Asenapine {#Sec2}\n=====================\n\nNeuropharmacology {#Sec3}\n-----------------\n\nAlthough the mechanism of action for asenapine in schizophrenia and bipolar disorder is unknown, it is suggested to be the result of combined antagonist activity at dopamine (D~2~) and serotonin (5-HT~2A~) receptors \\[[@CR11]\\]. The receptor profile of asenapine is complex, and beyond the common features of D~2~ and 5-HT~2A~ receptor affinity, it is different than other second-generation antipsychotics \\[[@CR25]\\]. Asenapine displays high antagonist activity for other dopamine (D~1~, D~3~, and D~4~), serotonin (5-HT~1A~, 5-HT~1B~, 5-HT~2B~, 5-HT~2C~, 5-HT~5~, 5-HT~6~, and 5-HT~7~), adrenergic (\u03b1~1~ and \u03b1~2~), and histamine (H~1~) receptors and moderate antagonistic affinity for the H~2~ receptor; it has no appreciable affinity for muscarinic receptors. This receptor binding profile may theoretically predict a low propensity for causing anticholinergic effects through low affinity at muscarinic receptors \\[[@CR25], [@CR26]\\]. Conversely, antagonism at the histamine H~1~ and \u03b1~1~-adrenergic receptors may cause sedation and H~1~ receptor blockade may be related to weight gain; antagonism of \u03b1~1~-adrenergic receptors may also be associated with cardiovascular effects \\[[@CR25], [@CR26]\\].\n\nPharmacokinetic Parameters {#Sec4}\n--------------------------\n\n### Pediatric Pharmacokinetic Studies {#Sec5}\n\nTwo phase I studies investigated the pharmacokinetic profile of asenapine in pediatric patients \\[[@CR27], [@CR28]\\]; a summary of findings is presented in Table\u00a0[1](#Tab1){ref-type=\"table\"}. Data from these two studies showed that the pharmacokinetic profile of asenapine in pediatric patients is similar to that in adult patients in other studies \\[[@CR11], [@CR25], [@CR29]\\]. This important clinical characteristic suggests that the recommended adult dosage of asenapine can be given to pediatric and adolescent patients without adjustment \\[[@CR11]\\].Table\u00a01Summary of pharmacokinetic observations in pediatric patientsDosePharmacokinetic parameters*C* ~max~ (ng/mL)*T* ~max~ (h)AUC (h\u00b7ng/mL)*t* ~\u00bd~ (h)Vd/F (L)Time to steady state (d)CL/F (L/h)Study 1 \\[[@CR28]\\]\u00a03\u00a0mg2.60.815.825.612,100--^a^272\u00a05\u00a0mg3.51.222.932.314,700--^a^409\u00a010\u00a0mg2.81.419.722.619,7008618Study 2 \\[[@CR27], [@CR28]\\]\u00a02.5\u00a0mg1.81.111.422.083206264\u00a05\u00a0mg3.51.923.618.559406218\u00a010\u00a0mg7.81.144.220.110,1007--11294Data on file, Allergan*AUC* area under the concentration--time curve, *CL/F* apparent clearance, *C* ~*max*~ peak plasma concentration, *T* ~*max*~ time to *C* ~*max*~, *t* ~*\u00bd*~ terminal half-life, *Vd/F* apparent volume of distribution^a^Time to steady state not calculated for 3- and 5-mg doses\n\nIn the first study, 40 adolescent patients (12--17\u00a0years of age) with a history of schizophrenia, bipolar disorder, or any condition for which chronic use of antipsychotic medication was warranted received 10\u00a0days of double-blind asenapine (1, 3, 5, or 10\u00a0mg BID) or matching placebo. Results indicated that asenapine exposure (maximum concentration \\[*C* ~max~\\] and area under the concentration--time curve, 0--12\u00a0h \\[AUC~0--12~\\]) tended to increase with increasing doses from 1\u00a0mg BID up to and including 5\u00a0mg BID, although the increase was slightly less than dose proportional \\[[@CR28]\\]. Lower asenapine exposure was observed in the 10-mg BID dose group than in the 5-mg BID dose group. Although the exact reason for this is unclear, it is possible that patients in this group swallowed a greater portion of the administered dose. Namely, asenapine undergoes significant first-pass metabolism when given orally; first-pass metabolism is minimized when asenapine is administered sublingually and higher plasma asenapine concentrations are achieved \\[[@CR25]\\]. For asenapine metabolites such as desmethylasenapine, exposure (AUC) following either route of administration is expected to be similar. As such, the finding that lower *C* ~max~ and AUC~0--12~ values for asenapine, but not for desmethylasenapine, were observed in the 10-mg dose group may support the oral ingestion hypothesis. Rapid absorption and an initial rapid decline in plasma concentrations was followed by a slower elimination phase; time to maximum asenapine plasma concentration was \\~1\u00a0h (range 0.7--1.3\u00a0h) and steady state was attained in \\~8 days (data on file, Allergan). Multiple sublingual doses of asenapine were safe and well tolerated in adolescent subjects with a disorder that warranted the use of antipsychotic medication \\[[@CR28]\\].\n\nTo further investigate the pharmacokinetics of asenapine 10\u00a0mg BID in subjects aged 12--17\u00a0years and to evaluate asenapine pharmacokinetics in younger subjects (10--11\u00a0years of age), a second study was conducted \\[[@CR28]\\]. Thirty pediatric patients (10--17\u00a0years of age) with a documented history of psychiatric disorder, including bipolar disorder, schizophrenia, autism, and conduct disorder, participated in this 7-, 8-, or 12-day open-label, sequential group, rising multiple-dose study \\[[@CR27]\\]. Multiple sublingual doses of asenapine 2.5\u00a0mg BID, 5\u00a0mg BID, or escalating doses up to 10\u00a0mg BID were administered to the 10- to 11-year-old subjects; after receiving 5\u00a0mg BID on day 1, multiple sublingual doses of asenapine 10\u00a0mg BID were administered to the 12- to 17-year-old subjects. Results indicated that asenapine was rapidly absorbed (time to maximum concentration \\[*T* ~max~\\] \\~1\u00a0h) and had an apparent terminal half-life of \\~20\u00a0h. At the 10-mg BID dose, exposure to asenapine was similar across the age groups from 10--17 years, but maximum asenapine concentrations were \\~30% higher in 10- to 11-year-old subjects compared with the older age group. Asenapine was generally safe and well tolerated in pediatric subjects. Of note, eight events of dystonia were reported in seven subjects (five events in the 10- to 11-year age group receiving 5\u00a0mg BID; three events in the 12- to 17-year age group receiving 10\u00a0mg BID). As such, a low initial dose (2.5\u00a0mg BID) followed by a short up-titration schedule became the recommended dosing schedule to circumvent initial sensitivity to asenapine treatment \\[[@CR11]\\].\n\nOf note in adult subjects, a dosage increase from 5 to 10\u00a0mg BID results in less than linear increases in the extent of exposure and *C* ~max~ (1.7 times) \\[[@CR11]\\]. Overall, data from the two designated pediatric studies showed that the pharmacokinetic profile of asenapine in pediatric patients is similar to that in adult patients in other studies \\[[@CR11], [@CR25], [@CR29]\\]. This is an important clinical characteristic that suggests that the recommended adult dosage of asenapine can be given to pediatric and adolescent patients without adjustment \\[[@CR11]\\].\n\n### Bioavailability and Metabolism {#Sec6}\n\nGiven its sublingual method of administration, the bioavailability of asenapine is dependent on various factors including the amount of saliva in the mouth, food and water intake, and dissolving versus ingestion of the tablet. It is particularly important to note that, in adults, the absolute bioavailability of asenapine decreases markedly (to \\<2%) if it is swallowed \\[[@CR11]\\]; bioavailability can also be reduced by \\~12--20% with the intake of water within 2--5\u00a0min after administration \\[[@CR25], [@CR30]\\]. Additionally, a high-fat meal immediately before sublingual administration in healthy adults reduced asenapine exposure by 20%; and, when food was given 4\u00a0h after asenapine was administered, AUC was reduced by 13% \\[[@CR31]\\]. These effects are likely due to increased clearance of asenapine related to increased hepatic blood flow following food intake and are expected to be similar in pediatric patients. As such, adult and pediatric patients should avoid drinking and eating for 10\u00a0min after asenapine is administered \\[[@CR11]\\].\n\nAlthough asenapine has multiple metabolites, none are active or considered clinically relevant, with asenapine activity predominantly due to the parent drug \\[[@CR25]\\]. The primary metabolic pathways for asenapine are direct glucuronidation by uridine 5\u2032-diphospho-glucuronosyltransferase 1A4 (UGT1A4) and oxidative metabolism by cytochrome P450 (CYP450) isoenzymes (predominantly CYP1A2) \\[[@CR11]\\]. Asenapine is rapidly distributed, has a large volume of distribution, and is highly bound to plasma proteins (95%) \\[[@CR11], [@CR32]\\]. Elimination of asenapine and its metabolites occurs approximately equally via hepatic and renal routes \\[[@CR30]\\].\n\n### Population Pharmacokinetics {#Sec7}\n\nThe population pharmacokinetics of asenapine in pediatric patients has been modeled to determine whether dose adjustments based on differences in age or weight are warranted \\[[@CR28]\\]. Pediatric patient data from the two dedicated phase I pediatric pharmacokinetic studies, one phase III acute study in bipolar I disorder \\[[@CR33]\\], and one phase III acute study in schizophrenia were included in the model (2451 samples from 561 pediatric patients) \\[[@CR24]\\]. The final population pharmacokinetic model was described as a two-compartment model with first-order absorption and first-order elimination. Building on an existing population pharmacokinetic model in adults \\[[@CR34]\\], analyses were conducted using age, body mass index (BMI), gender, race, and dose as covariates. Simulations were performed to determine the effect of covariates on asenapine pharmacokinetics and compare steady-state exposure between pediatric and adult subjects. Results showed no statistically significant differences or clinically meaningful changes in asenapine exposure for age, BMI, race, and gender. Predictive checks indicated that the model adequately described the observed asenapine data and indicated that no dose adjustments are required based on the covariates tested.\n\n### Drug--Drug Interactions and Contraindications {#Sec8}\n\nAsenapine is a substrate for UGT1A4, CYP1A2, and to a lesser extent CYP3A4 and CYP2D6; it is a weak inhibitor of CYP2D6 \\[[@CR25]\\]. Asenapine does not cause induction of CYP1A2 or CYP3A4. The dose of paroxetine (CYP2D6 substrate and inhibitor) should be reduced by half when it is coadministered with asenapine. When coadministered with a strong CYP1A2 inhibitor (e.g., fluvoxamine), asenapine exposure may be increased, and a dosage reduction may be necessary based on clinical response. However, no asenapine dosage adjustments are required if asenapine is concomitantly administered with imipramine (CYP1A2/2C19/3A4 inhibitor), cimetidine (CYP3A4/2D6/1A2 inhibitor), valproate, lithium, or a CYP34A inducer (e.g., carbamazepine, phenytoin, rifampin) \\[[@CR11]\\]. Asenapine is contraindicated for patients with severe hepatic impairment because exposure was more than\u00a07-fold greater in these patients than in patients with normal hepatic function; however, no dose adjustment is needed in patients with mild or moderate hepatic impairment or in patients with renal impairment \\[[@CR11]\\].\n\nPediatric Clinical Trial Experience {#Sec9}\n===================================\n\nAsenapine has been studied in pediatric and adolescent patients with manic and mixed episodes associated with bipolar I disorder and schizophrenia (paranoid, disorganized, or undifferentiated subtype) (Table\u00a0[2](#Tab2){ref-type=\"table\"}).Table\u00a02Summary of asenapine clinical trials in pediatric patientsStudy/Clinical Trials.gov identifierStudy duration and designPatientsNumber of patientsEfficacy outcomeSafety findings*Bipolar I disorder*Findling et al. \\[[@CR33]\\]\\\nNCT012448153 weeks, randomized (1:1:1:1), double-blind, placebo-controlled, parallel-group10--17 years of age with a current manic or mixed episode, with or without psychotic features, associated with bipolar I disorderRandomized: 403\\\nPlacebo: 101\\\nAsenapine BID\\\n2.5\u00a0mg: 104\\\n5\u00a0mg: 99\\\n10\u00a0mg: 99LSMD (95% CI) in YMRS total score at day 21 was statistically significant for each asenapine dose versus placeboThere were no deaths and few SAEs; common asenapine TEAEs (\\>5% and twice placebo) were somnolence, sedation, oral hypoesthesia, oral paresthesia, and increased appetiteFindling et al. \\[[@CR35]\\]\\\nNCT0134990750 weeks, open-label safety extension10--18 years of age who completed the acute mania study322 entered open-label treatment (321 patients were included in the analyses); 80 had received placebo in the acute study (placebo/asenapine group) and 241 had received asenapine (asenapine/asenapine group)Descriptive statistics suggest that improvement in mania seen during acute treatment was maintained over the course of the extension trialThere were no deaths; SAEs were reported in 6.9% of patients; somnolence, weight gain, sedation, and headache were the most commonly reported TEAEs*Schizophrenia*Findling et al. \\[[@CR24]\\]\\\nNCT011902548 weeks, randomized (1:1:1), double-blind, placebo-controlled, parallel-group12--17 years of age with schizophreniaRandomized: 306\\\nPlacebo: 102\\\nAsenapine BID\\\n\u00a02.5\u00a0mg: 98\\\n\u00a05\u00a0mg: 106LSMD (95% CI) in PANSS total score at week 8 was not statistically significant for either asenapine dose versus placeboThere were no deaths and the incidence of SAEs was low and similar across treatment groups (\u22643%); the most common TEAEs overall were nervous system events (e.g., sedation and somnolence)Findling et al. \\[[@CR24]\\]\\\nNCT0119026726 weeks, open-label safety extension12--17 years of age who completed the acute schizophrenia study196 patients entered open-label treatment; 62 had received placebo in the acute study (placebo/asenapine group) and 134 had received asenapine (asenapine/asenapine group)Numerical improvements in PANSS total score that were seen in the acute trial were maintained during open-label treatmentOne death (accidental; not related to the study drug) occurred in a placebo/asenapine patient; SAEs were reported in 4% of patients; somnolence/hypersomnia/sedation combined was the most common TEAE of interest in long-term treatment*BID* twice daily, *CI* confidence interval, *LSMD* least squares mean difference, *PANSS* Positive and Negative Syndrome Scale, *SAE* serious adverse event, *TEAE* treatment-emergent adverse event, *YMRS* Young Mania Rating Scale\n\nBipolar I Disorder {#Sec10}\n------------------\n\n### Acute Study {#Sec11}\n\nAsenapine was approved for use in pediatric patients with bipolar I disorder based on efficacy and safety demonstrated in one short-term study (ClinicalTrials.gov: NCT01244815) \\[[@CR33]\\]; long-term safety has also been investigated in a 50-week, open-label extension trial (NCT01349907) \\[[@CR35]\\]. The short-term study was a 3-week, double-blind, placebo-controlled, parallel-group trial in pediatric patients with a current manic or mixed episode, with or without psychotic features, associated with bipolar I disorder. Patients were randomized (1:1:1:1) to placebo or asenapine 2.5\u00a0mg BID, 5\u00a0mg BID, or 10\u00a0mg BID; the primary and secondary efficacy measures were change from baseline to day 21 in Young Mania Rating Scale (YMRS) \\[[@CR36]\\] total score and Clinical Global Impressions-Bipolar Version (CGI-BP) \\[[@CR37]\\], respectively.\n\nStudy completion was similar among groups (placebo 86.1%; asenapine: 2.5\u00a0mg BID 84.6%, 5\u00a0mg BID 88.9%, 10\u00a0mg BID 87.9%); adverse events (AEs) were the most common reason for discontinuation in each group (placebo 4.0%; asenapine: 2.5\u00a0mg BID 6.7%, 5\u00a0mg BID 5.1%, 10\u00a0mg BID 5.1%). Statistically significant improvement in manic symptoms was demonstrated by the least squares mean difference (LSMD) and associated 95% confidence interval (CI) in YMRS total score at day 21 for each dose of asenapine versus placebo (LSMD \\[95% CI\\]: 2.5\u00a0mg BID \u22123.2 \\[\u22125.6 to \u22120.8\\], *p*\u00a0=\u00a00.008; 5\u00a0mg BID \u22125.3 \\[\u22127.7 to \u22122.9\\], *p*\u00a0\\<\u00a00.001; 10\u00a0mg BID \u22126.2 \\[\u22128.6 to \u22123.8\\], *p*\u00a0\\<\u00a00.001). The odds ratio (OR) and associated 95% CI for YMRS response (50% total score improvement from baseline) was statistically significant in favor of asenapine versus placebo across the dose groups (asenapine: 2.5\u00a0mg BID 1.9 \\[1.0 to 3.4\\], *p*\u00a0=\u00a00.042, number needed to treat \\[NNT\\] 8; 5\u00a0mg BID 3.2 \\[1.7 to 5.8\\], *p*\u00a0\\<\u00a00.001, NNT 4; 10\u00a0mg BID 2.9 \\[1.6 to 5.3\\], *p*\u00a0\\<\u00a00.001, NNT 5). Similarly, a greater reduction in severity of illness was shown by statistically significant LSMDs versus placebo in change from CGI-BP baseline for each asenapine group (2.5\u00a0mg BID \u22120.6 \\[\u22120.9 to \u22120.3\\], *p*\u00a0\\<\u00a00.001; 5\u00a0mg BID \u22120.8 \\[\u22121.0 to \u22120.5\\], *p*\u00a0\\<\u00a00.001; 10\u00a0mg BID \u22120.7 \\[\u22121.0 to \u22120.4\\], *p*\u00a0\\<\u00a00.001).\n\nAsenapine was generally safe and well tolerated in pediatric patients who participated in the study. No deaths and few serious adverse events (SAEs) occurred (placebo 3 \\[3.0%\\]; asenapine: 2.5\u00a0mg BID 0 \\[0%\\], 5\u00a0mg BID 2 \\[2.0%\\], 10\u00a0mg BID 2 \\[2.0%\\]); all SAEs were related to psychiatric disorders (worsening of bipolar 1 disorder/bipolar disorder \\[2 placebo patients, 1 asenapine 5.0-mg BID patient, 1 asenapine 10.0-mg BID patient\\], mania \\[1 asenapine 5.0-mg BID patient\\], suicidal behavior and suicidal ideation \\[2 events in 1 placebo patient\\], suicide attempt \\[1 asenapine 10.0-mg BID patient\\]). AEs occurred with greater frequency in the asenapine groups (2.5\u00a0mg BID 76.0%, 5\u00a0mg BID 72.7%, 10\u00a0mg BID 86.9%) than in the placebo group (55.4%). Some AEs were predefined as events of special interest; of these AEs, incidences of oral hypoesthesia/dysgeusia combined and somnolence/sedation/hypersomnia combined were significantly higher for all three asenapine treatment groups compared with the placebo treatment group (*p*\u00a0\\<\u00a00.05). The incidence of dizziness was greater in the 5-mg BID asenapine treatment group compared with the placebo group (*p*\u00a0\\<\u00a00.05). Of potential clinical consequence, body weight increase \u22657% from baseline was significantly higher in patients in each asenapine treatment group (2.5\u00a0mg BID 12.0%, 5\u00a0mg BID 8.9%, 10\u00a0mg BID 8.0%) compared with patients in the placebo treatment group (1.1%) (*p*\u00a0\\<\u00a00.05 for each group). There were no significant differences between the asenapine and placebo groups for extrapyramidal symptoms (EPS) defined by the narrow risk set, which included akathisia, dyskinesia, dystonia, and Parkinson-like events; similarly, no significant differences between groups were observed for akathisia or insomnia.\n\nTreatment-emergent AEs (TEAEs) that occurred at an incidence \u22655% and twice the rate of placebo in all three asenapine groups were somnolence, sedation, oral hypoesthesia, oral paresthesia, and increased appetite. The oral TEAEs, which are related to the local anesthetic properties of the sublingual formulation of asenapine, are generally transient and only caused treatment discontinuation in one patient (paresthesia in the 2.5-mg group). Treatment-emergent EPS were reported as AEs for three (3.0%) placebo-treated patients and six (5.8%), five (5.1%), and six (6.1%) asenapine-treated patients in the 2.5-, 5-, and 10-mg BID dose groups, respectively; these AEs were considered mild to moderate in intensity in all but one patient (asenapine 2.5\u00a0mg BID: severe akathisia resulting in discontinuation).\n\nLarger mean increases in weight, BMI, fasting cholesterol, triglycerides, fasting glucose, and fasting insulin were seen in asenapine- versus placebo-treated patients (Table\u00a0[3](#Tab3){ref-type=\"table\"}). A potential dose--response relationship with asenapine was observed for increases in aspartate aminotransferase and alanine aminotransferase. Insulin level shifts \\>1.2 times the upper limit of normal occurred in greater percentages of asenapine-treated patients (2.5\u00a0mg BID 31.8%, 5\u00a0mg BID 32.2%, 10\u00a0mg BID 24.4%) than in placebo-treated patients (17.7%), suggesting a treatment-related effect.Table\u00a03Acute bipolar I disorder study: changes in laboratory parametersLaboratory valuesPlacebo\\\n*n*\u00a0=\u00a0101\\\nMean (SD)Asenapine2.5\u00a0mg BID\\\n*n*\u00a0=\u00a0104\\\nMean (SD)5\u00a0mg BID\\\n*n*\u00a0=\u00a099\\\nMean (SD)10\u00a0mg BID\\\n*n*\u00a0=\u00a099\\\nMean (SD)Cholesterol, fasting, mmol/L\u22120.060 (0.526)0.095 (0.493)0.186 (0.535)0.240 (0.462)Triglycerides, fasting, mmol/L\u22120.074 (0.443)0.098 (0.544)0.151 (0.605)0.166 (0.526)Glucose, fasting, mmol/L\u22120.12 (0.49)0.08 (0.56)\u22120.02 (0.70)0.02 (0.72)Hemoglobin A1c, m/L\u22120.0005 (0.002)\u22120.0002 (0.002)\u22120.0002 (0.002)\u22120.0001 (0.002)Insulin, fasting, pmol/L3.69 (93.61)73.38 (304.66)114.04 (265.92)59.85 (211.30)Prolactin, \u00b5g/L2.53 (12.89)3.21 (13.94)2.09 (12.71)6.41 (16.49)Aspartate aminotransferase, U/L\u22121.2 (8.1)0.6 (8.1)0.4 (9.4)3.4 (12.9)Alanine aminotransferase, U/L\u22121.0 (9.1)1.3 (7.4)3.9 (12.8)8.8 (26.7)Weight increase, kg0.48 (1.57)1.72 (2.04)1.62 (2.17)1.44 (1.91)BMI increase, kg/m^2^0.06 (0.72)0.60 (0.79)0.57 (0.89)0.49 (0.81)Adapted from Findling et al. \\[[@CR33]\\]*BID* twice daily, *BMI* body mass index, *SD* standard deviation\n\nSuicidal ideation assessed by the Columbia Suicide Severity Rating Scale (C-SSRS) \\[[@CR38]\\] was reported in five (5.0%) placebo patients and five (4.8%), six (6.1%), and eight (8.1%) asenapine-treated patients in the 2.5-, 5-, and 10-mg BID groups, respectively; suicidal behavior was reported by two patients (1 placebo and 1 asenapine 10\u00a0mg BID). TEAEs of suicidal ideation were reported in nine patients (1 placebo, 4 asenapine 2.5\u00a0mg BID, 1 asenapine 5\u00a0mg BID, 3 asenapine 10\u00a0mg BID); each TEAE had a corresponding positive C-SSRS response. One SAE of suicidal behavior was reported (placebo patient; discontinued); one suicide attempt, which did not result in study discontinuation and was not considered related to the study drug, was reported in a patient receiving asenapine 10\u00a0mg BID.\n\n### Long-Term Safety Study {#Sec12}\n\nPatients who completed the acute study could enroll in a flexible-dose (2.5--10\u00a0mg BID), 50-week, open-label extension study \\[[@CR35]\\]. Although the primary objective of this trial was to assess the long-term safety and tolerability of asenapine, efficacy outcomes were collected, including change from open-label baseline on the YMRS and CGI-BP. Of the 350 patients who completed the acute trial, 322 (92.0%) patients entered the open-label extension and were treated with asenapine; 321 patients were included in the analyses, of whom 80 had been treated with placebo (placebo/asenapine) and 241 had been treated with asenapine (asenapine/asenapine) in the acute trial. A total of 181 (56.4%) patients prematurely discontinued from the study; the most common reasons for discontinuation were AEs (15.0%), nonadherence to the study protocol (14.3%), and withdrawal of consent (11.8%). Overall asenapine exposure ranged from 2 to 380 days, with a mean duration of treatment of 187 days. The mean average daily dose was 12.8\u00a0mg for placebo/asenapine patients and 14.1\u00a0mg for asenapine/asenapine patients.\n\nNo deaths occurred during the trial; SAEs were reported in 22 (6.9%) patients. Most SAEs belonged to the psychiatric disorders class, including suicidal ideation, aggression, bipolar disorder, depression, and agitation. SAEs resulted in the discontinuation of 13 (4.0%) patients (aggression \\[2 patients\\], suicidal ideation \\[3 patients\\], worsening of attention-deficit/hyperactivity disorder \\[ADHD\\], agitation, anxiety, worsening of bipolar disorder, worsening of depression, dystonia, intentional overdose, and suicidal behavior \\[1 patient each\\]). Few SAEs were considered by the investigator to be possibly or probably related to treatment (depression/suicidal ideation \\[1 patient\\]; swollen tongue \\[1 patient\\]; hypersensitivity to asenapine \\[1 patient\\]; dystonia \\[1 patient\\]).\n\nTEAEs were reported in 83.2% of patients; the most common TEAEs leading to discontinuation were somnolence (1.9%), sedation (1.2%), and fatigue (1.2%). The most commonly reported TEAEs were somnolence (26.8%), weight gain (18.1%), sedation (15.9%), and headache (12.8%); these TEAEs, in addition to dizziness and oral hypoesthesia, occurred more frequently in placebo/asenapine patients than in asenapine/asenapine patients. TEAEs of special interest were also assessed; most of these events also occurred with greater frequency in the placebo/asenapine group versus the asenapine/asenapine group (akathisia 3.8% vs 2.9%; dizziness 11.3% vs 2.5%; oral hypoesthesia/dysgeusia combined 18.8% vs 3.7%; insomnia 2.5% vs 2.5%; EPS \\[narrow risk set\\] 7.5% vs 5.8%; somnolence/hypersomnia/sedation combined 60.0% vs 36.5%; weight gain \u22657% from baseline to endpoint 37.3% vs 34.0%). All TEAEs reported for EPS (narrow) were not serious and considered mild or moderate in severity; akathisia led to one treatment discontinuation.\n\nNo clinically meaningful changes were noted in the prespecified laboratory parameters of interest (Table\u00a0[4](#Tab4){ref-type=\"table\"}). Overall mean weight increase from baseline to endpoint was 3.5\u00a0kg over a mean duration of approximately 27\u00a0weeks of treatment. There were no reported TEAEs of specific hyperglycemia or new-onset diabetes; however, when the broad definition of hyperglycemia/new-onset diabetes was applied (i.e., increased weight, increased appetite, increased blood glucose, central obesity, hyperlipidemia, and loss of consciousness), 73 (22.7%) patients reported TEAEs. There were also two TEAEs of increased blood glucose and six events of increased blood insulin, all of which were considered mild or moderate in intensity. The post-baseline criteria for new-onset metabolic syndrome were met by 18 patients during the extension study; 10 patients who met metabolic syndrome criteria at baseline did not meet the criteria at endpoint and 12 patients met the criteria at both baseline and endpoint.Table\u00a04Long-term bipolar I disorder study: changes in laboratory parametersLaboratory valuesAsenapinePlacebo/asenapine\\\n*n*\u00a0=\u00a080\\\nMean (SD)Asenapine/asenapine\\\n*n*\u00a0=\u00a0241\\\nMean (SD)Asenapine overall\\\n*n*\u00a0=\u00a0321\\\nMean (SD)Cholesterol, fasting, mmol/L2.67 (22.93)\u22123.40 (22.31)\u22121.89 (22.54)Triglycerides, fasting, mmol/L0.62 (50.97)\u22120.97 (58.76)\u22120.53 (56.81)Glucose, fasting, mmol/L0.90 (13.33)1.26 (13.69)1.26 (13.51)Hemoglobin A1c (\\<7.0% to \u22657.0%), *n*000Insulin, fasting, pmol/L4.51 (19.72)\u22126.35 (55.72)\u22123.58 (49.27)Prolactin, \u00b5g/L2.31 (12.30)\u22120.39 (14.07)0.27 (13.69)Adapted from Findling et al. \\[[@CR35]\\]*SD* standard deviation\n\nNo suicidal ideation or behavior based on C-SSRS responses was reported for the 295 (91.9%) patients with data (one patient did not have C-SSRS data). Most suicidal ideation was reported as passive or nonspecific ideation; eight patients reported self-injurious behavior without suicidal intent.\n\nAlthough assessing safety was the primary objective of this extension study, descriptive statistics suggested that improvement in mania was maintained over the course of the extension trial; week 26 was used for efficacy analyses due to limited available week-50 data because of a study protocol amendment. Mean change from open-label baseline to week 26 in YMRS total score was \u22126.9 points in the total treatment group (observed cases); a greater magnitude of change was seen in the placebo/asenapine group (\u221213.0 points) than in the asenapine/asenapine group (\u22124.9 points). Mean change in CGI-BP severity overall score was similar among all treatment groups at week 26 (total treatment group \u22121.2; placebo/asenapine group \u22121.8; asenapine/asenapine group \u22120.9), further suggesting that the effectiveness of asenapine was maintained over the course of treatment.\n\nSchizophrenia {#Sec13}\n-------------\n\n### Acute Study {#Sec14}\n\nThe efficacy and safety of asenapine in adolescents with schizophrenia (12--17 years of age) were evaluated in an 8-week randomized, double-blind, placebo-controlled trial (NCT01190254) \\[[@CR24]\\]. In the acute study, patients were randomized (1:1:1) to placebo (*n*\u00a0=\u00a0102), asenapine 2.5\u00a0mg BID (*n*\u00a0=\u00a098), or asenapine 5\u00a0mg BID (*n*\u00a0=\u00a0106); similar percentages of patients across treatment groups completed the study (placebo 79%; asenapine: 2.5\u00a0mg BID 83%, 5\u00a0mg BID 79%). The most common reasons for study discontinuation in the 2.5-mg BID and 5-mg BID asenapine groups, respectively, were AEs (6.1% and 7.5%), withdrawal of consent (5.1% and 6.6%), and treatment failure (4.1% and 4.7%).\n\nAlthough change from baseline to week 8 in Positive and Negative Syndrome Scale (PANSS) \\[[@CR39]\\] total score was numerically greater for asenapine-treated patients compared with placebo-treated patients, the LSMD (95% CI) was not statistically significant for either asenapine dose versus placebo (2.5\u00a0mg BID \u22124.8 \\[\u22129.9 to 0.4\\], *p*\u00a0*=*\u00a00.070, 5\u00a0mg BID \u22125.6 \\[\u221210.7 to \u22120.5\\], *p*\u00a0*=*\u00a00.064). Some differences in favor of asenapine over placebo were observed on exploratory secondary endpoints including change from baseline in the Clinical Global Impressions--Severity (CGI--S) scale (asenapine 5\u00a0mg BID: LSMD \\[95% CI\\] \u22120.3 \\[\u22120.6 to \u22120.0\\], *p*\u00a0=\u00a00.024) and PANSS response, defined as 30% improvement from baseline (placebo 36%; asenapine: 2.5\u00a0mg BID 50%, *p*\u00a0=\u00a00.028 and 5\u00a0mg BID 49%, *p*\u00a0=\u00a00.048).\n\nNo deaths occurred during the study; the incidence of SAEs was low and similar across treatment groups (placebo 2.9%; asenapine: 2.5\u00a0mg BID 3.1%, 5\u00a0mg BID 2.8%), with most events related to worsening of schizophrenia. The incidence of TEAEs was higher in the asenapine groups (2.5\u00a0mg BID 62.2%, 5\u00a0mg BID 67.0%) than in the placebo group (47.1%). AEs resulted in the discontinuation of six patients in the asenapine 2.5-mg BID group (schizophrenia, agitation, depression, auditory hallucination, somnolence, pneumonia \\[1 patient each\\]) and eight patients in the 5-mg BID group (worsening schizophrenia \\[3 patients\\], insomnia and akathisia, dysgeusia, sedation, polycythemia, abnormal alanine aminotransferase \\[1 patient each\\]). The most common TEAEs overall were nervous system events (placebo 13.7%; asenapine: 2.5\u00a0mg BID 36.7%, 5\u00a0mg BID 48.1%); somnolence and sedation, respectively, occurred in 6.9% and 2.0% of placebo patients, 20.4% and 4.1% of asenapine 2.5-mg BID patients, and 17.0% and 11.3% of asenapine 5-mg BID patients. Oral hypoesthesia occurred more than twice as frequently in asenapine patients (2.5\u00a0mg BID 5.1% and 5\u00a0mg BID 4.7%) than in placebo patients (1.0%). Predefined AEs of interest that occurred significantly more frequently in at least one asenapine group versus the placebo group were akathisia (placebo 1.0%, asenapine 5\u00a0mg BID 6.6%, *p*\u00a0\\<\u00a00.05), dizziness (placebo 1.0%, asenapine 2.5\u00a0mg BID 7.1%, *p*\u00a0\\<\u00a00.05), combined somnolence/sedation/hypersomnia (placebo 8.8%, asenapine 2.5\u00a0mg BID 24.5%, asenapine 5\u00a0mg BID 29.2%, *p*\u00a0\\<\u00a00.05 each), and weight increase \u22657% (placebo 3.1%, asenapine 2.5\u00a0mg BID 9.5%, asenapine 5\u00a0mg BID 10.1%, *p*\u00a0\\<\u00a00.05 each). AEs likely to be associated with EPS occurred in more asenapine-treated patients (2.5\u00a0mg BID 5 \\[5.1%\\], 5\u00a0mg BID 11 \\[10.4%\\]) than placebo-treated patients (4 \\[3.9%\\]), but the difference was not statistically significant (*p*\u00a0\u2265\u00a00.05).\n\nThere were no noteworthy differences between treatment groups in fasting cholesterol, fasting glucose, fasting triglycerides, or HbA1c (Table\u00a0[5](#Tab5){ref-type=\"table\"}). Mean weight increase was 0.1\u00a0kg in the placebo group; in the asenapine 2.5- and 5-mg BID groups, weight increase was 1.3 and 1.4\u00a0kg, respectively. Using the hyperglycemia/new-onset diabetes (broad) criteria, potential new-onset diabetes was reported in four (4%) placebo patients, seven (7%) asenapine 2.5-mg BID patients, and seven (7%) asenapine 5-mg BID patients; new-onset metabolic syndrome was reported by one patient in the asenapine 2.5-mg BID group and two patients in the 5-mg BID group. Mean change in laboratory values showed a potential dose response for asenapine in fasting insulin. Mean changes in prolactin were similar among groups; however, prolactin levels meeting the predefined limit of change for a high value (\u22651.1 times the upper limit of normal) were more common with asenapine (2.5\u00a0mg BID 23%, 5\u00a0mg BID 19%) than placebo (13%).Table\u00a05Acute schizophrenia trial: changes in laboratory parametersLaboratory valuesPlacebo\\\n*n*\u00a0=\u00a0102\\\nMean (SD)Asenapine2.5\u00a0mg BID\\\n*n*\u00a0=\u00a098\\\nMean (SD)5\u00a0mg BID\\\n*n*\u00a0=\u00a0106\\\nMean (SD)Cholesterol, fasting, mmol/L\u22120.31 (0.62)\u22120.05 (0.79)\u22120.09 (0.64)Triglycerides, fasting, mmol/L\u22120.11 (0.87)\u22120.01 (0.69)\u22120.02 (0.76)Glucose, fasting, mmol/L\u22120.14 (0.60)\u22120.01 (0.72)0.12 (0.69)Hemoglobin A1c (\\<7.0% to \u22657.0%), *n*000Insulin, fasting, pmol/L\u22122.86 (102.54)15.67 (5244)34.98 (76.92)Prolactin, \u00b5g/L\u22129.10 (23.62)\u22129.09 (28.92)\u221210.04 (26.64)Adapted from Findling et al. \\[[@CR24]\\]*BID* twice daily, *SD* standard deviation\n\n### Long-Term Safety Study {#Sec15}\n\nLong-term safety and tolerability were assessed in adolescent patients who completed the 8-week acute trial in schizophrenia and continued in a 26-week, open-label, flexible-dose (2.5--5\u00a0mg BID) extension trial (NCT01190267) \\[[@CR24]\\]. Although safety was the primary focus of the long-term study, exploratory efficacy findings were also collected. A total of 196 patients participated in the extension study; 62 of these patients had received placebo in the acute study (placebo/asenapine) and 134 patients had received asenapine (asenapine/asenapine). Discontinuations were recorded for 9 (14.5%) placebo/asenapine patients and 32 (23.9%) asenapine/asenapine patients; AEs were the most common reason for discontinuation in both groups (6.5% and 4.5%, respectively; four events were considered SAEs). The mean duration of treatment in the open-label extension was 171\u00a0days for placebo/asenapine patients and 163\u00a0days for the asenapine/asenapine patients; the mean average daily dose was 9.8\u00a0mg for both groups.\n\nOne death, which resulted from a fall from a sixth-floor window, occurred in a placebo/asenapine patient; it was determined by the investigator and the patient's family to be accidental and not related to the study drug. SAEs, including the death, were reported for seven (4%) patients (2 placebo/asenapine, 5 asenapine/asenapine). Six patients were hospitalized due to worsening psychiatric disorders (1 placebo/asenapine patient \\[aggression and anxiety\\]; 5 asenapine/asenapine patients \\[3 with worsening of schizophrenia, 1 with aggression, 1 with agitation\\]). Consistent with findings from the acute study, somnolence/hypersomnia/sedation combined was the most common TEAE of interest in longer-term treatment (21.4%). Except for the TEAE of weight gain \u22657%, all other TEAEs of interest were reported with greater frequency in placebo/asenapine patients versus asenapine/asenapine patients (akathisia 3.2% vs 1.5%; dizziness 4.8% vs 2.2%; oral hypoesthesia/dysgeusia combined 9.7% vs 0.7%; insomnia 4.8% vs 3.7%; EPS \\[narrow risk set\\] 9.7% vs 2.2%; somnolence/hypersomnia/sedation combined 27.4% vs 18.4%; weight gain \u22657% 12.9% vs 14.9%). Somnolence, the most common TEAE in acute treatment, occurred with greater frequency in placebo/asenapine patients (22.6%) than in asenapine/asenapine patients (11.2%); sedation occurred in 5.1% of patients overall (placebo/asenapine 3.2%; asenapine/asenapine 6.0%). Oral hypoesthesia also occurred with greater frequency in placebo/asenapine patients (8.1%) than in asenapine/asenapine patients (0.7%), suggesting a transitory effect.\n\nMean changes in metabolic parameters were generally modest (Table\u00a0[6](#Tab6){ref-type=\"table\"}). Mean weight increase was 1.6\u00a0kg for patients in the placebo/asenapine group and 1.0\u00a0kg for the asenapine/asenapine group. New-onset hyperglycemia/diabetes (broad) occurred in eight (4%) patients, five of whom also met the weight gain \u22657% criterion (asenapine/asenapine group). New-onset metabolic syndrome was also observed in seven patients (2 placebo/asenapine, 5 asenapine/asenapine). Mean changes in prolactin levels during open-label treatment were small; however, prolactin increases \u22651.1 times the upper limit of normal occurred in 32% of placebo/asenapine patients and 31% of asenapine/asenapine patients.Table\u00a06Long-term schizophrenia trial: changes in laboratory parametersLaboratory valuesAsenapinePlacebo/asenapine\\\n*n*\u00a0=\u00a062\\\nMean (SD)Asenapine/asenapine\\\n*n*\u00a0=\u00a0134\\\nMean (SD)Asenapine overall\\\n*n*\u00a0=\u00a0196\\\nMean (SD)Cholesterol, fasting, mmol/L0.06 (0.62)\u22120.05 (0.58)\u22120.02 (0.59)Triglycerides, fasting, mmol/L0.05 (0.40)\u22120.02 (0.58)0.01 (0.52)Glucose, fasting, mmol/L0.19 (0.75)\u22120.02 (0.83)0.05 (0.81)Hemoglobin A1c (\\<7.0 to \u22657.0%), *n* (%)0 (0)1 (0.8)1 (0.5)Insulin, fasting, pmol/L16.93 (81.27)1.28 (99.09)6.43 (93.64)Prolactin, \u00b5g/L5.27 (19.23)2.92 (23.7)3.69 (22.33)Adapted from Findling et al. \\[[@CR24]\\]*SD* standard deviation\n\nC-SSRS--assessed suicidal ideation was recorded in eight (4%) patients during open-label treatment; none of the events was associated with intent or a formal plan, nor were suicidal behaviors recorded. TEAEs of suicidal ideation were reported for three (1.5%) of these events, and parental consent was withdrawn for six subjects with suicidal ideation.\n\nNumerical improvements in PANSS total and CGI--S scores that were observed in the acute phase of the study were maintained during open-label treatment, suggesting that some treatment effect persisted over the course of the long-term study. From the acute baseline to the extension endpoint, mean change in PANSS total score was \u221236.7 points for asenapine/asenapine patients and \u221234.4 points for placebo/asenapine patients. Of the 103 patients who were PANSS responders at the end of double-blind treatment and continued into extension treatment, 79% of patients (placebo/asenapine 76% and asenapine/asenapine 80%) maintained response.\n\nAsenapine Subgroup and Post Hoc Analyses {#Sec16}\n----------------------------------------\n\nAlthough results of subgroup and post hoc analyses should be interpreted with caution due to inherent limitations associated with this type of analysis, exploratory information about a drug can be informative for clinicians. For example, in subgroup analyses from the acute asenapine bipolar study in pediatric patients \\[[@CR33]\\], the efficacy of asenapine versus placebo (YMRS total score change from baseline to day 21) was not significantly different between patients with and without ADHD, which was the most common comorbid Axis I disorder in the study (54.6% of patients), and between patients with and without concomitant stimulant use. Additionally, efficacy did not differ significantly with respect to patient gender or for patients with bipolar disorder onset before or after the age of 11\u00a0years \\[[@CR33]\\]. No noticeable between-group differences in TEAEs were reported.\n\nIn post hoc analyses conducted on data from the acute bipolar study, patient body weight and BMI did not strongly influence the efficacy of asenapine; as such, adjusting for body weight or BMI is unlikely to be an effective dosing strategy in adolescent patients \\[[@CR40]\\]. Additional analyses showed that although asenapine-treated patients had significantly greater increases in body weight and BMI compared with placebo-treated patients, weight increases were only slightly greater in asenapine-treated patients with baseline BMI percentiles above the median (1.4--2.2\u00a0kg) than in patients with BMI percentiles below the median (1.4--1.7\u00a0kg) across the dose range. This suggests that weight gain in pediatric patients treated with asenapine was not substantively correlated to baseline BMI \\[[@CR41]\\]. Asenapine has also been found to be generally effective in pediatric patients with bipolar disorder regardless of the number of previous manic episodes; small sample sizes in these analyses precluded definitive conclusions about differences by dose \\[[@CR42]\\]. Finally, asenapine appeared to be generally effective in subsets of patients with manic episodes as well as in patients with mixed episodes as shown by significantly greater improvements in YMRS total score for asenapine-treated patients compared with placebo-treated patients \\[[@CR43]\\].\n\nDosing and Administration {#Sec17}\n=========================\n\nFor bipolar mania monotherapy in pediatric patients aged 10--17\u00a0years, the recommended dose is 2.5--10\u00a0mg BID, which may be adjusted based on individual response and tolerability \\[[@CR11]\\]. As discussed previously, since initial sensitivity to asenapine and the occurrence of dystonia in some pediatric patients has been noted, a short up-titration period is recommended when treatment is initiated. After starting treatment with the recommended 2.5\u00a0mg BID dose, the dose may be increased to 5\u00a0mg BID after 3\u00a0days and to 10\u00a0mg BID after an additional 3 days. Drinking and eating should be avoided for 10\u00a0min after asenapine is administered since this can remove drug from the oral cavity and prevent further mucosal absorption \\[[@CR11]\\]. The safety of doses \\>10\u00a0mg BID has not been evaluated in clinical trials; safety and efficacy have not been evaluated in patients with bipolar disorder under the age of 10\u00a0years or in patients with schizophrenia under the age of 12\u00a0years.\n\nUsing Asenapine in Pediatric Patients {#Sec18}\n=====================================\n\nAsenapine monotherapy is FDA approved for the treatment of bipolar mania in adolescent patients aged 10--17\u00a0years; the recommended dosage is 2.5--10\u00a0mg BID (Table\u00a0[7](#Tab7){ref-type=\"table\"}). Asenapine has also been studied in pediatric patients with schizophrenia, but it is not approved for this indication. Asenapine, with its unique sublingual formulation, is absorbed through the oral mucosa when a tablet placed under the tongue completely dissolves, which occurs within seconds. In pediatric patients, sublingual administration may be considered particularly advantageous since difficulty swallowing a pill may make conventional oral tablets less acceptable \\[[@CR44]\\]. Complete instructions for sublingual administration are available in the prescribing information (Table\u00a0[7](#Tab7){ref-type=\"table\"}) \\[[@CR11]\\]. Patients should be informed that numbness or tingling of the mouth or throat may occur directly after asenapine administration and that these sensations usually resolve within 1\u00a0h.Table\u00a07Asenapine in pediatric patients: key pediatric dosing and administration considerationsAsenapine is FDA approved for the treatment of pediatric patients (10--17 years of age) with bipolar maniaThe recommended dosage is 2.5--10\u00a0mg sublingually twice daily; the dose may be adjusted for individual response and tolerabilityAn up-titration schedule is recommended in pediatric patients (starting dose of 2.5\u00a0mg BID may be followed by an increase to 5\u00a0mg BID after 3\u00a0days and to 10\u00a0mg BID after 3\u00a0additional days)The asenapine tablet should be placed under the tongue and allowed to dissolve completely; the tablet should not be split, crushed, chewed, or swallowedTingling, numbness in the mouth or throat, or distortions in taste may occur directly after asenapine administration and usually resolve within an hourFood and drink should not be consumed for 10\u00a0min after administrationWorsening psychiatric condition and signs of suicidality should be closely monitored in pediatric patients taking antipsychotic medicationsWeight gain and changes in metabolic parameters should be monitored*BID* twice daily, *FDA* US Food and Drug Administration\n\nAlthough swallowing an asenapine tablet is difficult to do because it immediately dissolves under the tongue, absolute bioavailability is markedly decreased if asenapine is swallowed, so it should not be ingested \\[[@CR25]\\]. Asenapine exposure can also be reduced by food or drink, so eating and drinking should be avoided for at least 10\u00a0min after administration. Additionally, consumption of a high-fat meal immediately prior to sublingual administration of asenapine has been shown to reduce drug exposure by 20%. Although there is no need to adjust the recommended adult dosage when asenapine is used in pediatric patients, young patients may be more sensitive to dystonia when asenapine is initiated, so a short up-titration is recommended in this population.\n\nAsenapine was generally safe and well tolerated in clinical trials of young patients with bipolar I disorder and schizophrenia. Serious adverse effects were generally related to worsening of the underlying psychiatric disorder, including bipolar disorder, schizophrenia, suicidal ideation, aggression, and agitation. Suicidal ideation was noted in acute and long-term asenapine studies, although most events occurred in patients with a history of suicidal ideation and no deaths from suicide were reported in any asenapine study. Given this vulnerable patient population, the high risk for relapse, and the high suicide attempt rate in young patients with bipolar disorder and schizophrenia \\[[@CR2], [@CR45]\\], close monitoring for worsening of the psychiatric condition and prompt response to signs of suicidality is imperative.\n\nSomnolence and sedation are commonly reported with asenapine. Although antipsychotic-related sedating effects may diminish over time in many patients, sedation can become a persistent issue that affects quality of life for a substantial proportion of patients \\[[@CR46]\\]. Since persistent somnolence and sedation may impair academic and social functioning, clinicians should be aware that these effects may be treatment-limiting in some young patients. Oral effects (hypoesthesia and dysgeusia), which are unique to asenapine and related to sublingual administration, do not typically result in treatment discontinuation and appear to be transient. Additionally, the risk of antipsychotic-related EPS in young patients requires special attention from clinicians when any antipsychotic agent is prescribed. As expected, the incidence of EPS was greater in asenapine- than in placebo-treated patients in pediatric clinical trials, although EPS-related effects were generally considered mild to moderate and resulted in few treatment discontinuations.\n\nFinally, elevations in weight and BMI are commonly seen in young patients treated with a second-generation antipsychotic, although metabolic effects may vary \\[[@CR47]\\]. Weight gain in asenapine-treated patients was accompanied by some metabolic changes, indicating that baseline values for lipid levels, BMI, and glucose levels should be established and regularly monitored during treatment. Counseling in healthy eating and lifestyle should accompany metabolic monitoring to encourage patient involvement in their treatment and overall health.\n\nConclusion {#Sec19}\n==========\n\nAsenapine is a unique sublingual second-generation antipsychotic that is approved in the US for use in patients with bipolar mania who are 10--17\u00a0years of age. In the acute bipolar I disorder trial, improvement in mania was seen for three BID doses of asenapine (2.5, 5, 10\u00a0mg) versus placebo. Although results of the acute schizophrenia trial were not robust enough to establish efficacy for asenapine in schizophrenia, valuable information was added to the drug's safety profile in pediatric patients. Asenapine was generally well tolerated in acute and longer-term treatment of bipolar I disorder and schizophrenia; safety findings were generally consistent across indications, with potentially clinically significant changes in some metabolic parameters and a high incidence of sedation noted. As with any antipsychotic administered to pediatric patients, the use of asenapine should be closely monitored for adequate treatment response and antipsychotic-induced adverse effects.\n\nWriting assistance and editorial support for this manuscript were provided by Carol Brown, MS, at Prescott Medical Communications Group (Chicago, IL, USA), with support from Allergan (Jersey City, NJ, USA) in accordance with Good Publication Practice (GPP3) guidelines (<http://www.ismpp.org/gpp3>). All authors had full control of content throughout development of the manuscript and participated in the writing, editing, and approval of this final version.\n\nSource of funding and material support {#FPar2}\n======================================\n\nSupported by funding from Allergan.\n\nConflict of interest {#FPar3}\n====================\n\nDr. Stepanova receives research support from PsychNostics, LLC. Dr. Grant has no conflicts of interest to disclose. Dr. Findling receives or has received research support, acted as a consultant and/or served on a speaker's bureau for Actavis, Akili, Alcobra, American Academy of Child & Adolescent Psychiatry, American Psychiatric Press, Bracket, CogCubed, Cognition Group, Coronado Biosciences, Elsevier, Epharma Solutions, Forest, Genentech, GlaxoSmithKline, Guilford Press, Ironshore, Johns Hopkins University Press, KemPharm, Lundbeck, Medgenics, Merck, NIH, Neurim, Novartis, Otsuka, PCORI, Pfizer, Physicians Postgraduate Press, Purdue, Rhodes Pharmaceuticals, Roche, Sage, Shire, Sunovion, Supernus Pharmaceuticals, Syneurx, Takeda, Teva, TouchPoint, Tris, Validus, and WebMD.\n"}
+{"text": "Background {#Sec1}\n==========\n\nMungbean (*Vigna radiata* \\[L.\\] R. Wilczek) is an important legume crop with high nutritional value in South and Southeast Asia. Because of its high content of easily digestible protein and relatively high iron and folate contents, it represents a nutrition-balanced food for cereal-based diets \\[[@CR1]--[@CR3]\\]. Mungbean is also consumed as sprouts, which are important sources of vitamins and minerals \\[[@CR4], [@CR5]\\].\n\nBruchids (*Callosobruchus* spp.), the bean weevils, cause serious damage to and loss of legume seeds, including mungbean, during storage \\[[@CR6]\\]. Infestation of the crop is generally low in the field. Only a few insect-infested seeds are needed for the initial inoculum for population build-up during grain storage \\[[@CR7]\\]. Bruchid development from eggs to pupae takes place in a single seed, the larva being the most destructive stage. The emerging adults deposit eggs on the seed, causing rapid multiplication of the pest during storage and resulting in up to 100\u00a0% of grain loss.\n\nOnly a few bruchid-resistant mungbean varieties are available today \\[[@CR8]\\] and resistant lines adapted to the tropics are lacking. Chemical controls with organophosphate compounds, synthetic pyrethroids or insect growth regulators widely used to protect mungbean against this pest \\[[@CR9]\\] are expensive, with risks to consumer health and the environment and development of insecticide tolerance by the pest \\[[@CR10], [@CR11]\\]. Biological control by the bruchid parasitoid *Dinarmus* sp. is less efficient than chemicals in reducing the storage pest effects \\[[@CR12]\\]. Therefore, host resistance would be the most sustainable and economical way to preserve mungbean seeds against destruction by bruchids during storage.\n\nThe wild mungbean accession *V. radiata* var. *sublobata* TC1966 from Madagascar is resistant to many bean weevil species, including *Callosobruchus chinensis*, *C. phaseoli*, *C. maculatus*, and *Zabrotes subfasciatus* \\[[@CR13], [@CR14]\\]. TC1966 is easily crossed with *V. radiata*, and the bruchid resistance of this accession was introduced into the cultivated gene pool \\[[@CR14]--[@CR16]\\]. The bruchid resistance of TC1966 was proposed to depend on a single dominant gene plus one or a few modifier factors \\[[@CR16]--[@CR19]\\]. A bruchid-resistant gene (*Br*) for this line has not yet been identified, although several candidate genes have been suggested and genetic markers co-segregating with the *Br* gene have been described. On restriction fragment length polymorphism (RFLP) analysis of 58\u00a0F2 progenies from a cross of TC1966 and a susceptible mungbean, VC3890, the *Br* was mapped to a single locus on linkage group VIII, approximately 3.6 centimorgans (cM) from the nearest RFLP marker \\[[@CR19]\\]. Later, 10 randomly amplified polymorphic DNA (RAPD) markers were found associated with the *Br* gene in a segregating population derived from TC1966 and NM92, a bruchid-susceptible mungbean \\[[@CR15]\\]. RNA-directed DNA polymerase, gypsy/Ty-3 retroelement and chloroplast NADH dehydrogenase subunit genes were highly associated with the proposed *Br* gene of mungbean \\[[@CR15]\\]. Further quantitative trait loci (QTL) analysis revealed one major and two minor QTL for bruchid resistance in TC1966 \\[[@CR17]\\]. Seed metabolite analysis in line BC20F4 derived from a cross between TC1966 and a susceptible cultivar Osaka-ryokuto suggested the involvement of cyclopeptide alkaloids named vignatic acids with bruchid resistance. The gene responsible for vignatic acid (*Va*) accumulation was mapped to a single locus, 0.2\u00a0cM away from the previously mapped *Br* gene \\[[@CR20]\\]. Additionally, a small cysteine-rich protein, VrCPR protein, which is lethal to *C. chinensis* larvae, was identified in TC1966 \\[[@CR21]\\]. Proteomic research has proposed that chitinase, beta-1,3-glucanase, peroxidase, provicilin and canavalin precursors play a role in bruchid resistance of mungbean \\[[@CR22]\\].\n\nThe implication of proteinase and amylase inhibitor activity in bruchid resistance in legumes remains controversial \\[[@CR23]--[@CR26]\\]. Whether these candidate factors indeed associated with previously described bruchid-resistant QTL \\[[@CR17]\\] and contributed to resistance remained unknown. Some of the putative *Br* factors of TC1966 may be harmful for human consumption \\[[@CR27]\\]. Because the chemical nature of the resistance factor is still unknown, the safety of using the resistance factors derived from TC1966 is difficult to assess. Despite much effort directed toward the identification of bruchid-resistant factors, physiological differences between bruchid-resistant and -susceptible mungbean have not been reported.\n\nA molecular marker associated with *Br* would facilitate breeding of bruchid-resistant varieties, and mapping of the resistance genes also would help identify factors underlying resistance. Available markers have not been validated for breeding, and more information on *Br* is required to generate reliable markers for breeding bruchid-resistant mungbean varieties. Gene-based or regulatory sequence-based markers would be the most efficient for selecting bruchid-resistant lines in breeding programs. In contrast to resistance locus-linked RFLP and RAPD markers, resistance-gene or regulatory sequence-based markers cannot be separated from the resistant phenotypes by recombination and thus are more reliable for selection. Bruchid resistance is assumed to be due to the expression of resistance factors. Resistance factors could be direct products of resistance genes that are absent in susceptible lines, or could result from activity changes of factors in susceptible and resistant lines due to sequence variation or from expression differences of resistance genes. Polymorphisms related to any of these differences would provide reliable markers for resistance.\n\nRecently, the whole-genome sequence of a bruchid-susceptible mungbean (*V. radiata* var. *radiata* VC1973A) was published \\[[@CR28]\\]. Here we report the whole-genome sequence of a bruchid-resistant recombinant inbred line (RIL) and an increased number of available gene annotations for mungbean, by 14,500 genes. We have identified differentially expressed genes (DEGs) and nucleotide variations (NVs) in the promotor regions of DEGs and in the exons of sequence-changed protein genes (SCPs). The putative effects of DEGs and SCPs on bruchid resistance of mungbean are discussed and molecular markers derived from NVs that can be used for selection of resistant lines are reported.\n\nResults {#Sec2}\n=======\n\nGenome size of different mungbean cultivars and wild relatives {#Sec3}\n--------------------------------------------------------------\n\nThe genome size estimated by cytometry ranged from about 494 to 555\u00a0Mb (mega base pairs) (Table\u00a0[1](#Tab1){ref-type=\"table\"}) in the lines under investigation. We found about a 20-Mb difference in genome size between wild mungbean TC1966 (494\u00a0Mb) and the cultivar NM92 (517\u00a0Mb). The genome size of RIL59, offspring of a cross between TC1966 and NM92, was similar to that of its female parent NM92, whereas k-mer frequency distribution analysis of RIL59 suggested a genome size of 452\u00a0Mb. The estimated genome size of the buchid-susceptible mungbean line VC1973A, recently sequenced \\[[@CR28]\\], was about 502\u00a0Mb, similar to the size of the bruchid-resistant mungbean line V2802; another bruchid-resistant mungbean line, V2709, had the largest genome size in our study.Table 1Genome size of mungbean varietiesTC1966NM92RIL59V1973AV2709V2802pg/2C^a^1.01\u2009\u00b1\u20090.021.06\u2009\u00b1\u20090.021.06\u2009\u00b1\u20090.001.03\u2009\u00b1\u20090.021.13\u2009\u00b1\u20090.011.04\u2009\u00b1\u20090.01Genome size (Mb)493.6\u2009\u00b1\u20093.3517.3\u2009\u00b1\u20093.1517.1\u2009\u00b1\u20090.9502.2\u2009\u00b1\u20092.9554.7\u2009\u00b1\u20092.8506.1\u2009\u00b1\u20092.3Data are mean\u2009\u00b1\u2009SE from six biological repeats^a^DNA content of diploid organisms (2C) represented in picograms (pg); 1\u00a0pg\u2009=\u2009978\u00a0Mb \\[[@CR33]\\]\n\n*De novo* genome assembly of RIL59 {#Sec4}\n----------------------------------\n\nThe previously published whole genome sequence for mungbean is derived from the bruchid-susceptible cultivar VC1973A \\[[@CR28]\\]. For genomic comparison and to facilitate research on bruchid resistance of mungbean, we sequenced and assembled the draft genome of the bruchid-resistant line RIL59, whose *Br* gene was inherited from the wild mungbean accession TC1966. Sequencing of four DNA libraries, including two paired-end and two mate-pair libraries with various fragment lengths (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1), resulted in 90.1 Gb (Giga base pairs) of sequence information, which corresponds approximately to a 174.2-fold sequencing coverage according to the genome size of RIL59 (Table\u00a0[1](#Tab1){ref-type=\"table\"}). *De novo* assembly of the sequence reads resulted in 2509 scaffolds with an N50 of 676.7\u00a0kb (kilo base pairs) comprising 455.2\u00a0Mb (Table\u00a0[2](#Tab2){ref-type=\"table\"}) and contributed to approximately 88\u00a0% of the estimated genome size of RIL59 (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The largest scaffold had a length of 4.4\u00a0Mb.Table 2Summary of *de novo* genome assembly of RIL59StageN50 (kb)Average Length (kb)Total Length (Mb)Longest (kb)No. sequencesContigs34.916.9437.9287.025,895Scaffolds676.7181.4455.24419.02509\n\nGene annotation {#Sec5}\n---------------\n\nIn total, 40.5\u00a0% of the draft genome was classified as repeat sequences and 23.3\u00a0% as long tandem repeat (LTR) elements. The repeat elements were annotated by using the TIGR plant repeat database (Table\u00a0[3](#Tab3){ref-type=\"table\"}). Sixteen paired-end RNA libraries (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1) constructed from RIL59 tissues and different RIL seeds represented 134.4 Gb. RNA-seq data for RIL59 (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1) and NCBI (<http://www.ncbi.nlm.nih.gov/>) soybean refseq protein sequences were aligned to the repeat-masked genome to identify splice junctions for gene prediction. Overall, 63.35\u00a0% of the RNA-seq data mapped uniquely to splice junctions. *Ab initio* gene prediction combined with protein alignment resulted in annotations for 36,939 protein-coding genes; 4493 of these encoded for multiple isoforms, for 42,223 transcripts in total. Overall, 85\u00a0% of the 49,952 predicted-gene models had matches in the NCBI non-redundant protein database. The predicted-gene models consisted of transcript lengths of 4108\u00a0bp, coding lengths of 1290\u00a0bp, and 5.76 exons per gene, on average.Table 3Repeated sequences annotation of repeat elements from the TIGR databaseClassNumberSize (bp)Retrotransposon54481,182Transposon14521,405Miniature Inverted-repeat Transposable Elements (MITE)2230Centromere satellite91161Unclassified centromere sequence71621Telomere sequence81358Telomere associated9918rDNA 45S2911,783rDNA 5S426914Unclassified (total)24324,467\n\nIdentification of bruchid-resistance--associated genes by transcriptome comparison {#Sec6}\n----------------------------------------------------------------------------------\n\nWe searched for bruchid-resistance--associated genes by comparing the seed transcriptome of bruchid-resistant (R) and -susceptible (S) mungbean lines (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1), including two parental lines of a population of NM92 (S) and TC1966 (R), and RILs derived from this population: RIL59 (R) and three pairs of RILs, each pair with contrasting bruchid resistance. Two methods to identify DEGs by RNA-seq were applied. The first approach, which involved calculating the number of transcripts per million (TPM), revealed 22 up- and 6 downregulated genes in seeds of bruchid-resistant mungbean (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} and Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). Three of the upregulated genes (*g4706*, *g34480* and *g42613*) were specifically detected in R mungbean, and two downregulated genes (*g40048*, *g41876*) were specifically detected in S mungbean.Fig. 1Transcriptome analysis of bruchid-resistant--associated genes in mungbean. Bruchid-resistant--associated genes were selected from transcripts per million (TPM) fold change comparison and DESeq analysis of transcriptomes between brucnid-resistant and -susceptible mungbean. The number of DEGs selected by each criterion is indicated. Up and down represent the genes up- and downregulated, respectively, in bruchid-resistant mungbean\n\nThe second approach by DESeq analysis \\[[@CR29]\\] of the same nine transcriptomes identified 81 transcripts of 80 DEGs; 31 were up- and 49 downregulated in bruchid-resistant mungbean (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} and Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). The downregulated gene *g16371* was present in two splice forms, *g16371.t1* and *g16371.t2*. Ten genes (*g24427*, *g34321*, *g4706*, *g34480*, *g28730*, *g17228*, *g9844*, *g39181*, *g39425*, *g42613*) were expressed only in bruchid-resistant lines and three (*g40048*, *g35775*, *g2158*) only in bruchid-susceptible lines. Together, the two approaches identified 91 DEGs most likely related to bruchid resistance. We classified the 17 consensus genes pinpointed by both approaches as major bruchid-resistance--associated genes and the other 74 as minor bruchid-resistance--associated genes (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}).\n\nThe 17 consensus genes are most likely highly related to bruchid resistance of mungbean, especially the 12 upregulated genes (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} and Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). However, five of these genes have unknown function, including three with no hits on Blastx analysis. The putative UBN2_2 domain of g34480 and RVT_2 domain of g4739 implying their transposase activity, together with the putative gag/pol polyprotein, g34458, represented transposable elements (TEs). The remaining genes encoded a putative MCM2-related protein, a putative adenylate cyclase, a senescence regulator and a resistant-specific protein (Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2).\n\nRT-qPCR analysis of the RNA-seq data verified the 17 consensus genes (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). Ten upregulated genes were in all R mungbean lines as compared with S lines, except *g9801* and *g17262* were undetected in the bruchid-resistant RIL153. Among the five downregulated genes, *g40048*, *g28764* and *g759* were consistently downregulated in all R lines as compared with S lines.Fig. 2RT-qPCR validation of differentially expressed genes (DEGs). RT-qPCR results of the pattern of gene expression between bruchid-resistant and -susceptible mungbean. The Y axis indicates the relative quantity (RQ) of gene expression with mungbean *VrActin* (*g12676*) used as a control. Data are RQ\u2009\u00b1\u2009SE of \u0394\u0394CT from three experimental repeats. The X axis indicates different bruchid-resistant (R) and -susceptible (S) mungbean lines. Asterisk indicates that the expression of the gene was not detected in the parental line NM92 with CT value set to 40\u00a0cycles for calculating the RQ of gene expression\n\nThe high consistency between RNA-seq and RT-qPCR results implied the DEGs might represent the biological difference between R and S mungbean seeds. In terms of functional categorization based on gene annotation combined with predicted protein domains, 36 of the 91 DEGs encoded proteins with enzymatic activities, four encoded resistant-related proteins and eight encoded TEs (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). Among the DEGs, 18 were involved in metabolic pathways, genetic information processing, environmental information processing and cellular processes (Additional file [3](#MOESM3){ref-type=\"media\"}: Table S3).Fig. 3Pie chart representing the functional categories of DEGs and sequence-changed-protein genes (SCPs). DEGs (**a**) and SCPs (**b**) were functionally classified into categories based on annotation and the putative protein domains they harbored. The number of genes in each category is indicated in parentheses\n\nTwo of the DEGs, *g728* and *g17654*, encoded a cysteinyl endopeptidase and a basic 7S globulin 2 precursor, respectively. The former has protease activity and the latter was implicated in bruchid resistance \\[[@CR22]\\]. Both proteins are predicted to contain an inhibitor domain (Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). However, we found their expression downregulated in bruchid-resistant mungbean (Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2), which suggests that these proteins have no role in resistance.\n\nNVs in promoter regions might affect the expression of genes. A survey of NVs including substitutions and insertions and deletions (indels) by comparing genomic sequences of bruchid-resistant and -susceptible lines revealed that 408 NVs located in the 2-kb region presumably included the promoter regions of 68 consensus DEGs (Additional file [4](#MOESM4){ref-type=\"media\"}: Table S4). The number of NV sites in the 2-kb regions ranged from 1 to 24 (Additional file [4](#MOESM4){ref-type=\"media\"}: Table S4).\n\nIdentification of bruchid-resistance--associated SCPs {#Sec7}\n-----------------------------------------------------\n\nIn addition to DEGs, NVs including nonsynonymous substitutions and indels in exon regions producing SCPs can modify protein functions, without necessarily changing gene expression. Because genetic codes stored in RNA are directly transmitted to proteins, we compared NVs of genes based on RNA-seq data between bruchid-resistant and -susceptible lines and found 282 consensus NVs on 149 transcripts (148 genes) (Additional file [5](#MOESM5){ref-type=\"media\"}: Table S5). The confidence of NVs was verified by genomic sequence comparison of a few genes between RIL59 and its parents. For illustration, seven NVs were proposed on *g662* cDNA by RNA-seq comparison (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). Genomic sequence results confirmed that these NVs consistently exist in R mungbean, lines RIL59 and TC1966, and S mungbean, NM92 (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}).Fig. 4Validation of nucleotide variations (NVs) identified by RNA sequence comparison. Gene *g662* was used to illustrate the verification of NVs. The upper panel shows the cDNA sequence of *g662* and the seven NVs (mark in red) identified by RNA sequence comparison of bruchid-resistant (R) and -susceptible (S) mungbean. The NVs in parentheses show the nucleotides in R mungbean (the former letter) changed to that in S mungbean (the latter letter). The lower panel shows the validation of NVs by genomic sequencing between R mungbean lines RIL59 andTC1966 and S line NM92. The color of the letter is synchronized with that of the chromatogram for easy reading. The box indicates the site of NVs. The order of NV sites starts from down-left then down-right panels\n\nOf the 148 SCPs, 134 could be functionally annotated by Blast analysis. Most encoded proteins harbored enzymatic activities, and 15 encoded transcription factors. Importantly, seven and four genes encoded resistant-related proteins and TEs, respectively (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"} and Additional file [6](#MOESM6){ref-type=\"media\"}: Table S6). Similar to DEGs, 28 of the 148 SCPs were involved in pathways of metabolism, genetic information processing, environmental information processing, cellular processes and organismal systems (Additional file [7](#MOESM7){ref-type=\"media\"}: Table S7). DEGs and SCPs involved in conserved pathways implied the conserved intrinsic difference between R and S mungbean (Additional file [3](#MOESM3){ref-type=\"media\"}: Table S3 and Additional file [7](#MOESM7){ref-type=\"media\"}: Table S7).\n\nTwo of the SCPs, *g29024* and *g4649*, encoded putative pectinesterase inhibitor 3-like and Kunitz trypsin inhibitor protein, respectively. Whether they are involved in bruchid resistance needs further investigation.\n\nMapping of bruchid-resistance--associated NVs in the mungbean genome {#Sec8}\n--------------------------------------------------------------------\n\nBruchid-resistance--associated DEGs and SCPs are potential *Br* genes. Hence, the NVs in the promoter region are potential regulatory-sequence--based markers, whereas NVs on SCPs are potential gene-based markers for resistance. We mapped the identified bruchid-resistance--associated NVs and genes to the mungbean genome of VC1973A \\[[@CR28]\\] to assess whether their genomic position co-localizes with previously reported bruchid-resistance--associated markers. The 2-kb promoter region considered to have putative regulatory sequences implicated in resistance for the 68 DEGs was mapped to pseudochromosomes of mungbean \\[[@CR28]\\]. The promoters of these DEGs were found unevenly distributed over the 11 chromosomes, and most sequences were mapped to chromosome 5 and to 10 scaffolds (Table\u00a0[4](#Tab4){ref-type=\"table\"} and Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). Similarly, 282 NVs of 148 SCP genes were unevenly distributed over the 11 chromosomes and 16 scaffolds of the mungbean reference sequence \\[[@CR28]\\]. Interestingly, most of these sequences were mapped to chromosome 5 (Table\u00a0[4](#Tab4){ref-type=\"table\"} and Additional file [6](#MOESM6){ref-type=\"media\"}: Table S6). Therefore, 690 bruchid-resistance--associated NVs were mapped to 11 chromosomes and 21 scaffolds of the reference sequence (Table\u00a0[4](#Tab4){ref-type=\"table\"}, Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2 and Additional file [6](#MOESM6){ref-type=\"media\"}: Table S6).Table 4Mapping of bruchid-resistance--associated genes on mungbean pseudochromosomeVr1Vr2Vr3Vr4Vr5Vr6Vr7Vr8Vr9Vr10Vr11ScaffoldsTotalDEGs4434163633221767SCPs737105565171244148Total1171013679114103360208The promoter 2-kb sequences of differentially expressed genes (DEGs) and sequences of sequence-changed-protein genes (SCPs) were mapped on the 11 pseudochromosomes (Vr1\u2009\\~\u2009Vr11) and scaffolds of mungbean \\[[@CR28]\\]. The total number of genes mapped to Vr4, Vr5, Vr11 and scaffolds were not equal to the sum of DEGs and SCPs because some SCPs also belonged to DEGs\n\nThe two published bruchid-resistance--associated markers, the cleaved amplified polymorphic DNA (CAP) marker OPW02a4 and the simple sequence repeat (SSR) marker DMB-SSR158 \\[[@CR15], [@CR17]\\], were mapped to scaffolds 298 and 227 of RIL59, respectively, and both mapped to chromosome 5 of the mungbean reference (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). In the present study, 67 bruchid-resistance--associated genes, including DEGs and SCPs, were mapped to chromosome 5 of mungbean. The mapping results revealed a striking difference in promoter region of *g39185* between RIL59 and VC1973A. A similar phenomenon was observed with the promoters of *g34480* and the gene body of *g28730* (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}). These results imply that the genome structure at these positions differs between RIL59 and VC1973A, which might be related to the difference in resistance against bruchids.Fig. 5Map of bruchid-resistant--associated genes on chromosome 5 (Vr5) of VC1973A. The corresponding scaffold for each gene in RIL59 is at both sides. The two bruchid-resistant markers are in red. The DEGs are indicated in blue and SCPs in black. The DEGs with an asterisk are also SCPs. For DEGs, the 2-kb promoter sequences were used for mapping, whereas for SCPs, the gene sequences were usedFig. 6Close-up map of *g39185* and *g34480* promoter sequence (**a**) and *g28730* gene (**b**) on mungbean Vr5. The 2-kb promoter sequences of *g39185* (*g39185_p*) and *g34480* (*g34480_p*) and *g27830* gene of RIL59 are strikingly different from that of VC1973A. The number on Vr5 of V1973A indicates the position on the chromosome. mb, million base\n\nGeneration of bruchid-resistance--associated markers {#Sec9}\n----------------------------------------------------\n\nFrom the bruchid-resistance--associated NVs, we selected long sequence indels and designed primers (Additional file [8](#MOESM8){ref-type=\"media\"}: Table S8) for PCR-based molecular markers. Three markers derived from NVs on promoters of DEGs could distinguish R and S mungbean well between RIL59, two parents and three sets of RILs (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}). Marker *g779p* produced a smaller band in R than S mungbean. Marker *g34480p* produced a band only in R mungbean, as expected, but a smaller size in RILs than TC1966. Marker *g34458p* produced a small band in R mungbean and a large band in S mungbean. Further applying these markers together with the two bruchid-resistance--associated markers, the CAP marker OPW02a4 and SSR marker DMB-SSR158 \\[[@CR15], [@CR17]\\], to 61 RILs revealed DMB-SSR158 with the highest accuracy, 98.3\u00a0%, in selecting mungbean with bruchid resistance. The CAP marker OPW02a4, analyzed by digesting the PCR products with *Hae*III restriction enzyme, exhibited 73.7\u00a0% accuracy. The new developed markers *g779p* and *g34480p* exhibited 93.4\u00a0% accuracy, which was better than the 80.3\u00a0% accuracy of marker *g34458p* (Additional file [9](#MOESM9){ref-type=\"media\"}: Table S9).Fig. 7Bruchid-resistant--associated markers of mungbean. Markers designed from promoter sequences *g779p*, *g34480p*, and *g34458p* were used for selecting bruchid-resistant (R) and -susceptible (S) mungbean. The numbers 1 to 9 indicate different mungbean lines, named TC1966, RIL59, NM92, RIL38, RIL39, RIL54, RIL55, RIL153 and RIL156, respectively. PCR products of *g779p* and *g34458p* were analyzed on 4\u00a0% agarose gel and that of *g34480p* on 1\u00a0% agarose gel\n\nDiscussion {#Sec10}\n==========\n\nGenome size of mungbean {#Sec11}\n-----------------------\n\nThe genome size, DNA quantity, or so-called C-value is important for genome polyploidy, phylogenetic and taxa research \\[[@CR30]--[@CR32]\\]. Recently, a reliable genome size for achieving the correct coverage and estimating the percentage of repeated sequences of a genome has become an important parameter for planning next-generation sequencing (NGS) experiments. Many methods have been used to estimate the genome size of organisms. Besides k-mer frequency distribution analysis together with NGS, flow cytometry has become the most popular method for estimating genome size \\[[@CR33]\\] and is superior to other methods such as DNA phosphate content measurement \\[[@CR34]\\], analysis of reassociation kinetics \\[[@CR35]\\], pulsed-field gel electrophoresis \\[[@CR36]\\] and image analysis of Feulgen photometry \\[[@CR37]\\] because of its convenience, fast processing and reliability \\[[@CR38], [@CR39]\\].\n\nThe same flow cytometry system should be used for comparing plant genome sizes \\[[@CR40]\\] and should avoid the use of an animal genome as a reference \\[[@CR33]\\]. With these recommendations, the genome size estimation for the mungbean lines in our study varied by more than 60\u00a0Mb, from 493.6 to 554.7\u00a0Mb (Table\u00a0[1](#Tab1){ref-type=\"table\"}), whereas previous reports estimated the genome sizes between 470 and 579\u00a0Mb \\[[@CR41], [@CR42]\\]. The large variation in estimations between the studies may be due to variation in mungbean lines and different strategies and methods used for analyses. The genome size of mungbean VC1973A was estimated at 579\u00a0Mb by flow cytometry with nuclei from chicken red blood cells used as an internal standard \\[[@CR28], [@CR41]\\], even though the use of animal genomes as a standard is not recommended for plant genome size prediction. The 25-base k-mer frequency distribution in NGS provided an estimated genome size of 548\u00a0Mb for V1973A \\[[@CR28]\\], which is slightly larger than that by flow cytometry (Table\u00a0[1](#Tab1){ref-type=\"table\"}). Similarly, the estimated genome size for the wild mungbean TC1966 was about 494\u00a0Mb in our flow-cytometry research but 501\u00a0Mb by 25-base k-mer frequency distribution \\[[@CR28]\\]. In contrast, our k-mer frequency distribution provided a size estimate of 452\u00a0Mb for RIL59, smaller than by flow cytometry (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The purity of the constructed DNA libraries for NGS would affect the reliability of k-mer frequency distribution used to estimate genome size \\[[@CR30]\\]. In our study, we used a flow-cytometry method with Arabidopsis nuclei as an internal standard to estimate the mungbean genome size and for a more reliable explanation of the NGS data obtained from RIL59.\n\nVariation in the quantity of repetitive DNA sequences is the main factor determining genome size \\[[@CR43]\\]. This fact was not true for the mungbean lines in our study. RIL59, with a larger genome than that for VC1973A or TC1966, contained only 40.45\u00a0% repetitive sequences as compared with 50.1\u00a0% and 46.9\u00a0% for VC1973A and TC1966, respectively. The difference in proportion of repetitive sequences in TC1966 and its offspring RIL59 suggests that heterozygosity of the genome in a hybrid can lead to loss of repetitive sequences.\n\nGenome assembly and gene annotation {#Sec12}\n-----------------------------------\n\nOur genome assembly of RIL59 is comparable to that of VC1973A, which had 2748 scaffolds with N50 length of 1.52\u00a0Mb and 80\u00a0% genome coverage \\[[@CR28]\\]. We made available a draft genome of a bruchid-resistant variety. Comparing this genome with the available sequence from bruchid-susceptible VC1973A can reveal genomic regions responsible for resistance. The annotated 36,939 genes in RIL59 are 14,512 genes greater than that reported for VC1973A \\[[@CR28]\\]. The larger number of annotated genes could be due to the inclusion of more varied tissues and developmental stages of RIL59 than in the previous study, for a broader capture of different genes. We included RNA from seeds, different developmental stages of pods, 2- to 7-day-old seedlings and 1-month-old whole plants, for broader range of developmental stages and tissues and probably a more complete RNA population. The genome sequence information for the bruchid-resistant RIL59 and a more complete gene annotation of mungbean will contribute to improving \"omic\" research and promoting the breeding of mungbean in all aspects.\n\nDEGs and SCPs together maintain transcript diversity in bruchid-resistant mungbean {#Sec13}\n----------------------------------------------------------------------------------\n\nResearch of bruchid resistance has focused on breeding and developing molecular markers \\[[@CR14]--[@CR16]\\]. Studies of resistance mechanisms at the transcriptomic and proteomic levels were attempted but did not reach final conclusions \\[[@CR21], [@CR22]\\]. We found that DEGs and SCPs might be strategies bruchid-resistant mungbean uses to retain transcript diversity and specificity. Therefore, further proteomic research of mungbean for bruchid resistance should consider both the effects of SCPs and quantity of differential proteins. The DEGs and SCPs related to bruchid resistance might be overestimated in our research for a few reasons. First, based on genetic and QTL studies \\[[@CR16]--[@CR19]\\], a major *Br* locus and two minor loci that might contain one to a few genes were proposed to be responsible for bruchid resistance in mungbean. Second, the RILs we used were not near-isogenic lines. Thus, not all of the identified DEGs and SCPs are likely to be directly associated with bruchid resistance. Third, nonsynonymous substitution due to NVs indeed may not affect the protein function and further biochemical characters in mungbean seeds. Further functional study of the selected DEGs and SCPs will help to evaluate their roles in bruchid resistance of mungbean.\n\nAlthough the transcriptome of mature mungbean seeds most likely reflects what seeds prepared for the upcoming germination and may not be directly related to bruchid resistance of mungbean, the conserved expression pattern of DEGs implies that NVs on their promoters and those on SCPs can be potential molecular markers for mungbean breeding. Use of the promoter-based markers designed from bruchid-resistance--associated genes can indeed offer a high selection rate of bruchid-resistant mungbean.\n\nBoth DEGs and SCPs are involved in similar functional gene categories and conserved metabolic pathways, which implies intrinsic differences between bruchid-resistant and -susceptible mungbean. Whether these intrinsic differences of mungbean represent modifier factors or minor loci modulating bruchid resistance described in previous investigations \\[[@CR16], [@CR17]\\] needs further evaluation. However, almost one-quarter of the DEGs do not harbor NVs on their promoter regions, which suggests alternative mechanisms for differential regulation than sequence variation in these regions involved in regulating the expression of these genes.\n\nThe search for a *Br* gene responsible for bruchid resistance of mungbean {#Sec14}\n-------------------------------------------------------------------------\n\nGreat efforts have been invested in mapping a *Br* gene in mungbean \\[[@CR19]\\]. First, a single *Br* gene was proposed in TC1966, and later additional minor modifier factors were postulated \\[[@CR16], [@CR17], [@CR44]\\]; nevertheless, the nature of the *Br* gene(s) of TC1966 remained unclear. Biochemically, several *Br* candidate genes highly associated with bruchid resistance have been reported, including the *Va* gene for vignatic acid biosynthesis \\[[@CR20]\\]; vicilins, also known as 7S storage globulins \\[[@CR22]\\]; and the cysteine-rich protein VrD1 \\[[@CR21]\\]. Also, resistance-related proteins including chitinase, beta-1,3-glucanase and peroxidase were proposed to play a role in bruchid resistance of mungbean \\[[@CR22]\\]. However, in cowpea and common bean, protease and amylase inhibitors, lectins, chitinases, and beta-1,3-glucanases are ineffective against *C. maculatus* and *Z. subfasciatus* \\[[@CR11]\\]. In our research, only two genes that likely contain inhibitor activities were among the DEGs and a further two among SCPs. The two DEGs were downregulated in bruchid-resistant mungbean, so they might not be involved in resistance, whereas the role of the two detected SCPs encoding inhibitor-like proteins needs further study. None of the other previously proposed *Br* candidate genes was identified as a DEG or SCP, which reduces the probability that they play an important role in bruchid resistance of mungbean. Thus, the identity of a *Br* gene remains elusive.\n\nFrom our research, the 91 DEGs and 148 SCPs are potential *Br* candidates. The mapping results narrowed the number of candidates to 67 on chromosome 5, where the two bruchid-resistant--associated markers, DMB-SSR158 and OPW02a4, are located. Hence, genes located near the markers are potential *Br* genes. However, the genome of mungbean is still incomplete, with many scaffold sequences that could not be assigned to the mungbean genome. Genes mapping to scaffolds cannot be eliminated from the putative *Br* gene list. Different mungbean accessions showed different genome sizes. Therefore, more genome sequencing is necessary to complete the genome assembly of mungbean, especially for TC1966 and other bruchid-resistant sources, for facilitating *Br* gene identification and mungbean breeding and improvement.\n\nTEs and bruchid resistance {#Sec15}\n--------------------------\n\nAlmost half of the genome of mungbean contained repetitive sequences in most TEs, including retrotransposons, transposons and miniature inverted-repeat TEs (MITEs). These elements are ubiquitous in eukaryotic genomes, although the content varies among the different organisms. They can represent 20\u00a0% of the genome, as for the *Drosophila melanogaster* genome \\[[@CR45]\\], or 85\u00a0% for the *Zea mays* genome \\[[@CR46]\\]. TEs are believed to be the major determinant of genome size \\[[@CR47]\\].\n\nRepetitive sequences, previously considered \"junk DNA\", were found to function in modifying the genome structure and gene function and regulating gene expression \\[[@CR48]\\]. The involvement of the non-LTR retrotransposon *CDT-1* in desiccation tolerance of *Craterostigma plantagineum* by mediating small RNA illustrated that TEs regulate plant stress resistance \\[[@CR49]\\]. Could TEs also be involved in bruchid-resistance? We found that genes encoding TEs were DEGs and SCPs when comparing bruchid-resistant and -susceptible lines. In addition, the markers derived from TE sequences well distinguished bruchid-susceptible from --resistant mungbean lines. More studies are required to clarify how TEs are implicated in mungbean resistance or represent modifier factors for this trait.\n\nConclusion {#Sec16}\n==========\n\nHere we provide whole-genome scaffold sequences for a bruchid-resistant mungbean line and increase the annotation of mungbean genes. We obtained a list of putative *Br* genes and candidates of molecular markers for selecting resistant lines and proposed that besides the *Br* gene, intrinsic differences caused by DEGs and SCPs of mungbean and TEs are most likely the modifier factors determining bruchid resistance. As expected, when comparing only a few selected lines with contrasting resistance phenotypes, the identified sequence variations spanned the whole chromosome. However, analysis of all DEG, SCP and NV data revealed factors located on chromosome 5 involved with resistance. More sequence information from different bruchid-resistant sources are needed to facilitate and promote mungbean research and crop improvement.\n\nMethods {#Sec17}\n=======\n\nPlant materials and assessment of bruchid resistance {#Sec18}\n----------------------------------------------------\n\nMungbean (*Vigna radiata* \\[L.\\] R. Wilczek) of the bruchid-susceptible variety NM92, bruchid-resistant accession TC1966 (*V. radiata* var. *sublobata*), their 12-inbred-generation progeny (F~12~) RILs \\[[@CR15]\\], the previously sequenced mungbean line VC1973A \\[[@CR28]\\] and the two BRUCHID-resistant lines *V. radiata* V2802 and V2709 \\[[@CR4]\\] were used as plant materials. All the plant materials used were from the support of AVRDC-The World Vegetable Center. Plants were grown and seeds without disease were harvested in greenhouses at AVRDC \\[[@CR17]\\]. Assay for bruchid resistance with 40 seeds was performed in three replicates as described \\[[@CR15]\\]. Seeds with 0\u00a0% damage were defined as resistant, with more than 80\u00a0% damage defined as susceptibility and damage between 0\u00a0% to 80\u00a0% defined as moderately resistant. DNA extracted from RIL59, TC1966 and NM92 was prepared for genome sequencing.\n\nEstimation of genome size by flow cytometry and k-mer distribution {#Sec19}\n------------------------------------------------------------------\n\nFresh tender leaves of TC1966, NM92, RIL59, VC1973A, V02802 and V02709 were harvested and the nuclear DNA content was estimated as described \\[[@CR33]\\] with minor modifications. Fresh leaf sections (1.0\u00a0cm^2^) were chopped with use of a new razor blade in 1\u00a0mL ice-cold Tris-MgCl~2~ buffer (200\u00a0mM Tris, 4\u00a0mM MgCl-6H~2~O, 0.5\u00a0% triton X-100, pH: 7.5). After filtering through a 20-\u03bcm nylon mesh, the sample was stained with propidium iodide solution (50\u00a0\u03bcg/mL) containing RNase at 50\u00a0\u03bcg/mL. *Arabidopsis thaliana* (Columbia, 0.412\u00a0pg/2C) was used as a reference for estimating the mungbean genome size \\[[@CR38]\\]. All samples were analyzed on a MoFlo XDP Cell Sorter (Beckman Coulter) coupled with a Quanta SC (Beckman Coulter) at the Flow Cytometry Analysis and Sorting Services, Institute of Plant and Microbial Biology (IPMB), Academia Sinica (AS), Taiwan. Additionally, the genome size of RIL59 after sequencing was estimated by analyzing the k-mer distribution by use of the KmerGenie program \\[[@CR50]\\].\n\nDNA and RNA extraction {#Sec20}\n----------------------\n\nDNA was extracted from 0.5\u00a0g of the first leaves of 7-day-old plants of TC1966, NM92, RIL 59 and the three sets of RILs with use of the Plant Genomic DNA Extraction Minprep kit (Viogene, Taipei). RNA was extracted from 1-month-old whole plants, 2- to 7-day-old seedlings, open flowers and flower buds and pods at a series of developmental stages and from seeds of the bruchid-resistant line RIL59. RNA was also extracted from seed of bruchid-resistant and -susceptible parents (TC1966, NM92) and pairs of resistant and susceptible RILs, RIL38 and RIL39, RIL54 and RIL55, and RIL153 and RIL156. Each RIL pair originated from the same F2 plant and possessed the allele of the bruchid-resistant or -susceptible parent at any locus. Total RNA was extracted by use of the pine tree method with minor modifications \\[[@CR51]\\]. DNA contamination of the RNA sample was removed with use of the TURBO DNA-free Kit (Ambion). RNA quality was analyzed on a Bioanalyzer RNA 6000 NanoChip (Agilent Technologies, Santa Clara, CA) coupled with an Agilent 2100 Bioanalyzer (Agilent Technologies) at the DNA Microarray Core Laboratory, IPMB, AS, Taiwan.\n\nNucleotide sequencing and genome assembly {#Sec21}\n-----------------------------------------\n\nDNA and RNA sequencing involved an Illumina Hiseq2000 platform. Two paired-end and two mate-pair libraries were constructed from RIL59 DNA (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1). The DNA was randomly fragmented and size-fractionated by electrophoresis. DNA fragments of 5\u00a0K, 2\u00a0K, 500 and 180\u00a0bp were purified and ligated with adapters to generate libraries for Hiseq sequencing. The DNA extracted from NM92 and TC1966 was used to construct a 500-bp library for Hiseq sequencing. The DNA sequence data were quality-filtered (25 of the first 35 bases at the 5\u2032 end with a Phred quality score\u2009\\>\u200930 for read retention), and reads with ambiguous base-calls and\u2009\\>\u200985\u00a0% low complexity sequences were discarded. Contigs and scaffolds were assembled by use of ALLPATHS \\[[@CR52]\\].\n\nTwo types of libraries were prepared for RNA sequencing (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1). The total RNA extracted from 1-month-old whole plants, 2- to 7-day-old seedlings, flowers and pods at a series of developmental stages of RIL59 underwent rRNA removal, fragmentation, first- and second-strand cDNA synthesis, adapter ligation, PCR amplification and sequencing, and RNA from seeds of two parents and 10 RILs (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1) was submitted to poly(A) RNA enrichment and fragmentation before sequencing.\n\nGene annotation {#Sec22}\n---------------\n\nRepeats were masked on the assembled genome by use of RepeatMasker \\[[@CR53]\\] and RepeatModeler \\[[@CR54]\\] with *de novo* repeat prediction along with the TIGR plant repeat database \\[[@CR55]\\] and Repbase (2014/01/31). Quality-filtered RNA-seq reads of RIL59 were aligned to the repeat masked genome by use of STAR \\[[@CR56]\\]. Transcript assembly for exon sequences involved use of Cufflinks \\[[@CR57]\\]. NCBI soybean refseq protein sequences were aligned to the repeat masked genome by using exonerate \\[[@CR58]\\]. The results of the alignments with the RNA sequences, the soybean refseq protein sequences and the assembled transcript sequences were used to generate extrinsic data for the gene prediction tool Augustus \\[[@CR59]\\]. Transcripts from predicted gene-models were aligned against the NCBI non-redundant set of proteins by using Blastx (E-value 1e^\u22125^) to find homologues. The best alignment for each transcript was retained as an annotation. For functional annotation, Blast analysis of bruchid-resistant--associated genes involved use of Blastn for nucleotides and Blastx for protein homologs and putative function domains. The functional categories of a gene were determined by referring to the gene and protein description and the putative functional domain, if present, and were based on information from the KEGG website (<http://www.genome.jp/kegg/>). The best Blastx hit plus the functional description is presented as Blast results.\n\nRNA-seq analysis {#Sec23}\n----------------\n\nThe RNA-seq reads for all samples (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1) were trimmed for low-quality bases and then aligned individually for each sample to the set of annotated transcripts by using BWA MEM \\[[@CR60]\\]. For each dataset, transcript quantification involved use of eXpress \\[[@CR61]\\] to calculate the transcripts per million (TPM) for each transcript. Fold change (FC) of gene expression was determined by calculating the ratio of TPM for resistant and susceptible samples. A transcript was denoted as up- or downregulated if the ratio was\u2009\\>\u20092 or\u2009\\<\u20090.5, respectively; otherwise, it was denoted as non-differentially expressed.\n\nWe also used DESeq for differential expression analysis of transcripts by calculating the total read counts of a gene in each sample. The DESeq analysis used the default parameter described previously \\[[@CR29]\\]. A transcript was denoted as a DEG with P~adj~\u2009\\<\u20090.1, *P*\u2009\\<\u20090.05 \\[[@CR29]\\], with FC resistance to susceptible\u2009\\>\u20092 or\u2009\\<\u20090.5, respectively. Otherwise, a transcript was denoted as non-differentially expressed.\n\nRT-qPCR validation and genomic PCR for genotyping and sequencing {#Sec24}\n----------------------------------------------------------------\n\nRNA samples extracted from seeds of RIL59, TC1966, NM92 and 3 sets of RILs were used as templates for RT-pPCR analysis. Primer pairs (Additional file [8](#MOESM8){ref-type=\"media\"}: Table S8) for each gene were designed with use of Primer Express v2.0 (Applied Biosystems). qPCR involved use of the QuantStudio 12\u00a0K Flex System (Applied Biosystems) under a cycling profile of 50\u00a0\u00b0C for 2\u00a0min, then 95\u00a0\u00b0C for 10\u00a0min, and 40\u00a0cycles of 95\u00a0\u00b0C for 15\u00a0s and 60\u00a0\u00b0C for 1\u00a0min. The relative quantity (RQ) of gene expression was determined by the 2^\u2212\u0394\u0394CT^ method (\u0394CT\u2009=\u2009CT of interest gene\u2009\u2212\u2009CT of control gene; \u0394\u0394CT\u2009=\u2009\u0394CT of experimental mungbean\u2009\u2212\u2009\u0394CT of control mungbean) \\[[@CR62]\\]. Some genes were not detected in all mungbean lines, so we denoted their CT value as 40 to easily compare the RQ of gene expression between different mungbean lines. The expression of the mungbean acting gene, *g12676*, was used as an internal control to normalize the expression of genes tested. Three experimental repeats were performed.\n\nGenomic PCR was performed with primer pairs (Additional file [8](#MOESM8){ref-type=\"media\"}: Table S8) designed for molecular marker validation and for DNA sequencing. For marker testing, the PCR products for the CAP marker OPW02a4 were further digested with the *Hae*III restriction enzyme (New England) at 37\u00a0\u00b0C following the user instructions. The markers were analyzed on an agarose gel or a 6\u00a0% polyacrylamide gel. For sequence comparison, the PCR products were sequenced and the obtained DNA sequences were aligned with use of ContigExpress in Vector NTI Suite 9 (Invitrogen) to indicate NVs between R and S mungbean.\n\nMapping of bruchid-resistance--associated genes on mungbean pseudochromosome {#Sec25}\n----------------------------------------------------------------------------\n\nThe BLAT program \\[[@CR63]\\] was used for gene location mapping. The coding sequences and promoter region (2-kb upstream if any) of genes were aligned against the draft mungbean VC1973A genome \\[[@CR28]\\]. The applied tile size was 11 and the step size was set to 5. The pslReps option was used to choose the possible location with use of the singleHit option.\n\nEthics approval and consent to participate {#Sec26}\n==========================================\n\nNot applicable.\n\nConsent for publication {#Sec27}\n=======================\n\nNot applicable.\n\nAvailability of data and materials {#Sec28}\n==================================\n\nThis Whole Genome Shotgun project of the genome assembly for bruchid-resistant mungbean RIL59 has been deposited at DDBJ/EMBL/GenBank under the accession LJIH00000000. The RNA-seq data for mungbean lines are available form the NCBI Bioproject: PRJNA276314.\n\nAdditional files {#Sec29}\n================\n\nAdditional file 1: Table S1.Sequencing libraries for RIL59 genome assembly and transcriptome analysis. (DOC 42\u00a0kb)Additional file 2: Table S2.Differentially expressed genes (DEGs) associated with bruchid resistance of mungbean. (XLS 67\u00a0kb)Additional file 3: Table S3.DEGs involved in metabolic pathways. (XLS 31\u00a0kb)Additional file 4: Table S4.Nucleotide variations (NVs) in the 2-kb promoter region of DEGs. (XLS 79\u00a0kb)Additional file 5: Table S5.Transcriptome-based NVs on exons of genes between bruchid-susceptible mungbean NM92 and bruchid-resistant mungbean TC1966 and RIL59. (XLS 57\u00a0kb)Additional file 6: Table S6.Bruchid-resistance--associated sequence-changed-protein genes (SCPs) of mungbean. (XLS 82\u00a0kb)Additional file 7: Table S7.SCPs involved in metabolic pathways. (XLS 35\u00a0kb)Additional file 8: Table S8.Primers used in the experiments. (DOCX 89\u00a0kb)Additional file 9: Table 9.Marker test of *g779p*, *g34480p*, *g34458p*, DMB-SSR158 and OPW02a4 on mungbean recombinant inbred lines (RILs). (XLSX 432\u00a0kb)\n\n*Br*\n\n:   bruchid-resistant gene\n\nCAP\n\n:   cleaved amplified polymorphic DNA\n\nDEG\n\n:   differential expressed gene\n\nFC\n\n:   fold change\n\nGb\n\n:   giga base pairs\n\nkb\n\n:   kilo base pairs\n\nLTR\n\n:   long tandem repeat\n\nMb\n\n:   mega base pairs\n\nMITE\n\n:   miniature inverted-repeat transposable elements\n\nNGS\n\n:   next generation sequencing\n\nNV\n\n:   nucleotide variation\n\nR\n\n:   bruchid-resistant\n\nRIL\n\n:   recombinant inbred line\n\nS\n\n:   bruchid-susceptible\n\nSCP\n\n:   sequence-changed-protein gene\n\nSSR\n\n:   simple sequence repeat\n\nTE\n\n:   transposable element\n\nTPM\n\n:   transcripts per million\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n\nMSL, LFOC, CYC, RS and HFL conceived of and designed experiments. YWW verified molecular markers in 61 RILs. CYK performed genome size estimation and prepared DNA and RNA samples for sequencing and PCR verification. TCYK and KYL performed genome assembly and gene annotation. TCYK and WJL performed transcriptome comparison with TPM fold-change and DESeq by using R software. DCW mapped putative *Br* genes and identified NVs. MSL drafted the manuscript, validated gene expression and NVs, and designed molecular markers. RS, TCYK, CYK, DCW, WJL and CPL participated in the coordination of the study and discussions on data interpretation and helped to draft the manuscript. All authors read and approved the final manuscript.\n\nWe thank Dr. Huei Mei Chen, AVRDC, for providing seeds for the experiments. We thank the Flow Cytometry Analysis and Sorting Service Lab, Institute of Plant and Microbial Biology (IPMB), Academia Sinica (AS), for assisting in flow-cytometry analysis; the DNA Microarray Core, IPMB, AS, for RNA quality analysis; the DNA Analysis Core, IPMB, AS, for help in RT-qPCR and small fragment DNA sequencing; and the High Throughput Genomics Core, AS, for performing RNA sequencing.\n\nFunding {#FPar1}\n=======\n\nThis work was supported by the Innovative Translational Agricultural Research Program (Project \\#2014CP04), Academia Sinica (AS), Taiwan.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nGrowth factor signaling supports cell survival through various pathways. Thus, deprivation of growth factor ultimately results in apoptosis. The decisive step for the induction of intrinsic apoptosis is the mitochondrial outer membrane permeabilization (MOMP). This results in the release of cytochrome c and other proteins from the mitochondrial intermembrane space into the cytosol, leading to apoptosome formation, caspase activation, and apoptosis. MOMP is controlled by proteins of the BCL-2 family. While the pro-apoptotic BCL-2 proteins BAX and BAK are required for the formation of a mitochondrial outer membrane pore, their activity is induced by BH3-only proteins (PUMA, BIM, Bid, and others). MOMP is prevented by related proteins with anti-apoptotic function (like BCL-2, MCL-1, BCL-xL)^[@CR1]^.\n\nMOMP is controlled by growth factor availability, which induces various pathways promoting cell survival. A key pro-survival pathway is the PI3K/AKT signaling pathway, which can prevent MOMP and apoptosis through regulating a number of substrates. For instance, AKT was shown to phosphorylate and inactivate the transcription factor FOXO3A as well as glycogen synthase kinase-3 (GSK-3). The inactivation of both FOXO3A and GSK-3 was shown to play an important role for the pro-survival activity of PI3K/AKT signaling^[@CR2]--[@CR4]^. More specifically, it was shown that the suppression of FOXO3A plays an essential role for the suppression of *Puma* induction and cell death by PI3K signaling^[@CR5]^.\n\nThe death promoting role of GSK-3 is instrumental for p53-mediated *Puma* induction and apoptosis: GSK-3 phosphorylates the histone acetyl transferase Tip60 (also known as KAT5), which stimulates Tip60 to acetylate p53 at K120, resulting in the transcriptional induction of *Puma* and apoptosis upon induction of p53^[@CR6]^. Interestingly, GSK-3 was also shown to modulate the transcriptional activity of FOXO3A^[@CR7],[@CR8]^.\n\nIn the present study, employing knockout by CRISPR/Cas9, we systematically investigated the role of GSK-3-dependent factors required for apoptosis induction by IL-3 deprivation. We show that PUMA is the main pro-apoptotic protein responsible for apoptosis in this context, and that the induction of *Puma* is mediated by a FOXO3A-, p53-, and GSK-3-dependent mechanism.\n\nResults {#Sec2}\n=======\n\nApoptosis induced by growth factor withdrawal requires GSK-3-dependent PUMA induction {#Sec3}\n-------------------------------------------------------------------------------------\n\nWhen IL-3-dependent cells such as Ba/F3 or FL5.12 cells (two murine pro B cell lines) are deprived of the growth factor, they undergo rapid apoptosis. Additional treatment with the highly selective GSK-3 inhibitor CT98014 completely blocked IL-3-withdrawal-induced apoptosis of Ba/F3 cells as observed previously^[@CR9]^ (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). We aimed at systematically defining the pro-apoptotic factors involved in IL-3 withdrawal-induced apoptosis and at investigating their link to GSK-3. To address the role of pro-apoptotic BH3-only proteins for growth factor-withdrawal-induced apoptosis, we transduced Ba/F3 cells with the lentiCRISPRv2 system targeting either *Puma* or *Bim*. As shown in Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}, apoptosis by IL-3 deprivation was substantially reduced in Ba/F3 cells expressing CRISPR/Cas9 targeting *Puma*, while loss of *Bim* conferred only moderate protection from cell death. This effect was even more pronounced in the IL-3-dependent cell line FL5.12 (Fig.\u00a0[S1A](#MOESM1){ref-type=\"media\"}). To further verify the role of PUMA in this system, clones derived from individual cells (single-cell clones) were generated from the CRISPR/Cas9-transduced cultures and cells with frameshift mutations on both *Puma* alleles or both *Bim* alleles were selected. Almost all *Puma*^\u2212/\u2212^ single-cell clones were strongly protected from IL-3 withdrawal-induced apoptosis (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}) while *Bim*^\u2212/\u2212^ single-cell clones exhibited no statistically significant survival advantage (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). The protective effect of *Puma* depletion lasted at least 24\u2009h, however, the cells committed to apoptosis at later time points. *Puma*^\u2212/\u2212^ Ba/F3 cells additionally treated with the GSK-3 inhibitor CT98014 showed a similar kinetic but the protective effect was more pronounced even after 40\u2009h (Fig.\u00a0[S1B](#MOESM1){ref-type=\"media\"}). Thus, the kinase activity of GSK-3 as well as the induction of *Puma*\u00a0are rate-limiting for apoptosis induced by IL-3 withdrawal. The stronger anti-apoptotic effect achieved by inhibition of GSK-3 than by the loss of PUMA\u00a0alone is consistent with other cell death regulatory factors being regulated by GSK-3, such as MCL-1^[@CR9]^ (Fig.\u00a0[S1C](#MOESM1){ref-type=\"media\"}). We next asked whether PUMA is transcriptionally regulated in a GSK-3-dependent manner upon growth factor withdrawal. Wild-type Ba/F3 cells were deprived of IL-3, treated with GSK-3 inhibitor (CT98014) and *Puma* mRNA levels were analyzed by quantitative RT-PCR. IL-3 withdrawal-induced *Puma* mRNA up to 2-fold after 7.5\u2009h while *Puma* mRNA was reduced upon treatment with CT98014 in the absence of IL-3 (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}). This effect was reflected by the protein levels of PUMA in Ba/F3 wt cells: PUMA was induced upon IL-3 withdrawal, but this upregulation was completely blocked by addition of CT98014 (Fig.\u00a0[1f](#Fig1){ref-type=\"fig\"}). Loss of PI3K is permitting GSK-3 activity by relieving the suppression of GSK-3 by AKT-mediated phosphorylation. Consistently, we found that the pharmacological inhibition of PI3K resulted in strong induction of PUMA (Fig.\u00a0[S1D](#MOESM1){ref-type=\"media\"}).Fig. 1Apoptosis induced by growth factor withdrawal requires GSK-3-dependent PUMA induction.**a** Ba/F3 cells were deprived of IL-3 in the presence or absence of CT98014 (0.75\u2009\u00b5M) and analyzed for apoptosis by Annexin V staining and flow cytometry analysis. Error bars represent SD from technical replicates. **b** Ba/F3 cells expressing CRISPR/Cas9 targeting *Luciferase* (crLUC), *Puma* (crPuma), or *Bim* (crBim) were deprived of IL-3 in presence or absence of CT98014 (0.75\u2009\u00b5M) and analyzed for apoptosis by Annexin V staining after 18\u2009h. Error bars represent SD from technical replicates. **c** Ba/F3 *Puma*^\u2212/\u2212^ single-cell clones and Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were deprived of IL-3 for 18\u2009h and analyzed for apoptosis by Annexin V staining. Each dot represents the mean of two independent experiments analyzing an individual single-cell clone. Error bars represent 95% confidence interval from two independent experiments (*n*\u2009=\u20092). Significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. **d** Ba/F3 *Bim*^\u2212/\u2212^ single-cell clones and Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were deprived of IL-3 for 18\u2009h and analyzed for apoptosis by Annexin V staining. Each dot represents the mean of two independent experiments analyzing an individual single-cell clone. Error bars represent the\u00a095% confidence interval from two independent experiments (*n*\u2009=\u20092). Significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. ns\u2009=\u2009not significant. **e** Ba/F3 cells were deprived of IL-3 in presence or absence of\u00a0CT98014 (0.75\u2009\u00b5M) and harvested after the indicated time. The 0\u2009h condition represents cells kept in medium with IL-3. RNA levels were analyzed by qRT-PCR with primers for *Puma* and *L32* as internal reference. Data points show relative (mRNA) *Puma* induction compared to cells kept in medium with IL-3. Error bars represent SD from technical replicates. **f** Cells from **a** were subjected to western blotting and analyzed with the antibodies indicated.\n\nWe next generated IL-2-dependent murine primary lymphocytes, which were deprived of the growth factor. Consistent with the results obtained with IL-3-dependent cell lines, removal of IL-2 induced PUMA and apoptosis, while this was abrogated in presence of the GSK-3 inhibitor (Fig.\u00a0[S1E](#MOESM1){ref-type=\"media\"}, Fig.\u00a0[S1F](#MOESM1){ref-type=\"media\"}).\n\nWe therefore conclude that PUMA represents the main BH3-only protein mediating IL-3 and IL-2 withdrawal-induced apoptosis induction and that the transcriptional induction of *Puma\u00a0*is dependent on GSK-3 activity.\n\np53 has a minor role for GSK-3-dependent PUMA induction {#Sec4}\n-------------------------------------------------------\n\nWe next investigated the identity of the transcription factor, which mediates the GSK-3-dependent transcriptional induction of *Puma*\u00a0upon loss of growth-factor-induced PI3K signaling. *Puma* is a crucial pro-apoptotic target of p53^[@CR10],[@CR11]^. While p53 is not stabilized upon growth factor withdrawal, p53 is nevertheless able to induce some of its target genes (such as *Mdm2*) when present at very low levels. Thus, even in the absence of DNA damage, we considered p53 as a candidate for GSK-3-dependent *Puma* induction upon loss of pro-survival signaling during IL-3 withdrawal. To address the role of p53 in this context, we inhibited PI3K by GDC-0941 in HCT116 *p53*^+/+^ and *p53*^\u2212/\u2212^ cells in order to induce PUMA, as described before (see Fig.\u00a0[S1D](#MOESM1){ref-type=\"media\"}). As shown in Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}, inhibition of PI3K-induced PUMA in *p53*^+/+^ as well as *p53*^\u2212/\u2212^ cells, indicating that p53 is not a major transcription factor responsible for *Puma* induction in this setting. Consistent with the data shown before, the induction of PUMA was prevented by pharmacological inhibition of GSK-3. To further investigate the role of p53, we knocked out *p53* by CRISPR/Cas9 in Ba/F3 cells, which have an intact p53 signaling pathway^[@CR6]^. Upon IL-3 withdrawal, Ba/F3 cells lacking *p53* exhibited some protection from apoptosis (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}), which was however not as strong as observed with cells lacking PUMA (see Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}). We generated *p53*^\u2212/\u2212^ single-cell clones and subjected them to IL-3 deprivation. We unexpectedly observed that individual Ba/F3 cell clones lacking *p53* exhibited quite some variation regarding the protection from IL-3 withdrawal-induced apoptosis. Nevertheless, on average, the protection conferred by lack of p53 was modest (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). Interestingly, a variation in IL-3-induced apoptosis was also observed in different single-cell clones expressing CRISPR/Cas9 targeting *Luciferase* and---of note---in wild-type Ba/F3 cells, indicating that individual clones always seems to exhibit some variation, even in the absence of previous manipulation. We therefore always analyzed a high number of clones (Fig.\u00a0[S2A+B](#MOESM2){ref-type=\"media\"}). Consistent with the modest protection of *p53*^\u2212/\u2212^ single-cell clones from IL-3 deprivation, PUMA induction was still observed in *p53*^\u2212/\u2212^ single-cell clones upon IL-3 withdrawal (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}) or PI3K inhibition by LY294002 (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}) and was still dependent on GSK-3 (Fig.\u00a0[S2C](#MOESM2){ref-type=\"media\"}). However, the quantification of PUMA expression levels revealed that, while *Puma* was still induced, the PUMA protein levels were generally somewhat reduced in *p53*^\u2212/\u2212^ cells (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}, right). Together, although there is a GSK-3-dependent effect of p53 on PUMA and apoptosis induction, p53 seems not to be the main GSK-3-dependent transcription factor responsible for the induction of PUMA and apoptosis upon IL-3 withdrawal.Fig. 2p53 has a minor role for GSK-3-dependent PUMA induction.**a** HCT116 *p53*^\u2212/\u2212^ or *p53*^+/+^ were treated with GDC-0941 (GDC, 10\u2009\u00b5M), CT98014 (CT, 0.75\u2009\u00b5M), or a combination of both for 7\u2009h. The cells were harvested, subjected to western blotting and analyzed by the antibodies indicated. **b** Ba/F3 expressing CRISPR/Cas9 targeting *p53* or *Luciferase* were deprived of IL-3, in presence or absence of\u00a0 CT98014 (0.75\u2009\u00b5M) for 18\u2009h and analyzed for apoptosis by Annexin V staining. Error bars represent SD from technical replicates. **c** Ba/F3 *p53*^\u2212/\u2212^ single-cell clones and Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were deprived of IL-3 for 18\u2009h and analyzed for apoptosis by Annexin V staining. Each dot represents the mean of two independent experiments analyzing an individual single-cell clone. Error bars represent 95% confidence interval from two independent experiments (*n*\u2009=\u20092). Significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. ns\u2009=\u2009not significant. **d** Ba/F3 *p53*^\u2212/\u2212^ single-cell clones and two independent cell lines of Ba/F3 cells infected with CRISPR/Cas9 targeting *Luciferase* were deprived of IL-3 for 12\u2009h, harvested and analyzed by western blotting by probing with antibodies as indicated. Bands were quantified using FusionCapt Advance Solo 4 16.08. PUMA levels were normalized to GSK-3 levels (loading control). **e** Ba/F3 *p53*^\u2212/\u2212^ single-cell clones and two independent cell lines of Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were treated with LY294002 (10\u2009\u00b5M) or DMSO (\u2212) for 18\u2009h, harvested and analyzed by western blotting with the antibodies indicated. Bands were quantified using FusionCapt Advance Solo 4 16.08. PUMA levels were normalized to GSK-3 levels (loading control).\n\nFOXO3A is an important *Puma* inducer upon growth factor deprivation {#Sec5}\n--------------------------------------------------------------------\n\nTo test other transcription factors for their role in inducing *Puma* upon IL-3 deprivation, we generated Ba/F3 expressing CRISPR/Cas9 targeting *Foxo1*, *Foxo3a*, and the p53 relatives *p63* and *p73*. We then subjected these cells to IL-3 withdrawal and analyzed apoptosis induction. Among the cells we tested, only *Foxo3a*-targeted cells showed a protection from IL-3 deprivation (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}, Fig.\u00a0[S3A](#MOESM3){ref-type=\"media\"}). Single-cell clones generated from these cells again showed a wide distribution of sensitivity to IL-3 withdrawal but were on average significantly protected compared to control cells (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}). The *Foxo3a*^\u2212/\u2212^ single-cell clones showed a diminished capacity to induce PUMA upon IL-3 withdrawal (Fig.\u00a0[S3B](#MOESM3){ref-type=\"media\"}). Because of the high variation of individual *Foxo3a*^\u2212/\u2212^ single-cell clones to induce PUMA, they were pooled and we observed a clear reduction of the signal for PUMA upon IL-3 withdrawal (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}). Likewise, *Foxo3a*^\u2212/\u2212^ single-cell clones treated with the PI3K inhibitor LY294002 exhibited reduced *Puma* induction, as evident from the quantification of the PUMA expression levels (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}).Fig. 3FOXO3A is an important *Puma* inducer upon growth factor deprivation.**a** Ba/F3 expressing CRISPR/Cas9 targeting *Foxo3a*, *Foxo1*, or *Luciferase* were deprived of IL-3 in the presence or absence of CT98014 (0.75\u2009\u00b5M) for 18\u2009h and analyzed for apoptosis by Annexin V staining. **b** Ba/F3 *Foxo3a*^\u2212/\u2212^ single-cell clones and Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were deprived of IL-3 for 18\u2009h and analyzed for apoptosis by Annexin V staining. Each dot represents the mean of three independent experiments analyzing an individual single-cell clone. Error bars represent 95% confidence interval from three independent experiments (*n*\u2009=\u20093). Significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. **c** Equal cell numbers of eight different Ba/F3 *Foxo3a*^\u2212/\u2212^ single-cell clones were pooled. This pool and Ba/F3 expressing CRISPR/Cas9 targeting *Luciferase* were deprived of IL-3 for 9\u2009h, harvested and analyzed by western blotting with the antibodies indicated. **d** Ba/F3 *Foxo3a*^\u2212/\u2212^ single-cell clones and two independent cell lines of Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* were treated with LY294002 (10\u2009\u00b5M) or DMSO (\u2212) for 18\u2009h, harvested and analyzed by western blotting with the antibodies indicated. Bands were quantified using FusionCapt Advance Solo 4 16.08. PUMA levels were normalized to GSK-3 levels (loading control).\n\nDouble knockout of *Foxo3a* and *p53* fully protects cells from IL-3-induced apoptosis and prevents *Puma* induction {#Sec6}\n--------------------------------------------------------------------------------------------------------------------\n\nWe next addressed the combined role of p53 and FOXO3A to induce apoptosis and PUMA upon IL-3 withdrawal. We generated cells expressing CRISPR/Cas9 targeting *p53* and *Foxo3a* simultaneously and identified two single-cell clones with frameshift mutations on both alleles of each *p53* and *Foxo3a* (double knock out, DKO). As shown in Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}, both DKO clones were protected from IL-3 deprivation-induced apoptosis. In line with this, PUMA induction was absent upon treatment with LY294002 (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}) or IL-3 withdrawal (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}) in these cells. Thus, p53 and FOXO3A together are required for PUMA upregulation and apoptosis upon IL-3 withdrawal, possibly also compensating for each other to some extent in cells lacking either p53 or FOXO3A.Fig. 4Double knockout of *Foxo3a* and *p53* fully protects cells from IL-3-induced apoptosis and prevents *Puma* induction.**a** Ba/F3 *p53*^\u2212/\u2212^ *Foxo3a*^\u2212/\u2212^ single-cell clones (DKO \\#1 and DKO \\#2) and Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* (crLUC) were deprived of IL-3 for 18\u2009h and analyzed for apoptosis by Annexin V staining. Error bars represent 95% confidence intervals from four independent experiments (*n*\u2009=\u20094). Significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. ns\u2009=\u2009not significant. \\*\\*\\**P*\u2009\\<\u20090,001. **b** The same cells were treated with LY294002 (10\u2009\u00b5M) for 18\u2009h or left untreated. The cells were harvested after the treatment and subjected to western blotting. The protein levels were analyzed by the antibodies indicated. **c** The same cells were deprived of IL-3 for 18\u2009h, harvested after the treatment and subjected to western blotting.\n\nThe induction of *Puma* by FOXO3A depends on GSK-3 {#Sec7}\n--------------------------------------------------\n\nTo confirm the major role of FOXO3A and its GSK-3 dependency in IL-3 withdrawal-induced apoptosis, we re-expressed FOXO3A with silent mutations, rendering it resistant to CRISPR/Cas9 cleavage, in a Ba/F3 bulk culture which expressed a CRISPR/Cas9 construct targeting *Foxo3a* (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}) as well as in a *Foxo3a*^\u2212/\u2212^ single-cell clone (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). Re-expression of FOXO3A re-established apoptosis induction to a comparable level as in control cells transduced with FOXO3A in bulk culture and the single-cell clone (Fig.\u00a0[5a, b](#Fig5){ref-type=\"fig\"}). Importantly, restoring FOXO3A and thereby the competence to undergo apoptosis did not relieve the requirement for GSK-3 activity, as inhibition of GSK-3 suppressed apoptosis induced by IL-3 withdrawal in FOXO3A overexpressing cells. Thus, FOXO3A-dependent PUMA and apoptosis induction upon growth factor withdrawal requires GSK-3. Of note, the induction of apoptosis as well as the induction of PUMA in *Foxo3a*^\u2212/\u2212^ cells was still dependent on GSK-3 (Fig.\u00a0[S4A+B](#MOESM4){ref-type=\"media\"}), confirming that the regulation of PUMA by p53 is also controlled by GSK-3^[@CR6]^.Fig. 5The induction of *Puma* by FOXO3A depends on GSK-3.**a** Ba/F3 cells expressing CRISPR/Cas9 constructs targeting *Luciferase* or *Foxo3a* were infected with retrovirus encoding CRISPR/Cas9-resistant human FOXO3A control retrovirus (empty vector, EV). The cells were deprived of IL-3 in the presence or absence of CT98014 (CT, 0.75\u2009\u00b5M) for 18\u2009h and analyzed for apoptosis by Annexin V staining. The significance was tested by one-way ANOVA with post hoc Tukey's multiple comparison test. Error bars represent 95% confidence intervals from four independent experiments (*n*\u2009=\u20094). **b** Ba/F3 expressing CRISPR/Cas9 constructs targeting *Luciferase* (crLUC) or a *Foxo3a*^\u2212/\u2212^ single-cell clone (\\#4) were infected with retrovirus expressing human FOXO3A or none (empty vector, EV). The cells were deprived of IL-3 in presence\u00a0or\u00a0absence of\u00a0CT98014 (0.75\u2009\u00b5M, CT) for 18\u2009h, then analyzed for apoptosis by Annexin V staining. Error bars represent SD from technical replicates. **c** Cells from **b** were harvested after 18\u2009h and analyzed by western blotting with the antibodies indicated. **d** HCT116 *p53*^\u2212/\u2212^ cells were transfected by Lipofectamine2000\u00ae with vectors encoding FOXO3A or control vector (empty). Twenty-four hours after transfection, the cells were treated with DMSO or GDC-0941 (10\u2009\u00b5M) for 5\u2009h. The cells were harvested and analyzed by western blotting, probing with antibodies as indicated.\n\nWhile PUMA protein levels were reduced in empty vector-transduced *Foxo3a*^\u2212/\u2212^ cells upon IL-3 withdrawal, PUMA induction was restored in *Foxo3a*^\u2212/\u2212^ cells re-expressing CRISPR/Cas9-resistant FOXO3A (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}). Consistently, FOXO3A overexpression in HCT116 *p53*^\u2212/\u2212^ cells increased PUMA induction in cells treated with the AKT inhibitor GDC-0941 (Fig.\u00a0[5d](#Fig5){ref-type=\"fig\"}). We therefore conclude that FOXO3A induces *Puma* when Ba/F3 are deprived of IL-3, while p53 contributes to this induction, and that both transcription factors require GSK-3 activity for the induction of PUMA and apoptosis.\n\nFOXO3A requires GSK-3 activity for full transcriptional activity {#Sec8}\n----------------------------------------------------------------\n\nAs shown above, the pharmacological inhibition of GSK-3 was sufficient to prevent growth factor withdrawal-induced PUMA induction and apoptosis. We previously demonstrated that when p53 is stabilized by DNA damage, GSK-3 activity regulates the capacity of p53 to induce *Puma* by phosphorylating Tip60 at S86, which stimulates Tip60 to acetylate lysine 120 (K120) of p53^[@CR6]^. K120-acetylated p53 then induces *Puma* which leads to apoptosis^[@CR12]--[@CR14]^. However, in *p53*^\u2212/\u2212^ cells we observed FOXO3A-dependent PUMA induction and apoptosis, which was also dependent on GSK-3 activity (Figs.\u00a0[2a--c](#Fig2){ref-type=\"fig\"} and\u00a0[5d](#Fig5){ref-type=\"fig\"}). To investigate the role of GSK-3 for the transcriptional regulation of *Puma* by FOXO3A, we generated\u00a0a promoter reporter system where Luciferase expression is under control of the *Puma* promoter, which includes a binding site for FOXO3A^[@CR5]^. It is well established that FOXO3A is controlled by the PI3K/AKT pathway as phosphorylation of FOXO3A residues T32, S253, and S315 by AKT results in retention of the transcription factor in the cytosol^[@CR2]^. We expressed a FOXO3A-triple mutant (TM; T32A-S253A-S315A), which cannot be phosphorylated by AKT and is considered to be constitutive active, along with the *Puma* promoter reporter constructs. As shown in Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}, overexpression of FOXO3A-TM induced Luciferase expression controlled by the *Puma* promoter. This was dependent on a FOXO3A binding site in the promoter region, as mutation of the binding site for FOXO3A prevented Luciferase induction. However, when we added the GSK-3 inhibitor CT98014, the ability of FOXO3A-TM to induce Luciferase through the *Puma* promoter was substantially decreased. We therefore conclude that FOXO3A, in order to fully exhibit its transcriptional activity at the *Puma* promoter, requires not only the dephosphorylation of AKT phosphorylation sites in FOXO3A, but also the activity of GSK-3.Fig. 6FOXO3A requires GSK-3 activity for full transcriptional activity.**a** 293T HEK cells were transfected with a reporter plasmid with Luciferase driven by a wild type (wt) or FOXO3A binding site mutant (FoxoBSmut) *Puma* promoter fragment, a Renilla reporter plasmid as internal control and a construct encoding FOXO3A-TM or empty vector. Eight hours after transfection, the cells were treated with CT98014 (CT, 0.75\u2009\u00b5M,) or DMSO for 18\u2009h. Luciferase activity was analyzed and normalized to Renilla activity. Error bars represent SD from technical replicates. **b** Ba/F3 cells expressing CRISPR/Cas9 constructs targeting *Luciferase* or a *Foxo3a*^\u2212/\u2212^ single-cell clone were infected with retrovirus expressing human FOXO3A-TM or none (empty vector, EV). The cells were deprived of IL-3+/\u2212 CT98014 (CT, 0.75\u2009\u00b5M) for 18\u2009h, then stained with Annexin-V-FITC and analyzed by flow cytometry. Error bars represent SD from technical replicates. This experiment was done together with the one shown in Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}, and the controls are identical. **c** HEK 293T cell were transfected with constructs encoding wild-type, FLAG-tagged FOXO3A (wt) or FLAG-tagged FOXO3A-TM (TM) or control vector (\u2212). Eight hours later, the cells were treated with CT98014 (0.75\u2009\u00b5M) or left untreated for 18\u2009h. The cells were lysed and 2% of the lysate were kept as input. FLAG immunoprecipitation was performed with the remaining lysate. After washing and elution with 3xFLAG peptide, eluate and input fractions were analyzed by western blotting with the antibodies indicated.\n\nWe next established Ba/F3 cell lines overexpressing FOXO3A-TM. We expressed FOXO3A-TM with silent mutations, mediating resistance to CRISPR/Cas9 cleavage, in a Ba/F3 bulk culture expressing a CRISPR/Cas9 construct targeting Foxo3a, or in control cells with CRISPR/Cas9 targeting Luciferase. This led to a slightly increased background of apoptosis even when IL-3 was available. However, although FOXO3A-TM is considered to be constitutively active, IL-3 deprivation was required to induce apoptosis (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}, left). Importantly, apoptosis could be inhibited by addition of CT98014 (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}, right), showing that FOXO3A-TM (which is not controlled by AKT) requires GSK-3 activity to induce apoptosis upon IL-3 withdrawal. As the inhibition of GSK-3 did not affect the binding to 14-3-3 proteins (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}) or the nuclear localization of FOXO3A (Fig.\u00a0[S5A](#MOESM5){ref-type=\"media\"}), the regulation of FOXO3A-mediated transcription by GSK-3 is independent of FOXO3A translocation. Interestingly, we observed an interaction of GSK-3 with FLAG-tagged FOXO3A overexpressed in 293T HEK cells (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}). Together, the results show that activity of GSK-3 is decisive for the FOXO3A and p53-mediated PUMA induction in the absence of PI3K signaling.\n\nDiscussion {#Sec9}\n==========\n\nWe had previously shown that GSK-3 phosphorylates the anti-apoptotic protein MCL-1 upon growth factor withdrawal or PI3K inhibition, which regulates lymphocyte survival^[@CR9],[@CR15]^. In this study, we aimed at systematically identifying the GSK-3-dependent pro-apoptotic factors, promoting apoptosis upon depletion from growth factor. We showed that PUMA rather than BIM mediates IL-3 withdrawal-induced apoptosis, although both proteins were upregulated upon IL-3 deprivation (data not shown). In line with our data, it was previously shown that PUMA and BIM are induced by IL-2 withdrawal^[@CR5],[@CR16]^. Our findings also confirm a previous report that IL-3-dependent myeloid cell lines from mice lacking PUMA, but not BIM, were protected from IL-3 withdrawal-induced apoptosis^[@CR17]^.\n\nFOXO3A and p53 were both described to be important for growth factor withdrawal-induced apoptosis. FOXO3A deficiency can protect cells from cytokine withdrawal-induced apoptosis^[@CR16],[@CR18]^, and FOXO3A was shown to be crucial for PUMA induction upon growth factor withdrawal^[@CR5]^. It was also found that p53, FOXO and E-box-binding transcription factors share many targeted genes in response to PI3K inhibition in Rat-1 cells and a dominant-negative p53 resulted in partial resistance to apoptosis upon PI3K inhibition^[@CR19]^. Likewise, HoxB8-transformed factor-dependent myeloid (FDM) cells from p53^\u2212/\u2212^ mice were less sensitive for IL-3 deprivation-induced cell death and showed reduced PUMA induction. In this study, FDM cells from FOXO3A^\u2212/\u2212^ mice showed an even higher susceptibility to cell death after IL-3 loss^[@CR20]^. The same authors reported that PUMA induction was independent of PI3K, a finding which is not confirmed by our data^[@CR17],[@CR21]^. Instead, we found that inhibition of the PI3K/AKT pathway upregulated PUMA in *p53*^\u2212/\u2212^ HCT116 cells and Ba/F3 cells.\n\nIn addition, p53 negative Ba/F3 cells were fully protected from apoptosis only when GSK-3 was inhibited, which further supports the important apoptosis-regulatory role of the PI3K/AKT pathway also in absence of p53.\n\nWe have shown that the induction of *Puma* by FOXO3A and/or p53 is dependent on GSK-3 activity. Importantly, we observed that even the transcriptional activity of a FOXO3A-TM mutant, which is considered constitutively active due to the loss of the AKT phosphorylation sites, required GSK-3. Our results suggest that the inhibition of GSK-3 is the key pro-survival function of PI3K signaling, being more important than the inactivation of FOXO3A. In line with this finding, full transcriptional activity of Foxo1/3/4 was shown to require GSK-3, as the induction of IGF-IR gene by serum starvation or AKT inhibition required both active GSK-3 and Foxo1/3/4^[@CR8]^. Supporting a PI3K-dependent regulation of FOXO3A independent of the AKT phosphorylation sites, the DNA-binding and transcriptional activity of a triple S/A mutant of the Foxo homolog in *C. elegans*, DAF-16, which is considered constitutive active, was shown to be further increased by PI3K inhibition by LY294002 and Wortmannin^[@CR22]^. Interestingly, promotion of FOXO3A activity to induce PUMA by GSK-3 was independent of the interaction of FOXO3A with 14-3-3 or its subcellular localization.\n\nIt is not clear however, how GSK-3 and FOXO3A cooperate. In a study of Terragni et al., upregulated genes upon inhibition of PI3K were co-regulated by FOXO3A, MITF, and USF1, with the latter two being regulated by GSK-3^[@CR23]^. Another explanation for the functional interaction of FOXO3A and GSK-3 could be provided by the regulation of TIP60. We could previously demonstrate the requirement of GSK-3 for *Puma* induction by p53 in the context of DNA damage and subsequent p53 stabilization by the phosphorylation of the histone acetyl transferase Tip60 (KAT5). Increased KAT activity of Tip60 is induced by GSK-3 mediated phosphorylation, resulting in the acetylation of p53 in the DNA-binding domain at lysine 120, which enables p53 to promote the transcription of *Puma*. It is possible that GSK-3 activated Tip60 acts on chromatin or FOXO3A directly, thereby promoting transcriptional *Puma* induction^[@CR6]^.\n\nA direct phosphorylation of FOXO3A by GSK-3 would also be a possibility as to how GSK-3 directly promotes the transcriptional activity of FOXO3A, which would be consistent with our data demonstrating a cooperation of FOXO3A-TM and GSK-3 for the regulation of the *Puma* promoter.\n\nTogether, in this study, we provide evidence that *Puma* induction by FOXO3A is an important step in IL-3 withdrawal-induced apoptosis, and that the *Puma* inducing function of FOXO3A is dependent on GSK-3 activity. Thus, GSK-3 is crucial for the full activation of FOXO3A transcriptional activity when the PI3K pathway is not active. As pharmacological modulation of kinase signaling pathways is a promising strategy for cancer therapy, a more detailed understanding of the underlying mechanisms will improve defining the crucial targets.\n\nMaterials and methods {#Sec10}\n=====================\n\nReagents and antibodies {#Sec11}\n-----------------------\n\nLY294002 was from Sigma-Aldrich (St. Louis, MO, USA), GDC-0941 and CT98014 were from Axon Medchem (Groningen, Netherlands), Annexin-V-FITC was generated in our lab. Annexin-V-APC (RUO) was from Becton Dickinson (Franklin Lakes, NJ, USA). 4-OHT was from Sigma-Aldrich (St. Louis, MO, USA).\n\nIL-2 and IL-3 were from Peprotech (Rocky Hill, NJ, USA).\n\nThe following antibodies were used for western blotting: Puma (\\#3043) was from Prosci (Poway, CA, USA), human Puma (\\#12450), Foxo3a (\\#2497) were from Cell Signaling Technologies (Danvers, MA, USA). GSK3a/b (sc-56913), 14-3-3 (sc-1657) and NFATc1 (sc-7294) were from Santa Cruz (Dallas, TX, USA). Tubulin (MCA77G) was from Bio-Rad (Hercules, CA, USA). MCL1 (600-401-394S) was from Rockland (Limerick, PA, USA).\n\nFLAG-M2 agarose affinity beads were from Sigma-Aldrich (St. Louis, MO, USA).\n\nGeneration of ko cell lines using lentiviral CRISPR/Cas9 {#Sec12}\n--------------------------------------------------------\n\nThe lentiCRISPRV2 system was used to generate ko cells as described by others^[@CR24]^. At least two different guide RNAs for each target were designed using crispr.mit.edu and a guide RNA targeting *Luciferase* was designed as control (see Table\u00a0[1](#Tab1){ref-type=\"table\"}: gRNAs). The guide RNAs were cloned into the lentiCRISPRV2 plasmid (a gift from Feng Zhang, Addgene plasmid \\# 52961) and lentiviral particles were produced by transfection of 5\u2009\u00b5g lentiCRISPRv2, 1.5\u2009\u00b5g pMISSIONVSV-G (Sigma-Aldrich) and 3\u2009\u00b5g pMISSION GAG POL (Sigma-Aldrich) with Attractene (Quiagen, Hilden, Germany) into 293T HEK cells seeded at 25% confluency in a 78-cm\u00b2 culture plate the day before transfection. The morning after transfection, fresh medium was added. In the evening, 4\u2009ml medium was added. The next day, viral supernatants were harvested, filtered (0,45\u2009\u00b5M) and supplemented with 5\u2009\u00b5g/ml polybrene (Sigma-Aldrich). The target cells were infected by spinfection at 400\u00d7*g* for 10\u2009min. The following day, selection was started using 4\u2009\u00b5g/ml puromycin (Sigma-Aldrich) for at least 3 days and until no viable cells were detected in an uninfected control. Mix cultures were tested for Cas9 cleavage by performing a Surveyor assay. The region of interest was amplified by PCR using the primers listed (Table\u00a0[2](#Tab2){ref-type=\"table\"}: Primers). The Surveyor assay was performed according to the manufacturer (Integrated DNA Technologies). The mix cultures were used for experiments and to generate single-cell clones by limited dilution. Single-cell clones were analyzed by sequencing and clones with a bi-allelic shift in the open reading frame generating a premature STOP codon were chosen.Table 1gRNAsTargetgRNA namegRNA sequence (20xN-NGG)p53 Exon3p53 gRNA \\#1 (759)AGTGAAGCCCTCCGAGTGTC-AGGp53 Exon3p53 gRNA \\#2 (760)AGGAGCTCCTGACACTCGGA-GGGp53 Exon3p53 gRNA \\#3 (761)GACACTCGGAGGGCTTCACT-TGGp63 Exon3p63 gRNA \\#1TCCACAAAGTTCAACTCGAT-GGGp63 Exon4p63 gRNA \\#2CCGTCACGCTATTCTGTGCG-TGGp63 Exon4p63 gRNA \\#3AGCCCCAGGTTCGTGTACTG-TGGp73 Exon3p73 gRNA \\#1CCGGGGTAGTCGGTATTGGA-AGGp73 Exon3p73 gRNA \\#2CGGGGTGTAGGGGCTCGCCG-GGGPuma Exon1Puma gRNA \\#1ATGGCCCGCGCACGCCAGGA-GGGPuma Exon1Puma gRNA \\#2AGCTCTCCGGAGCCCGTAGA-GGGPuma Exon1Puma gRNA \\#3GGAAGGGGCGCGGACTGTCG-CGGBim Exon1Bim gRNA \\#1ACTTACATCAGAAGGTTGCT-TGGBim Exon1Bim gRNA \\#2TTGCGGTTCTGTCTGTAGGG-AGGFoxo1 Exon1Foxo1 gRNA \\#1TCGTCGCGCCGCAACGCGTG-GGGFoxo1 Exon1Foxo1 gRNA \\#2GGAGAGTGAGGACTTCGCGC-GGGFoxo3a Exon1Foxo3a gRNA \\#1CACGCCGCCACCGATCACCA-TGGFoxo3a Exon1Foxo3a gRNA \\#2TCTCGATGGCGCGGGTGATC-AGGLuciferaseLuciferase gRNAACCGCTCCGGCGAAGGCGAA-NGGTable 2PrimersPrimerSequence (5\u2032-3\u2032)p53 forTGTCTGTAAATCCTGCGGGGp53 revGAGGCTAAAAAGGTTCAGGGCp63 Exon3 forTAAGACGGTGAGCCACTCCAp63 Exon3 revCCCACTGCAGAAAGCTGAGAp63 Exon4 forGATGGGTGGCTTTACTTGGGAp63 Exon4 revACACACCCTGGAACCTGTCTp73 forCTACTCACTGTCCAGGTGGCp73 revACAAGTAGTGGCCTGTTGGGPuma forTTCCTGGGTGGGAGTGACTTPuma revAGGGACTTCCCACTCGACTTBim forACGAAATGTAGACGTCCCGCBim revCCCACAGCCTTGAAACCGATFoxo1 forAACCAGTCCAACTCGACCACFoxo1 revAAGTTCCCAAACGAGCCCTGFoxo3a forGGAGAGAGCAAGAGCCCAAGFoxo3a revGACCCTCCCTTCCCACTTTGqRT-PCR Puma forGCCCAGCAGCACTTAGAGTCqRT-PCR Puma revGGTGTCGATGCTGCTCTTCTqRT-PCR L32 forTTAAGCGAAACTGGCGGAAACqRT-PCR L32 revTTGTTGGTCCCATAACCGATG\n\nCell culture and treatment {#Sec13}\n--------------------------\n\nHCT116 *p53*^\u2212/\u2212^, HCT116 *p53*^+/+^ (kindly provided by Bert Vogelstein) and HCT116 (ATCC) were maintained in DMEM containing 10% FCS and Penicillin/Streptomycin (P/S). FL5.12 and BaF3 were maintained in RPMI-1640 with 10% FCS, P/S and 1\u2009\u00b5g/l recombinant IL-3. To generate IL-2-dependent lymphoid cells, lymph node cells from three C57/Bl6 mice were isolated and activated by 20\u2009ng/ml PMA and 0.5\u2009\u00b5g/ml ionomycin for 48\u2009h in RPMI-1640 with 10% FSC and P/S. The cells were then cultured in the medium containing 100\u2009U/ml recombinant IL-2 for 24\u2009h and then used for experiments. For growth factor withdrawal, cells were washed twice with 50\u2009ml PBS and resuspended in RPMI-1640 with 10% FCS and P/S.\n\nFlow cytometry {#Sec14}\n--------------\n\nFor apoptosis quantification by flow cytometry, BaF3 or FL5.12 were deprived of IL-3 in the presence or absence of CT98014 (0.75\u2009\u00b5M) for 18\u2009h or as indicated. Activated lymphocytes were deprived of IL-2 in the presence or absence of CT98014 (0.75\u2009\u00b5M) for 22\u2009h. The cells were washed once with Annexin-V Binding Buffer (10\u2009mM HEPES, 150\u2009mM NaCl, 150\u2009\u00b5M MgCl, 2.5\u2009mM CaCl~2~) stained for 15\u2009min in the dark with Annexin-V-FITC or Annexin-V-APC in the same buffer. The fraction of Annexin-V-positive cells was measured using a FACS Calibur (BD Bioscience) or FACS LSRII (BD Bioscience).\n\nWestern blotting {#Sec15}\n----------------\n\nCells were harvested and washed with ice-cold PBS. Cells pellets were lysed for 5\u2009min on ice with lysis buffer (20\u2009mM Tris-HCl pH 7.5, 150\u2009mM NaCl, 1% Triton X-100, 1\u00d7 protease inhibitor cocktail (Roche), phosphatase inhibitor cocktail 1 (1:50, Sigma-Aldrich), MG132 (20\u2009\u00b5M, Alexis Biochemicals)) or by nuclear fractionation as described^[@CR6]^. Lysates were cleared by centrifugation at 16.100\u00d7*g* for 10\u2009min, 4\u2009\u00b0C. Protein concentration was determined using Bradford reagent (Bio-Rad). L\u00e4mmli buffer was added to 10--100\u2009\u00b5g of protein lysate and samples were boiled at 95\u2009\u00b0C for 5\u2009min. In some cases, the same amount of lysate was loaded on a second gel to probe for several antibodies. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were blocked in 3% dry milk in TBS-Tween20 (0.1%) (TBS-T) and then incubated at 4\u2009\u00b0C overnight with the primary antibody diluted in 3% dry milk in TBS-T. The membranes were washed three times with TBS-T, incubated for 1\u2009h with horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature and washed three times with TSB-T. Band visualization was achieved using Super Signal West Pico Chemiluminescent Substrate (Thermo Scientific, Waltham, USA) and the Fusion Solo imaging system (Vilber Lourmat, Eberhardzell, Germany).\n\nLuciferase reporter assay {#Sec16}\n-------------------------\n\n293T HEK cells in 24-well plates were transfected with a total amount of 1\u2009\u00b5g DNA by Attractene. A renilla luciferase expression plasmid (Promega, Madison, WI, USA) as internal control, a luciferase reporter plasmid driven by a 1.7\u2009kb Puma promotor region (\u22121200 to +500 from ATG) containing a conserved binding site for FOXO3A (CAAACAAT or mutated to CAGGGAAT)^[@CR25]^ and FOXO3A-TM (T32A, S253A, S315A) pcDNA3.1 expression plasmid were used in a 1:1:1 ratio. Empty vector was added as needed to keep the total DNA constant. Eight hours after transfection DMSO or CT98014 (0.75\u2009\u00b5M) were added. Cells were lysed using 100\u2009\u00b5l luciferase-lysis buffer (50\u2009mM Tris-phosphate pH 7.8, 250\u2009mM KCl, 10% Glycerol, 0.1% NP-40), lysates cleared by centrifugation at 16.100\u00d7*g* for 5\u2009min, 4\u2009\u00b0C. A volume of 10\u2009\u00b5l of the lysate was added in two black 96-well plates, renilla buffer (25\u2009mM Tris-phosphate, 100\u2009mM NaCl, 1\u2009mM EDTA, 0.05\u2009mM Coelenterazin) or luciferase buffer (25\u2009mM Tris-phosphate, 10\u2009mM MgSO~4~, 2\u2009mM ATP, 0.05\u2009mM Luciferin) were auto-injected and firefly and renilla luciferase activities were determined by a plate reader.\n\nImmunoprecipitation {#Sec17}\n-------------------\n\n293T HEK cells of a 78\u2009cm\u00b2 culture plate were transfected by PEI with pECE-FLAG-Foxo3a or pECE-FLAG-FOXO3A-TM (gifts from Michael Greenberg (Addgene plasmid \\# 8360 and \\# 8361)^[@CR26]^) and treated with CT98014 (0.75\u2009\u00b5M) 8\u2009h later. The next day, the cells were lysed in 1\u2009ml lysis buffer (20\u2009mM Tris-HCl pH 7.5, 150\u2009mM NaCl, 1% Triton X-100, 1\u00d7 protease inhibitor cocktail (Roche), phosphatase inhibitor cocktail 1 (1:50, Sigma-Aldrich), MG132 (20\u2009\u00b5M, Alexis Biochemicals)). A volume of 20\u2009\u00b5l were kept as input control and the remaining lysate was rotated for 2\u2009h with 20\u2009\u00b5l FLAG-M2 agarose affinity beads (Sigma Aldrich). Beads were washed three times with lysis buffer for 3\u2009min. Immunoprecipitates were eluted in 50\u2009\u00b5l lysis buffer with 3\u00d7 FLAG\u00ae peptide (150\u2009ng/\u00b5l, Sigma-Aldrich) by rotation for 30\u2009min at 4\u2009\u00b0C. Eluates were then subjected to western blotting.\n\nRT-PCR {#Sec18}\n------\n\nTotal RNA was extracted using Trizol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. 2\u2009\u00b5g RNA were transcribed into cDNA using the SuperScript First Strand Synthesis Kit (Invitrogen). Transcript levels of Puma and L32 were quantified by quantitative RT-PCR on an CFX96TM Real Time System thermocycler (Bio-Rad) with SYBR Green (Eurogentec, L\u00fcttich, Belgium). The following primers were used: mPuma: 5\u2032-GCCCAGCAGCACTTAGAGTC-3\u2032 and 5\u2032-GGTGTCGATGCTGCTCTTCT-3\u2032; mL32: 5\u2032-TTAAGCGAAACTGGCGGAAAC-3\u2032 and 5\u2032-TTGTTGGTCCCATAACCGATG-3\u2032.\n\nRe-expression of FOXO3A {#Sec19}\n-----------------------\n\nProduction of retroviral particles and infection were performed as described above for lentiviral particles, but using 1.5\u2009\u00b5g Hit60, 1.5\u2009\u00b5g pVSV-G (Clontech) and either 1.5\u2009\u00b5g pLXIN-hFOXO3A, plXIN-hFOXO3A-TM, or pLXIN-emtpy. Only BaF3 expressing lentiCRISPRv2 gRNA \\#1 targeting mouse *Foxo3a* were used, as human *Foxo3a* is also targeted by mouse *Foxo3a* gRNA \\#2. Selection was achieved with 50 ng/mL Geneticin\u2122 (ThermoFisher, Waltham, MA, USA). Experiments were performed immediately after successful selection.\n\nOverexpression of FOXO3A in HCT116 *p53*^\u2212/\u2212^ {#Sec20}\n---------------------------------------------\n\nHCT116 *p53*^\u2212/\u2212^ in six-well plates were transfected by Lipofectamine\u00ae 2000 with pECE-FLAG-FOXO3A or empty vector according to the manufacturer's protocol. Twenty-four hours after transfection, the cells were treated with GDC-0941 (10\u2009\u00b5M) for 5\u2009h and then subjected to western blotting.\n\nStatistical analysis {#Sec21}\n--------------------\n\nAll data are shown as mean\u2009\u00b1\u2009S.D. in case of technical replicates and as mean\u2009\u00b1\u200995% confidence interval in case of biological replicates. The GraphPad Prism 5 software (GraphPad Software Inc., La Jolla, CA, USA) was applied for statistical analysis using the one-way ANOVA with post hoc Tukey's multiple comparison test. *P*\u2009\u2264\u20090.05 was regarded as the threshold value for statistical significance. \\**P*\u2009\u2264\u20090.05; \\*\\**P*\u2009\u2264\u20090.01; \\*\\*\\**P*\u2009\u2264\u20090.001.\n\nElectronic supplementary material\n=================================\n\n {#Sec22}\n\nFig.S1 Fig.S2 Fig.S3 Fig.S4 Fig.S5 Supplementary Figure Legends\n\nEdited by G. Raschell\u00e0.\n\n**Electronic supplementary material**\n\n**Supplementary Information** accompanies this paper at 10.1038/s41419-018-0502-4.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nWe thank Martin Schuler for the p53-ER^TAM^ plasmid, Andreas Hecht for advice and discussions, Karin Neubert for excellent technical assistance and Celia Jakob for initial experiments. This study was supported by Grants Ma 1967/1 and Ma 1967/2 from the Deutsche Forschungsgemeinschaft to U.M., Grants 109199, 107397, and 112140 from the Deutsche Krebshilfe to U.M., by funding from the Centre for Biological Signalling Studies (BIOSS, EXC\u2010294), Freiburg, Germany to C.B. and U.M., and the Spemann Graduate School of Biology and Medicine (SGBM, GSC\u20104), Freiburg, Germany, to C.B. and U.M. both funded by the Excellence Initiative of the German Federal and State Governments, Germany.\n\nConflict of interest {#FPar1}\n====================\n\nThe authors declare that they have no conflict of interest.\n"}
+{"text": "At present, Iran has been known one of the up-warding countries in the world in regenerative medicine using stem cells therapy. In fact, the outcomes of some clinical trials on stem cell therapy of myocardial infarction, vitiligo, decompensated cirrhosis, and osteoarthritis narrate the feasibility of stem cell-based therapy for treatment of human diseases [@B1], [@B2]. However, in a similar manner with global configuration, the commercialization and translation of tissue engineering products into clinical phase has been restricted. It might be due to weak collaboration of different specialties for technology transfer of the multidisciplinary projects of tissue engineering field into clinical phase. Basic tissue engineers mostly prefer elegant studies, whereas physicians have tendency to solve medical problems with products indicating efficiency, easy to use, and cost benefit. Actually, a surgeon encountered with a dilemma between a partially effective tissue-engineered product that is both expensive and difficult to apply and a more traditional approach may choose the latter option. Therefore, a coherent teamwork between basic sciences and medicine as well as acquisition of competent knowledge about target tissue is necessary to conduct tissue engineering in the clinic. Moreover, it should be considered that in developing countries including Iran the high cost of high-tech biomedical research necessitates government investment [@B3]. Currently, the policy makers have established some action plans to support of science-based companies financially. This is a suitable opportunity to ligature basic research and market for commercialization of tissue engineering products. However, because private investors beyond academic laboratories should provide financing of tissue engineering products, incentive of private companies for investment should not be neglected. It is notable that tissue-engineered products will fulfill small market size unless they could indicate much superior results than competitive alternatives.\n\nIt is noticeable that tissue engineers should determine the requirements of community and develop strategies to penetrate the products into clinic. Indeed, the communication between scientists and policy makers should be increased to better definition of national research priorities. On the other hand, considering local necessities and natural resources should be rather than subjective experts' notions or international superiorities. Finally, ethical and legal regulations should be actually defined that indubitably make great profits to the society.\n"}
+{"text": "Y. Li, H. Neumann, M. Beller, *Chem. Eur. J.* **2020**, *26*, 6784.\n\nThe incorporation of (per)fluoroalkyl moieties into (hetero)arenes has been demonstrated to significantly improve their physical and biological properties for various applications.[1](#chem202001439-bib-0001){ref-type=\"ref\"}, [2](#chem202001439-bib-0002){ref-type=\"ref\"} Hence, in recent years there is an increasing interest to prepare such molecules by thermal[3](#chem202001439-bib-0003){ref-type=\"ref\"}, [4](#chem202001439-bib-0004){ref-type=\"ref\"} and photochemical reactions,[5](#chem202001439-bib-0005){ref-type=\"ref\"}, [6](#chem202001439-bib-0006){ref-type=\"ref\"}, [7](#chem202001439-bib-0007){ref-type=\"ref\"} as well as others.[8](#chem202001439-bib-0008){ref-type=\"ref\"}, [9](#chem202001439-bib-0009){ref-type=\"ref\"}, [10](#chem202001439-bib-0010){ref-type=\"ref\"}, [11](#chem202001439-bib-0011){ref-type=\"ref\"} Despite these notable achievements, still more practical and convenient methodologies are lacking and there is a need using sophisticated reagents and/or tedious purification due to selectivity problems. Indeed, regio\u2010 and chemoselective functionalizations are of crucial importance, because isolation of the resulting pure fluorinated isomers is often very difficult. To avoid these problems, synthetic methods have been developed for the synthesis of perfluoroalkyl arenes starting from the corresponding aryl halides.[10](#chem202001439-bib-0010){ref-type=\"ref\"} However, the direct preparation of the desired compounds from arenes via C\u2212H functionalization would be more desirable due to their availability and process step economy.\n\nAmong the different perfluoroalkylation reagents, especially the corresponding halides R~f~X constitute valuable substrates. Since the original work by Fuchikami and Ojima using perfluoroalkyl halides in the presence of copper bronze,[12](#chem202001439-bib-0012){ref-type=\"ref\"} several transition metal\u2010catalyzed perfluoroalkylations of arenes have been reported.[13](#chem202001439-bib-0013){ref-type=\"ref\"}, [14](#chem202001439-bib-0014){ref-type=\"ref\"}, [15](#chem202001439-bib-0015){ref-type=\"ref\"} However, in general products are obtained as mixtures, which were often not even isolated and purified. In this respect, the recent work of Zhao and co\u2010workers is noteworthy, who reported a *para*\u2010selective perfluoroalkylation of anilides in the presence of Mo(CO)~6~ as catalyst.[16](#chem202001439-bib-0016){ref-type=\"ref\"}\n\nBased on our interest in perfluoroalkylation reactions,[17](#chem202001439-bib-0017){ref-type=\"ref\"}, [18](#chem202001439-bib-0018){ref-type=\"ref\"} herein we present the first general and practical methodology for highly *ortho*\u2010selective, direct C\u2212H perfluoroalkylation of anilines under mild reaction conditions (Scheme\u2005[1](#chem202001439-fig-5001){ref-type=\"fig\"}). To the best of our knowledge, there have been no examples reported, which allow introducing *R* ~f~ groups with high site selectivity.\n\n![Selected methods for introducing trifluoromethyl and perfluoroalkyl groups into arenes.](CHEM-26-6784-g001){#chem202001439-fig-5001}\n\nRecently, our group reported platinum\u2010, nickel\u2010 and cobalt\u2010based catalysts for such reactions. However, in no case the regioselectivity problem could be resolved. Hence, we continued to search for more suitable transition metal catalysts. Following this strategy, we used the reaction of 4\u2010methoxyaniline (**1\u2009a**) and *n*\u2010C~4~F~9~I (**2\u2009a**) as a model system. In an initial screening of catalysts, ruthenium carbene complexes revealed to be active. In general, 5\u2005mol\u2009% of the catalyst and 2.0\u2005equiv of base (to trap HX) were used in organic solvents at 100\u2009\u00b0C for 12\u2005h. As shown in Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, optimization of reaction conditions, including pre\u2010catalysts, bases, and solvents gave the desired product 4\u2010methoxy\u20102\u2010(perfluorobutyl)\u2010aniline (**3\u2009a**) in \\>80\u2009% yield. ^19^F\u2005NMR investigations after the reaction clearly demonstrated that only the mono\u2010substituted aniline **3\u2009a** is formed. However, in the presence of an excess of **2\u2009a**, tiny amounts (\\<5\u2009%) of double perfluoroalkylation products can be formed.\n\n###### \n\nRu\u2010catalyzed perfluoroalkylation of 4\u2010methoxyaniline.^\\[a\\]^\n\n  ![](CHEM-26-6784-g007.jpg \"image\")                                                                  \n  ------------------------------------ ---------------------------- --------------- ----------------- --------------------\n  1                                    RuCl~3~ **\u22c5**3\u2009H~2~O         NaHCO~3~        1,4\u2010dioxane       6\n  2                                    \\[RuCl~2~(*p\u2010*Cymene)\\]~2~   NaHCO~3~        1,4\u2010dioxane       19\n  3                                    **Ru\u20101**                     NaHCO~3~        1,4\u2010dioxane       45\n  4                                    **Ru\u20102**                     NaHCO~3~        1,4\u2010dioxane       36\n  5                                    **Ru\u20103**                     NaHCO~3~        1,4\u2010dioxane       31\n  6                                    **Ru\u20104**                     NaHCO~3~        1,4\u2010dioxane       26\n  7                                    **Ru\u20105**                     NaHCO~3~        1,4\u2010dioxane       19\n  8^\\[c\\]^                             **Ru\u20105**                     NaHCO~3~        1,4\u2010dioxane       38\n  9^\\[c\\]^                             RuCl~3~ **\u22c5**3\u2009H~2~O         NaHCO~3~        1,4\u2010dioxane       41\n  10                                   **Ru\u20101**                     KHCO~3~         1,4\u2010dioxane       51\n  11                                   **Ru\u20101**                     K~2~CO~3~       1,4\u2010dioxane       72\n  12                                   **Ru\u20101**                     Na~2~CO~3~      1,4\u2010dioxane       58\n  13                                   **Ru\u20101**                     Na~2~HPO~4~     1,4\u2010dioxane       16\n  14                                   **Ru\u20101**                     Et~3~N          1,4\u2010dioxane       22\n  15                                   **Ru\u20101**                     K~2~CO~3~       THF               49\n  16                                   **Ru\u20101**                     K~2~CO~3~       MeCN              23\n  **17**                               **Ru\u20101**                     K~2~CO~3~       MePh              35\n  18                                   **Ru\u20101**                     K~2~CO~3~       DMF               14\n  19^**\\[d\\]**^                        **Ru\u20101**                     **K~2~CO~3~**   **1,4\u2010dioxane**   **83 (79)^\\[e\\]^**\n  20                                   RuCl~3~ **\u22c5**3\u2009H~2~O         K~2~CO~3~       1,4\u2010dioxane       73\n  21                                   free                         K~2~CO~3~       1,4\u2010dioxane       8\n\n\\[a\\]\u2005Reaction conditions: **1\u2009a** (1.0\u2005mmol), **2** (1.2\u2005mmol), catalyst (5\u2005mol\u2009%), base (2.0\u2005mmol), solvent (2.0\u2005mL), 100\u2009\u00b0C, 12\u2005h. \\[b\\]\u2005Determined by ^19^F\u2005NMR analysis using (trifluoromethoxy)benzene as internal standard. \\[c\\]\u20057\u2005mol\u2009% NHC ligand was added. \\[d\\]\u2005**2** (1.3\u2005mmol). \\[e\\]\u2005Isolated yield in brackets. Structure of catalysts and ligands:\n\nWiley\u2010VCH Verlag GmbH & Co. KGaA\n\nAmong the different ruthenium salts and complexes applied, a variety of defined metathesis catalysts proved to be effective for this transformation. In line with this observation, adding 1,3\u2010bis(2,6\u2010diisopropylphenyl)\u2010imidazolium bromide as NHC carbene ligand to simple ruthenium trichloride led to a reasonable active catalyst system (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entries\u20058--9). Nevertheless, commercially available **Ru\u20101** provided the highest product yield. Thus, in the presence of this precursor the influences of base (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entries\u200510--14) and solvents (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entries\u200515--18) including THF, MeCN, MePh and DMF were studied. Best results (86\u2009% of **3\u2009a**) were obtained using K~2~CO~3~ in 1,4\u2010dioxane with an increased amount of **2\u2009a** (1.3\u2005mmol; Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entry\u200519).\n\nObviously, the stoichiometry of the perfluoroalkylation reagent and the choice of base (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entries\u200510--11) have the major influence on the efficiency of this reaction. Thus, a control experiment using simple RuCl~3~\u22c5H~2~O (5\u2005mol\u2009%) in the presence of the carbene ligand (10\u2005mol\u2009%) and K~2~CO~3~ as base was performed using 1.3\u2005equiv of R~f~I. Indeed, the desired product **3\u2009a** is obtained in 73\u2009% yield (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entry\u200520). As expected without the metal catalyst, the background reaction occurred only to a negligible extent (Table\u2005[1](#chem202001439-tbl-0001){ref-type=\"table\"}, entry\u200521).\n\nRegarding the mechanism, we propose the initial formation of a perfluoroalkyl radical via a metal\u2010catalyzed single electron transfer process (SET) is occurring. Such SET processes are well known for several ruthenium complexes.[19](#chem202001439-bib-0019){ref-type=\"ref\"} The excellent regioselectivity observed in the model reaction should be a result of the coordination of aniline to the metal center, which determines preferential formation of the *ortho*\u2010perfluoroalkylated product **2\u2009a**. To understand this selectivity and the underlying reaction mechanism, several control experiments were carried out under standard conditions (see SI, Table\u2005S1). In contrast to aniline, phenol and anisole do not react under identical conditions, which highlights the importance of the amino group for this transformation.\n\nIn these latter cases the starting materials were retained in \\>98\u2009%. Interestingly, the use of *N*,*N*\u2010dimethylaniline led to a mixture of perfluoroalkylated products, which could not be isolated in pure form due to the similar physical properties of the regioisomers. To isolate and/or identify any organometallic intermediate of the catalytic cycle, the complex **Ru\u20101** (0.1\u2005mmol) was mixed with n\u2010C~4~F~9~I (0.1\u2005mmol) in 1,4\u2010dioxane (1\u2005mL) and stirred at 100\u2009\u00b0C for several hours. Analysis of the crude mixture by ^19^F\u2005NMR revealed unfortunately no obvious changes (see SI for more details).\n\nNext, we were interested to explore the substrate scope of anilines of this novel methodology. Thus, reactions of various substituted anilines with *n*\u2010C~4~F~9~I were examined. Based on the optimization vide supra, the following conditions were applied: 1.3\u2005equiv of **2**, 5\u2005mol\u2009% **Ru\u20101**, 2.0\u2005equiv of K~2~CO~3~ in 1,4\u2010dioxane (2.0\u2005mL) at 100\u2009\u00b0C for 12\u2005h under N~2~. Initially, we focused on the functional group tolerance of the catalyst. For this purpose, several electron\u2010rich *para*\u2010substituted anilines (\u2010OMe, \u2010SMe, \u2010OEt, \u2010OPh, \u2010OH, \u2010*t*Bu, \u2010Me) were reacted with *n*\u2010C~4~F~9~I to afford smoothly the corresponding products in good to excellent yields (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009a**--**3\u2009g**). Apart from *n*\u2010C~4~F~9~I, also using *n*\u2010C~4~F~9~Br gave the target compounds, albeit in somewhat lower yields. In addition, anilines with electron\u2010withdrawing groups (\u2010Cl, \u2010Br, \u2010Ac) provided the perfluoroalkylated products in moderate to good yields (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009h**--**3\u2009j**). Notably, halide and ketone substituents on the aniline ring remained untouched demonstrating the excellent chemo\u2010selectivity of the system. In general, the reactivity of the latter substrates is lower than those of electron\u2010rich anilines. Next, we investigated reactions of aniline, 1\u2010naphthylamine, 1,3\u2010disubstituted, and 1,3,5\u2010trisubstituted anilines. In all these cases, the control of regioselectivity is significantly more challenging. Nevertheless, using parent aniline **1\u2009k** as substrate perfluoroalkylation occurred specifically in the *ortho*\u2010position giving **3\u2009k** in 73\u2009% isolated yield. Other isomers (\\<5\u2009%) could not be observed. Next, anilines with *meta*\u2010substitution pattern were tested in this process. Again, both electron\u2010rich (\u2010OMe) and \u2010withdrawing (\u2010CN, \u2010NO~2~) substrates expressed good activities for the perfluoroalkylation reaction and only one regioisomer was obtained (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009l**--**3\u2009n**). In addition, anilines with two substituents, such as 3,5\u2010dimethyl and 3,5\u2010dimethoxy can be conveniently employed in this reaction with high site\u2010selectivity (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009o**, **3\u2009p**). Finally, when using 2,3\u2010dihydrobenzo\\[b\\]\\[1,4\\]dioxin\u20106\u2010amine **1\u2009q** we obtained two *ortho*\u2010substituted regioisomers in 49 and 22\u2009% yield (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009q**, **3\u2009q\u2032**), respectively. Similarly to the model system, also for these substrates *n*\u2010C~4~F~9~Br could be successfully used as perfluoroalkyl source and the *ortho*\u2010perfluoroalkylated anilines were obtained in moderate to good yields (Scheme\u2005[2](#chem202001439-fig-5002){ref-type=\"fig\"}, **3\u2009k**, **3\u2009l**, **3\u2009p**). In all these cases the perfluoroalkylated products are easily obtained in pure form due to the high selectivity of the process.\n\n![Selective Ru\u2010catalyzed perfluoroalkylation of anilines. \\[a\\]\u2005Reaction conditions: aniline (1.0\u2005mmol), *n*\u2010C~4~F~9~I (1.3\u2005mmol), **Ru\u20101** (5\u2005mol\u2009%), K~2~CO~3~ (2.0\u2005equiv). and 1,4\u2010dioxane (2.0\u2005mL), 100\u2009\u00b0C, 12\u2005h. \\[b\\]\u2005Isolated yield. \\[c\\]\u2005*n*\u2010C~4~F~9~Br (1.5\u2005mmol) used for the reaction in brackets.](CHEM-26-6784-g002){#chem202001439-fig-5002}\n\nTo demonstrate the synthetic utility of our protocol on multi\u2010g\u2010scale, the reaction of 4\u2010methoxyaniline (10.0\u2005mmol) with *n*\u2010C~4~F~9~I (13.0\u2005mmol) was performed and gave the corresponding product only in slightly lower yield (Scheme\u2005[3](#chem202001439-fig-5003){ref-type=\"fig\"}).\n\n![Gram\u2010scale synthesis of **3\u2009a**.](CHEM-26-6784-g003){#chem202001439-fig-5003}\n\nIndoles are an important class of heterocyclic compounds. Indeed, numerous natural products contain this electron\u2010rich scaffold. Thus, we explored the reactivity of **4\u2009a**--**c** with **2\u2009a** in this process. We were pleasured to find that selective perfluorobutylation occurred selectively at the C\u20102 position with or without N protection (Scheme\u2005[4](#chem202001439-fig-5004){ref-type=\"fig\"}).\n\n![Ru\u2010catalyzed perfluoroalkylation of indoles.](CHEM-26-6784-g004){#chem202001439-fig-5004}\n\nDuring the initial optimization of this procedure, we found small amounts of double perfluoroalkylation. Following this original observation, we investigated the reaction of aniline **1\u2009a** with a larger excess of **2\u2009a**. Surprisingly, the di\u2010*ortho*\u2010perfluoroalkylated aniline **5\u2009a** is obtained in 55\u2009% isolated yield using 3.0\u2005equiv of **2\u2009a**. As shown in Scheme\u2005[5](#chem202001439-fig-5005){ref-type=\"fig\"}, other anilines **1\u2009b**--**d**, **q** presented similar reactivity and gave the desired *ortho*/*ortho*\u2010disubstituted compounds in 53--66\u2009% isolated yield. It should be noted that such 2,6\u2010difunctionalized anilines are otherwise very difficult to access.\n\n![Ru\u2010catalyzed double\u2010perfluoroalkylation of anilines. \\[a\\]\u2005Reaction conditions: aniline (1.0\u2005mmol), *n*\u2010C~4~F~9~I (3.0\u2005mmol), **Ru\u20101** (5\u2005mol\u2009%), K~2~CO~3~ (2.0\u2005equiv.) and 1,4\u2010dioxane (2.0\u2005mL), 100\u2009\u00b0C, 12\u2005h. \\[b\\]\u2005Isolated yield.](CHEM-26-6784-g005){#chem202001439-fig-5005}\n\nRegarding the various perfluoroalkylation reactions, certainly trifluoromethylation is the most important transformation, especially for the synthesis of new bio\u2010active compounds. Hence, diverse state\u2010of\u2010the\u2010art reagents for laboratory scale synthesis including Umemoto\\'s,[20](#chem202001439-bib-0020){ref-type=\"ref\"} Togni\\'s,[21](#chem202001439-bib-0021){ref-type=\"ref\"} Langlois\\',[22](#chem202001439-bib-0022){ref-type=\"ref\"} and Ruppert\u2010Prakash\\'s reagent[23](#chem202001439-bib-0023){ref-type=\"ref\"} as well as CuCF~3~ [24](#chem202001439-bib-0024){ref-type=\"ref\"} were developed in the past decades. However, due to their price and availability, the use of these reagents on \\>kg\u2010scale is highly problematic. In contrast, CF~3~Br is available on multi\u2010100\u2005kg\u2010scale and less expensive compared to the above\u2010mentioned reagents. Nevertheless, we want to make clear that also this reagent has ozone depleting properties, which are forbidden by the Montreal protocol. Hence, appropriate safety measures should be taken. As shown in Scheme\u2005[6](#chem202001439-fig-5006){ref-type=\"fig\"}, CF~3~Br allows for *ortho*\u2010selective functionalization of *para*\u2010substituted anilines with acceptable yields (Scheme\u2005[6](#chem202001439-fig-5006){ref-type=\"fig\"}, **6\u2009a**--**6\u2009f**). Similarly, using **Ru\u20101** other perfluoroalkyl iodides (C~2~F~5~I, C~3~F~7~I, C~6~F~13~I, C~8~F~17~I, and C~10~F~21~I) reacted well with 4\u2010methoxyaniline (**1\u2009a**) to give the corresponding target compounds in moderate to good yields and excellent selectivities (Scheme\u2005[6](#chem202001439-fig-5006){ref-type=\"fig\"}, **6\u2009g**--**6\u2009l**). In the context of building blocks for bio\u2010active compounds, it is worth mentioning that heptafluoroisopropyl iodide **2\u2009i** afforded the corresponding product in 63\u2009% yield (Scheme\u2005[6](#chem202001439-fig-5006){ref-type=\"fig\"}, **6\u2009i**).\n\n![Substrate scope of trifluoromethyl and perfluoroalkyl halides. \\[a\\]\u2005Reaction conditions: **1** (1.0\u2005mmol), R~*f*~X (1.3\u2005mmol), **Ru\u20101** (5\u2005mol\u2009%), K~2~CO~3~ (2.0\u2005equiv). and 1,4\u2010dioxane (2.0\u2005mL), 100\u2009\u00b0C 12\u2005h. \\[b\\]\u20054\u2005mL solution of CF~3~Br in 1,4\u2010dioxane (0.06\u2005mol\u2009L^\u22121^) was used, and yield based on CF~3~Br.](CHEM-26-6784-g006){#chem202001439-fig-5006}\n\nIn conclusion, we have developed an efficient and practical *ortho*\u2010selective mono\u2010 and di(per)fluoroalkylation methodology of free anilines and indoles. In the presence of the stable and commercially available pre\u2010catalyst **Ru\u20101** various *meta*\u2010 and *para*\u2010substituted anilines, including mono\u2010 and disubstituted ones, gave the corresponding products without the need for protecting groups in high purity. The availability of the starting materials and the high selectivity make the process attractive for the synthesis of all kinds of perfluoroalkyl\u2010substituted anilines, including trifluormethyl derivatives.\n\nExperimental Section {#chem202001439-sec-0002}\n====================\n\nPerfluoroalkylation of anilines with *n*\u2010C~4~F~9~I: aniline (1.0\u2005mmol), *n*\u2010C~4~F~9~I (1.3\u2005mmol), **Ru\u20101** (5\u2005mol\u2009%), K~2~CO~3~ (2.0\u2005equiv). and 1,4\u2010dioxane (2.0\u2005mL) were added to a reaction tube. Then, the tube was degassed with argon three times and heated at 100\u2009\u00b0C for 12\u2005h. After cooling to room temperature, the solvent was removed under vacuum conditions, and the products were purified through silica gel chromatography by using hexane and ethyl acetate as the eluents.\n\nConflict of interest {#chem202001439-sec-0004}\n====================\n\nThe authors declare no conflict of interest.\n\nSupporting information\n======================\n\nAs a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re\u2010organized for online delivery, but are not copy\u2010edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.\n\n###### \n\nSupplementary\n\n###### \n\nClick here for additional data file.\n\nWe are grateful for the financial support from the BMBF and the State of Mecklenburg\u2010Western Pomerania, Germany. Y.L. would like to thank the Chinese Scholarship Council, China for funding.\n"}
+{"text": "Introduction {#S0001}\n============\n\nPharmacy is the health profession that is committed to ensuring safe and effective use of medication and it links the health sciences with the basic sciences.[@CIT0001] Pharmacy is among the respected professions worldwide, and it has experienced tremendous developments over the years.[@CIT0002]\n\nNowadays pharmacists' scope of practice widened from previous dispensing and suppling of medications to a variety of health services straddling prevention, addressing patient needs and expectations. All transition of focus is related to change of health care system, aging populations, health care workforce supply issues, emerging chronic diseases and changing expectations have re-engineered pharmacy in many developed nations.[@CIT0003],[@CIT0004]\n\nFor advancing pharmacy knowledge in a specific field or area of interest, postgraduate education (PGE) opportunities are the major options for pharmacy students and graduates. Often PGE is needed as a prerequisite for certain employment opportunities as well as to diversify pharmacists' skills and knowledge within the health care environment.[@CIT0005] Postgraduate education in Ethiopia is popular and it enhances job prospects. Also, there is Higher Education Relevance and Quality Agency (HERQA) which ensures whether higher education curriculum supports the countries' development needs. Postgraduate Master's programs under the school of pharmacy in Ethiopia include Clinical Pharmacy, Pharmacology, Pharmacognosy, Medicinal chemistry, Pharmaceutical Analysis, Social pharmacy and administrative pharmacy, and Pharmaceutics. The programs are to be completed within two years, one year is for theory and the other is for research. The theory part of the programs has supplementary courses for the development of academic research and prepare graduates to become an independent researcher.\n\nChoosing pharmacy as an area of specialization after high school graduation may have an impact on following PGE or future areas of the students. The study done by Abdelhadi Nadine et al reported that their choice is related to their desire to work with patients, their desire of profession with a high level of income, and others attracted with its flexible working hours.[@CIT0006] There have been other influential factors which have been reported to affect their choice including their relatives, friends, or teachers, or through career fairs.[@CIT0007] Difference in the selection of pharmacy profession as the first choice between developed and developing countries probably linked to low public awareness and appreciation of pharmacy as a valid health care profession in developing nations.[@CIT0008]--[@CIT0010] A review of available literature suggests that the proportion of students who selected pharmacy as their first choice of study ranged from 39--51.1% in Africa[@CIT0011],[@CIT0012] and 71.5% to 77.4% in developed countries, UK and the United States.[@CIT0007],[@CIT0013]\n\nFurther studies were done to identify influential factors for future engagement in research and academia. Introducing online courses and preparing summer programs has shown remarkable outcomes for students' future choices. It has been described that these practices have the potential to open the eyes of students to a new window and may show a clear picture of available options.[@CIT0014],[@CIT0015]\n\nStudent engagement represents the amount of time and effort students devote to their studies as well as the resource allocated by the institution to promote participation in educational activities, has been recommended to help future motivation on learning environment.[@CIT0016]\n\nGlobally, studies have investigated the reasons and motivations of pharmacy students to study pharmacy[@CIT0006],[@CIT0017] as well as their plans.[@CIT0018],[@CIT0019] In Ethiopia, there was no research done on what motivates pharmacy students to study pharmacy as well as pursuing postgraduate education and factor associated. This study aims to assess Pharmacy student's motivation to study postgraduate programs and potential barriers to postgraduate education.\n\nMethods and Materials {#S0002}\n=====================\n\nStudy Area, Design and Period {#S0002-S2001}\n-----------------------------\n\nInstitutional-based cross-sectional study was conducted on undergraduate regular pharmacy students of the University of Gondar. The study was conducted from February 1--30, 2019.\n\nInclusion and Exclusion Criteria {#S0002-S2002}\n--------------------------------\n\nAll volunteer undergraduate regular pharmacy students attending the bachelor of the pharmacy degree program at the University of Gondar were included in the study.\n\nPopulation {#S0002-S2003}\n----------\n\nUndergraduate regular pharmacy students who were attending the bachelor of pharmacy degree program were the study population.\n\nStudy Variables {#S0002-S2004}\n---------------\n\n### Dependent Variables {#S0002-S2004-S3001}\n\nChoice of pharmacy as first choice, interest for PGE, previous experience in research, pharmacy-related work experience, previous leadership position, and previous opportunity to attend a conference related to pharmacy school were dependent variables.\n\n### Independent {#S0002-S2004-S3002}\n\nSocio-demographic variables and other characteristics: age, sex, religion, region, marital status, level of study, and GPA.\n\nData Collection and Management {#S0002-S2005}\n------------------------------\n\nA structured questionnaire[@CIT0020] for this study was adapted and developed based on previous studies.[@CIT0021],[@CIT0022] The questionnaire involved four main sections. The first part includes Socio-demographic information of students includes age, sex, year of study, religion, and marital status. The second section wanted to assess a student's motivation and desire to study pharmacy as a first-choice career option. We asked whether pharmacy was the first-choice field of study using a \"Yes or No\" option. If yes, they were requested to choose from a list of factors that they think inclined their decision the most. The third section of the questionnaire assessed previous experiences of students, by asking them to choose from a list of areas of pharmacy practice. Information like research experience, pharmacy-related work experience, and factors influence interest for PGE were also collected. After explaining the purpose of the study, questionnaires were self-administered to volunteer students. They were allowed to fill the questionnaire within 15 to --20 min and their response was kept confidential.\n\nData Quality Assurance and Analysis {#S0002-S2006}\n-----------------------------------\n\nThe pretest was done before the actual data collection on 15 volunteer pharmacy students who were not considered in the final analysis. Content validity and reliability were checked with senior professionals. Content validity was checked by reviewing literature and preparing senior professionals' panels. The reliability test was done through Cronbach's alpha, and the score was 7.4. Necessary Correction and modification were instituted based on feedback before commencing the final survey collection. The collected data were cleared, arranged, coded, checked for completeness, and then analyzed by SPSS version 21. Descriptive statistics were used to summarize socio-demographic and other baseline characteristics. Binary-Logistic regression analysis was used to determine the association between variables and P-values were reported as statistically significant if \\<0.05 or 5%.\n\nResult {#S0003}\n======\n\nSocio-Demographic Characteristics of Study Subjects {#S0003-S2001}\n---------------------------------------------------\n\nThe questionnaire was distributed to all undergraduate regular pharmacy students (n=440) and there was a total of 388 pharmacy students who completed the survey (response rate =88.2%). Around two-third of 244 (62.9%) of participants were interested to pursue their postgraduate education. The majority of participants were male 215 (55.4%) and in the age group 22--25, 204 (52.6%). Students studying the first and second years are more interested in postgraduate programs, 60 (24.6%) and 56 (23.9%), respectively ([Table 1](#T0001){ref-type=\"table\"}).Table 1Demographics and Background Characteristics of RespondentsVariablesInterest for PGEAOR95% CIp-valueYes,\\\nN (%)\\\n244 (62.9)No,\\\nN (%)\\\n144 (37.1)Age18--21124 (50.8)50 (34.7)2.105\\[0.526--8.429\\]0.31522--25113 (46.3)91 (63.2)2.623\\[0.681--10.099\\]26--307 (2.9)3 (2.1)11SexMale124 (50.8)91 (63.2)1.436\\[0.835--2.468\\]0.190Female120 (49.2)53 (36.8)11ReligionOrthodox195 (79.9)112 (77.8)2.679\\[0.855--8.391\\]0.184Muslim27 (11.1)25 (17.4)3.041\\[0.901--10.260\\]Protestant22 (9.0)7 (4.9)11Marital StatusSingle238 (97.5)140 (97.2)2.589\\[0.87--35.855\\]0.181Married4 (1.6)3 (2.1)13.823\\[0.539--354.794\\]Divorced2 (0.8)1 (0.7)11Level of study1st Year60 (24.6)17 (11.8)0.274\\[0.100--0.751\\]0.0612nd Year56 (23.0)25 (17.4)0.334\\[0.134--0.832\\]3rd Year51 (20.9)40 (27.8)0.689\\[0.298--1.597\\]4th Year47 (19.3)33 (22.9)0.597\\[0.257--1.386\\]5th Year30 (12.3)29 (20.1)11RegionAmhara108 (44.3)69 (47.9)0.251\\[0.078--0.808\\]0.007\\*Tigray13 (5.3)10 (6.9)0.167\\[0.036--0.783\\]Oromia26 (10.7)18 (12.5)0.574\\[0.153--2.150\\]Somalia2 (0.8)3 (2.1)3.212\\[0.209--49.469\\]SNNP16 (6.6)14 (9.7)0.377\\[0.094--1.504\\]Addis Ababa70 (28.7)21 (14.6)0.141\\[0.041--0.485\\]Dire Dawa9 (3.7)9 (6.3)11GPA1.1--25 (2.0)0 (0.0)0.0000.0000.9962.1--3107 (43.9)67 (46.5)0.976\\[0.583--1.636\\]3.1--4132 (54.1)77 (53.5)11Was pharmacy your number one choice?Yes165 (67.6)118 (81.9)3.557\\[1.906--6.637\\]\\<0.001\\*\\*No79 (32.4)26 (18.1)11Have You ever been participated in the research?Yes87 (35.7)31 (21.5)0.551\\[0.275--1.107\\]0.094No157 (64.3)113 (78.5)11Have you ever been in a leadership position?Yes86 (35.2)45 (31.3)2.864\\[1.454--5.642\\]0.002\\*\\*No158 (64.8)99 (68.8)11Do you have pharmacy related work experience?Yes89 (36.5)28 (19.4)0.320\\[0.165--0.620\\]\\<0.001\\*No155 (63.5)116 (80.6)11Have you ever got an opportunity to attend a conference related to pharmacy school?Yes101 (41.4)26 (18.1)0.181\\[0.093--0.354\\]\\<0.001\\*\\*No143 (58.6)118 (81.9)11[^1]\n\nFactors Affecting Interest in Postgraduate Education {#S0003-S2002}\n----------------------------------------------------\n\nThe region, pharmacy as a first choice, experience in a leadership position, previous pharmacy-related work, and previous attendance in the conference have shown statistically significant association for the interest of postgraduate education with p-value 0.007, \\<0.001, 0.002, 0.001, and \\<0.001, respectively ([Table 1](#T0001){ref-type=\"table\"}).\n\nThose who have pharmacy-related work experience and who have got an opportunity to attend a conference related to pharmacy school have decreased interest for postgraduate education 68%, AOR: 0.32 \\[0.165--0.620\\] ([Figure 1](#F0001){ref-type=\"fig\"}) and 81.9%, AOR: 0.181\\[0.093--0.354\\] ([Figure 2](#F0002){ref-type=\"fig\"}), respectively ([Table 1](#T0001){ref-type=\"table\"}).Figure 1Influence of previous pharmacy-related work experience on interest for PGE.Figure 2Influence of previous exposure on pharmacy-related conference on interest for PGE.\n\nOn the other hand, those participants who are from Somalia region, who choose pharmacy as number one choice and who have experience in a leadership position were interested in postgraduate program AOR: 3.212\\[0.209--49.469\\], AOR: 3.557\\[1.906--6.637\\] ([Figure 3](#F0003){ref-type=\"fig\"}), and AOR:2.864 \\[1.454--5.642\\] ([Figure 4](#F0004){ref-type=\"fig\"}), respectively ([Table 1](#T0001){ref-type=\"table\"}).Figure 3Influence of choosing pharmacy as a first choice on interest for PGE.Figure 4Influence of previous leadership experience on interest for PGE.\n\nPrevious experience in research does not show a significant association with interest for PGE (p-0.094) ([Figure 5](#F0005){ref-type=\"fig\"}).Figure 5Influence of previous participation in research on interest for PGE.\n\nReasons for Choosing Pharmacy as a First Choice, Pharmacy-Related Work Experience and Confidence for Application and Focus of Interest {#S0003-S2003}\n--------------------------------------------------------------------------------------------------------------------------------------\n\nAmong the reasons for choosing pharmacy as a first choice, family's/relatives/friend's encouragement and university perspective are the most selected with 110 (38.9%) and 49 (17.3%), respectively ([Table 2](#T0002){ref-type=\"table\"}). More than half 70 (59.8%) of participants have experience in community pharmacy followed by hospital pharmacy 38 (32.5%) ([Table 3](#T0003){ref-type=\"table\"}). Among postgraduate interested students nearly half of them interested to study clinical pharmacy108 (44.3%) followed by pharmacology 38 (15.6%) and social pharmacy 30 (12.3%). Most of them rate their level of preparation for the program 66 (27%) and level of confidence for the application process110 (45.1%) at good ([Table 4](#T0004){ref-type=\"table\"}).Table 2Reason for Choosing Pharmacy as a First Choice (n=283)VariablesN (%)A subject teacher at school/college39 (13.8)University prospective49 (17.3)Radio or TV program about the pharmacy profession32 (11.3)My family/relative/friend encouraged me to choose the pharmacy110 (38.9)I was influenced by a pharmacist I know, as a role model24 (8.4)I was influenced by pharmacy work experience27 (9.5)Someone in my family who owns a pharmacy influences me0 (0)I wanted to work in a well-respected profession23 (8.1)I wanted the job with good career opportunities41 (14.5)I wanted a job where I am socially useful34 (12.0)I wanted the opportunity for self-employment31 (11)I wanted a profession where you can always get a job.37 (13.1)I was good at science in school so I wanted to have a carrier in the health care profession where I can work with patients16 (5.7)I wanted the opportunity for part-time work25 (8.8) Table 3Pharmacy-Related Experience (n=117)N (%)Community pharmacy70 (59.8)Hospital pharmacy38 (32.5)Clinical pharmacy9 (7.7)Clinical research0 (0)Regulation and academic activity0 (0)Manufacturing0 (0) Table 4The Focus of Interest, and Level of Preparation and Confidence for Application (n=244)Interest in Different Types of PGE ProgramsN (%)Clinical pharmacy108 (44.3)Pharmacology38 (15.6)Analytical chemistry15 (6.15)Medicinal chemistry13 (5.3)Pharmaceutics27 (11.1)Social pharmacy30 (12.3)Pharmacoepidemiology9 (3.4)Forensic pharmacy1 (0.4)Industrial pharmacy3 (1.2)Level of preparation for the postgraduate program\u2003Poor25 (10.2)\u2003Average64 (26.2)\u2003Neutral46 (18.9)\u2003Good66 (27)\u2003Very good43 (17.6)Level of confidence for the application process\u2003Poor22 (9)\u2003Average53 (21.7)\u2003Neutral38 (15.6)\u2003Good110 (45.1)\u2003Very good33 (13.5)\n\nReasons for Not Pursuing a Postgraduate Program {#S0003-S2004}\n-----------------------------------------------\n\nAmong the reasons for not pursuing a postgraduate program the most selected one is lack of information about the application and interview process 44 (30.6%) followed by financial obligation 39 (27.1%) and not feeling prepared 20 (13.9%) ([Table 5](#T0005){ref-type=\"table\"}).Table 5Reasons for Not Pursuing Postgraduate Education (n=144)ReasonsN (%)Financial obligations39 (27.1)Family obligations2 (1.4)Not feeling prepared20 (13.9)Lack of information about the application and interview process44 (30.6)No contacts to PGE personnel12 (8.3)Geographical limitations8 (5.6)The belief that grades were not good enough to apply7 (4.9)Other graduation options were stressed2 (1.4)Afraid of the competition2 (1.4)The disappointment of not getting a PGE0 (0)Advised that PGE are NOT necessary0 (0)Information on options for PGE are not readily available0 (0)The emphasis of these programs did not exist in school; not felt to be important6 (4.2)Deadlines were too soon to apply to PGE2 (1.4)\n\nDiscussion {#S0004}\n==========\n\nThis study was conducted to assess the motivation of pharmacy students for postgraduate programs and associated factors that affect their motivation.\n\nAmong all participants, more than two-third 283 (72.9%) choose pharmacy as a first choice. Similarly, on studies done in South Jordan and South Africa, the pharmacy has been selected by nearly half of the students as a first choice.[@CIT0007],[@CIT0017] The increased number in our study may be related to the inclusion of students who choose pharmacy other than other health sciences after losing the score which can make them a candidate to enter medicine.\n\nThe most selected influential factor to choose pharmacy as the first choice was family's/relatives/friend's encouragement 114 (38%). In line with this, another study also shows peer and family influencing a career path of students.[@CIT0007]\n\nIn our study, the region was significantly associated with interest for PGE, those who come from Somalia region have shown increased interest for PGE (p-0.007). Somali region was performing the worse education than other regions and students coming from Somalia region usually are the most top scorers from the region or they are from a more supportive community[@CIT0023] and commonly they are few (between 0--2 in each batch). So, their interest in PGE may be related to their passion to increase regional educational structure.\n\nIn the study, those who choose pharmacy as their number one choice were more interested to follow PGE (p-0.001). This may be linked to the previous awareness about the pharmacy-related courses and previously settled goals which may have a great impact on following PGE. Additionally, a Previous leadership position has been observed to have a positive impact on students to pursue PGE (p-0.002). Motivation is typically believed to be affected by a combination of individual and contextual factors. Here, previous leadership experience may be a powerful contextual factor for their motivation to study PGE.[@CIT0024]\n\nPrevious pharmacy-related work was among the factors that were associated with decreased interest for PGE (p-0.001). As we observed on the result part most of the participants have community pharmacy 70 (59.8%) and hospital pharmacy38 (32.5%) work experience which can attract participants to the market place or business, and students practice site may determine which practice settings they might like to enter in the future.[@CIT0025]\n\nFurthermore, those students who attend pharmacy-related conferences have shown decreased interest for PGE (p-0.001).On the contrary, the study conducted in the USA has shown, increased interest of students who have ever attended pharmacy-related conferences.[@CIT0022] This may be because, on the conference, they may get informed about the challenges they will have on following their PGE including financial requirements and country (region) of the institution where PGE will be followed. Following PGE in the higher institution is more difficult for students who did not get any scholarship opportunity to cover their financial requirements.[@CIT0026]\n\nAmong postgraduate interested students nearly half 108 (44.3%) of them have been interested to study clinical pharmacy. Another study conducted in Ethiopia shows student's interest in hospital pharmacy.[@CIT0012] Furthermore, a study conducted in the United Arab Emirates shows similar interest, and students\\' passion for helping people and being interested in science were observed as key motivators.[@CIT0027]\n\nMost of them rate their level of preparation for the program 66 (27%) and level of confidence for the application process110 (45.1%) at good. Providing information about PGE may have an impact to increase the level of confidence on preparation as well as on application process for PGE.[@CIT0028],[@CIT0029]\n\nAmong the reasons for not pursuing a postgraduate program the most depicted was lack of information about the application and interview process 44 (30.6%). On the other hand, financial obligation 39 (27.1%) was the second reason for not pursuing a PGE which was also the major reason in other studies.[@CIT0022],[@CIT0030]\n\nStudy Limitations {#S0004-S2001}\n-----------------\n\nAs the study is done only in one institute, the results of this study may not be generalizable to other areas of the country. Besides, the survey fails to deeply assess the reason for student's interest and non-interest for PGE.\n\nConclusion {#S0005}\n==========\n\nGenerally, PGE is mandatory to develop knowledge in a specified field of study. The region, pharmacy as a first choice, experience in a leadership position, previous pharmacy-related work, and previous attendance in the conference have shown statistically significant association with interest to follow PGE. Further works including preparing motivational speeches and details on admission requirements about available PGE programs has to be prepared in collaboration with the Ministry of Education (MOE) to enhance students' interest. Besides, the potential development of scholarships for PGE may enhance the ability of some students to apply for PGE, who otherwise would be unable to.\n\nAbbreviations {#S0006}\n=============\n\nAOR, adjusted odds ratio; COR, crude odd ratio; GPA, grade point average; MOE, Ministry of Education; PGE, postgraduate education; SPSS, Statistical Package for the Social Sciences; TV, television; UK, United Kingdom; USA, United States of America; UOG, University of Gondar.\n\nData Sharing Statement {#S0007}\n======================\n\nThe corresponding author has all materials and data of this study and it is available upon request.\n\nEthical Declaration and Informed Consent {#S0008}\n========================================\n\nEthical clearance was obtained from the School of Pharmacy, University of Gondar research, and the ethics committee before the commencement of the study. Verbal informed consent was obtained and it was acceptable and approved by the School of Pharmacy, University of Gondar Research and Ethics Committee.\n\nAuthor Contributions {#S0009}\n====================\n\nAll authors made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; took part in drafting the article or revising it critically for important intellectual content; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.\n\nDisclosure {#S0010}\n==========\n\nThe authors declared that they have no competing interests.\n\n[^1]: **Notes:** \\*Significant at 0.05 levels; \\*\\*significant at 0.01 levels.\n"}
+{"text": "Introduction\n============\n\nCongenital hypothyroidism (CH), which is defined by inadequate thyroid hormone production in newborn infants, is the most common neonatal metabolic disorder worldwide, with an incidence of 1 in 2000--4000 live births ([@B33]). Most neonates born with CH have a normal appearance and no detectable physical signs. In the past, we have often overlooked the harmfulness of hypothyroidism during the newborn period. Patients with the disease suffer from delayed diagnosis and treatment, and severe CH can lead to growth retardation and permanent intellectual disability. CH screening is an important component of the newborn screening (NBS) program, which is widely used as the third prevention intervention of birth defects ([@B17]; [@B2]). Using the NBS program, we can detect symptomless children with CH early. Children can receive a definitive diagnosis, and the proper treatment can be applied in time to prevent further complications and sequelae. The NBS program was established in Changzhou city in 2004, and approximately 430,000 infants have since been tested.\n\nCongenital hypothyroidism screening has been carried out all over the world for nearly 50 years, but the pathogenesis of CH remains unclear. An increasing amount of evidence suggests that genetic mutations are an important factor of CH ([@B29]). At present, more than 600 genomic variations have been recorded in the ClinVar database. CH is divided into two main types: thyroid dysgenesis and thyroid dyshormonogenesis. According to previous reports, the cause of CH in approximately 80--85% of patients is thyroid dysgenesis (including agenesis, ectopy, and hypoplasia), which is related to gene mutations in thyroid-stimulating hormone receptor (*TSHR*), paired box gene 8 (*PAX8*), thyroid transcription factor 1 (*TTF1/NKX2-1*), thyroid transcription factor 2 (*TTF2/FOXE1*), and NK2 transcription factor related locus 5 (*NKX2-5*). Otherwise, 10--15% of cases are caused by thyroid dyshormonogenesis, which is associated with mutations in thyroid oxidase 2 (*DUOX2*), dual-oxidase maturation factor 2 (*DUOXA2*), thyroglobulin (*TG*), thyroid peroxidase (*TPO*), solute carrier family 5 member 5 (*SLC5A5*), solute carrier family 26 member 4 (*SLC26A4*) and iodotyrosine deiodinase (*IYD*) ([@B29]; [@B5]). These genes play important roles in the growth and development of the thyroid gland. Genomic variations can prevent or destroy normal development of the gland or disturb the production of thyroid hormones. However, most previous studies have focused on Western populations. Few similar studies have been reported in a Chinese population and have focused on one or two pathogenicity genes ([@B11]; [@B14]; [@B18]). There are few reports on the mutation spectrum of CH-causing genes in the Chinese population.\n\nIn the present study, we retrospectively analyzed the clinical data of CH screening over the last 14 years and performed mutation screening of CH-causing genes in CH infants using next-generation sequencing (NGS). We hope to improve CH neonatal screening and better characterize the mutations of CH-causing genes in a Chinese population.\n\nMaterials and Methods {#s1}\n=====================\n\nPatients and Design\n-------------------\n\nFrom January 2004 to December 2016, all newborn infants who accepted the NBS program in Changzhou Maternity and Child Health Care Hospital Affiliated to Nanjing Medical University were recruited for this study. All subjects received CH screening via collection of dried blood spots (DBSs). In 2012, we began our search for CH-causing genes. One hundred and six non-consanguineous patients diagnosed with CH consented to undergoing the gene mutation test.\n\nThe study design and protocol were reviewed and approved by the ethics committee of Changzhou Maternity and Child Health Care Hospital Affiliated to Nanjing Medical University.\n\nNBS Program\n-----------\n\nThe methods of screening, diagnosis, and treatment were carried out according to the \"[@B7], which was promulgated by the Chinese Preventive Medicine Association. Briefly, DBS were collected from all infants on 903 filter paper (Wallac OY, Turku, Finland) at 72 h after birth. The levels of neonatal thyroid-stimulating hormone (NTSH) of DBS were detected by a time-resolved fluoroimmunoassay using the Wallac 1420 or Wallac 1235 AutoDELFIA (Perkin Elmer, Waltham, MA, United States). If the NTSH level was \\<9.0 mIU/L, the infants were considered normal. If the NTSH level was \u22659.0 and \\<20.0 mIU/L, the infants were recalled, and DBSs were collected for a second time (within 1 week) and the NTSH level re-determined. If the NTSH level was still \\>9.0 mIU/L, the infants were recalled again. The infant was considered normal if the second NTSH level was \\<9.0 mIU/L. If the NTSH level was \u226520.0 mIU/L, the infants were recalled, and the levels of serum thyroid-stimulating hormone (TSH) and free thyroxine (FT4) were determined.\n\nDiagnosis of CH\n---------------\n\nVenous blood from the recalled infants in the NBS program was sampled to test the concentrations of TSH and FT4. Diagnosis of CH was based on elevated TSH levels (TSH \u2265 10 mIU/L) and decreased FT4 levels (FT4 \\< 7.77 pmol/L). Serum TSH and FT4 levels were determined by electrochemistry immunoassay (ECL) using the COBAS e601 automated analyzer (Roche Diagnostics, Mannheim, Germany).\n\nTargeted Next-Generation Sequencing\n-----------------------------------\n\nGenomic DNA was extracted from peripheral blood leukocytes using the QIAamp DNA blood kit according to the manufacturer's protocol. A total of 10 ng DNA per sample was used for sequencing using the CH capture panel, which was designed based on the Illumina Truseq Custom Amplicon v1.5 kit. Thirteen pathogenic genes (*DUOX2*, *TG*, *TPO*, *TSHR*, *TTF1*, *TTF2*, *PAX8*, *NKX2-5*, *GNAS*, *THRA*, *TSHB*, *IYD,* and *SLC5A5*) were screened in all patients, including the entire coding regions and exon-intron boundaries. The genetic fragments were between 250 and 280 bp and were prepared using the Covaris and Agencourt AMPure XP kits, which include purified and captured gene fragments. Adaptor-ligated amplicons were prepared using the Illumina Paired-End Sample Preparation kit. Illumina multi-PE-adaptors were bound to terminal genes and target enrichment was performed by multiplex PCR. After 12 PCR cycles, amplicons were purified using Agencourt AMpurc SPRI XP beads and captured on the Illumina MiSeq 2000 instrument.\n\nMutation Analysis\n-----------------\n\nIllumina Amplicon Viewer v1.3 and MiSeq Reporter v2.3 software were used for data analysis, and SnpEff was used for mutation annotation. We also used automatic tools (including SIFT, PolyPhen-2, and MutationTaster) to predict the impact of mutations on the function and structure of their respective proteins. Briefly, mutations with frequencies \\> 1% or synonymous mutations were filtered. The evaluation of potential pathogenic mutations was based on the combination of the three tools, and mutations for which 2 of the 3 tools predicted damaging effects were selected. We also searched the selected mutations in other published studies to evaluate their potential pathogenicity.\n\nStatistical Analysis\n--------------------\n\nData that were not normally distributed are expressed as medians (M), 25^th^ percentiles (P25), and 75^th^ percentiles (P75). All data were analyzed using EmpowerStats x64 software ([@B38]).\n\nResults\n=======\n\nA total of 437,342 newborns, including 236,820 males and 200,522 females, underwent CH screening. In total, 3,931 infants had positive results, and their NTSH levels were in the range of 9.0--20.0 mIU/L. After being recalled, 3,768 infants underwent the second DBS test. Otherwise, 289 infants were considered as positive because their NTSH levels were \\>20.0 mIU/L. The positive rate of initial screening was 0.96% (4220/437342); 181 cases were missing in the first recall. The positive recall rate of initial screening was 95.7% (4039/4220). The NTSH levels of 501 recalled cases were still more than 9.0 mIU/L, and the positive rate of the recalled cases was 13.3% (501/3768). A total of 714 cases underwent the serological confirmation test, and 192 infants (105 males and 87 females) were diagnosed with CH. The incidence of CH was 1:2278 (1:2255 for males and 1:2304 for females). The positive predictive value was 4.8% (192/4039). A total of 239 infants were lost to follow-up. The loss to follow-up rate was 0.05% (239/437342). Reasons for loss to follow-up included incomplete contact information, dissent of the parents, or the patient moved to a new residence.\n\nWe examined the distribution of thyroid hormone levels (Table [1](#T1){ref-type=\"table\"}). The median NTSH level of the CH infants was 46.10 mIU/L (P25--P75: 17.90--120.00). The NTSH level of the healthy infants was 2.39 mIU/L (P25--P75: 1.37--3.93). The NTSH level of 99.04% of the normal infants was \\<9.0 mIU/L. The NTSH level was \\>100 mIU/L in 30.73% and 30 to 100 mIU/L in 28.13% of the CH infants. After serological examination, the median serum TSH level of CH infants was 75.0 mIU/L (P25--P75: 75.00--75.00), and that of FT4 was 5.14 pmol/L (P25--P75: 2.83--8.39). In total, 80.73% of CH cases had TSH levels \\>75 mIU/L.\n\n###### \n\nDistribution of thyroid hormone levels.\n\n  Range                                          *n*      \\%      Median   P25--P75\n  ---------------------------------------------- -------- ------- -------- ---------------\n  **NTSH in dried blood spots in all infants**                             \n  Total                                          437342   100     2.39     1.37--3.93\n  \\<3                                            220272   61.80            \n  3--\\<6                                         132383   30.27            \n  6--\\<9                                         30467    6.97             \n  9--\\<12                                        3208     0.73             \n  \u226512                                            1012     0.23             \n  **NTSH in dried blood spots in CH infants**                              \n  Total                                          192      100     46.10    17.90--120.00\n  \\<9                                            5        2.60             \n  9--\\<15                                        35       18.23            \n  15--\\<50                                       59       30.73            \n  50--\\<100                                      34       17.71            \n  00--\\<200                                      42       21.88            \n  \u2265200                                           17       8.85             \n  **Serum TSH in CH infants**                                              \n  Total                                          192      100     75.00    75.00--75.00\n  10--\\<20                                       2        1.04             \n  20--\\<30                                       4        2.08             \n  30--\\<50                                       15       7.81             \n  50--\\<75                                       16       8.33             \n  \u226575                                            155      80.73            \n  **Serum FT4 in CH infants**                                              \n  Total                                          192      100     5.14     2.83--8.39\n  \\<2.0                                          36       18.75            \n  2.0--\\<4.0                                     36       18.75            \n  4.0--\\<6.0                                     48       25.00            \n  6.0--\\<8.0                                     27       14.06            \n  8.0--\\<10.0                                    23       11.98            \n  10.0--\\<20                                     22       11.46            \n                                                                           \n\nCongenital hypothyroidism-causing genes were detected by targeted NGS in 106 CH infants. Based on our literature review, we designed a target sequencing panel, which included 13 causative genes: *DUOX2*, *TG*, *TPO*, *TSHR*, *TTF1*, *TTF2*, *PAX8*, *NKX2-5*, *GNAS*, *THRA*, *TSHB*, *IYD*, and *SLC5A5*. Among 106 CH infants, 69 (65.1%) had more than one gene mutation, and a total of 132 mutations were identified in nine genes (*DUOX2*, *TG*, *TPO*, *TSHR*, *TTF1*, *TTF2*, *NKX2-5*, *PAX8*, and *GNAS*). No mutation in *THRA*, *TSHB*, *IYD*, or *SLC5A5* was detected. In total, 92 of 132 (69.70%) mutations were related to thyroid dyshormonogenesis \\[*DUOX2* (*n* = 49), *TG* (*n* = 35), and *TPO* (*n* = 8)\\]. Additionally, 21.21% (28/132) of mutations were related to thyroid dysgenesis \\[*TSHR* (*n* = 19), *TTF1* (*n* = 5), *TTF2* (*n* = 1), *PAX8* (*n* = 2), and *NKX2-5* (*n* = 1)\\]. Moreover, 9.09% (12/132) of mutations were related to *GNAS*, which is associated with thyrotropin resistance ([@B21]). We made a list of the top 10 gene mutations detected in our study (Table [2](#T2){ref-type=\"table\"}). It is worth noting that the majority of mutations in the CH-causing genes were in *DUOX2*, *TG*, and *TSHR*, especially *DUOX2*. There were a total of 48 *DUOX2* mutations detected in our study, and 33 of 106 CH infants harbored *DUOX2* mutations. Twenty-four different types of *DUOX2* mutations were detected, including 20 reported mutations and four novel mutations (c.3721A \\> T, c.3321delC, c.1007_1009del, and c.1300_1320del). We also found 29 novel mutations, each with a rare frequency (Table [3](#T3){ref-type=\"table\"}).\n\n###### \n\nTop 10 of genes mutations in our study.\n\n  Gene_symbol   CytoBand   Exon position   Nucleotide position   Amino acid position   Mutation types   Number   RS ID\n  ------------- ---------- --------------- --------------------- --------------------- ---------------- -------- -------------\n  *TG*          8q24.22    Exon45          c.7847A \\> T          p.N2616I              Non-synonymous   11       rs10091530\n  *DUOX2*       15q21.1    Exon14          c.1588A \\> T          p.K530X               Stopgain         6        rs180671269\n  *DUOX2*       15q21.1    Exon30          c.4027C \\> T          p.L1343F              Non-synonymous   5        rs147945181\n  *DUOX2*       15q21.1    Exon25          c.3329G \\> A          p.R1110Q              Non-synonymous   5        rs368488511\n  *DUOX2*       15q21.1    Exon28          c.3632G \\> A          p.R1211H              Non-synonymous   4        rs141763307\n  *TG*          8q24.22    Exon42          c.7364G \\> A          p.R2455H              Non-synonymous   4        rs2272707\n  *DUOX2*       15q21.1    Exon4           c.227C \\> T           p.P76L                Non-synonymous   3        rs767705906\n  *DUOX2*       15q21.1    Exon26          c.3478_3480del        p.1160_1160del        Non-frameshift   3        rs758318135\n  *DUOX2*       15q21.1    Exon6           c.605_621del          p.Q202fs              Frameshift       3        rs769318570\n  *TSHR*        14q31.1    Exon10          c.1349G \\> A          p.R450H               Non-synonymous   3        rs189261858\n                                                                                                                 \n\nNCBI Reference Sequence: TG (NM_003235), DUOX2 (NM_014080), TSHR (NM_000369).\n\n###### \n\nNovel mutations in our study.\n\n  Gene_symbol                   CytoBand   Exon position   Nucleotide position   Amino acid position   Mutation types   Number\n  ----------------------------- ---------- --------------- --------------------- --------------------- ---------------- --------\n  ***TG* (NM_003235)**                                                                                                  \n                                8q24.22    Exon34          c.6185G \\> A          p.W2062X              Stopgain         2\n                                8q24.22    Exon8           c.976C \\> T           p.Q326X               Stopgain         1\n                                8q24.22    Exon8           c.1000delG            p.G334fs              Frameshift       1\n                                8q24.22    Exon10          c.2593C \\> A          p.P865T               Non-synonymous   1\n                                8q24.22    Exon16          c.3457A \\> T          p.K1153X              Stopgain         1\n                                8q24.22    Exon16          c.3538C \\> T          p.Q1180X              Stopgain         1\n                                8q24.22    Exon18          c.3994C \\> T          p.Q1332X              Stopgain         1\n                                8q24.22    Exon25          c.5020C \\> A          p.P1674T              Non-synonymous   1\n                                8q24.22    Exon45          c.7799G \\> A          p.W2600X              Stopgain         1\n  ***DUOX2* (NM_014080)**                                                                                               \n                                15q21.1    Exon9           c.1007_1009del        p.336_337del          Non-frameshift   1\n                                15q21.1    Exon12          c.1300_1320del        p.434_440del          Non-frameshift   1\n                                15q21.1    Exon25          c.3321delC            p.T1107fs             Frameshift       1\n                                15q21.1    Exon29          c.3721A \\> T          p.I1241F              Non-synonymous   1\n  ***TSHR* (NM_000369)**                                                                                                \n                                14q31.1    Exon1           c.152C \\> A           p.P51Q                Non-synonymous   1\n                                14q31.1    Exon6           c.501C \\> G           p.I167M               Non-synonymous   1\n                                14q31.1    Exon9           c.700T \\> C           p.S234P               Non-synonymous   1\n                                14q31.1    Exon10          c.1384T \\> C          p.C462R               Non-synonymous   1\n  ***TTF1* (NM_007344)**                                                                                                \n                                9q34.13    Exon2           c.269G \\> A           p.R90K                Non-synonymous   2\n                                9q34.13    Exon2           c.515A \\> G           p.Q172R               Non-synonymous   1\n                                9q34.13    Exon4           c.1598C \\> T          p.A533V               Non-synonymous   1\n  ***GNAS* (NM_000516)**                                                                                                \n                                20q13.32   Exon4           c.308T \\> C           p.I103T               Non-synonymous   2\n                                20q13.32   Exon6           c.478C \\> T           p.R160C               Non-synonymous   1\n                                20q13.32   Exon12          c.1018T \\> C          p.F340L               Non-synonymous   1\n  ***PAX8* (NM_013953)**                                                                                                \n                                2q13       Exon4           c.275T \\> C           p.I92T                Non-synonymous   1\n                                2q13       Exon5           c.398G \\> A           p.R133Q               Non-synonymous   1\n  ***TPO* (NM_175722)**                                                                                                 \n                                2p25.3     Exon13          c.2080T \\> C          p.S694P               Non-synonymous   1\n  ***NKX2-5* (NM_001166176)**                                                                                           \n                                5q35.1     Exon2           c.416G \\> A           p.S139N               Non-synonymous   1\n                                                                                                                        \n\nAmong the 69 CH infants with mutations, 28 (40.58%) had one potentially functional mutation, 26 (37.68%) had two mutations, and 15 (21.74%) had three or more mutations. Similarly, 39 (56.52%) infants had mutations in one gene, and 30 (43.483%) had mutations in two or three genes.\n\nDiscussion\n==========\n\nThe NBS program for CH is a major method used in preventive medicine. In this study, we retrospectively analyzed the clinical data from NBS over the last 14 years with the goal of improving CH neonatal screening. According to our results, the incidence of CH in Changzhou is 1:2278, which is the average level in China ([@B41]). The National Centre for Clinical Laboratories reported an incidence of CH was 1:2281 based on the data of the 202 laboratories around China. [@B37] reported that the apparent incidence of CH has more than doubled in recent years ranging from 1:2800 to 1:1400. Meanwhile, countries, such as the United States ([@B25]), Canada ([@B9]), New Zealand ([@B13]), Scotland ([@B20]), Brazil ([@B34]), reported slight differences. In our study, there was no significant difference in the incidence of CH between males and females, consistent with the study by [@B39].\n\nIn the present study, we detected mutations of CH-causing genes in a Chinese population by targeted NGS and found that the abnormal rates of these related genes in Chinese CH patients was 65.1%. A total of 132 gene mutations were detected (69.70, 21.2, and 9.09% mutations were related to thyroid dyshormonogenesis, thyroid dysgenesis, and thyrotropin resistance, respectively). This result was quite different from previous reports. According to research in Western countries, the primary pathology of CH is thyroid dyshormonogenesis ([@B29]; [@B5]). The pathogenic factors of CH in our Chinese population may differ from those in Western populations. It is very important to screen the pathogenic genes and pathogenic factors of CH in this region. In addition, the current study indicated that a considerable proportion of Chinese CH patients had mutations at multiple sites or in multiple genes. Multiple mutations may cause a more serious phenotype in CH patients. Recent studies have also revealed that a significant proportion of CH patients have multiple gene variations in more than one thyroid-specific gene ([@B8]). Moreover, heritable variations were found in more than half of our CH patients, as well as in the general population, albeit at a significantly lower prevalence. Together, these studies indicate that the pathogenesis of CH may be due to the sum effect of rare alleles ([@B32]). A previous study also indicated that patients with one or two *DUOX2* pathogenic mutations developed subclinical or transient CH, whereas patients with three or more *DUOX2* pathogenic mutations were associated with permanent CH ([@B24]). The coexistence of multiple pathogenic mutations may contribute to the severity of the hypothyroid condition, and mutations in multiple genes may lead to genotype--phenotype variability ([@B26]; [@B30]; [@B40]). Therefore, further studies are needed to enlarge the mutation spectrum of CH and to verify the functions of the associated mutations, which may provide more profound insight into the etiology of CH.\n\nIn the present study, *DUOX2* was the most commonly mutated gene in Chinese CH infants. According to previous studies, mutations in *DUOX2* are responsible for thyroid dyshormonogenesis ([@B27]). Most patients with *DUOX2* pathogenic mutations have an ectopic thyroid gland with an increased or normal size ([@B18]). However, the mutational spectrum of the *DUOX2* gene and the correlations between phenotype and genotype have not yet been fully established. The c.1588A \\> T mutation in *DUOX2*, which is responsible for thyroid dyshormonogenesis, was highly recurrent, with a prevalence of 1/40,000. The c.1588A \\> T mutation is population specific and has been reported mainly in Asian populations, including Chinese ([@B12], [@B10]; [@B35]), Japanese ([@B23], [@B22]), and Malaysian ([@B6]) populations. The c.4027C \\> T ([@B4]), c.3329G \\> A ([@B12]; [@B31]), c.3632G \\> A ([@B3]), c.2335G \\> A ([@B15]; [@B22]), and c.2654G \\> A ([@B40]) mutations are also predominant in Asians, mostly in the Chinese Han population c.1883delA ([@B23], [@B22]; [@B31]; [@B35]), c.3478_3480del ([@B28]; [@B10]; [@B31]), and c.605_621del ([@B16]; [@B24]; [@B35]) show a scattered distribution in Asian populations, including China, Japan, and South Korea. Six other mutations, including c.2048G \\> T ([@B10]), c.227C \\> T ([@B19]), c.2894C \\> T ([@B15]), c.3391G \\> T ([@B36]; [@B10]), c.2202G \\> A ([@B36]) and c.2104_2106del ([@B10]), were reported only in China, and the missense mutation of c.4405G \\> A, which was previous reported in Korean ([@B31]), was first identified among Chinese population in our study. The c.1873C \\> T mutation was identified as a novel pathogenic mutation by the qCarrier test in a reproductive carrier-testing program ([@B1]), but no direct evidence has shown that the mutation is related to CH. In addition, c.1265G \\> A, c.2413G \\> A, c.1717C \\> T, c.3721A \\> T, c.3321delC, c.1300_1320del, and c.1007_1009del mutations, which may be related to CH, were identified in our study for the first time.\n\nConclusion\n==========\n\nThe incidence of CH in Changzhou city is 1:2278. Some related quality control indicators indicate that the NBS program of CH in Changzhou is effective. Meanwhile, we preliminarily identified the pathogenic genes in infants with CH by targeted NGS. The rate of abnormal gene mutations was 65.1%, and most mutations were related to thyroid dyshormonogenesis, which differs from that observed in Western populations. A considerable proportion of the population had mutations at multiple sites, and *DUOX2* was the most common gene mutation in Chinese CH infants.\n\nAuthor Contributions\n====================\n\nBY, HW, and WL carried out the assays and participated in designing the study. HW and YW carried out clinical consultations. BY, YY, WL, and LJ carried out laboratory tests and performed the statistical analysis. ZC conceived the study, participated in its design and coordination, and helped draft the manuscript.\n\nConflict of Interest Statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This study was supported by grants from the Key Research and Development Plan Project of Jiangsu Province (BE2017650 and BE2018677), Changzhou Science and Technology Support Project (CE20175021), Jiangsu Maternal and Child Health Research Project (F201671 and FXK201754), and Project Funding for the Training of High Level Health Professionals in Changzhou (2016CZLJ013).\n\nWe thank Yan Li for assistance of samples, Fang Guo and Rui Yang for comments on this manuscript, and Dr. Jianbin Liu for bioinformatics analysis.\n\n[^1]: Edited by: Antonio Brunetti, Universit\u00e0 degli Studi Magna Gr\u00e6cia di Catanzaro, Italy\n\n[^2]: Reviewed by: Giuseppe Damante, Universit\u00e0 degli Studi di Udine, Italy; Georges Jacques Casimir, Free University of Brussels, Belgium\n\n[^3]: ^\u2020^These authors have contributed equally to this work\n\n[^4]: This article was submitted to Genomic Endocrinology, a section of the journal Frontiers in Genetics\n"}
+{"text": "Chandra A, Pius C, Nabeel M, et al. Ovarian cancer: Current status and strategies for improving therapeutic outcomes. Cancer Med. 2019;8:7018--7031. 10.1002/cam4.2560 31560828\n\n**Funding information**\n\nSupported by NIMHD (grant \\#: 2U54 MD006882\u201006), NCI (grant \\#: 1P20CA233355\u201001) and NHLBI (grant \\#: R25HL125447).\n\n1. INTRODUCTION {#cam42560-sec-0001}\n===============\n\nOvarian cancer (OC) is the deadliest cancer among women placing it with 4th place for all the fatal disease among women. Cancer statistics from 2019 show that the estimated number of new cases is 22\u00a0240 with deaths around 14\u00a0170 cases.[1](#cam42560-bib-0001){ref-type=\"ref\"} There are three histological types associated with the disease. The most common is epithelial OC (EOC). Patients with this fatal disease have only 45.6% 5\u2010year survival rate.[2](#cam42560-bib-0002){ref-type=\"ref\"} The survival rate in general increases up to 70% if effective early stage detection is possible. Early\u2010stage detection rate for this disease is as low as 20%. For most of the patients the late stage detection with advanced stage of cancer leads to low survival rate of 35%. In the case of recurrent EOC there is no satisfactory cure till to date.\n\nSeveral aspects influence the progression of the disease. Genetic and epigenetic factors are the most important ones among them. Nearly 10%\u201015% of familial OCs result from breast cancer gene mutations BRCA1 and BRCA2.[3](#cam42560-bib-0003){ref-type=\"ref\"} The characteristic feature of these cancers is that they are multifocal and progress rather quickly. Mutations and the loss of the TP53 function are found in 60%\u201080% of the familial and sporadic cases of the disease.[3](#cam42560-bib-0003){ref-type=\"ref\"} These oncogenes will turn on different signaling pathways that leads to pathogenicity. Higher rate of thrombosis associated with OC is due to such activation of coagulation pathways by OC.[4](#cam42560-bib-0004){ref-type=\"ref\"}, [5](#cam42560-bib-0005){ref-type=\"ref\"}\n\n2. OVARIAN CARCINOMA\u2010PATHOBIOLOGY {#cam42560-sec-0002}\n=================================\n\nOvarian carcinoma is heterogeneous in nature. The disease progresses through several molecular level changes. Mainly there are three areas in the ovary where the tumor is developed. Surface epithelium is where majority of the malignancy is developed from. It is presented in different type of histology. Serous ovarian carcinoma (SOC) is the most common one and is presented at old age. Endometrioid carcinoma is presented at young age and associated with endometriosis. Mucinous carcinoma and clear cell carcinoma are also presented at young age. The other areas where the OC is developed are the germ cells and stroma.\n\nThe complexity in these malignancies arises from the microenvironment affected by changes in genetic factors. The degree of complexity varies according to the changes in epigenetic factors too. Understanding the tumor microenvironment is key for its diagnosis, treatment options, and survival. The microenvironment varies for different type of ovarian carcinomas with changes in gene expression leads to different tumor markers. The tumor markers play crucial role for the development of targeted therapies.[3](#cam42560-bib-0003){ref-type=\"ref\"} Abnormal expression of homeobox (HOX) has been shown in histologic types developed at the embryonic stage.[6](#cam42560-bib-0006){ref-type=\"ref\"} HOXA9 is absent in normal ovarian cells. High\u2010level expression of HOXA9 in SOC is found at the embryonic stage during fallopian tube formation. Abnormal level of HOXA10 is linked to endometroid carcinoma and HOXA11is linked to mucinous carcinoma.[6](#cam42560-bib-0006){ref-type=\"ref\"} As far as the treatment goes, platinum\u2010 and taxane\u2010based chemotherapy have been shown to be successful for serous and endometrioid cancers, compared with clear\u2010cell cancers and mucinous histology type cancers.[3](#cam42560-bib-0003){ref-type=\"ref\"}\n\n3. OC SCREENING TESTS {#cam42560-sec-0003}\n=====================\n\nLow endurance rate for OC patients is because of the late\u2010stage detection and diagnosis of the disease. Early stage detection and diagnostic tools for screening OC are not efficient. Research is in progress for developing efficient diagnostic processes for OC. Transvaginal ultrasound (TVUS) is one of the current screening process for OC. Blood test for CA125, a tumor marker for OC is another common screening test for OC.[7](#cam42560-bib-0007){ref-type=\"ref\"} Transvaginal ultrasound will identify the growth and masses in the scanned area. It will not differentiate between the malignant and benign masses. CA125 is elevated in ovarian carcinoma. But it is not specific to OC, hence a combination of TVUS in patients with high levels of CA125 can be a better screening tool for OC diagnosis.[8](#cam42560-bib-0008){ref-type=\"ref\"}\n\n4. BIOMARKERS FOR OC {#cam42560-sec-0004}\n====================\n\nAs the disease progresses, it gets even harder to treat and manage the patients. Only 20% of those affected cases have an early detection of the ailment. Many healthcare professionals confused OC with other urologic, abdominal, and gynecologic diseases because of the overlap in signs and symptoms, resulting in late detections. Ovaries do not have a peritoneal covering; therefore, the cancer spreads locally to the peritoneal cavity, resulting in symptoms. Absence of effective testing tools and equipments further delay the detection process for OC.[9](#cam42560-bib-0009){ref-type=\"ref\"} As noted earlier, the early detection is crucial in increasing survival rates for advanced\u2010stage OC patients. Biomarkers are divided into diagnostic, prognostic, predictive, and response categories. Poor sensitivity and lack of specificity are the challenges for majority of biomarkes that have been studied. Although, the common biomarkers currently used are CA125, Human Epididymis Protein 4 (HE4), and mesothelin,[9](#cam42560-bib-0009){ref-type=\"ref\"}, [10](#cam42560-bib-0010){ref-type=\"ref\"} and their use in combination is often feasible.\n\n4.1. CA125 {#cam42560-sec-0005}\n----------\n\nCurrently, the disease progression and treatment efficacy in OC patients is monitored using TVUS and elevated CA125 expression.[11](#cam42560-bib-0011){ref-type=\"ref\"}, [12](#cam42560-bib-0012){ref-type=\"ref\"} Elevated CA125 levels are present in about 80% of advance stage OC patients. For early stage OC patients elevated CA125 level is present in 50% only.[13](#cam42560-bib-0013){ref-type=\"ref\"} New studies from Jennings group demonstrated the association of Neu5Gc\u2010glycans and SubB2M for detecting CA125 and using as an effective tool for the diagnosis and outcomes in stage II and IV patients.[14](#cam42560-bib-0014){ref-type=\"ref\"}\n\n4.2. HE4 {#cam42560-sec-0006}\n--------\n\nThis is another maker for OC. Elevated HE4 expression is present in OC patients compared with normal and other nonmalignant diseases for women.[15](#cam42560-bib-0015){ref-type=\"ref\"}, [16](#cam42560-bib-0016){ref-type=\"ref\"} HE4 and CA125 are the biomarkers used in a study of women (n\u00a0=\u00a0531) who has pelvic masses. 93.8% of these women were predicted for high\u2010risk ovarian carcinoma.[15](#cam42560-bib-0015){ref-type=\"ref\"} In the US, HE4 is only approved as maker for OC for disease recurrence or progression.\n\n4.3. Mesothelin {#cam42560-sec-0007}\n---------------\n\nA 40\u2010kD protein associated with cell survival, tumor progression, and adherence. It is present in normal mesothelial cells. Increased levels of mesothelin is presented in blood samples of 40%\u201067% of patients with OC.[16](#cam42560-bib-0016){ref-type=\"ref\"}, [17](#cam42560-bib-0017){ref-type=\"ref\"}, [18](#cam42560-bib-0018){ref-type=\"ref\"}, [19](#cam42560-bib-0019){ref-type=\"ref\"} The high expression level of mesothelin in OC identifies it as strong candidate for targeted therapy.\n\n4.4. OVA1 {#cam42560-sec-0008}\n---------\n\nA multiple biomarker\u2010based test OVA1 (Ovarian Malignancy Algorithm) is currently used for the evaluation of risk level of OC patients.[20](#cam42560-bib-0020){ref-type=\"ref\"} Microglobulin Beta2, CA125, transthyretin (pre\u2010albumin), ApoA1, and transferrin are the biomarkers in OVA1. OVA1 analyze serum levels of these biomarkers. The OVA1 algorithm combines the results of these levels with information on the menopausal status of the patient for OC risk group classification.\n\n4.5. DOvEEgene {#cam42560-sec-0009}\n--------------\n\nIt is an ongoing clinical trial (NCT02288676) study sponsored by McGill University, Canada. In advanced OC, treatment efficiency was studied using computed tomography (CT) perfusion.[21](#cam42560-bib-0021){ref-type=\"ref\"}\n\nNumerous studies, and trials involving OC biomarkers that are being conducted across the globe. A summary of the ongoing biomarker studies in OC detection is given in Table [S1](#cam42560-sup-0001){ref-type=\"supplementary-material\"} (Source: <http://Clinicaltrial.gov>).\n\n5. OC TREATMENT STRATEGIES {#cam42560-sec-0010}\n==========================\n\nThe treatment strategies for different type of cancer depends on its pathological stages. Early detection will help to have the treatment options that are promising and effective. Current treatment options are combining debulking surgery and drug treatment and radiation therapy. Some of the advanced level treatment options include targeted therapy, immunotherapy, and hormone therapy. Chemotherapy is the most vital part of OC treatment. Chemotherapeutic agents can be administered via intravenously (IV), intraperitonially (IP), or by IV/IP combination. In neo adjuvant treatment plan chemotherapy was done before the surgery. IP/IV combination delivery of chemotherapeutic agents is the preferred mode drug administration for OC patients with cytoreduced disease.[22](#cam42560-bib-0022){ref-type=\"ref\"}, [23](#cam42560-bib-0023){ref-type=\"ref\"}, [24](#cam42560-bib-0024){ref-type=\"ref\"} Treatment of peritoneal area is most effective when the chemotherapeutic agents are administered via IP route.[25](#cam42560-bib-0025){ref-type=\"ref\"} Compared with the IV carboplatin chemotherapy, the IP carboplatin chemotherapy is well tolerated in advance stage OC patients undergoing surgery followed by neoadjuvant.[26](#cam42560-bib-0026){ref-type=\"ref\"}\n\nChemotherapeutic agents will be selected for treatment based on the stage of OC. Platinum containing drugs (cisplatin and carboplatin) and taxane family (paclitaxel and docetaxel) are frequently used chemotherapeutic agents for treatment of OC.[27](#cam42560-bib-0027){ref-type=\"ref\"} Carboplatin is the preferred choice over cisplatin due to its reduced toxicity, and side effects with equivalent response rate and survival outcomes.[28](#cam42560-bib-0028){ref-type=\"ref\"}, [29](#cam42560-bib-0029){ref-type=\"ref\"}, [30](#cam42560-bib-0030){ref-type=\"ref\"} Sensitivity of the chemotherapeutic agent is important during the drug selection process of OC. Gemcitabine, doxorubicin, and bevacizumab are the drugs used for treatment for cisplatin and carboplatin\u2010resistant ovarian carcinoma.[31](#cam42560-bib-0031){ref-type=\"ref\"}, [32](#cam42560-bib-0032){ref-type=\"ref\"}, [33](#cam42560-bib-0033){ref-type=\"ref\"} Usage of high\u2010dose chemotherapeutic agent will lead to complications due to side effects and can result in termination of treatment plan. Since the OC cells undergo molecular level changes over the time and may lead to resistance to chemotherapy. A list of currently approved chemotherapeutic agents for OC therapy and their mechanism(s) for anti\u2010cancer activity is summarized in Table [S2](#cam42560-sup-0002){ref-type=\"supplementary-material\"}.\n\n6. ROLE OF APOPTOTIC GENES AND TARGETED THERAPY {#cam42560-sec-0011}\n===============================================\n\nThe biological phenomena by which the body gets rid of unnecessary cells in order to maintain homeostasis is known as apoptosis. OC, among others, has several genes working against apoptosis, which allows cancerous cells to flourish instead of being killed off. Candidates involved in both intrinsic and extrinsic pathways were studied. Bcl\u20102 family proteins and Tyrosine\u2010protein kinases, respectively, facilitates intrinsic and extrinsic apoptosis, while inhibitor of apoptosis (IAP) proteins are associated with both intrinsic and extrinsic pathways. Bcl2 is anti\u2010apoptotic [34](#cam42560-bib-0034){ref-type=\"ref\"} and is expressed in high concentration in OC. [35](#cam42560-bib-0035){ref-type=\"ref\"}, [36](#cam42560-bib-0036){ref-type=\"ref\"} Additionally, Bcl2 modulates resistance to chemotherapy and decreases survival, along with Bcl\u2010X and Mcl\u20101 in OC patients.[36](#cam42560-bib-0036){ref-type=\"ref\"}, [37](#cam42560-bib-0037){ref-type=\"ref\"}, [38](#cam42560-bib-0038){ref-type=\"ref\"} Conversely, Bid, Bad, Bax, and Kak all respond to the treatment by inducing apoptosis [39](#cam42560-bib-0039){ref-type=\"ref\"} and improve the survival. Clinical trials for treatment with Bcl\u20102 inhibitors improved the response to cisplatin, and this has also been seen in preclinical models of OC studies.[39](#cam42560-bib-0039){ref-type=\"ref\"}, [40](#cam42560-bib-0040){ref-type=\"ref\"}\n\nAnother anti\u2010apoptotic gene family is the IAP proteins. Survivin is a well characterized inhibitor for apoptotic proteins present in ovarian and other type of cancer cells.[41](#cam42560-bib-0041){ref-type=\"ref\"} Survivin plays a significant role in cell division and thus control apoptosis. Animal studies have shown that targeting survivin with suppressor drugs resulted in tumor growth suppression and enhanced sensitivity to chemotherapeutic agents.[41](#cam42560-bib-0041){ref-type=\"ref\"}\n\nTyrosine\u2010protein kinase Met (c\u2010Met) linked to poor treatment outcomes for cancer chemotherapy is upregulated in OC.[42](#cam42560-bib-0042){ref-type=\"ref\"}, [43](#cam42560-bib-0043){ref-type=\"ref\"}, [44](#cam42560-bib-0044){ref-type=\"ref\"}, [45](#cam42560-bib-0045){ref-type=\"ref\"} Increased levels of c\u2010Met impacts cell proliferation, infiltration, angiogenesis, and endurance.[46](#cam42560-bib-0046){ref-type=\"ref\"}, [47](#cam42560-bib-0047){ref-type=\"ref\"}, [48](#cam42560-bib-0048){ref-type=\"ref\"} Antiapoptotic activity of c\u2010Met linked to chemo resistance for therapies.[46](#cam42560-bib-0046){ref-type=\"ref\"} Radiotherapy induces c\u2010Met expression and triggers the series of signals that increases the pro\u2010survival process and spreads the response of treatments.[49](#cam42560-bib-0049){ref-type=\"ref\"} An in vitro study has shown that by treating OC cells with c\u2010Met inhibitors, cell proliferation has been significantly reduced and increased apoptosis of cancer cells was observed.[50](#cam42560-bib-0050){ref-type=\"ref\"}\n\nA class of transcription factors known as specific proteins (Sp) regulates VEGF (vascular endothelial growth factor) expression with functional variation. Thus, Sp transcription factors have crucial impact on tumor expansion and metastasis.[51](#cam42560-bib-0051){ref-type=\"ref\"} Association of Sp transcription factors in anti\u2010cancer activity is illustrated in Figure [1](#cam42560-fig-0001){ref-type=\"fig\"}.\n\n![Association of Sp transcription factors in anti\u2010cancer activity. Small molecules like tolfenamic acid (TA), Mithramycin A are shown to inhibit specificity protein (Sp) family of transcription factors and will result in increased apoptosis of cancer cells](CAM4-8-7018-g001){#cam42560-fig-0001}\n\n7. MULTIPLE DRUG RESISTANCE {#cam42560-sec-0012}\n===========================\n\nResearch is in progress for understanding the mechanism of drug resistance in OC. Some of these mechanisms include increased DNA repair, overexpression of surface p 170\u2010glycoprotein, increased cellular levels of glutathione (GSH) and glutathione *S*\u2010transferases causing de\u2010toxification of platinum agents and taxol.[52](#cam42560-bib-0052){ref-type=\"ref\"} Cancer cells develop certain transport proteins that help them to eliminate the effective dosages of drugs from the cells causing multiple drug resistance (MDR).[53](#cam42560-bib-0053){ref-type=\"ref\"}, [54](#cam42560-bib-0054){ref-type=\"ref\"}\n\nThe \"Classical\" MDR is resulted from higher level of MDR\u20101 gene that code for 170\u2010kD ATP\u2010dependent glycoprotein Pgp. Pgp causes reduction of cellular levels of cytotoxic drugs within the cells by transporting the drug outside the cells against the concentration gradient. Chemotherapy often upregulates the expression of P\u2010gp on cancer cells resulting in MDR. Resistance to multiple drugs is associated with P\u2010gp overexpression and includes paclitaxel, vincristine, and doxorubicin.[55](#cam42560-bib-0055){ref-type=\"ref\"}, [56](#cam42560-bib-0056){ref-type=\"ref\"}, [57](#cam42560-bib-0057){ref-type=\"ref\"}, [58](#cam42560-bib-0058){ref-type=\"ref\"} Cells that do not express P\u2010gp acquire other methods for drug resistance. Amplification of MDR\u2010associated protein gene (MRP) has been found in such cells that encodes a protein MRP, which expelling the drug out of the cells.[58](#cam42560-bib-0058){ref-type=\"ref\"}, [59](#cam42560-bib-0059){ref-type=\"ref\"} The MRP1 coded by ABCC1 and MRP2 coded by ABCC2 genes. They induce resistance to many cancer drugs, especially with the widely used cisplatin in OC.[55](#cam42560-bib-0055){ref-type=\"ref\"}, [56](#cam42560-bib-0056){ref-type=\"ref\"}, [57](#cam42560-bib-0057){ref-type=\"ref\"}, [58](#cam42560-bib-0058){ref-type=\"ref\"} A schematic representation of the proteins involved in drug resistance mechanism of commonly used chemotherapeutic agents for OC is given in Figure [2](#cam42560-fig-0002){ref-type=\"fig\"}.\n\n![Dug resistance mechanisms: A schematic representation of proteins involved in drug resistance mechanism of commonly used chemotherapeutic agents for ovarian cancer chemotherapy. The classical multiple drug resistance produced by ABC transporters and non\u2010ABC transporters are illustrated](CAM4-8-7018-g002){#cam42560-fig-0002}\n\nIn yeast and mammalian system, copper transporter CTR1 is one of the transporters mediating uptake of platinum compounds.[60](#cam42560-bib-0060){ref-type=\"ref\"} Cisplatin reduces its cellular influx by rapid degradation of CTR1 resulting in drug resistance.[61](#cam42560-bib-0061){ref-type=\"ref\"}, [62](#cam42560-bib-0062){ref-type=\"ref\"} Increased CTR1 expression by the cells results in an increased platinum concentration and decreased resistance to cisplatin.[62](#cam42560-bib-0062){ref-type=\"ref\"} Combination of platinum drugs with bortezomib, a modulator for copper transporter expression, is a current option for platinum\u2010resistant solid tumors.[63](#cam42560-bib-0063){ref-type=\"ref\"}\n\nSeveral epigenetic changes have been observed in cisplatin\u2010resistant human cells that open new avenues to study drug resistance among the OC cells. It has been observed that certain individual cells among the cancer cell population attain the reversible state of drug tolerance to prevent the eradication of the population by potential lethal exposure.[64](#cam42560-bib-0064){ref-type=\"ref\"} Altering the chromatin state and engaging IGF\u20101, insulin\u2010like growth factor will increase drug resistance. Treating with Inhibitors of IGF\u20101 receptor will reverse this process and can be an important therapeutic strategy. In addition, by inactivating the cytotoxic genes like folate binding gene (FBP) in cancer cells, DNA hypermethylation plays critical role(s) in generating the multiple drug\u2010resistant phenotypes.[65](#cam42560-bib-0065){ref-type=\"ref\"}\n\nAnother well\u2010known carrier, breast cancer resistant protein coded by ABCG2 is found to be overexpressed in ovarian [66](#cam42560-bib-0066){ref-type=\"ref\"} and breast cancers.[67](#cam42560-bib-0067){ref-type=\"ref\"} Upregulated BRCP is known to protect cancer cells from topotecan [66](#cam42560-bib-0066){ref-type=\"ref\"}, [68](#cam42560-bib-0068){ref-type=\"ref\"} and mitoxantrone.[67](#cam42560-bib-0067){ref-type=\"ref\"}, [68](#cam42560-bib-0068){ref-type=\"ref\"} The mechanism of normal epithelial cells changes to drug resistant cancerous cells is by activating epithelial\u2010mesenchymal transition (EMT).[34](#cam42560-bib-0034){ref-type=\"ref\"} Beginning of their transformation malignant epithelial cells go through angiogenesis and massive propagation.[69](#cam42560-bib-0069){ref-type=\"ref\"} Cellular level changes occurring during the transition process transform epithelial cells to mesenchymal cells. These mesenchymal cells have anti apoptotic with increased migratory capacity and invasiveness.[70](#cam42560-bib-0070){ref-type=\"ref\"} E\u2010cadherin a suppressor of insensitivity and motility is downregulated by transcription factors like Twist, Snail and Slug which are key coordinators for EMT.[71](#cam42560-bib-0071){ref-type=\"ref\"} Snail and Twist are overexpressed in paclitaxel\u2010resistant EOC cells which is predicted by the molecular level modifications during EMT.[72](#cam42560-bib-0072){ref-type=\"ref\"}, [73](#cam42560-bib-0073){ref-type=\"ref\"}\n\nSince the EMT is mediated by several signaling pathways,[74](#cam42560-bib-0074){ref-type=\"ref\"} it has become clearer that by halting these pathways, EMT can be reverted as well as some biological effects like drug sensitivity.[75](#cam42560-bib-0075){ref-type=\"ref\"} Overexpression of endothelin\u20101 and endothelin A receptor has been shown to enable EOC cells with increased resistance to chemo\u2010drugs, and thus, increase their relative survival capacity.[76](#cam42560-bib-0076){ref-type=\"ref\"} In advanced\u2010stage EOC, ET (Endothelin) 1 and ET A receptor, ETAR pathways are overexpressed.[77](#cam42560-bib-0077){ref-type=\"ref\"} The ET\u20101 and ETAR are overexpressed with increased MAP kinase (MAPK) and protein kinase B phosphorylation, c\u20101ell proliferation, in drug resistant EOC cells.[76](#cam42560-bib-0076){ref-type=\"ref\"} In a study, treatment of cancer cells with the drugs that can block ETAR\u2010driven EMT, inhibition of tumor progression was seen, and chemo resistance has been overcome. EMT markers are used as a tool in several randomized clinical trials to develop personalized therapies. Clinical trial based on the aspirin treatment (NCT02602938) for metastatic breast and colorectal is an example of circulating tumor cells (CTC)s with EMT features.[78](#cam42560-bib-0078){ref-type=\"ref\"}\n\n8. IMMUNOTHERAPIES {#cam42560-sec-0013}\n==================\n\nImmunotherapy involves various methods enhancing immune system. Exploiting the immune system for tumor recession is an ancient procedure as in 2600 BC The Pharaoh Imhotep self\u2010infected to enable tumor recession.[79](#cam42560-bib-0079){ref-type=\"ref\"} Native and adaptive immunity of the patient are activated in the beginning of the tumor formation.[80](#cam42560-bib-0080){ref-type=\"ref\"}, [81](#cam42560-bib-0081){ref-type=\"ref\"} In the later stages of tumor growth, the tumor microenvironment inhibits the immune system in targeting cancer.\n\nAnti\u2010tumor lymphocytes from healthy adults and patients are used in treatment using adoptive cell transfer to stimulate cancer decline.[82](#cam42560-bib-0082){ref-type=\"ref\"} In this approach, the autologous T cells are collected from patient\\'s peripheral blood or resected tumor tissue or tumor\u2010infiltrating lymphocytes (TIL) and those cells are expanded or manipulated ex vivo, an environment different from the patients tumor microenvironment (TME). These T cells cultured ex vivo and recombinant interleukin 2 are given back to patients.[83](#cam42560-bib-0083){ref-type=\"ref\"}\n\nUse of cancer vaccines is another approach to bring about immune activation. T\u2010cell responses were stimulated by activating the antigen\u2010presenting cells.[84](#cam42560-bib-0084){ref-type=\"ref\"} More recently, 11 heavily treated patients with platinum\u2010resistant OC (PROC) have been treated with GL\u2010ONC1 on a phase Ib protocol \\[NCT02759588\\]. GL\u2010ONC1 is a modified vaccinia virus developed by Genelux Corporation (San Diego, CA) that causes tumor cell oncolysis, immune activation through release of oncoproteins, presentation of both foreign and tumor antigens by dendritic cells, and durable anti\u2010cancer T cell tumor\u2010specific memory. In this phase Ib trial \\[NCT02759588\\], patients received a minimum of three prior lines of therapy, with five patients having had at least five prior lines. Nine patients had PROC, one was platinum refractory, and one was intermediate platinum sensitive (7\u00a0months prior PFS). Ten patients progressed on their prior line of therapy and nine had ascites or pleural effusions. The trial involved intraperitoneal (IP) infusion of GL\u2010ONC1 monotherapy that was given at higher dosages and contained provisions for dose\u2010escalation every three patients. Two separate IP instillations were performed 24\u00a0hours apart through a tunneled catheter system. The primary objectives were measurement of toxicity and secondary endpoints were anti\u2010tumor response. Encouraged by these preliminary outcomes, a clinical trial for first dose cohort is currently underway.\n\nManipulation of immune checkpoints has become the modern revolution in cancer immunotherapy. Cytotoxic T\u2010Lymphocyte\u2010Associated Protein 4 (CTLA\u20104) and Programmed Cell Death Protein 1 (PD\u20101) are Immune checkpoint linked to T cells. These proteins control the equilibrium of immune response and tolerance upon influencing the T lymphocyte activity. Activated T lymphocyte functions as an inhibitor via \"negative feedback loop\" mechanism and protect normal tissues from tumor\u2010derived immune response.[85](#cam42560-bib-0085){ref-type=\"ref\"} These proteins are over expressed in OC patients and their natural anti\u2010cancer immunity is at disadvantage. Currently, several monoclonal antibodies (mAb) against CTLA\u20104 and other proteins and its ligand are used in clinics.[86](#cam42560-bib-0086){ref-type=\"ref\"}, [87](#cam42560-bib-0087){ref-type=\"ref\"}, [88](#cam42560-bib-0088){ref-type=\"ref\"} FDA approved mAb\\'s targeted against immune checkpoints for various cancers are given in Table [1](#cam42560-tbl-0001){ref-type=\"table\"}.\n\n###### \n\nCurrently approved monoclonal antibodies targeted against Immune checkpoint proteins\n\n  Drug name       Immune checkpoint target   Current approval as of June, 2019\n  --------------- -------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Ipilimumab      CTLA\u20104                     Melanoma, renal cell carcinoma (combined with nivolumab), colorectal cancer\n  Pembrolizumab   PD\u20101                       Nonsmall cell lung cancer (NSCLC), squamous cell carcinoma of head and neck (SCCHN), classic Hodgkin\\'s lymphoma, large B\u2010cell lymphoma, urothelial cancer, micro\u2010satellite instability\u2010 high (MSI_H)or mismatch repair deficient (dMMR) cancers, gastric or GEJ adenocarcinoma and cervical ca\n  Nivolumab       PD\u20101                       NSCLC, melanoma, RCC, classic HL, squamous cell carcinoma of head and neck (SCCHN), urothelial cancer (UC), MSI\u2010H or dMMR colorectal cancer, hepatocellular cancer\n  Avelumab        PD\u2010L1                      Merkel cell carcinoma (MCC), UC\n  Durvalumab      PD\u2010L1                      UC, NSCLC\n  Atezolizumab    PD\u2010L1                      UC, NSCLC\n  Tremelimumab    CTLA\u20104                     Awaiting approval\n\nAbbreviations: CTLA\u20104, cytotoxic T lymphocyte\u2010associated antigen 4; PD\u20101: programmed cell death 1; PD\u2010L1: programmed cell death ligand 1.\n\nSource: FDA Approved drugs:\n\nhttps://www.accessdata.fda.gov/scripts/cder/daf/index.cfm\n\n.\n\nJohn Wiley & Sons, Ltd\n\n9. IMMUNOTHERAPY FOR OC {#cam42560-sec-0014}\n=======================\n\nThe success rate for immunotherapy for OC treatment is very low and there is not yet any FDA approval for immune therapies for OC. Motivated by some of the recent encouraging results in other closely\u2010related tumor types, the scientific community has also started adapting immunotherapy to treat gynecologic cancers. Antibodies and T cells responsive for cancer are detected from ascites, blood, and tumor of advanced\u2010stage OC patients.[89](#cam42560-bib-0089){ref-type=\"ref\"} Since it is known that the TILs expression level is linked to increased survival rate in OC patients, immunotherapies are highly potential for effective treatment outcomes, similar to other cancers.[90](#cam42560-bib-0090){ref-type=\"ref\"} Table [S1](#cam42560-sup-0001){ref-type=\"supplementary-material\"} includes current clinical trials listed on <http://www.ClinicalTrial.gov> that are recruiting patients. Schematic representation for the emerging immunotherapies for OC is given in Figure [3](#cam42560-fig-0003){ref-type=\"fig\"}.\n\n![Emerging immunotherapies for ovarian cancer. A schematic representation with the details on cancer vaccines, dendritic cell vaccines, adoptive T\u2010cell transfer, immunostimulatory cytokines are some of the techniques explained in this review](CAM4-8-7018-g003){#cam42560-fig-0003}\n\n10. INHIBITORS AND MODULATOR {#cam42560-sec-0015}\n============================\n\nImportant factors that help predict the tumor responses to inhibitors for immune checkpoint are the effector immune cell availability, approachability, and the tumor dependence on immune checkpoint pathways. TILs and PD\u2010L1 are identified as the markers to predict immune response to inhibitors. Using these markers, higher than 50% of advanced SOC estimated to act in response to immune checkpoint inhibitors via adaptive immune resistance. It is nearly absent low\u2010grade SOC, about 25% in other pathological types cancers.[91](#cam42560-bib-0091){ref-type=\"ref\"} Currently, several immune checkpoint inhibitors are tested in animal studies and some of them are in clinical studies for OC treatment.[82](#cam42560-bib-0082){ref-type=\"ref\"}, [92](#cam42560-bib-0092){ref-type=\"ref\"}\n\nPreliminary data from clinical studies with all the current inhibitors reveals limited efficacy in OC with 15% rates of objective response, ORR. Hence it is very important future studies are required to identify more biomarkers for immune checkpoint inhibitors. 17% of ORR with a rate of 83% disease control is reported for a phase I study of olaparib (Poly ADP ribose polymerase \\[PARP\\] inhibitor) and durvalumab (anti\u2010PD\u2010L1).[93](#cam42560-bib-0093){ref-type=\"ref\"} Table [S3](#cam42560-sup-0003){ref-type=\"supplementary-material\"} shows the currently recruiting clinical trials (as of August, 2019) that utilizes immunotherapy for patients with OC.\n\n11. VACCINES USED IN OC THERAPY {#cam42560-sec-0016}\n===============================\n\nVaccines used in cancer therapy will activate the immune cells for the elimination of cancerous cells. Specific tumor\u2010associate antigens (TAAs) are administered using various methods. Some of the common vaccines tested as cancer vaccines are developed using various methods, epigenetic, and genetic. Vaccines are administered as it or along with cytokines or other accelerating factors.[94](#cam42560-bib-0094){ref-type=\"ref\"}\n\nCA125, p53 protein, folate receptor\u2010alpha (FR\u03b1), human epidermal growth factor receptor\u20102 (HER2), and cancer\u2010testis antigens, like melanoma\u2010associated antigen A4 (MAGE\u2010A4) and New York\u2010esophageal squamous cell carcinoma11 (NY\u2010ESO\u20101) are potential TAA molecules found in OC..[95](#cam42560-bib-0095){ref-type=\"ref\"} Cancer Vaccine therapeutic investigation is an actively growing area in OC research. Currently, there are mainly pilot and phase I or II trials on the use of therapeutic vaccines in OC.[85](#cam42560-bib-0085){ref-type=\"ref\"}, [96](#cam42560-bib-0096){ref-type=\"ref\"} A list of ongoing vaccine studies in OC (Source: <http://www.ClinicalTrial.gov>) is given in Table [S4](#cam42560-sup-0004){ref-type=\"supplementary-material\"}.\n\nNY\u2010ESO\u20101, a potential molecule for targeted vaccine due to its over expression in OC exhibited stimulated immune response specific to T cells.[97](#cam42560-bib-0097){ref-type=\"ref\"} Combination therapy of NY\u2010ESO\u20101 with DNA methylation inhibitors and chemotherapy was administered in patients with recurrent disease for therapeutic efficacy enhancement.[83](#cam42560-bib-0083){ref-type=\"ref\"} This increases the NY\u2010ESO\u20101 antibody availability, T\u2010cell responses that lead to clinical response in OC patients.\n\nAs noted earlier, oncolytic virus (OV) has shown synergy when combined with checkpoint inhibitor antibodies.[98](#cam42560-bib-0098){ref-type=\"ref\"} Oncolytic virus immunotherapy or combination of OV with other molecules that are immune\u2010stimulatory or induce immunogenic responses are encouraging avenues to explore as novel therapeutic options for OC.[82](#cam42560-bib-0082){ref-type=\"ref\"}\n\nHer2/Neu, tumor antigen presents in almost 90% of the recurrent OC cases. Clinical study of 11 patients were given Her2/Neu packed antigen autologous dendritic cells combined with antigens of telomerase reverse transcriptase (human) and pan\u2010DR peptides.[99](#cam42560-bib-0099){ref-type=\"ref\"} 90% 3\u2010year overall survival was reported as the outcome for this study of patients in remission with advanced OC.\n\nWhole tumor cell vaccines will induce immunologic reaction for a larger range of antigens compared to specific TAA.[100](#cam42560-bib-0100){ref-type=\"ref\"} Broader reaction with T cells can also be induced using whole vaccine.[100](#cam42560-bib-0100){ref-type=\"ref\"} Another important avenue in vaccine therapy is personalized peptide vaccines developed from individual tumor depending on the human leukocyte antigen and IgG expressions.[101](#cam42560-bib-0101){ref-type=\"ref\"} Clinical research showed that patient with Pt\u2010 sensitive cancer have 39.3\u00a0months overall survival rate. For patients with Pt resistant cancer overall survival rate is 16.2\u00a0months.\n\nAnother novel therapeutic method is to explore the use cancer cells from the patient to deliver viruses to the tumor.[102](#cam42560-bib-0102){ref-type=\"ref\"} Tumor microenvironment can be manipulated using cancer cells as virus vehicles. Changing a \"cold\" cancer to a \"hot\" cancer potentiates anti\u2010cancer reaction. This a promising strategy for OC patients who do not get benefitted from current therapies due to suboptimal immune infiltration. Schematic representation for the emerging immunotherapies for OC is given in Figure [3](#cam42560-fig-0003){ref-type=\"fig\"}.\n\n12. ADOPTIVE CELLULAR THERAPY\u2010ADOPTIVE T\u2010CELL TRANSFER {#cam42560-sec-0017}\n======================================================\n\nPeripheral blood lymphocytes (PBLs) are separated from patients' blood and will be used for the isolation of tumor\u2010specific lymphocytes. Tumor\u2010specific PBLs will be grown to supply back to the patient. Anti\u2010cancer action of PBLs can also be enhanced by genetic modification.[103](#cam42560-bib-0103){ref-type=\"ref\"} Clinical trials of ACT for OC are ongoing. ACT has shown around 72% reaction rates that last for more than three years in metastatic melanoma occurring at.[83](#cam42560-bib-0083){ref-type=\"ref\"} This is a very promising outcome that can be translated into OC therapy by optimizing the conditions for Adoptive cellular therapy (ACT) in OC therapy.\n\n13. CAR\u2010T\u2010CHIMERIC ANTIGEN RECEPTOR T {#cam42560-sec-0018}\n=====================================\n\nThe major limitation to the ACT trials in the beginning stages were the need for isolation and culturing of functional cancer responsive T cells. The emergence of engineered T cells has become a promising tool to enhance the cancer immune therapy.[82](#cam42560-bib-0082){ref-type=\"ref\"}, [104](#cam42560-bib-0104){ref-type=\"ref\"}, [105](#cam42560-bib-0105){ref-type=\"ref\"} Tumor\u2010specific targeting can be gained using patients receptors of T cell and CARs. The tumor recognition in a major histocompatibility complex (MHC) can be achieved by CAR\u2010T cells. T\u2010cell stimulation and selectivity of antigen features are combined in one combination molecule.[105](#cam42560-bib-0105){ref-type=\"ref\"} The initial group of CARs was examined in OC and other tumors inducing modest responses.[105](#cam42560-bib-0105){ref-type=\"ref\"}, [106](#cam42560-bib-0106){ref-type=\"ref\"} At the very first CAR T\u2010cells trial in OC, cancer load was not reduced in patients.[106](#cam42560-bib-0106){ref-type=\"ref\"} FR\u03b1,[107](#cam42560-bib-0107){ref-type=\"ref\"} HER\u20102,[108](#cam42560-bib-0108){ref-type=\"ref\"} CA125 (MUC16),[109](#cam42560-bib-0109){ref-type=\"ref\"}, [110](#cam42560-bib-0110){ref-type=\"ref\"} and mesothelin [111](#cam42560-bib-0111){ref-type=\"ref\"} are ensuring antigens for CARS. CAR\u2010T therapeutic efficacy still need to be improved in OC. Combination treatment modalities to overcome these problems may be a novel approach. Combining inhibitors for immune checkpoint with CAR\u2010T cells is a better therapeutic option for OC.[112](#cam42560-bib-0112){ref-type=\"ref\"} A schematic representation of emerging immunotherapies for OC is given in Figure [3](#cam42560-fig-0003){ref-type=\"fig\"}.\n\n14. FUTURE PERSPECTIVES {#cam42560-sec-0019}\n=======================\n\nRegardless of the extensive developments in OC therapy, it is still the deadliest malignancy in women. The biggest hurdle is the shortage of efficient screening procedure that helps to detect the tumor at an early stage. Even though the curing rate for beginning stage OC patient is 90%, about 20% of OC is detected as early as stage1. This reveals the need for future research for finding biomarkers that are more responsive and specific for detection of OC at an early stage.\n\nSurgery and chemotherapy are conventional therapy for ovarian carcinoma. Poor prognosis with a recurring progressive cancer is the major challenge to the treatment. The prognosis varies for each patient and it depends on the level of response to preliminary therapy. Drug administration by IP is the efficient way to target OC cells situated the peritoneal area. Removal of all the residual tissues by surgery followed by chemotherapy, is the most ideal cure for OC.\n\nSensitivity to chemotherapeutic agents is a crucial parameter in therapeutic efficacy. Research to identify biomarkers for apoptosis and chemo resistance in OC therapy should be one of the prime goals in cancer research (eg, Caris^\u00ae^ Assays). Newly identified microRNA biomarkers linked to platinum drug resistance are let\u20107[109](#cam42560-bib-0109){ref-type=\"ref\"} and ATP11B.[108](#cam42560-bib-0108){ref-type=\"ref\"} Another upcoming are of research in OC is combination therapy of drugs and other small molecules that can enhance therapeutic efficacy by increasing drug sensitivity and reducing drug resistance. Currently, several studies are in progress with the compounds that modify Bcl2 proteins family,[82](#cam42560-bib-0082){ref-type=\"ref\"}, [85](#cam42560-bib-0085){ref-type=\"ref\"} and agents for targeting DNA repair and inhibit PARP.[113](#cam42560-bib-0113){ref-type=\"ref\"}, [114](#cam42560-bib-0114){ref-type=\"ref\"}, [115](#cam42560-bib-0115){ref-type=\"ref\"} To improve drug sensitivity to cisplatin, small molecules like triethylenetetramine, genistein, butathione sulfoximine, and rapamycin are under the stage of preclinical testing.[116](#cam42560-bib-0116){ref-type=\"ref\"} Triethylenetetramine inhibits telomerase, induces anti\u2010antiogenesis and acts as anti\u2010cancer agent. Triethylenetetramine reversed cisplatin resistance in OC cells. Genistein showed anti\u2010cancer activity in both pediatric and adult cancer models. Triethylenetetramine sensitized OC cells against cisplatin. Butathione sulfoximine also showed sensitizing effects in gastric and OC cells against cisplatin. Laboratory studies also demonstrated the effect of rapamycin for inducing the effect of cisplatin in OC breast cancer and lung cancer cells.\n\nAnother promising technique for OC therapy is silencing gene expression. By this technique, specific sets of genes can be targeted and altered, and it requires less dosage. This technique is still the subject of ongoing research to overcome its challenges like stability and compound delivery to a target site.[110](#cam42560-bib-0110){ref-type=\"ref\"}, [111](#cam42560-bib-0111){ref-type=\"ref\"} MicroRNAs are used as targeted molecules for diagnosis is another growing field. Developing clinical trials with more molecules similar to what we explained above will help to overcome the challenges in OC therapy.\n\nDiab et al have done a review of targeted therapy for OC for 2010\u20102017.[117](#cam42560-bib-0117){ref-type=\"ref\"} Targeted therapies are evolving in three main fields, angiogenesis, signaling, and apoptosis. The VEGF pathway is focused for angiogenesis. PI3K/Akt and the MAP kinase pathways critically involved signaling cascades. McCabe et al[118](#cam42560-bib-0118){ref-type=\"ref\"} have conducted examination of trials in EOC to investigate the association between platinum\u2010resistance and response to anti\u2010angiogenic agents. The analysis revealed that novel anti\u2010angiogenic therapies would be beneficiary for the patients with platinum\u2010resistant EOC.\n\nDose\u2010dense chemotherapy is the promising option currently available for the patients with poor responses to chemotherapy. PARP inhibitors are the most emerging class of new drugs in combination therapy with the traditional chemotherapy drugs listed in Figure [3](#cam42560-fig-0003){ref-type=\"fig\"}. Bevacizumab is recently approved for EOC treatment. The common challenge with anti\u2010angiogenic agents is the non\u2010availability of efficient biomarkers. Folate receptor targeting required further research to consider as one of the treatment options. Despite cost issues, regular Breast cancer susceptibility gene (BRCA) screening should be done for all OC patients for a better selection of targeted therapy. Knowledge about tumor microenvironment and immune suppressive pathways are crucial for newer immunotherapeutic approaches toward OC. For a personalized medical treatment, systematic data analysis of molecular and genetical categorization of various types of OC with precision is required.[119](#cam42560-bib-0119){ref-type=\"ref\"} A schematic representation of some of the most common strategies for improving therapeutic responses is given in Figure [4](#cam42560-fig-0004){ref-type=\"fig\"}.\n\n![Current strategies for improving therapeutic response. Effective Targeted therapy, Usage of PARP inhibitors, combination therapy, immunotherapy, and usage of chemosensitizers are some of the future strategies for improved therapeutic responses for ovarian cancer treatment](CAM4-8-7018-g004){#cam42560-fig-0004}\n\n15. CONCLUSIONS {#cam42560-sec-0020}\n===============\n\nOvarian cancer affects the lives of many women around us. Despite continued efforts and steady improvements in treatment over the past few decades, OC still remains the deadliest malignancy in women. The poor clinical outcome is due to the deficiency of effective tools for detecting the disease at an early stage, chemotherapy resistance and increased heterogeneity of the disease. The vast majority of cases have high\u2010grade papillary serous histology marked by p53 mutations and 25% of cases have either inherited or acquired mutations in BRCA. Primary therapy is initiated with cytoreductive surgery and chemotherapy. Even with optimization of treatment protocols that have improved PFS, only limited gains in OS. Ultimately, approximately 80% of patients develop PROC. Once this occurs, further chemotherapy response rates are about 10%\u201015% and survival averages 9\u201012\u00a0months. Therefore, we are critically in need of developing novel therapies to improve cure rates and provide effective long\u2010term disease stability for PROC.\n\nTargeted therapy is the fast growing modalities for cancer treatment. For targeted therapy drugs or small molecules will be used to block tumor growth. More studies should be done on combination therapies involving one or more of these small molecules as modulators for OC treatment. Such research should be augmented to fight chemo resistance better treatment outcomes. Inhibitors for various genes involved in the signal pathway in tumor growth should be another area of focus for future studies for OC treatment. Study of the various molecules involved in tumor micro environment at various stages of tumor metastasis is crucial for the development of better immunotherapies for OC.[120](#cam42560-bib-0120){ref-type=\"ref\"} How each one of these molecules control various treatment strategies and the immune system of the patient. Immune and cellular therapies coupled with genetic testing and precision assays (biomarkers) are promising strategies for better clinical outcomes. Novel strategies and rapid growth of research in medical field will lead to better therapeutic schemes to minimize ill health and improved life expectancy for patients with OC.\n\nCONFLICT OF INTEREST {#cam42560-sec-0021}\n====================\n\nNone declared.\n\nSupporting information\n======================\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n"}
+{"text": "Heart failure--Definition {#sec1-1}\n=========================\n\nHeart failure is a syndrome in which structural or functional cardiac conditions impair heart\\'s ability to supply sufficient **blood flow** in order to meet the body\\'s needs, or to do that at an elevated diastolic pressure\\[[@R2]\\].\n\nThere are many definitions of this complex syndrome, but none is satisfactory, due to the lack of a universally agreed definition and challenges in definitive diagnosis. Until now, only some selective features of this extremely complex physiological state were highlighted in the definitions--oxygen consumption, cardiac preload and afterload, left ventricular remodeling and dysfunction, ventricular filling pressures, neurohormonal responses, exercise capacity, etc.\n\nThe new American and European guidelines and recommendations include new information and have the declared intention to simplify and clarify the previous recommendations\\[[@R1]\\].\n\nHeart failure is a clinical syndrome in which patients have featured symptoms typical of heart failure (breathlessness at rest or on exercise, fatigue, tiredness, ankle swelling) and typical signs of heart failure (tachycardia, tachypnoea, pulmonary rales, pleural effusion, raised jugular venous pressure, peripheral oedema, hepatomegaly) and objective evidence of a structural or functional abnormality of the heart at rest (cardiomegaly, third heart sound, cardiac murmurs,abnormality on the echocardiogram, raised natriuretic peptide concentration)\\[[@R1]\\].\n\nA clinical response to a pharmacological therapy directed to heart failure is not sufficient for the diagnosis of heart failure, although the usefulness/efficacy of the treatment may be established by the improvement in symptoms or signs (e.g. diuretic administration)\\[[@R3]\\].\n\nHeart failure may be classified by structural abnormality (ACC/AHA), or by symptoms relating to functional capacity (NYHA).\n\nACC/AHA stages of heart failure (based on structure and damage to heart muscle)\\[[@R4]\\]\n\n-   Stage A: At high risk for developing heart failure. No identified structural or functional abnormality; no signs or symptoms.\n\n-   Stage B: Developed structural heart disease that is strongly associated with the development of heart failure, but without signs or symptoms.\n\n-   Stage C: Symptomatic heart failure associated with underlying structural heart disease.\n\n-   Stage D: Advanced structural heart disease and marked symptoms of heart failure at rest despite maximal medical therapy.\n\nNYHA functional classification (severity based on symptoms and physical activity)\\[[@R5]\\] (NYHA classification refers to stages C and D)\n\n-   Class \u2160:No limitation of physical activity. Ordinary physical activity does not cause fatigue, palpitation, or dyspnoea.\n\n-   Class \u2161: Slight limitation of physical activity. Comfortable at rest, but ordinary physical activity results in fatigue, palpitation, or dyspnoea.\n\n-   Class \u2162: Marked limitation of physical activity. Comfortable at rest, but less than ordinary activity results in fatigue, palpitation, or dyspnoea.\n\n-   Class \u2163: Unable to carry on any physical activity without discomfort. Symptoms at rest. If any physical activity is undertaken, discomfort is increased.\n\nDescriptive terms in heart failure {#sec1-2}\n==================================\n\nAcute and chronic heart failure\n-------------------------------\n\nAcute heart failure (ICA) is a clinical syndrome caused by the action of a factor with brutal effect, often reversible, over the functional capacity of the heart.\n\nAcute heart failure is defined by the rapid onset of signs and symptoms (secondary to cardiac dysfunction) resulting from impaired heart. It may occur in the presence or absence of preexisting heart disease. Acute heart failure may be an expression of systolic or diastolic dysfunction, heart rhythm abnormalities, or disturbances of preload or afterload. It is often a threat, life threatening, requiring emergency treatment.\n\nAcute heart failure may present as acute de novo heart failure (a patient without known preexisting heart disease) or acute decompensation of chronic heart failure. In practice, the most common form is decompensation of chronic heart failure.\n\nClassification of heart failure\\[[@R1]\\]\n\n-   New onset: First presentation, Acute or slow onset\n\n-   Transient: Recurrent or episodic\n\n-   Chronic: Persistent and Stable, worsening, or decompensated\n\nOther forms of acute heart failure (ICA) include: acute heart failure with hypertension, pulmonary edema, cardiogenic shock, heart failure with increased cardiac output and right heart failure.\n\nThere is a number of well-known classifications that are used in the context of acute heart failure secondary to a myocardial infarction: Killip classification (designed to clinically assess the severity of myocardial dysfunction)\\[[@R6]\\] and Forrester classification (evaluate clinical and hemodynamic patients with acute myocardial infarction)\\[[@R7]\\]. In the original publication \\[[@R8]\\] therapeutic strategy depended on clinical and hemodynamic status -- Forrester classification. Patients are classified according to signs of peripheral hypoperfusion (weak pulse, moist skin, cold, peripheral cyanosis, hypotension, tachycardia, confusion, oliguria) and to signs of pulmonary congestion (rales, chest X--ray changes). The mortality rate is different according to the class, respectively 2.2% in Class \u2160, 10.1% in Class \u2161, 22.4% Class \u2162, to 55.5% in Class \u2163.\n\nKillip classification \\[[@R6]\\]: Designed to provide a clinical estimate of the severity of circulatory derangement in the treatment of acute myocardial infarction.\n\n-   Stage \u2160: No heart failure. No clinical signs of cardiac decompensation (PCWP--estimate of left atrial pressure)\n\n-   Stage \u2161: Heart failure. Diagnostic criteria include rales, S3 gallop, and pulmonary venous hypertension. Pulmonary congestion with wet rales in the lower half of the lung fields.\n\n-   Stage \u2162: Severe heart failure. Frank pulmonary oedema with rales throughout the lung fields\n\n-   Stage \u2163: Cardiogenic shock. Signs include hypotension (SBP b90 mmHg), and evidence of peripheral vasoconstriction such as oliguria, cyanosis and sweating\n\nForrester classification\\[[@R7]\\]; Designed to describe clinical and haemodynamic status in acute myocardial infarction.\n\n1.  Normal perfusion and pulmonary wedge pressure\n\n2.  Poor perfusion and low PCWP (hypovolaemic)\n\n3.  Nearly normal perfusion and high PCWP (pulmonary oedema)\n\n4.  Poor perfusion and high PCWP (cardiogenic shock)\n\nAnother classification of severity of acute heart failure, used in some intensive care and coronary units, has been validated for cardiomyopathy\\[[@R9]\\],based on clinical signs\\[[@R10]\\], applicable in case of chronic decompensated heart failure\\[[@R9]\\].\n\nPatients are divided into 4 classes based on two clinical criteria: assessment of perfusion/peripheral circulation and skin appearance and pulmonary congestion (appreciated by auscultation).\n\n-   Class \u2160 warm skin without rales\n\n-   Class \u2161 warm skin, rales present\n\n-   Class \u2162 cold skin, without rales\n\n-   Class \u2163 cold skin, rales present\n\nThis classification allows an accurate assessment of prognosis regarding the patients with cardiomyopathy\\[[@R11]\\].\n\nHeart failure is a clinical and functional diagnosis, expression of supply / demand balance alteration; a relation in which both terms are equally important. A simple, objective, definition of chronic heart failure is difficult as long as there are no precise boundaries between ventricular and heart dysfunction (intracavitary pressure limit values, changes of flow, size and volume cavities)\\[[@R3]\\].\n\nDiastolic versus systolic heart failure\n---------------------------------------\n\nThe cardiologist\\'s interest all over the world has been directed, for a long time, to the study of the systolic function of heart, being considered the key in the pathophysiology of heart failure. In the last decade, numerous studies have attempted to shed light on natural history, pathophysiology, diagnosis, prognosis and treatment of heart failure with diastolic dysfunction.\n\nEpidemiological and clinical studies applied on hospital populations revealed that a percentage of 30--50% of patients with heart failure have preserved left ventricular function. Epidemiological data in patients with diastolic dysfunction are still very limited compared to available data on epidemiology of heart failure with systolic dysfunction, much more studied and well--documented.\n\nDiastolic heart failure is characterized by 3 basic elements: the presence of signs and symptoms of heart failure, preserved left ventricular systolic function (ejection fraction \\>45%) and the presence of abnormal diastolic function (alterations of relaxation and / or compliance) \\[[@R12]\\].\n\nWe must make a clear delimitation between two different concepts: diastolic dysfunction and diastolic heart failure. Diastolic dysfunction characterized by abnormal mechanical properties and diastolic heart failure, is a clinical syndrome characterized by signs and symptoms of heart failure and evidence of altered diastolic function.\n\nThe concept of diastolic dysfunction defined by the existence of an abnormality of one or both of the constituent processes of diastole: relaxation and compliance. They may suffer elongation, delays or they may be incomplete, changes that can be measured with Doppler echography or cardiac catheterization\\[[@R13]\\]. These abnormalities of diastolic function may be detected in the presence or absence of the clinical syndrome of diastolic heart failure; they can also be noticed alone or with concomitant abnormalities of systolic function \\[[@R14], [@R15]\\].\n\nPopulation studies estimate that approximately 5% of the general population have isolated diastolic dysfunction; systolic dysfunction has a similar rate, about half being asymptomatic\\[[@R16]\\].\n\nThe main etiology of heart failure in developed societies is ischemic heart disease. In this context, many signs of heart failure are associated with left ventricular systolic dysfunction, although diastolic impairment at rest is usually present\\[[@R3]\\].\n\nDiastolic dysfunction in young patients is rare (about 15% in populations under 50 years), but it increases in importance in the elders (50% in patients over 70 years) \\[[@R17]--[@R22]\\].\n\nDiastolic dysfunction is more common in women in whose systolic hypertension and myocardial hypertrophy contribute to cardiac dysfunction\\[[@R23], [@R24]\\].\n\nAlthough most patients with acute pulmonary edema have a systolic dysfunction, many cases of developing a clinical picture of acute pulmonary edema in patients with diastolic heart failure with preserved systolic function are described in literature\\[[@R25]\\].\n\nAmong the patients with chronic heart failure, a significant percentage has diastolic heart failure with preserved systolic function, if we take into account left ventricular ejection fraction at rest\\[[@R23], [@R26]\\].\n\nDiastolic and systolic heart failure cannot be considered as separate pathophysiological entities, they coexist in most patients with the dominance of one or the other\\[[@R3]\\]. Most patients with systolic dysfunction associate changes in diastolic function \\[[@R3]\\]. In some patients with hypertension or hypertrophic cardiomyopathy, the dominant opinion is that diastolic dysfunction and abnormalities of relaxation and compliance are rather an earlier and more sensitive marker of cardiac damage than the alteration of systolic function, measured by decreased ejection fraction\\[[@R3], [@R23], [@R27]\\].\n\nOther descriptive terms in heart failure\n----------------------------------------\n\nA number of other descriptive terms was frequently used in the past, but now they are used only occasionally, their clinical value being reduced. These terms do not provide information on the etiology and they have a reduced influence in establishing a modern heart failure therapy.\n\nThe terms of right and left heart failure, forward and backward HF, high and low output HF, mild, moderate, or severe HF are old terms used somehow arbitrary, expressing an imprecise measure of this complex clinical syndrome.\n\nCurrently, the assessment of dyspnea and the limitation of the daily work are done by classifying the patient in functional class NYHA--New York Heart Association and the decline of functional assessment is performed using the Katz ADL Scale-Activities of Daily Living\\[[@R28]\\].\n\nDescriptive epidemiology {#sec1-3}\n========================\n\nThe importance and extent of the problem. Incidence. Prevalence\n---------------------------------------------------------------\n\nHeart failure is a clinical syndrome, which, epidemiologically, is in a continuous growth in the economic developed countries and the developing ones, in contrast to ischemic heart disease, which is declining in the developed countries, due to the prevention programs. This fact is due to an aging population and to significant advances in the diagnosis and pharmacological treatment of heart disease.\n\nCongestive heart failure (CHF) is a major public healthcare problem, being an invalidating condition, with a bad long--term prognosis, which reduces the quality of life and has a considerable economic cost to the individual and the society in general.\n\nCurrent estimates show that CHF affects over 5 million people in the U.S., approximately 2% of adults and 10% of elderly people beeing affected by this 'cardiovascular epidemic' which was designated as national research priority\\[[@R29]\\].\n\nWith an incidence of over 400,000 new cases diagnosed each year\\[[@R17]\\], and approximately 1 million hospital admissions annually, out of which over 80% patients aged over 65 years\\[[@R17], [@R30]\\], heart failure is the only major cardiovascular disease that increases in the United States.\n\nToday it is estimated that the prevalence in the European countries ranges between 0.4 and 2%\\[[@R31]\\], the studies estimating that about 14 million of the approximately 900 million inhabitants of the 51 European countries suffer from heart failure.\n\nThe prevalence of heart failure increases with age\\[[@R32]\\], the average age of this population being 74--75 years and it is going due to increase in time, the data for the elderly population show percentage increased prevalence of disease\\[[@R33]--[@R36]\\].\n\nThe large number of studies carried out on large lots of heart failure patients shows the attention given to heart failure in clinical practice. We must note that most studies that refer to therapeutic management addresses to patients with heart failure and systolic disfunction, fewer to patients with preserved systolic function and heart failure.\n\nIn the last decade special attention has been paid to diastolic heart failure\\[[@R30]\\], pathophysiological condition in which heart failure syndrome is caused by abnormalities of diastolic ventricular function. Epidemiological and clinical studies conducted on hospital populations revealed that a percentage of 30--50% of patients with CHF have preserved left ventricular function. Epidemiological data on diastolic dysfunction are still very limited compared to available data on epidemiology of heart failure with systolic dysfunction, carefully studied and well--documented.\n\nDiastolic heart failure receives increasing interest from clinicians around the world because its clinical and epidemiological significance proved to be growing\\[[@R37]\\]. In the last decade a series of data has been accumulated that provide conclusive information on the epidemiology of diastolic dysfunction, about its magnitude and complexity. The studies performed in the recent years revealed an annual mortality rate of patients with preserved systolic function heart failure of 8--17%\\[[@R38]\\] and 9--28%\\[[@R39]\\]. Although this is a significantly lower mortality rate than that of patients with heart failure and systolic dysfunction (about half), it represents 3--4 times the mortality of patients with the same age enrolled in control group\\[[@R39]\\].\n\nAge and prevalence of heart failure\n-----------------------------------\n\nAs mentioned above the prevalence of heart failure increases with age\\[[@R30]\\], the average age of this population being 74--75 years. Because of the increasing proportion of elderly population an increase in the prevalence of this disease is registrated\\[[@R33]--[@R36]\\]. Data from literature show an increase in prevalence from 6.6% in 65--69 years age group to 14% in over 85 years age group\\[[@R29]\\].\n\nLooking at the percentage of patients with diastolic heart failure from the people with symptomatic CHF we can notice a clear increase with age. Diastolic HF percentage varies from a rate of approximately 15% in populations under 50 years, to 33% in the 50--70 years age group and 50% in patients over 70 years\\[[@R17], [@R18], [@R21], [@R22], [@R40], [@R41]\\].\n\nRecent clinical studies in patients with CHF and population observations estimate that heart failure with preserved systolic function represents approximately 30--50% of cases hospitalized for symptomatic heart failure and more than 50% for people over 65 years\\[[@R26], [@R42]--[@R44]\\].\n\nIncreasing prevalence of diastolic dysfunction in the elderly people is due to: age related to physiological changes (increased interstitial fibrosis and myocardial hypertrophy, which causes alteration of ventricular relaxation and compliance) and induced pathophysiological changes associated comorbidities -- essential hypertension (most important factor), ischemic heart disease, diabetes and left ventricular hypertrophy. With age there is a higher prevalence of these factors which may explain an increased percentage of patients with distolic dysfunction in the elderly people\\[[@R45]\\]. The risk factors mentioned above determine structural and functional heart adaptations that along with age--induced changes lead to alterations in diastolic function.\n\nSex, etiology and the prevalence of heart failure\n-------------------------------------------------\n\nIn younger age groups, patients under 50 years, heart failure as a whole (patients with systolic dysfunction and diastolic dysfunction) is more common in men. At this age the most common etiology is ischemic coronary disease, more common in males. The main causes of heart failure are: coronary heart disease (70% of patients with HF), valve disease and cardiomyopathies each with approximately 10% of cases. In elderly people prevalence is equal between the sexes\\[[@R1]\\].\n\nNumerous epidemiological reports have noticed that women with heart failure have an increased percentage (than men) of preserved systolic function\\[[@R40], [@R44], [@R46], [@R47]\\].\n\nIn similar age groups and associated clinical conditions female patients have more diastolic heart failure than men. Out of the 2.4 million female patients with heart failure in the U.S. more than 50% of them have normal systolic function \\[[@R40]\\].\n\nThere is no registry of heart failure in Romania as there is in other countries, so there is no precise data on the epidemiology of heart failure. Older data estimate a heart failure prevalence of 0.2--0.5% in the general population, that means approximately 150,000--200,000 patients with heart failure diagnosis\\[[@R48]\\]. Recent data from Statistics by Country for congestive Heart Failure\\[[@R49]\\] estimate an incidence of 0.146% (32 875) cases of heart failure and a prevalence of 1.76% (394 509) cases of heart failure in Romania. The President of the Romanian Heart Failure Working Group estimates the number of heart failure patients, at approximately 800--900 000, a number that increases permanently\\[[@R50]\\].\n\nMorbidity. Quality of life. Costs {#sec1-4}\n=================================\n\nHF morbidity is particularly important, patients with heart failure require frequent medical visits at home or rehospitalization, which represent a significant expenditure of health resources. In the first year after hospital discharge approximately 50% of patients with heart failure require rehospitalization, the data are similar in those with systolic dysfunction or diastolic dysfunction.\n\nHeart failure is the main cause of hospitalization in the Medicare population in the USA\\[[@R51]\\]. Data from Scotland show that the number of hospitalizations in which heart failure is primary or secondary diagnosis of hospitalization is increasing in Europe. According to a hospital record, 4.7% of women admissions and 5.1% of men admissions are due to heart failure (at any time to diagnose it)\\[[@R52]\\]. While only a few cases are due to acute heart failure, first presentation, most cases are due to decompensation of chronic heart failure. The HF incidence, regardless of the degree of severity, varies between 2.3 and 3.7 per 1000 inhabitants per year\\[[@R53]--[@R54]\\].\n\nEpidemiological data show that morbidity is decisively influenced by age, rehospitalization rate (at one year) increases from 25% for under 50 years age group, to 50% for 50--70 years age group \\[[@R18]\\].\n\nSome European studies show that about 1% of the national health budget is allocated to the therapy of heart failure, in the U.S. about 2% of the national health budget is allocated to this problem. The large number and the long length of hospitalization for acute heart failure or decompensation of chronic heart failure are a substantial economic burden for the healthcare budgets. In U.S., the first cause of hospitalization in patients over 65 years is heart failure. In Scotland, the number of patients with heart failure in hospital increased by 60% between 1980 and 1990. The United States spends about 20 billion dollars annually, 10% of healthcare budget allocated for the management of cardiovascular disease with heart failure, 75% of the amount being allocated to hospital care. Heart failure is the most expensive cardiological syndrome \\[[@R56],[@R57]\\].\n\nPrognosis. Mortality {#sec1-5}\n====================\n\nPrognosis is heterogeneous and depends on the heart failure class, its etiology and the patient\\'s age. The prognosis of untreated heart failure is unknown. Multiple studies and metaanalyses highlighted various predictive criteria: clinical--radiological (NYHA functional class and heart size on chest X--ray examination), presence of atrial fibrillation, ejection fraction of VS, maximal O2 consumption during exercise, 6--minute walk test, pulmonary capillary pressure, serum catecholamine, natriuretic peptides and ventricular arrhythmias.\n\nThe latest ESC guidelines summarize the knowledge about the conditions associated with a poor prognosis in heart failure:\n\n###### \n\nConditions associated with a poor prognosis in heart failure--according to ESC guidelines 2008\\[[@R1]\\](\\*Powerful predictors)\n\n  Demographics                  Clinical                             Electrophysiological                Functional/exertional          Laboratory                                       Imaging\n  ----------------------------- ------------------------------------ ----------------------------------- ------------------------------ ------------------------------------------------ ------------------------------------------------------------\n  Advanced age\\*                Hypotension\\*                        Tachycardia Q waves                 Reduced work, low peak         Marked elevation of BNP/NT pro--BNP\\*            Low LVEF\\*\n  Ischaemic etiology\\*          NYHA functional class \u2162--\u2163\\*         Wide QRS\\*                          VO2\\*                          Hyponatraemia\\*                                  Increased LV volumes\n  Resuscitated sudden death\\*   Prior HF hospitalization\\*           LV hypertrophy                      Poor 6--minute walk distance   Elevated troponin\\*                              Low cardiac index\n  Poor compliance               Tachycardia                          Complex ventricular arrhythmias\\*   High VE/VCO2 slope             Elevated biomarkers, neurohumoral activation\\*   High LV filling pressure\n  Renal dysfunction             Pulmonary rales                      Low heart rate variability          Periodic breathing             Elevated creatinine/BUN                          Restrictive mitral filling pattern, pulmonary hypertension\n  Diabetes                      Aortic stenosis                      Atrial fibrillation                                                Elevated bilirubin                               Impaired right ventricular function\n  Anaemia                       Low body mass index                  T--wave alternations                                               Anaemia                                          \n  COPD                          Sleep--related breathing disorders                                                                      Elevated uric acid                               \n  Depression                                                                                                                                                                             \n\nFramingham\\'s study reported a 65% survival rate at 1 year after specifying the diagnosis of heart failure, respectively 25% in men, and, 35--40% in women, meaning 5 years survival. Another study--Mayo Clinic--Minnesota, shows a similar 66% one--year survival after the diagnosis indication. The main causes of death are sudden death or terminal heart failure (50% of the cases). Depending on the functional NYHA class in which the patients are reported, different mortality rates are recorded: CONSENSUS \u2160 studies elderly patients with heart failure; NYHA class \u2153, reports a mortality rate of 44% at 6 months, and the \u2164Heft and SOLVD study shows a mortality rate of 15--20% at 1 year, in patients in NYHA class \u2161 and \u2162.\n\nPatients with acute heart failure have a very severe prognosis, large randomized trials with hospitalized patients for decompensated heart failure have shown a 9.6% mortality rate at 60 days and a 35.2% combined mortality rate and rehospitalization at 60 days\\[[@R58], [@R59]\\].\n\nIschemic etiology proves to be a negative prognostic factor in AIRE and TRACE studies on patients with post--myocardial infarction heart failure, a 14--25% mortality rate in the 1 year registered in these patients. Mortality is particularly high in patients with acute myocardial infarction associated with severe HF, with a mortality rate close to 30% in the 1 year\\[[@R59], [@R60]\\].\n\nImportant observational studies that include patients with both systolic and diastolic heart failure show that although short--term mortality rate may be lower in patients with diastolic dysfunction, long--term survival analysis shows a similar mortality rate for both groups of patients, especially those over 65 years\\[[@R14], [@R46]\\]. Different results regarding mortality in diastolic heart failure may be explained by differences in etiology and age of the patients taken into survey.\n\nAnother determinant of mortality is the age, mortality in diastolic heart failure increases significantly with age. The data show that mortality at 5 years is 15% in below the 50--year--old group, 33% in 50--70 year--old group, reaching 50% in patients over 70 years old. Thus, in the elderly people over 70 years old the mortality rate for heart failure diastolic and heart failure systolic is practically equivalent\\[[@R15], [@R22], [@R46]\\].\n\nPatterns of morbidity and mortality in Romania have undergone significant changes in the recent decades, due to the increased incidence and prevalence of chronic disease and mortality associated with these causes, in the context of the increasing percentage of elderly population, coupled with the action of multiple biological, environmental, behavioral risk factors as well as the influence of the socio--economic conditions and healthcare.\n\nCompared to the average of the first 15 states included in the European Union (EU--15) and the average of the new member states (NMS--10), the overall mortality, has shown a slow downward trend (EU--15 average was 9.93 deaths per 1000 inhabitants in 1999 and 9.67 deaths per 1000 inhabitants in 2001, and the NMS--10 average was 10.78 deaths per 1,000 inhabitants in 1999 and 10.36 deaths per 1,000 inhabitants in 2002)\\[[@R61]\\]. Romania is recording a substantial increase in overall mortality rate, which, combined with the dynamics of birth and fertility leads to a long--term aging population process, with major negative consequences for both the healthcare system and the social security one.\n\nThere are not sufficient epidemiological data yet on diastolic heart failure in Romania, some studies show a high mortality rate, at 1 year up to 29% in the diastolic heart failure lot and 33% in systolic (note patients included in the study are severe cases that required hospitalization in an emergency hospital)\\[[@R62]\\].\n\nPrognosis of individual patients differs considerably, outcomes in highly variable trial data often do not give an adequate direction; taking into consideration the magnitude of this syndrome in the society and its complexity, we need a model to predict the risk of death, to estimate survival of heart failure patients.\n\nSurvival analysis is a branch of statistics dealing with the life eextent of biological organisms. More generally, the survival analysis involves estimating the duration of survival until a certain event. In this context, in the specialized literature devoted to survival analysis, death is considered an 'event'. Survival analysis seeks to answer the questions related to the survival of an individual or a lot from a population, over a certain period of time, and which are the causes of increase or decrease in the extent of survival.\n\nA key element in this area is the **survival function**, usually noted by S and defined as: S(t)=Pr(T\\>t) in which--t represents one moment in time; T is a random variable designating the time (or age) of death;Pr() notes the probability of occurrence of an event. Pr(T\\>t) means the probability that the time of death T occurs some time after the specified time, t.\n\nUsually, survival function S(t) has the following two properties: S(0)=1; S(t~2~)\\< S(t~1~) for the moments of time t~2~\\>t~1~. Last property shows that the survival function is monotonously decreasing in time. Survival at a later age is only possible if the subject has survived to all younger ages. Also, for the usual mathematical functions used for survival models, the value of S (t) function tends to 0 (zero) at a time when t tends to infinity.\n\nIn connection with the survival function we can define the **lifetime distribution function** (life extension) by the relation: F(t)=Pr(T\\<t)=1--S(t). Pr(T\\<t) means the probability that death time T occurs before or at a specified time t, at the latest.\n\nAnother extremely important concept is the **hazard function**, conventionally denoted by h (t). Hazard function can be defined in mathematical terms as in the relation [Figure 1](#F1){ref-type=\"fig\"}:\n\n![Hazard Function](JMedLife-03-421-g001){#F1}\n\nin which S\\'(t) represents time derivative of survival function. Hazard function is the mortality rate at time t conditioned by survival until time t or later (T\\>t). Hazard function is non negative, this meaning that h(t)\\>0.\n\nSurvival models commonly use hazard function or hazard function logarithm. For example, a parametric model of the hazard function logarithm, based on a multiple linear distribution can be written as [Figure 2](#F2){ref-type=\"fig\"} or [Figure 3](#F3){ref-type=\"fig\"}:\n\n![A parametric model of the hazard function logarithm, based on a multiple linear distribution](JMedLife-03-421-g002){#F2}\n\n![Another form of the hazard function logarithm, based on a multiple linear distribution](JMedLife-03-421-g003){#F3}\n\nIn these equations the index i represents the number of observation, the index n is the number of independent variables noted x~i1~, x~i2~, ..., x~in~ and the a~1~, a~2~,..., a~n~, are the coefficients of the model. In this relation alpha is a kind of reference function/value (baseline function), because logh~i~(t)=alpha when all independent variables are zero. The purpose of multiple regression is to highlight the relationship between a dependent variable and many independent variables (control variables, predictors).\n\nOne of the most used functions of survival is the Cox multiple regression model \\[[@R2]\\], proportional hazards model.\n\n![Cox\\'s survival model](JMedLife-03-421-g004){#F4}\n\nIn this case we have:\n\n![](JMedLife-03-421-g005){#F5}\n\n![](JMedLife-03-421-g006){#F6}\n\nwhich will result in a hazard function[Figure 8](#F8){ref-type=\"fig\"}:\n\n![](JMedLife-03-421-g007){#F7}\n\nh~0~(t)is called baseline hazard function. In Cox\\'s model it is unspecified.\n\nTaking two observations, i and j in equation[Figure 8](#F8){ref-type=\"fig\"}, the following formula results\n\n![](JMedLife-03-421-g008){#F8}\n\nSince hazard functions as hazard ratio for any pair of observations, according to [Figure 8](#F8){ref-type=\"fig\"}, it does not depend on time t, Cox\\'s model is often referred to as proportional hazards model.\n\nCox\\'s model can be actually extended to include time--dependent variable x. In this case, however, Cox\\'s model is no longer a proportional hazards model. In this paper as in the vast majority of used applications, Cox\\'s model is used for variable x independent of time. The equation [Figure 8](#F8){ref-type=\"fig\"} can be written in the general form [Figure 9](#F9){ref-type=\"fig\"}:\n\n![](JMedLife-03-421-g009){#F9}\n\nin which a is n--dimensional column\\'s vector of model parameters, with components a~1~, a~2~,..., a~n~, x is n--dimensional column\\'s vector of independent variables with components x~1~,x~2~, ..., x~n~, and T designates the sign of transposition of a vector.\n\nSurvival function for Cox\\'s proportional hazards model [Figure 9](#F9){ref-type=\"fig\"} is [Figure 10](#F10){ref-type=\"fig\"}\n\n![Cumulative hazard function](JMedLife-03-421-g0010){#F10}\n\n![](JMedLife-03-421-g0011){#F11}\n\nFor baseline hazard function h~0~(t) respectively for H~0~(t) we can determine the values on experimental bases or we may use known statistical models: exponential, Weibull or Gompertz\\[[@R64]\\].\n\nTo determine the major factors that allow the forecasting survival (or mortality) we may use Cox\\'s multiple regression model, the proportional hazards model\\[[@R63]\\].\n\nThe next step is to fit the requested regression model to all data and to create a heart failure prognostic model.\n\nThe number of individuals living with heart failure is steadily increasing and there are many applications of such a model that may estimate the survival of a patient with heart failure.\n"}
+{"text": "All relevant data are within the paper.\n\nIntroduction {#sec001}\n============\n\nChronic beryllium disease (CBD) develops in up to 16% of individuals exposed to beryllium (Be) and is characterized by granulomatous inflammation and the accumulation of CD4+ T cells in the lung \\[[@pone.0117276.ref001]\\]. Using the Be lymphocyte proliferation test (BeLPT) we have identified workers with beryllium sensitization (BeS), demonstrating an immune response to Be with an abnormal BeLPT, but no evidence of CBD \\[[@pone.0117276.ref002], [@pone.0117276.ref003]\\]. Studies show that Be persists within the lungs of individuals many years after exposure has ceased \\[[@pone.0117276.ref004]\\], suggesting a failure to clear Be antigen from the lungs. This retention of Be may perpetuate an ongoing Be-specific immune response in CBD and/ or progression from BeS to CBD. It has been hypothesized that alveolar macrophages may undergo apoptosis upon exposure to Be and that this may contribute to retention of Be in the lungs of those with CBD \\[[@pone.0117276.ref004], [@pone.0117276.ref005]\\].\n\nIn human peripheral blood, monocyte subpopulations with distinct functional properties have been defined by their expression of CD14 and CD16. CD14 is the receptor for complexes of lipopolysaccharide (LPS) and LPS-binding protein \\[[@pone.0117276.ref006]\\]. CD16 is the low-affinity receptor for the Fc region of IgG (Fc\u03b3 receptor type III \\[Fc\u03b3RIII\\]) and plays an important role in the clearance of immune complexes \\[[@pone.0117276.ref007]\\]. Monocyte subsets were initially defined as CD14+CD16+ and CD14++CD16+ based on work in healthy control subjects. CD14+CD16+ monocytes, which compose a minor population of monocytes in human peripheral blood mononuclear cells (PBMC), are more mature than CD14++CD16+ classic monocytes \\[[@pone.0117276.ref008], [@pone.0117276.ref009]\\]. Previous studies have shown that the number of CD14+CD16+ monocytes are expanded during severe infectious and inflammatory conditions, such as Rheumatoid arthritis (RA) \\[[@pone.0117276.ref010]\\], tuberculosis \\[[@pone.0117276.ref011]\\], asthma \\[[@pone.0117276.ref012]\\] and sarcoidosis \\[[@pone.0117276.ref013]\\]. In addition, these monocytes have been implicated in the pathogenesis of several inflammatory diseases, such as RA as they produce higher levels of tumor necrosis factor-\u03b1 (TNF- \u03b1) and IL-1\u03b2 and preferentially differentiate into macrophages \\[[@pone.0117276.ref014]\\]. However, this finding has not been confirmed in other diseases, such as systemic lupus erythematosus (SLE) \\[[@pone.0117276.ref015]\\]. Moreover, several investigators demonstrated that CD14+CD16+ cells express surface markers and exhibit functional activity characteristic of dendritic cells \\[[@pone.0117276.ref016], [@pone.0117276.ref017]\\] and have suggested that CD14+CD16+ cells may differentiate into proinflammatory mononuclear cells. More recent evidence suggests that this subpopulation can be further subdivided into CD14^dim^CD16+ and CD14^bright^CD16+ cells. The CD14^bright^CD16+ monocyte subpopulation has been reported to contain the majority of interleukin-10 (IL-10) producing cells and to produce high levels of proinflammatory cytokines such as TNF- \u03b1 and IL-1 \u03b2 \\[[@pone.0117276.ref018], [@pone.0117276.ref019]\\]. In contrast, CD14^dim^CD16+ monocytes appear to have high migratory capacity but only limited phagocytic potential \\[[@pone.0117276.ref020]\\]. So far, this has not been investigated in patients with CBD.\n\nCurrent studies indicate that alveolar macrophages (AMs) arise from circulating blood monocytes, which colonize the tissues under inflammatory and non-inflammatory states. The AMs may improve function in maintaining homeostasis in the lung. On one hand, they fight against pathogens by activating multiple immunological pathways and serve as the first line of defense, while on the other hand, they manifest an anti-inflammatory response to protect excessive tissue damage. Cigarette smoke (CS) induces an accumulation of macrophages in the terminal airways of the lung and this is noted in alterations of BAL cell numbers and % as well as on histological lung samples from smokers \\[[@pone.0117276.ref021]\\]. Numerous epidemiologic studies and biological studies have shown that smoking maybe protective or reduce the development of granulomatous lung disease \\[[@pone.0117276.ref022]--[@pone.0117276.ref025]\\]. However, the molecular mechanism by which this occurs is not clear.\n\nSince AMs constitutes an important link between pulmonary innate and adaptive immunity due to their antigen-presenting capacity and ability to express different immunomodulating mediators \\[[@pone.0117276.ref026]\\], the migration of monocytes from blood to lung and their differentiation into macrophages may be an important step in disease pathogenesis. However, the role of AMs in CBD pathogenesis, and in the protective effects of smoking on CBD has yet to be fully determined. As altered functional capacity of AMs may be reflected in cell surface antigen expression, we hypothesized that CBD will have a different AMs phenotype compared to BeS Non-smoker (BeS-NS) patients and healthy subjects (HS). In addition, since Fc\u03b3R cross-linking on macrophages potentially initiates phagocytosis, antigen presentation, antibody-dependent cell-mediated cytotoxicity (ADCC) and release of pro-inflammatory cytokines and mediators of tissue destruction \\[[@pone.0117276.ref027]\\], we further hypothesized that some of these functions would differ between CBD, BeS-NS, BeS smokers (BeS-S) and HS, as well as in the lung and blood of these subjects and explain the protective effect of cigarette smoke on CBD. The aim of this study was to investigate the relevance of CD14+CD16+ monocyte phenotype as well as the AMs' cell surface receptor and function in patients with CBD, BeS-NS, BeS--S with normal lung function and HS.\n\nMaterials and Methods {#sec002}\n=====================\n\nStudy population {#sec003}\n----------------\n\nAll participants gave informed written consent in accordance with the Declaration of Helsinki, and the study was approved by National Jewish Health (NJH) Institutional Review Board (IRB) for Human Subjects. Twenty patients with a diagnosis of CBD, twenty with BeS-NS and ten BeS-S were enrolled in this study from patients seen clinically at NJH. The diagnosis of CBD was established using previously defined criteria, including the presence of granulomatous inflammation on lung biopsy, and a positive proliferative response of blood and/or BAL cells to BeSO~4~ in vitro. The diagnosis of BeS was established based on a positive proliferative response of blood cells to BeSO~4~ in vitro on the BeLPT, and the absence of granulomatous inflammation or other abnormalities on bronchoscopic lung biopsy. Individuals were considered to be current smokers if they had smoked daily for at least the last 3 months. PBMCs were isolated from 10 HS, 10 CBD, 10 BeS-NS subjects, while BAL cells were isolated from 14 CBD, 14 BeS-NS and 10 BeS-S at NJH. Paired PBMCs and BALs samples where obtained from four of the same CBD subjects, and four BeS-NS subjects and none of the HS subjects. Instead, the BAL cells from six HS were obtained as described previously \\[[@pone.0117276.ref028]\\]. Briefly, de-identified donor lungs that were not suitable for transplant were obtained through the National Disease Research Interchange (NDRI, Philadelphia, PA, USA) and the International Institute for the Advancement of Medicine (Edison, NJ, USA). Demographic data was obtained on all subjects as part of our research protocol and from the NDRI for the HS. No CBD smokers were available to participate in this research study, as we have very few subjects with CBD who currently smoke.\n\nPeripheral blood mononuclear cells (PBMCs) {#sec004}\n------------------------------------------\n\nPeripheral blood mononuclear cells (PBMCs) were isolated by Ficoll density gradient centrifugation. The viability of PBMCs was always \\>95%, as determined by trypan blue staining. The viable cells were quantified in a Neubauer chamber (Zeiss, Oberkochen, Germany) and leftover cells were cryopreserved in liquid nitrogen.\n\nBronchoscopy {#sec005}\n------------\n\nBronchoscopy was performed to obtain bronchoalveolar lavage (BAL) cells from patients, as previously described \\[[@pone.0117276.ref029]\\]. Briefly, four 60 ml aliquots of sterile normal saline (at room temperature) were instilled into the airway with a syringe then aspirated using low suction. For each collection from an individual patient the aspirated BAL specimens were pooled, kept on ice, and processed within one hour of collection.\n\nThe BAL cells from HS were isolated as described previously \\[[@pone.0117276.ref028]\\]. The middle lobe was perfused and lavaged with HEPES-buffered saline containing 2 mM EDTA and then with HEPES-buffered saline alone. The lavage was retained for isolation of AMs and the lavage fluid was centrifuged at 4\u00b0C for 10 min and washed in Dulbecco's modified Eagle's medium (DMEM).\n\nTo obtain AMs, the BAL cells were resuspended in DMEM supplemented with 10\u200a% fetal calf serum (FCS), 2 mM glutamine and antibiotics (100 g/ml streptomycin, 100 U/ml penicillin G) and plated on tissue-culture plates. After a 2 h incubation to allow adherence, the cells were washed with DMEM to remove the non-adherent cells. The purity of isolated AMs was close to 100% after adherence in culture, as measured by CD68 staining.\n\nFlow cytometry analysis {#sec006}\n-----------------------\n\nThe AMs and PBMCs were plated at a density of 1\u00d710^6^ /ml for culture with and without BeSO~4~ at 37\u00b0C for 24 hours. In fresh cells and those after 24h treatment with 10\u03bcM BeSO~4~, the expression of CD14, CD16, CD86, and HLA-DR on monocytes and CD11c, CD14, CD16, CD40, CD80, CD86, HLA-DR on AMs were measured using flow cytometry. Cells were washed in FACS medium (PBS containing 1% BSA) and stained at 4\u00b0C for 20 min by antibodies directly conjugated with FITC, PE, PerCP or APC. Thereafter, cells were washed three times with PBS and analysed by FACScalibum (Becton Dickinson, Heidelberg, Germany) using the CellQuest software (Becton Dickinson).\n\nPhagocytosis Assays {#sec007}\n-------------------\n\nPhagocytosis assays were performed as described previously \\[[@pone.0117276.ref030]\\]. In brief, one-millilitre aliquots of BAL cells at a concentration of 1\u00d7 10^6^ /ml were adhered to 12-well plates. After 24h treatment with 10\u03bcM or 100\u03bcM BeSO~4~, the fluid was removed and adherent AMs were placed in 1 ml fresh DMEM medium with 10%FCS. Flash Red carboxylate-modified beads (5mm) were added to the AMs monolayers (at a ratio of 5:1 beads/plated AMs for a total of 5\u00d7 10^6^ beads) in 12-well plates for 1 hour at 37\u00b0C/5% CO~2~. Phagocytosis of beads was determined by flow cytometry.\n\nStatistical Analysis {#sec008}\n--------------------\n\nA Welch ANOVA was used to determine the effect of treatment. After the data were checked for significant treatment differences, individual contrasts were calculated to compare treatment means of interest. Normalizing transformations were made in cases where the data was non-Gaussian. When data transformations were unsuccessful, we used suitable nonparametric tests. All other comparisons were made using Wilcoxon rank sum test and the paired student's t-Test. A p \\< 0.05 was used to determine statistical significance and all tests were two sided.\n\nResults {#sec009}\n=======\n\nDemographics of participants {#sec010}\n----------------------------\n\nThe demographics of the HS, CBD, BeS-NS and BeS-S patients are shown in [Table 1](#pone.0117276.t001){ref-type=\"table\"}. No difference was seen in the age of the HS, CBD, BeS-NS and BeS-S enrolled in this study and the majority were male. CBD subjects had a statistically significant increase in the percentage of BAL lymphocytes compared with BeS-NS patients (median 16, range 4--38.5, versus median 6.7, range 1.5--16.5; P \\< 0.05) and BeS-S (median 3.95, range 1--25.3, versus median 6.7, range 1.5--16.5; P \\< 0.05). The total BAL cell lymphocytes counts were not available for the HS. BeS-S subjects had a statistically significant increase in the total BAL WBC count (median 100, range 21--365) compared with BeS-NS patients (median 58.3, range 27.5--66.7).\n\n10.1371/journal.pone.0117276.t001\n\n###### Clinical characteristics of BeS and CBD subjects.\n\n![](pone.0117276.t001){#pone.0117276.t001g}\n\n                                         HS blood (n = 10)   BeS-NS blood (n = 10)   CBD blood (n = 10)   NDRI HS BAL (n = 6)   CBD BAL (n = 14)   BeS-NS BAL (n = 14)   BeS-S BAL (n = 10)\n  -------------------------------------- ------------------- ----------------------- -------------------- --------------------- ------------------ --------------------- --------------------\n  Age (yr)                               67(55--73)          64(50--73)              64(52--75)           52 (43--57)           60 (49--73)        63 (38--72)           50(40--65)\n  Gender (M/F)                           8/2                 9/1                     9/1                  4/2                   12/2               10/4                  8/2\n  Race (W/AA/H)                          10/0/0              9/1/0                   9/1/0                6/0/0                 11/0/3             13/1/0                9/0/1\n  Smoking status (CS/FS/NS)              0/0/10              0/0/10                  0/0/10               0/0/6                 0/0/14             0/0/14                10/0/0\n  BAL cells Lymphocytes (%)              NA                  NA                      NA                   NA                    16 (4--38.5)\\*     6.7 (1.5--16.5)       3.95 (1--25.3)\n  BAL cells Total WBC count (millions)   NA                  NA                      NA                   NA                    44(33--50.1)       58.3 (27.5--66.7)     100(21--365)\\*\n\nData are expressed as median (range). W, white; AA, African American; H, Hispanic; CS, current smoker; FS, former smoker, NS, never smoker. \\*p\\<0.05\n\nCD14^dim^CD16+ cells were decreased in fresh PBMCs of BeS and CBD subjects {#sec011}\n--------------------------------------------------------------------------\n\nFrequencies of CD14^bright^CD16+ and CD14^dim^CD16+ cells from PBMCs from 10 CBD, 10 BeS-NS and 10 age and gender matched HS were determined by flow cytometry. The representative staining patterns of surface expression of CD14 and CD16 from a CBD patient and an HS is shown in [Fig. 1A](#pone.0117276.g001){ref-type=\"fig\"}. The mean\u00b1 SD frequency of CD14^dim^CD16+ cells was significantly decreased in CBD patients (5.8\u00b11.5%, p\\<0.001) and BeS-NS patients (6.5\u00b11.4%, p\\<0.01) compared to the HS (9.1\u00b10.8%) ([Fig. 1B](#pone.0117276.g001){ref-type=\"fig\"}). There was no significant difference in the frequency of CD14^bright^CD16+ cells in CBD (9.0\u00b12.4%) and BeS-NS patients (8.6\u00b12.2%) compared with HS (7.2\u00b1 2.2%) ([Fig. 1C](#pone.0117276.g001){ref-type=\"fig\"}). In comparison with the classic monocyte CD14^bright^CD16- subpopulations, CD14^bright^CD16+ and CD14^dim^CD16+ cells showed higher HLA-DR mean fluorescent intensity (MFI, [Fig. 1D](#pone.0117276.g001){ref-type=\"fig\"}), with the most prominent expression noted in the BeS-NS participants (p\\<0.05).\n\n![CD14^dim^CD16+ cells were at decreased frequency in fresh PBMCs of both BeS-NS and CBD.\\\n1a. Representative dot plots of CD14 and CD16 expression on monocytes in CBD (Left) and HS (Right). 1b. Percentages of CD14^bright^CD16+monocytes among total monocytes from HS and patients with BeS-NS and CBD. Each dot represents an individual subjects. 1c. Percentages of CD14^dim^CD16+ monocytes among total monocytes from HS and patients with BeS-NS and CBD. 1d. MFI of HLA--DR expression on CD14^bright^CD16-, CD14^bright^CD16+ and CD14^dim^CD16+ monocytes from patients with HS, BeS-NS and CBD. The median percentage is shown as a solid line. MFI = mean fluorescence intensity. \\*p\\<0.05.](pone.0117276.g001){#pone.0117276.g001}\n\nCD14^dim^CD16+ cells were at increased frequency in AMs of CBD, while CD14^bright^CD16+ cells were at increased frequency in AMs of BeS, compared to HS {#sec012}\n-------------------------------------------------------------------------------------------------------------------------------------------------------\n\nA representative example of the AMs from CBD demonstrates a CD14^dim^CD16+ phenotype. Specifically, 86% of CBD AMs had a CD14^dim^CD16+ phenotype while only 6% demonstrated a CD14^bright^CD16+ phenotype ([Fig. 2A-C](#pone.0117276.g002){ref-type=\"fig\"}). The mean\u00b1 SD frequency of CD14^dim^CD16+ cells was significantly increased in CBD patients (57.9 \u00b16.6%, p\\<0.01) compared to the HS (22.1\u00b18.9%) ([Fig. 2D](#pone.0117276.g002){ref-type=\"fig\"}). The mean\u00b1 SD frequency of CD14^bright^CD16+ cells was significantly increased in BeS patients (30.9\u00b17.0%, p\\<0.01) compared to the HS (6.2\u00b12.5%) ([Fig. 2E](#pone.0117276.g002){ref-type=\"fig\"}).\n\n![CD14^dim^CD16+ cells were at increased frequency in AMs of CBD, while CD14^bright^CD16+ cells at increased frequency in AMs of BeS.\\\n2a. A representative example of AMs gates from fresh BAL cells from CBD patients. 2b. The expression of CD14 and CD16 using the AMs gate. In the representative examples shown, 85.9% of AM were with a CD14^dim^CD16+ phenotype while only 6.06% of AM with a CD14^bright^CD16+ phenotype. 2c. The expression of HLA-DR using the AMs gate. 95.2% of AMs highly expressed HLA-DR. 2d. Frequency of CD14^dim^CD16+ cells among total AMs from HS and patients with BeS-NS and CBD. Each dot represents an individual subjects. 2e. Frequency of CD14^bright^CD16+ cells among total AMs from HS and patients with BeS-NS and CBD. Each dot represents an individual subjects. \\*\\*p\\<0.01.](pone.0117276.g002){#pone.0117276.g002}\n\nIncreased expression of CD16, CD40, CD86 and HLA-DR is present on fresh BeS-NS and CBD AMs compared to HS and BeS smokers {#sec013}\n-------------------------------------------------------------------------------------------------------------------------\n\nCBD and BeS-NS AMs demonstrated different cell surface markers in comparison with HS and BeS-S ([Fig. 3](#pone.0117276.g003){ref-type=\"fig\"}). Specifically, the expression of CD16 (p\\<0.01), CD40 (p\\<0.05), CD86 (p\\<0.05) and HLA-DR (p\\<0.05) were significantly increased in CBD and BeS-NS in comparison with HS and BeS-S. No change was detected in the expression of CD11c and CD14 between participants. The AMs expression of CD80 was weak and did not differ between groups (data not shown). There was no significant difference in any of the markers between HS and BeS-S.\n\n![Percent of fresh AMs cells expression CD11c, CD14, CD16, CD40, CD86 and HLA-DR in HS, CBD, BeS-NS and BeS-S.\\\nEach dot represents an individual subjects. \\*p\\<0.05, \\*\\*p\\<0.01.](pone.0117276.g003){#pone.0117276.g003}\n\nBe exposure increases CD86 on circulating monocytes and decreases CD16 on AMs {#sec014}\n-----------------------------------------------------------------------------\n\nThe phenotypic pattern of circulating monocytes from PBMC and AMs were altered after 10\u03bcM BeSO~4~ treatment for 24h ([Fig. 4](#pone.0117276.g004){ref-type=\"fig\"}). The expression of CD86 on monocytes was increased significantly in CBD, BeS-NS and HS participants (p\\<0.05, [Fig. 4A](#pone.0117276.g004){ref-type=\"fig\"}) after stimulation with 10\u03bcM BeSO~4~ for 24h, while no change in the expression of CD14, CD16 and HLA-DR was noted before and after treatment with BeSO~4~ (data not shown). Based on the MFI, increased HLA-DR was observed on circulating monocytes after treatment in all three groups ([Fig. 4A](#pone.0117276.g004){ref-type=\"fig\"}). The expression of CD16 on AMs from both CBD and BeS-NS subjects was downregulated significantly after stimulation with 10\u03bcM BeSO~4~ for 24h, while the expression of CD86 and HLA-DR was unaltered with Be treatment in all subjects ([Fig. 4B](#pone.0117276.g004){ref-type=\"fig\"}).\n\n![Expression of phenotypic makers on monocytes and AMs with or without 10\u03bcM BeSO~4~ stimulation.\\\n4a. Expression of phenotypic makers on monocytes. Upper Left. monocytes CD86(%); Upper Right. monocytes HLA DR MFI. 4b. Expression of phenotypic makers on AMs. Lower Left. AMs CD16 (%); Lower middle. AMs CD86(%); Lower Right. AMs HLA DR(%). Bars in a and b show the mean +/- SEM. \\*p\\<0.05, \\*\\*p\\<0.01, \\*\\*\\*p\\<0.001.](pone.0117276.g004){#pone.0117276.g004}\n\nPhagocytosis by CBD and BeS-NS AMs is significantly downregulated after stimulation with BeSO~4~ {#sec015}\n------------------------------------------------------------------------------------------------\n\nInvestigation of phagocytic capacity of adherent AMs from CBD and BeS-NS revealed higher phagocytosis compared with HS and BeS-S (p\\<0.05) ([Fig. 5](#pone.0117276.g005){ref-type=\"fig\"}). However, after the 10\u03bcM and 100\u03bcM BeSO~4~ treatment for 24h, only CBD and BeS-NS AMs demonstrated a significant reduction in the percentage of AMs displaying phagocytosis, with the AMs from BeS-NS demonstrating a 26.4% reduction after treatment with 10\u03bcM BeSO~4~ (p\\<0.05) and a 48.3% reduction with 100\u03bcM BeSO~4~ (p\\<0.0001), while with the AMs from CBD demonstrating a 25.1% reduction after treatment with 10\u03bcM BeSO~4~ (p\\<0.05) and a 31.4% reduction with 100\u03bcM BeSO~4~ (p\\<0.05). No difference was found in the phagocytosis by the HS and BeS-S AMs with Be treatment.\n\n![The phagocytosis of AMs in CBD and BeS-NS were downregulated significantly after stimulation with BeSO~4~ for 24h.\\\nBars show the mean +/- SEM. \\*p\\<0.05, \\*\\*\\*p\\<0.001.](pone.0117276.g005){#pone.0117276.g005}\n\nDiscussion {#sec016}\n==========\n\nIn the present work we have demonstrated that the circulating CD14^dim^CD16+ monocytes, which represent a pro-inflammatory subpopulation, are present at a lower frequency in PBMCs than the lung of CBD and BeS-NS patients. We have also shown that AMs of CBD and BeS-NS display an activated cell surface phenotype with higher frequencies of the markers CD40, CD86 and HLA-DR in contrast to AMs of HS and BeS-S. Furthermore, the AMs from CBD and BeS-NS have an increased phagocytic ability compared to HS and BeS-S AMs. Interestingly, Be treatment down regulates the CD16 expression on AMs as well as the phagocytic capacity of CBD and BeS-NS AMs. Collectively, these findings may indicate that CD14^dim^CD16+ monocytes are lower in peripheral blood of CBD and BeS-NS patients as they compartmentalize and differentiate into mature and activated CD16+ AMs in the lung. Furthermore, these results may also suggest that Be exposure alters the balance between an innate and acquired immune response, with the promotion of activated AMs with reduced phagocytic capacity, which in turn result in a Th1 inflammatory response and deficient antigen clearance in the lung.\n\nCD14+CD16+ monocytes are a subpopulation of cells that have acquired features consistent with mature tissue macrophages while in the circulation \\[[@pone.0117276.ref016]--[@pone.0117276.ref017]\\]. The CD14^bright^CD16+ monocyte subpopulation has been reported to contain the majority of IL-10-producing cells \\[[@pone.0117276.ref018]\\] and to produce high levels of proinflammatory cytokines, such as TNF-\u03b1 and IL-1\u03b2. In addition, this monocyte subpopulation has been noted to be selectively expanded in a number of inflammatory processes, including coronary artery disease \\[[@pone.0117276.ref019]\\], severe asthma \\[[@pone.0117276.ref031]\\], and sarcoidosis \\[[@pone.0117276.ref013]\\]. In contrast, CD14^dim^CD16+ monocytes were recently shown to have high migratory but only limited phagocytic potential \\[[@pone.0117276.ref020]\\], and did not produce ROS or cytokines in response to cell surface toll-like receptors. Instead, they selectively produced TNF-\u03b1, IL-1\u03b2, and CCL3 in response to viruses and immune complexes containing nucleic acids, via a proinflammatory TLR7-TLR 8-MyD88-MEK pathway \\[[@pone.0117276.ref020]\\]. In the present study, we demonstrated that CD14^dim^CD16+ cells were at decreased frequency in PBMCs of both BeS-NS and CBD compared to HS, while CD14^bright^CD16+ monocytes did not differ between groups. In addition, we observed that the expression of HLA-DR was higher in both CD14^bright^CD16+ and CD14^dim^CD16+ monocyte subsets, suggesting that these cells are activated. We therefore hypothesized that blood CD14^dim^CD16+ monocytes compartmentalized to the lung and differentiated into AMs in CBD and BeS-NS, accounting for the decrease in the CD14^dim^CD16+ monocyte subsets we observed in blood cells. In support of this hypothesis, we investigated the phenotype of AMs derived from BAL. Indeed, our data demonstrate that CD14^dim^CD16+ cells were increased in CBD AMs, while CD14^bright^CD16+ cells were increased in AMs of BeS, compared to HS. Not surprising, the CBD AMs displayed an activated cell surface phenotype, with increased expression of CD16, CD40, CD86 and HLA-DR compared to HS AMs. As a robust Th1 cytokine response is noted in CBD \\[[@pone.0117276.ref032], [@pone.0117276.ref033]\\], our data supports that these cells contribute to the proinflammatory cytokine production in CBD lung, similar to that noted for sarcoidosis \\[[@pone.0117276.ref034]\\]. Surprisingly, the fresh AMs cells from BeS-NS demonstrated a similar patter to those of CBD, with increased expression of activation markers despite the fact that BeS-NS subjects are known to have normal BAL WBC counts and lymphocyte %, reduced or absent lung Th1 cytokine production as well as normal pathology (similar to data not shown from our subjects in this study). Since there are no difference in the AM phenotype between BeS-NS and CBD, there must be other differences in BeS-NS and CBD AMs, such as the presence of pivotal immune response genes and networks, which impact the Th1 immune response to Be and account for progression from BeS-NS to CBD.\n\nThe AMs are not thought to be the most important cell type involved in antigen presentation in the lower respiratory tract. Lung dendritic cells are thought to be the crucial APC, picking up the antigen and carrying it to local lymph nodes, where a specific immune response can be initiated \\[[@pone.0117276.ref035]\\]. AMs may act as silencers of an immune response in the lung \\[[@pone.0117276.ref036]\\]. However, in some diseases such as asthma, hypersensitivity pneumonitis (HP), and sarcoidosis \\[[@pone.0117276.ref037]--[@pone.0117276.ref039]\\], AMs become involved in the maintenance and further expansion of the immune response in the target organ, the lung. Previous studies from our group have shown that CBD BAL cells induce T cell proliferation and produce Th1-type cytokines, including TNF-\u03b1 and IFN-\u03b3 in response to Be-stimulation \\[[@pone.0117276.ref043]\\]. We speculated that BeS-NS AMs would have been active in silencing the immune response to Be and perpetuated maintenance of sensitization versus progression to CBD. Our finding of overexpression of CD16, CD40, CD86 and HLA-DR on AMs in CBD suggests increased AMs activation which in turn an increased Th1 cytokine production and immune response in CBD, likely with increased antigen presentation to T cells. Since we did not find differences in the AM activation between BeS-NS and CBD, this might suggest that AMs in BeS-NS and CBD are not able to silence the Be-immune response and instead perpetuate inflammation in the lung once started. Alternatively, other mechanisms besides the AM phenotype and activation may downregulate the immune response to Be in the BeS-NS lung.\n\nWe found that the phenotypic pattern of circulating monocytes and AMs were altered after Be treatment. Specifically, following Be challenge, the %CD86 on circulating monocytes increased in all the groups and %CD16 on AMs in both BeS-NS and CBD decreased ([Fig. 4A and 4B](#pone.0117276.g004){ref-type=\"fig\"}). The HLA-DR MFI on circulating monocytes also increased with Be exposure ([Fig. 4A](#pone.0117276.g004){ref-type=\"fig\"}). Other costimulatory markers were not altered with Be-stimulation. These results indicate that Be has the potential to trigger innate AM immune cell activation whether from CBD, BeS-NS or non-CBD individuals. It is likely that this alteration contributes to the induction of non-specific lung inflammation and injury, as well as to the development of BeS-NS in genetically susceptible individuals. Supporting this notion, prior studies have demonstrated that Be exposure in non-diseased human and animal cell models stimulates a variety of cellular responses, including cell migration \\[[@pone.0117276.ref040]\\], and cytokine regulation \\[[@pone.0117276.ref041]\\]. However, not all cells respond to Be in the same manner, as our data indicate that only CBD and BeS-NS AMs and circulating monocytes demonstrate a reduction in CD16, and thus a different phenotypic pattern with exposure to Be.\n\nTo further better define the impact of the loss in CD16 expression with Be challenge, we evaluated alterations in the phagocytic capability of AMs after Be treatment, as cross-linking of CD16 (a member of the of Fc\u03b3R family) on macrophages initiates phagocytosis. We found that the phagocytic capacity of both BeS-NS and CBD AMs was downregulated, while no difference was found in the phagocytic capacity of HS and BeS-S with Be stimulation. While the reduced phagocytic ability of BeS-NS and CBD AMs is likely related to CD16 reduction, our previous work has shown that Be-stimulates macrophage apoptosis. Specifically, Be stimulates apoptosis in mouse and human macrophage cell lines and in CBD and BeS-NS BAL macrophages \\[[@pone.0117276.ref004], [@pone.0117276.ref042]\\] which is dependant on the activation of intracellular caspases and the up-regulation of ROS, but independent of Be-stimulated TNF-\u03b1 production \\[[@pone.0117276.ref005]\\]. CD16-positive monocyte-derived macrophages showed a higher phagocytosis for opsonized Escherichia coli \\[[@pone.0117276.ref043]\\]. We speculate that this results in CD16 reduction on macrophages as they undergo apoptosis. In addition, the loss of surface expression of CD16 will contribute to reduced phagocytosis. Meanwhile, the phagocytic ability of AMs in HS did not change, which correlates with the lack of alteration in CD16 noted in these cells. Our data may provide at least a partial explanation for the protective effect of smoking on non-infectious granulomatous inflammatory diseases, such as sarcoidosis, HP and Crohn's disease \\[[@pone.0117276.ref022]--[@pone.0117276.ref025]\\], as smoking reversed the expression of AMs activated surface markers to normal ([Fig. 3](#pone.0117276.g003){ref-type=\"fig\"}) and reduced the expression of costimulatory molecules and phagocytosis. These changes may in turn inhibit the AMs ability to mount a Be-specific immune and inflammatory response in the lung as well as impair antigen presentation and T cell activation in the lung. That is maybe, at least in part, responsible for the protection that smokers have against developing the granulomatous disease CBD.\n\nIn summary, our data suggest that there is an inappropriate Be-stimulated immune response in CBD and BeS-NS with alterations in an AM phenotype that is activated compartmentalized to the lung. These AMs demonstrate reduced phagocytic ability likely resulting in the inability to clear a Be antigen in BeS-NS and CBD. We speculate that the Be-stimulation of a more mature CD16+ AMs in turn results in increased Th1 cytokine and chemokine production, and perpetuates the Be immune response in CBD. While it is clear that Be stimulates an innate immune response with the activation of AMs markers in HS, ultimately there must be an alteration in the balance between an innate and acquired immune response to develop BeS-NS and or CBD; this imbalance may result from deficient antigen clearance with reduced phagocytosis and promotion of Th1 cytokine inflammatory responses in the lung. The latter alterations are likely dependent on AMs in contrast to circulating monocytes. However, progression from CBD and BeS-NS likely result from differences in these cells (and or others) that are beyond what we demonstrated in our study, since we found no significant difference in AMs activation and phenotype between BeS-NS and CBD. Understanding the scope of the dysregulated immune response from Be-stimulated AMs in BeS-NS and CBD and defining ways of reversing this immune response is a focus of our future studies to prevent and/or treat CBD.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: LL NH MG JE JW EBG PM LM. Performed the experiments: LL MG JE JW. Analyzed the data: LL NH LM. Contributed reagents/materials/analysis tools: LL NH LM. Wrote the paper: LL NH LM.\n"}
+{"text": "Introduction {#section1-2158244019827717}\n============\n\nThe human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) pandemic (pandemic) was a major crisis at the end of the 20th and beginning of the 21st century. Such a defining moment in the history of health-related infections led to transformations in its proponents, as well as their medical practice. This research article aims to offer insights into the lived trajectories/experiences (journeys) of six Canadian gay physicians while they attempted to treat, care for, and cure/heal (treat) gay HIV/AIDS patients from 1981 to 2009.\n\nAccording to [@bibr55-2158244019827717], in June 1981, a major event, referred to as a tempest, crisis, lethal threat, or new outbreak (among other terms), shocked the world and baffled the medical establishment ([@bibr7-2158244019827717]; [@bibr22-2158244019827717]; [@bibr31-2158244019827717]; [@bibr36-2158244019827717]; [@bibr47-2158244019827717]; [@bibr54-2158244019827717]). An unknown virus, disease, or infection was reported for the first time by the Centers for Disease Control and Prevention (CDC) in the *Morbidity and Mortality Weekly Report* ([@bibr54-2158244019827717]). It was first called gay cancer or gay plague by the media, and gay-related immune deficiency (GRID) by the CDC ([@bibr4-2158244019827717]; [@bibr14-2158244019827717]; [@bibr15-2158244019827717]; [@bibr34-2158244019827717]). In 1982, the CDC renamed the infection AIDS.^[1](#fn1-2158244019827717){ref-type=\"fn\"}^ It was not until 1986 that HIV was termed ([@bibr3-2158244019827717]). The infection soon grew into an epidemic, as it was prevalent within a larger community, mostly gay ([@bibr7-2158244019827717]; [@bibr53-2158244019827717]). As cases appeared in other parts of the world, besides North America, such as Europe and Africa, it was eventually recognized as a pandemic ([@bibr5-2158244019827717]; [@bibr52-2158244019827717]).\n\nThroughout the more than four decades since the appearance of the HIV infection, many narratives have been published (e.g., [@bibr2-2158244019827717]; [@bibr11-2158244019827717]; [@bibr38-2158244019827717]; [@bibr49-2158244019827717]; [@bibr52-2158244019827717]). They speak of \"courage, cowardice, hope, despair, compassion, and bigotry\" ([@bibr53-2158244019827717], p. 26). They describe the fear of the unknown, the spread of the disease, death, sexuality, and alternative lifestyles, such as bisexuality, homosexuality, transgender, and drug use ([@bibr60-2158244019827717]; [@bibr61-2158244019827717]). Because of the attitudes, beliefs, interpretations, opinions, perceptions, and values that individuals held about HIV/AIDS, many social problems arose, such as stigmatization, discrimination, and prejudice ([@bibr19-2158244019827717]; [@bibr60-2158244019827717]; [@bibr61-2158244019827717]). Moreover, because of the convergence of the myriad of uncertainties regarding HIV prevention, HIV/AIDS treatment, and opportunistic infections, many HIV-infected patients developed comorbidities with the above-mentioned issues or fears. Furthermore, some health professionals \"succumb\\[ed\\] to hysteria and abandon\\[ed\\] their objectivity and compassion in the care of patients with HIV/AIDS\" ([@bibr20-2158244019827717], p. 114; [@bibr60-2158244019827717]; [@bibr61-2158244019827717]).\n\nIn response to the shock provoked by the sudden and still unknown infection, as well as the hysteria, abandonment, and victimization of some HIV-infected gay members by their community, \"the gay, lesbian, bisexual, and transgender (LGBT) communities quickly came to realize that if they didn't take action, nobody would\" ([@bibr54-2158244019827717], p. 3). Many gay health professionals, and gay physicians in particular, embraced the same view. They responded to the call to action because they had friends or family members who were HIV-infected, or as a personal response to the abandonment and rejection of their patients who were either gay, or HIV-infected, or both simultaneously ([@bibr46-2158244019827717]). By heeding the call to action, these physicians modified their public image and, at the same time, transformed their personal identity. They came out of their professional sexual neutrality and became \"openly\" gay physicians, profoundly implicated in the political, social, and medical causes of their patients, friends, and community ([@bibr29-2158244019827717]).\n\nUnfortunately, over time, memories of what it was like to meet head-on a grim, contagious, disfiguring, lethal, and sexually transmitted threat like HIV/AIDS have begun to fade. It was a \"time when medicine was all but powerless\" ([@bibr7-2158244019827717], p. 3) and when \"people with HIV \\[/AIDS\\] were fired from their jobs, kicked out of their apartments, denied health care and abandoned by their families\" ([@bibr1-2158244019827717], p. 4). However, those who survived this period appear to be ready to speak about it again. There is a new generation of activists, archivists, artists, historians, and researchers, among others, who were born during the HIV/AIDS pandemic, who now need to make sense of this grim period ([@bibr22-2158244019827717]). Moreover, there is a growing number of archival projects, books, documentaries, and gallery exhibitions that are now devoted to the topic, with more underway.\n\nOne such undertaking is this scientific research article which presents an unpublished study conducted between 2008 and 2009 with six Canadian gay HIV/AIDS physicians (study). Their stories are compelling journeys that transformed their personal and professional identities, and their medical practice. While it aims to offer insights into the lived journeys of unsung heroes (physicians), it also serves as a reminder of their courage as lobbyists and advocates (militants) for their HIV/AIDS patients (patients) during the pandemic.\n\nLiterature Review {#section2-2158244019827717}\n-----------------\n\nAt the time of the study, only two types of literature pertaining to the study's topic, that is, The Use of Experiential Learning in the Transformation of the Medical Practice of Six Professionals in the Context of the HIV/AIDS Pandemic,^[2](#fn2-2158244019827717){ref-type=\"fn\"}^ were available. They included books written by several nonmedical authors setting the study's context, as well as autobiographies written by medical and nonmedical authors providing a firsthand account of physicians' lived journeys. These two types of literature served to establish a preliminary understanding of the patients' trajectories from the onset of the infection to the patients' death. They also provided an account of physicians' journeys from the onset of the infection to their patients' death and beyond, including attending their funerals, for example, at different key periods of the pandemic's trajectory. These key periods were identified as follows: *Period 1*, HIV/AIDS Emergence; *Period 2*, Recognized Pandemic and Militancy Rise; and *Period 3*, HIV Chronicity and Militancy Decrease (see [appendix](#table1-2158244019827717){ref-type=\"table\"}---HIV-AIDS Pandemic Trajectory: Three Periods). A few examples of the books and autobiographies from nonmedical and medical authors are highlighted here for their significant input into the study.\n\nIn their book titled *Good Doctor Good Patients: Partners in the HIV Treatment*, [@bibr48-2158244019827717] present patients' and their physicians' beliefs about the characteristics of and expectations from the physician--patient therapeutic relationship. They discuss medical and psychological challenges faced by both physicians and patients from the initial HIV/AIDS diagnosis up to and including the end-stage of the disease, and the dying process. They also describe specific skills and strategies that physicians can use to meet these challenges and discuss the role that friends, family, and professional counselors play in addressing the ongoing challenges of living with HIV/AIDS. In *Reports From the Holocaust: The Story of an AIDS Activist*, [@bibr35-2158244019827717] shares his views as an AIDS activist and advocate, as well as on AIDS and LGBT civil rights from the beginning of the pandemic. In addition, he documents his time spent working in community organizations against HIV/AIDS (e.g., Gay Men's Health Crisis and ACT UP) and beyond. As for [@bibr51-2158244019827717], in his book titled *And the Band Played On: Politics, People and the AIDS Epidemic*, he discusses the discovery and spread of HIV/AIDS and comments on the government's indifference and political infighting during the early 1980s.\n\nIn his autobiography titled *To the Friend Who Did Not Save My Life*, [@bibr25-2158244019827717] shares his own journey on his AIDS-related physical and psychological suffering, and reveals his physician--patient therapeutic relationship. [@bibr28-2158244019827717], in her autobiography titled *Life Lessons: Stories of Hope, Love and Laughter in the Face of AIDS*, discloses stories of love, betrayal, forgiveness, and joy from the counseling experiences which impacted her life. She also reveals her efforts to find meaning in every positive and negative life experiences, including when individuals are confronted with AIDS. As for [@bibr59-2158244019827717], in *Someone Was Here: Profiles in the AIDS Epidemic*, he tries to humanize the notions of illness, epidemic, and courage by reaching out to men and women whose lives were changed by the HIV/AIDS epidemic. In fact, these autobiographies, and many more, present the human toll of the HIV/AIDS pandemic (see also [@bibr10-2158244019827717]; [@bibr12-2158244019827717]; [@bibr16-2158244019827717]; [@bibr24-2158244019827717]; [@bibr33-2158244019827717]; [@bibr42-2158244019827717]; [@bibr45-2158244019827717]).\n\nIn his autobiography titled *The Least of These My Brethren: A Doctor's Story of Hope and Miracles on an Inner-City AIDS Ward*, [@bibr6-2158244019827717] shares the life stories of suffering and redemption of HIV/AIDS patients. These stories capture the hope and miracles he encountered while attempting to treat HIV/AIDS patients during his three and a half years in a New York AIDS center. In *My Own Country: A Doctor's Story of a Town and Its People in the Age of AIDS*, [@bibr57-2158244019827717] tells a story of the AIDS epidemic inside a hospital in rural America while confronting his own prejudices and fears. In *Strong Shadows: Scenes From an Inner City AIDS Clinic*, [@bibr64-2158244019827717] presents stories of HIV/AIDS patients she attempted to treat by illustrating their struggles with infections and medications, as well as the effects on them and their families. As for the book titled *AIDS Doctors: Voices From the Epidemic. An Oral History*, authors [@bibr7-2158244019827717] recount the stories of 73 physicians whose lives were centered on the HIV/AIDS pandemic (see also [@bibr11-2158244019827717]; [@bibr17-2158244019827717]; [@bibr38-2158244019827717]; [@bibr44-2158244019827717]; [@bibr49-2158244019827717]; [@bibr50-2158244019827717]).\n\nThemes {#section3-2158244019827717}\n------\n\nThe books and autobiographies mentioned in the literature review helped to identify the themes needed for the preparation of the semi-structured interview grid and the creation of the initial conceptual taxonomy that was used to organize and analyze data. The following themes emerged initially, as they were frequently mentioned and emotionally charged. Physicians highlighted the impact of the emergence of HIV/AIDS, their recognition of the epidemic and the pandemic on them, their patients, and their profession. They mentioned, among other themes, the uncertainties generated by a multiplication of scientific and medical discoveries/findings, as well as the developments in the treatment of opportunistic infections related to the HIV infection. Physicians also talked about the perceptions they and other physicians held about sexuality, drug use, and HIV/AIDS. Finally, they disclosed how they modified the way they practiced medicine at the outset of their careers, moving from a physician-centered practice toward a more patient-centered practice.\n\nStudy Methodology {#section4-2158244019827717}\n=================\n\nThe data presented in this research article were gathered through the descriptive, interpretive, and exploratory qualitative case study conducted by [@bibr29-2158244019827717]. The case study methodology approach was chosen by Jacob because it helped researchers learn about the global meaning of the physicians' experiences in the context of the pandemic through detailed descriptions pertaining to their lived journeys. It was also preferred because the study was situated at the heart of physicians' everyday life while interacting with and attempting to treat their patients. Finally, it was privileged because the HIV/AIDS pandemic is a contemporary event with facts and meaning related to the physicians' journeys inserted into their professional--patient relationships.\n\nThe study conducted by [@bibr29-2158244019827717] identified six Canadian gay HIV/AIDS dedicated physicians' experiential learning and captured their transformations during the moments they recalled while attempting to treat their patients. It also explained the processes they used and the motivations that compelled them to learn from their experiences and transform their personal and professional identities, as well as their practice. Jacob outlines how they rethought their relationship with their patients, other professionals, and community networks during those same gripping moments.\n\nThe approach adopted when recruiting physicians for that study was purposive because it selected as participants Canadian gay dedicated HIV/AIDS physicians who met specific criteria ([@bibr9-2158244019827717]; [@bibr37-2158244019827717]). These criteria were inspired by the criteria used by patients when choosing to consult a physician during the pandemic ([@bibr29-2158244019827717]). They mainly sought out physicians who could and were willing to provide information by virtue of their expertise and competence in sexually and blood-transmitted infections and in the HIV/AIDS prevention and treatment field. They looked for those who were involved and had experience, since the beginning of the pandemic, in the treatment of patients, and who had the reputation of having been confronted by the devastating effects of the pandemic throughout its trajectory by the HIV/AIDS research community. Moreover, they sought out physicians with the same sexual preference as theirs and identified as being gay, as well as individuals who possessed a militancy background, that is to say, mobilized and advocated for their patients and their community. Finally, patients migrated toward locations where these specific physicians had established their practice in large Canadian urban centers, mainly in private and public health clinics specialized in the treatment of patients with sexually and blood-transmitted infections, and in the HIV/AIDS prevention and treatment field.\n\nAn invitation to participate in semi-structured interviews was sent via email to a pool of physicians who met the above selection criteria. Their credentials and reputation were well-known by the HIV/AIDS research community through prior research studies conducted with them and/or through the physicians' presentations at national and international conferences (e.g., Canadian Conference on HIV/AIDS Research, and the International AIDS Conference). Six physicians responded to the invitation, thus forming the six cases. Their median age was 48 years old and their experience in the treatment of HIV/AIDS patients spanned an average of 19 years, with their medical practice consisting of 200 HIV/AIDS patients, in mean terms. Finally, the six physicians practiced medicine in large Canadian urban centers: four in private health clinics and two in public health clinics specialized in the prevention, treatment, and care of patients with sexually and blood-transmitted infections (e.g., HIV).\n\nThe following sections explain the interview process, as well as the data collection, production, and analysis as it is suggested by many specialists in the case study methodology ([@bibr26-2158244019827717], [@bibr27-2158244019827717]; [@bibr32-2158244019827717]; [@bibr41-2158244019827717]). After the standard procedures for qualitative studies, that is, explanation of the study, interview procedure, and consent form, the second phase of study began, that is, data collection through interviews. Interviews were conducted either at the physician's workplace (e.g., office) or in a room on the university campus. They were to last roughly 1 hr. However, the physicians' generosity would not be dampened by a time constraint! The interviews were conducted using an interview grid, that is, closed-ended questions for quantitative information gathering (e.g., age, number of years of experience, number of HIV/AIDS patients, and work environment) and open-ended questions for qualitative information gathering about the physicians' lived journeys. Besides being audio-taped, a researcher journal was maintained to capture the researcher's impressions and perceptions following each interview. The study used the period that included the start of the epidemic (around 1981) through to the said infection chronicity (around 1998), as well as up to the interview period, 2008-2009.\n\nData were produced using the complete audio-taped transcription of each interview assisted by \"word processing software.\" To allow for a preliminary analysis, the complete transcription of the audio tape for each interview was performed within 24 hr following the interview.\n\nBased on [@bibr40-2158244019827717], as well as her peers ([@bibr26-2158244019827717], [@bibr27-2158244019827717]; [@bibr32-2158244019827717]; [@bibr41-2158244019827717]), the data analysis steps included, in a successive and cumulative manner, reading over the interview transcript (verbatim) and researcher journal after each interview session, to develop a certain familiarity with the data; creating headings manually and using a data processing software (e.g., Atlas-ti); and indicating recurrent ideas, that is, notions, concepts, themes, and so on. This process was used after each interview to readjust the interview grid so that new information could be obtained where gaps were identified. This back-and-forth movement between data collection and analysis was maintained throughout the data collection phases.\n\nDuring the data analysis for each transcript, an iterative process was used to separate the data into the following categories: units of meaning, coding, categorization, and established relationships between them (e.g., commonalities, differences, and patterns) ([@bibr29-2158244019827717]). Following the last interview, an in-depth data analysis was conducted for each transcript and between transcripts.\n\nThe study's scientific rigor and trustworthiness were established ([@bibr40-2158244019827717]) through a process which included peer reviews (the research team) of every interview transcript, discussions of the research process, and presentations of the data's interpretation. If any modifications to the established research protocol were needed, they were suggested and agreed upon by the research team before implementation. Finally, physicians who were interviewed were asked to review the draft data interpretation of their interview for their comments and suggestions. Two physicians presented their comments and suggestions.\n\nResults {#section5-2158244019827717}\n=======\n\nThe following two broad-meaning categories emerged from the data analysis of the semi-structured interviews conducted during [@bibr29-2158244019827717] study: namely, *Category 1*: the transformation of the physicians' identities, both personal and professional; and *Category 2*: the transformation of the relationships physicians nurtured with their patients (e.g., physician--patient therapeutic relationship) and with various networks (e.g., professional, pharmaceutical, and community networks). These categories are explained and illustrated in the following sections.\n\nCategory 1: Transformation of the Physicians' Identities: Personal and Professional {#section6-2158244019827717}\n-----------------------------------------------------------------------------------\n\nFollowing the growth of the pandemic, physicians reported noticing the transformation of the perspective they held of their personal and professional identities. At the personal level, physicians were led to reconsider and alter their medical practice. For some physicians, this meant, first, that they would devote more time to the treatment and care of patients from the gay community to which they belonged and, second, that they would demonstrate their commitment to and solidarity with their community by advocating and lobbying personally on their behalf. In other words, at the personal level, these physicians were involved not only as physicians but also as openly gay physicians.\n\n\". . ., we got involved . . ., that was our involvement (Doctor Antoine). (p. 150). . ., at the start of my career, I was always open about my sexuality, but then, I was openly gay. I was no longer scared of being an openly gay physician . . . There were many gay physicians who decided to get involved in the treatment of HIV-infected gay patients. We took control of our patients. We became their spokesperson (Doctor Denis).^[3](#fn3-2158244019827717){ref-type=\"fn\"}^ (p. 150)\"\n\nAt the professional level, a paradigm shift took place concerning their role. First, they became committed to the development of professional knowledge related to medications (drugs) and treatments. In Doctor Claude's words, \"We participated in the development of almost all new drugs\" (p. 151). They started conducting research protocols with the patients' voluntary participation. Therefore, they went from being medical practitioners to being medical researchers practitioners.\n\nSecond, physicians were called upon to play other roles besides the traditional doctor role, such as social worker, psychologist, guide, counselor, and trainer, as well as educator.\n\n\"At that time, we were not only physicians . . .; we were also social workers (Doctor Antoine). (p. 156). . . even if you are a doctor, all you can be, is a guide. . . . In many cases, we are also the psychological support, the moral support (Doctor Denis). (p. 156). . . our role is to offer advice (Doctor Edgar). (p. 156). . . I am a guide. I am a trainer for my patients (Doctor Fran\u00e7ois). (p. 156)I take care of . . . their medical education . . . I tell them to protect themselves and others with whom they are having unprotected sex. I talk to them about the potential risks to themselves, such as being contaminated by another HIV strain during unprotected sex. I then talk to them about the possible impacts that they were not aware of, one day, that they may have transmitted the HIV infection to someone else. My reaction was to do some sex education (Doctor Claude). (p. 167)\"\n\nBecause of the unusual high occurrence of death among their patients, physicians faced disorienting dilemmas that led them to modify some of the perceptions that they held about their medical practice. The following are examples of various dilemmas they encountered.\n\n### A dilemma about their perception of death and time {#section7-2158244019827717}\n\nThis dilemma takes into account their felt inability to do anything within their medical practice to reduce the infection's occurrence or mitigate its psychological and physical effects on their patients.\n\n\"We saw what it was like to lose patients, as we lost many. My worst episode was when I lost fourteen patients in ten days. It was awful. It meant that when a patient who entered your office was sick, you would send him to the hospital. The next patient was the friend of the previous patient. He would tell you that the previous patient died. And him as well, you would send him to the hospital and he would die. . . . It was like that non-stop. Afterward, it was the hospital staff who would call you to tell you that one of your patients died. . . . I learned to live one day at a time (Doctor Antoine). (pp. 152-153)There was this kind of fatalism in the air non-stop (Doctor Bernard). (p. 152)What we told patients was that they were going to die in a few months, or in a few years, if all went well. Therefore, it was a very dark era (Doctor Claude). (p. 152)Like everyone's life, it is the length that is the difference (Doctor Denis). (p. 153)\"\n\n### A dilemma about their perception of euthanasia {#section8-2158244019827717}\n\nThis dilemma led them to consider, at the risk of their own career, the demands made by their patients, their patients' friends, or their relatives, to assist them in dying so they may die in dignity.\n\n\"What are they \\[the patient's friends\\] going to do with what I gave them? Are they going to cause the patient's death? At the same time, I was thinking, if the patient is not well, maybe it is better that he dies. But at the same time, I was aware that euthanasia was illegal (Doctor Claude). (p. 153)\"\n\n### A dilemma about their perception of risk {#section9-2158244019827717}\n\nPhysicians realized that they would incur risk in exercising their profession as long as lasted the scientific and medical uncertainties, at the time of the pandemic, concerning the way that the infection was transmitted, often placing themselves in situations that could possibly endanger their own health and life.\n\n\"I was scared for my own health every time I was in contact with a patient because at that time, I did not know the modes of transmission (Doctor Claude). (p. 154)There was no reason why I could not fall sick to this disease. At the beginning, I didn't know if I could catch it while touching a patient. I never stopped being close to them. But, I didn't know if, by accident, I would get infected and what impact it would have on me (Doctor Denis). (p. 154)\"\n\n### A dilemma about their perception of aggressive and futile therapies {#section10-2158244019827717}\n\nPhysicians realized that they should cease to impose their will on their patients and begin to determine when they should stop therapies, treatments, and medications as they did not improve the patient's health.\n\n\"At a certain moment, I realized that I had to stop wanting to cure a patient at all cost. As long as I thought that there was a chance that I could cure the patient, I would try other medications. I would behave this way because when I would succeed in healing a patient, it was well perceived by other doctors (Doctor Denis). (p. 154)I thought that it was important to realize that there was a limit to how much medication I could administer to a patient when the medication was no longer efficient to cure the disease (Doctor Edgar). (p. 154)\"\n\n### A dilemma about the traditional images that physicians held about places to practice medicine {#section11-2158244019827717}\n\nPhysicians became aware that they had to go back and forth between private clinics, local community health centers and hospitals, and less conventional places to treat their patients, such as shelters, homes and hospices, private residences, and prisons.\n\n\"We monitored patients at home. . . . We went to . . . shelters for AIDS patients (Doctor Denis). (p. 156)I also worked in a prison where some of my patients were (Doctor Claude). (p. 156)\"\n\n### A dilemma about their physical and professional limitations {#section12-2158244019827717}\n\nIn a context where the number of HIV/AIDS cases was constantly increasing, physicians became acutely aware of their limitations as their patients were dying, and they neither knew the virus that infected them nor the many opportunistic diseases that were affecting them.\n\n\"I am not all mighty . . . (Doctor Claude). (p. 157). . . a doctor can do a lot, but there are limits to what we can do. We cannot stop everything, death or the infection. It is sometimes difficult to change the course of the infection (Doctor Edgar). (p. 157)\"\n\n### A dilemma about scientific and medical certainties {#section13-2158244019827717}\n\nPhysicians were practicing in an environment where medication was more and more numerous and constantly changing ([@bibr23-2158244019827717]), where data (evidence-based) to support the decisions they were making (e.g., on the start and stop of treatment) were missing, and where the knowledge about the medication they were prescribing and its food interactions was not available.\n\n\"Often, I would tell my patients that the drug they are taking now has a specific value in time. Possibly one day, this drug will be completely obsolete (Doctor Antoine). (p. 159)When I did not have information on the interaction between a drug I was prescribing and certain foods, such as garlic, I would tell my patients that it was better to be careful and not to eat it every day (Doctor Antoine). (p. 159)\"\n\n### A dilemma about their perception of the Faculty of Medicine and the College of Physicians {#section14-2158244019827717}\n\nPhysicians recognized the limits of the training they received from the Faculty of Medicine, as well as the College of Physicians' poor understanding of the physicians' daily realities when facing this new medical situation that required, among other, alternative practices.\n\n\"At the Faculty of Medicine, we did not learn to manage a large number of deaths and moreover a large number who were young (Doctor Antoine). (p. 159)In the field of complementary and alternative medical therapies, doctors do not have much training (Doctor Edgar). (p. 160)\"\n\n### A dilemma about medicine in general {#section15-2158244019827717}\n\nAccording to these physicians, the medical profession was still paternalistic ([@bibr21-2158244019827717]), often blaming patients for their mishaps and sustaining prejudice against the gay community.\n\n\"Medicine is quite paternalistic. The doctor in his medical practice is quite paternalistic. . . . in the medical domain, patients are often blamed for what happens to them (Doctor Edgar). (p. 160)I was in a system that was old school. You could be gay, but you could not show it. You did not invite your lover to suppers at the Department of Medicine. You went with a girl, you went alone, or you didn't go at all (Doctor Denis). (p. 160)\"\n\nCategory 2: Physician--Patient, Professional, Pharmaceutical, and Community Network Relationships {#section16-2158244019827717}\n-------------------------------------------------------------------------------------------------\n\nFollowing the growth of the pandemic, physicians mentioned perspective transformations in the physician--patient therapeutic relationship, patient empowerment, and medical therapeutic conversations, as well as in their professional, pharmaceutical, and community network relationships.\n\nIn their medical practice, physicians were compelled to transform the physician--patient therapeutic relationship because of the significant role social movements against HIV/AIDS played in HIV/AIDS information dissemination to its membership, and the growing number of informed and empowered patients they treated. This situation compelled physicians to adopt a more egalitarian therapeutic relationship. In fact, they started engaging in sustained and challenging therapeutic conversations with their patients. Physicians realized that they could not actively continue to confront, judge, or even ignore them. Therefore, they had to explain the decisions they made regarding their patients' health and pay more attention to their choice of words when having to provide advice, discuss options, and negotiate treatments and medications, as well as the start or cessation of a treatment and/or medication.\n\n\"When you treat those patients, it is more stimulating. When you tell those patients to take such and such a drug, they ask you why they must take that one and not another one. You must explain your decision (Doctor Antoine). (p. 207)I explain more my train of thought and the drug that I am going to prescribe when I treat those patients. I take more time to explain how I arrived at that decision (Doctor Bernard). (p. 207)What I learned is that it is important to be careful not to be confrontational with those patients. . . . I am sometimes in a situation where I must learn to negotiate with those patients, for example, when to start or to end a treatment (Doctor Claude). (pp. 168, 207)We do not judge patients (Doctor Antoine). (p. 168)What I learned is that I must . . . respect the patient who got infected and who lives with the HIV infection (Doctor Denis). (p. 168)\"\n\nPhysicians also modified their medical practice to strengthen the criteria surrounding privacy, while continuing to comply with the requirements prescribed by law (mandatory reporting), to address the new patients' demand (non-disclosure).\n\n\"The doctor-patient relationship is one that stays between four walls. Very often, we are the only ones who know that someone is HIV-positive. Their workplace doesn't know it. Their friends don't know it (Doctor Antoine). (p. 165)The patient did not want the reason for his death mentioned on his death certificate (Doctor Denis). (p. 165)\"\n\nPhysicians realized that the border (therapeutic distance) between themselves and their patients was gradually blurring as more and more of the patients who sought their HIV/AIDS expertise had often been their friends, lovers, colleagues, or co-gay community members.\n\n\"It was a clientele to which I identified a lot with because these men were like me, at that time, young gay men who were discovering themselves a little, who were starting to live their gay life (Doctor Claude). (p. 165)The doctor-patient border was not very strong. Often there was none. I began to see friends at the clinic . . . And I remember when a great friend of mine, a work colleague, came to see me at the clinic (Doctor Denis). (p. 166)\"\n\nMoreover, physicians often had a \"history\" with these individuals. They were their patients before they were diagnosed, up to their death and beyond, as many physicians started attending their patients' funerals, a rare occurrence before the pandemic.\n\n\"I am close enough to my patients. . . . I developed a stronger bond with some of them. When you have been treating them for over twenty years, you develop a stronger bond with them (Doctor Claude). (p. 166)It was important that I be there, when HIV was discovered, during the period of opportunistic infections, and at the end of the HIV trajectory, to be able to say goodbye. I attended my patients' funerals (Doctor Denis). (p. 166)\"\n\nThrough this renewed physician--patients therapeutic relationship, physicians developed a greater understanding of their patients' distress/helplessness. Moreover, they became cognizant of their non-inhibited sexual behaviors.\n\n\"My patients tell me everything, their intimacy, the emotions related to their sex life, their preferred drugs, and their cruising habits. . . . A patient started telling me that he went to the sauna . . ., where everyone, it seems, goes to have sex. It is not well looked upon for someone to put on a condom in that sauna (Doctor Claude). (pp. 170-171)A patient told me that he had had many partners who did not protect themselves . . . (Doctor Fran\u00e7ois). (p. 170)Patients would tell me about the rave parties they attended . . . (Doctor Denis). (p. 170)\"\n\nThey also learned about their patients' rejection, on a regular basis. In Doctor Edgar's words, \"Patients would tell me about the rejection they face from society in general and from the gay community in particular . . .\" (p. 171).\n\nAt some point during the pandemic, some physicians understood that they had to start rebuilding the boarder (therapeutic distance) which had once existed between themselves and their patients to be more effective and efficient. In fact, this therapeutic distance was essential in repositioning the physicians' role, in changing certain habits, or to educate their patients on health, sexuality, and blood-transmitted infections, among other topics.\n\n\". . . physicians had to start rebuilding the therapeutic distance, a distance between a doctor and the infection of his patients, the distance between a doctor and his patients . . . (Doctor Bernard). (p. 167)\"\n\nPhysicians also noted and understood the potential role reversal that could take place. At times, the traditional physician--patient role was reversed when patients became aware that their physician was possibly more ill (e.g., workload or HIV-infected) than they were, or when they were simply concerned about their physician's physical and mental well-being.\n\n\". . . it felt like the patient was almost trying to comfort me . . . (Doctor Claude). (p. 168). . . it was the patient who was taking care of the doctor. As if he thought that the situation was too intense for the doctor (Doctor Denis). (p. 169)\"\n\n### Professional, pharmaceutical, and community network relationships {#section17-2158244019827717}\n\nThe data showed that physicians were also obligated to transform the way that they viewed and dealt with other key networks. They had to develop new contacts with professionals who were sympathetic to their cause, fighting HIV/AIDS (e.g., pharmacists and medical specialists), not only to find technically relevant information but also to obtain reliable and supportive collaborations.\n\n\". . it was really through those years that we learned which doctors were sympathetic to our cause. They were the doctors who were not prejudicial to our patients. We developed a network, people we knew, teams or specialists to whom we had easier access. . . . I had to let go of control to enable the specialists to do what needed to be done (Doctor Denis). (pp. 173-174)The neurologist who evaluated the patient was very good with patients who were HIV-infected (Doctor Antoine). (p. 173)\"\n\nTheir professional networks were also extended to pharmaceutical companies. However, they needed to modify the perception they held about them, from an industry whose primary focus was profit maximization to one whose focus was more humanitarian, focused on patients. They also needed to realize that the pharmaceutical industry was moving from an industry reliant on universities and private sector medical research and development to one more reliant on physicians for the same purpose.\n\n\"If you have a new patient who does not yet have access to the drugs he needs, the pharmaceutical companies are there to help. I once phoned all the pharmaceutical companies to get access to a drug before it was available because I had a patient who was really sick. . . . It was possible (Doctor Edgar). (p. 174). . . if we hadn't participated in that research protocol, if we didn't have a structure that allowed us to do research, which is not always possible in a private clinic, our patients would have had to wait three years. Some would have died because of the lack of access to medication. Therefore, it was fantastic to save all those patients . . . (Doctor Claude). (p. 175)\"\n\nFinally, physicians also had to add to their networks gay community organizations, as they were also involved in the prevention and fight against HIV/AIDS. In fact, these organizations became partners and information disseminators, which often came directly from the physicians themselves.\n\n\". . . activists, we knew them well. We kept in contact, with the gay community organizations against HIV and we worked with them (Doctor Denis). (p. 177)We knew the information that they had in that organization, because it was us, the group of doctors, who wrote the articles circulating there. We work in close collaboration with them. They are very useful (Doctor Antoine). (p. 177)I used, for a long time, the services like those offered by that organization or those support groups. . . . I did many information sessions directly to patients, thanks to that organization . . . (Doctor Claude). (p. 177)\"\n\nTo sum up, the data showed that the physicians transformed their personal and professional identities, as well as their practice to take into account the growing number of HIV-infected patients they treated who were empowered and informed, because of the social movements against HIV/AIDS that supported them.\n\nThey also transformed the relationships they nurtured with their patients (e.g., physician--patient therapeutic relationship) and with various networks (e.g., professional, pharmaceutical, and community networks).\n\nDiscussion {#section18-2158244019827717}\n==========\n\nThe following section brings together three types of findings for discussion. The first section compares differences between the themes that emerged from the study and those that resulted from the analysis of the previously mentioned autobiographies. The second one presents new findings obtained through a recent literature review. Finally, the third section shows that some themes raised by this research article still persist today (recurring themes).\n\nDifferences {#section19-2158244019827717}\n-----------\n\nThe themes (identified in the autobiographies that were used to prepare the semi-structured interview grid and to create the initial conceptual taxonomy that was used to organize and analyze data) are almost a mirror image of the themes that emerged from the data analysis of the interviews. These themes highlighted the impact of the emergence of HIV/AIDS; emphasized the uncertainties generated by a multiplication of scientific and medical discoveries/findings; talked about the perceptions physicians held about sexuality, drug use, and HIV/AIDS; and have drawn attention to the way they practiced medicine, going from a physician-centered practice toward a patient-centered practice.\n\nHowever, the following themes emerged from the data of the study of some Canadian gay HIV/AIDS dedicated physicians' journeys but did not appear in the physicians' autobiographies. Under the first category mentioned in the study, the transformation of the physician's personal and professional identities, the study brought to light the militant implication of these gay physicians as they advocated and lobbied for their HIV/AIDS gay patients. As they recognized themselves as being gay physicians, they reconsidered and altered their involvement in their patients' medical, political, and social causes/battles (e.g., treatment, medication, and rights). In doing so, they demonstrated their solidarity to the gay community. To strengthen their ties with their community, they also modified their medical practice by treating their patients in conventional (e.g., private clinics, community health centers, and hospitals) and unconventional work environments, namely, in their patients' residences/homes.\n\nThe study also underlines that during the pandemic, these gay physicians recognized that pharmaceutical companies had acquired a more open attitude regarding their participation in research protocols. Thereafter, a paradigm shift took place: from medical professionals they became professional-practitioner researchers. They became committed to the development of professional knowledge related to medications (drugs) and treatments.\n\nFinally, as they were constantly confronted with a fatal disease, they recognized that there might be a place and a time where euthanasia would be appropriate, even though it was still illegal at the time.\n\nUnder the second category mentioned in the study, the transformation of the physician--patient relationship, as well as the professional, pharmaceutical and community networks, the study brought to light how gay HIV/AIDS dedicated physicians made more place for discussion about the HIV/AIDS patients' lifestyle (e.g., sexuality, homosexuality, cruising, raves, drugs, and saunas). They also modified their therapeutic conversations by being less paternalistic, refraining from blaming patients for what happened to them, and by not discriminating and stigmatizing their patients. While doing so, they also blurred the traditional border that existed between themselves and their patients, such as by going to their patients' funerals. They built networks of professionals who were sympathetic to their causes/battles and developed links with social movements to increase the support networks for their patients. They became directly involved in gay community organizations against HIV/AIDS and learned to play an active role as openly gay informants and activists to disseminate information about HIV/AIDS.\n\nThe results from the study, unlike the physician's autobiographies, make no mention of the role spirituality or religion played while patients were facing a life-threatening and limiting infection such as HIV/AIDS, neither did it mention the reflections patients and physicians provided in relation to it. Yet, the place of spirituality and religion in the lives of patients facing a life-threatening and limiting infection such as HIV/AIDS and their physicians is often mentioned in the autobiographies examined for the study (e.g., [@bibr6-2158244019827717]; [@bibr11-2158244019827717]; [@bibr49-2158244019827717]; [@bibr57-2158244019827717]).\n\n\"Then, one day, tired of seeing my friends and patients who were only 25, 30, 35 years old die one after the other from AIDS, I questioned my faith and felt anger towards God. So, I looked at the sky and, whether God exists or not, I spoke to him that day. I asked him, \"Why is this happening? Why is everyone I love dying? If you existed, you would not be doing this. It's too unfair.\" ([@bibr11-2158244019827717], pp. 59-60)\"\n\nNew Findings {#section20-2158244019827717}\n------------\n\nThrough a recent literature review pertaining to gay nurses caring for gay HIV-infected patients, three research papers surfaced about physicians who attempted to treat patients in the context of the pandemic. The first one, by [@bibr61-2158244019827717], is titled *Physicians Reflect on Their Lived Experiences in Long-Term Care*. The second one, by [@bibr43-2158244019827717], speaks about *M\u00e9decins et patients au temps du sida: le cas de Montr\u00e9al*; and the third one, by [@bibr36-2158244019827717], addresses the *Pratiques professionnelles m\u00e9dicales et VIH/sida: des t\u00e9moignages \u00e0 la fiction romanesque*. However, they never openly breach the topic from the same specific angle as Jacob's study, namely, the lived experiences of gay HIV/AIDS dedicated physicians attempting to treat gay HIV/AIDS patients.\n\n[@bibr61-2158244019827717] present the physicians' reflection on \"their professional histories in caring for patients with AIDS\" (p. 278). Even though the physicians they interviewed included one gay physician, the data were not segregated using that variable. This research article's interest resides in the additional information pertaining to the following categories that are described below.\n\n### Reasons for entering the AIDS care frontline {#section21-2158244019827717}\n\nThese physicians adopted the AIDS care domain as a natural progression of knowledge and skills as they had \"prior experience and knowledge in virology, epidemiology, or infectious disease\" (p. 278), and because it was at the \"cutting edge of medicine\" (p. 279). They also mentioned that they chose that domain of work because they felt that \"there was a heightened sense of connection with themselves, their patients, and the personal and professional values and beliefs that guided their lives and medical practices\" (p. 279).\n\n### Rewarding aspects of AIDS care {#section22-2158244019827717}\n\nOnce physicians had learned to deal with the numerous stressors and emotional consequences of AIDS care, they felt that \"the rewards in caring for patients with AIDS outweighed the stressors\" (p. 281). The rewards that they mentioned included \"developing long-term, reciprocal relationships with patients, developing partnerships in care, experiencing personal and professional growth, and a sense of gratification\" (pp. 281-282).\n\n### Aspects of the physician--patient therapeutic relationship and patient characteristics that promote a willingness and commitment to care {#section23-2158244019827717}\n\nPhysicians mentioned that many aspects of the physician--patient therapeutic relationship and patients' characteristics may influence a physician's willingness, desire, and commitment to attempting to treat patients, such as \"'patients who demonstrate self-care,' 'patients who trust my \\[their\\] care,' 'patients who value our \\[their\\] relationship,' and 'patients who teach us \\[them\\] how to die with grace and dignity'\" (p. 283).\n\n### Recommendations offered by physicians specializing in the AIDS care domain {#section24-2158244019827717}\n\nPhysicians offered many recommendations to their colleagues considering the AIDS care domain as a future profession. Such examples include working on caring (e.g., empowering patients, listening to them, and talking to them in a language they can understand), learning to turn mistakes into lessons (e.g., using reflection on action, and realizing that one is human), being realistic yet optimistic (e.g., admitting that one doesn't have all the answers, and dealing with uncertainty about the patient's condition with honesty), and learning to say no when you need to protect oneself (e.g., valuing and augmenting one's personal relationships) ([@bibr61-2158244019827717]).\n\n### Lessons learned {#section25-2158244019827717}\n\nSometimes, physicians learned more from what was not said than from what was said. In fact, they realized that there was no \"justice in this epidemic\" (p. 284), and as such, the virus did not discriminate. Such realization helped them to accept patients the way they were and, more importantly, that all patients deserved to be treated and cared for, and to receive quality care. As there was no mention of God or spirituality---like in Jacob's study---in either \"causing the epidemic, or alleviating the suffering\" (p. 284), physicians seem to have chosen this path of life based on \"conscious choices,\" and to be dedicated to their field for \"humanistic values\" (p. 284). However, such a life choice and dedication did not prevent them from burning out, experiencing depression, or marital conflicts.\n\n[@bibr43-2158244019827717] presents a study conducted in Montreal, in 1989-1990, where the data were not segregated using the variable of homosexuality. It attempted to comprehend how physicians and patients perceive their relation, the disease and death, and care and medical research. The interest of this research resides in the fact that it mentions experiences not previously accounted for in other studies.\n\n### Medical personnel obligations {#section26-2158244019827717}\n\nDuring the pandemic, a large societal debate was underway to determine the medical personnel's obligations, and, in particular, the physician's obligation to attempt to treat patients, as well as their freedom to choose their patients.\n\n### Reasons physicians entered the AIDS care domain {#section27-2158244019827717}\n\nLike [@bibr61-2158244019827717], [@bibr43-2158244019827717] provides reasons why physicians would decide to specialize in the AIDS-care domain. They perceived that they would have access to interesting medical cases, as AIDS was a major scientific event that could bring them peer recognition. Finally, some new specialists considered the AIDS domain as a means to access research funds and consolidate their research notoriety.\n\n### Cultural differences and alternative medicine {#section28-2158244019827717}\n\nIt was noted that most physicians adopted a degree of openness to diverse alternative medicines, and even supposedly \"miracle medications\" (p. 77), but without severing bridges with traditional medicine. However, there might have been a cultural bias in looking for such alternatives, as Francophone physicians were more prone to have confidence in \"natural\" treatments then their Anglophone counterparts.\n\n### Honesty in the context of research trials {#section29-2158244019827717}\n\nThe need to be \"very\" honest (p. 80) with HIV patients was more than necessary, as new research trials had a \"50/50 chance\" of success (p. 80). In this sphere of uncertainty, physicians, as well as patients, were desperate to try something---anything new. They had to be honest to themselves, as physicians, and their patients needed to be at the forefront of these research trials because they had never been done before.\n\n[@bibr36-2158244019827717] present a paper that triangulates sociological, autobiographies, and novel narratives to extract the transformations of the medical practice, the place of the affect related to the illness and death, the physician--patient therapeutic relationship, and its representation during the pandemic's evolution.\n\nThe data extracted from these narratives were not segregated using the variable of homosexuality. This research is interesting because it speaks about the emotional sphere in terms of its connection to the disease, death, and the relationship with their patients.\n\n### Affective challenges {#section30-2158244019827717}\n\nThe emotional sphere mentioned above speaks directly to the ambivalent feelings physicians experienced while attempting to treat their patients. While these feelings are comparable to those mentioned by patients, such as rage, agony, dissension, sadness, acceptance, revolt, and hope, they also include compassion, fear, and culpability.\n\nThrough the same literature review, two books also surfaced: one by Dr. Johnson, titled *Working on a Miracle*, explaining his combat against the HIV infection and his search for a treatment, and the other by Dr. Waddell, called *Gay Olympian: The Life and Death of Dr. Tom Waddell*, describing his life quests (e.g., \"coming out,\" a career, a family), as well as his battle against the infection.\n\nIn his book, [@bibr30-2158244019827717] describes his odyssey and \"apparently successful fight against HIV\" ([@bibr13-2158244019827717], p. 1). Four years earlier, \"an autopsy on an HIV-infected corpse\" (p. 1) went wrong. When his scalpel slipped, he went from being a doctor to being a patient ([@bibr30-2158244019827717]).\n\nThe interest of this book resides in the fact that Johnson formulated his \"battle plan to fight the disease\" ([@bibr13-2158244019827717], p. 1), which \"ran contrary to conventional treatments\" (p. 1). It consisted in immediately starting an aggressive treatment, at a time when the tendency was not to start treatment immediately, of a \"veritable alphabet soup of drug therapies\" (p. 1) which included 28 pills per day.^[4](#fn4-2158244019827717){ref-type=\"fn\"}^\n\nJohnson did not allege having found a cure for AIDS. In fact, at the time of Johnson's discovery, other AIDS experts,\"Warn\\[ed\\] against unrealistic optimism, fearing the disease may be dormant or hiding out in another part of his body, such as the lymph nodes. ([@bibr13-2158244019827717], p. 2)\"\n\nTwenty-one years later, Dr. Johnson is still working as a pathologist and a Professor at the University of Rochester School of Medicine & Dentistry in Rochester, New York.\n\nFinally, [@bibr58-2158244019827717] book presents, in the first person, the trials and tribulations of a teenager realizing that he might be gay, and his subsequent coming out, as a gay physician, once he had full-blown AIDS. This book is interesting because it is written by a gay physician who passed away from an AIDS-related infection. Waddell used his book to transmit to his daughter, and to the world, his vast knowledge, using his extraordinary ability to teach and translate knowledge. Waddell explains the full chronology of the progression of the infection and his attempts to cure/heal it. He also gives explicit examples of his lifestyle before and during the epidemic, up until his death. The book also confirms that Waddell, like some gay physicians, wrote \"articles on the subject for the California *Voice*, and I am trying \\[tried\\] to bring some common sense to the crisis and offer \\[offered\\] people ways to begin to alter their behaviour in the least traumatic way\" (p. 33). There was discrimination, stigmatization, and prejudice toward AIDS victims but, as Waddell points out, there were also poignant reunions with past lovers and friends because of the disease. His book also eloquently presents the state of panic that was present, at that time, in Castro's gay community, in San Francisco, California.\n\nWaddell also talked about how the gay community had to evolve to stay alive. \"Those who do not change their sex practices will die\" (p. 169). \"In this era of AIDS, it's sensible to be monogamous or even celibate\" (p. 180). He also mentioned how he was \"inundated with 'cures' from friends and strangers---some bizarre, some logical, all ineffective\" (pp. 204-205, 220). Spirituality, or the lack thereof, was also commented on by the author in the following manner, before closing his eyes for the last time, \"Well this should be interesting\" (p. 223).\n\nWithout a doubt, these narratives, and the information they contain, would have positively impacted [@bibr29-2158244019827717] study, by providing more information in the preparation of the interview grid, the categorization for data analysis, and more opportunity for disclosure.\n\nRecurring Themes {#section31-2158244019827717}\n----------------\n\nSome themes raised by this research article still persist today (recurring themes). Among them are the absence of services to key populations, discrimination and stigmatization, the lack of regulatory framework, and the importance of communication in the physician--patient therapeutic relationship. In [@bibr62-2158244019827717], the World Health Organization (WHO) acknowledged that health services to key populations are often not tailored to or integrated into other health and social services, nor do they meet the needs of key populations in terms of their values, desires, or preferences. As well, they do not address their practical constraints.\n\nFurther in the report, WHO recognized that structural barriers to health services, such as stigmatization and discrimination, continue to exist and they must be removed to improve the knowledge of \"serostatus\" in sex workers, men who have sex with men, transgender people, and drug injection users (key populations), while promoting and protecting human rights and promoting access to health services.\n\nIn the same report, WHO highlighted that many countries lack a regulatory framework, such as laws, policies, regulations, or interventions to protect human rights, to enable the rapid regulatory approval of new and generic medicines and diagnostics, and to expedite their marketing approval.\n\nFinally, in [@bibr63-2158244019827717], a study by Young reminded us of the importance of a physician--patient therapeutic relationship based on communication.\n\n\". . . optimal health requires attention to provider-patient communication across the entire HIV care cascade. While care cascade metrics begin with testing and end with viral suppression, healthy aging with HIV requires attention to aspects of health and healthcare communication. (p. 1)\"\n\nThese recurrent themes converge toward one central theme that is the patient's needs. It is through the patient's needs that health and social services are recognized and should be improved; that discrimination and stigmatization happen and should be addressed; that human rights are challenged and should be fought; and that communication with physicians occurs and should be evaluated.\n\nResearch Implications {#section32-2158244019827717}\n=====================\n\nIt is important to emphasize that the study conducted by Jacob in 2012, and on which this research article is based, revealed research implications that are equivalently associated with the patients' needs, the medical practice, and various stakeholders. A closer look at these research implications shows that the treatment of patients, especially in time of crisis, challenges the traditional evidence-based medicine and provokes new practices. As seen in this research article, these practices, based on lived journeys, have produced new scientific knowledge that increases the professional--patient relationship and the value of the patient who becomes more involved in the process of healing (patient informant and patient empowerment).\n\nThe lived journeys of these six Canadian gay physicians will serve the medical community and the public health domain when they attempt to treat, care for, and cure/heal patients who are infected with future diseases. This is especially true where at first, little institutionalized, evidence-based, knowledge is available and patient stigmatization and discrimination is prevalent, for instance, the severe acute respiratory syndrome (SARS), the human infected H1N1 influenza, and the mosquito-borne ZIKA. This reference to past outbreaks will help today's physicians approach these new diseases with a more human touch and a view to the possibility of being transformed personally and professionally (e.g., militancy and researcher practitioner). It will also encourage them to remain open to modified relationships (e.g., a new approach to their patients as informants and trainers, their colleagues, other professionals, and community organizations) that could help them in their battle against the infection.\n\nOther Research Interests {#section33-2158244019827717}\n========================\n\nBecause of the interest created by [@bibr29-2158244019827717] study, as well as the research papers and books that surfaced since, we are now moving our research interest from gay physicians to gay nurses' lived journeys while caring for gay HIV-infected patients during the HIV/AIDS pandemic. This new research article's interest lies in the fact that nurses have an increased proximity with patients; therefore, they may have had similar or different lived experiences with gay HIV-infected patients than their gay physician counterparts.\n\nOur literature review has already uncovered one book written by a gay nurse and one written by a lesbian nurse ([@bibr18-2158244019827717]; [@bibr56-2158244019827717]). No scientific papers using gay nurses as subjects have surfaced yet. However, one book and a few articles were found with nurses caring for HIV/AIDS patients as subjects ([@bibr8-2158244019827717]; [@bibr39-2158244019827717]; [@bibr60-2158244019827717]).\n\nConclusion {#section34-2158244019827717}\n==========\n\nThe objective of this research article was to provide the lived journeys of Canadian gay HIV/AIDS dedicated physicians in the context of the HIV/AIDS pandemic in an effort to draw a better picture of what it was like to be a gay physician during that particular period. The research background was based on books and autobiographies that set the stage for the HIV/AIDS context. In addition, they provided an account of the physicians' trajectories/journeys/experiences as they to treat, care for, and cure/heal HIV-infected patients throughout the pandemic.\n\nIn the context of the HIV/AIDS pandemic, the Canadian gay physicians' personal and professional identities and the way they viewed their practice were transformed by their lived journeys. They became openly gay HIV/AIDS dedicated and militant physicians, advocating and lobbying for their patients. They altered their physician--patient therapeutic relationship, making it more open and egalitarian. They modified their relationship to their networks, namely, community organizations, other colleagues, pharmacists, and the pharmaceutical industry, by increasing their personal and professional involvement in gay community organizations against HIV/AIDS, building a network of professionals able and willing to care for their patients, and becoming physician researchers practitioners.\n\nThe data presented and the results interpreted offered a candid insight into the lives of Canadian gay HIV/AIDS physicians, at a time when the medical world was not prepared for such a devastating disease. Moreover, this research has highlighted issues that were critical to bringing changes in the way physicians saw their medical practice. The therapeutic relationship between physicians and patients is a good example of this type of change, as both, physicians and patients, had to re-define what to expect from each other.\n\nAs people read these physicians' lived journeys, they are reminded of a time when gay physicians who practiced during the HIV/AIDS pandemic were militants, advocated for their gay HIV/AIDS patients, and lobbied for a cure. This article lifted the veil of darkness that shrouded the contribution made by these unsung heroes, shining of spotlight on their selfless dedication to their patients in the hope that readers will recognize their sacrifices and come to realize that they were personally and professionally affected and forever transformed because of their involvement in their attempts to treat, care of, and cure/heal gay HIV-infected patients during the HIV/AIDS pandemic.\n\nFinally, according to [@bibr22-2158244019827717], these narratives are a reminder that the full \"story of \\[HIV/\\] AIDS cannot be written yet \\[until a\\] cure, curative, functional cure, or vaccine\" (pp. 219-220) against HIV/AIDS is found, there is a need to re-count the way it was and how it is now translated into our present.\n\n**Declaration of Conflicting Interests:** The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.\n\n**Funding:** The author(s) received no financial support for the research, authorship, and/or publication of this article.\n\n###### \n\nHIV-AIDS Pandemic Trajectory: Three Periods.\n\n![](10.1177_2158244019827717-table1)\n\n  -----------------------------------------------------------------------------------------------------------------------------------------\n  **Period 1---HIV/AIDS Emergence**\\\n  [Patients]{.ul}: Gay individuals who contracted the HIV infection\\\n  \u2003--\u2003Isolated and destitute\\\n  \u2003--\u2003Little or badly informed\\\n  [Medical Practitioners]{.ul}: Generalists\\\n  \u2003--\u2003Little solidarity with patients\\\n  \u2003--\u2003Generally little militancy activities\\\n  \u2003--\u2003Interested by knowledge, but little engagement in new required knowledge production\\\n  [Gay Social Movements]{.ul}: Against HIV\\\n  \u2003--\u2003Non-existent or non-visible\n\n  **Period 2---Recognized Pandemic and Militancy Rise**\\\n  [Patients]{.ul}: Gay individuals who contracted the HIV infection\\\n  \u2003--\u2003Militancy rise\\\n  \u2003--\u2003Active search for information\\\n  \u2003--\u2003Participation in conferences, seminars, and workshops\\\n  [Medical Practitioners]{.ul}: Gay generalists\\\n  \u2003--\u2003New medical practice mode (experimental medicine, experiential learning used)\\\n  \u2003--\u2003Committed to knowledge production, not only its use\\\n  \u2003--\u2003Institutionalized knowledge and traditional continuing medical education modes challenged\\\n  \u2003--\u2003Patient solidarity, militancy\\\n  [Gay Social Movements:]{.ul} Against HIV\\\n  \u2003--\u2003Visions and alternative medical knowledge production\\\n  \u2003--\u2003Provides members with a range of services and training\n\n  **Period 3---HIV Chronicity and Militancy Decrease**\\\n  [Patients]{.ul}: Gay individuals who contracted the HIV infection\\\n  \u2003--\u2003Militancy and empowerment decline\\\n  \u2003--\u2003Conferences, seminars, and workshop participation decline\\\n  [Medical Practitioners]{.ul}: Gay generalists\\\n  \u2003--\u2003Initial and continuing medical education integrate new knowledge (knowledge institutionalization)\\\n  \u2003--\u2003Medical practice centered mainly on the preservation of health and the cure of diseases\\\n  \u2003--\u2003Experimental medicine continues, but less engaged in knowledge production\\\n  \u2003--\u2003HIV infection increasingly considered chronic (an \"implicit expectation\" that infected individuals will survive the HIV infection)\\\n  [Gay Social Movements:]{.ul} Against HIV\\\n  \u2003--\u2003Fewer militancy organizations\\\n  \u2003--\u2003Have difficulty recruiting volunteers\\\n  \u2003--\u2003Focused on specialized service provision\n  -----------------------------------------------------------------------------------------------------------------------------------------\n\nAccording to the University of California in San Francisco, a diagnosis of AIDS is made whenever a person is HIV-positive, the CD4+ cell count is below 200 cells per microliter, their CD4+ cells account for fewer that 14% of all lymphocytes, or they have been diagnosed with one or more of the AIDS-defining illnesses (infections, cancers, or neurological disorders).\n\nThe exact title is as follows: Recours \u00e0 l'apprentissage exp\u00e9rientiel dans la transformation de la pratique m\u00e9dicale de six professionnels dans le contexte de la pand\u00e9mie du VIH-Sida (see [@bibr29-2158244019827717]).\n\nAll citations presented in the results section of this article have been loosely translated from French to English.\n\n6 Ritonavir, 3 Saquinavir, 2 gel caps of AZT and 1 gel cap of 3TC at breakfast and at supper; plus 1 gel cap of AZT, 1 gram of vitamin C, a cocktail of vitamin B and one gel cap of beta-carotene at bedtime.\n\n**Carl GA Jacob** is attached to the University of Ottawa's Nursing History Research Unit (NHRU) and is a senior researcher at the Department of Health Canada. His main interests are Physicians and Nurses' Experiences with HIV/AIDS, and Education and Development of Leaders.\n\n**Daniel Lagac\u00e9-Roy** is an associate professor at the Royal Military College of Canada (RMC) and the head of the Military Psychology and Leadership (MPL) Department. His main interests are Military Ethics, Psychology of Religious Conflicts, and Mentoring and Development of Leaders.\n"}
+{"text": "1. Introduction {#sec1-nutrients-08-00243}\n===============\n\nDoxorubicin (DOX), a quinone-containing antitumor drug, is widely used for the treatment of cancer, especially for treating breast and esophageal carcinomas \\[[@B1-nutrients-08-00243]\\]. However, it has been noted that long-term use of DOX tends to induce neurotoxicity and may cause neuropsychiatric diseases including depression, anxiety, and impaired cognition function \\[[@B2-nutrients-08-00243],[@B3-nutrients-08-00243],[@B4-nutrients-08-00243]\\]. Previous studies of patients undergoing chemotherapy for breast cancer have consistently shown their depressed mood and decreased interest in surroundings \\[[@B5-nutrients-08-00243]\\], which highlight the importance of further understanding on the neurotoxic effects of DOX and seeking potential therapeutic strategies.\n\nThe impairment of neurogenesis and increased neural apoptosis in the limbic brain regions, including the prefrontal cortex and hippocampus, is considered as one of the leading causes of depression. It was reported that DOX-mediated generation of free radicals in the brain tissues increases lipid peroxidation and protein oxidation, and alters the antioxidant defense system, eventually leading to neuropsychological changes \\[[@B3-nutrients-08-00243],[@B6-nutrients-08-00243]\\]. Moreover, increased generation of superoxide anions induced by DOX may elevate the level of circulating tumor necrosis factor-alpha (TNF-\u03b1) which can directly pass blood brain barrier (BBB), and activate glial cells to initiate the local production of pro-inflammatory cytokines which exacerbate the oxidative stress and neural apoptosis \\[[@B1-nutrients-08-00243],[@B7-nutrients-08-00243],[@B8-nutrients-08-00243]\\]. In addition, many inflammatory mediators such as TNF-\u03b1 and nuclear factor-kappa B (NF-\u03baB) have been shown to be critically involved in neuroinflammation both in animal models and in patients undergoing chemotherapy \\[[@B9-nutrients-08-00243],[@B10-nutrients-08-00243]\\]. Therefore, in the treatment of DOX-induced depression, an approach involving reduced production of reactive oxygen species (ROS) and inhibited release of neurotoxic pro-inflammatory mediators might be beneficial.\n\nAmong the potent approaches, recent studies highlight the effectiveness of \u03c9-3 polyunsaturated fatty acids (\u03c9-3 PUFAs) supplementation as a potential treatment strategy for brain damage \\[[@B11-nutrients-08-00243],[@B12-nutrients-08-00243],[@B13-nutrients-08-00243],[@B14-nutrients-08-00243]\\]. Dietary intake of \u03c9-3 PUFAs, such as eicosapentaenoic acid (EPA, 20:5, \u03c9-3) and docosahexaenoic acid (DHA, 22:6, \u03c9-3), is known to be beneficial for mental health \\[[@B15-nutrients-08-00243]\\]. Many pharmacological studies demonstrated that EPA and DHA possess chemopreventive and antidepressant activities \\[[@B16-nutrients-08-00243],[@B17-nutrients-08-00243],[@B18-nutrients-08-00243],[@B19-nutrients-08-00243]\\]. These effects have been mainly attributed to their anti-oxidant and anti-inflammatory properties, which had been described in a variety of models \\[[@B20-nutrients-08-00243],[@B21-nutrients-08-00243],[@B22-nutrients-08-00243]\\]. \u03c9-3 PUFAs were shown to lead to a decrease in lipid peroxidation and suppress oxidative stress in brain tissues \\[[@B23-nutrients-08-00243],[@B24-nutrients-08-00243]\\]. Additionally, Irene *et al*. (2014) showed that DHA suppresses production of inflammatory mediators, such as NF-\u03baB, TNF-\u03b1, and iNOS in a mouse model of spinal cord injury \\[[@B12-nutrients-08-00243]\\]. In addition, another study found that \u03c9-3 PUFAs pretreatment could effectively protect the testicular cells from DOX-induced apoptotic injury, suggesting the anti-apoptotic activity of \u03c9-3 PUFAs \\[[@B17-nutrients-08-00243]\\].\n\nBased on the above findings, the aim of this study was to investigate the potential protective effects of \u03c9-3 PUFAs against DOX-induced neurotoxicity and depression-like behaviors in rats. In addition, the possible underlying mechanisms, including anti-oxidant, anti-inflammatory, as well as anti-apoptotic effects of \u03c9-3 PUFAs in brain tissues, were investigated.\n\n2. Materials and Methods {#sec2-nutrients-08-00243}\n========================\n\n2.1. Animals {#sec2dot1-nutrients-08-00243}\n------------\n\nSprague-Dawley rats (Male, 150--180 g; the Experimental Animal Center of the Second Xiangya Hospital) were initially housed in groups in a temperature-controlled environment under a 12/12 h light/dark cycle with free access to food and water except during experimental procedures. This study was approved by the Animal Care and Use Committee of Central South University (protocol number 036/2015). All experiments were performed in accordance with the Guide for Care and Use of Laboratory Animals (Chinese Council).\n\n2.2. Experimental Design {#sec2dot2-nutrients-08-00243}\n------------------------\n\nAnimals were divided randomly into four groups (*n* = 8): (1) control; (2) PUFA; (3) DOX; and (4) DOX + PUFA. The untreated control group was injected with the appropriate volume of the normal saline. Rats in PUFA group received daily gavage of 1.5 g/kg \u03c9-3 PUFAs (EPA 34%, DHA 24%, Sheng Tianyu Biotechnology, Wuhan, China) for three weeks, serving as another control group to exclude any toxic effects. The DOX group was given DOX (Zhejiang Hisun Pharmaceutical Company Limited, Taizhou, China) every two days for a total of seven injections via intraperitoneal injection at a dose of 2.5 mg/kg. The DOX + PUFA group received \u03c9-3 PUFAs daily for three weeks starting one week before giving DOX. The doses of \u03c9-3 PUFAs and DOX were based on previous studies \\[[@B1-nutrients-08-00243],[@B15-nutrients-08-00243]\\]. The body weight of these rats was monitored throughout the experiment, and drug doses were adjusted accordingly.\n\nAt the end of the three weeks, behavioral tests were carried out and the rats were anesthetized with 10% chloral hydrate (4 mL/kg) \\[[@B15-nutrients-08-00243]\\]. Blood was collected and the brain was quickly removed after cardiac perfusion with phosphate-buffered saline (PBS) (pH = 7.2). The left hemisphere of the brain was maintained in 4% paraformaldehyde and then embedded in paraffin, prepared for histopathological examination and immunohistochemical staining. For the right hemisphere, the prefrontal cortex and hippocampus were dissected and used for oxidative stress measurement and Western blot and PCR analysis.\n\n2.3. Behavioral Test {#sec2dot3-nutrients-08-00243}\n--------------------\n\n### 2.3.1. Open-Field Test (OFT) {#sec2dot3dot1-nutrients-08-00243}\n\nThe test was performed in a square arena (90 cm \u00d7 90 cm \u00d7 40 cm) with the floor divided into 25 equal squares by black lines. The rat was placed into the center of the open field and allowed to move freely over a 5-min period. The apparatus was cleaned with 75% ethanol prior to each test session to eliminate odors. Behavioral parameters to assess locomotor activity and exploratory behavior (latency period, number of crossings, and rearing) were recorded.\n\n### 2.3.2. Forced Swimming Test (FST) {#sec2dot3dot2-nutrients-08-00243}\n\nFST is widely employed to screen antidepressant efficacy and depressive-like behavior in rodents. The test was performed as previously described \\[[@B15-nutrients-08-00243]\\]. In brief, each rat was placed in a plastic drum (45 cm height, 25 cm diameter) containing approximately 35 cm of water (24 \u00b1 1 \u00b0C) for a 15-min pretest. After swimming, rats were dried with towels and placed back in their home cage. 24 h later, the rat was exposed to the same experimental conditions outlined above for a 5-min FST. The measured parameters were immobility, swimming and struggling scores. Water was changed before each trial. Immobility was defined as floating passively and only making slight movements to keep the head above water. Each test session was videotaped and the duration of immobility was scored by an experienced observer blind to the experiment design.\n\n2.4. Measurement of Oxidative Stress {#sec2dot4-nutrients-08-00243}\n------------------------------------\n\nThe prefrontal cortex and hippocampus were homogenized using Precellys 24 multifunctional homogenizer (Bertin Technologies, Aix-en-Provence, France). An aliquot of the homogenate was used for the assay of malondialdehyde (MDA) formation and superoxide dismutase (SOD) activity. MDA levels and SOD activity were determined using the Lipid Peroxidation MDA Assay Kit and Total Superoxide Dismutase Assay Kit with WST-1 (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) respectively, following the manufacturer's instructions.\n\n2.5. Western Blot Analysis {#sec2dot5-nutrients-08-00243}\n--------------------------\n\nFor Western blot analysis, total protein was prepared from the prefrontal cortex and hippocampus, and the concentration was determined using Bradford method. Samples were loaded on a precast 12% SDS-PAGE gel with 10 \u03bcg protein in each lane. Proteins in the gels were transferred to a PVDF membrane and blocked for 1 h in 5% non-fat dry milk in TBS-T (25 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween-20). The following antibodies and concentrations were used over night at 4 \u00b0C; iNOS (Proteintech; 1:500), NF-\u03baB (Proteintech; 1:500), and \u03b2-actin (Proteintech; 1:4000). It was then probed with HRP-conjugated secondary antibody for 40 min. The film signal was digitally scanned and then quantified using Image J software (National Institutes of Health, Bethesda, MD, USA). The signals were normalized to \u03b2-actin as an internal standard.\n\n2.6. Real-Time PCR Analysis {#sec2dot6-nutrients-08-00243}\n---------------------------\n\nTotal RNA was extracted from the prefrontal cortex and the hippocampus using Trizol reagent (Invitrogen Corp., Carlsbad, CA, USA) following the manufacturer's instructions for detection of the gene expression of IL-1\u03b2, IL-6, TNF-\u03b1, Bax, Bcl-2, and Bcl-xl. RNA concentration was determined for quantity and integrity using the spectrophotometry (Jingke, Ningbo, China). cDNA was produced using Revert Aid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Tewksbury, MA, USA). Quantitative PCR was performed on Bio-rad Cx96 Detection System (Bio-rad, Hercules, CA, USA) using SYBR green PCR kit (Applied Biosystems Inc., Woburn, MA, USA) and gene-specific primers. A 5 ng cDNA sample was used with 40 cycles of amplication. Each cDNA was tested in triplicate. Relative quantitation for PCR product was normalized to \u03b2-actin as an internal standard. The sequences of gene-specific primers are listed in [Table 1](#nutrients-08-00243-t001){ref-type=\"table\"}.\n\n2.7. Histopathological Examination {#sec2dot7-nutrients-08-00243}\n----------------------------------\n\nFor light microscopy, autopsy samples were taken from the brain of rats in different groups and embedded in paraffin. The paraffin tissue blocks were prepared for sectioning at 5 micron thickness by sledge microtome. The obtained tissue sections were collected on glass slides, deparaffinized, stained by hematoxylin and eosin stain for routine examination, and then examination was done through the light electric microscope.\n\n2.8. Immunohistochemical Study {#sec2dot8-nutrients-08-00243}\n------------------------------\n\nNeurocyte apoptosis was evaluated by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay. The TUNEL method, which detects fragmentation of DNA in the nucleus during apoptotic cell death *in situ*, was employed using an apoptosis detection kit (Keygen Biotech, Nanjing, China). The average ratio of the total TUNEL-positive neurocyte number was calculated from randomly selected 10 microscopic high-power fields for each rat in all of the groups. This ratio represented the apoptotic index of the sample and was compared between groups.\n\n2.9. Statistical Analysis {#sec2dot9-nutrients-08-00243}\n-------------------------\n\nAll statistical procedures were performed on Statistical Package for the Social Science (SPSS) version 18 (SPSS Inc., Chicago, IL, USA). Data were expressed as mean \u00b1 SD, and analyzed statistically by one-way analysis of variance (ANOVA) with Tukey *post hoc* multiple comparisons. The prior level of significance was established at *p* \\< 0.05.\n\n3. Results {#sec3-nutrients-08-00243}\n==========\n\n3.1. Effect of DOX and \u03c9-3 PUFAs on Body Weight Gain and Behavioral Changes {#sec3dot1-nutrients-08-00243}\n---------------------------------------------------------------------------\n\nAs shown in [Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}A, DOX-treated rats showed significantly decreased body weight gain when compared to control animals ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}A, *p* \\< 0.01), whereas \u03c9-3 PUFAs pretreatment had no influence on the body weight gain in both DOX or vehicle treated rats, which is consistent with the results of the previous studies \\[[@B15-nutrients-08-00243],[@B25-nutrients-08-00243],[@B26-nutrients-08-00243]\\]. The numbers of crossings and rearing, as well as latency time, were observed in the OFT. The DOX group exhibited a significant decrease in numbers of crossings ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}B, *p* \\< 0.01) and rearing ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}C, *p* \\< 0.05), as well as a significantly increased latency time ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}D, *p* \\< 0.01) when compared to control group. Pretreatment with \u03c9-3 PUFAs significantly increased the numbers of crossings ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}B, *p* \\< 0.05) and rearing ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}C, *p* \\< 0.01), and markedly decreased latency time ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}D, *p* \\< 0.01) in DOX-treated rats, showing more locomotor activity and exploratory behavior. In the FST, two weeks of DOX administration led to a significant decrease in the swimming ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}E, *p* \\< 0.05) and struggling ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}F, *p* \\< 0.05) times as well as a significant increase in the immobility time ([Figure 1](#nutrients-08-00243-f001){ref-type=\"fig\"}G, *p* \\< 0.01) when compared to the control group. On the other hand, \u03c9-3 PUFAs mitigated the behavioral changes in FST, except that the struggling time in the \u03c9-3 PUFAs pretreated group was slightly, but not significantly, increased as compared with DOX-treated group.\n\n3.2. Effects of DOX and \u03c9-3 PUFAs on Oxidative Stress Markers {#sec3dot2-nutrients-08-00243}\n-------------------------------------------------------------\n\nIn the DOX-treated rats, MDA level was significantly enhanced both in the prefrontal cortex ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}A, *p* \\< 0.05) and hippocampus ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}B, *p* \\< 0.01), while the SOD level was significantly reduced only in the hippocampus ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}D, *p* \\< 0.01) with no remarkable change in the prefrontal cortex ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}C). On the other hand, \u03c9-3 PUFAs supplementation significantly ameliorated the changes of SOD and MDA levels in the DOX-treated group with a significant increase of SOD level ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}D, *p* \\< 0.01) in the hippocampus and a significant decrease of MDA level both in the prefrontal cortex ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}A, *p* \\< 0.05) and hippocampus ([Figure 2](#nutrients-08-00243-f002){ref-type=\"fig\"}B, *p* \\< 0.01), showing the protective effect of \u03c9-3 PUFAs against DOX-induced oxidative stress in brain tissues.\n\n3.3. Effects of DOX and \u03c9-3 PUFAs on Neuroinflammation Biomarkers {#sec3dot3-nutrients-08-00243}\n-----------------------------------------------------------------\n\nThe DOX group showed a significant increase in gene expressions of IL-1\u03b2 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}A, *p* \\< 0.01) and IL-6 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}C, *p* \\< 0.01) in the prefrontal cortex, while in the hippocampus, gene expressions of IL-6 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}D, *p* \\< 0.01) and TNF-\u03b1 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}F, *p* \\< 0.05) were remarkably enhanced as compared with control group. However, these elevated gene expressions were significantly attenuated by \u03c9-3 PUFAs supplementation. The DOX + PUFA group showed significantly decreased gene expressions of IL-1\u03b2 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}A, *p* \\< 0.01) and IL-6 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}C, *p* \\< 0.01) in the prefrontal cortex, and reduced gene expressions of IL-6 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}D, *p* \\< 0.01) and TNF-\u03b1 ([Figure 3](#nutrients-08-00243-f003){ref-type=\"fig\"}F, *p* \\< 0.01) in the hippocampus when compared to DOX group.\n\nIn the prefrontal cortex and the hippocampus of DOX-treated rats, the protein expression of NF-\u03baB ([Figure 4](#nutrients-08-00243-f004){ref-type=\"fig\"}A,B; *p* \\< 0.01, both) and iNOS ([Figure 4](#nutrients-08-00243-f004){ref-type=\"fig\"}C,D; *p* \\< 0.01, both) was markedly increased as compared to the vehicle-treated rats. In correspondence to the modulating effects of \u03c9-3 PUFAs on the inflammatory cytokines, \u03c9-3 PUFAs downregulated protein level of NF-\u03baB ([Figure 4](#nutrients-08-00243-f004){ref-type=\"fig\"}A,B; *p* \\< 0.01, both) and iNOS ([Figure 4](#nutrients-08-00243-f004){ref-type=\"fig\"}D; *p* \\< 0.01) when compared to the animals treated with DOX alone.\n\n3.4. Effects of DOX and \u03c9-3 PUFAs on Histopathological Changes {#sec3dot4-nutrients-08-00243}\n--------------------------------------------------------------\n\nHistopathological alterations in brain specimens from different treated groups are shown in [Figure 5](#nutrients-08-00243-f005){ref-type=\"fig\"}. Sections from the control group showed normal histology ([Figure 5](#nutrients-08-00243-f005){ref-type=\"fig\"}A,E). DOX-administered brain tissues showed more frequent nuclear pyknosis ([Figure 5](#nutrients-08-00243-f005){ref-type=\"fig\"}C,G), whereas rats pretreated with \u03c9-3 PUFAs showed a reduced number of nuclear pyknosis and almost usual architecture similar to those of the normal brain tissues ([Figure 5](#nutrients-08-00243-f005){ref-type=\"fig\"}D,H).\n\n3.5. Effects of DOX and \u03c9-3 PUFAs on Neural Apoptotic Markers {#sec3dot5-nutrients-08-00243}\n-------------------------------------------------------------\n\nAs revealed in [Figure 6](#nutrients-08-00243-f006){ref-type=\"fig\"}, after two weeks of treatment, brain tissues exposed to DOX contained much more TUNEL-positive cells ([Figure 6](#nutrients-08-00243-f006){ref-type=\"fig\"}C,G) in contrast to those pretreated with \u03c9-3 PUFAs ([Figure 6](#nutrients-08-00243-f006){ref-type=\"fig\"}D,H), indicating the pro-apoptotic effects of DOX and anti-apoptotic effects of \u03c9-3 PUFAs in brain tissues. [Figure 6](#nutrients-08-00243-f006){ref-type=\"fig\"}I,J show the apoptotic index in the prefrontal cortex and hippocampus of rats in different groups. The percentage of apoptotic neurocytes was significantly increased by DOX (*p* \\< 0.01), but was partly restored by \u03c9-3 PUFAs pretreatment (*p* \\< 0.01).\n\nConsistent with the results of the TUNEL staining, a significant increase in gene expression of pro-apoptotic Bax was observed both in the prefrontal cortex ([Figure 7](#nutrients-08-00243-f007){ref-type=\"fig\"}A, *p* \\< 0.01) and the hippocampus ([Figure 7](#nutrients-08-00243-f007){ref-type=\"fig\"}B, *p* \\< 0.01) of DOX-treated rats, but this enhancement was largely normalized by \u03c9-3 PUFAs pretreatment ([Figure 7](#nutrients-08-00243-f007){ref-type=\"fig\"}A,B; *p* \\< 0.01, *p* \\< 0.05, respectively). However, in our study, gene expression of anti-apoptotic factors (Bcl-2, Bcl-xl) were not statistically changed, except for the increased gene expression of Bcl-2 ([Figure 7](#nutrients-08-00243-f007){ref-type=\"fig\"}D, *p* \\< 0.05) in the hippocampus of rats pretreated with \u03c9-3 PUFAs, and an increase in gene expression of Bcl-xl ([Figure 7](#nutrients-08-00243-f007){ref-type=\"fig\"}E, *p* \\< 0.05) in the prefrontal cortex of DOX-treated rats.\n\n4. Discussion {#sec4-nutrients-08-00243}\n=============\n\nThe present study firstly demonstrated the protective role of \u03c9-3 PUFAs against DOX-induced neurotoxicity in rats. We observed that DOX administration induced depressive-like behaviors in rats and pretreatment with \u03c9-3 PUFAs normalized behavioral changes in rats treated with DOX. We also found that DOX caused oxidative stress, neuroinflammation, and cell death in the brain tissues and \u03c9-3 PUFAs could partly alleviate these changes, suggesting the potentially protective role of \u03c9-3 PUFAs in the brain from this pathophysiology. These findings are significant since patients following DOX chemotherapy are prone to develop depression and the findings in the present study raise the possibility that \u03c9-3 PUFAs might be an adjuvant therapy and help to prevent this neurotoxic side effect of DOX in clinical practice.\n\nInduction of depression-like behavior in rats using DOX has been mentioned in previous studies \\[[@B3-nutrients-08-00243],[@B27-nutrients-08-00243]\\]. Our results indicated that DOX markedly promoted the depressive-like behaviors in rats, confirming the neurotoxical effect of DOX. In our preliminary work, a dose of 1.5 g/kg \u03c9-3 PUFAs was effective to attenuate depressive behaviors in rats exposed to chronic unpredictable mild stress (CUMS) \\[[@B15-nutrients-08-00243]\\]. Similarly, in the present study, \u03c9-3 PUFAs supplementation effectively restored these behavioral changes induced by DOX, showing robust antidepressant-like effects. However, it should be noted that although both EPA and DHA contain robust antidepressant properties \\[[@B28-nutrients-08-00243],[@B29-nutrients-08-00243],[@B30-nutrients-08-00243]\\], previous studies showed that EPA is more effective in mitigating the behavioral changes \\[[@B30-nutrients-08-00243],[@B31-nutrients-08-00243]\\]. In this context, future studies are warranted to investigate which nutritional ingredient plays a major role in the neuroprotective effects of \u03c9-3 PUFAs against DOX-induced behavioral changes.\n\nAccording to the previous findings \\[[@B11-nutrients-08-00243],[@B13-nutrients-08-00243],[@B15-nutrients-08-00243],[@B32-nutrients-08-00243],[@B33-nutrients-08-00243]\\], the mechanisms underlying the behavioral changes following DOX treatment and the antidepressant-like and neuroprotective effects of \u03c9-3 PUFAs might be related to the oxidative stress, inflammatory, and apoptotic status of the brain tissues. Thus, we further assessed the various markers of oxidative stress, inflammation, and apoptosis in different groups.\n\nThe oxidative stress as a plausible pathomechanism of neuropsychological alterations is strongly supported by previous findings \\[[@B2-nutrients-08-00243],[@B27-nutrients-08-00243]\\], showing that DOX increases oxidative stress and reduces the total antioxidant capacity. In accordance with previous results, we observed that lipid peroxidation, a downstream chain reaction initiated by free radicals, was activated by DOX and the endogenous antioxidant enzyme SOD, responsible for scavenging superoxide radicals, was markedly suppressed by DOX in both prefrontal cortex and hippocampus, confirming the pro-oxidative effect of DOX on the brain tissues \\[[@B34-nutrients-08-00243]\\]. Furthermore, in this work, we clearly demonstrated the capability of \u03c9-3 PUFAs in attenuating brain lipid peroxidation and protecting antioxidant enzyme activity in rats exposed to DOX, illustrating the effective anti-oxidative actions of \u03c9-3 PUFAs against a potent free radical-producing chemotherapeutic agent. This is consistent with previous results when \u03c9-3 PUFAs attenuated the oxidative damage to the heart or the testis of rats treated with DOX \\[[@B17-nutrients-08-00243],[@B18-nutrients-08-00243]\\].\n\nMoreover, previous studies have shown that ROS activates pro-inflammatory mediators, such as TNF-\u03b1 and NF-\u03baB, and subsequently induces brain neuroinflammation \\[[@B35-nutrients-08-00243],[@B36-nutrients-08-00243]\\]. Pro-inflammatory cytokines such as IL-1\u03b2, IL-6, TNF-\u03b1, NF-\u03baB, and iNOS have been demonstrated to induce abnormal behaviors, such as decreased locomotor activity, exploration, and depression \\[[@B1-nutrients-08-00243],[@B3-nutrients-08-00243],[@B37-nutrients-08-00243]\\], which was confirmed in our experiment. We found that DOX provoked generation of TNF-\u03b1 and subsequently caused the activation of NF-\u03baB and iNOS and increased the expression of genes required to control infection and injury, such as IL-1\u03b2 and IL-6, indicating severe inflammatory conditions in the brain. Nevertheless, the current study elaborated the effective inhibition of generation of pro-inflammatory mediators by \u03c9-3 PUFAs in brain tissues, including IL-1\u03b2, IL-6, TNF-\u03b1, NF-\u03baB, and iNOS. This anti-inflammatory effect of \u03c9-3 PUFAs has been reported elsewhere, when \u03c9-3 PUFAs antagonize the NF-\u03baB signaling pathway, and inhibit the expression of inflammatory genes downstream of NF-\u03baB \\[[@B12-nutrients-08-00243],[@B38-nutrients-08-00243],[@B39-nutrients-08-00243]\\].\n\nAdditionally, we found that DOX caused a significant increase in TUNEL-positive neurocytes, suggesting severe DNA damage and neuronal death. Previous study conducted by Kreisel and his co-workers (2014) showed that CUMS-induced depression-like behavior in rodents exerted apoptotic cell death in hippocampus, indicating the close relationship between neural apoptosis and depression \\[[@B40-nutrients-08-00243]\\]. Therefore, apoptosis might play a causative role in the development of DOX-induced depression. As previous findings in rat cortical neurons \\[[@B41-nutrients-08-00243]\\], we also observed that the pro-apoptotic Bax was responsive to DOX stimulation in neurocytes, indicating the involvement of Bcl-2 family proteins in the pro-apoptotic effect of DOX. On the other hand, the present study elucidated the effective suppression of apoptosis and rescue of neurocytes by \u03c9-3 PUFAs in brain tissues exposed to DOX. This finding implied that \u03c9-3 PUFAs can block apoptosis, which is in accordance with previous research \\[[@B12-nutrients-08-00243],[@B42-nutrients-08-00243]\\]. Although the mechanism responsible for the anti-apoptotic effect of \u03c9-3 PUFAs was not yet defined, it might be partially associated with regulation of the expression of the Bcl-2 family of proteins, as gene expression of Bax in brain tissues was significantly inhibited by \u03c9-3 PUFAs in our study. In line with our results, Paterniti *et al*. (2014) had demonstrated that the treatment with DHA reduced Bax expression in the brain tissues of mice \\[[@B12-nutrients-08-00243]\\]. Moreover, previous studies suggested that pro-inflammatory cytokines appear to contribute to depression-associated cell death through intrinsic apoptotic pathways and that neurotoxic free radicals are a second apoptosis-mediating factor associated with depressive disorder \\[[@B12-nutrients-08-00243],[@B20-nutrients-08-00243],[@B43-nutrients-08-00243]\\], suggesting that anti-oxidant and anti-inflammatory effects of \u03c9-3 PUFAs could, in turn, indirectly contribute to its anti-apoptotic effect.\n\nMore importantly, there are a large number of studies reporting the anticancer and anti-cachectic effects of \u03c9-3 PUFAs in a variety of model systems \\[[@B44-nutrients-08-00243],[@B45-nutrients-08-00243],[@B46-nutrients-08-00243]\\]. Eltweri *et al*. (2016) \\[[@B45-nutrients-08-00243]\\] provided evidence of the effectiveness of \u03c9-3 PUFAs in cancer management with favorable outcomes, including better quality of life, less toxicity, and even improved survival, confirming the anti-cancer activity of \u03c9-3 PUFAs. Additionally, Kim *et al*. (2016) \\[[@B46-nutrients-08-00243]\\] showed that the combination of regorafenib and DHA results in a synergistic effect upon tumor invasiveness and a reduction in tumor weights. These results indicate that adjunctive therapy with \u03c9-3 PUFAs is more likely to enhance anti-tumor properties of DOX and reduce its neurotoxicity.\n\n5. Conclusions {#sec5-nutrients-08-00243}\n==============\n\nCollectively, we demonstrated that \u03c9-3 PUFAs supplementation alleviated DOX-induced depressive-like behaviors and neurotoxicity in rats. The possible mechanisms underlying these behavioral-modulating and neuroprotective effects are proved to be at least partially associated with the anti-oxidant, anti-inflammatory, and anti-apoptotic actions of \u03c9-3 PUFAs in brain tissues. Furthermore, this study provided us with a new potential treatment for brain damage induced by chemotherapeutic drugs, and had paved the way for further studies to investigate other mechanisms underlying the behavior modulating and neuroprotective effects of \u03c9-3 PUFAs.\n\nThis study was supported by National Natural Science Foundation of China (NSFC: 81401113).\n\nHuan-De Li, Pei Jiang and Rui-Li Dang designed the study and wrote the protocol. Yan-Qin Wu, Rui-Li Dang, Mi-Mi Tang and Hua-Lin Cai performed the experiments and analyzed the data. De-Hua Liao, Xin He, Ling-Juan Cao and Ying Xue contributed to the reagents and materials. Yan-Qin Wu drafted the manuscript. Pei Jiang, Rui-Li Dang and Hua-Lin Cai revised the manuscript content. All authors read and approved the final manuscript.\n\nThe authors declare no conflict of interest.\n\n![Body weight gain and behavioral test. Effect of DOX and \u03c9-3 PUFAs on body weight gain (**A**) and open field test: numbers of crossing (**B**); number of rearing (**C**); and latency time (**D**); effect of DOX and \u03c9-3 PUFAs on forced swimming test: swimming time (**E**); struggling time (**F**); and immobility time (**G**). Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to DOX-injected group.](nutrients-08-00243-g001){#nutrients-08-00243-f001}\n\n![Effects of DOX and \u03c9-3 PUFAs on MDA and SOD in the prefrontal cortex and hippocampus. Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to the DOX-injected group.](nutrients-08-00243-g002){#nutrients-08-00243-f002}\n\n![Effects of DOX and \u03c9-3 PUFAs on gene expression of IL-1, IL-6, and TNF-\u03b1 in the prefrontal cortex and hippocampus. Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to the DOX-injected group.](nutrients-08-00243-g003){#nutrients-08-00243-f003}\n\n![Effects of DOX and \u03c9-3 PUFAs on protein expression of NF-\u03baB and iNOS in the prefrontal cortex and hippocampus. Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to the DOX-injected group.](nutrients-08-00243-g004){#nutrients-08-00243-f004}\n\n![Effects of DOX and \u03c9-3 PUFAs on histological changes in the prefrontal cortex and hippocampus (magnification 400\u00d7). Note the presence of degenerating neurons (arrows) in the DOX-treated group.](nutrients-08-00243-g005){#nutrients-08-00243-f005}\n\n![Effects of DOX and \u03c9-3 PUFAs on TUNEL-positive cells in the prefrontal cortex and hippocampus. (**A**)--(**H**) show the TUNEL-staining positive cells through the light electric microscope followed by observation at 400\u00d7; (**I**) and (**J**) show the quantitative analysis of TUNEL results in the prefrontal cortex and hippocampus. Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to the DOX-injected group.](nutrients-08-00243-g006){#nutrients-08-00243-f006}\n\n![Effects of DOX and \u03c9-3 PUFAs on gene expression of Bax, Bcl-2, and Bcl-xl in the prefrontal cortex and hippocampus. Data are expressed as means \u00b1 SD (*n* = 6--7). \\* *p* \\< 0.05, \\*\\* *p* \\< 0.01 compared to the control group. ^\\#^ *p* \\< 0.05, ^\\#\\#^ *p* \\< 0.01 compared to the DOX-injected group.](nutrients-08-00243-g007){#nutrients-08-00243-f007}\n\nnutrients-08-00243-t001_Table 1\n\n###### \n\nPrimers used in real-time PCR analyses of mRNA expression.\n\n  Target Gene                                    Primers Sequences               Size (bp)\n  ------------- -------------------------------- ------------------------------- -----------\n  IL-1\u03b2         Forward                          5\u2032-AGGTCGTCATCATCCCACGAG-3\u2032     119\n  Reverse       5\u2032-GCTGTGGCAGCTACCTATGTCTTG-3\u2032                                   \n  IL-6          Forward                          5\u2032-CACAAGTCCGGAGAGGAGAC-3\u2032      167\n  Reverse       5\u2032-ACAGTGCATCATCGCTGTTC-3\u2032                                       \n  TNF-\u03b1         Forward                          5\u2032-GAGAGATTGGCTGCTGGAAC-3\u2032      82\n  Reverse       5\u2032-TGGAGACCATGATGACCGTA-3\u2032                                       \n  Bax           Forward                          5\u2032-CCAGGACGCATCCACCAAGAAGC-3\u2032   135\n  Reverse       5\u2032-TGCCACACGGAAGAAGACCTCTCG-3\u2032                                   \n  Bcl-xl        Forward                          5\u2032-CAGCTTCATATAACCCCAGGGAC-3\u2032   207\n  Reverse       5\u2032-GCTCTAGGTGGTCATTCAGGTAGG-3\u2032                                   \n  Bcl-2         Forward                          5\u2032-AGCCCTGTGCCACCTGTGGT-3\u2032      93\n  Reverse       5\u2032-ACTGGACATCTCTGCAAAGTCGCG-3\u2032                                   \n  \u03b2-Actin       Forward                          5\u2032-CATCCTGCGTCTGGACCTGG-3\u2032      116\n  Reverse       5\u2032-TAATGTCACGCACGATTTCC-3\u2032                                       \n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "All relevant data are within the manuscript and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nLow-back pain is a highly prevalent disease, and a major problem for performing activities of daily living and health economics. Many studies have reported that cigarette smoking is a risk factor for low-back pain; i.e., 70% of persons with low-back pain are cigarette smokers, and the frequency, amount and duration of cigarette smoking correlate with the incidence of low-back pain \\[[@pone.0218298.ref001]--[@pone.0218298.ref009]\\]. In contrast, smoking cessation has been reported to improve patient-reported pain and is related to increased fusion rates \\[[@pone.0218298.ref010], [@pone.0218298.ref011]\\]. Moreover, cigarette smoking is also associated with degenerative disc disease, including middle-aged disc herniation \\[[@pone.0218298.ref012]--[@pone.0218298.ref015]\\]. Cigarette smoking leads to the formation of carboxy-hemoglobin \\[[@pone.0218298.ref016]\\], vasoconstriction \\[[@pone.0218298.ref016]\\] and arteriosclerosis \\[[@pone.0218298.ref016], [@pone.0218298.ref017]\\], and thus decreases oxygen transport and blood flow \\[[@pone.0218298.ref016], [@pone.0218298.ref017]\\]. These events are considered to lead to malnutrition of the intervertebral discs (IVD) and promote IVD degeneration. Studies using animal models and *in vitro* culturing of disc cells suggested that nicotine and tobacco smoke exposure induces degenerative changes in the spine \\[[@pone.0218298.ref018]--[@pone.0218298.ref022]\\]. Although there have been significant advances in our understanding of the biology underlying IVD degeneration \\[[@pone.0218298.ref023]\\] \\[[@pone.0218298.ref024]\\], the molecular mechanisms underlying the IVD changes induced by cigarette smoking remain to be elucidated.\n\nTo directly clarify the influence of passive cigarette smoking on IVD, we established a rat model of passive cigarette smoking. In passive cigarette smoking rats, slight structural changes were noted on haematoxylin-eosin and alcian blue + periodic acid-Schiff staining of the IVD tissue \\[[@pone.0218298.ref025]\\], and the expression of type I and IX collagen mRNA was reduced \\[[@pone.0218298.ref026]\\]. Comprehensive investigation with gene expression microarrays revealed increased expression of heat shock protein 70 and protein tyrosine phosphatase \\[[@pone.0218298.ref027]\\]. The expression of these genes was also increased in the isolated nucleus pulposus (NP) and annulus fibrosus (AF) \\[[@pone.0218298.ref028]\\], suggesting that the passive cigarette smoking-induced stress response occurs similarly in the NP and AF, and induces anti-apoptotic responses. Recently, we observed that passive cigarette smoking also changes the circadian rhythm of clock genes in rat IVD \\[[@pone.0218298.ref029]\\]. To further investigate the relationship between the morphological and molecular changes in passive cigarette smoking-induced IVD degeneration, we examined the histological changes and molecular events in the extracellular matrix (ECM) and chondrocytes of the IVD. Based on these findings, we proposed a model of passive cigarette smoking-induced IVD degeneration in rats.\n\nMaterials and methods {#sec002}\n=====================\n\nAnimals {#sec003}\n-------\n\nTwelve male, 8-week-old Sprague-Dawley rats (CLEA Japan, Inc., Tokyo, Japan) were subjected to passive cigarette smoking using our cigarette smoking device, which has been described previously \\[[@pone.0218298.ref026], [@pone.0218298.ref029]\\]. During housing, the rat health status was monitored twice daily. The monitoring items were food and water intake difficulty, anxiety symptoms such as self-mutilation, respiratory disorder and vocalization, and weight loss without rapid recovery. No rats showed these symptoms during the experiment. Rats underwent passive cigarette smoke exposure for 4 or 8 weeks, designated as the S4 and S8 groups, respectively (n = 6/group). As the respective control groups, non-smoking control rats were established as the N4 and N8 groups, respectively (n = 6/group).\n\nEuthanasia was performed under deep anaesthesia with intraperitoneal administration of 30 mg of pentobarbital sodium and efforts were made to minimize the animal's discomfort. After euthanasia, a longitudinal incision was made immediately above the dorsal spine and the entire spine was excised. The lumbar IVD excluding CEP were separated from the vertebrae and immediately frozen at -80\u00b0C. These samples were used for mRNA measurement and Western blotting. The lower thoracic vertebrae were also excised *en bloc*, fixed in 10% formalin, decalcified with EDTA for 2 months and embedded in paraffin. The study protocol was approved by the Animal Experimentation Committee of the Nihon University School of Medicine.\n\nQuantification of mRNA {#sec004}\n----------------------\n\nTotal RNA extraction, cDNA synthesis, and quantitative PCR were performed as described previously \\[[@pone.0218298.ref030]\\]. Briefly, total RNA was extracted from two IVDs of each rat. PCR was performed in 20 \u03bcl of reaction mixture containing TaqMan Universal PCR master mix (Applied Biosystems, Foster, USA), TaqMan Gene Expression Assay probe and primers ([S1 Table](#pone.0218298.s006){ref-type=\"supplementary-material\"}) and 10 ng of cDNA template by Rotor-Gene 3000 (Corbett Life Science, Mortlake, Australia) at 95\u00b0C for 10 min, followed by 45 cycles of 95\u00b0C for 15 s, 60\u00b0C for 1 min.\n\nImmunohistochemistry and specific staining {#sec005}\n------------------------------------------\n\nThin paraffin-embedded sections (4 \u03bcm) were cut and treated with 600 U/ml of hyaluronidase (Sigma-Aldrich, St. Louis, USA) at 37\u00b0C for 1 hr to activate antigens. Sections were stained using the CSA II kit (Biotin-free catalysed signal amplification system; DAKO, Glostrup, Denmark) and 3,3'-diaminobenzidine (DAB). Type II collagen was detected using a mouse monoclonal antibody (10 \u03bcg/ml; Daiichi Fine Chemical, Toyama, Japan) and aggrecan was detected using a mouse monoclonal antibody (10 \u03bcg/ml; Thermo Fisher Scientific, Massachusetts, USA) at 37\u00b0C for 1 hr, followed by incubation with horse radish peroxidase (HRP)-labelled goat anti-mouse IgG (DAKO) at room temperature for 15 min. Haematoxylin was used for counterstaining. To quantify staining, the stained tissues were imaged under a microscope (OLYMPUS BX51, Tokyo, Japan) and the positive areas were measured using Win ROOF version 5.6 (Mitani, Co., Fukui, Japan). To evaluate the staining in each region, the tissue images were divided into the NP, the AF, and the peripheral and central regions of the cartilage end-plate (CEP). The percent areas that stained positive were calculated for each region ([S1 Fig](#pone.0218298.s001){ref-type=\"supplementary-material\"}). Four IVD were observed for each of the 6 animals per group.\n\nThe above-described thin sections were subjected to Elastica van Gieson (EVG), safranin O and alcian blue staining, and positivity was evaluated. On safranin O staining, the red-stained area, representing acidic proteoglycan (PG) in the CEP, was measured using Win ROOF Version 5.6, and the percent positive area was calculated for the peripheral and central regions of the CEP ([S1 Fig](#pone.0218298.s001){ref-type=\"supplementary-material\"}).\n\nDNA fragmentation {#sec006}\n-----------------\n\nDNA fragmentation was detected by immunohistochemistry using an antibody against single-strand DNA (ssDNA). After blocking with 5% skim milk at 37\u00b0C for 1 hr, the sections were incubated with anti-ssDNA rabbit IgG (Immuno-Biological Laboratories Co., Gunma, Japan) at 37\u00b0C for 1 hr, followed by reaction with HRP-labelled anti-rabbit IgG antibody (Immuno-Biological Laboratories Co.) at 25\u00b0C for 30 min and colour development using DAB. Haematoxylin was used for counterstaining. Four IVD were observed for each of the 6 animals per group. The numbers of ssDNA-positive and -negative cells in the CEP were measured, and the positive rate was calculated by dividing the number of positive cells by the total number of cells. The specificity of the staining was confirmed using the thymus tissues from rats treated with and without dexamethasone.\n\nStatistical analysis {#sec007}\n--------------------\n\nThe Mann-Whitney U test was used to determine the significance of differences between two groups (group N4 vs. group S4, group N8 vs. group S8). Differences were considered significant when the *p*-values were less than 0.05.\n\nResults {#sec008}\n=======\n\nHistological changes of the IVD induced by passive cigarette smoking {#sec009}\n--------------------------------------------------------------------\n\nEVG staining was employed to investigate changes in the ECM of the IVD, and specifically demonstrated fibrous structural changes in the NP that were induced by passive cigarette smoking ([Fig 1A](#pone.0218298.g001){ref-type=\"fig\"}). As expected, the AF and CEP were covered with collagen fibres, which stained red, and were not affected by passive cigarette smoking. In contrast, no red-staining of the collagen fibres was noted in the NP, and dark purple stained elastic fibres were present in the surrounding region, exhibiting a closed reticular structure. In the central region, NP cells (notochordal cells) containing pink-stained cytoplasm were present in clusters, and cytoplasm that appeared to outline vacuoles was evident in some cells. Passive cigarette smoking markedly altered these characteristic structures of the NP. In particular, marked destruction of the reticular structure and condensation of NP cells were observed after 8-weeks of passive cigarette smoking ([Fig 1A](#pone.0218298.g001){ref-type=\"fig\"}). To identify the components of this reticular structure, safranin O and alcian blue staining, and immunostaining for type II collagen and aggrecan were performed for the IVD from the non-cigarette smoking rats (group N4) ([Fig 1B](#pone.0218298.g001){ref-type=\"fig\"}). The reticular structures that stained dark purple on EVG staining were positive for all stains (i.e., safranin O, alcian blue, type II collagen and aggrecan). Sulphated polysaccharide-containing PG and type II collagen fibres comprised the interstitium-supportive structure in the NP ([Fig 1B](#pone.0218298.g001){ref-type=\"fig\"}), which was destroyed by passive cigarette smoking ([Fig 1A](#pone.0218298.g001){ref-type=\"fig\"}). We therefore investigated how the individual components of the NP interstitium were affected by passive cigarette smoking.\n\n![Histological changes of the rat IVD induced by passive cigarette smoking.\\\n(A) EVG staining of IVD from control non-smoking (N4) and smoking (S4) rats for 4 weeks or 8 weeks (N8 and S8, respectively). Upper and lower panels represent low and high magnification, respectively. Bars in the left and right panels indicate 500 \u03bcm and 100 \u03bcm, respectively. NP, nucleus pulposus; AF, annulus fibrosus; CEP, cartilage end-plate. (B) Staining of NP of IVD from control non-smoking rat N4. a, EVG staining; b, safranin O staining; c, alcian blue staining; d, immunohistochemical staining for Type II collagen; e, immunohistochemical staining for aggrecan. Bar indicates 100 \u03bcm.](pone.0218298.g001){#pone.0218298.g001}\n\nIntense staining for type II collagen was observed in the reticular structure of the NP, which was consistent with the EVG staining (Figs [1B-a, 1B-d](#pone.0218298.g001){ref-type=\"fig\"} and [2A](#pone.0218298.g002){ref-type=\"fig\"}). Type II collagen-positive areas were subsequently evaluated in the NP, and compared between the S8 and N8 groups. Although the percent positive area decreased significantly in the S8 group, no significant differences were noted in the AF ([Fig 2B](#pone.0218298.g002){ref-type=\"fig\"}). Similarly, when the positive areas were measured in the CEP, significant decreases were observed in the central region in the S8 group ([Fig 2B](#pone.0218298.g002){ref-type=\"fig\"}). Therefore, type II collagen expression was reduced in the NP and CEP. Synthesis ability was investigated by measuring mRNA expression, but no significant decreases were noted ([S2A Fig](#pone.0218298.s002){ref-type=\"supplementary-material\"} and [S1 Table](#pone.0218298.s006){ref-type=\"supplementary-material\"}). Moreover, there were no changes in the expression of the type II collagen-degrading enzyme, Mmp13 ([S2A Fig](#pone.0218298.s002){ref-type=\"supplementary-material\"} and [S1 Table](#pone.0218298.s006){ref-type=\"supplementary-material\"}). Intense staining for aggrecan was also observed in the interstitium-supportive structure in the NP, which decreased significantly in the S8 group ([Fig 3](#pone.0218298.g003){ref-type=\"fig\"}). Synthesis and degradation of aggrecan were also investigated at the mRNA level. No changes were noted in the expression of aggrecan or Mmp3 mRNA, but Adamts4 mRNA was reduced in both the 4- and 8-week passive cigarette smoking groups ([S2B Fig](#pone.0218298.s002){ref-type=\"supplementary-material\"}).\n\n![Immunohistochemical staining for type II collagen.\\\n(A) Immunostaining for Type II collagen in the IVD from control non-smoking (N8) and smoking (S8) rats for 8 weeks. The IVD from N8 were stained without the primary antibody as a negative control. Upper and lower panels are low and high magnification, respectively. Bars indicate 500 \u03bcm and 100 \u03bcm, respectively. (B) The type II collagen-positive area was measured and the positive rate was calculated. NP, nucleus pulposus; AF, annulus fibrosus; CEP (central), central region of cartilage end-plate. The area of NP, AF and CEP (central) is specified as shown in [S1A Fig](#pone.0218298.s001){ref-type=\"supplementary-material\"}. *P*-values were determined by the Mann-Whitney U test.](pone.0218298.g002){#pone.0218298.g002}\n\n![Immunohistochemical staining for aggrecan.\\\n(A) Immunostaining for aggrecan in the IVD from control non-smoking (N8) and smoking (S8) rats for 8 weeks. Upper and lower panels represent low and high magnification, respectively. Bars indicate 500 \u03bcm and 100 \u03bcm, respectively. (B) The aggrecan-positive area of the NP (nucleus pulposus) was measured and the positive rate was calculated. The area of NP is specified as shown in [S1B Fig](#pone.0218298.s001){ref-type=\"supplementary-material\"}. *P*-values were determined by the Mann-Whitney U test.](pone.0218298.g003){#pone.0218298.g003}\n\nSafranin O-stained IVD are shown in [Fig 4](#pone.0218298.g004){ref-type=\"fig\"}. Passive cigarette smoking markedly affected the ECM staining of the CEP. The red staining of the acidic PG bound by safranin O was markedly reduced by passive cigarette smoking, whereas the green staining of non-collagen protein bound by fast green became more apparent. The quantitative results are shown in [Fig 4B](#pone.0218298.g004){ref-type=\"fig\"}. The percent PG positive area decreased significantly in all regions of the CEP in both the 4- and 8-week passive smoking groups compared with non-smoking controls. This led us to question the mechanism by which the PG decreased in the CEP in the early stage. Focusing on functional inactivation of chondrocytes in the CEP, we subsequently investigated apoptosis in these cells.\n\n![Safranin O staining of rat IVD and CEP.\\\n(A) Representative histological features from control non-smoking (N4) and smoking (S4) rats for 4 weeks or 8 weeks (N8 and S8, respectively). Upper and lower panels represent low and high magnification, respectively. Bars indicate 500 \u03bcm and 100 \u03bcm, respectively. NP, nucleus pulposus; CEP, cartilage end-plate. (B) The safranin O-positive area of the CEP was measured and the positive rate was calculated. The CEP was divided into peripheral and central regions, and the safranin O-positive area was measured. CEP (peripheral, central), peripheral or central region of the CEP specified as shown in [S1C Fig](#pone.0218298.s001){ref-type=\"supplementary-material\"}. *P*-values were determined by the Mann-Whitney U test.](pone.0218298.g004){#pone.0218298.g004}\n\nInduction of apoptosis in the CEP by passive cigarette smoking {#sec010}\n--------------------------------------------------------------\n\nApoptosis was investigated using immunohistochemical staining for ssDNA. ssDNA-positive cells were present in the IVD tissues of healthy rats ([S3 Fig](#pone.0218298.s003){ref-type=\"supplementary-material\"}). Specifically, 60% and 25% of chondrocytes were positive in the CEP in the S8 and N8 groups, respectively, demonstrating a significant increase in the S8 group. Similarly, the number of positive cells increased significantly in the S4 group compared with the N4 group ([Fig 5](#pone.0218298.g005){ref-type=\"fig\"}). In contrast, no passive cigarette smoking-induced changes were noted in the NP or AF cells, although positive cells were present ([S3 Fig](#pone.0218298.s003){ref-type=\"supplementary-material\"}). The lack of increase in apoptotic reactions was also confirmed by Western blotting for \u03b2-actin in the NP and AF ([S4 Fig](#pone.0218298.s004){ref-type=\"supplementary-material\"}). Fragmentation of \u03b2-actin by caspase 3 was evident even in healthy IVD tissue (NP and AF); however, no significant increases in response to passive cigarette smoking were found by quantitation of the fragmentation ([S4 Fig](#pone.0218298.s004){ref-type=\"supplementary-material\"}). Furthermore, no significant changes in apoptosis-related gene expression were observed in the NP or AF in response to passive cigarette smoking ([S1 Table](#pone.0218298.s006){ref-type=\"supplementary-material\"}). Apoptosis was observed to some extent even in healthy IVD cells, which was due to the disappearance and replacement of notochordal cells by chondrocyte-like cells with aging \\[[@pone.0218298.ref031]\\].\n\n![Apoptotic reaction of the CEP induced by passive cigarette smoking.\\\n(A) Immunostaining for ssDNA of CEP from control non-smoking (N4) and smoking (S4) rats for 4 weeks or 8 weeks (N8 and S8, respectively). Arrows indicate representative cells with ssDNA-positive brown nuclei. The bar indicates 100 \u03bcm. (B) The numbers of ssDNA-positive and--negative cells in the CEP were measured and the positive rate was calculated. *P*-values were determined by the Mann-Whitney U test.](pone.0218298.g005){#pone.0218298.g005}\n\nDiscussion {#sec011}\n==========\n\nUsing this model, we previously observed that rat body weight gain was suppressed by passive cigarette smoking ([S5 Fig](#pone.0218298.s005){ref-type=\"supplementary-material\"}). A similar effect was observed in rats allowed to self-administer nicotine, independent of food intake, which corresponded to 15 to 60 micrograms/kg/infusion nicotine \\[[@pone.0218298.ref032]\\]. We also measured blood nicotine levels in our previous study: 36.5 to 124.8 ng/ml (mean: 72.1 ng/ml) \\[[@pone.0218298.ref025]\\]. This concentration range corresponds to 10 to 70 ng/ml (mean: 33 ng/ml) reported for 330 human smokers who smoked 20.7 cigarettes/day on average \\[[@pone.0218298.ref025], [@pone.0218298.ref033]\\]. Therefore, this model is comparable to humans who are smoking 20 to 40 cigarettes/day. Under these exposure conditions, the NP architecture was destroyed by passive cigarette smoking and the supporting structure composed of the ECM had degenerated based on EVG staining. This supportive structure comprised type II collagen and PG, which were both destroyed by passive cigarette smoking. Type II collagen and aggrecan were decreased at the protein level, but this was not supported by quantitative mRNA analysis. Although there were no increases of *Mmp13* or *Mmp3* mRNA in this study, *Adamts4* expression was slightly attenuated in the cigarette smoking groups. Activation of other degrading enzymes may have been involved in the decrease of these matrix proteins. Recently, Ngo *et al*. demonstrated that ADAMTS5 is the primary aggrecanase mediating smoking-induced IVD degeneration in mouse models of chronic tobacco smoking using *ADAMTS5*-deficient mice \\[[@pone.0218298.ref034]\\]. Thus, ADAMTS5 may also be the enzyme responsible for the degradation in the current study. Wang *et al*. reported marked loss of disc matrix in a mouse cigarette smoking model using direct smoke inhalation \\[[@pone.0218298.ref021]\\]. As their model utilized direct vs. passive inhalation, the smoke conditions were more severe than those used in our study, and the degradation of aggrecan, and reduced synthesis of PG and collagen were also demonstrated. Thus, these findings support the conclusion that tobacco smoke alone is sufficient to affect peripheral tissues and lead to IVD degeneration.\n\nPG and type II collagen were also decreased in the CEP following passive cigarette smoking. These structural constituents are produced and maintained by chondrocytes, suggesting that passive cigarette smoking inhibited cellular function in the CEP. We demonstrated that apoptotic responses of CEP cells were stimulated by passive cigarette smoking. In the CEP, the stimulation of apoptosis and reduction of the ECM may have both occurred in the early stage after 4 weeks of passive cigarette smoking. In contrast, in the NP, the reduction of the ECM was notable by the eighth week. This time lag suggests that early functional changes in the CEP are involved in the changes in the NP ECM. Arana et al. observed that when NP cells were co-cultured with cartilage tissue, expression of PG and type I and II collagen increased in the NP cells, and expression of ECM-degrading enzymes was decreased, suggesting that chondrocytes in the CEP maintain homeostasis of the IVD tissue \\[[@pone.0218298.ref035]\\]. Similarly, in our study, the passive cigarette smoking-induced dysfunction of CEP cells may have led to the decrease in type II collagen and PG, and caused changes in the NP or changes in the architecture through NP cells. Ariga *et al*. also reported similar findings, i.e., aging-induced apoptosis and destruction of the CEP structure in mouse IVD, followed by NP and IVD degeneration \\[[@pone.0218298.ref036]\\]. Wang *et al*. reported that Fas receptor expression and apoptotic cells were increased in the CEP in addition to degeneration of human IVD \\[[@pone.0218298.ref037]\\]. These studies support the involvement of apoptosis of chondrocytes in the CEP during the course of IVD degeneration.\n\nBased on this study, we hypothesized the following molecular mechanisms underlying IVD degeneration ([Fig 6](#pone.0218298.g006){ref-type=\"fig\"}): 1) Passive cigarette smoking reduces blood flow, which most significantly influences the CEP. 2) Consequently, the potential for apoptosis is increased, and type II collagen and PG levels decrease around the chondrocytes. 3) This influence is transmitted to the NP, leading to further reduction of the production of type II collagen and PG. At the same time, structural changes of the NP cells and destruction of the tissue structure occurs. Regarding 1), a porcine model of cigarette smoking shows a significant reduction in solute transport from the blood capillaries in the IVD \\[[@pone.0218298.ref038]\\]. In our previous study on a rabbit model, nicotine treatment results in delineation of vascular buds in the vicinity of the vertebral endplate and a reduction of their numbers \\[[@pone.0218298.ref018]\\]. it has been reported that cigarette smoking induces carbon monoxide production, which promotes degradation of hypoxia inducible factor-1 (HIF-1\u03b1) and inhibits vascularization \\[[@pone.0218298.ref039], [@pone.0218298.ref040]\\]. Blood flow into the IVD may be likely decreased in our passive cigarette smoking rat model due to vasoconstriction induced by nicotine. Regarding 2), it is well known that hypoxia and ischemia can cause apoptosis \\[[@pone.0218298.ref041]\\]. HIF-2\u03b1 regulates Fas-mediated chondrocyte apoptosis during osteoarthritic cartilage destruction \\[[@pone.0218298.ref042]\\], and the expression of HIF-1\u03b1 has been reported to correlate significantly with apoptosis in human herniated discs \\[[@pone.0218298.ref043]\\]. As such, reduced blood flow-induced hypoxia may also have induced chondrocyte apoptosis in the CEP in our model. However, we did not assess mRNA expression in isolated CEP cells. Thus, further analyses of apoptosis-related gene expression in isolated CEP cells may more clearly demonstrate such changes, as observed in IVD cells. Regarding 3), as similar findings have been reported \\[[@pone.0218298.ref036]\\], apoptosis of the CEP may lead to degeneration of the NP and IVD. Wang *et al*. suggested that the occurrence of Fas-mediated apoptosis, which is promoted within the CEP, is not unidirectional, but indeed represents mutual interactions between these tissues and cells \\[[@pone.0218298.ref037]\\]. Recently, Elmasry *et al*. suggested that both direct and indirect effects of smoking play significant roles in IVD degeneration: the nicotine-mediated down-regulation of cell proliferation and anabolism mainly affects GAG levels in the CEP, and the reduction of solute exchange between blood vessels and disc tissue mainly affects GAG levels and cell density in the NP \\[[@pone.0218298.ref024]\\]. Thus, there are possible alternative mechanisms responsible for the effects of cigarette smoke on the CEP and NP: the direct effects of nicotine on the CEP and the reduction of transport of nutrients through the CEP to the NP.\n\n![Schema of the molecular mechanisms underlying IVD degeneration induced by passive cigarette smoking in rats.](pone.0218298.g006){#pone.0218298.g006}\n\nWe demonstrated the possibility of chondrocyte apoptosis within the CEP of rat IVD in response to passive smoking. Changes were accompanied by decreases in type II collagen and PG in the NP, leading to destruction of the NP architecture. Apoptosis was suggested by the detection of chondrocytes that were positive for ssDNA; however, definitive morphological features of apoptosis were not observed in this study \\[[@pone.0218298.ref043]\\]. Therefore, further studies are needed to elucidate the extent of true apoptosis within this region of the IVD in rats exposed to passive cigarette smoking.\n\nSupporting information {#sec012}\n======================\n\n###### Measurement area for histological examination.\n\n\\(A\\) Immunohistochemistry for type II collagen. (B) Immunohistochemistry for aggrecan. (C) Safranin O staining. The upper and lower sides of each panel represent the cranial and caudal directions, respectively. The left and right sides of each panel represent the ventral and dorsal directions, respectively. The positive NP area ratio was the ratio of the positive area to the entire NP, and the positive AF area ratio was the ratio of the positive area to the dorsal AF with a diameter of 300 \u03bcm. The CEP was divided into the peripheral and central regions, and the positive area percentage was calculated in each region. The bar indicates 200 \u03bcm.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Quantitative mRNA analysis.\n\n\\(A\\) Type II collagen and Mmp13 mRNA. (B) Aggrecan, Mmp3, and Adamts4 mRNA. Two IVD were excised from each rat and homogenized in TRIZOL (Invitrogen, Carlsbad, USA) to extract total RNA. The extracted RNA was treated with DNase I, and then reacted with random hexamer primers and Prime Script reverse transcriptase (TAKARA, Kyoto, Japan) at 30\u00b0C for 10 min, 45\u00b0C for 60 min, and 70\u00b0C for 15 min to synthesize cDNA. Using the TaqMan Gene Expression Assay (Applied Biosystems, Foster, USA), PCR was performed according to the manufacturer's instructions using a Rotor-Gene 6000 real-time analyzer (Corbett Life Science QIAGEN, Alabama, USA). 18S rRNA was measured as an endogenous control for correction of the gene expression levels. For quantification, the absolute quantification method was employed, in which a calibration curve was prepared for each gene from 5-fold serial dilutions using cDNA with the highest expression level as the standard. The measured expression level of each gene was divided by the measured 18S rRNA expression level to calculate the normalized value. This normalized value was compared between the groups. The significance of the differences was analyzed using the Mann-Whitney U test.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Immunostaining for ssDNA in the NP and AF.\n\nLeft and right panels represent low and high magnification, respectively. Bars indicate 1 mm and 200 \u03bcm, respectively. N4, non-smoking control for 4 weeks; S4, passive smoking for 4 weeks; N8, non-smoking control for 8 weeks; S8, passive smoking for 8 weeks.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Fragmentation of \u03b2-actin.\n\n\\(A\\) Cleavage sites of caspase 1 and caspase 3 in rat \u03b2-actin. The forty-two-kDa \u03b2-actin is cleaved by caspases-1 at 2 aspartic acid (Asp) residues at positions 11 and 244, producing a 29-kDa fragment. It is also cut at Asp 244 by caspase 3 to produce a 32-kDa fragment. The thick bar from amino acid residues 1 to 100 indicates the epitope of the anti-\u03b2-actin antibody used in this study. (B) Immunoblot analysis of \u03b2-actin from the IVD (NP and AF). Three IVD from each of 5 rats were combined, and protein was extracted. The IVD were mechanically ground using a mortar, cooled in liquid nitrogen, and extracted with shaking in 1 ml of guanidine hydrochloride extraction solution (4 M guanidine HCl, 50 mM sodium acetate, 65 mM DTT, 10 mM EDTA, Complete Mini Protease Inhibitor Cocktail (Roche), pH 8.5) at 4\u00b0C overnight. After centrifugation at 30,000*(x g)* for 5 min, precipitated collagen fibers were removed, and the supernatant was centrifuged again to remove macromolecular proteins that were 100 kDa or larger using a 100 kDa molecular weight cut off centrifugal filter (Millipore, CA, USA). The filtrate was used as the protein extract. The extract was mixed with 9 volumes of 100% ethanol to precipitate any protein. The precipitate was washed and resolved with 1xSDS-PAGE loading buffer. Forty \u03bcg of protein was applied to a 12.5% polyacrylamide gel and electrophoresed, followed by blotting onto a nitrocellulose membrane using the iBlot Dry Blotting System (Carlsbad, USA). The membrane was blocked with 5% skim milk/PBS at room temperature for one hour and then reacted with 0.3 \u03bcg/ml of mouse monoclonal anti-\u03b2-actin antibody (Abcam, Cambridge, UK) at room temperature for one hour, followed by reaction with 0.02 \u03bcg/ml of HRP-conjugated anti-mouse IgG goat antibody at room temperature for 30 minutes. Chemiluminescence was induced using ECL Advance (GE Healthcare, Buckingham, UK) and detected using Light-Capture (Atto, Tokyo, Japan). Two bands corresponding to \u03b2-actin, 42- and 32-kDa, were also detected in the IVD in the non-smoking control groups, suggesting that physiological cleavage of \u03b2-actin by caspase 3 occurs in the normal rat IVD. N4, non-smoking control for 4 weeks; S4, passive smoking for 4 weeks; N8, non-smoking control for 8 weeks; S8, passive smoking for 8 weeks. (C) Quantification of immunoblotting for \u03b2-actin. Immunoblot signals were subjected to molecular weight measurement and quantitative analysis using CS Analyzer 2.0 (Atto) and MultiGauge (FUJIFILM, Tokyo, Japan), respectively. The proportion of \u03b2-actin fragmentation by caspase 3 was calculated. Passive cigarette smoking did not induce any change in the fragmentation rate.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Rat body weight during passive cigarette smoking (grey) compared with that of non-smoking control (white).\n\n\\* indicates significant decrease in body weight gain in smoking rats (p\\<0.05 by Mann-Whitney U test).\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Comparison of mRNA levels in the intervertebral disc (IVD) between passive smoking and non-smoking control rats.\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\nWe thank Dr. Yoshiaki Kusumi for valuable discussions regarding the histological examination, and Rie Takahashi and Mika Sakamoto for their assistance in tissue staining.\n\n10.1371/journal.pone.0218298.r001\n\nDecision Letter 0\n\nSampen\n\nHee-Jeong Im\n\nAcademic Editor\n\n\u00a9 2019 Hee-Jeong Im Sampen\n\n2019\n\nHee-Jeong Im Sampen\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n13 Jun 2019\n\nPONE-D-19-14928\n\nDetection of apoptosis and matrical degeneration within the intervertebral discs of rats due to passive cigarette smoking.\n\nPLOS ONE\n\nDear Prof. Esumi,\n\nThank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE's publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.\n\nWe would appreciate receiving your revised manuscript by Jul 28 2019 11:59PM. When you are ready to submit your revision, log on to <https://www.editorialmanager.com/pone/> and select the \\'Submissions Needing Revision\\' folder to locate your manuscript file.\n\nIf you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.\n\nTo enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: <http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols>\n\nPlease include the following items when submitting your revised manuscript:\n\nA rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled \\'Response to Reviewers\\'.A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled \\'Revised Manuscript with Track Changes\\'.An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled \\'Manuscript\\'.\n\nPlease note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.\n\nWe look forward to receiving your revised manuscript.\n\nKind regards,\n\nHee-Jeong Im Sampen, PhD\n\nAcademic Editor\n\nPLOS ONE\n\nJournal Requirements:\n\nWhen submitting your revision, we need you to address these additional requirements.\n\n1\\. Please ensure that your manuscript meets PLOS ONE\\'s style requirements, including those for file naming. The PLOS ONE style templates can be found at\n\n<http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf> and <http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf>\n\n2.\u00a0 Thank you for stating the following in the Financial Disclosure section:\\\"The Department of Therapeutics for Aging Locomotive Disorders is funded by a combination of grants and individual contributions. Funding for this study was provided by charitable donations from Ono Pharmaceutical Co., Ltd., Stryker Japan K. K. and Nakashima Medical Co., Ltd. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\\\"\n\nWe note that you received funding from a commercial source: \\\"Ono Pharmaceutical Co., Ltd\\\" and \\\"Nakashima Medical Co., Ltd\\\"\n\nPlease provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other [relevant declarations](http://www.plosone.org/static/competing.action) relating to employment, consultancy, patents, products in development, marketed products, etc.\n\nWithin this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: \\\"This does not alter our adherence to PLOS ONE policies on sharing data and materials.\" (as detailed online in our guide for authors <http://journals.plos.org/plosone/s/competing-interests>). \u00a0If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.\n\nPlease include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf.\n\nPlease know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: <http://journals.plos.org/plosone/s/competing-interests>\n\n3.\u00a0 Please note that all PLOS journals ask authors to adhere to our policies for sharing of data and materials: <https://journals.plos.org/plosone/s/data-availability>. According to PLOS ONE's Data Availability policy, we require that the minimal dataset underlying results reported in the submission must be made immediately and freely available at the time of publication. As such, please remove any instances of \\'unpublished data\\' or \\'data not shown\\' in your manuscript and replace these with either the relevant data (in the form of additional figures, tables or descriptive text, as appropriate), a citation to where the data can be found, or remove altogether any statements supported by data not presented in the manuscript.\n\n4\\. To comply with PLOS ONE submissions requirements, in your Methods section, please provide additional information on the animal research and ensure you have included details on the frequency of welfare monitoring and any efforts to alleviate suffering.\n\nAdditional Editor Comments (if provided):\n\n\\[Note: HTML markup is below. Please do not edit.\\]\n\nReviewers\\' comments:\n\nReviewer\\'s Responses to Questions\n\n**Comments to the Author**\n\n1\\. Is the manuscript technically sound, and do the data support the conclusions?\n\nThe manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.\n\nReviewer \\#1: Partly\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n2\\. Has the statistical analysis been performed appropriately and rigorously?\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n3\\. Have the authors made all data underlying the findings in their manuscript fully available?\n\nThe [PLOS Data policy](http://www.plosone.org/static/policies.action#sharing) requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data---e.g. participant privacy or use of data from a third party---those must be specified.\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n4\\. Is the manuscript presented in an intelligible fashion and written in standard English?\n\nPLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n5\\. Review Comments to the Author\n\nPlease use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)\n\nReviewer \\#1: In this study, Nakahashi et al. established a rat model of passive cigarette smoking induced IVD degeneration, and investigated the cytohistological changes in the IVD.\n\nThey found that passive cigarette smoking could stimulate chondrocyte apoptosis in the end plate and reduce the extracellular matrix (ECM) of in the NP, which can then progress to IVD degeneration. This study was interesting and well written. There are few comments for this study.\n\n1\\. After euthanasia, the spine samples were harvested and then fixed in formalin or frozen at -80\u00b0C. The authors should consider doing the perfusion (first with saline to remove blood cells and followed by formalin to preserve the antigen). After perfusion, the samples would not contain the blood cells, which may affect the immunohistochemical staining. By the way, what are the samples frozen at -80\u00b0C use for? Are these samples for mRNA measurement? The authors should clarify.\n\n2\\. The authors found that type II collagen and aggrecan were decreased in the NP and CEP, while the synthesis of these two proteins were not affected by the smoking, as reflected by the mRNA repression measurement. How did the authors measure the mRNA? The authors should state the specific methods in the Methods section.\n\n3\\. The evidence about the passive cigarette smoking-induced apoptosis in the CEP was not strong. The authors should add some other experiments, such as TUNEL staining. Also, the authors found that no significant changes in apoptosis-related gene expression were observed in the NP or AF in response to passive cigarette smoking. What about the protein expression? they should consider measuring the related protein expression via immunohistochemical staining.\n\n4\\. In the conclusion, the authors stated passive cigarette smoking-induced stress stimuli first affect the CEP through blood flow. However, there was no results reflecting the blood flow. How did they prove that the blood flow was affected?\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n6\\. PLOS authors have the option to publish the peer review history of their article ([what does this mean?](https://journals.plos.org/plosone/s/editorial-and-peer-review-process#loc-peer-review-history)). If published, this will include your full peer review and any attached files.\n\nIf you choose \"no\", your identity will remain anonymous but your review may still be made public.\n\n**Do you want your identity to be public for this peer review?** For information about this choice, including consent withdrawal, please see our [Privacy Policy](https://www.plos.org/privacy-policy).\n\nReviewer \\#1: No\n\n\\[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link \\\"View Attachments\\\". If this link does not appear, there are no attachment files to be viewed.\\]\n\nWhile revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, <https://pacev2.apexcovantage.com/>. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at <figures@plos.org>. Please note that Supporting Information files do not need this step.\n\n10.1371/journal.pone.0218298.r002\n\nAuthor response to Decision Letter 0\n\n26 Jul 2019\n\nJuly 26, 2019\n\nManuscript ID PONE-D-19-14928\n\nTitle: Detection of apoptosis and matrical degeneration within the intervertebral discs of rats due to passive cigarette smoking.\n\nAuthor: Masahiro Nakahashi et al.\n\nDear Dr. Joerg Heber, Editor-in-Chief, PLOS ONE\n\n:\n\nWe appreciate the time and efforts by the editor and referees in reviewing this manuscript. We also thank the reviewers for valuable comments and suggestion. We have addressed all issues indicated in the review report, and believed that the revised version can meet the journal publication requirements. Please find our point-by-point responses to the editor's and reviewer's comments and the revised manuscript highlighting the revisions made.\n\nIn accordance with the editor's comment, we attached our revised Competing Interests Statement here:\n\nThis work was funded in part by Ono Pharmaceutical Co., Ltd., Stryker Japan K. K. and Nakashima Medical Co., Ltd. received by YT. There are no patents, products in development or marketed products associated with this research to declare. This does not alter our adherence to PLOS ONE policies on sharing data and materials.\n\nSincerely yours,\n\nMariko Esumi\n\nDepartment of Biomedical Sciences\n\nNihon University School of Medicine\n\nOhyaguchikami-cho, Itabashi-ku\n\nTokyo 173-8610, Japan\n\nPhone: +81-3-3972-8111 Ext. 2241\n\nFax: +81-3-3972-8199\n\nE-mail: <esumi.mariko@nihon-u.ac.jp>\n\nPoint-by-point responses to editor's and reviewer's comments:\n\nTo Editor\n\nComment 1:\n\nPlease ensure that your manuscript meets PLOS ONE\\'s style requirements, including those for file naming.\n\nResponse:\n\nWe have reviewed the PLOS ONE style templates and corrected our manuscript.\n\nComment 2:\n\nWe note that you received funding from a commercial source: \\\"Ono Pharmaceutical Co., Ltd\\\" and \\\"Nakashima Medical Co., Ltd\\\"\n\nPlease provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc.\n\nWithin this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: \\\"This does not alter our adherence to PLOS ONE policies on sharing data and materials.\"\n\nResponse:\n\nWe have corrected the Competing Interests Statement as follows:\n\nThis work was funded in part by Ono Pharmaceutical Co., Ltd., Stryker Japan K. K. and Nakashima Medical Co., Ltd. received by YT. There are no patents, products in development or marketed products associated with this research to declare. This does not alter our adherence to PLOS ONE policies on sharing data and materials.\n\nPlease include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf.\n\nResponse:\n\nWe have included our amended COI statement within our cover letter.\n\nComment 3:\n\nPlease note that all PLOS journals ask authors to adhere to our policies for sharing of data and materials: <https://journals.plos.org/plosone/s/data-availability>. According to PLOS ONE's Data Availability policy, we require that the minimal dataset underlying results reported in the submission must be made immediately and freely available at the time of publication. As such, please remove any instances of \\'unpublished data\\' or \\'data not shown\\' in your manuscript and replace these with either the relevant data (in the form of additional figures, tables or descriptive text, as appropriate), a citation to where the data can be found, or remove altogether any statements supported by data not presented in the manuscript.\n\nResponse:\n\nIn accordance with the editor's comment, we have revised the Introduction on page 3, line 67 by adding a reference 27 (written in Japanese) on page 19, lines 475-478. We have also revised Fig 2B by adding a graph of AF; accordingly, we have revised the Result on page 8, line 192-194 and Fig 2B legend on page 8, lines 212 and 213.\n\nComment 4:\n\nTo comply with PLOS ONE submissions requirements, in your Methods section, please provide additional information on the animal research and ensure you have included details on the frequency of welfare monitoring and any efforts to alleviate suffering.\n\nResponse:\n\nIn accordance with the editor's comment, we have revised the Materials and Methods on page 4, lines 84-88, and lines 93-94.\n\n\u2003\n\nTo Reviewer \\#1\n\nComment 1:\n\nAfter euthanasia, the spine samples were harvested and then fixed in formalin or frozen at -80\u00b0C. The authors should consider doing the perfusion (first with saline to remove blood cells and followed by formalin to preserve the antigen). After perfusion, the samples would not contain the blood cells, which may affect the immunohistochemical staining.\n\nResponse:\n\nIn order to obtain the intact RNA and exactly analyze the mRNA expression, we did not perform the perfusion. Instead, we performed cardiac puncture for blood removal, and quickly excised the entire spine. IVD is\u3000avascular tissue by nature, and there was no problem about the immunohistochemical staining in our experiment.\n\nBy the way, what are the samples frozen at -80\u00b0C use for? Are these samples for mRNA measurement? The authors should clarify.\n\nResponse:\n\nYes, the samples frozen at -80\u00b0C were used for mRNA and protein analysis. In accordance with the reviewer's comment, we have revised the Materials and Methods on page 4, lines 96 and 97.\n\nComment 2:\n\nThe authors found that type II collagen and aggrecan were decreased in the NP and CEP, while the synthesis of these two proteins were not affected by the smoking, as reflected by the mRNA repression measurement. How did the authors measure the mRNA? The authors should state the specific methods in the Methods section.\n\nResponse:\n\nIn accordance with the reviewer's comment, we have added the section 'Quantification of mRNA' in the Materials and Methods on page 4-5, lines 103-110. We have also added a reference 30 on page 19, lines 487-490.\n\nComment 3:\n\nThe evidence about the passive cigarette smoking-induced apoptosis in the CEP was not strong. The authors should add some other experiments, such as TUNEL staining. Also, the authors found that no significant changes in apoptosis-related gene expression were observed in the NP or AF in response to passive cigarette smoking. What about the protein expression? they should consider measuring the related protein expression via immunohistochemical staining.\n\nResponse:\n\nWe agree with the reviewer that other apoptosis assessments are necessary to confirm the apoptosis of the CEP. Actually, we tried to assess the apoptosis of the CEP by cleaved caspase-3 staining and TUNEL method. Unfortunately, the two methods did not work in our system, even in the positive control specimen, the thymus of rats treated with dexamethasone. It seems to be due to the tissue preparation such as fixation in our experiment. Therefore, we have not done the immunohistochemistry of other apoptosis-related protein. Only the ssDNA staining worked in the positive control tissue and was used here.\n\nThus, considering the limitation, we dampened the statement and discussed on the apoptosis of CEP, followed by several supportive data of other groups.\n\nComment 4:\n\nIn the conclusion, the authors stated passive cigarette smoking-induced stress stimuli first affect the CEP through blood flow. However, there was no results reflecting the blood flow. How did they prove that the blood flow was affected?\n\nResponse:\n\nAs the reviewer pointed out, we have no evidence for the reduction of blood flow in our experiment. We have revised the Abstract on page 2, line 36: from 'concluded' to 'hypothesized'.\n\nHowever, there have been several reports on blood flow reduction in the IVD induced by cigarette smoke and nicotine. We have revised the Discussion on page 13, lines 329-332 and 335 by citing these reports (references 38 and 18). We have also added a reference 38 on pages 22-23, lines 520-523.\n\n###### \n\nSubmitted filename: Response to Reviewers.docx\n\n###### \n\nClick here for additional data file.\n\n10.1371/journal.pone.0218298.r003\n\nDecision Letter 1\n\nSampen\n\nHee-Jeong Im\n\nAcademic Editor\n\n\u00a9 2019 Hee-Jeong Im Sampen\n\n2019\n\nHee-Jeong Im Sampen\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n5 Aug 2019\n\nDetection of apoptosis and matrical degeneration within the intervertebral discs of rats due to passive cigarette smoking.\n\nPONE-D-19-14928R1\n\nDear Dr. Esumi,\n\nWe are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.\n\nWithin one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.\n\nShortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at <https://www.editorialmanager.com/pone/>, click the \\\"Update My Information\\\" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at <authorbilling@plos.org>.\n\nIf your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact <onepress@plos.org>.\n\nWith kind regards,\n\nHee-Jeong Im Sampen, PhD\n\nAcademic Editor\n\nPLOS ONE\n\nAdditional Editor Comments (optional):\n\nReviewers\\' comments:\n\nReviewer\\'s Responses to Questions\n\n**Comments to the Author**\n\n1\\. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the \"Comments to the Author\" section, enter your conflict of interest statement in the \"Confidential to Editor\" section, and submit your \\\"Accept\\\" recommendation.\n\nReviewer \\#1: All comments have been addressed\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n2\\. Is the manuscript technically sound, and do the data support the conclusions?\n\nThe manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n3\\. Has the statistical analysis been performed appropriately and rigorously?\n\nReviewer \\#1: N/A\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n4\\. Have the authors made all data underlying the findings in their manuscript fully available?\n\nThe [PLOS Data policy](http://www.plosone.org/static/policies.action#sharing) requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data---e.g. participant privacy or use of data from a third party---those must be specified.\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n5\\. Is the manuscript presented in an intelligible fashion and written in standard English?\n\nPLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.\n\nReviewer \\#1: Yes\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n6\\. Review Comments to the Author\n\nPlease use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)\n\nReviewer \\#1: The authors have addressed all the comments that I raised. This paper is interesting. There is no further comments.\n\n\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\n\n7\\. PLOS authors have the option to publish the peer review history of their article ([what does this mean?](https://journals.plos.org/plosone/s/editorial-and-peer-review-process#loc-peer-review-history)). If published, this will include your full peer review and any attached files.\n\nIf you choose \"no\", your identity will remain anonymous but your review may still be made public.\n\n**Do you want your identity to be public for this peer review?** For information about this choice, including consent withdrawal, please see our [Privacy Policy](https://www.plos.org/privacy-policy).\n\nReviewer \\#1: No\n\n10.1371/journal.pone.0218298.r004\n\nAcceptance letter\n\nSampen\n\nHee-Jeong Im\n\nAcademic Editor\n\n\u00a9 2019 Hee-Jeong Im Sampen\n\n2019\n\nHee-Jeong Im Sampen\n\nThis is an open access article distributed under the terms of the\n\nCreative Commons Attribution License\n\n, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.\n\n20 Aug 2019\n\nPONE-D-19-14928R1\n\nDetection of apoptosis and matrical degeneration within the intervertebral discs of rats due to passive cigarette smoking.\n\nDear Dr. Esumi:\n\nI am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.\n\nIf your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact <onepress@plos.org>.\n\nFor any other questions or concerns, please email <plosone@plos.org>.\n\nThank you for submitting your work to PLOS ONE.\n\nWith kind regards,\n\nPLOS ONE Editorial Office Staff\n\non behalf of\n\nDr. Hee-Jeong Im Sampen\n\nAcademic Editor\n\nPLOS ONE\n\n[^1]: **Competing Interests:**This work was funded in part by Ono Pharmaceutical Co., Ltd., Stryker Japan K. K. and Nakashima Medical Co., Ltd. received by YT. There are no patents, products in development or marketed products associated with this research to declare. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nCiclopirox olamine (CPX) (also called Batrafen, Loprox, Penlac and Stieprox), a synthetic hydroxypyridone derivative, is currently used for the treatment of superficial fungal infections and available in a variety of formulations, including cream, lotion, gel, nail lacquer and shampoos \\[[@R1]\\]. CPX has a broad spectrum of action against dermatophytes, yeast, filamentous fungi and bacteria. Mechanistically, CPX has been proposed to act as an iron chelator, forming complexes with trivalent metal cations, such as Fe^3+^, and inhibiting metal-dependent enzymes, such as catalase and peroxidase, which play an essential role in the intracellular degradation of toxic peroxides, though this remains to be determined \\[[@R2]\\]. Recent studies have shown that CPX induces cell death in different cancer cells, such as primary human acute myeloid leukemia cells, human breast cancer MDA-MB231 cells and human rhabdomyosarcoma Rh30 cells \\[[@R3], [@R4]\\]. However, the mechanism by which CPX induces cancer cell death is only at the beginning to be investigated. Eberhard *et al*. reported that CPX displays preclinical anticancer activity against hematologic malignancies and induces cell death through its ability to chelate intracellular iron and inhibit the iron-dependent enzyme ribonucleotide reductase \\[[@R4]\\]. We showed that CPX inhibits tumor growth in human breast cancer MDA-MB231 xenografts and induces cell death through caspase-dependent and caspase-independent mechanisms in Rh30 cells \\[[@R3]\\].\n\nAutophagy is an evolutionarily conserved catabolic process that involves the degradation of the components of a cell through the lysosomal machinery \\[[@R5]\\]. During autophagy, a portion of the cytoplasmic materials and intracellular organelles are sequestered into a double membrane organelles known as autophagosomes, which degrade the sequestered contents by fusion with lysosomes to form autolysosomes \\[[@R5]\\]. Under some conditions, autophagy contributes to cellular survival by providing nutrients and energy to help cells adapt to starvation or stress (such as hypoxia, X-ray and anticancer drugs) \\[[@R6]\\]. However, under other settings, activated autophagy leads to cell death, called autophagic cell death or type II programmed cell death \\[[@R6]\\].\n\nIncreasing evidence has implicated that multiple stress conditions such as oxidative stress \\[[@R7]\\], endoplasmic reticulum (ER) stress \\[[@R8]\\] and pathogen infection \\[[@R9]\\], can induce autophagy through different molecular pathways. Among them, reactive oxygen species (ROS) function as signaling molecules not only in cell growth, differentiation, proliferation and apoptosis \\[[@R10]\\], but also in autophagy \\[[@R11]\\]. For example, hydrogen peroxide and 2-methoxyestradiol, two ROS-generating agents, can induce autophagic cell death in transformed HEK293 cells and cervical cancer (HeLa) cells \\[[@R12]\\]. Under starvation or oxidative stress conditions, ROS are increased and play a critical role in autophagosome formation through targeting cysteine protease Atg 4 \\[[@R13]\\]. Although the role of ROS in the regulation of autophagy has been confirmed, what signaling molecules involved in ROS-induced autophagy are still not well understood.\n\nMitogen-activated protein kinases (MAPKs) are evolutionarily conserved dual (Tyr and Ser/Thr) protein kinases, and play an important role in signal transduction from the cell surface to the nucleus \\[[@R14]\\]. To date, different groups of MAPKs have been characterized in mammals: the extracellular signal-regulated kinases ERK1/2, ERK3/4, ERK5, ERK7/8, the Jun N-terminal kinases JNK1/2/3 and the p38 MAPKs p38\u03b1/\u03b2/\u03b3/\u03b4 \\[[@R14]\\]. Not only receptor-ligand interactions, but also different stress stimuli such as the oxidative stress caused by ROS, can induce potential activation of MAPK pathways \\[[@R15]\\]. Depending on the cell type and the stimulus, ERK signaling pathway mediates different cell responses, such as proliferation, apoptosis and autophagy \\[[@R14]--[@R16]\\]. Generally, growth factors may activate the RAS-RAF-MEK1/2-ERK1/2 pathway \\[[@R14]\\]. More recently, ERK5, also termed big mitogen-activated protein kinase-1 (BMK1), has been identified as a component of a parallel MAPK pathway, which is associated with a diverse range of cellular processes including cellular proliferation, migration, survival and angiogenesis \\[[@R17]\\]. The demonstration that the commonly used MEK1/2 inhibitors inactivated ERK5 suggested that ERK5 might regulate some of the cellular functions originally attributed to MEK1/2 \\[[@R17], [@R18]\\]. Studies have demonstrated that mammalian p38 MAPK has four isoforms: p38\u03b1, p38\u03b2, p38\u03b3 and p38\u03b4, of which p38\u03b1 is ubiquitously expressed, and best characterized \\[[@R14],[@R19]\\]. p38 MAPK plays an essential role in the regulation of many cellular events including inflammation, cell growth, death, and differentiation \\[[@R14], [@R19], [@R20]\\]. Recently it has been found that p38 MAPK can also mediate autophagy in response to chemotherapeutic agents \\[[@R21]\\]. JNK, also known as the stress-activated protein kinase, has been implicated in apoptosis and autophagy \\[[@R21]--[@R23]\\]. In mammals, there are three JNK genes: JNK1, JNK2, and JNK3 \\[[@R24]\\]. JNK1 and JNK2 are ubiquitously expressed, while JNK3 is mainly expressed in brain, cardiac smooth muscle and testes \\[[@R24]\\]. In addition to apoptosis, JNK also contributes to autophagic induction in response to stress signals \\[[@R21]\\]. Particularly, ROS can induce JNK-dependent autophagy \\[[@R25], [@R26]\\].\n\nHere we found that CPX induced autophagy in rhabdomyosarcoma (Rh30 and RD) cells, which was mediated by ROS induction, leading to activation of JNK cascade. Further, inhibition of autophagy by chloroquine (CQ) increased the cell death induced by CPX, indicating that CPX-induced autophagy played a pro-survival role in human rhabdomyosarcoma cells. Our findings suggest that combination treatment with CPX and pharmacological autophagy inhibitors might be a promising strategy for rhabdomyosarcoma therapy.\n\nRESULTS {#s2}\n=======\n\nCPX reduces cell viability and alters morphology in rhabdomyosarcoma cells {#s2_1}\n--------------------------------------------------------------------------\n\nTo investigate the cytotoxicity of CPX in rhabdomyosarcoma cells, RD and Rh30 cells were treated with various concentrations of CPX for 72 h, followed by MTS assay. As shown in Figure [1A](#F1){ref-type=\"fig\"}, CPX decreased the cell viability in both RD and Rh30 cells in a concentration-dependent manner. By phase-contrast microscopic observation, more cells became round or detached from the culture plate when exposed to 20 \u03bcM of CPX (Figure [1B](#F1){ref-type=\"fig\"}). To determine whether CPX causes cell death by inducing apoptosis, cells were analyzed by flow cytometry following Annexin V-FITC and propidium iodide (PI) staining. As shown in Figure [1C](#F1){ref-type=\"fig\"}, 20 \u03bcM of CPX treatment for 48 and 72 h significantly increased the percentage of apoptotic cells compared with control cells.\n\n![CPX decreases cell viability in rhabdomyosarcoma cells\\\n**(A)** RD and Rh30 cells were treated with indicated concentrations of CPX for 72 h. Cell viability was determined by MTS assay. **(B)** Rh30 and RD cells were treated with 5 and 20 \u03bcM of CPX for 72 h, followed by taking images under an Olympus inverted phase-contrast microscope equipped with the Quick Imaging system. **(C)** RD cells were treated with 0 and 20 \u03bcM of CPX for 0--72 h or incubated with 0--10 \u03bcM of CPX for 72 h. The cells were then harvested and processed for Annexin V-FITC/PI staining and flow cytometry. Results are presented as mean \u00b1 SD (n=3). \\**P* \\< 0.05, \\*\\**P* \\< 0.01, \\*\\*\\**P* \\< 0.001, difference versus control group.](oncotarget-05-10140-g001){#F1}\n\nCPX induces autophagy in rhabdomyosarcoma cells {#s2_2}\n-----------------------------------------------\n\nRecently we have demonstrated that CPX induces caspase-dependent and independent cell death in Rh30 cells \\[[@R3]\\]. Since autophagy contributes to cell death in some cases \\[[@R6]\\], here we studied whether CPX induces autophagy in the cancer cells. Autophagy is characterized by the increased acidic vesicular organelles (AVOs), which are correlated with increased autophagosomes. We therefore investigated whether CPX could induce autophagy by staining with acridine orange (AO) in RD cells. As shown in Figure [2A](#F2){ref-type=\"fig\"}, 20 \u03bcM of CPX treatment for 24 h or 48 h induced the accumulation of AVO in the cytoplasm of RD cells.\n\n![CPX induces autophagy in rhabdomyosarcoma cells\\\n**(A)** Representative images of AO staining of RD cells following treatment with 0--20 \u03bcM of CPX for 24 and 48 h. Red color intensity shows acidic vesicular organelles, representing autophagolysosomes. **(B)** Representative micrographs of cells that show GFP-LC3 localization. Rh30 cells stably expressing GFP-LC3 were treated without or with 20 \u03bcM of CPX for indicated time (Upper panel), or with indicated concentrations of CPX for 24 h (Bottom panel), and then visualized under a fluorescent microscope. Control cells presented a diffuse distribution of GFP-LC3, whereas CPX-treated cells displayed a punctate pattern of GFP-LC3 expression (arrows), indicating formation of autophagosomes. **(C)** Western blot analysis of autophagy-related proteins. RD cells were treated without or with 20 \u03bcM of CPX for indicated time, or with indicated concentrations of CPX for 24 h. The cells were harvested and subjected to Western blot analysis using indicated antibodies. \u03b2-tubulin was used as a loading control.](oncotarget-05-10140-g002){#F2}\n\nLC3 is associated with the formation of the autophagosome membrane and is a specific marker for autophagy initiation \\[[@R5]\\]. Under normal conditions, LC3 is distributed homogeneously in the cytoplasm; when autophagy is induced, LC3 is recruited to the autophagosomal membrane and shows characteristic of GFP-LC3 puncta \\[[@R5]\\]. To elucidate whether CPX induced autophagy in rhabdomyosarcoma cells, the effect of CPX on the cellular localization of LC3 using Rh30 cells stably expressing GFP-LC3 was evaluated. As shown in Figure [2B](#F2){ref-type=\"fig\"}, CPX treatment induced a punctuated fluorescent pattern of LC3, whereas untreated cells manifested a diffuse distribution of GFP-LC3, suggesting that CPX indeed induced autophagy in the cells.\n\nSince LC3-I is converted to the hallmark autophagosome-associating protein LC3-II, expression of LC3-II has been widely used for monitoring autophagy \\[[@R27]\\]. To verify the above finding, we further tested whether CPX increases LC3-II protein level. The Western blotting results indicated that CPX potently increased LC3-II level in a time- and concentration-dependent manner in RD cells (Figure [2C](#F2){ref-type=\"fig\"}). Moreover, induction of autophagy by CPX was identified by assessing the expressions of Beclin-1, a key regulator of autophagosome formation and p62/SQSTM1, a protein facilitating autophagic degradation of ubiquitinated protein aggregation \\[[@R28]\\]. CPX treatment decreased the expression of p62, but did not affect Beclin-1 expression (Figure [2C](#F2){ref-type=\"fig\"}). Taken together, these results demonstrate that CPX induces autophagy in rhabdomyosarcoma cells.\n\nCPX induces autophagy via ROS induction {#s2_3}\n---------------------------------------\n\nGrowing evidence shows that ROS are important regulators of autophagy under various conditions \\[[@R11]\\]. To determine whether CPX-induced autophagy is associated with ROS induction, we first measured ROS level in both RD and Rh30 cells treated with CPX using the ROS-detecting fluorescent dye CM-H~2~DCFDA. ROS accumulation was observed after CPX treatment with the culturing time, and increased \\~2-fold after 24 h of CPX (20 \u03bcM) treatment in Rh30 cells (Figure [3A](#F3){ref-type=\"fig\"}). Similar results were seen in RD cells (Figure [3B](#F3){ref-type=\"fig\"}). Addition of 5 mM of NAC, a ROS scavenger, almost completely blocked CPX-induced ROS in both Rh30 and RD cells (Figure [3C](#F3){ref-type=\"fig\"} and [3D](#F3){ref-type=\"fig\"}). Interestingly, pretreatment with NAC remarkably attenuated CPX-induced GFP-LC3 puncta formation (Figure [3E](#F3){ref-type=\"fig\"}) and LC3-II expression in the cells (Figure [3F](#F3){ref-type=\"fig\"}), indicating that CPX-induced autophagy is mediated by ROS induction.\n\n![CPX increases intracellular level of ROS, thereby inducing autophagy\\\n**(A)** *and* **(B)** RD and Rh30 cells were treated without or with different concentrations of CPX (1--20 \u03bcM) for 30 min, followed by loading with 10 \u03bcM of CM-H~2~DCFDA for indicated time. Fluorescent intensity was detected using a microplate reader. Results are presented as mean \u00b1 SD (n=3). \\**P* \\< 0.05, \\*\\**P* \\< 0.01, \\*\\*\\**P* \\< 0.001, difference versus control group. **(C)** *and* **(D)** RD and Rh30 cells were pre-incubated with or without NAC (5 mM) for 30 min, and then treated with or without various concentrations of CPX (0--20 \u03bcM) for 30 min, followed by loading with 10 \u03bcM of CM-H~2~DCFDA for 8 h. Fluorescent intensity was detected using a microplate reader. Results are presented as mean \u00b1 SD (n=3). ^a^*P* \\< 0.05, ^b^*P* \\< 0.01, difference versus control group. ^c^*P* \\< 0.001, difference versus 1 \u03bcM CPX group, ^d^*P* \\< 0.001, difference versus 5 \u03bcM CPX group, ^e^*P* \\< 0.001, difference versus 10 \u03bcM CPX group, ^f^*P* \\< 0.001, difference versus 20 \u03bcM CPX group. **(E)** Rh30 cells stably expressing GFP-LC3 were pretreated with 5 mM of NAC for 1 h, and then incubated with 0 and 20 \u03bcM of CPX for 24 h. The cells were visualized under a fluorescent microscope. **(F)** RD and Rh30 cells were pretreated with 5 mM of NAC for 1 h, and then incubated with 0--20 \u03bcM of CPX for 24 h. The cells were harvested and subjected to Western blot analysis with indicated antibodies. \u03b2-tubulin was used as a loading control.](oncotarget-05-10140-g003){#F3}\n\nROS-mediated JNK activation contributes to CPX-induced autophagy {#s2_4}\n----------------------------------------------------------------\n\nROS have been demonstrated as an inducer or mediator for the activation of MAPK family members, including JNK, p38 and ERK1/2 \\[[@R14]\\]. Also, studies have shown that MAPKs play a pivotal role in autophagy \\[[@R16], [@R21], [@R29]\\]. In this study, we observed that CPX induced phosphorylation of p38\u03b1 (Thr180/Tyr182), ERK1/2 and JNK1/2 (Thr183/Tyr185) in a concentration-dependent manner (Figure [4A](#F4){ref-type=\"fig\"}). Thus, we next asked whether CPX-induced autophagy involves these MAPKs. To this end, selective inhibitors of ERK1/2, p38 and JNK were employed. As shown in Figure [4B](#F4){ref-type=\"fig\"}, pretreatment with U0126, a highly selective inhibitor of MEK1/2 (upstream kinases of ERK1/2), or doramapimod (also named BIRB 796), a highly selective p38\u03b1 MAPK inhibitor \\[[@R30]\\], markedly suppressed CPX-induced phosphorylation of ERK1/2 or p38, respectively. However, neither of the inhibitors apparently affected CPX-induced LC3-II expression, suggesting that ERK1/2 and p38 MAPK are not involved in CPX-induced autophagy. By contrast, presence of JNK inhibitor, SP600125, potently inhibited the activation of JNK pathway and LC3-II expression induced by CPX (Figure [4D](#F4){ref-type=\"fig\"}). To further confirm the role of JNK signaling pathway in CPX-induced autophagy, RD cells were infected with recombinant adenovirus encoding FLAG-tagged dominant negative c-Jun (Ad-c-Jun-DN) or Ad-GFP (as control), and then treated with CPX for 24 h. As shown in Figure [4E](#F4){ref-type=\"fig\"}, ectopic expression of dominant negative c-Jun substantially attenuated CPX-induced LC3-II, demonstrating that activation of JNK pathway contributes to CPX-induced autophagy.\n\n![ROS-activated MAPKs pathway and JNK pathway contributes to CPX-induced autophagy\\\n**(A)** RD cells were incubated with different concentrations of CPX for 24 h, followed by Western blot analysis with indicated antibodies. **(B)** *and* **(C)** RD cells were pretreated with U0126 (5 \u03bcM) or doramapimod (0.5 \u03bcM) for 1 h, and then incubated with 0--20 \u03bcM of CPX for 24 h, followed by Western blot analysis with indicated antibodies. **(D)** RD cells were pretreated with SP600125 (20 \u03bcM) for 1 h, and then incubated with 0--20 \u03bcM of CPX for 24 h. The cells were harvested and subjected to Western blot analysis with indicated antibodies. **(E)** RD cells were infected with Ad-c-Jun-DN or Ad-GFP for 24 h, and then treated with CPX for 24 h. The cells were harvested and subjected to Western blot analysis with indicated antibodies. **(F)** NAC attenuated CPX-induced JNK activation and autophagy. RD cells were pretreated with NAC (5 mM) for 1 h, and then incubated with 0--20 \u03bcM of CPX for 24 h. The cells were harvested and subjected to Western blot analysis with indicated antibodies. \u03b2-tubulin was used as a loading control in A-F.](oncotarget-05-10140-g004){#F4}\n\nTo further investigate the role of ROS induction in CPX-induced autophagy and activation of JNK pathway, ROS scavenger NAC was utilized. We found that pre-treatment with NAC (5 mM) obviously attenuated CPX-induced c-Jun phosphorylation and LC3-II expression (Figure [4F](#F4){ref-type=\"fig\"}), indicating that CPX induces autophagy through ROS-mediated activation of JNK signaling pathway.\n\nAutophagy plays a protective role in CPX-treated rhabdomyosarcoma cells {#s2_5}\n-----------------------------------------------------------------------\n\nSince autophagy has dual roles in cell survival and cell death \\[[@R6], [@R31]\\], to determine whether CPX-induced autophagy is pro-survival or pro-death in rhabdomyosarcoma cells, CQ, a pharmacologic autophagy inhibitor \\[[@R5], [@R6]\\], which inhibits the late autophagic process by blocking the fusion of autophagosomes and lysosomes, was used. As expected, treatment with CQ alone increased LC3-II and p62 expressions (Figure [5A](#F5){ref-type=\"fig\"}). CPX-induced LC3-II expression and p62 degradation were elevated in the presence of 5 \u03bcM of CQ (Figure [5A](#F5){ref-type=\"fig\"}). Importantly, MTS assay showed that inhibition of the late autophagic process with CQ (5 \u03bcM) for 48 h reduced the cell viability very marginally, but potentiated the CPX-reduced cell viability significantly (Figure [5B](#F5){ref-type=\"fig\"}). Furthermore, Annexin V-PI staining also demonstrated that CQ enhanced CPX-induced apoptotic cell death (Figure [5C](#F5){ref-type=\"fig\"}). These data suggest that autophagy protects rhabdomyosarcoma cells from CPX-induced cell death.\n\n![Inhibition of autophagy by CQ enhances CPX-induced autophagy and cell death\\\n**(A)** RD cells were pretreated with 5 \u03bcM of CQ for 1 h, and then incubated with 0--20 \u03bcM of CPX for 24 h. The cells were harvested and subjected to Western blot analysis with indicated antibodies. \u03b2-tubulin was used as a loading control. **(B)** RD cells were pretreated with 5 \u03bcM of CQ for 1 h, and then incubated with 0--20 \u03bcM of CPX for 48 h. Cell viability was measured by MTS assay. Results are presented as mean \u00b1 SD (n=3). \\**P* \\< 0.05, difference versus 1 \u03bcM CPX group, \\*\\**P* \\< 0.01, difference versus 5 \u03bcM CPX group, \\*\\*\\**P* \\< 0.001, difference versus 20 \u03bcM CPX group. **(C)** RD cells were pretreated with 5 \u03bcM of CQ for 1 h, and then incubated with 0--20 \u03bcM of CPX for 48 h. The cells were then harvested and processed for Annexin V-FITC/PI staining and flow cytometry. Results are presented as mean \u00b1 SD (n=3). \\**P* \\< 0.05, difference versus 1 \u03bcM and 5 \u03bcM CPX group, respectively, \\*\\**P* \\< 0.01, difference versus 20 \u03bcM CPX group.](oncotarget-05-10140-g005){#F5}\n\nDISCUSSION {#s3}\n==========\n\nCPX is an off-patent drug and currently used for the treatment of skin and nail fungal infection \\[[@R1]\\]. Recently, CPX has also been found to possess anticancer properties, by inhibiting cell proliferation, inducing cell death, as well as inhibiting angiogenesis and lymphangiogenesis \\[[@R3], [@R4], [@R32], [@R33]\\]. Oral administration of CPX at a dose of 20--25 mg/kg/day for 24--28 days inhibits the growth of xenografted leukemia (MDAY-D2, K562 and OCI-AML2) and primary human acute myeloid leukemia (AML) cells, as well as breast cancer (MDA-MB231) cells by 65--75% compared with vehicle control, but does not display evidence of weight loss or gross organ toxicity in mice \\[[@R3], [@R4]\\]. Furthermore, a recent phase I clinical trial has demonstrated that oral administration of CPX at a dose of 40 mg/m^2^ once daily for 5 days is well tolerated in all patients without dose-limiting toxicity, displays persistent pharmacodynamic activity in the patients, and induces disease stabilization and/or hematologic improvement in 2/3 patients with advanced hematologic malignancies \\[[@R34]\\]. Taken together, these findings highlight that CPX, as a fungicide, is a very promising anticancer agent and may be repurposed for cancer therapy.\n\nThough CPX has been used as a fungicide for over 20 years, its antifungal mechanism is still not well understood. However, it has been proposed that CPX acts as an iron chelator, forming complexes with trivalent metal cations, such as Fe^3+^, and inhibiting iron-dependent enzymes, such as catalase and peroxidase, which catalyze degradation of toxic peroxides, resulting in oxidative toxicity in fungi \\[[@R35]\\]. Recently, it has been found that CPX induces cell death by chelating intracellular iron and inhibiting the iron-dependent enzyme ribonucleotide reductase in human leukemia and myeloma cells \\[[@R4]\\]. CPX induces apoptosis by increasing caspase-3/7 activity and down-regulating of protein expressions of Bcl-xL and survivin in human rhabdomyosarcoma cells \\[[@R3]\\]. Furthermore, CPX inhibits Wnt/\u03b2-catenin pathway \\[[@R36]\\], the mammalian target of rapamycin (mTOR) activity \\[[@R37]\\], and the eukaryotic translation initiation factor 5A (eIF5A) function \\[[@R32], [@R38]\\]. It is unknown whether iron chelation is implicated in these effects of CPX. Clearly, the antitumor mechanism of CPX is complex and needs to be further investigated.\n\nIn the present study, for the first time, we found that CPX induced autophagy in human rhabdomyosarcoma RD and Rh30 cells. Although the precise mechanism of autophagy remains unclear, growing evidence has implicated that ROS play a critical role in controlling autophagy. ROS are highly reactive molecules formed by the incomplete one-electron reduction of oxygen, including oxygen ions and peroxides \\[[@R39]\\]. ROS form as a natural byproduct of the normal metabolism of oxygen and participate in cell signaling and homeostasis at low levels. However, under environmental stress, high levels of ROS can cause irreversible oxidative damage to cell structures \\[[@R39]\\]. Several stimuli that induce ROS generation can also induce autophagy, such as nutrient starvation, hypoxia, oxidative stress and some chemotherapeutic agents \\[[@R7]\\]. Alterations in ROS levels and autophagy play a crucial role in cancer initiation and progression, and both are recognized as the potential targets for cancer treatment \\[[@R31], [@R40]\\]. Our present data showed that CPX induced ROS, which were detectable in 4 h and increased by approximately 2-fold in 24 h of treatment. NAC, the ROS scavenger, remarkably attenuated CPX-induced GFP-LC3 puncta formation and LC3-II expression. Our results suggest that CPX triggers autophagy by induction of ROS.\n\nMultiple signaling molecules, such as MAPKs, mTOR, and class III PI3K, have been shown to regulate autophagy \\[[@R24], [@R41]\\]. The MAPKs, including ERK1/2, JNK and p38 MAPK, are a family of serine/threonine kinases that regulate a variety of cellular events such as proliferation and apoptosis \\[[@R42]\\]. It is well known that activation of JNK contributes to stress-induced apoptosis \\[[@R42]\\]. Recent studies have further revealed that activation of JNK is also associated with autophagy induction \\[[@R24]\\]. Endoplasmic reticulum (ER) stress induces autophagosome formation and accumulation by activation of inositol-requiring enzyme 1 (IRE1)-JNK pathway \\[[@R43]\\]. Bufalin induces autophagy via ROS induction and JNK activation \\[[@R44]\\]. Here we found that CPX was able to activate p38\u03b1, ERK1/2, and JNK1/2, but only activation of JNK pathway was responsible for CPX-induced autophagy. This is strongly supported by the findings that inhibition of JNK with SP600125 or ectopic expression of dominant negative c-Jun potently inhibited CPX-induced autophagy, whereas inhibition of ERK1/2 and p38\u03b1 using U0126 and doramapimod, respectively, failed to prevent CPX-induced autophagy. These results suggest that JNK pathway plays a critical role in CPX-induced autophagy. Our further study showed that pre-treatment with NAC prevented CPX-induced phosphorylation of c-Jun and autophagy, indicating that CPX-induced autophagy is dependent on ROS-activated JNK cascade. Further research is needed to address whether CPX induces autophagy also by targeting mTOR and class III PI3K.\n\nRecent studies have shown that several anticancer agents including tamoxifen, 5-fluorouraci and rapamycin can induce autophagy \\[[@R45]--[@R47]\\]. However, the role of autophagy in cancer therapy is complex and controversial. It has been reported that increased autophagy may be pro-apoptotic \\[[@R48]\\] or pro-survival \\[[@R45]\\], depending on experimental conditions (cell types and anticancer agents). Some efforts have been made to find under what conditions autophagy can be exploited for cancer therapy. Of note, inhibition of autophagy enhances vorinostat or tamoxifen-induced apoptosis \\[[@R46], [@R49]\\]. Here we observed that disruption of autophagy using CQ increased CPX-induced apoptosis, indicating that CPX-induced autophagy is a pro-survival mechanism in rhabdomyosarcoma cells.\n\nIn conclusion, we show that CPX induced autophagy in human rhabdomyosarcoma cells. CPX-induced autophagy was mediated by ROS induction and JNK activation. Disruption of autophagy using CQ enhanced CPX-induced cell death, indicating that CPX-induced autophagy is a pro-survival mechanism in the cells. Our findings suggest that combination of CPX and an autophagy inhibitor (e.g. CQ) may be a promising strategy for cancer therapy.\n\nMATERIAL AND METHODS {#s4}\n====================\n\nCell culture, agents and antibodies {#s4_1}\n-----------------------------------\n\nHuman rhabdomyosarcoma (Rh30 and RD) cell lines (gifts from Peter J. Houghton, Nationwide Children\\'s Hospital, Columbus, OH) were grown in antibiotic-free RPMI 1640 (Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals, Lawrenceville, GA). Rh30 cells stably expressing green fluorescent protein-light chain 3 (GFP-LC3) were generated by transfection of Rh30 cells with pEGFP-LC3 plasmid (Addgene, Cambridge MA) using Lipofectamine 2000 (Life Technologies, Grand Island, NY) and selection with G418 (500 \u03bcg/ml) (Mediatech). Cells were maintained in a humid incubator (37\u00b0C, 5% CO~2~). Ciclopirox olamine (CPX) (Sigma, St. Louis, MO) was dissolved in 100% ethanol to prepare a stock solution (100 mM), then aliquoted and stored at \u221220\u00b0C. Enhanced chemiluminescence solution was from Perkin-Elmer Life Science (Boston, MA, USA). CellTiter 96^\u00ae^ A~Queous~ One Solution Cell Proliferation Assay kit was from Promega (Madison, WI). Annexin V-FITC Apoptosis Detection Kit I was purchased from BD Biosciences (San Jose, CA). U0126, SP600125 and doramapimod were obtained form LC Laboratories (Woburn, MA). N-acetylcysteine (NAC), AO and CQ were purchased from Sigma. 5--6-chloromethyl-2,7-dichlorodihydrofluorescein diacetate (CM-H~2~DCFDA) was from Life Technologies. The following antibodies were used: p-JNK (Thr183/Tyr185), JNK, p-c-Jun (Ser63), c-Jun, ERK2, p-p38 (Thr180/Tyr182), p38\u03b1/\u03b2, p62, Beclin-1, MAP LC3\u03b2 (Santa Cruz Biotechnology, Santa Cruz, CA), phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) (Cell Signaling, Beverly, MA), \u03b2-tubulin, FLAG (Sigma), goat anti-rabbit IgG-horseradish peroxidase (HRP), goat anti-mouse IgG-HRP, and rabbit anti-goat IgG-HRP (Pierce, Rockford, IL).\n\nCell viability assay {#s4_2}\n--------------------\n\nCell viability was determined by MTS (3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) assay, according to the protocol of CellTiter 96^\u00ae^ A~Queous~ One Solution Cell Proliferation Assay kit (Promega). Briefly, 100 \u03bcl of cell suspensions were seeded into each well of 96-well plates (1\u00d710^4^ cells/well) and incubate overnight. The cells were then exposed to the tested compounds in triplicates for 72 h, followed by adding 20 \u03bcl of One Solution Reagent per well. After further incubation at 37\u00b0C for 1--2 h, cell viability was determined by measuring the optical density (OD) at 490 nm using a Wallac 1420 Multilabel Counter (PerkinElmer Life Sciences, Wellesley, MA).\n\nCell morphological analysis {#s4_3}\n---------------------------\n\nRh30 and RD cells were seeded in 6-well plates (2 \u00d7 10^5^ cells/well). The next day, the cells were treated with CPX (0, 5 and 20 \u03bcM). After incubation for 72 h, images were taken with an Olympus inverted phase-contrast microscope (Olympus Optical Co., Melville, NY) (200\u00d7) equipped with the Quick Imaging system.\n\nApoptosis assay {#s4_4}\n---------------\n\nCells were pre-incubated with or without 5 \u03bcM of CQ for 1 h, and then treated with or without CPX (1, 5 and 20 \u03bcM) for 48 h. Cells were then collected and stained using Annexin V-FITC Apoptosis Detection Kit I (BD Biosciences) according to the manufacturer\\'s instruction. In brief, cells were washed with cold PBS, and then resuspended in 100 \u03bcl of Annexin-V binding buffer, followed by incubation with FITC conjugated Annexin V and propidium iodide (PI) for 15 min at room temperature in the dark. Flow cytometry was performed using a FACS Calibur flow cytometer (Becton Dickinson, San Jose, CA). Cells treated with vehicle alone (100% ethanol) were used as a control.\n\nROS detection {#s4_5}\n-------------\n\nIntracellular level of ROS was measured by detecting the fluorescent intensity of oxidant-sensitive probe CM-H~2~DCFDA, which is taken up by cells, cleaved by esterases to DCFH and trapped intracellularly. Briefly, Rh30 and RD cells were seeded at a density of 1 \u00d7 10^4^ cells/well in 96-well plates. The next day, the cells were pre-incubated with or without NAC (5 mM) for 30 min, and then treated with or without CPX (0--20 \u03bcM) for 30 min, followed by loading with 10 \u03bcM of CM-H~2~DCFDA. At different time points (2, 4, 8 and 24 h), the fluorescent intensity was detected by excitation at 485 nm and emission at 535 nm using a Wallac 1420 Multilabel Counter (Perkin-Elmer Life Sciences, Wellesley, MA).\n\nGFP-LC3 puncta assay {#s4_6}\n--------------------\n\nRh30 cells stably expressing GFP-LC3 were seeded at a density of 2 \u00d7 10^5^ cells/well in 6-well plates. The next day, the cells were treated without or with 20 \u03bcM of CPX for different time (8 and 24 h), or with different concentrations of CPX (0--20 \u03bcM) for 24 h, and examined under a fluorescence microscope.\n\nWestern blot analysis {#s4_7}\n---------------------\n\nCells were seeded in 6-well plates at a density of 5 \u00d7 10^5^ cells/well. The next day, the cells were treated with CPX (0--20 \u03bcM) for 24 h, or with 20 \u03bcM CPX for 0--24 h. For experiments with NAC or MAPK inhibitors, cells were pre-incubated with or without NAC or MAPK inhibitors for 1 h, respectively. The cells were then treated with or without CPX (10 and 20 \u03bcM) for 24 h. Cell lysis and immunoblotting were performed as described previously \\[[@R3]\\]. \u03b2-tubulin served as a loading control.\n\nInfection of cells with recombinant adenovirus {#s4_8}\n----------------------------------------------\n\nRecombinant adenoviruses expressing FLAG-tagged dominant negative c-Jun (FLAG-\u0394169) (Ad-c-Jun-DN) and green fluorescence protein (GFP) (Ad-GFP) were described \\[[@R50]\\]. For experiments, RD cells were seeded in 6-well plates at a density of 5 \u00d7 10^5^ cells/well. The next day, the cells were infected with the Ad-c-Jun-DN or Ad-GFP (as a control) for 24 h at 5 of multiplicity of infection (MOI = 5). The cells were then treated with CPX (0--20 \u03bcM) for 24 h, followed by cell viability, morphology, and Western blot analysis. Expression of FLAG-tagged dominant negative c-Jun was confirmed by Western blotting with antibody to FLAG.\n\nStatistical analysis {#s4_9}\n--------------------\n\nResults were expressed as mean values \u00b1 standard deviation (mean \u00b1 SD). The data were analyzed by one-way analysis of variance (ANOVA) followed by post-hoc Dunnett\\'s *t*-test for multiple comparisons. A level of *P* \\< 0.05 was considered to be statistically significant.\n\n**Conflict of interest statement**\n\nThe authors declare no conflict of interest.\n\nThis work was in part supported by National Institutes of Health (CA115414; S. Huang), American Cancer Society (RSG-08-135-01-CNE; S. Huang), Carroll-Feist Predoctoral Fellowship Award (T. Shen), the West Light Foundation of The Chinese Academy of Sciences (H. Zhou), the National Natural Science Foundation of China (No. 81302807, H. Zhou), 100 Talents Program of the Chinese Academy of Sciences (Y. Li), Top Talents of Sciences and Technology of Yunnan Province (No. 2009CI120, Y. Li) and Basic Research Program of Yunnan Province (No. 2013FA047, Y. Li).\n"}
+{"text": "All relevant data are within the manuscript and the whole version of the study is available at the public repository of Sao Paulo State University: <https://repositorio.unesp.br/handle/11449/146728>.\n\nIntroduction {#sec001}\n============\n\nOver the past five decades, world fish production has grown twice as fast as human population and the consumption *per capita* increased from 9.9 kg in 1960 to 20.1 kg in 2014 \\[[@pone.0222626.ref001]\\]. In Brazil, pacu (*Piaractus mesopotamicus*) is one of the most common reared freshwater fish species \\[[@pone.0222626.ref002]\\]. Both growth and intensification of fish production has increased the incidence and severity of diseases, especially those of bacterial origin, such as aeromonosis caused by *Aeromonas* spp., which has a great significance in intensive rearing systems, due to its high mortality rates \\[[@pone.0222626.ref003]\\].\n\nAmong motile *Aeromonas* species, strains of *A*. *hydrophila* are considered the most virulent for teleosts \\[[@pone.0222626.ref004],[@pone.0222626.ref005]\\] and it is also a zoonosis that causes diarrhea and septicemia in humans \\[[@pone.0222626.ref006],[@pone.0222626.ref007]\\]. In fish, the infection by *Aeromonas* spp. causes rupture of small blood vessels, leading to cutaneous and fins hemorrhage, progressing to ulcerations with loss of epithelium, anemia, anorexia, lethargy, hemorrhagic sepsis and death \\[[@pone.0222626.ref008],[@pone.0222626.ref009]\\].\n\nThe deleterious effects caused by Gram-negative bacteria, such as *A*. *hydrophila*, are consequences of endotoxins, triggering sepsis or endotoxemia; complex syndromes that are defined by the presence of a systemic inflammatory response (SIRS) \\[[@pone.0222626.ref010],[@pone.0222626.ref011]\\]. The consequence of this exacerbated and uncontrolled inflammatory response is the appearance of the lesions in multiple organs and high mortality \\[[@pone.0222626.ref012]\\]. Internal hemorrhagic injuries associated to high mortality are common manifestations in acute and super-acute aeromonosis caused by *A*. *hydrophila* in fish \\[[@pone.0222626.ref013]\\]. Similarly, Claudiano et al. \\[[@pone.0222626.ref011]\\] observed the firsts signs of *Aeromonas* infection, wich were petechiaes and suffusions, 9 hours post-inoculation of *A*. *hydrophila* in pacus, showing the importance of understanding aeromonosis.\n\nBased on the aforementioned, this study aimed to characterize by morphological, ultrastructural and microbiological studies, the early stages of the septic process in different tissues of pacus experimentally infected with *A*. *hydrophila*.\n\nMaterial and methods {#sec002}\n====================\n\nFish and maintenance conditions {#sec003}\n-------------------------------\n\nA total of 50 healthy pacus (*P*. *mesopotamicus*) (250 \u00b1 99.3 g and 15 \u00b1 2 cm), originating from the same spawning of fish rared at Fish Repoduction Laboratory of Aquaculture Center of Unesp (Caunesp), were randomly distributed in five fiber tanks of 250 L (*n* = 10 fish/tank), supplied with chlorine-free running water from spring, at a flow of one liter per minute and constant aeration. Fish were acclimated for two weeks before the experiment. During this period, fish were fed to apparent satiety twice a day. Tanks were siphoned once a week and water quality was monitored daily using multiparameters probes (YSI\u00ae model 55 and model 63---YSI, Yellow Springs, OH, USA). Water quality parameters were kept as follows: dissolved oxygen = 5.1 \u00b1 0.6 mg.L^-1^; temperature = 29.47 \u00b1 1.58\u00b0C; pH = 7.66 \u00b1 0.36 and electrical conductivity = 117.96 \u00b1 6.12 \u03bcScm^-1^. Parameters remaining within the adequate ranges for this species \\[[@pone.0222626.ref014],[@pone.0222626.ref015]\\]. Ethical protocol for this study was approved by Ethics committee (CEUA-UNESP) under protocol number (01471/15) in accordance with guidelines for care and use of laboratory animals of Sao Paulo State University, Jaboticabal, Brazil.\n\nBacterial strain and preparation of challenge suspension {#sec004}\n--------------------------------------------------------\n\n*Aeromonas hydrophila* strain isolated from skin lesion of naturally infected pacus during an outbreak was used in this study. For genetic characterization, the bacterial mass originated from the culture of pure colonies underwent a DNA extraction process, according to the manufacturer's methodology (\"Genomic DNA Purification Kit---Wizard\u00ae\"). DNA concentration was 1690.9 ng/\u03bcL and absorbance ratio 260/280 and 260/230, varying between 2.02 and 2.04. Subsequently obtaining the DNA, ribosomal gene 16S rRNA was amplified according to Sarcar et al. \\[[@pone.0222626.ref016]\\]. Sequences were analyzed by the BLAST algorithm ([http://www.ncbi.nlm.nih.gov](http://www.ncbi.nlm.nih.gov/)), which presented 99% similarly with *A*. *hydrophila* (Accession number: CP007518-2). The challenge strain was grown on tryptic soy agar (TSA) (Kasvi, S\u00e3o Jos\u00e9 do Pinhais, PR, Brazil) and incubated at 28\u00b0C for 24 h. Bacterial suspensions were prepared by transferring a single colony of *A*. *hydrophila* to Falcon tubes with 50 mL of tryptic soy broth (TSB) (Kasvi, Brazil) and reincubated at 28\u00b0C in a bacteriological incubator (New Brunswick Scientific, New Jersey, NJ, USA). Twenty-four hours after incubation bacterial suspensions were centrifuged (Sorvall Legend Mach 1.6R, Germany) at 1792 *g*/10 min/4\u00b0C. Bacterial mass was resuspended in sterile 0.65% NaCl solution. Bacterial concentration was estimated by spectrophotometry and confirmed by determining the colony-forming units (CFU) of the original culture (serial dilutions, plating, and colony counting), reaching a concentration of 1.78 x 10^9^ CFU/mL.\n\nExperimental design and sepsis induction {#sec005}\n----------------------------------------\n\nAfter acclimation period, 50 pacus were randomly distributed into five groups for biological samples obtainment, according to time in hours post-inoculation (hpi): control group or 0 hpi; group 1 hpi; group 3 hpi; group 6 hpi and group 9 hpi (*n* = 10). Control group fish were injected by intraperitonal route with 0.5 mL of sterile 0.65% NaCl solution, while fish from the other groups were inoculated by the same route with 0.5 mL of *A*. *hydrophila* solution (1.78 x 10^9^ CFU/mL) \\[[@pone.0222626.ref011]\\]. Specimens were maintained in fasting during the 24 h previous challenge. Before the procedures, all fish were anesthetized by immersion in benzocaine solution (0.1 gL^-1^) \\[[@pone.0222626.ref017]\\].\n\nSampling of biological material {#sec006}\n-------------------------------\n\nAfter 0, 1, 3, 6 and 9 hpi, fish were anesthetized for blood sampling collection by puncture of the caudal vein using with sterile needles and syringes. Blood was immediately used for microbiological examination. After blood collection, anesthetized fish were euthanized by pithing and submitted to necropsy. Fragments of the spleen, heart, brain, liver and anterior kidney were aseptically sampled after the procedures of antisepsis and necropsy using sterile surgical instruments (e.g., scissors, tweezers, blades), aiming to perform microbiological, histopathological and ultrastructural examinations. Additionally, intestine, gills and pancreas were sampled for the histopathological exam.\n\nClinical trial and macroscopic alterations {#sec007}\n------------------------------------------\n\nAnimals were evaluated continually throughout the experiment, to verify possible macroscopic alterations and/or clinical behavior.\n\nMicrobiological analyzes {#sec008}\n------------------------\n\nIn all sampling times, an aseptically removed fragment of each sampled organ and fresh blood were immediately streaked with a platinum loop in TSA plates, supplemented with sodium ampicillin to a final concentration of 10 mgL^-1^, and incubated for 24 h at 28\u00b0C, in aerophilic atmosphere \\[[@pone.0222626.ref018]\\]. A second tissue fragment was used to characterize the bacterial population. Tissues were weighed (wet weight) and triturated in sterilized micro-tubes, filled with sterile saline phosphate buffer (0.1 M, pH 7.0) at the proportion 1:10 and serially diluted up to 10^\u22124^. By means of the \"Spread-plate\" method, 50 \u03bcL of each dilution (10^\u22121^, 10^\u22122^, 10^\u22123^ and 10^\u22124^) and the fresh blood were spread on TSA containing 10 mgL^-1^ of sodium ampicillin. Plates were incubated for 24 h, at 28\u00b0C. Then, colony counting was made in an electronic counter. Random samples of the bacterial colonies were confirmed as *A*. *hydrophila* by Polymerase Chain Reaction (PCR).\n\nHistopathological examination {#sec009}\n-----------------------------\n\nTissue samples were fixed in 3.5% neutral buffered formalin for 24 h, and then preserved in alcohol 70%. Then, samples were placed in plastic cassettes and processed by gradual dehydration in 70--100% alcohol, clearing in xylene, and embedding in paraffin wax. Five-micron thick sections were cut using a microtome (Slee, Mainz, Germany), and then stained with hematoxylin and eosin (H&E) \\[[@pone.0222626.ref019]\\]. Slides were examined with a light microscope (Olympus BX51 with DP72 digital camera and CellSens Standard 1.5 software package) (Olympus Corporation, Tokyo, Japan).\n\nUltrastructural examination {#sec010}\n---------------------------\n\nTissue fragments of spleen, heart, brain, liver and head kidney were fixed in Karnovsky's solution (2% formaldehyde plus 2.5% glutaraldehyde in phosphate buffer 0.1 M, pH 7.0) for 24 h at 4 \u00baC, bathed three times in phosphate buffer 0.1 M, pH 7.0, post-fixed in 1% osmium tetroxide in sodium cacodylate buffer (0.1 M, pH 7.2--7.4) for 1 h, and then processed for TEM \\[[@pone.0222626.ref020]\\]. The sections were examined using a JEOL-100CXII electron microscope (Jeol, Peabody, MA, USA).\n\nStatistical analyzes {#sec011}\n--------------------\n\nData from microbiological analyzes were tested for normality (Kolmogorov-Smirnov, Anderson-Darling, Shapiro-Wilk and Watson) and then submitted to a variance analysis, and Dunn test was applied in order to verify differences within the medians (p \\< 0.05).\n\nResults {#sec012}\n=======\n\nClinical trial and macroscopic alterations {#sec013}\n------------------------------------------\n\nNo behavioral alterations were observed in any of the groups. Macroscopic lesions observed during necropsies were evident only at 6 and 9 hpi, with the last being characterized by a greater severity of lesions. Petechial cutaneous hemorrhages were evident, especially close to the inoculation area ([Fig 1B](#pone.0222626.g001){ref-type=\"fig\"}), and in fins and operculum ([Fig 1C](#pone.0222626.g001){ref-type=\"fig\"}). Gills congestion ([Fig 1E](#pone.0222626.g001){ref-type=\"fig\"}) and the presence of blood in the anterior chamber of the ocular globe (hyphema) were also observed ([Fig 1F](#pone.0222626.g001){ref-type=\"fig\"}). The main alterations identified in cavities were petechial hemorrhage at the coelomic and visceral walls of the organs ([Fig 2B](#pone.0222626.g002){ref-type=\"fig\"}); congestion of the coelomic wall vessels ([Fig 2B](#pone.0222626.g002){ref-type=\"fig\"}); presence of serum-sanguineous liquid in the coelom ([Fig 2B--2D](#pone.0222626.g002){ref-type=\"fig\"} and [2E](#pone.0222626.g002){ref-type=\"fig\"}); hepatomegaly and hepatic congestion ([Fig 2D](#pone.0222626.g002){ref-type=\"fig\"}); splenomegaly and splenic congestion ([Fig 2E](#pone.0222626.g002){ref-type=\"fig\"}); and hemorrhagic enteritis ([Fig 2F](#pone.0222626.g002){ref-type=\"fig\"}).\n\n![Macroscopic alterations of *Piaractus mesopotamicus* challenged with *Aeromonas hydrophila*.\\\n(A) Control (no external changes). (B) Cutaneous haemorrhage (arrows). (C) Hemorrhage at fins (arrow) and operculum (asterisk). (D) Normal gills (arrow). (E) Gill congestion (arrow). (F) Hyphema (arrow). Bars A and B: 4 cm, bars C-F: 1 cm.](pone.0222626.g001){#pone.0222626.g001}\n\n![Macroscopic alterations of *Piaractus mesopotamicus* challenged *with Aeromonas hydrophila*.\\\n(A) Normal celoma wall. (B) Petechial haemorrhage on the coelomic and visceral walls of the organs (arrows), vessel congestion (asterisk) and presence of serum-sanguineous liquid in coelom (arrowhead). (C) Liver (asterisk), spleen (arrowhead) and bowel (arrow) normal. (D) Hepatic congestion and hepatomegaly (arrow), presence of serum-sanguineous liquid in the coelom (arrowhead). (E) Splenic congestion and splenomegaly (arrow), presence of serum-sanguineous liquid in coelom (arrowhead). (F) Hemorrhagic enteritis (white arrow). Bars: 1 cm.](pone.0222626.g002){#pone.0222626.g002}\n\nMicrobiological analyzes {#sec014}\n------------------------\n\nIsolation frequency results were expressed as percentage, characterized by the isolation of pure colonies of *A*. *hydrophila* in different organs and blood after the pre-established experimental times. The data has shown that only at 9 hpi 100% of animals inoculated with the bacteria obtained positive results for *A*. *hydrophila* re-isolation in all analyzed tissues ([Fig 3](#pone.0222626.g003){ref-type=\"fig\"}). However, it is noteworthy that at 1 hpi 90% and 100% of animals showed positive results in bacterial recovery from spleen and kidney samples, respectively ([Fig 3](#pone.0222626.g003){ref-type=\"fig\"}).\n\n![Isolation frequency of *Aeromonas hydrophila* in tissues of infected *Piaractus mesopotamicus*.\\\nVertical columns express the percentages of positive isolation for *A*. *hydrophila* (n = 10) at different times after challenged. hpi = hours post inoculation.](pone.0222626.g003){#pone.0222626.g003}\n\nAn ascendant bacterial growth was verified in samples of all organs and blood of infected fish throughout time. The greatest amounts of bacteria were verified at 9 and 6 hpi (p \\< 0.05) compared with the control group (0 hpi), which did not present bacterial growth, in all examined organs and blood ([Fig 4](#pone.0222626.g004){ref-type=\"fig\"}). Moreover, at 3 hpi the samples of spleen also displayed significant differences from the control group ([Fig 4](#pone.0222626.g004){ref-type=\"fig\"}). It must be emphasized that the tissues with larger bacterial quantities in all experimental times were spleen, kidney and liver.\n\n![Bacterial population in the organs and blood of *Piaractus mesopotamicus* challenged with *Aeromonas hydrophila*.\\\nBacterial count at 0 (control), 1, 3, 6 and 9 hours post-inoculation (hpi) in each organ and blood is expressed in log of colony forming units per gram (CFU/g). The corresponding amount of bacteria for each group represents the median value and the bars its respective range (n = 10). Asterisks represent significant difference (p \\< 0.05) between the analyzed times and control group (Dunn\\'s test 5%).](pone.0222626.g004){#pone.0222626.g004}\n\nMacroscopic alterations and microbiological exams showed that the isolated strain of *A*. *hydrophila* from diseased pacus was capable of inducing aeromonosis in healthy fish, which presented characteristic lesions of the disease and presented a positive re-isolation of this bacterial strain in the afflicted tissues of these animals. The strain was confirmed by PCR.\n\nHistopathological examination {#sec015}\n-----------------------------\n\nIn general, lesions were observed at 6 and 9 hpi in all evaluated organs, being more frequent and intense in the spleen, liver and kidney. The 1 and 3 hpi groups presented no lesions in most of the organs or rare and discrete histopathological alterations, which were characterized by discrete congestion and cellular edema in some organs (e.g., liver, kidney, and gill). Control group was used as a parameter to compare the lesions to the other groups, since it did not show significant histopathological alterations in all organs.\n\nIn the heart, bacterial colonies were observed adhered to the pericardium, as well as the presence of leukocyte infiltrate surrounding these areas ([Fig 5C](#pone.0222626.g005){ref-type=\"fig\"}); a discrete congestion of pericardium vessels; thrombocyte agglomerates between the ventricle and pericardium; necrosis of ventricle cardiomyocytes and leukocyte infiltrate ([Fig 5B](#pone.0222626.g005){ref-type=\"fig\"}). The splenic tissue showed extensive congestion areas ([Fig 5E](#pone.0222626.g005){ref-type=\"fig\"}), cytoplasmic vacuolization and cellular edema in the sub-capsular region of the organ. Bacterial colonies were also observed adhered to the splenic capsule, as well as the presence of leukocyte infiltrate ([Fig 5F](#pone.0222626.g005){ref-type=\"fig\"}).\n\n![Photomicrograph of control and infected *Piaractus mesopotamicus* heart and spleen.\\\n(A) Normal cardiac tissue. (B) Necrosis of ventricular cardiomyocytes and aggregation of inflammatory cells (asterisk). (C) Bacterial colonies adhered to the pericardium and presence of inflammatory cells infiltrates surrounding the area (arrow). (D) Normal splenic tissue. (E) Congestion of spleen vessels (arrows). (F) Bacterial colonies adhered to the splenic capsule (asterisk) and presence of leukocyte infiltrates (arrows). Bars: 20 \u03bcm, H&E.](pone.0222626.g005){#pone.0222626.g005}\n\nThe pancreatic tissue was found in the mesentery of pyloric sacks, in a capsular region of the spleen and liver. We observed extensive areas of cellular death with disorganization of the tissue's architecture, with the main alterations being the congestion of large vessels ([Fig 6B](#pone.0222626.g006){ref-type=\"fig\"}), hemorrhage, extensive areas of necrosis with leukocyte infiltrate ([Fig 6C](#pone.0222626.g006){ref-type=\"fig\"}) and the presence of bacterial colonies ([Fig 6D](#pone.0222626.g006){ref-type=\"fig\"}).\n\n![Photomicrograph of control and infected *Piaractus mesopotamicus* exocrine pancreas.\\\n(A) Normal exocrine pancreatic tissue. (B) Congestion of large vessels (asterisk). (C) Pancreatic necrosis (asterisks) with dilated vessel and presence of leukocytes inside and surrounding tissue (arrows). (D) Bacterial colonies (arrow) and adjacent pancreatic necrosis (asterisk). Bars A-C: 20 \u03bcm and bar D: 10 \u03bcm, H&E.](pone.0222626.g006){#pone.0222626.g006}\n\nIn the kidney, larger quantities of melanomacrophages were found ([Fig 7B](#pone.0222626.g007){ref-type=\"fig\"}), as well as hemorrhage ([Fig 7C](#pone.0222626.g007){ref-type=\"fig\"}), necrosis with karyolysis and loss of the cytoplasmic delimitation of cells of the renal tubules ([Fig 7D](#pone.0222626.g007){ref-type=\"fig\"}), congestion of large vessels and presence of bacterial colonies surrounded by leukocytes. Discrete capillary congestions were also verified in the encephalon ([Fig 7F](#pone.0222626.g007){ref-type=\"fig\"}).\n\n![Photomicrograph of control and infected *Piaractus mesopotamicus* kidney and encephalon.\\\n(A) Normal organization of renal tubules. (B) Abundant melanomacrophages accumulation between renal tubules (arrows). (C) Interstitial hemorrhage in renal tissue (asterisk). (D) Necrosis with karyolysis of tubular cell's nucleus (arrows) and loss of cytoplasmic delimitation between renal tubule epithelial cells (asterisk). (E) Capillary of normal nervous tissue without congestion (arrow). (F) Discrete capillary congestion in nervous tissue (arrow). Bars A and D: 10 \u03bcm, bars B, C, E and F: 20 \u03bcm, H&E.](pone.0222626.g007){#pone.0222626.g007}\n\nSinusoidal capillaries and large vessels congestion was verified in the liver ([Fig 8B](#pone.0222626.g008){ref-type=\"fig\"}), as well as hemorrhage ([Fig 8C](#pone.0222626.g008){ref-type=\"fig\"}), cellular edema in hepatocytes; disorganization of the hepatic tissue's architecture, especially below the capsule and in perivascular regions ([Fig 8D](#pone.0222626.g008){ref-type=\"fig\"}), presence of bacterial colonies adhered to the hepatic capsule with leukocyte infiltrates and hepatocyte necrosis ([Fig 8F](#pone.0222626.g008){ref-type=\"fig\"}).\n\n![Photomicrograph of control and infected *Piaractus mesopotamicus* liver.\\\n(A) Control liver tissue. (B) Congestion of hepatic sinusoids (arrows). (C) Hepatic hemorrhage (asterisk). (D) Hepatic necrosis and disorganization of hepatic tissue architecture in perivascular region (arrows). (E) Normal hepatic capsule (arrow). (F) Bacterial colonies adhered to the hepatic capsule (arrow) with leukocyte infiltrate and hepatocyte necrosis (asterisk). Bars A and B: 20 \u200b\u200b\u03bcm, bar C: 50 \u03bcm, bar D: 100 \u03bcm and bars E and F: 10 \u03bcm, H&E.](pone.0222626.g008){#pone.0222626.g008}\n\nInfected fish presented congestion of large vessels of gill filaments and extensive areas of epithelial cells detachment at the base of interlamellar spaces ([Fig 9B](#pone.0222626.g009){ref-type=\"fig\"}). Other findings in gills were congestion ([Fig 9C](#pone.0222626.g009){ref-type=\"fig\"}), interlayer hyperplasia and subendothelial edema ([Fig 9D](#pone.0222626.g009){ref-type=\"fig\"}) of secondary lamellae. The intestines presented villi and intestinal mucous necrosis ([Fig 9F](#pone.0222626.g009){ref-type=\"fig\"}) and presence of bacterial colonies adhered to both serum and muscular layers.\n\n![Photomicrograph of control and infected *Piaractus mesopotamicus* gill and intestine.\\\n(A) Gill filaments and normal secondary lamellae tissue. (B) Detachment of epithelial cells from the lamella base that are thinned (arrows) and congestion of a large vessel (asterisk). (C) Congestion of secondary lamellae (arrows). (D) Secondary lamella edema (arrow). (E) Normal villi and intestinal mucosa (arrow). (F) Necrosis of the villi and intestinal mucosa (arrow). Bars A-C: 20 \u03bcm, bar D: 50 \u03bcm and bars E and F: 200 \u03bcm, H&E.](pone.0222626.g009){#pone.0222626.g009}\n\nUltrastructural examination {#sec016}\n---------------------------\n\nExtensive areas of cellular death were identified in all examined tissues, especially at 6 and 9 hpi, in which the findings were more severe, in comparison to groups 1 hpi, 3 hpi and control. The most common ultrastructural alterations observed at the present study were karyolysis or chromatolysis ([Fig 10B](#pone.0222626.g010){ref-type=\"fig\"}); condensation of chromatin or pyknosis ([Fig 10C](#pone.0222626.g010){ref-type=\"fig\"}); fragmentation of chromatin or karyorrhexis ([Fig 10D](#pone.0222626.g010){ref-type=\"fig\"}); cytoplasmic and organelle membranes integrity loss and organelle dissolution ([Fig 11B](#pone.0222626.g011){ref-type=\"fig\"}); dilation and detachment of ribosomes from the endoplasmic reticulum (ER) ([Fig 11D](#pone.0222626.g011){ref-type=\"fig\"}); edema of ER, nuclear membrane and mitochondria ([Fig 12](#pone.0222626.g012){ref-type=\"fig\"}); presence of erythrocytes out of vessels, indicating hemorrhage and presence of bacteria inside dead and disorganized cells ([Fig 13](#pone.0222626.g013){ref-type=\"fig\"}); and phagocytosis of the products of erythrocyte degradation by melanomacrophages ([Fig 14](#pone.0222626.g014){ref-type=\"fig\"}).\n\n![Transmission electron microscopy images of control and infected spleen cells of *Piaractus mesopotamicus*.\\\n(A) Normal nucleus (N) in a lymphocyte. (B) Nuclei in karyolysis (N) in a lymphocyte. (C) Pynotic nucleus (N) in a lymphocyte. (D) Nucleus in karyorrhexis (N) in a lymphocyte. Bars: 1 \u03bcm.](pone.0222626.g010){#pone.0222626.g010}\n\n![Transmission electron microscopy images of control and infected kidney and heart cells of *Piaractus mesopotamicus*.\\\n(A) Normal renal tubule cell showing mitochondria (MT), endoplasmic reticulum (ER) and nucleus (N). (B) Infected renal tubule cell with loss of membrane integrity and dissolution of mitochondria (MT) and endoplasmic reticulum (ER). (C) Normal cardiomyocytes showing endoplasmic reticulum (ER) with adhered ribosomes (arrows). (D) Dilatation and detachment of ribosomes (arrows) from endoplasmic reticulum (ER) of infected cardiomyocytes. Bars: 1 \u03bcm.](pone.0222626.g011){#pone.0222626.g011}\n\n![Transmission electron microscopy image of infected kidney cells of *Piaractus mesopotamicus*.\\\nA renal tubule cell is showed in the upper right portion of the image, with normal nucleus (N), nuclear membrane (NM) and endoplasmic reticulum (ER). Separated by the cytoplasmic membrane (CM), another renal tubule cell located in the lower left portion of the image, shows endoplasmic reticulum edema (SER), nuclear membrane (SNM) and mitochondria (SMT). Bar: 1 \u03bcm.](pone.0222626.g012){#pone.0222626.g012}\n\n![Transmission electron microscopy image of infected splenic hemopoietic precursor cells of *Piaractus mesopotamicus*.\\\nBacteria presence (arrows) inside a compromised cytoplasm of a dead cell or in the process of death (DC). A normal cytoplasm of a living cell (LC) is shown for comparison. Bar: 1 \u03bcm.](pone.0222626.g013){#pone.0222626.g013}\n\n![Transmission electron microscopy image of infected kidney cells of *Piaractus mesopotamicus*.\\\nPhagocytosis process of the products of erythrocyte degradation, note the presence of a melanomacrophage (arrow) probably engulfing the products of a degraded erythrocyte (DE) and forming siderosomes (arrowhead) inside its cytoplasm, it also can be seen non-degraded erythrocytes (NDE) and renal tubule cells (RTC). Bar: 1 \u03bcm.](pone.0222626.g014){#pone.0222626.g014}\n\nDiscussion {#sec017}\n==========\n\nThis study showed the kinetics of histopathological lesions in the acute stage of experimental septicemic aeromonosis in pacus, revealing that the appearance of macro and microscopic lesions initiates at 6 hours post-inoculation (hpi), with more severe results at 9 hpi. The development and severity of lesions are related to bacteria pathogenicity, seen that *Aeromonas* are pathogens that have several virulence factors, thus the infection is complex and multifactorial \\[[@pone.0222626.ref021]--[@pone.0222626.ref023]\\].\n\nAltered behavioral manifestations were not shown by fish throughout the clinical observation period up to and including 9 hpi, which may be related to the absence of severe injuries in its nervous system; neurological disturbances such as erratic swimming are one of the main behavioral alterations in fish affected by aeromonosis \\[[@pone.0222626.ref024],[@pone.0222626.ref025]\\]. Another hypothesis that may explain the absence of behavioral alterations is related to the incubation period and neurological signs in this disease, in other words only 9 hpi may not be enough for such manifestations to appear in pacus. It was showed that in the acute phase of *A*. *salmonicida* infection in zebrafish, the signs related to swimming disturbances occurred only 12 hours after the experimental infection \\[[@pone.0222626.ref026]\\].\n\nConcerning macroscopic examinations, inoculated fish presented petechiae hemorrhage in the body, fins and operculum surfaces, gills congestion and ocular hemorrhage. Similarly, studies of *Aeromonas* spp. infection in cyprinids, demonstrated external macroscopic findings, such as skin darkening and hemorrhage in the body surface, fins, gills and eyes \\[[@pone.0222626.ref027],[@pone.0222626.ref028]\\].\n\nThe findings regarding the necropsy performed using fish evidenced hemorrhages in serous and organs, a discrete accumulation of serosanguinolent liquid in the cavity, congestion and a slight increase of the organs and superficial vessels volume, corroborating the results described in the literature for others species of fish affected by aeromonosis \\[[@pone.0222626.ref029]\\]. Similarly, Fichi et al. \\[[@pone.0222626.ref028]\\] observed the presence of hemorrhage in internal organs (heart, kidney, swim bladder and ovary) of crucian carps (*Carassius carassius*) infected by *A*. *sobria*.\n\nMacroscopic hemorrhagic alterations possibly occurred due to the acute inflammatory response to the infection. Leukocytes release several vasoactive substances that cause vasodilation and increased vascular permeability, leading to hemorrhages \\[[@pone.0222626.ref030]--[@pone.0222626.ref032]\\].\n\nMicrobiological examination revealed that, even in low percentages, bacteria were detected at the first hour post-inoculation (1 hpi), reaching 100% of positive isolations at 9 hpi in all organs and blood of infected fish. This rapid dissemination of *A*. *hydrophila* through the tissues supposedly occurred as a function of the motile capacity of these bacteria, which moves with the aid of flagella.\n\nOne of the most important virulence factors of *Aeromonas* spp. in the initial colonization of tissues relates to the presence of polar flagella, which allows the organism to move in liquid environments, and lateral flagella, aiding in viscous surfaces, giving a greater invasive capacity for these bacteria \\[[@pone.0222626.ref033]\\]. In studies of *Tenacibaculum maritimum* infection, a motile microorganism that does not have flagella and moves by sliding, Fa\u00edlde et al. \\[[@pone.0222626.ref034]\\] detected its presence in internal organs of infected fish (*Psetta maxima* L.) only six hours after experimental inoculation. By comparing these data with our findings, it can be inferred that the presence of flagella may have conferred a greater precocity in the initial colonization of pacus tissues, seen that *A*. *hydrophila* was already found in the first hour after infection.\n\nThe upward curve of bacterial quantification in organs and blood, emphasizing the greatest values for spleen, liver and kidney in all experimental times. This fact may indicate that these organs are most susceptible to infection by *A*. *hydrophila* in pacus, thus facilitating the diagnosis of this pathology, even in early cases through the preventive diagnosis of outbreaks. In studies of acute experimental streptococcosis, Abdullah et al. \\[[@pone.0222626.ref035]\\] isolated bacterial colonies from the encephalon, eyes and kidney, four hours after coelomatic inoculation of *Streptococcus agalactiae* in red tilapia (*Oreochromis* spp.), determining the encephalon as the target-organ of this microorganism in this fish.\n\nBacterial virulence factors are related to the capacity of invasion, replication and evasion of the host immune system, causing lesions during pathogenesis \\[[@pone.0222626.ref036]\\]. Farto et al. \\[[@pone.0222626.ref025]\\] studied the capacity of invasion and distribution of three strains of *A*. *salmonicida* subspecies *salmonicida*, two avirulent and one considered virulent, in organs of experimentally infected preached fish (*Psetta maxima* L.). The results indicated that the virulent strain was isolated in internal organs, especially liver and kidney, in the first 12 hpi, which remained present in all experimental times until death, after seven days. Conversely, avirulent strains were rarely isolated from internal organs, being eliminated four days after the challenge.\n\nHistopathological findings showed that *A*. *hydrophila*, when directly inoculated in the coelom of pacus, causes lesions in multiple organs. These lesions were observed at 6 hpi and were more severe at 9 hpi. As *Aeromonas* spp. possesses the capacity of adhering to tissue cells, the illness severity depends on the types of virulence factors involved and the immune *status* of the host, which leads to several degrees of injury on tissues \\[[@pone.0222626.ref037]--[@pone.0222626.ref039]\\].\n\nHistopathological lesions occurred likely due to the liberation of bacterial toxins and other virulence factors inherent to aeromonosis. This bacterium produces endotoxins and exotoxins, such as hemolysin and aerolysin, which cause the rupture of cellular membranes \\[[@pone.0222626.ref040],[@pone.0222626.ref041]\\], enterotoxins \\[[@pone.0222626.ref042]\\], dermonecrotic factors, proteases, phospholipases and DNAses that causes tissue damage and facilitate bacterial invasion and multiplication in the hosts' cells \\[[@pone.0222626.ref040],[@pone.0222626.ref042]\\].\n\nThe present results showed that, after the bacteria inoculation in the coelom, probably, there was colonization of serosa or capsular compartment of the organs in this cavity (e.g., liver, spleen, heart, kidney, intestine, and pancreas). Following that colonization, bacteria crossed the epithelial barriers and leaded to the infection and, consequently, inducing damage to the cells inside these organs. Concomitantly, some bacteria crossed mesothelium of peritoneum vessels and reached the blood circulation and colonized even distant organs (e.g., brain and gills). Farto et al. \\[[@pone.0222626.ref025]\\] observed the same colonization mechanisms of turbot tissues infected with *A*. *salmonicida* subsp. *salmonicida*.\n\nNo available reports of histopathological findings in the acute phase of aeromonosis were found until this time, however studies on advanced stages of septicemic diseases caused by pathogenic bacteria in fish showed similar histopathological alterations to the ones found in this study. Aguado-Urda et al. \\[[@pone.0222626.ref043]\\] observed necrosis in the renal tubular epithelium and hemorrhage in compact layers of the heart after experimental infection with *Lactococcus garvieae* in zebrafish. Chong et al. \\[[@pone.0222626.ref044]\\] found, in Australian eels inoculated with *Erysipelothrix rhusiopathiae*, moderate brain and hepatic vascular congestion. Carraschi et al. \\[[@pone.0222626.ref045]\\] reported sub-epithelial edema in secondary gill lamellae in pacus infected with *A*. *hydrophila*, and Avci et al. \\[[@pone.0222626.ref046]\\] found the same histopathological findings in rainbow trouts infected with *Lactococcus garvieae*.\n\nBacteria that causes sepsis are frequently observed in injured areas or associated to inflammatory processes in teleosts previously challenged or naturally infected \\[[@pone.0222626.ref043],[@pone.0222626.ref044],[@pone.0222626.ref047]\\], corroborating with the presence of the observed bacterial colonies in the lesions of this study. This occurs due to the structural components of *Aeromonas* spp., such as flagella, fimbria, membrane proteins and lipopolysaccharides (LPS), which allows the adherence of bacteria to the fish tissues \\[[@pone.0222626.ref043]\\].\n\nSimilar to our histopathological findings in kidney and liver, other studies of bacterial septicemia in fish reported that hybrid catfish (*Clarias macrocephalus* x *Clarias gariepinus*) infected with *Edwardsiella ictaluri* presented lymphocyte infiltrate in the renal glomeruli \\[[@pone.0222626.ref048]\\] and cyprinids (*Puntius sarana*) challenged with *A*. *hydrophila* presented increased number of melanomacrophages and congestion of central vessels in the liver \\[[@pone.0222626.ref027]\\].\n\nThe increased quantities of melanomacrophages in pacus infected with *A*. *hydrophila* is possibly related to the lysis of red blood cells, caused by bacterial toxins such as hemolysin and aerolysin, seen that these cells are responsible for the phagocytosis of the products of erythrocyte degradation, such as hemosiderin \\[[@pone.0222626.ref049]\\]. This is another indicative of the action of toxins in organs of pacu, throughout the acute stage of aeromonosis, since the increase of melanomacrophages relates to the protective response of tissues against damages caused by free radicals \\[[@pone.0222626.ref027]\\].\n\nIn this study, the observed lesions in transmission electron microscopy reinforce the histopathological observations, cellular death was verified, and organelle degeneration and a strict association between lesions and the presence of bacteria in infected organs. Likewise, eels (*Anguilla anguilla*) infected with *Vibrio vulnificus* manifested the presence of the bacteria in a close association with necrotic cells, which presented alterations in the cell membrane and organelle damage \\[[@pone.0222626.ref020]\\].\n\nIn a study concerning the resistance of *Streptococcus iniae* inside macrophage phagosomes of gilthead seabream (*Sparus aurata*) and red porgy (*Pagrus pagrus*), the authors demonstrated by means of ultrastructural analyzes, that these cells may be important sources of bacteria dissemination in the nervous system, since bacteria may cross the blood-brain barrier without any resistance \\[[@pone.0222626.ref050]\\].\n\nThe organs that presented more severe microscopic lesions were liver, spleen and kidney. These organs also presented the greatest bacterial quantifications, indicating that these are the target organs of the acute stage of septicemic aeromonosis in pacu. Monaghan et al. \\[[@pone.0222626.ref051]\\] determined through *in situ* hybridization, that both gills and intestines are target organs of the acute infection by herpesvirus in carps (*Cyprinus carpio*) after challenged by immersion. According to these authors, these organs were relevant as a gateway in the initial stage of the illness, reinforcing the importance of the determination of target organs in the understanding of infectious diseases.\n\nThe findings showed that this bacterium spreads in the first hpi through the pacus organs. Morphological and microbiological findings showed necrosis, degenerative processes, vascular alterations and an association between bacteria and lesions, especially at 6 and 9 hpi, demonstrating the action of the virulence factors of *A*. *hydrophila* in the pathophysiology of the acute stage of septicemic aeromonosis in pacu organs. Spleen, liver and kidneys were the most affected organs and are suggested as the target organs of the infection, which may assist the diagnosis of acute cases of this pathology in pacus. These findings help to improve the knowledge about the infection process model and may orientate and assist the diagnosis of acute cases of this pathology in pacus.\n\nThe authors would like to thank: Fl\u00e1vio Ruas de Moraes (*In memoriam*) for the mentorship and friendship;Microscopy laboratory at S\u00e3o Paulo University, campus Ribeir\u00e3o Preto, S\u00e3o Paulo, Brazil, for the technical assistance during TEM analyzes;Rog\u00e9rio Salvador, Marita Vedovelli Cardozo and Laura Satiko Okada Nakaghi for their great recommendations that considerably improved this paper.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nIn the mammalian ovary, folliculogenesis is critically regulated to provide fertilizable oocytes for the duration of a female's reproductive life. In vitro growth (IVG) of mammalian follicles could be a promising tool for use with assisted reproductive technologies (ART) to derive large quantities of fertilizable oocytes derived from immature follicles. Moreover, this system could provide a valuable research tool for further elucidation of the poorly understood molecular mechanisms of folliculogenesis. Therefore, establishing an in vitro follicle culture model that can accurately mimic the in vivo ovarian growth environment has been crucial. Until now many different approaches have been used with varying success, however, a single standard protocol has yet to be established, even in a model species such as the mouse \\[[@pone.0143114.ref001]\\]. The difficulties of in vitro follicle development originate in the unique features of the follicle. Individual follicles within the ovary feature a vascular system built up around a 3-D structure in which the oocyte is surrounded by follicular cells \\[[@pone.0143114.ref001]\\]. Growth of the oocyte is critically dependent on autocrine and paracrine bidirectional signaling between the oocyte and the surrounding follicular cells \\[[@pone.0143114.ref002]--[@pone.0143114.ref003]\\]. Due to the requirement for this crosstalk between the oocyte and follicular cells, maintenance of follicle architecture during the growth process is crucial for the correct acquisition of developmental competence of oocytes.\n\nIn general, ovarian tissue culture as a technology for in vitro follicle development has not been highly effective. While primordial follicle growth can be supported in this system, proper development of follicles beyond the pre-antral stage has been inhibited \\[[@pone.0143114.ref004]\\]. Instead of culturing the tissue itself, individual follicles have been isolated from the ovary and cultured on a dish/membrane to maintain their proper development (2-D follicle culture). Actually, the majority of early groundbreaking work on in vitro follicle culture was undertaken using conventional 2-D follicle culture systems. Eppig and Schroeder (1989) were able to obtain live offspring after in vitro culture of isolated mouse pre-antral follicles on collagen gel and the use of IVF to fertilize these IVM oocytes \\[[@pone.0143114.ref005]\\]. This same group further reported that primordial follicles can be cultured to complete maturation using a combination of ovarian tissue culture and 2-D follicle culture to generate developmentally competent oocytes \\[[@pone.0143114.ref006]\\]. In general, 2-D follicle culture systems have utilized a 2-D surface (dish /membrane) coated with extracellular matrix (ECM) proteins \\[[@pone.0143114.ref007]--[@pone.0143114.ref010]\\]. In mammalian ovaries, ECM proteins play key roles in structural support, cell aggregation and communication, cell survival, proliferation, and biochemical signal regulation \\[[@pone.0143114.ref009], [@pone.0143114.ref011]--[@pone.0143114.ref012]\\]. For example, laminin and type IV collagen have been observed in the basal lamina of follicles in several different species \\[[@pone.0143114.ref013]--[@pone.0143114.ref016]\\], suggesting addition of ECM proteins to 2-D follicle cultures may help mimic normal environmental signaling that occurs in vivo to promote follicle growth in 2-D culture systems. However, it has been recognized that 2-D follicle culture is not sufficient to sustain the normal architecture of follicles in vitro \\[[@pone.0143114.ref001], [@pone.0143114.ref004], [@pone.0143114.ref017]\\].\n\nIn recent years, further technical advancement has emerged in the form of 3-D follicle culture, which better sustains normal follicle structure in vitro. In 3-D follicle culture systems, isolated follicles are individually encapsulated in various types of substrates such as collagen, alginate, hyaluronan hydrogel, PEG, or Matrigel to maintain the normal 3-D architecture of the follicle in vitro \\[[@pone.0143114.ref001], [@pone.0143114.ref004], [@pone.0143114.ref017]\\]. To date, encapsulation with alginate gel has been the most widely applied system for 3-D follicle culture \\[[@pone.0143114.ref004], [@pone.0143114.ref017]\\]. For example, secondary---preantral follicles encapsulated with alginate gel showed growth in vitro \\[[@pone.0143114.ref018]--[@pone.0143114.ref020]\\], and resulted in the production of live offspring from oocytes grown under these 3-D follicle culture conditions \\[[@pone.0143114.ref021]\\]. However, 3-D follicle culture is typically considered a complex and time-consuming procedure. For example, the protocol for the culture of murine follicles in an alginate matrix typically involves the following steps--dissociation of the ovary to isolate follicles, selection of healthy follicles, placement of single follicles into each alginate droplet, immersion of alginate + follicle complexes in calcium chloride solution to crosslink the alginate, and culture of encapsulated follicles in individual wells of a 96-well plate \\[[@pone.0143114.ref019]--[@pone.0143114.ref023]\\].\n\nHere we describe a simple and easy 3-D ovarian tissue culture system for growth and maturation of follicles from mice. To create a more robust *in vivo*-like ovarian culture system, we utilized a solid Matrigel drop (1:1 dilution with culture medium) to support the proper 3-D structure of follicles in the ovarian tissue. An ovary was cut into 8 pieces and one piece of the tissue was placed on top of each Matrigel drop supported on a membrane. This procedure is much simpler than either the 2-D or 3-D isolated follicle culture systems described above. Because our goal was to generate fertilizable follicular oocytes in vitro, the developmental competence of the resulting oocytes was compared to that of oocytes derived from follicles cultured on membranes without Matrigel drop. We also evaluated the effect of activin A treatment on follicle growth in ovarian tissue cultured with or without Matrigel support. Activin A, a member of the transforming growth factor \u03b2 (TGF-\u03b2) superfamily, is normally generated by granulosa cells in primary to antral stage follicles \\[[@pone.0143114.ref024]\\] and regulates proliferation, differentiation, and steroidogenesis within the follicles \\[[@pone.0143114.ref025]\\]. Despite evidence from previous studies indicating that activin A treatment enhances preantral follicle development and oocyte growth in various species such as mouse \\[[@pone.0143114.ref026]--[@pone.0143114.ref029]\\], human \\[[@pone.0143114.ref030]\\] and sheep \\[[@pone.0143114.ref031]\\], it has not been clear how activin A treatment affects the quality of oocytes grown in vitro. Here we describe a new technology for IVG of follicles that yields high quality oocytes competent to yield live offspring after IVF, thus validating the utility of the oocytes produced by this simple and easy ovarian tissue culture method.\n\nMaterials and Methods {#sec002}\n=====================\n\nMice {#sec003}\n----\n\nB6D2F1 (C57BL/6 x DBA/2) female mice were mated with B6D2F1 males to obtain (B6D2F1 x B6D2F1) F2 female pups. Prepubertal ovaries were collected from these F2 females at 14 days post partum (14 dpp). For IVF, sperm were obtained by dissection of the cauda epididymides of adult B6D2 F1 male mice. CD-1 female mice were used as surrogate mothers after mating with vasectomized CD-1 males. All relevant experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Hawaii.\n\nMedia {#sec004}\n-----\n\nOvarian tissue culture was performed using in vitro growth (IVG) medium consisting of \u03b1MEM (Invitrogen, CA) supplemented with 5% FBS, 100 mIU/ml rhFSH (MERCK, NJ), 5 \u03bcg/ml insulin, 5 \u03bcg/ml transferrin, 5 ng/ml selenium (Sigma-Aldrich, MO). Activin A (Sigma-Aldrich, MO) was used as an IVG supplement (30 ng/ml). IVM medium consisted of \u03b1MEM supplemented with 5% FBS, 100 mIU/ml rhFSH and 5 ng/ml rhEGF (Promega, WI). TYH medium was used for IVF. After IVF, fertilized oocytes and embryos were cultured in mCZB medium.\n\nOvarian tissue culture {#sec005}\n----------------------\n\nOvaries at 14 dpp were cut into 8 pieces and subjected to organ culture as described below. As a control, ovarian tissue pieces were directly placed onto a floating membrane filter (0.4 \u03bcm HTTP; Millipore, MA) (4 pieces/membrane) on IVG medium and cultured with or without activin A treatment (Membrane culture). Matrigel is a soluble extract of basement membrane protein derived from the Engelbreth-Holm-Swarm tumor that forms a 3-D gel at 37\u00b0C and supports cell morphogenesis, differentiation, and tumor growth \\[[@pone.0143114.ref032]\\]. For our 3-D culture system, Matrigel (BD Biosciences, MA) was diluted with IVG medium (1:1) and 5 \u03bcl were placed on a floating membrane filter to make a small Matrigel drop (4 drops/membrane). Then, a single ovarian tissue piece was placed on top of each drop (1 piece/drop) and cultured with or without activin A treatment. Each ovarian tissue piece partially sank into the Martigel but the top surface of each piece retained access to surrounding air. We tested four different culture conditions: Control (C: Membrane/activin-), with activin A (A: Membrane/activin+), with Matrigel but no activin A (M: Matrigel/activin-), and with Matrigel and activin A (M+A: Matrigel/activin+). Ovarian tissues were cultured under each the four conditions for 10 days at 37\u00b0C, 5%CO~2~ in air. Half of the culture media was replaced every other day.\n\nFollicle and oocyte growth {#sec006}\n--------------------------\n\nOn Days 0, 6, and 10 of ovarian tissue culture, growing follicle diameter and oocyte diameter were measured. Under a stereomicroscope, the tissues were dissected to isolate well-advanced individual follicles. Follicles were visualized using an inverted microscope (Olympus, PA) and measured with an ocular micrometer calibrated at 400x magnification. The longest and shortest diameters of each follicle were averaged to yield the respective value. These follicles were then punctured to obtain germinal vesicle (GV) stage oocytes surrounded by granulosa cells. To measure oocyte size, surrounding granulosa cells were mechanically removed by gentle pipetting. Each oocyte diameter was measured as an average of the longest and shortest dimensions at 400X magnification.\n\nClassification of oocyte GV chromatin configurations {#sec007}\n----------------------------------------------------\n\nChromatin in the GV oocyte is initially decondensed with the nucleolus not surrounded by heterochromatin (Non Surrounded Nucleolus: NSN). During oocyte growth, the GV chromatin condenses into perinucleolar rings (Surrounded Nucleolus: SN) \\[[@pone.0143114.ref033]\\]. To determine the status of the chromatin in GVs, we isolated oocytes from well-advanced follicles on Days 0, 6, and 10 of ovarian tissue culture. After being freed from the surrounding granulosa cells, the oocytes were treated with Hoechst dye (1:1000) for 8 minutes at 37\u00b0C. Under a fluorescence microscope, the oocytes were categorized into three chromatin configurations: NSN, SN, and an intermediate stage between NSN and SN configurations (Int).\n\nHistological analysis {#sec008}\n---------------------\n\nOn Days 0, 6, and 10 of ovarian tissue culture, tissues were fixed overnight in Bouin's solution and subsequently treated with ascending concentrations of ethanol (70--100%) and xylene. Serial 5 \u03bcm paraffin sections were cut and stained with hematoxylin and eosin.\n\n*In vitro* maturation (IVM) {#sec009}\n---------------------------\n\nAfter 10 days of ovarian tissue culture, full-grown GV oocytes were subjected to IVM. Largely grown follicles were punctured to obtain GV oocytes surrounded by granulosa cell layers. These oocytes with granulosa cells or cumulus cells were cultured in IVM media for 15--17 hrs at 37\u00b0C, 5% CO~2~ in air. After culture, oocytes with first polar bodies were recognized as matured metaphase II (MII) oocytes.\n\n*In vitro* fertilization (IVF) {#sec010}\n------------------------------\n\nAfter IVM, in vitro matured oocytes with expanded granulosa cells/cumulus cells were subjected to IVF. Sperm mass collected from the caudal epididymides of adult males were incubated in TYH medium at 37\u00b0C, 5% CO~2~ in air. After 1 hr of pre-incubation, a small aliquot of sperm suspension was added to a TYH drop containing oocytes for IVF. Six hrs after IVF, oocytes were washed and transferred into mCZB media. Twenty-four hrs after IVF, embryo development to the 2-cell stage was assessed.\n\nEmbryo transfer (ET) {#sec011}\n--------------------\n\nEmbryos at the 2-cell stage were transferred into the oviducts of surrogate mothers that had been mated with vasectomized males the previous night (= Day 0.5 of gestation). On Day 19.5 of gestation, full-term pups were delivered by Cesarean section or naturally delivered. In vivo control pups were derived from natural mating between B6D2F1 female and male mice. Body weight and placenta weight of each pup were measured.\n\nStatistical analysis {#sec012}\n--------------------\n\nFollicle and oocyte sizes were analyzed by variance analysis (one-way multiple comparison ANOVA) and the differences were subjected to Tukey's test. Chi-square testing was used to analyze differences in oocyte chromatin configuration, IVM, IVF, and ET rates.\n\nResults {#sec013}\n=======\n\nFollicle and oocyte growth {#sec014}\n--------------------------\n\nIn the 14 dpp murine ovary, most advanced follicles are in the multilayered secondary follicle stage. We first focused on the developmental capacity of secondary follicles under our four different culture conditions. Ovarian tissue pieces were directly placed on a membrane (Membrane culture) and cultured with (A condition) or without (C condition) activin A ([Fig 1](#pone.0143114.g001){ref-type=\"fig\"}). Other ovarian tissue pieces were placed on top of a Matrigel drop (3-D culture) and cultured with (M+A condition) or without (M condition) activin A treatment ([Fig 1](#pone.0143114.g001){ref-type=\"fig\"}). All tissues were cultured for up to 10 days. We first examined follicle growth by measuring follicle diameters. Before culture (Day 0), the average diameter of the secondary follicles was 131 \u03bcm. During tissue culture, the secondary follicles gradually increased in size and expanded granulosa cell layers in all culture conditions. On Day 6 of culture, the average diameters of well-developed follicles at the preantral stage were similar among the four different culture conditions (271\\~290 \u03bcm) ([Fig 2A](#pone.0143114.g002){ref-type=\"fig\"} and [S1 Table](#pone.0143114.s001){ref-type=\"supplementary-material\"}). On Day 10 of culture, the ovarian tissues dramatically increased their size under all four conditions ([Fig 3A--3J](#pone.0143114.g003){ref-type=\"fig\"}). Each tissue piece contained 3--10 large follicles and average follicle sizes reached 299--356 \u03bcm, although the follicles in the C condition were larger compared to those grown in the other three conditions (*p*\\<0.05) ([Fig 2A](#pone.0143114.g002){ref-type=\"fig\"} and [S1 Table](#pone.0143114.s001){ref-type=\"supplementary-material\"}).\n\n![Ovarian tissue culture under 4 different culture conditions.\\\nA mouse ovary at 14 dpp was cut into 8 pieces and subjected to organ culture. Ovarian tissue pieces were directly placed onto a floating membrane filter on IVG medium and cultured with or without activin A (Membrane culture). For our 3-D culture system, Matrigel diluted with IVG medium (1:1) was placed on a membrane filter to make a Matrigel drop. A single ovarian tissue was placed on top of each drop and cultured with or without activin A treatment. The tissues were cultured for 10 days under four different culture conditions: Control (C: Membrane/activin-), with activin A (A: Membrane/activin+), with Matrigel but no activin A (M: Matrigel/activin-), and with Matrigel and activin A (M+A: Matrigel/activin+).](pone.0143114.g001){#pone.0143114.g001}\n\n![Follicle and oocyte growth in ovarian tissues cultured under 4 different conditions.\\\n**(a, b)** The ovarian tissues isolated from 14 dpp mice were cultured under 4 different conditions (C, A, M, M+A). On Days 0, 6 and 10 of culture, the growing follicle diameters (a) were measured. In all culture conditions, the advanced follicles increased their sizes on Day 6 (C vs A, M, M+A; *P*\\<0.01--0.05). On Day 10, the follicles reached 299--356 \u03bcm in all conditions (C vs A, M, M+A, A vs M, M+A, M vs M+A; *P*\\<0.05)., On Days 0, 6 and 10 of culture, oocytes were isolated from these follicles and measured their diameters (b). The oocytes increased their sizes in all conditions on Day 6 (C, A vs M, M+A; *P*\\<0.05). On Day 10, the isolated oocytes similarly reached full-size regardless of the conditions (A vs M, M+A; M vs M+A; *P*\\<0.05). All bar graphs show the mean \u00b1 SD.](pone.0143114.g002){#pone.0143114.g002}\n\n![Morphology of ovarian tissues cultured under 4 different conditions.\\\n**(a, b)** A whole murine ovary at 14 days old was cut into 8 pieces before culture (a). The most advanced follicles in the tissue was at the preantral stage (b). (**c-j)** Ovarian tissues after 10 days of culture in the C (c, d), A (e, f), M (g, h) and M+A (i, j) conditions. The antral cavity formation occurred only in the follicles under the 3-D Culture system (M, M+A) (arrows). The circles of dots represent the localization of oocytes. Regular bar = 500 \u03bcm. Bold bar = 100 \u03bcm.](pone.0143114.g003){#pone.0143114.g003}\n\nNext, we examined the growth of germinal vesicle (GV) stage oocytes within the developing follicles cultured under the four different conditions. On Day 6 of culture, oocyte sizes in the 3-D culture systems (M, M+A) were relatively larger than those maintained in the membrane culture systems (C, A) (*p*\\<0.05) ([Fig 2B](#pone.0143114.g002){ref-type=\"fig\"} and [S1 Table](#pone.0143114.s001){ref-type=\"supplementary-material\"}). On Day 10 of culture, GV oocytes in the larger developed follicles (average diameters of 299--356 um) shown in [Fig 2A](#pone.0143114.g002){ref-type=\"fig\"} reached full-size (73.5--76.3 um) ([Fig 2B](#pone.0143114.g002){ref-type=\"fig\"} and [S1 Table](#pone.0143114.s001){ref-type=\"supplementary-material\"}), matching those observed in full-grown GV oocytes in vivo isolated from antral follicles in the 24 dpp ovaries (74.0\u00b11.9 \u03bcm in diameter), regardless of the presence or absence of either the Matrigel drop or activin A supplement.\n\nHistological analysis of ovarian tissues {#sec015}\n----------------------------------------\n\nTo investigate morphological details of follicular development in culture, we prepared paraffin sections of the cultured ovarian tissues ([Fig 4](#pone.0143114.g004){ref-type=\"fig\"}). Before culture (Day 0), the 14 dpp in vivo ovary contained multilayered secondary follicles as the most advanced stage follicles ([Fig 4A](#pone.0143114.g004){ref-type=\"fig\"}). By 24 dpp, the most advanced follicles had formed antral cavities to develop to antral follicles ([Fig 4B](#pone.0143114.g004){ref-type=\"fig\"}). On Day 10 under the membrane culture system (C, A), the largest follicles were observed throughout the tissue without any bias in localization ([Fig 4C](#pone.0143114.g004){ref-type=\"fig\"}). These large follicles consisted of multi-layered granulosa cells and full-grown GV oocytes regardless of whether or not activin A was added. Under the 3-D culture system (M, M+A), large follicles were also detected in the tissue, but these were predominantly localized at the periphery of each tissue ([Fig 4D](#pone.0143114.g004){ref-type=\"fig\"}), while the central area seemed to undergo gradual degeneration accompanied by loss of healthy follicles after 10 days of culture. Most importantly, however, antral cavity formation occurred only within the large follicles generated in the 3-D culture system ([Fig 4D and 4F](#pone.0143114.g004){ref-type=\"fig\"}), and was not observed in follicles cultured in the membrane culture system ([Fig 4C and 4E](#pone.0143114.g004){ref-type=\"fig\"}). In the M and M+A conditions, large follicles typically generated an antral cavity to advance their developmental stage regardless of the presence or absence of activin A (Figs [3H, 3J](#pone.0143114.g003){ref-type=\"fig\"} and [4D, 4F](#pone.0143114.g004){ref-type=\"fig\"}). In tissues maintained under the membrane culture conditions, no cavity formation was observed despite steady follicle growth (Figs [3D, 3F](#pone.0143114.g003){ref-type=\"fig\"} and [4C, 4E](#pone.0143114.g004){ref-type=\"fig\"}). In addition, tissues cultured on the Matrigel drop maintained a thickness nearly 3 times larger than those cultured on the membrane filter alone ([Fig 4G and 4H](#pone.0143114.g004){ref-type=\"fig\"}), suggesting tissue structure may critically affect follicle development in vitro. These results clearly demonstrate that our 3-D culture system is able to support development of secondary follicles to the antral follicle stage by maintaining normal ovarian architecture.\n\n![Histological analysis of ovarian tissues under Membrane or 3-D culture system.\\\n**(a, b)** Paraffin sections of a whole murine ovary at 14 dpp (a) and 24 dpp (b) with a HE staining. **(c, e, g)** The ovarian tissues cultured in the Membrane culture system (C condition) for 10 days. **(d, f, h)** The ovarian tissues cultured in the 3-D culture system (M+A condition) for 10 days. Follicle growth was detected in both culture systems. Importantly, the antral cavity formation (arrows) occurred only in the 3-D culture system (M, M+A) regardless of the presence or absence of activin A treatment. Regular bar = 200 \u03bcm. Bold bar = 100 \u03bcm.](pone.0143114.g004){#pone.0143114.g004}\n\nChromatin configuration in GV oocytes grown *in vitro* {#sec016}\n------------------------------------------------------\n\nMammalian oocyte development is characterized by dynamic changes in chromatin structure and function within the germinal vesicle (GV). Chromatin in the GV is initially decondensed with the nucleolus not surrounded by heterochromatin (NSN). During oocyte growth, GV chromatin condenses into perinucleolar rings (SN) \\[[@pone.0143114.ref033]\\]. Thus, chromatin condensation provides a good indicator of acquisition of developmental competence in GV oocytes \\[[@pone.0143114.ref033]\\]. On the basis of GV chromatin configuration (NSN, Int, SN) ([Fig 5A](#pone.0143114.g005){ref-type=\"fig\"}), we assessed nuclear maturity of oocytes grown under the four different culture conditions ([Fig 5B](#pone.0143114.g005){ref-type=\"fig\"}). Before culture (Day 0), all GV oocytes isolated from multilayered secondary follicles exhibited the immature NSN chromatin configuration ([Fig 5B](#pone.0143114.g005){ref-type=\"fig\"} and [S2 Table](#pone.0143114.s002){ref-type=\"supplementary-material\"}). On Day 6 after culture, oocytes showed various patterns of chromatin configuration that varied among the culture conditions. In the membrane culture system (C, A condition), 38--48% of oocytes retained the NSN configuration. In contrast, under the 3-D culture without activin A (M), 84% of oocytes had progressed to the Int or SN configuration. In the presence of activin A (M+A), chromatin condensation was further advanced, with almost all oocytes (98%) having reached the Int or SN configuration ([Fig 5B](#pone.0143114.g005){ref-type=\"fig\"} and [S2 Table](#pone.0143114.s002){ref-type=\"supplementary-material\"}). By Day 10 of culture, GV oocytes isolated from the larger follicles had reached full-size as shown in [Fig 2B](#pone.0143114.g002){ref-type=\"fig\"}. Almost all of these full-sized GV oocytes had progressed to the nuclear matured status of SN regardless of the culture condition used ([Fig 5B](#pone.0143114.g005){ref-type=\"fig\"} and [S2 Table](#pone.0143114.s002){ref-type=\"supplementary-material\"}).\n\n![Chromatin configuration of GV oocytes isolated from follicles in ovarian tissues cultured under 4 different conditions.\\\n**(a)** Chromatin configurations of mouse GV oocytes after Hoechst staining (400x). By the presence or absence of a ring of condensed chromatin around the nucleorus, the oocytes were categorized into three groups: NSN (non-surrounded nucleolus), SN (surrounded nucleolus) and Int (intermediate stage). The SN configuration represents a stage of GV oocyte that is more advanced toward ovulation. Each square represents a magnified chromatin configuration. **(b)** Distribution of chromatin configuration patterns in the oocytes isolated from the follicles on Days 0, 6 and 10 of culture. On Day 6, the GV chromatin condensation was more advanced under the 3-D Culture (M, M+A) compared to under the Membrane Culture (C, A) (a-c: *P*\\<0.05; b,c: *P*\\<0.02). By Day 10 of culture, almost all of full-size GV oocytes progressed to the SN condfiguration regardless of the 4 different culture conditions.](pone.0143114.g005){#pone.0143114.g005}\n\nIVM {#sec017}\n---\n\nWe next examined the quality of the full-size GV oocytes grown under the four different conditions. First, we assessed maturation competence of these oocytes ([Table 1](#pone.0143114.t001){ref-type=\"table\"}). On Day 10 of tissue culture, GV oocytes surrounded by granulosa cells were isolated from the larger follicles ([Fig 2A and 2B](#pone.0143114.g002){ref-type=\"fig\"}) to be subjected to IVM ([Fig 6A and 6B](#pone.0143114.g006){ref-type=\"fig\"}). The GV oocytes in the membrane culture (C, A) were isolated from the follicles without antral cavities ([Fig 3D and 3F](#pone.0143114.g003){ref-type=\"fig\"}). However, GV oocytes in the 3-D culture (M, M+A) were collected from the follicles with cavities ([Fig 3H and 3J](#pone.0143114.g003){ref-type=\"fig\"}). After IVM, the granulosa cell layers surrounding the oocytes underwent expansion regardless of the culture conditions ([Fig 6C](#pone.0143114.g006){ref-type=\"fig\"}). In the membrane culture system (C, A), about 40% of the GV oocytes matured to reach the MII stage in vitro ([Fig 6D](#pone.0143114.g006){ref-type=\"fig\"}) regardless of the presence or absence of activin A treatment ([Table 1](#pone.0143114.t001){ref-type=\"table\"}). In the 3-D culture system (M, M+A), in contrast, the maturation rates were significantly increased relative to the membrane culture ([Table 1](#pone.0143114.t001){ref-type=\"table\"}). More than half of the GV oocytes grown under the M condition reached the MII stage in vitro (54%). With the addition of activin A (M+A), the MII rate was further increased to about 66% after IVM.\n\n![Developmental competence of full-size GV oocytes grown under 3-D culture system.\\\n**(a)** After 10 days of the 3-D ovarian tissue culture (M+A), GV oocytes surrounded by granulosa cells were isolated from the antral follicles for in vitro maturation (IVM). **(b)** To observe germinal vesicles (arrows) in GV oocytes, surrounding granulosa cells were mechanically removed. **(c)** After IVM, granulosa cell layers surrounding the oocytes underwent expansion. **(d)** To observe first polar bodies (arrows) in matured MII oocytes, expanded granulosa cell layers were removed after IVM. **(e)** In vitro matured oocytes with expanded granulosa cells were subjected to in vitro fertilization (IVF). Six hrs after IVF, two pronuclei (arrows) were observed in each fertilized oocyte. **(f)** Twenty-four hrs after IVF, some of fertilized oocytes developed to the 2-cell stage embryos. **(g)** The 2-cell embryos were transferred to surrogate mothers to obtain full-term offspring with black coat color. Bar = 40 \u03bcm.](pone.0143114.g006){#pone.0143114.g006}\n\n10.1371/journal.pone.0143114.t001\n\n###### *In Vitro* Maturation of GV Oocytes Grown under 4 Different Ovarian Culture Conditions.\n\n![](pone.0143114.t001){#pone.0143114.t001g}\n\n                  No. of oocytes after *in vitro* maturation              \n  ----- --------- -------------------------------------------- ---- ----- -------------\n  C     252 (8)   7                                            13   127   103 (40.9)a\n  A     60 (3)    0                                            17   19    24 (40.0)a\n  M     184 (4)   0                                            18   65    100 (54.3)b\n  M+A   262 (6)   2                                            11   75    172 (65.6)c\n\nValues with different superscripts within the same column are significantly different (a-c: *P*\\<0.01; b, c: *P*\\<0.02).\n\nIVF {#sec018}\n---\n\nAfter IVM, in vitro matured oocytes surrounded by granulosa cell layers were subjected to IVF and examined for development to the 2-cell embryo stage ([Table 2](#pone.0143114.t002){ref-type=\"table\"}, [Fig 6E and 6F](#pone.0143114.g006){ref-type=\"fig\"}). Under the membrane culture without activin A treatment (C), only 11% of the MII oocytes developed to the 2-cell stage. In the presence of activin A (A), the 2-cell rate was slightly increased to 20%. However, using oocytes derived in the 3-D culture system, the 2-cell developmental rates were dramatically improved compared to those derived with membrane culture. Thus, without activin A (M), 45% of 3-D culture system MII oocytes developed to 2-cell embryos after IVF. Activin A treatment (M+A) further improved the developmental capacity of MII oocytes, with 58% developing to the 2-cell stage. These results demonstrate that our 3-D culture system enhances fertilization rates of oocytes grown in vitro, and activin A treatment increases this quality even further.\n\n10.1371/journal.pone.0143114.t002\n\n###### Two-Cell Development after *In Vitro* Fertilization of Oocytes Grown under 4 Different Ovarian Culture Conditions.\n\n![](pone.0143114.t002){#pone.0143114.t002g}\n\n  Condition   No. of MII oocytes (replicates)   No. of 2-cell (%)\n  ----------- --------------------------------- -------------------\n  C           438 (7)                           49 (11.2)a\n  A           54 (4)                            11 (20.4)a\n  M           118 (3)                           53 (44.9)b\n  M+A         208 (4)                           120 (57.7)c\n\nValues with different superscripts within the same column are significantly different (a-c: *P*\\<0.01; b, c: *P*\\<0.02).\n\nET {#sec019}\n--\n\nTo determine the full-term developmental capacity of oocytes derived by IVG, 2-cell stage embryos were transferred into surrogate mothers (Tables [3](#pone.0143114.t003){ref-type=\"table\"} and [S3](#pone.0143114.s003){ref-type=\"supplementary-material\"}). One of forty 2-cell embryos derived from oocytes grown in the membrane culture system without activin A (C) developed into a full-term pup after ET. None of the 2-cell embryos derived from the membrane culture in the presence of activin A (A) developed to full-term pups, suggesting activin A treatment does not, by itself, promote the developmental competence of oocyes grown on a membrane filter. In contrast, oocytes derived from the 3-D culture system supported by a Matrigel drop showed significantly improved rates of full-term development of transferred embryos. Thus, in the absence of activin A (M), we obtained 5 full-term pups from 80 transferred embryos (6%). In the presence of activin A (M+A), developmental competence to term rose to 15%, with 8 full-term pups recovered from 55 transferred embryos ([Fig 6G](#pone.0143114.g006){ref-type=\"fig\"}). The average body weights (1.51--1.86 g) and placenta weights (0.15 g) of the pups derived from IVG oocytes were slightly higher than those of in vivo control pups derived from natural mating (Tables [4](#pone.0143114.t004){ref-type=\"table\"} and [S4](#pone.0143114.s004){ref-type=\"supplementary-material\"}). After ET, a number of low-quality embryos derived from IVG oocytes were degenerated to make smaller litters (1--3 pups per surrogate) ([S3](#pone.0143114.s003){ref-type=\"supplementary-material\"} and [S4](#pone.0143114.s004){ref-type=\"supplementary-material\"} Tables). It has been demonstrated that pups in smaller litters tend to be larger at birth than those in larger litters. All pups were otherwise healthy and normal.\n\n10.1371/journal.pone.0143114.t003\n\n###### Newborn Offspring Derived from Oocytes Grown under 4 Different Ovarian Culture Conditions.\n\n![](pone.0143114.t003){#pone.0143114.t003g}\n\n  Condition   No. of 2-cell embryos transferred (No. of surrogates)   No. of pups (%)\n  ----------- ------------------------------------------------------- -----------------\n  C           40 (5)                                                  1 (2.5)a\n  A           20 (3)                                                  0 (0)a\n  M           80 (5)                                                  5 (6.3)a, b\n  M+A         55 (6)                                                  8 (14.5)b\n\nValues with different superscripts within the same column are significantly different (a, b: *P\\<*0.0 5).\n\n10.1371/journal.pone.0143114.t004\n\n###### Body and Placenta Weights of Newborn Offspring Derived from Oocytes Grown under 4 Different Ovarian Culture Conditions.\n\n![](pone.0143114.t004){#pone.0143114.t004g}\n\n  Condition                                         No. of pups   Male/Female   Body weight (g) \u00b1 SD   Placenta weight (g) \u00b1 SD\n  ------------------------------------------------- ------------- ------------- ---------------------- --------------------------\n  C                                                 1             1/0           1.62                   0.15\n  M                                                 5             1/4           1.86 \u00b1 0.38            0.15 \u00b1 0.014\n  M+A                                               8             2/6           1.51 \u00b1 0.24            0.15 \u00b1 0.026\n  *In Vivo* [\\*](#t004fn001){ref-type=\"table-fn\"}   7             2/5           1.31 \u00b1 0.08            0.11 \u00b1 0.014\n\n\\* *In vivo* control pups were derived from natural mating between B6D2F1 female and male mice.\n\nDiscussion {#sec020}\n==========\n\nCurrently 3-D isolated follicle culture is a popular method to develop secondary \\~ preantral follicles to the antral follicle state in vitro. However, this system requires very complicated and time-consuming processes \\[[@pone.0143114.ref001], [@pone.0143114.ref019], [@pone.0143114.ref021]--[@pone.0143114.ref023], [@pone.0143114.ref034]\\]. In this study, we have developed a simple and easy ovarian tissue culture system to yield high quality oocytes derived from secondary follicles in mice. A thick Matrigel drop placed under the tissue supports proper follicle development during culture. Oocytes grown in our 3-D culture system acquired much higher developmental competence than those grown without the Matrigel drop, and developed to full-term offspring after IVM and IVF. Activin A supplement further enhanced the developmental competence of oocytes grown in our 3-D ovarian culture system. These results demonstrate that our 3-D culture system provides an optimal environment for proper development of secondary follicles to the antral stage.\n\nIn situ ovarian tissue culture is a promising technology for in vitro follicle growth, because this system: i) involves a simple and easy procedure; ii) is not damaged by chemical or mechanical dissociation to isolate follicles; iii) maintains an intact ovarian environment; and iv) could provide an in vitro model for ovarian toxicology testing \\[[@pone.0143114.ref035]\\]. Regardless of these advantages, it has been argued that in vitro ovarian culture does not consistently support proper development of advanced follicles \\[[@pone.0143114.ref004], [@pone.0143114.ref035]\\]. In this study, support afforded by a thick Matrigel drop successfully improved the outcome of ovarian tissue culture to yield developmentally competent oocytes. Thus, it appears the multiple ECM proteins present in Matrigel provide a mechanical scaffold that supports spatial and temporal growth of the multiple follicles within a segment of ovarian tissue. Importantly, the quality of the resulting oocytes was critically improved by our 3-D culture system compared to that detected in oocytes derived from ovarian tissue cultured without the Matrigel drop. These results are consistent with previous reports suggesting that oocytes acquire developmental competence only when the follicle is able to maintain its normal architecture \\[[@pone.0143114.ref001]\\]. Normal folliculogenesis is critically regulated by cell-cell communications between granulosa cells and the oocyte, and among granulosa and theca cells through a highly organized network of connections \\[[@pone.0143114.ref036]\\]. Our results suggest that a microenvironment conducive to normal folliculogenesis was provided by our 3-D culture system yielding developmentally competent oocytes. Currently encapsulation of follicles in alginate hydrogel is the most widely used 3-D culture system to promote advanced follicle growth \\[[@pone.0143114.ref001], [@pone.0143114.ref019], [@pone.0143114.ref021]--[@pone.0143114.ref023], [@pone.0143114.ref034]\\]. Xu et al. (2006) assessed the quality of murine oocytes grown in alginate-encapsulated secondary follicles (150--180 \u03bcm) from 16 dpp mice after 8 days in culture \\[[@pone.0143114.ref021]\\]. After IVM, about 71% of GV oocytes reached the MII stage, and 68% of those were fertilized after IVF. Four normal offspring (20% of fertilized embryos) were obtained after ET. Our ovarian culture system (M+A condition), yielded success rates of IVM, IVF and ET of 66%, 59%, and 15%, respectively), which were very close to the results reported by Xu et al. (2006), even though our data were derived from younger stage follicles (average: 131 \u03bcm, all less than 150 \u03bcm in size) recovered from ovaries at 14 dpp and cultured for 10 days, demonstrating that our 3-D culture system reliably generates fertilizable oocytes from secondary follicles in vitro.\n\nIn our 3-D system, the tissues have not been enclosed by either Matrigel or the medium, therefore they have easily accessed to air in culture. Recently, alginate encapsulation has been utilized with 3-D ovarian tissue culture to study the growth of the ovarian surface epithelium \\[[@pone.0143114.ref037]--[@pone.0143114.ref039]\\]. Small pieces of immature ovary were encapsulated into alginate hydrogel and cultured within medium. As a result, the tissues successfully maintained 3-D structure and grew ovarian epithelium, however, secondary follicles in these tissues did not develop or survive in this culture system \\[[@pone.0143114.ref039]\\]. In our preliminary experiment, we also observed that ovarian tissues cultured within medium quickly degenerated and lost all growing follicles (data not shown), suggesting that a hypoxic condition is not optimal for follicle growth in ovarian tissue culture. These observations suggest that optimal culture conditions for follicles within ovarian tissues are quite different than those for isolated follicles. In contrast, it was reported that human cortical ovarian tissues encapsulated within alginate and cultured for six weeks showed at least one primordial follicle that had developed to an antral follicle in the cortex \\[[@pone.0143114.ref040]\\]. Taken together, these results suggest that much work remains to be done to completely optimize follicle culture systems specific for each species that can yield normal follicle stages and developmentally competent oocytes in each case.\n\nIn this study, we also demonstrated that the activin A treatment further enhances the quality of oocytes grown in 3-D culture. In the ovary, activin A is generated by granulosa cells in primary to antral stage follicles to regulate their proliferation, differentiation, and steroidogenesis \\[[@pone.0143114.ref024]--[@pone.0143114.ref025]\\]. As a member of the TGF-\u03b2 superfamily, activin A stimulates Smad signaling pathways \\[[@pone.0143114.ref041]\\] and increases the number of FSH receptors in granulosa cells to promote FSH action on these cells \\[[@pone.0143114.ref042]\\]. FSH autocrine signaling allows GTP coupled receptors to activate the PKA pathway leading to granulosa cell differentiation \\[[@pone.0143114.ref043]\\]. In our 3-D system, additional activin A treatment appeared to further stimulate these endogenous phenomena to enhance the developmental competence of the resulting oocytes. Zhang *et al*. (2012) found that activin A and preantral granulosa cells cooperated to enhance the growth and maturation of immature oocytes in vitro \\[[@pone.0143114.ref044]\\]. Their data further demonstrated that activin A enhances a complex effect on the oocyte maturation, such as oocyte cytoplasmic maturation and oocyte-granulosa cell communication via gap junction \\[[@pone.0143114.ref044]\\]. Interestingly, our results showed that addition of activin A to our membrane culture system without Matrigel (our \"A condition\") did not enhance oocyte quality, suggesting that exogenous activin A functions only when the ovarian microenvironment is properly organized and maintained. Previous studies have demonstrated that a combination of Matrigel and activin A synergistically enhances in vitro follicle growth compared to conditions utilizing individual ECM proteins plus activin A \\[[@pone.0143114.ref008]\\], or Matrigel alone \\[[@pone.0143114.ref045]\\]. The ECM proteins act as a reservoir for growth factors, communicate with the intracellular cytoskeleton, and transmit growth factor signals \\[[@pone.0143114.ref009], [@pone.0143114.ref012]\\]. Indeed, many ECM proteins have binding sites for both growth factors and cell adhesion functions, increasing local concentrations of the growth factors near their cell surface receptors and cell adhesion sites \\[[@pone.0143114.ref012]\\]. It was reported that activin A selectively binds to specific ECM proteins such as heparan sulfate proteoglycans \\[[@pone.0143114.ref046]\\]. Oktay et al. (2000) demonstrated that an interaction between laminin and activin A promotes the transition from the primary to the multilayer follicle stage, whereas the combination of type IV collagen and activin A suppresses this transition \\[[@pone.0143114.ref047]\\]. As previously mentioned, the ovarian microenvironment provides stage-specific ECM composition and regulates actions of multiple factors such as activin A to organize follicle development. Therefore, the interaction between multiple ECM proteins (Matrigel) and activin A might synergistically up-regulate relevant signal transduction pathways in the follicular cells to eventually generate high quality oocytes in vitro. It will be important to determine what ECM component(s) are responsible for enhancing the quality of oocytes in a stage-specific manner.\n\nIn summary, the 3-D ovarian culture system we describe here has the potential to be a key technical advance for use in both clinical and laboratory settings. This system can serve as a biological model for studying complex follicle-follicle and follicle-epithelium interactions within the ovary. Furthermore, this model could be useful for toxicological studies involving different animal models both in vivo and in vitro to assess harmful effects of specific compounds or conditions on the female reproductive system including the ovary \\[[@pone.0143114.ref048]--[@pone.0143114.ref049]\\]. Currently, reproductive toxicology testing is primarily performed in vivo, however, our culture system could offer a novel in vitro option. For assisted reproductive technologies (ART), the application of in vitro folliculogenesis will contribute to an ultimate goal of the ability to generate fertilizable oocytes in vitro. Especially for prepubertal cancer patients, establishment of this technology will provide a novel approach to fertility preservation. Although it is still very challenging to generate high quality human oocytes in vitro, the 3-D ovarian culture system we describe here for use with mouse tissues will pave the way to advancement of similar technologies enabling fertility preservation in humans.\n\nSupporting Information {#sec021}\n======================\n\n###### Follicle and oocyte growth during the ovarian tissue culture.\n\nDuring the ovarian tissue culture in the membrane (C, A) and in the 3-D (M, M+A) culture systems, we measured the follicle and oocyte sizes on Days 0, 6 and 10 of culture.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Chromatin configuration of GV oocytes.\n\nOn Days 0, 6 and 10 of the ovarian tissue culture, GV oocytes were isolated and subjected to Hoechst staining for 8 minutes at 37\u00b0C to determine the status of their chromatin. The GV oocytes were categorized into NSN, Int and SN.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Newborn offspring after 2-cell embryo transfer.\n\nThe 2-cell stage embryos derived from 4 different culture conditions were transferred into the surrogate mothers to obtain full-term pups.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Body and placenta weights of newborn offspring.\n\nThe body and placenta weights of newborn pups were examined and compared to those of *in vivo* derived pups.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nThe authors thank Dr. John R. McCarrey for reviewing the original manuscript. We also acknowledge Drs. W. Steven Ward and Tomas Huang for their advice on this manuscript and experiment.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: YY CMH YM TH. Performed the experiments: CMH YM YY. Analyzed the data: YY CMH YM. Contributed reagents/materials/analysis tools: YY CMH YM TH. Wrote the paper: YY CMH.\n"}
+{"text": "Introduction {#s1}\n============\n\nSocial anxiety disorder (SAD) is one of the most common mental disorders with a lifetime prevalence of up to 12% in Western countries (Fehm et al., [@B17]; Kessler et al., [@B28]). Hallmark characteristics are intense fear and avoidance of being evaluated or criticized resulting in extreme discomfort and self-consciousness in everyday social situations (American Psychological Association, [@B1]). Theoretical models highlight the importance of cognitive biases in the processing of ambiguous or negative cues during social interactions for the etiology and/or maintenance of social anxiety (Rapee and Heimberg, [@B42]; Clark and Mcmanus, [@B10]). More specifically, studies show that socially anxious individuals have attentional biases in the processing of negative, rejection-related cues (Bar-Haim et al., [@B6]) and interpret ambiguous social situations as more threatening and negative than healthy controls (e.g., Stopa and Clark, [@B50]; Beard and Amir, [@B7]).\n\nRelatively little attention, however, has been paid to biases in gaze perception. This is particularly surprising since individuals with SAD experience intense feelings of being looked at by other individuals and show a marked avoidance and fear of eye contact during social interactions (Schneier et al., [@B48]). Biases in the self-referential perception of gaze directions, for instance, might more easily elicit feelings of mutual gaze and being the center of attention, which then will activate fears of being scrutinized by others. Here, we review studies with clinical and non-clinical socially anxious samples on self-referential and threat-related biases in the perception of mutual gaze.\n\nFirst, mutual gaze perception in healthy human beings will be discussed. Next, biases in the perception of other individuals' gaze in social anxiety will be reviewed with a focus on: (a) whether mutual gaze is more readily perceived; and (b) whether mutual gaze is avoided and perceived as threatening.\n\nGaze perception in healthy human beings {#s2}\n=======================================\n\nMost mammals generally interpret direct gaze as threatening or as a sign of dominance. Humans in contrast often associate mutual gaze with positive interest, such as love and attraction. A preference for direct gaze seems to be present at a very early age: Farroni et al. ([@B16]) found that infants as young as 2 days old prefer to look at faces that gazed directly at them compared to faces with averted gaze. Yet, humans sometimes find eye contact uncomfortable, for example if a stranger keeps staring at them.\n\nDifferent sources of information are taken into account when processing gaze direction. The most obvious cue lies in the eye itself. Kobayashi and Kohshima ([@B30], [@B31]) compared the eyes of a large number of primates and found that the morphology of the human eye is rather unique. Of all compared species human eyes have the highest width to height ratio and the highest index of exposed sclera size. The amount of visible sclera provides information about the orientation of the eyeball (Gibson and Pick, [@B20]; Cline, [@B11]; Anstis et al., [@B3]; Langton et al., [@B35]; Ando, [@B2]). Ando ([@B2]) provided direct evidence that the iris/sclera ratio is an important cue for eye gaze perception. By darkening one side of the sclera of eyes directed straight ahead, he found a substantial shift of the perceived gaze direction towards the darkened side.\n\nAnother factor influencing gaze perception is the head direction of the looker. Langton ([@B34]) (see also Wollaston, [@B56]) found that the orientation of another person's head strongly influenced the perceived direction of the person's gaze. Body posture is yet another cue that can provide information about where someone is attending (Perrett et al., [@B40]).\n\nStudies that focus on the ability to distinguish between direct and averted eye gaze are relatively numerous. All these studies generally report that human observers are highly accurate at determining mutual eye gaze. In their classic study, Gibson and Pick ([@B20]) asked observers to indicate whether a \"looker\" who was sitting opposite was making eye contact or looking at a peripheral target. The authors found that an angular deviation of the eye by only 2.8\u00b0 was correctly detected as not making eye contact. Cline ([@B11]) replicated and extended these findings and reported that an angular deviation of as little as 0.75\u00b0 was readily detected by an observer. Such high accuracy rates in detecting mutual gaze are not undisputed, since a number of studies found relatively poor discrimination of gaze direction, especially when the distance between looker and observer was large (e.g., Vine, [@B51]). With decreasing security (i.e., when visual information was reduced through distance or noise) observers tended to assume mutual gaze. There thus seems to be a considerable range wherein a person feels being looked at. Gaze direction might hence be better described as a cone rather than a ray (as assumed by e.g., Gibson and Pick, [@B20]; Cline, [@B11]). Consequently, Gamer and Hecht ([@B18]) introduced the cone of direct gaze (CoDG) as a concept to measure mutual gaze perception. The authors found an average width of the CoDG of between 4\u00b0 and 9\u00b0 of visual angle, depending on the distance between looker and observer.\n\nFacial emotional expression is another cue taken into account when judging gaze directions. Lobmaier et al. ([@B37]) presented participants with three-dimensional models that were either facing the observer, or were rotated 2\u00b0, 4\u00b0, 6\u00b0, 8\u00b0, and 10\u00b0 to the left and right. In this study eye gaze and head direction were aligned with each other (i.e., the whole head was rotated keeping the eyes relative to the head direction constant). Participants were asked to judge whether the face was looking at them or not. The results revealed a remarkable positivity bias: happy faces were more likely perceived as looking at the observer than angry, fearful, or neutral faces. The authors interpreted this finding in favor of self-esteem preservation: perceiving other's happiness as directed at oneself is socially rewarding (see also Lobmaier and Perrett, [@B36]). This interpretation is compatible with the assumption that human beings have a prior expectation that other people's gaze is directed towards them (Mareschal et al., [@B38]).\n\nEwbank et al. ([@B15]) employed the CoDG metaphor to further test the influence of emotional expression on perception of direct gaze. Using the method of constant stimuli (see also Mareschal et al., [@B38]) angry, fearful and neutral faces were presented in which the direction of eye gaze was manipulated. They found that the CoDG was significantly wider for angry faces compared to neutral and fearful faces.\n\nThe studies reviewed above reveal that gaze perception plays an important role in social interactions and is modulated by several factors, such as head direction, interpersonal distance, or emotional facial expressions (see also reviews by Graham and LaBar, [@B21]; Carlin and Calder, [@B8]; for behavioral and neuroscientific findings of gaze processing and gaze-emotion interactions). Given that social interactions are affected in SAD, it is conceivable that social anxiety might be associated with impeded gaze perception. In the following sections we discuss gaze perception in the context of SAD.\n\nGaze perception in social anxiety and social anxiety disorder {#s3}\n=============================================================\n\nSelf-directed perception of gaze {#s3-1}\n--------------------------------\n\nIn recent years, several studies have investigated the perception of self-directed gaze in order to quantify the perception of mutual gaze in social anxiety. Initial work used the previously described \"cone of gaze\" paradigm to investigate the self-directed perception of gaze cues in SAD (Gamer et al., [@B19]). In half of the trials an additional task-irrelevant looker was presented. The results provided support that patients with SAD exhibit an enlarged self-directed perception of gaze directions, but only in the presence of a second virtual looker. The magnitude of this effect was positively correlated with the severity of social anxiety symptoms.\n\nSubsequent work investigated dimensional relations between social anxiety and the perception of gaze directions in a non-clinical sample, while also addressing the specific role of facial emotional expressions (Schulze et al., [@B49]). Severity of social anxiety was positively correlated with the self-directed perception of other individuals' gaze, especially when the \"lookers\" exhibited a neutral or negative (i.e., angry, fearful) facial expression. In addition, response latencies negatively interacted with symptoms of social anxiety, presumably reflecting an increased avoidance of direct gaze. Similar findings were reported by Jun et al. ([@B27]) who assessed self-directed gaze perception using male facial stimuli in students with high and low social anxiety. An increased cone of gaze was found only in male students with marked social anxiety, possibly because male students experienced greater discomfort when being looked at than females (see also Jun et al., [@B27], for a discussion of possible interactions between the sex of \"lookers\" and \"observers\" in mutual gaze perception).\n\nNotably, enhanced self-referential perception of gaze directions was also demonstrated in more ecologically valid experimental setups with alive target stimuli. Harbort et al. ([@B22]) studied the effects of real persons and virtual heads on gaze perception. The findings underpinned that the CoDG was generally increased in SAD, but that effect sizes were larger in the *Real-Person-Condition* than in the *Virtual-Head-Condition*. The widening of the gaze cone in the *Real-Person-Condition* was suggested to be a consequence of higher arousal in SAD patients when confronted with a real person. In line with the proposed role of arousal, stress-induced increases in cortisol levels were previously shown to increase feelings of being looked at (Rimmele and Lobmaier, [@B43]). A face-to-face situation was also used by Honma ([@B23]) who found the range of gaze directions perceived as self-directed to be much larger than the actual amount of eye contact and perception of mutual gaze was accompanied by greater pupil dilations (see also Honma et al., [@B24]). In this study, severity of social anxiety was positively correlated with perceived eye contact and pupil dilation.\n\nHarbort et al. ([@B22]) assessed the effects of Cognitive Behavioral Therapy (CBT) on gaze perception. Patients with SAD were tested prior to standardized CBT and again after approximately 24 therapy sessions had been completed. Prior to psychotherapeutic treatment, patients with SAD were characterized by increased perceptions of gaze as being self-directed. Intriguingly, after CBT patients with SAD did not differ from healthy controls, suggesting that interventions aiming at reducing SAD symptoms lead to a normalization of the gaze cone. These findings still need to be considered preliminary since the interaction of group and assessment time failed to reach significance; several alternative explanations might thus account for the observed pattern.\n\nIn sum, available studies in SAD demonstrated an abnormal perception of mutual gaze, providing a quantification of the intense feelings of being looked at. Findings unanimously demonstrated an enhanced self-directed perception of gaze, particularly for negative and neutral facial expressions. Further studies are needed to investigate whether the cone of gaze changes due to psychotherapeutic interventions.\n\nThreat perception and avoidance of mutual gaze {#s3-2}\n----------------------------------------------\n\nClinical observations suggest fear and avoidance of *direct* eye contact to be prominent characteristics of SAD. Yet, empirical evidence on threat-related perception and avoidance of direct gaze compared with averted gaze is still scarce.[^1^](#fn0001){ref-type=\"fn\"}\n\nInitial studies provided some support that mutual gaze is feared and avoided in social interactions (e.g., Daly, [@B13]; Baker and Edelmann, [@B5]). These findings are however limited because subjective observations were used as dependent measures. Objective evidence for an avoidance of salient facial features was first provided by studies using eye-tracking to investigate visual responses to static images with direct gaze. Comparing visual scanpaths of emotional facial expressions in patients with SAD and healthy controls yielded an active avoidance of salient facial features such as the eye region in SAD. This was particularly reflected in reduced number and duration of fixations of the eye region while a \"hyperscanning\" strategy was exhibited for remaining facial features (Horley et al., [@B25], [@B26]). This distinct visual scanning behavior was most prominent for expressions of threat, whereas group differences were least pronounced in response to neutral or happy facial expressions (Horley et al., [@B25], [@B26]). Moukheiber et al. ([@B39]) later replicated these results, finding less fixations and shorter dwell times on the eye region in SAD compared to healthy individuals. Again, group differences were most notable for expressions of social threat (i.e., anger and disgust). A reduced number and duration of fixations upon the eye region were also reported when SAD patients received social feedback (Weeks et al., [@B52]).\n\nWhile these studies demonstrate an avoidance of the eye region, questions remained unanswered to what extent others' gaze directions differentially affect avoidance behavior in SAD. This question was recently addressed by means of the Approach-Avoidance Task. In social anxiety, behavioral avoidance of angry faces was present only when coupled with direct gaze (Roelofs et al., [@B44]). Notably, administration of oxytocin facilitated approach behavior towards angry faces with direct gaze in socially anxious individuals (Radke et al., [@B41]). In a related line of research, fixation behavior was investigated in response to animated video clips of faces with direct or averted gaze (Wieser et al., [@B53]). In high socially anxious participants longer fixations on the eye region were observed, although effects were only marginally significant. Additionally, heightened physiological arousal in socially anxious individuals was found for direct compared to averted gaze suggesting that mutual gaze is perceived as threatening. In line with this interpretation, increased startle reactivity was observed the very moment a virtual audience directed their eye gaze and attention towards individuals with SAD who had to deliver a speech (Cornwell et al., [@B12]). A virtual-reality environment was also used by Wieser et al. ([@B54]) to scrutinize the interplay between gaze directions, interpersonal distance, and sex of the interaction partner on avoidance behavior. Socially anxious individuals were found to avoid eye contact and to show increased backward head movements in response to male avatars with direct gaze.\n\nFurther evidence for the threatening quality of direct gaze was obtained by functional neuroimaging studies in SAD (see Etkin and Wager, [@B14] for a meta-analysis of neuroimaging and emotion processing in SAD). In a preliminary study comparing neural responses to direct and averted gaze, patients with SAD were found to exhibit greater activation in parts of the fear circuitry including the amygdala, insula, and anterior cingulate cortex (Schneier et al., [@B46]). Additional eye tracking results indicated that SAD patients show a greater avoidance of the eye region in stimuli with direct compared to averted gaze than healthy controls. In a subsequent study, neural responses to direct and averted gaze were assessed before and after intervention with paroxetine in patients with generalized SAD (gSAD; Schneier et al., [@B47]). At baseline, gSAD patients showed greater activation than healthy controls in brain regions related to self-referential processing and emotion regulation such as cortical midline structures of the ventromedial prefrontal cortex and the posterior cingulate cortex, when looking at direct versus averted gaze. However, fixation of the eye region did not differ significantly between gSAD patients and healthy controls. Pharmacological treatment resulted in a normalization of brain activation in response to direct gaze.\n\nIn contrast to the studies reviewed above, recent electrophysiological evidence suggested a specific processing bias for averted gaze in social anxiety as implied by enhanced late positive potentials and (marginally significant) higher amplitudes of the P100 in response to averted gaze (Schmitz et al., [@B45]). These authors proposed that direct gaze might only be perceived as threatening when coupled with negative facial expressions, whereas neutral expressions with averted gaze might rather signal disinterest.\n\nTaken together, there is ample evidence that mutual gaze is perceived as threatening by socially anxious individuals. However, findings are less consistent regarding the avoidance of mutual gaze in SAD. While several studies demonstrated an avoidance of the eye region when coupled with direct gaze, some studies failed to observe group differences and one study even reported prolonged fixation of the eye region. Future research suggestions will be discussed in the final section.\n\nSummary and Future Research {#s4}\n===========================\n\nClinical observations suggest abnormalities in gaze perception to be important for SAD. In accordance with such claims, findings from analog samples and clinical populations demonstrated a greater cone of gaze and a pronounced fear of direct eye contact in social anxiety. In addition, recent findings suggest that individuals with SAD avoid mutual gaze, but these results are less consistent.\n\nIn socially anxious individuals, a biased self-referential perception of gaze directions may underlie the fear of being the center of attention and cause uneasiness and discomfort. Specifically, biased perceptions of mutual gaze may lead socially anxious individuals to appraise a situation as social, which results in a heightened processing of the self as a social object, ultimately resulting in a negative cascade of somatic, cognitive, and behavioral consequences (Clark and Mcmanus, [@B10]). The avoidance of eye contact in social anxiety may be understood as an attempt to avoid signs of social threat and to regulate excessive fears of being evaluated. This avoidance behavior may contribute to the maintenance of SAD by negatively reinforcing expectations and fears of social encounters. Alternatively, taking into account findings of gaze aversion in social anxiety, it is also conceivable that SAD patients fail to extract relevant cues from the eye region. This factor may lead to abnormal perceptions of being looked at. A promising direction for future studies may therefore be to combine eye-tracking methods with paradigms of mutual gaze perception to further disentangle causes and consequences of abnormal gaze perception/behavior in social anxiety.\n\nMeasuring scan paths by means of eye-tracking is a highly ecologically valid method to assess overt gazing behavior. Hence, eye-tracking methods seem highly suitable to study avoidance of mutual gaze in individuals with social anxiety. In addition, such methods also allow studying approach-avoidance behavior in response to more ecologically valid stimuli, such as films or crowds of individuals (Lange et al., [@B32]). Ultimately, gaze measures may present objective benchmarks for the evaluation of psychotherapeutic treatment approaches for SAD. More specifically, scan paths may potentially be used as objective measures for avoidance behavior in social anxiety. Although the avoidance of mutual gaze is considered a behavioral marker of SAD (cf. Weeks et al., [@B52]), current findings are less consistent in this regard. A possible explanation for these inconsistencies may be that in most studies only *time-averaged* fixation behavior in response to direct gaze was analyzed. However, behavioral studies mainly suggest a hypervigilant-avoidant *time-course* of attention in social anxiety. In comparison to non-anxious individuals, threatening social information is detected earlier by socially anxious individuals (hypervigilance) and is followed by attentional avoidance of such stimuli (e.g., Wieser et al., [@B55]). More fine-grained analyses and paradigms might thus help to disentangle differential effects of early and late processes on fixation behavior in SAD (see also Bar-Haim et al., [@B6]; Armstrong and Olatunji, [@B4]).\n\nFurther research is needed to assess the diagnostic value of abnormalities in gaze perception as possible behavioral indicators of SAD. To date, statements regarding the diagnostic potential of such measures are substantially limited since none of the studies included a clinical comparison group, comprising for instance individuals with symptoms of autism, or schizophrenia who also exhibit abnormal gaze perception (Kliemann et al., [@B29]; Clark et al., [@B9]). It remains therefore unclear whether avoidance and fear of gaze are specific for socially anxious individuals or whether they are general signs of psychopathology and interpersonal dysfunction. Furthermore, the specific functions of gaze avoidance and its effects on states of social anxiety remain to be clarified. Langer and Rodebaugh ([@B33]) recently demonstrated avoidance of eye-to-eye contact to be an ineffective strategy for the regulation of anxiety in social phobic individuals.\n\nIn sum, recent findings highlighted abnormal gaze perceptions in social anxiety. In particular, socially anxious individuals were characterized by a greater self-referential perception of gaze direction along with a pronounced fear of direct eye contact.\n\nConflict of interest statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nJanek S. Lobmaier was supported by a grant from the Swiss National Science Foundation (grant number PP00P1_139072).\n\n^1^Note that the present review is focused specifically on *gaze perception*. For studies using eye-tracking in SAD to investigate attention mechanisms in general, see [@B4]\n\n[^1]: Edited by: Wolf-Gero Lange, Radboud University Nijmegen, Netherlands\n\n[^2]: Reviewed by: Matthias J. Wieser, University of W\u00fcrzburg, Germany; Matthias Gamer, University Medical Center Hamburg-Eppendorf, Germany\n\n[^3]: This article was submitted to the journal Frontiers in Human Neuroscience.\n"}
+{"text": "**Correction to:** *Hypertension Research* (2012) **35**, 585--591; doi:[10.1038/hr.2011.233](/doifinder/10.1038/hr.2011.233); published online 2 February 2012\n\nThe authors of the above article noticed an error in publication of this paper (online publication 2 February 2012) in the accompanying Supplementary Information. In Supplementary Table S4, female and male labels were incorrectly applied. Upon publication of the article in this issue, Supplementary Information accompanying to the html version has been rectified, and now Supplementary Table S4 is presented with correct labels.\n"}
+{"text": "Summary points\n==============\n\n1.  Barrett's oesophagus usually occurs as a consequence of chronic gastro-oesophageal reflux disease\n\n2.  The incidence of Barrett's oesophagus is increasing: the condition is present in 2% of the adult population in the West\n\n3.  The incidence of oesophageal adenocarcinoma related to Barrett's oesophagus is also increasing. In the United Kingdom, especially Scotland, oesophageal adenocarcinoma rates are higher than anywhere else in the world\n\n4.  Patients detected with early cancer related to Barrett's oesophagus might have surgically or endoscopically curable disease. Endoscopic therapy is recommended as an alternative to oesophagectomy for patients with dysplasia\n\n5.  The value of protocol based endoscopic surveillance to detect early cancer is yet to be established and is the subject of a major randomised clinical trial.\n\n6.  Other cancer prevention strategies being tested are chemoprevention of Barrett's oesophagus by aspirin in the 2513 patient AspECT trial and genome-wide identification of inherited risk factors in the 4500 patient EAGLE consortium study\n\nBarrett's oesophagus affects 2% of the adult population in the West, which makes it one of the most common premalignant lesions after colorectal polyps. Conversion to oesophageal adenocarcinoma is the most important complication of the condition, with a lifetime risk of 5% in men and 3% in women.[@ref1] [@ref2] [@ref3] [@ref4] Several large trials investigating surveillance (Barrett's Oesophagus Surveillance Study (BOSS)), chemoprevention (the Aspirin Esomeprazole Chemoprevention Trial (AspECT)), genetic stratification (EArly Genetics and Lifecourse Epidemiology (EAGLE) consortium), and endotherapy for high risk individuals are under way to determine the best way to prevent progression to adenocarcinoma.\n\nThere are now several endoscopic alternatives to the long established technique of radical surgical oesophagectomy for treating high grade dysplasia and early mucosal cancer, which avoid the mortality and morbidity of surgery. Recently consensus on optimal management of the condition was reached after a National Institute of Health and Clinical Excellence (NICE) review. It is recommended that clinicians, after discussion within the multidisciplinary team, consider offering endoscopic ablative therapy as an alternative to oesophagectomy for patients with high grade dysplasia and intramucosal cancer.[@ref1] [@ref2] [@ref5]\n\nA diagnosis of Barrett's oesophagus has important ramifications for the patient because of the uncertainty of prognosis, possible anxiety about cancer in the future, the need for repeated endoscopy in a surveillance programme, and the costs of drugs and repeated investigations.[@ref2] [@ref4] We review evidence from epidemiological studies, observational studies, and randomised trials, and draw on expert opinion to discuss the importance of early recognition and optimal treatment of Barrett's oesophagus.\n\nSources and selection criteria\n==============================\n\nWe searched Medline using the keywords \"Barrett's oesophagus,\" \"epidemiology,\" \"high grade dysplasia,\" \"medical therapy,\" \"surgery,\" \"histology,\" and \"endoscopic ablation,\" and found 12\u2009000 relevant articles. We also searched the Cochrane central register of controlled trials and the BMJ Clinical Evidence database. We went through the reference lists of articles identified from the Medline search to identify further relevant papers. Observational studies, epidemiological studies, and randomised controlled trials were extracted, and expert opinion was sought in areas where no trials existed. In addition, we consulted national and international guidelines on the management of Barrett's oesophagus, including guidelines from the British Society of Gastroenterology and the American College of Gastroenterology.\n\nTwo of the authors (JJ and HB) served on the National Institute for Health and Clinical Excellence review board for the management of dysplastic Barrett's oesophagus, which allowed us to do an extensive data search and seek independent advice on the robustness of the evidence available. All four authors are members of the consensus panel for the BArrett's Dysplasia and Cancer Taskforce (BAD CAT), which is composed of approximately 100 individuals and endorsed by 14 international societies.[@ref4] Although this taskforce has not completed its deliberations, the four authors exploited a small part of this resource to compile sections of this review.\n\nWhat is Barrett's oesophagus and who gets it?\n=============================================\n\nBarrett's oesophagus is a change in the lining of the oesophagus from normal stratified (multilayered) squamous mucosa to single layered, inflamed, premalignant, mucin secreting mucosa with variable degrees of goblet cell differentiation, termed intestinal metaplasia.[@ref3]\n\nBarrett's oesophagus develops in 5% of people with gastro-oesophageal reflux disease, which affects as many as 30% of adults in the Western world.[@ref6] [@ref7] Evidence from one case series suggests that at least 60% of patients with Barrett's oesophagus develop the disease as a result of chronic reflux, although other forms of mucosal inflammation in the lower oesophagus (such as from damage by chemotherapy, non-steroidal anti-inflammatory drugs, and viral infections) could be linked to the condition.[@ref3] [@ref7] [@ref8]\n\nCommunity studies have estimated the prevalence of Barrett's oesophagus to be just under 2% among adults in the West, which corresponds with approximately one million cases in the United Kingdom and four million in the United States. It is especially prevalent in middle aged to older men of Anglo-Saxon origin.[@ref3] [@ref8]\n\nThe annual incidence of Barrett's oesophagus in the adult population is probably around 0.1% (1 new case a year for every 1000 people)---approximately 60\u2009000 new cases in the UK and 240\u2009000 in the US a year---but evidence from case series suggests that the global rate of diagnosis of Barrett's oesophagus is increasing by 2% a year.[@ref7] [@ref8] [@ref9] This high rate may be in part because of increased endoscopic recognition, but it probably reflects a true increased incidence.[@ref7] [@ref9]\n\nConditions associated with the development of Barrett's oesophagus\n------------------------------------------------------------------\n\n1.  Chronic oesophageal reflux (\\>60% of cases)\n\n2.  Congenital retardation syndromes (1%)\n\n3.  Non-steroidal anti-inflammatory drugs (1%)\n\n4.  Chemotherapy (\\<1%)\n\n5.  Viral oesophagitis (\\<1%)\n\nWhat is the natural history of the condition?\n=============================================\n\nComplete resolution of Barrett's oesophagus rarely occurs except in very small segments, despite early reports suggesting otherwise. However, it is not uncommon to see modest shrinkage of the segment length in patients treated with acid suppression. The majority of cases stay constant, neither progressing to oesophageal adenocarcinoma nor regressing.\n\nCase series have indicated that the risk of patients with Barrett's oesophagus developing oesophageal adenocarcinoma is small in absolute terms (\\~5% lifetime risk in men and \\~3% in women).[@ref1] [@ref2] [@ref3] [@ref10] [@ref11] A recent decision analysis has suggested that in secondary referral centres this risk could be higher at 14% lifetime risk---a 30-100-fold higher risk of adenocarcinoma of the oesophagus compared with the general population's risk of 0.1% .[@ref1] [@ref2] [@ref3] [@ref11] The rates of oesophageal adenocarcinoma related to Barrett's oesophagus in west Scotland are the highest in the world (16 per 100\u2009000 population) compared with lower rates in eastern Europe, Africa, and Asia.[@ref11] [@ref12] Once a patient is in a surveillance programme, the risk of developing oesophageal adenocarcinoma varies from 0.4% a year in the US to 1% a year in the UK.[@ref11]\n\nHow does Barrett's oesophagus progress to adenocarcinoma of the oesophagus?\n===========================================================================\n\nFigure 1[](#fig1){ref-type=\"fig\"} illustrates the stages of progression of Barrett's oesophagus, from oesophagitis through metaplasia and dysplasia to adenocarcinoma.[@ref13] [@ref14] [@ref15] The sequence is thought to involve damage to stem cells deep in the oesophageal mucosa, an increase in number of abnormal but non-malignant cells, development of precancerous (dysplastic) cells, and, finally, progression to invasive cancer.\n\n![**Fig 1** The standard and alternative models of progression of Barrett's oesophagus to adenocarcinoma of the oesophagus. The standard pathway to cancer is through the oesophagitis-metaplasia-dysplasia-adenocarcinoma sequence. Recently, however, it has been recognised that submucosal glands can also develop into metaplastic cells (alternative pathway A). In addition, squamous oesophagitis can conceivably develop directly into adenocarcinoma via \"microscopic metaplasia\" without apparently transitioning through endoscopically evident metaplasia (alternative pathway B). The column on the left shows the environmental factors that help facilitate progression of the Barrett's oesophagus. The column on the right shows the genetic (blue) and epigenetic (red) changes in the evolution of cancer. APC, adenomatous polyposis coli gene](janj758763.f1_default){#fig1}\n\nThe steps of progression to cancer all involve genetic (damage to the DNA in cells) and epigenetic (reversible alterations to cell function) changes. For example, the development of metaplasia is associated with alterations in genes controlling stem cells, and progression to dysplasia is reflected by loss of heterozygosity or methylation of the adenomatous polyposis coli (APC) gene. Further progression entails loss of expression or mutations in P16 and P53, which decrease their function.[@ref13] [@ref14] However, none of these biological alterations can yet replace conventional histology for diagnosis and staging, because their exact relation with clinical progression has not been robustly tested in large randomised clinical trials.[@ref15]\n\nWhat influences the risk of developing adenocarcinoma?\n======================================================\n\nThe major factors associated with progression to cancer are: male gender; white ethnicity; length of Barrett's segment in centimetres, as seen during endoscopy (higher risk for length greater than 8 cm); diet poor in vegetables and fruit and high in fats; cigarette smoking; and obesity.[@ref3]\n\nCase-control studies have shown that symptoms of gastro-oesophageal reflux disease are associated with a significant increase in the risk of developing cancer (odds ratio 40\u00b115), but also that as many as 40% of those with adenocarcinoma do not report a history of reflux symptoms.[@ref8] [@ref9] [@ref12]\n\nHow is Barrett's oesophagus diagnosed?\n======================================\n\nCurrent evidence based guidelines on the management of dyspepsia from the National Institute for Health and Clinical Excellence advise that patients with long term symptoms of reflux (more than 5-10 years) should be referred for screening endoscopy to check for Barrett's oesophagus or its complications.[@ref16] [@ref17] On endoscopy, if the distal oesophagus looks pink or crimson in colour and is clearly distinguishable from the appearance of a hiatal hernia (fig 2[](#fig2){ref-type=\"fig\"}) using accepted criteria such as the Prague endoscopic criteria,[@ref17] then mucosal biopsies should be performed and the samples examined histopathologically. Biopsy samples are graded as \"diagnostic of Barrett's oesophagus,\" \"corroborative of Barrett's oesophagus,\" \"consistent with Barrett's oesophagus,\" or \"Barrett's oesophagus not present.\" The first three classifications should qualify the patient for entry into an endoscopic surveillance programme.[@ref18]\n\n![**Fig 2** Endoscopic image of Barrett's oesophagus. The two pictures are from the same patient but were taken five seconds apart. The panel on the right shows correct air insufflation during endoscopy, whereas the panel on the left shows the oesophagus suboptimally distended. As a consequence, the picture on the left may be misdiagnosed by inexperienced endoscopists as a hiatal hernia, because the folds in the oesophageal lining extend to the gastro-oesophageal junction (broken arrow). The panel on the right indicates circumferential Barrett's oesophagus, which can easily be seen above the folds of the hiatal hernia (solid arrow). Pictures taken with full informed written consent](janj758763.f2_default){#fig2}\n\nThe exact protocol for surveillance programmes varies, but they conventionally consist of biennial endoscopies (that is, every two years) with random circumferential biopsies, ideally four quadrants every 2 cm for flat mucosa and additional targeted biopsies for any areas that appear abnormal on endoscopy. The vast majority of patients will be assessed according to this protocol unless dysplasia is found, when more frequent intervals of endoscopy a few months apart coupled with more intensive endoscopic pinch biopsies should be used. Alternatively, those who are no longer fit for any intervention may be discharged. However, age alone should not be the sole criterion for removing patients from surveillance, because even octogenarians can cope easily with endoscopy.[@ref9] [@ref18] [@ref19] [@ref20] [@ref21] [@ref22]\n\nDoes surveillance prevent the development of adenocarcinoma?\n============================================================\n\nData from several medium sized case series suggest that patients with Barrett's oesophagus enrolled in surveillance programmes have cancer detected at an earlier (and hence more curable) stage than patients not in a surveillance programme who present with symptoms of oesophageal cancer.[@ref19] [@ref20] Other evidence suggests that most patients with cancer related to Barrett's oesophagus do not benefit from surveillance endoscopy.[@ref23] [@ref24] BOSS is a randomised trial aimed at identifying both the objective value of endoscopic surveillance and the best protocol. Data from the 2500 patient trial will be used to explore the benefits, in terms of preventing oesophageal cancer, of a regular two year upper gastrointestinal endoscopic surveillance programme versus endoscopy at time of need.[@ref20] Without evidence from randomised trials such as the BOSS trial to guide surveillance, current empirical random biopsy protocols may be suboptimal. In addition, several audits have shown that many specialists do not adhere to international surveillance guidelines.[@ref23] [@ref24]\n\nThe cost effectiveness of surveillance is still highly uncertain in the absence of real cost estimates from randomised controlled trials such as BOSS. Costs have been estimated to be about \u00a340\u2009000 (\u20ac50\u2009000; \\$60\u2009000) per cancer diagnosed for less than one quality adjusted life year (QALY) gained.[@ref25] [@ref26] The cost effectiveness is arguably better in the US. Although the country has a lower incidence of oesophageal adenocarcinoma than in the UK, endoscopic surveillance is undertaken less often (three yearly in the US *v* two yearly in the UK). In addition, in the US endoscopic surveillance is undertaken only in patients with proven intestinal metaplasia on biopsy, because such patients are threefold more likely to develop cancer than those without proven intestinal metaplasia.[@ref1] [@ref23] [@ref24] [@ref26]\n\nSurveillance related prevention of oesophageal adenocarcinoma, even if optimised, might not dramatically increase the longevity of patients because Barrett's oesophagus has also been associated with an increased risk of other potentially fatal conditions. For example, Barrett's oesophagus might be associated with obesity and gastropulmonary aspiration, which increase the risk of ischaemic heart disease and bronchopneumonia, respectively.[@ref10] The principal concern for health systems is how to manage patients at greatest risk of oesophageal cancer and distinguish them from those who are more likely to die of other causes.\n\nWhat treatments can prevent progression of Barrett's oesophagus to adenocarcinoma?\n==================================================================================\n\nCase series have suggested that as many as 10% of patients with Barrett's oesophagus develop high grade dysplasia in their lifetime.[@ref3] Cohort studies have shown that such patients have an increased risk of progression to adenocarcinoma compared with those who have non-dysplastic Barrett's oesophagus (30-55% in 8 years).[@ref18]\n\nData from several case control series indicate that management of multifocal areas of high grade dysplasia can be technically difficult and may require multiple interventions.[@ref19] [@ref20] Expert consensus indicates that because of their increased risk of cancer, such patients warrant intervention with either several sessions of endoscopic ablation therapy or, in exceptional cases, oesophagectomy.[@ref4] [@ref5] [@ref18] Arguably these patients represent a bigger burden to healthcare providers than those with cancer [@ref4].\n\nProton pump inhibitors\n----------------------\n\nA recent large randomised controlled trial found that early effective therapy for gastro-oesophageal reflux disease with proton pump inhibitors both manages symptoms effectively and heals oesophageal ulceration.[@ref27] These findings have given rise to a strategy whereby acid suppressant drugs such as proton pump inhibitors are used not only to heal and maintain healing of oesophagitis but also for \"chemoprevention\" in patients with Barrett's oesophagus. Proton pump inhibitor therapy for Barrett's oesophagus has been shown to be well tolerated and safe in both case-control studies and randomised controlled trials.[@ref28] They do not seem to promote elongation of Barrett's oesophagus, which was an initial fear following reports of hypergastrinaemia caused by proton pump inhibitors.[@ref29]\n\nHowever, case reports have speculated about a possible link between use of proton pump inhibitors and intestinal infections---especially *Clostridium difficile*---deficiencies of nutrients like folate and vitamin B~12~, and osteoporosis. Proton pump inhibitors also reduce the effectiveness of clopidogrel, and co-administration of the two drugs should be avoided if possible.\n\nSome practitioners have attempted to reduce costs and potential for side effects of proton pump inhibitors by treating patients who have gastro-oesophageal reflux disease with on demand medication.[@ref18] However, this approach might be the worst of all options because intermittent treatment could in fact increase the risk of Barrett's oesophagus and adenocarcinoma. Partial treatment might prevent the oesophagitis from healing completely and might also conceivably regulate the inflammation sufficiently for the metaplastic Barrett's cells at the ulcer base, which can tolerate a low pH, to colonise the residual ulcerated oesophageal mucosa.[@ref30] [@ref31] Selective mechanisms that allow Barrett's cells to grow preferentially in low inflammatory conditions when compared with native oesophageal squamous cells have already been demonstrated.[@ref31]\n\nDetecting significant differences between interventions for relatively rare outcomes in Barrett's oesophagus such as adenocarcinoma would need a controlled study with a very large number of subjects. Future developments in linking routine clinical data with research in the community could potentially facilitate this type of large scale study. A large randomised trial in secondary care, AspECT, ---is currently evaluating the long term value of low dose (20 mg) esomeprazole (a proton pump inhibitor) compared with high dose (80 mg) esomeprazole, either with or without aspirin.[@ref32] Aspirin is arguably the best drug to prevent cancer of the gastrointestinal tract, such as cancers of the colon, stomach, and oesophagus. So far 2513 patients have been recruited into the AspECT trial, and an interim analysis in one large centre has found a low rate of major side effects, suggesting that any interaction between esomeprazole and aspirin is acceptable.[@ref32]\n\nNissen fundoplication\n---------------------\n\nModerately sized randomised controlled trials have shown that surgical repair of the oesophageal sphincter by buttressing the stomach onto the oesophagus (fundoplication) offers good symptom control in patients with severe reflux disease and Barrett's oesophagus. In addition, this approach might be cheaper than proton pump inhibitors when drug use over many years is anticipated.[@ref27] Other randomised trials have confirmed that surgery controls reflux more completely than does medical therapy.\n\nFurthermore, fundoplication may prevent all constituents of the refluxate from entering the oesophagus, in particular the contents of the duodenum such as bile. Evidence from case series has suggested that these agents may not be suppressed by proton pump inhibitor therapy.[@ref3]\n\nNewer endoscopic therapies\n--------------------------\n\nEndoscopic mucosal resection for the eradication of early cancers (by definition confined to the mucosal lining) is highly effective---five year survival is 98% in patients with early adenocarcinoma confined to the mucosa and high grade dysplasia.[@ref1] [@ref19] [@ref21] [@ref24] [@ref25] The type of epithelium that re-grows is in part determined by the depth of injury that occurs as a consequence of treatment. In order to ensure squamous cell regeneration as opposed to recurrence of Barrett's oesophagus, some of the superficial squamous lined ducts of the oesophageal mucous glands must survive.[@ref30]\n\nPhotodynamic therapy comprises systemic administration of photosensitising agents that are retained selectively in malignant tissue. When exposed to appropriate wavelength laser light, a cytotoxic reaction occurs that causes cellular destruction. The strongest evidence for the effectiveness of photodynamic therapy comes from the five year follow-up of a randomised, multicentre, multinational, pathology blinded trial that evaluated the usefulness of the technique to eradicate dysplasia. Photodynamic therapy was significantly more effective at eradicating high grade dysplasia than omeprazole only (odds ratio 2\u00b10.7) and reduced the likelihood of developing cancer by half, with a significantly longer time to progression in the photodynamic therapy group compared with the omeprazole group.[@ref33] It may be necessary to repeat ablation at intervals, and patients treated this way should remain in lifelong surveillance.[@ref4]\n\nA further randomised trial compared thermal ablation and argon plasma coagulation with surveillance in 40 patients who had undergone surgical reflux control.[@ref34] Significant reversal of Barrett's oesophagus occurred in patients treated with argon plasma coagulation ablation (63% *v* 15% in patients under surveillance (odds ratio 4.1\u00b11.2)). Most recently, a randomised trial of radiofrequency ablation showed that this strategy is very effective in ablating both non-dysplastic and dysplastic Barrett's oesophagus, with complete eradication in 90.5% and 81.0% of cases, respectively.[@ref21] The immediate side effects of ablation are minor retrosternal discomfort in 30% of patients, but full functional activity is possible in almost all patients. Stricture, bleeding, and perforation occur in 10%, 1%, and less than 1% of patients, respectively.\n\nRecently published National Institute of Health and Clinical Excellence guidelines from the UK recommend that clinicians consider offering endoscopic ablative therapy as an alternative to oesophagectomy for people with high grade dysplasia and intramucosal cancer, according to individual patient preferences and their suitability for the procedure.[@ref4] [@ref5] National Institute of Health and Clinical Excellence guidelines consider endoscopic therapy---especially endoscopic resection and radiofrequency ablation---to be particularly suitable for patients who are considered unsuitable for surgery and those who do not wish to undergo oesophagectomy.[@ref5] [@ref21]\n\nWhat does the future hold?\n==========================\n\nConsensus has not yet been reached on the value of either tissue or blood biomarkers to stratify patients with Barrett's oesophagus in terms of risk of developing cancer.[@ref14] [@ref15] However, researchers hope that data from genome-wide association studies may assist with the understanding of the inherited basis of Barrett's oesophagus and its progression. Such knowledge might allow patient centred stratification of already known risk factors such as ethnicity, gender, and mucosal phenotype and facilitate individual tailoring of management. In fact, diagnosis and stratification may very well move to another level when the first genome-wide assessment study of Barrett's oesophagus is published in 2011. Several genetic consortiums are being set up to replicate these genetic data once published and validate them for clinical use. Perhaps the largest in Europe is the Esophageal Adenocarcinoma Genetic LinkagE (EAGLE) consortium, which incorporates both the Chemoprevention Of Premalignant Intestinal Neoplasia (ChOPIN) trial and the Inherited Predisposition of Oesophageal Diseases (IPOD) study.\n\nConclusion\n==========\n\nFrom diagnosis through to management of all stages of Barrett's oesophagus, early prompt action is important. Expert consensus and evidence based guidelines recommend that for patients with Barrett's oesophagus confirmed on histology, two yearly endoscopic surveillance is warranted along with either medical or surgical treatment to prevent gastric reflux. In patients with confirmed dysplasia, ablation therapy should be considered with endoscopic resection either alone or coupled with ablation therapy. For patients with non-dysplastic disease, the risk-benefit equation for ablation therapy has not yet determined and stratification of the likelihood of progression should be undertaken using conventional histological and endoscopic criteria. A large specialist and patient international consensus on the management of high grade dysplasia (BArretts's Dysplasia and CAncer Taskforce (BAD CAT)) is due in 2011, and the National Institute of Health and Clinical Excellence has published management guidelines this year. In the meantime patients with Barrett's oesophagus are strongly recommended to join patient support organisations with expertise in this disease, such as Fight Oesophageal Reflux Together (FORT), so they can be helped to have an informed opinion of their options at each stage in the pathway.[@ref35]\n\nTips for non-specialists\n------------------------\n\n### Who should be referred for routine endoscopy?\n\n1.  Patients with reflux for more than five years and who are aged over 50 years[@ref16] [@ref22]\n\n### What are the alarm symptoms for immediate referral for endoscopy?\n\n1.  Dysphagia[@ref16] [@ref22]\n\n2.  Weight loss[@ref16] [@ref22]\n\n3.  Vomiting blood[@ref16] [@ref22]\n\n4.  Anaemia[@ref16] [@ref22]\n\n### What other comorbid diseases should be screened for?\n\n1.  Ischaemic heart disease[@ref4] [@ref35]\n\n2.  Hypercholesterolaemia[@ref4] [@ref35]\n\n### What is the best treatment approach for patients diagnosed with Barrett's oesophagus?\n\n1.  90% can be managed by acid suppression therapy [@ref1] [@ref6] [@ref27]\n\n2.  5% may benefit from Nissen fundoplication [@ref1] [@ref6] [@ref27]\n\n3.  5% may develop oesophageal adenocarcinoma after at least 15 years [@ref1] [@ref3] [@ref27]\n\n### What dose of proton pump inhibitors should be used?\n\n1.  Use the lowest effective dose that suppresses symptoms so that heartburn occurs less than once a week[@ref16] [@ref22]\n\n### When should patients be reviewed?\n\n1.  **Primary care physician**---Dose of proton pump inhibitors should be reviewed annually, and healthy living messages---such as maintaining a low fat diet, exercising, and maintaining a BMI of less than 30---should be reinforced regularly\n\n2.  **Secondary care physician**---Patients should be reviewed endoscopically every two years in the UK and every three years elsewhere (because of higher incidence of cancer in the UK)[@ref22]\n\nAdditional educational resources\n--------------------------------\n\n### For healthcare professionals\n\n-   CORE ([www.corecharity.org.uk](www.corecharity.org.uk))---Charity specifically geared towards funding research into gastrointestinal diseases\n\n-   National Institute for Health and Clinical Excellence ([www.nice.org.uk](www.nice.org.uk))---National body providing evidence based guidance on specific diseases and conditions\n\n-   Barrett's Dysplasia and Cancer Taskforce ([www.worldgastroenterology.org/international-consensus-of-management-of-dysplastic-barretts-and-cancer.html](www.worldgastroenterology.org/international-consensus-of-management-of-dysplastic-barretts-and-cancer.html))---Taskforce producing evidence based guidelines for best clinical and cost effective management of high grade dysplasia and early mucosal cancer in Barrett's oesophagus\n\n### For patients\n\n-   Oesophageal Patients Association ([www.opa.org.uk](www.opa.org.uk))---Largest patients' support group dedicated to oesophageal cancer\n\n-   Patient UK ([www.patient.co.uk](www.patient.co.uk))---Comprehensive source of health and disease information for patients\n\n-   Fight Oesophageal Reflux Together ([refluxhelp.org](refluxhelp.org))---Largest UK patient support group, with online resources\n\n-   American College of Gastroenterology ([www.gi.org/patients/patientinfo/barretts.asp](www.gi.org/patients/patientinfo/barretts.asp))---Patient information on Barrett's oesophagus from one of the largest clinical organisations dealing with digestive care\n\n-   MacMillan Cancer Support and Cancer Backup ([www.macmillan.org.uk/Cancerinformation/Cancertypes/Gulletoesophagus/Pre-cancerousconditions/Barrettsoesophagus.aspx](www.macmillan.org.uk/Cancerinformation/Cancertypes/Gulletoesophagus/Pre-cancerousconditions/Barrettsoesophagus.aspx))---Patient information from one of the largest cancer patient information websites\n\n-   British Society of Gastroenterology ([www.bsg.org.uk/patients/patients/general/oesophageal-cancer.html](www.bsg.org.uk/patients/patients/general/oesophageal-cancer.html))---Patient information from one of the largest gastroenterology organisations in Europe\n\nWe thank Cathy Bennett of the Cochrane Collaboration Upper Gastrointestinal and Pancreatic Diseases Group and the BArrett's Dysplasia Cancer Taskforce (BAD CAT) for her help. We also thank Rebecca Harrison, Leicester, UK, for discussions during writing and proof reading.\n\nContributors: JJ came up with the concept for this article and undertook the research, writing, and coordination. HB contributed to researching and writing the article, whereas KW and BD both contributed to the writing. JJ and HB contributed equally to the manuscript. JJ acts as the guarantor and accepts full responsibility for the work, had full access to the data, and controlled the decision to publish.\n\nFunding: The authors received funding for this review from the following organisations: AstraZeneca; Cancer Research UK; the National Institute for Health Research Health Technology Assessment programme; National Institutes of Health; BAD CAT; Queen Mary University of London; and the Wellcome Trust.\n\nCompeting interests: All authors have completed the Unified Competing Interest form at [www.icmje.org/coi_disclosure.pdf](www.icmje.org/coi_disclosure.pdf) (available on request from the corresponding author) and declare: (1) Financial support from AstraZeneca; Cancer Research UK; the National Institute for Health Research Health Technology Assessment programme; National Institutes of Health; BAD CAT; Queen Mary University of London; and the Wellcome Trust for the submitted work. (2) JJ has acted as a consultant to AstraZeneca, has received educational grants, and is chief investigator for the AspECT and ChOPIN trials; HB has been a consultant for AstraZeneca and Axcan Pharma, and is chief investigator for the BOSS trial; and KW is a consultant to various companies. (3) No spouses, partners, or children with relationships with commercial entities that might have an interest in the submitted work. (4) No non-financial interests that may be relevant to the submitted work.\n\nProvenance and peer review: Commissioned; externally peer reviewed.\n\nCite this as: *BMJ* 2010;341:c4551\n"}
+{"text": "Resource table {#s0005}\n==============\n\nName of stem cell lineKCL040InstitutionKing\\'s College London, London UKDerivation teamNeli Kadeva, Victoria Wood, Glenda Cornwell, Stefano Codognotto, Emma StephensonContact person and emailDusko Ilic, email: <dusko.ilic@kcl.ac.uk>Date archived/stock dateFeb 03, 2012Type of resourceBiological reagent: cell lineSub-typeHuman pluripotent stem cell lineOriginHuman embryoKey marker expressionPluripotent stem cell markers: NANOG, OCT4, TRA-1-60, TRA-1-81, alkaline phosphatase (AP) activityAuthenticationIdentity and purity of line confirmedLink to related literature (direct URL links and full references)1)Jacquet, L., Stephenson, E., Collins, R., Patel, H., Trussler, J., Al-Bedaery, R., Renwick, P., Ogilvie, C., Vaughan, R., Ilic, D., 2013. Strategy for the creation of clinical grade hESC line banks that HLA-match a target population. EMBO Mol. Med. 5 (1), 10--17.doi: 10.1002/emmm.201201973 <http://www.ncbi.nlm.nih.gov/pubmed/23161805>\\\n2)Canham, A., Van Deusen, A., Brison, D.R., De Sousa, P., Downie, J., Devito, L., Hewitt, Z.A., Ilic, D., Kimber, S.J., Moore, H.D., Murray, H., Kunath, T., 2015. The molecular karyotype of 25 clinical-grade human embryonic stem cells lines. Sci. Rep*.* 5, 17258.doi: 10.1038/srep17258 <http://www.ncbi.nlm.nih.gov/pubmed/26607962>\\\n3)Ilic, D., Stephenson, E., Wood, V., Jacquet, L., Stevenson, D., Petrova, A., Kadeva, N., Codognotto, S., Patel, H., Semple, M., Cornwell, G., Ogilvie, C., Braude, P., 2012. Derivation and feeder-free propagation of human embryonic stem cells under xeno-free conditions. Cytotherapy. 14 (1), 122--128.doi: 10.3109/14653249.2011.623692\\\n<http://www.ncbi.nlm.nih.gov/pubmed/22029654>\\\n4)Stephenson, E., Jacquet, L., Miere, C., Wood, V., Kadeva, N., Cornwell, G., Codognotto, S., Dajani, Y., Braude, P., Ilic, D., 2012. Derivation and propagation of human embryonic stem cell lines from frozen embryos in an animal product-free environment. Nat. Protoc. 7 (7), 1366--1381.doi: 10.1038/nprot.2012.080\\\n<http://www.ncbi.nlm.nih.gov/pubmed/22722371>Information in public databasesKCL040 is a National Institutes of Health (NIH) registered hESC line\\\nNIH Registration Number: NIHhESC-14-0272\\\n<http://grants.nih.gov/stem_cells/registry/current.htm?id=678>EthicsThe hESC line KCL040 is derived under license from the UK Human Fertilisation and Embryology Authority (research licence numbers: R0075 and R0133) and also has local ethical approval (UK National Health Service Research Ethics Committee Reference: 06/Q0702/90).\\\nInformed consent was obtained from all subjects and the experiments conformed to the principles set out in the WMA Declaration of Helsinki and the NIH Belmont Report. No financial inducements are offered for donation.\n\nResource details {#s0010}\n================\n\nConsent signedSep 03, 2010Embryo thawedJan 17, 2012UK Stem Cell Bank Deposit ApprovalReference: SCSC12-37SexFemale 46, XXGradeClinicalDisease statusHealthy/UnaffectedKaryotype (aCGH)Reduced copy number at 5q13.2 (69,705,561--70,388,844).SNP ArrayCopy-neutral loss of heterozygosity (CN-LOH) at 2q11.1--11.2 (94,871,756--98,412,364), gain at 12p11.21 (31,116,366--31,248,444), loss at 16p11.2 (32,491,547--33,993,220) ([@bb0005])DNA fingerprintAllele sizes (in bp) of 16 microsatellite markers specific for chromosomes 13, 18 and 21 ([@bb0030])HLA typingHLA-A 03, 24; B 07, 15; Bw 4, 6; C 03, 07; DRB1 04, 15; DRB4 01; DRB5 01; DQB1 03, 06 ([@bb0030], [@bb0005])Viability testingPassMycoplasmaNegativeSterilityPassPluripotent markers (immunostaining)\\\n([Fig. 1](#f0005){ref-type=\"fig\"})NANOG, OCT4, TRA-1-60, TRA-1-81, AP activityThree germ layers differentiation in vitro (immunostaining)\\\n([Fig. 2](#f0010){ref-type=\"fig\"})Endoderm: AFP\\\nEctoderm: TUBB3 (tubulin, beta 3 class III)\\\nMesoderm: ACTA2 (actin, alpha 2, smooth muscle)Sibling lines availableNo\n\nWe generated KCL040 clinical grade hESC line following protocols, established previously ([@bb0025], [@bb0060]), and now adapted to cGMP conditions. The expression of the pluripotency markers was tested after freeze/thaw cycle ([Fig. 1](#f0005){ref-type=\"fig\"}). Differentiation potential into three germ layers was verified in vitro ([Fig. 2](#f0010){ref-type=\"fig\"}).\n\nMolecular karyotyping using array comparative genomic hybridization aCGH identified reduced copy number at 5q13.2 (69,705,561--70,388,844). The imbalance was not called by software. Whole-genome single nucleotide polymorphism (SNP) array analysis detected CN-LOH at 2q11.1--11.2 (94,871,756--98,412,364), gain at 12p11.21 (31,116,366--31,248,444), loss at 16p11.2 (32,491,547--33,993,220) ([@bb0005]).\n\nThis CN-LOH at 2q11.1--11.2 contains multiple genes: *TEKT4*, *MAL*, *MRPS5*, *ZNF514*, *ZNF2*, *PROM2*, *KCNIP3*, *FAHD2A*, *TRIM43*, *ANKRD36C*, *GPAT2*, *ADRA2B*, *ASTL*, *DUSP2*, *STARD7*, *TMEM127*, *CIAO1*, *SNRNP200*, *ITPRIPL1*, *NCAPH*, *NEURL3*, *ARID5A*, *KANSL3*, *FER1L5*, *LMAN2L*, *CNNM4*, *CNNM3*, *ANKRD23*, *ANKRD39*, *SEMA4C*, *FAM178B*, *FAHD2B*, *ANKRD36*, *ANKRD36B*, *COX5B*, *ACTR1B*, *ZAP70*, *TMEM131*, *VWA3B*, and *CNGA3*. Genetic size of this interstitial CN-LOH is relatively small and the double recombination event required to this to happen would be difficult to explain ([@bb0035], [@bb0045]). Therefore, it is unlikely that is acquired ([@bb0005]).\n\nThe gain on chromosome 12p11.21 was also found in KCL033. The region contains no genes and it has been also reported in at least 14 submissions at Database of Genomic Variants (DGV; <http://dgv.tcag.ca>), which has collected structural variations in more than 14,000 healthy individuals from worldwide population ([@bb0040]). Estimated frequency in the human population is 4.70% ([@bb0005]).\n\nThe loss at 16p11.2 contains three related genes *TP53TG3*, *TP53TG3C*, and *TP53TG3B* and it was reported previously in healthy population ([@bb0055], [@bb0010]). Estimated frequency in the human population is 5.14% ([@bb0005]).\n\nThe KCL040 line was negative for Human Immunodeficiency Virus 1 (HIV1), Hepatitis B (HepB, HCB), C Virus (HepC, HCV), Cytomegalovirus (CMV) and Epstein--Barr Virus (EBV) by PCR. Mycoplasma was also not detected.\n\nWe also generated research grade of KCL040 line that is adapted to feeder-free conditions.\n\nMaterials and methods {#s0015}\n=====================\n\nConsenting process {#s0020}\n------------------\n\nWe distribute Patient Information Sheet (PIS) and consent form to the in vitro fertilization (IVF) patients if they opted to donate to research embryos that were stored for 5 or 10\u00a0years. They mail signed consent back to us and that might be months after the PIS and consent were mailed to them. If in meantime new versions of PIS/consent are implemented, we do not send these to the patients or ask them to re-sign; the whole process is done with the version that was given them initially. The PIS/consent documents (FRO-V.8) were created on Mar. 11, 2010. HFEA Code of Practice that was in effect at the time of document creation: Edition 8 --- R.1 (<http://www.hfea.gov.uk/2999.html>). The donor couple signed the consent on Sep. 03, 2010. HFEA Code of Practice that was in effect at the time of donor signature: Edition 8 --- R.2. HFEA Code of Practice Edition 8 --- R.1 was in effect: Oct. 01 2009--Apr. 06, 2010, whereas 8 --- R.2 was in effect: Apr. 07, 2010--Apr. 06, 2011.\n\nEmbryo culture and micromanipulation {#s0025}\n------------------------------------\n\nEmbryo culture and laser-assisted dissection of inner cell mass (ICM) were carried out as previously described in details ([@bb0025], [@bb0060]). The cellular area containing the ICM was then washed and transferred to plates containing mitotically inactivated human neonatal foreskin fibroblasts (HFF).\n\nCell culture {#s0030}\n------------\n\nICM plated on mitotically inactivated HFF was cultured as described ([@bb0025], [@bb0060]). TE cells were removed mechanically from outgrowth ([@bb0015], [@bb0020]). hESC colonies were expanded and cryopreserved at the third passage.\n\nViability test {#s0035}\n--------------\n\nStraws with the earliest frozen passage (p.2--3) are thawed and new colonies are counted three days later. These colonies are then expanded up to passage 8, at which point cells were part frozen and part subjected to standard battery of tests (pluripotency markers, in vitro and in vivo differentiation capability, genetics, sterility, mycoplasma).\n\nPluripotency {#s0040}\n------------\n\nPluripotency in vitro was assessed using two different techniques: enzymatic activity assay \\[alkaline phosphatase (AP) assay\\] and immunostaining as described ([@bb0025], [@bb0060]).\n\nDifferentiation {#s0045}\n---------------\n\nSpontaneous differentiation into three germ layers was assessed in vitro and in vivo as described ([@bb0050], [@bb0060]). Targeted differentiation in cardiomyocytes followed the protocols described earlier ([@bb0030], [@bb9560]).\n\nGenotyping {#s0050}\n----------\n\nDNA was extracted from hESC cultures using a Chemagen DNA extraction robot according to the manufacturer\\'s instructions. Amplification of polymorphic microsatellite markers was carried out as described ([@bb0025]). Allele sizes were recorded to give a unique fingerprint of each cell line.\n\nArray comparative genomic hybridization (aCGH) {#s0055}\n----------------------------------------------\n\naCGH was performed as described in details ([@bb0025]).\n\nWhole-genome single nucleotide polymorphism (SNP) array {#s0060}\n-------------------------------------------------------\n\nSNP array was performed as described in details ([@bb0005]).\n\nHLA typing {#s0065}\n----------\n\nHLA-A, -B and -DRB1 typing was performed with a PCR sequence-specific oligonucleotide probe (SSOP; Luminex, Austin, TX, USA) hybridization protocol at the certified Clinical Transplantation Laboratory, Guy\\'s and St Thomas\\' NHS Foundation Trust and Serco Plc. (GSTS) Pathology (Guy\\'s Hospital, London, UK) as described ([@bb0030]). HLA typing was also performed independently by other group ([@bb0005]).\n\nSpecial pathology {#s0070}\n-----------------\n\nThe Doctors Laboratory London (UK) tested the line for HIV1, HepB, HepC, CMV and EBV by PCR.\n\nAuthor disclosure statement {#s0075}\n===========================\n\nThere are no competing financial interests in this study.\n\nThis work was supported by the UK Medical Research Council grants G0701172 and G0801061. We thank Dr. Yacoub Khalaf, Director of the Assisted Conception Unit of Guy\\'s and St Thomas\\' NHS Foundation Trust and his staff for supporting the research program. We are especially indebted to Prof Peter Braude and patients who donated embryos.\n\n![Expression of pluripotency markers. Pluripotency is confirmed by immunostaining (Oct4, Nanog, TRA-1-60, TRA-1-81) and alkaline phosphatase (AP) activity assay. Actin stress fibers, visualized with rhodamine-phalloidin (red), are present in both feeders and hES cell colonies, whereas AP activity (green) is detected only in hES cells. Scale bar, 50\u00a0\u03bcm.](gr1){#f0005}\n\n![Differentiation of three germ layers in vitro is confirmed by detection of markers: smooth muscle actin (red) for mesoderm, \u03b2-III tubulin (red) for ectoderm and \u03b1-fetoprotein (red) for endoderm. Nuclei are visualized with Hoechst 33,342 (blue). Scale bar, 50\u00a0\u03bcm.](gr2){#f0010}\n"}
+{"text": "![](edinbmedj74572-0087){#sp1 .66}\n\n![](edinbmedj74572-0088){#sp2 .67}\n\n![](edinbmedj74572-0089){#sp3 .68}\n\n![](edinbmedj74572-0090){#sp4 .69}\n\n![](edinbmedj74572-0091){#sp5 .70}\n\n![](edinbmedj74572-0092){#sp6 .71}\n\n![](edinbmedj74572-0093){#sp7 .72}\n\n![](edinbmedj74572-0094){#sp8 .73}\n"}
+{"text": "As a child I remember playing in a friend's somewhat dilapidated house and enjoying racing around an upstairs gallery and making out the curious shapes on the royal arms in plaster over the fireplace. Little did I know then that New Hall, Elland, had been the site of some of the earliest experiments to test Harvey's theory of the circulation, or that a hill visible from my junior school's playing field had witnessed important experiments on barometric pressure. Indeed, the name of Henry Power, doctor, scientist and one of the early members of the Royal Society, was unknown in the small town where he once lived and practised. Dr Hughes has done a great service in rescuing this important Restoration scientist from an obscurity that is more than local.\n\nThis engaging book clearly describes Power's scientific and medical work, emphasising his wide-ranging curiosity as well as his experimentalism. Indeed, given the author's own expertise, one would have liked to have had more quotations from Power's notebooks, as well as a more detailed comparison of his medical practice with that of other contemporaries. He developed a wide clientele, although several of the places cited by Dr Hughes on p. 69 are closer to New Hall than he implies, and it is unclear whether he went specifically to treat some of his farthest-flung patients or whether they were seen while Power was visiting other acquaintances or on his way to London. His move to Wakefield in 1663--4 will also have been more convenient for a wealthier clientele as well as being half a day's journey closer to the main road south. Wakefield was also at this time developing into a regional centre with claims to gentility. But Power did not live long there; he seems to have abandoned his scientific observations almost at once, perhaps because of increasing ill health, and he died there in December 1668.\n\nThere is more that can be said about Henry Power and, indeed, Dr Hughes in his earlier articles has shown how important are the Power notebooks in the British Library for an understanding of science and medicine in the Cromwellian and early Carolean period. Local pride, which compels me to point out that, *pace* p. 28, Halifax is on the tiny Hebble Brook (not river Hebden), which joins the Calder a mile or so upstream of Elland, must also acknowledge a paradox. A distinguished doctor, with metropolitan connections, chose to return, probably for family reasons, to a small and relatively isolated community, ill-served by medical men. He may have prospered, but a cynic might wonder whether his passion for experiment was not also fostered by a lack of patients and the absence of a wider local intellectual community.\n\nWith the publication of this elegant volume Dr Hughes has brought Henry Power to a broader notice, and one may hope that future inhabitants of Elland, and others, will not be as ignorant of this important scientist as I was for much of my life.\n"}
+{"text": "**To the Editor:** Hantavirus pulmonary syndrome (HPS) is characterized by fever, gastrointestinal symptoms, respiratory distress, elevated hematocrit, hypoalbuminemia, and thrombocytopenia. Most cases in North America are acquired from rodent vectors and are caused by the Sin Nombre virus. Person-to-person transmission has been reported for Andes virus ([@R1]*,*[@R2]) but not for Sin Nombre virus ([@R3]). We describe a patient with fatal hantavirus pulmonary syndrome.\n\nThe patient was a previously healthy 15-year-old Canadian girl. In the spring of 2006, she had traveled to the Santa Cruz-San Jose de Chiquitas corridor of Bolivia with her parents and siblings for a 4-week visit ([Figure](#F1){ref-type=\"fig\"}), where they stayed with family and friends on their farms. The family noted rodent droppings outside but no rodents were seen. The patient had no known exposure to rodents or rodent droppings after her return to Canada.\n\n![Map of Bolivia with an inset map of North America showing the location of British Columbia (BC) and its relation to Bolivia.](07-0708-F){#F1}\n\nOn day 26 after her return from Bolivia, the patient sought treatment at a community hospital at 6:30 [am]{.smallcaps} for malaise and mild fever. During the night before seeking treatment, she had mild confusion. Her initial blood pressure was 99/50, heart rate 97, and oxygen saturation 96% on room air. Her hemoglobin was 192 g/L (reference range 117--149), platelets 82 \u00d7 10^9^/L (reference range 165--400), and leukocyte count 7.5 \u00d7 10^9^/L (reference range 3.9--10.2). She was initially treated with 3 liters of normal saline and repeat hemoglobin tests showed a value of 206 g/L. Due to ongoing hypotension and hypoxia, she was intubated and sedated. Rocuronium was used as a paralytic agent to facilititate high pressure mechanical ventilation and maintain patient-ventilator synchrony. Dopamine and epinephrine were given as intravenous drips. Arrangements were made for the patient to be transferred to a tertiary pediatric care center for possible extracorporeal membrane oxygenation. During air transport, she had an asystolic cardiac arrest. While administering cardiopulmonary resuscitation, members of the healthcare team were exposed to a considerable volume of pulmonary edema fluid expelled from the patient's endotracheal tube. Few, if any, were able to maintain adequate protection with face shields or protective eyewear. Resuscitation efforts were unsuccessful, and the patient was pronounced dead on arrival at the tertiary care center at 7:26 [pm]{.smallcaps}. Postmortem examination showed evidence of marked pulmonary edema, diffuse alveolar damage, and lymphoid inflammation in the pulmonary interstitium.\n\nSerologic examination of an acute blood sample was immunoglobulin M (IgM) positive for Sin Nombre virus, but low optical densities indicated potential for an infection with a related hantavirus rather than Sin Nombre virus. Subsequently, reverse transcription--PCR (RT-PCR) on blood in EDTA and lung tissue followed by sequence analysis confirmed an Andes-like hantavirus infection. None of the 40 household and healthcare contacts of the patient had symptoms compatible with HPS during an 8-week monitoring period. Two contacts with nonspecific symptoms were tested and found to be negative for hantavirus-specific immunoglobulin (Ig) M and IgG and negative by RT-PCR.\n\nAdditonally, a seroprevalence survey of close contacts and assessment of level of contact was conducted. Close contacts were defined as persons who lived in the same household as the patient, were in the same enclosed space for \\>2 hours, or provided healthcare to her while she was symptomatic. Twenty-eight (62%) of 45 close contacts provided serum over the next 5 months. All serum samples were negative for Sin Nombre and Andes IgG and Sin Nombre IgM by ELISA. Fourteen (50%) of the 28 completed a self-administered questionnaire which assessed the type and intensity of contact. Of these, 12 were healthcare workers and 2 were friends. One friend had contact with the patient 3 days before she died, a friend and a healthcare worker had contact with her on the day before her death, and the rest of the healthcare workers had contact with the patient on the day she died.\n\nTo our knowledge, this is the first imported case and the tenth case of HPS reported in British Columbia, Canada, since 1994 (2005 BC Annual Summary of Reportable Diseases at [www.bccdc.org/content.php?item=33](http://www.bccdc.org/content.php?item=33)) ([@R4]). Six of these 10 cases were fatal. All cases except the 1 described here have been locally acquired Sin Nombre infections. Sin Nombre virus is endemic in the *Peromyscus maniculatus* (deer mice) population in most of British Columbia ([@R5]).\n\nWorldwide, imported cases of HPS are unusual, although HPS has been reported in countries that are in close geographic proximity or in travelers to disease-endemic areas ([@R6]*--*[@R8]). Fortunately, none of the persons exposed to the patient reported symptoms consistent with HPS during the incubation period, and none who were tested seroconverted. Seroprevalence surveys in Chile among healthcare worker contacts of patients with HPS caused by the Andes virus showed a prevalence of 0% ([@R9]). A report from Argentina showed that cases due to secondary transmission occurred mostly in non-healthcare workers after prolonged close contact in the prodromal period ([@R10]). In conclusion, we describe an imported case of fatal HPS due to an Andes-like hantavirus with no evidence of secondary transmission.\n\n*Suggested citation for this article*: Reynolds S, Galanis E, Krajden M, Morshed M, Bowering D, Abelson W, et al. Imported fatal hantavirus pulmonary syndrome \\[letter\\]. Emerg Infect Dis \\[serial on the Internet\\]. 2007 Sep \\[*date cited*\\]. Available from <http://www.cdc.gov/eid/content/13/9/1424.htm>\n\nWe thank Heinz Feldman and Harvey Artsob for coordinating the virologic workup and providing feedback on manuscript drafts; Deborrah McFadden for completing the autopsy; Bonnie Anderson for coordinating the local serosurveys; Sunny Mak for making the map used in the publication; and the public health staff at Northern Health Authority for conducting active surveillance and helping to coordinate the serosurvey.\n"}
+{"text": "Introduction {#s1}\n============\n\nMurine lung development is initiated at embryonic day 9 (E9) by separation of the lung primordium from the anterior foregut endoderm [@pone.0044871-Cardoso1], [@pone.0044871-Morrisey1]. During the pseudoglandular stage from E10.5-16.5, the bronchiolar tree is established by branching morphogenesis [@pone.0044871-Metzger1] providing the foundation of the developing lung. Moreover, the epithelium begins to differentiate into secretory (Clara), neuroendocrine (NE), or ciliated cells. During the canalicular stage (E16.5-17.5), the undifferentiated distal epithelium continues to branch and gives rise to terminal sacs. During the terminal sac stage \\[E17.5-postnatal day (P) 5\\], the sacs increase in number and type 1 and 2 alveolar epithelial cells begin to differentiate. During the alveolar stage (P5-30), the terminal sacs develop into mature alveoli [@pone.0044871-Minoo1].\n\nMolecularly, the pseudoglandular stage is marked by the expression of the transcription factor Nkx2.1 [@pone.0044871-Minoo1], [@pone.0044871-Monaghan1] and the surfactant gene, surfactant protein-c (Sftpc) in the progenitor cells of the distal epithelium. In conjunction with other transcription factors such as Gata6, Nkx2.1 appears to control transcriptional programs of the respiratory system [@pone.0044871-Morrisey1]. In the mouse lung, complete loss of either Gata6 or Nkx2-1 results in severe defects in branching morphogenesis and cell lineage differentiation [@pone.0044871-Kimura1], [@pone.0044871-Zhang1], [@pone.0044871-Que1]. As the distal progenitor epithelium grows out into the surrounding mesenchyme, a current model suggests that the progeny that are left behind in the proximal stalks change their pattern of gene expression and begin to differentiate into the various epithelial lineages. This model is supported by the downregulation of Sox9 [@pone.0044871-Perl1]--[@pone.0044871-Shu1] and Foxp1/2 expression [@pone.0044871-Shu1], [@pone.0044871-Rawlins1] in the progeny cells and the upregulation of markers characteristic of the proximal epithelium such as Sox2 [@pone.0044871-Que1]. Differentiation of epithelial cells in the airways proceeds in a proximal to distal direction, with the first evidence at E14.5. Ciliated cells, marked by the expression of Foxj1, are detected in the proximal epithelium in a \"salt and pepper\" pattern [@pone.0044871-Rawlins1]. Cells that express the Clara cell marker, Scgb1a1 (CCSP, CC10) are detected around E15.5, in a similar pattern to Foxj1 [@pone.0044871-Rawlins1].\n\nThe signaling pathways that dictate proximal-distal patterning and differentiation of both the lung endoderm and the mesoderm have been well described and include WNTs, SHH, BMPs and FGFs [@pone.0044871-Cardoso1], [@pone.0044871-Morrisey1]. Much less is known about the molecular details and the cellular pathways that are regulated by these signaling pathways. To discover genes that provide a better molecular understanding of early lung development and specification, we undertook an unbiased ENU-mutagenesis screen in mice and identified the gene Nubp1 to be required for distal progenitor survival and lung branching morphogenesis.\n\nResults {#s2}\n=======\n\nThe Mutation in Line3-2 Causes Lung Hypoplasia, Syndactyly, and Eye Cataracts {#s2a}\n-----------------------------------------------------------------------------\n\nTo identify novel genes that control lung development, we performed an ENU mutagenesis screen in mice. Embryos scored at E18.5 from line3-2 exhibited syndactyly, gross malformation and reduction in the number of digits in both hindlimbs and forelimbs ([Figure 1A, G, H](#pone-0044871-g001){ref-type=\"fig\"}) and severe hypoplasia of all lung lobes ([Figure 1B--F](#pone-0044871-g001){ref-type=\"fig\"}). Both, the limb and lung phenotypes were 100% penetrant and newborns did not survive due to rapid suffocation. Sections of E18.5 lungs revealed a significant decrease in airway number in homozygous mutants ([Figure 1I--M](#pone-0044871-g001){ref-type=\"fig\"}), concurrent with extensive airway dilation as determined by mean linear intercept analysis ([Figure 1N](#pone-0044871-g001){ref-type=\"fig\"}). The architecture of the epithelium was also disrupted. Wild type lungs showed a typical columnar epithelium, whereas the mutant epithelial cells arranged in a mushroom-like appearance and seemed to partially delaminate from the basal lamina ([Figure 1K, L](#pone-0044871-g001){ref-type=\"fig\"}).\n\n![Line3-2 reveals lung hypoplasia and limb syndactyly.\\\n(A--H) Embryos dissected at E18.5 (A) reveal that line3-2 mutants are slightly reduced in size and have significantly smaller lung lobes (B--F). (G--H) E18.5 embryos fixed and stained for bone (red) and cartilage (blue) in the forelimb demonstrate gross malformation and reduction in the number of digits in the line3-2. (I--N) Histological analysis of E18.5 lung lobes show decreased airway number, increased airway diameter and defects in distal lung epithelial morphology in line3-2. (I--L) E18.5 lungs were dissected, paraffin embedded and sectioned at 10 um. (M) Airway number and (N) the mean linear intercept (MLI) of wild-type and mutant lungs were determined from 4 different embryos at E18.5. Error bars are indicated as standard error of the mean (SEM).](pone.0044871.g001){#pone-0044871-g001}\n\nAt a lower frequency (\u223c30%) we observed eye cataracts in homozygous mutants (data not shown). None of these phenotypes were observed in heterozygous embryos.\n\nLine3-2 Mutant Lungs Exhibit a Defect in Distal Cell Specification and Disruptions in Proliferation and Apoptosis {#s2b}\n-----------------------------------------------------------------------------------------------------------------\n\nGiven the severe defect in branching morphogenesis, we first explored whether airway specification was disrupted. Since airway specification does not significantly start until E14.5, we decided to focus primarily on later developmental stages. In E18.5 lungs, proximal airway specification was assessed by immuno-labeling for acetylated-tubulin (all ciliated cells), CCSP (Clara cell secretory protein), and K5 (keratin-5; basal cells). The expression patterns of these markers were similar between mutant and wild type lungs, suggesting that proximal airway specification was not perturbed ([Figure 2A--H](#pone-0044871-g002){ref-type=\"fig\"}). In contrast, the expression of the distal airway and progenitor marker Sftpc was significantly reduced in line3-2 mutant lungs ([Figure 2I, J](#pone-0044871-g002){ref-type=\"fig\"}). The decrease in distal epithelial markers was verified by quantitative PCR (qPCR) for Sftpc, Sox9 and Foxp2 at three different embryonic stages during lung development ([Figure 2K](#pone-0044871-g002){ref-type=\"fig\"} and [Figure S1](#pone.0044871.s001){ref-type=\"supplementary-material\"}). Moreover, we noticed a decrease in expression levels of the transcription factor Gata6 and that the proximal epithelial marker Sox2 labeled nearly all airway epithelial cells by in situ hybridization in homozygous mutant lungs ([Figure S1](#pone.0044871.s001){ref-type=\"supplementary-material\"}). These results suggest that the mutation in line3-2 is important for lung development and primarily affects the distal epithelial cell population, causing a shift towards proximal epithelial specification in the mutant lung.\n\n![Line3-2 mutants show a reduction in distal epithelial cell marker Sftpc but normal proximal differentiation.\\\n(A--J) Lung sections from E18.5 embryos were analyzed by immunohistochemistry for the indicated proximal differentiation markers (A--H) as well as for the distal epithelial cell marker Sftpc (I--J) revealing no significant changes in proximal epithelial differentiation but a marked decrease in Sftpc expression (insets in each panel show higher magnification views). (K) Expression of Sftpc was quantified by qPCR at the indicated stages, confirming the reduced expression of Sftpc in mutant lungs. Error bars represent SEM from 4 different embryos.](pone.0044871.g002){#pone-0044871-g002}\n\nWe next investigated whether the mutation also impacts proliferation and/or apoptosis in epithelial cells. To our surprise, a short pulse of EdU at E18.5 indicated an increase in proliferating cells in mutant lungs compared to wild type ([Figure 3A--H](#pone-0044871-g003){ref-type=\"fig\"}). However, at E14.5 and E16.5, proliferation was reduced \u223c30% in the mutant lung compared to wild type ([Figure 3I](#pone-0044871-g003){ref-type=\"fig\"}). There was also a significant increase in apoptosis as determined by TUNEL staining ([Figure 3J--R](#pone-0044871-g003){ref-type=\"fig\"}). Around 5% of total nuclei in the mutant lung were TUNEL positive whereas apoptosis was barely detected in wild type at E14.5 ([Figure 3R](#pone-0044871-g003){ref-type=\"fig\"}).\n\n![Line3-2 mutation perturbs cell proliferation and causes increased apoptosis in the lung.\\\n(A--I) In mutant lungs, cell proliferation is decreased during the pseudoglandular stage in mutant lungs, but is higher at the terminal sac stage as detected by EdU incorporation in lung sections. (I) Data presented as percentage of total nuclei. The results are representatives of 6 different sections. (J--R) Apoptosis is markedly increased in mutant lungs. (J--R) Lungs from E14.5 embryos were sectioned and processed for TUNEL staining. (R) The number of TUNEL positive cells were determined on sections and presented as percentage of total nuclei. The results are representatives of 7 different sections. Error bars are presented as SEM. (High) indicates higher magnification view.](pone.0044871.g003){#pone-0044871-g003}\n\nTaken together, our results indicate that the mutation in line3-2 affects proliferation and cell survival, as well as expression of distal cell markers. These data suggest that the severe reduction of the lung may be due to loss of distal progenitor cells leading to reduced branching and is consistent with proximalization of the developing lung.\n\nLine3-2 has Two Point Mutations in the Nubp1 Gene on Chromosome 16 {#s2c}\n------------------------------------------------------------------\n\nTo identify the ENU-induced mutation in line3-2 we used meiotic recombination mapping to determine association between the mutant phenotype and the C57BL/6J genetic background on which the mutation was generated. This method delimited the line3-2 mutation to a 5 Mb region between 5.3 and 11.36 Mb on chromosome 16.\n\nDue to an infrequent recombination rate in this region (the interval was not further refined in examination of 76 meioses), the large number of genes within the interval and the potential for mutations in regulatory elements, we decided to use genomic sequence capture followed by high-throughput sequencing [@pone.0044871-Pyrgaki1]. Comparative analysis of the lin3-2 sequencing results with C57BL/6J wild type sequences revealed two C to A transversions at position 743nt and 803nt within the coding sequence of the gene encoding nucleotide binding protein 1 (Nubp1) ([Figure 4A](#pone-0044871-g004){ref-type=\"fig\"}). The two mutations were confirmed by genomic DNA sequencing using locus specific primer sets for the Nubp1 gene. The mutations resulted in a Threonine to Lysine and Proline to Glutamine change, respectively. These amino acids are highly conserved among human, mouse and zebrafish ([Figure 4B](#pone-0044871-g004){ref-type=\"fig\"}) and their disruption most likely would alter protein function.\n\n![Line3-2 contains two missense mutations in the Nubp1 gene.\\\n(A) Schematic representation of the genomic region around the Nubp1 gene, which contains two missense mutations at nucleotide position 743 and 803 of the coding sequence. The mutations leads to a Threonine to Lysine and Proline to Glutamine change in the amino acid sequence of Nubp1. (B) Protein sequence alignment of the indicated genes show the mutations in line3-2 to be in a highly conserved region. Red boxes and asterisks highlight mutated regions. (C--H) E14.5 embryos from a complementation cross between line3-2 and Nubp1 knock-out reveals similar defects in lung (H) and limb development (G) as lin3-2 homozygous mutants (E, F).](pone.0044871.g004){#pone-0044871-g004}\n\nTo verify that mutation of Nubp1 is responsible for the phenotypes, we obtained ES cells containing a targeted mutation in Nubp1 from EUCOMM (Nubp1^tm1(KOMP)Vlcg^, considered a definitive null allele design). After generating animals with the targeted allele, they were crossed with heterozygous line3-2 mice for a complementation assay. The resulting transheterozygous line3-2/Nubp1^tm1(KOMP)Vlcg^ embryos exhibited similar defects as line3-2 homozygous mutants (syndactyly and severely hypoplastic lungs) ([Figure 4C--H](#pone-0044871-g004){ref-type=\"fig\"}). Thus, we renamed line3-2 as *Nubp1^m1Nisw^*.\n\nNubp1^m1Nisw^ is Stably Expressed in the Distal Lung Epithelium and Disrupts the Interaction with Nubp2 {#s2d}\n-------------------------------------------------------------------------------------------------------\n\nNubp1 and its family member Nubp2 are highly conserved proteins in eukaryotes and their homologs in Saccharomyces cerevisiae (Nbp35 and Cfd1, respectively) are essential for viability [@pone.0044871-Leipe1], [@pone.0044871-Vitale1]. Vertebrate Nubp1 and Nubp2 have a high degree of amino acid similarity (51% amino acid identity and 71.5% similarity; MATCHER) and both proteins share a nucleotide binding motif near the N-terminus. Nubp1 and Nubp2 have been reported to form a hetero-complex, suggesting that these proteins function cooperatively in vivo [@pone.0044871-Netz1].\n\nIn order to test this hypothesis during lung development, we initially examined whether both genes are expressed in the same tissue. In situ hybridization for *Nubp1* and *Nubp2* showed strong expression levels for both genes in the distal lung epithelium of E11.5 wild type lungs ([Figure 5A, C](#pone-0044871-g005){ref-type=\"fig\"}). We confirmed the predominantly distal expression of Nubp1 and Nubp2 by PCR using RNA from isolated distal epithelial cells and comparing it to RNA expression from the remaining cell populations in E18.5 lungs ([Figure 5E](#pone-0044871-g005){ref-type=\"fig\"}).\n\n![Localization, expression and interaction of Nubp1.\\\n(A--E) Nubp1 and its homolog Nubp2 are expressed in the distal lung epithelium. (A, C) In situ hybridization of Nubp1 and Nubp2 on E11.5 lung whole mounts. (B, D) Q-PCR for Nubp1 and Nubp2 at different developmental stages reveals no expression differences in mutant versus wild-type lungs. (E) PCR amplification of the indicated genes on FACS sorted E18.5 distal lung epithelial cells (GFP +) versus remaining GFP (-) cells. (F) Nubp2 interacts with Nubp1 but not with the mutant protein (Nubp1-m1Nisw). A549 cells overexpressing the indicated genes were processed for immunoprecipitation and analyzed by western blotting. (G--H) Mutation in the Nubp1 gene (Nubp1-m1Nisw) does not alter its cellular localization. A549 cells were stably transfected with GFP-Nubp1 or GFP-Nubp1(m1Nisw) and processed for immunofluorescence.](pone.0044871.g005){#pone-0044871-g005}\n\nWe next determined whether the mutation in the *Nubp1* gene causes destabilization of either *Nubp1* or *Nubp2* message. To address this, we performed qPCR for *Nubp1* and *Nubp2* on wild type and *Nubp1^m1Nisw^* lungs at E14.5 and E18.5. However, we did not detect a significant change in expression of either gene in mutant compared to wild type lungs ([Figure 5B, D](#pone-0044871-g005){ref-type=\"fig\"}).\n\nTo test whether mutant Nubp1 can still interact with Nubp2, we co-expressed myc-tagged Nubp2 with either GFP-tagged wild type or *Nubp1^m1Nisw^* mutant in HeLa cells, performed immunoprecipitation with anti-GFP antibody and analyzed the immunoprecipitated lysate by western blot. As shown in [Figure 5F](#pone-0044871-g005){ref-type=\"fig\"}, Nubp2 was detected in the wild type Nubp1 precipitate as reported previously [@pone.0044871-Christodoulou1], but the mutant protein did not interact with Nubp2. However, we did not detect a difference in the cellular localization of the GFP-tagged mutant fusion protein relative to GFP-tagged wild-type Nubp1, suggesting that the mutant protein is properly localized within the cell ([Figure 5G, H](#pone-0044871-g005){ref-type=\"fig\"}).\n\nOur observations that both Nubp1 and Nubp2 are expressed in distal lung epithelium and that the Nubp1 mutation disrupts its interaction with Nubp2 raises the possibility that Nubp2 may also function in lung development.\n\nNubp1 is Required for Proper Centrosome Positioning {#s2e}\n---------------------------------------------------\n\nTo identify potential cellular pathways that may be disrupted in the Nubp1 mutant, we used a knock-down strategy in a lung epithelial cell culture line (A549). Previous studies have suggested that mouse Nubp1 and Nubp2 are responsible for regulating centrosome duplication in NIH 3T3 mouse fibroblasts [@pone.0044871-Christodoulou1]. However, the mechanism underlying Nubp1/Nubp2 function in centrosome duplication or whether these proteins act in other cellular processes remains largely unknown and has never been studied in the context of lung development.\n\nTo visualize centrosomes in wild type and Nubp1 knock-down cells we stably expressed GFP-Centrin in A549 cells [@pone.0044871-Piel1]. As shown before in fibroblasts, we observed multiple centrosomes in Nubp1 knock-down cells. Moreover, by visualizing centrosome dynamics over time we found that centrosome behavior was severly affected in Nubp1 knock-down cells. In control cells the GFP-tagged centriole pair stayed together during the entire image sequence ([Figure 6A](#pone-0044871-g006){ref-type=\"fig\"}; [Movie S1](#pone.0044871.s002){ref-type=\"supplementary-material\"}), whereas in Nubp1 knock-down cells many GFP-centrin labeled structures moved separately from each other in a seemingly random fashion ([Figure 6B](#pone-0044871-g006){ref-type=\"fig\"}; [Movie S2](#pone.0044871.s003){ref-type=\"supplementary-material\"}).\n\n![Knock-down of Nubp1 causes centrosome multiplication and alteration in centrsome dynamics.\\\n(A--B) A549 cells stably transfected with GFP-Centrin were either transfected with control or Nubp1 specific siRNA. Represented is a time sequence showing the movement of the centriole pair in control cells and the aberrant numbers, increased and irregular centrosome movements in Nubp1 knock-down cells. (C--F) Knock-down of Nubp1 leads to multiple microtubule organization points and decreased acetylated-tubulin. A549 cells were transfected with either control or Nubp1 siRNA and stained for the indicated antigens.](pone.0044871.g006){#pone-0044871-g006}\n\nSince centrosomes are nucleation centers for microtubules, we sought to determine whether the changes in centrosome numbers and positioning would also alter the microtubular network. Indeed, in Nubp1 knock-down cells there were more distinct foci from where microtubules radiated ([Figure 6C, D](#pone-0044871-g006){ref-type=\"fig\"}). We also stained the cells for acetylated-tubulin as an indicator of stable microtubules and this showed a marked decrease in Nubp1 knock-down cells ([Figure 6E, F](#pone-0044871-g006){ref-type=\"fig\"}), suggesting that Nubp1 affects microtubule dynamics.\n\nWhether the microtubular defects primarily occur due to centrosome perturbation or whether there are other cellular defects remains to be determined.\n\nNubp1 Perturbs the Distribution of Par3 and Numb Proteins in the Developing Lung {#s2f}\n--------------------------------------------------------------------------------\n\nWe next sought to determine how the mutation in Nubp1 causes centrosome misregulation in the lung. Centrosome positioning is tightly coupled with the localization of the membrane associated proteins, Numb and Par3. Numb is distributed uniformly around the cell cortex in interphase cells but during mitosis Numb becomes redistributed to a small cortical region overlying one of the centrosomes [@pone.0044871-Knoblich1]. Par3 associates with Dynein to regulate local microtubule dynamics and centrosome orientation [@pone.0044871-Schmoranzer1].\n\nTo determine whether the mutation in Nubp1 affects distribution of Par3 and Numb during lung development we performed immuno-stainings on E14.5 mutant and wild-type sections. In wild type lungs, Par3 ([Figure 7C](#pone-0044871-g007){ref-type=\"fig\"}) and Numb ([Figure 7M](#pone-0044871-g007){ref-type=\"fig\"}) localized in a typical cortical pattern whereas in Nubp1 mutant lungs, both proteins lost this pronounced staining pattern and instead became more randomly distributed throughout the distal epithelial cells ([Figure 7D, N](#pone-0044871-g007){ref-type=\"fig\"}). However, these patterns seemed to be unaffected in proximal epithelial cells in the mutant lung, consistent with our findings that Nubp1 function is primarily required in distal lung epithelial cells.\n\n![Nubp1^(m1Nisw)^ lungs show altered distribution of Par3 and Numb1.\\\n(A--R) E16.5 lungs were sectioned and stained by immunohistochemistry for the indicated proteins. (A--H) The cellular localization of the polarity protein Par3 is significantly disrupted in Nubp1^(m1Nisw)^ mutant lung. (I--R) The mutation in Nubp1^(m1Nisw)^ disrupts the cellular localization of Numb1. Note the defect is restricted to distal epithelial cells. Images K-R are higher magnifications of the regions indicated in images I and J.](pone.0044871.g007){#pone-0044871-g007}\n\nTaken together, this data suggests that Nubp1 may regulate centrosome behavior in the lung due to misregulation of Par3 and Numb at the cell cortex.\n\nDiscussion {#s3}\n==========\n\nLung development is governed by complex and highly coordinated processes that couple growth, patterning, and cell fate specification of the lung endoderm and mesenchyme in the formation of the tree-like branched organ. Many studies have explored the key developmental regulators that guide lung development including signaling molecules such as WNTs, BMPs, and SHH and transcription factors [@pone.0044871-Cardoso1], [@pone.0044871-Morrisey1]. These studies have built a robust framework within which to understand lung development. However, to fully grasp the underlying mechanisms it is necessary to also define the downstream effector genes that control the cellular processes to guide lung development. Here we have performed an unbiased forward genetic screen in mice to identify Nubp1 as an important gene for proper lung, as well as limb and eye development.\n\nNubp1 in Lung Development {#s3a}\n-------------------------\n\nNubp1 function in mammalian development had not been previously determined. During lung development, Nubp1 is expressed in the distal epithelium, which gives rise to the numerous airways. A defect in Nubp1 function, as in the *Nubp1^m1Nisw^* mouse line, interferes with the formation of distal progenitor cells as indicated by an increase in apoptosis of lung epithelial cells and a decrease in expression of the distal progenitor cell markers Sftpc, Sox9 and Foxp2. Proliferation is also significantly perturbed and the number of airways are markedly decreased. However, the differentiation of proximal epithelial cells into various cell lineages is not obviously affected.\n\nPremature differentiation of distal progenitor cells into proximal cell fates is a possible explanation for the altered airway formation in the Nubp1 mutant, however, this would not explain the increase in apoptosis and probably would change the distribution of proximal cell markers. At E18.5 proliferation is increased in the mutant lung, suggesting that some undifferentiated cells remain proliferative, whereas the wild-type lung epithelium has ceased proliferation and switched to predominantly differentiation events.\n\nThe reduced number of distal epithelial cells likely contributes to the smaller lung size in mutant lungs, yet the airway diameter seems to be increased as indicated by the mean linear intercept analysis (MLI). Two not mutually exclusive reasons may account for this phenomena. (1) The airway diameter determined by MLI reflects predominantly the diameter of the remaining, bigger proximal airways in the mutant lung. (2) The branching defect in the mutant lungs may result in an expanded distal airway due to a defect in bifurcation and therefore contributes to the general increase in airway diameter. Our immunofluorescence analysis did not detect any significant changes in proximal airway diameter between wild type and mutant lungs, however, there was a striking loss of SftpC positive distal epithelial cells. This is consistent with the first possibility, presence of proximal airways and reduction in distal airways, as a primary reason for the apparent increase in airway diameter.\n\nCellular Function of Nubp1 {#s3b}\n--------------------------\n\nOur studies also begin to address the cellular function of Nubp1 in mammalian development. Knock-down studies in tissue culture cells have demonstrated that loss of Nubp1 as well as its homologue Nubp2 result in abnormal centrosome duplication [@pone.0044871-Christodoulou1].\n\nUsing human lung epithelial cells we see a similar defect upon Nubp1 knock-down. In addition, dynamic visualization of centrosome behavior in Nubp1 knock-down cells shows that Nubp1 regulates centrosome dynamics and positioning in the cell. Nubp1 may act directly on the centrosome or it may coordinate cellular events at the cell cortex as our data show a redistribution of the cortical proteins Numb and Par3 in Nubp1 mutants. Both Numb and Par3 play a vital role in cell division and centrosome positioning. Par3, as a member of the polarity proteins, functions in asymmetrically positioning the mitotic spindle [@pone.0044871-Macara1], [@pone.0044871-EtemadMoghadam1] and recently has been demonstrated to position centrosomes in migrating cells [@pone.0044871-Schmoranzer1], whereas Numb1 is a key target for Par proteins during cell division [@pone.0044871-Lu1]. In addition, the Nubp1 homologue, MinD associates with the cortical membrane to position cytoskeletal elements that dictate bacterial cell division [@pone.0044871-Lutkenhaus1].\n\nBased on this information, Nubp1 may assist in orchestrating the proper assembly of the cortical machinery to guide centrosome positioning and cell division. Malfunction due to Nubp1 mutation could lead to abnormal behaviors such as defects in cell polarity, the number and behavior of centrosomes and ultimately to apoptosis as shown in our studies as well as previously exemplified for the yeast homologue NBP35 (Vitale, 1996, Gene).\n\nAlthough, a function of Nubp1 at the level of the centrosome cannot be ruled out, cellular localization studies of Nubp1 by Okuno [@pone.0044871-Okuno1] and our lab do not reveal a predominant centrosomal or spindle localization.\n\nMolecular Function of Nubp1 {#s3c}\n---------------------------\n\nNubp1 is a family member of small ATPases and Nubp1 has been demonstrated to form a hetero-complex with Nubp2 [@pone.0044871-Christodoulou1]. Nubp1 and Nubp2 show a similar subcellular pattern of localization [@pone.0044871-Okuno1] and both genes are co-expressed in the distal lung epithelium suggesting that these proteins act together. The *Nubp1^m1Nisw^* mutation does not affect Nubp1 expression and distribution in cells and lung tissue but it does disrupt the interaction with Nubp2. Whether a disruption in this interaction is the primary consequence of the Nubp1 mutation can only be suggested and remains to be directly addressed in the future by comparing the Nubp1 mutant phenotype to that of a Nubp2 mutation. Based on previous studies which showed similar defects in Nubp1 or Nubp2 knock-down cells [@pone.0044871-Christodoulou1], it seems possible that the activity of both proteins are required for full functionality.\n\nMany ATPases have multiple binding partners and therefore it seems unlikely that Nubp2 is the only interaction that Nubp1 mediates. Instead we propose a model where Nubp1 acts as a molecular switch to maintain cell polarity and proper centrosomal behavior. Both cellular processes seem to be essential for proper lung development as shown by recent studies on cell division orientation and asymmetric cell division during lung branching morphogenesis although the molecular outcome of this cellular function remains poorly understood [@pone.0044871-ElHashash1], [@pone.0044871-Tang1]). The Nubp1 mutant described here provides another powerful tool to shed new insights into aforementioned processes and to dissect the interconnections between cellular behaviors and lung morphogenesis.\n\nMaterials and Methods {#s4}\n=====================\n\nMouse Strains, Ethics Statement, Genotyping, and Nubp1 Mutation Identification {#s4a}\n------------------------------------------------------------------------------\n\nAll mice were bred and kept in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Steps were taken to minimize suffering. The research protocol was approved by the Institutional Animal Care and Use Committee of the University of Colorado at Denver. The mutation in Nubp1 was identified in a screen for recessive ENU-induced mutations that cause morphological changes in lung development in E18.5 embryos [@pone.0044871-GarciaGarcia1], [@pone.0044871-Kasarskis1]. Mutations were induced with N-ethyl-N-nitrosourea on a C57BL/6J genetic background and outcrossed to 129S1/SvImJ (129). Line3-2 was outcrossed for more than ten generations to establish a congenic line and no phenotypic variation was observed on the different backgrounds. The line3-2 mutation was initially mapped between Massachusetts Institute of Technology (MIT) SSLP markers D16Mit33 and D16Mit56. The mutation was ultimately identified by genomic sequence capture technology [@pone.0044871-Pyrgaki1] using Roche-Nimblegen custom designed hybrid capture microarray chip designed against unique sequence within the 5 Mb interval between 5.3 and 11.36 Mb on chromosome 16. The two mutations within the Nubp1 gene were verified by sequencing Nubp1 cDNAs generated by RT-PCR (SuperScript One-Step RT-PCR; Invitrogen) using RNA from 6 different E14.5 mutant embryos compared to wildtype control cDNA from E14.5 embryos of C57BL/6J background. A Nubp1 mouse embryonic stem (ES) cell line (JM8.F6) with a targeted mutation in Nubp1 (tm1(KOMP)Vlcg) was obtained from KOMP Repository at the University of California Davis (CloneID EPD0089_1\\_C03)and was used to generate Nubp1 knock-out mice by blastocyst injection. Genotyping was performed by PCR with lacZ primers [@pone.0044871-Liu1].\n\nSkeletal Staining, Histology, RNA and Protein Localization {#s4b}\n----------------------------------------------------------\n\nAlcian-blue and alizarin-red staining of cartilage and bone were performed as described [@pone.0044871-Barna1]. Whole-mount and section RNA in situ hybridization were performed as described [@pone.0044871-Holmes1], [@pone.0044871-Liu2]. Immunohistochemistry on cryosections was performed as described [@pone.0044871-Kim1]. Antibodies: CCSP (gift from S. Reynolds; dilution 1\u22365000), Keratin-5 (gift from S. Reynolds; dilution 1\u2236500), acetylated-Tubulin (Sigma; dilution 1\u2236200), human pro-surfactant protein-c (Seven Hills, dilution 1\u2236200), Par3 (Millipore, dilution 1\u2236100), GM130 (BD Biosciences, dilution 1\u2236200), Numb (Cell Signaling, dilution 1\u2236200), Smooth muscle actin (Sigma, dilution 1\u2236500), Cy5 labeled beta-tubulin (Sigma, dilution 1\u2236500), Hoechst (Molecular probes, dilution 1\u2236500), E-cadherin (Transduction Laboratories; dilution 1\u2236500). Immunofluorescence images were taken on a LSM510 META confocal microscope from Zeiss. EDU (1 mg/20 g body weight) was injected intraperitoneally into pregnant females 1 hr before sacrifice and detected according to the manufacturer's procedure (Invitrogen).?EDU incorporation and TUNEL staining for apoptotic cells, were quantified in sections from the indicated stages of wild-type or Nubp1 mutants, comparing the number of positive nuclei to the total number of nuclei in the field. Mean linear intercept was determined as described previously [@pone.0044871-Thurlbeck1]. To determine the number of airways, sections of mutant and wild-type lungs were obtained and the number of airways in representative areas were counted.\n\nFACS Sorting and PCR Analysis {#s4c}\n-----------------------------\n\nTransgenic mice in which epithelial GFP was expressed under the human Sftpc-promoter were harvested at E18.5, mildly homogenized in BGjB medium to obtain single cell suspension and FACS sorted for GFP expressing (+) versus non-expressing cells (-). RNA was extracted using Phenol/Chloroform and transcribed into DNA using SuperScript One-Step RT-PCR from Invitrogen. 50 ng of DNA were used for PCR analysis using the following primers: Nubp1 forward primer: CCCCAACCAGAGGCTGTGCG, reverse primer: CCGGACATCCTGGAGGGCCA; Nubp2 forward primer: CTGCCGGTGAGCGTGCAGAA, reverse primer: AGGGGCGCAGGGCTTCCATA; Sox9 forward primer: CCCTTCGTGGAGGAGGCGGA, reverse primer: CGTCGCGGAAGTCGATGGGG; Sftpc forward primer: TCCCAGGAGCCAGTTCCGCA, reverse primer: TGCCCTTCCTCCTGGCCCAG; Gapdh forward primer: CTGGCACTGCACAAGAAGATGC, reverse primer: GGGTTCCTATAAATACGGACTGCAGC.\n\nCloning, Coimmunoprecipitation and Cell Cultures {#s4d}\n------------------------------------------------\n\nFull length cDNAs for Nubp1, Nubp2 and Nubp1^m1Nisw^ were obtained by PCR from wild-type 129S1/SvImJ mice or *Nubp1^m1Nisw^* mutant embryos using the following primers. Nubp1 (wt and mutant), forward primer: 5?-actgagaattcatggaggaggcgccccac-3\u00a2, reverse primer: 5\u00a2-tgtgtctcgagtcagggactgatgagggtctcag-3\u00a2. Nubp2, forward primer: 5?- actgactcgag**atggaggctgctgccggtg**-3\u00a2, reverse primer: 5\u00a2- tgtgtctcgagtcaggagcacagggcagaca -3\u00a2. cDNAs were subcloned into pEGFP and pCMV3-myc expression vectors. Hela cells or A549 cells were obtained from ATCC and transfected with the indicated plasmids using Lipofectamine 2000 Plus (Invitrogen). For immunoprecipitation, cell lysates were incubated with 1 ug of anti-GFP antibody (Sigma, 1\u22361000 dilution) and western blotting was performed with anti-myc or anti-Nubp2 antibody (Sigma, 1\u22361000 dilution). For life cell imaging, A549 cells were transfected with a pEGFP-Centrin plasmid (kindly provided by Dr. Bornens). Stable clones were selected using G418 (4 ug/ml). Nubp1 knock-down was generated by RNAi using 4 different siRNA targets for Nubp1 which were transfected with RNAifect (Qiagen).\n\nTime-Lapse Microscopy {#s4e}\n---------------------\n\nConfocal time-lapse movies were collected on a Zeiss 7 LIVE microscope by using a C-Apochromat 40x/1.2 W Korr objective lens. Temperature was held at 37\u00b0C and CO~2~ was held at 5% by using a CTI Controller 3700 and Temperature Control 37.2 combination. Images were acquired every 2 min. Images collected were analyzed using Imaris (Bitplane) and Zen software (Zeiss). Representative single experiments are shown.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Line3-2 reveals a decrease in distal lung epithelial markers.** (A--F) Sections from lungs of E16.5 wild type and mutant embryos were prepared for in situ hybridization using the indicated genes. (G--I) Q-PCR was performed on wild type and mutant embryos for the indicated embryonic stages. Note the decrease of the distal lung epithelial markers Sox9 and Foxp2 by q-PCR and the labeling of all airways with the proximal marker Sox2 by in situ hybridization in mutant lungs. The results are representatives from at least 4 different embryos. Error bars are presented as SEM.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Confocal Time-Lapse Movie of GFP-Centrin stably expressing A549 cells and transiently transfected with control siRNA.** A549 cells stably expressing GFP-Centrin were transiently transfected with 50 nM non-silencing control siRNA for 3 days. Images were taken every 2 min. Movie plays at 7fps. Still images from this movie are shown in [Figure 6A](#pone-0044871-g006){ref-type=\"fig\"}.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Confocal Time-Lapse Movie of GFP-Centrin stably expressing A549 cells and transiently transfected with Nubp1 specific siRNA.** A549 cells stably expressing GFP-Centrin were transiently transfected with 50 nM of Nubp1 specific siRNA for 3 days. Images were taken every 2 min. Movie plays at 7fps. Still images from this movie are shown in [Figure 6B](#pone-0044871-g006){ref-type=\"fig\"}.\n\n(MOV)\n\n###### \n\nClick here for additional data file.\n\nWe gratefully acknowledge Dr. Susan Reynolds for generously providing us with antibodies, Dr. Michel Bornens for providing the GFP-Centrin expression vector and Angela Minic for excellent technical assistance. L.N. is an investigator of the Howard Hughes Medical Institute.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: CS LN. Performed the experiments: CS. Analyzed the data: CS LN. Contributed reagents/materials/analysis tools: LN. Wrote the paper: CS LN.\n"}
+{"text": "Introduction\n============\n\nGlioblastoma multiforme (GBM) is the most aggressive malignant brain tumour ([@B39]; [@B25]). Despite of numerous new strategies for targeted treatment of glioblastoma, progress in the field remains insufficient ([@B22]; [@B40]). After the surgical resection of the tumour node, the main purpose of all consequent treatment strategies is to eliminate the remaining glioblastoma cells and prevent the disease relapse being a main cause of the patient's deaths.\n\nHowever, being concentrated to destroy the glioblastoma cells, one can overlook a danger coming from a significant impairment of the surrounding normal brain tissue during the adjuvant therapy. In fact, the most dangerous trait of glioblastoma is its' active invasion into the surrounding healthy brain tissue ([@B28]; [@B4]), and the invasiveness of GBM cells and tumour development depend on not only migration capabilities of the proliferating glioblastoma cells but also structure of the surrounding normal brain tissue ([@B36]; [@B17]). One of the key invasion-related component of normal brain tissue is extracellular matrix (ECM) which occupies near 20% of its volume and serve as a main basic element of tissue structure and physiology ([@B24]). ECM is not only a physical non-specific barrier but is actively involved in cell--cell and cell--matrix interactions and signalling through the numerous ligands like chemokines, growth factors, and adhesion molecules ([@B19]). Unlike in other organs, brain ECM composes mainly of glycosylated molecules such as proteoglycans (PGs) and glycosaminoglycans (GAGs), predominantly chondroitin sulphate PGs (CSPGs) ([@B35]; [@B5]), hyaluronic acid (HA) ([@B27]; [@B19]), and heparan sulphate PGs (HSPGs) ([@B45]).\n\nAt present, most of GBM tumours are treated according to a common scheme based on maximally safe surgery, followed by a combination of radiotherapy and chemotherapy with temozolomide (TMZ) and dexamethasone (DXM) as accompanying anti-oedema drug ([@B32]; [@B31]). During the adjuvant anti-glioblastoma therapy, both the GBM cancer cells and a surrounding normal brain ECM are exposed to those drugs but their effects towards the brain ECM remain unclear.\n\nTMZ, by definition, is a drug to eliminate cancer cells and its' molecular effects on glioblastoma cells and clinical effects on GBM tumour development are well known ([@B2]; [@B44]). However, TMZ effects towards the normal brain tissue surrounding the GBM tumour and its' ECM (especially PGs) remain completely uninvestigated.\n\nMore information can be found in the literature on corticosteroid drug DXM, routinely used to prevent/treat peritumoural brain oedema during GBM chemotherapy, despite significant systemic side effects ([@B21]). DXM do affects glioblastoma cell biology *in vitro* and tumour development *in vivo*, and an overall impression denotes rather negative action of the drug for GBM prognosis and patients survival. DXM use correlates with low overall survival and progression-free survival of GBM patients ([@B34]); retrospective clinical analyses in three independent patient cohorts and mouse experimental data suggest that DXM may decrease the effectiveness of treatment and shorten patients' survival in glioblastoma, highlighting the importance of identifying alternative agents and substantiating the request for restricted use of corticosteroids in glioblastoma ([@B29]). However, as to glioma cells *in vitro*, DXM demonstrates opposite effects as it decreases TMZ-induced apoptosis in human gliobastoma T98G cells ([@B38]); inhibits glioma cell proliferation in a concentration and species-dependent manner and reduces tumour-induced angiogenesis ([@B6]), decreases MMP-2 secretion and invasiveness of human U87MG glioma cells ([@B16]) and suppresses the dispersal of GBM cells through the stimulation of fibronectin secretion and inhibition of the glioma cells motility ([@B33]). These data suggest that the negative effect of DXM treatment *in vivo* seem to be related to some unknown molecular mechanisms (possible related to microenvironmental issues) rather then its' direct action towards glioblastoma cells, however, the issue remain uninvestigated.\n\nIn this work, we aim to study effects of TMZ and/or DXM treatments on PGs expression and ECM structure in normal rat brain tissue in organotypic system *ex vivo* and animals experimental model *in vivo*.\n\nMaterials and Methods {#s1}\n=====================\n\nAnimals\n-------\n\nWistar rats aged 8, 60, and 120 days were used in the experiment on age-dependency (totally 15 animals, 5 animals/group). All other experiments *in vivo* were performed on male Wistar rats aged 9--10 weeks and weighing 200--250 g at the beginning of the experiment (totally 35 animals). Animals were housed in polycarbonate cages (36 \u00d7 50 \u00d7 28 cm) with free access to food and water, natural light/dark cycle, temperature of 25 \u00b1 1\u00b0C, humidity of 50--60% and weighed once/day. All the studied animals were adapted to the experimenter for 5 days prior to the start of the experiment. All procedures were conducted in accordance with European Communities Council Directive 2010/63/EU. All efforts were made to minimise animal suffering and to reduce the number of animals used. Animals were sacrificed by decapitation using guillotine according AVMA Guidelines for the Euthanasia of Animals ([@B1]).\n\nOrganotypic Hippocampal Slice Culture\n-------------------------------------\n\nOrganotypic hippocampal slice cultures (OHSCs) were prepared according to the previously described protocols ([@B8]) with modification ([@B26]). Briefly, neonatal Wistar rat pups (post-natal day 7--8) were decapitated, the brains were rapidly removed under aseptic conditions and placed into ice-cold Hank's balanced solution. The hippocampi were removed and placed into agarose blocks and cut rapidly with a tissue chopper into 400 \u03bcm transversal slices. The slices were transferred to collagen-coated coverslips and placed into Petri dishes containing specialised pedestals. Hank's solution was placed on the bottom of the Petri dishes to supply additional humidity. 100 uL of culture medium, consisting of 25% Hank's balanced solution, 65% DMEM, and 10% foetal bovine serum, was added to each cover slip. The OHSCs were cultivated in a 90% humidified atmosphere with 5% CO2 at 36\u00b0C. The medium was changed twice a week; the state of the OHSCs was controlled visually. At day 7 of incubation, DXM was added to the culture medium to final concentrations of 10 nM--200 \u03bcM for 24 h. To determine the effects of DXM on OHSC, the cultures were collected into RNAlater solution and used in RT-PCR analysis. All experimental procedures involving rats were approved by the Institutional Animal Care and Use Committee and performed according to the Directive 2010/63/EU.\n\nDrug Administration\n-------------------\n\nRats were randomly divided into groups (5 animals/group). The synthetic glucocorticoid agonist DXM (KRKA) was administered subcutaneously (s.c.). Two groups of rats received a single injection of DXM (0.1 or 5 mg/kg) and were sacrificed after 24 h to determine the short-time effects of the drug. Third group received 1 mg/kg of DXM daily for 1 week. TMZ-based drugs (Temodal, Temozolomide-Teva, and Temozolomide-Rus) were administrated peroral 30 mg/kg per day for 5 days. Control group received saline injections (s.c.) of the same volume as the experimental group. Animals were sacrificed by decapitation and one hemisphere from each animal was collected in RNAlater for further RT-PCR analysis, while the other one was fixed in 4% paraformaldehyde and embedded into paraffin blocks.\n\nRT-PCR Analysis\n---------------\n\nTotal RNA was extracted from the brain samples using the TRIzol Plus RNA Purification Kit (Thermo Fisher Scientific, United States) according to the manufacturer's instructions. cDNA was synthesised from 1 \u03bcg of total RNA using a First Strand cDNA Synthesis kit (Fermentas, United States) and 1/10th of the product was subjected to PCR analysis. Quantitative RT-PCR (qRT-PCR) was performed using the CFX96^TM^ Real-Time PCR Detection System (Bio-Rad, United States) and the Taq-pol (IMCB, Russia) Maxima SYBR Green/RO master mix (Thermo Fisher Scientific) under the following conditions: 95\u00b0C for 3 min, followed by 40 cycles at 95\u00b0C for 10 s, 59\u00b0C for 20 s, and 72\u00b0C for 30 s. The total reaction volume was 25 \u03bcl. The relative amount of mRNA was normalised against Gapdh mRNA, and the fold change for each mRNA was calculated by the 2^-\u0394Ct^ method. Primer sequences for rat proteoglycan genes are presented in **Table [1](#T1){ref-type=\"table\"}**.\n\n###### \n\nSequences of primers used in PCR analysis.\n\n  Gene      Sequence\n  --------- ------------------------------\n  *Sdc1*    5\u2032-GAACCCACCAGCAGGGATAC-3\u2032\n            5\u2032-CACACTTGGAGGCTGATGGT-3\u2032\n  *Gpc1*    5\u2032-GCCAGATCTACGGGGCTAAG - 3\u2032\n            5\u2032-AGACGCAGCTCAGCATACAG-3\u2032\n  *Hspg2*   5\u2032-TGATGACGAGGACTTGCTGG-3\u2032\n            5\u2032-ACACCACACTGACAACCTGG-3\u2032\n  *Vcan*    5\u2032-ATGTGGATCATCTGGACGGC-3\u2032\n            5\u2032-GTTTCGATGGTGGTTGCCTC-3\u2032\n  *Bcan*    5\u2032-AGGGGACCTCACAAGTTCTTC-3\u2032\n            5\u2032-ATTTGACTCGGGGAAAGCCC-3\u2032\n  *Cspg4*   5\u2032-ATCTGGGAGGGGGCTATTGT-3\u2032\n            5\u2032-GTACGCCATCAGAGAGGTCG-3\u2032\n  *Dcn*     5\u2032-AATGCCATCTCCGAGTGGTG-3\u2032\n            5\u2032-TTGTCGTGGAGTCGAAGCTC-3\u2032\n  *Bgn*     5\u2032-GAACAGTGGCTTTGAACCCG-3'\n            5\u2032-CCTCCAACTCGATAGCCTGG-3\u2032\n  *Lum*     5\u2032-AATTTGACCGAGTCCGTGGG-3\u2032\n            5\u2032-GCCTTTCAGAGAAGCCGAGA-3\u2032\n  *Gapdh*   5\u2032-ATGGCCTTCCGTGTTCCTAC-3\u2032\n            5\u2032-TCCAGGGTTTCTTACTCCTTGG-3\u2032\n\nImmunostaining\n--------------\n\nFor immunohistochemistry, 3,5-\u03bcm sections of formalin-fixed paraffin-embedded samples were deparaffinised in xylene twice for 5 min following stepwise rehydration in 100, 95, and 70%, ethanol for 5 min each and 5 min in deionised water. Antigen retrieval was performed at 99\u00b0C for 20 min in citrate buffer (pH 6.0). Non-specific binding was blocked with 1% BSA and 10% foetal bovine serum in phosphate buffered saline (PBST) at room temperature for 1 h. After blocking, the slides were incubated with anti-decorin (Abcam ab175404), anti-syndecan-1 (Abcam ab34164), anti-NG2 (Abcam ab83178), anti-HS (Millipore MAB2040), and anti-CS (Sigma-Aldrich C8035) primary antibodies for 1 h at room temperature. After rinsing with PBST three times for 15 min, slides were incubated with secondary antibodies anti-mouse Alexa Fluor 488 (Abcam ab150117) or anti-rabbit Alexa Fluor 647 (Abcam ab 150063) at room temperature for 1 h. After PBST rinse 3 \u00d7 10 min, slides were mounted using SlowFade Gold (Thermo Fisher Scientific, Untied States) mounting medium with DAPI and imaged on a Confocal Laser Scanning Biological Microscope Fluoview FV1000 (Olympus, United States). The images were processed using background subtraction to remove shading due to non-uniform illumination and inhomogeneous staining effects and using colour compensation to minimise the effects of spectral bleed-through among the three-color channels (red, green, and blue). Quantitative analysis of the images was performed using CellProfiler 2.2.0 software ([@B15]).\n\nStatistical Analysis\n--------------------\n\nStatistical analyses were performed using ORIGIN 8.1 software; a value of *p* \\< 0.05 was considered to indicate a statistically significant difference. Data are expressed as the means \u00b1 SD.\n\nResults\n=======\n\nConventional anti-glioblastoma chemotherapy includes TMZ as a basic \"first-line\" drug accompanied by supporting therapy with DXM to prevent brain oedema. In spite of the established effects of the drugs towards cancer cells, their effects to the surrounding normal brain tissue need additional investigation and dictate specific methodological approach. Cultured cross-sections of rat hippocampus *ex vivo* (OHSC) were used instead of GBM cell culture model *in vitro*, as it represents 3D structure of brain tissue and can be used for ECM study ([@B10]; [@B18]). In this study, DXM only but not TMZ effects were investigated in the OHSC experimental model *ex vivo* because the last one represents pro-drug and needs to be activated in the entire organism.\n\nDXM Affects PGs Expression in Organotypic Hippocampal Culture *ex vivo*\n-----------------------------------------------------------------------\n\nRat brain sections were incubated with various concentrations of DXM (0.01--200 \u03bcM) for 24 h. The concentration lies between the minimal (0.001 \u03bcM) and maximal (500 \u03bcM) concentrations described in the literature ([@B14]; [@B20]; [@B10]). Expression of main proteoglycan core proteins in the OHSC samples before and after DXM treatment was determined using RT-PCR (**Figure [1](#F1){ref-type=\"fig\"}**). Low and high doses of DXM resulted in different changes in the expression of individual PGs compared with the control organotypic culture-low dose (0.01--0.5 \u03bcM) treatments increased expression levels of glypican-1 and versican core proteins (5-fold and 10-fold, respectively), whereas high dose (50--200 \u03bcM) treatments suppressed syndecan-1 and biglycan expression (5-fold and 3-fold, respectively). The results demonstrate a complex attenuation of PGs expression in normal rat brain by different DXM concentrations in the experimental model *ex vivo* and warrant further investigation of the effects *in vivo*.\n\n![PGs expression levels in organotypic rat hippocampus cultures *ex vivo* before and after treatment with various concentrations of DXM. RT-PCR analysis, intensity of the amplified DNA fragments normalised to that of *Gapdh*. Bars represent the mean \u00b1 SD from triplicate experiments (OriginPro 8.1). Student's *t*-test, ^\u2217^*p* \\< 0.05.](fphar-09-01104-g001){#F1}\n\nProteoglycans Expression in Rat Brain Is Age- and Brain Zone-Dependent\n----------------------------------------------------------------------\n\nAs *in vivo* experiments are usually performed using 2-month old rats whereas OHSC study *ex vivo* was performed using hippocampi from 7 to 8 days old rat pups. Age-specificity as well as brain zone-specificity of the PGs expression levels were determined for a correct comparative analysis of the obtained data. Wistar rats aged 8, 60, and 120 days were used in the experiment and expression of main PG core proteins was profiled in various brain zones like hippocampus, cortex, cerebellum, and olfactory bulbs (**Figure [2](#F2){ref-type=\"fig\"}**). According to the RT-PCR data, PGs demonstrated brain zone-specific expression patterns and overall transcriptional activities of the PGs core proteins in 8-day old rat pups. During post-natal development, significant changes were revealed for both parameters, consisting in continuous overall decrease of the PGs expression in the aged brains mainly due to decorin and biglycan down-regulation and tendency to syndecan-1 and glypican-1 up-regulation (especially at the age of 60 days).\n\n![Proteoglycans expression in different brain zones of 8, 60, and 120-day old Wistar rats. RT-PCR analysis, intensity of the amplified DNA fragments normalised to that of *Gapdh* from triplicate experiments, Student's *t*-test (OriginPro 8.1). Stacked column compares the contribution of each value to a total across categories.](fphar-09-01104-g002){#F2}\n\nThe described effects were characteristic for most of the brain structures studied (hippocampus, cerebellum, and olfactory bulbs) except cortex, where moderate (1.5--2-fold) activation of the PGs expression was observed over 8--120 days development of the animals (**Figure [2](#F2){ref-type=\"fig\"}**; **Table [2](#T2){ref-type=\"table\"}**). Taking into account a long period of drug administration, younger mice (60 days) look more suitable for the planned long-term experiments. So, 2-month old Wistar rats were used for further analysis of DXM and TMZ effects on PGs expression in hippocampus and cortex of normal rat brain *in vivo*.\n\n###### \n\nExpression of proteoglycans in various Wistar rat brain zones at different age.\n\n                   days      *Sdc1*        *Gpc1*      *Hspg2*     *Vcan*      *Bcan*      *Cspg4*     *Dcn*       *Bgn*       *Lum*\n  ---------------- --------- ------------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- -----------\n  Hippocampus      **8**     **1.5** 0.2   0.33 0.05   0.08 0.00   2.0 0.2     NO          0.52 0.03   11.0 0.5    5.8 0.3     1.92 0.09\n                   **60**    3.4 0.1       5.9 0.2     0.12 0.00   1.9 0.7     0.06 0.00   1.71 0.06   3.4 0.1     2.93 0.7    0.72 0.01\n                   **120**   2.4 0.6       0.6 0.2     0.16 0.00   0.34 0.07   0.01 0.00   0.7 0.2     1.9 0.4     1.6 0.2     1.2 0.3\n  Cortex           **8**     1.17 0.06     0.2 0.02    0.05 0.00   0.91 0.04   NO          0.37 0.05   0.22 0.01   1.9 0.2     0.31 0.05\n                   **60**    1.9 0.7       2.24 0.08   0.2 0.1     0.47 0.03   0.21 0.05   1.0 0.1     3.16 0.01   1.69 0.03   0.73 0.06\n                   **120**   0.14 0.02     1.2 0.4     0.02 0.01   2.4 0.6     0.6 0.2     1.4 0.4     0.55 0.08   1.4 0.2     1.6 0.4\n  Cerebellum       **8**     2.5 0.1       0.22 0.00   0.05 0.00   2.8 0.1     NO          0.30 0.04   7.63 0.01   6.6 0.3     4.6 0.7\n                   **60**    3.9 0.6       0.9 0.1     0.05 0.01   0.15 0.02   0.08 0.01   0.30 0.06   1.2 0.1     2.0 0.2     1.1 0.1\n                   **120**   6.6 0.9       0.25 0.06   0.08 0.01   0.29 0.04   NO          0.3 0.1     1.2 0.2     1.30 0.01   0.7 0.1\n  Olfactory bulb   **8**     1.08 0.01     0.07 0.01   0.08 0.01   0.34 0.07   0.01 0.00   0.52 0.08   37.4 3.6    20.0 0.9    7.9 0.4\n                   **60**    5.9 0.7       3.4 0.1     0.08 0.02   0.6 0.03    0.01 0.00   1.66 0.01   14.5 0.5    27.7 0.9    2.46 0.05\n                   **120**   0.30 0.01     0.27 0.04   0.49 0.05   0.47 0.00   NO          0.23 0.03   3.3 0.6     5.03 0.01   0.9 0.2\n\nExpression calculated using 2\n\ndCt\n\nmethod as gene/Gapdh ratio, NO, extremely low PG expression. Average expression levels (bold) \u00b1 SD are presented (OriginPro 8.1).\n\nHigh-Dose DXM Treatment Significantly Affects Pattern and Transcriptional Activity of PG-Coding Genes in Rat Brain, Unlike Low DXM Doses or TMZ\n-----------------------------------------------------------------------------------------------------------------------------------------------\n\nThe experimental animals were subjected to the treatment with different DXM concentrations (single injection with 0.1 or 2, 5 or 5 mg/kg DXM; repetitive treatment with 1 mg/kg DXM for a week, once/day), or TMZ treatment at the regimen close to that used for GBM patients, or both in combination. DXM concentrations range was taken from the literature data and adjusted in each part of the study (*ex vivo* and *in vivo*) because of the significant difference between the tissue level of DXM in human brain tumors and its' cytotoxic concentration in cell culture ([@B23]). After that, rat brains were dissected into the different morphological zones, and proteoglycan expression was detected in hippocampus and cortex by RT-PCR analysis (**Figure [3](#F3){ref-type=\"fig\"}**). Syndecan-1, glypican-1, decorin, biglycan, and lumican were identified as the most expressed proteoglycan-coding genes both in normal rat hippocampus and cortex (**Figure [3A](#F3){ref-type=\"fig\"}**). High-dose DXM treatments (2.5 and 5 mg/kg) resulted in the increased overall transcriptional activity (**Figure [3B](#F3){ref-type=\"fig\"}**) and specific expression patterns (**Figure [3C](#F3){ref-type=\"fig\"}**) of the PGs in hippocampus in dose-dependent manner, mainly due to changes in syndecan-1 (+4-fold), glypican-1 (+3-fold), brevican (+7-fold), CSPG4/NG2 (+2-fold), decorin (-2-fold), and lumican (+3-fold) expression.\n\n![Proteoglycan mRNA expression levels in hippocampus and cortex of 2-month Wistar rats before and after treatments with DXM and/or TMZ. **(A)** mRNA expression levels of individual proteoglycan genes. **(B)** Overall transcriptional activity of the total pool of the studied PGs. **(C)** Expression pattern of the individual PG core protein-coding gene as percentage of 100%. RT-PCR analysis, intensity of the amplified DNA fragments normalised to that of *Gapdh*. Bars represent the mean \u00b1 SD from triplicate experiments, Student's *t*-test (OriginPro 8.0).](fphar-09-01104-g003){#F3}\n\nInterestingly, the PGs expression in cortex was more resistant to DXM treatment, and only highest DXM concentration (5 mg/kg) significantly increased the transcriptional activity of the genes (**Figure [3B](#F3){ref-type=\"fig\"}**). In common, long-term treatment by low doses of DXM (1 mg/kg) looks more favourable in terms of PGs expression in normal brain tissue for both hippocampus and cortex than high DXM doses.\n\nTMZ almost did not affect overall transcriptional activities of the PG core protein-coding genes in both brain zones-neither alone nor in combination with the selected DXM concentration (2.5 mg/kg) (**Figure [3B](#F3){ref-type=\"fig\"}**), although that modified PG expression patterns both in hippocampus (increase of glypican-1 and decrease of lumican expression levels) and cortex (increase of HSPG2/perlecan and CSPG4/NG2 expression levels) (**Figure [3C](#F3){ref-type=\"fig\"}**).\n\nTotally, high doses of DXM (5 and 2.5 mg/kg in a less extent) demonstrated pronounced effects towards the activation of transcriptional activity of PG-coding genes and changes in their expression pattern in rat brain tissue, while moderate DXM dose (1 mg/kg) did not affect significantly the mRNA levels of the studied PGs. TMZ alone or in combination with 2.5 mg/kg DXM treatment also did not result in significant changes in the expression of PG core proteins in normal rat brain (both hippocampus and cortex) at the transcriptional level.\n\nSuppression of the Overall Proteoglycans Expression in Normal Brain Tissue Varies for Different TMZ-Based Drugs\n---------------------------------------------------------------------------------------------------------------\n\nInterestingly, various TMZ-based drugs demonstrated different effects on the overall transcriptional activity and expression pattern of proteoglycan-related genes (**Figure [4](#F4){ref-type=\"fig\"}**). Common trend was a down-regulation of the overall proteoglycan expression in both hippocampus and cortex, up to 2-fold depending on the drug (TMZ, TMZ1, or TMZ2). The most affected PG type were HSPGs (syndecan-1, glypican-1, and perlecan), being the main contributors in the overall inhibition of the PGs expression by some of the drugs (TMZ1 and TMZ2). The demonstrated effects of the TMZ-based drugs were more pronounced for hippocampus than the cortex of normal rat brain.\n\n![Proteoglycan expression levels in hippocampus and cortex of 2-month Wistar rats before and after treatments with different TMZ-based drugs. **(A,B)** Hippocampus, **(C,D)** Cortex. **(A,C)** mRNA expression levels of individual proteoglycan genes. RT-PCR analysis, intensity of the amplified DNA fragments normalised to that of *Gapdh*, bars represent the mean \u00b1 SD from triplicate experiments, Student's *t*-test (OriginPro 8.1). **(B,D)** Overall transcriptional activity of PG core proteins in different TMZ-treated groups of animals. Stacked column compares the contribution of each value to a total across categories. Proteoglycan mRNA expression levels in hippocampus and cortex of 2-month Wistar rats before and after treatments with DXM and/or TMZ. mRNA expression levels of individual proteoglycan genes in hippocampus **(A)** and cortex **(C)**. Overall transcriptional activity of the total pool of all studied PGs in hippocampus **(B)** and cortex **(D)**. Expression pattern of the individual PG core protein-coding gene as percentage of 100%. RT-PCR analysis, intensity of the amplified DNA fragments normalised to that of *Gapdh*. Bars represent the mean \u00b1 SD from triplicate experiments (OriginPro 8.0). ^\u2217^*p* \\< 0.05.](fphar-09-01104-g004){#F4}\n\nLow Dose of DXM and TMZ Affect Proteoglycan Content in Brain ECM at Core Protein Level\n--------------------------------------------------------------------------------------\n\nTo investigate a potential effect of the treatments onto the PGs expression at the protein level, immunofluorescence analysis of the selected PG core proteins (syndecan-1, decorin and CSPG4/NG2) was performed for control rat brains and those treated with DXM (2.5 mg/kg on days 1 and 4 of the experiment), TMZ (30 mg/kg per day for 5 days) or both (**Figure [5](#F5){ref-type=\"fig\"}**; **Table [3](#T3){ref-type=\"table\"}**). These PGs were chosen from the PGs with the highest expression levels in the normal rat brain tissue (both in hippocampus and cortex), and as representatives from main proteoglycan subtypes (heparan sulphate proteoglycan, chondroitin sulphate proteoglycan and dermatan sulphate proteoglycan). The obtained results revealed differential effects of the drugs onto PG core proteins content in different brain zones. DXM treatment resulted in the decrease of decorin staining both in hippocampus and cortex (5-fold, *p* \\< 0,001 and 4,5-fold, *p* \\< 0,001, respectively), whereas TMZ significantly increased expression of syndecan-1 (3-fold, p \\< 0,001) in hippocampus but not cortex. There was a tendency for increased expression of CSPG4/NG2 in hippocampus upon the TMZ treatment as well; however, the change was not statistically significant (**Figure [5](#F5){ref-type=\"fig\"}**; **Table [3](#T3){ref-type=\"table\"}**).\n\n![Immunofluorescence analysis of PG core proteins content in rat brain before and after treatments with DXM and/or TMZ. **(A)** Immunostaining with anti-decorin, anti-CSPG4/NG2, or anti-syndecan-1 antibodies. Visualization with secondary antibodies Alexa Fluor 647 (red, for decorin and CSPG4/NG2) or Alexa Fluor 488 (green, for syndecan-1). Magnification ^\u2217^400. **(B)**. Quantitative analysis of the PG core proteins content in the control and treated rat brain tissues (CellProfiler 2.2.0 software). Student's *t*-test, ^\u2217^*p* \\< 0.05, ^\u2217\u2217^*p* \\< 0.01 and ^\u2217\u2217\u2217^*p* \\< 0.001.](fphar-09-01104-g005){#F5}\n\n###### \n\nImmunofluorescence analysis of individual PG core proteins and GAG content in hippocampi and cortex of control and DXM and/or TMZ-treated Wistar rats.\n\n                            Decorin      Syndecan-1   Cspg4/NG2    HS           CS\n  ------------- ----------- ------------ ------------ ------------ ------------ ------------\n  Hippocampus   Control     130.2 8.9    22.1 1.9     385.4 35.8   22.7 2.5     36.4 3.2\n                DXM         24.0 1.9     25.9 1.8     330.9 24.7   63.2 5.8     63.3 4.5\n                TMZ         158.5 12.4   68.9 5.4     481.0 36.1   53.1 6.3     65.0 5.2\n                TMZ + DXM   36.9 3.1     32.0 2.3     570.4 44.9   34.4 2.7     86.5 6.2\n  Cortex        Control     93.6 11.4    45.8 4.7     222.4 15.2   99.0 13.5    41.7 6.2\n                DXM         20.49 1.7    18.3 1.8     144.3 12.6   58.3 6.1     20.6 2.0\n                TMZ         44.2 7.4     22.9 2.4     220.3 21.4   41.1 5.3     46.9 5.7\n                TMZ + DXM   89.9 10.5    95.3 10.9    444.6 33.6   250.9 34.1   202.7 25.1\n\nAverage fluorescence intensity calculated using Cell Profiler 2.2.0 software. HS, heparan sulphate and CS, chondroitin sulphate.\n\nInteresting, the most profound effects were detected for the combined treatment (DXM + TMZ), with brain zone-specific characteristics. In hippocampus, significant decrease of decorin core protein content (3-4-fold, *p* \\< 0,001) along with an evident increase of CSPG4/NG2 (0,5-fold, *p* \\< 0,01) were observed. In cortex, up-regulation of CSPG4/NG2 and syndecan-1 (2-fold, *p* \\< 0,001 and 2-fold, *p* \\< 0,001, respectively) and no changes in decorin content were detected.\n\nThe results demonstrate an overall deterioration of the proteoglycan composition in the DXM and/or TMZ treated normal brain tissues at core protein level mainly due to changes in decorin, CSPG4/NG2 and syndecan-1 content. The changes in the proteoglycan core proteins upon DXM/TMZ treatment(s) stay in line with the RT-PCR data on more profound effects of the studied drugs towards hippocampus rather than cortex, and represent side effects of the conventional anti-glioblastoma therapy towards the normal rat brain tissue in the experimental system *in vivo*.\n\nBoth DXM and TMZ Significantly Increase Glycosaminoglycan Content in Rat Brain, Especially in Combination\n---------------------------------------------------------------------------------------------------------\n\nAlong with core proteins, polysaccharide chains of GAGs \\[heparan sulphate (HS) and chondroitin sulphate (CS)\\] represent an important structural and functional part of the entire proteoglycan molecules. To determine potential effects of DXM and TMZ onto polysaccharide content of normal brain tissue, specific antibodies to the polysaccharide epitopes on HS and CS molecules were used for immunofluorescence analysis of the control and DXM and/or TMZ treated rat brain tissues (**Figure [6](#F6){ref-type=\"fig\"}**).\n\n![Immunofluorescence analysis of GAGs content in rat brain before and after treatments with DXM and/or TMZ. **(A)** Immunofluorescence analysis with anti-HS and anti-chondroitin sulphate antibodies. Visualization with secondary antibodies Alexa Fluor 488 (green). Magnification ^\u2217^400. **(B)** Quantitative analysis of the GAGs content in the control and treated rat brain tissues (CellProfiler 2.2.0 software). Student's *t*-test, ^\u2217^*p* \\< 0.05, ^\u2217\u2217^*p* \\< 0.01 and ^\u2217\u2217\u2217^*p* \\< 0.001.](fphar-09-01104-g006){#F6}\n\nSeparate treatment with DXM or TMZ resulted in increase of HS and CS content in hippocampus (2-3-fold and 2-fold, respectively, *p* \\< 0,001) but not cortex. However, the most pronounced effects were observed for the combined DXM/TMZ treatments, which resulted in the significant increase of HS and CS content in hippocampus (1,5-fold, *p* \\< 0,01 and 2-2,5-fold, *p* \\< 0,001, respectively) and especially cortex (2,5-fold and 5-fold, respectively, *p* \\< 0,001).\n\nTaken together, the obtained results for the first time demonstrate an ability of DXM and TMZ to affect PG/GAG expression and composition in normal rat brain tissue in the experimental system *in vivo*. The revealed changes might contribute to transformation of normal brain ECM into tumorigenic niche, susceptible for the enhanced invasion of the residual glioma cells and tumour progression.\n\nDiscussion\n==========\n\nThe presented results demonstrate that anti-glioblastoma drugs TMZ and DXM affect the content and expression pattern of PGs in normal rat brain tissue *ex vivo* and *in vivo*. As to TMZ, in spite of the known anticancer effects of the drug, there are no published data on its' possible action onto the normal cells or tissues to compare with. A different situation is about DXM, which is often used as anti-oedema drug during anticancer therapy. Its toxic side effects are demonstrated for many different cancers including glioma ([@B34]; [@B21]), where DXM may decrease the effectiveness of treatment and shorten survival of GBM patients, substantiating the request for restricted use of corticosteroids in glioblastoma ([@B29]). However, molecular mechanisms of the DXM side effects remain unclear.\n\nIn the literature, we did not find any direct data about DXM or TMZ effects onto PGs content/composition in normal brain tissue to compare with, unlike their effect towards glioblastoma cells or GBM tumours. It is known that the expression levels of many PGs are changed significantly in glioblastoma cells and during the tumour development. For example, multiple changes occur in gliomas with a common tendency to overall activation of the expression of various PG core proteins: syndecan-1 and perlecan is significantly up-regulated in malignant glioma cell lines and GBM specimens ([@B42]); high expression of glypican-1 in glioma endothelial cells is associated with cell cycle progression ([@B30]) and high expression of perlecan is associated with poor GBM prognosis ([@B13]); brevican is highly up-regulated in gliomas and promote cancer cells motility in primary brain tumours ([@B11]); CSPG4/NG2 is expressed in gliomas in oncofoetal manner and is involved in disease progression ([@B37]); significant (3--15-fold) up-regulation of CD44 and NG2 expression during the experimental glioma development in mouse glioma model *in vivo* suggest the molecules as possible molecular markers of tumour invasion ([@B43]), however, human GBM specimens study supports NG2 but not CD44 as a possible prognostic markers of glioblastoma progression ([@B41]); CSPGs can serve as critical regulators of glioma cells invasion and play an important role in organisation of tumour microenvironment ([@B35]).\n\nThe demonstrated ability of the conventional anti-glioblastoma chemotherapy to modulate PG and GAG expression and/or composition in normal brain tissue, for the first time reveals these brain ECM molecules as potential microenvironmental biomarkers for DXM side effects or molecular targets to search for perspective anti-DXM neuroprotective drugs. From the practical point of view, it underlines a necessity for the possible revision of the current DXM use during the GBM treatment. Moderate dose (1 mg/kg) seems to be the most appropriate in terms of consistency of PGs expression in the brain of the control and treated animals and possibly should be taken as a desirable dose. It means long-term treatment with low dose of DXM could be more appropriate that single high-dose DXM injection to preserve the transcriptional activity of PG-coding genes and resulted structure of brain ECM. The results are indirectly supported by the demonstrated DXM effects in organotypic culture of animal normal brain *ex vivo*: DXM exposure after oxygen-glucose deprivation potentiates oxygen-glucose deprivation-mediated cytotoxicity in organotypic cerebellar slice cultures prepared from neonatal rat pups ([@B20]); DXM is able to suppress transcription of AVP gene during stimulation by cAMP ([@B14]).\n\nAnother interesting finding is that the high or lower doses of DXM have different \"targets\" on the complex proteoglycan molecules--whether the first one affects transcriptional activity of the PG core proteins, the last one points to polysaccharide GAG chains of the molecule. This fact makes the revealed GAG changes worse detected in the routine research or diagnostic studies, hiding this aspect of the side effects of the widely used drugs like DXM. Important aspect could be related as well to a potential contribution of the DXM-induced changes in normal brain tissue (and especially its ECM) into the neurological, psychological, and social problems for long-term survivors of glioblastoma ([@B9]). The demonstrated increase of the GAGs content in TMZ/DXM-treated normal brain tissue can be related to an unfavourable transformation of brain ECM into tumourigenic niche supporting glioma cells growth. The data stay in line with the available data on this issue: HS content is significantly increased in the relapsed GBM tumours ([@B13]); disruption of cell--cell and cell--ECM interactions by chondroitinase ABC treatment enhances the chemotherapeutic availability and sensitivity of glioma cells to TMZ ([@B12]); glioblastoma- and neuroblastome-derived CSC-like cells with high expression of decorin and lumican are resistant to TMZ suggesting a novel pivotal role for the PGs in drug resistance and cell plasticity of glioma stem cells ([@B7]); blocking of GAG function during anticancer therapy could be a perspective strategy to improve cancer treatment ([@B3]).\n\nComparative analysis of the results, obtained at the transcriptional level of PGs expression (RT-PCR), core proteins expression (IHC) and HS/CS polysaccharide content (IHC with anti-HS/CS antibodies to the polysaccharide epitopes) suggests complex multi-level molecular mechanism of DXM/TMZ effects onto PGs content and distribution in brain tissue in rat experimental model *in vivo* (**Figure [7](#F7){ref-type=\"fig\"}**). The effects are clearly dose-dependent, high DXM doses result in deregulation of the transcriptional activity of PG-coding genes, whereas low DXM doses and TMZ affect predominantly polysaccharide GAG chains of the molecules.\n\n![Schematic representation of the effects of DXM and TMZ onto PGs expression and content in normal rat brain tissue in the experimental system *in vivo*. Arrows--significant effect and dotted arrows--weak effect.](fphar-09-01104-g007){#F7}\n\nIn summary, the demonstrated disorganization of proteoglycan expression and content in normal rat brain tissue under the DXM/TMZ pressure might contribute to formation of favourable microenvironment for proliferation and invasion of the remaining post-surgery glioblastoma cells, resulting in relapse of the disease. The obtained results extend our knowledge in this field and for the first time reveals the DZM/TMZ-induced changes in PGs expression and ECM organization in normal brain tissue as one of the possible molecular mechanisms of known toxic side effects of the conventional anti-glioblastoma therapy.\n\nAuthor Contributions\n====================\n\nAT and TP carried out organotypic culture experiments *ex vivo*. AT and AS performed the animal experiments. AT and GdLB performed RT-PCR analyses. CB, AS, GK, and SA performed immunostaining. EG and AT designed the experiments and wrote the manuscript.\n\nConflict of Interest Statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** This study was funded by the Russian Science Foundation (Grant No. 16-15-10243). AT was supported by scholarship of Russian Federation President for young scientists (SP-5435.2018.4).\n\n[^1]: Edited by: Ahmed Lasfar, Rutgers, The State University of New Jersey, United States\n\n[^2]: Reviewed by: Sujuan Guo, Dana--Farber Cancer Institute, United States; Shiv K. Gupta, Mayo Clinic, United States\n\n[^3]: This article was submitted to Cancer Molecular Targets and Therapeutics, a section of the journal Frontiers in Pharmacology\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nA number of functional electrical stimulation (FES) systems have been proposed to date that are intended to allow individuals with neurological disorders to stand \\[[@B1]--[@B14]\\], with some of them applying closed-loop control strategies in order to facilitate arm-free standing \\[[@B1]--[@B5], [@B10], [@B11], [@B13], [@B14]\\]. While some of these systems have been partially successful in facilitating arm-free standing, they may not necessarily become practical applications in the near future due to a number of remaining challenges and limitations. One important challenge is that an optimal control strategy for a closed-loop controlled FES system has not been identified to date \\[[@B15]\\]. Among other reasons, this can be attributed to the fact that it is very difficult to directly test the isolated performance of an FES control strategy in human subjects. For example, contributions of other body parts that are not controlled by the FES system are oftentimes not foreseeable while differences or changes in a stimulated muscle\\'s capacity can modify or even compromise the performance \\[[@B15]\\].\n\nThe current study is set out to develop a platform for testing different control strategies for a closed-loop FES system for ankle joint control during quiet standing, by isolating this joint\\'s action in the standing posture. For this purpose, we developed a device called Inverted Pendulum Standing Apparatus (IPSA) that consists of a human-size inverted pendulum with a point mass at the height of body center of mass and a standing frame that supports the standing posture. IPSA was developed based on the notion that the dynamics of human stance can be modeled as an inverted pendulum with a point mass at the height of the body\\'s center of mass rotating around the ankle joint \\[[@B16]\\]. The ankle joint is the primary joint that controls the equilibrium of the body\\'s center of mass regardless of other joints\\' movement as the ankle joint is the first joint contacting the external environment via rigid feet segments. Thus, controlling the ankle joint is the first priority when controlling the entire body. We therefore developed a platform that focuses on stabilizing the ankle joint. Subjects with and without disability in lower-limb muscles can stand on the IPSA with support of a standing frame. In this condition, the plantar- and dorsiflexors are completely relaxed \\[[@B17]\\] such that we can test the isolated performance of a given FES control system. By regulating the FES intensity, muscle contractions of the user\\'s plantar- and dorsiflexors are modified, modulating in turn the inclination angle of the inverted pendulum. By means of this setup, only the subject\\'s muscles around the ankle joint are utilized, implying that no other body segments contribute to the behavior of the inverted pendulum. Thus, by assessing the \"performance\" of the inverted pendulum, we are also able to directly evaluate the performance of the tested FES control strategy.\n\nIt should be emphasized that a similar device has been developed by Loram\\'s team \\[[@B18], [@B19]\\] with the goal of studying human balance control in healthy people. In addition, we have reported on the use of IPSA in a pilot study, which demonstrated that a PID control strategy can successfully stabilize the human size pendulum via ankle joint control \\[[@B20]\\].\u3000In light of this, the objectives of the present study were to (1) report on the design of the IPSA and (2) demonstrate the use of the IPSA including principles of experimental setup and execution, including the feasibility of using the IPSA to test various FES controllers. For the latter objective, proportional-derivative (PD) control strategies were used with the goal of regulating the ankle joint position during standing via FES. The PD controller was chosen based on our previous studies suggesting that it can be a good model of the physiological controller for quiet standing \\[[@B21]--[@B23]\\] and that a PD-controlled FES system can ensure stable stance by modulating calf muscle contractions \\[[@B13]\\].\n\n2. Methods {#sec2}\n==========\n\n2.1. Design of IPSA {#sec2.1}\n-------------------\n\n[Figure 1](#fig1){ref-type=\"fig\"} shows a photograph of the IPSA setup. The IPSA consisted of (1) an inverted pendulum; (2) a standing frame that is used to secure the subject during the experiments; (3) two connected foot plates on which the subject stands and places his or her feet. Note that the ankle joints are aligned with the center of rotation of the foot plates, which are attached to the inverted pendulum and controlling its movement; (4) a torque transducer that measures the resultant torque at the rotational axis; and (5) a sensor system for measuring the pendulum angle. [Figure 2](#fig2){ref-type=\"fig\"} shows schematic diagrams of the major components of the IPSA. The static components of the IPSA included a base plate, a support structure, bearings, and stoppers ([Figure 2(a)](#fig2){ref-type=\"fig\"}). The base plate was a solid steel plate. Three metal tubes, a wall, and two safety stoppers were securely bolted to the base plate. The stoppers were placed to prevent the pendulum from swaying beyond the normal ankle joint range and, thus, reduced the risk of injury to the subject from excessive ankle joint movements. The moving components of the IPSA included the main shaft, end shaft, foot plates, and inverted pendulum ([Figure 2(b)](#fig2){ref-type=\"fig\"} and [2(c)](#fig2){ref-type=\"fig\"}). The main shaft had two sections: a 2\u2009in. (5.08\u2009cm) diameter section and a 4\u2009in. (10.16\u2009cm) diameter section. The latter served as the support into which the inverted pendulum was inserted and locked into place in an upright position. Two bearings supported the main shaft and one bearing supported the end shaft. The foot platform, held in place by two side supports, acted as the base for the foot plates onto which the subject\\'s feet were placed. The two foot plates were separated by 15\u2009cm and were bolted to the foot platform. The feet were secured to the foot plates using Velcro straps.\n\nThe inverted pendulum was designed so that its mass and the location of its center of mass could be changed. The mass was changed by stacking lifting weights onto the pendulum as shown in [Figure 1](#fig1){ref-type=\"fig\"}, and spacers could be used to translate the added mass to varying heights in 1\u2009in. (2.54\u2009cm) increments. Without any weights, the pendulum weighed 17.2\u2009kg but was designed to hold a maximum of 90\u2009kg of additional weight, that is, a total maximum weight of 107.2\u2009kg. In order to determine the total moment of inertia of the moving components of the IPSA about the common rotational axis, the moments of inertia of the inverted pendulum, shafts, foot platform, and side supports were individually computed and then summated, resulting in a total moment of inertia of 4.98\u2009kg\u00b7m^2^.\n\nA laser displacement sensor (LK-2500, Keyence, Japan) with an accuracy of 10\u2009*\u03bc*m was utilized to measure the angle of the IPSA\\'s inverted pendulum. A reaction torque sensor (TS11-200, Durham Instruments, Germany) with an accuracy of 0.05\u2009Nm was implemented to record the resultant torque at the rotational axis of the inverted pendulum.\n\n2.2. Experimental Performance Testing of IPSA {#sec2.2}\n---------------------------------------------\n\nFES was applied by means of an electrical stimulator (Compex Motion, Compex SA, Switzerland). 5\u2009cm \u00d7 5\u2009cm electrodes (StimTrode, Axelgaard Co., Ltd., USA), coated with hypoallergenic gel, were used to deliver the electric current to the nerves innervating the plantarflexors. Two channels of stimulation were used, one channel for each plantarflexor (right and left legs). Note that the dorsiflexors (e.g., tibialis anterior muscles) were not required to be activated since only plantarflexors are active during natural quiet standing as the body accelerates forward in the natural standing position \\[[@B20]\\]. When larger body sway fluctuations are expected (such as during perturbed standing), the dorsiflexors should be stimulated as well.\u3000The electrodes for the plantarflexors were placed along the midline of the posterior side of the calf muscles. The anode was placed above the gastrocnemius muscles and the cathode above the Achilles tendon. A rectangular, biphasic, asymmetric, charge-balanced, and monopolar stimulation waveform with a pulse duration of 300\u2009*\u03bc*s and a stimulation frequency of 40\u2009Hz were used in the experiments. The stimulation current was either fixed or controlled by the applied PD controller, depending on the experimental conditions described below.\n\nThree healthy and young individuals participated in this study (Subject A: age 22 yrs, height 180\u2009cm, and weight 80\u2009kg; Subject B: age 26 yrs, height 180\u2009cm, and weight 83\u2009kg; Subject C: age 21 yrs, height 172\u2009cm, and weight 60\u2009kg). Each subject gave written informed consent to the experimental procedure, which was approved by the local institutional ethics committee in accordance with the Declaration of Helsinki on the use of human subjects in experiments. Four conditions were tested with each subject: (1) no stimulation (i.e., voluntary control of the ankle muscles) (NOstim); (2) a PD controller with a larger derivative gain (PDstimLD); (3) a PD controller with a smaller derivative gain (PDstimSD); and (4) stimulation with a constant current (CONSTstim). \"Adequate\" controller gains for PDstimLD were selected based on our previous finding that the postural control system during quiet stance adopts a control strategy that relies heavily on velocity information \\[[@B22], [@B23]\\]. \"Inadequate\" controller gains for PDstimSD were determined by reducing the derivative gain arbitrarily. Each task lasted for 60 seconds or until the pendulum movement was interrupted by the safety stoppers. Three trials were performed for each condition. For the voluntary condition, the subject voluntarily stabilized the IPSA. In the two closed-loop conditions (PDstimLD and PDstimSD), a PD control strategy was employed with a constant offset value corresponding to the gravity toppling torque of the IPSA ([Table 1](#tab1){ref-type=\"table\"}). In CONSTstim, the average stimulation intensity during PDstimLD was used as the constant stimulation intensity ([Table 1](#tab1){ref-type=\"table\"}). The first trials were performed in the order of NOstim, PDstimLD, PDstimSD, and CONSTstim to identify the stimulation intensity for CONSTstim, whereas the remaining 2\u22174 trials were randomized. Sufficient resting time was provided in between two trials. Prior to the testing, the subjects were allowed to familiarize themselves with both the IPSA and the electrical stimulation. Since the subjects were mechanically supported by the standing frame, the plantarflexors can be completely relaxed during standing \\[[@B17]\\]. For the FES conditions, subjects were instructed not to use voluntary efforts to control the inverted pendulum. Although we believe that voluntary and reflexive contributions to controlling the pendulum were minimal during the FES tasks, this was not verified.\n\nThe pendulum angle and torque (sampled at 100\u2009Hz and low-pass filtered using an 8th-order Butterworth filter with a cutoff frequency of 10\u2009Hz), measured using a customized LabVIEW program (version 8.0, National Instruments, TX, USA), were used to assess the performance within each condition. In cases of successfully stabilized trials (i.e., trials that lasted for 60 seconds), we calculated the root mean square (RMS) of the measured pendulum angle and torque and compared the averages of the three trials per subject across the different conditions.\n\n3. Results {#sec3}\n==========\n\n[Figure 3](#fig3){ref-type=\"fig\"} exemplifies the angle and torque fluctuation of the pendulum during the four different conditions. For the CONSTstim and PDstimSD conditions, the generated ankle torque was inadequate for stabilizing the system, causing the pendulum to drop. The unstable behavior was observed for all 18 trials (3 subjects, 3 trials, and 2 conditions) associated with these two conditions. The time duration from trial start to pendulum drop was 8.10 \u00b1 2.12 seconds for CONSTstim and 13.21 \u00b1 3.23 seconds for PDstimSD (mean \u00b1 one standard deviation for all subjects).\n\nFor both the NOstim and PDstimLD conditions, the generated ankle torque dynamically stabilized the pendulum for the entire trial length of 60 seconds. This behavior was observed for all 18 trials (3\u22173\u22172) associated with these two conditions. [Table 2](#tab2){ref-type=\"table\"} presents the standard deviations of the measured angle fluctuation for all subjects and both conditions. The standard deviations indicate that the pendulum in PDstimLD was stabilized in a similar manner as in NOstim. These experiments demonstrate that, by examining the angle and torque fluctuations of the IPSA pendulum, we can evaluate the ability of the controller to adequately regulate plantarflexor activity.\n\n4. Discussion {#sec4}\n=============\n\nThe purpose of this study was twofold. First, we reported on the design and setup of IPSA. Second, we tested IPSA, showing that it ensures the safety of the subject and at the same time keeps the subject in the standing posture as required for evaluating FES control systems. We demonstrated that PDstimLD stabilized the subject\\'s body sufficiently well (i.e., equivalently well to NOstim) while PDstimSD and CONSTstim did not, probably due to the use of inappropriate control strategies (i.e., insufficient amount of derivative gain or not facilitating time-varying calf muscle contractions). These results suggest that IPSA can be a useful platform for designing FES control systems in a well-controlled manner. That is, using the IPSA, we can directly assess the performance of a particular FES controller for ankle joint regulation while eliminating the effect of contributions from other body segments. Using the IPSA for studying different FES control strategies for standing has the following benefits: (1) supported by the standing frame, individuals with paralysis in their lower limbs can safely participate in the testing; (2) since the standing frame allows the lower limb muscles to be completely relaxed during standing \\[[@B17]\\], also able-bodied individuals can participate in the testing; (3) since the weight of the inverted pendulum can be gradually increased, the performance of different FES control strategies can be tested without the risk of damaging joints and/or falling; and (4) in contrast to testing an FES control strategy in supine or prone position, we can accurately replicate natural muscle damping and stiffness during quiet standing, which has been shown to significantly contribute to the control mechanism of that postural task \\[[@B21], [@B24]\\].\n\nAlso Loram\\'s team designed an inverted pendulum that was stabilized via ankle torque modulation during standing \\[[@B18], [@B19]\\]. While their system was capable of measuring pendulum torque and angle as well, their body support mechanism did not include passive knee stabilization for ensuring complete elimination of natural plantarflexor activity during standing \\[[@B17]\\]. In addition, the device did not feature safety stoppers and lacked flexibility with respect to weight application. These differences are not surprising as their system was not intended to be used for developing FES control strategies in different populations but rather to study human balance control in healthy people.\n\nIn this study, we solely focused on the ankle joint and its movement in the anterior-posterior direction using an artificial inverted pendulum. The reason for choosing the ankle joint was that it plays a key role in maintaining balance of the entire body, being the closest joint to the contact surface during standing \\[[@B17], [@B22]--[@B25]\\]. In addition, we adopted an artificial pendulum rather than a human-body pendulum (with braces for all joints except for the ankle joint \\[[@B4]\\]) in order to ensure the rigidity of the pendulum. Such a setup also allows researchers to modify the load relatively easily, which is beneficial for investigating adequate FES control strategies. However, to realize an FES controller for arm-free standing, it is obvious that we need to also examine adequate FES control strategies for other joints (such as knee and hip \\[[@B7], [@B8], [@B14]\\]) and other degrees of freedom (such as the medial-lateral direction \\[[@B25]\\]). For this purpose, more sophisticated systems such as the multipurpose rehabilitation frame developed by Matjaci\u0107 et al. \\[[@B26]\\] are more adequate as they allow studying other joints and joint activity in lateral movements. Another challenge that could be encountered when developing a practical FES controller for arm-free standing is the effect of residual muscle activity, which could happen, for example, in case of individuals with incomplete spinal cord injury. As the IPSA with its standing frame provides an environment that allows zero muscle activity \\[[@B17]\\], it can be used to develop FES controller without residual muscle activity. At the same time, residual muscle activity, if present, can pose an unexpected disturbance for the developed FES controller. While there is no conclusive solution for this issue at this time, we demonstrated that a PD control strategy facilitated a patient\\'s stability during quiet standing who exhibited residual muscle activity \\[[@B13]\\]. Nevertheless, future studies are needed to investigate its effect in more detail.\n\n5. Conclusion {#sec5}\n=============\n\nThe IPSA was developed to help engineers and scientists test, characterize, and select controllers that are suitable for closed-loop FES control. As demonstrated in this paper, the IPSA represents an adequate platform for studying FES control of standing in general and for ankle joint regulation in particular. However, the IPSA can also be used for examining FES control strategies intended for other applications and for the regulation of other joints, for example, to characterize temporal dynamics, noise rejection, or fatigue compensation during FES.\n\nThis work was supported by the Canadian Institutes of Health Research (MOP-69003), the Natural Sciences and Engineering Research Council of Canada (249669), the Ministry of Health and Long Term Care in Ontario, and the Toronto Rehabilitation Institute, University Health Network. Each subject participating in the experiment gave written informed consent to the experimental procedure, which was approved by the Research Ethics Board at Toronto Rehabilitation Institute, University Health Network (REB12-011-DE). The authors would like to thank Navid Javadi for his assistance throughout the study.\n\nCONSTstim:\n\n:   Stimulation condition with a constant current\n\nFES:\n\n:   Functional electrical stimulation\n\nIPSA:\n\n:   Inverted Pendulum Standing Apparatus\n\nNOstim:\n\n:   No stimulation condition\n\nPD:\n\n:   Proportional-derivative\n\nPDstimLD:\n\n:   Stimulation condition with a PD controller with a large derivative gain\n\nPDstimSD:\n\n:   Stimulation condition with a PD controller with a small derivative gain\n\nRMS:\n\n:   Root mean square.\n\nConflict of Interests\n=====================\n\nThe authors declare that there is no conflict of interests regarding the publication of this paper.\n\n![Overall view of the IPSA setup. Depicted are (1) the inverted pendulum (with a shaft fitted to accept additional weights); (2) the standing frame; (3) the subject in a standing posture; (4) foot pedals on which the subject stands; and (5) the main shaft of the IPSA which translates the torque generated by the subject to the pendulum\\'s axis of rotation.](ISRN2014-192097.001){#fig1}\n\n![(a) Static components of the IPSA: this figure illustrates the components of the IPSA that remain fixed in place throughout the experiments: the base plate, metal tubes, bearings, and stoppers; (b) moving components of the IPSA: this figure illustrates the main shaft, end shaft, and the foot plates; (c) inverted pendulum: this figure illustrates the dimensions of the inverted pendulum, which locks into place through the two notches on the 4\u2009in. (10.16\u2009cm) end of the main shaft.](ISRN2014-192097.002){#fig2}\n\n![Examples of the pendulum fluctuation (angle and torque) during the four different conditions (single subject): (1) voluntary control---NOstim (top graph); (2) adequate controller---PDstimLD (second graph from top, using a P-gain of 100\u2009Nm/rad and a D-gain of 85\u2009Nm\u00b7s/rad); (3) inadequate controller---PDstimSD (third graph from top, using a P-gain of 100\u2009Nm/rad and a D-gain of 30\u2009Nm\u00b7s/rad); and (4) constant stimulation---CONSTstim (bottom graph, using a stimulation amplitude of 33\u2009mA). The left column shows the angular displacements and the right column the measured torque. Positive angles and torques in the panels correspond to plantarflexor contractions of the subject. For the CONSTstim and PDstimSD conditions, the pendulum movement was interrupted by the built-in safety stoppers after approximately 10 seconds into the trials.](ISRN2014-192097.003){#fig3}\n\n###### \n\nGains for the PD controller in the PDstimLD and PDstimSD conditions and stimulation intensity in the CONSTstim condition. P-gain stands for the proportional gain and D-gain stands for the derivative gain of the PD controller.\n\n  Subject   PDstimLD gains   PDstimSD gains   CONSTstim        \n  --------- ---------------- ---------------- ----------- ---- ----\n  A         100              85               100         30   33\n  B         110              80               110         50   33\n  C         70               60               70          20   34\n\n###### \n\nThe average standard deviation values of the pendulum angle and torque fluctuations for the PDstimLD and NOstim conditions for each subject. Each value represents the mean \u00b1 one standard deviation for three trials in each condition.\n\n  Subject   Angle standard deviation \\[deg\\]   Torque standard deviation \\[Nm\\]                 \n  --------- ---------------------------------- ---------------------------------- ------------- -------------\n  A         0.504 \u00b1 0.222                      0.382 \u00b1 0.121                      1.38 \u00b1 0.55   0.96 \u00b1 0.23\n  B         0.565 \u00b1 0.194                      0.621 \u00b1 0.510                      1.41 \u00b1 0.48   1.64 \u00b1 1.21\n  C         1.486 \u00b1 0.411                      0.418 \u00b1 0.126                      3.55 \u00b1 1.09   1.05 \u00b1 0.29\n\n[^1]: Academic Editor: Silvia Conforto\n"}
+{"text": "1.. Introduction {#s1}\n================\n\nSocial species inevitably experience within-group competition. These social conflicts may escalate into aggression leading to negative consequences, which range from the risk of injury to potential damage to social relationships and, as a consequence, disruption of group cohesion\u00a0\\[[@RSOS171553C1]\\]. One way to cope with these negative effects is to repair the damaged relationships after conflicts. De Waal & van Roosmalen\u00a0\\[[@RSOS171553C2]\\] first used the term 'reconciliation\\' to refer to affiliative contacts between former opponents, which occur soon after the end of an agonistic conflict and may help to repair a relationship damaged by the aggressive encounter.\n\nSince the original study documenting the occurrence of reconciliation in chimpanzees (*Pan troglodytes*)\u00a0\\[[@RSOS171553C2]\\], a large number of studies have shown that reconciliation occurs in many other primate (reviewed by\u00a0\\[[@RSOS171553C3],[@RSOS171553C4]\\]) and non-primate species (e.g. *Capra hircus*\u00a0\\[[@RSOS171553C5]\\]; *Tursiops* spp.\u00a0\\[[@RSOS171553C6]\\]; *Crocuta crocuta*\u00a0\\[[@RSOS171553C7]\\]; a review\u00a0\\[[@RSOS171553C4]\\]; *Corvus frugilegus*\u00a0\\[[@RSOS171553C8]\\]; *Canis lupus familiaris*\u00a0\\[[@RSOS171553C9]\\]; *Canis lupus*\u00a0\\[[@RSOS171553C10],[@RSOS171553C11]\\]; *Corvus corax*\u00a0\\[[@RSOS171553C12]\\]).\n\nWhether reconciliation occurs is thought to be affected by a species\\' reliance on cooperation\u00a0\\[[@RSOS171553C1],[@RSOS171553C4],[@RSOS171553C13],[@RSOS171553C14]\\]: species characterized by high levels of cooperation are expected to engage in frequent reconciliation following conflicts (e.g.\u00a0\\[[@RSOS171553C15]--[@RSOS171553C17]\\]) in order to re-establish peaceful relationships and maintain pack cohesiveness, potentially essential for future cooperative success\u00a0\\[[@RSOS171553C4],[@RSOS171553C18]\\]. In contrast, in species with low reliance on cooperation reconciliation is expected to be less frequent (e.g.\u00a0\\[[@RSOS171553C19]--[@RSOS171553C21]\\]).\n\nWolves (*Canis lupus*) and domestic dogs (*Canis lupus familiaris*) are particularly good species in which to investigate and compare patterns of reconciliation. Wolves are highly dependent on cooperation because they rely on group hunting, cooperative breeding including alloparental care (e.g.\u00a0\\[[@RSOS171553C22]--[@RSOS171553C25]\\]) and cooperative territorial defence (e.g.\u00a0\\[[@RSOS171553C26]\\]). The cohesiveness and functionality of the packs are crucial for allowing wolves to successfully forage, raise pups and defend their territories\u00a0\\[[@RSOS171553C22],[@RSOS171553C27],[@RSOS171553C28]\\]. Conversely, though free-ranging dogs have been observed to engage in joint territorial defence (e.g.\u00a0\\[[@RSOS171553C29]--[@RSOS171553C31]\\]), alloparental care (e.g.\u00a0\\[[@RSOS171553C32]--[@RSOS171553C34]\\]) and group hunting\u00a0\\[[@RSOS171553C35],[@RSOS171553C36]\\], they are mainly scavengers and exhibit a flexible and more promiscuous mating system\u00a0\\[[@RSOS171553C30],[@RSOS171553C32],[@RSOS171553C37],[@RSOS171553C38]\\] with pups being raised mostly by their mothers\u00a0\\[[@RSOS171553C32],[@RSOS171553C38],[@RSOS171553C39]\\]. They appear to be 'facultatively social', e.g. forming close relationships with other dogs mainly if food abundance is high\u00a0\\[[@RSOS171553C30],[@RSOS171553C40]\\] and/or females are receptive\u00a0\\[[@RSOS171553C41]\\]. Overall, such differences in their social ecologies demand less cooperation with conspecifics than wolves\u00a0\\[[@RSOS171553C42]\\]. As such, we would expect wolves to show a higher reconciliatory tendency than dogs. Indeed, so far, reconciliation has been demonstrated in all wolf packs in which the phenomenon has been studied (two captive packs: European wolves\u00a0\\[[@RSOS171553C10]\\]; Arctic wolves\u00a0\\[[@RSOS171553C43]\\]; and two wild packs\u00a0\\[[@RSOS171553C11]\\]), with the mean conciliatory tendency (i.e. the willingness to reconcile) ranging from 44% to 53%, which is relatively high compared with other species (e.g. hyenas: 16.6%\u00a0\\[[@RSOS171553C7]\\]; chimpanzees: 16.3%\u00a0\\[[@RSOS171553C44]\\]; bonobo: 35.6%\u00a0\\[[@RSOS171553C45]\\]; ravens: 0.16%\u00a0\\[[@RSOS171553C12]\\]; common marmosets: 31%\u00a0\\[[@RSOS171553C46]\\]; tamarins: from 24.7% to 48.3%\u00a0\\[[@RSOS171553C47]\\]; meerkats: 0%\u00a0\\[[@RSOS171553C18]\\]). In the only study investigating conflict management in domestic dogs, reconciliation was shown to occur \\[[@RSOS171553C9]\\], and, although not reported by the authors, the mean conciliatory tendency based on the data published by Cools *et al*. (\\[[@RSOS171553C9]\\], [table\u00a04](#RSOS171553TB4){ref-type=\"table\"}, p. 56) shows that, in line with our hypothesis, it is lower than that reported for wolves, ranging from 16.6% to 29.7%.\n\nThe purpose of our study was twofold. Our first aim was to investigate and compare the pattern of aggression and post-conflict (PC) interactions between former opponents in wolves and dogs living in similarly composed captive packs under identical conditions at the Wolf Science Center, Vienna, Austria. In line with previous findings on reconciliation in captive wolves and dogs and based on the observation that wolves, and under certain conditions also dogs, live in stable social packs, we expect conciliatory behaviours to occur in both species. Nevertheless, because dog packs rely on cooperation less than wolves, we expect to find a higher frequency of reconciliation in wolves than in dogs. We further investigated whether dogs and wolves might also use alternative PC strategies (e.g.\u00a0\\[[@RSOS171553C18]\\]) by including an analysis of the patterns of proximity between opponents after conflicts.\n\nIn order to compare our results with previous findings and to extend our knowledge of reconciliation in both species, our second aim was to investigate the potential factors affecting reconciliation, in terms of its occurrence, latency, duration, intensity as well as its initiator. Previous studies on both dogs and wolves limited their analyses to factors affecting whether or not reconciliation occurred and who was more likely to initiate reconciliation. It has been shown in other species, however, that when and how reconciliation occurs, as well as who initiates reconciliation, may be influenced by variables related to the preceding conflict and to the quality of the two opponents\\' relationship (e.g.\u00a0\\[[@RSOS171553C48]--[@RSOS171553C53]\\]).\n\nIn relation to the conflict characteristics, we included in our analyses (i) the intensity of the conflict (with versus without physical contact) and (ii) its context (whether it occurred in a feeding versus non-feeding situation).\n\nWe characterized the social relationship between opponents, based on daily observations of the animals, in terms of (i) the difference in their competitive abilities (measured by their rank distance and the asymmetry in the exchange of aggressive interactions) and (ii) the security of their relationship (measured by the asymmetry in the exchange of affiliative interactions).\n\nIn general, in line with previous findings in wolves and dogs, as well as in other species, we expect variation in reconciliation to be dependent on its potential costs and benefits as well as on the extent of the damage to the opponents\\' relationships (e.g.\u00a0\\[[@RSOS171553C10],[@RSOS171553C11],[@RSOS171553C43],[@RSOS171553C52],[@RSOS171553C53]\\]). In particular, one of the main costs which opponents may incur when engaging in reconciliation is the risk of further aggression. In this respect, we expect rank distance and aggression asymmetry. In fact, the risk of escalation into further aggression is expected to be higher between opponents closer in rank and those showing a high symmetry in the exchange of aggressive behaviours (i.e. with more similar competitive abilities), which may be more likely to compete for a hierarchical position. When the risk of renewed aggression is high, opponents are expected to hesitate to engage in reconciliation. Therefore, we expect reconciliation to occur less frequently, later and with a shorter duration when conflicts involve opponents closer in rank and with higher aggression symmetry.\n\nWhen competing over food resources, opponents need to find a balance between the benefits of restoring their relationships through reconciliation and the cost of a potential loss of time spent in accessing the food. Accordingly and in line with previous findings in other studies, we expect reconciliation to occur less frequently, sooner and with a shorter duration after conflicts over food (e.g.\u00a0\\[[@RSOS171553C43],[@RSOS171553C52]\\]) than after conflicts occurring in the absence of food.\n\nIn several species, including wolves and dogs, some evidence suggests that the damage to the opponents\\' relationship may be related to the intensity of the conflict, with higher intensity aggression causing higher levels of stress and tension in both opponents\u00a0\\[[@RSOS171553C9],[@RSOS171553C43],[@RSOS171553C52]\\]. Therefore, we expect opponents to reconcile more often and potentially with a more intense interaction after a high-intensity conflict than after a low intensity one. However, as stress and tension decrease with time\u00a0\\[[@RSOS171553C54]\\], aggression can be revived when reconciliation is initiated too quickly after the conflict\u00a0\\[[@RSOS171553C55]\\]. Accordingly, reconciliation should be initiated later after higher intensity conflicts than after low intensity ones.\n\nAlthough previous studies on wolves did not find an influence of the affiliative relationships on the occurrence of reconciliation, this was measured in terms of the frequency of affiliative behaviours exchanged by the dyad. However, another potentially more relevant measure of the social bond between opponents is the security of the relationship. This has been measured in previous studies in terms of the asymmetry of exchanged affiliative behaviours, assuming that those dyads with a high inequality in the exchange of affiliative behaviours are less secure\u00a0\\[[@RSOS171553C53],[@RSOS171553C56],[@RSOS171553C57]\\]. Accordingly, we predict that opponents with a higher symmetry in the exchange of affiliative behaviours (i.e. more secure relationships) will engage earlier and longer in reconciliation than dyads with lower symmetric affiliative relationship (i.e. less secure relationships). Finally, according to previous results on both wolves and dogs, as well as on other species, we expect reconciliation to be initiated more often by victims than aggressors\u00a0\\[[@RSOS171553C11],[@RSOS171553C43],[@RSOS171553C58]\\].\n\n2.. Material and methods {#s2}\n========================\n\n2.1. Subjects and study site {#s2a}\n----------------------------\n\nThis study was conducted on four captive wolf packs and four captive dog packs ([table\u00a01](#RSOS171553TB1){ref-type=\"table\"} for pack composition) at the Wolf Science Center ([www.wolfscience.at](www.wolfscience.at)). The wolves that participated in this study originated from North America but were born in captivity, while the dogs were adopted from Hungarian shelters. All subjects were hand-raised in peer groups from the age of 10 days. They were bottle-fed and later hand-fed by humans, and had continuous access to humans in the first 5 months of their life. After 5 months they were introduced into the packs of adult animals, but still had daily social contact with humans during training and/or cognitive and behavioural experiments. All packs observed were composed of artificially assembled unrelated individuals (except for one wolf pack in which a sibling pair was present; see [table\u00a01](#RSOS171553TB1){ref-type=\"table\"} for details). Table\u00a01.Description of the packs studied.packspeciespack compositionobservation periodtotal minutes of observation per pack per monthKasparwolvesKaspar \u2642Aragorn \u2642Tala \u2640Shima \u2640Chitto \u2642Jan 2013--Mar 2015423GeronimowolvesGeronimo \u2642Amarok \u2642Kenai \u2642Jan 2013--Nov 2014454NanukwolvesNanuk \u2642Yukon \u2640Una \u2640Apr 2013--Jul 2014473WambleewolvesWamblee \u2642Tatonga \u2640Jun 2013--Oct 2013406RafikidogsRafiki \u2642Maisha \u2642Binti \u2640Hakima \u2642Jan 2013--Nov 2013418KiliodogsKilio \u2642Meru \u2642Nia \u2640Bashira \u2640Jan 2013--Jul 2013335NurudogsNuru \u2642Layla \u2640Zuri \u2640Jan 2013--Mar 2014431AsalidogsAsali \u2642Bora \u2640Jan 2013--Nov 2014470\n\nPacks live in large enclosures (4000--8000\u2009m^2^ for wolves, 3000--4000\u2009m^2^ for dogs) equipped with trees, bushes and shelters. Water for drinking is permanently available. Animals receive a diet of meat, fruits, milk products and dry food. During the first months of their lives, they were fed several times per day, which was slowly reduced to being fed major meals daily (dogs) or only two or three times per week (wolves) according to their natural rhythm.\n\n2.2. Data collection {#s2b}\n--------------------\n\nData were collected from January 2013 until March 2015. All packs were observed over 2 days per week for approximately 1\u2009h a day, either in the morning or in the afternoon ([table\u00a01](#RSOS171553TB1){ref-type=\"table\"}) for a total of 927.13 h of observation (523.87\u2009h for wolf packs and 403.26 for dog packs).\n\nTo ensure that the relationships between the animals were characterized based on data independent of the PC behaviour, we adopted two sampling methods: (i) we carried out 'focal animal\\' sampling of all individuals in the pack, focusing on the social behaviours exchanged with all other pack members, and on the basis of this dataset calculated the relationship indices used in the analyses, and (ii) to collect data on the PC behaviours we adopted 'behavioural' sampling methods\u00a0\\[[@RSOS171553C59]\\], video-recording the victim of a conflict.\n\nFocal animal sampling lasted 10\u2009min and was carried out using the program Pocket Observer v. 2.1.23.2 (Noldus Information Technology) on a hand-held device (Samsung Galaxy Note 2). We recorded all occurrences of the focal animal\\'s aggressive, dominant, submissive (used to assess the hierarchical rank) and affiliative interactions and towards whom they were directed (see electronic supplementary material, table S1) (for a total of 421.83\u2009h of observation; 17.17\u2009\u00b1\u200913.99 mean\u2009\u00b1\u2009s.d. per subject).\n\nBehavioural sampling sessions for PC behaviour were started when an aggressive encounter began and ended 10\u2009min after it ended. During this time we continuously filmed the victim as the focal individual for a 10\u2009min PC period. Control observations (matched controls, MCs) took place the next possible day at the same time as the original PC, on the same focal animal, in the absence of agonistic interactions occurring during the 10\u2009min before the beginning of the MC\u00a0\\[[@RSOS171553C60],[@RSOS171553C61]\\]. For each aggressive encounter we recorded: (i) the opponents, (ii) the context (presence or absence of food), (iii) the outcome of conflict (decided or undecided), and (iv) aggressive behavioural patterns (see electronic supplementary material, table S2). Decided conflicts were defined as those conflicts in which one individual performed an aggressive behaviour and the other responded with a submissive behaviour. Cases in which both were aggressive or in which the recipients of the aggression did not show any clear submissive behaviour or answered by showing dominant behaviours were defined as undecided. For undecided conflicts, we defined as the aggressor the wolf/dog that was first seen being aggressive.\n\nAggressive behaviours were then categorized according to two levels of aggressive intensity: level 1: low intensity---aggressions without physical contact (threat, chase, jaw spar and snap); and level 2: high intensity---aggressions with physical contact (attack, knock down, stand over aggressive, pin, fight and bite) (see the electronic supplementary material, table S2, for detailed descriptions of the behaviours).\n\nFinally, video records of the PC and MC observations were analysed using the software Solomon Coder^\u00ae^ (Andr\u00e1s P\u00e9ter). For both PCs and MCs we recorded (i) starting time (min), (ii) the minute of the first affiliative behaviour, (iii) type and duration of the affiliative interaction, (iv) the initiator of the affiliative behaviour, and (v) the duration of the time spent by the victim in close proximity to the aggressor (see the electronic supplementary material, table S2, for a detailed description of the behaviours).\n\n2.3.. Inter-observer reliability {#s2c}\n--------------------------------\n\nForty-two PC--MC pairs (i.e. 23.7% of the total observations) were video-recorded and subsequently analysed independently by the two main observers (S.C. and M.L.) to obtain inter-observer reliability coding. We calculated a \u03ba coefficient\u00a0\\[[@RSOS171553C62],[@RSOS171553C63]\\] for all relevant behaviours (affiliative and proximity interactions, *k*\u2009=\u20090.81 and *k*\u2009=\u20090.74, respectively), timing (*k*\u2009=\u20090.76) and partner identities (*k*\u2009=\u20090.97). The overall averaged \u03ba coefficient was *k*\u2009=\u20090.82.\n\n2.4.. Data analyses {#s2d}\n-------------------\n\n### 2.4.1.. Measure of relationship quality {#s2d1}\n\nIn order to characterize the relationship quality between aggressor and victim within the packs, we used data from our focal animal observations to calculate the following behavioural variables: (i) the asymmetry in the exchange of affiliative and (ii) aggressive behaviours and (iii) the dyadic rank distance.\n\nFor (i) and (ii) we calculated the asymmetry in the exchange of affiliative (AFFav) and aggressive (AGGav) behaviours for each dyad according to the following formula (adapted from\u00a0\\[[@RSOS171553C53]\\]): $$\\text{AFFav(\\ AGGav)~} = \\frac{Ba\\rightarrow Bv}{Ba\\leftrightarrow Bv} - \\frac{Bv\\rightarrow Ba}{Ba\\leftrightarrow Bv},$$ where *Ba\u2009*\u2192\u2009*Bv* is the total number of behaviours that *the aggressor* directed at *the victim* during the focal observations, *Bv\u2009\u2192\u2009Ba* is the amount of behaviours that *the victim* directed at *the aggressor*, and *Ba\u2009*\u2194\u2009*Bv* is the total amount of behaviour exchanged between *aggressor* and *victim*.\n\nA value close to '0\\' means that the two animals exchange the behaviours equally frequently, that is, their relationship is symmetric. A value close to '\u22121\\' means that the victim shows more behaviours towards the aggressor than vice versa, whereas a value close to '+1\\' means that the reverse is true.\n\nThe asymmetry in the exchange of *affiliative behaviours* has been used in other species as a measure of the security of the relationships\u00a0\\[[@RSOS171553C53],[@RSOS171553C56],[@RSOS171553C57]\\]; in particular, dyads showing a high asymmetry are assumed to have a less secure relationship. Asymmetry in the exchange of *aggressive behaviours* is instead considered a measure of the differences in competitive abilities between two subjects; with dyads showing a high asymmetry with highly diverse competitive skills.\n\nFor (iii), for each individual we calculated the David\\'s score\u00a0\\[[@RSOS171553C64]\\] based on the direction and frequency of submissive behaviours recorded during the focal observations. Then, we used the value of the difference between the individual scores to represent the dyadic rank distance (see the electronic supplementary material for more details).\n\n### 2.4.2.. Measures of reconciliation {#s2d2}\n\nTo investigate the occurrence of reconciliation, we used two kinds of analyses and took only those dyads that were involved in at least three conflicts into consideration (for a total of 15 dyads and 175 conflicts for wolves and five dyads and 26 conflicts for dogs). First, to compare the timing of the first affiliative interaction between former opponents during one PC with that during the corresponding MC, we followed de Waal & Yoshihara\\'s\u00a0\\[[@RSOS171553C60]\\] PC--MC method. A PC--MC pair was labelled 'attracted\\' if the former opponents affiliated only in the PC or earlier in the PC than in the MC. Similarly, a PC--MC pair was labelled 'dispersed\\' if the former opponents affiliated only in the MC, or earlier in the MC than in the PC. When affiliative contacts occurred during the same minute in the PC and the MC, or no contact occurred in either the PC or the MC, the PC--MC pair was labelled 'neutral\\'. For each individual, we compared the proportion of 'attracted\\' PC--MC pairs with the proportion of 'dispersed\\' pairs by using a Wilcoxon matched-pair signed-ranks test (corrected for ties\u00a0\\[[@RSOS171553C65]\\]). Second, in order to establish the specific time frame in which an affiliative behaviour is more likely to function as a 'reconciliation\\' event, we used the 'time rule\\' method (compares the frequency of PC affiliative interactions between opponents occurring within a specified timeframe, i.e. the 10\u2009min PC focal, in PC versus MC periods; see\u00a0\\[[@RSOS171553C3]\\], pp. 22 and\u00a0\\[[@RSOS171553C66]\\]). Following this method, we determined for each PC and MC observation the minute in which the first affiliative contact between opponents occurred. Next, we compared the distribution of first PC events with first MC events using the Kolmogorov--Smirnov test. If this test produced a significant result, we ran a Wilcoxon matched-pair signed-ranks test comparing individual PC and MC scores within the time period in which the PC distribution differed from that for the MC. In this way we checked the generality of this phenomenon at the individual level.\n\nOverall, reconciliation is considered to occur when following both the PC--MC method and the 'time rule\\' method, former opponents made friendly contacts earlier in PCs than in MCs.\n\nFollowing Veenema *et al*.\u00a0\\[[@RSOS171553C67]\\], to calculate each individual\\'s corrected conciliatory tendency (CCT), we followed the formula: attracted minus dispersed pairs divided by the total number of PC--MC pairs (attracted\u2009+\u2009dispersed\u2009+\u2009neutral pairs). Individual CCTs were then used to determine the mean group CCT for species comparison.\n\nAll non-parametric tests (two-tailed) were conducted in STATISTICA 7.1 edition (StatSoft Italy s.r.l. 2005). The probability level for rejection of the null hypothesis was set at 0.05.\n\n### 2.4.3.. Dog--wolf comparison of aggression and proximity patterns {#s2d3}\n\nIn order to achieve our first aim, that is, to compare dogs and wolves in regard to their proportion and intensity of aggression, the occurrence of reconciliation and proximity between opponents after conflicts, we used the Mann--Whitney test (two-tailed) conducted in STATISTICA 7.1 edition (StatSoft Italy s.r.l. 2005). The probability level for rejection of the null hypothesis was set at 0.05.\n\n### 2.4.4.. Exploring variation in reconciliation {#s2d4}\n\n#### 2.4.4.1.. Measures of variation in reconciliation {#s2d4a}\n\nA reconciliation event, defined as either the victim or the aggressor displaying an affiliative behaviour towards its opponent, was further characterized in a number of ways: (i) the latency of its occurrence, that is, the time elapsed between the end of the conflict and the beginning of the first PC affiliative interaction, (ii) the duration of the reconciliation event, that is, the time elapsed from the start of the first affiliative behaviour to the end of the last consecutive affiliative behaviour displayed by the initiator of reconciliation, and (iii) the intensity of the reconciliation event based on the affiliative intensity displayed: low-intensity reconciliation---only one affiliative behavioural pattern was displayed without involving any physical contact between the opponents (e.g. approach friendly and stand friendly)---and high-intensity reconciliation---one or more affiliative behavioural patterns were displayed with at least one of these involving physical contact between opponents (e.g. nose touch, muzzle-licking and body rubbing), (see the electronic supplementary material, table S2, for a detailed description). Both the duration and intensity of reconciliation were, to some extent, dependent on the number of affiliative behavioural patterns displayed during the reconciliatory interaction. However, high-intensity reconciliations did not last significantly longer than low-intensity ones (Mann--Whitney test: *U~39,30~*\u2009=\u2009518.5, *Z*~adj~\u2009=\u20090.1, *p*\u2009=\u20090.42; mean durations\u2009\u00b1\u2009s.d.: high-intensity reconciliations, 7.33\u2009\u00b1\u20098.81 s; low-intensity reconciliations, 9.33\u2009\u00b1\u200913.17\u2009s), suggesting that the duration and intensity of reconciliation were independent from each other.\n\n#### 2.4.4.2.. Variables potentially affecting variation in reconciliation: test models {#s2d4b}\n\n[Table\u00a02](#RSOS171553TB2){ref-type=\"table\"} outlines the variables considered as potential predictors of the occurrence, latency, duration and intensity as well as the initiator (aggressor versus victim) of reconciliation. To test which factor may affect the occurrence of reconciliation (model 1), its initiator (model 2) and its intensity (model 3), we ran generalized mixed-effect models (GLMMs) with a binomial distribution. To detect which independent variables may affect the latency (model 4) and duration (model 5) of reconciliation we ran linear mixed-effect models (LMMs). Models 2, 3, 4 and 5 were restricted to 'attracted\\' PC--MC pairs (for a total of 69 pairs), thus only PCs in which reconciliation occurred (i.e. all PCs in which a friendly contact occurred earlier than in MCs). Model assumptions (i.e. over-dispersion for GLMMs and normally distributed residuals and homogeneity of variances for LMMs) were met. We included, as random factors, the identity of the victim for each dyad nested into dyad and pack identity. Random factors allow the inclusion in the model of multiple data collected on the same victim or dyad/pack, thus controlling for non-independence of the samples\u00a0\\[[@RSOS171553C68]\\]. Table\u00a02.Name and description of the variables considered in the test models. In Model 1, we analysed 175 PC--MC pairs for a total of 15 dyads. Models 2, 3, 4 and 5 were restricted to 'attracted\\' PC--MC pairs, for a total of 69 pairs and 13 dyads.variabletype and description*response variable*Model 1. occurrence of reconciliationcategorical (yes--no)Model 2. initiator of reconciliationcategorical (victim--aggressor)Model 3. intensity of reconciliationcategorical (low intensity--high intensity)Model 4. latency of reconciliationcontinuous (the seconds from the end of a conflict to the occurrence of reconciliation)Model 5. duration of reconciliationcontinuous (the seconds from the beginning of the first affiliative behaviour to the last one)*independent variable* (for all models)1. rank distancecontinuous (differences of individual David\\' scores)2. degree of affiliation symmetrycontinuous (AFFav)^a^3. degree of aggression symmetrycontinuous (AGGav)^a^4. conflict intensitycategorical (high intensity--low intensity)5. context of the conflictcategorical (feeding context--non-feeding context)[^2]\n\nFor all analyses to evaluate and compare different models, we used the Akaike information criterion, corrected for small sample sizes (AICc), and adopted a model-averaging approach (using the package MuMIn in R), allowing us to evaluate also the 'relative importance\\' of each variable (RVI) across models\u00a0\\[[@RSOS171553C69],[@RSOS171553C70]\\].\n\nAll model analyses were performed using R v. 3.3.3. We implemented linear mixed-effects models using the 'lmer\\' function of the 'lmerTest\\' package\u00a0\\[[@RSOS171553C71]\\] and generalized mixed-effects models using the 'glmer\\' function in the 'lme4\\'package\u00a0\\[[@RSOS171553C72]\\]. All data are provided in the electronic supplementary material.\n\n3.. Results {#s3}\n===========\n\n3.1.. Pattern of aggressions in wolf and dog packs {#s3a}\n--------------------------------------------------\n\nIn wolf packs, we recorded 419 aggressive interactions (0.80 aggressions per observation hour) and a total of 177 PC--MC pairs^[1](#FN1){ref-type=\"fn\"}^ involving 12 subjects as victim (one adult male, Nanuk, was never recorded as a conflict victim). The number of PCs per wolf ranged from 1 to 47 with an average number of conflicts per focal of 13.62\u2009\u00b1\u200914.07. Of the total 177 PC episodes, 45 occurred in the food context and 132 outside of it. Furthermore, most of the conflicts had a clear outcome (155 were decided and only 22 were undecided). Finally, 59.3% (*N*\u2009=\u2009105) of conflicts were characterized by aggression of high intensity and 40.7% (*N*\u2009=\u200972) by low-intensity aggression.\n\nIn dogs, the number of aggressive interactions recorded was much lower than in wolves. In fact, during the observation period, we recorded only 55 aggressive interactions (0.14 aggressions per observation hour) and a total of 30 PC--MC pairs involving eight subjects as victim (four adult male dogs, Rafiki, Kilio, Nuru and Asali, and one young female, Nia, were never recorded as a conflict victim). The number of PCs per dog ranged from one to seven with an average number of conflicts per focal of 3.75\u2009\u00b1\u20092.44. Of the total 30 PC episodes, six occurred in the food context and 24 outside of it. As for wolves, most of the conflicts had a clear outcome (27 were decided and only three were undecided), but, in contrast to wolves, the higher proportion of conflicts were characterized by high-intensity aggressions (86.6% (*N*\u2009=\u200928) were of high intensity while only 12.4% (*N*\u2009=\u20092) were of low intensity). In fact, we found that dog dyads showed significantly more high-intensity conflicts than wolf dyads (Mann--Whitney test: *U*~15,8~\u2009=\u200911, *Z*~adj~\u2009=\u2009\u22123.23, *p*\u2009=\u20090.0008; [figure\u00a01](#RSOS171553F1){ref-type=\"fig\"}). Figure 1.Pattern of aggressive interactions in wolves and dogs.\n\n3.2.. Does reconciliation occur in wolves and dogs? {#s3b}\n---------------------------------------------------\n\nFollowing the PC--MC method, wolf victims and aggressors were significantly more likely to interact affiliatively with former opponents in PCs than MCs (42.4% versus 19.8%; [figure\u00a02](#RSOS171553F2){ref-type=\"fig\"}). In fact, in the wolves we found significantly more attracted than dispersed PC--MC pairs (Wilcoxon signed-ranks test: *Z*\u2009=\u20092.55, *n*\u2009=\u200910, *p*\u2009=\u20090.01). This indicates that the majority of affiliative contacts between the opponents occurred earlier in PC than in MC periods. The time rule method confirmed this effect. The temporal distribution of first affiliative interactions between former opponents in the PCs was significantly different from that in the MCs (Kolmogorov--Smirnov test: *D*\u2009=\u20090.01, *p*\u2009\\<\u20090.001; [figure\u00a03](#RSOS171553F3){ref-type=\"fig\"}). Figure 2.Percentage of PCs and MCs followed by affiliation for both wolves and dogs. Figure 3.Temporal distribution of first affiliative interactions in PC and MC periods.\n\nThe greatest difference in the cumulative distributions was during the first minute. Importantly, this result is not due to a few individuals, as most wolf victims were involved in affiliation with former aggressors in the first minute of more PCs than in the first minute of their MCs (Wilcoxon signed-ranks test: *Z*\u2009=\u20091.96, *n*\u2009=\u200910, *p*\u2009=\u20090.04). The mean CCT of all focal individuals was 22.46%, ranging from \u221225.00% to 63.16% ([table\u00a03](#RSOS171553TB3){ref-type=\"table\"}). Based on the two methods outlined above, we could consider 71 conflicts out of a total of 177 as reconciled. Table\u00a03.Corrected conciliatory tendencies (CCTs) in wolves, number of attracted, dispersed and neutral pairs for each victim. Only victims involved in more than three conflicts were included (thus, two subjects, Wamblee and Kaspar, who were involved in one and two conflicts, respectively, were not included in the analysis). Group CCT%, the mean CCTs of the pack; s.e., standard error.focal/victimattracteddispersedneutraltotalCCT%Aragorn12482433.33%Tala13151963.16%Shima5424333.03%Chitto2511114729.79%Geronimo102333.33%Amarok3271208.33%Kenai20101216.67%Yukon312633.33%Una5181428.57%Tatonga0134\u221225.00%**group CCT%22**.**46%s.e.0**.**07%** Table\u00a04.Number of attracted, dispersed and neutral pairs for each dog victim.focal/victimattracteddispersedneutraltotalMaisha0077Binti0044Hakima0011Meru2147Bashira0022Bora2215Layla1001Zuri1203\n\nContrary to wolves, dog victims and aggressors were not more likely to interact affiliatively with former opponents in PCs than in MCs (23.3% versus 26.7%; [figure\u00a02](#RSOS171553F2){ref-type=\"fig\"}). Although we had only 30 PC--MC pairs, we found no difference between the number of attracted and dispersed PC--MC pairs (Wilcoxon signed-ranks test: *Z*\u2009=\u20090.54, *n*\u2009=\u20098, *p*\u2009=\u20090.59; [table\u00a04](#RSOS171553TB4){ref-type=\"table\"}).\n\n3.3.. Patterns of proximity during post-conflict observations in wolves and dogs {#s3c}\n--------------------------------------------------------------------------------\n\nThe analyses of proximity revealed that, after a conflict (i.e. during PC periods), dog victims spent significantly less time in close proximity to their aggressor than wolves (Mann--Whitney test: *U*~15,8~\u2009=\u200913.5, *Z*~adj~\u2009=\u20093.02, *p*\u2009=\u20090.001), while no difference between the two species emerged during MC periods (Mann--Whitney test: *U*~15,8~\u2009=\u200955, *Z*~adj~\u2009=\u2009\u22120.32, *p*\u2009=\u20090.78). This result indicates that dog opponents do not seek an opportunity to reconcile but rather choose to avoid each other. Furthermore, during PCs, the occurrence of reconciliation did not influence the proximity between wolf opponents because there was no difference in the duration of proximity between reconciled versus un-reconciled PCs (Mann--Whitney test: *U*~106,71~\u2009=\u20093244, *Z*~adj~\u2009=\u2009\u22121.70, *p*\u2009=\u20090.13). Therefore, the finding that, after conflicts, wolf opponents spent more time in close proximity than dogs is not just a consequence of reconciliation.\n\nSince we could not find any evidence of the occurrence of reconciliation in dogs, data on PC behaviour of dogs were not further analysed.\n\n3.4.. Variation of reconciliation in wolves {#s3d}\n-------------------------------------------\n\n### 3.4.1.. What factors affect the occurrence of post-conflict friendly interactions (model 1)? {#s3d1}\n\nWe found an effect of the rank distance between opponents, context and intensity of aggression with a higher likelihood of reconciliation occurring when the victim and the aggressor were more distant in rank, when the conflict happened without food present and after high-intensity aggression.\n\nIn particular, of the 45 conflicts occurring when food was available, 5.08% (*n*\u2009=\u20099) were followed by a friendly interaction, whereas of the 130 conflicts occurring in the non-food context 46.92% (*n*\u2009=\u200961) were followed by a friendly interaction. In the same way, of the 104 conflicts characterized by high-intensity aggression, 48 were followed by a friendly interaction (46.15%), whereas of the 71 conflicts with low-intensity aggression only 22 were followed by a friendly interaction (30.99%).\n\nThe asymmetry in the exchange of affiliative and aggressive behaviours had no effect on the likelihood of a friendly interaction occurring. Model comparison is shown in [table\u00a05](#RSOS171553TB5){ref-type=\"table\"} while estimated effect sizes, relative variable importance and *p*-values are shown in [table\u00a06](#RSOS171553TB6){ref-type=\"table\"}. Table\u00a05.Model comparison for factors affecting the occurrence of post-conflict friendly interactions (model 1). Models are reported up to three AICc points from the best model.d.f.logLikAICcdeltaweightcontext, conflict intensity and rank distance6\u2212101.64215.790.000.34context, conflict intensity, rank distance and degree of aggression symmetry7\u2212101.56217.792.000.12context, conflict intensity, rank distance and degree of affiliation symmetry7\u2212101.62217.912.120.12context, conflict intensity and degree of aggression symmetry6\u2212103.07218.642.850.08 Table\u00a06.Estimated effect size, adjusted standard error (s.e.), *Z*-value and relative variable importance (RVI) estimated by a generalized linear mixed model to determine the effects of each variable on the occurrence of reconciliation (model 1).effect sizeadjusted s.e.*Z*-value*p*RVIcontext of conflict1.590930.513533.0980.001950.99conflict intensity\u22120.835390.386862.1590.030820.82rank distance0.118860.054472.1820.029110.77degree of aggression symmetry0.802781.084870.7400.459320.34degree of affiliation symmetry\u22120.356951.210230.2950.768040.26\n\n### 3.4.2.. What factors affect the initiator identity of post-conflict friendly interactions (model 2)? {#s3d2}\n\nWhen reconciliation occurred between wolf opponents, it was more likely to have been initiated by the victim (58) than the aggressor (13) (Mann--Whitney test: *U*~71,71~\u2009=\u2009958.5, *Z*~adj~*\u2009*=\u20097.36, *p*\u2009\\<\u20090.000001). In the majority of cases, the victim was also the subordinate (56) individual (Mann--Whitney test: *U*~71,71~\u2009=\u2009923, *Z*~adj~*\u2009*=\u20097.53, *p*\u2009\\<\u20090.000001). Only the context of the conflict had a significant effect on the identity of the initiator of PC friendly interaction. In particular, the majority of PC affiliative interactions initiated by the victim occurred in the non-feeding context (victims: 53; aggressors: 9), while both victim and aggressor were equally responsible for the initiation of reconciliation in the feeding context (victims: 5; aggressors: 4). Intensity of aggression as well as the asymmetry in the exchange of affiliative and aggressive behaviours had no effect on the initiator identity of friendly interactions. Model comparison is shown in [table\u00a07](#RSOS171553TB7){ref-type=\"table\"}, while estimated effect sizes, relative variable importance and *p*-values are shown in [table\u00a08](#RSOS171553TB8){ref-type=\"table\"}. Table\u00a07.Model comparison for factors affecting the initiator identity of post-conflict friendly interactions (model 2). Models are reported up to 2 AICc points from the best model.d.f.logLikAICcdeltaweightcontext and rank distance5\u221229.1569.260.000.25context, conflict intensity and rank distance6\u221228.9771.302.040.09 Table\u00a08.Estimated effect size, adjusted standard error (s.e.), *Z*-value and RVI estimated by a generalized linear mixed model to determine the effects of each variable on the initiator identity of post-conflict friendly interactions (model 2).effect sizeadjusted s.e.*Z*-value*p*RVIcontext of conflict2.028150.947652.1400.03230.80conflict intensity0.330990.789800.4190.67520.26rank distance0.125510.070891.7710.07660.68degree of aggression symmetry0.262180.964160.2720.78570.25degree of affiliation symmetry\u22120.861651.820490.4730.63600.25\n\n### 3.4.3.. What factors affect the intensity of post-conflict friendly interactions (model 3)? {#s3d3}\n\nHigh-intensity PC affiliative interactions occurred more frequently than low-intensity ones (high intensity: 56.52%; low intensity: 43.48%). The intensity of PC friendly interaction was affected by context, intensity of aggression and asymmetry in the exchange of aggressive behaviours. In fact, conflicts in the non-feeding context were followed more often by highly intense friendly interactions (*N*\u2009=\u200938) than low-intensity ones (*N*\u2009=\u200922), whereas the reverse was true in the feeding context (high intensity: 1; low intensity: 8). Furthermore, high-intensity, but not low-intensity, reconciliations occurred more often after high-intensity (*N*\u2009=\u200932) than low-intensity aggressive interactions (*N*\u2009=\u20097). Finally, reconciliation between former opponents was more frequently of high intensity for those dyads with higher aggression asymmetry, thus in dyads in which the aggressor showed a higher frequency of aggressive behaviours towards the victim than vice versa.\n\nRank distance between opponents and asymmetry in the exchange of affiliative behaviours had no effect on the intensity of friendly interactions. Model comparison is shown in [table\u00a09](#RSOS171553TB9){ref-type=\"table\"}, while estimated effect sizes, relative variable importance and *p*-values are shown in [table\u00a010](#RSOS171553TB10){ref-type=\"table\"}. Table\u00a09.Model comparison for factors affecting the intensity of post-conflict friendly interactions (model 3). Models are reported up to three AICc points from the best model.d.f.logLikAICcdeltaweightcontext, conflict intensity and degree of aggression symmetry6\u221236.6086.550.000.28context, conflict intensity, degree of aggression symmetry and degree of affiliation symmetry7\u221235.7687.360.810.19context, conflict intensity, rank distance and degree of aggression symmetry7\u221236.4088.642.090.10context, conflict intensity, rank distance, degree of aggression symmetry and degree of affiliation symmetry8\u221235.2988.982.440.08 Table\u00a010.Estimated effect size, adjusted standard error (s.e.), *Z*-value and RVI estimated by a generalized linear mixed model to determine the effects of each variable on the intensity of reconciliation (model 3).effect sizeadjusted s.e.*Z*-value*p*RVIcontext of conflict2.690491.171212.2970.02160.95conflict intensity\u22121.343950.639632.1010.03560.76rank distance0.047490.060580.7840.43300.30degree of aggression symmetry2.611941.135962.2990.02150.92degree of affiliation symmetry2.145321.688101.2710.20380.42\n\n### 3.4.4.. What factors affect the latency of post-conflict friendly interactions (model 4)? {#s3d4}\n\nReconciliation mainly took place within the first 2\u2009min from the end of a conflict although its latency was highly variable, ranging from 0 to 582\u2009s (mean 99.4\u2009\u00b1\u2009145.8\u2009s between the end of a conflict and the occurrence of reconciliation\u2009\u00b1\u2009s.d.). None of the predictor variables had an effect on the timing of reconciliation. Model comparison is shown in [table\u00a011](#RSOS171553TB11){ref-type=\"table\"}, while estimated effect sizes, relative variable importance and *p*-values are shown in [table\u00a012](#RSOS171553TB12){ref-type=\"table\"}. Table\u00a011.Model comparison for factors affecting the latency of post-conflict friendly interactions (model 4). Models are reported up to 2 AICc points from the best model.d.f.logLikAICcdeltaweightcontext5\u2212439.59890.120.000.12context and conflict intensity6\u2212438.68890.710.590.09null model4\u2212441.19891.000.870.08context and degree of affiliation symmetry6\u2212439.39892.132.000.05 Table\u00a012.Estimated effect size, adjusted standard error (s.e.), *Z*-value and RVI estimated by a generalized linear mixed model to determine the effects of each variable on the latency of reconciliation (model 4).effect sizeadjusted s.e.*Z*-value*p*RVIcontext of conflict\u221294.71053.4491.7720.07640.56conflict intensity\u221244.41537.9321.1710.24160.38rank distance\u22122.5043.9030.6420.52120.28degree of aggression symmetry78.63161.1981.2850.19880.39degree of affiliation symmetry98.21799.9050.9830.32560.32\n\n### 3.4.5.. What factors affect the duration of post-conflict friendly interactions (model 5)? {#s3d5}\n\nThe duration of reconciliation was highly variable, ranging from 0.6 to 65\u2009s (8.2\u2009\u00b1\u200910.8, mean seconds\u2009\u00b1\u2009s.d.). We found an effect of asymmetry in the exchange of affiliative behaviours with reconciliation between former opponents lasting longer for those dyads in which the aggressor showed more affiliative behaviours towards the victim than vice versa.\n\nAll other variables had no effect on the timing of reconciliation. Model comparison is shown in [table\u00a013](#RSOS171553TB13){ref-type=\"table\"}, while estimated effect sizes, relative variable importance and *p*-values are shown in [table\u00a014](#RSOS171553TB14){ref-type=\"table\"}. Table\u00a013.Model comparison for factors affecting the duration of post-conflict friendly interactions (model 5). Models are reported up to three AICc points from the best model.d.f.logLikAICcdeltaweightcontext, conflict intensity, degree of aggression symmetry and degree of affiliation symmetry8\u2212247.73513.870.000.39conflict intensity, degree of aggression symmetry and degree of affiliation symmetry7\u2212250.03515.892.030.14context, conflict intensity and degree of affiliation symmetry7\u2212250.34516.512.650.10context, degree of aggression symmetry and degree of affiliation symmetry7\u2212250.48516.802.930.09 Table\u00a014.Estimated effect size, adjusted standard error (s.e.), *Z*-value and RVI estimated by a generalized linear mixed model to determine the effects of each variable on the duration of reconciliation (model 5).effect sizeadjusted s.e.*Z*-value*p*RVIcontext of conflict0.3196834.0353590.0790.93690.74conflict intensity3.5956572.8832781.2470.21240.81rank distance\u22120.0016790.3217530.0050.99580.17degree of aggression symmetry3.1049184.4292290.7010.48330.79degree of affiliation symmetry17.1467207.3616612.3290.01980.98\n\n4.. Discussion {#s4}\n==============\n\n4.1.. Comparing patterns of aggression and post-conflict behaviours in wolves and dogs {#s4a}\n--------------------------------------------------------------------------------------\n\nOur results highlight a number of interesting differences between dogs and wolves in both their conflict and PC behaviours. Dogs showed significantly lower rates of aggression than wolves; however, when aggression did occur it was of higher intensity (i.e. involving physical contact). Interestingly, these results are in line with early studies of similarly raised packs of wolves and dogs, which lead to the suggestion that ritualized aggression in domestic dog may have been progressively lost during the course of domestication\u00a0\\[[@RSOS171553C73]--[@RSOS171553C75]\\].\n\nAfter conflicts, wolves were more likely to reconcile, whereas dogs appeared to adopt an avoidance strategy, because dogs but not wolves spent more time away from their partners after a conflict than in control periods. This is also reflected in the animal management at the Wolf Science Center as we had to remove some dogs from the packs because conflicts between dogs escalated into severe aggression more often than conflicts between wolves. During the study, three dogs had to be removed from the packs, whereas no wolf was taken out. While this may lead to an underestimation of the frequency of aggressive interactions/conflicts that occurred in dogs, it does not affect the estimation of the tendency to reconcile because this measure is proportional to the conflicts that occurred in the respective species.\n\nIn wolves, the number of attracted pairs was significantly higher than the number of dispersed pairs, indicating that former opponents show a higher affiliative tendency in PC situations than without a preceding conflict. Supporting this suggestion, reconciliation occurred mainly in the first minute. The mean conciliatory tendency of our wolf packs (22.46%) was lower than that found in two wild packs of Canadian timber wolves (CCT\u2009=\u200944.1%\u00a0\\[[@RSOS171553C11]\\]), and in captive families of European (CCT\u2009=\u200953.2%\u00a0\\[[@RSOS171553C10]\\]) and Arctic wolves (CCT\u2009=\u200946.87%\u00a0\\[[@RSOS171553C43]\\]). However, our packs differ from all previous studied packs in two respects. First, our packs were smaller in size (3.5\u2009\u00b1\u20091.26 individuals) than the captive pack of European wolves (nine individuals) and Arctic wolves (19 individuals). Second, our packs were composed of artificially assembled unrelated individuals, whereas all the other packs studied were families composed of the breeding pair and their offspring of the previous year (with the exception of the European wolf pack in which only the alpha-male was present together with the offspring\u00a0\\[[@RSOS171553C10]\\]). While the number of available social partners does not seem to influence the occurrence and frequency of reconciliation (chimpanzees\u00a0\\[[@RSOS171553C44]\\]; wolves: nine individuals\u00a0\\[[@RSOS171553C10]\\]; 19 individuals\u00a0\\[[@RSOS171553C43]\\]), in a number of species reconciliation is more frequent among related than unrelated individuals (the kinship hypothesis\u00a0\\[[@RSOS171553C12],[@RSOS171553C76]--[@RSOS171553C78]\\]). Considering that wild wolf packs are composed of related members that cooperate in a wide variety of activities, relatedness is likely to be an important aspect of wolves\\' sociality. As such, in the current study pack composition is more likely than pack size to account for the low level of CCT.\n\nIn contrast to wolves, in our packs of domestic dogs, we found no evidence of reconciliation after aggressive encounters. These results are in contrast with those found in the only other study investigating reconciliation in three captive groups of dogs\u00a0\\[[@RSOS171553C9]\\]. The study differed from ours in a number of ways---two in particular, which are thought to be crucial as regards the contrasting results, are (i) the relatedness between pack members and (ii) the behaviours considered as aggressive and affiliative interactions.\n\nOur dog packs were not composed of related individuals (with the only exception of two siblings, Nuru and Zuri, which were both members of the same pack), while several dogs in the study of Cools *et al*. \\[[@RSOS171553C9]\\] were related to each other. Relatedness is known to affect reconciliation rates\u00a0\\[[@RSOS171553C12],[@RSOS171553C76]--[@RSOS171553C78]\\], and indeed Cools *et al.* also found that reconciliation occurred more in 'familiar\\' than in 'unfamiliar\\' dogs (where familiarity and relatedness overlapped to a large extent in the study). It is, therefore, possible that reconciliation rates in our pack were particularly low because no such relatedness bonds were present. Second, the ethogram used by Cools *et al*.\u00a0\\[[@RSOS171553C9]\\] was different from the one used in the current study. In particular, they considered 'directed barking\\' and 'anogenital sniffing\\' as light aggression and affiliation, respectively. Actually, both behavioural patterns may have other functions (e.g.\u00a0\\[[@RSOS171553C79],[@RSOS171553C80]\\]), and this may have led Cools *et al*.\u00a0\\[[@RSOS171553C9]\\] to overestimate the frequency of both conflicts (1711 conflicts over two months of observation on three groups composed, respectively, by seven, seven and six dogs) and reconciliations (*N*\u2009=\u2009606), compared with our study. Even though the factors affecting the occurrence of reconciliation in dogs are still unclear, it is interesting to note that, in both studies, reconciliation tendencies were markedly lower than any reported in wolf studies. Indeed, where wolves engage in conflict management, dogs seem to avoid each other. This is in line with results from Cools *et al*.\u00a0\\[[@RSOS171553C9]\\], who found that, when a conflict was not immediately followed by reconciliation, opponents usually avoided each other for some time\u00a0\\[[@RSOS171553C9]\\]. Actually, this result suggests that dogs may use a different PC strategy from that used by wolves. Interestingly this 'avoidance strategy\\' has also been found in dogs, but not in wolves in a feeding context in that dogs are less tolerant of proximity during feeding on a monopolizable food source than wolves, tending to avoid conflicts by maintaining distance rather than using communication to negotiate access as wolves do\u00a0\\[[@RSOS171553C81],[@RSOS171553C82]\\].\n\nTaken together current results on PC behaviour and results on food tolerance with similarly raised captive wolves and dogs suggest that wolves may use communication (including aggressive displays) with each other more frequently to negotiate and resolve conflict than dogs, which only engage in reconciliation under specific conditions (e.g. if relatives are involved).\n\nThe difference in the social ecology of dogs compared with wolves may explain the difference in their tendency to reconcile. In fact, dogs appear to depend mostly on scavenging rather than group hunting and rarely show cooperative breeding (e.g. provisioning of pups by all pack members). This reduced dependence on pack members may have relaxed the selection for communicative complexity\u00a0\\[[@RSOS171553C42],[@RSOS171553C73]--[@RSOS171553C75]\\], resulting in less frequent communicative threats, more intense physical aggression when conflicts occur and avoidance strategies remaining as the main alternative.\n\n4.2.. Factors affecting reconciliation in timber wolves {#s4b}\n-------------------------------------------------------\n\nAs regards the characteristics of the conflict itself, variation in wolf reconciliation was mainly influenced by two factors: the context in which the conflict occurred and the intensity of the aggression. According to our predictions and in line with previous findings in wolves\u00a0\\[[@RSOS171553C43]\\] and other species (e.g.\u00a0\\[[@RSOS171553C83]--[@RSOS171553C87]\\]), reconciliation occurred more often and was more intense (showing multiple affiliative behaviours) in the non-feeding contexts than during feeding sessions. In contrast to our prediction, the latency and duration of reconciliation were not influenced by the context. Overall, because in the feeding context animals are busy eating, opponents often do not engage in reconciliation, but if they do they invest the same amount of time in the interaction as in the non-feeding context. Nevertheless, during feeding they just reconcile in a simple way. We also found that victims were more likely to initiate reconciliation than aggressors in the non-feeding context, whereas, after conflict over food, both victim and aggressor initiated friendly reunions at the same frequency. Indeed, in both non-feeding and feeding contexts victims were subordinate to the aggressors in most of the conflicts (non-feeding context: 86% of conflicts; feeding context: 89.9% of conflicts). Therefore, it may be suggested that, during feeding, both subordinate victims and dominant aggressors perceive a stronger advantage from the maintenance of a peaceful relationship, because restoring the relationship quickly can allow them to return to feeding, reducing the risk of further interruptions (see\u00a0\\[[@RSOS171553C52]\\]).\n\nThe intensity of aggression affected the likelihood of reconciliation occurring. According to our prediction, high-intensity aggressions (i.e. conflict with physical contact) were followed by PC affiliative interactions more often than low-intensity aggressions, but, contrary to our expectations, the latency of reconciliation did not differ between them. The intensity of reconciliation also varied according to the intensity of the preceding conflict, such that reconciliation was more intense after higher intensity conflicts. This suggests that relationships between opponents may be increasingly disturbed with increasing conflict intensity and that the occurrence of reconciliation and its intensity may help to restore the disrupted relationships.\n\nBoth the security of the relationship and the competitive difference between opponents influenced some aspects of reconciliation. In particular, the security of the relationship (measured by the asymmetry in the exchange of affiliative behaviours) affected the duration of reconciliation. Our results showed that longer PC affiliative interactions occurred in dyads with a low relationship security (i.e. where the affiliative asymmetry was high and biased towards the aggressor). Therefore, apparently and in contrast to our prediction and previous findings\u00a0\\[[@RSOS171553C53],[@RSOS171553C57]\\], the duration of reconciliation increases with a decrease in the security of the opponent\\'s affiliative relationship. Nevertheless, because victims were responsible for the initiation of reconciliation after the majority of conflicts, they indeed engaged in longer reconciliations with those aggressors from whom they have a higher probability of being reciprocated.\n\nFurthermore, dyads in which the asymmetry of aggression was more heavily biased towards the aggressor reconciled with more intense interactions. This result may be explained by the rank relationship between the two opponents; in fact, the intensity of reconciliation may increase with the increasing risk of re-aggression by the dominant aggressor towards the subordinate victim, thus with a higher difference in the competitive abilities of the two opponents. Indeed, we found that the higher the rank distances between opponents, the more likely the occurrence of reconciliation. Since in most of the conflicts the victim was subordinate to the aggressor (91.5%), the higher occurrence of reconciliation between opponents more distant in rank may be directly linked to the potentially higher costs for the subordinate victims in case of a renewed aggression, and thus the higher benefits of restoring their relationship with the dominant aggressors.\n\nOverall, we found evidence that conflicts in wolves may disrupt the relationship between opponents and that the variation of reconciliation is a consequence of a compromise between costs and benefits of the interaction and for the particular relationship of the two opponents. Our results are in line with the suggestion that reconciliation functions to repair relationships because it reduces the probability of renewed aggression (*relationships repair hypothesis*, e.g.\u00a0\\[[@RSOS171553C20],[@RSOS171553C84],[@RSOS171553C88]--[@RSOS171553C91]\\]; see also\u00a0\\[[@RSOS171553C43],[@RSOS171553C92]\\]).\n\n5.. Conclusion {#s5}\n==============\n\nIn conclusion, the results of our dog--wolf comparison suggest that wolves are more communicative in their social interactions after conflicts, seeking their partner out to re-establish the relationship, whereas dogs appear to adopt an avoidance strategy. This is in line with the avoidance observed in a feeding context when potential for conflicts is high\u00a0\\[[@RSOS171553C81]\\] and with the different social ecology of the two species\u00a0\\[[@RSOS171553C42]\\]. Nevertheless, contrasting results found in other studies highlight the need for further investigations in order to clarify which factors may promote and modulate reconciliation in domestic dogs. Contrary to dogs, we confirmed the occurrence of reconciliation in wolves, providing, for the first time, new insight into the factors affecting its variation.\n\nSupplementary Material\n======================\n\n###### Cafazzo_et_al_S1_S2_tables_ESM.doc\n\nSupplementary Material\n======================\n\n###### Cafazzo_et_al_David_Score_Calculation_ESM\n\nSupplementary Material\n======================\n\n###### Cafazzo_et_al_Data_used_for_analyses_ESM\n\nThe Wolf Science Center was established by Zs\u00f3fia Vir\u00e1nyi, Kurt Kotrschal and Friederike Range and we thank all the helpers who made this possible, hence indirectly supporting this research. We thank Els Weinans, Andrea Sommesse and Mafalda Velez de Castro for help with data collection. We are also grateful to the Wolf Science Center staff and many students and volunteers for their devotion and assistance. We further thank many private sponsors including Royal Canin for financial support and the Game Park Ernstbrunn for hosting the Wolf Science Center. Finally, we also would like to thank the editor, two anonymous referees and Alexandra Protopopova, whose suggestions greatly improved the manuscript.\n\nWe recorded fewer PC--MC pairs than the total number of aggressive interactions observed for two main reasons. First, most of the aggression came from data recorded during the 'focal animal sampling', which was used to independently assess the relationship between the pack animals. Second, during the 'behavioural sampling' periods, it was not always possible to follow the victim for the entire 10 min period, because the subjects sometimes moved fast and went out of sight. Furthermore, for the same reasons, it was not always possible to collect the MC for each PC period collected. As a result of all these factors, the final number of PC--MC was lower than the total number of aggressive interactions recorded.\n\nEthics {#s6}\n======\n\nThe study was purely observational with no manipulation of the animals. The relevant committee that allows research without special permissions regarding animals is: Tierversuchskommission am Bundesministerium fuer Wissenschaft und Forschung (Austria).\n\nData accessibility {#s7}\n==================\n\nThe datasets supporting this article have been uploaded as part of the electronic supplementary material.\n\nAuthors\\' contributions {#s8}\n=======================\n\nS.C. conceived, designed and coordinated the study, participated in data analysis and drafted the manuscript. S.M.P., Z.V. and F.R. helped to design the study and draft the manuscript. S.C. and M.L. collected the data, coded the videos and did the initial statistical analyses. All authors gave final approval for publication.\n\nCompeting interests {#s9}\n===================\n\nWe have no competing interests.\n\nFunding {#s10}\n=======\n\nS.C. was supported by the Austrian Science Funds (<https://www.fwf.ac.at/en/>) project no. M1400-B19. S.C. and S.M.P. were supported by funding to F.R. from the European Research Council under the European Union\\'s Seventh Framework Programme (<https://ec.europa.eu/research/fp7/index_> en.cfm) by ERC Grant Agreement no. (311870). Z.V. was supported by the Wiener Wissenschafts-Forschungs- und Technologiefonds (<http://www.wwtf.at/>) project CS11-026. S.M.P. was further supported by funding from the Vienna Science and Technology Fund (WWTF CS15-018). The authors further thank many private sponsors, including Royal Canin, for financial support. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.\n\n[^1]: Electronic supplementary material is available online at <https://dx.doi.org/10.6084/m9.figshare.c.4135796>.\n\n[^2]: ^a^See text for further details.\n"}
+{"text": "Introduction {#S1}\n============\n\nFrailty in older adults is characterized by a nonspecific state of vulnerability, specifically, reduced multisystem physiological reserve, decreased resistance to stressors, and increased risk for adverse health outcomes ([@B1]--[@B3]). The relationship between physical frailty (PF) and cognitive impairment has been recognized for decades and thought to be connected by their similar pathophysiological mechanisms ([@B4]). There is a frequent coexistence of frailty and cognitive impairment, which also had cumulative effect mortality ([@B5], [@B6]). Hence, demanding the need of a novel entity to discriminate associated risk factors of both PF and cognitive impairment, as well as to provide better prevention and therapy strategies for those frail patients who are prone to cognitive disorders ([@B7]). Accordingly, the International Academy on Nutrition and Aging (IANA) and the International Association of Gerontology and Geriatrics (IAGG) proposed a new construct \"cognitive frailty\" (CF) ([@B8]), to define a condition characterized by the simultaneous presence of PF and cognitive impairment in the absence of dementia, which might be marked as a promising target for the prevention of age-related disorders ([@B9]). In this new concept, PF precedes the onset of cognitive impairment ([@B8], [@B9]), thus, intervention programs targeted to improve frailty may prevent late-life cognitive disorders. Several studies have investigated the concept of CF, and reported the prevalence of CF to be ranging between 10.7 and 40% in clinical settings and 0.9--12.0% in community-based population ([@B7]). A recent study claimed that CF is a precursor of neurodegenerative processes and could be potentially reversed ([@B10]). Furthermore, other research reported that CF was a useful predictor of mortality and dementia, even after adjusting for vascular risk factors and depressive symptoms ([@B11]).\n\nAccordingly, this new concept poses challenges as well as opportunities for geriatricians. Nonetheless, the validity and utilization of CF in the Chinese population which represents the largest and fastest growing aging population in the world remains unclear. Hence, we aimed to explore the prevalence and associated factors of CF in the Chinese population.\n\nMaterials and Methods {#S2}\n=====================\n\nParticipants {#S2-1}\n------------\n\nData were obtained from the China Comprehensive Geriatric Assessment Study (CCGAS) (2011--2012) which used a two-step statistical sampling techniques including cluster, stratification, and random selection ([@B12]). In the first step, seven cities representing the six main regions of China were selected: Beijing, Xi'an, Harbin, Chengdu, Chongqing, Changsha, and Shanghai. Then, participants from the above seven cities were selected regarding urban--rural areas, age, and gender in the second step. Further details regarding the CCGAS have been reported ([@B12]). Finally, 9,694 elderly participants were enrolled including 6,867 elderly adults living in community and 2,827 inpatients. Of the 6,867 community-dwelling older adults, 5,708 of those without a self-history of diagnosed dementia and with Comprehensive Geriatric Assessment-Frailty Index (CGA-FI) and Mini-Mental State Examination (MMSE) data were included. This study was reviewed and approved by the ethics committee of Xuanwu Hospital of Capital Medical University. All subjects gave written informed consent in accordance with the Declaration of Helsinki.\n\nThe Construct of CF {#S2-2}\n-------------------\n\nCognitive frailty was operationalized using CGA-FI for the evaluation of PF and MMSE for the evaluation of cognitive status. Participants were stratified by educational level to determine thresholds for global cognition. The thresholds for those who were illiterate, or attended at most primary school, middle school, or university were \u226417, \u226420, \u226422, and \u226424, respectively ([@B13]). Participants who scored below the threshold value for their education group were recorded as cognitive impairment. As we previously published ([@B14]), CGA-FI was measured by 68 parameters, but in the current study the \"cognition\" variable of FI was excluded, thus 67 parameters from the following five variables remains: demographic characteristics, physical health, physical function, living behavior and social function, and mental health. Further detail on CGA-FI is in Table S1 in Supplementary Material. The FI score was defined as the cumulative sum of the total score of each index divided by 67. PF was defined as FI\u2009\u2265\u20090.25 ([@B15], [@B16]). Participants positive for both instruments were classified as having CF. Dementia was defined by a reported disease history diagnosed by a doctor. All of the participants were free of dementia.\n\nSociodemographics {#S2-3}\n-----------------\n\nUsing face-to-face interviews, we examined the sociodemographic variables, medical conditions, and physical function based on the standard CGA instrument ([@B12]). Area of residence was classified into urban and rural. Northern cities included Beijing, Xi'an, and Harbin, and southern cities included Chengdu, Chongqing, Changsha, and Shanghai. Participants were divided into the following five age groups: 60--64, 65--69, 70--74, 75--79, and \u226580\u2009years. Education status was recorded as illiterate or literate. Participants were also stratified by monthly income: USD\u2009\\<\u2009180 and USD\u2009\u2265\u2009180. Marital status was listed as married or widowed.\n\nMedical Conditions {#S2-4}\n------------------\n\nParticipants were considered to have a medical condition if they had a self-reported history of chronic disease diagnosed by a doctor. Clinical syndromes and geriatric syndromes were also recorded. Functional ability was assessed on the basis of activities of daily living (ADL) and instrumental activities of daily living (IADL) ([@B17]). Participants with one or more impaired ADL or IADL were defined as having a disability. The Geriatric Depression Scale was used to assess depression ([@B18]), with a total score ranging 0--30. A score of \u226511 typically indicates clinical depression. Comorbidity was defined as having \u22652 chronic diseases. A walking speed below the height-adjustment threshold value in 20\u2009m walking test was considered a slow pace. Regular exercise was defined as exercising for \u22653\u2009h/week over the past 12\u2009months. We also screened for a history of spontaneous fractures that occurred over the past 2\u2009years and falls that occurred twice or more often in the past year.\n\nStatistical Analysis {#S2-5}\n--------------------\n\nEpiData was used to establish the database, input, and automatically verify the data. Statistical analyses were performed by SPSS version 11.5 (Inc., Chicago, IL, USA). Count data were expressed as percentages, with standardized rates (weighted values) calculated using the national standard population composition ratio as at the Sixth National Census (2010). The bootstrap confidence interval for the prevalence was estimated based on 1,000 bootstrap samples and was bias-corrected. Chi-square tests were performed to compare percentages. Forward stepwise logistic regression was done to explore the association between the various factors as independent variables and CF as the dependent variable. Adjustments were made for sociodemographic variables and age-related factors. All statistical tests were two-sided and statistical significance was set at a *P*-value\u2009\\<\u20090.05.\n\nResults {#S3}\n=======\n\nFigure [1](#F1){ref-type=\"fig\"} presents the prevalence of CF in older adults. Of the 5,708 older adults, 187 individuals demonstrated CF, accordingly, the overall crude prevalence and standard prevalence of CF were 3.3% \\[95% confidence interval (CI)\u2009=\u20093.0--4.0%\\] and 2.7% (95% CI\u2009=\u20092.0--3.0%), respectively. The prevalence of CF was significantly higher in women and those living in rural areas. Moreover, the prevalence of CF increased with age, with the highest values recorded for participants aged \u226580\u2009years (9.8%) and the lowest values observed among those aged 60--64\u2009years (1.1%, Figure [1](#F1){ref-type=\"fig\"}).\n\n![Prevalence of cognitive frailty (CF) in older adults by gender, area, and age. The prevalence of CF among adults aged 60+ in China by China Comprehensive Geriatric Assessment Study, 2011--2012. Data were weighted by the national standard population composition ratio based on the Sixth National Census (2010). The total sample population in the analysis was 5,708.](fmed-04-00174-g001){#F1}\n\nThe effect of sociopsychological factors and physical function on CF are shown in Table [1](#T1){ref-type=\"table\"}. The prevalence of CF was higher among participants who were illiterate, had a low income, or widowed, and it was also found to be relatively higher among participants who had depression (Table [1](#T1){ref-type=\"table\"}). We observed a higher prevalence of CF among participants with comorbidities, disabilities, slow walking speed, vision impairment, and hearing impairment. Those who had less exercise and a low body mass index demonstrated a higher prevalence of CF. Participants who reported spontaneous fractures or falls showed a higher prevalence of CF (Table [1](#T1){ref-type=\"table\"}).\n\n###### \n\nEffect of sociopsychology factors and physical function on cognitive frailty (CF).\n\n                                Total   NC, *n* (%)    CF, *n* (%)   Weighted (%)   \u03c7^2^      *P*\n  ----------------------------- ------- -------------- ------------- -------------- --------- ---------\n  **Sociopsychology factors**                                                                 \n  Education                                                                                   \n  \u2003Illiterate                   1,040   941 (90.5)     99 (9.5)      9.1            156.423   \\<0.001\n  \u2003Not illiterate               4,668   4,580 (98.1)   88 (1.9)      1.6                      \n  Monthly income (US\\$)                                                                       \n  \u2003\\<180                        2,633   2,502 (95.0)   131 (5.0)     4.2            56.035    \\<0.001\n  \u2003\u2265180                         2,922   2,879 (98.5)   43 (1.5)      1.2                      \n  Marital status                                                                              \n  \u2003Married                      4,398   4,286 (97.5)   112 (2.5)     2.1            31.065    \\<0.001\n  \u2003Widowed                      1,306   1,232 (94.3)   74 (5.7)      5.3                      \n  Smoking                       1,628   1,573 (96.6)   55 (3.4)      2.4            0.075     0.784\n  Depression                    691     587 (84.9)     104 (15.1)    13.3           343.966   \\<0.001\n  **Physical function**                                                                       \n  Comorbidity                   3,249   3,089 (95.1)   160 (4.9)     4.4            64.678    \\<0.001\n  Disability                    414     292 (70.5)     122 (29.5)    29.1           966.413   \\<0.001\n  Slow walking speed            647     616 (95.2)     31 (4.8)      4.2            38.596    \\<0.001\n  Vision impairment             383     338 (88.3)     45 (11.7)     10.6           93.019    \\<0.001\n  Hearing impairment            272     231 (84.9)     41 (15.1)     15.1           125.446   \\<0.001\n  Less exercise                 1,210   1,093 (90.3)   117 (9.7)     8              198.067   \\<0.001\n  Fall                          252     218 (86.5)     34 (13.5)     12.9           86.831    \\<0.001\n  Fracture                      190     179 (94.2)     11 (5.8)      5.2            3.918     0.048\n  Low body mass index           317     296 (93.4)     21 (6.6)      5.9            11.876    0.001\n\nThe results of logistic regression models are shown in Table [2](#T2){ref-type=\"table\"}. In the context of the robust elderly individuals, comorbidity, depression, less exercise, hearing impairment, disability, and falls were independent factors influencing CF. Furthermore, when referred to the elderly individuals with PF, depression and hearing impairment were independently associated with CF.\n\n###### \n\nForward stepwise logistic regression for associated factors with CF.\n\n                       Model 1   Model 2                                            \n  -------------------- --------- --------------- --------- -------- --------------- ---------\n  Age (\u226575\u2009years)      4.237     1.955--9.183    \\<0.001   4.918    1.845--13.107   0.001\n  Area (rural)         5.670     2.454--13.099   \\<0.001   22.196   8.258--59.659   \\<0.001\n  Comorbidity          11.761    4.041--34.231   \\<0.001   /        /               /\n  Depression           11.371    5.302--24.387   \\<0.001   2.462    1.066--5.687    0.035\n  Less exercise        3.213     1.529--6.754    0.002     /        /               /\n  Hearing impairment   3.519     1.410--8.779    0.007     2.713    1.114--6.608    0.028\n  Disability           13.418    5.317--33.865   \\<0.001   /        /               /\n  Fall                 6.653     2.651--16.697   \\<0.001   /        /               /\n\n*Model 1: Logistic regression for risk factors associated with CF in the robust and CF population. The variables not in the equation were gender, smoking, marital status, education, income, walking speed, vision impairment, low body mass index, and fracture. Adjusted for sociodemographic variables and age-related factors*.\n\n*Model 2: Logistic regression for risk factors associated with CF in the population with physical frailty. The variables not in the equation were gender, smoking, marital status, education, income, walking speed, vision impairment, comorbidity, exercise, disability, fall, low body mass index, and fracture. Adjusted for sociodemographic variables and age-related factors*.\n\n*OR, odds ratio; CI, confidence interval; CF, cognitive frailty*.\n\nDiscussion {#S4}\n==========\n\nOur results showed that the standard prevalence of CF was 2.7%, and increased with age in the Chinese older population. Women and participants living in rural areas were found to be at higher risk for CF. Currently, owing to different definitions of CF, the prevalence varies from 0.9 to 40% across countries ([@B7], [@B19]--[@B21]). In the Singapore Longitudinal Ageing Studies, the estimated prevalence of PF coexisting with cognitive impairment was 1.8%; moreover, the prevalence of pre-frailty and frailty coexisting with cognitive impairment was 10.7% and was associated with more severe functional disability, hospitalization, poor quality of life, and mortality ([@B19]). In an Italian study, the prevalence of CF was 4.4% among older adults, and those with CF showed more severe disability than those without frailty ([@B22]). Similarly, the findings of our study also showed that older participants with comorbidity, disability, and fall were independently associated with CF.\n\nPast studies have shown PF to be associated with cognitive decline in older adults ([@B23], [@B24]). Compared to the individuals with only cognitive impairment (i.e., without PF), those with CF showed poorer scores on executive and attention tests ([@B25]). Furthermore, baseline frailty was found to be strongly associated with subsequent changes in cognition assessed by MMSE ([@B26], [@B27]) and higher risk for non-AD dementia ([@B28]). In addition, studies have shown frailty state transitions to be associated with cognitive deterioration in participants with mild to moderate Alzheimer disease ([@B29]). Another study reported that PF was a stronger indicator of cognition than age ([@B30]).\n\nHearing impairment is one of the principal causes of chronic disability in older adults ([@B31]), and our study showed that old individuals with hearing impairment were independently associated with CF either in robust or frail population. A previous study suggested that hearing impairment to be a prognostic marker of frailty in older age and could identify older persons with adverse health outcomes ([@B31]--[@B33]). Recently, CF was considered to embody two different manifestations: slow gait and low cognition, which may share a common underlying mechanism ([@B34]). Furthermore, Verghese et al. validated a new Motoric Cognitive Risk syndrome, which was defined as the presence of cognitive complaints and slow gait, and found it was associated higher risk of developing dementia ([@B35]). Our study also showed that participants with slow gait speed demonstrated a higher prevalence of CF; however, gait speed was not an independent factor per the logistic regression analysis. Nevertheless, other studies found that gait speed was associated with severity of cognitive impairment, after adjusting for age, gender, and age-related factors ([@B36]).\n\nWe used the CGA-FI to assess for PF in this study, while majority of the other studies used the Fried criteria ([@B11], [@B19], [@B21], [@B22], [@B25], [@B34], [@B37], [@B38]). However, it is noteworthy that in fact, the construct of CF itself is rather controversial, and the past studies on CF implemented non-uniform operational criteria both for assessing PF and cognitive impairment ([@B7]). Moreover, the operational definition of PF still remains unresolved ([@B8]), which might partially explain the non-uniformity. Although the consensus paper of IANA/IAGG definition of CF has been described by Fried criteria for PF ([@B8]), an obvious question emerges: can frailty be defined by FI in the construct of CF? It has not yet fully explored in the literature. Fried criteria and FI are the two most commonly used measurements in the world, and they share common characteristics and complementarity when applied to the Chinese older population ([@B39]); moreover, the preliminary results of our study further demonstrated the feasibility of this method in a Chinese population. A previous study reported that using a multi-dimensional FI, both baseline status and within-person changes in frailty were predictive of cognitive trajectories ([@B40]); furthermore, this tool was shown to be effective in identifying individuals at high risk for cognitive decline ([@B41]). Thus, FI may be a promising instrument for determining the vulnerability of dementia and was also recommended to be used for assessing CF ([@B42]). CGA can be used to identify the medical, psychosocial, and functional capabilities of older adults ([@B43]), in addition, CGA-FI can predict both cognitive changes and mortality ([@B27]); therefore, CGA-FI has applications in frailty measurements in elderly individuals with cognitive impairment.\n\nThis is the first study to report the prevalence of CF and the associated factors in China; furthermore, our results show that the CGA-FI is a feasible tool for defining CF. CGA is regularly used as an assessment tool for old individuals. In older adults, most health deficits are known to be associated with late-life cognitive impairment ([@B5]). Our study provides a quick and simple method to identify CF in any individual with CGA data; furthermore, this approach allows for rapid diagnosis of CF, such that prevention of and intervention for dementia and disability can be established at an early stage ([@B44]). However, these results must be interpreted in light of several limitations. First, we chose only seven cities in China; although our methodology was strong, the small number of cities and participants included may have biased our results. Second, this is a cross-sectional study, further longitudinal studies that incorporate frailty and cognition and randomized controlled trials are needed to provide more information on the cause-and-effect relationship of frailty and cognition, risk factors of CF and the transition to dementia. Third, this is a study designed for screening tools, so dementia was defined by a reported disease history diagnosed by a doctor, and a lack of some important examinations specific for dementia such as neuroimaging tests and other neuropsychiatric scales, in this perspective, some patients with potential dementia might be included in this population. Besides, the relationship between PF and cognitive impairment was not explored in the study. The existing cognitive decline in this study is uncertain to be driven by the physical domain makes the criteria arbitrary to be defined as CF, which indicates that there is a disparity in our operational construct and the construct recommended by IANA/IAGG consensus. However, it is worthwhile to note that both the clinical diagnosis of dementia and the identification of non-neurodegenerative cognitive impairment require a comprehensive neuropsychological battery which is hard to apply in busy daily clinical practice. Fourth, only one kind of frailty measurements was used in this study, so further research on the comparison between FI model and Fried model needs to be conducted to confirm with our findings. Additionally, biomedical variables were not included in this study. Last, in spite of the fact that CGA is most evidence-based for detection and severity grade of frailty, it is bounded by the resource-intensive and time-consuming process, thus further simple and more efficient instruments are expected to be developed for daily clinical work ([@B45]).\n\nIn conclusion, while preliminary, this work contributes to expanding the knowledge that CF may be a promising new concept for the assessment of vulnerability in patients with cognitive impairment, as well as identifying individuals at high risk for negative outcomes. Our study identified that depression and hearing impairment were independent associated factors of CF in elderly individuals with frailty in China, showing the possibility of controlling further cognitive deterioration in a population with PF. Our results shed new light on the identification and related factors for CF and suggest that many health deficits are associated with CF. Therefore, in order to narrow the gap between the hopefully promising concept and the limited evidence from current studies, especially in the situation that CF was still considered to be far away from clinical and research scenario ([@B7]), the reliability and predictive validity of the operational definition of CF should be clarified in future studies, as well as the underlying biological characteristics. Prospective studies will be needed to address the early intervention strategies to integrate physical and cognitive function.\n\nEthics Statement {#S5}\n================\n\nThis study was carried out in accordance with the recommendations of the ethics committee of Xuanwu Hospital of Capital Medical University. All subjects gave written informed consent in accordance with the Declaration of Helsinki.\n\nAuthor Contributions {#S6}\n====================\n\nPC and ZT contributed to the design of the work. LM drafted the manuscript and wrote it together with PC, ZT, and YZ. LZ and YL contributed to the analysis and interpretation of data. All the authors contributed to writing the paper and revising it critically and gave final approval of this version.\n\nConflict of Interest Statement {#S7}\n==============================\n\nThere are no ethical/legal conflicts involved in the article. The reviewer BF and handling editor declared their shared affiliation.\n\nWe would like to thank the CCGAS participants and their families for taking part in the study. We acknowledge all the colleagues who participated in the study from the following hospitals: Xuanwu Hospital, Beijing Hospital, Beijing Institute of Geriatrics, General Hospital of PLA, Peking University Third Hospital, The First Affiliated Hospital of Xi'an Jiaotong University, Xijing Hospital of The Fourth Military Medical University, The First Affiliated Hospital of Harbin Medical University, West China Hospital of Sichuan University, First Affiliated Hospital of Chongqing Medical University, Xiangya Hospital of Central South University, The Third Xiangya Hospital of Central South University, and Ruijin Hospital of Shanghai Jiao Tong University School of Medicine.\n\n**Funding.** This work was supported by the following grants: National Natural Science Foundation of China (81600927), the Ministry of Health Welfare Industry Special Fund (201002011), and Beijing Municipal Health Bureau Research Fund (Jing 17-12).\n\nSupplementary Material {#S8}\n======================\n\nThe Supplementary Material for this article can be found online at <http://www.frontiersin.org/article/10.3389/fmed.2017.00174/full#supplementary-material>.\n\n###### \n\nClick here for additional data file.\n\nAbbreviations {#S9}\n=============\n\nCF, cognitive frailty; PF, physical frailty; CGA, Comprehensive Geriatric Assessment; FI, frailty index; MMSE, Mini-Mental State Examination.\n\n[^1]: Edited by: Matteo Cesari, Centre Hospitalier Universitaire (CHU) de Toulouse, France\n\n[^2]: Reviewed by: Jagadish K. Chhetri, G\u00e9rontop\u00f4le-Institute on Aging, Toulouse University Hospital, France; Bertrand Foug\u00e8re, Centre Hospitalier Universitaire (CHU) de Toulouse, France\n\n[^3]: Specialty section: This article was submitted to Geriatric Medicine, a section of the journal Frontiers in Medicine\n"}
+{"text": "Introduction {#sec1}\n============\n\n*Nocardia* is a soil-dwelling, aerobic, facultative intracellular bacteria with a challenging growth pattern to discern when sampled from cutaneous infections.[@bib1], [@bib2] This difficulty is partly attributed to similarity with the *Actinomyces* species, slow growth rate, and low incidence.[@bib3] Although considered opportunistic, immunocompetent hosts may become infected with a primary cutaneous infection following penetrative trauma.[@bib4] *Nocardia* causes localized destruction or systemic infection, with a particular affinity for the pulmonary and central nervous systems once disseminated. More prevalent in tropical countries, *Nocardia* has various cutaneous presentations, belonging with differential diagnoses including *Actinomyces*, sporotrichosis, and tuberculosis because of similarity in presentation.[@bib2] Cutaneous infections left untreated can progress into a mycetoma, a chronic subcutaneous, granulomatous infection with fistulous tracts leading to invasion of the musculoskeletal system, lymphatics, or the central nervous system.[@bib3], [@bib5]\n\nCase report {#sec2}\n===========\n\nThis patient is an immunocompetent 56-year-old Hispanic man with no significant medical history, presenting with painful posterior cervicothoracic skin lesions ([Fig 1](#fig1){ref-type=\"fig\"}).Fig 1Prior to laminectomy.\n\nOnset was estimated to be 20\u00a0years ago; however, in the last 6\u00a0months, pain and purulent drainage from fistulous tracts had progressively increased. Evaluations and laboratory analysis by previous institutions\\' dermatology and infectious disease departments were nonrevealing. Earlier differential diagnoses including acne conglobata, *Actinomyces*, and eumycetoma were addressed; however, a definitive diagnosis was never established because of persistently negative tissue biopsies, blood cultures, or response to treatment.\n\nThree months before presentation, the patient had motor and sensory deficits in bilateral lower extremities with resultant gait disturbances. Magnetic resonance imaging (MRI) found epidural abscesses, fistulous tracts, vertebral osteomyelitis, and lower cervical/upper thoracic spinal cord compression. He subsequently underwent laminectomy of C7 and T1-4, removal of phlegmon, and osseous biopsies. Tissue biopsies found neutrophilic clusters that failed to culture. Suspicion was maintained for *Actinomyces* infection due to a single Gram stain demonstrating its branching formation. The patient was discharged empirically on ceftriaxone, metronidazole, and daptomycin thereafter. Lack of improvement 1\u00a0month postoperatively prompted regimen alteration to 8\u00a0weeks of metronidazole with 3\u00a0weeks of penicillin V and doxycycline. Despite this extensive course, symptoms persisted, and a suspicion for an infectious etiology remained.\n\nAt presentation to our service, lesions had expanded with increased drainage and worsening pain. New symptoms included right-sided rib pain, cough, low-grade fever, fatigue, and an unintentional 25-pound weight loss. Immunocompetence was assessed via negative rapid plasma reagin, p24 HIV, coccidioides antigen, QuantiFERON, hepatitis panels, and HgA1C (5.8%). MRI and computed tomography of the thorax found that the spinal cord remained preserved; however, there was contiguous advancement of abscesses and fistulous tracts into the thoracic cavity with osteomyelitis of the vertebral column and rib cage ([Fig 2](#fig2){ref-type=\"fig\"}).Fig 2**A**, MRI, T1 weighted with contrast axial view. Shows anterior advancement of phlegmon into right thoracic cavity. **B**, MRI, T1 weighted with contrast sagittal view. Shows anterior advancement of phlegmon into right thoracic cavity.\n\nAbscess fluid and tissue were obtained for repeat culture. Because of failure of prior *Actinomyces* sp therapies and culture techniques, the differential diagnosis broadened to include *Nocardia*. Clinical suspicion prompted treatment despite lacking confirmation from molecular techniques, culture, or staining (Gram, acid-fast bacilli, Fite, periodic acid--Schiff, or Gomori methenamine-silver nitrate stain). Specific consideration to *Nocardia\\'s* particular growth pattern showed eventual growth of *Nocardia* with preliminary confirmation per microbiology and pathology report ([Fig 3](#fig3){ref-type=\"fig\"}). With a successful culture, a reference laboratory performed gene sequencing further classifying the sample as *Nocardia vulneris*.Fig 3Fite stain shows weakly acid-fast organisms. (Original magnification: \u00d7100.)\n\nClinical suspicion was imperative, as culture necessitated 27\u00a0days for identifiable growth. Trimethoprim-sulfamethoxazole (TMP-SMX) and meropenem were initiated, resulting in substantial improvement of leukocytosis, from 17.3 to 12.2 in 2\u00a0days, alongside subjective improvement in pain and drainage. Responsiveness to *Nocardia* therapy encouraged the addition of amikacin on day 3 of treatment, further accelerating his improvement. Appearance of his fistulous tracts markedly improved over the lengthy inpatient antibiotic administration: 3\u00a0months of TMP-SMX and meropenem, with 4\u00a0weeks of amikacin, with frequent metabolic panels. The patient was discharged home on 6\u00a0months of daily oral moxifloxacin and lifetime oral TMP-SMX. Primary care provided follow-up and monitored for nephrotoxic effects of associated therapies.\n\nDiscussion {#sec3}\n==========\n\nMycetomas are considered a neglected tropical disease and rare within developed nations.[@bib6] Creating extensive disfigurement, they are defined by chronic, destructive granulomas, tumefaction, and draining tracts from subcutaneous nodules.[@bib3], [@bib4], [@bib6] Requiring extensive time for growth, *Nocardia* cultures rarely develop before 48\u00a0hours, which further complicates the diagnosis.[@bib1] Because of this constraint, clinical suspicion remains imperative for prompt therapy initiation. Suspicion for a Nocardial mycetoma should ideally involve consulting a tropical identification team if available. Prior antibiotic application reduces the sensitivity of microbial cultures producing additional diagnostic difficulties. Serologic or other molecular techniques such as the B-cell P61 marker, 16S rRNA polymerase chain reaction--based assay, or 54-kDa antigen, are more expedient diagnostic options for laboratories with molecular capabilities.[@bib7] These techniques may require referral to reference laboratories outside of an institution\\'s standard referral. Reference laboratories may provide specific susceptibilities furthering patient safety and response to therapies.[@bib7] Molecular options and tropical identification teams were unavailable at this patient\\'s institutions, exemplifying this report as an educational reminder of these options and emphasis on clinical suspicion.\n\n*Nocardia* resides within the soil, which facilitates primary cutaneous infections from penetrative injury to predominantly involve the foot.[@bib4], [@bib5], [@bib6], [@bib8] Large-scale studies of *Nocardia* mycetoma, 400 to 3,000 cases, report affecting the back at only 7.88% compared with the foot at 60.29% to 62.44%.[@bib3], [@bib5], [@bib8] In this patient\\'s native country of Mexico, *Nocardia brasiliensis* is the most common etiologic agent of bacterial mycetoma (65.58%-78.21%).[@bib5], [@bib8] *N vulneris* species in this case, is closely related to *N brasiliensis* phylogenetically at 99.37%.[@bib9]\n\nPrimary cutaneous nocardiosis has multiple presentations, mimicking other diseases including tuberculosis, neoplasms, or lymphocutaneous infections such as *Sporotrichosis.*[@bib1], [@bib6] Diagnosis may be further complicated by seeding of a disseminated infection precipitating distal cutaneous infectious sites.[@bib3] The presence of pain does not distinguish an origin; however, it may provide additional clinical suspicion for a nocardial agent. Differentials such as *Sporotrichosis* and leishmaniasis typically manifest as painless ulcers.[@bib2]\n\nTreatments using sulfonamides are highly effective in simple cutaneous infections; however, alternatives exist in minocycline, imipenem, and moxifloxacin.[@bib10] Complicated cases involving dissemination, neural tissue, or bone may require combination therapy with amikacin and meropenem as in this case.[@bib10] Combination therapy, such as TMP-SMX and amikacin, has been successful when used for several months.[@bib9], [@bib10] Aggressive therapy lends concern for nephrotoxic effects; therefore, strict monitoring of electrolytes and kidney function remain an additional cause for inpatient treatment beyond that of intravenous application.\n\nUnderstanding of serologic laboratory confirmation techniques in combination with clinical suspicion may aid in proper treatment of recalcitrant cutaneous *Nocardia* infections. Treatments may need to be aggressive while awaiting serologic or culture confirmation, exemplified by the exorbitant timeframe necessitated for growth and definitive diagnosis. Although mycetomas are rare within developed nations, cognizance of the possible introductions and apparent difficulties in standard confirmation is necessary. This finding is especially pertinent within communities possessing increasing transient populations from endemic countries.\n\nFunding sources: None.\n\nConflicts of interest: None disclosed.\n"}
+{"text": "The editors are deeply grateful to the following reviewers (members and nonmembers of the Botanical Society of America) who have generously given of their time to review manuscripts submitted to\u00a0*Applications in Plant Sciences*. The list includes those who reviewed manuscripts from December 21, 2016, to December 31, 2017. If any reviewer\\'s name has been omitted, we apologize, and ask that a note be sent to the Managing Editor so the list can be corrected.\n\nAgrawal, Anurag\n\nAllen, Julia\n\nBack, Kyoungwhan\n\nBallard, Harvey\n\nBlackmon, Heath\n\nBlazier, Chris\n\nBlischak, Paul\n\nBonnet, Pierre\n\nBrosi, Berry\n\nBruns, Tom\n\nCampbell, Lesley\n\nChambers, Alan\n\nChambers, Sally\n\nChaney, Lindsay\n\nChang, Peter\n\nCruse\u2010Sanders, Jennifer\n\nde Vere, Natasha\n\nDeyholos, Michael\n\nDrummond, Emily\n\nDrummond, Francis\n\nEdger, Patrick\n\nEllwood, Elizabeth\n\nEndresen, Dag\n\nEscribano\u2010Rocafort, Adri\u00e1n G.\n\nEstep, Matt\n\nFaivre, Amy\n\nFine, Paul\n\nFishbein, Mark\n\nFolk, Ryan\n\nForrest, Laura\n\nFrom, Margaret\n\nGardner, Dale\n\nGoodell, Karen\n\nGrote, Paul\n\nHamilton, Matthew\n\nHegeman, Adrian\n\nHeyduk, Karolina\n\nHodel, Richard\n\nHufford, Matthew\n\nJarvis, David\n\nJinga, Percy\n\nJohnson, Matthew\n\nJolles, Diana\n\nKimura, Mitsuhiro\n\nLagomarsino, Laura\n\nLandis, Jacob\n\nLatvis, Maribeth\n\nLavoie, Claude\n\nLavor, P\u00e2mela\n\nLeonard, Joan\n\nLiang, Chengzhen\n\nLu\u2010Irving, Patricia\n\nMacklin, James\n\nMacLeod, Norman\n\nMandel, Jennifer\n\nMarx, Hannah\n\nMcDonnell, Angela\n\nMei, Wenbin\n\nMerritt, Benjamin\n\nMichaels, Helen\n\nMiller, Andrew\n\nMiller, Ian\n\nMoore, Abigail\n\nMoore, Gloria\n\nMorse, Clinton\n\nMueller, Rebecca\n\nMuellner\u2010Riehl, Alexandra\n\nMunson, Richard\n\nMusah, Rabi\n\nNagel, Dawn\n\nOgutcen, Ezgi\n\nOrbovic, Vladimir\n\nOrli, Sylvia\n\nPanchen, Zoe\n\nParducci, Laura\n\nParks, Matthew\n\nPease, James\n\nPec, Gregory\n\nPetit, Sophie\n\nPhilpott, Megan\n\nPyle, Richard\n\nRafferty, Nicole\n\nRanathunge, Chathurani\n\nReynolds, Richard\n\nRiccardi, Greg\n\nRichards, Chris\n\nRoulston, T\\'ai\n\nRuas, Claudete\n\nSarangi, Debalin\n\nSchade, Sven\n\nScherm, Harald\n\nSchneider, Adam\n\nSchori, Melanie\n\nSerrano Serrano, Martha\n\nSparks, Erin\n\nStucky, Brian\n\nSweeney, Patrick\n\nTank, David\n\nTepe, Eric\n\nTheroux\u2010Rancourt, Guillaume\n\nThompson, Pamela\n\nTunison, Robert\n\nUribe\u2010Convers, Simon\n\nVanBuren, Bob\n\nVieira, Leila\n\nVonderheide, Anne\n\nWeitemier, Kevin\n\nWhipple, Clinton\n\nWilf, Peter\n\nWinkler, Daniel\n\nWoods, Justin\n\nXiao, Yuguo\n\nZavada, Michael\n\nZenil\u2010Ferguson, Rosana\n\nZhang, Wenheng\n\nZweck, Justin\n"}
+{"text": "Introduction {#s1}\n============\n\nThe Court of Justice for the European Union[^1^](#FN1){ref-type=\"fn\"} ('Court') has been an important driver of EU integration. The legal order of the European Union is an interlinked network of European and national courts, all effectively functioning as European courts. Through this system, most of the costs of enforcing EU laws are shifted onto the legal systems of the Member States. Enforcement within the Member States happens largely through the actions of litigants and national courts, essentially outsourcing governance ([@R22]).\n\nThe reach of the Court is limited to the cases that are brought before it. Its caseload is comprised of enforcement actions brought by the Commission and preliminary references from the national courts of the Member States. National courts thus function as 'agenda-setters' in the European legal system ([@R10], p. 158). However, preliminary references are unevenly distributed across the Member States of the EU, and some Member States' preliminary references are concentrated in particular policy areas.\n\nHow do we account for the different usage of the European legal process across the Member States? This paper presents part of a larger project completed for the author\\'s doctoral thesis, which sought to explain variation in rates of social policy preliminary references[^2^](#FN2){ref-type=\"fn\"} across the member states of the EU-15 from 1996--2009. The study began with a quantitative analysis, which was followed by a Qualitative Comparative Analysis ([@R34], [@R35]). Finally, two case studies were conducted in the UK and France. These case studies helped to unpack the complexities of how the preliminary references process functions in the Member States.\n\nExplaining Variation in Rates of Preliminary References {#s2}\n=======================================================\n\nPotential explanations for variation in national rates of preliminary references have been much theorised (See, e.g. [@R1], [@R2], [@R3], [@R8], [@R9], [@R11], [@R17], [@R30], [@R43], [@R42], [@R44], [@R45], [@R46]). Because most prior studies have not been specific to an issue area, the explanations are largely general and relate to the legal system as a whole. Factors considered have include what Alter calls the 'friendliness' of a legal system to litigation (2000); judicial review ([@R1], [@R30]); the effects of exposure to the EU ([@R43]); 'containment' by the preferences of political and judicial actors ([@R12]); and the power of organised interest groups ([@R17], [@R11], [@R3], [@R38]). Although these more general factors play an important role in determining rates of preliminary references, they do not address the effects of policy misfit, which has been identified as an important factor in the Europeanisation of domestic policy ([@R7], [@R36]). The more general factors also do not explain why references from some Member States are concentrated in particular issue areas.\n\nBy focusing on preliminary references in the area of social policy, it was possible t consider the effects of underlying policy fit on rates of references. Social policy is an important area for EU law. The Court has been a driver in the increasing Europeanisation of social policy. It has chipped away at Member State control over who has the right to receive services under the national welfare state, where those services may be consumed and who can provide those services ([@R24]). The role of the Court is essential to a governance-based account of the development of the European Union social dimension, especially in terms of regulatory policy-making ([@R26], [@R27]).\n\nThe findings of the quantitative and QCA elements of this study support the importance of national legal culture, judicial activism from lower level courts ([@R25]; [@R37]); and compliance with the EU ([@R15]; [@R14]). In terms of policy misfit, Bismarckian welfare states[^3^](#FN3){ref-type=\"fn\"} had higher rates of social policy preliminary references. This reinforces the findings of [@R28] in the her study of cases in the area of social security for migrant workers. The percentage of the population that belongs to a union ('union density') was also important, confirming the importance of the organizing power of interest groups ([@R3]; [@R11]; [@R17]; [@R29]; [@R38]). The case studies presented in this paper were used to inform and explain these findings.\n\nMethods {#s3}\n=======\n\nThe theoretical argument of this study is that the different usage of the preliminary reference procedure arises from the institutional structures of the national legal system and of the welfare state, as well as because of the agency of actors within the national system. France and the UK were chosen through purposive sampling because they provide the best insight into the research question. They are alike enough to be comparable but have differing rates of referral. The UK has a rate of social policy preliminary references twice that of France. The UK and France are comparably sized and have similar long histories with the EU. However, they differ in key ways in their legal systems and welfare states.\n\nEvidence was collected via documentary analysis and 25 expert interviews. A starting point for the documentary analysis was the cases that comprise the preliminary references sample. There were 27 relevant preliminary references from France, and 69 from the UK in the period. Additionally, searches were conducted in national media, legal and academic databases in order to identify relevant documents and news items.\n\nThen, 25 semi-structured expert interviews were conducted with lawyers, interest group officials, court officials and legal experts familiar with the history of preliminary references in the Member States. 8 were conducted in France, 11 in the UK, 3 with officials at the Court of Justice, and 3 with officials associated with the Commission in Brussels.[^4^](#FN4){ref-type=\"fn\"} Participants were offered anonymity. Those who are named in this article explicitly consented to be named. The circumstances of those participants who want to remain anonymous have been masked to ensure that anonymity.\n\nThe resulting data were hand-coded for themes and sub-themes. The themes fall into the three types of explanatory factors. First, there are themes related to the underlying policy structure of the Member State. Second, there are themes related to structural factors of the national legal system. Third, there are themes related to the actions of individuals and groups.\n\nResults for the United Kingdom {#s4}\n==============================\n\nPolicy Structure Themes {#s5}\n-----------------------\n\n### Free Movement of Persons/European Citizenship {#s6}\n\nThe UK is classified as a low-spending Beveridgean welfare state ([@R4]). It provides low-level, often means-tested, support. Apart from meeting any means test requirements, eligibility to receive benefits is based upon UK residence.\n\nThe UK was also one of only three established EU Member States that permitted nationals of the new accession states to move to the UK for work after 2004. It was at that point that the net inflows and outflows of EU migrants into the UK made a large shift well into a positive balance, after years of being at or near stasis.\n\n![Net Migration from the EU to the UK, 1991--2010 Source: Oxford Migration Observatory 2012. [www.migrationobservatory.ox.ac.uk](http://www.migrationobservatory.ox.ac.uk)](rjsf34_489_f1){#F1}\n\nThese elements lead to the first UK policy theme, free movement of workers/persons and European citizenship. 21 of the 69 cases from the UK were in the area of free movement of workers, European citizenship or both. Which migrants are able to use European rights and to what benefits they are entitled have been very active questions. These have been really significant references from the UK, which have shaped the EU citizenship debate. These references contributed to the development of the jurisprudence gradually extending the free movement of persons. The right to free movement has been interpreted to include derived rights for third country nationals (TCNs) based upon their relationship to EU nationals. In the process, tiers of rights have developed according to citizenship status, with the broadest rights for nationals of the pre-2004 EU Member States, followed by nationals of the A8[^5^](#FN5){ref-type=\"fn\"}/A2[^6^](#FN6){ref-type=\"fn\"} Member States that joined the EU in the 2004 and 2007 accessions, and then third TCNs.\n\nThis tiering is significant for determining welfare claims. Whether and to what extent migrants are able to access a national welfare state has large financial implications. Paul Eden, who co-represented the plaintiff in the landmark *Collins* case concerning access to Job Seekers Allowance (JSA) for unemployed EU migrants said: \"I was as aware as everyone of the potential financial dimensions to this case. The courts did not want to put the financial burden on the state. If everyone coming to the country for the first time had the right to seek JSA that has obviously a huge potential cost. There were Home Office lawyers sitting there in the phalanx of government lawyers, monitoring the implications for asylum and immigration (Interview, 15/09/11).\"\n\nMr *Collins* had been refused the JSA on grounds that he was not habitually resident in the UK. In the *Collins* case, the Court found that the UK\\'s requirement of a genuine link with the labour market constituted indirect discrimination, because UK nationals will be able to fulfil the requirements more easily. However, it was justified because the JSA had been designed to address unemployment for those living long term in the UK.\n\nThe Child Poverty Action Group (CPAG) is an advocacy organisation using litigation as part of a broader strategy to end child poverty. In addition to EU migrants to the UK, their strategy has also benefitted UK migrants to the rest of the EU. CPAG has been involved in cases about UK nationals' right to export their benefits from the UK when they move within the EU. Sarah Clarke, a solicitor at CPAG, said: \"That\\'s become quite a big issue since the European Court of Justice made a decision that one of the UK\\'s major disability benefits and the carer\\'s allowanc... are sickness benefits, not special non-contributory benefits. There are different categories of benefits that are basically treated differently in European law and special non-contributory benefits, if you move within the EU you can\\'t take them with you, but sickness benefits you can take with you (Interview, 02/09/11).\"\n\nThe EU rights to free movement were grounded in the free movement of workers. This strong nexus between the right to residency and economic self-sufficiency limits who is able to access these European rights. As Clarke noted that: \"I think it\\'s striking that the people who tend to lose out under European law, the people who don\\'t get their rights recognised, tend to be women, and often women with children, because they fall out of the movement of workers, which is all based upon worker status. (Interview, 02/09/11).\"\n\n### Right to Reside {#s7}\n\nThe first sub-theme involves the Right to Reside Test. It was created in May 2004, in response to concerns about the EU enlargement ([@R21]). It applies to applicants for most benefits. Under the Right to Reside Test, other EU nationals have the right to reside so long as they remain 'qualified' persons or are the family member of a qualified person. According to UK immigration regulations[^7^](#FN7){ref-type=\"fn\"}, a 'qualified person' is an European Economic Area (EEA) national who is in the UK and is: a jobseeker, a worker, a self-employed person, a self-sufficient person, or a student. Who qualifies for these sub-categories has been the subject of much litigation.\n\nThe European Commission commenced formal infringement proceedings against the UK in 2010 over the application of the Right to Reside Test[^8^](#FN8){ref-type=\"fn\"} On 29^th^ September 2011, the European Commission announced that it had sent the UK a reasoned opinion that the Right to Reside Test is in contravention of EU law. By requiring non-UK citizens who are EEA nationals to pass the Right to Reside Test, the UK indirectly discriminates against those EU citizens.[^9^](#FN9){ref-type=\"fn\"} Chris Grayling, the Minister for Employment, has said that the UK has formally rejected the Commission\\'s opinion, 'in the strongest possible manner.[^10^](#FN10){ref-type=\"fn\"}\n\nThe Supreme Court previously had considered the test in March 2011 in the *Patmalniece* case[^11^](#FN11){ref-type=\"fn\"}, rejecting the argument that the test directly discriminated against EU nationals, but agreeing that it constituted indirect discrimination. However, the Supreme Court found that the indirect discrimination was justified as a proportionate means of achieving a legitimate aim, namely protecting the public purse\n\nCPAG run a welfare rights advice hotline. Two-thirds of all calls to the hotline are for assistance with the Right to Reside Test. It is a convoluted area and many are ill informed. One welfare rights adviser said that: \"Most clients only know as much as the DWP/HMRC/Local Authority tells them. This information is usually dispensed by frontline staff who only have a cursory understanding of EEA nationals' rights... In particular, we find that clients are often advised to claim income support... and are refused because, by doing so, they become economically inactive... and lose their 'right to reside\" (Interview, 07/09/11).\"\n\nA further complication is that the Member States may discriminate against newer Member State nationals for up to seven years after the date of accession. Although the UK allowed nationals from the newer Member States to relocate to the UK for work, Workers Registration Scheme (WRS) was instituted for A8 and A2 nationals, requiring them to be in registered paid employment for 12 months before they were eligible for benefits. The WRS ended in May 2011 for A8 nationals, but was extended until December 2013 for A2 nationals.\n\n### Rights of family members of UK or EU national {#s8}\n\nThe next sub-theme is the derived rights of family members of UK or EU nationals. Some of the most important European citizenship cases have recently arisen from the UK in this sub-theme, including the *Baumbast*[^12^](#FN12){ref-type=\"fn\"} case. In Baumbast, the Court found that German immigrant to the UK was entitled to remain because his children were in education, even though he effectively had ceased to be economically active by becoming employed outside the EU. More recently, in the *Teixeira* and *Ibrahim*[^13^](#FN13){ref-type=\"fn\"} cases, the Court found that subsequently TCNs may claim a right of residence as the primary carer of an EU child in education in the Member State even though they are not economically active. CPAG has been granted a preliminary reference in the *Punakova*[^14^](#FN14){ref-type=\"fn\"} case, about whether previously self-employed A8 nationals have *Baumbast*-type rights of residence. This is illustrative of the process by which the freedom of movement has been extended via derived rights.\n\nThere is confusion about who qualifies as a family member. A welfare benefits adviser described a case where the client, a Dutch national living with her daughter, was refused Pension Credit despite having a right to reside as an extended family member of her daughter who was a worker. (Interview, 07/09/11).\n\nThe process by which the freedom of movement for persons has been extended through derived rights has created uncertainty as to the limit of the right. As different claimants present, their specific factual situations have had to be addressed via litigation. This has resulted in increased preliminary references in the area of free movement of persons.\n\n### Anti-discrimination {#s9}\n\nThe second major policy structure theme is anti-discrimination cases. This reflects the fact that there was a misfit between the UK\\'s policies in this area and Europe in some ways. Most of the cases address matters at the edges of rights and protections. Although later than other Member States, notably France, anti-discrimination had been extensively legislated in the UK. This earlier UK anti-discrimination was consolidated into the Equality Act 2010, which closely follows the EU Equal Treatment Directives[^15^](#FN15){ref-type=\"fn\"}.\n\nThere were 19 discrimination cases in the UK sample. Seventeen of those concerned equal treatment of men and women. The majority of these cases were brought on behalf of women. Two were brought on behalf of both men and women, and two were on behalf o transsexuals.[^16^](#FN16){ref-type=\"fn\"} Despite the protections afforded in UK law, examples of discriminatory treatment are still found. Christine Boch, who was involved with the *Brown*[^17^](#FN17){ref-type=\"fn\"} case on pregnancy discrimination, said that at the Citizens' Advice Bureau, 'We do still have loads of people who are the victims of the most blatant form of, for example, pregnancy discrimination.' (Interview, 23/09/11). In her experience, it happens primarily but not exclusively in two sectors -- catering and cleaning.\n\nThere was also an age discrimination test case from the UK, the *Age Concern* case, challenging the UK\\'s default retirement age. The Court found that a default retirement age could be justified if it was a proportionate means of achieving a legitimate aim. However, the aims of the strategic litigation went beyond the immediate result. Andrew Harrop, formerly of Age Concern, said that, 'We wanted to get a tight interpretation of 'objective justification,' in order to show that the prohibition of age discrimination could only be disregarded in special circumstances' (Interview, 07/09/11). Moreover, although the initial result was a setback, the legal challenge worked in the long run: \"It was an interesting example of the usefulness of litigation in a political lobbying context. The ECJ opinion was somewhat helpful in that it left some room for favourable interpretation, but it guided the High Court to dismiss the case. However, over the course of the case the issue had garnered huge interest and entered the public debate. It turned around stakeholder views... We lost the case in the High Court in the autumn of 2009, but going into the 2010 General Election we had all three parties committed to ending the practice, against what remained pretty robust lobbying from the CBI (Interview, 07/09/11).\"\n\nFinally, in the *Coleman*[^18^](#FN18){ref-type=\"fn\"} case the Court agreed that discrimination against an employee because she is the carer for a disabled person is in violation of the Equal Treatment Framework Directive 2000/78/EC ('Framework Directive'). Although associative discrimination was not covered on the face of the Disability Discrimination Act 1995 (DDA), the claimant argued that the DDA should be construed in light of the Framework Directive. Beyond the immediate result, the *Coleman* case was very significant for UK anti-discrimination law. Prior to the case, associative discrimination had not been included in the draft legislation for the Equality Act 2010. After the decision, the government announced that it would 'extend the prohibition against associative and perceptive direct discrimination and harassment to other strands and areas where this does not currently apply.'[^19^](#FN19){ref-type=\"fn\"}\n\nStructural Elements {#s10}\n-------------------\n\nThe next set of themes relates to the elements of the UK legal system that make it friendly or unfriendly to litigation to enforce rights. The relatively smooth functioning of the judicial system for administering EU law related to social policy was affirmed by this study. Some of these elements have benefitted references, some were barriers.\n\nStructural Benefits {#s11}\n-------------------\n\n### Acceptance of courts {#s12}\n\nParticipants reported that UK courts are generally accepting of EU rights and courts are open to preliminary references. Liz Barratt, a solicitor, said that, 'UK courts are very receptive to preliminary references. On immigration issues, the courts are very prepared to sort of knock it off to Europe' (Interview, 15/09/11). There are differences by courts. Sarah Clarke reported that, 'The Upper Tribunal \\[Social Security\\] has become quite good about referring... the Court of Appeal is probably a bit more hostile' (Interview, 02/09/11). Liz Barratt said: \"Once you get to the higher courts then the judges and barristers are more familiar with EU law. Lower courts may be nervous about sending a reference off, compared to the Supreme Court. I know people who have asked at the tribunal and been bounced up and only got their reference higher up (Interview, 15/09/11).\"\n\nVictoria Phillips, solicitor for the *Stringe*[^20^](#FN20){ref-type=\"fn\"} and British *Airways v. William*[^21^](#FN21){ref-type=\"fn\"} cases, said that in her experience it is easier to get a reference from a higher court. She did note few cases that had gone directly to the Court from the tribunal stage, 'but I think it\\'s a bold employment judge who\\'s prepared to go for a reference (Interview, 16/09/11).\n\nSome reported that there used to be more resistance to EU law in UK courts. In Advocate General Eleanor Sharpston\\'s experience as a barrister before 2006, 'when I was in practice, resisting a reference on behalf of the government was relatively easy for experienced counsel' (Interview, 08/09/11). She said that: \"It used to be almost a dirty word in English, you know, if you\\'ve got nothing else you run the Euro Defence... Obviously, if you were running an EU law point then it was because you had nothing better to say. In effect of course, that was a real caricature (Interview, 08/09/11).\"\n\n### Employment and Social Security Tribunals {#s13}\n\nThe second structural benefit in the UK legal system is the tribunal system of first instance courts that litigants can access with few barriers. The tribunals for employment and social security cases were cited as important to promote social policy cases. These systems handle an incredible volume of cases. There were 557,100 cases in the Employment tribunal and Social Security and child support Tribunals in 2011--12.[^22^](#FN22){ref-type=\"fn\"} In the *Preston*[^23^](#FN23){ref-type=\"fn\"} case, the Court noted that after its earlier judgments about the rights of part time employees to join pensions systems, 'some 60,000 part-time workers in the United Kingdom in both the public and the private sector commenced proceedings before industrial tribunals' (para. 17).[^24^](#FN24){ref-type=\"fn\"}\n\nAlthough plaintiffs do not need a lawyer to access the tribunals, the tribunals will also sometimes act to ensure that plaintiffs have better representation as the cases move up through the system on appeal. Sarah Clarke of CPAG said that, 'Sometimes the Upper Tribunal refers cases to us, if the claimant isn\\'t represented and it\\'s a complicated issue' (Interview, 02/09/11). Paul Eden, the co-counsel in *Collins*, which rose up through the social security tribunal system, discussed how he suddenly found himself out of his depth when the Commissioner decided to send a preliminary reference: \"Neither the government nor our side said 'wouldn\\'t it be a good idea to go to the ECJ.' We weren\\'t lawyers at that level... It was an entirely different matter to go off to the ECJ, rather than the Commissioner in London (Interview, 15/09/11).\"\n\nHowever, help came from the tribunal, which put them in touch with potential *pro bono* legal support (Interview, 15/09/11).\n\n### Legal Aid {#s14}\n\nA Judge at the Court stated that 'of course you need finance,' in order for cases to b brought (Interview, 08/09/11). The availability of legal aid was mentioned often as benefit to preliminary references.\n\nLegal aid only applies in some types of cases. There is no legal aid in the employment tribunals or for third party interventions. Legal aid is only available to persons, so organisations have to find funding for strategic litigation. There is some limited legal aid available at the Court for indigent litigants. Paul Eden said that, 'I should say that the Court was tremendously helpful.... They were very good about the legal aid process' (Interview, 15/09/11). He described support to understand the processes of the Court, as well as financial support.\n\nThreats to legal aid were frequently mentioned as a potential problem. Sarah Clarke said: \"There\\'s a network of advice workers and advice agencies in the UK, so there are Citizens' Advice Bureaux, Law Centres, local authorities might have welfare rights units, and there are independent advice agencies as well, all of whom might refer cases to us.... But we don\\'t know quite what will happen about that, because the government is proposing to take legal aid away for social welfare cases, which would devastate the advice sector. I would think it would be catastrophic. I think legal aid is quite an important part of how the entitlement system works in the UK (Interview, 02/09/11).\"\n\nThe proposal mentioned by Clarke at the time of this interview has materialised in the Legal Aid, Sentencing and Punishment of Offenders Act 2012 (LASPO), approved in April after 14 defeats in the House of Lords. The legislation makes deep cuts to the provision of legal aid services for social welfare from April 2013. It will eliminate advice provision to approximately 500,000 individuals annually by the Government\\'s estimates. By one legal aid group\\'s calculation, legal aid income to non-profit agencies will be reduced by 92% ([@R40]). These cuts will be implemented over two years. As legal aid was cited frequently as important for the advancement of preliminary references in the area of social welfare, these cuts have major implications for the ability to enforce EU social rights.\n\nStructural Barriers {#s15}\n-------------------\n\nMoney is the biggest identified structural barrier. There were two different issues. First, it is expensive to bring a test case. Victoria Phillips identified the availability of finance as crucial, listing the sources of funding for some key UK preliminary references: \"You\\'d have to have some big funding. It\\'s interesting to look at those cases that have been referred. My colleague... was involved in the *Preston*[^25^](#FN25){ref-type=\"fn\"} litigation. That was... all trade union-funded. We had another colleague who had the *UK BECTU*[^26^](#FN26){ref-type=\"fn\"} reference. Now that went straight from the High Court to the ECJ, and again that was trade union funded.... EHRC or the EOC have funded a few cases, but I can\\'t think of anything in the employment or industrial relations world that has gone without funding. *Coleman*[^27^](#FN27){ref-type=\"fn\"} was EOC-funded (Interview, 16/09/11).\"\n\nHowever, she pointed out that it would be considerably less expensive if the reference were made by the lower courts to the Court without having to appeal a case up to the higher courts, 'it\\'s not costly -- other than the costs of travel -- it\\'s not costly to make an application to the ECJ... There\\'s less documentation, there\\'s less fuss.' (Interview, 16/09/11). Nicola Smith, a solicitor in Scotland, said, 'Money is another big barrier. You\\'re always relying on someone to do it *pro bono*' (Interview, 16/09/11).\n\nThe second element of funding barriers is the possibility of a large award of costs against the losing party. In some cases lawyers have managed to agree to a protective costs order, essentially agreeing that the litigation is in the public interest and therefore the costs will not be passed along to the losing party. However, barring that, the prospect of bearing all of the costs for an expensive action is a deterrent. Andrew Harrop, formerly of Age Concern, said: \"The main challenge for us as a litigant was that the government refused to agree to a protective costs agreement. All along we were more afraid of an award of costs against us than we were of our legal fees, because costs would be so unpredictable. At one point we had two QCs against us. (Interview, 07/09/11).\"\n\nNicola Smith also identified potential costs as a barrier, particularly in light of the sorts of cases that result in preliminary references. 'There\\'s always a substantial risk of losing and getting stuck with all the costs. In these cases there\\'s always a genuine disagreement as to the area of law. It could go either way, usually' (Interview, 14/09/11). Roger Smith of JUSTICE has identified the fact that costs are not awarded in Employment Tribunals as a benefit to strategic litigation arising from the system (2003). However, as legal aid is not available in the employment tribunals, this may balance out this benefit\n\nActor {#s16}\n-----\n\nNGOs, the Equality and Human Rights Commission (EHRC), its predecessor commissions and unions are crucial actors to advance preliminary references. They help to overcome the barriers to social policy preliminary references.\n\n### Test case strategy {#s17}\n\nFor these actors, support for strategic litigation, or test cases, is part of an organised plan. CPAG\\'s strategy is well-developed and they openly solicit on their website for the fact patterns for test cases that they\\'re seeking. Sarah Clarke said that the test case strategy was one part of the organisation\\'s broader efforts to lobby for an end to child poverty in the UK (Interview, 02/09/11).\n\nSome legal NGOs run training sessions or advice lines for client advisers on European rights. Nicola Smith collaborated with other disability rights NGOs in other Member States to write Legal Strategy *A Good Practice Guide for Lawyers*: \"We had identified that there were very few cases on people with learning disabilities in the Member States involved and none at the European Court. We looked at the types of cases that it would be useful to have in the employment field. We also considered how NGOs could help with access to justice because it is difficult for this particular group to access courts, more so than for other groups. (Interview 14/09/11).\"\n\nPaul Eden said that both sides in test case litigation are looking for favourable cases. 'The government is looking for their test cases and CPAG is actively seeking their test cases. Both sides are looking for the best facts they can' (Interview, 15/09/11).\n\nOrganisations often have larger purposes behind test cases. In addition to the legal result desired, they seek publicity for a cause or to exert political pressure. One participant said of the *Age Concern* case: \"It was primarily there to change the world, and to change the situation with regard to retirement, but actually it had another very powerful benefit, to show Age Concern and Heyday as champions of older people, particularly of the baby boomers, who were coming up to an age where they could be booted out just because of their age' (Interview, 02/09/11).\"\n\nSeveral types of groups emerged as important. NGO support for litigation can come from either specialist legal NGOs that regularly take on strategic litigation or from general NGOs that provide support for an occasional strategic case. Within the legal NGOs category there are also community law centres, which occasionally take test case strategies. Unions also provide support for test cases. Victoria Phillips, who has had two union-funded preliminary references, said 'This is all entirely strategic litigation' (Interview, 16/09/11). Finally, the EHRC and its predecessor the EOC have provided support for cases.\n\nThe impact of these actors who have experience with the system is very important for preliminary references, especially considering the frequent inequity between the experience levels of the government and plaintiffs. Having the support of a pressure group can level the playing field. Paul Eden noted that, 'There\\'s the socio-legal theory about one shot versus repeat players Brian \\[*Collins*\\] was a classic one-shot player. CPAG and the Government were repeat players' (Interview, 15/09/11).[^28^](#FN28){ref-type=\"fn\"}\n\nResults for France {#s18}\n==================\n\nGenerally speaking, there was less data available in France. Although media and academic coverage of the Court grew in France over the period studied, it was still much lower than in the United Kingdom. There were fewer potential interviewees, and they tended to have had isolated experiences with the Court. The cases and expert interviews did not reveal equivalent legal NGOs with test cases strategies in the time period studied.[^29^](#FN29){ref-type=\"fn\"}\n\nPolicy Structure themes {#s19}\n-----------------------\n\n### Dualization of the welfare state {#s20}\n\nThe first policy theme for France is that France\\'s strongly dualized welfare state does not generate circumstances conducive to social policy preliminary references. France has 'dual social protection system.' Its primary component, the *S\u00e9curit\u00e9 sociale*, is a social insurance system financed through employment-related contributions and providing benefits linked to contributions. A secondary component, the *Solidarit\u00e9 nationale*, is a non-contributory scheme 'generally designed to cater for those who have been unable to build up an adequate contribution record ([@R5], p. 123--124). France\\'s social insurance system had the largest proportion of financing from employer and employee contributions in Europe ([@R6]). The social insurance funds that manage these contributions are relatively independent from the state. Through the principle of 'management by interested parties,' unions, employers and mutualist representatives were granted joint management ([@R23], p. 54). Unions are involved in managing these funds and as such have 'a *de facto* veto power against welfare state reforms' ([@R6], p. 339.) 'The trade unions act as the representatives and defenders of the systems. They defend both the interests of the salaried population and their own interests ([@R32], p. 111). The unions have become integrated into institutions that have 'broader constituency' than just union members, and the unions increasingly must intermediate between the short-term interests of their membership and the broader general interests ([@R23], p. 61).\n\nThe French welfare system is generous. Hall notes that, 'at 53 percent of GDP, France\\'s public expenditure has reached Scandinavian levels, although the redistributive impact of its tax and transfer systems is more meagre' (2008, p. 8). The system is regressive, and most social benefits go to the richest half of society ([@R41]).\n\nThe contrast between the traditional Bismarckian system of social insurance, and the newer system of basic social protection is increasingly leading to welfare 'dualization' and 'two distinct worlds of welfare' ([@R33], p. 96). The national solidarity scheme consists of healthcare, family benefits, and policies against social exclusion ([@R33], p. 96). Eligibility is based upon citizenship, benefits are either universal or means-tested, and the funding is through taxation ([@R33]). About one-third of the French population does not participate in the 'normal' labour market and social insurance arrangements, and therefore is covered by this second world. ([@R33]).\n\nIn 2001 Palier noted the increasing distance between the social insurance schemes which were becoming less solidaristic, and the emerging state-financed health care, family policy and social minima policies, which were increasingly subject to means tests and activation principles geared toward getting recipients into employment. Smith, in considering why it has been so difficult to reform the French welfare state, answered that, 'a key reason is that so many comfortable people resist change' (2004, p. 2). This may perhaps also be revealing of why France has so few social policy preliminary references.\n\n### Reverse Discrimination {#s21}\n\nFrance already had laws providing for gender equality, at least at the level of grand principles, before the European legal order began. It was at the insistence of France that Article 119 (now Article 157 TFEU) on Equal Pay was inserted into the Treaty of Rome ([@R11]). However, the EU has been important in ensuring a more strict application of rights. In an interview, the President of the former French equality agency, the *Haute autorit\u00e9 de lutte contre les discriminations et pour l\\'\u00e9galit\u00e9* (HALDE) said that they had quickly noted that, 'There were not enough complaints from women on the grounds of gender and pregnancy.' (Interview with [@R13]). The second policy theme identified is in line with this observation, as reverse discrimination cases were important in the period studied. None of the five cases in the area of equal treatment of men and women in France was brought on behalf of female plaintiffs. Every case involved men challenging more protective treatment for women.\n\nThe two most important cases were the *Griesmar*[^30^](#FN30){ref-type=\"fn\"} and *Mouflin*[^31^](#FN31){ref-type=\"fn\"} cases, with judgements released in close succession at the end of 2001. In Griesmar, a male magistrate challenged provisions in the scheme that granted service credits to female civil servants for each child but not to male civil servants. In *Mouflin*, Mr Mouflin wished to retire and claim his pension early in order to care for his wife, who had a terminal illness. However, the pensions scheme only allowed female civil servants to retire to care for an invalid spouse. In these cases the Court confirmed that pensions provided under the French civil service retirement scheme fell within the scope of Article 157 (ex Articles 141 and 119).\n\nAfter Griesmar there was more public awareness of the Court of Justice and EU rights in France (Interviews, 22/08/11 and 24/08/11). There was a proliferation of newspaper and other media coverage of *Griesmar*, in particular, and responses within France to the Court\\'s decision. They triggered a series of online commentaries, with participants in the discussions discussing potential strategies to enforce their rights or to challenge discriminatory policies.\n\nFrance continues to have gender discrimination issues for women. In May 2012 the *Conseil Constitutionnel* repealed the new law on sexual harassment, forcing the cancellation of all ongoing prosecutions under the prior law. There were demonstrations in the streets of Paris. It will likely be a delay of several months before new legislation can be adopted. Furthermore, speaking of gender discrimination, one Court official said: \"Particularly in France, it was very difficult to make French judges understand what indirect discrimination is. Because they have a very formal understanding of equality in France, and indirect discrimination is a much more substantive concept. So it really took years to make them understand and to get the concept into the writings (Interview, 08/09/11).\"\n\n### Opening up the French Civil Service/Protected Occupations {#s22}\n\nThe third policy theme is challenges to the closed French systems for the civil service and French protectionism for certain occupations. Access to the French civil service is tightly controlled and regimented through national examinations and the elite French universities.\n\nFive of the cases were related to efforts to open up the French civil service, or other protected occupations, to nationals of other Member States who have credentials from their home Member State. One such case was that of Isabel Burbaud[^32^](#FN32){ref-type=\"fn\"}, a Portuguese national, who wished to become a public hospital administrator in France based upon her qualifications in Portugal. The Court held that she could not be required to pass the public competitive examination for this position in the French civil service. In another case, Josep Pe\u00f1arroja Fa challenged a decision denying him enrolment in the court registries of translators. He believes that the restriction is related to nationality.[^33^](#FN33){ref-type=\"fn\"} He successfully brought a test action, although at the time of interview he was still awaiting the response of the French authorities to the Court\\'s decision in his favour. He noted that one likely response was for the French authorities to now say that the register is full (Interview, 25/09/11). The controlled French system for its civil service and professions conflicts with the opening logic of European integration.\n\nStructural Themes {#s23}\n-----------------\n\n### Resistance of Judges and Courts {#s24}\n\nResistance of French judges and courts to preliminary references specifically, or European rights and law generally was the first theme related to the French legal structure. Every lawyer reported resistance to preliminary references by French courts as a barrier. A Court official said that, 'I think what you need at the national level is of course judges who are willing to start a preliminary reference procedure. What is generally well known is that if they can avoid it, they often prefer to avoid it' (Interview, 08/09/11). Another respondent said that it was 'absolutely not' his experience that French courts were open to the idea of European law or European rights (Interview, 25/09/11). Jean-Edouard Robiou de Pont, the counsel in the *Wood*[^34^](#FN34){ref-type=\"fn\"} case, said that in his experience it was 'very rare' for a French court to agree to send a preliminary reference that had been requested of it\n\nRespondents also commented on a resistance to European law among French judges and courts. Mr. Robiou de Pont said, 'I think that French judges think that a lawyer only uses the European rights if he has nothing serious to say' (Interview, 23/08/11). Another participant said, 'The judges do not like European law' (Interview, 24/08/11). A variant of a common response was, 'They wish to keep cases in France' (Interview, 23/08/11). Philippe Derouin, a lawyer who has brought three preliminary references to the Court, including one in which he was the named party,[^35^](#FN35){ref-type=\"fn\"} noted that, 'I think that the Paris administrative court, at the lower level and the court of appeals, pride themselves on never having referred a case to the ECJ' (Interview, 24/08/11). One lawyer remembered that this was not a new trend, 'even in the old days the *Conseil d\\'\u00c9tat* refused to acknowledge that there were EU competencies for a long time' (Interview, 22/08/11).\n\nMr Derouin noted that there are two routes to the Court: to file a case and hope for referral, or to make a complaint to the Commission. If the Commission takes it seriously enough it will start an infringement procedure (Articles 258--260 TFEU). However: \"In France, in the areas that I know better, there are almost as many infringement procedures as there are referrals. Whereas in countries like the UK, Germany or the Netherlands, there are three times as many referrals as there are infringement procedures (Interview, 24/08/11).\"\n\nThis ratio suggests that there is much more resistance in the French system to preliminary references than in these other Member States.\n\nMr Derouin described a lengthy legal process for preliminary reference in which he was the named plaintiff, at the end of which, 'because this had been lasting for ten years, we could convince the *tribunal de sociale s\u00e9curit\u00e9* that there really was an issue' (Interview, 24/08/11). However, the tribunal did not make a referral directly to the Court. It instead referred the case to the *Cour de Cassation*, which also did not refer the case, although it noted that there was an EU question. The *Cour de Cassation* referred the case back to the tribunal, which finally sent the preliminary reference to the Court (Interview, 24/08/11).\n\n### Wilful misinterpretation by courts {#s25}\n\nThis resistance to EU law also manifests in the way that French courts interpret and apply EU law. Participants observed many instances of French courts interpreting EU directives in opposition to Court precedent. Mr Derouin observed that: \"One of the ways of the French court to resist is to disobey themselves, and the administrative courts tend to decide in favour of the administration. They tend to give a neutralising interpretation to European law or the European legislation or to the Treaties (Interview, 24/08/11).\"\n\nAnother interviewee said that: \"The French judges do what they want with the text. There are famous cases of the French courts, where the judges say that it is obvious what we must do based upon European law. But if you look at what they have said, it is nothing like the law. It is very different (Interview, 23/08/11).\"\n\n### Improvement over time {#s26}\n\nIn the second major structural theme, the interviewees indicated that the situation is improving. Several respondents said that the French legal system had become more open to European law and rights over the past decade. Mr Derouin attributed this change to the leadership of the *Conseil d\\'\u00c9tat*, stating that it 'showed \\[lower courts\\] the way' (Interview, 24/08/11). Seven of the 23 cases in this sample were referred by the *Conseil d\\'\u00c9tat* Another participant said that: \"Certainly the Griesmar case has triggered some changes in the litigation in France. And I feel like there is a lot more understanding of French law and EU law in France... I would say that in the last ten years it has changed a lot (Interview, 22/08/11).\"\n\nActor {#s27}\n-----\n\n### Lack of union or NGO support for litigation {#s28}\n\nIn France there were very few of the actors that were identified in the UK case study. France has the lowest union density, or percentage of the population that belongs to union, in Europe. Moreover, in the broader study union density was serving as a proxy for having interest groups willing to intervene to enable cases to get to the Court to enforce rights. Because French unions are so heavily interlinked with the management of French social insurance funds, unions are an unlikely source of support to challenge the system. Most of the French cases challenging social insurance provisions were brought against the social insurance funds, meaning that the interests of the unions were likely to be on the opposing side to the plaintiffs. There is very limited assistance available to individuals wishing to access the French legal system. This makes it difficult to overcome the substantial barriers in the French system\n\nThere was briefly an agency that could intervene on behalf of individuals in antidiscrimination matters. In May 2005 the French equality agency HALDE came into being. However, its staff was 'ridiculously small' and by aggregating all types of discrimination it 'dilutes each issue' ([@R19], p. 146). In 2006 its powers were expanded and it gained a statutory ability to be a party to cases. Louis Schweitzer, President of the HALDE, said that the number of complaints received by the agency had grown from 1,400 in 2005 to 10,549 in 2009 (Interview with Equinet Europe 18/11/11). The largest group of these complaints was in the area of discrimination based upon origin. However, 70% of the claims processed by HALDE in 2009 proved to be outside of its areas of competence. In the interview, Mr. Schweitzer discussed a problem of interest: \"The HALDE considered that the fact that children\\'s access to family benefits was attached to the regularity of their right of residence was discriminatory. The HALDE considered it was contrary to articles 8 and 14 of the European Convention on Human Rights as well as article 3 of the International Convention on Children\\'s Rights. Many courts echo the HALDE\\'s analysis and cancel the payment refusals. However, despite the repeated recommendations of the HALDE, at this very stage, the government will not amend the law on this issue despite it being clearly contrary to the international commitments of our country. This illustrates the limits of our powers (Interview with Equinet Europe 18/11/11).\"\n\nIt is striking that the HALDE was so powerless to effect change in one of the major issue areas that has generated most of the UK case law -- being legally resident for benefits purposes. On 1^st^ May 2011 the HALDE was dissolved and its functions were incorporated into another, more general, organisation, the *D\u00e9fenseur des droits* (DDD). The structure and remit of the organisation is still developing.\n\nWithout legal NGOs or unions or an equality agency to intervene as 'repeat players' ([@R16]) to help overcome the structural barriers on plaintiffs' behalf, individuals become crucial to the success of claims. However, without the benefit of being an elite or very determined lawyer, individual plaintiffs seem to have difficulties accessing the system.\n\n### Importance of individuals or elites {#s29}\n\nA key theme in the category of actors is the necessity of having individuals who will push for change through litigation. Without these protagonists it is unlikely that preliminary references will happen. This was highlighted by a Judge at the Court, who said, 'For certain important changes, you need people who pursue that aim with much litigation and force. If those people are not there you can forget about it, I think' (Interview, 08/09/11). One French lawyer said that most people would not consider taking a test case, as 'That requires really serious people' (Interview, 22/09/11). As an example of the determination required, Philippe Derouin had tried at least once or twice a year for ten years to obtain preliminary reference in a tax area before he was successful (Interview, 24/08/11). Jean-Edouard Robiou de Pont, the counsel in the *Wood*[^36^](#FN36){ref-type=\"fn\"} case, described how 'both courts in town refused to pose the question to the CJE[^37^](#FN37){ref-type=\"fn\"},' but he persisted until 'one very brave judge' sent the preliminary reference (Interview, 23/08/11).\n\nThe specialist knowledge required to be able to bring a preliminary reference to the Court can be a barrier: \"For a long time there were two different types of lawyers, there were the national lawyers and there were the European lawyers... It takes a certain type of lawyer to have the ambition to take a European case... to bring a test case, and also to have the two types of knowledge together -- European and national law -- it is quite rare, this type of person. So in France it is not unusual that there were not these types of cases, these types of people (Interview, 22/08/11).\"\n\nAnother interviewee said, 'I think a lot depends on the involvement of the lawyers and how far they are familiar with EU law' (Interview, 22/08/11). As an example of protagonist, several interviewees mentioned H\u00e9l\u00e8ne Masse-Dessin[^38^](#FN38){ref-type=\"fn\"}, the attorney in number of gender equity preliminary references as a driving force behind legal strategies. One interviewee described her as, 'certainly looking at legal strategies that might be useful to generate preliminary questions' (Interview, 15/09/11). Ms Masse-Dessin appeared for the party requesting the reference in three matters in this sample.\n\nThe key individual can be a lawyer, but it can also be the party bringing the suit, who is determined to push the action. In six cases the party requesting the preliminary reference represented him or herself in written or oral proceedings before the Court. When the stature of the Court is considered, it seems extraordinary that so many matters should have been argued by the parties themselves.\n\nA related sub-theme is the importance of elites. Many of the parties to the actions that generated the preliminary references were elites by virtue of professional status, who presumably have better access to resources or knowledge. Josep Pe\u00f1arroja Fa, who was the plaintiff in a matter that resulted in a preliminary reference, is the President of the Association of Sworn Translators. He is a lawyer and a translator from French to Spanish who passed a competitive examination in his home Member State, Spain. Having heard o difficulties for his colleagues with 'free circulation in France,' he applied to be admitted to the register of court translators in France for both the *Cour d\\'Appel de Paris* and the *Cour de Cassation*. Both applications were rejected. Subsequently he brought a case to challenge policies of the French courts as a restriction on the free movement of services. As there were insufficient funds for a lawyer, he brought the cases himself and represented himself at the Court. He said of the cases, 'legally speaking it was me against France, but really it was all the Spanish sworn translators represented by my association against the French authorities' (Interview, 24/09/11). Philippe Derouin is a partner at major multinational law firm in Paris. He was the party to a case about the proper social security contributions when an individual has income in more than one Member State of the EU. The *Griesmar* case had a French magistrate as the plaintiff. Another case, *Olympique Lyonnais SASP v Olivier Bernard and Newcastle UFC*[^39^](#FN39){ref-type=\"fn\"}, was brought on behalf of professional football player, with support of his Premier League club, over the breach of training contract with his previous club. Clearly there were resources available to him that would not be available to the average citizen with an employment dispute.\n\nDiscussion and Conclusion {#s30}\n=========================\n\nThis article has explored factors affecting national rates of preliminary references in the UK and France. The results clarify how policy structures can influence the rates of preliminary references and the areas in which they arise. The larger study indicated that Bismarckian welfare state contributed to a higher rate of social policy preliminary references, making the UK anomalous. The themes from the policy structure category in the UK case study offer some insight. Beveridgean welfare states can offer other points where European cases arise. As eligibility for benefits in Beveridgean systems is primarily based upon residence establishing a nexus to the state, who is 'resident' for the purposes of benefits becomes a crucial and, in the UK, highly litigated issue. Also, the fact that the UK became a destination state for migration for many nationals of new accession Member States after 2004 contributed to the rise in these cases. Because the UK\\'s equal protection legislation lagged behind that of Europe in some areas, there also were potential points to challenge the policy structure in these areas. This policy misfit created opportunities for challenges via preliminary references.\n\nFrance has a Bismarckian welfare state that is very generous to insiders, who are seldom motivated to challenge the system. The high degree of dualization in the system creates a strong insider/outsider dynamic. Without help, outsiders will have difficulty accessing legal systems to enforce their rights. In France there was little help of this nature to outsiders.\n\nEven in the newer areas of French social policy, such as social protection, antidiscrimination and disability, there are few cases. France does have cases in the area of reverse discrimination and opening up the civil service and protected professions. France had policies about equal pay before the Treaty of Rome. The areas in which French law was out of sync with EU law in the period studied concerned the disadvantage that French policies benefitting women caused to men. Anti-discrimination cases were a major category of cases from France, and all of the cases during the period studied were brought by or on behalf of men\n\nThe French system is characterised by being very closed for its civil service and regulated professions. These restrictions put France in conflict with EU law on free movement of persons and the right to establishment. Therefore, cases related to opening up the civil service and the regulated professions are an important policy theme.\n\nIn the second category, there are themes related to structural factors in the Member State legal system. These factors can present barriers to social policy preliminary references or they can benefit them. In the UK, although some structural themes are barriers, there are beneficial structural factors that helped to offset those barriers. The acceptance of courts is a broad theme. Courts have become accustomed to greater oversight over the policy process. The cost of litigation, both the fees to bring legislation and the threat of costs shifting onto the losing party are barriers. Finally, the employment and social security tribunal systems and legal aid are benefits to preliminary references, and help to overcome some of the barriers. However, the LASPO may radically shift the balance between structural barriers and benefits, as it eliminates legal aid for most social welfare cases, with cuts beginning in April 2013.\n\nIn contrast, in France, although there was reported improvement over time, the resistance of courts and judges to European law and European rights is a barrier to references. Wilful misinterpretation of EU law by French courts also creates barriers to references.\n\nThe third category of themes relate to the actions of individuals and groups. These actors whether they are plaintiffs, supporting groups or judges in the national system, affect rates of preliminary references. They can act to overcome structural barriers in the system.\n\nIn the UK organised test case strategies are backed by different types of organisations. These organised actions can be more strategic and advance the case law by soliciting particular cases needed to clarify points of law. This arises from a culture of activism in the NGO sector. The efforts are backed by NGOs, unions, the EHRC and its predecessor commissions. These actors benefit from legal aid and provide advice through benefits advisers and assist plaintiffs to access the tribunal systems and courts.\n\nIn France, the major theme in the category of actors is the importance of individuals. few determined lawyers have accounted for a number of the French preliminary references. Individuals are also important as plaintiffs in the actions. These individual plaintiffs are often elites themselves, who presumably have better access to information and resources to overcome the substantial barriers in the system.\n\nIn both France and the United Kingdom, the themes related to policy structures affect the types of cases that arise. These are the opportunities in the systems for strategic actions. In both Member States, cases arose that were the most significant to that national policy context. The interlinked nature of the EU legal system will then ensure that these established precedents affect every EU Member State. This has implications for the development of EU social welfare law. The Court\\'s justification for decisions based upon the circumstances of a particular case, or Member State, is then applied in a different policy context, with sometimes haphazard results.[^40^](#FN40){ref-type=\"fn\"}\n\nI am grateful for the research participants, who gave so generously of their time. I\\'d also like to thank colleagues and reviewers for their helpful suggestions to improve this article.\n\nThe Court was previously known as the European Court of Justice (ECJ), and some interview respondents refer to it in this way.\n\nSocial policy preliminary references were defined as those in the areas of: free movement of persons; European citizenship; social provisions; social security for migrant workers, and free movement of services, but only those services that are social policy-related, including health and education.\n\nBismarckian social security systems are primarily financed through social insurance contributions by employers and employees. They are focused on horizontal redistribution across the recipient\\'s lifespan, rather than vertical redistribution across society. They are earnings-related and the right to receive benefits is tied to work or its equivalent.\n\nSome interviewees fell into more than one category. For example, a Judge at the Court might also be a lawyer in France or the United Kingdom. In those instances, the individual was interviewed about all of his or her relevant experiences.\n\nCzech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and Slovenia.\n\nBulgaria and Romania.\n\nThe Immigration (European Economic Area) Regulations 2006, SI 2006/1003, reg 6(1)(a)--(e).\n\nEuropean Parliament Committee on Petitions, *Notice to Members: Subject: Petition 1119/2009 by Piotr Kalisz (Polish) on the British Authorities' refusal of his application for unemployment benefit ('Jobseeker\\'s Allowance')*, CM\\\\829426EN.doc, PE448.691, 2 September 2010.\n\nEuropean Commission press release IP/11/1118, 29 September 2011.\n\nHansard HC Deb 28 November 2011, vol 536 col 667.\n\n*Patmalniece (FC) (Appellant) v Secretary of State for Work and Pensions (Respondent)* \\[2011\\] UKSC 11\n\nCase C-413/99, *Baumbast and R. v Secretary of State for the Home Department*, \\[2002\\] ECR I-7091.\n\nCase C-480/08, *Teixeira v Lambeth LBC and Secretary of State* \\[2010\\] ECR I-01107 and Case C-310/08, *Harrow LBC v Ibrahim and Secretary of State*, \\[2010\\] ECR I-01065.\n\nCase C-148/11, *Secretary of State for Work and Pensions v Margita Punakova*.\n\nCouncil Directive 2000/43/EC implementing the principle of equal treatment between persons irrespective of racial or ethnic origin \\[2000\\] OJ L180/22; Council Directive 2000/78/EC establishing a general framework for equal treatment in employment and occupation \\[2000\\] OJ L303/16; European Parliament and Council Directive 2006/54/EC on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation \\[2006\\] OJ L204/23.\n\nThere have been prior anti-discrimination cases brought on behalf of men, notably Case C-262/88, *Barber v Guardian Royal Exchange Assurance Group* \\[1990\\] ECR I-1889, on pensionable ages for men and women.\n\nCase C-394/96, *Mary Brown v Rentokil Ltd* \\[1998\\] ECR I-4185.\n\nCase C-303/06, *Coleman v Attridge Law and Steve Law* \\[2008\\] ECR I-5603.\n\nHansard HC Deb 02 April 2009, vol 490, col 88WS.\n\nCase C-520/06, *Stringer and others v HMRC* \\[2009\\] ECR I-179.\n\nCase C-155/10, *British Airways v Williams* \\[2011\\].\n\nMinistry of Justice, HM Courts & Tribunals Service, *Annual Tribunals Statistics, 2011--12.*\n\nCase 78/98, *Shirley Preston and Others v Wolverhampton Healthcare NHS Trust and Others* \\[2000\\] ECR I-3201.\n\nThe industrial tribunals were the predecessors to the employment tribunals.\n\nCase 78/98, *Shirley Preston and Others v Wolverhampton Healthcare NHS Trust and Others* \\[2000\\] ECR I-3201.\n\nCase 173/99, *The Queen v Secretary of State for Trade and Industry, ex parte Broadcasting, Entertainment, Cinematographic and Theatre Union* (BECTU) \\[2001\\] ECR I-4881.\n\nCase C-303/06, *Coleman v Attridge Law and Steve Law* \\[2008\\] ECR I-5603.\n\n*See* [@R16].\n\nAlthough it was not a preliminary reference, the French NGO *Group d\\'information et de soutien des immigr\u00e9s (GISTI)* was involved in triggering the European Parliament to challenge several provisions of the Directive on the Right to Family Reunification (2003/86/EC) in the Court in Case C-540/03, *European Parliament v Council of the European Union* \\[2006\\] ECR I-05769.\n\nCase C-366/99, *Joseph Griesmar v Ministre de l\\'Economie, des Finances et de l\\'Industrie et Ministre de la Fonction publique, de la R\u00e9forme de l\\'Etat et de la D\u00e9centralisation* \\[2001\\] ECR I-9383.\n\nCase 206/00, *Henri Mouflin v Recteur de l\\'acad\u00e9mie de Reims* \\[2001\\] ECR I-10201.\n\nCase C-285/01, *Burbaud v Minist\u00e8re de l\\'Emploi et de la Solidarit\u00e9* \\[2003\\] ECR I-821.\n\nJoined Cases C-372/09 and C-373/09, *Josep Pe\u00f1arroja Fa* \\[2011\\] ECR I0000.\n\nCase C-164/07, *James Wood v Fonds de garantie des victimes des actes de terrorisme et d\\'autres infractions* \\[2008\\] ECR I-4143. Mr Wood, a UK national, had resided in France for twenty years with his French partner. One of their children was killed while on holiday outside the EU. French law provides for a payment to French family members in such a situation. Although other family members received payment, it was denied to Mr Wood because he was not French. The Court held that this constituted discrimination based upon nationality.\n\nCase C-133/06, *Philippe Derouin v Urssaf de Paris -- R\u00e9gion parisienne* \\[2008\\] ECR I-1853.\n\nCase C-164/07, *James Wood v Fonds de garantie des victimes des actes de terrorisme et d\\'autres infractions* \\[2008\\] ECR I-4143.\n\nThe French acronym for the prior name of the Court.\n\nUnfortunately, Ms Masse-Dessin\\'s caseload did not permit her to be interviewed at the time that she was approached.\n\nCase C-325/08, *Olympique Lyonnais SASP v Olivier Bernard and Newcastle UFC* \\[2010\\] ECR I-2177.\n\nSee [@R31] on the 'monocular' approach whereby the Court precedents around migrant workers have developed.\n"}
+{"text": "1. Introduction {#sec1-molecules-23-01659}\n===============\n\nUkgansan (UGS), which is also called yokukansan in Japan and Yi-gan san in China, is a traditional Oriental herbal formula composing seven medicinal herbal plants including *Uncaria sinensis*, *Atractylodes japonica*, *Poria cocos*, *Bupleurum falcatum*, *Angelica gigas*, *Cnidium officinale*, and *Glycyrrhiza uralensis*. UGS has been utilized to manage various diseases such as neurosis, insomnia, and irritability in children and is a herbal medicine that is approved by the Ministry of Health, Labor and Welfare of Japan. Modern pharmacological and clinical studies have suggested that UGS has the potential to improve insomnia \\[[@B1-molecules-23-01659],[@B2-molecules-23-01659]\\], borderline personality disorder \\[[@B3-molecules-23-01659]\\], neuroleptic-induced tardive dyskinesia \\[[@B4-molecules-23-01659]\\], drug-induced parkinsonism \\[[@B5-molecules-23-01659]\\], cognitive impairment \\[[@B6-molecules-23-01659]\\], and the behavioral and psychological symptoms of dementia (BPSD) \\[[@B7-molecules-23-01659],[@B8-molecules-23-01659],[@B9-molecules-23-01659]\\]. It is well known that the occurrence of these diseases is closely related to oxidative stress \\[[@B10-molecules-23-01659],[@B11-molecules-23-01659],[@B12-molecules-23-01659]\\]. Thus, antioxidant activity should be considered, in addition to the efficacy of therapeutic drugs.\n\nThe quality of herbal formulas depends on environmental factors, such as harvesting, storing, processing, and formulating methods. Because of their multiple components and the effect of the environment, quality assessment of the major components of herbal formulas is required for investigating their efficacy and safety \\[[@B13-molecules-23-01659]\\]. Although the identification and simultaneous determination of the components of UGS using high-performance liquid chromatography (HPLC) and HPLC-Q-TOF-MS have been reported \\[[@B14-molecules-23-01659],[@B15-molecules-23-01659]\\], there are few studies on simultaneous determination of the compounds in UGS using HPLC. Therefore, the current study conducted a simultaneous determination of the marker compounds of UGS for the quality control of UGS using a photodiode array HPLC detector (denoted the HPLC-PDA method), as it is the most popular analytical method. The HPLC-PDA method is a convenient and rapid analytical method to separate and identify the multiple compounds in herbal formulas \\[[@B16-molecules-23-01659],[@B17-molecules-23-01659]\\].\n\nIn this study, simultaneous analysis of seven compounds (liquiritin apioside, liquiritin, ferulic acid, glycyrrhizin, decursin, decursinol angelate, and atractylenolide I) in a UGS extract was performed, and method validation was carried out using the HPLC-PDA method. Moreover, the antioxidant activities of the seven marker compounds were determined using in vitro radical scavenging assays and the biological effects in neuronal cell lines HT22 hippocampal cells and BV-2 microglia. A variety of drugs have shown the neuroprotective activity against damage-induced HT22 cells and the inhibitory effects on neuroinflammation in lipopolysaccharide-stimulated BV-2 cells \\[[@B18-molecules-23-01659],[@B19-molecules-23-01659],[@B20-molecules-23-01659]\\].\n\n2. Results {#sec2-molecules-23-01659}\n==========\n\n2.1. Optimization of HPLC Separation {#sec2dot1-molecules-23-01659}\n------------------------------------\n\nHPLC analysis was performed for separating the seven marker compounds ([Figure 1](#molecules-23-01659-f001){ref-type=\"fig\"}) from a 70% ethanol extract of UGS ([Table 1](#molecules-23-01659-t001){ref-type=\"table\"}). To establish an efficient separation of the seven compounds, various mobile phases were evaluated, including water, acetonitrile, and methanol with trifluoroacetic acid (TFA), acetic acid, and phosphoric acid. The results showed good separation chromatograms using mobile phases consisting of 1.0% (v/v) aqueous acetic acid (A) and acetonitrile (B). The UV wavelength used for quantitative analysis was 250 nm for glycyrrhizin; 276 nm for liquiritin apioside, liquiritin, and atractylenolide I; and 325 nm for ferulic acid, decursin, and decursinol angelate. Using the established methods of HPLC, the seven marker compounds were resolved within 50 min. The retention times of liquiritin apioside, liquiritin, ferulic acid, glycyrrhizin, decursin, decursinol angelate, and atractylenolide I were 12.36, 12.86, 13.95, 33.77, 42.74, 43.25, and 49.99 min, respectively. HPLC chromatograms of the UGS extract in 70% ethanol and the standard compound mixture are shown in [Figure 2](#molecules-23-01659-f002){ref-type=\"fig\"}.\n\n2.2. Regression Equation, Linearity, and Limits of Detection (LOD) and Quantification (LOQ) {#sec2dot2-molecules-23-01659}\n-------------------------------------------------------------------------------------------\n\nThe linear relationships between the concentrations (*x*, \u03bcg/mL) and peak areas (*y*) of each compound were expressed by the regression equations (*y* = a*x* + b) ([Table 2](#molecules-23-01659-t002){ref-type=\"table\"} and [Supplementary Figure 1](#app1-molecules-23-01659){ref-type=\"app\"}). Calibration curves of the marker compounds revealed good linearity (*r*^2^ \u2265 0.9998). The LODs and LOQs for the tested compounds were 0.015--0.925 and 0.046--2.804 \u03bcg/mL, respectively.\n\n2.3. Determination of the Seven Marker Compounds of the UGS Extract {#sec2dot3-molecules-23-01659}\n-------------------------------------------------------------------\n\nThe established analytical method using HPLC was used to the simultaneous quantification of the seven marker compounds of UGS extract. The amounts of the seven marker compounds ranged from 0.190 mg/g to 16.431 mg/g. As shown in [Table 3](#molecules-23-01659-t003){ref-type=\"table\"}, decursin (16.431 mg/g) was the most abundant compound among these seven compounds.\n\n2.4. Precision, Accuracy, and Recovery {#sec2dot4-molecules-23-01659}\n--------------------------------------\n\nPrecision was represented as the relative standard deviations (RSDs) of the concentrations of marker compounds in mixed standard solutions, and repeated five times at low, middle and high concentration levels. The results for the intra- and inter-day precision and accuracy are shown in [Table 4](#molecules-23-01659-t004){ref-type=\"table\"} The intra- and inter-day precision for the seven marker compounds in mixed standard solutions was \\<3.36%, and the accuracy ranged from 95.12% to 105.12%. The RSD values for repeatability for the seven compounds ranged from 0.16% to 0.55% for retention times and from 0.4% to 0.76% for peak areas ([Table 5](#molecules-23-01659-t005){ref-type=\"table\"}). Recovery tests of the seven marker compounds were performed by adding the three different known amounts (low, middle, and high concentrations) of standard solutions to a certain amount of UGS extract. The recoveries of the seven marker compounds were between 95.99% and 104.94%, with RSD \u2264 3.21% ([Table 6](#molecules-23-01659-t006){ref-type=\"table\"}). These results demonstrate that the established method of HPLC had satisfactory precision, accuracy, repeatability, and recovery for simultaneous analysis.\n\n2.5. Antioxidant Activity of the Marker Compounds of UGS {#sec2dot5-molecules-23-01659}\n--------------------------------------------------------\n\nTo assess the antioxidant activity of UGS, the current study measured their scavenging activities against the 2,2\u2032-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and 2,2\u2032-diphenyl-1-picrylhydrazyl (DPPH) radicals. As shown in [Figure 3](#molecules-23-01659-f003){ref-type=\"fig\"}A,B, UGS extract dose-dependently increased the ABTS and DPPH radical scavenging activities. The effect on DPPH was not significant compared to ABTS. Then, antioxidant activity of the seven marker compounds of UGS was tested. These results revealed that ferulic acid dramatically increased the scavenging activity for ABTS in a dose-dependent manner. The concentration of ferulic acid required for a 50% reduction (RC~50~) in ABTS radicals was 16.22 \u03bcM ([Table 7](#molecules-23-01659-t007){ref-type=\"table\"} and [Figure 4](#molecules-23-01659-f004){ref-type=\"fig\"}A). The antioxidant activities obtained for ferulic acid using the DPPH assay are shown in [Table 8](#molecules-23-01659-t008){ref-type=\"table\"} and [Figure 4](#molecules-23-01659-f004){ref-type=\"fig\"}B. Similar to the results observed for ABTS, ferulic acid reduced the formation of the DPPH radical in a dose-dependent manner. The RC~50~ of ferulic acid against DPPH radicals was 41.21 \u03bcM. L-ascorbic acid was used as a positive antioxidant control compound.\n\n2.6. Biological Activitis of the UGS Extract in Neuronal Cell Lines {#sec2dot6-molecules-23-01659}\n-------------------------------------------------------------------\n\nThe current study examined the biological effects of the UGS extract on the neurodegenerative diseases. The cell viability test was carried out to measure the viability of HT22 hippocampal cells against UGS extract. Treatment with UGS extract did not affect the cell viability at \u2264100 \u03bcg/mL ([Figure 5](#molecules-23-01659-f005){ref-type=\"fig\"}A). Then, the neuroprotection activity of UGS was investigated. HT22 cells were stimulated with H~2~O~2~ in the absence or presence of the UGS extract. H~2~O~2~ treatment significantly decreased the viability of HT22 whereas UGS extract significantly reversed the H~2~O~2~-mediated cell death. The generation of lactose dehydrogenase (LDH) was further confirmed the neuroprotective effect of UGS. Consistent with the results of viability assay, H~2~O~2~ stimulation significantly enhanced the release of LDH compared with the untreated cells. In contrast, the UGS extract significantly blocked the release of LDH in H~2~O~2~-stimulated HT22 cells compared with the cells treated with H~2~O~2~ alone ([Figure 5](#molecules-23-01659-f005){ref-type=\"fig\"}A).\n\nAdditionally, the inhibitory effect of UGS on neuroinflammation was studied using microglia cell line. The cell viability assay was performed to assess cytotoxicity of UGS against BV-2 cells. As shown in [Figure 5](#molecules-23-01659-f005){ref-type=\"fig\"}B, no cytotoxicity of UGS extract was observed up to 100 \u03bcg/mL. Subsequent experiments were conducted at the range of nontoxic concentrations. To investigate the effects of UGS on the pro-inflammatory cytokine production, ELISAs for tumor necrosis factor-alpha (TNF-\u03b1) and interleukin-6 (IL-6) was performed using culture supernatant from the lipopolysaccharide (LPS)-stimulated BV-2 cells. Results showed that stimulation with LPS- significantly increased TNF-\u03b1 and IL-6 levels. In contrast, UGS treatment significantly reversed the LPS effect on production of TNF-\u03b1 and IL-6 ([Figure 5](#molecules-23-01659-f005){ref-type=\"fig\"}B).\n\n3. Discussion {#sec3-molecules-23-01659}\n=============\n\nIn these study, a simultaneous analysis of the seven marker compounds of UGS was performed using the HPLC-PDA method. The main ingredients of each medicinal herb forming UGS, are as follows: alkaloids (e.g., corynoxeine and hirsutine) from *Uncaria sinensis* \\[[@B21-molecules-23-01659]\\], sesquiterpenes (e.g., atractylon and atractylenolide I-III) from *Atractylodes japonica* \\[[@B22-molecules-23-01659]\\], triterpenes (e.g., pachymic acid, dehydrotumulosic acid, and dehydrotrametenolic acid) from *Poria cocos* \\[[@B23-molecules-23-01659]\\], triterpene saponins (e.g., saikosaponin A, C, and D) from *Bupleurum falcatum* \\[[@B24-molecules-23-01659]\\], coumarins (e.g., decursin, decursinol angelate, and nodakenin) from *Angelica gigas* \\[[@B25-molecules-23-01659]\\], alkylphthalides (e.g., cnidilide, ligustilide, butylphthalide, and neocnidilide) and phenol (e.g., ferulic acid) from *Cnidium officinale* \\[[@B26-molecules-23-01659],[@B27-molecules-23-01659]\\], and flavonoids (e.g., liquiritin, liquiritin apioside and liquiritigenin) and triterpene saponins (e.g., glycyrrhizin) from *Glycyrrhiza uralensis* \\[[@B28-molecules-23-01659]\\]. Among the various ingredients, the current study conducted simultaneous determination of the seven components liquiritin apioside, liquiritin, glycyrrhizin (*Glycyrrhiza uralensis*), ferulic acid (*Cnidium officinale*), decursin, decursinol angelate (*Angelica gigas*), and atractylenolide I (*Atractylodes japonica*) in the formulation of UGS by the established and validated analytical HPLC-PDA method. Consequently, decursin (16.431 mg/g), marker compound of *Angelica gigas*, was found as major compound in the UGS.\n\nThere are increasing evidence on the powerful antioxidant activity of herbal medicines \\[[@B29-molecules-23-01659],[@B30-molecules-23-01659]\\]. Herbal medicines contain various free radical scavenging molecules that mediate oxidative stress, which ultimately causes a variety of diseases such as cancer, inflammation, and metabolic disorders \\[[@B31-molecules-23-01659],[@B32-molecules-23-01659],[@B33-molecules-23-01659]\\]. Thus, antioxidant therapy is considered as an attractive approach to treat various human diseases \\[[@B34-molecules-23-01659],[@B35-molecules-23-01659]\\]. To date, previous studies reported significant antioxidant activities of various herbal formulas such Galkeun-tang \\[[@B36-molecules-23-01659]\\] and Samchulgeonbi-tang \\[[@B37-molecules-23-01659]\\] by regulating the free radical scavenging activity were reported. The current study data revealed that UGS extract activate the free radical scavenging effects against ABTS and DPPH. Among seven marker compounds, ferulic acid showed antioxidant activity compared with others. The scavenging activities of UGS extract were 100 \u00b1 0.1% and 87.8 \u00b1 0.5% for ABTS and DPPH, respectively, at 100 \u03bcg/mL. In contrast, the activities of other six compounds were below 43.6% and 4.1% for ABTS and DPPH, respectively, at 100 \u03bcg/mL. These results suggest that ferulic acid may be an active compound of UGS possessing the potent antioxidant activity.\n\nSimilarly, Liang et al. reported in vitro antioxidant effect of ferulic acid from *Spiranthes sinensis* \\[[@B38-molecules-23-01659]\\]. Additionally, to investigate the biological activity of UGS on the neurodegenerative diseases, neuronal cell lines were treated with various concentrations of UGS extract in the presence of neuronal damage inducers H~2~O~2~ and LPS, respectively. The results demonstrated that UGS has the effects on neuroprotection and anti-neuroinflammation in vitro. Further studies are required to verify the UGS effect using additional in vitro and in vivo neuronal diseases-related models.\n\nIn conclusion, the current study established a HPLC method for the quantitative analysis of the seven marker compounds present in extracts of UGS. Validation results of the method displayed good linearity, repeatability, intra- and inter-day precision, and recovery, indicating a successful application for the simultaneous analysis of marker compounds for the quality control of UGS. In addition, the results of antioxidant activity assays demonstrate the potent antioxidant activity of ferulic acid as an active compound of UGS.\n\n4. Materials and Methods {#sec4-molecules-23-01659}\n========================\n\n4.1. Plant Materials {#sec4dot1-molecules-23-01659}\n--------------------\n\nThe seven crude herbal medicines forming UGS, Uncariae Ramulus et Uncus, Atractylodis Rhizoma Alba, Poria Sclerotium, Bupleuri Radix, Angelicae Gigantis Radix, Cnidii Rhizoma, and Glycyrrhizae Radix et Rhizoma, were purchased at the Kwangmyungdang herbal market (Ulsan, South Korea). Voucher specimens (SCD-B-032) have been deposited at the Clinical Medicine Division, Korea Institute of Oriental Medicine.\n\n4.2. Chemicals and Reagents {#sec4dot2-molecules-23-01659}\n---------------------------\n\nThe marker compound glycyrrhizin was purchased from ChemFaces Biochemical Co., Ltd. (Wuhan, China); atractylenolide I, ferulic acid, and decursinol angelate were purchased from Biopurify Phytochemicals (Chengdu, China); and liquiritin apioside, liquiritin, and decursin were purchased from Sunny Biotech Co., Ltd. (Shanghai, China). The chemical structures of compounds are shown in [Figure 1](#molecules-23-01659-f001){ref-type=\"fig\"}. The purity of them was \u226598.0% according to HPLC analysis. HPLC-grade acetonitrile, methanol, and water were purchased from J. T. Baker Chemical Co. (Phillipsburg, NJ, USA), and analytical-grade acetic acid was purchased from Sigma-Aldrich (St. Louis, MO, USA).\n\n4.3. Apparatus and Chromatographic Conditions {#sec4dot3-molecules-23-01659}\n---------------------------------------------\n\nA Waters Alliance e2695 HPLC system (Waters Corp., Milford, MA, USA) equipped with a pump, degasser, column oven, auto sample injector, and photodiode array (PDA) detector (\\#2998, Waters Corp. Milford, MA, USA) was used in the quantitative analysis and Empower software (version 3; Waters Corp, Milford, MA, USA) was used to data processing. The chromatographic separation of the seven marker compounds was performed at 30 \u00b0C using a Gemini C-18 analytical column (250 \u00d7 4.6 mm, 5 \u03bcm; Phenomenex, Torrance, CA, USA) with a gradient solvent system of 1.0% (*v*/*v*) aqueous acetic acid (A) and acetonitrile (B). The elution conditions were as follows: 12--42% B for 0--25 min, 42--52% B for 25--30 min, 52--65% B for 30--55 min, 65--100% B for 55--56 min, and 100% B for 56--63 min. The flow rate and injection volume were 1.0 mL/min and 10 \u03bcL, respectively. The wavelength range of the PDA detector was 190 nm to 400 nm.\n\n4.4. Preparation of Standard Solutions {#sec4dot4-molecules-23-01659}\n--------------------------------------\n\nThe seven marker compounds were dissolved in methanol at a concentration of 1.0 mg/mL. Then, these stock solutions were diluted to make series of standard solutions with different concentrations for quantitative analysis.\n\n4.5. Preparation of the UGS 70% Ethanol Extract and Sample Solutions {#sec4dot5-molecules-23-01659}\n--------------------------------------------------------------------\n\nThe UGS composed of the seven crude herbal medicines, Uncariae Ramulus et Uncus, Atractylodis Rhizoma Alba, Poria Sclerotium, Bupleuri Radix, Angelicae Gigantis Radix, Cnidii Rhizoma, and Glycyrrhizae Radix et Rhizoma, was mixed as indicated in [Table 8](#molecules-23-01659-t008){ref-type=\"table\"} (41 g) and extracted using 70% aqueous ethanol (twice each with 246 mL) by refluxing for 2 h at 100 \u00b0C. The 70% ethanol extract was then filtered through a filter paper (5 \u03bcm) and concentrated using a rotary evaporator system (EYELA N-1000, Rikakikai Co., Tokyo, Japan) under vacuum to make powdered extract (8.613 g). The yield of UGS extract was 21%. The extract of UGS was weighed accurately and dissolved in methanol at 10 mg/ml for simultaneous determination. Then, the sample solution was filtered through a syringe filter (0.45 \u03bcm) and used for HPLC analysis. For testing biological activities, the extract of UGS was dissolved in dimethylsulfoxide (DMSO).\n\n4.6. Calibration Curve, LOD, and LOQ {#sec4dot6-molecules-23-01659}\n------------------------------------\n\nThe calibration curves of compounds were calculated from the peak areas of the standard solutions at different concentrations. The concentration ranges of marker compounds were as follows: liquiritin apioside (3.125--50 \u03bcg/mL), liquiritin (1.5625--25 \u03bcg/mL), ferulic acid (0.78125--25 \u03bcg/mL), glycyrrhizin and decursinol angelate (6.25--200 \u03bcg/mL), decursin (12.5--400 \u03bcg/mL), and atractylenolide I (0.78125--12.5 \u03bcg/mL). These solutions were measured in triplicate for the preparation of the calibration curves. The slope of the calibration curve and the standard deviation (SD) of the intercept were used to calculate the LOD and LOQ for the seven marker compounds, as follows:\n\nLOQ = 10 \u00d7 (SD of the response/slope of the calibration curve)\n\n4.7. Precision, Accuracy, and Recovery {#sec4dot7-molecules-23-01659}\n--------------------------------------\n\nTo evaluate the precision of the established HPLC conditions, intra- and inter-day variations were measured using the mixed standard solutions with low, middle, and high concentration levels of marker compounds. For the measurement of intra-day precision and accuracy, the mixed standard solutions were analyzed five times in a single day. The inter-day precision was assessed by repeating the analysis of the mixed standard solutions for three consecutive days. The intra- and inter-day RSD (%) was used to express the precision, and the percentage of the observed concentration for the fortified concentration were used to present the accuracy. To confirm the repeatability, six replicates were measured using the mixed standard solutions and the RSDs of retention times and peak areas for each compound were used. The recoveries of the seven compounds were determined by adding standard solutions at three different concentration levels (low, middle, and high) to the extract of UGS samples (100 mg for liquiritin apioside, liquiritin, ferulic acid, and atractylenolide I, and 10 mg for glycyrrhizin, decursin, and decursinol angelate); the 90% aqueous methanol was added to volume metric flask to make 10mL sample solution. The recovery was performed five times at each level and calculated as follows:$${Recovery}~\\left( \\% \\right) = \\frac{{found}~{concentration} - ~{original}~{concentration}}{{spiked}~{concentration}}~ \\times ~100$$\n\n4.8. ABTS-Scavenging Activity {#sec4dot8-molecules-23-01659}\n-----------------------------\n\nThe ABTS radical-scavenging activity of the UGS extract was assessed according to the previous study \\[[@B36-molecules-23-01659]\\]. The ABTS radical cation was prepared by reaction with a 7 mM ABTS solution and 2.45 mM potassium persulfate, followed by keeping in the dark at room temperature for 16 h. The absorbance of the reaction mixture was adjusted to 0.7 at 734 nm. A 100 \u03bcL of sample solution at various concentrations (12.5--200 \u00b5g/mL) was mixed with ABTS^\u2022+^ solution. The mixture was incubated in the dark at room temperature for 5 min and the absorbance at 734 nm was measured using a spectrophotometer (Benchmark Plus, Bio-Rad, CA, USA). The ABTS radical-scavenging capacity of the UGS extract was measured using the following equation:$${ABTS}~{scavenging}~{activity}~\\left( \\% \\right) = \\frac{1 - {absorbance}~{at}~734~{nm}~{of}~{UGS}~{extract}}{{absorbance}~{at}~734~{nm}~{of}~{control}}~ \\times ~100$$\n\n4.9. DPPH-Scavenging Activity {#sec4dot9-molecules-23-01659}\n-----------------------------\n\nThe DPPH radical-scavenging activity of the UGS extract was assessed according to the previous study \\[[@B37-molecules-23-01659]\\]. In brief, a 100 \u03bcL aliquot of sample solution at various concentrations was mixed with 100 \u03bcL of 0.15 mM DPPH solution in methanol. The mixture was incubated in the dark at room temperature for 30 min and the absorbance at 517 nm was measured on a spectrophotometer (Benchmark Plus, Bio-Rad, CA, USA). The DPPH radical-scavenging capacity of the UGS extract was measured using the following equation:$${DPPH}~{scavenging}~{activity}~\\left( \\% \\right) = \\frac{1 - {absorbance}~{at}~517~{nm}~{of}~{UGS}~{extract}}{{absorbance}~{at}~517~{nm}~{of}~{control}}~ \\times ~100$$\n\n4.10. Cell Lines and Culture {#sec4dot10-molecules-23-01659}\n----------------------------\n\nHT22 and BV-2 cells were maintained in Dulbecco's modified Eagle's medium (Hyclone/Thermo, Rockford, IL, USA) supplemented with 10% fetal bovine serum (Hyclone/Thermo, Rockford, IL, USA) and penicillin/streptomycin in 5% CO~2~ at 37 \u00b0C.\n\n4.11. Cell Counting Kit (CCK) Assay {#sec4dot11-molecules-23-01659}\n-----------------------------------\n\nCells were plated on 96-well microplates and treated with various concentrations of UGS extract in DMSO for 24 h. After adding CCK-8 solution (Dojindo, Kumamoto, Japan) to each well, and the cells were maintained for 4 h at 37 \u00b0C. The absorbance at 450 nm was measured on an Epoch microplate spectrophotometer (Bio-Tek Instruments, Inc., Winooski, VT, USA). The cell viability was determined using the following equation:$${Cell}~{viability}~\\left( \\% \\right) = \\frac{{Mean}~{OD}~{in}~{UGS}~{extract} - {treated}~{cells}}{{Mean}~{OD}~{in}~{untreated}~{cells}}~ \\times ~100$$\n\nTo examine the neuroprotective effect of UGS extract, HT22 cells were co-treated with UGS and H~2~O~2~ (Sigma-Aldrich, St. Louis, MO, USA) for 6 h.\n\n4.12. LDH Release Assay {#sec4dot12-molecules-23-01659}\n-----------------------\n\nThe release of LDH was measured using the CytoTox 96 nonradioactive cytotoxicity assay kit (Promega, Madison, WI, USA). Cell lysates and supernatants were prepared to induce maximal LDH release and experimental LDH release, respectively, and incubated with substrate mixture in the dark at room temperature for 30 min. Stop solution was added to each well and absorbance at 490 nm was determined on an Epoch microplate spectrophotometer (Bio-Tek Instruments, Inc., Winooski, VT, USA). The cytotoxicity of the UGS extract was calculated using the following formula:$${Cytotoxicity}~\\left( \\% \\right) = \\frac{{Experimental}~{LDH}~{release}~}{{Maximum}~{LDH}~{release}}~ \\times ~100$$\n\n4.13. Enzyme-Linked Immunosorbent Assays (ELISAs) for Cytokine Production {#sec4dot13-molecules-23-01659}\n-------------------------------------------------------------------------\n\nBV-2 cells were pretreated with UGS extract for 2 h and treated with LPS (1 \u03bcg/mL) for an additional 22 h. Culture supernatants were collected and the levels of TNF-\u03b1 and IL-6 were assessed using ELISA kits from R&D Systems (Minneapolis, MN, USA).\n\n4.14. Statistical Analysis {#sec4dot14-molecules-23-01659}\n--------------------------\n\nThe data are expressed as the mean \u00b1SEM. Data were analyzed using one-way analysis of variance and Dunnett's multiple comparisons test and student's t-test. *P* \\< 0.05 was considered significant.\n\nThis study was supported by a research grant (C17120 and K18293) from the Korea Institute of Oriental Medicine (KIOM).\n\n**Sample Availability:** Samples of the compounds liquiritin apioside, liquiritin, ferulic acid, glycyrrhizin, decursin, decursinol angelate, and atractylenolide I are commercially available.\n\nSupplementary Figure 1 is available online.\n\n###### \n\nClick here for additional data file.\n\nY.J.K., H.-S.L., and S.-J.J. participated in the design of the study data analyses and manuscript preparation. Y.J.K., J.Y. and C.-S.S. conducted JDS extract and simultaneous determination of seven constituents in JDS extract using HPLC-PDA method. H.-S.L., E.S., B.-Y.K., and S.-J.J. conducted the biological examination of in vitro. All authors read and approved the final manuscript.\n\nThis research received no external funding.\n\nThe authors declare no conflict of interest.\n\n![Chemical structures of the seven marker compounds of UGS.](molecules-23-01659-g001){#molecules-23-01659-f001}\n\n![HPLC chromatograms of the 70% ethanol extract of UGS (**A**) and a standard mixture (**B**) at 250 nm, 276 nm, and 325 nm. Liquiritin apioside (**1**), liquiritin (**2**), ferulic acid (**3**), glycyrrhizin (**4**), decursin (**5**), decursinol angelate (**6**), and atractylenolide I (**7**).](molecules-23-01659-g002){#molecules-23-01659-f002}\n\n![Effects of UGS on free radical-scavenging activities. The antioxidant activity of UGS against ABTS (A) or DPPH (B) was assessed using a radical-scavenging method. The quantitative data are presented as the mean \u00b1SEM of triplicate experiments. \\*\\* *P* \\< 0.01\" or \\*\\*\\* *P* \\< 0.001 vs vehicle control cells *n* = 3/sample. '0\u2032 in x-axis represents vehicle control. AA: L-ascorbic acid, a positive control.](molecules-23-01659-g003){#molecules-23-01659-f003}\n\n![Effects of ferulic acid on free radical-scavenging activities. The antioxidant activity of different concentrations of ferulic acid and L-ascorbic acid against ABTS (A) or DPPH (B), as assessed using a radical-scavenging method. ~L~-ascorbic acid was used as a positive control of antioxidant. The quantitative data are presented as the mean \u00b1 SEM of triplicate experiments. \\* *P* \\< 0.05, \\*\\* *P* \\< 0.01 or \\*\\*\\* *P* \\< 0.001 vs vehicle control cells n = 3/sample.](molecules-23-01659-g004){#molecules-23-01659-f004}\n\n![Biological effects of UGS extract on neuroprotection and anti-inflammation in HT22 neuronal hippocampal cells and BV-2 microglia. (A) Cell viability was performed to assess the cytotoxicity of HT22 cells against UGS extract using the cell counting Kit (CCK)-8 assay. Neuroprotective activity of UGS was tested using the CCK assay (middle) and LDH release assay (right). The results are expressed as the mean \u00b1SEM of three independent experiments. ^\\#\\#\\#^ *P* \\< 0.001 vs vehicle control cells; \\*\\*\\* *P* \\< 0.001 and \\*\\* *P* \\< 0.01 vs H~2~O~2~-treated cells. (B) Cell viability was performed to assess the cytotoxicity of BV-2 cells against UGS extract using the CCK-8 assay. The UGS effect on lipopolysaccharide (LPS)-induced production of proinflammatory cytokines were assessed in BV-2 cells using ELISA. Cells were pretreated with UGS for 2 h and then stimulated with LPS for an additional 22 h. The results are expressed as the mean \u00b1 SEM of three independent experiments. ^\\#\\#\\#^ *P* \\< 0.001 vs vehicle control cells; \\*\\*\\* *P* \\< 0.001, \\*\\* *P* \\< 0.01 and \\* *P* \\< 0.05 vs LPS-treated cells.](molecules-23-01659-g005){#molecules-23-01659-f005}\n\nmolecules-23-01659-t001_Table 1\n\n###### \n\nComposition of UGS.\n\n  Latin Name                      Scientific Name           Amount (g)   Origin\n  ------------------------------- ------------------------- ------------ ----------------\n  Uncariae Ramulus et Uncus       *Uncaria sinensis*        6            China\n  Atractylodis Rhizoma Alba       *Atractylodes japonica*   8            China\n  Poria Sclerotium                *Poria cocos*             8            China\n  Bupleuri Radix                  *Bupleurum falcatum*      4            China\n  Angelicae Gigantis Radix        *Angelica gigas*          6            Bonghwa, Korea\n  Cnidii Rhizoma                  *Cnidium officinale*      6            China\n  Glycyrrhizae Radix et Rhizoma   *Glycyrrhiza uralensis*   3            China\n  Total amount                                              41           \n\nmolecules-23-01659-t002_Table 2\n\n###### \n\nLinear range, regression equation, correlation coefficients, LODs, and LOQs for compounds.\n\n  -------------------------------------------------------------------------------------------------------\n  Compound.             Linear Range\\   Regression Equation\\    *r* ^2^   LOD ^b)^\\   LOQ ^c)^\\   \n                        (\u03bcg/mL)         (*y* = a*x* + b) ^a)^             (\u03bcg/mL)     (\u03bcg/mL)     \n  --------------------- --------------- ----------------------- --------- ----------- ----------- -------\n  Liquiritin apioside   3.125--50       15290                   4536.9    1.0000      0.177       0.537\n\n  Liquiritin            1.5625--25      18759                   2614.9    0.9999      0.052       0.157\n\n  Ferulic acid          0.78125--25     56995                   1865.6    1.0000      0.039       0.118\n\n  Glycyrrhizin          6.25--200       4882.5                  1533      1.0000      0.619       1.876\n\n  Decursin              12.5--400       30409                   77457     0.9998      0.925       2.804\n\n  Decursinol angelate   6.25--200       35125                   38848     0.9998      0.232       0.705\n\n  Atractylenolide I     0.78125--12.5   62615                   1322.9    1.0000      0.015       0.046\n  -------------------------------------------------------------------------------------------------------\n\n**^a^**^)^*y* = a*x* + b, y means peak area and *x* means concentration (\u03bcg /mL). **^b)^** LOD (Limit of detection): 3.3 \u00d7 (SD of the response/slope of the calibration curve). **^c)^** LOQ (Limit of quantitation): 10 \u00d7 (SD of the response/slope of the calibration curve).\n\nmolecules-23-01659-t003_Table 3\n\n###### \n\nThe content of marker compounds in UGS.\n\n  Compound              Content (mg/g)\n  --------------------- ----------------\n  Liquiritin apioside   1.671 \u00b1 0.004\n  Liquiritin            2.014 \u00b1 0.004\n  Ferulic acid          0.605 \u00b1 0.002\n  Glycyrrhizin          10.267 \u00b1 0.05\n  Decursin              16.431 \u00b1 0.04\n  Decursinol angelate   7.606 \u00b1 0.002\n  Atractylenolide I     0.190 \u00b1 0.001\n\nmolecules-23-01659-t004_Table 4\n\n###### \n\nPrecision and accuracy of seven marker compounds in UGS.\n\n  Compound              Fortified Conc. (\u03bcg/mL)   Intra-Day (n = 5)   Inter-Day (n = 5)                           \n  --------------------- ------------------------- ------------------- ------------------- -------- ------- ------ --------\n  Liquiritin apioside   5                         4.84                2.11                96.88    5.01    1.25   100.18\n                        10                        10.14               1.49                101.36   10.19   0.96   101.87\n                        20                        20.10               0.96                100.50   20.15   0.61   100.76\n  Liquiritin            5                         4.99                2.52                99.78    5.09    0.74   101.84\n                        10                        10.50               1.65                105.04   10.36   1.14   103.63\n                        20                        20.49               1.13                102.46   20.44   0.73   102.18\n  Ferulic acid          1.5                       1.46                3.36                97.63    1.47    2.73   97.81\n                        3                         3.15                1.36                105.12   2.97    2.98   98.93\n                        6                         6.28                0.84                104.63   5.85    2.90   97.42\n  Glycyrrhizin          12.5                      12.26               0.97                98.08    12.25   0.96   97.98\n                        25                        24.98               2.00                99.92    24.47   0.86   97.90\n                        50                        49.44               0.38                98.88    49.51   0.42   99.02\n  Decursin              20                        19.74               1.42                98.71    19.68   1.48   98.42\n                        40                        41.51               0.97                103.78   40.47   1.22   101.17\n                        80                        81.37               0.72                101.71   81.34   0.53   101.67\n  Decursinol angelate   10                        9.58                0.93                95.76    9.51    2.61   95.12\n                        20                        19.55               0.73                97.73    19.50   0.45   97.51\n                        40                        39.13               0.43                97.83    39.11   0.46   97.78\n  Atractylenolide I     1                         0.98                2.53                98.40    1.00    1.14   100.07\n                        2                         2.05                1.48                102.60   2.02    0.95   101.23\n                        4                         4.02                0.98                100.50   4.01    0.69   100.37\n\n**^a^**^)^ Precision is expressed as RSD (%) = (SD/Mean) \u00d7 100. ^b)^ Accuracy (%) = (Observed concentration/Fortified concentration) \u00d7 100.\n\nmolecules-23-01659-t005_Table 5\n\n###### \n\nRepeatability of retention times and peak areas for the seven analytes (n = 6).\n\n  Compound              Retention Time (min)   Peak Area (AU)                           \n  --------------------- ---------------------- ---------------- ----------------------- ------\n  Liquiritin apioside   12.48 \u00b1 0.05           0.43             234009.83 \u00b1 1419.74     0.61\n  Liquiritin            13.00 \u00b1 0.06           0.49             143858.00 \u00b1 903.60      0.63\n  Ferulic acid          14.09 \u00b1 0.08           0.55             212067.33 \u00b1 1604.31     0.76\n  Glycyrrhizin          33.46 \u00b1 0.18           0.52             144019.50 \u00b1 726.13      0.50\n  Decursin              42.93 \u00b1 0.07           0.16             1905525.83 \u00b1 12044.19   0.63\n  Decursinol angelate   43.45 \u00b1 0.07           0.16             1094712.83 \u00b1 4431.44    0.40\n  Atractylenolide I     50.27 \u00b1 0.12           0.24             234499.67 \u00b1 1056.72     0.45\n\nSD: Standard deviation; RSD: Relative standard deviation.\n\nmolecules-23-01659-t006_Table 6\n\n###### \n\nRecovery of seven marker compounds in UGS.\n\n  Compound              Original Conc. (\u03bcg/mL)   Spiked Conc. (\u03bcg/mL)   Found Conc. (\u03bcg/mL)   Recovery ^a)^ \u00b1 SD (%)   RSD (%)\n  --------------------- ------------------------ ---------------------- --------------------- ------------------------ ---------\n  Liquiritin apioside   17.75                    4                      21.59                 95.99 \u00b1 0.75             0.78\n                                                 10                     27.86                 101.14 \u00b1 1.34            1.33\n                                                 20                     37.56                 99.07 \u00b1 1.28             1.29\n  Liquiritin            21.27                    4                      25.46                 104.94 \u00b1 0.84            0.80\n                                                 10                     31.61                 103.49 \u00b1 1.72            1.66\n                                                 20                     41.42                 100.80 \u00b1 1.41            1.40\n  Ferulic acid          6.63                     1.5                    8.16                  101.49 \u00b1 1.55            1.53\n                                                 3                      9.73                  103.14 \u00b1 0.76            0.74\n                                                 6                      12.79                 102.70 \u00b1 0.61            0.60\n  Glycyrrhizin          9.21                     2.5                    11.63                 97.04 \u00b1 1.32             1.36\n                                                 5                      14.08                 97.45 \u00b1 1.21             1.24\n                                                 10                     18.90                 96.94 \u00b1 1.40             1.44\n  Decursin              14.81                    4                      18.84                 100.68 \u00b1 3.23            3.21\n                                                 8                      22.56                 96.77 \u00b1 1.05             1.08\n                                                 16                     30.48                 97.91 \u00b1 0.75             0.77\n  Decursinol angelate   6.90                     2                      8.85                  97.26 \u00b1 1.16             1.20\n                                                 4                      10.76                 96.53 \u00b1 0.47             0.48\n                                                 8                      15.22                 104.00 \u00b1 0.35            0.33\n  Atractylenolide I     2.08                     1                      3.11                  102.93 \u00b1 1.30            1.26\n                                                 2                      4.14                  102.68 \u00b1 1.49            1.46\n                                                 4                      5.97                  97.24 \u00b1 0.42             0.43\n\n**^a^**^)^ Recovery (%) = (Found concentration -- Original concentration)/spiked concentration \u00d7 100.\n\nmolecules-23-01659-t007_Table 7\n\n###### \n\nABTS/DPPH radical scavenging activity of marker compounds of UGS.\n\n  \u03bcM       1             2             3             4             5             6             7             nM        L-ascorbicAcid \\*\n  -------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- --------- -------------------\n  ABTS                                                                                                                 \n  0        0.0 \u00b1 0.9     0.0 \u00b1 0.9     0.0 \u00b1 0.9     0.0 \u00b1 0.9     0.0 \u00b1 0.9     0.0 \u00b1 0.9     0.0 \u00b1 0.9     0         0.0 \u00b1 0.9\n  1.5625   2.3 \u00b1 0.6     1.9 \u00b1 0.3     20.8 \u00b1 0.7    0.7 \u00b1 0.9     0.1 \u00b1 0.3     --0.2 \u00b1 0.3   --1.1 \u00b1 0.6   0.78125   --0.9 \u00b1 0.9\n  3.125    3.2 \u00b1 0.6     3.5 \u00b1 0.5     36.7 \u00b1 0.4    0.0 \u00b1 0.8     0.3 \u00b1 0.2     --1.1 \u00b1 0.2   2.2 \u00b1 1.3     1.5625    0.8 \u00b1 0.6\n  6.25     7.0 \u00b1 0.6     5.5 \u00b1 0.4     38.0 \u00b1 0.8    --0.1 \u00b1 0.7   0.7 \u00b1 0.6     0.0 \u00b1 0.2     --0.4 \u00b1 1.1   3.125     -0.2 \u00b1 0.5\n  12.5     10.1 \u00b1 0.3    10.5 \u00b1 0.1    64.4 \u00b1 0.2    --0.8 \u00b1 0.9   --1.5 \u00b1 0.7   --0.3 \u00b1 0.2   0.4 \u00b1 0.6     6.25      2.7 \u00b1 0.8\n  25       18.8 \u00b1 0.8    19.7 \u00b1 0.8    99.7 \u00b1 0.1    1.3 \u00b1 0.1     0.1 \u00b1 0.4     --0.4 \u00b1 0.6   1.0 \u00b1 1.1     12.5      7.4 \u00b1 0.5\n  50       28.3 \u00b1 0.4    29.0 \u00b1 0.6    99.8 \u00b1 0.0    2.7 \u00b1 1.0     0.5 \u00b1 0.9     1.5 \u00b1 0.2     2.0 \u00b1 0.5     25        22.4 \u00b1 0.6\n  100      43.6 \u00b1 0.0    43.1 \u00b1 0.7    100.0 \u00b1 0.1   3.6 \u00b1 0.2     0.0 \u00b1 0.1     2.6 \u00b1 0.3     --0.7 \u00b1 0.3   50        65.9 \u00b1 1.0\n  DPPH                                                                                                                 \n  0        0.0 \u00b1 2.5     0.0 \u00b1 2.5     0.0 \u00b1 2.5     0.0 \u00b1 2.5     0.0 \u00b1 2.5     0.0 \u00b1 2.5     0.0 \u00b1 2.5     0         0.0 \u00b1 2.5\n  1.5625   0.9 \u00b1 0.2     0.9 \u00b1 0.2     4.8 \u00b1 0.5     --0.1 \u00b1 0.7   0.9 \u00b1 0.7     --0.2 \u00b1 1.0   2.7 \u00b1 0.4     0.78125   5.7 \u00b1 1.1\n  3.125    --0.2 \u00b1 0.3   --0.2 \u00b1 0.3   8.6 \u00b1 0.5     --0.7 \u00b1 0.7   0.1 \u00b1 1.4     3.7 \u00b1 1.0     --0.5 \u00b1 0.5   1.5625    9.1 \u00b1 3.0\n  6.25     2.5 \u00b1 1.0     2.5 \u00b1 1.0     16.9 \u00b1 0.1    --1.7 \u00b1 1.2   0.3 \u00b1 1.3     2.3 \u00b1 0.5     0.4 \u00b1 0.8     3.125     12.4 \u00b1 1.0\n  12.5     2.8 \u00b1 0.5     2.8 \u00b1 0.5     34.6 \u00b1 2.0    --0.9 \u00b1 0.6   0.6 \u00b1 0.8     3.6 \u00b1 1.4     1.0 \u00b1 0.7     6.25      21.3 \u00b1 1.0\n  25       1.7 \u00b1 0.7     1.7 \u00b1 0.7     55.1 \u00b1 1.7    0.5 \u00b1 1.3     0.5 \u00b1 0.3     5.5 \u00b1 1.4     1.3 \u00b1 1.0     12.5      41.5 \u00b1 3.6\n  50       4.1 \u00b1 1.2     4.1 \u00b1 1.2     75.5 \u00b1 0.9    1.9 \u00b1 0.9     0.4 \u00b1 0.2     4.1 \u00b1 1.2     3.1 \u00b1 0.5     25        61.8 \u00b1 1.4\n  100      2.2 \u00b1 1.3     2.2 \u00b1 1.3     87.8 \u00b1 0.5    4.1 \u00b1 0.4     0.2 \u00b1 1.6     0.9 \u00b1 1.6     --1.8 \u00b1 0.6   50        82.0 \u00b1 1.2\n\nLiquiritin apioside (1), liquiritin (2), ferulic acid (3), glycyrrhizin (4), decursin (5), decursinol angelate (6), and atractylenolide I (7); \\* L-ascorbic acid was used as a positive control of antioxidant; The quantitative data are presented as the mean \u00b1 SEM of triplicate experiments.\n\nmolecules-23-01659-t008_Table 8\n\n###### \n\nComposition of UGS.\n\n  Latin Name                      Scientific Name           Amount (g)   Origin\n  ------------------------------- ------------------------- ------------ ----------------\n  Uncariae Ramulus et Uncus       *Uncaria sinensis*        6            China\n  Atractylodis Rhizoma Alba       *Atractylodes japonica*   8            China\n  Poria Sclerotium                *Poria cocos*             8            China\n  Bupleuri Radix                  *Bupleurum falcatum*      4            China\n  Angelicae Gigantis Radix        *Angelica gigas*          6            Bonghwa, Korea\n  Cnidii Rhizoma                  *Cnidium officinale*      6            China\n  Glycyrrhizae Radix et Rhizoma   *Glycyrrhiza uralensis*   3            China\n  Total amount                                              41           \n"}
+{"text": "Introduction {#s1}\n============\n\nOver the last decades, the role that biodiversity plays in ecosystem functioning has emerged as a key issue in ecology [@pone.0044247-Unsicker1], [@pone.0044247-Bertheau1], [@pone.0044247-Hooper1]. Although a majority of studies have focussed on the effect of plant diversity on primary production [@pone.0044247-Cardinale1], a growing attention is being paid on other ecosystem services provided by biodiversity such as pest regulation.\n\nThe diversity -- resistance hypothesis states that species rich plant communities suffer less feeding damage by herbivores than plant monocultures [@pone.0044247-Andow1], [@pone.0044247-Balvanera1], [@pone.0044247-Jactel1]. However two opposite effects of plant diversity on herbivory have been observed [@pone.0044247-Barbosa1]. A given focal plant species can experience more damage when associated with other plant species that are more attractive or palatable for herbivores [@pone.0044247-Barbosa1], [@pone.0044247-White1]. This pattern is known as associational susceptibility and seems to mainly involve generalist herbivore species [@pone.0044247-Jactel1]. Conversely, a focal plant species can have less herbivore damage (*i.e.* associational resistance) when the presence of non conspecific neighbours (i) reduces host plants concentration and the probability to be located by specialist herbivores [@pone.0044247-Root1]; (ii) provides physical or chemical barriers to host colonisation [@pone.0044247-Finch1], [@pone.0044247-Floater1], [@pone.0044247-Jactel2] and (iii) increases the abundance, the diversity and/or the efficiency of natural enemies [@pone.0044247-Hambck1], [@pone.0044247-Jactel3], [@pone.0044247-Riihimaki1], [@pone.0044247-Castagneyrol1]. Several meta-analyses have shown that associational resistance is more frequent than associational susceptibility but the balance between these two mechanisms is likely to depend on the identity of host plant species, herbivore feeding guilds or the way herbivory is assessed (abundance of herbivores *vs* biomass removed) [@pone.0044247-Andow1], [@pone.0044247-Jactel1], [@pone.0044247-Barbosa1], [@pone.0044247-Vehvilainen1].\n\nIntraspecific diversity (*i.e.* genetic diversity) is a key component of biodiversity. Recent research in the field of community genetics has shown that host plant genotype is one of the ecological filters shaping the structure of insect species assemblages [@pone.0044247-Whitham1], [@pone.0044247-Wimp1], [@pone.0044247-Bangert1] and that insect species diversity increases with the genetic diversity in host plant populations [@pone.0044247-Crutsinger1], [@pone.0044247-Crawford1], [@pone.0044247-Johnson1]. The question of the effects of genetic diversity on ecosystem functioning has also attracted considerable interest in recent years. It has been shown that most of the ecosystem functions provided by species diversity are also supported by genetic diversity, including plant productivity [@pone.0044247-CookPatton1], [@pone.0044247-Crutsinger2] nutrient cycling [@pone.0044247-Madritch1], temporal stability [@pone.0044247-Booth1], [@pone.0044247-Gamfeldt1], [@pone.0044247-Reusch1] and resistance to invasion [@pone.0044247-Crutsinger1]. Despite the similarity between the effects of plant species and plant intraspecific diversity on ecosystem properties, the mechanisms underlying the biodiversity-ecosystem functioning relationship may be different at the two scales. For example, Cook-Patton *et al.* [@pone.0044247-CookPatton1] showed that the increase in arthropod species richness with plant genetic diversity is mediated by arthropod abundance while resource specialisation is the main factor explaining the increase in arthropod species richness with plant species richness. It is therefore necessary to verify whether the ecological processes leading to associational resistance or susceptibility in plant species assemblages also apply to assemblages of plant genotypes.\n\nBecause herbivore species richness and abundance generally increase with genetic diversity in host plant populations [@pone.0044247-Crutsinger1], [@pone.0044247-CookPatton1], [@pone.0044247-Crutsinger2], associational susceptibility may be more likely to occur than associational resistance in mixtures of host plant genotypes. In addition, generalist insect herbivores (such as grasshoppers and many leaf chewers) are known to develop better on plant species mixtures due to food resources complementation or toxins dilution, a phenomenon known as diet mixing [@pone.0044247-Karban1], [@pone.0044247-Bernays1] that has been reported for mixtures of plant genotypes [@pone.0044247-Kotowska1], [@pone.0044247-Mody1]. Generalist herbivores are then expected to cause higher damage in genotype mixtures. Recent studies have shown positive [@pone.0044247-Crawford1], [@pone.0044247-Utsumi1] or neutral [@pone.0044247-Genung1] effects of host plant genetic diversity on the abundance of specialist herbivores. It is therefore still uncertain whether plant genetic diversity might have different effects on herbivores with different diet breadth or feeding behaviour [@pone.0044247-Kaplan1], [@pone.0044247-Ali1].\n\nWith a few exceptions [@pone.0044247-Madritch1], [@pone.0044247-Tack1], [@pone.0044247-Tack2], [@pone.0044247-Madritch2], studies on the functional consequences of genetic diversity for ecosystem functioning have focused on hybrids [@pone.0044247-Wimp1], [@pone.0044247-Dungey1], [@pone.0044247-Hochwender1] or clones [@pone.0044247-Crawford1], [@pone.0044247-CookPatton1], [@pone.0044247-Utsumi1], [@pone.0044247-Barbour1], [@pone.0044247-Johnson2], [@pone.0044247-Tack3]. Because these studies were designed so as to increase the contrast between plant genotypes, they may not be relevant to more complex processes occurring in more natural conditions [@pone.0044247-Tack3]. We present here one of the first attempt to assess the effect of casual intraspecific plant diversity on natural insect herbivory. Using an experimental plantation of pedunculate oak saplings, we tested the following hypotheses: (i) the genetic diversity of young trees tends to increase insect herbivory (*i.e.* associational susceptibility) and (ii) the magnitude of the effect depends on host specialization of insect herbivores, being higher for more generalist species. To test these hypotheses we designed a common garden experiment with 90 synthetic mixtures of oak saplings composed of one to four half-sib families. We genotyped all saplings and evaluated the amount of damage caused by ectophagous insect herbivores (less specialized) and endophagous leaf miners (more specialized) on each individual sapling. We assessed the level of genetic diversity in each mixture and estimated the correlation between diversity and insect herbivory.\n\nMaterials and Methods {#s2}\n=====================\n\nNo specific permits were required for the described field studies. The site on which the experimental common garden was established is owned by our institute (INRA) and is no subjected to any protection scheme. This work did not involve any endangered or protected species or area.\n\nExperimental design {#s2a}\n-------------------\n\nIn autumn 2007, we collected acorns from the canopy of four mature pedunculate oaks (*Quercus robur*), referred to hereafter as 'mother trees', sampled at random within a 10 km radius at a site 40 km south of Bordeaux (44\u00b0440 N, 00\u00b0460 W). In March 2008, we sowed the acorns at the nursery of the forest research centre of the French National Institute for Agricultural Research (INRA), to produce four half-sib families of oak seedlings. The seedlings were grown in individual pots containing peat and were treated with fungicide and insecticide during the first growing season (*i.e.* 2008), to prevent damage before planting. In March 2009, the seedlings were transplanted to a clearing surrounded by pine trees (*Pinus pinaster*) and broadleaved species (*Quercus robur, Quercus rubra* and *Betula pendula*).\n\nSix different blocks were established, with 15 plots in every block, each plot corresponding to one of the 15 possible combinations of one (*n*\u200a=\u200a4 plots, *i.e.* one per family), 2 (*n*\u200a=\u200a6), 3 (*n*\u200a=\u200a4) and 4 (*n*\u200a=\u200a1) families per plot. Each plot contained four rows of three seedlings; the seedlings were 0.2 m apart and the plot area was 0.24 m^2^ (0.60\u00d70.40). Within each plot, oak families were planted at equal density in a regular alternate pattern, such that seedlings from the same family were never adjacent in mixed plots. The plots were separated by a distance of 3 m and were randomly distributed within the blocks. Blocks were 14 m \u00d76 m in size and were located 4 m apart ([Figure S1](#pone.0044247.s001){ref-type=\"supplementary-material\"}).\n\nThe experimental site was fenced to prevent grazing by mammalian herbivores. The herbaceous plants growing between plots were removed by mowing, twice yearly. Pine bark chips were spread on the soil of each plot to control the vegetation and limit evaporation. Plots were watered during the summer of 2009, to minimise seedling mortality. In August 2011, 25 out of the 1080 planted seedlings were dead (*i.e.* 1055 survived).\n\nHerbivory assessment {#s2b}\n--------------------\n\nInsect herbivory was assessed by the visual inspection of 20 leaves on each four-year-old sapling, in August 2011. Five leaves were sampled at the tip and five at the base of two branches randomly chosen at the top and two branches randomly chosen towards the bottom of the sapling. We also recorded the total height of each sapling during this herbivory assessment.\n\nHerbivore damage on oak leaves was assigned to four different trophic guilds: *chewers* (mostly adult Curculionidae or Chrysomelidae and Lepidoptera caterpillars), *skeletonisers* (adult grasshoppers and Tenthredinoidea larvae), *rollers* (mostly Lepidoptera larvae) and *miners* (mostly Microlepidoptera larvae). No gall makers were observed. The percentage leaf area affected was visually estimated for each leaf and each guild using six classes (0%, 1--5%, 6--15%, 16--25%, 26--50%, 51--75% and \\>76%) and then averaged per sapling.\n\nDamage due to skeletonisers and leaf-rollers were very rare. We therefore pooled these two guilds with the chewers and classified the damage inflicted as being due to 'ectophagous insects'. Previous work by Giffard *et al.* [@pone.0044247-Giffard1] in the same study area showed that most of ectophagous insect herbivores found feeding on *Q. robur* are polyphagous species able to consume plant tissues from different genera and families and may be then considered as generalists (see [@pone.0044247-Giffard1] for a list of the commonest species). Leaf miners are different from the other insect herbivores found on oak saplings in that they are endophagous and much more specialized (they develop on a narrow spectrum of species within the Fagaceae family). Damage by leaf miners was quite frequent but minor in term of leaf area impacted. In addition, the leaf surface affected by a mine is dependent on the timing of assessment, while the presence or absence of a mine is not. We therefore used the density of mines per sapling (number of mines/20 leaves) to quantify damage due to these specialist insects.\n\nGenotyping of oak saplings {#s2c}\n--------------------------\n\nAll oak saplings and the four mother trees were genotyped with 12 microsatellite markers (see Guichoux *et al*. [@pone.0044247-Guichoux1] for details), using one leaf per sapling and per mother tree collected in August 2010. Leaves were dried and stored separately before DNA extraction and gene amplification. We isolated DNA from five leaf discs, each 5 mm in diameter, from each sample with the Invisorb DNA plant HTS 96 kit (Invitek, Berlin, Germany). We used the 12plex SSR (Single Sequence Repeats) kit developed by Guichoux *et al.* [@pone.0044247-Guichoux1] for genotyping. We scored SSR profiles, using real allele sizes and alleles were binned with the Microsoft Excel macro AUTOBIN program (available from <http://www4.bordeaux-aquitaine.inra.fr/biogeco/Ressources/Logiciels/Autobin>) developed by Guichoux *et al.* [@pone.0044247-Guichoux2].\n\nAmong the 1059 surviving individuals in 2010, 1032 were successfully genotyped. The mean proportion of loci succesfully typed was 99.7%. The mean number of alleles per locus was 11 (range: 6--19). More detailed information about genetic structure of the oak seedlings population is provided in [Table S1](#pone.0044247.s005){ref-type=\"supplementary-material\"}. Seventeen offspring (1.7%) were excluded from the analysis because their genotype at multiple loci did not match that of any mother tree. 134 offsprings showed only one mismatch with the corresponding mother tree. These offsprings were used to identify loci with genotyping errors before correction. Error rates based on these comparisons were low for 10 markers (\\<2%), high for one single-nucleotide marker (1.92% for the PIE258 marker) and high for another marker (10.49% for the PIE020 marker). Comparisons of the genotypes of mother trees and offspring revealed that manual binning was incorrect for the single-nucleotide marker and a null allele in the offspring of one mother tree, for the PIE020 marker. Single-nucleotide errors were corrected for further analysis and manual binning was repeated for the PIE258 marker. The PIE020 marker was removed from the data set. We finally retained 11 markers for the genotyping of 1016 offsprings plus the four mother trees.\n\nEstimation of genetic diversity {#s2d}\n-------------------------------\n\nWe initially used the number of maternal lineages per plot as a measure of genetic diversity. However, as a given mother tree could have been pollinated by several father trees, the offspring may be half-sibs or full-sibs and the proportion of the two types of saplings could vary within families and within sapling assemblages. The number of maternal lineages per plot may therefore underestimate genetic diversity and be poorly correlated with variation in insect damage.\n\nWe then determined SSR genotypes, to calculate the genetic relatedness between oak saplings, thereby improving estimates of genetic diversity per plot and switching from an almost categorical (1, 2, 3 or 4 maternal lineages per plot) to a more continuous (90 individual scores of genetic diversity) variable. Hereafter, *genetic diversity* (GD) refers to the number of maternal lineages per plot, whereas *genetic relatedness* (GR) refers to the mean between-saplings relatedness per plot.\n\nGenetic relatedness was calculated with CoAncestry software [@pone.0044247-Wang1]. We used the DyadML estimator (a dyadic likelihood estimator described in [@pone.0044247-Milligan1]) because the simulated values of relatedness it provided were the closest to expected values (*i.e.* 0.5 for full sibs, 0.25 for half sibs and 0 for unrelated saplings). GR was calculated for all pairs of individuals (*n*\u200a=\u200a\u00bd(1016\u00d7(1016--1)) \u200a=\u200a515,620 pairs) and we used these values to calculate a mean genetic relatedness for each plot. Mean genetic relatedness significantly differed between plots with different numbers of maternal lineages (Kruskal-Wallis test: *K~calc~* \u200a=\u200a845.55, *df* \u200a=\u200a3, *p*\\<0.001), decreasing with increasing number of lineages ([Figure S2](#pone.0044247.s002){ref-type=\"supplementary-material\"}). However, genetic relatedness also varied considerably within each level of genetic diversity, thus supporting the use of the two indices. As they were highly correlated (*r*\u200a=\u200a--0.80, [Figure S2](#pone.0044247.s002){ref-type=\"supplementary-material\"}), GD and GR were introduced separately in further models.\n\nOwing to missing genotypes (dead saplings, unamplified DNA, mismatch between observed genotype and mother tree), mean relatedness was averaged across a variable number of individual saplings per plot (9 to 12). For the sake of consistency, missing genotypes were also removed before the analysis of insect damage data. The final dataset contains 1002 individuals (6 blocks \u00d715 plots \u00d712 trees --25 dead saplings --53 unamplified or mismatched genotypes).\n\nStatistical analyses {#s2e}\n--------------------\n\nResponse variables (*i.e.* herbivory by ectophagous insects, abundance of leaf miners and sapling height) were analysed using each individual tree as a replicate to make it possible to test for the possible effects of interactions between mother tree identity (MT) and GD or GR. We accounted for spatial replication by nesting the 'population effect' (*i.e.* 1 population \u200a=\u200a1 plot \u200a=\u200a12 saplings) within the block effect, both factors being treated as random effects in all mixed models, in order to specify that individual observations were correlated within blocks and within plots.\n\nWe first tested the MT effect on response variables in monocultures alone, to avoid confounding factors. Mother tree identity was part of the experimental design and we were interested in its influence on the mean of herbivory by ectophagous insects, abundance of leaf miners and tree height. We therefore treated this factor as a fixed effect because there were not enough levels on which to base an estimate of the variance of the total population (only four different mother trees).\n\nIn order to determine potential genetic effects on sapling height, we first performed two sets of linear mixed models with MT and GD or GR, separately, and their interactions as fixed effects. We then carried out linear mixed models to test the effect of sapling height, MT, GD or GR, and their interactions on herbivory by ectophagous insects (% leaf area damaged) and by endophagous insects (abundance of leaf miners), separately. Prior to analyses, continuous explanatory variables (GD, GR and sapling height) were centred (*i.e.* subtracting the sample mean from all observations) and reduced (*i.e.* dividing centred variables by their sample standard deviation) in order to make model coefficients comparable within and between models [@pone.0044247-Schielzeth1] and to allow estimating the magnitude of effects. Centring variables also makes main effects biologically interpretable even when involved in interactions [@pone.0044247-Schielzeth1].\n\nIn all mixed models, we applied a model simplification procedure and reduced each maximal mixed model by removing non significant interaction terms, starting with the highest order interaction, to finally retain the least parameterized models including only simple terms and significant interaction terms.\n\nTest statistics for fixed effects were based on F values for linear mixed models (herbivory by ectophagous insects and sapling height) and on \u03c7^2^ values (loglikelihood ratio tests with one degree of freedom) for generalised linear mixed models performed on the abundance of leaf miners. Log-likelihood R^2^ values were calculated to estimate the amount of variance explained by each independent variable [@pone.0044247-KramerMR2Statisticsformixed1].\n\nData for sapling height and damage due to ectophagous insects were analysed with linear mixed models with the *lme* procedure [@pone.0044247-Pinheiro1] in R [@pone.0044247-R1]. Tree height was square-transformed and percentage data were transformed with the *logit* function [@pone.0044247-Warton1] to meet the assumptions of these tests, which were checked by graphical analyses and Shapiro-Wilk tests on model residuals. The abundance of leaf miners per tree was expressed as counts, which were analysed with generalized linear mixed models by specifying a Poisson error structure, with the *lmer* procedure in the lme4 package in R [@pone.0044247-Bates1].\n\nWe used the method developed by Loreau and Hector [@pone.0044247-Loreau1] and adapted by Unsicker *et al.* [@pone.0044247-Unsicker1] to quantify the net genetic diversity effect on herbivore damage. We first calculated the observed relative forage of the family *i* (*RF* ~O*i*~) as the ratio of the damage observed on each family *i* (*i* from 1 to 4) in a mixture (*C~i~*) to that observed on this family in monoculture (*M~i~*) [@pone.0044247-Unsicker1]:\n\nThe expected relative forage of the family *i* (*RF* ~E*i*~) under the null hypothesis (*i.e.* no effect of genetic diversity on damage) was simply its proportion in the mixture, *i.e.* 1/*n* where *n* is the number of families in the mixture [@pone.0044247-Unsicker1], [@pone.0044247-Loreau1].\n\nThe deviation of the observed relative damage in a mixture from the relative damage expected in the corresponding monoculture was thus:\n\nThe total observed damage in the mixture was calculated as:\n\nThe total expected damage in the mixture was calculated as:\n\nA positive NGDE indicates associational susceptibility (higher level of damage observed in mixtures than expected from mean damage levels in the corresponding monocultures), whereas a negative NGDE indicates associational resistance (lower level of damage observed in mixtures than expected from mean damage levels in the corresponding monocultures).\n\nThe NGDE can be further divided into two additive components: a complementarity effect (CE) and a selection effect (SE) [@pone.0044247-Unsicker1], [@pone.0044247-Loreau1].\n\nThe CE is assessed by calculating the mean *\u0394RC~i~* over all families at the plot level:\n\nCE measures the change in mean relative forage of the species. CE is positive when the mean relative forage increases *i.e.* when oak families are, on average, consumed more in mixtures than it would be expected from their consumption in monocultures.\n\nThe calculation of SE takes into account the covariance between \u0394RC~i~ and M~i~:\n\nSE values are used to determine whether there is a relationship between consumption in the monoculture and relative forage in mixtures. SE is positive when plant species that are consumed in larger amounts in monocultures (less resistant) also have higher relative forage values in mixtures, thus making a greater contribution to total plot damage.\n\nNGDE, CE and SE were calculated for all levels and combinations of mixtures within each block, giving a total of 66 comparisons between observed and expected values. The significance of each effect (NGDE, CE, SE) was determined by one sided t-tests [@pone.0044247-Loreau1]. We first tested grand mean values across all mixtures against zero, to determine whether they differed significantly from the weighted average of the response variable in monocultures. We also assessed the significance of the NGDE, CE and SE against zero for each level of genetic diversity. We used analyses of variance to assess change in these three effects along the gradient of GD [@pone.0044247-Loreau1].\n\nResults {#s3}\n=======\n\nEffects of genetic diversity and relatedness on sapling height {#s3a}\n--------------------------------------------------------------\n\nMean sapling height significantly differed between oak families (*F* ~3,909~ \u200a=\u200a2.88, *p*\u200a=\u200a0.035) but we observed no significant effect of genetic diversity (GD: *F* ~1,83~ \u200a=\u200a0.02, *p*\u200a=\u200a0.880) or genetic relatedness (GR: *F* ~1,83~ \\<0.01, *p*\u200a=\u200a0.984) on sapling height.\n\nEffects of genetic diversity and relatedness on insect herbivory {#s3b}\n----------------------------------------------------------------\n\nDamage due to ectophagous insects was significantly affected by MT, GD, GR and sapling height (H), but not by interactions between these factors ([Table 1](#pone-0044247-t001){ref-type=\"table\"}). Significant differences in damage levels between families were observed in monocultures (on average 5.5 and 8.3% of leaf area was removed in the more and the less resistant families, respectively), suggesting a genetic control of oak saplings resistance to ectophagous insects (*F* ~15,244~ \u200a=\u200a4.81; *P*\u200a=\u200a0.015; *R* ^2^\u200a=\u200a0.04, [Figure 1A](#pone-0044247-g001){ref-type=\"fig\"}). Damage also increased significantly with the GD of saplings ([Figure 2B](#pone-0044247-g002){ref-type=\"fig\"}) and decreased significantly with increasing GR, regardless of the family considered ([Figure 2C](#pone-0044247-g002){ref-type=\"fig\"}), indicating that the presence of more genetically diverse neighbours increased the risk of damage and that this risk increased with the diversity of conspecific neighbours. However, the magnitude of this effect was low, leaf area removed being on average 6.9% in monocultures and 7.9% in 4-families mixtures. Damage by ectophagous insects also increased significantly with sapling height ([Table 1](#pone-0044247-t001){ref-type=\"table\"}, [Figure 2A](#pone-0044247-g002){ref-type=\"fig\"}). The effects of GD, GR and H on damage by ectophagous herbivores were comparable in terms of magnitude, as shown by standardized model coefficients ([Table 1](#pone-0044247-t001){ref-type=\"table\"}). The effects of GD and GR on herbivory seemed to be direct rather than mediated by the genetic control of sapling height as (i) GD and GR had no effect on height and (ii) MT\u00d7H (*F* ~3,905~ \u200a=\u200a1.56; *p-value* \u200a=\u200a0.198), GD\u00d7H (*F* ~1,901~ \u200a=\u200a0.08; *p-value* \u200a=\u200a0.775) and GR\u00d7H (*F* ~1,907~ \u200a=\u200a2.92; *p-value* \u200a=\u200a0.088) interactions had no significant effect on damage ([Table 1](#pone-0044247-t001){ref-type=\"table\"}).\n\n![Effect of mother tree identity on insect herbivores in monocultures.\\\n(A) Effect of mother tree identity on damage (% leaf area removed) due to ectophagous herbivores. (B) Effect of mother tree identity on the abundance of endophagous insect herbivores. Semi transparent coloured circles represent individual saplings. Darkest circles represent overlapping datapoints. Solid black circled dots indicate the mean values in monocultures for all saplings and all blocks. Same letter above two lines of dots indicates that the corresponding means were not significantly different (LMM and GLMM on monoculture plots).](pone.0044247.g001){#pone-0044247-g001}\n\n![Effects of sapling height, genetic diversity and genetic relatedness on ectophagous and endophagous insects.\\\nEffects of sapling height (A, D), genetic diversity (B, E) and genetic relatedness (C, F) on damage due to ectophagous insects (A, B, C) and on the abundance of endophagous insects (D, E, F). The different colours indicate regression lines for different families (MT1: green, MT2: blue, MT3: red, MT4, purple). The shaded areas indicate the corresponding 95% confidence intervals.](pone.0044247.g002){#pone-0044247-g002}\n\n10.1371/journal.pone.0044247.t001\n\n###### Summary of the results of linear mixed models assessing the effect of sapling height (H), mother tree identity (MT), genetic diversity (GD) and genetic relatedness (GR) between oak saplings and their interactions on herbivory by ectophagous insects and on abundance of endophagous insects (leaf-miners).\n\n![](pone.0044247.t001){#pone-0044247-t001-1}\n\n                                  Ectophagous insects   Endophagous insects                                                                         \n  --------------------- -------- --------------------- --------------------- ------- ------------- ------- ---------------- --------- ------------- -------\n  Genetic diversity        H            1, 908              0.05 \u00b1 0.02       6.51     **0.011**    0.006    0.31 \u00b1 0.052    127.53    **\\<0.001**  \n                           MT           3, 908                                18.39   **\\<0.001**   0.052                     7.65        0.054     \n                           GD            1, 83              0.06 \u00b1 0.03       4.45     **0.038**    0.004   \u22120.001 \u00b1 0.080   \\<0.001      0.995     \n                         H \u00d7 GD           \\-                                                                \u22120.003 \u00b1 0.088    12.50     **0.014**    0.004\n  Genetic relatedness      H            1, 908              0.05 \u00b1 0.02       6.57     **0.011**    0.007    0.31 \u00b1 0.05     128.48    **\\<0.001**  \n                           MT           3, 908                                19.22   **\\<0.001**   0.054                     7.51        0.057     \n                           GR            1, 83             \u22120.06 \u00b1 0.03       4.49     **0.037**    0.004    \u22120.04 \u00b1 0.08     1.08        0.300     \n                         H \u00d7 GR           \\-                                                                 0.08 \u00b1 0.27      15.49     **0.004**    0.007\n\nResults are given from LMM and Poisson GLMM for ectophagous and endophagous herbivores respectively.\n\n*df* degrees of freedom (numerator, denominator).\n\nLog-likelihood R^2^ were not estimated in case of significant H \u00d7 GR and H \u00d7 GD interactions.\n\nFor endophagous herbivores (*i.e.* leaf miners), sapling height emerged as the main factor determining their abundance on individual saplings ([Table 1](#pone-0044247-t001){ref-type=\"table\"}, [Figure 2D](#pone-0044247-g002){ref-type=\"fig\"}). GD\u00d7H and GR\u00d7H interactions also significantly affected the abundance of leaf miners ([Table 1](#pone-0044247-t001){ref-type=\"table\"}) while H\u00d7MT did not (\u03c7^2^\u200a=\u200a1.05; *p-value* \u200a=\u200a0.789). The coefficient estimate of GD\u00d7H interaction term was negative ([Table 1](#pone-0044247-t001){ref-type=\"table\"}) which means that the strength of the effect of sapling height on abundance of leaf miners decreased when increasing GD. The opposite was true for GR\u00d7H ([Table 1](#pone-0044247-t001){ref-type=\"table\"}), which is consistent with the negative covariation between GD and GR. However, standardized coefficients of regression of both GD\u00d7H and GR\u00d7H were low compared to the coefficient of regression for H. The simple effects of MT ([Figure 1B](#pone-0044247-g001){ref-type=\"fig\"}), GD ([Figure 2E](#pone-0044247-g002){ref-type=\"fig\"}) and GR ([Figure 2F](#pone-0044247-g002){ref-type=\"fig\"}) on leaf miner abundance were not significant ([Table 1](#pone-0044247-t001){ref-type=\"table\"}).\n\nNet genetic diversity effect {#s3c}\n----------------------------\n\nThe net genetic diversity effect (NGDE) on herbivory by ectophagous insects was overall significantly positive ([Table 2](#pone-0044247-t002){ref-type=\"table\"}), indicating a higher level of damage in mixtures than expected from monocultures (*i.e.* associational susceptibility). Both complementarity and selection effects (CE and SE) were significantly different from zero ([Table 2](#pone-0044247-t002){ref-type=\"table\"}) but had opposite signs ([Figure 3](#pone-0044247-g003){ref-type=\"fig\"}): mean CE was positive and more than three times higher than mean SE, which was negative. The resulting positive NGDE was therefore principally due to the positive complementarity effect. Mean NGDE and CE were consistently positive at each level of genetic diversity, and SE was significantly negative at all but the higher level of GD ([Figure 3](#pone-0044247-g003){ref-type=\"fig\"}, [Table 2](#pone-0044247-t002){ref-type=\"table\"}). A negative Selection Effect indicates a negative covariation between damage in monocultures and the deviation between observed and expected relative damage in mixtures ([Figure S3](#pone.0044247.s003){ref-type=\"supplementary-material\"}). For families with lower levels of damage in monocultures (*i.e.* intrinsically more resistant), herbivory in mixtures was much higher than expected ([Figure S3](#pone.0044247.s003){ref-type=\"supplementary-material\"}).\n\n![Non-additive effect of genetic diversity insect herbivores.\\\n(A) Test of the non-additive effect of genetic diversity on ectophagous insects. (B) Test of the non-additive effect of genetic diversity on endophagous insects. Semi transparent circles represent individual values per plot for net genetic diversity effect (NGDE, blue), complementarity effect (CE, red) and selection effect (SE, green). Solid black circled dots are the averaged values for all plots (grand mean) and each level of genetic diversity (GD). The '\\*' symbol are for means value significantly different from zero.](pone.0044247.g003){#pone-0044247-g003}\n\n10.1371/journal.pone.0044247.t002\n\n###### Summary of *t* values from *t*-tests for net genetic diversity effect (NGDE), complementarity effect (CE) and selection effect (SE) on damage of ectophagous and abundance of endophagous (leaf-miners) insects, and on sapling height, for all mixtures (grand mean) and for each level of genetic diversity (GD).\n\n![](pone.0044247.t002){#pone-0044247-t002-2}\n\n                                  Grand mean          GD\u200a=\u200a2           GD\u200a=\u200a3         GD\u200a=\u200a4\n  --------------------- ------ ----------------- ----------------- --------------- -------------\n  Ectophagous insects    NGDE     **2.44**\\*           1.30             1.79           1.43\n                          CE     **3.37**\\*\\*          1.95          **2.39**\\*        1.57\n                          SE    **\u22125.51**\\*\\*\\*   **\u22124.07**\\*\\*\\*   **\u22123.51**\\*\\*      \u22121.38\n  Endophagous insects    NGDE        \u22120.08             0.28             \u22120.61          0.01\n                          CE         1.70              1.40             0.89           0.37\n                          SE    **\u22125.46**\\*\\*\\*   **\u22123.96**\\*\\*\\*   **\u22123.16**\\*\\*   **\u22122.69**\\*\n  Sapling height         NGDE        0.57              1.01             \u22121.12         2.76 \\*\n                          CE         1.05              1.34             \u22120.77         2.90 \\*\n                          SE    **\u22124.31**\\*\\*\\*    **\u22123.51**\\*\\*     **\u22122.46**\\*       \u22120.91\n\nSignificant t-values are in bold: (\\*\\*\\*) P-value \\<0.001, (\\*\\*) 0.001\\< P-value \\<0.01, (\\*) 0.01\\< P-value \\<0.05.\n\nThere was no significant NGDE or CE on the abundance of endophagous herbivores ([Table 2](#pone-0044247-t002){ref-type=\"table\"}). By contrast, SE was significant and negative at all levels of genetic diversity ([Figure 3](#pone-0044247-g003){ref-type=\"fig\"}, [Table 2](#pone-0044247-t002){ref-type=\"table\"}). Thus, for the families showing a tendency for higher infestation in monocultures (more susceptible), relative infestation levels were lower than expected in mixtures, the opposite being true for less susceptible families. CE and SE were of similar magnitude but of opposite signs, accounting for the null NGDE.\n\nNGDE and CE were not significant for sapling height, for either the grand mean, or for any of the levels of genetic diversity, with exception of the 4-families mixtures ([Table 2](#pone-0044247-t002){ref-type=\"table\"}). SE was consistently and significantly negative (but for GD\u200a=\u200a4).\n\nDiscussion {#s4}\n==========\n\nBased on a large number of samples and a manipulative experiment this study shows for the first time that genetic diversity can trigger associational susceptibility to insect herbivory [@pone.0044247-Barbosa1], [@pone.0044247-White1] in tree saplings. This process describes an increase in insect herbivory with increasing genetic diversity in host population.\n\nThe relative importance of genetic diversity *vs.* other ecological factors as drivers of ecosystem processes is a central issue for community genetics [@pone.0044247-Tack3], [@pone.0044247-Hughes1], [@pone.0044247-HerschGreen1]. In our study we found that the effects of genetic diversity or relatedness on insect herbivory were overall significant but low in terms of magnitude. These results are consistent with the small effects of tree genetic diversity on structuring the insect community associated with pedunculate oak, as recently reported by Tack *et al.* [@pone.0044247-Tack1], [@pone.0044247-Tack2]. In addition, we showed that sapling height was as important as genetic diversity for predicting generalist herbivore damage and the best predictor of endophagous herbivores abundance. These findings suggest that the influence of host tree genetic diversity on insect herbivores may originate in the variance of particular functional traits.\n\nIf genetically based differences in tree susceptibility to herbivores is now well documented [@pone.0044247-Tack1], [@pone.0044247-Tikkanen1], [@pone.0044247-Smith1], [@pone.0044247-Silfver1], [@pone.0044247-Ito1], the effect of genetic diversity on insect damage has rarely been investigated and most often on crops or herbaceous plants [@pone.0044247-Crawford1], [@pone.0044247-Genung1], [@pone.0044247-Hajjar1]. Recently, Tack *et al.* [@pone.0044247-Tack1], [@pone.0044247-Tack2] studied the effects of genotype identity and diversity on the structure of endophagous insect communities on *Quercus robur*, but they did not measure corresponding herbivory. We are aware of only two studies that investigated the relationship between tree genetic diversity and insect damage. They reported a trend towards higher levels of pest damage in monocultures than in mixtures of willow clones [@pone.0044247-Peacock1], [@pone.0044247-Peacock2] but they focused on only two specialized leaf beetles. The consequences of tree genetic diversity on total herbivory remain largely unknown. A similar concern is currently emerging about the effect of plant species diversity on insect herbivory. The diversity-resistance relationship has been clearly demonstrated by meta-analyses focusing on individual species-species interactions [@pone.0044247-Jactel1], [@pone.0044247-Letourneau1], [@pone.0044247-Tonhasca1]. However, the diet breadth of insect herbivores emerged as a key factor accounting for differences in insect response to plant diversity. Herbivory by oligophagous species is often reduced in mixed-species forests in comparison to monospecific forests, whereas the response of polyphagous insect species is more variable [@pone.0044247-Jactel1]. Several examples of such opposite patterns have recently been reported, with higher levels of damage caused by polyphagous insects [@pone.0044247-Schuldt1], [@pone.0044247-Sobek1] and lower abundance of oligophagous insects [@pone.0044247-Otway1] in more diverse plant communities. The effect of plant diversity on total insect herbivory may then primarily depend on the share of generalist and specialist herbivore species. Here we tentatively addressed this issue by considering two guilds of herbivores of contrasting diet breadth. Oak leaf-miners are oligophagous species that develop a narrow range of species within the *Quercus* and the *Castanea* genera while all the ectophagous insects we observed in the field were polyphagous species able to feed on host plants belonging to different families (see [@pone.0044247-Giffard1] for the list of insect herbivore species found on oak trees in the study area). As endophagous herbivores, leaf-miners have an intimate relationship with their host and are expected to be more dependent on host genotype than ectophagous insects that can move freely and exploit several hosts during their development [@pone.0044247-Wimp1], [@pone.0044247-Tack3].\n\nResponse of ectophagous herbivores {#s4a}\n----------------------------------\n\nIn the present study, we observed that the four oak families displayed different levels of resistance to generalist insect herbivores but also differed significantly in sapling height. Herbivory by generalist insects increased with sapling height. However there was no significant interaction between the effect of sapling height and genetic identity on damage by ectophagous herbivores. Moreover the genetic diversity and relatedness had only weak effects on sapling height whereas they significantly affected ectophagous insect herbivory. So the observed increase in damage caused by these herbivores in genetically diverse oak sapling mixtures was not mediated by differences in height. Yet, four ecological mechanisms may account for the observed relationship between genetic diversity and herbivory by ectophagous insects.\n\n### (i) Herbivore abundance {#s4a1}\n\nHerbivore abundance has been reported to increase with genetic diversity [@pone.0044247-CookPatton1]. Associational susceptibility may then have been driven by an increase in abundance of generalist herbivores in genotype mixtures. However, as we did not sampled insects, we cannot validate this hypothesis. In addition, the relationship between herbivore density and herbivory damage remains unclear and we are not aware of any study that convincingly demonstrated an increase in herbivory with the abundance of herbivores.\n\n### (ii) Mixing diet {#s4a2}\n\nHerbivory by ectophagous insects increased in genotype mixtures because of a higher consumption of the four oak families, regardless their intrinsic susceptibility in monocultures, as evidenced by a significant and positive complementarity effect. This is consistent with the observation that generalist insect herbivores can increase their fitness by feeding on different host plants [@pone.0044247-Unsicker1], [@pone.0044247-Bertheau1]. Different plant genotypes may provide insects with feeding resources of different qualities [@pone.0044247-Madritch2]. Mixtures of genotypes are therefore likely to improve diet mixing, which is known to benefit generalist herbivores [@pone.0044247-Unsicker1], [@pone.0044247-Kotowska1], [@pone.0044247-Mody1]. It has been also proposed that feeding on different host plants results in the dilution of toxic compounds present in the plant tissues, allowing a more balanced input of nutrients [@pone.0044247-Karban1], [@pone.0044247-Bernays1]. It should be of great interest now to investigate leaf chemistry and check whether the blends of secondary metabolites involved in plant defence can explain herbivory patterns in genotypes mixtures and possibly changes with the genetic diversity of mixtures.\n\n### (iii) Spill over {#s4a3}\n\nThe higher damage by herbivorous insects in plant species mixtures (*i.e.* associational susceptibility) has been initially attributed to a spill over of generalist herbivores from their preferred host plants to nearby suitable but less suitable host plants [@pone.0044247-White1]. Despite the fact we did not monitor the temporal dynamic of ectophagous insects on individual oak saplings, the negative selection effect we report, though low, may account for such a spill over. Indeed, a negative selection reveals the existence of negative covariance between observed damage in mixtures and observed damage in monocultures: the increase in damage with genetic diversity was higher for families that suffered less damage when growing in monocultures. This is consistent with the hypothesis of greater colonisation through contagion, with the transfer of insects from more to less palatable families, in sapling mixtures. A similar behaviour was recently reported by Utsumi *et al.* [@pone.0044247-Utsumi1] who observed a shift of insect herbivores from more to less preferred host genotypes in mixtures of annual plants.\n\n### (iv) Host location {#s4a4}\n\nThe way insect herbivores perceive their host plants may change with their genetic diversity as recently suggested by Crawford *et al.* [@pone.0044247-Crawford1] who showed a non additive increase in gall abundance on patches of *Solidago altissima* with higher genetic diversity. At the plot scale, associational susceptibility may be explained by a better patch detection by foraging herbivores. For example, the mixture of sapling genotypes may have increased the probability of incorporating tall saplings that could be easier to detect and colonise, as suggested by our observation of a significant effect of sapling height on insect damage. Consistent with this hypothesis, the difference in height between taller (75^th^ percentile of heights distribution) and medium-sized (median of heights distribution) saplings tended to increase with genetic diversity at the plot level ([Figure S4](#pone.0044247.s004){ref-type=\"supplementary-material\"}). In addition to visual cues that shape plant \"physical\" apparency, host plant location by insect herbivores is most often mediated by olfactory cues [@pone.0044247-Visser1]. Host-plant recognition depends on ratios of plant volatiles and not just on detection of the presence or absence of particular compounds [@pone.0044247-Bruce1]. Insects use blends of volatile compounds to distinguish between host and non-host plant species. It has recently been suggested that there is redundancy in the composition of host odour blends, with some components being substitutable to others [@pone.0044247-Bruce2]. It is therefore possible that a mix of host plant genotypes is more likely to produce the right combination of attractants than a monoculture of a single plant genotype [@pone.0044247-Glinwood1].\n\nHowever, as we did not sample insect herbivores, it is difficult to determine which one of these mechanisms is the more likely or if they operate synergistically. For example, for a single herbivore, diet mixing might actually lead to less herbivory if that individual is able to acquire more nutrients with a variety of host genotype consumed. But on the other hand, if mixed diets are more preferable, and if mixed diets are associated with mixed host finding cues, it might attract more individuals and lead to greater overall herbivore damage. As a result, the abundance of herbivores may have ultimately been the primary driver of associational susceptibility.\n\nResponse of endophagous herbivores {#s4b}\n----------------------------------\n\nNone of the genetic attributes (identity, diversity or relatedness) had a significant effect on the abundance of specialist herbivores (leaf-miners). This finding is consistent with previous studies showing that genotype [@pone.0044247-Tack1], and genetic diversity [@pone.0044247-Tack1] are poor predictors of the diversity of specialist herbivores (leaf-miners and gall-makers) feeding on oaks. Instead, sapling height emerged as the key determinant of leaf miner abundance. Consistently, the maternal lineages that produced the tallest saplings (MT2 and MT1) were also more infested by leaf miners (although not significantly) than those in which saplings were significantly smaller suggesting that genetically based differences in sapling height may drive differences in the abundance of leaf miners. The negative selection effect on abundance of leaf-miners indicates that the oak families that tended to be less infested in monocultures also tended to be more often colonised by leaf miners in mixtures than in monocultures. As the larval stages of leaf miners cannot relocate after oviposition and cannot shift from one host plant to the next in order to find new (spillover) or complementary (mixed diet) feeding resources, the distribution of leaf miners between and within plots thus reflects the choice of oviposition site by females. As proposed for ectophagous herbivores, a possible explanation of the negative selection effect is that mixing genotypes resulted in a greater probability of including taller and then more attractive saplings to endophagous insects ([Figure S4](#pone.0044247.s004){ref-type=\"supplementary-material\"}). One cannot exclude that the combination of relevant attractants was also more likely to occur in more diverse genotypes mixtures.\n\nThe mean abundance and species richness of leaf miners were very low in our experiment, with abundance scores of 0 to 10 mines in 20 leaves per sapling and 85% of total abundance represented by only two of the nine observed species (*Phyllonorycter* sp. and *Stigmella* sp.). Separate analyses of each species of leaf miner would have generated too many zero counts, so we decided to pool data into a single category of \"insect specialists\". However, Tack and Roslin [@pone.0044247-Tack1] showed that *Phyllonorycter* sp. and *Stigmella* sp. responded differently to genetic and environmental treatments. Considering the abundance of several leaf-miner species together may therefore have prevented the detection of species-specific abundance patterns. Further investigation, after the oak saplings have been colonised by a larger number of insect species, are required for more detailed comparisons of the responses of generalist and specialist herbivores to genetic diversity.\n\nConclusion {#s5}\n==========\n\nUnlike plant species richness, plant genetic diversity may not provide sufficient functional contrast to prevent host colonization by specialist herbivores while enhancing the foraging of generalist herbivores. Overall, we observed a significant effect of tree genetic diversity on generalist herbivores but not on specialists. Increasing genetic diversity resulted in higher damage by generalist herbivores because of (i) a general increase in leaf consumption in more diverse genotype mixtures (*i.e.* positive complementarity effect) and (ii) an increased damage exposure of individuals from more resistant genotypes in the vicinity of individuals from more susceptible genotypes (*i.e.* negative selection effect). To date many studies have shown that herbivore diversity increases with host plant genetic diversity [@pone.0044247-Crutsinger1], [@pone.0044247-Crawford1], [@pone.0044247-Johnson1], [@pone.0044247-CookPatton1], [@pone.0044247-Crutsinger2]. There is a need now to reconcile the two approaches and investigate the relationship between the diversity of herbivores and the resulting herbivory along gradients of plant genetic diversity. In addition, when saplings develop into young trees, they may be more easily located and infested by more specialized herbivores, therefore benefiting from being part of a mixed-genotype community [@pone.0044247-Peacock1], [@pone.0044247-Peacock2]. We would therefore advocate long-term monitoring of the dynamics of sapling colonisation by insects with various degrees of host plant range limitation, to determine whether the magnitude and direction of the effect of genetic diversity on associational herbivory change with the ontogeny of focal tree species.\n\nSupporting Information {#s6}\n======================\n\n###### \n\n**Experimental design.** Each colored square represents an individual oak sapling.\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Effect of the number of half-sib families per plot (Genetic Diversity, GD) on the mean genetic relatedness among oak seedlings within plots (Genetic relatedness, GR).** Open circles represent individual plots (*n*\u200a=\u200a90); filed circles represent mean genetic relatedness per level of genetic diversity.\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Negative selection effect of genetic diversity on exophagous herbivores.**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Effects of genetic diversity on oak height heterogeneity within plots.** Each dot represents the mean difference (\u00b1 SE) between the 75^th^ percentile and the median (50^th^ percentile) of sapling heights distribution within plots. This difference indicates how far taller trees were apparent to herbivores within plots.\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Summary of genetic data describing the genetic structure of the population used to constructed experimental plots.**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\nWe thank the INRA experimental units at Pierroton for their assistance in the establishment of the common garden experiments. We are also grateful to Audrey Lefran\u00e7ois and Rodolphe Lauverjon for their help in the field. Genotyping was carried out at the genome/transcriptome facility at Pierroton.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: HJ BC. Performed the experiments: BC LL. Analyzed the data: BC LL BG AK HJ. Contributed reagents/materials/analysis tools: BC LL BG. Wrote the paper: BC HJ. Obtained fundings: HJ AK.\n"}
+{"text": "In the article titled \"Five-Year Clinical Outcomes of Local versus General Anesthesia Deep Brain Stimulation for Parkinson\\'s Disease\" \\[[@B1]\\], there was an error in Methods where the IRB \"no. 097-32\" was missing. The sentence \"The institutional review board of Tzu Chi General Hospital approved this study (no. 097-08)\" should be corrected as follows: \"The Institutional Review Board of Tzu Chi General Hospital approved the studies (no. 097-08 and no. 097-32).\"\n"}
+{"text": "Pathological micro-domains have been proposed to underpin the generation of local pathological activity, as seen in focal seizures in the epileptic cortex \\[[@B1]-[@B3]\\]. Specifically, so-called micro-seizures have been suggested to be markers for these micro-domains \\[[@B2],[@B3]\\]. Astonishingly, micro-seizures have also been observed in non-epileptic control patients \\[[@B3]\\]. This suggests that local activity, such as micro-seizures, only become pathological when in a specific arrangement.\n\nWe hypothesize that pathological dynamics could be due to an increased density of micro-domains. To test this, we introduce a computational model on the mesoscopic scale of a 5 \u00d7 5 mm^2^cortical sheet \\[[@B4]\\]. The units are modelled as excitable minicolumns. This model also incorporates realistic connectivity schemes observed at this spatial scale \\[[@B5]\\].\n\nThe model shows occasional, non-pathological micro-seizure occurrences, as well as recruitment of normal tissue into full-blown seizure activity in the presence of dense clusters of hyperactive micro-seizure domains. A specific prediction of this model is that the transition to full-blown seizures can be prevented by using micro-incisions to separate the clusters of abnormally active micro-domains (Figure. [1](#F1){ref-type=\"fig\"}) \\[[@B6]\\].\n\n![**First column indicated the location of the pathological clusters in black and the position of the micro-incision using the labeled line**. The second column shows the temporal average activity in 10000 simulation steps. The third column indicates the spatial average of the same simulation using the black line and the purple lines show the spatial average of 10 \u00d7 10 subsquares, which represent macro-columns in the model. (A) Resulting simulation after a 5 minicolumn long micro-incision. (B) Resulting simulation after a 10 minicolumn long micro-incision. (C) Resulting simulation after a 45 minicolumn long micro-incision.](1471-2202-14-S1-P14-1){#F1}\n"}
+{"text": "Background {#Sec1}\n==========\n\nAdaptive immunity is an effective strategy for cancer treatment. T cells are the main force to combat microbes and cancer cells in adaptive cells \\[[@CR1], [@CR2]\\]. However, T cell is heterogeneous, which exists in a continuum of differentiation states. According to differentiation stages, T cells can be divided into naive T cells (T~N~), memory stem cells (T~SCM~), central memory cells (T~CM~), effector memory cells (T~EM~) and differentiated effectors (T~EF~) \\[[@CR3], [@CR4]\\]. Adoptive T cell therapy (ACT) for clinical application is normally based on the use of terminally differentiated T~EF~ cells, which have short lifespan and inferior engraftment capacities \\[[@CR5]--[@CR7]\\]. T~SCM~ is defined as the earliest developmental stage of memory T cells, which possesses capacities of self-renewal and differentiation \\[[@CR5], [@CR8]\\]. Phenotypically, T~SCM~ can be identified by the expression of CD3+ CD45RA+ CD62L+ CD95+ CCR7+ CD28+ \\[[@CR5], [@CR9]--[@CR11]\\]. T~SCM~ has longer lifespan, as it is reported T~SCM~ is able to be tracked in vivo for 12\u2009years after infusion of genetically modified lymphocytes \\[[@CR7]\\]. In particular, T~SCM~ mediates longer and more robust efficiency of tumor rejection in vivo compared to other memory and effector subsets \\[[@CR5], [@CR6], [@CR8], [@CR12]--[@CR14]\\]. It is these characteristics that make T~SCM~ a good candidate for immunological cytotherapy of cancer.\n\nT~SCM~ constitutes a small proportion of the T cell subset, 2--4% of the total T cells in peripheral blood \\[[@CR5], [@CR8]\\]. The premise of antigen-specific T~SCM~ used for clinical therapy is to accomplish in vitro T~SCM~ preparation in a large number. It is demonstrated that small molecule chemical inhibitor TWS119, a potent inhibitor of the serine--threonine glycogen synthase kinase 3\u03b2 (GSK-3\u03b2) able to induce the Wnt-\u03b2-catenin signaling, helps the enrichment of T~SCM~ through differentiation inhibition both in mice and in human \\[[@CR15]--[@CR17]\\]. On the other hand, T~SCM~ can be amplified with cytokines IL-7 and IL-15; IL-21 is also reported to promote the generation of T~SCM~ under ex vivo culture conditions \\[[@CR6], [@CR18]--[@CR26]\\].\n\nAllogeneic hematopoietic stem cell transplantation (allo-SCT) is an effective immunotherapeutic approach with curative potential in patients with malignancies. The therapeutic basis is mainly dependent on the donor T cell alloresponses against the recipient's malignant cells named as graft versus leukemia (GVL) or graft versus tumor (GVT) effect \\[[@CR27]\\]. T cell responses to alloantigen are of peptide-MHC complex (pMHC) specificity in the same way as those to nominal antigen \\[[@CR28], [@CR29]\\]. Transfer of alloreactive T cells with defined specificity, such as for leukemia- or tumor-associated antigens, is proposed to separate the GVL or GVT effect from the deleterious graft versus host disease (GVHD) \\[[@CR30], [@CR31]\\]. As the implantation and long-term existence of tumor-specific T cells in host are the prerequisite for adoptive immunotherapy, it is of importance to prepare alloantigen-specific T~SCM~ cells in vitro.\n\nIn this study, we explored a methodology of alloantigen-specific T~SCM~ (allo-specific T~SCM~) preparing in vitro. Although both CD4+ and CD8+ T~SCM~ cells are reported, we focused on CD8+ T~SCM~ cells. The T~SCM~ cells were generated in an allogeneic co-culture, enriched by TWS119, then sorted with proliferation, and finally expanded by IL-7 and IL-15. Our protocol prepared 2\u2009\u00d7\u200910^7^ allo-specific T~SCM~ cells from 1\u2009\u00d7\u200910^7^ PBLs. An LCL burden mouse model was introduced in for the T~SCM~ cell behavior in vivo. LCL cells were human B lymphoblastoid cells immortalized by EB virus infection. Although this model would reflect lymphoproliferative disorders, the LCL cells in mice could act as allogeneic targets to measure eradication efficacy of T cells in vivo in our study. Importantly, the prepared T~SCM~ cells exhibited both stem cell and memory T cell properties, and were able to implant and effectively eradicate the targets after adoptive transfer into mice. This study provided a practical method for generation of allo-specific T~SCM~ cells in vitro, which would be expected to be referred in preparation of allogeneic T~SCM~ grafts with defined antigen specificity for the purpose of adoptive immunotherapy.\n\nMaterials and methods {#Sec2}\n=====================\n\nPeripheral blood lymphocyte isolation {#Sec3}\n-------------------------------------\n\nPeripheral blood was obtained from healthy donors after informed consent under a protocol approved by the Ethics Committee of Tongji Medical College, Wuhan, China. Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation through a ficoll-hypaque gradient (Ficoll-Hypaquedensity 1.077\u2009g/ml) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), then placed in dish for 2\u2009h to remove the adherent cells. The non-adherent cells were collected as peripheral blood lymphocytes (PBLs) for the co-culture.\n\nCell lines and antibodies {#Sec4}\n-------------------------\n\nThe EB virus (EBV) transformed B lymphoblastoid cell lines (LCLs) E007 and E001 were established in our lab according to reported protocol \\[[@CR32]\\]. HLA typing for LCLs was performed with PCR-SSP \\[[@CR33]\\]. The HLA class I alleles of E007 and E001 are mismatched, E007 carried A\\*110x, 310x; B\\*510x, 550x; C\\*030x, 150x; and E001 A\\*020x, 240x; B\\*460x, 540x; C\\*010x, 080x.\n\nFluorescent antibodies for cells staining included human CD3-APC-Cy7 (clone HIT3a), CD8-BV510 (clone SK1), CD3-Percp-Cy5.5 (clone HIT3a), CD62L-PE-Cy7 (clone DREG-56), CD45RA-APC (clone HIT100), CD45RA-BV421 (clone HI100), CD95-PE (clone DX2), CXCR7-PE (clone 10D1-J16), CD28-APC (clone CD28.2), IL-2-APC (clone MQ1-17H12), IFN-\u03b3-PE (clone 4S.B3), TNF-\u03b1-PE (clone MAb11), the above antibodies were purchased from BioLegend, USA. Corresponding isotype for each antibody were used as isotype control. Cells were analyzed on a BD LSR flow cytometer. T cell subsets were determined using fluorescence minus one (FMO) controls for interesting antibody.\n\nAllogeneic co-culture and allo-specific T~SCM~ preparation {#Sec5}\n----------------------------------------------------------\n\nAllogeneic PBLs were co-cultured with the E007 for raising alloreactive T~SCM~ cells. PBLs were stained with 5\u00a0\u03bc\u039c carboxyfluorescein diacetate succinimidyl ester (CFSE, sigma) for 8\u2009min at 37\u2009\u00b0C and washed with RPMI1640 supplemented with 10% FBS three times. E007 were inactivated by irradiating (2.0\u2009Gy). PBLs were mixed with E007 on day 0 at a ratio of 5:1 with 5\u2009\u03bcM TWS119 in the co-culture for the first week. PBLs cultured alone used as control. TWS119 and medium were replaced on day 4, and cells were counted by Trypan blue dye exclusion. As the allo-specific T cells showed proliferation in the co-culture, CFSEdim cells were sorted by FACS with BD FACS AriaII on day 7. Sorted cells were expanded by cytokines IL-7 and IL-15 (Peprotech, USA) of 25\u2009ng/ml each for a further week.\n\nT cell subsets in the co-culture bulks were identified by their surface markers, CD3+ CD8+ CD45RA+ CD62L+ CCR7+ CD95+ CD28+ for T~SCM~, CD3+ CD8+ CD45 RA- CD62L+ for T~CM~, CD3+ CD8+ CD45RA- CD62L- for T~EM~, and CD3+ CD8+ CD45 RA+ CD62L- for T~EF~.\n\nPreparation of allo-specific T~EM~ and T~EF~ for mouse transfer {#Sec6}\n---------------------------------------------------------------\n\nAllo-specific T~EM~ and T~EF~ cells were prepared by a co-culture of allogeneic PBLs with E007 set up as above, but no TWS119 was added. CFSEdim cells were sorted on day 7. Sorted cells were further expanded by 300\u2009U/ml IL-2 (Peprotech, USA) instead of IL-7 and IL-15 for a further week.\n\nIn vitro T~SCM~ cells differentiation assay and intracellular cytokine staining {#Sec7}\n-------------------------------------------------------------------------------\n\nThe prepared T~SCM~ cells were incubated with E007, E001 cells and Dynabeads Human T-Activator CD3/CD28 (\u03b1-CD3/CD28) (Gibco, USA), respectively, in a 5:1 ratio. Cells were collected at 6, 12, and 24\u2009h, and labeled with corresponding fluorescent antibodies for differentiation assay. The T~SCM~ cells incubated with the above stimulations in the presence of 1\u2009\u00d7\u2009BFA (eBioscience, USA) were collected for intracellular IFN-\u03b3, IL-2, and TNF-\u03b1 staining after 4\u2009h incubation. Samples were analyzed on BD Verse flow cytometer. The data analysis was performed with Flowjo software version 10.0 (Tree Star).\n\nQuantitative real-time PCR {#Sec8}\n--------------------------\n\nT cell subsets in the co-culture were sorted by BD Aria II flow cytometer based on the corresponding phenotypes on day 7. Total DNA of the T cells was isolated using a genomic Extraction kit (Takara, Japan), following manufacturer's instructions. TRECs of T cell subsets were detected using fluorescently quantitative PCR kit (Takara, Japan) with TREC-specific primers \\[[@CR34]\\] on Bio-Rad CFX Sequence detection system.\n\nAnimal experiments {#Sec9}\n------------------\n\n6-week-old female NOD-SCID mice were purchased from Vital River Lab Animal Co, Ltd. (Beijing; a distributor of the Jackson Laboratory). Animal experiments in this study were approved by the Ethical Committee of Tongji Medical College. The mice were irradiated with 2.0\u2009Gy and randomized in five groups (*n*\u2009=\u20095), then intravenous injection with 8\u2009\u00d7\u200910^6^ E007 or E001 on day \u2212\u20093. The mice were injected with the 1\u2009\u00d7\u200910^7^ allo-specific T~SCM~ or allo-specific T~EM~ and T~EF~ on day 0. Blood samples were taken every week from mice after the T cell transfer. The number and phenotype of human T cells in mouse peripheral blood were determined by flow cytometry. The signs of GVHD were monitored daily. Mice were euthanized on day 35 and bone marrow and spleen were removed to detect human T cells and residual LCL cells. The LCL cells were detected by the latent membrane protein 1 (LMP1) of EB virus using antibodies LMP1-FITC (primary antibody: mouse anti-EBV LMP1, clone CS1--4, Abcom; secondary antibody: goat anti-mouse IgG-FITC, clone poly 4053, BioLegend) and flow cytometry.\n\nImmunofluorescence analysis for mouse spleen {#Sec10}\n--------------------------------------------\n\nFormaldehyde-fixed spleen specimens were embedded in paraffin and cut into sections for immunofluorescent detection of human CD3+ T cells with antibodies CD3-488 (primary antibody: rabbit anti-human CD3, clone SP7, ThermoFisher; secondary antibody: goat anti-rabbit lgG H&L-Alexa Fluor 488, Abcom) and LMP1-CY3 (secondary antibody: goat anti-mouse lgG H&L-Cy3, Abcom) respectively. Nuclei were stained with DAPI (BioLegend).\n\nStatistical analysis {#Sec11}\n--------------------\n\nThe statistical significance of differences between two groups was assessed with a 2-tailed paired or unpaired *t* test. Comparisons of more than two groups were performed by one-way ANOVA with multiple comparison tests. Data are shown as the mean\u2009\u00b1\u2009standard deviation (SD). Difference were marked as NS, *P*\u2009\\>\u20090.05; \\**P*\u2009\\<\u20090.05; \\*\\**P*\u2009\\<\u20090.01, and \\*\\*\\**P*\u2009\\<\u20090.001. All the data obtained from the study was analyzed using SPSS 22.0 (IBM, USA).\n\nResults {#Sec12}\n=======\n\nOur in vitro protocol is able to prepare allo-specific CD8+ T~SCM~ cells effectively {#Sec13}\n------------------------------------------------------------------------------------\n\nTo prepare allo-specific T~SCM~, this study began with a co-culture of a simulator cells and allogeneic PBLs on day 0. The stimulator was an LCL named as E007 with defined HLA allotyping. Due to the difference in HLA alleles among random donors, allo-specific T~SCM~ cells were generated from T~N~ through proliferation during the co-culturing (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). The T~SCM~ were enriched in the presence of differentiation inhibitor TWS119, of which the optimal concentration was 5\u2009\u03bcM in the allogeneic co-culture (Additional\u00a0file\u00a0[1](#MOESM1){ref-type=\"media\"}: Figure S1A B). On day 7, the T~SCM~ cells in the co-culture bulks were defined by the phenotype CD3+ CD8+ CD45RA+ CD62L+ CD95+ CCR7+ CD28+ (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}). The inhibition of differentiation in the allogeneic co-culture enriched the T~SCM~ numbers by 100 folds on day 7 (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}). As the allo-specific T cells showed proliferation in the co-culture, sorting of proliferation cells ensured the antigen-specificity of the prepared T~SCM~ cells. The sorting reached above 98% purity of the proliferative cells (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). It would be rational that the prepared T~SCM~ cells were E007 specific. After sorting, the cells were cultured in the presence of IL-7 and IL-15 (25\u2009ng/ml each) for the next 7\u2009days. The allo-specific T~SCM~ increased by another 150 folds on day 14 (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}). The lymphocyte distribution at this time not only showed the cultural bulks were mainly CD8+ T~SCM~ (60.1\u2009\u00b1\u200911.2%), but also contained a few CD4+ T~SCM~ (10.4\u2009\u00b1\u20098.16%), CD3- cells (6.15\u2009\u00b1\u20095.23%), CD8+ non-T~SCM~(12.6\u2009\u00b1\u20093.48%), and CD4+ non-T~SCM~ (10.2\u2009\u00b1\u20098.66%) cells (Additional\u00a0file\u00a0[1](#MOESM1){ref-type=\"media\"}: Figure S1D E F). By allogeneic activating, inhibiting differentiation with TWS119, sorting CFSEdim cells, and expansion with IL-7 and IL-15, our in vitro protocol was able to prepare about 2\u2009\u00d7\u200910^7^ allo-specific CD8+ T~SCM~ cells from 1\u2009\u00d7\u200910^7^ PBLs. The number of the T~SCM~ cells was sufficient to meet the needs of the following studies.Fig. 1Allo-specific T~SCM~ cells are effectively prepared in vitro. **a** Protocol for preparation of allo-specific T~SCM~. The method began with setting up an allo-reactive co-culture by mixing of 2\u2009\u00d7\u200910^6^ E007 cells and 1\u2009\u00d7\u200910^7^ allogeneic PBLs (allo-PBLs) on day 0 (allogeneic activation), to generate allo-specific T cells. The presence of 5\u2009\u03bcM TWS119 enriched T~SCM~ in the co-culture (differentiation inhibition). The CFSE diluted cells in the co-culture bulks were sorted by FACS on day 7 (proliferation sorting). The sorted cells were cultured with IL-7 and IL-15 (25\u2009ng/ml each) to expand the T~SCM~ for the next 7\u2009days (cytokine expansion). Then, the cultural bulks were used as the prepared T~SCM~ cells for the tests in vitro and in vivo included in this study. **b** Gating strategy for identification of CD8+ T~SCM~ subset in the co-culture bulks. Lymphocytes were identified based on SSC versus FSC, and proliferating cells were on a reduced fluorescence intensity of CFSE (CFSEdim). In the gated CFSEdim population, cells expressing CD3 and CD8 were selected for further CD45RA and CD62L analysis. CD8+ T~SCM~ showed the phenotype of CFSEdim CD3+ CD8+ CD45RA+ CD62L+, T~CM~ of CD45RA- CD62L+, T~EM~ of CD45RA- CD62L-, T~EF~ of CD45RA+ CD62L-, and CD95 was further detected to distinguish T~SCM~ from T~N~, the latter showed the phenotype of CFSE high CD95- in the co-culture bulks, but T~SCM~ of CFSEdim CD95+. The illustration was a representative of co-culture bulks analyzed by FCM on day 7 before sorting, the white peaks in CD95, CCR7, and CD28 plots were isotype controls. The broken arrows indicated the sequential gating strategy. **c** T~SCM~ numbers increased by 100 folds in the co-culture bulks with TWS119 enrichment for 7\u2009days. Data are represented as mean\u2009\u00b1\u2009SD of six individual experiments. **d** E007-specific T cells underwent proliferation in the co-culture, the CFSEdim cells were sorted by FACS with purity above 98%. **e** The sorted cells were further expanded by IL-7 and IL-15 for the next 7\u2009days, the allo-specific T~SCM~ increased by another 150 folds. Data are represented as mean\u2009\u00b1\u2009SD of six individual experiments\n\nThe T~SCM~ preparation strategy used in this study could be translated into a single epitope-specific T~SCM~ preparation. When the E007 co-cultured with allogeneic PBLs, the precursor frequencies were much higher than those of T cell responses to a single allogeneic epitope. To prepare a single epitope-specific T~SCM~ cells in a large number, a modification with a prolonged cytokine expansion was required (Additional\u00a0file\u00a0[2](#MOESM2){ref-type=\"media\"}: Figure S2).\n\nThe prepared T~SCM~ cells show stem cell and memory T cell characteristics in vitro {#Sec14}\n-----------------------------------------------------------------------------------\n\nTo examine the self-renewal capacity, T~N~, T~SCM~, T~CM~, and T~EM~ cells were sorted by their corresponding phenotypes from the co-culture on day 7 (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}). The content of T cell receptor rearrangement excision circles (TRECs) in each subset was examined by real-time qPCR. As TRECs cannot be replicated while cell division, the content of TRECs reflects the frequency of T cell proliferation. Results showed that T~N~ had the highest content of TRECs. T~SCM~, T~CM~, and T~EM~ cells possessed less TRECs, confirming they were differentiated from T~N~ through T cell clonal proliferation (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). The content of TRECs in T~SCM~ cells was between those of T~N~ and T~CM~ cells, suggesting T~SCM~ was at the earliest stage after T~N~ activation.Fig. 2The prepared T~SCM~ cells bear the ability of self-renewing. **a** T~N~, T~SCM~, T~CM~, and T~EM~ cell subsets in CD8+ T cells were sorted by their corresponding phenotypes from the co-culture on day 7. TREC copy number in the sorted subsets relative to T~N~ cells was shown. Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments. **b** The prepared T~SCM~ cells were relabeled with CFSE and exposed to 200\u2009U/ml of IL-2 for 10\u2009days, the proliferating daughter cells divided into 7 generations (left panel). In each generation, there were a portion of CD8+ T cells expressing CD62L and CD45RA (right panel). Representative FCM plots were shown. **c** The frequency of T~SCM~ (CD45RA+ CD62L+) cells in CD8+ T cells was analyzed at each CFSE dilution peak. Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments. (\\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01, and NS, *p*\u2009\\>\u20090.05)\n\nThe prepared T~SCM~ cells were relabeled with CFSE and exposed to IL-2 (200\u2009U/ml) for 10\u2009days. As IL-2 is a potential promoter for T cell proliferation and differentiation, seven generations of CFSE-diluted daughter cells developed. A portion of T~SCM~ cells (CD45RA+ CD62L+) were found in each generation, indicating that the prepared T~SCM~ cells were able to maintain a stem cell phenotype during proliferation and differentiation (Fig.\u00a0[2b, c](#Fig2){ref-type=\"fig\"}). Collectively, the prepared T~SCM~ cells showed characteristics of self-renewal.\n\nTo check memory T cell characteristics, the prepared T~SCM~ cells were incubated with the cognate stimulator E007, HLA class I mismatched E001, and \u03b1-CD3/CD28 beads. Phenotypic analysis revealed that the E007-restimulated T~SCM~ cells showed vigorous differentiation in 6\u2009h, which became more intensive with prolonged stimulation time. A similar response was observed in the \u03b1-CD3/CD28-stimulated T~SCM~ cells. While the E001-stimulated T~SCM~ cells showed no significant differentiation after 24\u2009h, the response was similar to that of non-stimulated T~SCM~ cells (Fig.\u00a0[3a, b](#Fig3){ref-type=\"fig\"}). In addition, after 24\u2009h of incubation with E007 and \u03b1-CD3/CD28, we found that T~SCM~ cells in the incubation showed no significant change in absolute number (Additional\u00a0file\u00a0[3](#MOESM3){ref-type=\"media\"}: Figure S3E)~,~ suggesting that T~SCM~ cells were able to self-renew during proliferation and differentiation.Fig. 3E007 cells stimulate the T~SCM~ differentiate rapidly and the daughter effectors produce cytokines. The prepared T~SCM~ were stimulated with E007, E001, and \u03b1-CD3/CD28, respectively. **a**, **b** The T~SCM~ differentiated into effector T cells upon the stimulation with E007. CD3+ CD8+ T cells expressing CD45RA and CD62L (T~SCM~ cells) were detected at indicated hours after stimulation. Representative FCM plots were shown (**a**). Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments (**b**). **c**--**f** Effector T cells differentiated from the T~SCM~ produced cytokines. Intracellular IL-2, TNF-\u03b1, and IFN-\u03b3 production of the T cell subsets were detected. Gating by CD62L, the T cells were divided into CD62L+ (T~SCM~ and T~CM~) and CD62L- (T~EM~ and T~EF~) cells, the T~EM~ and T~EF~ subsets showed the cytokine positively stained cells. Representative FCM plots (the percentages indicated the cytokine producing cells in CD62L+ and CD62L- cells, respectively) (**c**) and the frequencies of IFN-\u03b3 (**d**), TNF-\u03b1 (**e**) and IL-2 (**f**) positive cells in corresponding CD8+ T cell subsets were shown. Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments (\\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001, and NS, *p*\u2009\\>\u20090.05)\n\nCD8+ T cell mediated immune effectors include production of cytokines and killing of target cells. To investigate production of cytokines, the T~SCM~ cells were stimulated with E007, E001, and \u03b1-CD3/CD28 beads, then intracellular IL-2, TNF-\u03b1, and IFN-\u03b3 production was measured. The E007-restimulated T~SCM~ cells showed more positive cells of these cytokines. Gating with CD62L, the incubated bulks were divided into CD62L+ (T~SCM~ and T~CM~) and CD62L- (T~EM~ and T~EF~) cells, most of the cytokine positive cells were T~EM~ and T~EF~ cells (Fig.\u00a0[3c](#Fig3){ref-type=\"fig\"}). Meanwhile, gating with CD45RA, we found that both T~EM~ and T~EF~ cells showed similar frequency of the cytokine positive cells (Additional\u00a0file\u00a0[3](#MOESM3){ref-type=\"media\"}: Figure S3A-D). Whereas, the T~SCM~ cells in the incubation produced almost background IL-2, and only a low level of TNF-\u03b1 and IFN-\u03b3. The T~SCM~ cells responded against \u03b1-CD3/CD28 in a similar profile of the cytokine production to that against E007. In contrast, the T~SCM~ cells stimulated with E001 showed frequency of cytokine-positive cells in the same way as that with non-stimulation (Fig.\u00a0[3c--f](#Fig3){ref-type=\"fig\"}). Next, we moved to cytotoxicity assay, the T~SCM~ co-cultured with E007 or E001 for 24\u2009h at a ratio 5:1. Killing rate of the T~SCM~ and daughter cells against E007 was higher than that against E001 after 8\u2009h (Additional\u00a0file\u00a0[3](#MOESM3){ref-type=\"media\"}: Figure S3F, G). These results showed that the prepared T~SCM~ cells were E007-specific and able to differentiate rapidly into effector T cells after stimulated by the same antigen.\n\nThe prepared T~SCM~ cells are able to implant and effectively remove target cells in vivo {#Sec15}\n-----------------------------------------------------------------------------------------\n\nTo evaluate capacity of persistence and rejecting target cells in vivo of the prepared T~SCM~ cells, human LCL cells E001 or E007 were inoculated intravenously into NOD/SCID mice to establish LCL-burden mouse model. After 3\u2009days, the LCL-burden mice were treated with either E007-specific T~SCM~ cells or E007-specific T~EM~ and T~EF~ cells (Fig.\u00a0[4a, b](#Fig4){ref-type=\"fig\"}). On days 7, 14, 21, 28, and 35 after T cell infusion, caudal vein peripheral blood samples were taken to detect the frequency and phenotype of human T cells. We found human CD3+ CD8+ T cells in the peripheral blood of the T~SCM~-treated mice at all sampling times. However, in the T~EM\u2009+\u2009EF~-treated mice (E007-T~EM\u2009+\u2009EF~), human T cells were detected in peripheral blood on day 7, but not on day 14 or later (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}). In the T~SCM~-treated mice, the frequency of T~SCM~ cells in the E007-burden mice (E007-T~SCM~) were lower than that in the E001-burden mice (E001-T~SCM~), but no difference was found in the absolute number of T~SCM~ cells between E007-burden and E001-burden mice (Fig.\u00a0[4d, e](#Fig4){ref-type=\"fig\"}). On day 35, more frequent human T cells were found in the spleen and bone marrow of the T~SCM~-treated mice revealed by flow cytometry (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}), and the immunofluorescent staining of the spleen sections showed the similar results (Fig.\u00a0[6d, e](#Fig6){ref-type=\"fig\"}). These findings indicated that the T~SCM~ cells were able to survive long through self-renewal in the mice, and differentiate into other T cell subsets as stimulated with the specific antigens of E007.Fig. 4The T~SCM~ are transferred into LCL-burden mice for stem and memory T cell behavior. **a b** Adoptive transfer protocol of the T~SCM~. On day \u2212\u20093, irradiated mice were inoculated with E007 or E001 cells to set up the LCL-burden mouse model. On day 0, E007-specific T~SCM~ or T~EM~ and T~EF~ (T~EM\u2009+\u2009EF~) were transferred into the mice intravenously (iv). Blood sampling was taken once a week for 5\u2009weeks after the T cell treatment. On day 35, the mice were sacrificed for detection of human T cells and LCL cells in spleen and bone marrow. **c**--**e** The prepared T~SCM~ cells had superior persistence in vivo. The frequency of human CD3\u2009+\u2009CD8+ T cells (**c**), the frequency of T~SCM~ cells (CD45RA+ CD62L+) in human CD3+ CD8+ T cells (**d**), and the absolute number of T~SCM~ cells in 50\u2009\u03bcl blood (**e**) were detected by FCM. Data are represented as mean\u2009\u00b1\u2009SD of five mice. (\\**p*\u2009\\<\u20090.05; NS, *p*\u2009\\>\u20090.05) Fig. 5The prepared T~SCM~ cells are able to reconstitute T cell subsets in vivo. The mice were euthanized on day 35. **a** Human CD3\u2009+\u2009CD8+ T cells were detected by FCM in peripheral blood, spleen, and bone marrow, respectively. The T~SCM~-treated mice (E007-T~SCM~ and E001-T~SCM~) showed more human CD3\u2009+\u2009CD8+ T cells. **b** The distribution of human CD3\u2009+\u2009CD8+ T cells subsets (T~SCM~,T~CM~, T~EM~,T~EF~) was detected by FCM in peripheral blood, spleen, and bone marrow, respectively. More effector T cell subsets (T~EM~ and T~EF~) were developed in the E007-T~SCM~ mice. Data are represented as mean\u2009\u00b1\u2009SD of five mice. (\\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01, and \\*\\*\\**p*\u2009\\<\u20090.001) Fig. 6The T~SCM~ cells are able to eradicate allo-antigen-specific targets in vivo. The spleen and bone marrow were detected for the residual LCL cells by intracellular LMP1 staining using FCM. **a**--**c** Representative FCM plots (**a**) and the absolute number of LCLs in bone marrow of two femurs (**b**) and in 100\u2009mg spleen tissue (**c**) were analyzed. **d** Spleen sections were observed for LCLs and human T cells by immunofluorescence. Representative illustrations of immunofluorescent labeling of human CD3 (green), LMP1 (red), and nucleus staining by DAPI (blue) for spleen sections were shown (\u00d7\u2009400). **e** Date showed the mean frequencies of human CD3 and LMP-1-positive cells in 4 fields chosen randomly (100 cells each) for each mouse spleen section. The residual LCLs in the E007-T~SCM~ mice were similar to that in the control mice without LCL inoculation, the rest mouse groups showed relatively more LCL cells. Data are represented as mean\u2009\u00b1\u2009SD of five mice (\\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001, and NS, *p*\u2009\\>\u20090.05)\n\nTo investigate the immune reconstitution capacity of the T~SCM~ cells, peripheral blood, spleen, and bone marrow of T~SCM~-treated and T~EM\u2009+\u2009EF~-treated mice were tested for human T cells on day 35. The human T cells in the above specimens consisted of T~SCM~ (CD45RA\u2009+\u2009CD62L+), T~CM~ (CD45RA-CD62L+), T~EM~ (CD45RA-CD62L-), and T~EF~ (CD45RA+ CD62L-). The distribution of T cell subsets revealed more T~EM~ and T~EF~ cells in the E007-burden mice, whereas more T~SCM~ and T~CM~ cells in the E001-burden mice (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). As T~EM~ and T~EF~ cells are at terminal differentiation stage while T~SCM~ and T~CM~ cells at early differentiation stage, our findings suggested the E007-specific T~SCM~ cells differentiated into the other T cell subsets when they came cross the same antigen in the E007 burden mice.\n\nTo analyze killing efficacy of the T~SCM~ cells in vivo, on day 35 of the T cell treatment, the residual LCL cells in mice were detected by intracellular LMP1 staining. The LCL cells were mainly found in the spleen and bone marrow of the LCL-burden mice. In the E007-burden mice, the residual LCL cells in spleen and bone marrow of the T~SCM~-treated mice were similar to that of the control mice without LCL infusion (control). Nevertheless, the T~EM\u2009+\u2009EF~-treated mice carried a significant amount of the residual LCL cells compared to the control mice, but the amount was lower than the E007-burden mice without T cell treatment (E007) (Fig.\u00a0[6a--c](#Fig6){ref-type=\"fig\"}). More residual LCL cells were also found in the immunofluorescent spleen sections of the T~EM\u2009+\u2009EF~-treated mice than those of the T~SCM~-treated mice (Fig.\u00a0[6d, e](#Fig6){ref-type=\"fig\"}) These data demonstrated the T~SCM~ cells equipped with superior capacity of killing targets of specific antigens in vivo compared to the T~EM~ and T~EF~ cells.\n\nIn contrast to the E007-burden mice, the E001-burden mice bore not only less T~EM~ and T~EF~ cells after the T~SCM~ treatment, but also the residual LCL cells as high as that of the E007-burden mice without T cell treatment (Fig.\u00a0[6a--c](#Fig6){ref-type=\"fig\"}). Results of immunofluorescence staining of spleen sections for human T cells and LCLs showed residual LCL cells in E007-burden mice treated with T~SCM~ were similar to those of the control mice. Although the spleen of the E001-burden mice showed a significant number of human T cells, the residual LCL cells were similar to that of the E007-burden mice without T cell treatment (Fig.\u00a0[6d, e](#Fig6){ref-type=\"fig\"}). By the way, the human T cell-treated mice showed no sign of xenogeneic graft-versus-host disease (xeno-GVHD), such as loss of body weight, back arched, and shed. Collectively, these results indicated that the prepared T~SCM~ cells were able to survive over a long-term in vivo and differentiate into effector T cells to eradicate the target bearing the cognate antigens.\n\nDiscussion {#Sec16}\n==========\n\nMature T cells are comprised of cells that are at various stages of differentiation, which are discernible by the expression of surface molecules. T~N~ cells are conventionally defined by the co-expression of the CD45RA, CCR7, and the lymph node homing molecules L-selectin (CD62L) \\[[@CR3]\\]. Similar to T~N~ cells, CD62L and CCR7 are maintained on T~CM~ cells, whereas these molecules are lost on more differentiated T~EM~ cells \\[[@CR3], [@CR35]\\]. T~SCM~ cells, the least differentiated of all distinct memory populations, are identifiable through the expression of markers, including CD62L, CCR7, CD95, and the chemokine (C-X-C motif) receptor 3 \\[[@CR5], [@CR10], [@CR36]\\]. Na\u00efve T cells downregulate CD62L expression after stimulation, as clonal expansion goes on, both CD62L^LOW^ and CD62L^HIGH^ T cell subsets are developed. With real-time tracking of CD8 T-cell divisions in vitro, Kinjyo and co-workers define memory T cells among CD62L^HIGH^ cells. The memory T cells stay CD62L^HIGH^ and proliferate in vitro driven by IL-2, although the division duration is slow, the phenotype and cell cycle duration are inherited by the progeny of the T cells \\[[@CR37]\\]. We identified CD8+ T~SCM~ subset in the co-culture bulks by the phenotypes CD3+ CD8+ CD45RA+ CD62L+ CCR7+ CD95+ CD28+, as our results showed the cells with the phenotypes in the co-culture bulks were of stem cell and memory T cell properties.\n\nA frequently used method to enrich T~SCM~ cells is differentiation inhibition. With antigen priming, antigen-specific CD8+ T~N~ cells proliferate and develop into T~SCM~ cells first and then the other subsets. It is reported that T~SCM~ differentiation can be inhibited by TWS119, which inhibits the GSK-3\u03b2 and activate Wnt/\u03b2-catenin. The inhibition improves the maintenance of 'stemness' in mature memory CD8+ T cells which mediated a better anti-tumor response after transferred into mice \\[[@CR15], [@CR38]\\]. Rapamycin is also reported to help the formation of T cell subset at an early stage of differentiation by modulation of mTOR signaling \\[[@CR39]--[@CR41]\\]. However, rapamycin tended to promote the generation of T~CM~ cells in our study (Additional\u00a0file\u00a0[1](#MOESM1){ref-type=\"media\"}: Figure S1C), in line with other report \\[[@CR41]\\]. In our co-culture of E007 with allogeneic PBLs, the E007-specific T~SCM~ cells were generated from T~N~ and enriched in the presence of TWS119 which inhibited the T~SCM~ further differentiation. As inhibition by TWS119 results in a T~SCM~ phenotype in human CD8+ T cells \\[[@CR5], [@CR15], [@CR16]\\], our protocol preferred expansion of CD8+ T~SCM~ to that of CD4+ T~SCM~.\n\nThe administration of cytokines is another method to expand T~SCM~. Cytokines have important functions related to T cell expansion, differentiation, survival, and homeostasis. Common \u03b3-chain (\u03b3c) family cytokines are commonly used in clinical trials, including IL-2, IL-7, IL-15, and IL-21. IL-7 is instrumental for the generation of T~SCM~ cells by binding to IL-7 receptor expressing naive and memory T lymphocytes \\[[@CR42], [@CR43]\\]. Expansion of T~SCM~ required either IL-15 or IL-2; IL-15 proves superior to IL-2 in supporting expansion coupled to preservation of the T~SCM~ phenotype \\[[@CR18], [@CR19]\\]. Activated naive T cells show a higher expression of IL-21 receptor, and IL-21 has been reported to be able to enrich less differentiated T~SCM~ within the total T~SCM~ subset \\[[@CR20], [@CR22], [@CR23], [@CR44]\\]. However, IL-21 exerted few effects on the T~SCM~ expansion in our study. Therefore, IL-7 and IL-15 were used to expand the T~SCM~ cells after the proliferation sorting, resulting in a large number of the T~SCM~ sufficient for the experiments in vitro and in vivo included in our study.\n\nAn important feature of stem cells is the self-renewal capacity. Although self-renewal abilities of the prepared T~SCM~ cells were revealed by TREC copy numbers and proliferative history driven by IL-2 in vitro, long-term in vivo implantation is the gold standard for identification of stem cells. Long-lasting survival of human T~SCM~ cells in mice is reported. In serial transplantations model, T~SCM~ cells prove able to persist in secondary recipients suggesting self-renewal abilities \\[[@CR5], [@CR6], [@CR19], [@CR45]\\]. Here, the prepared T~SCM~ migrated to secondary lymphoid organs, such as bone marrow and spleen, showed long-lasting survival potential at least for 35\u2009days after transferred in our model. Importantly, the number of T~SCM~ cells in the blood samples was observed consistent during this study, suggesting the self-renewal capacity of the transferred T~SCM~ cells in vivo. The implantation and long-term existence of the T~SCM~ cells in the host would be expected to mount durable T cell responses after transferring.\n\nGoverning antigen specificity is important for T~SCM~ preparation; various strategies are used for this purpose. For example, HCMV-specific T~SCM~ cells are isolated from peripheral blood T~SCM~ cells of HCMV seropositive donors and enriched in vitro by incubation with HCMV antigen \\[[@CR26]\\]. CD19-specific T~SCM~ cells are prepared from peripheral blood T~N~ and T~SCM~ by transduction with a \u03b3-retroviral vector encoding the CD19-chimeric antigen receptor (CAR) \\[[@CR24]\\]. TCR-transgenic mice are also a source of antigen-specific T~SCM~ cells \\[[@CR46]\\]. In our approach, the E007-specific T cells underwent proliferation in the co-culture bulks, making it feasible to be sorted by CFSE dilution. In our pre-experiment, IL-7 or IL-15 was able to make PBLs proliferation without the E007 stimulation. Hence, no cytokine was added to the co-culture before the sorting, to ensure the E007-specificity of the prepared T~SCM~ cells.\n\nTwo HLA class I mismatched LCLs, E007 and E001, were used to examine the antigen specificity of the prepared CD8+ T~SCM~ cells. Rapid differentiation into effector T cells and robust effector functions were observed when the prepared T~SCM~ cells were stimulated with the E007, but not with the E001. Although the E007-specificity was suggested by these in vitro findings, it would be more important to observe the behavior of the prepared T~SCM~ cells in vivo. In our LCL burden mouse model, LCL cells could be found in spleen and bone marrow up to 45\u2009days after transfer into the immunodeficiency mice in our preliminary test. The LCL burden mice showed no measurable side effect during our study, such as loss of body weight, back arched, and shed. Interestingly, the prepared T~SCM~ cells exhibited intensive differentiation into other T cell subsets and effective eradication of LCL targets in the E007-burden mice instead of the E001-burden mice. These observations reflected the prepared T~SCM~ cells were E007-specific.\n\nConclusions {#Sec17}\n===========\n\nFunctional allo-specific CD8+ T~SCM~ cells were prepared from human PBLs in a procedure of allogeneic co-culture, differentiation inhibition, proliferation sorting, and cytokine expansion. Although our study provided a practical protocol for allo-specific T~SCM~ cell preparation, this method would be adapted to prepare T~SCM~ cells specific for antigen of interesting. As T~SCM~ cells show the implantation and long-term existence in the host after transferring, the preparation of antigen-specific T~SCM~ cells is crucial for T cell adoptive immunotherapy.\n\nAdditional files\n================\n\n {#Sec18}\n\nAdditional file 1:Figure S1. Enrichment of T~SCM~ cells by differentiation inhibitors and the lymphocyte distribution changes in the cultural bulks over the T~SCM~ preparation course. **A**, **B** TWS119 exhibited concentration-dependent enrichment of T~SCM~ in the alloreactive co-culture, 5\u2009\u03bcM of TWS119 was the best concentration. T~SCM~ cells in frequency (**A**) and in absolute number (**B**) were shown. **C** Differentiation inhibition by rapamycin was more likely to enrich T~CM~ instead of T~SCM~. **D**--**F** After allogeneic activation, differentiation inhibition, proliferation sorting, and cytokine expansion, lymphocyte distribution in the cultural bulks was revealed by FCM over the T~SCM~ preparation course. CD3\u2009+\u2009CD4+ T cells, CD3\u2009+\u2009CD8\u2009+\u2009T cells and CD3- cells in the cultural bulks were detected, and the proportion of CD3\u2009+\u2009CD8+ T cells was increased over time (**D**). The majority of cultural bulks were CD3\u2009+\u2009CD8+ T~SCM~ (60.1\u2009\u00b1\u200911.2%) on day 14 (**E**). The proportion of CD3\u2009+\u2009CD8+ T~SCM~ in CD8+ T cells was increased over time, about 80% of CD8+ T cells were T~SCM~ on day 14 (**F**). Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments. (JPG 170\u00a0kb) Additional file 2:Figure S2. Our T~SCM~ preparation strategy can be used to generate a single epitope-specific T~SCM~ cells. **A** Procedure for preparation of AFP-specific, allogeneic T~SCM~. T2 cells express only empty HLA-A2 allele and no other HLA allele. Alpha fetoprotein (AFP) is a tumor associated antigen of hepatocarcinoma, the hAFP~158--166~ (FMNKFIYEI) is an HLA-A2 restricted peptide. When pulsed with the AFP peptide, T2 cells were able to present the AFP/HLA-A2 complex. The AFP-specific T~SCM~ were raised by co-culturing HLA-A2 negative (HLA-A2-ve) PBLs and the T2 cells pulsed with the AFP peptide (T2/AFP). In a procedure of an allogeneic co-culture, differentiation inhibition, proliferation sorting and cytokine expansion, the AFP/HLA-A2 complex-specific T~SCM~ cells were produced. **B** Co-culture by mixing of 1\u2009\u00d7\u200910^7^ PBLs and 2\u2009\u00d7\u200910^6^ T2/AFP on day 0, a prolonged cytokine expansion was required to generate 1\u2009\u00d7\u200910^6^ AFP-specific T cells. Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments. **C**--**E** The prepared T~SCM~ cells were AFP-specific. The prepared T~SCM~ cells were incubated with the T2/AFP and T2 cells pulsed with an irrelevant peptide HBcAg~18--27~ (T2/HBC), respectively. After 4\u2009h incubation, the T cell subsets and their intracellular IFN-\u03b3 production were detected. Representative FCM plots (**C**). T~SCM~ cells differentiated more when incubated with the T2/AFP (**D**). The daughter cells showed more frequent IFN-\u03b3 positive cells when incubated with the T2/AFP (**E**). Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments (\\*\\* *p*\u2009\\<\u20090.01). (JPG 404\u00a0kb) Additional file 3:Figure S3. The prepared T~SCM~ differentiated into effector T cells stimulated by E007. The prepared T~SCM~ were stimulated with E007, E001 and \u03b1-CD3/CD28, respectively. The T cell subsets and their intracellular IL-2, TNF-\u03b1 and IFN-\u03b3 production were detected. The T~SCM~ differentiated into effector T cells upon the stimulation with E007 and \u03b1-CD3/CD28. **A**--**D** After 4-h stimulation, both T~EM~ and T~EF~ cells exhibited the similar frequency of the cytokine positive cells. Representative FCM plots (**A**). Gating by CD3+ CD8+ CD62L-, the T cells were divided into CD45RA- (T~EM~) and CD45RA+ (T~EF~) cells. The T~EM~ and T~EF~ subsets showed the cytokine positively stained cells of IFN-\u03b3 (**B**), TNF-\u03b1 (**C**) and IL-2 (**D**). **E** After 24-h stimulation, the absolute number of T~SCM~ remained stable during differentiation. **F**, **G** The cytotoxicity of the T~SCM~ and the daughter cells were E007-specific. E007 and E001 labeled with celltrace, co-cultured with the T~SCM~ at ratio 1:5. Dead cells stained by PI dye were detected by FCM. Representative FCM plots (**F**) and the frequencies of dead cells (**G**) were shown. Data are represented as mean\u2009\u00b1\u2009SD of four individual experiments (NS, *p*\u2009\\>\u20090.05; \\*\\**p*\u2009\\<\u20090.01, and \\*\\*\\**p*\u2009\\<\u20090.001). (JPG 722\u00a0kb)\n\nACT\n\n:   Adoptive cell therapy\n\nGVHD\n\n:   Graft versus host disease\n\nHLA\n\n:   Human leukocyte antigen\n\nLCL\n\n:   B lymphoblastoid cell lines\n\nLMP1\n\n:   Latent membrane protein 1\n\nPBL\n\n:   Peripheral blood lymphocytes\n\nPBMC\n\n:   Peripheral blood mononuclear cells\n\npMHC\n\n:   Peptide major histocompatibility complex\n\nTCR\n\n:   T cell receptor\n\nTREC\n\n:   T cell receptor rearrangement excision circle\n\nThe authors apologize to colleagues whose work could not be cited because of space limitation. They thank Dr. Jinghui Zhang (The Union Hospital, Tongji Medical College, Huazhong University of Science and Technology) for the FACS sorting in this study.\n\nFunding {#FPar1}\n=======\n\nThis project was supported by the National Natural Science Foundation of China (no. 31370885).\n\nAvailability of data and materials {#FPar2}\n==================================\n\nPlease contact author for data requests.\n\nXWW designed the study, analyzed the data, and wrote the manuscript. LG was a major contributor in the experimental performance, analyzed the data, and wrote the manuscript. XL performed and analyzed the cell culture experiments. JW contributed to the animal experimental performance. XFW contributed to the design of study. ZL contributed to the data analysis. JY contributed to the manuscript preparation. All authors read and approved the final manuscript.\n\nEthics approval {#FPar3}\n===============\n\nAll donors' samples were obtained after informed consent according to a protocol approved by the Ethics Committee of Tongji Medical College, Wuhan, China. Animal experiments were approved by the Ethics Committee of Tongji Medical College. All animal procedures were performed in strict accordance with the guidelines of the Chinese Council on Animal Care.\n\nConsent for publication {#FPar4}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar5}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar6}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nMature blood cell lineages originate from a pool of self-renewing hematopoietic stem cells (HSCs) and are an attractive source for stem-cell-based therapies like hematopoietic stem cell transplantation (HSCT) that offer a potential cure for various malignant (leukemia, lymphoma, and myeloma) and non-malignant (aplastic anemia) hematologic disorders. Currently, bone marrow (BM), umbilical cord blood (UCB), and peripheral blood from G-CSF (granulocyte-colony stimulating factor) treated donors are the major sources of stem cells for transplantation, and peripheral blood stem cell transplantation (PBSCT) is the most common and widely used procedure in the clinical setting^[@CR1]^. However, issues regarding the yield of transplantable HSCs still prevail, especially in the context of UCB transplantation^[@CR2]^, despite the recent increase in the number of suitable donors and the success of haploidentical HSCT^[@CR3]^. As low HSC numbers at transplantation have been associated with greater incidence of graft failure, delayed hematopoietic recovery, slow immune reconstitution, and early mortality, even in PBSCT recipients^[@CR4]^, protocols that facilitate the *ex vivo* expansion of HSCs represent an important step to overcome these limitations. Further, efficient *ex vivo* expansion of genetically modified HSCs, obtained using novel gene editing techniques, can potentially be applied in patients with inborn genetic diseases (e.g. hemoglobinopathies)^[@CR5]^.\n\nNumerous attempts have been made to identify conditions and/or chemicals that allow the *ex vivo* expansion of functional hematopoietic stem and progenitor cells (HSPC), including cytokine cocktails, feeder layer of mesenchymal stromal cells (MSCs), and proteins or chemicals (e.g. notch ligand, aryl hydrocarbon receptor antagonists, PGE~2~, all-trans retinoic acid, and other epigenetic modulators)^[@CR6]--[@CR13]^. However, amplification of HSPCs is not necessarily associated with preservation of HSPC function as some studies have reported loss of self-renewal capabilities, measured by the long-term repopulating capacity of these cells^[@CR14]^. While certain procedures for *ex vivo* expansion have been shown to retain HSC function and clinical trials have attested to the feasibility of this approach^[@CR15]^, successful hematopoietic recovery after HSC transplantation not only relies on self-renewal and differentiation capacity but also on homing to the bone marrow and subsequent lodging in hematopoietic stem cell niches^[@CR16]^. Such migration and lodging of HSCs in specific niches are tightly regulated processes that are controlled by the expression and function of various molecules, including integrins (VLA-4, VLA-5, and LFA-1), selectins (P- and E-selectin), and certain chemokines (SDF-1)^[@CR17]^.\n\nIn this study transgenic *c-myb:EGFP* zebrafish were used to screen and identify small molecules that modulated HSPC activity^[@CR18]^. Histone deacetylase inhibitors (HDACIs), namely, valproic acid (VPA), resminostat, and entinostat, significantly increased HSPC numbers, and their functional relevance was validated by analyzing runx1^+^ expression in the zebrafish embryos. HDACIs also produced similar effects in human HSPCs as human CD34^+^ cells could be extensively expanded *in vitro* using various HDACIs, especially, VPA. *In vivo*, even though VPA-expanded CD34^+^ HSPCs displayed impaired homing to the bone marrow of immunodeficient mice that resulted in reduced short-term engraftment as monitored by peripheral blood donor chimerism, they however, retained their long-term engraftment potential and maintained their differentiation ability both *in vitro* and *in vivo*. These results imply that HDACIs, particularly VPA, can be used for the *in vitro* expansion of G-CSF mobilized hHSPCs, but their use in clinical transplantation protocols should consider impaired homing and lower short-term-engraftment.\n\nResults {#Sec2}\n=======\n\nHDACIs increase c-myb^+^ HSPC number and *runx1* expression in zebrafish embryos {#Sec3}\n--------------------------------------------------------------------------------\n\nA recently developed semi-automated imaging assay^[@CR18]^ was used on transgenic zebrafish embryos expressing *c-myb:EGFP* in HSPCs to screen 550 compounds and identify small molecules that modulate HSPC activity. In zebrafish hematopoiesis, long-term HSCs occur in the aorta-gonad-mesonephros (AGM) at approximately 30\u2009hours post fertilization (hpf) and migrate to the caudal hematopoietic tissue (CHT) region, colonize the thymus, and finally translocate to the kidney marrow, which is the equivalent of mammalian bone marrow^[@CR19]^ (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). *c-myb* is expressed in the cells of the AGM and CHT regions in zebrafish during hematopoiesis^[@CR20],[@CR21]^. In the assay, embryos were exposed to compounds at concentrations of 20 or 40\u2009\u00b5M and between 12 and 36 hpf. Compared to DMSO-treated controls, three HDACIs, namely valproic acid (VPA), resminostat, and entinostat, significantly increased the number of c-myb^+^ cells in the AGM and CHT regions (DMSO 93\u2009\u00b1\u20094, VPA 137\u2009\u00b1\u200922, resminostat 194\u2009\u00b1\u200929, entinostat 150\u2009\u00b1\u200919, p\u2009\\<\u20090.001 for all; Fig.\u00a0[1b and c](#Fig1){ref-type=\"fig\"}). These observations were validated by whole-mount *in situ* hybridization (WISH) for *runx1* on wild type embryos as *runx1*, a transcription factor, is an essential regulator of definitive hematopoiesis and is consistently expressed in HSCs at all sites of embryonic and adult hematopoiesis^[@CR22]^. All three HDACIs triggered an increase in runx1^+^ cells in the AGM compared to minimal expression in control embryos (Fig.\u00a0[1d,e](#Fig1){ref-type=\"fig\"}), suggesting that HDAC inhibition by VPA, resminostat, or entinostat leads to the *in vivo* expansion of HSPCs in zebrafish.Figure 1HDACIs increase c-myb^+^ HSPC number and *runx1* expression in zebrafish embryos. (**a**) Schematic representation of HSPC development in the AGM and CHT regions of a zebrafish embryo. YS -- yolk sac; YE -- yolk extension; DA -- dorsal aorta; AV -- axial vein; AGM -- aorta-gonad-mesonephros; CHT -- caudal hematopoietic tissue. Small green circles between the DA and AV and in CHT regions represent HSPCs. (**b**) Image based identification of c-myb^+^ cells in the AGM and CHT region identified HDACIs (valproic acid, resminostat, and entinostat) as enhancers of HSPC cell-count at 36 hpf (Bars\u2009=\u2009200\u2009\u00b5m). The region-of-interest (AGM and CHT) marked by white line and the c-myb^+^ cells are marked by red circle (indicated by red arrows) and the false positive objects that are excluded from the quantification are marked by yellow circle (indicated by yellow arrows). (**c**) Quantification of relative number of c-myb^+^ cells in the AGM and CHT region showing increased cell-count after 40\u2009\u00b5M VPA, resminostat or entinostat treatment compared to DMSO (n\u2009=\u20095). (**d**) Validation of identified hits through whole-mount *in situ* hybridization for *runx1* expression. (**e**) Quantification of *runx1* relative stain intensity shows significantly higher *runx1* expression in HDACI treated fish compared to DMSO controls. Intensity was calculated using area under curve analyses in ImageJ. Single images were split into 5 regions of interest (ROIs) and intensity was normalized to background signal. SD displays deviation among 5 ROIs in one image (n\u2009=\u20093). Data are shown as mean\u2009\u00b1\u2009SD, \\*p\u2009\\<\u20090.05; \\*\\*p\u2009\\<\u20090.01; \\*\\*\\* p\u2009\\<\u20090.001.\n\nHDAC inhibitors increase zebrafish hematopoietic cell engraftment efficiency {#Sec4}\n----------------------------------------------------------------------------\n\nOf the identified hits, VPA was chosen for validation using an *ex vivo* transplantation procedure. Zebrafish adult whole kidney marrow (WKM) cells from *Tg(ubi:GFP)* animals^[@CR23]^ were treated with VPA (50\u2009\u00b5M) or PBS (control) *in vitro*, transplanted into irradiated Casper^[@CR24]^ adult zebrafish (Fig.\u00a0[S1](#MOESM1){ref-type=\"media\"}). This incubation time was chosen based on data from pilot experiments that showed a sharp reduction in cell viability after longer time-periods of *ex vivo* culture. After 28 days, the WKM cells from recipient Casper fish were analyzed by flow cytometry for engraftment of donor cells (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). VPA treatment significantly increased WKM cell engraftment compared to PBS treatment (donor chimerism 46.5\u2009\u00b1\u200926.8% vs. 0.7\u2009\u00b1\u20090.2; p\u2009=\u20090.0056; Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}), suggesting that *ex vivo* VPA treatment maintains the stem cell potential of WKM cells and increases the engraftment capability after transplantation in this model.Figure 2*Ex vivo* VPA treatment preserves engraftment capacity of whole kidney marrow cells. (**a**) A representative plot showing gating strategy used to determine the degree of chimerism in the recipient and the FACS plot of PBS and VPA treated conditions. (**b**) PBS-treated control donor WKM cells failed to engraft the kidney of the recipient, but VPA treated cells demonstrated an engraftment capacity similar to that of freshly isolated cells (uncultured; n\u2009=\u20095 per group). Data are shown as mean\u2009\u00b1\u2009SD, \\*\\*p\u2009\\<\u20090.01, \\*\\*\\*p\u2009\\<\u20090.001.\n\nHDACIs increase CD34^+^ and CD34^+^CD90^+^ cell populations in human G-CSF mobilized peripheral blood stem cells {#Sec5}\n----------------------------------------------------------------------------------------------------------------\n\nNext, the effects of HDACIs on human CD34^+^ G-CSF-mobilized peripheral blood stem cells (purity \\>95%) were analyzed by treating them *in vitro* with VPA, resminostat, or entinostat for five consecutive days in the presence of cytokines. The number and phenotype of the cultured cells were assessed at the end of the treatment period. The proportion of CD34^+^ cells was 79.4\u2009\u00b1\u20092.4%, 70.2\u2009\u00b1\u20092.8%, and 72.2\u2009\u00b1\u20092.2% for VPA, resminostat, and entinostat, respectively, compared to 35.6\u2009\u00b1\u20093.2% and 29.4\u2009\u00b1\u20093.3% for PBS- or DMSO-treated controls (p\u2009\\<\u20090.001 for all; Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). Importantly, the fraction of CD34^+^CD90^+^ cells, a population known to be enriched in repopulating cells^[@CR25]^, was higher in VPA, resminostat, or entinostat treated samples (59.3\u2009\u00b1\u20093.4%, 71.5\u2009\u00b1\u20093.8%, and 67.9\u2009\u00b1\u20094.3%, respectively) compared to PBS- or DMSO-treated cells (2.2\u2009\u00b1\u20090.7% and 2.0\u2009\u00b1\u20090.2%, respectively (p\u2009\\<\u20090.001 for all; Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). All three HDACIs also amplified the starting cell numbers compared to controls and resulted in significantly higher absolute cell numbers of both CD34^+^ and CD34^+^CD90^+^ cell populations after 5 days of *in vitro* treatment (Fig.\u00a0[3b and c](#Fig3){ref-type=\"fig\"}). Thus, HDACIs identified from the zebrafish screen expanded phenotypic human HSPCs *in vitro*. As VPA exerted similar effects as the other two compounds but had lower toxicity over a broader range of concentrations, all subsequent experiments on human HSPCs were performed with VPA alone.Figure 3*Ex vivo* expansion of G-CSF mobilized CD34^+^ HSPCs treated for 5 days with HDACIs. (**a**) Representative flow-cytometry analysis of CD34 and CD90 expression after 5 d of *ex vivo* HDACI treatment. (**b**) Absolute cell numbers of CD34^+^ cells. (**c**) Absolute cell numbers of CD34^+^CD90^+^ cells. Both CD34^+^ and CD34^+^CD90^+^ cells were significantly increased after 5 days treatment with VPA (1\u2009mM), resminostat (1.5\u2009\u00b5M) or entinostat (1.5\u2009\u00b5M) compared to controls (PBS and DMSO; n\u2009=\u20095). Data are shown as mean\u2009\u00b1\u2009SD, \\*\\*\\*p\u2009\\<\u20090.001.\n\nVPA increases attachment of hHSPCs on MSCs during *ex vivo* expansion {#Sec6}\n---------------------------------------------------------------------\n\nAs MSCs are one of the supporting factors of HSPC maintenance^[@CR26]^ and proliferation^[@CR11]^, we investigated if VPA influences the interaction between hHSPCs and MSCs *in vitro*. Similar to suspension cultures, VPA induced hHSPC expansion in co-cultures with MSCs. In addition, a higher density of hHSPCs were attached to the MSC layer in the VPA-treated group than the control group (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}), and 3-fold greater number of hHSPCs were attached to MSCs in the VPA-treated group compared to controls (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}). These findings prompted us to investigate the adhesion strength between VPA-treated CD34^+^ cells and MSCs using atomic force microscopy-based single-cell force spectroscopy (AFM-SCFS) under defined conditions of contact force and time. The adhesion between hHSPCs and the substrate-bound MSC was quantified during retraction (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}). In accordance with the bulk assays, higher detachment forces, i.e. stronger cell-cell interactions, between HSPCs and MSCs were measured in VPA-treated cells compared to PBS-treated controls (Fig.\u00a0[4d](#Fig4){ref-type=\"fig\"}). These observations suggest that VPA-treated CD34^+^ cells are more adhesive and that this adhesiveness can result in greater interaction with MSCs. Thus, VPA-induced expansion of phenotypic HSPCs *in vitro* is also associated with enhanced attachment to MSCs.Figure 4*Ex vivo* VPA-expanded CD34^+^ cells exhibit increased adhesion to MSCs. (**a**) Freshly isolated CD34^+^ cells were seeded onto a confluent MSC layer. Representative images after 5 days of co-culture with MSCs showed an increase in the number of VPA-treated adherent cells (right) compared to control cells (left). Images were taken after sufficient washing with PBS. (Scale bar: 250\u2009\u00b5m). (**b**) In the adhesion assay, a significantly higher number of VPA expanded cells were attached to MSCs compared to control (n\u2009=\u20094). (**c**) Schematic representation of atomic force microscopy-based single-cell force spectroscopy used to measure adhesive strength of VPA treated and control cells. **(d)** Plot showing detachment force measurement after 5 days of VPA or PBS (control) treatment. Adhesive strength of VPA treated cells was 2--3 fold higher than control (n\u2009=\u20093). Data are mean\u2009\u00b1\u2009SD, \\*\\*\\*p\u2009\\<\u20090.001.\n\nVPA treatment preserves self-renewal capacity and differentiation potential {#Sec7}\n---------------------------------------------------------------------------\n\nTo determine whether VPA-expanded cells retained their differentiation capacity, we performed colony forming unit (CFU) assays and compared their clonogenicity to freshly isolated human CD34^+^ cells from GCS-F mobilized blood. VPA-treated cells formed significantly fewer total number of colonies compared to the PBS treated cells (Fig.\u00a0[S3a](#MOESM1){ref-type=\"media\"}). This difference was mainly due to an increase in BFU-E (burst forming unit-erythroid) colonies in the PBS-treated CD34^+^ cells as there were no statistically significant differences in the total number of CFU-G, CFU-M, CFU-GM, and CFU-GEMM populations between the VPA- and PBS-treated CD34^+^ cells. These observations imply that VPA treatment does not affect clonogenicity and differentiation capacity of the hHSPCs (Fig.\u00a0[S3b](#MOESM1){ref-type=\"media\"}). We also analyzed the cobblestone area forming cell (CFAC) potential of *ex vivo* expanded CD34^+^ cells to characterize their *in vitro* stem cell functions and found no significant differences between VPA-treated and control cells (Fig.\u00a0[S3c](#MOESM1){ref-type=\"media\"}).\n\nVPA suppresses CXCR4 expression and HSPC migration toward SDF-1 *in vitro* {#Sec8}\n--------------------------------------------------------------------------\n\nSuccessful HSC transplantation depends on effective HSPC homing and engraftment in the bone morrow after transplantation, and stromal-derived factor-1 (SDF-1, CXCL12) and its receptor CXCR4 (CD184) play essential roles^[@CR27]^. To investigate migration and homing potential of *ex vivo* VPA-expanded hHSPCs, we measured CXCR4 expression using flow cytometry (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}) and found a drastic reduction in CXCR4 expression upon VPA treatment compared to control cells (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). This reduction was confirmed by real-time PCR (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}) and verified by RNA-seq data (Fig.\u00a0[5h](#Fig5){ref-type=\"fig\"}). RNA-seq data also revealed that VPA treatment induced SDF-1 expression in CD34^+^ cells (Fig.\u00a0[5h](#Fig5){ref-type=\"fig\"}). The functional relevance of these findings was tested by quantifying the migration capacity of VPA-expanded CD34^+^ cells toward SDF-1 using a transwell migration assay where we detected significantly lower numbers of migratory CD34^+^ cells compared to PBS-treated control hHSPCs (Fig.\u00a0[5d](#Fig5){ref-type=\"fig\"}).Figure 5Valproic acid affects adhesion of HSPCs and suppresses their migration toward SDF-1 *in vitro*. G-CSF mobilized CD34^+^ HSPCs were treated *in vitro* for 5 days with VPA or PBS and analyzed for the expression of molecules that are involved in cell adhesion and migration. The functional consequence of VPA-treatment on the migratory capacity toward SDF-1 was also evaluated. (**a**) Representative dot-plot of CD184 (CXCR4) expression on the cell surface of CD34^+^ cells as determined by flow cytometry. (**b**) VPA-treatment significantly reduced the expression of CXCR4 on the cell surface of CD34^+^ cells compared to control cells (n\u2009=\u20093), measured by flow cytometry. (**c**) Reduced CXCR4 expression was confirmed by quantitative PCR (n\u2009=\u20093). (**d**) Trans-well migration assay showed that VPA-treatment significantly reduced the migration capacity of CD34^+^ cells toward an SDF-1 gradient (100 ng/ml) compared to control cells (n\u2009=\u20094). (**e**) Representative plot of CD146 (MCAM) expression on the cell surface of CD34^+^ cells as determined by flow cytometry. (**f**) Flow cytometric analysis showed that VPA-treatment significantly increased surface expression of MCAM on CD34^+^ cells compared to controls (n\u2009=\u20094). (**g**) Significantly higher expression of MCAM in VPA expanded CD34^+^ cells was verified by quantitative PCR (n\u2009=\u20094). (**h**) RNA sequencing revealed that VPA-treatment substantially changed the expression of molecules involved in cell adhesion and migration in CD34^+^ cells compared to control cells, including CXCR4 and MCAM (n\u2009=\u20094). Data are mean\u2009\u00b1\u2009SD, \\*\\*p\u2009\\<\u20090.01, \\*\\*\\*p\u2009\\<\u20090.001.\n\nNext, we used flow cytometry In order to identify other molecular pathways that could mediate the increased adhesiveness of VPA-treated cells and found that the adhesion molecule CD146 (MCAM) was expressed on 40% of VPA-treated CD34^+^ cells while control cells did not express this molecule (Fig.\u00a0[5e and f](#Fig5){ref-type=\"fig\"}). Quantitative PCR analysis (Fig.\u00a0[5g](#Fig5){ref-type=\"fig\"}) and RNA-seq (Fig.\u00a0[5h](#Fig5){ref-type=\"fig\"}) also showed significantly elevated MCAM transcript levels in VPA-treated cells compared to PBS-treated controls. However, RNA-seq data also showed that VPA induced the expression of several cell adhesion molecules along with MCAM (Fig.\u00a0[5h](#Fig5){ref-type=\"fig\"}). Thus, VPA not only decreases CXCR4 expression on CD34^+^ cells that is associated with reduced migratory potential toward SDF-1, but also triggers the upregulation of specific adhesion molecules including MCAM, which probably mediate the VPA-induced increase in the adhesion capacity.\n\nVPA treatment reduces bone marrow homing efficiency of CD34^+^ cells {#Sec9}\n--------------------------------------------------------------------\n\nAs VPA-expanded CD34^+^ cells exhibited reduced migration towards SDF-1 *in vitro*, we tested their bone marrow homing capacity in NSG mice. Equal numbers of re-isolated CD34^+^ cells (1\u2009\u00d7\u200910^6^ cells/mouse), treated with either VPA or PBS, were intravenously injected into sub-lethally irradiated NSG mice (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}), and the bone marrow of recipient mice analyzed for the presence of human CD45^+^ cells 24\u2009hours after injection. Mice injected with VPA-treated human CD34^+^ cells had significantly lower number of human leucocytes in the bone marrow compared to those that received control cells (1.6\u2009\u00b1\u20090.2\u2009\u00d7\u200910^2^/femur vs. 4.4\u2009\u00b1\u20090.7\u2009\u00d7\u200910^2^, p\u2009=\u20090.0001; Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}), implying reduced *in vivo* bone marrow homing capacity of these cells.Figure 6Valproic acid decreases bone marrow homing capacity of *in vitro* expanded CD34^+^ cells. (**a**) Schematic representation of the homing assay performed in NSG mice. CD34^+^ cells were re-isolated after 5 days *ex vivo* treatment with VPA or PBS. 1\u2009\u00d7\u200910^6^ CD34^+^ cells were transplanted by intravenous injection into the retro-orbital venous plexus of sub-lethally irradiated (100\u2009cGy) NSG mice. Homing was quantified by flow cytometric analysis of human leucocytes (human CD45^+^) in the bone marrow of recipient mice 24\u2009hours after injection. (**b**) The absolute number of human leukocytes homing to the femur of recipient mice was significantly reduced by *ex vivo* treatment of CD34^+^ with VPA compared to control cells (n\u2009=\u20098--14). Data are shown as mean\u2009\u00b1\u2009SD, \\*\\*\\*p\u2009\\<\u20090.001.\n\nVPA-expanded hHSPCs retain their long-term repopulating potential in immunocompromised mice {#Sec10}\n-------------------------------------------------------------------------------------------\n\nNext, we investigated the repopulating potential of *in vitro* VPA-expanded human CD34^+^ cells by transplanting equal numbers of VPA-expanded or PBS-treated re-isolated CD34^+^ cells into sub-lethally irradiated immunocompromised mice. Peripheral blood chimerism and the phenotype of circulating human leucocytes were monitored periodically and human leucocyte repopulation of the bone marrow was assessed at 20 weeks after transplantation (Fig.\u00a0[7a](#Fig7){ref-type=\"fig\"}). While there were no differences in peripheral blood chimerism at the 4-week time point or the 12-week and later time points, at 8-weeks mice receiving PBS-treated CD34^+^ cells tended to have higher overall chimerism compared to those given VPA-treated CD34^+^ cells (14\u2009\u00b1\u20093.5% vs. 5\u2009\u00b1\u20093.1%, p\u2009=\u20090.0913, unpaired t-test) (Fig.\u00a0[7b](#Fig7){ref-type=\"fig\"}). Thus, despite impaired short-term engraftment, as measured by peripheral blood chimerism, the phenotype of the circulating human leucocytes was indistinguishable between the two groups, suggesting that *in vitro* VPA treatment did not affect the differentiation potential of these cells for the main hematopoietic lineages, namely, CD3^+^ T cells, CD19^+^ B cells, and CD33^+^ myeloid cells (Fig.\u00a0[7c](#Fig7){ref-type=\"fig\"}). Similar to the results of peripheral blood chimerism, long-term bone marrow engraftment in recipient mice showed no significant differences in overall BM chimerism (PBS: 11.3\u2009\u00b1\u20094.2%, VPA: 8\u2009\u00b1\u20097.4%; Fig.\u00a0[7d](#Fig7){ref-type=\"fig\"}). Further, the numbers of human leucocyte-derived T lymphocytes (CD3^+^), B lymphocytes (CD19^+^), and myeloid cells (CD33^+^) were indistinguishable between the two groups (Fig.\u00a0[7e](#Fig7){ref-type=\"fig\"}). These results confirm *in vitro* data (Fig.\u00a0[S3](#MOESM1){ref-type=\"media\"}) that VPA treatment does not negatively affect the long-term engraftment capacity and differentiation potential of human CD34^+^ cells. Importantly, as VPA treatment increases the overall number of CD34^+^ cells by 2--3 fold compared to controls (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}), *ex vivo* VPA treatment results in a net expansion of mouse-repopulating HSCs.Figure 7Valproic acid modifies short-term engraftment but does not influence long-term engraftment and differentiation capacity of VPA expanded CD34^+^ cells. (**a**) Schematic representation of the engraftment assay performed in NSG mice. Sub-lethally irradiated (100\u2009cGy) mice were intravenously transplanted with 3.5\u2009\u00d7\u200910^5^ CD34^+^ VPA-treated or control cells. Engraftment of human CD34^+^ cells was monitored by analyzing chimerism and phenotype of circulating human leucocytes (human CD45^+^) in the peripheral blood of recipient mice every four weeks by flow cytometry. Long-term *in vivo* marrow repopulation capacity was determined at 20 weeks after transplantation by quantification and phenotyping of human leucocytes in the femur of NSG mice (n\u2009=\u20095 per group). (**b**) Overall peripheral blood chimerism increased by week 12 and subsequently declined. Mice transplanted with PBS-treated CD34^+^ cells showed elevated overall chimerism compared to mice that received VPA-treated CD34^+^ cells. Differences were most pronounced at week 8 but did not reach statistical significance (p\u2009=\u20090.913). Data represent mean\u2009+\u2009SD. (**c**) Lineage commitment of circulating human leucocytes was examined by analyzing CD3, CD19, and CD33 cell surface expression with no significant differences in the proportion of T-cells, B-cells or myeloid cells. (**d**) Absolute numbers of human leucocytes (human CD45^+^) per femur at week 20 after transplantation did not differ between groups (p\u2009=\u20090.61). (**e**) Lineage diversification of long-term marrow repopulating human leucocytes was similar in mice injected with VPA-treated or PBS-treated human CD34^+^ cells at 20 weeks after transplantation.\n\nVPA-induced changes in the gene-expression profile of CD34^+^ HSPCs {#Sec11}\n-------------------------------------------------------------------\n\nTo test how VPA affects gene expression in CD34^+^ cells, we performed mRNA sequencing. Freshly isolated CD34^+^ cells were treated with either VPA or PBS for 5 days and RNA-seq performed on total RNA isolated from both sets of CD34^+^ cells. The clustering of individual biological replicates revealed high homology between samples (Fig.\u00a0[8a](#Fig8){ref-type=\"fig\"}). A comparison of gene expression profiles between the two sets revealed that VPA treatment upregulated 1215 genes and downregulated 162 genes (Fig.\u00a0[8b](#Fig8){ref-type=\"fig\"}). GO (gene ontology) pathway analysis revealed an increase in the expression of genes predominantly involved in cell adhesion due to VPA treatment. This observation is in agreement with data from attachment assays, including AFM-SCFS, which showed greater adhesiveness of VPA-expanded cells toward MSCs. In addition, VPA-treated CD34^+^ cells also significantly downregulated genes that participate in chemotaxis. This result is in line with the observed impairment of chemotactic behavior *in vitro*, especially the SDF-1/CXCR4 axis. VPA-expanded cells also increased the expression of genes involved in Notch and wnt signaling (Fig.\u00a0[S6](#MOESM1){ref-type=\"media\"}). Pluripotency genes such as *OCT4*, *SOX2*, and *NANOG* were not upregulated (Fig.\u00a0[S5](#MOESM1){ref-type=\"media\"}), and the changes in the expression of *CXCR4*, *MCAM*, *OCT4*, *SOX2*, and *NANOG* were further verified by quantitative real-time PCR.Figure 8Gene expression profiling revealed induction of cell adhesion pathways and reduced expression of genes involved in chemotaxis. (**a**) Heat map of a sample-to-sample Pearson correlation and dendrogram showing sample-to-sample correlation. All biological replicates cluster well with each other, and all samples from different populations are clearly separated from each other (analysis by R). (**b**) The red dots represent differentially expressed genes (DEG) by VPA treatment (1% false discovery rate). (**c**,**d**) GO pathway analysis for differentially expressed genes. Plots show biological processes that are associated with genes that are up- or down-regulated in VPA treated CD34^+^ cells.\n\nDiscussion {#Sec12}\n==========\n\nWe have identified HDAC inhibitors, particularly VPA, resminostat, and entinostat, as potent *in vitro* multipliers of hHSPCs. Initially, a transgenic zebrafish model was used to screen for molecules capable of expanding the HSPC pool *in vivo*, and although several HDACIs were included in the screen, only VPA, resminostat, and entinostat increased the number of c-myb^+^ HSPCs, suggesting that HSPC expansion is specific to these three drugs rather than a general group effect of HDACIs. Importantly, the results from the zebrafish screen were reproduced in human HSPCs as *in vitro* treatment of CD34^+^ cells isolated from G-CSF mobilized peripheral blood with these HDACIs led to a significantly greater expansion of both phenotypic and functional HSPCs. Even though *in vitro* VPA-induced expansion of G-CSF mobilized CD34^+^ cells was associated with functional changes that affected homing and short-term engraftment, their *in vivo* long-term engraftment (and thereby self-renewal) and differentiation capacity remained unaltered.\n\nThe use of G-CSF mobilized HSPCs has many advantages over allogeneic bone marrow transplantation such as earlier hematopoietic recovery in the recipient, higher anti-leukemic activity, and a less invasive procedure for the donor. Engraftment is also much faster compared to cord blood stem cells and the greater numbers of HSPCs obtained by G-CSF ad apheresis predestine mobilized HSPCs as starting material for future genetic-engineering approaches. Therefore, identification and validation of small molecules that significantly expand mobilized HSPC numbers *ex vivo* represent a clinically attractive approach, and our results imply that *ex vivo* treatment of mobilized HSPCs with VPA can increase the absolute number of CD34^+^ cells with preserved *in vivo* repopulating potential.\n\nWhile our results have further contributed to the body of evidence that the zebrafish model represents a valuable tool to study vertebrate hematopoiesis^[@CR12],[@CR28]^, more importantly, its use as an *in vivo* screening tool is valuable as large drug libraries can be efficiently screened and any systemic effects made readily apparent.\n\nThe fate of any cell, including HSPCs, is substantially influenced by epigenetics^[@CR29],[@CR30]^. Among several functionally relevant epigenetic modifications, DNA methylation and histone acetylation play important roles in health and disease as they regulate gene expression, and drugs that target or modify methylation and acetylation patterns are associated with a range of conditions, including cancer progression^[@CR31],[@CR32]^. Importantly, HDACIs and hypomethylating agents (HMAs) have been shown to alter the epigenetic status of HSPCs^[@CR33],[@CR34]^, and epigenetic modifiers such as 5-Aza 2\u2032-deoxycytidine (an HMA) and trichostatin-A (an HDACI) are capable of expanding CD34^+^CD90^+^ BM cells with preserved long-term repopulating potential^[@CR35]^. A recent report also showed that VPA can epigenetically reprogram cord blood (CB)-derived CD34^+^ cells *in vitro* that leads to extensive expansion of functional CB HSCs; these cells have improved repopulating capacity in NSG mice compared to control treated cells. Furthermore, they detected that higher number of VPA treated cells residing within G0/G1 and G2/M phase than the cells that are exposed to control^[@CR10]^. However, they also found that *in vitro* VPA treatment led to greater CXCR4 expression that was associated with enhanced migration toward a SDF-1 gradient *in vitro* and increased homing to the bone marrow of NSG mice *in vivo*. Contrarily, our cell cycle analyses revealed no difference in the proportion of CD34\u2009+\u2009cells in G0/G1 vs S/M-phase after VPA treatment compared to control (Fig.\u00a0[S2](#MOESM1){ref-type=\"media\"}) and we show that CXCR4 downregulation upon VPA treatment, identified by flow cytometry and confirmed by direct RNA-Seq, and complementary simultaneous upregulation of SDF-1, indicating the presence of an autocrine feedback-loop^[@CR36]^. These disparities in CXCR4 expression between the two studies can be explained in part by differences in HSPC source and experimental conditions such as cytokine concentration/combination and culture medium. Further, G-CSF mobilized CD34^+^ cells exhibit robust CXCR4 expression levels probably because their metalloproteinase-induced degradation declines during differentiation^[@CR37]^. Dynamic changes in CXCR4 expression should also be considered as a previous study reported that CXCR4 expression is heterogeneous in the human HSPC compartment and that its stem cell repopulating potential does not necessarily correlate with CXCR4 expression^[@CR38]^.\n\nThe success of HSCT by intravenous infusion of the graft also relies on the homing of HSCs to recipient bone marrow. This process is regulated by multiple molecular events such as cell-cell interaction, migration and adhesion. We observed reduced homing of the VPA-primed CD34^+^ cells that was associated with either unaffected or lower overall peripheral blood chimerism in recipient mice at 8 weeks after transplantation. Conversely, at this time point; Chaurasia *et al*. (2004) found a significant increase in bone marrow overall chimerism in NSG mice transplanted with VPA-treated CB CD34^+^ cells; however, these results cannot be directly compared, as peripheral blood chimerism does not necessarily reflect levels of bone marrow chimerism.\n\nApart from reduced migration, we observed an increase in the adhesiveness of VPA-treated CD34^+^ cells toward MSCs in co-culture experiments and quantified this force using AFM--SCFS. Cell adhesion is typically mediated by multiprotein complexes that connect individual cells to the extracellular matrix (ECM) and to other cells. Several specialized adhesion molecules such as ICAM-1, NCAM, VCAM-1, and N-cadherin are expressed on subsets of HSCs and are known to be crucial for HSC function and niche interaction^[@CR39]--[@CR42]^. The gene expression data reported here show that several molecules known to be involved in cell adhesion, including NCAM, are upregulated in VPA-treated CD34^+^ cells. Unexpectedly, the expression of melanoma cell adhesion molecule (MCAM, CD146) was upregulated in VPA-treated CD34^+^ cells and its surface expression was verified by flow cytometry. This result is surprising as while it is known that MCAM is commonly expressed within the vascular wall, including in the vascular endothelial cells, the vascular smooth muscle cells, and the MSCs^[@CR43]^, its expression has not been described in HSCs. Further, MCAM expression on MSCs is known to influence the fate of HSCs^[@CR44]^. MCAM expression has also been demonstrated on subsets of activated T- and B-lymphocytes, mainly in the context of inflammatory conditions^[@CR45]^. We hypothesize that this induction of MCAM expression is involved in the increased adhesiveness of HSPCs on MSCs and is supported by AFM measurements.\n\nIn contrast to the observed reduction in bone marrow homing and short-term engraftment, long-term engraftment and differentiation capacity of VPA-treated CD34^+^ cells was not different from controls. Also, the relative and absolute numbers of human phenotypic HSPCs in the bone marrow of NSG mice at 20 weeks after transplantation were indistinguishable between VPA and controls, suggesting that the *in vitro* VPA-expanded CD34^+^ cells retained their stem cell properties. These results show that VPA significantly influences the functional activity of HSPCs but the effect on stem cell frequency still needs to be determined. A previous report showed that VPA-induced expansion of CB CD34^+^ cells *in vitro* was associated with a transient increase in the expression of genes associated with chromatin remodeling and pluripotency in induced pluripotent stem cells (iPS) and embryonic stem cells (ES), such as *SOX2*, *OCT4*, and *NANOG* ^[@CR10]^. Interestingly, we found that VPA did not induce the expression of these pluripotency genes in G-CSF mobilized adult CD34^+^ cells (Fig.\u00a0[S5](#MOESM1){ref-type=\"media\"}) but rather upregulated *Jag2*, *Notch3*, *Hes1*, and *DLL1* (Fig.\u00a0[S6a](#MOESM1){ref-type=\"media\"}). Such recurrent variations in results between these two studies point to the differential effects of VPA on the ontogenetically more immature CB CD34^+^ HSCs compared to the mature G-CSF mobilized adult CD34^+^ cells. Importantly, these genes (*Jag2*, *Notch3*, *Hes1*, and *DLL1)* and their related pathways are all known to be involved in the preservation of stem cell properties during the *ex vivo* expansion of HSCs^[@CR46]^. Furthermore, *DLL1* induces erythroid differentiation and inhibits myeloid skewing^[@CR47],[@CR48]^.\n\nThe role of wnt signaling in the regulation of hematopoiesis has been extensively studied^[@CR49]^, and both the canonical and non-canonical pathways of wnt signaling are crucial for HSC maintenance and activation^[@CR50]^. Gene expression data presented here (Fig.\u00a0[S6b](#MOESM1){ref-type=\"media\"}) show upregulation of several genes that are involved in both canonical and non-canonical wnt signaling, such as Frizilled (Fzd) 8 and Flemingo (Fmi), members of the non-canonical wnt signaling pathway that are crucial for long-term quiescent HSC maintenance^[@CR51]^.\n\nTo conclude, we show that certain HDACIs enhance the *in vitro* expansion of CD34^+^ HSPCs derived from G-CSF mobilized peripheral blood and that, functionally, these cells are capable of effective long-term engraftment and multi-lineage differentiation, even though VPA treatment led to reduced homing and short-term engraftment that are probably related to lower migratory capacity and enhanced adhesiveness. Importantly, as our observations are in contrast to previous work with cord blood derived CD34^+^ cells, we conclude that VPA-mediated effects on HSPCs are distinct and depend on the graft source used. Lastly, impaired short-term engraftment that might lead to prolonged cytopenia after transplantation of *in vitro* VPA expanded G-CSF mobilized peripheral blood stem cells has to be considered when translating HDACI-based expansion protocols into the clinical setting.\n\nMaterials and Methods {#Sec13}\n=====================\n\nDetailed materials and methods are described in the supporting information materials and methods. All experiments were performed in accordance with the local and national regulations and guidelines and were approved by the local institutional review board (Ethical committee of Technischen Universit\u00e4t Dresden, Fiedlerstra\u00dfe 33, 01307 Dresden). The xenotransplant experiments were performed according to the national guidelines after approval by the local animal protection committee.\n\nSmall molecule screen and validation {#Sec14}\n------------------------------------\n\nAll animal experiments were carried out in accordance with live animal handling and research regulations under protocols approved by the animal ethics committees of the Technische Universit\u00e4t Dresden and the Landesdirektion Sachsen (approval number no: AZ 24D-9168.11-1/2008-4). Fish were maintained at 28\u2009\u00b0C^[@CR52]^, and synchronized embryos from homozygous zebrafish *c-myb*:EGFP^[@CR12]^ transgenic line were collected in E3. The semi-automated *in vivo* chemical screen was performed as described previously^[@CR18]^. Several small focused libraries were screened, including an epigenetic modulator library (Selleckchem) containing several HDACIs. Identified hits from the chemical screen were validated by whole-mount *in situ* hybridization for *runx1* expression as described elsewhere^[@CR53]^. Compounds were also tested at a concentration of 40\u2009\u00b5M and imaged as reported previously^[@CR54]^.\n\nAdult zebrafish kidney marrow transplantation {#Sec15}\n---------------------------------------------\n\nAdult kidney marrow cells from multiple *Tg(ubi:GFP)* fish were harvested, WKM cells dissociated, pooled, and filtered^[@CR55]^. Unfractionated WKM cells were incubated *in vitro* at 28\u2009\u00b0C in ZKS (zebrafish kidney stromal) medium^[@CR56]^ and treated with either VPA (50\u2009\u00b5M) or PBS control for 48\u2009hours. One day before transplantation, three-month-old Casper recipient fish were sub-lethally irradiated at 15\u2009cGy. The *ex vivo* chemical treatment of WKM cells and their transplantation were performed as previously reported^[@CR57]^ (Fig.\u00a0[S1](#MOESM1){ref-type=\"media\"}). Fresh WKM cells from *Tg(ubi:GFP)* animals were used as a positive control to measure engraftment efficiency. Donor chimerism was assessed on a BD LSR II flow cytometer (BD Biosciences) and its analysis limited to myelomonocytes, as described previously^[@CR58]^.\n\n*Ex vivo* culture and co-culture assay {#Sec16}\n--------------------------------------\n\nG-CSF (granulocyte colony-stimulating factor) mobilized peripheral blood for CD34^+^ HSPC isolation and bone marrow for MSC isolation was obtained from healthy donors after informed consent (ethical approval no. EK221102004, EK47022007) and human cells were isolated according to the Ethical committee of Technischen Universit\u00e4t Dresden (<https://www.tu-dresden.de/tu-dresden/organisation/gremien-und-beauftragte/kommissionen/ethikkommission>). Purified CD34^+^ cells (1\u2009\u00d7\u200910^4^ cells/well) were cultured in CellGro serum free medium (CellGenix) supplemented with 10 ng/ml stem cell factor (SCF), 10 ng/ml Fms-related tyrosine kinase 3 ligand (FLT3 ligand), 10 ng/ml interleukin 3 (IL-3), and incubated at 37\u2009\u00b0C in 5% CO~2.~ After 24\u2009h, the cells were exposed to PBS, VPA (1\u2009mM), DMSO, resminostat (1.5\u2009\u00b5M), or entinostat (1.5\u2009\u00b5M) for 5 days. After treatment, expanded cells were stained for CD34 and CD90 expression. For further experiments, CD34^+^ cells were re-isolated from *ex vivo*-expanded cells and their purity analyzed as stated above. For co-culture experiments, freshly isolated CD34^+^ cells were suspended in CellGro serum free medium supplemented with cytokines and plated onto a confluent MSC layer. Co-culture of CD34^+^ cells with human MSCs was performed as previously described^[@CR11]^.\n\nAtomic force microscopy-based single-cell force spectroscopy (AFM-SCFS) {#Sec17}\n-----------------------------------------------------------------------\n\nA NanoWizard II AFM equipped with a CellHesion module (JPK Instruments), mounted on an inverted light microscope (Axiovert 200, Carl Zeiss), was used to perform AFM-SCFS. All measurements were performed in PBS containing Ca^2+^/Mg^2+^ at 37\u2009\u00b0C using a temperature-controlled sample chamber (PetriDish Heater, JPK Instruments). Tipless cantilevers with a nominal spring constant of 0.08\u2009N/m (PNP-TR-TL-Au, Nanoworld) were coated with 1\u2009mg/ml wheat germ agglutinin (Vector Laboratories) as described previously^[@CR59]^. Cantilevers were calibrated before every experiment using built-in functions of the AFM software (JPK Instruments) based on the equipartition theorem^[@CR60]^. For SCFS experiments, MSCs were grown to confluence in 30-mm Petri dishes (TPP-Sigma-Aldrich). The Petri dish was placed in the PetriDish Heater and subsequently, 1\u2009\u00d7\u200910^3^ CD34^+^ HSPCs, either PBS- or VPA-treated for 5 days, were seeded into the Petri dish. A single HSPC was attached to the AFM cantilever as described^[@CR59]^. Force-distance (F-D) curves were acquired by approaching the cantilever-bound HSPC at a constant speed of 5\u03bcm/sec onto a single MSC in the Petri dish until a contact force of 2 nN was reached. Before retraction of the cantilever-bound cell at constant speed (5 \u03bcm/sec), cells were maintained in contact for 60\u2009sec in constant force mode. Approximately 10 F-D curves per cell and 10 cells/condition were measured, yielding 100 data points per condition. Data processing software provided by the AFM manufacturer was used to extract maximum detachment force (F~detach~) from retract F-D curves.\n\nHoming Assay {#Sec18}\n------------\n\nAll animal experiments were performed according institutional guidelines and the German animal protection law (Landesdirektion Sachsen, 24-9168.11-1/2013-54). Re-isolated CD34^+^ cells (1\u2009\u00d7\u200910^6^ cells/mouse), both VPA-treated and controls, were transplanted by retro-orbital injection into NSG mice. The recipient mice were irradiated (100\u2009cGy) 24\u2009h prior to transplantation. Bone marrow of recipient mice (one femur/mouse) was analyzed 24\u2009h after transplantation by flow cytometry for the presence of human CD45^+^. Eight NSG mice received PBS expanded cells (control) while 14 mice received VPA-expanded cells. Eleven mice did not receive any cells and their bone marrow was similarly analyzed to determine thresholds for the analysis of human cells.\n\nLong-term engraftment assay {#Sec19}\n---------------------------\n\nCD34^+^ cells were re-isolated after 5 days *ex vivo* treatment with VPA or PBS (control) and 3.5\u2009\u00d7\u200910^5^ cells/mouse were transplanted by intravenous injection in the retro-orbital venous plexus of the recipient mice. NSG mice were sub-lethally irradiated with 100\u2009cGy 24\u2009h prior to transplantation. To assess engraftment, peripheral blood from recipient mice was obtained every 4 weeks by retro-orbital bleeding and collected in heparinized hematocrit capillaries (Brand GMBH). Peripheral blood was analyzed by flow cytometry using anti-mouse CD45, anti-human CD45, CD3, CD19 and CD33 antibodies. Mice were sacrificed at 20 weeks after transplantation and the bone marrow of each mouse (2 femurs and 1 tibia) was analyzed for the presence of human cells using anti-mouse CD45, anti-human CD45, CD3, CD19, and CD33 antibodies.\n\nStatistics {#Sec20}\n----------\n\nData were analyzed using student's t-test by GraphPad Prism 6 (GraphPad Software, Inc.). All data are from at least three independent experiments, unless otherwise specified. Results are expressed as the mean\u2009\u00b1\u2009SD or the mean\u2009\u00b1\u2009SEM of varying numbers in individual experiments. Statistical significance was defined as \\*P\u2009\\<\u20090.05; \\*\\*P\u2009\\<\u20090.01; \\*\\*\\*P\u2009\\<\u20090.001.\n\nData availability {#Sec21}\n-----------------\n\nThe authors declare that the data supporting the findings of this study are available within the paper and its supplementary information files or can be obtained from the authors upon reasonable request. The gene expression data reported in this paper have been deposited in the Gene Expression Omnibus (GEO) database, [www.ncbi.nlm.nih.gov/geo](http://www.ncbi.nlm.nih.gov/geo) (Reference super-series GSE90552).\n\nElectronic supplementary material\n=================================\n\n {#Sec22}\n\nSupplementary Information\n\nGuruchandar Arulmozhivarman, Martin Kr\u00e4ter, Malte von Bonin and Martin Bornh\u00e4user contributed equally to this work.\n\n**Electronic supplementary material**\n\n**Supplementary information** accompanies this paper at 10.1038/s41598-017-12360-0.\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nWe thank David Traver for the *Tg(c-myb:eGFP)* transgenic line used in this study; Leonard I. Zon, Harvard Stem Cell Program, Boston Children's Hospital, for supporting Nona Shayegi during her tenure as a visiting scientist in his laboratory; Naharajan lakshmanaperumal for assistance with Figures 8 and S4; Mathias Lesche for assistance with gene expression data analyses, and Dr. Vasuprada Iyengar for editing. The authors disclose receipt of the following financial support for conduct of this research, its authorship, and publication. This work was supported by the European Union (EU ERC Advanced grant 'Zf BrainReg' to Michael B, German Research Association (DFG) Cluster of Excellence 168, Centre for Regenerative Therapies Dresden ([www.crt-dresden.de](http://www.crt-dresden.de)) and the Collaborative Research Grant (SFB 655, Project B2 to MB and CW).\n\nG.A., M.K., M.v.B. and M.Bo. designed the project outline d and carried out most experiments, interpreted results and co-wrote the manuscript. M.W. provided reagents and supported study design and interpretation. J.F. conducted and performed atomic force microscopy-based single-cell force spectroscopy experiments and analyzed data. K.M. performed and analyzed colony-forming cell and cobblestone area-forming cell assays. K.L. performed adult zebrafish kidney marrow transplantation. D.A. and A.D. were responsible for mRNA-sequencing and bioinformatics analysis. M.S. and M.Bi. supported and performed the chemical screen in transgenic zebrafish embryos. E.P.B. provided intellectual contributions, interpreted results and assisted with animal experiments. J.H., N.S. and F.S. and M.Br. were involved in study design and interpreted results.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Introduction\n============\n\n\u03b1-Actinin is an actin-binding cytoskeletal protein present in a variety of cells \\[[@B1]\\] and in focal adhesion sites where cells adhere to the substrate \\[[@B2]\\]. There is biochemical \\[[@B3]\\] and histologic \\[[@B4]\\] evidence that focal adhesion complexes, containing \u03b1-actinin and other footpad material, are left behind as a result of normal movement of cells \\[[@B2]\\], perhaps at increased rates when neutrophils and monocytes move into inflammatory tissue. We have shown that \u03b1-actinin is abundant in the bone marrow stroma matrix, presumably at focal adhesion sites \\[[@B5]\\]. We have also reported that a 31 kDa amino-terminal \u03b1-actinin fragment, which we have named mactinin, is generated by the degradation of extracellular \u03b1-actinin by monocyte-secreted urokinase \\[[@B6]\\]. Furthermore, we have demonstrated that mactinin is present in inflammation caused by *Pneumocystis carinii*pneumonia, by examining bronchoalveolar lavage fluid from mice with infection \\[[@B6]\\]. It was not present in mice not challenged with *P. carinii*, suggesting that inflammaton is necessary for mactinin formation. We have also reported that mactinin promotes monocyte/macrophage maturation \\[[@B7]\\]. For example, \u03b1-actinin fragments significantly increase lysozyme secretion and tartate-resistant acid phosphatase staining in peripheral blood monocytes. In contrast, intact \u03b1-actinin has no maturation-promoting activity. We proposed that mactinin is present in the microenvironment at sites of various types of inflammation, perhaps owing to migrating cell populations, and there it might contribute to the recruitment and maturation of monocytes.\n\nMonocyte/macrophage infiltration has a key role in the pathogenesis of chronic arthritis \\[[@B8]\\]. The release of pro-inflammatory cytokines, chemokines, growth factors, and enzymes by the synovial lining macrophages is important for the onset, propagation, and flare of arthritic inflammation \\[[@B9]\\]. The finding that the number of synovial tissue macrophages is correlated with joint destruction in rheumatoid arthritis is evidence of their importance \\[[@B9],[@B10]\\]. Monocytes and macrophages are believed to have a similar role in other chronic inflammatory joint diseases, such as gout \\[[@B11]\\] and psoriatic arthropathy \\[[@B12]\\]. Therefore in this study we assessed the effects of mactinin on monocyte chemotaxis *in vitro*. We have also tested synovial fluid from patients with various types of arthritis, including rheumatoid arthritis, psoriatic arthritis, reactive arthritis, gout, and ankylosing spondylitis, for the presence of the monocyte/macrophage maturation-promoting fragment, mactinin. We have also investigated whether mactinin is present in the antigen-induced arthritis model in rabbits \\[[@B13],[@B14]\\]. Macrophages are believed to be important in this model of rheumatoid arthritis \\[[@B15],[@B16]\\], and both the arthritic and control joint fluid can be tested for mactinin.\n\nMaterials and methods\n=====================\n\nSource of mactinin\n------------------\n\nAs described previously \\[[@B6]\\], a pGEX2 vector, encoding the actin-binding domain, residues 2--269 of chicken smooth muscle \u03b1-actinin, fused with the carboxy terminus of glutathione S-transferase (GST) with an engineered thrombin cleavage site, was kindly provided by Dr DR Critchley of the University of Leicester, UK. Fusion protein was expressed in *Escherichia coli*, and the cleavage products of the fusion protein were purified by affinity chromatography of cell extracts on immobilized glutathione. The fusion protein was then cleaved with thrombin (Calbiochem, San Diego, CA) to yield the actin-binding domain of \u03b1-actinin and the GST carrier. The cleavage products were then separated by reverse-phase high-performance liquid chromatography on a C-4 column \\[[@B6]\\]. SDS--PAGE demonstrated that the \u03b1-actinin fragment was more than 90% of the total protein of pooled fractions, with the remaining 10% being carrier GST. The calculated molecular mass of this \u03b1-actinin fragment was 30,700 Da. In this report, both the active product of urokinase degradation of \u03b1-actinin formed *in vivo*\\[[@B6]\\] and the active recombinant actin-binding domain, which are of similar molecular masses, will be referred to as mactinin. Mactinin and GST routinely assay negative for protein endotoxin with a Pyrotell chromogenic assay kit, which can detect more than 0.25 endotoxin units/ml (Associates of Cape Cod, Woods Hole, MA).\n\nIsolation of peripheral blood monocytes\n---------------------------------------\n\nMononuclear cells were isolated from buffy coat preparations of healthy blood donors by density gradient centrifugation with Histopaque 1077 (Sigma). Contaminating red cells were then lysed in distilled water, and the sample was applied to an LS separation column with a magnetic monocyte isolation kit in accordance with the manufacturer\\'s instructions (Miltenyi Biotec, Auburn, CA). This negative selection method resulted in a cell population containing more than 90% monocytes as determined by CD14 expression.\n\nChemotaxis assay\n----------------\n\nCell migration was assessed by a 48-well micro-chemotaxis chamber (NeuroProbe, Gaithersburg, MD). An aliquot of peptide was placed in the lower compartment, and a suspension of monocytes (30,000--35,000) was placed in the upper compartment of the well. The two compartments were separated by a polyvinylpyrrolidone-free polycarbonate filter with a pore size of 5 \u03bcm. The chamber was incubated at 37\u00b0C for 90 minutes. At the end of the incubation period the filter was removed, fixed, and stained with a Hema 3 stain set (Fisher, Pittsburgh, PA). The cells that migrated through the membrane pore in three high-power fields (\u00d7400) were counted by light microscopy. Three chamber membranes were counted for each concentration.\n\nAssessment of mactinin concentrations necessary for HL-60 cell maturation\n-------------------------------------------------------------------------\n\nHL-60 myeloid leukemia cells were seeded at a density of 10^5^/ml and grown for 3 days in RPMI medium with 50 \u03bcg/ml gentamicin and 15% fetal calf serum at 37\u00b0C and 5% CO~2~, in the presence of various concentrations of recombinant mactinin. We have previously reported that mactinin promotes monocyte maturation as measured by morphology, non-specific esterase activity, and Fc rosette formation in this leukemia cell line \\[[@B7]\\]. Here we report the concentrations necessary to induce maturation as measured by non-specific esterase staining used as a maturation marker.\n\nAntisera generation\n-------------------\n\nTo generate antisera with sensitivity for detecting mactinin, purified recombinant chicken \u03b1-actinin peptide was modified by coupling with dinitrophenol \\[[@B17]\\] and injected into two New Zealand white rabbits along with complete Freund\\'s adjuvant, as described previously \\[[@B6]\\]. Boosts were done with peptide and incomplete Freund\\'s adjuvant. These animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of the Minneapolis Veterans Affairs Medical Center. Antisera were screened for their ability to detect the immunizing peptide and were immunoaffinity-purified with columns of the recombinant fragment covalently bound to a Affi-Gel 15 matrix (Bio-Rad, Hercules, CA). We expected cross-reactivity of the purified antisera with fragments from rats, mice, or humans because of the highly conserved amino acid sequence.\n\nSynovial fluid samples\n----------------------\n\nFluid from patients with arthritis undergoing therapeutic arthrocentesis was collected and tested for the presence of the fragment by Western blot analysis. The use of these fluid samples in this study was approved by the subcommittee on human studies of the Minneapolis VAMC. In brief, fresh samples were centrifuged at 895*g*for 10 minutes and frozen at -80\u00b0C until the time of analysis. Because mactinin is a very small fraction of the total protein content (more than 50 \u03bcg/10 \u03bcl) of the fluid, immunoaffinity purification of mactinin was performed before Western blotting. Samples were thawed, dialyzed against phosphate-buffered saline, and centrifuged again; the total sample volume was then applied to a column of immunoaffinity-purified antibody covalently bound to a Affi-Gel 15 matrix, and eluted with 0.1 M sodium citrate in 0.3 M NaCl, pH 3.0. Fractions (2 ml) were collected and neutralized with 2 M Tris-HCl. Protein-containing fractions were pooled, dialyzed against phosphate-buffered saline, concentrated, and subjected to electrophoresis on a 12% SDS--PAGE gel under reducing conditions. Each lane contained 100 \u03bcl, representing about 1% of the total sample.\n\nImmunoblot analysis\n-------------------\n\nThe proteins separated by SDS--PAGE were transferred electrophoretically to poly(vinylidene difluoride) membranes (Bio-Rad). The membrane was blocked with 5% nonfat milk in 50 M Tris-HCl/150 M NaCl, pH 7.5, and sequentially treated with affinity-purified rabbit antisera raised against recombinant mactinin, followed by second antibody conjugated with alkaline phosphatase (ICN, Costa Mesa, CA). Immunoreactive proteins were detected by alkaline phosphatase reaction with 5-bromo-4-chloro-3-indoyl phosphate/nitroblue tetrazolium. Control analyses were performed with rabbit IgG (Santa Cruz Biochemistry, Santa Cruz, CA).\n\nDissociation of immune complexes in rheumatoid arthritis fluid\n--------------------------------------------------------------\n\nTo examine whether mactinin was present in immune complexes in rheumatoid arthritis fluid, some samples were acidified by dialysis against 0.1 M sodium acetate, pH 4.1, to dissociate complexes \\[[@B18]\\]. The acidified samples were first fractionated on an 800 ml G-75 Sephadex (Pharmacia, Piscataway, NJ) size-exclusion column. Fractions identified by Western blotting as containing mactinin were pooled, neutralized, and further purified by fractionation on a C-4 column. The C-4 column was equilibrated with 0.1% trifluoroacetic acid and eluted with an acetonitrile gradient (0--100%) run at 1%/ml/min. Aliquots of protein-containing fractions were used for Western blots.\n\nMeasurement of mactinin in the antigen-induced arthritis model\n--------------------------------------------------------------\n\nTo produce antigen-induced arthritis, New Zealand white rabbits were immunized by subdermal injection of ovalbumin emulsified in complete Freund\\'s adjuvant in accordance with modifications \\[[@B14]\\] of the method of Dumonde and Glynn \\[[@B13]\\] under IACUC approval. Three weeks later, animals with positive skin tests to ovalbumin received intra-articular injections of 1 mg of sterile ovalbumin into one knee and an equal volume of sterile saline in the contralateral control knee weekly for 3 weeks. Arthritic and control knee joints were then lavaged with saline and aspirated at the time of killing, when animals had developed chronic synovitis 10 weeks after the intraarticular injections were complete. Samples were frozen at -80\u00b0C until the time of analysis. Samples were then concentrated and used for Western blots.\n\nResults\n=======\n\nMactinin is a chemoattractant for peripheral blood monocytes\n------------------------------------------------------------\n\nFor analysis of the effect of mactinin on monocyte chemotaxis, peripheral blood monocytes were placed in the upper chambers of a 48-well micro-chamber plate with various concentrations of mactinin. The lower compartment of the wells also contained various concentrations of mactinin and was separated from the upper chamber by a polycarbonate filter. As shown in Table [1](#T1){ref-type=\"table\"}, 1--10 nM mactinin had significant chemotactic activity for monocytes. Intact \u03b1-actinin at 10 nM had no activity. Because our mactinin preparations were contaminated with up to 10% GST (thus containing 0.1 nM GST in 1 nM mactinin), we tested 0.1 nM GST alone and also found no activity. The number of concentrations tested per assay was limited by the 48 wells (16 combinations for triplicate wells), but in separate assays we have seen significant chemotactic activity at mactinin concentrations as low as 0.5 nM. The concentration of mactinin necessary for activity is similar to that of FMLP (0.1--10 nM) in our assay system, and mactinin and FMLP attract similar numbers of monocytes.\n\nMeasurement of mactinin\\'s maturation-promoting activity\n--------------------------------------------------------\n\nRecombinant mactinin has maturation-promoting activity *in vitro*in HL-60 leukemia cells at the concentrations shown in Figure [1](#F1){ref-type=\"fig\"}. That is, it induces non-specific esterase staining at 2.5 pM and activity reaches a plateau at 25 pM (0.8 ng/ml). GST controls were run at a similar range of concentrations and showed no significant activity.\n\nDetection of mactinin in arthritis fluid\n----------------------------------------\n\nAffinity-purified rabbit antiserum raised against a recombinant chicken \u03b1-actinin fragment detected picogram amounts of the immunizing amino-terminal protein fragment (Fig. [2](#F2){ref-type=\"fig\"}). As also shown in Figure [2](#F2){ref-type=\"fig\"}, this antiserum reacted with a protein of the expected molecular mass in representative samples of immunoaffinity-purified synovial fluid from patients with psoriatic arthritis, reactive arthritis, gout, and ankylosing spondylitis. In all, six of seven samples from patients with these arthritides contained mactinin. In contrast, mactinin was detected in none of five rheumatoid arthritis samples (*P*\\< 0.05; Fisher\\'s exact test). When the detected levels of mactinin are corrected for the recovery rate by the isolation procedure, they are 1.7--15 ng/ml, and are in the range of maturation-promoting activity *in vitro*(\u2265 25 pM or 0.8 ng/ml) and the level needed for chemotactic activity (0.5 nM or 15 ng/ml). The lack of mactinin in rheumatoid arthritis fluid was surprising, and we decided to pursue this further.\n\nOwing to the autoimmune nature of rheumatoid arthritis, we examined whether mactinin was present in immune complexes. That is, antibody-bound mactinin might not bind to the antibody-matrix column used in the isolation protocol to decrease the total protein load, resulting in mactinin being undetectable but potentially active. To dissociate immune complexes, an aliquot of a rheumatoid arthritis fluid sample was acidified and the proteins were fractionated on a C-4 column. As shown in Figure [3](#F3){ref-type=\"fig\"}, mactinin is detectable by Western blot analysis after dissociation from immune complexes.\n\nTo confirm the presence of mactinin in rheumatoid arthritis fluids, some frozen aliquots were thawed and immediately subjected to electrophoresis on a 12% SDS--PAGE gel under reducing conditions. As shown in Figure [4](#F4){ref-type=\"fig\"}, affinity-purified mactinin antisera reacted with several 30--40 kDa proteins, including 31 kDa mactinin in each of the samples tested. We have previously demonstrated that only the 31 kDa \u03b1-actinin degradation product has maturation-promoting activity \\[[@B5]\\]. These proteins were not seen in non-reduced samples, which did have immunoreactive material at the top of the gel (data not shown). Hence, the detected proteins might represent mactinin bound to various immune complex fragments.\n\nMactinin is present in inflammatory fluid from antigen-induced arthritis\n------------------------------------------------------------------------\n\nArthritic and control knee joints of rabbits with chronic antigen-induced arthritis were lavaged with saline and aspirated at the time of killing. Samples were concentrated and used for Western blots run under reducing conditions. As seen in Figure [5](#F5){ref-type=\"fig\"}, mactinin is present in arthritis joint fluid but absent from control fluid, suggesting that mactinin formation is dependent on the inflammatory response. As also seen in Figure [5](#F5){ref-type=\"fig\"}, another \u03b1-actinin fragment of slightly higher molecular mass is present in both samples.\n\nDiscussion\n==========\n\nDuring inflammatory processes, various mediators, such as cytokines and chemokines, regulate the recruitment of monocytes. Once in the tissue, monocytes undergo the poorly understood process of transformation to macrophages with altered morphology and function \\[[@B19]\\]. In arthritis, synovial macrophages might cause joint destruction by differentiating to bone-resorbing osteoclasts \\[[@B20]\\] or by releasing cartilage-degrading enzymes and cytokines, such as interleukin-1 and tumor necrosis factor-\u03b1 \\[[@B8]\\]. It has therefore been suggested that therapies for chronic arthritis should be aimed at depleting joint mononuclear cells or controlling the activation of synovial macrophages \\[[@B21]\\]. Indeed, elimination of macrophages by clodronate-laden liposomes in rat models of adjuvant \\[[@B22]\\] and antigen-induced arthritis \\[[@B23]\\] induces amelioration of the arthritis.\n\nOf the many mediators of inflammation, mactinin is the first example of a fragment of a cytoskeletal component that might be released during leukocyte influx into inflammatory tissue. Further, mactinin might have a role in promoting the response of mononuclear phagocytes to inflammation. The monocyte functional studies *in vitro*demonstrate that 0.5--10 nM levels of the fragment have significant chemotactic activity. We have previously reported that mactinin promotes monocyte maturation, as measured by lysozyme secretion and tartrate-resistant acid phosphatase staining \\[[@B7]\\]. Here we show that 25 pM levels of mactinin promote monocytic maturation of the HL-60 leukemia cell line.\n\nMactinin is present at sites of various types of arthritic inflammation at levels that are active *in vitro*, including synovial fluid samples from patients with psoriatic arthritis, reactive arthritis, gout, and ankylosing spondylitis. Although it was not initially detected in five immunoaffinity-purified rheumatoid arthritis samples, it was detected after the acid dissociation of immune complexes. Girard and Senecal \\[[@B24]\\] have reported that sera from patients with autoimmune diseases such as rheumatoid arthritis contain antibodies against microfilament-associated proteins, including \u03b1-actinin. In addition, auto-antibodies against actin, vinculin, integrins, or fibronectin could also form complexes with mactinin \\[[@B24],[@B25]\\]. Our results suggest that mactinin is bound to immune complexes in rheumatoid arthritis joint fluid, which prevents its binding to the antibody-matrix during the isolation procedure. The finding that mactinin is detected by Western blotting of samples run under reducing conditions without immunoaffinity purification seems to confirm this. It is noteworthy that, unless the antibody is neutralizing, even antibody-bound mactinin might still be active.\n\nThe antigen-induced arthritis model of rheumatoid arthritis in rabbits demonstrates persistent active inflammation for several months after the intra-articular injection of antigen, including hypertrophy and hyperplasia of the synovial lining cells, pannus formation with articular cartilage erosion, and chronic infiltration of synovium by lymphocytes and plasma cells \\[[@B13]\\]. In addition, Dijkstra and colleagues \\[[@B15]\\] found macrophages in the superficial layer of the synovium, where they might secrete enzymes and oxygen radicals into the joint space, which can lead to cartilage erosion \\[[@B9]\\]. The protein levels in lavage fluid from arthritic joints in this model are low enough to allow direct testing by Western blot analysis without immunoaffinity purification, as was needed with the human aspirates. Mactinin was found in the lavage fluid from arthritic knee joints of rabbits with this immune arthritis and might contribute to macrophage function in this arthritis model.\n\nBecause this study was done using waste fluid from therapeutic synovial fluid aspirates, we did not have samples from noninflamed joints. The low mactinin recovery rate during the purification process and the low concentration of mactinin needed for activity make it necessary to assay at least 1 ml of fluid, and this amount is not available from any tissue and fluid bank. However, mactinin was not present in the control joint fluid in the rabbit antigen-induced arthritis model, suggesting that mactinin is specific for the inflammation process. Similarly, we have reported that bronchoalveolar lavage fluid from uninfected mice contains no mactinin, in contrast to fluid from mice infected with *P. carinii*\\[[@B6]\\].\n\nThe plasminogen activators, tissue type and urokinase type, have been reported to be both deleterious in inflammation, owing to the proteolysis of tissue proteins, and beneficial because of fibrinolytic activity \\[[@B26]\\]. The presence of both increased urokinase and plasmin inhibitors in rheumatoid arthritis synovial tissue suggests a complex role for urokinase in this disease \\[[@B27]-[@B29]\\] that seems pertinent to our finding of urokinase-generated mactinin in the arthritis fluid samples. The overall effect of urokinase in the antigen-induced arthritis model seems to be beneficial, because chronic joint inflammation and bone erosion are significantly worse in urokinase-deficient mice \\[[@B30]\\]. However, the ability of urokinase-generated mactinin to enhance proteolysis might be deleterious. Hence, future testing of specific mactinin inhibitors in animal models of arthritis seems warranted.\n\nConclusion\n==========\n\nWe conclude that mactinin is present in arthritic synovial fluid in levels that can promote mononuclear phagocyte chemotaxis and maturation. There, increased numbers of mature monocytes might increase cartilage and bone destruction. These results lead us to speculate that inhibitors of mactinin might be of benefit in the treatment of some forms of chronic arthritis and form the basis for our plans to test the efficacy of mactinin antisera to ameliorate antigen-induced arthritis.\n\nCompeting interests\n===================\n\nNone declared.\n\nAbbreviations\n=============\n\nGST = glutathione S-transferase; IACUC = Institutional Animal Care and Use Committee.\n\nAcknowledgements\n================\n\nThis study was supported by the Department of Veterans Affairs.\n\nFigures and Tables\n==================\n\n![Percentage of HL-60 cells staining positive for nonspecific esterase after treatment with various concentrations of recombinant mactinin. Cells were incubated for 3 days with mactinin, then harvested and stained. The percentage of untreated HL-60 cells positive for staining was subtracted. Each value is the mean \u00b1 SD for a minimum of two assays of 100 cells each. The result of treatment with 100 nM 12-*O*-tetradecanoylphorbol-13-acetate is also shown (circle).](ar799-1){#F1}\n\n![Western blot analysis with affinity-purified rabbit antisera. Each of the first seven lanes contains the immunizing peptide in the amount shown (in nanograms). The second band seems to be due to an alternative cleavage site in the fusion protein at amino acid 262. Lanes A--E contain synovial fluid from patients with various types of arthritis; it was immunoaffinity-purified to decrease the protein load before Western blotting. Lane A, psoriatic arthritis; lane B, reactive arthritis; lane C, gout; lane D, ankylosing spondylitis; lane E, rheumatoid arthritis. The samples in lanes A--D contained mactinin. Controls with rabbit IgG were negative for all samples.](ar799-2){#F2}\n\n![Dissociation of immune complexes in rheumatoid arthritis fluid. An aliquot of rheumatoid arthritis fluid was acidified to dissociate immune complexes, then fractionated on a C-4 column before Western immunoblotting (lane A). Another aliquot of the same sample (lane B) was immunoaffinity-purified as in figure [2](#F2){ref-type=\"fig\"} but was not subjected to immune complex dissociation.](ar799-3){#F3}\n\n![Western blot analysis of rheumatoid arthritis fluid. Synovial fluid (10 \u03bcl) from two patients with rheumatoid arthritis was subjected to Western blot analysis under reducing conditions.](ar799-4){#F4}\n\n![Western blot analysis of synovial fluid of rabbits with antigen-induced arthritis. Affinity-purified rabbit antisera raised against recombinant mactinin was used to detect the immunizing protein in the amounts shown in the first four lanes and mactinin from arthritic joint fluid in lane A. Control joint fluid is shown in lane B.](ar799-5){#F5}\n\n###### \n\n\\'Checkerboard\\' analysis of mactinin as a chemotactic factor\n\n        Mactinin concentration above membrane (nM)                          \n  ----- -------------------------------------------- ----------- ---------- ---------\n  0     21 \u00b1 7                                       25 \u00b1 8      17 \u00b1 6     32 \u00b1 10\n  0.1   25 \u00b1 8                                       39 \u00b1 12     21 \u00b1 9     29 \u00b1 5\n  1     70 \u00b1 7\\*                                     77 \u00b1 15\\*   48 \u00b1 9\\*   28 \u00b1 6\n  10    78 \u00b1 7\\*                                     82 \u00b1 19\\*   NT         26 \u00b1 7\n\nDifferent concentrations of mactinin in the upper and lower compartments of chemotactic chambers define the \\'checkerboard\\' analysis of mactinin as a chemotactic or chemokinetic factor. Results (means \u00b1 SEM) are the average number of migrated cells per oil field (counting three fields) from three filters. Significant results, compared with controls with no mactinin below the membrane and either an equivalent amount of mactinin or no mactinin above the membrane by Student\\'s *t*-test, are indicated by asterisks (*P*\\< 0.05). This represents one of two experiments with similar results. Neither 10 nM intact \u03b1-actinin nor 0.1 nM GST had significant activity. NT, not tested.\n"}
+{"text": "Introduction\n============\n\nLung cancer is a respiratory disease that is responsible for the highest rates of cancer-associated mortality and air contamination worldwide ([@b1-etm-0-0-6274]). Small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) are two main types of human lung cancer that are characterized by tumor morphology ([@b2-etm-0-0-6274]), and they account for \\~95% of all lung cancer ([@b3-etm-0-0-6274]). NSCLC which has the highest incidence among all kinds of cancer originates from non-small cells in the lungs ([@b4-etm-0-0-6274]). NSCLC, which is also the most frequent type of lung cancer, can be divided into squamous cell carcinoma, large cell carcinoma and adenocarcinoma determined by tumor cell genetics. According to the clinical statistics investigations of lung cancer cases \\>80% of newly diagnosed NSCLC patients were in middle or severe stage ([@b5-etm-0-0-6274]--[@b7-etm-0-0-6274]).\n\nThough increasing research has endeavored to improve the efficacy of treatment for patients with NSCLC, the survival rate remained poor, with \\<15% survival observed in the 5 years after clinical treatment ([@b6-etm-0-0-6274],[@b8-etm-0-0-6274],[@b9-etm-0-0-6274]). In addition, the majority of newly diagnosed patients with NSCLC are in the advanced stage. Previous research has reported that migration and invasion in NSCLC are predominantly responsible for the poor survival rate during treatment and recurrence for patients with NSCLC ([@b10-etm-0-0-6274],[@b11-etm-0-0-6274]). As a result, the exploration for effective agents for the inhibition of migration and invasion has become critically important for the treatment of cancer patients ([@b12-etm-0-0-6274],[@b13-etm-0-0-6274]). The present study investigated H358 NSCLC cell migration and invasion. Furthermore, the inhibitory effects of anti-cysteine-rich angiogenic inducer-61 (CYR-61) on H358 NSCLC cell migration and invasion were analyzed *in vitro* and *in vivo*.\n\nCYR-61 is a member of the CYR-61/connective tissue growth factor/nephroblastoma overexpressed (CCN) protein family ([@b14-etm-0-0-6274]). Previous studies have demonstrated that CYR-61 promotes human lung cancer cell migration and metastasis and it is closely related to patient prognosis in NSCLC ([@b15-etm-0-0-6274],[@b16-etm-0-0-6274]). In addition, CYR-61 is involved in tumor cell mitogenesis, cellular adhesion, migration, differentiation, wound healing, angiogenesis and survival ([@b17-etm-0-0-6274]). Previous reports have demonstrated the important roles of CYR-61 in cancer development and metastasis, indicating that CYR-61 may be an important target for gene therapy and tumor suppression (16, 17). In addition, CYR61 has been demonstrated to induce angiogenesis by supplying oxygen for tumor cells during proliferation ([@b18-etm-0-0-6274]).\n\nMany studies have focused on the functions of the CCN protein family in cancer biology ([@b17-etm-0-0-6274],[@b19-etm-0-0-6274]). A study by Barnett *et al* ([@b7-etm-0-0-6274]) reported that CYR-61 demonstrated potential as an oncogene or a tumor suppressor, depending on tumor cell type. Clinically, expression of CYR-61 has been associated with the prognosis of breast cancer and prostate cancer ([@b20-etm-0-0-6274]). However, few studies have investigated the function of CYR-61 in NSCLC. Therefore, the present study investigated the expression of CYR-61 in NSCLC cells and tumors. Results indicated that CYR-61 was expressed at higher levels in NSCLC cells, when compared with normal lung cells of MRC-5. Furthermore, an antibody against CYR-61 (anti-CYR-61) was constructed and its therapeutic effects in mice with NSCLC were investigated.\n\nRecently, numerous studies have indicated that mechanistic target of rapamycin (mTOR) may regulate tumor cell growth, migration and cancer metastasis ([@b21-etm-0-0-6274],[@b22-etm-0-0-6274]). Epithelial-mesenchymal transition (EMT) has an essential role in tumor growth, migration and cancer metastasis. In addition, the EMT process reduces tumor cell adhesion and results in tumor cells gaining migratory and invasive properties through cell-cell connections ([@b23-etm-0-0-6274]). Previous research has indicated that CYR-61 is associated with NSCLC migration and cancer metastasis ([@b17-etm-0-0-6274]). However, little is known about the signaling mechanisms regulating mTOR, CYR-61 and EMT in NSCLC. Therefore, the present study examined the association between CYR-61 and EMT in NSCLC cells. EMT biomarker expression levels of vimentin, fibronectin, \u03b1-smooth muscle actin (SMA) and N-cadherin were analyzed. Mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT)/mTOR signaling pathways in EMT were also investigated *in vitro* and *in vivo* in NSCLC cells and tissues, respectively.\n\nThe aim of the present study was to determine the effects of anti-CYR-61 on CYR-61-associated invasion and metastasis in NSCLC through MAPK/EMT signaling pathways. It was concluded that CYR-61 may be considered as a potential prognostic biomarker for NSCLC, and anti-CYR-61 may provide a potential minimally invasive therapy for NSCLC.\n\nMaterials and methods\n=====================\n\n### Ethics statement\n\nThe present study was carried out in strict accordance with the approval and recommendations from the Ethics Committee of the Care and Use of Laboratory Animals of Qilu Hospital of Shandong University (Jinan, China). All surgery and euthanasia were performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.\n\n### Cell culture\n\nThe H358 NSCLC cell line and MRC-5 normal lung cell line were purchased from American Type Culture Collection (Manassas, VA, USA). The cell lines were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal bovine serum (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) at 37\u00b0C, 5% CO~2~ and 100% humidity.\n\n### Construction of full-length anti-CYR-61 antibody\n\nA mouse anti-human CYR-61 monoclonal antibody was constructed using a conventional approach and screened by fluorescence-activated cell sorting (FACS). The full length of the anti-CYR-61 antibody was constructed, as previously described ([@b24-etm-0-0-6274]). The single chain variable fragments of the mouse anti-human CYR-61 monoclonal antibody (Sino Biological, Beijing, China) were cloned and inserted into a Pklight vector (termed Pklight-anti-CYR-61 vector; Biovector NTCC, Inc., Beijing, China). The constant domain heavy chain Fc and light chain fragments of mouse anti-human CYR-61 monoclonal antibody were subcloned into the Pklight-anti-CYR-61 vector. Pklight-anti-human CYR-61 monoclonal antibody and IREX-enhanced green fluorescent protein (EGFP) were subcloned into the Peedual12.4 vector (BioVector NTCC, Inc.), which contained the glutamine synthetase gene. The CHO-K1SV cell line (American Type Culture Collection) was cultured in Iscove\\'s modified Dulbecco\\'s medium (Sigma-Aldrich; Merck KGaA) supplemented with 10% fetal bovine serum and 2 mM L-glutamine. Peedual12.4-anti-human CYR-61 monoclonal antibody was transfected into the fluorescein isothiocyanate (Shanghai Xinyu Biotechnology Pharmaceutical Co., Ltd.)-labeled CHO-K1SV cells (1\u00d710^5^ cells/ml) using Lipofectamine 2000 (Tiangen Biotech Co., Ltd., Beijing, China). CHO-K1SV cells were washed and resuspended in 0.01 mol/l pH=7.4 PBS twice. The cells were identified and sorted in a flow cytometer at 488 nm. Thus, anti-human CYR-61 monoclonal antibody and EGFP were stably expressed in the CHO-K1SV cells. Stable and high expression of anti-CYR-61 antibodies in bacterial cells was screened using FACS (BD Biosciences, Franklin Lakes, NJ, USA). Data analysis and statistics were performed using BD Accuri^\u2122^ C6 Plus (BD Biosciences) and GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA).\n\n### MTT assay\n\nA total of 3,000 of H358 cells were cultured into each well of a 96-well plate and cells were used to investigate the inhibitory effects of anti-CYR-61 on cell viability when \\~90% cell confluence was reached. Anti-CYR-61 (1:1,000) or 30 \u00b5l PBS (control) were added into each well of the 96-well plate and incubated at 37\u00b0C for 12 h. Subsequently, 10 \u00b5l MTT (5 mg/ml; Sigma-Aldrich; Merck KGaA) was added to the cells and incubated at 37\u00b0C for 4 h. Following this, dimethylsulfoxide (Amresco, LLC, Solon, OH, USA) was added for incubation for 30 min to dissolve the precipitate after the supernatant had been removed. Results were determined using a spectrophotometer (Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 540 nm.\n\n### Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)\n\nTotal RNA was extracted from H358 cells and a normal lung cell line MRC-5 cells with or without treatment of anti-CYR-61 using an RNeasy mini kit (Qiagen Sciences, Inc., Gaithersburg, MD, USA), according to the manufacturer\\'s protocol. RNA (1.0 \u00b5g) was reverse transcribed into cDNA using QuantiTect Reverse Transcription kit (Qiagen Sciences, Inc.), according to the manufacturer\\'s protocol. The primers ([Table I](#tI-etm-0-0-6274){ref-type=\"table\"}) were designed using Primer Express software (version 2.0; Thermo Fisher Scientific, Inc.) qPCR analysis was performed using the SYBR^\u00ae^ *Premix*^\u2122^ *Ex Taq*^\u2122^ (Perfect Real Time; Takara Biotechnology Co., Ltd., Dalian, China) in a total volume of 20 \u00b5l using a 7300 Real-Time PCR System (Thermo Fisher Scientific, Inc.), according to the manufacturer\\'s protocol. The thermocycling conditions were as follows: 95\u00b0C for 30 sec, and 40 cycles of 95\u00b0C for 5 sec and 60\u00b0C for 30 sec. All relative mRNA expression levels were calculated using the by 2^\u2212\u0394\u0394Cq^ method ([@b25-etm-0-0-6274]). Results were expressed as the n-fold relative to the housekeeping gene, \u03b2-actin.\n\n### ELISA\n\nThe affinity of anti-CYR-61 was examined for its target antigens, CYR-61, using ELISA in H358 cells. CYR-61 (0.2--1.4 mg/ml) was added into an enzyme-linked-immuno microplate and incubated at 4\u00b0C for 12 h. Anti-CYR-61 (4 \u00b5g/ml) was added to the wells and incubated for 60 min at 37\u00b0C, and bovine serum albumin (BSA; Atlanta Biologicals, Inc., Flowery Branch, GA, USA) was used as a control. Subsequently, 100 \u00b5l human horseradish peroxidase-conjugated CYR-61 antibodies (1:1,000; eBioscience, Inc.; Thermo Fisher Scientific, Inc.) was added and incubated at 37\u00b0C for 60 min. The 3,3\u2032-diaminobenzidine/H~2~O~2~ system was used for the detection of anti-CYR-61 affinity. Results were analyzed at 450 nm using an ELISA plate reader (Bio-Rad Laboratories, Inc.).\n\n### Cell invasion and migration assays\n\nH358 cells were treated with anti-CYR-61 and non-treated H358 cells were used as control. H358 cells were adjusted to a density of 1\u00d710^6^ cells in 500 \u00b5l serum-free RPMI-1640 medium for the invasion assay. H358 cells were treated with anti-CYR-61 (1:1,000) for 12 h at room temperature and then added to the tops of BD BioCoat Matrigel Invasion Chambers (BD Biosciences), according to the manufacturer\\'s protocol. Transwell chambers (Costar; Corning Incorporated, Corning, NY, USA) with 8 \u00b5m diameter pores were utilized. Matrigel (100 \u00b5l; BD Biosciences, San Jose, CA, USA) was added to the Transwell apparatus. A total of 106 H358 cells in 100 \u00b5l serum-free culture medium (Gibco; Thermo Fisher Scientific, Inc.) was placed in the upper chamber and 500 \u00b5l complete culture medium containing 20% fetal bovine serum (Sigma-Aldrich; Merck KGaA) was added to the lower chamber as a chemoattractant. Following incubation at 37\u00b0C for 24 h, Transwell chambers were stained with 0.4% crystal violet for 5 min at room temperature and washed with PBS three times. Non-migrating and non-invading cells were carefully wiped from the upper chambers with cotton wool. Results were examined using a CX21 Olympus light microscope (Olympus Corporation, Tokyo, Japan; magnification, \u00d7100).\n\nFor the migration assay, H358 (1\u00d710^4^) cells were inoculated with anti-CYR-61 (1:1,000) for 12 h at room temperature and Control inserts (BD Biosciences) were used instead of a Matrigel Invasion Chamber. Tumor cell invasion and migration were observed in at least three stained fields in every membrane using a light microscope at a magnification of \u00d7100. The invasion and migration assays used 24 well dishes.\n\n### Apoptosis assay\n\nApoptosis was determined by staining H358 cells with Annexin V-PE/ and 7-aminoactinomycin (BD Biosciences) for 15 min at 25\u00b0C in the dark, and flow cytometry analysis was performed. The Annexin V-positive cells were counted as early apoptotic cells. 7-amino actinomycin-positive cells were counted as necrotic cells. Double positive cells were counted as late apoptotic cells. Double negative cells were counted as live cells.\n\n### Animal study\n\nA total of 50 female specific pathogen-free C57BL/6 mice (aged 6--8 weeks; weighing 20\u00b12 g) were purchased from Shanghai Slack Experimental Animals Co., Ltd., (Shanghai, China). The mice were given free accessible to food and water, and were housed at 20\u00b0C with 60% humidity and 12 h light/dark cycle. C57BL/6 mice were subcutaneously implanted with H358 tumor cells and were divided into two groups (20 per group). Treatments were initiated on day 5 after tumor implantation when the tumor diameter had reached 6--8 mm. H358-bearing mice were intravenously injected with anti-CYR-61 or 10 \u00b5l PBS as a control. The treatment was administered once daily for 14 days. Tumor volumes were calculated according to previous study ([@b26-etm-0-0-6274]). The survival rate for animals treated with anti-CYR-61 (1:1,000) or 10 \u00b5l PBS was detected during a 120-day observation period.\n\n### Western blotting\n\nTotal protein was extracted using CytoBuster^\u2122^ Protein Extraction Reagent (EMD Millipore, Billerica, MA, USA) and protein concentration was measured by the *DC*^\u2122^ Protein Assay (Bio-Rad Laboratories, Inc.) using the Bradford method ([@b27-etm-0-0-6274]). A total of 40 \u00b5g of protein from each sample were separated by 10% SDS-PAGE and transferred to an equilibrated polyvinylidene difluoride membrane (GE Healthcare, Chicago, IL, USA). The membrane was blocked by 5% milk in 0.1% Tween 20 in Tris-buffer solution (TBS) at room temperature for 1 h. After incubation with specific primary antibody at 4\u00b0C overnight and the membranes were washed four times with TBS, 10 min each time. The membranes were then incubated with horseradish peroxidase-conjugated secondary antibodies (1:800; cat. no. ab7090; Abcam, Cambridge, MA, USA) at 37\u00b0C for 30 min and the proteins were detected by enhanced chemiluminescence (GE Healthcare) and quantified using an image analyzer Quantity One System (Bio-Rad Laboratories, Inc.). The primary antibodies directed against CYR-61 (cat. no. ab24448), extracellular signal-regulated kinase (cat. no. ab54230), AKT (cat. no. ab8805), phosphorylated (p)ERK (cat. no. ab79483), pAKT (cat. no. ab8933), fibronectin (cat. no. ab23750), SMA (cat. no. ab21027) and N-cadherin (cat. no. ab98952) were purchased from Abcam (Cambridge, UK) all at a dilution of 1:500 and were incubated at 4\u00b0C overnight. Primary antibodies directed against Caspase 8 (1;800; cat. no, sc81656), 9 (1;800; cat. no. sc56076) and 10 (1;800; cat. no. sc134299), Fas ligand (FasL; 1;800; cat. no. sc33716), Fas-associated protein with Death Domain (FADD; 1:800; cat. no. sc5559), apoptotic protease activating factor 1 (Apaf-1; 1:800; cat. no. sc135836), B cell lymphoma-2 antagonist/killer (Bak; 1:800; cat. no. sc517390), B cell lymphoma-2-associated X protein (Bax; 1:800; cat. no. sc20067), vimentin (1:800; cat. no. sc80975) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). The housekeeping proteins \u03b2-actin (1:500; cat. no. 3700) and GAPDH (1:500; cat. no. 97166) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Density analysis was performed using Image J v.1.49 (National Institutes of Health, Bethesda, MD, USA).\n\n### Histological immunostaining\n\nFor immunostaining, H358 cells or tumors from xenograft mice with NSCLC were fixed using 10% formaldehyde at room temperature for 24 h and subsequently embedded in paraffin. Following this, tumor samples were sliced into 4-\u00b5m-thick sections and antigen retrieval was also performed with 0.1 M citrate buffer for 15 min at 100\u00b0C, rehydrated in a descending series of ethanol (100, 95, 85, 80 and 75%) watched in xylene in tumor sections. After blocking by 10% BSA at room temperature for 1 h, H358 cells and tumor sections were incubated with pERK (cat. no. ab192591), pAKT (cat. no. ab38449) and CYR-61 (cat. no. ab24448; all 1:600; Abcam) primary antibodies at 37\u00b0C for 2 h. Following washing three times with TBST, HRP-conjugated goat anti-rabbit Immunoglobulin G secondary antibodies (1:800; cat. no. ab97051; Abcam) were incubated 30 min at 37\u00b0C. Specimens were then visualized using a binocular light microscope (Eclipse E100-LED; Nikon Corporation, Tokyo, Japan) at a magnification of \u00d7200. A Ventana Benchmark automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA) was used for observation of CYR-61.\n\n### Statistical analysis\n\nAll data were presented as the mean \u00b1 standard deviation of triplicates. Unpaired data were compared using Student\\'s t-tests and comparisons of data between multiple groups were made using one-way analysis of variance with a post-hoc Tukey\\'s test using SPSS 17.0 software (SPSS, Inc., Chicago, IL, USA). Kaplan-Meier tests were used to estimate the survival rate during 120-day long-term treatment. P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\n### CYR-61 expression and anti-CYR-61 characteristics\n\nIn order to analyze the role of CYR-61 in NSCLC, mRNA and protein expression levels of CYR-61 were detected in the NSCLC cell line H358 and a normal lung cell line MRC-5 by RT-qPCR analysis and western blotting. The results demonstrated that CYR-61 mRNA expression levels were significantly higher (P\\<0.01) and protein expression levels were markedly higher in the H358 cell line compared with the levels in the MRC-5 cell line ([Fig. 1A](#f1-etm-0-0-6274){ref-type=\"fig\"}). These findings suggested that CYR-61 was a potential molecular target for NSCLC. Therefore, a full-length antibody was constructed to target CYR-61, termed anti-CYR-61, which was screened by flow cytometry. The population of CYR-61 positive cells began to reduce 1 h after anti-CYR-61 treatment, and reduced to its lowest level 3 h after the treatment ([Fig. 1B](#f1-etm-0-0-6274){ref-type=\"fig\"}). An ELISA assay of anti-CYR-61 demonstrated a high affinity with CYR-61 ([Fig. 1C](#f1-etm-0-0-6274){ref-type=\"fig\"}). Western blotting revealed that anti-CYR-61 was 76 kDa under non-reducing conditions ([Fig. 1D](#f1-etm-0-0-6274){ref-type=\"fig\"}). Western blotting also demonstrated that anti-CYR-61 was able to specifically bind with CYR-61. These results indicated that CYR-61 was upregulated in NSCLC cell lines and that the anti-CYR-61 antibodies we constructed exhibited high affinity for CYR-61.\n\n### Effect of anti-CYR-61 on CYR-61-induced NSCLC viability, migration and invasion\n\nTo identify the function of CYR-61 in NSCLC, we investigated the inhibitory effects of anti-CYR-61 on cell viability, migration and invasion of H358 cells. As demonstrated in [Fig. 2A and B](#f2-etm-0-0-6274){ref-type=\"fig\"}, H358 cell viability was significantly inhibited following treatment with anti-CYR-61, which decreased CYR-61 expression (P\\<0.01). In migration and invasion assays, anti-CYR-61 treatment inhibited the migration and invasion of H358 cells ([Fig. 2C and D](#f2-etm-0-0-6274){ref-type=\"fig\"}). These results indicated that anti-CYR-61 may have potential therapeutic effects by inhibiting NSCLC cell viability, migration and invasion by targeting CYR-61.\n\n### Effect of anti-CYR-61 on NSCLC cell migration and the AKT and ERK signaling pathways\n\nTo further investigate the effects and mechanism of anti-CYR-61 on cell migration, we analyzed ERK and AKT expression and phosphorylation levels in H358 cells prior to and following treatment with anti-CYR-61. As demonstrated in [Fig. 3A and B](#f3-etm-0-0-6274){ref-type=\"fig\"}, anti-CYR-61 treatment significantly suppressed AKT and ERK protein expression and phosphorylation compared with untreated controls (P\\<0.01), indicating a reduction in AKT and ERK activity in NSCLC cells treated with anti-CYR-61. Immunofluorescence also indicated that expression and phosphorylation levels of AKT and ERK were inhibited in H358 cells treated with anti-CYR-61 ([Fig. 3C and D](#f3-etm-0-0-6274){ref-type=\"fig\"}). These results suggested that the inhibitory effect of anti-CYR-61 on migration may involve AKT and ERK phosphorylation and the ERK and AKT signaling pathways.\n\n### Role of anti-CYR-61 on cell apoptosis\n\nAs the results demonstrated that anti-CYR-61 regulated NSCLC cell migration through phosphorylation of ERK and AKT process, the apoptotic effects of anti-CYR-61 on H358 cells by blocking the CYR-61 were analyzed. NSCLC H358 cells were treated with or without anti-CYR-61. Results demonstrated that anti-CYR-61 significantly promoted H358 cell apoptosis compared with the control cells (P\\<0.01; [Fig. 4A](#f4-etm-0-0-6274){ref-type=\"fig\"}). In addition, the relationship between anti-CYR-61 and apoptosis signaling pathways in H358 cells was investigated. Results indicated that FasL, FADD, caspase-8 and \u221210 mRNA expression levels were significantly elevated by anti-CYR-61 treatment compared with the control cells (P\\<0.01; [Fig. 4B](#f4-etm-0-0-6274){ref-type=\"fig\"}). Western blot analysis also revealed that FasL, FADD, caspase-8 and \u221210 protein expression levels were markedly upregulated following treatment with anti-CYR-61 compared with control cells ([Fig. 4C](#f4-etm-0-0-6274){ref-type=\"fig\"}). However, no significant difference in the expression levels of Bak, Bax, Apaf-1 and caspase-9 were observed between anti-CYR-61-treated cells and the control cells ([Fig. 4D and E](#f4-etm-0-0-6274){ref-type=\"fig\"}). These results suggested that anti-CYR-61 has an important role in regulating exogenous cell apoptosis signaling pathways for NSCLC cells.\n\n### Anti-tumor effects of anti-CYR-61 in mice with NSCLC\n\nFollowing observation of the inhibitory effects of anti-CYR-61 on NSCLC cell viability *in vitro*, the anti-tumor efficacy of anti-CYR-61 in H358-bearing mice was investigated *in vivo*. As illustrated in [Fig. 5A](#f5-etm-0-0-6274){ref-type=\"fig\"}, tumor growth was significantly inhibited 21 days after anti-CYR-61 treatment, as determined via tumor volume, compared with PBS-treated mice (P\\<0.01). In addition, immunohistology demonstrated that CYR-61 expression was downregulated in anti-CYR-61-treated tumors compared with PBS-treated tumors ([Fig. 5B](#f5-etm-0-0-6274){ref-type=\"fig\"}). In agreement with the *in vitro* results, the findings demonstrated that AKT and ERK expression were decreased in tumors from experimental mice treated with anti-CYR-61 on day 25 compared with control mice treated with PBS ([Fig. 5C](#f5-etm-0-0-6274){ref-type=\"fig\"}). To analyze the relationship between anti-CYR-61 and the EMT process *in vivo*, EMT biomarker expression levels of vimentin, fibronectin, SMA and N-cadherin were analyzed. The results in [Fig. 5D](#f5-etm-0-0-6274){ref-type=\"fig\"} demonstrated that the protein expression levels of EMT markers were decreased in anti-CYR-61-treated tumors compared with PBS-treated tumors. Furthermore, it was observed that the survival rate of Anti-CYR-61 treatment is significantly higher than that of PBS treatment after day 40 during a 120-day observation period (P\\<0.01; [Fig. 5E](#f5-etm-0-0-6274){ref-type=\"fig\"}). Anti-CYR-61 treatment significantly inhibited tumor metastasis compared with PBS-treated mice (P\\<0.01; [Fig. 5E and F](#f5-etm-0-0-6274){ref-type=\"fig\"}). In conclusion, our findings suggested that anti-CYR-61 presented potential anti-cancer efficacy in NSCLC treatment in murine model.\n\nDiscussion\n==========\n\nThe incidence and mortality rate of human lung cancer has been growing rapidly in recent years and is one of the most threatening types of malignant tumors to health and survival ([@b28-etm-0-0-6274]). NSCLC represents more than 85% of lung cancer cases according to statistical clinical data ([@b29-etm-0-0-6274]). Therefore, NSCLC has attracted much attention to identify anti-cancer agents, ranging from molecular markers to immunotherapy, in order to improve the treatment and prognosis of patients with NSCLC. Notably, the majority of newly diagnosed NSCLC cases are often in the moderate or severe stage, decreasing the recovery probability and survival period ([@b30-etm-0-0-6274]). Therapeutic protocols for advanced NSCLC have expanded to include a number of targeted interventions, including chemotherapy, radiotherapy, small molecule target therapy, personalized treatment and immunotherapy ([@b31-etm-0-0-6274]). Various types of treatment for NSCLC have been introduced with different mechanisms of action; they have been shown to change tumor architecture and the tumor microenvironment, however the outcomes were not promising as an early reduction in tumor mass was not achieved ([@b32-etm-0-0-6274]). The reasons of treatment failure were predominantly that these treatments failed to control NSCLC migration and inhibit invasion in patients during treatment periods.\n\nCYR-61 is a secreted protein of the CCN family that is associated with the extracellular matrix signaling pathway. Previous research has indicated that CYR-61 is potent in regulating tumor cell activities, such as tumor cell growth, apoptosis, proliferation, migration, adhesion, differentiation and the EMT process, in the majority of human cancer cells ([@b17-etm-0-0-6274]). A study by Sabile *et al* ([@b33-etm-0-0-6274]) reported that CYR-61 signaling was regulated by phosphorylation of AKT and ERK in osteosarcoma tumor and lung cancer cells. In addition, a study by Chen *et al* ([@b34-etm-0-0-6274]) described that phosphorylation of AKT and ERK was significantly associated with the migration and invasion of prostate carcinoma PC-3 cells. In the present study, a full-length antibody target for CYR-61 was constructed and its anti-tumor efficacy in a murine model of lung cancer was investigated. In agreement with results from a previous report ([@b35-etm-0-0-6274]), the present study demonstrated that anti-CYR-61 treatment inhibited phosphorylation of the AKT and ERK signaling pathway in human NSCLC cells. Furthermore, it was demonstrated that anti-CYR-61 inhibited tumor growth and downregulated EMT biomarker expression of vimentin, fibronectin, SMA and N-cadherin.\n\nIn a previous report, CYR-61 expression was demonstrated to be downregulated via inhibition of the ERK and AKT pathways, resulting in suppression of colon cancer cell migration ([@b19-etm-0-0-6274]). A study by Lee *et al* ([@b36-etm-0-0-6274]) suggested that activities of CYR-61 protein were modulated through extracellular acidification and the PI3K/AKT signaling pathway in prostate carcinoma cells. Furthermore, Chen *et al* ([@b34-etm-0-0-6274]) indicated that transforming growth factor-\u03b2 induced CYR-61 production to enhance tumor cell migration and invasion *in vitro* and *in vivo*. These observations were supported in the present study in NSCLC cells and the results of the present study also suggested that inhibition of CYR-61 production is beneficial for inhibiting NSCLC cell viability, migration and invasion via the AKT and ERK signaling pathways.\n\nIn the present study, CYR-61 production in human NSCLC cells and normal lung cells was investigated. The critical role of the CYR-61 signaling pathway for the EMT process was confirmed in NSCLC H358 cells. Previous research has reported that activation of ERK and the PI3K/AKT signal pathway has an important role in regulation of tumor cell migration ([@b37-etm-0-0-6274]). In the present study, antibody targeting of CYR-61 not only inhibited viability of NSCLC H358 cells, it also suppressed migration of NSCLC H358 cells *in vitro* and *in vivo*.\n\nIn conclusion, the aim of the present study was to investigate the association between CYR-61 and the prognosis of NSCLC in a murine model. Anti-CYR-61was demonstrated to be a potential anti-tumor agent for NSCLC by targeting CYR-61, which led to inhibition of migration by decreasing the production of CYR-61 via suppressing phosphorylation of AKT and ERK. Taken together, the therapeutic efficacy of anti-CYR-61 was examined and the results suggested that this targeted strategy may represent an attractive method of inhibiting tumor cell migration for the treatment of NSCLC.\n\nNot applicable.\n\nFunding\n=======\n\nNo funding was received.\n\nAvailability of data and materials\n==================================\n\nThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\n\nAuthors\\' contributions\n=======================\n\nXL and YQ constructed and performed the cell invasion and migration assays, and were major contributors in writing the manuscript. NY analyzed and interpreted the RT-qPCR, western blotting and histological immunostaining data. LL performed and analyzed the animal study and ELISA assay. LY performed and analyzed the MTT and apoptosis assays.\n\nEthics approval and consent to participate\n==========================================\n\nThe present study was carried out in strict accordance with the approval and recommendations from the Ethics Committee of the Care and Use of Laboratory Animals of Qilu Hospital of Shandong University (Jinan, China). All surgery and euthanasia were performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.\n\nConsent for publication\n=======================\n\nNot applicable.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\n![Expression of CYR-61 in NSCLC and normal lung cell line and the characteristics of anti-CYR-61. (A) Reverse transcription-quantitative polymerase chain reaction and western blot analyses were used to analyze the expression of CYR-61 in H358 NSCLC cells and MRC-5 normal lung cells. Data are presented as the mean \u00b1 SD of triplicate samples. (B) Fluorescence-activated cell sorting was used to screen anti-CYR-61affinity. 1st, 2nd and 3rd represent the population of CYR-61 positive cells 1, 2 and 3 h after anti-CYR-61 treatment, respectively. (C) ELISA was used to analyze the affinity of anti-CYR-61 for its target protein, CYR-61. Data are presented as the mean \u00b1 SD of triplicate samples. (D) Western blotting was used to analyze the purified anti-CYR-61. \\*\\*P\\<0.01 vs. MRC-5 cells. CYR-61, cysteine-rich angiogenic inducer-61; SD, standard deviation; TERTR-FAM96A, telomerase reverse transcriptase-family with sequence similarity 96 member A; NSCLC, non-small cell lung cancer.](etm-16-02-0730-g00){#f1-etm-0-0-6274}\n\n![Inhibitory effects of anti-CYR-61 on H358 cells *in vitro*. (A) MTT assays were used to analyze the inhibitory effects of anti-CYR-61 on H358 cell viability. (B) Immunofluorescence (using Alexa Fluro^\u00ae^ conjugated Goat anti-rabbit Immunoglobulin G) was used to detect CYR-61 expression after anti-CYR-61 treatment. (C) Migration analysis was performed to determine the effect of anti-CYR-61 on H358 cell migration. (D) Invasion analysis was performed to determine the effect of anti-CYR-61 on H358 cell invasion. C and D were stained using hematoxylin and eosin. Data was presented as the mean \u00b1 standard deviation of triplicate samples. \\*\\*P\\<0.01 vs. the control. CYR-61, cysteine-rich angiogenic inducer-61.](etm-16-02-0730-g01){#f2-etm-0-0-6274}\n\n![Downregulation of ERK and AKT phosphorylation induced by anti-CYR-61. (A) Expression levels of ERK were analyzed using western blotting after treatment with anti-CYR-61. (B) Expression levels of AKT were analyzed by western blotting after treatment with anti-CYR-61. (C) Phosphorylation levels of ERK were analyzed in tumor sections following treatment with anti-CYR-61. Red arrows demonstrate positive pERK expression. (D) Phosphorylation levels of AKT were analyzed in tumor sections after treatment with anti-CYR-61. Red arrows demonstrate positive pAKT expression. Data are presented as the mean \u00b1 standard deviation of triplicate samples. \\*\\*P\\<0.01 vs. the Control. ERK, extracellular signal-regulated kinase; AKT, protein kinase B; CYR-61, cysteine-rich angiogenic inducer-61; p, phosphorylated.](etm-16-02-0730-g02){#f3-etm-0-0-6274}\n\n![Effect of anti-CYR-61 on apoptosis in NSCLC cells. (A) Fluorescence-activated cell sorting was used to evaluate the apoptosis of H358 cells after anti-CYR-61 treatment. Apoptosis-related (B) mRNA and (C) protein expression levels in an exogenous cell apoptosis signaling pathway were analyzed in anti-CYR-61-treated cells. Apoptosis-related (D) mRNA and (E) protein expression levels in the mitochondrial apoptosis pathway were analyzed in anti-CYR-61-treated cells. Data are presented as the mean \u00b1 standard deviation of triplicate samples. \\*\\*P\\<0.01 vs. the control. CYR-61, cysteine-rich angiogenic inducer-61; NSCLC, non-small-cell lung cancer; FasL, Fas ligand; FADD, Fas-associated protein with Death Domain; Bak, B cell lymphoma-2 antagonist/killer; Bax, B cell lymphoma-2-associated X protein; Apaf-1, apoptotic protease activating factor 1.](etm-16-02-0730-g03){#f4-etm-0-0-6274}\n\n![Therapeutic and metastasis effects of anti-CYR-61 in H358-bearing mice. (A) Tumor growth was analyzed during a 24-day short-term observation period. Data are presented as the mean \u00b1 SD. \\*\\*P\\<0.01 vs. PBS on day 21. (B) CYR-61 expression was analyzed after anti-CYR-61 treatment and DAB staining. Black arrows, *in situ* CYR-61 (C) Phosphorylation levels of ERK and AKT were analyzed by western blotting in tumors following treatment with anti-CYR-61. (D) EMT biomarker expression levels were analyzed by western blotting in tumors following treatment with anti-CYR-61. (E) Long-term survival probability was compared over a 120-day observation between the anti-CYR-61 and PBS treatment groups. It was demonstrated that the survival rate of Anti-CYR-61 treatment was significantly higher than that of PBS treatment after day 40 during the 120 day observation period. (F) NSCLC metastasis was analyzed in experimental mice. Data are expressed as the mean \u00b1 SD. \\*\\*P\\<0.01 vs. PBS. CYR-61, cysteine-rich angiogenic inducer-61; ERK, extracellular signal-regulated kinase; AKT, protein kinase B; EMT, epithelial-mesenchymal transition; PBS, phosphate-buffered saline; NSCLC, non-small cell lung cancer; p, phosphorylated; SMA, \u03b1-smooth muscle actin; SD, standard deviation.](etm-16-02-0730-g04){#f5-etm-0-0-6274}\n\n###### \n\nPrimer sequences utilized in RT-qPCR.\n\n  Gene         Primer    Sequence\n  ------------ --------- --------------------------------\n  CYR-61       Forward   5\u2032-CAAGGAGCTGGGATTCGATG-3\u2032\n               Reverse   5\u2032-AAAGGGTTGTATAGGATGCGAG-3\u2032\n  Bak          Forward   5\u2032-CACCTTACCTCTGCAACCTAG-3\u2032\n               Reverse   5\u2032-TGCAACATGGTCTGGAACTC-3\u2032\n  Bax          Forward   5\u2032-AGTAACATGGAGCTGCAGAG-3\u2032\n               Reverse   5\u2032-AGTAGAAAAGGGCGACAACC-3\u2032\n  Caspase-9    Forward   5\u2032-GTTTGAGGACCTTCGACCAG-3\u2032\n               Reverse   5\u2032-GCATTAGCGACCCTAAGCAG-3\u2032\n  Apaf-1       Forward   5\u2032-CCTCTCATTTGCTGATGTCG-3\u2032\n               Reverse   5\u2032-TCACTGCAGATTTTCACCAGA-3\u2032\n  Caspase-10   Forward   5\u2032-AATCTGACATGCCTGGAG-3\u2032\n               Reverse   5\u2032-ACTCGGCTTCCTTGTCTAC-3\u2032\n  Caspase-8    Forward   5\u2032-ATGCAAACTGGATGATGACA-3\u2032\n               Reverse   5\u2032-TTCATATCTTCAGCAGGTCT-3\u2032\n  FasL         Forward   5\u2032-AACCAAGTGGACCTTGAGACCACA-3\u2032\n               Reverse   5\u2032-TTCACATGGCAGCCCAGAGTTCTA-3\u2032\n  FADD         Forward   5\u2032-CCTGGTACAAGAGGTTCAGC-3\u2032\n               Reverse   5\u2032-CTGTGTAGATGCCTGTGGTC-3\u2032\n  \u03b2-actin      Forward   5\u2032-ACCTTCTACAATGAGCTGCG-3\u2032\n               Reverse   5\u2032-CCTGGATAGCAACGTACATGG-3\u2032\n"}
+{"text": "Related literature {#sec1}\n==================\n\nFor amides as functional groups in biologically relevant mol\u00adecules, see: Allen *et al.* (2010[@bb1]). For the synthesis of this and similar compounds, see: Montalbetti *et al.* (2005[@bb4]).\n\nExperimental {#sec2}\n============\n\n {#sec2.1}\n\n### Crystal data {#sec2.1.1}\n\nC~11~H~14~N~2~O~2~*M* *~r~* = 206.24Monoclinic,*a* = 8.0907 (10) \u00c5*b* = 15.754 (2) \u00c5*c* = 8.4529 (11) \u00c5\u03b2 = 104.205 (2)\u00b0*V* = 1044.5 (2) \u00c5^3^*Z* = 4Mo *K*\u03b1 radiation\u03bc = 0.09 mm^\u22121^*T* = 293 K0.29 \u00d7 0.21 \u00d7 0.19 mm\n\n### Data collection {#sec2.1.2}\n\nBruker SMART APEX CCD diffractometerAbsorption correction: multi-scan (*SADABS*; Bruker, 2001[@bb2]) *T* ~min~ = 0.976, *T* ~max~ = 0.9815309 measured reflections2056 independent reflections1633 reflections with *I* \\> 2\u03c3(*I*)*R* ~int~ = 0.016\n\n### Refinement {#sec2.1.3}\n\n*R*\\[*F* ^2^ \\> 2\u03c3(*F* ^2^)\\] = 0.040*wR*(*F* ^2^) = 0.105*S* = 1.042056 reflections139 parametersH atoms treated by a mixture of independent and constrained refinement\u0394\u03c1~max~ = 0.13 e \u00c5^\u22123^\u0394\u03c1~min~ = \u22120.18 e \u00c5^\u22123^\n\n {#d5e407}\n\nData collection: *SMART* (Bruker, 2007[@bb3]); cell refinement: *SAINT* (Bruker, 2007[@bb3]); data reduction: *SAINT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb5]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb5]); molecular graphics: *SHELXTL-Plus* (Sheldrick, 2008[@bb5]); software used to prepare material for publication: *SHELXL97*.\n\nSupplementary Material\n======================\n\nCrystal structure: contains datablocks global, I. DOI: [10.1107/S1600536810052207/su2236sup1.cif](http://dx.doi.org/10.1107/S1600536810052207/su2236sup1.cif)\n\nStructure factors: contains datablocks I. DOI: [10.1107/S1600536810052207/su2236Isup2.hkl](http://dx.doi.org/10.1107/S1600536810052207/su2236Isup2.hkl)\n\nAdditional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?su2236&file=su2236sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?su2236sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?su2236&checkcif=yes)\n\nSupplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [SU2236](http://scripts.iucr.org/cgi-bin/sendsup?su2236)).\n\nWe thank the Natural Science Foundation of Shanxi (No. 2010011018) for support.\n\nComment\n=======\n\nAmides are one of the most important and prolific functional groups found in biologically relevant molecules (Allen *et al.*, 2010), which lead to the synthesis of *N*-phenylmorpholine-4-carboxamide. In this study, this new acetamide derivative was prepared and its structure is presented herein.\n\nIn the title compound all the bond lengths and angles are within normal ranges. The molecule of the title compound is markedly non-planar (Fig. 1). The urea-type moiety \\[atoms N1,C7,O1,N2 - planar to within 0.0002\u00a0(13) \u00c5\\] is inclined to the phenyl ring by 42.88\u00a0(8) \u00c5. The morpholine ring has a chair conformation.\n\nIn the crystal, intermolecular N---H\u00b7\u00b7\u00b7O hydrogen bonds link the molecules into infinite one-dimensional chains propagagting in \\[001\\] (see Fig. 2 and Table 1).\n\nExperimental {#experimental}\n============\n\nThe title compound was synthesized according to the literature procedure (Montalbetti *et al.*, 2005). To a solution of isocyanatobenzene (1.19 g, 10 mmol) and morpholine (0.87 ml, 10 mmol) in CH~2~Cl~2~ (25 ml) was added triethylamine (1.20 ml, 10 mmol) in one portion. The reaction mixture was stirred at room temperature for 3 h, and then poured into ice-water (100 ml) under stirring. The combined organic phase was washed with water (3 \u00d7 20 ml), dried over MgSO~4~, and filtered. Colourless single crystals were obtained by slow evaporation of the filtrate at room temperature.\n\nRefinement {#refinement}\n==========\n\nThe NH H-atom was located in a difference Fourier map and was refined with *U*~iso~(H) = 1.2*U*~eq~(N). The C-bound H-atoms were positioned geometrically and refined as riding: C---H = 0.93 \u00c5 (CH) and 0.97 \u00c5 (CH~2~) with *U*~iso~(H) = 1.2*U*~eq~(C).\n\nFigures\n=======\n\n![Molecular structure of the title molecule. Displacement ellipsoids are drawn at the 30% probability level.](e-67-0o225-fig1){#Fap1}\n\n![View along the b-axis of the one-dimensional polymeric chain of the title compound formed by hydrogen bonding (green dashed lines). H-atoms not involved in hydrogen bonding have been omitted for clarity.](e-67-0o225-fig2){#Fap2}\n\nCrystal data {#tablewrapcrystaldatalong}\n============\n\n  ------------------------- ---------------------------------------\n  C~11~H~14~N~2~O~2~        *F*(000) = 440\n  *M~r~* = 206.24           *D*~x~ = 1.312 Mg m^\u22123^\n  Monoclinic, *P*2~1~/*c*   Mo *K*\u03b1 radiation, \u03bb = 0.71073 \u00c5\n  Hall symbol: -P 2ybc      Cell parameters from 5309 reflections\n  *a* = 8.0907 (10) \u00c5       \u03b8 = 2.6--26.1\u00b0\n  *b* = 15.754 (2) \u00c5        \u00b5 = 0.09 mm^\u22121^\n  *c* = 8.4529 (11) \u00c5       *T* = 293 K\n  \u03b2 = 104.205 (2)\u00b0          Block, colourless\n  *V* = 1044.5 (2) \u00c5^3^     0.29 \u00d7 0.21 \u00d7 0.19 mm\n  *Z* = 4                   \n  ------------------------- ---------------------------------------\n\nData collection {#tablewrapdatacollectionlong}\n===============\n\n  ------------------------------------------------------------ --------------------------------------\n  Bruker SMART APEX CCD diffractometer                         2056 independent reflections\n  Radiation source: fine-focus sealed tube                     1633 reflections with *I* \\> 2\u03c3(*I*)\n  graphite                                                     *R*~int~ = 0.016\n  \u03c9 scans                                                      \u03b8~max~ = 26.1\u00b0, \u03b8~min~ = 2.6\u00b0\n  Absorption correction: multi-scan (*SADABS*; Bruker, 2001)   *h* = \u22128\u21929\n  *T*~min~ = 0.976, *T*~max~ = 0.981                           *k* = \u221212\u219219\n  5309 measured reflections                                    *l* = \u221210\u21929\n  ------------------------------------------------------------ --------------------------------------\n\nRefinement {#tablewraprefinementdatalong}\n==========\n\n  ------------------------------------- ------------------------------------------------------------------------------------------------\n  Refinement on *F*^2^                  Primary atom site location: structure-invariant direct methods\n  Least-squares matrix: full            Secondary atom site location: difference Fourier map\n  *R*\\[*F*^2^ \\> 2\u03c3(*F*^2^)\\] = 0.040   Hydrogen site location: inferred from neighbouring sites\n  *wR*(*F*^2^) = 0.105                  H atoms treated by a mixture of independent and constrained refinement\n  *S* = 1.04                            *w* = 1/\\[\u03c3^2^(*F*~o~^2^) + (0.051*P*)^2^ + 0.1766*P*\\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3\n  2056 reflections                      (\u0394/\u03c3)~max~ \\< 0.001\n  139 parameters                        \u0394\u03c1~max~ = 0.13 e \u00c5^\u22123^\n  0 restraints                          \u0394\u03c1~min~ = \u22120.18 e \u00c5^\u22123^\n  ------------------------------------- ------------------------------------------------------------------------------------------------\n\nSpecial details {#specialdetails}\n===============\n\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Geometry. All e.s.d.\\'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.\\'s are taken into account individually in the estimation of e.s.d.\\'s in distances, angles and torsion angles; correlations between e.s.d.\\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.\\'s is used for estimating e.s.d.\\'s involving l.s. planes.\n  Refinement. Refinement of *F*^2^ against ALL reflections. The weighted *R*-factor *wR* and goodness of fit *S* are based on *F*^2^, conventional *R*-factors *R* are based on *F*, with *F* set to zero for negative *F*^2^. The threshold expression of *F*^2^ \\> \u03c3(*F*^2^) is used only for calculating *R*-factors(gt) *etc*. and is not relevant to the choice of reflections for refinement. *R*-factors based on *F*^2^ are statistically about twice as large as those based on *F*, and *R*- factors based on ALL data will be even larger.\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nFractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (\u00c5^2^) {#tablewrapcoords}\n==================================================================================================\n\n  ------ -------------- --------------- -------------- -------------------- --\n         *x*            *y*             *z*            *U*~iso~\\*/*U*~eq~   \n  C1     0.32831 (18)   0.27851 (9)     0.53510 (15)   0.0363 (3)           \n  C6     0.21670 (19)   0.23577 (9)     0.60820 (17)   0.0423 (4)           \n  H006   0.2379         0.1796          0.6410         0.051\\*              \n  C2     0.2934 (2)     0.36162 (9)     0.48415 (18)   0.0446 (4)           \n  H007   0.3679         0.3909          0.4358         0.054\\*              \n  C7     0.56729 (17)   0.17667 (8)     0.59477 (16)   0.0354 (3)           \n  C11    0.79490 (19)   0.07168 (10)    0.61799 (19)   0.0466 (4)           \n  H00A   0.7747         0.0651          0.7258         0.056\\*              \n  H00B   0.9098         0.0932          0.6306         0.056\\*              \n  C5     0.0740 (2)     0.27684 (11)    0.6320 (2)     0.0533 (4)           \n  H010   0.0011         0.2486          0.6838         0.064\\*              \n  C9     0.6653 (2)     0.04917 (11)    0.2800 (2)     0.0532 (4)           \n  H01A   0.5540         0.0256          0.2783         0.064\\*              \n  H01B   0.6755         0.0528          0.1683         0.064\\*              \n  C8     0.6786 (2)     0.13675 (10)    0.35298 (18)   0.0515 (4)           \n  H01C   0.7849         0.1629          0.3450         0.062\\*              \n  H01D   0.5854         0.1716          0.2929         0.062\\*              \n  C10    0.7776 (2)     \u22120.01294 (10)   0.5331 (2)     0.0535 (4)           \n  H01E   0.8650         \u22120.0511         0.5929         0.064\\*              \n  H01F   0.6674         \u22120.0373         0.5327         0.064\\*              \n  C4     0.0387 (2)     0.35911 (12)    0.5799 (2)     0.0633 (5)           \n  H014   \u22120.0584        0.3862          0.5951         0.076\\*              \n  C3     0.1479 (2)     0.40080 (11)    0.5054 (2)     0.0587 (5)           \n  H015   0.1236         0.4561          0.4687         0.070\\*              \n  N2     0.67253 (16)   0.13193 (8)     0.52338 (14)   0.0450 (3)           \n  O1     0.55772 (13)   0.16176 (6)     0.73527 (11)   0.0438 (3)           \n  O2     0.79327 (14)   \u22120.00542 (7)    0.36958 (15)   0.0565 (3)           \n  N1     0.47270 (16)   0.23906 (8)     0.50123 (15)   0.0425 (3)           \n  H1N    0.505 (2)      0.2590 (11)     0.421 (2)      0.051\\*              \n  ------ -------------- --------------- -------------- -------------------- --\n\nAtomic displacement parameters (\u00c5^2^) {#tablewrapadps}\n=====================================\n\n  ----- ------------- ------------- ------------- ------------ ------------ -------------\n        *U*^11^       *U*^22^       *U*^33^       *U*^12^      *U*^13^      *U*^23^\n  C1    0.0432 (8)    0.0365 (7)    0.0307 (6)    0.0074 (6)   0.0117 (6)   \u22120.0014 (5)\n  C6    0.0503 (9)    0.0360 (7)    0.0430 (8)    0.0046 (6)   0.0162 (7)   0.0020 (6)\n  C2    0.0550 (9)    0.0384 (8)    0.0453 (8)    0.0065 (7)   0.0213 (7)   0.0050 (6)\n  C7    0.0396 (7)    0.0314 (7)    0.0375 (7)    0.0000 (6)   0.0135 (6)   \u22120.0023 (6)\n  C11   0.0415 (8)    0.0472 (9)    0.0516 (9)    0.0114 (7)   0.0123 (7)   0.0019 (7)\n  C5    0.0504 (9)    0.0546 (10)   0.0616 (10)   0.0050 (8)   0.0265 (8)   0.0052 (8)\n  C9    0.0503 (9)    0.0595 (10)   0.0538 (9)    0.0068 (8)   0.0202 (7)   \u22120.0086 (8)\n  C8    0.0658 (11)   0.0506 (9)    0.0463 (8)    0.0153 (8)   0.0293 (8)   0.0033 (7)\n  C10   0.0486 (9)    0.0437 (9)    0.0701 (11)   0.0086 (7)   0.0185 (8)   0.0023 (8)\n  C4    0.0603 (11)   0.0606 (11)   0.0781 (12)   0.0253 (9)   0.0347 (9)   0.0106 (9)\n  C3    0.0698 (11)   0.0421 (9)    0.0711 (11)   0.0222 (8)   0.0306 (9)   0.0134 (8)\n  N2    0.0531 (8)    0.0446 (7)    0.0417 (7)    0.0159 (6)   0.0199 (6)   0.0024 (5)\n  O1    0.0544 (6)    0.0438 (6)    0.0369 (5)    0.0098 (5)   0.0182 (5)   0.0034 (4)\n  O2    0.0538 (7)    0.0500 (7)    0.0693 (8)    0.0108 (5)   0.0220 (6)   \u22120.0118 (6)\n  N1    0.0525 (8)    0.0407 (7)    0.0411 (7)    0.0138 (6)   0.0245 (6)   0.0091 (5)\n  ----- ------------- ------------- ------------- ------------ ------------ -------------\n\nGeometric parameters (\u00c5, \u00b0) {#tablewrapgeomlong}\n===========================\n\n  --------------------- -------------- --------------------- --------------\n  C1---C2               1.386\u00a0(2)      C5---H010             0.9300\n  C1---C6               1.387\u00a0(2)      C9---O2               1.414\u00a0(2)\n  C1---N1               1.4132\u00a0(18)    C9---C8               1.504\u00a0(2)\n  C6---C5               1.380\u00a0(2)      C9---H01A             0.9700\n  C6---H006             0.9300         C9---H01B             0.9700\n  C2---C3               1.379\u00a0(2)      C8---N2               1.4553\u00a0(18)\n  C2---H007             0.9300         C8---H01C             0.9700\n  C7---O1               1.2315\u00a0(16)    C8---H01D             0.9700\n  C7---N2               1.3559\u00a0(18)    C10---O2              1.424\u00a0(2)\n  C7---N1               1.3711\u00a0(18)    C10---H01E            0.9700\n  C11---N2              1.4601\u00a0(18)    C10---H01F            0.9700\n  C11---C10             1.504\u00a0(2)      C4---C3               1.372\u00a0(2)\n  C11---H00A            0.9700         C4---H014             0.9300\n  C11---H00B            0.9700         C3---H015             0.9300\n  C5---C4               1.376\u00a0(2)      N1---H1N              0.844\u00a0(17)\n                                                             \n  C2---C1---C6          119.46\u00a0(13)    H01A---C9---H01B      108.0\n  C2---C1---N1          117.91\u00a0(13)    N2---C8---C9          109.95\u00a0(13)\n  C6---C1---N1          122.52\u00a0(12)    N2---C8---H01C        109.7\n  C5---C6---C1          119.76\u00a0(14)    C9---C8---H01C        109.7\n  C5---C6---H006        120.1          N2---C8---H01D        109.7\n  C1---C6---H006        120.1          C9---C8---H01D        109.7\n  C3---C2---C1          119.86\u00a0(15)    H01C---C8---H01D      108.2\n  C3---C2---H007        120.1          O2---C10---C11        111.67\u00a0(13)\n  C1---C2---H007        120.1          O2---C10---H01E       109.3\n  O1---C7---N2          121.67\u00a0(13)    C11---C10---H01E      109.3\n  O1---C7---N1          122.31\u00a0(12)    O2---C10---H01F       109.3\n  N2---C7---N1          116.02\u00a0(12)    C11---C10---H01F      109.3\n  N2---C11---C10        110.11\u00a0(13)    H01E---C10---H01F     107.9\n  N2---C11---H00A       109.6          C3---C4---C5          119.41\u00a0(15)\n  C10---C11---H00A      109.6          C3---C4---H014        120.3\n  N2---C11---H00B       109.6          C5---C4---H014        120.3\n  C10---C11---H00B      109.6          C4---C3---C2          120.78\u00a0(15)\n  H00A---C11---H00B     108.2          C4---C3---H015        119.6\n  C4---C5---C6          120.69\u00a0(15)    C2---C3---H015        119.6\n  C4---C5---H010        119.7          C7---N2---C8          126.22\u00a0(12)\n  C6---C5---H010        119.7          C7---N2---C11         120.64\u00a0(12)\n  O2---C9---C8          111.64\u00a0(14)    C8---N2---C11         113.14\u00a0(12)\n  O2---C9---H01A        109.3          C9---O2---C10         110.01\u00a0(11)\n  C8---C9---H01A        109.3          C7---N1---C1          124.79\u00a0(12)\n  O2---C9---H01B        109.3          C7---N1---H1N         119.4\u00a0(12)\n  C8---C9---H01B        109.3          C1---N1---H1N         115.7\u00a0(12)\n                                                             \n  C2---C1---C6---C5     \u22121.2\u00a0(2)       O1---C7---N2---C11    \u22128.0\u00a0(2)\n  N1---C1---C6---C5     \u2212177.35\u00a0(13)   N1---C7---N2---C11    172.05\u00a0(13)\n  C6---C1---C2---C3     \u22120.5\u00a0(2)       C9---C8---N2---C7     \u2212128.09\u00a0(16)\n  N1---C1---C2---C3     175.89\u00a0(14)    C9---C8---N2---C11    51.77\u00a0(18)\n  C1---C6---C5---C4     1.8\u00a0(2)        C10---C11---N2---C7   128.49\u00a0(15)\n  O2---C9---C8---N2     \u221255.97\u00a0(18)    C10---C11---N2---C8   \u221251.38\u00a0(18)\n  N2---C11---C10---O2   54.81\u00a0(17)     C8---C9---O2---C10    60.27\u00a0(17)\n  C6---C5---C4---C3     \u22120.8\u00a0(3)       C11---C10---O2---C9   \u221259.74\u00a0(17)\n  C5---C4---C3---C2     \u22120.9\u00a0(3)       O1---C7---N1---C1     \u221216.0\u00a0(2)\n  C1---C2---C3---C4     1.5\u00a0(3)        N2---C7---N1---C1     163.93\u00a0(13)\n  O1---C7---N2---C8     171.82\u00a0(14)    C2---C1---N1---C7     150.53\u00a0(14)\n  N1---C7---N2---C8     \u22128.1\u00a0(2)       C6---C1---N1---C7     \u221233.2\u00a0(2)\n  --------------------- -------------- --------------------- --------------\n\nHydrogen-bond geometry (\u00c5, \u00b0) {#tablewraphbondslong}\n=============================\n\n  ------------------ ------------ ------------ ------------- ---------------\n  *D*---H\u00b7\u00b7\u00b7*A*      *D*---H      H\u00b7\u00b7\u00b7*A*      *D*\u00b7\u00b7\u00b7*A*     *D*---H\u00b7\u00b7\u00b7*A*\n  N1---H1N\u00b7\u00b7\u00b7O1^i^   0.844\u00a0(17)   2.130\u00a0(18)   2.9543\u00a0(16)   165.3\u00a0(16)\n  ------------------ ------------ ------------ ------------- ---------------\n\nSymmetry codes: (i) *x*, \u2212*y*+1/2, *z*\u22121/2.\n\n###### Hydrogen-bond geometry (\u00c5, \u00b0)\n\n  *D*---H\u22ef*A*        *D*---H      H\u22ef*A*        *D*\u22ef*A*       *D*---H\u22ef*A*\n  ------------------ ------------ ------------ ------------- -------------\n  N1---H1*N*\u22efO1^i^   0.844 (17)   2.130 (18)   2.9543 (16)   165.3 (16)\n\nSymmetry code: (i) .\n"}
+{"text": "**Author Information** An event is serious (based on the ICH definition) when the patient outcome is:\\* death\\* life-threatening\\* hospitalisation\\* disability\\* congenital anomaly\\* other medically important event\n\nA 39-year-old man developed severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) during treatment with mycophenolate mofetil, prednisone and tacrolimus as immunosupression therapy \\[*routes not stated; not all dosages stated*\\].\n\nThe man with a history of idiopathic dilated cardiomyopathy and chronic kidney disease, and underwent combined heart/kidney transplant in 2017, presented for COVID-19 testing. He had symptoms of fever, headache, dyspnoea, fatigue, sore throat, dry cough and myalgia. On the nest day of the testing, he was found positive for COVID-19. As the symptoms were minimal, he was quarantined at his home. On the next day, his symptoms worsened. Therefore, he was advised to return to the hospital for further evaluation. Previously, he had antibody-mediated cardiac graft rejection, which was being treated with methylprednisolone, plasmapheresis, IV immunoglobulin \\[Gamunex\\] and antithymocyte-globulin \\[rabbit anti-thymocyte globulin\\]. He had completed this treatment approximately 8\u00a0months prior to the admission. His history was significant for poorly controlled insulin-dependent diabetes mellitus, morbid obesity, hypertension and chronic diabetic foot ulcers. Upon his repeat presentation, he had headache, fatigue and dyspnoea. He was afebrile, normotensive and mildly tachycardic. Oxygen saturation was of 97% on room air. Another nasopharyngeal swab also found positive for COVID-19. Laboratory test showed elevated levels of C-reactive protein, serum lactate, D-dimer, lactate dehydrogenase and troponin I. Chest x-ray (CXR) showed mild pulmonary vascular congestion. Therefore, he was admitted for monitoring and supportive care. Upon admission, his mycophenolate mofetil was held due to his pronounced lymphopenia. However, his treatment with tacrolimus, losartan and prednisone 9mg daily kept continued. On the following day, he remained afebrile with minimal dyspnoea upon exertion. He was discharged as he insisted to. He was advised to continue holding mycophenolate mofetil until further notice. After three days, he was readmitted due to worsening of symptoms with hypoxia. At admission, he was treated with paracetamol, and he became afebrile. He was normotensive, tachycardic and tachypneic with an oxygen saturation of 95% on room air. His D-dimer, LDH, CRP, BNP and troponin I levels were found to be elevated. Therefore, his tacrolimus doses were adjusted to tacrolimus trough level of 5-8 ng/mL. A severely low immune cell response was noted upon examination. His treatment with tacrolimus and prednisone were continued, while his losartan and mycophenolate mofetil remained held. He was started on off label hydroxychloroquine 400mg twice daily. On the next day, he developed low-grade fever with hypoxia, which required 4\u00a0lpm of oxygen via nasal cannula. Due to rapid respiratory decompensation, he was shifted to the ICU. He was enrolled in the Adaptive COVID-19 Treatment Trial (NCT04280705) evaluating remdesivir versus placebo and started on the study drug. His treatment with hydroxychloroquine was discontinued. He had mild fever and chills over the next two days, which were controlled with aspirin. His oxygen requirements decreased steadily. A left ventricular ejection fraction was found to be normal upon transthoracic echocardiography. Therefore, he was shifted to the general ward. He was slowly weaned off oxygen supplementation. Eventually, he was discharged.\n"}
+{"text": "INTRODUCTION\n============\n\nChronic constipation is a symptom-based disorder of unsatisfactory defecation. Patients complain of mainly infrequent stools and/or difficult stool passage. There are two principal etiologies for such symptoms: delayed transit through the colon and impaired evacuation of the rectum. Impaired rectal evacuation can result from mechanical or structural obstruction, but the more common cause are functional pelvic outlet obstructions, which include paradoxical contraction or inadequate relaxation of pelvic floor muscle during defecation (pelvic floor dyssynergia, PFD) and failure to increase intrarectal pressure during evacuation \\[[@B1],[@B2]\\].\n\nDynamic defecography is a fluoroscopic procedure used to observe rectal evacuation of contrast paste, in order to assess changing anatomic relationships of the pelvic floor and associated organs. Thus, it is particularly relevant in those patients who complain of difficult stool passage and usually performed to diagnose functional pelvic outlet obstruction. This test is generally known to be sensitive and specific for PFD, and recommended in situations when other anorectal physiologic test results differ from clinical impression \\[[@B2]-[@B4]\\]. However, even in defecography, standardized parameters for diagnostic analysis is still incomplete, and there are some controversies about what kind of defecographic finding is necessary for diagnosing PFD \\[[@B5]-[@B7]\\].\n\nIn this study, we tried to determine which defecographic findings are most significant for PFD, and how closely they match other physiologic tests such as anorectal manometry and electromyography (EMG), and how accurately they reflect clinical symptoms of constipation.\n\nMETHODS\n=======\n\nAmong patients with chronic constipation who visited our colorectal clinic from March 2009 to April 2012, patients who completed work-up of their symptoms with dynamic defecography, anorectal manometry and EMG testing were included to the study. Chronic constipation was defined according to the Rome III criteria \\[[@B8]\\]. Patients with drug-induced or organic lesion-associated constipation were excluded. Ninety-six (mean age, 54 years; range, 17 to 83 years; female 72, male 24) constituted final study population.\n\nPatients\\' data including dynamic video files of defecography were reviewed retrospectively. Patients\\' symptoms were evaluated by reviewing our own questionnaire sheet, which was routinely checked during initial interview. The questionnaire form was made with reference to the Wexner constipation scoring system \\[[@B9]\\], and the obstructed defecation syndrome score questionnaire \\[[@B10]\\]. Among the more than 10 items of our form, 5 items which were empirically believed related to obstructed defecation were selected and scored from 0 to 4 for stratifying the symptoms of pelvic outlet obstruction. The sum of all those points was termed as obstructive symptom score (OSS) with a maximum possible of 20 points ([Table 1](#T1){ref-type=\"table\"}). When it was 16 or more, the patient was considered to have obvious obstructive symptoms of constipation.\n\nDefecographic examination was done by radiologic technician. With patient in left lateral position, liquid barium of 50 mL was inserted first into the rectum using a catheter syringe. For female patients, the vaginal wall was coated with an appropriate amount of water-soluble contrast. Then, the barium paste, which was made with our own recipe (barium powder 180 mL, potato starch 90 mL, hot water 180 mL), was then inserted to the point where patients felt rectal filling and desire to defecate. Then the patient was asked to sit down on a specially designed commode and pose as usual for defecation. Dynamic image of fluoroscopy was then obtained while the patient was trying to evacuate the rectum. Defecographic findings of interest were those already known to be suggestive of PFD. They were: poor opening of the anal canal (PO), persistent posterior angulation of the rectum (PA), poor emptying of the rectum (PE). Anatomical changes such as rectocele (RE) and intussusception (IN) were also checked. PO was defined when the anal canal was not widened enough until 20 seconds after effective push action began. PA was defined when posterior wall of the rectum was not or poorly straightened with rectal tenesmus. PE was defined when zone of evacuation (a portion of the rectum below distal transverse fold) was not or poorly emptied until 40 seconds after the anal canal opened. Rectal tenesmus was defined effective when perineal descent and concave change of upper border of the rectum were noted. Anatomical changes were defined when there was obvious anterior bulging (RE), or funnel-shaped deformity (IN) of the rectum with rectal tenesmus. All these findings were observed only in dynamic image, not in static image.\n\nAnorectal manometry was performed by nurse practitioner using a water-perfused catheter with eight channels attached to a hydraulic capillary infusion system (Medtronics, Minneapolis, MN, USA). The catheter was 4.5 mm in diameter with side-holes of 0.8 mm in diameter. Side-holes of each channel were spirally aligned with 7 mm-intervals along the longitudinal axis. Examination was performed in left lateral position without preparation. After calibration of systems, the catheter was inserted into the anal canal until the most distal side-hole was located at the anal verge. While the catheter was in position, simultaneous pressure measurements of the rectum and anal canal were done at squeeze and push with intervening resting stage. Rectal channels and anal channels were distinguished by pressure change during squeeze. Channels recording obvious pressure increase by squeeze were considered to be anal channels, and those proximal to them were considered rectal channels. Manometric defecation index (MDI) \\[[@B11]\\] was given by dividing maximum rectal pressure into minimum anal pressure during push. MDI was considered compatible for PFD when 1.2 or smaller. Manometric evacuation index (MEI) \\[[@B12]\\] was given by dividing evacuation pressure into pressure increase during squeeze. Evacuation pressure was herein defined as the pressure difference between the baseline resting pressure and the maximal relaxation or contraction during a given push effort with pressure variables measured at high-pressure zone of the anal canal. MEI was considered compatible with PFD when greater than 0.25.\n\nEMG testing was done by nurse practitioner with patient in sitting position. An anal electrode of plug type was inserted into the anal canal and another electrode of surface type was attached to the abdominal wall. After patient\\'s accommodation to the sensation of having a plug in the anal canal, EMG was recorded by system (HMT2000, HMT Inc., Seoul, Korea) at rest, squeeze, and push. EMG finding was considered compatible with PFD when reproducible increase of electrical activity was present while patient was attempting to push the plug out.\n\nStatistical analysis was completed with SPSS ver. 17.0 (SPSS Inc., Chicago, IL, USA). Crohnbach\\'s \u03b1 test was used to define internal consistency of defecographic findings. Cohen\\'s \u03ba test was used to define agreements between each defecographic finding and each result of other tests, and agreements between results of other tests.\n\nRESULTS\n=======\n\nObstructive symptoms of constipation were considered obvious by OSS in 35 patients (36.5%). For defecographic findings, PO was found in 33 (34.4%), PA was in 33 (34.4%), and PE was in 61 cases (63.5%). Patients with at least one of three findings totaled 64 (66.7%), and those with all three findings totaled 26 (27.1%). RE and IN were present in 61 (63.5%) and 37 patients (38.5%), respectively. As for findings of other physiologic tests, MDI, MEI and EMG pattern compatible with PFD were in 81 (84.4%), 72 (75%), and 73 cases (76%), respectively.\n\nInternal consistency of three defecographic findings was generally good (\u03b1 = 0.78), although consistency between PO and PA was highest ([Table 2](#T2){ref-type=\"table\"}). Agreements between each defecographic finding and each result of other tests (OSS, MDI, MEI, EMG) were all poor, and agreements between simultaneous occurrence of all three defecographic findings and each result of other tests were also poor ([Table 3](#T3){ref-type=\"table\"}). Agreements within results of other tests were moderate between MDI and EMG, but fair or poor between others ([Table 4](#T4){ref-type=\"table\"}).\n\nDISCUSSION\n==========\n\nChronic constipation is a symptom-based disorder of unsatisfactory defecation. Patients complain of mainly infrequent stools, difficult stool passage, or both. Except organic ones, etiologies of such symptoms are known to be colonic slow transit, pelvic outlet obstruction and irritable bowel syndrome (IBS) \\[[@B1],[@B2]\\]. But IBS was classified separately in Rome criteria of functional constipation, and thus colonic slow transit and pelvic outlet obstruction could be deemed two major etiologies of chronic constipation.\n\nPelvic outlet obstruction can be either structural or functional \\[[@B1]\\]. Functional outlet obstruction is newly termed in Rome III criteria as functional defecation disorders, which include dyssynergic defecation and inadequate defecatory propulsion. Dyssynergic defecation is defined as inappropriate contraction of the pelvic floor or insufficient relaxation of the anal sphincter with adequate propulsive forces during attempted defecation. Inadequate defecatory propulsion is defined as inadequate propulsive forces with or without inappropriate contraction or insufficient relaxation of the anal sphincter during attempted defecation \\[[@B8],[@B13]\\]. PFD was used previously as equivalent terminology to functional defecation disorders, without including the concept of inadequate defecatory propulsion. Strictly, it is equivalent to dyssynergic defecation in Rome III criteria.\n\nRome III criteria for diagnosing functional defecation disorders are as follows; 1) The patient must satisfy diagnostic criteria of functional constipation, and 2) During repeated attempts to defecate must have at least 2 of the following: (1) evidence of impaired evacuation, based on balloon expulsion or imaging, (2) inappropriate contraction of the pelvic floor muscles or less than 20% relaxation of basal resting sphincter pressure by manometry, imaging, or EMG, (3) inadequate propulsive forces assessed by manometry or imaging \\[[@B8],[@B13]\\].\n\nAs shown above, usual tools for diagnosis of functional defecation disorders have been imaging, manometry and EMG, with recent addition of balloon expulsion test. Among them, defecography as imaging test has been considered to be a sensitive and specific measurement of PFD when it documents nonrelaxation of the puborectalis and the anal sphincter \\[[@B3],[@B4]\\]. While balloon expulsion test, EMG, or anorectal manometry is usually chosen as a first-line test in diagnostic steps for PFD, defecography may be useful as a second-line test, especially when those tests are discrepant or differ from the clinical impression \\[[@B2]-[@B4]\\].\n\nBut defecography also has its own limitations. Contrast paste used in defecography is so different from typical condensed feces in patients with pelvic outlet obstruction that it may mislead to a different physiology of defecation. Particularly, such paste cannot always be a standardized commercial product, depending on the medico-social environment. Then, its consistency can vary enough to lead to a different physiology of defecation in certain cases. It is also well known that the reproducibility of measurement value is lacking, the concept of normal value is ambiguous, and standardized parameters for defecographic analysis are still incomplete. Besides, patient\\'s feeling of embarrassment about being observed while defecating may lead to false positive findings.\n\nUsual diagnostic findings of defecographic PFD are poor opening of the anal canal, persistent posterior angulation of the rectum, and/or poor emptying of the rectum. Traditionally, efforts to diagnose PFD have been focused on a patient\\'s inappropriate puborectalis contaction during rectal evacuation. As the puborectalis cradles the anorectal junction, it has been suggested that a prominent muscular impression during evacuation, reflected by persistent posterior angulation of the rectum, is indicative of PFD. However, several studies have suggested that poor rectal emptying is more important in many of these patients and that this impairment may be more reliable in prediction of PFD than abnormal puborectalis configuration. They insisted that although the anorectal angle may sometimes reflect puborectalis tone, it is more often a secondary sign \\[[@B6],[@B7]\\].\n\nIn our study, internal consistency of defecographic variables was calculated as good. Then, it is not pertinent to treat PO, PA, and PE separately in defecographic analysis. In other words, it is hard to say one of the above findings of defecography is more important than the others for the diagnosis of PFD. Moreover, individual agreements between each defecographic finding and each result of other tests were all poor. None of the defecographic findings were correlated to PFD-compatible findings in MDI, MEI or EMG. None of them agreed with clinical obstructive symptoms assessed by OSS. Even simultaneous occurrence of all three defecographic findings did not show better agreement with the results of other tests. Thus, it is hard to expect consistent results of various diagnostic tests for PFD.\n\nAlthough defecography has often been considered as sensitive and specific for PFD, it is better to believe no single test is specific enough to confirm it, because other physiologic tests could also have their own measuring accuracy. In this sense, PFD should not depend on the results of a single test, but depend on the consistent results of 2 or more tests for its diagnosis. However, in another sense, those diagnostic tests for PFD including defecography may not be testing the same pathophysiologic condition, or the pathophysiology of PFD is much more complicated than generally accepted.\n\nBased on this kind of interpretation, it was suggested that careful differentiation of testing results suggestive of PFD may actually lend itself to different forms of treatment \\[[@B5]\\]. Although theoretical, patients with poor rectal emptying on defecography, yet with normal EMG may have a rectal propulsion problem, rather than a relaxation problem, and thus may benefit from a laxative treatment. But patients with PFD-compatible EMG, irrespective of defecographic findings, could have more chance of improvement with biofeedback therapy. Further studies will elucidate issues about whether patients with PFD should be treated according to their individual patterns of testing abnormalities.\n\nThere are some problems or limitations in our study. First, the number of patients with abnormal findings in MDI, MEI, or EMG was high enough to suppose some errors in measuring technique, or interpreting method. But it does not seem possible, because agreement between MDI and EMG was moderate and it is very unusual for such errors to occur simultaneously in two different tests. In addition, the defecographic contrast paste used in our study was made using our own recipe at each time of examination. Although we tried to keep constancy in the making process, there could have been some errors in maintaining equable consistency of the paste. These can lead to different results under similar conditions.\n\nIn conclusion, internal consistency of defecographic variables was so good that it is not pertinent to treat PE, PO, and PA separately in defecographic analysis. Then, among known defecographic findings for PFD, one specific finding cannot be deemed more important than the others for its diagnosis. Individual agreements between each defecographic finding and each result of other tests were all poor. Thus, it is hard to expect consistent results of various diagnostic tests and to predict the presence of defecographic PFD by use of anorectal manometry, EMG, or even by clinical symptoms. Further studies will elucidate issues about whether patients with PFD should be treated according to their individual patterns of testing abnormalities.\n\nThis study was supported by Konkuk university, 2012.\n\nNo potential conflict of interest relevant to this article was reported.\n\n###### \n\nThe obstructive symptom score\n\n![](jkss-84-225-i001)\n\n###### \n\nInternal consistency of defecographic findings (Crohnbach\\'s \u03b1 value)\n\n![](jkss-84-225-i002)\n\nPO, poor opening of the anal canal; PA, persistent posterior angulation of the rectum; PE, poor emptying of the rectum.\n\nExcellent, \u03b1 \u2265 0.9; good, 0.9 \\> \u03b1 \u2265 0.8; acceptable, 0.8 \\> \u03b1 \u2265 0.7; questionable, 0.7 \\> \u03b1 \u2265 0.6; poor, 0.6 \\> \u03b1 \u2265 0.5; unacceptable, 0.5 \\> \u03b1.\n\n###### \n\nAgreement between defecography findings and results of other tests (Cohen\\'s \u03ba value)\n\n![](jkss-84-225-i003)\n\nPO, poor opening of the anal canal; PA, persistent posterior angulation of the rectum; PE, poor emptying of the rectum; POAE, simultaneous occurrence of PO, PA, and PE; OSS, obstructive symptom score; MDI, manometric defecation index; MEI, manometric evacuation index.\n\nGood, \u03ba \\> 0.8; substantial, 0.8 \u2265 \u03ba \\> 0.6; moderate, 0.6 \u2265 \u03ba \\> 0.4; fair, 0.4 \u2265 \u03ba \\> 0.2; poor, 0.2 \u2265 \u03ba.\n\n###### \n\nAgreement between other tests (Cohen\\'s \u03ba value)\n\n![](jkss-84-225-i004)\n\nOSS, obstructive symptom score; EMG, electromyography; MDI, manometric defecation index; MEI, manometric evacuation index.\n\nGood, \u03ba \\> 0.8; substantial, 0.8 \u2265 \u03ba \\> 0.6; moderate, 0.6 \u2265 \u03ba \\> 0.4; fair, 0.4 \u2265 \u03ba \\> 0.2; poor, 0.2 \u2265 \u03ba.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nCancer stem cells (CSCs) are resistant to conventional therapeutic approaches \\[[@R1]-[@R3]\\]. As a consequence, they have been directly implicated in the disease pathogenesis of tumor recurrence and distant metastasis \\[[@R4], [@R5]\\]. In addition, drug-resistant CSCs have been linked to unfavorable clinical outcomes, across different tumor types \\[[@R6]-[@R8]\\]. As only a very small percentage of cancer cells have \"stem-like\" and \"tumor-initiating\" properties, they are difficult to study and their key distinguishing features remain relatively uncharacterized, although they appears to resist both chemo-therapy and radiation.\n\nInterestingly, CSCs share many properties with normal stem cells, including immortality and resistance to stress, as well as asymmetric cell division \\[[@R9], [@R10]\\]. A particular distinguishing characteristic of CSCs is their ability to initiate tumors and to undergo anchorage-independent growth, when cultured in suspension \\[[@R11]\\]. Under these particular cell culture conditions, CSCs proliferate and form 3D-spheroid-like structures, containing CSCs and progenitor cells, which are known as \"tumor-spheres\" or \"onco-spheres\" \\[[@R12], [@R13]\\]. In striking contrast, the vast majority of non-CSCs undergo a specialized form of apoptosis in suspension cultures, called anoikis. Importantly, each 3D-spheroid originates from the clonal proliferation of a single CSC, is not to due to the self-aggregation of cancer cells. Thus, tumor-sphere formation is an efficient means to selectively enrich for CSCs. The CSC population is resistant to DNA-damage, and shows lower levels of ROS production, as well \\[[@R14], [@R15]\\]. 3D-tumor-spheres derived from breast cancer cells are also known as mammo-spheres \\[[@R12], [@R13]\\].\n\nClinically, there is an urgent need to identify new therapeutic strategies for selectively targeting CSCs. Here, we show that graphene oxide (GO) demonstrates this selectivity. As such, our current study provides a new rationale for exploiting graphene oxide itself as an anti-cancer therapeutic, rather than simply as a drug-delivery agent \\[[@R16]\\].\n\nRESULTS {#s2}\n=======\n\nGO flakes target breast cancer stem cells {#s2_1}\n-----------------------------------------\n\nHere, we tested the efficacy of graphene oxide as a potential new anti-cancer agent, for the selective targeting of CSCs. Graphene, described as two-dimensional sheets of carbon atoms without any additional functional groups, does not forms stable dispersions in water or other biologically relevant solvents \\[[@R17]\\]. On the other hand, graphene oxide \\[[@R18]\\] is the water-soluble derivative of graphene which can be produced with various sizes and bearing varied functional groups and which can be more easily manipulated experimentally, especially in biological systems. Sterile GO dispersions were prepared in a 5% mixture of DMSO in double-distilled water for these studies.\n\nInitially, we tested the ability of GO to affect CSC proliferation, using MCF7 cells, a well-established ER(+) breast cancer cell line. Two grades of GO were used, small GO (s-GO) with flake sizes of 0.2--2 \u03bcm and big GO (b-GO) with flake sizes of 5--20 \u03bcm, to represent two size classes, where the flakes are either smaller or larger than the target cells (Figure [1](#F1){ref-type=\"fig\"}). For this purpose, we assessed the effects of graphene oxide on the anchorage-independent clonal expansion of MCF7 CSCs, using the tumor-sphere assay. This functional assay directly measures CSCs proliferative expansion \\[[@R12]\\]. Figure [2](#F2){ref-type=\"fig\"} shows the results of this analysis. Using s-GO flakes, we observed a dose-dependent inhibition of tumor-sphere formation, in the range of 1.25 to 25 \u03bcg/ml, with an IC-50 of \\~12.5 \u03bcg/ml. Similarly, using b-GO flakes, we also observed a dose-dependent inhibition of tumor-sphere formation, in the range of 6.25 to 100 \u03bcg/ml, again with an IC-50 of \\~12.5 \u03bcg/ml. Importantly, both small and big GO flakes did not affect the viability of the bulk non-CSC population of MCF7 cells, indicating selectivity towards CSCs (Figure [2](#F2){ref-type=\"fig\"}).\n\n![Graphene oxide (GO) grades\\\nLeft and Right panels show atomic force microscopy images of graphene oxide on a silicon dioxide substrate indicating the flake size distribution and monolayer thickness of the flakes. Inset shows a vial of b-GO stock dispersion in DMSO and water, at concentration of 2.3 mg/ml.](oncotarget-06-3553-g001){#F1}\n\n![Graphene oxide (GO) selectively targets cancer stem cells (CSCs) in breast cancer cells\\\nUpper Panels. Note that GO (big and small flakes) inhibits the anchorage-independent proliferation of MCF7 CSCs, as evidenced by inhibition of mammosphere formation. Lower Panels. In contrast, GO (big and small flakes) does not affect cell viability of the total MCF7 cell population. An \\* indicates p \\< 0.05 (Student\\'s t-test).](oncotarget-06-3553-g002){#F2}\n\nGO flakes target CSCs, across multiple cancer types {#s2_2}\n---------------------------------------------------\n\nSince both the small and big GO flakes showed similar potency, we focused more on evaluating the efficacy of b-GO flakes. We next evaluated whether GO also showed efficacy against CSCs from multiple cancer types, such as ovarian, prostate, pancreatic and lung cancers, as well as glioblastoma (brain) (the six cell lines tested are summarized in Table [1](#T1){ref-type=\"table\"}). For simplicity, we tested b-GO flakes at two doses, namely 25 and 50 \u03bcg/ml.\n\n###### Six Cancer Cell Models With Board Applicability\n\n  Cancer Types      Cell Lines\n  ----------------- ------------\n  **Breast(ER+)**   MCF7\n  **Ovarian**       SKQV3\n  **Prostate**      PC3\n  **Pancreatic**    MIA-PaCa-2\n  **Lung**          A549\n  **Ghoblastoma**   U87MG\n\nFigure [3](#F3){ref-type=\"fig\"} shows that b-GO flakes also effectively inhibited tumor-sphere formation in these 5 other cell lines. Thus, our results indicate that GO must be targeting a relatively specific and highly-conserved phenotypic property of CSCs, across multiple cancer types. Representative images of tumor-sphere inhibition by GO treatment are shown in Figure [4](#F4){ref-type=\"fig\"}. Interestingly, the viability of bulk non-CSCs from these five cancer cell lines (SKOV3, PC3, A549, Mia PaCa2, and U-87) was not affected by GO (Figure [5](#F5){ref-type=\"fig\"}), further highlighting its specificity and selectivity for CSCs.\n\n![Graphene oxide (GO) selectively targets cancer stem cells (CSCs) of multiple cancer cell types\\\nGO (big flakes) inhibits the anchorage-independent proliferation of SKOV3 ovarian cancer cells (A), U87 glioblastoma cells (B), PC3 prostate cancer cells (C), A549 lung cancer cells (D), as well as pancreatic cancer cells (E), in a concentration-independent manner. These results indicate that GO inhibits sphere formations of multiple cancer types. An \\* indicates p \\< 0.05 (Student\\'s t-test).](oncotarget-06-3553-g003){#F3}\n\n![Graphene oxide (GO) selectively targets cancer stem cells (CSCs) of multiple cancer cell types\\\nRepresentative images showing that GO (big flakes, 25\u03bcg/ml) inhibits the anchorage-independent proliferation of MCF7 breast cancer cells, SKOV3 ovarian cancer cells, PC3 prostate cancer cells, U87 glioblastoma cells, A549 lung cancer cells as well as MIA-PaCa-2 pancreatic cancer cells.](oncotarget-06-3553-g004){#F4}\n\n![Graphene oxide (GO) does not affect cell viability of the total population of cancer cells\\\nCell viability was assessed using an SRB assay. Note that GO (big flakes) does not affect cell viability of the total cell population of SKOV3 ovarian cancer cells (A), U87 glioblastoma cells (B), PC3 prostate cancer cells (C), A540 lung cancer cells (D) as well as MIA-PaCa-2 pancreatic cancer cells.](oncotarget-06-3553-g005){#F5}\n\nImportantly, b-GO flakes also did not affect the viability of a normal skin fibroblast cell line (hTERT-BJ1), indicating that GO is relatively non-toxic for normal body cells (Figure [6](#F6){ref-type=\"fig\"}). This is consistent with the findings of many other laboratories, i.e., that GO is non-toxic for multiple normal cell types \\[[@R19], [@R20]\\].\n\n![Graphene oxide (GO) does not affect the cell viability normal fibroblasts\\\nCell viability of hTERT-BJ1 fibroblasts was assessed using an SRB assay. Note that GO (big flakes) does not affect cell viability of the total cell population of normal fibroblasts.](oncotarget-06-3553-g006){#F6}\n\nGO-based mechanistic studies: Effects on well-established CSC signaling pathways {#s2_3}\n--------------------------------------------------------------------------------\n\nTo gain mechanistic insights into how GO flakes target cancer stem cells, we next analyzed their effects on a series of well-established signal transduction pathways, which have been shown to contribute towards \"stemness\" \\[[@R21]-[@R23]\\].\n\nFor this purpose, we used a panel of MCF7 cell lines that were stably-transfected with different luciferase reporters, that allows one to quantitatively measure the activation-state of a given signal transduction pathway. Interestingly, Figure [7](#F7){ref-type=\"fig\"} shows that a number of signaling pathways were significantly inhibited by GO treatment. More specifically, GO treatment inhibited WNT- and Notch-driven signaling, as well as STAT1/3 signaling and the NRF2-dependent anti-oxidant response. However, little or no effect was observed on TGF-beta/SMAD-signaling (Figure [7](#F7){ref-type=\"fig\"}).\n\n![Graphene oxide (GO) inhibits signaling pathways related to cancer stem cells, antioxidant responses and interferon\\\nMCF7 breast cancer cells carrying luciferase-reporters (Cignal, QIAGEN) were generated to monitor the activation of a variety of signaling networks, including Wnt, STAT3, Notch, NRF2-dependent antioxidant responses, Interferon\u03b3-STAT1 and SMAD-TGF\u03b2 pathways. MCF7-Luc reporter cells were treated with GO (big flakes) for 48 hours and luminescence was determined as a measure of pathway activation status. Note that GO inhibits cancer stem cell signaling (WNT, STAT3 and Notch), NRF2-antioxidant responses, as well as INF\u03a5-STAT1 signaling. No effects were observed for the SMAD-TGF\u03b2-pathway. An \\* indicates p \\< 0.05; \\*\\* indicates p \\< 0.01 (Student\\'s t-test).](oncotarget-06-3553-g007){#F7}\n\nThus, it appears that GO treatment somehow targets several different signal transduction pathways in cancer cells, to reduce overall \"stemness\".\n\nGO promotes the differentiation of breast cancer stem cells {#s2_4}\n-----------------------------------------------------------\n\nTo further validate the idea GO was reducing stemness in MCF7-derived CSCs, we used a series of well-established breast cancer stem cell markers (CD44 and CD24), and quantitatively analyzed their expression by FACS analysis. The results of these studies are presented in Figure [8](#F8){ref-type=\"fig\"}.\n\n![Graphene oxide (GO) promotes the differentiation of breast cancer stem cells\\\nMCF7 cells were treated as monolayer cultures with small or big GO (50\u03bcg/ml) for 48 hours or left untreated (vehicle alone control). Then, cells were trypsinized and plated on low-attachment plates for 10 hours to induce anoikis and enrich for cancer stem cells. Single cells were then analyzed by FACS to quantitate the CD44(+)CD24-/low population, which represents the cancer stem cells. A. Note that, as expected, the CD44(+)CD24-/low population is greatly enriched after 10 hours in low-attachment conditions (vehicle alone control). B. Interestingly, GO does not reduce the total number of anoikis-resistant cells (data not shown), but rather induces the expression of CD24, thereby significantly reducing the CD44(+)CD24-/low population. This suggests that GO inhibits mammosphere formation by promoting the differentiation of breast cancer stem cells. An \\* indicates p \\< 0.05 (Student\\'s t-test).](oncotarget-06-3553-g008){#F8}\n\nBriefly, MCF7 cells were treated as monolayer cultures with GO (50\u03bcg/ml) for 48 hours or left untreated (vehicle alone control). Then, cells were trypsinized and plated on low-attachment plates for 10 hours to induce anoikis and enrich for CSCs. Single cells were then analysed by FACS to quantitate the CD44(+)CD24-/low population, which represents the breast CSCs. As predicted, the CD44(+)CD24-/low population is greatly enriched after 10 hours in low-attachment conditions (vehicle alone control) (Figure [8A](#F8){ref-type=\"fig\"}).\n\nInterestingly, GO does not reduce the total number of anoikis-resistant cells (data not shown), but rather induces the expression of CD24, thereby significantly reducing the CD44(+)CD24-/low population (Figure [8B](#F8){ref-type=\"fig\"}). This suggests that GO inhibits mammosphere formation by promoting the differentiation of breast cancer stem cells, supporting our results from the analysis of multiple signal transduction pathways.\n\nDISCUSSION {#s3}\n==========\n\nHere, we show that treatment with GO is sufficient to inhibit tumor-sphere formation in six independent cancer cell lines, across multiple tumor types (breast, ovarian, prostate, lung, and pancreatic cancer, as well glioblastoma (brain cancer)). These results suggest that GO specifically targets a global phenotypic property of CSCs that is highly conserved in multiple tumor types. Moreover, using MCF7 cells expressing a panel of luciferase reporters, we observed that GO treatment was indeed sufficient to inhibit a number of different signal transduction pathways, including WNT, Notch, STAT1/3 and NRF-2, but did not effect TGB-beta/SMAD signaling. Finally, using a panel of specific well-established breast CSC markers (namely CD44 and CD24), we show that GO appears to reduce the number of CSCs by inducing their differentiation, as they now begin to express CD24. Thus, GO may reduce the number of bonafide CSCs that are capable of forming tumor-spheres, by inducing their differentiation and inhibiting their proliferation. However, additional mechanistic studies are clearly warranted.\n\nImportantly, our preliminary results indicate that GO treatment does not significantly affect oxidative mitochondrial metabolism (OXPHOS) in this context (data not shown), suggesting that GO does not target mitochondria. This is in contrast to our previous studies where a number of mitochondrially-targeted FDA-approved antibiotics effectively eradiated CSCs \\[[@R24]\\]. Thus, GO and mitochondrially-targeted antibiotics appear to work differently, via separate and distinct molecular mechanism(s).\n\nAlso, since b-GO flakes are 5-to-20\u03bcm in size, they must be exerting their effects at the cell surface, as they are too large to be internalized within cells and are actually larger than a single cell. This is consistent with our findings that GO-treatment dampens the activation of several stem cell associated signal transduction pathways, which are initiated at the cell surface. This could then mechanistically induce CSC differentiation, which we observed experimentally (summarized in Figure [9](#F9){ref-type=\"fig\"}).\n\n![Graphene oxide (GO): Targeting cancer stem cells with differentiation-based nano-therapy\\\nOur current mechanistic studies suggest that GO could directly be used as a therapeutic for targeting CSCs, because of its ability to induce differentiation. In this context, we might envision that GO could used to clear residual CSCs, with the aim of preventing tumor recurrence and distant metastasis, thereby providing a practical means for achieving \"differentiation-based nano-therapy\".](oncotarget-06-3553-g009){#F9}\n\nPrevious studies have shown that GO (or its related derivatives) can inhibit \"bulk\" cancer cell migration or prevent tumor growth in xenograft models \\[[@R25]-[@R28]\\]. However, none of these studies connected GO treatment to the CSC phenotype or indicated that it could be used for \"differentiation\" therapy.\n\nInterestingly, several studies have shown that GO is non-toxic for normal stem cells, and indeed GO promotes their differentiation. More specifically, it was demonstrated that culturing normal pluripotent stem cells on GO as a substrate induces their terminal differentiation towards multiple cell lineages, including neurons, chondrocytes and adipocytes \\[[@R29]-[@R33]\\]. These properties are currently being actively exploited for tissue engineering and regenerative medicine, by using GO as a scaffold for bone reconstruction and neural regeneration.\n\nOur new mechanistic studies suggest that GO could directly be used as a therapeutic for targeting CSCs, possibly as a differentiation agent. In this context, we envision that GO could be delivered i.v. or p.o., as a new anti-cancer therapeutic, depending on the location of the tumor. Alternatively, GO flakes could also be used as a lavage solution during surgery, to clear the tumor excision site or the peritoneal cavity (as in ovarian or other peritoneal cancers) of residual CSCs, with the aim of preventing tumor recurrence and distant metastasis, via differentiation-based nano-therapy (Figure [9](#F9){ref-type=\"fig\"}).\n\nMATERIALS & METHODS {#s4}\n===================\n\nMaterials {#s4_1}\n---------\n\nCancer cell lines were purchased from the ATCC or other commercially available sources. Gibco-brand cell culture media (DMEM/F12) was purchased from Life Technologies.\n\nGraphene oxide {#s4_2}\n--------------\n\nGraphene oxide was prepared by using the Hummers method with modifications \\[[@R34], [@R35]\\]. The individual graphite oxide flakes contain carboxyl groups mainly at the edges, and epoxide, hydroxide and ketone groups mainly on the basal plane. The C to O ratio is usually slightly lower or slightly higher than 1 as determined by X-ray photoemission spectroscopy. The graphene oxide flakes of different sizes were separated by centrifuging graphene oxide suspensions at various rpm and collecting different phases of the suspension. The AFM characterization of graphene oxide flakes was performed on a Bruker Dimension FastScan AFM system by using taping mode. The substrates were prepared by spin-casting the suspension on a Si/SiO2 substrate to yield monolayer film, followed by AFM imaging. Concentrations were obtained from UV-Vis spectra, which were recorded in 10 mm path length quartz cells using a PerkinElmer Lambda -- 1050 UV-Vis-NIR spectrometer. The dispersions were diluted to give the absorption intensity lower than 1.\n\nTumor-sphere culture {#s4_3}\n--------------------\n\nA single cell suspension was prepared using enzymatic (1x Trypsin-EDTA, Sigma Aldrich, \\#T3924), and manual disaggregation (25 gauge needle) to create a single cell suspension. Cells were plated at a density of 500 cells/cm^2^ in mammosphere medium (DMEM-F12/B27/20ng/ml EGF/PenStrep) in non-adherent conditions, in culture dishes coated with (2-hydroxyethylmethacrylate) (poly-HEMA, Sigma, \\#P3932) \\[[@R12]\\]. Cells were grown for 5 days and maintained in a humidified incubator at 37\u00b0C at an atmospheric pressure in 5% (v/v) carbon dioxide/air. After 5 days for culture, spheres \\>50 \u03bcm were counted using an eye graticule, and the percentage of tumor-sphere formation was normalized to 100% for vehicle alone control (1 = 100 % TSF) \\[[@R12]\\]. All experiments were performed in triplicate, three times independently, such that each data point represents the average of 9 replicates.\n\nEvaluation of CSC signalling pathways {#s4_4}\n-------------------------------------\n\nThe Cignal Lenti reporter assay (luc) system (Qiagen) was chosen for monitoring the activity of several signal transduction pathways in MCF7 cells. The responsive luciferase constructs encode the firefly luciferase reporter gene under the control of a minimal (m)CMV promoter and tandem repeats of response elements for each pathway. The following constructs were used: TCF/LEF(luc) for Wnt signal transduction (CLS-018L); STAT3(luc) for transcriptional activity of STAT3 (CLS-6028L); RBP-Jk(luc) for Notch-induced signaling (CLS-014L); ARE(luc) for Nrf2- and Nrf1-mediated antioxidant response (CLS-2020L); GAS(luc) for Interferon gamma-induced Stat1-signal transduction (CLS-009L); SMAD(luc) for TGF\u03b2-induced signal transduction (CLS-017L). Briefly, 1 \u00d7 10^5^ MCF7-GFP cells were seeded in 12-well plates. Once cells were attached, the viral particles were diluted 1:10 in complete culture media containing polybrene (sc-134220, Santa Cruz), and added to the cells. Puromycin treatment (P9620, Sigma) was started 48 hours later in order to select stably infected cells.\n\nLuciferase assays {#s4_5}\n-----------------\n\nThe Luciferase Assay System (E1501, Promega Kit) was used on all luciferase reporter MCF7 cells treated with GO. Briefly, 6 \u00d7 10^3^ MCF7-GFP cells were seeded in black-walled 96-well plates and then were treated with GO (50 \u03bcg/ml). As controls, vehicle-treated cells were run in parallel. Four replicates were used for each condition. After 48 hours of treatment, luciferase assays were performed according to the manufacturer\\'s instructions. Light signal was acquired for 2 minutes in photons/second in the Xenogen VivoVision IVIS Lumina (Caliper Life Sciences), and the results were analysed using Living Image 3.2 sofware (Caliper Life Sciences). Luminescence was normalized using SRB (to determine total cellular protein), as a measure of MCF7 cell viability.\n\nAnoikis and CSC differentiation analysis {#s4_6}\n----------------------------------------\n\nFollowing GO treatment, the CSC population was enriched by seeding on low-attachment plates. Under these conditions, the non-CSCs undergo anoikis (a form of apoptosis induced by lack of proper attachment) and CSCs are believed to survive. The expression of differentiation markers by the surviving \"CSC fraction\" was analyzed by FACS analysis. Briefly, 1 \u00d7 10^4^ MCF7 cells were treated with GO (50\u03bcg/ml) for 48h in 6-well plates, grown as a monolayer. Then, the monolayer cells were trypsinized and seeded in low-attachment plates in mammosphere media. After 10h under low-attachment conditions, MCF7 cells were spun down and incubated with CD24 (IOTest CD24-PE, Beckman Coulter) and CD44 (APC mouse Anti-Human CD44, BD Pharmingen cat.559942) antibodies for 15 minutes on ice. Cells were rinsed twice and incubated with LIVE/DEAD dye (Fixable Dead Violet reactive dye; Invitrogen) for 10 minutes. Samples were then analyzed by FACS (Fortessa, BD Bioscence). Only the live population, as identified by the LIVE/DEAD dye staining, was analyzed for CD24/CD44 expression. Data were analysed using FlowJo software. Virtually identical results were also obtained using 7-AAD (7-Aminoactinomycin D; Life Technologies) to distinguish between the live and dead populations of cells (cell viability), during anoikis.\n\nWe thank the University of Manchester for providing start-up funds that contributed to the success of this study (to Federica Sotgia and Michael P. Lisanti). Aravind Vijayaraghavan, Maria Iliut and Andrea F. Verre were supported by Engineering and Physical Sciences Research Council (EPSRC) grants EP/G03737X/1 and EP/G035954/1.\n"}
+{"text": "One of the several ways in which microbiology puts the neo-Darwinian synthesis in jeopardy is by the threatening to \"uproot the Tree of Life (TOL)\" \\[[@pgen.1005912.ref001]\\]. Lateral gene transfer (LGT) is much more frequent than most biologists would have imagined up until about 20 years ago, so phylogenetic trees based on sequences of different prokaryotic genes are often different. How to tease out from such conflicting data something that might correspond to a single, universal Tree of Life becomes problematic. Moreover, since many important evolutionary transitions involve lineage fusions at one level or another, the aptness of a tree (a pattern of successive bifurcations) as a summary of life's history is uncertain \\[[@pgen.1005912.ref002]--[@pgen.1005912.ref004]\\].\n\nDarwin's Tree of Life Hypothesis {#sec001}\n================================\n\nBefore 1859, hierarchical (tree-like) patterns of organismal relationships were commonly drawn and taken to mirror some natural order, most likely divine \\[[@pgen.1005912.ref005]\\]. What Darwin gave us when he published *The Origin of Species* was a nontheistic reason for the discoverability and utility of such classifications. \"Community of descent is the hidden bond which naturalists have been unconsciously seeking, and not some unknown plan of creation,\" he wrote \\[[@pgen.1005912.ref006]\\]. There was to be an actual TOL whose \"ramifying branches may well represent the classification of all extinct and living species in groups subordinate to groups\" \\[[@pgen.1005912.ref006]\\]. Bapteste and one of us have called this claim by Darwin his \"TOL Hypothesis\"---that is, that the tree-like pattern of relationships recognized by systematists reflects an underlying tree-like evolutionary process \\[[@pgen.1005912.ref007]\\].\n\nAfter Darwin, classifications were most often assumed without proof to be evolutionary: phenetics was taken to be identical to phylogenetics. As De Querioz noted, \"...the relationships expressed in existing taxonomies were merely reinterpreted as the result of evolution, and evolutionary concepts were developed to justify existing methods\" \\[[@pgen.1005912.ref008]\\]. Not much could be done to improve on this as long as the methods of comparative biology (mostly anatomical) remained the basis for classification, as they had been for centuries.\n\nIn the mid-1960s, molecular sequencing (first proteins, then ribosomal RNAs, then genes, and now genomes) appeared to offer a way out of such circularity, a possible independent proof of the TOL hypothesis, some thought. Emile Zuckerkandl and Linus Pauling wrote:\n\n> \"...molecular phylogenetic trees should in principle be definable in terms of molecular information alone. It will be determined to what extent the phylogenetic tree, as derived from molecular data in complete independence of organismal biology, coincides with the tree constructed on the basis of organismal biology...\"\n\nAnd thus,\n\n> \"...If the two trees are mostly in agreement with respect to the topology of branching, the best available single proof of the reality of macro-evolution would be furnished\" \\[[@pgen.1005912.ref009]\\].\n\nThis made a lot of sense at the time. The way in which genes replicate and mutate is, barring recombination, tree-like. One consequence is that ancestral nodes in a tree constructed from gene sequences are interpreted as corresponding to real ancestors (actual genes). Nothing like that would necessarily be the case for trees based on \"organismal biology,\" that is, on phenotypic similarities and differences. If evolution were not the cause of phenotypic similarities and differences, then phenetic classifications would be like those used to order books in a library: all the books on Canadian cuisine might (arguably) belong on the same shelf, but no one would claim that they descended from one first book on that subject. Nodes in such a phenetic classification need not represent ancestors \\[[@pgen.1005912.ref007]\\].\n\nSo, agreement between trees would indeed have been some sort of \"proof\" of Darwin's TOL hypothesis, with two caveats. The first is that no widely accepted prokaryotic tree \"constructed on the basis of organismal biology\" was actually available for comparison: microbiologists had given up the attempt to make one in the mid-1950s, and no one since has been foolish enough to reboot the effort. The second is that agreement of molecular and organismal trees is not really a disproof of the theistic explanation Darwin wanted to supplant. Any sensible creator would surely use similar genes to make similar organisms.\n\nThe molecular phylogenetic tree-making enterprise has grown more spectacularly than Zuckerkandl and Pauling could have dared hope, and, in particular, the use of small-subunit ribosomal RNA (ssu rRNA) as \"molecular chronometer\" has revolutionized systematics, most importantly making it possible to put all prokaryotes and all eukaryotes into a single universal TOL. The story of how Carl Woese courageously pioneered the use of this molecule has been often told \\[[@pgen.1005912.ref010],[@pgen.1005912.ref011]\\]. And the discovery by Woese and George Fox that prokaryotes were deeply divided into Bacteria and Archaea, so that life appeared to comprise three domains, not two (Prokaryotes and Eukaryotes), has by now made it into all the textbooks.\n\nThat LGT might be a spoiler for the TOL was only a suspicion at the beginning of the era of molecular phylogenetics. Japanese clinical microbiologists found in the early 1960s that antibiotic resistance could spread between bacterial species via \"resistance transfer factors,\" now called plasmids \\[[@pgen.1005912.ref012]\\]. But at that time, such cases were thought to be special, disease-related, and possibly human-caused. Sporadic cases of incongruence between protein-coding and rRNA gene trees were subsequently discovered \\[[@pgen.1005912.ref013],[@pgen.1005912.ref014]\\], but it was probably the evidence from complete bacterial genome sequences that gave rise to the first serious questioning of the TOL. The claim that a quarter of the genes of the thermophilic bacterium *Thermotoga maritima* had been transferred from Archaea \\[[@pgen.1005912.ref015]\\] stunned many of us when it appeared in 1999, and prompted one of us to write that if...\n\n> \"...different genes give different trees, and there is no fair way to suppress this disagreement, then a species (or phylum) can 'belong' to many genera (or kingdoms) at the same time: There really can be no universal phylogenetic tree of organisms based on such a reduction to genes\" \\[[@pgen.1005912.ref002]\\].\n\nNearly Universal Trees and the Statistical Tree of Life {#sec002}\n=======================================================\n\nA decade and a half of prokaryotic tree-making has not produced general agreement on just how much LGT there is, other than \"much more than we expected,\" although there have been many serious attempts \\[[@pgen.1005912.ref016]\\]. Indeed, there is little agreement on how to quantify LGT as a process and its impact on genomes over all of evolutionary time. The most reasonable way, we think, is to ask of any contemporary genome how many of its genes have been strictly vertically inherited along a lineage of replicating genomes, tracing all the way back to that of the last universal common ancestor (LUCA). There are only about 100 genes that are found in all or nearly all prokaryotic genomes and thus have a strong claim to having been present continuously in all lineages \\[[@pgen.1005912.ref017],[@pgen.1005912.ref018]\\]. These are largely involved with the ribosome and its functions (or transcription), and are considered to be relatively immune to LGT, according to the \"complexity hypothesis\" \\[[@pgen.1005912.ref019]\\].\n\nThe complexity hypothesis is the notion that genes whose products interact with the products of many others will have so many coevolved molecular dependencies that function in a distantly related cellular environment is impaired. \"Connectivity\" (number of protein--protein interactions), not functional category, is in fact the key determinant \\[[@pgen.1005912.ref020]\\], and there are highly connected systems other than translation in many cells. So it may be equally relevant that translation and transcription are conservative and essential cellular activities, while many frequently transferred genes function in \"optional\" metabolic systems, representing alternative and therefore evolutionarily more changeable ways cells can make livings \\[[@pgen.1005912.ref018]\\].\n\nPuigbo et al. \\[[@pgen.1005912.ref021],[@pgen.1005912.ref022]\\] have for some time been conducting large-scale gene family tree comparisons across bacteria and archaea, the Forest of Life (FOL) project involving thousands of such families. Even among their 102 Nearly Universal Trees (NUTs), made from genes found in all or nearly all bacteria and archaea, there are very few with identical histories. However, NUT topologies are \"far more congruent than expected by chance,\" so \"they appear to reflect a significant central trend, an attractor in the tree space that could be equated with the STOL \\[the Statistical Tree of Life\\]\" \\[[@pgen.1005912.ref022]\\].\n\nPangenomics {#sec003}\n===========\n\nMost gene families are not among the NUTs, however, and even within a designated species, a sizeable fraction of any genome participates in rapid gene loss and gain (by LGT). The pangenome concept, which aims to describe the gene repertoire of a bacterial species by comparing gene contents of several to many of its strains, supports this notion. A typical bacterial genome comprises a \"core\" of genes present in all or nearly all of the strains of its species and often many more \"dispensable\" or \"accessory\" genes \\[[@pgen.1005912.ref023]\\] present in only some strains (as few as one). With *Escherichia coli* (several thousand genomes sequenced), the average strain carries about 5,000 genes. Genes shared between all or almost all strains number little more than 3,000, but the number of gene families with a representative in at least one *E*. *coli* genome is approaching 100,000 \\[[@pgen.1005912.ref024],[@pgen.1005912.ref025]\\]. *Prochlorococcus*, the world's most plentiful organism and most important oxygen-provider, shows an average genome content of only about 2,000 genes but a pangenome calculated at 85,000, so far \\[[@pgen.1005912.ref026],[@pgen.1005912.ref027]\\]. Archaea, too, boast pangenomes \\[[@pgen.1005912.ref028]\\], and we have come to think in terms of \"distributed\" prokaryotic genomes and \"shared genomic resource\" models, in which genes are lost when superfluous and regained when needed, by LGT.\n\nAccording to a \"strong version\" \\[[@pgen.1005912.ref029]\\] of the popular Black Queen hypothesis \\[[@pgen.1005912.ref030]\\], different strains (or species) will lose different genes for \"leaky\" synthetic functions---those whose products leak out of cells and thus allow \"cross-feeding.\" They become mutually dependent. Several recent metagenomic surveys have uncovered heterogeneous populations of previously unknown small-genomed bacteria or archaea that may be metabolically dependent on each other in this way \\[[@pgen.1005912.ref031],[@pgen.1005912.ref032]\\]. In line with this, Wolf and Koonin \\[[@pgen.1005912.ref033]\\] suggest a general model in which gene loss is the dominant mode of prokaryotic evolution: what saves genomes from extinction is periodic genomic expansion, in part via multiple simultaneous (or nearly simultaneous) LGT events. Whether such a punctuational pattern holds or gene gain and loss are more regularly ongoing processes, the result is that often less than half of the genes in any strain's genome are likely to have been continuously present along a lineage of genomes stretching back to the first member (common ancestor) of that species.\n\nAs we go deeper than species into the TOL, the fraction of vertically inherited genes can only get smaller. In an influential 2009 analysis, Lapierre and Gogarten extended and deepened the pangenome analysis to 573 bacterial genomes spanning the then available diversity \\[[@pgen.1005912.ref034]\\]. The core, about 250 genes, which would have included the NUTs, comprises 8% of the typical bacterial genome and would shrink further if archaea were included. Among dispensable genes, they called 64% \"character genes\"---\"essential for colonization and survival in particular environmental niches\" and found in related species---while the 28% more rapidly turning-over genes were strain-specific \"accessory genes.\"\n\nDreams of LUCA {#sec004}\n==============\n\nWhat, then, did LUCA's genome contain? At one extreme, one might imagine that all contemporary gene families have their coalescents (roots) or direct ancestors thereof, in the genome of LUCA. On this model, either (i) that genome was large, and descendant lineages have differentially lost gene families, or (ii) many important and widely distributed gene families have evolved since LUCA and are not detectably homologous to other gene families with which they share common origins. Mixtures of (i) and (ii) might also be entertained.\n\nAt the other extreme, one could hold that LUCA as a cell had a \"normal\" or even smallish prokaryotic genome (1,500--5,000 genes) with only 100 or so genes (represented in NUTs) that might potentially have been passed down directly to all or almost all contemporary genomes. This would be a \"uniformitarian\" model, explaining the past in terms of processes familiar from the present. In all descendant lineages, the majority of LUCA's genes will have been replaced via LGT---from closely related and distant lineages of other genes, homologous or not, carrying out the same or quite different cellular functions (structural and metabolic). A key difference from the first model is that some contemporary gene families will have had coalescents that predate LUCA \\[[@pgen.1005912.ref035]--[@pgen.1005912.ref037]\\]. That is, the donors for many LGT events would have been cellular lineages that have since gone extinct, because LUCA is by this and the previous conceptualization the last universal common cellular (not necessarily genomic) ancestor.\n\nMore explicitly, according to the second model, the fraction of any particular contemporary prokaryote's genes that have been strictly vertically inherited along a lineage of replicating genomes tracing all the way back to LUCA would comprise a universally conserved core (roughly the NUTs) plus \"character genes\" defining its metabolic lifestyle---methanogenesis genes if both LUCA and the contemporary prokaryote were methanogens, for instance. If the contemporary prokaryote we picked was instead a cyanobacterium while LUCA was methanogenic, non-core shared genes would be substantially fewer. Given that even within the NUTs there is some incongruence of trees, the fraction of a randomly chosen or \"average\" contemporary prokaryote's genes that have been inherited only vertically since LUCA could easily be only a small percentage.\n\nA third model, favored by some, is to redefine LUCA as a community, not a single cell or species \\[[@pgen.1005912.ref038],[@pgen.1005912.ref039]\\]. Although in our view conflating \"having common ancestry\" with \"having a common ancestor,\" this model points out the basic conflict implicit in reconstructing the TOL from the sequences of genes that disagree. We are using genes as a proxy. What we have really been after is a tree of entities at some level of the biological hierarchy higher than genes or genomes---cells, organisms, populations, or (most likely) species. Perhaps our quest should be recast as simply an attempt to retrace life's history at such a higher level, and not as an effort to achieve the proof of the \"TOL Hypothesis\" that Zuckerkandl and Pauling thought gene sequences could provide. Perhaps we have already unknowingly done this: Tassy recently mused that \"we enter a world of non systematic phylogenetics, a surprising oxymoron\" \\[[@pgen.1005912.ref040]\\].\n\nTrees of Cells and Species {#sec005}\n==========================\n\nIf one assumes, almost as an entailment of Schleiden and Schwann's cell theory, that all cells derive from previous cells, then it could be the case that there is a single tree-like pattern relating all cells, although one would need to look up to the level of \"species\" in lineages where sex became a reproductive necessity. Although incongruence of gene trees due to LGT means that no single one can be taken as topologically equivalent to such a tree of cells and species, the STOL might still be the best (though not a guaranteed) proxy for it. Patrick Forterre endorses this common view when he writes:\n\n> \"The universal tree should depict evolutionary relationships between domains defined according to the translation apparatus reflecting the history of cells (and their envelope) and not according to the global genome composition that is influenced by LGT, virus integration and endosymbiosis, the history of which is incredibly complex\" \\[[@pgen.1005912.ref041]\\].\n\nBut we might also ask what warrant there is to believe in such a tree-like history of cells, coupled or not to molecular phylogenies. Reticulating events that would compromise trees at the cellular level are occasional or regular whole-cell fusions, mixing entire genomes and, more contentiously, conjugation and the compromising of cellular integrity represented by phage attachment or transformation. These latter introduce only a little genetic material each time, but when summed over many events, can replace every gene: the argument would be over whether the pattern of material as well as genetic relationships might be seen as non-tree-like.\n\nInterspecies whole-cell fusion resulting in heterodiploid cells does occur in haloarchaea \\[[@pgen.1005912.ref042]\\] but is probably rare in other prokaryotes. Rare does not mean evolutionarily unimportant, however. Whatever theory we might support as to the origin of eukaryotes, we must acknowledge that something like a fusion of (presumably unicellular) archaeal and bacterial cells was instrumental in this most \"basic\" divergence. Although a claim that many major archaeal phyla were independently founded by the simultaneous importation of many bacterial genes \\[[@pgen.1005912.ref043]\\] is now in doubt \\[[@pgen.1005912.ref044]\\], such events as the acquisition of suites of genes for respiration by the anaerobic methanogenic ancestors of haloarchaea are hard to reconcile as one-gene-at-a-time LGT events. And there is no doubt that a fusion of cyanobacterial and eukaryotic cellular lineages ultimately gave rise to higher plants and, through secondary and tertiary endosymbiotic events, to several algal lineages.\n\nSo, even a tree of cellular lineages is not an unproblematic concept. Students of animals and plants have long accepted that incomplete lineage sorting, introgression, and full-species hybridization pose difficulties for the sorts of trees that Darwin might have had us draw. But it is microbes, with their promiscuous willingness to exchange genes between widely separated branches of any \"tree,\" that have most seriously jeopardized the neo-Darwinian synthesis, in the oversimplified form that we have often presented it to the public \\[[@pgen.1005912.ref045]\\]. More sophisticated understandings do remain possible \\[[@pgen.1005912.ref003]\\] and should be debated in a more conceptually and science-historically self-aware context \\[[@pgen.1005912.ref004]\\].\n\nWe thank Laura Eme and Olga Zhaxybayeva for comments on this manuscript.\n\n[^1]: The authors have declared that no competing interests exist.\n"}
+{"text": "1. Introduction {#sec1-animals-09-00599}\n===============\n\nThe intensity and specialization of beef cattle production make off-site fattening, introduce new breed and commercial trade more and more common, and need transportation to achieve the goals \\[[@B1-animals-09-00599]\\]. During animal's transportation, the nutrition-metabolism balance, hormone secretion levels, and immune competence are imbalanced due to animals' transport stress. It affects meat quality, risk of disease, and even causes death \\[[@B2-animals-09-00599]\\]. In the south of China, the land is unevenly distributed, especially in the Sichuan provenience, and there are a number of feed ingredients which are used for feeding of animals. Therefore, Sichuan province is an ideal place for beef cattle production. A previous study indicated that the health of cattle was seriously affected by the transportation of 15 h \\[[@B3-animals-09-00599]\\]. Breed is one factor can affect animals production, immune function and anti-reversion \\[[@B4-animals-09-00599],[@B5-animals-09-00599]\\].\n\nThere are a number of different breeds of beef cattle in China, from the three breeds were used in the present study. Native Yellow Cattle (NY: Xuanhan Yellow Cattle) are a unique local breed which were domesticated in China for thousands of years ago and have adapted to the local natural environmental conditions; they account for about 80% of indigenous breeds \\[[@B6-animals-09-00599]\\]. Xuanhan Yellow Cattle is one of the indigenous yellow cattle live in southern parts of China, and the body of yellow cattle is not massive compared with the introduced breeds such as Simmentals. Cattle Yak (CY: Jersey\u00d7Maiwa Yak) is a hybrid of yak and NY. This breed is found at Qinghai-Tibet Plateau and adapted the cold and hypoxia environmental conditions. These animals provide milk, labor force, meat, and pelage with the local herdsman. Finally, Simmental Crossbred Cattle (SC: Simmental \u00d7 Xuanhan) is an improved crossbreed of Simmental cattle and Xuanhan cattle that generally has better production performance. However, no previous research was conducted on different anti-stress abilities among these breeds. Transport stress reduces cattle immunity. Immune stimulation for most cell activation causes inflammatory factors, such as TNF-\u03b1, IL-4, IL-6, and IL10, to be released to the serum, as well as oxidative stress increases the expression and production of pro-inflammatory cytokines (IL-1\u03b2 and TNF-\u03b1) \\[[@B2-animals-09-00599]\\]. Similarly, animals exposed to stressful environments emit gastrointestinal pathogens that can induce increases in the secretion of virulence factors \\[[@B7-animals-09-00599]\\]. The central nervous system via the hypothalamic-pituitary-adrenal system controls the secretion of hormones under stress conditions, stimulating ACTH and releasing cortisol \\[[@B8-animals-09-00599]\\]. More than 100 trillion bacteria, including thousands of different species, have been reported in mammalian gastrointestinal tracts. In ruminants, microbiota and organ systems react to secreted hormones and stimulate the body lead to immune response. Rumen microbes play a very important rule in digest fiber and maintain homeostasis, (transport stress has been reported to affect ruminant microbiota abundance (cellulolytic bacteria, *Ruminococcus amylophilus,* and *Prevotella albensis*) \\[[@B9-animals-09-00599],[@B10-animals-09-00599],[@B11-animals-09-00599]\\]. Microbiome are signaling hub that can integrate many information such as diet, stress and genetic, commensal bacteria and epithelial cells are the most important elements for the immune system, especially the bacteria reside on a single layer of the intestinal epithelial cells \\[[@B12-animals-09-00599],[@B13-animals-09-00599]\\].\n\nFactors such as epithelial permeability and the expression of critical factors, including tight-junction and antimicrobial proteins, are regulated by intestinal epithelial cells and influenced by CD4^+^T helper cells \\[[@B13-animals-09-00599],[@B14-animals-09-00599]\\]. Accordingly, the present study examined the anti-transport stress ability among the three different cattle breeds, in terms of nutrition-metabolism balance, hormone secretion level, and immunocompetence. It is also evaluated the changes in rumen microbiota through high-throughput sequencing.\n\n2. Materials and Methods {#sec2-animals-09-00599}\n========================\n\n2.1. Animal Care and Study Design {#sec2dot1-animals-09-00599}\n---------------------------------\n\nAll procedures of the experiment were approved by the Animal Care and Use Committee of Animal Nutrition Institute, Sichuan Agriculture University, and followed the current laws of animal protection (Ethics Approval Code: SCAUAC201408-3).\n\nA total of 18 male beef cattle with similar body condition (20 months of age with consistent feeding methods) were used: 6 SC (Simmental \u00d7 Xuanhan yellow cattle crossbred cattle), 6 NY (NY: Xuanhan Yellow Cattle) cattle, and 6 CY (Jersey \u00d7 Maiwa yak crossbred cattle), and selected 5 cattle used for rumen microorganism detection in every groups. All cattle originated a farm located 350 km from the Sichuan Agricultural University in Ya'an city (coordinates 29\u00b059\u203258.96\u2033 N, 103\u00b00\u203233.65\u2033 E, and 580 m altitude) and were transported for 6 h (h: hours) at an average speed of 50--60 km/h to Nine Cattle Company in Changning County, located in Sichuan province (coordinates 28\u00b040\u203224.46\u2033 N, 104\u00b058\u203253.55\u2033 E, and 261.12 m altitude). The journey started at 18:00 on 19 November 2016, and the cattle arrived at 24:00 on 20 November 2016. The outside temperature was between 7--15 \u00b0C with a relative humidity of 75--90%. Animals were deprived of food and water during transportation. Samples were taken 4 h before transportation and immediately after arrival.\n\n2.2. Blood and Rumen Fluid Sample Collection and Storage {#sec2dot2-animals-09-00599}\n--------------------------------------------------------\n\nRumen fluid and blood were sampled at 09:00 on 19 November 2016, and at 24:00 on 20 November 2016. The blood samples were collected from the jugular vein of the beef cattle. Samples were kept refrigerated on ice until all the samples were collected, then processed immediately at the laboratory and were centrifuged at 4000 rpm for 15 min to obtain serum and frozen at \u221220 \u00b0C for the testing index. The rumen fluid was sampled through the oral cavity into the rumen with an oral collector. The first 300 mL of rumen fluid was discarded to avoid reticulum fluid or salivary contaminated fluid or body surface bacteria, and then 400 mL of rumen fluid was collected and squeezed through four layers of gauze and tested pH (HJ-90B, Aerospace Computer Company, Beijing, China) using a precision strip test immediately. 0.25 mL of metaphosphoric acid was added to 1 mL rumen fluid and centrifuged at 15,000\u00d7 *g* for 15 min by gas chromatography (GC-MS, Agilent Technologies, Palo Alto, CA, USA) for detection of acetic, propio nic and butyric acids concentrations. Finally, the samples were frozen in liquid nitrogen and stored at \u221280 \u00b0C until DNA extraction. Serum Cortisol (COR), adrenocorticotrophic hormone (ACTH) detected by double-antibody radioimmunoassay (RIA); Serum T3, T4, IgG, IgA, TNF-\u03b1, IL-1\u03b2, IL-6, IL-10, IL-4 and lipopolysaccharide (LPS) a double antibody sandwich ELISA (ELISA kit: Shang HaiLengton Bioscience Co, Ltd., Shanghai, China) and LPS was detected both in rumen fluid and serum.\n\n2.3. DNA Extraction {#sec2dot3-animals-09-00599}\n-------------------\n\nRumen fluid (1 mL) was centrifuged at 12,000\u00d7 *g* for 10 min at 4 \u00b0C for DNA extraction, using a QIAamp DNA kit (Omega Bio-Tek, Norcross, GA, USA) according to the manufacturer instructions. DNA extracts were dissolved in 200 mL EB buffer, and then the quality and quantity of the extracted DNA were determined by UV spectrophotometric analysis using a NanoDrop ND-1000 Spectrophotometer (Nyxor Biotech, Paris, France). DNA used in subsequent experiments had to present an A260/A280 ratio between 1.7 and 1.9, indicating intact and highly pure DNA. All DNA samples were stored at \u221280 \u00b0C.\n\n2.4. PCR Amplification, Library Construction and Illumina Sequencing {#sec2dot4-animals-09-00599}\n--------------------------------------------------------------------\n\nThe V4 regions of the 16S rRNA gene were amplified using primers Arc Primer5'\u21923':515F(5'-GTGCCAGCMGCCGCGGTAA-)and Arc 806R(5'-GGACTACH VGGGTW TCTA AT-3') \\[[@B15-animals-09-00599]\\] (PCR adoption KOD-401B: TOYOBO KOD-Plus-Neo DNA Polymerase, PCR instrument: Applied Biosystems^\u00ae^ Gene Amp^\u00ae^ PCR System 9700 (Thermo Fisher Scientific, Massachusetts, MA, USA). 16S rRNA genes were amplified using the specific primer with a 12 nt unique barcode. The PCR mixture (25 \u03bcL) contained 1\u00d7 PCR buffer, 1.5 mM MgCl~2~, 0.4 \u03bcM deoxynucleoside triphosphate, 1.0 \u03bcM of each primer, 0.5 \u03bcM of KOD-Plus-Neo (TOYOBO, New England Biolabs, Beijing, China), and 10 ng template DNA. The PCR amplification procedure consisted of initial denaturation at 94 \u00b0C for 1 min, followed by 30 cycles (denaturation at 94 \u00b0C for 20 s, annealing at 54 \u00b0C for 30 s, and elongation at 72 \u00b0C for 30 s), and a final extension at 72 \u00b0C for 5 min. Three replicates of PCR reactions for each sample were combined. PCR products mixed with 1/6 the volume of 6\u00d7 loading buffer were loaded on 2% agarose gel for detection. Samples with a bright main strip between 200--400 bp were chosen for further experiments.\n\n2.5. Library Preparation and Sequencing {#sec2dot5-animals-09-00599}\n---------------------------------------\n\nSequencing libraries were generated using a TruSeq DNA PCR-Free Sample Prep Kit (Illumina, San Diego, CA, USA), following the manufacturer's recommendations, and index codes were added. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Fisher Scientific, Massachusetts, MA, USA) and Agilent Bioanalyzer 2100 system. Lastly, the library was applied to paired-end sequencing (2 \u00d7 250 bp) with the Illumina Hiseq apparatus at Rhonin Biosciences Co (Rhonin Biotechnology Ltd., Chengdu, China) \\[[@B16-animals-09-00599]\\]. Each PCR reaction terminated in a linear amplification period was repeated three times. After PCR amplification, the PCR product of the same sample was mixed then detected with 5 V/cm, 20 min (1% agarose gel electrophoresis), using an OMEGA Gel Extraction Kit (Rhonin Biotechnology Ltd., Chengdu, China) gel to cut the PCR products and TE buffer elution recovery to target the DNA fragment. Finally, library construction was completed by the Illumina TruSeq DNA PCR-Free Sample Prep Kit (New England Biolabs, Beijing, China). Next-generation sequencing was performed by the Illumina Hiseq 2500 PE250, which was conducted by an Illumina Hiseq Rapid SBS Kit v2 (New England Biolabs, Beijing, China).\n\n2.6. Bioinformatics and Statistical Analysis {#sec2dot6-animals-09-00599}\n--------------------------------------------\n\n1\\. Paired end reads assembly. The sequences were analyzed according to Usearch (v8.1.1756, <http://www.drive5.com/usearch/>) and the quantitative insights into microbial ecology (QIIME) \\[[@B15-animals-09-00599]\\] pipeline. Paired-end reads from the original DNA fragments were merged using FLASH \\[[@B17-animals-09-00599]\\]. Then, sequences were assigned to each sample according to the unique barcode.\n\n2\\. OTUs clustering and taxonomy assignment. We adopted relatively stringent quality controls. We first filtered low-quality reads (length \\<200 bp, more than two ambiguous base 'N' or average base quality score \\<30) and truncated sequences where quality scores decayed (score \\<11). After finding the duplicated sequences, we discarded all the singletons, which may be bad implications (<http://www.drive5.com/usearch/manual/singletons.html>) and lead to overestimation of diversity. Sequences were clustered into operational taxonomic units (OTUs) at a 97% identity threshold using UPARSE (version 7.1 <http://drive5.com/uparse/>) algorithms \\[[@B18-animals-09-00599]\\]. We picked representative sequences and removed potential chimeras using the Uchime algorithm \\[[@B19-animals-09-00599]\\]. Taxonomy was assigned using the Silva database \\[[@B20-animals-09-00599]\\] and the uclust classifier in QIIME. Representative sequences were aligned using PyNAST \\[[@B21-animals-09-00599]\\] embedded in QIIME.\n\nIn case the sequencing depth influenced community diversity, the OTU table was rarified to make all samples hold the same sequence number. We tested 10611 sequences for every sample and total 13,797 OTU. All data analyses were performed using R or Python \\[[@B22-animals-09-00599],[@B23-animals-09-00599]\\]. The random seed number was fixed at 1234. Phylogenetic diversity (PD) \\[[@B16-animals-09-00599]\\] was calculated following Picante \\[[@B22-animals-09-00599]\\]. Weighted and Un-weighted Unifrac distances were calculated in GUniFrac \\[[@B23-animals-09-00599]\\]. Other alpha- and beta-diversity metrics were calculated in Vegan \\[[@B23-animals-09-00599]\\]. Rarefaction curves were generated based on these three metrics. Principal component analysis (PCoA) was applied to reduce the dimensions of the original community data. Principal Coordinate Analysis (PCoA) and Non-metric Multi-Dimensional Scaling (NMDS) were performed using the Ape \\[[@B24-animals-09-00599]\\] and Vegan packages, respectively. Hierarchical cluster analysis was done using the R function hclust. Random Forest analyses were done using.\n\n2.7. Correlation Between Rumen Microbiota and Cattle Physiological Variables {#sec2dot7-animals-09-00599}\n----------------------------------------------------------------------------\n\nNon-parametric Spearman rank correlation coefficient analysis implemented in PAST software was used to analyze the relationship between Serum hormones, immune function and Rumen fluid characteristics and bacterial communities in rumen fluid. The resulting correlation matrix was visualized in a heat-map format generated by the corr-plot package of R (Corrplot: visualization of a correlation matrix, R package version 02-0. 2010) \\[[@B22-animals-09-00599]\\].\n\nThe results were analyzed with the Statistical Package for Social Sciences (SPSS, version 22.0, IBM, Armonk, NY, USA). A single-factor analysis of variance (ANOVA) and two-factor analysis variance were used to analyze all data, and Duncan's multiple comparisons were used to separate means where significant differences were found. A value of *p*-values \\< 0.05 was considered statistically significant. Data are shown as mean, mean \u00b1 standard deviations (SD).\n\n3. Results {#sec3-animals-09-00599}\n==========\n\n3.1. Serum Hormones {#sec3dot1-animals-09-00599}\n-------------------\n\nThe results of serum hormones of the present study are presented in [Table 1](#animals-09-00599-t001){ref-type=\"table\"}, there were no significant differences (*p* \\> 0.05) between different breeds of cattle in terms of the concentration of COR and ACTH before transportation, the concentration of ACTH was significantly (*p* \\< 0.05) lower in the CY group than the NY and SC groups after transportation. Transport stress led to consistent and significant (*p* \\< 0.05) increases in the concentrations of COR and ATCH, but there was no noticeable difference across breeds in COR levels after transportation were observed [Table 1](#animals-09-00599-t001){ref-type=\"table\"}. Transport stress significant (*p* \\< 0.05) affects the secretion of COR while the breed factor affects the secretion of COR and ACTH were not significant (*p* \\> 0.05) Table 5. Meantime, after transportation, the concentration of T3 and T4 were decreased among all the cattle breed groups, but only T3 were significant (*p* \\> 0.05) decreased in SC group.\n\n3.2. Rumen Fluid Characteristics {#sec3dot2-animals-09-00599}\n--------------------------------\n\nThe results of rumen fluid characteristics are presented in [Table 2](#animals-09-00599-t002){ref-type=\"table\"}. There was no significant difference (*p* \\> 0.05) for rumen pH among three different breeds before transportation, but after transportation the pH was reduced in the NY, SC, and CY groups and there was significant (*p* \\< 0.05) reduced in SC, and CY groups, in the NY had higher (*p* \\< 0.05) pH than SC and CY groups after transportation. The concentration of serum LPS (lipopolysaccharide) was increased in the NY, SC, and CY groups after transportation, but only in CY group significantly increased (*p* \\< 0.05). The factor of breed affects serum LPS significant (*p* \\< 0.05). Meanwhile, there were higher concentrations of rumen LPS after transportation in NY, CY, and SC groups than before transportation, and in SC and CY groups were significantly increased (*p* \\< 0.05), but there were no significant (*p* \\> 0.05) differences in NY group after transportation. For rumen volatile fatty acid, the concentration of acetic acid, propionic acid, butyric acid, and the acetic to propionic acid ratio were increased after transportation, furthermore, after transportation in NY, SC, and CY groups had a significant (*p* \\< 0.05) higher concentrations of total volatile fatty acid (TVFA) than before transportation. The concentration of acetic acid was significant (*p* \\< 0.05) affected by the factors of transportation and breed, but there was non-significant (*p* \\> 0.05) differences among three cattle breeds before transportation. The concentration of propionic acid was non-significant (*p* \\> 0.05) differences among three cattle breeds before transportation, meanwhile after transportation all groups had a significant (*p* \\< 0.05) higher concentrations of propionic acid than before transportation. After transportation the ratio of acetic to propionic acids was significant (*p* \\< 0.05) increased in NY and CY groups, while there was non-significant (*p* \\> 0.05) in SC group. The CY group had a significant (*p* \\< 0.05) higher concentrations of rumen lactic acid than SC and NY groups, meantime, SC had a significant (*p* \\< 0.05) higher concentrations of rumen lactic acid than NY group before transportation, and after transportation the concentrations of rumen lactic acid was decreased in SC group (*p* \\< 0.05), while the concentration of serum lactic acid was significant (*p* \\< 0.05) increased in all cattle breeds after transportation.\n\n3.3. Immunity Levels {#sec3dot3-animals-09-00599}\n--------------------\n\nAfter transportation the concentration of IgG significantly (*p* \\< 0.05) decreased in all groups and presented in [Table 3](#animals-09-00599-t003){ref-type=\"table\"}, but there were non-significant (*p* \\> 0.05) difference among the NY, SC and CY groups before and after transportation. The factor of transportation significantly (*p* \\< 0.05) affects the level of IgG in serum (Table 5). Similarly, the concentration of IgA was reduced in three different cattle breeds after transportation; however, there was non-significant (*p* \\> 0.05) effect by the factor of transportation. For the pro-inflammatory cytokines, TNF-\u03b1, IL-1\u03b2, and IL-6, the concentrations were increased significantly (*p* \\< 0.05) after transportation in all cattle breeds. The CY group had a lower level of IL-1\u03b2 than other groups before transportation (*p* \\< 0.05). The anti-inflammatory cytokines IL-10 and IL-4 were increased after transportation, but IL-10 did not increase significantly (*p* \\> 0.05) for all cattle breeds, and the concentration of IL-4 was increased significantly (*p* \\< 0.05) only in CY group after transportation. However, there was significant (*p* \\< 0.05) higher concentration of IL-4 in NY than CY groups before transportation, and for IL-10 in NY group had a significant (*p* \\< 0.05) higher concentration than both SC and CY groups before transportation.\n\n3.4. Alpha-Diversity Measures and OTU Analysis {#sec3dot4-animals-09-00599}\n----------------------------------------------\n\nThe alpha diversity index estimation of the 16S rRNA gene libraries of the beef cattle rumen emerged from the sequencing analysis are presented in [Table 4](#animals-09-00599-t004){ref-type=\"table\"}. To obtain the taxonomic information for each OTU, the uclust taxonomy was used for analysis, and the default algorithm sought 97% OTU representative sequences of similar levels.\n\nThe richness of rumen microbiota was based on 97% similarity cut-off and normalization. The continued to be total 1,263,210-row sequences and total 1,217,669 clean sequences; every sample includes 33,824.13 clean sequences, 84,852 reads after filtration, with a mean of 2357 reads per sample were detected. In the current study, the results showed significantly (*p* \\< 0.05) different by the chao1 index in SC group before and after transportation, while both breed and transport factors were non-significant (*p* \\> 0.05) affect the chao1 index. After transportation, the Shannon index was increased for CY and SC groups, but only the SC group changed significantly (*p* \\< 0.05) and had the highest Shannon index after transportation. Additionally, NY group was significantly (*p* \\< 0.05) higher than other groups about OTUs before transportation, all OTUs were increased among all cattle breeds after transportation, and only SC group increased significant (*p* \\< 0.05), meanwhile, the factor of breeds affects the number of OTUs very significant (*p* \\< 0.05) [Table 5](#animals-09-00599-t005){ref-type=\"table\"}. The beta diversity was further analyzed by principal coordinate analysis (PCoA). In [Figure 1](#animals-09-00599-f001){ref-type=\"fig\"}, PCO1 plots of bacterial 16S rRNA showed obvious clusters among different cattle breeds before transportation ([Figure 1](#animals-09-00599-f001){ref-type=\"fig\"}A), and percent variation explained 20.3% and 12.9%, it also showed visible clusters between BNY and ANY, BSC and ASC, BCY and ACY before and after transportation in [Figure 1](#animals-09-00599-f001){ref-type=\"fig\"}B--D, respectively. According to the Venn diagram at a 97% similarity level, 2444 and 2357 OTUs (86.87% and 85.21% of the total sequences) were common to before and after transportation among NY, SC and CY groups, respectively. The NY group shared 787 (2.19% of sequences) and 789 (2.31% of sequences) OTUs, with SC and CY groups before transportation, respectively. Compared with before transportation, the OTUs (sequences) unique in NY, SC and CY groups were 2801(5.04% of sequences), 3333 (8.02% of sequences), and 1815 (2.98% of sequences) after transportation, respectively ([Figure 2](#animals-09-00599-f002){ref-type=\"fig\"}).\n\n### Microbiota Composition of Rumen\n\nAll sequences were classified from phylum to species based on the SILVA (From Latin silva, forest, <http://www.arb-silva.de>) taxonomic database and using the analytical program QIIME. We assigned the OTUs to 22 phyla, 38 classes, 59 orders, 97 families, 237 genus, and 338 species. High-throughput sequencing was showed 2477 and 2376, 2255 and 2755, 2093 and 2184 OTUs (defined at 3% dissimilarity in UPARSE) before and after transportation in the NY, SC, and CY groups, respectively. A few common kinds of bacteria and phyla were found by taxonomic analysis such as Bacteroidetes and Firmicutes (\\>18%) were the most dominant phyla before and after transportation in groups. The highest relative abundance reached to 80%, the lowest reached to 70% for Bacteroidetes for some samples, there was no significant (*p* \\> 0.05) difference among NY, SC, and CY groups before transportation, while after transportation the relative abundance of Bacteroidetes were reduced in all groups, but only in NY group reduced very significantly (*p* \\< 0.05) (from 78.43% to 73.47%) ([Figure 3](#animals-09-00599-f003){ref-type=\"fig\"} and [Supplementary Table S1](#app1-animals-09-00599){ref-type=\"app\"}). However, the relative abundance of Firmicutes were increased in all groups after transportation, and there were increased significantly (*p* \\< 0.05) in SC group (from 18.80% to 22.87%), but there was no significant (*p* \\> 0.05) difference among NY, SC, and CY groups before transportation. SC group had the highest relative abundance about proteobacteria (up to 1.73%) but non-significant (*p* \\> 0.05) difference before transportation among the groups.\n\nAt the genus level, 237 genera were observed in the rumen, and dominant genera included *Prevotella 1* (all sample average 25.21%), *Rikenellaceae RC9* gut group (18.87%), *Christensenellaceae R-7* group (2.46%), **Lactobacillus* (2.29%), Lachnospiraceae NK4A136* group (1.44%), the *Ruminococcaceae NK4A214* group (1.36%), *Butyrivibrio 2* (1.14%), and *Ruminococcus* 1 (0.82%). Transportation leads to the abundance of *Prevotella 1* significant (*p* \\< 0.05) reduced in all NY, SC, and CY groups. Meanwhile the NY group had the significant (*p* \\< 0.05) lower (only 25.266%) than both SC (30.81%) and CY (29.645%) groups before transportation. Before transportation in SC group (8.45%) for the abundance of *Rikenellaceae RC9* gut group was significant (*p* \\< 0.05) lower than both NY (17.717%) and CY (15.684%) groups before transportation, and the abundance of *Christensenellaceae R-7* was increased after transportation in all groups including NY (from 2.143% to 3.261%) and SC (from 2.115% to 2.72%) groups were changed significant (*p* \\< 0.05). For *Lactobacillus* CY group (2.488%) was higher (*p* \\< 0.05) than NY group (1.923%), and transportation leads the abundance increased in all groups but only in NY group increased significant (*p* \\< 0.05). There was reduced for the abundance of *Butyrivibrio 2* in all three different cattle breeds, and in NY and CY groups were significant (*p* \\< 0.05) decreased. To identify the specific bacterial taxa associated with breeds and transportation, we compared microbiota among NY, SC, and CY groups before transportation, and between BNY and ANY, BSC and ASC, BCY and ACY using the linear discriminant analysis effect size (LEfSe) method. [Figure 4](#animals-09-00599-f004){ref-type=\"fig\"}, (A) shows a representative cladogram of the structure of predominant bacteria, showing the most significant differences in taxa between the three groups (NY, SC and CY before transport). The data indicated that eight specific genera, five specific Family, three specific Order, and two specific Class belong to two dominant phyla (Spirochaetae and Tenericutes) in BSC groups, 11 specific genus, five specific Family, two specific Order, and two specific Class belong to two specific dominant phyla (Saccharibacteria and Lentisphaerae) in BNY groups and in BCY group only five specific genus belong to one specific Family. In addition, to identify the specific bacterial taxa associated with transport stress ([Figure 4](#animals-09-00599-f004){ref-type=\"fig\"}B--D), predominant bacteria between BNY and ANY, BSC and ASC, BCY and ACY were shown. 9 specific genus belong to one and one specific phyla Spirochaetae and Lentisphaerae were found in BNY, and ANY groups ([Figure 4](#animals-09-00599-f004){ref-type=\"fig\"}B), 6 and 11 specific genera belong to two specific families (Bacteroidales and Christensenellaceae) and one specific class (Gammaproteobacteria) was found in BCY and ACY groups ([Figure 4](#animals-09-00599-f004){ref-type=\"fig\"}C) respectively. Fifteen and 16 specific genera belong to two (Firmicutes and Actinobacteria), and four (Euryarchaeota, Deferribacteres, Spirochaetae and Proteobacteria) specific phyla were found in BSC and ASC groups ([Figure 4](#animals-09-00599-f004){ref-type=\"fig\"}D).\n\n3.5. Correlation Between Rumen Microbiota and Physiological Variables {#sec3dot5-animals-09-00599}\n---------------------------------------------------------------------\n\nThe relationship between ruminal microbiota abundance \\[representing at least 0.1% of the bacterial community in at least one sample (in phyla and genus level)\\] and physiological parameters were used to analyze the evaluated correlations.\n\nThe results showed ([Figure 5](#animals-09-00599-f005){ref-type=\"fig\"}) that the rumen pH was negatively correlated with *f. Ruminococcaceae* (R = \u22120.514, *p* \\< 0.05), *Ruminococcaceae UCG-002* (R = \u22120.782, *p* \\< 0.01), *Saccharofermentans* (R = \u2212734, *p* \\< 0.01) *Lactobacillus* (R = \u22120.596, *p* \\< 0.01), f. *Bacteroidales BS11 gut group* (R = \u22120.695, *p* \\< 0.01), positively correlated with *Prevotella1*(R = 0.547, *p* \\< 0.01), *Butyrivibrio2* (R = 0.521, *p* \\< 0.01). Rumen LPS were negatively correlated with *Prevotella 1* (R = \u22120.627, *p* \\< 0.01), and *Butyrivibrio2* (R = \u22120.528, *p* \\< 0.01), positively correlated with *Ruminococcaceae UCG-002* (R = 0.472, *p* \\< 0.05), *Lactobacillus* (R = 0.521, *p* \\< 0.01). The rumen fermentation characteristics of acetic acid were negatively correlated with *Prevotella 1* (R = \u22120.527, *p* \\< 0.01), *Butyrivibrio2* (R = \u22120.577, *p* \\< 0.01), and positive correlated with *Ruminococcaceae UCG-002*(R = 0.543, *p* \\< 0.01), *Saccharofermentans* (R = 0.648, *p* \\< 0.01). Acetic:propionic acids were negatively correlated with *Prevotella 1* (R = \u22120.467, *p* \\< 0.01) and f.*Bacteroidales S24-7 group* (R = \u22120.577, *p* \\< 0.01). TVFA was negatively correlated with *Prevotella 1* (R = \u22120.649, *p* \\< 0.01) and *Butyrivibrio 2* (R = \u22120.565, *p* \\< 0.01) while positive correlated with *Saccharofermentans* (R = 0.648, *p* \\< 0.01) and *Ruminococcaceae UCG-002* (R = 0.539, *p* \\< 0.01). The concentration of rumen lactic acid was positive correlated with *f.Bacteroidales BS11 gut group* (R = 0.529, *p* \\< 0.01), *Lactobacillus* (R = 0.555, *p* \\< 0.01), *Ruminococcaceae UCG-002* (R = 0.609, *p* \\< 0.01). For the immune system, the abundance of *Prevotella 1* was positively correlated with IgA (R = 0.547, *p* \\< 0.01) while negatively correlated with pro-inflammatory cytokines IL-6 and TNF-\u03b1 (R = \u22120.476, *p* \\< 0.01, R = \u22120.452, *p* \\< 0.05 respectively). The abundance of f.*Bacteroidales BS11 gut group* was positively correlated with pro-inflammatory cytokines IL-6 and IL-1\u03b2 (R = 0.585, *p* \\< 0.01, R = 0.378, *p* \\< 0.05 respectively), meantime positively correlated with anti-inflammatory cytokines IL-4 and IL-10 (R = 0.618 and 602, *p* \\< 0.01, respectively). *Lactobacillus* was positively correlated with IL-6 (R = 0.445, *p* \\< 0.05), serum lactic acid (R = 0.602, *p* \\< 0.01), and IL-4 (R = 0.557, *p* \\< 0.01), while negatively correlated with IgA (R = \u22120.596, *p* \\< 0.01), *Butyrivibrio 2* negatively correlated with serum lactic acid (R = \u22120.561, *p* \\< 0.01), anti-inflammatory cytokines IL-4 and IL-10 (R = \u22120.579 and -0.565, *p* \\< 0.01, respectively). While *Saccharofermentans* was positively correlated with IL-4 and IL-10 (R = 0.623 and 0.648, *p* \\< 0.01, respectively) serum lactic acid (R = 0.544, *p* \\< 0.01).\n\n4. Discussion {#sec4-animals-09-00599}\n=============\n\nThe microbiota influences the brain-gut axis and immunity. The present data was the first to confirm the immunity and microbiota status as affected by transport stress and breeds factors.\n\n4.1. Effect of Hormone Balance in Serum {#sec4dot1-animals-09-00599}\n---------------------------------------\n\nThe previous work showed elevated levels of COR and ACTH among animals in stressful conditions \\[[@B25-animals-09-00599]\\]. Transport stress leads to central nervous system (CNS) excitement and stimulation of the sympathetic nervous system. As a result, the secretion of hormones got out of control and increased ACTH and COR are released \\[[@B26-animals-09-00599]\\]. Up-regulated COR concentration stimulates carbohydrate, fat, protein metabolism, and provides energy for the body under stressful condition \\[[@B27-animals-09-00599]\\]. Under the transportation, the concentration of COR and ACTH increased, and in CY group had the lowest level of ACTH, while in SC group had the highest concentration of COR after transportation. T3, T4 can regulate growth in central nervous system, while cattle under heat-stress the concentration was reduced, it was the same with present research, and in SC group had a higher level compare with NY and CY groups before transportation, it means different breeds have effect on the concentration of the hormone, and 6 h transport leads to cause stress to all cattle breeds\n\n4.2. Effects of Transport Stress on Rumen Fermentation Characteristics {#sec4dot2-animals-09-00599}\n----------------------------------------------------------------------\n\nThere were non-significant differences in volatile fatty acid (VFA) among NY, SC, and CY groups before transportation, which might be owing to all groups had the same diet, and environment condition or all breeds had the same ratio for VFA production to absorption in the rumen. The concentration of TVFA was increased after transportation because during the transportation cattle were very nervous, and under stress, the hormone level changed such as, COR, it leads to increased metabolic strengthening and the respiratory rate and sweat secretion, while deprived of water, which might lead to cattle dehydration \\[[@B27-animals-09-00599]\\]. The concentration of propionic acid was increased and can leads to the increase the concentration of glucose because glucose can be produced by gluconeogenesis from propionic acid \\[[@B28-animals-09-00599]\\], and during the heat stress cattle have a higher respiratory rate and energy consumption \\[[@B29-animals-09-00599],[@B30-animals-09-00599]\\]. This research supported that during transportation stress, both propionic acid levels were increased. It has been reported that acetic acid can reduce the efficiency of rumen energy utilization \\[[@B31-animals-09-00599]\\]. Under the transportation stress the ratio of acetic: propionic acids were increased, it means rumen fermentation mode changed and efficiency reduced, and SC groups had the lowest acetic: propionic acids and the highest TVFA level, it means during the transportation stress SC's energy efficiency was higher when compared with NY and CY groups. All these means, during transportation stress, cattle required more energy, but the efficiency reduced. While butyric acid can effects bacteria and it also have a protective effect on gastrointestinal mucosa, promoting the function of the immune system \\[[@B32-animals-09-00599]\\]. Our data indicated that with the increase of butyric acid concentrations in the rumen, the anti-inflammatory cytokines and pro-inflammatory cytokines were synchronously increased and that *Butyrivibrio 2* was negatively correlated with IL-1\u03b2, IL-6. Both lactic acid and volatile fatty acids affect rumen pH, as an essential indicator of animal health: low or high pH will lead to rumen metabolic disorder, based on the number of gram-negative bacterium deaths under a low-pH environment \\[[@B27-animals-09-00599]\\]. The data support in that view, after transportation pH and the abundance of Bacteroidetes, were decreased, and pH was positively correlated with Bacteroidetes. LPS are a major component of the outer membrane of gram-negative bacteria: when gram-negative bacteria disintegrate, and dissolve lipopolysaccharides are released. LPS are among the main components of bacterial endotoxin \\[[@B33-animals-09-00599]\\], but it can also stimulate the host's innate immune system and enhance the resistance of the body. Specifically, the collateral pathway of the complement can be directly activated under low concentrations of LPS. On the other hand, high concentrations of LPS cause a systemic inflammatory response and activation of mononuclear macrophages. The endothelial fine cells will release inflammatory mediators such as TNF-\u03b1, IL-1, IL-6, IL-8, oxygen-free radicals, and histamine, but also directly or indirectly induce apoptosis of the immune cells, inhibiting the body's immune function \\[[@B34-animals-09-00599]\\]. In the present study, after transportation the concentration of LPS were increased in both blood and rumen, means during the transportation stress cattle may be suffering an inflammatory reaction, while there were reduced concentrations of some gram-negative bacteria such as the dominant members of Bacteroidetes and Proteobacteria. There were effects of transportation, breeds, and transportation **\u00d7** breeds for the increased concentration of LPS in serum and rumen LPS.\n\n4.3. Effect of Transport Stress on Rumen Microorganisms {#sec4dot3-animals-09-00599}\n-------------------------------------------------------\n\nThe previous study reported that in cattle, the abundance of rumen Bacteroidetes reduced in stress condition \\[[@B35-animals-09-00599],[@B36-animals-09-00599]\\]. Pre-weaning stress does not change the microbial community constantly in the rumen and feces by the analyzed species richness (chao1 index) \\[[@B35-animals-09-00599]\\]. In this study, richness estimates and diversity indices indicated that transport stress affects rumen microbial diversity, and after transport, the index about chao1, Shannon, and OTUs in SC group was higher than any other groups. The factor of breed significant effect the Shannon and OTUs index supported by Paz, H. A. \\[[@B37-animals-09-00599]\\].\n\nOur data showing that Bacteroidetes, Firmicutes, Proteobacteria, and Lentisphaerae were the most common microbial flora in phyla levels among different breeds of beef cattle (NY, SC, and CY groups) before and after transportation. These microbes are essential for rumen fermentation \\[[@B38-animals-09-00599],[@B39-animals-09-00599],[@B40-animals-09-00599],[@B41-animals-09-00599]\\]. This research showed that the relative abundance of Bacteroidetes were higher than Firmicutes in the rumen, which was consisted of previous research of Jami, E \\[[@B42-animals-09-00599]\\]. Before transportation there were higher abundant of Bacteroidetes in NY group than other groups, and after transportation SC's abundance of Firmicutes were the highest, meaning, the major microbial flora were the same for different breeds, but each microbial flora's proportion was different in rumen, so for different breeds have a different fermentability and lead to different digestive capacity.\n\nIt has been reported that Bacteroidales can degrade cellulose and their genomes can encode the decomposition of plant polysaccharide \\[[@B43-animals-09-00599],[@B44-animals-09-00599]\\]. It means that the ability to degrade cellulose might be influenced by transport stress, and it should be carefully fed more coarse fodder while cattle under the transport stress. *Prevotella* was the most abundant genus belongs to Bacteroidetes phyla. The previous research suggested that dietary balance between carbohydrates and protein were the primary factor based on shifting microbial community, and the abundance of *Bacteroides* are associated with high protein and fat diets \\[[@B45-animals-09-00599],[@B46-animals-09-00599]\\], and our study shows that transport stress leads to reduced *Prevotella,* and *Prevotella* can influence the activity of dipeptidyl peptidase type IV rate-limiting and promote oligopeptides degradation, so it can promote protein degradation \\[[@B47-animals-09-00599]\\]. In this research, *Prevotella 1* and *Prevotella UCG-003* were negatively correlated with the acetic to propionic acid ratio, and previous research showed that *Prevotella* which predominates associated with greater carbohydrate intake \\[[@B48-animals-09-00599]\\], a high abundance of *Prevotella* can increase propionate concentration and reduced the acetic to a propionic acid ratio in rumen fermentation \\[[@B49-animals-09-00599]\\]. It has been reported that the relative abundance of *Prevotella 1* was negatively correlated with serum BHBA but positively correlated with rumen pH \\[[@B35-animals-09-00599]\\]. This is consistent with previous findings that the abundance of *Prevotella 1* was reduced as the acetic to propionate ratio was increased \\[[@B35-animals-09-00599]\\]. *Ruminococcus* is a predominant microbial flora at the genus level, and consist of *Ruminobacter albums* and *Ruminococcus flavefaciens,* and within the phylum Firmicutes play an important rule for fiber degradation by secreting cellulase and hemicellulase to decompose plant fiber \\[[@B50-animals-09-00599]\\]. In this study, the abundant of Firmicutes were increased after transportation and SC group had a maximum variation, increased *Ruminococcus 1* was positively correlated with the concentration of acetic acid it means *Ruminococcus 1* effects the digestion and rumen fermentation.\n\n4.4. Effect of Transportation Stress on Immunity {#sec4dot4-animals-09-00599}\n------------------------------------------------\n\nAt present, it is clear that microbiota greatly contributes to the metabolic and immune homeostasis and influences health and disease \\[[@B51-animals-09-00599]\\]. Transport stress can induce immunosuppressive efficacy and increase the risk of disease or infection. It has been reported that over-production of IL-6 will induce inflammation, chronic metabolic disease, and severely impact on organism health \\[[@B52-animals-09-00599],[@B53-animals-09-00599]\\]. The neuro-immune- endocrine interface can secrete IL-6 and induce inflammation that contributes or caused by other diseases, and it can be activated by the hypothalamo-pituitary-adrenocortical (HPA) axis \\[[@B54-animals-09-00599],[@B55-animals-09-00599]\\], and it has been reported that under an acute stress environment the concentration of IL-6 and ACTH in plasma were accompanied by an increase in the phosphorylation of STAT3 in the anterior pituitary \\[[@B56-animals-09-00599]\\]. The exercise can up-regulate the concentration of IL-6, and that the secretion of COR was regulated by ACTH \\[[@B57-animals-09-00599]\\]. In the current study, after transportation the concentrations of COR, ACTH and inflammatory cytokines IL-6, IL-1\u03b2 and TNF-\u03b1 had similar variation trends, both higher than before transportation (for IL-6 increased 21.58%, 16.57% and 44.28%, for IL-1\u03b2 increased 34.06%,65.19% and 46.74% compared with before transport in NY, SC and CY groups, respectively). This also suggests that IL-6 might directly stimulate the secretion of ACTH when under acute stress condition.\n\nTransport stress not only affects hormone secretion but also directly or indirectly influences the body's immune performance. The concentration of white blood cells and neutrophil significant decreased after transportation \\[[@B3-animals-09-00599]\\]. In germ-free (GF) mice the concentration of IgA was lower, but fecal flora containing commensal species like Alcaligenes can indirectly send signals to lymphocytes to induce IgA production, and improve the level of IgA, this means microbiota can significantly effect on the immune level of the body \\[[@B58-animals-09-00599],[@B59-animals-09-00599]\\]. It has been reported that microbial-fermented concentrate can significantly improve the levels of IgG, IgA, and IFN-\u03b3 in the cattle serum during the heat stress and IL-6 and CCL2 correlated with stressor-induced changes in *Coprococcus, Pseudobutyrivibrio,* and *Dorea* \\[[@B60-animals-09-00599]\\]. The present study also indicates that immunoglobulin (IgA and IgG) were decreased after transportation (for IgG decreased 27.18%, 7.08%, and 37% compared with before transportation in NY, SC, and CY groups, respectively). IgA positive correlated with *Butyrivibrio 2* and *Prevotella 1* and negatively correlated with *f.Bacteroidales* BS11 gut group and *Ruminococcaceae* UCG-002, these correlations suggest that transportation stress induces microbial composition and physiological characteristics and immune activities in different cattle breeds.\n\nMeanwhile, the factors of breeds significant effect on the level of IgA, although there were no obvious differences among all cattle breeds, SC group had a higher concentration of IgA before or after transportation. There are many microbiotas associated with immune response-related diseases and health conditions. Clostridia and SCFA can induce directly T (reg) anti-inflammatory bowel disease, while the abundance of Bacteroidetes and Lachnospiraceae were decreased or the abundance of *Actinobacteria* and *Proteobacteria* were increased the risk of Crohn's disease or ulcerative colitis. Our data suggested that butyric acid was positively correlated with *Butyrivibrio 2* and *f.Bacteroidales* BS11 gut group, TVFA was positively correlated with *Lactobacillus* \\[[@B61-animals-09-00599],[@B62-animals-09-00599]\\].\n\nAllergy was greater associated with a decrease of abundance *Bifidobacterium adolescentis* and *Lactobacillus* \\[[@B63-animals-09-00599]\\]. It was evidenced that gram-negative Bacteroides and gram-positive Firmicutes are the vast majority of commensals \\[[@B54-animals-09-00599]\\], they can stimulate the host immune system and maintain the metastable number to help enhance immunity, and also influence the metabolism and providing nutrients to the host \\[[@B54-animals-09-00599],[@B64-animals-09-00599]\\]. The risk of metabolic syndrome type 2 diabetes and cardiovascular disease due to obesity were increased when the balance of Firmicutes and Bacteroidetes changed \\[[@B65-animals-09-00599]\\]. Clostridium induced T cells to express more IL-10, which is an essential cytokine for anti-inflammatory reaction, and we found *Ruminococcaceae* NK4A214 group, *Butyrivibrio* 2, *Saccharo -fermentans, Ruminococcaceae* UCG-005, and *Ruminococcaceae* UCG-002 were belonged to Clostridiales order, while IL-10 was positive correlated with *Saccharofermentans* and *Ruminococcaceae* UCG-002 \\[[@B66-animals-09-00599]\\].\n\nIn our research, both pro-inflammatory and anti-inflammatory cytokines were increased after transportation. One possible cause is that transport stress leads to pro-inflammatory cytokines TNF-\u03b1, IL-1\u03b2, and IL-6 induce an inflammatory reaction in all groups and stimulate the immune system to increase more IL-4 and IL-10 to an anti-inflammatory response. In fact, this can permit the colonization of pathogens in the gastrointestinal tract and can limit their excessive multiplication \\[[@B67-animals-09-00599]\\]. There was a maximum variety for pro-inflammatory IL-1\u03b2 and IL-6 in CY groups, and the factor of breeds significant effect the concentration of TNF-\u03b1 and IL-6 it means different breeds have a different inflammatory reaction during transport stress. Meanwhile, transport stress increased the anti-inflammatory cytokines IL-10 and IL-4, in CY groups the level of anti-inflammatory cytokines were the lowest in all groups after transportation. It means CY had a more serious inflammatory reaction in three groups. *Lactobacillus* can induce anti-inflammatory \\[[@B68-animals-09-00599]\\], and the abundance of *Lactobacillus* was positively correlated with IL-4 may be to induce an anti-inflammatory response. The abundance of *Lactobacillus* had the same variation tendency, Bacteroides and Clostridium can ferment fiber carbohydrate and produce short-chain fatty acids as acetic acid, propionic acid, and butyric acid \\[[@B69-animals-09-00599]\\]. This was mostly consistent with the changed in rumen volatile fatty acids, in this study the concentration of butyric acid was increased it can restrain of pro-inflammatory cytokine expression and inhibition of the NFkB pathway \\[[@B70-animals-09-00599]\\], and produce mucin and antimicrobial peptides, while up-regulating the expression of tight junction proteins in the epithelial barrier \\[[@B71-animals-09-00599]\\].\n\n5. Conclusions {#sec5-animals-09-00599}\n==============\n\nIn the present study, we concluded that transportation stress caused to change the levels of the hormone in all cattle, the concentrations of pro-inflammatory cytokines TNF-\u03b1, IL-1\u03b2, and IL-6 were increased in all cattle breeds meantime, in order to adapt the environment changed anti-inflammatory cytokines IL-10 and IL-4 increased to regulate endocrine balance. Transport stress affects the microbial flora, and the relative abundance of Bacteroidetes were decreased, while Firmicutes were increased at the phyla level, and the metabolites of microorganisms were also changed. In addition the results indicate that under transport stress cattle may suffer from inflammatory response through modulate HPA axis, and also affects the secretion of hormone levels and cytokines, in the meantime, microbiota, through regulating metabolite and indirectly affecting the immune functions, it is also concluded from the present study that breeds factor affects the performance of stress resistance.\n\nThis work was funded by the projects of National key research and development program: research and development of key techniques of fine feeding economize cost and increasing efficiency of cattle and sheep (2017YFD0502005); Supported by Technical System of National Beef cattle/Yak Industry(CARS-37); Supported by Sichuan Science and Technology Program (2018NZ0002). We would like to thank all staff members of Ruminant Nutrition Laboratory, Sichuan Agricultural University, the authors gratefully acknowledge the technical support and guidance of Xianwen Dong, Yaqun Shao, Zhaoxi Pei and Shaoyu Zeng of the senior fellow apprentice.\n\nThe following are available online at <https://www.mdpi.com/2076-2615/9/9/599/s1>.\n\n###### \n\nClick here for additional data file.\n\nData curation, Z.W., A.M.S., Q.P., R.H., C.T., X.Z., Y.L. and B.X.; formal analysis, A.M.S., Q.P., C.T., X.W. and B.X.; funding acquisition, Z.W.; Investigation, F.L., Z.W. and X.W.; methodology, F.L., A.M.S., R.H. and H.Z.; project administration, Z.W.; software, A.M.S., R.H., H.Z., X.Z. and L.W.; supervision, Z.W.; validation, Q.P., Y.W. and L.Z.; writing---original draft, F.L.; writing---review & editing, A.M.S.\n\nThe authors declare no conflict of interest.\n\n###### \n\nPrincipal Coordinate Analysis (PCoA) of rumen bacterial community structures of cattle in NY, SC and CY groups before (**A**), rumen bacterial community before and after transportation in NY (**B**), rumen bacterial community before and after transportation in SC (**C**), rumen bacterial community before and after transportation in CY (**D**). The PCoA plots were constructed using the weighted UniFrac method.\n\n![](animals-09-00599-g001a)\n\n![](animals-09-00599-g001b)\n\n![Venn diagram representation of the shared and exclusive OTUs at the 97% similarity level among the three parts of the rumen bacterial community before (**A**) and after (**B**) transportation in NY, SC, and CY groups: between BNY and ANY (**C**), BCY and ACY (**D**) and BSC and ASC (**E**).](animals-09-00599-g002){#animals-09-00599-f002}\n\n![Bar plots showing the average relative abundance of bacterial phyla (%) in the rumen at the phylum level (**A**) and genus level (**B**). For each breed, B indicates before transport, and A indicates after transport. (For example, the Native Yellow (NY) group includes BNY and ANY, Simmental Crossbred Cattle (SC) group includes BSC and ASC, and Cattle Yak (CY) includes BCY and ACY) and transportation (before vs. after). Data represent the abundance at greater than 0.1% of the community between three beef breeds or between treatment groups (before and after transportation).](animals-09-00599-g003){#animals-09-00599-f003}\n\n![LEfSe identified the most differentially abundant taxons in NY, SC, and CY (**A**) before transport, BNY and ANY (**B**), BSC and ASC (**C**), BCY and ACY (**D**). Taxonomic cladogram obtained from LEfSe analysis of 16 S sequences (relative abundance \u22650.5%). (red) BNY-enriched taxa, (blue) taxa enriched in BSC, (green) taxa enriched in BCY (**A**), (red) BNY-enriched taxa, (green) taxa enriched in ANY (**B**), (red) BCY-enriched taxa, (green) taxa enriched in ACY (**C**), (red) BSC-enriched taxa, (green) taxa enriched in ASC (**D**). The brightness of each dot is proportional to its effect size. Phylum: p (put \"p\" in front of the microbes or don't use italics), class: c(put \"c\" in front of the microbes and use italics), order: o(put \"o\" in front of the microbes and use italics), family: f (put \"f\" in front of the microbes and use italics), genus: g(put \"g\" in front of the microbes and use italics or not put \"g\" in front of the microbes but use italics) and species: s (put \"s\" in front of the microbes and use italics), all data were same.](animals-09-00599-g004){#animals-09-00599-f004}\n\n![Correlation between rumen microbiota and cattle physiological variables. Spearman non-parametric Rank correlation matrix between serum index, characteristics of rumen fermentation parameters and microbiota abundance (representing at least 1% of the bacterial community in at least one sample). The red color represents a negative correlation; the blue color represents a positive correlation, and the white color represents no correlation. The OTU count data were subjected to variance stabilizing transformation, then pairwise serum index, characteristics of rumen fermentation parameters. Spearman correlations between bacterial and biological parameters at corresponding were analyzed.](animals-09-00599-g005){#animals-09-00599-f005}\n\nanimals-09-00599-t001_Table 1\n\n###### \n\nHormones levels before and after transportation.\n\n  Item      Treatment   NY                    SC                    CY\n  --------- ----------- --------------------- --------------------- ---------------------\n  COR       B           130.77 \u00b1 34.45 ^B^    133.06 \u00b1 23.78 ^B^    125.57 \u00b1 12.37 ^B^\n  (ng/mL)   A           158.37 \u00b1 22.26 ^A^    160.54 \u00b1 25.17 ^A^    156.89 \u00b1 18.71 ^A^\n  ACTH      B           338.86 \u00b1 70.49 ^B^    340.58 \u00b1 58.31 ^B^    295.15 \u00b1 48.59 ^B^\n  (pg/mL)   A           422.71 \u00b1 61.02 ^Aa^   437.84 \u00b1 70.92 ^Aa^   357.32 \u00b1 66.26 ^Ab^\n  T 3       B           6.89 \u00b1 0.778 ^a^      6.91 \u00b1 0.885 ^Aa^     5.37 \u00b1 0.776 ^b^\n  (ng/mL)   A           7.18 \u00b1 0.386 ^a^      5.14 \u00b1 0.684 ^Bb^     5.51 \u00b1 0.618 ^b^\n  T 4       B           243.89 \u00b1 35.286 ^a^   231.14 \u00b1 35.594 ^a^   184.32 \u00b1 70.257 ^b^\n  (ng/mL)   A           253.56 \u00b1 37.249 ^a^   206.31 \u00b1 43.599 ^b^   186.63 \u00b1 47.442 ^b^\n\nNote: Data are shown as means \u00b1 SD (n = 6), the serum index includes COR (cortisol), ACTH (adrenocorticotropic hormone), T3 (Triiodothyronine), T4 (thyroxine). For treatment column Before transport: B. After transport: A. In the same row, values with the same or no small letter superscripts (a, b and c) mean no significant difference (*p* \\> 0.05), while with different small letter superscripts (a, b and c) mean significant difference (*p* \\< 0.05), in the same column, values of the same index with no capital letter superscripts (A and B) mean no significant difference (*p* \\> 0.05), while with different capital superscripts (A and B) mean significant difference (*p* \\< 0.05). SC (Simmental Crossbred Cattle: Simmental \u00d7 Xuanhan), NY (Native Yellow Cattle: Xuanhan Yellow Cattle), and CY (Cattle Yak: Jersey \u00d7 Maiwa Yak).\n\nanimals-09-00599-t002_Table 2\n\n###### \n\nFluid characteristics before and after transportation.\n\n  Item                 Treatment         NY                   SC                   CY\n  -------------------- ----------------- -------------------- -------------------- --------------------\n  Serum LPS            B                 13.63 \u00b1 3.144        14.61 \u00b1 2.022        13.25 \u00b1 2.199 ^B^\n  (ng/mL)              A                 15.77 \u00b1 2.08         17.56 \u00b1 2.186        17.54 \u00b1 1.69 ^A^\n  Rumen LPS            B                 13.03 \u00b1 2.589        10.93 \u00b1 2.557 ^B^    10.65 \u00b1 2.55 ^B^\n  (ng/mL)              A                 15.26 \u00b1 2.323 ^a^    14.86 \u00b1 2.383 ^Ab^   14.22 \u00b1 1.923 ^Ac^\n  Serum lactic acid    B                 0.91 \u00b1 0.349 ^B^     0.99 \u00b1 0.464 ^B^     1.27 \u00b1 0.479\n  (mmol/L)             A                 1.74 \u00b1 0.192 ^A^     1.85 \u00b1 0.363 ^A^     1.7 \u00b1 0.31\n  Rumen lactic acid    B                 0.26 \u00b1 0.049 ^c^     0.78 \u00b1 0.641 ^Ab^    1.97 \u00b1 0.373 ^Aa^\n  (mmol/L)             A                 0.25 \u00b1 0.066 ^b^     0.43 \u00b1 0.099 ^Ba^    0.54 \u00b1 0.168 ^Ba^\n  Acetic acid          B                 21.52 \u00b1 0.986        23.51 \u00b1 7.56 ^B^     21.69 \u00b1 4.495 ^B^\n  (mmol/L)             A                 33.36 \u00b1 5.628 ^c^    45.22 \u00b1 7.21 ^Aab^   36.01 \u00b1 2.486 ^Ab^\n  Propionic acid       B                 6.03 \u00b1 1.35          6.17 \u00b1 2.2 ^B^       5.17 \u00b1 0.883 ^B^\n  (mmol/L)             A                 7.2 \u00b1 1.608 ^b^      11.43 \u00b1 2.93 ^Aa^    7 \u00b1 1.06 ^Ac^\n  Butyric acid         B                 2.3 \u00b1 0.411 ^B^      3.2 \u00b1 1.73 ^B^       2.46 \u00b1 0.181 ^B^\n  (mmol/L)             A                 5.06 \u00b1 1.557 ^Ac^    7.7 \u00b1 1.406 ^Aa^     6.85 \u00b1 1.969 ^Ab^\n  Acetic : propionic   B                 3.72 \u00b1 0.691 ^B^     3.84 \u00b1 0.093         4.16 \u00b1 0.218 ^B^\n  acids                A                 4.59 \u00b1 0.311 ^Aab^   4.05 \u00b1 0.372 ^b^     5.13 \u00b1 0.267 ^Aa^\n  TVFA                 B                 29.85 \u00b1 1.49 ^B^     32.87 \u00b1 3.32 ^B^     29.31 \u00b1 5.55 ^B^\n  (mmol/L)             A                 45.62 \u00b1 4.34 ^Ab^    64.345 \u00b1 5.52 ^Aa^   47.93 \u00b1 3.48 ^Ab^\n  PH                   B                 7.27 \u00b1 0.09          7.3 \u00b1 0.06 ^A^       7.22 \u00b1 0.11 ^A^\n  A                    6.98 \u00b1 0.03 ^a^   6.75 \u00b1 0.09 ^Bb^     6.79 \u00b1 0.12 ^Bb^     \n\nNote: Data are shown as means \u00b1 SD (n = 6). Before transport: B. After transport: A. In the same row, values with the same or no small letter superscripts (a, b and c) mean no significant difference (*p* \\> 0.05), while with different small letter superscripts (a, b and c) mean significant difference (*p* \\< 0.05), in the same column, values of the same index with no capital letter superscripts (A and B) mean no significant difference (*p* \\> 0.05), while with different capital superscripts (A and B) mean significant difference (*p* \\< 0.05). SC (Simmental Crossbred Cattle: Simmental \u00d7 Xuanhan), NY (Native Yellow Cattle: Xuanhan Yellow Cattle), and CY (Cattle Yak: Jersey \u00d7 Maiwa Yak).\n\nanimals-09-00599-t003_Table 3\n\n###### \n\nImmunity levels before and after transportation.\n\n  Item      Treatment   NY                      SC                      CY\n  --------- ----------- ----------------------- ----------------------- -----------------------\n  IgG       B           1.03 \u00b1 0.05 ^A^         1.13 \u00b1 0.01 ^A^         1.32 \u00b1 0.06 ^A^\n  (mg/mL)   A           0.75 \u00b1 0.08 ^B^         1.05 \u00b1 0.02 ^B^         0.95 \u00b1 0.02 ^B^\n  IgA       B           14.98 \u00b1 5.31            19.22 \u00b1 2.19            15.21 \u00b1 3.39\n  (ug/mL)   A           13.82 \u00b1 5.04 ^ab^       18.03 \u00b1 3.16 ^a^        13.64 \u00b1 2.34 ^b^\n  TNF-\u03b1     B           12.65 \u00b1 1.86 ^Ba^       9.77 \u00b1 1.48 ^Ab^        11.58 \u00b1 1.03 ^Ba^\n  (pg/mL)   A           15.61 \u00b1 1.24 ^A^        16.02 \u00b1 1.31 ^B^        15.38 \u00b1 2.81 ^A^\n  IL-1\u03b2     B           2000.61 \u00b1 455.54 ^Bb^   1716.15 \u00b1 239.57 ^Bb^   2225.56 \u00b1 391.49 ^Ba^\n  (pg/mL)   A           2682.02 \u00b1 511.69 ^Ab^   2834.96 \u00b1 310.28 ^Ab^   3265.82 \u00b1 335.59 ^Aa^\n  IL-6      B           695.02 \u00b1 135.81 ^Ba^    688.45 \u00b1 57.96 ^Ba^     594.1 \u00b1 45.63 ^Bb^\n  (pg/mL)   A           845.49 \u00b1 68.06 ^A^      802.35 \u00b1 82.38 ^A^      857.58 \u00b1 45.92 ^A^\n  IL-10     B           783.66 \u00b1 97.12 ^a^      621.33 \u00b1 70.11 ^b^      621.03 \u00b1 127.17 ^b^\n  (pg/mL)   A           806.59 \u00b1 78.53 ^a^      689.08 \u00b1 111.145 ^ab^   648.91 \u00b1 171.76 ^b^\n  IL-4      B           498.49 \u00b1 113.535 ^a^    444.01 \u00b1 50.79 ^ab^     361.82 \u00b1 64.257 ^Bb^\n  (pg/mL)   A           504.49 \u00b1 204.821        501.25 \u00b1 47.275         477.85 \u00b1 50.824 ^A^\n\nNote: Data are shown as means \u00b1 SD (n = 6). Before transport: B. After transport: A. In the same row, values with the same or no small letter superscripts (a, b and c) mean no significant difference (*p* \\> 0.05), while with different small letter superscripts (a, b and c) mean significant difference (*p* \\< 0.05), in the same column, values of the same index with no capital letter superscripts (A and B) mean no significant difference (*p* \\> 0.05), while with different capital superscripts (A and B) mean significant difference (*p* \\< 0.05). SC (Simmental Crossbred Cattle: Simmental \u00d7 Xuanhan), NY (Native Yellow Cattle: Xuanhan Yellow Cattle), and CY (Cattle Yak: Jersey \u00d7 Maiwa Yak).\n\nanimals-09-00599-t004_Table 4\n\n###### \n\nAlpha diversity indices.\n\n  Item      Treatment              NY                      SC                      CY\n  --------- ---------------------- ----------------------- ----------------------- ---------------------\n  OTUs      B                      2610.75 \u00b1 100.24 ^a^    2255.8 \u00b1 372.33 ^Bb^    2093.6 \u00b1 146.02 ^b^\n  A         2410.25 \u00b1 387.46 ^b^   2755.6 \u00b1 177 ^Aa^       2184 \u00b1 113.74 ^b^       \n  Chao1     B                      4219.42 \u00b1 340.24        3678.18 \u00b1 1014.29 ^B^   3343.85 \u00b1 375.36\n  A         3947.86 \u00b1 849.05 ^a^   4315.88 \u00b1 606.76 ^Aa^   3688.9 \u00b1 426.59 ^b^     \n  Shannon   B                      6.5 \u00b1 0.14              6.38 \u00b1 0.19 ^B^         6.15 \u00b1 0.18\n  A         6.41 \u00b1 0.3 ^b^         6.67 \u00b1 0.07 ^Aa^        6.2 \u00b1 0.08 ^b^          \n\nNote: Data are shown as means \u00b1 SD (n = 5). Before transport: B. After transport: A. In the same row, values with the same or no small letter superscripts (a, b and c) mean no significant difference (*p* \\> 0.05), while with different small letter superscripts (a, b and c) mean significant difference (*p* \\< 0.05), in the same column, values of the same index with no capital letter superscripts (A and B) mean no significant difference (*p* \\> 0.05), while with different capital superscripts (A and B) mean significant difference (*p* \\< 0.05). SC (Simmental Crossbred Cattle: Simmental \u00d7 Xuanhan), NY (Native Yellow Cattle: Xuanhan Yellow Cattle), and CY (Cattle Yak: Jersey \u00d7 Maiwa Yak).\n\nanimals-09-00599-t005_Table 5\n\n###### \n\nTwo-way analysis of transportation stress and beef breed.\n\n  Item                Transport   Breeds    Transport \u00d7 Breeds                      \n  ------------------- ----------- --------- -------------------- --------- -------- ---------\n  COR                 13.856      \\<0.001   0.262                0.77      0.192    0.826\n  ACTH                10.057      0.0812    0.817                0.445     14.668   \\<0.001\n  Serum LPS           21.157      0.029     24.027               \\<0.001   0.593    \\<0.001\n  Rumen LPS           9.094       0.006     0.654                0.529     0.149    0.862\n  Rumen lactic acid   24.048      \\<0.001   1.634                0.216     9.769    0.001\n  Serum lactic acid   21.823      \\<0.001   20.274               \\<0.001   0.82     0.452\n  pH                  33.873      \\<0.001   22.273               \\<0.001   4.392    \\<0.001\n  Acetic acid         42.38       \\<0.001   14.98                \\<0.001   1.684    0.207\n  Propionic acid      16.37       \\<0.001   3.349                0.052     3.73     0.039\n  Butyric acid        47.527      \\<0.001   16.458               \\<0.001   0.89     0.424\n  Acetic: propionic   22.877      \\<0.001   3.658                0.041     2.733    0.085\n  TVFA                39.951      \\<0.001   8.164                0.002     1.934    0.166\n  IgA                 0.365       0.547     10.874               \\<0.001   0.138    0.871\n  IgG                 26.078      \\<0.001   0.109                0.897     0.308    0.738\n  TNF-\u03b1               10.252      0.0617    16.084               \\<0.001   10.169   \\<0.001\n  IL-1\u03b2               15.095      \\<0.001   7.98                 0.044     14.218   \\<0.001\n  IL-6                1.78        0.186     15.228               \\<0.001   1.025    0.363\n  IL-10               1.878       0.174     1.917                0.153     10.522   \\<0.001\n  IL-4                13.701      \\<0.001   1.112                0.334     2.305    0.106\n  OTUs                2.351       0.138     4.388                0.024     2.765    \\<0.001\n  Chao1               1.529       0.228     1.297                0.292     0.414    0.665\n  Shannon             1.621       0.215     7.464                0.003     1.905    0.171\n\n**Note:** Two-way analysis including transport, breeds and interaction factors, *F* (*F*-value), *p* (*p*-value), when (*p* \\< 0.05) means that factor effect the index significant and (*p* \\> 0.05) means that factor effect the index not significant.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nThe prevalence of diabetes mellitus is increasing, with the estimated number of affected adults growing to 693 million by 2045 from 451 million in 2017 globally^[@CR1]^. Diabetes mellitus patients have an increased risk of cardiovascular diseases. Cardiovascular complications remain a key cause of diabetes-related morbidity and mortality^[@CR2]^. On the heart, the unfavorable influence of diabetes inevitably causes myocardial dysfunction which, initially clinically silent, might then lead to overt diabetic cardiomyopathy if early treatment is insufficient^[@CR3]^. Diabetic cardiomyopathy is characterized by the occurrence of ventricular dysfunction independent of coronary artery disease and hypertension^[@CR4]^. The proposed metabolic impairments contributing to diabetic cardiomyopathy include deposition of advanced glycation end products, atherosclerosis, subclinical microinfarctions, mitochondrial dysfunction, and lipotoxicity^[@CR5]^. These impairments not only lead to left ventricle (LV) impairment but also might inevitably hamper right ventricle (RV) function because of the systematic nature of these impairments^[@CR6],[@CR7]^.\n\nRV function is known to have diagnostic and prognostic values in multiple cardiovascular diseases and pulmonary disorders^[@CR8],[@CR9]^. The association between RV dysfunction with aggravation of myocardial function and prognosis in distinct cardiac diseases has been confirmed^[@CR10]^. The favorable effect of healthy diet and physical activity on RV mechanics indicates that RV myocardial abnormalities are probably modifiable through adequate interventional strategies^[@CR11]^. Most previous studies on myocardial dysfunction in diabetes have paid more attention to the LV^[@CR12],[@CR13]^. The role of the RV in diabetic cardiomyopathy is still under-investigated. Cardiovascular magnetic resonance (CMR) has been regarded as the golden standard for the accurate quantification of RV function and chamber size compared with the echocardiographic assessment because of the complex anatomical and contraction patterns of the RV^[@CR8],[@CR14],[@CR15]^. Further, the accuracy of echocardiographic results depends substantially on operator skill and is limited by narrow acoustic windows. There were several echocardiographic studies focused on the RV strain in T2DM patients^[@CR3],[@CR14],[@CR15]^. In recent, cardiac magnetic resonance myocardial feature tracking (CMR-FT) has been treated as a novel tool with high accuracy for quantitatively assessing left and right ventricular myocardial deformation^[@CR16]--[@CR18]^. However, to the best of our knowledge, the information of CMR-FT for assessing the RV deformation is limited. Therefore, this study aimed to evaluate the feasibility of using CMR-FT for quantifying global and regional RV myocardial deformation and to test whether CMR-FT can detect subclinical RV dysfunction in T2DM patients.\n\nMethods {#Sec2}\n=======\n\nStudy population {#Sec3}\n----------------\n\nThis study was approved by the Institutional Ethics Review Board of West China Hospital (No. 2016-24) and we pledged to abide by the declaration of Helsinki (2000 EDITION) in accordance with the relevant medical research rules in the study. All participants provided written informed consent prior to study commencement. For this prospective study, we recruited 128 consecutive T2DM patients visiting the outpatient Department of Endocrinology and Metabolism at our institution and undergoing routine CMR examination between June 2016 and July 2018. The inclusion criteria included an established T2DM diagnosis based on the American Diabetes Association criteria^[@CR19]^; no symptoms, signs, or history of heart disease (known coronary artery disease, cardiomyopathy, or valvular heart disease); sinus rhythm; and no contraindications to MR imaging. Subjects were excluded if they had severe renal impairment (estimated glomerular filtration rate \\<30\u2009mL/min/1.73 mm^2^; n\u2009=\u20096), uncontrolled blood pressure at rest (systolic blood pressure\u2009\\>\u2009180\u2009mmHg and/or diastolic blood pressure\u2009\\>\u2009100\u2009mmHg; n\u2009=\u20096), or poor CMR image quality (n\u2009=\u200912). Known coronary heart disease included self-reported prior myocardial infarction, percutaneous coronary artery intervention, and/or coronary artery bypass grafting. Consequently, 104 T2DM patients (mean age, 52 years; range 25--77 years; male mean age, 55 years; range 29--77 years; female mean age, 54 years; range 25--74 years; P\u2009=\u20090.63, independent-samples t-test) remained. Patients were divided into two groups according to the recent task force criteria^[@CR20]^: reduced right ventricular ejection fraction (RVEF) group (RVEF\u2009\\<\u200945%) and preserved RVEF group (RVEF\u2009\u2265\u200945%). Concurrently, 26 individuals from our healthy volunteer database, with similar sex and age distribution to those of our patients, were recruited to constitute the control group.\n\nCMR imaging {#Sec4}\n-----------\n\nAll participants underwent CMR imaging in the supine position using a 3.0-T whole-body scanner (Skyra; Siemens Medical Solutions) fitted with an 18-element body phased array coil for signal detection. We used the manufacturer's standard ECG-triggering device and the end-inspiratory breath-holding technique to monitor ECG values and breathing, respectively. Data were acquired during the breath-holding period. In the short-axis view, 8--12 continuous CMR cine images were obtained from the level of the mitral valve to the apex using steady-state free-precession sequences (field of view \\[FOV\\] 340\u2009\u00d7\u2009284 mm^2^; repetition time (TR) 39.3\u2009ms (echo spacing 2.8\u2009ms); echo time (TE) 1.2\u2009ms; slice thickness 8.0\u2009mm; flip angle 40\u00b0; temporal resolution 25--40\u2009ms; matrix size 208\u2009\u00d7\u2009174; voxel size 1.6\u2009\u00d7\u20091.6\u2009\u00d7\u20098\u2009mm; bandwidth 1145\u2009Hz/Px) with generalized autocalibrating partial parallel acquisition (GRAPPA, acceleration factor: 3) reconstruction. The horizontal four-chamber view cine series were also acquired.\n\nCMR data analysis {#Sec5}\n-----------------\n\nAll CMR data were uploaded and analyzed offline using dedicated commercial software (cvi42; Circle Cardiovascular Imaging, Inc.). Cvi42 is based on an incompressible volume-based algorithm, which has been validated previously to perform accurate biventricular anatomical tracking^[@CR21]^. In the serial short-axis slices, the endocardial and epicardial borders were manually outlined at the end-diastolic and end-systolic phases by an experienced radiologist (with 3 years of experience in CMR). Global parameters of RV geometry and function, including RV end-diastolic volume, RV end-systolic volume, RV stroke volume (SV), and RVEF, were computed based on serial short-axis slices; similar parameters for the LV were also computed. Both long-axis four-chamber and short-axis slices were loaded into the feature-tracking module for analyzing RV myocardial strain. In all series, the endocardial and epicardial contours were delineated manually in each slice at the end-diastole phase (reference phase), and the papillary muscles and moderator bands were carefully excluded (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The accuracy of feature tracking for both RV endocardial and epicardial contours was visually checked following automated strain analysis on the CMR-FT model, and good quality tracking was gained in all subjects following a maximum of two observer adjustments. Subsequently, the following three-dimenasional (3D) feature-tracking parameters were automatically acquired: longitudinal, circumferential, and radial peak strain (PS), defined as the absolute value of maximum strain measured over an entire cardiac cycle; peak diastolic strain rate (PDSR), defined as the absolute value of the maximum strain rate over all phases starting from the end systole until the end diastole; and peak systolic strain rate (PSSR), defined as the absolute value of the maximum strain rate over all phases starting from the end diastole until the end systole.Figure 1CMR feature tracking using cmr42 (Circle Cardiovascular Imaging Inc., Calgary, Canada) in short-axis, and four-chamber long-axis cine images at the end-diastole (**A**,**B**) and end-systole (**C**,**D**). The yellow and cyan curves delineate the endocardial and epicardial contours, respectively. The yellow dots represent the right ventricle myocardial voxel points, and the yellow short line on the images shows the tracking of the ventricle myocardial voxel points. The orange line is used to define the base and apex of tricuspid valve plane, and the apical plane. Abbreviations: CMR, cardiovascular magnetic resonance; T2DM: type 2 diabetes mellitus.\n\nReproducibility {#Sec6}\n---------------\n\nIntra-observer variability in RV strain parameters was calculated by an experienced investigator by comparing the measurements from 20 randomly selected cases analyzed by the same observer after an interval of 1 month. Inter-observer variability was estimated by comparing the measurements with those calculated by another independent double-blinded experienced observer (with more than 3 years of experience in CMR).\n\nStatistical analysis {#Sec7}\n--------------------\n\nStatistical analyses were performed using SPSS statistics software for MAC (version 24.0; SPSS Institute, Inc.). Normality was assessed using Kolmogorov--Smirnov test. Normally distributed data are presented as mean\u2009\u00b1\u2009standard deviation for continuous variables and as frequencies (percentage) for categorical values. Non-parametric variables are expressed as median (interquartile range, 25--75%). One-way ANOVA or the Kruskal-Wallis test was used to assess the differences in continuous variables among the three groups. The chi-square or Fisher exact test was performed for categorical values. Spearman's and Pearson's correlation coefficients were used for nonparametric and normally distributed data, respectively. Binary logistic regression was used to identify predictors of RV dysfunction. Receiver operating characteristic analysis was performed to determine optimal cut-off values for RV strain parameters to identify RV dysfunction in T2DM patients. The Bland--Altman analysis was used to assess the inter- and intra-observer variabilities between acquisitions by calculating the bias (mean difference) and the 95% limits of agreement (1.96 standard deviations around the difference). Two-tailed *p* values of \\<0.05 were considered statistically significant.\n\nResults {#Sec8}\n=======\n\nBaseline characteristics and biventricular geometry and function {#Sec9}\n----------------------------------------------------------------\n\nAll 142 subjects (116 T2DM patients and 26 healthy subjects) underwent CMR imaging; data sets of 104 T2DM patients and 26 healthy subjects were included in the analysis. We divided the T2DM patients into a preserved RVEF group (RVEF\u2009\u2265\u200945%) and a reduced RVEF group (RVEF\u2009\\<\u200945%). Data on baseline characteristics, biventricular dimensions, and cardiac functional parameters of T2DM patients and healthy subjects are presented in Table\u00a0[1](#Tab1){ref-type=\"table\"}. Of the 104 T2DM patients, 14 (13.46%) patients presented with impaired systolic function (RVEF\u2009\\<\u200945%). Two and eight patients in the impaired RVEF group and preserved RVEF group had LV dysfunction (left ventricular ejection fraction (LVEF)\u2009\\<\u200955%), respectively. T2DM patients with or without reduced RVEF had significantly higher systolic blood pressure, diastolic blood pressure, glycosylated hemoglobin, fasting blood glucose, and high-density lipoprotein than healthy controls (*p*\u2009\\<\u20090.05). T2DM patients with reduced RVEF had significantly higher RV end-systolic volume and lower RV SV values than normal subjects or T2DM patients with preserved RVEF (*p*\u2009\\<\u20090.05).Table 1Baseline differences of clinical, biomedical, and CMR characteristics between control subjects, T2DM patients with preserved and impaired RVEF.VariableControl subjects (n\u2009=\u200926)DM with preserved RVEF (n\u2009=\u200990)DM with\\\nreduced RVEF (n\u2009=\u200914)Age (years)53\u2009\u00b1\u20091055\u2009\u00b1\u20091153\u2009\u00b1\u200911Male, n (%)16 (61.5%)55 (61.1%)11 (78.6%)Height (cm)162.0\u2009\u00b1\u20097.4163.9\u2009\u00b1\u20098.3166.8\u2009\u00b1\u20097.5Weight (kg)61.2\u2009\u00b1\u20098.864.6\u2009\u00b1\u20099.465\u2009\u00b1\u200911Body mass index (kg/m^2^)23.2\u2009\u00b1\u20092.324.0\u2009\u00b1\u20092.823.1\u2009\u00b1\u20093.1Systolic blood pressure (mmHg)118.1\u2009\u00b1\u20097.9131\u2009\u00b1\u200915\\*128\u2009\u00b1\u200912\\*Diastolic blood pressure (mmHg)78.7\u2009\u00b1\u20098.080\u2009\u00b1\u20091181.6\u2009\u00b1\u20098.4Heart rate (beats/min)73.0\u2009\u00b1\u20099.576\u2009\u00b1\u20091171\u2009\u00b1\u200912Duration of diabetes (years)NA8.0\u2009\u00b1\u20096.66.5\u2009\u00b1\u20095.9HbA1c (%)5.38\u2009\u00b1\u20090.397.6\u2009\u00b1\u20092.2\\*7.8\u2009\u00b1\u20092.1\\*Fasting blood glucose (mmol/L)5.14\u2009\u00b1\u20090.428.6\u2009\u00b1\u20093.9\\*9.2\u2009\u00b1\u20093.5\\*Plasma triglycerides (mmol/L)1.4\u2009\u00b1\u20091.21.44\u2009\u00b1\u20090.811.41\u2009\u00b1\u20090.55Total cholesterol (mmol/L)4.24\u2009\u00b1\u20090.843.8\u2009\u00b1\u20091.73.9\u2009\u00b1\u20091.8High-density lipoprotein (mmol/L)1.28\u2009\u00b1\u20090.331.9\u2009\u00b1\u20091.2\\*2.5\u2009\u00b1\u20092.0\\*Low-density lipoprotein (mmol/L)2.61\u2009\u00b1\u20090.612.30\u2009\u00b1\u20090.992.17\u2009\u00b1\u20090.89Plasma creatinine (mmol/L)63\u2009\u00b1\u20091655\u2009\u00b1\u20093352\u2009\u00b1\u200945eGFR (mL/min/1.73\u2009m^2^)100\u2009\u00b1\u20091293\u2009\u00b1\u20092682\u2009\u00b1\u200921LVEDV (ml)128\u2009\u00b1\u200927120\u2009\u00b1\u200927120\u2009\u00b1\u200919LVESV (ml)53\u2009\u00b1\u20092045\u2009\u00b1\u20091447\u2009\u00b1\u200912LVSV (ml)75\u2009\u00b1\u20091975\u2009\u00b1\u20091773\u2009\u00b1\u200913LVEF (%)58.6\u2009\u00b1\u20099.062.8\u2009\u00b1\u20096.461.3\u2009\u00b1\u20096.9LV myocardial mass (g)83\u2009\u00b1\u20092691\u2009\u00b1\u20093089\u2009\u00b1\u200922RVEDV (ml)100\u2009\u00b1\u200932102\u2009\u00b1\u20092899\u2009\u00b1\u200917RVESV (ml)48\u2009\u00b1\u20091643.7\u2009\u00b1\u20091.560.1\u2009\u00b1\u20098.9\\*^\u00a7^RVSV (ml)53\u2009\u00b1\u20091858\u2009\u00b1\u20091839\u2009\u00b1\u200911\\*^\u00a7^RVEF (%)52.5\u2009\u00b1\u20097.156.8\u2009\u00b1\u20096.538.5\u2009\u00b1\u20097.7\\*^\u00a7^RV myocardial mass (g)18.6\u2009\u00b1\u20096.117.4\u2009\u00b1\u20096.618.4\u2009\u00b1\u20093.6Notes: The values are the mean\u2009\u00b1\u2009SD, Numbers in the brackets are percentages.\\*P\u2009\\<\u20090.05 vs. normal group; ^\u00a7^P\u2009\\<\u20090.05 vs. T2DM with normal RVEF. One-way ANOVA test was used to assess the differences in continuous variables among the three groups. The chi-square test was performed for categorical values.DM, diabetes mellitus; RVEF, right ventricular ejection fraction; BSA, body surface area; HbA1c, glycosylated hemoglobin; eGFR, estimated glomerular filtration rate; LV, left ventricular; RV, right ventricular; EDV, end diastolic volume; ESV, end systolic volume; EF, ejection fraction; SV, stroke volume.\n\nGlobal and regional strain analysis in T2DM patients and normal controls {#Sec10}\n------------------------------------------------------------------------\n\nData on RV global strain parameters for all participants are shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}. The magnitude of both global longitudinal PS (GLPS) and global circumferential PS (GCPS) were significantly decreased in T2DM patients, with or without reduced RVEF, than in healthy subjects (*p*\u2009\\<\u20090.05). Among T2DM patients, the magnitude of GLPS was significantly reduced in the reduced RVEF group than in the preserved RVEF group \\[\u22125.00(\u22128.00--(\u22121.40))% vs. \u22126.85(\u221210.78--(\u22125.02))%; *p*\u2009\\<\u20090.05\\]. There was a strong trend toward decreased magnitude of global longitudinal PDSR of T2DM patients with decreased RVEF than that of healthy subjects and T2DM patients with preserved RVEF (*p*\u2009\\<\u20090.05). The magnitude of Global radial PSSR in T2DM patients with reduced RVEF was reduced compared to that in normal subjects, while the global circumferential PSSR was higher in T2DM patients with or without preserved RVEF than that in healthy subjects (*p*\u2009\\<\u20090.05).Figure 2Box-and-whisker plot for comparison of global longitudinal PS, PSSR, PDSR (**A**,**D**,**G**), circumferential PS, PSSR, PDSR (**B**,**E**,**H**), and radial PS, PSSR, PDSR (**C**,**F**,**I**) in control subjects versus T2DM patients with preserved (T2DM1) and impaired RVEF (T2DM2). Kruskal-Wallis test was used to assess the differences in continuous variables among the three groups. Box plot: box length represents the interquartile range; horizontal box line represents the median; whiskers represent the maximum and minimum values; asterisk denotes P\u2009\\<\u20090.05 compared to controls; section denotes P\u2009\\<\u20090.05 T2DM compared to T2DM patients with normal RVEF. T2DM indicates, type 2 diabetes mellitus; RVEF, right ventricular ejection fraction; PS, peak strain; PSSR, peak systolic strain rate; PDSR, peak diastolic strain rate.\n\nData on regional RV deformation parameters among T2DM patients and healthy subjects are provided in Table\u00a0[2](#Tab2){ref-type=\"table\"}. Individual-slice analysis showed that the magnitude of PS (longitudinal and circumferential) was significantly decreased at mid-ventricular and apical slices in T2DM patients with or without reduced RVEF than in healthy subjects (*p*\u2009\\<\u20090.05). The magnitude of longitudinal PDSR at mid-ventricular segments in T2DM patients with or without systolic RV dysfunction, as well as the longitudinal PDSR at apical slices in T2DM patients with reduced RVEF, was significantly reduced than that of controls (*p*\u2009\\<\u20090.05). Among the T2DM patients, the magnitude of longitudinal PS and PDSR at mid-ventricular segments were significantly decreased in the reduced RVEF group relative to the preserved RVEF group \\[0.73(\u22128.62--7.22)% vs. \u22127.69(\u221211.81--(\u22123.89))%; \u22120.22(\u22121.00--0.54) 1/s vs. 0.68(\u22120.27--0.90) 1/s; *p*\u2009\\<\u20090.05, respectively\\]. Interestingly, the magnitude of circumferential PS and PDSR at basal segments were significantly higher in T2DM patients with preserved RVEF than in healthy subjects (*p*\u2009\\<\u20090.05).Table 2Values for Right ventricular deformation parameters obtained using feature tracking for slice values of all study groups (basal, mid-ventricular, apical).BasalMid-ventricularApicalHealthy Controls\\\n(n\u2009=\u200926)DM with preserved\\\nRVEF (n\u2009=\u200990)DM with reduced\\\nRVEF (n\u2009=\u200914)Healthy Controls\\\n(n\u2009=\u200926)DM with preserved\\\nRVEF (n\u2009=\u200990)DM with reduced\\\nRVEF (n\u2009=\u200914)Healthy Controls\\\n(n\u2009=\u200926)DM with preserved\\\nRVEF (n\u2009=\u200990)DM with reducedvRVEF (n\u2009=\u200914)**Longitudinal**PS (%)\u221210.48(\u221211.74--(\u22128.41))\u22126.83(\u221210.18--9.43)\u22126.95(\u22128.36--9.43)\u221211.27(\u221213.57--(\u22128.86))\u22127.69(\u221211.81--(\u22123.89))\\*0.73(\u22128.62--7.22)^\u00a7^\\*\u221214.87(\u221217.30--(\u221213.21))\u221211.97(\u221214.25--(\u22127.63))\\*\u22129.88(\u221213.13--(\u22125.38))\\*PSSR (1/s)\u22120.58(\u22120.77--0.82)\u22120.56(\u22120.92--1.25)\u22120.83(\u22121.02--0.94)\u22120.67(\u22120.95--(\u22120.49))\u22120.71(\u22120.95--0.54)\u22120.14(\u22120.80--0.72)\u22120.86(\u22121.03--(\u22120.71))\u22120.84(\u22121.05--(\u22120.63))\u22120.75(\u22120.98--(\u22120.22))PDSR (1/s)0.64(0.23--0.97)0.63(\u22121.05--1.04)0.59(\u22121.01--0.77)0.92(0.58--1.09)0.68(\u22120.27--0.90)\\*\u22120.22(\u22121.00--0.54)^\u00a7^\\*0.92(0.73--1.13)0.86(0.69--1.09)0.66(0.26--0.95)\\***Circumferential**PS (%)10.90(5.77--14.58)14.26(8.23--20.47)\\*12.90(5.09--25.35)\u22127.15(\u22128.71--(\u22124.99))\u22123.83(\u22126.16\u22127.14)\\*\u22123.71(\u22126.27--4.62)\\*\u221210.55(\u221212.93--(\u22126.86))\u22127.08(\u221211.06--(\u22121.00))\\*\u22126.66(\u22128.19--(\u22122.44))\\*PSSR (1/s)0.68(0.44--1.12)1.02(0.53--1.51)0.80(0.52--1.95)\u22120.55(\u22120.69--(\u22120.35))1.02(0.53--1.51)\\*\u22120.51(\u22120.64--0.44)\u22120.71(\u22120.94--(\u22120.42))\u22120.63(\u22120.89--0.48)\u22120.53(\u22120.70--0.39)PDSR (1/s)\u22120.83(\u22121.01--(\u22120.56))\u22121.09(\u22121.65--(\u22120.70))\\*\u22120.97(\u22121.23--0.82)0.54(0.36--0.84)0.28(\u22120.84--0.61)\\*0.37(\u22120.48--0.58)0.63(0.51--0.87)0.66(\u22120.49--0.92)0.66(0.47--0.84)**Radial**PS (%)39.98(28.33--50.51)38.89(25.46--54.71)18.52(11.49--26.30)22.83(15.40--29.45)18.70(11.07--28.57)18.52(11.48--26.30)9.23(\u22122.67--16.68)8.84(\u22122.82--13.20)7.64(\u22123.17--11.08)PSSR (1/s)2.28(1.68--2.62)2.03(1.34--2.97)1.80(1.18--2.37)1.44(1.03--1.68)1.08(0.77--1.64)\\*1.37(0.92--1.70)0.92(0.45--1.31)0.88(0.44--1.24)0.72(0.52--0.91)PDSR (1/s)\u22122.30(\u22123.40--(\u22121.76))\u22122.17(\u22123.22--(\u22121.60))\u22122.00(\u22122.79--(\u22121.62))\u22121.61(\u22121.89--(\u22121.08))\u22121.21(\u22121.59--(\u22120.86))\\*\u22121.11(\u22121.59--0.85)\u22120.87(\u22121.35--(\u22120.43))\u22120.92(\u22121.48--(\u22120.40))\u22120.85(\u22121.08--(\u22120.34))Notes: Data are presented as the median (25th, 75th percentile).\\*P\u2009\\<\u20090.05 vs. normal group. ^\u00a7^P\u2009\\<\u20090.05 vs. T2DM with normal RVEF. The Kruskal-Wallis test was used to assess the differences in continuous variables among the three groups. DM, diabetes mellitus; RVEF, right ventricular ejection fraction; PS, peak strain; PSSR, peak systolic strain rate; PDSR, peak diastolic strain rate.\n\nCorrelations of RV global strain parameters with RV and LV dimensional and functional parameters and biochemical markers {#Sec11}\n------------------------------------------------------------------------------------------------------------------------\n\nThe associations of the global strain parameters with HbA1c, fasting blood glucose, plasma triglycerides, total cholesterol, high-density lipoprotein, low-density, lipoprotein, plasma creatinine, estimated glomerular filtration rate, LV end diastolic volume (EDV), LV end systolic volume (ESV), LVSV, LVEF, LV and RV myocardial mass, RVEDV, RVESV, RVSV, and RVEF are tested. The significant associations are shown in Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}. RV GLPS was correlated to LV mass (r\u2009=\u20090.209, *p*\u2009=\u20090.035). RV GCPS was associated with LV mass (r\u2009=\u20090.230, *p*\u2009=\u20090.020).Figure 3Scattergrams show the results of Pearson correlation analysis between LV mass and global (**A**) longitudinal, and (**B**) circumferential peak strain measurements.\n\nPredictors of RV dysfunction {#Sec12}\n----------------------------\n\nAs shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}, a binary logistic regression model that used RV dysfunction (RVEF\u2009\\<\u200945%) as the dependent variable identified GLPS as an independent predictor of RV dysfunction (odds ratio: 1.246, 95% CI: 1.037--1.496; *p*\u2009=\u20090.019), adjusted for age, sex, body mass index, glycosylated hemoglobin, high-density lipoprotein, global circumferential PS, and global radial PS.Table 3Univariate and multivariable stepwise backward logistic regression analysis for the prediction of RV dysfunction (n\u2009=\u2009130).VariableUnivariateMultivariateRP valueOdds ratio95% CIP valueAge (years)\u22120.0550.5330.9750.906--1.0490.497Sex (male)0.1120.2060.0510.002--1.2380.067BMI (kg/m^2^)\u22120.0820.3570.8780.639--1.2060.421Height (cm)0.1270.150Weight (kg)0.0220.807Systolic blood pressure (mmHg)0.0120.904Diastolic blood pressure (mmHg)0.0560.563Heart rate (beats/min)\u22120.1080.225Duration of diabetes (years)\u22120.0770.479HbA1c (%)0.1120.2421.1600.857--1.5700.337Fasting blood glucose (mmol/L)0.1510.159Plasma triglycerides (mmol/L)\u22120.0060.957Total cholesterol (mmol/L)\u22120.0020.986High-density lipoprotein (mmol/L)0.1990.042\\*1.3320.805--2.2050.264Low-density lipoprotein (mmol/L)--0.0730.460Plasma creatinine (mmol/L)\u22120.0520.595eGFR (mL/min/1.73\u2009m^2^)\u22120.0850.419GLPS (%)0.2220.011\\*1.2461.037--1.496**0.019\\***GCPS (%)0.0490.5830.9730.859--1.1040.674GRPS (%)\u22120.1350.1260.9780.899--1.0640.606Notes: GCPS, global circumferential peak strain; GRPS, global radial peak strain; GLPS, global longitudinal peak strain; HbA1c, glycosylated hemoglobin; eGFR, estimated glomerular filtration rate; BMI, body mass index.\n\nAnalysis of diagnostic performance {#Sec13}\n----------------------------------\n\nReceiver operating characteristic (ROC) analysis showed the predictive value of RV strain parameters, including global longitudinal, circumferential, and radial PS, for RV dysfunction in T2DM patients (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). The area under the ROC curve and sensitivity and specificity of the RV global strain parameters used for the discrimination of RV deformation between patients and controls are summarized in Table\u00a0[4](#Tab4){ref-type=\"table\"}.Figure 4Bland--Altman plots with limits of agreement (95% confidence intervals) demonstrating the intra-observer (**A**--**C**) and inter-observer (**D**--**F**) reproducibility of CMR myocardial feature tracking strain parameters: GRPS\u2009=\u2009global radial peak strain; GCPS\u2009=\u2009global circumferential peak strain; GLPS\u2009=\u2009global longitudinal peak strain. Solid lines represent bias (blue) and 95% limits of agreement (orange).Table 4Inter- and intra-observer variability of tissue tracking (n\u2009=\u200920).Inter-observer variabilityIntra-observer variabilityMean bias\u2009\u00b1\u2009SD95%CIMean bias\u2009\u00b1\u2009SD95%CIGLPS (%)\u22120.39\u2009\u00b1\u20091.65\u22123.62--2.85\u22120.02\u2009\u00b1\u20091.85\u22123.64--3.61GCPS (%)\u22120.19\u2009\u00b1\u20094.64\u22129.28--8.90\u22120.54\u2009\u00b1\u20092.6\u22125.63--4.56GRPS (%)1.08\u2009\u00b1\u20093.71\u22126.19--8.353.00\u2009\u00b1\u20096.03\u22128.81--14.81Note: 95%CI, 95% Confidence interval. GRPS\u2009=\u2009global radial peak strain; GCPS\u2009=\u2009global circumferential peak strain; GLPS\u2009=\u2009global longitudinal peak strain.\n\nInter-observer and intra-observer variabilities {#Sec14}\n-----------------------------------------------\n\nThe Bland--Altman analysis indicated moderate-to-excellent agreement between intra- and inter-observer acquisitions in measuring global radial, circumferential, and longitudinal PS (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). The mean bias\u2009\u00b1\u2009standard deviations and 95% confidence interval of the Bland--Altman analysis are summarized in Table\u00a0[5](#Tab5){ref-type=\"table\"}.Figure 5ROC analysis of global longitudinal PS (blue), circumferential PS (orange), and radial PS (green) between patients with T2DM and controls. Abbreviations: T2DM, type 2 diabetes mellitus; PS, peak strain; ROC, receiver operating characteristic.Table 5ROC analysis of CMR feature tracking for detecting RV dysfunction between T2DM patients and normal controls.Cut offAUCSensitivity (%) (95%CI)Specificity (%) (95%CI)Global Longitudinal PS (%)\u22125.140.71657.14 (28.9--82.2)79.31 (70.8--86.3)Global Circumferential PS (%)5.230.57950.00 (23.1--76.9)72.41 (63.3--80.3)Global Radial PS (%)27.050.617100.00 (76.7--100.0)28.45 (20.5--37.6)Notes: ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval. PS, peak strain.\n\nDiscussion {#Sec15}\n==========\n\nIn this study, we describe the results of a prospective study on detailed RV global and regional myocardial strain analysis in adult T2DM patients using CMR-FT. To the best of our knowledge, this is the first report of its kind. We demonstrated that CMR-FT can be used to quantify RV myocardial deformation in T2DM patients and that these measurements can provide objective evaluations of subclinical RV dysfunction before clinically manifested reduction in RVEF; specifically, we showed that GLPS has the greatest diagnostic potential among the tested RV feature-tracking parameters. Further, we concluded that GLPS might provide additional information in predicting RV dysfunction in T2DM patients.\n\nRecently, CMR-FT has emerged as a novel tool to quantitatively assess biventricular myocardial function^[@CR16]--[@CR18]^. CMR-FT utilizes tissue voxel motion-tracking technology based on CMR cine images. As its main advantage, it has a relatively simple post-processing procedure. Further, the estimated 10-min post-processing time for each subject implies that this method can be used as a routine procedure^[@CR22]--[@CR24]^. Intra-observer reproducibility of CMR-derived strain analysis does not depend on the magnetic field strength^[@CR25]^, and in our study, both intra- and inter-observer reproducibilities for global PS were excellent. Further, the reproducibilities of the radial PS measurements were better than that reported previously^[@CR26]^. Previous CMR-FT studies using two-dimensional imaging model have shown that normal RV strain was almost twice as much as ours; the reason is that our study was based on 3D CMR-FT model with totally different algorithm^[@CR27],[@CR28]^.\n\nRV function is often difficult to evaluate because of its complex shape and motion. CMR imaging is indeed regarded as the gold standard for assessing RV dimension and function. However, because of the complex RV contraction patterns, regional wall motion abnormalities are still inadequately analyzed. Measures of myocardial deformation, such as strain analysis, are emerging modalities to quantitatively assess regional and global RV function^[@CR29]^. The usefulness of myocardial strain imaging to quantify RV function in T2DM has been previously demonstrated using tissue Doppler and speckle-tracking echocardiography studies, which consistently showed reduced RV myocardial strain in T2DM patients compared with that in healthy subjects. However, these studies were of limited value in the exact quantification of RV function and chamber size^[@CR14],[@CR30]--[@CR32]^. However, despite the large volume of information on echocardiographic imaging data, information on the applicability of deformation imaging in T2DM using CMR is still limited. Thus, our results provide complementary and more accurate information on detailed global and regional RV strain measurements with respect to echocardiographic data. Myocardial tagging is considered the standard or reference CMR technique to quantitatively evaluate LV regional strain; however, applying tagging lines to the thin myocardial wall of the RV and the time-consuming and complex procedure of data acquisition and post-processing have limited its popularity^[@CR4]^. Conversely, based on the results of a previous study that are comparable to those obtained using CMR tagging, CMR-FT imaging is a promising technique to overcome these limitations and challenges^[@CR33]^. Furthermore, recent advances in feature-tracking technology offer powerful modalities for evaluating RV contractile function and diastolic relaxation in patients without RV hypertrophy^[@CR34]^. However, Wehner *et al*. reported that CMR-FT inaccurately quantifies LV strain values when compared to displacement encoding with stimulated echoes (DENSE) imaging as a reference standard. However, they only evaluated the agreement between measurements of two techniques, and they could not evaluate the prognostic utility of the measures^[@CR35]^. These results could not deny the diagnostic and prognostic values of strain measurements derived from CMR-FT which has been established^[@CR9],[@CR36]^. Furthermore, when applied to RV, DENSE imaging was limited by its lengthy scan time, lower resolution, difficulty to distinguish the minimum from the intermediate principal strains, and the reduction in signal to noise ratio associated with the stimulated echo^[@CR37]^.\n\nWe showed impaired RV global and regional strain in T2DM patients compared with those in healthy individuals and notably the ability of GLPS to detect and predict RV dysfunction. These RV strain changes in T2DM patients seemed to precede overt RV dysfunction, as they were recognized even in patients without impaired RVEF. As described elsewhere, the deep muscle layer of the RV myocardial wall is primarily composed of longitudinal fibers^[@CR33]^; this architectural pattern of the RV contributes to the predominantly longitudinal shortening that leads to blood ejection during systolic phase. Importantly, this structure can also explain the predominance of longitudinal RV strain changes observed in our study. Specifically, GLPS with optimal cut-off value of \u22125.14% had a higher diagnostic accuracy in identifying RV dysfunction in T2DM patients without clinical symptoms of heart failure, thus fulfilling the requirement for more sensitive CMR parameters that can be used for early diagnosis of this condition^[@CR38]^. Data obtained from the ROC analysis demonstrated the diagnostic performance of global radial, circumferential, and longitudinal PS, as the criteria for differentiating RV dysfunction was moderate in this patient group. Peter *et al*. have established that the LV global longitudinal strain could be used for prediction of silent myocardial infarction which is associated with significant mortality and morbidity in T2DM patients^[@CR39]^. Whether RV GLPS in diabetes is predictive of silent myocardial infarction should be clarified in the future. Interestingly, the circumferential strain at basal slices increased in T2DM patients with preserved RVEF. We hypothesized that in the early stages of diabetes, epicardial fibers remaining are spared, which compensate for the longitudinal dysfunction and thus preserve SV and ejection fraction^[@CR25]^. Blood pressure control in hypertensive patients has been shown to decrease circumferential strain, which also might be helpful to explain it as our cohort included T2DM patients with higher blood pressure than controls^[@CR25]^. As subclinical RV dysfunction reflects the structural and metabolic milieu of the myocardium, RV strain data might be useful to develop targeted therapeutic strategies that modulate cardiac metabolism and prevent heart failure in T2DM^[@CR40]^. Besides, further studies are required to determine whether global and regional RV strain parameters, especially of GLPS, can be used as risk factors to predict outcomes in T2DM patients.\n\nIn this study, we have not found any significant correlations between the HbA1c level and RV volumes and function parameters, which does not agree with the findings of previous studies^[@CR41]^. To the best of our knowledge, significant relationships between RV GLPS, GCPS, and LV mass were observed in our study, which has not been demonstrated. The concentric LV hypertrophy, defined as increased LV mass, has shown to be a strong predictor of adverse cardiovascular events in T2DM patients^[@CR7]^. In our cohort, we did not observe signifancant increased LV mass in diabetes patients. We may hypothezised that impaired RV GLPS and GCPS in T2DM patients might be predictors for adverse clinical outcomes. Further follow-up studies in this patient population are warranted.\n\nOur study has several limitations. First, this was a single-center study with a limited sample size and included only 14 subjects with impaired RVEF. Therefore, these findings need further validation using larger cohort or multi-center results. Second, several software can be used to analyze RV deformation, data on RV myocardial strain quantification using CMR-FT is insufficient. Thus, reference values for RV strain should be determined. Finally, as clinical follow-up data were not available for the study patients, the clinical correlation and prognostic implications of RV strain parameters in T2DM patients should be verified in our futher studies.\n\nIn summary, we showed that RV myocardial strain assessment can be used to identify RV deformation in T2DM patients, even in those with preserved RVEF, implying that RV strain parameters is more sensitive than RVEF values for the early detection of subclinical RV dysfunction in T2DM patients. Further, RV strain parameters may have diagnostic value for RV dysfunction and might be useful predictors of RV dysfunction in T2DM patients.\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nBi-yue Hu, Jin Wang, Zhi-gang Yang and Ying-kun Guo contributed equally.\n\nThis work was supported by the National Natural Science Foundation of China (81471721, 81471722), Program for New Century Excellent Talents in University (No: NCET-13-0386), Program for Young Scholars and Innovative Research Team in Sichuan Province (2017TD0005) of China, and Program of Scientific and Technology Bureau of Chengdu City (2015-HM01-00473-SF), 1\u00b73\u00b75 project for disciplines of excellence, West China Hospital, Sichuan University.\n\nConceived and designed the experiments: B.Y. Hu, J. Wang, Z.G. Yang, Y.K. Guo. Performed the experiments: B.Y. Hu, J. Wang, Z.G. Yang, Y. Ren, L. Jiang, L.J. Xie, X. Liu, Y. Gao, M.T. Shen, H.Y. Xu, K. Shi, Z.L. Li, C.C. Xia, W.L. Peng, M.Y. Deng, H. Li, Y.K. Guo. Analyzed the data: B.Y. Hu, J. Wang, Z.G. Yang, Y. Ren, L. Jiang, L.J. Xie, X. Liu, Y. Gao, M.T. Shen, Y.K. Guo. Contributed reagents/materials/analysis tools: B.Y. Hu, J. Wang, Z.G. Yang, H.Y. Xu, K. Shi, Z.L. Li, C.C. Xia, W.L. Peng, M.Y. Deng, H. Li, Y.K. Guo. Wrote and review the paper: B.Y. Hu, J. Wang, Z.G. Yang, Y. Ren, L. Jiang, L.J. Xie, X. Liu, Y. Gao, M.T. Shen, Y.K. Guo.\n\nThe datasets used during the current study are available from the corresponding author on reasonable request.\n\nThe authors declare no competing interests.\n"}
+{"text": "Editor: M. Paul\n\nIntroduction {#sec1}\n============\n\nChest computed tomography (CT) has shown promise as a diagnostic tool for coronavirus disease 2019 (COVID-19) \\[[@bib1],[@bib2]\\]. Initial studies have described a stereotypical time course with successive radiological stages, ground-glass opacities (GGOs) being the main initial lesion \\[[@bib3],[@bib4]\\]. Quantitative CT lung analysis shows a good association with patient prognosis in those with non-COVID-19 acute respiratory distress syndrome \\[[@bib5]\\]. From March 2020, patients admitted to the Strasbourg University Hospital with suspected COVID-19 were managed using a specific protocol, including reverse-transcription polymerase chain reaction (RT-PCR) on respiratory samples and a systematic chest CT to improve the triage of patients. We aimed to determine the early prognostic value of systematic chest CT with quantification of lung lesions in COVID-19 patients performed at the time of admission.\n\nMethods {#sec2}\n=======\n\nWe conducted a retrospective study of prospectively collected data. We included patients with positive PCR and CT performed, this study was not interventional so the chest CT was not realized after inclusion in the study. All consecutive patients aged \u226518 years---hospitalized at the Strasbourg University Hospital in March 2020 with COVID-19 confirmed using RT-PCR and a chest CT performed with quantitative evaluation of the lesions---were enrolled in this cohort study. Patients for whom the CT was realized \u226548\u00a0h after admission were excluded. Non-contrast-enhanced chest CT images were acquired on an 80-row scanner (Aquilion Prime SP, Canon Medical Systems), with parameters based on the patient\\'s morphotype (tension 100--135 kV and maximum mAs 2--50) \\[[@bib6]\\]. Images were reconstructed with a slice thickness of 1 mm in mediastinal and parenchymal windows using an iterative reconstruction algorithm (AIDR-3D, Canon Medical Systems) and read on dedicated workstations with multiplanar and maximum intensity projection reconstructions. CT angiography was performed secondarily for patients with suspected pulmonary embolism. Visual quantification of the lung lesions was performed at the time of the CT by two different radiologists who were blinded to the patients\\' clinical condition. Evaluations were made independently, but discrepancies were resolved by consensus. The CT images were classified as per the percentage of the whole lung parenchyma affected by COVID-19 lesions---GGO and/or consolidations---in the following six groups: normal CT (no lesion), minimal (0--10%), moderate (11--25%), important (26--50%), severe (51--75%), and critical (\\>75%) \\[[@bib7]\\]. To simplify the analysis of the clinical data, the patients were divided into three subgroups: \u226425%, 26--50%, and \\>50%.\n\nThe primary endpoint was early severe disease, defined as death or intensive care unit (ICU) admission in the 7\u00a0days after hospital admission. Statistical analyses were performed using R software (version 3.5.2). This study was approved by the Ethics Committee of Strasbourg University Hospital (N\u00b0CE--2020--51). Oral informed consent was obtained from all the patients.\n\nResults {#sec3}\n=======\n\nDuring the study period, 572 patients were assessed out of 854 patients with positive RT-PCR ([Supplementary Material Fig.\u00a0S1](#appsec1){ref-type=\"sec\"}). The mean patient age was 66.0\u00a0years (standard deviation 16.0; range 20--95\u00a0years), and 343/572 patients (60.0%) were men. Among all 572 patients, chest CT was normal in 14 patients (2.4%), and showed minimal, moderate, important, severe, and critical lesions in 68 (11.9%), 224 (39.2%), 171 (29.9%), 82 (14.3%), and 13 patients (2.3%), respectively. Most patients had bilateral involvement (524/572, 91.6%) and GGO (540/572, 94.4%). Consolidations were observed in 372/572 patients (65.0%).\n\nThere were no significant differences in the prevalence of comorbidities based on the extent of the lesions on CT ([Table\u00a01](#tbl1){ref-type=\"table\"} ). Patients with lung involvement \\>50% had a significantly higher C-reactive protein (CRP) level and neutrophil count, lower lymphocyte count, and more consolidations on CT compared to those with lung involvement \u226425% (p\u00a0\\<\u00a00.01 for each comparison). Finally, 16/95 patients (16.8%) with lung involvement \\>50% were diagnosed with pulmonary embolism.Table\u00a01Baseline characteristics of the 572 COVID-19 patients according to the extent of lesions on CTTable\u00a01Extent of lesions on CTp\u226425% (*n*\u00a0=\u00a0306)26--50% (*n*\u00a0=\u00a0171)\\>50% (*n*\u00a0=\u00a095)\u00a0Age, mean\u00a0\u00b1\u00a0SD (years)66.5\u00a0\u00b1\u00a016.265.2\u00a0\u00b1\u00a016.265.6\u00a0\u00b1\u00a014.90.69\u00a0Male sex153 (50.0)114 (66.7)76 (80.0)\\<0.01\u00a0Body mass index, mean\u00a0\u00b1\u00a0SD (kg/m^2^)28.7\u00a0\u00b1\u00a06.0 (*n*\u00a0=\u00a0265)29.0\u00a0\u00b1\u00a05.9 (*n*\u00a0=\u00a0151)29.6\u00a0\u00b1\u00a04.3 (*n*\u00a0=\u00a086)0.13Comorbidity\u00a0Diabetes76 (24.8)44 (25.7)25 (26.3)0.95\u00a0Hypertension161 (52.6)87 (50.9)49 (51.6)0.93\u00a0Chronic heart failure30 (9.8)21 (12.3)5 (5.3)0.18\u00a0Chronic lung disease58 (19.0)23 (13.5)18 (18.9)0.28\u00a0Immunodepression8 (2.6)5 (2.9)3 (3.2)0.94\u00a0Active malignancy20 (6.5)9 (5.3)4 (4.2)0.66Clinical findings\u00a0Fever224 (73.2)146 (85.4)67 (70.5)\\<0.01\u00a0Dyspnea180 (58.8)140 (81.9)82 (86.3)\\<0.01\u00a0Cough192 (62.7)126 (73.7)58 (61.1)0.03\u00a0Chest pain28 (9.2)17 (9.9)7 (7.4)0.78\u00a0SpO~2~ (%)94\u00a0\u00b1\u00a04 (*n*\u00a0=\u00a0302)92\u00a0\u00b1\u00a06 (*n*\u00a0=\u00a0166)90\u00a0\u00b1\u00a08 (*n*\u00a0=\u00a095)\\<0.01\u00a0Maximal oxygen level (L/min)2\u00a0\u00b1\u00a03 (*n*\u00a0=\u00a0288)4\u00a0\u00b1\u00a05 (*n*\u00a0=\u00a0159)9\u00a0\u00b1\u00a012 (*n*\u00a0=\u00a080)\\<0.01Time between symptom onset and CT performance (days)6\u00a0\u00b1\u00a067\u00a0\u00b1\u00a067\u00a0\u00b1\u00a04\\<0.01Imaging findings\u00a0Bilateral involvement260 (85.0)169 (98.8)95 (100.0)\\<0.01\u00a0Ground-glass opacities277 (90.5)169 (98.8)94 (98.9)\\<0.01\u00a0Consolidations174 (56.9)127 (74.3)71 (74.7)\\<0.01\u00a0Micronodules20 (6.5)4 (2.3)7 (7.4)0.10\u00a0Pulmonary embolism[a](#tbl1fna){ref-type=\"table-fn\"}7 (2.3)6 (3.5)16 (16.8)\\<0.01Laboratory findings\u00a0C-reactive protein (mg/L)59\u00a0\u00b1\u00a085 (*n*\u00a0=\u00a0300)104\u00a0\u00b1\u00a084 (*n*\u00a0=\u00a0169)154\u00a0\u00b1\u00a0114 (*n*\u00a0=\u00a091)\\<0.01\u00a0Neutrophil count (cells/mm\u00b3)4000\u00a0\u00b1\u00a02877 (*n*\u00a0=\u00a0302)5100\u00a0\u00b1\u00a02900 (*n*\u00a0=\u00a0169)6375\u00a0\u00b1\u00a04592 (*n*\u00a0=\u00a094)\\<0.01\u00a0Lymphocyte count (cells/mm\u00b3)900\u00a0\u00b1\u00a0527 (*n*\u00a0=\u00a0302)930\u00a0\u00b1\u00a0590 (*n*\u00a0=\u00a0169)740\u00a0\u00b1\u00a0475 (*n*\u00a0=\u00a095)\\<0.01\u00a0Serum creatinine (\u03bcmol/L)74\u00a0\u00b1\u00a035 (*n*\u00a0=\u00a0301)77\u00a0\u00b1\u00a035 (*n*\u00a0=\u00a0170)84\u00a0\u00b1\u00a047 (*n*\u00a0=\u00a095)0.2\u00a0Aspartate aminotransferase (U/L)37\u00a0\u00b1\u00a026 (*n*\u00a0=\u00a0242)47\u00a0\u00b1\u00a031 (*n*\u00a0=\u00a0143)58\u00a0\u00b1\u00a042 (*n*\u00a0=\u00a084)\\<0.01\u00a0Lactate (mmol/L)0.9\u00a0\u00b1\u00a00.5 (*n*\u00a0=\u00a0186)1.1\u00a0\u00b1\u00a00.7 (*n*\u00a0=\u00a0123)1.2\u00a0\u00b1\u00a00.9 (*n*\u00a0=\u00a083)\\<0.01Outcome\u00a0Severe disease on day 7[b](#tbl1fnb){ref-type=\"table-fn\"}70 (22.9)70 (40.9)66 (69.5)\\<0.01\u00a0Severe disease on day 30[b](#tbl1fnb){ref-type=\"table-fn\"}82\u00a0(26.8)74\u00a0(43.3)71 (74.7)\\<0.01\u00a0Death on day 719 (6.2)20 (11.7)16 (16.8)\\<0.01\u00a0Death on day 3033 (10.8)29 (17.0)27 (28.4)\\<0.01[^1][^2][^3][^4]\n\nOverall, 206/572 patients (36.0%) met the criteria for early severe disease, including 55/572 (9.6%) who died. The extent of lesions on the initial CT was associated with severe disease ([Fig.\u00a01](#fig1){ref-type=\"fig\"} A). Among the 14 patients with normal chest CT at admission, none developed severe disease. Most patients with lung involvement \\>50% were admitted to the ICU or died (66/95, 69.5%), and this rate was lower in patients with lung involvement of 26--50% (70/171, 40.9%) and \u226425% (70/306, 22.9%) ([Table\u00a01](#tbl1){ref-type=\"table\"}). In multivariate analysis, lung involvement \\>50% was significantly associated with early severe disease (odds ratio 2.35, confidence interval 1.24--4.46; p\u00a0\\<\u00a00.01) ([Supplementary Material Table\u00a0S1](#appsec1){ref-type=\"sec\"}). Survival analysis showed a significantly reduced 30-day event-free survival in patients with lung involvement of 26--50% and \\>50% (p\u00a0\\<\u00a00.001) ([Supplementary Material Fig.\u00a0S2](#appsec1){ref-type=\"sec\"}).Fig.\u00a01Outcome and time from onset of symptoms as per the quantification of lesions on computed tomography (CT). (A) Histogram showing the outcome according to the extent of lesions on CT. (B) Box plot showing the time between symptom onset and CT performance as per the extent of lesions on CT. (C) Scatter plot showing the extent of lesions on CT as per the time between symptom onset and CT performance. In order to ease visualization, noise was randomly added to each point. Curves were fitted through points with the locally weighted scatterplot smoothing (LOESS) method using the 'ggplot 2' R package. Shaded area represents the standard error.Fig.\u00a01\n\nThe median time between the onset of the symptoms and CT was 7\u00a0days (interquartile range 6.0). This duration was shorter in patients with normal CT or minimal lesions ([Fig.\u00a01](#fig1){ref-type=\"fig\"}B). None of the patients with minimal lesions and symptoms for more than 10\u00a0days has developed severe disease, while 10/12 (83.3%) of those with minimal lesions and severe disease presented with symptoms for \u22645\u00a0days ([Fig.\u00a01](#fig1){ref-type=\"fig\"}C).\n\nDiscussion {#sec4}\n==========\n\nOur study has shown that visual quantification of CT lung lesions is associated with early death or ICU admission in hospitalized patients, especially in patients with lung involvement \\>50%.\n\nSeveral risk factors for severe COVID-19 have been reported, such as older age, male sex, and chronic diseases \\[[@bib8],[@bib9]\\]. The chest CT has shown benefit in the diagnosis of COVID-19 pneumonia; however, its relevance as a prognostic factor remains unclear \\[[@bib1]\\]. Based on a study of 134 COVID-19 patients, Liu et\u00a0al. showed that CT quantification of pneumonia lesions can predict early progression to severe illness \\[[@bib10]\\].\n\nAlthough our study has employed one of the largest cohorts on COVID-19 imaging, it has some limitations. We chose to evaluate early prognosis with the outcome on day 7; a longer endpoint may have increased the number of patients with severe disease. However, most deaths and ICU admissions occurred within 7\u00a0days after admission in our study, and peak lung involvement was reached before 2\u00a0weeks of evolution in previous studies \\[[@bib3],[@bib11]\\]. Furthermore, we performed visual quantification, while other studies have used dedicated software to quantify lung lesions \\[[@bib10],[@bib11]\\]. Although this makes our evaluation dependent on the experience of radiologists, this facilitates its generalization to centres that are not equipped with such software.\n\nFour radiological stages have been described, with progressive extent of GGO and the secondary onset of consolidations \\[[@bib3],[@bib12]\\]. The higher CRP level, neutrophilia and lymphopenia in our severe patients suggested an inflammatory profile that appears to be associated with lung consolidations and subsequent worsening of their respiratory condition \\[[@bib13]\\]. Consolidations are associated with poor outcome, as previously described \\[[@bib14]\\]. Patients with lung involvement \\>50% were significantly more often diagnosed with pulmonary embolism. This higher risk of thrombosis in patients with severe COVID-19 has been reported, although the pathophysiology remains unclear and may possibly involve several mechanisms \\[[@bib15]\\].\n\nThe timing between the onset of symptoms and the performance of CT was lower for patients with lung involvement \u226425%, indicating that these patients could have presented at an earlier disease stage. However, this difference was only 1\u00a0day, raising a question about its clinical relevance. Fourteen patients (2.4%) had no lesions on the initial chest CT, and among these 10/14 had symptoms for \u22643\u00a0days, as has previously been reported \\[[@bib1],[@bib2]\\]. None of these patients died or was admitted to the ICU, suggesting that normal CT at the time of hospital admission could predict a good prognosis.\n\nIn conclusion, in addition to its diagnostic value, chest CT could predict severe COVID-19 pneumonia as visual quantification of the lesions appears to be associated with early prognosis. Whether this strategy should be systematically implemented remains to be evaluated in further studies.\n\nAuthor contributions {#sec5}\n====================\n\nAll authors have made substantial contributions to this work and have approved the final manuscript. Concept and design: YR, CK, FD, MO and AL. Acquisition and interpretation of imaging data: MO and AL. Collection of clinical data: YR, CK, FD, YH, NL, VG, PB and SK. Virological analysis: MS. Analysis and interpretation of clinical data: YR, CK and FD. Statistical analysis: YR, TF, FD and VG. Writing of the original draft: YR and FD.\n\nTransparency declaration {#sec6}\n========================\n\nThe authors report no conflicts of interest. MO reports personal fees from 10.13039/100015650Canon Medical Systems, outside the submitted work. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\n\nAppendix A. Supplementary data {#appsec1}\n==============================\n\nThe following are the Supplementary data to this article:Multimedia component 1Multimedia component 1 figs1figs1 figs2figs2\n\nSupplementary data to this article can be found online at <https://doi.org/10.1016/j.cmi.2020.07.030>.\n\n[^1]: Data are given in *n* (%) or median\u00a0\u00b1\u00a0interquartile range, otherwise specified. The inferential analysis for the categorical data was performed using the \u03c7^2^ test or Fisher\\'s exact test (2\u00a0\u00d7\u00a03 comparison), as per the theoretical size of the samples. Continuous data were compared using a non-parametric test (Kruskal--Wallis test).\n\n[^2]: COVID-19, coronavirus disease 2019; CT, computed tomography; SD, standard deviation; SpO~2~, peripheral oxygen saturation.\n\n[^3]: Only 10/29 pulmonary embolisms were diagnosed at admission.\n\n[^4]: Defined as intensive care unit admission or death.\n"}
+{"text": "The authors confirm that all data underlying the findings are fully available without restriction. All rawdata and historic data are supplied as electronic supplements.\n\nIntroduction {#s1}\n============\n\nPresent day atmospheric carbon dioxide (CO~2~) concentrations are now over 30% higher than the maximum observed in the previous 2 million years [@pone.0109092-Hnisch1]. Approximately 28% of this additional CO~2~ is absorbed by the world\\'s oceans [@pone.0109092-Tyrrell1], [@pone.0109092-Khatiwala1], leading to lower seawater pH (Ocean Acidification; OA) with reduced carbonate ion concentrations (CO~3~ ^2\u2212^) and a reduced saturation state (\u03a9) of calcium carbonate minerals (CaCO~3~). Surface seawater pH has decreased by 0.1 units since pre-industrial times and is predicted to fall by a further 0.3**--**0.5 units in the next 100 years [@pone.0109092-Caldeira1]. Large-scale spatial and temporal variations (seasonal, inter-annual) in surface seawater CO~2~ concentrations are known to be caused by biogeochemical and air-sea exchange processes. Knowledge of this variability is critical to understand the current state of the carbon cycle and to predict how the ocean will react to future increases in atmospheric CO~2~ concentration. A recent review indicated that the partial pressure of CO~2~ (*p*CO~2~) in coral reef waters is increasing more rapidly than in the atmosphere, most likely due to other anthropogenic impacts on water quality [@pone.0109092-Cyronak1]. Coral reefs in tropical and subtropical regions contribute to the ocean carbon cycle through the processes of photosynthesis, respiration, CaCO~3~ production and dissolution [@pone.0109092-Barnes1], [@pone.0109092-Barnes2]. Coral reef ecosystems are vulnerable to OA and climate change induced ocean warming [@pone.0109092-HoeghGuldberg1] with a range of effects on the ecosystem and associated biota e.g., [@pone.0109092-Fabricius1]--[@pone.0109092-Uthicke1]. In particular, increases in oceanic CO~2~ will reduce the aragonite saturation state (\u03a9~ar~), which decreases the ability of many coral species to produce their carbonate skeletons [@pone.0109092-Anthony1], [@pone.0109092-Death1], [@pone.0109092-Reynaud1]. As a consequence, future coral reefs may exhibit net-carbonate dissolution as opposed to the net-accretion witnessed today [@pone.0109092-Albright1].\n\nBecause many coral reefs are net-autotrophic, these reefs may have an increased buffering capacity towards OA [@pone.0109092-Anthony2], [@pone.0109092-Kleypas1]. Shallow reef flat areas are dominated by respiratory processes at night (increasing CO~2~ in the water) and autotrophic processes during the day (decreasing CO~2~ and thus increasing pH). This can lead to considerable fluctuations of pH, \u03a9~ar~ and *p*CO~2~ [@pone.0109092-Albright1], [@pone.0109092-Shaw1], [@pone.0109092-Uthicke2].\n\nThe Great Barrier Reef (GBR), situated on the NE Australia continental shelf between 9 and 24\u00b0S, is the largest contiguous coral reef system in the world. The GBR contains approximately 3,700 individual coral reefs on a shallow shelf with an area close to 250,000 km^2^. Reefs occupy approximately 10% of the shelf area, while most of the remaining shelf is covered with carbonate sediments. The majority of the coral reefs are located on the outer half of the shelf, which is primarily under oceanic influence; however, approximately 20% of reefs lie within 10 km of the coast and are under direct terrestrial influence from freshwater, sediment, nutrient and organic carbon runoff.\n\nResearch to date on reef calcification and inorganic carbon dynamics within the GBR system has largely focused upon on-reef processes on mid- and outer-shelf reefs [@pone.0109092-Barnes1], [@pone.0109092-Barnes2], [@pone.0109092-Albright1], [@pone.0109092-Shaw2]. Relatively little work has been done on the shelf-scale dynamics of inorganic carbon in the GBR system [@pone.0109092-Suzuki1], [@pone.0109092-Kawahata1] and almost no consideration has been given to the many inshore reefs close to the coast that are under the greatest threat from increases in runoff of sediment, nutrients and pesticides [@pone.0109092-Fabricius2]--[@pone.0109092-Fabricius3]. The ratio of primary productivity and respiration (P/R) of inshore reefs are often lower than on reefs further from the coast due to decreased light availability associated with greater turbidity inshore [@pone.0109092-Fabricius4]--[@pone.0109092-Uthicke4]. Because of this, inshore reefs may be less able to buffer rising dissolved inorganic carbon (DIC) by photosynthesis.\n\nHere, we present broad-scale carbon chemistry data from inshore reefs of the GBR, collected six times over two years covering a comprehensive latitudinal range. We tested if there were any persistent regional and seasonal differences between inorganic carbon system parameters in the coastal waters of the GBR. In addition, we compared the carbon chemistry on inshore reefs to a smaller sample set from mid- and outer-shelf reefs and to historical data collected 18 and 30 years ago.\n\nMaterials and Methods {#s2}\n=====================\n\nSampling design {#s2a}\n---------------\n\nAll work described was covered under a permit obtained from the Great Barrier Reef Marine Park Authority (G12/35236.1).\n\nWater sampling for inshore chemical characteristics was carried out at 14 nearshore fringing reefs at islands between 16 and 23\u00b0 S ([Fig. 1](#pone-0109092-g001){ref-type=\"fig\"}). Twelve of the 14 core sites are within 15 km of the mainland and all are directly affected on a seasonal or episodic basis by terrestrial runoff. Sampling at the inshore core reef sites (Visits, n\u200a=\u200a6) was conducted at four-monthly intervals over two years (September 2011--June 2013) in the late dry season (September--October), wet season (February) and early dry season (June). The GBR region has a monsoonal climate with most (ca. 60**--**80%) rainfall falling in the January to March period. All samples were collected during the day time. In order to test if values differed between times of the day all samples were grouped into four time brackets (0600**--**0900, 0901**--**1200, 1201**--**1500, and 1501**--**1800) and an overall one factor analysis of variance (ANOVA) was conducted to test if average Total Alkalinity (TA) and dissolved inorganic carbon (DIC) were different between sampling times. This analysis illustrated that there was no significant difference in TA (F~1,\\ 165~\u200a=\u200a0.55, p\u200a=\u200a0.6462) or DIC (F~1,\\ 165~\u200a=\u200a1.41, p\u200a=\u200a0.2418) values between the four time brackets. The majority (\u223c65%) of the samples were collected between 0900 and 1500. We therefore concluded that the time of sampling did not bias our spatial or long-term temporal comparisons.\n\n![Map showing the sampling stations (\u2022) where surface samples were collected during cruises aboard R/V *Cape Ferguson* over the period September 2011 to November 2013.\\\nInshore monitoring stations, from north to south, are: Snapper Island, Fitzroy Island, High Island, Frankland Island, Dunk Island, Pelorus Island, Pandora Reef, Geoffrey Bay, Double Cone Island, Daydream Island, Pine Island, Humpy and Halfway Islands, Pelican Island, Barren Island, with separate colours representing separate regions as indicated in the legend. Yellow dots indicate mid- and outer-shelf reefs, from north to south: Mantis Reef, Wreck Bay, Fairway Channel, Shipping Channel, Tydeman Reef, Lizard Island, Arlington Reef, Twin Cays, Elusive Reef, Swains Reefs, Inner Swains.](pone.0109092.g001){#pone-0109092-g001}\n\nTo provide comparison with mid- and outer-shelf reef water we collected additional samples in two GBR regions (the Northern section and the Swains region) on five occasions. Samples from the Northern section were collected in November 2011 (Wreck Bay, Tydeman Reef, and Arlington Reef), June 2012 (Mantis Reef \\[7 samples\\] and Lizard Island), and November 2013 (Tydeman Reef and Fairway Channel). Samples from the Swains region were collected in April 2012 (Elusive Reef, Twin Cays Mooring Site, Swains Reefs, and Inner Swains) and September 2013 (Elusive Reef and Swains Reefs). All locations are shown in [Fig. 1](#pone-0109092-g001){ref-type=\"fig\"}. Water masses at the outer-shelf sites are primarily influenced by mixing with the oceanic waters of the Coral Sea. For most of its length, the outer-shelf reef matrix is separated from the mainland (and inshore reefs) by an open water body known as the GBR lagoon. South of 15\u00b0 S, the dominant (non-tidal) water movement on the outer-shelf is to the south under the forcing of the geostrophic pressure gradient of the East Australian Current. Northward surface flows on the shelf may occur during periods of strong SE trade winds. On the inner-shelf, this northward flow driven by SE trade winds is stronger and more persistent. To the north of 15\u00b0 S, shelf flows are primarily wind-driven.\n\nSample collection and analysis {#s2b}\n------------------------------\n\nWater samples for analysis of TA and DIC were collected at the 14 core reefs in conjunction with a range of standard oceanographic (temperature and salinity) and water quality (nitrate/nitrite (NO~3~ ^\u2212^/NO~2~ ^\u2212^), ammonium (NH~4~ ^+^), phosphate (PO~4~ ^3\u2212^), and chlorophyll *a* (Chl *a*)) parameters. The latter parameters are only summarised here to provide a background on the biogeochemical setting of the sites; a more detailed description of these parameters is given elsewhere [@pone.0109092-Schaffelke1]. At each of the inshore locations, surface (\u223c1 m water depth) and near-bottom (average depth 9.4 m, 1 SD\u200a=\u200a3.1 m) water samples were collected from the R/V *Cape Ferguson* using 10 L Niskin bottles. These open water stations were 0.3**--**2 km from the neighbouring reef. In addition, divers collected water near-bottom (average depth 6.5 m) on the reef slopes of the coral reef at each inshore site.\n\nDuplicate aliquots (250 ml) were carefully drawn from the Niskin bottles for TA and DIC analysis, taking care to avoid bubble formation and minimize headspace. Samples were fixed with 125 \u00b5l of saturated HgCl~2~. Samples for TA and DIC were analysed using a VINDTA 3C titrator (Marianda, Germany) at the Australian Institute of Marine Science (AIMS). Alkalinity was determined by acid titration [@pone.0109092-Dickson1] and DIC by acidification and coulometric detection (UIC 5105 Coulometer) of the evolved CO~2~. The VINDTA titrator was calibrated with Certified Reference seawaters (A. G. Dickson, Scripps Institute of Oceanography, Dixon, Batch 106). Raw-data for all TA and DIC samples are given in Table S1 in [File S1](#pone.0109092.s001){ref-type=\"supplementary-material\"}.\n\nHistorical data {#s2c}\n---------------\n\nOur data collected in 2011 to 2013 were compared to data collected by Kawahata et al. [@pone.0109092-Kawahata1] which were only collected during the dry season (May 1996) and on locations further from the reef than those obtained here. In addition, we obtained historical data from 1982/83 from the AIMS data archive. These were collected by Dr Dave Barnes and colleagues following detailed methods described in [@pone.0109092-Barnes1], [@pone.0109092-Barnes2]. The carbon chemistry calculations from the 1982/83 dataset were based on precision measurements of pH and TA, and those from 1996 on CO~2~ measurements in equilibrator chambers and TA measurements. Although individual methods may vary in their measuring certainty, all are still accepted methods [@pone.0109092-Riebesell1] and there is therefore no reason to assume that these data are not comparable. See Table S2 in [File S1](#pone.0109092.s001){ref-type=\"supplementary-material\"} for more detailed descriptions and a transcript of the raw data used.\n\nData analysis {#s2d}\n-------------\n\nCarbon chemistry parameters (the partial pressure of CO~2~ \\[*p*CO~2~\\], pH on the total scale \\[pH~Total~\\]), and the saturation state for aragonite \\[\u03a9~ar~\\]) were calculated using the Excel macro CO2SYS [@pone.0109092-Pierrot1], taking salinity and temperature into consideration.\n\nSalinity normalization to a constant salinity is commonly used to correct for differences between source water masses and local effects from evaporation and precipitation on the marine carbon chemistry [@pone.0109092-Friis1], [@pone.0109092-Robbins1]. We used the method proposed by Friis et al. [@pone.0109092-Friis1] for TA and DIC, using the annual average salinity during the sampling period of 34.5 and a non-zero freshwater end member \\[TA~S\u200a=\u200a0~\u200a=\u200a309.13 \u00b5mol Kg^\u22121^; DIC~S\u200a=\u200a0~\u200a=\u200a288.48 \u00b5mol Kg ^\u22121^\\].\n\nTo separate the seasonal effect of biological processes (B) and temperature (T) on the *p*CO~2~ dynamics, we used the method developed by Takahashi et al. [@pone.0109092-Takahashi1] and calculated the effect as:where T is temperature (\u00b0C) and the subscripts \"mean\" and \"obs\" indicate the annual mean temperatures (reported in [Table 1](#pone-0109092-t001){ref-type=\"table\"}) or *p*CO~2~ for each region and the observed values, respectively. The relative importance of each effect is expressed as the ratio between *p*CO~2,\\ Temp~ and *p*CO~2,\\ bio~ (T/B). A ratio \\>1 suggests a dominance of temperature effects over biological processes on the *p*CO~2~ dynamics.\n\n10.1371/journal.pone.0109092.t001\n\n###### Biological, chemical and physical properties of water samples at the time of collection between 2011 and 2013.\n\n![](pone.0109092.t001){#pone-0109092-t001-1}\n\n                                  Sal.   Temp.    Chl *a*    NH~4~ ^+^   NO~3~ ^\u2212^/NO~2~ ^\u2212^   PO~4~ ^3\u2212^\n  ------------- ----------- ---- ------ ------- ----------- ----------- --------------------- ------------\n  Wet-tropics    Early dry   20   33.7   23.2    0.33\u00b10.11   0.07\u00b10.04        0.17\u00b10.14        0.11\u00b10.05\n                 Late dry    20   35.2   25.3    0.28\u00b10.02   0.06\u00b10.05        0.20\u00b10.19        0.11\u00b10.02\n                    Wet      20   33.7   29.9    0.38\u00b10.17   0.11\u00b10.10        0.18\u00b10.17        0.05\u00b10.02\n                 All year         34.2   26.1    0.33\u00b10.15   0.08\u00b10.07        0.18\u00b10.16        0.11\u00b10.04\n                 Amplitude        3.6     8.0      0.75        0.37             0.80              0.22\n  Burdekin       Early dry   12   34.6   22.5    0.25\u00b10.06   0.07\u00b10.07        0.14\u00b10.09        0.11\u00b10.02\n                 Late dry    12   35.3   24.7    0.48\u00b10.27   0.11\u00b10.12        0.24\u00b10.22        0.11\u00b10.03\n                    Wet      12   33.9   30.0    0.34\u00b10.13   0.11\u00b10.07        0.16\u00b10.18        0.05\u00b10.02\n                 All year         34.6   25.7    0.38\u00b10.22   0.10\u00b10.09        0.18\u00b10.17        0.09\u00b10.04\n                 Amplitude        3.2    10.0      0.75        0.37             0.76              0.16\n  Whitsundays    Early dry   12   34.3   22.2    0.45\u00b10.16   0.10\u00b10.05        0.25\u00b10.13        0.15\u00b10.04\n                 Late dry    12   35.2   23.0    0.41\u00b10.17   0.06\u00b10.04        0.11\u00b10.05        0.14\u00b10.03\n                    Wet      12   34.8   29.0    0.58\u00b10.11   0.19\u00b10.15        0.26\u00b10.22        0.09\u00b10.03\n                 All year         34.8   24.7    0.48\u00b10.17   0.13\u00b10.12        0.21\u00b10.16        0.12\u00b10.04\n                 Amplitude        1.6     8.4      0.59        0.41             0.56              0.16\n  Fitzroy        Early dry   12   35.1   21.0    0.46\u00b10.28   0.03\u00b10.02        0.11\u00b10.06        0.10\u00b10.06\n                 Late dry    12   35.6   22.0    0.36\u00b10.42   0.05\u00b10.05        0.16\u00b10.13        0.13\u00b10.05\n                    Wet      12   33.5   28.0    0.59\u00b10.18   0.11\u00b10.11        0.18\u00b10.22        0.11\u00b10.07\n                 All year         34.7   23.7    0.47\u00b10.31   0.07\u00b10.08        0.15\u00b10.15        0.12\u00b10.06\n                 Amplitude        4.6    10.4      1.17        0.28             0.64              0.28\n  Offshore       Early dry   18   34.9   25.8    0.39\u00b10.18   0.03\u00b10.05        0.12\u00b10.07        0.09\u00b10.04\n                 Late dry    23   35.4   26.7    0.29\u00b10.23   0.03\u00b10.05        0.13\u00b10.06        0.06\u00b10.02\n                    Wet           n.d     n.d       n.d         n.d              n.d              n.d\n                 All year         n.d     n.d       n.d         n.d              n.d              n.d\n                 Amplitude        1.7     7.2      1.02        0.18             0.31              0.21\n\nAverage values for salinity (Sal.), temperature (Temp.), chlorophyll *a* (Chl *a*), ammonium (NH~4~ ^+^), Nitrate/Nitrite (NO~3~ ^--^/NO~2~ ^--^) and phosphate (PO~4~ ^3\u2212^) and their amplitude (maximum minus minimum level) are shown. Standard deviations are shown for chlorophyll *a* and nutrient data; N: number of samples used to calculate the average.\n\nSalinity-normalized data for TA (TA~S~) and DIC (DIC~S~) were also used to create TA~S~ vs DIC~S~ plots to examine the impact of calcification on the annual changes in the carbon system. This approach follows on from the assumption that net primary production of one mole of organic C reduces DIC by one mole, while calcification reduces TA by two moles and DIC by one mole for each mole of CaCO~3~ precipitated [@pone.0109092-Suzuki2]. In systems where calcification is dominating, there should therefore be a linear relationship between DIC and TA with a slope approaching 2.0. The slope of this relationship can be used to calculate the net ecosystem production (NEP) to net ecosystem calcification (NEC) ratio, which is given by the function: (2/slope)-1 [@pone.0109092-Suzuki2].\n\nWe used mixed model ANOVAs to examine sources of variation in observed levels of TA, DIC, *p*CO~2~, pH~Total~, and \u03a9~ar~. The fixed main factor \"Region\" was used to test for differences between the five regions (Wet-tropics, Burdekin, Whitsundays, Fitzroy and offshore). We considered replicate \"Islands\" as a random nested factor within Regions. To evaluate if samples taken from the reef slopes were different from those collected from the research vessel \"Location\" was included in the model as a second fixed factor. The main factor \"Visit\" tested for differences between the six sample periods. With the exception of pH (that is already on a log-scale) all data were log-transformed prior to analysis. Boxplot and residual plots indicated no deviation from ANOVA assumptions for the transformed variables. We tested for correlations between several parameters using Pearson\\'s product moment correlations. To further investigate differences between seasons and inshore *vs* offshore reefs, we conducted a principal component analysis (PCA) with TA, DIC, *p*CO~2~ and \u03a9~ar~ as carbon chemistry parameters. pH was omitted from this analysis as it is highly correlated with *p*CO~2~. For the comparison of the historical data, we calculated a Bayesian 95% confidence interval for the rate of change based on averages and standard deviations from present day and historical data. This was achieved using Markov Chain Monte Carlo sampling. After a burn-in period of 2000 steps, 5000 steps were sampled and three parallel chains run. These models were also used to calculate the probabilities that one rate is larger than another. All statistical analyses were conducted using NCSS [@pone.0109092-Hintze1] or the R environment [@pone.0109092-RDevelopmentCoreTeam1].\n\nResults {#s3}\n=======\n\nEnvironmental conditions {#s3a}\n------------------------\n\nSalinity at the core reef sites ranged from 31.4 to 36.0, being highest at the end of the dry season (September 2011, October 2012) and lowest during the two wet seasons ([Table 1](#pone-0109092-t001){ref-type=\"table\"}). This pattern is most distinct in the two northern regions (Wet-tropics and Burdekin) that had the highest levels of terrestrial runoff. During most seasons, salinities in the southern-most region (Fitzroy) were slightly higher compared to other regions. The one exception was during the 2013 wet season when major flooding in the Fitzroy River catchment resulted in low (*ca*. 31) and variable salinities. Temperatures were highest during the summer wet season and lowest in June and declined from the north to the south, regardless of season ([Table 1](#pone-0109092-t001){ref-type=\"table\"}).\n\nMeasured Chl *a* concentrations varied between 0.11 and 1.3 \u00b5g l^\u22121^. Inorganic nitrogen and phosphorus concentrations were of the order of 0.1 \u00b5mol kg^\u22121^, with elevated NH~4~ ^+^ and NO~3~ ^\u2212^/NO~2~ ^\u2212^ concentrations during the wet season. The highest wet season NO~3~ ^\u2212^/NO~2~ ^\u2212^ levels were generally found in the more freshwater influenced geographic regions (Wet-tropics and Fitzroy). In contrast, the highest PO~4~ ^3\u2212^concentrations were measured during the dry season ([Table 1](#pone-0109092-t001){ref-type=\"table\"}).\n\nInorganic carbon dynamics {#s3b}\n-------------------------\n\nWith few exceptions, there was no appreciable or consistent difference between data derived from samples collected from the research vessel (near-surface, near-bottom) and by diver on the adjacent coral reef slope (diver collected; [Fig. 2](#pone-0109092-g002){ref-type=\"fig\"}). Exceptions were observed during the wet seasons, where the surface sample differed from the two near-bottom samples (e.g. Snapper Island February 2012, Dunk Island February 2013). These samples were characterized as having lower salinity than the contemporaneous near-bottom samples, so the most likely cause was either recent rainfall or freshwater runoff affecting these sites. To facilitate interpretation, we restricted further analyses and statistical tests to samples from the reef slope and the surface samples from the research vessel\\'s anchorage and averaged replicate sub-samples.\n\n![Total Alkalinity (TA) and dissolved inorganic carbon (DIC) data for each island station, sorted from north to south, during each Visit and for each of the three Locations.\\\nRegions are colour coded as per [Fig. 1](#pone-0109092-g001){ref-type=\"fig\"}. Note that duplicates are plotted but in most cases cannot be distinguished because of the low between-duplicate variance.](pone.0109092.g002){#pone-0109092-g002}\n\nTotal Alkalinity values ranged between 2069 and 2315 \u00b5mol kg^\u22121^. Higher TA levels were generally measured during the late dry season (September--October) and in the regions more influenced by freshwater (Wet-tropics and Fitzroy). There was a general increase from north to south, with the exception of low values in the Fitzroy region in February 2013 ([Fig. 2](#pone-0109092-g002){ref-type=\"fig\"}). The ANOVA for inshore TA values showed significant effects of geographic regions and amongst visits ([Table 2](#pone-0109092-t002){ref-type=\"table\"}) with a significant interaction between these factors, indicating that regional trends were dependent on the season sampled ([Fig. 3](#pone-0109092-g003){ref-type=\"fig\"}). In general, TA values were closely correlated with salinity (r^2^\u200a=\u200a0.94, p\\<0.0001; [Fig. 2](#pone-0109092-g002){ref-type=\"fig\"}). Thus, salinity normalization (TA~S~) removed a large part of the seasonal variability in TA at the inshore stations; particularly in the Wet-tropics (198 \u00b5mol kg^\u22121^) and Fitzroy (106 \u00b5mol kg^\u22121^) regions where freshwater inputs from rivers were largest ([Table 3](#pone-0109092-t003){ref-type=\"table\"}). In contrast, no salinity related variation was found for the outer shelf reefs ([Table 3](#pone-0109092-t003){ref-type=\"table\"}).\n\n![Measured Temperature, Salinity, Total Alkalinity (TA) and dissolved inorganic carbon (DIC) on 14 inshore reefs during six research trips in four geographical regions along the length of the Great Barrier Reef.\\\nThe box denotes the inter-quartile range, whiskers denote 1.5\u00d7 the inter-quartile range, the black line indicates the mean, and circles are outliers\\>1.5\u00d7 the inter-quartile range.](pone.0109092.g003){#pone-0109092-g003}\n\n10.1371/journal.pone.0109092.t002\n\n###### Mixed model ANOVA for measured parameters Total Alkalinity (TA) and dissolved inorganic carbon (DIC).\n\n![](pone.0109092.t002){#pone-0109092-t002-2}\n\n                                          TA                                    DIC    \n  ---------------- ----- -------------- ------- -------------- -------------- -------- --------------\n  **R**egion         3    1.01 10^\u221203^   22.26    **0.0001**    1.02 10^\u221203^   21.64     **0.0001**\n  **I**sland (R)    10    4.54 10^\u221205^   2.28       0.0179      4.72 10^\u221205^    2.26       0.0189\n  **L**ocation       1    2.12 10^\u221205^   1.06       0.3045      6.38 10^\u221205^    3.06       0.0830\n  R x L              3    2.62 10^\u221205^   1.32       0.2724      2.74 10^\u221205^    1.32       0.2732\n  **V**isit          5    1.63 10^\u221203^   81.85   **\\<0.0001**   2.19 10^\u221203^   105.11   **\\<0.0001**\n  R x V             15    2.98 10^\u221204^   14.97   **\\<0.0001**   2.45 10^\u221204^   11.77    **\\<0.0001**\n  L x V              5    2.22 10^\u221205^   1.11       0.3568      2.05 10^\u221205^    0.98       0.4306\n  R x L x V         15    8.33 10^\u221206^   0.42       0.9709      8.46 10^\u221206^    0.41       0.9748\n  Residual          110   1.99 10^\u221205^                          2.09 10^\u221205^           \n\nThe model tests for differences between four Regions of the Great Barrier Reef (factor \"Region\"), the vessel\\'s anchorage (0 m) and dive site (\"Location\") and six visits over two years of monitoring (\"Visit\"). \"Island\" is nested as a random factor in \"Region\". Significant (p\\<0.05) fixed factors are highlighted in bold. All data are log-transformed for analysis. DF: degrees of freedom; MS: mean square; F: F-value for F test.\n\n10.1371/journal.pone.0109092.t003\n\n###### A summary of carbon chemistry of water samples collected during 2011 to 2013 in the Great Barrier Reef region.\n\n![](pone.0109092.t003){#pone-0109092-t003-3}\n\n                                    TA       TA~S~      DIC     DIC~S~    *p*CO~2~   *p*CO~2,\\ Bio~   *p*CO~2,\\ Temp~\n  ------------- ----------- ---- --------- --------- --------- --------- ---------- ---------------- -----------------\n  Wet-tropics    Early dry   20   2196\u00b140   2244\u00b16    1936\u00b132   1978\u00b112    398\u00b124        478\u00b126            380\u00b18\n                 Late dry    20   2302\u00b110   2262\u00b19    2021\u00b120   1986\u00b119    445\u00b135        490\u00b137           415\u00b112\n                    Wet      20   2195\u00b140   2239\u00b18    1916\u00b130   1955\u00b111    470\u00b129        401\u00b127            513\u00b15\n                 All year         2231\u00b160   2249\u00b113   1957\u00b153   1973\u00b120    437\u00b142        437\u00b139           441\u00b154\n                 Amplitude          269       71        253       123       198           220               150\n  Burdekin       Early dry   12   2262\u00b19    2256\u00b18    1991\u00b112   1986\u00b119    399\u00b125        489\u00b129           367\u00b112\n                 Late dry    12   2320\u00b111   2274\u00b19    2038\u00b123   1998\u00b120    442\u00b135        493\u00b137            403\u00b16\n                    Wet      12   2234\u00b115   2270\u00b131   1941\u00b117   1972\u00b131    458\u00b114        383\u00b115           518\u00b115\n                 All year         2272\u00b138   2266\u00b120   1990\u00b144   1985\u00b126    433\u00b136        434\u00b144           437\u00b161\n                 Amplitude          131       100       164       106       134           162               186\n  Whitsundays    Early dry   12   2248\u00b123   2258\u00b18    1991\u00b116   2000\u00b113    414\u00b126        487\u00b123           367\u00b112\n                 Late dry    12   2300\u00b114   2261\u00b110   2022\u00b111   1988\u00b113    411\u00b114        467\u00b115           380\u00b113\n                    Wet      12   2271\u00b118   2252\u00b17    1971\u00b17    1954\u00b15     451\u00b18         377\u00b115           509\u00b115\n                 All year         2273\u00b128   2257\u00b19    1995\u00b125   1981\u00b122    425\u00b125        427\u00b10            429\u00b159\n                 Amplitude          107       34        80        68         90           162               157\n  Fitzroy        Early dry   12   2303\u00b14    2268\u00b127   2023\u00b113   1993\u00b133    376\u00b123        446\u00b122           363\u00b120\n                 Late dry    12   2332\u00b115   2271\u00b114   2052\u00b121   1998\u00b125    406\u00b131        461\u00b129            379\u00b16\n                    Wet      12   2214\u00b170   2271\u00b140   1940\u00b162   1990\u00b135    454\u00b139        377\u00b124           496\u00b119\n                 All year         2283\u00b165   2270\u00b128   2005\u00b161   1994\u00b131    412\u00b145        411\u00b134           416\u00b160\n                 Amplitude          257       151       242       155       175           145               187\n  Offshore       Early dry   18   2290\u00b116   2267\u00b119   1969\u00b111   1950\u00b114    368\u00b115        377\u00b122           374\u00b124\n                 Late dry    23   2308\u00b114   2257\u00b112   1986\u00b112   1943\u00b18     392\u00b127        385\u00b110           389\u00b124\n                    Wet             n.d       n.d       n.d       n.d       n.d           n.d               n.d\n                 All year           n.d       n.d       n.d       n.d       n.d           n.d               n.d\n                 Amplitude          79        86        51        55        103            73               111\n\nThe average values (\u00b1 standard deviation) and amplitude (maximum minus minimum level) for Total Alkalinity (TA), dissolved inorganic carbon (DIC) and salinity normalized TA and DIC (TA~S~, DIC~S~) values are shown, together with partial pressure of carbon dioxide (*p*CO~2~) and effect of biological processes (*p*CO~2,Bio~) and temperature (*p*CO~2,\\ Temp~) on *p*CO~2~ dynamics. N: number of samples used to calculate the average.\n\nRegional averages of DIC concentrations at inshore reefs ranged between 1930 and 2050 \u00b5mol kg^\u22121^. The within-region variation in DIC concentrations was between 340 and 554 \u00b5mol kg^\u22121^, with the highest concentrations and variation measured in September 2011 and October 2012. Again, there was a distinct north to south DIC gradient during most sampling campaigns ([Table 3](#pone-0109092-t003){ref-type=\"table\"}; [Fig. 3](#pone-0109092-g003){ref-type=\"fig\"}). The DIC ANOVA was similar to that for TA, with a strong interaction between Region and Visit ([Table 2](#pone-0109092-t002){ref-type=\"table\"}). As evident in the raw-data plots ([Fig. 2](#pone-0109092-g002){ref-type=\"fig\"}), there was no effect of Location (surface water at the anchorage and the reef slope sites) or interactions of this factor with other fixed factors for either TA or DIC ([Table 2](#pone-0109092-t002){ref-type=\"table\"}). Dissolved inorganic carbon was also strongly correlated with salinity (r^2^\u200a=\u200a0.88, p\\<0.0001) and TA (r^2^\u200a=\u200a0.95, p\\<0.0001). Salinity normalization of the DIC data (DIC~S~) removed a large part of the seasonal variation, especially in the Wet-tropics (130 \u00b5mol kg^\u22121^) and Fitzroy (87 \u00b5mol kg^\u22121^) region, leaving a residual seasonal variability of 123 and 155 \u00b5mol kg^\u22121^ in those areas ([Table 3](#pone-0109092-t003){ref-type=\"table\"}). The linear relationships between DIC~S~ and TA~S~ had the steepest slopes in the offshore and Fitzroy regions and flattest in the Whitsunday region ([Fig. 4](#pone-0109092-g004){ref-type=\"fig\"}). In all cases, the slope value was \\<2, which resulted in NEP/NEC ratios varying between 7.3 (Whitsundays) and 0.4 (Offshore). This suggests that the importance of calcification in controlling the carbon cycle varies regionally with decreasing importance and larger influence of primary production/respiration in the sequence Offshore (NEP/NEC\u200a=\u200a0.4) \\> Fitzroy (1.2) \\> Burdekin (1.7) \\> Wet-tropics (3.0) \\> Whitsundays (7.3).\n\n![Relationships between salinity normalized dissolved inorganic carbon (DIC~S~) and salinity normalized Total Alkalinity (TA~S~) for the a) Wet-tropics, b) Burdekin, c) Whitsundays, d) Fitzroy and e) offshore shelf regions of the Great Barrier Reef.\\\nThe regression lines and corresponding equations were obtained using model II linear regression. r^2^: coefficient of determination; p: significance level.](pone.0109092.g004){#pone-0109092-g004}\n\nBased on the observed patterns of variability for the measured parameters (DIC, TA), the derived parameters *p*CO~2~, pH and \u03a9~ar~ also exhibited significant interactions between Region and Visit ([Table 4](#pone-0109092-t004){ref-type=\"table\"}). *p*CO~2~ reached concentrations between 340 and 554 \u00b5atm, with a decline from north to south during three of the visits ([Table 3](#pone-0109092-t003){ref-type=\"table\"}; [Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}). The most distinct temporal pattern in the *p*CO~2~ data ([Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}) was an elevation during the wet seasons (total average February 2012: 460 \u00b5atm, 1 SD\u200a=\u200a19 \u00b5atm; February 2013: 460 \u00b5atm, 1 SD\u200a=\u200a33 \u00b5atm), compared to the early dry seasons (June 2012: 383 \u00b5atm, 1 SD\u200a=\u200a21; June 2013: 410 \u00b5atm, 1 SD\u200a=\u200a25 \u00b5atm) and late dry seasons (September 2011: 416 \u00b5atm, 1 SD\u200a=\u200a35 \u00b5atm; October 2012: 440 \u00b5atm, 1 SD\u200a=\u200a29 \u00b5atm) ([Table 3](#pone-0109092-t003){ref-type=\"table\"}; [Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}). Dry season *p*CO~2~ concentrations were slightly elevated relative to present atmospheric values, whereas wet season concentrations were distinctively (\u223c15%) above atmospheric values. The seasonal fluctuation of *p*CO~2~ concentrations was highest in the Wet-tropics (198 \u00b5mol kg^\u22121^) and Fitzroy (175 \u00b5mol kg^\u22121^) regions ([Table 3](#pone-0109092-t003){ref-type=\"table\"}). Derived pH values varied significantly among trips and a Region x Visit interaction was also significant ([Table 4](#pone-0109092-t004){ref-type=\"table\"}). As expected, seasonal and spatial variations for pH were reversed compared to those of *p*CO~2~ ([Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}). Derived pH values were lowest during the wet seasons (February 2012: 7.97, SD\u200a=\u200a0.01; February 2013: 7.97, SD\u200a=\u200a0.03) and slightly higher in both the early dry seasons (June 2012: 8.04, 1 SD\u200a=\u200a0.02; June 2013: 8.02, 1 SD\u200a=\u200a0.02) and the late dry seasons (September 2011: 8.02, 1 SD\u200a=\u200a0.03, October 2012: 8.00, 1 SD\u200a=\u200a0.02). With one exception (February 2012), there was a slight increase in pH from north to south ([Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}).\n\n![Derived parameters at 14 inshore reefs during six research trips in four geographic regions along the length of the Great Barrier Reef.\\\n\u03a9~ar~\u200a=\u200aAragonite saturation state. The box denotes the inter-quartile range, whiskers denote 1.5\u00d7 the inter-quartile range, the black line indicates the mean, and circles are outliers \\>1.5\u00d7 the inter-quartile range.](pone.0109092.g005){#pone-0109092-g005}\n\n10.1371/journal.pone.0109092.t004\n\n###### Mixed model ANOVA for derived parameters partial pressure of carbon dioxide (*p*CO2), pH~\\[Total\\]~ and aragonite saturation state (\u03a9~ar~).\n\n![](pone.0109092.t004){#pone-0109092-t004-4}\n\n                                        *p*CO~2~                                pH~\\[total\\]~                                \u03a9~ar~  \n  ---------------- ----- ------------- ---------- -------------- ------------- --------------- -------------- ------------- ------- --------------\n  **R**egion         3    5.53 10^\u22123^     3.11        0.0757      6.97 10^\u22123^       6.29         **0.0114**    1.16 10^\u22123^   1.08       0.4001\n  **I**sland (R)    10    1.78 10^\u22123^     4.06        0.0001      1.11 10^\u22123^        3.7           0.0003      1.07 10^\u22123^   4.32      \\<0.0001\n  **L**ocation       1    2.18 10^\u22123^     4.97      **0.0278**    1.53 10^\u22123^        5.1         **0.0259**    4.98 10^\u22124^   2.02       0.1586\n  R x L              3    1.60 10^\u22124^     0.37        0.7784      6.39 10^\u22125^       0.21           0.8869      2.92 10^\u22125^   0.12       0.9494\n  **V**isit          5      0.0255       58.29     **\\<0.0001**   2.25 10^\u22122^       75.28       **\\<0.0001**   1.29 10^\u22122^   52.09   **\\<0.0001**\n  R x V             15    2.26 10^\u22123^     5.15     **\\<0.0001**   1.41 10^\u22123^       4.72        **\\<0.0001**   1.53 10^\u22123^   6.19    **\\<0.0001**\n  L x V              5    9.84 10^\u22125^     0.22        0.9512      5.35 10^\u22125^       0.18           0.9701      5.87 10^\u22125^   0.24       0.9453\n  R x L x V         15    1.82 10^\u22124^     0.41        0.9721      1.21 10^\u22124^        0.4           0.9753      9.76 10^\u22125^   0.39       0.9780\n  Residual          110   4.38 10^\u22124^                             2.99 10^\u22124^                                  2.47 10^\u22124^          \n\nThe model tests for differences between four Regions of the Great Barrier Reef (factor \"Region\"), the vessel\\'s anchorage (0 m) and dive site (\"Location\"), and six visits over two years of monitoring (\"Visit\"). \"Island\" is nested as a random factor in \"Region\". Significant (p\\<0.05) fixed factors are highlighted in bold. With the exception of pH, all data are log-transformed for analysis. DF: degrees of freedom; MS: mean square; F: F-value for F test.\n\nAragonite saturation state (\u03a9~ar~) varied between 2.6 and 3.8, with highly significant differences between sampling trips and a significant Region x Visit interaction ([Table 4](#pone-0109092-t004){ref-type=\"table\"}; [Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}). Average \u03a9~ar~ values in the wet seasons (February 2012: 3.39, 1 SD\u200a=\u200a0.15; February 2013: 3.25, 1 SD\u200a=\u200a0.16) were higher than in the early dry seasons (June 2012: 2.99, 1 SD\u200a=\u200a0.13; June 2013: 2.98, 1 SD\u200a=\u200a0.16) or in the late dry seasons (September 2011: 3.17, 1 SD\u200a=\u200a0.14; October 2012: 3.16, 1 SD\u200a=\u200a0.10). Thus, despite higher *p*CO~2~ and lower pH values during the summer wet season, \u03a9~ar~ were not reduced, but actually increased. This is likely due to higher water temperatures; resulting in lower aragonite solubility in the summer wet season.\n\nIn contrast to TA and DIC, the derived parameters *p*CO~2~ and pH exhibited small but significant differences between sampling locations at individual sites ([Table 4](#pone-0109092-t004){ref-type=\"table\"}). The average *p*CO~2~ for reef slope and adjacent open water samples were 432 (1 SD\u200a=\u200a42) and 424 (1 SD\u200a=\u200a36) \u00b5atm, respectively. The overall average pH for the reef slope sites (8.00, 1 SD\u200a=\u200a0.04) was slightly lower than the mean for the anchorage sites (8.01, 1 SD\u200a=\u200a0.04).\n\nWater samples taken at the mid - to outer-shelf reefs over the same study period were less variable than those collected on inshore reefs ([Table 3](#pone-0109092-t003){ref-type=\"table\"}) and the *p*CO~2~ was always closer to atmospheric equilibrium, with resulting higher pH values. The mid - to outer-shelf \u03a9~ar~ was also clearly higher than most inshore values. A principal component analysis of measured and derived parameters ([Fig. 6](#pone-0109092-g006){ref-type=\"fig\"}) separates inshore and mid -- to outer-shelf sites, with the main distinguishing factors being \u03a9~ar~ and *p*CO~2~. In addition, samples taken within the three defined 'seasons' (wet, early dry, late dry) clearly group together, with wet season samples distinguished by lower DIC and TA values.\n\n![Principal component analysis of carbonate chemistry data for four inshore regions of the Great Barrier Reef (GBR) over three different seasons.\\\nData are contrasted to those from the mid- and outer-shelf reefs of the GBR. Inshore data were pooled over regions per sampling Visit.](pone.0109092.g006){#pone-0109092-g006}\n\nIn order to determine whether temperature (T) or biological (B) processes (T/B ratio) primarily controlled *p*CO~2~ dynamics, we used the method proposed by Takahashi et al. [@pone.0109092-Takahashi1]. The T/B ratios showed that over an annual cycle, biological effects primarily control *p*CO~2~ dynamics in the Wet-tropics (T/B ratio\u200a=\u200a0.7). In the Whitsundays region, temperature and biology contributed equally (1.0), while in the Burdekin (1.2), Fitzroy (1.3) and at the offshore reefs (1.5) the *p*CO~2~ concentrations appear to be primarily controlled by temperature. Thus, for the largest part of the GBR, seasonal *p*CO~2~ changes are most likely controlled by temperature.\n\nDiscussion {#s4}\n==========\n\nLike other coastal areas [@pone.0109092-Duarte1], OA is part of a suite of factors that influence the coral reefs found on the GBR. We analyzed a two year data set collected over a large portion of the GBR to describe seasonal and broad-scale spatial changes in carbon chemistry and contrast those to outer reef areas. Overall, regional variability in carbon system parameters is relatively small. Within the inshore reefs, the largest amount of variation occurred seasonally. In addition, *p*CO~2~ was distinctly higher on inshore compared to offshore reefs.\n\n*p*CO~2~ in the GBR and other coastal waters is influenced by a number of processes, including thermodynamics; air-sea exchange; biological metabolism (photosynthesis, respiration, calcification); and freshwater inputs. Reasons for elevated *p*CO~2~ in the GBR inshore waters especially during the wet season are not fully resolved. In European estuaries, high DIC levels in freshwater can elevate *p*CO~2~ in inshore areas [@pone.0109092-Abril1]. Although salinity normalization removed a large part of the seasonal amplitude in DIC and TA fluctuations, there is no significant correlation between salinity and *p*CO~2~ in our dataset (r^2^\u200a=\u200a0.02, p\u200a=\u200a0.0940). Thus, it is unlikely that freshwater inflow resulted in the elevated *p*CO~2~ in the inshore GBR.\n\nIt is also possible that pCO~2~ increase is a consequence of higher calcification (benthic -- e.g., corals, foraminifera; or pelagic -- e.g., coccolithophorids), resulting from elevated temperatures during the wet seasons [@pone.0109092-Kawahata1], [@pone.0109092-Frankignoulle1]. It is difficult to judge the net effect of this on the areas studied, because it would also require more detailed knowledge on accompanying respiration (CO~2~ source) and primary production (CO~2~ sink). The expected slope of the linear regression between DIC~S~ and TA~S~ in systems where calcification is dominating is close to 2.0 e.g., [@pone.0109092-Kawahata1], [@pone.0109092-Suzuki2]. We found slopes \\<2 and high (1.2 to 7.3) NEP/NEC ratios for all the inshore regions investigated ([Fig. 5](#pone-0109092-g005){ref-type=\"fig\"}), suggesting that processes other than calcification (e.g., photosynthesis, respiration) are largely controlling the carbon cycle on inshore GBR reefs.\n\nChanging *p*CO~2~ outside the equilibrium may also be caused by thermodynamic effects. A seasonal increase at a similar range as observed here (albeit from a lower baseline) was observed near Lady Elliot Island in the southern GBR [@pone.0109092-Shaw2]. Based on the slope of the *p*CO~2~ -- temperature relationship, the authors of that study suggested thermodynamic effects as the most likely explanation for elevated *p*CO~2~ values in summer. There was a significant correlation between temperature and *p*CO~2~ in our dataset (r^2^\u200a=\u200a0.53, p\\<0.0001, *p*CO~2~\u200a=\u200a2.07% \\[1 SE\u200a=\u200a0.1%\\] \\* Temp (\u00b0C)+5.53), but our value was lower than that found at Lady Elliot Island \\[3.8 \\[1 SE\u200a=\u200a0.4%\\], 20\\], and only about 50% of the expected slope of 4**--**4.2% *p*CO~2~ [@pone.0109092-Takahashi2]. However, the regression clearly levels off at \\>26\u00b0C. If data \\>26\u00b0C were excluded, the slope of the regression (r^2^\u200a=\u200a0.42, p \\<0.0001, *p*CO~2~\u200a=\u200a3.6% \\[1 SE\u200a=\u200a0.4%\\] \\* Temp (\u00b0C)+5.19) is closer to the theoretical value for temperature controlled systems.\n\nWe applied the method proposed by Takahashi et al. [@pone.0109092-Takahashi1] to distinguish if seasonal changes in *p*CO~2~ concentrations were controlled by temperature or biological processes. The definition of \"biology effect\" applied includes biogeochemical processes (e.g., primary production, respiration, calcification), and other processes influencing the CO~2~ such as air-sea exchange and lateral vertical mixing processes [@pone.0109092-Thomas1]. The application of this method suggested a strong temperature control in the southern GBR, as was also proposed by Shaw and McNeil [@pone.0109092-Shaw2]. Northern GBR reefs were impacted more by biological processes. In addition, temporal difference in the controlling factor also existed, with stronger biological and temperature control during wet and dry seasons, respectively ([Table 3](#pone-0109092-t003){ref-type=\"table\"}).\n\nAnalysis and modelling of historical trends in coral reef *p*CO~2~ levels worldwide [@pone.0109092-Cyronak1] suggests that *p*CO~2~ is sensitive to the P/R ratio of the system and also to overall increases in production and respiration. Sediment and nutrient loads through riverine input into the GBR lagoon have increased several fold since European settlement [@pone.0109092-Kroon1]. Thus, the inshore reefs studied here are characterized by elevated chlorophyll and nutrient values and increased turbidity, especially during the summer months [@pone.0109092-Schaffelke2]. Higher near-shore turbidity associated with enhanced wet season runoff on the GBR [@pone.0109092-Cooper1] may reduce in situ light availability and thus restrain benthic primary production and favour heterotrophic processes. Although coastal waters in the GBR system are net-autotrophic throughout the year [@pone.0109092-McKinnon1], pelagic and benthic heterotrophic processes may also be seasonally enhanced by greater inputs of organic carbon in terrestrial runoff, thus shifting the P/R ratio and increasing the *p*CO~2~ levels in coastal GBR waters.\n\nDissolved inorganic carbon and TA levels in samples taken during the daytime directly over the coral reef slopes on inshore reefs did not vary greatly from levels taken at the water surface or at \u223c9 m depth in nearby open water at the research vessel\\'s anchorages. We consider the small differences in pH and *p*CO~2~ measured as biologically insignificant. This casts some doubt on whether inshore reefs (or at least reef slope areas investigated) on the GBR can take up sufficient DIC during light periods to alter carbon chemistry and buffer increased DIC, as has been suggested for larger offshore reefs [@pone.0109092-Albright1]--[@pone.0109092-Shaw1]. We assume that currents and wave induced mixing are too high on inshore reefs for the benthos to affect the water carbon chemistry. On reef flats of Lady Elliot Island (southern GBR) *p*CO~2~ can vary between 89 and 1325 \u00b5atm and pH between 7.59 and 8.56 depending on time of day and season. pH and *p*CO~2~ in similar ranges were measured on the reef flat of One Tree Island, southern GBR [@pone.0109092-Uthicke2] and in lagoonal waters of Heron Island for pH [@pone.0109092-Santos1], [@pone.0109092-Cyronak2]. Diurnal and seasonal changes on a mid-shelf reef flat in the central GBR were also distinct, albeit somewhat less than reported in the latter studies [@pone.0109092-Albright1]. Compared to that, changes at a back-reef area of another mid-shelf reef were rather small [@pone.0109092-Uthicke2]. Thus, it is possible that highly fluctuating values on reef flats reported are extremes and not representative for all habitats, even on mid-shelf reefs. However, further studies including finer scale temporal (i.e., sampling during day and night) and spatial (i.e., comparing different reef habitats) investigations are required to further investigate these dynamics.\n\nAlthough we did not sample mid- and outer-shelf reefs contemporaneously with the inshore reefs, outer-shelf carbon parameters were less variable than encountered near to the coast. From these data there is clear evidence that inshore reefs at present are subjected to higher *p*CO~2~ and lower \u03a9~ar~ than on the outer-shelf. Recent measurements at Lady Elliot Island and Davies Reef [@pone.0109092-Albright1], [@pone.0109092-Shaw1] are in a similar range as observed for mid-shelf reefs here. Both these studies indicated that, even on mid-shelf reefs, night time \u03a9~ar~ values on reef flats can fall below 3, but all day time values were above the value of the surrounding water, and often above 4. Similarly, carbon chemistry analyses from samples collected in surface waters near mid- and outer-shelf reefs of the GBR showed little spatial or temporal variation ([Table 3](#pone-0109092-t003){ref-type=\"table\"}). *p*CO~2~ values on these reefs were generally at or slightly below equilibrium with the atmosphere, and thus somewhat lower than on inshore reefs in the wet-season (\u223c20%) and the late dry season (\u223c10%). Aragonite saturation state on mid- and outer-shelf reefs was in the range of 3.6**--**3.8, also higher than on the inshore reefs studied (\u223c10**--**20%, depending on the season). The PCA analysis clearly separated data from inshore and mid- and outer-shelf reefs, primarily based on their \u03a9~ar~ and *p*CO~2~ values.\n\nIn general, DIC was lower and less variable on mid- and outer-shelf reefs. As discussed above for seasonal differences, possible sources for CO~2~ are calcification, respiration or input through runoff, but it remains unresolved which factor(s) elevate DIC and *p*CO~2~ inshore. It has been shown from a global dataset that reefs closer to the shore generally have elevated *p*CO~2~ levels [@pone.0109092-Cyronak1]. The latter authors suggested that human induced changes in productivity and P/R ratios are the most likely explanation for this. *p*CO~2~ measurements conducted in the GBR lagoon in 1996 [@pone.0109092-Suzuki1], [@pone.0109092-Kawahata1] showed that all samples had *p*CO~2~ concentrations below atmospheric values. Offshore *p*CO~2~ data collected in the present study were also lower than atmospheric levels, although the difference between atmospheric and near-surface *p*CO~2~ is smaller ([Table 5](#pone-0109092-t005){ref-type=\"table\"}). Kawahata et al. [@pone.0109092-Kawahata1] also included data from two inshore stations in the Burdekin and Whitsunday regions, which had *p*CO~2~ close to 325 \u00b5atm and \u03a9~ar~ of \u223c3.8. Data collected in the GBR 30 years ago also showed no indication of elevated *p*CO~2~ compared with the atmospheric values ([Table 5](#pone-0109092-t005){ref-type=\"table\"}, Table S2 in [File S1](#pone.0109092.s001){ref-type=\"supplementary-material\"}) [@pone.0109092-Barnes1], [@pone.0109092-Barnes2]. The rate of increase of *p*CO~2~ on outer reefs measured in this study was on a similar level to that in the atmosphere (1.7 and 1.6 \u00b5atm yr ^\u22121^, respectively, based on dry season data only), which is distinctly below that reported for global reefs based on changes over the last 20 years [@pone.0109092-Cyronak1]. The rate of increase in *p*CO~2~ measured on inshore reefs was much closer to the latter estimates. According to the Bayesian modelling there was a 91% and 98% probability that the increase inshore is higher than offshore for the dry and wet seasons, respectively. This comparison suggests that elevated *p*CO~2~ levels on inshore reefs are a relatively recent phenomenon, and it is possible that these are caused by anthropogenically increased sediment and nutrient runoff.\n\n10.1371/journal.pone.0109092.t005\n\n###### Summary table of historic (1982/83; 1996) and present (2011 to 2013) data from inshore, mid-and outer-shelf reefs for partial pressure of carbon dioxide in the water (*p*CO~2\\ water~), aragonite saturation state (\u03a9~ar~) and partial pressure of carbon dioxide in the atmosphere (*p*CO~2\\ atmosphere~).\n\n![](pone.0109092.t005){#pone-0109092-t005-5}\n\n  Shelf position                                           Season           *p*CO~2~ ~water~ (\u00b5 atm)      \u03a9 ~ar~      *p*CO~2\\ atmosphere~ [\\*\\*](#nt109){ref-type=\"table-fn\"} (\u00b5 atm)\n  ----------------------------------------------- ------------------------ -------------------------- -------------- ------------------------------------------------------------------\n  **1982/83** [\\*](#nt106){ref-type=\"table-fn\"}                                                                      \n  Mid-/Outer-shelf reefs                                Wet (N\u200a=\u200a3)                 350 (13)            4.1 (0.2)                            Mauna Loa: 343 (2)\n                                                        Dry (N\u200a=\u200a4)                 329 (29)            3.7 (0.9)    \n  Inshore                                               Wet (N\u200a=\u200a3)                 324 (24)            4.0 (0.1)    \n                                                        Dry (N\u200a=\u200a2)                 309 (7)             3.7 (0.3)    \n  **1996**\u2227                                                                                                          \n  All lagoonal samples                                  Dry (N\u200a=\u200a27)                332 (13)            3.8 (0.1)                          Cape Ferguson: 361 (1)\n  Inshore                                               Dry (N\u200a=\u200a2)                 325 (1)             3.8 (0.1)                            Mauna Loa: 363 (2)\n  **2011 -13+**                                                                                                      \n  Mid-/Outer-shelf reefs                                Dry (N\u200a=\u200a5)                 380 (15)            3.6 (0.1)                          Cape Ferguson: 394 (1)\n                                                                                                                                             Mauna Loa: 396 (2)\n  Inshore                                               Dry (N\u200a=\u200a8)                 411 (23)            3.1 (0.1)    \n                                                        Wet (N\u200a=\u200a4)                 458 (9)             3.3 (0.2)    \n  Rate of increase:                                *p*CO~2~ (\u00b5atm yr^\u22121^)                              Atmospheric:                          1.7 (1.6**--**2.0)\n                                                                                                        Outer Dry:                        1.6 (**\u2212**0.4**--**3.8)\n                                                                                                       Inshore Dry:                          3.4 (1.8**--**5.0)\n                                                                                                       Inshore Wet:                          4.5 (2.8**--**6.1)\n\nStandard deviations are given in brackets. Average annual rates of increase of atmospheric *p*CO~2~ and water *p*CO~2~ values are calculated from present and 1982/83 data; ranges are 95% Bayesian confidence intervals.\n\n\\*1982/83: Barnes et al. unpublished, for raw data see Table S2 in [File S1](#pone.0109092.s001){ref-type=\"supplementary-material\"}, N\u200a=\u200anumber of reefs surveyed.\n\n\u22271996: Kawahata et al. [@pone.0109092-Kawahata1].\n\n+2011 - 13: present study, data are averages of the annual regional averages.\n\n\\*\\*Atmospheric data are given as annual averages from Mauna Loa, Hawaii (available at: <http://www.esrl.noaa.gov/gmd/ccgg/trends/#mlo_data>) and data for inshore GBR areas at AIMS Townsville (Cape Ferguson: available at: <http://ds.data.jma.go.jp/gmd/wdcgg/cgi-bin/wdcgg/accessdata.cgi?index=CFA519S00-CSIRO&select=inventory>; only available from 1991 onwards); based on raw data averaged over months (N\u200a=\u200a12).\n\nIn conclusion, further investigation of the differences between inshore and offshore carbon chemistry and the trophic status of the surrounding waters are crucial for our understanding of the vulnerability of inshore GBR reefs to climate change and OA. However, given the two year dataset presented here and data from offshore reefs it is apparent that, in addition to sporadically enhanced nutrient levels, decreased light and increased sedimentation [@pone.0109092-Cooper1], [@pone.0109092-Fabricius3], [@pone.0109092-Fabricius5], [@pone.0109092-Schaffelke3], inshore reefs are subjected to elevated *p*CO~2~ values similar to those expected in the near future under OA scenarios [@pone.0109092-Moss1]. The present study also confirms that the rate of increase on inshore reefs is faster than for offshore reefs and atmospheric values [@pone.0109092-Cyronak1]. Increased *p*CO~2~ can be beneficial for primary producers such as benthic algae, seagrasses and phytoplankton [@pone.0109092-Kroeker1], [@pone.0109092-Russell1] but might decrease coral calcification due to reduced aragonite saturation states. Thus, there is a potential for a further shift from coral reefs to algal-dominated areas. However, it should also be acknowledged that at least to some extent corals on inshore reefs can grow under conditions previously predicted as detrimental for coral reef existence (*p*CO~2~ \\>450, \u03a9~ar~\\<3.3) [@pone.0109092-HoeghGuldberg1], [@pone.0109092-Veron1]. To detangle what controls these processes, further water quality and community studies, as well as detailed measurements of calcification and growth of corals and other coral reefs organisms, are needed.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Table S1**, Raw data of all inshore samples analyzed in the present publications. Station: a unique station code from the Australian Institute of Marine Science (AIMS) database; Island: sample location; Code: a depth related code; Depth (in m): actual sampling depth, 0 m\u200a=\u200asurface sample, assumed to be on average from 1 m depth; Dupl.: duplicate number; Temp.: temperature (\u00b0C); Sal.: salinity; DIC: dissolved inorganic carbon (\u00b5mol kg^\u22121^); TA: Total Alkalinity (\u00b5mol kg^\u22121^); Date: collection date; pH: calculated pH on total scale; *p*CO~2~: calculated partial pressure of CO~2~ (\u00b5atm); \u03a9~Ar~: aragonite saturation state; Time: time of sample collection. **Table S2,** Historic water chemistry data from inshore (Bowling Green Bay, Pandora Reef), mid-shelf (Rib Reef, Davies Reef) and outer-shelf (Myrmidon Reef) reefs of the Great Barrier Reef. Samples are the first sample from the cross-reef transects, and are thus close to the windward reef edge and presumed under little influence of reef metabolism. Samples in Cape Bowling Green are not associated with reefs, but represent inshore water close to the coastline. Methods are the same as described in [@pone.0109092-Barnes1], [@pone.0109092-Barnes2].\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\nWe are indebted to all team members of the marine monitoring program group, especially Irena Zagorskis and Johnstone Davidson for their help in collecting water samples. We also thank Stephen Boyle for water chemistry analysis, Dr. Murray Logan for help with statistical analysis and Samantha Talbot for editorial comments on the manuscript.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: SU. Performed the experiments: SU CL MF. Analyzed the data: SU MF CL. Contributed reagents/materials/analysis tools: SU. Wrote the paper: SU MF CL.\n"}
+{"text": "I[NTRODUCTION]{.smallcaps} {#sec1-1}\n==========================\n\nPolycystic ovarian syndrome (PCOS) is a common endocrine system disorder among the women of reproductive age, yet the etiology of PCOS remains unclear. Infertility in females with PCOS can be caused by anovulation, high luteinizing hormone (LH) levels, and hyperandrogenism. In PCOS, follicle arrest occurs, and it prevents full maturation of ovum. Granulosa cells are essential in the production of steroid hormones, providing nutrition and other growth factors that may interact with the oocyte during its development in the ovarian follicle. Follicle-stimulating hormone (FSH) has a role in inducing granulosa cell proliferation, recruitment of secondary follicles, and selection of the dominant follicle and regulates the aromatase activity in the granulosa cells. This research analyses the role of the aromatase gene (CYP19A1) in PCOS pathogenesis.\\[[@ref1][@ref2][@ref3][@ref4][@ref5][@ref6][@ref7][@ref8][@ref9][@ref10]\\]\n\nS[UBJECTS AND]{.smallcaps} M[ETHODS]{.smallcaps} {#sec1-2}\n================================================\n\nThe study protocols were reviewed and approved by the Ethics Committee of the Faculty of Medicine, University of Indonesia (No. 106/H2.F1/ETIK/2013), and all participants provided written informed consent. The design of the study used an analytical, observational, cross-sectional study which was carried out to reach the target of the study that is along with the purpose of the study. The participants of the study that fulfill the criteria of inclusion to analyze the expression of mRNA were 110 people consisting of 55 women with PCOS and 55 women non-PCOS.\n\nThe number of participants in this study fulfilled the minimum number of samples needed for research in a health study of 30 samples. So, the data generated is expected to be normally distributed in calculating statistics. As a consideration, the inclusion criteria for the participants of this study were women with PCOS aged 18--40 years who underwent an *in vitro* fertilization (IVF) program. This is a factor that limits the acquisition of samples in this study because there are not many PCOS women who can undergo IVF programs.\n\nFollicular fluid was collected from the women with PCOS and non-PCOS who were carrying out the fertility program. The granulosa cell samples were obtained from follicle fluid that was aspirated during ovum pick-up (OPU) procedure. The sample of follicle fluid is about 0.5--1.5 ml. Granulosa cell is the excess from the OPU process which is not used for the next IVF steps. However, the granulosa cell is important for research in support of ovum quality.\n\nOPU procedure is an standard operating procedure for the stages of the process that must be undertaken by the patients undergoing IVF. Those who carry out the procedure are medical personnel and embryologists who are competent in their field and have certificates to carry out these actions. The researchers only took and examined the sucking fluid follicles left in the aspiration process during the OPU. While the ovum obtained in the OPU process will be treated for the next IVF process.\n\nGranulosa cell is isolated from follicle fluid by using Ficoll solution. Then, it was multiplied by the cell culture method by using the combination of DMEM F12 media (Gibco) and fetal bovine serum (Gibco). The measurement of the expression mRNA aromatase in the granulosa cell uses real-time quantitative polymerase chain reaction (RT-qPCR).\n\nThe analysis of the expression gene was measured in relative quantification by using the Livak method from the value of cycle threshold. Level of mRNA expression was obtained. The ratio of mRNA CYP19A1 expression is rated based on the Livak formula: *n* = 2^-AACt^.\\[[@ref11]\\]\n\nThe analysis of statistics is to determine the comparison between the relative expression of mRNA CYP19A1 gene. It was done on the sample of granulosa cells in PCOS patients and non-PCOS patients with a *t*-test independently upon the limit of significance of *P* \u2264 0.05 for data which are normal. Meanwhile, the Mann--Whitney U test was used for the data which are not normal (SPSS program version 17, IBM Statistics, New York, USA).\\[[@ref12]\\]\n\nR[ESULTS]{.smallcaps} {#sec1-3}\n=====================\n\nThe measurement of mRNA CYP19A1 expression (aromatase) in the PCOS group and the non-PCOS group used RT-qPCR technique and Livak method. The result of the ratio measurement of mRNA aromatase in granulosa cells in women with PCOS group which was normalized toward non-PCOS (comparison) is shown in [Figure 1](#F1){ref-type=\"fig\"}.\n\n![mRNA CYP19A1 expression ratio in women with polycystic ovarian syndrome group toward the nonpolycystic ovarian syndrome group](JHRS-13-100-g001){#F1}\n\nThe average expression of mRNA aromatase in granulosa cells in women with the PCOS group is 0.38 \u00b1 0.25, while non PCOS group is 1.00 \u00b1 0.00. The result using the Mann--Whitney test shows the difference (*P* \\< 0.01) between the average of mRNA aromatase expression of granulosa cells in women with PCOS compared with women of non-PCOS. The result of this study showed that the average mRNA expression in granulosa cells was lower than the mRNA expression of aromatase in granulosa cells in women with non-PCOS.\n\nFollicle granulosa cells are cells which express the aromatase. One of the main functions of FSH is regulating the activity of enzyme aromatase in granulosa cells through the mechanism which is mediated by cyclic adenosine monophosphate (cAMP) which will stimulate the transcription of aromatase in granulosa cells. FSH acts in granulosa cells to aromatize the androgens from theca cells into estrogen, via its action through follicle stimulating hormone receptor (FSHR). In PCOS cases, Asn680Ser genes possibly affects the regulation of aromatase enzyme in granulosa cells. The aromatase expression is regulated by signaling FSHR. The variety of genotypes in FSHR affects the role of FSHR in regulating the aromatase gene expression which depends on the FSHR signal.\n\nD[ISCUSSION]{.smallcaps} {#sec1-4}\n========================\n\nThe subject of this study were PCOS women in Indonesia who were undergoing an IVF program aged 18--40 years as compared to those women who underwent IVF for male factor infertility, tubal factor or unexplained infertility.\n\nThis study has differences from previous studies on the topic of the CYP 19A1 gene from its purpose of analyzing the role of granulosa cells in the pathogenesis of PCOS in terms of the expression of the CYP19A1 gene mRNA which encodes the aromatase enzyme. This enzyme plays a role in the estrogen biosynthesis in granulosa cells. The ability of granulosa cells to carry out their functions determines the quality of oocytes production which will determine the quality of the embryo.\n\nThe average mRNA CYP19A1 expression for aromatase in granulosa cells in women with PCOS was lower than women with non-PCOS \\[[Figure 1](#F1){ref-type=\"fig\"}\\]. CYP19A1 gene was located on chromosome 15q21.1, encodes the aromatase (P450arom), which is steroidogenesis enzyme key that catalyzes the last steps of estrogen biosynthesis from androgens. It transforms androstenedione to estrone and testosterone to estradiol (E2) and estrone in gonads and extragonadal tissue (brain and adipose tissue).\\[[@ref13][@ref14][@ref15]\\] The previous result also showed that the ratio of mRNA expression for aromatase in granulosa cells in women with PCOS was lower than non-PCOS women.\n\nThe difference of mRNA expression for aromatase in granulosa cells between both groups was index significant statistically of *P* \\< 0.05. It can be assumed that the protein in the aromatase of PCOS group declines because of the excretion compared to the non-PCOS group. It also occurred in the previous result study that states that there was a significant relationship between the expression of mRNA and protein expression. It showed that gene which was expressed high would increase the production of protein more highly than the gene which was expressed lower.\n\nThe biological factors which affect the relationship between mRNA expression and abundance of protein are, protein synthesis, degradation speed of m RNA and protein as well as posttranscriptional m RNA regulation. The involvement of posttranscriptional mechanism and complicated posttranslation techniques possibly determines the abundance of protein.\\[[@ref16]\\]\n\nThe decrease in aromatase activity contributes to the increase of testosterone. This situation has a role in hyperandrogenism which is a typical characteristic of PCOS. In PCOS, as the consequence of the increased LH secretion, it stimulates theca cells to synthesize the testosterone abundantly, while the decrease of aromatase activity causes disorder in the aromatization process. Granulosa cells are unable to aromatize the androgen into estrogen. Hence there is inadequate amount of estrogen for oocyte maturation resulting in chronic anovulation, typical of PCOS.\n\nIn this study, no androgen profile was examined. Previous study stated that the most common phenotype encountered for PCOS was normoandrogenic with the phenotype of anovulation and the picture of polycystic ovaries.\\[[@ref17]\\] Anti-M\u00fcllerian hormone (AMH) levels of women in the PCOS group was almost twice as high as those of women who were not PCOS. This condition is caused by follicular arrest in small preantral and antral stages, where AMH is expressed in high concentrations. Such condition ultimately causes an increase in AMH levels. The research subjects have fulfilled the criteria for the characteristics of patients with PCOS, including AMH levels in the PCOS group, almost twice higher than those in the non-PCOS group. This is also a characteristic of PCOS.\n\nThe result of the study is in line with the report of Naessen *et al*.\\[[@ref18]\\] which states that the activity of aromatase enzyme is lower in women with the PCOS group. The converting activity of testosterone into estrogen by aromatase is lower in follicle fluid for PCOS compared to the control group, and then granulosa cells which were obtained from antral follicles of PCOS patients show low aromatase activity.\\[[@ref19][@ref20]\\] It gets more interesting to analyze forward about the role of CYP19A1 gene with pathogenesis PCOS. Mainly, it is compared to hyperandrogenism in PCOS women. Aromatase has a closer relationship with the concentration of androgen and estrogen, and it is essential to develop the ovary in teenage. The lack of stimulation by FSH will cause inadequacy induction toward enzyme aromatase activity which is very important for synthesizing estrogen, and also it will cause anovulation. FSH can induce aromatase activity which has a positive correlation with E2. Therefore, inadequate E2 can stimulate to produce FSH as a negative feedback. High FSH shows lower aromatase activity.\\[[@ref5][@ref21][@ref22]\\]\n\nThe role of the CYP19A1 gene is possibly involved in pathophysiology syndrome. It can be explained that the relation between FSH and the receptor in the membrane of granulosa cells causes the stimulation of promoter II activity by FSH. The intercellular cAMP will increase, and it will give more significant chance to bind both critical transcription factors that are steroidogenic factor-1 and cAMP response element-binding protein, in proximal promoter II gene aromatase. Aromatase activity gets expressed so that it causes the secretion of estrogen from the preovulatory follicle.\n\nC[ONCLUSION AND]{.smallcaps} S[UGGESTION]{.smallcaps} {#sec1-5}\n=====================================================\n\nExpression of mRNA CYP19A1 in granulosa cells of PCOS patients is lower than non PCOS patient. Further we need to analyze aromatase expression in protein to support the explanation about the role of CYP19A1 towards pathogenesis PCOS.\n\nFinancial support and sponsorship {#sec2-1}\n---------------------------------\n\nNil.\n\nConflicts of interest {#sec2-2}\n---------------------\n\nThere are no conflicts of interest.\n\nMy appreciation and thanks for Rector of Lambung Mangkurat University Kalimantan Selatan, Dean of Medical Faculty Lambung Mangkurat University Kalimantan Selatan and also Universitas Indonesia, Jakarta and lecturers who have given their knowledge during this time, my friends; Maman Saputra who helped in publication process and also Yasmin Clinic-RSCM Kencana with permission and willingness in this study.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nThe liver possesses an enormous capacity for regeneration. After the loss of hepatic tissue caused by toxins or partial surgical removal, remnant hepatocytes markedly proliferate to restore the liver's original mass and functions^[@CR1]^. This process is a compensatory response overcoming the disturbance of whole-body homeostasis provoked by the functional organ defect. Thus, uncovering the precise mechanism of this regenerative response would lead to better understanding of the whole-body homeostatic system.\n\nWhile the roles of humoral factors, including cytokines and growth factors, in liver regeneration have been extensively investigated^[@CR2]^, several studies have also suggested a role of vagal nerve signals in liver regeneration^[@CR3]--[@CR5]^. After partial hepatectomy (PHx), the liver predominantly regenerates through the proliferation of pre-existing hepatocytes^[@CR6]^. Hepatocyte proliferation after PHx reportedly begins at 12--16\u2009h and peaks between 36 and 48\u2009h postoperatively in rodents^[@CR7]^. Reflecting this rapid hepatocyte response, the weight of the remnant liver doubles and has been restored to \\~60--70% of the original weight within the first 72\u2009h after 70% PHx, with eventual restoration of the original weight by 14 days after the operation^[@CR8],[@CR9]^. These findings suggest that the liver regenerative process after injury is mechanistically divided into at least two phases, i.e., an early phase characterized by rapid regeneration and a late phase in which the final organ size is determined. In particular, rapid regeneration after severe organ injury leads to prompt functional recovery which is critical for maintaining homeostasis. Hepatic branch vagotomy reportedly suppresses post PHx liver regeneration, especially in the early phase of regeneration^[@CR3],[@CR4]^. However, the mechanisms by which vagal nerve signals induce rapid hepatocyte proliferation after injury remain largely unknown.\n\nHerein, using PHx as an acute organ defect model, we investigated the molecular mechanism underlying acute liver regeneration. Experiments employing blockade of the hepatic branch of the vagus, inducible hepatocyte-specific *FoxM1* deficiency and adenoviral supplementation of hepatic FoxM1 revealed vagal signals to be involved in rapid activation of the hepatocyte FoxM1 pathway, leading to prompt hepatocyte proliferation after PHx. Importantly, while hepatic branch vagotomy increased postoperative mortality, hepatic supplementation of FoxM1 blocked the increase in mortality, indicating the unraveled mechanism to be critical for promoting survival after liver injury. It is noteworthy that macrophages serve as intermediaries for vagal signals and acute activation of the hepatocyte FoxM1 pathway. Acetylcholine, secreted from the vagus, and interleukin (IL)-6, from macrophages, are involved in the molecular mechanism of the vagal nerve--macrophage--hepatocyte transduction. We recently reported vagal signals to directly activate the FoxM1 pathway in pancreatic \u03b2-cells, thereby promoting compensatory proliferation^[@CR10],[@CR11]^. In contrast to richly innervated pancreatic islets^[@CR12],[@CR13]^, vagal nerves in the liver were very scarce and were observed only around the portal region^[@CR14]^. Thus, this elaborate multistep mechanism, consisting of neuronal, immune, and parenchymal cells, may allow urgent regenerative signals to be amplified and spread throughout the entire organ, thereby promoting prompt liver regeneration and assuring whole-body survival after severe liver injury.\n\nResults {#Sec2}\n=======\n\nVagal signals are critical for post PHx survival {#Sec3}\n------------------------------------------------\n\nFirst, we performed 70% PHx concomitantly with hepatic branch vagotomy (HV) (PHx-HV) or sham operation for HV (PHx-sham) in mice, followed by analyzing the survival rates. Notably, the postoperative mortality of PHx-HV mice, within 3 days after surgery, was markedly higher than that in PHx-sham mice or non-hepatectomized mice subjected to HV alone (HV alone) (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). Histological analysis immediately after their deaths showed that, in contrast to PHx-sham mice, focally necrotic areas with marked bleeding were diffusely located in the livers of PHx-HV mice (Supplementary Fig.\u00a0[1a](#MOESM1){ref-type=\"media\"}). In contrast, regardless of whether or not they had undergone HV, none of the mice, which had survived the initial 3 postoperative days, died during the period of 14 days after the operation (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). In surviving animals, recovery of the hepatic weights of PHx-HV mice was significantly suppressed, as compared with those of PHx-sham mice on day 3 (Supplementary Fig.\u00a0[1b](#MOESM1){ref-type=\"media\"}). Until day 5, the weights of livers from both PHx-sham- and PHx-HV mice were remarkably increased, with an edematous appearance (Supplementary Fig.\u00a0[1c](#MOESM1){ref-type=\"media\"}), and hepatic weights in PHx-HV mice had almost caught up with those in PHx-sham mice on day 7 (Supplementary Fig.\u00a0[1b](#MOESM1){ref-type=\"media\"}). Thus, vagal signals are essential for assuring survival after PHx.Fig. 1Vagal signals are critical for post PHx survival. **a** Cumulative survival of mice that underwent PHx-sham (*n*\u2009=\u200926), PHx-HV (*n*\u2009=\u200926), and HV alone (*n*\u2009=\u200918). **b** Representative images of liver sections immunostained for BrdU on days 2, 5, and 7 after sham operation for PHx (SO), PHx-sham, or PHx-HV. Scale bars indicate 100\u2009\u00b5m. **c** BrdU-positive hepatocyte ratios in SO mice (*n*\u2009=\u20095--6 per group), PHx-sham mice (*n*\u2009=\u20094--6 per group), and PHx-HV mice (*n*\u2009=\u20094--6 per group) on postoperative days 2, 5, and 7. The magnified graphs on days 5 and 7 are shown in framed boxes with lowered scale ranges. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by log-rank test with Bonferroni correction (**a**) or one-way ANOVA followed by Bonferroni's post hoc test (**c**). n.s., not significant\n\nBrdU-positive hepatocytes in PHx-sham mice were remarkably increased as compared with those in mice that had received the sham operation for PHx, by \\~30-fold, on day 2 after PHx. In contrast, HV markedly suppressed the increases in BrdU-positive proliferating hepatocytes (Fig.\u00a0[1b, c](#Fig1){ref-type=\"fig\"}). Similar results were obtained when HV was performed 10 days prior to PHx (Supplementary Fig.\u00a0[1d](#MOESM1){ref-type=\"media\"}). Consistent with hepatic weight changes after PHx, the difference in BrdU-positive hepatocyte ratios between PHx-sham- and PHx-HV mice lost significance on days 5 and 7 (Fig.\u00a0[1b, c](#Fig1){ref-type=\"fig\"}). There were no significant differences in body weights or food intakes, during the entire 14-day observation periods after PHx, between surviving PHx-sham mice and PHx-HV mice (Supplementary Fig.\u00a0[1e, f](#MOESM1){ref-type=\"media\"}), indicating that neither HV-induced increments in post-injury mortality nor suppression of the liver regenerative response was attributable mainly to general conditions, including nutritional states. These findings suggest that vagal signals play important roles in acute-phase liver regeneration, which appears to be critical for the post PHx survival.\n\nVagal signals are involved in activation of hepatic FoxM1 after PHx {#Sec4}\n-------------------------------------------------------------------\n\nWe next investigated the molecular mechanism underlying the vagal signal-mediated acute liver regeneration after PHx. We recently reported that vagal signals promote pancreatic \u03b2-cell proliferation via an intracellular FoxM1-dependent mechanism^[@CR11]^. FoxM1 affects several aspects of cell cycle progression^[@CR15]^, such as promoting the G1/S transition by triggering the transcriptions of several cyclins, including cyclin A2 (*Ccna2*)^[@CR16]--[@CR18]^, regulation of the G2/M transition by transactivating cyclin-dependent kinase 1 (*Cdk1*)^[@CR17],[@CR19]^, and control of proper progression of mitosis by increasing the expressions of several mitogenic genes, such as polo-like kinase 1 (*Plk1*)^[@CR20]^. Therefore, we explored whether FoxM1 expression in hepatocytes is involved in vagal signal-mediated cellular proliferation. PHx markedly upregulated *Foxm1* and cell cycle-related genes downstream from FoxM1, such as *Cdk1*, *Ccna2,* and *Plk1*, as well as a cell proliferation marker, *Mki67*, by several dozen to 100-fold on postoperative days 2 and 3 (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"} and Supplementary Fig.\u00a0[1g](#MOESM1){ref-type=\"media\"}). Importantly, PHx-induced upregulations of *FoxM1* and its related genes were almost completely blocked by HV (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"} and Supplementary Fig.\u00a0[1g](#MOESM1){ref-type=\"media\"}). Consistently, immunoblotting revealed that expressions of hepatic FoxM1 and its downstream proteins, such as Cyclin A2, Cdk1, and PLK1, were markedly increased on postoperative day 2 and that these increases were blocked by HV (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}). Thus, vagal signals are essential for activation of the hepatic FoxM1 pathway in the early phase after PHx, leading to acute hepatocyte replication. Again, in contrast to the early postoperative periods, HV had much weaker and minimal, respectively, impacts on PHx-induced upregulations of FoxM1-related gene expressions on days 5 and 7 (Supplementary Fig.\u00a0[1h, i](#MOESM1){ref-type=\"media\"}), although the magnitudes of increases differed due to using different detection procedures. Thus, hepatocyte regeneration in the late phase is unlikely to be mediated by vagal signals.Fig. 2Vagal signals are involved in activation of hepatic FoxM1 after PHx. **a** Relative expressions of *Foxm1*, its target genes. and *MKi67* in the liver 2 days after SO (*n*\u2009=\u20094), PHx-sham (*n*\u2009=\u20096), and PHx-HV (*n*\u2009=\u20096). **b** Images of liver extract immunoblottings with anti-FoxM1, Cyclin A2, Cdk1, PLK1, and actin in SO, PHx-sham, and PHx-HV mice on postoperative day 2. Arrowhead indicates bands for FoxM1. Data are presented as means\u2009\u00b1\u2009SEM. \\**P*\u2009\\<\u20090.01 assessed by one-way ANOVA followed by Bonferroni's post hoc test (**a**). n.s., not significant\n\nFoxM1 activation promotes liver regeneration and survival {#Sec5}\n---------------------------------------------------------\n\nTo explore whether the FoxM1 upregulation observed in the early phase after PHx is actually responsible for post PHx liver regeneration, we generated tamoxifen-inducible liver-specific FoxM1 knockout mice (iFoxM1LKO mice) by crossing serum albumin promoter-CreER mice^[@CR21]^ and FoxM1-floxed mice^[@CR22]^. In contrast to congenital liver-specific FoxM1 knockout mice^[@CR22]^, iFoxM1LKO mice are not deficient in hepatocyte FoxM1 during developmental and growth stages. Tamoxifen administration to these mice markedly decreased *Foxm1* expressions in the liver, by 71% (Supplementary Fig.\u00a0[2](#MOESM1){ref-type=\"media\"}). As expected, PHx-induced upregulations of the cell cycle-related genes (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}) and increases in BrdU-positive hepatocytes (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}) on day 2 after the surgeries were markedly blunted in iFoxM1LKO mice, showing that FoxM1 expressed in hepatocytes is essential for accelerating the hepatocyte cell cycle, thereby promoting hepatocyte replication in the early phase after PHx.Fig. 3FoxM1 activation promotes liver regeneration and whole-body survival. **a** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after surgery in control mice that underwent SO (*n*\u2009=\u20095) and PHx (*n*\u2009=\u20095), and iFoxM1LKO mice that underwent PHx (*n*\u2009=\u20095). **b** (Upper panels) Representative images of liver sections immunostained for BrdU 2 days after PHx from control and iFoxM1LKO mice. Scale bars indicate 100\u2009\u00b5m. (Lower panel) BrdU-positive hepatocyte ratios in control (*n*\u2009=\u20095) and iFoxM1LKO (*n*\u2009=\u20095) mice 2 days after PHx. **c** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after SO (*n*\u2009=\u20095), PHx-sham (*n*\u2009=\u20096), and PHx-HV (*n*\u2009=\u20095) in mice receiving Ad-LacZ and after PHx-HV in mice receiving Ad-hFoxM1 (*n*\u2009=\u20096). **d** (Upper panels) Representative images of liver sections immunostained for BrdU 2 days after SO, PHx-sham, and PHx-HV from mice receiving Ad-LacZ and after PHx-HV from mice receiving Ad-hFoxM1. Scale bars indicate 100\u2009\u00b5m. (Lower panel) BrdU-positive hepatocyte ratios in the liver 2 days after SO (*n*\u2009=\u20095), PHx-sham (*n*\u2009=\u20096), and PHx-HV (*n*\u2009=\u20095) in Ad-LacZ-treated mice and PHx-HV in Ad-hFoxM1-treated mice (n\u2009=\u20096). **e** Cumulative survival of mice that underwent PHx-sham (*n*\u2009=\u200921) and PHx-HV (*n*\u2009=\u200934) procedures after receiving Ad-LacZ and mice that underwent PHx-HV after receiving Ad-hFoxM1 (*n*\u2009=\u200934). Data are presented as means\u2009\u00b1\u2009SEM. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by one-way ANOVA followed by Bonferroni's post hoc test (**a**, **c**, **d**), unpaired *t* test (**b**), or log-rank test with Bonferroni correction (**e**). n.s., not significant\n\nNext, to explore the role of FoxM1 in vagal signal-mediated liver regeneration, we examined the effects of HV on liver regeneration after PHx using mice, in which hepatic FoxM1 had been supplemented by adenoviral gene transduction. We used adenovirus containing the human *Foxm1* gene to distinguish endogenous FoxM1 from exogenously expressed FoxM1. We selected an adenoviral titer at which gene transduction increased exogenous *Foxm1* expression in the livers of mice to levels similar to those of hepatic endogenous FoxM1 in PHx-sham mice induced on day 2 after PHx (Supplementary Fig.\u00a0[3a](#MOESM1){ref-type=\"media\"}). Strikingly, in hepatically FoxM1-supplemented mice, HV failed to suppress PHx-induced increments in expressions of the genes downstream from FoxM1 (Fig.\u00a0[3c](#Fig3){ref-type=\"fig\"}). This effect was also observed in BrdU-positive hepatocytes (Fig.\u00a0[3d](#Fig3){ref-type=\"fig\"}). Notably, hepatic FoxM1 supplementation completely blocked increases in mortality after PHx in vagotomized mice (Fig.\u00a0[3e](#Fig3){ref-type=\"fig\"}). Interestingly, exogenous (human) FoxM1 expression in normal mice, without PHx or HV, did not alter the expressions of either endogenous mouse *FoxM1* or its target genes, nor that of *Mki67* (Supplementary Fig.\u00a0[3b](#MOESM1){ref-type=\"media\"}). Furthermore, hepatic FoxM1 overexpression in PHx-sham mice did not yield further enhancements of hepatocyte proliferative responses, suggesting that the hepatic FoxM1 pathway had already been fully activated in response to PHx (Supplementary Fig.\u00a0[3c, d](#MOESM1){ref-type=\"media\"}). In addition, administration of carbachol to normal mice failed to upregulate hepatic FoxM1-related genes (Supplementary Fig.\u00a0[3e](#MOESM1){ref-type=\"media\"}). These results indicate that, under early-phase post-PHx conditions, activation of the hepatic FoxM1 pathway mediates vagal signal-induced promotion of hepatocyte replication and the resultant whole-body survival.\n\nMacrophages mediate vagus-induced hepatic FoxM1 activation {#Sec6}\n----------------------------------------------------------\n\nOur next goal was to elucidate the mechanism by which vagal signals activate the FoxM1 pathway in hepatocytes. First, to examine whether the muscarinic receptor signaling induced by acetylcholine, the main neurotransmitter released by the vagal nerve, is involved, mice were treated with atropine, a muscarinic receptor antagonist, followed by PHx. With atropine treatment, the increases in BrdU-positive hepatocytes were markedly reduced (Fig.\u00a0[4a](#Fig4){ref-type=\"fig\"}) and recovery of hepatic weights in the early phase after PHx was suppressed (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}). Consistently, upregulations of hepatic *Foxm1* and its target genes, as well as *Mki67* after PHx were significantly blocked by atropine treatment (Fig.\u00a0[4c](#Fig4){ref-type=\"fig\"}). These data indicate that muscarinic signals are involved in promoting acute liver regenerative responses after PHx.Fig. 4Muscarinic signals are involved in acute liver regenerative responses after PHx. **a** (Upper panels) Representative images of liver sections immunostained for BrdU on day 2 after sham operation for PHx (SO) and PHx in vehicle-treated mice and after PHx in atropine-treated mice. Scale bars indicate 100\u2009\u00b5m. (Lower panel) BrdU-positive hepatocyte ratios in SO\u2009+\u2009vehicle mice (*n*\u2009=\u20094), PHx\u2009+\u2009vehicle mice (*n*\u2009=\u20095), and PHx\u2009+\u2009atropine mice (*n*\u2009=\u20095). **b** Fold changes in hepatic weights 3 days after surgery in PHx\u2009+\u2009vehicle mice (*n*\u2009=\u20095) and PHx\u2009+\u2009atropine mice (*n*\u2009=\u20095). Hepatic weights were divided by those obtained immediately after surgery. **c** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after SO (*n*\u2009=\u20094) and PHx (*n*\u2009=\u20095) in vehicle-treated mice and PHx (*n*\u2009=\u20095) in atropine-treated mice. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by one-way ANOVA followed by Bonferroni's post hoc test (**a** and **c**) or assessed by unpaired *t* test (**b**). n.s., not significant\n\nWe then attempted to elucidate the molecular mechanism underlying vagal signal-mediated hepatocyte proliferation in vitro by using a murine hepatoma cell line, Hepa1--6 cells. Unexpectedly, however, direct stimulation with carbachol, a cholinergic receptor agonist, did not upregulate *Foxm1* and *Mki67* gene expressions in the cultured hepatocytes (Supplementary Fig.\u00a0[4a](#MOESM1){ref-type=\"media\"}). Therefore, we examined the possibility that an indirect mechanism(s) mediates the effects of vagal signals on hepatocyte proliferation. Since resident macrophages are reportedly involved in tissue repair in several tissues^[@CR23]^, we examined post PHx hepatocyte proliferation using mice in which hepatic macrophages had been depleted by treatment with clodronate liposomes. Clodronate liposome administration significantly decreased hepatic expressions of macrophage markers (Fig.\u00a0[5a](#Fig5){ref-type=\"fig\"}), indicating successful depletion of hepatic macrophages. Intriguingly, clodronate liposome administration markedly blunted activation of the FoxM1 pathway in the liver on day 2 after PHx, as compared with control liposome treatment (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}). In addition, the increases in BrdU-positive hepatocytes after PHx were markedly inhibited by clodronate liposome administration (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}). Of note, HV failed to exert further inhibitory effects on the acute hepatocyte proliferation responses of clodronate liposome-treated mice (Fig.\u00a0[5b, c](#Fig5){ref-type=\"fig\"}). Furthermore, adenoviral supplementation of hepatic FoxM1 significantly blunted the inhibitory effects of macrophage depletion on acute hepatocyte proliferation responses after PHx (Fig.\u00a0[5d, e](#Fig5){ref-type=\"fig\"}). Notably, the post-PHx mortality of macrophage-depleted mice, within 2 days postoperatively, was markedly higher than those of mice without macrophage depletion (Fig.\u00a0[5f](#Fig5){ref-type=\"fig\"}), observations similar to those in mice undergoing HV (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). These findings, taken together, suggest that hepatic macrophages contribute to mediating vagal signals to hepatocytes, thereby activating the hepatocyte FoxM1 pathway and promoting hepatocyte proliferation.Fig. 5Macrophages mediate vagus-induced hepatic FoxM1 activation. **a** (Upper panels) Representative images of liver sections immunostained for F4/80 from mice 24\u2009h after receiving control and clodronate liposomes. Scale bars indicate 100\u2009\u00b5m. (Lower panel) Relative gene expressions of *F4/80* in the liver 24\u2009h after control (*n*\u2009=\u20094) and clodronate (*n*\u2009=\u20094) liposome administration. **b** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after surgery in SO\u2009+\u2009control mice (*n*\u2009=\u20096), PHx-sham\u2009+\u2009control mice (*n*\u2009=\u20098), PHx-sham\u2009+\u2009clodronate mice (*n*\u2009=\u20096), and PHx-HV\u2009+\u2009clodronate mice (*n*\u2009=\u20096). **c** (Left panels) Representative images of liver sections immunostained for BrdU on day 2 in SO\u2009+\u2009control-, PHx-sham\u2009+\u2009control-, PHx-sham\u2009+\u2009clordronate-, and PHx-HV\u2009+\u2009clodronate mice. Scale bars indicate 100\u2009\u00b5m. (Right panel) BrdU-positive hepatocyte ratios in SO\u2009+\u2009control mice (*n*\u2009=\u20096), PHx-sham\u2009+\u2009control mice (*n*\u2009=\u20098), PHx-sham\u2009+\u2009clodronate mice (*n*\u2009=\u20096,) and PHx-HV\u2009+\u2009clodronate mice (*n*\u2009=\u20096) 2 days after surgery. **d** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after SO in mice receiving control liposome and Ad-LacZ (*n*\u2009=\u20096) and PHx in mice receiving control liposome and Ad-LacZ (*n*\u2009=\u20096), clodronate liposome and Ad-LacZ (*n*\u2009=\u20096), and clodronate liposome and Ad-hFoxM1 (*n*\u2009=\u20096). **e** (Left panels) Representative images of liver sections immunostained for BrdU on day 2 after SO in mice receiving control liposome and Ad-LacZ and PHx in mice receiving control liposome and Ad-LacZ, clodronate liposome and Ad-LacZ, and clodronate liposome and Ad-hFoxM1. Scale bars indicate 100\u2009\u00b5m. (Right panel) BrdU-positive hepatocyte ratios in the liver 2 days after SO in mice receiving control liposome and Ad-LacZ (*n*\u2009=\u20096) and PHx in mice receiving control liposome and Ad-LacZ (*n*\u2009=\u20096), clodronate liposome and Ad-LacZ (*n*\u2009=\u20096), and clodronate liposome and Ad-hFoxM1 (*n*\u2009=\u20096). **f** Mortality rates within 2 postoperative days in PHx\u2009+\u2009control mice (*n*\u2009=\u200915) and PHx\u2009+\u2009clodronate mice (*n*\u2009=\u200915). Data are presented as means\u2009\u00b1\u2009SEM. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by one-way ANOVA followed by Bonferroni's post hoc test (**b**, **c**, **d**, **e**) or assessed by unpaired *t* test (**a**). n.s., not significant\n\nVagal signals increase macrophage IL-6 production after PHx {#Sec7}\n-----------------------------------------------------------\n\nWhat then is the mediator of the macrophage--hepatocyte link? Activated macrophages are well-known to produce and release pro-inflammatory cytokines, including IL-6^[@CR24]^, and the cellular source of IL-6 after PHx was reported to be resident macrophages^[@CR25]^. Therefore, we focused on IL-6 production after PHx. *Il-6* expression in the remnant liver of PHx-sham mice was markedly increased at 6\u2009h after PHx and, importantly, this *Il-6* upregulation was almost completely abolished by HV (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}). Plasma IL-6 concentrations were also increased after PHx, and these increases were suppressed by vagotomy (Supplementary Fig.\u00a0[4b](#MOESM1){ref-type=\"media\"}). In addition, upregulation of *Il-6* by PHx was completely absent in clodronate liposome-treated mice (Fig.\u00a0[4b](#Fig4){ref-type=\"fig\"}). Consistent with the results of hepatocyte proliferative responses (Fig.\u00a0[5b, c](#Fig5){ref-type=\"fig\"}), HV failed to exert further inhibitory effects on upregulation of *Il-6* in clodronate liposome-treated mice (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}). Therefore, these findings indicate that, after PHx, vagal signals enhance IL-6 production by hepatic macrophages in the remnant liver.Fig. 6Vagal signals increase macrophage IL-6 production after PHx. **a** Relative gene expressions of *Il-6* after surgery in PHx-sham mice (*n*\u2009=\u20094 per group) and PHx-HV mice (*n*\u2009=\u20094 per group). **b** Relative gene expressions of *Il-6* after PHx-sham in control liposome-treated (*n*\u2009=\u20095--6 per group) and clodronate liposome-treated mice (*n*\u2009=\u20095--6 per group), and after PHx-HV in clodronate liposome-treated mice (*n*\u2009=\u20095--6 per group). **c** Relative gene expression of *Il-6* in primary macrophages after 4\u2009h of stimulation with vehicle (*n*\u2009=\u20096), 100\u2009\u00b5M carbachol (*n*\u2009=\u20096), or both 100\u2009\u00b5M carbachol and 100\u2009\u00b5M atropine (*n*\u2009=\u20096). **d** Relative gene expressions of *Il-6* after PHx in vehicle-treated (*n*\u2009=\u20094 per group) and atropine-treated (*n*\u2009=\u20094 per group) mice. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by unpaired *t* test (**a** and **d**), or assessed by one-way ANOVA followed by Bonferroni's post hoc test (**b** and **c**). n.s., not significant\n\nWe next investigated the effects of cholinergic signals on macrophage IL-6 production employing ex vivo experiments using primary macrophages isolated from the peritoneal cavity after intraperitoneal thioglycollate administration. Strikingly, carbachol treatment markedly increased *Il-6* expression in these primary macrophages in a dose-dependent manner (Supplementary Fig.\u00a0[4c](#MOESM1){ref-type=\"media\"}), and this increase was significantly blunted by co-treatment with atropine (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}). In contrast, carbachol treatment of primary hepatocytes failed to increase *Il-6* expression (Supplementary Fig.\u00a0[4d](#MOESM1){ref-type=\"media\"}). Moreover, in vivo atropine treatment significantly suppressed PHx-induced increases in *Il-6* expression in the liver (Fig.\u00a0[6d](#Fig6){ref-type=\"fig\"}). Collectively, vagus-derived cholinergic signals can directly upregulate macrophage IL-6 production through a muscarinic receptor-dependent mechanism.\n\nIL-6 mediates vagal signal-induced hepatic FoxM1 activation {#Sec8}\n-----------------------------------------------------------\n\nTo examine whether IL-6 activates the FoxM1 pathway in hepatocytes, we treated primary hepatocytes with IL-6. Baseline *Foxm1* levels in primary hepatocytes were approximately eightfold higher than those in non-treated quiescent livers (Supplementary Fig.\u00a0[5a](#MOESM1){ref-type=\"media\"}). The isolation procedure may intrinsically enhance *Foxm1* expression. Even under these conditions, IL-6 treatment of primary hepatocytes further increased expressions of *Foxm1*, its target genes, and *Mki67* (Fig.\u00a0[7a](#Fig7){ref-type=\"fig\"}). It should be noted that IL-6-induced upregulations of these cell cycle-related genes were markedly blocked in primary hepatocytes from iFoxM1LKO mice (Fig.\u00a0[7a](#Fig7){ref-type=\"fig\"}). These findings indicate that IL-6 accelerates cell cycling of hepatocytes through a FoxM1-dependent mechanism.Fig. 7IL-6 mediates vagal signal-induced hepatic FoxM1 activation. **a** Relative expressions of *Foxm1*, its target genes, and *MKi67* in primary hepatocytes isolated from control and iFoxM1LKO mice treated with 100\u2009ng/ml IL-6 (*n*\u2009=\u20097--8 per group) and vehicle (*n*\u2009=\u20097--8 per group) for 6\u2009h. **b** (Left panels) Representative images of liver extract immunoblottings with anti-phospho-STAT3, total STAT3, and actin. (Right panel) Relative intensities of phospho/total STAT3 in livers from sham operation for PHx (SO)- (*n*\u2009=\u20096), PHx-sham- (*n*\u2009=\u20096), and PHx-HV mice (*n*\u2009=\u20096). **c** Relative expressions of *Foxm1*, its target genes, and *MKi67* in the liver 2 days after surgery from SO\u2009+\u2009IgG (*n*\u2009=\u20095), PHx\u2009+\u2009IgG (*n*\u2009=\u20095), and PHx\u2009+\u2009anti-IL-6 antibody- (*n*\u2009=\u20095) treated mice. Data are presented as means\u2009\u00b1\u2009SEM. \\**P*\u2009*\\<*\u20090.05; \\*\\**P*\u2009\\<\u20090.01 assessed by unpaired *t* test (**a**), or assessed by one-way ANOVA followed by Bonferroni's post hoc test (**b** and **c**). n.s., not significant\n\nIL-6 is well-known to exert its effects through signal transducer and activator of transcription 3 (STAT3) signaling^[@CR26]^. IL-6 promotes phosphorylation and homodimerization of STAT3, thereby increasing STAT3 translocation to the nucleus, leading to transactivation of its target genes^[@CR26]^. Since STAT3 reportedly transactivates *Foxm1* gene expression^[@CR27]^, we next examined hepatic STAT3 phosphorylation of PHx-sham- and PHx-HV mice. Compared with control mice, hepatic STAT3 phosphorylation was significantly increased at 24\u2009h after PHx, and this enhancement of phosphorylation was blunted by HV (Fig.\u00a0[7b](#Fig7){ref-type=\"fig\"}). Next, we pretreated primary hepatocytes with STAT3 inhibitor peptide, which inhibits both homodimerization of STAT3 and heterodimerization of STAT3 and STAT1^[@CR28]^, followed by treatment with IL-6. Pretreatment with the STAT3 inhibitor blocked IL-6-mediated increases in *FoxM1* and its target genes as well as *Mki67* (Supplementary Fig.\u00a0[5b](#MOESM1){ref-type=\"media\"}). In addition, we examined the phosphorylation of STAT1 as well as that of STAT5 after PHx. In contrast to STAT3, phosphorylations of STAT1 and STAT5 were not altered after PHx (Supplementary Fig.\u00a0[5c](#MOESM1){ref-type=\"media\"}). These results suggest that IL-6 promotes phosphorylation and homodimerization of STAT3, thereby upregulating FoxM1-related genes in hepatocytes.\n\nFinally, to examine the significance of IL-6 in activation of the hepatic STAT3--FoxM1 pathway in vivo, we blocked IL-6 signaling by administering anti-IL-6 antibody, followed by PHx surgery. IL-6 neutralization significantly suppressed the hepatic STAT3 phosphorylation enhanced by PHx (Supplementary Fig.\u00a0[5d](#MOESM1){ref-type=\"media\"}). Under these conditions, upregulations of *Foxm1* and its target cell cycle-related genes, as well as *Mki67* on day 2 after PHx were markedly suppressed in anti-IL-6 antibody-treated mice (Fig.\u00a0[7c](#Fig7){ref-type=\"fig\"}). Thus, IL-6 signaling plays key roles in activation of the FoxM1 pathway in hepatocytes and in the proliferation of hepatocytes during the acute phases after PHx.\n\nTaken together, our observations suggest that vagal nerve-derived cholinergic signals are likely to stimulate IL-6 production by resident macrophages in the liver, and that, in a paracrine manner, IL-6 enhances hepatocyte proliferation based on a FoxM1-dependent mechanism (Fig.\u00a0[8](#Fig8){ref-type=\"fig\"}).Fig. 8The proposed mechanism of vagus-mediated liver regeneration. The multistep regulatory mechanism of acute liver regeneration after PHx via a vagus--macrophage--hepatocyte link\n\nDiscussion {#Sec9}\n==========\n\nUsing the PHx model, we have clarified that vagal signal-regulated acute liver regenerative responses are critical for promoting survival after severe liver injury. Postoperative mortality was high in PHx-HV mice but, regardless of whether or not HV had been performed, all of the mice which had survived the initial 3 postoperative days survived for at least 14 days after the operation (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). Thus, acute recovery of liver mass within a few days after PHx is critical for life maintenance. Increases in postoperative mortality in PHx-HV mice were accompanied by areas of focal necrosis with marked bleeding in the livers of these mice. Further studies are necessary to elucidate the mechanisms by which HV leads to such pathological changes. Meanwhile, hepatic FoxM1 supplementation in PHx-HV mice completely blocked the increases in mortality (Fig.\u00a0[3e](#Fig3){ref-type=\"fig\"}) with recovery of acute liver regeneration (Fig.\u00a0[3c, d](#Fig3){ref-type=\"fig\"}). On the other hand, adenoviral FoxM1 expression or carbachol administration to normal mice had no effects on the FoxM1 pathway (Supplementary Fig.\u00a0[3b, e](#MOESM1){ref-type=\"media\"}). These findings suggest that additional and unknown signals, induced by PHx, are necessary for promoting hepatocyte proliferation. Thus, vagus-mediated activation of the FoxM1 pathway is critical, especially when prompt hepatocyte regeneration is required, and functions as a preventive mechanism against whole-body death after injury. In contrast, chronic liver regeneration occurred even in PHx-HV mice, indicating different mechanisms and roles of liver regeneration in the acute versus chronic phases. Late-phase liver regeneration was proposed to be promoted by humoral factors including several growth factors^[@CR7]^, and could be important in regulating liver size. On the other hand, neuronal signals may be advantageous for inducing prompt compensatory regeneration, which confers a survival advantage for animals suffering severe crises such as an organ defect. Recovery of liver weight within 72\u2009h after PHx is reportedly reduced in 40- as compared to 4-week-old mice^[@CR29]^. It is noteworthy that 40-week-old mice showed a low survival rate within 72\u2009h after PHx^[@CR29]^, similar to that of PHx-HV mice in the present study. Taken together with our results, the lower survival rate after PHx associated with aging might be attributable to impairment of the mechanism, underlying acute liver regenerative responses, elucidated in the present study.\n\nIn this study, we have delineated a full picture of vagal signal-regulated acute hepatic regenerative responses. We have unmasked several novel mechanisms involved in this process. First, vagal nerve signals play critical roles in upregulating the hepatocyte FoxM1 pathway, thereby promoting rapid liver regeneration in the early phases after injury. Second, resident macrophages mediate processes involving vagal signals and prompt hepatocyte proliferation in these regenerative responses. Third, IL-6 is a critical mediator in the mechanism linking macrophages and hepatocytes.\n\nConsistent with the results of a previous study using congenital FoxM1 knockout mice^[@CR22]^, liver-specific induction of FoxM1 deficiency prior to PHx markedly suppressed liver regeneration in the early phase after PHx (Fig.\u00a0[3a, b](#Fig3){ref-type=\"fig\"}). Notably, we further showed that HV did not suppress liver regeneration and whole-body survival in hepatically FoxM1-supplemented mice (Fig.\u00a0[3d, e](#Fig3){ref-type=\"fig\"}). Thus, vagal signal-induced FoxM1 upregulation in hepatocytes is necessary and sufficient for accelerating the hepatocyte cell cycle, thereby promoting hepatocyte replication in the early phase after PHx. Given that vagus-mediated FoxM1 activation is involved in compensatory \u03b2-cell proliferation^[@CR11]^, the vagus-FoxM1 linking mechanism may be shared by a variety of tissues for in vivo regeneration.\n\nDirect carbachol stimulation did not upregulate FoxM1-related genes in hepatocytes (Supplementary Fig.\u00a0[4a](#MOESM1){ref-type=\"media\"}), but did markedly enhance IL-6 production in macrophages (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}), suggesting the targets of vagal signals to be hepatic macrophages, rather than hepatocytes. As reported^[@CR24],[@CR30]^, hepatic macrophage-depleted mice showed marked impairment of liver regeneration (Fig.\u00a0[5c](#Fig5){ref-type=\"fig\"}). Importantly, this impairment was accompanied by blunting of FoxM1 pathway activation (Fig.\u00a0[5b](#Fig5){ref-type=\"fig\"}), and vagotomy failed to exert further inhibitory effects on either acute liver regenerative responses (Fig.\u00a0[5b, c](#Fig5){ref-type=\"fig\"}) or the upregulation of hepatic *Il-6* (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}). Vagal nerve signals reportedly regulate resident macrophages, thereby modulating several biological conditions in peripheral organs, such as inflammation^[@CR31],[@CR32]^ and metabolism^[@CR33]^. In this study, we clarified that the vagal nerve--macrophage link is further involved in the organ regenerative process. In this process, the cholinergic signal is identified as a stimulator of hepatic macrophages. Cholinergic signals reportedly promote anti-inflammatory responses of macrophages through the \u03b17-nicotinic receptor^[@CR31],[@CR32]^. In addition, activation of muscarinic signals in the central nervous system has systemic anti-inflammatory properties, and these effects are reportedly elicited by efferent vagal nerve-mediated nicotinic^[@CR34]^, rather than muscarinic, actions^[@CR35]^ on peripheral immune cells. Meanwhile, since atropine treatment blunted the carbachol-mediated (Fig.\u00a0[6c](#Fig6){ref-type=\"fig\"}) and PHx-induced (Fig.\u00a0[6d](#Fig6){ref-type=\"fig\"}) upregulations of IL-6 in macrophages ex vivo and in the remnant liver in vivo, respectively, muscarinic signaling is involved in liver regeneration. Intriguingly, inflammatory responses opposite those of hepatic macrophages to cholinergic signals, depending on whether the signals are mediated by muscarinic or nicotinic receptors, were recently reported^[@CR33]^. Thus, responses of macrophages to vagal signals are likely to differ depending on physiological or pathological situations. Expression levels of muscarinic and nicotinic receptors on macrophages might be altered according to various situations, although further examinations are needed to elucidate the molecular mechanism underlying the opposite responses of macrophages to cholinergic stimulation.\n\nIn terms of the macrophage--hepatocyte link, IL-6 was revealed to be a key mediator. *Il-6* expression in the liver was rapidly and markedly increased after PHx, and this increase was found to be suppressed by vagotomy (Fig.\u00a0[6a](#Fig6){ref-type=\"fig\"}) and in hepatic macrophage-depleted mice (Fig.\u00a0[6b](#Fig6){ref-type=\"fig\"}). IL-6 treatment activated the FoxM1 pathway in hepatocytes ex vivo (Fig.\u00a0[7a](#Fig7){ref-type=\"fig\"}), and activation of the FoxM1 pathway after PHx was inhibited by antibody-mediated blockade of hepatic IL-6 signaling in vivo (Fig.\u00a0[7c](#Fig7){ref-type=\"fig\"}). In addition, PHx enhanced phosphorylation of hepatic STAT3 (Fig.\u00a0[7b](#Fig7){ref-type=\"fig\"}), a downstream target of the IL-6 receptor, and this enhancement of STAT3 phosphorylation was suppressed by anti-IL-6 antibody treatment (Supplementary Fig.\u00a0[5d](#MOESM1){ref-type=\"media\"}). Consistently, hepatocyte-specific STAT3 knockout mice reportedly show retardation of liver regeneration after PHx^[@CR36]^. Considering that STAT3 binds to the *Foxm1* gene promoter and that a STAT3 inhibitor was shown to reduce *Foxm1* expression in a leukemic cell line^[@CR27]^, STAT3 phosphorylated by IL-6 signaling may transactivate *Foxm1* gene expression in regenerating hepatocytes.\n\nCollectively, muscarinic signals from the vagal nerve may stimulate IL-6 production by hepatic macrophages, thereby activating the hepatocyte FoxM1 pathway in a paracrine manner. Thus, acute liver regenerative responses are achieved by a multistep mechanism comprised of neuronal, immune, and parenchymal cells. We recently reported that vagal signals directly activate the FoxM1 pathway in pancreatic \u03b2-cells, thereby enhancing compensatory proliferation^[@CR10],[@CR11]^. Pancreatic islets are richly innervated by vagal nerves^[@CR11]--[@CR13]^, and this anatomical feature may allow vagal signals to reach islet cells efficiently. In contrast, it was reported that vagal nerves in the liver were observed only around the portal region^[@CR14]^. Therefore, to achieve prompt regeneration of remnant hepatocytes, there must be a distinct mechanism whereby neuronal signals are conveyed to individual cells in the liver, which is a large organ. In this context, taking advantage of resident macrophages as intermediaries, regenerative signals from the vagal nerve can be amplified and spread throughout the entire remnant liver by IL-6 secretion (Fig.\u00a0[8](#Fig8){ref-type=\"fig\"}).\n\nThe findings obtained in the present study illustrate the novel concept that a vagal signal--macrophage link promptly promotes organ regeneration which is necessary especially after severe organ injury. This concept may lead to better understanding of both the organ regenerative process and the whole-body homeostasis-maintaining mechanisms. Organ regeneration is reportedly regulated by neuronal signals in other tissues as well, such as the limbs^[@CR37],[@CR38]^ and pancreatic \u03b2-cells^[@CR10],[@CR11]^. Macrophages also play regulatory roles in the regeneration of several tissues other than the liver, such as the spinal cord and the heart^[@CR23]^. Thus, our results raise the possibility of elucidating whether the vagal signal--macrophage link is involved in the regenerations of other organs/tissues and may be useful for developing novel strategies for regenerative medicine.\n\nMethods {#Sec10}\n=======\n\nAnimals {#Sec11}\n-------\n\nAnimal studies were conducted in accordance with the Tohoku University institutional guidelines. Ethical approval has been obtained from the Institutional Animal Care and Use Committee of the Tohoku University Environmental & Safety Committee.\n\nMale C57BL/6N mice were purchased from SLC Japan (Shizuoka, Japan). Transgenic mice expressing a Cre recombinase transgene fused to mutated estrogen receptor ligand-binding domains under the control of the mouse serum albumin promoter/enhancer (SA-CreER^T2^)^[@CR21]^ were gifts from Prof. Pierre Chambon and Dr. Daniel Metzger (Institute of Genetics and Molecular and Cellular Biology, Illkirch-Cedex, France). To obtain tamoxifen-inducible liver-specific FoxM1 knockout (iFoxM1LKO) mice, we crossed SA-CreER^T2^ mice and FoxM1^flox/flox^ mice^[@CR22]^. These mice have mixed backgrounds. At 8 weeks of age, SA-CreER^T2^; FoxM1^flox/flox^ mice and FoxM1^flox/flox^ mice (as controls) were injected intraperitoneally with 1\u2009mg of tamoxifen (Sigma, St. Louis, MO, USA) dissolved in corn oil (Sigma) every other day for 5 days, and were subjected to PHx 7 days after the end of tamoxifen administration. All mice were housed individually in a controlled environment (room temperature 25\u2009\u00b0C) with a 12-h light--dark cycle, and received standard chow and drinking water ad libitum. Animal studies were conducted in accordance with Tohoku University institutional guidelines.\n\nSurgical procedures {#Sec12}\n-------------------\n\nAll operations were carried out on 8--10-week-old male mice, anesthetized with an intraperitoneal injection containing a mixture of medetomidine (0.3\u2009mg/kg), midazolam (4\u2009mg/kg), and butorphanol tartrate (5\u2009mg/kg). For HV, after incising the abdominal wall, the stomach was retracted to expose the anterior vagal trunk and hepatic branch, and then only the hepatic branch was transected with fine forceps. Immediately after HV, 70% PHx was performed. The left lateral and median lobes of the liver were securely ligated with 6-0 silk suture and resected^[@CR39],[@CR40]^. For the sham operations, only the abdominal incision was made, and the liver tissues and hepatic vagal branch were left intact. At the completion of surgery, the abdominal muscles and skin were both sutured layer by layer with 6-0 silk. In the experiments shown in Supplementary Fig.\u00a0[1d](#MOESM1){ref-type=\"media\"}, PHx was performed 10 days after HV or sham operation for HV.\n\nRecombinant adenovirus {#Sec13}\n----------------------\n\nRecombinant adenoviruses carrying the human *Foxm1* gene (Ad-hFoxM1) and \u03b2-galactosidase gene (Ad-LacZ) were prepared^[@CR41]^. Then, 1\u2009\u00d7\u200910^8^ plaque-forming units (PFU) per mouse of the Ad-hFoxM1 were injected intravenously 1\u2009day prior to PHx. The control mice were given 1\u2009\u00d7\u200910^8^ PFU per mouse of the Ad-LacZ. In the experiments shown in l Fig.\u00a0[5d, e](#Fig5){ref-type=\"fig\"}, recombinant adenoviruses were injected into mice immediately after control or clodronate liposome administration.\n\nIn vivo carbachol treatment {#Sec14}\n---------------------------\n\nCarbachol (Nacalai Tesque, Kyoto, Japan) was dissolved in saline to 2\u2009mmol/l, and mice were intraperitoneally administered 200\u2009nmol carbachol (0.1\u2009ml of 2\u2009mmol/l solution). Livers were harvested 6\u2009h after carbachol administration.\n\nIn vivo atropine treatment {#Sec15}\n--------------------------\n\nAtropine (Sigma) was dissolved in saline with ethanol (10% v/v) to 625\u2009mg/dl. Mice were intraperitoneally administered atropine, at 25\u2009\u00b5g/g body weight, twice a day from 2 days before the surgical interventions until kill.\n\nStudies with Hepa1--6 cells {#Sec16}\n---------------------------\n\nMurine hepatocyte cell line Hepa1--6 cells (CRL-1830) were obtained from ATCC and maintained in Dulbecco's modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS), 100\u2009U/ml penicillin, and 100\u2009\u00b5g/ml streptomycin (P/S) at 37\u2009\u00b0C with 5% CO~2~ and 95% air. Hepa1--6 cells were incubated with 100\u2009\u00b5M carbachol (Nacalai Tesque) for 24\u2009h, and then collected for RNA extraction. Water was used as the vehicle.\n\nStudies with peritoneal macrophages {#Sec17}\n-----------------------------------\n\nPeritoneal macrophages were harvested from lavage of 8-week-old male C57BL/6\u2009N mice 4 days after intraperitoneal injection with 4% thioglycollate (Sigma)^[@CR42]^, plated onto 24-well plates with 1.0\u2009\u00d7\u200910^5^ cells per well and cultured in DMEM with 10% FBS and P/S. At 24\u2009h after isolation, the cells were incubated with 100\u2009\u00b5M carbachol (Nacalai Tesque)^[@CR43]^ and 100\u2009\u00b5M atropine (Sigma) for 4\u2009h, and finally collected for RNA extraction. The carbachol concentration was determined by examining dose-dependent effects of carbachol on IL-6 expression in primary macrophages. The atropine concentration was determined according to previously reported in vitro experiments^[@CR11]^.\n\nMacrophage depletion {#Sec18}\n--------------------\n\nTo deplete hepatic macrophages, mice were intraperitoneally administered 140\u2009\u00b5g/body clodronate or control liposomes (FormuMax Scientific, Sunnyvale, CA, USA) 24\u2009h prior to surgical interventions.\n\nStudies with primary hepatocytes {#Sec19}\n--------------------------------\n\nPrimary hepatocytes were isolated by a collagenase perfusion method^[@CR44]^ from 10-week-old iFoxM1LKO and control male mice. After making an abdominal incision, the portal vein was cannulated and the inferior vena cava was cut, the liver was then perfused with Liver Perfusion medium (Thermo Fisher Scientific, Waltham, MA, USA) at 5\u2009ml/min for 9.5\u2009min, followed by perfusion with HEPES buffer containing 0.5\u2009mg/ml collagenase (Sigma) at 5\u2009ml/min for an additional 9.5\u2009min. The liver was excised and dissociated in DMEM with 10% FBS and P/S, filtered through a 70-\u00b5m nylon mesh, and centrifuged at 50*g* for 5\u2009min three times to clear the non-parenchymal fraction. Then, fractions containing hepatocytes were suspended with 35% Percoll (GE Healthcare, UK) solution and centrifuged at 60*g* for 10\u2009min. The pellets were washed and suspended with DMEM and plated onto 96-well plates with 2.5\u2009\u00d7\u200910^4^ cells per well. At 48\u2009h after isolation, hepatocytes were incubated with 100\u2009ng/ml recombinant murine IL-6 (PEPROTECH, London, UK) for 6\u2009h, and then collected for RNA extraction. In experiments employing ex vivo carbachol treatment, to avoid macrophage contamination, hepatocytes were incubated with 1\u2009mM clodronate liposomes for 1\u2009h starting at 24\u2009h after isolation. At 24\u2009h after starting this treatment with clodronate liposomes, the hepatocytes were stimulated with 100\u2009\u00b5M carbachol for 4\u2009h. In STAT3 inhibition experiments, at 48\u2009h after isolation, hepatocytes were incubated with 100\u2009\u00b5M cell-permeable STAT3 inhibitor peptide (EMD Millipore, Billerica, MA, USA) for 1\u2009h, and then stimulated with recombinant murine IL-6 for 6\u2009h.\n\nRNA extraction and quantitative real-time PCR {#Sec20}\n---------------------------------------------\n\nTotal RNA was extracted from isolated macrophages, hepatocytes, and Hepa1--6 cells using an RNeasy Micro Kit (Qiagen, Hilden, Germany), and from the liver using an RNeasy Mini Kit (Qiagen). cDNA synthesis was performed with a QuantiTect Reverse Transcription Kit (Qiagen) using 100\u2009ng of RNA from cultured cells and 1\u2009\u00b5g of total RNA from the liver. Real-time PCR was performed using the Light Cycler Quick System 350S (Roche Diagnostic, Mannheim, Germany)^[@CR45]^. The relative amount of mRNA was calculated with \u03b2-actin mRNA serving as the invariant control. The oligonucleotide primers are presented in Supplementary Table\u00a0[1](#MOESM1){ref-type=\"media\"}.\n\nHistological analysis {#Sec21}\n---------------------\n\nExcised liver specimens were fixed in 10% formalin, embedded in paraffin, and sectioned. Sections were stained with hematoxylin and eosin. For bromodeoxyuridine (BrdU) in situ detection, mice were injected intraperitoneally with 1\u2009mg of BrdU (BD Bioscience, San Jose, CA, USA) diluted to 10\u2009mg/ml with phosphate-buffered saline 2\u2009h before liver extraction. The labeled cells were immunostained with anti-BrdU antibody (51-75512\u2009L, BD Bioscience). The number of BrdU-positive nuclei per 1000 hepatocytes per liver specimen was counted. For detection of macrophages, liver sections were stained with anti-F4/80 antibody (12-4801-80, Affymetrix eBioscience, Santa Clara, CA, USA).\n\nImmunoblotting {#Sec22}\n--------------\n\nLiver samples were homogenized in lysis buffer containing 100\u2009mM Tris, pH 8.5, 250\u2009mM NaCl, 1\u2009mM EDTA, 1\u2009mM phenylmethylsulfonyl fluoride, aprotinin at 1:5000 dilution, and leupeptin at 1:5000 dilution^[@CR46]^. Tissue homogenates were centrifuged and the supernatants including tissue protein extracts were boiled in Laemmli buffer containing 10\u2009mM dithiothreitol, then subjected to SDS-polyacrylamide gel electrophoresis. Separated proteins were transferred to nitrocellulose membranes and blocked in Tris-buffered saline containing 3% FBS. Immunoblot analyses were performed using antibodies to phospho-STAT3 (Tyr705) (\\#9131, Cell Signaling Technology, Danvers, MA, USA), total STAT3 (\\#4904, Cell Signaling Technology), phospho-STAT1 (Tyr701) (\\#9167, Cell Signaling Technology), phospho-STAT5 (Tyr694) (\\#9314, Cell Signaling Technology), FoxM1 (13147-1-AP, Proteintech, Rosemont, IL, USA), Cyclin A2 (ab181591, Abcam, Cambridge, UK), Cdk1 (\\#28439, Cell Signaling Technology), and PLK1 (\\#4535, Cell Signaling Technology) at 1:2000 dilution and actin (A2066, Sigma) at 1:5000 dilution. Quantitative data were obtained employing a ChemiDoc Touch Imaging System (Bio-Rad Laboratories, Hercules, CA, USA). Uncropped membrane images are presented in Supplementary Fig.\u00a0[6](#MOESM1){ref-type=\"media\"}.\n\nPlasma IL-6 measurement {#Sec23}\n-----------------------\n\nBlood samples were collected from tail veins. Plasma IL-6 concentrations were measured using a Mouse IL-6 ELISA kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions.\n\nIn vivo IL-6 neutralization {#Sec24}\n---------------------------\n\nWe administered 500\u2009\u00b5g/body neutralizing antibodies to IL-6 and control rat IgG (Bio X Cells, West Lebanon, NH, USA) intraperitoneally 1\u2009hr before PHx.\n\nStatistical analysis {#Sec25}\n--------------------\n\nAll data are presented as means\u2009\u00b1\u2009standard error. For experiments in which data differences among three or four groups needed to be assessed, we used one-way ANOVA followed by Bonferroni's post hoc test. In experiments in which data differences between two groups were assessed, results were analyzed using the unpaired *t* test. Survival curves were estimated using the Kaplan--Meier method and differences in survival among the groups were tested by applying the log-rank test with Bonferroni correction. Differences were considered to be significant at *p*\u2009\\<\u20090.05. No statistical method was used to predetermine sample size. Most sample sizes were chosen based on data from previous publications. Statistical analyses were performed assuming a normal distribution in all experiments.\n\nSupplementary information\n=========================\n\n {#Sec26}\n\nSupplementary Information Peer Review File Reporting Summary\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nElectronic supplementary material\n=================================\n\n**Supplementary Information** accompanies this paper at 10.1038/s41467-018-07747-0.\n\nProf. Pradip Raychaudhuri of the University of Illinois at Chicago contributed to generation of the FoxM1-floxed mice. Prof. Pierre Chambon and Prof. Daniel Metzger of the Institute of Genetics and Molecular and Cellular Biology, Illkirch-Cedex contributed to generation of the SA-CreER^T2^ mice. This work was supported by Grants-in-Aid for Scientific Research to H.K. (17H01565 and 16K15486), J.I. (17K09816), and T.I. (18K16221) from the Japan Society for the Promotion of Science, the CREST to H.K. (17gm0610001h0006) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, Takeda Science Foundation to J.I. This research was also supported by the Japan Agency for Medical Research and Development, AMED, under Grant Numbers JP17gm0610001 and JP17gm5010002 to H.K. and the PRIME to J.I. (18gm6210002h0001). We thank T. Takasugi, J. Fushimi, H. Hatakeyama, and A. Iwama (all belong to the Department of Metabolism and Diabetes, Tohoku University Graduate School of Medicine) for technical support.\n\nT.I. and J.I. conducted the research and obtained the data, contributed to relevant discussions, wrote the manuscript, and reviewed/edited the paper. J.Y., Y.K., A.E., H.S., M.K., Y.A., K.T., S.K., K.K., J.G., K.U., S.S., Y.I. and T.Y. contributed to the relevant discussions. V.V.K. provided FoxM1-floxed mice. H.K. contributed to the relevant discussions, writing the paper, and reviewing/editing the paper.\n\nThe data that support the findings of this study are available from the corresponding author upon reasonable request. A Reporting Summary for this Article is available as a Supplementary Information file.\n\nCompeting interests {#FPar1}\n===================\n\nThe authors declare no competing interests.\n"}
+{"text": "1.. Introduction\n================\n\nNumerous cell signaling pathways are regulated by the ubiquitin-proteasome system. Ubiquitination of the target protein generally requires the joint action of three enzymes. First, E1 activates the ubiquitin molecule; E2 then binds the ubiquitin molecule; and finally, an E3 ubiquitin ligase binds to a specific substrate and E2 covalently transfers the ubiquitin protein to one or more lysine residues of the target protein ([@b1-ol-0-0-11728]--[@b3-ol-0-0-11728]). Thousands of ubiquitination processes are ongoing in a cell at any moment ([@b4-ol-0-0-11728]). The high specificity of this degradation mechanism targeting specific proteins is determined mainly by the E3 ubiquitin ligase ([@b5-ol-0-0-11728]). Therefore, the E3 ligase plays a key role in degradation of the target protein ([@b6-ol-0-0-11728]). E3 ubiquitin ligases with a RING domain comprise a large group of E3 ubiquitin ligases that is responsible for \\~20% of ubiquitin-mediated protein degradation ([@b6-ol-0-0-11728]); additionally, degradation events regulated by Cullin RING ligases (CRLs), which are composed of Cullin (Cul) proteins, account for a large proportion ([@b7-ol-0-0-11728]). Previous studies have shown that CRLs not only are the primary factors in protein degradation but also are involved in the initiation and progression of some cancers including colorectal cancer, cholangiocarcinoma, etc ([@b8-ol-0-0-11728],[@b9-ol-0-0-11728]).\n\nThe Cul protein family was first reported in 1996 to form an active complex that can regulate the cell cycle ([@b10-ol-0-0-11728]). The human genome contains 7 Cul proteins, including, Cul1, Cul2, Cul3, Cul4A, Cul4B, Cul5, Cul7 and Cul9 ([@b11-ol-0-0-11728]). All six have conserved, homologous Cul domains, which function to bind the subunit RING domain proteins RING-box protein 1 (Rbx)1 or Rbx 2 in the whole complex. In addition to Cul7 and Cul9, five other members have three serial N-terminal Cul repeats, which are used to recognize subunits in the E3 ubiquitin ligase complex. Since it was discovered and defined nearly 20 years ago, subsequent studies have revealed that the multipart E3 ligase complex consisting of this family of proteins has an intricate structure and serves important functions in the cell cycle, signal transduction, cell development and other physiological processes ([@b12-ol-0-0-11728],[@b13-ol-0-0-11728]).\n\nAmong the CRLs, CRL1 \\[also known as S phase kinase- associated protein 1-Cul 1-F-box (SCF)\\] has been extensively studied. The SCF complex is composed of Cul1, Rbx1, the linker protein Skp1 and variable F-box (FBX) proteins. Within the SCF structure, differences in the FBX proteins, determines the specificity of substrate binding to accomplish the degradation of different substrates. To date, 69 FBX proteins have been found and identified in the human genome, and most can form the CRL1 complex ([@b14-ol-0-0-11728]). FBX proteins can be roughly divided into three types according to the other domains present: i) FBXW, which contain a tryptophan-aspartic acid 40 (WD40) repeat domain; ii) FBXL, which contain a leucine-rich repeat domain; and iii) FBXO, which contain other domain motifs. Accumulating evidence has indicated that dysregulation of FBX proteins is involved in the development, angiogenesis, proliferation and metastasis of a number of malignancies ([@b9-ol-0-0-11728]). F-box/WD repeat-containing protein 7 (FBXW7) is also called AGO, hCDC4 and SEL-10; SEL-10 was first identified from yeast, AGO was first found in *Drosophila*, and CDC4 is a yeast gene ([@b15-ol-0-0-11728]). The SCF E3 ubiquitin ligase complex contains FBXW7 which targets several important oncoproteins including c-Jun, c-Myc and Notch1 etc. for ubiquitylation ([Fig. 1A](#f1-ol-0-0-11728){ref-type=\"fig\"}). The human FBXW7 gene is located on chromosome 4 (4q31.3) and is mainly expressed in the brain, heart and testes, but at lower levels in the liver, lungs and kidneys ([@b16-ol-0-0-11728]).\n\nThe FBXW7 gene produces three mRNA transcripts (FBXW7\u03b1, FBXW7\u03b2 and FBXW7\u03b3) through selective splicing following transcription. The tissue expression and subcellular localization differ among the three subtypes; FBXW7\u03b1 mRNA is expressed in almost all tissues, FBXW7\u03b2 mRNA can be detected only in the brain and testes, and FBXW7\u03b3 mRNA mainly exists in the heart and skeletal muscle ([@b17-ol-0-0-11728]). FBXW7\u03b1 and FBXW7\u03b2 are localized in the cytoplasm, whereas FBXW7\u03b3 is in the nucleolus ([@b18-ol-0-0-11728]). FBXW7\u03b1 plays the most important role among the three subtypes and can occasionally replace the other two ([@b19-ol-0-0-11728]). However, the \u03b2 and \u03b3 subtypes may also play unique roles under specific conditions, as their abnormal localization may cause cell functional defects ([@b20-ol-0-0-11728]). The human FBXW7\u03b1 contains 707 amino acid residues. The N-terminus of FBXW7\u03b1 contains an FBX motif composed of \\~40 amino acids, which mainly contributes to ubiquitin ligase activity ([@b21-ol-0-0-11728]). In addition, FBXW7\u03b1 has a seven-tandem repeat WD40 domain that is necessary for substrate recognition, and a dimerization domain that is associated with the binding of FBXW7 to the substrate ([Fig. 1B](#f1-ol-0-0-11728){ref-type=\"fig\"}). FBXW7 recognizes and binds substrates with the specific phosphorylated amino acid sequence, (L)-X-pT/pS-P-(P)-X-pS/pT/E/D (where X can be any amino acid residue) ([@b22-ol-0-0-11728]). In numerous cases, the conserved Cdc4 phosphodegron (CPD) motif of FBXW7 substrates is phosphorylated by glycogen synthase kinase 3 (GSK3), which initiates FBXW7-mediated substrate ubiquitination ([@b23-ol-0-0-11728],[@b24-ol-0-0-11728]).\n\nFBXW7 is generally recognized as a cancer suppressor that regulates the ubiquitination and degradation of various proliferation- and survival-related proteins, including cyclin E ([@b25-ol-0-0-11728]), Notch ([@b15-ol-0-0-11728]), c-Myc ([@b26-ol-0-0-11728]) and mammalian target of rapamycin (mTOR) ([@b27-ol-0-0-11728]). The mutation rate of FBXW7 in colorectal cancer ([@b28-ol-0-0-11728]), breast cancer ([@b29-ol-0-0-11728]) and other malignant tumors is \\~10%, whereas in some types of leukemia, the mutation rate is as high as 60% ([@b30-ol-0-0-11728]). Moreover, the mutational status of FBXW7 is directly related to the therapeutic effect of chemotherapeutic drugs ([@b31-ol-0-0-11728]). In addition, FBXW7 plays a role in embryonic development, as mice with FBXW7 knockout mutation die on day 11, the middle stage of embryonic development, mainly because of a developmental deformity of the digestive organs ([@b32-ol-0-0-11728]). FBXW7 also serves an important role in the proliferation and maturation of stem cell populations ([@b33-ol-0-0-11728]).\n\n2.. Transcriptional and translational regulation of FBXW7 expression\n====================================================================\n\n### Transcriptional regulation\n\nOn the one hand, FBXW7 affects the ubiquitination state and protein content of various substrates; on the other hand, this process is tightly regulated from DNA to protein expression ([Fig. 1B](#f1-ol-0-0-11728){ref-type=\"fig\"}) ([@b21-ol-0-0-11728]). FBXW7 expression can be regulated at a number of levels, including at the transcriptional level through the regulation of multiple transcription factors. FBXW7\u03b1 is negatively regulated by CCAAT/enhancer-binding protein \u03b4 (CEBP\u03b4), a transcription factor involved in adipocyte differentiation as well as inflammation reactions ([@b34-ol-0-0-11728]). In mammary tumors, CEBP\u03b4 expression is induced under a hypoxic microenvironment, which can lead to tumor metastasis by directly binding to the promoter of FBXW7 to inhibit its expression ([@b35-ol-0-0-11728],[@b36-ol-0-0-11728]). Additionally, presenilin, a regulator of Notch processing and \u03b2-catenin signaling, can also indirectly downregulate FBXW7\u03b1 mRNA expression ([@b37-ol-0-0-11728]). FBXW7\u03b2 and FBXW7\u03b3 are upregulated in a p53-dependent manner, and their upregulation is required for the apoptotic response and tumor suppression induced by p53 ([@b38-ol-0-0-11728]).\n\n### Translational regulation\n\nA number of previous studies have indicated that the direct binding of multiple non-coding microRNAs (miRNAs) to the 3\u2032 untranslated region of the mRNA can prevent protein translation of FBXW7 ([@b39-ol-0-0-11728],[@b40-ol-0-0-11728]). Overexpression of miRNA (miR)-223 in T cell acute lymphoblastic leukemia (T-ALL), colorectal cancer and gastric cancer has been reported to downregulate FBXW7 ([@b39-ol-0-0-11728],[@b40-ol-0-0-11728]). In T-ALL, high levels of miR-223 promote the proliferation of tumor cells, and its inhibition increases the sensitivity to \u03b3-secretase inhibitor drugs ([@b39-ol-0-0-11728]). FBXW7 is also repressed by miR-27a overexpression in an adenomatous polyposis coli protein (APC) mutation-mediated murine model of colorectal adenocarcinoma and in human high-grade colorectal adenocarcinomas associated with preinvasive adenomas ([@b41-ol-0-0-11728],[@b42-ol-0-0-11728]). Inhibition of miR-27a increases FBXW7 expression and downregulates FBXW7 substrates, which delays tumor formation in model systems and inhibits the proliferation of colorectal cancer cells ([@b43-ol-0-0-11728]). miR-92a is reported to decrease the expression levels of FBXW7 mRNA and protein, and to increase c-Myc expression, which facilitates Myc-mediated apoptosis and proliferation in a model of B-cell lymphoma ([@b44-ol-0-0-11728]). Knockdown of miR-92a suppresses cancer cell invasion and proliferation through the upregulation of FBXW7 ([@b45-ol-0-0-11728]).\n\nAdditional miRNAs, including miR-548, miR-544a, miR-367, miR-182, miR-503, miR-155-3p and miR-32, have also been demonstrated to modulate FBXW7 activity via various mechanisms ([@b46-ol-0-0-11728]--[@b51-ol-0-0-11728]). Although individual miRNAs are weakly related to FBXW7 expression in patients with cancer, suggesting that more than one miRNA is involved in the regulation of FBXW7, miRNA-mediated suppression of FBXW7 synchronously targets the three FBXW7 isoforms, which may influence additional FBXW7 substrates ([@b52-ol-0-0-11728]).\n\n### Epigenetic regulation\n\nThe transcription and translation of FBXW7 are regulated by epigenetics; specifically, histone modifications have been reported to regulate FBXW7. For example, the histone methyltransferase enhancer of zeste homolog 2 (EZH2), is associated with epigenetic inactivation of genes, including FBXW7; EZH2 promotes trimethylation of histone H3 Lys27 residue of FBXW7, which results in inactivation of FBXW7 gene function ([@b53-ol-0-0-11728]). Notably, FBXW7 has been reported to target EZH2 for ubiquitination and degradation in pancreatic cancer cells and to be negatively associated with the expression of EZH2 in human pancreatic cancer samples ([@b54-ol-0-0-11728]). In addition to histone modifications, DNA modifications are reported to regulate FBXW7 expression. In contrast to FBXW7\u03b1, which exhibits ubiquitous expression in cell lines and a broad tissue distribution, FBXW7\u03b2 is expressed in specific cell lines and tissues ([@b55-ol-0-0-11728]). Histone and DNA modifications have been reported to epigenetically regulate the FBXW7\u03b2 promoter, which is methylated in 51% of primary breast cancer tumors and 43% of 60 human cancer cell lines originating from brain, breast, kidney, prostate, cervix, blood, skin, lung, thyroid and bone ([@b56-ol-0-0-11728]). The expression of the FBXW7\u03b2 gene is negatively associated with its methylation level ([@b56-ol-0-0-11728]). Patients with lymph node-positive breast cancer with higher FBXW7 methylation levels have longer overall survival times, although methylation of FBXW7 correlates with high-grade tumors ([@b56-ol-0-0-11728]). In addition, patients with ovarian cancer with p53 mutations have been reported to exhibit lower FBXW7 expression compared with those with wild-type p53, and hypermethylation of the FBXW7 promoter associated with mutations in p53 leads to decreased FBXW7 expression ([@b57-ol-0-0-11728]).\n\n3.. Post-translational modifications of FBXW7\n=============================================\n\n### Autoubiquitination\n\nPost-translational modifications of FBXW7 are involved in its autoubiquitination, deubiquitination, dimerization and localization. In addition to targeting substrates for ubiquitination and degradation, FBXW7 can also be regulated by autoubiquitination. Peptidyl-prolyl *cis*-*trans* isomerase NIMA-interacting 1 (Pin1) has been reported to destabilize and downregulate FBXW7 by mediating a decrease in dimerization and promoting FBXW7 self-ubiquitination and degradation ([@b58-ol-0-0-11728]). In addition, phosphorylation at Thr205 by extracellular signal-regulated kinase and at Ser176 by polo-like kinase 2 leads to destabilization and degradation ([@b59-ol-0-0-11728],[@b60-ol-0-0-11728]). By contrast, phosphorylation mediated by serum- and glucocorticoid-inducible kinase 1 (SGK1) and phosphoinositide 3-kinase at Ser227 stabilizes FBXW7 but increases ubiquitination of cyclin E, Notch and Myc ([@b53-ol-0-0-11728],[@b54-ol-0-0-11728]). FBXW7 stability can also be controlled by SCF-dependent mechanisms. For instance, COP9 signalosome complex subunit 6, a member of the COP9 signalosome complex, increases FBXW7 autoubiquitination and proteasome-mediated degradation by regulating Cul1 neddylation ([@b61-ol-0-0-11728]).\n\n### Deubiquitination\n\nAutoubiquitination and degradation of FBXW7 can be reversed by the deubiquitinating enzyme 28 (USP28). Overexpression of USP28 not only allows the degradation of several FBXW7 substrates, inhibits progenitor cell proliferation and delays tumor formation ([@b62-ol-0-0-11728]), but also represses autocatalytic ubiquitination of FBXW7 ([@b63-ol-0-0-11728]). Genetic ablation of USP28 negatively regulates FBXW7 and its substrates in the pancreas, liver and lungs. In addition, abrogation of USP28 facilitates the transformation of mouse fibroblasts due to FXBW7 destabilization ([@b63-ol-0-0-11728]).\n\n### Dimerization\n\nFBXW7 is characterized by its ability to form dimers through a conserved D domain. Previous studies of endogenous FBXW7 with mutations preventing dimer formation have revealed that dimerization facilitates ubiquitination of FBXW7 substrates with low-affinity degrons, but it is unimportant to substrates with high-affinity degrons, including cyclin E and Myc ([@b64-ol-0-0-11728],[@b65-ol-0-0-11728]). Dimer-deficient mutants of CDC4 in yeast variant show increased autocatalytic ubiquitination and instability *in vivo* ([@b66-ol-0-0-11728]). In addition, isomerization mediated by Pin1 and phosphorylation of human FBXW7 at Ser205 repress its dimerization and enhance its autoubiquitination ([@b58-ol-0-0-11728]).\n\n### Location\n\nAbnormal cellular distribution of FBXW7 impairs its interaction with substrates. For example, nucleophosmin is necessary for the nucleolar localization of FBX7\u03b3 and is often mutated in acute myelogenous leukemia, leading to FBX7\u03b3 instability and an increase in c-Myc expression ([@b67-ol-0-0-11728]). Additionally, phosphorylation of FBXW7\u03b1 at Ser10 inhibits one of its nuclear localization signals ([@b68-ol-0-0-11728]).\n\n4.. Genetic alterations of FBXW7 in cancers\n===========================================\n\nThe important role of FBXW7 in tumor suppression has been confirmed by different genetic alterations of FBXW7 in various human cancers. Functional silencing of FBXW7 by deletion, mutation and hypermethylation ultimately leads to tumorigenesis and cancer progression ([@b69-ol-0-0-11728]). Although rare, FBXW7 mutations occur often in breast, pancreatic, gastric and cervical cancers but rarer promoter hypermethylation and genetic deletion of FBXW7 occurs in bladder, cervical and breast cancers ([@b65-ol-0-0-11728]). However, missense point mutations are the most common type of genetic alterations in FBXW7 and are observed on three arginine residues (R465, R479 and R505) at the \u03b2 propeller ([@b65-ol-0-0-11728]). FBXW7 is usually expressed as a functional wild-type protein because the second wild-type allele is retained. Consistent with this observation, mouse models with monoallelic deletion of FBXW7 show a milder tumor phenotype compared with biallelic gene deletions ([@b70-ol-0-0-11728]). Therefore, biallelic FBXW7 mutations are assumed to silence the function of the wild-type protein as dominant negative alleles ([@b65-ol-0-0-11728]).\n\nIn the intestine and the hematopoietic system, knock-in mice with a heterozygous FBXW7 mutation show accelerated tumorigenesis compared to mice with heterozygous wild-type FBXW7 (FBXW7^+/\u2212^) ([@b70-ol-0-0-11728],[@b71-ol-0-0-11728]). The hematopoietic stem cells in FBXW7^Mut/+^ mice show a significant increase in Myc but do not exhibit the characteristic hyperproliferative phenotype of those in FBXW7^\u2212/\u2212^ animals ([@b70-ol-0-0-11728]). Mice with T-cell leukemias induced by an activated Notch allele show accumulation of sterol regulatory element-binding protein 1 and Myc ([@b70-ol-0-0-11728]). Moreover, additional depletion of p53 does not promote the onset of disease, indicating the functional difference between complete loss and mutation of FBXW7 ([@b70-ol-0-0-11728]). By contrast, the observation that only Kr\u00fcppel-like factor 5 (KLF5) and homeobox protein TGIF1 (TGIF1), instead of most tested FBXW7 substrates, are substantially affected by mutated FBXW7, reveals that the influence of FBXW7 mutations on substrate regulation is strongly context dependent ([@b71-ol-0-0-11728]). Although FBXW7 mutations enhance intestinal tumorigenesis driven by the multiple intestinal neoplasia, a mutant allele of APC, the levels of Notch, Jun and Myc remain normal in these mice with developing adenomas ([@b71-ol-0-0-11728]). Therefore, heterozygous mutations in FBXW7 may promote tumorigenesis by the regulation of non-canonical substrates such as TGIF1 and KLF5.\n\n5.. Substrates and mechanisms of FBXW7 involved in tumor suppression\n====================================================================\n\nUbiquitin ligases play a biological role by affecting the expression levels of specific target proteins. The most common mechanism by which ubiquitin ligases participate in cancer processes is through regulating the content of cell cycle-related factors. FBXW7 can effectively recognize a variety of substrates, such as cyclin E, KLF5, mTOR, Aurora A, c-Myc, c-Jun and induced myeloid leukemia cell differentiation protein Mcl-1 (MCL-1). Most of these protein substrates are involved in the regulation of cell cycle processes or homeostasis. Additionally, they are the expressed products of oncogenes or potential oncogenes ([Fig. 2](#f2-ol-0-0-11728){ref-type=\"fig\"}; [Table I](#tI-ol-0-0-11728){ref-type=\"table\"}).\n\n### Cyclin E\n\nCyclin E binds and activates the cell cycle-dependent protein kinase cyclin-dependent kinase 2 (CDK2) to promote transition of the cell cycle from G~1~ to S phase ([@b72-ol-0-0-11728]). Increased expression of CDK2 can cause chromosomal instability and accelerate the occurrence of cancer ([@b73-ol-0-0-11728]). Cyclin E is expressed mainly in tissues exhibiting vigorous cell division, in direct contrast with FBXW7, which is expressed mainly in non-proliferative tissues ([@b74-ol-0-0-11728]). Cyclin E has a typical CPD, in which phosphorylation of Thr380 and Ser384 is necessary for degradation ([@b32-ol-0-0-11728]). FBXW7 can catalyze the degradation of cyclin E2 in addition to E1, the most common cyclin E subtype. Cyclin E2 contains two typical CPDs, that are specifically recognized by FBXW7, and the double-site phosphorylation of Thr392 and Ser396 initiates ubiquitination and proteasomal degradation ([@b74-ol-0-0-11728]), explaining why the decrease in FBXW7 expression and the increase in cyclin E in primary invasive breast cancer lead to the appearance of a large number of cells with chromosomal polyploidy ([@b75-ol-0-0-11728]).\n\n### Aurora A\n\nAurora A (also known as serine/threonine protein kinase 15) serves a role in mitosis and meiosis by regulating the phosphorylation of specific substrates, and its activity is the highest during G~2~/S phase transition ([@b76-ol-0-0-11728]). Aurora A is overexpressed in a variety of tumors, resulting in abnormal centrosome expansion, increased chromosomal instability and eventual carcinogenic transformation ([@b76-ol-0-0-11728],[@b77-ol-0-0-11728]). FBXW7-deficient HCT116 and HeLa cells exposed to vincristine, paclitaxel and spindle toxin exhibit extensive mitotic delay and nuclear duplication, which are important causes of polyploidy. Loss of FBXW7 can increase the content of cyclin E and Aurora A, but the increase of single cyclin E or Aurora A could not cause drug-induced polyploidy. This observation shows that the increase in cyclin E and Aurora A is a cause of polyploidy ([@b78-ol-0-0-11728]).\n\n### Notch\n\nNotch is a highly conserved signaling system in multicellular organisms, playing an important role in cell proliferation, differentiation and apoptosis ([@b79-ol-0-0-11728]). In mammals there are four isoforms of Notch receptors, all of which are single-pass transmembrane receptors in which the N-terminus is located outside the cell and accounts for most of the structure and only a small part (C-terminus) is inside the cell ([@b80-ol-0-0-11728]). When a ligand binds to the extracellular domain of the Notch receptor, it can induce a proteolysis reaction and release the intracellular domain ([@b81-ol-0-0-11728]). The intracellular domain of the Notch receptor acts as a transcription factor to regulate the expression of specific genes ([@b81-ol-0-0-11728]). Overactivation of the Notch signaling pathway can cause abnormal cell proliferation and cancer ([@b82-ol-0-0-11728]). Ubiquitination of the intracellular domain of Notch 1 and Notch 4 by FBXW7 significantly weakens Notch signal transduction, whereas inhibition of FBXW7 enhances Notch-mediated activation of downstream signaling ([@b83-ol-0-0-11728]). FBXW7 gene knockout results in the abolishment of Notch 4 signal transmission, which can lead to abnormal development of blood vessels in mouse embryos and may result in death at \\~11 days ([@b84-ol-0-0-11728]). Similarly, mice with brain-specific FBXW7 knockout succumb after birth owing to the accumulation of Notch 1 and Notch 3 proteins in the brain, which results in increased expression of target genes and abnormal differentiation of neural stem cells, causing abnormal brain development and morphology in these mice ([@b85-ol-0-0-11728]).\n\n### MCL-1\n\nMCL1 is an antiapoptotic protein of the BCL2 family that can promote cancer development by reducing apoptosis ([@b86-ol-0-0-11728]). A number of hematological tumors, such as B-cell lymphoma and chronic myeloid leukemia, exhibit an abnormal increase in MCL1 expression, which is considered an important cause of chemotherapeutic resistance. In normal cells, the half-life of the MCL1 protein is short, and it is easily degraded by ubiquitination modification. However, in tumor cells, the MCL1 protein content is increased, although a detailed understanding of this mechanism is lacking. MCL1 can be phosphorylated by GSK3, which initiates ubiquitination and degradation of MCL1 ([@b86-ol-0-0-11728]). Loss of FBXW7 in a human T-ALL cell line is accompanied by an increase in MCL1 content. This cell line is sensitive to a variety of kinase inhibitors, such as sorafenib but is resistant to ABT-737, an inhibitor of BCL2 ([@b22-ol-0-0-11728]). When FBXW7 function is restored or MCL1 is lost, sensitivity to ABT-737 can be restored. Therefore, these findings confirm that the increase in the MCL1 protein content caused by FBXW7 deficiency is a mechanism underlying tumor chemoresistance ([@b22-ol-0-0-11728]). In addition, paclitaxel treatment induces phosphorylation modification of MCL1, which can be recognized by FBXW7, leading to MCL1 ubiquitination and degradation, consequently increasing apoptosis ([@b31-ol-0-0-11728]). By contrast, when FBXW7 is inactivated or expression of FBXW7 is decreased, the protein stability of MCL1 is increased. Correspondingly, resistance to microtubule-targeted drugs, such as paclitaxel, is increased in tumor patients, and the its chemotherapeutic effect is significantly reduced ([@b31-ol-0-0-11728]).\n\n### c-Myc\n\nc-Myc is an important oncogenic protein that can regulate cell growth and division, serving a number of roles in human cancer. In lymphoid tumor cell lines, mutation of the Thr58 site in c-Myc is the most frequent mutation and results in the failure of FBXW7 to regulate c-Myc protein content, accumulation of c-Myc protein and eventual tumor development ([@b87-ol-0-0-11728]). In addition to GSK3, NEMO-like kinase (NLK) also regulates the c-Myc protein content. NLK can directly bind to c-Myc and catalyze the phosphorylation of multiple C-terminal sites. This modification promotes ubiquitination and proteasomal degradation of FBXW7. Mutation of these sites can abrogate the c-Myc-FBXW7 interaction and protein ubiquitination ([@b88-ol-0-0-11728]).\n\nAbnormal localization of c-Myc proteins can cause its accumulation and may lead to tumorigenesis ([@b63-ol-0-0-11728]). USP28 can bind to FBXW7 and inhibit ubiquitination modification of c-Myc by the latter, thus increasing the protein stability of c-Myc ([@b89-ol-0-0-11728]). DNA damage caused by ultraviolet radiation can reduce c-Myc protein content due to dissociation of USP28 and FBXW7, thus increasing FBXW7-mediated ubiquitination and degradation of c-Myc protein. Because both cyclin E and c-Myc are positive regulators of the cell cycle, decreases in their levels can cause cell cycle exit. However, FBXW7 deficiency increases the protein levels of these two factors and promotes cell cycle re-entry and G~1~/S phase transition, which is conducive to cell division ([@b90-ol-0-0-11728]). This mechanism is an important reason for cancer driven by FBXW7 mutations.\n\n### mTOR\n\nmTOR is a protein kinase that promotes cell growth and division by regulating protein synthesis and cell autotropism. The increase in mTOR content and activity is a common feature of tumorigenesis, and mTOR is widely used as a drug target in tumor therapy ([@b27-ol-0-0-11728]). In human breast cancer cell lines and patients with primary breast cancer, FBXW7 can contain a gene deletion or a functional inactivation mutation, which leads to an increase in the mTOR protein content and activation of its downstream signaling pathway ([@b27-ol-0-0-11728]). The sensitivity of breast cancer cells harboring wild-type FBXW7 to rapamycin is significantly increased, which suggests that loss of FBXW7 function is a mechanism underlying the resistance of tumor patients to mTOR pathway inhibitors ([@b27-ol-0-0-11728]).\n\n### KLF5\n\nKLF5 is a development-related transcription factor that can promote the expression of multiple development-related genes and may serve a role in cell proliferation, cell cycle, apoptosis and cell migration and differentiation ([@b73-ol-0-0-11728]). KLF5 is overexpressed in various cancers and can promote proliferation and deterioration of breast cells ([@b91-ol-0-0-11728]). KLF5 is a short-lived protein that can be rapidly degraded by ubiquitination. KLF5 contains three CPDs recognized by FBXW7; simultaneous mutation of these sites can inhibit the interaction and ubiquitination of FBXW7 and KLF5, whereas the point mutations in FBXW7 can significantly delay degradation of the KLF5 protein, leading to accumulation of KLF5 in cells ([@b92-ol-0-0-11728]). The binding and ubiquitination of KLF5 and FBXW7 are dependent on phosphorylation of KLF5 at Ser303 by GSK3\u03b2 ([@b93-ol-0-0-11728]). In tumor cells, mutation or abnormal activation of FBXW7 results in a decrease in the ubiquitination and the protein degradation of KLF5. In turn, excessive accumulation of KLF5 promotes the occurrence of cancer by increasing the expression of target genes.\n\n6.. Therapeutic exploitation of FBXW7 signaling\n===============================================\n\nGiven the substantial tumor control exhibited by FBXW7 in mouse models, therapeutic exploitation of FBXW7 and its related pathways for several types of cancer has attracted intense interest. Various oncogenic transcription factors and oncoproteins are degraded through FBXW7-mediated ubiquitination. Most substrates with increased levels due to FBXW7 silencing mutations are oncoproteins. Such substrates are likely to alter regulation of the cell cycle, perturb stress response and rewire metabolic pathways ([@b28-ol-0-0-11728]). Previous studies have reported that the accumulation of c-Myc and MCL1 induce synthetic lethal interactions between non-functionalFBXW7 and tumor necrosis factor-like death ligands or mitotic inhibitors in tumor cells ([@b31-ol-0-0-11728],[@b94-ol-0-0-11728]). Similarly, FBXW7 mutant cells with increased c-Myc are more sensitive to suppression of additional enzymes, such as CDK1, and energy-sensing enzymes, such as AMP-activated protein kinase ([@b95-ol-0-0-11728],[@b96-ol-0-0-11728]). Alternatively, post-translational modifications are important for FBXW7 E3 ligase activity, suggesting that aberrant alteration of upstream signaling might impair the antitumor effect of FBXW7. One possibility is to target deubiquitinases, considering that USP36 and USP28 can antagonize FBXW7 activity. Growth of established tumors can be inhibited by acute decrease of USP28, showing that the development of tumors depends on USP28 ([@b56-ol-0-0-11728]), and that certain types of tumor could be effectively treated using inhibition of USP28 with small molecules. USP7 inhibitors successfully stabilize p53 and promote apoptosis in myeloma cells that are not sensitive to available therapies ([@b97-ol-0-0-11728]). Additionally, several studies have demonstrated that loss of p53, or mutation or deletion of FBXW7 promotes tumorigenesis ([@b10-ol-0-0-11728],[@b98-ol-0-0-11728]). Since frequent mutant FBXW7 alleles result in impaired substrate recognition, recovering substrate binding by FBXW7 may effectively inhibit tumor development. This may be potentially achieved in a manner analogous to p53, in which mutant forms of the protein can be reactivated by small molecules ([@b99-ol-0-0-11728]). Moreover, since mutant FBXW7 has the wild-type allele in most human tumors, several enzymes, such as Pin1 and SGK1, which modulate the activity of FBXW7, may provide a reasonable approach for tumor suppression by increasing FBXW7 activity ([@b53-ol-0-0-11728],[@b100-ol-0-0-11728]). Therefore, restoring the antitumor function of FBXW7 by blocking oncogenic upstream mediators may be an effective therapeutic strategy.\n\n7.. Conclusions\n===============\n\nUbiquitination of specific proteins plays an important role in tumor initiation, development, metastasis and chemoresistance. Currently, the drugs designed to target ubiquitin-modified molecules exhibit good potential in cancer treatment. FBXW7 has been suggested to have tumor-suppressive effects in various tumors. On the one hand, mutation of FBXW7 or a reduction in its activity promotes the occurrence and progression of tumors; on the other hand, it also increases the chemoresistance of tumors. These developments are important to improving our understanding of tumor occurrence mechanisms, the development of diagnostic reagents, and the optimization and design of therapeutic drugs. In the study of FBXW7, a series of important problems remain to be solved. For example, are there still unidentified FBXW7 substrates? Is there a network mechanism underlying the regulation of multiple substrates by FBXW7? What are the precise mechanisms of FBXW7 regulation? In conclusion, the present review examined the crucial role and molecular mechanism of FBXW7 in tumor inhibition, and may offer a putative therapeutic approach for multiple cancers.\n\nNot applicable.\n\nFunding\n=======\n\nThis work was jointly supported by The National Science Foundation of China (grant nos. 81871950, 81702871 and 81602085), The Shanghai Municipal Commission of Health and Family Planning (grant nos. 20154Y0090 and 2018YQ06) and The Shanghai Sailing Program (grant no. 16YF1401800).\n\nAvailability of data and materials\n==================================\n\nNot applicable.\n\nAuthors\\' contributions\n=======================\n\nXY, ZZ, XX, SJ and WX were involved in the conception of the review. ZZ, QH, WL and ML were involved in the writing of the article. QS, GF, ZY, and YQ were involved in critically revising and proofreading the manuscript. All authors have read and approved the final manuscript.\n\nEthics approval and consent to participate\n==========================================\n\nNot applicable.\n\nPatient consent for publication\n===============================\n\nNot applicable.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\n![Schematic illustration presenting the functional model and regulation of FBXW7. (A) Functional model of FBXW7. Schematic illustration showed that SCF E3 ubiquitin ligase complex containing FBXW7 can target several important oncoproteins including c-Jun, c-Myc, and Notch1 *et al* for ubiquitylation. (B) Regulation of FBXW7 from gene to protein. FBXW7 contains a D-domain for dimer formation and two key domains of F-box protein family including WD repeat and F-box for substrate binding and enzymatic activity. Regulatory mechanisms and related proteins that induce FBXW7 dysfunction at different levels. FBXW7, F-box/WD repeat-containing protein 7; RBX1, RING-box protein; KLF5, Kr\u00fcppel-like factor 5; E2, enzyme 2; ub, ubiquitin; Pi, phosphate; SKP1, S phase kinase-associated protein 1; CUL1, cullin 1; Hes-5, hairy and enhancer of split 5; C/EBP-\u03b4, CCAAT enhancer binding protein \u03b4.](ol-20-02-1526-g00){#f1-ol-0-0-11728}\n\n![Regulators and substrates of FBXW7 in human cancers. Several upstream proteins including EBP2, p53 and Numb-4 can positively regulate FBXW7 while other upstream regulators including Pin1, Hes-5 and C/EBP\u03b4 etc. can negatively regulate FBXW7. The specific substrates of FBXW7 including c-Myc, c-Jun and Mcl-1 can promote development of some tumors including lymphomas, intestinal cancer and hematological tumors. FBXW7, F-box/WD repeat-containing protein 7; EBP2, eIF4E-binding protein 2; C/EBP-\u03b4, CCAAT enhancer binding protein \u03b4; Pin1, peptidyl-prolyl *cis-trans* isomerase NIMA-interacting 1; Hes-5, hairy and enhancer of split 5; KLFs, Kr\u00fcppel-like factor; T-ALL, T cell acute lymphoblastic leukemia; Mcl-1, induced myeloid leukemia cell differentiation protein Mcl-1. Numb-4, NUMB endocytic adaptor protein.](ol-20-02-1526-g01){#f2-ol-0-0-11728}\n\n###### \n\nSubstrates targeted by F-box/WD repeat-containing protein 7.\n\n  Author, year               Substrate   Genomic location   Function/signalling pathways                                           Molecular size (amino acids)   Phospho-degron (phosphorylation sites)   Putative modifying kinase   (Refs.)\n  -------------------------- ----------- ------------------ ---------------------------------------------------------------------- ------------------------------ ---------------------------------------- --------------------------- -----------------------\n  Finkin *et al*, 2008       Aurora-A    20q13              Protein kinase; cell cycle                                             403                            LGTVYREL                                 GSK3                        ([@b78-ol-0-0-11728])\n  Klotz *et al*, 2009        Cyclin E1   19q12              Cyclin, cell cycle                                                     410                            LLTPPQSG                                 GSK3                        ([@b74-ol-0-0-11728])\n  Tsunematsu *et al*, 2004   Notch1      9q34               Transcription factor; Notch signalling pathway                         2,555                          FLTPSPES                                 Cdk8                        ([@b84-ol-0-0-11728])\n  Yang-Yen *et al*, 2006     Mcl-1       1q23               BCL2-family protein; cell survival                                     350                            IMSPEEEL, DGSLPSTP                       GSK3                        ([@b86-ol-0-0-11728])\n  Welcker *et al*, 2004      c-Myc       8q24               Transcription factor; cell proliferation                               439                            LPTPPLSP                                 GSK3                        ([@b87-ol-0-0-11728])\n  Mao *et al*, 2008          mTOR        1p36               Protein kinase; PI3K signalling pathway                                2,549                          LLTPSIHL                                 --                          ([@b27-ol-0-0-11728])\n  Liu *et al*, 2010          KLF5        13q22              Transcription factor; adipocyte differentiation and lipid metabolism   457                            PPSPPSSE, LNTPDLDM, NLTPPPSY             GSK3                        ([@b92-ol-0-0-11728])\n  Wei *et al*, 2005          c-Jun       1p32               Transcription factor; cell proliferation                               331                            GETPPLSP                                 GSK3                        ([@b23-ol-0-0-11728])\n\nKLF5, Kr\u00fcppel-like factor 5; GSK3, glycogen synthase kinase 3; Mcl-1, induced myeloid leukemia cell differentiation protein Mcl-1.\n\n[^1]: Contributed equally\n"}
+{"text": "Introduction {#Sec1}\n============\n\nCardiovascular diseases are currently one of the leading causes of death in the Western world. An abdominal aortic aneurysm (AAA) is defined as a weakening and dilatation of the abdominal aorta, prevailing in the infrarenal portion of the artery. The prevalence of AAA has been on the rise during the last decades as a result of demographic ageing, screening programs and improved clinical imaging techniques^[@CR1]^. The main risk of undetected, asymptomatic aneurysms is progressive expansion, followed by rupture, hemorrhage and death in approximately 80% of the cases^[@CR2]^. The development of AAA involves inflammation as a fundamental process. Chronic inflammation is characterized by the infiltration of inflammatory cell types, mainly macrophages and monocytes in the thrombus and throughout all layers of the aortic wall^[@CR3]^. These cells release several proteolytic enzymes such as matrix metalloproteinases, cytokines and oxidation-derived free radicals, leading to vascular smooth muscle cell apoptosis and degradation of the aortic tunica media^[@CR3]^. In AAA, elastolysis as a result of the chronic inflammation leads to reduced stability and consequently gradual dilatation of the aorta^[@CR4]^. An increase in the inflammatory response and extracellular matrix (ECM) degradation results in an increased risk of rupture^[@CR5]^.\n\nMagnetic nanoparticles (MNPs) represent a molecular imaging probe type, mainly used for magnetic resonance imaging (MRI). MNPs are comprised of small iron oxide crystals, typically surface-modified by coating (polysaccharides, polyethylene glycol) or capping (organic acids). Their sensitivity to external magnetic fields and small size, ranging from 20 to 150\u00a0nm makes them ideal candidates for tracking tumor cells, drug delivery and detecting endothelial inflammation^[@CR6],[@CR7]^. MNPs are internalized into macrophages/monocytes and represent a promising target for molecular imaging of inflammation and predicting aneurysm growth^[@CR8]--[@CR10]^. Ferucarbotran (RESOVIST) is a clinically approved MNP for MRI, designed for detecting liver lesions. It consists of magnetic nanoparticles (Magnetite--Fe~3~O~4~/Maghemite--Fe~2~O~3~) coated with carboxydextran. Previous studies have proven ferucarbotrans efficiency for imaging AAA in clinical studies^[@CR11]^ as well as in animal models^[@CR12]^. No adverse reactions are associated with rapid intravenous (i.v) injection of ferucarbotran^[@CR13]^.\n\nSo far, there are several different approaches for detecting inflammation in AAA. 18F-fluorodeoxyglucose (FDG) is a radiopharmaceutical tracer for positron emission tomography (PET) and PET computed tomography (CT), targeting high-glucose-using cells including macrophages in aneurysm inflammatory sites, yet it relies on ionizing radiation^[@CR14]^. In recent years, molecular MRI targeting macrophages and monocytes via various types of MNPs has gained momentum^[@CR9]^. However, the detection of MNPs in MRI is highly dependent on the surrounding tissue and can be hard to distinguish from other sources (e.g. air, imaging artifacts or pathological tissue changes). Additionally, a reference scan before MNP injection is required to perform quantification.\n\nMagnetic particle imaging (MPI) is a novel tomographic imaging modality for the highly sensitive detection and quantification of magnetic nanoparticles^[@CR15]^. First described by Gleich and Weizenecker in 2005 MPI is currently used in preclinical studies and is not yet in clinical practice. Opposed to MRI, where MNPs cause signal void, MPI detects MNP tracers directly, causing positive contrast without any background signal from the surrounding tissue. MPI images are comparable to the images known from\u00a0nuclear medicine imaging modalities such as PET and single-photon emission computed tomography (SPECT). Compared to MRI, MPI has a much higher sensitivity in detection of MNPs. MPI enables imaging and quantification of MNPs with higher specificity and without the need of additional measurements before injection of MNPs. MPI scanners have a spatial resolution in the millimeter range, which compares well with the resolution of clinical PET and SPECT. In contrast to PET or SPECT, MPI provides a much higher temporal resolution by using non-radiating tracers. Therefore, MPI makes a promising candidate for vascular imaging.\n\nIn this study, we assessed the potential of sensitive ex vivo MP imaging for the characterization of relevant parameters in AAA development and progression. A MPI-suitable MNP, ferucarbotran was used to evaluate the inflammatory processes in the aortic wall.\n\nResults {#Sec2}\n=======\n\nNo side effects or adverse reactions to the imaging agents were observed in the investigated animals. In the control group, consisting of sham-operated mice (n\u2009=\u20099) that received a continuous saline infusion for 28\u00a0days, AAA development was not observed. In the experimental group (n\u2009=\u200923), the continuous infusion of angiotensin II (Ang II) via osmotic minipumps led to the formation of suprarenal aortic aneurysms (Figs.\u00a0[1](#Fig1){ref-type=\"fig\"}, [2](#Fig2){ref-type=\"fig\"}). Animals developing no abdominal aneurysms were excluded from the study prior to data acquisition (n\u2009=\u20094).Figure 1Experimental setup. Native in vivo MR imaging (N\u2009=\u200932) was performed after three (group 1, n\u2009=\u20098) and four weeks (group 2, n\u2009=\u200915) of angiotensin II infusion. Following the scan, 50\u00a0\u00b5l dose of macrophage-specific iron-oxide particles (ferucarbotran, 46.66\u00a0\u00b5g iron per kg body weight) was administered via the tail vein. Ex vivo analysis (magnetic particle spectroscopy, magnetic particle imaging, histology, immunohistochemistry, laser ablation coupled to inductively coupled plasma-mass spectrometry) was performed 24\u00a0h after MNP administration. The control group consisted of apolipoprotein E-deficient mice (n\u2009=\u20099) implanted with osmotic minipumps filled with sodium chloride, serving as the control group. *MPI* magnetic particle imaging, *MPS* magnetic particle spectroscopy, *MNP* magnetic nanoparticles. Figure 2In vivo MRI of inflammatory activity during the development of aortic abdominal aneurysm compared to an animal from the control group. **(A1)** Time-of-flight angiogram showing the suprarenal abdominal aorta, including the right renal artery, of a male apolipoprotein E-deficient (Apo E \u2212/\u2212) mouse after four weeks of angiotensin II (Ang II) infusion. (**A2)** A pronounced dilatation of the arotic lumen was observed on the T1 weighted sequence after 4\u00a0weeks of angiotensin infusion. (**A3--A7)** Ex vivo histological measurements using EvG (**A3**), LA-ICP-MS (**A4**), HE (**A5**), Perls stain (**A6**) confirmed the in vivo findings. (**A4, A6, A7)** A strong correlation between the areas positive for iron-oxide particles in LA-ICP-MS (**A4**), Perls' stain (**A6**) and immunofluorescence for macrophage accumulation (**A7**) in corresponding histological sections was observed. The scale bars represent 100\u00a0\u03bcm. (**B1)** Time-of-flight angiogram showing the suprarenal abdominal aorta, including the right renal artery, of a male apolipoprotein E-deficient (Apo E \u2212/\u2212) control group mouse after four weeks of sodium chloride solution infusion. (**B2)** No dilatation of the aortic lumen was observed on the T1 weighted sequence after 4\u00a0weeks of sodium chloride solution infusion. (**B3--B6)** Ex vivo histological measurements using EvG (**B3**), HE (**B4**), Perls' Prussian Blue (**B5**) and immunofluorescence for macrophage accumulation **(B6)** in corresponding histological sections revealed neither MNP accumulation nor AAA development. The scale bars represent 100\u00a0\u03bcm. *TOF* arterial time of flight, *aA* suprarenal abdominal aorta, *rRA* right renal artery, *MRA* magnetic resonance angiography, *HE* hematoxylin--eosin-staining, *EvG* Miller's elastica van Gieson staining, *LA-ICP-MS* laser ablation coupled to inductively coupled plasma-mass spectrometry, *MNP* magnetic nanoparticles.\n\nMR angiography of abdominal aortic aneurysms {#Sec3}\n--------------------------------------------\n\nCross-sections of the abdominal aorta were assessed after 3 and 4\u00a0weeks of Ang II infusion (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The protocol included scans prior to administration of ferucarbotran. A significant aortic diameter increase was visible in T1 3D TOF (p\u2009\\<\u20090.05) (Figs.\u00a0[3](#Fig3){ref-type=\"fig\"}, [4](#Fig4){ref-type=\"fig\"}), while no difference was seen in the control animal group. The aortic diameter increased by 88% percent in the 3-week group and 175% in the 4-week group.Figure 3Correlation of in vivo MRI and ex vivo histological cross sectional AAA area measurements. To investigate the presence of AAA, in vivo MRI findings were compared to histological cross sections from the same region of the aorta. Time-of-flight angiogram detected the development of AAA in the experimental group. A strong correlation (R\u2009=\u20090.87) between the in vivo MRA and ex vivo histology images was shown. Overall, these measurements indicate an excellent agreement between in vivo and ex vivo measurements of the lumen dilatation in AAA. *MRA* magnetic resonance angiography. Figure 4In vivo MRI and ex vivo MPI of inflammatory-activity during the development of aortic abdominal aneurysm. (**A1)** Time-of-flight angiogram showing the suprarenal abdominal aorta of a male apolipoprotein E-deficient (Apo E \u2212/\u2212) mouse after four weeks of angiotensin II (Ang II) infusion; (**A2)**: a pronounced dilatation of the arotic lumen was observed on the T1 weighted sequence after 4\u00a0weeks of angiotensin infusion; (**A3)** ex vivo MPI of the AAA region of the same mouse; (**A4)**: ex vivo aortic MPI---in vivo whole body MRI signal manual fusion overlay based on anatomical landmarks; (**A5)** Perls' Prussian Blue; the scale bar represents 100\u00a0\u03bcm. *TOF* arterial time of flight. *aA* suprarenal abdominal aorta, *MRA* magnetic resonance angiography, *MNP* magnetic nanoparticles.\n\nEx vivo magnetic particle imaging of AAA {#Sec4}\n----------------------------------------\n\nTo evaluate the potential of MPI for measuring inflammatory response in AAA, ex vivo MPI images of the aorta were acquired 24\u00a0h after i.v. ferucarbotran administration. The AAAs with an overall iron content above 0.3\u00a0\u00b5g were visible in MPI (Fig. [4](#Fig4){ref-type=\"fig\"}A3). The reconstruction of images of ex vivo AAAs with low iron content was possible only with the immobilized MNP system function (SF). The decay of higher harmonics in freeze-dried ferucarbotran is rapid in comparison to the fluid sample. Hence, fewer frequency components were used in the reconstruction, which is not beneficial in sense of MPI resolution. However, the immobilized state of MNPs in the AAA after phagocytosis by macrophages ^[@CR16]^ resembles more the immobilized state of MNPs in sugar matrix of mannitol SF and results in a more reliable reconstruction. The magnetic particle spectroscopy (MPS) quantification results are in good agreement and validate the MPI iron mass quantification (R\u2009=\u20090.99) (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). The average deviation of total iron amount determined by MPI\u00a0from MPS is 8.3% for samples above 1\u00a0\u00b5g and 20.6% for samples below 1\u00a0\u00b5g and 15.2% overall. The slight over-- or underestimation of iron amount in comparison to MPS results might arise from partial volume effects due to the limited resolution of MPI and low signal-to-noise ratio (SNR).Figure 5Correlation of ex vivo MPS and MPI iron oxide particle measurements. MPS is a standard method for validation of MPI results, combining static and dynamic magnetic properties of iron oxide nanoparticles. A strong correlation (R\u2009=\u20090.99) between the amount of iron-oxide particles measured in ex vivo MPS and ex vivo MPI was measured. *MPS* magnetic particle spectroscopy, *MPI* magnetic particle imaging.\n\nHistological analysis {#Sec5}\n---------------------\n\nElastica van Gieson (EvG) stained histological sections revealed strong extracellular matrix remodeling, visible in the 3\u00a0weeks- as well as in the 4\u00a0weeks Ang II infusion groups (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}A3). Degradation of elastic fibers and following dilatation of the aortic lumen was accompanied by the formation of a thrombus. On the other hand, elastogenesis characterized by a higher amount of elastic fibers in the areas adjacent to the vascular lumen was observed, indicating a repair process in late stage aneurysm. The Perls' Prussian blue stained histopathologic sections revealed abundant iron within AAA, while no or little iron was detected in the control group (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}A6). Anti-CD68 monoclonal antibody immunohistology analysis revealed abundant macrophage accumulation in the adventitial area of the aneurysm (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}A7).\n\nCorrelation of magnetic particle spectroscopy, magnetic particle imaging, histology and immunohistochemistry {#Sec6}\n------------------------------------------------------------------------------------------------------------\n\nMPS functions as a zero-dimensional MPI with higher sensitivity. The accumulation of iron in the AAAs was measured with MPS and results were used for validation of the MPI quantification.\n\nEx vivo MPI and MPS measurements revealed abundant iron within AAAs (Figs.\u00a0[4](#Fig4){ref-type=\"fig\"}, [5](#Fig5){ref-type=\"fig\"}), while no or little iron was detected in the control group (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}B5).\n\nA strong correlation was found between the areas positive for CD68 immunohistology stain and Perls' Prussian blue stain (Figs.\u00a0[2](#Fig2){ref-type=\"fig\"}A6, A7), confirming the co-localization of macrophages and MNPs. There is a positive correlation between the amount of iron measured in Perls' Prussian Blue and ex vivo MPS as well as ex vivo MPI (Figs.\u00a0[5](#Fig5){ref-type=\"fig\"}, [6](#Fig6){ref-type=\"fig\"}) iron quantification. The MPS quantification of the intact harvested AAA verified that the stained iron in the sections does not originate from endogenous iron nor from cutting blades during tissue processing.Figure 6Correlation of MPS iron oxide particle measurements and Perls' Prussian Blue stained histological measurements. In order to investigate the absolute amount of iron oxide particles in AAA samples, ex vivo MPS measurements were performed prior to histological processing. A strong correlation between the amount of iron-oxide particles measured in ex vivo MPS and Perl's Prussian Blue staining was shown. The combined assessment of both analytical methods verify the feasibility of MPS for detection of inflammation sites in AAA. *MPS* magnetic particle spectroscopy.\n\nElemental bioimaging by means of laser ablation-inductively coupled plasma-mass spectrometry for spatial localization of iron {#Sec7}\n-----------------------------------------------------------------------------------------------------------------------------\n\nTo determine the spatial distribution of MNPs within the aneurysmal wall, laser\u00a0ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) measurements were performed in two mice after 4\u00a0weeks of Ang II infusion. A pronounced colocalization of iron, macrophages (CD68) and Perls' Prussian blue positive areas was found (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}A4, A6, A7).\n\nDiscussion {#Sec8}\n==========\n\nThis study demonstrates the feasibility of ferucarbotran-enhanced ex vivo MPI for the detection of vascular inflammation in AAA. Ex vivo iron measurements via LA-ICP-MS, immunohistology, histopathology, MPI and MPS demonstrated a strong correlation, thus confirming the inflammatory activity and resulting MNP accumulation in the aneurysmal wall.\n\nInflammation is a key process in the emergence and development of many pathological conditions such as vascular disease, cancer, immune and neurologic disorders. Tracking inflammatory processes usually involves low specificity biomedical imaging or invasive methods such as biopsies^[@CR17]^. Previous research on inflammation detection has tended to focus on MR imaging, demonstrating the MNP uptake at inflammation sites^[@CR18],[@CR19]^. A major disadvantage of this method is that iron oxide nanoparticles cause a signal intensity decrease in MRI, which could easily be missed or mistaken for an artifact^[@CR17]^. MPI allows the detection of inflammation through MNPs with a higher specificity and SNR.\n\nFerucarbotran is currently considered the standard MPI agent due to its commercial availability and excellent signal properties^[@CR20],[@CR21]^. Uptake of MNPs in AAAs has previously been investigated^[@CR22]^ and imaged with MRI in murine model as well as in human patients^[@CR10],[@CR16],[@CR23]^. It has been demonstrated that MNP-enhanced MRI could identify aortic wall inflammation in patients with AAAs and predicts the rate of aneurysm growth and clinical outcome^[@CR10],[@CR22]^. In a clinical study, the growth rate of AAA of patients correlated significantly with amount of MNP in the aneurysmal wall, despite comparable aneurysm diameters^[@CR10]^.\n\nFurther experimental investigations should focus on enhancing the quality of angiographic and vascular MPI. MNPs are administered mostly i.v., allowing the direct visualization of the blood flow^[@CR24]^. The first dynamic MPI-static MRI imaging sequence of a beating mouse heart was acquired in 2009^[@CR25]^. Since most MPI scanners have no tissue depth limitation, imaging the whole cardiovascular system without signal attenuation in a similar manner is theoretically possible^[@CR24]^. Regarding AAA, so far only hemodynamic aneurysm phantom experiments have been conducted, paving the way for in vivo MPI AAA imaging in the future. Once MNPs are administered i.v., there is a very limited imaging time frame before the particles escape the cardiovascular system and start accumulating in the spleen and liver^[@CR26],[@CR27]^. However, for many diseases models, including AAA a prolonged blood circulation time would be beneficial. Ferumoxytol for example, an ultrasmall superparamagnetic iron oxide (USPIO) circulates longer^[@CR28]^, avoiding uptake by Kupffer cells in the liver and also reducing the chances for shadowing effect in MPI^[@CR29]^. The prolonged circulation is favorable for access to the AAA site and uptake by macrophages. Future studies should target identifying further, suitable, high resolution MPI tracer. Furthermore, we should also consider the application of MPI for AAA rupture prediction. In a MRI study by Brangsch et al.in a murine AAA model, the prediction of aneurysm rupture with an iron oxide based contrast agent was associated with a sensitivity of 80% and specificity of 89% . It is a question of future research to investigate whether these results are also applicable to MPI.\n\nAlthough the pathophysiology of AAA is not completely deciphered, it is a well-known fact that both inflammatory activity, characterized by a pronounced proinflammatory cell infiltration, as well as ECM degradation, defined by breakdown of cross-linked elastin and collagen are key^[@CR16]^. While those two mechanisms appear to be autonomous, the formation of AAA is most likely when they co-occur.\n\nInflammatory activity is an excellent in vivo indicator for the characterization of AAAs. Previous studies from the past years have showcased the potential of MNPs for imaging macrophage activity in AAAs^[@CR10],[@CR23],[@CR30],[@CR31]^. There are two different types of macrophages, namely M1 and M2^[@CR32]^. While M1 macrophages are associated with proinflammatory cytokines, classically activated by interferon \u03b3 (IFN-\u03b3) or lipopolysaccharides (LPS) from viral and bacterial pathogens, M2 macrophages are responsible for tissue repair and wound healing, inducing collagen production^[@CR33]^. AAAs are marked by a high M1:M2 cell ratio^[@CR34]^, suggesting that various inflammatory cytokines secreted by M1 macrophages such as migration inhibitory factor (MIF-1) and tumor necrosis factor (TNF) in addition to matrix metalloproteinases (MMP-9) are responsible for the degradation of ECM proteins in the vessel wall, causing continuous dilatation of the aortic lumen^[@CR35],[@CR36]^. In this study, we could detect and quantify the macrophage activity in AAA through several different ex vivo methods. Future work should concentrate on targeting the long-term quantification of inflammatory activity at multiple stages of the AAA development through MPI.\n\nThis study has two primary limitations. First, the strong shadowing effect^[@CR29]^ observed in ferucarbotran-enhanced MPI is responsible for the absence of in vivo MPI in our experiment. When two objects with a large difference in iron concentration simultaneously present in the field of view, the object with lower iron content is suppressed and thus not visible ^[@CR37]^. This is the case with the abdominal portion of the aorta in vicinity of the liver, where ferucarbotran is mainly sequestered by the reticuloendothelial system of the body.\n\nSecond, since no topological anatomical information is obtained from the MPI scans, a reference image is required from another imaging modality such as CT or MRI for accurate localization of the imaging area. Co-registration via fiducial markers^[@CR38]^ is an option, however the transferring of animals from the MPI scanner to the other modality might cause spatial confidence issues and requires image post-processing. An integrated hybrid system combining MPI and MRI will assure spatial co-registration accuracy. A MPI/MRI Hybrid imaging system has been realized successfully, yet only used for static images and 2D phantom measurements^[@CR39],[@CR40]^.\n\nConclusion {#Sec9}\n==========\n\nTo our knowledge, this is the first study that demonstrates the potential of a combined in vivo MR---ex vivo MP imaging for the assessment of inflammatory response in the aneurysmal wall of an Ang II-infused ApoE \u2212/\u2212 mouse model using the MNP ferucarbotran. Future research could examine the feasibility of a combined MR-MP imaging in order to improve the in vivo characterization of AAAs.\n\nMethods {#Sec10}\n=======\n\nAnimal experiments {#Sec11}\n------------------\n\nThe animal experiments were approved and performed according to the local Guidelines and Provisions for Implementation of the Animal Welfare Act by Charite Universitaetsmedizin Berlin, the regulations of the Federation of Laboratory Animal Science Associations (FELASA) and the local animal protection committee of the LaGeSo, Berlin, Germany.\n\nAll procedures in this study were conducted by a veterinarian, and all possible steps were taken to avoid suffering at each stage of the experiment. The animals were fed with a standard lab diet and housed in a clean barrier. For surgery and for the imaging sessions, mice were anesthesized with an intraperitoneal (i.p.) combined injection of 500\u00a0\u00b5g/kg medetomidin, 50\u00a0\u00b5g/kg fentanyl, and 5\u00a0mg/kg midazolam. In order to accelerate recovery time, anesthesia was antagonized using an i.p. combination of Atipamezole (2.5\u00a0mg/kg), Naloxone (1,200\u00a0\u00b5g/kg), Flumazenil (500\u00a0\u00b5g/kg) following MRI imaging. AAAs were induced in 23 male, Apolipoprotein E deficient mice (B6.129P2-ApoE^tm1Unc^/J) mice (8\u00a0weeks old)**.** Osmotic minipumps (Alzet model 2004, Durect Corp) were implanted subcutaneously in the dorsal neck area. Angiotensin II was continuously infused with a rate of 1,000\u00a0ng/kg/min for 3\u00a0weeks (group 1, n\u2009=\u20098) or 4\u00a0weeks (group 2, n\u2009=\u200915), respectively. Sham-operated ApoE\u2212/\u2212 mice (n\u2009=\u20099) delivered saline over 4\u00a0weeks serving as the control group.\n\nIn order to verify the development of AAA, native MR imaging was performed after 3\u00a0weeks (group 1) or 4\u00a0weeks (group 2 and control group), followed by i.v. injection of ferucarbotran to the tail vein (50\u00a0\u00b5l ferucarbotran, 46.66\u00a0\u00b5g iron per kg body weight). 24\u00a0h later, animals were sacrificed and the abdominal part of the aorta was harvested in order to correlate the in vivo MRI findings with ex vivo data (MPI, MPS, histology, immunohistochemistry and LA-ICP-MS).\n\nMagnetic nanoparticles {#Sec12}\n----------------------\n\nFerucarbotran (RESOVIST, I'rom Pharmaceutical Co Ltd, Tokyo, Japan) is the second clinically approved MNP developed for contrast-enhanced MRI of the liver. It is a hydrophilic colloidal solution of \u03b3-Fe~2~O~3~ coated with carboxydextran, composed of clusters of single-domain nanoparticles. Ferucarbotran has a bimodal size distribution with mean core diameters of about 4\u00a0nm and 16\u00a0nm (electron microscopy)^[@CR13],[@CR20]^ and shows a mean hydrodynamic diameter (D~H~) of 60\u00a0nm (photon correlation spectroscopy)^[@CR41]^. The carboxydextran coating (27--35\u00a0mg/ml with an iron to carboxydextran ratio of 1:1 (w/w)) ensures aqueous solubility of the microparticles and prevents aggregation. Ferucarbotran contains 0.5\u00a0mol Fe/l, including 40\u00a0mg/ml mannitol and 2\u00a0mg/ml of lactatic acid, adjusted to a pH of 6.5. At 37\u00a0\u00b0C, the solution has an osmolality of 0.319\u00a0osmol/kg H~2~O and a viscosity of 1,031 MPas. Upon i.v. application, ferucarbotran is taken up by the reticuloendothelial system (RES), mostly in the liver (80%) and spleen (8--9%). Following uptake in RES cells, the carboxydextran coating decomposes and the iron is conveyed to the iron pool via transferrin^[@CR42]^.\n\nIn vivo magnetic resonance imaging {#Sec13}\n==================================\n\nMice were imaged in supine position after induction of anesthesia. Using a clinically approved single loop coil (47\u00a0mm Siemens Healthcare Solutions, Erlangen, Germany)), the imaging sessions were performed on a clinical 3\u00a0T Siemens system (Biograph-mMR, Siemens Healthcare Solutions, Erlangen, Germany). Body temperature (37 \u00b0C) was monitored using a MR-compatible heating system (Model 1025, SA Instruments Inc, Stony Brook, NY). In conclusion to the acquisition of native scans, ferucarbotran was administered via a 30G cannula attached to a small diameter tube inserted into the tail vein of the animals.\n\nA non-contrast-enhanced two-dimensional time-of-flight angiography (2D TOF) following a three-dimensional (3D) gradient echo scout scan was performed in transverse orientation for visualization of the abdominal aorta. Following parameters were used: field of view (FOV) of 200\u2009\u00d7\u2009200\u00a0mm, matrix of 960\u2009\u00d7\u2009960, resolution of 0.2\u2009\u00d7\u20090.2x0.5\u00a0mm, 40 slices, repetition time(TR)/echo time (TE) of 35\u00a0ms/4.44\u00a0ms, flip angle of 90\u00b0, and bandwidth of 124\u00a0Hz/Px. To obtain an arterial angiogram of the abdominal aorta for planning the subsequent MR angiography, a maximum intensity projection (MIP) was automatically produced.\n\nEx vivo magnetic particle imaging {#Sec14}\n---------------------------------\n\nEx vivo MP imaging of the aorta was performed 24\u00a0h post-i.v. administration of 50\u00a0\u00b5l ferucarbotran after 3 (group 1, n\u2009=\u20098) respectively 4 (group 2, n\u2009=\u200915) weeks of Ang II perfusion. MPI images were acquired on a commercial preclinical MPI system (Bruker MPI 25/20 FF), equipped with a separate gradiometric receive coil ^[@CR43],[@CR44]^ for improved sensitivity and image quality.\n\nThis field-free-point (FFP) based MPI system requires a pre-recorded (SF) for image reconstruction^[@CR45],[@CR46]^. The SPIONs inside the FOV are excited with three drive fields at amplitudes of 12 mT, orthogonal to one another, operating at slightly different excitation frequencies (2.5\u00a0MHz divided by 102/96/99 in x-/y-/z-direction) to generate the Lissajous trajectory of the FFP movement in a selection field gradient at strength of 2.5\u00a0T/m in z-direction, and 1.25\u00a0T/m in x and y direction. The FFP scans the FOV along this Lissajous trajectory. In this matter the FOV is scanned into a 3D image of the MNP distribution at a temporal resolution of 21.5\u00a0ms.\n\nMagnetic particle imaging system function {#Sec15}\n-----------------------------------------\n\nThe SF was measured for ferucarbotran in two forms: one in an aqueous suspension and one in a freeze-dried mannitol sugar matrix. Both SF samples, the aqueous suspension SF sample and the immobilized SF sample (in mannitol sugar matrix) ^[@CR47]^ had an iron concentration (c(Fe)) of 100\u00a0mM and volume of 13.5\u00a0\u03bcl. The immobilized SF sample was prepared with a mannitol solution (10% w/v). The SF sample was measured in a container with dimension of 3\u2009\u00d7\u20093\u2009\u00d7\u20091.5 mm3 in a cuboid shape. The measurement grid had a size of 25\u2009\u00d7\u200925\u2009\u00d7\u200913 in FOV of size 25\u2009\u00d7\u200925\u2009\u00d7\u200913 mm3 leading to an overscan of the physical FOV of 19.2\u2009\u00d7\u200919.2\u2009\u00d7\u20099.6 mm^3^ ^[@CR48]^.\u00a0The SF measurements were acquired with the same aforementioned drive and gradient fields for image acquisition.\n\nImage reconstruction\u00a0and analysis by magnetic particle imaging {#Sec16}\n--------------------------------------------------------------\n\nThe images were reconstructed to 25\u2009\u00d7\u200925\u2009\u00d7\u200913 voxels via Kaczmarz's algorithm with Tikhonov regularization in ParaVision 6 MPI software (Bruker Biospin, Ettlingen, Germany). The hardware background noise limits the used frequency components to a bandwidth to 0.09--125\u00a0MHz so the 3rd harmonic is as well filtered out. A number of 487 frequency components were chosen automatically according to the applied SF and the SNR (SNR\u2009=\u20097) threshold determined from the SF. Frequency components with a mixing order above 25 were removed from the further analysis. For reconstruction, a block average of 20 repetitions was applied to the measurement to reduce visible noise in the images. A regularization of \u03bb\u2009=\u200910^\u22121[@CR49]^ and five iterations were performed for the final reconstructions. These parameters were kept constant in all reconstructions to exclude the influence of variations in these parameters on the intensity values.\n\nFor quantification of iron mass in ex vivo MPI images, they were analysed in\u00a0MATLAB (Mathworks, Natick, MA, USA). First a 50% cut-off threshold of maximum voxel intensity value was applied to the 3D dataset; to eliminate background noise and artefacts, and to minimize the effect of image blurring due to the regularization in image reconstruction^[@CR44]^. Thereafter the calculation of iron mass in AAA was performed by integration of overall iron in volume of interest (VOI) over the AAA.\n\nQuantification by magnetic particle spectroscopy {#Sec17}\n------------------------------------------------\n\nEx vivo MPS measurements of harvested abdominal aortas (n\u2009=\u200932) were performed using a commercial magnetic particle spectrometer (Bruker, Germany) with a sinusoidal magnetic signal excitation using an amplitude of 25 mT, a frequency of 25\u00a0kHz and a sample temperature of 37\u00a0\u00b0C. The nonlinear magnetization response of MNPs in AAA was measured for 10\u00a0s by a pickup coil (sensitivity: 10^--12^ A m^2^). For accurate MPS iron mass quantification, the reference was chosen according to the harmonic ratio A5/A3 of measured sample, immobilized (freeze-dried in 10% mannitol) or in water dispersion and the background signal of the empty sample holder was subtracted from the MPS spectra. For quantitation, the amplitude of the 3rd harmonic of the MPS spectra of measured samples was normalized to the amplitude of the known reference sample.\n\nHistological analysis of aortic aneurysms and aortic aneurysm morphometry {#Sec18}\n-------------------------------------------------------------------------\n\nHistological analysis was performed in the same region of the aorta that was imaged in MRI and MPI. Aortic aneurysm samples were divided in half for paraffin- and cryosectioning, tissues were either processed overnight in MorFFFix (MORPHISTO, Frankfurt am Main, Germany) or frozen at -20\u00b0 C. 9\u00a0\u00b5m thick sections of the vessels were stained with Perls\\` Prussian Blue staining, Miller's Elastica van Gieson staining, Hematoxylin and Eosin staining. Resulting histological slices were scanned and photographed using a light microscope (Keyence BZ-X800, Keyence Corporation of America, USA). The morphometrical analysis of the aortic region was performed using Keyence BZ-X800 Analyzer software (Keyence BZ-X800, Keyence Corporation of America, USA). To measure the iron oxide percentage in the tissue in a single digitized image, the color profile of iron oxide as seen using Perls' Prussian Blue was set as reference. All structures within this specific color profile were automatically recorded and divided by the overall tissue area in order to acquire the iron oxide ratio.\n\nImmunofluorescence analysis {#Sec19}\n---------------------------\n\nImmunofluorescence staining was performed to assess the localization of macrophages. Frozen AAA samples fixed in optimal cutting temperature compound (OCT) at \u2212 25\u00b0 were cut into 9\u00a0\u00b5m thick cryosections and subsequently mounted on SuperFrost microscope slides (Thermo Scientific). The slides were first incubated overnight at 4\u00b0 using a monoclonal CD68 antibody (Rat anti-Mouse CD68, clone FA-11, Bio-Rad, 1:100) diluted in Dako REAL Antibody Diluent (Dako, Denmark) and subsequently washed with phosphate-buffered saline (PBS, pH\u20097.4) three times. Slides were incubated with polyclonal secondary antibody AlexaFluor 568 (Goat anti Rat IgG, Thermo Fisher Scientific, Germany, 1:200) for one hour at room temperature, followed by counterstaining and mounting with (DAPI Solution, Roti---Mount FlourCare (CARL ROTH, Germany). Co-localization of macrophages and Perl's Prussian Blue positive areas were assessed in serials section of AAAs.\n\nElemental bioimaging by means of laser ablation-inductively coupled plasma-mass spectrometry for spatial localization of iron {#Sec20}\n-----------------------------------------------------------------------------------------------------------------------------\n\nAAA samples were cut at \u2212 25\u00a0\u00b0C into 9\u00a0\u00b5m cryosections and immediately mounted on SuperFrost adhesion slides (Thermo Scientific). The LA-ICP-MS analysis was performed with a LSX 213 G2\u2009+\u2009laser system (CETAC Technologies, Omaha, USA) equipped with a two volume HelEx II cell connected via Tygon tubing to an ICPMS-2030 (Shimadzu, Kyoto, Japan). Samples were ablated via line-by-line scan with a spot size of 7\u00a0\u00b5m, a scan speed of 21\u00a0\u00b5m/s and 800\u00a0mL/min He as transport gas. The analysis was performed in collision gas mode with He as collision gas and 50\u00a0ms integration time for the ^57^Fe isotope. For the quantification of Fe, matrix-matched standards based on gelatin were used. Nine gelatin standards (10% w/w) including a blank, were spiked with different Fe concentrations ranging from 1 to 5.000\u00a0\u00b5g/g. Averaged intensities of the scanned lines of the standards showed a good linear correlation with a regression coefficient R^2^\u2009=\u20090.9999 within this concentration range. Limit of detection (LOD) and limit of quantification (LOQ), calculated with the 3\u03c3- and 10\u03c3-criteria, were 26\u00a0\u00b5g/g and 86\u00a0\u00b5g/g Fe. The quantification and visualization were performed with an in-house developed software (WWU M\u00fcnster, M\u00fcnster, Germany).\n\nStatistical analysis {#Sec21}\n--------------------\n\nFor the comparison of continuous variables, a Student's t test (unpaired, two-tailed) was applied. p\u2009\\<\u20090.05 was regarded to be statistically significant.\n\n**Publisher\\'s note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThis study was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)---Projektnummer 372486779---SFB 1340; MA 5943/3-1/4-1/9-1, Magnetic Particle Imaging and Research Training Group 2260 (BIOQIC),\n\nD.B.M. performed and analyzed the animal and ex vivo experiments and wrote the manuscript. M.R.M. designed and supervised the study. R.B. and U.K. designed and carried out the LA-ICP-MS experiments. A.M, J.B., H.P. and O.K. participated in ex vivo experiments and reviewed the manuscript. J.B., A.M., O.K., H.P., F.W., R.K., R.B., U.K., M.T., J.S. and B.H reviewed and commented on the manuscript.\n\nThe\u00a0datasets\u00a0generated\u00a0during\u00a0and/or\u00a0analysed\u00a0during\u00a0the\u00a0current\u00a0study\u00a0are\u00a0available\u00a0from\u00a0the\u00a0corresponding\u00a0author\u00a0on\u00a0reasonable\u00a0request.\n\nThe authors declare no competing interests.\n"}
+{"text": "The online version of this article contains supplemental material.\n\nIntroduction\n============\n\nCytoskeletal--cell membrane attachments are critical for establishing the proper relationships between cells during tissue morphogenesis and for maintaining the integrity of tissues during differentiation. In vertebrates, intermediate filaments (IFs)[\\*](#fn1){ref-type=\"fn\"} provide the tensile strength required for tissue integrity ([@bib13]). In complex epithelia and other tissues such as heart that experience mechanical stress, the association of IFs with intercellular adhesive junctions known as desmosomes generates a stress-resistant scaffolding that integrates all of the cells within the tissue ([@bib26]; [@bib46]). Perturbation of either the IF cytoskeleton or its desmosome anchor compromises tissue integrity ([@bib16]; [@bib46]); however, the respective contributions and molecular mechanisms by which these integrated components serve their mechanical roles have not been fully elucidated. Specifically, the question of whether IFs strengthen desmosomal cadherin-mediated adhesion, much in the way actin has been proposed to strengthen classical cadherin-mediated adhesion, has not been directly addressed.\n\nDiseases targeting IFs as well as proteins that tether IFs to desmosomes demonstrate the importance of the desmosome--IF scaffolding. On one hand, point mutations that interfere with the assembly of IFs into networks lead to epidermolysis bullosa simplex, which is characterized by cell fragility and cytolysis in response to mild trauma ([@bib14]; [@bib17]; [@bib45]). On the other hand, patients with mutations in desmosomal plaque components involved in IF tethering also exhibit skin fragility, accompanied by retraction of IFs from the cell surfaces and perinuclear accumulation ([@bib46]). Histological observations also show widened intercellular spaces, raising the possibility that such mutations might cause an adhesive defect, but making such a conclusion is complicated by possible reductions in desmosome number ([@bib46]).\n\nDesmosomes are highly organized structures composed of members from three families, the cadherins, armadillo proteins, and plakins (for reviews see [@bib55]; [@bib21]; [@bib9]; [@bib40]; [@bib26]). The desmosomal cadherins, subdivided into the desmocollins and desmogleins, form the adhesive core of the desmosome ([@bib36]; [@bib3]). The cadherins interact, in turn, with members of the armadillo family, including plakoglobin and, in certain cases, the plakophilins ([@bib23]; [@bib31]). The latter comprises a subclass of armadillo proteins related to the classic cadherin-associated protein p120^ctn^ ([@bib31]). Armadillo family members bind to desmoplakin (DP) and possibly other members of the plakin family of cytolinkers, which also includes plectin, envoplakin, and periplakin ([@bib52]; [@bib18]; [@bib42]). DP is an obligate desmosomal component that closes the link by mediating interactions with the IF cytoskeleton ([@bib25]). The submembranous desmosomal plaque is further bolstered by interactions of a lateral nature among armadillo family members from different subclasses, between DP and the cadherins and possibly between IFs and armadillo family members ([@bib26]).\n\nDP plays an essential role in anchoring IFs to the membrane through its COOH-terminal plakin repeat domain ([@bib59]; [@bib37]). DP\\'s role is underscored by the existence of a dominantly inherited DP haploinsufficiency ([@bib5]) as well as recessively inherited truncations within the COOH-terminal IF-binding domain of DP, which lead to striate palmoplantar keratoderma and skin, hair, and cardiac defects, respectively ([@bib48]). Genetically engineered mice have provided additional insight into the fundamental importance of DP. DP-null mice are unable to proceed past day 6.5 of embryonic development due to defects of the embryonic endoderm and failure of the egg cylinder to elongate ([@bib19]). Chimeric morulae expressing DP in extraembryonic tissue survived somewhat longer, until E10.5--E12.5, exhibiting defects in heart, neuroepithelium, and epidermis and finally dying as a result of defects in the microvascular system ([@bib20]). Most recently, conditional knockouts in which DP is specifically ablated in the epidermis provided compelling data that this molecule plays a key role in epidermal sheet formation ([@bib67]). However, keratinocytes derived from conditional null animals have a greatly reduced number of desmosomal junctions, precluding a determination of whether severing the IF--desmosome connection, while still leaving the remaining plaque intact, affects cell integrity versus adhesive strength.\n\nDP-null keratinocytes also displayed abnormalities in microfilament organization, suggesting that the proper maturation of the cortical actin cytoskeleton is dependent on the presence of desmosomes ([@bib67]). Other studies have suggested that the reverse is also true; that is, desmosome assembly is dependent on the prior initiation of adhesion through the classic cadherin system ([@bib28]; [@bib43]), which triggers assembly of actin and reorganization of the cortical actin cytoskeleton ([@bib66]; [@bib65]; [@bib38]). In addition, studies have reported that loss or mutation of desmosomal plaque constituents can result in the intermingling of adherens junction and desmosomal components ([@bib7]; [@bib53]). The observed interdependence of these junction--cytoskeleton systems raises the question of their relative contributions to adhesive strength.\n\nTo address these questions and to specifically test the role of the IF connection in cellular integrity and adhesion, we developed Tet-On A431 cell lines expressing a dominant-negative DP polypeptide (DPNTP) that uncouples IFs from the desmosome while leaving junctional plaques intact ([@bib7]). Cell aggregates in hanging drop culture as well as epithelial cell sheets generated from cells expressing this polypeptide both exhibited increased dissociation without evidence of cytolysis when subjected to mild mechanical stress. In addition, time-lapse imaging analysis revealed that dissociation of DPNTP-expressing cells occurred more rapidly and with less cell distension compared with control cells. This adhesive defect occurred without alteration of the cell surface expression of desmosomal or classic cadherins or effect on E-cadherin--dependent adhesion in capillary shear flow assays. However, the adhesive defect was accompanied by an approximately twofold decrease in desmoglein 2 (Dsg 2) present in the detergent-insoluble pool, suggesting that desmosomal cadherins are specifically affected by loss of cytoskeletal attachment. Latrunculin A (LtnA), which interferes with the cortical actin cytoskeleton and associated adherens junctions, decreased adhesive strength as well, but importantly, DPNTP expression together with LtnA treatment resulted in an almost complete dissociation of the cell sheet. Keratinocytes derived from patients with a COOH-terminal truncation of DP also exhibit an adhesive defect, supporting the physiological relevance of these findings. Together these data provide the first direct in vitro evidence for a specific role of IF attachment in cell--cell adhesive strength.\n\nResults\n=======\n\nDPNTP expression leads to uncoupling of IFs from cell junctions without altering cell surface expression of desmosomal or classic cadherins\n-------------------------------------------------------------------------------------------------------------------------------------------\n\nPrevious studies demonstrated that ectopic expression of a DP deletion mutant, DPNTP, which retains plakoglobin and plakophilin binding sites but lacks the central rod and COOH-terminal IF-binding domains ([@bib7]) ([Fig. 1](#fig1){ref-type=\"fig\"} A), results in uncoupling of IFs from the plasma membrane while leaving the IF network intact. To further study the consequences of severing the IF--desmosome connection without potential complications arising from clonal variability, tetracycline-inducible cell lines were generated (Tet-On; CLONTECH Laboratories, Inc.). Single stable (SS) A431 clones expressing the tetracycline-dependent transactivator were first generated and then selected for optimum morphology coupled with the ability to robustly induce expression of a luciferase reporter construct in a tightly regulated fashion. A431/SS clones were stably transfected with an expression cassette driven by a tetracycline response element containing a coding sequence for a COOH-terminally FLAG-tagged DPNTP. Double stable clones were obtained by puromycin selection and induction of DPNTP expression was scored by Western blot analysis. Of the stable clones identified, two clones, A431/C2 and G4, were characterized in detail. These clones behaved similarly in all assays used in this study but occasionally differed in the magnitude of response (e.g., see [Fig. 5](#fig5){ref-type=\"fig\"} B). The observed differences might be explained by the fact that the G4 clone tended to exhibit a low level of leaky expression of DPNTP and, in addition, exhibited a lower level of E-cadherin expression compared with the C2 line. Data for the C2 line are shown throughout, and where instructive, we have included and discussed representative data for the G4 line.\n\n![**Characterization of inducible A431 cell lines expressing DPNTP.** (A) Comparison of DPNTP and human JD-1 deletion mutant DP to wild-type DP. DPNTP encodes the NH~2~-terminal 584 amino acids of DP containing binding sites for both plakoglobin and the plakophilins. The JD-1 mutation removes the C subdomain and includes 18 amino acids not found in wild-type DP. (B) Dose response of DPNTP expression. A431/C2 cells were induced for 48 h with increasing concentrations of Dox and then analyzed by Western blotting using antibodies directed against the NH~2~ terminus of DP. (C) Western blot analysis of junctional proteins in Dox-treated and untreated A431/C2 cells. Whole cell lysates were analyzed using antibodies directed against Dsg 2, E-cadherin, \u03b1-catenin, \u03b2-catenin, plakoglobin, and DP/DPNTP. Immunoblotting for keratin 18 was used as a loading control in both B and C. Note that the levels of desmosomal and classic cadherins and catenins are comparable in DPNTP-expressing and control cells.](200206098f1){#fig1}\n\n![**DPNTP weakens adhesion in epithelial sheets.** (A) Dox-treated and untreated monolayers, plated in triplicate, were separated from culture dishes via incubation with dispase. Monolayers were transferred to 15-ml conical tubes containing 5 ml of DPBS. After 50 inversions, the degree of fragmentation of the monolayer was observed. (B) The dissociation assay was quantified by counting the number of total particles. Dox-treated A431/C2 and G4 monolayers showed increased dissociation relative to uninduced and A431/SS monolayers (*t* test, P \\< 0.001 and P \\< 0.005, respectively). Each bar represents the SEM from three independent experiments in which each condition was tested in triplicate. (C) LDH release was not increased despite increased fragmentation of monolayers expressing DPNTP. Values are compared with controls obtained from a mechanically lysed cell population.](200206098f5){#fig5}\n\nNorthern and Western blot analysis showed that in the absence of doxycycline (Dox), a tetracycline analogue, DPNTP mRNA and protein were undetectable in the A431/C2 cell line (unpublished data). Incubation of cells for 24--48 h with 2 \u03bcg/ml of Dox led to robust expression of DPNTP mRNA, and protein could be detected as early as 12 h after induction ([@bib22]). Maximum levels of DPNTP expression were observed at Dox concentrations between 2 and 6 \u03bcg/ml ([Fig. 1](#fig1){ref-type=\"fig\"} B) and reached levels comparable to endogenous DP expression. Levels of endogenous DP decreased up to twofold upon expression of DPNTP, consistent with our previous observations ([Fig. 1](#fig1){ref-type=\"fig\"} B) ([@bib7]).\n\nTo examine whether DPNTP affected the expression of other junctional proteins, immunoblot analysis was performed using antibodies against the desmosomal molecules DP, Dsg 2, desmocollin 2, and plakoglobin and the adherens junction proteins E-cadherin, \u03b2-catenin, and \u03b1-catenin ([Fig. 1](#fig1){ref-type=\"fig\"} C; unpublished data). No detectable changes in expression of any of the desmosomal components, besides DP, or of the adherens junction components were observed. Similar data were obtained for the G4 line (unpublished data). In addition, no discernible difference in total classic or desmosomal cadherin turnover was observed over a period of 2 d in 10 \u03bcg/ml cycloheximide to block protein synthesis (unpublished data).\n\nIn the absence of Dox, endogenous DP was localized at cell borders in a punctate pattern typical of parental A431 cells ([Fig. 2](#fig2){ref-type=\"fig\"}). Using an antibody directed against the FLAG epitope on DPNTP, cell border staining colocalizing with endogenous desmosomes was rarely detected in the C2 line and it was only occasionally detected in the leaky G4 line. Staining for IFs revealed an intact network extending out to and anchoring at desmosomes at sites of cell--cell contact. After Dox treatment, DPNTP was localized at cell borders in a punctate pattern, whereas endogenous DP largely disappeared from cell borders. IF bundles were dramatically uncoupled from membrane sites where DPNTP was localized, although a few plasma membrane sites still retained some endogenous DP with associated IF bundles.\n\n![**DPNTP localizes to intercellular contacts in a punctate pattern and uncouples IFs from the cell surface.** A431/C2 cells were induced for 24 h in 2 \u03bcg/ml Dox and either immunostained using antibodies directed against endogenous DP (DP2.15) or the FLAG epitope (poly-FLAG) on the COOH terminus of DPNTP (A and B), or antibodies directed against DP/DPNTP (NW161) and keratins (KS-B17.2) (C and D). Dox induction led to the accumulation of DPNTP in a junctional distribution at cell--cell borders accompanied by loss of endogenous DP and detachment of K8/18-containing IF bundles.](200206098f2){#fig2}\n\nOur results described above indicated that total expression levels of desmosomal and classic cadherins did not change in DPNTP-expressing cells. However, because previous observations showed that expression of DPNTP can cause intermixing of E-cadherin and Dsg 2 ([@bib7]), we sought to ensure that cadherins were properly delivered to the plasma membrane and present in a normal distribution. We compared the steady-state expression level and distribution of desmosomal and classic cadherins on the cell surface in uninduced and induced cells by labeling cell surface proteins with a membrane-impermeable biotin reagent. Levels of biotinylated Dsg 2 and E-cadherin from induced and uninduced A431/SS and C2 cells were compared after precipitation with streptavidin-conjugated agarose beads. Cell surface levels of both cadherins remained the same after DPNTP expression ([Fig. 3](#fig3){ref-type=\"fig\"} A). In addition, the amounts of cadherin-associated \u03b2-catenin and plakoglobin were unchanged.\n\n![**Cell surface distribution of desmosomal and classic cadherins is unaltered but Dsg 2 is reduced in the detergent-insoluble fraction of DPNTP-expressing cells.** (A) Dox-induced and uninduced A431/SS and A431/C2 were biotinylated using a cell-impermeable biotin reagent. After precipitation using streptavidin beads, precipitates were separated by SDS-PAGE and immunoblotted using antibodies against E-cadherin (795), Dsg 2 (6D8), plakoglobin (1407), and \u03b2-catenin (C2206). To verify complete biotinylation, twice the concentration of biotin reagent was used in a separate experiment. Induction of DPNTP was confirmed by immunoblot analysis of whole cell lysates using an anti-FLAG antibody. No difference was seen in the amount of either E-cadherin or Dsg 2 or their associated proteins at the cell surface. (B) Immunofluorescence analysis was performed in Dox-induced and uninduced A431/C2 cells using antibodies against E-cadherin (HECD-1) and desmoglein (Dg3.10). (C) Fluorescence density was determined by dividing the fluorescent signal at cell borders (area in square pixels, see boxed examples) by the length of the border (pixels). The distribution and intensity for these cadherins was not detectably altered by DPNTP expression. (D) The detergent-insoluble fractions from Dox-treated and untreated cells were analyzed by SDS-PAGE and immunoblotted using antibodies against E-cadherin (795) and Dsg 2 (6D8). Note the decrease in the Triton-insoluble pool for Dsg 2, but not E-cadherin.](200206098f3){#fig3}\n\nThe fluorescence intensity distribution of E-cadherin and Dsg 2 at regions of cell--cell contact was determined by immunofluorescence and morphometric analysis ([Fig. 3, B and C](#fig3){ref-type=\"fig\"}). Fluorescence intensity was quantified and normalized for the length of the region of cell contact. The distributions ([Fig. 3](#fig3){ref-type=\"fig\"} B) and fluorescence intensities ([Fig. 3](#fig3){ref-type=\"fig\"} C) of E-cadherin and Dsg 2 in Dox-treated and untreated cells were both comparable. Similar data were obtained for desmocollin 2 (unpublished data).\n\nA decrease in the Triton-insoluble fraction of desmoglein in DPNTP-expressing cells is accompanied by reductions in intercellular adhesion and cell sheet integrity\n-------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAs described above, DPNTP did not detectably alter the cell surface expression of classic or desmosomal cadherins. However, as IFs are uncoupled from these adhesive structures, we predicted that the loss of IF attachment might alter the amount of Dsg 2 in the IF-associated Triton-insoluble pool. To test this possibility, untreated or treated A431/SS, C2, and G4 lines were lysed in 1% Triton X-100 buffer and the insoluble pellet was analyzed by SDS-PAGE ([Fig. 3](#fig3){ref-type=\"fig\"} D). Both the A431/C2 and G4 lines exhibited an approximately twofold decrease in the Triton-insoluble fraction, whereas no change was detected in the solubility of E-cadherin.\n\nWe next sought to test whether this desmosomal cadherin-specific alteration in solubility translated into a functional difference in intercellular adhesion and epithelial sheet integrity. We first employed an adaptation of a previously described hanging drop aggregation assay using EDTA-dissociated single cells to generate clusters of cells not adherent to the substrate ([@bib62]). Cells allowed to aggregate for at least 6 h in a hanging drop on the underside of a culture dish were subjected to trituration through a 200-\u03bcl Gilson pipette tip in an attempt to disrupt intercellular adhesion. The degree of dissociation was quantified by counting the particles that dissociated from the original cluster. Cells induced with Dox to express DPNTP were more susceptible to dissociation than untreated cells ([Fig. 4](#fig4){ref-type=\"fig\"} A) and exhibited on average an approximately fourfold increase in particle number when compared with the untreated cells after trituration ([Fig. 4](#fig4){ref-type=\"fig\"} B). Dox-treated and untreated A431/SS cells did not show any differences in the degree of dissociation, which was comparable to untreated C2 cells.\n\n![**DPNTP weakens adhesion in cell clusters formed in hanging drop culture.** (A) 1 \u00d7 10^3^ cells that had previously been Dox induced for at least 24 h or left untreated were seeded into hanging drop cultures without (top left) or with Dox-containing (top right) media and allowed to aggregate for 6 h. After trituration by passing the cell cluster 30 times through a 200-\u03bcl pipette tip, the degree of dissociation of the cell cluster was visualized by microscopy (A) and quantified by manual counting in a dissecting microscope (B). Note that cell clusters were imaged while suspended in a hanging drop and that dark patterns at the edge of field are caused by refraction of light. Due to cropping of the images, some of the dissociated small particles at the periphery in the bottom right panel of A are not visible but are still reflected in the graph in B. Error bars represent the standard deviation from an experiment in which each condition was tested in triplicate.](200206098f4){#fig4}\n\nBecause cell clusters grown in suspension were not well suited for analysis by immunofluorescence microscopy to monitor desmosome status, we sought to use an assay in which we could both verify formation of intercellular junctions and test adhesive strength. Confluent monolayers of Dox-induced and uninduced A431/SS and C2 cells were harvested from tissue culture dishes by incubation with dispase as previously described ([@bib11]). Monolayers were then transferred to 15-ml conical tubes containing 5 ml of Dulbecco\\'s PBS (containing 0.9 mM calcium chloride and 0.5 mM magnesium chloride \\[DPBS\\]). After inverting the tubes 50 times on a rocker, monolayer fragments were counted ([Fig. 5](#fig5){ref-type=\"fig\"} A). DPNTP-expressing monolayers dissociated into numerous smaller fragments, whereas both Dox-treated and untreated control cell lines and uninduced A431 monolayers exhibited only minimal dissociation. Quantification of the number of fragments showed a 23-fold (*t* test, P \\< 0.001) and fourfold (P \\< 0.005) increase in fragmentation by the DPNTP-expressing A431/C2 and G4 monolayers, respectively ([Fig. 5](#fig5){ref-type=\"fig\"} B).\n\nDPNTP-dependent dissociation of cell sheets is not accompanied by lactate dehydrogenase (LDH) release\n-----------------------------------------------------------------------------------------------------\n\nTo address whether the observed increase in fragmentation in DPNTP-expressing monolayers was due to increased cellular fragility and cytolysis, we assessed cellular damage by measuring the release of lactate dehydrogenase (LDH) from monolayers after application of mechanical stress. LDH release measured from DPNTP-expressing monolayers after mechanical disruption was not statistically altered compared with that released from the uninduced and control cell lines, which had minimal fragmentation ([Fig. 5](#fig5){ref-type=\"fig\"} C). Thus, despite a dramatically increased degree of monolayer dissociation, the DPNTP-expressing cell lines did not exhibit a corresponding increase in LDH release.\n\nTime-lapse microscopy of living cells reveals that DPNTP-expressing cells dissociate more rapidly and with less cell distortion under stress than control cells\n---------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nTo examine at higher resolution the behavior of IF-uncoupled cells undergoing stress within epithelial cell sheets, time-lapse imaging of living cells was performed. Tet-On cell lines expressing DPNTP COOH-terminally tagged with EGFP were generated to confirm that DPNTP was being expressed and incorporated into cell junctions specifically in the cells under observation. These cell lines behave similarly in all respects to the DPNTP--FLAG-expressing cells. One cell line, termed A431/I3, which homogeneously expressed DPNTP--EGFP and exhibited the best morphological characteristics for imaging, was selected for this analysis. Data obtained with Dox-induced I3 cells were compared with either uninduced cells or control cell lines treated with Dox to rule out Dox-dependent effects.\n\nThe I3 cell line inducibly expressing DPNTP--EGFP was grown to confluence in the presence or absence of Dox, wounded, and then mounted in a living cell viewing chamber as described in the Materials and methods. To initiate stress-induced tearing of the cell sheet, the wounded monolayers were released from the center of the coverslip by flowing dispase through the chamber. The edges of the cell sheet remained tethered by the circumferential gasket associated with the living cell chamber; thus, cells at the edge of the wound began to stretch and tear due to contraction of the sheet as it lifted from the coverslip. Time-lapse images were captured at 4- ([Fig. 6](#fig6){ref-type=\"fig\"} A) or 8-s ([Fig. 6](#fig6){ref-type=\"fig\"} B) intervals and are shown here as selected individual still images and as four supplemental video clips (Videos 1--4, available online at <http://www.jcb.org/cgi/content/full/jcb.200206098/DC1>). Aspect ratios, which describe the geometry or \"squareness\" of a cell, were determined by dividing the diameter along the axis of stretch by the diameter orthogonal to the axis of stretch.\n\n###### \n\n**DPNTP expression leads to rapid dissociation of cells within an epithelial sheet with less cell distortion than control cells.** (A) The A431 I3 DPNTP--EGFP line treated with Dox (top panels) or without (bottom panels) was imaged during stress-induced tearing. Expression of DPNTP was confirmed by visualizing EGFP fluorescence (top, second panel from right). Individual still images from the time-lapse record have been enlarged here to include the boxed regions and are shown with a time stamp. The entire time-lapse record for each movie can be viewed at <http://www.jcb.org/cgi/content/full/jcb.200206098/DC1> as Video 1 (top panels) and Video 2 (bottom panels). (B) A Dox-induced DPNTP--EGFP line (top panels) was imaged during stress-induced tearing and compared with control cells treated with Dox (bottom panels) to rule out Dox- dependent effects. Individual still images from the time-lapse record are shown here and the complete movies can be viewed as Video 3 (top panels) and Video 4 (bottom panels). Expression of DPNTP was confirmed by visualizing GFP fluorescence (unpublished data). Quantitative information for the cells marked with the asterisk is shown in C, left. Note that in the examples in both A and B, cell sheets expressing DPNTP tear more rapidly, and cells become less distended before cell--cell dissociation occurs, reflecting that less stress is required for dissociation to occur. This can be appreciated best by watching the video clips. (C) Cell aspect ratios for the cells in B were determined as described in the Materials and methods and plotted versus time. Arrows represent the time points shown in B. Note that the DPNTP--EGFP cells dissociated more quickly with less distension along the axis of stretch. Population data summarizing the average time to breakage are provided in C, right. Error bars represent the SEM for a population of 10--15 cells.\n\n![](200206098f6ab)\n\n![](200206098f6c)\n\nAlthough the precise degree of stretch could not be controlled using this method, we observed a dramatic difference between DPNTP-expressing cells and control cells, both in the length of time it took for tearing to occur and the degree of cell distortion observed at the time of dissociation ([Fig. 6](#fig6){ref-type=\"fig\"}, A--C). DPNTP-expressing cells dissociated much more rapidly than control cells ([Fig. 6, C and D](#fig6){ref-type=\"fig\"}) and did so without undergoing as much distension along the axis of stretch ([Fig. 6](#fig6){ref-type=\"fig\"} C).\n\nDPNTP expression does not alter E-cadherin function\n---------------------------------------------------\n\nAs shown above, total and cell surface expression of E-cadherin were not altered by expression of DPNTP. To verify that DPNTP did not interfere with E-cadherin function, a capillary tube flow assay was employed ([@bib70]; [@bib24]). Uninduced and induced A431/C2 and G4 cells were first allowed to adhere to the inner surface of capillary tubes precoated with the extracellular domain of E-cadherin and then subjected to shear flow through the capillaries after either 10 or 45 min ([Fig. 7](#fig7){ref-type=\"fig\"}), the latter time frame having previously been shown to allow for a strengthening phase of adhesion ([@bib70]). The percentage of cells still adhering to the capillary tube after being subjected to various flow rates was recorded. CHO cells, which do not express E-cadherin, were used as a negative control, whereas transfected CHO cells expressing E-cadherin were used as a positive control. Although previous studies suggested that the \u03b2-catenin binding site is not necessary for adhesion or strengthening, when assays were performed in the presence of LtnA, E-cadherin--dependent adhesion was abrogated, supporting the importance of the actin cytoskeleton in this assay ([Fig. 7](#fig7){ref-type=\"fig\"}). Both uninduced and induced A431/C2 and G4 cells exhibited adhesion comparable to the E-cadherin--expressing CHO cells at both time points (unpublished data). These results suggest that DPNTP expression does not impair the early phase of E-cadherin--dependent adhesion or the strengthening phase, both of which are also dependent on an intact actin cytoskeleton.\n\n![**DPNTP expression does not alter the adhesive function of E-cadherin.** Capillary flow assays were performed on induced (2 \u03bcg/ml Dox) and uninduced (0 \u03bcg/ml Dox) A431/C2 and G4 cells using glass capillary tubes coated with recombinant E-cadherin extracellular domain as described in the Materials and methods. Adhesion of both treated and untreated cells was similar under various flow rates. In addition, no difference was seen in the strengthening of adhesion after cells were allowed to adhere to capillary tubes for longer periods of time (45 min). Note that pretreatment with 2 \u03bcM latrunculin A abrogated E-cadherin--dependent adhesion. Note that the G4 line exhibited a somewhat greater loss of binding in response to flow than the C2 line in both the Dox-treated cultures and controls. It is possible this could be explained in part by the lower expression of E-cadherin in this line.](200206098f7){#fig7}\n\nContribution of microfilaments to adhesive strength in uninduced and induced cell lines\n---------------------------------------------------------------------------------------\n\nTo further verify that DPNTP expression did not interfere with the cortical actin cytoskeleton--dependent strengthening of adhesion, Dox-treated and untreated A431 monolayers were incubated with LtnA before application of mechanical strain. LtnA disrupts the actin cytoskeleton by sequestering G-actin, which subsequently causes depolymerization of F-actin and disrupts adherens junctions ([@bib10])*.* If the observed DPNTP-dependent loss of adhesive strength was solely due to interfering with adherens junction function, one would not expect a dramatic increase in stress-induced dissociation when DPNTP is expressed in the presence of LtnA.\n\nAs expected, Dox-treated A431/C2 monolayers exhibited significant dissociation ([Fig. 8](#fig8){ref-type=\"fig\"} A). LtnA alone also resulted in the dissociation of cell sheets. When the A431/C2 line was treated with 2 \u03bcM LtnA after Dox treatment to induce DPNTP expression, the monolayer dissociated into \\>500 particles. Measurement of LDH release indicated that there was no statistical difference among the samples and, in all cases, the amount of LDH released after mechanical strain was \\<5% of the total amount of LDH expressed.\n\n![**LtnA and DPNTP act synergistically to weaken adhesion.** (A) Dox-treated and untreated A431/SS and A431/C2 monolayers were treated with dispase and subjected to mechanical stress as described above for Fig. 4, with or without a 30-min treatment with 2 \u03bcM LtnA. (B) Again, LDH release was not significantly increased despite the dramatic fragmentation of LtnA-treated monolayers expressing DPNTP. Error bars represent the SEM from an experiment in which each condition was tested in triplicate.](200206098f8){#fig8}\n\nKeratinocytes derived from patients with a recessively inherited COOH-terminal truncation of DP exhibit adhesive defects in vitro\n---------------------------------------------------------------------------------------------------------------------------------\n\nTo further examine the physiological relevance of COOH-terminal DP truncations, we employed keratinocytes that had been isolated from patients with a COOH-terminal truncation of DP due to deletion of the G at position 7622 within the coding sequence of DP (position 7901 of Genbank/EMBL/DDBJ accession no. [M77830](M77830)) ([@bib48]). Histological analysis of patient epidermis revealed widening of intercellular spaces, suggesting the possibility that this truncated form of DP might lead to an adhesive defect. This mutation deletes the C-subdomain containing the plakin repeats and part of the upstream linker region and introduces 18 new amino acids downstream of the deletion ([Fig. 1](#fig1){ref-type=\"fig\"} A). Based on earlier work, we would predict that this deletion would compromise interactions between keratin IFs and the plaque ([@bib60]; [@bib47]). However, the remaining upstream plakin repeat domain and linking motifs may be sufficient to allow a partial interaction with IFs, whereas the recruitment of plakophilins and/or plakoglobin via the DP NH~2~ terminus may provide additional stabilization of IFs ([@bib58]; [@bib41]; [@bib32]).\n\nConsistent with this, immunofluorescence analysis of unstressed keratinocytes growing in vitro did not reveal a dramatic uncoupling of IF tonofibrils from the desmosomal plaque, even though the presence of the truncated DP was confirmed by immunoblot analysis (unpublished data). In addition, no difference in the fluorescence intensity of plaque or transmembrane cadherins was detected ([Fig. 9](#fig9){ref-type=\"fig\"} A) in these cells compared with normal human epidermal keratinocytes (NHEK). However, when subjected to mechanical stress, dispase-released monolayers generated from these keratinocytes exhibited dissociation into particles greater than that seen with NHEK cells ([Fig. 9](#fig9){ref-type=\"fig\"} B). In contrast to A431 cell monolayers, both the control NHEK cells and JD-1 cells exhibited release of LDH, indicating that some cytolysis was occurring ([Fig. 9](#fig9){ref-type=\"fig\"} C). However, there was no difference in the degree of cytolysis between these cultures, suggesting that the greater particle number was due to an additional adhesive defect existing in the JD-1, and not the control, cell lines.\n\n![**JD-1 keratinocytes derived from patients harboring a COOH-terminal truncation of DP exhibit an adhesive defect.** (A) Fluorescence density measurements performed on JD-1 and NHEK cell lines using antibodies directed against \u03b1-catenin (C2081) and DP (DP2.15) were found to be comparable. Confluent JD-1 and NHEK monolayers were analyzed using the dissociation assay. Particle numbers were quantified (B) and LDH release was measured (C). Although both NHEK and JD-1 keratinocytes exhibited more cell lysis compared with the DPNTP-expressing A431 cells, JD-1 cells fragmented into a larger number of particles, supporting the existence of an adhesive defect in the JD-1 cells. Error bars represent the SEM from an experiment in which each condition was tested in triplicate.](200206098f9){#fig9}\n\nDiscussion\n==========\n\nRole of the cytoskeleton in adhesion: uncoupling IFs from desmosomes decreases intercellular adhesive strength\n--------------------------------------------------------------------------------------------------------------\n\nSeveral lines of evidence support a role for the actin cytoskeleton in strengthening adhesion. The catenin complex of intercellular adherens junctions, which mediates the link with actin, is required for E-cadherin--dependent adhesion in vitro and in vivo ([@bib50]; [@bib66]), and the actin-capping drug cytochalasin D prevents cadherin-mediated adhesion ([@bib2]; [@bib33]). These data are consistent with our observation that LtnA abrogates adhesion in the capillary flow assay. Each of the three major catenins, \u03b1-catenin ([@bib68]), \u03b2-catenin ([@bib12]), and p120^ctn^ ([@bib62]; [@bib71]), has been implicated in playing important roles in cell--cell adhesion, although the regulatory role of p120^ctn^ has been somewhat controversial ([@bib49]). In addition, determining the respective contributions of lateral cadherin clustering and physical linkage to the cytoskeleton has been challenging. It has been proposed that linkage to the actin cytoskeleton through \u03b1- and \u03b2-catenin may strengthen initial p120^ctn^ -dependent clustering by subsequent recruitment or nucleation of actin filament attachments and that continued tension between contacting cells requires associated myosin activity ([@bib1]).\n\nLess attention has been focused on the role of IF--plasma membrane associations in regulating adhesive strength. Previous studies demonstrated that desmosomal cadherin tails play a critical role in regulating IF--desmosomal connections and demonstrated that a desmoglein 1--connexin chimera can act in a dominant-negative fashion to uncouple IFs from cell--cell contacts ([@bib63], [@bib64]). In that case, however, a disappearance of endogenous desmosomes was observed. The strategy used here, which employed a truncated DP molecule, allowed us to test the role of the IF--membrane connection in epithelial sheet integrity and adhesive strength in vitro by severing the IF connection while retaining clustered desmosomal cadherins associated with their armadillo family members. Furthermore, as E-cadherin function appeared to be normal in this system, it allowed us to test the relative contributions of IFs and actin cytoskeleton linkages to epithelial sheet integrity. A431 cells induced to express dominant-negative DPNTP exhibited a decreased ability to withstand trituration after aggregation in cell suspension. In addition, DPNTP-expressing confluent cell sheets demonstrated a decreased ability to withstand turbulent mechanical strain as compared with uninduced monolayers, dissociating into many fragments without an increase in cell lysis. In other experiments not shown here, even greater cell sheet dissociation was achieved by using a trituration protocol; however, data from these experiments were difficult to interpret, as trituration often led to cell lysis. Here we show that gentle inversion of the epithelial sheets is sufficient to cause cell detachment without resulting in lysis. Furthermore, higher resolution time-lapse analysis of strained cell sheets revealed that dissociation of DPNTP-expressing cells is more rapid and occurs with less cell distension than control cells. Together these data suggest that although the presence of an intact IF cytoskeleton plays a role in maintaining cellular integrity, connection of this network to the plasma membrane via desmosomes provides strong intercellular adhesion.\n\nContributions of the desmosomal plaque and IF attachment to adhesive strength\n-----------------------------------------------------------------------------\n\nDrawing a parallel to adherens junctions, one might envision that desmosomal strengthening of adhesion would involve at least two major steps: (1) lateral clustering and proper orientation of desmosomal cadherins through proper assembly of the plaque and (2) physical anchoring of the cadherin complex via IFs. In the case of adherens junctions, it has been suggested that lateral cadherin clustering of the extracellular domain, whether through artificial means or via p120^ctn^, can increase adhesive strength ([@bib8]; [@bib70], [@bib71]; [@bib62]). We showed previously that the p120^ctn^-related protein plakophilin 1 cooperates with plakoglobin to cluster desmosomal cadherins in the presence of DP ([@bib6]). Furthermore, the NH~2~-terminal 584 amino acids of DP, which comprise DPNTP, are sufficient for such clustering in L cell fibroblasts ([@bib39]). It seems likely that this clustering event constitutes a necessary step in desmosomal cadherin--dependent adhesive strengthening. Consistent with this idea is the observation that the DP NH~2~ terminus appeared to restore sealing of epithelial sheets and possibly even some measure of IF attachment in DP-null keratinocytes ([@bib30]; [@bib67]). However, it is clear from the present work that even though DPNTP is capable of clustering desmosomal cadherins and forming junctional plaques in A431 cells, adhesive strength is severely compromised in epithelial sheets derived from these cells.\n\nAs DPNTP-dependent clustering is not sufficient to support adhesion comparable to wild-type DP, this ability must require the remaining DP domains, which include the rod and the COOH-terminal IF-binding domains. Previous work suggests that the DP rod contributes to the formation of oligomeric coassemblies with IF polypeptides that build the innermost portion of the plaque ([@bib59]; [@bib37]). Thus, it could be envisioned that loss of these domains may compromise the ability of the plaque to withstand mechanical stress. However, the JD-1 cell experiments suggest that even small deletions of the DP COOH terminus that interfere with IF anchorage are enough to compromise adhesive strength when stress is applied to cells in vitro ([Fig. 9](#fig9){ref-type=\"fig\"} A) and in vivo ([@bib48]).\n\nIF--membrane attachments may enhance adhesive strength in several ways. As has been suggested for actin, IFs may stabilize desmosomal cadherin clusters formed by the armadillo--DP NH~2~-terminal complex, and thus, the proper disposition of desmosomal cadherin extracellular domains. Another possibility is that IFs anchor desmosomal cadherins within the plane of the membrane and that uncoupling of this connection results in molecular extraction of desmosomal cadherins out of the plane of the membrane. Similar molecular extraction has been reported previously for agglutinin receptors from red blood cells ([@bib15]), L-selectin from neutrophils ([@bib56]), and integrins from fibroblasts moving across a substratum ([@bib51]). Another possibility is that the entire plaque might be extracted, as might be suggested by the plaque-associated membrane blebs seen in electron micrographs of the epidermis of DP-null animals ([@bib67]). Whether such blebs seen in vivo would be enough to lead to cytolysis is not clear, but if this sort of extraction occurred in A431 cells, membrane resealing would have to be rapid so that LDH release would not occur.\n\nThe actin and IF cytoskeletons contribute synergistically to provide strong intercellular adhesion through classic and desmosomal cadherins\n-------------------------------------------------------------------------------------------------------------------------------------------\n\nDPNTP was previously shown in stable A431 clones to lead to the codistribution of certain adherens junction and desmosome components ([@bib7]), raising the possibility that DPNTP might interfere with adhesive strength solely by interfering with E-cadherin function. However, as shown here, the fluorescence patterns for E-cadherin and Dsg 2 did not detectably differ in uninduced and induced cells ([Fig. 3, B and C](#fig3){ref-type=\"fig\"}), and the dramatic codistribution between desmosomal and adherens junction components that was previously reported in those clonal lines was not observed here. The basis for this difference is not currently known, although it may be related to DPNTP expression levels, clonal variability, or both.\n\nNevertheless, we sought to ensure that the observed adhesive defect was not through effects on E-cadherin. The experiments outlined here show that, in addition to the fact that neither cell surface expression nor distribution was altered, the fraction of E-cadherin present in the detergent-insoluble pool remained constant. This was in contrast to Dsg 2, which was reduced in the Triton-insoluble pool, likely due to the loss of attachment to the IF cytoskeleton. The data presented here also support the idea that E-cadherin is still functional in DPNTP-expressing cells. The ability of DPNTP-expressing cells adhering to capillary tubes coated with the E-cadherin extracellular domain to resist shear stress was no different than that of uninduced cells at early or later time points, even though adhesion was totally abrogated by LtnA. In addition, DPNTP-expressing epithelial sheets treated with LtnA exhibited a dramatically increased dissociation, which would be predicted only if IF attachments have an additional strengthening effect beyond that contributed by the cortical actin--adherens junction complex. Thus, the decreased structural integrity of DPNTP-expressing monolayers is not simply due to interference with E-cadherin function.\n\nIn fact, as the combined effect of LtnA treatment and DPNTP expression is much greater than the sum of the effects of the separate treatments, our data suggest that the actin and IF cytoskeletons function synergistically to provide strong intercellular adhesion and maintain structural integrity of cell sheets. These data are consistent with earlier work demonstrating that the microfilament and IF cytoskeletons are physically associated at the ultrastructural level in cells. Indeed, our earlier work in primary mouse keratinocytes demonstrated an intimate relationship between the cortical actin circumferential ring associated with the perpendicular bundles emanating from adherens junctions and keratin IF bundles that looped in and out of the cortical actin ([@bib27]). More recent work suggested that F-actin serves as a template for cytokeratin organization in vitro ([@bib69]), and such interactions might in fact be stabilized by other members of the plakin family of cytolinkers, which may cross-link these systems laterally ([@bib18]).\n\nObservations from DP-null keratinocytes support the idea that the DP--IF network plays an important role in the maturation of the keratinocyte microfilament cytoskeleton ([@bib67]). Cells cultured from these animals had severely reduced desmosome numbers and, in addition, failed to undergo the normal maturation of the cortical actin cytoskeleton, which might in turn contribute to the observed failure to form epithelial sheets in vitro and skin fragility in vivo. In the current work, the actin microfilament system did not appear to be significantly altered in DPNTP-expressing cells or JD-1 keratinocytes (unpublished data), possibly due to restoration of certain DP NH~2~-terminal--dependent functions ([@bib30]). Together these previous observations and the data presented here suggest two related but independent points. First, DP, possibly through its NH~2~-terminal head domain, plays an important role in the development and maturation of the actin cytoskeleton. Second, actin and IF cytoskeleton connections to cadherins function to synergistically strengthen adhesion, possibly through anchoring cadherins in the plane of the membrane, thus maintaining epithelial cell sheet integrity.\n\nDP truncations in human disease\n-------------------------------\n\nEpithelial sheets formed from JD-1 keratinocytes derived from patients harboring a short COOH-terminal truncation of the IF-binding domain also show an adhesive defect. This is supported by data showing an increase in cell sheet dissociation compared with normal keratinocytes, without any additional increase in LDH release. That some LDH release occurs in this case is not surprising based on the fact that these cells have a much milder defect than the genetically modified A431 cells, and thus some cells may break apart during the increased number of inversions required to release particles in the assay. The JD-1 data suggest that perturbation of the cytoplasmic plaque and its IF connection leads to the weakening of cell adhesion that underlies the cell separation observed in patients harboring this truncated protein ([@bib48]) and that the cellular consequences are distinct from the largely cytolytic defects caused by mutant IF proteins.\n\nThis idea is consistent with observations seen in patients with epidermolytic palmoplantar keratoderma (EPPK) or non-EPPK (NEPPK), a noncytolytic form of EPPK. Some EPPK patients possess mutations in keratin K9 that cause skin blistering on the palms of the hand and soles of the feet with characteristic cytolysis in the suprabasal layers ([@bib35]). On the other hand, in one NEPPK study, a mutation was discovered in keratin K1 at a position in the head domain that, in keratin K5, is believed to interact with DP ([@bib34]; [@bib37]). These patients do not appear to undergo the same sort of skin blistering exhibited by the EPPK patients. Similar phenotypic traits were observed in patients with mutations in desmosomal molecules, such as plakophilin 1 ([@bib44]). These observations are consistent with a hypothesis that the presence of an IF cytoskeleton is responsible for maintaining the structural integrity of the cell, whereas connection of this network to the plasma membrane via DP may be responsible for maintaining or modulating intercellular adhesion.\n\nTogether the results presented in this paper demonstrate the first direct physical evidence to support a role for the IF--desmosome connection, independent of desmosomal cadherin clustering, in regulating the adhesive strength of desmosomes. Furthermore, the data suggest that the IF--desmosome connection works synergistically with the actin--adherens junction complex in providing adhesive strength to epithelia.\n\nMaterials and methods\n=====================\n\nGeneration of cDNA constructs\n-----------------------------\n\np700, a CMV promotor--based construct expressing DPNTP with a COOH-terminal FLAG tag (DPNTP--FLAG), was previously described ([@bib41]). DPNTP--FLAG was inserted into a shuttle vector by subcloning the EcoRI--HindIII fragment of p700 into pGEM7Zf (p766) (Promega). To generate the construct with DPNTP--FLAG under the control of the tetracycline-responsive element, p766 was digested with EcoRI and BamHI, and the resulting fragment was ligated into the same sites in the MCS of pTRE (CLONTECH Laboratories, Inc.), yielding p801.\n\nTo create DPNTP--EGFP under the control of the tetracycline response element, PCR was used to amplify a 1.755-kb DPNTP-containing fragment from pTRE-DPNTP--FLAG (p801). This fragment was cloned into the EcoRI/KpnI sites of pEGFP-N1 (CLONTECH Laboratories, Inc.) to create pEGFP-N1 DPNTP. Next, a 2.5-kb DNA fragment containing DPNTP plus EGFP was removed from pEGFP-N1 DPNTP by sequential NotI/NheI digestion. The resulting fragment was subcloned into the NotI/NheI sites of the pTRE2 response plasmid (CLONTECH Laboratories, Inc.), producing vector p992. Constructs were DNA sequenced to verify correct structure.\n\nCell culture and generation of inducible lines\n----------------------------------------------\n\nParental A431 epithelial cells were cultured as previously described ([@bib7]). These cells were transfected with pTet-On (CLONTECH Laboratories, Inc.) using calcium phosphate precipitate and selected using 400 \u03bcg/ml G418 (Mediatech). Expression of the tetracycline-responsive transactivator in drug-resistant lines was determined by a luciferase assay (pTRE-Luc; CLONTECH Laboratories, Inc., and pRL-CMV; Dual-Luciferase Reporter Assay System; Promega). To establish tetracycline-inducible DPNTP--FLAG lines, a single stable line, 68Q71, was cotransfected with p801 and pSV2pac\u0394P. After selection with 400 \u03bcg/ml G418 and 1 \u03bcg/ml puromycin, resistant clones were screened for DPNTP--FLAG by treating with 2 \u03bcg/ml Dox for 24 h. The cells were then lysed in Laemmli sample buffer and analyzed via SDS-PAGE and immunoblotting. To establish a cell line inducibly expressing EGFP-tagged DPNTP, the 68Q71 cell line was transfected with vectors p992 and pSV2pac\u0394P, and puromycin-resistant clones were selected as described above.\n\nJD-1 keratinocytes isolated from patients were immortalized with an HPV16 plasmid (pJ45216) that has early region genes driven by MoMLV-LTR as previously described ([@bib61]). Cells were cultured in DME/Ham\\'s F12 (3:1) supplemented with 10% FCS, 4 mM glutamine, 0.4 \u03bcg/ml hydrocortisone, 0.1 nM cholera toxin, 5 \u03bcg/ml insulin, and 10 ng/ml EGF. NHEK were isolated from neonatal foreskin as previously described ([@bib29]) and were cultured in media M154 (Cascade Biologics, Inc.).\n\nAntibodies\n----------\n\nFull-length DP and DPNTP--FLAG were detected using a rabbit polyclonal antibody, NW161, against the DP NH~2~ terminus ([@bib7]) or poly-FLAG (Sigma-Aldrich) against the FLAG-tagged constructs. The following rabbit polyclonal antibodies were also used: C2206 against \u03b2-catenin (Sigma-Aldrich), C2081 against \u03b1-catenin (Sigma-Aldrich), 795 against E-cadherin (provided by R. Marsh and R. Brackenbury, University of Cincinnati, Cincinnati, OH), and NW6 against the DP COOH terminus ([@bib4]). 1407 is a polyclonal chicken antibody recognizing plakoglobin ([@bib22]). Mouse monoclonal antibodies used in this study were: DP2.15 (a gift from P. Cowin, New York University, New York, NY) against the central rod domain of DP, Dg3.10 against Dsg 2 ([@bib54]), HECD-1 against human E-cadherin ([@bib57]), 6D8 against Dsg 2, 11E4 against plakoglobin, 5H10 against \u03b2-catenin (gifts from M. Wheelock, University of Nebraska, Omaha, NE), and KSB17.2 against keratin 18 (Sigma-Aldrich).\n\nPreparation of whole cell lysates and Triton-insoluble pool\n-----------------------------------------------------------\n\nTo obtain whole cell lysates, cells were homogenized in Laemmli sample buffer, resolved on 7.5% SDS-PAGE gels, transferred to nitrocellulose, and immunoblotted using the following antibodies: 6D8 (1:1,000), 795 (1:1,000), C2081 (1:2,000), 5H10 (1:100), 11E4 (1:1,000), and NW161 (1:5,000). Sample loading was normalized to keratin that was visualized using KSB17.2 (1:2,000). For analysis of the Triton X-100--insoluble pool, cells were lysed in 1% Triton X-100 buffer (1% Triton X-100, 145 mM NaCl, 10 mM Tris-HCl, pH 7.4, 5 mM EDTA, 2 mM EGTA, and 1 mM PMSF) followed by centrifugation (16000 *g*, 30 min). The Triton X-100--insoluble pellet was solubilized in Laemmli sample buffer, resolved by SDS-PAGE, transferred to nitrocellulose, and labeled using 795 (1:1,000) and Dg3.10 (1:100).\n\nImmunofluorescence analysis\n---------------------------\n\nCells were seeded onto poly-[l]{.smallcaps}-lysine--coated (0.1 mg/ml, 3 h) glass coverslips. 1 d after plating, samples were induced with Dox (2 \u03bcg/ml, 24 h). Coverslips were then rinsed in DPBS and fixed in methanol (\u221220\u00b0C, 2 min). Indirect immunofluorescence was performed using KSB17.2 (1:200), poly-FLAG (1:50), NW6 (1:50), DP2.15 (1:100), Dg3.10 (1:10), and HECD-1 (undiluted supernatant). Primary antibodies were detected using Alexa Fluor^\u00ae^488 goat anti--mouse IgG and Alexa Fluor^\u00ae^568 goat anti--rabbit IgG at a dilution of 1:400 (Molecular Probes). Images were captured using a Leica DMR microscope/Orca Hamamatsu digital camera system and analyzed using OpenLab software (Improvision).\n\nFluorescence density at cell borders was determined by measuring fluorescent signal (OpenLab) from regions of intercellular contact in induced and uninduced A431/SS, C2, and G4 cells stained using 795 (1:1,000) and Dg3.10 (1:10) or JD-1 and NHEK cells stained using DP2.15 (1:500) and C2081 (1:2,000). Fluorescence density was calculated by dividing the fluorescent signal (pixels^2^) by the length (pixels) of the measured cell border. Calculations were determined from 10 representative regions of intercellular contact for each cell line.\n\nCell surface biotinylation\n--------------------------\n\nCells treated with or without 2 \u03bcg/ml Dox were grown to confluency in 60-mm dishes. Cultures were washed twice with DPBS and labeled with cell-impermeable EZ-Link\u2122 Sulfo-NHS-LC-Biotin (4 ml at 500 \u03bcg/ml in DPBS, 4\u00b0C, 30 min) (Pierce Chemical Co.) on a rocker. Cultures were then washed twice with DPBS and incubated in 4 ml of DPBS containing 100 mM glycine (4\u00b0C, 20 min) on a rocker. Finally, cells were washed twice with DPBS and lysed with 1 ml RIPA buffer (10 mM Tris-HCl, pH 7.5, 140 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 5 mM EDTA, 2 mM EGTA, 1mM PMSF). Biotin-labeled proteins were precipitated using 40 \u03bcl of UltraLink\u2122 Immobilized Streptavidin (Pierce Chemical Co.) in 1.5-ml Eppendorf tubes overnight at 4\u00b0C on a rotator. The streptavidin beads were washed five times using lysis buffer and then centrifuged (10,000 *g*, 1 min, 4\u00b0C). Precipitated proteins were eluted using Laemmli buffer, boiled, resolved by SDS-PAGE, and immunoblotted using 795 (1:1,000), C2206 (1:4,000), Dg3.10 (1:100), and 1407 (1:5,000).\n\nHanging drop assay\n------------------\n\nHanging drop cultures of aggregated cells were generated from 1 \u00d7 10^3^ cells that were either untreated or pretreated with Dox for 24 h and then dissociated in 1mM EDTA in DPBS without Ca^2+^ and Mg^2+^. Cells were allowed to aggregate over times ranging from 2 h to overnight on the underside of a culture dish as previously described ([@bib62]). Resulting cell clusters were subjected to 30 rounds of pipetting through a 200-\u03bcl Gilson pipette, and the degree of dissociation was quantified by counting the particles after trituration.\n\nDispase-based dissociation assay\n--------------------------------\n\nA431/SS, C2, and G4 cultures were seeded in triplicate onto 60-mm dishes. At \u223c50% confluency, cells were treated with or without 2 \u03bcg/ml Dox. 24 h after reaching confluency, cultures were washed twice in DPBS and then incubated in 2 ml of dispase (2.4 U/ml; Roche Diagnostics GmbH) for more than 30 min. Released monolayers were carefully washed twice with DPBS and transferred to 15-ml conical tubes. Enough DPBS was added for a final volume of 5 ml. Tubes were secured to a rocker and subjected to 50 inversion cycles. Fragments were counted using a dissecting microscope (Leica MZ6). For experiments using LtnA, cell monolayers released by dispase were further incubated for 30 min in 2 ml of dispase solution containing 2 \u03bcM LtnA. Statistical analysis was performed on the average of three separate experiments in the cases of the A431/SS and C2 cell lines and one experiment using G4 cells. Statistical significance (*t* test) was defined as P \\< 0.05. For experiments with JD-1 and NHEK, application of mechanical stress included 50 inversion cycles with either five additional inversion cycles in which the sample tubes were turned rapidly end over end or shaking at 400 rpm for 1 min in a rotary shaker.\n\nLDH release measurements\n------------------------\n\nLDH measurements using the Cytotoxicity Detection Kit (Roche Diagnostics GmbH) were performed according to the manufacturer\\'s instructions on 200 \u03bcl of cell-free supernatant collected after monolayers were subjected to the dispase-based dissociation assay.\n\nVideomicroscopy of dispase-treated wounded cell monolayers\n----------------------------------------------------------\n\nCell lines inducibly expressing DPNTP--FLAG or DPNTP--EGFP were seeded onto 40-mm diameter circular glass coverslips and allowed to grow to confluence either in the presence or absence of 2--4 \u03bcg/ml of Dox. Monolayers were then wounded at the center of the coverslips using a razor blade. Coverslips were then quickly mounted into the FCS2 closed system live-cell chamber (Bioptechs). The cell chamber was then filled with DPBS containing 48 U/ml dispase. Phase contrast and/or fluorescence time-lapse record of tearing of the monolayer was obtained at 40\u00d7 magnification using a Leica DMIRBE inverted microscope fitted with an Orca Hamamatsu (C4742--95) digital camera and analyzed using Openlab software. Images were captured at 4- or 8-s intervals. Aspect ratios of measured cells were determined by dividing the diameter along the axis of stretch by the diameter orthogonal to the axis of stretch. Quicktime movies of the time-lapse record were assembled using Adobe Premier software.\n\nShear flow assay\n----------------\n\nGlass capillary tubes coated with 1 mg/ml protein A (Amersham Biosciences) in DPBS (4--7 h, 4\u00b0C) and blocked with 0.5% casein enzymatic hydrolysate were coated with 25 \u03bcg/ml E-cadherin extracellular domain--Fc fusion protein in HBSS plus 1 mM calcium chloride (HBSS^+^) (overnight, 4\u00b0C). Tubes were then blocked with 5% nonfat dry milk (in HBSS^+^, 2 h, 4\u00b0C) followed by a wash in HBSS^+^. A431/C2 or G4 cells treated with 2 \u03bcg/ml Dox for 36 h were dispersed with 2 mM EDTA in PBS without calcium or magnesium (37\u00b0C, 15 min) and washed in HBSS^+^ containing 0.2 mg/ml RGD peptide. Cells were loaded onto tubes and allowed to incubate for 10 or 45 min. Flow of HBSS^+^ started at 0.5 ml/min and doubled every 30 s, ending with a final flow rate of 25 ml/min. The percentage of cells bound to substrate was determined at each flow rate by video microscopy. For experiments involving LtnA, cells were incubated in 2 \u03bcM LtnA for 30 min before loading onto coated capillary tubes.\n\nOnline supplemental material\n----------------------------\n\nThe supplemental videos are available at <http://www.jcb.org/cgi/content/full/jcb.200206098/DC1>. Videos 1 and 2 correspond to the top and bottom panels of [Fig. 6](#fig6){ref-type=\"fig\"} A, respectively. Dox-treated or untreated DPNTP--EGFP I3 line monolayers were wounded and treated with dispase to release the sheets from the substrate and initiate contraction of the cells surrounding the wound. Tearing of the sheet and cell--cell dissociation were recorded in exposures taken every 4 s with a display rate of 15 frames per second for both Videos 1 and 2. The time stamp in the lower right hand corner of each video clip is displayed in min:sec. See [Fig. 6](#fig6){ref-type=\"fig\"} A for static images at selected time points. Videos 3 and 4 correspond to the top and bottom panels of [Fig. 6](#fig6){ref-type=\"fig\"} B, respectively. Dox-treated DPNTP--EGFP I3 line and a Dox-treated control line were wounded and treated with dispase to release the epithelial sheets as described above for [Fig. 6](#fig6){ref-type=\"fig\"} A. Exposures were taken every 8 s and the movie is displayed at 15 frames per second. See [Fig. 6](#fig6){ref-type=\"fig\"} B for static images at selected time points and [Fig. 6](#fig6){ref-type=\"fig\"} C for quantitative analysis of the cells indicated with asterisks.\n\nThe authors would like to thank all those who have generously contributed antibodies, plasmids, and other reagents, including P. Cowin, R. Marsh, R. Brackenbury, M. Wheelock, K. Johnson, and P. Purkis. Thanks also S. Getsios and A. Kowalczyk for helpful discussion and critical reading of the manuscript.\n\nThis work is supported by National Institutes of Health grants RO1 AR43380 and PO1 DE12328 (project No. 4) to K. Green and R01 GM52717 to B. Gumbiner.\n\nAbbreviations used in this paper: Dox, doxycycline; DP, desmoplakin; Dsg 2, desmoglein 2; EPPK, epidermolytic palmoplantar keratoderma; IF, intermediate filament; LDH, lactate dehydrogenase; Ltn A, latrunculin A; NEPPK, non-EPPK; NHEK, normal human epidermal keratinocytes; SS, single stable.\n\n[^1]: Address correspondence to Dr. Kathleen J. Green, Department of Pathology, Northwestern University Medical School, 303 East Chicago Ave., Chicago, IL 60611. Tel.: (312) 503-5300. Fax: (312) 503-8240. E-mail: <kgreen@northwestern.edu>\n"}
+{"text": "Mitochondria are among the prime suppliers of energy (ATP) in virtually all living cells[@b1][@b2] and empirical evidence suggests that their dysfunction is involved in the pathomechanism of Parkinson's disease (PD)[@b3][@b4][@b5]. Despite considerable progress, the potential role of mitochondrial (dys)function during PD development remains poorly understood. One of the main reasons for this lack of insight is the fact that there is no direct *in vivo* access to nigrostriatal dopaminergic neurons, the loss of which is one of the main pathological features of PD[@b6]. The latter study hypothesized the gastrointestinal tract may be the starting point of the disease and thus enteric neurons may reliably reflect fundamental disease characteristics. Indeed, both Lewy neurites and Lewy bodies (LB), abnormal protein aggregates that develop in nerve cells during PD, have been demonstrated in the submucosal colon nerve layer of patients with early PD[@b7][@b8][@b9][@b10][@b11]. This suggests that pathological aberrations observed in enteric neurons might represent an early stage of PD manifestation[@b7][@b8]. Furthermore, animal studies of enteric neurons suggest a direct involvement of mitochondrial dysfunction in PD. It was suggested that inhibitory enteric neurons are particularly vulnerable to mitochondrial dysfunction induced by Parkinsonian neurotoxins[@b12]. In this sense, mice treated with a chemical inhibitor of the first mitochondrial oxidative phosphorylation (OXPHOS) complex (complex I), displayed delayed gastric emptying in a dosage-dependent manner[@b12][@b13]. Similarly, severe gastrointestinal dysmotility associated with degeneration of enteric neuronal and glial cells was demonstrated in a transgenic mouse model displaying impaired mitochondrial metabolism in these cells[@b14]. Accumulating evidence suggests that cell metabolism and mitochondrial (dys)function are linked to mitochondrial (ultra)structure and motility[@b1][@b15][@b16][@b17][@b18]. Changes in mitochondrial morphology have been demonstrated in fibroblasts of PD patients with LRRK2 or Parkin mutations[@b19][@b20]. This suggests that mitochondrial morphology changes might be similarly indicative of the ongoing disease process in idiopathic PD. Here we investigated this hypothesis by quantitatively comparing 3D mitochondrial morphology and ganglion volume between colon biopsies from PD patients without clinically apparent gastrointestinal dysfunction and age-matched control subjects. We developed an image quantification procedure allowing unbiased 3D analysis of mitochondrial morphology and ganglion volume in immunofluorescence images obtained by confocal microscopy. This revealed that ganglion volume and mitochondrial morphology were aberrant in PD patients and that these changes allowed discrimination between PD patients and healthy individuals.\n\nResults\n=======\n\nThe enrolled subjects were of similar age (PD patients: 70\u2009\u00b1\u20096 years; controls: 65\u2009\u00b1\u20095 years; P\u2009=\u20090.17). There were five male and six female patients and one male and three female controls. In patients the average disease duration at the time of colon biopsy was 5.4\u2009\u00b1\u20095.7 years. The average Unified Parkinson's Disease Rating Scale (UPDRS) motor score describing disease progression equaled 10.7\u2009\u00b1\u20098.1. The mean levodopa equivalent dosage was 630.3\u2009\u00b1\u2009483\u2009mg[@b21]. Following random selection of the biopsies, we analyzed 65 biopsies from PD patients and 41 biopsies from healthy subjects in detail.\n\nImage analysis algorithm and parameters\n---------------------------------------\n\nThe biopsies were processed as illustrated in [Fig. 1](#f1){ref-type=\"fig\"}. For automated segmentation of mitochondria within a ganglion, deconvolved Alexa-488 channel images were adjusted to 16\u2009bit by multiplying each pixel value with the ratio between the theoretical maximum of a 16\u2009bit pixel, namely 2^16^\u2009\u2212\u20091, and the maximum pixel value of the 12\u2009bit image to be adjusted. Mitochondrial pixels were defined using two rules: first, local thresholds had to confirm that local signals obtained from a 5\u2009pixel sized Gaussian filter with a standard deviation of 2 were at least 25% brighter than the surrounding background, as defined by a 10\u2009\u00d7\u200910 average filter. Second, foreground signals in non-uniformly corrected image stacks had to be confirmed via global thresholds according to the Otsu method, in which an optimal threshold is selected by maximizing the measure of separability between foreground and background in terms of gray levels[@b22]. Correction of non-uniform illumination was performed by subtracting 20\u2009\u00d7\u200920 average filtered images from the corresponding deconvolved images. To remove shot noise, which occurs in photon counting devices, connected components fulfilling both mitochondrial pixel recognition criteria were filtered for a volume of at least eight pixels[@b23]. Objects with more than 10^6^ pixels were removed from the mitochondria mask. Mitochondrial feature extraction was restricted to connected components intersecting with the ganglion volume. For segmentation of mitochondrial surface and mitochondrial body, the mitochondrial mask was eroded with a structuring element defined by a center pixel and its 6-connected neighbourhood. Skeletonization of the mitochondrial 3D mask was performed as described before[@b24]. [Figure 1](#f1){ref-type=\"fig\"} and the [Supp. Movie 1](#S1){ref-type=\"supplementary-material\"} shows the resulting mitochondrial volumes, surfaces, eroded bodies, and skeletons in 3D. In addition, we show the mitochondrial state in both healthy controls and PD patients ([Supplementary Figure 1](#S1){ref-type=\"supplementary-material\"}). Overall, 7 features describing mitochondrial morphology in a ganglion were analyzed [Fig. 2](#f2){ref-type=\"fig\"}.\n\nMitochondrial morphometrics\n---------------------------\n\n3D immunofluorescence image stacks were acquired by confocal microscopy from fixed colon biopsies, as described above. These stacks were used to first reconstruct and subsequently quantify 3D ganglion and mitochondrial structural features. The number of mitochondrial objects per ganglion, corrected for ganglion volume ([Fig. 2](#f2){ref-type=\"fig\"}; *MitoCount*), was higher in PD patients than in control subjects ([Fig. 3A](#f3){ref-type=\"fig\"}; P\u2009=\u20090.043). In the whole cohort (including PD patients and controls), there was no difference in *MitoCount* between the right and left colon ([Fig. 3B](#f3){ref-type=\"fig\"}; P\u2009=\u20090.11). Also in PD patients *MitoCount* was not significantly different between the right and left colon ([Fig. 3C](#f3){ref-type=\"fig\"}; P\u2009=\u20090.61). In control individuals, in contrast, *MitoCount* was higher in the right colon than in the left colon ([Fig. 3D](#f3){ref-type=\"fig\"}; P\u2009=\u20090.022). Total mitochondrial volume normalized to ganglion volume ([Fig. 2](#f2){ref-type=\"fig\"}; *MitoVolumeTotal*) is a measure of mitochondrial mass[@b25]. In the pooled colon samples from right and left colon *MitoVolumeTotal* was higher in PD patients than controls ([Fig. 3F](#f3){ref-type=\"fig\"}; P\u2009=\u20090.017).\n\nWhile there was no difference in mean mitochondrial volume ([Fig. 2](#f2){ref-type=\"fig\"}; *MitoVolumeMean*) between PD patients and controls ([Fig. 3E](#f3){ref-type=\"fig\"}; P\u2009=\u20090.197), these volumes showed a significant gradient from left to right colon in PD patients, but not in controls ([Fig. 3G](#f3){ref-type=\"fig\"}; P\u2009=\u20090.028 and [Fig. 3H](#f3){ref-type=\"fig\"}; P\u2009=\u20090.076). Taken together, PD patients displayed a lower *MitoVolumeMean* in the right colon than in the left colon, whereas controls did not.\n\nWe compared the mitochondrial features with the clinical parameters of PD patients. This revealed that decreasing mean mitochondrial volume was linked to an increasing UPDRS motor score, which reflects reduced motor performances ([Fig. 4](#f4){ref-type=\"fig\"}; P\u2009=\u20090.04; rho\u2009=\u2009\u22120.258, Spearman ranked test). Importanly, none of the features was linked to the age of the participants.\n\nGanglion morphometrics\n----------------------\n\nAverage ganglion volume ([Fig. 2](#f2){ref-type=\"fig\"}; *GanglionVolume*) was significantly smaller in PD patients (17233\u2009\u00b1\u200914627\u2009\u03bcm^3^) than in controls (27384\u2009\u00b1\u200922255\u2009\u03bcm[@b3]) ([Fig. 3I](#f3){ref-type=\"fig\"}; P\u2009=\u20090.005). Comparing the right (n\u2009=\u200950) and the left colon (n\u2009=\u200956) of the whole cohort (including PD patients and controls), revealed that the *GanglionVolume* was larger in the right than in the left colon ([Fig. 3J](#f3){ref-type=\"fig\"}; P\u2009=\u20090.004). However, the *GanglionVolume* for the right colon was smaller in PD patients than in controls ([Fig. 3K](#f3){ref-type=\"fig\"}; P\u2009=\u20090.0006). A significant gradient in *GanglionVolume* along the colonic tract was observed in controls but not in patients ([Fig. 3](#f3){ref-type=\"fig\"}N,M). ROC curve analysis using the right-colon *GanglionVolume* revealed a high AUC value of 0.81, indicating that this feature displays high specificity and sensitivity in discriminating between PD patients and controls ([Fig. 3L](#f3){ref-type=\"fig\"}).\n\nResults extended by principal component analysis and support vector machine learning\n------------------------------------------------------------------------------------\n\nFeature-by-feature comparison for the pooled ganglion data revealed that five morphometric features differed between PD patients and controls ([Fig. 5A](#f5){ref-type=\"fig\"}). Also five features differed between PD patients and controls in the left colon ([Fig. 5B](#f5){ref-type=\"fig\"}). In contrast, a single feature differed between PD patients and controls in the right colon ([Fig. 5C](#f5){ref-type=\"fig\"}). Next, we performed a principal component analysis (PCA), that simultaneously considers all variables in a multidimensional space[@b26][@b27][@b28]. In the context of data visualization, PCA is widely used to reduce data dimensionality while retaining most of the variance of the data, thereby allowing graphic visualization[@b26]. It accomplishes this reduction by identifying directions, called Principal Components (PCs), along which the variance of the data is maximal. PCA revealed that a single outlier was present in the dataset ([Fig. 6A](#f6){ref-type=\"fig\"}; green circle). This information was used for the training of the SVM classifier, where parameters were set to assume 1% of outliers. Separate SVM analysis of the overall mitochondrial morphological phenotype in the right (AUC\u2009=\u20090.74) and in the left colon (AUC\u2009=\u20090.71), using the seven mitochondrial features, indicates that mitochondrial morphometrics provide useful information for the discrimination between PD patients and healthy control subjects ([Fig. 6](#f6){ref-type=\"fig\"}B--D).\n\nIntegrating mitochondrial and ganglion morphometrics\n----------------------------------------------------\n\nIntegrating mitochondrial *and* ganglion morphometric information might leverage the classification of PD patients and control subjects. To test this hypothesis we determined whether such an integration increases classification accurary. Right colon samples from patients displayed reduced ganglion volume ([Fig. 5C](#f5){ref-type=\"fig\"}) and an AUC-value with respect to mitochondrial morphometrics of 0.74 ([Fig. 6B](#f6){ref-type=\"fig\"}). When ganglion volume information was combined with information from mitochondrial morphometrics, two thresholds were required for classifying ganglia from PD patients or healthy subjects: one for the SVM score representing the predictive information from mitochondrial morphometrics, and one for ganglia volumes. These thresholds were set using the prior knowledge that 40% of the analyzed ganglia were derived from healthy controls and 60% from PD patients ([Fig. 6E](#f6){ref-type=\"fig\"}). The predictive power of a classifier combining both sets of information reached a sensitivity (true positive rate) of 0.63 and a specificity (1 - false positive rate) of 0.9 ([Fig. 6E](#f6){ref-type=\"fig\"}). This means that this classifier, independently from clinical diagnosis, could identify 63% of patients in this study as true positives. Importantly, 90% of the predicted patients would be correctly diagnosed by this combinatorial classifier while the remaining 10% would be false positives. At the same rate of 10% false positives, the sensitivity for ganglion volume based classification ([Fig. 3L](#f3){ref-type=\"fig\"}) and mitochondria based classification ([Fig. 6D](#f6){ref-type=\"fig\"}) was below 0.63 in the right colon. This reveals that combination of morphometric features from ganglia and mitochondria increases the classification accuracy of PD patients and healthy subjects.\n\nDiscussion\n==========\n\nAccumulating evidence suggests that aberrations in enteric neurons might be an early feature of PD[@b29][@b30][@b31]. To address this hypothesis we compared mitochondrial morphology between *ex vivo* enteric ganglia from PD patients (11 patients, 65 ganglia) and healthy age-matched control subjects (4 subjects, 41 ganglia).\n\nMorphometrics in 3D reveal mitochondrial changes in the enteric nervous system\n------------------------------------------------------------------------------\n\nMicrodissection of colonic submucosa, introduced by Lebouvier *et al*.[@b9][@b32] leveraged the immunofluorescence based study of ganglia. Here, we developed an image analysis algorithm for the quantification of mitochondrial morphometrics within ganglia volumes. To maximize spatial information retrieval and establish unbiased and reproducible morphometric measurements, the whole 3D image data set was analysed[@b33][@b34].\n\nOur data confirms that progressive mitochondrial fragmentation occurs in parallel with progressing motor deficits. An even more robust correlation with clinical findings may be foreseen in larger cohorts, with recruitment of patients in late stages of the disease, and with directly PD-linked gastrointestinal symptoms. The observed reduced ganglion volumes in PD patients are in line with the idea of enteric neurodegeneration as an early pathogenic event. Mitochondrial fragmentation, increased number of mitochondria, and increased total mitochondrial mass suggest a functional connection between mitochondrial morphofunction and ganglion degeneration.\n\nRostrocaudal gradient of ganglion morphometrics in PD patients\n--------------------------------------------------------------\n\nThe observed changes differed between the right and left colon of PD patients. Although PD patients did not display clinically apparent gastrointestinal dysfunction, their right colon ganglia had a reduced volume. Importantly, this right colon ganglion degeneration had predictive power to discriminate between PD patients and control subjects. The selective ganglion volume reduction in the right colon is in line with the previously proposed rostrocaudal gradient of enteric neuropathology along the digestive tube of PD patients[@b35][@b36], characterized by a more pronounced pathology in proximal parts of the gastrointestinal tract, such as esophagus, than in distal parts of the gastrointestinal tract, such as the colon. We observed in this study that even within the colon such a gradient can be shown. Both healthy individuals and patients displayed a lower mean mitochondrial volume in the right colon relative to the left colon. This difference was significant for patients and borderline significant in the control group. To elucidate if mitochondrial changes are predominant in the right or left colon, each colon region was directly compared between patients and controls, as discussed in the next section.\n\nLack of neuroprotective mitochondrial compensation in the right colon\n---------------------------------------------------------------------\n\nStatistical comparison of single mitochondrial features between patients and healthy subjects only revealed significant differences in the left colon. Mitochondrial changes in the right colon could only be demonstrated using machine learning techniques that combined multiple features. This might suggest that significant mitochondrial structural changes in the right colon are absent or have already occurred at an earlier timepoint. Mitochondrial fragmentation has been proposed as one of the earliest subcellular markers of neurodegeneration[@b37]. This is in line with the observed mitochondrial fragmentation in the present cohort composed of PD patients at a rather early disease stage. However, the findings should be confirmed in PD patients at a yet earlier motor stage or in subjects with REM sleep behavior syndrome, an established forerunner syndrome of PD[@b38]. The increase in total mitochondrial mass in the left colon of PD patients might reflect a compensatory mechanism to sustain mitochondrial functioning[@b39][@b40]. Alternatively, increased mitochondrial mass might be due to impaired mitophagy[@b41][@b42].The normal ganglion volumes observed in the left colon of PD patients support the idea of successful mitochondrial adaptation. Indeed, in fibroblasts derived from patients with Parkin mutation, paraquat exposure increased the mitochondrial mass but directly reduced mitochondrial branching and ATP production[@b43].\n\nThe finding of intact ganglia in the left colon and shrunken ganglia in the right colon suggests a two-stage process: at first mitochondrial changes are efficiently compensatory and neuroprotective, but they are inexorably followed by the breakdown of cellular homeostasis further triggering cellular apoptosis and ganglia shrinking. Such a pathophysiological cascade is in line with the presented rostrocaudal gradient in mitochondrial fragmentation in PD. The rostrocaudal gradient theory is also supported by the fact that microbiota composition varies along the digestive tract[@b44]. A radial gradient in oxygene pressure from the intestinal tissue interface to the intestinal lumen indicates oxygene diffusion from host to microbiota[@b45], which might interfere with oxidative phosphorylation in the enteric submucosa. Indeed, it is recognized that the brain-gut axis including central, autonomic, and enteric nervous systems is significantly modulated by gut microbiota[@b46].\n\nPotential links between morphometric changes, alpha-synuclein, and pathogenesis\n-------------------------------------------------------------------------------\n\nIn complement to morphological changes in ganglion mitochondria, reported in the present study, prior studies reported abnormal protein aggregates in the colonic neurons of early-stage PD patients[@b7][@b8][@b9][@b10][@b11]. It has been proposed that misfolded alpha-synuclein directly alters mitochondria and their associated membranes[@b47][@b48]. In this perspective, the observed mitochondrial morphology changes are also in line with established neuropathological involvement of the enteric nervous system in PD[@b29][@b30][@b31].\n\nDue to lack of material, we did not evaluate the role of LB burden, so far considered to be primordial[@b7][@b8][@b9]. Although not uncontested, it had been proposed that ganglionic cell degeneration may be secondary to LB burden. Esophageal LB were first identified in PD patients with dysphagia[@b49]. Later LB were found in ganglia cells from the colonic myenteric plexus in a PD patient suffering from megacolon[@b29]. The presence of LB in other parts of the gastrointestinal tract was confirmed by others ever since[@b30][@b32]. Clinically asymptomatic patients with Braak stage 2 brain pathology also show LB pathology in ganglia located in the intramural and submucosa layers[@b50]. However, none of these studies analyzed mitochondrial alterations[@b51][@b52]. To the best of our knowledge, the present study is the first *ex vivo* study exploring mitochondrial morphometrics in the colon of PD patients.\n\nThe methodological approach is innovative for the field, but, importantly, requires deeper biopsy samples than usually collected. The thereby increased risk for injury and the true positive rate of only 63% vetoes the use of this method for large scale diagnostic screening. However, as colonoscopy is, at least in some countries, systematically used for colon cancer screening, and as the methods that we present here provide low false positive rates, 3D ganglion and mitochondria morphometrics offer new opportunities for the early validation of preliminary PD diagnosis. Furthermore, the presented approach could leverage the study of mitochondrial morphometrics in other neurodegenerative diseases such as Alzheimers's disease, and Huntington's disease, which have also been linked to mitochondrial dysfunction[@b53]. While mitochondrial sizes are at the resolution limit of conventional microscopy, our sophisticated method using advanced confocal microscopy, deconvolution, and computational image analysis, successfully manages this barrier and yields reproducible quantification of mitochondrial morphometrics within the ganglia. However, confirmation in larger cohorts and by another imaging approach such as electron microscopy or indirect proof by metabolic studies are essential.\n\nWe are aware of the caveats and restrictions of such a pioneer study. Mitochondrial morphometrics give only a snapshot view on the complex and rapidly ongoing process of mitochondrial dynamics. Within ganglia the view is limited to the direct surroundings of the nucleus and does not extend beyond the ganglia limits, to distal parts of the axons, and most importantly to the synaptic level[@b54]. Furthermore, the enteric neurons at the submucosal level essentially modulate mucosal processes, while those at the myenteric level control the motor activity, i.e., are responsible for dysmotility syndromes including constipation as seen in PD[@b55]. Separate analysis of the different enteric neuron subpopulations, such as inhibitory motor neurons, ascending interneurons and excitatory longitudinal muscle motor neurons may further enhance the discriminative power[@b56]. Future studies should include patients with overt dysmotility syndrome and patients at later motor stages of disease.\n\nConclusion\n==========\n\nStudying mitochondrial morphology in enteric neurons of PD patients is appealing as there are numerous similarities between enteric neurons and striatal dopaminergic neurons: high complexity and branching, low level of myelination, and long axon size. In both cell types mitochondria have to travel a long distance from the cell body to the dendritic or synaptic processes. Mitochondrial dynamics have to be performant, both in terms of adequate organelle fusion/fission and transport to distant cellular locations. High adaptability mediating robustness against mitochondrial stress has thus to be presumed, already in the healthy condition. With this analogy between enteric and striatal neurons in mind, our study is the first *ex vivo* proof in humans on mitochondrial dysfunction in Parkinson's disease by detecting altered mitochondrial morphometrics in the colon.\n\nMethods\n=======\n\nRecruitment procedure, tissue preparation, staining and fluorescence microscopy\n-------------------------------------------------------------------------------\n\nColonoscopy in Luxemburg is offered within a colon cancer-screening program to individuals between the age of 50 and 80. Here this procedure was carried out with consecutive, non-demented patients with idiopathic PD, according to the London Brain Bank criteria[@b57]. These PD patients displayed no clinically apparent gastrointestinal symptoms and their age was within the above range. Healthy, age-matched volunteers were recruited as controls. All participants had given informed written consent prior to participating in this study. The study, the detailed experimental protocols as well as the information sheet for the patients were authorized and approved by the National Research Ethics Committee of Luxembourg (decision 201107/03), according to the rules established by the International Conference on Harmonisation (ICH) and the Declaration of Helsinki. All subjects underwent a standardized examination program as previously described[@b58]. All biopsies were taken by the same experienced gastroenterologist (W.J.). For each individual, at least one biopsy was obtained from the proximal/right colon tract and at least one biopsy was taken from the distal/left colon tract. Biopsies were stretched and pinned flat using needles (FST\\#26002-10, Minutien Pins, FST) using established protocols[@b9][@b32], as illustrated ([Fig. 1B](#f1){ref-type=\"fig\"}). Submucosa was mechanically dissected from the mucosal tissue using a stereomicroscope and watchmaker's forceps. The submucosa samples were fixed in 4% w/v paraformaldehyde for 2\u2009h at room temperature. After three subsequent washing steps (10\u2009min each), the specimen was permeabilized for 1.5\u2009h in Dulbecco's Phosphate Buffered Saline (D8537, SigmaAldrich, MO, USA) containing 1% v/v Triton X-100 (T8787, SigmaAldrich, MO, USA) and 1% w/v bovine serum albumin (A2058, SigmaAldrich, MO, USA). After another short wash, it was incubated for 45\u2009minutes in a blocking solution (dPBS\u2009+\u20095% v/v goat serum, S-1000, Vector laboratories, Inc., CA, USA). We applied primary antibodies including anti-neurofilament-L (C-term) rabbit monoclonal antibody (Cat. \\#04-1112, Millipore), and anti-mitochondria mouse monoclonal antibody (Cat. \\#MAB1273, Millipore). These were diluted in dPBS containing 1% v/v Triton X-100\u2009+\u20091% w/v BSA and incubated for 12\u2009hours at a temperature of 4\u2009\u00b0C. Following incubation with the primary antibodies, samples were washed three times with dPBS (10\u2009minutes each) and then incubated for three hours at room temperature with secondary antibodies: Alexa-555 goat anti-rabbit (Cat. \\#A21428, Invitrogen) and Alexa-488 goat anti-mouse (Cat. \\#A11001, Invitrogen), diluted 1:500 in dPBS. After three additional washing steps with dPBS, the submucosa was mounted on a microscope glass slide (Fluoroshield with DAPI, Sigma-Aldrich) and covered with a cover glass. Images were acquired with a confocal microscope (LSM 710, Zeiss, Germany). Each biopsy slide was visually inspected by the operator using a 20\u00d7 air objective (Zeiss, Plan-Apochromat, NA\u2009=\u20090.8) to locate ganglia. After this step, the area containing ganglion material was scanned using an objective with a higher magnification (63\u00d7, oil, Zeiss, Plan-Apochromat, NA\u2009=\u20091.4), resulting in a 3D image stack. The pinhole was set to acquire 1.5\u2009\u03bcm sections. The size of z-steps was set to 0.18\u2009\u03bcm. DAPI was excited using the 405\u2009nm laser line. Fluorescence between 410--483\u2009nm was filtered using a 3-Channel Quasar Detection Unit and detected using a photomultiplier tube (PMT). Alexa-488 was excited using the 488\u2009nm laser line and fluorescence was detected using a 500--550 bandpass filter and a gallium arsenide phosphide detector. Alexa-555 was excited with a 561\u2009nm laser and fluorescence was detected using a 565--610 bandpass filter and a gallium arsenide phosphide detector. All lightpaths passed a 488/561 main beam splitter. For DAPI, an additional 405 beamsplitter placed between 405 laser, main beam splitter, and sample was used. The gallium arsenide phosphide detectors were placed behind a 545 long pass filter while the light reflected from this filter was targeting the DAPI detector. All used filters, beamsplitters, and detection units are from Zeiss. The detection mode was set to photon counting. The size of acquired images was adjusted to ganglion size. Alexa-488 fluorescence was acquired in a first scan. DAPI and Alexa-555 fluorescence were acquired simultaneously in a second scan.\n\nImage analysis\n--------------\n\n3D image stacks were imported into AutoQuant X (version X3.0.3 64-bit edition, Media Cybernetics, Rockville, MD, USA) and deconvolved using an adaptive point spread function (10 iterations). Deconvolved stacks were further processed in Imaris 8.1.2 (Bitplane, Switzerland). Iso-surfaces of ganglia were reconstructed by manually selecting regions of interest (ROIs), corresponding to the convex hull of cells containing neurofilaments, for each individual slice. To export the ROI to Matlab 2014a (Mathworks, MA, USA), 3D surfaces were converted to ganglion volume masks. The image analysis procedure for mitochondrial analysis in Matlab, as well as the underlying settings for image processing, are presented in detail in the Results section.\n\nStatistical analysis\n--------------------\n\nAll investigators (except statisticians) were blinded for the identity of the barcoded samples. In total, a number of 89 anonymized biopsies (48 from the right colon and 41 from the left colon) were analyzed. In these biopsies 329 ganglia were identified (175 from the right and 154 from the left colon), 106 of which were randomly chosen for analysis using the default pseudorandom number generator provided by the R package, which uses the Mersenne-Twister algorithm. Whenever possible, at least two ganglia from the right and left colon were included for each subject. Significance of the differences between features were evaluated using a permutation test[@b59]: For comparing vectors of features, the true absolute mean difference was calculated. Then, the labels between compared vectors were randomly shuffled and the absolute mean difference was calculated. The random shuffling was repeated 100000 times. The P-value was calculated by comparing the true absolute mean with the absolute means from random shuffling. The classic approach to evaluate the diagnostic performance of a potential biomarker(s) is the receiver operating characteristic (ROC) analysis[@b60][@b61], which plots the true-positive rate (sensitivity) against the false-positive rate (1-specificity) for a given biomarker or a combination of biomarkers. We applied a ROC analysis to determine whether mitochondrial and/or ganglial features could be used for discriminating between PD patients and healthy subjects. The area under the ROC curve (AUC) is a measure of the diagnostic performance (1 indicating a perfect diagnostic power while 0.5 indicates a futile biomarker). In addition to the above approaches, we used principal component analysis (PCA) for the data visualization, and linear support vector machines (SVM)[@b62] for binary classification of PD patients and healthy subjects. To this end the SVM model was first trained using the whole dataset. The predictions obtained were then validated by running 100 experiments of 5-fold cross-validation using a linear SVM classifier. In this approach, each iteration uses 4/5 of the dataset to train the SVM classifier that determines how to split between classes (i.e., PD patients and healthy subjects), and then evaluates the rates of true and false positives in the remaining dataset. This way we ensure the independence between the training and test set, essential to ensure the generalization of the approach. Both PCA and SVM analysis were done in Matlab.\n\nAdditional Information\n======================\n\n**How to cite this article**: Baumuratov, A. S. *et al*. Enteric neurons from Parkinson's disease patients display *ex vivo* aberrations in mitochondrial structure. *Sci. Rep.* **6**, 33117; doi: 10.1038/srep33117 (2016).\n\nSupplementary Material {#S1}\n======================\n\n###### Suplementary Figure 1\n\n###### Supplementary video file\n\n###### Supplementary movie legend\n\nWe thank Prof. Dr. Pieter VandenBerghe, KU Leuven (Belgium) for constructive discussions. This study has been supported in part by Foundation Think, Luxembourg.\n\n**Author Contributions** Conceived and designed the experiments: B.A.S., A.P.M.A. and D.N.J. Performed colonoscopy: W.J. and L.L. Performed the experiments: B.A.S., A.P.M.A. and L.L. Analyzed the data: B.A.S., A.P.M.A., O.M., H.F. and S.L. Contributed reagents/materials/analysis tools: A.L. and L.L. Wrote the manuscript and prepared figures: B.A.S., A.P.M.A., O.M., H.F., S.L., D.P., D.N.J. and K.W.J.H. All authors reviewed the manuscript.\n\n![Workflow and image analysis procedure: The summarized workflow from colonoscopy to the analysis of 3D mitochondrial morphology is shown in (**A**). Schematic location of the submucosa layer and colon biopsy processing steps (**B**). Stretched biopsy represented as mosaic composed of 77 stitched images from the neurofilament channel (**C**). The arrows point to individual ganglia. Scale bar\u2009=\u2009500\u2009\u03bcm. Image analysis steps (**D**): Mitochondrial volumes were defined via computational image analysis (D1abc). Ganglion masks were defined via manual surface selection from the neurofilament channel (D2ab) and transformed to volume information (D2c). All mitochondria not in contact with ganglia were removed (D3). Morphological analysis on segmented mitochondrial volumes was based on the distinction between mitochondrial surface and body (D3b) and on the analysis of mitochondrial branching (D3c). Branching analysis was based on the count of mitochondrial nodes. In the example highlighted in red, not only the circled branchpoint but also the mitochondrial endpoints labeled 1, 2, and 3 are counted as nodes. Node degrees, as defined by the adjacency matrix proposed by Kerschnitzki *et al*.[@b24], do not only count branches connected to a node (3 in this example), but the cumulated count of skeleton pixels in these branches. Scale bars for D1abc, D2abc, D3: 20\u2009\u03bcm. Scale bars for D3abc: 0,5\u2009\u03bcm.](srep33117-f1){#f1}\n\n![Morphometric and clinical features used in this work.](srep33117-f2){#f2}\n\n![Mitochondria and ganglia morphometrics of enteric neurons from patients and controls.\\\n(**A**) Mitochondria count per volume of patients and controls (p\u2009=\u20090.043). (**B**) Mitochondria count per volume between left and right colon location independently of disease status (p\u2009=\u20090.11). (**C**) Mitochondria count per volume compared between left and right colon location from patients only (p\u2009=\u20090.61). (**D**) Mitochondria count per volume compared between left and right colon location from controls only (p\u2009=\u20090.022). (**E**) Mean mitochondria volume of patients and controls (p\u2009=\u20090.197). (**F**) Normalized mitochondrial mass of patients and controls (p\u2009=\u20090.017). (**G**) Mean mitochondria volume comparison between left and right colon location from patients only (p\u2009=\u20090.028). (**H**) Mean mitochondria volume comparison between left and right colon location from controls only (p\u2009=\u20090.076). (**I**) Ganglia volume of patients and controls (p\u2009=\u20090.005). (**J**) Ganglia volume comparison between left and right colon location independently of disease status (p\u2009=\u20090.004). (**K**) Ganglia volumes from the right colon of patients and controls (p\u2009=\u20090.0006). (**L**) ROC analysis. The corresponding p-value 2.3e-4 indicates the random chance that the AUC is not different from 0.5 (null hypothesis: AUC\u2009=\u20090.5). (**M**) Ganglion volumes of patients compared between left and right colon (p\u2009=\u20090.33). (**N**) Ganglion volumes of controls compared between left and right colon (p\u2009=\u20090.002). All features are plotted on log scales (**A**--**K**,**M**,**N**). AUC\u2009=\u2009area under the curve; PD\u2009=\u2009Parkinson's disease; ROC\u2009=\u2009receiver operating curve.](srep33117-f3){#f3}\n\n![Analysis of correlation between clinical patient data, mitochondrial features, and ganglia volumes.\\\nMitochondrial features are described in [Fig. 2](#f2){ref-type=\"fig\"}. GV, Ganglion volume; SumMMSE, Score from Mini-Mental State Examination; UPDRSIII, Motor part of the Unified Parkinson's Disease Rating Scale; SchwabEngland, Score for individual levels of independence. Spearman's rank correlation coefficients are color coded as indicated in the colorbar. Grouping of similar rows and columns in the matrix of coefficients, as illustrated by dendrograms, was done via hierarchical clustering with Euclidean distance metric and average linkage in Matlab.](srep33117-f4){#f4}\n\n![Ganglion and mitochondrial morphometric features.\\\nComparision of pooled ganglion data from left and right colon between patients and controls (**A**). Comparison of left colon ganglia between patients and controls (**B**). Comparison of right colon ganglia between patients and controls (**C**). All features are plotted on log scale axes and described in [Fig. 2](#f2){ref-type=\"fig\"}.](srep33117-f5){#f5}\n\n![Evaluation of disease status classification.\\\nPrinciple component analysis of mitochondrial features in the whole study cohort (**A**). Each data point represents a single ganglion. Shapes and color codes indicate disease state and colonic sample location as shown in the legend. Evaluation of ganglion classification via mitochondrial morphometrics and support vector machines (**B**--**D**); AUC analysis was based on one hundred five-fold cross validations via support vector machine classification on patient or control ganglia. The combination of underlying mitochondrial scores and ganglion volumes in the right colon is used with the prior information that 60% of samples in this study are derived from patients and 40% from controls. Setting the MitoSVM score threshold to the quantile 0.6 of MitoSVM scores and the threshold for GangliaVolume - larger in patients than in controls - to the quantile 0.4 of GangliaVolumes provides a biomarker based classifier with a sensitivity of 0.63 and a specificity of 0.9 (**E**).](srep33117-f6){#f6}\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction\n============\n\nThe rumen microbiota enables the ruminant to digest cellulose-rich feedstuff and to convert them into a wide range of metabolites that are used by both microbes for their own proliferation and also for the supply of nutrients to the host animal. This symbiotic relationship between the rumen microbiota and host is vital to the ruminant's survival and performance ([@B30]). The diverse microbial community adapts to a wide array of dietary feedstuffs and management strategies ([@B17]). Elucidating the rumen microbiome composition, adaptation, and function is of international interest due to its implications, ranging from climate change to applied animal production and animal health ([@B21]).\n\nClassical knowledge of rumen microbiology was previously based on anaerobic, culture-dependent methods. However, in recent years, the development of next-generation sequencing (NGS) has permitted in-depth sequencing and data analysis of various types of environmental samples ([@B39]). Amplicon sequencing targeting the 16S rRNA in bacteria and archaea remains the most cost-effective and facile tool to provide valuable phylogenetic information for the comparison of microbial diversity in large numbers of environmental samples. Numerous research groups investigating the rumen microbiome have utilized amplicon sequencing as an aid in identifying potential community compositional trends that could be used as an estimation of various performance traits, including methane emission, animal protein production, and ruminant health status ([@B13]; [@B25]; [@B34]; [@B29]; [@B33]). It has been identified that there is no standardization between amplicon studies ([@B36]) and it has been emphasized that this lack of standardization is causing issues with research replication as well as problems translating research into applications ([@B35]). The assessment of US microbiome research also identified a need for standardized protocols between the microbiome research communities in order to facilitate comparisons between studies ([@B39]).\n\nThe method utilized for DNA extraction can lead to dramatic differences in microbial output composition ([@B12]; [@B2]). Hence, validating extraction methods with a mock microbial community is an absolute necessity to ensure accurate representation of microbial communities in samples. The extraction method assessed in the current study is the repeated bead beating and column (RBBC) method ([@B47]) as it is a common method utilized for extraction of DNA from rumen contents ([@B12]; [@B43]; [@B20]). However, this DNA extraction method has not been validated with a mock microbial community standard (MC) to verify its efficiency at extracting accurate representative quantities of DNA from both Gram-positive and Gram-negative bacteria.\n\nComparisons of the rumen bacterial and archaeal community compositions are typically investigated by sequencing one or more of the nine different hypervariable regions of the 16S rRNA gene, due to conservation across the domains of bacteria and archaea ([@B24]; [@B41]). Most NGS amplicon studies in recent years have been conducted on the Illumina MiSeq as this platform gives the highest number of reads at the lowest cost. However, the length of reliable high-quality reads remains at 250 bp for the forward read and approximately 230 bp for the reverse read. Consequently, this allows sequencing of only one or two 16S rRNA variable regions which limits the analysis to genus level identification at best. Despite this limited resolution, amplicon sequencing on the Illumina MiSeq is a very useful approach for screening large numbers of complex microbial community samples to identify samples that show major perturbations in the composition of microbial genera. The choice of a specific hyper variable region can affect both the accuracy and specificity of phylogenetic assignment and the community diversity estimation ([@B19]; [@B46]). In the current study, we tested two commonly used primer pairs -- 515F/806R ([@B4]) and Pro341F/Pro805R ([@B40]) -- that both simultaneously amplify bacteria and archaea. It has been previously shown that simultaneous amplification of archaea and bacteria using multi-domain specific universal primers may not be as accurate as separate analysis of the archaeal and bacterial domain ([@B16]; [@B9]). However, simultaneous amplification of archaea and bacteria is particularly appropriate for rumen microbiomic diversity studies as there are major efforts to develop strategies in cattle and sheep production that reduce the numbers of rumen archaea relative to bacteria ([@B29]). It also facilitates investigation into relationships between microbial taxa from both archaeal and bacterial domains in a convenient, one step-approach, for example, [@B20] previously deciphered an inverse relationship between a member of *Methanobrevibacter gottschalkii* clade and uncharacterized Succinivibrionaceae species utilizing this method. The primer pair Pro341F/Pro805R targets V3--V4 and is a subsequent modification of the primers reported by [@B16]. Pro341F/Pro805R was reported to generate microbial profiles that were more accurate than those obtained with other commonly used bacterial and archaeal primer pairs ([@B40]). The primer pair 515F/806R targets the V4 region of the 16S rRNA gene and was previously validated with a mock community. However, [@B4] utilized an alternative amplification protocol with a single round of polymerase chain reaction (PCR) rather than the two-step PCR protocol used in this investigation.\n\nPolymerase chain reaction cycle conditions are another contentious issue that can cause misrepresentation of microbial community composition. A two-step PCR protocol is routinely used to initially amplify the hypervariable region of interest and, second, to barcode samples in order to facilitate multiplex sequencing ([@B14]). It was our aim to assess our current methodology ([@B20]) against a lower number of PCR cycle conditions for the first amplicon generation step. Superfluous PCR cycles can lead to the formation of PCR by-products ([@B28]) such as chimeras ([@B44]), heteroduplexes, and single-stranded DNA molecules ([@B15]). The classification of sequences is yet another element of consideration for 16S rRNA phylogenetic analysis of bacterial and archaeal communities. For the present study, we used multiple sequence alignment with Basic Local Alignment Search Tool (BLAST) to identify species from DNA that were extracted from mock community cells. This BLAST approach was successful as the mock community is composed of only eight well-characterized bacterial species and two well-characterized yeast species. We also tested the Ribosomal Database Project (RDP) classifier ([@B45]), which is a popular na\u00efve Bayesian classifier ([@B27]) for taxonomic assignment of rumen bacterial and archaeal species ([@B20]; [@B22]; [@B26]). The taxonomic classifier can be used independently and can be encapsulated into other programs and pipelines. The overall objective of the study was to evaluate standard methodology frequently utilized for phylogenetic surveys of the rumen bacterial and archaeal community and assess whether it is giving an accurate representation of microbial communities using a commercially available mock MC and a DNA standard (DS). The mock microbial community utilized for this evaluation is commercially available and is composed of five Gram-positive bacteria (i.e., difficult to lyze): *Lactobacillus fermentum*, *Enterococcus faecalis*, *Staphylococcus aureus*, *Listeria monocytogenes*, and *Bacillus subtilis*, three Gram-negative bacteria: *Escherichia coli*, *Salmonella enterica*, and *Pseudomonas aeruginosa*, and two yeast species: *Saccharomyces cerevisiae* and *Cryptococcus neoformans*.\n\nMaterials and Methods {#s1}\n=====================\n\nDNA Extraction\n--------------\n\nIn order to replicate the DNA extraction protocol identically to our rumen extraction protocol, the ZymoBIOMICS^TM^ Microbial Community Standard (Zymo Research Corp., Irvine, CA, United States) was diluted in a 2-mL microfuge tube (DNA/RNA/DNase/RNase-free; Merck KGaA, Darmstadt, Germany) with molecular grade phosphate buffered saline (PBS; Merck KGaA, Darmstadt, Germany), and the tube and its contents were snap frozen in liquid nitrogen. The frozen tube contents were then homogenized to a fine frozen powder under liquid nitrogen using a pestle and mortar and the frozen powder was stored in a 2-mL microfuge tube at -80\u00b0C. Approximately 250 mg homogenized frozen powder was used for DNA extraction from the mock community. DNA was also extracted from 250 mg of homogenized frozen rumen solid digesta powder sample and molecular grade water as a positive and a negative control (NC), respectively. DNA was extracted using the RBBC method ([@B47]). The DNA quality (i.e., approximation of high molecular weight DNA:degraded DNA and presence/absence of RNA) was assessed on an agarose gel at a normalized concentration of 100 ng/\u03bcL. DNA concentration and purity were assessed with two consecutive readings on the Nanodrop 1000 spectrophotometer.\n\n16S rRNA Gene Amplicon Library Preparation\n------------------------------------------\n\nFive replicate amplicon libraries were generated for both the ZymoBIOMICS^TM^ Microbial Community Standard cells (MC; Zymo Research Corp., Irvine, CA, United States), which were extracted using the RBBC extraction method ([@B47], and the ZymoBIOMICS^TM^ Microbial Community Standard DNA (DS; Zymo Research Corp., Irvine, CA, United States). The ZymoBIOMICS^TM^ Microbial Community Standard DNA (DS) was generated at Zymo by extracting DNA from pure colonies of each of eight bacteria species and two yeast strains and combining the resulting 10 DNA extractions in carefully defined amounts (12% for each of the bacterial DNA extracts and 2% for each of the two yeast DNA extracts). Libraries were also generated for positive (rumen digesta) and negative extraction (molecular grade water) controls, respectively.\n\nPolymerase chain reactions were performed for amplicon libraries with 20 ng of rumen DNA 515F/806R (V4; [@B4]) and Pro341F/Pro805R (V3--V4) primers ([@B40]) and designed with Nextera overhang adapters, using 1X KAPA HiFi HotStart ReadyMix DNA polymerase (Roche Diagnostics, West Sussex, United Kingdom). Both 515F/806R and Pro341F/Pro805R matched to all bacterial DNA targets within MC and DS with no mismatch. Two sets of cycle conditions were applied for the initial PCR (PCR1) stage:\n\n1.  95\u00b0C for 3 min, 20 cycles; 95\u00b0C for 30 s, 55\u00b0C for 30 s, 72\u00b0C for 30 s, and then 72\u00b0C for 5 min\n\n2.  95\u00b0C for 3 min, 28 cycles; 95\u00b0C for 30 s, 55\u00b0C for 30 s, 72\u00b0C for 30 s, and then 72\u00b0C for 5 min.\n\nAmplicons were purified using the Qiaquick PCR Purification Kit (Qiagen, Manchester, United Kingdom). A second PCR step was used to attach dual indices and Illumina sequencing adapters using Nextera XT index kit (Illumina, San Diego, CA, United States). Cycle conditions were 95\u00b0C for 3 min, 8 cycles; 95\u00b0C for 30 s, 55\u00b0C for 30 s, 72\u00b0C for 30 s, and then 72\u00b0C for 5 min. Amplicon generation was validated through visualization on a 2% (w/v) agarose gel ([@B47]). Amplicons were pooled in equal concentrations and gel purified to remove adapter dimers using the Qiagen Gel Extraction Kit (Qiagen, Manchester, United Kingdom). An extra purification with the QIAquick purification kit (Qiagen, Manchester, United Kingdom) was used to remove residual agarose. The pooled purified libraries were measured for purity and quantity on the Nanodrop 1000 and further quantified using the KAPA SYBR FAST universal kit with Illumina Primer Premix (Roche Diagnositics, West Sussex, United Kingdom). The library pool was then diluted and denatured as recommended by the Illumina MiSeq library preparation guide. Sequencing was conducted using 500 cycle MiSeq reagent kits (Illumina, San Diego, CA, United States).\n\nZymo also provided the fastq files (ZE) containing 16S rRNA sequencing data from their extraction of the microbial mock community with the ZymoBIOMCS^TM^ DNA Mini Kit (Zymo Research Corp., Irvine, CA, United States). These data had been generated with primers targeting V3--4 region and the amplicons were sequenced on Illumina MiSeq (2 \u00d7 250 bp; Illumina, San Diego, CA, United States).\n\nSequence Analysis\n-----------------\n\nRead 1 and Read 2 of 16S rRNA gene from all amplicons were quality checked with FASTQC ([@B1])^[1](#fn01){ref-type=\"fn\"}^ and overlapping reads (\\>Q30) were merged using BBMerge from the BBMap package^[2](#fn02){ref-type=\"fn\"}^. The merged sequence outputs were processed and classified using four different methods:\n\n1.  A high throughput NCBI BLAST (version 2.5) of individual sequences of merged reads was conducted against the NCBI 16S ribosomal RNA database.\n\n2.  Merged reads from all samples were also subsequently combined into a single dataset for processing with QIIME (version 1.9). Chimeric sequences were removed via USEARCH using the ChimeraSlayer GOLD database ([@B8]). A combination of *de novo* and reference-based operational taxonomic unit (OTU) identification was carried out using the open reference calling method implemented within QIIME. Sequences were clustered at 97% identity into individual OTUs using UCLUST ([@B6]). A single representative sequence from each clustered OTU was aligned to the Greengenes (version gg_13_8), RDP (version 16), SILVA (Silva_128_release), and NCBI 16S ribosomal RNA database independently. Taxonomic classification for each OTU was determined using the RDP classifier (version 2.12).\n\n3.  Identical sequences were identified from merged sequence reads and a unique representative sequence of the original sequences was used for sequence classification. Classification was conducted with the R package; microclass, using the taxMachine tool and the contax_trim database. Merged ZE sequence reads were also classified using the RDP classifier (version 2.9) in R, with the contax_trim database.\n\nStatistical Analysis\n--------------------\n\nSequence data from amplicon libraries (*n* = 48) were compared to the theoretical composition of the Zymo standards. 16S rRNA gene abundance was calculated from theoretical genomic DNA composition with the following formula: 16S copy number = total genomic DNA (g) \u00d7 unit conversion constant (bp/g)/genome size (bp) \u00d7 16S copy number per genome ([@B48]). Relative abundance was calculated for identified sequences in each sample at genus level. The experiment was conducted as a 2 \u00d7 2 factorial design. MC and DS replicates for each primer set and both cycle conditions examined were analyzed as an experimental unit for statistical analysis (*n* = 5 per unit). Group means were calculated for MC and DS replicates R Studio (v3.1). Spearman's correlation analysis was conducted with hmisc package R Studio (v3.1; [@B11]). A power analysis was conducted prior to experimentation to ensure that technical replication was sufficient to deduce a significant correlation between the technical treatments examined in the current experiment and the theoretical composition of the mock community supplied by Zymo. This utilized a significance level of 0.05, an expected correlation of 0.7, and a power of 0.2 ([@B5]). Similarity was declared as significant if *P* \\< 0.05 and as a trend if 0.05 \\> *P* \u2264 0.1.\n\nSummary of Methods\n------------------\n\nDNA was extracted from MC using the RBBC method commonly used for microbial DNA extraction from rumen digesta samples. 16S rRNA amplicon libraries were generated for the MC and DS using primers routinely used for rumen bacterial and archaeal community analysis. The primers targeted the V4 and V3--V4 region of the 16S rRNA gene and samples were subjected to both 20 and 28 PCR cycles under identical cycle conditions. Sequencing was conducted using the Illumina MiSeq platform. As the bacteria contained in the microbial mock community were well-classified species, and for ease of explanation, we used the results of the BLAST classification to assess the DNA, PCR cycle number, and primer type. Sequence classification methodology was assessed independently.\n\nResults\n=======\n\nDNA Extraction\n--------------\n\nWhen BLAST analysis was used to analyze the MiSeq amplicon sequence data, all eight expected bacterial genera were detected for all combinations of DNA extraction/primer pair/PCR cycle number (**Figure [1](#F1){ref-type=\"fig\"}**). The largest effect on accuracy of genera relative abundances was DNA extraction method. As expected, DS (DNA that was pre-extracted by Zymo from each species then combined in equal amounts) showed significant (*P* \\< 0.05) Spearman correlation with the theoretical mock community relative abundances for all combinations of PCR cycle numbers and primers. MC (DNA extracted using RBBC method) showed non-significant Spearman correlations with theoretical mock community relative abundances (*P* \\> 0.05; **Figure [2](#F2){ref-type=\"fig\"}**). The best result achieved for correlations between MC and the theoretical composition was a tendency toward significance (Spearman = 0.67, *P* = 0.07) when 20 PCR cycles were used in combination with the V4 primers.\n\n![Stack plot comparing the relative abundance of the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequence taxonomy was classified through a high-throughput BLAST search against the NCBI 16S rRNA database.](fmicb-09-01365-g001){#F1}\n\n![Correlation matrix of Spearman correlation coefficient values between the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequence taxonomy was classified through a high throughput BLAST search against the NCBI 16S rRNA database. Values with a positive significant correlation are illustrated with shaded color while insignificant correlations are blank.](fmicb-09-01365-g002){#F2}\n\nInterestingly, correlation of ZE (FASTQ files from a MiSeq amplicon run conducted at Zymo using DNA they extracted from the mock community with their Zymo kit) with the theoretical mock community relative abundance values showed a Spearman correlation of 0.69 (*P* = 0.058) which was similar to the correlation between MC/20 PCR cycles/V4 primers and the theoretical mock community (Spearman correlation of 0.67, *P* = 0.071; **Figure [2](#F2){ref-type=\"fig\"}**). This shows that the RBBC DNA extraction method, amplicon library generation, and MiSeq sequencing that we conducted produced bacterial relative abundance profiles that were equivalent to those produced with the procedures that were conducted by Zymo.\n\nBackground amplification was determined to be the presence of any genera other than the eight expected to be present in the mock community. Background amplification was present in all samples including ZE (**Figure [3](#F3){ref-type=\"fig\"}**). Background amplification was increased in the MC samples in comparison to the DS samples (*P* \\< 0.05; **Table [1](#T1){ref-type=\"table\"}** and **Figure [3](#F3){ref-type=\"fig\"}**).\n\n###### \n\nMean relative abundance of background amplification from mock community cells extracted using the RBBC method for DNA extraction (MC) in comparison to mock community DNA (DS) amplified with two primer sets targeting the V3--V4 and V4 region of the 16S rRNA gene at both 20 and 28 cycles.\n\n                    DNA extraction                        \n  ----------------- ---------------- ------ ------ ------ --------\n  BLAST_V3--V4_20   0.16             0.01   0.57   0.03   \\<0.05\n  BLAST_V3--V4_28   0.96             0.03   1.47   0.06   \\<0.05\n  BLAST_V4_20       0.17             0.01   0.56   0.08   \\<0.05\n  BLAST_V4_28       0.63             0.03   1.19   0.04   \\<0.05\n\nP-values were calculated using a Mann--Whitney U-test to evaluate background amplification from each treatment group (n = 5) resulting from DNA extraction method\n\n.\n\n![Bar chart illustrating background amplification from Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) compared to the theoretical mock community composition. Background amplification was determined to be the presence of any genera beyond the eight expected to be present in the mock community.](fmicb-09-01365-g003){#F3}\n\nPrimer Selection\n----------------\n\nTwo primer pairs were investigated in this study; 515F/806R and Pro341F/Pro805R which targeted the V4 region and the V3--V4 region of the 16S rRNA gene, respectively. As expected, both primer sets detected eight bacterial genera (**Figure [1](#F1){ref-type=\"fig\"}**) and both archaeal and bacterial species were detected in our positive control samples. The relative abundances of the genera present in the DS and the MC amplified using the two primer pairs were evaluated via correlation analysis with the theoretical mock community. The DS bacterial relative abundance profiles for V4 and V3--V4 showed strong positive correlations with the theoretical mock community profiles (*P* \\< 0.05; **Figure [2](#F2){ref-type=\"fig\"}**). Both DS and MC amplified with V3--V4 primers showed a positive correlation to their counterpart amplified with V4 primers (*P* \\< 0.01; **Figure [2](#F2){ref-type=\"fig\"}**).\n\nAt 28 PCR cycle, there was an increase in background amplification of both MC and DS amplified with V3--V4 primers compared to MC and DS amplified with V4 primers (*P* \\< 0.01; **Table [2](#T2){ref-type=\"table\"}**). Background amplification was comparable between MC and DS amplified with V3--V4 primers and MC and DS amplified with V4 primers at 20 PCR cycles (*P* \\> 0.5; **Table [2](#T2){ref-type=\"table\"}**).\n\n###### \n\nMean relative abundance of amplification of unexpected bacteria which results from two primer sets targeting the V3-V4 and V4 region of the 16S rRNA gene which were used for amplification of mock community cells extracted with the RBBC (MC) and mock community DNA amplified at both 20 and 28 cycles.\n\n                Primer comparison                        \n  ------------- ------------------- ------ ------ ------ --------\n  BLAST_DS_20   0.16                0.01   0.17   0.01   NS\n  BLAST_DS_28   0.96                0.03   0.63   0.03   \\<0.05\n  BLAST_MC_20   0.57                0.03   0.56   0.08   NS\n  BLAST_MC_28   1.47                0.06   1.19   0.04   \\<0.05\n\nP-values were calculated using a Mann--Whitney U-test to compare background amplification from each treatment group (n = 5) which results from the primer set utilized\n\n.\n\nPCR Cycle Number\n----------------\n\nTwo rounds of PCR were used for library preparation. The first round of PCR used primers that amplified 16S rRNA gene and the second round added the sequencing adapters and barcodes. Too many PCR cycles are thought to generate amplicon libraries that do not accurately represent microbiome composition. We tested the effect of increasing the number of PCR cycles in the first round of PCR (PCR1) from 20 PCR1 cycles to 28 PCR1 cycles. Relative abundances of taxa derived from 16S rRNA amplicon sequences from DS and MC amplified with 20 and 28 PCR cycles were compared against those of the theoretical mock community composition. The DS amplified at both 20 and 28 PCR1 cycles were significantly correlated with the theoretical mock community composition regardless of which primer pair was used (*P* \\< 0.05; **Figure [2](#F2){ref-type=\"fig\"}**). DS and MC amplified at 20 PCR1 cycles were significantly correlated with their 28 PCR1 cycle counterpart (*P* \\< 0.01; **Figure [2](#F2){ref-type=\"fig\"}**). There was an increase in detection of unexpected taxa at 28 PCR1 cycles in comparison to samples amplified at 20 PCR1 cycles for both DS and MC (*P* \\< 0.01; **Figure [3](#F3){ref-type=\"fig\"}** and **Table [3](#T3){ref-type=\"table\"}**).\n\n###### \n\nMean relative abundance of background amplification resulting from the number of PCR cycles during library preparation; 20 and 28 PCR cycles, constructed from DNA extracted from mock community cells extracted using the RBBC method (MC) and mock community DNA (DS) amplified with two primer sets targeting the V3--V4 and V4 region of the 16S rRNA gene.\n\n                    PCR cycle number                        \n  ----------------- ------------------ ------ ------ ------ --------\n  BLAST_DS_V4       0.17               0.01   0.63   0.03   \\<0.05\n  BLAST_DS_V3--V4   0.16               0.01   0.96   0.03   \\<0.05\n  BLAST_MC_V4       0.56               0.08   1.19   0.04   \\<0.05\n  BLAST_MC_V3--V4   0.57               0.03   1.47   0.06   \\<0.05\n\nP-values were calculated using a Mann--Whitney U-test to evaluate background amplification from each treatment group (n = 5) resulting from PCR cycle number\n\n.\n\nDetection of Unexpected Taxa\n----------------------------\n\nUnexpected taxa were detected in all samples at relative abundances that varied between 0.12% and 1.6% (**Figure [3](#F3){ref-type=\"fig\"}**). A higher number of PCR1 cycles (i.e., 28 PCR1 cycles versus 20 PCR1 cycles) clearly caused higher relative abundances of unexpected taxa (**Figure [3](#F3){ref-type=\"fig\"}**). The highest and lowest unexpected taxa detection was in the MC_V3-V4_28 and DS_V4_20, respectively. Unexpected genera detection was one to three times higher in MC than DS in all cases when the same primer/PCR cycle number combinations were compared. This indicates that this \"background noise\" was influenced by the RBBC extraction method. Primers also influenced unexpected taxa detection. At 28 PCR1 cycles (but not 20 cycles), there was a significant (*P* \\< 0.01) increase in detection of unexpected genera in both MC and DS amplified with V3--V4 primers compared to MC and DS amplified with V4 primers. V4 primers amplified different unexpected taxa than V3--V4 (**Figure [3](#F3){ref-type=\"fig\"}**).\n\nSequence Classifier\n-------------------\n\nBasic Local Alignment Search Tool analysis of amplicon sequences against the NCBI 16S rRNA database identified the eight bacterial genera present in the mock community from DS, MC, and ZE (**Figure [1](#F1){ref-type=\"fig\"}**). The ZE, DS, and MC amplified with the V4 primers all showed significant correlation to the theoretical mock community (**Figure [2](#F2){ref-type=\"fig\"}**). Classification was consistent with the theoretical composition with minimal background amplification detected (\\<2%; **Figure [3](#F3){ref-type=\"fig\"}**). Although suitable for this experiment due to the well-defined nature and convenient number of the mock community components, BLAST in conjunction with the NCBI database is seldom used for 16S rRNA phylogenetic analysis of less well-defined communities, such as the rumen.\n\nOTU-Based Approach\n------------------\n\nMore commonly used for taxonomic classification of 16S rRNA amplicon sequences in rumen phylogenetic studies is the generation of OTUs from sequences that share 97% identity and classification of the OTUs with the RDP classifier. This approach was investigated independently in combination with four databases: Greengenes, NCBI-16S-rRNA, RDP, and SILVA. The RDP classifier identified all eight bacterial genera from DS, MC, and ZE (**Figure [4](#F4){ref-type=\"fig\"}**). However, a large proportion of the 16S rRNA OTUs that were classified as members of the Enterobacteriaceae family were not classified as far as genus level in DS, MC, and the ZE using the RDP classifier. These OTUs were assigned to \"Enterobactericeae_other\" (**Figure [4](#F4){ref-type=\"fig\"}**). There were two genera from this family within the mock community: *Escherichia* and *Salmonella*. Representation of these two genera was inaccurate in all of the relative abundance profiles that were generated from OTUs.\n\n![Stack plot comparing the relative abundance of the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequences were clustered into OTUs at 97% identify and taxonomic classification of OTUs was conducted with the RDP classifier within QIIME analyses utilizing the **(A)** Greengenes, **(B)** SILVA, **(C)** RDP, and **(D)** NCBI 16S rRNA databases.](fmicb-09-01365-g004){#F4}\n\n*Salmonella* was identified using the RDP database for the MC (\u223c13%) and DS (\u223c12%) at 20 and 28 PCR cycles when amplified with V3--V4 primers. However, *Salmonella* was not identified for ZE (V3--V4 amplicon sequence files supplied by Zymo) using the RDP database for sequence identification despite the same region (V3--V4) of the 16S rRNA gene being amplified (**Figure [4C](#F4){ref-type=\"fig\"}**). The majority of OTUs from DS and MC amplified with the V4 primers which were identified as belonging to the Enterobactericeae family with the RDP database were not classified to genus level, with less than 0.2% of sequences identified as *Salmonella* (**Figure [4C](#F4){ref-type=\"fig\"}**).\n\nUtilizing the Greengenes and NCBI databases also showed that OTUs spanning the V3--V4 region of the 16S rRNA gene increased *Salmonella* identification in comparison to the V4 sequences from MC and DS but this increase was only small and the *Salmonella* genus was severely underrepresented with both V4 and V3--V4 primers for Greengenes and NCBI (**Figures [4A,D](#F4){ref-type=\"fig\"}**). *Salmonella* was equally underrepresented in all samples (\\<0.1%) regardless of OTU sequence length when the SILVA database was used (**Figure [4B](#F4){ref-type=\"fig\"}**).\n\n*Enterococcus* remained almost absent from the MC and DS samples identified using the Greengenes database irrespective of primer type used for amplification (**Figure [4A](#F4){ref-type=\"fig\"}**). *Enterococcus* was identified in the correct proportions from the ZE (**Figure [4A](#F4){ref-type=\"fig\"}**) when classified with the Greengenes database (**Figure [4A](#F4){ref-type=\"fig\"}**). *Enterococcus* was also underrepresented in DS and MC samples amplified with V3--V4 primers when identified with the SILVA, RDP, and NCBI databases in comparison to DS and MC samples amplified with V4 primers and ZE (**Figures [4B](#F4){ref-type=\"fig\"}--[D](#F4){ref-type=\"fig\"}**). *Pseudomonas* was underrepresented from DS and MC samples amplified with V4 primers when classified with the Greengenes database (\\<1%; **Figure [4A](#F4){ref-type=\"fig\"}**).\n\nSpearman's rank correlation analysis was used to compare the DS, MC, and ZE sequences classified with the RDP classifier and Greengenes, RDP, SILVA, and the NCBI 16S rRNA database (**Figure [5](#F5){ref-type=\"fig\"}**). Zymo amplified their fastq files with primers targeting the V3--V4 region (which were designed by Zymo) of the 16S rRNA gene; a similar correlation between this amplified community and the theoretical mock community (*r* = 0.64, *P* \\< 0.1) is observed for Greengenes and SILVA databases (**Figure [5](#F5){ref-type=\"fig\"}**). ZE classified with the RDP and NCBI databases gave a reduced correlation to the theoretical mock community; *r* = 0.55 and *r* = 0.62, respectively, no tendency toward the theoretical mock community was observed (**Figure [5](#F5){ref-type=\"fig\"}**). The Zymo mock community gave a similar community structure regardless of database used for sequence classification (**Figure [5](#F5){ref-type=\"fig\"}**). *Salmonella* was underrepresented in all of the ZE. Approximately 7% of the sequences belonging to the Enterococcaceae family were left unclassified at the genus level when classified with the RDP database (**Figure [4](#F4){ref-type=\"fig\"}**).\n\n![Correlation matrix of Spearman's correlation coefficient values between the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequences were clustered into OTUs at 97% identify and taxonomic classification of OTUs was conducted with the RDP classifier within QIIME analyses utilizing the **(A)** Greengenes, **(B)** SILVA, **(C)** RDP, and **(D)** NCBI 16S rRNA databases. Values with a positive significant correlation are illustrated with shaded color while insignificant correlations are blank.](fmicb-09-01365-g005){#F5}\n\nThe DS amplified with V3--V4 primers with 28 PCR1 cycles showed a similar correlation for the Greengenes, SILVA, NCBI, and RDP databases and the community compositions tended toward the theoretical mock community (*P* \\> 0.05; **Figure [5](#F5){ref-type=\"fig\"}**). This correlation extended for the DS amplified with V3--V4 primers with 20 PCR1 cycles when sequences were classified with the RDP database (*P* \\< 0.05; **Figure [5C](#F5){ref-type=\"fig\"}**) but not for the Greengenes, SILVA, and NCBI databases (*P* \\> 0.05; **Figures [5A,B,D](#F5){ref-type=\"fig\"}**). All four databases showed a significant correlation for the MC and DS amplified with the V4 primers at 20 and 28 PCR1 cycles (**Figure [5](#F5){ref-type=\"fig\"}**).\n\nDNA extracted from mock community cells amplified with V4 primers with 20 PCR cycles and classified with the Greengenes database were the only condition where a tendency toward the theoretical community for sequenced MC samples was observed (*P* \\< 0.1). The V4 sequences amplified from DS and MC samples at 20 and 28 PCR1 cycles were significantly correlated to their V3--V4 counterparts (*P* \\< 0.05) when classified with using the Greengenes, SILVA, and NCBI databases (**Figures [5A,B,D](#F5){ref-type=\"fig\"}**). This correlation was not apparent in the MC samples classified with the RDP database (*P* \\< 0.05; **Figure [5C](#F5){ref-type=\"fig\"}**). For Greengenes, SILVA, and NCBI databases, DS samples were correlated with their MC counterpart amplified with identical primers with the same PCR1 cycle conditions (*P* \\< 0.05; **Figures [5A,B,D](#F5){ref-type=\"fig\"}**). This was also observed for MC and DS samples classified with the RDP databases at 28 PCR1 cycles for both primers and for samples amplified with V4 primers at 20 PCR1 cycles (*P* \\< 0.05; **Figure [5C](#F5){ref-type=\"fig\"}**). A significant correlation was not observed between MC and DS both amplified with V3--V4 primers at 20 PCR1 cycles when classified with the RDP database (*P* \\> 0.5; **Figure [5C](#F5){ref-type=\"fig\"}**).\n\ntaxMachine\n----------\n\nTaxonomic classification of unique sequences from DS, MC, and ZE with taxMachine identified all eight bacterial genera (**Figure [6](#F6){ref-type=\"fig\"}**). Similarly to the BLAST classification, the classification was consistent with the theoretical composition. With taxMachine as taxonomic classifier, the Spearman's correlation coefficients (**Figure [7](#F7){ref-type=\"fig\"}**) were similar to that of the BLAST classification (**Figure [2](#F2){ref-type=\"fig\"}**). ZE tended toward the theoretical mock community (*r* = 0.69, *P* \\< 0.1). The DS samples amplified with the V3--V4 and V4 at 20 and 28 PCR1 cycles were all significantly correlated with the theoretical mock community. Samples derived from DNA extracted from mock community cells showed a lower correlation to the theoretical mock community with primer choice and PCR1 cycle number influencing this correlation (**Figure [7](#F7){ref-type=\"fig\"}**). MC amplified with 20 PCR1 cycles showed increased correlation to the theoretical mock community over MC amplified with 28 PCR1 cycles. DNA extracted from mock community cells and DS amplified with V4 primers showed increased correlation to the theoretical mock community over MC and DS amplified with V3--V4 primers. DNA extracted from mock community cells amplified with V4 primers with 20 PCR1 cycles tended toward the theoretical mock community.\n\n![Stack plot comparing the relative abundance of the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC) and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequence taxonomy was classified through a high throughput taxMachine tool and contax.trim database within R.](fmicb-09-01365-g006){#F6}\n\n![Correlation matrix of Spearman correlation coefficient values between the theoretical mock community compositions (Theoretical), Zymo 16S rRNA sequencing data (Zymo), mock community cells extracted using the RBBC method (MC), and mock community DNA (DS) amplified using primers targeting the V3--V4 and V4 region of the 16S rRNA gene with 20 and 28 cycles. Sequences were assigned taxonomic classification using the taxMachine tool and the contax.trim database within R. Values with a positive significant correlation are illustrated with shaded color while insignificant correlations are blank.](fmicb-09-01365-g007){#F7}\n\nAs a comparative tool, unique sequences from ZE were also classified with a version of the RDP classifier within R package rRDP. Spearman's correlation coefficients for unique sequences classified with RDP were significantly correlated for the theoretical mock community (*r* = 0.71, *P* \\< 0.05).\n\nNegative Controls\n-----------------\n\nDNA was extracted from water as a NC and was subjected to 20 and 28 PCR1 cycles with primers targeting the V3--V4 and V4 region of the 16S rRNA gene (**Figure [3](#F3){ref-type=\"fig\"}**). Compared to DS and MC, read numbers were negligible (\\<1000 sequences) for NC amplified with V4 primers at 20 and 28 PCR1 cycles and NC amplified with V3--V4 at 20 PCR1 cycles. One NC sample (amplified V3--V4 at 28 PCR1 cycles) had \u223c14,500 merged sequences after quality filtering. Although this was much higher than the other NC samples, it was less than a 16th of the average number of sequences obtained for DS and MC community samples. NC amplified with 28 PCR1 cycles had a higher number of merged sequence reads when amplified with both V4 and V3--V4 primers.\n\nDiscussion\n==========\n\nRuminants are unique in their ability to convert cellulose in plant cell walls into high-quality meat and milk protein suitable for human consumption. Despite their importance in animal protein production, ruminants are a major contributor to global anthropogenic greenhouse gas emissions by producing methane via enteric fermentation. In consideration of this, the majority of current rumen microbiomic studies aim to elucidate the microbial community present in the rumen in order to increase nutrient utilization and reduce methane emissions. It is crucial that these studies are giving an accurate representation of the rumen environment and that the studies are reproducible and suitable for meta-analysis. Currently, there is a lack of concordance between rumen microbiomic studies with research groups, including our own, failing to include negative extraction controls and positive controls in their analysis ([@B20]; [@B23]; [@B22]; [@B33]; [@B26]). Among the human microbiome community, concerns have been raised in response to the recent uncertainty surrounding the accuracy, consistency, and reliability of 16S rRNA phylogenetic studies ([@B35], [@B36]). In our present study, we subjected some of the methods that are currently used for investigation of the rumen microbiota using mock community cells and DSs. Power analysis inferred that technical replication of five was sufficient to deduce a significant correlation between the technical treatments examined in the current experiment and the theoretical composition of the mock community supplied by Zymo.\n\nZymo maintain that they provide an unbiased approach for DNA extraction. The fastq files provided by Zymo containing 16S rRNA sequencing data from their extraction of the microbial mock community with the ZymoBIOMCS^TM^ DNA Mini Kit were used initially to confirm the method of taxonomic classification. All eight genera were detected from ZE using BLAST as a method of classification, confirming that the method of classification was suitable for methodological comparative analysis. The results obtained for the ZE showed a compositional trend toward that of the theoretical mock community. 16S rRNA sequences amplified from the DS under the amplification conditions regularly utilized in rumen microbiomic studies showed better correlation to the theoretical mock community than ZE. Results from the sequenced mock community DNA indicated that both the V3--V4 and V4 primer set in this study accurately identified the genera present in the mock microbial community. All eight genera were detected and in relatively accurate proportions. The DS amplified with primers targeting the V4 region of the 16S rRNA gene showed an improved correlation to the mock microbial community than the DS samples amplified with primers targeting the V3--V4 region of the 16S rRNA gene. This may have been due to the increased length of the V3--V4 primer sequence. The base quality of the sequencing deteriorates at the end of the read ([@B31]). The increased length of the V3--V4 sequence allows only around 20 bp of overlap between sequence reads when a 250-cycle MiSeq reagent kit is used (Illumina had serious QC issues with the 300-cycle kits at the time of writing) and requires utilization of lower quality read ends. This may have contributed to the marginally lower correlation coefficient values discerned for the V3--V4 sequences relative to the V4 sequences. The correlation of DS to the mock community did not differ greatly between 20 and 28 PCR1 cycles. However, non-specific background amplification was increased for samples generated with 28 PCR1 cycles in comparison to amplicons generated with 20 PCR1 cycles indicating that 20 PCR1 cycles was favorable for 16S rRNA gene amplicon generation. The results provide confirmation that both primer types are suitable for amplification of bacteria for phylogenetic community analysis and that 20 PCR1 cycles was the preferable option for the amplicon amplification when utilizing these primer sets.\n\nHaving confirmed that the amplification process was satisfactory, we examined the DNA extraction methodology commonly utilized for extraction of microbial DNA from rumen microbiota. All eight genera were produced from the sequences output from the MC, inferring that the extraction protocol is suitable for both Gram-positive and Gram-negative bacteria. The MC did not reflect the theoretical mock community as precisely as the DS; however, the correlation coefficient obtained for the MC samples was similar to that obtained when comparing the ZE to the theoretical mock community. Zymo deemed this result to be an accurate representation of the theoretical mock community profile. The increase in background amplification in MC in comparison to DS may have been a contributing factor in the reduced correlation with the theoretical mock community. Addition of any steps to the full NGS protocol is invariably going to contribute to non-specific amplification, due to the sensitivity of \"universal\" bacterial primers ([@B4]; [@B40]) and possible residual bacterial DNA present in extraction reagents. Non-specific amplification could have been generated during the PCR step of the protocol due to the KAPA Taq DNA polymerase which is purified from *E. coli*. However, this is not likely as the number of sequence reads in the NCs was low and there was a high correlation to the theoretical mock community in the DS samples. Therefore, the reduced correlation coefficient for MC in comparison to DS is likely to be due to minor extraction inefficiencies. 16S rRNA gene amplification of DNA extraction from microbial cells would inevitably be less efficient than amplification from DNA directly; however, the DS and MC did show strong correlations to each under the same conditions highlighting the minor variation between the two amplified communities. It was also noted that MC followed the same pattern as DS, with samples amplified with V4 primers showing a higher correlation to the theoretical composition of the mock microbial community than samples amplified with primers targeting the V3--V4 region of the 16S rRNA gene, further validating the results from both the sampled MC and DS. The results provided confirmation that both primer types are suitable for amplification of bacteria for phylogenetic community analysis and that 20 PCR1 cycles was the preferable option for the amplicon amplification when utilizing these primer sets in this two PCR step protocol. The results indicate that MC provided the best simulation of the theoretical mock community when amplified with primers targeting the V4 region of the 16S rRNA gene at 20 PCR1 cycles.\n\nThe major area of concern arose not with the DNA extraction or amplicon generation process but with the taxonomic classification of the 16S rRNA sequences. NCBI BLAST in conjunction with the NCBI 16S database was chosen as a method for this assessment as the Zymo mock community being examined contained well-characterized bacteria. However, BLAST is computationally expensive, slow and not refined for the phylogenetic assessment of rumen microbiota which are largely uncharacterized to species or even family level. A more common method for handling of 16S rRNA gene sequences from rumen phylogenetic studies is the publicly available pipelines such as QIIME ([@B3]). QIIME is a wrapper for applications such as quality filtering, OTU picking, and taxonomic assignment. A general approach to dealing with rumen 16S rRNA gene sequences is to cluster sequences at 97% identity utilizing QIIME's open reference-based OTU picking approach and use the RDP classifier algorithm for taxonomic classification of 16S rRNA amplicons ([@B45]). Representative OTUs are then assigned to bacterial taxonomies using RDP classifier via QIIME. This taxonomic-independent, OTU-based approach was developed to overcome the underrepresentation of 16S rRNA sequences originating in the publicly available 16S rRNA databases. This solution to cluster sequences at 97% identity was proposed by [@B37] when few full-length 16S rRNA sequences were available and sequencing studies were still in their infancy. Due to the generic nature of the mock microbial community, this taxonomic-independent approach was assumed not to be a necessity but was undertaken in order to assess rumen microbiomic approaches to phylogenetic assessment. Our results indicated that the RDP classifier struggled to identify 16S rRNA V3--V4 and V4 OTUs to genus level, using the default confidence threshold of 80%. For all the databases we assessed, the RDP classifier failed to discern between the members of the Enterobacteriaceae family within ZE, DS, and MC samples when sequences were clustered into OTUs at 97% identity. It should be noted that, at a species level, the sequences of *E. coli* and *S. enterica* are two of the most well-characterized bacterial species with numerous genomic assemblies for each on the NCBI database ([@B7]); reducing these sequences to representative OTU sequences may not necessarily emulate the diversity of the original sequences. It is therefore possible the confidence threshold was too high for accurate identification at a genus level.\n\nAlthough not the initial aim of our study, taxonomic classification became a concern. [@B18] highlight that while the RDP classifier tool is a good method for classification, it was not perfect and that they had found that other methods outperformed the RDP classifier ([@B42]). A new ready-to-use tool for taxonomic classification method for 16S rRNA sequences, taxMachine, within the microclass R-package, became available during the course of our study ([@B18]). This was used to classify samples sequenced within the experiment as an alternative tool to the standard QIIME pipeline. The taxMachine tool used an internal optimized database contax.trim database for sequence classification. [@B18] emphasized that there is no comprehensive gold standard databases available for taxonomic classification of 16S rRNA sequences due to our growing knowledge of the microbial community. This variability and lack of comprehensiveness were evident in database coverage and were visible in our own results of the RDP classification with the Greengenes, SILVA, RDP, and NCBI databases, giving varied correlation coefficient values for the comparison of ZE, DS, and MC to the theoretical mock community. The microcontax dataset was designed based on consensus taxonomy assignment from several data repositories and was highlighted to be the closest to a gold standard currently available ([@B18]). For this element of the investigation, we moved away from the convention of clustering sequences at 97% identity ([@B7]). There was improvement in the community profiles from amplicon sequences derived from ZE, DS, and MC when identical sequences were aligned and classified with the taxMachine tool. Interestingly, the same result was observed when the OTU clustering step was removed and identical ZE sequences were classified with the RDP classifier within the rRDP package in R ([@B10]). This indicates that the standard conventional pipeline we have been utilizing for 16S rRNA may not be providing the most accurate results. Clustering sequencing into OTUs of 97% identity provided a facile tool for management of the vast amount of sequence data obtained from rumen amplicon studies but it may be time to move away from this concept; however, this will need to be explored further with a mock community which more closely resembles the diversity and genera present in the rumen.\n\nDespite the limitations in taxonomic classification highlighted in the current study, it is acknowledged that global efforts are ongoing to improve our interpretation of rumen microbial communities ([@B32]; [@B38]).\n\nConclusion/Future Work\n======================\n\nMethods for phylogenetic analysis are fundamental for translation of research findings into application. This study highlights that we are following the right path for accurate community analysis; however, still have improvements to make. The analysis indicated that utilizing the RBBC method for DNA extraction in combination with primers targeting the 16S rRNA gene using 20 PCR cycles was sufficient for amplicon sequencing to generate a relatively accurate depiction of the bacterial communities present in rumen samples. However, this study was conducted using a low diversity bacterial mock community. In order to obtain a more tangible representation of the efficiency of our DNA extraction and PCR methodology, it is important to conduct a similar study using a mock community which simulates the microbiota present in the rumen. However, this is currently unavailable and common pipelines for bioinformatics analysis will also need to be appraised. Although beyond the scope of this current study, we identified some major flaws in the rumen microbiomic pipelines we and others had been utilizing. This study also highlights the need for positive mock community controls within all rumen microbiomic studies in order to discern errors which may arise at any step during an NGS protocol.\n\nAuthor Contributions\n====================\n\nMM, EM, and SW conceived and designed the experiments. EM and MM performed the experiments and wrote the paper. EM, MM, and GB analyzed the data. EM, MM, SW, and GB contributed reagents, materials, and analysis tools, interpreted the results.\n\nConflict of Interest Statement\n==============================\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding**. The study was funded by a FACCE-JPI \"Rumen Stability\" grant (2014). EM was funded by a Walsh Fellowship award (Ref. No. 2014231).\n\nThe authors acknowledge Paul Cormican for the bioinformatic advice.\n\n<http://www.bioinformatics.babraham.ac.uk/projects/fastqc/>\n\n<https://sourceforge.net/projects/bbmap/>\n\n[^1]: Edited by: Stuart Edward Denman, Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia\n\n[^2]: Reviewed by: Marcell Nikolausz, Helmholtz-Zentrum f\u00fcr Umweltforschung (UFZ), Germany; Alex V. Chaves, The University of Sydney, Australia\n\n[^3]: This article was submitted to Microbial Symbioses, a section of the journal Frontiers in Microbiology\n"}
+{"text": "![](hosplond74126-0004){#sp1 .150}\n"}
+{"text": "Although leukemia is the most common childhood cancer, the only established modifiable risk factor is prenatal or childhood exposure to ionizing radiation ([@b8-ehp-117-1505]; [@b25-ehp-117-1505]). Acute lymphocytic leukemia (ALL) comprises about 80% of all childhood leukemia cases, the remainder being mainly acute myeloid leukemia (AML) ([@b11-ehp-117-1505]). All known risk factors, including ionizing radiation, sex, race, Down syndrome, and other genetic syndromes, account for \\< 10% of all childhood leukemia cases ([@b18-ehp-117-1505]). A narrative review concluded that recent epidemiologic studies are consistent with those reviewed previously ([@b98-ehp-117-1505]) and support associations between childhood leukemia and parental pesticide exposure before and during pregnancy and childhood exposure to household insecticides ([@b45-ehp-117-1505]). In a recent meta-analysis, childhood leukemia was weakly associated with preconceptual and overall paternal smoking ([@b52-ehp-117-1505]). Other potential risk factors include preconceptual paternal occupational exposure to solvents ([@b17-ehp-117-1505]), motor exhaust fumes ([@b95-ehp-117-1505]), or electromagnetic fields ([@b29-ehp-117-1505]; [@b73-ehp-117-1505]); prenatal maternal alcohol consumption (for AML) ([@b85-ehp-117-1505]); and reduced occurrence of common infections during childhood ([@b56-ehp-117-1505]). Prenatal maternal occupational electromagnetic field exposure was linked to childhood leukemia in a Canadian case--control study ([@b40-ehp-117-1505]) but not in two other case--control studies ([@b29-ehp-117-1505]; [@b86-ehp-117-1505]).\n\nPreconceptual paternal occupational or environmental exposures have not been established as causes of any childhood cancer. Although prenatal maternal exposure to ionizing radiation can cause childhood leukemia, there is little evidence that preconceptual paternal ionizing radiation exposure is a risk factor ([@b27-ehp-117-1505]; [@b46-ehp-117-1505]; [@b61-ehp-117-1505]; [@b92-ehp-117-1505]). Paternal smoking has also been linked to increased risks of childhood brain cancer and lymphomas ([@b19-ehp-117-1505]).\n\nOccupational exposures of reproductive-age adults to pesticides may substantially exceed those from other sources. Serum hexachlorobenzene (HCB) levels among men occupationally exposed to airborne HCB in Spain were 6-fold higher than those of unexposed men ([@b81-ehp-117-1505]). Geometric mean peak daily urinary pesticide levels in agricultural applicators were notably higher than those of their spouses \\[2,4-D (2,4-dichlorophenoxyacetic acid), 61 vs. 1 ppb; glyphosate, 3 vs. \\< 0.5 ppb; chlorpyrifos, 19 vs. 5 ppb\\] ([@b59-ehp-117-1505]). Pregnant women employed as farm fieldworkers in California had significantly higher prenatal urinary organophosphate insecticide metabolite levels compared with pregnant women in the general U.S. population; mean total dialkyl phosphate urinary metabolite levels in these two groups were 113 and 70.5 nmol/L, respectively ([@b14-ehp-117-1505]).\n\nAlthough the present study focuses on preconceptual paternal and prenatal maternal exposure, children may have relatively high exposures to certain pesticides. In the 1999--2000 cycle of the National Health and Nutrition Examination Survey, children 6--11 years of age had higher levels of urinary 3,5,6-trichloro-2-pyridinol (TCPy, a chlorpyrifos metabolite) than did adolescents or adults ([@b6-ehp-117-1505]). Farm children in Iowa had higher urinary atrazine levels compared with nonfarm children (0.71 vs. 0.46 \u03bcg/L, *p* \\< 0.001) ([@b20-ehp-117-1505]). Among inner-city children 3--6 years of age in Minneapolis, the highest measured blood levels of heptachlor epoxide, oxychlordane, 1,1-dichloro-2,2-bis(*p*-chlorophenyl)ethylene (*p*,*p*-DDE), and *trans*-nonachlor approached or exceeded the 95th percentile levels of older children and adults in national surveys ([@b83-ehp-117-1505]). In North Carolina and Ohio, preschool children had urinary pentachlorophenol (PCP) levels more than 10-fold those predicted from multimedia PCP levels in homes and daycare centers ([@b97-ehp-117-1505]).\n\nIn this systematic review and meta-analysis we synthesize currently available epidemiologic evidence on the relationships between childhood leukemia and paternal or maternal occupational pesticide exposure. A related report addresses childhood leukemia and parental or childhood residential pesticide exposure ([@b89-ehp-117-1505]).\n\nMaterials and Methods\n=====================\n\nThis systematic review and meta-analysis was conducted according to a protocol designed by two of us (D.T.W. and M.C.T.).\n\nLiterature search\n-----------------\n\nThe literature search and selection processes were conducted simultaneously for studies of childhood leukemia and parental occupational and parental or childhood residential pesticide exposure. The search strategy \\[see Supplemental Material, Appendix 1, available online (doi:10.1289/ehp.0900582.S1 via <http://dx.doi.org/>)\\] was applied to OVID MEDLINE database (1950--2009 March week 3) and OVID MEDLINE database of in process and other nonindexed citations (1950 to March 31 2009) ([@b71-ehp-117-1505]) and then adapted to search the OVID EMBASE (1980--2009 week 13) ([@b71-ehp-117-1505]), TOXNET ([@b70-ehp-117-1505]), OpenSigle (2009), and [@b76-ehp-117-1505]. We used the following MeSH (medical subject heading) terms and key words:\n\n-   Exposure: exp Environmental Exposure/, exp Environmental Pollutants/, exp Pest Control/, exp Pesticides/, pesticid\\$.tw, herbicid\\$.tw, insecticid\\$.tw, fungicid\\$.tw\n\n-   Population: exp Child/, exp Adolescent/, exp Infant/, child\\$.tw, adolescen\\$.tw, infant?.tw, newborn?.tw, youth.tw, teenage\\$.tw\n\n-   Outcome: exp Hematologic Neoplasms/, exp Leukemia/, leuk?emia\\$.tw.\n\nSearch terms were grouped according to the Boolean operators \"OR\" and \"AND.\" We screened all titles and abstracts to determine their suitability and then applied inclusion/exclusion criteria to the complete articles and resolved discrepancies by consensus. We attempted to contact the corresponding author of reports that did not include confidence intervals (CIs) or other essential information. We also searched the reference lists of all included studies.\n\nInclusion and exclusion criteria\n--------------------------------\n\nInclusion criteria were *a*) original epidemiologic research on childhood leukemia, *b*) use of an analytic design (case--control or cohort), and *c*) availability of at least one index of paternal or maternal occupational pesticide exposure. Studies that included a history of occupation in farming or other jobs with likely pesticide exposure and those with self-reported or documented information on occupational pesticide exposure were included. Reports were excluded if only ecologic data were collected and analyzed, or if more recent/relevant reports of the same study were available; case reports and cluster investigations were also excluded. No language criteria restrictions were applied. A flowchart of the selection process is provided in [Figure 1](#f1-ehp-117-1505){ref-type=\"fig\"}.\n\nData abstraction\n----------------\n\nD.T.W. and M.C.T. independently extracted key data from all included studies using a data abstraction form piloted before the present study was undertaken. Data categories comprised referencing, study design, subject selection, exposure assessment, statistical analysis, and results. For each included study, we identified a single exposure index per parent and pesticide category (unspecified, insecticides, herbicides, fungicides). In 15 of 27 included studies, paternal occupational pesticide exposure during the period up to 2 years before conception was well defined; maternal occupational pesticide exposure during pregnancy was well defined in 15 of 16 included studies \\[[Table 1](#t1-ehp-117-1505){ref-type=\"table\"}; also see Supplemental Material, Appendix 4 (doi:10.1289/ehp.0900582.S1)\\]. For studies reporting more than one risk estimate relevant to a given meta-analysis, a single odds ratio (OR) was selected based on *a*) specificity of the exposure index (e.g., a self-reported occupational pesticide exposure was preferred to job title alone and *b*) intensity or duration of exposure (e.g., an index based on frequency or duration of use was used instead of one based on ever vs. never exposed). Although analysis of exposure duration or intensity reduces the numbers of exposed subjects in the highest exposure category (compared with analyses of ever/never exposed), only three studies had such data, and the numbers of highly exposed case parents included 11 mothers and 27 fathers ([@b17-ehp-117-1505]), 5 fathers ([@b37-ehp-117-1505]), and 2 mothers ([@b87-ehp-117-1505]) \\[for key characteristics of these and other included studies, see Supplemental Material, Appendix 2 (doi:10.1289/ehp.0900582.S1)\\].\n\nQuality assessment\n------------------\n\nWe modified the assessment tool of [@b26-ehp-117-1505], a checklist for assessing the methodological quality of health care interventions, by adding three new assessment factors focusing on the quality of exposure assessment (robustness of exposure measurement, variability of exposure intensity or duration, and specificity) and the ability to identify exposure windows (preconception, pregnancy, childhood). Because this tool was developed mainly for randomized clinical trials, we developed ad hoc guidelines to apply the 15 quality rating factors to observational studies \\[see Supplemental Material, Appendix 3 (doi:10.1289/ehp.0900582.S1)\\]. D.T.W. and M.C.T. independently scored the studies without blinding to authorship or publication status of the original studies and resolved any scoring differences by consensus. The maximum total possible quality score was 20; the assigned scores ranged from 4 to 17, with a median of 12. Median scores for total quality and its components were based on all studies.\n\nAnalysis\n--------\n\nWe conducted meta-analyses using the software package Comprehensive Meta Analysis version 2 ([@b10-ehp-117-1505]). Random-effects summary ORs and 95% CIs were estimated to provide an indicator of the overall strength of associations between pesticide exposure indices and childhood leukemia. We assessed heterogeneity across individual studies using Cochran's Q-test. Subgroup analyses assessed summary ORs stratified by total quality score, the four quality-score components (external validity, control of bias, exposure assessment, and control of confounding), exposure time window definition (well- or ill-defined preconceptual paternal or prenatal maternal exposure), timing of report of occupational exposure (exposure reported before vs. after offspring leukemia diagnosis), place of occupational exposure (farm, nonfarm, mixed, or unknown), cell type (ALL, AML, or unspecified acute leukemia), type of pesticide (unspecified pesticides, insecticides, herbicides, or fungicides), date of study report (pre-1990, \u2265 1990), and study design (population-based case--control, hospital-based case--control, or cohort). Critical exposure time windows were defined as pregnancy for mothers and up to 2 years before conception for fathers. Some studies of paternal pesticide exposure only assessed exposure during pregnancy or paternal occupation at birth; we deemed these to be reasonable proxies for preconceptual exposure, assuming that paternal occupations likely did not change from preconception to pregnancy (23 of the 27 paternal occupations were in farming). Analysis of the quality component for bias control was limited to paternal exposure because no studies of maternal exposure had median or higher scores. Publication bias was assessed on the assumption that smaller studies are more likely to be published if they suggest elevated risks. We used Begg and Mazumdar's test based on the rank correlation (as gauged by Kendall's tau statistic) between standardized effect sizes and their variances to assess this potential source of bias ([@b7-ehp-117-1505]). Asymmetry caused by publication bias is expected to produce higher standard errors for smaller studies with larger effects (producing a larger Kendall's tau *Z*-score).\n\nResults\n=======\n\nStudy identification\n--------------------\n\nThe results of the search strategy and study selection process are detailed in [Figure 1](#f1-ehp-117-1505){ref-type=\"fig\"}. From a total of 1,775 studies identified, 111 were retained from the primary screening of abstracts; most excluded studies were irrelevant (*n* = 1,178), duplicates (*n* = 380), or review articles (*n* = 93). After the secondary screening of full reports, a total of 35 studies were deemed eligible.\n\nStudy characteristics\n---------------------\n\nWe included 31 of the 35 eligible studies in the meta-analyses: 26 case--control studies and 5 cohort studies \\[see citations in [Table 1](#t1-ehp-117-1505){ref-type=\"table\"} and study summaries in Supplemental Material, Appendix 2 (doi:10.1289/ehp.0900582.S1)\\]. Among the four excluded studies, three did not present ORs and CIs or sufficient data to enable their calculation (nor were we able to obtain these from the corresponding author) ([@b16-ehp-117-1505]; [@b35-ehp-117-1505]; [@b38-ehp-117-1505]). The other excluded study reported an exceptionally strong association between childhood ALL and pesticide exposure (crude OR = 126.4; 95% CI, 22.2--2,657; calculated from data in the report and assuming one exposed control father rather than none as reported) ([@b100-ehp-117-1505]). This study did not distinguish occupational versus residential pesticide exposure.\n\nFour of the included studies had data for maternal exposure only, and 16 studies had data for paternal exposure only. We conducted meta-analyses of 27 studies with any paternal occupational pesticide exposure with a total of 30 ORs because three studies reported data separately for ALL and AML. For any prenatal maternal occupational pesticide exposure, we analyzed 14 studies with a total of 16 ORs because two studies reported data separately for ALL and AML.\n\nParental occupational pesticide exposure indices reported by individual studies are shown in [Table 1](#t1-ehp-117-1505){ref-type=\"table\"} \\[see also Supplemental Material, Appendix 4 (doi:10.1289/ehp. 0900582.S1)\\]. Paternal occupational pesticide exposure information was collected from fathers or proxies after offspring were diagnosed with leukemia in 18 case--control studies, and before offspring leukemia diagnosis in 9 studies, the sources being paternal occupation on birth records in four case--control studies and census, employer, or pesticide applicator records in five cohort studies. Preconceptual paternal occupational pesticide exposure was well defined or reasonably inferable in 15 studies and ill defined in the remaining 12 studies \\[[Table 1](#t1-ehp-117-1505){ref-type=\"table\"}; see also Supplemental Material, Appendix 4 (doi:10.1289/ehp.0900582.S1)\\]. Paternal occupational exposure to unspecified pesticides was usually based on employment in farming or job titles where pesticide exposure commonly occurs. Exposure to specific or broad classes of pesticides was limited to five studies with relevant data collection and results ([@b22-ehp-117-1505]; [@b37-ehp-117-1505]; [@b44-ehp-117-1505]; [@b67-ehp-117-1505]; [@b96-ehp-117-1505]).\n\nMaternal occupational pesticide exposure during pregnancy was well defined in 15 studies and ill defined in 1 study.\n\nQuality assessment\n------------------\n\nThe quality factor scores for included studies are in the Supplemental Material \\[Appendix 5 (doi: 10.1289/ehp.0900582.S1)\\]. Compared with lower ranking studies, those with median or higher total quality scores tended to have higher scores for factors related to exposure measurement and bias control (Appendix 5). All five cohort studies had median or higher quality scores.\n\nPublication bias\n----------------\n\nWe attempted to identify all relevant original studies, including thesis dissertations, in any language (one report was translated from Japanese) ([@b48-ehp-117-1505]). Inverse funnel plots of the main findings from studies of any paternal and maternal pesticide exposure and childhood leukemia risk revealed no clear evidence of publication bias; Kendall's tau *Z*-scores and one-tailed *p*-values for paternal and maternal exposure, respectively, were 0.45, *p* = 0.33, and 0.18, *p*= 0.43 ([Table 2](#t2-ehp-117-1505){ref-type=\"table\"}).\n\nData synthesis\n--------------\n\n### Paternal occupational pesticide exposure\n\nResults for the 27 studies of any paternal occupational pesticide exposure are shown in [Figure 2](#f2-ehp-117-1505){ref-type=\"fig\"}, arrayed by year of publication; 30 ORs are shown because three studies reported data for ALL and AML separately. Childhood leukemia was not associated with paternal occupational exposure to any pesticides (i.e., exposure to specified or unspecified types of pesticides) (random effects summary OR = 1.09; 95% CI, 0.88--1.34) or unspecified pesticides (random effects summary OR = 1.04; 95% CI, 0.83--1.31) ([Table 2](#t2-ehp-117-1505){ref-type=\"table\"}). For both analyses, there was evidence of heterogeneity. There was a weak inverse association of borderline statistical significance between the year of publication and ORs of individual studies \\[e.g., for any paternal occupational pesticide exposure, meta-regression slope = \u22120.012 (weighted average change in OR per year), *p* = 0.09\\].\n\nThere was an association of borderline statistical significance between childhood leukemia and any paternal occupational pesticide exposure among studies with below-median total quality scores (summary OR = 1.39; 95% CI, 0.99--1.95) but not in those with higher scores (summary OR = 0.93; 95% CI, 0.71--1.21) ([Table 3](#t3-ehp-117-1505){ref-type=\"table\"}). In analyses of the four quality-score components (external validity, control of bias, exposure measurement, and control of confounding), there was an inverse association between childhood leukemia and any paternal occupational pesticide exposure among studies with median or higher bias control scores (summary OR = 0.73; 95% CI, 0.53--0.99) and a positive association in studies with median or higher pesticide exposure measurement scores (summary OR = 1.37; 95% CI, 1.00--1.89). Childhood leukemia was associated with any paternal occupational pesticide exposure among studies with ill-defined preconceptual exposure windows (summary OR = 1.36; 95% CI, 1.00--1.85) but not those with well-defined windows (summary OR = 0.89; 95% CI, 0.67--1.19).\n\nThere was no association between any paternal occupational pesticide exposure and unspecified acute leukemia (summary OR = 0.99; 95% CI, 0.74--1.33) or AML (summary OR = 1.12; 95% CI, 0.60--2.13); the summary OR for ALL was elevated but was not statistically significant (summary OR = 1.30; 95% CI, 0.86--1.94). Childhood leukemia and any paternal occupational pesticide exposure were associated in studies in which exposure information was collected after diagnosis of offspring leukemia (summary OR = 1.34; 95% CI, 1.05--1.70) but not when pesticide exposure information was collected before off-spring leukemia diagnosis (OR = 0.73; 95% CI, 0.54--1.00). Childhood leukemia risk was not elevated in studies of paternal farm-related pesticide exposure (summary OR = 1.04; 95% CI, 0.82--1.32) and was statistically nonsignificantly elevated in studies of nonfarm workplace exposure (summary OR = 1.41; 95% CI, 0.66--3.00) and mixed or unknown workplace exposure (summary OR = 1.30; 95% CI, 0.65--2.60). Childhood leukemia was not associated with any paternal occupational pesticide exposure in pre-1990 or more recent studies \\[summary ORs, 1.23 (95% CI, 0.76--2.00) and 1.06 (95% CI, 0.83--1.35), respectively\\] in population-based or hospital-based case--control studies \\[summary ORs, 1.17 (95% CI, 0.87--1.58) and 1.11 (95% CI, 0.72--1.69), respectively\\], or in cohort studies (summary OR = 0.88; 95% CI, 0.55--1.40) ([Table 3](#t3-ehp-117-1505){ref-type=\"table\"}). There were elevated childhood leukemia risks for paternal occupational exposure to the broad pesticide classes of insecticides (summary OR = 1.43; 95% CI, 1.06--1.92), herbicides (summary OR = 1.25; 95% CI, 0.94--1.66), and fungicides (summary OR = 1.66; 95% CI, 0.87--3.17) ([Table 3](#t3-ehp-117-1505){ref-type=\"table\"}).\n\n### Prenatal maternal occupational pesticide exposure\n\nChildhood leukemia was associated with prenatal maternal occupational exposure to any pesticides (summary OR = 2.09; 95% CI, 1.51--2.88) and unspecified pesticides (summary OR = 2.16; 95% CI, 1.51--3.08), with no evidence of significant heterogeneity (e.g., for any pesticide exposure, *Q* = 19.6, *p* = 0.19) ([Table 2](#t2-ehp-117-1505){ref-type=\"table\"}). There was no association between year of publication and ORs of individual studies (regression slope = \u22120.013, *p* = 0.48). Results for each of the studies are shown in [Figure 3](#f3-ehp-117-1505){ref-type=\"fig\"}, sorted by year of publication. The strength of the association was somewhat weaker among studies with median or higher total quality scores (summary OR = 1.86; 95% CI, 1.11--3.14) and those with high confounding scores (summary OR = 2.38; 95% CI, 1.56--3.62) compared with those with lower scores ([Table 4](#t4-ehp-117-1505){ref-type=\"table\"}). In analyses of quality-score components, the summary ORs in studies with below-median scores for external validity and confounding control were similar to those for studies with higher scores ([Table 4](#t4-ehp-117-1505){ref-type=\"table\"}). The association was stronger in studies with median or higher exposure measurement scores (summary OR = 2.45; 95% CI, 1.68--3.58) compared with those with lower scores (summary OR = 1.44; 95% CI, 0.83--2.51). All studies of prenatal maternal occupational pesticide exposure had below-median bias control scores.\n\nAfter excluding one study with an ill-defined prenatal exposure window, the association between maternal exposure to any pesticides during pregnancy and childhood leukemia was little changed (summary OR = 2.06; 95% CI, 1.47--2.90). The association was somewhat stronger for both ALL (summary OR = 2.64; 95% CI, 1.40--5.00) and AML (summary OR = 2.64; 95% CI, 1.48--4.71), compared with unspecified acute leukemia (summary OR = 1.59; 95% CI, 1.02--2.47). The association between childhood leukemia and any prenatal maternal occupational pesticide exposure was somewhat stronger in studies of farm-related exposures (summary OR = 2.44; 95% CI, 1.53--3.89) compared with studies of mixed or unknown pesticide exposure place (summary OR = 1.81; 95% CI, 1.17--2.81). Summary ORs were similar for pre-1990 compared with more recently reported studies of prenatal maternal occupational pesticide exposure and childhood leukemia ([Table 4](#t4-ehp-117-1505){ref-type=\"table\"}). On removal of the only hospital-based case--control study, the summary OR for any prenatal maternal occupational pesticide exposure was virtually unchanged. All of the studies of prenatal maternal occupational pesticide exposure collected exposure information after offspring leukemia diagnosis. Childhood leukemia was also associated with prenatal maternal occupational exposure to the broad pesticide classes of insecticides (summary OR = 2.72; 95% CI, 1.47--5.04) and herbicides (summary OR = 3.62; 95% CI, 1.28--10.3), but these estimates are based on few studies ([Table 4](#t4-ehp-117-1505){ref-type=\"table\"}).\n\nDiscussion\n==========\n\nAfter a systematic retrieval and screening of the literature on the relationships between parental occupational pesticide exposure and childhood leukemia, we evaluated the overall evidence using a quantitative meta-analytic approach. Childhood leukemia was not associated with any paternal occupational pesticide exposure in our analyses of all relevant studies or in subgroups of studies with median or higher total quality or bias control scores, well-defined or reasonably inferable preconceptual exposure windows, exposure information collected before offspring leukemia diagnosis, farm-related exposure, data for the major leukemia subtypes, population-based case--control or cohort design, or a publication date of 1990 or later. Childhood leukemia was associated with paternal occupational exposure to insecticides and herbicides, but none of the few relevant studies assessed exposure--risk gradients.\n\nChildhood leukemia was associated with prenatal maternal occupational pesticide exposure with no evidence of statistically significant heterogeneity or publication bias. The association was somewhat stronger among studies with higher exposure measurement quality scores and those with farm-related pesticide exposure. Summary ORs were similar for studies of ALL and AML and for pre-1990 or more recent studies. There were moderately strong associations between childhood leukemia and prenatal maternal occupational exposure to insecticides or herbicides based on the few available studies. All of the eligible studies of prenatal maternal occupational pesticide exposure were based on information collected after offspring leukemia diagnosis. There were too few relevant studies for meaningful analyses of maternal occupational exposure to fungicides or for exposure of either parent to individual pesticides.\n\nInterpretation of our meta-analyses is constrained by limitations in the original studies, particularly exposure assessment and potential sources of bias. We attempted to address these issues by conducting a comprehensive literature search (to reduce publication bias) and independent data extraction and study quality assessment by two persons. We also conducted meta-analyses stratified by parent exposed, study quality scores (total and major components), exposure window definition, leukemia subtype, exposure index, farm versus other workplace exposure, study design, publication period, and broad pesticide class. We assessed study quality with a modified Downs and Black tool ([@b26-ehp-117-1505]) \\[see Supplemental Material, Appendix 3 (doi:10.1289/ehp.0900582.S1)\\]. The main limitations incurred during quality assessment were incomplete descriptions of study methods and findings in reports of original research, lack of a direct method to assess recall bias, and the largely unknown etiology of childhood leukemia, reducing our ability to assess the control of potential confounders. Studies with median or greater quality scores generally had better exposure assessment and control of potential sources of bias compared with lower ranking studies. Few eligible studies collected exposure information for specific or toxicologically related pesticides. Only three studies collected and assessed exposure frequency or intensity information, and little is known about the etiology of childhood leukemia, apart from ionizing radiation. Accordingly, our results should be interpreted cautiously.\n\nPotential sources of bias in observational epidemiologic studies are well described elsewhere ([@b79-ehp-117-1505]). Although case--control studies dependent on parental recall of potentially hazardous exposures may be subject to recall bias, nondifferential misclassification of exposure status may be a bigger problem ([@b41-ehp-117-1505]). The latter source of bias tends to reduce the chance of detecting a true association between a potential causal factor and an adverse health outcome. In a large case--control study, pesticide exposure was the only self-reported paternal occupational exposure associated with childhood AML; this association persisted when paternal occupational pesticide exposure was inferred from a job--exposure matrix ([@b17-ehp-117-1505]). Such findings argue against a major bias arising from self-reported occupational pesticide exposure information. In controlled biomoni-toring field studies of farmers, self-reported pesticide exposure information was a fairly good predictor of body burden, if subjects noncompliant for urine collection or reporting incomplete or inconsistent pesticide use information were removed from analysis ([@b82-ehp-117-1505]). However, other controlled field studies of farm children and farmers revealed poor correlations between biomonitoring and self-reported pesticide exposure data ([@b4-ehp-117-1505]; [@b75-ehp-117-1505]).\n\nPrevious reviewers concluded that there were fairly consistent associations between childhood leukemia and parental occupational or residential pesticide exposure ([@b21-ehp-117-1505]; [@b98-ehp-117-1505]). Recent reviewers noted that associations were strongest for parental pesticide exposure before and during pregnancy and for childhood exposure to household insecticides ([@b18-ehp-117-1505]; [@b45-ehp-117-1505]) and that prenatal maternal pesticide exposure may be more important than paternal exposure ([@b15-ehp-117-1505]). A recent meta-analysis of seven case--control studies of adult leukemia and occupational pesticide exposure published during 1990--2005 showed a summary OR of 1.35 (95% CI, 0.91--2.0) ([@b66-ehp-117-1505]). A meta-analysis of 17 studies of adult myeloid leukemia and occupational pesticide exposure published during 1979--2005 revealed a slightly elevated risk (summary OR = 1.21; 95% CI, 0.99--1.48) ([@b93-ehp-117-1505]); their subgroup analyses showed stronger associations in the five studies of pesticide applicators (summary OR = 2.14; 95% CI, 1.39--3.31) and the two studies of manufacturing workers (summary OR = 6.32; 95% CI, 1.90--21.0). Although these studies suggest a role for pesticides in adult leukemia, their relevance to childhood leukemia is not clear because the mechanisms may differ.\n\nChildhood leukemia is associated with genetic polymorphisms in genes encoding enzymes or other proteins involved in DNA repair, membrane transport, cell cycle regulation, and phase I and II metabolism of chemical toxicants ([@b47-ehp-117-1505]). As noted in a recent review ([@b3-ehp-117-1505]), associations between childhood hematopoietic cancers and genetic polymorphisms in genes encoding phase I and II enzymes are consistent with potential chemical causes of these cancers. For instance, a large Quebec case-only analysis reported relatively large interaction ORs between childhood leukemia and *CYP1A1m1* and *CYP1a1m2* variants and prenatal maternal or childhood pesticide exposure ([@b42-ehp-117-1505]).\n\nMost childhood leukemia cases have gross chromosomal abnormalities, including translocations caused by faulty repair of double-strand DNA breaks. Double-stranded DNA breaks may be caused directly by ionizing radiation and certain mutagenic chemicals or indirectly by modulation of type II topoisomerase enzymes. Analysis of routinely collected neonatal blood samples revealed leukemia clones with specific chromosomal translocations in children who later developed ALL, suggesting that many such cases originate *in utero* ([@b33-ehp-117-1505]). About half of all childhood leukemia cases occur by 3 years of age, and most cases probably have a clonal origin, developing from a single abnormal precursor cell over a period of several months ([@b32-ehp-117-1505]; [@b55-ehp-117-1505]; [@b68-ehp-117-1505]; [@b88-ehp-117-1505]). In a small study of infants born in an agricultural region with high pesticide use in the Philippines, the prevalence of the t(8;21) translocation in cord blood samples was 20.5% among those with detectable meconium levels of the methylcarbamate insecticide propoxur, compared with 10% among infants with undetectable levels (crude OR = 2.32; 95% CI, 0.30--57.4; calculated from data given in the report) ([@b50-ehp-117-1505]). It appears that preleukemic clones can persist during childhood and that only a minority progress to leukemia, suggesting that postnatal exposures could influence progression ([@b58-ehp-117-1505]).\n\nThe biological plausibility of potential causal relationships between cancer and pesticide exposure is supported by reviews of available evidence, mainly from animal studies. The U.S. Environmental Protection Agency (EPA) and other national and international bodies have identified about 165 pesticidal active ingredients as known, probable, or possible human carcinogens, some of which have been banned or restricted ([@b36-ehp-117-1505]). The 15 most intensely used pesticides in the United States during 2001, based on the amount of active ingredient sold ([@b90-ehp-117-1505]), include three probable human carcinogens (alachlor, metam sodium, and chlorothalonil) and five possible human carcinogens (acetochlor, malathion, metolachlor, pendimethalin, and trifluralin) ([@b91-ehp-117-1505]). Among 60 pesticides still used in Canada but banned in one or more Organisation for Economic Co-operation and Development member countries because of health and environmental concerns ([@b23-ehp-117-1505]), the insecticides carbaryl and propoxur and the fungicides captan, mancozeb, maneb, and metiram are recognized as probable human carcinogens ([@b91-ehp-117-1505]).\n\nIn experimental animals, exposure of pregnant females to carcinogens can produce cancer in offspring ([@b5-ehp-117-1505]). Lymphomas in mice were induced by transplacental exposure to the fungicides carbendazim or dodecylquanidine acetate together with sodium nitrite ([@b12-ehp-117-1505], [@b13-ehp-117-1505]). Among men, lymphocyte or sperm DNA damage detectable using the comet assay has been associated with background exposure to chlorpyrifos or carbaryl ([@b62-ehp-117-1505]), with occupational exposure to carbofuran ([@b99-ehp-117-1505]) or multiple pesticides ([@b53-ehp-117-1505]), and with occupations in pesticide production ([@b9-ehp-117-1505]) and farming ([@b69-ehp-117-1505]). Male mice preconceptually exposed to ionizing radiation had increased sperm DNA strand breaks, and their offspring demonstrated an increased risk of hematopoietic cancers ([@b39-ehp-117-1505]). These studies suggest potential mechanisms for relationships between childhood hematopoietic cancers and prenatal maternal or preconceptual paternal pesticide exposures.\n\nConclusion\n==========\n\nBased on the present meta-analysis of original epidemiologic studies of childhood leukemia and parental occupational pesticide exposure, we concluded that there was no overall association between childhood leukemia and any paternal occupational pesticide exposure among all studies combined or subgroups of studies with high total-quality scores, well-defined or reasonably inferable preconceptual exposure windows, pesticide exposure information collected before offspring leukemia diagnosis, farm-related exposures, or cohort design. We found elevated childhood leukemia risks in relation to paternal occupational exposure to the broad pesticide classes of insecticides and herbicides; however, there were few relevant studies and they did not address exposure--risk relationships, precluding firm conclusions.\n\nWe also concluded that there was an overall association between childhood leukemia and prenatal maternal occupational pesticide exposure; this association was somewhat stronger among the subgroups of studies with high exposure-measurement-quality scores or farm-related exposures and those that assessed ALL and AML subtypes. We also found associations between childhood leukemia and maternal occupational exposure to insecticides and herbicides; however, because these were based on few available studies, further research in this area is needed.\n\nAlthough the evidence for associations between parental occupational pesticide exposure and childhood leukemia is limited, precautionary public health policies that will minimize such exposures may be warranted. The epidemiologic and biological evidence summarized here suggests that avoidance of prenatal maternal occupational pesticide exposure may be particularly important in this regard.\n\nImportant research needs include *a*) validated self-reported pesticide exposure indices for both parents, including specific pesticide exposure questions; *b*) biomonitoring of pesticide levels in occupationally exposed men and women; *c*) continued follow-up of existing well-designed cohort studies, such as the Agricultural Health Study in the United States; *d*) follow-up studies of the children of parents in such cohorts; *e*) new case--control and cohort studies with sufficient statistical power to assess childhood leukemia subtypes, leukemia before 5 years of age, potential precursors of childhood leukemia, exposure--risk gradients, specific or toxicologically related groups of pesticides, and genetic susceptibility markers (including preservation of DNA samples from parents and children to permit future analyses of genetic markers); and *f*) basic research on potential biomarkers of pesticide exposure and mechanisms of childhood leukemia initiation and progression.\n\nC[orrection]{.smallcaps}\n========================\n\nMany of the values (e.g., ORs, 95% CIs, *p*-values) were slightly different in the manuscript originally published online; they have been corrected here.\n\nSupplemental Material is available online (doi:10.1289/ehp.0900582.S1 via <http://dx.doi.org/>).\n\nThis review was funded in part by a financial contribution from the Public Health Agency of Canada through the National Collaborating Centre for Environmental Health. M.C.T. holds a Canada Graduate Scholarship from the Canadian Institutes of Health Research.\n\nThe views expressed herein do not necessarily represent the views of the Public Health Agency of Canada or the National Collaborating Centre for Environmental Health.\n\n![Literature search results.](ehp-117-1505f1){#f1-ehp-117-1505}\n\n![Random effect ORs for childhood leukemia in relation to paternal occupational exposure to any or unspecified pesticides. See [Table 1](#t1-ehp-117-1505){ref-type=\"table\"} for list of studies. Some studies reported data separately for AML and ALL.](ehp-117-1505f2){#f2-ehp-117-1505}\n\n![Random effect ORs for childhood leukemia in relation to maternal occupational exposure to any or unspecified pesticides. See [Table 1](#t1-ehp-117-1505){ref-type=\"table\"} for list of studies. Some studies reported data separately for AML and ALL.](ehp-117-1505f3){#f3-ehp-117-1505}\n\n###### \n\nPesticide exposure by parent, exposure definition, source, and exposure window rating.\n\n  Reference                         Parent                                          Pesticide exposure definition                                                Exposure source                                    Exposure rating[a](#tfn2-ehp-117-1505){ref-type=\"table-fn\"}\n  --------------------------------- ----------------------------------------------- ---------------------------------------------------------------------------- -------------------------------------------------- -------------------------------------------------------------\n  [@b28-ehp-117-1505] (Fabi74)      P                                               Occupation in farming                                                        Birth records                                      1\n  [@b94-ehp-117-1505] (vanS85)      P                                               Occupational pesticide exposure during pregnancy                             Self[b](#tfn3-ehp-117-1505){ref-type=\"table-fn\"}   1\n                                    M                                               Same                                                                         Self                                               1\n  [@b54-ehp-117-1505] (Low87)       P                                               Occupation in farming 1 year before conception to 1 year before diagnosis    Self                                               2\n  [@b84-ehp-117-1505] (Shu88)       P                                               Occupation in farming during pregnancy                                       Self                                               1\n                                    P                                               Same                                                                         Self                                               1\n  [@b51-ehp-117-1505] (Lava88)      P[c](#tfn4-ehp-117-1505){ref-type=\"table-fn\"}   Occupational pesticide exposure, timing not stated                           Self                                               2\n  [@b17-ehp-117-1505] (Buck89)      P                                               Occupational pesticide exposure 1 year before birth to diagnosis             Self                                               2\n                                    M                                               Same                                                                         Self                                               2\n  [@b22-ehp-117-1505] (Dani89)      P                                               Agricultural pesticide use since 16 years of age                             Self                                               2\n                                    M                                               Prenatal agricultural pesticide use                                          Self                                               1\n  [@b34-ehp-117-1505] (Gard90)      P                                               Occupation in farming                                                        Birth records                                      1\n  [@b57-ehp-117-1505] (Magn90)      P                                               Occupation in farming before child's birth                                   Self                                               2\n  [@b43-ehp-117-1505] (Infa91)      M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b48-ehp-117-1505] (Kish93)      P                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n                                    M                                               Same                                                                         Self                                               1\n  [@b78-ehp-117-1505] (Roma93)      P                                               Occupation in farming                                                        Birth records                                      1\n  [@b87-ehp-117-1505] (Stein94)     M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b49-ehp-117-1505] (Kris96)      P                                               Occupation as farmer and information on pesticide purchases                  Census                                             2\n  [@b63-ehp-117-1505] (Mein96)      P                                               Occupational pesticide exposure during year before conception                Self                                               1\n                                    M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b42-ehp-117-1505] (Inf&Sin99)   P                                               Preconceptual occupational pesticide exposure, duration not given            Self                                               2\n  [@b37-ehp-117-1505] (Heac00)      P                                               Cumulative chlorophenate exposure hours                                      Employee records                                   2\n  [@b64-ehp-117-1505] (Mein00)      P                                               Occupational pesticide exposure during year before conception                Self                                               1\n                                    M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b96-ehp-117-1505] (Wen00)       P                                               Occupational herbicide exposure up to \u2265 15 years before conception           Self                                               2\n  [@b30-ehp-117-1505] (Feyc01)      P                                               Job title with likely pesticide exposure 2--26 months before child's birth   Census                                             1\n  [@b2-ehp-117-1505] (Alex01)       M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b60-ehp-117-1505] (McKi03)      P                                               Agricultural chemical use during 1 year before child's birth                 Self                                               1\n                                    M                                               Same                                                                         Self                                               1\n  [@b77-ehp-117-1505] (Rodv03)      P                                               Pesticide applicator up to 29 years before child's birth                     License                                            2\n  [@b24-ehp-117-1505] (Dell04)      P                                               Occupational pesticide exposure during 2 years before conception             Self                                               1\n  [@b31-ehp-117-1505] (Flow04)      P                                               Farm pesticide applicator during wide preconceptual period                   License                                            2\n  [@b1-ehp-117-1505] (Abad06)       P                                               Occupational pesticide exposure before date of diagnosis                     Self                                               2\n  [@b65-ehp-117-1505] (Mene06)      M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b72-ehp-117-1505] (Pear06)      P                                               Occupation in farming                                                        Birth records                                      1\n  [@b67-ehp-117-1505] (Mong07)      P                                               Occupational pesticide exposure during year before conception                Self                                               1\n                                    M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b80-ehp-117-1505] (Ruda07)      P                                               Occupation in farming during pregnancy                                       Self                                               1\n                                    M                                               Occupational pesticide exposure during pregnancy                             Self                                               1\n  [@b74-ehp-117-1505] (Pere08)      P                                               Occupational pesticide exposure during 2 years before conception             Self                                               1\n\nAbbreviations: M, maternal; P, paternal.\n\nRatings: 1, reported to be exposed during 2 years before conception (for fathers) or pregnancy (for mothers) or such exposure was reasonably inferable; 2, ill-defined exposure time window.\n\nReported by given parent or spouse.\n\nStudy reported paternal or maternal occupational pesticide exposure, assumed here to be mainly paternal.\n\n###### \n\nRandom effects summary ORs for childhood leukemia in relation to parental occupational pesticide exposure.\n\n  Exposure (no. of risk estimates)[a](#tfn5-ehp-117-1505){ref-type=\"table-fn\"}          Summary OR (95% CI)   Heterogeneity *Q*-value   Publication bias (Kendall's tau *Z*-score)   Meta-regression slope[b](#tfn6-ehp-117-1505){ref-type=\"table-fn\"}\n  ------------------------------------------------------------------------------------- --------------------- ------------------------- -------------------------------------------- -------------------------------------------------------------------\n  Paternal occupational exposure                                                                                                                                                     \n  \u2003Any pesticide exposure[c](#tfn7-ehp-117-1505){ref-type=\"table-fn\"} (*n* = 30)        1.09 (0.88--1.34)     81.0, *p* \\< 0.001        0.45, *p* = 0.33                             \u22120.012, *p* = 0.09\n  \u2003Unspecified pesticides only[d](#tfn8-ehp-117-1505){ref-type=\"table-fn\"} (*n* = 26)   1.04 (0.83--1.31)     76.9, *p* \\< 0.001        0.51, *p* = 0.31                             \u22120.011, *p* = 0.13\n  Prenatal maternal occupational exposure                                                                                                                                            \n  \u2003Any pesticide exposure (*n* = 16)                                                    2.09 (1.51--2.88)     19.6, *p* = 0.19          0.18, *p* = 0.43                             \u22120.013, *p* = 0.48\n  \u2003Unspecified pesticides only (*n* = 14)                                               2.16 (1.51--3.08)     19.2, *p* = 0.12          0.05, *p* = 0.48                             \u22120.016, *p* = 0.41\n\nNumber of ORs summarized (one per study unless a study reported data separately for ALL and for AML).\n\nRegression of OR versus calendar year: weighted average change in OR per year.\n\nExposed to specified or unspecified types of pesticides.\n\nExcludes studies that reported only exposure to specific types of pesticides.\n\n###### \n\nRandom effects summary ORs for childhood leukemia in relation to paternal occupational pesticide exposure: subgroup analyses.\n\n  Exposure (no. of risk estimates)[a](#tfn9-ehp-117-1505){ref-type=\"table-fn\"}     Summary OR (95% CI)   Heterogeneity *Q*-value\n  -------------------------------------------------------------------------------- --------------------- -------------------------\n  Total quality score \\< median (14)                                               1.39 (0.99--1.95)     19.3, *p* = 0.11\n  Total quality score \u2265 median (16)                                                0.93 (0.71--1.21)     53.4, *p* \\< 0.001\n  External validity score \\< median (12)                                           1.06 (0.75--1.51)     42.0, *p* \\< 0.001\n  External validity score \u2265 median (18)                                            1.10 (0.84--1.43)     33.7, *p* = 0.009\n  Bias score \\< median (20)                                                        1.33 (1.05--1.69)     35.6, *p* = 0.012\n  Bias score \u2265 median (10)                                                         0.73 (0.53--0.99)     23.8, *p* = 0.005\n  Exposure measurement score \\< median (19)                                        0.92 (0.71--1.19)     46.5, *p* \\< 0.001\n  Exposure measurement score \u2265 median (11)                                         1.36 (1.00--1.89)     20.0, *p* = 0.03\n  Confounding score \\< median (14)                                                 1.17 (0.84--1.63)     24.7, *p* = 0.03\n  Confounding score \u2265 median (16)                                                  1.03 (0.77--1.38)     54.9, *p* \\< 0.001\n  Ill-defined exposure window (12)                                                 1.37 (1.00--1.85)     21.1, *p* = 0.10\n  Well-defined exposure window[b](#tfn10-ehp-117-1505){ref-type=\"table-fn\"} (15)   0.89 (0.67--1.19)     52.1, *p* \\< 0.001\n  Unspecified acute leukemia (18)                                                  0.99 (0.74--1.33)     32.4, *p* = 0.01\n  ALL (8)                                                                          1.30 (0.86--1.94)     36.0, *p* \\< 0.001\n  AML (4)                                                                          1.12 (0.60--2.13)     12.5, *p* = 0.006\n  Exposure reported after diagnosis (19)                                           1.34 (1.05--1.70)     35.6, *p* = 0.008\n  Exposure reported before diagnosis (11)                                          0.73 (0.54--1.00)     24.1, *p* = 0.007\n  Exposure in farming (23)                                                         1.04 (0.82--1.32)     73.6, *p* \\< 0.001\n  Nonfarm exposure (4)                                                             1.41 (0.66--3.00)     1.2, *p* = 0.76\n  Mixed or unknown exposure place (3)                                              1.30 (0.65--2.67)     4.5, *p* = 0.11\n  Pre-1990 (7)                                                                     1.23 (0.76--2.00)     13.6, *p* = 0.04\n  \u2265 1990 (23)                                                                      1.06 (0.83--1.35)     67.3, *p* \\< 0.001\n  Population-based case--control studies (14)                                      1.17 (0.87--1.58)     30.7, *p* = 0.004\n  Hospital-based case--control studies (10)                                        1.11 (0.72--1.69)     49.7, *p* \\< 0.001\n  Cohort studies (6)                                                               0.88 (0.55--1.40)     4.9, *p* = 0.43\n  Insecticides (3)                                                                 1.43 (1.06--1.92)     0.33, *p* = 0.85\n  Herbicides (5)                                                                   1.25 (0.94--1.66)     1.9, *p* = 0.75\n  Fungicides (4)                                                                   1.66 (0.87--3.17)     4.64, *p* = 0.20\n\nNumber of ORs summarized (one per study unless a study reported data for ALL and AML separately).\n\nOr reasonably inferable.\n\n###### \n\nRandom effects summary ORs for childhood leukemia in relation to prenatal maternal occupational exposure: subgroup analyses.\n\n  Exposure (no. of risk estimates)                                                 Summary OR (95% CI)   Heterogeneity *Q*-value\n  -------------------------------------------------------------------------------- --------------------- -------------------------\n  Total quality score \\< median (12)                                               2.25 (1.49--3.42)     13.3, *p* = 0.27\n  Total quality score \u2265 median (4)                                                 1.86 (1.11--3.14)     5.0, *p* = 0.17\n  External validity score \\< median (7)                                            1.99 (1.12--3.51)     10.7, *p* = 0.10\n  External validity score \u2265 median (9)                                             2.18 (1.43--3.31)     9.0, *p* = 0.34\n  Exposure measurement score \\< median (7)                                         1.44 (0.83--2.51)     5.6, *p* = 0.47\n  Exposure measurement score \u2265 median (9)                                          2.45 (1.68--3.58)     11.9, *p* = 0.16\n  Confounding score \\< median (7)                                                  1.71 (0.99--2.96)     8.9, *p* = 0.18\n  Confounding score \u2265 median (9)                                                   2.38 (1.56--3.62)     10.3, *p* = 0.24\n  Well-defined[a](#tfn11-ehp-117-1505){ref-type=\"table-fn\"} exposure window (15)   2.06 (1.47--2.90)     19.2, *p* = 0.16\n  Unspecified acute leukemia (7)                                                   1.59 (1.02--2.47)     11.0, *p* = 0.09\n  ALL (5)                                                                          2.64 (1.40--5.00)     1.8, *p* = 0.77\n  AML (4)                                                                          2.64 (1.48--4.71)     2.8, *p* = 0.42\n  Exposed on farm (9)                                                              2.44 (1.53--3.89)     8.7, *p* = 0.37\n  Mixed or unknown exposure place (7)                                              1.81 (1.17--2.81)     9.2, *p* = 0.16\n  Pre-1990 (5)                                                                     2.12 (1.05--4.25)     4.0, *p* = 0.40\n  \u2265 1990 (11)                                                                      2.10 (1.44--3.08)     15.5, *p* = 0.11\n  Population-based case--control studies (15)                                      2.10 (1.50--2.94)     19.6, *p* = 0.14\n  Insecticides (6)                                                                 2.72 (1.47--5.04)     6.2, *p* = 0.29\n  Herbicides (2)                                                                   3.62 (1.28--10.3)     0.8, *p* = 0.37\n\nOr reasonably inferable.\n\n[^1]: D.K. is chief executive officer and chief risk scientist for Risk Sciences International. The authors declare they have no competing financial interests.\n"}
+{"text": "Introduction {#s1}\n============\n\nUnder normal volemia conditions, increased natremia or plasma osmolality involves compensatory responses such as water intake, sodium appetite inhibition, plasma release of oxytocin (OT) and vasopressin (AVP), and consequently renal sodium excretion and water retention [@pone.0074689-Vivas1]. The central circuit involved in the control of these behavioral and physiological responses induced by hypernatremia has been studied, analyzing the expression of brain immediate early genes [@pone.0074689-Morgan1]--[@pone.0074689-Rinaman1]. Enhanced *c-fos* expression as shown by increased Fos immunoreactivity was identified within the magnocellular groups of OT and AVP hypothalamic paraventricular and supraoptic nuclei (PVN and SON), and matching the observed increase of both peptides in plasma [@pone.0074689-Rinaman1], [@pone.0074689-Verbalis1]. The increased natremia and osmolality is centrally detected by the circumventricular organs of the lamina terminalis (CVOs of LT), which have osmo- and sodium-sensitive cells and have shown increased Fos immunoreactivity (Fos-ir) after different paradigms of body salt loading [@pone.0074689-Rinaman1]--[@pone.0074689-Oldfield1]. At brainstem level, this stimulus involved the activity of the lateral parabrachial nucleus (LPBN), locus coeruleus (LC), ventrolateral medulla (VL), nucleus of the solitary tract (NTS) and area postrema (AP). Together, all these areas work in coordination to mediate the autonomic, endocrine and behavioral responses inherent in osmoregulation [@pone.0074689-Rinaman1], [@pone.0074689-Olson1], [@pone.0074689-Hochstenbach1].\n\nOur previous results involved the serotonergic neurons of the dorsal raphe nucleus (DRN) in the regulation of fluid and electrolyte balance after particular hydroelectrolyte disturbances [@pone.0074689-Franchini1]--[@pone.0074689-Margatho1]. These studies analyzed the double immunoreactive cells for Fos and serotonin along the raphe system, before and after sodium intake induced by sodium depletion and after extracellular volume expansion (EVE). The data indicate that body sodium status modulates the activity of serotonergic DRN neurons, since Fos-5HT-ir decreased after sodium depletion and increased in animals in positive sodium balance or in process of reestablishing sodium balance. Serotonergic cells of the DRN were also activated after body sodium status was reestablished, independently of the concentration of NaCl consumed, suggesting that this system is involved in the inhibition of sodium appetite under conditions of satiety [@pone.0074689-Godino2]. Significant increased activity was also observed in serotonergic DRN cells after EVE [@pone.0074689-Godino1]. Moreover, our connectional studies using a retrograde marker, Fluorogold, in combination with Fos have demonstrated an important functional projection from the DRN to the LPBN and the LT, which are involved in different models of fluid and sodium balance regulation [@pone.0074689-Godino3]--[@pone.0074689-Margatho1].\n\nIn agreement with the above evidence, other authors have demonstrated that DRN lesion and 5HT antagonist (Methysergide) injection into the LPBN increased sodium intake induced by sodium depletion and decreased sodium and potassium renal excretion and endocrine responses (OT and AVP) after EVE. Similarly, the 5HT agonist (5HT2A/2C, 2,5-dimethoxy-4-iodoamphetamine hydrobromide), injected into the LPBN, increased the renal and endocrine response after EVE and inhibited sodium intake induced by different experimental models [@pone.0074689-Menani1]--[@pone.0074689-Margatho3].\n\nTogether these results suggest that body sodium levels modulate the activity of serotonergic DRN cells. However, their specific participation in a model of systemic hypernatremia/hyperosmolarity without blood volume expansion, directly recording \"in vivo\" DRN cell electrical activity during sodium overload, has not yet been studied. For this purpose, anesthetized animals were used, subjected to a single unit extracellular recording of serotonergic DRN cells and receiving a s.c. infusion of either 2 M or 0.15 M NaCl (0.6 ml/100 g b w) solution for one minute. The distribution of neurons along the forebrain and brainstem that express Fos, the double immunolabeling of Fos-5HT, and Fos-OT were also identified under this sodium overload model, and behavioral, renal and endocrine responses were characterized.\n\nOur electrophysiological study demonstrates that the firing rate of putative 5HT DRN neurons increases after body sodium overload, matching the pattern of Fos-5HT double immunostaining. This suggests that the serotonergic system within the DRN participates in the brain circuit that controls this homeostatic response. Oxytocinergic neural activity and OT plasma concentration also increased after sodium overload without EVE, confirming their involvement in this physiological condition. The central circuit activated after body sodium overload involves specific nuclei along the brainstem, lamina terminalis, hypothalamic and central extended amygdala areas.\n\nMaterials and Methods {#s2}\n=====================\n\nThe experiments used adult, Wistar-derived male rats, born and reared in the breeding colony at the Instituto Ferreyra (INIMEC, C\u00f3rdoba, Argentina). Animals weighing 250--300 g were housed singly in metabolic cages. Room lights were on for 12 h/day, and temperature was controlled at 23\u00b0C.\n\nEthical Statement {#s2a}\n-----------------\n\nAll experimental procedures were approved and carried out in accordance with the guidelines of the Ethical Committee of the Instituto Ferreyra for the use and care of laboratory animals. The protocol designed to minimize animal suffering was approved by this Committee (Permit Number: PIP 2009/2011, 11220080102561).\n\nSodium Overload (So) {#s2b}\n--------------------\n\nAfter 4 h of food and water deprivation, each adult rat was weighed, injected subcutaneously (s.c.) on the back with xilocaine (in an effort to reduce non-specific treatment-associated pain and stress), and 5 minutes later with hypertonic saline (2 M NaCl) delivered through a 26-gauge needle (0.6 ml/100 g body weight, for 1 minute) using an infusion pump (B. Braun, Germany), and then returned to its metabolic cage. Control rats were injected similarly but with isotonic saline (0.15 M NaCl). Individual rats from each group were used for urine and plasma measurements, OT radioimmunoassay or brain immunohistochemical detection of Fos, Fos-OT and Fos-5HT expression as described below.\n\nAnother group of animals was used for the electrophysiological study, and were anesthetized and subjected to SO (as previously described) during the extracellular recording.\n\nExperiment 1: Analysis of the Firing Frequency of Serotonergic DRN Neurons during Sodium Overload {#s2c}\n-------------------------------------------------------------------------------------------------\n\nThe rats were anesthetized with urethane 50% (0.3 ml/100 g b w) and then were catheterized in the jugular vein. Body temperature was maintained at 36--37\u00b0C, with a feedback-controlled heating pad. The animals were mounted onto a stereotaxic frame, their skulls exposed, and a hole was drilled above the DRN into which an electrode was lowered (7.8 mm posterior to bregma, 1.0 mm lateral to midline suture, 5.5--7.0 mm below the dura mater) [@pone.0074689-Paxinos1] by means of a hydraulic microdrive. The saggital sinus was ligated and displaced to avoid puncturing this sinus during the descent of the electrode. Single unit DRN discharge was recorded with glass micropipettes (1 \u00b5m tip diameter recording electrode, World Precision Instruments, Inc., item number 1B120F-4) filled with 2 M NaCl solution containing 2% Pontamine Sky Blue. The firing rate was obtained from cells which displayed a signal: noise ratio 2\u22361 or more. DRN serotonergic neurons were tentatively identified on the basis of characteristic firing patterns and pharmacological characterization for an \"*in vivo\"* or \"*in vitro\"* spontaneously active neurons described previously by Baraban & Aghajanian (1980), Kirby et al., (2003) and others [@pone.0074689-Aghajanian1]--[@pone.0074689-Allers1]. These features includes: a slow (0.5--2.5 Hz), regular firing rate and biphasic action potentials of 1.0--3.0 ms duration, while pharmacological phenotype characterization was performed by 5HT1A stimulation induced by iv injection of fluoxetine (a serotonin reuptake inhibitor which causes a decrease in the firing frequency of serotonergic neurons mainly by 5HT1A autoreceptors stimulation). Thus, we only recorded the spontaneously active neurons which have a slow and regular firing rate that are expected to be inhibited by 5HT1A autoreceptors stimulation [@pone.0074689-Kirby1]. Electrode potentials, which had been previously passed through a high-impedance amplifier, were passed through a window discriminator and screened on an audio amplifier. Integrated histograms generated by the analogy output of the window discriminator were analyzed on-line with a computer.\n\nOnly one neuron was recorded per animal in order to analyze the response to 2 M NaCl or 0.15 M NaCl injection. The baseline activity of each neuron was recorded for 3--5 min before any treatment, and changes in neuronal firing were observed for 15 min after s.c. saline infusions (2 M or 0.15 M NaCl). The mean discharge rate was determined over 1 min intervals. Pharmacological phenotype characterization was then performed by iv injection of fluoxetine. The recording site was marked by injecting Pontamine sky blue dye using a pneumatic pressure pump (Medical Systems Corp. NY MS-2).\n\nFor the qualitative analysis, the following criteria were used to characterize the neuronal response to hypertonic or isotonic infusions:\n\n1.  Excitatory response: when firing frequency increases 0.2 Hz or more after infusion in comparison to baseline activity.\n\n2.  Inhibitory response: when firing frequency decreases 0.2 Hz or more after infusion in relation to baseline.\n\n3.  Neutral response: the absolute changes in firing frequency after infusion do not exceed 0.2 Hz in relation to baseline activity.\n\nFor quantitative analysis, the registered putative 5HT cells were analyzed using a two-way ANOVA mixed with repeated measures (treatment as main factor and time as repeated measures) was used. Post-hoc comparisons were made with the least significant difference test.\n\nExperiment 2: Physiological Characterization of the Response to Sodium Overload {#s2d}\n-------------------------------------------------------------------------------\n\n### Water intake induced by SO {#s2d1}\n\nFood- and water-deprived animals were s.c. saline infused (2 M NaCl or 0.15 M NaCl), returned to metabolic cages without food and provided immediate access to water. Cumulative intake was measured in calibrated burettes at 0, 30, 60, 120, and 180 min. Drinking latencies were recorded immediately after saline infusion. A one-way repeated measures ANOVA was used for the analysis of water intake. Post-hoc comparisons were made with the least significant difference test.\n\n### Determination of plasma OT concentration {#s2d2}\n\nFor the plasma OT concentration assay, different groups of animals were used from those used in the immunohistochemical and electrophysiological studies.\n\nAnimals were decapitated and bled before and 5, 10 and 15 minutes after SO. Trunk blood was collected in chilled plastic tubes containing heparin. Plasma OT level was measured by radioimmunoassay as described by Morris and Alexander (1989) [@pone.0074689-Morris1]. For the assay, OT was extracted from 1 ml of plasma with acetone and petroleum ether. The percentage of recovery after extraction was 85%. The assay sensitivity and intra- and inter-assay coefficients of variation were 0.9 pg/ml, 7% and 12.6%. A two-way ANOVA was used for plasma oxytocin concentration analysis.\n\n### Electrolytes and protein assays {#s2d3}\n\nTo analyze the plasma electrolyte concentration in the 2 M NaCl or 0.15 M NaCl groups, animals were decapitated and bled before and 5, 10 and 15 minutes after SO. Samples were centrifuged and 1 ml of plasma was extracted and stored at \u221220\u00b0C. Urinary samples were taken after the intake test (180 min. after s.c. infusion) and immediately centrifuged, and 1 ml was extracted and stored at \u221220\u00b0C. Electrolyte concentrations of these samples were analyzed by flame photometry (Hitachi 911, automatic analyzer). Plasma volume was inferred from the plasma protein concentration measured according to Lowry [@pone.0074689-Lowry1]. Plasma and urine osmolality were analyzed by vapor pressure osmometer (VAPRO 5520). A two-way ANOVA (treatment and time factors) was used for the analysis of plasma electrolytes, protein concentration and osmolality. Post-hoc comparisons were made with the least significant difference test. A t-test was used for the urinary electrolytes and osmolality after treatments.\n\nExperiment 3: Brain Pattern of Fos, Fos-OT and Fos-5HT Immunoreactivity after SO {#s2e}\n--------------------------------------------------------------------------------\n\nAnimals were perfused 90 min after s.c. 2 M NaCl or 0.15 M NaCl infusion and their brains extracted for immunohistochemical detection of Fos, Fos-OT and Fos-5HT. For this purpose, the different groups of rats were anesthetized with thiopentone (100 mg/kg ip) and perfused transcardially with \u223c100 ml normal saline followed by \u223c400 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.2). The 90 minutes interval after stimulation was chosen in order to be able to compare the distribution of sodium overload-induced c-fos in these animals with that observed and reported previously after induced-sodium consumption [@pone.0074689-Franchini1], [@pone.0074689-Godino2], [@pone.0074689-Franchini2]. The brains were removed, fixed in the same solution overnight, and then stored at 4\u00b0C in PB containing 30% sucrose. Coronal sections were cut into two series of 40 \u00b5m using a freezing microtome and were placed in a mixture of 10% H~2~O~2~ and 10% methanol until oxygen bubbles ceased appearing. They were incubated in 10% normal horse serum (NHS) in PB for 1 h to block non-specific binding sites.\n\nAll the series of the free-floating sections from each brain were first processed for Fos immunoreactivity (Fos-ir), using an avidin-biotin-peroxidase procedure. The sections of the midbrain were then also stained for 5HT immunoreactivity (5HT-ir) and those from the hypothalamus for OT (OT-ir). The staining procedures followed the double-labeling procedures previously described [@pone.0074689-Franchini2], [@pone.0074689-Franchini1]. Briefly, free-floating sections were incubated overnight at room temperature in a rabbit anti-fos antibody (produced in rabbits against a synthetic 14-amino acid sequence, corresponding to residues 4--17 of human Fos) (Ab-5, Calbiochem), diluted 1\u223610,000 in PB containing 2% NHS (Gibco, Auckland, NZ) and 0.3% Triton X-100 (Sigma Chemical Co., St. Louis, MO, USA). The sections were then washed with PB and incubated with biotin-labeled anti-rabbit immunoglobulin and avidin-biotin-peroxidase complex (Vector Laboratories Inc., Burlingame, CA USA, 1\u2236200 dilutions in 1% NHS-PB) for 1 h at room temperature. The peroxidase label was detected using diaminobenzidine hydrochloride (DAB, Sigma Chemical Co., St. Louis, MO, USA), intensified with 1% cobalt chloride and 1% nickel ammonium sulphate. This method produces a blue-black nuclear reaction product. The series of Fos-labeled sections, also processed for immunocytochemical localization of 5HT and OT, were incubated for 72 h at 4\u00b0C with their corresponding antibodies: polyclonal rabbit anti-5HT antibody (ImmunoStar Inc, WI, USA, dilution 1\u223610,000) and polyclonal rabbit anti-OT antibody (Calbiochem, dilution: 1\u223625,000). After incubation, the sections were rinsed and incubated with biotin-labeled anti-rabbit immunoglobulin and avidin-biotin-peroxidase complex for 1 h at room temperature. Cytoplasmic 5HT-ir and OT-ir were detected with unintensified DAB that produces a brown reaction product. Finally, the free-floating sections were mounted on gelatinized slides, air-dried overnight, dehydrated, cleared in xylene and placed under a coverslip with DePeX (Fluka, Buchs, Switzerland).\n\nFos-ir controls were conducted by placing sections in primary Fos antibody that had been preadsorbed with an excess of Fos peptide, or by processing sections without the primary antiserum. No Fos-ir neurons were observed following either of these control procedures.\n\nCytoarchitectural and Quantitative Analysis {#s2f}\n-------------------------------------------\n\nBrain nuclei exhibiting Fos-ir were identified and delimited according to the rat brain atlas of Paxinos and Watson [@pone.0074689-Paxinos1]. The different PVN subnuclei were counted at three different levels of PVN, anterior, medial, posterior (distance from the bregma of the corresponding plates: \u22121.30 mm, \u22121.80 mm, \u22122.12 mm respectively). The distance from the bregma of the corresponding plates is indicated between brackets: accumbens core (AcbC, 1.00 mm), SON (\u22121.3 mm), central amygdaloid nucleus (CeA, \u22122.3 mm), bed nucleus of the stria terminalis, laterodorsal subdivision (BSTLD, \u22120.26 mm), subfornical organ (SFO, \u22120.92 mm), organum vasculosum of the lamina terminalis (OVLT, \u22120.20 mm), median preoptic nucleus (MnPO, \u22120.30 mm), thalamic anterodorsal (AD, \u22121.8 mm), median raphe nucleus (MnR) and DRN (\u22128.00 mm), LPBN (\u22129.3 mm), NTS (\u221213.24 mm) and AP (\u221213.68 mm).\n\nFos-ir nuclei were quantified using a computerized system that includes a Zeiss microscope equipped with a DC 200 Leica digital camera attached to a contrast enhancement device. Images were digitalized and analyzed using Scion Image PC, based on the NIH 1997 version. Fos-ir cells in each section were counted by setting a size range for cellular nuclei (in pixels) and a threshold level for staining intensity. Representative sections in each group were acquired at exactly the same level, with the aid of the Adobe Photoshop Image Analysis Program, version 5.5. The counting was done in four animals of each condition, and was repeated at least twice on each section analyzed, to ensure that the number of profiles obtained was similar. The investigator who conducted the counting of Fos-ir cells was blinded for the experimental groups. Immunohistochemical study was analyzed by Student t-test.\n\nResults {#s3}\n=======\n\nExperiment 1: Effects of a Body Sodium Overload on Putative 5HT-DRN Neuronal Firing {#s3a}\n-----------------------------------------------------------------------------------\n\nThe effect of a 2 M NaCl infusion was tested in 20 putative serotonergic neurons of the DRN. Of the 20 neurons tested with s.c. NaCl 2 M infusion, 14 (70%) displayed an excitatory response, 3 (15%) an inhibitory response and 3 (15%) did not respond. Of the 16 neurons infused with isotonic saline, 3 (19%) displayed an excitatory response, 6 (37.5%) an inhibitory response and 7 (43.5%) did not respond. Taking into account this heterogeneity of responses, the quantitative analysis was done comparing the excitatory putative 5HT neurons of the sodium overload vs control group. The analysis showed a significant interaction between the treatment and time (F~(18,504)\u200a=\u200a~5.82; P\u200a=\u200a0.001). That is, there is a significant differential response in excitatory neuronal firing of the 5HT cells of rats infused with a hypertonic solution of sodium (2 M NaCl) compared to those infused with isotonic saline (0.15 M NaCl) ([Figs. 1](#pone-0074689-g001){ref-type=\"fig\"} and [2](#pone-0074689-g002){ref-type=\"fig\"}). Post hoc analysis indicated that the firing rate of putative serotonergic neurons in rats infused with 2 M NaCl increased significantly 3 and 4 minutes after the infusion started compared to the 0.15 M NaCl injected group and to baseline, respectively. In contrast, the treatment with isotonic saline induced a tendency to decrease the firing rate of serotonergic neurons, being significantly different in the last point of time. Besides, in those neurons that showed an inhibitory response induced by 2 M NaCl, inhibition was observed 4 minutes after the infusion began compared to baseline level.\n\n![Effect of 2-DRN neurons.\\\nValues are means of 1 minutes intervals from 2(n\u200a=\u200a14, black circles) or 0.15 M NaCl (n\u200a=\u200a16, white circles) groups, before and after NaCl infusion indicated by the arrow. Values are means \u00b1 SE. \\*P\\<0.05 vs. 0.15 M NaCl group+P\\<0.05 vs. basal recording. (2-way ANOVA post hoc LSD test).](pone.0074689.g001){#pone-0074689-g001}\n\n![Histograms showing the effect of NaCl sc infusion on electrical activity of putative 5HT-DRN neurons.\\\nFragments recording activity at basal, and after NaCl and fluoxetine infusion are also shown (mV vs ms). A)- 2 M NaCl sc infusion, B)- 0.15 M NaCl sc infusion.](pone.0074689.g002){#pone-0074689-g002}\n\n[Figure 3](#pone-0074689-g003){ref-type=\"fig\"} shows the representative plots at a rostral (3.A) and caudal level (3.B) of DRN from subjects with sodium or control overload, to illustrate the precise location of the 14 recording sites where the dye injection was successfully delivered, specifying the different responses: excitatory (+), inhibitory(\u2212), control () and neutral (\u039f) responses. In the remaining cases we have only an estimation of the recording site location; however all were approximately sited within the dorsal, ventral and ventrolateral subdivisions of the DRN between \u22127.8 mm and \u22128.0 mm from bregma.\n\n![Location of recording sites within the DRN from 2 M NaCl or 0.15 M NaCl infused rats.\\\nThe recording sites are represented within plates 48 and 50 from Paxinos and Watson (1997) corresponding with \u22127.64 mm (A) and \u22128.00 mm (B) distance from bregma, specifying the excitatory (+), inhibitory (\u2212), control () and neutral (\u039f) responses. DRD: dorsal subdivision of dorsal raphe nucleus, DRVL: ventrolateral subdivision of dorsal raphe nucleus, DRV: ventral subdivision of dorsal raphe nucleus, Aq: cerebral aqueduct.](pone.0074689.g003){#pone-0074689-g003}\n\nExperiment 2: Physiological Characterization of the Response to Sodium Overload {#s3b}\n-------------------------------------------------------------------------------\n\n### Water intake induced by SO {#s3b1}\n\nAs previously described, water intake significantly increased in the 2 M NaCl injected group compared to the 0.15 M NaCl group ([Fig. 4](#pone-0074689-g004){ref-type=\"fig\"}). That is, the interaction between treatment factor and time was statistically significant (F~4,64~\u200a=\u200a4.059; p\u200a=\u200a0.005), increasing the volume of water drunk by 2 M NaCl rats.\n\n![Cumulative volume of water drunk during the intake test (3 h) after 2 M NaCl or 0.15 M infusion.\\\nValues are means \u00b1 SE. \\*P\\<0.05 vs. Control group (n\u200a=\u200a9).](pone.0074689.g004){#pone-0074689-g004}\n\nSodium overload rats started to drink water 31 minutes after 2 M NaCl infusion (means of 9 rats). Most of the control rats did not drink water during the test of 180 minutes (6 rats), and those who drank water started to drink 82 minutes after 0.15 M NaCl injection (means of 3 animals).\n\nPlasma Protein Concentration. Osmolality and Plasma Sodium and Urinary Concentration {#s3c}\n------------------------------------------------------------------------------------\n\nIn order to infer whether there are blood volume changes after s.c. saline infusions, plasma protein concentration was examined. No significant differences were observed in plasma protein concentration for treatment factor or time, suggesting that there are no changes in blood volume between groups or baseline levels ([Table 1](#pone-0074689-t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0074689.t001\n\n###### Plasma sodium and protein concentration and plasma osmolality after 2\n\n![](pone.0074689.t001){#pone-0074689-t001-1}\n\n  Time (min)    Plasma Sodium Concentrationmeq/l            Plasma Osmolalitymosmol/KgH~2~O             Plasma Protein Concentrationg/dl                                                                                   \n  ------------ ---------------------------------- ---------------------------------------------------- ---------------------------------- ------------------------------------------------------ ------------ ------------ --\n  0                        146.2\u00b11.0                                   146.2\u00b11.0                                  296.91\u00b10.55                                  296.91\u00b10.55                        6.23\u00b10.28    6.23\u00b10. 28  \n  5                        145.1\u00b11.2                   148.4\u00b11.1[\\*](#nt102){ref-type=\"table-fn\"}                 287.81\u00b10.89                  298.38\u00b10.68[\\*](#nt102){ref-type=\"table-fn\"}       6.62\u00b10. 34   6.14\u00b10.30   \n  10                       144.5\u00b11.1               151.4\u00b11.0[\\*](#nt102){ref-type=\"table-fn\"} **^+^**             298.06\u00b10.78              309.35\u00b10.61[\\*](#nt102){ref-type=\"table-fn\"} **^+^**   6.44\u00b10.32    6.10\u00b10.28   \n  15                       142.7\u00b11.2                   148.2\u00b11.1[\\*](#nt102){ref-type=\"table-fn\"}                 296.78\u00b10.89              303.16\u00b10.68[\\*](#nt102){ref-type=\"table-fn\"} **^+^**   6.18\u00b10.34    6.32\u00b10.30   \n\nValues are means \u00b1 SE; n\u200a=\u200a5.\n\nP\\<0.05 Significantly different between NaCl 2 M and NaCl 0.15 M groups.+P\\<0.05 Significantly different from baseline levels (time 0).\n\n### Plasma osmolality and electrolytes {#s3c1}\n\nAs expected, 2 M NaCl infusions significantly increased plasma sodium concentration and osmolality. The ANOVA for plasma sodium concentration indicated a significant interaction between treatment factor and time (F~3,\\ 64~\u200a=\u200a4.3; p\u200a=\u200a0.008). Plasma sodium concentration significantly increased at 5, 10, 15 minutes compared to the control group and also increased at 10 minutes compared to baseline levels ([Table 1](#pone-0074689-t001){ref-type=\"table\"}).\n\nThe analysis of plasma osmolality also showed a significant interaction between treatment factor and time (F~3,\\ 64~\u200a=\u200a3.95; p\u200a=\u200a0.012). As shown in [Table 1](#pone-0074689-t001){ref-type=\"table\"}, an increase was observed in the sodium overload group at 5, 10, 15 minutes after 2 M NaCl s.c. infusion compared to the 0.15 M NaCl group, and at 10 and 15 minutes compared to baseline levels. The percentage increase of plasma osmolality was 3--4% compared to baseline levels.\n\n### Renal response induced by sodium overload {#s3c2}\n\nSodium overloaded rats had a different renal response compared to the control group during the 180 minutes after s.c. injection ([Table 2](#pone-0074689-t002){ref-type=\"table\"}). Three of the nine rats injected with isotonic saline did not even urinate, while all the rats infused with hypertonic NaCl showed this response. As shown in [Table 2](#pone-0074689-t002){ref-type=\"table\"}, a significant increase in sodium, chloride and potassium renal excretion was observed after 2 M NaCl s.c. infusion compared to the control infused group. However, no difference was found between groups in urine osmolality.\n\n10.1371/journal.pone.0074689.t002\n\n###### Renal response after 2\n\n![](pone.0074689.t002){#pone-0074689-t002-2}\n\n                               Unit                0.15 M NaCl                         2 M NaCl\n  ----------------- -------------------------- ------------------- -------------------------------------------------\n  Urinary volume     ml\\. \u2022 100 g bw^\u22121^ \u00b73 h    0.533 (\u00b10.128)     3.105 (\u00b10.255)[\\*](#nt104){ref-type=\"table-fn\"}\n  Na^+^ Excretion      meq. \u2022 100 g bw^\u22121^       0.047 (\u00b10.015)     0.699 (\u00b10.057)[\\*](#nt104){ref-type=\"table-fn\"}\n  Cl^\u2212^ Excretion      meq. \u2022 100 g bw^\u22121^       0.446 (\u00b10.016)     0.798 (\u00b10.071)[\\*](#nt104){ref-type=\"table-fn\"}\n  K^+^ Excretion       meq. \u2022 100 g bw^\u22121^       0.038 (\u00b10.015)     0.143 (\u00b10.024)[\\*](#nt104){ref-type=\"table-fn\"}\n  Osmolality             mosmol. \u2022 bw^\u22121^       692.579 (\u00b165.895)                  693.826 (\u00b127.407)\n                                                     *n\u200a=\u200a*6                            *n\u200a=\u200a*9\n\nValues are means \u00b1 SE; n\u200a=\u200a5.\n\nP\\<0.05 Significantly different between groups.\n\nPlasma OT Concentration {#s3d}\n-----------------------\n\nAs previously demonstrated [@pone.0074689-Verbalis1], [@pone.0074689-Rinaman1], hypertonic sodium infusion changed plasma OT concentration compared to the control group.The ANOVA indicated that the main treatment factor was statistically significant (F~1,\\ 45~\u200a=\u200a8.305; p\u200a=\u200a0.006), but the interaction between the treatment factor and time did not reach significant levels ([Fig. 5](#pone-0074689-g005){ref-type=\"fig\"}).\n\n![Plasma OT concentration before and 5, 10 and 15 minutes after 2 M NaCl or 0.15 M NaCl infusion.\\\nValues are means \u00b1 SE. The 2 way ANOVA indicated that the main treatment factor was statistically significant (F1, 45\u200a=\u200a8.3; p\u200a=\u200a0.006) \\*P\\<0.05 difference from 0.15 M NaCl group.](pone.0074689.g005){#pone-0074689-g005}\n\nExperiment 3: Brain Pattern of Fos-ir and Double--immunolabeled Cells (Fos-5HT and Fos-OT) after s.c. Infusion of 2 M NaCl vs 0.15 M NaCl Solutions {#s3e}\n---------------------------------------------------------------------------------------------------------------------------------------------------\n\n### DRN {#s3e1}\n\nAs observed in our previous studies, greater activation was found at medial level of the DRN, which includes its dorsal, ventral and ventrolateral subdivisions. As expected based on previous results, sodium overload produced a significant increase in the number of Fos immunoreactive neurons in the DRD, DRV, and DRVL subdivisions of DRN. The number of double-labelled (Fos-5HT) cells is also increased in the DRD and DRVL regions of DRN ([Fig. 6](#pone-0074689-g006){ref-type=\"fig\"}). No significant differences were observed in the Fos-ir and Fos-5HT neurons in the other raphe nuclei such as MnR ([Figs. 6](#pone-0074689-g006){ref-type=\"fig\"} B and C respectively). We also analyzed the number of 5HT neurons in both groups; however we did not observe any significant differences in these analyzed areas ([Fig. 6A](#pone-0074689-g006){ref-type=\"fig\"}).\n\n![Brain Pattern of Fos-ir and Double--immunolabeled cells, Fos-5HT, after SO.\\\nA)- Average number of neurons single-labeled for 5HT, B)- Fos and C)-double-labeled for Fos and 5HT, along the median raphe nucleus and dorsal, ventrolateral and ventral subdivisions of DRN, after s.c. injections of 2 M NaCl or 0.15 M NaCl. Values are means \u00b1 SE. \\*P\\<0.05 significant differences between sodium overload (2 M NaCl) vs control (0.15 M NaCl) groups. D)- Schematic photomicrographs of DRN and MnR analyzed sections (left panel), illustrating the effects of s.c. injections of 2 M NaCl or 0.15 M NaCl on *c-fos* expression in serotonergic neurons of DRN (upper and bottom right panels, respectively). Small square within DRD indicates the region photographed at higher magnification (40X), and shown in right panels. Fos-5HT immunolabeled cells within the DRD are indicated by arrows (upper right panel). Scale Bar: 50 \u00b5m.](pone.0074689.g006){#pone-0074689-g006}\n\n### SON and PVN {#s3e2}\n\nFos expression in the SON was significantly increased in sodium overloaded animals compared to the control group. As previously shown [@pone.0074689-Verbalis1], [@pone.0074689-Rinaman1], the number of double-immunolabeled neurons for Fos and oxytocin was also significantly increased in the 2 M NaCl group ([Fig. 7](#pone-0074689-g007){ref-type=\"fig\"}).\n\n![Brain Pattern of Fos-ir and Double--immunolabeled cells, Fos-OT, after SO.\\\nA)- Average number of Fos immunoreactive neurons and B)- Fos-OT immunolabeled neurons in the supraoptic nucleus (SON), and paraventricular hypothalamic nucleus along the lateral magnocellular and medial subdivisions (PaLM and PaMM, respectively), after s.c. injections of 2 M NaCl or 0.15 M NaCl. Values are means \u00b1 SE. \\*P\\<0.05 significant differences between overload (2 M NaCl) vs. control (0.15 M NaCl) groups. The bottom panels photomicrographs (C,D,E) are showing the pattern of double Fos-OT immunoreactive cells in the SON, C)-, and paraventricular hypothalamic nucleus along the medial (PaMM, D) and lateral magnocellular (PaLM, E) subdivisions in control (upper sections) and sodium overload animals (bottom sections). The left panels illustrate the distribution of these immunoreactive cells at low magnification (10x). Small squares in these panels indicate regions photographed at higher magnification (40x, right panels), and indicated by arrows are the cells photographed at higher magnification (100\u00d7). Scale Bar: 100 \u00b5m.](pone.0074689.g007){#pone-0074689-g007}\n\nWithin the medial (PaMM) and lateral (PaLM) magnocellular subdivisions of the PVN, a significant increase in Fos and Fos-OT immunoreactive cells was found, 90 min after s.c. 2 M NaCl infusion ([Fig. 7](#pone-0074689-g007){ref-type=\"fig\"}).\n\nBrain Pattern of Fos-ir Neurons in Other Nuclei Involved in Sodium Balance Regulation {#s3f}\n-------------------------------------------------------------------------------------\n\n*Brainstem:* The *t-test* analysis indicated that sodium overload produces a significant increase in the number of Fos immunoreactive cells along the AP, NTS and LPBN compared to control animals ([Fig. 8](#pone-0074689-g008){ref-type=\"fig\"}).\n\n![Brain Pattern of Fos-ir in the brainstem nuclei after SO.\\\nA)- Average number of Fos-immunoreactive neurons in the NTS, AP and LPBN after s.c. injections of 2 M NaCl or 0.15 M NaCl. Values are means \u00b1 SE. \\*P\\<0.05 significantly different from 0.15 M NaCl group. B)- Photomicrographs showing the pattern of Fos-immunoreactivity within the LPBN (A--B), AP (C--D) and NTS (E--F), after s.c. injections of 2 M NaCl (B,D,F) or 0.15 M NaCl (A, C, E). Scale Bar: 100 \u00b5m.](pone.0074689.g008){#pone-0074689-g008}\n\n### Lamina terminalis {#s3f1}\n\nThe 2 M NaCl group showed a significant increase in the number of Fos-ir neurons along the circumventricular organs of the lamina terminalis, OVLT and SFO, compared to the control group. A significant increase in the number of Fos-ir cells of experimental group was also observed in the ventral part of the MnPO ([Fig. 9](#pone-0074689-g009){ref-type=\"fig\"}).\n\n![Brain Pattern of Fos-ir in the Lamina Terminalis nuclei after SO.\\\nA)- Average number of Fos-immunoreactive neurons in the SFO, OVLT and MnPO after s.c. injections of 2 M NaCl or 0.15 M NaCl. Values are means \u00b1 SE. \\*P\\<0.05 significantly different from 0.15 M NaCl group. B)- Photomicrographs showing the pattern of Fos-immunoreactivity within the SFO (A--B), OVLT (C--D) and MnPO (E--F), after s.c. injections of 2 M NaCl (B,D,F) or 0.15 M NaCl (A, C, E). Scale Bar: 100 \u00b5m.](pone.0074689.g009){#pone-0074689-g009}\n\n### Central Extended Amygdala (ExA) {#s3f2}\n\nA significantly increased number of activated neurons were observed within the central extended amygdala nuclei, CeA and BSTLD of sodium-overloaded rats, in comparison with control rats ([Fig. 10](#pone-0074689-g010){ref-type=\"fig\"}).\n\n![Brain Pattern of Fos-ir in the Central Extended Amygdala nuclei after SO.\\\nA)- Average number of Fos-ir neurons in the CeA and BSTLD after s.c. injections of 2 M NaCl or 0.15 M NaCl. Values are means \u00b1 SE. \\*P\\<0.05 significantly different from 0.15 M NaCl group. B)- Photomicrographs showing the pattern of Fos-immunoreactivity within the CeA (A--B) and BSTLD (C--D), after s.c. injections of 2 M NaCl (B,D,F) or 0.15 M NaCl (A, C, E). Scale Bar: 100 \u00b5m.](pone.0074689.g010){#pone-0074689-g010}\n\nBrain Pattern of Fos-ir Neurons in other Nuclei not Involved in Sodium Balance Regulation {#s3g}\n-----------------------------------------------------------------------------------------\n\nIn order to demonstrate that the sodium-overload induced effect is selective to brain regions of interest, and not a global elevation in neuronal activity throughout the brain we included the analysis of others nuclei such as AD and AcbC. In both areas we did not observed significantly differences between sodium overload and control groups (AD: mean of 2 M NaCl: 15.17, mean of 0.15 M NaCl: 15.4, p\u200a=\u200a0.97 and AcbC: mean of 2 M NaCl: 20.8, mean of 0.15 M NaCl: 22, p\u200a=\u200a0.91).\n\nDiscussion {#s4}\n==========\n\nThe results of this study allow us to determine in an \"*in vivo*\\\" model that DRN 5HT neurons increase their firing frequency during an increase in systemic sodium concentration and osmolality, matching the pattern of Fos-5HT double immunostaining. Therefore, both electrical activity (analyzed by *in vivo* extracellular recording) and c-fos expression within serotonergic cells of the DRN increased after a hypernatremic/hyperosmolar state.\n\nThe endocrine, renal and behavioral responses were also analyzed in the same model and the brain areas and the oxytocinergic PVN-SON groups of neurons activated after a sodium overload were identified (by single and double immunolabeling).\n\nOur data provide new evidence regarding the activation of specific groups of serotonergic cells within the DRN during increases of plasma sodium concentration and osmolarity not associated to volume expansion.\n\nThe electrophysiological evidence supports previous immunohistochemical and pharmacological results that involved serotonergic pathways modulating sodium intake and renal excretion after different hydroelectrolytic disturbances. Our previous reports demonstrated serotonergic system involvement in the inhibitory control of sodium appetite induced by peritoneal dialysis (PD). Fos-ir decreased in serotonergic cells 24 hs after PD (during the appetitive phase) compared to non-depleted animals, while Fos-ir increased during the satiety phase when animals were in process of reestablishing body sodium status by ingesting sodium salts [@pone.0074689-Franchini1], [@pone.0074689-Godino2], [@pone.0074689-Godino3]. According to these, it has been shown that fos expression depends on the temporal features of action potential patterns. For example in cultured dorsal root ganglion cells, immediate-early gene activation was inversely correlated with the burst --intervals of action potentials [@pone.0074689-Kovcs1]. Our previous results also demonstrate that the 5HT system is involved in the regulation of renal responses. Serotonergic receptor blockade with LPBN injections of methysergide reduced the increase in urinary volume and sodium and potassium excretion induced by EVE, while injections of the serotonergic 5HT2A and 5HT2C receptor agonist, 2.5-dimethoxy-4-iodoamphetamine hydrobromide, enhanced the effects of BVE on Na\\_ and K\\_ excretion and urinary volume [@pone.0074689-Margatho3]. Decreased levels of 5HT and its metabolite 5-hydroxyindoleacetic acid were also observed within the raphe nucleus 15 min after EVE [@pone.0074689-Margatho1]. Finally, the activation and the increase in serotonergic DRN firing frequency after a body sodium overload may be interpreted as these cells participate in the behavioral osmoregulatory response [@pone.0074689-Franchini1]--[@pone.0074689-Godino2]. It would also influence renal and endocrine responses, increasing plasma OT and atrial natriuretic peptide and consequently increasing urine output and sodium and potassium renal excretion, as often reported after an EVE or after induced sodium intake [@pone.0074689-Godino1], [@pone.0074689-Godino2], [@pone.0074689-Margatho3].\n\nThese data together suggest that the serotonergic system at DRN level is modulated by body sodium status and is therefore participating in its regulation. In sum, the increased activity and firing frequency of putative serotonergic cells of the DRN after body sodium overload may reflect how this system mediates the behavioral, renal and endocrine responses for reestablishing body sodium balance.\n\nNeuroanatomical evidence indicates that the DRN sends 5HT projections to the LPBN [@pone.0074689-Margatho1], [@pone.0074689-Petrov1], forming a key pathway to regulate homeostatic responses under hydroelectrolyte balance alterations. Previous investigations also demonstrated that serotonergic mechanisms in the LPBN play an inhibitory role in controlling sodium appetite following a variety of dipsogenic and/or natriorexigenic stimuli [@pone.0074689-Menani3], [@pone.0074689-Tanaka1]--[@pone.0074689-Menani4]. Both the LPBN and the DRN receive afferents from the LT nuclei and their cells are activated not only after SO but also during different body sodium balance changes [@pone.0074689-BadauPassos1], [@pone.0074689-Godino3]. That is, hypertonicity would be detected by the LT, which sends projections directly to the DRN and LPBN, increasing the activity of DRN 5HT cells, which also send efferents to the LPBN, modulating its activity and consequently sodium appetite.\n\nAs expected, SO significantly increased plasma sodium concentration and osmolality. The highest increase was observed at 10 minutes of NaCl injection. However, no differences in plasma protein concentration were observed, suggesting that this protocol of SO does not involve an EVE.\n\nAs previously described [@pone.0074689-Stricker1]--[@pone.0074689-Schoorlemmer1], an increase of 1--2% in plasma osmolality produces the activation of a central circuit that induces thirst in order to reestablish normal osmolality. Similarly, we observed that the SO group consumed three times more water than the control group during the 3 h drinking test. Renal response is also implicated in plasma tonicity regulation. In agreement with previous reports, we found that SO enhanced volume and electrolyte excretion [@pone.0074689-Schoorlemmer1], [@pone.0074689-CheemaDhadli1]. High sodium excretion is in part a consequence of the natriuretic effect of plasma oxytocin, which was also found to be elevated. These data confirm previous studies which showed that OT is released after hyperosmotic stimulus [@pone.0074689-Stricker2]--[@pone.0074689-Bisset1].\n\nThe observed variations in osmolarity and plasma sodium concentration produced by subcutaneous sodium overload are detected by highly specialized neurons, able to translate these changes to electrical signals which activate CNS areas involved in the control of water and salt intake and excretion [@pone.0074689-Noda1]--[@pone.0074689-Richard1]. There is a general agreement that osmosensory transduction is primarily mediated by cells of the two circumventricular organs located in the lamina terminalis: the SFO and the OVLT [@pone.0074689-Noda1], [@pone.0074689-Watanabe1], [@pone.0074689-Johnson1]--[@pone.0074689-McKinley3]. These areas send information to the DRN [@pone.0074689-BadauPassos1] and hypothalamic nuclei such as PVN, SON and MnPO [@pone.0074689-McKinley3]--[@pone.0074689-Kawano1]. Then they may activate the magnocellular neurons of the PVN and SON, increasing the plasma concentration of OT. As previously described in the results section, sodium overload also increased Fos-ir and Fos-OT positive cells in the SON after 2 M NaCl infusion; however, isotonic infusion did not produce any change in oxytocinergic cell activity.\n\nDifferent areas of the brainstem such as the NTS, AP, LPBN, which mediate peripheral satiety and osmoregulatory signals to modulate fluid intake and neurohypophyseal hormone secretion, showed increased Fos immunoreactivity after SO, and structures of the extended amygdala complex, such as the CeA and bed nucleus of the stria terminalis (BST), which are involved in the processing of integrated signals related to sodium appetite behavior, are activated by SO [@pone.0074689-Rinaman1]--[@pone.0074689-Hochstenbach1], [@pone.0074689-Kovcs2], [@pone.0074689-Larsen1]. These data give new evidence for BST involvement during systemic hypertonicity states associated with normo-volemia. Previous studies indicated that both the CeA and BST sub-nuclei are also activated by isotonic and hypertonic EVE and are involved in thirst and sodium appetite control [@pone.0074689-Godino1], [@pone.0074689-Johnson2]--[@pone.0074689-Li1].\n\nIn conclusion, taking into account these results and other studies, we can speculate on the possible brain circuit involved in regulatory responses during a SO without volume expansion. Lamina terminalis structures detect humoral changes produced by SO such as plasma hormones, Na concentration and osmolality changes. The circumventricular organs of the LT send projections to the hypothalamus (SON and PVN), brainstem nuclei (LPBN, NTS, AP and DRN) and central extended amygdala complex. LT projections, among others, activate the oxytocinergic neurons of SON and PVN, increasing plasma OT concentration in order to promote sodium excretion. SO produces an increase in Fos-ir and electrical activity of 5HT DRN neurons. These cells may control body sodium status, inhibiting sodium appetite and increasing natriuresis.\n\nThe authors thank Graciela Jahn and Marta Soaje for their help on measuring oxytocin plasma concentration.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: AG HFC LV. Performed the experiments: AG SP. Analyzed the data: AG SP HFC. Contributed reagents/materials/analysis tools: AG HFC LV. Wrote the paper: AG SP HFC LV. Technical assistance of experiments electrophysiology: HC.\n"}
+{"text": "Background {#Sec1}\n==========\n\nA large and growing body of evidence indicates that postprandial hyperglycemia is a contributing factor for the development of several health related disorders \\[[@CR1], [@CR2]\\]. Firstly, postprandial hyperglycemia has been shown to begin prior to type 2 diabetes occurring \\[[@CR3]--[@CR5]\\] and its control is essential for achieving recommended glycosylated hemoglobin levels (representing mean glycemia over a prolonged period of time) since control of fasting hyperglycemia only is necessary but usually insufficient \\[[@CR4]\\]. The human and economic cost of diabetes could be significantly reduced through prevention, and the control of postprandial glycemia is a good area to target for this purpose. Secondly, several studies suggest that low glycemic index diets, inducing lower postprandial glycemic responses, improve lipid profiles and body weight gain patterns \\[[@CR6]--[@CR9]\\]. These diets decrease total fat mass \\[[@CR7], [@CR8]\\], total cholesterol and LDL cholesterol, thus improving plasma lipid profiles in healthy subjects \\[[@CR6], [@CR7], [@CR9]\\]. Thirdly, a causal relationship has been demonstrated between postprandial hyperglycemia and oxidative stress \\[[@CR10]\\], carotid intima-media thickness and endothelial dysfunction \\[[@CR11]\\], and increased the risk of cardiovascular diseases occurring \\[[@CR12], [@CR13]\\]. Moreover, results from the Baltimore Longitudinal Study on aging showed that impaired glucose tolerance (IGT), by itself, increases the risk of cardiovascular diseases occurring \\[[@CR14]\\]. Accordingly, each of these studies provide relevance for reducing postprandial hyperglycemia in order to prevent several metabolic disorders.\n\nOne interesting approach for limiting postprandial hyperglycemia is to reduce or slow down dietary carbohydrate digestion. Inhibiting the enzymes involved, such as the \u03b1-amylase and \u03b1-glucosidase enzymes, is a powerful therapeutic target for managing the postprandial glycemic response. Also, \u03b1-glucosidase inhibitors are one of the anti-diabetic drug families, of which Acarbose is the most well-known. These medications have a very strong effect and are suitable for treating type 2 diabetes \\[[@CR15]\\] but also induce gastrointestinal side effects that limit their use in a preventive approach \\[[@CR16]\\]. Consequently, several scientists are researching and developing nutritional strategies to finely regulate postprandial glycemia, without inducing adverse events in the gastric tract \\[[@CR17]\\].\n\nCinnamon, one of the oldest spices known world-wide, is obtained from the inner bark of different tree species of the genus *Cinnamomum*. Chinese cinnamon (*Cinnamomum cassia* or *Cinnamomum aromaticum*), coming from China and Southern and Eastern Asia, is more widely used, notably in food, but it contains high level of coumarin, a potentially harmful molecule \\[[@CR18]\\], unlike Ceylon cinnamon (*Cinnamomum zeylanicum* or *Cinnamomum verum*), coming from Sri Lanka and Madagascar, which contains only traces of coumarin \\[[@CR19]\\]. Cinnamon has been extensively studied for the regulation of blood glucose. Several clinical trials have evaluated the chronic effects of cinnamon \\[[@CR20], [@CR21]\\] or cinnamon extracts \\[[@CR22], [@CR23]\\] on subjects suffering from type 2 diabetes. However, the results of several meta-analyses remain controversial, some indicating that cinnamon causes a significant decrease in fasting blood glucose \\[[@CR24], [@CR25]\\] where another reports that cinnamon does not improve fasting blood glucose, HbA1c or lipid parameters in type 2 diabetic subjects \\[[@CR26]\\].\n\nRegarding the acute effect of cinnamon on postprandial glycemia, the available literature is quite puzzling. *In vivo*, a methanol extract of Ceylon cinnamon reduced the glycemic response to maltose and sucrose in normal and diabetic rats but had no effect on glucose loading, thus suggesting a specific effect of this extract on disaccharide digestion through the inhibition of the \u03b1-glucosidase enzyme \\[[@CR27]\\]. Some clinical studies have evaluated the acute effects of ground cinnamon bark on postprandial plasma glucose. Wickenberg *et al.* did not observe any significant effect with 6\u00a0g of Ceylon cinnamon during a standard oral glucose tolerance test (OGTT) in subjects with IGT \\[[@CR28]\\], and Markey *et al.* saw no effect of 3\u00a0g of Ceylon cinnamon in response to a high fat breakfast in healthy volunteers \\[[@CR29]\\]. For Cassia species, Hlebowicz *et al.* did show that the ingestion of 6\u00a0g of Cassia cinnamon reduced the blood glucose response to a test meal \\[[@CR30]\\] while 3\u00a0g did not \\[[@CR31]\\]. Magistrelli *et al.* confirmed the acute effect of 6\u00a0g Cassia cinnamon on postprandial glycemia in both: normal-weight and obese adults \\[[@CR32]\\].\n\nAs illustrated by all of the above, some published clinical studies have assessed the acute effect of raw cinnamon powder from different cinnamon species on postprandial glycemia. However, to our knowledge, the acute effect of cinnamon extracts has not been clinically researched. As a result, we have identified a specific hydro-alcoholic Ceylon cinnamon extract. This publication presents the effects that the extract has on the pancreatic \u03b1-amylase activity and on the glucose response to starch in rodents as well as humans as tested in a randomized, placebo-controlled, cross-over clinical trial in healthy subjects.\n\nMethods {#Sec2}\n=======\n\nCinnamon extracts {#Sec3}\n-----------------\n\nThe Ceylon cinnamon extract (CCE) tested in this study is manufactured by Dialpha (commercially available under the MealShape trademark). It is a hydro-alcoholic extract of Ceylon cinnamon bark (10:1), also called true Cinnamon, *Cinnamomum zeylanicum*, or *Cinnamomum verum*. The dried and milled cinnamon barks were extracted with a 50:50 water-ethanol solution during a 2\u00a0hour period. The extraction mixture was filtrated in order to remove any remaining solids, then concentrated under reduced pressure through evaporation of the ethanol and most of the water, before finally being dried in a vacuum drier to obtain a fine brown powder: CCE.\n\nCCE is standardized at a minimum of 40% of polyphenols, CCE's main active compounds. These polyphenols are mainly constituted of procyanidin oligomers which are composed of catechin and epicatechin monomers. The total polyphenol content was measured by a colorimetric method using a Folin-Ciocalteu reagent as per the ISO14502-1 method.\n\nIn order to compare the potential impact of the extraction solvent on the acute reduction of the starch glycemic response, an aqueous cinnamon extract was also produced from the same Ceylon cinnamon raw material as CCE, using the same extraction process.\n\n*In vitro*inhibition of the pancreatic \u03b1-amylase enzyme activity {#Sec4}\n----------------------------------------------------------------\n\nInhibition of \u03b1-amylase activity was assayed using the enzymatic assay and reagents proposed by Sigma Aldrich. Briefly, the pre-incubation mix was composed of 1.5\u00a0mL of buffer (50\u00a0mmol/L sodium phosphate, 50\u00a0mmol/L sodium chloride, 0.5\u00a0mmol/L calcium chloride, pH\u00a06.9), 0.4\u00a0mL of the solutions to be tested containing the inhibitors at different concentrations, and 0.1\u00a0mL of the enzymatic solution consisting of 100 U/mL of \u03b1-amylase from porcine pancreas (Sigma Aldrich). It was pre-incubated for 30\u00a0min at 25\u00b0C. Then, 0.5\u00a0mL of this pre-mix was incubated with 0.5\u00a0mL of a substrate composed of 1% (w:v) gelatinized potato starch (Sigma Aldrich) in 20\u00a0mmol/L sodium phosphate buffer and 6.7\u00a0mmol/L sodium chloride (pH\u00a06.9) for 20\u00a0min at 25\u00b0C. The reaction was stopped by the addition of 0.5\u00a0mL of a solution composed of sodium potassium tartrate and 96\u00a0mmol/L 3,5-dinitrosalicylic acid, then by boiling for 15\u00a0min, then by cooling on ice. The absorbance was read at 540\u00a0nm. The assays were conducted in triplicates. Acarbose (Sigma Aldrich) was used as a positive control.\n\nAcute starch tolerance test in the rat {#Sec5}\n--------------------------------------\n\nFive-week old male Wistar Han IGS rats weighing between 50 and 75\u00a0g were purchased from the Charles Rivers Laboratories. Two rats were housed per cage. Filtered tap water and regular animal non-purified diet comprised of 51.7% carbohydrates, 21.4% proteins, and 5.1% lipids (diet A03, SAFE) were supplied *ad libitum*. The animal room environment was controlled with a temperature of 22\u2009\u00b1\u20092\u00b0C and day/night cycles of 12\u00a0hours light, 12\u00a0hours dark (19:00--7:00). Animals were allowed to acclimate to the laboratory environment for 10\u00a0days, weighed 3 times per week, and randomly assigned to the different study groups (8 to 20 animals per group depending on the study, as indicated in the figure legends). The protocols were approved by the following ethics committee: \"*Comit\u00e9 R\u00e9gional d'Ethique sur l'Exp\u00e9rimentation Animale*\", Pharmacology Laboratory, Pharmacy University, 15 Avenue Charles Flahault, 34093 Montpellier, France. Starch tolerance tests (STT) were conducted in animals which had fasted overnight through the acute administration by oral gavage of a 7.5% wheat starch solution at 1.5\u00a0g/kg or 20\u00a0mL/kg of body weight containing the products to be tested. The control group was administered starch only. The actual volume administered to each rat was calculated and adjusted based on the most recent body weight of each animal. Blood samples were collected via the tail vein before and 15, 30, 60, 90, and 120\u00a0min after starch administration. One drop of blood was used for the glucose determination using a hand-held glucometer (OneTouch Ultra 2, LifeScan) and, when insulin was measured, 50\u00a0\u03bcL of blood were withdrawn using heparin as an anticoagulant for the plasma preparation. Insulin concentrations were determined using an ELISA assay (Ultrasensitive Mouse Insulin ELISA, Mercodia). At the end of the studies, animals were euthanized by CO2 inhalation.\n\nThree studies were performed, all following this protocol and referred to in the following sections of the Results section of this publication: effect of CCE at 50\u00a0mg/kg of body weight on blood glucose and insulin response, dose-effect of CCE at 6.25, 12.5, 25, 50, and 100\u00a0mg/kg of body weight on blood glucose response and differences of effects between hydro-alcoholic (CCE), and aqueous extracts at 50\u00a0mg/kg of body weight.\n\nHuman clinical trial {#Sec6}\n--------------------\n\nThis clinical study was conducted by Naturalpha, a Contract Research Organization (CRO), within their facilities of the Clinic Nutrition Center, H\u00f4pital Saint Vincent de Paul, Lille, France. It consisted of a monocentric, randomized, double-blind, placebo-controlled, crossover clinical trial which was approved by the National Agency for Medicines and Health Products Safety (ANSM, 143/147 boulevard Anatole France, 93285 Saint-Denis, France) and the Ethics Committee (CPP Nord-Ouest III, CHU, avenue C\u00f4te de Nacre, 14033 Caen, France).\n\nThe primary endpoint was the area under the curve (AUC) 0--120\u00a0min of glycemia compared between CCE and placebo after the absorption of a standard meal. The secondary endpoints were the AUC 0--60\u00a0min of glycemia, the AUC 0--120 and 0--60\u00a0min of insulinemia, maximal glucose and insulin concentration, glycemia and insulinemia values at each point in time. The occurrence of adverse events (AE) was recorded during the entire study. AE have been either spontaneously reported by the subjects, collected through Investigator questioning, clinical examination or laboratory testing, or from Investigator reviews of the subject's diary. The intensity of all AE was recorded on the Case Report Form as 'mild', 'moderate' or 'severe'.\n\nBefore participating in the study, each volunteer signed an informed consent form. Inclusion criteria included age between 18 and 45, BMI between 18.5 (limit included) and 25\u00a0kg/m^2^ (limit excluded), good physical condition, and stable body weight within the last 3\u00a0months prior to screening (\\<5% variation). Exclusion criteria included fasting capillary blood glucose levels\u2009\\>\u2009110\u00a0mg/dL, history of diabetes, smokers, intake of dietary supplements or drugs having an effect on glycemia and insulinemia. The subjects were required to maintain the same lifestyle throughout the study and to eat the same meal at dinner the evening before the two tests. The subjects eligible for the study in terms of inclusion and non-inclusion criteria had an OGTT at the screening visit (V1) in order to exclude subjects with atypical glycemic responses so as to have a homogeneous healthy population with a minimal glycemic response variation. After an overnight fast of at least 10\u00a0hours, subjects were given 50\u00a0g of glucose in 250\u00a0mL of water to drink within a 3\u00a0minute timeframe. Capillary blood samples were collected before and 15, 30, 45, 60, 90, and 120\u00a0min after the glucose load. Subjects with fasting capillary blood glucose levels\u2009\\>\u2009110\u00a0mg/dL and/or 2\u00a0hours postprandial capillary blood glucose levels\u2009\\>\u2009140\u00a0mg/dL were excluded from the study as well as subjects with AUC\u2009\\<\u2009Mean AUC\u2009\u00b1\u20092 SD or AUC\u2009\\>\u2009Mean AUC\u2009\u00b1\u20092 SD.\n\nEvery selected subject came to the clinical center after one night of fasting to complete two study visits (V2 and V3) at least 72\u00a0hours apart. Randomization and product assignment occurred during V2. It was performed using an allocation list prepared by a biostatistics company (Biostatem, France). The subjects were assigned in a random order to receive one of the following combinations: CCE, then placebo or placebo, then CCE. With this cross-over design, all subjects received both: the test product and the placebo, in a random order. The allocation list was used to label the study products accordingly. The study products were packaged and labeled by the Promoter (Dialpha, France), with the packaging and labeling enabling unconditional double blind administration by the individuals conducting the trial (investigator and staff; Naturalpha, France) or the subjects of the study.\n\nThe test product was delivered as two 500\u00a0mg capsules of pure CCE. The matching placebo was provided in two 500\u00a0mg capsules, visually identical to the CCE and containing a mix of 20% microcrystalline cellulose and 80% di-calcium phosphate. The test meal consisted of 103\u00a0g of white bread containing 52.2% of carbohydrates (of which 3.6% of sugars), 7.4% of proteins, 0.1% of lipids, and 3.3% of fibers (the bread thus corresponding to 50\u00a0g of complex carbohydrates, i.e. carbohydrates equivalent to three or more sugars). One drop of capillary blood was used for glucose determination using a hand-held glucometer (Accu-Check Performa, Roche). A catheter was placed in the antecubital vein of the subject's arm to collect venous blood in dried tubes for serum preparation for insulin measurement by ELISA (Unicel DXi, Beckman). Capillary blood samples were collected at -35\u00a0minutes prior to meal absorption to check the fasting status (i.e. blood glucose levels\u2009\u2264\u2009110\u00a0mg/dL). The capsules (test product or placebo) were consumed with 125\u00a0mL of water at -30\u00a0minutes. At 0\u00a0minutes, the test meal was served with 250\u00a0mL of water for consumption within a period of 8\u00a0minutes. Capillary and venous blood samples were collected at -10, -5, 15, 30, 45, 60, 90, and 120\u00a0minutes. The averages of plasma glucose levels at -10 and -5\u00a0minutes were used as the fasting (time \"zero\") glycemia value, whereas insulin level at time -5\u00a0min was used as the fasting (time \"zero\") insulinemia value.\n\nThe study was regularly monitored to ensure execution in accordance with the protocol, Good Clinical Practice (GCP), and other applicable regulations. The person in charge of monitoring reviewed and compared the Case Report Form (CRF) entries with original source data in order to check protocol adherence and to detect any data inconsistency or discrepancy.\n\nStatistical analyses {#Sec7}\n--------------------\n\nFor the rat studies, statistical analyses were conducted using Microsoft Excel software. Student's t-test (two-sample equal variance with a two-tailed distribution) was executed as indicated in the legends of the figures. Values of P\u2009\\<\u20090.05 were considered statistically significant when comparing the control group to the treated group.\n\nFor the clinical trial, the required sample population size was calculated considering a power of 80%, a standard deviation of 25%, an expected difference between placebo and CCE (based on *in vivo* CCE preclinical data) of 25%, and using Student's t table. According to these hypotheses, 18 subjects had to be enrolled in the trial. Statistical analyses were performed on ITT (Intention To Treat) and PP (Per Protocol) populations using SAS\u00ae software version 9.3. Since we were evaluating the ability of an active extract to reduce the glycemic and insulinic responses to white bread, we decided to measure the \"net incremental AUC\". The Net incremental AUC takes into account the negative increments of the curve. The trapezoid rule is applied for all increments, whether positive and negative. All tests were two-sided and significance was declared at a 0.05 threshold. An analysis of variance was performed on the AUC by testing the subject, period, and product effects. Global product effect was tested and, subsequently, product comparisons were performed on CCE compared to Placebo. The comparison tests followed the intra-individual design using a paired Student's t-test or a non-parametric method in case of non-normality: the Koch method \\[[@CR33]\\]. The difference between the two products was tested with the Mann Whitney-Wilcoxon test, taking into account the period effect. First, the equality of residual effects was tested on the sum of the two periods, then in case of no difference, the equality of direct effect between CCE and Placebo was tested using the difference of the two periods. The period effect could be tested using the same test on the crossover difference.\n\nResults {#Sec8}\n=======\n\nInhibition of the pancreatic \u03b1-amylase enzyme activity by CCE {#Sec9}\n-------------------------------------------------------------\n\nCCE was tested for its ability to inhibit the activity of the \u03b1-amylase from the porcine pancreas. Acarbose, a well-known drug that inhibits \u03b1-glucosidase and \u03b1-amylase activity, was used as positive control. Both CCE and Acarbose inhibited mammalian \u03b1-amylase activity, with an IC50 of 25 and 18\u00a0\u03bcg/mL respectively (Figure\u00a0[1](#Fig1){ref-type=\"fig\"}).Figure 1**Alpha-amylase inhibitory effect of CCE.** Values represent mean of triplicate measures.\n\nEffect of CCE on blood glucose and insulin response to starch in the rat {#Sec10}\n------------------------------------------------------------------------\n\nThe administration of 1.5\u00a0g/kg of body weight of starch in rats fasting overnight induced a quick and significant rise in glycemia, i.e. an increase of 73.5\u2009\u00b1\u20092.8\u00a0mg/dL above the pre-STT value (T0) 30\u00a0min after the starch load (Figure\u00a0[2](#Fig2){ref-type=\"fig\"}A). When added to starch, CCE at 50\u00a0mg/kg of body weight reduced the glycemic response by 20.4% compared to starch alone (AUC 0--120, P\u2009\\<\u20090.05). This effect is particularly significant during the peak of glycemia at 30\u00a0min (Student's t-test P\u2009\\<\u20090.001; Figure\u00a0[2](#Fig2){ref-type=\"fig\"}A).\n\nWithout the CCE and during the same lapse of time, following the rise of glycemia-induced insulin secretion by the pancreas, the blood insulin level quickly rose reaching 2.3\u2009\u00b1\u20090.5\u00a0ng/mL above pre-STT value at the peak, 15\u00a0min after starch load (Figure\u00a0[2](#Fig2){ref-type=\"fig\"}B). With CCE, the insulin peak (AUC 0--120\u00a0min) was blunted by 40.6%, reaching 1.0\u2009\u00b1\u20090.4\u00a0ng/mL above pre-STT value at 15\u00a0min. This effect on insulin is statistically significant for the AUC 0--60\u00a0min (P\u2009\\<\u20090.05) but not for the AUC 0--120\u00a0min (P\u2009=\u20090.13).Figure 2**Effect of CCE on blood glucose and insulin response during a STT in rats.** Acute effect of CCE at the dose of 50\u00a0mg/kg of body weight on blood glucose response **(A)** and on insulin response **(B)** during STT in normal rats. Values represent mean\u2009\u00b1\u2009SEM, n\u2009=\u20098. Time points different from control, Student's t-test \\*P\u2009\\<\u20090.05; \\*\\*P\u2009\\<\u20090.01; \\*\\*\\*P\u2009\\<\u20090.001.\n\nDose-effect of CCE on blood glucose response to starch in the rat {#Sec11}\n-----------------------------------------------------------------\n\nThe dose--response of CCE was evaluated on blood glucose levels during STT in normal fasting rats. CCE reduced the glycemic response to starch in a dose-dependent manner (Figure\u00a0[3](#Fig3){ref-type=\"fig\"}A). Compared to the control (starch alone), the AUC between 0 and 120\u00a0min was reduced by 3.2%, 10.1%, 14.5%, 22.5%, and 30.7% with CCE doses of 6.25, 12.5, 25, 50, and 100\u00a0mg/kg of body weight, respectively. This effect is significant from 12.5\u00a0mg/kg of body weight (Student's t-test P\u2009\\<\u20090.05). A regression analysis of the dose--response of CCE were performed. The best fit was obtained using logarithmic equation (R^2^\u2009=\u20090.99; Figure\u00a0[3](#Fig3){ref-type=\"fig\"}B) suggesting that CCE effect on STT may reach a plateau at doses higher than 100\u00a0mg/kg of body weight.Figure 3**Dose-effect of CCE on blood glucose during a STT in rat.** Acute dose-effect of CCE on blood glucose response during STT in normal rats, **(A)** AUC 0--120\u00a0min, **(B)** Logarithmic regression of CCE dose-effect. Values represent mean\u2009\u00b1\u2009SEM, n\u2009=\u20098. AUC different from control, Student's t-test \\*P\u2009\\<\u20090.05; \\*\\*\\*P\u2009\\<\u20090.001.\n\nDifferences of effects between hydro-alcoholic and aqueous extracts {#Sec12}\n-------------------------------------------------------------------\n\nThe magnitude of effects on the reduction of glycemic response to starch in rats with the hydro-alcoholic extract (CCE; 14.9% reduction of AUC 0--60\u00a0min) is double compared with the aqueous extract (7.8% reduction of AUC 0--60\u00a0min) (Figure\u00a0[4](#Fig4){ref-type=\"fig\"}A). The effect of CCE is significantly higher at time point 15\u00a0min (P\u2009\\<\u20090.05), 30\u00a0min (P\u2009\\<\u20090.01), and for the AUC 0--60\u00a0min (P\u2009\\<\u20090.05) compared to the aqueous cinnamon extract.Figure 4**Comparison of the effects of hydro-alcoholic and aqueous cinnamon extracts. (A)** Blood glucose response during STT in normal rats. Values represent mean\u2009\u00b1\u2009SEM, n\u2009=\u200920. Student's t-test: \\*P\u2009\\<\u20090.05; \\*\\*P\u2009\\<\u20090.01; \\*\\*\\*P\u2009\\<\u20090.001 different from control; ^\\#^P\u2009\\<\u20090.05; ^\\#\\#^P\u2009\\<\u20090.01 different from aqueous cinnamon extract. **(B)** *In vitro* \u03b1-amylase inhibition. Values represent mean of triplicate measures.\n\nThe effects obtained on the \u03b1-amylase inhibition *in vitro* (Figure\u00a0[4](#Fig4){ref-type=\"fig\"}B) seem consistent with the effects obtained on the glycemic response in the rat: the hydro-alcoholic extract (CCE; IC50\u2009=\u200930\u00a0\u03bcg/mL) presented a stronger inhibition than the aqueous extract (IC50\u2009=\u200940\u00a0\u03bcg/mL).\n\nEffect of CCE in human clinical trial {#Sec13}\n-------------------------------------\n\nTwenty-two subjects were screened for this study. Three subjects were declared non-eligible (blood glucose levels or BMI not compliant with inclusion criteria) and one subject was not included because the total number of subjects required for the study was reached. In total, 18 randomized subjects were included in the crossover study (Table\u00a0[1](#Tab1){ref-type=\"table\"}). All 18 selected subjects completed the study. Compliance with the treatment reached 100% and no side effects were reported during the study. Two subjects were excluded from Per Protocol statistical analyses as they presented major deviations from the protocol (delay within blood sampling time points and delay in standard meal intake; the CONSORT flow diagram of participants is described in Figure\u00a0[5](#Fig5){ref-type=\"fig\"}). Sequence effect was investigated to compare \"Product then Placebo\" and \"Placebo then Product\" sequences. No statistical difference between sequences was observed.\n\nThe intake of 1\u00a0g of CCE 30\u00a0minutes before the test meal reduced the glycemic response to this test meal compared to placebo (Figure\u00a0[6](#Fig6){ref-type=\"fig\"}A). CCE reduced the AUC 0--120\u00a0min by 14.8%, although a significant difference was not reached (P\u2009=\u20090.15; Figure\u00a0[6](#Fig6){ref-type=\"fig\"}B). However, a significant difference was observed at the peak of glycemia corresponding to the AUC 0--60\u00a0min. In these first 60\u00a0min, the consumption of CCE was associated with a 21.2% reduction of blood glucose compared to placebo (P\u2009\\<\u20090.05; Figure\u00a0[6](#Fig6){ref-type=\"fig\"}C). At the same time, no significant difference was observed between groups regarding all the other measured parameters, notably insulin response (Figure\u00a0[6](#Fig6){ref-type=\"fig\"}D). These results suggest that CCE ingestion prior to a standard meal helps limit the glucose response in the 60\u00a0minutes following the meal absorption without affecting insulin response.Table 1**Baseline characteristics of the healthy volunteers***n*18Age (years)29.9\u2009\u00b1\u20091.8Gender11 male/7 femaleBMI (kg/m^2^)21.4\u2009\u00b1\u20090.4Heart rate (bpm)67.8\u2009\u00b1\u20092.4Systolic blood pressure (mmHg)127.3\u2009\u00b1\u20092.7Diastolic blood pressure (mmHg)73.2\u2009\u00b1\u20092.4Cross-over design. Values represent mean\u2009\u00b1\u2009SEM.Figure 5**Flow diagram of clinical study participants.**Figure 6**Effect of 1\u00a0g CCE on blood glucose and insulin response after a standard meal in humans. (A)** Blood glucose concentration over 120\u00a0min, **(B)** Glycaemia AUC 0--120\u00a0min, **(C)** Glycemia AUC 0--60\u00a0min, and **(D)** Insulin concentrations over 120\u00a0min. Values represent mean\u2009\u00b1\u2009SEM, n\u2009=\u200916.% of inhibition of test meal glycemic response is indicated. CCE different from placebo, \\*P\u2009\\<\u20090.05.\n\nDiscussion {#Sec14}\n==========\n\nThis study investigated for the first time the acute effect of a Ceylon cinnamon water-ethanol extract on blood glucose response to starch.\n\n*In vivo* data provide evidence of the effect of CCE on the reduction of postprandial glycemia following starch administration in normal rats. The effect increased gradually with the dose and is significant from 12.5\u00a0mg/kg of body weight. Interestingly, at the same time, CCE also tended to lower insulin response to starch, showing that the effect of the extract is not due to a stimulation of insulin secretion. Next to that, other experiments that we have conducted in rats showed that if starch is being replaced with glucose, maltose or saccharose in the acute tolerance test, CCE has no effect on the glycemic response (data not shown). These results indicate that CCE acutely inhibits the digestion of starch *in vivo* but has no effect on the digestion of disaccharides (maltose and saccharose), nor on the passage of glucose from the intestine to the blood stream. Also, CCE showed no acute effect on glucose tolerance, i.e., the capacity of the body to manage glycemia after the passage of glucose into the blood.\n\nTaken together, these results support the fact that CCE seems to specifically inhibit \u03b1-amylases. One published study demonstrating the acute effect of a methanol Ceylon cinnamon extract at 300\u00a0mg/kg of body weight on the glycemic response to maltose and sucrose in rats suggested an effect on alpha-glucosidase \\[[@CR27]\\]. We could not reproduce these results with CCE at the dose of 50\u00a0mg/kg of body weight, a dose which is clearly effective on starch response and which constitutes, in our point of view, a more physiological dose regarding further applications in humans.\n\nThe effects of CCE on postprandial blood glucose and insulin response were assessed in healthy male and female volunteers. Since the extract seems to act by inhibiting \u03b1-amylase enzymes, the first enzymes involved in carbohydrate digestion, and since CCE was formulated in capsules, we wanted to make sure that the capsules would dissolve and that the enzymes would be inhibited before the arrival of the test meal in the upper intestine. For this reason, we decided to administrate the capsules 30\u00a0minutes before the meal consumption. In doing so, 1\u00a0g of CCE reduced the post-meal glycemic response by\u2009\\~\u200920% and\u2009\\~\u200915% between 0 and 60\u00a0min and between 0 and 120\u00a0min post meal absorption respectively. This effect is statistically significant (P\u2009\\<\u20090.05) at the peak of glycemia between 0 and 60\u00a0min post meal absorption. In this trial, we addressed young healthy volunteers with normal glucose tolerance (IGT subjects were excluded at screening). Accordingly, in these subjects, glycemia is mainly increased between 0 and 60\u00a0min post meal absorption and returns to normal value between 60 and 120\u00a0min, almost linearly (Figure\u00a0[6](#Fig6){ref-type=\"fig\"}A). In these conditions, it is understandable that the effect of CCE is higher on the peak where the \"therapeutic window\" is larger. By increasing the number of subjects, our observed glycemic response reduction could possibly also have become statistically significant between 0 and 120\u00a0min. Another option could have been to select IGT subjects presenting an exacerbated glycemic response which remains elevated until 2\u00a0hours after the meal absorption.\n\nNo significant difference was observed in terms of insulin response between CCE and placebo, showing that the effect of CCE in reducing postprandial glycemia does not seem to be due to a stimulation of insulin secretion. This is consistent with the findings from the rat study.\n\nRegulatory bodies consider that the improvement of post-prandial glycemia should not be accompanied by an increase in insulinemia which is in accordance with this clinical trial's findings.\n\nIn conclusion, women and men who have taken CCE have reduced their glycemia following the absorption of the test meal, yet with no variation of their insulin demand.\n\nInterestingly, as a side finding, this work allows us to compare the acute effect of CCE, an \u03b1-amylase inhibitor acting in the intestine, on the glycemic response to starch in rats and in humans. To our knowledge, this correlation is poorly described in existing literature. 1\u00a0g of CCE (\\~12.5\u00a0mg/kg of body weight for a man weighing 75\u00a0kg) resulted in the reduction of the glycemic response to starch in humans by 15 to 20%. Given that a dose of 25 to 50\u00a0mg/kg of body weight is needed to observe an equivalent reduction in the rat, it appears that the human equivalent dose of CCE or, by extension, a \"CCE-like\" \u03b1-amylase inhibitor, to the rat may be the dose effective in the rat divided by 2 to 4.\n\nHere we tested the acute effect of a specific hydro-alcoholic Ceylon cinnamon extract on postprandial glycemia. This is a new discovered property of cinnamon extracts which are usually recognized for their chronic effect on glycemia. Other investigators have tested the effect of up to 6\u00a0g of raw cinnamon bark powder on the glycemic response in human clinical trials. No effect was observed with Ceylon cinnamon \\[[@CR28], [@CR29]\\] whereas *cassia* cinnamon presented significant effect in 2 studies \\[[@CR30], [@CR32]\\].\n\nThese interesting results bring us to another important consideration regarding the safety of cinnamons and their derived extracts with respect to their levels of coumarin, a potentially toxic molecule \\[[@CR18]\\]. Ceylon cinnamon contains only traces of coumarin and the corresponding extract, even using ethanol as the extraction solvent, contains far less coumarin amounts than the tolerable daily intake established by the European Food Safety Authority (0.1\u00a0mg/day/kg of body weight; 7\u00a0mg/day for a person of 70\u00a0kg). The CCE coumarin contents is less than 200\u00a0mg/kg (i.e. 1\u00a0g of CCE/day contains less than 0.2\u00a0mg/day of coumarin). The coumarin content in the CCE batch used for the clinical trial was 112\u00a0mg/kg. For the cassia species however, coumarin levels are of concern \\[[@CR19]\\]. Furthermore, since coumarin is more alcohol soluble than water soluble, an ethanol extract of this species would enrich the extract in coumarin, thus worsening coumarin related problem altogether. An aqueous extraction of the *Cinnamomum cassia* species would allow to reduce the level of coumarin contained in the raw material. However, according to the *in vitro* results on \u03b1-amylase inhibition and as evidenced by the *in vivo* results on starch tolerance tests in rats, CCE was significantly more efficient than an aqueous cinnamon extract.\n\nTaken all together, these results show that a hydro-alcoholic Ceylon cinnamon extract appears to be an optimal solution in terms of coumarin content and efficacy as far as cinnamon extracts for use in the management of postprandial glycemia are concerned.\n\nConclusions {#Sec15}\n===========\n\nIn summary, this study demonstrates that a well-defined hydro-alcoholic Ceylon cinnamon bark extract reduces the glycemic response to starch in normal rats and healthy male and female subjects by inhibiting pancreatic \u03b1-amylase starch digestion. This extract may be of great interest with regards to the many recognized benefits associated with the reduction of postprandial hyperglycemia. These benefits namely include reducing the risk of developing type 2 diabetes in people with IGT, helping with the lipid profile management, helping with the control of fat mass, body weight, and oxidative stress, and finally helping reduce the risk for cardiovascular diseases occurring. Other human clinical trials will be required to confirm these results on postprandial glycemia and in order to directly assess the other health benefits.\n\n**Competing interests**\n\nVB, MPI, CM, and NC are employees of the private company Dialpha SAS which researches, develops, and commercializes active plant extracts for the prevention of metabolic disorders. The Ceylon cinnamon hydro-alcoholic extract (CCE) tested in this study is commercialized by Dialpha.\n\nDMP, XD, and CR are employees of Naturalpha SAS, a private company that provides research and development services and which was mandated and financed by Dialpha to conduct the human clinical trial presented in the present publication. DMP, XD, and CR declare that they have no competing interests.\n\n**Authors' contributions**\n\nVB, MPI, CM, and NC were involved in the preclinical studies design, data analysis, and review of the manuscript. VB and CM performed the preclinical experiments. MPI, DMP, XD, CR, and NC were involved in the clinical study design, data analysis, and review of the manuscript. MPI and NC wrote the manuscript. All authors read and approved the final manuscript.\n\nWe would like to thank Dr. Xiaoming Chien for her pertinent advice during the clinical trial design and manuscript building and Dr. Eric Chappuis and Pascal Doll for their critical reviews of this manuscript.\n\nThis research was supported by the French Ministry of Education and Research, the BpiFrance, the R\u00e9gion Languedoc-Roussillon, and the Business & Innovation Center of Montpellier Agglom\u00e9ration.\n"}
+{"text": "Introduction\n============\n\nPreparing the public health workforce to mitigate, respond to, and recover from natural and man made disasters is not a minor undertaking. Both governmental and non governmental organizations have called on universities and other educational institutions to develop programs to efficiently and effectively train our public health workforce \\[[@b1-ojphi-02-20],[@b2-ojphi-02-20],[@b3-ojphi-02-20],[@b4-ojphi-02-20]\\]. Many educational institutions across the country have responded to this by developing and offering virtual or online programs that incorporate 'canned courses'---i.e. courses that do not require an instructor and instead allow students to download the materials and self-learn at their time of convenience \\[[@b5-ojphi-02-20]\\].\n\nThe effectiveness of such online courses depends on delivering rich learning experiences for the students. However, unlike traditional classroom-based education, the online environment is not under the control of an instructor. Students' learning experience in such online situations could be affected by not only the structure and content of the course but also the student interactions facilitated by the technology-based infrastructure and the usability of such infrastructure. Thus, to measure the effectiveness of online courses, we need to go beyond the evaluation tools that are currently used to evaluate offline or classroom-based courses and use tools that provide a more holistic view of users' online learning experience.\n\nSpecifically, to understand and evaluate the learning experience in an online program, we need to draw on our understanding of people's behavior in online environments. Prior studies in consumer psychology and human computer interaction offer an appropriate foundation for this. Research in consumer psychology indicate that experience has two primary dimensions---a utilitarian (or cognitive) dimension and a hedonic (or affective) dimension \\[[@b6-ojphi-02-20],[@b7-ojphi-02-20],[@b8-ojphi-02-20],[@b9-ojphi-02-20],[@b10-ojphi-02-20]\\]. However, in an online environment, factors that are either related to the technology itself or to the interactions of the people with the technology could also shape such experience. Prior studies in the area of human-computer interaction and computer-mediated communication \\[[@b11-ojphi-02-20], [@b12-ojphi-02-20], [@b13-ojphi-02-20]\\] indicate the relevance of two other dimensions---sociability experience and usability experience. In this study, the online offerings of an emergency preparedness program offered by a public university in the Northeast was evaluated on the above four dimensions of online user experience.\n\nIn addition, in this study we also examine whether online user experience had any impact on course completion. Prior research in this area has shown that online distance education courses often have higher non-completion rates than traditional in-class courses \\[[@b14-ojphi-02-20], [@b15-ojphi-02-20]\\]. The reasons cited for this include student isolation and technological barriers which in turn de-motivates students and lead to course drop out \\[[@b16-ojphi-02-20], [@b17-ojphi-02-20], [@b18-ojphi-02-20]\\]. The current study will provide insights into how the technological environments can be developed so that users (i.e. students) would not only learn but also have a positive experience that in turn enhances the probability of course completion.\n\nFurther, we empirically show that higher levels of student self-motivation do not translate into course completion, which in turn emphasizes the need to focus on student's experience during the course to enhance program success.\n\nThe remainder of this paper is organized as follows: The next section reviews the background literature and theories for this study: 1) workforce development for public health emergency preparedness, 2) online learning environment and online consumer experience, 3) cognitive affective learning, 4) social Learning theory, 5) usability in distance-learning environments, and 6) motivation and course completion. Following that we formally define our study research questions. This is followed by the methodology section which includes details on data collection and data analysis. Next, we discuss the study results and their implications. The report ends with a brief conclusion and key recommendations for improving the users' online experience and thereby enhancing program effectiveness. .\n\nBackground\n==========\n\nWorkforce development for public health emergency preparedness\n--------------------------------------------------------------\n\nIn 1997, the U.S. Department of Health and Human Services issued a report titled 'The Public Health Workforce: An Agenda for the 21^st^ Century', which highlighted the gap in training and preparation for public health professionals for emergency preparedness \\[[@b2-ojphi-02-20]\\]. It is estimated that there are around 500,000 people in the public health workforce at the federal, state and local levels. In addition, there are around 3 million people working in the healthcare system (private and non profit) who play a key role in public health emergencies \\[[@b1-ojphi-02-20]\\]. In case of an emergency situation, be it an epidemic, terrorist attack or a natural disaster, these are the people who will be deployed to the front lines and the report raised concerns regarding their training and readiness. According to their assessment this \"compelling and urgent programmatic forces are making enhanced training and education opportunities for public health professionals a necessity.\" \\[[@b1-ojphi-02-20]\\].\n\nAs a result, in September 2000, the Center for Disease Control (CDC) and the Association of Schools of Public Health together brought out a plan to develop a national network of public health preparedness centers. As part of the plan, they funded several University programs to start Centers of Public Health Preparedness (CPHP) in around 10 regions across the country. CPHPs in all these regions have been offering relevant courses to train the public health workforce for emergency preparedness \\[[@b4-ojphi-02-20]\\].\n\nThe Institute of Medicine later released a report in 2003 titled \"*Who Will Keep the Public Healthy?\"* that not only reiterated the need for education and training for the public health workforce, but also stated that online distance learning was the best solution to train this large number of diverse public health workers in a cost-effective manner. This has led several state universities and local governmental agencies to start their own online education programs \\[[@b19-ojphi-02-20], [@b20-ojphi-02-20]\\] for training the public health workforce in emergency preparedness. Despite the growing numbers of such programs, there have been very few initiatives focused on evaluating the online learning environments of these programs, especially for the CPHP offerings, other than the evaluations done by CDC itself.\n\nOnline (or distance) education is definitely a cost effective and efficient way of training such large numbers of public health workforce. However, in order to evaluate such programs, one needs to adopt an interdisciplinary perspective as diverse aspects (technology, social, etc.) assume importance. This study offers a theory based framework drawn from multiple disciplines to evaluate the online environments of such distance education programs.\n\nOnline learning environment and online experience\n-------------------------------------------------\n\nOnline classrooms and learning environments are inevitable to meet the demands of training requirements for public health emergency preparedness. It provides the economies of scale and convenience that will not be available in traditional classroom settings. There are innumerable benefits for students from online distance learning, flexibility and convenience perhaps being the most important \\[[@b16-ojphi-02-20]\\]. However, this is not without many disadvantages and problems \\[[@b21-ojphi-02-20],[@b22-ojphi-02-20],[@b23-ojphi-02-20]\\]. Student isolation \\[[@b17-ojphi-02-20]\\] and student frustrations \\[[@b18-ojphi-02-20]\\] have been found to be two of the major disadvantages with online distance education. A recent study on an online medical self-paced course noted that the major challenges were technological problems and the fact that opportunities for social interaction was much lower \\[[@b16-ojphi-02-20]\\]. Another study found that there were eight main factors that impede online distance education: administrative issues, lack of social interaction, academic skills, technical skills, learner motivation, time and support for studies, cost and access to the Internet, and technical problems \\[[@b24-ojphi-02-20]\\].\n\nSustained frustrations and isolation can impede learning, especially the *cognitive and affective* dimensions of the learning experience \\[[@b25-ojphi-02-20]\\]. Studies in this area show that these would also decrease the storage and processing capacity of working memory \\[[@b26-ojphi-02-20], [@b27-ojphi-02-20]\\]. In addition, frustration and anxiety are major factors that lead to de-motivation among students \\[[@b25-ojphi-02-20]\\]. Motivation is critical for this kind of online learning environments \\[[@b28-ojphi-02-20]\\]. It becomes even more critical when training public health emergency preparedness workers as many students are much older, have full time jobs and other work and family commitments, as compared to young college students \\[[@b20-ojphi-02-20]\\].\n\nCognitive Affective learning\n----------------------------\n\nSimilar to the research in consumer psychology, where pragmatic and hedonic component of experience received much attention, in the education and learning literature, the cognitive (pragmatic) and affective (or hedonic) dimensions of learning has been the focus of many researchers. The cognitive dimension was considered most critical for learning in many of the earlier studies.\n\nWhile the cognitive dimension is critical, researchers also began discovering that there is an affective dimension that impacts learning, memory, retention and inference making. More recently this component received even more focus in the context of online learning which led researchers in the MIT Media Lab to work on affective agents where a robotic computer aims to improve user's motivation to learn. The robotic computer is capable of expressing affect by rewarding or showing pleasure when the learner does something right, and when the learner gets distracted, it would try to entertain the learner and so on. There has also been significant work done in developing affective interface agents that are capable of working as teaching assistants in monitoring and managing online distance learning \\[[@b29-ojphi-02-20], [@b30-ojphi-02-20], [@b31-ojphi-02-20]\\]. The objective of this line of research is to detect the affective or emotional state of the learner and provide appropriate affective or hedonic support to keep the learner engaged in the content and also motivate them to complete the tasks before them.\n\nResearch in consumer online behavior shows that when users are engrossed in the online activity, they do not keep track of time and get into a state of \"flow\" \\[[@b32-ojphi-02-20], [@b33-ojphi-02-20], [@b34-ojphi-02-20]\\]. This stream of research suggests that when people are provided with activities that they get engrossed in and start deriving fun from, they reach a state of flow \\[[@b34-ojphi-02-20]\\]. In the online learning environment, if students are provided with activities that they could get immersed in and achieve a state of flow, it would not only improve learning but also enhance the course completion rates and the student retention rate. Hence, while the cognitive dimension of learning leads the student to evaluate the pragmatic value or the usefulness of the course content, it is the affective component that enables the user to have a better hedonic experience.\n\nSocial Learning theory\n----------------------\n\nAnother relevant stream of research adopts the social learning perspective in which the conjecture is that knowledge is socially constructed and occurs when individuals engage in discourse about a subject matter \\[[@b35-ojphi-02-20], [@b36-ojphi-02-20], [@b37-ojphi-02-20]\\]. Knowledge is embedded in individuals, and by providing effective communication channels and opportunities to interact with one another---either socially or in a classroom setting---it would lead to more knowledge transfer and creation, and in turn offer a richer learning environment \\[[@b36-ojphi-02-20], [@b37-ojphi-02-20]\\].\n\nThis perspective has been widely accepted in the context of online distance education and it is often emphasized that student interactions are central and critical for a successful learning experience and consequently the success of online courses \\[[@b35-ojphi-02-20], [@b38-ojphi-02-20]\\]. These interactions could be with other students or with the instructor. In the context of online public health emergency preparedness courses, especially CPHP courses, almost all the courses are 'canned courses' without an instructor or fellow students. This could potentially affect the sociability experience and thereby impact learning and course completion rates.\n\nUsability in distance-learning environments\n-------------------------------------------\n\nAs mentioned previously, technological barriers and usability issues are the two most often cited reasons for student frustrations and poor completion rates. Several studies have considered the usability issues of different online courses \\[[@b39-ojphi-02-20], [@b40-ojphi-02-20], [@b41-ojphi-02-20]\\] and have broadly concluded that usability is a critical factor in determining the success of any online course.\n\nUsability is the extent to which a user can successfully accomplish the tasks with effectiveness and efficiency \\[[@b42-ojphi-02-20]\\]. In the distance education context, usability would be the effective and efficient accomplishment of learning related tasks or goals in the online environment (with or without using specified tools for that system). In the context of emergency preparedness training courses, it is a critical evaluation component as users' interaction with the system is more than users' interaction with the instructor.\n\nUsability issues are more widely accepted by course providers as a potential problem and many understand the need to rectify them. However, usability issues are much more difficult to evaluate as users often attribute usability issues to their own lack of skills or a problem at their end (for example, their problematic home computer or Internet connection). In addition, many specific usability questions such as \"is navigation through the website easy or difficult?\" can be answered in two different ways -- navigation through the website is easy or difficult for 'everybody else who is skilled in computing technology'; or navigation through the website is easy or difficult for 'me' specifically. Analysis of the results also becomes difficult as users may hold different technology standards, different levels of skills, and access to different levels of technological assistance.\n\nTo overcome these measurements issues, in the current study, we used a simple pre-validated scale to evaluate whether the overall technological environment was easy/difficult; confusing/not confusing; consistent/inconsistent; stressful/not stressful; simple/complicated and tiring/not tiring. This usability tool has been found to be effective in understanding whether the overall usability experience was satisfactory to the user \\[[@b43-ojphi-02-20], [@b44-ojphi-02-20]\\].\n\nMotivation and course completion\n--------------------------------\n\nLack of motivation has been cited as one of the major impediments to online learning \\[[@b15-ojphi-02-20], [@b24-ojphi-02-20], [@b45-ojphi-02-20], [@b46-ojphi-02-20]\\]. Motivation to enroll for courses could come from both internal forces and external forces \\[[@b47-ojphi-02-20]\\].\n\nIntrinsic (or internal) motivation has been indicated as one of the key factors that drive people to register for courses as it reflects a person's need to enhance their skill set, their market value, self-esteem, etc. Extrinsic (or external) motivation relates to one's profession including mandatory job requirements, CPE credits, suggestion from one's boss and colleagues, etc. Extrinsic motivation would also include motivation from educational institutions or the course providers (e.g. instructors, universities, and program administrators). However, the CPHP is not organized to provide this kind of motivation. Hence, the main sources of extrinsic motivation seemed to be from their own professional life.\n\nWhile both intrinsic and extrinsic motivation could certainly lead students to register for courses, there is no evidence yet that this would lead to course completion. In this study we empirically examine whether there is any significant difference in the intrinsic and extrinsic motivation levels of students who completed the courses and that of students who did not complete the courses.\n\nResearch Questions:\n===================\n\nThe above literature review suggests that poor online course experience (that in turn may arise from a lack of instructors, lack of social interactions, technological problems in the online courses, etc) could de-motivate students and lead them to drop the courses that they had registered for. The discussion also suggests that motivation to enroll for a course, while an important factor, may not be enough to ensure that the student completes the course.\n\nThus, in our empirical study, we address two research questions that reflect the above two issues. First, are there any significant differences in students' online course experiences (pragmatic, hedonic, sociability and usability) based on their course completion status? Second, are there any significant differences in students' intrinsic and extrinsic motivation levels based on their course completion status?\n\nBased on the theories and concepts outlined previously, we define student's online course experience---i.e. the overall experience a student derives from his or her interaction in the online course environment---along four dimensions: pragmatic, hedonic, sociability, and usability.\n\nPragmatic experience is the pragmatic or utilitarian value the student experiences in the online learning environment. This dimension is related to goal-oriented behavior \\[[@b33-ojphi-02-20]\\] of the student and reflects whether the student found the experience in the online learning environment useful, valuable, and/or worthwhile \\[[@b43-ojphi-02-20], [@b44-ojphi-02-20]\\]. The hedonic dimension is the intrinsic value the customer derives from the interactions in the online learning environment. It reflects the enjoyment and excitement students derive during the learning process as well as during their interactions in the online learning environment. The sociability dimension is the social experience students derive from the interactions in the online learning environment. It captures students' perceptions regarding the overall openness, friendliness and politeness of the community in the learning environment \\[[@b11-ojphi-02-20],[@b48-ojphi-02-20]\\]. Even though there weren't much human-human interactions in this study context, there were human-computer interactions and such interactions can also lead to sociability experience \\[[@b48-ojphi-02-20]\\]. The usability dimension is defined as the students' experience in navigating and using the online materials. As such, this dimension captures the ease of use and clarity of the technological features of the online learning environment. Higher levels of usability experience reflect the ability of the student to navigate and participate in the online learning environment smoothly and effortlessly and without any obstructions or annoyances that might distract them from their goals or interests \\[[@b11-ojphi-02-20]\\].\n\nNext, we describe our empirical study..\n\nMethod\n======\n\nData collection and Data analysis\n---------------------------------\n\nData was collected using a Web-based questionnaire from students who had registered for the courses offered by a CPHP based in a large public university in upstate NY. Emails were sent to approximately 2700 students who had enrolled in one or more of the courses during the past one year. Each email briefly described the study and invited the student to respond to a survey---the link to the survey was included in the email (the survey was available from the CPHP's Web site). There were 415 responses to the email invite. 38 responses had to be excluded from the analysis due to high amount of missing data. Thus, there were a total of 377 usable responses.\n\nData was collected on different aspects of the online program, including, student motivations, student profile, and their overall experience with the CPHP Web site and with the courses (specifically, the 4 dimensions of user experience---pragmatic, hedonic, sociability and usability). The questionnaire was built using existing scales for measuring each of the variables. Student's online course experience was measured using an existing validated scale designed to measure online experience \\[[@b43-ojphi-02-20], [@b44-ojphi-02-20]\\]. A tool to measure student motivation to enroll was developed by the CPHP staff for an earlier study and was adapted and used in this survey. Course completion data was collected using a simple yes/no question as to whether they completed all the courses they had registered for.\n\nA factor analysis of the data related to student motivation yielded two distinct factors--- intrinsic motivation and extrinsic motivation. See [Table 1](#t1-ojphi-02-20){ref-type=\"table\"} for items and their factor loadings. The intrinsic motivation factor included 4 items and extrinsic motivation factor also had 4 items.\n\nSimilarly factor analysis for each of the online experience dimensions were done separately. Items and factor loadings are provided in [Table 2](#t2-ojphi-02-20){ref-type=\"table\"}. As can be seen in [Table 2](#t2-ojphi-02-20){ref-type=\"table\"}, pragmatic experience was measured using a 7 item scale (reliability \u03b1 = 0.96), hedonic experience by using a 9 item scale (\u03b1 = 0.95), sociability experience by using a 5 item scale (\u03b1 = 0.87), and usability experience by using 6 item scale (\u03b1 = 0.91).\n\nAn independent sample t test was used to compare the means of the four experience dimensions (pragmatic, hedonic, sociability and usability) between students who completed all the courses they registered for and students who did not complete one or more of the courses they registered for. The data was analyzed using SPSS, all the experience dimensions were entered as test variables and the item 'Did you complete all the courses you registered for' was entered as the grouping variable. Similarly, an independent sample t test was used to compare the means of intrinsic and extrinsic motivation between students who completed all the courses they registered for and students who did not complete all the courses they registered for.\n\nResults and Discussion\n======================\n\nMajority of students had registered for just one course. Specifically, 157 people (41.6%) registered for 1 course; 73 people (19.4%) registered for 2 courses, and 24 people (6.4%) had enrolled for a course that was not listed in the survey. The study sample also included students from 31 countries although the large majority was from the United States. The number of female students was much higher (61%). This represents the actual student population ratio at this CPHP. Racial distribution was as follows: 73.7% White non Hispanic, 7.6% black non-Hispanic, 5.4% Hispanic or Latino and 5% South East Asian. This distribution also mirrors the student population distribution at this CPHP.\n\nThe mean and standard deviation for all the 4 dimensions of online experience and the two factors of motivation are provided in [Table 3](#t3-ojphi-02-20){ref-type=\"table\"}.\n\nOnline experience & course completion\n-------------------------------------\n\nAnalysis of the data indicates that, overall, participants had higher levels of pragmatic and usability experience compared to hedonic and sociability experience (see mean values in [Table 3](#t3-ojphi-02-20){ref-type=\"table\"}).\n\nThe results from the independent sample t-test showed that there was significant difference in the scores for all the 4 dimensions of experience between students who completed the course and students who did not. Results are shown in [Table 4](#t4-ojphi-02-20){ref-type=\"table\"}.\n\nThe mean scores for pragmatic experience for students who completed the courses (M =6.2, SD =.96) was significantly higher than for those who did not complete the courses (M=5.5, SD=1.48); t(79.9)=3.89, p\\<.001. The mean scores for hedonic experience for students who completed the courses (M =5.1, SD =1.15) was significantly higher than for those who did not complete the courses (M=4.4, SD=1.53); t(84.9)=3.67, p\\<.001. Similarly, the mean scores for sociability experience for students who completed the courses (M =5.1, SD = 1.07) was significantly higher than for those who did not complete the courses (M=4.3, SD=1.57); t(81.6)=3.68, p\\<.001. Finally, the mean scores for usability experience for students who completed the courses (M =5.8, SD =1.04) was significantly higher than for those who did not complete the courses (M=4.9, SD=1.32); t(87.3)=4.92, p\\<.001.\n\nOverall, the results support the broader study thesis that students who experience higher levels of pragmatic, hedonic, sociability and usability experiences are more likely to complete the course(s) they registered for compared to those who report lower levels. In other words, these results indicate that people who dropped out had less positive online experience on all the four dimensions---pragmatic, hedonic, sociability and usability. The four-dimensional online experience questionnaire is useful in such situations where one can capture the underlying experience and derive insights on what aspect of the user experience really leads to non completion.\n\nOur analysis also shows that students rated 'hedonic' experience and 'sociability' experience lower than 'pragmatic' and usability experience. For sociability experience, a sizeable number of the students gave a rating of 4 (neutral) on a scale of 1 to 7, which indicates that they did not perceive sociability to be either negative or positive. It could also indicate the lack of sociability experience in this CPHP online program.\n\nMotivation and course completion\n--------------------------------\n\nThere was no statistically significant difference in students' extrinsic motivation levels between those who completed all the courses they registered for and those who didn't (See [Table 5](#t5-ojphi-02-20){ref-type=\"table\"}). There was mild statistically significant difference in students' intrinsic levels between those who completed all the courses they registered for and those who did not (M=5.2, SD =2.01); t(106.7) = \u22123.2, p\\<.05. However, the results from the independent sample t test indicate a negative effect. In other words, students who completed all the courses they registered for had lower intrinsic motivation levels compared to those who did not complete all the courses they registered for. This indicates that lower levels of intrinsic motivation do not imply that they would drop out from the course. On the same lines, higher levels of intrinsic motivation do not imply that they would complete the course. In short, the results from this study indicate that student motivation (both intrinsic and extrinsic) is not a good predictor of course completion.\n\nThis finding combined with the earlier finding further indicates the importance of students' online experience (all the four dimensions) for maintaining student interest and ensuring that they complete the courses. In other words, while motivation may play a key role in bringing the student to the program (i.e. enrolling for the course), it is their perceived experience during the online course that critically determines whether or not they would complete the course.\n\nStudy Implications\n==================\n\nThe results from this study have several implications for CPHPs, and more generally, for similar online training programs. First, this study indicates the need to focus on the four key dimensions of user's online experience (i.e. their underlying feelings and perceptions) rather than on ad-hoc issues. Prior studies and evaluations have mainly considered specific problems perceived by the course provider rather than the actual user experience. The evaluation tool described here brings out users' sentiments about different aspects of the program and gives a much more fundamental and holistic understanding of the program's potential weaknesses and areas for improvement.\n\nThe study also highlights the importance of hedonic and sociability experience for students in such online training environments. Many online courses focus mainly on the pragmatic value of a particular course for the students and neglect the potential hedonic experience. While pragmatic experience is important and should be the primary focus, boredom and lack of fun can make students weary and de-motivated. Including elements that enhance fun and entertainment as part of the learning experience would be invaluable. Even in face-to-face classroom situations, instructors constantly try to include fun activities such as field trips, role-playing, including videos and movie clips etc that are relevant for the content of the course. The type of fun activities would be different in an online course (a few suggestions are provided in [Table 6](#t6-ojphi-02-20){ref-type=\"table\"}), but necessary especially for training programs that use 'canned courses'.\n\nAs discussed previously, social learning theory suggests the importance of, sociability experience in learning; the current study findings indicate that sociability experience is equally important to ensure higher course completion rates. Good sociability experience prevents students from \"feeling lonely\", and more importantly, enables them to engage in \"active learning\". Indeed, student interactions have been found to be critical for the success of many online distance education courses \\[[@b49-ojphi-02-20], [@b50-ojphi-02-20], [@b51-ojphi-02-20]\\]. Such interactions allow students to feel that they are part of a community of learners and share experiential knowledge that enhance the overall quality of learning.\n\nFinally, this study found that majority of the students who enrolled in these programs were self-driven or self-motivated. Intrinsic factors such as professional and personal development seem to drive these public health professionals to enroll in such training courses. At the same time, such motivation did not translate into ensuring course completion. This implies that rather than depend on student motivation, course providers would need to provide such self-motivated individuals with a positive and engaging online learning experience to ensure high levels of course completion.\n\nConclusions and Key Recommendations\n===================================\n\nKey recommendations that follow from the study findings are given below (also summarized in [Table 6](#t6-ojphi-02-20){ref-type=\"table\"}).\n\n1.  ***1) Improving pragmatic experience***: In this study, the majority of the students found the courses to be useful and valuable (the mean score for pragmatic experience was higher than those for the other three experience dimensions). However, this is still relative to the very low hedonic and sociability experiences and indicates the potential for improvement. An important means to enhance pragmatic experience is effective student expectations management. Students should be able understand upfront what they will be getting out of each course. This can be done by bringing more clarity to course descriptions and also detailing as to what specific goals students will be able to accomplish by taking each course. It will also help to indicate who would benefit by taking a particular course.\n\n2.  ***2) Improving Hedonic experience***: Hedonic or fun and entertainment from these courses were rated quite low. It is true that fun and entertainment is not one of the primary objectives of these courses. However, as mentioned previously, when people get engrossed in the learning material, their learning and retention of the material are typically much higher \\[[@b52-ojphi-02-20]\\]. In addition, they would try to finish the courses, instead of procrastinating and/or getting distracted with other things. An effective way to improve hedonic experience is to create more interactive and fun user interface. For example, one could incorporate video clips made in 'Second Life' that will give the user a personal view of a disaster and how things could be as he/she approaches a disaster area in addition to being a fun experience. Play2train <http://play2train.us/wordpress/> developed by Idaho Bioterrorism Awareness and Preparedness Program using 'Second Life' is a good example of this.\n\n3.  ***3) Improving Sociability experience***: Sociability experience was another weak factor in CPHP courses\\... As noted previously, positive sociability experience would enable students to feel that they are part of the overall community of students who are enrolled in the program. One solution would be to provide students with access to an online community/forum within the CPHP that will enable interactions with fellow students as well as with the CPHP staff. This would not only improve students' sociability experience, but also enhance their learning and networking potential, and in turn, improve student retention. Developing such forums is a very cost-effective solution with proven benefits given the low cost of associated information technologies.\n\n4.  ***4) Improving Usability experience***: It is important to ensure that the design of the online environment provides seamless and enjoyable navigation experience for the user. Best practices in usability include offering simple and clutter less user interface, intuitive navigational features, and avoiding technological jargons in user guidance. . In addition, usability can be significantly improved by offering online programs on mobile platforms and thereby catering to today's public health worker who is likely to be very mobile. If courses can be accessed through smart phones (this would require redesigning the interface to fit the mobile device) it would improve the convenience factor significantly.\n\nIn conclusion, there are ample opportunities to improve the pragmatic, hedonic, sociability and usability experiences of the target audience. This is critical to improve the participants' learning and retention as well as the completion rates for the courses offered. Online emergency preparedness programs are likely to play a crucial role in preparing emergency responders at all levels in the future and their success has critical implications for public health informatics.\n\nHowever, we need more studies in the future to understand the factors that affect students' overall experience in the online learning environment of CPHP courses. Future research could focus on understanding how the experience (pragmatic, hedonic, sociability and usability) would impact student performance in the courses, student learning and retention of the subject matter, and more importantly, their real life job performance. In addition, conducting qualitative studies with a cohort group of students could help us better understand the factors that shape the overall experience specific to this set of population as well as whether such online training is an effective long term solution for training our public health workers.\n\n###### \n\n**Factor scores for 'motivation'**\n\n  ---------------------------------------------------- ------\n  **Intrinsic motivation**                             \n  I registered to gain more knowledge                  .922\n  I registered myself for personal development         .922\n  I registered myself for professional development     .887\n  I registered to do something useful/constructive     .810\n  **Extrinsic motivation**                             \n  It was required for other educational programs       .839\n  It was recommended by someone outside my workplace   .703\n  It provided continuing education credit              .620\n  It was required/highly recommended for my job        .588\n  ---------------------------------------------------- ------\n\n###### \n\n**Factor scores for Online Experience**\n\n  ----------------------------------------- --------\n  **Pragmatic**                             Scores\n                                            \n  Valuable/Not Valuable                     .938\n  Practical/Impractical                     .918\n  Relevant/Irrelevant                       .915\n  Informative/Not Informative               .905\n  Worthwhile/Worthless                      .904\n  Productive/Not Productive                 .903\n  Useful/Not Useful                         .893\n                                            \n  **Hedonic**                               \n                                            \n  Stimulating/Boring                        .914\n  Exciting/Not exciting                     .892\n  Captivating/Not captivating               .872\n  Fun/Not fun                               .856\n  Satisfying/Unsatisfying                   .846\n  Enjoyable/Not enjoyable                   .831\n  Entertaining/Not entertaining             .809\n  Deeply engrossing/Not deeply engrossing   .803\n  Pleasant/Unpleasant                       .802\n                                            \n  **Sociability**                           \n                                            \n  Inviting/Uninviting                       .851\n  Friendly/Unfriendly                       .840\n  Polite/Impolite                           .808\n  Personal/Impersonal                       .799\n  Social/Unsocial                           .748\n                                            \n  **Usability**                             \n                                            \n  Simple/Complicated                        .866\n  Easy/Difficult                            .858\n  Confusing/Not Confusing                   .827\n  Not Tiring/Tiring                         .827\n  Consistent/Inconsistent                   .826\n  Stressful/Not Stressful                   .807\n  ----------------------------------------- --------\n\n###### \n\n**Means and Standard Deviation of Study Variables**\n\n  Variables                    Mean   S.D\n  ---------------------------- ------ ------\n  1\\. Pragmatic experience     6.1    1.12\n  2\\. Hedonic experience       5.0    1.26\n  3\\. Sociability experience   4.9    1.22\n  4\\. Usability experience     5.6    1.15\n  5\\. Intrinsic motivation     5.4    1.96\n  6\\. Extrinsic motivation     3.8    1.52\n\n###### \n\n**T test results for online experience & course completion**\n\n  ---------------------------------------------------------------------------------------------------------------\n  Experience    Means       N      Std. deviation   DoF    T value\n  ------------- ----------- ------ ---------------- ------ ------------------------------------------------------\n  Pragmatic     Yes\u2003\u20036.2\\   275\\   .96\\             79.9   3.89[\\*\\*\\*](#tfn4-ojphi-02-20){ref-type=\"table-fn\"}\n                No\u2003\u20035.5     67     1.48                    \n\n  Hedonic       Yes\u2003\u20035.1\\   274\\   1.15\\            84.9   3.67[\\*\\*\\*](#tfn4-ojphi-02-20){ref-type=\"table-fn\"}\n                No\u2003\u20034.4     67     1.53                    \n\n  Sociability   Yes\u2003\u20035.1\\   275\\   1.07\\            81.6   3.68[\\*\\*\\*](#tfn4-ojphi-02-20){ref-type=\"table-fn\"}\n                No\u2003\u20034.3     67     1.57                    \n\n  Usability     Yes\u2003\u20035.8\\   275\\   1.04\\            87.3   4.92[\\*\\*\\*](#tfn4-ojphi-02-20){ref-type=\"table-fn\"}\n                No\u2003\u20034.9     67     1.32                    \n  ---------------------------------------------------------------------------------------------------------------\n\nDoF -- Degrees of freedom\n\nYes -- Completed all the courses they enrolled\n\nNo -- Did not complete all the courses they enrolled\n\np\\<.001;\n\np\\<.01;\n\np\\<.05\n\n###### \n\nT test results for Motivation and course completion\n\n  ------------------------------------------------------------------------------------------------------------\n  Motivation   Means       N      Std. deviation   DoF     T value\n  ------------ ----------- ------ ---------------- ------- ---------------------------------------------------\n  Intrinsic    Yes\u2003\u20035.2\\   238\\   2.01\\            106.7   \u22123.2[\\*](#tfn12-ojphi-02-20){ref-type=\"table-fn\"}\n               No\u2003\u20036.0     59     1.62                     \n\n  Extrinsic    Yes\u2003\u20033.8\\   213\\   1.45\\            66.7    \u2212.259\n               No\u2003\u2003 3.9    51     1.78                     \n  ------------------------------------------------------------------------------------------------------------\n\nDoF -- Degrees of freedom\n\nYes -- Completed all the courses they enrolled\n\nNo -- Did not complete all the courses they enrolled\n\np\\<.001;\n\np\\<.01;\n\np\\<.05\n\n###### \n\nKey strategies for improving course completion rates and overall program success\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  **Key Strategies**                        **How to**                                                                                                                                                                                                                                                                  **Benefits for students**                                                                                                                                                                                                                                                                                                                                                                                                                                 **Benefits for CPHP**\n  ----------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  **1) Improving pragmatic experience**     [Expectations management:]{.ul} - make clear what the content of the course is- make it clear upfront who would benefit from the course and who should be taking it.- collect feedback from students at the end of each course on how useful and valuable the course was.   \\- will not unnecessarily register for a course that they do not need.                                                                                                                                                                                                                                                                                                                                                                                    \\- will be able to target courses better at the right individuals- will be able to improve the content of the course\n\n  **2) Improving hedonic experience**       [Improve fun and entertainment:]{.ul} - Add more interactive elements in the courses- Include video clips made with 'Second Life'. E.g. Play2train <http://play2train.us/wordpress/>- Include pictures and graphics (pictures speak a thousand words)                       \\- will capture the attention of students- will improve learning and retention of the material.- will see courses as more fun than as a chore.- will keep student engrossed (time flies when you are deeply engrossed).- will keep them from getting distracted.                                                                                                                                                                                          \\- will improve the success of the overall program.- will improve student ratings- will be able to attract more students (such online programs don't have any boundaries, so the potential is immense).- will be able to retain students and get them to come back for more courses.\n\n  **3) Improving sociability experience**   [Improve possibilities for social interaction.]{.ul} - Provide an online community/forum for students to interact- Allow students as well as CPHP staff to interact in the community- Offer some courses in 'blended format' -- i.e. part online and part in-class.         \\- will improve [networking potential]{.ul}- will improve their social experience- will improve learning and retention (collective learning seems to improve information processing)- will feel part of the CPHP community- will not feel that they are on their own- Blended format offers the convenience of online courses but will provide some f2f time that will enhance sociability experience. However, this will be limited to local students.   \\- will improve CPHP's relationship with students (strong ties).- will be able to attract more students through 'word-of-mouth' marketing (which is a potential outcome of such online communities).- will be able to understand student needs by keeping abreast of the ongoing discussions in the community (instant feedback loop).- online communities have been found to improve motivation as well (Huett et al, 2007)- Blended format will allow CPHPs to improve the variety of courses offered.- It will allow CPHPs to get to know their students better.- Will improve student retention in the local region.\n\n  **4) Improving usability experience**     [1) Improve usability experience by using some of the standard usability practices (Nielson, 2000).]{.ul}\\                                                                                                                                                                  \\- improved usability would make it easier for students to access the course materials and reduce the learning curve related to the course technologies-Convenience would be the biggest benefit for students.- Beneficial for public health workers who are always travelling.- Very beneficial for people who use public transportation and have lot of time while travelling as well as during wait times.                                             \\- improved usability can improve student retention and continued enrollment.- Benefits for CPHP include improving versatility of courses.- Offering courses using more than one platform will improve the reach and enrollment levels- Will improve completion rates and continued enrollment.\n                                            - Update the websites regularly (at least every 2 years or so if not more frequently) using the latest technologies- Use simple designs (Nielson, 2000)- Remove unnecessary content and avoid clutter.\\                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n                                            [3) Make CPHP courses accessible through Mobile phones]{.ul}\\                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         \n                                            Mobile phones are now very well equipped with fast connections and readable interfaces.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               \n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n"}
+{"text": "1. Introduction {#sec1-materials-12-00418}\n===============\n\nMetal component additive manufacturing (AM) is the general designation for a series of advanced manufacturing processes based on a three-dimensional digital model that are used to fabricate metal components by melting and depositing metal powders or wires layer upon layer. AM can be divided into two main categories according to the different supply methods of the raw metal materials: powder bed fusion (PBF) and direct energy deposition (DED). PBF processes, such as selective laser melting (SLM), use thermal energy to selectively fuse regions of a powder bed, while DED processes use focused thermal energy to fuse materials by melting during deposition \\[[@B1-materials-12-00418]\\]. DED methods, including laser metal deposition (LMD), are more suitable for fabricating large titanium structural components than PBF processes \\[[@B2-materials-12-00418],[@B3-materials-12-00418],[@B4-materials-12-00418],[@B5-materials-12-00418],[@B6-materials-12-00418]\\]. Large titanium structural components have been successfully fabricated by many researchers and research institutions by using LMD and have been widely used as load-bearing structures in aeronautical engineering \\[[@B7-materials-12-00418],[@B8-materials-12-00418]\\].\n\nAeronautical structures experience repeat loads over their long periods of service, and fatigue is one of the main damage phenomena in aerospace metallic structures \\[[@B9-materials-12-00418]\\]. Therefore, the fatigue behavior of LMD titanium components must be further elucidated, and fatigue performance characterization methods must be established. The major factors causing the fatigue failure of LMD and other AM titanium alloys include the microstructure, internal defects, and residual stresses \\[[@B10-materials-12-00418]\\].\n\nThe specific thermodynamic process used during manufacturing, i.e., cyclic heating and rapid cooling, results in the unique microstructure of the LMD titanium alloy, which includes relatively large columnar grains, multiple melt layers between the processing layers and unique microstructures \\[[@B11-materials-12-00418]\\]. These specific characteristics in the microstructures may have some influence on the fatigue performance of the parts \\[[@B12-materials-12-00418],[@B13-materials-12-00418],[@B14-materials-12-00418],[@B15-materials-12-00418]\\]. Wang et al. \\[[@B12-materials-12-00418]\\] studied the fatigue strength of LMD TC18(Ti-5Al-5Mo-5V-1Cr-1Fe) titanium alloy and determined that its strength is much lower than that of wrought TC18 titanium alloy. Sterling et al. \\[[@B13-materials-12-00418]\\] reported that the fatigue lives of LMD Ti-6Al-4V specimens, including as-built and annealed specimens, are shorter than those of wrought specimens. Zhai et al. \\[[@B14-materials-12-00418],[@B15-materials-12-00418]\\] also determined that LMD Ti-6Al-4V has a lower fatigue crack growth threshold than wrought Ti-6Al-4V. Studies have also been conducted on the fatigue mechanism and the effects of the microstructures of titanium alloys manufactured by LMD and other AM processes. The studies of Rafi et al. \\[[@B16-materials-12-00418]\\] and Zhai \\[[@B13-materials-12-00418]\\] et al. showed that, compared to EBM Ti-6Al-4V, LMD and SLM Ti-6Al-4V parts have a higher fatigue strength but lower fatigue toughness, and this phenomenon may result from the presence of fine \u03b1' martensite.\n\nInternal defects are another major factor affecting the fatigue performance. Such defects introduce a stress concentration around them and easily become a nucleation site of crack \\[[@B17-materials-12-00418]\\]. Biswas et al. \\[[@B18-materials-12-00418]\\] and Li et al. \\[[@B19-materials-12-00418]\\] revealed that the internal defects, such as lack of fusion (LOF) and pores, are likely to become nucleation sites for shear bands and micro-cracks. The presence of internal defects may also lead to mixed failure behaviors that were observed in many studies \\[[@B12-materials-12-00418],[@B13-materials-12-00418],[@B17-materials-12-00418],[@B20-materials-12-00418],[@B21-materials-12-00418],[@B22-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418],[@B26-materials-12-00418]\\], inconsistent with the case of traditionally manufactured titanium alloy. The crack may initiate from both surface and internal defects of specimens. The influences of different types of defects are also investigated. Akerfeldt et al. \\[[@B20-materials-12-00418]\\] stated that the internal LOF is more likely to result in cracks than pores, and this phenomenon is particularly significant in specimens with a force directed parallel to the powder deposition direction. Prabhu et al. \\[[@B21-materials-12-00418]\\] showed that unmelted particles in LMD titanium alloy specimens have a significant effect on the fatigue life, while the pores have a minimal effect. Leuders et al. \\[[@B22-materials-12-00418]\\] considered that the pores within specimens have a drastic effect on the fatigue behavior of SLM titanium alloys in the high-cycle-fatigue (HCF) regime.\n\nThe location and size of internal defects also affect the fatigue properties to a great extent. Many investigations have shown that the nearer pores are to the surface, the shorter the fatigue life \\[[@B13-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418],[@B26-materials-12-00418]\\]. This phenomenon can be partly explained by Xu et al.'s \\[[@B27-materials-12-00418]\\] study. The cross-sectional area of a defect normal to the applied stress has a large effect on the fatigue life properties \\[[@B28-materials-12-00418]\\]. Several studies reported that fatigue failure occurs first at the largest defect, so there is a shorter fatigue life for AM specimens in the presence of larger internal defects \\[[@B13-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418]\\]. However, He et al. \\[[@B26-materials-12-00418]\\] indicated a different conclusion that a larger pore may lead to a longer fatigue life when the pores have similar distances to the surface.\n\nResidual stresses are also a direct cause of crack initiation that can affect the fatigue properties of AM specimens. AM processes, particularly LMD and other laser-based AM processes, are prone to induce significant residual stresses due to their large inherent temperature gradients \\[[@B29-materials-12-00418]\\]. Different process parameters and deposition strategies may lead to different residual stress distributions. Tensile residual stresses are always distributed at the surface and near-surface zone of as-built AM titanium alloy parts and may have a significant influence on the fatigue properties \\[[@B30-materials-12-00418]\\]. Although the residual stress can be partly relieved by some heat treatment processes, it is hardly eliminated completely \\[[@B22-materials-12-00418],[@B31-materials-12-00418],[@B32-materials-12-00418],[@B33-materials-12-00418]\\].\n\nThe fatigue behaviors of AM materials are usually complicated due to their unique microstructure, internal defects, and residual stress distributions. As a new manufacturing method, the fatigue life properties must be precisely described. The experimental techniques and description methods for *S-N* curves have long been a concern \\[[@B34-materials-12-00418]\\]. The *S-N* curves of materials show the pre-conditions and inputs for anti-fatigue structure design. Fatigue life data always have significant variations, and the fatigue life under a specific stress level is closely related to the survival rate, *P*. In many cases, especially for the reliability design of components, the relationships between the fatigue stress and life at different survival rates, namely, the *P-S-N* curves, must be determined. The previously mentioned *S-N* curve is the median fatigue life curve, i.e., the *P-S-N* curve with a 50% survival rate. Scholars have conducted studies on the fatigue properties of AM materials in recent years, and the *P-S-N* curves (*S-N* curves) of different materials manufactured by several different AM processes have been tested \\[[@B35-materials-12-00418],[@B36-materials-12-00418],[@B37-materials-12-00418],[@B38-materials-12-00418],[@B39-materials-12-00418],[@B40-materials-12-00418]\\]. Regulations and standards such as ASTM E739-10 \\[[@B41-materials-12-00418]\\] and ISO 12107:2012 \\[[@B42-materials-12-00418]\\] have been issued with test procedures and method descriptions for *P-S-N* curves (*S-N* curves) based on a lognormal distribution. However, the unique AM fabrication process leads to various failure behaviors, and the fatigue life of AM materials therefore displays significantly greater uncertainty and variation than that of their conventionally manufactured counterparts \\[[@B26-materials-12-00418],[@B40-materials-12-00418],[@B43-materials-12-00418],[@B44-materials-12-00418]\\]. The mixed failure behaviors and large variation make the existing description methods unable to accurately describe the fatigue properties of AM materials, and the following problems may occur when the existing description method is used to determine the fatigue *P-S-N* curves of AM materials:\n\n\\(i\\) Because of the complex fatigue failure behaviors, the fatigue life distribution of AM materials may be different from that of traditional materials, and the traditional distribution model thus cannot accurately describe the fatigue life variation under specific stress levels.\n\n\\(ii\\) Due to the large variation in the fatigue life, the reliability life in the high-reliability region will be very short using the traditional *P-S-N* curve model to describe the fatigue properties of AM materials.\n\n\\(iii\\) The large fatigue life variation requires a larger number of fatigue test specimens to determine the *P-S-N* curve, which will inevitably increase the time and economic cost of the process.\n\nIn this paper, two sets of fatigue tests under peak stresses of 720 and 760 MPa were conducted, and fatigue life data under three different stress levels were obtained, including the data under 800 MPa published in \\[[@B26-materials-12-00418]\\]. Sufficient data, no fewer than 15 specimens under each stress level, were used to develop and prove the conclusions. Compared with the existing research, especially the previous study (\\[[@B26-materials-12-00418]\\]), there are four main innovations and contributions in this paper:\n\n\\(i\\) This paper examines more data (at least 15 specimens under each stress level) than in previous studies, which generally reported the data of 5--10 specimens under the same stress \\[[@B23-materials-12-00418],[@B26-materials-12-00418],[@B35-materials-12-00418],[@B37-materials-12-00418],[@B38-materials-12-00418],[@B39-materials-12-00418],[@B40-materials-12-00418],[@B45-materials-12-00418]\\]; the data herein were used to illustrate the mixed failure behavior and fatigue life properties.\n\n\\(ii\\) A *P-S-N* curve description method of the LMD Ti-6.5Al-2Zr-1Mo-1V titanium alloy was established based on a bimodal lognormal distribution (BLG).\n\n\\(iii\\) Considering the disadvantages of the parameter estimation method based on rank distribution theory developed in \\[[@B26-materials-12-00418]\\], especially the weakness in robustness, the maximum likelihood estimation (MLE) method was used to estimate the parameters, and the Newton--Raphson algorithm was used to solve the equations.\n\n\\(iv\\) Compared to previous studies \\[[@B13-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418],[@B26-materials-12-00418]\\], the mixed failure behavior and the influence of internal pores on the fatigue life of LMD Ti-6.5Al-2Zr-1Mo-1V specimens were discussed in more detail, and the conclusion regarding the influence of the size and location of the pore defects on the fatigue life was corrected by analyzing the greater amount of data in this paper.\n\n2. Materials and Experiments {#sec2-materials-12-00418}\n============================\n\n2.1. Specimens {#sec2dot1-materials-12-00418}\n--------------\n\nThe Ti-6.5Al-2Zr-1Mo-1V titanium alloy specimens used in this paper were fabricated in the same batch as those in \\[[@B26-materials-12-00418]\\] with the same manufacturing procedures. A Ti-6.5Al-2Zr-1Mo-1V titanium alloy rectangular plate was deposited on a substrate made of wrought Ti-6.5Al-2Zr-1Mo-1V titanium alloy, as shown in [Figure 1](#materials-12-00418-f001){ref-type=\"fig\"}a. The spherical Ti-6.5Al-2Zr-1Mo-1V powder used for the LMD was prepared by a plasma rotation electrode process. The Ti-6.5Al-2Zr-1Mo-1V titanium alloy plate was machined into standard smooth cylindrical specimens after duplex annealing (850 \u00b0C AC 1 h, 650 \u00b0C FC 2 h), as shown in [Figure 1](#materials-12-00418-f001){ref-type=\"fig\"}b, and had a surface roughness, *Ra*, of 0.32 \u03bcm.\n\n[Figure 2](#materials-12-00418-f002){ref-type=\"fig\"} shows the microstructure of the material. The material mainly consists of coarse columnar grains with the growth direction in the Z direction. [Figure 2](#materials-12-00418-f002){ref-type=\"fig\"}a shows that the size of the columnar grains is approximately 900 \u03bcm. As shown in [Figure 2](#materials-12-00418-f002){ref-type=\"fig\"}b, the columnar grains mainly consist of a basket-weave microstructure, and a large number of *\u03b1* colonies are present near the grain boundary. However, [Figure 2](#materials-12-00418-f002){ref-type=\"fig\"}b also shows that the microstructure of the material is inhomogeneous. In addition to the fine basket-weave organization, an *\u03b1* lamellar microstructure also exists in the material.\n\n2.2. Experiments {#sec2dot2-materials-12-00418}\n----------------\n\nThe fatigue test was performed on an Instron 8801-100kN electrohydraulic servo fatigue system in an atmospheric environment at room temperature under a sine-wave constant-amplitude (CA) stress with a stress ratio of 0.06 and frequency of 10 Hz, the same experimental procedures and conditions used in \\[[@B26-materials-12-00418]\\]. Two sets of fatigue tests with peak stresses of 720 and 760 MPa were carried out. Therefore, three sets of fatigue test data were obtained in total, including the data under 800 MPa in \\[[@B26-materials-12-00418]\\]. Fracture surface examinations and fractographic analyses were performed by scanning electron microscopy (SEM, JEOL, Tokyo, Japan/Carl Zeiss, Jena, Germany) after the fatigue tests, and the crack initiation and failure forms were identified.\n\n3. Tests Results {#sec3-materials-12-00418}\n================\n\nTwo sets of fatigue tests were conducted under two different stress levels and combined with the test results in \\[[@B26-materials-12-00418]\\]. Fifteen, 17, and 22 specimens under 720, 760, and 800 MPa, respectively, were analyzed in total. Fracture surface examinations and fractographic analyses were performed, and mixed failure behaviors appeared in all three sets of fatigue tests. All the fracture surfaces were classified into two categories according to the crack initiation locations, and the fatigue life distribution was preliminarily estimated based on a lognormal distribution.\n\n3.1. Fractographic Analysis {#sec3dot1-materials-12-00418}\n---------------------------\n\nThe fatigue lives of all specimens were in the range of the medium- and high-cycle fatigue regimes. The fractographic, in [Figure 3](#materials-12-00418-f003){ref-type=\"fig\"}, shows that mixed failure behaviors appeared under the three stress levels. Parts of the specimens, with internal crack initiation, fractured due to cracks initiating from internal defects such as internal pores were denoted as SII. The other parts of the specimens that fractured due to cracks initiating from the surface or subsurface of the specimens were denoted as SIS.\n\nFor SII, the crack origins were all pores, and no LOF was observed in this investigation. A typical fracture photograph of SII is shown in [Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}a,c. The center of the region close to the initiation of the crack is bright, while the surrounding area is dark. For SIS, the crack origins are usually mechanical scratches and *\u03b1* lamellar structures located on the surface or subsurface of the specimens. A typical fracture photograph of SIS is shown in [Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}b,d. Significant radial ridges can be seen on the fracture surface, extending radially from the origin site to the surrounding area.\n\nThe fracture surfaces of SII and SIS can be divided into three regions, as shown in [Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}a,b:\n\n\\(i\\) crack initiation site (CIS): the crack initiation location;\n\n\\(ii\\) crack propagation region (CPR): a flat plane vertical to the load direction that forms as the result of the crack-stabilized propagation;\n\n\\(iii\\) fast fracture region (FFR): the area produced by fast crack propagation and fracture.\n\nCIS and CPR are relatively flat and nearly in the same plane. As shown in [Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}c,d, the FFR can be divided into two parts, denoted as FFR I and FFR II. FFR I is relatively flat and roughly in the same plane as CIS and CPR, while FFR II is a plane inclined at an angle of approximately 45\u00b0 relative to the plane of the CIS and CPR. Different stress--strain states when fracture occurs may lead to the differences in the morphology of the two parts of the FFR. FFR I is close to the ideal plane strain state due to constraints inside the materials, while FFR II is similar to the plane stress state because of the small distance from the surface of the specimen. Obvious dimple features can be seen in the FFR ([Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}e).\n\nThe characteristics of the CPR are almost the same for SIS and SII. The typical characteristics are shown in [Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}e,f. Fatigue striations and secondary cracks were found in the CPR ([Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}e). Unmelted particles were observed, and their location and shape were irregular ([Figure 4](#materials-12-00418-f004){ref-type=\"fig\"}f). The formation of secondary cracks may be related to the laser metal deposition (LMD) parameters, heat treatment process, and stress state. Some powder was not completely melted and remained as unmelted particles, which may have a spherical shape or be present as partially melted irregular particles. These particles produced irregular voids inside the specimen, and these voids could affect the fatigue properties of the specimen.\n\nFor SIS, cracks initiate from coarse \u03b1 lamellar grains or mechanical scratches located on the surface or subsurface of the specimens, as shown in [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}a,b. Mechanical scratches may be traces left by tools during the machining process. [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}a,b show apparent cleavage faces near the CIS, and cleavage feather-like features can be easily identified.\n\nFor SII specimens, as shown in [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}, cracks originate from internal pores. The sizes of the pores inducing the cracks vary greatly between different specimens. [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}c,d are typical fracture photographs of two SII specimens. In [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}c, the pore area is relatively large, and many discontinuous cleavage faces can be seen near the pore. In [Figure 5](#materials-12-00418-f005){ref-type=\"fig\"}d, the pore area is small, and a continuous cleavage face can be seen around the pore. Thus, the crack initiation mechanism induced by the different pore sizes may be different.\n\n3.2. Fatigue Life Data {#sec3dot2-materials-12-00418}\n----------------------\n\nThe fatigue life data under three stress levels obtained from the fatigue tests are shown in [Figure 6](#materials-12-00418-f006){ref-type=\"fig\"}. The pentagons represent the fatigue lives of the specimens under 720 MPa, while the triangles and diamonds represent those of the specimens under 760 and 800 MPa, respectively. The fatigue lives of all specimens are in the range of medium and long fatigue lifetimes. The ranges of the fatigue lives are 108,089--1,559,872 cycles (720 MPa), 49,758--645,171 cycles (760 MPa), and 30,715--544,102 cycles (800 MPa), respectively. The data dots representing the fatigue life of SII are blue, while those of SIS are red. The fatigue lives of SII and SIS overlap with each other. Therefore, the fatigue lives of SII and SIS cannot be differentiated using the location and type of CIS.\n\nThe fatigue life is denoted as the random variable *Y*. Conventionally, a lognormal distribution (LG) can be used to describe the fatigue life distribution of metal materials. Assuming that the fatigue life follows an LG, the probability density function (PDF), $f_{Y}\\left( y \\right)$, and the cumulative distribution function (CDF), $F_{Y}\\left( y \\right)$, of *Y* are given by Equations (1) and (2), respectively. $$f_{Y}\\left( y \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\mathsf{\\pi}} \\cdot \\sigma_{LG} \\cdot y \\cdot \\ln 10}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{\\lg y - \\mu_{LG}}{\\sigma_{LG}} \\right)^{2}} \\right\\rbrack} & {y > 0,} \\\\\n0 & {y \\leq 0.} \\\\\n\\end{cases}$$ $$F_{Y}\\left( y \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\mathsf{\\pi}} \\cdot \\sigma_{LG} \\cdot \\ln 10}{\\int_{0}^{y}{\\frac{1}{y}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{\\lg y - \\mu_{LG}}{\\sigma_{LG}} \\right)^{2}} \\right\\rbrack}}dy} & {y > 0,} \\\\\n0 & {y \\leq 0.} \\\\\n\\end{cases}$$ $$\\left\\{ \\begin{array}{l}\n{\\mu_{LG} = \\frac{1}{n}{\\sum\\limits_{i = 1}^{n}{\\lg y_{i}}},} \\\\\n{\\sigma_{LG} = \\sqrt{\\frac{1}{n - 1}{\\sum\\limits_{i = 1}^{n}\\left( {\\lg y_{i} - \\mu_{LG}} \\right)^{2}}},} \\\\\n \\\\\n{N_{50,{LG}} = 10^{\\mu_{LG}}.} \\\\\n\\end{array} \\right.$$\n\nTwo parameters, the logarithmic expectation (*\u03bc*~LG~) and logarithmic life standard deviation (*\u03c3*~LG~), can be used to determine the distribution. The parameters under the three stress levels were estimated by using the MLE method by using Equation (3) and are listed in [Table 1](#materials-12-00418-t001){ref-type=\"table\"}. The table also contains the maximum and minimum values of the fatigue life and the logarithmic median life, $N_{50,{LG}}$ under the three stress levels. [Table 1](#materials-12-00418-t001){ref-type=\"table\"} shows that the *\u03c3*~LG~ values under all three stress levels are no less than 0.3 and increase with the stress level, indicating that the fatigue life has a large variation. The maximum fatigue life under each stress level is approximately 15 times the minimum value. The variation may be caused by the mixed failure behavior described in [Section 3.1](#sec3dot1-materials-12-00418){ref-type=\"sec\"}.\n\nTo further elucidate the distribution of the fatigue life, frequency distribution histograms (FDHs) of the logarithmic lives under the three stress levels are plotted in [Figure 7](#materials-12-00418-f007){ref-type=\"fig\"} with an interval of 0.2. The estimated PDF curves are also plotted in the figure (red curves). The PDF curves of the LG have significant differences from the FDHs under the three stress levels, and the FDHs have apparent double-peak characteristics. Therefore, the fatigue life distribution of LMD titanium alloy cannot be well fitted by using an LG.\n\n4. Bimodal Lognormal Distribution {#sec4-materials-12-00418}\n=================================\n\nSince the PDFs of the LG in [Figure 7](#materials-12-00418-f007){ref-type=\"fig\"} cannot fit the FDHs of the fatigue life data of the specimens, the LG distribution is not suitable to describe the fatigue life of the LMD titanium alloy and will lead to large fatigue life variations. The poor description result of the LG and the significant variation induced by the unsuitable distribution may lead to very short reliability lives under high-reliability demands and may increase the requirement of the number of specimens in the fatigue tests. Thus, a distribution model that can describe the fatigue life of LMD titanium alloy more precisely is needed.\n\nIn this section, a bimodal lognormal distribution (BLG), proposed in \\[[@B26-materials-12-00418]\\], is used to describe the fatigue life data of LMD titanium alloys. Considering the disadvantages of the parameter estimation method developed in \\[[@B26-materials-12-00418]\\], the Newton--Raphson algorithm is used to estimate the parameters based on the MLE method.\n\n4.1. Distribution Model {#sec4dot1-materials-12-00418}\n-----------------------\n\nAssume that the fatigue life, *Y*, follows a BLG, $Y \\sim BLG(\\alpha,\\mu_{1},\\sigma_{1}^{2},\\mu_{2},\\sigma_{2}^{2})$, where $\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}$ are the five distribution parameters of the BLG. The PDF, $f_{Y}\\left( y \\right)$, and the CDF, $F_{Y}\\left( y \\right)$, of *Y* are given by Equation (4), where *\u03b1* represents the weight and $0 \\leq \\alpha \\leq 1$. $$\\left\\{ \\begin{array}{l}\n{f_{Y}\\left( y \\right) = \\alpha f_{Y_{1}}\\left( y \\right) + (1 - \\alpha)f_{Y_{2}}\\left( y \\right),} \\\\\n{F_{Y}\\left( y \\right) = \\alpha F_{Y_{1}}\\left( y \\right) + (1 - \\alpha)F_{Y_{2}}\\left( y \\right).} \\\\\n\\end{array} \\right.$$\n\n*Y*~1~ and *Y*~2~ are two independent random variables that follow a lognormal distribution $\\left\\{ \\begin{array}{l}\n{Y_{1} \\sim LG(\\mu_{1},\\sigma_{1}^{2})} \\\\\n{Y_{2} \\sim LG(\\mu_{2},\\sigma_{2}^{2})} \\\\\n\\end{array} \\right.$. The PDF and CDF of *Y*~1~ and *Y*~2~ are shown in Equations (5) and (6), respectively. $$f_{Y_{i}}\\left( y \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\mathsf{\\pi}} \\cdot \\sigma_{i} \\cdot y \\cdot \\ln 10}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{\\lg y - \\mu_{i}}{\\sigma_{i}} \\right)^{2}} \\right\\rbrack} & {y > 0} \\\\\n0 & {y \\leq 0} \\\\\n\\end{cases}(i = 1,2),$$ $$F_{Y_{i}}\\left( y \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\mathsf{\\pi}} \\cdot \\sigma_{i} \\cdot \\ln 10}{\\int_{0}^{y}{\\frac{1}{y}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{\\lg y - \\mu_{i}}{\\sigma_{i}} \\right)^{2}} \\right\\rbrack}}dy} & {y > 0} \\\\\n0 & {y \\leq 0} \\\\\n\\end{cases}(i = 1,2).$$\n\n4.2. BLG Parameter Estimation Method {#sec4dot2-materials-12-00418}\n------------------------------------\n\nIn \\[[@B26-materials-12-00418]\\], a parameter estimation method based on the rank statistics theory is proposed. The basic principle of this method is that the minimization of the Sum of Squared Error (SSE) is used as a criterion for determining the distribution parameters. This method is easy to perform but lacks the rigor of mathematical logic. It also has some weakness in robustness and can easily be affected by the sample alteration. A relatively large sample size is needed. Considering the above shortcomings of the method, the MLE method was used to estimate the parameters. The likelihood function is constructed, and the Newton--Raphson method is used to obtain an estimated value for each parameter. The following are the detailed steps.\n\nFor fatigue life data under a specific stress level, the MLE method can be used to estimate the five distribution parameters $\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}$. To simplify the expression, set *X* = lg*Y*. The random variable *X* follows a bimodal normal distribution (BNG), denoted as $X \\sim BNG(\\alpha,\\mu_{1},\\sigma_{1}^{2},\\mu_{2},\\sigma_{2}^{2})$. The PDF and CDF are given by Equation (7), and the PDF and CDF of *Y*~1~ and *Y*~2~ are determined by Equations (8) and (9), respectively. Obviously, the BLG has the same parameters as the BNG. $$\\left\\{ \\begin{array}{l}\n{f_{X}\\left( {x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}} \\right) = \\alpha f_{X_{1}}\\left( {x;\\mu_{1},\\sigma_{1}} \\right) + (1 - \\alpha)f_{X_{2}}\\left( {x;\\mu_{2},\\sigma_{2}} \\right),} \\\\\n{F_{X}\\left( {x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}} \\right) = \\alpha F_{X_{1}}\\left( {x;\\mu_{1},\\sigma_{1}} \\right) + (1 - \\alpha)F_{X_{2}}\\left( {x;\\mu_{2},\\sigma_{2}} \\right),} \\\\\n\\end{array} \\right.$$ $$f_{X_{i}}\\left( {x;\\mu_{i},\\sigma_{i}} \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\pi} \\cdot \\sigma_{i} \\cdot x}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{x - \\mu_{i}}{\\sigma_{i}} \\right)^{2}} \\right\\rbrack} & {x > 0} \\\\\n0 & {x \\leq 0} \\\\\n\\end{cases}(i = 1,2),$$ $$F_{X_{i}}\\left( {x;\\mu_{i},\\sigma_{i}} \\right) = \\begin{cases}\n{\\frac{1}{\\sqrt{2\\pi} \\cdot \\sigma_{i}}{\\int_{0}^{x}{\\frac{1}{x}\\exp\\left\\lbrack {- \\frac{1}{2}\\left( \\frac{x - \\mu_{i}}{\\sigma_{i}} \\right)^{2}} \\right\\rbrack}}dx} & {x > 0} \\\\\n0 & {x \\leq 0} \\\\\n\\end{cases}(i = 1,2).$$\n\nThe likelihood function, $L(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2})$, can be expressed as Equation (10), $$L(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\prod\\limits_{i = 1}^{n}\\left\\lbrack {\\alpha f_{X}(x;\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x;\\mu_{2},\\sigma_{2})} \\right\\rbrack}.$$ By taking the logarithm of both sides of Equation (10), the logarithmic likelihood function can be obtained as Equation (11), $$\\ln L(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}{\\ln\\left\\lbrack {\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})} \\right\\rbrack}}.$$ The first partial derivative of $\\ln L(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2})$ can then be taken with respect to $\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}$, respectively. If the partial derivatives are equal to zero, then $$\\left\\{ \\begin{array}{l}\n{\\frac{\\partial\\ln L}{\\partial\\alpha} = {\\sum\\limits_{i = 1}^{n}\\frac{f_{X}(x;\\mu_{1},\\sigma_{1}) - f_{X}(x;\\mu_{2},\\sigma_{2})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} = 0,} \\\\\n{\\frac{\\partial\\ln L}{\\partial\\mu_{1}} = {\\sum\\limits_{i = 1}^{n}\\frac{\\alpha}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\frac{(x_{i} - \\mu_{1})}{\\sqrt{2\\pi}\\sigma_{1}^{3}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{1})}^{2}}{2\\sigma_{1}^{2}}} \\right\\rbrack = 0,} \\\\\n{\\frac{\\partial\\ln L}{\\partial\\sigma_{1}} = {\\sum\\limits_{i = 1}^{n}\\frac{\\alpha}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\left\\{ {- \\frac{1}{\\sqrt{2\\pi}\\sigma_{1}^{2}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{1})}^{2}}{2\\sigma_{1}^{2}}} \\right\\rbrack + \\frac{{(x_{i} - \\mu_{1})}^{2}}{\\sqrt{2\\pi}\\sigma_{1}^{4}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{1})}^{2}}{2\\sigma_{1}^{2}}} \\right\\rbrack} \\right\\} = 0,} \\\\\n{\\frac{\\partial\\ln L}{\\partial\\mu_{2}} = {\\sum\\limits_{i = 1}^{n}\\frac{1 - \\alpha}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\frac{(x_{i} - \\mu_{2})}{\\sqrt{2\\pi}\\sigma_{2}^{3}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{2})}^{2}}{2\\sigma_{2}^{2}}} \\right\\rbrack = 0,} \\\\\n{\\frac{\\partial\\ln L}{\\partial\\sigma_{2}} = {\\sum\\limits_{i = 1}^{n}\\frac{1 - \\alpha}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\left\\{ {- \\frac{1}{\\sqrt{2\\pi}\\sigma_{2}^{2}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{2})}^{2}}{2\\sigma_{2}^{2}}} \\right\\rbrack + \\frac{{(x_{i} - \\mu_{2})}^{2}}{\\sqrt{2\\pi}\\sigma_{2}^{4}}\\exp\\left\\lbrack {- \\frac{{(x_{i} - \\mu_{2})}^{2}}{2\\sigma_{2}^{2}}} \\right\\rbrack} \\right\\} = 0} \\\\\n\\end{array} \\right.$$ The equations in Equation (12) can be simplified and denoted as $f_{L,i}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}),i = 1,2,\\cdots,5$: $$\\left\\{ \\begin{array}{l}\n{f_{L,1}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}\\frac{f_{X}(x;\\mu_{1},\\sigma_{1}) - f_{X}(x;\\mu_{2},\\sigma_{2})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} = 0} \\\\\n{f_{L,2}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}\\frac{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\frac{(x_{i} - \\mu_{1})}{\\sigma_{1}^{2}} = 0} \\\\\n{f_{L,3}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}{\\frac{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})} \\bullet \\frac{1}{\\sigma_{1}} \\bullet \\left\\lbrack {\\frac{{(x_{i} - \\mu_{1})}^{2}}{\\sigma_{1}^{3}} - 1} \\right\\rbrack = 0}}} \\\\\n{f_{L,4}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}\\frac{(1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}} \\bullet \\frac{(x_{i} - \\mu_{2})}{\\sigma_{2}^{2}} = 0} \\\\\n{f_{L,5}(x;\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2}) = {\\sum\\limits_{i = 1}^{n}{\\frac{(1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})}{\\alpha f_{X}(x_{i};\\mu_{1},\\sigma_{1}) + (1 - \\alpha)f_{X}(x_{i};\\mu_{2},\\sigma_{2})} \\bullet \\frac{1}{\\sigma_{2}} \\bullet \\left\\lbrack {\\frac{{(x_{i} - \\mu_{2})}^{2}}{\\sigma_{2}^{3}} - 1} \\right\\rbrack = 0}}} \\\\\n\\end{array}. \\right.$$ Equation (13) is thus a nonlinear system with five variables, and obtaining an analytical solution is difficult. Thus, the Newton--Raphson iterative algorithm was used to solve this nonlinear system. Let $\\mathbf{F}_{L}(\\mathbf{\\mathbf{\\theta}}) = \\left( {f_{L,1}(\\mathbf{\\mathbf{\\theta}}),f_{L,2}(\\mathbf{\\mathbf{\\theta}}),f_{L,3}(\\mathbf{\\mathbf{\\theta}}),f_{L,4}(\\mathbf{\\mathbf{\\theta}}),f_{L,5}(\\mathbf{\\mathbf{\\theta}})} \\right)^{T} = 0$ and $\\mathbf{\\mathbf{\\theta}} = {(\\alpha,\\mu_{1},\\sigma_{1},\\mu_{2},\\sigma_{2})}^{T}$. Assuming that the system has been iterated k times, $\\mathbf{\\mathbf{\\theta}}^{(k{+ 1})} = {(\\alpha^{(k{+ 1})},\\mu_{1}{}^{(k{+ 1})},\\sigma_{1}{}^{(k{+ 1})},\\mu_{2}{}^{(k{+ 1})},\\sigma_{2}{}^{(k{+ 1})})}^{T}$ can be calculated by Equation (14). $$\\mathbf{\\mathbf{\\theta}}^{(k + 1)} = \\mathbf{\\mathbf{\\theta}}^{(k)} - \\mathbf{J}^{- 1}\\mathbf{F}_{L}(\\mathbf{\\mathbf{\\theta}}^{(k)}),k = 0,1,2\\cdots$$ where **J** represents the derivative of $\\mathbf{F}_{T}(\\mathbf{\\mathbf{\\theta}})$ at $\\mathbf{\\mathbf{\\theta}}^{(k)}$(Equation (15)). $$\\mathbf{J} = \\left\\lbrack \\begin{array}{l}\n{\\frac{\\partial f_{L,1}}{\\partial\\alpha},\\frac{\\partial f_{L,1}}{\\partial\\mu_{1}},\\frac{\\partial f_{L,1}}{\\partial\\sigma_{1}},\\frac{\\partial f_{L,1}}{\\partial\\mu_{2}},\\frac{\\partial f_{L,1}}{\\partial\\sigma_{2}}} \\\\\n{\\frac{\\partial f_{L,2}}{\\partial\\alpha},\\frac{\\partial f_{L,2}}{\\partial\\mu_{1}},\\frac{\\partial f_{L,2}}{\\partial\\sigma_{1}},\\frac{\\partial f_{L,2}}{\\partial\\mu_{2}},\\frac{\\partial f_{L,2}}{\\partial\\sigma_{2}}} \\\\\n{\\frac{\\partial f_{L,3}}{\\partial\\alpha},\\frac{\\partial f_{L,3}}{\\partial\\mu_{1}},\\frac{\\partial f_{L,3}}{\\partial\\sigma_{1}},\\frac{\\partial f_{L,3}}{\\partial\\mu_{2}},\\frac{\\partial f_{L,3}}{\\partial\\sigma_{2}}} \\\\\n{\\frac{\\partial f_{L,4}}{\\partial\\alpha},\\frac{\\partial f_{L,4}}{\\partial\\mu_{1}},\\frac{\\partial f_{L,4}}{\\partial\\sigma_{1}},\\frac{\\partial f_{L,4}}{\\partial\\mu_{2}},\\frac{\\partial f_{L,4}}{\\partial\\sigma_{2}}} \\\\\n{\\frac{\\partial f_{L,5}}{\\partial\\alpha},\\frac{\\partial f_{L,5}}{\\partial\\mu_{1}},\\frac{\\partial f_{L,5}}{\\partial\\sigma_{1}},\\frac{\\partial f_{L,5}}{\\partial\\mu_{2}},\\frac{\\partial f_{L,5}}{\\partial\\sigma_{2}}} \\\\\n\\end{array} \\right\\rbrack.$$\n\nIn this calculation procedure, the iterative formula in Equation (14) is needed to calculate the inverse matrix of **J**, resulting in a large calculation. Thus, Equation (16) was used in the actual calculation. $$\\left\\{ \\begin{array}{l}\n{\\mathbf{J}(\\mathbf{\\mathbf{\\theta}}^{(k)})\\Delta\\mathbf{\\mathbf{\\theta}}^{(k)} = - \\mathbf{F}_{L}(\\mathbf{\\mathbf{\\theta}}^{(k)})} \\\\\n{\\mathbf{\\mathbf{\\theta}}^{(k + 1)} = \\mathbf{\\mathbf{\\theta}}^{(k)} + \\Delta\\mathbf{\\mathbf{\\theta}}^{(k)}} \\\\\n\\end{array} \\right..$$\n\n4.3. BLG Parameter Estimation Method {#sec4dot3-materials-12-00418}\n------------------------------------\n\nUsing the MLE method in [Section 4.2](#sec4dot2-materials-12-00418){ref-type=\"sec\"}, the parameters of the BLG distribution were estimated for the fatigue life data obtained from the fatigue tests under three stress levels. The estimated parameters are shown in [Table 2](#materials-12-00418-t002){ref-type=\"table\"}. The PDF curves are shown in [Figure 7](#materials-12-00418-f007){ref-type=\"fig\"} (blue curves), and the figure shows that the BLG can more effectively and accurately reflect the double-peak characteristics of the fatigue life distribution of LMD titanium alloys. [Table 2](#materials-12-00418-t002){ref-type=\"table\"} shows that *\u03bc*~1~ and *\u03bc*~2~ both decrease with an increase in the stress level. However, *\u03c3*~1~ and *\u03c3*~2~ show different behaviors: the *\u03c3*~1~ values decrease while the *\u03c3*~2~ values increase with an increase in the stress level.\n\n5. *P*-*S*-*N* Curve of LMD Titanium Alloy {#sec5-materials-12-00418}\n==========================================\n\nA BLG can reflect the fatigue life distribution of LMD titanium alloys more realistically than an LG, so the *P-S-N* curve description method was established based on a BLG model. In the calculation of the *P-S-N* curve, the fatigue life under a specific reliability should be calculated first. Due to cost and time constraints, about only 20 specimens were tested under each stress level. Therefore, the confidence level of the fatigue life under a specific reliability should be considered.\n\n5.1. Fatigue Life Under a Specific Reliability Value {#sec5dot1-materials-12-00418}\n----------------------------------------------------\n\nFor a specific reliability value, *P*, the fatigue life is denoted as $N_{P}$ and satisfies Equation (17). $$P(y \\geq N_{P}) = P,$$ so $$1 - P = \\alpha F_{Y_{1}}\\left( N_{P} \\right) + (1 - \\alpha)F_{Y_{2}}\\left( N_{P} \\right).$$\n\nSeveral $N_{P}$ values under the three stress levels were calculated by using a numerical method according to Equation (18) and plotted as double logarithmic coordinates in [Figure 8](#materials-12-00418-f008){ref-type=\"fig\"}.\n\n5.2. P-S-N Curves {#sec5dot2-materials-12-00418}\n-----------------\n\n[Figure 8](#materials-12-00418-f008){ref-type=\"fig\"} shows that the reliability fatigue lives, $N_{P}$, of three stress levels at the same reliability value, *P*, are approximately linear. Thus, the *P-S-N* curve of LMD titanium alloy can be described by the Basquin equation, Equation (19), where S represents the peak stress, and *C~P~* and *m~P~* are two undetermined parameters. Particularly, when *P* = 50%, Equation (19) simplifies to an *S-N* curve. $$S^{m_{P}}N_{P} = C_{P},$$ and, by taking the logarithm of both sides of Equation (19), the resulting Equation (20) is a linear function. For the specific reliability value *P*, the undetermined parameters *C~P~* and *m~P~* in Equation (20) can be estimated using the linear regression method. $$m_{P}\\lg S + \\lg N_{P} = \\lg C_{P}.$$\n\nUsing the fatigue life data obtained in this paper, the parameters of the *P-S-N* curve under several typical reliabilities were estimated based on the above method and are listed in [Table 3](#materials-12-00418-t003){ref-type=\"table\"} with their correlation coefficients, *r*. [Table 3](#materials-12-00418-t003){ref-type=\"table\"} shows that both parameters, *C~P~* and *m~P~*, decrease with an increase in the reliability value, *P*. The *P-S-N* curves are plotted in [Figure 8](#materials-12-00418-f008){ref-type=\"fig\"}.\n\n6. Discussion {#sec6-materials-12-00418}\n=============\n\n6.1. Mixed Failure Behaviors and BLG {#sec6dot1-materials-12-00418}\n------------------------------------\n\nThe previous studies show that the fatigue behaviors of LMD titanium alloys are different from those of traditional materials, and mixed failure behaviors appear \\[[@B20-materials-12-00418],[@B21-materials-12-00418],[@B22-materials-12-00418],[@B26-materials-12-00418],[@B36-materials-12-00418],[@B46-materials-12-00418],[@B47-materials-12-00418]\\]. The investigations in this paper confirmed this view and obtained a more general conclusion. Through the analysis of an abundance of data, the relationship between the stress level and the proportion of SII in the overall population, which has rarely been examined, was obtained herein.\n\nAs shown in [Figure 3](#materials-12-00418-f003){ref-type=\"fig\"}, mixed failure behaviors appeared under the three stress levels. Thus, in the medium- and high-cycle fatigue regimes, it may be an intrinsic characteristic of LMD titanium alloys under CA stress and independent of the stress level. However, the proportion of SII in the overall population may be related to the stress level. In [Table 4](#materials-12-00418-t004){ref-type=\"table\"}, *n*~SIS~ and *n*~SII~ represent the numbers of SIS and SII, respectively. The value of *n*~SII~/*n* decreases with an increase in the stress level in an approximately linearly manner. This phenomenon may be caused by the combined influences of the microstructure, the types and geometric parameters of the defects, and the local stress state, although further research is needed on the interpretation of this phenomenon.\n\n[Table 4](#materials-12-00418-t004){ref-type=\"table\"} also shows the distribution parameters of SIS and SII estimated by an LG. *\u03bc*~SIS~ and *\u03c3*~SIS~ represent the logarithmic expectation and logarithmic life standard deviation of SIS under an LG distribution, respectively, and *\u03bc*~SII~ and *\u03c3*~SII~ represent those of SII. In [Table 4](#materials-12-00418-t004){ref-type=\"table\"}, it can be found that *\u03bc*~SII~ is larger than *\u03bc*~SIS~ under all three stress levels. This phenomenon is different from the conclusion reported by many investigations that the internal defects such as pores and LOFs have harmful influences on the fatigue life of an AM material \\[[@B39-materials-12-00418],[@B44-materials-12-00418]\\]. [Figure 9](#materials-12-00418-f009){ref-type=\"fig\"}a,b show the FDHs of the fatigue lives of SIS under 800 MPa and SII under 720 MPa, respectively, and the PDF curves of LG~SII~ under 720 MPa and LG~SIS~ under 800 MPa are also plotted in the figures. It can be inferred that, even for a single failure behavior, SIS or SII, the LG cannot reflect the fatigue life distribution well.\n\n[Figure 10](#materials-12-00418-f010){ref-type=\"fig\"} shows the PDF curves of SIS and SII under an LG and for all specimens under a BLG. The two peak positions of the BLG significantly differ from those of the LG~SII~ and LG~SIS~. Therefore, the double-peak characteristic of the LMD titanium alloy fatigue cannot be simply explained by the difference in the CIS. The difference in the CIS is one of the reasons for the large variation in the fatigue life of LMD titanium alloys and the double-peak characteristic, but it is not the only reason.\n\n6.2. Influence of Size and Location of Pore Defects on the Fatigue Life {#sec6dot2-materials-12-00418}\n-----------------------------------------------------------------------\n\nMany studies have reported the influence of the size and location of the internal defects on the fatigue life of AM specimens \\[[@B13-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418],[@B26-materials-12-00418]\\]. A positive or negative correlation between the fatigue life and the size and location of the pore defects appears to exist based on these studies. However, a different conclusion has been obtained in this paper.\n\n[Figure 11](#materials-12-00418-f011){ref-type=\"fig\"} shows the influence of the size and location of a pore on the fatigue life under three stress levels. The left figures show the relationship between the fatigue life and pore area, while the right figures show the relationship between the fatigue life and the distance from the pore to the surface. Under a specific stress level, the data points that have similar areas were marked in the same color and were connected with straight lines to indicate the influence of the distance to the specimen surface in the right figures. Likewise, the data points that share a similar distance to the surface were circled and were connected with a dashed line to indicate the influence of the pore area. The previous conclusion in \\[[@B26-materials-12-00418]\\] was based on the data points under 800 MPa shown in [Figure 11](#materials-12-00418-f011){ref-type=\"fig\"}a. However, some of the data points under the other two stress levels showed different phenomena. In [Figure 11](#materials-12-00418-f011){ref-type=\"fig\"}b, the red data points show that when the areas are similar, a shorter distance does not necessarily lead to a longer fatigue life, and the blue point data in [Figure 11](#materials-12-00418-f011){ref-type=\"fig\"}c show a similar phenomenon. In [Figure 11](#materials-12-00418-f011){ref-type=\"fig\"}c, the data points circled in yellow and pink also show that, when the distances are similar, a larger area does not necessarily lead to a shorter fatigue life. Therefore, there is no significant positive or negative correlation between the fatigue life and the size and location of a pore defect. The conclusions from previous studies about the influences of the size and location of pore defects on the fatigue life \\[[@B13-materials-12-00418],[@B23-materials-12-00418],[@B24-materials-12-00418],[@B25-materials-12-00418],[@B26-materials-12-00418]\\] lack generalizability and thus need to be further investigated.\n\n6.3. Comparisons of the P-S-N Curves Based on a BLG and LG {#sec6dot3-materials-12-00418}\n----------------------------------------------------------\n\nThe logarithmic fatigue life distribution intervals (0.13%, 99.87%) of the BLG under three different stress levels equal the intervals of \u00b13\u03c3 under the LG and are plotted in [Figure 12](#materials-12-00418-f012){ref-type=\"fig\"}. The intervals (0.13%, 99.87%) of the BLG are obviously narrower than those of the LG, indicating that the BLG model can better fit the fatigue life distribution of LMD titanium alloys without excessively amplifying their variation. Therefore, the fatigue properties of LMD titanium alloys can be more accurately described by using the *P-S-N* curve description model based on a BLG.\n\nThe *P-S-N* curve based on an LG was estimated by using a method similar to that in [Section 5](#sec5-materials-12-00418){ref-type=\"sec\"}. [Figure 13](#materials-12-00418-f013){ref-type=\"fig\"} shows the (0.13%, 99.87%) distribution bands of the *P-S-N* curves based on both an LG and BLG. The distribution band of the *P-S-N* curves based on the BLG is significantly narrower than that of the *P-S-N* curves based on the LG.\n\n[Table 5](#materials-12-00418-t005){ref-type=\"table\"} gives the estimated parameters of the *P-S-N* curves based on an LG under several reliability values. The parameters show different laws compared with those of the *P-S-N* curves based on a BLG, as shown in [Table 3](#materials-12-00418-t003){ref-type=\"table\"}. Both *C~P~* and *m~P~* increase with the reliability value, *P*. *P-S-N* curves based on an LG and BLG are plotted in [Figure 14](#materials-12-00418-f014){ref-type=\"fig\"}. [Table 6](#materials-12-00418-t006){ref-type=\"table\"} shows the fatigue life at a typical reliability value under three stress levels, 720, 760, and 800 MPa, which was obtained from the two types of *P-S-N* curves. The figures and tables above indicate that the *P-S-N* curve model based on a BLG can not only more accurately reflect the fatigue life properties but also significantly increase the fatigue life at high-reliability values. The fatigue life at a reliability value of 99.9% is improved by almost two-fold in the high stress level regime (800 MPa). Therefore, describing the fatigue properties of LMD titanium alloys by using the *P-S-N* model based on a BLG can greatly promote the applications of LMD titanium alloys in engineering, especially in load-bearing structures, significantly improve the structural design flexibility, and reduce the structure weight.\n\n7. Conclusions and Recommendations for Further Work {#sec7-materials-12-00418}\n===================================================\n\nIn this paper, fatigue tests of standard smooth Ti-6.5Al-2Zr-1Mo-1V specimens were conducted, and abundant data were obtained. The mixed failure behavior and the influence of internal pores on the fatigue life of LMD Ti-6.5Al-2Zr-1Mo-1V specimens were discussed in detail, and a novel *P-S-N* curve method based on a BLG was established to describe the fatigue properties of LMD titanium alloys. Through the method, the fatigue life at a high reliability value is increased. This result may promote the application of LMD titanium alloys in engineering and can significantly improve the structural design flexibility and reduce the structure weight. From the test results and analyses, the following conclusions can be drawn:\n\n\\(i\\) Mixed failure behaviors of Ti-6.5Al-2Zr-1Mo-1V titanium alloy, with cracks initiating from both internal pores and the surface or subsurface, are observed in the medium- and high-cycle fatigue regimes. These mixed failure behaviors may be an intrinsic characteristic of LMD titanium alloys under CA stress and independent of the stress level. However, the proportion of SII in the overall population may be related to the stress level: the proportion decreases with an increase in the stress level.\n\n\\(ii\\) For the internal pore failure model, there is no significant positive or negative correlation between the fatigue life and the size and location of the pore defects.\n\n\\(iii\\) The BLG is reasonable for the fatigue life description in the medium- and high-cycle fatigue regimes of LMD titanium alloys. The Newton--Raphson algorithm was used to estimate the BLG parameters based on the MLE method, and the Basquin equation can be used to describe the *P-S-N* curve of the LMD titanium alloy.\n\nThere are also some recommendations for further investigations:\n\n\\(i\\) In-depth research on the relationship between the proportion of SII and the stress level is needed. This phenomenon may be caused by the combined influences of the microstructure, the types and geometric parameters of the defects, and the local stress state.\n\n\\(ii\\) The influence of the sizes and locations of the pore defects on the fatigue life requires further research to arrive at a more general conclusion.\n\nConceptualization, X.H.; methodology, X.H. and T.W.; formal analysis, X.H. and Y.L.; investigation, T.W.; writing-original draft preparation, T.W.; writing-review and editing, X.W.\n\nThis research was funded by the National Key Research and Development Program of China (No. 2017YFB1104003), the National Natural Science Foundation of China (No. 11772027), and the Aeronautical Science Foundation of China (No. 28163701002).\n\nThe authors declare no conflict of interest.\n\n![(**a**) Schematic drawing of Ti-6.5Al-2Zr-1Mo-1V blank fabrication by Laser Metal Deposition (LMD) and cylindrical specimen extraction. (**b**) Schematic illustration of the specimen dimensions in mm \\[[@B26-materials-12-00418]\\].](materials-12-00418-g001){#materials-12-00418-f001}\n\n![Microstructure of laser metal deposition (LMD) Ti-6.5Al-2Zr-1Mo-1V, (**a**):50X; (**b**):500X.](materials-12-00418-g002){#materials-12-00418-f002}\n\n![Typical fracture surfaces of specimens with internal crack initiation (SII) and specimens with crack initiating from suface (SIS) under three stress levels.](materials-12-00418-g003){#materials-12-00418-f003}\n\n![Typical fracture photographs of (**a**) SII and (**b**) SIS and fracture surface region details of (**c**) SII and (**d**) SIS; (**e**) typical dimple feature in the fast fracture region (FFR); (**f**) typical fatigue striations and (**g**) secondary cracks and unmelted particles in the crack propagation region (CPR).](materials-12-00418-g004){#materials-12-00418-f004}\n\n![Details of CIS of (**a**,**b**) SIS; (**c**,**d**) SII.](materials-12-00418-g005){#materials-12-00418-f005}\n\n![Fatigue life data under three stress levels.](materials-12-00418-g006){#materials-12-00418-f006}\n\n![Probability density function (PDF) curves and frequency distribution histograms (FDHs) of specimens under (**a**) 720, (**b**) 760, and (**c**) 800 MPa \\[[@B26-materials-12-00418]\\].](materials-12-00418-g007){#materials-12-00418-f007}\n\n![Fatigue lives under typical reliabilities based on a BLG.](materials-12-00418-g008){#materials-12-00418-f008}\n\n![(**a**) FDH of SII under 720 MPa and its PDF; (**b**) FDH of SIS under 800 MPa and its PDF.](materials-12-00418-g009){#materials-12-00418-f009}\n\n![Comparisons of the PDFs of LG~SII~, LG~SIS~, and BLG under (**a**) 720, (**b**) 760, and (**c**) 800 MPa.](materials-12-00418-g010){#materials-12-00418-f010}\n\n![Influence of the size and location of a pore on the fatigue life under (**a**) 800, (**b**) 760, and (**c**) 720 MPa.](materials-12-00418-g011){#materials-12-00418-f011}\n\n###### \n\nVariations in the BLG and LG. (**a**): 720; (**b**):760; (**c**):800 MPa.\n\n![](materials-12-00418-g012a)\n\n![](materials-12-00418-g012b)\n\n![Variation of P-S-N curves based on an LG and BLG.](materials-12-00418-g013){#materials-12-00418-f013}\n\n###### \n\n*P-S-N* curves based on an LG and BLG under reliabilities of (**a**) 50%, (**b**) 80%, (**c**) 90%, (**d**) 95%, (**e**) 99%, (**f**) 99.5%, (**g**) 99.87%, and (**h**) 99.9%.\n\n![](materials-12-00418-g014a)\n\n![](materials-12-00418-g014b)\n\nmaterials-12-00418-t001_Table 1\n\n###### \n\nFatigue life distribution under a lognormal distribution (LG).\n\n  Stress Level/MPa   Number of Specimens   *\u03bc* ~LG~   *\u03c3* ~LG~   *y*~max~/Cycles   *y*~min~/Cycles   $\\mathbf{\\mathbf{N}_{50,\\mathbf{L}\\mathbf{G}}/{Cycles}}$\n  ------------------ --------------------- ---------- ---------- ----------------- ----------------- ----------------------------------------------------------\n  720                15                    5.66       0.316      108,089           1,559,872         457,088\n  760                17                    5.33       0.348      49,758            645,171           213,796\n  800                22                    4.98       0.370      30,715            544,102           95,499\n\nmaterials-12-00418-t002_Table 2\n\n###### \n\nBimodal lognormal distribution (BLG) parameters estimated by maximum likelihood estimation (MLE).\n\n  Stress Level/MPa   *\u03b1*     *\u03bc* ~1~   *\u03c3* ~1~   *\u03bc* ~2~   *\u03c3* ~2~\n  ------------------ ------- --------- --------- --------- ---------\n  720                0.508   5.43      0.235     5.90      0.144\n  760                0.296   4.87      0.143     5.53      0.158\n  800                0.528   4.69      0.138     5.31      0.242\n\nmaterials-12-00418-t003_Table 3\n\n###### \n\nParameters of the *P-S-N* curve based on a BLG.\n\n  *P*      *m~P~*   *C~P~*          *r*\n  -------- -------- --------------- -------\n  50%      17.01    2.07 \u00d7 10^54^   0.965\n  80%      15.62    9.51 \u00d7 10^49^   0.996\n  90%      14.45    3.08 \u00d7 10^46^   0.991\n  95%      13.52    5.56 \u00d7 10^43^   0.990\n  99%      11.88    7.33 \u00d7 10^38^   0.992\n  99.5%    11.29    1.37 \u00d7 10^37^   0.994\n  99.87%   10.29    1.49 \u00d7 10^34^   0.996\n  99.9%    10.11    4.39 \u00d7 10^33^   0.997\n\nmaterials-12-00418-t004_Table 4\n\n###### \n\nLG parameters of SIS and SII under three stress levels.\n\n  *S*/MPa   *n*   *n* ~SIS~   *\u03bc* ~SIS~   *\u03c3* ~SIS~   *n* ~SII~   *\u03bc* ~SII~   *\u03c3* ~SII~   *n*~SII~/*n*\n  --------- ----- ----------- ----------- ----------- ----------- ----------- ----------- --------------\n  720       15    2           5.61        0.819       13          5.67        0.246       0.867\n  760       17    7           5.16        0.414       10          5.46        0.248       0.588\n  800       22    15          4.89        0.317       7           5.18        0.424       0.318\n\nmaterials-12-00418-t005_Table 5\n\n###### \n\nParameters of the P-S-N curve based on an LG.\n\n  *P*      *m~P~*   *C~P~*          *r*\n  -------- -------- --------------- -------\n  50%      14.85    1.28 \u00d7 10^48^   0.999\n  80%      15.85    4.82 \u00d7 10^50^   1.000\n  90%      16.36    1.07 \u00d7 10^52^   1.000\n  95%      16.98    4.65 \u00d7 10^53^   1.000\n  99%      17.60    1.68 \u00d7 10^55^   1.000\n  99.5%    17.90    9.74 \u00d7 10^55^   1.000\n  99.87%   18.41    2.09 \u00d7 10^57^   1.000\n  99.9%    18.51    3.64 \u00d7 10^57^   1.000\n\nmaterials-12-00418-t006_Table 6\n\n###### \n\nFatigue life calculated from P-S-N curves under specific reliabilities (cycles).\n\n  *P*      720 MPa   760 MPa   800 MPa                      \n  -------- --------- --------- --------- --------- -------- --------\n  50%      463,447   523,842   207,587   20,8930   96,917   87,337\n  80%      250,726   219,381   106,365   94,276    47,206   42,306\n  90%      181,635   159,221   75,024    72,912    32,389   34,738\n  95%      142,167   126,241   56,755    60,758    23,747   30,367\n  99%      84,723    84,489    32,689    44,422    13,256   24,165\n  99.5%    70,599    73,384    26,816    39,847    10,705   22,325\n  99.87%   51,095    59,168    19,143    31,282    7,341    20,030\n  99.9%    48,440    55,208    17,816    31,945    6,893    19,020\n"}
+{"text": "Background\n==========\n\nBesides Hong Kong, Macao is the other special administrative region (SAR) belongs to the People's Republic of China. As the only government hospital in Macao, Centro Hospitalar Conde de S\u00e3o Janu\u00e1rio is responsible for all of the peritoneal dialysis (PD) patients. Benefitting from the superior end-stage renal disease (ESRD) funding system provided by Macao SAR, almost all the resident patients on chronic dialysis can get free treatment including dialysis and medication. Different from Hong Kong, there is no PD-first policy in Macao. Nevertheless, PD has been rapidly developed during the past decade in Macao with about 40% of the PD penetration rate in chronic dialysis patients \\[[@B1]\\].\n\nCardiovascular disease (CVD) is the leading cause of mortality and morbidity among patients with ESRD on chronic dialysis \\[[@B2]\\]. According to U.S. Renal Data System and Hong Kong Renal Registry reports, CVD accounts for approximately 40% mortality in dialysis patients and also is the main cause of hospitalization \\[[@B3],[@B4]\\]. In keeping with these facts, data from the Macao Renal Registry show that 36.2% of the mortality in PD patients is attributable to CVD. Currently, many available studies have shown that increased arterial stiffness, which can be examined by pulse wave velocity (PWV), is a powerful and independent predictor of all-cause and cardiovascular mortality in ESRD patients \\[[@B5],[@B6]\\]. Available study \\[[@B7]\\] has clearly shown that the eGFR slope was negatively associated with baPWV in patients with chronic kidney disease stages 3 to 5. Moreover, higher baPWV was independently associated with progression to commencement of dialysis or death.\n\nGenerally, carotid-femoral PWV (cfPWV), calculated on the basis of pulse transit time and the distance travelled by a pulse between carotid artery and femoral artery, is considered as a well-established index of central arterial stiffness \\[[@B8]\\]. However, there are some obvious limitations for routine use cfPWV measurement in clinics \\[[@B9]\\]. Firstly, it is somewhat difficult for clinical operator to use pressure transducers on target arteries. Additionally, some subjects may feel uncomfortable and generally hesitate to exposing inguinal area during the acquisition of femoral pressure waveforms. Recently, the brachial--ankle PWV (baPWV) technique has been developed due to its simple way of measurement by only wrapping the four extremities with blood pressure cuffs. Since most CVD events in PD patients are related with peripheral arteries but not aortic artery, it is important that baPWV reflects both central and peripheral arterial stiffness \\[[@B10]\\]. However, available reports on arterial stiffness in PD patients were limited and cfPWV was used for assessment in all studies \\[[@B11]-[@B14]\\]. The clinical value of baPWV in chronic PD patients has not yet been fully evaluated.\n\nAlthough an increasing PWV of PD patients is reported to be significantly associated with age, malnutrition, pulse pressure, and peritoneal transport status by a single-center study in North China \\[[@B11],[@B12]\\], it has not yet been elucidated whether there are some dissimilar impact factors on arterial stiffness among Macao PD patients who live in South China with inherent variations in different geographic characteristics, primary diseases, and underlying medical funding systems. Therefore, this present study aims to investigate the relationship between baPWV and its associated risk factors among chronic PD patients in Macao.\n\nMethods\n=======\n\nPatients\n--------\n\nWe studied all chronic PD patients (n=107) in our center in this cross-sectional study. All patients were undergoing continuous PD therapy. Exclusion criteria were:(1) treatment time \\<3 months prior to enrollment; (2) age younger than 18 years; (3) presence of clinically overt congestive heart failure (NYHA class III-IV); (4) peritonitis less than 1 month before the study; (5) persistent hypotension despite pharmacological therapy which was defined as systolic blood pressure (BP) \\<90 mmHg or diastolic BP \\<60 mmHg; (6) unwillingness to participate in our study. Finally, total 96 patients were included in the study. On enrollment, demographic and clinical data were collected, including age, gender, height, weight, body mass index (BMI), BP, presence of diabetes mellitus, medication history and etiology of ESRD. The causes of ESRD were as follows: chronic glomerulonephritis (n=38), diabetes mellitus (n=34), essential hypertension or ischemic nephropathy (n=21), obstructive nephropathy (n=1), lupus nephritis (n=1), polycystic kidney disease (n=1). Of total 96 patients, 32 patients were ongoing automatic PD (APD) therapy and 35 patients were using low glucose degradation product peritoneal dialysis fluid (Low-GDP PDF). Charlson comorbidity index (CCI) was scored as described by Beddhu S et al. \\[[@B15]\\]. Hyperlipidemia was defined as diagnosed according to Adult Treatment Panel III criteria or use of statins. This study was approved by the Ethical Committee of Centro Hospitalar Conde de S\u00e3o Janu\u00e1rio and written informed consent was obtained from all participants.\n\nBiochemical analysis\n--------------------\n\nFasting blood samples were collected in the morning. Measurements were performed using routine laboratory methods for such serum parameters as creatinine, calcium, phosphate, albumin, pre-albumin, total cholesterol, triglyceride, low- and high-density lipoprotein, ferritin, hematocrit and hemoglobin. Serum C-reactive protein (CRP) was measured by a high-sensitivity commercial assay. Serum intact parathyroid hormone (iPTH) was measured by Nichols immunoradiometric assay. Blood, urine and dialysate samples were collected in order to calculate weekly Kt/V and creatinine clearance (CCr).\n\nPWV measurement\n---------------\n\nThe baPWV was assessed using VP-1000 vascular profiler (Nippon Colin Ltd., Komaki City, Japan), which allowed on-line pulse wave recording and automatic calculation of PWV. Briefly, baPWV was calculated from the equation: (D1 - D2)/T. D1 is the distance between the heart and ankle, D2 is the distance between the heart nd brachium, and T is the transit between the right brachial arterial wave and right tibial arterial wave. The distances between the sampling points are automatically calculated from the patient's height and are divided by the time interval for the waveform from each measuring point. The baPWV was performed in PD patients with empty abdomen after drainage of dialysate and at least 15 minutes supine rest. Two measurements were performed in each arm, and the average value was used for the analysis. BaPWV is used as arterial stiffness markers due to ease of measurement, reproducibility and validity in previous studies \\[[@B16]\\]. All the PWV measurements were performed by one experienced operator and the intra-observer coefficient of variation was about 1.58--3.36%. Patients were divided into two groups according to mean baPWV value: those above the mean baPWV value were in the high baPWV group, while those below the mean baPWV value were in the low baPWV group.\n\nStatistical analysis\n--------------------\n\nContinuous variables with normal distribution were expressed as means \u00b1 standard deviation, while those without normal distribution were shown as median and interquartile range. Comparisons between the two groups were done by student's *t* test or *\u03c7*^2^ test. Non-parametric data were compared with using Mann--Whitney *U* test. Univariate analysis was done to explore relationships between baPWV and other variables by Pearson correlation test for normally distributed data and Spearman Rank correlation test for non-parametric data (Model 1). Age-adjusted partial correlation test (Model 2) also was performed. Step-wise multiple linear regression analysis was used to assess the independent determinants of increased baPWV. We included all significant variables with respect to the univariate analysis. Variables recognized to present clinical relevance in the current literature but not presenting statistical significance in our study were also included. A two tailed *[p]{.smallcaps}* \\< 0.05 was considered as statistically significant. All statistical analyses were performed using the SPSS statistical software 17.0 for Windows (SPSS, Chicago Davis, IL. USA).\n\nResults\n=======\n\nPatient characteristics and comparisons between two subgroups\n-------------------------------------------------------------\n\nTable [1](#T1){ref-type=\"table\"} shows the demographic and clinical characteristics of enrolled PD patients and the comparisons between two subgroup patients divided according to the mean baPWV value. Compared with low baPWV group patients, high baPWV group patients were significantly older (*p*\\<0.001) and more likely to have a high proportion of female gender (*p*=0.004) as well as previous CVD history (*p*=0.008). Table [2](#T2){ref-type=\"table\"} shows the laboratory parameters of enrolled PD patients and the comparisons between two subgroup patients. Serum albumin level (*p*\\<0.001), pre-albumin level (*p*=0.004) and residual renal CCr (*p*=0.008) were significantly lower but the serum ferritin level (*p*=0.009) was significantly higher in high baPWV group patients compared with low baPWV group patients. However, there were no significant differences in CCI, PD duration, BP, BMI, weekly erythropoietin dosage, rennin-angiotensin system inhibitor or \u03b2-blocker use, APD use, low-GDP PDF use, diabetic status, CRP, serum phosphate and iPTH levels between two subgroups (all *p*\\>0.05).\n\n###### \n\nDemographic and clinical characteristics of study population and comparisons between subgroups\n\n  **Variables**                **Total**      **Low baPWV Group**   **High baPWV Group**   ***p*Value**\n  ------------------------- ---------------- --------------------- ---------------------- --------------\n  Female Gender                  50.0%               37.5%                 67.5%              0.004\n  Age (years)                63.92 \u00b1 14.24       58.34 \u00b1 14.13         71.73 \u00b1 10.26         \\<0.001\n  CCI (score)                5.0 (2.0, 9.0)     4.0 (2.0, 8.0)         5.0 (3.0, 9.0)         0.054\n  PD duration (Months)       44.47 \u00b1 27.89       40.86 \u00b1 26.12         49.53 \u00b1 29.80          0.134\n  SBP (mmHg)                 135.75 \u00b1 20.13     137.07 \u00b1 20.58         133.90 \u00b1 19.59         0.450\n  DBP (mmHg)                 76.21 \u00b1 14.58       78.39 \u00b1 13.32         73.15 \u00b1 15.84          0.082\n  MAP (mmHg)                 96.06 \u00b1 14.81       97.95 \u00b1 13.95         93.40 \u00b1 15.73          0.138\n  PP (mmHg)                  59.54 \u00b1 16.07       58.68 \u00b1 17.08         60.75 \u00b1 14.68          0.536\n  BMI (kg/m^2^)               23.44 \u00b1 3.65       23.26 \u00b1 3.47           23.70 \u00b1 3.92          0.559\n  Dose of EPO (U/kg.week)    136.95 \u00b1 98.18     137.35 \u00b1 90.53        136.39 \u00b1 109.18         0.963\n  RAS inhibitor use              62.5%               67.9%                 55.0%              0.200\n  \u03b2-Blocker use                  13.5%               14.3%                 12.5%              0.801\n  Hyperlipidemia                 40.6%               39.3%                 42.5%              0.752\n  Previous CVD history           26.0%               16.1%                 40.0%              0.008\n  APD use                        33.3%               41.1%                 22.5%              0.057\n  Low-GDP PDF use                36.5%               33.9%                 40.0%              0.542\n  Diabetes mellitus              34.4%               32.1%                 37.5%              0.586\n  baPWV (m/s)                 21.20 \u00b1 5.63       17.49 \u00b1 2.41           26.39 \u00b1 4.66         \\<0.001\n\nData were divided into two groups according to mean baPWV value.\n\nAbbreviations: *baPWV*, brachial-ankle pulse wave velocity; *CCI*, Charlson comorbidity index; *PD*, peritoneal dislysis; *SBP*, systolic blood pressure; *DBP*, diastolic blood pressure; *MAP*, mean arterial pressure; *PP*, pulse pressure; *BMI*, body mass index; *EPO*, erythropoietin; *RAS*, renin-angiotensin system; *APD*, automatic peritoneal dialysis; *GDP*, glucose degradation product; *PDF*, peritoneal dialysis fluid.\n\n###### \n\nLaboratory parameters of study population and comparisons between subgroups\n\n  **Variables**                                      **Total**            **Low baPWV Group**      **High baPWV Group**     ***p*Value**\n  ------------------------------------------ ------------------------- ------------------------- ------------------------- --------------\n  Serum Albumin(g/L)                               38.72 \u00b1 4.85              40.41 \u00b1 3.50              36.35 \u00b1 5.48           \\<0.001\n  Pre-Albumin(g/L)                                 40.16 \u00b1 10.84             42.75 \u00b1 9.61              35.92 \u00b1 11.53           0.004\n  Cholesterol(mmol/L)                               4.82 \u00b1 1.17               4.87 \u00b1 1.35               4.77 \u00b1 0.88            0.682\n  Triglycerides(mmol/L)                          1.92 (0.43, 8.70)         1.79 (0.43, 5.38)         2.08 (0.89, 8.70)         0.314\n  LDL-cholesterol(mmol/L)                           2.72 \u00b1 1.13               2.73 \u00b1 1.23               2.70 \u00b1 0.97            0.909\n  HDL-cholesterol(mmol/L)                           1.22 \u00b1 0.44               1.25 \u00b1 0.50               1.18 \u00b1 0.35            0.451\n  Hemoglobin(g/dL)                                 11.17 \u00b1 1.39              11.32 \u00b1 1.41              10.96 \u00b1 1.35            0.214\n  Hematocrit(%)                                    32.76 \u00b1 4.26              33.19 \u00b1 4.47              32.15 \u00b1 3.91            0.239\n  Ferritin(ug/L)                                  658.89 \u00b1 457.06           549.51 \u00b1 343.48           812.04 \u00b1 548.87          0.009\n  C-reactive protein(mg/L)                      0.41 (0.10, 15.70)         0.37 (0.10, 4.46)        0.59 (0.10, 15.70)         0.104\n  Calcium(mmol/L)                                   2.40 \u00b1 0.25               2.37 \u00b1 0.21               2.44 \u00b1 0.30            0.186\n  Phosphate(mmol/L)                                 1.53 \u00b1 0.43               1.60 \u00b1 0.44               1.44 \u00b1 0.40            0.071\n  Calcium-phosphate product\\[(mmol/L)^2^\\]          3.67 \u00b1 1.12               3.79 \u00b1 1.18               3.49 \u00b1 1.02            0.198\n  intact-PTH(pg/ml)                           342.50 (11.50, 1761.00)   360.75 (11.50, 1486.00)   306.20 (56.45, 1761.00)      0.970\n  Kt/V (total)                                      2.11 \u00b1 0.58               2.13 \u00b1 0.68               2.09 \u00b1 0.40            0.783\n  Kt/V (renal)                                      0.35 \u00b1 0.48               0.43 \u00b1 0.52               0.24 \u00b1 0.41            0.052\n  CCr (total)                                  59.04 (42.23, 138.22)     60.58 (42.23, 138.22)     58.37 (42.51, 117.01)       0.476\n  CCr (renal)                                   6.16 (0.00, 90.00)        14.52 (0.00, 90.00)       0.46 (0.00, 75.49)         0.008\n\nData were divided into two groups according to mean baPWV value.\n\nAbbreviations: *baPWV*, brachial-ankle pulse wave velocity; *LDL*, low-density lipoprotein; *HDL*, high-density lipoprotein; *PTH*, parathyroid hormone; *CCr*, creatinine clearance.\n\nCorrelations between baPWV and related parameters\n-------------------------------------------------\n\nTable [3](#T3){ref-type=\"table\"} shows the correlations between baPWV and related parameters by means of univariate analysis. In brief, Model 1 showed that baPWV was positively associated with patients' age (*r*=0.534, *p*\\<0.001), CCI (*r*=0.350, *p*\\<0.001) and serum ferritin level (*r*=0.340, *p*=0.001). Meanwhile, baPWV negatively correlated with serum albumin (*r*=\u22120.479, *p*\\<0.001), pre-albumin levels (*r*=\u22120.320, *p*=0.003) and residual renal CCr (*r*=\u22120.177, *p*=0.048). However, no significant correlation was found between baPWV and CRP (*r*=0.182, *p*=0.076). Results of partial correlation test from Model 2 by controlling for age showed that the correlation between CCI and baPWV was missing, but a significant correlation was found between baPWV and CRP (*r*=0.462, *p*\\<0.001).\n\n###### \n\nCorrelation between pulse wave velocity and related parameters\n\n  **Variables**                               **Model 1**   **Model 2**           \n  ------------------------------------------ ------------- ------------- -------- ---------\n  Age (years)                                    0.534        \\<0.001       \\-       \\-\n  CCI (score)                                    0.350        \\<0.001     0.059     0.592\n  PD duration (Months)                           0.063         0.539      0.128     0.239\n  SBP (mmHg)                                    \u22120.003         0.976      0.089     0.418\n  DBP (mmHg)                                    \u22120.125         0.226      0.045     0.682\n  MAP (mmHg)                                    \u22120.083         0.420      0.070     0.522\n  PP (mmHg)                                      0.109         0.290      0.073     0.504\n  BMI (kg/m^2^)                                  0.104         0.314      0.009     0.937\n  Serum Albumin(g/L)                            \u22120.479        \\<0.001     \u22120.451   \\<0.001\n  Pre-Albumin(g/L)                              \u22120.320         0.003      \u22120.161    0.139\n  Cholesterol(mmol/L)                           \u22120.064         0.538      \u22120.090    0.408\n  Triglycerides(mmol/L)                          0.022         0.829      \u22120.018    0.870\n  LDL-cholesterol(mmol/L)                        0.047         0.647      0.064     0.560\n  HDL-cholesterol(mmol/L)                        0.003         0.974      \u22120.041    0.711\n  Hemoglobin(g/dL)                              \u22120.106         0.306      \u22120.166    0.126\n  Hematocrit(%)                                 \u22120.078         0.451      \u22120.129    0.235\n  Ferritin(ug/L)                                 0.340         0.001      0.293     0.006\n  C-reactive protein(mg/L)                       0.182         0.076      0.462    \\<0.001\n  Calcium(mmol/L)                                0.071         0.494      \u22120.002    0.984\n  Phosphate(mmol/L)                             \u22120.147         0.154      0.010     0.924\n  Calcium-phosphate product\\[(mmol/L)^2^\\]      \u22120.121         0.241      0.004     0.974\n  intact-PTH(pg/ml)                             \u22120.135         0.189      \u22120.012    0.911\n  Kt/V (total)                                  \u22120.025         0.807      \u22120.093    0.394\n  Kt/V (renal)                                   0.110         0.288      \u22120.294    0.005\n  CCr (total)                                   \u22120.013         0.901      \u22120.276    0.010\n  CCr (renal)                                   \u22120.177         0.048      \u22120.301    0.005\n\nAbbreviations: *CCI*, Charlson comorbidity index; *PD*, peritoneal dialysis; *SBP*, systolic blood pressure; *DBP*, diastolic blood pressure; *MAP*, mean arterial pressure; *PP*, pulse pressure; *BMI*, body mass index; *LDL*, low-density lipoprotein; *HDL*, high-density lipoprotein; *PTH*, parathyroid hormone; *CCr*, creatinine clearance.\n\nMultivariate regression analysis of determinants of baPWV\n---------------------------------------------------------\n\nIn a stepwise multiple linear regression analysis, we employed baPWV value as a dependent variable, while using age, CCI, MAP, calcium-phosphate product, serum albumin level, ferritin level, total cholesterol level, CRP and residual renal CCr as independent variables. Table [4](#T4){ref-type=\"table\"} shows the multivariate linear regression analysis results. Age, serum albumin level, CRP and residual renal CCr were independently associated with baPWV (*p*\\<0.001, =0.015, =0.019 and =0.045, respectively) and together explained 51.1% of the total variance.\n\n###### \n\nMultiple linear regression analysis of factors associated with pulse wave velocity\n\n  \u00a0                 **B**     **SE**    **Standardized coefficient**   ***t*value**   ***P*value**\n  --------------- ---------- --------- ------------------------------ -------------- --------------\n  Constant         1538.293   604.077                \u00a0                    2.547          0.013\n  Age               14.814     3.684               0.375                  4.021         \\<0.001\n  Serum albumin    \u221227.891    11.233               \u22120.240                 \u22122.483         0.015\n  CRP               53.088    22.299               0.210                  2.381          0.019\n  CCr (renal)       \u22123.806     1.874               \u22120.169                 \u22122.031         0.045\n\nAbbreviations: *CRP*, C reactive protein; *CCr*, creatinine clearance.\n\nDiscussion\n==========\n\nChronic PD patients are at a high risk of developing accelerated atherosclerosis, vascular stiffness and CVD incidence secondary to a multitude of traditional and uremia-specific risk factors \\[[@B17]\\]. The present study investigated the arterial stiffness and its associated factors among stable PD patients in Macao. We found that baPWV was independently correlated with patient's age, serum albumin level and residual renal CCr.\n\nArterial stiffness has taken on great importance in the pathophysiology of CVD. In previous studies of general population as well as ESRD patients, increased arterial stiffness assessed by PWV was well established as an independent predictor of all-cause and cardiovascular mortality \\[[@B5],[@B18]\\]. More recently, study by Sipahioglu et al. \\[[@B19]\\] reported that arterial stiffness was an independent risk predictor of mortality and adverse CVD outcome in PD patients. Although epidemiological data show that mortality rate in PD patients continues to decline, long-term survival remains poor. CVD accounts for most deaths, therefore, strategies aimed at reducing modifiable risk factors of CVD have been highlighted for enhancing long-term survival in PD patients \\[[@B20]\\]. Given the multivariate regression analysis findings in this study, patients' age, as one of the independent associated factors with baPWV, cannot be modified. However, attention needs to be placed on other modifiable risk factors which are strongly correlated with increasing arterial stiffness in PD patients, including improvement of malnutrition status and prevention of residual renal function.\n\nHypoalbuminemia is an independent predictor of increased CVD and mortality in dialysis patients, although the exact mechanisms remain unclear. In consistent with the previous study reported by Gu et al. \\[[@B12]\\], we showed that serum albumin level was independently associated with baPWV in PD patients. Several factors may contribute to malnutrition in PD patients such as low protein or energy intakes, psychosocial factors, catabolic effects of acidosis, dialytic losses of protein or amino acids, bio-incompatibility of PDF, and infection \\[[@B21]\\]. Low serum albumin level may induce micro or clinical systemic inflammation which may play an additive role on atherosclerotic vascular disease progression \\[[@B22]\\]. Moreover, growing evidences have suggested that hypoalbuminemia was associated with increased oxidative stress which could accelerate atherosclerosis process in dialysis patients \\[[@B23]\\]. In addition, a number of studies showed that nutritional deficiencies may also play an important role on endothelial dysfunction in ESRD as well as dialysis patients \\[[@B24]\\].\n\nGenerally, various factors associated with PD procedure, such as peritonitis, exit site infection, use of bio-incompatible PDF may promote inflammation. Protein-energy malnutrition with micro-inflammation presenting in a large proportion of chronic PD patients is widely accepted to be a strong risk factor for cardiovascular mortality in this patient group \\[[@B25]\\]. Meanwhile, inflammation has been proposed to be a critical promoter of atherosclerosis, interacting with many pathophysiologic pathways to lead to vascular stiffness. Although the precise link between inflammation and CVD mortality in PD patients remains unknown, endothelial dysfunction has been proposed to play an important role in inflammation-mediated artherosclerosis \\[[@B26]\\]. CRP, which is one of the prototypic markers of inflammation, has been showed to be an important predictor of mortality and CVD death in PD patients \\[[@B27]\\]. Inconsistent with previous single-centre studies in North China \\[[@B11],[@B12]\\], we found that CRP was an independent risk factor for baPWV. Besides the different measurement for arterial stiffness (baPWV vs cfPWV), it may also partly be explained by the higher proportion of DM (35.4% vs 30.0%) and the longer average duration of PD (44.5 months vs 9.5 months) in our studied population.\n\nIt has been clearly established that residual renal function is related to all-cause mortality and risk of cardiovascular death in PD patients \\[[@B28]\\]. The reanalysis of the CANUSA study \\[[@B29]\\] demonstrated that patient survival was linked with the magnitude of residual renal function. Each 5-L per week per 1.73 m^2^ increase in residual glomerular filtration rate corresponded to a 12% decrease in the relative risk of death. Neither peritoneal CCr nor net peritoneal ultrafiltration was associated with patient survival. In the present study, residual renal CCr, instead of peritoneal CCr, was independently associated with baPWV in PD patients. In addition to better solute clearance and volume removal, residual renal function was also associated with decreased levels of circulating inflammatory markers and free radicals, reduced BP, increased phosphorus removal, and reduced left ventricular hypertrophy \\[[@B30]\\]. Taken together, these multiple factors may contribute to improvement on vascular endothelial dysfunction and atherosclerosis.\n\nSeveral limitations of this study must be taken into consideration when interpreting the data. Firstly, its cross-sectional design of the study did not allow us to determine causality. Secondly, although many potential confounding factors have been assessed, the existence of other unrecognized variables should be noted. Finally, due to the technical limitation of baPWV measurement, patients with atrial fibrillation or amputated extremity were excluded. However, these PD patients generally are relevant to high risk of arterial stiffness.\n\nConclusions\n===========\n\nIn conclusion, the presented data clearly showed that arterial stiffness, which was assessed by baPWV, had an independent correlation with age, serum albumin level, CRP level and residual renal CCr in PD patients. Further scaled and prospective studies are warranted to investigate whether amelioration of malnutrition and micro-inflammatory status and preserving residual renal function might prevent the progression of arterial stiffness and reduce the CVD incidence among chronic PD patients.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors' contributions\n======================\n\nKDW collected data, analyzed, interpreted data and drafted the manuscript. LCL collected data and drafted the manuscript. CK, LWI and XJ collected data, analyzed and revised the manuscript. KUI conceived the study, participated in its design and coordination and helped to draft the manuscript and had full access to all the data and assume responsibility for the integrity of the data and the accuracy of the analysis. All authors read and approved the final manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2369/13/143/prepub>\n\nAcknowledgements\n================\n\nThis study was supported by a grant sponsored from Science and Technology Development Fund of Macao SAR (No. 093-2011-A). Nursing staffs of the peritoneal dialysis center at Centro Hospitalar Conde de S\u00e3o Janu\u00e1rio are appreciated for their excellent technical assistance.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nInfectious bronchitis virus (IBV), which is a gamma corona virus classified in the family *Coronaviridae*, causes respiratory signs, egg-drops, nephritis, and proventriculitis in domestic fowl, resulting in serious economic losses \\[[@CR7]\\]. The single-stranded positive-sense RNA genome is approximately 27\u00a0kb, with eight subgenomic mRNAs (sg-mRNAs) that are transcribed under the regulation of transcription regulatory sequences (TRSs). The leader TRS is located at the 5'-end of the viral genomic RNA and 8 body TRSs are located upstream of the coding regions of structural and nonstructural genes. The homology between the leader and body TRSs may affect the copy number of subgenomic mRNA transcripts and the level of gene expression \\[[@CR4], [@CR44]\\]. The viral genome encodes RNA polymerase/transcriptase (1ab), spike (S), envelope (E), membrane (M), nonstructural proteins (nsps), 3a, 3b, 4b, 4c, 5a, 5b and 6b, and nucleocapsid (N) \\[[@CR7], [@CR35]\\]. Large 1a and 1ab proteins are cleaved by viral proteases into approximately 15 nsps, which are involved in virus replication and pathogenicity \\[[@CR2], [@CR13]\\]. Spike is cleaved into S1 and S2 by proteases, and most epitopes for virus neutralization are within the hypervariable regions of S1 and S2 \\[[@CR19], [@CR20]\\]. Six epitopes, D (24-60), E (132-149), C/A/B (291-398) and F (497-543), in S1 and 1 epitope, G (548-574), in S2 have already been defined \\[[@CR19]\\]. Different IBV serotypes do not confer protective effects against each other. E and M proteins are important for virion assembly and are membrane-anchored glycoproteins \\[[@CR9]\\]. The short ectodomain of the M protein protrudes out from the viral envelope and contains a protective epitope \\[[@CR14]\\]. Viral variation and diversity are attributed to the long RNA genome and frequent nucleotide substitutions, deletions, insertions, and recombination events occur due to the poor proofreading capability of RNA polymerases \\[[@CR16], [@CR27], [@CR31], [@CR43]\\].\n\nIBVs are distributed worldwide in poultry, and most countries have their own indigenous variants and common genotypes \\[[@CR38]\\]. According to a phylogenetic analysis based on the short variable region of S1, several indigenous and common IBV genotypes (e.g., K-I, KM91-like, New cluster 1, QX-like) have co-circulated, and in the past recombination events are suspected to have occurred in Korea \\[[@CR16], [@CR24], [@CR25]\\]. Nephropathogenic KM91-like viruses were first isolated in 1991 and became prevalent in the field. To prevent the economic losses caused by KM91-like virus infection, oil emulsion-inactivated and attenuated live vaccines have been used in Korea \\[[@CR25]\\]. QX-like IBVs are rampant throughout Asia and Europe. They were initially isolated in 1997 in China and might have been introduced to Korea in 2002--2003 \\[[@CR1], [@CR8], [@CR18], [@CR24], [@CR25]\\]. Subsequently, novel recombinant viruses (New Cluster 1, NC1), which were suspected to have originated through recombination between KM91-like and QX-like viruses, emerged in Korea in 2005 and became prevalent between 2007 and 2010 \\[[@CR27]\\]. Recent recurrences of QX-like viruses were observed in 2008 and 2011, and they have become prevalent in Korea \\[[@CR31]\\]. Phylogenetic and computational recombination analyses using genome sequence data have shed light on the occurrence of complex recombination events in the genomes of IBVs and the precise donor and acceptor relationships between concurrent field viruses during recombination \\[[@CR6], [@CR11], [@CR15], [@CR16], [@CR35], [@CR42]\\]. However, the minimum essential core replication and pathogenicity-related 1ab gene is not typically included in phylogenetic analyses. In the present study, we analyzed coding and non-coding IBV sequences and conducted computational recombination analyses as well as phylogenetic analyses with S1, 1ab, and full genome sequences of two early, three recombinant and two recurrent QX-like viruses to understand the molecular epidemiology and molecular evolution of QX-like viruses at the genome level. Consequently, we clarified the molecular epidemiology of infectious bronchitis caused by QX-like viruses in Korea. Furthermore, we classified IBVs into three distinct lineages, including North America-Europe, China/Taiwan, and China, by using a 1ab gene-based phylogenetic analysis. We found that the early Korean and European QX-like viruses were generated by recombination between the North America-Europe and China lineages.\n\nMaterials and methods {#Sec2}\n=====================\n\nVirus isolation and propagation {#Sec3}\n-------------------------------\n\nQIA-03342, QIA-KR/D79/05, and QIA-Q43/06 were isolated in 2003, 2005, and 2006, respectively, and form a cluster with the QX strain according to a partial S1 gene-based phylogenetic tree \\[[@CR24]\\]. SNU9106 and SNU10043 as well as SNU8065 and SNU11045 have been classified as NC1 and recurrent QX-like viruses, respectively \\[[@CR31]\\]. All isolates were inoculated into 10-day-old SPF embryonated chicken eggs (ECE; Valo BioMedia, Adel, IA, USA) via the allantoic cavity route and incubated for 48\u00a0h. Next, the eggs were chilled at 4\u00b0C overnight. The allantoic fluid was harvested, and the supernatant was stored at -70\u00b0C after centrifugation at 3,000\u00a0rpm for 10\u00a0min.\n\nRNA extraction and RT-PCR {#Sec4}\n-------------------------\n\nViral genomic RNA was extracted from infectious allantoic fluid using the Viral Gene-Spin Kit (iNtRON Biotechnology, Seongnam, Korea), and RT-PCR was performed using a One-step RT-PCR Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. The cDNA was synthesized at 50\u00b0C for 30\u00a0min and then heated to 95\u00b0C for 15\u00a0min to inactivate reverse transcriptase. The PCR conditions were as follows: 40 repetitions of denaturation at 94\u00b0C for 30\u00a0s, annealing at 50\u00b0C for 30\u00a0s, extension at 72\u00b0C for 2\u00a0min, and a final extension at 72\u00b0C for 5\u00a0min. The primer sets for the amplification of overlapping fragments of the IBV genomes and for sequencing are listed in Supplementary data 1.\n\nSequencing and genome sequence analysis {#Sec5}\n---------------------------------------\n\nThe amplicons were purified using the MEGA-quick-spinTM Total Fragment DNA Purification Kit (iNtRON Biotechnology) and sequenced with PCR primers using an ABI3711 automatic sequencer (Macrogen Co., Seoul, Korea). The overlapping gene fragments were assembled to obtain a single complete genome sequence using ChromasPro version 1.5 (Technelysium Pty Ltd., Brisbane, Australia). Nucleotide and amino acid identity estimates and amino acid translations were obtained using BioEdit (ver. 5.0.9.1.). All sequences were submitted to the GenBank database (accession numbers, KU900738--KU900744).\n\nPhylogenetic analyses based on S1, 1ab, and the complete genomes sequences were conducted using MEGA (ver. 6.0.6) and the neighbor-joining method with Tamura-Nei distances and 1,000 bootstrap replicates. The GenBank accession numbers of IBV isolates included in the analysis are presented in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}.\n\nComputational recombination analysis {#Sec6}\n------------------------------------\n\nPutative recombination events between the QIA-03342, QIA-KR/D79/05, QIA-Q43, SNU-9106, SNU-10043, KM91, and YX10 strains \\[[@CR40]\\] were analyzed with RDP4 (ver. 4.14; neighbor-joining, Kimura distance correction; P\u00a0\\<\u00a00.05) \\[[@CR30]\\]. Seven methods in RDP 4.14, including RDP, GENECONV, Bootscan, MaxChi, Chimaera, SiScan, and 3Seq were used to detect recombination events. We inputted genomic nucleotide sequences of each Korean QX-like virus with the nucleotides of the KM91 and YX10 strains. The recombination detection methods sequentially tested every combination of the three sequences for evidence that one of the three sequences was a recombinant and the other two were its parents.\n\nResults {#Sec7}\n=======\n\nCharacterization of the genome sequences {#Sec8}\n----------------------------------------\n\nThe genome sizes of QIA03342, QIA-KR/D79/05, QIA-Q43, SNU9106, SNU10043, SNU8065, and SNU11045 were 27684, 27682, 27675, 27684, 27683, 27679, and 27679, respectively, and the GC contents ranged from 38.08--38.23% (Table\u00a0[1](#Tab1){ref-type=\"table\"}). Most of the open reading frames overlapped, and the ribosomal frameshifting/slippery sequences were conserved in all of the QX-like viruses. The YX10 strain (CK/CH/Zhejiang/06/10), a recombinant nephropathogenic QX-like virus isolated in the Zhejiang province of China in 2010, was selected due to its high nucleotide identity with the recurrent QX-like viruses \\[[@CR40]\\]. The whole genomes of QIA-03342, QIA-D79/05, QIA-Q43/06, SNU9106, and SNU10043 had higher nucleotide identities for KM91 (95--96%) than for YX10 (88%) but had identities to SNU8065 and SNU11045 of 87% and 98%, respectively (Table\u00a0[2](#Tab2){ref-type=\"table\"}). The 5\u2032- and 3\u2032-UTRs of QIA-03342, QIA-D79/05, QIA-Q43/06, SNU9106, and SNU10043 were more similar to KM91 (98--99% and 95--98%, respectively) than to YX10 (94--95% and 80--83%, respectively), and SNU8065 and SNU11045 were more similar to YX10 (99% and 95--96%, respectively) than to KM91 (94--95% and 84--85%, respectively) (Table\u00a0[2](#Tab2){ref-type=\"table\"}).Table\u00a01Genome size, GC content, and IBV proteins with differing amino acid lengths for the examined QX-like virusesQIA-03342QIA-KR/D79/05QIA-Q43SNU9106SNU-10043SNU-8065SNU11045Full length27,68427,68227,67527,68427,68327,67927,679G\u00a0+\u00a0C content38.08%38.2%38.08%38.23%38.13%38.16%38.14%S1,1651,1651,1631,1651,1651,1651,1653b65656565656262E109109109109109108108M226226226226226225225 Table\u00a02Nucleotide and amino acid identities (%) of QX-like viruses, when compared to KM91 and YX10Nucleotide and amino acid identity (nt/aa; %)QIA-\\\n03342QIA-D79/05QIA-Q43/06SNU-\\\n9106SNU-10043SNU-\\\n8065SNU-11045GenomeKM9195959696958787YX10888888888898985\u2032-UTRKM9199989999999594YX10959495949599991abKM9197/9796/9796/9796/9796/9787/9187/91YX1087/9187/9187/9187/9187/9198/9998/99S1KM9185/8685/8590/9092/9091/9185/8585/85YX1096/9595/9491/8988/8888/8998/9898/98S2KM9192/9494/9498/9798/9793/9393/9593/95YX1080/7180/7180/7179/7178/7099/10099/1003aKM9163/5263/5263/5263/5262/5072/6972/69YX1074/6374/6374/6375/6576/66100/100100/1003bKM9180/7280/7280/7280/7281/7474/6774/67YX1086/8587/8586/8585/8386/8499/9999/99EKM9198/10097/9899/10097/9694/9586/8586/85YX1086/8587/8586/8585/8386/8499/9999/99MKM9194/9696/9894/9595/9698/10089/9389/93YX1090/9491/9490/9389/9289/9499/9999/994bKM9185/8293/9485/8294/9385/8189/8789/87YX1085/8490/9185/8489/8785/83100/100100/1004cKM9188/7787/74887787/7288/7771/537153YX1075/5874/5374/5874/5875/58100/100100/1005aKM9196/9797/9895/9595/9294/9483/8383/83YX1083/8583/8581/8182/8180/8098/9898/985bKM9197/9399/9898/9598/9499/9891/8891/88YX1093/9192/9092/9092/9093/9193/8892/88NKM9197/9797/9797/9697/9797/9789/9289/92YX1088/9288/9288/9288/9288/9288/9288/926bKM9197/9696/9394/9297/95100/10084/7885/77YX1084/8183/7881/7784/8084/7896/9595/933\u2032-UTRKM9198979695988485YX1083828080829596\n\nComparison of open reading frames (orfs) {#Sec9}\n----------------------------------------\n\nExcept for the S, 3b, E, and M proteins, the amino acid lengths of the proteins did not differ among strains (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The nucleotide and amino acid sequences of each virus were compared to those of the KM91 and YX10 strains (Table\u00a0[2](#Tab2){ref-type=\"table\"}). The 1ab, S2, E, M, 5a, 5b, N, and 6b genes of QIA-03342, QIA-D79/05, QIA-Q43/06, SNU9106, and SNU10043 were more similar to KM91 (nucleotide (nt) 92--99%, amino acid (aa) 92--100%) than to YX10 (nt 78--93%, aa 70--94%). S1 of QIA-03342 and QIA-D79/05 had higher identities to YX10 (nt 96% and 95%, aa 95% and 94%, respectively) than to KM91 (nt 85% and 85%, aa 86% and 85%, respectively), but QIA-Q43/06, SNU9106, and SNU10043 with respect to KM91 and YX10 were more similar (nt 90--92%, aa 90--91% and nt 88--91, aa 88--89%, respectively). The 3a, 3b, 4b, and 4c genes of QIA-03342, QIA-D79/05, QIA-Q43/06, SNU9106, and SNU10043 showed relatively low nucleotide and amino acid identities when compared to KM91 (nt 62--94%, aa 50--94%) and YX10 (nt 74--90%, aa 53--91%), but they showed relatively high identities when compared to other viruses, e.g., 3a (Connecticut/KF696629, M41/FJ904713, TCoV/EU022526, CK/CH/LDL/110931/KJ425485, 2994/02/GU386375; 94--95%), 3b (Delaware/GU393332, TCoV, and Gray/GU393334; 93--97%), 4b (CK/CH/LTJ/95I/EF602448; 98--100%), and 4c (CK/CH/LTJ/95I, 93--97%). The 1ab, S1, S2, 3a, 3b, E, M, and 5a genes of SNU8065 and SNU11045 were more similar to YX10 (nt 98--100%, aa 98--100%) than KM91 (nt 72--93%, aa 67--95%), but 5b and N had similarly low identities to KM91 and YX10 (nt 91% and 89%, aa 88% and 92%; nt 92--93% and 88%, aa 88% and 92%, respectively). SNU8065 and SNU11045 showed relatively high nucleotide identities when compared to other viruses; specifically for 5b (CK/CH/LHLJ/141105/KP790145 and CK/CH/IBYZ/2011/KF663561, 99%) and N (SD10/KF625030, CK/CH/SD09/006/HM230751, and GX-NN120084/KF996278, 99%). The amino acid identities of nsps were examined with respect to the nsps of KM91 and YX10 (Table\u00a0[3](#Tab3){ref-type=\"table\"}). The identities were different among nsps and were higher for nsp 2, 3, 4, 5, and 6 than for the other nsps, which indicated that the former were more variable.Table\u00a03Nonstructural protein (nsp) amino acid identities between the QX-like viruses and KM91 and YX10NSPPutative function^b^Protease^\\*^Size (aa)Amino acid identity with respect to KM91/YX10 (%)Percent differences^b^QIA-03342QIA-D79/05QIA-Q43/06SNU-9106SNU-10043SNU-8065SNU-110452UnknownPLP67395/8596/8696/8695/8695/8686/9886/989--12%3Papain-like viral proteasePLP1,59495/8695/8694/8594/8594/8584/9884/989--14%4UnknownPLP/3CLpro51498/8898/8998/8998/8897/8887/10087/1009--13%5Coronavirus endopeptidase C303CLpro30798/9198/9098/9197/9098/9089/9990/997--10%6Hydrophobic domain3CLpro29399/8699/8797/8698/8798/8687/9987/9911--13%7nsp7 superfamily3CLpro83100/9899/9699/9699/9696/9496/9996/992--4%8nsp8 superfamily3CLpro21098/9499/9399/9398/9397/9294/9894/984--6%9nsp9 superfamily3CLpro111100/100100/100100/100100/100100/10098/98100/1000--2%10nsp10 superfamily, RNA synthesis3CLpro14597/9499/9697/9498/9499/9595/10094/993--5%12RNA-dependent RNA polymerase3CLpro94099/9699/9698/9699/9698/9696/9996/992--3%13Viral RNA helicase3CLpro60098/9798/9799/9799/9899/9897/10097/991--3%14nsp11 superfamily; exoribonuclease3CLpro52199/9799/9799/9799/9798/9797/9997/992%15Nidoviral uridylate-specific endoribonuclease3CLpro33897/9597/9598/9596/9496/9394/9694/962--3%1623S rRNA methylase3CLpro30299/9699/9698/9798/9698/9696/9996/991--3%^\\*^PLP, papain-like protease; 3CLpro, 3C-like protease; ^b^ difference between identities with respect to KM91 and YX10\n\nComparison of regulatory sequences {#Sec10}\n----------------------------------\n\nWe examined the TRSs of each IBV and compared nucleotide sequences between the leader and body TRSs (Table\u00a0[4](#Tab4){ref-type=\"table\"}). The leader TRS sequences of all viruses were identical (CTTAACAA). The leader TRS sequence of each virus had 0--3 nucleotide mismatches with the body TRS sequences, and the body TRS sequences of the 4 and 6 sg-mRNAs were identical to the leader TRS sequence.Table\u00a04Comparison of the transcription regulatory sequences (TRSs) in the full IBV genomesSg mRNATRSKM91Sequence and location of TRSQIA-03342QIA-KR/D79/05QIAQ-43/06SNU9106SNU10043YX10SNU8065SNU11045LeaderCTTAACAA(57-64)CTTAACAA(57-64)CTTAACAA(57-64)CTTAACAA(57-64)CTTAACAA(57-64)CTTAACAA((57-64)CTTAACAA(57-64)CTTAACAA(57-64)CTTAACAA(57-64)1(S)BodyAAG\\-\\-\\-\\--(20,311-20,318)AGG\\-\\-\\-\\--(20,314-20,321)AGG\\-\\-\\-\\--(20,314-20,321)AAG\\-\\-\\-\\--(20315-20322)T--T\\-\\-\\-\\--(20,314-20,321)AGG\\-\\-\\-\\--(20316-20323)\\--G\\-\\-\\-\\--(20,311-20,318)\\--G\\-\\-\\-\\--(20315-20322)\\--G\\-\\-\\-\\--(20,314-20,321)2(3ab)Body\\--G\\-\\-\\-\\--(23,831-23,838)\\--G\\-\\-\\-\\--(23840-23847)\\--G\\-\\-\\-\\--(23840-23847)\\--G\\-\\-\\-\\--(23835-23842)\\--G\\-\\-\\-\\--(23840-23847)\\--G\\-\\-\\-\\--(20342-23849)\\-\\-\\-\\-\\-\\-\\--(23837-23844)\\--G\\-\\-\\-\\--(23835-23842)\\--G\\-\\-\\-\\--(23,838-23,845)3(E)Body\\--C\\-\\-\\-\\--(24,035-24,042)T-C\\-\\-\\-\\--(24088-24094)T-C\\-\\-\\-\\--(24087-24094)T-C\\-\\-\\-\\--(24082-24089)T-C\\-\\-\\-\\--(24087-24094)T-C\\-\\-\\-\\--(24089-24096)TAC\\-\\-\\-\\--(24084-24091)TAC\\-\\-\\-\\--(24089-24096)TAC\\-\\-\\-\\--(24,085-24,092)4(M)Body\\-\\-\\-\\-\\-\\-\\--(24,377-24,384)\\-\\-\\-\\-\\-\\-\\--(24,435-24,442)\\-\\-\\-\\-\\-\\-\\--(24,435-24,442)\\-\\-\\-\\-\\-\\-\\--(24430-24437)\\-\\-\\-\\-\\-\\-\\--(24,435-24,442)\\-\\-\\-\\-\\-\\-\\--(24437-24444)\\-\\-\\-\\-\\-\\-\\--(24429-24436)\\-\\-\\-\\-\\-\\-\\--(24434-24441)\\-\\-\\-\\-\\-\\-\\--(24,435-24,442)IR(4bc)BodyG-C\\-\\-\\-\\--(24850-24857)G-C\\-\\-\\-\\--(24905-24912)G-C\\-\\-\\-\\--(24905-24912G-C\\-\\-\\-\\--(24899-24906)G-C\\-\\-\\-\\--(24934-24941)G-C\\-\\-\\-\\--(24905-24912)G-C\\-\\-\\-\\--(24896-24903)G-C\\-\\-\\-\\--(24896-24903)G-C\\-\\-\\-\\--(24896-24903)5(5ab)Body\\-\\-\\-\\-\\-\\-\\--(25,485-25,492)\\-\\-\\-\\-\\-\\-\\--(25543-25550)\\-\\-\\-\\-\\-\\-\\--(25540-25547)\\-\\-\\-\\-\\-\\-\\--(25538-25545)\\-\\-\\-\\--TAA(25543-25550)\\-\\-\\-\\-\\-\\-\\--(25545-25552)\\-\\-\\-\\--G-G(25534-25541)\\-\\-\\-\\--G-G(25539-25546)\\-\\-\\-\\--G-G(25,535-25,542)6(N)Body\\-\\-\\-\\-\\-\\-\\--(25,786-25,793)\\-\\-\\-\\-\\-\\-\\--(25844-25851)\\-\\-\\-\\-\\-\\-\\--(25841-25848)\\-\\-\\-\\-\\-\\-\\--(25839-25846)\\-\\-\\-\\-\\-\\-\\--(25844-25851)\\-\\-\\-\\-\\-\\-\\--(25846-25854)\\-\\-\\-\\-\\-\\-\\--(25839-25846)\\-\\-\\-\\-\\-\\-\\--(25844-25851)\\-\\-\\-\\-\\-\\-\\--(25,840-25,847)6bBodyAGG\\-\\-\\-\\--(27,004-27,011)AAG\\-\\-\\-\\--(27,059-27,066)AGG\\-\\-\\-\\--(27,056-27063)AAG\\-\\-\\-\\--(27,053-27060)AGG\\-\\-\\-\\--(27,059-27,066)AGG\\-\\-\\-\\--(27059-27066)--GG\\-\\-\\-\\--(27,054-27,061)AGG\\-\\-\\-\\--(27,054-27061)AGG\\-\\-\\-\\--(27,054-27,061)\n\nComputational recombination analysis {#Sec11}\n------------------------------------\n\nRecombination events were detected in the S1 gene of QIA-03342, QIA-KR/79/05, QIA-Q43/06, SNU9106, and SNU10043, with high significance (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The genomic backgrounds of QIA-03342, QIA-KR/79/05, QIA-Q43/06, SNU9106, and SNU10043 were KM91, and the partial or complete S1 gene might have been acquired from QX-like viruses. However, the patterns of S1 recombination differed among the viruses. Additional recombination events were suspected in S2 (QIA-03342 and SNU10043), 3a/3b (all), and 4b/4c (QIA-03342), but statistical significance was not obtained for these analyses.Fig.\u00a01Using the full genome sequences computational recombination analyses were performed with RDP (ver. 4.14; neighbor-joining, Kimura-distance measure method; P\u00a0\\<\u00a00.05). Seven algorithms in RDP 4.14, including RDP, GENECONV, Bootscan, MaxChi, Chimaera, SiScan and 3Seq were used to evaluate the recombination events. P values for the Bootscan are represented. The locations of the S1 and S2 genes are represented by the red boxes\n\nPhylogenetic analysis {#Sec12}\n---------------------\n\nBased on a phylogenetic analysis using the nucleotide sequences of S1, QIA-03342, QIA-KR/D79/05, SNU8065, and SNU11045 these isolates were classified as QX-like while QIA-Q43/06, SNU9106, and SNU10043 were classified as NC1 sub-genotypes, as previously reported \\[[@CR24], [@CR31]\\]. Interestingly, the European QX-like viruses ITA/90254/2005 and CK/SWE/0658946/10 clustered with the early QX-like viruses QIA-03342 and QIA-KR/D79/05 (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a) \\[[@CR1]\\]. QIA-03342, QIA-KR/D79/05, QIA-Q43/06, SNU-9106, and SNU-10043 clustered with KM91, and SNU8065 and SNU-11045 clustered with YX10 in the phylogenetic tree constructed with full genome sequences. ITA/90254/2005 and CK/SWE/0658946/10 clustered together, but they did not form a cluster with the Belgian nephropathogenic reference strain B1648 (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}b). The Europe and American strains formed a cluster, but the European QX-like viruses clustered with strains from China, including QX-like viruses. To remove the effects of highly variable and frequently recombining genes, i.e., S, 3a/3b, and 4b/4c, on the phylogeny, we constructed a phylogenetic tree using the minimum essential core replication gene 1ab (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}c). The topology and clustering pattern of the 1ab-based tree were similar but not identical to those of the whole genome tree. Based on the clustering patterns, we classified IBVs into North America-Europe, China/Taiwan, and China lineages. Unexpectedly, KM91, the early QX-like viruses, and the NC1 viruses formed a cluster with the North America-Europe lineage viruses. CK/SWE/0658946/10 clustered with the French strain 4/91, but ITA/90254 clustered with B1648. Thus, viruses with the 1ab genes of the North America-Europe lineage might have acquired Spike genes directly or indirectly from QX-like viruses in the China lineage. SNU-8065 and SNU-11045 clustered with viruses from the China lineage, including LX4, DY07, and YX10. Ck/CH/LSD/11235, GX-YL9, and Ck/CH/LGD/120723 clustered with viruses from the China lineage but did not form a cluster with QX-like viruses in the S1 tree. Thus, they possessed QX-like 1ab but non-QX-like spike genes.Fig.\u00a02Phylogenetic trees based on: 1.5\u00a0kb nucleotide sequences of S1 (a), full genome (b), and 1ab genes (c), were constructed with the neighbor-joining method using MEGA 5.05. The bootstrap values were determined from 1000 replicates of the original data. The branch number represents the percentage of times that the branch appeared in the tree. Bootstrap values greater than 90% are shown. The p-distance is indicated by the bar at the bottom of the figure. The QX-like viruses characterized in the present study are marked with closed circles\n\nMolecular characterization of structural proteins {#Sec13}\n-------------------------------------------------\n\nThe amino acid sequences of cleavage sites between S1, S2, and the fusion peptide were compared (Table\u00a0[5](#Tab5){ref-type=\"table\"}) \\[[@CR41]\\]. The NC1 viruses QIA-043/06, SNU9106, and SNU10043 possessed the same amino acid sequence, RRFRR/S, as KM91 at the cleavage site between S1 and S2, but the other viruses had HRRRR/S. The early QX-like and NC1 viruses possessed PSGR/S at the cleavage site of the fusion peptide except Q43/06 (PRRR/S); however, the recurrent QX-like viruses had PRGR/S. The partial or whole amino acid sequences of B cell epitopes, D (24--61), E (132--149), C/A/B (291--398), and F (497--543) in S1 and G (548--574) in S2 were compared, as summarized in Fig.\u00a0[3](#Fig3){ref-type=\"fig\"} \\[[@CR19], [@CR20]\\]. The early QX-like viruses QIA-03342 and QIA-KR/D79/05 exhibited amino acids differences in all of the epitopes, but QIA-043/06, SNU9106, and SNU10043 acquired similar amino acids to KM91 for the C/A/B, F, and G epitopes. However, the C/A/B epitopes of SNU9106 and SNU10043 were more similar at the amino acid level to KM91 than QIA-043/06 was, and they acquired additional amino acid changes in the F and G epitopes. The recurrent QX-like viruses SNU8065 and SNU11045 had different amino acid sequences in all of the epitopes. Similar to KM91, QIA-Q43 had two amino acid deletions, 24G and 25N, in epitope D.Table\u00a05Comparison of functional motifs in the spike (S), envelope (E) and membrane (M) proteinsSEMProteolytic cleavage siteER retention signalVariable amino acid residue in ectodomainS1/S2S2'(fusion peptide)561112131617KM91~534~RRFRR/S~539~^a^~688~ PSGR/S~692~^a^~103~RHGKLHS~109~^b^GESTQAEQIA-03342~536~HRRRR/S~541690~ PSGR/S~694~-^c^-N-EVQQIA-KR/D79/05D\\-\\-\\-\\-\\--QIA-Q43/06~534~RRFRR/S~539~^a^~688~ PRRR/S~692~^a^\\--N-EVQSNU9106~536~RRFRR/S~541690~ PSGR/S~694~\\-\\-\\-\\-\\-\\--SNU10043\\-\\-\\-\\-\\-\\--YX10~536~HRRRR/S~541690~ PRGR/S~694103~RDKLHP~108~E.^d^DSEILSNU8065E.DSEILSNU11045E.DSEIL^a^The different location numbers of KM91 and QIA-Q43/06 are due to the 24G and 25N deletion^b^The different location number is due to an insertion at 105G^c^Same amino acid sequence as KM91^d^Non-comparable amino acid residue due to 6E insertion Fig.\u00a03Comparison of the amino acid sequences of each B cell epitope in the IBV spike proteins. Identical amino acids are represented with dots, while deleted amino acids are represented with hyphens. The amino acids unique to KM91 are represented with red letters, and the amino acids unique to each NC1 virus are represented with blue letters. The putative N-linked glycosylation sites are underlined\n\nThe endoplasmic reticulum (ER) retention signal E protein localizes to the ER \\[[@CR26]\\]. We compared the amino acid sequences of QX-like viruses with KM91 and YX10 (Table\u00a0[5](#Tab5){ref-type=\"table\"}). The early QX-like and NC1 viruses were one amino acid longer and had a different ER retention signal motif, RHGKLHS (105G insertion), when compared to the recurrent QX-like viruses, RDKLHP. The ectodomain amino acid sequences (1-23 amino acid residues) of M proteins were also compared (summarized in Table\u00a0[5](#Tab5){ref-type=\"table\"}). Compared to KM91 the early QX-like and NC1 viruses shared a 6E insertion and were one amino acid longer than the recurrent QX-like and YX10 viruses. The amino acid sequences of the recurrent QX-like viruses were more different from KM91 than the early QX-like and NC1 viruses.\n\nDiscussion {#Sec14}\n==========\n\nIn contrast to our expectations, our results revealed that early QX-like viruses were recombinant KM91-like and QX-like viruses. The absence of pure QX-like viruses before the appearance of the early QX-like viruses suggests that recombinant viruses were introduced from other countries. A computational recombination analysis did not support recombination events in the S2, 3a, 3b, 4b, and 4c genes of the early QX-like viruses. However, their low nucleotide and amino acid identities, to both KM91 and YX10, and relatively high identities to other viruses in the North America group suggest potential recombinant events with other unknown viruses. The NC1 viruses may be products of recombination between the early QX-like and KM91-like viruses, as previously reported \\[[@CR27]\\]. Furthermore, the low nucleotide and amino acid identities of the recurrent QX-like virus 5b and N genes to both KM91 and YX10, but high identities to other viruses in China also suggest frequent recombinant events in the 5b and N genes. A high number of recombination breakpoints were identified in the 1a region, but the probabilities for these hotspots were not significant \\[[@CR39]\\]. However, recombinant events among IBVs may occur more frequently in the latter fourth of the viral genome than the remainder of the genome containing the 1ab gene \\[[@CR22]\\]. Thus, phylogenetic analyses based on 1ab may provide more information on the evolution of IBVs. According to the results of the present study, IBVs were classified into the North America-Europe, China/Taiwan, and China genotypes. The clustering of KM91, the early QX-like viruses, and the NC1 viruses with the viruses of the North America-Europe lineage was unexpected, but recent reports on the spread of H5N8 highly pathogenic avian influenza viruses from Korea to North America and Europe as well as their reintroduction to Korea by migratory birds may provide a hint \\[[@CR21], [@CR23]\\]. Siberia and Beringia are breeding sites for migratory birds from Eurasia and parts of North America, China, Korea and Japan are wintering sites \\[[@CR23]\\]. The high prevalence and diversity of gamma coronaviruses in Beringia wild birds may contribute to the intercontinental spread and evolution of these viruses \\[[@CR32]\\]. The presence of the European QX-like viruses ITA/90254/2005 and CK/SWE/0658946/10 in a cluster with the early QX-like viruses in the S1 gene tree may also reflect recurrent recombination events between viruses of the North America-Europe and China lineages. The recent detection of IBV-like gene fragments in the wild bird population may support a possible role for wild birds as reservoirs or long distance carriers \\[[@CR12], [@CR17]\\]. However, it is not clear whether they play a role in the recombination events of IBV. In addition, it is not clear why the QX-like viruses provide the spike gene rather than the 1ab gene.\n\nThe 1ab and N genes are the minimum essential core elements for genomic RNA replication and persistence of coronaviruses in infected cells. The 1ab gene encodes nsps 2--16, which are essential for viral RNA replication and pathogenicity \\[[@CR2], [@CR13], [@CR29]\\]. Thus, 1ab genotypes that increase replication may be selected for in the competitive conditions of superinfected cells. The early QX-like viruses possessed the 1ab genes of KM91-like rather than QX-like viruses. The predominance of KM91-like 1ab genes supports the hypothesis that they conferred more favorable biological properties to the recombinants than the QX-like 1ab genes. For example, they may have conferred efficient virus replication and extended persistence in the trachea \\[[@CR34]\\]. Based on the nucleotide and amino acid sequence identities, nsp 2, 3, 4, 5, and 6 were more variable than the other nsps among the viruses (Table\u00a0[4](#Tab4){ref-type=\"table\"}). Nsp 2 is a papain-like viral protease and nsp5 is a coronavirus endopeptidase. They are important for the post-translational cleavage of nsps \\[[@CR13]\\]. Nsp 3, 4, and 6 possess transmembrane domains and form an internal membrane scaffolding complex that facilitates the assembly of the membrane-associated replication complex \\[[@CR33], [@CR37]\\]. The interactions between these molecules and their role in complex formation may facilitate the coevolution of nsp 3, 4, and 6. The deubiquitinating activity of nsp 3 and its frequent mutations during the attenuation of virulent strains as well as the autophagosome-generating activity of nsp 6 may be directly, or indirectly, related to viral competence. This in turn may influence how recombinant viruses overcome host responses and promote their own growth \\[[@CR3], [@CR10], [@CR34]\\].\n\nThe compatibility of the leader TRS sequence with the body TRS sequences is a determinant of viral gene expression levels and viral replication efficiency \\[[@CR4]\\]. The 2--3 nucleotide mismatches of the body TRS sequences of the early and the NC1 sg-mRNA 1 encoding Spike, which is the major target of humoral immunity, were comparable to the single nucleotide mismatch in recurrent QX-like viruses. The 1--2 mismatches in the sg-mRNA 3 (encoding the E protein) and lack of mismatch in the sg-mRNA 5 (encoding the 5a and 5b proteins) of the early QX-like and the NC1 viruses, except SNU9106, are also consistent with 3 and 2 mismatches, respectively, in the corresponding sg-mRNAs of recurrent QX-like viruses. Differences in the expression of the E, 5a, and 5b proteins may affect virus replication efficiency \\[[@CR36]\\]. Thus, these mismatches between the leader and the body TRSs may be related to the predominance of certain genomic backgrounds. Furthermore, the higher copy number of the 1(S) sg-mRNA in recurrent QX-like viruses than KM91-like viruses may increase the probability of RNA-dependent RNA polymerase mediated spike gene recombination \\[[@CR22]\\]. Therefore, further studies to investigate these mechanisms may be valuable in the future.\n\nThe proteolytic cleavages of the spike protein are essential for virus infection of host cells. The cleavage site between S1 and S2 contained the furin-recognition motif R-X-R(K)-R, and was shared by all of the compared viruses \\[[@CR5]\\]. The second proteolytic cleavage site near the fusion peptide of the spike protein is important for virus entry, syncytium formation and virus infectivity \\[[@CR41]\\]. The second cleavage sites of all the compared viruses contained only 1 to 3 arginines, and they may be cleaved by other serine proteases \\[[@CR41]\\]. The effects of the single amino acid insertions in the ER motif of E and ectodomain of M proteins are unknown. However, the differing amino acid sequences of the M protein ectodomains, identified when examining recurrent QX-like viruses, encourages further study of a role for IBV antigenic variation \\[[@CR14]\\].\n\nThe prevalence of the early QX-like viruses may be explained by amino acid variation in the Spike protein's epitopes, brought about by immune-selection pressure induced by KM91 vaccination. However, host immune pressure induced by infections with QX-like viruses between 2002 and 2005, and vaccination, might have resulted in the selection of NC1 viruses. Compared to the early NC1 viruses, QIA-Q43/06, SNU9106, and SNU10043 possessed the C/A/B epitopes, which were more similar to KM91 than QX-like viruses; however, they also acquired additional mutations in the F and G epitopes, which were similar, but not identical, to KM91. Despite pressure for the NC1 viruses to evolve new epitopes, a live-attenuated vaccine strain, K2, was successful in the field, and the NC1 has rarely been isolated since 2011 \\[[@CR28], [@CR31]\\]. However, the prevalence of the recurrent QX-like viruses and their epitope structures may reflect incomplete protective effects or vaccine breaks when conventional vaccines are applied in the field.\n\nConclusion {#Sec15}\n==========\n\nIn conclusion, the early QX-like viruses are recombinant viruses between KM91-like and QX-like viruses, and NC1 viruses subsequently evolved via recombination and the accumulation of missense mutations in spike genes. The predominance of KM91-like 1ab and QX-like spike genes in the recombinant viruses suggests that these loci are beneficial with respect to replication efficiency and the capacity to evade immunity, respectively.\n\nElectronic supplementary material\n=================================\n\n {#Sec16}\n\nBelow is the link to the electronic supplementary material. Supplementary material 1 (PDF 37 kb)\n\nThis study was supported by the BK 21 plus Program for Veterinary Science Research.\n\nThis work was partially supported by a QIA grant (No. B-1543084-2016-18-1)\n\nThe authors declare no conflicts of interest.\n\nThis article does not contain any studies with human participants or animals performed by any of the authors.\n"}
+{"text": "Introduction {#fsn3367-sec-0001}\n============\n\nAccording to Kenya Demographic Health Survey 2014 (KNBS and ICF Macro [2015](#fsn3367-bib-0020){ref-type=\"ref\"}), the proportion of stunted and wasted children in Kenya was 26% and 4%, respectively. Children aged 6--11\u00a0months had the highest wasting rates of 7%. These children fail to attain their full potential of growth and development, suffer long\u2010term deprivation of energy, nutrients, and consequently chronic protein energy malnutrition (PEM), often accompanied by micronutrient deficiencies. The period 6--24\u00a0months of age is one of the most critical periods in the growth of the infant. At this age, their demand for nutrients relative to their body size is high. However, there are limitations in the quality and quantity of available complementary foods (Dewey and Adu\u2010Afarwuah [2008](#fsn3367-bib-0009){ref-type=\"ref\"}; Owino et\u00a0al. [2008](#fsn3367-bib-0030){ref-type=\"ref\"}; Imdad et\u00a0al. [2011](#fsn3367-bib-0016){ref-type=\"ref\"}).\n\nSome of the challenges during this period include the use of plant\u2010based complementary foods that are too bulky for the weanling with a tiny stomach to eat the necessary quantities that provide adequate nutrients and energy to meet their requirements. Besides plant\u2010based foods notably unrefined cereals, legumes, and nuts contain high levels of phytates and at times polyphenols (Gibson and Ferguson [2008](#fsn3367-bib-0101){ref-type=\"ref\"}). These components vary in the degree to which they inhibit iron and zinc absorption (Gibson and Ferguson [2008](#fsn3367-bib-0101){ref-type=\"ref\"}; WHO/FAO [2004](#fsn3367-bib-5000){ref-type=\"ref\"}). Therefore, the children who mainly rely on these unprocessed foods may develop PEM and micronutrient deficiencies. Hence, there is a need to develop nutrient\u2010dense complementary foods for such children.\n\nThere is currently, a lot of interest in the nutritional value of amaranth plant, whose leaves are eaten as a vegetable in many parts of Kenya (Mlakar et\u00a0al. [2010](#fsn3367-bib-0028){ref-type=\"ref\"}). Amaranth seeds (Fig.\u00a0[1](#fsn3367-fig-0001){ref-type=\"fig\"}) contain more protein than most grains such as wheat, maize, rice, and sorghum (Mlakar et\u00a0al. [2010](#fsn3367-bib-0028){ref-type=\"ref\"}; Mburu et\u00a0al. [2012](#fsn3367-bib-0024){ref-type=\"ref\"}). Further, they contain relatively high levels of micronutrients especially iron, phosphorus, magnesium, vitamin A and E (Kauffman and Weber [1990](#fsn3367-bib-0019){ref-type=\"ref\"}; Alemu [2008](#fsn3367-bib-0002){ref-type=\"ref\"}). Besides, amaranth grain is recommended for infants because of its high protein digestibility, absorption, and retention by the baby\\'s body system (Kauffman and Weber [1990](#fsn3367-bib-0019){ref-type=\"ref\"}). In addition, amaranth has satisfactory lysine and tryptophan contents by FAO/WHO standards. This makes it a valuable complement to staples like maize or sorghum which are limited in amino acid lysine and tryptophan, respectively. Amaranth seed is, however, deficient in the amino acid leucine. Kauffman and Weber ([1990](#fsn3367-bib-0019){ref-type=\"ref\"}) reported threonine as the limiting amino acid in amaranth seeds. However, amaranth grain is not commonly consumed in Kenya, though it could give a nutrient\u2010dense complementary food. In this study, sorghum grain was used to complement amaranth grain because it is a drought\u2010tolerant traditional staple cereal that is widely utilized in Kenya.\n\n![Amaranth grain.](FSN3-5-86-g001){#fsn3367-fig-0001}\n\nA number of traditional food processing technologies such as germination and lactic acid fermentation have been proposed as ways to improve nutrient density of complementary foods and reduce antinutrients (Gibson and Hotz [2007](#fsn3367-bib-0013){ref-type=\"ref\"}). This study, therefore, aimed at developing a nutrient\u2010dense high\u2010quality complementary food from amaranth and sorghum grains through processing (steeping and germination).\n\nMethodology {#fsn3367-sec-0002}\n===========\n\nAmaranth\u2010sorghum grains product development (processing) {#fsn3367-sec-0003}\n--------------------------------------------------------\n\nThe ingredients were processed to optimize the nutrient density while reducing the antinutrient content. Then product formulation was done followed by nutrient and antinutrient content determination.\n\n### Preparation of raw materials {#fsn3367-sec-0004}\n\nOne batch, each of amaranth and sorghum grains were prepared as shown in Figure\u00a0[2](#fsn3367-fig-0002){ref-type=\"fig\"}. The grains were sorted, washed, steeped, and germinated at room temperature. They were then dried in an oven at 60\u00b0C for 48\u00a0h and the rootlets were removed. The grains were then boiled until soft for (30\u00a0min), dried, and ground into flour. A second batch of amaranth grain was roasted immediately after germination to find out if the nutrient content would be better than when it was only boiled after germination. The flours were stored in sealed cellophane bags in a cold room at 10\u00b0C.\n\n![Processing of amaranth and sorghum grains for product development.](FSN3-5-86-g002){#fsn3367-fig-0002}\n\n### Formulation of the complementary food from amaranth\u2010sorghum flours {#fsn3367-sec-0005}\n\nAmaranth grains that were steeped for 5\u00a0h and germinated for 24\u00a0h as well as sorghum grains steeped for 24\u00a0h and germinated for 72\u00a0h, were ground into flour, and then mixed in various ratios as shown in Table\u00a0[1](#fsn3367-tbl-0001){ref-type=\"table-wrap\"} (Okoth et\u00a0al. [2011](#fsn3367-bib-0029){ref-type=\"ref\"}). The product formulation was aimed at providing adequate energy and protein content while minimizing antinutrients content. The formulations were then stored in cellophane bags at 0\u00b0C prior to laboratory analyses.\n\n###### \n\nFormulation of nutrient\u2010dense complementary food\n\n  Formulations of the grains                  Amaranth:sorghum ratios\n  ------------------------------------------- -------------------------\n  Amaranth boiled\u00a0+\u00a0sorghum boiled            90:10\n  Amaranth boiled\u00a0+\u00a0sorghum boiled            80:20\n  Amaranth boiled\u00a0+\u00a0sorghum boiled            70:30\n  Amaranth boiled\u00a0+\u00a0sorghum boiled            60:40\n  Amaranth roasted, boiled\u00a0+\u00a0sorghum boiled   90:10\n  Amaranth roasted, boiled\u00a0+\u00a0sorghum boiled   80:20\n  Amaranth roasted, boiled\u00a0+\u00a0sorghum boiled   70:30\n  Amaranth roasted, boiled\u00a0+\u00a0sorghum boiled   60:40\n\nJohn Wiley & Sons, Ltd\n\nDetermination of energy and protein content of the formulations {#fsn3367-sec-0006}\n---------------------------------------------------------------\n\nThe gross energy content was determined using a bomb calorimeter (C400) (Gallenkamp, Loughbrough, UK) while the protein content was determined using the semi\u2010micro kjeldahl method (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}; Method 20.87\u201032.1.22).\n\nDetermination of the nutritional profile of the developed amaranth and sorghum grains product {#fsn3367-sec-0007}\n---------------------------------------------------------------------------------------------\n\nMoisture and fat contents were determined by hot air drying (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}, Method 925. 10\u201032.1.03) and Soxhlet extraction (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}; Method 920.85\u201032.1.13) methods, respectively. Fatty acids were converted to methyl esters using 5% HCl in methanol (w/v), followed by extraction with hexane (Ichihara and Fukubayashi [2009](#fsn3367-bib-0015){ref-type=\"ref\"}). The methyl esters were then separated and quantified by gas chromatography and flame ionization detection (Ciftci et\u00a0al. [2009](#fsn3367-bib-0008){ref-type=\"ref\"}). Crude Fiber content was quantified using the Henneberg--Stohmann method (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}; Method 920.86\u201032.1.15). Total mineral content was determined by dry ashing method (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}; Method 923.03\u201032.1.05). The total iron and zinc contents were determined by wet digestion and with an atomic absorption spectrophotometer (AAS) (AOAC [1995](#fsn3367-bib-0004){ref-type=\"ref\"}; Method 970.12). Product carbohydrate content was calculated by difference method as described by FAO ([2010](#fsn3367-bib-0012){ref-type=\"ref\"}), Shahnawaz et\u00a0al. ([2009](#fsn3367-bib-0032){ref-type=\"ref\"}) and James ([1995](#fsn3367-bib-0017){ref-type=\"ref\"}) given below:$$\\text{Total\\ carbohydrate\\ in\\ 100\\ g\\ of\\ food} = \\text{100}\\,\\text{\u2010}\\,\\text{weight\\ in\\ g\\ of\\ (Protein}\\,\\text{+}\\,\\text{Water}\\,\\text{+}\\,\\text{Ash}\\,\\text{+}\\,\\text{Fat)}$$where g\u00a0=\u00a0grams.\n\nThe HCl extractable iron and zinc in the product were determined as described by Mbithi\u2010Mwikya et\u00a0al. ([2000](#fsn3367-bib-0023){ref-type=\"ref\"}). Iron and zinc availability in the product was determined by the method described by Svanberg et\u00a0al. ([1993](#fsn3367-bib-0034){ref-type=\"ref\"}) and modified by Matuschek et\u00a0al. ([2001](#fsn3367-bib-0022){ref-type=\"ref\"}). Sugars were separated and quantified using AOAC ([1995](#fsn3367-bib-0004){ref-type=\"ref\"}), method 980.13. *\u03b2*\u2010carotene was determined using UV\u2010visible spectrophotometric method as described by Srivastava and Sanjeev ([1998](#fsn3367-bib-0033){ref-type=\"ref\"}).\n\nDetermination of the antinutrient content of the amaranth\u2010sorghum product {#fsn3367-sec-0008}\n-------------------------------------------------------------------------\n\nThe tannin content of the amaranth\u2010sorghum product was determined using the modified vanillin--HCl method (Makokha et\u00a0al. [2002](#fsn3367-bib-0021){ref-type=\"ref\"} adapted from Price et\u00a0al. [1978](#fsn3367-bib-0031){ref-type=\"ref\"} and Burns [1963](#fsn3367-bib-0006){ref-type=\"ref\"}). The phytate content was determined by the Camire and Clydesdale ([1982](#fsn3367-bib-0007){ref-type=\"ref\"}) method. In vitro\u2010protein digestibility was determined by the Mertz et\u00a0al. ([1984](#fsn3367-bib-0025){ref-type=\"ref\"}) method.\n\nResults and Discussion {#fsn3367-sec-0009}\n======================\n\nFormulation of the amaranth\u2010sorghum (complementary) food product {#fsn3367-sec-0010}\n----------------------------------------------------------------\n\nThe various formulations that were made, their moisture, energy, and protein content are given in Table\u00a0[2](#fsn3367-tbl-0002){ref-type=\"table-wrap\"}. The mean energy content of the formulations was 4.5\u00a0kcal with, a minimum of 3.5\u00a0kcal and a maximum of 5\u00a0kcal. The mean protein content of the formulations was 14.7% with a minimum of 12.9% and a maximum of 17.4%. The formulation with germinated and boiled amaranth grain and germinated and boiled sorghum grain in the ratio of 90:10 had the highest energy (5.0\u00a0\u00b1\u00a00.2\u00a0kcal per g). This is higher as compared to energy obtained from 1\u00a0g of both carbohydrate (4\u00a0kcal) and protein (4\u00a0kcal). The energy content of the formulation with 90% amaranth grain was significantly higher than that of the other formulations (*F*\u00a0=\u00a032.133, *P*\u00a0\\<\u00a00.05). However, there were no significant difference between the protein content of the 90% amaranth formulation and the other formulations (*F*\u00a0=\u00a02.319, *P*\u00a0=\u00a00.127). On the basis of energy content, the 90:10 amaranth to sorghum content formulation was chosen as the best for the complementary food. The moisture content was used to express the results on dry weight basis (dwb).\n\n###### \n\nThe mean of moisture, energy, and protein content of the formulations (dwb)\n\n  Nutrients   Formulation amaranth:sorghum ratio   Moisture content (%)   Energy (kcal per g dwb)   Protein content (% dwb)\n  ----------- ------------------------------------ ---------------------- ------------------------- -------------------------\n  ABSI        90:10                                4.8\u00a0\u00b1\u00a00.3              5.0\u00a0\u00b1\u00a00.2^1^              14.4\u00a0\u00b1\u00a00.6\n  ABS2        80:20                                4.8\u00a0\u00b1\u00a00.3              4.7\u00a0\u00b1\u00a00.1^1^              14.9\u00a0\u00b1\u00a00.7\n  ABS3        70:30                                5.2\u00a0\u00b1\u00a00.0              4.5\u00a0\u00b1\u00a00.2^1^              14.2\u00a0\u00b1\u00a00.3\n  ABS4        60:40                                4.4\u00a0\u00b1\u00a00.0              3.5\u00a0\u00b1\u00a00.2                 14.0\u00a0\u00b1\u00a00.1\n  ARBS1       90:10                                2.7\u00a0\u00b1\u00a00.3              4.8\u00a0\u00b1\u00a00.1^1^              17.4\u00a0\u00b1\u00a00.0\n  ARBS2       80:20                                2.9\u00a0\u00b1\u00a00.1              4.7\u00a0\u00b1\u00a00.2^1^              15.0\u00a0\u00b1\u00a00.2\n  ARBS3       70:30                                3.2\u00a0\u00b1\u00a00.7              4.6\u00a0\u00b1\u00a00.1^1^              14.1\u00a0\u00b1\u00a00.0\n  ARBS4       60:40                                3.8\u00a0\u00b1\u00a00.4              4.4\u00a0\u00b1\u00a00.1                 12.9\u00a0\u00b1\u00a00.0\n\n*N*\u00a0=\u00a03. Values are means\u00a0\u00b1\u00a0standard deviation of triplicate analysis. dwb, dry weight basis. Key: ABS1, Amaranth boiled\u00a0+\u00a0Sorghum, steeped, germinated, boiled; ABS2, Amaranth boiled\u00a0+\u00a0Sorghum, steeped, germinated, boiled; ABS3, Amaranth boiled\u00a0+\u00a0Sorghum, steeped, germinated, boiled; ABS4, Amaranth boiled\u00a0+\u00a0Sorghum, steeped, germinated, boiled; ARBS1, Amaranth roasted, boiled\u00a0+\u00a0sorghum steeped, germinated, boiled; ARBS2, Amaranth roasted, boiled\u00a0+\u00a0sorghum steeped, germinated, boiled; ARBS3, Amaranth roasted, boiled\u00a0+\u00a0sorghum steeped, germinated, boiled; ARBS4, Amaranth roasted, boiled\u00a0+\u00a0sorghum steeped, germinated, boiled. Significantly different *P*\u00a0\\<\u00a00.05.The energy content of the formulation with 90% amaranth grain was significantly higher than that of the other formulations (F=32.133, *P*\u00a0\\<\u00a00.05).\n\nJohn Wiley & Sons, Ltd\n\n\"Energy value should be viewed in light of its density in the ready to eat form of a complementary food because of differences in the viscosity in foods leading to bulk densities\" (Mbithi\u2010Mwikya et\u00a0al. [2000](#fsn3367-bib-0023){ref-type=\"ref\"}). According to Wardlaw and Kessel ([2002](#fsn3367-bib-0104){ref-type=\"ref\"}), a child at 6\u00a0months requires 700\u00a0kcal of energy per day. To meet this need, a 6\u2010month\u2010old child will need to take two cups (350\u00a0mL each) of the complementary food in a day with the assumption that 1\u00a0mL/g of the food provides about 1\u00a0kcal. A child between 1 and 3\u00a0years requires 1022\u00a0kcal of energy per day (Burgess, A., and P, Glasauer. [2004](#fsn3367-bib-0005){ref-type=\"ref\"}). Therefore, they would need to take two cups of 300\u00a0mL of the amaranth\u2010sorghum food in a day with a dilution of 1:2 (Amaranth\u2010sorghum product to water).\n\nThe formulation with a ratio of 90 (steeped, germinated, roasted, and boiled amaranth grain):10 (germinated and boiled sorghum grain) had the highest protein content. However, the differences in protein content among the different formulations were not significant. The formulation chosen for the product had a protein content of 14.4%. This is higher than the estimated protein needs from complementary foods even for a 12--23\u2010month\u2010old child of low breast milk intake (9.1\u00a0g/d). The formulation chosen contains 28.8\u00a0g protein per 1000\u00a0kcal. This is considerably higher than that in the World Health Organization/Food and Agriculture Organization (WHO/FAO) report on protein requirements (WHO [2007](#fsn3367-bib-0105){ref-type=\"ref\"}), which recommended a minimum of 6.9% energy contributed by protein.\n\nNutritional properties of developed amaranth\u2010sorghum product and the effect of processing on the nutrients {#fsn3367-sec-0011}\n----------------------------------------------------------------------------------------------------------\n\nThe proximate composition of the product is given in Table\u00a0[3](#fsn3367-tbl-0003){ref-type=\"table-wrap\"}. The carbohydrate content was 71.4%. It provides an energy content of 1.2\u00a0kcal per 1\u00a0g wet weight basis. The product provides 28.8\u00a0g protein per 1000\u00a0kcal of energy. It had a glucose content of 14.8\u00a0g per 100\u00a0g, fructose 9.3\u00a0g per 100\u00a0g, and sucrose content of 12.3\u00a0g per 100\u00a0g. Michaelsen et\u00a0al. ([2008](#fsn3367-bib-0102){ref-type=\"ref\"}) reported that the most important dietary mono\u2010 and disaccharides are glucose, fructose, lactose, and sucrose (sugar). These sugars are good sources of energy and will typically increase the energy density of a diet (Michaelsen et\u00a0al. [2008](#fsn3367-bib-0102){ref-type=\"ref\"}).\n\n###### \n\nNutritional profile of the developed complementary food on dwb\n\n  Nutrient            Quantity\n  ------------------- ------------\n  Moisture %          5.2\u00a0\u00b1\u00a00.0\n  Energy kcal per g   5.0\u00a0\u00b1\u00a00.2\n  Protein %           14.4\u00a0\u00b1\u00a00.3\n  Fat %               6.8\u00a0\u00b1\u00a00.4\n  Ash %               2.2\u00a0\u00b1\u00a00.1\n  Crude fiber %       3.0\u00a0\u00b1\u00a00.1\n\nValues are means\u00a0\u00b1\u00a0standard deviation of triplicate analysis. dwb, dry weight basis.\n\nJohn Wiley & Sons, Ltd\n\nThe proximate composition of fortified blended foods (FBF) commonly used by World Food Programme (WFP) to rehabilitate moderately malnourished children is given in Table\u00a0[4](#fsn3367-tbl-0004){ref-type=\"table-wrap\"}. The FBF nutritional value was compared with that of the developed food product (Table\u00a0[3](#fsn3367-tbl-0003){ref-type=\"table-wrap\"}). The developed complementary food has a higher energy content compared to the FBF. The developed amaranth\u2010sorghum product protein (14.4%) provides 28.8\u00a0g protein per 1000\u00a0kcal of energy. This is higher than the protein per kcal in FBF (Table\u00a0[4](#fsn3367-tbl-0004){ref-type=\"table-wrap\"}). Further, there was no indigestible protein detected in the product. The inability to detect indigestible protein in the product can be attributed to the decrease in the level of antinutrients (Singh et\u00a0al. [1982](#fsn3367-bib-0103){ref-type=\"ref\"}; Kataria et\u00a0al. [1989](#fsn3367-bib-0018){ref-type=\"ref\"}; Mbithi\u2010Mwikya et\u00a0al. [2000](#fsn3367-bib-0023){ref-type=\"ref\"}).\n\n###### \n\nNutritional value of commonly used food aid commodities\n\n                    Energy (kcal per 100\u00a0g)   Fat (MJ/100\u00a0g)   Fat (E %)   Protein (g per 100\u00a0g)   Protein (E %)\n  ----------------- ------------------------- ---------------- ----------- ----------------------- ---------------\n  Corn Soy blend    3.8                       1.6              14.2        18.0                    18.9\n  Wheat Soy blend   3.7                       1.55             14.6        20.0                    21.1\n  Corn Soy milk     3.8                       1.6              14.2        20.0                    21.1\n\nSource: WFP (\n\n2000\n\n); Michaelsen et al. (\n\n2008\n\n).\n\nJohn Wiley & Sons, Ltd\n\nIn the study, no fat was added to the product, as is usually done in commercially blended foods. Usually in FBFs, the oil is added at the time of distribution to minimize use of oil for other purposes. Fat tends to reduce the shelf life of food. The energy content supplied by fat in the amaranth\u2010sorghum product was 12.2% of the total energy in the product. Linoleic acid (LA) was the highest in amount compared with the other fatty acids in the product as given in Table\u00a0[5](#fsn3367-tbl-0005){ref-type=\"table-wrap\"} while lauric acid (0.34\u00a0mg/100\u00a0g) was the least. The essential fatty acids in the product provided 7.4% energy (E %) of the total energy on dwb. The n\u20103 polyunsaturated fatty acids (PUFA) provided by the developed product was 1.8 E %.\n\n###### \n\nEssential fatty acids and fatty acid profile of the complementary food\n\n  Fatty acid composition   Abbreviation   Quantity (mg per 100\u00a0g dwb)\n  ------------------------ -------------- -----------------------------\n  Caprylic acid            C8:0           0.68\n  Lauric acid              C12:0          0.034\n  Myristic acid            C14:0          47.6\n  Palmitic acid            C16:0          115.6\n  Oleic                    C18:1          176.8\n  Linoleic acid            C18:2          231\n  Linolenic acid           C18:3          10.2\n  Stearic acid             C18:0          11.6\n\nValues are means\u00a0\u00b1\u00a0standard deviation of triplicate analysis. dwb, dry weight basis.\n\nJohn Wiley & Sons, Ltd\n\nThere are two types of essential fatty acids, the n\u20106 and the n\u20103 PUFA, which in most diets are provided by vegetable oils in the form of LA (25 C18:2n\u20106) and *\u03b1*\u2010linolenic acid (ALA, C18:3n\u20103), respectively. According to Wulf et al. ([2008](#fsn3367-bib-0500){ref-type=\"ref\"}) and Michaelsen et al. ([2009](#fsn3367-bib-0027){ref-type=\"ref\"}), the fat intake of young children (1--2\u00a0years) should have a quality that provides 5--10 E % as essential fatty acids including at least 1 E % of n\u20103 PUFA and have a ratio of n\u20106 to n\u20103 PUFA between 3 and 9. For complementary feeding of children who are well nourished, a level of 30--45 fat E % was recommended, including the fat from breast milk (Dewey and Brown [2003](#fsn3367-bib-0010){ref-type=\"ref\"}). Several authors have recommended that fat should supply 30--45% of energy intake for children less than 2\u00a0years of age (Michaelsen and Jorgensen [1995](#fsn3367-bib-0026){ref-type=\"ref\"}; Mbithi\u2010Mwikya et\u00a0al. [2000](#fsn3367-bib-0023){ref-type=\"ref\"}; Wardlaw and Kessel [2002](#fsn3367-bib-0104){ref-type=\"ref\"}).\n\nThe crude fiber was less than 5%. It is recommended that insoluble fibers should be present in low amounts in the diet because they increase bulk and reduce gastrointestinal transit time (Michaelsen et\u00a0al. [2009](#fsn3367-bib-0027){ref-type=\"ref\"}). Dietary fibers may reduce energy intake through a suppressing effect on appetite and they may increase fecal losses of energy due to reduced absorption of fat and carbohydrate (Aggett et\u00a0al. [2003](#fsn3367-bib-0001){ref-type=\"ref\"}; Michaelsen [2009](#fsn3367-bib-0027){ref-type=\"ref\"}). The amaranth\u2010sorghum product had a total iron content of 6\u00a0mg/100\u00a0g, a value slightly lower than that of raw amaranth grains (6.6\u00a0mg per 100\u00a0g) (Fig.\u00a0[3](#fsn3367-fig-0003){ref-type=\"fig\"}). It had a higher HCl extractability of iron than its ingredients (93.3%) and its bioavailable iron was 5.5\u00a0mg per 100\u00a0g (dwb) (11.2\u00a0mg per 1000\u00a0kcal).\n\n![Total iron and HCl extractable iron content of the product and its ingredients (dry weight basis).](FSN3-5-86-g003){#fsn3367-fig-0003}\n\nThe amaranth\u2010sorghum product had a total zinc content of 3.2\u00a0mg per 100\u00a0g as given in Figure\u00a0[4](#fsn3367-fig-0004){ref-type=\"fig\"}. The HCl extractable zinc in the food product was 84% which was higher than both its ingredients. Also, it had 2.8\u00a0mg per 100\u00a0g (dwb) (5.6\u00a0mg per 1000\u00a0kcal) bioavailable zinc.\n\n![The total and HCl extractable zinc content in the product and its ingredients.](FSN3-5-86-g004){#fsn3367-fig-0004}\n\nThe developed amaranth\u2010sorghum food product had a higher availability of iron and zinc compared to raw amaranth and sorghum grains. For a food\u2010based approach and for most formulated diets, it is important that the basic ingredients be such that the iron is more available (Golden [2009](#fsn3367-bib-0014){ref-type=\"ref\"}). According to Wardlaw and Kessel ([2002](#fsn3367-bib-0104){ref-type=\"ref\"}), the recommended daily intake of iron and zinc for children 7--12\u00a0months is 11\u00a0mg per day (mg/d) while those 1--3\u00a0years it is 7\u00a0mg/d. HCl extractability of minerals and trace elements under simulated gastric conditions is an indicator of bioavailability in foods (Antony and Chandra [1998](#fsn3367-bib-0003){ref-type=\"ref\"}). The improvement in in vitro iron and zinc bioavailability observed in this study could be attributed to a decrease in phytic acid among other antinutrients.\n\nThe complementary food contained 4.2\u00a0*\u03bc*g per 100\u00a0g of *\u03b2*\u2010carotene (7.1\u00a0I.U of vitamin A; 0.35 retinol activity equivalents per 100\u00a0g) on dwb. It also provided 0.7 retinol activity equivalents/1000\u00a0kcal. The recommended daily allowance of vitamin A is 400, 500, and 300 retinol equivalents for children 0--6\u00a0months, 7--12\u00a0months, and 1--3\u00a0years, respectively. The complementary food, therefore, will not supply the estimated needs of vitamin A for children below 23\u00a0months.\n\nAntinutritional components of the amaranth\u2010sorghum product (dwb) {#fsn3367-sec-0012}\n----------------------------------------------------------------\n\nRaw amaranth grain had a tannin content of 0.8% CE while sorghum had a tannin content of 24.9% CE. No tannins could be detected in the product. Phytates levels (inositol hexaphosphate) were 7.9 and 252\u00a0mg per 100\u00a0g in raw amaranth and sorghum grains, respectively. No phytates were detected in the developed complementary food. The most important antinutritional food constituent in diets in low\u2010income countries in terms of negative nutritional impact is phytate, primarily contributed by cereal staples and secondarily by legumes and other plant foods (Michaelsen et\u00a0al. [2009](#fsn3367-bib-0027){ref-type=\"ref\"}). Phytate forms insoluble complexes with a range of nutrients and thereby inhibits the absorption of protein and minerals in particular iron, zinc, and calcium. One of the most widespread groups of polyphenols with antinutritional properties is soluble tannins (Michaelsen et\u00a0al. [2009](#fsn3367-bib-0027){ref-type=\"ref\"}). The antinutritional effect of polyphenols is complex formation with iron and other minerals, and precipitation of protein, which reduces absorbability. The fact that the antinutrients in the product could not be detected in the steeped and germinated grain implies that minerals and nutrients usually bound by them could be more bioavailable.\n\nConclusion {#fsn3367-sec-0013}\n==========\n\nA complementary food from amaranth and sorghum grains with an energy density of 5\u00a0kcal per g sufficient to meet the recommendations for children 6--23\u00a0months of age at 2--3\u00a0servings per day was developed. The food product was also low in antinutrients and therefore of correspondingly high nutrient availability and digestibility.\n\nConflict of Interest {#fsn3367-sec-0015}\n====================\n\nNone declared.\n\nWe express special thanks to Jomo Kenyatta University of Agriculture and Technology for the study leave and financial support during the research. We acknowledge the Kenya Agricultural Productivity Project, Project number KAPP06/PRC\u2010SECBCCI\u201006FP2006023 under N. K. Gikonyo, the Principal Investigator for the financial support too.\n"}
+{"text": "INTRODUCTION {#SEC1}\n============\n\nCellular senescence is defined as a stable cell cycle arrest elicited in response to a variety of stressors. Intense oncogenic signaling, telomere loss, radiation, chemotherapeutic drugs, bacterial toxins and oxidative stress have all been linked to the induction of the senescent phenotype through direct DNA damage or replication stress-induced DNA damage ([@B1]--[@B9]). Interestingly, oxidative stress has been shown to induce cellular senescence ([@B8],[@B10],[@B11]) and replication stress independently ([@B12],[@B13]). There is a lack of evidence to implicate replication stress-induced DNA damage as the driver for the initiation of cellular senescence in response to oxidative stress. The acquisition of cellular senescence is a dynamic process in which changes take place over an extended period of time ([@B14]--[@B17]). These changes are necessary for the permanent halt of proliferation, failing which cells might escape from senescence to a pro-oncogenic state ([@B5],[@B18]). The senescent phenotype is associated with the activation of the tumor suppressor p53 through its phosphorylation at Ser15 residue, which prevents cells carrying genomic lesions from progressing through the cell cycle ([@B19]--[@B22]) and the acquisition of persistent DNA damage foci or DNA segments with chromatin alterations reinforcing senescence (DNA-SCARS) ([@B5],[@B16],[@B21]). DNA-SCARS contain mediators of the DNA damage response (DDR) such as CHK2 and p53, but lack DNA repair proteins and single-stranded DNA (ssDNA)-binding proteins such as Rad51 and the replication protein A (RPA) ([@B21]). The absence of DNA repair proteins such as Rad51 in DNA-SCARS has led to the proposal that DNA-SCARS formation is due to an ineffective DNA repair process, which is accelerated in cells deficient in DNA repair proteins of the homologous recombination (HR) repair system, such as Rad51 ([@B21]). However, induction of persistent foci and cell growth arrest is insufficient to complete the acquisition of the senescent phenotype. After the establishment of growth arrest, senescent cells undergo extensive changes in chromatin, which contribute to the progression of senescence into a deep senescent state ([@B23]). Among these important changes is the formation of senescence-associated heterochromatin foci (SAHF), which are regions of highly condensed chromatin structures observed *in vitro* and *in vivo* ([@B23]--[@B26]). As SAHF sequesters genes controlling proliferation and cell cycle, SAHF formation is an important step leading to the deepening of the senescent phenotype ([@B23],[@B26],[@B27]). Along these lines, an increase in expression of High Mobility Group AT-Hook 2 (HMGA2) is associated with the formation of SAHF ([@B28]). Furthermore, an important chromatin remodeling process during the establishment of cellular senescence is the formation of cytoplasmic chromatin fragments (CCFs). CCFs are heterochromatin structures that are extruded from the nucleus and processed by lysosomes, leading to the general loss of histones in the senescent cells ([@B17]). This extrusion process is facilitated by the disintegration of the nuclear membrane upon the repression of Lamin B1 protein expression, a rapid and early event in the deepening of the cellular senescent state. Lamin B1 downregulation triggers both global and local modifications in chromatin, inducing an extensive chromatin remodeling and consequently enhancing senescent characteristics and deepening of the senescent phenotype ([@B14],[@B17],[@B29]). Furthermore, senescent cells secrete cytokines such as interleukin (IL)-6 as part of the senescence-associated secretory phenotype (SASP) ([@B3],[@B30]).\n\nIn the present report, using increasing concentrations of exogenous H~2~O~2~, we demonstrate that replication stress-dependent formation of endogenous DNA damage is responsible for the initiation and establishment of the senescent phenotype induced by sub-lethal oxidative stress. Moreover, we show the critical role of p53 in the inhibition of Rad51 and Lamin B1 but not in the increase in senescence-associated \u03b2-galactosidase (SA-\u03b2-Gal) activity and HMGA2 expression upon the establishment of the senescent state.\n\nMATERIALS AND METHODS {#SEC2}\n=====================\n\nCell lines and cultures {#SEC2-1}\n-----------------------\n\nL6 rat myoblasts were obtained from Dr Larry Fliegel (Department of Biochemistry, University of Alberta, Canada). L6 myoblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 0.25 mg/ml Geneticin (G418 sulfate) and 1 mM Gentamicin Sulfate at 37\u00b0C, with 5% CO~2~ in a humidified atmosphere. Human retinal pigmental cells-1 (RPE1-hTERT) cells were grown in DMEM-F12 medium, supplemented with 10% FBS and 1 mM Gentamicin sulfate at 37\u00b0C, with 5% CO~2~ in a humidified atmosphere. After 24 h of seeding, the cells were treated with desired H~2~O~2~ concentration (Merck-107209) for 1 h and then replenished with fresh media and collected at the respective time intervals as per the experimental requirements. In the case of synchronized condition, after 24 h of seeding, the cells (L6 and RPE1) were treated with 2 mM thymidine for 24 h to synchronize the cells at late G1 phase, followed by H~2~O~2~ treatment. L6 cells were also synchronized in G1 using 1 \u03bcM lovastatin (Merck Millipore \\#438185) and late G1 phase using 10 \u03bcM aphidicolin (Merck Millipore \\#178273). In the case of gene silencing experiments, after 24 h of seeding, the cells were maintained in SiRNA mixture for 6 h and then followed by 24 h of synchronization and 1 h of H~2~O~2~ treatment. The cell morphology images were captured using Nikon Eclipse TS100 as per the experimental requirements.\n\nReagents and chemicals {#SEC2-2}\n----------------------\n\nThe chemicals used in the study and the corresponding catalog numbers are: hydrogen peroxide (H~2~O~2~) (Merck \\#107209), 1,10-phenanthroline monohydrate (Phen) (Sigma \\#131377), 4-hydroxy TEMPO (Tempol) (Sigma \\#176141), MnTMPyP (Calbiochem \\#475872), thymidine (Sigma \\#T1895), etoposide (Sigma \\#E1383), KU55933 (Selleck \\#S1092) and AZ20 (\\#S7050).\n\nCrystal violet assay: cell viability {#SEC2-3}\n------------------------------------\n\nBriefly, after washing with 1\u00d7 phosphate buffered saline (PBS), cells were stained with 0.5 ml crystal violet solution (0.75% (w/v) crystal violet, 50% (v/v) ethanol, 1.75% (v/v) formaldehyde, 0.25% (w/v) NaCl) for 10 min and then lysed with 1% sodium dodecyl sulphate (SDS)/PBS to solubilize the dye retained in the adherent cells. The absorbance was measured using Spectrafluor Plus spectrofluorometer (TECAN, GmbH, Gr\u00f6dig, Austria) at 595 nm.\n\nSDS-PAGE and western blot analysis {#SEC2-4}\n----------------------------------\n\nCell lysates were prepared in radio-immunoprecipitation assay (RIPA) lysis buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM egtazic acid (EGTA), 1% Triton X-100, supplemented with 1 mM Na~3~VO~4~ (Sigma, S6508), 1 \u03bcg/ml leupeptin, 1 \u03bcg/ml pepstatin A, 1 \u03bcg/ml aprotinin and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cell lysates were resolved by SDS-PAGE and probed with protein of interest using SuperSignal chemiluminescent substrate (Thermo Scientific, 34080) with Kodak Biomax MR X-ray film or Biorad ChemiDoc\u2122 MP System. For re-probing of the same membrane for the detection of different proteins, membrane was stripped with Restore western blot stripping buffer (Thermo Scientific, 21059). The primary antibodies used: rabbit-\u03b3H2AX (Ser139) (\\#2577), rabbit-p-p53 (Ser15) (\\#9284), rabbit-total p53 (\\#9282) rabbit- \u03b2-tubulin (\\#2146) and horseradish peroxidase (HRP)-conjugated goat anti-rat secondary antibody (\\#7077) were purchased from Cell Signaling. Rabbit-p16INK4a (\\#sc-1207) antibody was purchased from Santa Cruz. Mouse-Rad51 (\\#ab1837), rabbit-Lamin B1 (\\#ab16048) and rabbit-MPG (\\#ab155092) were purchased from Abcam. Rabbit-HMGA2 antibody (\\#GTX100519) was purchased from GeneTex. Mouse-PAR antibody (\\#ALX-804--220-R100) was purchased from Enzo lifesciences. Mouse-RB (\\#554136) and mouse-p21 (\\#ab7960) were purchased from BD Bioscience. Mouse-\u03b2-actin antibody (\\#A5441) was purchased from Sigma. HRP-conjugated goat anti-mouse secondary antibody (\\#31430) was purchased from Pierce. HRP-conjugated goat anti-rabbit secondary antibody (\\#P0448) was purchased from DakoCytomation.\n\nImmunofluorescence assay using confocal microscopy {#SEC2-5}\n--------------------------------------------------\n\nCells were seeded on coverslips and were fixed with 4% paraformaldehyde for 30 min at room temperature. After permeabilization with 0.2% TX-100 (Sigma, X100) for 10 min at room temperature, cells were incubated with primary antibody overnight at 4\u00b0C. The primary antibodies used: rabbit-Lamin B1 (\\#ab16048) was purchased from Abcam; Rabbit-\u03b3H2AX (Ser139) (\\#2577), rat-RPA32 (\\#2208), mouse-Lamin A\\\\C (\\#4777) and rabbit-histone H3 (\\#4499) were purchased from Cell Signaling; Rabbit-Rad51 (\\#sc8349), rabbit-PML (\\#sc-H238) and rabbit anti-53BP1(\\#sc-22760) were purchased from Santa Cruz; Mouse-\u03b3H2AX (Ser139) (\\#05--636) was purchased from Merck millipore. After washing with PBS, cells were incubated with either one of the following secondary tagged antibodies: Rhodamine RedTM-X goat anti-rabbit IgG (\\#R6394), fluorescein isothiocyanate (FITC) goat anti-rabbit IgG (\\#65611), FITC goat anti-mouse IgG (\\#F2761), FITC goat anti-rat IgG (\\#A11006) and Hoechst 34580 (Molecular Probes, H21486) for 1 h at room temperature (RT). The images were captured using Zeiss LSM 710 confocal microscope (Carl Zeiss, Jena, Germany) and pictures were analyzed with Carl Zeiss ZEN 2010 software.\n\nCell cycle analysis and PI staining {#SEC2-6}\n-----------------------------------\n\nThe cells were fixed with 90% methanol for 30 min at room temperature. The cells were then stained with 500 \u03bcl of propidium iodide (PI)/RNaseA staining solution for 40 min at 37\u00b0C. The stained cells were then acquired using FACSCanto II-PE channel to measure the PI intensities. Flow cytometry data were analyzed using Flowing software V 2.5.1 (<http://www.flowingsoftware.com>) or FlowJo software.\n\nMeasurement of senescence-associated SA-\u03b2-Gal activity using C~12~FDG staining {#SEC2-7}\n------------------------------------------------------------------------------\n\nSenescence-associated \u03b2-galactosidase (SA-\u03b2-Gal) activity was measured as directed by the nature-protocols, Chainiaux *et\u00a0al.* ([@B31]).\n\nRNA interference (RNAi) assay {#SEC2-8}\n-----------------------------\n\nCellular transfection of SiRNA was performed using Lipofectamine RNAiMAX (Invitrogen, 13778150) in Opti-MEM reduced serum medium (Invitrogen, 31985070) according to the manufacturer's protocol. SiRNA for p53 and MPG (ON-TARGETplus SMARTpool---Rat TP53 and Rat MPG) were purchased from Dharmacon (Thermo Scientific), and control SiRNA (QIAGEN, Valencia, CA, USA) that is non-homologous to any known gene sequence was used as a negative control.\n\nNeutral comet assay {#SEC2-9}\n-------------------\n\nThe neutral comet assay and comet analysis were performed using in-house developed OpenComet V 1.3 software as described previously ([@B32]).\n\nMeasurement of rat interleukin-6 (IL-6) using ELISA {#SEC2-10}\n---------------------------------------------------\n\nThe conditioned media in the wells were collected at 72 h post-H~2~O~2~ treatment. Rat IL-6 secretion in the media was measured using rat IL-6 enzyme-linked immunosorbent assay (ELISA) kit (Invitrogen, CA, USA) according to manufacturer's protocol.\n\nRNA isolation, reverse transcription and real-time PCR {#SEC2-11}\n------------------------------------------------------\n\nTotal RNA was isolated using RNeasy Mini Kit (Qiagen, \\#74104) according to the manufacturer's instructions. Reverse transcription was performed using the TaqMan Reverse Transcription Reagents kit (Life technologies, N8080234). Real-time quantitative polymerase chain reaction (PCR) reaction was carried out with SYBR-Green (Applied Biosystems, \\#4309155), detection using ABI PRISM 7300 (Applied Biosystems). The sequences of the primers used for PCR were as follows: rat p53 (FP:GCTCCCCTGAAGACTGGATAA, RP: ATTAGGTGACCCTGTCGCTG), rat Lamin B1 (FP:AAGGCTCTCTACGAGACCGA, RP: TCCTTCTTAGCATAATTGAGCAGC), rat RAD 51 (FP:CTGCGAAGTGTGTTTGAGCC, RP:AGCCATTACTGTCTCGCAGC), rat 18S (FP:CATTCGAACGTC TGCCCT, RP:GTTTCTCAGGCTCCCTCTCC). Relative gene expression was obtained after normalization with endogenous 18S and determination of the difference in threshold cycle (*C*~t~) between treated and untreated cells using 2^-\u0394\u0394*C*^~t~ method ([@B33]).\n\nChromatin extraction {#SEC2-12}\n--------------------\n\nThe chromatin extraction was done as described previously ([@B34]). At the end of required time point, the cells were harvested and the cytosolic protein fraction was extracted by incubation in hypotonic buffer (10 mM hydroxyethyl piperazineethanesulfonic acid (HEPES), pH 7, 50 mM NaCl, 0.3 M sucrose, 0.5% Triton X-100, supplemented with protease inhibitor; Roche) for 15 min on ice and centrifuged at 9000 rpm for 5 min. The obtained supernatant was the cytosolic fraction and was transferred into a new tube. The remaining pellet was re-suspended with nuclear buffer (10 mM HEPES, pH 7, 200 mM NaCl, 1 mM EDTA, 0.5% NP-40 and protease inhibitor cocktail) and incubated on ice for 10 min and then centrifuged at 14 000 rpm for 5 min. The supernatant obtained was the nuclear fraction. The remaining pellet was re-suspended in the chromatin lysis buffer (10 mM HEPES, pH 7, 500 mM NaCl, 1 mM EDTA, 1% NP-40 and protease inhibitor cocktail), sonicated at low amplitude for 30 s and then centrifuged for 5 min at 14000 rpm; the supernatant obtained was the chromatin extract. The total protein concentration in the chromatin extract was determined using Coomassie Plus protein assay reagent (Pierce, Thermo Fisher Scientific Inc, Rockford, IL, USA), according to manufacturer's instructions and then proteins of interest were detected using western blot.\n\nImageJ analysis {#SEC2-13}\n---------------\n\nNuclear morphometric analysis (NMA): the nuclear analysis of H~2~O~2~-treated cells were performed using the NMA plugin as directed by the authors ([@B35]).Foci counting in H~2~O~2~-treated cells: the H2AX, 53BP1, Rad51 and RPA32 foci in H~2~O~2~-treated cells were counted using Focipicker 3D ImageJ plugin (<http://rsb.info.nih.gov/ij/plugins/foci-picker3d/index.html>) or FoCo a matlab-based software (<https://sourceforge.net/projects/focicount>) ([@B36]). The parameters of the both the plugins were adjusted dynamically as per the images requirements. A total of at least 50 nuclei were counted per treatment and per time point. The data were represented as average number of foci per nucleus.Co-localization finder: the co-localization between RPA32 and 53BP1 foci was found using co-localization finder plugin (<http://rsb.info.nih.gov/ij/plugins/colocalization-finder.html>). The respective channel images representing the RPA32 and 53BP1 foci were used to find the overlapping pixels between the two proteins localization at the DNA damage foci and the Pearson's correlation ratio, the masked and overlapping pixels were generated.\n\nMetaphase chromosome spreads preparation {#SEC2-14}\n----------------------------------------\n\nAfter 12 h of H~2~O~2~ treatment, colcemid to a final concentration of 150 ng/ml was added for 4 to 6 h. The cells were then collected and 6 ml of 75 mM KCl (pre-warmed to 37\u00b0C) was added dropwise for first 1 ml, while gently vortexing and then incubated for 16 min at 37\u00b0C. After spinning at 1000 rpm for 10 min, the pellet was re-suspended and 5 ml of fixative (3:1 solution of methanol:glacial acetic acid (freshly prepared)) was added and incubated for 20 min at 4\u00b0C. The fixative step was repeated another two times and after spinning at 1000 rpm for 10 min, the cell pellet was re-suspended with small volume of fixative and then added dropwise onto the slides and allowed to dry overnight. The slides were then stained with Hoechst and the images were captured using Zeiss LSM 710 confocal microscope (Carl Zeiss, Jena, Germany) and pictures were analyzed with Carl Zeiss ZEN 2010 and ImageJ software.\n\nStatistical analysis {#SEC2-15}\n--------------------\n\nStudent's *t*-test was performed when appropriate, using the Microsoft Excel software. The *P*-value (\\**P*-value \\< 0.05, \\*\\**P*-value \\< 0.01, \\*\\*\\**P*-value \\< 0.001 and ns---not significant) of \\<0.05 considered significant. The values represent the average of at least three independent experiments and error bars represent \u00b1 SD of the mean.\n\nRESULTS {#SEC3}\n=======\n\nSub-lethal oxidative stress induces two waves of DNA damage, growth arrest and senescence-like morphology {#SEC3-1}\n---------------------------------------------------------------------------------------------------------\n\nExposure of L6 cells to 25, 50 and 150 \u03bcM of exogenous H~2~O~2~ resulted in a significant increase in \u03b3H2AX, 1 h after exposure to the oxidant (Figure [1A](#F1){ref-type=\"fig\"} and\u00a0[B](#F1){ref-type=\"fig\"}). Interestingly, while \u03b3H2AX receded to baseline level 6 h after the initial oxidative stress when 25 \u03bcM H~2~O~2~ was used, a second wave of \u03b3H2AX between 12 and 24 h was detected in cells exposed to 50 and 150 \u03bcM H~2~O~2~ (Figure [1A](#F1){ref-type=\"fig\"} and\u00a0[B](#F1){ref-type=\"fig\"}). In order to confirm that \u03b3H2AX detected using western blot reflected DNA damage, neutral comet assay was performed in cells exposed to 50 \u03bcM H~2~O~2~. Two peaks of DNA damage were detected at 1 and 12 h following cells' exposure to the oxidant (Figure [1C](#F1){ref-type=\"fig\"} and\u00a0[D](#F1){ref-type=\"fig\"}). Interestingly, while cells exposed to 25 \u03bcM H~2~O~2~ showed a slight delay in the rate of proliferation following cells' exposure to the oxidant, no significant change in the cells' morphology could be detected up to 72 h after the redox stress (Figure [1E](#F1){ref-type=\"fig\"} and\u00a0[F](#F1){ref-type=\"fig\"}); however, increasing the concentration of H~2~O~2~ to 50 \u03bcM resulted in the inhibition of cells' proliferation, as evidenced by absence of an increase in cell number as well as the acquisition of enlarged and flattened cells, a cell morphology reminiscent of a senescent-like phenotype. Interestingly, inhibition of the first wave of \u03b3H2AX using the iron chelator phenanthroline and two reactive oxygen species (ROS) scavengers Tempol and MnTMYp prevented the second increase in \u03b3H2AX in cells exposed to 50 \u03bcM H~2~O~2~ compared to control cells ([Supplementary Figure S2A](#sup1){ref-type=\"supplementary-material\"}). Furthermore, pre-incubation of the cells with phenanthroline, Tempol or MnTMPyP rescued the cells' growth arrest observed in control cells 72 h after exposure to the sub-lethal oxidative stress ([Supplementary Figure S2E](#sup1){ref-type=\"supplementary-material\"}). Taken together, these results support that the first wave of \u03b3H2AX is required to drive the onset of the second wave of \u03b3H2AX and the cells' growth arrest that follows. On the other hand, 150 \u03bcM H~2~O~2~ induced cell death with apoptotic characteristics such as cell size shrinkage (Figure [1E](#F1){ref-type=\"fig\"} and\u00a0[F](#F1){ref-type=\"fig\"}). Intrigued by the association between the appearance of a second wave of DNA damage and different cell fate, we set out to gain a deeper understanding of the origin of the second wave of DNA damage in cells exposed to 50 \u03bcM H~2~O~2~ and asked if the senescent-like morphology observed reflected a true cellular senescent state.\n\n![Sub-lethal oxidative stress induces two waves of DNA damage, growth arrest and senescence-like morphology. L6 cells were treated with indicated concentration of H~2~O~2~ for 1 h and assessed for \u03b3H2AX and DNA damage by comet assay, cell growth and cell morphology. (**A**) Western blot of \u03b3H2AX kinetics in 25, 50 and 150 \u03bcM H~2~O~2~-treated cells. (**B**) Graph showing the densitometry plots of \u03b3H2AX from (A) in 25, 50 and 150 \u03bcM H~2~O~2~-treated cells. (**C**) Representative comets of Neutral comet assay measurement of DNA damage in control and H~2~O~2~-treated cells. (**D**) Percentage of DNA in the comet tails. Data are representative of three independent experiments, and at least 100 comets were scored for each time point. The box center lines show the medians; box limits indicate the 25th and 75th percentiles; whiskers extend to 5th and 95th percentiles; outliers are represented by dots; crosses represent sample means. (**E**) Adherent cell density using crystal violet in 25, 50 and 150 \u03bcM H~2~O~2~-treated cells. Values are normalized to the day of trigger. (**F**) Cell morphology at 72 h observed using phase contrast microscope. Cell shape and morphology at 72 h were zoomed and highlighted in the second column.](gkx684fig1){#F1}\n\nDNA damage response pathways are activated upon the induction of DNA damage by a sub-lethal oxidative stress {#SEC3-2}\n------------------------------------------------------------------------------------------------------------\n\nIn order to gain a better understanding of the origin of the second wave of DNA damage in cells exposed to 50 \u03bcM H~2~O~2~, L6 cells were synchronized using lovastatin, aphidicolin or thymidine before the addition of the oxidant. Independent of the approach used to synchronize the cells, a consistent S and G2/M phase extension in response to H~2~O~2~ compared to control cells was observed (Figure [2A](#F2){ref-type=\"fig\"} and\u00a0[Supplementary Figure S1A and C](#sup1){ref-type=\"supplementary-material\"}). Moreover, western blot analysis of \u03b3H2AX in lovastatin and aphidicolin synchronized cells or the detection of \u03b3H2AX foci in cells synchronized using thymidine block confirmed the induction of two waves of \u03b3H2AX in cells exposed to 50 \u03bcM H~2~O~2~ (Figure [2B](#F2){ref-type=\"fig\"} and\u00a0[C](#F2){ref-type=\"fig\"}; [Supplementary Figure S1B and D](#sup1){ref-type=\"supplementary-material\"}). Interestingly, note that cell cycle analysis of non-synchronized cells exposed to 50 \u03bcM H~2~O~2~ also showed that the second increase in \u03b3H2AX (at 12 h) corresponded to an increase in cells at the late S to G2/M boundary ([Supplementary Figure S9A and B](#sup1){ref-type=\"supplementary-material\"}). Moreover, in aphidicolin-synchronized cells exposed to a sub-lethal oxidative stress, detection of cyclin A1, a S-phase cyclin and the detection of cyclin B1, a G2\\\\M cyclin show an overlap between cyclin A1 and cyclin B expression. Indeed, cyclin A1 is detected up to 14 h and cyclin B1 from 9 to 16 h following cells exposure to the sub-lethal oxidative stress ([Supplementary Figure S1E and F](#sup1){ref-type=\"supplementary-material\"}). Taken together, these data confirmed that the second increase in \u03b3H2AX occurs at the S to G2/M boundary of the cell cycle ([Supplementary Figure S1A and C](#sup1){ref-type=\"supplementary-material\"}). To further support DNA damage as the origin of the two waves of \u03b3H2AX, detection of the p53-binding protein 1 (53BP1), a well-known DDR factor recruited to nuclear structures at the site of DNA damage and phosphorylation of p53 at ser15 were then assessed in cells exposed to 50 \u03bcM H~2~O~2~. Similar to the results obtained with the detection of \u03b3H2AX, two waves of 53BP1 foci were detected in cells synchronized with the thymidine block technique. The first wave of 53BP1 foci occurred 1 h after the sub-lethal redox stress and the second wave of 53BP1 foci was detected at 10 h after the exposure of the cells to the oxidant (Figure [2D](#F2){ref-type=\"fig\"} and\u00a0[E](#F2){ref-type=\"fig\"}). In addition, two waves of p53 expression and p53 phosphorylation at ser15 were also detected in cells exposed to 50 \u03bcM H~2~O~2~ (Figure [2F](#F2){ref-type=\"fig\"}). Finally, to assess if the ATM- or the ATR-dependent DDR pathways were activated at the time of the second wave of DNA damage, ataxia telangiectasia mutated (ATM) and ataxia telangiectasia and Rad3 (ATR) kinases were inhibited before \u03b3H2AX and p53 phosphorylation were detected. Figure [3A](#F3){ref-type=\"fig\"} shows that addition of the ATM inhibitor (KU55933) 3 h after the cells' exposure to H~2~O~2~ prevented the increase in \u03b3H2AX and p53 phosphorylation at 6, 9 and 12 h after the initial exposure of the cells to the oxidant (Figure [3A](#F3){ref-type=\"fig\"}). On the other hand, addition of the ATR inhibitor (AZ20) 2 h before each time point (6, 9 and 12 h) inhibited only the phosphorylation of p53 but not the increase in \u03b3H2AX (Figure [3B](#F3){ref-type=\"fig\"}). As a control for the specificity of the inhibitors, it is shown that phosphorylation of CHK1 (S345) was only inhibited by AZ20 (Figure [3B](#F3){ref-type=\"fig\"}) and not by KU55933 (Figure [3A](#F3){ref-type=\"fig\"}).\n\n![Sub-lethal oxidative stress induces two waves of \u03b3H2AX, 53BP1 foci and phosphorylated p53 (Ser15) in synchronized L6 cells. L6 cells were synchronized using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h and cell cycle analysis, \u03b3H2AX foci, 53BP1 foci, p53 expression and p53 phosphorylation were assessed. (**A**) Line plots capturing the percentage of cells in each cell cycle phase at the indicated time points. (**B**) Representative nucleus showing \u03b3H2AX foci in control and H~2~O~2~-treated cells. (**C**) Quantification of the average number of \u03b3H2AX foci per nuclei from the analysis of at least 30 to 40 nuclei. (**D**) Representative nucleus showing 53BP1 foci in control and H~2~O~2~-treated cells. (**E**) Quantification of the average number of 53BP1 foci per nuclei from the analysis of at least 30--40 nuclei. (**F**) Kinetics of phosphorylated p53 (ser15) and total p53 in control and 50 \u03bcM H~2~O~2~-treated cells. Plots shown in (C) and (E) are representative of at least two independent experiments and error bars represent \u00b1 SD of number of foci per nucleus counted within the experiment.](gkx684fig2){#F2}\n\n![ATM and ATR are activated after endogenous DNA damage induction. Cells were synchronized using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h. (**A**) Western blot analysis of the levels of \u03b3H2AX and phosphorylated p53 (ser15) at 6, 9 and 12 h following synchronized cells exposure to 50 \u03bcM H~2~O~2~ and the inhibition of ATM using KU55933 3 h after 50 \u03bcM H~2~O~2~ treatment. (**B**) Western blot analysis of \u03b3H2AX and phosphorylated p53 (ser15) after inhibition of ATR using AZ20 at the time of second wave of DNA damage (6--12 h). AZ20 was added 2 h before the collection of cells at the indicated time points.](gkx684fig3){#F3}\n\nThe second wave of \u03b3H2AX in cells exposed to sub-lethal oxidative stress is due to an unresolved replication stress leading to replication forks collapse and the conversion of single-strand DNA breaks to double-strand DNA breaks {#SEC3-3}\n------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nInterestingly, irrespective of the methods used to synchronize the cells, the second wave of \u03b3H2AX foci consistently corresponded with the cells' transition between the S and the G2/M phases of the cell cycle. Moreover, the overlap of cyclin A1 and cyclin B1 detected in aphidicolin-synchronized cells further supported that the second increase in \u03b3H2AX occurred at the late S phase--G2/M boundary suggested that the second wave of DNA damage might have originated in S phase before the cells rolled into G2/M. We, hence, reasoned that the second wave of DNA damage detected in cells exposed to 50 \u03bcM H~2~O~2~ might result from the generation of endogenous DNA damage as a result of unresolved replication stress ([@B9],[@B19],[@B37]--[@B41]). The RPA complex is an essential regulator of eukaryotic DNA metabolism. RPA avidly binds to ssDNA through multiple oligonucleotide/oligosaccharide-binding folds and coordinates the recruitment and exchange of genome maintenance factors to regulate DNA replication, recombination and repair. At the time of the second wave of DNA damage, we show a transient recruitment of the replication protein A32 (RPA32) at the chromatin (Figure [4A](#F4){ref-type=\"fig\"} and\u00a0[B](#F4){ref-type=\"fig\"}). Furthermore, the detection of RPA32 nuclear foci from 4 to 12 h exposure to 50 \u03bcM H~2~O~2~ and the co-localization of RPA32 and 53BP1 (shown by the Pearson's correlation ratio) support the presence of ssDNA and DSB at the time of the detection of the second wave of \u03b3H2AX (Figure [4C](#F4){ref-type=\"fig\"} and\u00a0[D](#F4){ref-type=\"fig\"}). Finally, incubation with aphidicolin, a DNA pol\u03b1 inhibitor, after 1 h of H~2~O~2~ treatment inhibited the second wave of \u03b3H2AX and the subsequent phosphorylation of p53 9 h after the initial oxidative stress. Together, these results provide evidence that the progression of the replication fork is required for the induction of the second wave of \u03b3H2AX (Figure [4E](#F4){ref-type=\"fig\"}).\n\n![Replicative stress induced by sub-lethal oxidative stress leads to the formation of DSB following replicative fork collapse. Cells were synchronized using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h. (**A**) Western blot showing the levels of RPA32 at the chromatin level at indicated time. (**B**) Densitometry plots of the chromatin level of RPA32 at 6, 10 and 24 h from (A). (**C**) Immunofluorescence images showing the levels of RPA32 foci in control and H~2~O~2~-treated cells in representative nucleus at indicated time points. (**D**) Co-localization analysis of 53BP1 and RPA32 foci. (**E**) Western blot of \u03b3H2AX and phosphorylated p53 (ser15) at 9 h, after addition of 10 \u03bcM aphidicolin (DNA pol\u03b1 inhibitor) 1 h after 50 \u03bcM H~2~O~2~ treatment.](gkx684fig4){#F4}\n\nSub-lethal oxidative stress induces a state of cellular senescence {#SEC3-4}\n------------------------------------------------------------------\n\nAs shown in Figure [1](#F1){ref-type=\"fig\"}, exposure of the cells to sub-lethal oxidative stress was associated with growth arrest and a senescent-like phenotype. Using an NMA that provides a direct and objective way of screening normal, senescent, apoptotic and nuclear irregularities that occur during failed mitosis or mitotic catastrophe, we show that 72 h following exposure to 50 \u03bcM H~2~O~2~, 25.5% of synchronized cells had a senescent nucleus, 1.4% of the nuclei were apoptotic and 2.6% showed a nucleus characteristic of mitotic catastrophe. Although this analytic tool predicted 69.1% of the cells to have normal nuclei following treatment with 50 \u03bcM H~2~O~2~, the average area of these nuclei was higher than the nuclei in control cells, indicating transition to the senescent phenotype (Figure [5A](#F5){ref-type=\"fig\"}). Indeed, passage to a senescent phenotype is corroborated by the flat and enlarged morphology of cells at 48 h, 120 h and 15 days after exposure to 50 \u03bcM H~2~O~2~ (Figure [5B](#F5){ref-type=\"fig\"}). Moreover, an increase in acidic SA-\u03b2-Gal activity (Figure [5C](#F5){ref-type=\"fig\"}) and activation of p16--RB (Figure [5D](#F5){ref-type=\"fig\"} and\u00a0[E](#F5){ref-type=\"fig\"}), but not the p21 axis ([Supplementary Figure S3A](#sup1){ref-type=\"supplementary-material\"}) and secretion of IL-6 (Figure [5F](#F5){ref-type=\"fig\"}), the most prominent cytokine of the SASPs was detected 48 h after the initial oxidative stress.\n\n![Exposure of L6 cells to 50 \u03bcM exogenous H~2~O~2~ induces senescence and senescence-associated phenotypic characteristics. Cells were synchronized using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h and senescence-associated characteristics were assessed. (**A**) NMA analysis of the nuclei of synchronized cells exposed to 50 \u03bcM H~2~O~2~ 72 h post-H~2~O~2~ treatment. Nuclei types and its corresponding phases are described as below. N---normal (interphase), S---small (mitosis), SR---small regular (apoptosis), SI---small irregular (mitotic abnormality), LR---large regular (senescence), LI---large irregular (mitotic catastrophe) and I---irregular (mitotic catastrophe). (**B**) Fifty micromolars H~2~O~2~-treated cells acquired enlarged cell shape and size. (**C**) C~12~FDG fluorescence staining of SA-\u03b2-Gal activity in synchronized cells exposed to 50 \u03bcM H~2~O~2~ at the indicated time points. (**D**) Western blot showing the kinetics of p16INK4a expression at indicated time points. (**E**) Western blot showing the kinetics of hypo-phosphorylated RB (pRB) and hyper-phosphorylated levels of RB (ppRB) upon H~2~O~2~ treatment. Hypo-phosphorylated RB was observed in 50 \u03bcM H~2~O~2~ from 12 h and sustained up to 48 h (**F**). Secreted IL-6 levels in the media were measured at 72 h following the cells exposure to 50 \u03bcM H~2~O~2~ using ELISA. The data shown in (C) and (F) represent the mean of at least three independent experiments and the error represents \u00b1 SD. *P*-value (one-tailed, Student's *t*-test, \\**P*-value \\< 0.05, \\*\\*\\**P*-value \\< 0.001, ns---not significant) calculation is based on the comparison between the respective time points as indicated in the figures.](gkx684fig5){#F5}\n\nThe cellular senescent state induced by sub-lethal oxidative stress is characterized by the formation of persistent DNA damage foci {#SEC3-5}\n-----------------------------------------------------------------------------------------------------------------------------------\n\nSevere or irreparable DNA damage causes cells to senesce with persistent DNA damage foci. These persistent changes precede the establishment of senescence-associated characteristics, including growth arrest ([@B20],[@B21],[@B42]) and SASP ([@B3],[@B43]). In order to show that the senescent state induced by sub-lethal oxidative stress is also associated with these changes, the presence of persistent \u03b3H2AX and 53BP1 foci at 24, 48 and 72 h after the initial cells exposure to 50 \u03bcM H~2~O~2~ was assessed. Results show that persistent \u03b3H2AX and 53BP1 foci were detected in cells exposed to sub-lethal oxidative stress (Figure [6A](#F6){ref-type=\"fig\"}). Persistent DNA damage foci/DNA-SCARS have been shown to exhibit promyelocytic leukemia (PML) nuclear bodies at their periphery and the absence of DNA repair protein Rad51 ([@B21]). Indeed, Figure [6B](#F6){ref-type=\"fig\"} shows the co-localization of PML nuclear bodies at the periphery of \u03b3H2AX foci 72 h following exposure to the oxidant. In addition, persistent \u03b3H2AX foci concomitant with a decrease in Rad51 foci (Figure [6D](#F6){ref-type=\"fig\"}--[F](#F6){ref-type=\"fig\"}) and repression of Rad51 expression at the whole chromatin level (Figure [6C](#F6){ref-type=\"fig\"}) were detected 72 h after exposure of the cells to H~2~O~2~.\n\n![Exposure of L6 cells to 50 \u03bcM exogenous H~2~O~2~ induced senescence-associated persistent foci/DNA-SCARS. Cells were synchronized in late G1 using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h and the appearance of senescence-associated persistent foci/DNA-SCARS assessed. (**A**) \u03b3H2AX and 53BP1 foci after 24, 48 and 72 h post-H~2~O~2~ treatment. (**B**) Immunofluorescence images showing the co-localization of PML and \u03b3H2AX 96 h after exposure to H~2~O~2~. (**C**) Total p53, Rad51 and \u03b3H2AX protein expression levels in a chromatin extract 72 h after cells exposure to 50 \u03bcM H~2~O~2~. Total histone H3 is used as a loading control. (**D**) Rad51 and \u03b3H2AX foci in control and 50 \u03bcM H~2~O~2~-treated cells at indicated time points. Representative immunofluorescence images of Rad51 and \u03b3H2AX foci in a nucleus are shown. Line plots showing the average of Rad51 foci (**E**) and \u03b3H2AX foci (**F**) from at least 30--40 nuclei at indicated time points are also shown. Error bars represent \u00b1 SD of number of foci per nucleus counted within the experiment.](gkx684fig6){#F6}\n\nThe cellular senescent state induced by sub-lethal oxidative stress is characterized by an increase in HMGA2 expression and formation of CCF-like micronuclei {#SEC3-6}\n-------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nHigh-level expression of HMGA2 is sufficient to induce SAHF formation, an important step leading to the deepening of the senescent phenotype, as SAHF sequesters genes controlling proliferation and cell cycle, thus reinforcing permanent cell cycle arrest ([@B26],[@B28],[@B44]). As shown in Figure [7A](#F7){ref-type=\"fig\"}, cells exposure to 50 \u03bcM H~2~O~2~ induces an increase in HMGA2 expression. Furthermore, extrusion of chromatin structures in the form of micronuclei staining positive for \u03b3H2AX and Histone H3 were also observed in cells treated with 50 \u03bcM H~2~O~2~ (Figure [7B](#F7){ref-type=\"fig\"}--[D](#F7){ref-type=\"fig\"}). Interestingly, an increase in the chromatin structures observed in the cytoplasm did not stain for 53BP1 (Figure [7E](#F7){ref-type=\"fig\"}), but were positive for Lamin A\\\\C (Figure [7F](#F7){ref-type=\"fig\"}), supporting these structures as micronuclei.\n\n![Exposure of L6 cells to 50 \u03bcM exogenous H~2~O~2~ induces HMGA2 and extrusion of chromatin as micronuclei structures. Cells were synchronized in late G1 using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h and then observed for senescence-associated chromatin changes. (**A**) Western blot images showing the HMGA2 protein levels. (**B**) The number of nucleus with extruded chromatins are counted manually and percentage of cells with chromatin extrusions in control and H~2~O~2~-treated cells are shown in (A). The data shown represent the mean of at least three independent experiments and the error bars represent \u00b1 SD. (**C** and **D**) Immunofluorescence images showing the blebbing of chromatin from nucleus to cytoplasm at 72 h. Arrows indicate the blebbing and extruded chromatin. Presence of chromatin in CCFs is confirmed by occurrence of (C) histone H3 and (D) \u03b3H2AX in control and H~2~O~2~-treated cells at 72 h. (**E**) Absence of 53BP1 staining in extruded chromatin in 50 \u03bcM H~2~O~2~-treated cells at 72 h. (**F**) Presence of Lamin A\\\\C staining in extruded chromatin in 50 \u03bcM H~2~O~2~-treated cells at 72 h.](gkx684fig7){#F7}\n\nAbsence of p53 mitigates growth arrest and prevents the decrease in Rad51 and Lamin B1 expression {#SEC3-7}\n-------------------------------------------------------------------------------------------------\n\nIn the absence of p53, the percentage of cells showing signs of nuclear senescence and morphology, assessed by NMA, decreased significantly (Figure [8A](#F8){ref-type=\"fig\"}) and cell growth arrest was mitigated (Figure [8B](#F8){ref-type=\"fig\"} and\u00a0[C](#F8){ref-type=\"fig\"}). Silencing p53 also prevented the repression of Rad51 protein (Figure [9A](#F9){ref-type=\"fig\"} and\u00a0[B](#F9){ref-type=\"fig\"}) as well as prevented the repression of Rad51 mRNA (Figure [9C](#F9){ref-type=\"fig\"}). In addition, in agreement with the repression of Rad51 expression downstream of the second wave of DNA damage ([Supplementary Figure S7A](#sup1){ref-type=\"supplementary-material\"}), inhibition of ATM 3 h after first wave of DNA damage prevented the phosphorylation of p53 and rescue of Rad51 expression at 9 and 12 h post-H~2~O~2~ treatment ([Supplementary Figure S7B](#sup1){ref-type=\"supplementary-material\"}). Interestingly, absence of p53 did not affect the appearance of the second wave of \u03b3H2AX and the S phase delay despite the maintenance of Rad51 expression ([Supplementary Figure S8A and B](#sup1){ref-type=\"supplementary-material\"}). Finally, silencing p53 expression prevented the repression of Lamin B1 protein and mRNA expression observed 6 h after cells exposure to 50 \u03bcM H~2~O~2~ (Figure [9D](#F9){ref-type=\"fig\"}--[F](#F9){ref-type=\"fig\"}) and was accompanied by a decrease of Lamin B1 in the perinuclear region at 24 and 48 h (Figure [9G](#F9){ref-type=\"fig\"}). On the other hand, a significant increase in genomic instability, as seen by an increase in micronuclei and aneuploid cells, was detected in p53 null cells exposed to 50 \u03bcM H~2~O~2~, compared to control cells ([Supplementary Figure S4A--D](#sup1){ref-type=\"supplementary-material\"}). Interestingly, the absence of p53 did not have a notable effect on the SA-\u03b2-Gal activity ([Supplementary Figure S4E](#sup1){ref-type=\"supplementary-material\"}) and HMGA2 expression ([Supplementary Figure S4F](#sup1){ref-type=\"supplementary-material\"}).\n\n![Endogenous DNA damage induced p53-dependent senescence growth arrest. Cells were synchronized with thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h. (**A**) NMA analysis of the nuclei of untreated and 50 \u03bcM H~2~O~2~-treated synchronized cells at 72 h in CoSi and Sip53-transfected cells. Nuclei types and its corresponding phases are described as below. N---normal (interphase), S---small (mitosis), SR---small regular (apoptosis), SI---small irregular (mitotic abnormality), LR---large regular (senescence), LI---large irregular (mitotic catastrophe) and I---irregular (mitotic catastrophe). (**B**) Phase contrast images showing the cell morphology of untreated and 50 \u03bcM H~2~O~2~-treated cells in CoSi and Sip53-transfected cells 48 h following cells exposure to the oxidant. (**C**) Cells were stained with crystal violet to quantify adherent cell density of untreated and 50 \u03bcM H~2~O~2~-treated cells at 48 h upon silencing with CoSi and Sip53. Values are normalized to the day of trigger. The data shown in (C) represent the mean of at least three independent experiments and the error represents \u00b1 SD. *P*-value (one-tailed, Student's *t*-test, \\*\\**P*-value \\< 0.01, ns---not significant) of the comparison groups are indicated in the figures.](gkx684fig8){#F8}\n\n![p53 transcriptionally repressed Rad51 and Lamin B1. Cells were synchronized using thymidine before being treated with 50 \u03bcM H~2~O~2~ for 1 h. (**A**) Western blot showing Rad51expression in CoSi and Sip53-transfected cells. (**B**) Rad51 protein expression at 24 and 48 h following cells exposure to 50 \u03bcM H~2~O~2~. (**C**) Rad51 mRNA expression in CoSi and Sip53-transfected cells 24 and 48 h after cells exposure to 50 \u03bcM H~2~O~2~. Data shown are the representative of four independent experiments and the error represents \u00b1 SD. (**D**) Western blot showing Lamin B1 expression in CoSi and Sip53-transfected cells 24 and 48 h following cells exposure to 50 \u03bcM H~2~O~2~. (**E**) Lamin B1 mRNA expression at 24 and 48 h upon silencing p53. Data shown are the representative of four independent experiments and the error represents \u00b1 SD. (**F**) Western blot showing Lamin B1 expression at indicated time points in 50 \u03bcM H~2~O~2~-treated cells. (**G**) Depletion of Lamin B1 in H~2~O~2~-treated cells at indicated time points, shown by immunofluorescence. White boxed nucleus are magnified to show the intensity of Lamin B1 at various time points indicated (right column). The Lamin B1 fluorescence intensity along the diameter of the nucleus is also quantified. Note the gradual decrease of Lamin B1 fluorescence intensity in 50 \u03bcM H~2~O~2~-treated cells (right bottom column).](gkx684fig9){#F9}\n\nTwo waves of DNA damage followed by cellular senescence are not unique to L6 cells and sub-lethal oxidative stress {#SEC3-8}\n------------------------------------------------------------------------------------------------------------------\n\nSo far, our results support that endogenous DNA damage is responsible for the induction of cellular senescence in L6 cells exposed to sub-lethal oxidative stress. [Supplementary Figure S5A and B](#sup1){ref-type=\"supplementary-material\"} shows that similar to what was observed in L6 cells, exposure of synchronized RPE1 cells to sub-lethal oxidative stress induced a prolonged S phase, two waves of DNA damage with the second wave coinciding with the transition between the S and the G2/M phase of the cell cycle ([Supplementary Figure S5C and D](#sup1){ref-type=\"supplementary-material\"}), and establishment of cellular senescence detected by an increase in SA-\u03b2-Gal activity ([Supplementary Figure S5E](#sup1){ref-type=\"supplementary-material\"}). Finally, supporting that other inducers of DNA damage that do not induce cell death might activate cellular senescence as seen with sub-lethal oxidative stress, we show that exposure of L6 cells to sub-lethal concentration of DNA topoisomerase II inhibitor, etoposide coincided with two waves of DNA damage and the induction of cell growth arrest ([Supplementary Figure S6A--C](#sup1){ref-type=\"supplementary-material\"}).\n\nDISCUSSION {#SEC4}\n==========\n\nThe present report shows that exposure of cells to sub-lethal oxidative stress induces two waves of \u03b3H2AX and a cellular senescence state that is detected 48 h after the initial oxidative stress. While the formation of DSB is responsible for the two waves of \u03b3H2AX, the induction of the senescent state is supported by changes in cells morphology, an increase in SA-\u03b2-Gal activity, establishment of DNA-SCARS, formation of CCF-like micronuclei, activation of the p16--RB axis, an increase in HMGA2 expression as a marker of SAHF and secretion of IL-6 as one of the SASP cytokines.\n\nOrigin of the initial wave of \u03b3H2AX {#SEC4-1}\n-----------------------------------\n\nIt is well accepted that H~2~O~2~ can react with Fe^2+^ leading to base alterations that can lead to DNA damage. The two basic groups of complex DNA damage are DSBs and non-DSB oxidative clustered DNA lesion (OCDL). OCDLs are formed due to closely spaced lesions in the form of abasic sites, oxypyrimidines, oxypurines and single strand breaks (SSBs) ([@B45],[@B46]). Interestingly, OCDLs can be converted to DSBs during the repair process when both strands are incised simultaneously in close proximity. In particular, activation of the base excision repair (BER) pathway has been shown to convert OCDLs to DSBs ([@B45]). Although in the present report we do not extensively address the nature of the first wave of DNA damage, we propose that similar to what has recently been shown by Sharma and colleagues ([@B45]) in DT40 cells exposed to 40 \u03bcM H~2~O~2~, the first wave of \u03b3H2AX detected after the exposure of L6 cells to 50 \u03bcM H~2~O~2~ can be explained by the conversion of OCDLs to DSBs. In support of this proposal, incubation of L6 cells with iron chelators and ROS scavengers prevented the rapid increase in \u03b3H2AX and the subsequent second wave of \u03b3H2AX after cells exposure to the oxidant ([Supplementary Figure S2A and B](#sup1){ref-type=\"supplementary-material\"}). In addition, formation of SSBs is shown by the activation of Poly (ADP-ribose) polymerase (PARP) detected by an increase in Poly (ADP-ribose) (PAR) 1 h after the exposure of cells to 50 \u03bcM H~2~O~2~ ([Supplementary Figure S2C](#sup1){ref-type=\"supplementary-material\"}). Finally, silencing *N*-methylpuridine DNA glycolsylase 1 (MPG1), one of the initiators of the BER pathway, partially reduced the initial H2AX phosphorylation after the cells exposure to 50 \u03bcM H~2~O~2~ ([Supplementary Figure S2D](#sup1){ref-type=\"supplementary-material\"}). Partial reduction of the initial \u03b3H2AX upon silencing MPG1 also reduced the second wave of phosphorylated H2AX ([Supplementary Figure S2D](#sup1){ref-type=\"supplementary-material\"}). Taken together, these results support that the first wave of \u03b3H2AX drives the onset of the second wave of \u03b3H2AX and the cells' growth arrest that follows.\n\nFrom an exogenous to an endogenous induction of DNA damage {#SEC4-2}\n----------------------------------------------------------\n\nUnlike other DNA damaging agents, H~2~O~2~ induces a variety of complex DNA ends including oxidized deoxyribose, DSBs, SSBs and OCDLs that need to be repaired to prevent potential replication stress when cells enter the replicative phase of the cell cycle. Contrary to the DNA damage induced by 25 \u03bcM H~2~O~2~ that was rapid and completely repaired only leading to a slight growth delay, the initial DNA damage induced by 50 \u03bcM H~2~O~2~ was only partially repaired. The partial repair of DNA damage following cells exposure to 50 \u03bcM H~2~O~2~ is first supported by the data showing that \u03b3H2AX does not go back to control level before the second wave is detected. Similarly, the number of 53BP1 foci detected in synchronized cells following exposure to 50 \u03bcM H~2~O~2~ stayed slightly higher than in control cells at the time preceding the second wave of \u03b3H2AX. Moreover, a delay in S phase is detected in synchronized cells exposed to 50 \u03bcM H~2~O~2~. Taken together, these data provide testimony that cells entered the S phase of the cell cycle with unrepaired DNA. Furthermore, despite the activation of the DDR pathways (including the ATR and ATM kinases), the inability to resolve the stress resulted in the stalling of the replication fork, replication fork collapse and the formation of DSB. Of note, the detection of RPA32 nuclear foci from 4 to 12 h in cells exposed to 50 \u03bcM H~2~O~2~ and presence of DSBs at the time of the detection of the second wave of \u03b3H2AX, as shown by the co-localization of RPA32 and 53BP1, demonstrate that replication fork collapse leading to DSB is the origin of the second wave of \u03b3H2AX upon sub-lethal oxidative stress. Furthermore, the second wave of \u03b3H2AX in cells exposed to sub-lethal oxidative stress corresponded with the junction between the extended S phase and entry into the G2/M phase of the cell cycle. The extended S phase is in agreement with replication stress and the extended G2/M suggests activation of checkpoints to prevent cells carrying DSB from entering mitosis. The latter is evidenced by the absence of histone H3 phosphorylation at ser10 in cells exposed to H~2~O~2~ ([Supplementary Figure S3D](#sup1){ref-type=\"supplementary-material\"}). In addition, the number of cells in metaphase was significantly reduced upon exposure to 50 \u03bcM H~2~O~2~, compared to untreated cells ([Supplementary Figure S3B and C](#sup1){ref-type=\"supplementary-material\"}). Hence, the clear demonstration of a replication stress and absence of mitotic entry as well as the necessity for replication fork progression for the formation of the second wave of \u03b3H2AX support that unresolved replication stress led to replication forks collapse.\n\nSub-lethal oxidative stress induces the formation of DNA-SCARS {#SEC4-3}\n--------------------------------------------------------------\n\nDespite the recognition that acquisition of persistent DNA damage foci such as DNA-SCARS is a key feature of the decision to enter cellular senescence ([@B5],[@B16],[@B21]), no study has previously linked replication stress-induced DNA damage to DNA-SCARS formation. Formation of DNA-SCARS is accelerated, when the DNA repair proteins are inhibited, leading to an ineffective and incomplete DNA repair process. Rad51 is one such important protein that is absent from the DNA-SCARS structures. Interestingly, upon cells exposure to sub-lethal oxidative stress, we show that the expression of Rad51 is inhibited by p53. Moreover, the decrease in Rad51 resulted in the appearance of \u03b3H2AX foci without Rad51. DNA-SCARS also lack ssDNA-binding proteins like RPA, but harbor PML at the site of DNA damage foci. Both characteristics were detected in cells exposed to 50 \u03bcM H~2~O~2~. Taken together, the existence of 53BP1 and \u03b3H2AX foci without Rad51, repression of total RPA32 and chromatin level of RPA32, and localization of PML bodies at the periphery of DNA damage foci are in agreement with the formation of DNA-SCARS in response to 50 \u03bcM H~2~O~2~.\n\nSub-lethal oxidative stress and the extrusion of CCF-like micronuclei {#SEC4-4}\n---------------------------------------------------------------------\n\nMicronuclei structures induced by replication stress inducers contain non-repaired or mis-repaired DSBs ([@B47],[@B48]). Increased micronuclei structures staining positive for \u03b3H2AX and Histone H3 were observed upon treatment with 50 \u03bcM H~2~O~2~. Interestingly, the increased chromatin structures observed in the cytoplasm did not stain for 53BP1, but were positive for Lamin A\\\\C. These data suggest that the cytoplasmic chromatin structures detected in cells exposed to a sub-lethal redox stress resemble micronuclei, as reported previously ([@B17],[@B47],[@B48]).\n\nSub-lethal oxidative induces a p16--RB-dependent cellular senescence {#SEC4-5}\n--------------------------------------------------------------------\n\nThe p53 target gene, p21, has often been considered critical for establishing senescence, whereas p16 may be more involved in the maintenance of the phenotype ([@B7],[@B49],[@B50]). In agreement with the role of p53 in the establishment of senescence by sub-lethal oxidative stress, two peaks of increase in p53 expression and p53 phosphorylation were detected in cells treated with 50 \u03bcM H~2~O~2~. Moreover, silencing p53 expression mitigated the cells growth arrest and the percentage of cells with senescent nuclei and morphology abnormality. However, the induction of senescence by exogenous H~2~O~2~ was not associated with a p53--p21 axis but with a p16--RB axis. It is interesting to note that the accumulation of p16 in tissues is implicated in cellular aging and clearance of p16-positive cells delays aging ([@B51]). Hence, the ability of a sub-lethal oxidative stress to induce a p16-dependent cellular senescence might explain the strong correlation between ROS and the development of an aging phenotype at both the cellular as well as at the organism level ([@B52]).\n\nSenescence induced by sub-lethal oxidative stress requires the activation of p53--Rad51 and p53--Lamin B1 feedback loops {#SEC4-6}\n------------------------------------------------------------------------------------------------------------------------\n\nThere is a plethora of genes activated by p53 that are associated with the establishment of the cellular senescent state ([@B22]). On the other hand, association of the transcriptional repression capacity of p53 and a cellular senescent phenotype is not well documented. In the present report, we show that following cells exposure to sub-lethal oxidative stress, p53 was responsible for the decrease in Rad51 and Lamin B1 expression. Rad51 is a key protein of the HR pathway involved in resolving replication stress ([@B53]). Silencing p53 rescued not only the growth arrest but also the expression of Rad51 in cells exposed to 50 \u03bcM H~2~O~2~. Knowing that absence of Rad51 is a characteristic of DNA-SCARS ([@B21]), we propose that upon a sub-lethal oxidative stress the repression of Rad51 by active p53 is critical for the formation of DNA-SCARS. In addition, reciprocally, DNA-SCARS formation is critical to maintain the low level of p53 activation necessary to maintain cell cycle arrest in senescent cells ([@B21]). On the other hand, Lamin B1 is a robust and easily detectable marker of senescence. Silencing Lamin B1 affects the proliferation rate of the cells ([@B54]) and contributes to chromatin extrusion ([@B14],[@B17]). Activation of p53 or p16 has been shown to result in loss of Lamin B1 in senescent cells ([@B14],[@B54]). However, in the present report repression of Lamin B1 expression was observed before the detection of an increase in p16 and the chromatin extrusion. Hence, we propose that in cells exposed to a sub-lethal oxidative stress, aggregation of DSB into micronuclei like structures or the formation of chromatin budding is initiated by the unresolved replication stress, while the p53-dependent repression of Lamin B1 might then facilitate the translocation of chromatin via micronuclei like structures to the cytoplasm. Furthermore, since Lamin B1 downregulation has been shown to activate p53 ([@B54]) and the expression of Lamin B1 and p53 levels were constitutively low in cells exposed to sub-lethal oxidative stress, similar to the p53--Rad51 feedback loop, we propose the establishment of a feedback mechanism between p53 and Lamin B1 to be critical to maintain growth arrest. Following the establishment of the p53--Rad51 and p53--Lamin B1 feedback loops, cells will then enter a deep senescent state, a key event to prevent the formation of tumorigenic cells. Indeed, in the absence of p53, cells exposed to 50 \u03bcM H~2~O~2~ showed severe genomic instability, aneuploidy and micronuclei formation ([Supplementary Figure S4B--D](#sup1){ref-type=\"supplementary-material\"}). Interestingly, the absence of p53 did not have a notable effect on the SA-\u03b2-Gal activity and HMGA2 expression compared to control cells. These results suggest that p53-independent pathways might also be important to complete the redox stress-induced senescent phenotype. In particular, it might be interesting to better understand what is activating the formation of SAHF. Regarding the increase in SA-\u03b2-Gal, we have recently shown that sub-lethal oxidative stress increases lysosome biogenesis through a caspase 3/TFEB pathway ([@B55]). It could be interesting to decipher whether the increase in lysosome biogenesis could be linked to the increase in SA-\u03b2-Gal.\n\nCONCLUSION {#SEC5}\n==========\n\nOur study contributes to the mechanistic understanding of the induction of cellular senescence following cells' exposure to sub-lethal oxidative stress (Figure [10](#F10){ref-type=\"fig\"}). Interestingly, the same pathways might be activated upon cells exposure to sub-lethal concentration of chemotherapeutic drug such as etoposide. In addition, evidence from the literature suggest that oncogenic stimuli that do not induce direct DNA damage also mediate senescence via replication stress-induced endogenous DNA damage ([@B56],[@B57]). Hence, activation of cellular senescence due to the formation of endogenous DSB could be a pathway common to a variety of inducers of cellular senescence. Finally, establishing the role of replication stress induced endogenous DNA damage in the initiation and the establishment of cellular senescence following cells exposure to a sub-lethal oxidative stress is critical for a better understanding of the molecular mechanisms/dynamics involved in the induction of aging and cancer by ROS.\n\n![Proposed model of the induction of cellular senescence by a sub-lethal oxidative stress. We propose that the acquisition of cellular senescence following cells' exposure to a sub-lethal oxidative stress occurs in three phases. **Phase 1** (1--3 h): upon exposure of cells to a sub-lethal oxidative stress, variety of complex DNA ends including oxidized deoxyriboses, DSBs, SSBs and OCDLs are generated. **Phase 2** (4--12 h): despite the presence of unrepaired DNA, cells progress to S phase. Upon the progression to S phase, presence of unrepaired DNA damage induces replication stress. The inability to resolve the stress results in the stalling of the replication fork, replication fork collapse leading to the formation of DSB. At the same time activation of p53 inhibits the transcription of the HR repair protein Rad51 and the nuclear protein Lamin B1. **Phase 3** (12--72 h): repression of Rad51 by active p53 is critical for the formation of DNA-SCARS. In addition, reciprocally, DNA-SCARS formation is critical to maintain the low level of p53 activation necessary to maintain cell cycle arrest. On the other hand, aggregation of DSB into micronuclei like structures or the formation of chromatin budding initiated by the unresolved replication stress associated with the p53-dependent repression of Lamin B1 facilitates the translocation of chromatin via micronuclei like structures to the cytoplasm. Furthermore, similar to the p53--Rad51 feedback loop, a feedback mechanism between p53 and Lamin B1 is critical to maintain growth arrest. Following the establishment of the p53--Rad51 and p53--Lamin B1 feedback loops, cells will then enter a deep senescent state, characterized by growth arrest, an increase in SA-\u03b2-Gal activity, activation of a p16-RB axis, establishment of DNA-SCARS, increase in HMGA2 expression as a marker of SAHF, formation of CCF-like micronuclei and secretion of IL-6.](gkx684fig10){#F10}\n\nSupplementary Material\n======================\n\n###### \n\nClick here for additional data file.\n\nWe would like to thank Dr Kathirvel Paramasivam and Ms Vishnupriya Manivannan for their help in performing experiments to complete the necessary revisions of the original manuscript.\n\nSUPPLEMENTARY DATA {#SEC6}\n==================\n\n[Supplementary Data](#sup1){ref-type=\"supplementary-material\"} are available at NAR Online.\n\nFUNDING {#SEC7}\n=======\n\nMinistry of Education (MOE), Singapore (An AcR tier 1 FRC Grant) \\[T1--2013 to M.V.C.\\]; Biomedical Research Council of A\\*STAR (Agency of Science Technology and Research) (to U.S.); NGS Ph.D. Research Scholarship, National University of Singapore (to G.V.). Funding for open access charge: MOE, Singapore (AcR tier 1 FRC) (to M.V.C.).\n\n*Conflict of interest statement*. None declared.\n"}
+{"text": "Background {#Sec1}\n==========\n\nTraumatic brain injury (TBI) is a leading cause of death and disability around the world \\[[@CR1], [@CR2]\\]. In the USA, the prevalence of TBI is estimated to be 2\u00a0% in the general population \\[[@CR3]\\]. The mortality rate was reported to be 18.4 per 100,000 persons with annual average of 53,014 deaths \\[[@CR4]\\]. A report from the Canadian Institute for Health Information (CIHI) report indicated that there were 16,811 hospitalizations annually for TBI with 1368 (8\u00a0%) related deaths \\[[@CR5]\\]. Among residents in a large Canadian health region, the annual incidence of severe TBI was 11.4 per 100,000 persons with a mortality rate of 5.1 per 100,000 persons per year \\[[@CR6]\\].\n\nClinical practice guidelines are developed to improve quality of care decrease discrepancy in practice and ensure that evidence is followed \\[[@CR7]\\]. Mostly, these guidelines are developed and distributed by well-recognized organizations. A guideline consists of systematically developed recommendations to guide practitioners in choosing the appropriate health care decision for specific clinical circumstances \\[[@CR8]\\]. A guideline recommendation is defined as \"any statement that promotes or advocates a particular course of action in clinical care\" \\[[@CR9]\\]. In the treatment of TBI, guidelines are proposed to be an important aspect of patient management.\n\nThere are several published guidelines in the management of TBI from different countries. These guidelines target different aspects of TBI management including management during pre-hospital at the emergency department, in-hospital and intensive care unit, indications for surgical management, and computed tomography (CAT) scan of the head \\[[@CR10]--[@CR14]\\].\n\nInternationally, Brain Trauma Foundation (BTF) guidelines are widely disseminated. They have been translated into over 15 different languages and applied in Europe, South America, and some parts of China \\[[@CR12]\\]. The BTF maintains and revises several TBI guidelines on an approximate 5-year cycles, including *Guidelines for Prehospital Management of Traumatic Brain Injury*, *Guidelines for the Management of Severe Traumatic Brain Injury*, *Guidelines for the Surgical Management of Traumatic Brain Injury*, *Guidelines for the Acute Medical Management of Severe Traumatic Brain Injury in Infants, Children, and Adolescents*, and *Guidelines for the Field Management of Combat Related Head Trauma and Early Indicators of Prognosis of Severe Traumatic Brain Injury*. These guidelines are developed and maintained through a collaborative agreement with the American Association of Neurological Surgeons (AANS) and the Congress of Neurological Surgeons (CNS), and in collaboration with the AANS/CNS Joint Section on Neurotrauma and Critical Care, European Brain Injury Consortium, and other stakeholders in TBI patient outcome \\[[@CR12]\\].\n\n*Guidelines for Management of Severe Traumatic Brain Injury* addresses key topics useful for in-hospital medical management of severe TBI in adult patients with a Glasgow Coma Scale (GCS) score of 3--8. These include blood pressure and oxygenation hyperosmolar therapy, prophylactic hypothermia, infection prophylaxis, deep vein thrombosis prophylaxis, intracranial pressure monitoring, cerebral perfusion thresholds, brain oxygen monitoring and thresholds, anesthetics, analgesics and sedatives, nutrition, antiseizure prophylaxis, and hyperventilation through steroids use. In 2007, the third edition of these *Guidelines* was released following the first and second editions in 1995 and 2000. \\[[@CR12], [@CR15], [@CR16]\\].\n\n*Guidelines for the Surgical Management of Traumatic Brain Injury* addresses acute surgical management of TBI including acute epidural and subdural hematomas, parenchymal mass lesions, depressed skull fractures through posterior fossa lesions with focus on indications, technique, and timing of surgery. These *Guidelines* were published in 2006 \\[[@CR13]\\].\n\nStudies suggest that implementation and strict adherence to BTF guidelines results in improvement in the neurological outcomes and reduction in mortality from severe traumatic brain injury \\[[@CR17], [@CR18]\\]. However, there is still significant variability and inconsistency in the management of traumatic brain injury patients \\[[@CR19], [@CR20]\\]. This review will be the first systematic review assessing the adherence to BTF guidelines and its effect on outcome.\n\nObjectives {#Sec2}\n----------\n\nThe first objective of this study is to present a systematic review of adherence by practitioners to the BTF guidelines for the management of severe TBI. The second objective is to explore the factors influencing adherence to the guidelines. Identification of these factors may provide valuable insight into the development of strategies to increase the adherence. The third objective is to study the outcome of guideline-based management in comparison to non-guideline based management to determine the effectiveness of these guidelines.\n\nMethods/design {#Sec3}\n==============\n\nProtocol and study overview {#Sec4}\n---------------------------\n\nMethods of this systematic review and meta-analysis have been developed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) \\[[@CR21]\\] and the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines \\[[@CR22]\\]. We will begin by developing a comprehensive database containing all published literature that addresses adherence to BTF guidelines in the management of severe TBI. This protocol has been registered in the PROSPERO International Prospective Register of Systematic Reviews (ID: [CRD42015017794](http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42015017794)).\n\nSelection criteria {#Sec5}\n------------------\n\n### Population {#Sec6}\n\nThe population of interest will include adult (\u226518\u00a0years old) hospitalized patients with blunt TBI. Whenever outcome measures are available, the patients who were treated based on the BTF guideline will be compared to the patients who were not treated based on the BTF guideline. Additionally, the population of this study will include the practitioners, mainly the neurosurgeons and critical care physicians, who will be assessed for adherences to guidelines. The assessed guidelines will be (a) in-hospital guidelines regarding blood pressure and oxygenation, hyperosmolar therapy, prophylactic hypothermia, infection prophylaxis, deep vein thrombosis prophylaxis, indications for intracranial pressure monitoring, intracranial pressure monitoring technology, intracranial pressure thresholds, cerebral perfusion thresholds, brain oxygen monitoring and thresholds, anesthetics, analgesics, sedatives, nutrition, antiseizure prophylaxis, hyperventilation, and steroids. (b) Guidelines for surgical management for acute epidural and subdural hematomas, parenchymal lesions, posterior fossa mass lesions, and depress cranial fractures. We will exclude (1) studies addressed adherence to pre-hospital guidelines (the result from studies on pre-hospital management may not reflect the adherence because failure to achieve target recommendation may be an indicator of severe injury), (2) studies focused on military/combat-related TBI, because the results would not be generalizable to the source population of civilian patients with TBI, and (3) studies with majority of pediatric patients.\n\n### Outcome {#Sec7}\n\nThe main outcome will be the adherence rate with BTF guidelines. In addition, we will identify factors influencing the adherence to the BTF guidelines. The effectiveness of adherence with the BTF guidelines on several clinical outcomes will be assessed. The measured clinical outcomes will include mortality (ICU, in-hospital mortality) and morbidity (Glasgow Outcome Scale (GOS), Modified Rankin Scale (MRS), ventilation days, ICU stay, and hospital stay).\n\n### Study design {#Sec8}\n\nOriginal searches will include RCT cohort, case-control, and case series. We will exclude studies that included fewer than ten patients.\n\n### Search strategy {#Sec9}\n\nThe primary search strategy was developed by the primary investigator (YK) and in collaboration with an expert searcher/librarian (SC). We will search the following electronic bibliographic databases: PROSPERO Medline (OVID), EMBASE (OVID), EBM Reviews---Cochrane Database of Systematic Reviews, EBM Reviews---ACP Journal Club, EBM Reviews---Database of Abstracts of Reviews of Effects, EBM Reviews---Cochrane Central Register of Controlled Trials, EBM Reviews---Cochrane Methodology Register, EBM Reviews---Health Technology Assessment, EBM Reviews---NHS Economic Evaluation Database, CINAHL Plus with Full Text, ProQuest Dissertations and Theses Full-text, SCOPUS, and Google Scholar using both controlled vocabulary (e.g., EMTREE and MeSH) and keywords to retrieve concepts including Brain Trauma Foundation or brain injur\\* and guideline\\* and adhere\\*. Searches will be limited to adult patients in non-military settings. Animal studies will be excluded. This systematic review will include searching gray literature, reviewing references lists, and contacting experts in the field. (See appendix in Additional file [1](#MOESM1){ref-type=\"media\"} for the final proposed MEDLINE, EMBASE, and EBM Reviews---Cochrane Database of Systematic Reviews search strategy).\n\n### Study selection {#Sec10}\n\nTwo investigators (YK and IG) will independently screen all title abstracts and articles to identify study meeting inclusion/exclusion criteria. Inclusion disagreement will be discussed and resolved by consensus or arbitration by other researchers (CO and DZ).\n\nData extraction {#Sec11}\n---------------\n\nTwo investigators (YK and IG) will independently extract data from eligible studies using a pre-designed and pilot-tested standardized electronic data extraction form. We will extract data on (1) publication details (year and language of publication name of the publishing journal and country in which the study was conducted). (2) Design: type of study (RCT, cohort, case-control, case series), study temporality (prospective, retrospective). (3) Study participant details: patient characteristics (age, sex, GCS, Injury severity score). (4) Data for percentage adherence to BTF guidelines. From each article, adherence percentages for each recommendation will be extracted. In case of a pre- and post-intervention design for evaluation of intervention (for example introducing a protocol or teaching program), only the post-intervention percentages will be extracted because our interest is in the current clinical practice. (5) Demographic and injury-related characteristics, which may influence adherence to the BTF guidelines: increase age, elevated blood alcohol level, normal CT scan, and planned neurosurgical intervention or other factors reported in the study will be extracted when a statistically significant relationship between these factors and adherence is demonstrated. (6) Outcomes including mortality or morbidity if they compared between patients treated according to the BTF guidelines and patients treated differently and 95% confidence interval are reported. Discrepancies will be discussed and resolved by consensus or arbitration by other researchers (CO and DZ).\n\nQuality assessment {#Sec12}\n------------------\n\n### Randomized Controlled Trials (RCT) {#Sec13}\n\nThe quality will be assessed using the Cochrane Handbook \"Risk of Bias\" assessment tool \\[[@CR23]\\]. Additionally, we will assess the quality of reporting using a checklist, which will be based on the CONSORT (Consolidated Standards of Reporting Trials).\n\n### Observational studies {#Sec14}\n\nThe quality will be assessed using the Cochrane Risk Of Bias Assessment Tool: for Non- Randomized Studies of Interventions (ACROBAT-NRSI) \\[[@CR24]\\] which evaluates the observational studies based on three domains: (1) Pre-intervention; evaluation of bias due to confounding and bias in selection of participants into the study. (2) At intervention: evaluation of bias in measurement of interventions. (3) Post-intervention: evaluation of biases due to departures from intended interventions bias due to missing data, bias in measurement of outcomes and bias in selection of the reported results. We will assess the quality of reporting of observational studies using a checklist, which will be based on the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement.\n\nTwo researchers (YK and IG) will address quality assessment of the included studies independently. Differences of opinion will be resolved by a discussion with other researchers (CO and DZ).\n\nData synthesis {#Sec15}\n--------------\n\nNarrative synthesis and where appropriate, quantitative meta-analysis will be used. Synthesis will be based on clustering the selected studies based on type of recommendation. Adherence to BTF-based protocol will be extracted as a separate category if the full description of the protocol and protocol adherence rate was reported. Data synthesis will include description of included studies.\n\nThe median adherence and interquartile range for each recommendation will be calculated as well as for overall adherence. Additionally, factors influencing adherence will be examined based on the type of recommendation.\n\nCalculation of pooled estimates of mortality among TBI patients managed based on BTF guidelines and patients managed differently {#Sec16}\n--------------------------------------------------------------------------------------------------------------------------------\n\nIn preliminary search, odds ratio was used as measure of association in several studies, and we will also use the odds ratio as the summary measure of association in our study. If only the relative risk is reported in a selected study, we will transform the relative risk into an odds ratio using the method described by Deeks and Altman \\[[@CR25]\\]. The cohort studies and RCTs will be pooled separately. We will conduct stratified analyses of pooled estimate of mortality by type of recommendation and outcome (for example in-hospital mortality intensive care unit mortality, 30\u00a0days mortality or 6\u00a0months mortality). We will examine heterogeneity separately in the pooled estimates by study design (RCT, observational) using the Cochran *Q* and *I*^2^ statistics \\[[@CR24]\\]. In the presence of heterogeneity, random effects models will be used instead of fixed effects models to account for the expected variability beyond the chance and obtain pooled effect estimates across studies \\[[@CR26]\\]. The pooled estimates obtained from these calculations will then be compared to determine if the results are different between experimental and cohort study designs. If the adjustment for confounding variables varies between studies, analysis will be stratified into two parts, one for studies adjusting for several confounding variables (e.g., age, GCS, injury severity score, pupillary response, and CAT scan head finding) and the other one for studies adjusting for a few confounding variables.\n\nIf an adequate number of studies are chosen for the meta-analysis, we will conduct meta-regression considering the following covariates: year of publication, country of origin, and study period.\n\nPublication bias will also be assessed using funnel plot and the methods described by Begg and Egger \\[[@CR27], [@CR28]\\]. Meta-analysis will be performed using Review Manager software (RevMan5.3.5 Cochrane Collaboration) and regression analysis will be conducted using Stata Statistical Software version 13.1. (StataCorp LP, College Station, TX, USA).\n\nAll data will be extracted by two independent investigators (YK IG). To assess inter-rater reliability, the percent agreement will be calculated on adherence percentage for number of guideline recommendations by third investigator (AS).\n\nDiscussion {#Sec17}\n==========\n\nThis systematic review and meta-analysis will be the first systematic review summarizing relevant literature on guidelines for management of severe traumatic brain injury. In this review, we will demonstrate the current level of professionals' adherence to BTF guidelines in patients with severe traumatic brain injury. In addition, we will describe the factors influencing adherence, which may provide valuable input for development of strategies to successfully increase adherence. Finally, we will describe the effect of these guidelines on patient outcome if data is sufficiently homogenous. Results of this review are expected to be available near the end of 2015.\n\nThe major strength of this systematic review will be the use of several electronic databases and other relevant sources based on established guideline for systematic review. An additional strength of the review will be the use of inter-rater reliability, standard protocol for reporting systematic reviews as well as quality assessment of the included studies. However, there are some limitations in this review. We may not be able to find the non-observational studies due to the nature of the measured effect. Furthermore, this review will be examining adherences with different recommendations at several locations using different clinical determinants. Therefore, a high level of heterogeneity will be expected and may limit our ability to perform meta-analysis.\n\nAdditional file {#Sec18}\n===============\n\nAdditional file 1:**Appendix 1.** Proposed Medline search strategy; **Appendix 2.** Proposed Embase search strategy; **Appendix 3.** Proposed EBM Reviews---Cochrane Database of Systematic Reviews search strategy. (DOCX 24 kb)\n\nBTF\n\n:   Brain Trauma Foundation\n\nTBI\n\n:   traumatic brain injury\n\nCINAHL\n\n:   Cumulative Index to Nursing and Allied Health Literature\n\nCENTRAL\n\n:   Cochrane Central Register of Controlled Trials\n\nRCT\n\n:   Randomized Controlled Trial\n\nAANS\n\n:   American Association of Neurological Surgeons\n\nCNS\n\n:   Congress of Neurological Surgeons\n\nGCS\n\n:   Glasgow Coma Scale\n\nMOOSE\n\n:   Meta-analysis of Observational Studies in Epidemiology\n\nPRISMA\n\n:   Preferred Reporting Items for Systematic Reviews and Meta-Analyses\n\nEBM\n\n:   evidence-based medicine\n\nACP\n\n:   American College of Physicians\n\nNOS\n\n:   Newcastle-Ottawa Scale\n\nSTROBE\n\n:   Strengthening the Reporting of Observational studies in Epidemiology\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n\nYK and DZ formulated the research question. YK, DZ, and CO designed the study. YK developed the preliminary search strategy and drafted the manuscript. SC refined and executed the search strategy and exported the results. YK, IG, and AS designed the statistical analysis plan. YK, DZ, CO, AS, IG, and SC critically reviewed the manuscript for content. All authors have approved the final version of the manuscript.\n\n**Authors' information**\n\nYK is a neurosurgery resident and Clinical Epidemiology Master of Sciene student at the University of Alberta, IG is an assistant professor at the Department of Community and Family Medicine, King Saud University, SC is a health sciences librarian at the University of Alberta. AS is a professor (statistics) at the School of Public Health, University of Alberta. CO is an assistant professor in the Division of Neurosurgery, University of Alberta. DZ is a Professor in the Division of Critical Care Medicine, University of Alberta.\n\nYK is not funded. IG is financially supported by the Vice Deanship of Research Chair, King Saud University, Riyadh, Kingdom of Saudi Arabia. SC is not funded. AS is not funded. CO is not funded. DZ is not funded for this project specifically but is appointed to the Division of Critical Care Medicine, Faculty of Medicine and Dentistry, University of Alberta.\n"}
+{"text": "INTRODUCTION\n============\n\nDiphyllobothriasis is a fishborne tapeworm infection caused by the genus *Diphyllobothrium*. These tapeworms are creamy white in color, and can grow as long as 2-15 m in the human small intestine \\[[@b1-kjp-53-1-109],[@b2-kjp-53-1-109]\\]. Generally, the incubation period of *Diphyllobothrium* plerocercoids until they develop into an adult tapeworm is 2-6 weeks \\[[@b3-kjp-53-1-109]\\]. It can cause gastrointestinal symptoms such as abdominal pain and diarrhea, but typically does not cause severe problems. Humans can be infected with diphyllobothriid species through consumption of raw or poorly cooked fish containing larval plerocercoids.\n\nTo date, at least 13 of about 50 species of genus *Diphyllobothrium* have been reported as human pathogens, and among them, *D. latum* had been reported to be the main pathogen in Korea \\[[@b1-kjp-53-1-109],[@b4-kjp-53-1-109]\\]. In 1919, human *D. latum* infection was first reported in Korea based on recovery of eggs in the stool passages of 2 residents in Jinju by Kojima and Ko \\[[@b5-kjp-53-1-109]\\]. In 1971, *D. latum* infection was first documented on the basis of morphological identification of the expelled proglottids \\[[@b6-kjp-53-1-109]\\]. The first Korean child case of *D. latum* infection was reported in 1980 by Jeong et al. \\[[@b7-kjp-53-1-109]\\]. Since then, 6 pediatric cases of *D. latum* infection have reported on the basis of morphological identification of the expelled proglottids \\[[@b4-kjp-53-1-109]\\]. However, in 2009, by Jeon et al. \\[[@b8-kjp-53-1-109]\\], all the 62 cases previously diagnosed as *D. latum* were verified to be actually *D. nihonkaiense* infection based on DNA analysis \\[[@b8-kjp-53-1-109]\\].\n\n*D. latum* and *D. nihonkaiense* are morphologically indistinguishable. Therefore, alternative genetic sequencing has been used as the most discriminative method for identifying each species. Recently, a cheap and rapid molecular test using a multiplex PCR with the *cox1* gene was developed for verifying the most common diphyllobothriid species infecting humans \\[[@b9-kjp-53-1-109]\\]. Although the treatment regimen for these 2 species is identical, the correct identification is necessary to acquire the epidemiologic information including intermediate hosts of this tapeworm so as to prevent further infections. We encountered a family (4 people) who experienced abdominal discomfort and diarrhea and discharged tapeoworm strobilae after eating raw fish in Korea. We could collect some segments discharged from a 7-year-old boy and genetically analyzed by *cox1* gene sequencing to determine the species.\n\nCASE RECORD\n===========\n\nA 7-year-old boy visited our outpatient clinic with a strobila of a tapeworm, about 1.2 m long. The patient was a resident of an urban area (Goyang, Korea) and was a first-grade student at an elementary school. His height was 123 cm (50-75th percentile) and weight was 24 kg (50-75th percentile). He had no previous medical history of note. He and his family members (4 persons), including his parents, had been on a trip to a coastal area (Mukho Harbor, Donghae, Korea) 16 days before they visited our hospital. There, they ate slices of 3 unknown kinds of raw fish and it was the first time when the child ate raw fish. Two weeks after the trip, 4 people who had been on the trip, including the boy, complained of lower abdominal pain and watery diarrhea. At day 16 after eating the fish, his mother found a noodle-like worm in his stool, and tried to pull out the tapeworm, but it snapped. His mother and the patient visited our outpatient clinic with some segment of the tapeworm. He had no other symptoms abide by the gastrointestinal problems. His bowel sounds were hyperactive, but other physical examinations were unremarkable, with no signs of anemia or neurological manifestations. We found no abnormal laboratory test results, including vitamin B~12~ level which was 1,033 *p*g/ml (normal range: 200-950 *p*g/ml). In stool examination, no parasite eggs or larvae were found. We treated the patient with a single oral dose of praziquantel. On the next day, he expelled the rest of the tapeworm, which was about 30 cm long and all of his symptoms disappeared. The other 3 members, i.e., his parents and a relative, also expelled strobilae in their stools (we could not collect the tapeworm specimens). They also took praziquantel and recovered from their illnesses.\n\nThe tapeworm specimen was fixed in 10% formalin, and sent to the Department of Parasitology and Tropical Medicine, Seoul National University College of Medicine, Seoul, Korea. Formalin was substituted with 1\u00d7 PBS to remove the formalin for 24 hr. Before the molecular identification, the sample was identified as *Diphyllobothrium* sp., according to the microscopic examinations (\u00d7400 magnification) of eggs extracted from the uteri. Genomic DNA was then extracted from a single proglottid following a spin-column protocol using DNeasy Blood & Tissue Kit (\u01eaiagen, Hilden, Germany). The partial *cox1* gene was amplified by PCR using the forward primer (D1/nco1f1 5'-TAG CTG CTG CTA TAC AAT GTT GTT ATT-3') and the reverse primer (D1/nco1r1 5'-ACG ACG TGG TAA ACG GCA CAC ACC AAA-3'). PCR was carried out with 1 cycle of 94\u02daC for 5 min, 35 cycles of 94\u02daC for 1 min, 46\u02daC for 1 min, and 72\u02daC for 1 min, plus a final cycle at 72\u02daC for 5 min. DNA sequencing and construction of a neighbor-joining tree using our samples were performed according to the protocol described by Jeon et al. \\[[@b8-kjp-53-1-109]\\]. The reference data we used for a phylogenetic study were *D. nihonkaiense* (GenBank no. AB684622, EF420138, AB268585), *D. latum* (AB504899, DQ985706), and *S. erinaceieuropaei* (AB369250). The *cox1* sequences (688 bp) of the tapeworm showed 99.9% (687/688 bp) similarity with the reference sequence of *D. nihonkaiense* (EF420138), whereas the similarity with *D. latum* (DQ985706) was 93.2% (641/688 bp). Thus, the pathogen was identified as *D. nihonkaiense* ([Fig. 1](#f1-kjp-53-1-109){ref-type=\"fig\"}).\n\nDISCUSSION\n==========\n\nIn the Far East Asian countries, the main pathogen in almost all human diphyllobothriasis had been considered to be *D. latum*. However, in 1986, Yamane et al. \\[[@b10-kjp-53-1-109]\\] in Japan characterized *D. nihonkaiense* for the first time, describing morphological differences in eggs, adult worms, and plerocercoids of so-called *D. latum* between Japan and Finland \\[[@b10-kjp-53-1-109]\\]. *D. nihonkaiense* in Korea was first mentioned in 1990 on the basis of the morphological findings by Rim et al. \\[[@b11-kjp-53-1-109]\\].\n\nIn 2009, genetic analysis of *Diphyllobothrium* tapeworms collected from patients with diphyllobothriasis that had been reported as *D. latum* infection between 1982 and 2007 was conducted in Korea, and all of the 62 specimens analyzed were confirmed to be *D. nihonkaiense* \\[[@b8-kjp-53-1-109]\\]. Thereafter, additional 10 cases of diphyllobothriasis in Korea were confirmed to be *D. nihonkaiense* by *cox1* gene sequencing \\[[@b12-kjp-53-1-109]-[@b15-kjp-53-1-109]\\]. The current case (boy) is the 73rd case of *D. nihonkaiense* infection in Korea confirmed by genetic analysis and the 8th pediatric case of *D. nihonkaiense* infection in Korea. The current case is valuable in that there are rare pediatric cases of *D.nihonkaiense* infection reported in Korea, and *D. nihonkaiense* infection as a family unit rarely occurs. This is the first report on *D. nihonkainens*e infection in a family unit.\n\nThe intermediate hosts of *D. nihonkaiense* are now known to be some kinds of Pacific salmon, which include *Oncorhynchus keta*, *O. masou*, *O. gorbuscha*, and *O.nerka* in Japan \\[[@b3-kjp-53-1-109]\\]. These kinds of salmon are also considered to be the main intermediate hosts of *D. nihonkaiense* in Korea because these fish migrate from Okhotsk, Bering, and the Pacific Ocean and return back to the East Sea, which Korea and Japan are sharing \\[[@b8-kjp-53-1-109]\\]. In Japan, a study investigating plerocercoids in wild Pacific salmon using PCR-based DNA sequence analysis targeting *cox1* and mitochondrial NADH dehydrogenase subunit 3 *(nad3)* genes was conducted; all plerocercoids obtained in salmon were identified as *D. nihonkaiense* \\[[@b16-kjp-53-1-109]\\]. However, there has been no reported trial analyzing genes of diphyllobothriid plerocercoids collected from intermediate host fish in Korea. Thus, there is a need for gene analysis of plerocercoids acquired from the fish that were ingested by patients to determine the source of infection in Korea.\n\nThe incidence of infection with *D. nihonkaiense* is increasing not only in the Far East Asian countries like Korea and Japan, but also in countries far from the East Asia, such as North America, New Zealand, and European countries due to ingestion of imported Pacific salmon \\[[@b17-kjp-53-1-109]-[@b20-kjp-53-1-109]\\]. For example, there were 9 cases of molecularly identified human diphyllobothriasis reported in Europe from 2005-2011. Among these, 6 cases were confirmed as *D. nihonkaiense* and most become infected with the pathogen by eating imported Pacific salmon \\[[@b17-kjp-53-1-109],[@b20-kjp-53-1-109]-[@b23-kjp-53-1-109]\\].\n\nInfection by diphyllobothriid species can be prevented by freezing and storing fish at -20\u02daC for 7 days or -35\u02daC for 15 hr before consumption \\[[@b24-kjp-53-1-109]\\]. Alternatively, larval plerocercoids can be destroyed by cooking the fish at 54-56\u02daC for 5 min \\[[@b3-kjp-53-1-109]\\]. However, people generally do not know about the risk of diphyllobothriasis when they eat raw or undercooked fish. Moreover, people tend to prefer never-frozen raw fish, like sushi or sashimi, to that prepared from frozen fish. Recently, consumption of raw fish or incompletely cooked fish, such as smoked salmon, has become popular globally. As a result, the chances of eating raw or undercooked fish have increased. This increased raw or insufficiently cooked fish consumption has resulted in increased risks for fishborne parasite infections such as diphyllobothriasis.\n\nBecause the treatment regimen of diphyllobothriasis is praziquantel, diagnosis through *cox1* gene sequencing can be considered too academic. Even though, accurate diagnosis for these species is necessary, regarding the epidemiology of these parasites and arrangement of protective methods. It led to a perception that dominant species is not *D. latum* but *D. nihonkaiense* in Korea. It would provide information whether the cause of diphyllobothriasis is imported cases or domestic species during its outbreak and lead to build protective methods for public health promotion.\n\nWe have no conflict of interest related to this work.\n\n![Neighbor-joining tree of human *Diphyllobothrium* tapeworms based on nucleotide sequences of the *cox1* gene. Numbers above the branches detect the bootstrap values (1,000 replicates). The scale bar represents the estimated number of nucleotide substitutions per nucleotide site. The phylogenetic tree reveals that the *cox1* gene from our case was closer to *D. nihonkaiense* than *D. latum*.](kjp-53-1-109f1){#f1-kjp-53-1-109}\n"}
+{"text": "1.. Introduction {#s1}\n================\n\nMultiple myeloma (MM) is a neoplastic plasma cell disease with high morbidity and mortality rates. MM reportedly accounts for approximately 10--15% of haematologic malignancies in the world \\[[@RSOS192152C1],[@RSOS192152C2]\\]. Almost 6--7 individuals out of every 100 000 people in the world are diagnosed with MM each year \\[[@RSOS192152C2]\\]. In 2015, 28 850 patients were diagnosed with MM, and the number of deaths due to this illness was more than 11 000 in the United States \\[[@RSOS192152C3]\\]. However, MM is still incurable and there is a great need to develop more effective therapeutic strategies.\n\nNuclear factor-kappa B (NF-\u03baB) was reported to be abnormally activated in many types of cancer cells, including MM in the 1990s \\[[@RSOS192152C4]\\], and it inhibits the apoptosis of cancer cells by upregulating anti-apoptotic genes \\[[@RSOS192152C4],[@RSOS192152C5]\\]. A series of studies have been performed to investigate the underlying mechanism of NF-\u03baB in the pathogenesis of MM, with the aim of exploring NF-\u03baB-targeted therapies \\[[@RSOS192152C6]--[@RSOS192152C11]\\]. However, to date, no appropriate NF-\u03baB-targeted treatments have been established due to their serious side effects \\[[@RSOS192152C4]\\]. The inhibitors of NF-\u03baB similarly affect the NF-\u03baB pathway in MM and normal cells leading to serious damage. This reflects the important roles of NF-\u03baB in cell survival, inflammatory and immune responses and other issues \\[[@RSOS192152C12]\\].\n\nGrowth arrest and DNA damage-inducible beta (GADD45\u03b2) forms part of the downstream module of the NF-\u03baB pathway and is essential to the survival of MM cells \\[[@RSOS192152C4],[@RSOS192152C13]\\]. GADD45\u03b2 binds to and inhibits MAP kinase kinase 7 (MKK7) to reduce the phosphorylation of c-Jun N-terminal kinase (JNK) pathway which usually acts to ensure the normal apoptosis of cell \\[[@RSOS192152C14]\\]. Therefore, the overexpression of GADD45\u03b2 due to an abnormally activated NF-\u03baB pathway in MM cells can inhibit the apoptosis of MM cells by suppressing the JNK pathway \\[[@RSOS192152C4],[@RSOS192152C14],[@RSOS192152C15]\\]. Hence, GADD45\u03b2\\|MKK7-targeted therapeutic strategies were suggested as a potential way to kill MM cells effectively, and crucially, these may not have side effects on normal cells \\[[@RSOS192152C4],[@RSOS192152C16]\\].\n\nA series of mathematical models have been constructed to study MM based on experimental findings \\[[@RSOS192152C17]--[@RSOS192152C23]\\]. These models have demonstrated great potential to improve our understanding of the complicated pathogenesis of the disease. However, the underlying mechanism through which GADD45\u03b2 promotes the survival of MM cells as downstream modules of the NF-\u03baB pathway was not included. Therefore, a mathematical model of MM is built in this paper by considering the NF-\u03baB and JNK pathways and their interaction through GADD45\u03b2 and MKK7. The model cannot only mimic the variations in bone cells, the bone volume and MM cells with time, but it can also examine how the NF-\u03baB pathway acts with the JNK pathway to promote the development of MM cells.\n\n2.. Model development {#s2}\n=====================\n\n2.1.. Basic structure of the model {#s2a}\n----------------------------------\n\n[Figure\u00a01](#RSOS192152F1){ref-type=\"fig\"} describes the schematic diagram of the constructed mathematical model in the paper. It demonstrates the mechanisms of the NF-\u03baB and JNK pathways and shows how the interaction between these two pathways promotes the survival of MM cells. It should be noted that the coupling between MM cells and bone cells (osteoclastic and osteoblastic lineages) is not included in [figure\u00a01](#RSOS192152F1){ref-type=\"fig\"}, but it is included in the electronic supplementary material, Figure A1 in Appendix A for convenience, since these cells have been previously studied (e.g. \\[[@RSOS192152C23]\\]). [Figure\u00a01](#RSOS192152F1){ref-type=\"fig\"} is made up of two parts: part A and part B, which involve the NF-\u03baB and JNK pathways, respectively. Figure 1.The schematic description of the mechanisms of the NF-\u03baB and JNK pathways, and their interaction.\n\nPart A shows how the NF-\u03baB pathway is activated and how it then upregulates anti-apoptotic genes in MM cells. In the unstimulated state, NF-\u03baB heterodimers are always kept in the I\u03baB\u03b1\\|NF-\u03baB complex in the cytoplasm. Upon the stimulation of TNF, IKK is transformed from its neutral state (denoted as IKKn) into its active state (denoted as IKKa). IKKa can phosphorylate and ubiquitinate I\u03baB\u03b1 and then release free NF-\u03baB heterodimers from the I\u03baB\u03b1\\|NF-\u03baB complex. The free NF-\u03baB heterodimers enter the nucleus and regulate the transcription of I\u03baB\u03b1, A20 and GADD45\u03b2 \\[[@RSOS192152C24]\\]. In addition to I\u03baB\u03b1, A20 also serves as an inhibitor of the NF-\u03baB pathway, promoting the transformation of IKKa into its unactivated IKKi form \\[[@RSOS192152C25]\\]. As a downstream factor of NF-\u03baB signalling, the newly produced GADD45\u03b2 is expressed at a high level in the MM cells and interacts with MKK7 in the JNK pathway, promoting the survival of MM cells \\[[@RSOS192152C4],[@RSOS192152C26],[@RSOS192152C27]\\] (detailed information is included in part B). It should be noted that I\u03baB\u03b2, I\u03baB\u03b3 and I\u03baB\u03b5 can also bind to NF-\u03baB heterodimers and inhibit their nuclear localization signals \\[[@RSOS192152C28]\\]. Only I\u03baB\u03b1 is considered in our model, as NF-\u03baB heterodimers are primarily sequestered by I\u03baB\u03b1.\n\nPart B describes how the overexpression of GADD45\u03b2 acts with MKK7 and dysregulates the JNK pathway to facilitate the survival of MM cells. There are three cascades, JNK (a member of the mitogen-activated protein kinase (MAPK) family), MKK7 (a member of the MAPK kinase family) and MAPK kinase kinase (MAPKKK) in the JNK pathway \\[[@RSOS192152C8],[@RSOS192152C29],[@RSOS192152C30]\\]. The activated kinase at the upper level phosphorylates the kinase at the next level down the cascade. TNF stimulation promotes the phosphorylation of MAPKKK mediated by reactive oxygen species (ROS) \\[[@RSOS192152C26],[@RSOS192152C31],[@RSOS192152C32]\\]. MAPKKK_P denotes phosphorylated MAPKKK. The full activation of both MKK7 and JNK requires the phosphorylation of two conserved sites \\[[@RSOS192152C33]\\]. MAPKKK_P promotes the single phosphorylation of MKK7 (denoted as MKK7_P) at first and then doubly phosphorylated (denoted as MKK7_PP). Similarly, MKK7_PP then facilitates the single and double phosphorylation of JNK (denoted as JNK_P and JNK_PP, respectively). JNK_PP is able to promote the apoptosis of MM cells by up-regulating pro-apoptotic genes \\[[@RSOS192152C15]\\]. The GADD45\u03b2 produced during NF-\u03baB signalling binds to MKK7 and inhibits its enzymatic activity, which limits the phosphorylation of JNK promoted by MKK7_PP. The decrease in JNK_PP inhibits the expression of pro-apoptotic genes and MM cell death.\n\n2.2.. Model equations {#s2b}\n---------------------\n\nTo investigate the role of GADD45\u03b2 in MM pathogenesis, the mathematical model proposed in \\[[@RSOS192152C23]\\] is extended by including the pathways of NF-\u03baB and JNK as well as their crosstalk. The proposed model consists of 29 ordinary differential equations (ODEs) with five state variables, namely osteoblast precursors (OBp), active osteoblasts and osteoclasts (OBa and OCa), MM and the bone volume (BV). Equations (2.1)--(2.5) represent temporal variations in the concentrations of OBp, OBa, OCa, MM and the BV, respectively. Equations (B1)--(B11) and (C1)--(C16) in the electronic supplementary material, describe variations in biochemical factors regarding the JNK and NF-\u03baB pathways with time, respectively.$$\\frac{\\text{d}}{\\text{d}t}OB_{p} = D_{OB_{u}} \\cdot \\pi_{act,OB_{u}}^{TGF\\beta} \\cdot OB_{u} - D_{OB_{p}} \\cdot \\pi_{rep,OB_{p}}^{TGF\\beta} \\cdot \\pi_{rep,OB_{p}}^{VCAM1} \\cdot OB_{p},$$$$\\frac{\\text{d}}{\\text{d}t}OB_{a} = D_{OB_{p}} \\cdot \\pi_{rep,OB_{p}}^{TGF\\beta} \\cdot \\pi_{rep,OB_{p}}^{VCAM1} \\cdot OB_{p} - \\, A_{OB_{a}} \\cdot \\pi_{act,OB_{a}}^{VCAM1} \\cdot OB_{a},$$$$\\frac{\\text{d}}{\\text{d}t}OC_{a} = D_{OC_{p}} \\cdot \\pi_{act,OC_{\\, p\\,}}^{RANKL} \\cdot OC_{p} - \\,\\pi_{act,OC_{a\\,}}^{TGF\\beta} \\cdot A_{OC_{a}} \\cdot OC_{a},$$$$\\begin{matrix}\n{\\frac{\\text{d}}{\\text{d}t}MM = \\,} & {D_{MM} \\cdot \\pi_{act,\\, MM}^{IL6} \\cdot \\pi_{act,\\, MM}^{VCAM1} \\cdot MM \\cdot \\left( {1 - \\frac{MM}{MM_{\\max}}} \\right)} \\\\\n & {\\quad - A_{MM} \\cdot \\pi_{rep,\\, MM}^{SLRPS} \\cdot \\pi_{act,\\, MM}^{JNK\\_ PP} \\cdot MM} \\\\\n\\end{matrix}$$$$\\text{and}\\qquad{\\frac{\\text{d}}{\\text{d}t}BV = - K_{res} \\cdot OC_{a} + \\, K_{form} \\cdot OB_{a}.}$$\n\n'Hill functions' are used to represent the cellular interaction via single ligand to receptor binding denoted by \u03c0 functions \\[[@RSOS192152C34]\\], with equations (2.6) and (2.7) describing the stimulating and inhibiting functions of ligand-receptor binding. Here. *L* represents the concentration of the ligand, *\u03b2* represents maximal expression level of the promoter, *n* is the coefficient which regulates the steepness of the function *\u03c0* and *k*~1~ and *k*~2~ represent the dissociation constant, respectively. Both *\u03b2* and *n* are assumed to equal 1 in the model following the work of Pivonka *et al*. \\[[@RSOS192152C34]\\].$$f(x) = \\beta\\pi_{act} = \\frac{\\beta{(L)}^{n}}{k_{1} + {(L)}^{n}}$$and$$f(x) = \\beta\\pi_{rep} = {\\frac{\\beta}{1 + {({L/k_{2}})}^{n}}.}$$\n\n*dOB~p~*/d*t*, *dOB~a~*/d*t*, *dOC~a~*/d*t*, (d/d*t*)*MM* and (d/d*t*)*BV* denote the variations of *OB~p~*, *OB~a~*, *OC~a~*, *MM* and *BV,* respectively. For example, *dOB~p~*/d*t* is the variation of *OB~p~* with time. $D_{OB_{u}}$ and $D_{OB_{p}}$ represent the differentiation rates of uncommitted osteoblast progenitors and osteoblast precursors. *OB~u~* and *OB~p~* are concentrations of uncommitted osteoblast progenitors and osteoblast precursors. $\\pi_{act,OB_{u}}^{TGF\\beta}$ represents the stimulation of uncommitted osteoblastic progenitors into osteoblastic precursors. $\\pi_{rep,OB_{p}}^{TGF\\beta}$ represents the inhibition of the differentiation of osteoblastic precursors into active osteoblasts. $\\pi_{rep,OB_{p}}^{VCAM1}$ represents BMSC-MM cell adhesion that blocks the differentiation of mature osteoblasts from their progenitors. Definitions of other Hill functions and variables are not repeated here but they are included in electronic supplementary material, Appendix A.\n\nThe newly added Hill function $\\pi_{act,{\\,\\,\\,}MM_{\\,}}^{JNK\\_ PP}$ in equation (2.4) represents the promotion of MM cell apoptosis by JNK_PP. The definition of $\\pi_{act,{\\,\\,\\,}MM_{\\,}}^{JNK\\_ PP}$ is as follows:$$\\pi_{act,{\\,\\,\\,}MM_{\\,}}^{JNK\\_ PP}\\, = {\\,\\,}{\\frac{JNK\\_ PP}{\\, JNK\\_ PP + \\, K_{JNK\\_ PP}{\\,\\,}},}$$where $K_{JNK\\_ PP}$ represents the activation coefficient related to JNK_PP promoting MM cell apoptosis, and its value is included in [table\u00a01](#RSOS192152TB1){ref-type=\"table\"}. JNK_PP represents the concentration of doubly phosphorylated JNK. The calculation of JNK_PP requires the mathematical modelling of the NF-\u03baB and JNK pathways together with their interaction through GADD45\u03b2. In accordance with the earlier work of \\[[@RSOS192152C35]\\] and \\[[@RSOS192152C25]\\], equations (B1)-- (B11) and (C1)--(C16) in the electronic supplementary material, were constructed to simulate the JNK and NF-\u03baB pathways. The distinct feature of these equations is that the interaction of the two pathways, which was ignored before, was included, with the addition of $E_{{MKK}7}^{{GADD}45\\beta}$ in electronic supplementary material, equations (B10) and (C16). $E_{{MKK}7}^{{GADD}45\\beta}$ represents the binding of GADD45\u03b2 to MKK7, which decreases the amount of MKK7 available for phosphorylation into MKK7_P. This action further leads to the inhibition of the JNK pathway and promotes the survival of MM cells as discussed before. The definition of $E_{{MKK}7}^{{GADD}45\\beta}$ is as follows:$$E_{{MKK}7}^{{GADD}45\\beta} = k_{GM} \\cdot GADD45\\beta \\cdot {{MKK}7} - d_{GM} \\cdot ({{{GADD}45\\beta}|{{MKK}7}})\\,,$$where the first term represents the formation of the GADD45\u03b2\\|MKK7 complex, and the second term describes its dissociation. The definitions of the variables in the JNK and NF-\u03baB pathways are included in electronic supplementary material, Appendices B and C. In summary, \u03c0 function $\\pi_{act,{\\,\\,\\,}MM_{\\,}}^{JNK\\_ PP}$ and 24 ODEs are new additions to our model, allowing us to investigate the essential role of the interactions between NF-\u03baB and JNK signalling in the development of MM. Table\u00a01.Descriptions and values of parameters used in the model.parameterdescriptionvalue\u00a0$K_{JNK\\_ PP}$activation coefficient related to effect of JNK_PP on MM3.2124 pM (fitted)\u00a0*k~G~*GADD45\u03b2 translation rate0.5 s^\u22121^ (estimated)\u00a0*d~G~*GADD45\u03b2 protein degradation rate0.0003 s^\u22121^ (estimated)\u00a0*k~GM~*GADD45\u03b2-MKK7 association rate5.3781 \u00d7 10^\u22125^ \u03bcM s^\u22121^ (fitted)\u00a0*d~GM~*GADD45\u03b2-MKK7 dissociation rate2.1046 \u00d7 10^\u22125^ \u03bcM s^\u22121^ (fitted)\u00a0*k~MAPKKK~*synthesis rate of MAPKKK0.5 nM s^\u22121^ (estimated)\u00a0*k~MKK~*~7~synthesis rate of MKK71.5 nM s^\u22121^ (estimated)\u00a0*k~JNK~*synthesis rate of JNK1.5 nM s^\u22121^ (estimated)\u00a0*d~MAPKKK~*degradation rate of MAPKKK2 \u00d7 10^\u22125^ nM s^\u22121^ (estimated)\u00a0*d~MKK~*~7~degradation rate of MKK72 \u00d7 10^\u22125^ nM s^\u22121^ (estimated)\u00a0*d~JNK~*Degradation rate of JNK2 \u00d7 10^\u22125^ nM s^\u22121^ (estimated)\u00a0*K~D~*~,*VCAM*1,*MM*,*act*~half-maximal concentration of VCAM-1 on promoting the MM cells production0.0022 pM (fitted)\u00a0*K~D~*~,*SLRPs*,*MM*,*rep*~half-maximal concentration of SLRPs on repressing the MM cells production2.3549 \u00d7 10^9^ pM (fitted)\u00a0*K~D~*~,*IL*6,*MM*,*act*~half-maximal concentration of IL-6 on promoting the MM cells production5.8217 \u00d7 10^\u22126^ pM (fitted)\u00a0*MM~max~*maximum possible MM concentration2.0836 pM (fitted)\n\n3.. Results {#s3}\n===========\n\nThe definitions and values of the model parameters are listed in [table\u00a01](#RSOS192152TB1){ref-type=\"table\"}. Several parameter values were reported in previous studies, while the remaining unknown parameters (i.e. those parameters where experimental data are unavailable or those that have no direct biological meaning) were fitted via the genetic algorithm (GA) in this paper. The initial values of the model variables are described in [table\u00a02](#RSOS192152TB2){ref-type=\"table\"}. The ode45 solver in the Matlab software package (R2014b, Mathworks, Natick, USA) is used to solve the model equations. Curves in figures\u00a0[2](#RSOS192152F2){ref-type=\"fig\"}--[10](#RSOS192152F10){ref-type=\"fig\"} represent numerical solutions of model equations. The Matlab code is included in the electronic supplementary material. The calculation of unknown model parameters based on GA are described as follows:$$F(X) = \\sum\\limits_{i = 1:3}abs({M{(X)}_{i}\\, - \\, P_{i}})$$and$$X = \\lbrack{K_{JNK_{PP}},\\,\\ldots,\\, MM_{\\max}}\\rbrack,$$where $\\text{X} = \\lbrack{K_{JNK\\_ PP},\\, k_{GM},\\, d_{GM},\\, K_{D,VCAM1,MM,act},\\, K_{D,SLRPs,MM,rep},\\, K_{D,IL6,MM,act},\\, MM_{max}}\\rbrack$ is a row vector consisting of the seven unknown model parameters and represents one point in the parameter space. *M*(*X*)*~i~* and *P~i~*(*i* = 1, 2, 3) represent model outputs corresponding to each point in the parameter space and the preferred model outputs, respectively. In this work, model outputs refer to the concentrations of *OB~p~*. Different combinations of parameter values have different outputs. GA is an effective way to search for parameter values in parameter space to make model outputs approach to preferred outputs. Figure 2.Model simulations of the normalized variation in the concentrations of OBp, OBa, OCa and MM cells with respect to their respective initial values (MM cells are injected at 10th hour and removed at 150th hour). Figure 3.Model simulation of the variation in the normalized ratio of OBa to OCa with respect to the initial ratio (MM cells are injected at 10th hour and removed at 150th hour). Figure 4.Model simulation of the variation in the normalized bone volume with respect to its initial value (MM cells are injected at 10th hour and removed at 150th hour). Figure 5.Model simulation of the variation in the GADD45\u03b2 concentration after the activation of the NF-\u03baB pathway at 10th hour. Figure 6.Model simulations of the variation in biochemical factors involved in JNK signalling after the pathway is activated at 10th hour. Figure 7.Model simulations of the variation in MM cells due to the inhibition of GADD45\u03b2 and MKK7 binding by 10% and 40% at 150th hour. Figure 8.Model simulations of the variation in JNK_PP due to the inhibition of GADD45\u03b2 and MKK7 binding by 10% and 40% at 150th hour. Figure 9.Model simulations of the variation in MM cells due to the inhibition of IKK activation by 10% and 40% at 150th hour. Figure 10.Model simulation of the variation in NF-\u03baB due to the inhibition of binding of GADD45\u03b2 and MKK7, and the IKK activation. (*a*) The binding of GADD45\u03b2 and MKK7 inhibited by 40% at 150th hour. (*b*) The IKK activation inhibited by 40% at 150th hour. Table\u00a02.The initial values of cell concentrations used in the model.variablesvaluesunitOBu3.27 \u00d7 10^\u22126^ \\[[@RSOS192152C36],[@RSOS192152C37]\\]pMOBp7.63 \u00d7 10^\u22124^ \\[[@RSOS192152C38]\\]pMOBa6.31 \u00d7 10^\u22124^ \\[[@RSOS192152C39],[@RSOS192152C40]\\]pMOCp1.28 \u00d7 10^\u22123^ \\[[@RSOS192152C41]\\]pMOCa1.05 \u00d7 10^\u22124^ \\[[@RSOS192152C39],[@RSOS192152C40]\\]pMMM3.26 \u00d7 10^\u22121^ \\[[@RSOS192152C42],[@RSOS192152C43]\\]pM\n\nFigures\u00a0[2](#RSOS192152F2){ref-type=\"fig\"}--[4](#RSOS192152F4){ref-type=\"fig\"} describe the variations in the concentrations of bone cells, MM cells, the OBa : OCa ratio, and the bone volume after the invasion and removal of the MM cells. As demonstrated in [figure\u00a02](#RSOS192152F2){ref-type=\"fig\"}, the concentrations of OBp, OBa and OCa remain in a steady state under normal conditions. This steady state is disturbed due to the invasion of the MM cells. OBp, OBa and OCa all increase to different degrees following the increase of MM cells. These simulation results are consistent with experimental observations \\[[@RSOS192152C44]--[@RSOS192152C47]\\]. OCa grows in a larger degree than that of OBa, which leads the OBa : OCa ratio to drop after the invasion of the MM cells, as demonstrated in [figure\u00a03](#RSOS192152F3){ref-type=\"fig\"}. The OBa : OCa ratio is a key factor in the variation of the bone volume, since OBa and OCa are for forming and resorbing bone, respectively, during the bone remodelling process. The decreased OBa : OCa ratio causes the decline in the bone volume, as shown in [figure\u00a04](#RSOS192152F4){ref-type=\"fig\"}, which is consistent with the observed loss of the bone volume in MM patients \\[[@RSOS192152C44]\\]. These simulation results match the work of \\[[@RSOS192152C23]\\]. Figures\u00a0[5](#RSOS192152F5){ref-type=\"fig\"} and [6](#RSOS192152F6){ref-type=\"fig\"} reveal the change in the GADD45\u03b2 concentration after the activation of NF-\u03baB signalling and its influence on JNK signalling. [Figure\u00a07](#RSOS192152F7){ref-type=\"fig\"} reveals the temporal variation in MM cells due to the inhibition of GADD45\u03b2 and MKK7 binding by 10% and 40%. Figures\u00a0[7](#RSOS192152F7){ref-type=\"fig\"} and [8](#RSOS192152F8){ref-type=\"fig\"} suggest how GADD45\u03b2-targeted therapy inhibits the growth of MM cells and JNK_PP. Thus, the inhibition of GADD45\u03b2 binding to MKK7 leads to an obvious drop of MM cells, which then leads to a new steady state at a lower level. The larger degree of inhibition results in a larger drop in MM cells. These simulation results can be confirmed by the experimental data \\[[@RSOS192152C4]\\]. [Figure\u00a09](#RSOS192152F9){ref-type=\"fig\"} shows the variation in MM cells after IKK activation is inhibited by 10% and 40%. As shown in [figure\u00a09](#RSOS192152F9){ref-type=\"fig\"}, IKK inhibition leads to a decrease in MM cells which is accentuated with a larger degree of IKK is inhibition. [Figure\u00a010](#RSOS192152F10){ref-type=\"fig\"} indicates how the inhibition of GADD45\u03b2 and IKK influence NF-\u03baB.\n\n4.. Discussion {#s4}\n==============\n\nThe proposed model in this paper aims to investigate how the interaction between NF-\u03baB and JNK signalling via the GADD45\u03b2\\|MKK7 complex promotes the development of MM. The model can not only simulate the temporal variation of bone cells, the bone volume and biochemical factors involved in the NF-\u03baB and JNK pathways, but it can also mimic the underlying mechanism in which the coupling between the two pathways inhibits apoptosis in MM cells. Additionally, the efficacies of GADD45\u03b2- and NF-\u03baB-targeted therapies were compared based on the model simulation. This work only demonstrates the temporal variation in the GADD45\u03b2 concentration after the activation of the NF-\u03baB pathway. The variations in other biochemical factors related to NF-\u03baB signalling were not repeated here as they have been analysed in \\[[@RSOS192152C25]\\] but are included in electronic supplementary material, Appendix C.\n\nAs mentioned above, JNK_PP plays an essential role in the apoptosis of MM cells, and JNK_PP is able to inhibit the growth of MM cells. According to [figure\u00a08](#RSOS192152F8){ref-type=\"fig\"}, the decrease in MM cells can be explained by the rising amount of JNK_PP after the inhibition of GADD45\u03b2 binding to MKK7. NF-\u03baB-targeted therapies are most often performed by inhibiting IKK, which is essential for NF-\u03baB activation \\[[@RSOS192152C8]\\]. Based on model simulations ([figure\u00a010](#RSOS192152F10){ref-type=\"fig\"}), the GADD45\u03b2-targeted therapy is suggested to be more sensitive than the NF-\u03baB-targeted therapy as it can produce a greater suppression of MM cells than IKK inhibition when each are suppressed to the same degree. In addition to the higher sensitivity, GADD45\u03b2-targeted therapy has a high MM cell specificity. This approach can not only kill MM cells effectively, but it also has no effect on NF-\u03baB levels. Therefore, it can be predicted that the side effects seen with NF-\u03baB-targeted treatment would be avoided. However, these *in silico* hypotheses generated from the mathematical model described in this paper will ultimately have to be validated by further experimental data to confirm the clinical potential of GADD45\u03b2-targeting therapies. It should also be noted that the proposed simplified model does not take into account the interactions with several other proteins and other cell structures, which represents a limitation of this study. Specifically, although the differentiation of progenitors into active osteoclasts and osteoblasts contains several intermediate stages, the model only considered four osteoblastic and three osteoclastic lineages, and contained three state variables: osteoblast precursors, active osteoblasts and active osteoclasts. Additionally, I\u03baBs, including I\u03baB\u03b1, I\u03baB\u03b2, I\u03baB\u03b3 and I\u03baB\u03b5, can also bind to NF-\u03baB and form a complex but we only considered I\u03baB\u03b1.\n\n5.. Conclusion {#s5}\n==============\n\nThis paper describes a mathematical model of MM that was used to investigate how NF-\u03baB signalling acts to include the interplay between NF-\u03baB and JNK signalling that was not included in the previous models. The model not only reconstructs how the invasion of MM cells disturbs the steady state of the bone microenvironment and triggers the variation in bone cells, but it also mimics the changes in biochemical factors involved in the NF-\u03baB and JNK pathways. In addition, the model also investigates the efficacies of GADD45\u03b2- and NF-\u03baB-targeted treatments, suggesting that GADD45\u03b2-targeted therapy is more effective but has no apparent side effects. The simulation results match the experimental observations. This model helps to illuminate the essential function of the crosstalk between the NF-\u03baB and JNK pathways during MM development. It is anticipated that this model could be employed as a useful tool to initially investigate and even explore potential therapies involving the NF-\u03baB and JNK pathways. More work is required to improve the built model to take into account many others parameters such as the interactions with other protein partners, with membranes and other cell structures.\n\nSupplementary Material\n======================\n\n###### APPENDIX.pdf\n\n###### Reviewer comments\n\nSupplementary Material\n======================\n\n###### tumor.m is used to simulate what happens after the introduction and removal of MM cells\n\nSupplementary Material\n======================\n\n###### tumor_GADD45B.m is used to reproduce what happens after inhibiting the formation rate of GADD45B\\|MKK7 complex.\n\nSupplementary Material\n======================\n\n###### tumor_IKK.m is used to reproduce what happens after inhibiting the activation rate of IKK.\n\nSupplementary Material\n======================\n\n###### Equation_Normal.m represents the normal condition without the introduction of MM cells.; Equation_Normal.m, Equation_Tumor1.m, Equation_Tumor2.m, Equation_Tumor3.m and parameters.m are subfunctions.\n\nSupplementary Material\n======================\n\n###### Equation_Tumor1.m represents the condition with the introduction of MM cells.\n\nSupplementary Material\n======================\n\n###### Equation_Tumor2.m represents the condition in which the formation rate of GADD45B\\|MKK7 complex is inhibited after the introduction of MM cells.\n\nSupplementary Material\n======================\n\n###### Equation_Tumor3.m represents the condition in which the activation rate of IKK is inhibited after the introduction of MM cells.\n\nSupplementary Material\n======================\n\n###### parameters.m represents the parameters in the NF-kB pathway.\n\nData accessibility {#s6}\n==================\n\nMatlab code is available in the electronic supplementary material.\n\nAuthors\\' contributions {#s7}\n=======================\n\nY.Z. carried out program coding and model simulations, participated in the design of the study and drafted the manuscript; C.Z. carried out the analysis of experimental data and critically revised the manuscript; Q.Y. collected the experimental data and critically revised the manuscript; B.J. conceived of the study, designed the study, coordinated the study and helped draft the manuscript. All authors gave final approval for publication and agree to be held accountable for the work performed therein.\n\nCompeting interests {#s8}\n===================\n\nWe declare that we have no competing interests.\n\nFunding {#s9}\n=======\n\nThis work was supported by National Key R&D Program of China (grant no. 2018YFB1305400); the National Natural Science Foundation of China (grant nos. 61673246 and 81301294); the key Research and Development Program of Shandong province (grant no. 2016GSF201168) and the Research and Development Program of Jinan (grant no. 201907064).\n\n[^1]: Electronic supplementary material is available online at <https://doi.org/10.6084/m9.figshare.c.4970645>.\n"}
+{"text": "*COI*\n\n:   *cytochrome c oxidase subunit I*\n\nDNA\n\n:   deoxyribonucleic acid\n\n*ITS*\n\n:   *internal transcribed spacer*\n\nPCR\n\n:   polymerase chain reaction\n\nRADSeq\n\n:   restriction site associated DNA sequencing\n\n*rRNA*\n\n:   *ribosomal ribonucleic acid*\n\nY2K\n\n:   year 2000\n\n1.. Introduction {#s1}\n================\n\n1.1.. DNA barcoding {#s1a}\n-------------------\n\nDNA barcodes were initially proposed as a solution to the worldwide shortage of taxonomic expertise for many groups of organisms [@b1]. Short stretches of sequenced DNA from a single gene from expertly diagnosed specimens from as many species as possible would serve as a database of identifiers or barcodes to facilitate the identification of additional unknown specimens collected in the future. For animals, a piece of the mitochondrial *cytochrome c oxidase subunit I* (*COI*) gene ([Table 1](#genetics-05-01-001-t01){ref-type=\"table\"}) was employed as the barcode sequence because of the availability of degenerate PCR primers that had been shown to consistently amplify a homologous DNA fragment in diverse organisms [@b2]. In fungi, a fragment of the *internal transcribed spacer 2* (*ITS2*) within the nuclear *ribosomal RNA* (*rRNA*) repeat was adopted as the barcode region [@b3]. In bacteria, variable portions of the *16S rRNA* are used for barcoding [@b4],[@b5] In plants, regions of 2 different chloroplast genes are used as a combinatorial barcode: the large subunit of *ribulose bisphosphate carboxylase* (*rbcL*) and *maturaseK* (*matK*) [@b6]. The choice and size of each of these barcode regions was based on the technologies for PCR amplification and DNA sequencing that were widespread when the barcode primers were designed [@b7]. DNA barcoding in animals has been the most controversial as discussed below and will be the focus of my comments here, but many of the challenges confronting DNA barcoding in animals also confront the barcoding strategies employed in other taxa. The suggested approach to overcoming these challenges should be broadly applicable to many groups of organisms.\n\nMitochondrial *COI-*based DNA barcodes are often incredibly helpful for associating morphologically distinct life stages or sexes within a species, identifying cryptic species, and understanding the diversity of species assemblages within a particular habitat or geographic region\u00a0[@b21]--[@b25]. The strengths of barcoding include the standardization of the identifiers across taxa (at least within Kingdoms---different barcode loci are used in animals, plants, fungi, and bacteria ([Table 1](#genetics-05-01-001-t01){ref-type=\"table\"})), the large number of species that have already been barcoded, and the fact that many different laboratories can contribute to a unified barcoding database using a variety of different experimental approaches and instrumentation [@b12],[@b26]--[@b30]. Because they are useful in many different contexts, a great deal of investment has been made in expanding the number of species and populations that have been DNA barcoded, as well as in computational resources for using the database of barcodes [@b12]. It is because of this utility that they have been so enthusiastically embraced by so many researchers\u00a0[@b17],[@b31]--[@b33].\n\n###### Commonalities between challenges presented by the Y2K problem and by DNA barcoding.\n\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                                                  **Y2K Problem**                                                                                                                                                               **DNA Barcoding**\n  --------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                                                  ![](genetics-05-01-001-i001.jpg)                                                                                                                                              ![](genetics-05-01-001-i002.jpg)\n\n  Identifier                                                      two-digit year notation                                                                                                                                                       animals: \u223c658 bp *COI* fragment\\\n                                                                                                                                                                                                                                                plants:\u223c553 bp *rbcL* & \u223c776 bp *matK* fragments\\\n                                                                                                                                                                                                                                                fungi:\u223c534 bp *ITS2* fragment\\\n                                                                                                                                                                                                                                                bacteria:\u223c1400 bp *16S rRNA* fragment\n\n  Identifier purpose                                              1\\. To distinguish between years\\                                                                                                                                             1\\. To facilitate species identification without reference to morphology\\\n                                                                  2. To sort data chronologically\\                                                                                                                                              2. To determine species limits and potentially to detect new species (by % divergence)\\\n                                                                  3. Mathematical operations (mostly addition/subtraction) to calculate time intervals [@b8]                                                                                    3. To understand species relationships [@b1],[@b9]\n\n  Constraints influencing the original design of the identifier   Extremely limited memory in early computers\\                                                                                                                                  1\\. \u223c500 bp maximum read length of radiolabeled dideoxy-terminated sequencing\\\n                                                                  (No longer applicable by 1983 when hard drives became common in personal computers [@b10])                                                                                    (No longer applicable by 1996 when automated dye-labeled sequencing radiolabeled sequencing [@b7])\\\n                                                                                                                                                                                                                                                2. Need for widely-conserved primer binding sites for PCR amplification from diverse organisms\\\n                                                                                                                                                                                                                                                (No longer applicable by 2007 when next generation sequencing methods were introduced which could recover high copy number genes without PCR [@b11])\n\n  Reason(s) for identifier maintenance                            To maintain backwards compatibility with older software applications\\                                                                                                         To take advantage of the large database of existing DNA barcodes\\\n                                                                  Reuse of old computer algorithms in newer code\\                                                                                                                               To facilitate comparisons of new results with those of previous studies [@b12]\\\n                                                                  Inertia/tradition/habit [@b10]                                                                                                                                                Cost\\\n                                                                                                                                                                                                                                                Inertia/tradition/habit [@b13]\n\n  Crisis                                                          At the turn of the 21st century: 2000 \\> 1999, but 00 \\< 99.\\                                                                                                                 1\\. Many recent species pairs cannot be separated by 658 bp barcodes because there has not been enough time for mutations to accumulate within the regions\u00a0[@b14].\\\n                                                                  Compromised sorting algorithms and mathematical operations [@b10]                                                                                                             2. Barcodes from organelle genomes in plants and animals are vulnerable to lateral transfer between species (through hybridization or other mechanisms) and reticulate evolution, sometimes resulting in misidentification [@b15],[@b16].\\\n                                                                                                                                                                                                                                                3. The barcode region reaches saturation quickly and cannot resolve deep phylogenetic nodes. (e.g., amphibians saturate at 10--11%, reptiles saturate at 9--10%, holometabolous insects saturate at 22%, and all hexapods saturate at 25% barcode sequence divergence) [@b17].\\\n                                                                                                                                                                                                                                                Collectively these issues compromise the general utility of DNA barcode application. It is highly desirable to produce more universal DNA barcodes that address these deficiencies [@b18]\n\n  Resolution                                                      Enlarge Identifiers: Worldwide effort to update software applications and change to four-digit identifiers in the late 1990s (acceptable until the year 9999) [@b10],[@b19]   Enlarge and diversify identifiers: The high copy number of organelle genomes and the nuclear rRNA repeat relative to the rest of the nuclear genome will cause these sequences to be very well represented among random reads of whole genome DNA extractions [@b20]. This permits routine assembly of complete organelle genomes (e.g., mitochondrial genome, \u223c15 kb) and complete rRNA repeat sequences (\u223c9 kb).\\\n                                                                                                                                                                                                                                                These longer sequences contain segments that evolve at different rates and have much higher information content than the short barcode sequences currently in use. Here I explore the use of these sequences as next generation barcodes to address the deficiencies of DNA barcodes currently in use.\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nHowever, other researchers have pointed out that while DNA barcoding uses the phylogenetic species concept (defining species as reciprocally monophyletic clades, with each clade possessing a distinct set of diagnostic characteristics [@b34]), its implementation often ran contrary to some of the guiding principles of phylogenetics [@b13],[@b35],[@b36]. For example, the tree-building algorithms implemented in the Barcode of Life Database (BOLD) are distance-based and do not use rigorous phylogenetic approaches to understanding relationships among barcodes [@b37],[@b38]. Also, by relying exclusively on sequences from organelle genomes, DNA barcoding activities in animals and plants are vulnerable to misleading associations between species and DNA barcodes because of the frequency of interspecific organelle capture due to hybridization or other kinds of lateral transfer events ([Table 1](#genetics-05-01-001-t01){ref-type=\"table\"}) [@b16],[@b39]--[@b42]. Finally, phylogenetic hypotheses (including the identification of monophyletic groups of barcodes associated with individual species) based on a small number of informative characters are prone to unresolved [@b43] or erroneous relationships [@b18],[@b44] between branches, as well as low bootstrap support [@b45]. This manifests both in recently diverged species that have not yet accumulated sequence variation within the barcode region [@b29] and in more distantly related species where multiple substitutions at the same sites obscure the phylogenetic signal within barcodes [@b17]. To the extent that any taxon defies the phylogenetic species concept\\'s fundamental criterion of reciprocal monophyly (due to lack of sequence divergence between species, retained ancestral polymorphisms, organelle capture, parallel evolution, etc.) the conventional DNA barcoding approach will fail, even for its original intended purpose of identifying species.\n\nThe question is whether it is possible to build upon the strengths of the DNA barcoding strategies currently in use, while also addressing these deficiencies. In this work, I will suggest that the Y2K problem from computer science [@b10] shares a number of common features with DNA barcoding, and so the solutions to the challenges of DNA barcoding in its current implementation may also share some similarities to how the Y2K problem was resolved.\n\n1.2.. The Y2K problem {#s1b}\n---------------------\n\nThe Y2K problem in computer sciences ([Table 1](#genetics-05-01-001-t01){ref-type=\"table\"}) refers a feature of many computer programs in the late 1990s that encoded year identifiers with only two digits (for example: such that the year \"1997\" would be identified by \"97\") [@b8]. The practice of using two-digit identifiers dates to the early days of computer science when computer memory was at a premium, and so the minimum number of digits necessary to encode year identifiers was employed. In turn, the adequacy of minimal two-digit year identifiers was due to the historical contingency of the birth of computer science in the 1940s and 1950s, decades away from both the turn of the 20^th^ century, and the turn of the 21^st^ century [@b10].\n\nThe practice of using two-digit identifiers was perpetuated over decades, long after the constraints of computer memory no longer applied, as code from earlier applications was reused in subsequent applications, perhaps abetted by habits and traditions adopted by programmers over time [@b10]. For much of the 20^th^ Century, this practice was unproblematic, but as the year 2000 approached it became apparent that this usage would be problematic for distinguishing between years from different centuries (1901 vs 2001), for sorting data chronologically (01 \\< 99, but 2001 \\> 1999), and for mathematical operations (01 \u2212 99 = \u221298, but 2001 \u2212 1999 = 2). The solution to the Y2K problem in the late 1990s, was a coordinated worldwide effort to update computer software so that they employed four-digit year-identifiers [@b10],[@b19]. The choice of four-digit identifiers was arbitrary, but should allow for upgraded computer software to function as expected until the year 9999.\n\nThe sizes of the DNA barcodes now in use are similarly arbitrary. They were selected on the basis of several factors including the availability of conserved primers that would amplify the barcode region in diverse organisms, and the availability of a large collection of previously sequenced examples of the region from many species [@b1]. Both of these factors were greatly influenced by the early experimental methods being used to acquire barcode sequences. The *COI* primers that came to be the standard for animal barcodes were described in 1994 [@b2] when most DNA sequencing was done by ^32^P-labelled dideoxy-terminated Sanger sequencing, which has a maximum read length of about 500 bp [@b7]. By sequencing in both directions, all of the 658 bp *COI* fragment could be covered, with bidirectional coverage over most of the more variable interval between the two more conservative binding sites. Radiolabelled Sanger sequencing was largely replaced by fluorescent dye-labeled Sanger sequencing, by about 1996, when the cost of the dye-labeled technology dropped below the cost of radio-labeled sequencing [@b7]. Yet, the same 658 bp region continued to be extremely popular for phylogenetic studies [@b46],[@b47], and for DNA barcoding initiatives (which were introduced in 2003) [@b1], even though fluorescent dye-labeled Sanger sequencing has much longer maximum read lengths.\n\nBeginning in 2007, when next generation sequencing technologies began to be adopted by the research community [@b11], it became possible to recover high copy number sequences like the mitochondrial genome (including the *COI* barcode region) by low-coverage shotgun sequencing of the whole genome or \"genome skimming\" [@b48]--[@b51] without requiring the use of conserved PCR primers flanking the barcode region. Similarly, the chloroplast genome of plants and the nuclear *rRNA* repeat (which contains the *18S*, *5.8S*, and *28S rRNAs* and *internal transcribed spacer* (*ITS)* 1 and 2 sequences) that also occur at high copy number are also easily recovered by genome skimming [@b49]. It has already been demonstrated that these high copy number sequences contain substantial phylogenetic signal [@b50],[@b51]. Complete mitochondrial genome sequences in particular have a good track record for reconstructing the phylogenetic history of organisms at a range of taxonomic (and sequence) divergence [@b52]--[@b56].\n\nWhile it is still more affordable to use traditional DNA barcodes, but this may always not be the case. In 2018, it costs roughly \\$10 USD to sequence a 658 bp *COI* barcode PCR product in both directions by dye-terminated Sanger sequencing. In comparison, to sequence a complete mitochondrial genome and a complete nuclear rRNA repeat by genome skimming using an Illumina MiSeq instrument as described in this paper costs approximately \\$250 USD per sample, while yielding a vastly larger pool of sequence data for analysis. At the same time, while the cost of Sanger sequencing has been fairly stable for the last 15 years (2002--2017), the cost of next generation sequences dropped from \\$0.08 USD per raw Megabase of DNA sequence to less than \\$0.02 USD per raw Megabase between 2014 and 2017 [@b57]. It is therefore becoming increasingly affordable to generate next generation sequence datasets for the purposes of species identification and phylogenetics.\n\nOnce a system of identifiers exists, it is often used for purposes not envisioned when it was first created. A fundamental element of the analogy between the Y2K problem and DNA barcoding is that in both cases, many of the fundamental limitations of the identifiers were not apparent until practitioners attempted to extend their use to novel situations. In the case of two-digit year identifiers, this was when then number of years being identified exceeded 100; while in the case of DNA barcodes, it was when researchers tried to use DNA barcode identifiers for molecular phylogenetic and population genetic analysis beyond mere species identification.\n\nIt is to address two issues that I am arguing that a major modification of the current DNA barcoding strategy may be warranted. First, that DNA barcodes as currently implemented are imperfect tools even for their original intended purpose of species identification. Second, that the current DNA barcodes are inadequate or inappropriate for many of the applications for which many researchers wish to use molecular species identifiers. As an explicit analogue to the decision to expand the number of digits in year identifiers in order to resolve the Y2K problem, this is an opportune time to see if easily (and increasingly inexpensively) obtained plastid genome and nuclear ribosomal repeat sequences might be used as larger \"next generation\" DNA barcodes that might be less vulnerable to some of the deficiencies of the short conventional barcode sequences that are currently in use [@b18]. That is not to say that the current DNA barcodes have no role to play (my research group uses them frequently in our own work [@b44],[@b58]--[@b60]), but rather that we may be able to design next generation barcodes that have all of the positive attributes of the current identifiers, while eliminating most of the limitations that have plagued DNA barcoding efforts to date.\n\n1.3.. The test case: the net-spinning caddisflies (Insecta: Trichoptera: Hydropsychidae) {#s1c}\n----------------------------------------------------------------------------------------\n\nFor a test case to explore whether enlarging barcodes can improve their performance for both species identification and phylogenetic analysis, I chose to examine caddisflies (Insect Order Trichoptera) with a focus on Family Hydropsychidae, the net-spinning caddisflies. The species in the Hydropsychidae are distinctive because they spin nets made of silk that they use to harvest food particles from the water column in their larval aquatic environment [@b61]. The Trichopera demonstrate some of the most frustrating limitations of current barcodes: the family-level phylogeny of the Trichoptera cannot be reconstructed on the basis of barcode sequences [@b62] and within the Hydropsychidae, there are many similar species that can be difficult to distinguish on the basis of morphology (especially as larvae) and that also cannot always be distinguished on the basis of *COI* barcodes due the presence of shared haplotypes [@b12],[@b63].\n\nThe family-level phylogeny of the Trichoptera is well established on the basis of multiple datasets [@b64],[@b65], as is the sister clade relationship between the Trichoptera and the Lepidoptera (butterflies and moths) [@b66],[@b67], but the species-level phylogeny of many of the 14,500 described trichopteran species is unknown. It has recently been proposed that *COI* barcodes can and should be used to arrange the terminal branches of the caddisfly phylogeny, in combination with more extensive sequence data from other genetic regions from select species to establish the backbone and deeper nodes of the tree [@b65]. Yet, in some other taxonomic groups, barcode-based phylogenies are not good predictors of the phylogeny of the mitochondrial genomes of which they are a part\u00a0[@b18],[@b44]. Assessing the predictive validity of *COI* barcode-based phylogenies in the Trichoptera would be very helpful to determine if the proposed strategy for resolving terminal branches [@b65] is likely to be successful. Therefore the explorations of conventional versus next generation barcode approaches considered here will yield valuable insights both for future phylogenetic research in the Trichoptera and for all taxa more generally.\n\nTo evaluate the effectiveness of these different approaches to barcoding, datasets of *COI* barcodes and mitochondrial genomes were assembled for 14 trichopteran species, and a dataset of *rRNA* repeats was assembled for 8 trichopteran species. All of the *rRNA* repeat sequences and all but 6 of the mitochondrial genome sequences were collected and assembled by my laboratory for this study. The data sets include 3 species of *Cheumatopsyche* (Hydropsychidae) that are not readily distinguishable by *COI* barcodes, 3 species of *Hydropsyche* (Hydropsychidae) that are distinguishable by *COI* barcodes, and 1 species of *Potamyia* (Hydropsychidae), as well as 7 caddisfly species from 6 other trichopteran families. Also included in the data sets were sequences from representatives of 2 lepidopteran families as outgroups. These analyses consider all of the publicly available complete mitochondrial genomes and *rRNA* repeats for the Trichoptera.\n\n###### Caddisfly species collected at the Living Prairie Museum and analyzed in this study.\n\n                                                                                              Million   Mitochondrial Genome   Nuclear *rRNA* Repeat                           \n  ------------------------------------------------------------ ----------- ----------------- --------- ---------------------- ----------------------- ------- ------- -------- ------\n  [Hydropsychidae]{.ul}                                                                                                                                                        \n  *Cheumatopsyche analis[^1^](#nt101){ref-type=\"table-fn\"}*     14-Aug-15   2015.08.14.065A    6.84            67275                  1266 X           15097   9412    308 X    7791\n  *Cheumatopsyche campyla[^1^](#nt101){ref-type=\"table-fn\"}*    17-Jul-15   2015.07.17.021A    5.89            67559                  1275 X           15100   6239    122 X    8323\n  *Cheumatopsyche speciosa*                                     14-Aug-15   2015.08.14.106A    7.75            37191                   184 X           15098   29449   548 X    8683\n  *Hydropsyche orris*                                           14-Aug-15   2015.08.14.066A    2.08            86392                   458 X           15185   29054   640 X    9228\n  *Hydropsyche simulans*                                        14-Aug-15   2015.08.14.067     6.30            15864                   326 X           15237   31093   301 X    7797\n  *Potamyia flava*                                              14-Aug-15   2015.08.14.070B    4.59            122730                  600 X           15160   53095   1222 X   9244\n  [Limnephilidae]{.ul}                                                                                                                                                         \n  *Anabolia bimaculata*                                         17-Jul-15   2015.07.17.018     8.29            40865                   482 X           15048   98766   3149 X   9400\n  [Leptoceridae]{.ul}                                                                                                                                                          \n  *Triaenodes tardus*                                           14-Aug-15   2015.08.14.077     8.36             6952                   35 X            14963   82832   168 X    9232\n\n^1^Read length for *C. analis* and *C. campyla* was 300 bp. For all other species, read length was 75 bp.\n\n2.. Materials and Methods {#s2}\n=========================\n\n2.1.. Specimen Collection and DNA Preparation {#s2a}\n---------------------------------------------\n\nAdult caddisflies (Insecta: Trichoptera) were collected by USDA blacklight trap containing ethyl acetate [@b68] deployed overnight as part of a taxonomic inventory of arthropods at the Living Prairie Museum in Winnipeg, Manitoba, Canada (GPS 49.889607 N, \u221297.270487 W) during the 2015 growing season. The Living Prairie Museum consists of 12.9 hectares of relict unplowed prairie maintained by periodic controlled burns and is home to over 160 native plant species, supporting a rich arthropod fauna [@b69]. Nearby aquatic habitats suitable for larval caddisflies include Sturgeon Creek (0.57 km) and the Assiniboine River (1.92 km). Light trap collections were brought back to the laboratory, sorted to species by morphology, and then stored in glassine envelopes at \u221220 \u00b0C before further processing. Specimens collected as part of the inventory that are used in this study include 6 species in trichopteran family Hydropsychidae, 1 species in family Limnephilidae [@b70], and 1 species in family Leptoceridae [@b71], and are listed in [Table 2](#genetics-05-01-001-t02){ref-type=\"table\"}.\n\nFor each caddisfly species, DNA was extracted from abdominal tissues from each specimen using the DNEasy Blood and Tissue kit (Qiagen, D\u00fcsseldorf, Germany) following the standard animal tissue extraction protocol with modifications as previously described [@b58]. Tissue was ground up in 180 \u00b5L of tissue lysis buffer ATL (Qiagen) using a mortar and pestle followed by 20 \u00b5L of protein kinase K (Qiagen, 600 mU/mL) which was added to the mixture and then incubated in a 55 \u00b0C water bath for 1 hour. Using the standard instrument protocol for purification of total DNA from animal tissue [@b59], the samples were processed on a QiaCube extraction robot (Qiagen) to complete the DNA extraction procedure. Extracted DNA was evaluated for yield and quality on a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, Delaware, USA) and a Qubit 2.0 fluorometer (Life Technologies, Carlsbad, California, USA). DNA was stored in Eppendorf tubes (Eppendorf, Hamburg, Germany) at \u221220 \u00b0C until required [@b44].\n\nThe morphology-based identification of each species was further examined by *cytochrome c oxidase I* DNA barcoding. Polymerase chain reaction (PCR) products for the *COI* barcode sequence were obtained and sequenced for each specimen using standard methods [@b44],[@b72]. Sequences for each specimen were compared with reference sequences in the BOLD database [@b12], and in all cases yielded a species diagnosis consistent with that previously determined from morphological characteristics (data not shown).\n\n2.2.. Sequence preparation, assembly, and annotation {#s2b}\n----------------------------------------------------\n\nDNA libraries were prepared and samples were sequenced at the Next Generation Sequencing (NGS) Platform facility at the Children\\'s Hospital Research Institute of Manitoba (Winnipeg, Manitoba, Canada). The DNA sample was sheared by sonication with an S220 Focused-Ultrasonicator (Covaris, Woburn, Massachusetts, USA). Fragment sizes were evaluated using a High Sensitivity DNA chip for the Bioanalyzer 2100 electrophoresis system (Agilent, Santa Clara, California, USA) using the standard manufacturer protocol. A TruSeq library preparation kit (NEB) was used to prepare an indexed library from each sheared sample for loading onto a MiSeq NextGen Sequencing Instrument equipped with either a MiSeq reagent V3 75X2 paired end reagent kit (6 samples) or a V3 300X2 paired end reagent kit (2 samples) (Illumina, San Diego, California, USA). In both cases, the specimens included in this study were processed simultaneously on the instrument with several other indexed libraries that will be described separately in future work. The sequences for each of the species included in this study represents about 10% of the data generated from a run of the MiSeq instrument.\n\nThe assembly process for *Anabolia bimaculata* (Trichoptera: Limnephilidae) and for *Triaenodes tardus* (Trichoptera: Leptoceridae) has already been described [@b70],[@b71]. For trichopteran species in family Hydropsychidae, the sequence reads for each species were assembled to the full mitochondrial genome reference sequence (GenBank voucher MF680449) and the complete ribosomal RNA repeat (GenBank voucher MF680448) from *A. bimaculata* [@b70] using Geneious version 10.1.2 [@b73]. In each case, MiSeq reads were mapped to the voucher sequence in 25 iterations at the \"Medium-Low Sensistivity/Fast\" setting of Geneious. In rare cases where the assembly produced large gaps (\\>30 bp), 1 kb of the consensus sequence of the assembly on each side of the gap were used as reference sequences for mapping the MiSeq reads in 5 iterations at the \"Medium-Low Sensistivity/Fast\" setting of Geneious. The resulting assemblies from both sides of the gap were then aligned with each other and with the original assembly that was mapped to *A. bimaculata* to produce a continuous sequence for the mitochondrial genome and the *rRNA* repeat for each species.\n\nSequences were annotated in Geneious. Secondary RNA structures were analyzed using the default settings of RNAstructure [@b74] and Mfold [@b75] software. Annotation of mitochondrial genes was facilitated by comparison with the mitochondrial genome sequences of *A. bimaculata* and *Eubasilissa regina* (Trichoptera: Phryganeinae, Genbank voucher NC_023374 [@b76]). Annotation of nuclear *rRNA* repeats was facilitated by comparison with *rRNA* repeat reference sequences from *A. bimaculata*, *T. tardus* (Genbank voucher MG201853 [@b71]), *Meroptera pravella* (Lepidoptera: Pyralidae, Genbank voucher MF073208 [@b69]), and *Samia cynthia ricini* (Lepidoptera: Saturniidae, Genbank voucher AF463459 [@b77]).\n\n2.3.. Phylogenetic analysis {#s2c}\n---------------------------\n\nMitochondrial genome sequences from specimens collected at the Living Prairie Museum (Genbank Vouchers MF680449, MG201852, MG669121-MG669126) were combined with Trichopteran mitochondrial genome sequences from other geographic regions obtained from Genbank (Vouchers AB971912, KF756944, KP455290, KP455291, KT876876, NC 023374) previously published by other research groups [@b76],[@b78],[@b79]. These full-length mitochondrial genome sequences were then aligned with lepidopteran outgroups *M. pravella* and *S. cynthia ricini* (Genbank vouchers NC 017869, MF073207) [@b69],[@b77] in CLUSTAL Omega [@b80]. The nuclear *rRNA* repeats from each of the Living Prairie Museum Trichoptera (Genbank Vouchers MF680448, MG201853, MG669127-MG669132) were aligned with only the repeats from Lepidoptera outgroups *M. pravella* and *S. cynthia ricini* (Genbank vouchers MF073208, AF463459) [@b69],[@b81] because these are the first complete Trichopteran *rRNA* repeats to be reported.\n\nThe aligned mitochondrial genome and nuclear rRNA repeat sequences were each analyzed using the parsimony and maximum likelihood heuristic and bootstrap search algorithms implemented in PAUP\\* version 4.0b8/4.0d78 using default settings unless otherwise specified [@b82]. The best model for maximum likelihood phylogenetic analysis of both datasets were identified using jModeltest 2.1.7 [@b83] and likelihood ratio tests [@b84] and were determined in both cases to be the GTR + I + G (General Time Reversible) model (mitochondrial genomes: I = 0.1770, G = 1.0130, *rRNA* repeats: I = 0.3260, G = 0.6360). Parsimony and maximum likelihood (GTR + I + G) heuristic searches were carried out on the 658 bp barcode region of *COI* within the mitochondrial genome, the complete mitochondrial genome, and the complete nuclear *rRNA* repeat. Searches of each dataset were conducted using the following settings: 1 million maximum search replicates with random sequence addition, tree bisection and reconnection branch swapping on only the best trees, multiple trees saved at each step, and retention of the best trees. The bootstrap searches were conducted using the following settings: 1 million random sequence addition fast addition search replicates and retention of all groups compatible with 50% bootstrap consensus.\n\n3.. Results {#s3}\n===========\n\n3.1.. Mitochondrial genome and nuclear rRNA repeat assemblies {#s3a}\n-------------------------------------------------------------\n\nMitochondrial genomes and nuclear *rRNA* repeats were assembled for six caddisfly species in family Hydropsychidae and one species each in families Limnephilidae [@b70] and Leptoceridae [@b71]. Assemblies of the mitochondrial genome sequences ranged from 14,963 bp (*Trianodes tardus*, family Leptoceridae) to 15,185 bp (*Hydropsyche orris*, family Hydropsychidae) ([Table 2](#genetics-05-01-001-t02){ref-type=\"table\"}). Most of the variation in sequence length in the mitochondrial genome between caddisfly species was in the control region, a noncoding sequence that services as an origin of replication for the mitochondrial genome and that is responsible for regulating transcription of mitochondrial genes [@b20]. The same gene order and arrangement was found in all of the mitochondrial genomes assembled in this study. It is the same as the ancestral mitochondrial gene order for all insects [@b76] and has been found in all other caddisfly species sequenced to date except for *Hydropsyche pellucidula* [@b71],[@b78].\n\nNuclear *rRNA* repeat assemblies varied in size from 7791 bp (*Cheumatopsyche analis*, family Hydropsychidae) to 9400 bp (*Anabolia bimaculata*, family Limnephilidae). Most of the variation in sequence length was in the 5\u2032 external transcribed spacer and the 5\u2032 non-transcribed spacer regions of the *rRNA* repeat [@b81]. Also present was some sequence length variation in ITS2, the most variable internal region of the repeat [@b85]. The gene order observed in the caddisfly *rRNA* repeats was identical in all species considered here and identical to that observed in most eukaryotic organisms [@b3].\n\n3.2.. Phylogenetic analyses {#s3b}\n---------------------------\n\nPhylogenetic analysis of the *COI* barcode dataset produced four most parsimonious trees (length 1020 steps), one of which was identical to the single maximum likelihood tree (likelihood score 5040.9131) produced by this dataset ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). The most parsimonious trees differed from one another in the relationship of *Hydropsyche* species to one another, and in the placement of genus *Potamyia* relative to the genera *Hydropsyche* and *Cheumatopsyche*. In all cases, analysis of barcode sequences resulted in unresolved relationships between species in genus *Cheumatopsyche*. Strong maximum likelihood and parsimony bootstrap support was observed for the monophyly of insect order Trichoptera, for family Hydropsychidae, and for genera *Hydropsyche* and *Cheumatopsyche.* Except for moderate bootstrap support for family Limnephilidae, all of the other relationships between taxa had only weak bootstrap support from the *COI* barcode dataset.\n\nPhylogenetic analysis of the mitochondrial genome dataset produced a single most parsimonious tree (length 31583 steps) with the same topology as the single maximum likelihood tree (likelihood score 158974.51910) ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). Species relationships in genera *Hydropsyche* and *Cheumatopsyche* were fully resolved by phylogenetic analysis of the mitochondrial genome. Bootstrap support was robust throughout, except for the node supporting the sister-clade relationship between *Potamyia* and *Cheumatopsyche*, where bootstrap analysis was more modest.\n\n![Phylogenetic tress reconstructed from *COI* barcodes (left) and complete mitochondrial genomes (right) using maximum likelihood and parsimony. Asterisks indicate where some of the four most parsimonious *COI* trees differ from the tree topology shown here. Portions of the phylogenetic tree that are congruent between the analyses of the *COI* and the mitochondrial genome datasets are indicated by bold lines on the tree. Maximum likelihood bootstrap values are shown above each node, parsimony bootstrap values are shown below the node.](genetics-05-01-001-g001){#genetics-05-01-001-g001}\n\nMonophyly of the Trichoptera, the Integripalpia, families Hydropsychidae and Limnephilidae, and genera *Hydropsyche* and *Cheumatopsyche* were supported by both *COI* barcode and mitochondrial genome datasets (bold branches on the phylogenetic trees, [Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). However, virtually all of the remaining relationships among taxa differ substantially between the trees generated from these 2 datasets derived from the mitochondrion.\n\nPhylogenetic analysis of the nuclear *rRNA* repeat dataset produced trees with very similar topologies with as the result of parsimony (length 13211 steps) and maximum likelihood (score 65074.87230) searches, but the two methods reconstructed different relationships within genus *Cheumatopsyche* ([Figure 2](#genetics-05-01-001-g002){ref-type=\"fig\"}). Bootstrap support was very strong at most nodes, except for monophyly of the Hydropsychidae, the sister-relationship between genera *Potamyia* and *Hydropsyche*, and relationships within *Cheumatopsyche* (maximum likelihood only). In most cases, the *rRNA* repeat-based analyses also produced relationships among taxa that were the same as was found in phylogenetic analysis of the mitochondrial genome (bold branches on the phylogenetic trees, [Figure 2](#genetics-05-01-001-g002){ref-type=\"fig\"}) including the monophyly of order Trichoptera, the Integripalpia, families Hydropsychidae, and genera *Hydropsyche* and *Cheumatopsyche.* Exceptions to the congruency between mitochondrial genome-based and *rRNA* repeat-based trees are the relationships within genus *Cheumatopsyche* and the relationship of genus *Potamyia* to the genera *Hydropsyche* and *Cheumatopsyche.*\n\n![Phylogenetic tress reconstructed from nuclear *rRNA* repeats using maximum likelihood (left) and parsimony (right) methods. Portions of the phylogenetic tree that are congruent between the analyses of the nuclear *rRNA* repeat and mitochondrial genome datasets are indicated by bold lines on the trees. Maximum likelihood bootstrap values are shown above each node, parsimony bootstrap values are shown below the node.](genetics-05-01-001-g002){#genetics-05-01-001-g002}\n\n4.. Discussion {#s4}\n==============\n\n4.1.. Caddisfly mitochondrial genome structure {#s4a}\n----------------------------------------------\n\nAssembled mitochondrial genomes and nuclear *rRNA* repeats can easily be recovered from shallow next generation sequencing of total cellular DNA (sometimes called genome skimming) due to their repetitive nature, as has been found in prior studies [@b49]--[@b51],[@b86]. The gene order and overall structure of caddisfly mitochondrial genomes is very consistent both among newly assembled sequences presented here and among previously reported mitochondrial genome sequences\u00a0[@b70],[@b71],[@b76],[@b79]. The only exception to this general pattern is the sequence reported from *Hydropsyche pellucidula* (Hydropsychidae) that differs from other Trichopteran mitochondrial genomes in size (25 kb versus the typical \u223c15 kb), in the arrangement of the mitochondrial rRNA genes (the *12S rRNA* was translocated from its usual position between the *16S rRNA* and the control region to a position between *cytochrome b* and *nad1*), in the atypical locations of *tRNA-P* and *tRNA-I*, and in topology (with a possibly linear mitochondrial genome structure) [@b78].\n\nThe mitochondrial genome sequences of *H. orris* and *H. simulans* share none of these features, suggesting that the reported rearrangements of the mitochondrial genome in *H. pellucidula* are probably either of very recent origin (occurring since the diversification of *Hydropsyche*) or may be attributable to experimental artifact. The *H. pellucidula* mitochondrial genome sequence is from an experiment involving the next generation sequencing of a metagenomic library containing multiple taxa followed by de novo assembly [@b78] so it is possible that contaminating sequences may have inadvertently been incorporated into the published sequence. It may be worthwhile to resequence and/or reassemble the *H. pellucidula* mitochondrial genome in order to verify the existence and timing of the reported rearrangements, since these features of the mitochondrial genome are often very useful for understanding relationships among insect taxonomic groups [@b87]. In any case, because the unique features of the reported *H. pellucidula* mitochondrial genome are autapomorphic, they are expected to have virtually no effect on the phylogenetic analyses performed in this study.\n\n4.2.. Performance of phylogenetic datasets {#s4b}\n------------------------------------------\n\n*COI* barcode-based phylogenetic analysis of the caddisfly species considered here produced multiple unresolved trees with poor bootstrap support both for species relationships within genera, and among the lineages representing different caddisfly families ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). The topology of the *COI* trees is also incongruent with relationships reported in previous analyses both between genera within family Hydropsychidae [@b88] and between trichopteran families [@b64],[@b65].\n\nIn contrast, phylogenetic analysis of complete mitochondrial genomes produces fully resolved trees with robust bootstrap support at nearly all nodes ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). With one exception, the mitochondrial genome-based phylogenetic relationships among the trichopteran families match those proposed previously on the basis of other data sets [@b64],[@b65]. The principal difference is that in this analysis, the Apantaniidae and the Limnephilidae are sister clades, with the Uenoidae as a near outgroup; while in prior analyses the Apantaniidae and Uenoidae were sister clades, with the Limnephilidae as the outgroup. This part of the trichopteran phylogenetic tree (which included several other families not sampled in the current study) was not fully-resolved by prior work based on smaller data sets [@b64], so it is not surprising that the larger number of informative characters found in the mitochondrial genomes introduces some changes in this portion of the topology.\n\nMaximum likelihood and parsimony methods for phylogenetic reconstruction produced nearly identical tree topologies from the *rRNA* repeat dataset, except for the arrangement of species within the genus *Cheumatopsyche* ([Figure 2](#genetics-05-01-001-g002){ref-type=\"fig\"}). Neither of these topologies matches the arrangement of *Cheumatopsyche* species in the mitochondrial genome-based tree (or the arrangement in the *COI* barcode tree where the basal *Cheumatopsyche* node is unresolved) ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}). There are several factors that might contribute to these patterns of phylogenetic discordance among these very closely related (and in all probability, recently diverged) species, including retained ancestral polymorphisms, lateral transfer of mitochondria between lineages, and possibly selection on the mitochondrial genome and/or the nuclear *rRNA* repeat [@b89]. Other than the relationships within *Cheumatopsyche*, the remainder of the nuclear *rRNA* repeat-based trees is topologically similar to the mitochondrial genome-based tree, with one exception. In the *rRNA* repeat tree, *Potamyia* is sister to genus *Hydropsyche*, in agreement with the weakly supported arrangement from the *COI* barcode tree, but incongruent with the strongly supported sister relationship between *Potamyia* and *Cheumatopsyche* from the mitochondrial genome data set in this study ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}) and by a prior study [@b88]. Bootstrap analysis of the *rRNA* dataset shows robust bootstrap support at most nodes, except for some of the nodes that differ between this data set and the mitochondrial genome-based dataset. Better taxon sampling of additional genera in the Hydropsychidae as well as additional families in the Trichoptera will help to break up the longer branches of the phylogenetic tree may aid in improving the robustness of the reconstructions for some of these nodes in the *rRNA* repeat trees [@b90].\n\n4.3.. Building a better barcode {#s4c}\n-------------------------------\n\nInvestigators have widely different perspectives on the value of conventional DNA barcoding strategies, much of which can be traced to whether these strategies are appropriate for answering the questions being addressed within their research programs [@b14]. Researchers whose primary interest is to identify unknown specimens, to associate morphologically disparate individuals (due to life stage, sexual dimorphism, or other kinds of variation) from the same species, or to quantify individuals of a given mitochondrial haplotype in the environment might be completely adequately served by the short barcode sequences (including the *COI* fragment) currently in use [@b29],[@b63],[@b91], at least in some taxonomic groups [@b13]. Others with research questions that require inferences about the relationships among organisms often find that the information content of *COI* barcodes is insufficient for their purposes [@b18],[@b37],[@b53]. However, even subjects such as species delimitation [@b92],[@b93], cryptic species identification [@b21],[@b37], or hybridization and organelle capture [@b16],[@b41],[@b94], research topics for which DNA barcoding is supposedly highly suitable [@b14], cannot be demonstrated without reference to sequences from the nuclear genome, examination of morphological or other phenotypic traits, or both [@b13],[@b95]. Simply stated, the limited information content of conventional plastid-based barcodes due to their small size, and their propensity for lateral transfer between lineages due to their location within the mitochondrial genome means that they cannot be used to effectively address certain research questions [@b13],[@b14],[@b18],[@b35],[@b37],[@b53].\n\nThe desirability of a standard set of genetic markers that can be used to identify nearly any organism is clear, and if the phylogenetic species concept is to be invoked in order to operationalize species identification, the markers used should possess qualities that will allow them to be phylogenetically informative. The short barcode sequences currently in use were developed in the context of experimental technologies that placed constraints on the size and location of the barcode regions that are no longer universally applicable ([Table 1](#genetics-05-01-001-t01){ref-type=\"table\"}). While some have argued that next generation sequencing may make DNA barcoding obsolete (e.g., [@b13],[@b53]), I am more convinced by the argument that these new sequencing technologies may revolutionize DNA barcoding [@b96], because, akin to the solution of the Y2K problem, they allow the expansion of sequence identifiers so as to increase their information content. In particular, by using next generation sequencing to expand the barcode sequence regions to encompass entire plastid genomes by genome skimming, this greatly increases our ability to distinguish between species that have recently diverged from one another by sampling more sites that might have undergone mutation [@b44]. Similarly, plastid genomes include genes that evolve at dramatically different rates [@b13],[@b72], and by sampling and sequencing them in their entirety it becomes possible to resolve deeper phylogenetic nodes with robust bootstrap support that are unresolved in conventional barcode-based trees ([Figure 1](#genetics-05-01-001-g001){ref-type=\"fig\"}) [@b18]. Even better, because mitochondrial genomes contain the *COI* sequence and chloroplast genomes include the *rbcL* and *matK* genes, they are \"backwards compatible\" with the conventional barcodes currently in use.\n\nIf combinatorial analyses involving sequences from more than one source are included within the definition of barcoding (as already in use for plant barcoding [@b6]), one can further expand the sources of DNA barcode information to include the nuclear *rRNA* repeat which in turn includes the *ITS2* region used for DNA barcoding in fungi [@b3] and the *28S rRNA*, the eukaryotic homologue of the *16S rRNA* used for barcoding in bacteria [@b4],[@b5]. Unlike plastid genomes, the nuclear *rRNA* repeat is bi-parentally inherited as a component of the nuclear genome [@b97], and thus may not be as prone to phylogenetic distortions due to lateral transfer and organelle capture. Even when lateral transfer is unlikely, the phylogenetic hypotheses produced by organelle genomes and the *rRNA* repeat may not always be congruent, (as is the case between the genera *Cheumatopsyche*, *Potamyia*, and *Hydropsyche*, in this study, see [Figures 1](#genetics-05-01-001-g001){ref-type=\"fig\"} and [2](#genetics-05-01-001-g002){ref-type=\"fig\"}), but each of these genera are clearly distinguished by both data sets, and this may set the stage for further work examining additional genetic regions if the relationships of these genera are of particular importance for addressing a given research question. In plants, where chloroplast genomes, mitochondrial genomes, and nuclear *rRNA* repeats can be recovered from genome skimming relatively easily [@b48],[@b51] it may be possible for researchers to \"triangulate\" and use the phylogenetic signal from all 3 sources to identify specimens and reconstruct the evolutionary history of a group.\n\nWhat I am proposing is by no means the only way to employ next generation sequencing methods for phylogenetic reconstruction, but it does have several advantages over some of the available alternatives. For example, a recently announced set of PCR primers for amplifying 30 nuclear genes in the Lepidoptera, comprising some 11 kb of DNA sequence, is based entirely on slowly evolving coding sequences [@b98], potentially making them valuable for resolving interfamilial relationships, but likely with limited applications for more recent species divergences. It is also unknown to what degree the primers for these nuclear regions will work in taxa outside of the Lepidoptera and there are few comparable sequences in the databases to compare with data generated from these primers. Conversely, restriction site associated DNA sequencing (RADSeq) can be used to generate extremely large phylogenetic data sets that are often quite valuable for resolving relationships among closely related species [@b40]. However, as more taxa and especially more distantly related data are included in the analysis, the proportion of missing data increases and the proportion of informative characters often decrease in RADSeq data sets. Thus, both of these approaches lack features of standardization that are present in next generation barcodes, making them less attractive for answering certain kinds of research questions.\n\nMore compatible with next generation barcodes are mitogenomic approaches which enrich target sequences by PCR [@b52],[@b54],[@b99] and target enrichment approaches which use anchored probes to pull target sequences out of genomic pools, followed by sequencing [@b100]. If probes are designed to match (or primers are designed to amplify) entire mitochondrial genomes, entire chloroplast genomes, and complete nuclear *rRNA* repeats, these approaches may allow larger numbers of species to be included within a single lane or run of a next generation sequencer. Target enrichment by anchored probes may be a particularly effective method for obtaining the sequences of these regions from rare species preserved as specimens in museum collections and may ultimately also be a cost-effective method for extracting fragments from these genetic regions from recently collected specimens as well. The opportunity to deploy next generation barcodes to answer research questions using all of these methods is expected to continue to increase as next generation sequencers become more available and as the cost of running the instruments continues to drop.\n\n5.. Conclusion {#s5}\n==============\n\nWhile several other authors have suggested that next generation sequencing will have profound effects on how barcoding is conducted [@b13],[@b26],[@b53],[@b96], this is the first explicit proposal that genome skimming by next generation shotgun sequencing be used to enlarge conventional DNA barcode identifiers and upgrade current barcoding strategies. This expansion increases the information content of barcodes, giving them properties that are appropriate for the statistical operations used in phylogenetic analysis. This is analogous to expanding the year identifiers to resolve the Y2K problem in computer science so that they have a sufficient number of digits to permit mathematical operations and chronological sorting [@b10]. The improved functionality of the proposed next generation barcodes over conventional *COI* barcodes was demonstrated by comparing their effectiveness in reconstructing the phylogenetic history of the Trichoptera.\n\nThank you to Xuhua Xia for inviting me to participate in this special issue. I am grateful to Sarah Semmler and Kyle Lucyk for permitting and encouraging my laboratory\\'s work at the Living Prairie Museum. Melissa Peters and Ashley Haverstick helped with fieldwork; Melanie Lalonde and Daniel Peirson helped in the laboratory; and Aleksandar Ilik and Debbie Tsuyuki (Children\\'s Hospital Research Institute of Manitoba Next Generation Sequencing Platform) assisted with library preparation and sequencing. Melane Lalonde and two anonymous reviewers provided many insightful comments on drafts of this paper. This work receives support from the University of Manitoba Research Grants Program, the Faculty of Science Field Work Support Program, and the CATL Teaching and Learning Enhancement Fund, as well as from NSERC under Grants RGPIN386337-2011 and RGPIN-2016-06012.\n\n**Conflicts of interest:** The author has no conflict of interest to declare.\n"}
+{"text": "Tuberculosis (TB) is a leading cause of death worldwide. According to a recent report by the World Health Organization (WHO), TB killed 1.5 million and caused illness in 9.6 million people in 2014. Bacille Calmette-Gu\u00e9rin (BCG) is the only licensed vaccine against TB. Although it is effective at reducing disseminated forms of TB (e.g., miliary TB and tubercular meningitis) in children[@b1][@b2], BCG has highly variable efficacy (0--80%) against adult pulmonary TB[@b3][@b4], the most contagious form of the disease. Another concern with BCG is its safety in immunocompromised individuals. Disseminated BCG disease has been observed in HIV-infected children following BCG vaccination[@b5], and the risk outweighs the benefit of TB prevention[@b5][@b6]. In 2007, the WHO revised its recommendation and declared that HIV infection is a contraindication for giving BCG[@b7]. In light of this, there is an urgent need to develop a more effective and safe TB vaccine.\n\nOne hypothesis to account for the highly variable protective efficacy of BCG observed in clinical trials concerns the heterogeneity of BCG strains[@b8]. Although colloquially referred to as BCG, there are a number of BCG substrains that have been used in different vaccination programs[@b9][@b10]. Genetic differences including deletions and duplications of genomic regions and single nucleotide polymorphisms (SNPs) among these BCG strains have been well documented, based on a number of studies including whole genome sequencing[@b9][@b11][@b12][@b13][@b14][@b15]. As such, it was suggested that the strain variation may contribute to the variable efficacy of BCG and that some BCG strains might have been over-attenuated during the *in vitro* passaging and consequently lost effectiveness[@b16]. However, this hypothesis has not been formally tested due to the paucity of clinical studies directly comparing different BCG strains. In addition, although genetic and biochemical differences among BCG strains are well established[@b17], whether and how these differences affect BCG effectiveness against TB are largely unknown and remain a matter of debate[@b8][@b18].\n\nPreviously, we found that BCG-Japan, -Moreau, and -Glaxo are naturally deficient in the production of phthiocerol dimycocerosates (PDIMs) and phenolic glycolipids (PGLs), whereas the other nine BCG strains tested, including BCG-Pasteur, produced abundant levels of PDIMs and PGLs[@b19]. PDIMs and PGLs are structurally related complex lipids in the mycobacterial cell wall and are critical for mycobacterial virulence[@b20]. PDIMs are present only in pathogenic mycobacteria such as *Mycobacterium tuberculosis* (*M. tb*)*, M. bovis,* and *M. marinum*. PGLs are also restricted to pathogenic mycobacteria except that in *M. tb*, only a subset of clinical isolates produces PGLs[@b20]. PDIMs were first implicated in virulence using signature-tagged transposon mutagenesis which identified *M. tb* PDIM mutants that were attenuated in mice[@b21][@b22]. Since then, PDIMs have been shown to mediate receptor-dependent phagocytosis of *M. tb*[@b23], and contribute to cell wall permeability[@b24] and protection against bactericidal effects of reactive nitrogen intermediates in activated macrophages[@b25]. PGLs have been implicated in dampening the immune response by inhibiting the release of pro-inflammatory cytokines and have been associated with a hypervirulent phenotype of certain *M. tb* clinical isolates[@b26][@b27]. The critical role PDIMs/PGLs in virulence has also been demonstrated in *M. bovis*[@b28] and *M. marinum*[@b29]. Recently, a study in *M. marinum* suggested that PDIMs and PGLs work in a concerted fashion to recruit permissive macrophages and restrict macrophages with high bactericidal activities, which favors mycobacterial survival and replication in the host[@b30].\n\nGiven that PDIMs/PGLs play important roles in host-pathogen interactions, it is of great interest to determine whether the loss of PDIMs/PGLs, which occurs naturally in a subset of BCG-strains, affects BCG vaccine properties in terms of safety and protective efficacy. In this study, we constructed a PDIM/PGL-deficient strain of BCG-Pasteur by targeted deletion of *fadD28*, a biosynthetic gene of PDIMs/PGLs, and performed comparative analyses of virulence and protective efficacy of the isogenic strains. We found that the PDIM/PGL-deficient strain was less virulent than the wild type (WT) strain of BCG-Pasteur in SCID mice, but was also less protective against *M. tb* infection in both BALB/c mice and guinea pigs.\n\nResults\n=======\n\nConstruction of an isogenic PDIM/PGL deficient mutant of BCG-Pasteur\n--------------------------------------------------------------------\n\nA BCG-Pasteur strain deficient in PDIMs/PGLs was generated by target deletion of *fadD28* ([Fig. 1A,B](#f1){ref-type=\"fig\"}), which encodes a fatty acyl-AMP ligase involved in PDIM/PGL biosynthesis[@b20]. Deletion of *fadD28* was confirmed by Southern blot using a 500\u2009bp probe against the upstream region of *fadD28* ([Fig. 1A,B](#f1){ref-type=\"fig\"}). The \u0394*fadD28* strain grew equally well as the WT strain in 7H9 medium ([Supplementary Fig. S1](#S1){ref-type=\"supplementary-material\"}). Analysis of cell wall lipids by two-dimensional thin layer chromatography (2D-TLC) showed that \u0394*fadD28* was defective in the synthesis of PDIMs/PGLs ([Fig. 1C](#f1){ref-type=\"fig\"}). Transformation of plasmid pFADD28, which contains intact *fadD28*, into the knockout strain restored the production of PDIMs and PGLs in the cell wall ([Fig. 1C](#f1){ref-type=\"fig\"}).\n\nLoss of PDIMs/PGLs reduces virulence of BCG-Pasteur\n---------------------------------------------------\n\nGiven the critical role of PDIMs/PGLs in pathogenic mycobacteria, the loss of PDIMs/PGLs in a BCG strain will likely reduce its virulence. On the other hand, since BCGs are already attenuated strains, the extent to which the additional loss of PDIMs/PGLs contributes to the attenuation of BCG remains unknown. To address this question, we compared the virulence of WT, \u2206*fadD28*, and the complemented strains in severely immunocompromised SCID mice, a mouse model that has been commonly used to assess the safety of BCG strains including the recombinant BCG and attenuated *M. tb* vaccine candidates[@b31][@b32][@b33]. The safety of a live vaccine is inferred from its virulence in SCID mice, which is reflected in the ability of the vaccine to replicate in the animal and to cause mortality. Groups of SCID mice were infected intravenously via the tail vein with \\~10^4^ colony forming units (CFU) of each strain. At 43 days post-infection (dpi), WT BCG-Pasteur reached an average of 7.05 log~10~ CFU in the lungs of SCID mice, whereas the \u2206*fadD28* strain exhibited reduced growth during the same period, with an average of 4.85 log~10~ CFU in the lungs, which is 2.2 log~10~ CFU lower than WT (*p*\u2009\\<\u20090.001, two-way ANOVA, [Fig. 2A](#f2){ref-type=\"fig\"}). The complemented strain had on average 6.25 log~10~ CFUs in the lungs of SCID mice at the same time point, which was not significantly different to that of WT ([Fig. 2A](#f2){ref-type=\"fig\"}). A similar trend was observed for BCG counts in the spleen of SCID mice, although there was no difference between the \u2206*fadD28* and the complemented strains ([Supplementary Fig. S2](#S1){ref-type=\"supplementary-material\"}).\n\nIn a separate experiment, SCID mice (4 mice per group) were infected with a higher dose (10^5^ CFU) of BCG strains (WT and \u2206*fadD28*) and monitored for morbidity over time. WT-infected mice began to lose body weight at 22\u2009dpi ([Fig. 2B](#f2){ref-type=\"fig\"}). By 52\u2009dpi, these mice exhibited severe dehydration and weight loss (\u226520%) and were euthanized to comply with our animal protocols. In contrast, none of the SCID mice infected with \u2206*fadD28* exhibited significant weight loss or other disease phenotypes at 52\u2009dpi and remained healthy until the experiment was terminated at 80\u2009dpi ([Fig. 2B](#f2){ref-type=\"fig\"}). SCID mice infected with WT BCG-Pasteur also exhibited gross pathological evidence of disease, with numerous surface nodules observed in the lungs, whereas few were seen in the \u2206*fadD28*-infected mice ([Fig. 2C](#f2){ref-type=\"fig\"}). Consistently, histological analysis of lung tissues from WT-infected mice had numerous acid-fast positive granulomatous lesions, whereas lungs from \u2206*fadD28*-infected mice showed few scattered lesions ([Fig. 2D](#f2){ref-type=\"fig\"}). Taken together, our results demonstrate that loss of PDIMs/PGLs reduces the virulence of BCG-Pasteur.\n\nLoss of PDIMs/PGLs does not affect immunogenicity of BCG-Pasteur\n----------------------------------------------------------------\n\nTo assess if the loss of PDIMs/PGLs affects the immunogenicity of BCG-Pasteur, we examined production of IFN-\u03b3 in vaccinated C57BL/6 mice. Currently, there is no proven immunological correlate of protection or \"biomarker\" for efficacy[@b34][@b35][@b36], however BCG has been shown to induce a T helper cell 1 (Th1) type response that is characterized by the production of IFN-\u03b3 from CD4^+^ T cells[@b37]. A critical role of IFN-\u03b3 in the control of TB has been demonstrated in mice[@b38][@b39] and humans[@b40][@b41]. As such, antigen specific IFN-\u03b3 produced by CD4^+^ T cells has been used most widely as a measure of protective immunity, even though IFN-\u03b3 alone is insufficient for protection against TB[@b42]. Thus, to examine the role of PDIMs/PGLs in immunogenicity of BCG, we used a C57BL/6 immunocompetent mouse model and measured antigen (PPD) specific IFN-\u03b3 production from both CD4^+^ and CD8^+^ T cells by intracellular cytokine staining. Interestingly, we found that the loss of PDIMs/PGLs in BCG-Pasteur did not significantly alter the amount of IFN-\u03b3 production from both CD4^+^ and CD8^+^ T cells, where comparable levels were observed between the WT- and \u2206*fadD28*-vaccinated groups ([Fig. 3](#f3){ref-type=\"fig\"}, [Supplementary Fig. S3](#S1){ref-type=\"supplementary-material\"}). Detection of IFN-\u03b3 production by ELISA also yielded similar results ([Supplementary Fig. S4](#S1){ref-type=\"supplementary-material\"}). Levels of additional Th1 markers, IL-2 and TNF, were also similar between the WT- and \u2206*fadD28*-vaccinated mice after PPD stimulation ([Supplementary Fig. S5](#S1){ref-type=\"supplementary-material\"}). Taken together, our results indicate that loss of PDIMs/PGLs does not affect BCG immunogenicity.\n\nLoss of PDIMs/PGLs reduces BCG-mediated protection against *M. tb*\n------------------------------------------------------------------\n\nTo determine if the PDIM/PGL-deficient mutant of BCG-Pasteur retained the same capacity to protect against *M. tb*, we used an aerosol challenge model in BALB/c mice. Immunocompetent inbred mice (BALB/c and C57BL/6) are widely used for TB vaccine studies because of the low cost and the availability of immunological reagents[@b43]. Groups of mice were vaccinated subcutaneously with \\~10^5^ CFU of BCG strains (BCG-Pasteur, \u2206*fadD28*, \u2206*fadD28*\u2009+\u2009pFADD28, BCG-Japan) or PBS as a control. BCG-Japan was included in this experiment for comparison because it is naturally deficient in PDIMs/PGLs[@b19]. At 8 weeks post-vaccination, the mice were aerogenically challenged with 400--600 CFU of *M. tb* H37Rv and bacterial burden in the lung and spleen was determined at 5 and 9 weeks post-challenge.\n\nAt week 5 post-challenge, the non-vaccinated group of BALB/c mice had a mean *M. tb* burden of 6.23 log~10~ CFU in the lungs ([Fig. 4A](#f4){ref-type=\"fig\"}). Mice vaccinated with WT BCG-Pasteur, \u2206*fadD28*, the complemented strain, and BCG-Japan had on average 5.49 log~10~, 5.93 log~10~, 5.51 log~10~, and 5.98 log~10~ CFU of *M. tb* in the lungs, respectively. Compared to the PBS group, mice vaccinated with BCG strains had significantly lower *M. tb* burdens, with a reduction of 0.3 log~10~ (\u2206*fadD28*, BCG-Japan) and 0.7 log~10~ CFU (BCG-Pasteur, the complemented strain). Interestingly, mice vaccinated with \u2206*fadD28* or BCG-Japan had significantly higher *M. tb* burdens than those vaccinated with WT BCG-Pasteur or the complemented strain, by \\~0.45 log~10~ CFU (*p*\u2009\\<\u20090.05, one-way ANOVA, Tukey's post hoc test).\n\nAt week 9 post-challenge, mice vaccinated with BCG-Pasteur, the complemented strain or BCG-Japan had significantly lower *M. tb* burdens in the lungs compared to the PBS control by 0.41 log~10~ (*p*\u2009\\<\u20090.05), 0.52 log~10~ (*p*\u2009\\<\u20090.01), and 0.34 log~10~ (*p*\u2009\\<\u20090.05), respectively ([Fig. 4A](#f4){ref-type=\"fig\"}, one-way ANOVA, Tukey's post hoc test). The difference between the \u2206*fadD28* and the PBS groups was not statistically significant nor was the difference between the WT and the \u2206*fadD28* groups.\n\nThe aerosol infection model best mimics natural infection, and the levels of *M. tb* disseminated to the spleen are anticipated to be much lower than in the lungs[@b43]. Consistently, our data showed that the amounts of *M. tb* disseminated to the spleen were lower than that in the lungs by \\~1.0--2.0 log~10~ CFU ([Fig. 4B](#f4){ref-type=\"fig\"}). At week 5 post-infection, mice vaccinated with BCG-Pasteur and the complemented strain had 0.35 and 0.38 log~10~ CFU fewer *M. tb* in the spleen than the PBS control group, respectively ([Fig. 4B](#f4){ref-type=\"fig\"}). There was no difference in the *M. tb* burden between mice vaccinated with \u2206*fadD28* and the PBS control (3.91 log~10~ vs. 3.98 log~10~ CFU). At week 9, compared to the PBS group, mice vaccinated with BCG-Pasteur (\u0394CFU\u2009=\u20090.77 log~10~, *p*\u2009\\<\u20090.01), \u2206*fadD28* (\u0394CFU\u2009=\u20090.55 log~10~, *p*\u2009\\<\u20090.01), the complemented strain (\u0394CFU\u2009=\u20090.73 log~10~, *p*\u2009\\<\u20090.01), or BCG-Japan (\u0394CFU\u2009=\u20090.37 log~10~, *p*\u2009\\<\u20090.05) had significantly lower *M. tb* burdens in the spleen. Mice vaccinated with BCG-Pasteur or the complemented strain had significantly lower *M. tb* burdens than those vaccinated with BCG-Japan (*p*\u2009\\<\u20090.01, one-way ANOVA, Tukey's post hoc test). Mice vaccinated with \u2206*fadD28* appeared to have higher *M. tb* burdens than those vaccinated with BCG-Pasteur or the complemented strain, but the differences were not statistically significant ([Fig. 4B](#f4){ref-type=\"fig\"}).\n\nHistological analysis of *M. tb* challenged mice showed consistent differences in lung pathology between the different BCG-vaccinated cohorts. Vaccination with WT BCG-Pasteur or the complemented strain appeared to reduce the number of granuloma-like lesions in the lung, whereas mice vaccinated with the BCG-Pasteur \u2206*fadD28* strain or BCG-Japan had lung pathology similar to the PBS control group ([Fig. 4C](#f4){ref-type=\"fig\"}).\n\nC57BL/6 and BALB/c mice are highly resistant to *M. tb* infection and do not form caseous granulomas in the lungs that are typical of human TB disease[@b43][@b44]. This represents a major limitation of the murine model for vaccine studies. Alternatively, the guinea pig model is considered a more stringent test of vaccine efficacy[@b45]. Guinea pigs are highly susceptible to *M. tb* infection and develop clinically relevant symptoms, including weight loss and decreased pulmonary function due to extensive pulmonary infiltration. A recent study comparing vaccine testing from three different laboratories highlighted the reliability and reproducibility of the guinea pig model in obtaining efficacy data[@b46].\n\nTo examine if the reduced protection of the PDIM/PGL-deficient mutant of BCG-Pasteur against *M. tb* can also be observed in the guinea pig model, we conducted a *M. tb* challenge experiment in Hartley guinea pigs. Groups of six guinea pigs were vaccinated subcutaneously with WT BCG-Pasteur, \u2206*fadD28*, or PBS and were aerogenically challenged 10 weeks later with *M. tb*. Prior to infection, guinea pigs of all groups exhibited similar weight gain. After challenge, one guinea pig in the PBS control group reached the humane end-points (loss of 20% maximal body weight and/or labored/heavy breathing) and was euthanized at week 10. Beginning at week 5 post-challenge, guinea pigs vaccinated with WT BCG-Pasteur gained significantly more weight than the PBS group ([Fig. 5A](#f5){ref-type=\"fig\"}). Guinea pigs vaccinated with \u2206*fadD28* also showed significant weight gain compared to the PBS group starting at week 9 ([Fig. 5A](#f5){ref-type=\"fig\"}). The experiment was terminated at week 12 post-challenge and the lungs and spleen of these animals were isolated for further analyses. Mean guinea pig lung weights in the PBS and \u0394*fadD28* groups were 4.81\u2009g and 3.71\u2009g, respectively. In contrast, the mean lung weight of WT BCG-Pasteur vaccinated guinea pigs was 2.62\u2009g, which was significantly lower than those vaccinated with \u0394*fadD28* or PBS (*p*\u2009\\<\u20090.05 and *p*\u2009\\<\u20090.01, Mann-Whitney test, [Fig. 5B](#f5){ref-type=\"fig\"}). Guinea pig spleen weights followed a similar trend, where the mean spleen weights in the PBS, \u0394*fadD28*, and WT groups were 3.08, 1.22, and 0.99\u2009g, respectively, although the difference between \u0394*fadD28* and WT groups were not statistically significant ([Fig. 5B](#f5){ref-type=\"fig\"}).\n\nMean CFU counts of *M. tb* in the lungs of the PBS, \u0394*fadD28*, and WT groups were 6.57 log~10~, 5.33 log~10~ and 4.63 log~10~, respectively ([Fig. 5C](#f5){ref-type=\"fig\"}). Notably, guinea pigs vaccinated with \u0394*fadD28* had 0.7 log~10~ CFU more *M. tb* than those vaccinated with WT BCG-Pasteur and the difference was approaching significance (*p*\u2009=\u20090.064, Mann-Whitney test, [Fig. 5C](#f5){ref-type=\"fig\"}). The *M. tb* burden in the spleen exhibited a similar trend, where guinea pigs vaccinated with \u0394*fadD28* had \\~1.5 log~10~ more *M. tb* counts than those vaccinated with WT BCG-Pasteur (*p*\u2009=\u20090.063, Mann-Whitney test, [Fig. 5C](#f5){ref-type=\"fig\"}).\n\nHistological examination of the lungs in the unvaccinated guinea pig group showed numerous granulomatous lesions with occasionally visible central necrosis ([Fig. 5D,E](#f5){ref-type=\"fig\"}), and large areas of the lungs were affected ([Fig. 5D](#f5){ref-type=\"fig\"}, mean\u2009=\u200918.02%). Lungs obtained from guinea pigs vaccinated with WT BCG-Pasteur or \u2206*fadD28* had smaller, primarily non-necrotic granuloma-like lesions ([Fig. 5D,E](#f5){ref-type=\"fig\"}). Importantly, the affected area of the lungs from WT-vaccinated guinea pigs (mean\u2009=\u20094.68%) was smaller than that from \u2206*fadD28*-vaccinated guinea pigs (mean\u2009=\u20099.29%) ([Fig. 5D](#f5){ref-type=\"fig\"}, *p*\u2009=\u20090.064, Mann-Whitney test). This difference is consistent with our observations that the bacterial burden in the lungs obtained from WT-vaccinated guinea pigs was reduced compared to that from \u2206*fadD28*-vaccinated guinea pigs ([Fig. 5C](#f5){ref-type=\"fig\"}). Taken together, our results from both mouse and guinea pig models suggest that the loss of PDIMs/PGLs compromises the ability of the BCG vaccine to protect against *M. tb* infection.\n\nDiscussion\n==========\n\nBCG was derived from a virulent strain of *M. bovis* through *in vitro* attenuation (230 passages) from 1908 to1921. Beginning in 1924, BCG was distributed to various countries worldwide, resulting in a number of genetically distinct substrains. The mechanisms of BCG attenuation remain incompletely understood[@b17]. The loss of RD1, which encodes the type VII secretion system ESX-1, contributes to the attenuation of BCG[@b14][@b15][@b47]. However, recombinant BCG strains complemented with RD1 only partially restored the virulence, suggesting additional mechanisms are involved[@b31][@b48]. Comparative genome analyses revealed a number of genetic polymorphisms including deletions, duplications, and SNPs in BCG strains[@b9][@b11][@b12][@b13][@b14]. Some of these genetic changes are shared by subgroups of BCG strains whereas others are specific to individual strains. Despite the vast number of publications on BCG, studies to evaluate the impact of genetic polymorphisms on BCG vaccine properties (safety and efficacy) have been scarce. BCG strains distributed after 1927 (i.e. late BCG strains) contains an additional deletion of RD2, which encompasses genes *Rv1978* to *Rv1988* and includes important antigens such as MPT62 (encoded by *Rv1980c*)[@b15]. Deletion of RD2 may have attenuated the virulence of late BCG strains since a RD2-deletion mutant of *M. tb* H37Rv was more attenuated than the parental strain[@b49]. However, recombinant BCG-Pasteur complemented with *Rv1979c-Rv1982* did not improve protection against pulmonary TB, although it reduced the dissemination of *M. tb* to the spleen[@b50]. Late BCG strains also contain a point mutation in *mmaA3*, which impairs the production of methoxymycolate. However, complementation of a late BCG strain (BCG-Danish) with wild type *mmaA3*, which restored the production of methoxymycolate, had no effect on the virulence of BCG and its effect on protection was not determined[@b51]. The lack of experimental and clinical evidence demonstrating the impact of genetic differences among BCG has led to the argument that strain variation is not a significant factor for BCG effectiveness[@b18]. In this study, we demonstrated that the loss of PDIMs/PGLs, which occurs naturally in a subset of BCG strains, had a significant effect on the safety and protective efficacy of BCG, providing evidence that differences in BCG strains can influence vaccine effectiveness. A randomized trial study comparing two BCG strain in 300,000 infants in Hong Kong found that BCG-Pasteur, administered at a lower dosage, provided a significantly greater (40%) protection against childhood forms of TB than BCG-Glaxo[@b52]. Based on the finding of current study, the loss of PDIMs/PGLs in BCG-Glaxo[@b19] is likely an important factor affecting its efficacy.\n\nThe loss of PDIMs/PGLs in BCG-Japan, -Moreau and -Glaxo correlates with their superior safety records in clinical studies over other BCG strains[@b17][@b19][@b53]. However, the existence of other mutations that distinguish BCG strains precludes a simple comparison of PDIM/PGL producers and non-producers to determine the importance, if any, of PDIMs/PGLs for vaccine safety and protection. To address this question, we constructed a PDIM/PGL-deficient strain from BCG-Pasteur, a PDIM/PGL producer, and performed a comparative study of the isogenic strains.\n\nConsistent with the well-established role of PDIMs/PGLs in mycobacterial virulence, the loss of PDIMs/PGLs reduced the virulence of BCG-Pasteur, as demonstrated in the SCID mouse model. The ability of BCG-Pasteur to replicate in SCID mice and to cause morbidity was compromised when the production of PDIMs/PGLs was abrogated ([Fig. 2](#f2){ref-type=\"fig\"}). However, the loss of PDIMs/PGLs did not affect the immunogenicity of BCG-Pasteur, specifically its ability to induce antigen-specific IFN-\u03b3 production by CD4^+^ and CD8^+^ T cells ([Fig. 3](#f3){ref-type=\"fig\"}). Unexpectedly, the loss of PDIMs/PGLs also decreased the efficacy of BCG against *M. tb* challenge. This was demonstrated in both mouse and guinea pig models. As classically demonstrated[@b54][@b55][@b56], the *M. tb* infection of BALB/c or C57BL/6 mice by aerosol challenge is followed by two phases. The progressive phase, in which *M. tb* grows essentially uninhibitedly for the first 3--4 weeks, results in 6--7 log~10~ CFU in the lungs. This is followed by the stationary phase in which further *M. tb* growth is inhibited by adaptive immunity. We found that at week 5 post-challenge, the *M. tb* burden in the lungs of BALB/c mice vaccinated with WT BCG-Pasteur was significantly lower than those vaccinated with the \u2206*fadD28* strain ([Fig. 4](#f4){ref-type=\"fig\"}). However, this difference diminished at the stationary phase of infection (week 9 post-challenge) presumably because mice have begun to control the *M. tb* infection at this time. Considering the drawbacks of the mouse model (e.g., highly resistant to *M. tb* infection), we moved to the guinea pig model, which allows the evaluation of a broader spectrum of disease phenotypes. Although the change in body weight was not sensitive enough to distinguish between guinea pigs vaccinated with different BCG strains, as demonstrated previously[@b57], the lung weights of guinea pigs vaccinated with \u2206*fadD28* was on average 41.6% higher than those vaccinated with WT BCG-Pasteur ([Fig. 5](#f5){ref-type=\"fig\"}). The difference in spleen weight between these two groups was less significant presumably because the *M. tb* burden in the spleen was much lower than in the lungs as a result of the aerosol challenge route. Increased organ weights (lungs and spleen) have been associated with more severe disease phenotypes and frequently observed in guinea pigs infected with virulent *M. tb*[@b58][@b59]. Consistently, the *M. tb* burden in \u2206*fadD28*-vaccinated guinea pigs was higher than those vaccinated with WT BCG-Pasteur, by 0.7 log~10~ and 1.5 log~10~ CFU in the lungs and spleen, respectively ([Fig. 5](#f5){ref-type=\"fig\"}). Collectively, these data provide strong evidence that the loss of PDIMs/PGLs reduces the protective efficacy of BCG.\n\nOur finding has practical implications for the clinical preparations of BCG vaccines. The natural loss of PDIMs/PGLs in BCG-Japan, -Moreau, and -Glaxo likely occurred randomly during *in vitro* passaging. This is consistent with several observations. Firstly, independent mutations in biosynthetic genes account for the defective biosynthesis of PDIMs/PGLs in these BCG-strains. BCG-Japan contains a frame-shift single nucleotide insertion within *ppsA*[@b60], and BCG-Moreau contains a deletion that disrupts both *fadD26* and *ppsA*[@b12]. The genetic mutation responsible for the loss of PDIMs/PGLs in BCG-Glaxo has yet to be identified. Secondly, spontaneous loss of PDIMs has been frequently observed in *M. tb* H37Rv (which produces PDIMs but not PGLs) during *in vitro* experiments[@b61][@b62][@b63]. Because this event occurs at such a high frequency, it is necessary to confirm the presence of PDIMs in all parental strains and recombinant clones before undertaking *in vivo* virulence studies[@b49][@b63]. Spontaneous loss of PDIMs in BCG-Pasteur during *in vitro* passage has also been reported[@b64]. Moreover, the clinical preparations of BCG-Japan actually contain two subpopulations, one producing PDIMs/PGLs and the other defective in PDIMs/PGLs[@b60][@b65]. Considering the high selective pressure under *in vitro* conditions for PDIM/PGL-negative clones, and our finding that the loss of PDIMs/PGLs has a significant impact on BCG vaccine safety and efficacy, it is essential that quality control programs in BCG manufacturers should include a regular test of PDIMs/PGLs of vaccine preparations.\n\nThe PDIM/PGL-deficient strain is less virulent but also less protective, suggesting a positive correlation between virulence and efficacy. Consistent with this notion, a previous study found that recombinant BCG strains complemented with the RD1 region exhibited increased virulence in SCID mice but also better protection in C57BL/6 mice and guinea pigs[@b31][@b48]. More recently, a comparative analysis of the virulence and efficacy of 13 different BCG strains in SCID and BALB/c mice, respectively, also revealed a general trend that more virulent BCG strains were also more effective in protection against *M. tb* challenge[@b66]. Currently, the strategies for developing the next generation of TB vaccines include live vaccines (recombinant BCG or attenuated *M. tb*) and subunit vaccines[@b67][@b68]. The lack of protective efficacy of MVA85A, the most advanced subunit vaccine candidate thus far, in a recent clinical trial study[@b69] further underlines the importance of live vaccine research[@b70]. The positive correlation between virulence and efficacy we observed suggests that when developing recombinant BCG or attenuated *M. tb*, there needs to be a fine balance between these two factors in order to achieve optimal protection while maintaining an acceptable level of safety.\n\nMaterials and Methods\n=====================\n\nBacterial strains and culture conditions\n----------------------------------------\n\n*Mycobacterium bovis* BCG strains, BCG-Pasteur and BCG-Japan, were grown at 37\u2009\u00b0C in Middlebrook 7H9 broth (Difco^\u2122^) supplemented with 0.2% glycerol, 10% albumin-dextrose-catalase (ADC; BD BBL^\u2122^), and 0.05% Tween80 or on Middlebrook 7H11 agar (Difco^\u2122^) supplemented with 0.5% glycerol and 10% oleic acid-albumin-dextrose-catalase (OADC; BD BBL^\u2122^). *Escherichia coli* strain DH5\u03b1 was used for routine manipulation and propagation of plasmid DNA. *E. coli* DH5\u03b1 was grown in LB broth or agar (BioShop). Antibiotics were added as required: kanamycin, 50 \u03bcg/ml for *E. coli* and 25\u2009\u03bcg/ml for BCG; hygromycin, 150\u2009\u03bcg/ml for *E. coli* and 75\u2009\u03bcg/ml for BCG.\n\nGeneration of a PDIM/PGL-deficient mutant of BCG-Pasteur\n--------------------------------------------------------\n\nSpecialized phage transduction was used to generate a PDIM/PGL-deficient mutant of BCG-Pasteur as described previously[@b71]. Briefly, the allelic exchange construct was made by amplifying upstream and downstream regions flanking the *fadD28* gene from BCG-Pasteur genomic DNA using the primer sets 5\u2032-[ACTAGT]{.ul}GATTTCGACACTCGGTAA-3\u2032 (SpeI)/5\u2032-[AAGCTT]{.ul}GTCTTCTTTGAAGGT-3\u2032 (HindIII) and 5\u2032-[TCTAGA]{.ul}GATTTTCACGCCTTT-3\u2032(Xba1)/ 5\u2032-[GGTACC]{.ul}AGTTCGATA ATG G-3\u2032 (KpnI), respectively (restriction sites are underlined). The upstream amplicon was digested and ligated into a SpeI/HindIII-digested pJSC284 cosmid, containing a hygromycin resistance marker (hyg^R^). The resulting vector was then digested with XbaI and KpnI and ligated to the downstream amplicon, creating the complete allelic exchange construct. Correct insertion of both amplicons was confirmed by PCR using locus-specific primers. The recombinant construct was cloned into a conditionally replicating TM4 shuttle phasmid, phLR, and specialized transducing mycobacteriophage were generated by electroporating *M. smegmatis* mc^2^155 at the permissive temperature (30\u2009\u00b0C). Putative knockout mutants were obtained by transducing BCG-Pasteur at the non-permissive temperature (37\u2009\u00b0C) and selecting hygromycin-resistant colonies. Deletion of *fadD28* was confirmed by Southern blot (Amersham) analysis using a 500\u2009bp probe against the upstream region of *fadD28*, generated with primers 5\u2032-TCCAACCTCGTCTCAGCT-3\u2032 and 5\u2032-CGCCAT GGGTCCACCA-3\u2032, following the manufacturer's protocol. The complementation plasmid was generated by amplifying a 2094\u2009bp fragment containing a wild type (WT) copy of *fadD28*, using the forward primer 5\u2032-[GGTACC]{.ul}AAGCCAGTTAGGGGC-3\u2032 (KpnI) and reverse primer 5\u2032-[AAGCTT]{.ul}CAGTCCG GGGAGGAC-3\u2032 (HindIII), and cloned into a KpnI/HindIII-digested pME shuttle vector to generate pFADD28. Three to five clones of each strain were tested.\n\nLipid analysis by thin layer chromatography\n-------------------------------------------\n\nProduction of PDIMs/PGLs was examined using two-dimensional thin layer chromatography (2D-TLC), according to published procedures[@b19][@b72]. Briefly, the apolar lipid fraction was extracted from 50\u2009mg (dry weight) of BCG and analyzed on silica gel 60 plates (EMD Chemicals Inc.). For detection of PDIMs, apolar lipids were developed with petroleum ether/ethyl acetate (98:2, 3\u00d7) in the first dimension and petroleum ether/acetone (98:2) in the second dimension. Lipids were visualized by staining plates with 5% phosphomolybdic acid followed by gentle charring. For detection of PGLs, the apolar lipid extract was developed with chloroform/methanol (96:4, v/v) in the first dimension and toluene/acetone (80:20, v/v) in the second dimension, followed by charring with \u03b1-naphthol. The productions of PDIMs/PGLs were periodically checked to ensure strain integrity.\n\nEthics statement\n----------------\n\nAll of the animal procedures were approved by the University of Toronto Animal Care Committee. All experimental procedures were performed in accordance with the Canadian Council on Animal Care (CCAC) and University of Toronto regulations.\n\nAnalysis of BCG virulence in SCID mice\n--------------------------------------\n\nFemale Fox Chase CB17^\u00ae^ SCID mice were purchased from Charles River Laboratories and the mice were age-matched (7--8 weeks) within each experiment. Mice (4--6 per group per time point) were infected intravenously via the tail vein with \\~10^4^ or \\~10^5^ CFU of the different BCG strains in 0.2\u2009ml PBS/0.01% Tween80. At 1, 7, 21, 42, 52, and 79 days post-infection, the lungs were harvested, homogenized in PBS, and plated on 7H11 agar to enumerate bacterial burden. CFU was counted after incubation at 37\u2009\u00b0C for 3 weeks. Bacterial counts from lung homogenates harvested at day 1 post-infection were used as an indicator of initial infection dose. These experiments were done in duplicate.\n\nImmunogenicity studies\n----------------------\n\nFemale C57BL/6 mice were purchased from Charles River Laboratories and were age-matched (6 weeks) within each experiment. Four to nine mice per group were inoculated subcutaneously on the scruff of the neck with approximately \\~10^4^ CFU in 0.2\u2009ml PBS/0.01%Tween80 of parental BCG-Pasteur or the *fadD28* knockout strain. Control mice were given 0.2\u2009ml of PBS/0.01% Tween80. After 9 weeks, mice were euthanized, splenocytes were isolated, and intracellular IFN\u03b3 was measured. Briefly, splenocytes were seeded at 2\u00d710^6^ cells/well in 100\u2009\u03bcl in triplicate and stimulated with 2.5\u2009\u03bcg/well of purified protein derivative (PPD) (Statens Serum Institute, Denmark) or complete RPMI (cRPMI; RPMI/10% FBS/1% L-glutamine/1% penicillin/streptomycin) as a control and incubated at 37\u2009\u00b0C and 5% CO~2~. After 19\u2009hours of stimulation, GolgiPlug (BD Biosciences) was added in a 1:1000 dilution and incubated for an additional 5\u2009hours. After a total of 24\u2009hours stimulation, plates were centrifuged and the cell pellet was washed in 200\u2009\u03bcl FACS Buffer (0.5% BSA/PBS) and incubated with Fc Block (eBiosciences) diluted in FACS Buffer (1:400) for 15\u2009minutes. The cells were then washed with FACS Buffer and stained for extracellular T-cell surface markers: CD3-PE, CD4-FITC, and CD8a-PercyPCy5.5 (BD Biosciences) diluted in FACS Buffer for 30\u2009minutes. Following extracellular marker staining, the cells were permeabilized and fixed with 1\u00d7CytoFix/CytoPerm (BD Biosciences) for 20\u2009minutes. Cells were then washed with 1\u00d7PermWash (BD Biosciences) and incubated with IFN\u03b3-APC (BD Biosciences) for 30\u2009minutes to stain for intracellular IFN\u03b3. Immediately following staining, cells were analyzed on a BD FACSCalibur^TM^ flow cytometer (BD Biosciences). A total of 300,000 events per sample were collected in the lymphocyte gate and analyzed using FlowJo V7.6. Gates for analysis were set based on isotype controls. These experiments were done in duplicate.\n\nProtection against *M. tb* challenge\n------------------------------------\n\nMouse model: Groups of 13--15 female BALB/c mice (Charles River Laboratories) were vaccinated subcutaneously on the scruff of the neck with \\~10^5^ CFU of the BCG strains in 0.2\u2009ml PBS/0.01% Tween80 or PBS/0.01% Tween80 alone as a control. At 8 weeks post-vaccination, mice were aerogenically challenged with 400--600 CFU of *M. tb* H37Rv using a GlasCol nebulizer. Mice were euthanized at 5 and 9 weeks post-challenge (6--7 mice per group per time point) to harvest the lungs and spleen. A portion of the organs were fixed in 10% formalin for histological analysis. The remaining portion was homogenized and plated on 7H11 agar to enumerate burden of *M. tb* in the lung and spleen. Plates were incubated at 37\u2009\u00b0C and counted after 2.5--3 weeks. These experiments were done in duplicate.\n\nGuinea pig model: Groups of six female Hartley guinea pigs (Charles River Laboratories) were vaccinated subcutaneously with 5\u2009\u00d7\u200910^4^ CFU of parental BCG-Pasteur or the *fadD28* knockout strain in 0.2\u2009ml PBS/0.01% Tween80 or PBS/0.01% Tween80 alone as a control. At 10 weeks post-vaccination, guinea pigs were infected with \\~1000 CFU of *M. tb* H37Rv by an aerosol challenge using a GlasCol nebulizer. At 12 weeks post-challenge, guinea pigs were euthanized to obtain the lungs and spleen. A portion of the spleen and the caudal lobe of the left lung were fixed in 10% formalin for histological analysis. The remaining portion of the spleen and the entire right lung lobes were homogenized separately and plated on 7H11 agar to quantify the *M. tb* burden in the lungs and spleen. Colonies were counted after incubation at 37\u2009\u00b0C for three weeks. This experiment was performed once.\n\nHistological analysis\n---------------------\n\nFixed tissues were embedded into paraffin blocks at the Centre for Modeling Human Disease (Toronto Centre for Phenogenomics). Serial sections (4\u2009\u03bcm thick for mouse tissues and 5\u2009\u03bcm thick for guinea pig tissues) were prepared and kept at 37\u2009\u00b0C for more than 12\u2009hours. The sections were deparafinized in three changes of xylene for 3\u2009minutes each and rehydrated in four consecutive washes of alcohol (100%, 100%, 95%, and 70%) for 3\u2009minutes each. Sections were stained with hematoxylin and eosin (EMD Chemicals) or Acid Fast staining kit (Surgipath) according to standard procedures and were examined using a Leica microscope (Life Technologies) or Cytation^TM^ 5 (BioTek). Perceived areas of granulomatous lesions were determined by ImageJ.\n\nStatistical Analysis\n--------------------\n\nMajority of the dataset passed the Kolmogorov--Smirnov normality test. One-Way Analysis of Variance (One-way ANOVA) with Tukey's multiple comparisons were performed for *M. tb* burdens (log~10~ transformed CFU data) when there are more than 3 groups. Two-Way ANOVA were performed on data (CFU or body weight) of 3 or more groups at multiple time points. Student's *t* test (Mann-Whitney test) was performed when there are 3 or fewer groups.\n\nAdditional Information\n======================\n\n**How to cite this article**: Tran, V. *et al.* Loss of Lipid Virulence Factors Reduces the Efficacy of the BCG Vaccine. *Sci. Rep.* **6**, 29076; doi: 10.1038/srep29076 (2016).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nThis work was supported by Canadian Institutes of Health Research (CIHR) Grant MOP-106559 (to JL). VT was the recipient of the CIHR Frederick Banting and Charles Best Canada Graduate Scholarship. The funding agency had no role in the study design, data collection and interpretation, or the decision to submit the work for publication.\n\n**Author Contributions** V.T., S.K.A., M.N. and M.L. performed experiments. V.T., J.L. and S.K.A. wrote the manuscript. All authors reviewed the manuscript.\n\n![Construction of a PDIM/PGL deficient strain of BCG-Pasteur.\\\n(**A**) Genomic organization of WT (Pasteur) and \u2206*fadD28* strains. Dashed lines indicate products of restriction digestion with *Cla*I and *Box*I. (**B**) Southern blot analysis. Chromosomal DNAs isolated from WT and three randomly picked \u2206*fadD28* clones were digested with *Cla*I and *Box*I and blotted with a 500\u2009bp probe of *fadD28*, which yielded a 3.5\u2009kb and 1.8\u2009kb fragment, respectively, and agreed with prediction (**A**). (**C**) 2D-TLC analysis of PDIMs and PGLs. For PDIM analysis, apolar lipids were developed with petroleum ether/ethyl acetate (98:2 v/v, 3 times) in the first dimension (1^st^) and petroleum ether/acetone (98:2, v/v) in the second dimension (2^nd^). Lipids were visualized by charring with 5% phosphomolybdic acid. For PGL analysis, the apolar lipid extract was developed with chloroform/methanol (96:4, v/v) and toluene-acetone (80:20, v/v), followed by charring with \u03b1-naphthol. PDIMs, phthiocerol dimycocerosates; PGLs, phenolic glycolipids.](srep29076-f1){#f1}\n\n![The PDIM/PGL deficient mutant of BCG-Pasteur is less virulent in SCID mice.\\\n(**A**) BCG burden in the lungs of infected SCID mice. SCID mice were infected intravenously with 10^4^ CFU of BCG-Pasteur (Pasteur), the \u2206*fadD28*, or complemented strain (\u2206*fadD28*\u2009+\u2009pFADD28). Bacterial burden in the lungs were determined at various time points. (\\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001, two-way ANOVA). (**B**) Body weight of SCID mice infected with 10^5^ CFU of WT BCG-Pasteur or \u2206*fadD28* (\\**p*\u2009\\<\u20090.05; student's *t*-test). (**C**) Lung pathology of mice infected with WT BCG-Pasteur and \u2206*fadD28* (10^5^ CFU) at 52 day post infection. (**D**) Lung histology of mice infected with WT BCG-Pasteur and \u2206*fadD28* (10^5^ CFU) at 52 day post-infection; arrow denotes magnified granuloma-like lesion. HE, hematoxylin-eosin stain; AF, acid fast stain.](srep29076-f2){#f2}\n\n![The loss of PDIMs/PGLs does not affect production of IFN-\u03b3.\\\nIntracellular cytokine staining analysis of IFN-\u03b3 production by (**A**) CD4^+^ and (**B**) CD8^+^ T-cells. C57BL/6 mice were immunized subcutaneously with the WT BCG-Pasteur, \u2206*fadD28*, or PBS/0.01% Tween 80. At 9 weeks post-vaccination, mice were sacrificed and splenocytes were harvested. Splenocytes were incubated with or without PPD for 24\u2009hr followed by staining for T-cell surface markers (CD3-PE, CD4-FITC, CD8a-PercyPCy5.5) and intracellular IFN-\u03b3 (IFN-\u03b3-APC). Samples were analyzed by BD FACSCalibur^TM^ and FlowJo^\u00a9^ Software. Pooled results from two independent experiments; each data point represents one mouse.](srep29076-f3){#f3}\n\n![Loss of PDIMs/PGLs reduces BCG-mediated protection against *M. tb* in mice.\\\nBALB/c mice were vaccinated subcutaneously with \\~10^5^ CFU of the BCG strains or PBS as a control. At 8-weeks post vaccination, mice were aerogenically challenged with *M. tb*. Mice were sacrificed at 5 and 9 weeks post-challenge and organs were examined for bacterial burden and pathology. The *M. tb* burden in the (**A**) lungs and (**B**) spleen was shown (6 mice per group per time point; data were plotted as box-whiskers in which the whiskers represent the minimum and maximum of all data. \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001; one-way ANOVA, Tukey's post hoc test). (**C**) Histological analysis of lung sections from mice in each group at 9 weeks post-challenge. Samples are stained with H&E. Arrows indicate regions of granuloma-like lesions. Top row is 50x magnification; bottom row is 200x magnification.](srep29076-f4){#f4}\n\n![Loss of PDIMs/PGLs reduces BCG-mediated protection against *M. tb* in guinea pigs.\\\nGuinea pigs were vaccinated subcutaneously with WT BCG-Pasteur, \u2206*fadD28* or PBS. At 10 weeks post-vaccination, guinea pigs were aerogenically challenged with *M. tb*. (**A**) Relative body weight (normalized to the body weight at time of *M. tb* challenge) of guinea pigs post infection. Data are plotted as mean \u00b1 SEM (*n*=6). \\*\\**p*\u2009\\<\u20090.01; \\*\\*\\**p*\u2009\\<\u20090.001, two-way ANOVA. (**B**) Lung and spleen weights of guinea pigs at 12-weeks post challenge. Data are plotted as box-whiskers in which the whiskers represent the minimum and maximum of all data (*n*=6). \\**p*\u2009\\<\u20090.05; \\*\\**p*\u2009\\<\u20090.01, Mann-Whitney test. (**C**) *M. tb* burden in the lungs and spleen plotted as box-whiskers (*n*=6). \\*\\**p*\u2009\\<\u20090.01, Mann-Whitney test. (**D**) Quantitation of lung area affected by granuloma (% of total lung area). Six slides from each group were analyzed and data are plotted as box-whiskers. \\*\\**p*\u2009\\<\u20090.01, Mann-Whitney test. (**E**) Histological analysis of lung sections from guinea pigs in each group. Samples are stained with HE. Arrows indicate regions of granuloma-like lesions (40x magnification). Scale bar represents 1000\u2009\u03bcm.](srep29076-f5){#f5}\n\n[^1]: Present address: Sandra A. Rotman Laboratories, Sandra Rotman Centre for Global Health, Toronto General Research Institute -- University Health Network, Toronto, ON, Canada.\n"}
+{"text": "Presentation slides available from: <http://www.bridgingknowledge.net/Presentations/Symp8_Wolfmayr.pdf>\n"}
+{"text": "Cancer Sci 108 (2017) 136--142\n\n**Funding Information**\n\nNational Cancer Center Research and Development Fund (23\u2010A\u201016); Japan Society for the Promotion of Science (KAKENHI 26108007).\n\nDespite advances in therapy, the prognosis of patients with advanced NSCLC remains poor with a 5\u2010year survival rate of \\<20%.[1](#cas13110-bib-0001){ref-type=\"ref\"} Distant metastasis is the most important prognostic factor and play a major role in choice of treatment. Squamous cell carcinoma accounts for approximately 30% of lung cancer and is the second most common histological type behind adenocarcinoma.[2](#cas13110-bib-0002){ref-type=\"ref\"} Lung SqCC has a lower rate of distant metastasis than other histological type, but not of LN metastasis. Furthermore, local recurrence including LN is more frequent after the resection of early stage of lung SqCC than adenocarcinoma.[3](#cas13110-bib-0003){ref-type=\"ref\"}, [4](#cas13110-bib-0004){ref-type=\"ref\"} As lymph node metastasis is an important prognostic factor,[5](#cas13110-bib-0005){ref-type=\"ref\"} an improved knowledge of the mechanism of lymphatic metastasis in lung SqCC might be of potential clinical benefit.\n\nMetastasis is a complex multistep process. In the PT, tumor cells detach from their surrounding cells by EMT and intravasate into blood or lymphatic vessels. The tumor cells float in the circulation and some of them survive without anoikis. The cells adhere to the endothelium, extravasate, and begin reproliferation. To facilitate an increase in tumor size, induction of angiogenesis and several microenvironmental factors are required.[6](#cas13110-bib-0006){ref-type=\"ref\"}, [7](#cas13110-bib-0007){ref-type=\"ref\"} Previously, Kirita *et\u00a0al*.[8](#cas13110-bib-0008){ref-type=\"ref\"} reported that intralymphatic adenocarcinoma cells expressing low levels of ALDH1, high levels of SOX2, and a high number of infiltrating CD204\u2010positive macrophages have a critical impact on LN metastasis. Therefore, it is suggested that the ILT microenvironment is an important determinant of LN metastasis.\n\nComparative transcription analysis has previously identified a different pattern of gene expression in LN metastasis compared to PT.[9](#cas13110-bib-0009){ref-type=\"ref\"}, [10](#cas13110-bib-0010){ref-type=\"ref\"}, [11](#cas13110-bib-0011){ref-type=\"ref\"} Wang *et\u00a0al*.[12](#cas13110-bib-0012){ref-type=\"ref\"} reported that, in NSCLC, the expression of E\u2010cadherin and Twist is lower and higher, respectively, in metastatic lesions than in the primary lesion. This result suggests that the tumor cells change from the epithelial to mesenchymal phenotype in metastatic lesions.\n\nAlthough there are many previous reports of LN\u2010Mic, its influence on prognosis is yet to be determined.[13](#cas13110-bib-0013){ref-type=\"ref\"}, [14](#cas13110-bib-0014){ref-type=\"ref\"}, [15](#cas13110-bib-0015){ref-type=\"ref\"} There may also be differences in histological and phenotypic features between LN\u2010Mic and LN\u2010Mac. Recently, Aramaki *et\u00a0al*. focused on the morphological and phenotypic differences between LN\u2010Mic and LN\u2010Mac in lung adenocarcinoma cases and found that dynamic microenvironmental changes occur during the growth of LN\u2010Mic into LN\u2010Mac. These included EMT\u2010related changes in tumor cells and structural changes in stromal cells.[16](#cas13110-bib-0016){ref-type=\"ref\"}\n\nTo explain how lung SqCC in the metastatic site acquires dynamic molecular changes, characteristics of ILT and the size of LN metastasis should be considered. In this study, to address the molecular mechanisms of lymphatic metastasis and growth of LN metastasis, we compared morphological and immunophenotypic features of the tumor cells and the stromal cells in the PT, ILT, LN\u2010Mic (\u22642\u00a0mm) and LN\u2010Mac (\u226510\u00a0mm) as a large cohort of patients.\n\nMaterials and Methods {#cas13110-sec-0002}\n=====================\n\nCase selection {#cas13110-sec-0003}\n--------------\n\nA total of 102 consecutive patients with surgically resected lung SqCC harboring lymphatic permeation, LN\u2010Mic, or LN\u2010Mac seen at our hospital between April 1994 and April 2014 were enrolled in this study (Fig.\u00a0S1, Table\u00a0S1). We defined LN metastasis with a maximum diameter \u22642\u00a0mm as LN\u2010Mic; LN metastasis with a maximum diameter \u226510\u00a0mm was defined as LN\u2010Mac. Morphological characteristics were microscopically analyzed in 102 PTs, 50 ILTs, 51 LN\u2010Mics, and 82 LN\u2010Macs (Fig.\u00a0[1](#cas13110-fig-0001){ref-type=\"fig\"}).\n\n![Schema of multistep process of lymphatic metastasis in primary lung cancer. In the primary tumor (PT), tumor cells detach from their surrounding cells and intravasate into lymphatic vessels. The tumor cells float on lymphatic flow. The cells adhere to the endothelium, extravasate, and start reproliferation in the lymph nodes (LNs). (a, e) PT; (b, f) ILT; (c, g) LN\u2010Mic (\u22642\u00a0mm); (d, h) LN\u2010Mac (\u226510\u00a0mm).](CAS-108-136-g001){#cas13110-fig-0001}\n\nAll the surgical specimens were collected and analyzed after receiving the approval of the Institutional Review Board of the National Cancer Center Hospital East (Kashiwa, Japan).\n\nCompliance with ethical standards {#cas13110-sec-0004}\n---------------------------------\n\nAll procedures involving human participants were carried out in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. All work reported herein was undertaken at the National Cancer Center Hospital East, Kashiwa, Chiba, Japan and approved by the Internal Review Board of the institution. Informed consent was obtained from all individual participants included in the study.\n\nHistological studies {#cas13110-sec-0005}\n--------------------\n\nSurgical specimens were fixed in 10% formalin or methanol, embedded in paraffin, and sectioned to 4\u2010\u03bcm thickness sequentially. The sections were stained using the H&E method. All the histological materials in this study were reviewed by two pathologists (S.I. and G.I.). The pathological stage was judged based on the TNM classification of the Union for International Cancer Control, 7th edition.[5](#cas13110-bib-0005){ref-type=\"ref\"} Histological features of each lesion were judged based on the WHO classification of cell types, 3rd edition.[17](#cas13110-bib-0017){ref-type=\"ref\"}\n\nWe investigated stratified differentiation, necrosis, mitosis, and stromal reaction in each lesion. Stroma is defined as organic change including fibrosis of surrounding tumor cells. Cases were deemed positive if the stromal component accounted for more than 10% of the entire tumor. Isolated single tumor cells and clusters composed of fewer than five tumor cells were defined as \"budding cells\" and grading of budding was not assessed.\n\nAntibodies and immunohistochemical staining {#cas13110-sec-0006}\n-------------------------------------------\n\nWe selected 23 cases of lung SqCC all with lymphatic vessel permeation (evaluated by immunohistochemical staining with D2\u201040 protein), LN\u2010Mic, and LN\u2010Mac from the 102 lung SqCC cases described above (Table\u00a0S2).\n\nImmunohistochemical staining was carried out according to the methods previously reported.[16](#cas13110-bib-0016){ref-type=\"ref\"} The slides were deparaffinized in xylene and dehydrated in a graded ethanol series, and endogenous peroxidase was blocked with 3% hydrogen peroxide in absolute methyl alcohol. After epitope retrieval, the slides were washed with PBS and incubated overnight at 4\u00b0C using primary antibodies at their final dilution in the blocking buffer. The primary antibodies used in this study are summarized in Table\u00a0S3. The slides were washed again and incubated with EnVision (Dako, Glostrup, Denmark) for 1\u00a0h at room temperature. The reaction products were stained with diaminobenzidine; finally, the slides were counterstained with Meyer hematoxylin.\n\nImmunohistochemical scoring {#cas13110-sec-0007}\n---------------------------\n\nAll the stained tissue sections were scored semiquantitatively and evaluated independently under a light microscope by two pathologists (S.I. and G.I.). The immunostaining scores for EGFR, FGFR2, E\u2010cadherin, S100A4, CD44 (variant), ALDH, D2\u201040, and ezrin were evaluated based on the staining intensity and the percentage of tumor cells that were stained. The following scoring system was used: 0 (negative staining, defined as no immunoreactivity); 1 (weak staining intensity, staining between 0 and 2); and 2 (strong staining intensity, staining almost the same as positive control). We also evaluated the area of staining in 10% units (0--100%) for each lesion. The staining scores were calculated by multiplying the percentage of stained area by the staining intensity, producing scores was from 0 to 200. The score for geminin was simply the percentage of positive tumor cells per number of tumor cells counted. For CD204 and CD34, we counted the number of positive macrophages or microvessels, respectively, in HPF (0.0625\u00a0mm^2^) and calculated the mean. To produce a score for SMA\u03b1, we undertook computer\u2010assisted analysis using WinROOF image processing software (Mitani Corp., Tokyo, Japan). We set the macroinstruction, which was composed of algorithms for color identification based on red--green--blue (RGB) and hue--luminosity--saturation (HLS) parameters, to identify the immunostained area. We calculated the average proportion of the SMA\u03b1\u2010positive area in the representative three parts we selected. We confirmed that the positive control tissues were stained by each antibody, and we also undertook negative control studies without the primary antigen for all the antibodies. When the evaluation results differed, the final report was determined based on a consensus reached between the two pathologists who evaluated the slides together under a conference microscope.\n\nStatistical analysis {#cas13110-sec-0008}\n--------------------\n\nDifferences between PT, ILT, LN\u2010Mic, and LN\u2010Mac were compared using Fisher\\'s exact test and the Mann--Whitney *U*\u2010test. The Wilcoxon signed\u2010rank test was used to calculate the statistical significance of the differences of immunohistochemical studies. All *P*\u2010values are two\u2010sided, and the significance level was set at \\<0.05. All statistical analyses were carried out using IBM [spss]{.smallcaps} Statistics 22 (IBM, Armonk, NY, USA).\n\nResults {#cas13110-sec-0009}\n=======\n\nMorphological features of tumor cells {#cas13110-sec-0010}\n-------------------------------------\n\nThe results of pathological analyses of 102 samples of PT, 50 of ILT, 51 of LN\u2010Mic, and 82 of LN\u2010Mac are summarized in Figure\u00a0[2](#cas13110-fig-0002){ref-type=\"fig\"}. The median mitotic index values were 15/HPF, 0/HPF, 1/HPF, and 11/HPF, respectively. The percentages of cases with necrosis were 99.0%, 18.0%, 51.0%, and 97.6%, respectively; 27.5%, 0.0%, 0.0%, and 14.6% had budding cells, respectively, and 99.0%, 2.0%, 37.3%, and 100.0% showed stromal reaction, respectively. The median mitotic index values and the percentage of necrosis, budding cells, and stromal reaction were all significantly less in ILT and LN\u2010Mic than PT and LN\u2010Mac (*P\u00a0*\\<*\u00a0*0.001).\n\n![Morphological features of tumor cells in primary tumor (PT), intralymphatic tumor (ILT), lymph node micrometastasis (LN\u2010Mic), and lymph node macrometastasis (LN\u2010Mac). (a) Proportion of cases with necrosis. (b) Median mitotic index. (c) Budding cells. (d) Stromal reaction. (a, c, d) Fisher\\'s exact test. (b) Mann--Whitney\\'s *U*\u2010test.](CAS-108-136-g002){#cas13110-fig-0002}\n\nImmunohistochemical scores of tumor cells {#cas13110-sec-0011}\n-----------------------------------------\n\nWe selected samples from 23 lung SqCC patients who all harbored lymphatic vessel permeation, LN\u2010Mic, or LN\u2010Mac and stained 12 molecular markers.\n\n### Growth factor receptors {#cas13110-sec-0012}\n\nThe average staining score for EGFR in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 86.5, 63.9, 58.7, and 80.0, respectively. The EGFR expression level in PT was significantly higher than in ILT and LN\u2010Mic (Wilcoxon signed\u2010rank test, *P*\u00a0=\u00a00.0164 and 0.0344, respectively; Fig.\u00a0[3](#cas13110-fig-0003){ref-type=\"fig\"}a). Expression of EGFR in LN\u2010Mic was significantly lower than in LN\u2010Mac (*P*\u00a0=\u00a00.0476). The average staining score for FGFR2 in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 33.3, 18.7, 27.6, and 35.2, respectively. The FGFR2 expression level in ILT was significantly lower than in PT (*P\u00a0*\\<*\u00a0*0.001).\n\n![Comparison of immunohistochemical staining scores of tumor cells in primary tumor (PT) (b, g), intralymphatic tumor (ILT) (c, h), lymph node micrometastasis (LN\u2010Mic) (d, i), and lymph node macrometastasis (LN\u2010Mac) (e, j). The expression levels of epidermal growth factor receptor (EGFR) (a) and geminin (f) in each tumor were examined using Wilcoxon\\'s signed\u2010rank test. Representative immunohistochemical staining results for PT tissue and LN\u2010Mic and LN\u2010Mac are shown using EGFR (b--e) and geminin (g--j).](CAS-108-136-g003){#cas13110-fig-0003}\n\n### Epithelial--mesenchymal transition\u2010related markers {#cas13110-sec-0013}\n\nThere was no significant difference between each lesion type in the expression level of E\u2010cadherin and S100A4.\n\n### Cancer stem cell\u2010related molecules {#cas13110-sec-0014}\n\nThere was no significant difference between each lesion type in the expression level of CD44, ALDH, and D2\u201040.\n\n### Others {#cas13110-sec-0015}\n\nThe average staining score for geminin in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 32.3%, 20.0%, 25.9% and 35.1%, respectively. The geminin expression level in PT was significantly higher than in ILT and LN\u2010Mic (*P\u00a0*\\<*\u00a0*0.001 and 0.0307, respectively; Fig.\u00a0[3](#cas13110-fig-0003){ref-type=\"fig\"}b). Geminin expression in LN\u2010Mic was significantly lower than in LN\u2010Mac (*P*\u00a0=\u00a00.0365). There was no significant difference between each lesion type in the expression level of ezrin.\n\nImmunophenotype of stromal cells {#cas13110-sec-0016}\n--------------------------------\n\n### Microvessel density {#cas13110-sec-0017}\n\nThe average number of vessels formed by CD34\u2010positive endothelial cells in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 7.8, 0.0, 1.9, and 4.6, respectively. There were significant differences in microvessel density between each lesion type (*P\u00a0*\\<*\u00a0*0.001; Fig.\u00a0[4](#cas13110-fig-0004){ref-type=\"fig\"}).\n\n![Comparison of immunohistochemical staining of stromal cells in primary tumor (PT) (b, g, l), intralymphatic tumor (ILT) (c, h, m), lymph node micrometastasis (LN\u2010Mic) (d, i, n), and lymph node macrometastasis (LN\u2010Mac) (e, j, o). The expression levels of CD34 (a), CD204 (f), and smooth muscle actin \u03b1 (SMA\u03b1) (k) in each tumor were examined using Wilcoxon\\'s signed\u2010rank test. Representative immunohistochemical staining results for PT,ILT,LN\u2010Mic, and LN\u2010Mac are shown using CD34 (b--e), CD204 (g--j), and SMA\u03b1 (l--o). HPF, high power field.](CAS-108-136-g004){#cas13110-fig-0004}\n\n### M2 macrophage density {#cas13110-sec-0018}\n\nThe average number of CD204\u2010positive TAMs in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 29.3, 2.8, 5.3, and 24.4, respectively. The density of CD204\u2010positive TAMs in PT and LN\u2010Mac was significantly higher than in ILT and LN\u2010Mic (*P\u00a0*\\<*\u00a0*0.001 and *P\u00a0*\\<*\u00a0*0.001, respectively; Fig.\u00a0[4](#cas13110-fig-0004){ref-type=\"fig\"}f).\n\n### Myofibroblast density {#cas13110-sec-0019}\n\nThe average area of SMA\u03b1\u2010positive myofibroblasts in PT, ILT, LN\u2010Mic, and LN\u2010Mac was 10.0%,\u00a00.13%, 2.1%, and 7.7%, respectively. The area of SMA\u03b1\u2010positive myofibroblasts in PT and LN\u2010Mic was larger than that in ILT (*P\u00a0*\\<*\u00a0*0.001 and *P\u00a0*\\<*\u00a0*0.001, respectively; Fig.\u00a0[4](#cas13110-fig-0004){ref-type=\"fig\"}k), and that in LN\u2010Mac was larger than that in LN\u2010Mic (*P\u00a0*\\<*\u00a0*0.001). Detailed information of each score is provided in Table\u00a0S4.\n\nDiscussion {#cas13110-sec-0020}\n==========\n\nIn this study, we compared the histological features and the immunohistochemical staining of PT, ILT, LN\u2010Mic, and LN\u2010Mac of lung SqCC in order to elucidate the mechanism of lymphatic metastasis. We found that the proliferative capacity of tumor cells in ILT and LN\u2010Mic was significantly lower than in PT and LN\u2010Mac. Moreover, stromal reaction was less prominent in ILT and LN\u2010Mic. This is the first report that has examined structural and phenotypical changes of lung SqCC in the course of lymphatic metastasis.\n\nBy histological analysis, the mitotic indices of ILT and LN\u2010Mic were significantly lower than PT and LN\u2010Mac. This phenomenon was also confirmed by immunohistochemical staining of geminin, known as an indicator of cell proliferation. Ting *et\u00a0al*.[18](#cas13110-bib-0018){ref-type=\"ref\"} reported that circulating tumor cells clustered separately from primary tumors and showed low proliferative signatures in a mouse model of pancreatic cancer, which is consistent with our findings. M\u00fcller *et\u00a0al*.[19](#cas13110-bib-0019){ref-type=\"ref\"} reported that all of 22 circulating tumor cells detected in breast cancer patients were negative for Ki\u201067 antibody, which is also known as an indicator of cell proliferation. Therefore, we consider that the tumor cells might switch to by decreasing the proliferative capacity in the period of intralymphatic and early metastatic phase. Then, the tumor cells again proliferate to create macrometastatic tumors. Expression of EGFR decreased in ILT and LN\u2010Mic and increased from LN\u2010Mic to LN\u2010Mac, which might explain the change in the proliferative capacity of tumor cells. We also consider that the cells of lung SqCC could temporarily suppress the proliferative capacity to survive as floating intralymphatic cells as well as circulating tumor cells. As another possibility, the cell clusters with lower proliferative capacity may invade the lymphatic vessels selectively and acquire a higher one after colonization in the LN. The stem cell characteristics in each lesion were considered to be no different compared to the primary tumor, because a quantitative increase of stem cell markers such as CD44 and podoplanin was not observed in this study.\n\nWei *et\u00a0al*.[20](#cas13110-bib-0020){ref-type=\"ref\"} reported that lung adenocarcinoma cell lines avoid anoikis by altered regulation of Src and detachment of these cell lines decreases the phosphorylation of EGFR. In our study, the expression of EGFR decreased in ILT and LN\u2010Mic and increased from LN\u2010Mic to LN\u2010Mac. Although not statistically significant, FGFR2 also showed a similar tendency (Table\u00a0S4). Because the proliferation ability is reduced with decreased expression of these growth factor receptors, their downregulation may be an important mechanism by which lymphatic vessels avoid anoikis. However, it was previously reported from *in\u00a0vitro* studies that the activation of signaling through growth factor receptors was involved in anoikis resistance.[21](#cas13110-bib-0021){ref-type=\"ref\"}, [22](#cas13110-bib-0022){ref-type=\"ref\"} These controversial results may be due to technical limitations and lack of an appropriate *in\u00a0vitro* model to mimic the early phase of metastasis. Because it was reported that cetuximab, an anti\u2010EGFR antibody, is effective against NSCLC expressing high levels of EGFR,[23](#cas13110-bib-0023){ref-type=\"ref\"}, [24](#cas13110-bib-0024){ref-type=\"ref\"} these findings may open new therapeutic avenues in the future.\n\nAlthough there was some reports that EMT was involved in the metastasis of tumor cells,[23](#cas13110-bib-0023){ref-type=\"ref\"}, [24](#cas13110-bib-0024){ref-type=\"ref\"}, [25](#cas13110-bib-0025){ref-type=\"ref\"} no significant changes in EMT markers were observed between the different tumor cell phases in this study. Aramaki *et\u00a0al*.[16](#cas13110-bib-0016){ref-type=\"ref\"} undertook similar studies with LN metastasis of lung adenocarcinoma and reported the involvement of EMT during lymphatic metastasis. It was suggested that the impact of EMT on lymphatic metastasis is less important in lung SqCC.\n\nMicroenvironmental factors, such as microvessels, TAMs, and cancer\u2010associated fibroblasts influence the growth and metastasis of cancer.[7](#cas13110-bib-0007){ref-type=\"ref\"}, [25](#cas13110-bib-0025){ref-type=\"ref\"}, [26](#cas13110-bib-0026){ref-type=\"ref\"} We found that the number of CD34, CD204, and SMA\u03b1\u2010positive stromal cells significantly decreased when the tumor cells infiltrated lymphatic vessels from PT, and then increased with the engraftment of tumor cells to LN growing as micrometastasis and macrometastasis. Aramaki *et\u00a0al*.[16](#cas13110-bib-0016){ref-type=\"ref\"} examined the morphological and phenotypical differences between LN micrometastatic and macrometastatic tumors of lung adenocarcinoma. They found that the numbers of SMA\u03b1\u2010positive fibroblasts, CD34\u2010positive microvessels, and CD204\u2010positive macrophages were significantly higher for LN\u2010Mac and PT than for micrometastasis, indicating that dynamic structural changes in stromal cells occur during the growth of LN\u2010Mic into LN\u2010Mac. The latter observations' results are in part consistent with our current results. These findings suggest that the biological functions of stromal cells within the lymphatic vessels and the early phase of LN metastasis may be less important, whereas in the later stages of progression, tumor cells need them to survive and proliferate.\n\nIn conclusion, by histological and immunohistochemical comparison of each stage of lymphatic metastasis of lung SqCC, this study revealed that proliferative capacity, EGFR expression, and stromal reaction is reduced in lymphatic vessel permeation and subsequently restored during growth in the lymph nodes (Fig.\u00a0[5](#cas13110-fig-0005){ref-type=\"fig\"}). Although further investigation is required, this study emphasizes the functional importance of microenvironmental factors in the formation of metastatic tumors.\n\n![Hypothetical schema of lymphatic metastasis of lung squamous cell carcinoma. Tumor cells undergo lymph node (LN) metastasis while changing their proliferative capacity and epidermal growth factor receptor (EGFR) expression in accordance with each step. Stromal reaction is also modulated during this process.](CAS-108-136-g005){#cas13110-fig-0005}\n\nDisclosure Statement {#cas13110-sec-0022}\n====================\n\nThe authors have no conflict of interest.\n\nAbbreviationsALDHaldehyde dehydrogenaseEGFRepidermal growth factor receptorEMTepithelial--mesenchymal transitionFGFR2fibroblast growth factor receptor 2HPFhigh power fieldILTintralymphatic tumorLNlymph nodeLN\u2010Maclymph node macrometastasisLN\u2010Miclymph node micrometastasisNSCLCnon\u2010small\u2010cell lung carcinomaPTprimary tumorSMAsmooth muscle actinSqCCsquamous cell carcinomaTAMtumor\u2010associated macrophage\n\nSupporting information\n======================\n\n###### \n\n**Fig.\u00a0S1.** Schema of case selection for the morphological and immunohistochemical studies.\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Table\u00a0S1.** Characteristics of patients in the morphological study.\n\n**Table\u00a0S2.** Characteristics of patients in the immunohistochemical study.\n\n**Table\u00a0S3.** Antibodies for immunohistochemical staining.\n\n**Table\u00a0S4.** Immunohistochemical staining scores of cancer cells and stromal cells.\n\n###### \n\nClick here for additional data file.\n\nThis work was in part supported by the National Cancer Center Research and Development Fund (23\u2010A\u201016) and the Japan Society for the Promotion of Science (KAKENHI 26108007).\n"}
+{"text": "Specifications tableSubjectStrategy and ManagementSubject AreaKey drivers of Port CompetitivenessMore specific subject areaDry Bulk Port OperationType of dataTable, Graph and FigureHow data was acquiredThe data was collected from all of Malaysian Dry Bulk Terminal Ports using questionnaire distributed through email. From the primary sources, the data was digitised from corresponding archive.Data formatRaw, Filtered and analysed dataData collection parametersMachines, Conventional labor oriented (CLO), Trucking efficiency minutes, Stockpile LocationsDescription of data collectionThe data was collected from around of Malaysian ports at dry bulk terminal using an online questionnaire distributed to 18 Ports through email. The dispersion of respondents corresponds approximately to the structure of Malaysian maritime at the dry bulk terminal. The data provides full responses from the head of the port management. Each respond assessed the level of efficiency factors by percentage rate given from the total of 100%.Data source locationNo.PortRegionCoordinate1SandakanSabah5.8120\u00b0 N, 118.0769\u00b0 E2Kota Kinabalu5\u00b058\u203260.00\u2032\\'N 116\u00b04\u20320.00\u2032\\'E3Kudat6\u00b052\u203260.00\u2033N 116\u00b050\u203260.00\u2033E4Labuan5.2765\u00b0 N, 115.2430\u00b0 E5Bintulu3\u00b016\u20320.00\u2033N 113\u00b04\u20320.00\u2033E6Tawau4.2460\u00b0 N, 117.8807\u00b0 E7Lahad Datu5.0202\u00b0 N, 118.3495\u00b0 E8Tanjung ManisSarawak2.1575\u00b0 N, 111.3391\u00b0 E9Kuching1\u00b033\u203213.76\u2033N 110\u00b020\u20327.00\u2033E10Sarikei2\u00b07\u203260.00\u2033N 111\u00b031\u203260.00\u2033E11Sibu02\u00b017\u203216\u2033N 111\u00b049\u203251\u2033E12KuantanCentral3.9767\u00b0 N, 103.4242\u00b0 E13KemamanEast Coast4\u00b024\u203258.48\u2033N 103\u00b015\u203218.02\u2033E\\\n14JohorSouthern1.4438\u00b0 N, 103.9064\u00b0 E15PenangNorthern5.4098\u00b0 N, 100.3679\u00b0 E16Lumut4\u00b013\u20320.01\u2033N 100\u00b037\u20320.01\u2033E17North27.0442\u00b0 N, 82.2359\u00b0 W18WestportWestern2.9833\u00b0 N, 101.4190\u00b0 EData accessibility<https://data.mendeley.com/datasets/jxj6dt54w6/1>Related research articleRozar, N. M., Razik, M. A., & Sidik, M. H. M. (2018). The Factor Analysis of the Antecedents of Dry Bulk Terminal for Port Operation Improvement in Malaysia. International Journal of Engineering and Md. 10(6), 1801--1805.\n\nValue of the data {#sec0001}\n=================\n\n\u2022In dry bulk terminal, the data encapsulates a large number of Malaysian ports efficiency dataset.\u2022The data offers insight for assessing Malaysian Ports efficiency in dry bulk terminal where it can be used to comprehend the other terminals of Malaysian ports (e.g. changes in coastal shipping services and port facilities) into regional economic change; in the long run, give broad geographical and temporal coverage of the data.\u2022The data uncovers the variances of efficiency factors in dry bulk terminal ports and for port managers in order to build a long-term action strategy.\n\n1. Data description {#sec0002}\n===================\n\n[Table 2](#tbl0002){ref-type=\"table\"} shows the normality test from four different techniques, namely Kolmogorov-Smirnov and Shapiro-Wilk. The normality test was conducted from 10 variables as at [Table 1](#tbl0001){ref-type=\"table\"}. The result demonstrated that the dataset of Machines (VA.1), Conventional labor oriented (VA.2), Trucking efficiency \\< 15\u00a0min (VA.3), Trucking efficiency 15 -- 30\u00a0min (VA.4), Trucking efficiency \\> 30\u00a0min VA.5). These are one of the facilities for Malaysians' port managers to achieve higher level of efficiency in the port operation and it was categorised of cargo handling technology and equipment, and port information technology. Thus, affected in port trade to take initiatives to expand port capacity for trucking efficiency \\[[@bib0001]--[@bib0002]\\].Table 1Summary of the variable\\'s descriptions.Table 1SymbolDescriptionsVA. 1MachinesVA. 2Conventional labor orientedVA.3Trucking efficiency \\< 15\u00a0minVA.4Trucking efficiency 15 -- 30\u00a0minVA.5Trucking efficiency \\> 30\u00a0minVA.6Stockpile Locations \\< 1kmVA.7Stockpile Locations 1\u202fkm -- 3kmVA.8Stockpile Locations 3\u202fkm -- 5kmVA.9Stockpile Locations 5\u202fkm -- 10kmVA.10Stockpile Locations \\> 10km\n\nWhile, at [Table 2](#tbl0002){ref-type=\"table\"} shows the normality test for Stockpile Locations as at [Table 1](#tbl0001){ref-type=\"table\"}**.** were consisted Stockpile Locations \\< 1\u202fkm (VA.5), Stockpile Locations 1\u202fkm -- 3\u202fkm (VA.6) Stockpile Locations 3\u202fkm -- 5\u202fkm (VA.7), Stockpile Locations 5\u202fkm -- 10\u202fkm (VA.8), Stockpile Locations \\> 10\u202fkm (VA.9) are normal. [Table 3](#tbl0003){ref-type=\"table\"} and [Fig. 2](#fig0002){ref-type=\"fig\"} show the variability of all variables, i.e. the minimum, maximum, interquartile, median, mean standard deviation, Variance, skewness and Kurtosis. [Figs. 1](#fig0001){ref-type=\"fig\"} and [2](#fig0002){ref-type=\"fig\"} show the normality test and histogram for each variable, respectively. The strategic location of a port significantly increases its efficiencies. From [Fig. 1](#fig0001){ref-type=\"fig\"}, the mean value for 18 ports are mostly equivalent for all types of variables. However, Stockpile Locations 5\u202fkm -- 10\u202fkm (VA.10) consistently showed low value. The results were related with the position refers to of \\\"diversion distance\\\" concept where ships deviate from main trunk routes to the port. It was discussed by [@bib0003] said that the centrality of shipping routes is vital not only because it acts a port gateway but also as a hub for transhipment.Table 2Summary of the Case Processing Summary/ normality test.Table 2Tests of NormalityKolmogorov-Smirnov[a](#tb2fn2){ref-type=\"table-fn\"}Shapiro-WilkVariablesDescriptionStatisticdfSig.StatisticdfSig.VA. 1Machines.12818.200[\\*](#tb2fn1){ref-type=\"table-fn\"}.90018.058VA. 2Conventional labor oriented.27118.001.77818.001VA.3Trucking efficiency \\< 15\u00a0min.23718.009.93818.270VA.4Trucking efficiency 15 -- 30\u00a0min.28718.000.90318.066VA.5Trucking efficiency \\> 30\u00a0min.26218.002.85818.011VA.6Stockpile Locations \\< 1km.25118.004.82218.003VA.7Stockpile Locations 1\u202fkm -- 3km.32318.000.73718.000VA.8Stockpile Locations 3\u202fkm -- 5km.35818.000.71018.000VA.9Stockpile Locations 5\u202fkm -- 10km.21118.034.85518.010VA.10Stockpile Locations \\> 10km.22218.019.81818.003[^1][^2]Table 3Descriptive analysis of Demographic factors in dry bulk terminal for port efficiency.Table 3StatisticStd. ErrorVA.1Mean524.722280.3805895% Confidence Interval for MeanLower Bound355.1340Upper Bound694.31045% Trimmed Mean494.6358Median475.0000Variance116,298.683Std. Deviation341.02593Minimum100.00Maximum1491.00Range1391.00Interquartile Range426.50Skewness1.341.536Kurtosis2.6811.038VA.2Mean19.72223.7151995% Confidence Interval for MeanLower Bound11.8839Upper Bound27.56065% Trimmed Mean18.0247Median17.5000Variance248.448Std. Deviation15.76222Minimum.00Maximum70.00Range70.00Interquartile Range12.50Skewness2.085.536Kurtosis5.6161.038VA.3Mean40.83335.3359495% Confidence Interval for MeanLower Bound29.5755Upper Bound52.09125% Trimmed Mean40.0926Median40.0000Variance512.500Std. Deviation22.63846Minimum5.00Maximum90.00Range85.00Interquartile Range32.50Skewness.378.536Kurtosis\u22120.0281.038VA.4Mean43.61115.1893295% Confidence Interval for MeanLower Bound32.6626Upper Bound54.55965% Trimmed Mean43.1790Median40.0000Variance484.722Std. Deviation22.01641Minimum5.00Maximum90.00Range85.00Interquartile Range25.00Skewness.698.536Kurtosis.1691.038VA.5Mean15.55561.9339195% Confidence Interval for MeanLower Bound11.4754Upper Bound19.63585% Trimmed Mean15.3395Median20.0000Variance67.320Std. Deviation8.20489Minimum5.00Maximum30.00Range25.00Interquartile Range11.25Skewness.160.536Kurtosis\u22120.9561.038VA.6Mean18.88893.0932095% Confidence Interval for MeanLower Bound12.3628Upper Bound25.41505% Trimmed Mean18.4877Median12.5000Variance172.222Std. Deviation13.12335Minimum5.00Maximum40.00Range35.00Interquartile Range25.00Skewness.316.536Kurtosis\u22121.6341.038VA.7Mean15.55563.2561995% Confidence Interval for MeanLower Bound8.6856Upper Bound22.42555% Trimmed Mean14.7840Median10.0000Variance190.850Std. Deviation13.81484Minimum5.00Maximum40.00Range35.00Interquartile Range25.00Skewness1.000.536Kurtosis\u22120.7091.038VA.8Mean13.05563.5732095% Confidence Interval for MeanLower Bound5.5168Upper Bound20.59445% Trimmed Mean11.1728Median7.5000Variance229.820Std. Deviation15.15982Minimum.00Maximum60.00Range60.00Interquartile Range10.00Skewness2.086.536Kurtosis4.6271.038VA.9Mean6.94441.1532195% Confidence Interval for MeanLower Bound4.5114Upper Bound9.37755% Trimmed Mean6.6049Median5.0000Variance23.938Std. Deviation4.89264Minimum.00Maximum20.00Range20.00Interquartile Range5.00Skewness.773.536Kurtosis1.7621.038VA.10Mean5.0000.9038895% Confidence Interval for MeanLower Bound3.0930Upper Bound6.90705% Trimmed Mean5.0000Median5.0000Variance14.706Std. Deviation3.83482Minimum.00Maximum10.00Range10.00Interquartile Range10.00Skewness.000.536Kurtosis\u22121.1901.038Fig. 1The normality test chart for port efficiency in dry bulk terminal.Fig 1Fig. 2Histogram analysis for port efficiency in dry bulk terminal.Fig 2\n\n2. Experimental design, materials, and methods {#sec0003}\n==============================================\n\nIn summary, our ports data includes 18 different places. These ports are appearing to be consistently important places for ocean shipping. Others appear in the data in different benchmark years, which indicates real changes in use and was similar with the concept of the study by [@bib0004], but in this data has also distinct recording practices at different times and between the sources. [Fig. 2](#fig0002){ref-type=\"fig\"} shows the aggregate distribution of the number of appearances of each variables for all ports.\n\nAppendix A. Supplementary data\n\nSupplementary data to this article can be found online at <https://data.mendeley.com/datasets/jxj6dt54w6/1>\n\nDeclaration of Competing Interest\n=================================\n\nThe authors declare that they have no known competing financial interests or personal relation-ships that could have appeared to influence the work reported in this paper.\n\nAppendix. Supplementary materials {#sec0005}\n=================================\n\nImage, application 1Image, application 2\n\nThe authors also want to gratitude the Faculty Maritime Studies, University Malaysia Terengganu (UMT), Faculty Entrepreneurship and Business, University Malaysia Kelantan (UMK) for all support and dedication given along the process.\n\nSupplementary material associated with this article can be found, in the online version, at doi:[10.1016/j.dib.2020.105858](https://doi.org/10.1016/j.dib.2020.105858){#interref0003}.\n\n[^1]: This is a lower bound of the true significance.\n\n[^2]: Lilliefors Significance Correction.\n"}
+{"text": "Prof. Aurelio Bairati passed away in Milan, in December 2017.\n\nHe was Professor of Histology and Human Anatomy at the University of Milan. His interest in research began in 1951, when, still a freshman in the medical school, he had the chance to work in Berne in the laboratory of Prof. F.E. Lehmann, who was, at that time, a pioneer of biological electron microscopy. In the 60s' of the last century, Aurelio Bairati was at the forefront of ultrastructural research in Italy, studying the organization of a variety of cells and tissues.\n\nHe published extensively on the ultrastructure of the male reproductive apparatus of *Drosophila melanogaster*, thus providing the basis for future studies which proposed this insect as a model for studying male gametogenesis. Later, he extended his scientific interest to the study of the structure of the human testis and spermatozoa in infertile patients. He made influential contributions to the study of the extracellular matrix of connective tissues in vertebrates and invertebrates from the structural and comparative point of view, combining polarized light microscopy, electron microscopy, immunocytochemistry, morphometry, and biochemistry. He was a mentor for a large number of students in the field of electron microscopy and published over 200 scientific papers.\n\nProf. Bairati was a member of The Italian Society of Human Anatomy, The Italian Society of Experimental Biology, The Italian Society of Electron Microscopy, The Italian Society of Histochemistry, Unione Zoologica Italiana, Association des Anatomistes, The International Society of Matrix Biology, The European Microscopy Society, and was on the Editorial Board of the Journal of Ultrastructure Research.\n\nHe will be remembered as a passionate researcher, as a brilliant and, at the same time, unassuming scholar, and as a fair and generous leader.\n\n![](ejh-62-1-2898-g001){#fig001}\n"}
+{"text": "###### Strengths and limitations of this study\n\n-   The AIDS Therapy Evaluation in the Netherlands (ATHENA) national observational HIV cohort was established in 1998, following the introduction of triple combination antiretroviral therapy\u00a0(cART) in 1996.\n\n-   The ATHENA cohort collects clinical data from 98% of all adults and children in HIV care in the Netherlands (2% *opt-out*).\n\n-   ATHENA is an open cohort, as new participants continue to be enrolled on entry into HIV care, following a positive HIV diagnosis.\n\n-   The data are extracted from medical records: the availability of the data depends on the frequency of patient visits and the completeness and detail of the medical record.\n\n-   Findings are limited to those observed during follow-up in HIV care.\n\nIntroduction {#s1}\n============\n\nThe AIDS Therapy Evaluation in the Netherlands (ATHENA) cohort is a nationwide observational cohort that monitors HIV management of all HIV-positive people in HIV care in the Netherlands, since the introduction of highly active antiretroviral therapy (HAART\u00a0nowadays known as combination antiretroviral therapy (cART)) when the first HIV-protease inhibitor (PI) became available. As a result of the decisive action taken by the late Dutch Minister of Health, Welfare and Sport, Professor Dr Els Borst, cART including this PI became available in the Netherlands in September 1996. However, its swift introduction was conditional on the following: only specialised HIV treatment centres were allowed to prescribe cART, and the effectiveness of cART in clinical practice had to be demonstrated. Accordingly, in November 1997, the Dutch National Health Insurance Council provided a grant to the Academic Medical Center of the University of Amsterdam for the ATHENA cohort to assess the effectiveness of cART in clinical practice. The ATHENA cohort was highly successfully in its mission to 'assess the implications for the course of HIV disease, public health and health\u00a0care of the introduction of new anti-HIV treatment, the monitoring of that new treatment, as well as the economic costs and benefits of the new treatment'.\n\nSubsequently, in 2000, the Ministry of Health, Welfare and Sport institutionalised the national registration and monitoring of HIV-positive people in care by establishing *Stichting HIV Monitoring* (*SHM*; HIV Monitoring Foundation). Under the responsibility of SHM, the ATHENA cohort was continued and extended to all HIV-positive people in HIV care in the designated HIV centres (irrespective of the use of cART). Today, the mission of SHM and the ATHENA cohort is to further the knowledge and understanding of all relevant aspects of HIV infection, including comorbidities and coinfections (notably viral hepatitis), in HIV-positive people in outpatient HIV care in the Netherlands.\n\nCohort description {#s2}\n==================\n\nSetting and location {#s2a}\n--------------------\n\nThe ATHENA cohort is an observational, open cohort of all HIV-positive adults and children who enter HIV care in the Netherlands. On 1 November 2000, prior to expansion, the cohort comprised 3449 HIV-positive people, of whom 554 participated in a focus group for substudies on antiretroviral drug resistance and therapeutic drug monitoring, adherence, quality of life and costs of treatment.[@R1] The original cohort enrolled people who had died or were otherwise no longer in care, but who had used any of the antiretroviral drugs newly licenced in 1996 or later. Subsequently, the cohort enrolled all HIV-positive people entering HIV care in one of the designated HIV\u00a0treatment centres for adults and children in the Netherlands ([figure 1](#F1){ref-type=\"fig\"}), including those not yet on cART.\n\n![HIV treatment centres in the Netherlands.\u00a0Legend: adult HIV treatment centres: (1) Noordwest Ziekenhuisgroep, Alkmaar; (2) Flevoziekenhuis, Almere; (3) Academic Medical Center of the University of Amsterdam (AMC-UvA), Amsterdam; (4) DC Klinieken Lairesse --\u00a0HIV Focus Centrum, Amsterdam; (5) OLVG, Amsterdam; (6) MC Slotervaart, Amsterdam; (7) Medisch Centrum Jan van Goyen (MC Jan van Goyen), Amsterdam; (8) VUmc, Amsterdam; (9) Rijnstate, Arnhem; (10) HagaZiekenhuis (Leyweg site), Den Haag; (11) HMC (Haaglanden Medisch Centrum), Den Haag; (12) Catharina Ziekenhuis, Eindhoven; (13) Medisch Spectrum Twente (MST), Enschede; (14) Admiraal De Ruyter Ziekenhuis, Goes; (15) Universitair Medisch Centrum Groningen (UMCG), Groningen; (16) Spaarne Gasthuis, Haarlem; (17) Medisch Centrum Leeuwarden (MCL), Leeuwarden; (18) Leids Universitair Medisch Centrum (LUMC), Leiden; (19) MC Zuiderzee, Lelystad; (20) Maastricht UMC+ (MUMC+), Maastricht; (21) Radboudumc, Nijmegen; (22) Erasmus MC, Rotterdam; (23) Maasstad Ziekenhuis, Rotterdam; (24) ETZ (Elisabeth-TweeSteden Ziekenhuis), Tilburg; (25) UMC Utrecht (Universitair Medisch Centrum Utrecht), Utrecht; and\u00a0(26) Isala Zwolle. Paediatric HIV treatment centres: (A) Emma Kinderziekenhuis (EKZ), AMC-UvA, Amsterdam; (B) Beatrix Kinderziekenhuis (BKZ) UMCG, Groningen; (C) Erasmus MC-Sophia Kinderziekenhuis, Rotterdam;\u00a0and (D) Wilhelmina Kinderziekenhuis (WKZ), UMC Utrecht, Utrecht.](bmjopen-2018-022516f01){#F1}\n\nSince November 2005, on the initiative of the Curacao Red Cross Blood Bank, SHM has also assisted in the registration and monitoring of HIV-positive people in care at St. Elisabeth Hospital in Willemstad, Curacao.[@R2] Because data collection on Curacao is not part of the ATHENA\u00a0cohort, this will not be discussed further as it is beyond the scope of this profile.\n\nEligibility criteria and recruitment {#s2b}\n------------------------------------\n\nRegistration with SHM and enrolment in ATHENA is an ongoing process. First, *all* newly diagnosed HIV-positive adults and children entering HIV care are registered at SHM by their treating HIV physician. Subsequently, enrolment in the ATHENA cohort is based on an *opt-out* principle (see also: *Ethical approval and Patient consent*). By 1 January 2017, SHM had registered a cumulative number of 25 564 HIV-positive adults and children in the Netherlands, of whom 525 (2.1%) adults and 3 (1.2%) children opted-out of further data collection[@R4] ([table 1](#T1){ref-type=\"table\"}). Subsequently, a total of 25 036 (97.9%) adults and children were enrolled in the cohort with a total follow-up time of 263 600 person-years. The majority were HIV-1 positive (n=24 773; 98.9%); the remainder were HIV-2 positive (n=97; 0.4%) or had HIV-1 and HIV-2 dual infection (n=60; 0.2%) or no serological results (n=106; 0.4%).\n\n###### \n\nRegistration at SHM and enrolment in the ATHENA cohort: cumulative number of HIV-positive adults and children by 1 January 2017\n\n  HIV-treatment centre             Location     Total        Alive       Deceased     Opt-out^\\*^                                    \n  -------------------------------- ------------ ------------ ----------- ------------ ------------- ----------- ---------- --------- ---------\n  **Adults**                                                                                                                         \n  \u2003Noordwest Ziekenhuisgroep       Alkmaar      360          1.4         328          91.1          32          8.9        5         1.4\n  \u2003Flevoziekenhuis                 Almere       200          0.8         191          95.5          9           4.5        3         1.5\n  \u2003AMC-UvA                         Amsterdam    3 025        12.0        2 608        86.2          417         13.8       13        0.4\n  \u2003Hiv Focus Centrum               Amsterdam    597          2.4         593          99.3          4           0.7        0         0.0\n  \u2003MC Jan van Goyen                Amsterdam    312          1.2         270          86.5          42          13.5       4         1.3\n  \u2003OLVG                            Amsterdam    3 458        13.7        3 042        88.0          416         12.0       158       4.6\n  \u2003Slotervaart-ziekenhuis          Amsterdam    860          3.4         701          81.5          159         18.5       12        1.4\n  \u2003St Lucas Andreas                Amsterdam    468          1.8         417          89.1          51          10.9       0         0.0\n  \u2003VUmc                            Amsterdam    659          2.6         569          86.3          90          13.7       12        1.8\n  \u2003Rijnstate                       Arnhem       850          3.4         767          90.2          83          9.8        3         0.4\n  \u2003HMC                             Den Haag     1 108        4.4         1 018        91.9          90          8.1        45        4.1\n  \u2003HagaZiekenhuis -- Leyweg        Den Haag     756          3.0         649          85.8          107         14.2       32        4.2\n  \u2003Catharina Ziekenhuis            Eindhoven    695          2.7         649          93.4          46          6.6        5         0.7\n  \u2003MST                             Enschede     619          2.4         505          81.6          114         18.4       4         0.6\n  \u2003Admiraal De Ruyter Ziekenhuis   Goes         211          0.8         195          92.4          16          7.6        2         0.9\n  \u2003UMCG                            Groningen    946          3.7         846          89.4          100         10.6       31        3.3\n  \u2003Kennemer Gasthuis               Haarlem      515          2.0         459          89.1          56          10.9       4         0.8\n  \u2003MCL                             Leeuwarden   310          1.2         278          89.7          32          10.3       1         0.3\n  \u2003LUMC                            Leiden       729          2.9         661          90.7          68          9.3        41        5.6\n  \u2003MC Zuiderzee                    Lelystad     81           0.3         80           98.8          1           1.2        1         1.2\n  \u2003MUMC+                           Maastricht   934          3.7         788          84.4          146         15.6       5         0.5\n  \u2003Radboudumc                      Nijmegen     765          3.0         675          88.2          90          11.8       19        2.5\n  \u2003Erasmus MC                      Rotterdam    2 636        10.4        2 331        88.4          305         11.6       13        0.5\n  \u2003Maasstad Ziekenhuis             Rotterdam    780          3.1         724          92.8          56          7.2        8         1.0\n  \u2003ETZ                             Tilburg      1 174        4.6         1 093        93.1          81          6.9        19        1.6\n  \u2003UMC Utrecht                     Utrecht      1 749        6.9         1 556        89.0          193         11.0       65        3.7\n  \u2003Isala -- Sophia                 Zwolle       509          2.0         472          92.7          37          7.3        20        3.9\n  \u2003**Total**                                    **25 306**   **100.0**   **22 465**   **88.8**      **2 841**   **11.2**   **525**   **2.1**\n  **Paediatric**                                                                                                                     \n  \u2003EKZ, AMC-UvA                    Amsterdam    71           27.5        71           100.0         0           0.0        1         1.4\n  \u2003BKZ, UMCG                       Groningen    28           10.9        28           100.0         0           0.0        0         0.0\n  \u2003Erasmus MC -- Sophia            Rotterdam    82           31.8        80           97.6          2           2.4        0         0.0\n  \u2003WKZ, UMC Utrecht                Utrecht      77           29.8        76           98.7          1           1.3        2         2.6\n  \u2003**Total**                                    **258**      **100.0**   **255**      **98.8**      **3**       **1.2**    **3**     **1.2**\n\n\\*Opt out: never gave permission to collect clinical information.\n\nAMC-UvA, Academic Medical Center of the University of Amsterdam; ATHENA, AIDS Therapy Evaluation in the Netherlands; BKZ, Beatrix Kinderziekenhuis; EKZ, Emma Kinderziekenhuis; ETZ, Elisabeth-TweeSteden Ziekenhuis; HMC, Haaglanden Medisch Centrum; LUMC, Leids Universitair Medisch Centrum; MCL, Medisch Centrum Leeuwarden; MST, Medisch Spectrum Twente; MUMC+, Maastricht UMC+; SHM, *Stichting HIV\u00a0Monitoring*; UMCG, Universitair Medisch Centrum Groningen; UMC Utrecht, Universitair Medisch Centrum Utrecht;\u00a0WKZ, Wilhelmina Kinderziekenhuis.\n\nAt initiation, the cohort was approved by the institutional review board of all participating centres. HIV-positive people in care can opt-out after being informed by their treating physician of the purpose of collection of data. Data are pseudonymised before being provided to investigators and may be used for scientific purposes. A designated quality management coordinator safeguards compliance with the European General Data Protection Regulation.[@R5]\n\nMethods of data collection and follow-up {#s2c}\n----------------------------------------\n\nDetailed clinical and laboratory data from all enrolled HIV-positive adults and children are extracted from medical records by trained data collectors in cooperation with the responsible HIV-treating physician using an extensive, standardised protocol. In addition, 13 laboratories electronically forward test results directly to SHM. The ATHENA cohort does not dictate scheduled study visits, examinations or blood work; that is, the frequency and type of visits, examinations and laboratory measurements depend on both the participant and physician. Consequently, the cohort reflects routine HIV care, and follow-up time in the cohort is equal to follow-up time in HIV care in the Netherlands. If participants move from one HIV treatment centre to another, this transfer is registered, and follow-up continues at the new centre. Likewise, children who transfer from paediatric to adult HIV care remain in follow-up. Participant registration and data collection are organised on a national level. When transferring to a new treatment centre within the Netherlands---in case of transfer from paediatric to adult care, or transfer to another treatment centre---the participants' study identifier is retained, and data collection is continued without interruption.\n\nKey epidemiological data are collected from all HIV-positive people on registration at SHM. For national HIV surveillance, SHM reports these statistics to the *Centrum voor Infectieziektebestrijding* (*CIb*; Centre for Infectious Disease Control) of the *Rijksinstituut voor Volksgezondheid en Milieu* (*RIVM*; National Institute for Public Health and the Environment). Through the RIVM-CIb, SHM also contributes to European HIV surveillance by the European Centre for Disease Prevention and Control (ECDC).[@R6]\n\nAfter registration at SHM and enrolment in ATHENA, clinical and laboratory data from all adults and children are prospectively collected as long as the patients remain in care in a Dutch HIV treatment centre. The date of HIV diagnosis is usually retrieved from the letter of referral from the general\u00a0practitioner or sexually transmitted infections clinic, or from health records in the HIV treatment centre, and can be self-reported on entry to care if there is no documentation available. All reported HIV tests, test dates and test outcomes (including the last negative HIV tests) are captured in the cohort. The data collection process is dynamic and has evolved over the years, with additions and removals of data items based on novel insights and clinical and epidemiological requirements. While the standard data collection protocol is already comprehensive, specific protocols exist for pregnant women, children and hepatitis B (HBV) and hepatitis C (HCV) virus coinfection. A complete overview of the items that are collected can be found online.[@R7]\n\nOn enrolment, the following information is collected: demographics, history of HIV infection (including the possibility of a primary infection), most likely route of HIV acquisition, smoking, alcohol and drug use. At each visit, the following information is collected: body weight and height, blood pressure, Centers for Disease Control and Prevention\u00a0classification for HIV infection - Category B\u00a0(CDC-B) and Centers for Disease Control and Prevention\u00a0classification for HIV infection - Category C (CDC-C) events,[@R8] antiretroviral therapy, adverse events (when resulting in a change of antiretroviral treatment or when specified in the protocol[@R7]), non-AIDS comorbidities (particularly diabetes mellitus, cardiovascular disease, chronic kidney disease and non-AIDS malignancies), comedication and participant's participation in clinical trials or studies. All available laboratory results are collected at each visit: virology, immunology, chemistry, haematology and results on viral coinfections or sexually transmitted diseases. HIV genome sequences and HCV and HBV genotypes are collected when available. Furthermore, additional liver diagnostic information (fibroscan, liver pathology and radiology), and additional information concerning HBV and HCV coinfection (HBV vaccination, hepatocellular carcinoma screening, diagnosis and treatment, and virological response to antiviral treatment) are collected. During pregnancy, obstetric and pregnancy-related data are also collected. If applicable, the cause of death is classified by the physician, using the Coding of Death format.[@R9]\n\nFor children, additional information is collected at enrolment and subsequent visits: nationality, country of birth, HIV status of both parents, HIV status of their siblings, perinatal information, head circumference, puberty stage, Centers for Disease Control and Prevention classification for HIV infection - Category A (CDC-A) events,[@R8] vaccinations, information regarding psychological, locomotor and puberty development, additional treatment, care and education. Additional laboratory results collected for children comprise HIV DNA and additional viral infections.\n\nPatient and public involvement {#s2d}\n------------------------------\n\nThe community of people living with HIV in the Netherlands, represented by the Dutch Association of People Living with HIV (*HIV Vereniging*), together with healthcare providers and policy makers was closely involved in the establishment and setting-up of the ATHENA cohort in 1997 to monitor the effectiveness of providing early access to life-saving PI\u00a0containing combination therapy by the Minister of Health. The involvement of the Dutch Association of People Living with HIV continues to the present, with the association being represented on both SHM's Governing Board and Scientific Advisory Board. Findings from the ATHENA cohort are published annually in the HIV Monitoring Report, which is publicly accessible through our website (<https://www.hiv-monitoring.nl/english/research/monitoringreports/>). Each year, the Dutch Association of People Living with HIV, together with other stakeholders, are asked for input on a press release by SHM in which key findings from the annual report are publicly disseminated.\n\nFindings to date {#s3}\n================\n\nSince 2001,[@R1] the ATHENA cohort provides a comprehensive overview of the trends over time in the HIV epidemic in the Netherlands including the impact of antiretroviral treatment (<https://www.hiv-monitoring.nl/english/research/monitoringreports/>). Furthermore, it has advanced our knowledge of HIV-associated comorbidities and coinfections, as well as insight on the spread and prevention of HIV. In addition, a complete and up-to-date list of scientific publications is available online (<https://www.hiv-monitoring.nl/english/research/publications/>).\n\nParticipant characteristics {#s3a}\n---------------------------\n\nOn 1 January 2017, 19 035 HIV-1-positive participants were known to be in HIV care: 18 824 were aged 18 years or older and 211 were below 18 years of age ([table 2](#T2){ref-type=\"table\"}). The remaining 6001 participants were no longer in care because they had died (n=2776; 46%), were lost to care (n=1734; 29%) or had moved abroad (n=1491; 25%).[@R2]\n\n###### \n\nThe ATHENA cohort: participant characteristics of the 19\u00a0035 HIV-1-positive people in HIV care by 1 January 2017\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                                         Men who have sex with men   Heterosexual   Injecting drug use   Blood or blood products   Other/unknown   Total                                                       \n  ------------------------------------------------------ --------------------------- -------------- -------------------- ------------------------- --------------- ----------- ----------- ----------- ----------- ----------- -----------\n  Current age in years                                                                                                                                                                                                         \n\n  \u20030--12                                                 0\\                          0\\             0\\                   0\\                        0\\              0\\          0\\          57\\         77\\         57\\         77\\\n                                                         0.0%                        0.0%           0.0%                 0.0%                      0.0%            0.0%        0.0%        7.2%        27.2%       0.4%        2.2%\n\n  \u200313--17                                                4\\                          0\\             1\\                   0\\                        0\\              1\\          2\\          43\\         26\\         48\\         29\\\n                                                         0.0%                        0.0%           0.0%                 0.0%                      0.0%            0.7%        2.2%        5.4%        9.2%        0.3%        0.8%\n\n  \u200318--24                                                199\\                        16\\            37\\                  0\\                        0\\              3\\          1\\          46\\         41\\         264\\        79\\\n                                                         1.7%                        0.7%           1.2%                 0.0%                      0.0%            2.0%        1.1%        5.8%        14.5%       1.7%        2.2%\n\n  \u200325--34                                                1432\\                       221\\           497\\                 4\\                        2\\              18\\         10\\         87\\         32\\         1762\\       541\\\n                                                         12.0%                       9.3%           16.0%                1.8%                      2.4%            11.9%       11.2%       10.9%       11.3%       11.4%       15.2%\n\n  \u200335--44                                                2547\\                       488\\           1017\\                36\\                       10\\             24\\         22\\         115\\        38\\         3210\\       1087\\\n                                                         21.4%                       20.5%          32.8%                16.1%                     11.9%           15.9%       24.7%       14.5%       13.4%       20.7%       30.5%\n\n  \u200345--54                                                4131\\                       853\\           955\\                 92\\                       34\\             55\\         29\\         214\\        37\\         5345\\       1055\\\n                                                         34.6%                       35.9%          30.8%                41.1%                     40.5%           36.4%       32.6%       26.9%       13.1%       34.5%       29.6%\n\n  \u200355--64                                                2548\\                       538\\           415\\                 79\\                       34\\             26\\         16\\         142\\        26\\         3333\\       491\\\n                                                         21.4%                       22.6%          13.4%                35.3%                     40.5%           17.2%       18.0%       17.9%       9.2%        21.5%       13.8%\n\n  \u200365--74                                                929\\                        222\\           144\\                 13\\                       4\\              22\\         6\\          73\\         5\\          1259\\       159\\\n                                                         7.8%                        9.3%           4.6%                 5.8%                      4.8%            14.6%       6.7%        9.2%        1.8%        8.1%        4.5%\n\n  \u2003\u226575                                                   138\\                        39\\            38\\                  0\\                        0\\              2\\          3\\          18\\         1\\          197\\        42\\\n                                                         1.2%                        1.6%           1.2%                 0.0%                      0.0%            1.3%        3.4%        2.3%        0.4%        1.3%        1.2%\n\n  Current age 50\u2009years or older                                                                                                                                                                                                \n\n  \u2003No                                                    6132\\                       1150\\          2085\\                74\\                       17\\             69\\         53\\         449\\        232\\        7874\\       2387\\\n                                                         51.4%                       48.4%          67.2%                33.0%                     20.2%           45.7%       59.6%       56.5%       82.0%       50.9%       67.1%\n\n  \u2003Yes                                                   5796\\                       1227\\          1019\\                150\\                      67\\             82\\         36\\         346\\        51\\         7601\\       1173\\\n                                                         48.6%                       51.6%          32.8%                67.0%                     79.8%           54.3%       40.4%       43.5%       18.0%       49.1%       32.9%\n\n  Current age 60\u2009years or older                                                                                                                                                                                                \n\n  \u2003No                                                    9899\\                       1911\\          2776\\                180\\                      68\\             117\\        75\\         648\\        267\\        12 755\\     3186\\\n                                                         83.0%                       80.4%          89.4%                80.4%                     81.0%           77.5%       84.3%       81.5%       94.3%       82.4%       89.5%\n\n  \u2003Yes                                                   2029\\                       466\\           328\\                 44\\                       16\\             34\\         14\\         147\\        16\\         2720\\       374\\\n                                                         17.0%                       19.6%          10.6%                19.6%                     19.0%           22.5%       15.7%       18.5%       5.7%        17.6%       10.5%\n\n  Region of origin                                                                                                                                                                                                             \n\n  \u2003Netherlands                                           8669\\                       1136\\          913\\                 132\\                      41\\             97\\         17\\         360\\        102\\        10 394\\     1073\\\n                                                         72.7%                       47.8%          29.4%                58.9%                     48.8%           64.2%       19.1%       45.3%       36.0%       67.2%       30.1%\n\n  \u2003Sub-Saharan Africa                                    164\\                        636\\           1304\\                4\\                        0\\              29\\         38\\         225\\        105\\        1058\\       1447\\\n                                                         1.4%                        26.8%          42.0%                1.8%                      0.0%            19.2%       42.7%       28.3%       37.1%       6.8%        40.6%\n\n  \u2003Western Europe                                        744\\                        80\\            71\\                  24\\                       27\\             3\\          3\\          31\\         25\\         882\\        126\\\n                                                         6.2%                        3.4%           2.3%                 10.7%                     33.1%           2.0%        3.4%        3.9%        8.8%        5.7%        3.5%\n\n  \u2003South America                                         749\\                        209\\           307\\                 11\\                       1\\              3\\          9\\          37\\         13\\         1009\\       330\\\n                                                         6.3%                        8.8%           9.9%                 4.9%                      1.2%            2.0%        10.1%       4.7%        4.6%        6.5%        9.3%\n\n  \u2003Caribbean                                             453\\                        121\\           162\\                 5\\                        1\\              4\\          4\\          35\\         1\\          618\\        168\\\n                                                         3.8%                        5.1%           5.2%                 2.2%                      1.2%            2.6%        4.5%        4,4%        0.4%        4.0%        4.7%\n\n  \u2003South and Southeast Asia                              355\\                        39\\            207\\                 9\\                        1\\              6\\          11\\         27\\         8\\          436\\        227\\\n                                                         3.0%                        1.6%           6.7%                 4.0%                      1.2%            4.0%        12.4%       3.4%        2.8%        2.8%        6.4%\n\n  \u2003Other                                                 748\\                        149\\           131\\                 39\\                       13\\             9\\          7\\          75\\         26\\         1020\\       177\\\n                                                         6.3%                        6.3%           4.2%                 17.4%                     15.5%           6.0%        7.9%        9.4%        9.2%        6.6%        5.0%\n\n  \u2003Unknown                                               46\\                         7\\             9\\                   0\\                        0\\              0\\          0\\          5\\          3\\          58\\         12\\\n                                                         0.4%                        0.3%           0.3%                 0.0%                      0.0%            0.0%        0.0%        0.6%        1.1%        0.4%        0.3%\n\n  Years aware of HIV infection                                                                                                                                                                                                 \n\n  \u2003\\<1                                                   473\\                        92\\            87\\                  1\\                        0\\              7\\          0\\          32\\         7\\          605\\        94\\\n                                                         4.0%                        3.9%           2.8%                 0.4%                      0.0%            4.6%        0.0%        4.0%        2.5%        3.9%        2.6%\n\n  \u20031--2                                                  1107\\                       211\\           226\\                 2\\                        0\\              8\\          4\\          75\\         18\\         1403\\       248\\\n                                                         9.3%                        8.9%           7.3%                 0.9%                      0.0%            5.3%        4.5%        9.4%        6.4%        9.1%        7.0%\n\n  \u20033--4                                                  1336\\                       224\\           227\\                 2\\                        3\\              14\\         2\\          65\\         28\\         1641\\       260\\\n                                                         11.2%                       9.4%           7.3%                 0.9%                      3.6%            9.3%        2.2%        8.2%        9.9%        10.6%       7.3%\n\n  \u20035--10                                                 3505\\                       624\\           720\\                 18\\                       3\\              21\\         20\\         191\\        75\\         4359\\       818\\\n                                                         29.4%                       26.3%          23.2%                8.0%                      3.6%            13.9%       22.5%       24.0%       26.5%       28.2%       23.0%\n\n  \u200310--20                                                3818\\                       1025\\          1502\\                85\\                       21\\             48\\         42\\         321\\        106\\        5297\\       1671\\\n                                                         32.0%                       43.1%          48.4%                37.9%                     25.0%           31.8%       47.2%       40.4%       37.5%       34.2%       46.9%\n\n  \u2003\\>20                                                  1683\\                       197\\           330\\                 116\\                      57\\             51\\         21\\         99\\         44\\         2146\\       452\\\n                                                         14.1%                       8.3%           10.6%                51.8%                     67.9%           33.8%       23.6%       12.5%       15.5%       13.9%       12.7%\n\n  \u2003Unknown                                               6\\                          4\\             12\\                  0\\                        0\\              2\\          0\\          12\\         5\\          24\\         17\\\n                                                         0.1%                        0.2%           0.4%                 0.0%                      0.0%            1.3%        0.0%        1.5%        1.8%        0.2%        0.5%\n\n  Current CD4 count in cells/mm^3^, median/IQR range                                                                                                                                                                           \n\n                                                         670\\                        571\\           670\\                 510\\                      590\\            558\\        737\\        550\\        750\\        650\\        680\\\n                                                         510--870                    410--790       490--870             335--827                  370--832        370--750    544--980    400--790    530--1050   480--854    490--880\n\n  Current CD8 count in cells/mm^3^, median/IQR range                                                                                                                                                                           \n\n                                                         860\\                        840\\           770\\                 857\\                      886\\            800\\        890\\        840\\        740\\        860\\        780\\\n                                                         640--1180                   600--1177      573--1050            570--1200                 671--1125       530--1100   635--1290   590--1170   480--1000   630--1180   573--1050\n\n  Current HIV RNA\u00a0\\<200 copies/mL                                                                                                                                                                                              \n\n  \u2003No                                                    503\\                        144\\           209\\                 14\\                       7\\              7\\          8\\          52\\         24\\         720\\        282\\\n                                                         4.2%                        6.1%           6.7%                 6.3%                      8.3%            4.6%        9.0%        6.5%        8.5%        4.7%        7.9%\n\n  \u2003Yes                                                   10 904\\                     2137\\          2755\\                193\\                      67\\             140\\        77\\         689\\        239\\        14\u2009063\\     3217\\\n                                                         91.4%                       89.9%          88.8%                86.2%                     79.8%           92.7%       86.5%       86.7%       84.5%       90.9%       90.4%\n\n  Current HIV RNA\u00a0\\<100 copies/mL                                                                                                                                                                                              \n\n  \u2003No                                                    605\\                        170\\           243\\                 14\\                       5\\              7\\          7\\          61\\         23\\         857\\        244\\\n                                                         5.1%                        7.2%           7.8%                 6.3%                      6.0%            4.6%        7.9%        7.7%        8.1%        5.5%        6.9%\n\n  \u2003Yes                                                   10\u2009802\\                     2111\\          2721\\                193\\                      69\\             140\\        78\\         680\\        240\\        13\u2009926\\     3142\\\n                                                         90.6%                       88.8%          87.7%                86.2%                     82.1%           92.7%       87.6%       85.5%       84.8%       90.0%       88.2%\n\n  Ever AIDS                                                                                                                                                                                                                    \n\n                                                         2125\\                       714\\           676\\                 81\\                       35\\             50\\         24\\         278\\        81\\         3248\\       816\\\n                                                         17.8%                       30.0%          21.8%                36.2%                     41.7%           33.1%       27.0%       35.0%       28.6%       21.0%       22.9%\n\n  AIDS at diagnosis                                                                                                                                                                                                            \n\n                                                         1129\\                       509\\           395\\                 18\\                       6\\              31\\         12\\         196\\        43\\         1883\\       456\\\n                                                         9.5%                        21.4%          12.7%                8.0%                      7.1%            20.5%       13.5%       24.7%       15.2%       12.2%       12.8%\n\n  Current HIV treatment                                                                                                                                                                                                        \n\n  \u2003Combination antiretroviral therapy                    11 642\\                     2307\\          3013\\                229\\                      84\\             146\\        89\\         773\\        278\\        15 088\\     3464\\\n                                                         97.6%                       97.1%          97.1%                98.2%                     100.0%          96.7%       100.0%      97.2%       98.2%       97.5%       97.3%\n\n  \u2003Currently not on combination antiretroviral therapy   18\\                         2\\             2\\                   0\\                        0\\              0\\          0\\          1\\          2\\          21\\         4\\\n                                                         0.2%                        0.1%           0.1%                 0.0%                      0.0%            0.0%        0.0%        0.1%        0.7%        0.1%        0.1%\n\n  \u2003Not started                                           268\\                        68\\            89\\                  4\\                        0\\              5\\          0\\          21\\         3\\          366\\        92\\\n                                                         2.2%                        2.9%           2.9%                 1.8%                      0.0%            3.3%        0.0%        2.6%        1.1%        2.4%        2.6%\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nATHENA,\u00a0AIDS Therapy Evaluation in the Netherlands.\n\nThe HIV epidemic in the Netherlands {#s3b}\n-----------------------------------\n\nThe ATHENA cohort has generated many insights into the evolution of the Dutch HIV epidemic over time, including information on diagnosed, and estimates of as yet undiagnosed HIV infections, HIV transmission and transmission risk groups. For example, from the 1990s onwards, the annual number of new HIV diagnoses among men who have sex with men (MSM) increased from approximately 400 to well over 800 by 2009. However, since then, the registered number of new diagnoses has steadily declined, and in 2016, the estimated number of new HIV diagnoses among MSM was 545. Moreover, in the heterosexual population, the number of new diagnoses has declined to between 200 and 250 per year.[@R2] Although the overall rate of late presentation has also been declining,[@R10] targeted programmes to reduce late HIV diagnoses remain needed for all risk groups and should be prioritised for heterosexual males and females, migrants from South and Southeast Asia and sub-Saharan Africa, people over 50 years of age and certain regions in the Netherlands.[@R11]\n\nBased on the ATHENA cohort and the ECDC HIV Modelling Tool,[@R12] there were an estimated 1000 new HIV infections each year between 2000 and 2010, which subsequently decreased to 500 in 2016. In MSM, the number of new HIV infections reached a peak of approximately 900 around 2007, and then decreased to approximately 400 in 2016. The estimated number of people living with undiagnosed HIV has also decreased since 2000, although this decrease was less pronounced among MSM.[@R2] Studies combining modelling and phylogenetic analysis have shown that HIV transmission among MSM can be averted if men undergo annual HIV testing and, if found to be HIV\u00a0positive, immediately initiate cART, or, if found to be HIV-negative, half of these men who are perceived to be at high-risk for HIV acquisition initiate pre-exposure antiretroviral prophylaxis.[@R14]\n\nResponse to cART {#s3c}\n----------------\n\nIn the early 2000s, ATHENA generated additional evidence that cART was efficacious for the treatment of HIV and provided insight into CD4 cell count and viral load dynamics in both treated and untreated HIV infection.[@R15] In recent years, initiation of cART following an HIV diagnosis has been taking place increasingly earlier and irrespective of CD4 cell count. By 2016, the median CD4 cell count at start of cART had increased to 410 cells/mm^3^, and the majority of people starting cART did so within 1\u2009month of diagnosis ([figure 2](#F2){ref-type=\"fig\"}). At the same time, the time spent on a first-line regimen is increasing, and both short-term and long-term virological suppression rates are high and continue to improve. Since 2000, the annual proportion of individuals on cART with a viral load\u00a0\\>200\u2009copies/mL has decreased to approximately 3%. Consequently, earlier diagnosis in combination with immediate treatment and very low rates of virological failure on treatment have contributed to a substantial decrease in the annual number of new HIV diagnoses among MSM in the Netherlands, despite an increase in risk behaviour.[@R18]\n\n![Time between HIV diagnosis and initiation of combination antiretroviral therapy (cART) among people initiating cART by year of entry into care.](bmjopen-2018-022516f02){#F2}\n\nAmong adults who do experience virological failure, the rate of acquired drug resistance remains low and continues to decline over time.[@R2] The overall prevalence of transmitted resistance remains low over time, in line with European surveillance data.[@R21]\n\nTrends in morbidity and mortality over time {#s3d}\n-------------------------------------------\n\nSince cART became available, AIDS-related deaths have decreased dramatically and life-expectancy has improved substantially.[@R22] Moreover, data from the cohort led to the Netherlands becoming one of the first countries where HIV-positive people could get life\u00a0insurance.[@R24] The limited and decreasing number of individuals who still die of AIDS each year mainly consist of those presenting late for care.\n\nAs a result of both increased life expectancy in people living with HIV, as well as an increase in age at HIV diagnosis, the HIV-positive population in the Netherlands is ageing ([figure 3](#F3){ref-type=\"fig\"}). Modelling studies have provided insight into the expected disease burden resulting from this age shift. By 2030, an estimated 73% of the HIV-positive population in the Netherlands will be aged 50 years or older, and 84% of HIV-positive people are expected to have at least one non-communicable disease (up from 29% in 2010).[@R25] Consequently, the annual cardiovascular disease incidence and costs are predicted to increase between 2015 and 2030. However, traditional cardiovascular disease prevention interventions can optimise cardiovascular health and avert future costs, particularly if targeting high-risk individuals.[@R26]\n\n![Increasing age distribution of the ATHENA cohort over time.\u00a0Legend: this figure includes all HIV-1-positive people in clinical care. In 1996, 14% of the patients in care were younger than 30 years of age, whereas 11% were 50 years or older. In 2016, these proportions were 7% and 46%, respectively, while 16% of patients in care were 60 years of age or older. The proportion of patients in clinical care as of 31 December of each calendar year is shown according to those who were\u00a0\\<30 years of age, 30--39 years, 40--49 years, 50--59 years and 60 years or older.\u00a0ATHENA,\u00a0AIDS Therapy Evaluation in the Netherlands.](bmjopen-2018-022516f03){#F3}\n\nContinuum of care {#s3e}\n-----------------\n\nThe Netherlands is well on its way to reaching the UNAIDS 90-90-90 targets for 2020,[@R27] and Amsterdam is among the first cities to have reached and exceeded these targets.[@R28] In 2016, an estimated 89% of all people living with HIV in the Netherlands were diagnosed and linked to care, 92% of these people were on cART and 95% of those on cART had a suppressed viral load ([figure 4](#F4){ref-type=\"fig\"}). Although time from acquiring HIV to reaching viral suppression has declined over time, further decline continues to be challenged mainly by the duration of undiagnosed HIV infection.[@R2]\n\n![Continuum of HIV care for the total estimated HIV-positive population in the Netherlands in 2016.](bmjopen-2018-022516f04){#F4}\n\nPregnant women, children and adolescents {#s3f}\n----------------------------------------\n\nThe combination of opt-out HIV screening and use of cART in HIV-positive pregnant women has reduced the risk of mother-to-child transmission to\u00a0\\<1% in the Netherlands.[@R2] At the same time, an increasing number of children who acquired HIV through vertical transmission are being adopted from sub-Saharan Africa by Dutch parents. In general, paediatric HIV care has resulted in favourable treatment outcomes, and viral suppression among children has improved substantially over time.[@R30] Of note, however, is that adolescents and young adults are particularly vulnerable to virological failure, especially around and shortly after the moment of transition from paediatric to adult care.[@R31]\n\nViral hepatitis coinfection {#s3g}\n---------------------------\n\nHBV and HCV coinfections are relatively common among HIV-positive people and are closely monitored in the cohort.[@R2] However, phylogenetic analysis has shown limited overlap between HCV and HIV transmission networks, which indicates a dynamic sexual network structure of Dutch HIV-positive MSM.[@R32]\n\nUp to 2016, an increase in the prevalence of active HCV (presence of detectable HCV RNA) was reported, especially among MSM. Since the introduction and reimbursement of direct-acting antiviral agents for HCV treatment in 2014, increasing numbers of HIV-positive people with both acute and chronic HCV\u00a0coinfection have been treated for, and successfully cured of, HCV (ie, sustained virological response) ([figure 5](#F5){ref-type=\"fig\"}).[@R2] This has resulted in a rapidly declining number of coinfected individuals in HIV care who still need to be cured of HCV.\n\n![Hepatitis C continuum of care.\u00a0Legend: out of a total of 1883 individuals linked to HIV care and diagnosed with HCV, 1439 individuals (76%) were retained in care, and of these 1439, 1302 (90%) had ever received treatment for HCV. Of the 1302 individuals treated for HCV, 1247 (96%) had completed HCV treatment and had data available to calculate their HCV treatment response. Overall, 1137 of the 1247 (91%) individuals who completed treatment had achieved an SVR. As a result, 302 of the 1439 individuals (21%) who were alive and in care as of 1 May 2017 in one of the Dutch HIV treatment centres remained untreated (n=137), not successfully treated (n=85), were still being treated or had insufficient time after treatment discontinuation to allow SVR calculation (n=80). All 80 individuals in whom SVR could not yet be calculated due to insufficient time since treatment discontinuation had been treated with novel direct-acting antiviral agents combinations. For that reason, we extrapolated the observed direct-acting antiviral agents SVR rate of 97% and assumed that 97% of these 80 individuals (n=77) will eventually be successfully treated. This resulted in an estimated number of 302--77=225 individuals who remain untreated or unsuccessfully treated. SVR,\u00a0sustained\u2009virological response (ie, HCV cure).](bmjopen-2018-022516f05){#F5}\n\nStrengths and limitations {#s4}\n=========================\n\nThe unique strength of the cohort is its completeness, both in terms of coverage (ie, 98% of all HIV-positive people in care in the Netherlands), as well as the quality and extensiveness of the data collection.[@R34] Furthermore, follow-up and data collection continues uninterrupted for children who transfer from paediatric to adult HIV care and for people who transfer to another HIV treatment centre within the Netherlands.\n\nTo improve data quality, SHM has developed a quality management system based on the principals of a plan--do--check--adjust cycle. New data collectors receive intensive training and follow a personal coaching programme. SHM also has an internal helpdesk that resolves questions from the data collectors. After collection, the data are automatically checked for inconsistencies and source-verified by the data management and quality control departments.\n\nSHM has become an integral part of HIV care and works closely with all the HIV-treating physicians and the *Nederlandse Vereniging van HIV Behandelaren* (Dutch Association of HIV-Treating Physicians). Additionally, SHM has contributed to the development of care indicators and provides these statistics for the *Zichtbare Zorg* (Visible Care) programme,[@R35] which is run by the Public Health Inspection Agency and was commissioned by the Ministry of Health, Welfare and Sport to allow comparison of quality of healthcare.\n\nAnother main strength is the participation of SHM and the ATHENA cohort in regional, national and international scientific research collaborations. There is close collaboration with the Academic Medical Center of the University of Amsterdam, the Public Health Service Amsterdam through the Amsterdam Cohort Studies (ACS)[@R36]; the MSM Observational Study of Acute Infection with Hepatitis C (MOSAIC)[@R32]; Co-morbidity and Ageing with HIV (AGE~h~IV) cohort study[@R37]; EU-funded Co-morbidity in Relation to AIDS (COBRA) programme[@R38]; and the HIV Transmission Elimination Amsterdam (H-TEAM) initiative.[@R39] ATHENA collaborates with other research groups and observational cohorts in western Europe, the USA and Canada. Data sharing and collaboration has taken place or is ongoing with the Collaboration of Observational HIV Epidemiological Research Europe (COHERE),[@R40] EuroSIDA,[@R41] the European Pregnancy and Paediatric HIV Cohort Collaboration (EPPICC)[@R42] within the Paediatric European Network for the Treatment of AIDS (PENTA), the European Coordinating Committee for the Integration of Ongoing Coordination Actions Related to Clinical and Epidemiological HIV Research (EuroCoord); A Collaboration on HIV-2 infection (ACHI~E~V~2E~)[@R43]; Antiretroviral Therapy Cohort Collaboration (ART-CC)[@R44]; Data collection on Adverse events of anti-HIV Drugs (D:A:D)[@R45]; HIV Cohorts Analysed Using Structural Approaches to Longitudinal data (HIV-CAUSAL) Collaboration[@R46]; Bridging the Evolution and Epidemiology of HIV in Europe (BEEHIVE)[@R47]; HIV Resistance Response Database Initiative (RDI)[@R48]; and International epidemiology Databases to Evaluate AIDS (IeDEA).[@R50]\n\nA limitation of the cohort is that the data are based on medical records and not on scheduled data collection through questionnaires. While the data collection protocol is structured, the availability of the data depends on the frequency of patient visits and the completeness and detail of the notes in the medical record. Although this provides insight into routine HIV care in a 'real life' setting, some aspects of HIV care and disease progression might go unmeasured and therefore unnoticed. Data collection largely depends on information captured by the medical records of HIV physicians and HIV nurse\u00a0consultants. This potentially limits obtaining more detailed data regarding less severe comorbid conditions and substance use, which may not be brought to the attention of the HIV physician. Furthermore, while the cohort captures 98% of all people in HIV care, we do not know well which (eg, migrants) and how many people do not enter HIV care. Another limitation is that HIV genotype sequence data are incomplete and not available from all cohort participants.\n\nCollaboration {#s5}\n=============\n\nHIV physicians can review the data of their own treatment centre and compare these data with the full cohort through an online report builder. Statistical information or data for own research purposes can be requested by submitting a research proposal (<https://www.hiv-monitoring.nl/english/research/research-projects/>). For correspondence: hiv.monitoring\\@amc.uva.nl.\n\nThe ATHENA national observational HIV cohort is not possible without the ongoing contribution of all HIV treating physicians, nurse consultants and staff at the diagnostic laboratories and facilities in the HIV treatment centres, along with the data collecting and monitoring staff both within and outside Stichting HIV Monitoring (listed under collaborators). We would like to extend our sincerest thanks and appreciation to representatives of the Dutch Association of People Living with HIV (HIV Vereniging) for being represented on SHM's Governing Board and Scientific Advisory Board.\n\n**Contributors:** TSB wrote the first draft of the manuscript. AvS, CJS, FWNMW and TSB performed the analyses for the key statistics, tables and figures of the cohort. PR provided overall supervision of the cohort profile. All authors contributed to interpreting the data and to the writing and revising of the manuscript.\n\n**Funding:** The ATHENA cohort is managed by Stichting HIV Monitoring and supported by a grant from the Dutch Ministry of Health, Welfare and Sport through the Centre for Infectious Disease Control of the National Institute for Public Health and the Environment.\n\n**Competing interests:** None declared.\n\n**Patient consent:** Not required.\n\n**Ethics approval:** At initiation, the cohort was approved by the institutional review board of all participating centres. People entering HIV care receive written material about participation in the ATHENA cohort and are being informed by their treating physician of the purpose of collection of data, after which they can consent verbally or elect to opt-out. Data are pseudonymised before being provided to investigators and may be used for scientific purposes. A designated quality management coordinator safeguards compliance with the European General Data Protection Regulation.\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n\n**Data sharing statement:** HIV physicians can review the data of their own treatment centre and compare these data with the full cohort through an online report builder. Statistical information or data for own research purposes can be requested by submitting a research proposal (<https://www.hiv-monitoring.nl/english/research/research-projects/>). For correspondence: hiv.monitoring\\@amc.uva.nl.\n\n**Collaborators:** CLINICAL CENTRES *\\* denotes site coordinating physician* Amsterdam UMC, AMC site, Amsterdam: *HIV treating physicians:* M. van der Valk \\*, S.E. Geerlings, M.H. Godfried, A. Goorhuis, J.W. Hovius, T.W. Kuijpers, F.J.B. Nellen, D. Pajkrt, T. van der Poll, J.M. Prins, P. Reiss, H.J. Scherpbier, M. van Vugt, W.J. Wiersinga, F.W.M.N. Wit. *HIV nurse consultants:* M. van Duinen, J. van Eden, A.M.H. van Hes, F.J.J. Pijnappel, A.M. Weijsenfeld. *HIV clinical virologists/chemists:* S. Jurriaans, N.K.T. Back, H.L. Zaaijer, B. Berkhout, M.T.E. Cornelissen, C.J. Schinkel, K.C. Wolthers. Amsterdam UMC, VUmc site, Amsterdam: *HIV treating physicians:* E.J.G. Peters\\*, M.A. van Agtmael, M. Bomers. *HIV nurse consultants:* M. Heitmuller, L.M. Laan. *HIV clinical virologists/chemists:* C.W. Ang, R. van Houdt, A.M. Pettersson, C.M.J.E. Vandenbroucke-Grauls. Emma Kinderziekenhuis (AmsterdamUMC): *HIV nurse consultants:* C. de Boer, A. van der Plas, A.M. Weijsenfeld. Admiraal De Ruyter Ziekenhuis, Goes: *HIV treating physicians:* M. van den Berge, A. Stegeman. *HIV nurse consultants:* S. Baas, L. Hage de Looff. *HIV clinical virologists/chemists:* B Wintermans, J Veenemans. Catharina Ziekenhuis, Eindhoven: *HIV treating physicians:* M.J.H. Pronk\\*, H.S.M. Ammerlaan. *HIV nurse consultants:* E.S. de Munnik. *HIV clinicalvirologists/chemists:* A.R. Jansz, J. Tjhie, M.C.A. Wegdam, B. Deiman, V. Scharnhorst. DC Klinieken Lairesse - Hiv Focus Centrum: *HIV treating physicians:* A. van Eeden\\*, M. van der Valk. *HIV nurse consultants:* W. Brokking, M. Groot, L.J.M. Elsenburg. *HIV clinical virologists/chemists:* M. Damen, I.S. Kwa. ETZ(Elisabeth-TweeSteden Ziekenhuis), Tilburg: *HIV treating physicians:* M.E.E. van Kasteren\\*, A.E. Brouwer. *HIV nurse consultants:* R. van Erve, B.A.F.M. de Kruijf-van de Wiel, S. Keelan-Pfaf, B. van de Ven. *Data collection:* B.A.F.M. de Kruijf-van de Wiel, B. van der Ven. *HIV clinical virologists/chemists:* A.G.M. Buiting, P.J. Kabel, D. Versteeg. Erasmus MC, Rotterdam: *HIV treating physicians:* M.E. van der Ende\\*, H.I. Bax, E.C.M. van Gorp, J.L. Nouwen, B.J.A. Rijnders, C.A.M. Schurink, A. Verbon, T.E.M.S. de Vries-Sluijs, N.C. deJong-Peltenburg. *HIV nurse consultants:* N. Bassant, J.E.A. van Beek, M. Vriesde, L.M. van Zonneveld. *Data collection:* H.J. van den Berg-Cameron, J. deGroot, M. de Zeeuw-de Man. *HIV clinicalvirologists/chemists:* C.A.B. Boucher, M.P.G Koopmans, J.J.A van Kampen, S.D. Pas. Erasmus MC--Sophia, Rotterdam: *HIV treating physicians:* P.L.A. Fraaij, A.M.C. van Rossum, C.L. Vermont. *HIV nurse consultants:* L.C. van der Knaap, E. Visser. Flevoziekenhuis, Almere: *HIV treating physicians:* J. Branger\\*, R.A. Douma. *HIV nurse consultant:* C.J.H.M. Duijf-van de Ven. HagaZiekenhuis, Den Haag: *HIV treating physicians:* E.F. Schippers\\*, C. van Nieuwkoop. *HIV nurse consultants:* J.M. van IJperen, J. Geilings. *Datacollection:* G. van der Hut. *HIV clinicalvirologist/chemist:* N.D. van Burgel. HMC (Haaglanden Medisch Centrum), Den Haag: *HIV treating physicians:* E.M.S. Leyten\\*, L.B.S. Gelinck. *HIV nurse consultants:* S. Davids-Veldhuis, A.Y. van Hartingsveld, C. Meerkerk, G.S. Wildenbeest. *HIV clinicalvirologists/chemists:* E. Heikens. Isala, Zwolle: *HIV treating physicians:* P.H.P. Groeneveld\\*, J.W. Bouwhuis, A.J.J. Lammers. *HIV nurseconsultants:* S. Kraan, A.G.W. van Hulzen, M.S.M. Kruiper. *Datacollection:* G.L. van der Bliek, P.C.J. Bor. *HIV clinicalvirologists/chemists:* P. Bloembergen, M.J.H.M. Wolfhagen, G.J.H.M. Ruijs. Leids Universitair Medisch Centrum, Leiden: *HIV treating physicians:* F.P. Kroon\\*, M.G.J. de Boer, H. Scheper, H. Jolink. *HIV nurse consultants:* W. Dorama, N. van Holten. *HIV clinical virologists/chemists:* E.C.J. Claas, E. Wessels. Maasstad Ziekenhuis, Rotterdam: *HIV treating physicians:* J.G. den Hollander\\*, K. Pogany, A. Roukens. *HIV nurse consultants:* M. Kastelijns, J.V. Smit, E. Smit, D. Struik-Kalkman, C. Tearno. *Data collection:* T. van Niekerk. *HIV clinicalvirologists/chemists:* O. Pontesilli. Maastricht UMC+, Maastricht: *HIV treating physicians:* S.H. Lowe\\*, A.M.L. Oude Lashof, D. Posthouwer. *HIV nurse consultants:* R.P. Ackens, K. Burgers, J. Schippers. *Data collection:* B. Weijenberg-Maes. *HIV clinical virologists/chemists:* I.H.M. van Loo, T.R.A. Havenith. MC Slotervaart, Amsterdam: *HIV treating physicians:* J.W. Mulder\\*, S.M.E. Vrouenraets, F.N. Lauw. *HIV nurse consultants:* M.C. van Broekhuizen, D.J. Vlasblom. *HIV clinical virologists/chemists:* P.H.M. Smits. MC Zuiderzee, Lelystad: *HIV treating physicians:* S. Weijer\\*, R. El Moussaoui. *HIV nurse consultant:* A.S. Bosma. Medisch Centrum Leeuwarden, Leeuwarden: *HIV treating physicians:* M.G.A. van Vonderen\\*, L.M. Kampschreur. *HIV nurse consultants:* K. Dijkstra, S. Faber. *HIV clinical virologists/chemists:* J Weel. Medisch Spectrum Twente, Enschede: *HIV treatingphysicians:* G.J. Kootstra\\*, C.E. Delsing. *HIV nurse consultants:* M. van derBurg-van de Plas, H. Heins. Noordwest Ziekenhuisgroep, Alkmaar: *HIV treating physicians:* W. Kortmann\\*, G. van Twillert\\*, R. Renckens. *HIV nurse consultant anddata collection:* D. Ruiter-Pronk, F.A. vanTruijen-Oud. *HIV clinicalvirologists/chemists:* J.W.T. Cohen Stuart, E.P. IJzerman, R. Jansen, W. Rozemeijer W. A. van der Reijden. OLVG, Amsterdam: *HIV treating physicians:* K. Brinkman\\*, G.E.L. van den Berk, W.L. Blok, P.H.J. Frissen, K.D. Lettinga W.E.M. Schouten, J. Veenstra. *HIV nurse consultants:* C.J. Brouwer, G.F. Geerders, K. Hoeksema, M.J. Kleene, I.B. vander Mech\u00e9, M. Spelbrink, A.J.M. Toonen, S. Wijnands. *HIV clinical virologists:* D. Kwa. *Data collection:* R. Regez (coordinator). Radboudumc, Nijmegen: *HIV treating physicians:* R. van Crevel\\*, M. Keuter, A.J.A.M. van der Ven, H.J.M. ter Hofstede, A.S.M. Dofferhoff, S.S.V. Henriet, M. van de Flier, K. van Aerde, J. Hoogerwerf. *HIV nurse consultants:* M. Albers, K.J.T. Grintjes-Huisman, M. de Haan, M. Marneef, R. Strik-Albers. *HIV clinical virologists/chemists:* J. Rahamat-Langendoen, F.F. Stelma. *HIV clinical pharmacology consultant:* D. Burger. Rijnstate, Arnhem: *HIV treating physicians:* E.H. Gisolf\\*, R.J. Hassing, M. Claassen. *HIV nurse consultants:* G. ter Beest, P.H.M. van Bentum, N. Langebeek. *HIV clinical virologists/chemists:* R. Tiemessen, C.M.A. Swanink. Spaarne Gasthuis, Haarlem: *HIV treating physicians:* S.F.L. van Lelyveld\\*, R. Soetekouw. *HIV nurse consultants:* L.M.M. van der Prijt, J. van der Swaluw. *Data collection:* N. Bermon. *HIV clinical virologists/chemists:* W.A. van der Reijden, R. Jansen, B.L. Herpers, D. Veenendaal. Medisch Centrum Janvan Goyen, Amsterdam: *HIV treating physicians:* D.W.M. Verhagen. *HIV nurse consultants:* M. van Wijk. Universitair Medisch Centrum Groningen, Groningen: *HIV treating physicians:* W.F.W. Bierman\\*, M. Bakker, J. Kleinnijenhuis, E. Kloeze, E.H. Scholvinck, Y. Stienstra, C.L. Vermont, K.R. Wilting, M. Wouthuyzen-Bakker. *HIV nurse consultants:* A. Boonstra, H. de Groot-de Jonge, P.A. van der Meulen, D.A. de Weerd. *HIV clinical virologists/chemists:* H.G.M. Niesters, C.C. van Leer-Buter, M. Knoester. Universitair Medisch Centrum Utrecht, Utrecht: *HIV treating physicians:* A.I.M. Hoepelman\\*, J.E. Arends, R.E. Barth, A.H.W. Bruns, P.M. Ellerbroek, T. Mudrikova, J.J. Oosterheert, E.M. Schadd, M.W.M. Wassenberg, M.A.D. van Zoelen. *HIV nurse consultants:* K. Aarsman, D.H.M. van Elst-Laurijssen, I. de Kroon, C.S.A.M. van Rooijen. *Data collection:* M. van Berkel, C.S.A.M. van Rooijen. *HIV clinical virologists/chemists:* R. Schuurman, F. Verduyn-Lunel, A.M.J. Wensing. Wilhelmina Kinderziekenhuis, UMC Utrecht, Utrecht: *HIV treating physicians:* L.J. Bont, S.P.M. Geelen, T.F.W. Wolfs. *HIV nurse consultants:* N. Nauta. COORDINATING CENTRE *Director:* P Reiss. *Deputy director:* S Zaheri. *Data analysis:* D O Bezemer, A I van Sighem, C Smit, F W M N Wit and T S Boender. *Data management and quality control:* M Hillebregt, A de Jong and T Woudstra. *Data monitoring:* D Bergsma, S Grivell, R Meijering, M Raethke and T Rutkens. *Data collection:* L de Groot, M van den Akker, Y Bakker, M Bezemer, E Claessen, A El Berkaoui, J Geerlinks, J Koops, E Kruijne, C Lodewijk, E Lucas, R van der Meer, L Munjishvili, F Paling, B Peeck, C Ree, R Regtop, Y Ruijs, M Schoorl, E Tuijn, L Veenenberg, S van der Vliet, A Wisse and E C Witte. *Patient registration:* B Tuk.\n"}
+{"text": "Background\n==========\n\nHypercholesterolemia represents one of the important and recognized risk factor for atherosclerosis \\[[@B1]\\]. There are compelling evidences indicating the importance of type of fats than the total amount of the fats with respect to the risk of the cardiovascular diseases \\[[@B2]\\]. Cholesterol is transported within lipoproteins in the blood stream. High density lipoprotein (HDL) cholesterol levels are inversely related to the risk for atherosclerotic events \\[[@B3]\\] and are found to possess anti-atherogenic activity \\[[@B4]\\]. Among the risk factors, total cholesterol/HDL cholesterol ratio is considered to be the most predictive for atherosclerosis \\[[@B5]\\]. The protective effect of the HDL is related partly to enzymes associated with HDL \\[[@B6]-[@B8]\\] and due to its participation in reverse cholesterol transport \\[[@B9]\\].\n\nParaoxonase1 (PON1) is one of the enzymes associated with HDL \\[[@B10]\\]. PON1 was shown to protect against oxidative stress \\[[@B11],[@B12]\\], a phenomenon that can be attributed to its ability to modulate oxidized lipids in LDL and HDL \\[[@B13],[@B14]\\], in macrophages \\[[@B15],[@B16]\\] and also in atherosclerotic plaques \\[[@B17]\\]. PON is capable of hydrolyzing lipid peroxides in LDL \\[[@B18]\\]. Serum HDL-associated PON1 reduces oxidative stress in lipoproteins, in macrophages and in atherosclerotic lesion, whereas PON2 acts as an antioxidant at the cellular and not humoral level. The attenuation to atherosclerosis is related to the nutritional anti-oxidative induced increase in HDL-PON activity \\[[@B19]\\].\n\nPAF-AH is the major enzyme responsible for the catabolism of PAF and PAF like lipids that are also the potent mediators of inflammation \\[[@B20],[@B21]\\]. Genetic deficiency of PAF-AH in defined human populations increases the severity of atherosclerosis and other syndromes \\[[@B22]\\]. PAF-AH has marked preference for phospholipids with short chain moieties at syn-2 position and, with the exception of PAF, it can equally hydrolyze oxidized phospholipids containing at syn-2 position polyunsaturated fatty acyl residues \\[[@B23]\\]. However, during hydrolysis the oxidized phospholipids, PAF-AH liberate the bioactive oxidized free fatty acids \\[[@B24]\\] and generates lysophopsphatidylcholine, both of which are implicated in the biological actions of ox-LDL \\[[@B25]\\]. Thus, PAF-AH could play both pro-atherogenic and anti-atherogenic role.\n\nSelenium, an essential trace element, is associated with cardiovascular diseases since years. Selenium deficiency is related to increase in plasma cholesterol levels \\[[@B26],[@B27]\\], cardiac myopathy \\[[@B28]\\], other cardiovascular disease and ischemic heart diseases \\[[@B29],[@B30]\\]. Selenium supplementation leads to decrease in total cholesterol and triglyceride levels \\[[@B31],[@B32]\\].\n\nKeeping these in view, in the present study influence of selenium was explored on HDL associated enzymes, PON1 and PAF-AH.\n\nResults\n=======\n\nSelenium levels\n---------------\n\nSelenium levels were estimated in the serum of rats from all the groups after 4 months of diet feeding schedule (Figure [1](#F1){ref-type=\"fig\"}). Significant decrease (P \\< 0.05) in the Se levels in HFD group was observed in comparison to the control group. However, an apparent increase (P \\< 0.001) in Se was observed on HFD + Se (1 ppm) supplementation in comparison to the rats fed on only HFD diet.\n\n![**Serum selenium levels**. \\*P \\< 0.05 represents comparison between control and HFD, ^aaa^P \\< 0.001 represents comparison between HFD and HFD + Se group.](1476-511X-8-55-1){#F1}\n\nLipid profile\n-------------\n\nLipid profile analysis was done in serum (Table [1](#T1){ref-type=\"table\"}). Total cholesterol (P \\< 0.001), triglycerides (P \\< 0.05) and HDL cholesterol (P \\< 0.05) levels were found to be significantly increased in HFD group in comparison to control group. A highly significant increase (P \\< 0.001) was observed in LDL-cholesterol and total cholesterol/HDL-cholesterol ratio in HFD group in comparison to the control group. However, no significant change in various lipid profile parameters except decrease in triglycerides levels was observed in HFD + Se group in comparison to the only HFD fed group.\n\n###### \n\nLipid profile in serum after 4 months of Control, HFD and HFD + Se feeding schedule\n\n  Lipid profile (mg/dL)         Control         HFD                    HFD + Se\n  ----------------------------- --------------- ---------------------- ------------------\n  **Total cholesterol**         71.79 \u00b1 13.66   451.66 \u00b1 26.73\\*\\*\\*   438.20 \u00b1 53.91\n                                                                       \n  **Triglycerides**             82.50 \u00b1 6.96    91.98 \u00b1 0.02\\*         81.26 \u00b1 11.96^a^\n                                                                       \n  **HDL-C**                     24.76 \u00b1 4.26    31.70 \u00b1 6.56\\*         37.58 \u00b1 5.55\n                                                                       \n  **LDL-C**                     32.66 \u00b1 2.81    395.52 \u00b1 26.51\\*\\*\\*   381.00 \u00b1 53.58\n                                                                       \n  **Total cholesterol/HDL-C**   2.90 \u00b1 0.22     14.86 \u00b1 3.86\\*\\*\\*     11.99 \u00b1 2.95\n\nData is represented as Mean \u00b1 S.D. from 5 independent observations. Data is statistically analyzed by student\\'s t-test. \\*\\*p \\< 0.01,\\*p \\< 0.05 represent the comparison between control and HFD group. ^a^p \\< 0.05 represents the comparison between HFD and HFD+ Se group.\n\nROS (Reactive oxygen species) levels\n------------------------------------\n\nROS levels in liver homogenates were estimated using a fluorescence probe DCFH-DA. Oxidation of DCFH-DA to DCF was measured as an index of total ROS. A highly significant (P \\< 0.001) increase in ROS levels was observed in HFD group in comparison to the control group. However, a significant (P \\< 0.01) decrease in ROS levels were observed in HFD+Se fed group in comparison to HFD fed group (Figure [2](#F2){ref-type=\"fig\"}).\n\n![**Reactive oxygen species levels**. \\*\\*\\*P \\< 0.001 represents comparison between control and HFD, ^aa^P \\< 0.01 represents comparison between HFD and HFD + Se group.](1476-511X-8-55-2){#F2}\n\nPON1 activity\n-------------\n\nPON1 activity was estimated in serum using substrate paraoxon. A highly significant (P \\< 0.001) decrease in the level of PON1 activity was observed in HFD fed group in comparison to control group. However, a significant (P \\< 0.01) increase in the level of PON1 was observed in HFD+Se fed group in comparison to HFD fed group (Figure [3](#F3){ref-type=\"fig\"}).\n\n![**PON1 enzyme activity**. \\*\\*\\*P \\< 0.001 represents comparison between control and HFD, ^aa^P \\< 0.01 represents comparison between HFD and HFD + Se group.](1476-511X-8-55-3){#F3}\n\nmRNA expression of PON1 and PAF-AH\n----------------------------------\n\n**S**tatistically no significant changes in the mRNA expression of PON1 and PAF-AH were observed in HFD or HFD + Se fed groups in comparison to the control group or HFD fed group respectively Figure [4a](#F4){ref-type=\"fig\"}, Figure [4b](#F4){ref-type=\"fig\"}.\n\n![**A - mRNA expression of PON1 and PAF-AH by RT-PCR.** B - Densitometric analysis of PON1 and PAF-AH mRNA expression. Data is represented as Mean \u00b1 S.D. from four independent observations. Data is statistically analyzed by student\\'s t-test.](1476-511X-8-55-4){#F4}\n\nProtein expression of PON1 and PAF-AH by ELISA\n----------------------------------------------\n\nPON1 protein levels were significantly (\\*\\*P \\< 0.01) decreased in HFD fed group in comparison to control group. However, PON1 levels significantly (\\*P \\< 0.05) increased in HFD + Se fed group in comparison to HFD fed group. On the other hand, PAF-AH protein levels were significantly (P \\< 0.05) decreased in HFD fed group in comparison to the control group. However, there was no significant change observed in HFD + Se fed group in comparison to the HFD fed group (Table [2](#T2){ref-type=\"table\"}).\n\n###### \n\nPON1 and PAF-AH levels by ELISA after 4 months of control, HFD and HFD + Se feeding schedule\n\n  A~405\\ nm~   Control       HFD               HFD + Se\n  ------------ ------------- ----------------- ----------------\n  **PON1**     1.55 \u00b1 0.13   1.06 \u00b1 0.01\\*\\*   1.28 \u00b1 0.12^a^\n                                               \n  **PAF-AH**   0.26 \u00b1 0.01   0.25 \u00b1 0.01\\*     0.26 \u00b1 0.01\n\nData is represented as Mean \u00b1 S.D. from 5 independent observations. Data is statistically analyzed by student\\'s t-test. \\*\\*p \\< 0.01,\\*p \\< 0.05 represent the comparison between control and HFD group. ^a^p \\< 0.05 represents the comparison between HFD and HFD+ Se group.\n\nDiscussion\n==========\n\nHypercholesterolemia represents one of the very important and recognized risk factor for atherosclerosis \\[[@B33]\\]. Abnormally high cholesterol levels (high concentration of LDL and low concentration of HDL) are strongly associated with the cardiovascular diseases. High cholesterol diet leads to the cholesterol deposition in the arterial walls \\[[@B34]\\]. Compelling evidences indicate the importance of the type of fats than the total amount of the fats with respect to the risk of the cardiovascular diseases \\[[@B35]\\]. Controlled clinical trials have shown that replacing saturated fats with unsaturated is more effective in lowering serum cholesterol and reducing the risk of cardiovascular diseases than simply reducing total fat consumption \\[[@B36]\\].\n\nSelenium, an essential trace element is proved to be protective against cardiovascular diseases \\[[@B37]\\]. In the earlier experiments in the authors laboratory, it was found that selenium supplementation at 1 ppm level along with high fat diet (HFD) feeding in rats inhibited the incidence of atherosclerosis as studied by Scanning electron microscopy (SEM) of aorta \\[[@B38]\\]. In the present study, interestingly the serum selenium levels decreased in HFD group in comparison to the control group. On the other hand, selenium level, as expected, increased when we gave external supplementation of 1 ppm as sodium selenite in HFD + Se fed group.\n\nExcess of lipids in the serum derived from endogenous synthesis/dietary sources initiate atherosclerosis by accumulation in the cells of the arterial wall and provoking atheroma growth \\[[@B39]\\]. In the present study, total cholesterol and triglycerides levels were found to be significantly increased in HFD group in comparison to control group. Triglyceride levels were found to be greatly reduced in HFD + Se fed group in comparison to HFD fed group. These findings suggest the hypercholesterolemic state as reported earlier \\[[@B40]-[@B42]\\]. The selenium potential against hypercholesterolemia is supported by some other research groups as well \\[[@B41],[@B43]\\].\n\nLDL has long been implicated in the development of atherosclerosis \\[[@B44]\\]. It has also been reported that selenium supplementation protect LDL from oxidation and other atherogenic changes \\[[@B45],[@B46]\\]. HDL has been found to protect against the oxidation of LDL by metal ions in-vitro \\[[@B47],[@B48]\\] and by reverse cholesterol transport \\[[@B49]\\]. Further, LDL and HDL levels were found to be significantly increased in HFD group in comparison to the control group. Increase in HDL levels \\[[@B41],[@B43]\\] on Se supplementation was also reported earlier. Moreover, total cholesterol/HDL ratio is considered to be most predictive for atherosclerosis \\[[@B48],[@B50]\\]. In the present study, this ratio was found to be increased in HFD group in comparison to the control group suggesting greatest risk for the clinical events related to hypercholesterolemia in HFD group in comparison to other groups.\n\nReactive oxygen species (ROS) levels were estimated in liver homogenates and were found to be increased 2-folds in HFD group in comparison to the control group. Se supplementation, however, diminished the HFD-induced ROS levels by 29%. This suggests the presence of oxidative stress in HFD fed group which decreased on 1 ppm selenium supplementation possibly due to the anti-oxidative property of Se. Previously also in the author\\'s laboratory, on analysis of glutathione peroxidase, lipid peroxidation, nitric oxide synthase (NOS) and reduced/oxidized glutathione ratio in aorta, liver and serum, it was demonstrated that the reduced incidence of atherosclerosis on selenium supplementation was due to the anti-oxidant function of selenium \\[[@B51],[@B52]\\]. Also in-vitro studies, in author\\'s lab, demonstrated the role of mitogen stimulated lymphocytes and macrophage NO production on selenium supplementation in HFD-induced atherogenesis in rats \\[[@B53]\\].\n\nWork from a number of laboratories have suggested that the protective effect of HDL may be due to the enzymes associated with HDL \\[[@B6],[@B48],[@B54]\\] i.e PON1 (Paraoxonase 1), LCAT (Lecithin acyl transferase), PAF-AH (platelet activating factor-acetyl hydrolase)\n\nPON1, an HDL associated enzyme, synthesized in liver, was shown to protect against oxidative stress \\[[@B11],[@B12]\\], a phenomenon that can be attributed to its ability to modulated oxidized lipids in LDL and HDL \\[[@B13],[@B14]\\], in macrophages \\[[@B15],[@B16]\\] and also in atherosclerotic lesions \\[[@B17]\\]. In the present study, biochemically, PON1 enzymes activity was found to be significantly reduced in HFD group in comparison to control group. In addition, protein expression of PON1 by Elisa was also found to be significantly reduced in HFD group in comparison to control group. However, Se supplementation improved the HFD-mediated reduction of serum PON1 enzyme activity by 34% and PON1 protein levels by 21%. This suggests that oxidative stress under hypercholesterolemic state leads to the reduced activity of PON1 whereas on selenium supplementation, levels started retrieving. Interestingly, the PON1 mRNA expression studies revealed that there was no significant change at transcriptional level. This suggests that PON1 may possibly be involved in anti-atherogenic activities at translational levels. In conclusion, Se supplementation appears to be protective in experimental hypercholesterolemia by restoring the antioxidant properties of the HDL associated enzyme, PON1.\n\nPAF-AH is the major enzyme responsible for the catabolism of PAF (platelet activating factor). Hypercholesterolemia and inflammation work as partners in atherogenesis \\[[@B55],[@B56]\\]. Evidences have suggested that genetic deficiency of PAF-AH in defined human population increases the severity of atherosclerosis and other syndromes \\[[@B57]\\]. PAF-AH possesses marked preference for hydrolyzing oxidized phospholipids containing at syn-2 position polyunsaturated fatty acyl residues \\[[@B58]\\]. However, during hydrolysis the oxidized phospholipids, PAF-AH liberate the bioactive oxidized free fatty acids \\[[@B24]\\] and generates lysophopsphatidylcholine, both of which are implicated in the biological actions of ox-LDL \\[[@B25]\\]. Thus, PAF-AH could play both atherogenic and anti-atherogenic role. In the present study, PAF-AH protein expression by ELISA was found to be significantly decreased in HFD group in comparison to control group whereas no statistically significant change was observed on selenium supplementation. Interestingly, on the other hand mRNA expression revealed that there was no significant change in PAF-AH at the transcriptional levels. From the present study, it is inferred that PAF-AH possibly participates in both pro and anti-atherogenic activities as here the system aims towards balancing the PAF-AH levels but it may need some further investigations in order to designate the exact role of PAF-AH.\n\nConclusion\n==========\n\nIn conclusion, from the results obtained by the present study, Se supplementation appears to be protective in experimental hypercholesterolemia by restoring the anti-oxidative properties of the HDL associated enzyme PON1. However, findings are inconclusive in determining the role of PAF-AH whether pro or anti-atherogenic in role or both and may need some further investigations. Therefore, Se supplementation might be a valuable approach to limit the adverse effects of hypercholesterolemia.\n\nMaterials and methods\n=====================\n\nChemicals\n---------\n\nSodium selenite (Na~2~SeO~3~), 2, 3-diaminonapthalene and Dichlorofluorescein diacetate (DCFH-DA) were purchased from Sigma-Aldrich (St. Louis, Missouri, USA). TRI-reagent and one step RT-PCR kits were obtained from Molecular research Centre (Inc. Cincinnati, Ohio) and QIAGEN, respectively. Antibodies against PON1 and PAF-AH were obtained from Santa-Cruz Biotechnology, Santa Cruz USA. All other chemicals were of analytical grade and were procured from Indian manufacturers.\n\nAnimals\n-------\n\nMale Sprague Dawley rats (100-150 g body weights) were obtained from the Central Animal House, Panjab University, Chandigarh. The animals were kept in plastic cages under the hygienic conditions and were fed on special diets. Before initiating the experiment, the animals were adapted to the laboratory conditions for a week. Necessary approvals were obtained from the Institutional Ethics Committee. The animal care and handling were done according to the guidelines set by the World Health Organization (WHO), Geneva, Switzerland, and the Indian National Science Academy (INSA), New Delhi, India.\n\nExperimental design\n-------------------\n\nRats were divided into three groups (five animals each group) for the present study. Group I animals served as controls. These animals were fed on Control Diet. Group II animals were fed on High fat Diet (HFD). Group III animals were fed on HFD with 1 ppm selenium supplementation.\n\nDiet Preparation\n----------------\n\nCasein based diets i.e. control diet, HFD was prepared in the laboratory itself according to the composition given by Abraham *et al*\\[[@B59]\\] Table [3](#T3){ref-type=\"table\"}.\n\n###### \n\nComposition of Control and High Fat diet (HFD):\n\n  Component               Control Diet (weight %)   High Fat diet (weight %)\n  ----------------------- ------------------------- -----------------------------\n  Corn Starch             71.0                      61.5\n                                                    \n  Casein                  16.0                      16.0\n                                                    \n  Groundnut Oil           8.0                       0.0\n                                                    \n  Coconut Oil             0.0                       15.0\n                                                    \n  Cholesterol             0.0                       2.0\n                                                    \n  Sodium Cholate          0.0                       0.5\n                                                    \n  Salt Mixture            4.0                       4.0\n                                                    \n  Vitamin mixture         1.0                       1.0\n                                                    \n  Potassium Perchlorate   0.0                       25 mg/100 g B.W./rat/24 hrs\n\nTo the group III animals i.e. HFD + Se group, 1 ppm selenium was given as sodium selenite and was administered along with the high fat diet. The selenium was so chosen that the level is in excess to adequate levels of selenium (0.2 ppm) but well below the sub toxic limits (2.0 ppm).\n\nSelenium levels\n---------------\n\nSelenium levels were estimated in serum by fluorimetric method\\[[@B60]\\]. The assay is based on the principle that Se content in samples on acid digestion is converted to selenous acid. The reaction between selenous acid and aromatic-o-diamines such as 2, 3-diaminonapathalene leads to the formation of 4, 5- benzopiazselenol which displays brilliant lime-green fluorescence when excited at 366 nm in cyclohexane. Fluorescence emission in cyclohexane extract was read on fluorescence spectrophotometer using 366 nm as excitation and 520 nm as emission wavelengths.\n\nLipid profile\n-------------\n\nLipid profile analysis was done in serum using colorimetric kits to estimate the total cholesterol, triglycerides, HDL, LDL levels. Total Cholesterol level was estimated using CHOP-PAP based kit (Human Diagnostic Germany). Triglyceride levels were estimated using GPO based kit (Accurex Biomedical India). HDL and LDL cholesterol levels were estimated using (Fortress direct kit) enzymatic kit.\n\nReactive oxygen species (ROS) levels\n------------------------------------\n\nDetermination of ROS was based on the modified method of Driver *et al*\\[[@B61]\\]. Liver homogenates were prepared in ice-cold Locke\\'s buffer (154 mM NaCl, 5.6 mM KCl, 3.6 mM NaHCO~3~, 2 mM CaCl~2,~10 mM d-glucose and 5 mM HEPES pH 7.4). The homogenates were allowed to warm at 21\u00b0C for 5 min. The reaction mixture containing 10 \u03bcM DCFH-DA and 5 mg tissue/mL was incubated for 15 min at room temperature (21\u00b0C). After another 30 min of incubation, the conversion of DCFH to the fluorescent product 2, 7 dichloroflourescein (DCF) was measured using fluorescence spectrophotometer with excitation at 485 nm and emission at 530 nm. Background fluorescence (Conversion of DCFH-DH to DCF in the absence of homogenate) was corrected by inclusion of parallel blanks. The relative fluorescence intensity was taken as the measure of amount of ROS in different groups.\n\nMeasurement of PON1 activity\n----------------------------\n\nPON1 activity was assessed in serum by measuring the initial rate of Paraoxon hydrolysis to yield p-Nitrophenol at 412 nm at 25\u00b0C. The basal assay mixture included 2 mM Paraoxon, 2 mM CaCl~2~and 0.5 ml serum in 100 mM Tris/Cl buffer. The extinction coefficient for the reaction is 18290 M^-1^cm^-1^. Activity is expressed as \u03bcM Paraoxon hydrolyzed/mg protein/min \\[[@B62]\\].\n\nRNA isolation and mRNA expression of PON1 and PAF-AH using RT-PCR\n-----------------------------------------------------------------\n\nTotal RNA was isolated from fresh liver using Tri-reagent (Molecular research Centre, Inc Ohio, USA) and the quality of isolated RNA was checked on 1.2% agarose gel electrophoresis. For reverse-transcriptase polymerase chain reaction (RT-PCR), primers for PON1, PAF-AH were designed with the aid of software and \u03b2- actin primer was taken from literature. The primer sequence for PON1 was Fav- 5\\'-TGGCATTGGCATTTCCCTTG-3\\', Rev- 5\\'-CAGTAGCTTTCACTCCGGTAA-3\\' and for PAF-AH Fav- 5\\'-CTGATGACAAGACCCTCCGTG-3\\', Rev- 5\\'-CCGTAACCAGTGTGGTCCGGAT-3\\' and for \u03b2-actin Fav- 5\\'-AGAGCTATGAGCTGCCTGAC-3\\', Rev-3\\'-CTGCATCCTGTCAGCCTACG-5\\'. After pilot experiments, it was found that PCR products for PON1, PAF-AH were progressively amplified till 35 cycles and hence 35 amplification cycles were performed. The RT-PCR reaction (Qiagen kit) used a template cDNA followed by PCR amplification with Accu Taq DNA polymerase in the same tube. PCR products were analysed by 1.5% agarose gel electrophoresis. Densitometric analysis from six independent observations was done by Image-J software.\n\nPON1 and PAF-AH protein expression by ELISA\n-------------------------------------------\n\nWells were coated with 5 \u03bcg of sample for PON1 and PAF-AH in 100 \u03bcl of 0.05 M carbonate buffer (pH 9.6) and kept overnight at 4\u00b0C, in a moist chamber. Plates were flicked to remove the unbound antigen solution and wells were blocked with 1% BSA in 0.1 M phosphate buffer saline (pH 7.2) for1 hr at 37\u00b0C. Plates were flicked and wells were washed with 200 \u03bcl of PBS containing 0.05% (v/v) Tween-20. Wells were then incubated with anti-PON1 and anti-PAF-AH respectively, diluted in PBS (containing 0.05% Tween and 1% BSA) and kept for 2 hr at 37\u00b0C. Plates were washed again and incubated with anti-goat secondary antibody (peroxidase labeled) for PON1 (1:1000) and PAF-AH (1:1000) for 2 hr at 37\u00b0C. Wells were washed further thrice as described above and color was developed by addition of 2, 2\\'-azino-di (3-ethyl)-benzothiozolinsulphonic acid reagent and absorbance at 405 nm was measured by ELISA reader.\n\nStatistical Analysis\n--------------------\n\nData is represented as mean \u00b1 S.D. Statistical analysis of the data was performed by student\\'s T-test.\n\nAbbreviations\n=============\n\nA~260~: Absorption at 260 nm; **A**~412~: Absorption at 412 nm; B.W: Body Weight; LCAT: Lecithin Cholesterol Acyl Transferase; LDL: Low Density Lipoproteins; OD: Optical Density; oxLDL: Oxidized Low Density Lipoprotein; PAF: Platelet activating Factor; PAF-AH: Platelet activating Factor-Acetyl Hydrolase; PON1: Paraoxonase1; Se: Selenium; SRB1: Scavenger receptor B (class) 1(type); VLDL: Very Low Density Lipoproteins.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nMPB designed the study and participated in drafting manuscript and result analysis. HDK carried out all the experimental work, participated in statistical analysis and drafting the manuscript under the guidance of MPB. All authors read and approved the final manuscript.\n\nAcknowledgements\n================\n\nAuthors acknowledge the support given by Director: Prof. Tapas Mukhopadhyay (Department of human genomics, Panjab University, Chandigarh, 160014 India) for using various equipments.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nSystemic lupus erythematosus (SLE) is an autoimmune disease characterized by multisystem involvements. The exact etiology of systemic lupus erythematosus (SLE) is still not clear, although multifactorial interaction with environmental and genetic factors has been associated. Immune complex formation along with activation of complement system has been postulated for various manifestations of systemic lupus erythmatosus \\[[@B1]\\].\n\nIt is uncommon in children and young adolescents, with 0.5-0.6 per 1,00,000 per year being the incidence rate in children younger than 15 years of age. Moreover, it is rarer among male subjects \\[[@B2]\\]. Only very few reports have described noncaseating epithelioid cell granulomas in necropsy specimens of lymph node, lung, spleen, and serous membranes \\[[@B3]--[@B5]\\].\n\nAlthough liver dysfunction may be found in SLE (25--50%), but SLE with hepatic and splenic granuloma is very rare. These features may also cause diagnostic confusion with other causes of granulomas. Here, we described a case presented with granulomas in liver and spleen.\n\n2. Case Report {#sec2}\n==============\n\nA 15-year-old male presented with complaints of moderate grade fever for 4 months and generalized body swelling with facial puffiness for 2 months. There was no history of cough, weight loss, joint pain, oral ulcerations, skin rash, photosensitivity, loss of hair, pain abdomen, jaundice, or any significant illness in the past. There was no past history of tuberculosis in the patient or in any other family member. General examination revealed moderate pallor and bilateral cervical and axillary lymph nodes, with the largest being of size 1\u2009cm \u00d7 1\u2009cm mobile, nontender, and firm. Systemic examination was insignificant except for the presence of palpable spleen, 2\u2009cm below the costal margin. Hemogram was suggestive of bicytopenia (Hb 6.9\u2009g/dL (normal: 12--14\u2009g/dL), TLC 2700/mm^3^ (normal: 4000--11000/mm^3^), and total platelet count 2,10,000/mm^3^ (normal: 150000--400000/mm^3^)). Serum amino transferases and alkaline phosphatase were raised (SGOT 316\u2009U/L (normal: 20--40\u2009U/L), SGPT 109\u2009U/L (normal: 20--50\u2009U/L), and S. ALP 2750\u2009IU/L (normal: 44 to 147\u2009IU/L)). Total serum protein (5.1\u2009g/dL, normal: 6.0--8.3\u2009g/dL) and albumin (2.3\u2009g/dL, normal: 3.5--5.5\u2009g/dL) were low. Renal function tests were normal. Mantoux test was negative and ESR was slightly raised. Urine routine and microscopy showed 4-5\u2009RBCs and 8--10 pus cells per high power field and 4+ albuminuria. 24-hour urinary protein was 2.1\u2009gm. Kidney biopsy was done. Chest roentgenogram was normal. Ultrasonography of abdomen revealed mild splenomegaly with a tiny hypoechoic mass in spleen suggestive of splenic granuloma, along with a few subcentimetric retroperitoneal lymph nodes. Bone marrow studies were suggestive of nutritional anemia. Cervical lymph node biopsy showed reactive lymphoid hyperplasia. Barium meal follow through showed jejunoileitis. IgM antibody for Epstein-Barr virus was negative. Serum angiotensin converting enzyme levels were also normal. Based on the clinical picture, the high prevalence of tuberculosis in this part of the world and the investigations, a provisional diagnosis of disseminated tuberculosis was kept. The patient was started on a trial of antituberculous treatment and was asked for review with renal biopsy report. He turned up early with high grade fever, swelling of upper lips, and oral ulceration. On further investigations, his antinuclear and antidouble stranded DNA titers were significantly raised (ANA 7.3\u2009IU/mL, N \\< 1, and anti-dsDNA 680\u2009IU/mL, N \\< 40).\n\nHistopathological report of kidney biopsy was suggestive of membranoproliferative (Type) lupus nephritis. Contrast enhanced computerized tomography of the abdomen revealed hypodense lesions in both liver and spleen (without contrast enhancement), suggestive of granulomas along with few retroperitoneal and mesenteric lymph nodes. Consequently, a revised diagnosis of SLE with lupus nephritis and granulomatous hepatitis was now evident, and the child was put on pulse methylprednisolone 500\u2009mg for 3 days. He showed dramatic improvement clinically.\n\n3. Discussion {#sec3}\n=============\n\nSLE is an autoimmune disease that presents usually with multiorgan manifestations \\[[@B2]\\]. It is uncommon in children and young adolescents, with 0.5-0.6 per 1,00,000 per year being the incidence rate in children younger than 15 years of age. Moreover, it is rarer among male subjects. This coupled with the initial varied and vague presentation of the disease often leads to missing the diagnosis of the disease at first instance. In patients with SLE, subclinical manifestations or biochemical abnormalities of hepatic involvement are usually described, with overt disease seen rarely. Liver disease is diagnosed usually after one year of diagnosis of SLE. Various forms of hepatic involvement have already been described in SLE. Granulomatous hepatitis in SLE is uncommon, although not rare \\[[@B6]\\]. It is usually associated with raised hepatic enzymes, especially alkaline phosphatase levels and appears as hypoechoic/hypodense lesions on CT study. Starting the patient on corticosteroid therapy (as the treatment of SLE) is the main concern as the lesions closely mimic a liver abscess \\[[@B7]\\]. Presence of hepatic and splenic granulomas with generalized lymphadenopathy is well known in tuberculosis and sarcoidosis (Figures [1](#fig1){ref-type=\"fig\"} and [2](#fig2){ref-type=\"fig\"}).\n\nHowever, these were ruled out by relevant investigations. Possibility of an autoimmune disease like SLE was not kept initially despite the presence of significant proteinuria and active urinary sediments because of rarity of the disease in a male child, as described earlier. The only viable option was to place the patient on a trial of antituberculous treatment and look for response. However, failure to improve on ATT combined with the reports of renal biopsy, CECT abdomen, and autoimmune markers made the diagnosis of SLE imperative. Pathogenesis of granuloma formation in systemic lupus erythmatosus (SLE) is still not clear. It was postulated as a response to tissue injury and considered as a manifestation of allergic tissue reaction \\[[@B3]\\]. Although systemic lupus erythmatosus is considered type III hypersensitivity reaction, but type IV mediated mechanism also may play a role in autoimmune nephritis in the pathogenesis granuloma in SLE \\[[@B8]\\]. Reduced numbers of macrophages and capacity for phagocytosis along with defective clearance of apoptotic bodies by the complement system lead to a high level of apoptotic bodies in patients with SLE \\[[@B9]--[@B11]\\]. Hence, persistence of these apoptotic bodies may stimulate granuloma formation.\n\nTherefore, in patients who present with granulomatous disease along with prolonged fever and constitutional symptoms, though uncommon, noninfectious causes like SLE should also be considered as a diagnostic possibility.\n\nConflict of Interests\n=====================\n\nThe authors declare that there is no conflict of interests regarding the publication of this paper.\n\n![Hepatic and splenic granuloma.](CRII2014-737453.001){#fig1}\n\n![Hepatic and splenic granuloma.](CRII2014-737453.002){#fig2}\n\n[^1]: Academic Editor: Maurizio Benucci\n"}
+{"text": "Hypertension is the number one attributable risk factor for death within the general population worldwide ([@B1]) and remains particularly prevalent among individuals with diabetes ([@B2]), despite the broad availability of effective treatment regimens ([@B3]). Among individuals with type 1 diabetes, the presence of hypertension has been associated with a significantly increased risk of both microvascular ([@B4]) and macrovascular ([@B5]) complications, and it also raises overall mortality risk ([@B6]). Given the increased incidence of cardiovascular and kidney complications in this population, the control of arterial blood pressure is of imminent importance, as is the management of risk factors for hypertension incidence itself.\n\nModifiable lifestyle factors, such as obesity and physical inactivity, and dietary factors, including excess alcohol consumption, increased dietary sodium intake, and inadequate fruit, vegetable, and potassium intakes, have been shown to significantly increase the risk of new-onset hypertension in the general population ([@B3],[@B7],[@B8]). Although, traditionally, individuals with type 1 diabetes were thought to be of normal or subnormal weight, the adoption of unhealthy lifestyle behaviors and/or intensive insulin therapy have led to an increasing prevalence of overweight and obesity in individuals with this diabetes type ([@B9]). Moreover, the presence of hyperglycemia has been suggested to further contribute to the excess risk of hypertension in these individuals by promoting vascular stiffness ([@B10]). Indeed, analyses of the Diabetes Control and Complications Trial (DCCT) and its observational follow-up study, Epidemiology of Diabetes Intervention and Complications (EDIC), demonstrated that hyperglycemia and intensive insulin therapy are associated with incident hypertension ([@B11]), although sex differences were not evaluated. Differences in the incidence of and/or risk factors for vascular complications associated with hypertension (i.e., kidney and heart disease) have been previously described by our group among individuals with type 1 diabetes ([@B12]--[@B14]). We therefore aimed to assess the association between glycemia (HbA~1c~), glycemic control (intensive therapy), and the development of hypertension in a well-characterized cohort study of individuals with childhood-onset type 1 diabetes, to confirm whether findings from the DCCT/EDIC study are apparent in the general type 1 diabetes population, and to determine whether any association between glycemic control or intensive insulin treatment with incident hypertension varies by sex.\n\nRESEARCH DESIGN AND METHODS {#s1}\n===========================\n\nParticipants from the Pittsburgh Epidemiology of Diabetes Complications (EDC) study with arterial blood pressure \\<140/90 at study initiation were selected for study (*n* = 510). The EDC is a historical cohort study based on incident cases of childhood-onset (prior to their 17th birthday) type 1 diabetes, diagnosed or seen within 1 year of diagnosis (1950--1980) at the Children's Hospital of Pittsburgh ([@B15]). This cohort has been previously shown to be epidemiologically representative of the type 1 diabetes population of Allegheny County, Pennsylvania ([@B16]). The first clinical assessment for the EDC study took place between 1986 and 1988, when the mean participant age and diabetes duration were 28 and 19 years, respectively. Subsequently, biennial examinations were conducted for 10 years, with a further examination at 18 years of follow-up. The University of Pittsburgh institutional review board approved the study protocol.\n\nPrior to each clinic visit, participants were sent questionnaires concerning demographic, health care, diabetes self-care, and medical history information. Leisure time physical activity was assessed by the Paffenbarger questionnaire ([@B17]), and a previously published algorithm ([@B18]) was used to calculate physical activity over the past week and over the past year based on the daily number of city block equivalents walked, the number of flights of stairs climbed, and the frequency and duration of leisure time activity. Three blood pressure measurements were taken by trained and certified personnel with a Hawksley random zero sphygmomanometer, after a 5-min rest in the sitting position according to the Hypertension Detection and Follow-up Program protocol ([@B19]). Hypertension was defined as \u2265140/90 mmHg (mean of the second and third readings) or use of antihypertensive medications. For the present analyses, participants were only considered \"hypertensive\" if they were positive on two consecutive examination cycles. Intensive insulin therapy was defined as multiple (three or more) daily insulin injections or continuous subcutaneous insulin infusion in addition to frequent (at least four times daily) glucose testing. Stable glycosylated hemoglobin (HbA~1~) was measured by ion-exchange chromatography (Isolab, Akron, OH) for the first 18 months, and the subsequent 10 years by automated high-performance liquid chromatography (Diamat; Bio-Rad, Hercules, CA); the two assays were highly correlated (*r* = 0.95). For follow-up beyond the 10 years, HbA~1c~ was measured with the DCA 2000 analyzer (Bayer, Tarrytown, NY). The DCA and Diamat assays were also highly correlated (*r* = 0.95). Original HbA~1~ (1986--1998) and HbA~1c~ values (1998--2004) were converted to DCCT-aligned standard HbA~1c~ values using regression formulae derived from duplicate assays (DCCT HbA~1c~ = \\[0.83 \u00d7 Diamat HbA~1~\\] + 0.14 and DCCT HbA~1c~ = \\[DCA HbA~1c~ \u2212 1.13\\]/0.81). HDL cholesterol was determined enzymatically after precipitation with heparin and manganese chloride, with a modification ([@B20]) of the Lipid Research Clinics method ([@B21]). Cholesterol and triglycerides were measured enzymatically ([@B22],[@B23]). Non-HDL cholesterol was calculated as total minus HDL cholesterol. White blood cell (WBC) count was obtained using a counter S-plus IV and fibrinogen using a biuret colorimetric procedure and a clotting method. Urinary albumin was measured by immunonephelometry ([@B24]), and creatinine was assayed by an Ectachem 400 Analyzer (Eastman Kodak Co., Rochester, NY). Glomerular filtration rate was estimated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation ([@B25]). It should, however, be noted that serum creatinine was not calibrated in this study. All assays were conducted during the cycle that samples were obtained, and thus, prolonged storage would not have affected measurements performed in this study.\n\nStatistical analysis {#s2}\n--------------------\n\nAnalyses were conducted stratified by sex. Univariate associations were determined using the Student *t* test for normally distributed continuous variables or Wilcoxon two-sample test for nonnormally distributed continuous variables. The \u03c7^2^ or Fisher exact test, as appropriate, was used for univariate analysis of categorical variables. Cox proportional hazards models with backward elimination were constructed to assess independent predictors of hypertension incidence among traditional risk factors and univariately significant variables. Cox proportional hazards models were also constructed using time-dependent updated means of variables significant in models using baseline characteristics. Survival time was defined as the time in years from study entry to either incident hypertension or censorship during the 18-year follow-up. Nonnormally distributed variables were logarithmically transformed for entry into multivariable models. Statistical analyses were conducted using Statistical Analysis Software (SAS), version 9.2 (SAS Institute, Cary, NC).\n\nRESULTS {#s3}\n=======\n\nIncidence and univariate predictors of hypertension in the entire cohort {#s4}\n------------------------------------------------------------------------\n\nDuring 18 years of follow-up, 39.2% (*n* = 200) of individuals developed incident hypertension, for an incidence rate of 31.2 per 1,000 person-years. Incidence was slightly lower in women (35.4%) than in men (43.2%, *P* = 0.07). [Table 1](#T1){ref-type=\"table\"} presents characteristics of male and female participants at study entry by incident hypertension. In both sexes, compared with participants whose arterial blood pressure remained within a normal range, those who subsequently developed hypertension were more likely to be older and have elevated baseline blood pressure, non-HDL cholesterol, albumin excretion rate (AER), and WBC count. In addition, male participants who developed hypertension were more likely to have a longer duration of type 1 diabetes, larger waist-to-hip ratio (WHR), and elevated HbA~1c~ and fibrinogen levels at study entry. Among women, those who subsequently developed hypertension had lower HDL cholesterol concentrations at study entry compared with women who maintained normal arterial blood pressure. No participant received angiotensin-converting enzyme inhibitors/angiotensin receptor blockers at study baseline.\n\n###### \n\nParticipant characteristics at study entry by incidence of hypertension during 18 years of follow-up\n\n![](77tbl1)\n\nThe effect of glycemic control on the development of hypertension {#s5}\n-----------------------------------------------------------------\n\nTo evaluate whether a dissimilar distribution of HbA~1c~ between male and female participants at study entry was responsible for the observed discrepancy in its association with hypertension incidence by sex, we assessed incidence by HbA~1c~ tertiles. As shown in [Fig. 1*A*](#F1){ref-type=\"fig\"}, the proportion of incident cases was similar by sex within the first HbA~1c~ tertile but appeared to be slightly increased among men compared with women within the second and third tertiles. Moreover, although hypertension incidence appeared to increase linearly with increasing HbA~1c~ tertile among men (*P* value for trend = 0.0008), a similar increase in risk was not apparent among women (*P* value for trend = 0.31). [Figure 1*B* and *C*](#F1){ref-type=\"fig\"} depicts the diabetes duration--adjusted 18-year survival free of hypertension for men and women, respectively, by tertiles of HbA~1c~ at study entry. These graphs clearly show a strong association between HbA~1c~ and hypertension incidence among men but a much weaker, nonsignificant relationship among women. Thus, compared with men with HbA~1c~ \\<8%, the estimated relative hazard of developing hypertension was 2.42 (95% CI 1.43--4.10) times higher among those with HbA~1c~ between 8 and 9.2% and four (2.35--6.80) times higher among men whose levels were \\>9.2%. Among women, however, even levels \\>9.2% were only associated with a borderline significantly increased hazard compared with HbA~1c~ levels \\<8% (*P* = 0.07).\n\n###### \n\n*A*: Incidence of hypertension by sex and tertiles of HbA~1c~ at study entry. Among men, *P* value = 0.004 and *P* value for trend = 0.0008. Among women, *P* value = 0.47 and *P* value for trend = 0.31. *B*: Diabetes duration--adjusted survival curves for hypertension by tertiles of HbA~1c~ at study entry among men. The Pittsburgh EDC study (hazard ratio~8.0\\ to\\ \\<9.2~ 2.42 \\[95% CI 1.43--4.10\\], *P* value = 0.0009; hazard ratio [~\\>~]{.ul}~9.2~ 4.00 \\[2.35--6.80\\], *P* value \\<0.0001; log-rank *P* value \\<0.0001). *C*: Diabetes duration--adjusted survival curves for hypertension by tertiles of HbA~1c~ at study entry among women. The Pittsburgh EDC study (hazard ratio ~8.0--\\<9.2~ 1.18 \\[0.70--1.98\\], *P* value = 0.54; HR[~\\>~]{.ul}~9.2~ 1.58 \\[0.97--2.58\\], *P* value = 0.07; log-rank *P* value = 0.22).\n\n![](77fig1-1)\n\n![](77fig1-2)\n\nThe presence of effect modification of HbA~1c~ by sex was confirmed with the results of a significant interaction term (*P* = 0.009) in a Cox proportional hazards model that included variables for sex, HbA~1c~, and the interaction term. Thus, separate multivariable models (with backward elimination) were constructed for male and female participants. Since only 4 men and 21 women followed an intensive insulin therapy protocol at study entry, this variable was not included in multivariable models of baseline predictors. Among men ([Table 2](#T2){ref-type=\"table\"}), elevated HbA~1c~, systolic blood pressure (SBP), and AER predicted hypertension development. In women, elevated SBP and AER were also predictors, along with non-HDL cholesterol. HbA~1c~, however, was not an independent predictor in women.\n\n###### \n\nCox proportional hazards models for the prediction of hypertension among male and female participants during 18 years of follow-up in the Pittsburgh EDC study\n\n![](77tbl2)\n\nTo better understand the modification of the effect of HbA~1c~ by sex, we also performed separate Cox proportional hazards models by HbA~1c~ tertile ([Table 3](#T3){ref-type=\"table\"}). Interestingly, these analyses revealed the presence of an over twofold significantly increased risk for hypertension incidence in women compared with men among those in the lowest HbA~1c~ tertile (hazard ratio 2.24 \\[95% CI 1.17--4.27\\]). No difference in risk by sex was noted in the second tertile of HbA~1c~ (1.30 \\[0.77--2.18\\]), whereas a significantly lower risk for women compared with men was observed in the third HbA~1c~ tertile (0.53 \\[0.32--0.86\\]). These findings could not be attributed to sex differences in the distribution of HbA~1c~ within each tertile (the *P* value for a difference by sex was 0.81, 0.60, and 0.47 for the first, second, and third tertile, respectively).\n\n###### \n\nCox proportional hazards models for hypertension incidence stratified by tertile of HbA~1c~ at study entry\n\n![](77tbl3)\n\nA significant interaction between sex and HbA~1c~ (*P* \\< 0.0001) was also noted when analyses were repeated using time-dependent updated means of baseline independent predictors of hypertension incidence. Thus, although a significant effect of HbA~1c~ was seen even among female participants (hazard ratio 1.21 \\[95% CI 1.00--1.46\\]), the effect of glycemic control on hypertension development was stronger in men (1.59 \\[1.29--1.97\\]) (data not shown). As intensive insulin therapy became more prevalent past study entry, this variable was allowed for in time-dependent models; however, intensive insulin therapy was not associated with incident hypertension in either sex. Although surprising, this finding is likely attributed to the fact that \\>60% of incident hypertension cases had developed by the sixth examination cycle (1996--1998), whereas intensive insulin therapy appears to have been largely adopted after that time period in this population.\n\nCONCLUSIONS {#s6}\n===========\n\nIn this cohort of individuals with long-standing type 1 diabetes, the effect of hyperglycemia, as measured by greater HbA~1c~ levels at study entry, on new-onset hypertension was much stronger in men compared with women. The very small number of men and women following intensive insulin therapy at study entry (4 and 21, respectively) hindered assessment of such practice in multivariable models. When included in time-dependent analyses, intensive insulin therapy did not predict hypertension in either sex, potentially because such therapy was largely adopted in the later examination cycles, by which time hypertension would have developed in a large proportion of participants.\n\nSimilar findings of a more prominent role of hyperglycemia in the incidence of hypertension in male compared with female participants were also noted when mean HbA~1c~ levels throughout follow-up were considered. Moreover, the observed protective effect of insulin dose per kilogram body weight against hypertension incidence was also restricted to men in these analyses. The weaker relationship between glycemia and hypertension in women did not appear to be explained by a different distribution of HbA~1c~ by sex.\n\nIn the current analyses, hypertension incidence was slightly higher in men than in women. Male sex per se was also associated with increased hypertension risk in the DCCT/EDIC study of individuals with type 1 diabetes ([@B11]), as well as in the biracial cohort of the Coronary Artery Risk Development in (Young) Adults (CARDIA) study ([@B26]) and the Framingham Heart Study ([@B27]), among many others in the general population. Although not all reports concur that risk is greater among men, blood pressure has generally been reported to be higher in men than in women within the general population ([@B28]), with women exhibiting a lower risk for hypertension especially in the years prior to menopause, whereas risk becomes comparable between age-matched men and postmenopausal women ([@B29]). This sex dimorphism and the observation of an important role of menopause have led to the hypothesis that sex hormones may act as modulators of vascular function and the pathogenesis of hypertension ([@B30]). Interestingly, a previous report also suggested that sex hormone--binding globulin and total testosterone are higher in male, but not female, children and young adults with diabetes compared with nondiabetic siblings, a finding apparently related to the absence of endogenous insulin ([@B31]). Whether such differences may account for the differential association between HbA~1c~ and hypertension incidence by sex could, unfortunately, not be evaluated, as hormone data are not available for EDC study participants. Moreover, as the associations reported in this manuscript were derived from a population where the majority of women are premenopausal, any conclusions made should be restricted to younger and middle-aged adults, as deviations from these relationships may be noted with longer follow-up, when more women would have reached menopause.\n\nIt has been suggested that the presence of hyperglycemia contributes to the excess risk of hypertension in individuals with diabetes by promoting vascular stiffness ([@B10]). Indeed, considerable evidence links hyperglycemia with increased flux through the polyol pathway and the reduction of glucose to sorbitol, increased formation of advanced glycosylation end products, and, importantly, generation of reactive oxygen species, which lead to vascular damage ([@B10]). Interestingly, our results suggest that the incidence of new-onset hypertension is higher in women compared with men among participants at better glycemic control but higher in men compared with women among those at worse glycemic control. The reason why glycemic stress may differentially affect a person's risk for developing hypertension based on their sex is currently unclear. As HbA~1c~ has improved over follow-up, it is possible that the impact of male predominance for hypertension at high HbA~1c~ has diminished in time, in a similar manner to what we have reported for renal disease ([@B13]) where the elimination of the male excess in advanced renal disease in the more recently diagnosed cohort may be potentially linked to improved glycemic control. A large body of literature has provided evidence that in addition to structural differences, developmental/environmental stressors may provoke a distinctive physiological response by sex ([@B32]). Thus, as here, although only a weak difference in hypertension incidence exists by sex, the pathogenesis of blood pressure elevation is likely to differ between men and women. This is consistent with our earlier observation of similar coronary artery disease incidences between men and women despite differences in risk factors by sex ([@B12]) and our recent report that HDL cholesterol shows a different relationship to coronary artery disease in women (U shaped) compared with men (inverse linear) in the EDC study, potentially explaining some of the loss of female protection against heart disease seen in type 1 diabetes ([@B14]).\n\nDespite previously reported associations between BMI and hypertension incidence in the general population ([@B26],[@B27]) as well as in type 1 diabetes ([@B11]), measures of body fatness (WHR and, in separate models, BMI) were not selected in the final prediction models in this study. However, AER, as measured at baseline and also as a time-dependent updated mean over the follow-up period, was significantly associated with increased risk for hypertension in both sexes. The categorization of study participants into normo- or microalbuminuric at the baseline assessment was also associated with increased risk of hypertension in both men (almost fourfold increased risk) and women (twofold increased risk). AER was strongly associated with the incidence of hypertension also in the report from the DCCT/EDIC study ([@B11]). A direct association between plasma lipid concentrations and hypertension incidence was also observed among women, although similarly strong associations were not seen among men.\n\nIn conclusion, the results of this study suggest that although hyperglycemia contributes to the development of hypertension, this effect is much stronger in men compared with women with type 1 diabetes. The reasons for this difference are not clear but merit further investigation as sex differences appear to be common in the natural history of type 1 diabetes complications, including coronary artery disease and kidney disease, both of which are closely related to hypertension.\n\nThis research was supported by National Institutes of Health grants DK-34818 and DK-082900.\n\nNo potential conflicts of interest relevant to this article were reported.\n\nT.C. researched and analyzed data and wrote the manuscript. T.J.O. researched data, contributed to the discussion, and reviewed and edited the manuscript. All authors have read and approved the final version of the manuscript. T.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.\n\nPreliminary analyses of this study were presented at the Scientific Sessions of the American Heart Association, Orlando, Florida, 14--18 November 2009.\n\nThe authors are indebted to the participants of the Pittsburgh EDC study who have tirelessly volunteered their time for more than 20 years.\n"}
+{"text": "The data is available at <https://www.dhsprogram.com/data/available-datasets.cfm>.\n\nIntroduction {#sec005}\n============\n\nThe health of children requires much attention since occurrence of morbidity during this tender age could have lifelong effects on their growth and development \\[[@pone.0221324.ref001], [@pone.0221324.ref002]\\]. To improve the health of children, it is necessary for health planners to effectively comprehend the spatial homogeneity and heterogeneity of top rated childhood morbidities. Children in developing countries, places with low socio-economic indicators or poor environmental conditions, commonly suffer from diarrhoea \\[[@pone.0221324.ref003]\\], respiratory infections \\[[@pone.0221324.ref004]\\], anaemia \\[[@pone.0221324.ref005]\\] and fever \\[[@pone.0221324.ref006]\\]. Diarrhoea is one of the major causes of mortality in children under five years old, and it accounts for about 760,000 mortality cases every year \\[[@pone.0221324.ref007]\\]. Approximately, four to five million children each year in developing countries die due to acute respiratory infections (ARI) \\[[@pone.0221324.ref008]\\]. Also, prevalence of anaemia was estimated to be 65 percent in Africa (in 2008) \\[[@pone.0221324.ref009]\\], and fever is a common symptom of most childhood morbidities \\[[@pone.0221324.ref010]\\].\n\nIn Ghana, diarrhoea, ARI, anaemia, and fever are also the leading causes of morbidity and mortality among children \\[[@pone.0221324.ref011]\\]. The prevalence of these childhood morbidities among under 5 years showed that 12 percent had all forms of diarrhoea, four percent had ARI symptoms, 66 percent (of children 6--59 months) were anaemic and 14 percent (of children under five) had fever in 2014 \\[[@pone.0221324.ref012]\\].\n\nElsewhere, empirical evidences indicate that studies have been conducted to establish hot spots of morbidities among children \\[[@pone.0221324.ref013], [@pone.0221324.ref014]\\]. Similarly in Ghana, a number of authors have examined spatio-temporal patterns of morbidities using facility-based data \\[[@pone.0221324.ref015], [@pone.0221324.ref016], [@pone.0221324.ref017], [@pone.0221324.ref018]\\]. All these studies commonly examined a single morbidity such as diarrhoea among children. In addition, a number of studies either selected towns or a small number of districts to spatially assess morbidities. The usage of more layers (thus all districts) in our study represents an increase in scale compared with the limited studies on spatial analysis of childhood morbidities in the country. We examined spatial clustering and hot spots of four most reported morbidities among infants and young children in Ghana (aged 0--23 months of age). Findings of the study may inform related stakeholders (such as Ghana Health Services and health related non-governmental organisations) on how to programme and allocate their resources effectively. The study, therefore, sought to geospatially analyse concurrently four the morbidities among children using five datasets (2014, 2008, 2003, 1998, 1993) from the Ghana Demographic and Health Surveys (GDHS).\n\nMethods {#sec006}\n=======\n\nStudy setting {#sec007}\n-------------\n\nGhana lies between latitudes 4\u00b045'N and 11\u00b0N, and longitudes 1\u00b015'E and 3\u00b015'W. The country is located in West Africa with a total land area of 238,537 square kilometres; and bordered by three countries: to the east by Togo; Cote d'Ivoire to the west; Burkina Faso to the north and south by the Gulf of Guinea ([Fig 1](#pone.0221324.g001){ref-type=\"fig\"}). The population of the country as recorded in the 2010 Population and Housing Census was 24,658,823 million people with an average annual growth of about 2.5 percent. Ghana is divided into ten main administrative regions with 216 sub-units (metropolis/municipal and districts) \\[[@pone.0221324.ref012]\\].\n\n![Map of Africa projected to show map of Ghana.\\\nGhana is located in the Western portion of Africa. The country is boarded to the: north by Burkina-Faso; west by Cote D'Ivoire; east by Togo; and south by Gulf of Guinea.](pone.0221324.g001){#pone.0221324.g001}\n\nData source {#sec008}\n-----------\n\nThe study used five different GDHS datasets (2014, 2008, 2003, 1998, 1993). These data were acquired from Demographic and Health Survey (DHS) programme website (<https://dhsprogram.com/>). The GDHS is a national-level population and health survey conducted in Ghana as part of the Global Demographic and Health Survey programme; and the surveys are conducted in the country in partnership with Ghana Statistical Service and Ghana Health Service. STATA data files (with individual rows containing health related details on women \\[15--49 years\\] including their children), and shapefiles for the location of clusters visited (during the surveys) were specifically used for this study. The data is available at <https://www.dhsprogram.com/data/available-datasets.cfm>. For each of the surveys conducted in the country, approval was sought from Ghana Health Service Research Ethical Committee.\n\nSampling technique {#sec009}\n------------------\n\nThe GDHS used a two-stage stratified nationally representative sample of households \\[[@pone.0221324.ref012]\\]. All the surveys usually used the previous sample points (clusters) comprising of enumeration areas of national population censuses as the sampling frame to produce separate estimates for key indicators for the ten administrative regions in the country. Sample points (clusters) were situated within districts, and the primary sampling units consisted both urban and rural clusters. The secondary sampling units included households. Households were listed in all enumerated areas and they were randomly selected from the list. A number of households were randomly selected from each cluster (found within districts) and different sample sizes were obtained at the end of each sample design for each survey period. More details on the sampling techniques used for each survey can be obtained from the various survey report documents at <https://dhsprogram.com/Publications/Publications-by-Country.cfm>.\n\nData collection {#sec010}\n---------------\n\nIn collecting the household data, enumerators also mapped the clusters in which they interviewed respondents. The clusters were mapped, rather than the house of the respondent, to protect actual identity and location of respondents. However, the mapped spatial clusters have primary fields which enable easy data merging with household records. Spatial data of the boundaries of 216 administrative sub-units were solicited from the Ghana Survey Department to enable the analysis to be made at a district level. We analysed at the district level because this level is more informative to policy makers rather than analysing at the cluster level.\n\nData preparation {#sec011}\n----------------\n\nThe study used GDHS mapped clusters, and 216 district administrative layers. The 216 administrative layers had their geographic coordinates as the Ghana Meter Grid (projected coordinate system) while clusters mapped from GDHS was in the World Geodetic System 1984 (geographic coordinate system). The GDHS spatial data was re-projected into the projected coordinate system of Ghana Meter Grid. The two spatial datasets were then later converted from shapefile to ERSI geodatabase format; this was done using ArcGIS 10.3. Then a spatial join was undertaken to transfer cluster names to the sub-administrative polygon layer. Some of the extracted sub-administrative units had more than one cluster. Such clusters were aggregated and their means were used to represent the respective district they fell within.\n\nFor the prevalence of childhood morbidity (diarrhoea, ARI, anaemia, and fever), data was extracted from the individual datasets from the various surveys. The extraction was restricted to children less than two years who were living with their mothers during the survey periods. This category of children is deemed to be more susceptible to episodes of morbidities compared to children more than two years. Mothers were asked whether any of their children under five years of age had diarrhoea during the two weeks preceding the survey. ARI was estimated by asking mothers whether their children under age five had been ill with a cough accompanied by short rapid breathing in the two weeks preceding the survey. For anaemia, children who stayed in the household on the night before the survey were tested for anaemia based on their haemoglobin levels. With fever, mothers were asked whether a child has been ill with fever at any time in the last two weeks preceding the survey. All the survey years contained the required variables of interest except surveys of 1993, and 1988 that did not have anaemia data. Extracted data on child morbidity were spatially joined with the sub-administrative unit polygon to generate geo-relational data to enable spatial analysis. Stata version 12 was used to process and extract the childhood morbidity.\n\nStatistical analyses {#sec012}\n--------------------\n\nLogistic regression was used to examine association between childhood morbidity and place of residence by adjusting for the effect of other covariates (place of residence, education, working status, and drinking water. These variables can potentially influence health outcomes in children. This was done to further explain the occurrences of childhood morbidity in the districts found to have significant geographic clusters. The place of residence was considered as the main independent variable since the GDHS data were mainly collected within clusters found within districts that could be either be described as an urban area or a rural area. The other covariates were selected based on their possible effects across all the childhood morbidities. For each morbidity, two models (unadjusted and adjusted) were generated. Results were presented as odds ratio and 95 percent confidence intervals.\n\nFor the geospatial analyses, three spatial statistical tools were applied to analyze the data. These tools are: spatial autocorrelation (Global Moran's *I*); cluster and outlier analysis (Anselin's local Moran's *I*); and hot spot analysis (Getis-Ord G). Spatial autocorrelation (spatial association) measures clustering or dispersion based on feature geographical locations and attribute values of a single variable. It is used to calculate correlation among neighbouring observations and to ascertain patterns and levels of spatial clustering in neighbouring districts \\[[@pone.0221324.ref019], [@pone.0221324.ref020]\\]. It computes a single summary value, a z-score (of -1 to +1), describing the degree of spatial concentration or dispersion for a measured variable \\[[@pone.0221324.ref021]\\].\n\nThe spatial autocorrelation was, therefore, used to examine whether childhood morbidity (diarrhoea, ARI, anaemia, and fever) had a clustering or dispersion pattern in the country using districts as features. The study hypothesized that the prevalence of study phenomena (childhood morbidity) are randomly distributed (no spatial dependency) across various districts in the country. The null hypothesis is rejected if a calculated *p*-value is very small (95% confidence interval); which implies an unlikely situation that observed spatial pattern is as a result of random processes. Hence, a district with a high *z*-score and a small *p*-value indicates a spatial clustering of high values.\n\nTo locate where childhood morbidity was clustered, the ArcGIS clusters and outlier analysis tool was applied. Cluster and outlier tool measures spatial autocorrelation based on both area locations and area values simultaneously. Given a set of areas and an associated attribute, it evaluates whether the pattern expressed is clustered, dispersed, or randomized \\[[@pone.0221324.ref022]\\].\n\nThe cluster and outlier analysis (Anselin Local Moran's *I*) was used to ascertain geographic composition of childhood morbidity clusters and outliers. This statistic calculates a *z*-score, a *p*-value, and a code representing each cluster type into four statistical significant outcomes. High positive z-score for a district means that its surrounding districts have similar values, High-High (districts showing high levels of morbidity surrounded by districts with similar high levels) for a statistically significant (0.05 level) cluster of high values and Low-Low (districts showing low levels of morbidity surrounded by districts with similar low levels) for a statistically significant (0.05 level) cluster of low values. Low negative z-score for a district means a statistically significant (0.05 level) spatial outlier, High-Low for a district a high value surrounded by low values Low-High for low value surrounded by high values \\[[@pone.0221324.ref022]\\]. The Moran's *I* identifies clusters of features with values similar in magnitude. Moran's *I* values range from -1 (disperse) to +1 (clustered). A calculated Moran's *I* value of 0 indicates complete spatial randomness. This study used *p* \\< 0.05 to determine statistical significance of the computed index values.\n\nFurther, hot spot analysis measures each area in a dataset within the context of neighbouring areas in the same dataset. It uses vectors to identify areas of statistically significant hot spots and cold spots in data. Also, hot spot analysis assumes that there is the presence of clustering within the data \\[[@pone.0221324.ref023]\\]. Hot spot analysis is used to identify statistically significant spatial clusters of high values (hot spots) and low values (cold spots). This analysis was used to define districts with high prevalence versus districts of low prevalence of childhood morbidity. The description of a district as being a hot spot of a childhood morbidity was expressed using 99%, 95% and 90% confidence levels (CL). The usage of these informs the reader about the prominence of hot spots. The CL indicates the significance of occurrence of a morbidity in a district. Only districts identified as hotspots at the 99% confidence level (high-lighted table in the results section) and then while all hot spots were represented on the maps.\n\nThe three analyses all adopted Inverse Distance Method (IDM) with the notion that near incidents are more related than those afar. However, the False Discovery Rate Correction (FDRC) was not applied. This was to allow the analysis to be conducted based on individual district first in isolation.\n\nResults {#sec013}\n=======\n\n[Table 1](#pone.0221324.t001){ref-type=\"table\"} shows the association between childhood morbidity and selected covariates after controlling some other socio-demographic factors. Children in rural areas (OR = 1.32, 95% CI = 1.25--1.52) were more likely to have had experienced diarrhoea compared to children in urban areas. After adjusting for other factors, the odds of children in rural areas to experience episodes of diarrhoea was still positive significant (OR = 1.19, 95% CI = 1.39). In addition, children with mothers who attained secondary or higher education (OR = 0.75, 95% CI = 0.64--0.88) were less likely to report occurrence of diarrhoea compared to those whose mothers had no formal education. Likewise, mothers who had access to unimproved sources of drinking water (OR = 0.82, 95% CI = 0.71--0.96) reported less occurrence of childhood diarrhoea compared to those who has access to improved sources of drinking water. For ARI, no significant associations were found with the covariates.\n\n10.1371/journal.pone.0221324.t001\n\n###### Association between childhood morbidity and covariates.\n\n![](pone.0221324.t001){#pone.0221324.t001g}\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Covariates               Diarrhoea                                          ARI                                                Anaemia          Fever                                                                                                                                   \n  ------------------------ -------------------------------------------------- -------------------------------------------------- ---------------- ---------------- ---------------- -------------------------------------------------- -------------------------------------------------- --------------------------------------------------\n  **Place of Residence**                                                                                                                                                                                                                                                                  \n\n  \u2003Urban                   Ref.                                               Ref.                                               Ref.             Ref.             Ref.             Ref.                                               Ref.                                               Ref.\n\n  \u2003Rural                   1.32[\\*](#t001fn002){ref-type=\"table-fn\"}(0.00)\\   1.19[\\*](#t001fn002){ref-type=\"table-fn\"}(0.03)\\   0.99(0.89)\\      0.94(0.42)\\      1.41(0.00)\\      1.37[\\*](#t001fn002){ref-type=\"table-fn\"}(0.00)\\   1.29[\\*](#t001fn002){ref-type=\"table-fn\"}(0.00)\\   1.16[\\*](#t001fn002){ref-type=\"table-fn\"}(0.05)\\\n                           \\[1.15--1.52\\]                                     \\[1.02--1.39\\]                                     \\[0.87--1.13\\]   \\[0.82--1.09\\]   \\[1.23--1.61\\]   \\[1.17--1.59\\]                                     \\[1.22--1.48\\]                                     \\[0.99--1.36\\]\n\n  **Education**                                                                                                                                                                                                                                                                           \n\n  No education                                                                Ref.                                                                Ref.                              Ref.                                                                                                  Ref.\n\n  \u2003Primary                                                                    0.95(0.59)\\                                                         1.11(0.26)\\                       1.03(0.76)\\                                                                                           0.92(0.35)\\\n                                                                              \\[0.797--1.14\\]                                                     \\[0.93--1.32\\]                    \\[0.85--1.24\\]                                                                                        \\[0.76--1.09\\]\n\n  \u2003Secondary/Higher                                                           0.75[\\*](#t001fn002){ref-type=\"table-fn\"}(0.00)\\                    1.01(0.94)\\                       0.88(0.12)\\                                                                                           0.83(0.02)\\\n                                                                              \\[0.64--0.88\\]                                                      \\[0.86--1.17\\]                    \\[0.74--1.04\\]                                                                                        \\[0.71--0.97\\]\n\n  **Working status**                                                                                                                                                                                                                                                                      \n\n  Not working                                                                 Ref.                                                                Ref.                              Ref.                                                                                                  Ref.\n\n  Working                                                                     1.19(0.06)\\                                                         1.13(0.15)\\                       1.04(0.68)\\                                                                                           1.47[\\*](#t001fn002){ref-type=\"table-fn\"}(0.00)\\\n                                                                              \\[0.71--0.96\\]                                                      \\[0.96--1.33\\]                    \\[0.86--1.25\\]                                                                                        \\[1.22--1.77\\]\n\n  **Drinking water**                                                                                                                                                                                                                                                                      \n\n  Improved                                                                    Ref.                                                                Ref.                              Ref.                                                                                                  Ref.\n\n  Unimproved                                                                  0.82[\\*](#t001fn002){ref-type=\"table-fn\"}(0.01)\\                    1.02(0.83)\\                       0.96(0.57)\\                                                                                           0.95(0.47)\\\n                                                                              \\[0.71--0.96\\]                                                      \\[0.88--1.17\\]                    \\[0.82--1.12\\]                                                                                        \\[0.81--1.09\\]\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nM-model; ARI-acute respiratory infection; Ref.-reference group;\n\n\\*-significant at p\\<0.05\n\np-values are in round brackets, and 95% confidence interval ranges are in square brackets\n\nChildren in rural areas (OR = 1.37, 95% CI = 1.17--1.59) were more likely to be anaemic compared to their counterparts in urban areas ([Table 1](#pone.0221324.t001){ref-type=\"table\"}). This significance was still there after holding other factors constant. More feverish children were likely to be found in rural areas (OR = 1.29, 95% CI = 1.22--1.48) compared to urban areas. After adjusting for other covariates, children in rural areas were still more likely to experience fever compared to those in urban areas. Also, children whose mothers were working (OR = 1.47, 95% CI = 1.22--1.77) at the time of survey were likely to have more episodes of fever compared to those whose mothers were not working.\n\nSpatial autocorrelation of childhood morbidity {#sec014}\n----------------------------------------------\n\nThe study revealed positive spatial autocorrelation for childhood morbidity in the country ([Table 2](#pone.0221324.t002){ref-type=\"table\"}). In 2014, there were clustering of similar values for diarrhoea (Moran's *I* = 0.096; *p* = 0.001), ARI (Moran's *I* = 0.073; *p* = 0.004) and anaemia (Moran's *I* = 0.056; *p* = 0.024), suggesting that there is less than 1 percent possibility that these clustered patterns could be attributed to random chance. In other words, the output was significant indicating clustering of acute respiratory infections at the district level. Also meaning, there were areas of hotspots in data. In 2008 and 2003, for diarrhoea, given the Moran's *I* of 0.078 (*p* = 0.007) and 0.063 (*p* = 0.019) respectively, the patterns appeared to be significantly significant thus cases of diarrhoea were not randomly distributed among the districts. The clustered patterns of diarrhoea, ARI, and fever were statistically significant in 1998 and 1993; thus suggesting that the clustered patterns were not the result of random chance.\n\n10.1371/journal.pone.0221324.t002\n\n###### Global Moran's *I* analysis of childhood morbidity (1993--2014).\n\n![](pone.0221324.t002){#pone.0221324.t002g}\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------\n  Morbidity   Year       Moran's\\    Expected\\   Variance   *z*-score                                       *p*-value\n                         Index       Index                                                                  \n  ----------- ---------- ----------- ----------- ---------- ----------------------------------------------- -----------------------------------------------\n  Diarrhoea   2014       0.096479    -0.004651   0.000728   3.748256                                        0.000178[\\*](#t002fn002){ref-type=\"table-fn\"}\n\n  ARI         0.072805   -0.004651   0.000726    2.874740   0.004044[\\*](#t002fn002){ref-type=\"table-fn\"}   \n\n  Anaemia     0.056534   -0.004651   0.000737    2.253649   0.024218[\\*](#t002fn002){ref-type=\"table-fn\"}   \n\n  Fever       0.028028   -0.004651   0.000724    1.214795   0.224444                                        \n\n  Diarrhoea   2008       0.078913    -0.006024   0.001004   2.681021                                        0.007340[\\*](#t002fn002){ref-type=\"table-fn\"}\n\n  ARI         0.027528   -0.006024   0.000973    1.075815   0.282010                                        \n\n  Anaemia     0.014738   -0.006024   0.665518    0.665518   0.505719                                        \n\n  Fever       0.017944   -0.000982   0.000982    0.764720   0.444438                                        \n\n  Diarrhoea   2003       0.063113    -0.005747   0.000859   2.349601                                        0.018794[\\*](#t002fn002){ref-type=\"table-fn\"}\n\n  ARI         0.041527   -0.005747   0.000877    1.596134   0.110459                                        \n\n  Anaemia     0.041527   -0.005747   0.000877    1.596124   0.110459                                        \n\n  Fever       0.026199   -0.005747   0.000877    1.078826   0.280665                                        \n\n  Diarrhoea   1998       0.083797    -0.005650   0.000900   2.981013                                        0.002873[\\*](#t002fn002){ref-type=\"table-fn\"}\n\n  ARI         0.067498   -0.005650   0.000865    2.486490   0.012901[\\*](#t002fn002){ref-type=\"table-fn\"}   \n\n  Fever       0.138308   -0.005650   0.000883    4.845580   0.000001[\\*](#t002fn002){ref-type=\"table-fn\"}   \n\n  Diarrhoea   1993       0.071159    -0.005780   0.000961   2.481752                                        0.013074[\\*](#t002fn002){ref-type=\"table-fn\"}\n\n  ARI         0.052478   -0.005780   0.000830    2.022656   0.043109[\\*](#t002fn002){ref-type=\"table-fn\"}   \n\n  Fever       0.130203   -0.005780   0.000974    4.357366   0.000013[\\*](#t002fn002){ref-type=\"table-fn\"}   \n  ---------------------------------------------------------------------------------------------------------------------------------------------------------\n\nARI-Acute respiratory infection;\n\n\\* statistically significant (p \\< 0.05)\n\nBecause of the low values for Moran's *I* indicating low positive correlation necessitated further analysis to understand at what spatial scale the data cluster. An incremental analysis was performed.\n\n[Table 3](#pone.0221324.t003){ref-type=\"table\"} indicates that the spatial scale at which each childhood morbidity indicated positive spatial autocorrelation varied. Diarrhoea did not have any peak distance for positive correlation for 2014, 2008 but was significant in 1993 (with a peak distance of 17135.07meters. The remaining morbidity had peak positive spatial correlation within minimum distance of 101951.24 meters and maximum of 185659.79 meters.\n\n10.1371/journal.pone.0221324.t003\n\n###### Incremental spatial autocorrelation Global Moran's *I*'s analysis of childhood morbidity (1993--2014).\n\n![](pone.0221324.t003){#pone.0221324.t003g}\n\n  Morbidity   Year        Peak distance (m)   Moran's Index   Expected Index   Variance   *z*-score                                        *p*-value\n  ----------- ----------- ------------------- --------------- ---------------- ---------- ------------------------------------------------ -----------------------------------------------\n  Diarrhoea   2014        \\-                  \\-              \\-               \\-         \\-                                               \\-\n  ARI         117331.86   0.059704            -0.005291       0.000374         3.360480   0.000778 [\\*](#t003fn002){ref-type=\"table-fn\"}   \n  Anaemia     178854.34   0.039277            -0.005291       0.000161         3.514823   0.000440 [\\*](#t003fn002){ref-type=\"table-fn\"}   \n  Fever       101951.24   0.066769            -0.005291       0.000459         3.361907   0.000774[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Diarrhoea   2008        \\-                  \\-              \\-               \\-         \\-                                               \\-\n  ARI         \\-          \\-                  \\-              \\-               \\-         \\-                                               \n  Anaemia     148214.05   0.080945            -0.006024       0.000293         5.083468   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Fever       162562.52   0.102688            -0.005952       0.000239         7.034425   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Diarrhoea   2003        139956.83           0.144554        -0.005747        0.000308   8.570825                                         0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}\n  ARI         139956.83   0.101640            -0.005747       0.000315         6.054973   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Anaemia     139956.83   0.101640            -0.005747       0.000315         6.054973   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Fever       139760.77   0.046944            -0.005682       0.000312         2.980574   0.002877[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Diarrhoea   1998        118649.23           0.107844        -0.005650        0.000418   5.552924                                         0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}\n  ARI         147790.71   0.072297            -0.005650       0.000263         4.804862   0.000002[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Fever       \\-          \\-                  \\-              \\-               \\-         \\-                                               \n  Diarrhoea   1993        171351.07           0.141159        -0.005780        0.000204   10.288151                                        0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}\n  ARI         171351.07   0.067507            -0.005780       0.000174         5.553500   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n  Fever       185659.79   0.124020            -0.005780       0.000170         9.967209   0.000000[\\*](#t003fn002){ref-type=\"table-fn\"}    \n\n-No peak distance results\n\n\\* statistically significant (p \\< 0.05)\n\nSpatial clusters of childhood morbidity {#sec015}\n---------------------------------------\n\nChildhood morbidity (diarrhoea, ARI, anaemia, and fever) clusters were identified. The clustered districts were classified as: high-high cluster (light brown coloured); high-low outlier (red coloured); low-high outlier (blue coloured); and low-low cluster (light blue coloured). Clustering of childhood morbidity was found largely in the middle and northern portions of the country.\n\nMost of the high clustered cases of diarrhoea where located in the Upper West, Upper East, and Northern. Except for the year 2003, most of the clustering occurred in Upper East, and Northern. In 2014, 19 districts were found to have highly clustered levels of diarrhoea cases. Fifteen of the highly clustered districts were found in Upper East (6) and Northern (9). A few of the highly clustered districts were found in Upper West (3), and Brong Ahafo (1) ([Fig 2](#pone.0221324.g002){ref-type=\"fig\"}). Nine districts were found to have clustered occurrence of diarrhoea with Northern having more of them in 2008. Six districts had high clustered cases of diarrhoea in Upper East and Northern in 1998 and ten districts exhibited high clustered episodes of diarrhoea in 1993.\n\n![Spatial cluster analysis (Local Moran I) of childhood diarrhoea.\\\n1993--Districts with high clusters of diarrhoea found were situated in Upper East, and Northern. 1998--High clustered districts were in Upper East, and Northern. 2003--High clustered districts were found in Upper West, and Northern. 2008--High clustered districts in Upper West, Upper East, and Northern. 2014--Districts with high clusters of diarrhoea located in Upper West, Upper East, Northern, and Brong Ahafo.](pone.0221324.g002){#pone.0221324.g002}\n\nClustering of ARI happened more in the northern portion (Upper West, Upper East, Northern) and some occurred in the middle and southern portions (Central, Accra, Eastern) of the country. [Fig 3](#pone.0221324.g003){ref-type=\"fig\"} displays districts identified with high clustered cases of ARI. In 2014, five districts in Upper East, three in Eastern, two in Greater Accra and three in Central were identified to have high clustered occurrence of ARI. Also, high clusters of ARI were found in three districts in 2008; 11 districts in 2003; six districts in 1998; and 10 districts in 1993.\n\n![Spatial cluster analysis (Local Moran I) of childhood ARI.\\\n1993--High clustered districts with ARI found in Upper East, Northern, and Volta. 1998--High clustered districts were in Upper East, and Northern. 2003--high clustered districts located in Upper West, and Northern. 2008--High clusters situated in Upper West, and Northern. 2014--Districts in Upper East, Northern, Brong Ahafo, Eastern, Accra, and Central showed high clusters of ARI.](pone.0221324.g003){#pone.0221324.g003}\n\nMost clustered cases of anaemia occurred in Upper West in the year 2003. For the following years, few clustered cases were still observed in Upper West in 2008 while some patches of clusters occurred in Northern, and Upper East in 2014. Also, six districts each were identified to have high clusters of anaemia in 2014 and 2008. In 2003, ten districts showed high clustered occurrence of anaemia in Upper West ([Fig 4](#pone.0221324.g004){ref-type=\"fig\"}). Almost the districts Upper West showed high levels of clustering in 2003. Some clustering was observed in 2008 and 2014 for Upper West, Upper East, and Northern.\n\n![Spatial cluster analysis (Local Moran I) of childhood anaemia.\\\n1993--High clustered districts in Upper East, Northern, Eastern, and Volta. 1998--High clustered districts in Upper West, Upper East, Northern, and Accra. 2003--All high clustered districts in Upper West. 2008--High clustered districts in Upper East, Northern, and Brong Ahafo. 2014--High clustered districts in Upper West, Upper East, Northern, and Brong Ahafo.](pone.0221324.g004){#pone.0221324.g004}\n\nAcross the years, more clusters of fever were in Upper West for the year 2003. For the other years, high clusters of fever were found in Upper East, Northern, and Brong Ahafo. In relation to fever, eight districts displayed high clusters in 2014 ([Fig 5](#pone.0221324.g005){ref-type=\"fig\"}). In the years 2008 and 2003, 10 districts demonstrated high clustering of children who experienced fever. Eleven districts displayed highly clustered levels of fever were found in 1998 and 17 districts equally were linked with clustered patterns in 1993. Most districts identified with fever were found within the northern portions of the country.\n\n![Spatial cluster analysis (Local Moran I) of childhood fever.\\\n1993--Upper East, Northern, Eastern, and Volta had districts with high clusters of Fever. 1998--High clustered districts were in Upper East, Northern, and Accra. 2003--All clustered districts in Upper West. 2008---Upper East, Northern, and Brong Ahafo had high clustered districts. 2014--Clustered districts found in Upper West, Upper East, Northern, and Brong Ahafo.](pone.0221324.g005){#pone.0221324.g005}\n\nHot spots of childhood morbidity {#sec016}\n--------------------------------\n\nThe results from the Local Moran I shows the outliers and clusters. Hotspot/Coldspot analysis furhter Local Morans I by indicating whether the observed spatial clustering of high or low values is more pronounced than one would expect in a random distribution. Statistical results for hotspots/coldspots explains that the hot spots of childhood morbidity were assessed with confidence intervals of: 99%; 95%; and 90% at district levels. The confidence level explains the level of spatial clustering of higher or lower values. In [Fig 6](#pone.0221324.g006){ref-type=\"fig\"}, it was found that 16 districts (99% CL), seven districts (95% CL) and four districts (90% CL) were significant hot spots of diarrhoea in 2014. In 2008, eight districts (99% CL), nine districts (95%) and one district (90% CL) were hot spots of diarrhoea. In 2003, 11 districts (99% CL), six districts (95% CL) and 10 districts (90% CL) were hot spots for diarrhoea. Ten districts (99% CL), five districts (95% CL) and three districts (90% CL) were significant hot spots in 1998. For the year 1993, six districts (99% CL), eight districts (95% CL) and two districts (95% CL) were the hot spots. In all the survey years, most diarrhoea hot spot districts were in Northern, Upper East and Upper West. Only hot spot districts with 99% CL are presented in [Table 4](#pone.0221324.t004){ref-type=\"table\"}.\n\n![Hot spots of childhood diarrhoea.\\\n1993--Diarrhoea hot spot districts were located in Northern. 1998--Hot spots were found in Upper East, and Northern. 2003--Hot spots in Upper West, and Northern. 2008--Districts identified as hot spots were in Upper East, and Northern. 2014--Districts in Upper West, Upper East, and Northern were hot spots for diarrhoea.](pone.0221324.g006){#pone.0221324.g006}\n\n10.1371/journal.pone.0221324.t004\n\n###### Hot spot districts of diarrhoea (2014--1993).\n\n![](pone.0221324.t004){#pone.0221324.t004g}\n\n  -----------------------------------------------------------------------\n  Year   Hot spot districts\n  ------ ----------------------------------------------------------------\n  2014   East Gonja (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Sagnarigu (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Tamale Metro (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Tolon (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Savelugu Nanton(N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Gushiegu (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal(UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Nabdam (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Binduri (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Garu-Tempane (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bawku Municipal (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Sissala West (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\n\n  2008   North Gonja (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Kumbungu (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Savelugu Nanton (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         West Mamprusi (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         East Mamprusi (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bunkpurugu Yunyoo (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\n\n  2003   Nandom (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Lawra (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Jirapa (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Sissala West (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Daffiama Bussie Issa(UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Nadowli (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Sissala East (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Wa Municipal (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Wa West (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Wa East (UW)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         East Gonja (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\n\n  1988   East Mamprusi (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Chereponi (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal(UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Nabdam (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Binduri (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Garu-Tempane (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Chereponi (UE)[\\*](#t004fn001){ref-type=\"table-fn\"}\n\n  1993   Mamprugu Moaduri (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         West Mamprusi (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         North Gonja (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Kumbungu (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Tolon (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\\\n         Tamale Metro (N)[\\*](#t004fn001){ref-type=\"table-fn\"}\n  -----------------------------------------------------------------------\n\n\\*---99% CL; N---Northern; UE---Upper East; UW---Upper West\n\n[Fig 7](#pone.0221324.g007){ref-type=\"fig\"} showed that 18 districts (99% CL), five districts (95% CL), and seven districts (90% CL) were found to be significant ARI hot spots in 2014. Most of the districts affected with ARI were in Upper East, Central and Eastern regions. In 2008, three districts (99% CL), nine districts (90% CL) and two districts (90% CL) were hot spots for ARI. Largely, in 2003, most of the ARI hot spot districts (9) were located in Upper West. In the years 1998 and 1993, Upper East recorded a larger number of ARI hot spot districts (8) compared to the other regions. ARI hot spot districts (99% CL) are summarised in [Table 5](#pone.0221324.t005){ref-type=\"table\"}.\n\n![Hot spots of childhood ARI.\\\n1993--Hot spots districts for ARI were in Upper East, and Volta. 1998--Hot spot districts in Upper East, and Northern. 2003--Districts in Upper West were mainly hot spots for ARI. 2008--Districts in Upper West, and Northern were hot spots for ARI. 2014--Hot spot districts were found in Upper East, Northern, Eastern, Accra, and Central.](pone.0221324.g007){#pone.0221324.g007}\n\n10.1371/journal.pone.0221324.t005\n\n###### Hot spot districts of ARI (2014--1993).\n\n![](pone.0221324.t005){#pone.0221324.t005g}\n\n  ------------------------------------------------------------------------------\n  Year   Hot spot district\n  ------ -----------------------------------------------------------------------\n  2014   West Mamprusi (N)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal(UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Builsa South (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Gomoa West (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Agona West Municipal (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Birim Central Municipal (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         West Akim Municipal (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Upper West Akim (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Nsawam Municipal (E)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Ga West Municipal (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         La Nkwantanang/Madina (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Adentan Municipal (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Ledzokuku Krowor Municipal (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Accra Metro (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Ga South Municipal (GA)[\\*](#t005fn001){ref-type=\"table-fn\"}\n\n  2008   Wa West (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Mion (N)[\\*](#t005fn001){ref-type=\"table-fn\"}\n\n  2003   Nandom (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Lawra (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Daffiama Bussie Issa (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Wa West (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Wa Municipal (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Wa East (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Nadowli (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Sissala West (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Sissala East (UW)[\\*](#t005fn001){ref-type=\"table-fn\"}\n\n  1998   East Mamprusi (N)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Kassena Nankani East (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Nabdam (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\n\n  1993   Bongo (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Garu-Tempane (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\\\n         Bawku Municipal (UE)[\\*](#t005fn001){ref-type=\"table-fn\"}\n  ------------------------------------------------------------------------------\n\n\\*-99% CI; N-Northern; UE-Upper East; UW-Upper West; GA-Greater Accra; E-Eastern\n\nIn relation to anaemia, for the year 2014, all hot spot districts were found in Northern. In 2008, six districts (99% CL), ten districts (95% CL) and two districts (90% CL) were anaemia hot spot districts. Most of them were located in Upper West. All of the anaemia hot spot districts (9) in 2003 were found in Upper West ([Fig 8](#pone.0221324.g008){ref-type=\"fig\"}). [Table 6](#pone.0221324.t006){ref-type=\"table\"} contains anaemia hot spot districts.\n\n![Hot spots of childhood anaemia.\\\n2003--Upper West had most hot spot districts for anaemia. 2008--Hot spot districts were found in Upper West, and Northern. 2014--Hot spot districts were found in Northern.](pone.0221324.g008){#pone.0221324.g008}\n\n10.1371/journal.pone.0221324.t006\n\n###### Hot spot districts of anaemia (2014--2003).\n\n![](pone.0221324.t006){#pone.0221324.t006g}\n\n  ------------------------------------------------------------------------\n  Year   Hot spot districts\n  ------ -----------------------------------------------------------------\n  2014   Saboba (N)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Yendi Municipal (N)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Zabzugu (N)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Nanumba North (N)[\\*](#t006fn001){ref-type=\"table-fn\"}\n\n  2008   Daffiama Bussie Issa (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Nadowli (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa Municipal (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa West (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa East UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         West Gonja (N)[\\*](#t006fn001){ref-type=\"table-fn\"}\n\n  2003   Nandom (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Lawra (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Daffiama Bussie Issa (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa West (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa Municipal (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Wa East (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Nadowli (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Sissala East (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\\\n         Sissala West (UW)[\\*](#t006fn001){ref-type=\"table-fn\"}\n  ------------------------------------------------------------------------\n\n\\*-99% CI; N-Northern; UE-Upper East; UW-Upper West\n\nNorthern, Upper East and Upper West hosted all of the fever hot spot districts. [Fig 9](#pone.0221324.g009){ref-type=\"fig\"} revealed that seven districts each in Upper East and Upper West, and six districts in Northern that significantly were linked with high occurence of fever in 2014. Nine districts (99% CL), 11 districts (95% CL) and 11 districts (95% CL) were found to be significant hot spot districts in 2008; and Upper West had more fever hot spot districts. In 2003, most of the hot spot districts were found in Upper West (9) and Upper East (6). In 1998, 10 districts (99% CL), five districts (99% CL) and three districts (90% CL) were fever hot districts; and most of them were located in Upper East. For the year 1993, fever hot spot districts were identified: nine districts each in Upper East and Northern; eight districts in Volta; and two districts in Eastern. Fever hot spot districts (99% CL) are shown in [Table 7](#pone.0221324.t007){ref-type=\"table\"}.\n\n![Hot spots of childhood fever.\\\n1993--Hot spot districts for fever were in Upper East, Northern, and Volta. 1998--Upper East, and Northen had hot spot districts. 2003--Hot spot districts were in Upper West, and Upper East. 2014--Hot spots for fever were in Upper West, Upper East, and Northern.](pone.0221324.g009){#pone.0221324.g009}\n\n10.1371/journal.pone.0221324.t007\n\n###### Hot spot districts of fever (2014--1993).\n\n![](pone.0221324.t007){#pone.0221324.t007g}\n\n  ------------------------------------------------------------------------\n  Year   Hot spot\n  ------ -----------------------------------------------------------------\n  2014   Nandom (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Lawra (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Daffiama Bussie Issa (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Jirapa (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Lambussie Karni (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Sissala West (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Nadowli (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bongo UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bawku Municipal (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Garu-Tempane (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         West Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         East Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\n\n  2008   Bawku West (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         East Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         West Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Kumbungu (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Savelugu Nanton (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Karaga (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Sagnarigu (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Tamale Metro (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Nanumba North (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\n\n  2003   Nadowli (UW)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Binduri (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\n\n  1988   Kassena Nankana West\\\n         Builsa North (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Kassena Nankani East (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Nabdam (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Binduri (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         East Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\n\n  1993   Kassena Nankani West (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Builsa North (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Kassena Nankani East (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bongo (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bolgatanga Municipal (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Talensi (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Bawku West (UE)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Mamprugu Moaduri (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         West Mamprusi (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Kumbungu (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Savelugu Nanton (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Sagnarigu (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Tamale Metro (N)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Keta Municipal (V)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Ketu North (V)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Akatsi South (V)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         South Tongu (V)[\\*](#t007fn001){ref-type=\"table-fn\"}\\\n         Central Tongu (V)[\\*](#t007fn001){ref-type=\"table-fn\"}\n  ------------------------------------------------------------------------\n\n\\*-99% CI; N-Northern; UE-Upper East; UW-Upper West; V-Volta\n\nDiscussion {#sec017}\n==========\n\nThe study was designed to mainly reveal hot spots of childhood morbidity in Ghana. In addition, the effect of rural versus urban residence on the occurrence of childhood morbidity was examined. This was done by considering four top morbidities (diarrhoea, ARI, anaemia, and fever) and using five different data sets (1993, 1998, 2003, 2008 and 2014) of GDHS. Children in rural versus urban areas were much significantly associated with reported cases of childhood morbidity. Districts identified to be burdened with each of the childhood morbidities would inform and guide major stakeholders such as Ghana Health Service, District Health Directorates, and health related NGOs to allocated their resources rationally for much desired outcomes.\n\nMost of the hot spot districts for the prevalence of diarrhoea were concentrated in Upper West, Upper East, and Northern. For instance, in the Multiple Indicator Cluster Survey in 2011 \\[[@pone.0221324.ref024]\\], Northern Region recorded the highest prevalence of childhood diarrhoea. Records of high prevalence of diarrhoea in this part of the country may be attributed to environmental factors such as insufficient access to potable water and poor sanitation. These factors expose children to diseases; and diarrhoea is the most common. A study conducted by Cheng et al. \\[[@pone.0221324.ref025]\\] documented that 50 percent of the population in Northern Region has access to improved drinking water. And in Savelugu-Nanton, less than 50 percent of its inhabitants had access to improved drinking water. Again open defecation is highly practiced among districts in the region despite the implementation of interventions such as Community-Led Total Sanitation \\[[@pone.0221324.ref026]\\]. Open defecation encourages the transfer of pathogens from faeces to food and water systems through run-off and houseflies which has the potential to cause diarrhoea in the in the region hence, Savelugu-Nanton. Osei (2017) attests to this poor sanitation as the major causes of diarrhoea in the region. These conditions cited in Northern are similar to those in Upper East and Upper West. Therefore, it is likely that these other districts are faced with sanitation and water challenges.\n\nAdditionally, hot spots for ARI were commonly clustered in Upper East, Upper West, and Northern. Some districts in Accra, in 2014, were equally identified as hot spots for episodes of ARI. The top rated hot spot for childhood ARI was Accra Metro (the capital city) in the Greater Accra Region. In densely populated urban areas, like Accra, overcrowding, vehicular fumes and industrial waste (gasses) could cause ARI among children. Children in urban areas might be more exposed to pollutants such as carbon monoxide, nitrogen oxide and sulfur dioxide emitted by cars and industries. Another reason for high rates of ARI in Accra could be attributed to low level of vegetation and tree life which can serve as air filters as the area is more of a 'concrete jungle'. For the occurrence of ARI in the other districts which are more in rural settings may be due to cooking arrangements where children are exposed to harmful gasses from fuelwood.\n\nFurther, hot spot districts identified for prevalence of anaemia were located along the north-eastern corridor, and north western portions of the country. And the unavailability the required foods for children to prevent anaemia may not be readily available in the hot spot districts. In a study by Ewusie et al. \\[[@pone.0221324.ref005]\\], they revealed that Upper East Region had the highest prevalence rate of anaemia. During the surveys, blood samples were taken from children to test for anaemia and a number of children who were anaemic had low haemoglobin levels less than 11 grams per deciliter (g/dl) \\[[@pone.0221324.ref012]\\]. This, accordingly, is caused by the kinds of low-quality foods children consume that predispose them to reduced red blood cells and consequently decreased levels of haemoglobin \\[[@pone.0221324.ref027]\\]. To avert this, it is recommended that children should consume additional to basic cereals such as maize, millet, sorghum, and soy \\[[@pone.0221324.ref028]\\]. This, therefore, suggests that most children in the identified hot spot districts consumed poor diets. Thus, diets that do not contain essential quantities of iron and micronutrients required to boost haemoglobin levels. Aside the aforementioned reasons, there are food insecurity issues in the districts and this might have compounded dietary deficiencies among children, thereby resulting in high prevalence of anaemia among children \\[[@pone.0221324.ref029]\\].\n\nIn the study, there were also districts identified as hot spots for fever; mainly in the north-western and north-eastern parts of the country. Most of the hot spot districts were situated in Upper West and Northern regions across the survey years. High rates of fever in these locations could be as a result of high malaria cases; since fever is a common symptom of malaria among children. Other infections caused by viruses and bacteria among children could have also contributed to high episodes of fever in the identified hot spot districts \\[[@pone.0221324.ref030]\\].\n\nThe hot spot districts identified are assumed to be homogeneous. Identification of specific target localities such as households, and more socio-economic and environmental factors within districts may further inform the formulation and implementation of interventions. Notwithstanding the stated limitations, this study has importantly examined the spatial patterns and hot spots of four top childhood morbidities in Ghana using districts as the main reference at five different time points. These findings should be interpreted with caution since seasonality of weather conditions were not part of the study variables.\n\nConclusion {#sec018}\n==========\n\nDiarrhoea and fever hot spots are in the north-western, north-eastern and middle portions of the country. For ARI, hot spots are in the north-eastern, middle, and along the coast. Childhood anaemia cases are common in the north-eastern corner, and hot spots for fever are mainly in the northern divide of the country. To minimize childhood morbidity in Ghana, various Ghana Health Directorates could design context-based interventions to target hot spot districts within their administrative areas. However, further geospatial studies on households within districts would be more beneficial for various local health offices to adequately target populations at risk.\n\nWe thank MEASURE DHS for making available freely the Ghana Demographic and Health Survey data set used for this study.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "Charcot-Marie-Tooth disease is a demyelinating peripheral neuropathy and is a heterogeneous inherited disorder (40 loci have been described so far) affecting peripheral nerves. Two forms, CMT1A and CMTX, account for 60 and 15% of patients, respectively, with clear familial transmission (Boerkel et al., [@B3]).\n\nCMTX is caused by mutations in the GBJ1 gene encoding the synthesis of connexin 32 (Cx32, Bergoffen et al., [@B2]), which is a myelin protein related to PNS and CNS (Scherer et al., [@B21]). Cx32 is a membrane protein located in gap junctions, which forms hexameric hemichannels called connexons. However, the mechanism by which anomalies in connexin 32 affect myelination and function of PNS remains is still unclear. Here, we share our opinion that CamKinase are involved in the pathophysiology of CMTX and likely involve genomic instability caused by mutations in Gjb1.\n\nConnexin, CamKII, and genomic stability {#s1}\n=======================================\n\nApart from its role in the synthesis of Cx32, an inappropriate expression of the gene Gjb1 has recently been associated with genomic instability during mitosis (Mones et al., [@B13], [www.mitocheck.org](http://www.mitocheck.org)) leading to the hypothesis that connexins may be linked, directly or indirectly, to genomic stability. The literature suggests that at least one signaling pathway is associated with both phenomena---calcium homeostasis. Indeed, connexin 32 oligomerizes in Schwann cells to form hemichannels known to facilitate the diffusion of metabolites across the layers of myelin sheath. It is now established that calmodulin (CaM) can directly bind the C-terminus of Cx32 and regulate gap junction channels (Torok et al., [@B26]; Ahmad et al., [@B1]; Dodd et al., [@B5]; Stauch et al., [@B23]). In response to increased intracellular levels of calcium, CaM then binds to Cx32 channels triggering its closure and allowing the extrusion of Ca^2+^ via others channels to maintain calcium homeostasis. (Lurtz et al., [@B10]) In CMTX, anomalies of Cx32 alter the binding of CaM, leading to disturbance of calcium homeostasis by maintaining high intracellular calcium concentrations. (De Vuyst et al., [@B4]) However, as Cx32 anomalies decreases the binding of CaM, we can hypothesize an increase of \"free\" CaM in the presence of high concentrations of calcium leading to autophosphorylation of CaMKII upon activation by Ca^2+^ and CaM (Hell Johannes, [@B8]). Acting via a retrocontrol mechanism, CaMKII can then phosphorylate Cx32, leading to modulations of the gap junctional communications (Lampe and Lau, [@B9]; Stauch et al., [@B23]). Recently, we have shown that KN93 and KN-62---two inhibitors of reference of CaMKII---could restore connexin activity (Mones et al., [@B14], p. 382), but as the level of phosphorylation of Cx32 has not been measured, we cannot make any conclusions about the mechanism involved in this overstimulation.\n\nGenomic instability occurs when multipolar spindles appear due to centrosome overduplication. In 2002, Matsumoto and Maller described in Xenopus egg extracts that inactivation of CaMKII could block centrosome overduplication suggesting that calcium oscillations in the cell may be linked to centrosome duplication. (Matsumoto and Maller, [@B12]) Also also Xenopus egg extracts, Reber et al. confirmed this hypothesis and showed that calcium or constitutive active CaMKII promotes microtubule destabilization (Reber et al., [@B17]). We recently showed that CamKII activity is overactivated in cells from a mouse model of CMTX involving mutations in Gjb1 (Mones et al., [@B14]) and in cells from CMTX1 patients (Mones et al., [@B15]). These cells presented centrosome overduplication and genomic instability. Interestingly, this cellular phenotype was corrected by treatment with CamKII inhibitors both in Gjb1-mutated mouse and human patient cells (Mones et al., [@B14], [@B15]).\n\nThese data suggest that anomalies of Cx32 due to GBJ1 mutations lead to unbalanced calcium homeostasis resulting in a genomic instability through the overactivation of CaMKII.\n\nConnexin and myelin {#s2}\n===================\n\nMyelin deficit is the first alteration observed in Cx32-deficient mice, supporting the link between connexin and myelin (Scherer et al., [@B22]). Cx32 is the most abundant connexin isoform in Schwann cells, and it oligomerizes to form gap junction channels between two cells. However, gap junctions in Schwann cells can also form intracellular channels through the layers of myelin providing a direct route for the diffusion of metabolites and second messengers from the Schwann cell to the axon. This is likely necessary for myelin formation (Ressot and Bruzzone, [@B18]). Our recent work with KN93 and KN62, two inhibitors of CaMKII, concur with this hypothesis because they may block the CaMKII-mediated phosphorylation of Cx32 and at least partially restore the activity of the connexon (Mones et al., [@B14], [@B15]).\n\nHowever, some facts challenge this mechanistic hypothesis. Indeed, Cx32 is the major component of liver gap junctions, and CMTX mutations of GJB1 do not result in a severe phenotype in this organ. Regarding CNS, although patients with central clinical signs have been reported (Hanemann et al., [@B7]), the majority of CMTX patients did not present central clinical signs. However, they did present infra-clinical and asymptomatic signs. This a major observation in inherited disorders caused by mutations in an ubiquitously expressed protein. There is usually no clear explanation.\n\nCamKII and myelin {#s3}\n=================\n\nAnother hypothesis to explain the CMTX phenotype due to GJB1 mutations is overactivation of CaMKII. A recent work by Weggener et al. supports this hypothesis. They showed that a mouse line invalidated for CamKII\u03b2 gene exhibits a deficit in myelination in the CNS. The authors suggested that this deficit is due to anomalies in phosphorylation of cytoskeletal proteins (Waggener et al., [@B27]). Earlier, Sahenk et al. reported that CMTX xenografts where Schwann cells bear mutations on Cx32 were presenting major cytoskeletal alterations (Sahenk, [@B20]). While the link between CaMKII, cytoskeleton anomalies and myelin deficits may be established, we do not yet know whether the CaMKII-related genomic instability observed in cell phenotypes of CMTX patients is part of the same signaling pathway leading to myelin deficit. However, this hypothesis is interesting because CMTX-induced genomic instability in the Schwann cells may affect the PNS myelination in a downstream process occurring over a long period of time before the appearance of clinical symptoms. In another form of CMT (CMT1A), infants were showed to bear the CMT1A mutation, but without clinical symptoms. They still had a reduced nerve conduction velocity indicating that a deficit in myelination is already present before the disease manifests (Garcia and Combarros, [@B6]).\n\nThis begs the question of calcium homeostasis and myelination. Nobbio et al. demonstrated the impact of perturbation of calcium homeostasis in a CMT1A model (Nobbio et al., [@B16]). This is strengthened by the report of Stevens et al. demonstrating that perturbation of calcium homeostasis inhibits differentiation of precursors of Schwann cells (Stevens and Fields, [@B24]). In the same report, they suggested that it involves Ca^2+^, calmodulin and calmodulin kinase.\n\nDemyelination and inflammation {#s4}\n==============================\n\nSeveral mechanisms are likely co-existing between the primary causes of the disease (including gene mutation and the resulting deregulation of the downstream cascade) and the symptomatic manifestations of the disease (locomotion, fatigue, pain, etc.) In this field, Martini et al. demonstrated that murine models of demyelinating CMT present inflammation, in which both innate and adaptive immune systems would act as amplifiers of the primarily genetically caused neuropathies (for review, Martini and Willison, [@B11]). However, in their putative mechanism, they argued propose a yet to be identified primary activator that would lead to myelin phagocytosis by macrophages and Schwann cell dedifferentiation. The inflammatory process would be an aggravating factor of the neuropathy. Considering that inflammation is very frequently associated with degenerative disorders (Alzheimer\\'s disease, diabetes ...), it is likely that inflammation acts in clinical picture of demyelinating CMT. The question of cause or consequence of the disorder is difficult to answer. However, the assumption that anomalies in tissues structure (abnormal myelin, decompacted myelin, etc), are leading to inflammation is very likely.\n\nOur opinion is that inherited demyelinating disorders are developmental disorders linked to anomalies of Schwann cells terminal differentiation and myelin formation. However, this putative role of inappropriate differentiation as the primary cause of inflammation is still to be proved. In addition, potential direct links between genomic instability and neuroinflammation remain still elusive, although Taupin reported an association between neuroinflammation in Alzheimer disease and genomic instability (aneuploidy) (Taupin, [@B25]). We did not find a report of a similar phenomenon in PNS disorders.\n\nCMTX and therapeutic approaches {#s5}\n===============================\n\nTherapeutic efficacy requires a good understanding of the mechanisms involved in the disease. Recent works in the CMTX field show that there are many questions still pending before this genetic disease can be fully understood. Inflammation is certainly a good target that could lower disease severity including pain. Targeting upstream events linked to the primary cause of the disorder would be also a promising track to treat CMTX patients. Indeed, we have demonstrated that *in vitro* and *in vivo* abnormal phenotypes are corrected in animal models by treatment with CamKII inhibitors (Mones et al., [@B14]). In addition, we have demonstrated that the same anomalies are observed in cells from CMTX patients and that these anomalies are corrected using CamKII inhibitors. CaMKII is a common target to the various mechanistic hypotheses presented earlier. In order to prevent potential toxicity, development of CaMKII modulators may allow down-activation of CaMKII in the PNS without strongly altering calcium homeostasis in other cells. The CamKII inhibitors KN93 and KN62 have never been tested in humans and deserve a full optimization process to improve their efficacy toward CamKII. However, although no toxic evaluation has been reported, no adverse effect has been reported in animals treated with this molecule.\n\nThe hypothesis dealing with genomic instability and myelin deficit is also interesting in terms of therapeutic options. Indeed, several kinases are already known to be involved in genomic instability. These likely act downstream to CamKinases. Pim1 is one of these kinases that is involved in genomic instability in cancer cells (Roh et al., [@B19]). We have shown that Pim1 inhibitors can correct genomic instability in GBJ1-mutated cell lines, but were unable to correct Cx32 activity (Mones et al., [@B14]). It would be interesting to test these Pim1 inhibitors in CMTX mice to evaluate their ability to restore a normal phenotype, and validate or rule out this mechanistic hypothesis. Moreover, downstream kinases could be an additional therapeutic target.\n\nAuthor contributions {#s6}\n====================\n\nFB and MF wrote the manuscript. BG critically reviewed the manuscript. All authors read and approved the final version of the manuscript.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nThis article is dedicated to S. Mones who published most of the cited papers. Saleh was a MD and defended his PhD thesis in 2014. He returns now to Syria, and was incorporated in Syrian army as military doctor. He was killed Saturday September 10th in a suicide terrorist attack of his hospital.\n\n[^1]: Edited by: Alessandro Tozzi, University of Perugia, Italy\n\n[^2]: Reviewed by: Georg Zoidl, York University, Canada\n"}
+{"text": "Background\n==========\n\nRapid advances in epigenetics have led to the discovery of complex mechanisms of gene regulation involving phenomena such as DNA methylation and chromatin modifications. Methylation of particular histone residues has been found to correlate with specific and often opposing cellular functions, e.g. trimethylation of histone H3 lysine 4 (H3K4me3) is associated with transcriptional start sites (TSSs) of active genes while trimethylation of histone H3 lysine 27 (H3K27me3) is found to mark broad genomic regions for repression. Recent studies have also suggested that characteristic combinations of histone modifications or 'chromatin states' define functional elements of the genome and determine their contribution to transcriptional regulation \\[[@B1]-[@B3]\\]. Moreover, the epigenetic state of host genes have been observed to be affected by viral infection leading to tumors in humans \\[[@B4]-[@B6]\\]. Thus, epigenetics constitute a dynamic regulatory framework linking genotypes with environmental factors that could play a major role in differential disease responses among individuals having high genetic similarity.\n\nMarek's disease (MD) is a highly contagious disease caused by an oncogenic \u03b1-herpesvirus MD virus (MDV) and characterized by T-cell lymphomas in chickens \\[[@B7]\\]. Major losses to the poultry industry as a result of MD have largely been averted due to the success of various vaccination strategies which, remarkably, is also the first instance of the successful control of a natural cancer-causing agent using vaccines \\[[@B8]-[@B10]\\]. However, the virulence of the virus may have progressively increased as a consequence of vaccine development \\[[@B11]-[@B13]\\]. Several reported instances of vaccine breaks or reduced efficacy of vaccination, therefore, underlines the importance of investigating resistance to the disease as a long-term strategy to control MDV.\n\nNatural resistance to MDV can be divided into two categories: major histocompatibility complex (MHC)-associated resistance, wherein different MHC haplotypes at the B blood group locus confer varying levels of resistance and non-MHC associated resistance in which birds having the same MHC haplotype exhibit vastly different responses to MDV infection. Inbred lines 6~3~ and 7~2~ developed at the Avian Disease and Oncology Laboratory (ADOL, East Lansing, MI) that we used in this study, fall into the latter category. These lines share a high degree of genetic similarity but have divergent responses to MDV infection completely independent of the MHC. Several studies have attempted to pinpoint factors responsible for conferring resistance \\[[@B14]-[@B16]\\], but confounding factors, such as, tissue types, virus strains and ages of birds have made it difficult to find a consensus. Multiple risk elements are possibly at play in this complex disease, and increased resistance or susceptibility is likely to be produced by a combination of such factors. In this study, we take a closer look at epigenetic factors behind different responses to MD with a view to a deeper understanding of the broader genomic impact of MDV infection.\n\nWe utilized the above population of inbred chickens -- line 6~3~ is highly resistant to MD, while line 7~2~ is highly susceptible -- and compared the epigenetic effects of MD. Genome-wide maps of H3K4me3 and H3K27me3 in thymus tissues of birds from these chicken lines at the latent stage of MDV infection were generated. We carried out systematic analyses to find differential chromatin marks induced by MDV infection. We also investigated co-localization patterns of the two chromatin modifications to detect putative bivalent domains and the effect of MDV on such domains. The results of our study confirm that Marek's disease has far-reaching effects on the epigenetic landscape of chicken lines with diverse responses to the virus and, thus, furthers our understanding of this complex disease.\n\nResults\n=======\n\nGenome-wide distribution of H3K4me3 and H3K27me3\n------------------------------------------------\n\nWe performed ChIP-Seq experiments on infected and uninfected birds from lines 6~3~ and 7~2~ to investigate the epigenetic effects of MDV infection. More than 76 million reads from eight samples were mapped to the chicken genome yielding 14418 and 24950 significantly enriched regions (SERs) for H3K4me3 and H3K27me3, respectively (Table [1](#T1){ref-type=\"table\"}). We further classified these regions as follows: Ubiquitous SERs were found in all samples and were likely due to similarities in the genetic background of the chickens. Line-specific SERs were present in only one line either before or after MDV infection, while condition-specific SERs appeared in both lines but only in individuals with the same infection status.\n\n###### \n\nSignificantly enriched regions (SERs) and associated genes in each sample\n\n  \u00a0                        \u00a0                     **H3K4me3**    \u00a0           **H3K27me3**   \u00a0\n  ------------------------ --------------------- -------------- ----------- -------------- ----------\n  **Line-Specific**        **63I**               647 (4.5)      78          3477 (13.9)    615\n  \u00a0                        **63N**               594 (4.1)      71          2514 (10.1)    896\n  \u00a0                        **63I,63N**           924 (6.4)      190         577 (2.3)      150\n  \u00a0                        **72I**               105 (0.7)      16          1658 (6.6)     451\n  \u00a0                        **72N**               126 (0.9)      11          2506 (10)      346\n  \u00a0                        **72I,72N**           73 (0.5)       17          330 (1.3)      89\n  **Condition-specific**   **63I,72I**           97 (0.7)       35          2061 (8.3)     579\n  \u00a0                        **63N,72N**           47 (0.3)       9           66 (0.3)       22\n  **Ubiquitous**           **63I,63N,72I,72N**   10691 (74.2)   9475        5831 (23.4)    2942\n  \u00a0                        **Total**             **14418**      **10206**   **24950**      **7904**\n\n63I: line 6~3~ infected, 63N: line 6~3~ control, 72I: line 7~2~ infected, 72N: line 7~2~ control.\n\nUbiquitous SERs formed the largest percentage of all enriched regions, accounting for 74.2% and 23.3% in H3K4me3 and H3K27me3 samples, respectively. In the case of H3K4me3, there were large differences in the number of specific SERs - more than 2000 line-specific SERs were found in line 6~3~, compared to about 300 in line 7~2~. Similarly, we found 50% more line-specific SERs of H3K27me3 in line 6~3~ (6568) compared to line 7~2~ (4494). However, upon closer examination, most of the line-specific and condition-specific SERs were revealed to have low read counts (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S1) corresponding to regions of low enrichment.\n\nGenes were divided into five regions -- promoter, 5' untranslated region (UTR), exons, introns and 3' UTR -- and the distribution of SERs across these elements was probed (Figure [1](#F1){ref-type=\"fig\"}A). We found a large number of intergenic regions marked by H3K27me3, consistent with high levels of this mark associated with areas of silent heterochromatin. In the case of H3K4me3, a larger proportion of SERs were found around the promoter, exons and the 5' UTR while similar proportions of H3K4me3 and H3K27me3 SERs were present in introns and 3' UTRs. A comparison of the genomic distributions of SERs in the different samples (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figures S2A, B) showed a similar number of H3K4me3 SERs across the promoter, exons and the 5' and 3' UTRs of genes. Line 6~3~ contained a higher number of intronic and intergenic SERs as compared to line 7~2~ although this did not appear to change as a result of MDV infection. On the other hand, a greater number of H3K27me3 SERs were found in the infected samples although these levels were similar in the two different lines.\n\n![**Genomic distribution of SERs and relationship between histone marks and gene expression.** (**A**) Distribution of SERs over different genomic elements shows increased levels of H3K4me3 around the promoter region and exons while there are increased levels of H3K27me3 on intergenic regions. (**B**-**E**) Relationship between gene expression and histone marks in infected line 6~3~ birds. Plots of histone modifications around the gene body (**B**, **C**) in genes having high (blue), medium (red), low (green) and no activity (brown): H3K4me3 shows positive correlation with gene expression levels while H3K27me3 exhibits a negative relationship. A comparison of epigenetic marks and transcriptional levels (**D**, **E**) confirms the same. Similar trends were observed in other experimental groups (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figures S3-5).](1471-2164-13-557-1){#F1}\n\nTo analyze the relationship between histone modifications and gene expression, histone modification profiles surrounding the TSS and gene body were plotted for genes ranked by their expression level (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figures [1](#F1){ref-type=\"fig\"}B-E and S3-5). As expected, a strong positive correlation was observed between gene expression and H3K4me3 marks with a distinct peak around the TSS and little to no enrichment within the gene body. On the other hand, H3K27me3 showed negative correlation with gene expression with a peak near the TSS followed by a broad plateau across the gene body. However, the latter relationship was non-linear -- genes with lower expression had similar levels of H3K27me3 marks and levels were markedly distinct only at higher expression levels (Figure [1](#F1){ref-type=\"fig\"}C, E).\n\nDifferential H3K4me3 marks on genes related to MD\n-------------------------------------------------\n\nWe conducted a comprehensive analysis of genome-wide chromatin marks to find significant differences in MDV-induced responses in line 6~3~ and 7~2~. We used two sets of comparisons: First, to assess the influence of MDV infection within each line, we compared the infected and the non-infected samples from the same line. Secondly, the non-infected samples from the two lines were compared to detect line-specific effects. As a result of this analysis we found 179 differential H3K4me3 SERs and 1116 differential H3K27me3 SERs that mapped to 59 and 66 genes, respectively (Table [2](#T2){ref-type=\"table\"}). A majority of differential SERs were found in the comparison between non-infected samples of the two lines (Additional files [2](#S2){ref-type=\"supplementary-material\"} and [3](#S3){ref-type=\"supplementary-material\"}) with several observed on genes that have been associated with MDV infection.\n\n###### \n\nDifferential SERs identified in thymus\n\n  \u00a0                 **H3K4me3**   **H3K27me3**             \n  ---------------- ------------- -------------- ---------- --------\n  **63I vs 63N**         9             4            42        1\n  **72I vs 72N**        30             13           5         0\n  **63N vs 72N**        148            46          1094       65\n  **Total**           **179**       **59**\u2020      **1116**   **66**\n\n\\*FDR \\< 0.4. \u2020 Some genes are shared between different comparisons. 63I: line 6~3~ infected, 63N: line 6~3~ control, 72I: line 7~2~ infected, 72N: line 7~2~ control.\n\n*MX1* is a zinc-finger transcription factor that has antiviral properties against a large number of viruses. *MX1* was upregulated after MDV infection \\[[@B17]\\] although its contribution to MD progression is unknown. MDV infection induced a highly significant increase in H3K4me3 enrichment in the promoter region of *MX1* in both lines (line 6~3~: p = 1.28x10^-7^, line 7~2~: p = 4.26x10^-9^; Figure [2](#F2){ref-type=\"fig\"}A). We observed a concurrent increase in transcript levels after MDV infection in line 7~2~ (p = 0.0264; Figure [2](#F2){ref-type=\"fig\"}B); *MX1* expression in line 6~3~ showed a similar trend (fold change = 38.22, p=0.085) although mRNA levels were much lower.\n\n![**Genes related to MD show differential H3K4me3 marks.***MX1* (**A**, **B**) and *CTLA-4* (**C**, **D**) show increased H3K4me3 marks and higher expression in infected individuals from both lines; *MMP2* (**E**, **F**) exhibits higher levels of H3K4me3 in susceptible line 7~2~. n = 4; \\* = significant at p \\< 0.05; \\*\\* = significant at p \\< 0.01; \\*\\*\\* = significant at p \\< 0.001.](1471-2164-13-557-2){#F2}\n\n*CTLA-4* is a cell surface glycoprotein expressed on CD4+ and CD8+ T lymphocytes that is a powerful negative regulator of T-cell activation \\[[@B18]\\]. The CTLA4 protein is expressed on T lymphocytes soon after activation and regulates T-cell proliferation, production of IL-2 and also supports the function of T~reg~ cells that suppress immune response \\[[@B19]\\]. Previous studies have reported an increase in *CTLA-4* expression after MDV infection \\[[@B20]\\]. We found an increase in H3K4me3 enrichment in line 7~2~ as a result of MDV infection (p = 0.0003) and there was a similar trend in line 6~3~ (Figure [2](#F2){ref-type=\"fig\"}C). Consistent with the above, there was a significant increase in transcript levels after MDV infection in line 7~2~ (p = 0.04) and a small increase in line 6~3~ (Figure [2](#F2){ref-type=\"fig\"}D).\n\n*MMP2* plays a key role in the degradation of the extra-cellular matrix, and an increase in expression has been associated with increasing tumor cell migration and tumor angiogenesis \\[[@B21],[@B22]\\]. *MMP2* was upregulated during the neoplastic stage of MD infection in susceptible birds \\[[@B23]\\] but downregulated in response to MDV infection during early lytic infection in susceptible and resistant chickens \\[[@B17]\\]. We observed a slight increase in H3K4me3 enrichment after MDV infection in both lines, while line 7~2~ exhibited significantly lower levels than line 6~3~ (p = 0.0016; Figure [2](#F2){ref-type=\"fig\"}E). This was coupled with increased *MMP2* expression in line 6~3~ after infection (p = 0.0068) while there was no such change in line 7~2~ (Figure [2](#F2){ref-type=\"fig\"}F).\n\nGenes related to cancers show epigenetic changes in response to MD\n------------------------------------------------------------------\n\nWe observed differential histone marks on several genes that have been associated with other cancers but not in the context of MDV infection. Insulin-like growth factor 2 binding protein 1 (*IGF2BP1*) is an RNA-binding factor that regulates the translation of mRNAs produced by certain genes like *IGF2* and *ACTB*. Increased expression of *IGF2BP1* has been implicated in the development and progression of cancers of various organs, e.g. lung, brain, breast and colon \\[[@B24]-[@B27]\\]. There was no change in the H3K4me3 enrichment levels induced by MDV infection although a significantly higher level of enrichment was present in line 7~2~ (p = 4.21x10^-13^; Figure [3](#F3){ref-type=\"fig\"}A). Transcript levels in line 7~2~ were much higher than in line 6~3~, but reduced in response to MDV infection (p = 0.044) (Figure [3](#F3){ref-type=\"fig\"}B).\n\n![**MD induces epigenetic changes in genes related to various cancers.***IGF2BP1* (**A**, **B**) shows differential H3K4me3 marks and increased expression in susceptible birds while *EAF2* (**C**, **D**) and *GAL* (**E**, **F**) have differential H3K27me3 levels on the gene body. n = 4; \\* = significant at p \\< 0.05; \\*\\* = significant at p \\< 0.01; \\*\\*\\* = significant at p \\< 0.001.](1471-2164-13-557-3){#F3}\n\nELL associated factor 2 (*EAF2*) is a testosterone regulated apoptosis inducer and tumor suppressor. Inactivation of *EAF2* has been shown to lead to tumors in multiple organs \\[[@B28]\\]. There was a significant increase in H3K27me3 levels after MDV infection in line 6~3~ (p = 0.0414) while among uninfected chickens these levels were markedly higher in line 7~2~ (p = 0.0138; Figure [3](#F3){ref-type=\"fig\"}C). However, *EAF2* expression was drastically reduced after MDV infection in line 7~2~ (p=0.0016) but showed only a small decrease in line 6~3~ (Figure [3](#F3){ref-type=\"fig\"}D).\n\nGalanin (*GAL*) is a neuropeptide that modulates various physiological functions, such as, inhibition of insulin secretion and stimulation of growth hormone secretion. Three galanin receptors are known (*GALR1*, *2* and *3*): the expression of *GALR1* has anti-proliferative effects while *GALR2* can be both anti- or pro-proliferative in function. Therefore, the *GAL* system is considered to be a promising candidate for detection and treatment of various cancers \\[[@B29],[@B30]\\]. We observed an increase in H3K27me3 levels on *GAL* after infection in both lines (Figure [3](#F3){ref-type=\"fig\"}E). Also, expression levels were significantly lowered after MDV infection in line 7~2~ (p = 0.00087) while there was a similar trend in line 6~3~ (p = 0.051; Figure [3](#F3){ref-type=\"fig\"}F). Interestingly, *GALR1* had both H3K4me3 and H3K27me3 enrichments (Figure [4](#F4){ref-type=\"fig\"}) although *GALR2* showed no significant histone marks (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S6).\n\n![**Significant H3K4me3 and H3K27me3 enrichment around*GALR1.*** The anti-proliferative *GAL* receptor *GALR1* exhibited both active and repressive marks. There is no change in H3K4me3 levels but a definite increase in H3K27me3 levels after infection in line 7~2~.](1471-2164-13-557-4){#F4}\n\nChromatin co-localization patterns reveal putative bivalent genes\n-----------------------------------------------------------------\n\nRegions of chromatin having both active and repressive marks are said to be bivalent and have been shown to play important roles in development and genetic imprinting \\[[@B31],[@B32]\\]. For example, bivalent domains have been shown to mark promoters of genes that are subsequently silenced in tumors by DNA hypermethylation indicating their importance in cancer \\[[@B33]\\]. A mono-allelic bivalent chromatin domain that controls tissue-specific genomic imprinting at a specific locus was recently found in mice \\[[@B32]\\]. To investigate the presence of such bivalent chromatin states and the possible effect of MDV infection, we defined bivalent genes as those having H3K4me3 reads (TSS \u00b1 500 bp) greater than 30 reads per million mapped reads (RPM) and H3K27me3 reads (gene body) greater than 2 RPM, respectively (\\~85^th^ percentile). This filtering process yielded a list of 99 putative bivalent genes (Additional file [4](#S4){ref-type=\"supplementary-material\"}).\n\nFunctional annotation clustering of the above genes using DAVID \\[[@B34],[@B35]\\] revealed significant enrichment of several immune-related functions. These included transcription factor *EGR1* which is reported to have tumor suppressor properties, genes involved in lymphocyte activation and differentiation such as *BCL6*, *CD4* and *SMAD3* and genes *TLR3* and *TIRAP* that are part of the toll-like receptor signaling pathway. Bivalent domains were also present on a variety of transcription factors with immune-related functions such as *CITED2*, a transactivator that regulates NF-\u03baB, *MYC* a transcription factor associated with hematopoetic tumors and *RHOB* a Ras family homolog that mediates apoptosis in tumor cells after DNA damage. Moreover, all genes involved in the top five functional annotation clusters showed higher chromatin levels in line 7~2~ primarily after MDV infection (Additional file [5](#S5){ref-type=\"supplementary-material\"}).\n\nBivalent domains are altered in response to MD\n----------------------------------------------\n\nWe further investigated the effect of MD on bivalent chromatin domains observed on *BCL6*, *CITED2*, *EGR1*, *CD4* and *TLR3* (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure [5](#F5){ref-type=\"fig\"} and S7). Interestingly, three of these genes, *CITED2, BCL6* and *EGR1*, showed identical epigenetic and transcriptional signatures.\n\n![**Bivalent domains on transcriptional regulators are altered by MD.** H3K4me3 (orange) and H3K27me3 (green) profiles and associated transcript levels of *BCL6* (**A**, **B**), *EGR1* (**C**, **D**) and *CITED2* (**E**, **F**). In all three cases we observe a slight increase in H3K27me3 induced by MDV infection in line 7~2~ and a concurrent significant decrease in transcript levels while increase in active and repressive marks appear to cancel each other out in line 6~3~.](1471-2164-13-557-5){#F5}\n\n*CITED2* is a member of the p300/CBP co-activator family that has intrinsic histone acetyltransferase activity and plays a major role in regulating and coordinating multiple complex cellular signals to determine the expression level of a gene \\[[@B36]\\]. B-cell CLL/lymphoma 6 (*BCL6*) is a zinc finger protein that functions as a transcriptional repressor which was downregulated at 15 dpi in spleen tissues from F1 progeny (15I~5~ X 7~1~) of MD-susceptible chickens \\[[@B20]\\]. *EGR1* belongs to a group of early response genes induced by a variety of signaling molecules such as growth factors, hormones and neurotransmitters that is believed to play a major role in cell proliferation and apoptosis \\[[@B37]\\]. Overexpression of *EGR1* promotes tumor growth in kidney cells \\[[@B38]\\] but suppresses growth and transformation in other cell types, e.g. fibroblasts and glioblastoma cells \\[[@B39]\\].\n\nIn each of the above genes, both active and repressive chromatin marks were increased in response to infection in line 6~3~ chickens. However, in line 7~2~, there was a definite increase in H3K27me3 marks but no change in H3K4me3 (Figures [5](#F5){ref-type=\"fig\"}A, C, E). Transcript levels were in agreement with this observation: infected line 7~2~ chickens showed a significant downregulation (*CITED2*: p=0.0004; *BCL6*: p=0.0048; *EGR1*: p=0.0005; Figures [5](#F5){ref-type=\"fig\"}B, D, F), while there were no such changes in line 6~3~.\n\nOn the other hand, *TLR3* and *CD4* showed a slight increase in H3K4me3 marks after MDV infection while there were no appreciable changes in H3K27me3 levels. In keeping with the epigenetic changes, there was a small increase in gene expression in infected birds from both lines (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S7).\n\nDiscussion\n==========\n\nImmune parameters that are known to play a major role in genetic resistance to MDV are correlated with innate immune responses, such as NK cell activity, production of nitric oxide and cytokines, such as, interferons. Recent studies have identified host cytokines such as IL-18 and IFN-\u03b3 that contribute to the initiation and continuation of latency \\[[@B40]\\]. However, cytokine levels can undergo rapid flux in response to infection, and consistent with this, we did not observe any epigenetic changes associated with these genes in the MHC-congenic lines used in our study (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S8). This suggests the existence of other extrinsic factors responsible for transcriptional variations between resistant and susceptible chickens at this stage of the disease.\n\nA global comparison of histone modification levels in the two inbred chicken lines yielded some interesting results. As expected, a majority of SERs were found in all the experimental groups, indicating a high level of epigenetic similarity between the lines in addition to inherent genetic similarity. In the case of H3K27me3, the percentage of ubiquitous SERs was relatively low (23.4%), although this was likely due to lower precision of the peak caller for broad chromatin marks. Besides, most of the SERs detected in a subset of samples corresponded to regions of low enrichment, which may also be the reason behind the relatively low number of differential SERs detected in our study.\n\nGenes that have been previously associated with MD and various other cancers showed differential marks that are either MD-induced (*MX1, CTLA-4, EAF2* and *GAL*) or line-specific (*IGF2BP1* and *MMP2*). The increase in H3K4me3 marks around the TSS of *MX1,* a gene with known antiviral properties, appeared to be correlated with upregulated expression in both lines in response to MDV infection. However, lowered overall mRNA levels in the resistant line suggest additional factors could be involved in the regulation of this gene. Similarly, increased mRNA levels of the lymphocyte surface marker *CTLA4* is possibly due to the presence of larger numbers of T cells in MDV infected birds as a result of higher levels of infection. *EAF2* functions as an apoptosis inducer in addition to being a tumor suppressor, and therefore, its downregulation could contribute to higher tumor incidence rates in line 7~2~. However, it is not clear why a significant increase in H3K27me3 levels did not have any effect on transcript levels in the resistant line.\n\n*IGF2BP1* is believed to act by stabilizing the mRNA of the *c-myc* oncogene and therefore, the higher expression of this gene and a more stable c-myc gene product might play a role in increasing MD susceptibility in line 7~2~ birds via increased cell proliferation and transformation. The matrix metalloprotease *MMP2* is upregulated after infection in the resistant line 6~3~, similar to the previously observed increase at the neoplastic stage of MD. However, mRNA levels were similar in the two lines before infection contrary to the difference in H3K4me3 levels suggesting that additional factors are responsible for regulating this gene.\n\nThe correlation between observed differential histone marks and transcript levels was moderate at best. Indeed, differential H3K4me3 marks were strongly predictive of gene expression levels but the correlation between H3K27me3 and mRNA levels was relatively poor. There could be various reasons for this -- indeed, H3K27me3 levels had a non-linear relationship with gene expression with higher marks showing little difference in the effect on expression. Therefore, in this tissue, the levels of H3K27me3 may not be a very good indicator of gene expression levels. Also, the transcription of these genes might be controlled by other factors with the change in H3K27me3 levels only incidental.\n\nBivalent domains were detected on transcriptional regulators *BCL6*, *CITED2* and *EGR1* and the galanin receptor *GALR1.* The epigenetic and transcriptional signatures observed on these genes indicated that they were poised at the latent stage of the disease, but with crucial differences in the two lines. Increased repressive marks in the susceptible line correlated with significant downregulation of the genes, while in line 6~3~, the increase in both marks appeared to compensate for each other with no eventual effect on gene transcription. This suggests that some 'poised' bivalent genes can become highly repressed even with a relatively small increase in H3K27me3 marks. The change in the chromatin levels could also be correlated with an increase in cell populations having the repressive mark. Taken together, the above findings point towards the existence of finely balanced epigenetic control of transcription, which may be necessary to mount a rapid response by the immune system. However, this machinery could potentially be hijacked by a pathogen and result in an aberrant phenotype. The effect of MDV infection on the bivalent domain on *GALR1*, in particular, suggests the repression and potential loss of its anti-proliferative effects. Thus, the galanin system possibly plays an important and hitherto unknown role in MD progression and could be a novel target for long-term control of the disease.\n\nOne of the major players in MDV-induced malignant transformation is Meq, a virus-encoded oncoprotein that has diverse functions, e.g. transactivation, chromatin remodeling and regulation of transcription. Meq interacts with and sequesters the tumor suppressor protein p53, resulting in the dysregulation of cell-cycle control \\[[@B7]\\] and inhibition of the transcriptional and apoptotic activities of the protein \\[[@B41]\\]. Several genes that show epigenetic changes in response to MDV infection have been associated with p53. Downregulation of *CITED2* stabilizes the p53 protein leading to its accumulation \\[[@B42]\\]. The *BCL6* gene product is believed to contribute to lymphomagenesis by inactivation of p53 \\[[@B43]\\]. Besides, *EAF2* has also been shown to interact with and suppress the function of p53 \\[[@B28]\\]. The downregulation of all of the above genes in susceptible birds after MDV infection points towards the increased production of p53 and a robust anti-tumor response. That we still observe higher tumor incidence rates in this line, suggests the presence of large amounts of inactivating viral Meq protein which, in turn, indicates that increased numbers of MD-infected cells are present in the susceptible line at this stage of the disease. A majority of tumors are believed to result from the viral transformation of CD4+ T cells some of which are infected at the latent stage of MD \\[[@B44]\\]. The larger number of virus-infected cells produced in the susceptible line is possibly due to lowered immunocompetence as a result of the early stages of infection. Thus, a more detailed investigation of the early cytolytic stage of MD is necessary to shed further light on the causes behind the divergent response to MD observed in these birds.\n\nWhole tissues represent a mixture of various cell populations, and observed epigenetic changes might be due to a change in chromatin marks in a particular cell type or a variation in the relative number of cells carrying active or repressive histone marks. However, such *in vivo* studies are representative of the host response at a systems level wherein different cell types might interact in a cooperative manner to fight infection. Thus, while the study of pure cell populations is likely to yield greater discriminative power, the investigation of tissue macroenvironments is, perhaps, closer to reality.\n\nThis study focused on the thymus tissue as it is a major immune organ and contains a significant population of T lymphocytes in various stages of differentiation. Our earlier study of the MDV-induced transcriptome in these birds indicated the presence of line-specific differences at the latent stage of MD \\[[@B45]\\]. In addition, birds susceptible to MD suffer thymic atrophy during the early stages of infection \\[[@B46]\\], indicating the importance of understanding changes in this organ to elucidate the mechanisms involved in disease progression. Ongoing studies in our lab include other tissues, e.g. spleen \\[[@B47]\\], and a time-course through the various stages of infection, to further investigate the systemic effects of MD and the epigenetic basis of MD resistance.\n\nConclusions\n===========\n\nWe studied the effect of latent MDV infection on two chromatin modifications in inbred chicken lines exhibiting different degrees of resistance to MD. Several genes showed changes in histone enrichment and this response was often significantly different between the two chicken lines. A detailed analysis of co-localization patterns of the chromatin marks revealed the presence of bivalent domains on a number of immune-related transcriptional regulators. More importantly, these domains showed marked changes in response to MDV infection and provide further evidence of the far-reaching epigenetic effects of MD. Our results suggest putative roles for the *GAL* system in MD progression. A majority of the differential chromatin marks are also suggestive of increased levels of viral infection in the susceptible line symptomatic of lowered immunocompetence in these birds at early stages of the disease. In summary, our study provides further insight into the mechanisms of MD progression while revealing a complex landscape of epigenetic regulatory mechanisms. Further work is necessary to fully elucidate the underlying mechanisms of MD, but our results suggest that this is a promising step towards a deeper understanding of this complex disease.\n\nMethods\n=======\n\nAnimals and viruses\n-------------------\n\nTwo specific-pathogen-free inbred lines of White Leghorn either resistant (6~3~) or susceptible (7~2~) to MD were hatched, reared and maintained in Avian Disease and Oncology Laboratory (ADOL, Michigan, USDA). Four chickens from each line were injected intra-abdominally with a partially attenuated very virulent plus strain of MDV (648A passage 40) at 5 days after hatch with a viral dosage of 500 plaque-forming units (PFU). Infected and control chickens from both lines (n = 4) were terminated at 10dpi to collect thymus tissues. All procedures followed the standard animal ethics and use guidelines of ADOL.\n\nQuantification of MDV loads in thymus\n-------------------------------------\n\nThe MDV gene *ICP4* was used for quantification of viral genomic DNA in thymus as previously described \\[[@B48]\\]. Quantitative PCR was performed by using 100 ng/\u03bcl of genomic DNA on the iCycler iQ PCR system (Bio-Rad, USA) and QuantiTect SYBR Green PCR Kit (Qiagen, USA) (Additional file [1](#S1){ref-type=\"supplementary-material\"}: Figure S9). The relative MDV loads were determined by normalizing to a single-copy gene *Vim* (vimentin) \\[[@B49]\\]. The primers for *Vim* are as follows: Forward: 5'-CAGCCACAGAGTAGGGTAGTC-3'; Reverse: 5'-GAATAGGGAAGAACAGGAAAT-3'.\n\nChromatin Immunoprecipitation and Illumina Sequencing\n-----------------------------------------------------\n\nChromatin immunoprecipitation (ChIP) was carried out using thymus samples from MDV infected and controls birds \\[[@B50]\\]. About 30 mg thymus samples from three individuals were cut into small pieces (1 mm^3^) and digested with MNase to obtain mononucleosomes. PNK and Klenow enzymes (NBE, Ipswich, MA, USA) were used to repair the ChIP DNA ends pulled down by the specific antibody. A 3\u2032 adenine was then added using Taq polymerase and a pair of Solexa adaptors (Illumina, USA) ligated to the repaired ends. Seventeen cycles of PCR was performed on ChIP DNA using the adaptor primers and fragments with a length of about 190 bp (mononucleosome + adaptors) were isolated from agarose gel. Subsequently, cluster generation and ChIP-sequencing (ChIP-Seq) using the purified DNA was performed on the Solexa 1G Genome Analyzer (Illumina, USA) following manufacturer protocols. The antibodies used and the total number of reads obtained for each sample is listed in Additional file [6](#S6){ref-type=\"supplementary-material\"}.\n\nRead mapping and summary counts\n-------------------------------\n\nSequence reads obtained from the Illumina 1G Genome Analyzer were aligned to the May 2006 version of the chicken genome (galGal3) using Maq version 0.7.1 \\[[@B51]\\]. Default alignment policies of Maq were enforced: a valid alignment could have a maximum of two mismatches and if a read aligned equally well to multiple places in the genome, one was chosen at random. If multiple reads mapped to the same genomic location only one was kept to avoid amplification bias. Summary read counts were calculated using non-overlapping windows of 200 bp for visualization and normalized to per million mapped reads in each sample for the purpose of comparisons.\n\nIdentification of enriched regions\n----------------------------------\n\nSummarized read counts were subjected to peak calling with SICER \\[[@B52]\\]. The source code was modified to include support for the chicken genome. Fragment length was specified to be 190. A window size of 200 bp and gap size of 400 bp was used for the analysis. The E-value for estimating significant peaks was set to 100. For the purposes of comparing different samples, SERs found in similar genomic regions of different samples were merged to obtain a consolidated list as follows: SERs from one sample were used to initialize the list. For each such region *M,* we searched for overlapping SERs in the next sample. In the case of an overlap between *M* and a significant region, *S*, the merged region was updated to include both *M* and *S*. This procedure was iterated over all samples to obtain a consolidated list of merged SERs.\n\nGene annotation and genomic distribution of SERs\n------------------------------------------------\n\nRefSeq and Ensembl gene annotations were downloaded from UCSC genome browser \\[[@B53]\\]. As there were only 4306 RefSeq genes in the database, we included Ensembl genes in our analysis to improve genome-wide coverage. There were 17858 annotated genes in the Ensembl database, which include validated and predicted genes. Redundancies between the databases were listed and accounted for, yielding a non-redundant list of 18198 genes with 4306 RefSeq genes and 13892 Ensembl genes. We then searched for overlaps between merged SERs and the non-redundant list of annotated genes. For H3K4me3, an SER was annotated with a gene if it overlapped the TSS region of the gene whereas in the case of H3K27me3, a valid overlap constituted an SER overlapping the gene body. To calculate the genomic distribution we counted all SERs having an overlap with one of the following regions: promoter (TSS \u00b1 1 kb), exons, introns, 5' UTR and 3' UTR.\n\nHistone modification profiles and differential chromatin marks\n--------------------------------------------------------------\n\nGenes were divided into 10 sets based on their absolute expression and representative sets corresponding to high, medium, low and no expression were chosen for visualization. We defined the gene body as the region between the transcription start site (TSS) and the transcription termination site (TTS). Histone modification profiles surrounding the gene body were calculated in 3 distinct regions: 5000 bp upstream of the 5' end, gene body and 5000 bp downstream of the 3' end of the gene. For reads falling within the gene body, read counts were obtained in bins 5% of the gene length while 1000 bp windows were used for the 5' and 3' flanking regions. The read counts in all cases were normalized to the total number of genes in the categories and total number of reads in the sample. We also compared gene expression to histone modification levels by plotting normalized microarray data (Zhang, H. unpublished data) against reads mapping to (i) TSS \u00b1 500 bp and (ii) the gene body for each gene.\n\nReads mapping to merged SERs were tested for epigenetic changes induced by MDV infection in lines 6~3~ and 7~2~ using DESeq \\[[@B54]\\]. We used the method 'blind' for variance estimation and p-values were corrected for multiple testing using the Benjamini-Hochberg FDR procedure \\[[@B55]\\]. Statistical significance was defined using a false discovery rate (FDR) threshold of 0.4.\n\nValidation of ChIP, ChIP-Seq and gene transcription by Q-PCR\n------------------------------------------------------------\n\nQuantitative real-time RT-PCR was used to validate the quality of the ChIP and gene transcript levels on the iCycler iQ PCR system (Bio-Rad, Hercules, CA, USA). The real-time RT-PCR reactions were performed with a QuantiTect SYBR Green PCR Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. An online primer system ( <http://frodo.wi.mit.edu/primer3/>) was used to design the primers and four biological and four technical replicates were performed. The primer sequences are shown in Additional file [7](#S7){ref-type=\"supplementary-material\"}.\n\nData access\n-----------\n\nRaw and processed ChIP-Seq data discussed in this manuscript were deposited in the NCBI Gene Expression Omnibus (GEO) ( <http://www.ncbi.nlm.nih.gov/geo/>) under Series accession number GSE24017.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthor contributions\n====================\n\nAM performed the data analysis and wrote the manuscript. JL performed the ChIP experiments, sequence library preparation, validation of ChIP-Seq results and revised the manuscript. HMZ collected samples and revised the manuscript. JZS designed the experiments and revised the manuscript. All authors read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\nFigures S1 to S9. Supplementary Figures S1 to S9.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**Differential H3K4me3 marks. Genome-wide differential H3K4me3 marks produced by DESeq (FDR \\< 0.4) and the associated genes.** P-values from three contrasts are displayed as follows: 63: 63I vs 63N, 72: 72I vs 72N, 63v72N: 63N vs 72N. 63I: line 6~3~ infected, 63N: line 6~3~ control, 72I: line 7~2~ infected, 72N: line 7~2~ control.\n\n###### \n\nClick here for file\n\n###### Additional file 3\n\n**Differential H3K27me3 marks.Genome-wide differential H3K27me3 marks produced by DESeq (FDR \\< 0.4) and the associated genes.** P-values from three contrasts are displayed as follows: 63: 63I vs 63N, 72: 72I vs 72N, 63v72N: 63N vs 72N. 63I: line 6~3~ infected, 63N: line 6~3~ control, 72I: line 7~2~ infected, 72N: line 7~2~ control.\n\n###### \n\nClick here for file\n\n###### Additional file 4\n\nPutative bivalent genes from colocalization analysis of H3K4me3 and H3K27me3.\n\n###### \n\nClick here for file\n\n###### Additional file 5\n\nFunctional annotation clustering of bivalent genes using DAVID.\n\n###### \n\nClick here for file\n\n###### Additional file 6\n\nSequencing results showing raw and mapped reads for each sample.\n\n###### \n\nClick here for file\n\n###### Additional file 7\n\nPrimers used for quantitative PCR validation.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis project was supported by National Research Initiative Competitive Grant no. USDA-NRI/NIFA 2008-35204-04660 and USDA-NRI/NIFA 2010-65205-20588 from the USDA National Institute of Food and Agriculture.\n"}
+{"text": "Introduction\n============\n\nOvarian cancer is the fifth-leading cause of worldwide cancer-related deaths among woman. It accounts for more deaths than other gynecologic tumors. Standard therapy for advanced ovarian cancer consists of cytoreductive surgery followed by chemotherapy ([@b1-mmr-18-04-3898]). The 5-year survival rate for ovarian cancer patients ranges from 30 to 92% ([@b2-mmr-18-04-3898]). Paclitaxel-platinum treatment is the standard first-line treatment in ovarian cancer, with typical response rate of over 70%. However, most patients eventually experience recurrence, with a median progression-free survival of 18 months ([@b2-mmr-18-04-3898]). Although efforts have been focused on overcoming resistance to chemotherapy drugs for years, mortality rates of patients with ovarian cancer remain high ([@b3-mmr-18-04-3898]).\n\nExtracellular signal-regulating kinase (ERK) 1/2 pathway plays a major role in the survival of cancer cells via inhibition of apoptosis controlling Bcl-2 family members such as Bim, Bad, Bcl-1, and Mcl-1 ([@b4-mmr-18-04-3898]). However, depending on cell type and stimulus, ERK1/2 activation can also mediate antiproliferative events such as apoptosis and autophagy in some studies ([@b4-mmr-18-04-3898],[@b5-mmr-18-04-3898]). ERK1/2 phosphorylates and activates direct downstream kinase p90 ribosomal S6 kinase (RSK). ERK1/2/p90RSK activation is involved in the progression of many cancers ([@b6-mmr-18-04-3898]). It is emerging as a potential therapeutic target ([@b6-mmr-18-04-3898]). The purpose of this study was to explore whether ERK1/2/p90RSK activation could mediate anticancer mechanism in ovarian cancer cells.\n\n*Croton tonkinesis* Gagnep. (Euphorbiaceae), commonly known as 'Kho sam cho la (or Kho sam Bac Bo)' in Vietnamese, is a small medicinal plant indigenous to northern Vietnam ([@b7-mmr-18-04-3898]). This plant has been used to treat stomach aches, abscesses, impetigoes, gastric and duodenal ulcers, and malaria ([@b8-mmr-18-04-3898]). Crude extract of *C. tonkinensis* has shown significant cytotoxicity against breast cancer, lung cancer, and glioblastoma ([@b9-mmr-18-04-3898]). Some *ent*-kaurane diterpenoids isolated from *C. tonkinensis* also possess cytotoxic and proapoptotic activities ([@b7-mmr-18-04-3898],[@b10-mmr-18-04-3898],[@b11-mmr-18-04-3898]). They can inhibit lipopolysaccharide (LPS)-induced nuclear factor-\u03baB (NF-\u03baB) activation with anti-inflammatory properties ([@b12-mmr-18-04-3898],[@b13-mmr-18-04-3898]). They can also inhibit silent information regulator two ortholog 1 (SIRT1) and stimulate osteoblast differentiation ([@b14-mmr-18-04-3898],[@b15-mmr-18-04-3898]). However, the underlying mechanism by which *ent*-kaurane diterpenoids inhibit cancer cell growth remains controversial. Multiple signaling mechanisms, including PI3K/PKB pathway, AMP-activated protein kinase pathway, Wnt/\u03b2-catenin pathway, and Akt/mTOR/p70S6K pathway, have been implicated in anticancer activities of this class of diterpenoids ([@b16-mmr-18-04-3898]--[@b20-mmr-18-04-3898]).\n\nThe objective of the present study was to investigate the effect of a natural *ent*-kaurane diterpenoid CRT1 isolated from *C. tonkinensis* on proliferation, apoptosis, migration, and invasion of human ovarian cancer cells. Molecular mechanisms associated with cellular changes were also examined in this study.\n\nMaterials and methods\n=====================\n\n### Plant material and chemical isolation\n\nThe major *ent*-kauranoid was previously isolated from leaves of plant *Croton tonkinensis* and identified as *ent*\u221218-acetoxy-7*\u03b2*-hydroxy kaur-15-oxo-16-ene (namely CRT1) ([Fig. 1A](#f1-mmr-18-04-3898){ref-type=\"fig\"}). Purification and structure identification of CRT1 were demonstrated elsewhere ([@b13-mmr-18-04-3898]). The purity of the compound was determined to be 97% by HPLC.\n\n### Cell culture\n\nHuman ovary adenocarcinoma SKOV3 was purchased from Korean Cell Line Bank (Seoul, Korea). McCoy\\'s 5A, fetal bovine serum (FBS), and penicillin/streptomycin were obtained from Gibco (Thermo Fisher Scientific, Inc., Waltham, MA, USA). Trypsin/EDTA was purchased from Thermo Scientific HyClone (GE Healthcare Life Sciences, Logan, UT, USA). Cells were grown in McCoy\\'s 5A media supplemented with 10% (v/v) FBS, penicillin (100 U/ml)/streptomycin (100 \u00b5g/ml) at 37\u00b0C in a humidified CO~2~ (5%)-controlled incubator.\n\n### Cell viability assay\n\nCells were seeded into 96-well microplates at density of 5\u00d710^3^ cells/ml and allowed to attach for 24 h. CRT1 and/or PD98059 were added to the medium at various concentrations. After treatment, cell cytotoxicity and/or proliferation was assessed using Cell Counting Kit-8 (CCK-8; Dojindo Laboratories, Kumamoto, Japan). Briefly, highly water-soluble tetrazolium salt, WST-8\\[2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt\\], produced an orange colored water-soluble product, formazan. The amount of formazan dye generated by dehydrogenases in cells was directly proportional to the number of living cells. CCK-8 (10 \u00b5l) was added to each well and incubated at 37\u00b0C for 3 h. Cell proliferation and cytotoxicity were then assessed by measuring the absorbance at wavelength of 450 nm using a microplate reader. Three replicated wells were used for each experimental condition.\n\n### Western blot analysis\n\nCells were incubated with CRT1 (10, 25, and 50 \u00b5M) and/or PD98059 (25 \u00b5g/ml) for 24 h, and washed twice in cold phosphate buffered saline (PBS). Cells were lysed with lysis buffer \\[10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% TritonX-100, 0.5% NP-40, 1 mM propidium iodide (PI), 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF)\\], placed on ice for 1 h with occasional vortexing, and centrifuged at 13,000 \u00d7 g for 10 min at 4\u00b0C to collect the supernatant. Pierce BCA Assay Kit (Pierce; Thermo Fisher Scientific, Inc.) was used to determine protein concentration. Cell lysates (50 \u00b5g) were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene difluoride (PVDF) membrane. Blots were blocked with 5% skim milk in PBS containing 0.05% Tween-20 (PBST) for 1 h at 25\u00b0C and then incubated with primary antibodies (1:1,000) overnight at 4\u00b0C. After washing with PBST, membranes were incubated with anti-rabbit or anti-mouse horseradish peroxidase-conjugated IgG (1:3,000) at room temperature for 2 h and visualized with enhanced chemiluminescence using Super Signal^\u00ae^ West Pico Chemiluminescent substrate purchased from Pierce; Thermo Fisher Scientific, Inc. Antibodies to rabbit polyclonal anti-human Bax (1:1,000; no. 2772), rabbit polyclonal anti-human Bcl-2 (1:1,000; no. 2876), rabbit polyclonal anti-human caspase-3 (1:1,000; no. 9662), mouse monoclonal anti-human cspase-7 (1:1,000; no. 9494), rabbit polyclonal anti-human caspase-9 (1:1,000; no. 9502), rabbit polyclonal anti-human poly-(ADP-ribose) polymerase (PARP) (1:1,000; no. 9542), rabbit polyclonal anti-human phospho-p44/p42 MAPK (ERK1/2) (Thr202/Tyr204) (1:1,000; no. 9101), rabbit monoclonal anti-human RSK1/2/3 (1:1,000; no. 9355), rabbit polyclonal anti-human phospho-p90RSK (Ser380) (1:1,000; no. 9314), and rabbit polyclonal anti-human \u03b2-actin (1:1,000; no. 4967) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Rabbit polyclonal anti-human ERK (1:1,000; sc-94) was obtained from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Horseradish peroxidase-conjugated anti-mouse and anti-rabbit antibodies were bought from Transduction Lab (Lexington, KY, USA). Band intensity was quantified by densitometry using ImageJ software and was normalized to loading controls. Quantification value was expressed as the fold change vs. band numbered 1.0. ImageJ was downloaded from NIH website (<http:rsbweb.nih.gov/ij/download.html>).\n\n### Annexin V-FITC/PI double staining assay\n\nAnnexin-V-FLUOS Staining kit was purchased from Roche Diagnostics GmbH (Penzberg, Germany). Cells were cultured in 6-well plates at cell density of 10^6^ cells/well in McCoy\\'s 5A medium and treated with CRT1 (10, 25, and 50 \u00b5M) and/or PD98059 (25 \u00b5M) for 24 h. Cells were centrifuged and washed three times with PBS. Cell pellet was then resuspended in 100 \u00b5l of Annexin V-FLUOS labeling solution. After incubating at room temperature for 30 min, samples were analyzed on a flow cytometer (BD FACSCanto^\u2122^II; BD Biosciences, Franklin Lakes, NY, USA).\n\n### Soft agar colony forming assay\n\nFor determination of anchorage-independent cell growth, 10^4^ cells were suspended in growth media supplemented with 10% FBS (3 ml) containing 0.3% agar. They were then applied onto pre-solidified 0.6% agar (3 ml) in FBS-free growth media in 60 mm culture dishes. After 2--3 weeks of incubation, colonies on soft agar were observed under a phase-contrast microscope (IX2-ILL100; Olympus Corporation, Tokyo, Japan) and photographed.\n\n### Wound healing assay\n\nCells were seeded into 6-well plates and incubated in serum-free McCoy\\'s 5A for 18 h. The cellular monolayer was wounded with a sterile 10 \u00b5l-pipette tip and washed with serum free McCoy\\'s 5A to remove detached cells from plates. These cells were incubated in the presence or absence of CRT1 for 48 h in McCoy\\'s 5A containing 10% FBS. The medium was replaced with PBS and cells were photographed using a phase-contrast microscope (IX2-ILL100; Olympus Corporation).\n\n### Matrigel invasion assay\n\nCell invasion assay was carried out with 24-well flat bottom plate with transparent PET membrane with pore size of 8.0 \u00b5m (FALCON^\u00ae^; Corning Incorporated, Corning, NY, USA). Each insert has been coated with Matrigel Matrix (BD Bioscience, San Jose, CA, USA). Cells (2.5\u00d710^4^) suspended in 300 \u00b5l of serum free McCoy\\'s 5A with or without drugs were added to the upper chamber while 500 \u00b5l of McCoy\\'s 5A containing 10% FBS were added to the lower chamber of 24-well flat bottom plate. After incubation for 24 h, non-invading cells were removed from the upper surface of the membrane by scrubbing while invading cells on the lower surface of the membrane were stained with hematoxylin. Membranes were then removed and invading cells were counted randomly by light microscopy. Each assay was performed in triplicates and repeated at least three times. Due to variation in the number of migrated cells from different experiment, results were normalized to control cells and relative invasion was expressed as mean \u00b1 SD of migrating cells relative to control cells.\n\n### Statistical analysis\n\nAll results presented were confirmed in at least three independent experiments. Data were presented as the mean \u00b1 standard deviation. Statistical differences were analyzed by one-way analysis of variance followed by a Tukey post hoc test using SPSS 24.0 software (IBM Corp., Armonk, NY, USA). P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\n### CRT1 reduces SKOV3 cellular proliferation in a dose- and time-dependent manner\n\nTo evaluate effects of CRT1 on growth, SKOV3 ovarian carcinoma cells were treated with increasing concentrations (1 to 100 \u00b5M) of CRT1 for various time periods (12, 12, and 48 h) and cell viabilities were assessed by CCK-8 assay. Viabilities of SKOV3 cells were decreased in a dose- and time-dependent manner after exposure to different concentrations of CRT1 ([Fig. 1B and C](#f1-mmr-18-04-3898){ref-type=\"fig\"}).\n\n### Mitochondrial pathway of apoptosis induced by CRT1 treatment of SKOV3 cells\n\nTo investigate the occurrence of apoptosis following CRT1 treatment, flow cytometric analysis was performed using a Annexin V-FITC/PI double staining assay. After SKOV3 cells were treated with 25 and 50 \u00b5M CRT1 for 24 h, the percentage of Annexin V-positive cells was increased from 0.4% (in the control) to 29.7 and 57.8%, respectively ([Fig. 2A and B](#f2-mmr-18-04-3898){ref-type=\"fig\"}). The percentage of necrotic cells was increased from 0.7% (in the control) to 20.5% (10 \u00b5M CRT1), 20.6% (25 \u00b5M CRT1), and 36.8% (50 \u00b5M CRT1) ([Fig. 2A](#f2-mmr-18-04-3898){ref-type=\"fig\"}). Effects of CRT1 treatment on the expression of apoptotic proteins, Bax and Bcl-2, were also investigated by western blot analysis. As shown in [Fig. 2C](#f2-mmr-18-04-3898){ref-type=\"fig\"}, expression levels of proapoptotic protein Bax were significantly increased in SKOV3 cells following treatment with CRT1 at 25 and 50 \u00b5M for 24 h compared to those in the control. However, expression levels of anti-apoptotic protein Bcl-2 were decreased in cells treated with CRT1 at 25 and 50 \u00b5M for 24 h compared to those in the control. Bax is a key proapoptotic molecule in the mitochondrially mediated apoptotic pathway. It is responsible for pore-opening of the mitochondria to release cytochrome c and activate caspase family of proteases as part of the apoptotic cascade ([@b21-mmr-18-04-3898]). Therefore, we examined the effect of CRT1 on the release of cytochrome c and caspase-3, \u22127, \u22129, and poly(ADP-ribose) polymerase (PARP) activation. CRT1 significantly enhanced cytochrome c release from the mitochondria to the cytosol ([Fig. 2D](#f2-mmr-18-04-3898){ref-type=\"fig\"}), leading to increased expression of cleaved caspase-3, cleaved caspase-7, and cleaved PARP but decreased expression of procaspase-9 compared with the control ([Fig. 2E](#f2-mmr-18-04-3898){ref-type=\"fig\"}). Pretreatment of SKOV3 cells with z-VAD-fmk, a caspase inhibitor, attenuated the ability of CRT1 to activate the caspase cascade ([Fig. 2E](#f2-mmr-18-04-3898){ref-type=\"fig\"}).\n\n### CRT1 increases ERK/p90RSK phosphorylation of SKOV3 cells\n\nTo determine whether the MAPK/ERK1/2 pathway was involved in the anticancer effect of CRT1, this study examined phosphorylation of ERK1/2 and its substrate, p90RSK, following treatment with CRT1. CRT1 significantly increased phosphorylation levels of ERK1/2 at Thr202/Tyr204 sites and p90RSK at Ser380 site in a dose- and time-dependent manner. However, there was no significant difference in total expression levels of ERK1/2 and p90RSK proteins between control and CRT1 treatment ([Fig. 3A and B](#f3-mmr-18-04-3898){ref-type=\"fig\"}). PD98059, a specific ERK inhibitor, failed to increase CRT1-induced ERK1/2 activation ([Fig. 3C](#f3-mmr-18-04-3898){ref-type=\"fig\"}) or suppress CRT1-induced cell viability ([Fig. 3D](#f3-mmr-18-04-3898){ref-type=\"fig\"}).\n\n### CRT1-induced anti-proliferation and apoptosis is reversed by ERK inhibitor\n\nTo study whether ERK1/2 was involved in CRT1-induced apoptosis of SKOV3 cells, PD98059, a specific ERK inhibitor, was used. PD98059 treatment failed to increase CRT1-induced caspase-3, caspase-7, and PARP cleavages ([Fig. 4A](#f4-mmr-18-04-3898){ref-type=\"fig\"}) or the percentage of annexin V-positive cells ([Fig. 4B and C](#f4-mmr-18-04-3898){ref-type=\"fig\"}). To evaluate the reversed effect of PD98059 on CRT1-induced cell growth, viability of SKOV3 cells was assessed by CCK8 cell-counting assay. CRT1 failed to decrease cell viability following pretreatment of SKOV3 cells with PD98059 ([Fig. 4D](#f4-mmr-18-04-3898){ref-type=\"fig\"}).\n\n### CRT1 inhibits migration and invasion through ERK1/2 activation in SKOV3 cells\n\nResults from colony survival assays of CRT1 are presented as photographs. Results showed that activation of ERK1/2 by 50 \u00b5M CRT1 treatment strongly diminished colony sizes ([Fig. 5A](#f5-mmr-18-04-3898){ref-type=\"fig\"}). To evaluate the effect of CRT1 on cell migration, wound healing assay and transwell migration assay were performed. Results showed that CRT1 (50 \u00b5M) suppressed the migration of SKOV3 cells ([Fig. 5B](#f5-mmr-18-04-3898){ref-type=\"fig\"}). Consistent with this result, transwell invasion assay also showed that 50 \u00b5M CRT1 significantly weakened the invasion capacity of SKOV3 cells ([Fig. 5C and D](#f5-mmr-18-04-3898){ref-type=\"fig\"}). To investigate whether ERK1/2 was involved in CRT1-induced metastasis of SKOV3 cells, PD98059, a specific ERK inhibitor, was used in the experiment. PD98059 reversed CRT1-induced migration and invasion ([Fig. 5B-D](#f5-mmr-18-04-3898){ref-type=\"fig\"}) as well as colony-forming ability ([Fig. 5A](#f5-mmr-18-04-3898){ref-type=\"fig\"}).\n\nDiscussion\n==========\n\nA variety of substances such as those present in dietary and medicinal plants all over the world have inhibitory effects on several cancers. The genus *Croton* consists of about 800 species that are widely distributed throughout tropical and subtropical regions. About 31 of these medicinal plant species are cultivated or grow wild in northern Vietnam ([@b9-mmr-18-04-3898]). An *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene is one of main substances isolated from *Croton tonkinensis*. A prior report has shown that this *ent*-kaurane diterpenoid compound can regulate cell viability of SK-HEP1 hepatoma cancer cell line at IC~50~ of \\<5 \u00b5M ([@b18-mmr-18-04-3898]). Another *ent*-kaurane diterpenoid compound, oridonin, can suppress the proliferation of human ovarian cancer cells such as SKOV3, OVCAR-3 and A2780 cells at IC~50~ values of 17.21, 13.9, and 12.1 \u00b5M, respectively ([@b22-mmr-18-04-3898]). Our results showed that CRT1, an *ent*-kaurane diterpenoid compound, significantly reduced proliferation of SKOV3 ovarian cancer cells, with IC~50~ value of 24.6 \u00b5M ([Fig. 1A](#f1-mmr-18-04-3898){ref-type=\"fig\"}). Several papers have suggested that *ent*-kaurane diterpenoid possesses apoptotic effects on various cancers such as colorectal carcinoma, pancreatic adenocarcinoma, esophageal squamous cell carcinoma, hepatocellular carcinoma, and ovarian cancer ([@b23-mmr-18-04-3898]--[@b26-mmr-18-04-3898]). Consistent with these reports, our study showed that CRT1 increased apoptosis of SKOV3 cells ([Fig. 2](#f2-mmr-18-04-3898){ref-type=\"fig\"}). Besides, the percentages of necrotic cells increased after treatment, suggesting that CRT1 could induce necrosis in SKOV3 cells ([Fig. 2A](#f2-mmr-18-04-3898){ref-type=\"fig\"}). Prior reports have indicated that ent-kaurane diterpenoids can produce antiproliferation effects on cancer cells by induction of apoptosis as well as necrosis ([@b27-mmr-18-04-3898],[@b28-mmr-18-04-3898]).\n\nTo further understand the mechanisms in the anticancer activity of CRT1 against ovarian cancer, we analyzed the effect of CRT1 on ERK1/2/p90RSK signaling pathway. Mitogen-activated protein kinases (MAPKs) are a family of serine/threonine protein kinases that include ERK, JNK, and p38. These kinases are involved in the activation of nuclear transcription factors that control cellular proliferation, cell cycle progression, apoptosis, and cell migration ([@b29-mmr-18-04-3898]). RSKs, a family member of serine/threonine protein kinases, are activated by ERK1/2 in response to many growth factors and hormones. ERK/1/2/p90RSK pathway is known to modulate a variety of cellular processes, including cell proliferation, survival, motility, and invasiveness in ovarian cancers ([@b30-mmr-18-04-3898]--[@b32-mmr-18-04-3898]). However, how this pathway activates proapoptotic signaling remains largely unknown. Oridonin, one *ent*-kaurane diterpenoid, has been previously reported to be able to induce phosphorylation of ERK, JNK, and p38 MAPK ([@b33-mmr-18-04-3898]). Our study revealed that CRT1 dramatically increased phosphorylation level of ERK1/2 in a dose-dependent ([Fig. 3A](#f3-mmr-18-04-3898){ref-type=\"fig\"}) and time-dependent ([Fig. 3B](#f3-mmr-18-04-3898){ref-type=\"fig\"}). However, no significant activation of JNK or p38 was observed under the same condition (data not shown). A specific ERK inhibitor, PD98059, prevented the activation of ERK1/2 in CRT1-induced cells ([Fig. 3C and D](#f3-mmr-18-04-3898){ref-type=\"fig\"}). Moreover, PD98059 blocked CRT1-mediated increase in cell growth ([Fig. 4D](#f4-mmr-18-04-3898){ref-type=\"fig\"}) and apoptosis ([Figs. 4A-C](#f4-mmr-18-04-3898){ref-type=\"fig\"}). These results strongly support that CRT1-induced anticancer activities are regulated by ERK1/2 activation in SKOV3 ovarian cancer cells. A prior report has suggested that ERK1/2 and p90RSK signaling can dynamically regulate cell motility in cancers ([@b34-mmr-18-04-3898]). We were interested in investigating the role of ERK1/2 activation in SKOV3 cells in relation to colony formation, migration, and invasion. Our results revealed that CRT1-induced ERK1/2 activation inhibited ovarian cancer cell metastasis while ERK inhibitor PD98059 prevented CRT1-induced anti-migration and anti-invasive activity of SKOV3 cells ([Fig. 5](#f5-mmr-18-04-3898){ref-type=\"fig\"}). These results suggest that CRT1 plays an important role in cell proliferation, migration, and the invasive potential of ovarian cancer cells via ERK1/2 activation.\n\nPrimary cells are believed to be more biologically relevant tools than cell lines for studying cancer biology ([@b35-mmr-18-04-3898]). Cell lines are known to grow faster with minimal care. There is a possibility that cell lines do not behave identically with primary cells. They should not be used to replace primary cells ([@b35-mmr-18-04-3898]). Although this study was performed with SKOV3 cell line, we should do confirmation with primary cells in the future.\n\nOverall, our study showed that CRT1 could inhibit cell viability, cell migration, and invasion of SKOV3 ovarian cancer cells through activating the ERK1/2/p90RSK signaling pathway. These results indicate that a natural *ent*-kaurane diterpenoid, CRT1, might have potential as a chemotherapeutic agent to prevent the spread of ovarian cancer by activating ERK1/2. Further studies are needed to investigate the biological efficacy of CRT1 for treating ovarian cancer using *in vivo* model.\n\nNot applicable.\n\nFunding\n=======\n\nThe present study was supported by a research fund from Chungnam National University (grant no. 2015-0881-01).\n\nAvailability of data and materials\n==================================\n\nAll data generated or analyzed during this study are included in this published article.\n\nAuthors\\' contributions\n=======================\n\nJSL designed the study and prepared the manuscript. MSL, EYC and JBP performed the experiments and analyzed the data. PTT isolated and identified the structure of CRT1. JYS and YBK conceived and designed the study, and revised the manuscript.\n\nEthics approval and consent to participate\n==========================================\n\nNot applicable.\n\nPatient consent for publication\n===============================\n\nNot applicable.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\n![CRT1 decreases the viability of SKOV3 cells. (A) Chemical structure of CRT1. (B) Cells were treated with CRT1 at the indicated concentrations (1, 5, 10, 25, 50 and 100 \u00b5M) for 24 h. The viability of cells was measured by a CCK-8 assay. (C) Following treatment with CRT1 at 25 \u00b5M for different time periods (12, 24 and 48 h), cell viability was determined using a CCK-8 assay. Each assay was performed in triplicate. Data presented as the mean \u00b1 standard deviation. \\*P\\<0.05 vs. control (0 \u00b5M or h). CRT1, *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene; CCK-8, Cell Counting Kit-8.](MMR-18-04-3898-g00){#f1-mmr-18-04-3898}\n\n![CRT1 induces apoptosis in SKOV3 cells. (A) Cells were treated with CRT1 at 25 or 50 \u00b5M for 24 h followed by Annexin V-FITC/PI double staining. Apoptotic cells were analyzed by flow cytometry. Data are representatives of three independent experiments. (B) Percentage of apoptotic cells is presented as the mean \u00b1 standard deviation. \\*P\\<0.05 and \\*\\*P\\<0.01 vs. control. (C) Cells were treated with various concentrations of CRT1 for 24 h followed by western blot analysis to measure protein expression levels of Bax and Bcl-2. \u03b2-Actin was used as a protein loading control. Protein expression level is presented as the mean \u00b1 standard deviation. \\*P\\<0.05 and \\*\\*P\\<0.01 vs. Lane no. 1. (D) Mitochondria and cytosolic fractions were prepared and western blot analysis was performed to determine cytochrome c levels. (E) Protein levels of caspase-3, \u22127 and \u22129 were determined by western blot analysis. Z-VAD-fmk at 100 \u00b5M was pretreated for 2 h followed by CRT1 treatment for an additional 24 h. CRT1, *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene; FITC, fluorescein isothiocyanate; PI, propidium iodide; Bcl-2, B-cell lymphoma 2; Bax, Bcl-2-associated X protein; PARP, poly-(adenosine diphosphate-ribose) polymerase.](MMR-18-04-3898-g01){#f2-mmr-18-04-3898}\n\n![Effect of CRT1 on ERK1/2 phosphorylation in SKOV3 cells. (A) Time-dependent and (B) concentration-dependent effects of CRT1 on the protein expression levels of ERK1/2, p-ERK1/2, p90RSK and p-p90RSK were analyzed by western blotting. Actin was used as a protein loading control. Each experiment was performed in triplicate. All data are presented as the mean \u00b1 standard deviation. \\*P\\<0.05 vs. the control. (C) ERK inhibitor PD98059 reversed the CRT1-induced phosphorylation of ERK1/2. PD98059 (25 \u00b5M) was preincubated for 2 h prior to the addition of CRT1 (50 \u00b5M). Phosphorylation levels of ERK1/2 and total ERK, and the expression level of actin were analyzed by western blot analysis with the indicated antibodies. (D) Protein expression level is presented as the mean \u00b1 standard deviation. \\*P\\<0.05 vs. Lane no. 3. CRT1, *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene; ERK, extracellular signal-regulated kinase; p-, phosphorylated.](MMR-18-04-3898-g02){#f3-mmr-18-04-3898}\n\n![ERK1/2 regulates CRT1-induced apoptosis and anti-proliferation. (A) ERK inhibitor PD98059 reversed CRT1-induced caspase cleavages. PD98059 (25 \u00b5M) was preincubated for 2 h prior to the addition of CRT1 (50 \u00b5M). Levels of cleaved caspase-3, caspase-7, PARP and \u03b2-actin were analyzed using western blot analysis with the indicated antibodies. (B) Cells were pretreated with PD98059 for 2 h prior to CRT1 (50 \u00b5M) treatment. Apoptotic cells were analyzed by Annexin V-FITC/PI double staining assay followed by flow cytometry. Data are representatives of three independent experiments. (C) Percentage of apoptotic cells. (D) Following pre-incubation of PD98059 prior to CRT1 treatment (50 \u00b5M), cell viability was determined using Cell Counting Kit-8 assay. Each assay was performed in triplicate. Data are presented as the mean \u00b1 standard deviation. \\*P\\<0.05 vs. CRT1 single treatment. ERK, extracellular signal-regulated kinase; CRT1, *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene; PARP, poly-(adenosine diphosphate-ribose) polymerase; FITC, fluorescein isothiocyanate; PI, propidium iodide.](MMR-18-04-3898-g03){#f4-mmr-18-04-3898}\n\n![CRT1 inhibits colony-forming ability, cell migration and invasion of SKOV3 cells by ERK activation. (A) Colony-forming assay for single cell culture of SKOV3 cells treated with CRT1 (50 \u00b5M) and/or PD98059 (25 \u00b5M). Cells were allowed to culture until visible colonies were evident. Data are presented as photographs: Phase-contrast microscopic examination of the morphology of cells in culture (magnification, \u00d7100). (B) Cell migration was measured by wound healing assay. Cells in culture were wounded and allowed to migrate into the denuded area for 48 h. Five random views were chosen along the scratch wound in each well (magnification, \u00d7100). (C) Transwell invasion assay. CRT1 (50 \u00b5M) and/or PD98059 (25 \u00b5M)-treated cells were fixed and stained following 48 h as described in the materials and methods. Three randomly selected fields of view were photographed in each well (magnification, \u00d7100). (D) Quantification of the migration of SKOV3 cells. Data represent replicates of three independent experiments and are expressed as the mean \u00b1 standard deviation. \\*P\\<0.05 vs. CRT1 single treatment; ^\\#^P\\<0.05 vs. untreated control. CRT1, *ent*\u221218-acetoxy-7\u03b2-hydroxy kaur-15-oxo-16-ene; ERK, extracellular signal-regulated kinase.](MMR-18-04-3898-g04){#f5-mmr-18-04-3898}\n"}
+{"text": "Introduction {#sec1-1}\n============\n\nThe increase in the life expectancy of the population has been in part driven by reduced mortality of aged people \\[[@ref1]\\]. The number of people aged 65 or more is expected to grow from ca. 524 million in 2010 to almost 1.5 billion in 2050, with most of the increase in the developing countries. In Brazil, for example, the same demographic ageing which took more than a century in developed France occurred in two decades \\[[@ref2]\\]. In a descriptive study covering the period from 1970 to 2010 the increase in the ageing index (AI) of 268% was determined for Brazil in general, with still higher AI established for the state of Rio Grande do Sul (RS) \\[[@ref3]\\]. According to the data from the National Public Health Care System, in the RS, during 1991-2012 absolute number of seniors increased by 652,655 corresponding to the rise of the proportion of elderly people from 8.9% to 13.6% \\[[@ref4]\\].\n\nDespite the increase in longevity, the higher proportion of comorbidities distinguishes the older from the younger population \\[[@ref1]\\]. It's known that the elderly people have 2-3 times higher incidence of acute myocardial infarction (AMI) than the younger ones. They also tend to have more complications associated with prolonged hospital stays, low physical activity and hence, suffer substantially higher fatality rates due to coronary artery disease (CAD) \\[[@ref5]\\]. Furthermore, cardiovascular disease (CVD) is by far the most important cause of hospitalisation among the elderly, and CAD is the leading cause of death in Brazil. A recent study showed that CVD accounted for 43.1% of all deaths that occurred between 2006 and 2010 in Brazil \\[[@ref6]\\]. Because of high rates of morbidity and mortality, primary and secondary prevention programs are important strategies not only for alleviating cardiovascular risk factors but also for decrease the mortality and improvement of life quality for elderly patients \\[[@ref5]\\].\n\nOver the past four decades, rehabilitation programs have been recognised as an important tool in the medical care of patients with CVD. Preventive strategies for clinical practice should be developed based on cardiovascular rehabilitation programs because:\n\n1)CVD is the leading cause of death in most countries; it is a major cause of disability and contributes significantly to increase healthcare costs.2)Atherosclerosis can develop slowly over decades and its clinical manifestations are only seen in advanced stages of the disease.3)Most CVD is closely associated with lifestyle.4)Risk factors such as obesity, smoking, diabetes mellitus and hypertension have increased in the last decades.5)The cardiac rehabilitation programs (CRP) are able to minimise risk factors and reduce morbidity and mortality \\[[@ref7]\\].\n\nDespite the low representation of older people in most experimental and observational studies on cardiac rehabilitation (CR), there seems to be clear that CR has the most beneficial effect of on younger patients compared to elder ones \\[[@ref8], [@ref9]\\]. At the same time, the risk of adverse reactions and complications for the elderly during the exercise sessions is similar to that of younger patients. Even though the age is not a determining factor for physical, functional and psychosocial response to a CRP, the existing stereotypes often prevent health professionals from prescription these exercises to the aged patients \\[[@ref8]\\]. It seems clear that for many reasons the most important predictive factor for not entering in a CRP is the age \\[[@ref9]\\]. In that respect, the strongest possible recommendations have an extremely high importance for entering the CRP.\n\nCR is indicated for all patients with different presentations of CAD as it is known to have a positive effect on the treatment \\[[@ref7]\\]. Nevertheless, elderly patients are less likely to be referred to a CRP \\[[@ref10]\\]. It was shown that attendance to a CRP in a long-term reduces all-cause mortality \\[[@ref11]\\]. However, the majority of older than 70 years referred to a CR program still avoids any training session \\[[@ref11], [@ref12]\\]. The term \"old\" often defines patients \\>65, despite the fact that most specialists consider patients 65-75 years old relatively young, while those \\>75 or 80 are \"very elderly\" and for them the CR data are still less available \\[[@ref13]\\].\n\nThe dropout rate of aged patients from CR programs in Brazil was not comprehensively analysed and no differentiation by age group was carried out.\n\nThereby, this study aims to analyse the dropout rate of old people in a Brazilian CRP on the example of the large CR centre affiliated with the Institute of Sports Medicine (University of Caxias do Sul, RS). Comparative analysis of different age groups was also the aim of the present work.\n\nMaterials and Methods {#sec1-2}\n=====================\n\nEthical aspects {#sec2-1}\n---------------\n\nThis historic cohort study was conducted in the Cardiac Rehabilitation Service of the Sports Medicine Institute (SMI) at the University of Caxias do Sul (UCS), Brazil. The research was approved by the Research and Ethics Committee of the Cenecista Faculty of Bento Gon\u00e7alves. All patients had given informed consent.\n\nStudy Population {#sec2-2}\n----------------\n\nAll patients were referred and attended the CRP between March 2011 and June 2016. They were divided into groups according to age: Non-Old (\\< 65 years); Young- Old (65-74 years); Middle-Old (75-84); and Oldest-Old (\u2265 85 years), which is consistent with previous studies \\[[@ref14]\\]. The exclusion criterion was the absence of data or discrepancy in the medical records.\n\nMeasures {#sec2-3}\n--------\n\nThe clinical characteristics assessed were: AMI with hospital report; CAD proved by cine coronary angiography; heart failure (HF) with an echocardiogram and coronary artery bypass grafting (CABG) and percutaneous transluminal coronary angioplasty (PTCA) with stent placement, both with hospital report. Also, risk factors analysed were: hypertension, diabetes mellitus, dyslipidemia, all three considered from a report of attending physician; tobacco was self-reported at medical interview.\n\nThe body mass index (BMI) was classified as corresponding to normal (\u2264 24.99), to overweight (25-29.99) and to obesity (\u2265 30). The exercise capacity (EC) was assessed according to the percentiles of maximal oxygen uptake (VO~2~max) of the study sample. The functional capacity classification varies according to age and gender. So, the calculation was done separately according to the type of assessment, exercise testing (ergometry) or cardiopulmonary exercise testing (ergo spirometry). Both tests were carried out using Micromed Biotecnologia Ergo PC Elite version 3.3.6.2. Protocol ramp was used in all exercises. Dropout was defined as attending 50% of the rehabilitation program or less, which is consistent with previous studies \\[[@ref15]\\].\n\nCardiac Rehabilitation {#sec2-4}\n----------------------\n\nA multidisciplinary rehabilitation program of 48 sessions of training was offered to each patient after being referred by an assistant physician. The program included: exercise training, dietary counselling by nutritionists, smoking cessation, and psychological support. Patients enrolled through National Health System (NHS) trained twice a week and those enrolled in Private Health Plan (PHP), trained 3 times a week. Before the training, the patients were seen by the nurse to assess vital signs. The session training lasted 60 min and it was conducted by physical educators. The exercise training program consisted of a combination of aerobic and strengthening exercises. In the end of the session, the physiotherapist coordinated stretching the major muscle groups.\n\nData collection and Statistical Methods {#sec2-5}\n---------------------------------------\n\nData collection was performed using a spreadsheet online of Google Drive\u00ae from March 2013 to June 2015, in the SMI of UCS, a university service of CR. Statistical analysis was made with SPSS\u00ae software 22.0 version with descriptive statistics for sample characterization. The results were presented as a mean and standard deviation, absolute numbers for frequency and percentages values.\n\nClinical characteristics between age groups at baseline were compared using ANOVA (quantitative analyses) and by Chi-square for categorical variables. The logistic regression binominal was performed to assess the risk of dropout according to age groups, and to verify the confounding factors.\n\nFirst, it was verified the correlation of the dropout with each variable by the obtaining the odds ratio. In the following steps, those variables were selected whose value level was less than 0.20 according to the descriptive test of significance. Then, the adjusted odds ratio was performed using logistic regression model. For all analyses was used a two-tailed P-value \\<0.05.\n\nResults {#sec1-3}\n=======\n\nClinical Characteristics According to Age Groups {#sec2-6}\n------------------------------------------------\n\nOut of 362 patients, 31 were excluded because of lack of information or disagreement of data in the medical records. Moreover, only 5 patients were Oldest-Old (\u2265 85 years) and because of no sufficiency of this sample, these patients were excluded as well. Thus, in total the sample consisted of 326 patients. 141 of them were Non- Old Patients (\\<65); 128 were Young-Old people (65-74 years), and 57 were Middle- Old people (75-84 years old). The patients enrolled by NHS constituted 52.3% and by PHP, 47.7%. The overall mean age was 63.82 years (\u00b1 11.66) and 180 patients (55.2%) were men. Clinical characteristics according to age groups are summarised in [Table 1](#T1){ref-type=\"table\"}.\n\n###### \n\nClinical characteristic according to age: non-old group (\\< 65 y), young-old-group (65-74 y) and of the middle-old group (75-84 y)\n\n  Mean (sd^2^)                          Non-Old (n = 141)   Young-Old (n = 128)   Middle-Old (n = 57)   P value    \n  ------------------------------------- ------------------- --------------------- --------------------- ---------- ------\n  Age                                   52.99 (8.57)        69.21 (2.85)          78.49 (2.61)          NA^1^      \n  Body Mass Index                       28.34 (5.29)        29.12 (5.97)          28.29 (4.89)          0.79       \n  Ergometry VO~2~ Max: ml/kg/min        26,09 (9,55)        21.30 (9.91)          14.42 (5.55)          \\<0.01     \n  Ergospirometry VO~2~ Max: ml/kg/min   20,47 (6,33)        17.18 (4.76)          14.02 (6.06)          \\<0.01     \n  Prevalence at baseline (%)                                                                                       \n  Gender Male                           77 (54.6%)          54 (57.8%)            29 (50.9%)            0.66       \n  Heart Failure                         30 (21.3%)          33 (25.8%)            13 (22.8%)            0.64       \n  Coronary Artery Disease               111 (78,7%)         98 (76.6%)            41 (71.9%)            0.32       \n  Acute Myocardial Infarction           68 (48,2%)          44 (34.4%)            20 (35.1%)            0.03       \n  CABG^3^                               36 (25.5%)          40 (31.3%)            12 (21.1%)            0.82       \n  PTCA^4^ with stent                    46 (32.9%)          34 (26.6%)            22 (38.6%)            0.73       \n  Diabetes                              34 (24.1%)          47 (36.7%)            22 (38.6%)            0.01       \n  Hypertension                          98 (69.5%)          106 (82.8%)           49 (86%)              \\<0.01     \n  Dyslipidemia                          80 (56.7%)          84 (65.6%)            33 (57.9%)            0.56       \n  Smoker                                Current             22 (15.7%)            19 (14.8%)            3 (2.5%)   0.05\n  Ex-smoker                             57 (40.7%)          40 (31.3%)            21 (36.8%)                       \n  Non-smoker                            61 (43.6%)          69 (53.9%)            34 (59.6%)                       \n\ny: years; 1: not applicable; 2: standard deviation; 3: coronary artery bypass grafting; 4: percutaneous transluminal coronary angioplasty.\n\nThe clinical characteristics were comparable and a similar clinical status can be seen in both groups. Most significantly older people differ from more often cases of hypertension, diabetes, and by a higher number of non-smokers (p \\< 0.05). According to enrollment, the Non-Old group more often was covered by NHS, while in the Middle- Old group PHP prevailed. These results tend to be significant (p = 0.07). Older people were less prone to AMI and had lower EC in both tests (p \\< 0.05).\n\nDropout in Cardiac Rehabilitation Program {#sec2-7}\n-----------------------------------------\n\nThe overall mean of attendance at exercise sessions of the rehabilitation program was 28.94 (\u00b1 13.94), a frequency mean of 60.29%. Regarding dropouts, 120 patients (36.8%) abandoned the exercise program (\u2264 24 sessions). [Table 2](#T2){ref-type=\"table\"} shows the clinical characteristics according to the dropout rates which differ significantly for different age groups.\n\n###### \n\nClinical characteristics according to drop out in cardiac rehabilitation\n\n                                             Dropout (n = 120)        Non-Dropout (n = 206)   P value       \n  ------------------------------------------ ------------------------ ----------------------- ------------- --------\n  Enrollment                                 National Health System   70 (58.3%)              101 (49%)     0.10\n  Private Health Plan                        50 (41.7%)               105 (51%)                             \n  Gender                                     Men                      64 (53.3%)              116 (56.3%)   0.60\n  Women                                      56 (46.7%)               90 (43.7%)                            \n  Age                                        Non-Old                  62 (51.7%)              79 (38.3%)    0.01\n                                             Young-Old                35 (29.2%)              93 (45.1%)    \n                                             Middle-Old               23 (19.2%)              34 (16.5%)    \n  Heart Failure                              28 (23.3%)               48 (23.3%)              0.99          \n  Coronary Artery Disease                    93 (77,5%)               157 (76.2%)             0.79          \n  Acute Myocardial Infarction                50 (41,7%)               82 (39.8%)              0.74          \n  CABG^1^                                                             19 (15.8%)              69 (33.5%)    \\<0.01\n  PTCA^2^ + stent                                                     42 (35%)                60 (29.1%)    0.27\n  Tobacco                                    Current                  20 (16.7%)              23 (11.2%)    0.20\n  Ex-smoker                                  46 (38.3%)               72 (35%)                              \n  Non-smoker                                 54 (45%)                 111 (53.9%)                           \n  Diabetes                                                            35 (29.2%)              68 (33%)      0.47\n  Hypertension                                                        92 (76.7%)              161 (78.2%)   0.75\n  Dyslipidemia                                                        70 (58.3%)              127 (61.7%)   0.55\n  Exercise capacity (Percentile VO~2~ max)   0-25^th^                 31 (25.8%)              50 (24.3%)    0.47\n  26-50^th^                                  32 (26.7%)               51 (25.5%)                            \n  51-75^th^                                  24 (20%)                 81 (24.8%)                            \n  76-100^th^                                 33 (27.5%)               81 (24.8%)                            \n  Body Mass Index                            Normal (\u2264 24.99)         27 (22.5%)              45 (21.8%)    0.82\n  Overweight (25-29.99)                      53 (44.2%)               98 (47.6%)                            \n  Obesity (\u2265 30)                             40 (33.3%)               63 (30.6%)                            \n\n1: coronary artery bypass grafting; 2: percutaneous transluminal coronary angioplasty.\n\nThus, the group Young-Old showed lower dropout rate compared to non-old patients (p = 0.01). Furthermore, patients underwent CABG, showed a lower rate of dropout at the end of the program (p = 0.001). The variables Enrolled, Tobacco, Age, and CABG and, thus, were selected to the binary logistic regression.\n\nLogistic Regression Model for Dropout in Cardiac Rehabilitation {#sec2-8}\n---------------------------------------------------------------\n\nLogistic backwards regression was performed to identify the confounding factors. The remaining significant variables are shown in [Table 3](#T3){ref-type=\"table\"}.\n\n###### \n\nLogistic regression model for dropout in cardiac rehabilitation\n\n                           Adjusted Odds Ratio               95% CI       P value      \n  ------------------------ --------------------------------- ------------ ------------ ------\n  Enrollment               Private Health Plan (reference)   1.00                      \n                                                                                       \n  National Health System   1.45                              0.89--2.35   0.12         \n                                                                                       \n  Tobacco                  Current (reference)               1.00                      \n                                                                                       \n  Ex-smoker                1.24                              0.59--2.63   0.56         \n                                                                                       \n  Non-smoker               1.69                              0.81--2.52   0.15         \n                                                                                       \n  Age                      Non-Old, \\<65 y (reference)       1.00                      \n                                                                                       \n                           Young-Old, 65-74 y                1.96         1.16--3.29   0.01\n                                                                                       \n                           Middle-Old, \u2265 75                  1.06         0.55--2.04   0.84\n                                                                                       \n  CABG^1^                  Yes (reference)                   1.00                      \n                                                                                       \n  No                       2.76                              1.53--4.95   \\<0.01       \n\n1: coronary artery bypass grafting.\n\nThe type of enrollment and tobacco status did not show any significant correlation with dropout. The age, on the other hand, showed that Young-Old people have 96% less risk of drop out compared to Non-Old (AOR: 1.96 \\[1.16--3.29\\]). The absence of CABG, in turn, increases the risk of dropout by almost three times (AOR: 2.76 \\[1.53--4.95\\]).\n\nDiscussion {#sec1-4}\n==========\n\nThe study consisted of 326 patients who attended the CRP of the SMI at the UCS. The patients were referred by physicians of the NHS and of PHPs. As expected, EC decreased with age of patients while comorbidities enhanced. The overall dropout was 36.8% and the age was seen as an important predictor for dropout. At the end of exercise program, the non-old patients showed higher dropout rate compared to young-old patients. The women did not show any difference regarding dropout. At the same time, CABG reduced the risk of dropout.\n\nThe prevalence of males in the literature varies due to differences in selection of participants in the studies, but it is clear that men, among older patients, are also most in this type of service \\[[@ref8], [@ref17]-[@ref19]\\]. Our findings show same results. Our results have shown high frequencies of the risks of CVDs, especially of hypertension which corroborates with data of other studies \\[[@ref5], [@ref17], [@ref20]\\]. The different clinical characteristics depending on age were expected and observed in our study. The ergo spirometry proved that VO~2~ max in very old patients is very rare. The functional capacity of a person decreases significantly with advanced age \\[[@ref21]\\]. Some studies showed lower functional capacities of older people at baseline in CRP. Nevertheless, this age group seems to respond better to an exercise program, and tend to have superior relative improvements of aerobic capacity \\[[@ref22]-[@ref24]\\]. The effects of aerobic capacity were not a goal of the present study which only described EC at baseline in the program. A significant decrease of VO~2~ max was observed in the linear test, either by exercise testing or cardiopulmonary exercise test. Other comorbidities, as diabetes and hypertension noted in [Table 1](#T1){ref-type=\"table\"}, were more common in older patients because they have more pathologies and this is in concordance with previous studies \\[[@ref9], [@ref25]\\].\n\nThe dropout rate of the elderly in CR programs worth discussing, since the patients, who abandoned the programs are more prone to cardiac complications than those who complete them \\[[@ref26]\\]. Also, a reduced mortality was documented after 14-years follow-up, with control of a dose-response. Beauchamp et al. showed that attendance at 10-24% of training sessions have a good correlation with the greater global cause of mortality if compared with an attendance rate of 75-100% \\[[@ref12]\\]. In our study, the overall dropout rate was 36.8% (120 of 326). The literature data shows some variations in these values. Thus Turk-Adawi et al. observed 49.7% of dropout \\[[@ref27]\\], Worcester et al. found 19.08% \\[[@ref28]\\] and Pardaens et al. registered about 20% of overall dropout from CR program \\[[@ref29]\\]. This difference should be due to the socio-cultural specificity of different samples as they belong obviously to different societies and countries.\n\nSuaya et al. evaluated outpatient CR training effect on the patients that had acute MI or CABG surgery using a statistical sample which comprised 267427 post- hospitalisation patients aged \u2265 65-year-old \\[[@ref9]\\]. They found CR and exercise training was prescribed to 13.9% of the patients hospitalised with acute MI and to 31.0% of the patients after CABG surgery \\[[@ref9]\\]. Patients underwent this kind of surgery shown higher adherence due to the severity of the medical conditions. Worcester et al. found the attendance rates of 66%, 51% and 25% respectively for CABG, AMI, and PTCA. In several studies, the CABG surgery was found to be a variable which can serve a predictor for non-dropout. In the work of Yohannes et al. \\[[@ref30]\\] the correlation between CABG cases and accomplishment of CR program was also established, however for AMI and PTCA such relation was not found. Similar results were observed in the work \\[[@ref28]\\] where it was noted that patients after CABG less likely abandon CR program than patients after acute coronary syndrome or HF. High degree of adhesion of CABG patients to CR program was also shown in the article \\[[@ref27]\\] (AOR: 1.54 \\[1.24-1.82\\]). In our study, 27% of all patients had CABG surgery and these cases were equally distributed between age groups. It may be noted that our results corroborate with the literature data establishing a correlation between CABG cases and adherence to CR program. The AOR showed that participants of CR program that never underwent a CABG surgery had almost three times more risk for dropout. These findings were not observed for AMI or PTCA patients. However, it is not associated with 'poor' adherence of AMI patients but rather with higher adherence to the program of CABG patients. This could be explained by the severity of CABG which serves as a motivator for changes in lifestyle \\[[@ref8]\\].\n\nThe multicenter cohort study presented in \\[[@ref27]\\] was performed in 39 CR services and analysed 4,442 patients. The authors noted that younger patients (aged \\<65 years) were more likely to abandon the program compared to older patients (aged \u2265 65 years). In this article, the notion of \"adherent\" (i.e. patients attended \\> 50% of sessions) was used as reciprocal to \"drop out\". Besides, this paper did not distinguish the old people and therefore our results should be compared with observations made in \\[[@ref27]\\] with attention to the methodological differences. A prospective study, in turn, consisted of 556 patients eligible to participate in secondary prevention program of exercises. There a mean age was 64,9 years old and the overall dropout rate were 23.4% \\[[@ref28]\\]. This paper studied predictors for the dropout and found no difference between patients \\< 70 and \u2265 70 years old. In another prospective research, the dropout rate of approximately 20% was documented for all cases where one or more sessions were analysed \\[[@ref21]\\]. However, a quantitative analyst was carried out in this article using only the mean age, while partitioning by age group has not been made and therefore higher rates of dropout for elderly patients could not be noticed.\n\nIn turn, Yohannes coordinated a study which included 189 post-AMI patients. They were recruited from a consecutive series of outpatient referrals prior to a CRP consisted of 6 weeks \\[[@ref30]\\]. The authors considered dropout those who abandon the CR program during first two weeks and completers were those who attended the entire program (12 sessions). The quantitative assessment made in \\[[@ref30]\\] has shown a trend for younger patients to drop out (61.4 y vs. 58.7 y, p = 0.08). In another large cohort study, which sought to assess the rate and predictors for dropout of CR program, age was not used as a factor \\[[@ref21]\\]. For a 12-week program overall dropout rate was found to be 12.9 %; however, there was no clear definition of what was considered \"dropout\" and what was \"completer\" in their study \\[[@ref21]\\]. On the other hand, another cohort of 872 patients with mean age of 67 years, did not define dropout objectively and, in turn, used subjective parameters for such a concept. In this study, it was seen dropout rates greater in patients over 65 years; however, it is difficult to consider these findings since the fall of methods \\[[@ref31]\\].\n\nThe notion of dropout (from CR program) is not strictly defined. Most often participation in \\< 50% of the training sessions is regarded as a dropout, but some other approaches, such as abandon during first weeks, also exist. At the same time, it seems that even participation in 50% of sessions may not be enough for adequate rehabilitation. So, the concept of dropout needs to be standardised, which is however not the goal of the present contribution. In any case, in the present research, it was shown that Young-Old (65-74 years) demonstrated the lowest rate of dropout (AOR 1.96 \\[1:16 to 3:29\\], p = 0.01) compared to non-elderly patients (\\< 65 years). On the other hand, the dropout rate of the Non-elderly group was close to that of the Middle-old group. It can be probably assumed that younger people are more often diverted from CR treatment by higher workload and family responsibilities \\[[@ref32]\\], while elderly are more dependent and more often suffer greater disability, not necessarily from CVD but also from co- morbidities such as arthritis \\[[@ref33]\\].\n\nThis study has some limitations which must be acknowledged. One of them is the nature of observational historic cohort study. Despite the broad spectrum of used variables, some potentially significant parameters are lacking, including for instance information on illness cognition (i.e. how people perceive the situation they experience). The latter has been recognised in the literature as a crucial determinant of health-promoting behaviour along with the financial situation \\[[@ref30], [@ref34]-[@ref36]\\]. Besides the factors such as self-motivation \\[[@ref37]\\] and work demands \\[[@ref32]\\], also have been recognised as important in the literature. Another limitation of the present study was exclusion the Oldest-Old group due to the inadequate number of patients in this age category. In addition, thirty-seven patients were not included in the study because they did not attend any session at all after being evaluated in CR centre. Further research could clarify the reasons of that dropout.\n\nBesides, the electronic database used in the present study was built from medical records collected in a single CR centre and therefore some measurement bias cannot be excluded. Despite the mentioned limitations, the present study is the first attempt to differentiate age groups in the analysis of dropout rate of aged people in a Brazilian CR program. It has shown that there are important differences between dropout rates in different age groups. We also consider this contribution quite representative because even with omitted Oldest-Old patients the statistical samples can be considered moderate in size, compared to other, not numerous studies based on still smaller sample sizes. Our main outcome was achieved objectively based on the attendance lists. In spite of being carried out in single-site, the study covered rather a long period of time and was performed in a large CR service, in a university centre of reference of southern Brazil.\n\nIn conclusion, the dropout rate in a large Brazilian CR centre was approximately 39% during 48 sessions of exercise program. The Non-Old and the Middle-Old patients showed increased dropout rates compared to Young-Old. Thus, there is a difference of dropout rate at elderly categories. To ensure the best possible rehabilitation and to improve patients\u00b4 participation in CR, the CR programs should be customised to patients' needs in terms of their age. It was also established that CABG procedure is associated with better adherence to the program. The study also confirmed that despite reduced functional capacity of older people, it should not be considered an obstacle for CR and these patients remain admissible to CR program.\n\nDue to limited literature data, more studies involving elder patients are needed, particularly those \\> 84 years, since the low presence of them in this type of service. This is necessary for analysis of the problems that came with ageing and could contribute to dropout.\n\nThe research project was funded by the University Program coordinated by the Coordinator of Research and Graduate Studies Stricto Sensu, Brazil. We thank the entire team of the Institute of Sports Medicine at the University of Caxias do Sul, mainly the medical staff, our co-worker Iara Maria Hunoff and the Professor Ricardo Rodrigo Rech for all the support. We thank Professor Sergei Mikhaelenko for the help with the writing of the article, especially regarding the English language. We are also grateful to the team of nursing, physical education, physical therapy, nutrition and psychology of Institute of Sports Medicine of UCS and we thank, finally, to the group of research of the LAMFEME.\n\n**Funding:** Research project was funded by the University Program coordinated by the Coordinator of Research and Graduate Studies Stricto Sensu, Brazil.\n\n**Competing Interests:** The authors have declared that no competing interests exist.\n"}
+{"text": "\n"}
+{"text": "Introduction\n============\n\nTelomeres, the nucleoprotein structures at the ends of chromosomes, consist of 5\u2032-TTAGGG-3\u2032 repeats bound by a dedicated set of proteins forming the shelterin complex ([@b54]). Three members of the six protein complex directly bind telomeric DNA and were the only direct telomere-specific binding proteins known so far: TRF1 and TRF2 bind double-stranded DNA (dsDNA; [@b79]; [@b9]; [@b11]), whereas POT1 binds single-stranded 5\u2032-TTAGGG-3\u2032 repeats ([@b7]). This complex constitutively associates with telomeres and shields the ends of linear chromosomes from being recognized as a double-stranded break, thus protecting telomeres from end-to-end fusions ([@b54]). While this solves the end-protection problem ([@b24]), maintaining telomere integrity itself is of outstanding importance and major factors may have remained elusive.\n\nTelomere length homeostasis, a crucial process in stem cell biology, aging and cancer, depends on the equilibrium between telomere lengthening (in most cases due to telomerase activity) and shortening reactions (generally due to replication and controlled processing) ([@b34]). Telomerase is capable of adding telomeric repeats to chromosome ends *de novo.* The enzyme works as a ribonucleoprotein complex, which consists of a catalytic subunit with reverse-transcriptase activity (called TERT), and an RNA serving as the elongation matrix for telomeres (called TR or TERC) ([@b29]). While these two core elements are sufficient for telomerase activity *in vitro*, biochemical analyses have shown that *in vivo* telomerase resides in a large complex of about 1 MDa ([@b57]). Some additional components of this large multi-subunit holoenzyme complex have been identified. In particular, the core components of box H/ACA small nucleolar ribonucleoprotein particles (snoRNPs), DKC1 (dyskerin), GAR1, NHP2 and NOP10, are part of the active telomerase complex, and are necessary for proper RNP assembly as well as for TERC stability ([@b50]; [@b68]). More recently, the ATPases RUVBL1 and RUVBL2 have been identified as factors essential for holoenzyme assembly ([@b66]), and TCAB1 (WDR79/WRAP53), identified as a DKC1 interaction partner, was shown to be required for proper localization of CAB box containing small Cajal body (CB)-specific RNPs (scaRNPs) to CBs, including TERC, and is part of the active telomerase complex ([@b63]; [@b65]).\n\nThe presence of major scaRNA processing and trafficking factors in the telomerase complex hints to an important aspect of telomerase cell biology: the orchestrated maturation of telomerase and interaction with telomeres in the CB. Telomere maintenance by telomerase requires that both TERT and TERC are recruited from distinct subnuclear sites to telomeres during S phase (synthesis phase) ([@b61]). Like other scaRNAs, TERC contains a common CB-specific localization signal and accumulates in CBs ([@b35]; [@b80]), where it is found together with TERT ([@b61]). In a cell cycle-dependent manner, telomerase-containing CBs are then recruited to telomeres, suggesting that CBs represent an enzymatic hub in which telomere elongation by telomerase takes place ([@b36]; [@b62]; [@b21]). This trafficking model is further supported by telomere elongation defects in the absence of TCAB1 or presence of dysfunctional TCAB1, disrupting TERC accumulation in the CB ([@b65]; [@b77]). Nevertheless, so far it remains elusive how telomeres are recruited to CBs, how this selective interaction is regulated and what drives the conversion from telomeres in a closed state, in which telomerase has little or no access, to telomeres in an open, accessible state.\n\nTelomerase is usually limiting and, under physiological conditions, acts preferentially on short telomeres ([@b31]; [@b10]), due to a well-established negative feedback loop mediated in *cis* by TRF1 and POT1, likely by hiding the 3\u2032-overhang, which serves as a template for telomerase ([@b46]). Indeed, diminished loading of POT1 or expression of a dominant-negative version lacking DNA-binding activity leads to telomere elongation by telomerase, and *in vitro* experiments have shown that POT1 is competing with telomerase for its substrate ([@b46]; [@b74]; [@b38]; [@b43]). However, POT1 also interacts with TPP1, and both proteins together promote telomerase activity *in vitro* ([@b41]). Furthermore, TPP1 has been shown to be required for the recruitment of telomerase to its substrate *in vitro* and to telomeric chromatin *in vivo* ([@b72]; [@b1]; [@b60]; [@b75]; [@b78]). While TPP1 has been proposed as a telomerase recruiter, it does not completely fit the definition, since it has initially been described as a negative regulator of telomere length in telomerase-positive cells ([@b45]; [@b74]), although these results may in part be attributed to secondary effects, such as the lack of POT1 tethering to telomeres in the absence of TPP1. Based on these bivalent results, a ying-yang model for telomerase recruitment and activity control has been proposed for TPP1--POT1 ([@b72]). Indeed, recently, a specific patch of amino acids on the surface of TPP1, the TEL patch (TPP1 glutamate (E)- and leucine (L)-rich patch), has been identified as crucial for the TPP1 function in telomerase recruitment and regulation ([@b52]). Analysis of point mutations within the TEL patch demonstrated that the TEL patch is physically and functionally distant from the portion of TPP1 engaged in end protection, separating these two functions ([@b52]).\n\nRegulating the amount of telomeric repeats added by telomerase also involves how long telomerase can act on a given chromosome end. Therefore, turning off telomerase needs to be regulated in addition to the recruitment step and processivity control. In a recent study, [@b15] described the human CST (CTC1, STN1 and TEN1) complex as a terminator of telomerase activity ([@b15]). CST competes with POT1--TPP1 for telomeric DNA and is increasingly enriched on telomeric DNA during late S/G2 phase, correlating with the period in the cell cycle when telomerase action is terminated. In agreement with a suppression of telomerase action, depletion of any of the three CST complex members led to a steady increase in telomere length in a telomerase-dependent manner. The authors suggested that CST binds to telomerase-extended 3\u2032-ends and thereby suppresses telomerase access and further elongation ([@b15]). However, depletion of STN1 has been reported not to affect telomere length in various telomerase-positive cellular contexts ([@b70]) and murine CTC1-null cells do not exhibit the reported telomere lengthening phenotype ([@b30]). The proposed model is nonetheless appealing, as human CST subunits stimulate DNA polymerase \u03b1-primase ([@b14]) and therefore CST binding to telomerase-extended 3\u2032-ends could initiate a switch from telomere elongation to fill-in synthesis. This mechanism could provide an autonomous end point to telomerase action at single telomeres, ensuring that every telomere is extended by telomerase once and only once during every cell cycle ([@b15]).\n\nWe reasoned that the identification of novel telomere-binding proteins would be a first step to the identification of additional factors implicated in telomere biology. SILAC-based quantitative mass spectrometry (MS) has been adapted for DNA--protein interactions ([@b51]) and has been used by us and others successfully to identify factors binding to particular functional DNA fragments ([@b48]; [@b13]; [@b6]; [@b12]). For our purpose, we applied this approach to telomeric DNA in order to screen for telomere repeat-binding proteins and identified the protein HOT1 (HMBOX1; homeobox telomere-binding protein 1). HOT1 had previously been described as a putative transcriptional repressor based on reporter gene assays ([@b16]) and had been identified as a telomere-associated protein by the proteomics of isolated chromatin segments (PICh) approach ([@b25]). Here we demonstrate that HOT1 directly binds to telomeric DNA, and characterize this binding in atomic detail by resolving a crystal structure of the HOT1 homeobox domain in a cocrystal with telomeric DNA. *In viv*o, HOT1 localizes to a subset of telomeres with a higher degree of HOT1--telomere association in cellular contexts of elevated telomere processing. In addition, we show that HOT1 associates with the active telomerase complex and that HOT1 is required for telomerase chromatin binding. These findings suggest that HOT1 contributes to the association of telomerase with telomeres and telomere length maintenance in various cellular settings, and classify HOT1 as the first direct telomere-binding protein that acts as a positive regulator of telomere length.\n\nResults\n=======\n\nIdentification of HOT1 as a direct telomere repeat-binding protein\n------------------------------------------------------------------\n\nTo identify telomere-binding proteins we used polymerized biotinylated double-stranded oligonucleotides of the telomeric sequence (5\u2032-TTAGGG-3\u2032) and a scrambled control sequence (5\u2032-GTGAGT-3\u2032), separately immobilized on paramagnetic streptavidin beads and incubated with heavy and light SILAC-labelled nuclear extracts from HeLa cells, respectively. Specific binding of proteins is detected by incubation of cell lysates encoded by 'heavy\\' amino acids (^15^N- and ^13^C-labelled Lys and Arg) with the bait sequence, while a control sequence is incubated with 'light\\', nonlabelled amino acids. Specific binders display a differential SILAC ratio, whereas background binders have a 1:1 ratio. After mild washing, bead fractions were combined and captured proteins were analysed by quantitative, high-resolution MS ([@b20]) ([Figure 1A](#f1){ref-type=\"fig\"}).\n\nWe identified all the six core shelterin components with a SILAC ratio of about 10 or higher in the 'forward\\' and about 0.1 in the 'reverse\\' experiment, in which we had switched the labels ([Figure 1B--D](#f1){ref-type=\"fig\"}, [Supplementary Figure S1](#S1){ref-type=\"supplementary-material\"} and [Supplementary Table S1](#S1){ref-type=\"supplementary-material\"}). In contrast, none of the proteins known to interact with shelterin were identified with SILAC ratios sufficiently high to be consistent with telomere binding, demonstrating that this approach was very stringent and exclusively detected telomere repeat-binding proteins and their strong interaction partners ([Figure 1C](#f1){ref-type=\"fig\"}). In addition to the shelterin components, we found the protein HOT1 with a high SILAC ratio that clustered with those of the shelterin components ([Figure 1B--D](#f1){ref-type=\"fig\"}, [Supplementary Figure S1](#S1){ref-type=\"supplementary-material\"} and [Supplementary Table S1](#S1){ref-type=\"supplementary-material\"}). This indicates that HOT1 must either strongly associate with the shelterin complex or directly bind to the 5\u2032-TTAGGG-3\u2032 repeats.\n\nTo verify that the HOT1 identification was not cancer-, cell- or species-specific, we repeated our telomere-binding assay with SILAC-labelled nuclear extracts derived from mouse embryonic stem cells (ES cells). Again, all components of the shelterin complex and HOT1 were identified with SILAC ratios, indicating specific binding to the telomere repeats ([Figure 1D](#f1){ref-type=\"fig\"} and [Supplementary Table S2](#S1){ref-type=\"supplementary-material\"}). Here we also identified the two paralogues, POT1a and POT1b, which result from a gene duplication of the *Pot1* gene in the rodent lineage ([@b33]), underscoring the specificity of our assay for direct telomere-binding proteins. Hence, HOT1 is a putative telomere repeat-binding protein conserved in mammalian cells.\n\nHOT1 contains a homeobox domain ([@b16]), suggesting that it may bind DNA directly. To determine whether HOT1 was detected in our assay due to direct binding to the 5\u2032-TTAGGG-3\u2032 repeats, we performed DNA-binding assays with HOT1 *in vitro*. Recombinant HOT1 bound specifically to telomeric repeats, whereas no binding to the negative control repeat fragments (5\u2032-GTGAGT-3\u2032) was detected ([Figure 2A](#f2){ref-type=\"fig\"}). Exhibiting similar binding behaviour as TRF1, HOT1 was not enriched on any of the subtelomeric variant repeats 5\u2032-TCAGGG-3\u2032, 5\u2032-TGAGGG-3\u2032 and 5\u2032-TTGGGG-3\u2032, nor on the *C. elegans* telomere 5\u2032-TTAGGC-3\u2032 repeat sequence ([@b71]; [Figure 2A](#f2){ref-type=\"fig\"}). To test whether HOT1 also associates with telomeres *in vivo*, we performed chromatin immunoprecipitation (ChIP) experiments with extracts from HeLa cells using an antibody directed against HOT1. Similar to TRF2, HOT1 IPs showed enrichment of telomeric DNA in comparison to two negative controls (anti-GFP antibody and IgG; [Figure 2B](#f2){ref-type=\"fig\"}). Thus, HOT1 is a direct and specific telomere repeat-binding protein.\n\nHOT1 recognizes telomeric DNA by means of its homeodomain\n---------------------------------------------------------\n\nDriven by these findings and with the aim to fully understand the molecular interactions between HOT1 and telomeric DNA, we crystallized the DNA-binding domain (DBD) of HOT1 with telomeric DNA. In order to identify a construct suitable for crystallization, we initially tested six different HOT1 fragments for their DNA-binding ability ([Supplementary Figure S2](#S1){ref-type=\"supplementary-material\"}). The three longer constructs Q144--A345, L156--A345 and G233--A345 all bound to immobilized telomeric dsDNA baits, demonstrating that the homeodomain of HOT1 is sufficient for recognizing telomeric DNA and that integrity of the predicted N-terminal POU-specific (POUs) domain ([@b17]) is not required. The three shorter constructs (P242--A345, P254--A345 and R271--A345) were not able to bind to the bait DNA ([Supplementary Figure S2](#S1){ref-type=\"supplementary-material\"}). For crystallization, we reconstituted and purified telomeric DNA complexes with all three binding constructs, but only one (G233--A345) yielded crystals when reconstituted with a duplex telomeric DNA (5\u2032-cTGTTAGGGTTAGGGTTAG-3\u2032 and 3\u2032-ACAATCCCAATCCCAATCt-5\u2032) similar to the one present in the crystal structures of the TRF1 and TRF2 homeodomains bound to telomeric DNA ([@b19]).\n\nThe optimized crystals diffracted to 2.9 \u00c5 resolution and we could solve the structure by molecular replacement using the NMR model of the human HOT1 homeodomain (residues 268--343, PDB entry 2CUF, unpublished data from RIKEN Structural Genomics/Proteomics Initiative). In the orthorhombic crystals, the DNA forms an infinite double helix via a C--T nonWatson--Crick base pairing of the single-base overhangs (not shown). Two copies of the HOT1 DBD are bound to one duplex DNA that comprises two and a half 5\u2032-TTAGGG-3\u2032 repeats in a regular and undistorted B-form conformation ([Supplementary Figure S3](#S1){ref-type=\"supplementary-material\"}). As expected, and in accordance with the NMR model, the homeodomain of HOT1 folds into a small structure of three consecutive helices, \u03b11 (res. 276--288), \u03b12 (res. 293--309) and \u03b13 (res. 322--342), separated by a loop (\u03b11--\u03b12) and a turn (\u03b12--\u03b13; [Figures 2C](#f2){ref-type=\"fig\"} and [3A](#f3){ref-type=\"fig\"}, and [Supplementary Figure S3](#S1){ref-type=\"supplementary-material\"}). The N-terminal residues 233--266 and the two C-terminal residues were not defined by electron density and were, thus, not built.\n\nAs reported for TRF1 and TRF2, each copy of the HOT1 homeodomain binds to the major groove of the DNA double helix around a 5\u2032-TTAGGG-3\u2032 motif ([Figure 2C](#f2){ref-type=\"fig\"} and [Supplementary Figure S3](#S1){ref-type=\"supplementary-material\"}). Typical for homeobox domains, binding to DNA is mainly mediated by an N-terminal unstructured arm (267--276), the loop between \u03b11 and \u03b12, and the C-terminal \u03b13 ([Figure 2C](#f2){ref-type=\"fig\"}). Binding and sequence recognition is achieved through a combination of either water-mediated or direct contacts with the phosphate backbone (on both sides of the major groove) and the DNA nucleobases in the major (\u03b13) and minor groove (N-terminal arm; [Figures 2C and D](#f2){ref-type=\"fig\"}; the structure is deposited under PBD entry 4J19; for crystallographic statistics please see [Table I](#t1){ref-type=\"table\"}).\n\nLysine 335 is a key residue for telomere sequence specificity of HOT1\n---------------------------------------------------------------------\n\nIn order to bind a specific sequence, the HOT1 homeodomain must recognize and directly contact individual bases of telomeric DNA. We could identify five such direct interactions (R271, Y327, N332, K335, R339; [Figures 2C, D and F](#f2){ref-type=\"fig\"}) in our structure. Four of these residues reside on helix 3, which is generally referred to as the recognition helix and one, R271, is part of the unstructured N-terminal arm.\n\nIn detail, Y327 binds to N6 of the adenine base of A6 via a water molecule ([Figure 2F](#f2){ref-type=\"fig\"}). K338 and R339 are in close proximity and oriented towards the phosphate of G8 and the nucleobase of G7, respectively, but their electron density was not as clearly defined, probably due to conformational flexibility. However, in our mutant analysis the single mutants K338A and R339A no longer bind DNA, and Y327A shows weakened binding---a clear indication that these residues are important for binding ([Figure 2E](#f2){ref-type=\"fig\"}). In addition, N332 makes two direct hydrogen bonds with the adenine base of A11\u2032 of the complementary strand, but mutating N332 to A had no or little effect in our binding assay.\n\nThe interaction of a lysine, K335, of this helix might be the most noteworthy. The \u025b-amino group of K335 makes a bifurcated hydrogen bond with carbonyl oxygens of two adjacent guanine bases (G8 and G9) in the major groove. Mutation of this lysine to alanine (K335A) not only completely abrogates binding of HOT1 to the 5\u2032-TTAGGG-3\u2032 motif, suggesting that this residue is essential for binding ([Figures 2E and F](#f2){ref-type=\"fig\"}), but this mutation also results in the 'gained\\' ability to bind a nontelomeric control sequence (5\u2032-GTGAGT-3\u2032) in our pull-down experiments. To our understanding, this 'gain-of-function\\' is a strong indication that K335 is a critical determinant for recognizing and binding to the telomeric sequence and, thus, for the specificity of HOT1.\n\nCooperatively, residues in helix 3 recognize a 5\u2032-AGGGT-3\u2032 (=A\u2032) motif in the telomeric DNA. Furthermore, residues K325, N328 and R334 of this helix bind to the ribose phosphate backbone either directly (K325 to A11\u2032, and R334 to A6) or through water (N328 to A11\u2032 and C12\u2032). Apart from helix 3, we observe sequence-specific interactions from residues in the \u03b11--\u03b12 loop (Y291 binding to the phosphate O of G7) and the N-terminal arm. In full-length HOT1, this arm is predicted to be part of an unstructured stretch of \u223c65 residues between the homeodomain and the preceding putative POUs domain. Notably, in the structures of TRF1 and TRF2 this arm was discussed to be important for extending the DNA-binding interface by adding specificity and affinity ([@b19]). In our structure, the hydroxyl group of S270 of this arm binds to the G14 backbone phosphate, and R271 specifically contacts both base moieties of the A12--T9\u2032 base pair, either directly (T9\u2032), or through a water molecule (A12; [Figure 2F](#f2){ref-type=\"fig\"}). The importance of R271 was also confirmed by the fact that mutating this residue to alanine (R271A) completely abolishes DNA binding ([Figure 2E](#f2){ref-type=\"fig\"}). As outlined in [Figure 2D](#f2){ref-type=\"fig\"}, the homeodomain of HOT1 docks onto telomeric DNA by specifically interacting with nucleobases and backbone groups of more than one telomeric repeat. In particular, R271 binds to the A12--T9\u2032 base pair of the following repeat. This overlapping binding, which was also seen in the structures of TRF1 and TRF2, accounts for the repetitive nature of telomeric DNA and probably provides an extra level of selectivity.\n\nWhile TRF1 and TRF2 also recognize telomeric dsDNA via their respective homeobox domains, they only share low sequence similarity with the homeodomain of HOT1 (13 and 5%, respectively, [Figure 3A](#f3){ref-type=\"fig\"}) and DNA binding is different in both cases. TRF1 and TRF2 both bind DNA via highly conserved interactions, involving direct interactions of \u03b13 to TTxGG (direct contacts with phosphate (x) or bases), or xAGGGTx when water-mediated contacts are included. In HOT1, \u03b13 of the homeodomain recognizes AxGGT directly (or AxGGTxAx with all resolved contacts; [@b19]). When aligning and superimposing the DNA moieties of the TRF1 and HOT1 crystal structures, the difference in binding is easily visualized ([Figure 3B](#f3){ref-type=\"fig\"}). With respect to TRF1, HOT1 is shifted 'down\\' (5\u2032\u21923\u2032) the major groove by about one base and towards the subsequent telomeric repeat.\n\nOur crystal structure reveals in atomic detail how HOT1 specifically recognizes and binds 5\u2032-TTAGGG-3\u2032 repeats by means of its homeodomain. We could underpin these findings with mutational analyses and DNA pull-down assays, and we conclude that K335 is a key residue for telomere recognition and binding in HOT1.\n\nHOT1--telomere associations correlate with the degree of telomere processing\n----------------------------------------------------------------------------\n\nTo investigate the interaction between HOT1 and telomeres *in vivo*, we analysed HOT1 intracellular localization by immunoFISH (immunohistochemistry combined with fluorescence *in situ* hybridization) microscopy in HeLa cells. In agreement with previous observations ([@b25]), HOT1 showed a nuclear-punctuated localization pattern, but in contrast to TRF1 and TRF2 only associated with a subset of telomeres. Colocalization between HOT1 and telomeric DNA was observed in about 90% of all cells, with on average 4.5 HOT1 and telomeric foci colocalizing per cell ([Figure 4A](#f4){ref-type=\"fig\"}, [Supplementary Figure S4a](#S1){ref-type=\"supplementary-material\"}). To study the HOT1 localization in another cell type, we performed IF stainings in mouse ES cells ([Figure 1D](#f1){ref-type=\"fig\"}). Interestingly, in this setting we observed a higher degree of colocalization of HOT1 with telomeres, with an average of 13.5 HOT1 and TRF1 foci colocalizing ([Figure 4B](#f4){ref-type=\"fig\"}). In telomerase-positive cancer cells, such as HeLa, only a small fraction of telomeres is actively processed, as telomerase is only associated with a few telomeres at any given time ([@b36]; [@b62]). Here most telomeres are believed to exist in a 'closed\\' state, while only some telomeres are in an 'open\\', actively processed state. In contrast, ES cells have high telomerase activity and maintain long telomeres ([@b64]). We reasoned that more colocalization events of HOT1 with telomeres might reflect the degree of active telomere extension in these cells. If this was true, then cells that very actively extend their telomeres should have the highest frequency of HOT1--telomere associations. To test this hypothesis, we carried out IF stainings for HOT1 on pachytene chromosome spreads from mouse spermatocytes, in which telomerase very actively elongates telomeres ([@b26]; [@b40]; [@b73]; [@b56]; [@b59]), using SYCP3 as a marker for the synaptonemal complex/the chromosome axis ([@b2]). In agreement with a correlation between telomere processing and HOT1--telomere association, HOT1 robustly localized to telomeres on all spreads analysed with on average 87% of chromosome ends being HOT1 positive ([Figure 4C](#f4){ref-type=\"fig\"}). We also confirmed this frequent association of HOT1 with telomeres in mouse testes sections using the same IF set-up. Again, we could confirm the prominent localization of HOT1 to telomeres in SYCP3-positive pachytene cells, highlighting the *in vivo* relevance of our discovery ([Supplementary Figure S5](#S1){ref-type=\"supplementary-material\"}). Together, this data validates the *in vivo* localization of HOT1 to telomeres and suggests that HOT1 is a dynamic telomere-binding protein, putatively associated with actively processed telomeres.\n\nHOT1 interacts with the active telomerase complex and associates with CBs\n-------------------------------------------------------------------------\n\nTo gain further insights about HOT1 function at telomeres, we performed HOT1 immunoprecipitation assays combined with quantitative MS ([@b67]), using both a polyclonal rabbit and a monoclonal mouse HOT1 antibody and SILAC-labelled nuclear protein extracts from HeLa cells. HOT1 itself was recovered with SILAC ratios indicative of specific binding to the antibody ([Figure 5A](#f5){ref-type=\"fig\"} and [Supplementary Figure S6a](#S1){ref-type=\"supplementary-material\"}). We also identified several proteins relevant to telomere homeostasis: the four core components of box H/ACA snoRNPs (DKC1, GAR1, NHP2 and NOP10; all part of the active telomerase RNP ([@b68])) and the Ku70--Ku80 heterodimer proteins. We validated several of these interactions by co-IP experiments, confirming a physical interaction of endogenous HOT1 with components of the active telomerase complex ([Figure 5B](#f5){ref-type=\"fig\"}, [Supplementary Figure S6](#S1){ref-type=\"supplementary-material\"} and [Supplementary Tables S3--S6](#S1){ref-type=\"supplementary-material\"}). In addition, CB proteins, notably the CB-scaffolding and -marker protein Coilin ([@b18]), were strongly enriched in our IP experiments ([Figures 5A, B](#f5){ref-type=\"fig\"}, [Supplementary Figure S6a](#S1){ref-type=\"supplementary-material\"} and [Supplementary Tables S3--S6](#S1){ref-type=\"supplementary-material\"}), indicating that HOT1 functions in the CB, where telomerase is assembled and potentially brought together with telomeres ([@b18]; [@b36]; [@b62]), or in association with CB components in the nucleoplasm. We could further validate these interactions with analogous immunoprecipitations using extracts from mouse ES cells ([Supplementary Figure S6b](#S1){ref-type=\"supplementary-material\"} and [Supplementary Table S7](#S1){ref-type=\"supplementary-material\"}). It is noteworthy that in none of our HOT1 IPs any of the shelterin components were enriched and that in agreement with this HOT1 was not identified in a reciprocal IP of the shelterin component POT1 ([Supplementary Figure S6c](#S1){ref-type=\"supplementary-material\"} and [Supplementary Table S8](#S1){ref-type=\"supplementary-material\"}), suggesting that HOT1 does not directly interact with the shelterin complex.\n\nThe interaction of HOT1 with the active telomerase complex components box H/ACA snoRNPs subunits ([@b68]) raises the possibility that HOT1 binds to active telomerase. To substantiate this hypothesis, we performed immunoprecipitation experiments followed by telomerase activity measurements using the quantitative TRAP (telomere repeat amplification protocol) assay for immunoprecipitates obtained from IPs for both HOT1 and the positive control DKC1. Indeed, we detected telomerase activity in both cases ([Figures 5C](#f5){ref-type=\"fig\"} and [5D](#f5){ref-type=\"fig\"}), while five other nuclear DNA-binding proteins (TBP, YY1, STAT3, Histone3\\[K4me3\\] and CENP-B) did not show any enrichment of telomerase activity compared to an IgG control. Antibodies against TRF1 and TRF2 immunoprecitated telomerase activity, putatively dependent on the previously established link between the shelterin component TPP1 and telomerase ([@b72]). In sum, we conclude that HOT1 associates with the active telomerase complex.\n\nTo corroborate the physical interaction between HOT1 and Coilin, we performed IF stainings for HOT1 and Coilin in nonsynchronized HeLa cells. After deconvolution and 3D reconstruction of the IF images the colocalization between both proteins was analysed. In about 85% of all cells analysed, we observed colocalization of one to seven HOT1 foci with CBs ([Figure 5E](#f5){ref-type=\"fig\"}). Remarkably, HOT1 foci preferentially localized to the periphery of Coilin, reminiscent of previous findings on the association of telomerase RNA and telomeres with CBs ([@b36]). This data underscores the idea that HOT1 might be associated with actively processed telomeres in a telomerase-positive context.\n\nHOT1 is a positive regulator of telomere length\n-----------------------------------------------\n\nThe differential degree of the HOT1--telomere colocalization, and the association with the active telomerase complex as well as the reminiscent localization to the periphery of CBs, suggested that HOT1 might be involved in telomere maintenance. To investigate this, we depleted both HOT1 and TCAB1 in HeLa cells with endoribonuclease-prepared siRNA (esiRNA; [@b39]; [Figure 6A](#f6){ref-type=\"fig\"} and [Supplementary Figure S7](#S1){ref-type=\"supplementary-material\"}). Here TCAB1, as a recently described DKC1 interaction partner necessary for proper trafficking of TERC ([@b65]; [@b77]), serves as a reference for the extent of telomere shortening that can be assigned as telomerase dependent. From HOT1- and TCAB1-depleted, as well as control-transfected cells, metaphase spreads were prepared 3 days post transfection and telomere length was determined using quantitative telomeric FISH ([@b47]). Knockdown of both HOT1 and TCAB1 resulted in significant telomere shortening ([Figure 6B](#f6){ref-type=\"fig\"} and [Supplementary Figure S7](#S1){ref-type=\"supplementary-material\"}) as indicated by reduced FISH signals. Here the entire populations of FISH signals shifted towards weaker signals, indicating that telomeres globally shortened in the absence of HOT1. In these experiments the distributions after HOT1 and TCAB1 knockdown were indistinguishable from each other, indicating that their effect on telomere homeostasis is similar. Consistent with this finding we also observed a significant increase in the appearance of signal-free chromosome ends both upon HOT1 and TCAB1 knockdown ([Figures 6B and C](#f6){ref-type=\"fig\"}, and [Supplementary Figure S7](#S1){ref-type=\"supplementary-material\"}). We also obtained qualitatively similar results by measuring telomere length after HOT1 and TCAB1 knockdowns 3 days post transfection by a universal, single telomere-elongation length analysis (STELA; [Figure 6D--F](#f6){ref-type=\"fig\"}). Again, depletion of both HOT1 and TCAB1 led to shorter telomere length with a shortening of on average 700 and 600 bp, respectively. Given the short nature of telomeres in this cell line (on average, 5 kb measured by STELA) and the substantial contribution of variant repeats and subtelomeric DNA to this value ([@b5]), our quantitative FISH and STELA data mutually confirm HOT1 as a positive regulator of telomere length that quantitatively behaves similar to the established telomerase pathway member TCAB1.\n\nNext we performed telomere length measurements after HOT1 depletion at different time points. We reasoned that if HOT1 is influencing telomere length in a mode dependent on telomerase activity, its depletion should lead to a gradual loss of telomeric tracts. To test this, we performed telomere length measurements upon HOT1 depletion as a time course over several days. Indeed, the telomeric signal was gradually reduced in the absence of HOT1 ([Figures 7A and B](#f7){ref-type=\"fig\"}). Hence, the telomere length changes observed due to manipulations of HOT1 levels are in agreement with telomerase-dependent changes.\n\nIn a complementary experiment, we transiently overexpressed FLAG--HOT1 in HeLa cells and analysed telomere length 3 days after transfection. Consistent with the observed telomere shortening upon HOT1 depletion, telomeres were elongated upon its overexpression as indicated by an increase of the entire population of FISH signals ([Figures 7C and D](#f7){ref-type=\"fig\"}). The qFISH data suggests that the telomere length regulation exerted by HOT1 is dependent on its telomere-binding capacity. To test whether HOT1-dependent telomere elongation indeed requires DNA binding, we used a HOT1\u0394Homeobox variant. *In vitro*-binding studies with recombinant HOT1\u0394Homeobox demonstrated that the deletion of this domain indeed abolished binding to telomeric 5\u2032-TTAGGG-3\u2032 repeats ([Figure 7E](#f7){ref-type=\"fig\"}) and the lack of binding also coincides with loss of telomeric localization *in vivo* ([Supplementary Figure S4b](#S1){ref-type=\"supplementary-material\"}). To examine functional consequences of DNA-binding-deficient HOT1 on telomere length regulation, we overexpressed FLAG--HOT1\u0394Homeobox in HeLa cells. In contrast to telomere elongation upon expression of full-length FLAG--HOT1, the expression of FLAG--HOT1\u0394Homeobox did not lead to an effect on telomere length in comparison to the control ([Figures 7C and D](#f7){ref-type=\"fig\"}). Thus, the HOT1--telomere interaction is essential for telomere elongation. Collectively, these findings demonstrate that HOT1 acts as a positive regulator of telomere length. The fact that overexpression of HOT1 can stimulate telomere lengthening further suggests that similar to TERT and TERC ([@b22]), HOT1 is a limiting factor, at least in this cellular context.\n\nTERT binding to chromatin is dependent on HOT1\n----------------------------------------------\n\nThe combined features of HOT1---direct binding to telomeric dsDNA, preferential localization to telomeres in settings of active processing, association with active telomerase, localization to the periphery of CBs and a positive effect on telomere length---suggest a putative role of HOT1 in telomerase recruitment. To test this possibility we assayed TERT binding to chromatin in the absence of HOT1, based on a previous experimental set-up that contributed to the identification of TPP1 as a telomerase recruiter ([@b60]). To avoid potential convolutions of experimental results by the presence of residual HOT1 protein, we established a HOT1 genetrap mouse. In *Hot1*^Gt(pU-21T)346Card/(Gt(pU-21T)346Card^ mice, the genetrap construct is inserted in the first intron of the *Hot1* gene. As the start codon resides in exon 2, we expected that protein synthesis would be entirely abrogated ([Figure 8A and B](#f8){ref-type=\"fig\"}). To verify this, we established mouse embryonic fibroblasts (MEFs) of the different genotypes from littermate embryos and performed immunofluorescence stainings with our HOT1 antibody. While wild-type cells showed the typical nuclear-punctuated pattern, HOT1 signals were absent in *Hot1*^Gt(pU-21T)346Card/(Gt(pU-21T)346Card^ MEFs ([Figure 8C](#f8){ref-type=\"fig\"}). Based on this data and for the sake of simplicity, we will refer to this as *Hot1*^\u2212/\u2212^ genotype. We then performed the chromatin fractionation assay and could, as expected, detect TERT on the chromatin fraction of wild-type MEFs, but not on chromatin fractions from *Tert*^\u2212/\u2212^ MEFs. Notably, the TERT signal was also absent when extracts of *Hot1*^\u2212/\u2212^ MEFs were used ([Figure 8D](#f8){ref-type=\"fig\"}), reminiscent of previous data on TPP1 ([@b60]). This data indicate that HOT1 indeed contributes to telomerase recruitment *in vivo*. Thus, HOT1 as a dynamic and differential telomere-binding protein contributes to telomerase--telomere association, mechanistically underlying its function as a positive regulator of telomere length.\n\nDiscussion\n==========\n\nWe have adapted our SILAC-based DNA--protein interaction screening technique to repetitive sequences as a straightforward and robust approach to identify novel interactions. In addition to the established telomere-binding proteins and their tight interaction partner, we have identified HOT1 as the first direct telomere-binder that positively regulates telomere length. HOT1 has previously been found as a telomere-associated factor that localizes to some telomeres in HeLa cells using the PICh approach ([@b25]). Interestingly, this is the only one of several recent large scale studies searching for telomere-associated factors, which has identified HOT1 ([@b25]; [@b28]; [@b53]; [@b42]), possibly because also in this approach telomeric DNA was used as bait, rather than using the shelterin components. While by the PICh approach several hundreds of putative telomere-associated factors were identified and the PICh technique is, in general, addressing the global composition of (telomeric) chromatin, our approach is more apt to rapidly identify the direct DNA binders and their tight interaction partners as illustrated by the identification of HOT1. Thus, our *in vitro* reconstitution assay seems particularly suited to discover novel direct telomere binders.\n\nIn contrast to the other direct telomere-binding proteins, TRF1 and TRF2, HOT1 does not localize to all telomeres and the degree of telomere association varies between cell types. While *in situ* telomerase activity is limited in HeLa cells to a simple maintenance of short telomeres, it is increased in mouse ES cells and is very high in mouse spermatogonia and spermatocytes, where telomere length is reset to the very long telomeres observed in mature sperm ([@b26]; [@b40]; [@b73]; [@b56]; [@b59]). The fact that the nature of the HOT1 association with telomeres seems to be more transient in normally cycling telomerase-positive cancer cells compared to an intermediate degree of association in mouse ES cells, and to a prominent and robust association with telomeres in mouse spermatocytes, is consistent with promoting telomerase-dependent telomere lengthening. This raises the possibility that HOT1 selectively interacts with actively processed telomeres. How exactly this interaction is regulated and which precise step in telomere lengthening HOT1 promotes will have to be addressed in the future.\n\nWhile HOT1 localizes to a subset of telomeres together with shelterin components, our HOT1 immunoprecipitation experiments failed to establish any putative association between HOT1 and shelterin complex members. Likewise, we could not retrieve HOT1 after a POT1--IP, which is in agreement with several other studies that searched for shelterin-associated factors either by immunoprecipitation or by bimolecular fluorescence complementation ([@b28]; [@b53]; [@b42]). Our comparison of the cocrystal structures of telomeric DNA with the HOT1, TRF1 and TRF2 homeodomains indicates that while HOT1 is shifted 'down\\' in 5\u2032\u21923\u2032 direction by one base towards the following telomeric repeat, the binding sites are largely overlapping. Hence, it is intriguing how these proteins coexist at telomeres, how they compete for binding sites and whether the shelterin proteins are found interspersed on telomeres or whether there are discrete, mutually exclusive patches along the telomeric tracts. This might further contribute to answering the question how HOT1 is selectively restricted to a subset of telomeres.\n\nHow does HOT1 fit into the current view of telomere homeostasis? TCAB1 (WDR79/WRAP53) has been shown to be required for proper localization of CAB box containing scaRNPs to CBs, including TERC ([@b63]; [@b65]). Similar to HOT1, TCAB1 acts as a positive regulator of telomere length by recruiting TERC to CBs ([@b65]). The telomere-binding properties of HOT1 and its functional consequences, which are quantitatively similar to those of TCAB1, suggest that HOT1 may act downstream of TCAB1. The putative mechanism may involve recruiting telomerase-containing CBs to telomeres, and thus promoting telomerase association with telomeres, perhaps through additional interactions involving TPP1 ([@b1]; [@b60]; [@b75]). The fact that overexpression of HOT1 further stimulates telomere lengthening suggests that HOT1, similar to TERT and TERC, is also a limiting factor. Therefore, the shelterin and CST complexes, and the controlled levels of TERT, TERC and HOT1, may all act cooperatively to define a mean telomere length in the cell.\n\nOur chromatin fractionation analysis indicates that HOT1 is required for TERT recruitment to chromatin and, therefore, qualifies as a telomerase recruitment factor. Similarly, TPP1 has been established as a telomerase recruiter likely by (direct or indirect) interaction with TERT via its TEL patch ([@b52]), but based on our immunoprecipitation data and previous studies there is no indication for an interaction between HOT1 and TPP1. This discrepancy may be explained by a step-wise telomerase recruitment model, in which HOT1 simply acts before TPP1. Previous reports have shown that POT1--TPP1 act as a processivity factor for telomerase ([@b69]; [@b41]; [@b75]). Enhanced *in vivo* processivity would also result in an increased residence time and a higher affinity to telomeres, thus explaining the reduction of telomerase--telomere association in TPP1 knockdown/knockout conditions. Given that HOT1 preferentially localizes to telomeres in settings of high *in situ* telomerase activity, one possibility is that HOT1 contributes to an 'open\\' telomere state that allows/promotes telomere elongation. In such a model, HOT1 might contribute to rendering telomeres accessible and/or contributes to the delivery of telomerase to 'open\\' telomeres. In this scenario, TPP1 would contribute more prominently to maintaining telomerase at telomeres and, thus, to processivity stimulation. Undoubtedly, the mechanistic contribution of both factors and potential genetic interactions are exciting avenues for future research.\n\nMaterials and methods\n=====================\n\nCell culture\n------------\n\nHeLa cells (Epitheloid carcinoma, cervix) and MEFs were cultivated in 4.5 g/l glucose Dulbecco\\'s modified Eagle\\'s medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 \u03bcg/ml streptomycin (Gibco) at 37\u00b0C and 5% CO~2~. For SILAC labelling, HeLa cells were incubated in RPMI 1640 (--Arg, --Lys) medium containing 10% dialysed fetal bovine serum (Gibco) supplemented with 84 mg/l ^13^C~6~^15^N~4~ [L-]{.smallcaps}arginine and 50 mg/l ^13^C~6~^15^N~2~ [L-]{.smallcaps}lysine (Sigma Isotec or Euriso-top) or the corresponding nonlabelled amino acids, respectively. R1/E murine ES cells were also grown in 4.5 g/l GlutaMAX DMEM containing sodium pyruvate, but supplemented with 20% FBS, 50 \u03bcM 2-mercaptoethanol (Invitrogen), LIF (MPI-CBG protein expression facility), as well as 100 U/ml penicillin and 100 \u03bcg/ml streptomycin (Gibco). For SILAC labelling of R1/E cells, DMEM (--Arg, --Lys) medium containing 20% dialysed FBS (Gibco) supplemented with 40 mg/l ^13^C~6~^15^N~4~ [L-]{.smallcaps}arginine and 80 mg/l ^13^C~6~^15^N~2~ [L-]{.smallcaps}lysine (Sigma Isotec or Euris-top) or the corresponding nonlabelled amino acids, respectively, was used. Cells were collected and nuclear extracts were prepared as described ([@b13]).\n\nTelomere pull-down\n------------------\n\nChemically synthesized oligonucleotides ([Table II](#t2){ref-type=\"table\"}; Metabion) were annealed and polymerized by T4 ligase (Fermentas) and biotinylated with biotin--dATP (Invitrogen) by Klenow fragment (Fermentas) following the manufacturer\\'s instructions. Twenty-five microgram baits were immobilized on 750 \u03bcg paramagnetic streptavidin beads (Dynabeads MyOne C1, Invitrogen) and subsequently incubated with either 400 \u03bcg of SILAC-labelled nuclear extract or 20 \u03bcl of supernatants of crude *E. coli* lysates (3 \u03bcg/ml) in PBB buffer (150 mM NaCl, 50 mM Tris--HCl pH 8.0, 10 mM MgCl~2~, 0.5% NP-40, Complete Protease Inhibitor without EDTA (Roche)) for 2 h at 4\u00b0C on a rotation wheel. After three washes with PBB buffer bead fractions were pooled, bound proteins boiled at 80\u00b0C in 1 \u00d7 LDS buffer (Invitrogen) and separated on a 4--12% gradient gel (Novex, Invitrogen; [@b13]).\n\nMS data acquisition\n-------------------\n\nIn-gel digestion and MS analysis was performed essentially as previously described by [@b13]. Peptides were desalted on StageTips and analysed by nanoflow liquid chromatography on an EASY-nLC system from Proxeon Biosystems coupled to a LTQ-Orbitrap XL or a Orbitrap-Velos (Thermo Electron). Peptides were separated on a C18-reversed phase column (15 cm long, 75 \u03bcm inner diameter, packed in-house with ReproSil-Pur C18-AQ 3 \u03bcm resin (provided by Dr Maisch)) directly mounted on the electrospray ion source. We used a 107-min gradient from 2 to 60% acetonitrile in 0.5% acetic acid at a flow of 200 nl/min. The LTQ-Orbitrap XL was operated with a Top5 MS/MS spectra acquisition method in the linear ion trap per MS full scan in the Orbitrap, while for the Orbitrap-Velos a Top10 acquisition method with HCD fragmentation was used.\n\nMS spectrum and data analysis\n-----------------------------\n\nThe raw files were processed with MaxQuant ([@b20]) and searched with the Mascot search engine (Matrix Science) against IPI human v3.37 protein database concatenated with a decoy of the reversed sequences. Carbamidomethylation was set as fixed modification, while methionine oxidation and protein N-acetylation were considered as variable modifications. The search was performed with an initial mass tolerance of 7 p.p.m. mass accuracy for the precursor ion and 0.5 Da for the MS/MS spectra obtained with CID fragmentation and 20 p.p.m. for the MS/MS spectra in the HCD fragmentation mode. Search results were processed with MaxQuant filtered with a false-discovery rate of 0.01. Before statistical analysis, known contaminants and reverse hits were removed. The protein ratios of a 'forward\\' experiment and the 'reverse\\' experiment were plotted in R (prerelease version 2.8.0). Only proteins identified with at least two unique peptides and two quantitation events were plotted for the telomere pull-down (quality filter).\n\nRecombinant protein expression and deletion variant construction\n----------------------------------------------------------------\n\nThe HOT1 clone was obtained from the ORFeome collection (IOH40784, Invitrogen). The sequence was subcloned into SLIC-compatible pETM44 or pETM14 vector via SLIC cloning ([@b44]; [@b58]) and was expressed in *E. coli* Rosetta at 18\u00b0C. *E. coli* extracts with overexpressed recombinant proteins were used for binding studies on variant DNA motives. For crystallization, protein was purified under nondenaturing conditions using Ni^2+^ sepharose high-performance columns (GE Healthcare) followed by size-exclusion chromatography on Superdex 75 (GE Healthcare), using 150 mM NaCl and 20 mM Tris--HCl (pH 7.5) as a running buffer.\n\nThe HOT1 ORFeome clone was LR recombined into a gateway compatible pcDNA3.1 vector with N-terminal FLAG-tag (gift of Christian Brandts). A FLAG-tagged HOT1 homeobox deletion variant was constructed by PCR amplification of the pcDNA3.1-FLAG vector using primers with site-specific overhangs ([Table III](#t3){ref-type=\"table\"}; Metabion) following the QuickChange II Site-directed mutagenesis kit protocol (Stratagene; [@b76]). Point mutants were generated similarly using the QuickChange protocol. For expression in ES cells, constructs were subcloned into a pCAGGS vector. The constructs were sequence verified using an Applied Biosystems 3730 Genetic Analyzer (Applied Biosystems), according to the manufacturer\\'s instructions.\n\nCrystallization and structure determination\n-------------------------------------------\n\nThe complex of human HOT1 DBD and double-stranded telomeric DNA (5\u2032-TTAGGGTTAGGGTTA-3\u2032) was obtained as follows: synthetic single-strand complementary oligonucleotides ([Table IV](#t4){ref-type=\"table\"}; Eurofins MWG) were dissolved in water and mixed to yield a 5-mM stock solution. For double-strand annealing, this solution was heated to 96\u00b0C for 10 min and allowed to cool down to room temperature (RT) in the switched-off heat block within 2--3 h. This solution was mixed in a 1--1.2 M excess with a solution of purified HOT1 DBD at 10 mg/ml and incubated on ice for 10 min. Crystals of the complex were grown at 4\u00b0C by sitting-drop vapour diffusion from drops formed by mixing equal volumes of the complex solution and the crystallization condition 24% PEG 3350, 100 mM Na Acetate (pH 4.5) and 100 mM KSO~4~. For data collection, crystals were cryoprotected by soaking in a mother liquor supplemented with 20% glycerol and was flash-frozen in liquid nitrogen. Diffraction data were collected at the PXII beam line of the Swiss Light Source (SLS; Villigen, Switzerland) and processed using XDS ([@b37]). The structure was solved by molecular replacement with the programme PHASER ([@b49]), using the NMR structure of the HOT1 DBD as a search model (PDB entry 2CUF). Model building and refinement was carried out with Phenix ([@b3]) and COOT ([@b27]), and models were validated with the programme Molprobity ([@b23]).\n\nThe alignment of the HOT1, TRF1 and TRF2 homeobox domain sequences was done in two steps. First, a secondary structure-based alignment was generated using the programme Chimera ([www.cgl.ucsf.edu/chimera](http://www.cgl.ucsf.edu/chimera)) followed by feeding it into the web application Espript (espript.ibcp.fr) for a simple sequence alignment scoring and colouring.\n\nAntibody production\n-------------------\n\nHis-MBP-tagged HOT1 (pETM44 vector construct) was expressed in *E. coli* Rosetta at 18\u00b0C in a fermentation tank. Cell pellet was lysed with Avestin and the soluble fraction subjected to affinity purification using Ni-sepharose. The elution fraction was concentrated with an Amicon Ultra 15 concentrator column and dialysed into buffer containing 50 mM K~2~PO~3~, 20 mM NaCl, 10% glycerol, 1 mM TCEP and protein inhibitors. Purified MBP-HOT1 was injected into rabbits for immunization and the rabbits were ultimately sacrificed. The antibody was affinity purified using His-MBP-HOT1 and MBP immobilized on HiTRAP desalting columns (GE Healthcare). First, the serum was applied to the His-MBP-HOT1 column, eluted and then applied to the MBP column. The flow trough was quantified and used for subsequent experiments.\n\nPurified MBP-HOT1 was injected into mice for immunization and the mice were ultimately sacrificed. Immortalized hybrid cells (hybridomas) were obtained by the fusion of B cells from the spleen of an immunized mouse with a myeloma cell line, which itself does not produce antibodies, using PEG and AH selection (Sigma Aldrich). Hybridoma clones were generated and screened using the Meso Scale Discovery platform (Meso Scale Diagnostics) by comparing affinity to His-MBP-HOT1 and His-MBP-Katanin as an unspecific negative control. Positive clones were subcloned by limiting dilution and retested using the MSD platform. Based on the subcloned hybridoma cell lines, antibodies were purified using HiTRAP protein G columns (GE Healthcare) followed by acid elution.\n\nChromatin immunoprecipitation\n-----------------------------\n\nCells were crosslinked with 1% formaldehyde (Thermo Scientific) for 10 min at 37\u00b0C, and the reaction was stopped by adding glycine to a final concentration of 0.125 M for 5 min at RT. Nuclei were prepared from fixed and washed cells by homogenization in cell lysis buffer (5 mM PIPES pH 8, 85 mM KCl, 0.5% NP-40) and centrifugation at 1800 *g* for 10 min. Finally, nuclei were lysed in 900 \u03bcl nuclei lysis buffer (50 mM Tris--HCl pH 8, 10 mM EDTA pH 8, 1% SDS) and lysates were sonicated for 30 min (30 s on/30 s off) in a Diagenode water-bath sonicator at speed 5. Following a centrifugation at 14 000 r.p.m. for 10 min, the cleared supernatants were snap-frozen in liquid nitrogen and stored at \u221280\u00b0C. Sonication efficiency was routinely monitored by DNA gel electrophoresis to ensure that the bulk of DNA fragments was between 100 and 500 bp.\n\nSonicated chromatin containing 50 \u03bcg DNA was diluted 10 times in ChIP dilution buffer (16.7 mM Tris--HCl pH 8, 167 mM NaCl, 1.2 mM EDTA pH 8, 1.1% Triton X-100, 0.01% SDS) and precleared for 2 h, rotating at 4\u00b0C, with 20 \u03bcl blocked beads (Dynabeads Protein A beads (Invitrogen) incubated for 2 h with 5 mg/ml BSA) before the overnight incubation with 5 \u03bcg of antibody. The following antibodies were used: mouse anti-HOT1 (MPI-CBG Antibody Facility), rabbit anti-TRF2 (NB110-57130, Novus), rabbit IgG (ab37415, Abcam) or mouse anti-GFP (ab1218, Abcam)). The bound material was recovered after a 2 h incubation, rotating at 4\u00b0C, with 30 \u03bcl blocked beads. The beads were washed for 10 min in each of the following wash buffers: low-salt buffer (20 mM Tris--HCl pH 8, 150 mM NaCl, 2 mM EDTA pH 8, 1% Triton X-100, 0.1% SDS), high-salt buffer (20 mM Tris--HCl pH 8, 500 mM NaCl, 2 mM EDTA pH 8, 1% Triton X-100, 0.1% SDS), LiCl buffer (10 mM Tris--HCl pH 8, 0.25 M LiCl, 1 mM EDTA pH 8, 1% NP-40, 1% Na deoxycholate) and twice, 5 min each, in TE. ChIPed material was eluted by two 15 min incubations at RT with 225 \u03bcl elution buffer (0.1 M NaHCO~3~, 1% SDS). Chromatin was reverse-crosslinked by adding 18 \u03bcl of 5 M NaCl and incubated overnight at 65\u00b0C, and DNA was submitted to RNase and proteinase K digestion and extracted by phenol--chloroform.\n\nPurified DNA recovered by ChIP was denatured in 0.2 M NaOH by heating to 100\u00b0C for 10 min and spotted onto a positively charged Biodyne B nylon membrane (Pall, VWR). Membranes were hybridized at 42\u00b0C in 6 \u00d7 SSC, 0.01% SDS, 0.1% milk with 20 pmol of DIG-labelled telomeric C-rich LNA probe (36 bp, Exiqon). Following hybridization washes (twice 5 min in 2 \u00d7 SSC, 0.01% SDS, and once 2 min in 0.1 \u00d7 SSC, 0.01% SDS), the signal was revealed using the anti-DIG-AP antibodies (Roche) and CDP-Star (Roche) following the manufacturer\\'s instructions. Images were obtained using the Luminescent image analyser LAS-4000 mini (GE Healthcare). Following the detection of a telomeric signal, membranes were stripped twice for 15 min each in 0.5% SDS at 60\u00b0C and twice for 15 min each in 0.2 N NaOH, 0.1% SDS at 37\u00b0C. Stripped membranes were then hybridized at 65\u00b0C in 6 \u00d7 SSC, 0.01% SDS, 0.1% milk with P^32^-random-primed labelled total genomic DNA from HeLa cells. Following hybridization washes (10 min in 2 \u00d7 SSC, 0.01% SDS, 10 min in 0.5 \u00d7 SSC, 0.01% SDS and 10 min in 0.1 \u00d7 SSC, 0.01% SDS) the signal was revealed using the phosphorimager and an image obtained by the Typhoon scanner.\n\nImmunofluorescence and immunoFISH stainings on cycling cells\n------------------------------------------------------------\n\nFor immunofluorescence stainings, cells were seeded 24 h before the treatment on either glass coverslips (0.17 mm, assorted glass, Thermo Scientific) or in LabTek II chambered coverglass chambers (Labtek). After a brief wash with 1x PBS cells were fixed in 10% formalin solution (Sigma Aldrich) for 10 min at RT, followed by two washes with 1 \u00d7 PBS+30 mM glycine. Cells were then permeabilized with 1 \u00d7 PBS+0.5% Triton X-100 for 5 min at 4\u00b0C, followed by again two washes with 1 \u00d7 PBS+30 mM glycine. Afterwards, cells were blocked in blocking solution (1 \u00d7 PBS, 0.2% fish-skin gelatine (Sigma Aldrich)) for 15 min at RT. Primary antibodies were diluted in blocking solution and incubated for 1 h at RT. The following primary antibodies were used: mouse anti-HOT1 (MPI-CBG Antibody Facility, 1:1000), rabbit anti-Coilin (sc-32860, Santa Cruz, 1:500), and goat anti-GFP (MPI-CBG Antibody Facility, 1:2000). Cells were washed three times for 3 min each in blocking solution followed by a 30-min incubation at RT with secondary antibodies, which were diluted in blocking solution. As secondary antibodies, fluorescent-labelled donkey anti-rabbit-IgG or donkey anti-mouse-IgG antibodies with either Alexa488, Alexa555, Alexa594 or Alexa647 as fluorochromes (1:500, Invitrogen) were used. After three final washes for 3 min each in blocking solution, slides were briefly rinsed in distilled water and mounted using DAPI Prolong Gold Antifade Reagent (Invitrogen). In the case of Labtek chambers, samples were incubated for 5 min with blocking solution containing 1 \u03bcg/ml DAPI, followed by one wash in blocking solution.\n\nIf FISH stainings were combined with immunofluorescence stainings (immunoFISH), the FISH labelling was carried out directly after the IF protocol. Cells were post-fixed in 10% formalin solution (Sigma Aldrich) for 10 min at RT, followed by three washes in 70, 90 and 100% ethanol for 5 min each, followed by a standard telomeric FISH protocol as previously described.\n\nAll images were acquired with a DeltaVision Core Microscope (Applied Precision, Olympus IX71 microscope) using a 100 \u00d7 /1.4 UPlanSApo oil-immersion objective. *Z*-stacks (0.2 \u03bcm optical sections) were collected and deconvolved using softWoRx (Applied Precision). *Z*-stacks were reconstructed in 3D using Imaris (Bitplane), and colocalization events were determined for signals above the background using the colocalization function.\n\nesiRNA synthesis\n----------------\n\nThe detailed protocol of esiRNA production has been previously published ([@b39]). Briefly, optimal regions for designing esiRNAs were chosen using the Deqor design algorithm ([@b32]) in order to fulfill two criteria: to obtain the most efficient silencing trigger in terms of silencing efficiency, and to get lowest chances to cross-silence other genes. The most favourable fragments were used to design gene-specific primers ([Table V](#t5){ref-type=\"table\"}) using the Primer3 algorithm (<http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi>). Two esiRNAs for HOT1 and one each for TRF1 and TCAB1 were designed and synthesized. PCR products for the esiRNA production were sequenced using an Applied Biosystems 3730 Genetic Analyzer (Applied Biosystems) according to the manufacturer\\'s instructions. All positions of sequence trace files were confirmed by manual inspection.\n\nesiRNA and plasmid transfection\n-------------------------------\n\nFor all transfections, HeLa cells were seeded in six-well plates (40 000 cells per well corresponding to 9.6 cm^2^) dish and incubated overnight before transfection. For esiRNA transfection, 30 \u03bcl Oligofectamine (Invitrogen) were diluted in 250 \u03bcl OptiMEM (Invitrogen) and incubated for 5 min at RT. In a separate tube, 2 \u03bcg esiRNA were diluted in 250 \u03bcl OptiMEM. Solutions were combined, mixed and incubated for 20 min at RT after which the transfection mix was evenly distributed over the dish.\n\nThe FLAG--HOT1 and the homeobox deletion variant construct were transfected using either Effectene (Qiagen) or Lipofectamine 2000 (Invitrogen) as the transfection reagent according to the manufacturers\\' instructions.\n\nTelomeric quantitative FISH\n---------------------------\n\nFor metaphase preparation, cells were incubated for 4 h with 200 nM nocodazole in order to induce mitotic arrest. A hypotonic shock was achieved in 0.03 M sodium citrate at 37\u00b0C for 40 min. Cells were fixed in an ethanol/acetic acid solution (3:1) and washed three times in this fixing reagent. Metaphase spreads were obtained by dropping suspensions of fixed cells onto clean glass slides.\n\nThe QFISH procedure was carried out as described, using an Alexa488-O-O-(CCCTAA)~3~ or Cy3-O-O-(CCCTAA)~3~ PNA probe (Panagene) ([@b47]).\n\nTelomeric signals were quantified using the iVision software (Chromaphor). Telomere signals were segmented manually and average pixel intensities from every segment were quantified. For each metaphase, the average background intensity was determined and subtracted from individual telomere signals. Statistical analyses were done using Student\\'s *t*-test.\n\nThe number of signal-free ends per metaphase was determined by manual inspection of the same metaphase images that were used for telomere signal-intensity quantification. Statistical analyses were done using Student\\'s *t*-test.\n\nUniversal STELA\n---------------\n\nOne hundred picogram of *Nde*I/*Mse*I (NEB)-digested DNA were amplified according to the previously published Universal STELA PCR amplification protocol ([@b8]) with the following cycling conditions: 68\u00b0C for 5 min, 95\u00b0C for 2 min, 26 cycles of 95\u00b0C for 15 s, 58\u00b0C for 30 s and 72\u00b0C for 12 min, and 72\u00b0C for 15 min. In brief, 10 ng of DNA were digested with 0.4 units of *Nsde*I and 0.4 units of *Mse*I, annealed at 16\u00b0C overnight to 4.2 \u03bcM of both 42-mer (5\u2032-TGTAGCGTGA-AGACGACAGAAAGGGCGTGGTGCGGACGCGGG-3\u2032)and 11+2-mer (5\u2032-TACCCGCGTCCGC-3\u2032) panhandle-oligos with 1.3 units of T4 ligase (NEB) and ligated to 1 nM of telorette3 (5\u2032-TGCTCCGTGCATCTGGCATCCCTAACC-3\u2032) at 35\u00b0C over night. One hundred picogram of ligated DNA was then used for PCR amplification with 0.1 \u03bcM of Adapter (5\u2032-TGTAGC-GTGAAGACGACAGAA-3\u2032) and of Teltail (5\u2032-TGCTCCGTGCATCTGGCATC-3\u2032) primers by the Failsafe Enzyme with the FailSafe buffer premix H (Epicentre). The PCR products were resolved on a 1.2% agarose gel, blotted onto a nylon membrane by capillary osmosis and hybridized at 42\u00b0C over night to a DIG-labelled telomeric C-rich LNA probe (36 bp, Exiqon). The blots were then washed with Maleic Acid/Tween solution and DIG signal was revealed with anti-DIG-AP antibody (1:20 000, Roche) and CDP-star solution (Roche) according to the manufacturer\\'s instructions. Statistical analyses were done using Student\\'s *t*-test.\n\nProtein immunoprecipation\n-------------------------\n\nImmunoprecipitations were carried out using the Dynabeads Protein G or A immunoprecipitation kit (Invitrogen). Fifty microlitre of beads were treated with 10 \u03bcg rabbit anti-HOT1, mouse anti-HOT1 (both MPI-CBG Antibody Facility), rabbit IgG (sc-66931, Santa Cruz; 2729s, Biolabs) or mouse IgG (ChromPure, Jackson ImmunoResearch) in PBB buffer (150 mM NaCl, 50 mM Tris--HCl pH 8.0, 10 mM MgCl~2~, 0.5% NP-40, Complete Protease Inhibitor--EDTA (Roche)) and subsequently incubated with 400 \u03bcg HeLa nuclear extract (SILAC-labelled if followed by MS analysis) for 2 h at 4\u00b0C on a rotation wheel, followed by three washes with PBB buffer. For MS, bead fractions were pooled, bound proteins were eluted and separated on a 4--12% gradient gel (Novex, Invitrogen). For co-IP experiments followed by western blot, bound proteins were simply eluted and subjected to western blot analysis.\n\nWestern blot\n------------\n\nFor western blot samples were boiled in Laemmli buffer (Sigma Aldrich) and subjected to SDS--PAGE (NuPage 4--12% Bis-Tris gels; Invitrogen). Gels were blotted to nitrocellulose (Protran; Schleicher & Schuell), blocked in 5% nonfat milk in PBST (PBS containing 0.1% Tween-20) for 1 h at RT and incubated over night at 4\u00b0C with primary antibody. The following primary antibodies were used: mouse anti-GFP (Roche Diagnostics, 1:4000 dilution), mouse anti-DM1alpha tubulin (MPI-CBG Antibody Facility, 1:50 000 dilution), mouse anti-Ku70 (sc17789, Santa Cruz, 1:1000 dilution), mouse anti-FLAG (M2, Sigma Aldrich, 1:5000) mouse anti-PCNA (sc-9847, Santa Cruz, 1:1000), mouse anti Histone3 (ab, 1791, Abcam, 1:10 000) and rabbit anti-TERT (gift from Madalena Tarsounas, 1:2000). The next day, membranes were washed three times for 10 min each in 5% milk PBST and were incubated for 1 h at RT with secondary antibody (goat anti-mouse antibody conjugated to horseradish peroxidase, Bio-Rad, 1:4000, or donkey anti-mouse IRDye 800CW or donkey anti-rabbit IRDye 800CW, both LI-COR Odyssey, 1:15 000). Membranes were washed three times for 10 min each in PBST followed by one PBS wash. Bands were visualized with enhanced chemiluminescence Western Blotting Detection Reagents (GE Healthcare) or with the LI-COR Odyssey imaging system. For detection of His-tagged HOT1, TRF1 and TBP, the Penta-His HRP Conjugate Kit (Qiagen) was used according to the manufacturer\\'s instructions. As a molecular weight standard, Spectra Multicolor Broad Range Protein Ladder (Fermentas), Seablue 2 (Invitrogen) and MagicMark XP Western Protein Standard (Invitrogen) were used.\n\nImmunoprecipitation of telomerase activity and quantitative TRAP assay\n----------------------------------------------------------------------\n\nFor the immunoprecipitation of telomerase activity HeLa cells were lysed in lysis buffer (50 mM Tris--HCl (pH 8.0), 150 mM NaCl and 1% NP-40 supplemented with Complete Protease Inhibitor--EDTA (Roche)) for 30 min on ice followed by a 30 min centrifugation step in a table-top centrifuge at 4\u00b0C and 20 000 *g*. Per IP assay, 25 \u03bcl of magnetic protein G beads (Invitrogen) were used. Beads were washed three times with 1 \u00d7 PBS before use and incubated with 5 mg/ml BSA (in 1 \u00d7 PBS) for 1 h at 4\u00b0C on a rotating wheel, while lysates were precleared with uncoated beads for 1 h at 4\u00b0C on a rotating wheel. Per IP assay, 1 mg lysate was incubated with 5 \u03bcg rabbit anti-HOT1 (MPI-CBG Antibody Facility), rabbit anti-DKC1 (ab64667, Abcam), rabbit anti-TRF1 (ab1423, Abcam), rabbit anti-TRF2 (NB110-57130, Novus), rabbit anti-TBP (sc-273, Santa Cruz), rabbit anti-YY1 (ab12132, Abcam), rabbit anti-STAT3 (9132, Cell Signaling), rabbit anti-Histone3K4tri-methylated (07-473, Upstate Antibodies), rabbit anti-CENP-B (ab25734, Abcam) or rabbit IgG (sc-66931, Santa Cruz) in PBS for 2 h at 4\u00b0C on a rotating wheel. BSA-coated beads were added, followed by a second incubation for 2 h at 4\u00b0C on a rotating wheel. Beads were then washed once with PBS, twice with lysis buffer and again once with PBS and finally recovered in 75 \u03bcl Chaps buffer (Chemicon).\n\nThe quantitative TRAP assay was carried out using GoTaq qPCR Master Mix (Promega) and both the TS (5\u2032-AATCCGTCGAGCAGAGTT-3\u2032) and ACX primer (5\u2032-GCGCGGCTTACCCTTACCCTTACCCTAACC-3\u2032) at 200 nM. The reaction was run on a Mx3000p real-time PCR system (Stratagene) with the following protocol: 25\u00b0C for 20 min, 95\u00b0C for 10 min and 32 cycles with 95\u00b0C for 30 s, 60\u00b0C for 30 s and 72\u00b0C for 1 min. Statistical analyses were done using Student\\'s *t*-test.\n\nFor verification that signals were due to the presence of the characteristic TRAP ladder, samples were run on a 20% precast TBE gel (Invitrogen) for 2 h at 200 V. For size reference, the GeneRuler Ultra Low Range DNA Ladder (Fermentas) was used. The gel was stained with EtBr for visualization.\n\nQuantitative real-time PCR\n--------------------------\n\nFor quantification of TCAB1 mRNA levels RNA was extracted with the RNeasy kit (Qiagen), including DNaseI digestion and from the eluted RNA cDNA was synthesized using the SuperScript III first-strand synthesis kit with oligodT according to manufacturer\\'s instructions. qPCR primers ([Table VI](#t6){ref-type=\"table\"}) were used at 70 nM concentration together with the Absolute qPCR SYBR green mix (Abgene) on a Mx3000p real-time PCR system (Stratagene). Target gene mRNA levels were normalized against quantification of GAPDH mRNA levels for housekeeping.\n\nStatistical analyses were done using student\\'s *t*-test.\n\nBAC TransgeneOmics\n------------------\n\nThe following BACs (bacterial artificial chromosomes) were used in this study: human HOT1 RP11-789B24 (Invitrogen), DKC1 RP11-107C18, Coilin RP23-375L19 and POT1 CTD-3053M7 and mouse TRF1 RP24-402F23 (BACPAC Resource Center). A LAP (localization and affinity purification) cassette was inserted as a C-terminal fusion using recombineering. Isolated BAC DNA was transfected and selected for stable integration as described ([@b55]). The BAC RP11-789B24 does not cover the entire *HOT1* gene and was complemented by insertion of a cDNA fragment covering the missing coding and 3\u2032-UTR sequence ([Table VII](#t7){ref-type=\"table\"}).\n\nImmunofluorescence stainings on testes chromosome spreads and testes sections\n-----------------------------------------------------------------------------\n\nTestes were isolated from wild-type 129T2/SvEms male mice, between 4--6 weeks old. To prepare chromosome spreads, the tunica albuginea was detached and the seminiferous tubules were incubated in 500 \u03bcl of 1 mg/ml Collagenase Type I (Gibco) in PBS for 10 min at 32\u00b0C. The tubules were agitated slightly, and the liquid was removed in order to remove a proportion of interstitial cells. A fresh 500 \u03bcl aliquot of 1 mg/ml collagenase was added and incubated at 32\u00b0C for 30 min, disaggregating by pipetting every 10 min. The cells were spun at 2000, r.p.m. for 5 min at 4\u00b0C. The pellet was resuspended in 1 ml cold PBS and filtered through a 40 \u03bcm nylon membrane to create a single-cell suspension.\n\nFifteen microlitre of fixation and permeabilization solution (1% PFA, 5 mM sodium borate pH 8.5, 0.2% Triton X-100) were added to each well (\u00d8 7 \u03bcm) on a 10-well glass slide (StarFrost coating, Engelbrecht). An ImmEdge pen (Vector Laboratories) was used to circle the wells and prevent leakage. The cells were diluted 1:5 into cold 100 mM sucrose in 1 \u00d7 PBS for 2--3 min to allow hypertonic swelling. One to two microlitre of cell suspension were added to each well. The slides were incubated in a humidified chamber for 30 min, followed by 2--3 h of drying in a laminar hood or on the bench at 22\u00b0C. The slides were washed briefly three times in 0.5% Photo-Flo (Kodak) then once in distilled water.\n\nTo prepare testes tissue sections, whole testes were immersed in O.C.T Compound (Tissue-Tek 4583) in specimen molds (Tissue-Tek 4566 Cyromold 15 mm \u00d7 15 mm \u00d7 5 mm) and frozen at \u221280\u00b0C. Seven-micrometer sections were cut using a Leica CM1900 and placed on microscope slides (StarFrost K078; 76 \u00d7 26 mm). Sections were fixed using 4% formaldehyde (Sigma F8775) in 1 \u00d7 PBS for 15 min at 22\u00b0C and permeabilized using 0.15% Triton X-100 for 10 min at 22\u00b0C and washed twice in 1 \u00d7 PBS.\n\nTo perform immunofluorescence for both chromosome spreads and tissue sections, blocking solution (0.2% fish gelatin (Sigma) in PBS\u22120.1% Tween-20) was added for 1 h at RT. Primary antibodies were added for 16 h at RT. The following primary antibodies were used: rabbit anti-HOT1 (MPI-CBG Antibody Facility, 1:1000), rabbit anti-TRF2 (sc-9143, Santa Cruz, 1:100) and mouse anti-SYCP3 (as previously described ([@b2])). The cells were washed 3 \u00d7 in blocking solution for 10 min each and secondary antibodies (goat anti-rabbit Alexa555 or goat anti-mouse Alexa488 at 1:500, Invitrogen) were added for 1 h at RT. After 3 \u00d7 washing in blocking solution for 10 min each, Vectashield (Vecta Laboratories) containing 1 \u03bcg/ml DAPI was added and a glass cover slip was placed on top of the wells and sealed closed. Slides were stored at \u221220\u00b0C.\n\nImages were acquired with a DeltaVision Core Microscope (Applied Precision, Olympus IX71 microscope) using a 100 \u00d7 /1.4 UPlanSApo oil-immersion objective. *Z*-stacks (0.2--0.5 \u03bcm optical sections) were collected and deconvolved using softWoRx (Applied Precision). *Z*-stacks were reconstructed in 3D using Imaris (Bitplane) and colocalization events were determined for signals above the background using the colocalization function. For imaging of entire tubules, separate images were acquired as one panel and stitched together using the stitch function in softWoRx for visualization.\n\n**Hot1**^***Gt(pU-21T)346Card/(Gt(pU-21T)346Card***^ ***mice and generation of MEFs***\n--------------------------------------------------------------------------------------\n\nC57BL/6-CBA-*Hot1*^Gt(pU-21T)346Card/+^ embryos were retrieved from the Center for Animal Resources and Development and are based on the Exchangeable Gene Trap Clones system using the exchangeable pU21 trap vector ([@b4]). Live animals were retrieved by embryo transfer and were outcrossed for several generations against C57BL/6 wild-type animals. MEFs were generated from littermate embryos at E13.5 by crossing C57BL/6-*Hot1*^Gt(pU-21T)346Card/+^ animals with each other.\n\nGenotyping PCR\n--------------\n\nFor genotyping of *Hot1*^Gt(pU-21T)346Card/(Gt(pU-21T)346Card^ mice, primers were designed to amplify the wild-type and mutant allele in one PCR reaction, both sharing the same forward primer. The PCR was carried out with a reaction containing 2.5 \u03bcl 10 \u00d7 PCR buffer, 0.8 \u03bcl 50 mM MgCl~2~, 0.2 \u03bcl 25 mM dNTP mix, 2 \u03bcl 10 \u03bcM WT-F primer, each 1 \u03bcl 10 \u03bcM WT-R and MUT-R primers, 1 \u03bcl TaqRed polymerase (Bioloine) and 2 \u03bcl crude genomic DNA extracts, filled up with HPLC-grade H~2~O to a total volume of 25 \u03bcl (for primer sequences see [Table VIII](#t8){ref-type=\"table\"}). PCR conditions were as follows: initial hotstart denaturation at 94\u00b0C for 3 min followed by 10 cycles with 94\u00b0C for 30 s, 62\u00b0C for 30 s with a 0.5\u00b0C touchdown decline per cycle and 72\u00b0C for 60 s, followed by 25 cycles with 94\u00b0C for 30 s, 57\u00b0C for 30 s and 72\u00b0C for 60 s, finished by a final elongation at 72\u00b0C for 5 min. The reaction was run on a DNA Engine Thermocycler (Bio-Rad). After the reaction the PCR products were checked by standard gel electrophoresis\n\nCell fractionation\n------------------\n\nCell fractionation analysis for TERT binding to chromatin was carried out as previously described, with minor modifications ([@b60]). Per sample one million cells were washed once in cold PBS and then resuspended in 200 \u03bcl ice-cold buffer A+ (10 mM HEPES pH 7.9, 10 mM KCl, 1.5 mM MgCl~2~, 0.34 M sucrose, 10% glycerol, Complete Protease Inhibitor--EDTA (Roche)). Samples were incubated for 5 min on ice and centrifuged for 5 min at 1300 *g* and 4\u00b0C. The supernatant was further cleaned by centrifugation for 25 min at 20 000 *g* and 4\u00b0C, and represents the cytoplasmic fraction. The pellet was washed twice in 500 \u03bcl buffer A+ and resuspended in 100 \u03bcl ice-cold buffer B (3 mM EDTA, 0.2 mM EGTA, Complete Protease Inhibitor--EDTA (Roche)). Samples were incubated for 30 min on ice and centrifuged for 5 min at 1700 *g* and 4\u00b0C. The supernatant was further cleaned by centrifugation for 25 min at 20 000 *g* and 4\u00b0C, and represents the nuclear soluble fraction. The pellet was washed twice in 500 \u03bcl buffer B, resuspended in 100 \u03bcl Laemmli buffer (Sigma Aldrich) and sonicated twice for 10 min each in a water-bath sonicator. This sample represents the chromatin fraction. The other fractions were equally mixed with Laemmli buffer, and all samples were boiled at 95\u00b0C for 5 min and separated on a 4--12% gradient gel (Novex, Invitrogen).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Figures\n\n###### Supplementary Table 1\n\n###### Supplementary Table 2\n\n###### Supplementary Table 3\n\n###### Supplementary Table 4\n\n###### Supplementary Table 5\n\n###### Supplementary Table 6\n\n###### Supplementary Table 7\n\n###### Supplementary Table 8\n\n###### Review Process File\n\nWe thank members from the Buchholz, Mann and Londo\u00f1o laboratories for helpful comments on the manuscript, Charlott-Am\u00e9lie Teutsch, an internship student selected by the F\u00f6rderverein der Biologieolympiade e.V., for performing DNA pull-down experiments, Sebastian Rose for technical assistance and Karla Neugebauer and Karl Lenhard Rudolph for critical reading of an earlier version of the manuscript. We are indebted to Sabine Suppmann and Claudia Franke of the MPI-B core facility for help with recombinant protein expression, the MPI-CBG antibody facility for raising anti-HOT1 antibodies, the MPI-CBG light microscopy facility for assistance with imaging and the MPI-B crystallization facility for help with crystallization screening. We thank Anthony Hyman for providing reagents, and Ina Poser, Martina Augsburg, Marit Leuschner and Andrea Ssykor for help with generating LAP cell lines. We also thank Anja Wehner and Bianca Splettstoesser for excellent technical assistance. The anti-SYCP3 antibody was a kind gift from Christa Heyting. *Tert*^\u2212/\u2212^ MEFs and the anti-TERT antibody were kindly provided by Madalena Tarsounas. Work in the Buchholz laboratory is supported by the Max Planck Society, the University of Technology Dresden, the DFG-Center for Regenerative Therapies Dresden (CRTD), the Bundesministerium f\u00fcr Bildung und Forschung grants Go-Bio (0315105) and NGFN-Plus (01GS1102), and the Deutsche Forschungsgemeinschaft (SFB 655 and SPP 1356). Work in the Mann laboratory is supported by the Max Planck Society, NGFN-Plus (01GS0859) and the 7th Framework Program PROSPECTS (HEALTH-F4-2008-021648). Work in the Londo\u00f1o laboratory was supported by grants from ARC, LCC, ANR and INCa, and is an 'Equipe Labellis\u00e9e Ligue\\'.\n\n*Author contributions*: DK and F Butter initiated the research, and planned and performed most experiments, with assistance from C Benda, M Scheibe, ID, M Stevense, CLN, C Basquin, DBK and RK; MA and KA generated the *Hot1*^Gt(pU-21T)346Card/(Gt(pU-21T)346Card^ strain. JAL and RJ contributed to the planning of the experiments; F Buchholz and MM supervised the research and contributed to the planning of the experiments; all authors contributed to the writing of the manuscript.\n\nDK, F Butter, MM and F Buchholz declare a patent application of this work as patent PCT/EP2011/065943. The remaining authors declare that they have no conflict of interest.\n\n![Detection of specific telomere-interacting proteins. (**A**) A schematic of the quantitative SILAC-based DNA interaction screen with DNA oligonucleotides containing either the telomeric repeat or a control sequence. Specific interaction partners are differentiated from background binders by a SILAC ratio other than 1:1. (**B**) MS spectra of representative peptides from the 'forward\\' pull-down experiment. The heavy peptide partners are easily detected (red dots), while the light partner is barely observable (blue dots) in the mass spectrum. (**C**) Two-dimensional interaction plot: known shelterin components cluster together with HOT1, demonstrating enrichment at the telomere sequence compared to the control sequence. (**D**) Summary of the MS data for HOT1 and the core shelterin components from the SILAC-based DNA--protein interaction screens carried out with nuclear extracts derived from HeLa and murine ES cells.](emboj2013105f1){#f1}\n\n![The DBD of human HOT1 recognizes telomeric DNA in a sequence-specific manner. (**A**) Sequence-specific pull-down of recombinant HOT1, TRF1 (positive control) and TBP (TATA-binding protein, negative control). Proteins were incubated with dsDNA of telomeric repeats (5\u2032-TTAGGG-3\u2032), the control sequence (5\u2032-GTGAGT-3\u2032), the subtelomeric repeat variants (5\u2032-TCAGGG-3\u2032, 5\u2032-TGAGGG-3\u2032 and 5\u2032-TTGGGG-3\u2032, as well as the *C. elegans* telomere repeat 5\u2032-TTAGGC-3\u2032). All DNA substrates were concatemerized from 60 bp oligonucleotides to larger DNA fragments (on average at least 1 kb). (**B**) ChIP of telomeric DNA using antibodies against HOT1, TRF2 (positive control), GFP and IgG (negative controls). Representative slotblot images are shown for ChIP from HeLa extracts after hybridization with a telomeric and genomic control. Input dilutions demonstrate the linearity of the signals acquired. (**C**) Structure of the DBD of HOT1 bound to double-stranded telomeric DNA. The protein is shown as a cartoon representation (orange), whereas DNA is shown as a stick model (grey). The interacting amino acid residues in HOT1 are shown as blue sticks, water molecules as red spheres and protein--DNA contacts are visualized as green dashed lines. (**D**) Schematic representation of all protein--DNA contacts in the complex. (**E**) Sequence-specific pull-down of FLAG--HOT1 and selected single mutations to investigate binding specificity. Proteins were incubated with either telomeric repeats (5\u2032-TTAGGG-3\u2032) or a control oligonucleotide (5\u2032-GTGAGT-3\u2032). (**F**) Atomic details of DNA sequence recognition by HOT1. K335 of helix 3 is involved in direct hydrogen bonding to O6 of G8 and O6 of G9. N332 of helix 3 specifically recognizes A11\u2032 of the complementary strand by forming two direct H-bonds with the bicyclic ring system of A11\u2032 (N6 and N7) (left panel). R271 of the N-terminal arm binds two bases of an AT base pair, directly to T9\u2032 and via a water-mediated H-bond to A12 (right panel).\\\nSource data for this figure is available on the online [supplementary information](#S1){ref-type=\"supplementary-material\"} page.](emboj2013105f2){#f2}\n\n![Comparison of the molecular recognition of telomeric DNA by HOT1 and TRFs. (**A**) Schematic representation of the domain structure of the homeobox domains of TRF1, TRF2 and HOT1. Residues involved in DNA binding are marked with an asterisk (HOT1) or diamond (TRF1). Strictly conserved residues are shown with white font on red background and conserved residues are written in red font. (**B**) Superposition of structures of the HOT1 DBD and TRF1 DBD bound to telomeric DNA. Both binding domains recognize a different set of DNA bases, resulting in a different positioning relative to the 5\u2032-TTAGGG-3\u2032 motif.](emboj2013105f3){#f3}\n\n![The degree of HOT1--telomere association varies between cell types. (**A**) Colocalization analysis of telomeres and HOT1 in HeLa cells by immunoFISH staining. A representative image illustrating the colocalization between several HOT1 foci (green) and telomeres (red) is shown. DAPI (blue) is used as nuclear counterstain. Colocalization events are indicated by arrows. Scale bars represent 5 \u03bcm. The quantification of the frequency of colocalization events was done after a 3D reconstruction of the acquired *Z*-stacks (*n*=147). The average value is indicated by a red bar. (**B**) Colocalization analysis of TRF1 and HOT1 in mouse ES cells by IF staining. To visualize TRF1 a LAP cell line ([@b55]) was used, expressing GFP-tagged TRF1 at endogenous expression levels. A representative image illustrating the colocalization between several HOT1 foci (green) and TRF1 (red) is shown. DAPI (blue) is used as a nuclear counterstain. Colocalization events are indicated by arrows. Scale bars represent 5 \u03bcm. The quantification of the frequency of colocalization events was done after a 3D reconstruction of the acquired *Z*-stacks (*n*=126). The average value is indicated by a red bar. (**C**) IF stainings of HOT1 at chromosome ends of mouse pachytene chromosome spreads. Representative images illustrating the localization of HOT1 and TRF2 (in green) to chromosome ends are shown. The synaptonemal complex/chromosome axis is marked by SYCP3 (red). The same field of view for the DNA counterstained by DAPI (greyscale) is shown in the bottom right corners. Scale bars represent 5 \u03bcm. The quantification of HOT1 foci at chromosome ends was done after a 3D reconstruction of the acquired *Z*-stacks (*n*=21). The average value is indicated by a red bar.](emboj2013105f4){#f4}\n\n![HOT1 associates with telomerase and CB complex components. (**A**) Summary of SILAC-based protein--protein interactions. Identification and normalized SILAC ratios are indicated for HOT1 (bait) and the identified interaction partner relevant for telomere biology from immunoprecipitation using both a rabbit and a mouse anti-HOT1 antibody. (**B**) Validation of the MS identifications by conventional immunoprecipitation. Nuclear HeLa extracts were subject to immunoprecipitation with either a polyclonal rabbit anti-HOT1 or an IgG antibody, and were immunoblotted for DKC1, Ku70 and Coilin. HOT1 IPs for the coprecipitation of DKC1 and Coilin were carried out in corresponding LAP cell lines ([@b55]) and both proteins were detected with anti-GFP antibody. FLAG--HOT1 was used to monitor the efficiency of the IP and a representative blot is shown. (**C**) Visualization of telomerase activity enrichment by gel electrophoresis in immunoprecipitations using antibodies against HOT1, DKC1 (positive control), TBP, YY1, STAT3, H3K4me3 and CENP-B (negative controls), as well as TRF1 and TRF2, using extracts from HeLa cells. All antibodies are rabbit polyclonal. A representative gel image of quantitative TRAP reaction products is shown. Samples were loaded on two gels and run in parallel represented by a gap between gel pictures. (**D**) Quantification of telomerase activity enrichment from the immunoprecipitations in panel **C**. Enrichments are normalized to immunoprecipitations using an IgG control. Error bars represent the s.d. of three independent experiments. Enrichments for DKC1, HOT1, TRF1 and TRF2 are statistically significant with *P*\\<0.05 (Student\\'s *t*-test). (**E**) Colocalization analysis of Coilin and HOT1 in HeLa cells by immunofluorescence staining. A representative image illustrating the colocalization between several HOT1 foci (green) and CBs (red; staining for Coilin) is shown. DAPI (blue) is used as nuclear counterstain. Colocalization events are indicated by arrows. An enhanced magnification of the boxed area is shown in the bottom right corners. Scale bars represent 5 \u03bcm. The quantification of the frequency of colocalization events was done after a 3D reconstruction of the acquired *Z*-stacks (*n*=179). The average value is indicated by a red bar.\\\nSource data for this figure is available on the online [supplementary information](#S1){ref-type=\"supplementary-material\"} page.](emboj2013105f5){#f5}\n\n![HOT1 regulates telomere length similar to the telomerase pathway member TCAB1. (**A**) Verification of HOT1 knockdown efficiency by western blot 48 h post transfection using the corresponding LAP cell line ([@b55]) as a reporter for protein expression. The TCAB1 knockdown was evaluated by quantitative PCR 24 h post transfection. (**B**) Quantification of telomere length by quantitative telomeric FISH after transient knockdown of HOT1 and TCAB1. The distributions of fluorescence intensities, in arbitrary units of fluorescence (a.u.f.), of individual telomeres from a total of 20 metaphases per treatment are displayed; the average intensity is indicated in red. For the gene-specific knockdowns, changes of average telomere signal intensity relative to the RLuc (Renilla Luciferase) control are shown (left). Representative FISH images are shown for each treatment and signal-free ends are indicated by arrows (right). Examples of individual chromosomes are magnified and the respective chromosomes are marked by rectangles (right). Scale bars represent 5 \u03bcm. (**C**) Summary of the quantification of signal-free ends per metaphase after gene-specific knockdown. (**D**) Quantification of telomere length by universal STELA after transient knockdown of HOT1 and TCAB1. The distributions of telomere length groups in kb of individual telomeres are displayed; the average length is indicated in red and averages are stated with the respective s.e.m. Changes of average telomere length are shown relative to the RLuc (Renilla Luciferase) control. (**E**) Raw data of STELA reactions for the quantification of telomere length. STELA products after gel electrophoresis, transfer and hybridization from 12 individual reactions (lanes) per treatment are shown. Samples were loaded on two gels and were run in parallel, represented by a gap within HOT1 esiRNA lanes. (**F**) Quantification of the frequency of short STELA products (\\<5 kb) relative to long STELA products (\\>5 kb). Five kilobases were used as a cut-off value based on this being the average telomere length as determined in the RLuc control.\\\nSource data for this figure is available on the online [supplementary information](#S1){ref-type=\"supplementary-material\"} page.](emboj2013105f6){#f6}\n\n![HOT1 acts as a positive regulator of telomere length. (**A**) Verification of HOT1 esiRNA knockdown efficiency using a HOT1-LAP cell line ([@b55]) as a reporter for HOT1 expression by western blot after 2, 3, 4 and 6 days. Cells analysed after 4 and 6 days were transfected twice: After the initial transfection, cells were transfected a second time on day 3 (72 h post transfection). All samples were run on the same gel, irrelevant lanes were spliced out. (**B**) Quantification of telomere length by quantitative telomeric FISH after transient knockdown of HOT1. The distributions of fluorescence intensities, in arbitrary units of fluorescence (a.u.f.), of individual telomeres from a total of 15--20 metaphases per treatment are displayed; the average intensity is indicated in red. Changes of average telomere signal intensity are shown relative to the RLuc (Renilla Luciferase) control. (**C**) Verification of FLAG--HOT1 and FLAG--HOT1\u0394Homeobox expression by western blot 48 h post transfection. (**D**) Results of telomere length measurements after transient overexpression of FLAG--HOT1 and FLAG--HOT1\u0394Homeobox. The distributions of fluorescence intensities, in a.u.f., of individual telomeres from a total of 30 metaphases per treatment are displayed; the average intensity is indicated in red. (**E**) Sequence-specific pull-down of FLAG--HOT1 and FLAG--HOT1\u0394Homeobox. Proteins were incubated with either telomeric repeats (5\u2032-TTAGGG-3\u2032) or a control oligonucleotide (5\u2032-GTGAGT-3\u2032).\\\nSource data for this figure is available on the online [supplementary information](#S1){ref-type=\"supplementary-material\"} page.](emboj2013105f7){#f7}\n\n![TERT binding to chromatin is dependent on HOT1. (**A**) Schematic of the genetrap insertion in the *Hot1* gene. (**B**) PCR genotyping. Genomic DNA from wild-type, heterozygous and homozygous mice were analysed by PCR. Genotypes were confirmed using three primers WT-F, WT-R and MUT-R. WT-F and WT-R amplify the wild-type allele (upper band), and WT-F and MUT-R amplify the genetrap allele (lower band). (**C**) Immunofluorescence stainings of HOT1 on wild-type and *Hot1*^Gt(pU-21T)346Card/(Gt(pU-21T)346Card^ MEFs (referred to as *Hot1*^\u2212/\u2212^ MEFs). HOT1 foci (green) are absent in *Hot1*^\u2212/\u2212^ MEFs. DAPI (blue) is used as a nuclear counterstain. Scale bars represent 5 \u03bcm. (**D**) Subcellular fractionation analysis of wild-type*, Hot1*^\u2212/\u2212^ and *Tert*^\u2212/\u2212^ MEFs. Tubulin, PCNA and Histone3 serve as loading controls and to monitor the cell fractionation.\\\nSource data for this figure is available on the online [supplementary information](#S1){ref-type=\"supplementary-material\"} page.](emboj2013105f8){#f8}\n\n###### Crystallographic statistics\n\n  -------------------------------------------------- --------------------------------\n  *Data collection*                                  \n  \u2003Data set                                                       Native\n  \u2003Beamline                                                      SLS PXII\n  \u2003Space group                                                 *C222*~*1*~\n  \u2003Unit cell parameters (\u00c5)                           *a*=111.4, *b*=116.5, *c*=75.7\n  \u2003Wavelength (\u00c5)                                                  1.00\n  \u2003Resolution range (\u00c5)[a](#tI-fn1){ref-type=\"fn\"}              46.2--2.9\n  \u2003Unique reflections                                             11 126\n  \u2003Multiplicity                                                 3.6 (3.6)\n  \u2003Completeness (%)[a](#tI-fn1){ref-type=\"fn\"}                 98.9 (97.4)\n  \u2003*I*/*\u03c3*(*I*)[a](#tI-fn1){ref-type=\"fn\"}                      18.8 (2.2)\n  \u2003*R*~merge~ (%)[a](#tI-fn1){ref-type=\"fn\"}                    3.3 (45.6)\n  \u00a0                                                  \n  *Refinement*                                       \n  \u2003Resolution range (\u00c5)                                         46.2--2.90\n  \u2003*R*~free~ (%)                                                   22.1\n  \u2003*R*~work~ (%)                                                   17.2\n  \u2003r.m.s.d. bond (\u00c5)                                              0.006\n  \u2003r.m.s.d. angle (deg)                                            1.0\n  \u2003B-factor 2 (\u00c5^2^)                                               79.5\n  \u00a0                                                  \n  *Ramachandran validation*                          \n  \u2003Favoured (%)                                                    99.3\n  \u2003Allowed (%)                                                     0.7\n  \u2003Outliers (%)                                                     0\n  -------------------------------------------------- --------------------------------\n\n![](emboj2013105t1)\n\n^a^Values in parentheses correspond to the highest resolution shell.\n\n###### Oligonucleotides used for pull-down experiments\n\n  No   Sequence motif (5\u2032\u21923\u2032)   Primer sequence (5\u2032\u21923\u2032)\n  ---- ------------------------ --------------------------------------------------------------\n  1a   TTAGGG for               TTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGG\n  1b   TTAGGG rev               AACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCT\n  2a   GTGAGT for               TTGACAGTGAGTGTGAGTGTGAGTGTGAGTGTGAGTGTGAGTGTGAGTGTGAGTGTGAGT\n  2b   GTGAGT rev               AAACTCACACTCACACTCACACTCACACTCACACTCACACTCACACTCACACTCACTGTC\n  3a   TTAGGC for               TTAGGCTTAGGCTTAGGCTTAGGCTTAGGCTTAGGCTTAGGCTTAGGCTTAGGCTTAGGC\n  3b   TTAGGC rev               AAGCCTAAGCCTAAGCCTAAGCCTAAGCCTAAGCCTAAGCCTAAGCCTAAGCCTAAGCCT\n  4a   TTGGGG for               TTGGGGTTGGGGTTGGGGTTGGGGTTGGGGTTGGGGTTGGGGTTGGGGTTGGGGTTGGGG\n  4b   TTGGGG rev               AACCCCAACCCCAACCCCAACCCCAACCCCAACCCCAACCCCAACCCCAACCCCAACCCC\n  5a   TCAGGG for               GTCAGGGTCAGGGTCAGGGTCAGGGTCAGGGTCAGGGTCAGGGTCAGGGTCAGGGTCAGG\n  5b   TCAGGG rev               ACCCTGACCCTGACCCTGACCCTGACCCTGACCCTGACCCTGACCCTGACCCTGACCCTG\n  6a   TGAGGG for               GTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGTGAGG\n  6b   TGAGGG rev               ACCCTCACCCTCACCCTCACCCTCACCCTCACCCTCACCCTCACCCTCACCCTCACCCTC\n\n![](emboj2013105t2)\n\n###### Oligonucleotides used for QuickChange\n\n  No   Cloning/BAC engineering purpose                 Primer sequence (5\u2032\u21923\u2032)\n  ---- ----------------------------------------------- ------------------------------------------\n  1    HOT1 homeobox domain deletion, forward primer   CTGCGACGAGGGAGTGCCAATATTGAAGCAGCAATCCTGG\n  2    HOT1 homeobox domain deletion, reverse primer   TGCTTCAATATTGGCACTCCCTCGTCGCAGTCGGAAAGTA\n\n![](emboj2013105t3)\n\n###### Oligonucleotides used for crystallization\n\n  No    Telomeric DNA oligos for crystallization  Primer sequence (5\u2032\u21923\u2032)\n  ---- ------------------------------------------ -------------------------\n  1                T1; forward oligo              CTGTTAGGGTTAGGGTTAG\n  2                T2; reverse oligo              TCTAACCCTAACCCTAACA\n\n![](emboj2013105t4)\n\n###### Oligonucleotides used for esiRNA production\n\n  No   Gene name                 ENSEMBL ID     Forward primer (5\u2032\u21923\u2032)   Reverse primer (5\u2032\u21923\u2032)\n  ---- ---------------------- ----------------- ------------------------ ------------------------\n  1    *Renilla Luciferase*           \u00a0         GATAACTGGTCCGCAGTGGT     CCATTCATCCCATGATTCAA\n  2    *HOT1* (1st esiRNA)     ENSG00000147421  ATTTCCCAAGCAGTTGTTGC     CCAGATCTGACAGCTTTTTGC\n  3    *HOT1* (2nd esiRNA)     ENSG00000147421  GTTACCTCC CTGAAAGTATA    GTTGGTTAT TATTTACTGGG\n  6    *TRF1*                  ENSG00000147601  TGGCAACTTTAAAGAAGCAGAA   AGCTTCAGTTTCCATTTCAACA\n  7    *TCAB1*                 ENSG00000141499  CGAATCGAGGAGCAAGAACT     GGGCTGAGGACATCAGAGAA\n\n![](emboj2013105t5)\n\nPlease note that esiRNA--template production consists of two consecutive PCRs. In the table only the transcript-specific primers for the first PCR round are listed. Universal tags have to be added to all primers that allow amplification with universal T7 primers in the second PCR round. The universal tags are 5\u2032-TGTAAAACGACGGCCAGT-3\u2032 (forward) and 5\u2032-AGGAAACAGCTATGACCAT-3\u2032 (reverse). The universal T7 primers are 5\u2032-GCTAATACGACTCACTATAGGGAGAG-3\u2032 (forward) and 5\u2032-GCTAATACGACTCACTATAGGGAGAC-3\u2032 (reverse).\n\n###### Oligonucleotides used for quantitative PCRs\n\n  No    Gene name     ENSEMBL ID     Forward primer (5\u2032\u21923\u2032)   Reverse primer (5\u2032\u21923\u2032)\n  ---- ----------- ----------------- ------------------------ ------------------------\n  1      *GAPDH*    ENSG00000111640  CAGCCTCAAGATCATCAGCA     TGTGGTCATGAGTCCTTCCA\n  2      *TCAB1*    ENSG00000141499  AGCCAGACACCTCCTACGTG     GGTTGAAGCCACAGAAGAGC\n\n![](emboj2013105t6)\n\n###### Oligonucleotides used for cloning and BAC engineering\n\n  No   Cloning/BAC engineering purpose                                            Primer sequence (5\u2032\u21923\u2032)\n  ---- -------------------------------------------------------------------------- -------------------------------------------------------------------------\n  1a   HOT1 BAC stitching primers to add missing CDS and 3\u2032-UTR, forward primer   CGATTTACCTGGAGAAAGGAGTGCCTGGCTGTTATGGAAAGT-TACTTCAATGAGAATCAATACCCAGATG\n  1b   HOT1 BAC stitching primers to add missing CDS and 3\u2032-UTR, reverse primer   TATAATACAGCATTTATGATATTCTAAAGTACTTTTAGAGAT-AGAACCAACCCTGTGCTGCTACATTGAA\n\n![](emboj2013105t7)\n\n###### Oligonucleotides used for genotyping PCRs\n\n  No    Primer name  Reverse primer (5\u2032\u21923\u2032)\n  ---- ------------- ------------------------\n  1        WT-F      CCTCCTGCTCTGTCTCCTTG\n  2        WT-R      CCCATTCAATCCCATCAAGG\n  3        MUT-R     CCCTTCAGTCTTCCTGTCCA\n\n![](emboj2013105t8)\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Specifications TableSubject areaComputer scienceMore specific subject areaComputer Vision and Image ProcessingType of dataImage, graph, video and spreadsheetHow data was acquiredMicrosoft Kinect RGB-D (Infrared and RGB camera) matricesData formateps, pdf, avi, xlsExperimental factorsData from the infrared and RGB sensor were converted into appropriate matrices with RGB channels.Experimental featuresImages produced for Detection and Tracking of Dynamic Objects and figures that detail the applied methodology in order to calculate effectiveness.Data source locationIndoor environment (SP-Brazil)Data accessibilityImage, video and spreadsheet (Supplementary files)Related research articleNatal Henrique Cordeiro, Emerson Carlos Pedrino,\\\nA new methodology applied to dynamic object detection and tracking systems for visually impaired people, Computers & Electrical Engineering, Volume 77, 2019, Pages 61--71,\\\nISSN 0045--7906, <https://doi.org/10.1016/j.compeleceng.2019.05.003>.**Value of the data**\u2022The data enable visually identifying which detections were effectively mapped using dynamic object detection and tracking techniques.\u2022The data allows (researcher) the chance of measuring the and effectiveness of each technique used for detection and tracking of dynamic objects.\u2022The data allow to understand how the detection of a dynamic object and error identification is performed by each technique.\u2022These data demonstrate how the positions of the detected dynamic objects can be mapped in millimeters.\u2022The quality of segmentation performed by each technique can be understood visually.\n\n1. Data {#sec1}\n=======\n\nThese data present a methodology applied to compare four techniques for the detection and tracking of dynamic objects (DTDO) regarding their effectiveness. The implemented techniques were customized from operational architectures. [Fig.\u00a01](#fig1){ref-type=\"fig\"} describes the methodology applied to the analysis of the effectiveness of each technique. The techniques used in this context were: 1- Pattern Recognition (PR); 2- Farneback Optical Flow (FOF); 3- Background Subtraction (BS); and 4- Continuously Adaptive Meanshift (CamShift). [Fig.\u00a02](#fig2){ref-type=\"fig\"} depicts two examples of the DTDO method using the PR technique. [Fig.\u00a03](#fig3){ref-type=\"fig\"} shows part of the detection from the processing of the dynamic objects implemented by the FOF technique. [Algorithm 1](#enun1){ref-type=\"statement\"} outlines the adaptation that has been made to detect dynamic objects by using FOF even when movements of the acquisition sensor take place. [Fig.\u00a04](#fig4){ref-type=\"fig\"} depicts two examples of the DTDO method by using the FOF technique. [Fig.\u00a05](#fig5){ref-type=\"fig\"} illustrates two examples of DTDO by using the BS technique. [Fig.\u00a06](#fig6){ref-type=\"fig\"} displays two examples of DTDO by using the CamShift technique. [Fig.\u00a07](#fig7){ref-type=\"fig\"} depicts a graph containing the effectiveness of each technique. Details of the implementation of these four architectures can be seen in the main article previously published. (A new methodology applied to dynamic object detection and tracking systems for visually impaired people [@bib1]).Fig.\u00a01Measuring the effectiveness using MAP (X,Z).Fig.\u00a01Fig.\u00a02Detection and Tracking of DO using the Architecture A1-PR.Fig.\u00a02Fig.\u00a03Dense map (Farneback).Fig.\u00a03Fig.\u00a04Detection and Tracking of dynamic object using the Architecture A2-FOF.Fig.\u00a04Fig.\u00a05Detection and Tracking of DO using the Architecture A3-BS.Fig.\u00a05Fig.\u00a06Detection and Tracking of dynamic object using the Architecture A4-CamShift.Fig.\u00a06Fig.\u00a07Dynamic object detected and mapped correctly.Fig.\u00a07\n\n1.1. Data acquisition for detection and tracking of dynamic objects {#sec1.1}\n-------------------------------------------------------------------\n\nThe data presented in this article show that the architectures acquire the necessary data for the 3D tracking and mapping of the dynamic objects. The methodology applied in the detection process to account for effectiveness is also given. All modules present in the architectures were implemented with the help of the OpenCV library [@bib2]. In this project, the Microsoft Kinect RGB-D was placed at the front of the waist of the VIP. When a dynamic object is detected, regardless of speed, a beep will be emitted. However, tracking may not be to the expected quality due to the frame capture rate and processing speed is low relative to that of the moving object.\n\nHence the importance of analyzing the speed of detection (SpD) of each architecture projected. SpD allows observing which architecture produces the longest run time to trigger a beep ([see supplementary file - spreadsheet](#appsec1){ref-type=\"sec\"}). Each beep indicates the position of the dynamic object detected. This comparison is independent on the processing power used. Effectiveness is another and more important feature the data (Maps) of this study allows calculating. The analysis of which detections were performed improperly is conducted based on effectiveness. This occurs when noises are generated by changes in lighting, strong movement in the acquisition sensor, or even when objects similar to those programmed for the techniques of pattern (PR) or color recognition (CamShift) are detected in the environment. Two indoor environments with different types of lighting, objects and passages were defined. For each environment, detection and tracking of the dynamic object were performed using the four architectures. For each architecture, six paths were reconstructed in each environment. Of these reconstructions, three had a static camera and the other three had camera movements. Thus, forty eight paths were obtained to perform a comparative analysis between the architectures implemented. Based on the reconstruction of the paths, the quality of detection and tracking of the dynamic object could be analyzed the efficiency of four approaches.\n\n2. Experimental design, materials and methods {#sec2}\n=============================================\n\nIn [Fig.\u00a01](#fig1){ref-type=\"fig\"}, the data (images) show the methodology applied to compare the four architectures with respect to the effectiveness. In these images (Map (X, Z)) the amount of detection (blank points) performed until the end of the tracking is shown. The green circles show the tracking of the dynamic object until its completion. As the system records the start and end times, the number of detections carried out per second can be calculated at the end of each tracking, for each architecture. This map also allows for error identification during the detection process (Effectiveness). In the images in [Fig.\u00a01](#fig1){ref-type=\"fig\"}, the yellow circles represent errors in the detection process. The images (a, b, c, d) show the following scans: Architecture 1 - Pattern Recognition (A1-PR), Architecture 2 - Farneback Optical Flow (A2-FOF), Architecture 3 - Background Subtraction (A3-BS) and Architecture 4 - Continuously Adaptive Meanshift (A4-CamShift) respectively. The spreadsheet containing the raw data obtained to compare the effectiveness of each architecture implemented can be seen in [Supplementary files (spreadsheet 1)](#appsec1){ref-type=\"sec\"}. Each point rendered on the map (X, Z) means that a beep was sounded to the VIP to show the position of the dynamic object in the environment.\n\nIn the images in [Fig.\u00a01](#fig1){ref-type=\"fig\"}, observing that experiments were carried out in the same region is possible. Even so, the points (blank points) allow observing that each architecture shows differences in the process of detection and tracking of the dynamic object.\n\nThe details of the experiments are divided into four topics. Each topic provides the input images, the segmented objects and the map (X, Z) used to compare effectiveness.\n\n2.1. Architecture 1 - pattern recognition (A1-PR) {#sec2.1}\n-------------------------------------------------\n\n[Fig.\u00a02 (images (a -- f))](#fig2){ref-type=\"fig\"} show two instances of data acquisition for detection and tracking of dynamic object using Architecture 1 - Pattern Recognition. In this architecture, the detection was performed using the Haar Feature technique based on Cascade classifiers [@bib3]. These acquisitions occur with the camera suffering little movement (see supplementary video and [tab A1-PR in spreadsheet 1](#appsec1){ref-type=\"sec\"}). Each sequence is composed of the following images: dynamic object detection; the depth map that only shows objects at less than 2\u00a0m; the position of the dynamic object on the map (X, Z) at the same instant as its detection.\n\nThe following are the supplementary data related to this article:Video 1Video 1Video 2Video 2\n\n2.2. Architecture 2 - Farneback Optical Flow (A2-FOF) {#sec2.2}\n-----------------------------------------------------\n\nThe Farneback Optical Flow method was the only one adapted regarding the way of detecting dynamic objects. These adaptations were made to detect the dynamic object even faced with movements in the acquisition sensor (IR and RGB sensors). The first step of this technique is to distribute a set of points that will determine the optical flow between subsequent frames. These can be seen in [Fig.\u00a03](#fig3){ref-type=\"fig\"} and are created in [Algorithm 1 (lines 3 and 4)](#enun1){ref-type=\"statement\"}.\n\nFor the FOF technique to detect the dynamic object, vectors were created for each point in the image (line 7). These vectors represent the velocity and direction respectively by their size and their angle (see [supplementary files -- image 5](#appsec1){ref-type=\"sec\"}). The larger the vector, the greater the flow detected at a certain point on the map. The size and angles of the vectors allow us to differentiate which are movements of the acquisition sensor and which are movements of a dynamic object. Because all vectors reproduce changes when the camera is moved, the vector sizes were averaged using Euclidean distances (line 8). In the same way, the mean (line 9) of the angles produced with these movements was created. The angles were generated by the function **atan2** and the parameters are the same as those already given. Thus, for each point, a check was made to see if there was a change of flow and direction above the average.\n\nWhen the acquisition sensor undergoes some movement or when a dynamic object is detected, the vectors have their size modified. By analyzing the mean it is possible to verify the points that have suffered major flows. Likewise, when there is camera movement, the angles are similar. However, if any angle has a high divergence with respect to the mean, the probability of a dynamic object being detected is large (lines 13 and 14). For this method, besides the averages, the concepts of covariance (line 15) were also implemented for the size (**Cov**~**Euclidian**~\u00a0\\>\u00a0**value**~**1**~) and the\u00a0direction (**Cov**~**Angle**~\u00a0\\>\u00a0**value**~**2**~) of the vectors produced, where values (**value**~**1**~ and **value**~**2**~) are the thresholds found (average) and differ when the flow detected at one point is more than the others.Algorithm 1Adaptation to detect dynamic object using FOF even with movements of the acquisition sensor).Image 1\n\n[Fig.\u00a04 (images (a -- f))](#fig4){ref-type=\"fig\"} show two instances of data acquisition for Detection and Tracking of dynamic object using Architecture 2 - Farneback Optical Flow. These acquisitions occur with the camera suffering little movement (see supplementary video and [tab A2-FOF in spreadsheet 1](#appsec1){ref-type=\"sec\"}). Each sequence is composed of the following images: dynamic object detection; the depth map that provides the distances; the position of the dynamic object on the map (X, Z) at the same instant as its detection.\n\n2.3. Architecture 3 - background subtraction (A3-BS) {#sec2.3}\n----------------------------------------------------\n\n[Fig.\u00a05 (images (a -- f))](#fig5){ref-type=\"fig\"} show two instances of data acquisition for detection and tracking of dynamic object using Architecture 3 - Background Subtraction. These acquisitions occur with the camera suffering little movement (see supplementary video and [tab A3-BS in spreadsheet 1](#appsec1){ref-type=\"sec\"}). Each sequence is composed of the following images: dynamic object detection using BS; the depth map that provides the distances; the position of the dynamic object on the map (X, Z) at the same instant as its detection.\n\n2.4. Architecture 4 - Continuously Adaptive Meanshift (A4-CamShift) {#sec2.4}\n-------------------------------------------------------------------\n\n[Fig.\u00a06 (images (a -- f))](#fig6){ref-type=\"fig\"} show two instances of data acquisition for detection and tracking of dynamic objects using Architecture 4 - Continuously Adaptive Meanshift. This architecture performs single hue object tracking. The acquisitions occur with the camera suffering little movement (see supplementary video and [tab A4-CamShift in spreadsheet 1](#appsec1){ref-type=\"sec\"}).\n\nEach sequence is composed of the following images: dynamic object detection; the outline of the image (a); the position of the dynamic object on the map (X, Z) at the same instant as its detection.\n\nWith the set of paths reconstructed, a comparative analysis could be done between some characteristics of the four detection and tracking architectures being tested. From all the data related to the effectiveness for each architecture was generated. The videos of the experiments conducted, the images of the maps (X,Z), and the data for analysis of effectiveness are available in [Supplementary files (tab effectiveness in spreadsheet 1](#appsec1){ref-type=\"sec\"}). [Fig.\u00a07](#fig7){ref-type=\"fig\"} present the averages obtained (effectiveness) for each architecture.\n\nConflict of interest {#appesc1}\n====================\n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n\nAppendix A. Supplementary data {#appsec1}\n==============================\n\nThe following is the Supplementary data to this article:Multimedia component 1Multimedia component 1\n\nWe are grateful to the Brazilian funding agency FAPESP (Project No. 2017/26421-3). Our thanks go as well to UFSCar-DC, and IFSP, SP, Brazil. This study was financed in part by the Coordena\u00e7\u00e3o de Aperfei\u00e7oamento Pessoal de N\u00edvel Superior - Brasil (CAPES) - Finance Code 001.\n\nSupplementary data to this article can be found online at <https://doi.org/10.1016/j.dib.2019.104403>.\n"}
+{"text": "Traditionally, the assessment of metabolic acidosis and alkalosis relies on measurement of the base excess, which is the difference between the \\'ideal\\' buffer base \\[[@B1]\\] (i.e. the sum of the negatively charged forms of weak acids, \\[A^-^\\] + \\[HCO~3~^-^\\] + \\[H~2~PO~4~^-^\\], at standard conditions (pH 7.4, temperature 37\u00b0C, partial carbon dioxide tension 40 mmHg) and the \\'actual\\' buffer base \\[[@B2]\\]):\n\nBase excess = buffer base~actual~- buffer base~ideal~\u00a0\u00a0\u00a0 (1)\n\nDuring the past few years a novel approach based on assessment of the strong ion difference (SID) has been introduced to evaluate metabolic acidosis and alkalosis. For simplicity, we limit our discussion to these two disturbances.\n\nPlease note that in the following discussion we will refer to the amount of strong ion difference as SID (mEq), while we will refer to the strong ion difference concentration as \\[SID\\] (mEq/l).\n\nBy definition, strong ions are always dissociated in a solution. In plasma, as well as in interstitial fluids, the sum of positively charged ions (primarily Na^+^, K^+^, Ca^2+^and Mg^2+^) exceeds the sum of the negatively charged strong ions (primarily Cl^-^and lactate^-^) of about 42 mEq/l. This difference is called the SID, and according to the Stewart model \\[[@B3]\\] its variation is one of the determinants of acid--base status. Looking at Figure [1](#F1){ref-type=\"fig\"}, the connection between base excess and SID is apparent. The buffer base and SID are equivalent. In fact, because the ideal SID is equal to 42 mEq/l (as is the normal buffer base), it follows that\n\nBase excess = SID~actual~- SID~ideal~= buffer base~actual~- buffer base~ideal~\u00a0\u00a0\u00a0 (2)\n\nBecause computation of the actual SID is rather complicated, requiring the determination of all of the strong ion concentrations, we believe that the base excess approach may be easier, more rapid and adequate for clinical purposes. Indeed, the frequent debate involving the comparison of the \\'SID approach\\' with the \\'base excess approach\\' to assessment of metabolic acidosis \\[[@B4],[@B5]\\] appears futile because their physiological meanings, as well as their variations, are identical. In other words, the two approaches look at the same thing from different points of view.\n\nThe picture is different when one considers the \\'understanding\\' of acid--base and electrolyte equilibria, which everyone has studied in separate chapters of the textbooks. The great merit of the Stewart approach is that it considers electrolytes and acid--base status in a common framework. Here, we would like to propose a comprehensive model that may explain, at least qualitatively, many of the findings observed in clinical practice and in the literature.\n\nThe SID reflects the difference in electrical charges of the strong ions in the volume of the extracellular compartment (V). At time 0, it will be equal to V(0) \u00d7 \\[SID(0)\\]. For example, if at time 0 the SID is normal (i.e. 42 mEq/l) then the net amount of electrical charge in the extracellular fluid (15 l) will be 630 mEq. During a given period of time there may be an addition of volume to the system (e.g. infusion of a solution) with its own SID (SID~infusion~). Consequently, a net amount of charge equal to V~infusion~\u00d7 \\[SID~infusion~\\] will be added to the system. Similarly, the urinary system will excrete a volume of urine (V~urine~) with its own SID (SID~urine~). The last variable that must be taken into account is endogenous production of SID (sulphates, phosphates, lactate and ketoacids, among other components). It follows that the SID at a given time \\'t\\' may be derived from a series of equations, which may appear to be complicated in their expression but simple in their meaning. Eqn 3 (below) indicates that, in a system, the net amount of electrical charges due to the strong ions is equal to the net electrical charge of the system at time zero plus the net electrical charge added as a result of metabolism plus the net electrical charge added with volume infusion minus the net electrical charge extracted via urine.\n\n![](cc4890-i1.gif)\n\nwhere EPR(t) is the \\'endogenous production rate\\' of SID (mEq/min), IR(t) is the volume infusion rate and UPR(t) is the urine production rate. At a given time \\'t\\', the net fluid volume of the extracellular compartment is equal to the initial volume of the system plus the volume added with infusion minus the volume extracted in the form of urine.\n\n![](cc4890-i2.gif)\n\nBecause what matters in terms of acid--base status is the concentration, rather than the net amount of electrical charge, the SID at a given time \\'t\\' may be expressed from the above equations as shown in equation 5 at the foot of the page:\n\n![](cc4890-i3.gif)\n\nIt is important to remember that an increase in SID will lead the system to become more basic whereas a decrease in SID will lead the system to become more acidic. In general, Eqn 5 indicates that metabolic acidosis or alkalosis may occur either by changing the net electrical charge at constant extracellular volume or by changing the extracellular volume at constant electrical charge.\n\nLooking at Eqn 5, we may make several comments. To maintain the metabolic acid--base status of a system (i.e. the baseline SID), two conditions must be satisfied: the input quantity of SID should equal the output quantity of SID; and the distribution volume of SID should remain constant. To the best of our knowledge, the only studies in which the strong ion balance (input and output) was investigated were conducted in cows \\[[@B6]-[@B8]\\]; different amounts of SID in the diet caused corresponding changes in urinary SID. Unfortunately, no such investigation has been conducted in critically ill patients. As discussed above, SID has been studied in comparison with base excess but without any physiological rationale \\[[@B9]\\]. The SID approach has been also proposed to explain metabolic acidosis during saline infusion (SID input) \\[[@B10]\\], but only a few papers have tackled and discussed the problem of urinary SID (SID output) \\[[@B11]-[@B13]\\]. What we lack is the entire picture of the system; unfortunately, this requires frequent assessment of urine electrolytes.\n\nSome clinical findings may be viewed from the perspective of the general framework of Eqn 5. It is well known that rapid infusion of saline induces metabolic acidosis. This has been attributed to changes in SID due to hyperchloraemia \\[[@B10]\\]. By looking at Eqn 3 we derive a different point of view. Because the SID of saline is equal to 0, it follows that, if the urinary output of electrical charge and metabolic production remain constant, the net difference of electrical charges in the system (i.e. the numerator in Eqn 5) does not change. What causes the acidosis is the expansion of the extracellular volume (volume input greater than volume output), which leads to decreased concentration of the net amount of electrical charge (i.e. the SID).\n\nUnfortunately, it is not easy to consider the urinary SID. In fact, although 40--42 mEq/l of plasmatic negative charge may be derived from the dissociated weak acids (\\[A^-^\\], \\[HCO~3~^-^\\] and \\[H~2~PO~4~^-^\\]), the amount of weak acids is far less in urine and, overall, the range of urinary pH is an order of magnitude greater than that in plasma. Once again, the problem is simpler when one considers the entire picture. In fact, as far as the plasmatic acid--base equilibrium is concerned, we must consider only the components of urinary \\[SID\\] that may affect the plasmatic \\[SID\\] (i.e. \\[K^+^\\], \\[Na^+^\\] and \\[Cl^-^\\]). In fact, in urine\n\n\\[Na^+^\\] + \\[K^+^\\] + \\[Un^+^\\] = \\[Cl^-^\\] + \\[Un^-^\\] \u00a0\u00a0\u00a0 (6)\n\nwhere Un^+^and Un^-^are the positive and negative unmeasured ions. It follows that\n\n\\[Na^+^\\] + \\[K^+^\\] - \\[Cl^-^\\] = \\[Un^-^\\] - \\[Un^+^\\] \u00a0\u00a0\u00a0 (7)\n\nQuantitatively, the most important anion in urine is SO~4~^2-^, which is derived from the metabolism of sulphur amino acids, whereas the most important cation is NH~4~^+^. In normal conditions, the sum of urinary \\[Na^+^\\] + \\[K^+^\\] - \\[Cl^-^\\] amounts to 42 mEq/l \\[[@B14]\\]. It follows that the concentration of unmeasured anions exceeds the concentration of unmeasured cations of 42 mEq/l. When a strong ion such as lactate is added to the plasma, the plasmatic SID will decrease. Consequently, the urinary system will react by increasing its excretion of chloride, thereby decreasing the plasma chloride concentration (while \\[Na^+^\\] and \\[K^+^\\] must be maintained within normal ranges). The increased excretion of chloride will decrease the urinary SID. Therefore, the difference between \\[Un^-^\\] and \\[Un^+^\\] should decrease (Eqn 7). This is accomplished by increasing the excretion rate of NH~4~^+^, which is a way to augment elimination of Cl^-^without Na^+^\\[[@B11],[@B15]\\].\n\nIndeed, the effects of any volume infusion or other interventions cannot be understood if the urinary SID and volume are not taken into account. A merit of the report by Moviat and colleagues \\[[@B13]\\] is that, for the first time in critical care, attention is focused on the urinary system, which is the main regulator of SID. The authors found that the increase in urinary SID (indirectly induced by blocking carbonic anhydrase) was the key driver for correction of metabolic alkalosis. The message is important -- urinary SID should be a key component of global acid--base assessment. We believe that urinary electrolyte monitoring may open a new perspective of research in critical care. Acid--base equilibrium, one of the oldest research areas in medicine, is still an open field for new discoveries and approaches.\n\nAbbreviations\n=============\n\nSID = strong ion difference.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nFigures and Tables\n==================\n\n![Gamblegram. The figure shows gamblegrams during ideal conditions and during acidosis. In ideal conditions the difference between positively and negatively charged strong ions is equal to 42 mEq/l (the strong ion difference \\[SID\\]) and, according to the law of electro-neutrality, is equivalent to the buffer base (BB; i.e. the sum of \\[HCO~3~^-^\\], \\[A^-^\\] and \\[H~2~PO~4~^-^\\], where A^-^are the weak acids in dissociated form, mainly albumin). During acidosis, SID decreases but the law of electro-neutrality is still satisfied. It follows that base excess = BB~actual~- BB~ideal~= SID~acidosis~- SID~ideal~.](cc4890-1){#F1}\n"}
+{"text": "Introduction {#Sec1}\n============\n\nParkinson's disease (PD) is clinically diagnosed by the presence of four cardinal motor symptoms: bradykinesia, tremor, rigidity and postural instability. The motor symptoms are caused by the loss of dopaminergic neurons in the substantia nigra pars compacta^[@CR1]^. However, non-motor symptoms can precede motor symptoms by more than a decade^[@CR2]^ and include olfactory dysfunction, rapid eye movement sleep behaviour disorder and autonomic disorders such as constipation. The non-motor symptoms and motor symptoms appear to coincide with the aggregation of the pathological protein \u03b1-synuclein (\u03b1-syn) in specific brain regions^[@CR2]--[@CR4]^. Phosphorylated \u03b1-syn pathology affects different brain regions in a sequential pattern that has been characterised into six stages^[@CR5]^. In stage I, the dorsal motor nucleus of the vagus nerve and the olfactory system -- olfactory mucosa, olfactory bulb (OFB) and regions of the anterior olfactory nucleus (AON) present \u03b1-syn pathology. As the disease progresses into stage II, \u03b1-syn pathology is evident in the brainstem where it reaches the substantia nigra in stage III, coinciding with the clinical motor symptoms associated with PD^[@CR5]^. As the disease continues through stages IV -- VI, \u03b1-syn begins to affect cortical regions.\n\nOne of the most common non-motor symptoms that is present in more than 90% of those suffering with PD is the loss of olfaction (hyposmia or anosmia depending on severity)^[@CR6]^. Prior to the changes seen in the substantia nigra, \u03b1-syn pathology is present throughout the OFB and tract and is especially abundant in each compartment of the AON^[@CR5],[@CR7],[@CR8]^. The AON is the neural conduit between the OFB and the piriform cortex, entorhinal cortex, amygdala and hippocampal formations^[@CR9],[@CR10]^. The olfactory system consists of olfactory mucosa in the roof of the nose that houses the olfactory sensory neurons whose axons traverse the cribriform plate and form\u00a0the olfactory nerve layer of the OFB. Evidence from studies of the rodent OFB structure suggests that glomeruli are formed from olfactory sensory neuron fibres and then project to mitral cells and mitral cells synapse in the AON. Thus, the olfactory sensory neurons, which are exposed to the mucus and external environment, are just one synapse between the external environment and the AON (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"})^[@CR11],[@CR12]^. In a theory called the olfactory vector hypothesis, it is proposed that the close connectional proximity of the external environment to the central brain structures allows inhaled environmental toxins and pathogens to enter the brain via the AON^[@CR6],[@CR11],[@CR12]^.Figure 1(**a**) Immunofluorescence image of a sagittal section from a human PD OFB stained with antibodies for PGP9.5, CNPase, NeuN, UEA-1, and a Hoechst counterstain to identify the AON regions. Phosphorylated \u03b1-syn (s129) antibody was used for the distribution of \u03b1-syn. (**b**) Illustration of the PD OFB and tract displaying the distribution of \u03b1-syn, highlighting that most of the \u03b1-syn inclusions are in the multiple AON regions. (**c**) Inset of AON 1 showing representative immunofluorescence images used to identify the AON region. The AON is characterized by a decreased density of cells represented with Hoechst staining, increased PGP9.5 immunoreactivity, a lack of CNPase immunoreactivity and increased density of large NeuN-positive cells. In PD OFBs, the increased abundance of \u03b1-syn can be used to identify the AON. Scale bar, 1\u2009mm (**a**) and 500\u2009\u00b5m (**c**). Graphs showing (**d**) area fraction of \u03b1-syn (%) inside versus outside of the AON of human PD OFBs (n\u2009=\u200911) and (**e**) area fraction of \u03b1-syn (%) inside versus outside of the AON by case. Data presented as mean\u2009\u00b1\u2009SD. *\\# p*\u2009\\<\u2009*0.0001, \\* p*\u2009\\<\u2009*0.05*.\n\nThe spread of \u03b1-syn from the OFB to deeper brain regions is evidenced by a previous study showing that injection of \u03b1-syn inclusions in the OFB of mice results in \u03b1-syn-positive inclusions developing in many interconnected brain regions over 12 months^[@CR13],[@CR14]^. Local OFB interneurons in the mouse and human AON are also particularly good at internalising \u03b1-syn^[@CR13],[@CR15],[@CR16]^. Indeed many *in vitro* and *in vivo* models have suggested that neuronal cells are the main facilitator for the spread of \u03b1-syn pathology following the major neuronal pathways in the brain^[@CR17]--[@CR19]^. These studies point toward the AON and OFB as being crucial structures for the spread of \u03b1-syn pathology. But what of other cell types that may spread, but may not aggregate \u03b1-syn?\n\nThe involvement of non-neuronal cells in the spread of \u03b1-syn pathology in the PD brain has been overlooked. Microglia and astrocytes are predominantly associated with inflammatory processes in the PD brain^[@CR20]--[@CR25]^. However, recent *in vitro* studies demonstrate that microglia and astrocytes efficiently take up and degrade \u03b1-syn from extracellular locations^[@CR26],[@CR27]^. In the human PD brain, the number of astrocytes and oligodendrocytes containing \u03b1-syn inclusions appear to correlate with the severity of nigral neuronal loss^[@CR28],[@CR29]^. Most recently, *in vitro* evidence suggests that pericytes, a blood-vessel associated cell involved in the maintenance of the blood brain barrier, together with astrocytes may be involved in the spread of \u03b1-syn from one cell to the next^[@CR30]--[@CR32]^.\n\nTaken together, current literature suggests that non-neuronal cells could play an active role in the progression of PD, but evidence of these cells containing \u03b1-syn in the human PD OFB is lacking. Here, we show that in the PD AON, \u03b1-syn is found within neurons, microglia, pericytes and astrocytes but not oligodendrocytes. Secondly, we observed that the \u03b1-syn structures in non-neuronal cells look similar to some of the \u03b1-syn inclusions seen in neuronal cells, suggesting that non-neuronal cells may play a more active role in the pathogenesis of PD than previously thought.\n\nResults {#Sec2}\n=======\n\nDistribution of phosphorylated \u03b1-synuclein in the human Parkinson's disease olfactory bulb {#Sec3}\n------------------------------------------------------------------------------------------\n\nPhosphorylated \u03b1-syn inclusions were present throughout the OFB and tract of the 11 PD cases used in this study (Fig.\u00a0[1a--c](#Fig1){ref-type=\"fig\"}). Of the 11 normal cases, only OFB55 contained phosphorylated \u03b1-syn in the glomerular layer of the OFB in small amounts. In the PD OFBs, phosphorylated \u03b1-syn-positive Lewy neurites and Lewy bodies were seen in the glomerular layer, external plexiform, mitral cell layer, internal plexiform layer, granule cell layers and in the multiple AON compartments. To confidently identify the AON regions across different sections and cases, we found that the co-labelling of Hoechst, NeuN, PGP9.5 and CNPase was sufficient. The AON has a reduced number of Hoechst positive cells, clusters of large NeuN positive neurons and increased PGP9.5 immunoreactivity. Regarding myelination, there is a lack of CNPase immunoreactivity in the AON but positive immunoreactivity labelling the myelinated fiber tracts in the acellular neuropil zone creating a definitive border around the AON. Lastly, in PD OFBs, increased abundance of phosphorylated \u03b1-syn staining can be seen (Fig.\u00a0[1a,c](#Fig1){ref-type=\"fig\"}).\n\nThe area fraction of \u03b1-syn was significantly more abundant in the AON regions covering 9.35%\u2009\u00b1\u20099.88% of the surface area compared with the area fraction of \u03b1-syn outside of the AON only covering 0.87%\u2009\u00b1\u20090.81% of the surface area in the 11 PD cases (p\u2009\\<\u20090.0001; Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). All 11 PD OFB cases had phosphorylated \u03b1-syn labelling in the AON, however, the amount of \u03b1-syn present varied greatly between cases (0.36--2.52% outside AON, 1.98--22.55% inside AON; Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}). As the majority of \u03b1-syn was found within the AON regions in the PD OFBs (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}), quantification of the cells that contained intracellular \u03b1-syn were confined to this region. None of the normal OFBs had phosphorylated \u03b1-syn in the AON, therefore, the number of cells containing \u03b1-syn inclusions were zero.Figure 2Representative confocal immunofluorescence images with orthogonal views of different cell types containing \u03b1-syn in the AON of human PD OFBs. Small intracellular \u03b1-syn inclusions were found in (**a**) microglia, (**b**) pericytes and (**c**) astrocytes. No intracellular \u03b1-syn was found in (**d**) oligodendrocytes. Neurons either contained (**e**) small \u03b1-syn inclusions or (**f**) large Lewy body like inclusions. Yellow arrows indicate \u03b1-syn inclusions that are intracellular. White arrows indicate \u03b1-syn inclusions that are extracellular. Scale bar, 10\u2009\u00b5m.\n\nQuantification of the different cell types containing phosphorylated \u03b1-synuclein inclusions in the human Parkinson's disease olfactory bulb {#Sec4}\n-------------------------------------------------------------------------------------------------------------------------------------------\n\n\u03b1-syn inclusions were present in microglia, pericytes and astrocytes but not in oligodendrocytes in the AON of the human PD OFB (Fig.\u00a0[2a--d](#Fig2){ref-type=\"fig\"}). We observed that these non-neuronal cells only contained small \u03b1-syn inclusions. However, neurons contained either small \u03b1-syn inclusions that resemble those seen in the non-neuronal cells (Fig.\u00a0[2e](#Fig2){ref-type=\"fig\"}) or the classical large Lewy body-like inclusions (Fig.\u00a0[2f](#Fig2){ref-type=\"fig\"}).\n\nIn this study, we used a two-step process to quantify cells containing \u03b1-syn. First, widefield tiled fluorescent images were used to identify cells that potentially contained intracellular \u03b1-syn. Cumulatively in the 11 PD cases, 783 microglia, 458 pericytes, 529 astrocytes and 80 oligodendrocytes were counted. In six PD cases, 495 neurons were identified that potentially contained \u03b1-syn. The presence of intracellular \u03b1-syn was difficult to quantify using simple co-localisation techniques. \u03b1-syn inclusions can appear in close proximity to cells of interest and co-localise with the specific cellular markers, however, it can be unclear whether the \u03b1-syn inclusions were intracellular, above or below the cell of interest (Supplementary Fig.\u00a0[1a,b](#MOESM1){ref-type=\"media\"}). To overcome this issue, we carried out a second quantification step by imaging each cell that potentially contained intracellular \u03b1-syn using confocal microscopy (Supplementary Fig.\u00a0[1a,a'](#MOESM1){ref-type=\"media\"}). Therefore, each cell that was identified to potentially contain \u03b1-syn was reimaged to confirm that the \u03b1-syn was intracellular (Supplementary Fig.\u00a0[1b,b'](#MOESM1){ref-type=\"media\"}). Following this, it was confirmed that 261 microglia, 89 pericytes, 119 astrocytes, zero oligodendrocytes and 182 neurons contained intracellular \u03b1-syn. Without the use of confocal microscopy, we would have overestimated the number of cells that contained intracellular \u03b1-syn by 3.5 times.\n\nThe total number of cells in each cell population was also counted in the AON regions of the OFB and tract to quantify the percentage of cells affected by \u03b1-syn. Cumulatively in the 11 PD cases, a total of 5288 microglia, 6925 astrocytes, 4419 pericytes,1621 oligodendrocytes and 2387 neurons were counted. Of the non-neuronal cell types, on average, 7.78%\u2009\u00b1\u20095.22% of microglia, 3.41\u2009\u00b1\u20092.74% of pericytes and 1.97\u2009\u00b1\u20091.17% of astrocytes contained intracellular \u03b1-syn (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). A total of 12.56\u2009\u00b1\u20097.98% of non-neuronal cells were affected by \u03b1-syn. On average, there were 2.25\u2009\u00b1\u20092.41% of neurons that contained small \u03b1-syn inclusions, 6.38\u2009\u00b1\u20096.83% of neurons that contained Lewy body like inclusions and 8.68\u2009\u00b1\u20099.34% of the total neuronal population contained intracellular \u03b1-syn inclusions (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}). There was a moderate positive correlation between the percentage of total cells containing intracellular \u03b1-syn in the AON and the area fraction of \u03b1-syn in the AON (r^2^\u2009=\u20090.55; p\u2009=\u20090.08).Figure 3Graphs showing (**a**) Percentage of cells with intracellular \u03b1-syn in the AON of PD OFBs (n\u2009=\u200911), where, 7.78%\u2009\u00b1\u20095.22% of microglia, 3.41%\u2009\u00b1\u20092.74% of pericytes, 1.97%\u2009\u00b1\u20091.17% of astrocytes and 0% of oligodendrocytes contained intracellular \u03b1-syn. Whereas, 2.25%\u2009\u00b1\u20092.41% of neurons contained small \u03b1-syn inclusions, 6.38%\u2009\u00b1\u20096.83% of neurons contained Lewy bodies, giving a total of 8.68%\u2009\u00b1\u20099.34% \u03b1-syn containing neurons overall. (**b**) Number of cells with intracellular \u03b1-syn in the AON/mm^2^ (n\u2009=\u200911), microglia (14.02\u2009\u00b1\u200910.23 cells/mm^2^), astrocytes (5.09\u2009\u00b1\u20093.55 cells/mm^2^), pericytes (4.93\u2009\u00b1\u20093.60 cells/mm^2^) and oligodendrocytes (0 cells /mm^2^). Neurons either contained small \u03b1-syn inclusions (8.84\u2009\u00b1\u20098.56 cells/mm^2^) or Lewy bodies (24.10\u2009\u00b1\u200925.25 cells/mm^2^) comprising a total number of neurons (32.49\u2009\u00b1\u200933.69 cells/mm^2^) containing intracellular \u03b1-syn. Data presented as mean\u2009\u00b1\u2009SD.\n\nAdditionally, the number of cells containing intracellular \u03b1-syn were normalised to the surface area of the AON. Of the non-neuronal cell types, on average, there were 14.02\u2009\u00b1\u200910.23 microglia/mm^2^, 4.93\u2009\u00b1\u20093.60 pericytes/mm^2^, and 5.09\u2009\u00b1\u20093.55 astrocytes/mm^2^ that contained intracellular \u03b1-syn inclusions (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}). A total of 22.43\u2009\u00b1\u200914.44 non-neuronal cells/mm^2^ contained intracellular \u03b1-syn. On average, there were 8.84\u2009\u00b1\u20098.56 neurons with small \u03b1-syn/mm^2^, 24.10\u2009\u00b1\u200925.25 neurons with Lewy bodies/mm^2^ and a total of 32.49\u2009\u00b1\u200933.69 neurons/mm^2^ containing intracellular \u03b1-syn inclusions (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}). There was a moderate positive correlation between the total number of cells/mm^2^ in the AON and the area fraction of \u03b1-syn in the AON (r^2^\u2009=\u20090.64; p\u2009=\u20090.05).\n\nPost-mortem human tissue often produces substantial variability that may be caused by a non-homogenous sample, where, interindividual variability, disease duration, age of death and post-mortem delay may all be contributing factors. We investigated this variability by looking for correlations between the total number of cells/mm^2^ containing \u03b1-syn compared to disease duration (R^2^\u2009=\u20090.09, p\u2009=\u20090.54), age at death (R^2^\u2009=\u20090.35, p\u2009=\u20090.21) and post-mortem delay (R^2^\u2009=\u20090.006217, p\u2009=\u20090.88), however, no significant correlations were found.\n\nDiscussion {#Sec5}\n==========\n\nConsistent with other studies, the majority of phosphorylated \u03b1-syn was present in the AON regions of the OFB and tract^[@CR5],[@CR16]^, however, \u03b1-syn pathology was also present in all the other layers of the PD OFB (Fig.\u00a0[1a,b](#Fig1){ref-type=\"fig\"}). The human AON is a complex structure that has been poorly characterised. Previous descriptions of the AON have typically subdivided it into four regions^[@CR16]^, however, more recently it has been suggested that the AON can be represented by 7 divisions along the rostro-caudal axis^[@CR33]^. Nissl-stained sections demonstrate differential cellular density and morphology between the AON and the rest of the OFB. Clusters of medium to large sized pyramidal neurons with a diameter of 15--20\u2009\u00b5m are often used to identify the AON region^[@CR7],[@CR16],[@CR34],[@CR35]^. In PD cases, the dense accumulation of \u03b1-syn has previously been used to delineate the AON regions, however, this does not allow for comparison between cases that do not have a significant amount of phosphorylated \u03b1-syn. Previous characterisations of the AON made it difficult to remain consistent when identifying multiple AON segments. The size and location of the AON can vary between sections within the same OFB and therefore significantly between different OFB cases. It also remains unclear whether the different AON segments are connected or have functionally distinctive roles^[@CR36]^. In this study, we used four cellular markers to target different structures of the OFB to more accurately delineate the AON, however, we did not make a distinction between the AON subdivisions (Fig.\u00a0[1a,c](#Fig1){ref-type=\"fig\"}).\n\nThe human AON is an area that is particularly affected by \u03b1-syn pathology early in the disease process^[@CR5],[@CR9]^. Evidence from rodent studies suggest that the AON receives direct projections from the different structures in the OFB including ipsilateral and contralateral centrifugal projections that include projections back to the OFB structures. There are projections to many of the secondary olfactory structures such as the entorhinal cortices, tertiary projections to brain regions such as the periamygdaloid cortex and rostral entorhinal cortex as well as projections from at least 27 non-olfactory regions^[@CR34],[@CR37],[@CR38]^. The structure of the OFB and the connections of the AON to the various other brain regions puts the AON in a position that could underlie the preferential involvement of \u03b1-syn pathology early in the disease. Previous studies highlight that neuronal cells such as mitral cells, neighbouring interneurons which express calcium-binding proteins, tyrosine hydroxylase expressing neurons and somatostatin neurons are affected by \u03b1-syn pathology and their projections to the tertiary structures deeper in the brain may provide a platform for the spread of \u03b1-syn along neural pathways^[@CR15],[@CR16],[@CR38]--[@CR40]^. In fact, *in vivo* studies in which various forms of \u03b1-syn inclusions have been injected into the OFB of mice have demonstrated that \u03b1-syn was detected in interconnected brain regions. These included the AON, frontal cortex, olfactory tubercle, periform cortex, striatum and amygdala. Local OFB interneurons were particularly good at internalising \u03b1-syn and it was likely that it would be propagating along these neural pathways into deeper brain regions^[@CR13]^.\n\nSupporting previous studies, this study demonstrates that neurons in the OFB are affected by \u03b1-syn pathology (Fig.\u00a0[3a,b](#Fig3){ref-type=\"fig\"})^[@CR15],[@CR16]^. However, we demonstrated that three of the four non-neuronal cells investigated -- microglia, pericytes and astrocytes contained intracellular \u03b1-syn in the AON of the human PD OFB. Microglia are excellent phagocytes and are likely to be the main cell involved in the clearing of \u03b1-syn from the extracellular space^[@CR19],[@CR26],[@CR41]^. *In vitro* evidence suggests that astrocytes can take up \u03b1-syn inclusions from the extracellular space and efficiently degrade it^[@CR19],[@CR26]^. Animal studies that have injected \u03b1-syn into different brain regions such as hippocampus and OFB, have shown that a significant amount of induced \u03b1-syn pathology was found in astrocytes and microglia^[@CR13],[@CR42]^. This suggests that they might be heavily involved in the clearance of \u03b1-syn (Fig.\u00a0[3a,b](#Fig3){ref-type=\"fig\"}). Interestingly, in tissue there is clear evidence to suggest that pericytes have macrophage-like properties^[@CR43]--[@CR45]^. Pericytes can actively phagocytose and may act as the last line of defence for the BBB, cleaning up the extracellular space and degrading foreign proteins and debris. Altogether, these data suggest that microglia, pericytes and astrocytes may have a role in the uptake and degradation of \u03b1-syn.\n\nAdditionally, non-neuronal cells may also be involved in the transfer of \u03b1-syn from one cell to the next. Primary human brain pericytes have been shown to transfer overexpressed \u03b1-syn from one pericyte to another^[@CR32]^. Through a similar mechanism, it has been demonstrated that astrocytes can transfer aggregated \u03b1-syn from one astrocyte to the next^[@CR30]^. More recently, primary rat astrocytes were shown to accept \u03b1-syn inclusions from neurons in culture and efficiently transfer it from astrocyte to astrocyte^[@CR27]^. However, evidence of these processes occurring *in vivo* is scarce. The fact that both astrocytes and pericytes contain intracellular \u03b1-syn in the AON of the human PD OFB suggests that they may have greater involvement in the disease processes. Further work needs to be done to understand whether these cells are actively involved in the transfer or degradation of \u03b1-syn in the human brain.\n\nOur study found that oligodendrocytes in the AON did not contain intracellular \u03b1-syn. Oligodendrocytes are implicated in multiple system atrophy, which is another synucleinopathy where \u03b1-syn pathology is predominantly present in neurons and oligodendrocytes^[@CR46]^. Previous studies have identified the presence of \u03b1-syn in oligodendrocytes in other regions of PD brain ^[@CR29],[@CR47]^. However, the findings in this study showed that oligodendrocytes were spared in the human PD OFB.\n\nPD is largely considered to be a disease of neurons, however, we observe that the \u03b1-syn inclusions in non-neuronal cells look similar to the \u03b1-syn inclusions seen in neuronal cells in the OFB. We identified that neurons can contain either the small \u03b1-syn inclusions, that resemble the types of inclusions seen in the non-neuronal cells (Fig.\u00a0[2a--c](#Fig2){ref-type=\"fig\"}) and the classical large Lewy body like inclusions (Fig.\u00a0[2e,f](#Fig2){ref-type=\"fig\"}). The difference in the size of the inclusions likely pertains to the ability of neurons to sequester \u03b1-syn inclusions into large Lewy body like inclusions^[@CR48]^. The process of Lewy body formation may develop over several stages, starting with diffuse cytoplasmic inclusions that ultimately develop into large Lewy body like inclusions ^[@CR48],[@CR49]^. Another point to consider regarding the size of the \u03b1-syn inclusions in the non-neuronal cells is their ability to divide and migrate^[@CR50],[@CR51]^. It is unclear whether this may alter \u03b1-syn inclusion size in non-neuronal cells in PD, however, there is evidence to suggest that cells that undergo proliferation may not be affected by aggregated protein inclusions that are found in degenerating mature neurons in diseases such as Huntington's disease and Machado--Joseph disease^[@CR52]--[@CR54]^. Therefore, the difference in \u03b1-syn inclusion size could be related to the amount of time that these cells contain the \u03b1-syn inclusions and because they are dividing, there is not enough time for Lewy body-like inclusions to develop.\n\nAlthough we demonstrate that microglia, pericytes and astrocytes contain \u03b1-syn in the human PD OFB, we acknowledge a particular caveat. All 11 PD cases used in this study have considerable pathology, suggesting that the AON regions of the OFB have had a prolonged and extensive exposure to \u03b1-syn pathology (Fig.\u00a0[1d,e](#Fig1){ref-type=\"fig\"}). Therefore, it is difficult to rule out whether these non-neuronal cells contain \u03b1-syn as a consequence of extremely dense \u03b1-syn pathology, or whether these non-neuronal cells are actively involved early in the disease processes. Interestingly, *in vivo* studies have demonstrated that injection of \u03b1-syn into the OFB in mice leads to a vast number of microglia being affected by \u03b1-syn pathology, but only at later time points. At earlier time points, few microglia cells in the OFB were positive for \u03b1-syn pathology^[@CR13]^. This brings into question whether a certain burden of pathology is needed before non-neuronal cells become involved.\n\nThis study demonstrates that microglia, pericytes and astrocytes in conjunction with neurons may play an important role in the pathogenesis of PD. Further studies are needed to elucidate the role of non-neuronal cells in the origins and spread of \u03b1-syn pathology in the brain. This may provide opportunities for therapeutic intervention by targeting non-neuronal cells.\n\nMaterials and Methods {#Sec6}\n=====================\n\nHuman brain tissue {#Sec7}\n------------------\n\nPost-mortem human OFBs were obtained from the Neurological Foundation Human Brain Bank and the Human Anatomy Laboratory within the Department of Anatomy and Medical Imaging, University of Auckland, New Zealand. Informed consent of the family was obtained prior to autopsy and the University of Auckland Human Participants Ethics Committee approved the protocols (Ref: 011654). All experiments were performed in accordance with relevant guidelines and regulations. The normal cases (n\u2009=\u200911) had no clinical history of neurological disease and no apparent pathological abnormalities upon post-mortem examination. The PD cases (n\u2009=\u200911) had a disease duration ranging from 9--23 years with an average of 16 years (Table 1). Although the post-mortem delay of the normal cases was on average higher than the PD cases, it did not appear to impact the ability to detect phosphorylated \u03b1-syn or any of the other markers used in this study (Table\u00a0[1](#Tab1){ref-type=\"table\"}). Pathological examination by a neuropathologist confirmed the clinical diagnosis of PD by observed presence of Lewy bodies in the substantia nigra as well as pigment incontinence and cell loss in the substantia nigra.Table 1Human OFB cases used in this study.Normal casesCase numberAgeSexPost-mortem delay (hours)Cause of deathOFB5185M20Carbon monoxide poisoningOFB5556M35Myocardial infarctionOFB5763F36Effects of diabetesOFB5860M36AsphyxiaOFB5967M20Complication of surgeryH19072F19Myocardial infarctionH24073M26.5Ruptured aneurysmH24377F13Ischaemic heart diseaseH24563M20AsphyxiaH24689M17Myocardial infarctionH25093F19PneumoniaAverage73 (Range: 56--93)7:4 (M:F)24 (Range: 13--36)**Parkinson's disease casesCase numberAgeSexPost-mortem delay (hours)Cause of deathDuration of PD (years)**PD5284M5Myocardial infarction12PD5379F25Renal failure9PD5478M6Aspiration pneumonia19PD5674M10.5End stage Lewy body disease12PD5882F18---15PD6391F5Parkinson's disease22PD6567M2.25Parkinson's disease9PD6673M17.5Aspiration pneumonia22PD6765M17Pneumonia12PD7776F6.5Abdominal carcinoma23PD7977M6.5End stage Lewy body disease22Average77 (Range: 65--84)7:4 (M:F)11 (Range: 5--25)16 (Range: 9--23)\n\nThe OFBs were obtained at autopsy and prepared as previously described^[@CR11]^. Briefly, the OFBs were fixed with 15% formaldehyde in 0.1\u2009M phosphate buffer for 24\u2009hours at room temperature. The OFBs were dehydrated in a graded ethanol series and embedded in paraffin wax using a Leica Tissue Processor. Dehydration comprised of sequential steps from 70%, 80% 2\u2009\u00d7\u200995% and 3\u2009\u00d7\u2009100% ethanol for 20\u2009minutes each at room temperature. The tissue was cleared in xylene for 2\u2009\u00d7\u200930\u2009minutes and inserted into molten paraffin wax during three cycles of 25\u2009minutes each. Paraffin blocks with embedded OFBs were sectioned at a thickness of 7\u2009\u00b5m on a rotary microtome (Leica Biosystems, RM2235). Sections were mounted individually on Superfrost Plus Slides (Menzel -- Gl\u00e4ser) and air-dried for at least 72\u2009hours at room temperature^[@CR11]^.\n\nThree OFB sections 500\u2009\u00b5m apart were selected per case for this study. The first section chosen was a mid-sagittal section of the OFB. The other two sections were then chosen 500\u2009\u00b5m either side of the first mid-sagittal section.\n\nImmunohistochemistry {#Sec8}\n--------------------\n\nSlides were heated to 60\u2009\u00b0C for 1\u2009hour to melt paraffin wax. Slides were cleared in xylene (2\u2009\u00d7\u200930\u2009mins) and rehydrated in an ethanol series: 2\u2009\u00d7\u2009100% (15\u2009mins), 1\u2009\u00d7\u200995% (10\u2009mins), 1\u2009\u00d7\u200980% (10\u2009mins), 1\u2009\u00d7\u200975% (10\u2009mins) followed by 3\u2009\u00d7\u20095\u2009mins in distilled H~2~O. Heat induced antigen retrieval was performed with a Tris -- EDTA (pH 9.0) buffer in a pressure cooker (2100 Antigen Retriever, Aptum Biologics Ltd.) for 20\u2009minutes at 121\u2009\u00b0C and left to cool for 1.5\u2009hours. Once cooled, slides were washed for 3\u2009\u00d7\u20095\u2009mins in phosphate -- buffered saline (PBS) and permeabilized in PBS - T (0.1% Triton X-100 in PBS) for 15\u2009minutes at 4\u2009\u00b0C. The sections were then blocked for non-specific secondary antibody binding for 1\u2009hour in 10% normal goat serum (Gibco \\#16210--072). Primary antibodies (Table\u00a0[2](#Tab2){ref-type=\"table\"}) were diluted in 1% normal goat serum and incubated on the sections overnight in a humidified chamber. Subsequently slides were washed 3\u2009\u00d7\u20095\u2009mins in PBS. Secondary antibodies (Table\u00a0[2](#Tab2){ref-type=\"table\"}) were diluted in 1% normal goat serum and incubated on the sections for 3\u2009hours at room temperature. Following this, sections were washed 3\u2009\u00d7\u20095\u2009mins in PBS. Sections were incubated for 5\u2009mins in PBS containing a 1:20 000 dilution of Hoechst 33342 (Molecular probes \\# H1399) to counterstain nuclei and subsequently washed 3\u2009\u00d7\u20095\u2009mins in PBS. Sections were coverslipped with Prolong Gold (Molecular Probes \\#P36930). Sections were sealed around the edges of the coverslip using nail polish and stored at 4\u2009\u00b0C in the dark until imaged^[@CR11]^.Table 2Primary and secondary antibodies used for immunohistochemistry.Primary AntibodiesAntibodySpeciesDilutionCatalogue \\#Manufacturer\u03b1-synuclein (phospho s129)Mouse1:3,000ab184674Abcam\u03b1-synuclein (phospho s129)Rabbit1:4,000ab190628AbcamPDGFR-\u03b2 \\[Y92\\]Rabbit1:200ab32570AbcamGFAPChicken1:4,000ab4674AbcamIba-1Chicken1:250ab139590AbcamNeuNChicken1:500MABN91EMD MilliporePGP9.5Mouse1:1,000ab8189AbcamUEA-1---1:500DL-1067Vector LaboratoriesCNPaseMouse1:500sc166558Santa Cruz Biotechnology**Secondary AntibodiesAntibodyDilutionCatalogue \\#Manufacturer**Goat anti -- mouse (488)1:400A11029Thermofisher ScientificGoat anti -- rabbit (594)1:400A11037Thermofisher ScientificGoat anti -- chicken (647)1:400A21449Thermofisher ScientificStrepavidin -- (647)1:500S21374Thermofisher Scientific\n\nImaging and quantification {#Sec9}\n--------------------------\n\nSections were imaged using an automated fluorescence microscope; Zeiss Z2 Axioimager equipped with MetaSystems VSlide slide scanner (MetaSystems) running MetaFer (V 3.12.1) coupled with MetaXpress using a 20x magnification objective lens (0.9 NA). Images were stitched using MetaCyte software. Following image capture, the total section scan was viewed using VSViewer (V 1.1.106) (MetaSystems) software. The AON regions were delineated using several antibodies (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}). Once the AON regions were identified, this process was applied to sequential sections. Cells with presumed intracellular \u03b1-syn were manually counted and marked for their location in the OFB using VSViewer (V 1.1.106) software. All cells with presumed intracellular \u03b1-syn were reimaged with a confocal microscope to confirm whether the \u03b1-syn was intracellular (Supplementary Fig.\u00a0[1](#MOESM1){ref-type=\"media\"}).\n\nConfocal images were acquired using a FV1000 confocal microscope (Olympus, Japan) with a 40x magnification oil immersion lens (1.00 NA), 60 x magnification oil immersion lens (1.35 NA) or 100 x magnification oil immersion lens (1.40 NA) in a Z-series using a step size of 0.5\u2009\u00b5m. Orthogonal projections with maximum intensity Z-projections were generated using ImageJ software.\n\nManual cell counts in the AON and calculation of the area fraction of \u03b1-synuclein (cell types) {#Sec10}\n----------------------------------------------------------------------------------------------\n\nTo count the total number of each cell type in the AON of the OFBs a manual counting method was used. Briefly, the AON regions were extracted using the VSViewer software and opened in ImageJ. Background intensity was measured using a 100 \u00b5m^[@CR2]^ box over three areas and averaged. The multi-point tool was used to pinpoint the brightest part of cellular labelling. The value of the integrated intensity of the point was recorded, background intensity was subtracted and if the value of the integrated intensity of the cell was above a determined threshold for that marker, it would be counted as a cell.\n\nThe total area fraction of \u03b1-syn in the OFB was determined by thresholding the \u03b1-syn labelling in the AON or across the entire OFB and tract. The thresholded area was subsequently normalised to total tissue area of the OFB or AON regions to obtain the percentage of \u03b1-syn coverage in a given area.\n\nStatistical analysis {#Sec11}\n--------------------\n\nIn general, data are presented as mean\u2009\u00b1\u2009standard deviation (SD) from the average of three different sections per case. Data visualization and statistical hypothesis testing was performed using GraphPad Prism Version 8.02. Linear regression was used to analyse correlations. One-way analysis of variance (ANOVA) was used when comparing across cell types with Tukey's multiple comparison adjustment and unpaired t-tests were used when comparing the area fraction of \u03b1-syn inside versus outside of the AON. Statistical significance was set as p\u2009\\<\u20090.05\n\nSupplementary information\n=========================\n\n {#Sec12}\n\nSupplementary Information. Supplementary Figure.\n\n**Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nThese authors contributed equally: Birger V. Dieriks and Maurice A. Curtis.\n\nSupplementary information\n=========================\n\nis available for this paper at 10.1038/s41598-020-63412-x.\n\nWe would like to acknowledge the generosity of the brain donors and their family for their generous gift of brain tissue for research. We also thank Marika Eszes at the Neurological Foundation Human Brain Bank and all technical staff involved in the collection and processing of the human brain tissue at the Centre for Brain Research and the Human Anatomy Labs. This work was supported by the Gillespie Neurological Foundation Postgraduate Scholarship, the Brain Research New Zealand Postgraduate Scholarship, the Deane Endowment Trust, the Neuro Research Charitable Trust and the Health Research Council of New Zealand (3710464).\n\nM.A.C. and B.V.D. contributed to the conception and design of the experiments. Material preparations were performed by M.A.C., B.V.D., T.J.S. and R.L.M.F. Pathological examination was performed by C.T. Data collection and analysis were performed by T.J.S., B.V.D. and H.C.M. The first draft of the manuscript was written by T.J.S. and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\n\nThe datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.\n\nThe authors declare no competing interests.\n"}
+{"text": "Sen\u00e9s\u2010Guerrero C, Col\u00f3n\u2010Contreras FA, Reynoso\u2010Lobo JF, Tinoco\u2010P\u00e9rez B, Siller\u2010Cepeda JH, Pacheco A. Biogas\u2010producing microbial composition of an anaerobic digester and associated bovine residues. MicrobiologyOpen. 2019;8:e854 10.1002/mbo3.854\n\n**Data Accessibility:** All data are included in the manuscript. DGGE sequences were deposited in the National Center for Biotechnology Information (NCBI) under accession numbers MH393448--MH393458 and metagenomic database as BioProject PRJNA378243. Additional data is provided in Appendix [A](#mbo3854-sup-0001){ref-type=\"supplementary-material\"}, Supplementary Material.\n\n1. INTRODUCTION {#mbo3854-sec-0001}\n===============\n\nBiogas plants are an attractive technology for sustainable generation of renewable energy. During anaerobic digestion a complex microbial community transforms organic wastes into biogas. Therefore, this practice exemplifies a sustainable solution for waste management and energy generation. The quantity and quality of the biogas, a mixture of methane, carbon dioxide and other trace gases, appears to be controlled by the type of biomass being digested and the microbial inoculum fed into the plant (Abendroth, Vilanova, G\u00fcnther, Luschnig, & Porcar, [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}; Sun, Pope, Eijsink, & Schn\u00fcrer, [2015](#mbo3854-bib-0023){ref-type=\"ref\"}; Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}).\n\nTraditionally, animal manure and sludge from wastewater treatment plants have been used to generate biogas (Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}). Animal manure and slurries from cattle and swine production have been estimated as one of the largest waste streams for biogas generation (Holm\u2010Nielsen, Al Seadi, & Oleskowicz\u2010Popiel, [2009](#mbo3854-bib-0010){ref-type=\"ref\"}). In the European Union, it is estimated that more than 1,284 million ton/year of manure is produced by cattle, according to an average of 38.5\u00a0kg manure days^\u22121^\u00a0head^\u22121^ (Holm\u2010Nielsen et al., [2009](#mbo3854-bib-0010){ref-type=\"ref\"}). Worldwide, in year 2016 livestock, represented 1,475 million heads of cattle, which would roughly account for 21 billion ton of animal manure (FAOStat, [2016](#mbo3854-bib-0007){ref-type=\"ref\"}). If left untreated or inadequately managed, animal manure becomes a major environmental problem because of nutrient leaching (N, P), ammonia evaporation and pathogen contamination. In addition, livestock production is estimated to be responsible of 18% of greenhouse gas (GHG) emissions and the anthropogenic source of 9% carbon dioxide, 37% methane and 65% nitrous oxide (Steinfeld et al., [2006](#mbo3854-bib-0021){ref-type=\"ref\"}). However, in most countries, only a small percentage of this waste is currently processed to generate biogas. As countries commit internationally to reduce GHG and incorporate renewable sources of energy, biogas generation and its optimization gains importance. Treated manure would also generate a residual solid and liquid fraction rich in bioavailable nutrients termed the digestate, considered a valuable end product and a mode to recycle nutrients from agriculture as it is commonly used as a biofertilizer (Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}).\n\nThe stable operation of an anaerobic digester is dictated by a dynamic equilibrium among four bacterial groups involved in the sequential digestion of the biomass from complex polymers to simpler components that are used by methanogens to generate methane. Therefore, it is crucial to elucidate the microbial community structure and function during digestion to understand and potentially optimize the process, as it can easily explain biodigester malfunctions. First, hydrolytic bacteria transform complex polymers into sugars and amino acids, followed by the process of acidogenesis and acetogenesis to generate organic acids that are transformed into acetate, H~2~ and CO~2~ for methanogenesis. Recently, it has been suggested that fungi might play an important role in the hydrolytic stage assisting bacteria in gaining access to recalcitrant plant materials (Bengelsdorf, Gerischer, Langer, Zak, & Kazda, [2012](#mbo3854-bib-0004){ref-type=\"ref\"}). Therefore, it is of interest to study the whole microbial community and not only the prokaryotic component. Usually, culture\u2010independent techniques that rely on the analysis of the 16S rRNA gene such as sequence analysis of clone libraries, fluorescence in situ hybridization, denaturing gradient gel electrophoresis (DGGE), restriction fragment length polymorphism, and 16S amplicon sequencing are used to study these complex microbial interactions (Ariesyady, Ito, & Okabe, [2007](#mbo3854-bib-0003){ref-type=\"ref\"}; Bengelsdorf et al., [2012](#mbo3854-bib-0004){ref-type=\"ref\"}; Goberna, Insam, & Franke\u2010Whittle, [2009](#mbo3854-bib-0008){ref-type=\"ref\"}; Sun et al., [2015](#mbo3854-bib-0023){ref-type=\"ref\"}; Tuan, Chang, Yu, & Huang, [2014](#mbo3854-bib-0026){ref-type=\"ref\"}; Whitford, Teather, & Forster, [2001](#mbo3854-bib-0028){ref-type=\"ref\"}; Zhou, Hernandez\u2010Sanabria, & Guan, [2010](#mbo3854-bib-0032){ref-type=\"ref\"}). However, these methods might present certain bias toward specific microbial groups and, since they are based on known sequences, do not cover all microbial diversity. Today, next generation sequencing allows a real\u2010time assessment of the whole microbial community involved in the process with applications such as metagenomics that do not rely on PCR\u2010targeted amplification. These data can be used to establish taxonomy, genome composition, and metabolic potential of the microorganisms in a sample. Nevertheless, as many microorganisms are still uncultured, not all microbial components of the community are identified, and bioinformatics methods and analysis platforms not always facilitate data interpretation (Menzel, Ng, & Krogh, [2016](#mbo3854-bib-0014){ref-type=\"ref\"}). Hence, it is still of relevance to compare results using different approaches.\n\nIn an interest to achieve higher biogas yields and improve the fertilizer value of the digestate, addition of other agricultural substrates or residues is being considered to increase the organic content of the treated biomass (Fantozzi & Buratti, [2009](#mbo3854-bib-0006){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}). Considering bovine residues, there has always been an interest in the methanogenic composition of the rumen not only to increase meat production but also as a microbial inoculum to digest plant\u2010material during anaerobic digestion (Fantozzi & Buratti, [2009](#mbo3854-bib-0006){ref-type=\"ref\"}; Zhou et al., [2010](#mbo3854-bib-0032){ref-type=\"ref\"}). Mexico is in the top ten producers and exporters of bovine meat in the world (SAGARPA, [2017](#mbo3854-bib-0020){ref-type=\"ref\"}). One of the largest meat producers in Mexico, in an effort to adopt sustainable practices, is implementing a biodigester to treat most of its wastes. Water from the oxidation lagoon of the company as well as fresh rumen and leachate from previous digester runs is incorporated into a dry mesophilic digester in order to efficiently treat animal manure. The objective of this study was to conduct a time\u2010lapse composition analysis of the biogas\u2010producing microbial community in a 10\u2010ton anaerobic digester and its associated bovine residues by an integrated approach that contemplates DGGE and shotgun metagenomics. By analyzing the microbial component of the different feedstocks, we pretend to establish the best substrates for biogas generation by taking into account particular microorganisms already adapted to this type of waste.\n\n2. MATERIALS AND METHODS {#mbo3854-sec-0002}\n========================\n\n2.1. Sample collection {#mbo3854-sec-0003}\n----------------------\n\nResidual biomass from the bovine industry associated with cattle raising and meat processing was evaluated as potential feedstocks and microbial inocula for biogas generation. Samples were taken from an oxidation lagoon (OL) treating wastewater from the meat production plant, leachate (L) from a previous biodigester run, 6\u2010month\u2010old dried cattle manure (M) and fresh rumen (R) fluid. In addition, a pilot\u2010scale dry anaerobic digester fed with these feedstocks and other supporting substrates (64.2% manure, 18.2% oxidation lagoon water, 3% rumen, 9.6% wood chips, 4.0% corn stover, 0.9% dust mill) was evaluated by sampling through time the recirculating leachate between two 10\u2010ton serial reactors. The digester was operated under batch mesophilic dry condition (60% total solids and 32% volatile solids). Biodigester (B) samples included seven points during the 22\u00a0days of operation in the following dates: 17.01.2014 (B17), 20.01.2014 (B20), 22.01.2014 (B22), 27.01.2014 (B27), 29.01.2014 (B29), 31.01.2014 (B31), and 04.02.2014 (B04). Sterile 1 L\u2010Nalgene bottles filled to the top were used for liquid samples (OL, L, B), while solid samples (M, R) were packed in sterile 18\u00a0oz. Whirl\u2010Pak bags. All samples were kept at 4\u00b0C until DNA extraction. Feedstock samples were processed during the following days (4--7\u00a0days) after arrival at the laboratory and samples from the biodigester run were processed all together after the run concluded (22\u00a0days). A preliminary evaluation of sample processing time was conducted by comparing DGGE profiles (data not shown), the same community profiles were obtained from samples stored for 4--7\u00a0days compared to around one month.\n\n2.2. DNA extraction {#mbo3854-sec-0004}\n-------------------\n\nDifferent volumes of the liquid samples (25 and 50\u00a0ml) were first evaluated to optimize extraction. Samples were filtered through a series of pore sizes to remove large particles (coffee filter, 20--25\u00a0\u03bcm and 2.5\u00a0\u03bcm) until a cell pellet was obtained at the last filtration step (0.45 or 0.22\u00a0\u03bcm). By comparing DGGE profiles (data not shown), it was established that 25\u00a0ml filtered to 0.45\u00a0\u03bcm was ideal to reduce sample manipulation in a timely matter. Filters from liquid samples and 0.5\u00a0g from solid samples were used for DNA extraction using the FastDNA Spin Kit for Soil (MP Biomedicals, USA). DNA quality was evaluated by agarose gel electrophoresis and spectrophotometry (Nanodrop 1000, Thermo Scientific, USA), and quantified by fluorometry (Qubit 2.0, Invitrogen, USA).\n\n2.3. Denaturing gradient gel electrophoresis analysis {#mbo3854-sec-0005}\n-----------------------------------------------------\n\nA nested PCR approach was used for targeted amplification of methanogens as described by Zhou et al. ([2010](#mbo3854-bib-0032){ref-type=\"ref\"}). First, a large 800\u00a0bp fragment of the 16S rRNA gene was amplified (94\u00b0C, 5\u00a0min; 30 cycles of: 94\u00b0C for 30\u00a0s, 57\u00b0C for 30\u00a0s, 68\u00b0C for 60\u00a0s; and 68\u00b0C, 7\u00a0min) with primers Met86f (5\u2010GCTCAGTAACACGTGG\u20103) and Met915r (5\u2010GTGCTCCCCCGCCAATTCCT\u20103). Then a nested PCR (1:100 dilution of previous PCR product) was performed with primers GC\u2010ARC344f (5\u2010ACGGGGYGCAGCAGGCGCGA\u20103) and 519r (5\u2010GWATTACCGCGGCKGCTG\u20103), the forward primer possessed a 40\u00a0bp GC\u2010clamp, which targeted a 191\u00a0bp fragment in the 16S rRNA\u2010V3 region (95\u00b0C, 5\u00a0min; 30 cycles of: 95\u00b0C for 30\u00a0s, 56.5\u00b0C for 30\u00a0s, 72\u00b0C for 30\u00a0s; and 72\u00b0C, 7\u00a0min). PCR products were loaded on a 1% (w/v) agarose gel with 1X Tris\u2010acetate\u2010EDTA (TAE) buffer and visualized after ethidium bromide (0.5\u00a0g/L) staining.\n\nDenaturing gradient gel electrophoresis of PCR products was performed using a DCode Universal Mutation System (Bio\u2010Rad, USA) in 1X TAE buffer with a 1.0\u00a0mm\u2010thick vertical gel containing 6% (w/v) polyacrylamide (37.5:1 acrylamide:bisacrylamide) and a 35--55% (w/v) linear gradient of denaturants (100% denaturation solution contained 7\u00a0M urea and 40% (w/v) formamide). Gel wells were loaded with 35--45\u00a0\u00b5l of the nested PCR product according to agarose gel band intensity and 1\u20444\u2010volume of loading buffer. Running conditions were 3.5\u00a0hr at 150\u00a0V. After, the gel was stained with ethidium bromide according to the manufacture\\'s protocol, visualized on a UV transilluminator at 312\u00a0nm using a Molecular Imager ChemiDoc XRS System (Bio\u2010Rad, USA). The most intense bands were excised in the middle with RNase/DNase clean scalpels and DNA was eluted according to Chory and Pollard ([2001](#mbo3854-bib-0005){ref-type=\"ref\"}). An aliquot (2\u00a0\u00b5l) was used for PCR re\u2010amplification using conditions described above and a second DGGE was run to confirm band purity. PCR products were cleaned with a PCR Clean\u2010Up System (Promega, USA) and sequenced with the same primer pair without the GC clamp at Eton Bioscience, Inc. (San Diego, USA). To determine the closest known relative species, sequences were blasted against the NCBI GenBank and MiDAS 2.1 database (Mcllroy et al., [2017](#mbo3854-bib-0013){ref-type=\"ref\"}). Sequences were deposited in NCBI under accession numbers\u00a0MH393448\u2010MH393458.\n\nSimilarities among DGGE community profiles were defined by analyzing gel images using ImageJ 1.48 (Rasband, [1997](#mbo3854-bib-0018){ref-type=\"ref\"}). Bands of each lane were detected automatically and their relative intensity measured by the peak area. Bands with \\<1% intensity with respect to the total intensity of the lane were removed from the analysis. A Manhattan distance matrix was generated for pairwise comparisons between lanes using MeV_4\\_8 v. 10.2 (Saeed et al., [2003](#mbo3854-bib-0019){ref-type=\"ref\"}). This matrix was used for hierarchical clustering using the unweighted pair group method with arithmetic mean.\n\n2.4. Shotgun metagenomic analysis {#mbo3854-sec-0006}\n---------------------------------\n\nGenomic DNA was used for library preparation using Nextera XT DNA Library Prep Kit (Illumina, USA) according to the manufacturer\\'s protocol. Libraries were sequenced (151\u00a0bp paired\u2010end) with MiSeq Reagent Kit v3, 600 cycles, using the MiSeq system (Illumina, USA) at the sequencing facilities of Tecnologico de Monterrey (Monterrey, Mexico). Raw sequenced data were processed by the FASTQ Toolkit v2.2.0 (BaseSpace Labs, Illumina, USA) to trim adapter sequences and remove reads below a mean quality of Q20, unpaired reads and reads \\<32\u00a0bp length. Taxonomic characterization was done by the metagenomic classifier Kaiju (Menzel et al., [2016](#mbo3854-bib-0014){ref-type=\"ref\"}) using the NCBI BLAST *nr*\u00a0+\u00a0euk reference dataset, searching for maximum exact matches with a minimum match length of 11. Kaiju classifies individual metagenomic reads using a reference database comprising the annotated protein\u2010coding genes of a set of microbial genomes. For this study, the reference dataset contained 89\u00a0M protein sequences from Bacteria, Archaea, virus, Fungi, and other\u00a0microbial eukaryotes. All sequences have been archived in the NCBI Sequence Read Archive under BioProject no. PRJNA378243. Relative read abundance as the proportion of raw reads of each taxon from the total amount of reads was used to assess the distribution of taxa across the different samples. Reads that were not assigned to any taxa at the phylum level were removed from the analysis.\n\nFor multivariate statistical analysis of metagenomic data, raw reads were normalized using DESeq2 (Love, Huber, & Anders, [2014](#mbo3854-bib-0012){ref-type=\"ref\"}) with the counts function and the parameter normalized=TRUE. A normalized data matrix of phyla from each microbial domain was used for hierarchical clustering using the package Pvclust (Suzuki & Shimodaira, [2006](#mbo3854-bib-0024){ref-type=\"ref\"}) with Ward\\'s method and a bootstrap number of 10,000. This data matrix was also analyzed by principal component analysis (PCA) with Euclidean distances and the total amount of inertia equal to number of species, using the vegan package (Oksanen et al., [2017](#mbo3854-bib-0016){ref-type=\"ref\"}). All statistical analyses were conducted in R v.3.3.2 (R Core Team, [2016](#mbo3854-bib-0017){ref-type=\"ref\"}).\n\n3. RESULTS {#mbo3854-sec-0007}\n==========\n\n3.1. Performance of the 10\u2010ton anaerobic digester {#mbo3854-sec-0008}\n-------------------------------------------------\n\nDuring sampling, the biodigester showed a lag\u2010phase of 6\u00a0days followed by biogas formation and 8\u00a0days of peak production until activity came to a halt (Figure [1](#mbo3854-fig-0001){ref-type=\"fig\"}). Residual oxygen in the biodigester decreased from day 1 to 8 from 3.8% to 2.1% v/v and remained around 2% during biogas production. Volatile fatty acids (VFAs) in the digester leachate increased from 2.4\u00a0\u00b1\u00a00.2\u00a0g acetic acid L^\u22121^, at the start of the run, to 14.4\u00a0\u00b1\u00a00.7\u00a0g/L at peak gas production and significantly decreased to a range of 4.6--7.6\u00a0g/L by the end of the run, indicating biomass breakdown and conversion of VFAs. Other performance parameters during the run showed a pH of 6.8--7.9, total inorganic carbon (TIC) of 5.6--9.3\u00a0g CaCO~3~ L^\u22121^ and VFAs/TIC ratios of 0.4--1.34. At peak production, 9.7\u00a0\u00b1\u00a00.7\u00a0m^3^\u00a0days^\u22121^ of biogas were produced with 49.6\u00a0\u00b1\u00a03.3% methane content. After 22\u00a0days of digestion, biogas and methane yield (46.1\u00a0L biogas kg\u00a0VS^\u22121^ and 25.5\u00a0L CH~4~ kg\u00a0VS^\u22121^, 55.3% CH~4~) were in the low range of reported values for mesophilic treatment of cattle manure due to mechanical problems with the run (i.e. pump failure during leachate recirculation at day 4 to day 14, when it was re\u2010establish). Although, biogas production peak during this period with an accumulation of VFAs that\u00a0were later consumed. Fantozzi and Buratti ([2009](#mbo3854-bib-0006){ref-type=\"ref\"}) reported productivities of 40\u00a0L CH~4~ kg\u00a0VS^\u22121^ for bovine fresh manure in a laboratory reactor (17\u00a0L working volume) and cited literature values of 170 to 220\u00a0L CH~4~ kg\u00a0VS^\u22121^. However, these reports contemplate thermophilic operations and not necessarily dry fermentation conditions.\n\n![Biogas production from anaerobic digestion of bovine residues in a 10\u2010ton pilot biogas plant under mesophilic dry fermentation conditions. Arrows indicate sampling time for microbial analysis](MBO3-8-e00854-g001){#mbo3854-fig-0001}\n\n3.2. Methanogen DGGE analysis {#mbo3854-sec-0009}\n-----------------------------\n\nOur first approach contemplated DGGE analysis of archaeal PCR products from the V3 region of the 16S rRNA gene. In this technique, community profiles from different samples were evaluated by studying the most abundant members, which corresponded to the most intense bands in the DGGE profile (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a). A total of 38 bands were cut from the gel and 68% of bands were successfully purified and sequenced. These bands were annotated based on their closest similarity to cultured or uncultured species in two gene databases. Nine methanogen species were identified, and five bands were classified as uncultured archaeons (Table [S1](#mbo3854-sup-0001){ref-type=\"supplementary-material\"}). Most sequences belonged to the *Methanobrevibacter* genus, which seemed to produce a multiple band pattern. Different strains of *Methanobrevibacter\u00a0smithii* (Mbs) were present in almost all samples (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a, Band 6, 11, 17) and *Methanobrevibacter\u00a0ruminantium* (Mbr) dominated the oxidation lagoon (OL),\u00a0while *Methanobrevibacter\u00a0boviskoreani* the rumen (R) (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a, Band 7, 10). Even though they were not as ubiquitous, intense bands of uncultured archaeon 1 (Uc1) and 3 (Uc3) (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a, Band 15, 18), and *Methanolinea mesophila* (Mlm; Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a, Band 27) characterized the biodigester run (B17\u2010B04). Leachate (L) also presented these bands and the oxidation lagoon (OL) was the only other feedstock where *M.\u00a0mesophila* was observed. At the end of the biodigester run, uncultured archaeon 2 (Uc2) and *Methanoculleus\u00a0marisnigri* (Mcm) became abundant (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a, Band 4, 24). The latter also dominated the manure (M) sample. Overall, methanogen richness (14--17 bands; Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a) was the highest in the biodigester after 10\u00a0days into the run when maximum biogas production rates were observed (Figure [1](#mbo3854-fig-0001){ref-type=\"fig\"}). Among feedstocks, oxidation lagoon (OL) was the most diverse (14 bands) and rumen (R) the least (9 bands) (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}a). By comparing the gel band pattern of each community, it was determined that the first days of the biodigester run (B17\u2010B22, Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}b) were more similar among each other than the middle (B27\u2010B31) and last sampled day (B04). All biodigester samples shared a high similarity to leachate (L), with the exception of the last day (B04) that was more similar to manure (M) (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}b). Feedstocks oxidation lagoon (OL) and rumen (R) shared similar methanogen compositions and differed from the biodigester run and other feedstocks.\n\n![Denaturing gradient gel electrophoresis gel of archaeal PCR products (a) and band pattern analysis (b) showing hierarchical clustering of samples and band intensity. Arrows indicate cut bands from gel. Samples: OL, oxidation lagoon; L, leachate; R, rumen; M, manure; B, biodigester time series (B17\u00a0=\u00a017.01.2014, B20\u00a0=\u00a020.01.2014, B22\u00a0=\u00a022.01.2014, B27\u00a0=\u00a027.01.2014, B29\u00a0=\u00a029.01.2014, B31\u00a0=\u00a031.01.2014, B04\u00a0=\u00a004.02.2014, B04\u2010R\u00a0=\u00a004.02.2014\u2010replicate). Band identity: NS, not sequenced; Mss, *Methanosphaera stadtmanae*; Uc, uncultured archaeon; Mbb, *Methanobrevibacter boviskoreani*; Mbs, *Methanobrevibacter smithii*; Mbr, *Methanobrevibacter ruminantium*; UMb, uncultured *Methanobrevibacter*; Msc, *Methanosaeta concilii*; Mcm, *Methanoculleus marisnigri*; Msm, *Methanosarcina mazei*; Mlm, *Methanolinea mesophila*; Msh, *Methanospirillum hungatei*](MBO3-8-e00854-g002){#mbo3854-fig-0002}\n\n###### \n\nNumber of paired\u2010end reads and read relative abundance in parenthesis (%) for each microbial domain (based on classified reads)\n\n  Domain                   L[a](#mbo3854-note-0002){ref-type=\"fn\"}              M                  OL               R                  Biodigester                                                         \n  ------------------------ ---------------------------------------------------- ------------------ ---------------- ------------------ ---------------- ---------------- ---------------- ---------------- ----------------\n  Eukaryota                14,726 (2.5)                                         13,378 (1.2)       8,081 (1.5)      22,984 (1.9)       9,288 (3.1)      16,575 (3.1)     19,823 (3.6)     27,846 (2.8)     17,706 (2.6)\n  Bacteria                 472,618 (79.9)                                       1,083,849 (97.9)   504,931 (95.9)   1,187,699 (96.1)   255,101 (84.5)   440,644 (82.1)   487,003 (88.1)   858,988 (86.6)   594,111 (88.2)\n  Archaea                  102,535 (17.3)                                       8,123 (0.7)        12,674 (2.4)     22,692 (1.8)       36,553 (12.1)    77,900 (14.5)    44,203 (8.0)     102,580 (10.3)   59,013 (8.8)\n  Virus                    2,019 (0.3)                                          1,960 (0.2)        893 (0.2)        2,571 (0.2)        1,074 (0.4)      1,438 (0.3)      1,922 (0.4)      2,912 (0.3)      2,797 (0.4)\n  Classified reads         591,898 (42)[b](#mbo3854-note-0003){ref-type=\"fn\"}   1,107,310 (61)     526,579 (49)     1,235,946 (45)     302,016 (47)     536,556 (52)     552,951 (42)     992,326 (44)     673,627 (44)\n  Total paired\u2010end reads   1,417,913                                            1,809,324          1,074,462        2,762,894          643,912          1,039,131        1,328,532        2,266,266        1,544,364\n\nSamples: L, leachate; M, manure; OL, oxidation lagoon; R, rumen; B, biodigester time series (B17\u00a0=\u00a017.01.2014, B22\u00a0=\u00a022.01.2014, B27\u00a0=\u00a027.01.2014, B29\u00a0=\u00a029.01.2014, B04\u00a0=\u00a004.02.2014).\n\nPercentage of classified reads out of the total amount of paired\u2010end reads passing quality filters.\n\nJohn Wiley & Sons, Ltd\n\n3.3. Shotgun metagenomic analysis {#mbo3854-sec-0010}\n---------------------------------\n\n### 3.3.1. Whole\u2010microbial community at high taxonomic level {#mbo3854-sec-0011}\n\nTo analyze all microbial components of the samples a metagenomic approach was performed. Sequencing of all sample libraries resulted in 27,773,646 single reads passing quality filters, which corresponded to 97.9% total reads (individual reads per sample are shown in Table [S2](#mbo3854-sup-0002){ref-type=\"supplementary-material\"}). Taxonomic identity per sample was possible in 42%--61% of reads (Table [1](#mbo3854-tbl-0001){ref-type=\"table\"}). The manure (M) feedstock showed the highest percentage of classified reads (61%), while leachate (L) along with the last days of the biodigester run presented the least classified (56%--58%). Table [1](#mbo3854-tbl-0001){ref-type=\"table\"} shows that the major microbial domain in all samples was Bacteria with a relative abundance of 79.9%--97.9%, followed by Archaea (0.7%--17.3%), Eukaryota (1.2%--3.6%) and a small not that variable percentage of virus (0.2%--0.4%). All biodigester run samples presented higher amounts of Archaea (8.0%--14.5%) compared to feedstocks (0.7%--2.4%), except for leachate (L) with a relative abundance of 17.3%. Also, the Eukaryota component was higher in the biodigester (2.6%--3.6%) compared to feedstocks (1.2%--2.5%).\n\nAs community dynamics of eukaryotes and bacteria influence the ability of methanogens to generate biogas, classified reads were further analyzed at lower taxonomic levels (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}). All samples enclosed a similar composition of major eukaryote groups, where fungi predominated (69%--73%) (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}a). The Bacteria domain was led by four phyla (Proteobacteria, Firmicutes, Bacteroidetes/Chlorobi and Actinobacteria) that differed in read relative abundance among samples. In the biodigester samples, differences in the Bacteria component seemed to response to time. However, as microbial biomass was not assessed per sample, direct comparisons among samples can be misleading. Proteobacteria at day B17 represented 47% of the community but by day B04 only 18%. On the contrary, the proportion of the Bacteroidetes/Chlorobi group was 8% at the start of the run and at the end was around 20%, similar to Actinobacteria (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}b). Firmicutes proportions in all biodigester samples were between 12%--18% and Spirochaetes seemed to be important components of the community at the end of the run (6%--10%). Among feedstocks, bacteria composition was distinct for each sample. Leachate (L) was the most uniform community, resembling the biodigester run where Proteobacteria and Bacteroidetes/Chlorobi were the main phyla with an abundance each of 25%--30% followed by Firmicutes and Actinobacteria (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}b). Actinobacteria dominated in manure (M) with an abundance of 44.7%, whereas Proteobacteria accounted for 43.8% of reads in the oxidation lagoon (OL). Finally, the rumen (R) consisted mostly of Firmicutes (47.9%).\n\n![Read relative abundance of Eukaryota major taxa (a), bacterial phyla (b), and methanogen orders (c). Samples: L, leachate; M, manure; OL, oxidation lagoon; R, rumen; B, biodigester time series (B17\u00a0=\u00a017.01.2014, B22\u00a0=\u00a022.01.2014, B27\u00a0=\u00a027.01.2014, B29\u00a0=\u00a029.01.2014, B04\u00a0=\u00a004.02.2014)](MBO3-8-e00854-g003){#mbo3854-fig-0003}\n\nThe Archaea community in all samples was predominantly of the phylum Euryarchaeota (89.5%--97.5%) with 1.4%--4.5% unclassified Archaea (data not shown). Interestingly, the manure sample also presented a proportion of Crenarchaeota (5.1%) and Thaumarchaeota (2.0%). Lower abundances of Crenarchaeota were also found in the oxidation lagoon and rumen (2.4%--3.0%). Methanogens, which belong to the phylum Euryarchaeota, were represented by the orders Methanosarcinales, Methanomicrobiales and Methanobacteriales (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}c). In addition, there was a significant contribution of reads classified as other Euryarchaeota (0.1%--52.3%). During the biodigester run, Methanomicrobiales was the most abundant order (26%--77%) followed by Methanosarcinales (11%--37%) and a low representation of Methanobacteriales (0.7%--2.3%). However, at day B22 there was a shift toward other Euryarchaeota but the community reestablished by day B27, when biogas was steadily produced (B27 and B29, Figure [1](#mbo3854-fig-0001){ref-type=\"fig\"}). Therefore, peak biogas production was characterized by the dominance of the order Methanomicrobiales (61%--64%), a slight reduction in Methanosarcinales (22%--24%) and the reestablishment of lower levels of other Euryarchaeota (12%--13%). Among feedstocks, leachate (L) was almost equally represented by Methanomicrobiales and Methanosarcinales, and the oxidation lagoon (OL) presented a similar profile as the biodigester but with a higher proportion of Methanobacteriales. Lastly, Methanobacteriales dominated the rumen (R) community (63.3%) and was also present in manure (M), along with Methanomicrobiales, Methanosarcinales and an important contribution of other Euryarchaeota.\n\nSimilarities among the metagenome of each sample, considering Eukaryota, Bacteria and Archaea, were evaluated by hierarchical clustering (Figure [4](#mbo3854-fig-0004){ref-type=\"fig\"}). At the highest taxon level, significant (*p*\u00a0\u2264\u00a00.05) microbial community profile similarities were found between leachate (L) and the middle days of the biodigester run (B27, B29), while a second cluster corresponded to the first days of the biodigester (B17, B22) (Figure [4](#mbo3854-fig-0004){ref-type=\"fig\"}a). PCA supported these groupings but showed that the last day of the biodigester run (B04) was more similar to oxidation lagoon (OL) than manure (M) (Figure [4](#mbo3854-fig-0004){ref-type=\"fig\"}b). Moreover, manure (M) and rumen (R) did not group with the biodigester samples. Although not statistically significant, methanogen community profiles at the order level clustered similarly to the whole community analysis, with the exception of leachate\u00a0(L) which grouped with the branch comprised of the last day of the biodigester run (B04), manure (M) and rumen (R) (Figure [4](#mbo3854-fig-0004){ref-type=\"fig\"}c). Methanogen PCA grouped all biodigester samples with leachate (L), while oxidation lagoon (OL), manure (M), and rumen (R) each formed an independent group (Figure [4](#mbo3854-fig-0004){ref-type=\"fig\"}d).\n\n![Hierarchical clustering and principal component analysis of the metagenomes from samples annotated at the phylum level (a--b) and of methanogens at the order level (c--d). Red squares represent clusters with \\>95% AU supported *P*\u2010values. AU, approximately unbiased; BP, bootstrap probability; Samples: L, leachate; M, manure; OL, oxidation lagoon; R, rumen; B, biodigester time series (B17\u00a0=\u00a017.01.2014, B22\u00a0=\u00a022.01.2014, B27\u00a0=\u00a027.01.2014, B29\u00a0=\u00a029.01.2014, B04\u00a0=\u00a004.02.2014)](MBO3-8-e00854-g004){#mbo3854-fig-0004}\n\n### 3.3.2. Methanogen community at genus and species level {#mbo3854-sec-0012}\n\nOverall, methanogen orders were represented by 12 genera and 33 species, which possessed \u22651% read relative abundance in at least one sample (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"} and Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}). Hydrogenotrophic and acetotrophic methanogens were found in high abundances, contrary to methylotrophs only represented by the genus *Methanosphaera* in the rumen (R) sample (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}). In the biodigester run, the most abundant orders (Methanomicrobiales, Methanosarcinales, and Methanobacteriales) were represented each by a genus (*Methanoculleus*, *Methanosaeta*, and *Methanolinea*, respectively) that seemed to change in abundance with time. *Methanoculleus*, a hydrogenotroph, dominated the biodigester community (70% relative abundance) by the end of the run, while *Methanosaeta,* an acetotroph, and *Methanolinea,* a hydrogenotroph, were minor components (7% and 5%, respectively) (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}a). At peak biogas production, these genera presented intermediate abundances in the observed range under the dominance of *Methanoculleus* (48%--38%). Other minor members were *Methanoregula* and *Methanosarcina* (2%--5%). All other genera showed lower abundances (\\<4%) and were present at specific sampling times, *Methanospirillium* and *Methanobacterium* during the first days, while *Methanofollis* increased at the end of the run. The proportion of genera below \\<1% read relative abundance (5%--6%) and unclassified methanogens (5%--9%) remained constant throughout the run.\n\n![Distribution of methanogens at the genus level (\u22651% relative abundance) in the biodigester run (a) and associated feedstocks (b--e). Proportions of the genus relative abundance are shown in the table. ^a^Relative abundance \\<1%, considered in category \\<1% RA. Samples: L, leachate; M, manure; OL, oxidation lagoon; R, rumen; B, biodigester time series (B17\u00a0=\u00a017.01.2014, B22\u00a0=\u00a022.01.2014, B27\u00a0=\u00a027.01.2014, B29\u00a0=\u00a029.01.2014, B04\u00a0=\u00a004.02.2014)](MBO3-8-e00854-g005){#mbo3854-fig-0005}\n\nThe feedstocks leachate (L), manure (M) and rumen (R) (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}b,c,e) were dominated each by a different methanogen, *Methanosaeta* (45%), *Methanosarcina* (39%) and *Methanobrevibacter* (72%), respectively. Leachate (L) and manure (M) also presented a high proportion of *Methanoculleus* (36 and 24%, respectively). On the contrary, oxidation lagoon (OL) presented a diverse methanogen community (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}d), where *Methanosaeta* (19%) and *Methanolinea* (18%) were the most abundant genera. Interestingly, *Methanobrevibacter* was not observed in the biodigester even though it was an abundant component of rumen (72%) and manure (20%).\n\nRelevant methanogen species in the biodigester run were consistent with the pattern described for genera. *Methanoculleus*, as one of the most abundant genera, was represented by eight species (Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}), where *M. marisnigri* was the most abundant at the end of the run (7%) followed by *M. horonobensis* and strain MH98A (6%). On the contrary, *Methanosaeta*, represented by *Methanosaeta concilii* and *Methanosaeta harundinacea*, were abundant at the start of the run (11%--13%). At peak biogas production the most abundant species in decreasing order were *M.\u00a0concilii* (15%--17%), *Methanolinea* sp. strain SDB (8%--14%), *Methanolinea tarda* (4%--5%) and *M.\u00a0marisnigri* (5%--4%). However, as mentioned above, *Methanoculleus* was represented by several species that add up to 20%--25% read relative abundance followed by *Methanosaeta* (18%--19%) and *Methanolinea* (12%--19%) species. Other species that appeared at the end of the run (B04) but in low abundance (around 2%) were *Methanofollis ethanolicus*, *Methanofollis liminatans* and *Methanosarcina mazei*. Interestingly, the most abundant species in the feedstock leachate (L), *M. harundinacea* (20%), was not the species of *Methanosaeta* that proliferated in the biodigester during peak production. Instead, *M.\u00a0concilii* was the main species, probably contributed by the oxidation lagoon (OL) where it was the major component (18%). In addition, leachate (L) together with manure (M) presented most of the species of the genus *Methanoculleus* that were observed in the biodigester. Still, manure (M) was dominated by *M.\u00a0mazei* (11%), which was not relevant in the digester. Similarly, most rumen (R) species also represented by an abundance of *M.\u00a0mazei* (12%) and 10 species of *Methanobrevibacter* (9% *M. ruminantium*, 6% *M. olleyae*, 6% *Methanobrevibacter millerae*) were not observed in the biodigester. On the contrary, the oxidation lagoon (OL) presented most of the species during peak biogas production (18% *M.\u00a0concilii*, 11% *M.\u00a0tarda*, 6% *Methanolinea* sp. SDB) except for *Methanospirillum hungatei* (13%), which was in low abundance (2%--3%) only at the beginning of the run. Nevertheless, species\u2010level interpretation should be cautious due to the high amount of sequences (23%--47%) that could not be classified to known species (Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}).\n\n4. DISCUSSION {#mbo3854-sec-0013}\n=============\n\nMethanogens play a key role in biogas production and, therefore, have been the focus of many microbial community studies (Guo et al., [2015](#mbo3854-bib-0009){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}; Traversi, Villa, Lorenzi, Degan, & Gilli, [2012](#mbo3854-bib-0025){ref-type=\"ref\"}; Whitford et al., [2001](#mbo3854-bib-0028){ref-type=\"ref\"}; Zhou et al., [2010](#mbo3854-bib-0032){ref-type=\"ref\"}). Current high\u2010throughput sequencing technologies allow a deeper insight into the whole microbial community structure and functioning, shifting the attention on understanding complex interactions to optimize biogas yield (Guo et al., [2015](#mbo3854-bib-0009){ref-type=\"ref\"}; Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}; Sun et al., [2015](#mbo3854-bib-0023){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}; Yang et al., [2014](#mbo3854-bib-0030){ref-type=\"ref\"}). In this study, two approaches were used to unravel the microbial community structure of a 10\u2010ton anaerobic digester designed to treat bovine residues under mesophilic dry fermentation conditions. DGGE was used to target methanogens and a metagenomic approach was used to study the whole microbial community. As expected, DGGE failed to cover and resolve many species that were assessed by the metagenomic study. Multiple DGGE bands were assigned to the same methanogen species and some incongruities between techniques were detected probably due to DGGE identity assignment by band positioning (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}). For example, the genus *Methanobrevibacter*, which characterized the most abundant members of the feedstock rumen (R), was represented by three species of which *M. smithii* and *M. boviskoreani* showed multiple DGGE band patterns (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}, Band 5--7, 10, 11, 17). Conversely, the metagenomic analysis identified ten *Methanobrevibacter* species and strains suggesting that some of these bands might correspond to different genotypes, underestimating this genus diversity (Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}). Zhou et al. ([2010](#mbo3854-bib-0032){ref-type=\"ref\"}) found similar DGGE patterns for several methanogen genera. For example, five bands appeared to correspond to different strain sequence types of *Methanobrevibacter gottschalkii*. Other unrepresented member in the DGGE analysis was the genus *Methanoculleus* with only one species identified (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}, Band 24), while the metagenomic study recovered eight abundant species and strains (Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}). Nonetheless, we were able to elucidate with both methods the most abundant methanogen genera that characterized each microbial community. Also, both approaches were consistent in the role that uncultured or unclassified methanogens might play during anaerobic digestion. DGGE bands classified as uncultured methanogens were among the most intense bands during the biodigester run (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}, Band 4, 15, 18). Accordingly, the metagenome study showed a high proportion of the reads (23%--48%) that corresponded to unclassified methanogens at the species level (Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}). Thus, the DGGE analysis was relevant to qualitatively identify changes in the microbial community related to the uncultured component of the community.\n\nUnderstanding how to maintain a balance among the four microbial metabolic stages of biogas production (hydrolysis, acidogenesis, acetogenesis, and methanogenesis) is key to improve productivity (Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}). However, the microbial process of generating biogas cannot be generalized as it has been shown that microbial diversity and shifts in the community strongly depend on the type of substrate being treated and the reactor system (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Bengelsdorf et al., [2012](#mbo3854-bib-0004){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}; Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}). Nevertheless, extremely stable bacterial and methanogenic community profiles have also been reported (Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}; Kampmann et al., [2012](#mbo3854-bib-0011){ref-type=\"ref\"}; Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}), generally associated with higher taxonomic levels and attributed to functional redundancy among phylogenetic groups or being defined by a crucial process parameter such as high salt content (Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}).\n\nIn this study, fungi were dominant players among the eukaryote microbial community (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}a). Surprisingly, their abundance was almost the same in every sample but during biogas production represented a higher percentage of the community, which might suggest a key role during biogas generation (Table [1](#mbo3854-tbl-0001){ref-type=\"table\"}). Stable fungal presence in biogas plants has been reported (Bengelsdorf et al., [2012](#mbo3854-bib-0004){ref-type=\"ref\"}), but knowledge of their role remains unclear. It has been suggested that fungi assist in lignocellulose decomposition, penetrating the lignified material first for cellulolytic bacteria to gain access. Contrary to what was observed with Eukaryota, Bacteria composition considerably varied during biogas generation and among feedstocks (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}b). It has been shown that for treatment of\u00a0solid feedstocks, the community is usually dominated by Firmicutes (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}; Tuan et al., [2014](#mbo3854-bib-0026){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}). Our results showed that Firmicutes only dominated the rumen (R) feedstock and, during biogas production, Firmicutes was third in frequency and characterized by a stable presence throughout time. Kampmann et al. ([2012](#mbo3854-bib-0011){ref-type=\"ref\"}) also reported Firmicutes as a stable phylum during liquid manure treatment. Other studies associated with the digestion of liquid feedstocks as sludge have reported Proteobacteria, Firmicutes and Bacteroidetes as dominant bacterial phyla, followed by Actinobacteria and Chloroflexi (Guo et al., [2015](#mbo3854-bib-0009){ref-type=\"ref\"}; Yang et al., [2014](#mbo3854-bib-0030){ref-type=\"ref\"}). Spirochaetes has also been observed as an abundant phylum along with Bacteroidetes when treating sludge (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}). These six phyla collectively characterized the biodigester community that changed in abundance during the run. Each feedstock seemed to contribute a different bacterial group to the biodigester as a particular phylum dominated each residue: oxidation lagoon (OL) by Proteobacteria, manure (M) by Actinobacteria and rumen (R) by Firmicutes. During peak biogas production, Proteobacteria decreased in abundance opposite to Bacteroidetes/Chlorobi, while Firmicutes and Actinobacteria remained stable and Spirochaetes, a minor component, also increased (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}b). It seems that hydrolytic bacteria were present and active from the start of the run, first represented by members of the phyla Firmicutes and Actinobacteria, and then assisted by an increase of the Bacteroidetes/Chlorobi group. Firmicutes and Bacteroidetes members possess cellulose and hydrogenase activity (Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}), while Actinobacteria produce lignin\u2010degrading enzymes that break down complex organic materials (Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}; Yang et al., [2014](#mbo3854-bib-0030){ref-type=\"ref\"}). In addition, Firmicutes and Bacteroidetes participate in the fermentation of the generated products into organic acids, CO~2~ and H~2~ (Traversi et al., [2012](#mbo3854-bib-0025){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}). Also, Firmicutes can proceed with the consumption of butyrate and various VFAs (Ariesyady et al., [2007](#mbo3854-bib-0003){ref-type=\"ref\"}). These hydrolytic bacteria seemed to be assisted in glucose degradation by Spirochaetes and Proteobacteria known to consume glucose, propionate, butyrate, and acetate. The observed decrease of Proteobacteria in the biodigester, probably incorporated by the use of water from the oxidation lagoon, might be related to the capacity of other feedstocks to contribute microbial species already adapted to the treated substrates in the digester as rumen and manure, which were abundant in Firmicutes and Actinobacteria, respectively. Abundance of Chloroflexi and Proteobacteria has been correlated to low biogas yield while Firmicutes and Bacteroidetes characterized high biogas production (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}). In this study, the latter phyla were observed to remain stable or even increase in abundance during biogas peak production while the former decreased.\n\nMethanogenesis as the last crucial step in anaerobic digestion is where the stability of the process is more susceptible (Traversi et al., [2012](#mbo3854-bib-0025){ref-type=\"ref\"}). Our results showed a high representation of Archaea in the microbial community associated to the biodigester and leachate (8%--17%, Table [1](#mbo3854-tbl-0001){ref-type=\"table\"}). Other metagenomic studies of biogas plants have reported 6%--10% archaeal abundance (Guo et al., [2015](#mbo3854-bib-0009){ref-type=\"ref\"}; Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}). Methanogens were represented by the orders Methanosarcinales, Methanomicrobiales and Methanobacteriales (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}c) and their abundance varied accordingly to feedstock and the biodigester time course, as observed with bacteria. In the biodigester, the order Methanomicrobiales dominated, and was represented by several species of the genus *Methanoculleus* with the most abundant species being *M. marisnigri* followed by *M. horonobensis* (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}, Table [S3](#mbo3854-sup-0003){ref-type=\"supplementary-material\"}). *Methanoculleus* is frequently described as the dominant Archaea in biogas plants, particularly in the treatment of solid feedstocks and in reactors with inorganic supports or high total solid content (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}; Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}; Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}; Zhao et al., [2013](#mbo3854-bib-0031){ref-type=\"ref\"}). The presence of this methanogen has also been correlated to critical process parameters such as high concentrations of ammonia and salt (Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}). *Methanoculleus* is capable of forming biofilms increasing its capability to attach to solids and tolerate inhibitor substances and reactor disturbances (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}). This characteristic might explain the increase in abundance of this genus with time and entire dominance by the end of the biodigester run (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}). As methanogenesis pathways are well known for methanogen genera, it can be suggested that the main pathway in the biodigester was the hydrogenotrophic, as *Methanoculleus* is known to use H~2~ and CO~2~ to generate methane (Anderson et al., [2009](#mbo3854-bib-0002){ref-type=\"ref\"}). However, the second most abundant order Methanosarciales mainly represented by *M.\u00a0concilii* uses the acetotrophic route for methanogenesis. Conversely, because the genus *Methanoculleus* was represented by several relevant species and the third most abundant order Methanobacteriales is also a hydrogenotroph, this pathway dominates the studied biogas reactor. *Methanosaeta* is favored under low acetate conditions commonly found in sludge digesters (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Guo et al., [2015](#mbo3854-bib-0009){ref-type=\"ref\"}; Yang et al., [2014](#mbo3854-bib-0030){ref-type=\"ref\"}). In the studied biodigester, it seems as the importance of the acetoclastic pathway shifts with time towards the hydrogenotrophic. Presence of *Methanoculleus* has been correlated with high biogas yield, whereas *Methanosaeta* has been linked to low biogas output (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}). Our results are congruent with the characteristics of the biogas plant under study and previous literature reports of similar biodigesters (Stolze et al., [2015](#mbo3854-bib-0022){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}). The plant operates under dry fermentation conditions with a high content of total solids and treats bovine residues that are rich in ammonia and alkaline (Goberna et al., [2009](#mbo3854-bib-0008){ref-type=\"ref\"}; Nettmann et al., [2010](#mbo3854-bib-0015){ref-type=\"ref\"}; Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}). High ammonia concentrations might inhibit susceptible methanogens, while alkaline substances help stabilize the reactor pH. All these conditions appeared to contribute to the predominance of *Methanoculleus*, which potentially forms biofilms over the treated substrates in close proximity to acetate\u2010oxidizing bacteria allowing an adequate syntrophic relationship between H~2~ producers and consumers (Abendroth et al., [2015](#mbo3854-bib-0001){ref-type=\"ref\"}; Weiland, [2010](#mbo3854-bib-0027){ref-type=\"ref\"}; Wirth et al., [2012](#mbo3854-bib-0029){ref-type=\"ref\"}; Zhao et al., [2013](#mbo3854-bib-0031){ref-type=\"ref\"}). This would ensure an optimum H~2~ balance and an efficient operation of the biogas\u2010producing microbial community.\n\nConcerning the associate bovine residues, each feedstock possessed a characteristic methanogenic community. As expected, leachate (L) composition was similar to the biodigester but maintained equal proportions of the orders Methanomicrobiales and Methanosarcinales (Figure [3](#mbo3854-fig-0003){ref-type=\"fig\"}c), which changed in the reactor towards a dominance of Methanomicrobiales. The manure (M) exhibited a uniform community of the three main methanogen orders including the Methanomicrobiales represented by the genus *Methanoculleus* that dominated the biodigester (Figure [5](#mbo3854-fig-0005){ref-type=\"fig\"}). Also, it appears to contribute the uncultured or unclassified component of the biogas\u2010producing community as the DGGE analysis revealed their significance during peak biogas production (Figure [2](#mbo3854-fig-0002){ref-type=\"fig\"}, Band 4, 15, 18). Of all feedstocks, the oxidation lagoon (OL) was the most diverse and almost exclusively possessed the genera *Methanosaeta* (Methanosarcinales) and *Methanolinea* (Methanobacteriales), observed as relevant members in the digester. In accordance to the literature (Zhou et al., [2010](#mbo3854-bib-0032){ref-type=\"ref\"}), rumen (R) was almost entirely dominated by the order Methanobacteriales represented by 10 species of the genus *Methanobrevibacter*, the most abundant being *M. ruminantium*. However, these methanogens did not proliferate in the biodigester, neither the second most abundant species in this community, *M.\u00a0mazei*. This species was also a major component of manure (M), probably associated to the diet of the animals. Overall, feedstocks that significantly contribute to the biogas\u2010producing microbial community in the biodigester were leachate (L), as expected, and oxidation lagoon (OL) and manure (M) more than rumen (R).\n\nFinally, during anaerobic digestion of the bovine residues three distinct microbial community profiles of the bacterial and archaeal component were observed, with the exception of Eukarya mainly represented by a stable presence of fungi. These changes appeared to correlate to each stage of biogas production. At the beginning of the run (samples B17 and B22), during the first 6\u00a0days, an adaptation phase was observed where no biogas was produced. This period was followed by biogas production, in the middle of the run (B27, B29), and, as a final phase, the last days of the run (B04) when biogas production reached zero.\n\n5. CONCLUSION {#mbo3854-sec-0014}\n=============\n\nAt high taxonomic level, the two sets of information from DGGE and metagenomics correlated to some extent. Relevant methanogen genera as *Methanoculleus* and *Methanobrevibacter* were underestimated in the DGGE analysis. However, this technique was indispensable to discern the role that uncultured or unidentified methanogens played during biogas generation. The 10\u2010ton dry digester presented a diverse and dynamic community of bacteria and methanogens, which correlated to a particular stage during biogas production. These community profiles appeared to be supported by specific members that characterized each feedstock or residue. Water from the oxidation lagoon and manure were the most relevant substrates, while rumen methanogenic members did not proliferate in the reactor. It was confirmed that leachate, as the biodigester microbial inoculum, adequately preserved the biogas\u2010producing microbial community. Therefore, we were able to correlate presence of certain microorganisms in the biodigester to type of feedstock, which could lead to bioaugmentation strategies by incorporating a higher proportion or an enriched microbial inoculum from the most relevant feedstocks. Process adjustments would help reduce the adaptation phase in the digester and, consequently, decrease retention time and increase biogas yield if augmented microorganisms could further breakdown the organic waste.\n\nCONFLICT OF INTERESTS {#mbo3854-sec-0016}\n=====================\n\nThe authors declare no conflict of interest.\n\nAUTHOR CONTRIBUTIONS {#mbo3854-sec-0015}\n====================\n\nCSG designed, analyzed and wrote metagenomic analysis, FACC performed DGGE experiment, JFRL, BTP and JHSC provided the biological material, designed the study and revised the manuscript, and AP designed the study, analyzed data and wrote the manuscript. All authors read and approved the final manuscript.\n\nETHICS STATEMENT {#mbo3854-sec-0017}\n================\n\nNone required.\n\nSupporting information\n======================\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\n###### \n\n\u00a0\n\n###### \n\nClick here for additional data file.\n\nThis work was funded by Productos Bioorganicos and SuKarne Sustainability Division (Culiacan, Sinaloa, Mexico); Tecnologico de Monterrey Research Funding Program (GIEE EICIM01); and CONACYT Mexican National Council for Research and Technology,\u00a0Postdoctoral Scholarship (CSG No. 253929).\n\nDATA ACCESSIBILITY {#mbo3854-sec-0019}\n==================\n\nAll data are included in the manuscript. DGGE sequences were deposited in the National Center for Biotechnology Information (NCBI) under accession numbers MH393448--MH393458 and metagenomic database as BioProject PRJNA378243. Additional data is provided in Appendix [A](#mbo3854-sup-0001){ref-type=\"supplementary-material\"}, Supplementary Material.\n"}
+{"text": "During armed conflicts, international humanitarian law (which regulates the conduct of parties engaged in war) protects health-care workers and health facilities, the wounded and the sick. In the first half of 2016, however, the international medical charity M\u00e9decins Sans Fronti\u00e8res (MSF) reported several attacks on health facilities and workers in Afghanistan, the Central African Republic, South Sudan, the Syrian Arab Republic and Yemen.[@R1] These events have attracted media attention to a phenomenon of contemporary armed conflict that has important ramifications for the health, humanitarian, legal and security sectors.[@R2] In December 2015, the Stockholm Peace Research Institute and the Conflict and Health Research Group at King's College London convened a workshop in London on Eliminating violence against health workers: from theory to practice. Participants from MSF, the International Committee of the Red Cross (ICRC), Medical Aid for Palestinians and academic organizations discussed current trends in violence against health workers and attacks on health facilities, presented research findings and highlighted key debates and research gaps in evidence.\n\nSome important lessons can be drawn from ICRC's Health Care in Danger campaign, MSF's Medical Care Under Fire campaign, as well as other organizations such as Physicians for Human Rights, which has recently documented mass atrocities in the Syrian Arab Republic as well as the impact of the Syrian conflict on the health sector.[@R3]^--^[@R5] There is a perception of an increase in the number of health workers being killed and facilities being accidentally destroyed (so-called collateral damage) or deliberated targeted during armed conflicts. Comprehensive databases have been set up by independent research organizations to record major incidents of violence against aid workers, such as the Aid Worker Security Database of Humanitarian Outcomes and the Security in Numbers Database from Insecurity Insight.[@R6] However, even these do not currently provide health-specific data. The absence of baseline and routine data relating to attacks on health workers and health facilities makes it difficult to identify actual rising trends. Most of the available data sources do not capture violence on local health workers, who seem to bear the brunt of most attacks. Data disaggregated by sex are also lacking.[@R6]\n\nThese gaps in the evidence seem incongruous in an era of increasingly accessible and globalized data. Yet there are many factors that inhibit systematic data collection: poor or non-existent data collection by those in the field (for a variety of reasons ranging from security risks to insufficient research capacity); bias in data collection; insufficient research funding for the topic; and a lack of developed method. Some efforts have been made to monitor and study attacks (both quantitatively and qualitatively), particularly by the ICRC and MSF. However, multidisciplinary, collaborative, long-term retrospective and prospective studies are absent -- often for valid reasons. The mandates of some prominent nongovernmental organizations (NGOs) and international organizations may not allow the collection or sharing of such data. If they do allow it, then for perceived reasons of confidentiality and protection, they may not allow these data to be aggregated and analysed, even for meta-analysis by independent organizations. [Box\u00a01](#B1){ref-type=\"boxed-text\"} summarizes some of the research gaps and needs for better documentation of attacks on health workers and facilities in conflicts.\n\n###### Key needs for documenting attacks on health workers and health facilities in armed conflicts\n\n-   Analysis of trends of attacks on patients and health-care workers, facilities and transport during armed conflict and other violent incidents.\n\n-   Collection of systematic routine data, prospective and retrospective, which are disaggregated by sex.\n\n-   Examination of the context of each conflict to understand the dynamics and motives for attacks.\n\n-   Disaggregation of data on humanitarian databases to distinguish between types of aid workers, including local and international health-care workers.\n\n-   Public availability of anonymized data collected by humanitarian organizations to support a global response on prevention and accountability.\n\n-   Assessment of open threats and impact on health facilities and health-care personnel by security staff both before deployment and immediately after conflict.\n\n-   Systematic analysis of the immediate and longer-term impact of violence on the providers of health care.\n\nAttacks on health staff may be intentional or unintentional and can take a range of forms: road blockades and checkpoints that delay or block ambulances; attacks against medical personnel, suppliers and patients; direct targeting of hospitals; and armed entry into health facilities.[@R7] Political motives for violence and attacks directed at health workers and facilities across several conflict-affected settings include: gaining military advantage; displacing or inducing fear in a population; denying health care to enemies; and seeking health care for the attacker's own soldiers or affiliates.[@R7] Personal disputes between patients and staff, disagreements over treatment and the individual behaviour of both perpetrators and health workers who might resist threats are also among the drivers of violence. The use of violence against health-care facilities to induce fear in the local population has had a resurgence during the current conflicts in the Syrian Arab Republic (which began in 2011) and South Sudan (which escalated in 2013), but this tactic has also been used for example in conflicts in the Democratic Republic of Congo (which began in 1998) and Somalia (from 1988).[@R3] Socioeconomic factors, such as poverty and inequality within local populations, as well as inequalities between health workers and local populations, can also be factors in violence against health workers and facilities, because these are viewed as easy targets for looting valuable medicines and equipment. These factors further complicate the task of collecting a robust evidence base.\n\nOverall, one of the most important lessons from research so far is that we must look closely at the context of each attack, to understand its dynamics and motives and to determine responses to attacks and solutions for prevention.[@R6] We acknowledge that conducting research on violence towards health workers and facilities during conflict is challenging, particularly in areas where terrorism-related violence is endemic. Humanitarian organizations have a strong operational and long-established local presence, while academic institutions have the capacity to conduct scientifically rigorous, multidisciplinary, quality-assessed and independently corroborated research. Strong research links between these two parties will foster a continuum of knowledge, evidence and practice. This could enable us to develop evidence-based, context-specific guidelines for more effective protection of health workers and facilities during armed conflict. A useful policy lesson comes from the systematic documentation of collateral damage resulting from anti-personnel mines,[@R8] whereby the data supported a reduction in the use of such weapons. This strategy could be a model for the health sector to make an effective case to those engaged in conflict (whether national armies or non-state armed groups) to exclude health workers, facilities and patients as deliberate targets. Unless attacks are systematically documented, there will continue to be an important gap in knowledge on the extent and severity of the damage to health-care systems in armed conflict.\n\nWe recognize that in many tactical situations, distinguishing between targeted and unintended attacks on health workers and facilities can be problematic. State and multinational armed forces must be encouraged to conduct thorough threat assessments before deployment of forces and during conflicts to prevent and mitigate unintended damage to health workers and facilities. Many armed forces have the technological and intelligence capacities to ensure that, even in fast-changing, unstable situations, tactical awareness of health facilities is possible. Furthermore, after conflicts have ended there should be a requirement for the collection and open distribution of any post-conflict threats of violence against health workers and facilities. This could be done for example through the leadership of the United Nations (UN) Office for the Coordination of Humanitarian Affairs.\n\nThere are strong global governance efforts focusing on health care in conflict. These include the 2016 UN Security Council Resolution (2286) on health care in armed conflict, the 2014 UN General Assembly Resolution (A/69/L.35) on global health and foreign policy, focusing on the protection of health workers, and the World Health Assembly's 2012 Resolution (WHA65.20) calling for leadership from the World Health Organization (WHO) to collect and disseminate data on attacks on health care in complex humanitarian emergencies.[@R9] A new system for collecting such data has been developed by WHO and is being tested in the Central African Republic, the Syrian Arab Republic and the West Bank and Gaza Strip. WHO is also field-testing tools to gather data on attacks, and establishing a repository for reports from governments, media and civil society organizations, which was due to be available for use in 2016.[@R6]^,^[@R10] At its 32nd international conference in 2015, the International Red Cross and Red Crescent Movement renewed its commitment to the Geneva Conventions and addressed attacks on health-care personnel and facilities.[@R11] These efforts need to be supported with evidence-based research across a variety of conflict-affected contexts to support the important initiatives by humanitarian and human rights NGOs, as well as local emergency responses. Academic institutions have a key part to play in supporting these efforts to improve the evidence base on this pressing humanitarian challenge.[@R12]\n\nWe thank the British International Studies Association Global Health Working Group for funding the workshop on Eliminating violence against health workers.\n\nNone declared.\n"}
+{"text": "Related literature \u00a0 {#sec1}\n====================\n\nFor the synthesis and structures of related tin and titanium complexes, see: Boyle *et al.* (2002[@bb1]); Eslava *et al.* (2010[@bb3]); Fric & Schubert (2008[@bb5]); Harrison *et al.* (1978[@bb6]); Mijatovic *et al.* (2001[@bb7]); Mokal *et al.* (1994[@bb8]); Vatsa *et al.* (1991[@bb12]); Verdenelli *et al.* (2000[@bb13]).\n\nExperimental \u00a0 {#sec2}\n==============\n\n {#sec2.1}\n\n### Crystal data \u00a0 {#sec2.1.1}\n\n\\[Sn~4~(C~3~H~7~O)~4~Cl~4~O~2~\\]*M* *~r~* = 884.98Monoclinic,*a* = 6.4423 (3) \u00c5*b* = 17.8302 (7) \u00c5*c* = 11.6843 (5) \u00c5\u03b2 = 105.474 (2)\u00b0*V* = 1293.5 (1) \u00c5^3^*Z* = 2Mo *K*\u03b1 radiation\u03bc = 4.25 mm^\u22121^*T* = 296 K0.36 \u00d7 0.17 \u00d7 0.12 mm\n\n### Data collection \u00a0 {#sec2.1.2}\n\nNonius KappaCCD diffractometerAbsorption correction: multi-scan (*DENZO*/*SCALEPACK*; Otwinowski & Minor, 1997[@bb10]) *T* ~min~ = 0.310, *T* ~max~ = 0.62914221 measured reflections3853 independent reflections2930 reflections with *I* \\> 2\u03c3(*I*)*R* ~int~ = 0.041\n\n### Refinement \u00a0 {#sec2.1.3}\n\n*R*\\[*F* ^2^ \\> 2\u03c3(*F* ^2^)\\] = 0.038*wR*(*F* ^2^) = 0.061*S* = 1.103853 reflections118 parametersH-atom parameters constrained\u0394\u03c1~max~ = 0.93 e \u00c5^\u22123^\u0394\u03c1~min~ = \u22120.80 e \u00c5^\u22123^\n\n {#d5e500}\n\nData collection: *COLLECT* (Nonius, 2000[@bb9]); cell refinement: *DENZO*/*SCALEPACK* (Otwinowski & Minor, 1997[@bb10]); data reduction: *DENZO*/*SCALEPACK*; program(s) used to solve structure: *SIR2004* (Burla *et al.*, 2005[@bb2]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb11]); molecular graphics: *ORTEP-3 for Windows* (Farrugia, 2012[@bb4]); software used to prepare material for publication: *SHELXL97*.\n\nSupplementary Material\n======================\n\nCrystal structure: contains datablock(s) I, global. DOI: [10.1107/S1600536814000816/hy2641sup1.cif](http://dx.doi.org/10.1107/S1600536814000816/hy2641sup1.cif)\n\nStructure factors: contains datablock(s) I. DOI: [10.1107/S1600536814000816/hy2641Isup2.hkl](http://dx.doi.org/10.1107/S1600536814000816/hy2641Isup2.hkl)\n\nCCDC reference: <http://scripts.iucr.org/cgi-bin/cr.cgi?rm=csd&csdid=981323>\n\nAdditional supporting information: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?hy2641&file=hy2641sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?hy2641sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?hy2641&checkcif=yes)\n\nSupporting information for this paper is available from the IUCr electronic archives (Reference: [HY2641](http://scripts.iucr.org/cgi-bin/sendsup?hy2641)).\n\nThe authors thank the Ministry of Education and Science of Ukraine for financial support (grant No. F28/241--2009).\n\n1. Comment\n==========\n\nCompound with the same symmetry and similar coordination environment as in the title compound was studied in the work of Boyle *et al.* (2002). However it has tin atoms with equal oxidation state, in opposite with our compound, which has different oxidation states and coordination environments of metal atoms. The same situation persists with other compounds of tin and titanium with similar coordination geometry and identical oxidation states of the metal atoms (Eslava *et al.*, 2010; Fric & Schubert, 2008; Harrison *et al.*, 1978; Mijatovic *et al.*, 2001; Mokal *et al.*, 1994; Vatsa *et al.*, 1991; Verdenelli *et al.*, 2000). The title compound has two types of tin atoms. The Sn^II^ atom is three-coordinated and has a trigonal-pyramidal environment. The Sn^IV^ atom is hexa-coordinated and has an octahedral environment. Literature search did not provide any information about similar compounds, which have simultaneously two tin atoms with different oxidation states. In the title compound this is possible due to two chloride atoms for each Sn^IV^ atom.\n\n2. Experimental {#experimental}\n===============\n\nTo a solution of (diisopropoxydichlorido)tin (1.54 g, 5 mmol) in 5 ml of toluene was added N,N,N-tris(trimethylsilyl)aminoiminophosphorane (0.695 g, 2.5 mmol) in 3 ml of toluene and stirred for 4 h. The resulting solution was concentrated to 4 ml and cooled to -25\u00b0C. After two days, resulting crystals were filtered from the solution and dried in vacuum (yield: 0.454 g, 41%). Analysis, calculated for C~12~H~28~Cl~4~O~6~Sn~4~: C 16.26, H 3.16, Cl 16.03, O 10.83, Sn 53.72%; found: C 16.01, H 3.56, Cl 16.25, O 10.94, Sn 53.24%.\n\n3. Refinement {#refinement}\n=============\n\nH atoms were positioned geometrically and refined as riding atoms, with C---H = 0.98 (CH~2~) and 0.96 (CH~3~) \u00c5 and with *U*~iso~(H) = 1.2(1.5 for methyl)*U*~eq~(C).\n\nFigures\n=======\n\n![The molecular structure of the title compound, with displacement ellipsoids drawn at the 50% probability level. H atoms have been omitted. \\[Symmetry code: (i) 1-x, 1-y, 1-z.\\]](e-70-00m45-fig1){#Fap1}\n\nCrystal data {#tablewrapcrystaldatalong}\n============\n\n  ---------------------------------- ---------------------------------------\n  \\[Sn~4~(C~3~H~7~O)~4~Cl~4~O~2~\\]   *F*(000) = 832\n  *M~r~* = 884.98                    *D*~x~ = 2.272 Mg m^\u22123^\n  Monoclinic, *P*2~1~/*c*            Mo *K*\u03b1 radiation, \u03bb = 0.71073 \u00c5\n  Hall symbol: -P 2ybc               Cell parameters from 3853 reflections\n  *a* = 6.4423 (3) \u00c5                 \u03b8 = 2.1--31.1\u00b0\n  *b* = 17.8302 (7) \u00c5                \u00b5 = 4.25 mm^\u22121^\n  *c* = 11.6843 (5) \u00c5                *T* = 296 K\n  \u03b2 = 105.474 (2)\u00b0                   Block, colorless\n  *V* = 1293.5 (1) \u00c5^3^              0.36 \u00d7 0.17 \u00d7 0.12 mm\n  *Z* = 2                            \n  ---------------------------------- ---------------------------------------\n\nData collection {#tablewrapdatacollectionlong}\n===============\n\n  ----------------------------------------------------------------------------------- --------------------------------------\n  Nonius KappaCCD diffractometer                                                      3853 independent reflections\n  Radiation source: fine-focus sealed tube                                            2930 reflections with *I* \\> 2\u03c3(*I*)\n  Graphite monochromator                                                              *R*~int~ = 0.041\n  \u03c6 and \u03c9 scans with \u03ba offset                                                         \u03b8~max~ = 31.1\u00b0, \u03b8~min~ = 2.1\u00b0\n  Absorption correction: multi-scan (*DENZO*/*SCALEPACK*; Otwinowski & Minor, 1997)   *h* = \u22129\u21929\n  *T*~min~ = 0.310, *T*~max~ = 0.629                                                  *k* = \u221225\u219221\n  14221 measured reflections                                                          *l* = \u221216\u219215\n  ----------------------------------------------------------------------------------- --------------------------------------\n\nRefinement {#tablewraprefinementdatalong}\n==========\n\n  ------------------------------------- ---------------------------------------------------------------------------------------------\n  Refinement on *F*^2^                  Primary atom site location: structure-invariant direct methods\n  Least-squares matrix: full            Secondary atom site location: difference Fourier map\n  *R*\\[*F*^2^ \\> 2\u03c3(*F*^2^)\\] = 0.038   Hydrogen site location: inferred from neighbouring sites\n  *wR*(*F*^2^) = 0.061                  H-atom parameters constrained\n  *S* = 1.10                            *w* = 1/\\[\u03c3^2^(*F*~o~^2^) + (0.*P*)^2^ + 2.1387*P*\\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3\n  3853 reflections                      (\u0394/\u03c3)~max~ = 0.001\n  118 parameters                        \u0394\u03c1~max~ = 0.93 e \u00c5^\u22123^\n  0 restraints                          \u0394\u03c1~min~ = \u22120.80 e \u00c5^\u22123^\n  ------------------------------------- ---------------------------------------------------------------------------------------------\n\nSpecial details {#specialdetails}\n===============\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.\n  Refinement. Refinement of F^2^ against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F^2^, conventional R-factors R are based on F, with F set to zero for negative F^2^. The threshold expression of F^2^ \\> 2sigma(F^2^) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F^2^ are statistically about twice as large as those based on F, and R- factors based on ALL data will be even larger.\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nFractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (\u00c5^2^) {#tablewrapcoords}\n==================================================================================================\n\n  ----- ------------- --------------- -------------- -------------------- --\n        *x*           *y*             *z*            *U*~iso~\\*/*U*~eq~   \n  Sn1   0.47306 (5)   0.589199 (15)   0.51770 (2)    0.02814 (8)          \n  Sn2   0.07243 (5)   0.494763 (16)   0.33749 (3)    0.03368 (8)          \n  Cl1   0.1984 (2)    0.65257 (7)     0.58158 (12)   0.0527 (3)           \n  Cl2   0.6589 (2)    0.69913 (7)     0.49108 (12)   0.0538 (3)           \n  O1    0.6730 (5)    0.57626 (15)    0.6906 (2)     0.0358 (7)           \n  O2    0.2858 (5)    0.58942 (15)    0.3422 (2)     0.0347 (7)           \n  O3    0.2987 (4)    0.48801 (13)    0.5045 (2)     0.0271 (6)           \n  C1    0.6845 (8)    0.6280 (3)      0.7884 (4)     0.0430 (11)          \n  H1    0.6029        0.6732          0.7566         0.052\\*              \n  C2    0.9156 (9)    0.6503 (3)      0.8431 (5)     0.0674 (17)          \n  H2A   0.9729        0.6733          0.7838         0.101\\*              \n  H2B   0.9987        0.6065          0.8737         0.101\\*              \n  H2C   0.9223        0.6852          0.9066         0.101\\*              \n  C3    0.5808 (11)   0.5926 (4)      0.8751 (5)     0.083 (2)            \n  H3A   0.4334        0.5810          0.8360         0.124\\*              \n  H3B   0.5860        0.6268          0.9393         0.124\\*              \n  H3C   0.6562        0.5474          0.9056         0.124\\*              \n  C4    0.2592 (8)    0.6511 (2)      0.2583 (4)     0.0423 (11)          \n  H4    0.3093        0.6973          0.3024         0.051\\*              \n  C5    0.3896 (12)   0.6379 (4)      0.1741 (6)     0.100 (3)            \n  H5A   0.5384        0.6332          0.2170         0.150\\*              \n  H5B   0.3728        0.6794          0.1199         0.150\\*              \n  H5C   0.3425        0.5926          0.1304         0.150\\*              \n  C6    0.0244 (10)   0.6592 (3)      0.1970 (6)     0.086 (2)            \n  H6A   \u22120.0541       0.6678          0.2549         0.130\\*              \n  H6B   \u22120.0269       0.6142          0.1536         0.130\\*              \n  H6C   0.0039        0.7009          0.1431         0.130\\*              \n  ----- ------------- --------------- -------------- -------------------- --\n\nAtomic displacement parameters (\u00c5^2^) {#tablewrapadps}\n=====================================\n\n  ----- -------------- -------------- -------------- --------------- -------------- ---------------\n        *U*^11^        *U*^22^        *U*^33^        *U*^12^         *U*^13^        *U*^23^\n  Sn1   0.02865 (15)   0.02228 (14)   0.03221 (15)   0.00089 (12)    0.00591 (11)   \u22120.00050 (12)\n  Sn2   0.03055 (16)   0.03210 (16)   0.03533 (16)   \u22120.00024 (13)   0.00347 (12)   0.00033 (13)\n  Cl1   0.0432 (7)     0.0410 (7)     0.0770 (9)     0.0083 (5)      0.0215 (6)     \u22120.0119 (6)\n  Cl2   0.0424 (7)     0.0356 (6)     0.0738 (9)     \u22120.0129 (5)     \u22120.0009 (6)    0.0116 (6)\n  O1    0.0380 (17)    0.0359 (17)    0.0309 (15)    0.0064 (13)     0.0044 (13)    \u22120.0096 (13)\n  O2    0.0364 (17)    0.0268 (15)    0.0370 (15)    \u22120.0058 (13)    0.0032 (13)    0.0050 (13)\n  O3    0.0297 (15)    0.0200 (14)    0.0302 (14)    0.0016 (11)     0.0058 (11)    \u22120.0004 (11)\n  C1    0.053 (3)      0.040 (3)      0.033 (2)      0.007 (2)       0.005 (2)      \u22120.014 (2)\n  C2    0.062 (4)      0.076 (4)      0.060 (3)      \u22120.013 (3)      0.011 (3)      \u22120.035 (3)\n  C3    0.103 (5)      0.098 (5)      0.058 (3)      \u22120.031 (4)      0.042 (4)      \u22120.035 (3)\n  C4    0.049 (3)      0.031 (2)      0.042 (3)      \u22120.003 (2)      0.003 (2)      0.014 (2)\n  C5    0.137 (7)      0.088 (5)      0.102 (5)      0.023 (5)       0.078 (5)      0.041 (4)\n  C6    0.079 (5)      0.064 (4)      0.093 (5)      0.004 (3)       \u22120.017 (4)     0.042 (4)\n  ----- -------------- -------------- -------------- --------------- -------------- ---------------\n\nGeometric parameters (\u00c5, \u00ba) {#tablewrapgeomlong}\n===========================\n\n  ---------------------------- -------------- --------------------------- --------------\n  Sn1---O1                     2.099 (3)      C2---H2A                    0.9600\n  Sn1---O2                     2.084 (3)      C2---H2B                    0.9600\n  Sn1---O3                     2.109 (2)      C2---H2C                    0.9600\n  Sn1---O3^i^                  2.081 (3)      C3---H3A                    0.9600\n  Sn1---Cl1                    2.3808 (12)    C3---H3B                    0.9600\n  Sn1---Cl2                    2.3604 (11)    C3---H3C                    0.9600\n  Sn1---Sn1^i^                 3.2382 (5)     C4---C5                     1.473 (7)\n  Sn2---O1^i^                  2.165 (3)      C4---C6                     1.498 (7)\n  Sn2---O2                     2.168 (3)      C4---H4                     0.9800\n  Sn2---O3                     2.105 (2)      C5---H5A                    0.9600\n  O1---C1                      1.456 (5)      C5---H5B                    0.9600\n  O2---C4                      1.453 (5)      C5---H5C                    0.9600\n  C1---C3                      1.494 (7)      C6---H6A                    0.9600\n  C1---C2                      1.508 (7)      C6---H6B                    0.9600\n  C1---H1                      0.9800         C6---H6C                    0.9600\n                                                                          \n  O3^i^---Sn1---O2             96.97 (10)     C3---C1---C2                113.2 (4)\n  O3^i^---Sn1---O1             76.99 (10)     O1---C1---H1                108.3\n  O2---Sn1---O1                173.28 (10)    C3---C1---H1                108.3\n  O3^i^---Sn1---O3             78.77 (11)     C2---C1---H1                108.3\n  O2---Sn1---O3                76.86 (10)     C1---C2---H2A               109.5\n  O1---Sn1---O3                98.87 (10)     C1---C2---H2B               109.5\n  O3^i^---Sn1---Cl2            97.63 (8)      H2A---C2---H2B              109.5\n  O2---Sn1---Cl2               92.84 (8)      C1---C2---H2C               109.5\n  O1---Sn1---Cl2               90.89 (9)      H2A---C2---H2C              109.5\n  O3---Sn1---Cl2               168.46 (7)     H2B---C2---H2C              109.5\n  O3^i^---Sn1---Cl1            164.20 (8)     C1---C3---H3A               109.5\n  O2---Sn1---Cl1               91.38 (9)      C1---C3---H3B               109.5\n  O1---Sn1---Cl1               93.83 (8)      H3A---C3---H3B              109.5\n  O3---Sn1---Cl1               90.17 (8)      C1---C3---H3C               109.5\n  Cl2---Sn1---Cl1              95.32 (5)      H3A---C3---H3C              109.5\n  O3^i^---Sn1---Sn1^i^         39.70 (7)      H3B---C3---H3C              109.5\n  O2---Sn1---Sn1^i^            85.98 (7)      O2---C4---C5                110.1 (4)\n  O1---Sn1---Sn1^i^            87.47 (7)      O2---C4---C6                108.6 (4)\n  O3---Sn1---Sn1^i^            39.06 (7)      C5---C4---C6                112.3 (5)\n  Cl2---Sn1---Sn1^i^           136.34 (4)     O2---C4---H4                108.6\n  Cl1---Sn1---Sn1^i^           128.33 (3)     C5---C4---H4                108.6\n  O3---Sn2---O1^i^             75.04 (10)     C6---C4---H4                108.6\n  O3---Sn2---O2                75.14 (10)     C4---C5---H5A               109.5\n  O1^i^---Sn2---O2             87.61 (11)     C4---C5---H5B               109.5\n  C1---O1---Sn1                125.2 (2)      H5A---C5---H5B              109.5\n  C1---O1---Sn2^i^             127.4 (3)      C4---C5---H5C               109.5\n  Sn1---O1---Sn2^i^            102.33 (11)    H5A---C5---H5C              109.5\n  C4---O2---Sn1                126.8 (2)      H5B---C5---H5C              109.5\n  C4---O2---Sn2                127.8 (2)      C4---C6---H6A               109.5\n  Sn1---O2---Sn2               102.71 (11)    C4---C6---H6B               109.5\n  Sn1^i^---O3---Sn2            105.05 (11)    H6A---C6---H6B              109.5\n  Sn1^i^---O3---Sn1            101.23 (11)    C4---C6---H6C               109.5\n  Sn2---O3---Sn1               104.02 (10)    H6A---C6---H6C              109.5\n  O1---C1---C3                 109.0 (4)      H6B---C6---H6C              109.5\n  O1---C1---C2                 109.7 (4)                                  \n                                                                          \n  O3^i^---Sn1---O1---C1        162.0 (3)      O1^i^---Sn2---O3---Sn1^i^   \u22125.91 (11)\n  O3---Sn1---O1---C1           \u2212121.8 (3)     O2---Sn2---O3---Sn1^i^      \u221297.42 (13)\n  Cl2---Sn1---O1---C1          64.4 (3)       O1^i^---Sn2---O3---Sn1      100.06 (12)\n  Cl1---Sn1---O1---C1          \u221231.0 (3)      O2---Sn2---O3---Sn1         8.55 (10)\n  Sn1^i^---Sn1---O1---C1       \u2212159.3 (3)     O3^i^---Sn1---O3---Sn1^i^   0.0\n  O3^i^---Sn1---O1---Sn2^i^    5.81 (11)      O2---Sn1---O3---Sn1^i^      99.98 (12)\n  O3---Sn1---O1---Sn2^i^       82.02 (12)     O1---Sn1---O3---Sn1^i^      \u221274.73 (11)\n  Cl2---Sn1---O1---Sn2^i^      \u221291.81 (10)    Cl2---Sn1---O3---Sn1^i^     72.8 (4)\n  Cl1---Sn1---O1---Sn2^i^      172.79 (10)    Cl1---Sn1---O3---Sn1^i^     \u2212168.64 (9)\n  Sn1^i^---Sn1---O1---Sn2^i^   44.54 (9)      O3^i^---Sn1---O3---Sn2      \u2212108.82 (14)\n  O3^i^---Sn1---O2---C4        \u2212112.5 (3)     O2---Sn1---O3---Sn2         \u22128.83 (11)\n  O3---Sn1---O2---C4           170.8 (3)      O1---Sn1---O3---Sn2         176.45 (11)\n  Cl2---Sn1---O2---C4          \u221214.4 (3)      Cl2---Sn1---O3---Sn2        \u221236.0 (4)\n  Cl1---Sn1---O2---C4          81.0 (3)       Cl1---Sn1---O3---Sn2        82.55 (10)\n  Sn1^i^---Sn1---O2---C4       \u2212150.7 (3)     Sn1^i^---Sn1---O3---Sn2     \u2212108.82 (14)\n  O3^i^---Sn1---O2---Sn2       85.22 (12)     Sn1---O1---C1---C3          108.1 (4)\n  O3---Sn1---O2---Sn2          8.53 (10)      Sn2^i^---O1---C1---C3       \u2212101.7 (4)\n  Cl2---Sn1---O2---Sn2         \u2212176.73 (10)   Sn1---O1---C1---C2          \u2212127.4 (4)\n  Cl1---Sn1---O2---Sn2         \u221281.33 (10)    Sn2^i^---O1---C1---C2       22.8 (5)\n  Sn1^i^---Sn1---O2---Sn2      47.00 (9)      Sn1---O2---C4---C5          102.9 (5)\n  O3---Sn2---O2---C4           \u2212170.6 (3)     Sn2---O2---C4---C5          \u221299.2 (5)\n  O1^i^---Sn2---O2---C4        114.2 (3)      Sn1---O2---C4---C6          \u2212133.8 (4)\n  O3---Sn2---O2---Sn1          \u22128.61 (10)     Sn2---O2---C4---C6          24.1 (5)\n  O1^i^---Sn2---O2---Sn1       \u221283.76 (12)                                \n  ---------------------------- -------------- --------------------------- --------------\n\nSymmetry code: (i) \u2212*x*+1, \u2212*y*+1, \u2212*z*+1.\n\n###### Selected bond lengths (\u00c5)\n\n  ------------- -------------\n  Sn1---O1      2.099\u2005(3)\n  Sn1---O2      2.084\u2005(3)\n  Sn1---O3      2.109\u2005(2)\n  Sn1---O3^i^   2.081\u2005(3)\n  Sn1---Cl1     2.3808\u2005(12)\n  Sn1---Cl2     2.3604\u2005(11)\n  Sn2---O1^i^   2.165\u2005(3)\n  Sn2---O2      2.168\u2005(3)\n  Sn2---O3      2.105\u2005(2)\n  ------------- -------------\n\nSymmetry code: (i) .\n"}
+{"text": "Correction to: *Scientific Reports* 10.1038/s41598-018-30559-7, published online 13 August 2018\n\nThis Article contains an error in the Methods section.\n\n\"(2) an abnormal CSF (i.e. reactive VDRL and/or greater than 4 leukocytes/mm^3^ and protein concentration less than 40\u2009mg/dl);\"\n\nshould read:\n\n\"(2) an abnormal CSF (i.e reactive VDRL and/or greater than 4 leukocytes/mm^3^ and protein concentration greater than 40\u2009mg/dl);\"\n"}
+{"text": "Introduction {#S0001}\n============\n\nRecent scientific advances have provided the possibility of fertility for infertile couples. Scientists and parents of the children born by such interventions are always faced with the concern that how their child\\'s visual health will be. Unfortunately, despite extensive researches in the field of assisted reproductive technology (ART), particularly in Iran, comprehensive studies on ophthalmic health and integrity of visual reflexes in such infants have not been done yet. With regard to the fact that these infants are born after several years of treatment and spending high costs, assessment and monitoring of ART infants is necessary.\n\nDevelopment of the visual system begins during embryonic period and continues until after birth ([@CIT0001]). Correct and complete development of the human visual system is affected by several intrinsic and extrinsic factors ([@CIT0002]). Infants born by ART may be susceptible to several factors with undesirable effects on development of the visual system ([@CIT0003]). On the other hand environmental conditions, before and after birth, strongly influence infant\\'s visual system ([@CIT0004]). Hence, being aware of their visual system condition can help the professionals to choose the proper technique for assisted reproduction technology and the effective care for it. Furthermore, a child\\'s first 12 month of life is considered as the critical period ([@CIT0005]); although, some studies have stated that the development of the visual reflexes may continue after age three ([@CIT0006]). Appropriate interventions in this period of time can reduce the influence of interruptive factors affecting the visual system development ([@CIT0007]). Therefore, as a rule in visual sciences, all infants experiencing hazardous conditions during prenatal period and those with unfavorable hereditary background must undergo various eye exams during infancy ([@CIT0008]). Accordingly, several screening programs are designated for this period in many countries ([@CIT0009]--[@CIT0011]). The aim of this study is to describe the impact of assisted reproductive technology (ART) on ocular and visual performances of infants born by these methods.\n\nMaterials and methods {#S0002}\n=====================\n\nThis cross-sectional descriptive study was accomplished at Child Health and Development Research Department (CHDRD) of Academic Center for Education, Culture and Research (ACECR), Tehran, Iran. Ethical considerations of this study were approved by Ethical Committee of Royan Research Institute and ACECR. We used non-random sequential sampling, and recorded history and the information about birth condition of all participants. Then, a representative was giving detailed explanations to the couples admitted to Royan Institute for the ophthalmic examination of their child. The inclusion criteria were infants conceived through one kind of ART methods (IVF, ICSI) and residence in Tehran, if infants did not come for the second examination, they were excluded from study. Infants were brought to the center from the age of one month, but clinical ophthalmic assessment was accomplished after the age of three months. Infants were re-examined to ensure the accuracy of the initial findings at age of nine months.\n\nTheir refractive status was evaluated using dry refraction by streak static retinoscopy. As the time consuming nature of cyclorefraction, and also, lack of permission from most parents for using cycloplegic agents, cyclorefraction did not performed.\n\nFurthermore, red reflex assessment, Hirschberg test, qualitative fixation evaluation, and external eye examination were performed. Family profile of the infants was also assessed. Since it was impossible to perform all tests on all infants due to crying, sleeping or their parents' impatience, a total of 320 infants were only evaluated. Infants who needed further assessments and interventions were referred to more specialized centers. Results were analyzed by using SPSS 16 software.\n\nResults {#S0003}\n=======\n\nEvaluation of refractive status was possible in 320 infants. Statistical comparison of the obtained results of refractive error between their right and left eyes showed no significant differences. [Table 1](#T0001){ref-type=\"table\"} demonstrates the results of distribution of ART infants regarding perinatal variable and reproductive techniques.\n\n###### \n\nDistribution of ART infants with consideration of prenatal variables and reproductive techniques\n\n  Prenatal variables                ART   ICSI         IVF         Total number\n  --------------------------------- ----- ------------ ----------- --------------\n  Total exam infants                      245(76.5%)   75(23.5%)   320(100%)\n  Preterm                                 75(23.5%)    26(8%)      101(31.6%)\n  Term                                    170(53.1%)   49(15.3%)   219(68.4%)\n  Low birth weigh                         91(28.4%)    20(6.2%)    111(34.6%)\n  Appropriate for gestational age         154(48.1%)   55(17.2%)   209(65.3%)\n\nIVF: In vitro fertilization\n\nICSI: Intra cytoplasmic sperm injection\n\nAmong participants, 101 (31.6%) infants and 219 (68.4%) infants were preterm and term, respectively ([Table 1](#T0001){ref-type=\"table\"}).\n\nThe findings confirmed 20.3% of the infants had poor fixation, and 2.9% showed manifest strabismus in Hirschberg test. The results also revealed the prevalences of myopia, hyperopia and emmetropia are 2.9%, 87%, and 10.1%, respectively ([Table 2](#T0002){ref-type=\"table\"}). External examination only showed 1% abnormality in infants' eyes. These findings did not show any significant difference between the sexes. But, the result of red reflex testing of participants confirmed abnormality in 17 male (5.3%) compared to 2 female (0.6%) (p = 0.003), indicating the involvement of male more than female. This abnormality was present significantly in preterm infants (15 cases, 4.7%) than term infants (4 cases, 1.2%) (p = 0.006). There were 60 preterm infants (18.8%) and 5 term infants (1.5%) with fixation deficiency, which indicates there were significantly more fixation deficient infants among preterm (p \\< 0.001). A significant correlation was found between refractive error and poor fixation, so failure of fixation control was observed more frequently with increasing refractive error (p \\< 0.001).\n\n###### \n\nComparison the relations between the descriptive results of red reflex, fixation deficiency and refractive error and sex and maturity of ART infants\n\n  Eye exam                                   Maturity & Sex   Sex          Maturity     Total                     \n  ------------------------------------------ ---------------- ------------ ------------ ------------ ------------ ------------\n  Total                                                       157(49%)     163(51%)     101(31.6%)   219(68.4%)   320 (100%)\n  Abnormal red reflex                                         17 (5.3%)    2 (0.6%)     15 (4.7%)    4 (1.2%)     19(5.9%)\n  Normal red reflex                                           140(43.7%)   161(50.4%)   86(26.9%)    215(67.2%)   301(94.1%)\n  Fixation deficient                                          37(11.5%)    28 (8.8%)    60 (18.8%)   5 (1.5%)     65(20.3%)\n  Fixation normal                                             120(37.5%)   135(42.2%)   41(12.8%)    214(66.9%)   255(79.7%)\n  Refractive error (myopia)                                   5 (1.5%)     4 (1.3%)     7 (2.2%)     2 (0.7%)     9(2.9%)\n  Refractive normal (hyperopia,emmetropia)                    152(47.5%)   159(49.7%)   94(29.3%)    217(67.8%)   311(97.1%)\n  Astigmatism                                                 58(47.1%)    65(52.8%)    82(66.6%)    41(33.3%)    123(38.6%)\n\nDiscussion {#S0004}\n==========\n\nOur results reflect the necessity of more comprehensive assessments and further follow-up of infants born by ART, especially for premature male ART infants. These results also suggest the probability of fixation condition and visual deficiencies in these infants. Many studies have demonstrated that low birth weight and/or premature neonates have higher refractive errors than full term neonates ([@CIT0012]--[@CIT0014]). Environmental influences besides hereditary factors may affect birth weight ([@CIT0003], [@CIT0004], [@CIT0015], [@CIT0016]). So, infants born by ART may have special genetic background and environmental conditions, which make them different than naturally conceived infants ([@CIT0003], [@CIT0004], [@CIT0015], [@CIT0016]). But, this study showed most of the ART-conceived neonates were not underweight and preterm ([Table 1](#T0001){ref-type=\"table\"}). Our finding confirmed that prevalence and severity of refractive errors in ART infants are caused by the factors involving in ART. These factors can influence ART infants in fetal period are laboratory environment, temperature, chemical materials, drugs for ovulation induction and drugs for continuity of pregnancy. However, the assisted reproductive technologies have been introduced only a few years ago, and there are not enough scientific reports regarding visual health of these infants. One of the most important studies in this field belongs to the Anteby et al. (2001) ([@CIT0017]). Their study was conducted on 47 boys and girls aged two months to five years born after *in vitro* fertilization (IVF) (82 eyes). Their results revealed the prevalences of myopia, hyperopia and emmetropia were 16%, 57%, and 27%, respectively. They performed cycloplegic refraction testing, and the result showed higher prevalence of hyperopia in comparison to prevalence of myopia. Nevertheless, in our findings showed that the prevalences of myopia, hyperopia and emmetropia are 2.9%, 87%, and 10.1%, respectively. Also, higher prevalence of hyperopia in our study in comparison to that in study of Anteby et al. may be due to either racial and genetic differences of two populations or differences in techniques used for ART ([@CIT0017]--[@CIT0025]). Changes in genetic background along with environmental influences can make significant differences in refractive status of newborn infants. Analysis of ocular and visual data of Iranian infants less than one year of age would be very helpful to come to a certain conclusion but unfortunately, there was not any research on infants and very young children in Iran. However, the studies that have been done in Tehran and Dezful cities by Fotohi et al (2007) may estimate the refractive status in some Iranian populations ([@CIT0026]). The result of study conducted in Tehran showed that among children over five years of age, there were 17.2% myopia and 56.5% hyperopia. While the result of study conducted in Dezful revealed that myopia was 3.4% and hyperopia was seen in 16.6% of the children between 7 to 15 years, indicating refractive error incidences were significantly different than our study. Therefore, considering decreasing of prevalence of hyperopia with age and emmetropization mechanism, different incidences of refractive error may be expected to happen in different ages ([@CIT0026]). However, precise conclusion may not be attainable in this regard.\n\nThe rate of astigmatism was 23.9% in Tehran and 18.7% in Dezful, whereas our result showed the rate of 38.6%.. Other studies in other countries mostly show lower prevalence for astigmatism. Axer et al. (2005) found no difference in incidence of astigmatism between IVF and naturally conceived infants ([@CIT0003]). In a study by Wikstrand (2006) on visual function and ocular morphology in children born after intra-cytoplasmic sperm injection (ICSI), he has demonstrates that there is no significant difference between ICSI and control group in the obtained results of performed following tests: hyperopia, myopia, astigmatism, heterotropia, stereo acuity, and convergence ([@CIT0018]).\n\nIn another study by Varghese et al. (2009), they examined the correlation of refractive error with birth weight, head circumference and birth age ([@CIT0020]). Their findings show a significant correlation between physical parameters of development and refractive status. They strictly recommend screening of refractive error for the infants who do not meet the criteria in developmental parameters. Management of refractive error in these infants can prevent many further visual defects. It is obvious that one of the main consequences of uncorrected refractive error, especially astigmatism, is amblyopia. In a study by Ludwig (2010), he compared vision and hearing function between children in ICSI and control groups, and did not find any significant difference between these two groups ([@CIT0027]).\n\nBasatemur et al. (2010) compared children conceived by IVF or ICSI with age matched naturally conceived children and found no significant difference regarding developmental parameters ([@CIT0028]). Nevertheless, in study by Basatemur (2010) and other similar studies have stated that further researches are needed for certain judgment ([@CIT0028], [@CIT0029]). It seems that more precise conclusion regarding refractive status of Iranian ART infants is required in order to gather sufficient information about normal infants (ART and naturally conceived infants) less than one year of age.\n\nIn our study, external examination of the eyes indicated abnormality in about 1% of the participants. This result seems better in comparison to some other studies (1% vs. 8.22%) ([@CIT0030]).\n\nOur findings also confirmed that incidence of red reflex abnormality was significantly more in boys than in girls. Red reflex abnormality may be due to high refractive error, media opacities or retinal defects ([@CIT0031]--[@CIT0033]). Examinations showed no significant difference in refractive error between boys and girls, so the difference found in incidence of red reflex abnormality may be due to media opacities or retinal defects ([@CIT0031], [@CIT0034]). Other studies showed no difference between boys and girls regarding media opacities and retinal defects. However, in a study by Eckstein et al. (1996), they stated that the reason of higher incidence of cataract in boys is due to a higher rate of boys referred to eye clinics. This is a scientific article, so has to be based on findings, documents and reasonable explanation.([@CIT0035]--[@CIT0037]). Nevertheless, sex dependent genetic disorders may show higher incidence in one sex, so these conditions may cause abnormalities in ocular media, retina, as well as changes in red reflex.\n\nIn another study by Bhatti (2003), he confirmed the higher prevalence of cataract in premature infants ([@CIT0037]), which is in agreement with our result that red reflex abnormality was significantly more common among premature infants.\n\nAnother parameter assessed in our study was ocular motility. We found that 20.3% of infants had poor fixation control. Unlike studies by Pedroso (2003) and Robaei (2006) on naturally conceived infants, our results showed that ocular reflexes of ART infants were not so desirable ([@CIT0038], [@CIT0039]). We found a significant difference in fixation control between premature and full-term infants. Premature infants had significantly poorer fixation control, which is a result of visual problems. In other words, health and integrity of visual system is necessary for proper fixation. Some studies have indicated that premature infants can have proper fixation control as in full-terms ([@CIT0040], [@CIT0041]). Nevertheless, other studies have demonstrated that premature infants have problem in fixation control in spite of their success in other visual and psychophysical examinations. In our study, disability to control fixation significantly increased with increasing refractive error in ART infants. This can be due to the fact that deterioration of visual acuity which is the result of high refractive error leading to undesirable effect on infant\\'s fixation function ([@CIT0042]).\n\nAfter performing Hirschberg test, we found that 2.9% of ART infants were afflicted with manifest strabismus. In other study on Iranian population by Fotouhi et al (2007), the prevalence of strabismus among children over five years old was about 0.8% ([@CIT0026]). In a study by Anteby et al. (2001) on ART infants, the prevalence of strabismus was declared about 4% ([@CIT0017]). Another study by Robaei et al. (2006) on Australian children population demonstrated that exo and eso deviations had different incidence rates ([@CIT0039]). They found significant correlation between prematurity and incidence of strabismus. Therefore, it seems that infants born after ART are more likely to have strabismus because of prematurity and their lower birth weight. However in this study, we did not find strabismus among participants. Binocular reflexes of these infants had not been completely developed, and it was possible that some of them would have better binocular condition in the future ([@CIT0039]). Using more accurate motility testing methods, like cover test, could lead to more reliable results, but it was not possible to perform cover test due to the age of the infants.\n\nAlthough some scientists did not find any differences in ocular and visual conditions between ART and naturally conceived infants ([@CIT0003], [@CIT0017]), special conditions of ART infants may affect development of the visual system, so some other scientists recommend vision screening for these infants ([@CIT0018], [@CIT0023], [@CIT0025]).\n\nConclusion {#S0005}\n==========\n\nThese results clearly reflect the necessity of more comprehensive assessments and further follow-up of infants born by ART, especially for premature male ART infants. These results also reflect the likelihood of finding refractive error in these children in older ages, and also, suggest the probability of fixation and visual deficiencies in these infants. It is recommended infertility specialists, ophthalmologists, optometrists, pediatricians, and parents of ART children to pay close attention to this preliminary report about the refractive and fixation condition of the infants born after ART.\n\nThis study was financially supported by Academic Center for Education, Culture and Research (ACECR). The authors express their gratitude to Ms. Sharareh Dadashloo for her cooperation. There is no conflict of interest in this article.\n"}
+{"text": "The authors confirm that all data underlying the findings are fully available without restriction. All data are included within the manuscript.\n\nIntroduction {#s1}\n============\n\nThe activator protein-1 (AP-1) family of transcriptional activators consists of dimeric combinations of Jun and Fos proteins that regulate a variety of transcriptional programs in response to various stimuli [@pone.0101015-Karin1]. AP-1 proteins share a common basic leucine zipper (bZip) motif that is responsible for dimerization and DNA-binding. AP-1 proteins recognize the AP-1 DNA binding site (consensus sequence: 5\u2032-TGA(C/G)TCG-3\u2032), also known as a phorbol 12-O-tetradecanoate-13-acetate (TPA) response element (TRE). Different subsets of AP-1 proteins have differing dimerization requirements. cJun, for example, can homo- and heterodimerize while cFos can only form heterodimers. These AP-1 dimers regulate a wide variety of cellular processes including the immune response, cell proliferation, apoptosis, and tumorigenesis [@pone.0101015-Eferl1].\n\nThe role of AP-1 proteins has been widely studied; however, discerning the distinct roles of individual dimer compositions remains challenging. Functions unique to cJun homodimers, but not cJun/cFos heterodimers have been identified. For example, cJun homodimers are not only capable of binding cis-elements on DNA to activate transcription but can also function as transcriptional co-activators by binding directly to other DNA-bound transcription factors, such as NFATc2 and PU.1 [@pone.0101015-Behre1]--[@pone.0101015-Walters1]. This function is unique to cJun/cJun and does not occur with cJun/cFos. Additionally, by expression of dimer specific mutants it was shown that cJun/cJun, cJun/Fra2, and cJun/ATF2 dimers have distinct functions in cJun induced transformation of chicken embryo fibroblasts [@pone.0101015-VanDam1]. Specifically, cJun/Fra2 induces anchorage independence and cJun/ATF2 induces growth factor independence. Another strategy to delineate unique functions of AP-1 dimers employed covalently tethering different combinations of Jun and Fos partners and testing their activities in cells. Different dimer compositions showed promoter-specific differences in activating transcription of reporter genes [@pone.0101015-Wisniewska1], [@pone.0101015-Bakiri1]. Together, these observations underscore the importance of developing tools to distinguish between different AP-1 dimer compositions in cells.\n\nCurrent strategies of gene knockout, siRNA knockdown, and transcription factor decoys have provided substantial insight into the role of AP-1 proteins in response to various stimuli [@pone.0101015-Eferl1], [@pone.0101015-Fennewald1]--[@pone.0101015-Hettinger1]. These strategies, however, do not discern the biological functions of different dimer compositions containing the same protein. For example, an AP-1 DNA decoy, which is an exogenous oligonucleotide containing the consensus AP-1 site, can sequester AP-1 proteins from gene promoters; however, this decoy targets all AP-1 dimers regardless of their composition. Moreover, knocking down cJun will inhibit the function of cJun/cJun homodimers as well as cJun heterodimers such as cJun/cFos. Similarly, ChIP assays against cJun cannot distinguish between sites of homo and heterodimer occupancy. Given the importance of AP-1 dimer composition on biological processes, research tools that allow us to discern between AP-1 dimers with different compositions would be very useful.\n\nSELEX (systematic evolution of ligands by exponential enrichment) is an iterative selection process to identify aptamers from a large DNA or RNA library that bind the desired target [@pone.0101015-Tuerk1], [@pone.0101015-Ellington1]. Here, we used SELEX to isolate a DNA aptamer that binds cJun; biochemical experiments found that the aptamer has high affinity and specificity for cJun/cJun homodimers compared to cJun/cFos heterodimers. The secondary structure and minimal binding region of the aptamer was determined. Using this aptamer we are able to specifically block cJun/cJun homodimers from binding AP-1 DNA. Moreover, the aptamer is capable of blocking cJun/cJun from cooperatively binding DNA with NFATc2, a common transcriptional partner of AP-1 proteins. We demonstrate that in cells this aptamer represses cJun/cJun-dependent IL-2 reporter activity. This work provides the foundation and selection method for specifically targeting different AP-1 dimers, which has the potential to provide insight into the unique biological roles of distinct AP-1 complexes.\n\nMaterials and Methods {#s2}\n=====================\n\nPlasmid construction {#s2a}\n--------------------\n\nThe pET-cJun, pET-\\\\is-cFos, pET-NFAT(DBD), and IL2-firefly-Luc reporter plasmids have been previously described [@pone.0101015-Walters1], [@pone.0101015-Ferguson1]. To prevent aggregation in EMSA, a cysteine at position 269 in cJun and 154 in cFos were mutated to serine by QuickChange site directed mutagenesis.\n\nProtein purification {#s2b}\n--------------------\n\ncJun was expressed and purified as previously described [@pone.0101015-Ferguson1]. The NFAT DNA-binding-domain was expressed and purified as previously described [@pone.0101015-Walters1]. For cFos expression and purification, cultures co-transformed with pET-6His-cFos and pSBET were grown in the presence of 100 \u00b5g/ml ampicillin and 50 \u00b5g/ml kanamycin in Luria-Bertani broth at 37\u00b0C to an optical density of 0.4 at 600 nm before expression was induced by the addition of isopropylthio-\u03b2-D-galactoside at a final concentration of 0.5 mM. After 2 hr, cells were harvested and the cell pellet was resuspended in a buffer containing 20 mM Tris (pH 7.9), 1 mM EDTA, 100 mM NaCl, 1 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and sonicated 4 times for 15 s. Samples were centrifuged for 30 min at 15,000 rpm and 4\u00b0C. Precipitated material containing 6His-cFos was resuspended in 10 ml of 20 mM Tris (pH 7.9), 1 mM EDTA, 100 mM NaCl, 5 mM DTT, 0.2 mM PMSF and sonicated 2 times for 30 s. The pellet was washed three more times by resuspending in 10 ml of 20 mM Tris (pH 7.9), 1 mM EDTA, 100 mM NaCl, 5 mM DTT, 0.2 mM PMSF followed by centrifugation. The pellet from the final wash was resuspended in 10 ml of buffer A (20 mM Tris (pH 7.9), 1 mM EDTA, 5 mM DTT, and 8 M urea, 0.1 M NaCl) containing 20 mM imidazole. Soluble material was loaded onto a Ni-NTA agarose column (Qiagen) and washed with 5 column volumes of buffer A containing 20 mM imidazole followed by 5 column volumes of buffer A containing 40 mM imidazole. cFos was eluted with buffer A containing 500 mM imidazole. Purified cFos was mixed with equimolar purified cJun in buffer A and subjected to three sequential dialyses in buffer B (20 mM Tris (pH 7.9), 0.1 mM EDTA, 10% glycerol, 5 mM DTT) containing the following additions: 1) 1 M urea and 1 M NaCl; 2) 1 M NaCl; 3) 0.1 M NaCl. After dialysis, the sample containing purified cJun/cFos heterodimers was separated into aliquots and stored at \u221280\u00b0C.\n\nIn vitro selection {#s2c}\n------------------\n\nThe ssDNA SELEX pool (IDT) was designed with a 40 nt randomized region flanked by the 5\u2032 constant region 5\u2032-GGGAGATCACTTACGGCACC-3\u2032 and the 3\u2032 constant region 5\u2032-CCAAGGCTCGGGACAGCG-3\u2032. Immobilized AP-1 DNA was made using biotin-5\u2032-AGGTCG[TGACTCA]{.ul}GCGG-3\u2032 annealed to 5\u2032-CCGC[TGAGTCA]{.ul}CGACCT-3 (the AP-1 site is underlined). 1 \u00b5mol double stranded DNA was incubated with 25 \u00b5l magnetic streptavidin beads (Invitrogen). Beads were washed three times with binding buffer (20 mM Tris (pH 7.9), 0.1 mM EDTA, 10% glycerol, 1 mM DTT, 150 mM KCl, 5 mM MgCl~2~, 0.04% NP-40). cJun (94 pmol) and 1\u00b5g poly(dI\u2022dC) were added to the immobilized AP-1 DNA and incubated at room temperature with nutation for 25 min in binding buffer. The poly(dI\u2022dC) served as a non-specific competitor to minimize background partitioning of the SELEX pool and any non-specific interactions that might occur between cJun and the DNA or beads. Unbound cJun was removed by washing the beads three times with binding buffer.\n\nFor the initial round of SELEX, 1.67 nmol of a ssDNA pool consisting of \u223c1\u00d710^15^ sequences was added in a volume of 3.4 \u00b5l to 150 \u00b5l of DNA-bound cJun beads and incubated for 30 min at room temperature in binding buffer with nutation. Unbound aptamers were removed by three 1 min washes in 100 \u00b5l binding buffer with nutation. ssDNA sequences were eluted by addition of 150 \u00b5l elution buffer (20 mM Tris (pH 7.9), 1.5 M NaCl) for 10 min at room temperature with nutation, then desalted using a G-25 column (GE Lifesciences). The ssDNA sequences in the eluates were amplified by PCR (using 25 \u00b5l of eluate) with forward primer 5\u2032-FAM-CGGGAGATCACTTACGGCACC-3\u2032 and reverse primer 5\u2032-AAAAAAAAAAAAAAAAAA-iSp9-CGCTGTCCCGAGCTTTGG-3\u2032 complementary to the constant regions. The poly-A stretch in the reverse primer added a poly-A tail to the strand complementary to the aptamer pool; iSp9 is a triethylene glycol spacer (IDT). The ssDNA aptamer pool was separated and purified from its complementary strand containing the poly-A stretch using a 8% denaturing polyacrylamide gel. Four rounds of positive selection were conducted using the conditions described for the first round with extension of the washing incubation to 2 min each for round 3, and 3 min each for round 4. This was followed by 2 rounds which included a negative selection against immobilized AP-1 DNA and washing incubation times of 3 and 5 min for rounds 5 and 6, respectively. For the negative selection the ssDNA SELEX pool was incubated with beads containing immobilized AP-1 DNA (30 \u00b5l beads for round 5 and 60 \u00b5l of beads for round 6) for 30 min in binding buffer at room temperature with nutation. The unbound ssDNA pool was then transferred to a tube containing 25 \u00b5l beads with DNA-bound cJun and the positive selection was performed as described. Aptamers from the final round of SELEX were amplified using a forward primer 5\u2032-GGGAGCTCACTTACGGCACC-3\u2032 containing a SacI restriction site and a reverse primer 5\u2032-GCCAAGCTTCGCTGTCCCGAGCCTTGG-3\u2032 containing a HindIII restriction site. The PCR product was digested with SacI and HindIII, gel purified, and ligated into a pUC18 vector for sequencing.\n\nElectrophoretic mobility shift assays {#s2d}\n-------------------------------------\n\n5\u2032-FAM or ^32^P-labeled aptamer or AP-1 DNA and purified cJun/cJun, cJun/cFos, and/or NFAT DBD were incubated together in 20 \u00b5l buffer containing 100 mM KCl, 10% glycerol, 4 mM MgCl~2~, 20 mM Tris (pH 7.9), 20 mM HEPES (pH 7.9), 0.1 mM DTT, 60 \u00b5g/ml BSA, and 0.06% NP-40 on ice for 20 min. For AP-1 DNA, the following oligos were annealed: 5\u2032-AGGTCG[TGACTCA]{.ul}GCGG-3\u2032 and 5\u2032-CCGC[TGAGTCA]{.ul}CGACCT-3 (the AP-1 site is underlined). 50 ng Poly(dI\u2022dC) was added to each reaction and incubated for an additional 5 min on ice. The reactions were subjected to electrophoresis through 4% polyacrylamide gels containing 1X Tris-glycine and 5% glycerol. For K~D~ determination the ^32^P-labeled DNA concentration was held constant at 50 pM and the protein was titrated over the indicated concentrations; data were fit with the following equation: Fraction bound\u200a=\u200aFraction bound~max~(\\[cJun/cJun\\]/(K~D~+\\[cJun/cJun\\])). For the EMSA with the round 6 pool the DNA concentration was held constant at 1 nM. The bands were visualized by fluorescent scanning (Typhoon 9400) for FAM-labeled DNA or phosphorimaging (Typhoon 9400) for ^32^P-labeled DNA. Data were quantified using ImageJ software. The sequences of the AP-1 DNA decoy and mutant AP-1 decoy are as follows: AP-1 DNA 5\u2032-GTCCA[TGACTCA]{.ul}GAAGAGACACACTCTTC[TGAGTCA]{.ul}TGGAC-3\u2032 (AP-1 sequence underlined) mutant AP-1 decoy 5\u2032-GTCCA[AATCTCA]{.ul}GAAGAGACACACTCTTC[TGAGATT]{.ul}TGGAC-3\u2032 (mutant AP-1 sequence underlined).\n\nNuclease probing and hydroxyl radical footprinting {#s2e}\n--------------------------------------------------\n\n20 fmol of ^32^P-labeled aptamer-19 was folded in RM buffer (20 mM HEPES (pH 7.9), 100 mM KCl, 8 mM MgCl~2~) by heating to 95\u00b0C and cooling on ice. Refolded aptamer was incubated with 30 units S1 nuclease in 1X S1 nuclease buffer (Promega) in a 20 \u00b5l reaction for 1 min on ice before adding 80 \u00b5l stop mix (200 mM KCl, 50 mM EDTA, 0.3 \u00b5g/\u00b5l yRNA). DNA was phenol-chloroform extracted and ethanol precipitated prior to resolving on a 10% denaturing polyacrylamide gel. Bands were visualized by phosphorimaging and quantified using ImageJ software. For the CviKI-1 digest, 20 fmol of ^32^P-labeled aptamer-19 folded in RM buffer was digested in 1X NEB4 buffer with 20 units CviKI-1 for 1 hr at room temperature. DNA was phenol-chloroform extracted and ethanol precipitated prior to resolving on a 10% denaturing polyacrylamide gel. Bands were visualized by phosphorimaging. The secondary structure drawing of aptamer-19 was created using VARNA software [@pone.0101015-Darty1].\n\nFor hydroxyl radical footprinting, glycerol was removed from the cJun prep using a G-25 column (GE lifescience). Protein was incubated with 20 fmol of folded ^32^P-labeled aptamer-19 in buffer containing 20 mM Tris (pH 7.9), 100 mM KCl, 4 mM MgCl~2~, 0.1 mM DTT, 20 \u00b5g/\u00b5l BSA, 0.06% NP-40, and 2.5 ng/\u00b5l Poly(dI\u2022dC) for 20 min on ice. Cleavage was initiated by mixing 1 \u00b5l each of 10 mM Fe(II)EDTA, 0.6% H~2~O~2~, and 10 mM sodium ascorbate for 3 min followed by quenching with 1 \u00b5l of 100 mM thiourea. Samples were ethanol precipitated and resolved on a 11.3% denaturing polyacrylamide gel. Bands were visualized by phosphorimaging and quantified using the software ImageJ.\n\nTransfection assays {#s2f}\n-------------------\n\nCOS-7 cells (purchased from ATCC) were maintained at 37\u00b0C and 5% CO~2~ in DMEM containing 10% FBS, 100 U/ml penicillin, 100 \u00b5g/ml streptomycin, and 2 mM GlutaMax. On the day of the transfection the cells were 75% confluent. Cells were transfected by Neon electroporation according to the manufacturers instructions (Invitrogen). Briefly, 500 ng of each protein expression construct, IL2-firefly-Luc reporter [@pone.0101015-Ferguson1], and pRL-TK-Renilla-luciferase (Promega) were mixed with the either aptamer-19 or its antisense sequence (AS) as a control; the oligos contained phosphorothioate linkages at the three terminal positions on both ends to increase cellular stability. Scrambled DNA (oligo of the same chemical composition as aptamer-19, but with a different sequence; 5\u2032-GTGACACGAATTGGGACCAGCGTATGGCTGATATAACATGTTTCGACCGAGCCTGACCGGTTG-3\u2032) was used to maintain an equal amount of oligonucleotide DNA for each reaction. Cells were electroporated and seeded into individual wells of a 6-well plate in the absence of antibiotics. 16 hr post transfection, cells were stimulated with 1 \u00b5M ionomycin and 20 ng/ml PMA for 6 hr. Cells were then harvested and lysed with 250 \u00b5l Passive Lysis Buffer (Promega). Firefly and renilla luciferase activities were determined using the Dual-luciferase kit (Promega).\n\nResults {#s3}\n=======\n\nSELEX targeting DNA-bound cJun homodimers yields aptamers that bind cJun/cJun with high affinity and specificity {#s3a}\n----------------------------------------------------------------------------------------------------------------\n\nOur goal was to use DNA SELEX to obtain aptamers that selectively bind cJun/cJun homodimers. Since, AP-1 proteins share a highly conserved bZip domain, responsible for protein dimerization and recognition of the consensus AP-1 element [@pone.0101015-Chinenov1], we hypothesized that blocking the DNA-binding domain during the selection process would facilitate obtaining DNA aptamers that have a higher binding affinity for cJun homodimers than cJun heterodimers such as cJun/cFos. cJun was incubated with dsDNA containing a consensus AP-1 site that was immobilized to magnetic beads via a biotin-streptavidin linkage. We selected DNA aptamers from a starting pool of \u223c1\u00d710^15^ unique sequences that were 78 nt long, consisting of a 40 nt randomized region flanked with constant regions of two different sequences. A schematic of the SELEX method is detailed in [Figure 1A](#pone-0101015-g001){ref-type=\"fig\"}. We performed 6 rounds of positive selection with rounds 5 and 6 also including a negative selection against beads and immobilized DNA. As shown in Figure S1A in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}, the DNA pool obtained after round 6 showed apparently stronger association with cJun/cJun that with cJun/cFos, despite the fact that cJun/cFos has a higher affinity than cJun/cJun for a consensus AP-1 DNA element (Figure S1B in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}).\n\n![Selection of ssDNA aptamers targeting DNA-bound cJun.\\\n**A)** cJun was incubated with the immobilized DNA and the complexes were washed prior to the addition of the ssDNA pool. Six rounds of SELEX were performed with rounds 5 and 6 including a negative selection as illustrated by the dashed arrows. **B)** Four individual sequences from round 6 of the selection, the conserved motifs are highlighted. **C)** EMSAs with aptamer-1, 16, 19, and 27. For each aptamer, either cJun/cJun or cJun/cFos was titrated from 0.4 nM to 6.5 nM.](pone.0101015.g001){#pone-0101015-g001}\n\nIndividual DNA molecules were isolated after round 6 and 46 clones were sequenced. All 46 sequences were different but contained families of over-represented motifs in the 5\u2032 and 3\u2032 ends of the randomized regions; no single sequence motif was present in all the aptamers. We experimentally tested 12 of the 46 sequenced aptamers that were representative of the sequence diversity present across all 46. All 12 aptamers bound cJun/cJun with high affinity. Four aptamers (1, 16, 19, and 27) that were amongst the highest affinity binders were chosen for further experimentation to address specificity for binding cJun/cJun homodimers compared to cJun/cFos heterodimers. Sequence alignment of these aptamers is shown in [Figure 1B](#pone-0101015-g001){ref-type=\"fig\"}, aptamers 1, 16, and 19 contain conserved motifs near the 5\u2032 and 3\u2032 ends of the randomized region, while aptamer 27 lacks the 3\u2032 conserved motif. As shown by the EMSA in [Figure 1C](#pone-0101015-g001){ref-type=\"fig\"}, all four aptamers bound cJun homodimers with low or sub-nanomolar binding affinity and did not appreciably bind cJun/cFos heterodimers over the concentration range tested. Moreover, the affinity of cJun/cJun homodimers for binding each of these aptamers is greater than the affinity with which cJun/cJun binds a consensus AP-1 element (see Figure S1B in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}). Because our primary goal was to obtain a single aptamer that bound cJun/cJun with high affinity and selectivity over cJun/cFos, we chose aptamer-19 for further characterization.\n\nAptamer-19 has a defined secondary structure, with three distinct elements required for binding cJun/cJun {#s3b}\n---------------------------------------------------------------------------------------------------------\n\nTo ultimately gain insight into the mechanism of aptamer binding to cJun/cJun, we interrogated the secondary structure of aptamer-19. First, to identify single stranded regions of the aptamer we used S1 nuclease, which preferentially degrades single stranded nucleotides. 5\u2032-^32^P-labeled aptamer-19 was digested with S1 nuclease and the products were resolved by denaturing electrophoresis ([Figure 2A](#pone-0101015-g002){ref-type=\"fig\"}). The relative band intensities were determined and plotted with the corresponding nucleotide position. Nucleotides that are predicted to be single stranded by S1 nuclease digestion are indicated by asterisks. To obtain a model for the secondary structure of the apatmer, the single stranded positions were used as input constraints for the mFold secondary structure prediction program ([Figure 2B](#pone-0101015-g002){ref-type=\"fig\"}) [@pone.0101015-Zuker1]. The predicted secondary structure contains three stem loops spanning from nucleotides 13--68. Portions of each constant region are predicted to anneal with the conserved motifs in the 5\u2032 and 3\u2032 ends of the randomized region (highlighted in blue) to form the flanking stems.\n\n![Secondary structure probing of aptamer-19.\\\n**A)** ^32^P-labeled aptamer was digested with S1 nuclease and resolved by denaturing gel electrophoresis. The relative band intensities of the +S1 nuclease lane are plotted according to nucleotide position at the right of the gel. **B)** The secondary structure of aptamer-19 as predicted by mFold given the single stranded constraints determined by S1 nuclease digestion. Conserved motifs are highlighted in blue. **C)** Restriction digest with CviKI-1 confirms the presence of the third stem-loop containing the double stranded recognition sequence 5\u2032-AGCC-3\u2032. The CviKI-1 site is labeled in panel B.](pone.0101015.g002){#pone-0101015-g002}\n\nTo further validate the proposed secondary structure we performed a restriction digest of 5\u2032-^32^P-labeled aptamer-19 with CviKI-1, which recognizes the double stranded motif 5\u2032-RGCY-3\u2032. If the third stem is present then CviKI-1 would create blunt-end cuts after nucleotides 54 and 66. As shown in [Figure 2C](#pone-0101015-g002){ref-type=\"fig\"} the predominant product created upon digestion with CviKI-1 is a 54 nt fragment, indicating the presence of the third stem-loop from nucleotides 53--68. The less abundant 66 nt product corresponds to CviKI-1 making a single cut between nucleotides 66 and 67.\n\nTo begin to identify the regions of aptamer-19 that interact with cJun we performed DNase I footprinting. As a positive control for DNase I footprinting we designed a dsDNA construct 78 nt long, the same length as the aptamer, with a high affinity AP-1 site from nucleotides 41--47. As shown in [Figure 3A](#pone-0101015-g003){ref-type=\"fig\"} both cJun/cJun and cJun/cFos protect a 15 nt region centered around the canonical AP-1 site, with each dimer inducing a unique profile of hypersensitive sites (raw data shown in Figure S2 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}). As shown in [Figure 3B](#pone-0101015-g003){ref-type=\"fig\"}, DNase I digestion of aptamer-19 without cJun/cJun or cJun/cFos present (sharp grey peaks, raw data shown in Figure S2 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}) results in fewer positions that show significant cleavage compared to digestion of the double stranded DNA shown in panel A. Despite less overall cleavage of aptamer-19, cJun/cJun protects nearly all of the cleaved nucleotide positions, revealing a much broader protection profile for aptamer-19 compared to AP-1 DNA. cJun/cFos only protected aptamer-19 near the 5\u2032 end, which further indicates that the aptamer binds cJun/cJun with specificity.\n\n![cJun/cJun homodimers, but not cJun/cFos heterodimers make extensive contacts with aptamer-19.\\\n**A)** Both cJun/cJun and cJun/cFos protect a discrete region surrounding the AP-1 site on a 78nt long dsDNA construct from DNase I digestion. The relative intensities of the digested products are plotted corresponding to nucleotide position. The digestion profile is presented in the absence of protein (solid gray), in the presence of 10 nM cJun/cJun (blue line), or the presence of 10 nM cJun/cFos (red line). **B)** cJun/cJun broadly protects aptamer-19 from DNase I digestion, while cJun/cFos does not. **C)** Digestion profile of aptamer-19 upon hydroxyl radical cleavage in the absence (solid gray) or presence of 5 nM cJun/cJun (blue line).](pone.0101015.g003){#pone-0101015-g003}\n\nTo obtain a higher resolution view of the regions of aptamer-19 that interact with cJun we performed hydroxyl radical footprinting. Hydroxyl radical cleavage targets the backbone of DNA with reactivity that is proportional to the solvent accessibility; as a result, the cleavage products are independent of the primary nucleotide sequence [@pone.0101015-Balasubramanian1]. Aptamer-19 was subjected to hydroxyl radical cleavage in the absence of cJun, and in the presence of 5 nM or 20 nM cJun (Figure S3 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}). The digested products were resolved by denaturing electrophoresis and the relative intensities of cleaved aptamer bands were plotted according to the corresponding nucleotide position ([Figure 3C](#pone-0101015-g003){ref-type=\"fig\"}). The addition of cJun/cJun protects the aptamer from hydroxyl radical cleavage at 4 distinct regions; nucleotides 11--13, 24--32, 52--56, and 68--72 all show a decrease in cleavage upon addition of cJun. When mapped onto the secondary structure of aptamer-19 these nucleotides occupy portions of the first and third stem-loops as well as the flanking single stranded regions (Figure S4 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}, circles). These data are consistent with the DNase I footprinting results (Figure S4 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}, triangles). Notably, the conserved motifs within the random region of aptamer-19 are contained within the regions that interact with cJun/cJun. Our results also show that cJun/cJun protects a broader region of aptamer-19 than the AP-1 consensus DNA elements, which suggests that the higher affinity of cJun/cJun for binding the aptamer compared to the consensus AP-1 element is due to a greater number of protein-nucleic acid contacts.\n\nAptamer-1\\\\12--74) has high specificity for binding cJun/cJun versus cJun/cFos and blocks cJun/cJun from binding an AP-1 DNA element {#s3c}\n------------------------------------------------------------------------------------------------------------------------------------\n\nTo identify the minimal region of aptamer-19 required to bind cJun/cJun with high affinity we used a series of truncations that systematically removed secondary structural elements (Figure S5 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}). These experiments showed the minimal binding region of aptamer 19 to be nucleotides 12--74. Aptamer-19 (12--74) was then used in EMSAs to measure the fold specificity for binding cJun/cJun homodimers over cJun/cFos heterodimers ([Figure 4A](#pone-0101015-g004){ref-type=\"fig\"}). Aptamer-19(12--74) bound cJun/cJun homodimers with a K~D~ of 0.5 nM and cJun/cFos heterodimers with a K~D~\\>85 nM. Hence, there is greater than 100-fold specificity of the aptamer for binding cJun/cJun homodimers over cJun/cFos heterodimers. These data imply that the aptamer targets a protein interface that is at least partially unique to the cJun homodimer.\n\n![Aptamer-19(12--74) binds cJun/cJun with high affinity and specificity, competes with AP-1 DNA for binding cJun, and inhibits cJun/cJun activated transcription in cells.\\\n**A)** Aptamer-19(12--74) is \\>100-fold more specific for binding cJun/cJun compared to cJun/cFos. Data from EMSAs were quantified and fit to binding curves; the equilibrium dissociation constant of the aptamer for cJun/cJun is 0.5 nM and for cJun/cFos is \\>85 nM. **B)** 2 nM cJun/cJun was incubated with 10 nM 5\u2032-Cy5-labeled AP-1 DNA and either unlabeled AP-1 decoy DNA, mutant AP-1 decoy, AS aptamer-19(12--74), or aptamer-19(12--74) at the concentrations indicated. **C)** Aptamer-19 prevents cJun/cJun from associating with NFAT bound to an NFAT/AP-1 composite site. dsDNA containing an NFAT/AP-1 composite element can be shifted by NFAT and cJun individually (lanes 2 and 3, respectively) or supershifted as a result of NFAT and cJun cooperativity (lane 4). Either an antisense of aptamer-19(12--74), aptamer-19(12--74), or AP-1 DNA was titrated into the reaction containing NFAT/AP-1 composite DNA, NFAT, and cJun. **D)** Overexpression of NFATc2 and cJun as well as stimulation with PMA and ionomycin are required for full activation of the IL-2 reporter in cells. Firefly luciferase was normalized to renilla luciferase; bars represent the average of three transfections and error bars represent one standard deviation. **E)** Aptamer-19(12--74) reduces transcription driven by cJun/cJun and NFATc2 from the IL-2 reporter in cells. Oligonucleotide concentrations were held constant under all conditions by cotransfecting a scrambled oligonucleotide of the same length. Firefly luciferase was normalized by the firefly luciferase plasmid copy number in each pool of transfected cells, as determined by real time PCR. Data were normalized to the IL-2 reporter activity in the presence of AS aptamer-19(12--74). Bars represent the average of three transfections and error bars represent one standard deviation. Asterisks represent statistical significance determined by a paired t-test. (\\*, p-value\\<0.032)](pone.0101015.g004){#pone-0101015-g004}\n\nWe asked whether aptamer-19 would compete with AP-1 DNA for binding cJun homodimers; given how the selection was performed we anticipated that cJun/cJun would still bind an AP-1 site in the presence of aptamer-19. We conducted an EMSA, allowing cJun/cJun to choose between ^32^P-labeled AP-1 DNA or unlabeled aptamer-19. As controls we also tested the following in competition with ^32^P-labeled AP-1 DNA: unlabeled AP-1 DNA decoy (an oligonucleotide that forms a hairpin containing the consensus AP-1 site), a mutant AP-1 DNA decoy (folds into the same structure but has the AP-1 site mutated), or a control oligo that is antisense (AS) to aptamer-19. As shown in [Figure 4B](#pone-0101015-g004){ref-type=\"fig\"}, aptamer-19 blocked cJun homodimers from binding AP-1 DNA, which we examine in more detail in the Discussion. The consensus and mutant AP-1 decoys functioned as expected, either blocking or not blocking cJun homodimers from binding DNA, respectively. Importantly, the AS aptamer control did not show any significant inhibition of cJun binding AP-1 DNA.\n\nTranscriptional regulation by AP-1 proteins often involves cooperative binding at composite promoter DNA sites with other transcription factors such as those in the NFAT family of transcriptional activators [@pone.0101015-Macian1]. At these composite sites, AP-1 and NFAT transcriptional activators cooperate to form a more stable DNA-bound complex. We hypothesized that due to the high binding affinity of aptamer-19 and its broad range of contacts with cJun homodimers, it would more effectively inhibit cJun homodimers from cooperatively binding DNA with NFAT compared to an AP-1 decoy. To test this we performed EMSAs using a high affinity NFAT-AP-1 composite element and asked whether aptamer-19 blocks the formation of the ternary NFAT/cJun/DNA complex. As shown in [Figure 4C](#pone-0101015-g004){ref-type=\"fig\"}, NFAT and cJun homodimers individually shifted the composite DNA element (lanes 2 and 3, respectively). Together, NFAT and cJun homodimers cooperatively bound DNA resulting in a slower migrating (supershifted) complex (lane 4). Either the antisense (AS) aptamer (lanes 5--7), aptamer-19 (lanes 8--10), or AP-1 DNA (lanes 11--13) were added at the concentrations indicated. Aptamer-19 showed the sharpest fold-decrease in the amount of NFAT/cJun/DNA complex, hence was more effective at inhibiting cJun homodimers from cooperatively binding DNA with NFAT than was the AP-1 DNA decoy. Additionally, when the cJun/NFAT/DNA complex was lost upon addition of aptamer-19, there was a corresponding increase in the NFAT/DNA band, demonstrating that aptamer-19 specifically targets cJun/cJun and not NFAT. The AS aptamer did not inhibit formation of the ternary complex to nearly the same extent as either the AP-1 decoy or aptamer-19.\n\nAptamer-19 inhibits IL-2 reporter activity in cells {#s3d}\n---------------------------------------------------\n\nWe previously reported that cJun homodimers and NFATc2 can achieve high levels of transcriptional synergy on the IL-2 promoter [@pone.0101015-Nguyen1]. The high level of synergy requires a unique interaction between the DNA-binding domain of cJun homodimers and the C-terminal activation domain of NFATc2 [@pone.0101015-Walters1]. We asked if the minimal binding domain of aptamer-19, which can specifically prevent cJun homodimers from binding DNA, could repress cJun/cJun-specific transcriptional synergy with NFATc2 at the IL-2 promoter. [Figure 4D](#pone-0101015-g004){ref-type=\"fig\"} illustrates transcriptional activation of an IL-2 reporter in response to overexpression of NFATc2 and cJun, and stimulation by PMA and ionomycin. Maximal IL-2 activity is only achieved when NFATc2 and cJun are overexpressed and cells are stimulated with PMA and ionomycin. When cotransfected under these conditions, aptamer-19 repressed IL-2 luciferase activity relative to its antisense sequence ([Figure 4E](#pone-0101015-g004){ref-type=\"fig\"}). Moreover, the repression mediated by aptamer-19 was attenuated by cotransfecting an equal amount of its antisense sequence, acting in essence as an antidote for aptamer-19. Thus, aptamer-19 appears to inhibit cJun homodimer-mediated transcription in cells. These results establish the potential for AP-1 dimer specific aptamers to serve as useful tools in dissecting the role of AP-1 dimer composition in transcriptional regulation.\n\nDiscussion {#s4}\n==========\n\nAP-1 transcription factors are key regulators of proliferation, tumorigenesis, apoptosis, and the immune response. Individual AP-1 family members are differentially expressed, resulting in a complex mixture of AP-1 proteins that are likely context and cell-type specific. Here, we used SELEX to identify DNA aptamers that bind cJun/cJun with high affinity and specificity compared to cJun/cFos. Biochemical assays revealed the secondary structure of an aptamer, the minimal region necessary for binding cJun homodimers, and showed that the aptamer competes with AP-1 DNA for binding cJun/cJun. In cells, the aptamer repressed cJun homodimer activated transcription from a reporter. This work sets the foundation and provides a selection method for interrogating the biological functions unique to distinct AP-1 dimer compositions.\n\nAn aptamer that binds cJun/cJun with high affinity and specificity {#s4a}\n------------------------------------------------------------------\n\nThe SELEX procedure has been used to isolate RNA and DNA aptamers that bind an extensive catalog of protein targets with high affinity and specificity [@pone.0101015-Nimjee1]. In particular, aptamers have proved to be extremely useful tools in dissecting the role of specific protein-protein and protein-DNA interactions of various transcription factors [@pone.0101015-Brody1]--[@pone.0101015-Park1]. In many of these cases the selected aptamers bind where DNA typically interacts, which is a high affinity nucleic acid binding site. It has been demonstrated, however, in a selection against the TATA-binding protein (TBP), that by masking the DNA binding domain aptamers can be directed to discrete binding surfaces on the protein [@pone.0101015-Shi1]. Here, we chose to mask the DNA binding domain of cJun homodimers with AP-1 DNA during the selection, with the goal of directing the aptamers to bind a surface on cJun homodimers that is not present on cJun/cFos heterodimers.\n\nGiven how we designed our selection process, however, we were surprised to find that aptamer-19 blocked cJun from binding AP-1 DNA ([Figure 4B](#pone-0101015-g004){ref-type=\"fig\"}). A likely model for this is that the region(s) of the aptamer not responsible for specific cJun recognition could, driven primarily by electrostatic interactions, occupy the basic region of the cJun DNA binding domain, thereby preventing it from subsequently recognizing AP-1 DNA. During the selection, when cJun/cJun was pre-bound to AP-1 DNA, specific high affinity contacts with cJun outside the DNA binding domain were adequate to recover aptamers. This model is consistent with the observation that during the SELEX the last aptamer pool did not displace cJun/cJun from immobilized AP-1 DNA, but aptamer-19 competes with AP-1 DNA when cJun/cJun is given the choice of which DNA to bind. In addition, this binding model suggests that the aptamer binds more than one site on cJun, which is consistent with the broad range of contacts between the aptamer and cJun homodimers as revealed by hydroxyl radical and DNase I footprinting ([Figures 3B and 3C](#pone-0101015-g003){ref-type=\"fig\"}). The DNase I footprint of DNA containing an AP-1 site showed a significantly smaller region of protection by cJun/cJun compared to the aptamer DNA ([Figure 3A](#pone-0101015-g003){ref-type=\"fig\"}). Moreover, the \u223c100-fold specificity for binding cJun homodimers compared to cJun/cFos heterodimers supports a binding mechanism in which the aptamer associates with regions of cJun/cJun outside the DNA binding domain.\n\ncJun/cJun binds aptamer-19 with a higher binding affinity compared to the consensus AP-1 sequence (\u223c0.5 nM versus \u223c15 nM, respectively). Aptamers that target the DNA binding domain of transcription factors tend to have a binding affinity similar to, or weaker than, the optimal DNA recognition site. For example, heat shock factor 1 (HSF-1) binds its respective DNA recognition element with a K~D~ of \u223c1 nM, yet binds an RNA aptamer isolated after 14 rounds of SELEX with an affinity of 20--40 nM [@pone.0101015-Zhao1], [@pone.0101015-Kroeger1]. Similarly, RNA aptamers targeting the DNA binding domain of NFATc2 bind with a K~D~ of \u223c10--100 nM, which is approximately 10-fold weaker than the NFAT DNA recognition sequence [@pone.0101015-Cho1], [@pone.0101015-Kim1]. One of the most widely studied transcription factor aptamers targets NFkB and inhibits NFkB from binding DNA [@pone.0101015-Lebruska1]. This RNA aptamer has a similar binding affinity for NFkB as the DNA recognition element, and crystal and NMR structures revealed that this aptamer structurally mimics the NFkB DNA recognition element [@pone.0101015-Huang1], [@pone.0101015-Reiter1]. By contrast, the observations that aptamer-19 binds cJun/cJun with at least 30 fold greater affinity than a consensus AP-1 DNA site and that the aptamer can block DNA binding by cJun/cJun is consistent with the aptamer interacting with cJun/cJun in the DNA binding domain as well as another region of the protein.\n\nIt was interesting to find that both of the constant regions used during the selection formed stem-loop structures with sequences derived from the randomized region in aptamer-19. We found these two stem-loops were required for binding cJun/cJun (Figure S5 in [File S1](#pone.0101015.s001){ref-type=\"supplementary-material\"}), although in both cases we do not know whether it is the structure or the sequence of the stem-loops that is important for mediating binding. This raises the question of whether using a different set of constant regions would also result in selecting high-affinity aptamers with critical stem-loops forming between sequences in the constant and randomized regions. If the primary determinant of high affinity binding is indeed the structures and not the sequences of the flanking stem-loops (see [Figure 2B](#pone-0101015-g002){ref-type=\"fig\"}), then choosing a different constant region would likely result in selection of aptamers that still contain sequences complementary to the constant region in order to form two flanking stem-loops. Alternatively, if both sequence and structure of the flanking stem-loops in aptamer-19 are important for binding, then performing the selection with a different constant region has the potential to result in aptamers that bind cJun/cJun with a different structure from that of aptamer-19.\n\nThe potential for aptamers to delineate AP-1 activation mechanisms in cells {#s4b}\n---------------------------------------------------------------------------\n\nAt the promoters of inducible genes, such as those involved in the immune response, AP-1 proteins cooperate with other transcription factors, namely those in the NFAT family. Given that aptamer-19 has higher affinity for cJun homodimers than AP-1 DNA we hypothesized that this aptamer could be an effective inhibitor of cJun/NFAT cooperative DNA binding. By EMSA, we showed that aptamer-19, but not its antisense sequence, significantly inhibited cJun/cJun from cooperatively binding DNA with NFAT.\n\nSince aptamer-19 functioned to potently inhibit cJun from binding AP-1 DNA, we asked if it could inhibit transcription from an IL-2 promoter in cells. We previously reported that cJun homodimers and NFATc2 cooperate to drive high levels of synergistic transcription from the IL-2 promoter [@pone.0101015-Walters1], [@pone.0101015-Nguyen1]. Since these high levels of synergy are unique to cJun homodimers, this allowed us to test for repression of cJun/cJun activated transcription. By transient transfection, we found that aptamer-19 repressed cJun/cJun and NFAT-dependent IL-2 reporter activity compared to transfection with an AS control oligo. Moreover, the repression by aptamer-19 was attenuated by cotransfection with the antisense control oligo, suggesting that the repression mediated by aptamer-19 necessitates the defined secondary structure of the aptamer.\n\nRecent studies have identified a potential coactivator function unique to cJun homodimers but not cJun/cFos heterodimers [@pone.0101015-Behre1]--[@pone.0101015-Walters1]. In this context, cJun utilizes its DNA binding domain to interact with DNA bound transcription factors. By specifically targeting cJun/cJun and obstructing DNA binding, aptamer-19 could prove to be a useful tool in probing the biological functions of cJun homodimers, including its role as a coactivator. We believe that this experimental approach also sets the foundation for targeting various AP-1 dimer compositions, helping to discern their distinct biological roles.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Figure S1,** Relative aptamer and AP-1 DNA binding affinities of cJun/cJun homodimers and cJun/cFos heterodimers. **Figure S2,** DNase I footprinting of AP-1 DNA and aptamer-19. **Figure S3,** Hydroxyl radical cleavage of aptamer-19 shows four distinct regions of protection upon addition of cJun. **Figure S4,** cJun footprint determined by DNase I and hydroxyl radical cleavage mapped onto the secondary structure of aptamer-19. **Figure S5,** All three stem-loops of aptamer-19 are required for binding cJun/cJun with high affinity.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: RDW DTM JFK JAG. Performed the experiments: RDW DTM. Analyzed the data: RDW DTM. Contributed reagents/materials/analysis tools: RDW DTM. Contributed to the writing of the manuscript: RDW JFK JAG.\n"}
+{"text": "1. Background {#sec140083}\n=============\n\nCarbapenems are considered to be one of the few drugs that are useful for the treatment of infections caused by multiresistant Gram-negative bacteria. The emergence of carbapenem-resistant *Enterobacteriaceae* is a serious public health concern due to the large spectrum of resistant genes and the lack of therapeutic options ([@A29314R1], [@A29314R2]). Therefore, there is an ongoing effort in the development of earlier and more sensitive detection of carbapenemase producers. With regard to carbapenem resistance, it must be taken into account that carbapenemase-producing bacteria may sometimes exhibit only a slight increase of minimal inhibitory concentration (MIC) values for carbapenems, which reflects the importance of molecular approaches to phenotypic tests ([@A29314R3]).\n\n*Enterobacteriaceae*, such as *Escherichia coli*, *Klebsiella pneumoniae* and *Enterobacter*spp., commonly cause nosocomial pneumonias and infections in the bloodstream, urinary tract and intra-abdominal region ([@A29314R4], [@A29314R5]). Carbapenem resistance mechanisms in *Enterobacteriaceae* include the following: (i) enzymatic inactivation by \u03b2-lactamases; (ii) modification of outer-membrane proteins (porins) and penicillin-binding proteins; and (iii) efflux pumps ([@A29314R6]). Carbapenem-hydrolyzing \u03b2-lactamases belonging to molecular class A (e.g., KPC, GES, IMI, SME), class B (e.g. IMP, VIM, NDM, GIM) and class D (e.g. OXA-23 and OXA-48), are the main source of antibiotic resistance in *Enterobacteriaceae*. Genes encoding these types of carbapenemases are extensively reported among *E. coli* and *K. pneumoniae* isolates from many European countries ([@A29314R1]).\n\nThe carbapenem-hydrolyzing class D \u03b2-lactamase OXA-48 has been identified in a *K. pneumoniae* isolate from Europe ([@A29314R7], [@A29314R8]), OXA-23 in *Proteus mirabilis* from France ([@A29314R9]), OXA-162 in *K. pneumoniae* isolates from Hungary ([@A29314R10]), OXA-181 in *K. pneumoniae* from Romania ([@A29314R11]) and *Citrobacter freundii* from France ([@A29314R12]), OXA-232 in *K. pneumoniae* from France ([@A29314R13]) and OXA-244 and 245, both in *K. pneumoniae* from Spain ([@A29314R14]). Additionally, OXA-247 was first described in *K. pneumoniae* in Argentina ([@A29314R15]). Recently, *bla*~OXA-51-like~, *bla*~OXA-58~ and *bla*~DIM-1~ carbapenemase genes have been found in a large variety of enterobacterial species ([@A29314R16]). Further, the presence of carbapenemases and extended-spectrum \u03b2-lactamases (ESBLs) were also described: (i) *bla*~TEM-1~, *bla*~SHV-11~, *bla*~CTX-M-15~ and *bla*~OXA-9~genes were present in the *K. pneumoniae* isolates harboring *bla*~OXA-48~; (ii) *bla*~TEM-1~and *bla*~OXA-1~genes were found in *E. coli* harboring *bla*~OXA-162~and *bla*~OXA-48~; (iii) *bla*~SHV-5~in *C. freundii* harboring bla~OXA-162~; and (iv) *bla*~TEM-1~ and *bla*~CTX-M-15~ in *Enterobacter cloacae* carrying *bla*~OXA-48~ ([@A29314R17]).\n\nMoreover, *E. coli*and *K. pneumoniae* clinical isolates producing CTX-M-2 and CTX-M-92 and a *K. pneumoniae* isolate with CTX-M-3 were detected, and each had the OXA-2 type beta-lactamase ([@A29314R18]). In addition, KPC-2 and TEM-1 enzymes were identified in *K. pneumoniae* isolates, and VIM-1 and TEM-1 in *P. mirabilis*, both of which were positive for *bla*~OXA-10~ ([@A29314R19]). In this context, a recent study has demonstrated that OXA-2 and OXA-10 are in fact carbapenem-hydrolyzing class D beta-lactamases (CHDLs), also called oxacillinases (OXA) ([@A29314R20]). Interestingly, development of carbapenem resistance in CTX-M-1-producing *Enterobacteriaceae* has been reported ([@A29314R21], [@A29314R22]). Furthermore, it has been reported that CMY-2 \u03b2-lactamase plays a role in carbapenem resistance (\\\"trapping\\\" of meropenem) ([@A29314R23]).\n\nThe rates of hydrolysis by OXA-type carbapenemases are weak. However, genes encoding OXA overcome this deficiency by possessing efficient promoters, leading to their overexpression and to increased carbapenem resistance ([@A29314R24]). Thus, rapid and useful detection methods are important for the implementation of appropriate infection control measures to prevent the further spread of ESBLs and carbapenemases.\n\n2. Objectives {#sec140084}\n=============\n\nThe aim of the present study was to develop the primers for single and/or multiplex PCR amplification assays for identification of all known genes that confer carbapenem resistance in *Enterobacteriaceae* and evaluate the amplification efficiency of the primers.\n\n3. Materials and Methods {#sec140088}\n========================\n\n3.1. Bacterial Isolates {#sec140085}\n-----------------------\n\nFifteen carbapenemase-producing strains were used as positive controls for optimizing the multiplex PCR assay; six reference strains and nine clinical isolates. The reference strains used were: (i) IMP-type (NCTC 13476), producing *E. coli*; (ii) KPC-3 positive *K. pneumoniae* (NCTC 13438); (iii) NDM-1 positive *K. pneumoniae* (NCTC 13443); (iv) VIM-10 positive *Pseudomonas aeruginosa* (NCTC 13437); (v) OXA-23 positive *Acinetobacter baumannii* (NCTC 13301); and (vi) OXA-48 positive *K. pneumoniae* (NCTC 13442). The clinical isolates were: (i--v) *K. pneumoniae* that produce VIM (V602B), KPC-2 (V117), KPC-3 (V514) and KPC (V601, V646); (vi--viii) P. ~aeruginosa~-producing IMP (V7424) and VIM (V109, V7393); and (ix) NDM-1 positive *A. baumannii* (V509). All the clinical strains were provided by Ing. J. Hrab\u00e1k, Ph.D. from the Department of Microbiology, Faculty of Medicine and University Hospital in Plzen, Charles University in Prague, Plzen, Czech Republic.\n\n3.2. Design of Group-Specific Primers for Single and Multiplex PCR Assays {#sec140086}\n-------------------------------------------------------------------------\n\nTwo multiplex PCRs were designed in this study: a *bla*~KPC~/*bla*~OXA-48-like~/*bla*~VIM~ multiplex PCR and a *bla*~NDM-1~/*bla*~IMP~ variants/*bla*~OXA-23-like~ mutliplex PCR. The sequences of genes that encode carbapenemases (KPC, VIM, IMP, SME, IMI, GES, NDM, OXA) so far described (http://www.lahey.org/studies/; last accessed February 2015) were downloaded from the GenBank databases and were aligned using Geneious Pro 4.8.5 (Biomatters Ltd, Newark, NJ, USA) to identify highly homologous regions suitable for designing primers. Two sets of primers were tested against reference standard strains, as well as clinical isolates, in a single PCR reaction and then in a multiplex format. These reference strains included NCTC 13476, NCTC 13438, NCTC 13443, NCTC 13437, NCTC 13301 and NCTC 13442. Representative V602B, V117, V514, V601, V646, V7424, V109, V7393 and V509 were used as clinical isolates. Further, eleven pairs of primers were designed but not tested. The primer sequences, concentrations and calculated lengths of the corresponding amplicons are listed in [Table 1](#tbl36048){ref-type=\"table\"}.\n\n###### Sequences of Primers Used for Single and Multiplex PCR for Detection of Genes Encoding Carbapenemases in *Enterobacteriaceae*\n\n  PCR Name                                                     Targeted Gene                                                                  Primer Name   Sequence (5\\' to 3\\' Direction) ^[a](#fn38479){ref-type=\"table-fn\"}^   Length (Bases)   Amplicon Size, bp   TM in \u00b0C   Primer CONCENTRATION, pmol/\u00b5L   Reference\n  ------------------------------------------------------------ ------------------------------------------------------------------------------ ------------- ---------------------------------------------------------------------- ---------------- ------------------- ---------- ------------------------------- ---------------------------------------\n  **Multiplex I KPC, OXA and VIM**                             *bla*~KPC~type                                                                 Forward       TGTTGCTGAAGGAGTTGGGC                                                   20               340                 56         15 pmol                         This study\n                                                                                                                                              Reverse       ACGACGGCATAGTCATTTGC                                                   20                                                                              \n                                                               *bla*~OXA-48-like~ including OXA-199 and OXA-370                               Forward       AACGGGCGAACCAAGCATTTT                                                  21               585 or 597                     15 pmol                         This study\n                                                                                                                                              Reverse       TGAGCACTTCTTTTGTGATGGCT                                                23                                                                              \n                                                               *bla*~VIM~ type                                                                Forward       CGCGGAGATTGARAAGCAAA                                                   20               247                            15 pmol                         This study\n                                                                                                                                              Reverse       CGCAGCACCRGGATAGAARA                                                   20                                                                              \n  **Multiplex II NDM, IMP and OXA**                            *bla* ~NDM-1~                                                                  Forward       TAAAATACCTTGAGCGGGC                                                    19               439                 52         15 pmol                         This study\n                                                                                                                                              Reverse       AAATGGAAACTGGCGACC                                                     18                                                                              \n                                                               *bla*~IMP~ variants except, IMP-3, IMP-16, IMP-27, IMP-31, IMP-34 and IMP-35   Forward       GAGTGGCTTAATTCTCRATC                                                   20               183                            15 pmol                         Monteiro et al., 2012 ([@A29314R25])\n                                                                                                                                              Reverse       CCAAACYACTASGTTATCT                                                    19                                                                              Ellington et al., 2007 ([@A29314R26])\n                                                               *bla* ~OXA-23-like~                                                            Forward       GTGGTTGCTTCTCTTTTTCT                                                   20               736                            15 pmol                         This study\n                                                                                                                                              Reverse       ATTTCTGACCGCATTTCCAT                                                   20                                                                              \n  **PCR/Multiplex** ^**[b](#fn38480){ref-type=\"table-fn\"}**^   *bla* ~SME(1-5)~                                                               Forward       TATGGAACGATTTCTTGGCG                                                   20               300                 56                                         This study\n                                                                                                                                              Reverse       CTCCCAGTTTTGTCACCTAC                                                   20                                   58                                         \n                                                               *bla* ~IMI\\ type~                                                              Forward       CTACGCTTTAGACACTGGC                                                    19               481                 57                                         This study\n                                                                                                                                              Reverse       AGGTTTCCTTTTCACGCTCA                                                   20                                   56                                         \n                                                               *bla* ~IMP-27~                                                                 Forward       AAAGGCACTGTTTCCTCACA                                                   20               169                 56                                         This study\n                                                                                                                                              Reverse       TCGCCAGCCAATAACTAACC                                                   20                                   58                                         \n                                                               *bla* ~NDM(2-10,\\ 12)~                                                         Forward       ATGACCAGACCGCCCAGAT                                                    19               380                 60                                         This study\n                                                                                                                                              Reverse       GAGATTGCCGAGCGACTTG                                                    19                                   60                                         \n                                                               *bla* ~IMP(3,\\ 34)~                                                            Forward       GTGGTTCTTGTAAATGCTGAGG                                                 22               532                 60                                         This study\n                                                                                                                                              Reverse       CCGCCTGCTCTAATGTAAGT                                                   20                                   58                                         \n                                                               *bla* ~GES(1-22,\\ 24)~                                                         Forward       GAAACCAAACGGGAGACGC                                                    19               207                 60                                         This study\n                                                                                                                                              Reverse       CTTGACCGACAGAGGCAACT                                                   20                                   60                                         \n                                                               *bla* ~GIM-1~                                                                  Forward       TCGACACACCTTGGTCTGAA                                                   20               477                 58                                         Ellington et al., 2007 ([@A29314R26])\n                                                                                                                                              Reverse       AACTTCCAACTTTGCCATGC                                                   20                                   56                                         \n                                                               *bla* ~OXA-2-like~                                                             Forward       ATTTCAAGCCAAAGGCACGA                                                   20               568                 56                                         This study\n                                                                                                                                              Reverse       GCCACTCAACCCATCCTACC                                                   20                                   63                                         \n                                                               *bla* ~OXA-10-like~                                                            Forward       ACGGAAAGCCAAGAGCCAT                                                    19               354 or 357          57                                         This study\n                                                                                                                                              Reverse       CCCACACCAGAAAAACCAGT                                                   20                                   58                                         \n                                                               *bla* ~OXA-51-like~                                                            Forward       TGTACCTGCTTCGACCTTCA                                                   20               435                 58                                         This study\n                                                                                                                                              Reverse       TCCCCAACCACTTTTTGCGT                                                   20                                   58                                         \n                                                               *bla* ~OXA-58-like~                                                            Forward       GCCATTCCCCAGCCACTTTTA                                                  21               470                 61                                         This study\n                                                                                                                                              Reverse       CACGCATTTAGACCGAGCAA                                                   20                                   58                                         \n                                                               *bla* ~DIM-1~                                                                  Forward       CGGCTGTTTTGTGCGTAG                                                     18               215                 56                                         This study\n                                                                                                                                              Reverse       GCGTTCGGCTGGATTGATT                                                    19                                   57                                         \n                                                               *bla* ~CMY-2~                                                                  Forward       GCCGTTGCCGTTATCTAC                                                     18               511                 56                                         This study\n                                                                                                                                              Reverse       AATCTTTTTGTTCGTTCTGCG                                                  21                                   55                                         \n\n^a^For degenerate primers: R = A or G; S = G or C; Y = C or T.\n\n^b^Designed primers but not tested.\n\n3.3. DNA Extraction and Multiplex PCRs {#sec140087}\n--------------------------------------\n\nDNA preparation was performed by suspending a colony of each bacterial strain in 100 \u03bcL of distilled water, boiling at 98\u00b0C for 10 minutes and centrifuging the cell extract for five minutes at 13,000 rpm. A multiplex PCR assay was designed to detect and differentiate one family of class A carbapenemase (KPC), three families of class B carbapenemases (IMP, NDM and VIM) and two families of class D carbapenemases (OXA) in two reactions. Both multiplex PCRs were performed with six pairs of specific primers ([Table 1](#tbl36048){ref-type=\"table\"}), which were used to amplify fragments (different in size) of 340 bp (KPC), 597 bp (OXA-48), 247 bp (VIM), 439 bp (NDM), 183 bp (IMP) and 736 bp (OXA-23). The PCR reaction mixture contained: 0.5 \u00b5L DNA (50 ng) in 24.5 \u00b5L complete reaction buffer with MgCl~2~ (containing 100 mmol/L Tris-HCl \\[pH 8.8\\], 500 mmol/L KCl, 1% Triton X-100, 15 mmol/L MgCl~2~) (Top-Bio, Czech Republic; 4 \u00b5L), dNTP (10 mM, 0.5 \u00b5L), 15 pmol of each primer (0.5 \u00b5L) and Taq DNA polymerase (Top-Bio, Czech Republic; 0.2 \u00b5L).\n\nThe PCR conditions were as follows: initial denaturation at 95\u00b0C for five minutes, 35 cycles at 95\u00b0C for one minute, at different annealing temperatures for one minute and 72\u00b0C for one minute, followed by a single, final elongation step at 72\u00b0C for five minutes. The annealing temperature was optimal at 56\u00b0C for amplification of the *bla*~KPC~, *bl*a~OXA~ and *bla*~VIM~ genes and optimal at 52\u00b0C for amplification of the *bla*~NDM~, *bla*~IMP~ and *bla*~OXA~ genes ([Table 1](#tbl36048){ref-type=\"table\"}). Amplicons were visualized after running at 100 V for 90 minutes on a 1.5% agarose gel containing ethidium bromide ([Figure 1](#fig26519){ref-type=\"fig\"}). A 200 -- 1500 bp DNA ladder (Top-Bio, Czech Republic) was used as a size marker.\n\n![Detection of Genes Encoding Carbapenemases in Carbapenem-resistant Control Strains and Clinical Isolates by Multiplex PCR\\\nNotes, lane 1, control *bla*~KPC~ gene; lane 2, control *bla*~OXA-48~gene; lane 3, control *bla*~VIM~ gene; lane 4, control *bla*~NDM-1~ gene; lane 5, control *bla*~IMP~ gene; lane 6, control *bla*~OXA-23~ gene. M, Molecular mass markers (200 - 1500 bp DNA ladder).](jjm-09-01-29314-g001){#fig26519}\n\n4. Results {#sec140089}\n==========\n\nAfter optimizing the PCR (amplification) conditions, all positive controls (reference strains and clinical isolates) yielded amplicons of the predicted sizes and confirmed the specificity of the primers used ([Figure 1](#fig26519){ref-type=\"fig\"}). The primer paris were tested in mixed ([Table 1](#tbl36048){ref-type=\"table\"}, [Figure 1](#fig26519){ref-type=\"fig\"}) and individual reactions (data not shown). The two multiplex PCR assays were 6 validated with a panel of fifteen characterized Gram-negative bacterial strains. This collection includes 5 class A carbapenemases (5 KPC), 8 class B carbapenemases (2 IMP, 2 NDM and 4 VIM) and 2 class D carbapenemases (2 OXA). The resistance genes of the control isolates were correctly determined by each multiplex PCR (accuracy 100%; [Figure 1](#fig26519){ref-type=\"fig\"}). Non-specific amplification generating fragments of unexpected size was not observed. In addition, eleven pairs of primers have been proposed but not tested experimentally in PCR and/or multiplex reactions ([Table 1](#tbl36048){ref-type=\"table\"}).\n\n5. Discussion {#sec140090}\n=============\n\nThe emergence of carbapenem resistance in *Enterobacteriaceae* has become a substantial clinical problem, since carbapenemase production cannot be easily inferred from the antimicrobial resistance profiles, thus, dissemination of these enzymes among nosocomial pathogens (e.g., *Enterobacteriaceae*) is a threat to public health and must be closely monitored (phenotypic and genotypic tests). In addition, delay in detection of multidrug-resistant *Enterobacteriaceae*results in longer hospitalizations and increased healthcare costs ([@A29314R26]).\n\nRecently, multiplex PCR assays for the detection of *bla*~IMP~, *bla*~VIM~, *bla*~OXA~, *bla*~NDM~ and *bla*~KPC~ have been described ([@A29314R25]-[@A29314R30]). Unfortunately, the primers used had lower detection ranges (many homologs are missed). For example, the multiplex PCR assay described by Kaase et al. ([@A29314R29]) targeted VIM and IMP enzymes, among other carbapenemases, and allowed the detection of 22 variants of IMP (IMP-1 to IMP-22) and 13 variants of VIM (VIM-1 to VIM-13). However, according to the Lahey Clinic website, to this day over 41 IMP and 39 VIM derivatives have been documented. This suggests that specificity of those primers may be impaired due to sequence variations. Nijhuis et al. ([@A29314R30]) targeted IMP, OXA-48-like types (OXA-48, 162, 163, 181, 204, 232, 244, 245, 370) and other carbapenemase genes. The assay reported in this study allowed the detection of 33.3% of IMP variants (14/42) and was not able to detect OXA-23 and OXA-247.\n\nThis study shows that primers in two multiplex PCRs are able to detect of KPC, VIM, NDM-1, OXA (OXA-23-like and OXA-48-like enzymes) and IMP variants (except IMP-3, IMP-16, IMP-27, IMP-31, IMP-34 and IMP-35) present in *Enterobacteriaceae* clinical isolates. Due to the success of the first primer sets, the additional untested sets ([Table 1](#tbl36048){ref-type=\"table\"}) are predicted to detect the remaining NDM- and IMP-type enzymes and other carbapenemases. Therefore, the combination of primer sets presented in this study will cover all known carbapenemase genes found in *Enterobacteriaceae*.\n\nThe multiplex PCR assay described in this study is a fast, low-cost, efficient and accurate test for rapid screening of *Enterobacteriaceae* isolates with respect to drug resistant genes for *bla*~KPC~, *bla*~VIM,~ *bla*~IMP~, *bla*~NDM~ and *bla*~OXA~ and cover 100% (19/19), 100% (39/39), 85.7% (36/42), 9.1% (1/11) and 8.5% (25/293) of the genes described in the Lahey database, respectively. The proposed untested primers may be also used for detection of the remaining carbapenemases in single and/or multiplex PCR reactions ([Table 1](#tbl36048){ref-type=\"table\"}). The resulting assays could collectively cover 92.9% (39/42), 100% (11/11), 100% (5/5), 100% (4/4), 95.8% (23/24) and 52.6% (154/293) of the *bla*~IMP~, *bla*~NDM~, *bla*~SME~, *bla*~IMI~, *bla*~GES~, *bla*~OXA~ and other genes, respectively.\n\nWe included here all known carbapenemase genes that have been identified in clinical isolates of *Enterobacteriaceae* (*bla*~KPC~, *bla*~OXA,~ *bla*~VIM~, *bla*~NDM~, *bla*~IMP~, *bla*~SME~, *bla*~IMI~, *bla*~GES~, *bla*~GIM~, *bla*~DIM~and *bla*~CMY~). In conclusion, primers tested in silico and in vitro may be used in single and/or multiplex PCR for screening encountered carbapenemases in *Enterobacteriaceae*, as well as for monitoring their emergence and spread (e.g., outbreaks).\n\nMany thanks to Ing. Jaroslav Hrab\u00e1k, Ph.D. for providing bacterial strains with positive carbapenemase production.\n\n**Authors' Contribution:**Patrik Mlynarcik designed the primers; Patrik Mlynarcik and Magdalena Roderova carried out the experiments; Patrik Mlynarcik, Magdalena Roderova and Milan Kolar prepared the manuscript.\n\n**Funding/Support:**This work was supported by project CZ.1.07/2.3.00/30.0004. This work was funded by the European Social Fund in the Czech Republic and by the state budget of the Czech Republic.\n"}
+{"text": "Introduction\n============\n\nObesity is a growing challenge for anesthesiologists because of many problems including an increased incidence of intra- and postoperative atelectasis,^[@B1]^ difficulty in tracheal intubation,^[@B2]^ an increase in airway resistance that may resemble asthma,^[@B3]^ an increased capacity to metabolize anesthetics such as halothane or enflurane^[@B4],[@B5]^ (but not apparently sevoflurane),^[@B6]^ a greater surgical demand for relaxation and decrease in functional residual capacity (FRC).^[@B7],[@B8]^ The obese or overweight patient presents kinetic issues that may delay recovery, thereby, aspiration and develop acute upper airway obstruction after tracheal extubation that add to the risks of anesthesia.^[@B9]-[@B11]^ Rapid recovery is, therefore, desirable to ensure early efficient coughing and to decrease the rate of postoperative respiratory complications.\n\nOne of the elements that determine recovery from anesthesia is the clearance of anesthetic from the effect site. Several factors influence clearance. Anesthetic in tissue depots, the solubility of the anesthetic in blood (the blood/gas partition coefficient) and tissue/blood solubility coefficients^[@B12],[@B13]^ will determine the rate of decrease of anesthetic in the arterial circulation during recovery from anesthesia because solubility determines the clearance of anesthetic at the lungs.^[@B14]^ If the fat solubility of the anesthetic is very small, most of the anesthetic will be cleared by ventilation and will thus not recirculate and delay recovery from anesthesia.\n\nMost Iranian sheep have large, fat tails, which accounts for up to 14.5% of the cold carcass weight.^[@B15]^ The blood supply to the tail originates from the median sacral artery (MSA) which is a branch of the abdominal aorta.^[@B16]^\n\nThe objective of the study reported here was to evaluate effectiveness of body fat on recovery times in obese patients. Therefore, the recovery times of anesthesia by isoflurane in intact fat-tailed lambs and lambs with a ligated MSA was conducted.\n\nMaterials and Methods\n=====================\n\nSeven healthy, 8 to 10-month-old, fat-tailed Lori-Bakhtiyari ewe lambs, with a body condition score of 4 (on a scale of 0 to 5 units)^[@B17]^ and with a mean \u00b1 SD weight of 34.3 \u00b1 1.2 kg, were used in the present study. The overall health of the sheep was monitored before and throughout the study. The animals were kept in barn and received anti-parasitic medications prior to the study and were acclimatized to the experimental conditions for 14 days. Body weight of the animals was recorded at arrival and before induction of the anesthesia. The animals had free access to hay and tap water throughout the study. Wool was clipped a week before the start of experiment.\n\nFood was withheld for 20 hr before induction of the anesthesia and the procedure. However the animals had free access to water. All practical parts of the study, including the ligating of MSA, the induction of inhalation anesthesia, and the monitoring of the animals, intra-operative and recovery parameters were performed with the same author. Sheep were anesthetized twice in a two-week interval (the experiment was performed in two stages).\n\n**Stage 1.**Sheep were anesthetized for 4 hr using isoflurane in oxygen. For induction of the anesthesia sheep were restrained gently and isoflurane in 100% oxygen (4 L min^-1^) was delivered via a fitted facemask without using of sedatives or tranquilizers. Prior to anesthesia, the right jugular vein was catheterized for subsequent fluid administration. For induction of the anesthesia, the concentration of isoflurane was increased gradually (0.5% every 30 sec) until a vaporizer setting of 4.5% was reached.\n\nLidocaine 10% spray (Astra, Sodertalje, Sweden) was used to anesthetize the larynx prior to tracheal intubation. Following intubation, sheep were connected to a re-breathing system and a medium plane of anesthesia, as determined by palpebral and pedal reflexes, was maintained with isoflurane (2.0 to 2.5%) in oxygen (1.5 L min^-1^).\n\n**Stage 2.**Two weeks later, prior to the induction of anesthesia, the MSA of lambs was ligated under epidural analgesia. The level of the lumbosacral junction was prepared aseptically, and the needle placed correctly into the epidural space for injection of 4 to 5 mL of a 2% lidocaine solution (Farvet, Bladel, The Netherlands).\n\nFollowing epidural anesthesia, sheep were positioned in dorsal recumbency to ligate the MSA. Using aseptic technique, a 4 cm midline incision was made at the base of the tail and after subcutaneous dissection the MSA was identified and ligated using No. 0 chromic catgut at its most proximal location ([Figs. 1A and 1B](#F1){ref-type=\"fig\"}). After careful hemostasis the subcutaneous tissue was closed with No. 0 chromic catgut and the skin with No. 1 polypropylene in a simple continuous pattern. An hour and half after epidural anesthesia, all sheep were anesthetized again as mentioned above.\n\nLactated Ringer's solution (10 mL kg^-1^ per hr) was administered during anesthesia in both stages.\n\nHeart rate was obtained using a continuous electro-cardiogram monitor (Model RS-2000; Gmedi Co., Seoul, Korea). Rectal temperature was measured at 1 hr intervals. Sheep were placed in left lateral recumbency on a padded surgical table during the maintenance of anesthesia. At the end of anesthesia, sheep were disconnected from the anesthetic circuit and were allowed to breathe room air. Induction time (time from the administration of isoflurane to tracheal intubation), extubation time (time from the discontinuation of isoflurane to swallowing reflex), time to sternal recumbency (time from the discontinuation of isoflurane to sternal recumbency) and attempts to stand (time from the discontinuation of isoflurane to time to standing) were recorded during anesthetic induction and recovery.\n\n![**A)**Identification of the median sacral artery (MSA) after subcutaneous dissection. **B)** Ligation of the MSA with absorbable suture at its most proximal location.](vrf-6-205-g001){#F1}\n\n**Data analysis.** Experimental results were expressed as mean \u00b1 standard deviation (SD). All data were analyzed by one-way analysis of variance to assess statistical significance between experimental groups using SPSS (Version 17; SPSS Inc., Chicago, USA) and *p* value less than 0.05 was considered significant.\n\nResults\n=======\n\nIt was relatively easy to locate the MSA at the base of the tail. In the stage 2, ligation of the MSA was performed without difficulties and no complications were encountered during surgery or post operation.\n\nMask induction was easily performed in sheep with minimal restraint. The induction of anesthesia was smooth and no struggling or objection to placement of the mask was observed.\n\nTimes to induce, extubate, sternal recumbency and attempts to stand are shown in [Fig. 2](#F2){ref-type=\"fig\"}. No significant differences were observed for the induction time between groups (*p* \\> 0.05). However, extubation, sternal recumbency and attempts to stand times were significantly longer in intact sheep (*p* \\< 0.05).\n\nIn both stages significant decreases in rectal temperature were observed in both groups, no significant changes in heart rate or rhythm were detected ([Table 1](#T1){ref-type=\"table\"}).\n\nIncreased salivation was occurred in most sheep during isoflurane anesthesia. One sheep in the control group showed moderate bloat during the first stage of the experiment. All sheep were recovered uneventfully from anesthesia and no adverse reactions or complications were encountered in the present study.\n\n###### \n\nMean (SD) cardiopulmonary measurements and body temperatures of seven sheep anesthetized with isoflurane at intervals after the induction of anesthesia in two stages\n\n  **Groups**   **Parameters**   **Time from induction of anesthesia (Min)**                               \n  ------------ ---------------- --------------------------------------------- ------------- ------------- -------------\n  Stage 1      HR (bpm)         82.0 \u00b1 15.5                                   78.5 \u00b1 13.5   77.0 \u00b1 14.5   75.2 \u00b1 13.7\n  Stage 1      TEMP (\u02daC)        38.6 \u00b1 0.5                                    38.0 \u00b1 0.7    37.4 \u00b1 0.7    37.0 \u00b1 0.7\n  Stage 2      HR (bpm)         83.0 \u00b1 16                                     80.3 \u00b1 13.2   78.0 \u00b1 15.1   76.0 \u00b1 13.5\n  Stage 2      TEMP (\u02daC)        38.7 \u00b1 0.5                                    38.1 \u00b1 0.6    37.3 \u00b1 0.7    37.1 \u00b1 0.6\n\nHR: Heart rate, TEMP: Temperature.\n\nThere were no significant differences between these data (*p*\\> 0.05).\n\n![Induction and recovery times in stage 1 (n=7) and stage 2, median sacral artery (MSA)-ligated sheep (n=7) following 4 hr isoflurane anesthesia.](vrf-6-205-g002){#F2}\n\nDiscussion\n==========\n\nThe results of this study supported the hypothesis that body fat is effective on recovery times in obese patients after prolonged periods of anesthesia. In the present study, anesthesia was induced with isoflurane delivered by mask, so any possible effects of premedications or injectable anesthetic agents on recovery times could be eliminated. Theoretical considerations suggest that if the inhalation anesthetics had been used without other medication, recovery should have been faster with desflurane, followed by sevoflurane and slower with isoflurane, but in experimental study Mohammadnia *et al*. compared the three inhalation agents in sheep after injection of xylazine as premedication agent that there were no differences between desflurane, sevoflurane and isoflurane in time of recovery.^[@B19]^ Although we did not measure the inhaled concentration of isoflurane, the vaporizer setting and the depth of anesthesia was the same in both groups. Because no surgery was performed during the anesthesia, it was easy to maintain a constant light level of anesthesia throughout the 4 hr period of anesthesia.\n\nFat bears a large capacity for anesthetic storage and may serve as a reservoir in obese patients. The impact of anesthetic stored in fat may be the result of a return of the anesthetic in blood perfusing the fat or of a transfer from fat to adjacent highly perfused tissues (e.g., omental/ mesenteric fat to intestine and liver).^[@B14]^ Such accumulation may delay recovery from anesthesia.\n\nOn average, the tail weight is 12.6% of carcass weight in Lori-Bakhtiyari rams.^[@B15]^ Therefore, the role of approximately 4.25 kg fat in anesthetic storage was abated by MSA-ligation. During the two weeks (Stage 1 to 2), on average, one kg was added to bodyweight of each sheep.\n\nIn this study the indices used to assess recovery were extubation, sternal recumbency and attempts to stand. For all species these include extubation,^[@B18]^ however, in human being they include awareness factors.^[@B1]^ In most veterinary studies final recovery is taken as the ability of the animal to stand, unless, standing might be limited such as orthopedic reasons.^[@B18],[@B19]^ Final recovery in the present study was taken as attempts to stand .When recovering from anesthesia, sheep should not be left unattended until they can safely maintain themselves in sternal recumbency; preliminary pilot studies showed that sheep could do attempts to stand, once they were sufficiently awake.\n\nTime to extubate, sternal recumbency and attempts to stand were significantly shorter in MSA-ligated sheep. Also, in our previous experimental study, we showed that the peak serum fluoride concentration and recovery time from anesthesia with halothane were decreased following ligation of MSA in fat-tailed sheep.^[@B20]^ Shorter recovery time has been reported in sheep without a fat tail.^[@B21]^ The more rapid recovery times may be attributed to the faster decline in the alveolar partial pressure of isoflurane in MSA-ligated sheep. We hypothesize that the rate of alveolar washout is faster in MSA-ligated sheep. Increased recovery time has also been observed in obese human patients compared with non-obese patients.^[@B22]^ Also Juvin *et al*.,^[@B23]^ showed that recovery was more rapid in such patients when they were anesthetized with desflurane versus isoflurane.\n\nThe blood-gas partition coefficient represents the partial pressure or solubility of an anesthetic between blood and gas at equilibrium. The higher the value of blood-gas partition coefficient, the greater the solubility of an anesthetic in the blood and vice versa and the lower the value, the lower the blood solubility of the anesthetic.^[@B19]^\n\nIsoflurane is available as a clear nonflammable, halogenated methyl ethyl ether with a blood/gas partition coefficient of 1.4 and fat/blood partition coefficient of 45.0. Its intermediate blood solubility rating along with its relatively high lipophilicity indicates that this agent will produce a fairly rapid induction.^[@B24]^\n\nIntermediate fat solubility of isoflurane (fat/gas partition coefficient approximately 45.0) limits elimination into alveolar gas. This implies that the total amount of isoflurane extracted by fat tissue is higher in intact sheep compared with MSA-ligated sheep.\n\nStudies showed that the newest volatile anesthetics such desflurane and sevoflurane, have significantly lower blood/gas partition coefficients (0.4 and 0.6, desflurane versus sevoflurane, respectively) than isoflurane (1.4) or halothane (2.4), predicting better intraoperative control of anesthesia and a more rapid recovery from anesthesia.^[@B12],[@B13],[@B25],[@B26]^\n\nIn conclusion, recovery times decreased following MSA ligation in fat-tailed sheep, which suggested that body fat had a major role in recovery times of isoflurane in sheep after prolonged periods of anesthesia. We developed an animal model to investigate fat drug solubility of isoflurane gas. Therefore, using of less-soluble in fat anesthetics is better than high-soluble anesthetics for prolonged anesthesia to decrease postoperative complication.\n\nThis research was supported by School of Veterinary Medicine, Shahrekord University, Shahrekord, Iran. The authors would like to thank Mr. Jamshid Kabiry for his expert technical help and Mr. Ezzatollah Ghaedi for his assistance with animal care during the course of the study.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nDuring the last decade the emergence of carbapenemase-producing strains among Enterobacteriaceae,*Pseudomonas* spp., and*Acinetobacter baumannii* is remarkable. A variety of carbapenemases have been reported such as the Ambler class A KPC-type (mostly identified in Enterobacteriaceae and*Pseudomonas aeruginosa*) and GES-type (mostly in*A. baumannii*), the Ambler class B metallo-*\u03b2*-lactamases (MBL) of VIM-, IMP-, GIM-, and NDM-types, and the Ambler class D carbapenemases of the OXA-48 type in Enterobacteriaceae and of OXA-23, OXA-24/-40, OXA-58, and OXA-143 types in*Acinetobacter* spp. The emergence of the most recently described carbapenemase, namely, the New Delhi metallo-*\u03b2*-lactamase (NDM-1), constitutes a critical medical issue. Indeed, this enzyme compromises the efficacy of almost all *\u03b2*-lactams (except aztreonam), including the last resort carbapenems. Although most of the NDM-producing strains identified are Enterobacteriaceae, this carbapenemase has also been reported from*Acinetobacter* spp. and more rarely from*P. aeruginosa*, both species causing severe nosocomial infections, including urinary tract infections, peritonitis, septicemia, and pulmonary infections. The Indian subcontinent, the Balkans regions, and the Middle East are considered to be the main reservoirs of NDM producers. Since therapeutical options are limited to very few antibiotics such as colistin, tigecycline, and fosfomycin, hospital- and community-acquired infections caused by NDM-1 producers are difficult to eradicate. Isolation of infected patients and carriers and rapid diagnostic techniques are the key factors that contribute to contain this outbreak that threatens the efficacy of the modern medicine.\n\n2. Clinical Impact of the Antibiotic Resistance Patterns of NDM Producers for the Treatment {#sec2}\n===========================================================================================\n\nCurrently, one of the most clinically significant carbapenemase is the recently described NDM-1 (New Delhi metallo-*\u03b2*-lactamase). This carbapenemase belongs to the class B of Ambler *\u03b2*-lactamases classification that includes the metallo-*\u03b2*-lactamases (MBLs). NDM-1 shares very little identity with other MBLs, the most similar being VIM-1/VIM-2 with only 32.4% amino acid identity. Compared to VIM-2, NDM-1 displays tighter binding to most cephalosporins, in particular to cefuroxime (*K* ~mNDM-1~ = 8\u2009*\u03bc*M, *K* ~mVIM-2~ = 22\u2009*\u03bc*M), cefotaxime (*K* ~mNDM-1~ = 10\u2009*\u03bc*M, *K* ~mVIM-2~ = 32\u2009*\u03bc*M), cephalothin (*K* ~mNDM-1~ = 10\u2009*\u03bc*M, *K* ~mVIM-2~ = 44\u2009*\u03bc*M), and penicillins (*K* ~mNDM-1~ = 16\u2009*\u03bc*M, *K* ~mVIM-2~ = 49\u2009*\u03bc*M). Like all other MBLs, NDM-1 efficiently hydrolyses a broad range of *\u03b2*-lactams including penicillins, cephalosporins, and carbapenems, just sparing monobactams such as aztreonam. NDM-1 does not bind to carbapenems as tightly as IMP-1 or VIM-2 does, and the turnover rate of carbapenem hydrolysis is similar to that of VIM-2 (*k* ~cat~/*K* ~m~ are 0.21, 1.2, and 0.99\u2009s^\u22121^\u00b7*\u03bc*M^\u22121^ for NDM-1, IMP-1, and VIM-2, resp.). Similar to the other MBLs, the active site of NDM-1 contains two metal ion binding sites: the His and Cys sites. Accordingly, a 3D-structure modelling of the NDM-1 enzyme showed that two zinc ions were present at both the His and Cys sites with a distance of 4.20\u2009\u00c5 \\[[@B1]\\]. Indeed, the hydrolysis activity of MBLs depends on the interaction of the *\u03b2*-lactam molecule with Zn^2+^ ion(s) in their active site. Consequently, their activity is inhibited by chelators of divalent cations, such as EDTA. Accordingly, the efficacy of EDTA (Ca-EDTA) has been evaluated in a mouse model of sepsis caused by an NDM-1-producing*Escherichia coli*. It has been shown that a combination therapy using imipenem/cilastatin sodium (IPM/CS) and Ca-EDTA reduced the bacterial inoculum, as compared to IPM/CS alone suggesting the possibility to use Ca-EDTA in clinical therapeutics \\[[@B2]\\]. Comparison of IMP-1, VIM-2, and NDM-1 by an*in silico* approach revealed that NDM-1 might have greater drug profile and catalytic efficiency than IMP-1 and VIM-2 due to a larger pocket opening and a lower distance between the Zn-I ion and *\u03b2*-lactam oxygen of the carbapenem \\[[@B3]\\].\n\nIt is noteworthy that a quite systematic association with other antibiotic resistance determinants is observed in almost all NDM producers (Enterobacteriaceae,*Acinetobacter,* and*Pseudomonas*). Those associated resistance determinants are AmpC cephalosporinases, clavulanic acid inhibited expanded-spectrum *\u03b2*-lactamases (ESBLs), other types of carbapenemases (OXA-48-, VIM-, and KPC-types), and resistance to aminoglycosides (16S RNA methylases), to quinolones (Qnr), to macrolides (esterases), to rifampicin (rifampicin-modifying enzymes), to chloramphenicol, and to sulfamethoxazole \\[[@B4]--[@B9]\\]. Consequently, most of the NDM-1 producers remain susceptible only to two bactericidal antibiotics (colistin and fosfomycin) and a single bacteriostatic antibiotic (tigecycline) \\[[@B10], [@B11]\\] ([Figure 1](#fig1){ref-type=\"fig\"}).*In vitro* synergy combination assays performed with NDM-1 producers with those three antibiotic molecules showed a synergistic activity of colistin and fosfomycin, of colistin and tigecycline in rare cases, whereas most of the antibiotic associations remain neutral for most of the tested isolates \\[[@B12]\\]. Since NDM-1 does not hydrolyze aztreonam, a combination therapy including aztreonam and avibactam (also named NXL-104), a novel serine *\u03b2*-lactamase inhibitor inhibiting the most frequent broad-spectrum hydrolyzing-*\u03b2*-lactamases hydrolyzing aztreonam has been suggested as a possible strategy against NDM-1-producing Enterobacteriaceae. This therapeutic option seems to be a very efficient combination therapy*in vitro*\\[[@B13], [@B14]\\].\n\n3. Infections Caused by NDM Producers {#sec3}\n=====================================\n\nSince NDM producers were mainly described in Enterobacteriaceae, infections caused by NDM producers include urinary tract infections, peritonitis, septicemia, pulmonary infections, soft tissue infections, and device-associated infections. As observed for other multidrug-resistant bacteria, it is highly probable that colonization of the gut flora might precede the infection by NDM producers and orofecal transmission in the community might occur mostly through hand contamination, food, and water. Among the NDM-1-producing Enterobacteriaceae,*Klebsiella pneumoniae* and*E. coli* are the most often described species. Both hospital- and community-acquired infections have been reported. However, this carbapenemase is also frequently described in other enterobacterial species including*Klebsiella oxytoca*,*Enterobacter cloacae*,*Citrobacter freundii*,*Proteus mirabilis*,*Salmonella* spp., and*Providencia* spp. Although most of NDM-producing bacteria are Enterobacteriaceae, this carbapenemase was also reported from*Acinetobacter* spp. \\[[@B15]--[@B32]\\] and in rare cases*Pseudomonas aeruginosa* \\[[@B33], [@B34]\\].\n\nSince, no specific virulence factor is known to be associated with *bla* ~NDM-1~-carrying plasmids \\[[@B6], [@B35]--[@B39]\\], there is no evidence that NDM-producing bacteria are more virulent than other strains \\[[@B40]--[@B42]\\]. However, some rare isolates of NDM-1-producing virulent enteric bacteria such as*Salmonella* \\[[@B43]--[@B45]\\] and*Vibrio cholerae*\\[[@B46], [@B47]\\] have been described.\n\n4. Epidemiology of NDM-Producing Bacteria {#sec4}\n=========================================\n\nNDM-1 was first identified in 2008 in a*K. pneumoniae* isolate recovered from a Swedish patient who has been previously hospitalized in New Delhi, India \\[[@B48]\\]. Since then, NDM carbapenemases are the focus of worldwide attention due to the rapid dissemination of the corresponding gene among Enterobacteriaceae and*Acinetobacter* spp. mainly ([Figure 2](#fig2){ref-type=\"fig\"}). Rapidly, a link between NDM-producing Enterobacteriaceae and the Indian subcontinent has been pointed out \\[[@B49]--[@B51]\\], and prevalence rates of NDM-producing Enterobacteriaceae were found to range from 5 to 18.5% in Indian and Pakistan hospitals \\[[@B52]--[@B55]\\]. In addition, the *bla* ~NDM-1~ gene was detected not only in patient samples, but also in drinking water and seepage samples in New Delhi \\[[@B47]\\]. The occurrence of NDM-1-producing bacteria in environmental samples in New Delhi is significant for people living in the city who often rely onto public water and poor sanitation facilities. A secondary reservoir of NDM-1 producers was then highlighted through several studies reporting patients colonized or infected with NDM-1 producers originating from the Balkan states \\[[@B50], [@B56]--[@B61]\\]. Recent reports also suggested that the Middle East might be an additional reservoir of NDM producers \\[[@B62]--[@B67]\\]. This dissemination of NDM producers in the Middle East could mostly be linked to the population exchange between the Middle East and the Indian subcontinent. However, NDM-1 producing bacteria have now been reported worldwide with a rapid dissemination from the two previously described reservoirs, namely, the Indian subcontinent and the Balkan countries.\n\nAs observed with the dissemination of NDM-1-producing Enterobacteriaceae, NDM-producing*Acinetobacter* has also been recovered from environmental samples in China \\[[@B31]\\]. Currently, the majority of NDM-producing*Acinetobacter* spp. are reported from China \\[[@B18], [@B19], [@B25], [@B29]--[@B32]\\] and Middle East \\[[@B17], [@B20]--[@B24], [@B26]\\].\n\n5. Genetic Features of the *bla* ~NDM~ Genes {#sec5}\n============================================\n\nIn Enterobacteriaceae, the *bla* ~NDM-1~ gene is located mostly onto conjugative plasmids belonging to several incompatibility groups \\[[@B6], [@B7], [@B25], [@B35], [@B37], [@B38], [@B63], [@B68]\\]. However, investigation of a worldwide collectionof NDM-1-producing enterobacterial isolates showed that the current spread of the *bla* ~NDM-1~ gene is not related to the spread of specific clones, specific plasmids, or single genetic structure \\[[@B7]\\]. In*Acinetobacter* spp. the *bla* ~NDM~-type genes are found to be either plasmid- or chromosome-located, and in the rare NDM-1-producing*P. aeruginosa*, the *bla* ~NDM-1~ gene was found to be chromosomally located \\[[@B33], [@B34]\\]. Investigations on the immediate genetic environment of *bla* ~NDM~ genes revealed the presence of a conserved structure that always associated the complete or truncated insertion sequence IS*Aba125*at the 5\u2032-end and the *ble* ~MBL~ gene (encoding resistance to the anticancer drug bleomycin) at the 3\u2032-end of the *bla* ~NDM~ genes \\[[@B69]\\] ([Figure 3](#fig3){ref-type=\"fig\"}). In addition, in several studies focusing on NDM-producing*A. baumannii*, the *bla* ~NDM~ gene was located between two copies of the IS*Aba125* element, forming a composite transposon named Tn*125* \\[[@B15]--[@B17], [@B20], [@B24], [@B26], [@B70], [@B71]\\] ([Figure 3](#fig3){ref-type=\"fig\"}). Systematic identification of a truncated form of this composite transposon in Enterobacteriaceae, while it was described in its entire form in*A. baumannii*, strongly suggesting that*Acinetobacter* spp. has been a reservoir of those *bla* ~NDM~ genes before targeting enterobacterial species. Those findings highlight that even though*A. baumannii* is usually recognized as a final acceptor for resistance genes, it may acquire several resistance determinants and then transfer them to Enterobacteriaceae and*Pseudomonas* spp.\n\n6. NDM Variants {#sec6}\n===============\n\nSince the first description of NDM-1, eight variants of this enzyme have been published (NDM-1 to -8) ([Figure 4](#fig4){ref-type=\"fig\"}) and ten have been assigned (<http://www.lahey.org>). The first variant NDM-2 is a point mutation variant having a C to G substitution at position 82 resulting in an amino acid substitution of a proline to an alanine residue at position 28 (Pro \u2192 Ala) ([Figure 4](#fig4){ref-type=\"fig\"}) \\[[@B26]\\]. Considering that this point mutation was located at the last amino acid of the peptide leader of the enzyme, MIC values of *\u03b2*-lactams including carbapenems showed no significant difference between NDM-1 and NDM-2 producers. NDM-2 has been identified in several*A. baumannii* strains \\[[@B20]--[@B22], [@B26]\\] but not yet in Enterobacteriaceae. The NDM-3 variant was described from an*E. coli* isolate and differs from NDM-1 by a single nucleotide change conferring a peptide sequence change at position 95 (Asp \u2192 Asn) that does not modify the hydrolytic activities of the enzyme ([Figure 4](#fig4){ref-type=\"fig\"}) \\[[@B11]\\]. The NDM-4 variant differs from NDM-1 by a single amino acid substitution at position 154 (Met \u2192 Leu) ([Figure 4](#fig4){ref-type=\"fig\"}). Kinetic data showed that this amino acid substitution is responsible for an increased hydrolytic activity of NDM-4 compared to NDM-1 toward cefalotin, ceftazidime, cefotaxime, imipenem, and meropenem, whereas cefepime was less hydrolyzed \\[[@B72]\\]. The NDM-5 variant shares the substitution at positions 154 (Met \u2192 Leu) with NDM-4, conferring enhanced hydrolytic activity against carbapenems and harbors a second amino acid substitution at position 88 (Val \u2192 Leu) ([Figure 4](#fig4){ref-type=\"fig\"}) \\[[@B73]\\]. The NDM-6 variant differs from NDM-1 by a single amino acid substitution at position 233 (Ala \u2192 Val) leading to no obvious modification in the hydrolytic activity of the enzyme ([Figure 4](#fig4){ref-type=\"fig\"}) \\[[@B74]\\]. The NDM-7 variant was concomitantly described from an*E. coli* isolate recovered from a French patient who had travelled to Burma \\[[@B75]\\] and an*E. coli* isolate recovered from a Yemeni patient previously hospitalized at the Frankfurt University Hospital in Germany \\[[@B76]\\]. The *bla* ~NDM-7~ gene differs from *bla* ~NDM-1~ by two amino-acid substitutions at positions 388 (G\u2192A) and 460 (A\u2192C) corresponding to amino acid substitutions at position 130 (Asp \u2192 Asn) and 154 (Met \u2192 Leu), respectively ([Figure 4](#fig4){ref-type=\"fig\"}). The amino acid substitution at position 154 (Met \u2192 Leu) increases the hydrolysis activity of the enzyme \\[[@B75], [@B76]\\]. The amino acid sequence of the last published NDM variant, namely, NDM-8, has substitutions at positions 130 (Asp \u2192 Gly) and 154 (Met \u2192 Leu) compared with NDM-1 ([Figure 4](#fig4){ref-type=\"fig\"}) \\[[@B77]\\]. This NDM variant possesses the amino acid substitution at position 154 (Met \u2192 Leu), but its critical impact on *\u03b2*-lactams hydrolysis has not been detailed.\n\n7. Identification of NDM Producers {#sec7}\n==================================\n\nDetection of carbapenemase producers, including NDM producers, in clinical specimens is based currently on a preliminary analysis of susceptibility testing results. The US guidelines (CLSI) (updated in 2013) retained as breakpoints for Enterobacteriaceae susceptibility (*S*) \u2264 1 and resistance (*R*) \\> 4\u2009mg/L for imipenem and meropenem and *S* \u2264 0.5 and *R* \\> 2\u2009mg/L for ertapenem. The European guidelines (EUCAST) (updated in 2013) are slightly different and propose breakpoints for imipenem and meropenem as follows: susceptible (*S*) \u2264 2 and resistant (*R*) \\> 8\u2009mg/L and for ertapenem *S* \u2264 0.5 and *R* \\> 1\u2009mg/L. Although some discrepancies might exist for several isolates depending on the reference used to interpret the antibiogram, MIC values of ertapenem are often higher than those of other carbapenems with NDM producers. Consequently, ertapenem would be the best molecule for suspecting most of the carbapenemase producers, including NDM producers, and constitutes good screening criteria. Notably, this greater sensitivity of ertapenem compared to the other carbapenems is counterbalanced by its lower specificity. Of note, susceptibility to carbapenems is observed for some NDM producers and additional tests for carbapenemase detection are needed to detect them accurately.\n\n7.1. Detection of a Carbapenemase Activity {#sec7.1}\n------------------------------------------\n\nOne of the commonly used techniques is the modified Hodge test (MHT), which has been used for years. Unfortunately, the MHT has been proved to lack sensitivity (50%) for detecting NDM-1 producers. Of note, ZnSO~4~ (100\u2009*\u03bc*g/mL) supplementation in the culture medium significantly increases the sensitivity to 85.7% \\[[@B78]\\]. However, this test has a low specificity with*Enterobacter* spp. often overexpressing their chromosomal cephalosporinase \\[[@B79]\\]. In addition, results of the MHT are obtained at least 72\u2009h after the bacterial identification.\n\nUV spectrophotometry analysis of carbapenem hydrolysis has been developed to detect carbapenem hydrolysis. This method is based on the detection of the decrease of imipenem absorbance with crude extracts of bacterial enzymes. Crude extracts can be obtained from an overnight culture of the tested strain after mechanical lysis. This UV spectrophotometry-based technique is cheap and has a 100% sensitivity and a 98.5% specificity for detecting carbapenemase activity \\[[@B80]\\]. However, it is time-consuming and requires trained microbiologists and expensive equipment.\n\nAnalysis of carbapenem hydrolysis by using the MALDI-TOF technology has been shown to be a useful technique to detect carbapenemase production in a few hours. This technique was based on detection of a carbapenem (imipenem, meropenem, or ertapenem) spectrum and of its main derivatives resulting from carbapenem hydrolysis. After 3 to 4 hours of incubation of the tested isolate with a carbapenem, the bacteria were pelleted by centrifugation and the supernatant containing the carbapenem and its metabolites was tested by MALDI-TOF mass spectrophotometry. Disappearance of the peak corresponding to the native carbapenem and appearance of peak(s) corresponding to the metabolite(s) resulting on the carbapenem hydrolysis sign a carbapenemase activity \\[[@B81]--[@B85]\\]. This test has excellent sensitivity and specificity. However, it again requires trained microbiologists and expensive equipment.\n\nThe most promising technique is the rapid Carba NP test. It is based on the detection of the hydrolysis of imipenem by a color change of a pH indicator ([Figure 5](#fig5){ref-type=\"fig\"}). This test is 100% sensitive and 100% specific for the detection of any type of carbapemenase produced by Enterobacteriaceae including NDM producers \\[[@B86]--[@B88]\\]. The Carba NP test has been also validated for the detection of most carbapenemase-producing*Pseudomonas* spp., including all NDM producers \\[[@B89]\\]. A second version of the Carba NP test (the Carba NP test II) has been developed to rapidly differentiate between the diverse carbapenemase types found in Enterobacteriaceae and*P. aeruginosa*. This Carba NP test II combines the inhibition properties of EDTA with the high efficiency of the Carba NP test for identification of any type of MBL producer, including all NDM producers \\[[@B90]\\]. Recently, the Carba NP test has been evaluated to detect carbapenemase-producing Enterobacteriaceae (*n* = 193) directly from spiked blood cultures. The proposed strategy allows detection of all NDM producers (*n* = 33) in less than 5 hours, with sensitivity and specificity of 100%, respectively \\[[@B91]\\]. This test has excellent sensitivity and specificity. However, it requires homemade reagents that are not yet commercially available.\n\n7.2. Phenotypic Detection of Metallo-*\u03b2*-Lactamase (MBL) Producing Isolates {#sec7.2}\n---------------------------------------------------------------------------\n\nDetection methods based on the inhibitory properties of several divalent ions chelators (e.g., EDTA and dipicolinic acid) may identify MBL producers. A disk-diffusion test based on the detection of a synergy between a carbapenem-containing disk (imipenem or meropenem) and a disk containing an MBL inhibitor (EDTA or mercaptopropionic acid or dipicolinic acid) has been proposed \\[[@B92]\\].\n\nA combined disk technique using a carbapenem disk and the same carbapenem disk supplemented with EDTA (10\u2009*\u03bc*L of a 0.1\u2009M solution at pH 8) has been also proposed \\[[@B93]\\]. Using this test, a 5\u2009mm increase of the inhibition diameter around the disk containing imipenem plus EDTA compared to imipenem alone likely indicates the production of a MBL. However, those two phenotypic methods are time-consuming and false-negative results often arise, in particular when low level of resistance is observed \\[[@B93]\\].\n\nAmong those phenotypic methods, the Etest MBL strip, a two-sided strip containing gradients of imipenem alone on one side and imipenem supplemented with EDTA on the other side, is also commonly used for the detection of MBL producers. Using this test, at least three doubling dilutions of the MIC in the presence of EDTA are considered as a positive result \\[[@B94]\\]. However, several NDM-producing isolates exhibit low MIC of carbapenems, leading to not interpretable results using the Etest MBL strip.\n\n7.3. Molecular Detection of NDM Producers {#sec7.3}\n-----------------------------------------\n\nAll the previous techniques can detect the carbapenemase production and, in some cases more precisely, production of an MBL, but none of them is able to specifically identify an NDM enzyme or its corresponding gene. Therefore a number of genotypic approaches have been reported, based on PCR techniques, including real-time PCR methods able to detect *bla* ~NDM~-positive isolates directly from clinical samples \\[[@B95], [@B96]\\]. Those methods, however, have the disadvantage to be unable to identify any novel carbapenemase gene and are quite expensive. Commercial DNA microarray methods are marketed and increase the convenience of those tests \\[[@B97]\\]. Although they cannot overcome general limitations of genotypic techniques those DNA microarrays are able to identify the presence of carbapenemase and the main extended-spectrum *\u03b2*-lactamase and acquired cephalosporinases genes. Accordingly, this technique is more adapted for an epidemiological purpose in order to control an outbreak. Finally, molecular amplification of the *bla* ~NDM~ gene followed by sequencing is needed to identify the exact nature of the NDM variant.\n\n8. Detection of Infected and Colonized Patients {#sec8}\n===============================================\n\nSince the prevention of dissemination of carbapenemase producers partially relies on an early and accurate detection of carriers, recommendations for the screening of colonized patients have been introduced in several countries. Commonly, \"at-risk\" patients, meaning those being colonized with carbapenemase producers, are patients transferred from a foreign hospital and those hospitalized in intensive-care units, in transplantation units, and immunocompromised patients.\n\nSince the intestinal flora is the main reservoir of Enterobacteriaceae, rectal swabs and stools are the most suitable clinical samples for performing this screening. These specimens may be plated on screening medium, either directly or after an enrichment step in broth containing imipenem 0.5--1\u2009*\u03bc*g/mL or ertapenem 0.5\u2009*\u03bc*g/mL \\[[@B98], [@B99]\\]. In outbreak situations, this enrichment step might increase the sensitivity of the screening and consequently reduce the number of potential false-negative results by increasing the inoculum of the targeted strain. On the opposite, its disadvantage is the induced delay (12--24\u2009h) needed to confirm or reject carbapenemase detection. Although the efficiency of this enrichment step has not been evaluated for NDM producers, it has already been shown to improve the detection of KPC producers.\n\nRegardless of the enrichment step, the specimens have to be plated on selective media. For that purpose, several screening media have been evaluated and compared to the screening of carriers of NDM producers. One of the first tested medium was the ChromID ESBL culture medium (bioM\u00e9rieux) containing cefpodoxime used as a selector and which is routinely used to screen ESBL producers. Since NDM enzymes have a broad-spectrum activity, they hydrolyze not only carbapenems but also expanded-spectrum cephalosporins very efficiently. Therefore, detection of NDM-producing isolates using ChromID ESBL (aimed do detect ESBL producers) is possible but with a low specificity since the selective agent is a cephalosporin and not a specific carbapenemase substrate (e.g., a carbapenem). Several media supplemented with a carbapenem have been developed and marketed for the screening of carbapenemase producers. The first screening medium targeting KPC producers was the CHROMagar KPC medium that contains meropenem (CHROMagar, Paris, France) \\[[@B100]\\]. Using this medium, carbapenem-resistant bacteria are well detected when they exhibit relatively high-level resistance to carbapenems. Its main disadvantage remains in its lack of sensitivity, since it does not detect carbapenemase producers with low-level carbapenem resistance. Indeed, although NDM producers have often high level resistance to carbapenems, several isolates that exhibited MICs comprised between 0.5 and 1\u2009*\u03bc*g/mL, making their detection difficult on screening media containing high concentration of carbapenems \\[[@B78], [@B93], [@B101]\\]. Colorex KPC (E&O laboratories, Bonnybridge, UK), another screening medium for carbapenemase producers, also contains meropenem. Since the content of this medium is reported to be identical to that of CHROMagar KPC, only NDM producers with high-level resistance to carbapenems may be detected, leading to an accurate detection of 57% to 64% of NDM-producing Enterobacteriaceae using this medium \\[[@B55], [@B102]\\]. A third commercially available screening medium also contains a carbapenem (CRE Brilliance, Thermo Fisher Scientific, UK). Depending on the study, sensitivities for detection of patients colonized with NDM producers were reported to be 63% to 85% using this medium \\[[@B102], [@B103]\\]. Another screening medium also containing a carbapenem is the ChromID CARBA (bioM\u00e9rieux, La Balmes-les-Grottes, France). This commercially available medium has been reported to be more sensitive (87.5% to 94%) than the others for the detection of NDM-producing Enterobacteriaceae \\[[@B55], [@B102], [@B103]\\]. Finally, a homemade screening medium containing ertapenem, cloxacillin, and zinc, namely, the SUPERCARBA medium, has an excellent sensitivity and specificity for the detection of carbapenemase producers, including NDM producers. The zinc supplementation and the low ertapenem concentration allow the efficient detection of all NDM producers regardless of their level of resistance to carbapenems \\[[@B101], [@B104]\\]. Consequently, using the SUPERCARBA medium and performing the Carba NP test on isolated colonies might be proposed as the recommended strategy for screening of carbapenemase producers \\[[@B105]\\].\n\nIn order to avoid the additional 24\u2009h to 48\u2009h before the carriage status of the patient can be established using those screening media an in-house quantitative real-time PCR assay using the TaqMan chemistry has been developed to detect the NDM-encoding genes directly from spiked stool samples. The bacterial extraction from stool samples was performed manually or adapted to a fully automated extraction system. This assay was found to be 100% specific and sensitive with detection limits reproducible below 1 \u00d7 10^1^\u2009CFU/100\u2009mg of feces \\[[@B95]\\]. However, this technology remains expensive and is thus considered to be a valuable tool in the follow-up of an outbreak and cohorting of colonized patients.\n\n9. Conclusion {#sec9}\n=============\n\nThe rapid diffusion of NDM producers is of particular concern since this now corresponds to a worldwide-located outbreak. Additional knowledge in relation to the driven forces behind the spread of those multidrug-resistant isolates is now required, in order to better understand the dynamics of the NDM producers.\n\nAmong the most important features of NDM producers, one may retain that those NDM producers are not only nosocomial enterobacterial pathogens, but also community-acquired Enterobacteriaceae or other Gram-negative species, such as*A. baumannii*. Such diffusion pattern of multidrug resistance is unique for NDM producers and not observed currently for none of the producers of other types of carbapenemases (OXA-48, KPC, IMP...).\n\nThe reservoir of those NDM producers is mainly located in Southeast Asia where the rate of carriers is estimated to be ca. 20%. However, it is difficult to predict the number of tourists in India, Pakistan, and Bangladesh (more than 10 million in 2012) that will bring back those NDM producers at a carrier stage in stools. The size of that reservoir may explain the rapidity of the dissemination of NDM producers worldwide. Accordingly, NDM producers are now on the top list of carbapenemase producers in European countries such as the UK and even in France.\n\nDue to the population demography of the NDM reservoirs coupled with the difficulties to implement hygiene measures or an efficient antibiotic stewardship program in those countries, the outbreak will not stop spontaneously. On the opposite, we will see an acceleration of the NDM producers spread worldwide. Hopefully the rate of NDM producers will not reach those currently observed for ESBL producers (20--80% worldwide). The spread of ESBL producers is an important driving force for usage of carbapenems that will enhance selection of carbapenemase producers. The only significant action we may actually take currently in Western countries which are not endemic for carbapenemase producers is to sustain the wide usage of rapid detection techniques now available and the extended screening of potential carriers at least in hospitals.\n\nIt is likely that novel antibiotic molecules such as the combination of aztreonam and avibactam may be launched in a near future and may bring some therapeutical openings. However there is an urgent need to finance novel research programs for discovering novel anti-Gram negatives molecules and to implement worldwide-located surveillance network of multidrug antibiotic resistance with sentinel labs.\n\nConflict of Interests\n=====================\n\nThe authors declare that there is no conflict of interests regarding the publication of this paper.\n\n![Antibiogram of a NDM-1-producing*K. pneumoniae* isolate. The *bla* ~NDM-1~ gene was located onto a IncHIIB plasmid of ca. \\~200\u2009kb in that strain that also harbored two additional *\u03b2*-lactamase genes (*bla* ~CTX-M-15~, *bla* ~SHV-12~, *bla* ~OXA-1~) and an aminoglycoside methylase (*armA*) responsible for high-level resistance to all aminoglycosides. PTZ, piperacillin + tazobactam; PIP, piperacillin; TIC, ticarcillin; AMX, amoxicillin; ETP, ertapenem; TCC, ticarcillin + clavulanic acid; CAZ, ceftazidime; CF, cefalotin; FOX, cefoxitin; IMP, imipenem; AMC, amoxicillin + clavulanic acid; CTX, cefotaxime; CMX, cefuroxime; MEM, meropenem; ATM, aztreonam; FEP, cefepime; FT, nitrofurantoin; NOR, norfloxacin; OFX, ofloxacin; CIP, ciprofloxacin; FOS, fosfomycin; TGC, tigecycline; TE, tetracycline; CS, colistin; SSS, sulfonamide; SXT, sulfamethoxazole + trimethoprim; C, chloramphenicol; NET, netilmicin; GM, gentamicin; AN, amikacin; TM, tobramycin; RA, rifampicin.](BMRI2014-249856.001){#fig1}\n\n![Geographical distribution of NDM producers.](BMRI2014-249856.002){#fig2}\n\n![Schematic representation of *bla* ~NDM~-associated genetic structures identified among Gram-negative clinical isolates. (a) Structure found in*A. baumannii*, where the *bla* ~NDM~ gene is part of the composite transposon Tn*125.* (b) Structures found in Enterobacteriaceae and*P. aeruginosa* where IS*Aba125* is presented as full or truncated element with *ble* ~MBL~ gene (bleomycin resistance encoding gene) also being present as full or truncated gene. Genes and their corresponding transcription orientations are represented by horizontal arrows.*ori*IS of IS*CR21*is indicated by a circle. The *bla* ~NDM~ promoter is indicated (P). IS, insertion sequence; gene names are abbreviated according to their corresponding proteins: *ble* ~MBL~, bleomycin resistance gene; \u0394*iso* for truncated phosphoribosylanthranilate isomerase; \u0394*pac* for truncated phospholipid acetyltransferase.](BMRI2014-249856.003){#fig3}\n\n![Alignment of the amino acid sequences of the eight reported NDM variants. Conserved residues of the active site of metallo-*\u03b2*-lactamases are highlighted in gray. The bolded leucine in position 154 has been described to be responsible to an increased carbapenem hydrolysis.](BMRI2014-249856.004){#fig4}\n\n![Principle (a) and interpretation (b) of the Carba NP test recently developed for the rapid identification of carbapenemase producers among Enterobacteriaceae and*Pseudomonas* spp.](BMRI2014-249856.005){#fig5}\n\n[^1]: Academic Editor: Karmen Torkar\n"}
+{"text": "Introduction {#s1}\n============\n\nRemembering the place of an object (object-location memory, OLM) is crucial for adapting to changing environments in every-day life. However, this ability is known to decline during aging (Hedden and Gabrieli, [@B34]; Kessels et al., [@B50]) and may represent an incipient marker of neurodegenerative disease (Iachini et al., [@B42]). Advanced age increases risk of cognitive impairment and other age-related diseases (Salthouse, [@B87]; Niccoli and Partridge, [@B68]; Harada et al., [@B30]). Hence, early preventative strategies aiming to activate cognitive and brain resources in order to retain cognitive health, autonomy, and a better quality of life (Depp and Jeste, [@B17]) are of paramount importance. A variety of approaches have been investigated for its therapeutic and neuro-enhancing potential, including cognitive training, dietary regimes, physical training, use of pharmacological agents, as well as non-invasive brain stimulation (Perceval et al., [@B78]).\n\nCognitive training (CT) usually involves strengthening of neural networks through repeated co-activation of specific neurocognitive circuits active during task performance (Santarnecchi et al., [@B89]). The literature indicates that CT is beneficial for older adults\\' memory, but gains are much smaller than in young subjects (c.f. Passow et al., [@B77]). Moreover, evidence for generalizing effects that go beyond trained domain (transfer effects) is scarce and inconsistent (Jaeggi et al., [@B43]). In addition, CT is generally time-consuming (applied over weeks) and might therefore suffer from low motivation and treatment adherence over time (Martin et al., [@B61]; Elmasry et al., [@B23]). Importantly, CT may be combined with and boosted by any of the other interventions, specifically non-invasive brain stimulation, offering a convenient application to further promote training effects (Prehn and Floel, [@B80]; Au et al., [@B4]).\n\nAnodal transcranial direct current stimulation (atDCS) is a non-invasive and painless technique increasingly used to modulate memory in aging (Bennabi et al., [@B7]). The rationale behind its use is the potential of atDCS to increase cortical excitability in targeted brain areas by subthreshold alteration of resting membrane potential (Nitsche et al., [@B73]; Giordano et al., [@B28]) and a modulation of glutamatergic neurotransmission which promote mechanisms of long-term-potentiation (LTP; Stagg and Nitsche, [@B95]). Of note, in a previous study, we demonstrated no impact of atDCS on immediate recall in OLM, but observed an improvement with 1-week delay after cessation of a single stimulation period (Floel et al., [@B26]). Consequently, and in accordance with others a consolidation mechanism susceptible to atDCS was suggested (Reis et al., [@B83], [@B82]; Prehn and Floel, [@B80]; Perceval et al., [@B78]; Sandrini et al., [@B88]). Hence, larger or more sustained improvements may be induced by repeated applications of a combined atDCS-training approach (e.g., Hsu et al., [@B41]; Nilsson et al., [@B69]). This principle has been nicely laid out in Holland et al. ([@B37]) suggesting that small gains would accumulate by each bout of stimulation on consecutive days.\n\nPrevious studies in the motor domain support this idea (Reis et al., [@B83], [@B82]). With regard to memory training, most recent evidence stems from studies combing working memory training and atDCS (Passow et al., [@B77]). However, the findings have not been unequivocal. While some studies demonstrated a small but significant positive effect of atDCS over the course of training relative to sham (Park et al., [@B76]; Jones et al., [@B47]; Au et al., [@B5]), others did not (Lally et al., [@B55]; Stephens and Berryhill, [@B96]; Nilsson et al., [@B70]). Moreover, little is known about a putative synergistic effect of combined intervention in episodic memory in older adults (Prehn and Floel, [@B80]; Bartr\u00e9s-Faz and Vidal-Pi\u00f1eiro, [@B6]; Perceval et al., [@B78]; Birba et al., [@B10]; Passow et al., [@B77]). Such synergistic effect would predict greater, prolonged or more persistent improvements when interventions are applied together compared to each intervention alone (see also Ditye et al., [@B18]). Most notably there is only one published training-plus-atDCS study of Cotelli et al. ([@B15]) on age-sensitive hippocampus-dependent associative memory like OLM. In this study, patients with Alzheimer\\'s disease underwent a 10-session face-naming association memory training paired with left dorsolateral prefrontal cortex (LDPFC) stimulation. Memory did not benefit from atDCS relative to sham, possibly due to substantial inter-individual variability in degree of cognitive decline and brain organization, which may thus have altered brain responsiveness to atDCS.\n\nWhat may account for the inconsistent findings across studies so far? A variety of factors have been discussed, such as age, gender, education, health status, genetic background, brain state, baseline performance, but also mood, motivation, activation, or quality of sleep (e.g., Krause and Cohen Kadosh, [@B52]; Santarnecchi et al., [@B89]; Hsu et al., [@B40]). Given the high functional relevance of OLM, which is vulnerable to decline during aging (Postma et al., [@B79]; Shih et al., [@B93]), and our promising finding after a single session application of atDCS delivered during a visuospatial task (Floel et al., [@B26]) we set out to assess the impact of a combined atDCS-OLM-training protocol. Therefore, healthy older subjects underwent an OLM-training on 3 consecutive days in a sham-controlled cross-over design. Performance immediately after the protocol (training success, primary outcome) and memory performance after 1 month (delayed memory, secondary outcome) were tested and controlled for main potential modulating factors (covariates: gender, age, education, sequence of stimulation and baseline performance). Moreover, interventions which are able to induce more generalized cognitive effects are of utmost behavioral relevance. Thus, we explored the impact of atDCS vs. sham stimulation on performance on trained (similar OLM task) and untrained (visuo-constructive and verbal) memory functions (transfer).\n\nMaterials and methods {#s2}\n=====================\n\nSubjects\n--------\n\nHealthy older adults between 50 and 90 years were recruited via advertisements in Berlin, Germany. Subjects were pre-screened by a structured phone interview for major exclusion criteria such as history of epilepsy or metal implants. Individuals that passed pre-screening underwent an on-site medical and neuropsychological screening, and a structural magnetic resonance imaging (MRI) to ascertain the following inclusion criteria: (1) native German language speaker; (2) no current intake of medication that affect the central nervous system (e.g., antipsychotics or antidepressants); (3) normal routine medical and neurological examinations; (4) no recreational drug use; and (5) no cognitive impairment as assessed by CERAD screening test (Consortium to Establish a Registry for Alzheimer\\'s Disease test battery; Memory Clinic Basel, [www.memoryclinic.ch](http://www.memoryclinic.ch)). Specifically, results of CERAD memory scales had to be within 1 SD of age/education norms and Mini Mental State Examination \u226526 points (Folstein et al., [@B27]). Then, subjects completed a comprehensive baseline assessment. Baseline tests comprised cognitive status obtained from standard neuropsychological tests, and non-cognitive functions acquired from standardized questionnaires (for details see Table [1](#T1){ref-type=\"table\"}). From 56 subjects that were screened, 19 declined participation because of time constraints and one did not met inclusion criteria. From the remaining 36 subjects, four subjects had to be excluded due to abnormal MRI findings (*n* = 2) and technical problems during training sessions (*n* = 2), thus leaving 32 healthy older subjects (mean age (SD) in years, \\[range\\]: 68 (7), \\[53--79\\], 22 females) for analysis. All subjects gave written informed consent before study-specific procedures and received a reimbursement for participation. The study was approved by the Ethics Committee of the Charit\u00e9 - Universit\u00e4tsmedizin Berlin, Germany, was conducted in accordance with the declaration of Helsinki, and was registered at <https://clinicaltrials.gov/> (NCT02110056).\n\n###### \n\nDemographic, non-cognitive, and cognitive baseline characteristics for healthy older adults grouped according to stimulation (atDCS, sham) applied in the 1st study block.\n\n                                                                            **atDCS Mean (SD)**                      **Sham Mean (SD)**\n  ------------------------------------------------------------------------- ---------------------------------------- --------------------\n  N (women)                                                                 16 (11)                                  16 (11)\n  Age (years)                                                               69 (6)                                   67 (8)\n  Education (no. of years)[^a^](#TN1){ref-type=\"table-fn\"}                  15 (3)                                   16 (3)\n  ApoE genotype \u03b54 allele carriers (N; %)                                   6[^b^](#TN2){ref-type=\"table-fn\"}; 20%   12; 37%\n  **NON-COGNITIVE CHARACTERISTICS**                                                                                  \n  Depression: BDI[^a^](#TN1){ref-type=\"table-fn\"}                           2.9 (2.4)                                3.7 (3.1)\n  Quality of life: WHOQoL (overall score)[^a^](#TN1){ref-type=\"table-fn\"}   76.6 (11.1)                              73.3 (17.6)\n  Sleep: PSQI[^b^](#TN2){ref-type=\"table-fn\"}                               5.3 (2.6)                                5.1 (3.6)\n  Coping with Stress: SVF: \u2022 positive strategies                            13 (3.5)                                 19.1 (21.5)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 negative strategies        8.1 (3.3)                                13.1 (23.2)\n  Motivation: NFC[^c^](#TN3){ref-type=\"table-fn\"}                           36 (22.0)                                49.9 (34.2)\n  **COGNITIVE DOMAINS**                                                                                              \n  MMSE                                                                      29.1 (1.3)                               29.1 (1.3)\n  CERAD \u2022 word recall                                                       7.9 (1.9)                                8.3 (1.8)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 figures recall                                           9.6 (2.6)                                9.3 (1.8)\n  Digit span \u2022 forward[^a^](#TN1){ref-type=\"table-fn\"}                      8.9 (1.7)                                8.6 (2.1)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 backwards[^a^](#TN1){ref-type=\"table-fn\"}            6.9 (2.4)                                6.3 (1.9)\n  TMT-A (sec)                                                               42 (14)                                  43 (14)\n  TMT-B (sec)                                                               81 (26)                                  75 (28)\n  Fluency: \u2022 s-words                                                        17.7 (4.7)                               17.4 (5)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 category animals                                        23.9 (6.7)                               23.2 (5.2)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 sport-fruits                                            15.3 (2.2)                               15 (3.1)\n  TAP: \u2022 Inhibition (Go/NoGo; median in ms)                                 414.5 (118.7)                            464.8 (78.9)\n  \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2022 Alertness                                                     0.01 (0.13)                              0.06 (0.08)\n  MWT                                                                       32.2 (2.0)                               33.7 (1.2)\n\nData are given as mean (SD). In some parameters N is reduced due to missing data:\n\n*N = 31*,\n\n*N = 30*,\n\n*N = 29*.\n\n*ApoE \u03b54 allele: Apolipoprotein E-DNA was extracted from whole blood using a blood mini-kit (Qiagen, Hilden, Germany) and genotyping on coded samples was performed by the lab of Prof. Dr. Dan Rujecscu, University Halle, Germany, procedure is described in more detail in Kerti et al. ([@B49]). BDI, Becks depression inventory (Hautzinger et al., [@B32]); WHOQoL, WHO Quality of life (Angermeyer et al., [@B1]); SVF120, stress coping strategies---habitual form (Erdmann and Janke, [@B24]); PSQI, habitual sleep score (Pittsburgh Sleep Quality, Buysse et al., [@B13]); NFC, Need for Cognition (Bless et al., [@B11]); MMSE, Mini Mental State Examination scores (Folstein et al., [@B27]); CERAD, Consortium to Establish a Registry for Alzheimer\\'s Disease test battery (Memory Clinic Basel, [www.memoryclinic.ch](http://www.memoryclinic.ch)); Digit span (H\u00e4rting et al., [@B31]); TMT, Trail Making Test (Tombaugh, [@B99]); Fluency, Regensburger Verbal Fluency Test (Aschenbrenner et al., [@B3]); TAP, computerized test battery to test attention (Zimmermann et al., [@B107]); MWT, Multiple-Choice Vocabulary Intelligence Test (Lehrl, [@B58])*.\n\nExperimental design\n-------------------\n\nIn this subject-blind, placebo-controlled, cross-over study all subjects were tested in two blocks. Each block comprised a 3-day visuospatial OLM-training, paired with either anodal or sham tDCS (combined intervention of atDCS+training and sham+training is abbreviated in the following to \"atDCS\" and \"sham\"). Order of stimulation (atDCS vs. sham) were pseudo-randomized and controlled for age and gender, with 16 subjects receiving atDCS first and 16 subjects receiving sham first. Additionally, sequence of intervention was balanced and separated by 3 months to prevent carry-over effects. Before (pre-training) and 1-day and 1-month after training (follow-up measurements FU1 and FU2), three other tasks to test trained (one task) and untrained (two tasks) memory functions were employed (transfer tasks). On average, subjects that started with atDCS first were comparable to subjects with sham first in terms of age, gender, distribution of ApoE \u03b54 allele carriers (a polymorphism that have been previously implicated in memory outcome, e.g., Wisdom et al., [@B105]; Matura et al., [@B62]), or baseline cognitive performance (see Table [1](#T1){ref-type=\"table\"} for details). An exception was found for estimator of premorbid intelligence (German version of Multiple-Choice Vocabulary Intelligence Test, MWT, Lehrl, [@B58]). Here, subjects with sham first scored on average 1.5 points \\[correct words, mean (SD): 33.7 (1.2) vs. 32.2 (2.0)\\] higher than subjects in the atDCS group prior to intervention.\n\nTraining task\n-------------\n\nFigure [1](#F1){ref-type=\"fig\"} provides an overview of procedure and tasks (for details see text below). To train visuospatial memory the OLM-task called LOCATO was used (Floel et al., [@B26]; Kulzow et al., [@B53]). In the present study, 30 pictures of real-life buildings (objects) were associated with different positions (locations) on a two-dimensional street map (\"LOCATO-30\"). Subjects had to learn the correct object-location pairings within five learning blocks on each of three training days. In detail, each learning block contained 120 trials (2 \u00d7 30 correct and 60 incorrect object-location associations) presented in randomized order resulting in a total of 1,800 trials across 3-day training. Thus, over the course of 3-day training correct object-location pairings were shown 30 times (10 per day) more frequently compared to \"incorrect\" positions (shown only once, respectively). Each trial comprised a picture of a schematized street map with one building presented for 3,000 ms and an inter-stimulus interval of 1,000 ms. Within this time frame subjects had to indicate by button press (\"YES,\" \"NO\") on a response pad as accurate as possible whether the building was presented at the \"correct\" or \"incorrect\" location (see Figure [1A](#F1){ref-type=\"fig\"}). Correct/incorrect responses were recorded during each learning trial in every learning block. No online feedback on performance was provided. Memory performance was tested shortly after the end the of the fifth learning block. Performance was assessed by cued recall using two different test formats, namely item recognition (IR) and 3-alternative forced choice (3-AFC) test (see also \"Recall Format\" of Figure [1A](#F1){ref-type=\"fig\"}). To avoid contaminations due to task order, and to reduce overall testing time, 50% of associations were tested via IR and the remaining 50% via 3-AFC. For IR, 15 correct object-location associations were intermixed with 15 new (not shown before) incorrect pairings. Stimulus presentation was identical to learning blocks and subjects had to indicate by button press if the position was \"correct\" (\"YES\") or \"incorrect\" (\"NO\"), timing was self-paced. In the subsequent 3-AFC test three possible locations for a particular building were shown on the street map marked with \"1,\" \"2,\" and \"3.\" The subject had to choose the building\\'s \"correct\" location by pressing the corresponding number on the keyboard, timing was self-paced. Two parallel versions (A,B) of LOCATO-30 were used, each with a different set of buildings, and with the street map rotated for 180\u00b0 for version B. Versions were assigned in counterbalanced manner to respective intervention. LOCATO was presented on a computer using Presentation software (Neurobehavioral Systems, Albany, CA, USA).\n\n![Study overview. **(A)** Schematic of the associative object-location learning paradigm (LOCATO). During acquisition, each trial comprised a picture of a schematized street map with one building. Buildings (objects) occurred on one correct and 10 incorrect positions (locations). Subjects had to learn correct object-location-pairings over the course of multiple learning blocks and indicate (button press) in each trial, whether a building was in a \"correct\" location (Yes or NO). Memory for object-location-associations were assessed by two different cued recall tests \\[item recognition (IR), 3-alternative-forced choice test (3-AFC)\\]. During IR correct object-location-associations were intermixed with new (incorrect) object-location associations and subjects indicated by button press if presented position is \"correct.\" In 3-AFC tests subjects specified by button press (\"1,\" \"2,\" or \"3\") the correct position of the building shown above the schematic street map. **(B)** Timeline Training. Each of the two study blocks (cross-over design) comprised 6 sessions (2--7; first session (not shown) included baseline testing) with 3 months in-between. Training (session 3--5) consisted of three consecutive days, each comprised five learning blocks and subsequent cued recall test (IR, 3-AFC). Memory was also tested follow-up on session 6 (1-day) and 7 (1-month) post-training. Overnight retention (offline effects) was assessed by applying memory tests (IR, 3-AFC) before next training (session 4 and 5). In addition, at the beginning of each training day subjects self-rated their affective state (\"Befindlichkeitsskalierung anhand von Kategorien und Eigenschaftsw\u00f6rtern\"; BSKE, Janke et al., [@B45]), and provided information about their sleep (number of slept hours, sleep quality) of previous night. In pre- and post-training sessions (2,6,7) we asked for positive and negative affective state (PANAS, Watson et al., [@B103]) and applied additional memory tests to assess transfer effects in trained (LOCATO-15; shorter and less complex version of training paradigm, Kulzow et al., [@B53]), and untrained (Rey--Osterrieth Complex Figure (ROCF) Test, Knight and Kaplan, [@B51]; Rey Auditory Verbal Learning Test (AVLT), Helmstaedter et al., [@B36]) memory tasks. **(C)** Stimulation protocol. Anode (7 \u00d7 5 cm^2^) was attached to T6 (according to EEG 10--20 System) and return electrode (cathode: 10 \u00d7 10 cm^2^) contralateral above the eyebrow (supraorbital). Connector of the anode was positioned at the posterior edge distant from the return electrode. Larger size of the cathode renders the stimulation density functionally ineffective. Anodal transcranial direct current stimulation (atDCS) of 1 mA was administered during beginning of OLM-training (Session 3--5) for 20 min (\"atDCS\") or 30 s (\"sham\") and current was ramped up and down within 10 s. Abbreviations: IR, item recognition; 3-AFC, 3 alternative forced choice task; PANAS, Positive and Negative Affect Schedule; LOCATO-15, short version of object-location-memory task; ROCF, Rey--Osterrieth Complex Figure Test; AVLT, German version of the Rey Auditory Verbal Learning Test, d-day; mo, month; FU, Follow-up.](fnins-11-00746-g0001){#F1}\n\nTransfer tasks\n--------------\n\nTransfer on trained function was measured by a LOCATO-15 (short version of training task). LOCATO-15 comprised object-location-learning (OLL) of 15 associations within three learning blocks on a less complex street-map compared to LOCATO-30 training task. Memory was tested immediately after learning by a 3-AFC cued recall test. Three parallel versions were used for pre- and FU1 and FU2 tests. The parallel versions consisted of different sets of buildings presented on different street maps, and were applied in balanced order across subjects and time points. Transfer effects on untrained memory functions were assessed by using the two following learning and memory tasks: (i) Rey--Osterrieth Complex Figure Test (ROCF; originally designed by Rey, [@B84]; Knight and Kaplan, [@B51]), which require to copy a complex figure from original and delayed (approx. 20 min) from memory, and (ii) German version of the Rey Auditory Verbal Learning Test (AVLT; Helmstaedter et al., [@B36]), which consisted of remembering 15 learned words after five immediate recall trials and again after a 30 min delay. For ROCF-Test and AVLT available parallel versions were used across subjects and time points to minimize test-retest effects, respectively.\n\nBrain stimulation\n-----------------\n\nDuring the beginning of OLM-training, either atDCS (20 min of anodal tDCS, 1 mA) or sham (30 s of anodal tDCS, 1 mA) was applied in a ramp-like fashion (fade in and fade out 10 s, respectively). Stimulation was delivered by a direct current stimulator (NeuroConn GmbH, Ilmenau, Germany) using two saline-soaked surface sponge electrodes. As different configurations can significantly affect the resulting electrical field (e.g., Saturnino et al., [@B91]) set up of electrodes (position and orientation; see also Figure [1C](#F1){ref-type=\"fig\"}) was precisely pre-defined by use of standard operation procedures. The anode (7 \u00d7 5 cm^2^, current density = 0.028 mA/cm^2^) was placed over right temporoparietal cortex, centered on T6 (according to the international 10--20 electroencephalography system). Another return electrode (cathode: 10 \u00d7 10 cm^2^, current density = 0.01 mA/cm^2^) was positioned contralateral above the left eyebrow (supraorbital) and was centered to the left eye pupil. Electrodes were attached to the scalp using rubber bands. Given that right temporoparietal region is implicated in the acquisition of OLM (Postma et al., [@B79]) and anodal tDCS over this area has been shown to improve performance on a similar version of the task employed in our study (Floel et al., [@B26]; Prehn et al., [@B81]) this site was selected for anodal simulation. Note, that the larger size of the cathode renders the stimulation density functionally ineffective. Moreover, the current density of cathode was below the required minimum (0.017 mA/cm^2^) to modify cortical excitability by tDCS in humans (Nitsche and Paulus, [@B72]; Nitsche et al., [@B71]). Perception of stimulation was prompted after application of atDCS or sham. Subjects had to indicate first, if they experienced the stimulation (YES/NO) and second, to rate their level of discomfort due to stimulation on a scale from 0 (not at all) to 6 (very strongly).\n\nProcedure\n---------\n\nProcedure was identical for study block 1 and 2. OLM performance was tested immediately after the end of each training day (see Figure [1B](#F1){ref-type=\"fig\"}). In addition, OLM performance was obtained at the next day before the respective intervention had started (to capture offline effects) and at FU1 and FU2 sessions. Moreover, before atDCS or sham was applied, potential confounders such as emotional state and sleep characteristics were assessed on every training day. Specific (e.g., anger, anxiety) and unspecific (e.g., activation, excitation) affective states (10 in total) were rated by means of the German questionnaire BSKE (\"Befindlichkeitsskalierung anhand von Kategorien und Eigenschaftsw\u00f6rtern\"; BSKE, Janke et al., [@B45]) on scales ranging from 0 (not at all) to 6 (very strongly). Subjective perception of sleep quality and sleep duration of prior night were determined by two questions, that is \"How did you sleep last night?\" \\[rated from 0 (lousy) to 6 (excellent)\\] and \"How many hours did you sleep last night?\" Besides performance measurement of trained and untrained memory functions (LOCATO-15, ROCF, AVLT), positive and negative affective state was self-rated by means of Positive and Negative Affect Schedule (PANAS, Watson et al., [@B103]) at pre- and FU1 and FU2 sessions. In contrast to aforementioned measurements quality of Life (WHO Quality of life, Angermeyer et al., [@B1]) and habitual subjective sleep quality (Pittsburgh Sleep Quality, Buysse et al., [@B13]) were controlled and assessed only once within the 1st study block by standardized questionnaires.\n\nData aggregation\n----------------\n\nPercent correct scores (PC) were calculated for every learning block (L1-L5, acquisition) and IR-test (memory) on the basis of hits and correct rejections in the respective trial. PC was defined as follows: PC = \\[number of hits + number of correct rejections\\] ^\\*^ 100/total number of presented buildings. Performance in the other used memory test (3-AFC) was measured by number of correct selected responses in %. Primary outcome \"training success\" was pre-specified before start of study (see clinicaltrials.gov: NCT02110056). Training success was operationalized by PC at fifth learning block (L5) on last training day (day3) and adjusted for baseline performance in the very first learning block \\[PC~L5\\ day3~ \u2212 PC~L1\\ day1~\\] to account for inter-individual differences. Secondary outcome comprised memory after 1-month post-training (FU 2). Therefore, cued recall performance (3-AFC, IR) at FU 2 was used and adjusted for learning performance after training day 1 (PC~L5~). For exploratory analyses, we computed indices for *online* (within-session), and *offline* (between-session) performance. Online scores were related to improvements (difference) within each training day \\[ON~day\\ n~ = PC~L5_day\\ n~ \u2212 PC~L1_day\\ n~\\], and offline scores to overnight changes in performance. Offline scores included cued recall performance (3-AFC and IR, respectively) before start of next OLM-training in relation to learning performance of previous day and were determined in the following way: OFF~day\\ n~ = Cued recall~day\\ n+1\\ before\\ OLM\u2212training~ \u2212 PC~L5_day\\ n~.\n\nStatistics\n----------\n\nAll Statistical analyses were conducted with IBM SPSS 24 ([www.ibm.com/software/de/analytics/spss/](http://www.ibm.com/software/de/analytics/spss/)). Perception of stimulation between training with and without atDCS was compared by non-parametric Wilcoxon test. Categorical variables (stimulus assignment) were analyzed by chi-square-tests (\u03c7^2^). The effect of atDCS concomitant to training was analyzed using separate linear mixed models (random intercept models; Verbeke and Molenberghs, [@B101]). To test effects on training success and delayed memory, repeated measurements (\"atDCS\" or \"sham\") were entered as level one unit nested in different individuals as level two units (64 data points in total). The stimulation effect was tested using the dichotomous variable INTERVENTION (atDCS, sham). Additional conducted analyses (exploratory) comprised on- and offline effects, and changes in emotional state and sleep characteristics across 3-day-training. For those analyses a further factor (factor \"DAY\") was added to the models (192 data points in total for each model). To statistically control for potential confounders such as age, gender, education, baseline performance (pre-training performance in LOCATO-15), sequence of intervention, and MWT-score, analyses were repeated with these variables as covariates. Pre- and post-training mood (PANAS) and changes in trained and untrained transfer tasks (LOCATO-15, ROCF, AVLT) as a function of intervention were analyzed by separate linear mixed models with three factors \"INTERVENTION\" (atDCS, sham), \"TIME\" (Baseline, FU1, FU2), and \"SEQUENCE\" (study block 1, study block 2). Impact of intervention was reported by regression coefficients \u03b2, 95% confidence interval \\[95% CI\\], and d as a measure of effect size if not otherwise mentioned. An effect size is typically defined as the ratio of a difference between treatment and control group means to a standard deviation (SD). According to Hedges ([@B35]) an appropriate SD in mixed model analysis can be obtained from the square root of summed covariance parameters (residual and intercept) in order to combine both, within- and between-subject variance. Estimate of d were then calculated by the ratio of estimated margin mean difference to SD. If appropriate, model-based *post-hoc* tests were computed to specify effects using *post-hoc* margin mean differences and 95% CI of these differences. Mean differences in % were reported and refer to atDCS---sham, if not otherwise mentioned. No adjustments were made to correct for multiple comparisons. The two-sided level of significance for all analyses was set at \u03b1 = 0.05.\n\nResults {#s3}\n=======\n\nMemory outcomes from training\n-----------------------------\n\n### Training success and delayed memory\n\nLearning and cued recall performance (3-AFC and IR) in LOCATO training task across days are presented in Figure [2](#F2){ref-type=\"fig\"}. For \"training success\" linear mixed model analysis revealed no significant difference between atDCS relative to sham (\u22121.3 \\[\u22124.4, 1.9\\], *d* = 0.1). Also, no significant difference in delayed memory was found after previously administered atDCS compared to sham in neither of the used cued recall tests (3-AFC: \u22125.8 \\[\u221215.1, 3.4\\], *d* = 0.3; IR: 2.0 \\[\u22125.6, 9.5\\], *d* = 0.1). Linear mixed model analysis adjusted for the different covariates likewise did not show a significant beneficial atDCS effect on performance compared to sham (for details see Table [2](#T2){ref-type=\"table\"}).\n\n![Performance during and after object-location-memory training. Response accuracy (% correct) during each learning (L) block and overnight cued recall performance (% correct) in 3-alternative forced choice (R~3AFC~) and item recognition (R~IR~) task assessed before next training on day 2 and 3 as well as cued recall at 1-day and 1-month follow-up is depicted. Dark filled circles (black, dark gray) represent performance of atDCS applied during training (\"atDCS\"), light filled circles represent performance of sham applied during training (\"sham\"). Behavioral online effects related to within-session performance and offline effects to between-session performance. Data are given as means and standard deviations. D, day; mo, month.](fnins-11-00746-g0002){#F2}\n\n###### \n\nResults of separate linear mixed models analysis with factor \"INTERVENTION\" (atDCS vs. sham) for training success and delayed recall after 1-month (measured by 3-AFC and IR) without (MODEL 0) and with adjustment (MODEL 1).\n\n  **Outcome**                                                                          **Training success**                  **Delayed recall (long-term memory)**                                                                                                                                \n  ------------------------------------------------------------------------------------ ------------------------------------- --------------------------------------- ---------------- ------------------------------------- ------- ---------------- ------------------------------------- ------ ----------------\n  **MODEL 0**                                                                                                                                                                                                                                                                                     \n  n (atDCS, sham)                                                                      32, 32                                                                                         32, 30                                                         31, 31                                       \n  Data points in total                                                                 64                                                                                             62[^\\*^](#TN4){ref-type=\"table-fn\"}                            62[^\\*^](#TN4){ref-type=\"table-fn\"}          \n  INTERVENTION                                                                                                               \u22121.3                                    \\[\u2212 4.4,1.9\\]                                          \u22125.8    \\[\u221215.1, 3.4\\]                                         2.0    \\[\u22125.6, 9.5\\]\n  **MODEL 1 (ADJUSTED FOR AGE, EDUCATION, GENDER, BASELINE (LOC15), SEQUENCE, MWT)**                                                                                                                                                                                                              \n  n (atDCS, sham)                                                                      31,31                                                                                          29,31                                                          30,30                                        \n  Data points in total                                                                 62[^\\*^](#TN4){ref-type=\"table-fn\"}                                                            60[^\\*^](#TN4){ref-type=\"table-fn\"}                            60[^\\*^](#TN4){ref-type=\"table-fn\"}          \n  INTERVENTION                                                                                                               \u22121.6                                    \\[\u22124.8, 1.7\\]                                          \u22125.3    \\[\u221214.5, 3.9\\]                                         1.9    \\[\u22125.3, 9.0\\]\n  Age                                                                                                                        0.4                                     \\[\u22120.1, 0.9\\]                                          0.9     \\[\u22120.1, 1.8\\]                                          0.02   \\[\u22120.8, 0.8\\]\n  Education                                                                                                                  \u22120.6                                    \\[\u22121.9, 0.8\\]                                          \u22120.02   \\[\u22122.2, 2.1\\]                                          \u22120.3   \\[\u22122.3, 1.8\\]\n  Gender                                                                                                                     \u22126.8                                    \\[\u221214.9, 1.2\\]                                         \u22126.1    \\[\u221218.5, 6.2\\]                                         \u22121.8   \\[\u221213.2, 9.7\\]\n  Baseline LOC15                                                                                                             \u22120.3                                    \\[\u22120.7, 0.1\\]                                          0.02    \\[\u22120.6, 0.6\\]                                          0.3    \\[\u22120.2, 0.9\\]\n  Sequence                                                                                                                   \u22121.9                                    \\[\u22125.2, 1.3\\]                                          7.9     \\[\u22121.4, 17.1\\]                                         9.1    \\[\u22122.0, 16.3\\]\n  MWT-score                                                                                                                  \u22120.5                                    \\[\u22122.4, 1.4\\]                                          \u22120.1    \\[\u22123.1, 2.9\\]                                          0.7    \\[\u22122.1, 3.5\\]\n\n*Three separate linear mixed models (MODEL 0: dependent variables: training success, 3-AFC and IR, respectively; independent variables: Intervention (atDCS, sham). MODEL 1 three separate linear mixed models with additional adjustment for age, education, gender, visuospatial baseline performance (LOC15), sequence of \"atDCS\" and \"sham\" and MWT (Multiple-Choice Vocabulary Intelligence Test)-score; n-number of individuals, \u03b2 = regression coefficient (sham = 0); CI = confidence interval; 3-AFC, 3 alternative forced choice task; IR, item recognition; LOC15, LOCATO-15*.\n\n*Reduced data points due to missing data in training with atDCS or training with sham session*.\n\n### Behavioral measures of on- and offline effects\n\nFor between-session (offline) measurements significant less forgetting overnight after atDCS compared to sham were found after first night in IR test (4.2 \\[0.3, 8.0\\], *d* = 0.5), but not after night two (0.5 \\[\u22123.4, 4.4\\], *d* = 0.1) and three (\u22121.84 \\[\u22125.7,2.1\\], *d* = 0.2). This 1st night effect remained also after adjustment for covariates (4.1 \\[0.2, 8.0\\], *d* = 0.5). For 3 AFC-test this 1st night offline effect was less clear (4.1 \\[\u22122.0, 10.2\\], *d* = 0.3), but was also observable after adjustment for above mentioned confounders (5.4 \\[\u22120.6, 11.4\\], *d* = 0.4). Analysis of within-session (online) improvements revealed no significant effects of atDCS (see Table [3](#T3){ref-type=\"table\"}), but a substantial training effect was evident. Because subjects started on higher performance level each day, magnitude of online effects significantly decreased across training days (ON~day1~ = 18.8 \\[16.3, 21.0\\], *d* = 2.2; ON~day2~ = 9.9 \\[7.7, 12.1\\], *d* = 1.1; ON~day3~ = 6.3 \\[4.1, 8.5\\], d = 0.7).\n\n###### \n\nMean Difference (Mean Diff) atDCS-sham and 95% CI of model based *post-hoc* comparisons between atDCS and sham for online (assessed as within-session difference performance score on each training day) and offline scores (assessed as overnight difference performance score between training sessions for 3-AFC and IR-scores) without (MODEL 0) and with adjustment (MODEL 1).\n\n  **Outcome**                                                                          **Online scores**                      **Offline scores**                                                                                                                                                                   \n  ------------------------------------------------------------------------------------ -------------------------------------- -------------------- --------------- -------------------------------------- -------------------------------------- ---------------- -------------------------------------- --------- ------------------\n  **MODEL 0**                                                                                                                                                                                                                                                                                                      \n  n (atDCS, sham)                                                                      30, 32                                                                      32, 32                                                                                         31, 32                                           \n  Data points in total                                                                 190[^\\*^](#TN5){ref-type=\"table-fn\"}                                        192                                                                                            191[^\\*^](#TN5){ref-type=\"table-fn\"}             \n  Day 1                                                                                                                       0.8                  \\[\u22123.4, 5.1\\]                                          4.1                                    \\[\u22122.0, 10.2\\]                                          **4.2**   **\\[0.3, 8.0\\]**\n  Day 2                                                                                                                       \u22122.2                 \\[\u22122.2, 6.5\\]                                          \u22123.1                                   \\[\u22129.1, 3.0\\]                                           0.5       \\[\u22123.4, 4.4\\]\n  Day 3                                                                                                                       \u22120.6                 \\[\u22124.8, 3.7\\]                                          1.1                                    \\[\u22125.0, 7.2\\]                                           \u22121.84     \\[\u22125.7, 2.1\\]\n  **MODEL 1 (ADJUSTED FOR AGE, EDUCATION, GENDER, BASELINE (LOC15), SEQUENCE, MWT)**                                                                                                                                                                                                                               \n  n (atDCS, sham)                                                                      30, 30                                                                      31, 31                                                                                         31, 30                                           \n  Data points in total                                                                 184[^\\*^](#TN5){ref-type=\"table-fn\"}                                        186[^\\*^](#TN5){ref-type=\"table-fn\"}                                                           185[^\\*^](#TN5){ref-type=\"table-fn\"}             \n  Day 1                                                                                                                       0.1                  \\[\u22124.1, 4.4\\]                                          5.4[^\\#^](#TN6){ref-type=\"table-fn\"}   \\[\u22120.6, 11.4\\]                                          **4.1**   **\\[0.2, 8.0\\]**\n  Day 2                                                                                                                       \u22122.3                 \\[\u22126.6, 2.0\\]                                          \u22121.6                                   \\[\u22127.6, 4.3\\]                                           0.7       \\[\u22123.2, 4.6\\]\n  Day 3                                                                                                                       \u22120.4                 \\[\u22124.7, 3.8\\]                                          1.9                                    \\[\u22124.1, 7.9\\]                                           \u22122.3      \\[\u22126.3, 1.6\\]\n\n*Model-based post-hoc tests resulted from three separate linear mixed models (MODEL 0: dependent variables: online effects, offline effects for 3-AFC and IR, respectively; independent variables: Intervention (atDCS, sham), Day (d1,d2,d3) and Intervention x Day; MODEL 1 with additional adjustment for age, education, gender, visuospatial baseline performance (LOC15), sequence of \"atDCS\" and \"sham\\\" and MWT (Multiple-Choice Vocabulary Intelligence Test-score as covariates); n, number of individuals; CI, confidence interval; 3-AFC, 3 alternative forced choice task; IR, item recognition; LOC15, LOCATO-15*.\n\n*Reduced data points due to missing data in training with atDCS or training with sham session. Positive difference scores indicate better performance (online: better learning, offline: less overnight forgetting) of atDCS relative to sham. Significant differences (p \\< 0.05) are bold*.\n\n*p \\< 0.10*.\n\nIn sum, training success and delayed memory was not affected by atDCS, but 3-day visuospatial training significantly improved OLM independent of atDCS. A small benefit of atDCS relative to sham was restricted to the first offline score (after 1st night) as indicated by less overnight forgetting after receiving atDCS compared to sham.\n\nControl of sleep characteristics and affective state during training\n--------------------------------------------------------------------\n\nSleep characteristic did not significantly differ between atDCs and sham (for details see Table [4](#T4){ref-type=\"table\"}). Subjects slept on average 7 h and reported good quality of sleep of prior night (scored \"4\" on average on a scale from 0 (lousy) to 6 (excellent) across training days. Also, no significant differences were found with regard to positive (relaxation, good mood, activation, confidence) and negative (excitation, bad mood, anger, anxiety, depressed, deactivation) affective states between atDCS and sham rated immediately before beginning of each training (BSKE: F-statistics and associated *p*-values of fixed effects: all *p*\\'s \\> 0.08). Overall subjects felt rather positive and rated themselves very low on negative affective states (scores on average \\< 1; scale range: 0 \"not at all\" to 6 \"very strongly\") across days.\n\n###### \n\nMean difference (Mean Diff) and 95% CI of model based *post-hoc* comparisons (mixed model analysis between atDCS and sham (atDCS-sham); *n* = 31 individuals in training with atDCS sessions, *n* = 30 individuals in training with sham session; data points in total 187[^\\*^](#TN7){ref-type=\"table-fn\"}) for reported sleep duration and sleep quality.\n\n  **Outcome**   **Sleep duration**   **Sleep quality**          \n  ------------- -------------------- ------------------- ------ ---------------\n  Day 1         \u22120.1                 \\[\u22120.5, 0.3\\]       0.1    \\[\u22120.2, 0.5\\]\n  Day 2         0.4                  \\[\u22120.1, 0.8\\]       0.04   \\[\u22120.3, 0.4\\]\n  Day 3         0.2                  \\[\u22120.2, 0.6\\]       0.3    \\[\u22120.7, 0.1\\]\n\n*Model-based post-hoc tests resulted from two separate linear mixed models (MODEL: dependent variables: sleep duration, and sleep quality, respectively; independent variables: Intervention (atDCS, sham), Day (d1,d2,d3) and Intervention \u00d7 Day. In both, atDCS and sham sessions, subjects slept 7 h on average across training days and reported good quality of sleep \\[scored \"4\" on average on a scale from 0 (lousy) to 6 (excellent)\\]*.\n\n*Reduced data points due to missing data*.\n\nAnalysis of pre- and post-training tasks and mood\n-------------------------------------------------\n\nWith regard to training gains on other tasks linear mixed model analysis with factors \"INTERVENTION,\" \"TIME,\" and \"SEQUENCE\" revealed no significant benefit of atDCS compared to sham neither for trained (LOCATO-15 task), nor for untrained (ROCF, AVLT) memory transfer tasks. However, a small inverse effect was observed for ROCF memory score (copy figure delayed from memory). Model based estimates indicated better performance 1-day after training for subjects receiving previously sham relative to atDCS (\u22122.9 \\[\u22125.1, \u22120.8\\], *d* = 0.4), but this difference did not persist after 1-month (\u22120.3 \\[\u22122.5, 1.8\\], *d* = 0.04). Although we used available parallel versions (three for LOCATO-15 and AVLT and two for ROCF) significant improvements in all outcomes of transfer tasks (except ROCF learning) were seen for performance in 2nd relative to 1st study block (significant \"SEQUENCE\" effects) probably due to practice and strategy learning during repeated testing.\n\nWith regard to mood no significant differences were found at pre- and FU1 and FU2 sessions between atDCS and sham. On average, subjects rated themselves higher on positive mood scale \\[mean (SD): 34.2(7.6)\\] than on negative mood scale \\[mean (SD): 11.3 (3.3)\\].\n\nStimulation perception\n----------------------\n\nDue to missing data, only data of 30 subjects were available for stimulation perception analysis. Analysis revealed that the majority of subjects could not discriminate atDCS from sham as 13 subjects believed to \"never have received atDCS,\" and 8 subjects reported that they received atDCS in both sessions. Only 9/30 subjects thought that they had received atDCS only once during the two study blocks, 7 of them assigned correctly and 2 incorrectly to the block comprising atDCS. However, cross-sectional analysis revealed that subjects neither in the first \\[$\\chi_{(1)}^{2}$ = 0.4\\] nor in the second \\[$\\chi_{(1)}^{2}$ = 2\\] study block could reliably differentiate stimulation conditions indicating that our sham procedure was successful in blinding subjects. Overall, subjects tolerated the procedure well. On average, subjects rated their \"feeling of discomfort caused by stimulation\" as very low on a scale ranging from 0 (not all) to 6 (extremely) under both, atDCS \\[mean (SD): 0.9 (1.2)\\] and sham \\[mean (SD): 0.7 (0.9)\\]. This difference was not significant (Z = \u22121.19).\n\nDiscussion {#s4}\n==========\n\nThe aim of this study was to investigate the effect of a combined intervention comprising 3 consecutive days of atDCS applied over right temporoparietal cortex and OLM-training in healthy older adults on training success, delayed memory (long-term effects after 1-month), and transfer tasks (generalization). First, we observed significant improvement in both training conditions, without additional gain induced by atDCS in training success or delayed memory performance. Second, exploratory analyses demonstrated a small benefit of atDCS for overnight forgetting rate, but this \"offline\" effect was confined to the first night. Third, results did not change after statistically controlling for a variety of proposed \"modulators.\" Forth, the combined intervention of atDCS and OLM-training did not promote performance on other trained and untrained memory tasks. In sum, in this cross-over study combining atDCS with an episodic memory training task (OLM), no improvement in the atDCS condition could be ascertained. However, a number of critical questions remain with regard to specifics of the experimental design, and also inter-individual differences in response to stimulation, which will be discussed. A more detailed understanding of potential modulators of the response to atDCS may help to boost episodic memory training with atDCS in older adults more successfully.\n\nSpecifics of the experimental design\n------------------------------------\n\nAs suggested by Holland et al. ([@B37]) a multi-day atDCS-training session protocol was applied, but in contrast to our recently reported results of beneficial effects of combined intervention on memory performance (Antonenko et al., [@B2]), we could not find a positive effect on training success. There are some decisive differences between both studies, which have to be considered when interpreting our findings. While Antonenko et al. ([@B2]) aimed to study brain-behavior associations in a sample of young and older adults, we solely tested older adults, and age has been recognized as one of the factors affecting responsiveness to atDCS (Meinzer et al., [@B64]; Krause and Cohen Kadosh, [@B52]; Summers et al., [@B97]). The studies differed in additional parameters assumed to be relevant in atDCS/cognition modulating protocols (Shin et al., [@B94]) such as primary cognitive outcome, time point of assessment and selected statistical approach. Specifically, Antonenko et al. ([@B2]) conducted resting state at baseline (pre-training) and with 1-day delay after OLM-training parallel to 3-AFC as recall test. In contrast, the present study determined \"training success\" by measuring performance (yes/no decision) during the last training block at third (last) training day. Both, the different test formats and time of testing (immediate vs. delayed) may involve different cognitive processes and physiological mechanisms of actions (Horvath et al., [@B38]) and could thus partly account for inconsistent results. Moreover, results from parallel group (Antonenko et al., [@B2]) may contradict results from within-group cross-over (present study) design, most notably because of temporal dynamics (carry-over effects). Most likely due to experience-based task-learning strategies, multi-day cognitive training resulted in strong practice effects during re-testing (cross-over design). Given that atDCS is a relatively weak form of modulation (Horvath et al., [@B38]), strong practice gains may have obscured subtle beneficial atDCS effects (see also Wang and Voss, [@B102]). Similarly, multi-day training would likely induce stronger practice effects during repeated testing than during single-session applications. This might explain different results between the present finding, and previous finding from our single-session approach (Floel et al., [@B26]). In addition, time course of atDCS after-effects are not fully understood and have not been systematically tested so far (Kuo et al., [@B54]). Together with possible carry-over effects, these after-effects might have masked beneficial effects of the stimulation condition on 1-month follow up testing. In sum, conceptual variability between different domains of cognition, known to be highly complex, have to be carefully taken into account when comparing studies. Further, more systematic work is required to understand temporal dynamics in multi-day cognitive training studies. For example, different experimental approaches should be employed in order to optimize methodological designs for atDCS-training approaches.\n\nAlternatively, number of sessions may still have been too small to observe significant differences between conditions immediately after training or after 1-month. However, there is no consensus about minimum/maximum number of sessions and studies systematically comparing number of sessions are lacking. Accordingly, the number of applied training+atDCS sessions varies widely across studies (e.g., Berryhill, [@B8]), and no consistent picture emerged so far. For example, a 3-day training+atDCS approach has yielded positive (Talsma et al., [@B98]) but a 20-session protocol negative effects (Nilsson et al., [@B70]). Thus, it remains unclear if increase of number of sessions would have led to more pronounced differential findings for memory performance in the present study.\n\nThe lack of beneficial atDCS effects on memory performance could also be related to the used low intensity of 1 mA during stimulation. Note that non-linear intensity-dependent effects of stimulation has been demonstrated in the motor (e.g., Jamil et al., [@B44]), but also in the cognitive domain (Hoy et al., [@B39]). For both, the lower intensities showed equal, if not greater effects in motor-cortical excitability (greatest at 0.5 and 1.0 mA compared to 1.5 mA and 2 mA) and working memory (1 mA better than 2 mA), respectively. Thus, the use of higher current might not necessarily mean a greater impact on performance, but can even invert stimulation effects (Woods et al., [@B106]). Moreover, the use of a higher dose might be problematic, because of safety and tolerability reasons in the meaning of increasing the risk of side effects which could in turn negatively affect subject blinding. Given that we have previously found positive effects with the application of 1 mA on OLM (Floel et al., [@B26]; Antonenko et al., [@B2]; Prehn et al., [@B81]), alternative factors, e.g., timing in relation to task, interindividual differences, or a combination, might contribute to null finding.\n\nAn impact of repeated atDCS on overnight performance that emerge between training sessions---possibly via affecting consolidation processes---has been previously suggested (Reis et al., [@B83], [@B82]). Similarly, we observed in our exploratory analysis at least a small atDCS related benefit on consolidation (\"offline\" effect), but after the first night only. Memory is a highly dynamic process (Schacter and Addis, [@B92]). After initial encoding memory traces are unstable (i.e., vulnerable to interference or modification), but become stabilized and more resistant to disruption over time, a process referred to as \"consolidation\" (McGaugh, [@B63]). Thus, the \"offline\" effect after the first night might indicate a relatively larger impact of atDCS on more \"labile phases\" of consolidation (Lally et al., [@B55]; Richmond et al., [@B85]). It is also consistent with the assumption that atDCS is most effective at near-threshold and fragile performance level (Berryhill et al., [@B9]). In contrast, Reis et al. ([@B83], [@B82]) reported increased offline skill gain over the course of five consecutive training days after receiving atDCS relative to sham. In fact, even consolidated memory can be returned into a transiently labile state again by reactivation (Lee, [@B57]; Dudai, [@B20]), e.g., by repeated retrieval, which then require another period of stabilization (reconsolidation; Sara, [@B90]; Nader and Hardt, [@B67]). However, given that the probability of being destabilized upon reactivation depend among others on strength of a memory (Dudai and Eisenberg, [@B21]; Bustos et al., [@B12]; Winters et al., [@B104]), susceptibility for modifications of a memory trace might change during training progress. Further, both training and atDCS implicate neuroplastic processes, which probably overlap in their underlying mechanisms, but may act on different time scales to boost task-related activity (Au et al., [@B4]). Besides, repeated sessions could affect intrinsic brain activity and might interact with training-induced plasticity (Moller et al., [@B66]). Therefore, non-optimal timing between those neuroplastic processes might have masked atDCS-related effects in the present study. Note also that optimal timing might differ for motor task and cognitive task (Summers et al., [@B97]). Even within the cognitive domain the optimal time gap between practice sessions seems to be moderated by the nature (here mental difficulty) of the task (Donovan and Radosevich, [@B19]). In sum, better synchronization of task- and stimulation-induced plasticity over training may help to improve atDCS impact in multi-day applications, a hypothesis to be tested in future studies. Moreover, *within*-session breaks of tDCS (spaced stimulation) may promote metaplasticity (Goldsworthy et al., [@B29]) and might also lead to longer-lasting effects. Hence, both spacing within- and between atDCS sessions seems to be critical in modulating neuroplasticity and should be carefully taken into account in designing multi-day atDCS training protocols.\n\nIndividual differences\n----------------------\n\nProfound variability amongst individuals in responsiveness to atDCS effects has been noted independent of protocol or electrode montage (e.g., Jantz et al., [@B46]). Accordingly, proposed moderators like gender, age, education, sequence of stimulation and baseline performance (e.g., Tseng et al., [@B100]; Meiron and Lavidor, [@B65]; Krause and Cohen Kadosh, [@B52]; Learmonth et al., [@B56]; Li et al., [@B59]; Santarnecchi et al., [@B89]; Hsu et al., [@B40]; Summers et al., [@B97]; Fertonani and Miniussi, [@B25]) were included in our statistical model. In addition, we considered variation in common genetic polymorphisms like APOE e4 carrier status (Elder and Taylor, [@B22]). However, no substantial impact on outcome of any of the selected \"modulators\" were found. Since intra-individual changes are known to increase with advanced age (Macdonald et al., [@B60]), we also monitored affective state (prior to each training) and sleep characteristics of the preceding night for each training+stimulation day, but found no differences between stimulation condition. However, most of the evidence related to inter-individual differences stems from single session studies, but both CT (Jaeggi et al., [@B43]; Katz et al., [@B48]) and CT in combination with atDCS can be potentially influenced by individual variability. Thus, it is important to further identify, control and/or counteract different sources of variability in multi-session studies, to allow for more reliable atDCS effects across individuals, see also (Shin et al., [@B94]; Davis, [@B16]) for further discussion.\n\nEffects on trained and untrained (transfer) material\n----------------------------------------------------\n\nThe present study also explored the translational potential of combined intervention on trained (LOCATO-15; OLL) and untrained (verbal and visuo-constructive) memory functions. Previous studies using a combined atDCS+training approach had demonstrated beneficial impact on delayed parameters of the trained task and untrained functions, even in the absence of immediate effects (Jones et al., [@B47]; Wang and Voss, [@B102]; Stephens and Berryhill, [@B96]; Antonenko et al., [@B2]; Ruf et al., [@B86]). In contrast, we could not substantiate these positive findings, in line with other studies on transfer effects (see Elmasry et al., [@B23]; Nilsson et al., [@B70]). Several factors may account for our present results. First, training schedule may have been too short to modulate gains on other tasks, although OLM performance was high at the end of 3-day training (on average 87%). Further, performing successfully in trained task might motivate to perform well on other tasks following training (Hayes et al., [@B33]) overriding subtle atDCS induced modulations. Second, since we aimed to study augmentation of combined atDCS+training intervention we did not include a \"atDCS only\" control group which might be necessary to disentangle effects above and beyond training itself. Third, studies showing positive transfer effects used a parallel-group design. It is conceivable that complex carry-over effects due to cross-over design emerged that may have obscured statistical differences between stimulation and sham (for a similar discussion see also Wang and Voss, [@B102]). Because long-lasting transfer effects are highly relevant in the field of neurorehabilitation, careful design including appropriate control is needed to ascertain the mechanisms underlying successful transfer gains.\n\nLimitations\n-----------\n\nSeveral limitations should be acknowledged when interpreting these findings. First, we did not study oppose polarities. Consequently, using a cathodal tDCS control might provide more comprehensive information. However, the dichotomous approach of anodal/cathodal stimulation and associated improvement vs. impairment via increased/decreased neuronal excitability is mainly based on results of primary motor cortex stimulation studies (Berryhill et al., [@B9]), and might not apply to the cognitive function under study here. Second, only subjects were blinded to the stimulation condition. Although double-blind approach is considered as gold-standard, the risk of biasing subject\\'s performance by investigator can be considered minor, because primary outcome was measured computerized. More importantly, the study was conducted in a randomized and sham controlled manner. Third, no modeling of current flow was employed (Chaieb et al., [@B14]). Thus, on the one hand stimulation site might not have been optimal for all subjects. However, the electrode-montage was not intended to induce focal stimulation. Instead, we aimed for stimulation of a cortical area connected to the fronto-hippocampal-parietal memory network sub serving the task under study (Nyberg, [@B74]). On the other hand, not only electrode site but also shape, size and connector position can interfere with the intensity and spatial distribution of the electric field generated around them. Direct measurements of electric fields are difficult to implement (Opitz et al., [@B75]). Nevertheless, compared to other tDCS studies the connectors were located relative far from each other (connector of the active electrode was positioned at the posterior edge and was relative distant from the return electrode). Hence, in line with the rough rule of thumb reported by Saturnino et al. ([@B91]) the remote location of both connectors should result in strengthening the electric field in the brain region underneath the temporoparietal anode.\n\nConclusion and outlook {#s5}\n======================\n\ntDCS, generally known to be a safe neuro-stimulation technique, was well-tolerated in healthy older subjects with stimulation and training-sessions over 3 consecutive days. As applied in its present form, the findings did not support the notion that the intervention combining atDCS and training improves memory formation in OLM. However, combined atDCS+training approach remains a fundamentally important goal in research on cognitive aging. Several factors may underlie the negative findings, as discussed in this report. Systematically addressing these factors in future studies may provide valuable information in order to advance in-depth knowledge in basic tDCS research to generate more robust results in individuals.\n\nAuthor contributions {#s6}\n====================\n\nNK contributed to conception, designed the study, performed analysis and interpretation, contributing to drafting and revising the paper; AVCdS contributed to acquisition, performed analysis and interpretation, drafting and revising the work; AF contributed to conception, designed the study, and critically revised the paper; UG contributed to analysis and interpretation of the data and revising the work; MC, J-MH, AG, SH, and SK contributed to acquisition and drafting of work; All authors provided final approval for the version to be published.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\nThe authors thank Anke Nie\u00dfen, Maria Meier, Almut D\u00fcnnebeil, Thorge Profitlich, Sonja Fabian, and Dr. Sven Passmann for help with data acquisition and the Lab of Prof. Dan Rujescu (University Halle Germany) for performing blood analysis.\n\n**Funding.** This work was supported by the Deutsche Forschungsgemeinschaft (DFG, FL 379-10) and the Bundesministerium f\u00fcr Bildung und Forschung (BMBF, 01GQ1424A).\n\n[^1]: Edited by: Gregor Thut, University of Glasgow, United Kingdom\n\n[^2]: Reviewed by: Ivan Alekseichuk, University of G\u00f6ttingen, Germany; Branislav Savic, University of Bern, Switzerland\n\n[^3]: This article was submitted to Perception Science, a section of the journal Frontiers in Neuroscience\n\n[^4]: \u2020Shared first authorship.\n"}
+{"text": "Introduction {#s1}\n============\n\nScience is disproportionately produced at research centers within a few select regions [@pbio.1001740-Bornmann1],[@pbio.1001740-National1]. This distribution contributes to \"brain drain\"---the cultural and geographical separation of researchers from their communities of origin [@pbio.1001740-Franzoni1]. In places lacking research centers, brain drain precludes achieving a critical mass of scientific expertise and the development of science, technology, engineering, and mathematics (STEM). Displaced scientists gradually become disconnected from their home communities and colleagues, presenting a challenge to maintaining research collaborations that could benefit their communities of origin. Insidiously, dispersion also presents socio-cognitive challenges to scientists who see themselves as underrepresented in the larger culture of science [@pbio.1001740-Payton1]--[@pbio.1001740-Chang1].\n\nSocial networks hold enormous promise for \"connecting\" dispersed groups and providing new opportunities for fellowship and mentorship among underrepresented communities in science. Faced personally with these obstacles, in 2006, we created Ciencia Puerto Rico (CienciaPR; [www.cienciapr.org](http://www.cienciapr.org)), an online network that connects scientists with geographic, academic, and/or cultural ties to Puerto Rico. CienciaPR was built to counteract the negative effects of scientific brain drain by: (1) promoting scholarly interaction among self-identified members of an otherwise dispersed community; (2) providing visibility to diverse scientific role models; and (3) supporting research and science education through initiatives that culturally resonate with our community of origin.\n\nHere, we present CienciaPR\\'s design and discuss how we leverage our membership to enhance science education and mentoring of Puerto Rican students. Looking beyond our own community, we suggest how our efforts can be translated to similarly dispersed populations. By growing and supporting scientific diversity, we believe social networking can democratize the scientific enterprise and more broadly distribute its benefits.\n\nConceptualization and Implementation of CienciaPR {#s2}\n=================================================\n\nThe Puerto Rican scientific community is highly dispersed---64% of Puerto Rican PhD STEM students and 44% of the Puerto Rican STEM doctorate workforce resides outside the Puerto Rican archipelago [@pbio.1001740-National2]--[@pbio.1001740-Special1]. Recent emigration trends have exacerbated dispersion [@pbio.1001740-RodrguezAyuso1]. A virtual space that connects the Puerto Rican scientific community thus represents a powerful means of addressing the unique challenges faced by this population, which shares elements of both diaspora and minority communities.\n\nTwo aspects were key in the design of CienciaPR. First, we conceptualized the network as a site for \"anyone interested in science and Puerto Rico.\" This definition was broad by design to account for the idiosyncrasies of cultural, ethnic, and national identification, and to promote the inclusion of anyone interested in contributing to the Puerto Rican scientific community regardless of their place of origin. While most of our membership consists of Puerto Rican scientists, a broad definition helps attract scientists around the world interested in research or educational collaborations with Puerto Rico.\n\nSecond, we gave important consideration to user profiles, a typical feature of social networking sites [@pbio.1001740-Boyd1], but tailored them for scientists, gathering information about research interests, publications, institutional affiliation, mentoring, entrepreneurial interests, and training history (see <http://bit.ly/1bXN6dy>, for example). At CienciaPR.org, users can open and populate a profile, free of cost. During the first 6 years of the website, name and email were the only fields required to open a profile. Despite this, a majority of CienciaPR members have chosen to list their institution (54%), field(s) of scientific interest (58%), and training or work position (70%). Information collected through CienciaPR profiles allows the creation of a map of the community\\'s geographic footprint and collective capacity ([Figure 1C](#pbio-1001740-g001){ref-type=\"fig\"}). Profiles serve both to identify individuals with specific expertise for mentoring or collaboration, and to provide visibility to a community of scientists, otherwise invisible due to geographic dispersion and underrepresentation.\n\n![CienciaPR website use and member characteristics.\\\n(A) Number of members who have registered with CienciaPR since 2006. (B) Average yearly visits to CienciaPR.org, from October 2008--September 2012, based on Webalizer data. CienciaPR switched to GoogleAnalytics after September 2012. (C) Map representation of CienciaPR members\\' geographic dispersion (portion of map is cropped for display purposes). CienciaPR members are in 48 countries, 47 states (not including the territory of Puerto Rico), and over 185 universities in the US. Map made with Tableau Public (D--F) Distribution of members based on (D) work sector, (E) training stage, and (F) broad scientific discipline of interest. Not all members indicated this information on their profiles. Percentages are based on total number of respondents (n). Data as of September 8, 2013. (G--I) Recent CienciaPR website activity from June 1, 2013--August 31, 2013 based on Google Analytics for (G) website visits by geographic region, (H) website visits by source, and (I) page views by website section. Percentages are based on (n), total number of visits (G,H) or pageviews (I).](pbio.1001740.g001){#pbio-1001740-g001}\n\nA number of social networking tools (peer-to-peer messaging, maps of nearby members, message boards, events calendar, personal scientific blogs, and automated matching of members with similar interests) facilitate interaction and keep the website dynamic and updated through member-driven content. In addition, we add value by creating listings---many updated in real-time through real simple syndication (RSS)---that display funding opportunities, job openings, and Puerto Rican scientific organizations. This approach allows us to crowd-source relevant content in a cost-effective manner, publishing an average of seven new pieces of member-driven content per week.\n\nThe original CienciaPR website, developed pro bono by one of the authors (DC), consisted of an SQL database with a custom-designed front-end and content management system. This proved sufficient for most of our operations during the first 6 years. The community\\'s interest in our website, as evidenced by member registrations and visits ([Figure 1A and 1B](#pbio-1001740-g001){ref-type=\"fig\"}), served to secure funding from the Puerto Rico Science, Technology, and Research Trust for the development of a more sophisticated Drupal7-based platform (launched September 2012). Both versions of the website were intentionally designed with open source platforms to facilitate sharing and use by other communities. Though CienciaPR chose Drupal for its customizability, there are other simple and affordable social networking solutions (e.g., Ning, SocialEngine, Elgg, BuddyPress) available to groups interested in replicating our approach.\n\nNews about the launch of the original website was initially sent to scientific leaders in Puerto Rico but thereafter membership continued to grow with no marketing efforts, based on word-of-mouth and organic website traffic. In 7 years CienciaPR\\'s membership has soared to 6,364 (as of September 8, 2013), making it the largest database of Puerto Rican scientists across the world ([Figure 1C--1F](#pbio-1001740-g001){ref-type=\"fig\"}). Based on national surveys, we estimate CienciaPR\\'s membership accounts for \u223c30% of all Puerto Rican doctorates in STEM and 65%--70% of all Puerto Rican graduate students across the US [@pbio.1001740-Special2].\n\nCienciaPR has an engaged following. Almost 29% of traffic to the website is from direct sources ([Figure 1H](#pbio-1001740-g001){ref-type=\"fig\"}; direct URL entries and member newsletter) and CienciaPR\\'s Facebook page has a weekly engagement average of 691 unique users. Member profiles and our message boards---displaying research training and funding opportunities---are our most frequented sections ([Figure 1I](#pbio-1001740-g001){ref-type=\"fig\"}). Membership engagement is reflected in the thousands of peer-messages and board posts over the lifespan of the website.\n\nFrom Connectivity to Action {#s3}\n===========================\n\nCienciaPR\\'s social network enables the identification of both needs and resources within our community for innovative research and educational interventions. Below, we enumerate online and offline initiatives, describe how they emerged organically from the needs of our membership, and how online tools were used to leverage the knowledge-base of the network to arrive at crowd-sourced, community-driven solutions.\n\nCulturally Relevant Science Communication and Education {#s4}\n=======================================================\n\nScience learning is enhanced when concepts are made relevant to an individual\\'s context and culture [@pbio.1001740-Osborne1]--[@pbio.1001740-Laughter1]. Yet in Puerto Rico, as in most minority communities and developing countries, scientific concepts are seldom illustrated in a culturally relevant fashion [@pbio.1001740-Massarani1]--[@pbio.1001740-GonzlezEspada1].\n\nScientists can play a key role in explaining the meaning and importance of research findings to their communities [@pbio.1001740-Gastel1]--[@pbio.1001740-LesliePelecky1]. CienciaPR has leveraged its unique network to enhance formal and informal science education in Puerto Rico and among Spanish-speaking communities through collaborations with eight national and international media outlets (<http://bit.ly/1a9xCzy>) [@pbio.1001740-FeliMjer1]. We established collaborations by emailing newspaper editors or bloggers with representative articles that could be published in their venue. In our experience, many media outlets lack access to scientific content and most venues we contacted responded eagerly. The main challenge was ensuring scientific topics were appropriately explained for a lay audience.\n\nTo address this hurdle, early on we identified two volunteer editors within our network (MFM and WGE) who could serve as key liaisons between CienciaPR members and media partners. Written work submitted by our members is first curated and edited by these expert science communicators before being sent to media collaborators. We prioritize articles that talk about science performed in Puerto Rico or by Puerto Rican scientists or that make universal science concepts meaningful and relatable by using examples from the Puerto Rican context. While partners reserve the right to publish the articles submitted by CienciaPR, 89% have been published (<http://bit.ly/1frwIoj>). Some articles are also distributed as podcasts (through CienciaPR.org and iTunes) and broadcasted through local radio stations (<http://bit.ly/16uTo3h>).\n\nIn 2006, before CienciaPR\\'s collaboration, very few of the science stories in El Nuevo D\u00eda (the main Puerto Rican newspaper) were contributed by scientific experts [@pbio.1001740-Massarani1]. CienciaPR\\'s media collaborations have increased the amount of culturally relevant scientific news content in Puerto Rico with 385 articles and 210 podcasts released as of October 2013, all authored by scientists.\n\nInspired by its successful media collaborations, in 2009 CienciaPR crowd-sourced a collection of short essays about science for the general public. Twenty-three scientists contributed 61 essays to the book *\u00a1Ciencia Boricua! Ensayos y An\u00e9cdotas del Cient\u00edfico Puertorro* [@pbio.1001740-GonzlezEspada2]. The articles showcase world-class research performed in Puerto Rico, or by Puerto Ricans, and research performed internationally. In all cases, essays illustrate scientific concepts using examples from the Puerto Rican context, culture, and landscape. This approach serves to underscore the relevance of international scientific discoveries locally, as well as celebrate local contributions to the international scientific enterprise.\n\nThrough our network we learned that teachers were using the book to help contextualize science in their curricula. Although not developed as a textbook, we followed teachers\\' lead and, with the book as centerpiece, undertook a pilot project to increase Puerto Rican elementary and middle school students\\' interest in science. Results from the pilot study suggest that students\\' interest in science increased after exposure to context-relevant material [@pbio.1001740-GonzalezEspada1].\n\nProviding Role Models and Advancing Careers {#s5}\n===========================================\n\nYoung people\\'s perception of scientists and their ability to envision a successful scientific future is influenced by access to diverse representations of scientists [@pbio.1001740-Painter1]--[@pbio.1001740-Bensimon1]. We have placed a high priority on increasing the visibility of Puerto Rican and Hispanic scientific role models by maintaining member profiles publicly accessible, celebrating their successes through our blog and social media accounts, and increasing their coverage in news media (see above). Additionally, we publish monthly bilingual online features (67 as of October 2013) highlighting the life and work of accomplished members, showcasing their successes, but also their challenges, drive to succeed, and interest in giving back to their communities.\n\nOur efforts extend beyond our website by organizing offline events that promote community and mentoring. For instance, we coordinate annual social events during national scientific conferences (e.g., Society for Neuroscience, Ecological Society of America, American Society for Microbiology) that give attendees the opportunity to network and thus mitigate the effects of dispersion. These events are often organized in collaboration with Puerto Rican chapters or members of the scientific societies.\n\nCienciaPR has also organized meetings to tackle issues such as the quality of science education in Puerto Rico and career advice to scientists-in-training. Connections with academic leaders and institutions in Puerto Rico have been crucial for these endeavors. For example, in 2011, in collaboration with the Research Centers in Minority Institutions (NIH-RCMI) of the University of Puerto Rico, CienciaPR held a symposium designed to guide students interested in STEM careers, but who lacked access to research-intensive environments. Over 120 students from 15 college campuses across Puerto Rico attended presentations that discussed how to apply, be successful, and find funding towards the completion of doctorate degrees in biomedical research. Eighty-eight percent of students felt the symposium \"helped clarify their interest in a graduate science program\" and 90% indicated that \"they were interested in pursuing a scientific career.\"\n\nChallenges and Lessons Learned {#s6}\n==============================\n\nThough our network is at the heart of CienciaPR, it would not have reached its full potential without a team of volunteers who emerged as CienciaPR\\'s governing and operating structure ([Box 1](#pbio-1001740-box001){ref-type=\"sec\"}). We have found two main challenges in running a grassroots, non-profit organization led by volunteers: limited funds and limited availability of human resources. Community members are eager to make time to volunteer, because they appreciate the potential of CienciaPR to help their careers and give back to their community. Nonetheless, academic and professional careers force time and effort commitments to be variable. Though most of CienciaPR\\'s initiatives have been implemented with minimal funding, in-kind contributions, and through the donation of volunteers\\' time, much more could be achieved with additional funds and full-time staff. Lack of these resources has restricted the potential reach and impact of the organization and resulted in lost opportunities. Though the use of social networking tools in combination with community-based volunteer leadership has resulted in the initiation of much needed interventions, we are currently working to establish more structured governance and a sustainable financial plan to build on the success of our initial years.\n\nBox 1. CienciaPR\\'s Volunteer Leadership at Work {#s6a1}\n------------------------------------------------\n\nCienciaPR has established a volunteer-based governing and operating structure (<http://bit.ly/18JLYHi>). To aid others interested in this approach we answer key questions regarding our leadership model:\n\n**How many volunteers are needed to run CienciaPR?** The current CienciaPR Team consists of nine PhD scientists. Additionally, seven volunteers help with the online platform and other initiatives.\n\n**How do you find volunteers?** Volunteers self-identify through our network (through email or by completing a webform). After discussing their interests, time, and skills, they are assigned to functions or projects as needed. Over the years, a group of leaders has emerged through their dedication and contributions. There is no lack of volunteers (we have a waitlist of 52), and in fact one challenge is the ability to manage the number of people interested in helping out.\n\n**How do volunteers find time?** It is challenging to find time amidst the demands of a scientific career, but we have found that volunteers are motivated by their desire to give back to the community. We avoid overextending our volunteers by discussing their time availability and matching their tasks with their career goals and interests.\n\n**How is the online platform managed?** The CienciaPR website has an easy-to-use content management system that keeps operational costs down. One of our team members (GGM) manages the website, develops new website sections, and performs basic web maintenance. She trains and supervises other volunteers who contribute curated content (e.g., news, events, training and funding opportunities).\n\n**How are initiatives created and implemented?** CienciaPR\\'s projects are conceived and implemented by members and volunteers. This contrasts markedly from conventional organizations, where staff devises initiatives without much participation from its audience. This crowd-sourced, community-driven approach means initiatives are directly responsive to perceived needs, breeds creativity, and assures that volunteers are matched to efforts that interest them. The participatory nature of volunteer leadership has driven the enhancement of the network and future projects.\n\n**How is the sense of community maintained among leadership?** Though the CienciaPR Team is as geographically dispersed as the CienciaPR network, volunteers are in almost daily contact via email, and hold monthly teleconferences where we discuss projects, roles, and assignments. CienciaPR\\'s open and participatory decision-making process enables volunteers to contribute ideas and voice concerns. The team\\'s bond is further solidified through regular celebrations and recognitions of volunteers\\' contributions and achievements (within or outside of CienciaPR).\n\nReplicability to Other Communities {#s7}\n==================================\n\nNetworks of diaspora or minority scientists are uniquely positioned to contribute to their communities by counteracting geographic dispersion and isolation [@pbio.1001740-Meyer1]--[@pbio.1001740-NIH1]. CienciaPR has demonstrated that an online approach can be leveraged to mitigate these dual challenges. We believe certain elements have enabled the growth and impact of CienciaPR: a social network approach, community-sourced initiatives, and collaboration with local scientists. While our efforts have focused on Puerto Rico, these aspects can enable other dispersed groups to connect with home communities for outreach, education, research collaboration, or entrepreneurial efforts. Organizations such as the Society for the Advancement of Hispanics/Chicanos and Native Americans in Science (SACNAS) (sacnas.org), Just Garcia Hill (justgarciahill.org), Minority Postdoc (minoritypostdoc.org), and Scientista Foundation (scientistafoundation.com), to mention a few, have also demonstrated the power of online networks to address the challenges facing minorities in science. With CienciaPR, we took this concept and extended it beyond mentoring and career advice, to connect scientists with their community for educational outreach and science communication.\n\nBroadening the participation of underrepresented groups across the world of science is critical for the progress of research and the development of countries [@pbio.1001740-National4],[@pbio.1001740-UN1]. CienciaPR stands as an example of a social networking model applied to geographically dispersed scientists that is cost-effective, feasible, and impactful. We extend our hand to other groups that want to use online technologies to apply their passion for science and desire to give back to their communities.\n\nWe would like to acknowledge other current and past CienciaPR volunteers. These include Nora Torres, Alexis Valent\u00edn Vargas, Francis Gonz\u00e1lez, Reyna I. Mart\u00ednez De Luna, Paola Giusti, Edwin Rosado, Elvin Estrada, Jorge Herrera, Aixa Alem\u00e1n, Odmara Barreto Chang, Wilfredo de Jes\u00fas Rojas, Idalia Massa, and Ver\u00f3nica A. Segarra. We thank James O\\'Malley from Altamente GS, LLC for the recent website redesign and support. None of this work would have been possible without the commitment and interest of the global Puerto Rican scientific community and all the scientists, friends, and family members who have contributed to CienciaPR initiatives.\n\n[^1]: The Community Page is a forum for organizations and societies to highlight their efforts to enhance the dissemination and value of scientific knowledge.\n\n[^2]: The authors have declared that no competing interests exist.\n\n[^3]: Current address: Department of Biostatistics, University of Washington, Seattle, Washington, United States of America\n\n[^4]: Current address: Department of Plant and Microbial Biology, University of California, Berkeley, Berkeley, California, United States of America\n\n[^5]: Current address: Department of Anatomy and Neurobiology, University of Puerto Rico, R\u00edo Piedras, Puerto Rico, United States of America\n"}
+{"text": "Introduction\n============\n\nEsophageal squamous cell carcinoma (ESCC) is one of the most malignant tumors worldwide ([@b1-ijo-52-05-1443]--[@b3-ijo-52-05-1443]) and the sixth most fatal cause of cancer-related death. Squamous cell carcinoma (SCC) is the main histological type of this cancer. Northern and central China is part of the \\'Asian belt\\' which has a very high incidence of ESCC; there are \\>100 cases per 100,000 annually. The urgency of preventing and curing ESCC is obvious ([@b4-ijo-52-05-1443],[@b5-ijo-52-05-1443]).\n\nSimilar to other types of cancer, the development of ESCC is believed to be a multi-step process caused by the accumulation of activated oncogenes and inactivated tumor suppressor genes (TSGs) ([@b6-ijo-52-05-1443]). TSGs can be inactivated by both genetic and epigenetic mechanisms. Genetic deletions and point mutations disrupt TSG functions, and epigenetic mechanisms, including CpG island promoter methylation and histone modifications, frequently lead to the loss of TSG functions and are involved in tumor development and progression ([@b7-ijo-52-05-1443]). The aberrant methylation of CpG islands leads to gene silencing, resulting in TSG inactivation, which can increase the rate of tumor formation by disabling multiple normal cellular processes, such as apoptosis and cell cycle progression ([@b7-ijo-52-05-1443]). In breast cancer and renal clear cell carcinoma ([@b8-ijo-52-05-1443]), methylation of erythrocyte membrane protein band 4.1 like 3 (EPB41L3) disables its tumor suppressive functions ([@b9-ijo-52-05-1443],[@b10-ijo-52-05-1443]), suggesting that this gene is a potential example of TSG inactivation in human cancers. However, the role of EPB41L3 in ESCC remains unclear. EPB41L3, which belongs to the protein 4.1 family, is a membrane skeletal protein 4 that is commonly expressed in various human tissues. It is localized to sites of cell-cell contact and functions as an adapter protein, linking the plasma membrane to the cytoskeleton or associated cytoplasmic signaling effectors. Subsequently, EPB41L3 facilitates their activities in different pathways ([@b11-ijo-52-05-1443]). Using a methylation microarray analysis, our group has demonstrated that EPB41L3 is methylated in ESCC tissue ([@b12-ijo-52-05-1443]). Our subsequent study revealed that EPB41L3 suppressed tumor cell invasion and inhibited matrix metallopeptidase (MMP)2 and MMP9 expression in ESCC cells ([@b13-ijo-52-05-1443]). However, the underlying mechanism by which EPB41L3 displays a tumor suppressive role has not been fully elucidated. Thus, aiming to improve clinical management and prolong the life expectancy of ESCC patients, the present study investigated the tumor suppressive functions of EPB41L3 in ESCC and its potential as a prognostic indicator. To explore the molecular function of EPB41L3 in ESCC, the expression of EPB41L3 in seven cell lines was measured. Then a tissue microarray with 97 pairs of tumor and paired normal tissues was used in order to assess its methylation status. The results demonstrated that expression of EPB41L3 was decreased in both ESCC cells and tissues and its methylation rate was increased compared with normal cells and tissues. Multiple tumor suppressors have been reported to function by inducing apoptosis, and loss of tumor suppression results in reduced apoptotic activity and increased tumor growth ([@b14-ijo-52-05-1443]--[@b17-ijo-52-05-1443]). Therefore the role of EPB41L3 in ESCC apoptosis and growth was examined. The results suggested that EPB41L3 may be a potential tumor suppressor and prognostic indicator in ESCC.\n\nMaterials and methods\n=====================\n\nTissue microarray and immunohistochemistry\n------------------------------------------\n\nA ESCC tissue microarray was used to examine the expression of EPB41L3. The tissue microarray was purchased from Shanghai Outdo Biotech Co., Ltd. (Shanghai, China) and contained 97 pairs of ESCC samples and their para-carcinoma tissues. The patients were selected based on a clear pathological diagnosis of relatively early stage (Stages IA-IIIA) ESCC and follow-up data dated back to July 2007. Their clinical features were listed in [Table I](#tI-ijo-52-05-1443){ref-type=\"table\"}. None of the patients had received chemotherapy or radiotherapy prior to surgery. Immunohistochemistry on these tissues was performed following a previously published protocol ([@b9-ijo-52-05-1443]). Scoring of EPB41L3 expression was based on the ratio of positive cells (score 0, 0--5% positive cells; 1, 6--35%; 2, 36--70%; 3, \\>70%) and staining intensity (score 0, no staining; 1, weak staining; 2, moderate staining; 3, strong staining). The final score was calculated as follows: positive cell score x staining intensity score. The samples were marked as follows: \\'\u2212\\' for a score of 0--1, \\'+\\' for a score of 2--3, \\'++\\' for a score of 4--6, and \\'+++\\' for a score of \\>6. Low expression was defined as a total score of \\<4 and high expression as a total score \u22654. For further analysis, \\'\u2212\\' and \\'+\\' were considered as low expression, while \\'++\\' and \\'+++\\' as high expression. These scores were determined independently by two senior pathologists (Yu Wang and Jingyu Li; Department of Pathology, Zhujiang Hospital) in a blinded manner ([@b18-ijo-52-05-1443]).\n\nCell lines\n----------\n\nThe non-neoplastic esophageal epithelial cell line Het-1a and six ESCC cell lines (Eca-109, TE-1, CaEs17, KYSE-150, KYSE-450 and KYSE-510) were obtained from the Research Center of Clinical Medicine at Nanfang Hospital in Guangzhou, China. Het-1a was cultured in RPMI-1640 medium (Life Technologies; Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 10% fetal bovine serum (Merck Millipore, Billerica, MA, USA) and 2 mM L-glutamine. CaEs17 and Eca-109 were cultured in DMEM (Life Technologies; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (Merck Millipore), and the other four cell lines were cultured in RPMI-1640 supplemented with 10% fetal bovine serum.\n\nDNA and RNA extraction\n----------------------\n\nTotal DNA and RNA were extracted using the Omega Tissue DNA kit and the Omega Total RNA kit I (both from Omega Bio-Tek, Inc., Norcross, GA, USA) respectively, according to the manufacturer\\'s instructions.\n\nReverse transcription-polymerase chain reaction (RT-PCR) and reverse transcription -quantitative PCR (RT-qPCR)\n--------------------------------------------------------------------------------------------------------------\n\nUsing the total RNA extracted from cells described above, reverse transcription was performed using the PrimeScript RT Reagent kit with gDNA Eraser (Takara Bio Inc., Otsu, Japan). qPCR was performed using SYBR Premix Ex Taq reagent (Takara Bio Inc.). The sequences of the reverse transcription-PCR primers were as follows: EPB41L3 forward, 5\u2032-GTAGTGGTCCATAA AGAGACAGAGA-3 and reverse, 5\u2032-GATACAAGTCAGTTGG GTTAGAAGA-3; GAPDH forward, 5\u2032-TGCTGAGTATGTCGT GGAGTCT-3\u2032 and reverse, 5\u2032-CCCTGTTGCTGTAGCCATA TTC-3\u2032. The gel concentration was 1% and the results were visualized under a UV lamp. qPCR was performed using a LightCycler 480 system (Roche, Basel, Switzerland). The qPCR reaction protocol was 30 sec at 95\u00b0C followed by 40 cycles of 5 sec at 95\u00b0C and 20 sec at 60\u00b0C. Quantification was performed using the 2^\u2212\u0394\u0394Cq^ method ([@b19-ijo-52-05-1443]).\n\nMethylation-specific PCR (MSP)\n------------------------------\n\nTotal DNA was extracted as aforementioned, and bisulfite-modified using the BisulFlash DNA Modification kit (EpiGentek, Farmingdale, NY, USA) according to the manufacturer\\'s protocol. The modified DNA was maintained at \u221220\u00b0C until PCR amplification. MSP was performed with 0.25 *\u00b5*l Taq HS, primers (500 nM), 50 ng total genomic DNA, and dNTP Mixture, and 10X PCR Buffer (both from Takara Biotechnology Co., Ltd., Dalian, China) in a total reaction volume of 20 *\u00b5*l. The primer pairs for EPB41L3 were as follows: to detect unmethylated sites, US, 5\u2032-TTTGTGTATTGT TGTTGAGGAGTG-3\u2032 and UAS, 5\u2032-CACAATCCCCCACTCCA AAAAACA-3\u2032; and to detect methylated sites, MS, 5\u2032-GCAGTG CAAAGTGATACTTC-3\u2032 and MAS, 5\u2032-TCTGGTGGATAAA ATTTCACAT-3\u2032. The thermocycling conditions for the MSP were: 95\u00b0C for 10 min, then 38 cycles of 94\u00b0C for 30 sec, 58\u00b0C or 60\u00b0C or 55\u00b0C for 30 sec and 72\u00b0C for 30 sec, followed by 72\u00b0C for 5 min, as previously described ([@b20-ijo-52-05-1443]). The results were visualized using a BiQ Analyzer (Max Planck Institute for Informatics, Saarbr\u00fccken, Germany).\n\nBisulfite treatment and pyrosequencing analysis\n-----------------------------------------------\n\nGenomic DNA was extracted from seven cell lines using the Tissue DNA kit (Omega Bio-Tek, Inc.). Extracted DNA was bisulfite-modified using the BisulFlash DNA Modification kit (EpiGentek). The average methylation rates of 33 CpG sites that are located at \u2212141551, \u2212141494, \u2212141439 and \u2212141379 bp in the EPB41L3 promoter were detected. For EPB41L3, methylation rates of \\>20%, 10--20% and \\<10% were defined as hypermethylation, partial methylation, and non-methylation respectively ([@b8-ijo-52-05-1443]).\n\nFor the methylation analysis, the extracted seven DNA samples were subjected to bisulfite conversion using the Bisulfite Conversion kit (Qiagen GmbH, Hilden, Germany), according to the manufacturer\\'s instructions. Then, the purified bisulfate-converted DNA was used to perform MSP with biotinylated primers and Platinum Taq DNA polymerase (Kapa Biosystems, Inc., Wilmington, MA, USA). All the primers used for PCR were designed using PyroMark Assay Design 2.0 (Qiagen GmbH), synthesized by BGI (Shenzhen, China), and presented in [Table II](#tII-ijo-52-05-1443){ref-type=\"table\"}. The PCR products were purified using a Qiaquick PCR purification kit (Qiagen GmbH), and single-stranded DNA was prepared using Dynabeads M280 streptavidin (Thermo Fisher Scientific, Inc.). The samples were analyzed using PyroMark Q96 ID (Qiagen GmbH). Analysis of the results was performed with the PyroMark Q96-CpG software (Qiagen GmbH).\n\n5-Aza-2-deoxygcytidine treatment\n--------------------------------\n\nTo determine whether demethylation could restore the expression of EPB41L3 in ESCC cells, 2\u00d710^5^ cells were seeded in each well of a 6-well plate and treated with 50 *\u00b5*mol/l 5-Aza-2\u2032-deoxycytidine (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) for 3 days. The medium was changed with fresh medium containing 5-Aza-2\u2032-deoxycytidine every 24 h. Total RNA and protein was then extracted, and EPB41L3 expression was measured by RT-PCR and western blotting ([@b6-ijo-52-05-1443]).\n\nCell counting kit-8 (CCK-8) cell viability assay\n------------------------------------------------\n\nCell viability was assessed by a CCK-8 assay (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). Vec-150, epb-150, vec-510, epb-510 cells, si-510 and si-NC (5\u00d710^3^ cells/well) were plated separately into 96-well plates. After being plated for 24, 48 or 72 h, the CCK-8 reagent was added to each well, and cells were incubated for 2 h at 37\u00b0C. The absorbance was measured at 450 nm by spectrophotometry.\n\nWestern blot analysis\n---------------------\n\nWestern blotting was performed according to the standard protocol ([@b21-ijo-52-05-1443]) with the following antibodies: EPB41L3 (1:2,000; Abcam, Cambridge, MA, USA), caspase-8 (1:1,000), caspase-9 (1:1,000), caspase-3 (1:1,000), CyclinA (1:2,000), CyclinD (1:1,000), CyclinE (1:1,000) (all from Cell Signaling Technology, Inc., Danvers, MA, USA), CyclinB1 (1:1,000), Cyclin-dependent kinase 1 (CDK1, also known as Cdc2; 1:1,000) (both from ProteinTech Wuhan Sanying Biotechnology, Wuhan, China). GAPDH (1:8,000; ProteinTech Wuhan Sanying Biotechnology) was used as loading control.\n\nFlow cytometry\n--------------\n\nTo measure dividing cells by flow cytometry, KYSE-150 and KYSE-510 cells transfected with Vec-NC or Vec-epb were fixed in 70% ethanol overnight, stained with propidium iodide the next day, and DNA content was analyzed using an LSR FORTESSA (BD Biosciences, Franklin Lakes, NJ, USA). To measure apoptosis, these cells were further were stained and analyzed using the Annexin V-allophycocyanin (APC)/7-am inoactinomycin D (7AAD) apoptosis kit (KeyGen Biotech Co., Ltd., Nanjing, China), according to the manufacturer\\'s instructions. Briefly, 2\u00d710^6^ cells were washed with PBS, stained with 7-AAD and Annexin V-APC for 15 min at room temperature and analyzed using an LSR FORTESSA. The results were analyzed using ModFit LT2.0 software (Verity Software House, Topsham, ME, USA).\n\nEPB41L3 plasmid construction and cell transfection\n--------------------------------------------------\n\nThe EPB41L3 expression vector pReceiver-M98-EPB41L3 was constructed by inserting the 417nt ORF sequence of EPB41L3 (MFIQIFPVIFLETSIAYSNVVWVYISYLHLLMKMFMR DHFGCLMNYLPCFSTETFSLTPTVLAVGWLLERREVS FSCSEWDKVASVVEQCLKYFFLSLFCSPFLSHLEHGIW TCQSSGNTLPLLVGTWIMWVSPEICV) into the pReceiver-M98 vector (Genecopoeia Inc., Rockville, MD, USA). One pair of specific oligonucleotides (EPB41L3 gene isoform 1) was annealed and then subcloned into the vector. Cell transfection was performed with Lipofectamine 2000 and KYSE-150 and KYSE-510 were selected with 500 *\u00b5*g/ml G418 sulfate (both from Thermo Fisher Scientific, Inc.) for 2--3 weeks, according to the manufacturer\\'s protocol for stable transfection. The KYSE-150 and KYSE-510 cells transfected with the EPB41L3 overexpression plasmid were termed vec-150-epb and vec-510-epb respectively. The KYSE-150 and KYSE-510 cells transfected with the empty control plasmid were termed vec-150-nc and vec-510-nc respectively.\n\nSmall interefering (si) RNA vector construction and cell transfection\n---------------------------------------------------------------------\n\nThree pairs of siRNA and a negative control siRNA were purchased from GenePharma Co., Ltd. (Shanghai, China) and subcloned into pcDNA6.2 GW/EmGFP vectors (Thermo Fisher Scientific, Inc.). KYSE-510 cells were transfected with the vectors carrying siRNA-EPB41L3 or the negative control siRNA. There were 3 siRNA-epb41l3 vectors were transfected separately. Silencing efficiency was measured at 48 h post-transfection by western blotting. The siRNA with the highest knockdown efficiency was selected for subsequent experiments (sense, GCUGCGAAUAAACAGAUUUTT and antisense, AAAUCUGUUUAUUCGCAGCTT). The negative control siRNA (sense, CUGCGGAAAUAAUUUCAGATT and antisense, AAUCAUUUUGACGUUAGCCTT). Cells with stable knockdown of EPB41L3 expression were established under selection with Blasticidin S HCl (Thermo Fisher Scientific, Inc.). KYSE-519 cells transfected with siRNA-EPB41L3 or the negative control siRNA were termed si-510-epb and si-510-nc respectively.\n\nClonogenic assays\n-----------------\n\nThe effects of EPB41L3 expression were analyzed by clonogenic assays using standard protocols. Eighteen h prior to transfection, cells were seeded in 6-well plates at a density of 2\u00d710^5^ per well in complete RPMI-1640 or DMEM (Thermo Fisher Scientific, Inc.). Cells were transfected with plasmid DNA using the Lipofectamine 2000 transfection reagent. Twenty-four h after transfection, cells were trypsinized and plated in 100 mm dishes at identical densities for each transfection. Transfected cells were selected in 500 *\u00b5*g/ml G418 for 10 to 14 days, and surviving colonies were stained with crystal violet and counted ([@b22-ijo-52-05-1443]). Transgene expression was confirmed for each experiment. Each experiment was conducted at least three times.\n\nIn vivo experiments\n-------------------\n\nTo assess the role of EPB41L3 in inhibiting tumor growth *in vivo*, suspensions of 1\u00d710^6^ EPB41L3-overexpressing KYSE-150 cells in 0.1 ml of PBS were injected subcutaneously into the backs of 4--5-week-old male immunodeficient SCID-Beige mice weighing 18--20 g (obtained from the Laboratory Animal Center, Southern Medical University, Guangzhou, China). As a control, suspensions of 1\u00d710^6^ KYSE-150 cells transfected with control vector were injected into the same mice on a different site. To assess the role of EPB41L3 in promoting tumor growth *in vivo*, suspensions of 1\u00d710^6^ EPB41L3-knockdown or control KYSE-510 cells in 0.1 ml PBS were injected subcutaneously into two sites of the backs of 4--5-week-old male immunodeficient SCID-Beige mice. The slightly upper position of the waist was selected as the optimal injection site. Four mice were used in each experimental group. Tumor growth was assessed by measuring the xenografts in two dimensions once a week. Tumor volumes were calculated according to the following formula: (volume) = 1/2 \u00d7 (long axis) \u00d7 (short axis)^2^. At 28 days post-injection, mice were sacrificed and the tumors were carefully dissected. All animal experiments were performed in accordance with the principles and procedures outlined in the Southern Medical University Guide for the Care and Use of Animals under the assurance number SCXK (Guangdong) 2008--0002. Approval was obtained from the Nanfang Hospital Animal Ethics Committee.\n\nStatistical analysis\n--------------------\n\nStatistical analysis was performed using the SPSS version 13.0 software (SPSS, Inc., Chicago, IL, USA). Each experiment was performed in triplicate. The results were expressed as the mean \u00b1 standard deviation. Significant differences were determined using Student\\'s t-test to compare two groups. The \u03c7^2^ test was used to analyze the correlations between EPB41L3 expression and clinicopathological features in patients with ESCC. Survival curves were evaluated using the Kaplan-Meier method, and differences between survival curves were assessed by the log-rank test. The Cox proportional hazards regression model was used to examine univariate and multivariate hazard ratios for the study variables that were dichotomized. Only significantly different variables in univariate analysis were entered into the subsequent multivariate analysis. P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\nEPB41L3 expression is downregulated in ESCC tissues\n---------------------------------------------------\n\nTo explore the expression status of EPB41L3 in ESCC tissues, immunohistochemistry analysis was performed using a tissue microarray which contained 97 pairs of ESCC samples and their adjacent non-tumor tissue samples. In general, the results demonstrated that EPB41L3 expression was significantly lower in tumor tissues compared with their matched adjacent non-tumor tissues ([Table III](#tIII-ijo-52-05-1443){ref-type=\"table\"}; [Fig. 1A](#f1-ijo-52-05-1443){ref-type=\"fig\"}). High immune reactivity to EPB41L3 was detected along cell membranes and cytoplasm in 91 out of 97 non-tumor tissue samples, while only 25 out of 97 ESCC tissue samples revealed high expression. Only 6 out of 97 adjacent non-tumor tissue samples exhibited low EPB41L3 expression or no expression, while 72 out of 97 ESCC tissue samples exhibited low expression or no expression ([Table III](#tIII-ijo-52-05-1443){ref-type=\"table\"}; [Fig. 1A](#f1-ijo-52-05-1443){ref-type=\"fig\"}).\n\nEPB41L3 expression is lower in ESCC cells compared with Het-1a cells due to methylation\n---------------------------------------------------------------------------------------\n\nRT-PCR analysis was performed to dete2ct EPB41L3 mRNA expression in six ESCC cell lines and in one non-neoplastic esophagus cell line, Het-1a. The results revealed that the non-neoplastic Het-1a cell line had markedly higher expression of EPB41L3 compared with all six esophagus cancer cell lines ([Fig. 1B](#f1-ijo-52-05-1443){ref-type=\"fig\"}). Western blotting was then used to detect EPB41L3 expression at the protein level. Similar to the RT-PCR results, higher EPB41L3 protein expression was observed in the normal esophagus cell line compared with the ESCC cell lines ([Fig. 1C](#f1-ijo-52-05-1443){ref-type=\"fig\"}).\n\nIn order to explore the reason why EPB41L3 was down-regulated in ESCC cells, methylation patterns were examined using MSP in the six esophagus cancer cell lines and the non-neoplastic esophagus Het-1a cell line. All six esophagus cancer cell lines developed both methylated and unmethylated bands, suggesting that EPB41L3 had partial methylation in the cancer cell lines. By contrast, Het-1a developed only an unmethylated band, demonstrating that EPB41L3 was unmethylated in the normal esophageal cells ([Fig. 2A](#f2-ijo-52-05-1443){ref-type=\"fig\"}).\n\nNext, the different promoter methylation levels were explored by pyrosequencing. A quantitative analysis of DNA methylation at 33 CpG sites of the EPB41L3 gene promoter was performed. The pyrosequencing results revealed high levels of EPB41L3 methylation in five cell lines, KYSE-150, KYSE-510, Eca-109, CaEs17 and TE-1, which were represented by dark circles in [Fig. 2B](#f2-ijo-52-05-1443){ref-type=\"fig\"}. For Het-1a, the methylation rate was \\<20%, which was considered to be unmethylated (represented by white circles in [Fig. 2B](#f2-ijo-52-05-1443){ref-type=\"fig\"}). These results were in accordance with the MSP results.\n\nTo identify whether the downregulated EPB41L3 mRNA and protein expression levels in ESCC were due to methylation, 5-aza-2-deoxycytidine, a potent inhibitor of DNA methyltransferase, was used to treat the TE-1, Eca-109, KYSE-150 and KYSE-510 cell lines. If the downregulation of EPB41L3 was due to methylation, treatment with 5-aza-2-deoxycytidine would be expected to increase EPB41L3 expression. Indeed, treating the cells with the inhibitor restored EPB41L3 mRNA expression, as identified by RT-PCR ([Fig. 2C](#f2-ijo-52-05-1443){ref-type=\"fig\"}) and qPCR ([Fig. 2D](#f2-ijo-52-05-1443){ref-type=\"fig\"}). Furthermore, western blotting confirmed restoration of EPB41L3 protein expression in KYSE-150 and KYSE-510 ([Fig. 2E](#f2-ijo-52-05-1443){ref-type=\"fig\"}). These results indicated that the decreased EPB41L3 expression in esophageal cancer cell lines occurred via promoter methylation.\n\nEPB41L3 inhibits ESCC cell proliferation and reduces colony formation in vitro\n------------------------------------------------------------------------------\n\nIn order to examine the functional role of EPB41L4 in ECC, its expression was silenced in ESCC cells by siRNA. First, the knockdown efficiency of the siRNAs was confirmed by western blotting. The results demonstrated that siRNA2 resulted in the highest knockdown efficiency ([Fig. 3A](#f3-ijo-52-05-1443){ref-type=\"fig\"}), and therefore siRNA2 as selected for subsequent experiments.\n\nTo investigate the effects of overexpressing EPB41L3 on human ESCC cell growth *in vitro*, the KYSE-150 and KYSE-510 cell lines were used to perform CCK-8 assays. EPB41L3 overexpression in KYSE-150 and KYSE-510 cells resulted in significant inhibition of cell proliferation compared with control cells (P\\<0.05; [Fig. 3B](#f3-ijo-52-05-1443){ref-type=\"fig\"}). By contrast, the transfection of EPB41L3 siRNA in KYSE-510 cells resulted in increased cell proliferation compared with control cells (P\\<0.05; [Fig. 3B](#f3-ijo-52-05-1443){ref-type=\"fig\"}). The ability of these two cell lines to form colonies in soft agar was then examined for two weeks. Overexpression of EPB41L3 significantly reduced the cell colony numbers in KYSE-150 and KYSE-510 cells, while EPB41L3 knockdown by siRNA in KYSE-510 cells had the opposite effect ([Fig. 3C](#f3-ijo-52-05-1443){ref-type=\"fig\"}). Collectively, these results indicated that EPB41L3 inhibited the growth of ESCC cells *in vitro*.\n\nEpb41l3 inhibits tumor formation in vivo\n----------------------------------------\n\nKYSE-510 cells transfected with EPB41L3 overexpression plasmid or control plasmid were termed vec-510-epb and vec-510-nc, respectively. KYSE-510 cells transfected with siRNA-EPB41L3 or control siRNA were termed si-510-epb or si-510-nc, respectively. To further examine the role of EPB41L3 in tumor growth *in vivo*, tumor xenografts were generated in mice by injecting vec-510-epb, vec-510-nc, si-510-epb and si-510-nc cells (four nude mice per group). Visible tumors began to form at 10 days post-injection. The volume of the tumors was measured twice a week and the mice were sacrificed at day 28 post-injection. At the end of the experiment, the mice weighted 168--180 g. The maximum diameter of a single tumor was 1.5 cm. None of the mice developed multiple tumors. The group injected with vec-510-epb cells formed significantly smaller tumors compared with the group injected with the control vec-510-nc cells ([Fig. 3D](#f3-ijo-52-05-1443){ref-type=\"fig\"}). The group injected with si-510-EPB cells formed significantly larger tumors compared with the group injected with the control si-510-NC cells ([Fig. 3D](#f3-ijo-52-05-1443){ref-type=\"fig\"}). Overall, these results data demonstrated that increased expression of Epb41L3 inhibited tumor growth, while decreased expression of Epb41L3 promoted tumor growth *in vivo*.\n\nEctopic expression of EPB41L3 promotes apoptosis by activating the caspase-3/8/9 pathway\n----------------------------------------------------------------------------------------\n\nApoptosis was detected using the 7-AAD/Annexin V-APC double staining method followed by flow cytometry analysis. Compared with the control group, KYSE-150 and KYSE-510 cells overexpressing EPB41L3 had significantly higher apoptotic rates compared with cells transfected with control plasmid (20.3 vs. 2.9%, and 19.9 vs. 5.2%, respectively) ([Fig. 4A](#f4-ijo-52-05-1443){ref-type=\"fig\"}). To further investigate the mechanisms underlying the higher apoptotic rates following EPB41L3 overexpression, western blot analysis was performed to determine the expression levels of apoptosis-related proteins. The results demonstrated that the expression levels of caspase-8, caspase-9 and caspase-3 were elevated in KYSE-150 and KYSE-510 cells following EPB41L3 over-expression compared with control cells ([Fig. 4B](#f4-ijo-52-05-1443){ref-type=\"fig\"}). Then, the expression levels of caspase-8, caspase-9 and caspase-3 were detected in KYSE-510 cells following EPB41L3 knockdown by siRNA. The results demonstrated that expression of caspase-8, caspase-9 and caspase-3 was markedly decreased following EPB41L3 knockdown compared with control cells ([Fig. 4B](#f4-ijo-52-05-1443){ref-type=\"fig\"}).\n\nEPB41L3 causes G2/M cell cycle arrest by activating the CDK1 pathway\n--------------------------------------------------------------------\n\nFlow cytometry experiments revealed that the % of cells in the G2 phase of the cell cycle was significantly increased in the EPB41L3-overexpressing group compared with the control group, in both the KYSE-150 ([Fig. 5A](#f5-ijo-52-05-1443){ref-type=\"fig\"}) and KYSE-510 ([Fig. 5B](#f5-ijo-52-05-1443){ref-type=\"fig\"}) cell lines. The % of cells in G2 was 36.43 and 6.77% for the EPB41L3-overexpressing and control KYSE150 cells respectively ([Fig. 5A](#f5-ijo-52-05-1443){ref-type=\"fig\"}), and 39.35 and 8.93% for the EPB41L3-overexpressing and control KYSE-510 cells respectively ([Fig. 5B](#f5-ijo-52-05-1443){ref-type=\"fig\"}). Western blotting was also performed in order to explore the mechanism of G2/M cell cycle arrest. The results demonstrated that overexpression of EPB41L3 decreased the protein expression levels of CDK1 and CyclinB1 in vec-150-epb and vec-510-epb cells, compared with their respective controls ([Fig. 5C](#f5-ijo-52-05-1443){ref-type=\"fig\"}). In addition, the expression levels of these proteins were examined following EPB41L3 knockdown in KYSE-510 cells. The expression levels of CDK1 and CyclinB1 were significantly increased in the knockdown cells compared with control cells ([Fig. 5C](#f5-ijo-52-05-1443){ref-type=\"fig\"}). Together with the flow cytometry results, these data suggested that inhibition of CDK1/CyclinB1 might be the main mechanism of the G2/M arrest induced by EPB41L3.\n\nEPB41L3 is negatively correlated with depth of tumor infiltration and may be a prognostic indicator\n---------------------------------------------------------------------------------------------------\n\nThe relationship between clinicopathological characteristics and EPB41L3 expression levels in patients with ESCC is summarized in [Table I](#tI-ijo-52-05-1443){ref-type=\"table\"}. The clinicopathological characteristics contained age, sex, pathology classification, numbers of lymph nodes, tumor infiltration and tumor-node-metastasis (TNM) stage (based on the 7th edition of American Joint Committee on Cancer clinical staging ([@b23-ijo-52-05-1443]). No significant association was observed between EPB41L3 expression levels and the patients\\' age, sex, pathologic type or lymph status in the 97 ESCC cases. However, the expression levels of EPB41L3 were significantly negatively correlated with depth of tumor infiltration (P=0.016; [Table I](#tI-ijo-52-05-1443){ref-type=\"table\"}). Next, univariate analysis was used to explore the association between clinicopathological characteristics and the overall survival. The results demonstrated that tumor infiltration depth (P=0.01), TNM stage (P\\<0.001) and EPB41L3 expression (P=0.004) were all positively correlated with overall survival ([Table IV](#tIV-ijo-52-05-1443){ref-type=\"table\"}). These three factors were further investigated using multivariate analysis, and the results revealed that only TNM stage (P=0.004) and EPB41L3 expression (P=0.033) were statistically significant ([Table IV](#tIV-ijo-52-05-1443){ref-type=\"table\"}). To investigate the prognostic value of EPB41L3 expression in ESCC, the association between the expression levels of EPB41L3 and the patients\\' survival was assessed using Kaplan-Meier analysis with the log-rank test. The results demonstrated that patients with lower expression of EPB41l3 had shorter survival time compared with the patients that had higher EPB41l3 expression (P=0.002; [Fig. 6](#f6-ijo-52-05-1443){ref-type=\"fig\"}). These findings revealed that EPB41L3 may have the potential to serve as a prognosis predictor in ESCC.\n\nDiscussion\n==========\n\nEsophageal carcinoma is a common malignancy with poor prognosis. The main reason for low survival is that the majority of ESCC are asymptomatic and undetected until they have spread beyond the esophageal wall and thus become unresectable ([@b24-ijo-52-05-1443]). EPB41L3, also known as 4.1B or DAL1, is located on chromosome 18p11.32, and it is expressed in multiple adult tissues, including the brain, lung, kidney, intestine and prostate. It has been demonstrated to serve as an antitumor factor in some cancers, primarily in non-small-cell lung carcinoma (NSCLC). Kikuchi *et al* ([@b30-ijo-52-05-1443]) noted that EPB41L3 methylation was involved in the development and progression of NSCLC and served as a poor prognosis predictor. Nevertheless, previous studies have not focused on the mechanisms underlying the tumor suppressive function and the clinical significance of EPB41L3 in ESCC. Our group has been studying the role of EPB41L3 in ESCC with the aim to explore its methylation status and to identify potential plasma biomarkers for early diagnosis. By using an Infinium Methylation 450k array, the methylation frequency of EPB41L3 was demonstrated to be higher in tumor tissues compared with normal surrounding tissues, and this was correlated with large tumor size and advanced pT tumor stage ([@b12-ijo-52-05-1443]). EPB41L3 has been demonstrated to be partly silenced due to hypermethylation ([@b13-ijo-52-05-1443]). Other genetic and epigenetic mechanisms may result in the regulation of EPB41L3 expression, including histone modifications and microRNAs, and these will need further investigation in the future.\n\nThere has been evidence suggesting that hypermethylation has a critical role in the inactivation of EPB41L3. Demethylation treatments were able to restore EPB41L3 expression in several types of cancer ([@b9-ijo-52-05-1443],[@b25-ijo-52-05-1443]--[@b27-ijo-52-05-1443]). Aberrant DNA methylation is one of the best-characterized epigenetic modifications, contributing to tumor initiation and progression ([@b28-ijo-52-05-1443],[@b29-ijo-52-05-1443]). This is because DNA methylation, which usually occurs on the cytosine residues of CpG dinucleotides, is key to tissue differentiation during early embryonic growth ([@b30-ijo-52-05-1443]). The present data provided evidence that EPB41L3 expression was downregulated in ESCC cells due to promoter methylation. First, EPB41L3 expression was demonstrated to be lower in ESCC tissues and cell lines compared with adjacent normal tissues and a non-neoplastic cell line. Second, MSP and pyrosequencing confirmed higher promoter methylation rate in ESCC cell lines compared with the normal esophageal cell line. Notably, although methylation was evident in all six cell lines, they all exhibited partial methylation rather than full methylation of the promoter. The methylation rates were different in different ESCC cell lines, demonstrating heterogeneity in these cancer cells. Pyrosequencing has been recently regarded as a better method for testing methylation in large-scale validation studies, biomarker development and clinical diagnostics ([@b31-ijo-52-05-1443]), while MSP has some disadvantages, including its inability of detecting mosaic DNA methylation ([@b32-ijo-52-05-1443],[@b33-ijo-52-05-1443]).\n\nDNA methylation provides an alternative way for tumor suppressor gene function loss ([@b12-ijo-52-05-1443]). Having detected EPB41L3 methylation in ESCC cells, the present study aimed to explore the mechanisms underlying the role of EPB41L3 as a tumor suppressor. Thus, experiments related to proliferation, apoptosis and cell cycle phase distribution were performed. CCK-8 cell viability and clonogenic assay results indicated that EPB41L3 overexpression inhibited tumor cell proliferation *in vitro*. Tumor xenograft experiments revealed that EPB41L3 over-expression suppressed tumor growth *in vivo*. Flow cytometry assay demonstrated higher rates of apoptosis in EPB41L3-overexpressing cells compared with control cells. These results offered a foundation for mechanistic exploration. Previous study has revealed that apoptosis in MCF-7 cells was primarily associated with the activation of Caspase-8, and inhibition of this activation blocked the ability of EPB41L3 to induce cell death in MCF-7 ([@b10-ijo-52-05-1443]). Cleaved-caspase-8 and cleaved-caspase-9 activate pro-caspase 3 into cleaved-caspase-3, resulting in apoptosis. Thus, the expression of the caspase family was detected by western blotting in order to explore the mechanism of apoptosis induction in ESCC. The results demonstrated that expression of caspase-8, caspase-9 and caspase-3 were elevated in the EPB41L3-overexpressing cells. By contrast, EPB41L3 knockdown by siRNA resulted in the opposite effects.\n\nKuns *et al* ([@b31-ijo-52-05-1443]) reported that, during pregnancy, the overexpression of EPB41L3 results in reduced CyclinA expression and Rb phosphorylation, which is accompanied by decreased tyrosine kinase receptor ErbB2 phosphorylation, causing the mammary epithelial cells to be arrested in the G1 phase ([@b34-ijo-52-05-1443]). The present data revealed no significant change in CyclinA expression, while expression of CyclinB1 and CDK1 decreased following EPB41L3 overexpression. High expression of CyclinB1-CDK1 only occurs in the M phase, not the G2 phase ([@b35-ijo-52-05-1443]). CDK1 controls entry into and exit from the M phase of the cell cycle and is commonly accompanied by CyclinA and CyclinB1 ([@b36-ijo-52-05-1443]). This is in accordance with the present cell cycle data. Cell death can occur during mitochondrial membrane permeabilization with the release of cell death effectors such as apoptosis-inducing factor ([@b37-ijo-52-05-1443]). The G2/M phase arrest may limit the proliferation of chromosomally unstable, pre-cancerous cells. Downregulation of the G2/M checkpoint may lead to tumorigenesis ([@b38-ijo-52-05-1443]). The present findings indicate that ectopic expression of EPB41L3 may inhibit cell proliferation by upregulating apoptosis and causing G2/M phase arrest.\n\nTo date, several genes have been reported to be hypermethylated in ESCC, including cyclin-dependent kinase inhibitor 2A (CDKN2A)/p16INK4a ([@b39-ijo-52-05-1443]--[@b41-ijo-52-05-1443]). These methylated genes have the potential to be used as diagnostic or prognostic molecular markers for ESCC. Recent study has focused on the correlation of epigenetic changes and ESCC patient survival. Zare *et al* ([@b42-ijo-52-05-1443]) and Lee *et al* ([@b43-ijo-52-05-1443]) separately reported that methylation of adenomatous polyposis coli (APC) or fragile histidine triad protein (FHIT) was closely correlated to poor outcome in patients with ESCC. The current study examined EPB41L3 expression in ESCC and its correlation with clinicopathological features and prognosis. A statistically significant difference was observed between EPB41L3 expression in ESCC tumor tissues and their matched adjacent non-tumor tissues. This was consistent in both RT-PCR and western blotting results. Survival analysis revealed that patients with higher expression of EPB41L3 had improved prognosis. The relationship between EPB41L3 expression levels and certain clinical features were then analyzed. Univariate analysis demonstrated that EPB41L3 expression was significantly correlated with tumor infiltration depth and TNM stage. These results indicate that EPB41L3 has a significant correlation with tumor progression. Through a tissue microarray, the current study revealed the clinical importance of EPB41L3 in ESCC patients, which may have the potential to be prognostic indicator. To date, only 20 types of tumor markers have been put into clinical use. No specific tumor markers exist to date for ESCC ([@b44-ijo-52-05-1443]). The main obstacle to developing clinically useful indicators is the lack of large clinical trials in ESCC, which are hampered by a lack of sizeable esophageal tissue repositories that include complete clinical annotation ([@b45-ijo-52-05-1443]). Furthermore, because chemotherapy and radiation remain important in ESCC treatment, predictive indicators may be critical for identifying patients who may benefit from molecular targeted agents or are more sensitive to radiotherapy ([@b46-ijo-52-05-1443]). Further studies are warranted to probe the clinical significance of EPB41L3 in ESCC.\n\nIn summary, low levels of EPB41L3 may be an unfavorable indicator of ESCC prognosis. Overexpressing EPB41L3 inhibited proliferation, and induced apoptosis and G2/M arrest in ESCC cells through the Caspase-3/8/9 and CDK1/CyclinB1 pathways. Detection of the tumor EPB41L3 methylation status may aid in predicting prognosis in patients with ESCC.\n\nNot applicable.\n\n**Funding**\n\nThis study was supported by the Natural Science Foundation of Guangdong Province, China (grant no. 2015A030313273).\n\n**Availability of data and materials**\n\nThe analyzed datasets generated during the study are available from the corresponding author on reasonable request.\n\n**Author\\'s contributions**\n\nRZ performed the cell culture, MSP, western blotting, transfection, CCK-8, clonogenic assays and was involved in the writing of the article. YL performed qPCR analysis. ZJJ performed the flow cytometry experiment. JPH performed the statistical analysis; YW performed the IHC analysis; XFL, WBX and XCW were involved in the animal experiments; JRZ and QEW performed pyrosequencing; YFZ contributed to the whole design of this research project. JRZ, QEW and YFZ all contributed to the writing of the manuscript. All authors have read and approved the final manuscript.\n\n**Ethics approval and consent to participate**\n\nAll animal experiments were performed in accordance with the principles and procedures outlined by the Southern Medical University Guide for the Care and Use of Animals under the assurance number SCXK (Guangdong) 2008--0002. Approval was obtained from the Nanfang Hospital Animal Ethics Committee.\n\n**Consent for publication**\n\nNot applicable.\n\n**Competing interests**\n\nThe authors declare that they have no competing interests.\n\n![EPB41L3 expression is downregulated in ESCC tissue samples and cell lines compared with their para-carcinoma tissues and a non-neoplastic esophageal epithelium cell line, respectively. (A) Representative images of EPB41L3 immunohistochemistry staining in ESCC tissues compared with CNE. Scale bar, \u00d7200. (B) Reverse transcription-polymerase chain reaction analyses of EPB41L3 mRNA expression in the ESCC cell lines TE-1, Eca-109, CaEs17, KYSE-150, KYSE-450 and KYSE-510 and one non-neoplastic esophageal epithelium cell line Het-1a. (C) Western blot analyses of EPB41L3 protein expression confirmed that EPB41L3 expression was lower in the ESCC cell lines compared with the non-neoplastic cell line. EPB41L3, erythrocyte membrane protein band 4.1 like 3; ESCC, esophageal squamous cell carcinoma; CNE, corresponding noncancerous esophagus.](IJO-52-05-1443-g00){#f1-ijo-52-05-1443}\n\n![Reduced expression of EPB41L3 in ESCC is due to methylation. (A) Methylation-specific polymerase chain reaction analysis of the seven cell lines revealed that EPB41L3 was partially methylated in all six ESCC cell lines while it was unmethylated in the normal Het-1a cells. (B) Representation schematic of the pyrosequencing results for the methylation status of the CpG promoter in the seven cell lines tested. Gray boxes indicate exons. Vertical bars indicate CpG sites examined for methylation. Black, gray, and white circles represent hypermethylation, partial methylation and nonmethylation, respectively. The schematic illustrates the representative results of pyrosequencing of a cytosine residue(s) at \u2212141551, \u2212141494, \u2212141439 and \u2212141379 bp of the EPB41L3 promoter. (C) Treatment with 5-Aza-2-deoxygcitidine restored the expression of EPB41L3 in ESCC lines, suggesting that downregulation of EPB41L3 in ESCC is due to methylation. (D) Reverse transcription-quantitative polymerase chain reaction analyses revealed that EPB41L3 expression was increased in ESCC cell lines following 5-Aza-2-deoxygcitidine treatment. The increase was most evident in the KYSE-510 cells. In the other cell lines, although an increase was observed, this was not as evident as in the KYSE-510 cells. (E) Western blot analyses revealed that EPB41L3 expression was increased in KYSE-150 and KYSE-510 cells following 5-Aza-2-deoxygcitidine treatment. EPB41L3, erythrocyte membrane protein band 4.1 like 3; ESCC, esophageal squamous cell carcinoma; M, methylated primer; U, unmethylated primer; 5aza, 5-Aza-2-deoxygcitidine.](IJO-52-05-1443-g01){#f2-ijo-52-05-1443}\n\n![EPB41L3 inhibits cell proliferation in ESCC *in vitro* and *in vivo*. (A) Western blot analysis of the siRNA transfection efficiency of 3 siRNAs. siRNA-2 resulted in the largest decrease in protein levels and was therefore selected for subsequent experiments. (B) Growth curves of cells transfected with either EPB41L3-ovexpression plasmid (vec-EBP), or siRNA (si-EBP) or respective controls by cell counting kit-8 assay. (C) Representative images and quantification of colony formation assays. (D) Photographic images of the xenograft tumors and quantification of tumor volumes in each experimental group. ^\\*^P\\<0.01 compared with control. EPB41L3, erythrocyte membrane protein band 4.1 like 3; ESCC, esophageal squamous cell carcinoma; si, small interfering; nc, negative control.](IJO-52-05-1443-g02){#f3-ijo-52-05-1443}\n\n![EPB41L3 has proapoptotic effects by activating the caspase-3/8/9 pathway. (A) Representative plots from the flow cytometry analysis following 7-AAD and Annexin V-APC double staining. (B) Western blot analysis for the expression levels of cleaved-caspase-3, cleaved-caspase-8 and cleaved-caspase-9 following EPB41L3 or knockdown in ESCC cells. GAPDH was used as a loading control. EPB41L3, erythrocyte membrane protein band 4.1 like 3; 7-AAD, 7-aminoactinomycin D; APC, allophycocyanin; ESCC, esophageal squamous cell carcinoma; si, small interfering; nc, negative control.](IJO-52-05-1443-g03){#f4-ijo-52-05-1443}\n\n![EPB41L3 induces G2/M arrest by activating the CyclinB/CDK1 pathway. (A and B) Flow cytometry analysis of cell cycle phase distribution revealed a G2/M arrest following EPB41L3 overexpression. (C) Western blot analysis of cell cycle-related proteins following EPB41L3 or knockdown in ESCC cells. ^\\*^P\\<0.05 compared with control. EPB41L3, erythrocyte membrane protein band 4.1 like 3; CDK1, Cyclin-dependent kinase 1; ESCC, esophageal squamous cell carcinoma; si, small interfering; nc, negative control.](IJO-52-05-1443-g04){#f5-ijo-52-05-1443}\n\n![EPB41L3 expression predicts overall survival in patients with ESCC. Kaplan-Meier analysis with the log-rank test was used to assess the association between levels of tumor EPB41L3 expression and patients\\' survival time. EPB41L3, erythrocyte membrane protein band 4.1 like 3; ESCC, esophageal squamous cell carcinoma.](IJO-52-05-1443-g05){#f6-ijo-52-05-1443}\n\n###### \n\nCorrelation between clinicopathological characteristics and EPB41L3 protein expression in esophageal squamous cell carcinoma.\n\n  Characteristic             Low   High   Total   P-value\n  -------------------------- ----- ------ ------- ---------\n  Sex                                             \n  \u2003Male                      49    21     70      0.125\n  \u2003Female                    23    4      27      \n  Age (years)                                     \n  \u2003\\<65                      34    12     46      0.947\n  \u2003\\>65                      38    13     51      \n  Pathology                                       \n  \u2003G1-G2                     55    22     77      0.216\n  \u2003G3                        17    3      20      \n  Lymph node metastasis                           \n  \u2003Negative                  39    11     50      0.381\n  \u2003Positive                  33    14     47      \n  Tumor infiltration depth                        \n  \u2003T1-T2                     12    10     22      0.016\n  \u2003T3                        60    15     75      \n  TNM stage                                       \n  \u2003C1-C2                     38    17     55      0.186\n  \u2003C3                        34    8      42      \n\nEPB41L3, erythrocyte membrane protein band 4.1 like 3; TNM, tumor-node-metastasis.\n\n###### \n\nSequences of primers used for pyrosequencing.\n\n  Primer      Sequence                    5\u2032 labeling\n  ----------- --------------------------- -------------\n  122-01-1F   AGTTTAGGGAGGAGGTTTGTAAG     \n  122-01-1R   ACACCTCCAAATTACCACCTACAA    Biotin\n  122-01-1S   GTTTGTAAGGAGATTTATATTTTG    \n  122-01-2F   GGAGGAGGTTTGTAAGGAGAT       \n  122-01-2R   CACCTCCAAATTACCACCTACA      Biotin\n  122-01-2S   TTTAGGGAGGGGTAGA            \n  122-01-3F   GAGGAGGTTTGTAAGGAGATTTAT    \n  122-01-3R   ACCTCCAAATTACCACCTAC        Biotin\n  122-01-3S   GGGTTAGAGAGGGTTGA           \n  122-01-4F   GGGAGGAGGTTTGTAAGGAGATTT    \n  122-01-4R   ATCCCTCCCCAAAAACTCTTTCCTT   Biotin\n  122-01-4S   GAAGAGAGTTGTAGGTGGTAAT      \n\n###### \n\nEPB41L3 protein expression in NE and ESCC samples.\n\n  Group   Cases   EPB41L3 protein expression   P-value   \n  ------- ------- ---------------------------- --------- ---------\n  NE      97      6                            91        \\<0.001\n  ESCC    97      72                           25        \n\nEPB41L3, erythrocyte membrane protein band 4.1 like 3; ESCC, esophageal squamous cell carcinoma; NE, non-cancerous esophagus.\n\n###### \n\nSummary of univariate and multivariate Cox regression analyses of overall survival.\n\n  Parameter                                     Univariate analysis   Multivariate analysis                                  \n  --------------------------------------------- --------------------- ----------------------- -------------- ------- ------- --------------\n  Sex                                           0.26                  1.368                   0.793--2.36                    \n  Age (years)                                   0.649                 1.114                   0.701--1.769                   \n  Pathology                                     0.91                  0.968                   0.548--1.71                    \n  Lymph nodes involved                          0.071                 0.651                   0.409--1.037                   \n  Tumor infiltration depth                      0.01                  0.425                   0.222--0.812   0.301   0.689   0.34--1.395\n  TNM stage                                     \\<0.001               0.398                   0.247--0.643   0.004   0.478   0.289--0.79\n  EPB41L3 expression in cancer tissue           0.004                 2.392                   1.329--4.304   0.033   1.942   1.054--3.577\n  EPB41L3 expression in para-cancerous tissue   0.327                 1.521                   0.657--3.524                   \n\nHR, hazard ratio; CI, confidence interval; EPB41L3, erythrocyte membrane protein band 4.1 like 3.\n\n[^1]: Contributed equally\n"}
+{"text": "1. Introduction {#sec1-sensors-18-04120}\n===============\n\nThe use of flexible or soft structures on the development of actuators, robots and devices is an emerging trend in the last few years \\[[@B1-sensors-18-04120]\\]. The so-called soft robotics involve the use of flexible fluidic actuators, shape memory materials, and electro active polymers as actuators in conjunction with rubbers, plastics and flexible cables, which result in a flexible robot \\[[@B2-sensors-18-04120]\\]. These advantages are especially desirable in the development of wearable robots, where the robot can be optimally designed and controlled for each user, achieving the so-called human-in-the-loop design \\[[@B3-sensors-18-04120]\\]. Soft robotics also find numerous applications in the biomedical field, which include wearable robots, prostheses, surgical and assistive devices due to the biocompatibility and biomimicry of such soft materials that are also employed as artificial organs as well as in body simulations, as summarized in \\[[@B4-sensors-18-04120]\\]. For the industrial environment, soft robots can be used to perform activities that the conventional rigid robots cannot, such as grasping fragile objects. In addition, due to their higher flexibility, such robots can be regarded as a safer solution in human--robot interaction factories or workplaces \\[[@B5-sensors-18-04120]\\].\n\n3D printing technology is an additive layer manufacturing (ALM) process, generally made by fused deposition modeling (FDM) in which hot or melted polymers are injected layer-upon-layer to form the desired structure \\[[@B6-sensors-18-04120]\\]. This technology has enabled the development of numerous flexible structures in soft robotics (summarized in \\[[@B7-sensors-18-04120]\\]) due to its advantages of relative low cost, flexibility in design, high repeatability and no need for performing additional post-fabrication process (such as sanding, milling and scraping) \\[[@B8-sensors-18-04120]\\]. Thus, many soft robotics applications rely on 3D-printed structures, which can be created using different materials such as acrylonitrile butadiene styrene (ABS), polycarbonate (PC) and polylatic acid (PLA) as well as materials with higher flexibility than the former, such as thermoplastic polyurethane (TPU).\n\nFor accurate control of any robotic device, the sensor's system plays a crucial role on the accuracy of control, which enable the robots to perform highly precise and demanding tasks such as in robotic assisted surgeries and human--robot interaction \\[[@B5-sensors-18-04120]\\]. In these cases, the conventional electronic sensors can suffer from lack of robustness, difficult installation and the flexibility requirements can inhibit the application of several electronic sensors \\[[@B9-sensors-18-04120]\\]. In addition, the electromagnetic field sensitivity of electronic sensors can harm their applications on the actuators instrumentation, which generally employs electric actuators being activated constantly \\[[@B10-sensors-18-04120]\\]. Furthermore, for invasive biomedical applications, e.g. drug delivery and surgical devices, the biocompatibility constraints of the whole system, including the sensors, can limit the application of most commercially available sensors \\[[@B11-sensors-18-04120]\\].\n\nOptical fiber sensors represent a continuously growing research field in both photonics and sensors communities \\[[@B12-sensors-18-04120]\\]. Advantages such as compactness, electromagnetic fields immunity, passive operation, multiplexing capabilities, chemical stability and biocompatibility \\[[@B13-sensors-18-04120]\\] have led to the widespread of this sensing technology. Many applications of optical fiber sensors are proposed including industrial \\[[@B14-sensors-18-04120]\\], structural health monitoring \\[[@B15-sensors-18-04120]\\], biochemical \\[[@B16-sensors-18-04120]\\] and medical \\[[@B17-sensors-18-04120]\\] applications. Optical fiber sensors also offer the possibility of using many different approaches for the sensor operation, which include intensity variation \\[[@B18-sensors-18-04120]\\], long period gratings \\[[@B19-sensors-18-04120]\\], fiber Bragg gratings (FBGs) \\[[@B20-sensors-18-04120]\\], non-uniform gratings \\[[@B21-sensors-18-04120]\\], nonlinear effects \\[[@B22-sensors-18-04120]\\] and interferometers \\[[@B23-sensors-18-04120]\\]. Among the aforementioned types of optical fiber sensors, FBGs present high multiplexing capabilities, where quasi-distributed sensor arrays can be obtained with high spatial resolution. These attractive features of FBGs have enabled the development of FBG-based sensors for temperature \\[[@B24-sensors-18-04120]\\], strain \\[[@B25-sensors-18-04120]\\], liquid level \\[[@B26-sensors-18-04120]\\], pressure \\[[@B27-sensors-18-04120]\\], force \\[[@B28-sensors-18-04120]\\] and torque monitoring \\[[@B29-sensors-18-04120]\\]. In addition, the FBG advantages are well-aligned with all the requirements of soft robotics and their applications on soft robotics bring important features and benefits to robotic devices based on flexible structures and actuators.\n\nAnother important advantage of optical fiber sensors is their ability to be embedded in different structures (rigid and flexible). Thus, FBG sensors are already embedded in 3D-printed structures \\[[@B30-sensors-18-04120],[@B31-sensors-18-04120],[@B32-sensors-18-04120]\\] and it is possible to foresee the 3D printing technology as the link between the FBG-based sensors and the soft robotics devices. In this way, the soft robotic device can be fabricated using 3D printing (as previously shown in \\[[@B7-sensors-18-04120]\\]) with an embedded FBG-based sensor system for measuring different parameters. Furthermore, the benefits of embedding FBG sensors in 3D-printed structures are already demonstrated, where it is possible to obtain a sensor that can withstand forces as high as 1 kN with enhanced temperature sensitivity \\[[@B32-sensors-18-04120]\\]. In addition, Homa et al. \\[[@B33-sensors-18-04120]\\] presented the spectral characteristics of FBGs embedded in PLA structures under different strain and temperature conditions.\n\nSince the flexible structures and actuators in soft robotics are generally made of polymers, temperature plays an important role in the robot's operation due to the polymer viscoelastic nature. The polymer viscoelasticity leads to a non-constant response to stress and strain \\[[@B34-sensors-18-04120]\\], in which the polymer's Young's modulus varies due to temperature deviations \\[[@B35-sensors-18-04120]\\]. For this reason, temperature monitoring can compensate these effects on the actuator structure, which enables an accurate and reliable sensor system. In addition, environmental temperature monitoring is important in biomedical, industrial and structural health monitoring applications \\[[@B13-sensors-18-04120]\\], which are application fields for soft robotics devices \\[[@B5-sensors-18-04120]\\].\n\nAiming at developing a fully embedded temperature sensor with high resolution and in compliance with soft robot application requirements, this paper presents the design of FBG-embedded, 3D-printed temperature sensors. To verify the sensor behavior in different materials, the FBG was embedded in 3D-printed structures made of PLA and TPU, where the sensitivity, linearity and hysteresis of each sensor were compared with each other and an unembedded FBG sensor used as reference. The proposed temperature sensor has a compact design, can be easily fabricated with commercially available 3D printers, has electromagnetic field immunity and can be readily employed on the next generation of soft robotics. Moreover, none of the aforementioned works \\[[@B32-sensors-18-04120],[@B33-sensors-18-04120],[@B36-sensors-18-04120]\\] present a thorough study on the sensor hysteresis, linearity and response times, as the one proposed in this work. Additionally, to the authors' best knowledge, this is the first FBG embedded in a flexible 3D-printed structure as the one proposed in TPU. The comparison of this material with other 3D-printed structures brings important knowledge for the design of flexible structures with embedded sensors, which also complements the work presented in \\[[@B32-sensors-18-04120]\\], where the infill density was evaluated, but for only one material (ABS in that case). It is also noteworthy that the temperature sensor proposed in this work also has much higher sensitivity (with negligible hysteresis) than the ones previously reported using similar methods.\n\nThis paper is divided as follows. [Section 1](#sec1-sensors-18-04120){ref-type=\"sec\"} presented the motivation and introductory aspects of the work. [Section 2](#sec2-sensors-18-04120){ref-type=\"sec\"} describes the operation principle of the sensor with analytical and numerical approaches. [Section 3](#sec3-sensors-18-04120){ref-type=\"sec\"} depicts the experimental setup used in the sensor characterizations. Results and discussions are presented in [Section 4](#sec4-sensors-18-04120){ref-type=\"sec\"}. Finally, the concluding remarks and future investigations are discussed in [Section 5](#sec5-sensors-18-04120){ref-type=\"sec\"}.\n\n2. Operation Principle and Numerical Simulations {#sec2-sensors-18-04120}\n================================================\n\nThe proposed temperature sensor is based on FBGs embedded in a 3D-printed structure. FBGs are created through the fiber exposure on a periodic intensity pattern, which results in a refractive index modulation \\[[@B37-sensors-18-04120]\\]. There are different ways to create such modulation: interference between two beams \\[[@B37-sensors-18-04120]\\], phase mask \\[[@B20-sensors-18-04120]\\] and direct writing using a femtosecond laser \\[[@B38-sensors-18-04120]\\]. The Bragg wavelength (*\u03bb~B~*) is directly related to the effective refractive index (refractive index with the modulation created by the intensity pattern) and the grating period as follows:$$\\lambda_{B} = 2n_{eff}\\Lambda,$$ where *n~eff~* is the effective refractive index and *\u039b* is the grating period. Thus, variations in the grating period and refractive index lead to a shift of the Bragg wavelength. The grating period changes with variations in the fiber length, which can be created through axial strain and thermal expansion, whereas the refractive index of a fiber also varies with temperature and strain through the thermo-optic and photoelastic effects, respectively. Therefore, a FBG is intrinsically sensitive to temperature and strain following Equation (2). $$\\Delta\\lambda_{B} = \\left\\lbrack {(1 - P_{e})\\varepsilon_{fiber} + (\\alpha + \\zeta)\\Delta T} \\right\\rbrack\\lambda_{B}.$$\n\nIn Equation (2), *\u03b5~fiber~* is the strain on the fiber, *P~e~* is the photoelastic constant, \u03b1 is the fiber's thermal expansion coefficient, *\u03b6* is the thermo-optic coefficient, and \u0394*T* is the temperature variation. In this way, if the fiber is submitted to only temperature variations (without strain), the wavelength shift (*\u03bb~B~*) is related to the initial Bragg wavelength, thermo-optic and thermal expansion coefficients, where typical values for FBG inscribed in silica fibers are 8--11 pm/\u00b0C \\[[@B39-sensors-18-04120]\\].\n\nIt is important to note that, if the fiber is embedded in a material, the temperature behavior of the material must be considered. In this work, we considered two polymers commonly used as filaments of 3D printers: PLA and TPU. The employed polymers have different thermal expansion coefficients as well as mechanical properties. In both cases, the FBG was embedded in the center of a cylinder with 10 mm diameter and 25 mm length to guarantee that the FBG is fully embedded in the 3D-printed structure. The FBG embedment leads to a slight decrease in the reflectivity and to a red-shift on the Bragg wavelength due to the strain induced by the coating material, as previously shown both numerically and experimentally in \\[[@B32-sensors-18-04120]\\].\n\nTo show the temperature behavior of each polymer used on the FBG embedment, a numerical simulation using the finite element method (FEM) was performed using the software Ansys Workbench 15.0 to show the behavior of each material under temperature. The simulations were performed considering an initial temperature of 20 \u00b0C, final temperature of 70 \u00b0C and a convection coefficient of 2000 W/(m^2^ \u00b0C), which is the one considering forced convection \\[[@B40-sensors-18-04120]\\]. The results for PLA and TPU are shown in [Figure 1](#sensors-18-04120-f001){ref-type=\"fig\"}, where the directional heat flux is presented as a function of the cylinder front view.\n\nIn [Figure 1](#sensors-18-04120-f001){ref-type=\"fig\"}, the colormap shows a positive heat flux in the regions with color green (or higher wavelengths, such as red), whereas a negative heat flux is shown in the light green region (or lower wavelengths, such as blue). As shown in [Figure 1](#sensors-18-04120-f001){ref-type=\"fig\"}, there is a heat flux towards the center of the cylinder for both polymers, following the fundamental law of thermodynamics, where the heat flows from the hotter environment (the ambient at 70 \u00b0C in this case) to the colder body region (center of the cylinder). Comparing the values of heat flux for both polymers, the PLA shows higher values of heat flux, which can indicate that the PLA reaches the steady state faster than TPU, leading to a lower response time for the temperature sensor. However, the difference between the structures is not high, which can also indicate that temperature sensors with TPU and PLA will present similar values of response time.\n\nThe temperature variation on the coating materials leads to a thermal expansion of the polymers, thus, from the results in [Figure 1](#sensors-18-04120-f001){ref-type=\"fig\"}, it is possible to infer that such thermal expansion will create a strain vector pointing towards the center of the cylinder, i.e., the region where the FBG is embedded. To verify this assumption, another simulation was performed. In this case, the analysis was made with respect to the vector of the principal strain (as shown in [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}a for the TPU case). It is worth mentioning that the same behavior occurs when the simulation was performed with PLA. In addition, in this simulation, we also considered a V-groove in the center of the cylinder, which is the region where the FBG was positioned. Thus, when the temperature increases, the FBG is also subjected to a strain from the thermal expansion of the 3D-printed structure. To estimate the strain which the FBG would be subjected, another simulation was performed, where the strain (as a function of the temperature) in the vicinity of the FBG for both PLA and TPU are presented in [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}b. The differences between the proposed sensors and the previous ones presented in the literature (e.g., \\[[@B32-sensors-18-04120]\\]) are related not only to the material used in FBG embedment, but also the sensor structure. As shown in [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}a, the sensor was embedded in a cylindrical structure with 10 mm diameter (in contrast with \\[[@B32-sensors-18-04120]\\], where a rectangular shape was used). This structure leads to a uniform distribution of the thermally induced strain on the structure, which can reduce the sensor hysteresis. In addition, this high diameter of the structure is also responsible for a higher strain transmitted to the FBG due to the polymer's thermal expansion, resulting in higher temperature sensitivity when compared with previously reported solutions. Regarding the material, we also proposed an embedment in a TPU structure, which has a Young's modulus of about 1.5 GPa \\[[@B41-sensors-18-04120]\\], three times lower than the one of PLA (3.5 GPa). It is noteworthy that the Young's modulus of PLA is also higher than the one of ABS (2.5 GPa) \\[[@B42-sensors-18-04120]\\]. If polymers are analyzed, the viscoelastic nature of such materials must be considered, as they not have constant response with stress or strain and may also present hysteresis \\[[@B34-sensors-18-04120]\\]. The viscoelasticity is defined as the combination of viscous and elastic behavior of polymers. In the viscous behavior, it is expected that the polymer has a viscous-like behavior, otherwise there is an elastic-like behavior \\[[@B43-sensors-18-04120]\\]. It is possible to estimate if a polymer will have a viscous or elastic behavior with its glass transition temperature (T~g~). Above this temperatures, the polymer behavior tends to be more viscous \\[[@B43-sensors-18-04120]\\], which can result in higher hysteresis. Since the T~g~ of TPU is \u221250 \u00b0C, whereas the one of the PLA is about 120 \u00b0C, one can expect a higher hysteresis of the TPU structure, as the tested temperatures are above the material T~g~. Thus, the analytical and numerical analyses indicate that the proposed sensor approach can enhance the sensors performance, which defines the novelty of this work, where, by performing an analytical and numerical analysis of the sensor design, structure and materials used on the embedment can lead to major improvements in the sensors' performance.\n\nThe strain in the center of the cylinder shown in [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}b for TPU and PLA shows a higher strain from the TPU material, which is also related to the thermal stress defined as the product of the material Young's modulus, thermal expansion coefficient and temperature variation. Even though PLA has higher thermal expansion coefficient (4.1 \u00d7 10^\u22125^ \u00b0C^\u22121^) than TPU (2.0 \u00d7 10^\u22125^ \u00b0C^\u22121^), the Young's modulus of the TPU is only 0.25 GPa, whereas the one of PLA is 2.30 GPa. In addition, the Poisson's ratios of TPU and PLA are 0.49 and 0.33, respectively. Thus, considering the thermal effects in a three-dimensional body, as shown in Equation (3) for the *y* plane, TPU will present higher strains than the PLA material. $$\\varepsilon_{yy} = \\frac{1}{E}\\left( {- \\upsilon\\sigma_{xx} + \\sigma_{yy} - \\upsilon\\sigma_{zz}} \\right) + \\alpha\\Delta T,$$ where *E* is the material Young's modulus, *\u03c5* is the Poisson's ratio and *\u03c3* is the stress.\n\nFollowing the strain simulations, analysis of the strain on the FBG was performed. It is noteworthy that the Young's modulus of silica (about 70 GPa) is an order of magnitude higher than the ones of TPU and PLA. Hence, one can assume that the silica will impose a restriction on the 3D-printed structure deformation proportional to the ratio between silica Young's modulus and 3D-printed structure material Young's modulus, as also described in \\[[@B44-sensors-18-04120]\\] for a similar case. Since the PLA Young's modulus is about ten times higher than the one of TPU, it is expected that the PLA structure will lead to the highest temperature sensitivity (due to the higher transmitted strain than TPU), even though the TPU structure shows higher strain variation with temperature. In addition, as already discussed, it is expected a lower response time from the PLA structure due to its temperature response shown in [Figure 1](#sensors-18-04120-f001){ref-type=\"fig\"}. Following this same assumption, it is also expected that the PLA structure will present higher temperature sensitivity than the one reported in previous works with ABS structures \\[[@B32-sensors-18-04120]\\].\n\n3. Experimental Setup {#sec3-sensors-18-04120}\n=====================\n\nTo verify the assumptions presented in [Section 2](#sec2-sensors-18-04120){ref-type=\"sec\"} and to validate the temperature sensors, experimental tests are performed with the FBG sensors embedded in both materials (TPU and PLA). An unembedded FBG is also tested to provide a comparison between all different approaches. In all tested sensors, the FBGs were inscribed using the phase mask technique with a KrF Excimer laser operating at 248 nm and present central wavelength in the C-band with a physical length of 10 mm.\n\nFor the 3D-printed structures fabrication, we used the 3D printer Sethi3D S3 (Sethi, Brazil), where all structures have a 99% infill density, which is the one that results in highest temperature sensitivity and can withstand higher forces (up to 1 kN), as discussed in \\[[@B32-sensors-18-04120]\\]. As shown in [Figure 3](#sensors-18-04120-f003){ref-type=\"fig\"}a, the 3D-printed structure has a v-groove for the fiber positioning. The embedment process comprised of pausing the printing when the v-groove was printed and, then, positioning the fiber on the v-groove with the FBG region in the center of the 3D-printed structure. Thereafter, the printing process was resumed and the FBG-embedded 3D-printed structure was fabricated. The steps for the FBG-embedded 3D-printed temperature sensor fabrication are depicted in [Figure 3](#sensors-18-04120-f003){ref-type=\"fig\"}a, where the process of printing and embedding the fiber on the 3D-printed structure took only about 15 min. Furthermore, there is also the possibility of process automation, which leads to lower production times with the potential of large-scale production. Furthermore, FBG spectra before and after embedded in the TPU flexible structure is shown in [Figure 3](#sensors-18-04120-f003){ref-type=\"fig\"}b, which shows a slight decrease on the reflectivity and a red-shift of about 0.1 nm. Similar results are reported in \\[[@B32-sensors-18-04120]\\] for rigid materials, such as PLA.\n\nFollowing the sensor fabrication, all three samples (PLA-embedded, TPU-embedded and unembedded) temperature sensors were placed inside the thermostatic bath ECO-RE630 (LAUDA, Germany) with closed loop temperature control (temperature accuracy of 0.01 \u00b0C), where water was used as bath fluid with forced circulation. In addition, the wavelength shift of each sensor was monitored by the FBG interrogator sm125 (Micron Optics, Atlanta, CA, USA) with 1 pm wavelength resolution.\n\nThe validation tests comprised of increasing the temperature from 20 \u00b0C to 70 \u00b0C in steps of 10 \u00b0C, where each temperature was kept constant for about 5 min. Then, the temperature was decreased from 70 \u00b0C to 20 \u00b0C with the same steps to evaluate the sensor hysteresis. The response of each sensor was evaluated and compared with respect to the sensitivity, linearity and hysteresis. In addition, we also analyzed the response time of each sensor. The sensitivity was defined as the ratio between the wavelength shift and temperature variation, whereas the linearity was the determination coefficient (R^2^) between the sensor response and a linear regression curve \\[[@B45-sensors-18-04120]\\]. In addition, hysteresis was the difference in the sensor response between temperature cycles (with temperature increase and decrease).\n\n4. Results and Discussion {#sec4-sensors-18-04120}\n=========================\n\nThe results for the temperature characterization (made with the experimental setup described in [Section 3](#sec3-sensors-18-04120){ref-type=\"sec\"}) are presented in [Figure 4](#sensors-18-04120-f004){ref-type=\"fig\"}. [Figure 4](#sensors-18-04120-f004){ref-type=\"fig\"}a shows the Bragg wavelength for the FBG embedded in the PLA structure as a function of the time. The comparison between the responses of each structure is presented in [Figure 4](#sensors-18-04120-f004){ref-type=\"fig\"}b, where the wavelength shift of each sensor is presented as a function of the temperature. In addition, the sensitivity and linearity of each sensor are presented.\n\nWhen the linearity of each sensor was analyzed, it was possible to note that, even though all sensors presented high linearity (R^2^ higher than 0.99), the highest linearity was achieved with the unembedded FBG, whereas the lowest one was found for the FBG embedded in the TPU structure. The reason for this behavior may be related to the material stiffness, which can isolate the sensor from perturbations caused by the fluid circulation inside the thermostatic bath. Thus, such effect was lower for the unembedded silica fiber due to its higher stiffness, followed by the PLA structure, which has higher stiffness than the TPU. This behavior can also be explained by the material anisotropy, which can create a non-uniform strain field on the grating when the temperature increases. For this reason, it is expected that a material with higher anisotropy will result in lower linearity for the FBG-embedded temperature sensor.\n\nComparing the results presented in [Figure 4](#sensors-18-04120-f004){ref-type=\"fig\"}b, the FBG embedded in the 3D-printed PLA structure shows far higher sensitivity than the other sensors analyzed (TPU structure and unembedded). In fact, the FBG embedded PLA structure has a temperature sensitivity of about 139 pm/\u00b0C, which is one of the highest reported temperature sensitivities for silica FBG-based temperature sensors. The unembedded FBG shows a temperature sensitivity of about 10.5 pm/\u00b0C, which is in accordance with previously published works in the literature \\[[@B13-sensors-18-04120]\\]. The FBG embedded in TPU structure shows a slight increase in temperature sensitivity, which is about 13.8 pm/\u00b0C. This sensitivity is higher than the theoretical limitation of the FBG temperature sensitivity (in silica fibers), which is the application of Equation (2) without strain. Thus, the theoretical limitation of temperature sensitivity for FBGs depends on the silica thermal expansion coefficient (*\u03b1* = 0.55 \u00d7 10^\u22126^ \u00b0C^\u22121^) and thermo-optic coefficient (*\u03b6* = 6.8 \u00d7 10^\u22126^ \u00b0C^\u22121^) as well as the Bragg wavelength. Considering the aforementioned values, the theoretical temperature sensitivity of FBGs in silica fibers is about 10.9 pm/\u00b0C. This higher temperature sensitivity of the FBG-embedded TPU structure indicates that the wavelength shift is due to not only the temperature effects in the fiber, but also to the thermally induced stress/strain of the TPU structure on the grating. The same principle occurs with the FBG embedded in the PLA structure, where more strain is transmitted to the grating due to the PLA higher Young's modulus, which is about ten times higher when compared to the one of TPU. Interestingly, the temperature sensitivity of the PLA structure is also about ten times higher than that of the TPU, which indicates, as assumed in [Section 2](#sec2-sensors-18-04120){ref-type=\"sec\"}, the strain transmission from the 3D-printed structure to the FBG follows a direct proportionality with the materials Young's modulus. Thus, even though the TPU structure shows the higher strain with temperature variation (see [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}b), only a small part of this strain is actually transmitted to the silica fiber due to its much higher Young's modulus, whereas the strain transmission of the PLA structure is higher, resulting in a higher temperature sensitivity.\n\nFor the hysteresis analysis, two temperature cycles were performed, where the temperature was increased from 20 \u00b0C to 70 \u00b0C and, then, decreased from 70 \u00b0C to 20 \u00b0C. The results of this test are presented in [Figure 5](#sensors-18-04120-f005){ref-type=\"fig\"}. To provide a better visualization of each curve, [Figure 5](#sensors-18-04120-f005){ref-type=\"fig\"}a depicts the results obtained in the temperature cycles for the FBG-embedded TPU structure and unembedded FBG, whereas [Figure 5](#sensors-18-04120-f005){ref-type=\"fig\"}b shows the response for the FBG-embedded PLA structure due to the large sensitivity difference between the 3D-printed structures.\n\nThe comparison between all three samples shows a lower hysteresis for the unembedded FBG, where a negligible hysteresis was obtained (below the FBG interrogator resolution). In contrast, a much higher hysteresis was found in the TPU structure, where a hysteresis as high as 5% was obtained. The hysteresis was estimated through the ratio of the highest difference between increasing and decreasing temperature with the wavelength shift obtained in the whole cycle (20 \u00b0C to 70 \u00b0C). This high hysteresis found for the TPU is related to the material response, which can present hysteresis due to its viscoelasticity \\[[@B34-sensors-18-04120]\\] and the material anisotropy. Same analysis was made for the FBG-embedded PLA structure, but a much lower hysteresis was obtained. In this case, the maximum hysteresis is below 0.5%. Therefore, the PLA is a much more suitable material for the FBG embedment in temperature sensors applications, since it provides ten times higher sensitivity, higher linearity and much lower hysteresis than the TPU.\n\nThe last performed analysis is the characterization of the sensor response time. In this case, a 10-\u00b0C step was applied and the deviation between the sensor response and the temperature rise was analyzed for each tested sample on the interval between 20 \u00b0C and 30 \u00b0C due to the higher stability of the thermostatic bath in this interval. To provide a better comparison between the sensors, the response of each sensor is shown with respect to the estimated temperature by applying the characterization equations obtained for each FBG sensor in [Figure 4](#sensors-18-04120-f004){ref-type=\"fig\"}b. The response time of each sensor is depicted in [Figure 6](#sensors-18-04120-f006){ref-type=\"fig\"}, where the response time of each sensor is the difference between the temperature rise time obtained from each sensor with the temperature increase time of the thermostatic bath. Thus, it is possible to observe the transient response of each sensor, where the unembedded FBG presented the highest slope on the temperature curve (as a function of time), indicating its lower response time, i.e., it reaches the steady state temperature faster than the other tested sensors. The response time of the thermostatic bath is about 140 s, i.e., the thermostatic bath takes 140 s to increase the temperature from 20 \u00b0C to 30 \u00b0C. In addition, [Figure 6](#sensors-18-04120-f006){ref-type=\"fig\"} inset depicts the high sensitivity difference between the sensors at this temperature range.\n\nThe response time of each sensor as a function of the temperature was obtained by subtracting the time that each sensor takes to increase the temperature from 20 \u00b0C to 30 \u00b0C with the time that the thermostatic bath takes to increase the temperature in this same range. Thus, this analysis already accounts for the dynamics of the thermostatic bath (2.8 \u00b0C/min), which is limited to the bath fluid (water) dynamics, harming the application of higher temperature rates. Nevertheless, analyzing a lower temperature rate (0.8 \u00b0C/min), similar response times were found. The unembedded FBG shows the lowest response time, since it only takes 0.3 s to respond a temperature variation of 1 \u00b0C. For the 3D-printed structures, the response time is higher, which is expected due to the higher thermal resistance imposed by both PLA and TPU structures. The response times in these cases are 1.6 s and 1.8 s for the TPU and PLA, respectively. The maximum response time variation, when compared with the case where the lower temperature variation rate was applied (0.8 \u00b0C/min), was below 3% (for the TPU-embedded FBG). Such lower variation on the response time indicates that the response times estimated with the employed method are not influenced by the thermostatic bath dynamics (since these dynamics are compensated for in the response time calculation). The reason for such lower response time for the TPU structure is related to its higher thermal conductivity (1.9 W/(m \u00b0C)) when compared with the one of PLA (1.3 W/(m \u00b0C)). Following Fourier's law, such higher thermal conductivity coefficient leads to a faster temperature response.\n\nTo summarize the analyses performed, [Table 1](#sensors-18-04120-t001){ref-type=\"table\"} shows a comparison between the sensors with respect to linearity, sensitivity, hysteresis and response time. In addition, we also compares the temperature resolution of each sensor, which was estimated considering the sensitivity of each sensor and the FBG interrogator wavelength resolution (1 pm). Thus, it is expected a higher resolution of the PLA-embedded FBG due to its higher sensitivity. The results presented in [Table 1](#sensors-18-04120-t001){ref-type=\"table\"} show that a PLA-embedded FBG sensor is a suitable solution when high precision, sensitivity and resolution is required and the 3D printing methods employed for its fabrication are well-aligned with the requirements for soft robotics applications. Thus, for temperature sensing applications, the PLA is the preferred material for the FBG embedment, since this material provides higher sensitivity and linearity with lower hysteresis than the other two tested options. The thermally induced strain sensitivity of each material for FBG embedment depicted in [Figure 2](#sensors-18-04120-f002){ref-type=\"fig\"}b is also shown in [Table 1](#sensors-18-04120-t001){ref-type=\"table\"}, since it indicates that the TPU structure may present some advantages in strain sensing applications due to its higher flexibility when compared with the PLA.\n\n5. Conclusions {#sec5-sensors-18-04120}\n==============\n\nThis paper presents the development of a FBG-embedded temperature sensor using ALM techniques, specifically, fused deposition modeling, the so-called 3D printing, where the FBG was embedded in 3D-printed structures made of two different materials, namely PLA and TPU. A numerical analysis of the material behavior under temperature and its effects on the FBG was performed prior to the sample fabrication using the finite element method. Thereafter, the FBG-embedded 3D-printed structures were fabricated using PLA and TPU, whereas an unembedded FBG was used as reference to provide a comparison between the different sensors. The experimental analysis shows a high temperature sensitivity of the PLA-embedded FBG sensor, which is one of the highest reported sensitivities for temperature sensors (139 pm/\u00b0C). In addition, the PLA-embedded sensor shows an almost negligible hysteresis, where the maximum hysteresis was 0.5%. In addition, the TPU-embedded structure also shows higher sensitivity than the unembedded FBG and lower response time than the PLA-embedded one. The results reported in this work show the feasibility of applying ALM in conjunction with optical fiber sensors to provide a substantial enhancement on the optical fiber sensor performance. Furthermore, as the soft robotics field moves towards the use of 3D printing in actuators and flexible structures design, the proposed FBG-embedded sensor is an important tool in this application field. Thus, it is possible to design actuators, structures and sensors with the same manufacturing process, which results in a portable and highly customizable system.\n\nA.L.-J conceived and designed the experiments, simulations and fabricated the samples. J.C. performed the experiments; A.F., M.J.P. and C.M. inscribed the FBGs and fabricated the samples. All authors analyzed the data, and were involved in the paper writing and revision.\n\nThe authors acknowledge the financial support from FCT through the fellowship SFRH/BPD/109458/2015, program UID/EEA/50008/2013 by the National Funds through the Funda\u00e7\u00e3o para a Ci\u00eancia e a Tecnologia/Minist\u00e9rio da Educa\u00e7\u00e3o e Ci\u00eancia, and the European Regional Development Fund under the PT2020 Partnership Agreement. The authors also acknowledge the CAPES (88887.095626/2015-01) and FAPES (72982608). A. Frizera and M. J. Pontes acknowledge CNPq for the research productivity fellowships 304192/2016-3 and 310310/2015-6, respectively. The authors also acknowledge the financial support of CAPES- financing code 001.\n\nThe authors declare no conflict of interest.\n\n![Directional heat flux simulation (in W/m^2^) of TPU and PLA structures, where the cylinders have 10 mm diameter and 25 mm length for an ambient temperature of 70 \u00b0C.](sensors-18-04120-g001){#sensors-18-04120-f001}\n\n###### \n\n(**a**) Principal strain vector for the TPU structure. (**b**) Strain on the cylinder center for TPU and PLA as a function of the temperature.\n\n![](sensors-18-04120-g002a)\n\n![](sensors-18-04120-g002b)\n\n![(**a**) Fabrication steps of the FBG-embedded temperature sensor and a photograph of the assembled sensor with PLA and TPU. (**b**) FBG spectra before and after embedding in the TPU structure.](sensors-18-04120-g003){#sensors-18-04120-f003}\n\n###### \n\n(**a**) Temperature response of the FBG-embedded 3D-printed PLA structure. (**b**) Wavelength shift as a function of the temperature for the unembedded, and TPU- and PLA-embedded FBGs.\n\n![](sensors-18-04120-g004a)\n\n![](sensors-18-04120-g004b)\n\n###### \n\n(**a**) Unembedded and FBG-embedded 3D-printed TPU structure response for temperature cycle. (**b**) FBG-embedded 3D-printed PLA structure response for temperature cycle.\n\n![](sensors-18-04120-g005a)\n\n![](sensors-18-04120-g005b)\n\n![Response time of embedded and unembedded FBG sensors to a 10 \u00b0C temperature step. Inset shows the wavelength shift of each FBG as a function of the time.](sensors-18-04120-g006){#sensors-18-04120-f006}\n\nsensors-18-04120-t001_Table 1\n\n###### \n\nComparison between the performance parameters of embedded and unembedded FBGs.\n\n                             PLA-embedded   TPU-embedded   Unembedded\n  -------------------------- -------------- -------------- ------------\n  Sensitivity                139.0 pm/\u00b0C    13.8 pm/\u00b0C     10.5 pm/\u00b0C\n  Linearity                  0.9997         0.9953         0.9998\n  Hysteresis                 \\<0.5%         5%             \\<0.1%\n  Response time              1.8 s          1.6 s          0.3 s\n  Resolution                 0.007 \u00b0C       0.07 \u00b0C        0.09 \u00b0C\n  Thermally induced strain   400 \u03bc\u03b5/\u00b0C      214 \u03bc\u03b5/\u00b0C      \\-\n"}
+{"text": "INTRODUCTION {#sec1}\n============\n\nSmall bowel obstruction (SBO) in pregnancy is a high risk situation for both mother and baby with maternal mortality ranging from 2 to 4% and fetal/neonatal loss from 13 to 17%. Preterm delivery occurs in \\~45% of the cases. Adhesions were the most common pathology in 30--50% followed by internal hernias and volvulus \\[[@ref1], [@ref2]\\].\n\nIn a systematic review \\[[@ref2]\\] of 78 studies describing 92 pregnancies from 2006, the most common risk factor for SBO in pregnancy was previous abdominal surgery (49%). The most common presenting symptoms were abdominal pain (88%), vomiting (67%), examination findings of tenderness (49%) and distension (28%). Abnormal laboratory parameters were seen only in 26%. Webster *et al.* \\[[@ref1]\\] reviewed cases from 1992 to 2014 with 46 cases of SBO identified. Of the case reports, five out of six (83%) women with adhesive SBO were treated surgically. One was treated successfully with conservative management. Of the case series, 16 cases of adhesive SBO were treated surgically (94%) with 1 case successfully managed conservatively.\n\nThis case occurred in public (state-funded) hospital, which is the only type of hospital of which acute surgery is managed in New Zealand.\n\nCASE REPORT {#sec2}\n===========\n\nA 37-year-old female was admitted under obstetrics, 30\u00a0weeks pregnant, with vomiting and abdominal pain. She had presented to her provincial hospital with increasing abdominal and back pain, and the following day she developed nausea and vomiting so was referred to our tertiary centre for further assessment. Her past surgical history included a lower segment caesarean section, laparoscopic ovarian cystectomy, laparoscopic cholecystectomy and laparoscopic posterior fundoplication with crus repair. Her only medical history was mild asthma. She was a smoker.\n\nShe described the pain 'feeling like contractions' with severe nausea and dry retching. She was passing flatus and her bowels had opened normally the day prior. She denied urinary symptoms. On examination her observations were within normal range. Her abdomen was soft with a nontender uterus; however, she had severe left-sided abdominal with localized percussion tenderness.\n\nLaboratory studies showed a leucocytosis of 15.56 \u00d7 10^9^/l and neutrophilia of 11.75 \u00d7 10^9^/l. Her C-Reactive Protein (CRP) was 29 mg/l. Her haemoglobin was 109 g/l and she had unremarkable liver function tests. Her urine dipstick indicated ketones 2+ and protein 1+.\n\nAn ultrasound scan (USS) of her abdomen had been completed in the provincial hospital and had reported moderate dilatation of fluid-filled small bowel loops with hyperperistalsis in the left abdomen.\n\nA nasogastric (NG) tube was placed which gave symptomatic relief.\n\nThe case was discussed with the consultant on call. A magnetic resonance imaging (MRI) abdomen was performed, which was unremarkable. She was managed conservatively with an NG tube on free drainage; however, her symptoms persisted and she stopped passing flatus. She was counselled about radiation exposure, and the decision was made to trial Gastrografin oral contrast media followed by an abdominal X-ray (see [Fig. 1](#f1){ref-type=\"fig\"}). Her symptoms resolved, and she was discharged home. She later gave birth to a healthy baby girl at term.\n\n![Abdominal X-ray with Gastrografin oral contrast media seen in the large bowel, with her 30-week-old baby seen in the lower abdomen and pelvis.](rjaa018f1){#f1}\n\nDISCUSSION {#sec3}\n==========\n\nA PubMed, MEDLINE and Google Scholar search of 'small bowel obstruction' and 'pregnancy' and 'Gastrografin' or 'oral contrast media' returned one result.\n\nBower *et al*. \\[[@ref3]\\] looked at SBO overall, with a section specific to pregnancy. It stated that safety analyses have not been completed on gastrointestinal contrast in pregnancy. The following recommendations were then made: indications for urgent operative management are the same as for nonpregnant adults; urgent MRI is advised if there are not immediate indications for surgery; SBO owing to a cause other than adhesions should undergo urgent operative management; adhesive SBO should undergo trial of nonoperative management with a low threshold for operation and surgery for partial or incomplete adhesive SBO can be delayed for foetal maturity if there is no clinical evidence of a complication.\n\nSBO in pregnancy is an uncommon condition with limited available literature. We were unable to find any literature describing oral contrast media in SBO in pregnancy. Majority of cases in the literature were secondary to adhesions. The safety and use of oral contrast media (Gastrografin) have been validated as a treatment of adhesive SBO \\[[@ref4]\\]. In light of this, we consider oral contrast media as a method of treating adhesive SBO in pregnancy under the proviso that there are no signs that indicate urgent surgical intervention.\n\nConflict of Interest statement {#sec4}\n==============================\n\nNone declared.\n\nFunding {#sec5}\n=======\n\nNo funding was received.\n\nEthical Approval {#sec6}\n================\n\nNo ethics approval is required.\n\nCONSENT {#sec7}\n=======\n\nThe patient gave written informed consent for their image and case to be published.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nParental income or income bracket is often used as a surrogate for child socioeconomic status (SES). Some cross-sectional and longitudinal studies have shown associations between parental income or SES and incident childhood asthma, but the associations have not been consistent \\[[@B1]--[@B6]\\]. Income alone may not take into account other home and family characteristics that represent the child\\'s true SES; the need for a comprehensive measure of SES provided impetus for the development of the Material Deprivation Index.\n\nMaterial deprivation, which takes absence of high school graduation, lone parent families, government transfers, unemployment, low income, and homes needing major repairs into account, as well as income, has been proposed as a more thorough measure of childhood SES and has been measured as a validated component of the Ontario Marginalization Index (ON-Marg) \\[[@B7], [@B8]\\]. Material deprivation allows examination of the effects of wide-ranging disparities in social well-being on chronic health conditions in children, addressing a gap in our ability to evaluate needs associated with lower SES among children.\n\nAdults with high material deprivation have increased odds of asthma (OR 1.23; 95% CI, 1.17--1.28) \\[[@B7]\\]. Material deprivation has been associated with poor glycemic control in children with type I diabetes \\[[@B9]\\]. Stronger associations between childhood emergency department visits for asthma and air pollution have been demonstrated among children living in neighborhoods with high levels of deprivation \\[[@B10]\\]. To our knowledge, this index has not been studied in association with childhood asthma development or ongoing asthma not specific to emergency department visits. Neighborhood deprivation may represent a helpful tool for evaluating the effects of disparities in health and social advantages on the likelihood of developing and continuing to need healthcare visits for ongoing childhood asthma. We evaluated the associations between material deprivation and new-onset and ongoing childhood asthma among children living in the Greater Toronto Area (GTA).\n\n2. Methods {#sec2}\n==========\n\n2.1. Data Sources {#sec2.1}\n-----------------\n\nWe analyzed prospectively collected Ontario health administrative data regarding clinic visit records from the Ontario Health Insurance Plan (OHIP), emergency department records from the National Ambulatory Care Reporting System (NACRS), and hospitalization data from the Canadian Institute for Health Information-Discharge Abstract Database (CIHI-DAD), housed at the Institute for Clinical Evaluative Sciences (ICES). Outpatient visits to clinics and emergency departments and admissions to hospital for asthma are covered under provincial healthcare for all Ontario residents and are recorded in OHIP, NACRS, and CIHI-DAD for billing purposes. These data are complete for asthma visits of Ontario residents billed in the province of Ontario. We used the Registered Persons Database (RPDB) to identify children who were born between 1997 and 2003 and had lived in the Greater Toronto Area (GTA) at any time during their lives; these children\\'s administrative data were followed from birth until March 31, 2012. Children born in 1997 had up to 15\u2009years of data, and children born in 2003 had up to 8\u2009years of data. This study was approved by Research Ethics Boards at the Hospital for Sick Children and the University of Toronto.\n\nMaterial deprivation was developed by Matheson et al. to contribute to our understanding of inequalities in measures of health and social well-being and to allow standardization across various population groups and geographical areas. Material deprivation was generated using principal components factor analysis \\[[@B7], [@B8]\\] and has been validated by comparing 2001 and 2006 census tract measures and dissemination areas within the census tracts \\[[@B7], [@B8]\\]. The index was developed using mandatory self-reported data from 52,973 dissemination areas and 5,017 census tracts. Data were included in the index in aggregates by dissemination area, the smallest standard geographic area for which census data may be reported. To maintain confidentiality, data would not have been included for the smallest dissemination areas. The index was housed at ICES, which would not have reported aggregates of smaller than 6 families to prevent identification. Only aggregate data without individual identifiers were available to the investigators of this study.\n\nThe Material Deprivation Index comprises high school graduation, lone parent families, government transfers, unemployment, low income, and homes needing major repairs and ranges from a score of \u22122 (lowest deprivation) to +6 (highest deprivation). The index was reported in quintiles, with the lowest quintile being the least deprived and the highest quintile the most deprived. To align with other studies, material deprivation was dichotomized into the two highest quintiles versus the two lowest quintiles for the principal analysis.\n\nThe Ontario Asthma Surveillance Information System (OASIS) database, created in 2005, tracks individuals with reported healthcare administrative codes for asthma. Entry into the database requires two outpatient visits for asthma within two consecutive years or ever being hospitalized for asthma \\[[@B11]\\]. Children who meet the OASIS database entry criteria are followed longitudinally in the database, regardless of the presence or absence of future asthma visits. This system has been validated with good sensitivity and specificity against a clinical diagnosis of childhood asthma made by a physician \\[[@B12]\\] and by parental reporting of physician-diagnosed asthma in children \\[[@B13]\\]. Asthma encounters were defined by the OHIP or the primary diagnostic ICD-9/ICD-10 code in NACRS and CIHI-DAD (Supplemental [Table 1](#supplementary-material-1){ref-type=\"supplementary-material\"}). For children in the OASIS database, asthma visits recorded in the OHIP, NACRS, and CIHI databases were available for each year of life during which the child lived in Ontario. Incident asthma was defined as the timing of entry into the OASIS database. Ongoing asthma, distinguished from asthma that has been outgrown, was defined as at least one healthcare visit for asthma in a given year of the child\\'s life.\n\nWe also considered covariables that had previously established association with incident asthma \\[[@B1]--[@B3], [@B14], [@B15]\\]; all potential covariables that were available in the administrative databases were included in the study. Data from the RPDB were used to determine sex, and data from the OHIP, NACRS, and CIHI databases were used to determine history of preterm birth, any diagnostic coding for obesity (weight or body mass index \\>95^th^ percentile), and atopic conditions other than asthma, including allergic rhinitis, eczema, and food allergy. Preterm birth, obesity, and atopic conditions were dichotomous variables, with codes for these conditions either present or absent for each child in the administrative datasets.\n\n2.2. Statistical Analyses {#sec2.2}\n-------------------------\n\nAll analyses were performed in SAS 9.3 (SAS Institute, Cary, NC). We conducted asthma-free survival analysis to compare the time to incident asthma for the highest 2 quintiles of home neighborhood deprivation versus the lowest 2 quintiles. Unadjusted and adjusted Cox proportional hazards models were used to determine the association between high deprivation at birth and age of incident childhood asthma diagnosis. The proportional hazards assumption was justified, no multicollinearity was observed between variables, and the Schoenfeld residuals showed good model fit.\n\nGeneralized linear mixed models (GLMM) with fixed intercepts were used to evaluate the associations between ongoing asthma (at least one asthma visit in each year of the child\\'s life) and ongoing high deprivation in each year of the child\\'s life. Results of fixed- and random-intercept GLMM were similar, and the random slope GLMM model had a negligible and nonsignificant slope for deprivation. GLMM is a subject-specific statistical method that accounts for the clustering or repeated measures of healthcare visits for asthma by year of life for each child. Generalized Estimating Equations (GEE) with an exchangeable covariance structure are population-average methods that were used to confirm the results of the GLMM analysis.\n\nEach covariable was compared with deprivation and asthma in unadjusted models. In the adjusted models, all covariables were included as potential confounders and remained significant (*p* \\< 0.05) in the final adjusted model. Effect modification with neighborhood deprivation was evaluated for all confounders with *a priori* likelihood of interaction. Interaction terms were included in the model, and for variables showing evidence of statistically significant interaction, the models were re-run using indicator variables for each of the four categories of deprivation (0, 1) and each covariable (0, 1).\n\nHistories of asthma, material deprivation, preterm delivery, obesity, and other atopic conditions were determined from Ontario administrative records and were complete for children who had lived in Ontario for all of their lives. Birth neighborhood deprivation data were complete for 94.3--97.3% of children depending on year of birth, and neighborhood deprivation data were complete for 94.3--96.2% of children in any given year. Children not born in an Ontario hospital (27.8%) may have been missing data regarding preterm birth, and children who did not live in Ontario throughout the study period may have been missing data regarding asthma, obesity, or other atopic conditions if they had been coded as having these diagnoses while living outside Ontario. To evaluate the possible effects of children moving into and out of the study area during the study period, we performed sensitivity analyses restricted to children born in an Ontario hospital and children living in the GTA during each year of their lives.\n\n3. Results {#sec3}\n==========\n\nThere were 326,383 children with administrative healthcare records born from 1997--2003, inclusive, and living in Toronto. The OASIS criteria for incident asthma were met by 69,628 children (21.3%), with a median age of diagnosis of 2.5\u2009years (interquartile range (IQR) 0.9--4.9); 14,050 children (4.3%) with a median age of diagnosis of 3.5\u2009years (IQR 1.4--6.7) had ongoing asthma symptoms by ages 8--15\u2009years, which would have excluded children with early-life wheezing that did not persist into later childhood.\n\nIn unadjusted proportional hazards models, high-deprivation birth neighborhood was associated with development of asthma (hazard ratio (HR) 1.14; 95% confidence interval (CI), 1.12--1.16) ([Table 1](#tab1){ref-type=\"table\"}). Male sex, preterm birth, obesity, and atopic conditions other than asthma were also associated with incident childhood asthma ([Table 1](#tab1){ref-type=\"table\"}). Obesity (HR 1.12; 95% CI, 1.11--1.13) and atopic conditions other than asthma (HR 1.09; 95% CI, 1.08--1.10) were also independently associated with high birth neighborhood deprivation. In proportional hazards models adjusted for sex, preterm delivery, obesity, and atopic conditions other than asthma, children with high birth neighborhood deprivation were at increased risk of incident asthma (HR 1.11; 95% CI, 1.09--1.13) ([Table 1](#tab1){ref-type=\"table\"}). The association between material deprivation and incident asthma persisted when the diagnosis of incident asthma was restricted to asthma development among children who continued to have visits for asthma at ages 8--15\u2009years (HR 1.11; 95% CI, 1.06--1.15), excluding children with early-life wheezing that resolved by school age.\n\nChildren with high neighborhood deprivation in any year of life had increased odds of healthcare visits for asthma in that year (OR 1.03; 95% CI, 1.02--1.05) ([Table 2](#tab2){ref-type=\"table\"}). The analyses appeared to be robust, and the subject-specific and population-average results were similar. The covariables did not modify the associations between neighborhood deprivation and incident asthma or ongoing asthma.\n\nSimilar associations between high neighborhood deprivation and asthma were seen after restricting the analyses to children who were born in an Ontario hospital or children who had lived their whole lives in the GTA (Supplemental [Table 2](#supplementary-material-1){ref-type=\"supplementary-material\"}). Neighborhood material deprivation was strongly associated with neighborhood income quintile (odds ratio 75.6; 95% CI, 72.1--79.2), which was obtained from the RPDB and evaluated as a predictor of asthma development in separate models. Income quintile was associated with asthma development in unadjusted (HR 1.08; 95% CI, 1.06--1.11) and adjusted (HR 1.06; 95% CI, 1.03--1.08) models.\n\n4. Discussion {#sec4}\n=============\n\nOur results showed longitudinal associations between incident childhood asthma and neighborhood material deprivation at birth. Material deprivation in any year of life was also significantly associated with increased odds of visits for ongoing asthma in that year. The presence of an association between material deprivation and incident asthma among children who continued to have ongoing asthma until ages 8--15\u2009years suggested that the association held true beyond transient early-childhood wheezing. The associations between childhood asthma development and material deprivation were robust and independent of the associations between childhood asthma development and male sex, prematurity, obesity, and other atopic conditions. Neighborhood material deprivation was associated with a higher hazard of childhood asthma development than neighborhood income quintile without overlapping confidence intervals and may have represented a more relevant measure for evaluations of incident asthma in children.\n\nStudies of adults have shown material deprivation to be associated with asthma diagnosis \\[[@B7], [@B8]\\]. Stronger associations between childhood emergency department visits for asthma and air pollution have been demonstrated among children living in neighborhoods with high levels of deprivation \\[[@B10]\\]. Our study extends the findings of these previous studies by determining an association between material deprivation and new development of childhood asthma, rather than asthma exacerbations. We have also demonstrated an association with ongoing asthma, even among children who have not required emergency department visits.\n\nParental or household income has been commonly evaluated as a predictor of incident childhood asthma. A longitudinal Canadian study showed an increased hazard of incident childhood asthma at age 5\u2009years among children with a low (HR 1.33; 95% CI, 0.98--1.78) and very low (HR 1.35; 95% CI, 1.01--1.82) SES index \\[[@B1]\\]. Neighborhood income has also been associated with incident childhood asthma in Toronto, Canada (HR 1.06; 95% CI, 1.03--1.09) \\[[@B2]\\]. However, a birth cohort of children in Manitoba, Canada \\[[@B4]\\], showed no association between family income and risk of incident asthma at age 7\u2009years, after adjusting for maternal physician visits, hospitalizations, or prescription medications for depression or anxiety, suggesting that some of the variability in associations between childhood asthma and income may be due to other covariables considered in the models.\n\nThere is also evidence that the timing of income status may influence its association with incident childhood asthma. In a prospective Australian birth cohort, family income trajectories were modeled and physician-diagnosed asthma at age 14\u2009years was associated with chronic low income (OR 2.21; 95% CI, 1.17--4.17) \\[[@B5]\\], while children of families with increasing and decreasing income did not have an increased odds of asthma at age 14\u2009years. The variability of associations between income and incident childhood asthma, depending on the population and on the other covariables included, suggests that a more comprehensive measure of childhood well-being and SES should be used as a predictor in these models.\n\nMold and moisture damage is a frequently-studied characteristic related to home repair. Systematic reviews have generally shown associations between childhood asthma and home exposure to mold or moisture \\[[@B16]--[@B19]\\] although the results of individual cohort studies have been more variable \\[[@B20]--[@B22]\\]. In a nested case-control study of Canadian children, parent-reported visible mold exposure in pregnancy and childhood was not associated with asthma \\[[@B20]\\]. However, in a study of Taiwanese children without asthma, mold odor (OR 2.09; 95% CI, 1.30--3.37), parent-reported visible mold (OR 1.76; 95% CI, 1.18--2.62), and water damage (OR 2.80; 95% CI, 0.59--13.3) were associated with new-onset asthma \\[[@B21]\\]. In a Swedish nested case-control study, inspector-observed moldy odor along the skirting board was associated with asthma \\[[@B22]\\]. Home structural damage has been evaluated as a predictor of asthma control, and its improvement has been associated with improvement of asthma morbidity \\[[@B23]\\]. Although individual characteristics suggesting home disrepair have shown associations with asthma development, the associations have not been consistent across all studies. These findings underscore the possible utility of a more global measure of home repair being included in the evaluation of SES and its relationship with childhood asthma.\n\nOther components of the deprivation index, including no parental high school graduation, lone parent families, government transfers, and unemployment, have been previously included as covariables in the longitudinal evaluation of associations with incident asthma in children. Single-parent families have shown an increased hazard of incident childhood asthma by age 5\u2009years (HR 1.43; 95% CI, 1.17--1.76) \\[[@B1]\\]. Maternal receipt of social welfare has been associated with incident childhood asthma in a Canadian nested case-control study (OR 1.87; 95% CI, 1.43--2.44) \\[[@B20]\\]. Among elementary school children in southern California, Title I Funding supporting academic achievement in schools with \\>40% of children living in poverty has been associated with incident asthma (HR 1.68; 95% CI, 1.10--2.56) \\[[@B24]\\]. Low parental education (OR 2.62; 95% CI, 1.07--6.39) and unemployment (OR 2.38; 95% CI, 1.16--4.90) have been associated with maternal depressive symptoms, \\[[@B25]\\] which have in turn been associated with incident childhood asthma (OR 1.25; 95% CI, 1.01--1.55) \\[[@B4]\\]. Lower parental education has also been associated with asthma diagnosis in the first year of life (HR 1.32, 95% CI 1.18--1.47) \\[[@B6]\\].\n\nOur study adds to the literature by demonstrating robust and novel associations between neighborhood deprivation and incident and ongoing childhood asthma in a large birth cohort of children. This study\\'s strengths include its large sample size, unselected population of children, and the utilization of prospectively collected administrative data for the exposure, outcome, and covariables. We investigated a validated, comprehensive measure of material deprivation in children, which takes into account parental education, single parenthood, government transfers, unemployment, and homes in need of major repairs, as well as parental income. We also used a validated outcome measure with good sensitivity and specificity (91.4% and 82.9%, respectively) \\[[@B12]\\]. Sensitivity analyses showed that movement of children into and out of Ontario or the GTA did not substantially alter the associations between material deprivation and childhood asthma.\n\nNeighborhood-level material deprivation data may not apply to all individuals within a neighborhood. However, individual measures of factors making up the deprivation index also have limitations. Studies of asthma symptoms have suggested that individual and neighborhood measures of poverty may be independent of each other in some populations \\[[@B26]\\]. Longitudinal evaluation of neighborhood has been shown to be critical for accurate determination of associations with health outcomes \\[[@B27]\\]. Individual-level family income and home repair data also pose potential problems, including higher rates of missing data if people prefer not to report their income or state of home disrepair. Parental or observer reports of mold and moisture damage are difficult to standardize among different studies, which may also detract from the usefulness of individual-level data regarding home disrepair.\n\nThis study was conducted in a large, cosmopolitan, Canadian, urban center and may not be generalizable to smaller urban centers or rural communities. We did not have information regarding family history of asthma and personal exposures such as second-hand smoke inside the home or at other locations. Individual-level cohort studies will also be needed to re-evaluate the associations of these covariables with material deprivation.\n\n5. Conclusions {#sec5}\n==============\n\nOur results show longitudinal associations between home neighborhood material deprivation and incident childhood asthma and ongoing asthma in each year of the child\\'s life. This study demonstrates the utility of a validated measure of material deprivation in studies of childhood asthma. Neighborhood material deprivation may represent a helpful tool for evaluating the effects of disparities in health and social advantages on the likelihood of developing and continuing to need healthcare visits for ongoing childhood asthma.\n\nThe authors gratefully acknowledge Jingqin Zhu for her assistance with generating the deidentified study dataset from the ICES databases.\n\nCI:\n\n:   Confidence interval\n\nCIHI-DAD:\n\n:   Canadian Institute for Health Information-Discharge Abstract Database\n\nGEE:\n\n:   Generalized Estimating Equations\n\nGLMM:\n\n:   Generalized linear mixed models\n\nGTA:\n\n:   Greater Toronto Area\n\nHR:\n\n:   Hazard ratio\n\nICES:\n\n:   Institute for Clinical Evaluative Sciences\n\nIQR:\n\n:   Interquartile range\n\nNACRS:\n\n:   National Ambulatory Care Reporting System\n\nOASIS:\n\n:   Ontario Asthma Surveillance Information System\n\nOHIP:\n\n:   Ontario Health Insurance Plan\n\nON-Marg:\n\n:   Ontario Marginalization Index\n\nOR:\n\n:   Odds ratio\n\nRPDB:\n\n:   Registered Persons Database\n\nSES:\n\n:   Socioeconomic status.\n\nData Availability\n=================\n\nData for this project were accessed and generated within the Institute for Clinical Evaluative Sciences (ICES) Data Repository. Approval from ICES is necessary to access the data.\n\nDisclosure\n==========\n\nDr. Elinor Simons received salary support from AllerGen NCE, the Canadian Thoracic Society, and the Hospital for Sick Children.\n\nConflicts of Interest\n=====================\n\nThe authors have no financial relationships relevant to this article or conflicts of interest relevant to this article to disclose.\n\nAuthors\\' Contributions\n=======================\n\nDr. Simons conceptualized and designed the study, conducted the data analysis, drafted and revised the manuscript, and approved the final manuscript as submitted. Dr. Dell validated the Ontario Asthma Surveillance Information System, mentored the study design and analysis, revised the manuscript, and approved the final manuscript as submitted. Dr. Moineddin mentored the statistical design and analysis, revised the manuscript, and approved the final manuscript as submitted. Dr. To created and validated the Ontario Asthma Surveillance Information System, mentored the study design and analysis, revised the manuscript, and approved the final manuscript as submitted.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\nSupplemental Table 1: OHIP, ICD-9, and ICD-10 diagnostic codes. Supplemental Table 2: Associations between incident asthma and birth neighborhood deprivation among subgroups of children living in the Greater Toronto Area (GTA).\n\n###### \n\nClick here for additional data file.\n\n###### \n\nAssociations between childhood asthma development and birth neighborhood deprivation and covariables.\n\n  Predictor                    *N* (%)^b^       Unadjusted HR (95% CI)^c^ of incident asthma   Adjusted HR (95% CI)^c^ of incident asthma\n  ---------------------------- ---------------- ---------------------------------------------- --------------------------------------------\n  High deprivation             125 664 (38.5)   1.14 (1.12, 1.16)                              1.11 (1.09, 1.13)\n  Male sex                     167 560 (51.3)   1.37 (1.35, 1.39)                              1.43 (1.41, 1.46)\n  Preterm birth                22 639 (6.9)     1.56 (1.53, 1.59)                              1.58 (1.54, 1.62)\n  Obesity                      18 071 (5.5)     1.55 (1.52, 1.59)                              1.46 (1.42, 1.51)\n  Other atopic conditions^a^   222 667 (68.2)   2.79 (2.73, 2.84)                              2.96 (2.89, 3.03)\n\n^a^Atopic conditions other than asthma (allergic rhinitis, food allergy, and atopic dermatitis). ^b^Percentages account for missing data. ^c^Hazard ratio (95% confidence interval). Results of multivariable Cox proportional hazards models.\n\n###### \n\nAssociations between asthma visits and neighborhood deprivation and covariables in each year of life among children living in the Greater Toronto Area.\n\n  Predictor                    OR (95% CI)^b^ of asthma visits by year using fixed-intercept GLMM^c^   OR (95% CI)^b^ of asthma visits by year using GEE^d^\n  ---------------------------- ----------------------------------------------------------------------- ------------------------------------------------------\n  High deprivation             1.03 (1.02, 1.05)                                                       1.03 (1.02, 1.05)\n  Male sex                     1.36 (1.34, 1.37)                                                       1.37 (1.36, 1.39)\n  Preterm birth                1.47 (1.44, 1.51)                                                       1.52 (1.48, 1.56)\n  Obesity                      1.48 (1.45, 1.51)                                                       1.47 (1.44, 1.51)\n  Other atopic conditions^a^   2.82 (2.77, 2.87)                                                       2.79 (2.74, 2.85)\n\n^a^Atopic conditions other than asthma (allergic rhinitis, food allergy, and atopic dermatitis). ^b^Adjusted odds ratio of asthma (95% confidence interval), also adjusted for year of birth. Results of ^c^generalized linear mixed model (subject-specific model) and ^d^Generalized Estimating Equations (population-average model).\n\n[^1]: Academic Editor: J\u00f6rg D. Leuppi\n"}
+{"text": "1. Introduction {#sec1-cells-07-00253}\n===============\n\nDuchenne muscular dystrophy (DMD) is the most common form of muscular dystrophy. It is an X-linked recessive disorder that affects one per 3500 live-born males \\[[@B1-cells-07-00253]\\]. It is caused by mutations in the *DMD* gene (cytogenetic location: Xp21.2--p21.1) which encodes dystrophin, a protein that is expressed at the muscle sarcolemma. Dystrophin is a basic component of the dystrophin-glycoprotein complex, which is involved in maintaining the stability of the plasma membrane of striated muscle fibers. Moreover, the dystrophin-glycoprotein complex plays important roles in mediating interactions between the cytoskeleton, membrane, and extracellular matrix components \\[[@B2-cells-07-00253]\\]. Defects in the dystrophin protein affect membrane integrity and lead to progressive degeneration and loss of skeletal and cardiac muscles \\[[@B3-cells-07-00253]\\].\n\nThe early stages of DMD are characterized by a process of gradual degeneration and regeneration of muscle fibers that is followed by the depletion of their regenerative ability, fibrosis, and the disruption of muscle tissue architecture. Clinically, DMD is accompanied by progressive muscle weakness and atrophy, which leads to disability in patients before the age of 12 years, and eventually to death caused by respiratory insufficiency \\[[@B4-cells-07-00253]\\]. In older patients with good management of respiratory failure, particular attention must be paid to the risk of heart failure, which represents the most frequent cause of death among adult DMD patients \\[[@B5-cells-07-00253]\\].\n\nUnfortunately, no effective therapy is available at present, and current therapeutic options are only palliative. Glucocorticoids, mainly prednisone and deflazacort, have been used to increase muscular strength and to retard the progression of disease. Moreover, they also reduce the need for scoliosis surgery, enhance lung function, and help maintain proper cardiac function \\[[@B6-cells-07-00253]\\]. More recent studies applying beta-blockers and angiotensin-converting enzyme inhibitors confirm their ability to delay the progression of DMD cardiomyopathy \\[[@B7-cells-07-00253]\\]. Great hopes have also been placed on gene therapy based on exon skipping to restore dystrophin production. In animal models, this technique resulted in a promising rescue of dystrophin expression in skeletal muscle tissue; however, the expression of dystrophin was much lower in cardiac muscle. \\[[@B8-cells-07-00253]\\].\n\nSeveral studies have used cell-based therapies to treat DMD. Pioneering studies employed myoblasts to promote the development of new or hybrid muscle fibers \\[[@B9-cells-07-00253],[@B10-cells-07-00253]\\]. However, this approach has many limitations, such as poor survival or low migratory ability of myoblasts \\[[@B11-cells-07-00253]\\]. In contrast to myoblasts, stem cells are multipotent and possess the capacity for long-term self-renewal, which makes them a unique tool for regenerative medicine, including for the regeneration of muscles. Their main advantage is that they may be obtained from different tissues and easily expanded to high quantities under in vitro conditions \\[[@B12-cells-07-00253],[@B13-cells-07-00253]\\]. In recent years, induced pluripotent stem cells (iPSCs) have also attracted significant interest from many researchers and clinicians. iPSCs can be generated from many specialized cells that have been reprogrammed by the ectopic expression of selected embryonic transcription factors (e.g., Oct4, Sox2, Lin28, Klf4 and L-Myc). This Nobel prize-winning technology can be used to produce patient-specific cells suitable for cell-based therapies of many pathological conditions, including DMD \\[[@B14-cells-07-00253]\\]. Moreover, iPSCs may be utilized for DMD modeling as well as for new drug discovery and testing \\[[@B15-cells-07-00253],[@B16-cells-07-00253]\\].\n\nIn this review, we briefly summarize the current state of knowledge on the preparation and biological features of iPSCs. We also discuss their potential for regeneration and the modeling of DMD.\n\n2. iPSCs Generation Techniques {#sec2-cells-07-00253}\n==============================\n\nThere are numerous cell reprogramming techniques to generate iPSCs. These techniques can be broadly divided into integrating and non-integrating delivery systems ([Table 1](#cells-07-00253-t001){ref-type=\"table\"}). Here, we provide a basic overview of these techniques and the history of iPSCs research.\n\nThe basic method of direct reprogramming of somatic cells was established by Yamanaka's group, who converted mouse fibroblasts to iPSCs through ectopic expression of the transcription factors Oct4, Sox2, Klf4, and c-Myc \\[[@B17-cells-07-00253]\\]. The same core set of factors has been used to prepare human iPSCs \\[[@B18-cells-07-00253]\\]. Variations on this basic reprogramming cocktail have also been used to successfully reprogram cells. For instance, Oct4, Sox2, Nanog, and Lin28 are sufficient to reprogram human fibroblasts \\[[@B21-cells-07-00253]\\]. It has been also shown that endogenous expression of certain transcription factors allows their exclusion from the reprogramming cocktail, such as c-Myc and Klf4 in mouse and human fibroblasts \\[[@B33-cells-07-00253]\\], or Sox2 and c-Myc in neural progenitor cells \\[[@B34-cells-07-00253]\\].\n\nRetroviral or lentiviral vectors were used in the first pioneering iPSCs studies. However, these integrating viral systems are controversial with respect to the clinical application of iPSCs, due to the increased probability of endogenous oncogene activation. For this reason, new non-integrating viral delivery systems have been introduced. Adenoviruses are an example of non-integrating viruses, so they have been tested as expression vehicles for producing iPSCs. However, the reprogramming efficiency of this method is very low \\[[@B19-cells-07-00253],[@B20-cells-07-00253]\\]. More recently, a non-integrating RNA virus, known as the Sendai virus, was used for the generation of iPSCs from blood cells and fibroblasts. They were reprogrammed after 25 days, with differing efficiencies (from 0.1% in blood cells to 1% in fibroblasts) \\[[@B22-cells-07-00253],[@B24-cells-07-00253]\\].\n\nIn addition to viral vector-based techniques, there are many non-viral approaches. For example, Okita et al. \\[[@B35-cells-07-00253]\\] generated iPSCs by the repeated transfection of mouse fibroblasts with plasmids containing the complementary DNA for Oct3/4, Sox2, and Klf4 and c-Myc. The ability to transfect cells with modified mRNAs for sustained expression of transcription factors provides another technique to generate footprint-free iPSCs (i.e., iPSCs that do not carry genomically integrated foreign DNA) from fibroblasts. Using this technique, a reprogramming efficiency of up to 4.4% was reported when Lin28 was added to the standard Yamanaka reprogramming factors in combination with culture medium containing valproic acid \\[[@B26-cells-07-00253]\\]. It was also found that several miRNAs (e.g., miR-302b, miR-372, miR367, miR200c and miR369) could reprogram cells at high efficiency, with or without the presence of the Yamanaka factors \\[[@B25-cells-07-00253],[@B27-cells-07-00253],[@B28-cells-07-00253]\\]. Episomal plasmids have also been used to produce footprint-free iPSCs, but they display only low levels of reprogramming efficacy \\[[@B23-cells-07-00253]\\]. The piggyBac transposon system also appears to be a promising method for iPSCs, although it is an integrating system that does not create footprint-free iPSCs. In addition, several obstacles must be circumvented to increase the efficiency of this approach \\[[@B30-cells-07-00253]\\].\n\nWhile the original set of four factors remains the standard for direct reprogramming, a group of small molecules and additional factors have been reported to increase reprogramming efficiency. Moreover, some of these seem to replace the effect of some of the transcription factors \\[[@B31-cells-07-00253]\\]. The majority of these act as epigenetic modifiers. Shi et al. \\[[@B32-cells-07-00253]\\] prepared iPSCs by using BIX01284, which acts as a specific inhibitor of histone methyltransferases, in combination with Oct4 and Klf4. Valproic acid (a histone deacetylase inhibitor), in combination with Oct4 and Sox2, was successfully used to induce pluripotency in human fibroblasts \\[[@B36-cells-07-00253]\\].\n\n3. Induction of Myogenic Progenitors and Precursor Cells from iPSCs {#sec3-cells-07-00253}\n===================================================================\n\nIt is well known that, during embryonic development, the process of myogenic activation is maintained by the up-regulation of myogenic regulatory factors (MYF5, MRF4, and MYOD), guided by the key paired-box transcription factors PAX3 and PAX7. The derivation of myogenic progenitors and precursor cells from iPSCs is based on the mimicking of embryonic mesodermal induction, followed by myogenic induction \\[[@B37-cells-07-00253]\\].\n\nMethods established for myogenic cell induction from embryonic stem cells (ESCs) have been used for iPSCs as well. There are two basic approaches for the generation of myogenic progenitors and precursor cells from iPSCs: direct reprogramming with muscle-specific transcription factors, including PAX3, PAX7, and MYOD (transgene method) ([Table 2](#cells-07-00253-t002){ref-type=\"table\"}), and the stepwise induction of skeletal muscle cells by means of small molecules and cytokines to inhibit or activate specific signaling pathways involved in the process of myogenesis (transgene-free method) ([Table 3](#cells-07-00253-t003){ref-type=\"table\"}) \\[[@B38-cells-07-00253]\\].\n\nDekel et al. described, for the first time, the induction of myogenic cells from ESCs by overexpression of MYOD in ESC cells that had been induced to form embryoid bodies (EBs). However, initial differentiation attempts showed low efficacy, caused by the inadequate recapitulation of paraxial mesoderm development during EB formation \\[[@B39-cells-07-00253]\\]. To overcome this hurdle, Darabi et al. used a lentiviral expression system to induce the expression of the transcription factor PAX3 during EB differentiation. Such PAX3 expression enhanced both the paraxial mesoderm formation and myogenic potential of cells \\[[@B40-cells-07-00253]\\]. In a following study, Darabi et al. \\[[@B37-cells-07-00253]\\] initiated the myogenic differentiation of ESCs/iPSCs by adding doxycycline on the second day of EB differentiation, to induce the expression of PAX7. The final maturation of cells was achieved by two-dimensional (2D) culture method in differentiation medium (2% horse serum and doxycycline withdrawal). After the implantation of iPSC-derived myogenic progenitors into the interior muscle of a mouse model for DMD, regeneration and the contractibility of transplanted muscle cells were detected. Mizuno et al. \\[[@B41-cells-07-00253]\\] also demonstrated the potential of murine iPSCs to differentiate into skeletal muscle progenitor cells, using a similar stepwise protocol involving EB formation by the hanging drop method. Differences were in the composition of the myogenic differentiation medium. Spindle-shaped fibers in the EBs were detected 7 days after they were plated onto Matrigel (day 13 of differentiation), and spontaneous contraction of these fibers was observed at day 27 of differentiation. The iPSC-derived myogenic cells were stained with an antibody against the anti-satellite cell marker SM/C-2.6, which is a cell surface marker for murine skeletal muscle. Cell sorting for SM/C-2.6-positive cells was performed by fluorescence-activated cell sorter (FACS). Results showed that the amount of iPSC-SM/C-2.6-positive cells significantly increased during myogenic differentiation. The cells were subsequently transplanted into the damaged muscle of mdx mice. Immunostaining analyses confirmed the successful engraftment of iPSC-SM/C-2.6+ cells and the absence of teratoma formation. Goudenege et al. \\[[@B29-cells-07-00253]\\] developed a two-step protocol to differentiate human iPSCs derived from DMD patient fibroblasts into myoblasts. Differentiation started with the culturing of iPSCs in myogenic medium (MB1), followed by infection with adenovirus expressing MYOD. Subsequently, the infected cells begun to form multinucleated myotubes. Four weeks after transplantation into the muscle of Rag/mdx mice, the fusion of myotubes with muscle fibers was observed. Moreover, there was no sign of teratoma formation.\n\nDespite high efficiency (more than 90%) and a fast process of differentiation for the approaches described above, the use of specific gene overexpression carries a risk of genetic recombination. Thus, myogenic cells generated by these methods are not suitable for possible clinical application. Therefore, there is a need for the development of appropriate transgene-free methods with comparable differentiation efficiency and a lower risk of genetic aberrations and tumor formation.\n\nFor cell reprogramming, Warren et al. \\[[@B26-cells-07-00253]\\] used a safer, non-integrating method based on repeated transfections of synthetic mRNA constructed to overcome innate anti-viral responses. The authors showed that this approach can be used for RNA-mediated direct differentiation of RNA-iPSCs (iPSCs derived by RNA transfections of cells with standard transcription factors) to mature myogenic cells. Myogenic differentiation was achieved by repeated daily transfections of RNA-iPSCs with MYOD-encoding modified RNA for 3 days, followed by 3 days of culturing in low serum medium. Subsequent immunostaining displayed high amounts of myogenin and MyHC-positive myotubes.\n\nMore suitable and safer alternatives for myogenic differentiation may also be achieved with methods using defined culture conditions, with the addition of specific molecules and growth factors that play essential roles in muscle development. Studies in embryonic development have established the essential roles of the Wnt signaling pathway and BMPs in myogenesis. In an extensive study, Xu et al. \\[[@B47-cells-07-00253]\\] examined the effects of 2400 chemical compounds on myogenesis and identified six potent myogenic inducers, including three GSK3\u03b2 inhibitors, two calpain inhibitors, bFGF, and forskolin. GSK3\u03b2 inhibitors are Wnt signaling activators, forskolin acts to stimulate cAMP signaling, and bFGF promotes myogenesis by activating the FGF receptor tyrosine kinase. The authors then tested the effects of such compounds on the differentiation of human iPSCs. In place of any of the GSK3\u03b2 inhibitors originally identified in the screen, the authors substituted the GSK3\u03b2 inhibitor BIO (6-bromoindirubin-3\u2032-oxime) because it displayed less toxicity. Embryoid bodies (EBs) derived from human iPSCs were cultured in a cocktail made of these molecules, and it was shown that the molecules induced cell differentiation into myogenic progenitors. After the first 7 days of cultivation, immunostaining was performed and showed the presence of muscle-specific proteins in the nuclei of EBs. To continue studying the process of muscle differentiation, the EBs were then cultured on Matrigel-coated dishes. At day 36 of culture, the EBs showed the formation of multinucleated myotubes. Moreover, the authors demonstrated the regenerative abilities of iPSC-derived muscle progenitors by successfully engrafting them into the pre-injured limb muscle of immunodeficient mice.\n\nHosoyama et al. \\[[@B48-cells-07-00253]\\] published a method of inducing muscle differentiation of iPSCs which relies upon the use of free-floating spherical cultures (EZ spheres) in a specific medium. The differentiation medium contained high concentrations of bFGF-2 and epidermal growth factor (EGF), stimulating the formation of EZ spheres. After 6 weeks of culture, iPSC-derived progenitors were detected and after 2 more weeks of terminal differentiation the authors observed multinucleated myotubes expressing PAX7, MYOD, MHC and myogenin. Based on these results, the authors concluded that a high concentration of bFGF-2 is a crucial factor for myogenic differentiation. In a follow-up study \\[[@B53-cells-07-00253]\\], the authors expanded their protocol in order to generate mature skeletal myotubes with organized sarcomeres. Moreover, the authors examined the influence of culture conditions, differentiation duration, culture surface coating, and medium components on the process of myogenic differentiation. Finally, the ability of three-dimensional (3D) cultures to form elongated and fully differentiated myotubes was tested by a bioengineering approach. The authors found that long-term differentiation (over 6 weeks) reduced the number of immature myogenic PAX7 positive (PAX7^+^) progenitors and MYOD/MyoG positive (MYOD^+^/MyoG^+^) myoblasts; on the other hand, the number of multinucleated myotubes was increased in comparison to iPSCs-derived myotubes at 2 weeks. Based on these results, authors suggested a minimal time of at least 6 weeks for the complete differentiation process of iPSCs into myocytes with sarcomere organization. However, examination of the myocytes under electron microscopy did not show the characteristic features of fully mature skeletal muscle fibers. Culture surface coatings of laminin and Matrigel displayed similar effects on the myogenic differentiation process. It was also shown that B27 serum-free supplement increased the efficacy of myogenesis compared to horse serum.\n\nThe selective GSK3\u03b2 inhibitor CHIR99021 is one of a number of specific compounds able to enhance muscle differentiation. CHIR99021 appears to do this by increasing the expression of mesoderm genes including *T*, *TBX6*, and *MSGN1* \\[[@B54-cells-07-00253]\\]. Shelton et al. \\[[@B55-cells-07-00253]\\] used CHIR99021, together with FGF2 treatment, to induce myogenic progenitors from ESCs, which subsequently underwent N2-mediated final differentiation. The resulting contractile skeletal myoblast population was observed at day 40. The total efficiency of differentiation, which was shown by the expression of *PAX7* and *MYH*, was 90%. However, it has also been shown that the exposure of ESCs/iPSCs to higher CHIR99021 concentrations leads to toxicity \\[[@B49-cells-07-00253]\\]. On the basis of this knowledge, van der Wall et al. \\[[@B50-cells-07-00253]\\] examined different concentrations of CHIR99021 in medium containing FGF2 as well as the time of treatment duration in order to find the least harmful treatment conditions for myogenic induction of iPSCs. Their results showed that the highest number of PAX7^+^ cells achieved without toxicity was seen after 4--5 days of treatment with 4 \u00b5M CHIR99021. The authors observed multinucleated myotubes between 30--40 days of differentiation. A senescent phenotype of differentiated cells was detected after 50 days of culture. Choi et al. \\[[@B51-cells-07-00253]\\] used CHIR99021 together with DAPT in their protocol for myogenic differentiation, resulting in successful the induction of ESCs/iPSCs into myogenic lineages in around 30 days. Several differentiated lineages contracted spontaneously in vitro and exhibited an ultrastructure similar to that of mature skeletal muscle fibers, including the presence of sarcomeres. However, it was necessary to purify cultured cells due to the presence of different cell types, such as myofibers, myotubes, and myoblasts, together with neurons and fibroblasts. Therefore, the desired myoblasts were purified by FACS using NCAM and HNK1 antibodies.\n\nOverall, transgene-free approaches represent a safer method of myogenic induction from pluripotent stem cells. However, they are still less efficient than transgene methods. Another issue is the ability to distinguish immature forms of myoblasts from mature ones \\[[@B54-cells-07-00253],[@B56-cells-07-00253]\\].\n\n4. iPSCs in Duchenne Muscular Dystrophy Modeling {#sec4-cells-07-00253}\n================================================\n\nOne of the most important properties of iPSC-derived cells is their utility in disease modeling. Remarkable progress has been made in the generation of human DMD iPSCs-derived myogenic cells in vitro, and in their usage for disease modeling. The generation of iPSCs with disease-specific phenotypes provides an unlimited source for the comprehensive study of DMD pathology, drug screening, and possible cell-based therapy ([Figure 1](#cells-07-00253-f001){ref-type=\"fig\"}).\n\nDarabi et al. \\[[@B37-cells-07-00253]\\] published one of the first studies on the generation of human iPSCs and their subsequent transplantation into DMD mice, leading to successful engraftment and improved contractility of treated muscle. Lin et al. \\[[@B57-cells-07-00253]\\] focused on studying the molecular mechanisms involved in dilated cardiomyopathy, which is a typical complication in DMD patients. These authors examined DMD-iPSC-derived cardiomyocytes and found dystrophin deficiency, mitochondrial damage, and elevated levels of resting Ca^2+^. Upon treatment with the membrane sealant Poloxamer 188, the resting cytosolic Ca^2+^ level considerably decreased, leading to the suppression of apoptosis in the DMD iPSC-derived cardiomyocytes. Shoji et al. \\[[@B46-cells-07-00253]\\] also observed in their iPSC-derived DMD model that excess influx of Ca^2+^ in DMD myotubes causes muscle damage. Due to this observation, they established an experimental system to recapitulate the early phase of DMD pathology and test the effect of an exon-skipping drug. They showed that the drug suppressed Ca^2+^ excess, thus leading to a decrease in cellular damage.\n\nMaffioletti et al. \\[[@B58-cells-07-00253]\\] created a three-dimensional (3D complex multilineage model of artificial skeletal muscle from DMD iPSCs. The micro-engineered artificial model mimicked features of human skeletal muscle tissue and was able to be engrafted into immunodeficient mice, thus representing an excellent model for studying DMD cellular pathology. Abujarour et al. \\[[@B16-cells-07-00253]\\] used DMD iPSC-derived myoblasts to study Wnt7a and IGF-1 as possible treatments for DMD. Wnt7a and IGF-1 are growth factors that are currently being evaluated in clinical studies for DMD treatment. Abujarour et al. tested these growth factors on iPSC-derived myotubes and showed that they caused significant hypertrophy to the cells, proving that iPSC-derived myotubes are functionally responsive to these factors and validating their potential use as a model for drug discovery. In summary, these studies provide evidence for high-throughput and cost-effective methods of DMD disease modeling and drug testing that may be sufficient to replace animal models in some circumstances. With the improvement of iPSCs reprogramming and subsequent myogenic differentiation methods, the possibilities for pre-clinical and clinical research will considerably expand.\n\n5. Genetic Correction {#sec5-cells-07-00253}\n=====================\n\nThe recent ability to correct known genetic defects in patient-derived iPSCs by various genetic-engineering methods provides hope for DMD treatment by autologous cell replacement therapy. In an example of such gene correction, Ousterout et al. used zinc finger nucleases (ZFNs) to delete exon 51 from the dystrophin transcript of DMD-derived myoblasts. This genetic alteration led to the restoration of the dystrophin open reading frame and the rescue of dystrophin expression \\[[@B59-cells-07-00253]\\]. Although this study was intriguing, the CRISPR/Cas9 technology is currently the most popular and flexible genetic correction method in use. CRISPR/Cas9 technology can be used to correct genetic alterations in mutant genes with relative ease and raises hope for the treatment of genetic disorders.\n\nYoung et al. \\[[@B60-cells-07-00253]\\] achieved the largest CRISPR/Cas9-mediated deletion to date in the *DMD* gene. Using the CRISPR/Cas9 technology, they deleted exons 45--55 of *DMD*. They showed that this deletion reframed the *DMD* transcript in human iPSC-derived skeletal myotubes and cardiomyocytes, which then expressed stable dystrophin protein that improved membrane stability. Moreover, the successful restoration of dystrophin was demonstrated in vivo by the engraftment of corrected iPSC-derived skeletal myotubes into a mouse model of DMD. The *DMD* deletion described has the potential to be clinically relevant in 60% of DMD patient mutations; thus, the authors highlighted the significant therapy potential of a single pair of guide RNAs (gRNAs) to treat a great number of DMD patients. Recently, Duch\u00eane et al. \\[[@B61-cells-07-00253]\\] published a similar therapeutic approach based on use of a pair of single guide RNAs to form a hybrid exon, resulting not only in the restoration of the dystrophin gene reading frame, but also in the production of dystrophin protein with a functional structure. Li et al. \\[[@B45-cells-07-00253]\\] also performed and demonstrated genetic correction of dystrophin in DMD patient-derived iPSCs using both the CRISPR/Cas9 and TALEN platforms, and subsequently performed a genome-wide analysis for off-target mutations. Their results showed the successful genetic correction of *DMD* in patient-derived iPSCs with minimal off-target mutagenesis.\n\nIn summary, several methods of gene editing have been applied for the correction of the *DMD* gene. Of these, the CRISPR/Cas9 system in particular has passed multiple proof-of-principle tests with successful use in a number of neuromuscular diseases and is now commonly used in preclinical studies.\n\n6. Conclusions {#sec6-cells-07-00253}\n==============\n\nRecent advances in cell-based therapy and tissue engineering represent significant promise for muscle regeneration. According to the abovementioned studies, iPSC-derived myocytes have an enormous potential for future therapy of DMD. Rapidly expanding genetic correction methods, such as CRISPR/Cas9 system, are able to correct desired mutations related to DMD. Such modified iPSC-derived myocytes could be possibly used as a source for transplantation into patients. However, the major obstacle to be overcome is the safety of the reprogramming and differentiation process. Another issue is the difficulty of recapitulating late-onset disease phenotypes; therefore, more detailed studies of molecular and cellular mechanisms underlying DMD in animal models are essential.\n\nWe thank Vanessa Fogg from ScienceDocs for English language editing.\n\nWriting---review and editing, L.D., M.C. (Martina Culenova), and M.C. (Maria Csobonyeiova); supervision, L.D.\n\nThis work was funded by the Slovak Research and Development Agency, APVV-14-0032 and the Research Grant Agency of the Ministry of Education, Science, Research and Sport of the Slovak Republic, VEGA-1/0153/15.\n\nThe authors declare no conflict of interest.\n\n![iPSCs potential for Duchenne muscular dystrophy.](cells-07-00253-g001){#cells-07-00253-f001}\n\ncells-07-00253-t001_Table 1\n\n###### \n\nOverview of current reprogramming techniques.\n\n  Integrating Systems                                            Non-Integrating Systems\n  -------------------------------------------------------------- ------------------------------------------------------------------------------------------------\n  Retroviruses \\[[@B17-cells-07-00253],[@B18-cells-07-00253]\\]   Adenoviruses \\[[@B19-cells-07-00253],[@B20-cells-07-00253]\\]\n  Lentiviruses \\[[@B21-cells-07-00253]\\]                         Sendai virus \\[[@B20-cells-07-00253],[@B22-cells-07-00253]\\]\n  piggyBac transposons \\[[@B23-cells-07-00253]\\]                 Plasmids \\[[@B24-cells-07-00253]\\]\n                                                                 Episomal vectors \\[[@B25-cells-07-00253]\\]\n                                                                 mRNA \\[[@B26-cells-07-00253],[@B27-cells-07-00253],[@B28-cells-07-00253]\\]\n                                                                 miRNA \\[[@B29-cells-07-00253]\\]\n                                                                 Proteins/small molecules \\[[@B30-cells-07-00253],[@B31-cells-07-00253],[@B32-cells-07-00253]\\]\n\ncells-07-00253-t002_Table 2\n\n###### \n\nTransgene myogenic induction of induced pluripotent stem cells (iPSCs).\n\n  Donor Cell Type   Cell Culture Method   Transgenes of Myogenic Cells   Differentiated Cell Type            Reference\n  ----------------- --------------------- ------------------------------ ----------------------------------- ---------------------------\n  Fibroblasts       2D culture            *MYOD-ERT*                     *MYOD1*-expressing mesangioblasts   \\[[@B42-cells-07-00253]\\]\n  Fibroblasts       EB culture            *PAX7*                         Myogenic precursors                 \\[[@B37-cells-07-00253]\\]\n  Fibroblasts       2D culture            *MYOD*                         Myocytes                            \\[[@B43-cells-07-00253]\\]\n  Fibroblasts       2D culture            *MYOD*                         Myotubes                            \\[[@B16-cells-07-00253]\\]\n  Fibroblasts       2D culture            *MYOD*                         Myocytes                            \\[[@B44-cells-07-00253]\\]\n  Fibroblasts       2D culture            *MYOD*                         Skeletal muscle fibers              \\[[@B45-cells-07-00253]\\]\n  Fibroblasts       2D culture            *MYOD*                         Myocytes                            \\[[@B46-cells-07-00253]\\]\n  Fibroblasts       EB culture            *MYOD*                         Myoblasts                           \\[[@B29-cells-07-00253]\\]\n  Fibroblasts       EB culture            *MYOD*                         Myogenic cells                      \\[[@B26-cells-07-00253]\\]\n\ncells-07-00253-t003_Table 3\n\n###### \n\nTransgene-free myogenic induction of iPSCs.\n\n  Donor Cell Type   Cell Culture Method   Factors                                           Differentiated Cell Type   Reference\n  ----------------- --------------------- ------------------------------------------------- -------------------------- ---------------------------\n  Fibroblasts       EB culture            GSK3\u03b2 inhibitor, bFGF, forskolin                  Myotubes                   \\[[@B47-cells-07-00253]\\]\n  Fibroblasts       EZ spheres culture    bFGF-2, EGF                                       Myotubes                   \\[[@B48-cells-07-00253]\\]\n  Fibroblasts       2D culture            GSK3\u03b2 inhibitor, bFGF                             Myoblasts                  \\[[@B49-cells-07-00253]\\]\n  Fibroblasts       EB culture            FGF-2, GSK3\u03b2 inhibitor (CHIR99021)                Myofibers                  \\[[@B50-cells-07-00253]\\]\n  Fibroblasts       2D culture            GSK3\u03b2 inhibitor (CHIR99021), DAPT                 Myoblasts                  \\[[@B51-cells-07-00253]\\]\n  Fibroblasts       2D culture            GSK3\u03b2 inhibitor, BMP inhibitor, bFGF, HGF, IGF1   Myogenic progenitors       \\[[@B52-cells-07-00253]\\]\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nBone mineral density (BMD) test measures the density of minerals present in the bones using a special scan. This can be used to assess the strength of bones. Bones naturally become thinner as you grow older because existing bone is broken down faster than the new bone made. As a result of this, calcium and other minerals decrease in the bones and they become light in weight, less dense, and more fragile. The bones might break if it goes thinner and weaker. Therefore, thicker bones take longer time to get osteoporosis. Although osteoporosis can occur in men, it is most common in women older than 65 years of age. Prospective studies by Ravn et al. \\[[@B1]\\] and Bjarnason and Christiansen \\[[@B2]\\] indicates that early postmenopausal women who have low BMI lose more bone compared to those with higher BMI tertiles. Studies conducted by Van der Voort et al. \\[[@B3], [@B4]\\] illustrates that thinness is related to both osteoporosis and increased fracture risk; hence, low BMI was included in the risk assessment tools for evaluation of osteoporosis and osteoporotic fracture risk as suggested by Eddy et al. \\[[@B5]\\] and National Osteoporosis Foundation \\[[@B6]\\] and Black et al. \\[[@B7]\\]. Studies from different authors emphasize the importance of exercise on BMD. Chan et al. \\[[@B8]\\] and Loud et al. \\[[@B9]\\] in their study indicate that exercise at a younger age helps to maximize the BMD. Family history of osteoporosis and fracture, tobacco consumption, and dietary habits are some of the factors related to the occurrence of osteoporosis. Among the anthropometric variables, low BMI and low weight are also interrelated with the occurrence of osteoporosis. A study conducted by Iqbal et al. \\[[@B10]\\] found that low BMI is a good indicator for referral of women less than 60\u2009yr for measurements of BMD.\n\nThe present study was conducted to assess the association between BMI and status of BMD among individuals referred to Department of Radio diagnosis for DEXA scan.\n\n2. Materials and Methods {#sec2}\n========================\n\nThe present retrospective study was conducted in Gulf Medical University, Ajman, UAE. For the study, people who were referred to the Department of Radio diagnosis for DEXA scan were selected. People who underwent DEXA scan from January till September 2009 formed the study subjects. Those patients referred for DEXA scan had metabolic syndromes, bone diseases, and few had difficulty in walking. As it is a case record analysis \"agree to participate\" does not arise. DEXA Scanning was performed by two technicians, and the interpretation was done by one radiologist. For the present study, 25- to 80-year old individuals of both genders who attended various departments of Gulf Medical College Hospital and Research Centre, Ajman, and whose densitometries recovered were included. The tool used for data collection consisted of sociodemographic details like age and gender. Height and weight were measured and respondents were dressed in light clothes and did not wear shoes. BMI was calculated from the height and weight recorded while performing DEXA scan. The BMI was calculated based on the formula weight (kg)/\\[height (m)\\]^2^. The standard categorisation of BMI by CDC \\[[@B11]\\] indicates less than 18.5 as Underweight, 18.5--24.9 as Normal, 25.0--29.9 as overweight, and 30.0 and above as obese.\n\nDual-Energy X-ray Absorptiometry (DEXA) is the most accurate way to measure BMD. A certified technician measured BMD at the femoral neck and the lumbar spine (L2 to L4) using a DEXA densitometer. WHO criteria were used for categorizing the respondents based on DEXA results. The statistical analysis included descriptive statistics to rule out the association between sociodemographic characteristics and DEXA scan result, BMI, and status of BMD. The statistical tests used in this study were chi-square test and logistic regression. The variables considered were gender, age, and BMI. The difference between the subjects was considered significant if the *P* value was less than\u2009\u2009.05.\n\n3. Results {#sec3}\n==========\n\nThe study population consisted of 101 individuals. Distribution of gender showed that 87.1% were females and the remaining were males. The sex ratio was found to be male\u2009:\u2009female = 1\u2009:\u20096.7. Age ranged from 25 to 80 years. The mean and standard deviation of age was 50.5 \u00b1 11.5 years. The respondents were classified in to three broad age groups, 25--39 years, 40--59 years, and 60 years and above. About 68% in the study were in the age group of 40--59 years. More than 50% of the subjects were obese. A detailed description is given in [Table 1](#tab1){ref-type=\"table\"}.\n\nThe results of the DEXA scan of 101 respondents show that 35.6% were osteopaenic and 25.7% had osteoporosis. 38.6% had normal DEXA scan results. Among patients with osteopaenia and osteoporosis, about 77.4% were overweight or obese. In the present study, none of the participants in the age group 25--39 years had osteoporosis. Maximum number of osteoporosis cases was seen in the age group of 60 years and above. In the age group of 40--59 years, only 23% had osteoporosis. The details of DEXA scan results are given in [Table 2](#tab2){ref-type=\"table\"}.\n\nFor further analysis, the whole group based on the DEXA scan result was divided as normal and low bone mineral density. There were 39 subjects with normal BMD and 62 with low BMD. BMD was low in 82.4% of people with normal BMI, 78.1% among overweight and 44.2% among obese. The association between BMI and BMD was found to be statistically significant (*P* \\< .001). Agewise variations showed that 33.3% of those in age group 25--39 years, 60.9% in 40--59 years of age, and 88.2% of those aged 60 years and above had low BMD. There was a statistically significant association between BMD and age (*P* \\< .01). The details are depicted in [Table 3](#tab3){ref-type=\"table\"}.\n\nIn the present study, age and BMI were the two factors found to be statistically associated with BMD. These two variables were put into a binary logistic regression model to find the crude and adjusted Odds Ratio (OR). The results of this analysis showed that compared to normal BMI, the chances of getting low BMD in overweight persons are 50% less. This association was not statistically significant. But among obese persons, the chance of getting low bone mineral density was 89% less compared to normal, which was statistically significant. As far as age is concerned, compared to those in age group 25--39 years the chance of getting low bone mineral density in persons aged 40--59 years was 4.24 times higher, and this was statistically significant. But those with age 60 years and above, the chance of low bone mineral density is 23 times higher compared to those in the age group 25--39 years. This association was found to be statistically significant. The details are depicted in [Table 4](#tab4){ref-type=\"table\"}.\n\n4. Discussion {#sec4}\n=============\n\nOsteoporosis is a major public health problem all over the world. The epidemiology of osteoporosis is, however, not well known in most of the Middle East countries. The present study was conducted in Gulf Medical University, Ajman, UAE, to identify the factors associated with the occurrence of low BMD. The three factors considered included age, gender, and BMI. DEXA scan uses two different X-ray beams to assess the bone density in the spine and hip. Although osteoporosis involves the whole body, measurements of BMD at one site can be predictive of fractures at other sites. Less X-ray beams pass through strong bones. DEXA can measure as little as 2% of bone loss per year. It is quick, and very low doses of radiation are used, but it is more costly when compared to ultrasound testing.\n\nThe correlation between BMD at the femoral neck and BMI observed was highly positive in a crosssectional study conducted among postmenopausal women by Steinschneider et al. \\[[@B12]\\]. The findings suggest that the increased BMD commonly reported in overweight women may result from soft tissue interference with BMD determination by DEXA. A hospital-based study conducted in elderly males by Paniagua et al. \\[[@B13]\\] observed that 37.3% were normal, 35.6% were osteopaenic and 27.1% were osteoporotic. In the same study, 35.6% of patients had normal BMI, 3.4% were underweight, 47.5% overweight, and 13.6% obese. Overweight and obese men were more likely to have osteoporosis and osteopaenia. Similar studies by Felson et al. \\[[@B14]\\], Nguyen et al. \\[[@B15]\\] and Baheiraei et al. \\[[@B16]\\] also reported the consistent finding that lower BMI was associated with lower BMD.\n\nThe previous literature of Baheiraei et al., \\[[@B16]\\] Jones et al. \\[[@B17]\\], and Nguyen et al. \\[[@B18]\\] indicated that advancing age was associated with low BMD. In this study also we observed an association between age and bone mineral density. The chance of low bone mineral density among people with age 60 and above is 23 times higher compared to those with age 25--39 years.\n\nThis study is an attempt to address one of the important public health problems which can be controlled if preventive measures are taken at an early stage. More about the risk factors could not be investigated as this is a retrospective case record analysis.\n\n5. Conclusion {#sec5}\n=============\n\nThe results of this study suggest that advancing age and lower BMI are important risk factors for the occurrence of low BMD. Further studies are required to investigate the effect of other factors like exposure to sunlight, calcium intake, and other habits like smoking, diet, and so forth.\n\nThe authors are heartily thankful to Mrs. Jancy Joshi, Secretary to the GMU Research Division for computation of data and help in secretarial work in the preparation of this article.\n\n###### \n\nDistribution of respondents according to age, gender, and BMI.\n\n  Variable       Group    No.    \\%\n  -------------- -------- ------ ------\n  Gender         Male     13     12.9\n  Female         88       87.1   \n                                 \n  Age            25--39   15     14.9\n  40--59         69       68.3   \n  60 and above   17       16.8   \n                                 \n  BMI            Normal   17     16.8\n  Overweight     32       31.7   \n  Obese          52       51.5   \n\n###### \n\nDistribution of bone mineral density according to independent factors.\n\n                            Status of bone mineral density                                      \n  ---------------- -------- -------------------------------- ------ ------ ------ ------ ------ -----\n  BMI              Normal   3                                17.6   7      41.2   7      41.2   17\n  Overweight       7        21.9                             12     37.5   13     40.6   32     \n  Obese            29       55.8                             17     32.7   6      11.5   52     \n                                                                                                \n  Age (in Years)   25--39   10                               66.7   5      33.3   ---    ---    15\n  40--59           27       39.1                             26     37.7   16     23.2   69     \n  60+              2        11.8                             5      29.4   10     58.8   17     \n                                                                                                \n  Total                     39                               38.6   36     35.6   26     25.7   101\n\n###### \n\nAssociation between bone mineral density and independent factors.\n\n                            Status of bone mineral density   *P*                        \n  ---------------- -------- -------------------------------- ------ ------ ------ ----- -------------\n  BMI              Normal   3                                17.6   14     82.4   17    *P* \\< .001\n  Overweight       7        21.9                             25     78.1   32           \n  Obese            29       55.8                             23     44.2   52           \n                                                                                        \n  Age (in Years)   25--39   10                               66.7   5      33.3   15    *P* \\< .01\n  40--59           27       39.1                             42     60.9   69           \n  60+              2        11.8                             15     88.2   17           \n                                                                                        \n  Total                     39                               38.6   62     61.4   101   \n\n###### \n\nCrude and adjusted odds ratio of independent variables.\n\n  Independent factors   Crude    Adjusted                             \n  --------------------- -------- ------------- ------- -------------- -----\n  BMI                   Normal   1             ---     1              ---\n  Overweight            0.77     0.17--3.44    0.50    0.10--2.55     \n  Obese                 0.17     0.04--0.66    0.11    0.02--0.50     \n                                                                      \n  Age (in Years)        25--39   1             ---     1              ---\n  40--59                3.11     0.95--10.10   4.24    1.09--16.53    \n  60+                   15.0     2.42--93.0    22.94   3.12--168.68   \n\n[^1]: Academic Editor: Stuarts L. Silverman\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nThe known relationship between the level of abdominal adipose tissue and metabolic syndrome and cardiovascular diseases (CVDs) is longstanding, including the risk for type 2 diabetes and sleep apnea. The developed world\\'s adult population is showing ever-increasing rates of obesity and, consequently, an increase in a wide range of health risks \\[[@B1]\\]. Developing a deeper understanding of the impact of obesity on health risk using more detailed quantitative traits of obesity, beyond body mass index (BMI), such as different types of adipose tissue, can provide more insights into health risks and the biology of the impact of obesity.\n\nAlthough BMI has been used for a wide range of research including genetic epidemiology, BMI has a distinct limitation in that the measurements assume a uniform contribution to risk by all adipose tissue, not taking into account the variation of the adiposity type and location from individual to individual. Location of adipose tissue has a critical role in the overall impact of obesity, with centralized adiposity having a higher impact on metabolic disorder health risks \\[[@B2]\\]. Furthermore, adipose is not a single homogeneous tissue and has regional deposits of both subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) \\[[@B3]\\]. While SAT lies directly under the skin, VAT is the adipose tissue around the organs. Past epidemiological studies have suggested that VAT is more strongly associated with CVD and metabolic syndrome \\[[@B4]\\].\n\nOther measures of adiposity such as waist-hip ratio (WHR) have proved to be predictive measures of diabetes risk in men \\[[@B5]\\] and coronary heart disease risk in both sexes \\[[@B6], [@B7]\\]. While WHR is appropriate for determining the regional distribution of adipose tissue \\[[@B8]\\], it has shown moderate associations with the amount of VAT accumulation in the abdomen \\[[@B9]--[@B11]\\] and large variability in distinguishing VAT from SAT \\[[@B8]\\]. Waist circumference measures have shown stronger associations with abdominal VAT compared to WHR in both men and women \\[[@B8], [@B11]--[@B13]\\].\n\nAlternatively, the levels of VAT and SAT can be more accurately measured through computed tomography (CT), ultrasound, and magnetic resonance imaging (MRI) \\[[@B14]\\]. Research on the impact of obesity using CT scans has provided valuable insights into the risk of CVD \\[[@B15]\\]. Additionally, CT measures of adipose tissue have facilitated the exploration of the relationship between accumulation of abdominal SAT or VAT and genetic variation, showing sex-specific loci associated with VAT levels \\[[@B16], [@B17]\\]. Although there are known genetic contributions to adipose distribution \\[[@B18]--[@B21]\\], identifying genetic biomarkers using quantitative adipose data from large-scale CT studies provides an opportunity for a better understanding of the nexus between genetics, adiposity, and an exhaustive list of health outcomes and laboratory measures.\n\nAn obstacle to large-scale imaging analyses is the financial investment in the collection of data. Geisinger has electronic health record (EHR) data and millions of biomedical images encompassing patient visits across health and disease that have been collected in EPIC\u00ae since 1996, with a stable patient population that uses primary and specialty care services. Furthermore, Geisinger has the MyCode Community Health Initiative, a biorepository with a growing collection of genome-wide array data and whole-exome sequencing data that will eventually surpass 200,000 individuals \\[[@B22]\\]. Thus, there is a unique opportunity to use the existing imaging, clinical, and genetic data from Geisinger for new discoveries on the impact of obesity and the identification of risk factors that can be used as biomarkers in translational medicine. Through the application of advanced image processing and computer vision methods that scale up the use of thousands of images, many challenges to the usage of existing EHR and imaging data for the purposes of research can be surpassed.\n\nAs a proof-of-principle study, we used 2,545 distinct CT images from the EHR of Geisinger and used a \"greedy snake\" segmentation algorithm to automatically measure individual levels of VAT, SAT, and total adipose tissue (TAT), as well as adipose ratio levels. Using ICD-9 codes to define cases and controls for diagnoses and 37 clinical laboratory measures, we characterized associations between these CT-derived adipose measures and a wide range of obesity-related diagnoses as well as quantitative traits. Subsequently, we also performed sex-stratified analyses to identify sex-specific associations with clinical diagnoses and obesity-related traits. Finally, we performed a genome-wide association study (GWAS) between the adipose measures and \u223c600,000 common genetic variants and identified suggestive genetic associations.\n\n2. Methods {#sec2}\n==========\n\n2.1. Study Samples {#sec2.1}\n------------------\n\nThe EHR and genetic data of this study for 2,545 individuals came from Geisinger and the MyCode Community Health Initiative of Geisinger. The genetic data were genotyped as a part of the DiscovEHR collaboration between Geisinger and Regeneron Genetics. Given that most patients within Geisinger are of European American (EA) ancestry (97%), only EA subjects were included in this study ([Table 1](#tab1){ref-type=\"table\"}).\n\n2.2. Image Data Collection {#sec2.2}\n--------------------------\n\nAdipose tissue measurements were obtained using an image segmentation technique called the \"greedy snake\" algorithm that analyzes delineated areas of TAT, VAT, and SAT from preexisting abdominal CT scans within Geisinger\\'s EHR. Data analyses were conducted for adipose measures extracted from the CT slice showing the largest area of total adipose. Also, these same measures from the slice with the largest waist circumference within abdominal CT scans were evaluated.\n\nOur previous work \\[[@B23]\\] describes the algorithm in great detail; here, we just briefly review the main steps. We formulated automatic fat quantification as an unsupervised contour minimization problem. The proposed algorithm comprised four major parts: (1) data preprocessing, (2) outer body contour estimation, (3) abdominal contour estimation, and (4) adipose quantification. Data preprocessing was done using standard image processing operations. Given an abdominal CT image, the first step was to separate the body from the whole image. This was done with a simple thresholding, taking into account the Hounsfield range of regions of interest. Next, a morphological opening operation was used to remove material artifacts like tables and trays. After segmenting the body from the image, the Moore-neighbor tracing algorithm was used to estimate the outer body contour. In computer vision, active contours are widely used over edge detection algorithms to locate the contour of an object, and they impose energy minimization of properties like smoothness and continuity to make the segmentation robust to noises and edge discontinuities.\n\nThe \"greedy snake\" algorithm differs from the original \"active contour\" algorithm by computing the movement of each contour point in a discrete manner. The outer body contour is provided as an initial contour to the \"greedy snake\" algorithm. At each iteration, the algorithm then makes a greedy choice and moves contour points to the position of minimum energy, a linear combination of image energy, elasticity energy, and the curvature energy of the image and the contour points, respectively. In this way, we could identify the body contour and the abdominal cavity contour.\n\nIn the last step, we quantified various fat tissues using the body contour and the abdominal cavity contour, respectively. Any pixel within the Hounsfield unit range (\u2212190, \u221230) was identified as adipose. The TAT region was calculated by the area inside the outer body contour. The VAT region was determined by the area inside the abdominal cavity contour, and finally, the SAT region was calculated by the area between the body and abdominal cavity contours. We also calculated visceral-to-subcutaneous fat ratios (VSRs) and visceral-to-total fat ratios (VTRs) using VAT and SAT. This algorithm identified TAT, VAT, and SAT segmentations with 0.885%, 3.55%, and 3.26% average error, respectively, as compared to a manual segmentation \\[[@B23]\\].\n\n2.3. Evaluation of Measures {#sec2.3}\n---------------------------\n\nFollowing the extraction of adipose tissue measurements from CT images, each measure type was evaluated for normality of data distribution and for identification of outliers. Adipose tissue measurements showing nonnormal distributions were Box-Cox transformed. Evaluation of all measures was done in R v3.4.0 \\[[@B24]\\]. Summary statistics of adipose measurements are given in Supplementary [Table 1](#supplementary-material-1){ref-type=\"supplementary-material\"}.\n\n2.4. ICD-9-Based Diagnoses {#sec2.4}\n--------------------------\n\nInternational Classification of Disease (ICD-9) diagnosis codes of the Geisinger EHR were used to define the case-control status for diagnoses. Cases were defined as individuals with three or more visits for a specific ICD-9 code at the 5-digit level (e.g., 250.60), whereas controls were defined as individuals having zero visits for the same code. Individuals were excluded from analysis for a given ICD-9 code if they had one to two visits for that specific ICD-9 code. We required at least 50 or more case subjects for the diagnosis to be included in our association testing. Based on our criteria for inclusion/exclusion, we used 1,233 ICD-9 codes for association testing, and those listed in Supplementary [Table 2](#supplementary-material-1){ref-type=\"supplementary-material\"} are the 263 phenotypes where the regression models converged.\n\n2.5. Clinical Lab Measures {#sec2.5}\n--------------------------\n\nWe had 37 different blood- and serum plasma-derived median/mean clinical laboratory measures from the Geisinger EHR. All clinical laboratory measures were evaluated for normality and outliers, where values greater than 2.5 standard deviation were removed. Log transformations were selectively applied after evaluating clinical laboratory data for normality. Supplementary [Table 3](#supplementary-material-1){ref-type=\"supplementary-material\"} shows the summary statistics of all laboratory measures and additional summary statistics of our dataset.\n\n2.6. Genotyping and Quality Control {#sec2.6}\n-----------------------------------\n\nGenotyping of Geisinger MyCode\u00ae participants was done using the Illumina HumanOmniExpress-12 v1.0 array through the DiscovEHR collaboration. Genotype quality control (QC) was performed prior to any association testing, using R 3.4.0 and PLINK \\[[@B25]\\] for the entire genetic dataset of MyCode (\u223c38,000 individuals at the time of this study). We filtered single-nucleotide polymorphisms (SNPs) for sample call rates (99%), genotyping (99%), and a minor allele frequency threshold of 1%. Additionally, highly related samples were then removed on the basis of their identity by descent kinship coefficient estimates (pi-hat\u2009\\>\u20090.125). Principal components were calculated using EIGENSOFT to confirm the EA ancestry status of individuals who also had imaging data. Following QC and filtering for samples that had both genotyping and adipose imaging data, we had 629,675 SNPs and 2,545 EA samples.\n\n2.7. Associations between CT-Derived Adipose Measures and Diagnoses and Clinical Laboratory Measures {#sec2.7}\n----------------------------------------------------------------------------------------------------\n\nAssociations between the CT-derived adipose measures and ICD-9-derived diagnoses or clinical laboratory measures were calculated using logistic regression and linear regression, respectively, using the software PLATO \\[[@B26]\\]. In both types of analyses, the responses were adjusted for age and sex. We also performed the analyses adjusted for age, sex, and type 2 diabetes status to determine if there were any appreciable differences in associations with those covariates. In addition to determining strength of associations between adipose measures and ICD-9 codes, odds ratios and 95% confidence intervals were also calculated in PLATO. Analyses were then repeated, stratified by sex, where age was used as a covariate.\n\n2.8. Associations between CT-Derived Adipose Measures and Diagnoses and Clinical Laboratory Measures for WC and TAT {#sec2.8}\n-------------------------------------------------------------------------------------------------------------------\n\nWe evaluated ICD-9 diagnosis associations from the optical slice with the greatest TAT, as well as ICD-9 diagnosis associations from the slice with the greatest WC. We did see a difference in the significance of associations when using the TAT measures compared to WC measures, for visceral adipose, but not subcutaneous adipose. All VAT and SAT association results presented within this manuscript are for associations from the imaging slices with greatest TAT.\n\n2.9. Bonferroni Correction and False Discovery Rate for Association Tests between Adipose Measures and EHR Data {#sec2.9}\n---------------------------------------------------------------------------------------------------------------\n\nA Bonferroni correction of 2.01\u2009\u00d7\u200910^\u22125^ (i.e., (1\u2009\u00d7\u200910^\u22122^)/(1,233\u2009\u00d7\u20094)) (the denominator corresponds to 1,233 ICD-9 codes and 4 phenotypes) was used as a significance threshold for association tests using ICD-9 code diagnoses and 7.14\u2009\u00d7\u200910^\u22125^ (i.e., (1\u2009\u00d7\u200910^\u22122^)/(35\u2009\u00d7\u20094)) (the denominator corresponds to 35 clinical lab measures and 4 phenotypes) was used in the association tests using clinical laboratory measures. In addition to the Bonferroni threshold, a false discovery rate (FDR) of 1% was determined. This translated to a level of significance of 2.09\u2009\u00d7\u200910^\u22124^ for analyses using logistic regression and 2.57\u2009\u00d7\u200910^\u22123^ for analyses using linear regression. A 1% FDR for the sex-stratified dataset translated into a level of significance of 1.58\u2009\u00d7\u200910^\u22124^ and 2.56\u2009\u00d7\u200910^\u22123^ for females and 1.38\u2009\u00d7\u200910^\u22124^ and 3.09\u2009\u00d7\u200910^\u22123^ for males for logistic regression and linear regression, respectively.\n\n2.10. Genetic Associations {#sec2.10}\n--------------------------\n\nGenetic associations between 629,675 SNPs obtained after QC and adipose tissue measures (VAT, SAT, VSR, and VTR) were calculated in PLATO using age and sex as covariates. Following the GWAS, SNPs were annotated using the NHGRI-EBI GWAS catalog \\[[@B27]\\] and GRASP \\[[@B28]\\] to determine if any associations within this study were replicated in previous studies, using a reported *p* value significance of at least 1\u2009\u00d7\u200910^\u22124^ from the GWAS catalog and GRASP.\n\n3. Results {#sec3}\n==========\n\nSummary statistics of all phenotypes can be found in [Table 1](#tab1){ref-type=\"table\"}. Our study had a total of 2,545 subjects, including 1,238 men and 1,307 women, with a mean age of 54\u2009\u00b1\u200917 (mean\u2009\u00b1\u2009standard deviation) years.\n\n3.1. Associations between Adipose Tissue Measures and Clinical Diagnosis Codes {#sec3.1}\n------------------------------------------------------------------------------\n\nThe results of all associations between adipose tissue and clinical diagnosis codes are presented in Supplementary [Figure 1](#supplementary-material-1){ref-type=\"supplementary-material\"}, and Supplementary [Table 4](#supplementary-material-1){ref-type=\"supplementary-material\"} shows association results, with *p* values \\<1\u2009\u00d7\u200910^\u22125^, of SAT and VAT with ICD-9 codes for sex-combined analyses. The most significant associations across all the diagnoses were between our CT-derived adipose tissue measures (VAT and SAT) and obesity-related ICD-9 diagnostic codes. [Figure 1](#fig1){ref-type=\"fig\"} shows both the *p* values plotted with \u2212log\u200910(*p* value) as well as the odds ratios (ORs) and confidence intervals (CIs) for the results passing our 1% FDR cutoff (refer to the full descriptions of the ICD-9 code abbreviations in Supplementary [Table 2](#supplementary-material-1){ref-type=\"supplementary-material\"}). The associations with morbid obesity were more significant by 10--20 orders of magnitude; therefore, the results were moved to a separate plot to maintain reasonable axes for the rest of the results. The results show the majority of the associations were positive and thus were associated with increased risk. [Figure 1(a)](#fig1){ref-type=\"fig\"} shows SAT had a stronger association with both morbid obesity (*p*\u2009=\u20093.29\u2009\u00d7\u200910^\u221283^; 95% CI\u2009=\u2009\\[5.45, 7.87\\]) and \"obesity unspecified\" (obesity NOS) (*p*\u2009=\u20091.41\u2009\u00d7\u200910^\u221268^; 95% CI\u2009=\u2009\\[2.98, 3.93\\]). However, out of the top results, VAT had a stronger association than SAT with other obesity-related ICD-9 codes such as \"diabetes mellitus without mention of complication, type II, or unspecified type, not stated as controlled\" (DMII wo cmp nt st uncrntr) (VAT: *p*\u2009=\u20091.49\u2009\u00d7\u200910^\u221258^; 95% CI\u2009=\u2009\\[2.29, 2.87\\], over SAT: *p*\u2009=\u20093.23\u2009\u00d7\u200910^\u221229^; 95% CI\u2009=\u2009\\[1.63, 2.00\\]) (Figures [1(a)](#fig1){ref-type=\"fig\"} and [1(b)](#fig1){ref-type=\"fig\"}). While the \u2212log(*p* values) and OR show trends of stronger associations with clinical diagnoses and VAT, the CIs for many of these same associations overlap between SAT and VAT.\n\nIn [Figure 2](#fig2){ref-type=\"fig\"}, we filtered the total associations for phenotypes that are related to obesity (e.g., comorbidities but not specifically obesity diagnoses). Again, the majority of associations showed an increased risk for obesity-related comorbidities. The only notable divergence from this was a decreased risk of osteoporosis, and there are conflicting reports regarding the relationship between obesity and bone mass \\[[@B29]\\]. As with the results of [Figure 1](#fig1){ref-type=\"fig\"}, VAT generally showed stronger associations with obesity-related comorbidities ([Figure 2(a)](#fig2){ref-type=\"fig\"}) as well as clinical diagnoses ([Figure 2(b)](#fig2){ref-type=\"fig\"}); however, their respective CIs showed an overlap between SAT and VAT ([Figure 2(b)](#fig2){ref-type=\"fig\"}).\n\n3.2. Associations between Adipose Tissue Measures and Clinical Laboratory Measures {#sec3.2}\n----------------------------------------------------------------------------------\n\nClinical lab measures obtained from outpatient clinics provided an opportunity to determine the relationship between adiposity and measures such as high-density lipoprotein (HDL) levels, low-density lipoprotein (LDL) levels, cholesterol (CHOL) levels, triglyceride (TRIG) levels, and white blood cell (WBC) counts. Supplementary [Table 5](#supplementary-material-1){ref-type=\"supplementary-material\"} presents results of sex-combined associations between SAT/VAT and clinical lab measures with *p* values \\<1\u2009\u00d7\u200910^\u22121^, and Supplementary [Table 6](#supplementary-material-1){ref-type=\"supplementary-material\"} provides summary statistics for the clinical lab measures. All obesity and adipose measures had similar directions of effect for associations with the clinical lab measures. However, VAT showed the most significant association for lipid levels, having the strongest association with both HDL (*p*\u2009=\u20091.42\u2009\u00d7\u200910^\u221236^; standard error (SE)\u2009=\u20095.80\u2009\u00d7\u200910^\u22123^) and TRIG (*p*\u2009=\u20091.44\u2009\u00d7\u200910^\u221243^; SE\u2009=\u20091.08\u2009\u00d7\u200910^\u22122^) ([Figure 3](#fig3){ref-type=\"fig\"}). Additionally, the adipose measures showed a negative direction of association with HDL (VAT *\u03b2*\u2009=\u2009\u22120.075) but a positive direction of association with TRIG (VAT *\u03b2*\u2009=\u20090.153). Interestingly, VAT also showed a significantly higher positive association with WBC (*p*\u2009=\u20097.37\u2009\u00d7\u200910^\u22129^; SE\u2009=\u20096.15\u2009\u00d7\u200910^\u22123^) compared to SAT (*p*\u2009=\u20091.21\u2009\u00d7\u200910^\u22124^; SE\u2009=\u20096.04\u2009\u00d7\u200910^\u22123^), although there was little difference from BMI (*p*\u2009=\u20092.21\u2009\u00d7\u200910^\u22129^; SE\u2009=\u20097.22\u2009\u00d7\u200910^\u22124^) ([Figure 3](#fig3){ref-type=\"fig\"}).\n\n3.3. Sex-Stratified Analyses {#sec3.3}\n----------------------------\n\nTo identify different trends of adipose deposition associated with health outcomes for men and women, we performed the same associations as above, sex stratified by both clinical diagnosis codes and lab measures. Summary statistics of sex-stratified clinical labs can be found in Supplementary [Table 7](#supplementary-material-1){ref-type=\"supplementary-material\"}. The results described here are limited to ICD-9 codes that were found to be associated in both sexes (some diagnoses are specific to women only and vice versa and thus would only be evaluated in one group or the other). Although the number of women and the number of men were very closely balanced, the strength of association for several different health outcomes was greater for women.\n\nMost notably, compared to males, SAT in females showed stronger associations with morbid obesity (*p*\u2009=\u20093.85\u2009\u00d7\u200910^\u221251^; 95% CI\u2009=\u2009\\[4.04, 6.15\\]), unspecified obesity (obesity NOS) (*p*\u2009=\u20091.32\u2009\u00d7\u200910^\u221239^; 95% CI\u2009=\u2009\\[2.63, 3.69\\]), and \"obstructive sleep apnea\" (*p*\u2009=\u20095.42\u2009\u00d7\u200910^\u221215^; 95% CI\u2009=\u2009\\[1.85, 2.79\\]) ([Figure 4(a)](#fig4){ref-type=\"fig\"}), although the ORs again had overlapping CIs ([Figure 4(b)](#fig4){ref-type=\"fig\"}). On the contrary, SAT in males, compared to females, showed a higher association with \"edema\" (*p*\u2009=\u20097.87\u2009\u00d7\u200910^\u221213^; 95% CI\u2009=\u2009\\[1.67, 2.46\\]) and \"cellulitis of the leg\" (*p*\u2009=\u20093.05\u2009\u00d7\u200910^\u22127^; 95% CI\u2009=\u2009\\[1.51, 2.52\\]) (Figures [4(c)](#fig4){ref-type=\"fig\"} and [4(d)](#fig4){ref-type=\"fig\"}). Beyond the top associations, neither sex showed additional associations for SAT that met at least the 1% FDR threshold (Supplementary Figures [2(a)](#supplementary-material-1){ref-type=\"supplementary-material\"} and [2(c)](#supplementary-material-1){ref-type=\"supplementary-material\"}).\n\nThe top associations between ICD-9 codes and VAT also showed sex specificity but differed from associations found in sex-combined analyses. Females showed strong evidence of VAT associations with \"diabetes mellitus without the mention of complication, type II, or unspecified type, not stated as controlled\" (*p*\u2009=\u20091.34\u2009\u00d7\u200910^\u221239^; 95% CI\u2009=\u2009\\[2.55, 3.55\\]), \"diabetes with renal manifestations, type II, or unspecified type, not stated as uncontrolled\" (DMII renl nt st uncntrld). (*p*\u2009=\u20091.18\u2009\u00d7\u200910^\u22129^; 95% CI\u2009=\u2009\\[1.77, 3.04\\]), and \"coronary atherosclerosis\" (Crnry athrscl natve vssl) (*p*\u2009=\u20091.11\u2009\u00d7\u200910^\u22126^; 95% CI\u2009=\u2009\\[1.35, 2.04\\]), with all relationships having significantly different ORs than males (Figures [5(a)](#fig5){ref-type=\"fig\"} and [5(b)](#fig5){ref-type=\"fig\"}). Outside of these top results, VAT associations for five ICD-9 codes (\"shortness of breath,\" \"gout\" (Gout NOS), \"osteoporosis\" (Osteoporosis NOS), \"neuropathy in diabetes,\" and \"osteoarthritis of lower leg\" (Loc prim osteoart-l/leg)) reached 1% FDR-level significance in females (1.58\u2009\u00d7\u200910^\u22124^) compared to none in males (Supplementary Figures [3(a)](#supplementary-material-1){ref-type=\"supplementary-material\"} and [3(c)](#supplementary-material-1){ref-type=\"supplementary-material\"}). Even beyond the FDR significant results, females showed a stronger and greater number of relationships between VAT and ICD-9 codes compared to males, although ORs did not significantly differ between sexes (Supplementary Figures [3(B)](#supplementary-material-1){ref-type=\"supplementary-material\"} and [3(D)](#supplementary-material-1){ref-type=\"supplementary-material\"}). Finally, like with SAT, males had a stronger VAT association with \"edema\" (*p*\u2009=\u20092.27\u2009\u00d7\u200910^\u22129^; 95% CI\u2009=\u2009\\[1.43, 2.04\\]) and \"cellulitis of the leg\" (*p*\u2009=\u20095.94\u2009\u00d7\u200910^\u22127^; 95% CI\u2009=\u2009\\[1.44, 2.31\\]); however, there were overlapping CIs for the ORs (Figures [5](#fig5){ref-type=\"fig\"} and [5](#fig5){ref-type=\"fig\"}). A full summary of sex-stratified association testing for ICD-9 codes, with *p* values\u2009\\<\u20091\u2009\u00d7\u200910^\u22122^, can be found in Supplementary [Table 8](#supplementary-material-1){ref-type=\"supplementary-material\"}.\n\nSimilar to the sex-combined analyses, key clinical laboratory measures were tested for their associations with adipose tissue measures. In females, VAT showed the strongest association with WBC count (*p*\u2009=\u20097.49\u2009\u00d7\u200910^\u221213^; SE\u2009=\u20097.58\u2009\u00d7\u200910^\u22123^), HDL (*p*\u2009=\u20095.94\u2009\u00d7\u200910^\u221226^; SE\u2009=\u20097.95\u2009\u00d7\u200910^\u22123^), and TRIG (*p*\u2009=\u20091.84\u2009\u00d7\u200910^\u221231^; SE\u2009=\u20091.39\u2009\u00d7\u200910^\u22122^), whereas in males, VAT showed the strongest associations with HDL (*p*\u2009=\u20095.50\u2009\u00d7\u200910^\u221213^; SE\u2009=\u20097.04\u2009\u00d7\u200910^\u22123^) and TRIG (*p*\u2009=\u20099.03\u2009\u00d7\u200910^\u221215^; SE\u2009=\u20091.39\u2009\u00d7\u200910^\u22122^) (Figures [6](#fig6){ref-type=\"fig\"} and [6](#fig6){ref-type=\"fig\"}). The direction of effect was similar to that of the sex-combined analyses. While both sexes showed a strong association between VAT and HDL as well as VAT and TRIG, females showed evidence of a stronger association with key obesity-related laboratory measures than males. This observation lends further support to the notion that visceral fat levels may impact obesity disease outcomes to a greater extent in females compared to males. A summary of the full results of sex-stratified clinical lab association testing can be found in Supplementary [Table 9](#supplementary-material-1){ref-type=\"supplementary-material\"}.\n\n3.4. Genome-Wide Association Studies of Adipose Tissue {#sec3.4}\n------------------------------------------------------\n\nManhattan plots for sex-combined analyses are presented in Supplementary Figures [4](#supplementary-material-1){ref-type=\"supplementary-material\"}--[7](#supplementary-material-1){ref-type=\"supplementary-material\"}, and a summary of the top results can be found in Supplementary [Table 10](#supplementary-material-1){ref-type=\"supplementary-material\"}. Among all genome-wide associations tests, only one SNP (rs10743966) reached genome-wide significance (*p*\u2009=\u20092.97\u2009\u00d7\u200910^\u22128^ for SAT; Supplementary [Figure 5](#supplementary-material-1){ref-type=\"supplementary-material\"}) and has previously been shown to be associated with WHR \\[[@B30]\\] and CVD \\[[@B31]\\]. Although no other SNPs reached statistical significance, GRASP annotations of the top results of SAT and VAT reveal previously established associations with obesity-related phenotypes. Additionally, SNPs associated with calculated adipose ratios (VSR and VTR) were also previously associated with obesity-related phenotypes. More specifically, the top SNP associated with VAT (rs933186; *p*\u2009=\u20091.29\u2009\u00d7\u200910^\u22127^) was previously shown to be related to HDL cholesterol \\[[@B32], [@B33]\\], while the top SNP associated with VSR (rs12950848; *p*\u2009=\u20096.45\u2009\u00d7\u200910^\u22127^) was previously shown to be related to lipid levels \\[[@B32], [@B33]\\] and total cholesterol \\[[@B32]\\]. Lastly, although VTR\\'s top association, rs7699631 (*p*\u2009=\u20091.99\u2009\u00d7\u200910^\u22126^), has no known associations with obesity-related traits, it was found to have a strong association with rs12950848 (*p*\u2009=\u20092.61\u2009\u00d7\u200910^\u22126^) which was earlier shown to be associated with LDL cholesterol \\[[@B32]\\].\n\n4. Discussion {#sec4}\n=============\n\nThe level of adipose tissue individuals carry has been previously shown to be an important risk factor for several obesity-related diseases. While BMI is typically used as a measurement of adipose levels, the location of adiposity is not reflected in this measure. Waist circumference is another measure that provides a reflection of the location of adiposity, in addition to the amount, as there is a known relationship between improved health outcomes for \"pear-shaped\" adiposity compared to \"apple-shaped,\" or more central, adiposity. Although BMI and waist-to-hip ratio have been shown to be predictors of CVD and mortality, they are not effective in distinguishing subcutaneous adipose from visceral adipose. Visceral adipose, the fat closest to the internal organs and more centrally deposited, is believed to have the most important impact on metabolic disorders, and increased visceral adipose has been linked to hypertension, atherosclerosis, and diabetes \\[[@B34]--[@B36]\\]. Individuals with high BMI can have a high level of physical fitness and a lower risk of mortality as compared to individuals with lower BMI but a different distribution of adiposity \\[[@B34]\\]. Women generally have a higher amount of body fat than men; however, women tend to carry excess adipose in their hips and thighs (gluteal-femoral region) \\[[@B37], [@B38]\\], whereas men carry it around their abdomen.\n\nIn this paper, we present a proof-of-principle study where we used imaging data from CT scans collected within a health system and then used a high-throughput automated approach to obtain measures of subcutaneous and visceral adipose tissue across individuals. An added advantage to this approach is that the imaging data can be coupled with the health data recorded on patients, including patients\\' diagnoses as well as their respective clinical laboratory measures. With this new approach, it is important to show we recapitulate known trends and relationships between adipose types, diagnoses, and clinical lab measures as a proof of principle in addition to potentially new discoveries.\n\nOur results show that obesity-related ICD-9 diagnostic codes have strong associations with VAT and SAT. In particular, morbid obesity was more strongly associated with SAT as compared to VAT. These results do reinforce that subcutaneous adipose levels provide an accurate reflection of the degree of obesity for an individual. Importantly, however, in line with the significant impact of visceral adipose on obesity-related conditions and comorbidities, our study shows the strength of the majority of associations is higher for VAT compared to SAT. The direction of effect of the majority of associations with obesity-related outcomes was for increased risk of these outcomes with increases in VAT.\n\nIntriguingly, even with balanced numbers of men and women and no major difference in power due to sample size, we saw the associations between SAT/VAT and obesity or obesity-related comorbidities to be consistently more significant for women than men. This will be an important area to research in a larger study with more CT-derived adipose measures to further determine the impact of these adipose levels when stratifying by sex. If these results replicate in a larger sample size, this could point to the increased importance on health outcomes of centralized adiposity for women versus men.\n\nIn the context of clinical lab measures, the negative direction of association for HDL, but a positive direction of association for triglycerides, particularly implies that the increase in VAT is associated with increased levels of triglycerides but decreased levels of HDL. HDL is considered a beneficial circulating lipid, while triglycerides are not, so this association points to the critical importance of VAT\\'s effect on obesity-related risks compared to BMI alone. Also, the significantly higher positive association of WBC with VAT compared to SAT may be related to increasing inflammation with obesity \\[[@B39], [@B40]\\]. Our clinical lab associations also suggest that the presence of a greater volume of visceral tissue in females may impact the outcomes of more obesity-related diseases than in males, as we again saw a more significant magnitude of association with these clinical lab measures in women compared to men. For this study, we used the median clinical lab measure each individual had across the lipid measures they had recorded in the EHR. In the future work, we will explore both the relationships of longitudinal lipid and blood cell count measures with visceral and subcutaneous adipose measures, characterizing the impact of medications such as lipid-lowering drugs on results when taken into account.\n\nNotably, for the majority of the associations of this study, except for the most statistically significant ones, there were overlapping confidence intervals for the association results of VAT and SAT. Thus, while we saw trends of the magnitude of association being higher for VAT compared to SAT for many diagnoses, the overlapping confidence intervals do not support a truly statistically significant difference. Power is always a consideration in association testing, and while ICD-9 codes have shown successful association testing in other studies \\[[@B41]\\], they can increase noise and decrease power. Thus, to strengthen these initial findings further, our future direction will be to deploy this strategy across thousands of CT images, thereby increasing our power to detect associations. We also had few statistically significant genetic associations and again that could be attributed to power for this small proof-of-principle study, and our future plan is to repeat these associations with larger sample sizes.\n\nOverall, the results of this study show the utility of repurposing existing imaging data for the study of the impact of visceral adipose levels on health outcomes. These images, while collected in large numbers, are an incredible untapped resource for a wide range of research projects. This project also shows the importance of using automated phenotype extraction from imaging datasets. Manual segmentation of images would have been considerably more time-consuming and will not scale up to the future work with measures obtained from additional CT scans. With the ever-increasing number of medical images available within electronic health record systems, using these images in a high-throughput manner is a powerful resource for research. Through this approach, we can obtain large sample sizes of understudied quantitative measures and subphenotypes beyond clinical lab measures, health screening measures such as BMI, and diagnoses. These measures obtained from images may better reflect the complexity of diseases and comorbidities, opening the door to new discoveries of the impact of these measures on health and the relationship to the genetic architecture.\n\nThis project was funded in part under a grant from the Pennsylvania Department of Health (SAP \\# 4100070267).\n\nData Availability\n=================\n\nThe electronic health record and genetic data used to support the findings of this study have not been made available because of protecting patient privacy.\n\nDisclosure\n==========\n\nThe Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions.\n\nConflicts of Interest\n=====================\n\nThe authors declare that there are no conflicts of interest regarding the publication of this paper.\n\nAuthors\\' Contributions\n=======================\n\nElliot D. K. Cha and Yogasudha Veturi jointly contributed to this work.\n\nSupplementary Materials {#supplementary-material-1}\n=======================\n\n###### \n\nSupplementary Figure 1: the plot shows \u2212log(*p*-values) for all phenotypic associations between ICD-9-based diagnoses on the *y*-axis for both VAT and SAT after controlling for age and sex. Point size is indicative of the number of cases (500, 1000, and 1500), and the direction of the point, upwards or downwards, represents the direction of the beta estimate (positive or negative). Supplementary Figure 2: the plot shows \u2212log(*p*-values) for sex-stratified phenotypic associations, outside of the top results, of ICD-9-based diagnoses (*y*-axis) associated with SAT after controlling for age. (A) Results of female phenotypic associations of ICD-9 codes with SAT, after controlling for age, which show a stronger relationship for females compared to males. (B) Results of odds ratios and 95% CIs of female phenotypic associations of ICD-9 codes with SAT after controlling for age. (C) Results of male phenotypic associations of ICD-9 codes with SAT, after controlling for age, which show a stronger relationship for males compared to females. (D) Results of odds ratios and 95% CIs of male phenotypic associations of ICD-9 codes with SAT after controlling for age. Supplementary Figure 3: the plot shows \u2212log(*p*-values) for sex-stratified phenotypic associations, outside of the top results, of ICD-9-based diagnoses (*y*-axis) associated with VAT after controlling for age. (A) Results of female phenotypic associations of ICD-9 codes with VAT, after controlling for age, which show a stronger relationship for females compared to males. (B) Results of odds ratios and 95% CIs of female phenotypic associations of ICD-9 codes with VAT after controlling for age. (C) Results of male phenotypic associations of ICD-9 codes with VAT, after controlling for age, which show a stronger relationship for males compared to females. (D) Results of odds ratios and 95% CIs of male phenotypic associations of ICD-9 codes with VAT after controlling for age. Supplementary Figure 4: Manhattan plots showing results for sex-combined analyses of genotypic associations of common variant SNPs with all adipose measures after controlling for age and sex. (A) Manhattan plot for genome-wide association tests for VAT. (B) Manhattan plot for genome-wide association tests for SAT. (C) Manhattan plot for genome-wide association tests for VSR. (D) Manhattan plot for genome-wide association tests for VTR. Blue line denotes a *p*-value of 1\u2009\u00d7\u200910^\u22125^. Red line indicates a GWAS significance of 5.0\u2009\u00d7\u200910^\u22128^. Supplementary Table 1: summary statistics of adipose measures. Supplementary Table 2: ICD-9 summary and descriptions. Supplementary Table 3: summary statistics of all EHR data. Supplementary Table 4: cross-phenotype associations of ICD-9 diagnostic codes. Supplementary Table 5: cross-phenotype associations of clinical lab measures. Supplementary Table 6: summary statistics of clinical laboratory measures. Supplementary Table 7: summary statistics of clinical laboratory measures stratified by sex. Supplementary Table 8: sex-stratified association tests for ICD-9 codes. Supplementary Table 9: sex-stratified association testing for clinical labs. Supplementary Table 10: top GWAS findings.\n\n###### \n\nClick here for additional data file.\n\n![(a) The plot shows \u2212log(*p* values) for the top phenotypic associations between ICD-9-based diagnoses on the *y*-axis for VAT and SAT after controlling for age and sex. Point size indicates the number of cases (approximately 400, 800, 1200, and 1600). The direction of the point, upwards or downwards, represents the direction of the corresponding beta estimate (positive or negative) (b) Odds ratios and 95% confidence intervals (CIs) of top phenotypic associations of ICD-9 codes with VAT and SAT after controlling for age and sex. In both panels, SAT is represented in orange and VAT in blue. The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold. Refer to the full descriptions of the ICD-9 code abbreviations in Supplementary [Table 2](#supplementary-material-1){ref-type=\"supplementary-material\"}.](JOBE2018-3253096.001){#fig1}\n\n![(a) The plot shows \u2212log(*p* values) for phenotypic associations of obesity-related ICD-9-based diagnoses with VAT and SAT on the *y*-axis after controlling for sex and age. Point size refers to the number of cases (250, 500, 750, and 1000). The direction of the point, upwards or downwards, represents the direction of the corresponding beta estimate (positive or negative) (b) Odds ratios and 95% CIs of phenotypic associations of obesity-related ICD-9 codes with VAT and SAT after controlling for sex and age. In both panels, SAT is represented in orange and VAT in blue. The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold.](JOBE2018-3253096.002){#fig2}\n\n![This plot shows \u2212log(*p* values) for adipose measures (*y*-axis: BMI in red, fat ratio in yellow, SAT in light blue, VAT in blue, and VSR in green) associated with obesity-related clinical laboratory measures (*x*-axis: natural log-transformed low-density lipoprotein (LDL), cholesterol (CHOL) levels, high-density lipoprotein (HDL), triglycerides (TRIG), and log-transformed white blood cell (WBC) counts after controlling for age and sex). The direction of the point, upwards or downwards, represents the direction of the corresponding beta estimate (positive or negative). The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold.](JOBE2018-3253096.003){#fig3}\n\n![The plot shows \u2212log(*p* values) for top sex-stratified phenotypic associations of ICD-9-based diagnoses with SAT on the *y*-axis after controlling for age. (a) Top female phenotypic associations of ICD-9 codes with SAT. Point size refers to the number of cases (250, 500, and 750). (b) Odds ratios and 95% CIs of top female phenotypic associations of ICD-9 codes with SAT after controlling for age. (c) Top male phenotypic associations of ICD-9 codes with SAT. Point size refers to the number of cases (100, 200, 300, 400, and 500). (d) Odds ratios and 95% CIs of top male phenotypic associations of ICD-9 codes with SAT after controlling for age. In both panels, females are represented in purple and males in green. The direction of the point, upwards or downwards, represents the direction of the corresponding beta estimate (positive or negative). The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold.](JOBE2018-3253096.004){#fig4}\n\n![The plot shows \u2212log(*p* values) for top sex-stratified phenotypic associations of ICD-9-based diagnoses with VAT on the *y*-axis after controlling for age. (a) Top female phenotypic associations of ICD-9 codes with VAT. Point size refers to the number of cases (250, 500, and 750). (b) Odds ratios and 95% CIs of top female phenotypic associations of ICD-9 codes with VAT after controlling for age. (c) Top male phenotypic associations of ICD-9 codes with VAT. Point size refers to the number of cases (100, 200, and 300). (d) Odds ratios and 95% CIs of top male phenotypic associations of ICD-9 codes with VAT after controlling for age. In all panels, females are represented in purple and males in green. The direction of the point, upwards or downwards, represents the direction of the corresponding beta estimate (positive or negative). The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold.](JOBE2018-3253096.005){#fig5}\n\n![(a) This plot shows \u2212log(*p* values) for adipose measures for females (*y*-axis: BMI in red, fat ratio in yellow, SAT in light blue, VAT in blue, and VSR in green) associated with obesity-related clinical laboratory measures (*x*-axis: natural log-transformed LDL, CHOL, HDL, TRIG, and log-transformed WBC) after controlling for age. (b) This plot shows \u2212log(*p* values) for adipose measures for males (*y*-axis: BMI in red, fat ratio in yellow, SAT in light blue, VAT in blue, and VSR in green) associated with obesity-related clinical laboratory measures (*x*-axis: natural log-transformed LDL, CHOL, HDL, TRIG, and log-transformed WBC) after controlling for age. The direction of the point (facing upwards or downwards) in both panels represents the direction of the corresponding beta estimate (positive or negative). The red line corresponds to the Bonferroni threshold, whereas the green line corresponds to the 1% FDR threshold.](JOBE2018-3253096.006){#fig6}\n\n###### \n\nDemographics and descriptive information of the cohort.\n\n  Covariate       Value    Count or (min, median, max)\n  --------------- -------- -----------------------------\n  Sex             Female   1307\n  Male            1238     \n  Diabetes        0        1796\n  1               749      \n  Age             \\<18     3\n  19--40          402      \n  41--60          919      \n  61--80          1046     \n  80+             175      \n  Weight (kg)     ---      (36.27, 88.45, 188.69)\n  Height (m)      ---      (1.09, 1.69, 2.52)\n  BMI (kg/m^2^)   ---      (12.56, 30.72, 68.48)\n\n[^1]: Academic Editor: Sharon Herring\n"}
+{"text": "1. Introduction {#s0005}\n===============\n\nAlzheimer\\'s disease (AD) is an age-related neurodegenerative disorder that has increased in prevalence as people live longer. It is estimated that the prevalence of AD may reach \\>\u00a074 million worldwide by 2030 ([@bb0165]). The cause and mechanism of AD is not fully elucidated, while progressive deposition of amyloid-\u03b2 and Tau protein is considered to be a neuropathological hallmark of AD. On the other hand, the close connection between plasmalogens (Pls) and AD has been indicated by the observations of decreased phosphatidyl ethanolamine Pls (PlsPE) in the affected brain regions of AD patients, such as the hippocampus and frontal cortex ([@bb0020], [@bb0035], [@bb0040]). Decreased levels of PlsPE in the blood and cerebrospinal fluid of AD patients have been reported ([@bb0025], [@bb0155], [@bb0160], [@bb0120], [@bb0170]). However, it is not clear whether the decrease of Pls in brain tissue and in plasma is the cause of the disease or merely a result of the disease. Recent studies of animal models of AD by our group indicated that intraperitoneal administration of purified Pls improved cognitive function ([@bb0065], [@bb0055], [@bb0060]).\n\nPls are a special class of glycerophospholipids characterized by a vinyl ether bond at the sn-1 position of glycerol backbone, and they are sometimes called plasmenyl phospholipid or alkenyl acryl phospholipid. They are found in almost all mammalian tissues and constitute about 18--20% of the total phospholipids in cell membranes. Predominant Pls in mammalian tissues are PlsPE and choline plasmalogen. PlsPE is much more abundant than choline plasmalogen except in heart and skeletal muscle. It is reported that Pls are abundant in the brain, retina, leukocytes (immune cells), sperm, heart, and skeletal muscle in mammals. This characteristic distribution of Pls indicates the importance of Pls in mammals ([@bb0015], [@bb0005]).\n\nPls are not only a structural component of animal cell membranes and reservoir for secondary messengers, but may also be involved in membrane fusion, ion transport, and cholesterol efflux, and act as antioxidants in cell membranes. Recently, the inhibitory effects of Pls on \u03b3-secretase activity have been reported ([@bb0125]). This study aimed to assess whether oral administration of Pls extracted from scallops would improve the cognitive function in patients with mild AD and mild cognitive impairment (MCI).\n\n2. Methods {#s0010}\n==========\n\n2.1. Study Design and Participants {#s0015}\n----------------------------------\n\nThis study was a multicenter, randomized, double-blind, placebo-controlled trial to evaluate the memory-improving efficacy of scallop-derived Pls in patients with mild AD and MCI. The study period consisted of a 24-week administration period and a four-week post-treatment period without administration (28\u00a0weeks in total).\n\nPatients were those who had 20 to 27 points of Mini Mental State Examination Japanese Version (MMSE-J) scores, i.e., patients with mild AD (20\u00a0\u2264\u00a0MMSE-J\u00a0\u2264\u00a023) or those with MCI (24\u00a0\u2264\u00a0MMSE-J\u00a0\u2264\u00a027) ([@bb0150]). All patients were required to meet the criteria for mild AD or MCI set out in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V). Patients were confirmed to have no arteriosclerotic dementia with MRI or CT scans performed within the previous six months. Patients were also ensured to have five or less points of Geriatric Depression Scale Short Version in Japanese (GDS-S-J) in order to exclude depressive pseudodementia. Patients taking anti-Alzheimer drugs had no change in the regimen during the previous three months. All caregivers were required to accompany patients to all visits throughout the study, and provided information on patients\\' daily life and health status. Patients were excluded if they were allergic to scallops, the raw materials for test substance. Written informed consent was obtained from either patients or their caregivers. The study protocol was approved by the Institutional Review Boards of Fukuoka University Hospital (Fukuoka), Nihonbashi Sakura Clinic (Tokyo), and BOOCS Clinic Fukuoka (Fukuoka). The study was implemented in compliance with the Declaration of Helsinki.\n\n2.2. Randomization and Masking {#s0020}\n------------------------------\n\nA computer-generated random allocation list was created by an expert at CAC Croit Corporation (Tokyo) based on the blocked randomization method with each block consisting of two placebo allocations and two Pls allocations. Each study site was provided with test substance kits numbered according to the allocation list, which were handed directly to patients. At some study sites without cold storage space, patients received test kits by courier every month. Thus enrolled patients were randomly assigned to receive either 1.0\u00a0mg\u00a0Pls/day or placebo. Test substance was jelly-like substance. Active substance and placebo were identical in appearance and taste. Patients, caregivers, study physicians, and clinical staff were masked to treatment allocation throughout the study period.\n\n2.3. Procedures {#s0025}\n---------------\n\nEnrolled patients were provided with the test substance at baseline visit or within one week after the baseline visit, and were instructed to take it orally twice/day for 24\u00a0weeks. They received one month\\'s extra test substance in case they missed the next scheduled visit. To confirm compliance, they were requested to return unconsumed test substance at their next visit. They were furthermore urged not to change the regimen during the study period as far as possible. If a patient changed his or her drug use, we terminated his or her observation at the point. We recorded any complications and adverse events reported by patients at each visit.\n\nThe primary outcome measure was MMSE-J. The secondary outcome measures included Wechsler Memory Scale-Revised (WMS-R), GDS-S-J, plasma PlsPE levels, and relative concentration of PlsPE in erythrocyte membrane, namely the percentage of Pls to the total phospholipids in erythrocyte membrane. Cognitive function was assessed at baseline, and at weeks 12, 24, and 28. Blood was drawn from patients in the fasting state at baseline, and at weeks 8, 16, 20, 24, and 28 for measuring erythrocyte PlsPE and plasma PlsPE. PlsPE measurement was performed using the previously reported method ([@bb0090], [@bb0100]).\n\nSafety assessment was conducted by recording adverse events and performing a physical examination and biochemical blood tests such as liver function, renal function, blood sugar, and lipid levels at each visit.\n\n2.4. Statistical Analysis {#s0030}\n-------------------------\n\nTo determine the sample size, we assumed that the MMSE-J score would improve by 5% in the placebo group and by 10% in the Pls treatment with SD of 15% in each. With a statistical power of 0.80 and a one-sided significance level of 0.01, the required sample size was estimated to be 181 in each group. The target number of enrollments was decided to be 200 in each with allowance for some extent of dropout.\n\nThe between-treatment difference was assessed by unpaired *t*-test, and the within-group change from the baseline was evaluated by paired *t*-test. In terms of after-treatment outcome of the cognitive function, the mean change from the baseline and 95% confidence interval (CI) were presented. Analyses were done with Stata version 13 (StataCorp, College Station, TX). The trial is registered at the University Hospital Medical Information Network as ID UMIN000014945.\n\n2.5. Role of the Funding Source {#s0035}\n-------------------------------\n\nB&S Corporation Co. Ltd. (Tokyo), one of the donors to The Japanese Plasmalogen Society, was involved in provision of Pls test substance and placebo and delivery to study sites, but it had no involvement in study design and planning, investigator training, data analysis, interpretation, and writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. All co-investigators also had full access to the data.\n\n3. Results {#s0040}\n==========\n\n3.1. Study Participants {#s0045}\n-----------------------\n\nA total of 328 patients were enrolled at 25 hospitals or clinics in Kyushu, Kanto, and Kansai regions in Japan from November 15, 2014 to October 8, 2015. They were randomly assigned to either of the two treatment groups (166 to Pls group and 162 to placebo group). Of the 328 enrolled patients, 285 continued to participate in the study for 12\u00a0weeks, and 276 completed the 24\u00a0week study period. Those who completed the study included 11 patients (4.0%) who failed to use 80% or more of the provided test substance. [Fig. 1](#f0005){ref-type=\"fig\"} shows the trial profile.\n\nNo significant difference was noted between the groups with respect to age, MMSE-J, GDS-S-J, WMS-R (0\u00a0min), WMS-R (30\u00a0min), erythrocyte PlsPE, and plasma PlsPE, at baseline, although the male-to-female ratio was significantly lower in the placebo group ([Table 1](#t0005){ref-type=\"table\"}).\n\n3.2. Clinical Efficacy {#s0050}\n----------------------\n\n### 3.2.1. Primary Analysis {#s0055}\n\nIn an intention-to-treat analysis, there was no significant difference between the treatment and placebo groups in the primary and secondary outcomes of cognitive function. The MMSE-J score showed a nearly significant improvement in the treatment group while no such improvement was observed in the placebo group, resulting no statistically significant between-group difference. The WMS-R (0\u00a0min) and WMS-R (30\u00a0min) score each showed a significant improvement in both groups, but there was no statistically significant between-group difference. Erythrocyte PlsPE increased and plasma PlsPE decreased to different degrees after treatment in both Pls and placebo groups. There was no statistically significant between-group difference at any points of time, while the decrease in plasma PlsPE seemed greater in the placebo group ([Table 2](#t0010){ref-type=\"table\"}).\n\n### 3.2.2. Subgroup Analysis {#s0060}\n\nWe examined the change in cognitive function in mild AD patients and MCI patients separately. No notable change was seen in MMSE-J in either group. WMS-R (0\u00a0min) and WMS-R (30\u00a0min) improved statistically significantly in the treatment group, and the improvement seemed greater in the Pls treatment group with respect to the 0\u00a0min value (P\u00a0=\u00a00.067) and the 30-min value (P\u00a0=\u00a00.078) ([Table 3](#t0015){ref-type=\"table\"}).\n\n### 3.2.3. Subgroup Analysis by Gender and Age {#s0065}\n\nFurthermore, we analyzed the cognitive function in patients with mild AD patients by gender and age ([Table 4](#t0020){ref-type=\"table\"}). Mild AD patients were divided into two groups based on the median value, 77\u00a0years or younger and 78\u00a0years or older. In the patients aged 77\u00a0years or younger, WMS-R (30\u00a0min) increased statistically significantly in the treatment group, showing a significant between-group difference (P\u00a0=\u00a00.029). In female patients with mild AD, WMS-R (30\u00a0min) improved significantly in the treatment group, with a significant between-group difference (P\u00a0=\u00a00.017). There was no notable between-group difference in the analysis of MCI patients by gender and age.\n\n### 3.2.4. Changes in Blood Plasmalogens {#s0070}\n\nThe changes in blood PlsPE levels among mild AD patients were examined ([Table 5](#t0025){ref-type=\"table\"}). Plasma PlsPE decreased in the placebo group, showing a significant between-group difference (P\u00a0=\u00a00.016). There was a statistically significant elevation of erythrocyte PlsPE at week 16 in the treatment group and at week 24 in placebo. However, the changes were not significantly different between the two groups.\n\n3.3. Clinical Safety {#s0075}\n--------------------\n\n[Table 6](#t0030){ref-type=\"table\"} summarizes adverse events. There was no notable difference in the occurrence of adverse events between the two groups.\n\n4. Discussion {#s0080}\n=============\n\nTo our knowledge, this is the unprecedented trial that addressed efficacy on cognitive function and blood Pls changes by oral administration of Pls in patients with mild AD and MCI. The present results showed no significant difference in the primary outcome (MMSE-J score) between the treatment and placebo groups. However, among patients with mild AD, WMS-R score significantly improved in the treatment group, and the improvement seemed greater than in the placebo group. In a subgroup analysis of mild AD patients, improvement of WMS-R was more evident in female patients and in patients below 77\u00a0years. These results indicate that oral administration of scallop-derived purified Pls may be effective to improve memory function of patients with mild AD. This study did not find the efficacy of Pls in mild AD patients over 78\u00a0years or male AD patients. The reason for the lack of efficacy in these aged patients may be ascribed to age-related irreversible degenerative changes in the brain ([@bb0070]). It is unknown why the efficacy was evident in females but not in males.\n\nThe present study did not find the efficacy of Pls in MCI patients. It should be noted that WMS-R improved significantly even at end of the study (week 24) in the placebo group as well. These findings suggest that MCI patients may have exhibited a stronger placebo effect than AD patients. MCI patients may retain higher brain functions such as \"expectation\" and \"hope\". AD patients are seemingly less affected by placebo effects because their cognitive function declines and the above-mentioned higher brain function diminishes more markedly than that of MCI patients. This line of conjecture indicates that the mental status of people can change Pls concentration in brain tissue.\n\nOur previous studies showed that Pls in erythrocyte membrane significantly increased with orally ingested Pls, and another study also revealed that plasma PlsPE increased shortly after being ingested orally ([@bb0095], [@bb0115]). However, in the present study, only 1.0\u00a0mg/day of purified Pls showed an effect on memory function of mild AD patients. The physiological mechanism of this effect with such a small amount of Pls by oral administration is unclear. One hypothesis is that Pls may work through some receptors like hormones. Lipid rafts of cell membrane are considered to be associated with cell signaling ([@bb0060], [@bb0110]). There are reports that lipid rafts are rich in PlsPE ([@bb0135]). G-protein coupled receptors (GPCR) are also localized to lipid rafts and caveolae ([@bb0010]). These may indicate the possibility that Pls work as a ligand of GPCR. Concentration of PlsPE in human plasma is about 100\u00a0\u03bcmol/l, but Pls in plasma may circulate as lipoproteins. Pls in the lipoproteins themselves are not likely to work as a ligand of receptors. However, it is possible that free Pls, which are derived from oral administration, even in a small amount, may work as a ligand of some receptors at intestinal cells before becoming lipoproteins. It is well known that there is close communication between the intestine and brain through neural, endocrine, and immune pathways ([@bb0175]).\n\nOn the other hand, there have been many reports that docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) relate closely to brain functions ([@bb0030], [@bb0050], [@bb0145]). Scallop-derived Pls used in the present study presumably contained relatively large amounts of DHA and EPA ([@bb0075], [@bb0045]). It is possible that these omega-3 polyunsaturated fatty acids of Pls may be effective for improvement of cognitive function in mild AD. Some studies suggest that DHA in the form of phospholipids passes through blood-brain barrier approximately ten times more efficiently than in the form of free fatty acid ([@bb0080], [@bb0085]), although other studies report that a diet high in omega-3 fatty acids such as DHA and EPA resulted in negligible beneficial effect on cognition in AD patients ([@bb0140], [@bb0130]).\n\nRegarding the change of blood Pls levels after treatment, there were significant differences between the treatment and placebo groups in mild AD patients. Plasma PlsPE in the placebo group decreased significantly after treatment, whereas that of the treatment group remained unchanged. Some studies suggest that statins, a class of cholesterol-lowering drugs, raise blood level of PlsPE ([@bb0105]). In the present study, 23 (23.5%) of 98 patients with mild AD used statins during the study period. However, there were still significant differences in the changes of plasma PlsPE between the treatment and placebo groups except for statin-users. Therefore, it is unlikely that statins protected the decrease of Plasma PlsPE in the treatment group. These results might indicate that Pls production in peroxisome was concurrently reduced with the progression of Alzheimer\\'s disease during the 24-week treatment period, and suggest that orally administered Pls may contribute to maintain the production of Pls in peroxisome.\n\nThe present study has several problems and limitations. The study duration may be too short to detect the efficacy of Pls in MCI patients with MMSE-J score over 24, whose condition was ill-defined and may have little differed from that of healthy individuals. To resolve this problem, we should conduct additional research to ascertain whether it is possible to delay progression from MCI to AD in a long-term follow-up of MCI patients in the treatment group. Further randomized controlled trials for moderate and severe AD patients are also required.\n\nFunding Sources {#s0085}\n===============\n\nThe study was funded by The Japanese Plasmalogen Society (Pls2014-01) (Fukuoka, Japan) as an investigator-initiated trial.\n\nConflicts of Interest {#s0090}\n=====================\n\nTF and SM have applied for patents on method for manufacturing ether phospholipid (patent application number: PCT/JP2015/63617, PCT/JP2015/63740). TY, TA, YT, CW, and SK declare that they have no conflicts of interest.\n\nAuthor Contributions {#s0095}\n====================\n\nTF was responsible for central management of the trial including literature search, study design, data collection, data interpretation and writing of the manuscript. TY, TA, and YT contributed to study design, recruitment and research monitoring. CW contributed to clinical management, data collection and writing of the manuscript. SM contributed to implementation of the study with laboratory measurements and writing of the manuscript. SK did study design, statistical analysis of the data and writing of the manuscript. All authors reviewed the manuscript and approved the final manuscript.\n\nWe thank all patients and caregivers who took part in the study. The following contributed to implementation of the trial as study physicians: E Nishikawa (Nishikawa-naika Clinic, Shimonoseki), H Fujino (Fujino Clinic, Yanagawa), H Matsuo (Matsuo Hospital, Fukuoka), H Nawata (Muta Hospital, Fukuoka), K Fukuyama (Hasami Hospital, Higashisonogi), K Irie (Hakujyuji Hospital, Fukuoka), K Saito, N Shinfuku (BOOCS Clinic Fukuoka, Fukuoka), K Serikawa (Monowasure Mental Clinic, Fukuoka), K Takasaki (Takasaki Neurosurgery Clinic, Kasuya), M Ichimaru (BOOCS Holistic Clinic Tokyo, Tokyo), M Kinoshita (Nagata Hospital, Yanagawa), M Munaka (Hayama Clinic, Munakata), N Araki (Utsunomiya Rehabilitation Hospital, Utsunomiya), S Nakamura (Sangenjaya Nakamura Mental Clinic, Tokyo), S Ouma (Fukuoka University Hospital, Fukuoka), S Nakano (Nihonbashi Sakura Clinic, Tokyo), T Asada (Olive Clinic Ochanomizu, Tokyo, Memory Clinic Ochanomizu, Tokyo), T Kaneko (Kaneko Hospital, Yanagawa), T Kinoshita (Nozomi Memory Clinic, Mitaka), T Yamada (Nishino Hospital, Kitakyushu, Tokorozawa Meisei Hospital, Tokorozawa), T Yosimatsu (Mito Hospital, Kasuya), Y Nakayama (Nakayoshi-Clinic, Kasuya), Y Sekine (Sekine Clinic, Hirakata).\n\n![Trial profile.](gr1){#f0005}\n\n###### \n\nBaseline characteristics.\n\nTable 1\n\n  Variable            Pls group (n\u00a0=\u00a0145)   Placebo group (n\u00a0=\u00a0140)                        P value[a](#tf0005){ref-type=\"table-fn\"}\n  ------------------- --------------------- ---------------------------------------------- ------------------------------------------\n  Male, n (%)         65 (44.8)             45 (32.1)                                      0.03\n  MCI, n (%)          62 (42.8)             59 (42.1)                                      0.92\n  Age in year         76.4 (6.0)            76.5 (5.6)                                     0.94\n  MMSE-J              24.0 (2.4)            24.2 (2.2)                                     0.38\n  GDS-S-J             2.32 (1.54)           2.36 (1.55)                                    0.83\n  WMS-R (0\u00a0min)       4.21 (3.71)           4.34 (4.03)[b](#tf0010){ref-type=\"table-fn\"}   0.78\n  WMS-R (30\u00a0min)      2.50 (3.73)           2.77 (4.07)[b](#tf0010){ref-type=\"table-fn\"}   0.56\n  Erythrocyte PlsPE   8.03 (1.05)           8.11 (0.96)                                    0.49\n  Plasma PlsPE        3.65 (1.31)           3.93 (1.28)                                    0.08\n\nValues are mean (SD) unless otherwise specified.\n\nMMSE-J\u00a0=\u00a0Mini Mental State Examination-Japanese, GDS-S-J\u00a0=\u00a0Geriatric Depression Scale-Short Version-Japanese, WMS-R\u00a0=\u00a0Wechsler Memory Scale-Revised, min\u00a0=\u00a0minute, PlsPE\u00a0=\u00a0phosphatidyl ethanolamine plasmalogens.\n\nChi-square test for proportion and unpaired *t*-test for mean.\n\nNumber of the patients was 139.\n\n###### \n\nMean difference from the baseline.\n\nTable 2\n\n  Variable            Week   Pls group   Placebo group      P value                                                           \n  ------------------- ------ ----------- ------------------ ------------------ -------- ------------------ ------------------ ------\n  MMSE-J              12     145         0.43               (0.04; 0.80)       140      0.33               (\u2212\u00a00.08; 0.73)     0.73\n  24                  140    0.40        (\u2212\u00a00.01; 0.81)     136                0.32     (\u2212\u00a00.14; 0.77)     0.79               \n  28                  81     0.69        (0.12; 1.26)       79                 0.44     (\u2212\u00a00.14; 1.03)     0.54               \n  GDS-S-J             12     145         0.37               (0.05; 0.68)       140      0.39               (0.06; 0.73)       0.91\n  24                  140    0.20        (\u2212\u00a00.12; 0.52)     136                0.18     (\u2212\u00a00.16; 0.51)     0.92               \n  28                  81     0.17        (\u2212\u00a00.32; 0.66)     80                 0.18     (\u2212\u00a00.34; 0.69)     1.00               \n  WMS-R (0\u00a0min)       12     145         0.66               (0.21; 1.20)       139      0.86               (0.36; 1.36)       0.55\n  24                  140    1.41        (0.91; 1.92)       135                1.39     (0.87; 1.92)       0.95               \n  28                  80     2.30        (1.48; 3.12)       78                 2.46     (1.67; 3.25)       0.78               \n  WMS-R (30\u00a0min)      12     145         0.48               (0.11; 0.83)       139      0.35               (\u2212\u00a00.05; 0.75)     0.65\n  24                  140    1.09        (0.59; 1.59)       135                1.10     (0.61; 1.59)       0.99               \n  28                  80     2.09        (1.19; 2.98)       78                 1.92     (1.19; 2.66)       0.78               \n  Erythrocyte PlsPE   8      145         0.13               (\u2212\u00a00.03; 0.30)     139      0.08               (\u2212\u00a00.09; 0.24)     0.64\n  16                  140    0.24        (0.10; 0.37)       138                0.22     (0.06; 0.37)       0.86               \n  24                  139    0.26        (0.10; 0.41)       135                0.35     (0.19; 0.52)       0.40               \n  28                  139    0.27        (0.12; 0.42)       133                0.26     (0.09; 0.43)       0.95               \n  Plasma PlsPE        8      116         \u2212\u00a00.48             (\u2212\u00a00.70; \u2212\u00a00.27)   116      \u2212\u00a00.77             (\u2212\u00a00.98; \u2212\u00a00.56)   0.06\n  16                  122    \u2212\u00a00.59      (\u2212\u00a00.80; \u2212\u00a00.39)   118                \u2212\u00a00.72   (\u2212\u00a00.93; \u2212\u00a00.50)   0.42               \n  24                  140    \u2212\u00a00.24      (\u2212\u00a00.42; \u2212\u00a00.06)   136                \u2212\u00a00.38   (\u2212\u00a00.60; \u2212\u00a00.17)   0.31               \n  28                  138    \u2212\u00a00.59      (\u2212\u00a00.77; \u2212\u00a00.40)   131                \u2212\u00a00.63   (\u2212\u00a00.81; \u2212\u00a00.44)   0.77               \n\nMMSE-J\u00a0=\u00a0Mini Mental State Examination-Japanese, GDS-S-J\u00a0=\u00a0Geriatric Depression Scale-Short Version-Japanese, WMS-R\u00a0=\u00a0Wechsler Memory Scale-Revised, min\u00a0=\u00a0minute, PlsPE\u00a0=\u00a0phosphatidyl ethanolamine plasmalogens.\n\n###### \n\nMean difference from the baseline, in patients with MMSE-J score 20 to 23.\n\nTable 3\n\n  Variable         Week   Pls group   Placebo group    P value                                                   \n  ---------------- ------ ----------- ---------------- ---------------- ------ ---------------- ---------------- ------\n  MMSE-J           12     55          0.31             (\u2212\u00a00.44; 1.06)   49     0.35             (\u2212\u00a00.43; 1.12)   0.94\n  24               50     0.06        (\u2212\u00a00.76; 0.88)   48               0.19   (\u2212\u00a00.59; 0.96)   0.82             \n  28               25     \u2212\u00a00.12      (\u2212\u00a01.48; 1.24)   25               0.36   (\u2212\u00a00.78; 1.50)   0.58             \n  GDS-S-J          12     55          0.33             (\u2212\u00a00.28; 0.94)   49     0.67             (0.02; 1.32)     0.44\n  24               50     0.20        (\u2212\u00a00.45; 0.85)   48               0.25   (\u2212\u00a00.28; 0.78)   0.91             \n  28               25     0.72        (\u2212\u00a00.66; 2.10)   25               0.18   (\u2212\u00a00.71; 1.03)   0.48             \n  WMS-R (0\u00a0min)    12     55          0.53             (\u2212\u00a00.13; 1.19)   48     0.73             (0.12; 1.34)     0.66\n  24               50     1.24        (0.49; 1.99)     47               0.30   (\u2212\u00a00.39; 0.99)   0.067            \n  28               24     1.29        (0.04; 2.55)     24               1.08   (0.46; 1.70)     0.76             \n  WMS-R (30\u00a0min)   12     55          0.47             (\u2212\u00a00.02; 0.96)   48     0.13             (\u2212\u00a00.25; 0.50)   0.27\n  24               50     0.78        (0.04; 1.52)     47               0.02   (\u2212\u00a00.37; 0.41)   0.078            \n  28               24     0.96        (\u2212\u00a00.33; 2.24)   24               0.29   (\u2212\u00a00.26; 0.84)   0.33             \n\nMMSE-J\u00a0=\u00a0Mini Mental State Examination-Japanese, GDS-S-J\u00a0=\u00a0Geriatric Depression Scale-Short Version-Japanese, WMS-R\u00a0=\u00a0Wechsler Memory Scale-Revised, min\u00a0=\u00a0minute.\n\n###### \n\nMean difference from the baseline, by gender and age, in patients with MMSE-J score 20 to 23.\n\nTable 4\n\n  Gender/age       Week            Pls group   Placebo group   P value                                                                      \n  ---------------- --------------- ----------- --------------- ---------------- ---------------- -------- ---------------- ---------------- ------\n  Female           WMS-R (0\u00a0min)   12          27              1.04             (0.02; 2.05)     30       0.80             (0.04; 1.56)     0.70\n  WMS-R (30\u00a0min)   12              27          0.93            (0.10; 1.75)     30               0.17     (\u2212\u00a00.37; 0.70)   0.11             \n  WMS-R (0\u00a0min)    24              25          1.52            (0.35; 2.69)     29               0.48     (\u2212\u00a00.47; 1.43)   0.16             \n  WMS-R (30\u00a0min)   24              25          1.08            (0.06; 2.10)     29               \u2212\u00a00.21   (\u2212\u00a00.68; 0.26)   0.02             \n  Male             WMS-R (0\u00a0min)   12          28              0.04             (\u2212\u00a00.84; 0.92)   18       0.61             (\u2212\u00a00.53; 1.76)   0.41\n  WMS-R (30\u00a0min)   12              28          0.04            (\u2212\u00a00.51; 0.58)   18               0.06     (\u2212\u00a00.47; 0.58)   0.96             \n  WMS-R (0\u00a0min)    24              25          0.96            (\u2212\u00a00.06; 1.98)   18               0.00     (\u2212\u00a01.05; 1.05)   0.19             \n  WMS-R (30\u00a0min)   24              25          0.48            (\u2212\u00a00.66; 1.62)   18               0.39     (\u2212\u00a00.32; 1.10)   0.90             \n  77 or younger    WMS-R (0\u00a0min)   12          20              1.30             (\u2212\u00a00.14; 2.74)   21       1.19             (0.37; 2.01)     0.89\n  WMS-R (30\u00a0min)   12              20          1.15            (0.05; 2.25)     21               0.10     (\u2212\u00a00.22; 0.41)   0.06             \n  WMS-R (0\u00a0min)    24              19          2.21            (0.61; 3.81)     21               0.48     (\u2212\u00a00.52; 1.47)   0.06             \n  WMS-R (30\u00a0min)   24              19          1.84            (0.10; 3.58)     21               0.00     (\u2212\u00a00.41; 0.41)   0.03             \n  78 or older      WMS-R (0\u00a0min)   12          35              0.09             (\u2212\u00a00.57; 0.74)   27       0.37             (\u2212\u00a00.54; 1.28)   0.60\n  WMS-R (30\u00a0min)   12              35          0.09            (\u2212\u00a00.37; 0.54)   27               0.15     (\u2212\u00a00.50; 0.79)   0.87             \n  WMS-R (0\u00a0min)    24              31          0.65            (\u2212\u00a00.07; 1.36)   26               0.15     (\u2212\u00a00.86; 1.17)   0.41             \n  WMS-R (30\u00a0min)   24              31          0.13            (\u2212\u00a00.42; 0.68)   26               0.04     (\u2212\u00a00.61; 0.69)   0.83             \n\nWMS-R\u00a0=\u00a0Wechsler Memory Scale-Revised, min\u00a0=\u00a0minute.\n\n###### \n\nMean difference from the baseline, in patients with MMSE-J score 20 to 23.\n\nTable 5\n\n  Variable            Week   Pls group   Placebo group      P value                                                         \n  ------------------- ------ ----------- ------------------ ---------------- -------- ------------------ ------------------ ------\n  Erythrocyte PlsPE   8      55          0.16               (\u2212\u00a00.13; 0.46)   50       0.28               (\u2212\u00a00.03; 0.48)     0.74\n  16                  53     0.33        (0.11; 0.57)       48               0.19     (\u2212\u00a00.07; 0.44)     0.39               \n  24                  49     0.24        (\u2212\u00a00.03; 0.52)     48               0.40     (0.11; 0.69)       0.42               \n  28                  49     0.22        (\u2212\u00a00.02; 0.46)     46               0.26     (\u2212\u00a00.02; 0.54)     0.85               \n  Plasma PlsPE        8      41          \u2212\u00a00.21             (\u2212\u00a00.50; 0.08)   39       \u2212\u00a00.64             (\u2212\u00a01.01; \u2212\u00a00.26)   0.05\n  16                  43     \u2212\u00a00.39      (\u2212\u00a00.68; \u2212\u00a00.11)   40               \u2212\u00a00.86   (\u2212\u00a01.29; \u2212\u00a00.43)   0.07               \n  24                  50     0.16        (\u2212\u00a00.07; 0.40)     48               \u2212\u00a00.34   (\u2212\u00a00.69; 0.00)     0.02               \n  28                  49     \u2212\u00a00.35      (\u2212\u00a00.63; \u2212\u00a00.08)   46               \u2212\u00a00.85   (\u2212\u00a01.13; \u2212\u00a00.57)   0.01               \n\nPlsPE\u00a0=\u00a0phosphatidyl ethanolamine plasmalogens.\n\n###### \n\nList of reported adverse events.\n\nTable 6\n\n                                       Pls group (n\u00a0=\u00a0169)   Placebo group (n\u00a0=\u00a0167)\n  ------------------------------------ --------------------- -------------------------\n  Neoplasms                                                  \n  \u00a0Esophageal cancer                   0                     1\n  \u00a0Ovarian cancer                      0                     1\n  \u00a0Bladder cancer                      1                     0\n  Nervous system                                             \n  \u00a0Parkinsonism                        1                     0\n  \u00a0Transient ischaemic attack          0                     1\n  Eye                                                        \n  \u00a0Ophthalmecchymosis                  0                     1\n  Circulatory system                                         \n  \u00a0Hypertension                        2                     1\n  \u00a0Ischaemic heart disease             1                     0\n  \u00a0Premature ventricular contraction   0                     2\n  \u00a0Stroke                              1                     0\n  \u00a0Subarachnoid hemorrhage             1                     0\n  Respiratory system                                         \n  \u00a0Cold                                2                     3\n  \u00a0Aspiration pneumonia                0                     1\n  Digestive system                                           \n  \u00a0Hepatic hemangioma                  1                     0\n  \u00a0Stomatitis                          1                     1\n  \u00a0Gastric ulcer                       1                     1\n  \u00a0Intestinal obstruction              1                     0\n  \u00a0Acute hepatic disorder              1                     0\n  \u00a0Cholecystolithiasis                 1                     0\n  Skin and subcutaneous tissue                               \n  \u00a0Herpes zoster                       0                     1\n  \u00a0Oral herpes                         0                     2\n  \u00a0Prurigo                             0                     1\n  \u00a0Rash                                1                     1\n  Musculoskeletal system                                     \n  \u00a0Gouty attack                        1                     0\n  Genitourinary system                                       \n  \u00a0Renal failure                       1                     0\n  Symptoms and signs                                         \n  \u00a0Diarrhoea                           3                     1\n  \u00a0Akathisia                           1                     0\n  \u00a0Hand numbness                       0                     1\n  \u00a0Ear pain                            1                     0\n  \u00a0Sore throat                         1                     0\n  \u00a0Constipation                        2                     0\n  \u00a0Knee pain                           2                     0\n  \u00a0Shoulder pain                       0                     1\n  \u00a0Muscle pain                         1                     0\n  \u00a0Cough                               1                     1\n  \u00a0Vomiting                            0                     1\n  \u00a0Dizziness                           0                     3\n  \u00a0Stomach pain                        1                     1\n  \u00a0Oedema                              1                     3\n  Injury                                                     \n  \u00a0Contusion                           1                     1\n  \u00a0Fracture                            4                     4\n  \u00a0Traumatic subarachnoid hemorrhage   1                     0\n  \u00a0Meniscus injury                     1                     0\n"}
+{"text": "**Suggested citation:** EFSA FEEDAP Panel\u00a0(EFSA Panel\u00a0on Additives and Products or Substances used in Animal Feed) , Bampidis V, Azimonti G, Bastos ML, Christensen H, Dusemund B, Kouba M, Kos Durjava M, L\u00f3pez\u2010Alonso M, L\u00f3pez Puente S, Marcon F, Mayo B, Pechov\u00e1 A, Petkova M, Ramos F, Sanz Y, Villa RE, Woutersen R, Galobart J, Holczknecht O, Manini P, Pizzo F, Call JT and Anguita M, 2019 Scientific Opinion on the safety and efficacy of APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase) as a feed additive for piglets (suckling and weaned) and growing minor porcine species. EFSA Journal 2019;17(11):5894, 9 pp. 10.2903/j.efsa.2019.5894\n\n**Requestor:** European Commission\n\n**Question number:** EFSA\u2010Q\u20102019\u201000192\n\n**Panel\u00a0members:** Giovanna Azimonti, Vasileios Bampidis Maria de Lourdes Bastos, Henrik Christensen, Birgit Dusemund, Maryline Kouba, Mojca Kos Durjava, Marta L\u00f3pez\u2010Alonso, Secundino L\u00f3pez Puente, Francesca Marcon, Baltasar Mayo, Alena Pechov\u00e1, Mariana Petkova, Fernando Ramos, Yolanda Sanz, Roberto Edoardo Villa and Ruud Woutersen.\n\n**Acknowledgements:** The Panel\u00a0wishes to acknowledge the contribution of Kostas Sofianidis and Elisa Pettenati to this opinion.\n\n**Note:** Relevant information or parts of this scientific output have been blackened in accordance with the confidentiality requests formulated by the applicant pending a decision thereon by the European Commission. The full output has been shared with the European Commission, EU Member States and the applicant. The blackening will be subject to review once the decision on the confidentiality requests is adopted by the European Commission.\n\nAdopted: 7 October 2019\n\n1. Introduction {#efs25894-sec-0002}\n===============\n\n1.1. Background and Terms of Reference {#efs25894-sec-0003}\n--------------------------------------\n\nRegulation (EC) No 1831/2003[1](#efs25894-note-1006){ref-type=\"fn\"} establishes the rules governing the Community authorisation of additives for use in animal nutrition. In particular, Article 4(1) of that Regulation lays down that any person seeking authorisation for a feed additive or for a new use of a feed additive shall submit an application in accordance with Article 7.\n\nThe European Commission received a request from Andr\u00e9s Pintaluba S.A.[2](#efs25894-note-1007){ref-type=\"fn\"} for authorisation of the product APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase), when used as a feed additive for piglets (suckling and weaned) and minor growing porcine species (category: zootechnical additives; functional group: digestibility enhancers).\n\nAccording to Article 7(1) of Regulation (EC) No 1831/2003, the Commission forwarded the application to the European Food Safety Authority (EFSA) as an application under Article 4(1) (authorisation of a feed additive or new use of a feed additive). EFSA received directly from the applicant the technical dossier in support of this application. The particulars and documents in support of the application were considered valid by EFSA as of 3 May 2019.\n\nAccording to Article 8 of Regulation (EC) No 1831/2003, EFSA, after verifying the particulars and documents submitted by the applicant, shall undertake an assessment in order to determine whether the feed additive complies with the conditions laid down in Article 5. EFSA shall deliver an opinion on the safety for the target animals, consumer, user and the environment and on the efficacy of the product APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase), when used under the proposed conditions of use (see Section\u00a0[3.1](#efs25894-sec-0009){ref-type=\"sec\"}).\n\n1.2. Additional information {#efs25894-sec-0004}\n---------------------------\n\nThe FEEDAP Panel\u00a0adopted an opinion on the safety and efficacy of APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase) as a feed additive for chickens for fattening or reared for laying and minor poultry species for fattening or reared for laying (EFSA FEEDAP Panel, [2019](#efs25894-bib-0005){ref-type=\"ref\"}).\n\n2. Data and methodologies {#efs25894-sec-0005}\n=========================\n\n2.1. Data {#efs25894-sec-0006}\n---------\n\nThe present assessment is based on data submitted by the applicant in the form of a technical dossier[3](#efs25894-note-1008){ref-type=\"fn\"} in support of the authorisation request for the use of APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase) as a feed additive.\n\nThe FEEDAP Panel\u00a0used the data provided by the applicant together with data from other sources, such as previous risk assessments by EFSA to deliver the present output.\n\nThe European Union Reference Laboratory (EURL) considered that the conclusions and recommendations reached in the previous assessment regarding the methods used for the control of the APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase) in animal feed are valid and applicable for the current application.[4](#efs25894-note-1009){ref-type=\"fn\"}\n\n2.2. Methodologies {#efs25894-sec-0007}\n------------------\n\nThe approach followed by the FEEDAP Panel\u00a0to assess the safety and the efficacy of APSA PHYTAFEED^\u00ae^ 20,000 GR/L (6\u2010phytase) is in line with the principles laid down in Regulation (EC) No 429/2008[5](#efs25894-note-1010){ref-type=\"fn\"} and the relevant guidance documents: Guidance on the assessment of the safety of feed additives for the target species (EFSA FEEDAP Pane, [2017](#efs25894-bib-0003){ref-type=\"ref\"}) and Guidance on the assessment of the efficacy of feed additives (EFSA FEEDAP Panel, [2018](#efs25894-bib-0004){ref-type=\"ref\"}).\n\n3. Assessment {#efs25894-sec-0008}\n=============\n\nThe additive APSA PHYTAFEED^\u00ae^ 20,000 GR/L contains 6\u2010phytase activity (EC 3.1.3.26; phytase) and is intended to be used in feed for piglets (suckling and weaned) and growing minor porcine species as a zootechnical additive (functional group: digestibility enhancers).\n\n3.1. Characterisation {#efs25894-sec-0009}\n---------------------\n\nThe phytase present in the additive is produced by a genetically modified strain of the yeast *Komagataella phaffii* that has been deposited in the China General Microbiological Culture Collection Centre (CGMCC) with the deposit number 12056.[6](#efs25894-note-1011){ref-type=\"fn\"} The additive is available in two formulations, a solid one APSA PHYTAFEED^\u00ae^ 20,000 GR and a liquid one APSA PHYTAFEED^\u00ae^ 20,000 L. The two formulations of the additive ensure a guaranteed minimum phytase activity of 20,000 U[7](#efs25894-note-1012){ref-type=\"fn\"}/g or mL of product. In a previous opinion, the Panel\u00a0characterised the additive and its manufacturing process including the production strain (EFSA FEEDAP Panel, [2019](#efs25894-bib-0005){ref-type=\"ref\"}). The applicant has provided new data on the taxonomic classification of the production strain and on the stability and the capacity of the additive to homogeneously distribute when added to feed for piglets.\n\n\u25a0\u25a0\u25a0\u25a0\u25a0 *K. phaffii* is considered by EFSA to be suitable for the Qualified Presumption of Safety (QPS) approach to safety assessment (EFSA, [2007](#efs25894-bib-0001){ref-type=\"ref\"}; EFSA BIOHAZ Panel, [2017](#efs25894-bib-0002){ref-type=\"ref\"}), when used for enzyme production.\n\nThe applicant has provided new data on the stability and the capacity to homogeneously distribute of the additive when added to feed for piglets.[8](#efs25894-note-1013){ref-type=\"fn\"} The two formulations of the additive (one batch each) were added to a mash feed to provide 1,000 U/kg feed. The mash feed supplemented with the solid formulation was also pelleted (60--65\u00b0C) in order to study the effect of processing. Recovery values after granulation showed no modifications of the initial enzyme activity. Samples of the mash and granulated feed were stored for 3 months at 20--25\u00b0C and 50--60% relative humidity (containers not specified). After 3 months of storage, recovery values showed reductions of the initial enzyme activity below 5%.\n\nThe capacity of the phytase to homogeneously distribute was studied in 10 subsamples of the feeds used in the stability study. Samples of the mash feeds showed a coefficient of variation (CV) of 8% for both the solid and liquid formulations and the samples of the pelleted feed showed a CV of 11%.\n\nThe additive is to be used in feed for suckling and weaned piglets and in growing minor porcine species at a minimum recommended enzyme activity of 250 U/kg feed.\n\n3.2. Safety {#efs25894-sec-0010}\n-----------\n\nThe safety aspects regarding the use of this additive in feed including the safety of the genetic modification of the production strain, the safety for the consumers, for the user and for the environment have been previously assessed (EFSA FEEDAP Panel, [2019](#efs25894-bib-0005){ref-type=\"ref\"}). The Panel\u00a0concluded that the use of the product as a feed additive raises no concerns for consumer safety and for the environment. Regarding the safety for the user, the Panel\u00a0concluded that additive is not irritant for skin or eye and it is not a dermal sensitiser, but it is considered a potential respiratory sensitiser.\n\nThe FEEDAP Panel\u00a0is not aware of any new information that would lead it to reconsider the conclusions drawn previously and considers that the extension of use to the new species for which the application is made would not have an impact on the safety aspects already considered. However, the safety for the new target species needs to be addressed.\n\n### 3.2.1. Safety for piglets {#efs25894-sec-0011}\n\nA tolerance trial in weaned piglets was submitted.[9](#efs25894-note-1014){ref-type=\"fn\"} A total of 144 weaned piglets (male and female, (Duroc \u00d7 Landrace) \u00d7 Pi\u00e9train)) were distributed in 36 pens in groups of 4 animals (mixed sexes). The first seven days after weaning the animals were in an adaptation period and received a common diet (see positive control). Then after, the experimental period started the animals (body weight 9 kg) received one of the six experimental diets (representing 6 replicates per treatment). The experimental diets were obtained from two basal diets (starter and grower) based on maize and soya bean meal (total phosphorus 4.2 and 3.9 g/kg, total calcium 8.0 and 7.1 g/kg, respectively) which were either not supplemented (control) or supplemented with APSA PHYTAFEED^\u00ae^ 20,000 GR to provide 250 (1\u00d7 recommended level), 500 (2\u00d7), 1,000 (4\u00d7) or 100,000 (400\u00d7) U per kg feed (confirmed by analysis). A positive control diet was also included (total phosphorus 6.7 and 6.2 g/kg, total calcium 8.0 and 7.1 g/kg, respectively). Diets were offered on *ad\u00a0libitum* basis in pelleted form for 42 days. Mortality and health status were checked daily. Animals were individually weighed on days 0, 14 and 42 under study, feed intake was registered per pen and feed to gain ratio calculated. Blood samples were obtained in 12 piglets on day 0 (6 males and 6 females) and 42 (random selection per treatment not specified how many per pen) for haematology[10](#efs25894-note-1015){ref-type=\"fn\"} and clinical chemistry analysis.[11](#efs25894-note-1016){ref-type=\"fn\"} An analysis of variance (ANOVA) was done with the performance data (pen basis) and considering the treatment as the effect. Group means were compared with Tukey test. Significance level was set at 0.05.\n\nNo piglets died during the study. The results on the feed intake, body weight and feed to gain ratio are presented in Table\u00a0[1](#efs25894-tbl-0001){ref-type=\"table\"}. The results showed a reduced growth of the piglets in the control diet compared to the other groups. The addition of the phytase restored, as compared to the positive control, the growth of the piglets from 1,000 U/kg feed. The performance registered in the control group does not allow to use it as a control to identify any impairment of the performance in animals receiving the phytase. However, the data from the groups receiving the phytase did not show a negative dose related trend with increasing levels of the phytase in the performance; in fact, significant improvements on the performance were seen with increasing levels of the phytase, from the level of 250 U/kg feed. These improvements resulted in no differences in the final body weight compared to the positive control but only when the supplementation reached 1,000 U/kg feed or higher.\n\n###### \n\nEffect of APSA PHYTAFEED^\u00ae^ 20,000 on the performance of weaned piglets\n\n  Groups (U/kg feed)   Daily feed intake (g)   Final body weight (Kg)                         Feed to gain ratio\n  -------------------- ----------------------- ---------------------------------------------- ----------------------------------------------\n  0                    595                     23.0[a](#efs25894-note-0006){ref-type=\"fn\"}    1.80[a](#efs25894-note-0006){ref-type=\"fn\"}\n  250                  671                     25.3[a](#efs25894-note-0006){ref-type=\"fn\"}    1.74[ab](#efs25894-note-0006){ref-type=\"fn\"}\n  500                  647                     25.5[a](#efs25894-note-0006){ref-type=\"fn\"}    1.65[ab](#efs25894-note-0006){ref-type=\"fn\"}\n  1,000                644                     26.4[ab](#efs25894-note-0006){ref-type=\"fn\"}   1.56[a](#efs25894-note-0006){ref-type=\"fn\"}\n  100,000              681                     27.3[a](#efs25894-note-0006){ref-type=\"fn\"}    1.57[ab](#efs25894-note-0006){ref-type=\"fn\"}\n  Positive control     685                     27.2[a](#efs25894-note-0006){ref-type=\"fn\"}    1.58[ab](#efs25894-note-0006){ref-type=\"fn\"}\n\n^a,b,c^ Values in the same column not sharing the same superscript are significantly different (p \\< 0.05).\n\nJohn Wiley & Sons, Ltd\n\nThe results in the blood parameters showed effects on the mean corpuscular volume, haematocrit, haemoglobin, erythrocytes, alkaline phosphatase and alanine aminotransferase. However, most of the effects did not show a dose response, the only one that showed a clear pattern was the alkaline phosphatase which was highest in the negative control 235.8 (U/L) compared to the diets supplemented with phytase from 500 U/kg onwards and the positive control. However, this effect can be related to the availability of phosphorus in the animals.\n\nA subchronic oral toxicity study in rats, formerly assessed by the FEEDAP Panel\u00a0(EFSA FEEDAP Panel, [2019](#efs25894-bib-0005){ref-type=\"ref\"}), was provided.[12](#efs25894-note-1017){ref-type=\"fn\"} The results of that study indicate a no\u2010observed\u2010adverse\u2010effect\u2010level (NOAEL) of 119,228 U/kg body weight in rats. Using this NOAEL, and applying the procedure detailed in the guidance on the safety for the target species (EFSA FEEDAP Panel, [2017](#efs25894-bib-0003){ref-type=\"ref\"}), the maximum safe level for piglets in feed is calculated to be 23,846 U/kg feed. This value is approximately 95 times higher than the proposed use level and would support the results of the tolerance study.\n\n#### 3.2.1.1. Conclusions on the safety for the target species {#efs25894-sec-0012}\n\nThe FEEDAP Panel\u00a0concludes that APSA PHYTAFEED^\u00ae^ 20,000 GR/L is safe for weaned piglets at the recommended level of 250 U/kg feed with a wide margin of safety. This conclusion is extended to suckling piglets. Considering the wide margin of safety the Panel\u00a0extrapolates the conclusion to growing minor porcine species.\n\n3.3. Efficacy for piglets {#efs25894-sec-0013}\n-------------------------\n\nTwo digestibility trials that included bone measurements and a performance trial were submitted for the assessment.\n\nThe two digestibility trials were conducted in the same trial site and followed the same trial design.[13](#efs25894-note-1018){ref-type=\"fn\"} In the two trials, male weaned piglets ((Duroc \u00d7 Landrace) \u00d7 Yorkshire) were subject to an adaptation period of 7 days after weaning and then after were individually caged and one of the five treatments were allocated. In trial 1, there were a total of 7 piglets per treatment and in trial 2 a total of 8. Basal diets (pre\u2010starter and starter) based on maize and soya bean meal were either not supplemented (control) or supplemented with APSA PHYTAFEED^\u00ae^ 20,000 GR to provide 250, 500 or 1,000 U/kg feed. The enzyme activities were confirmed by analysis. In the two trials, a positive control diet with higher content of phosphorus was also considered. Diets were offered in mash form for 24 or 21 days, respectively, and the ones given in the starter phase (from day 17 in trial 1 or from day 14 in trial 2) contained an external marker (titanium dioxide in trial 1 and insoluble ash in trial 2).\n\nMortality and health status of the piglets were checked every day. Animals were weighed and feed intake was measured throughout the study period. Spot samples of faeces were collected in days 22--24 in trial 1 and days 15--19 in trial 2 (no adaptation period to the diet). Feed and the faeces collected were analysed for the phosphorus content and external marker in order to study the digestibility. On the last day of the study, all animals were killed and the *os metacarpale* was collected from each animal to analyse it for ash and phosphorus content. An analysis of variance was done with the data and group means were compared with Tukey test. Significance level was set at 0.05.\n\nThe performance trial is the tolerance trial described in Section\u00a0[3.2.1](#efs25894-sec-0011){ref-type=\"sec\"}. In that trial, the performance of the piglets was measured and measurements on the digestibility of phosphorus and bone content of minerals were also done. Faecal samples were collected on days 19\u201021 under study. The diets contained titanium dioxide as an external marker and faecal samples were collected daily from the pens. Feed and excreta samples were analysed for the marker, ash, dry matter, calcium and phosphorus to determine the utilisation. *Os metacarpale* from one piglet per pen was collected at the end of the study (random selection) and the bones were analysed for ash content.\n\nThe results of the balance trials are presented in Table\u00a0[2](#efs25894-tbl-0002){ref-type=\"table\"} and the performance parameters of trial 3 in Table\u00a0[1](#efs25894-tbl-0001){ref-type=\"table\"} (see Section\u00a0[3.2.1](#efs25894-sec-0011){ref-type=\"sec\"}). The piglets that received the phytase showed higher faecal apparent digestibility of phosphorus in the three trials compared to the control group, from the enzyme activity of 250 U/kg feed in trials 1 and 2 and from 1,000 U/kg feed in trial 3. The bone content of ash and/or phosphorus was higher in the piglets receiving the phytase from 250 U/kg feed in trial 2,500 U/kg feed in trial 1 and from 1,000 U/kg feed in trial 3. The Panel\u00a0considers that the performance data in the tolerance trial can only support the efficacy of the additive due to the performance of the piglets. The results on the utilization of phosphorus show that the additive has a potential to be efficacious as a zootechnical additive in weaned piglets at the dose of 1,000 U/kg feed.\n\n###### \n\nEffect of APSA PHYTAFEED^\u00ae^ 20,000 on the apparent faecal digestibility of phosphorus and *os metacarpale* bone mineralisation\n\n  Trial              Diets      Faecal apparent digestibility of phosphorus (%)   Bone content (%)                                                                             \n  ------------------ ---------- ------------------------------------------------- ---------------------------------------------- --------------------------------------------- --------------------------------------------\n  1                  0          3.0--3.4                                          41.6[c](#efs25894-note-0008){ref-type=\"fn\"}    37.4                                          3.2[b](#efs25894-note-0008){ref-type=\"fn\"}\n  250                3.0--3.4   63.8[b](#efs25894-note-0008){ref-type=\"fn\"}       41.1                                           3.6[ab](#efs25894-note-0008){ref-type=\"fn\"}   \n  500                3.0--3.4   74.9[a](#efs25894-note-0008){ref-type=\"fn\"}       42.5                                           3.8[a](#efs25894-note-0008){ref-type=\"fn\"}    \n  1,000              3.0--3.4   77.2[a](#efs25894-note-0008){ref-type=\"fn\"}       42.4                                           3.8[a](#efs25894-note-0008){ref-type=\"fn\"}    \n  Positive control   3.6--4.5   62.9[b](#efs25894-note-0008){ref-type=\"fn\"}       42.2                                           3.5[ab](#efs25894-note-0008){ref-type=\"fn\"}   \n  2                  0          3.7--4.8                                          50.5[c](#efs25894-note-0008){ref-type=\"fn\"}    37.8[b](#efs25894-note-0008){ref-type=\"fn\"}   3.2[b](#efs25894-note-0008){ref-type=\"fn\"}\n  250                3.7--4.8   65.7[ab](#efs25894-note-0008){ref-type=\"fn\"}      42.9[a](#efs25894-note-0008){ref-type=\"fn\"}    3.7[a](#efs25894-note-0008){ref-type=\"fn\"}    \n  500                3.7--4.8   73.9[a](#efs25894-note-0008){ref-type=\"fn\"}       42.8[a](#efs25894-note-0008){ref-type=\"fn\"}    3.8[a](#efs25894-note-0008){ref-type=\"fn\"}    \n  1,000              3.7--4.8   77.0[a](#efs25894-note-0008){ref-type=\"fn\"}       42.6[a](#efs25894-note-0008){ref-type=\"fn\"}    3.9[a](#efs25894-note-0008){ref-type=\"fn\"}    \n  Positive control   5.4--6.9   61.7[bc](#efs25894-note-0008){ref-type=\"fn\"}      41.4[a](#efs25894-note-0008){ref-type=\"fn\"}    3.6[ab](#efs25894-note-0008){ref-type=\"fn\"}   \n  3                  0          3.9--7.1                                          22.9[c](#efs25894-note-0008){ref-type=\"fn\"}    32.6[c](#efs25894-note-0008){ref-type=\"fn\"}   --\n  250                3.9--7.1   33.0[c](#efs25894-note-0008){ref-type=\"fn\"}       36.2[c](#efs25894-note-0008){ref-type=\"fn\"}    --                                            \n  500                3.9--7.1   43.6[bc](#efs25894-note-0008){ref-type=\"fn\"}      37.2[bc](#efs25894-note-0008){ref-type=\"fn\"}   --                                            \n  1,000              3.9--7.1   55.0[ab](#efs25894-note-0008){ref-type=\"fn\"}      41.8[ab](#efs25894-note-0008){ref-type=\"fn\"}   --                                            \n  100,000            3.9--7.1   68.9[a](#efs25894-note-0008){ref-type=\"fn\"}       44.3[a](#efs25894-note-0008){ref-type=\"fn\"}    --                                            \n  Positive control   6.2--7.1   26.2[c](#efs25894-note-0008){ref-type=\"fn\"}       42.8[a](#efs25894-note-0008){ref-type=\"fn\"}    --                                            \n\n^1^ Intended values for the diets administered during the period in which collection of faeces was done.\n\n^a,b,c^Values in the same column not sharing the same superscript are significantly different (p \\< 0.05).\n\nJohn Wiley & Sons, Ltd\n\nThe FEEDAP Panel\u00a0concludes that APSA PHYTAFEED^\u00ae^ 20,000 GR/L is efficacious as a zootechnical additive for weaned piglets at 1,000 U/kg feed. This conclusion is extended to suckling piglets. Considering that the mode of action of the phytases is well known and it is reasonably assumed to be the same among porcine species, the Panel\u00a0extrapolates the conclusion on the efficacy to growing minor porcine species.\n\n3.4. Post\u2010market monitoring {#efs25894-sec-0014}\n---------------------------\n\nThe FEEDAP Panel\u00a0considers that there is no need for specific requirements for a post\u2010market monitoring plan other than those established in the Feed Hygiene Regulation[14](#efs25894-note-1019){ref-type=\"fn\"} and Good Manufacturing Practice.\n\n4. Conclusions {#efs25894-sec-0015}\n==============\n\nAPSA PHYTAFEED^\u00ae^ 20,000 GR/L is safe for piglets (suckling and weaned) and for growing minor porcine species at the recommended enzyme activity of 250 U/kg feed with a wide margin of safety.\n\nThe FEEDAP Panel\u00a0concludes that there are no concerns for consumer safety and no risks for the environment are expected from the use of APSA PHYTAFEED^\u00ae^ 20,000 GR/L in piglets and growing minor porcine species. The additive is not a skin or eye irritant, and it is not a dermal sensitizer but it should be considered a respiratory sensitizer.\n\nThe FEEDAP Panel\u00a0concludes that APSA PHYTAFEED^\u00ae^ 20,000 GR/L is efficacious for piglets (suckling and weaned) and for growing minor porcine species at an enzyme activity of 1,000 U/kg feed.\n\nDocumentation as provided to EFSA/Chronology {#efs25894-sec-0016}\n============================================\n\nDateEvent04/03/2019Dossier received by EFSA. APSA PHYTAFEED^\u00ae^ 20,000 GR/L for piglets and minor growing porcine species. Submitted by Andr\u00e9s Pintaluba S.A.18/03/2019Reception mandate from the European Commission03/05/2019Application validated by EFSA -- Start of the scientific assessment08/07/2019Request of supplementary information to the applicant in line with Article 8(1)(2) of Regulation (EC) No 1831/2003 -- Scientific assessment suspended. *Issues: characterisation and efficacy*14/08/2019Clarification teleconference during risk assessment with the applicant according to the \"EFSA\\'s Catalogue of support initiatives during the life\u2010cycle of applications for regulated products\"26/09/2019Reception of supplementary information from the applicant \u2010 Scientific assessment re\u2010started7/10/2019Opinion adopted by the FEEDAP Panel. End of the Scientific assessment\n\nAbbreviations {#efs25894-sec-0017}\n=============\n\nANOVAanalysis of varianceCGMCCChina General Microbiological Culture Collection CentreCVcoefficient of variationEURLEuropean Union Reference LaboratoryFEEDAPEFSA Panel\u00a0on Additives and Products or Substances used in Animal FeedNOAELno\u2010observed\u2010adverse\u2010effect\u2010levelQPSQualified Presumption of Safety\n\nRegulation (EC) No 1831/2003 of the European Parliament and of the Council of 22 September 2003 on additives for use in animal nutrition. OJ L 268, 18.10.2003, p. 29.\n\nAndr\u00e9s Pintaluba S.A. Pol. Ind. Agro Reus, c/Prudenci Bertrana, 5, Reus 43206, Spain.\n\nFEED dossier reference: FAD\u20102019\u20100020.\n\nThe full report is available on the EURL website: <https://ec.europa.eu/jrc/sites/jrcsh/files/fin_report_fad-2018-0031_phytafeed.pdf>\n\nCommission Regulation (EC) No 429/2008 of 25 April 2008 on detailed rules for the implementation of Regulation (EC) No 1831/2003 of the European Parliament and of the Council as regards the preparation and the presentation of applications and the assessment and the authorisation of feed additives. OJ L 133, 22.5.2008, p. 1.\n\nTechnical dossier/Section\u00a0II/Annex II.2.1.2.\n\nOne Unit (U) is defined as the amount of enzyme that releases 1 \u00b5mol of inorganic phosphate from phytate per minute at pH 5.5 and 37\u00b0C.\n\nTechnical dossier/Supplementary information September 2019/Annex II.2.1.2.7.\n\nTechnical dossier/Section\u00a0II/Annex II.4.1.3.1 and II.4.1.3.2.\n\nTechnical dossier/Section\u00a0III/Annex III.1.1.1.\n\nParameters measured included: mean corpuscular haemoglobin concentration, mean corpuscular haemoglobin mean corpuscular volume, haematocrit, haemoglobin, erythrocytes, platelets, leukocytes, eosinophils, basophils, lymphocytes, monocytes, segmented neutrophils and band neutrophils.\n\nParameters measured included: alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, gamma glutamine transpeptidase, total proteins, albumin and Uric acid.\n\nTechnical dossier/Section\u00a0III/Annex III.2.2.3.\n\nTechnical dossier/Section\u00a0IV/Annex IV.2.1 and Annex IV.2.2 and supplementary information September 2019/Annexes IV.2.1_updated and IV.2.2_updated.\n\nRegulation (EC) No 183/2005 of the European Parliament and of the Council of 12 January 2005 laying down requirements for feed hygiene. OJ L 35, 8.2.2005, p. 1.\n"}
+{"text": "Background\n==========\n\nUnlike mammals, adult newts have the remarkable ability to recover function after they are paralyzed by a spinal cord injury (SCI). After a complete transection injury, newts regenerate their spinal cords and regain use of their hindlimbs in as little as 4 weeks \\[[@B1]\\] (Additional file [1](#S1){ref-type=\"supplementary-material\"}). This recovery requires supraspinal axons to regenerate across the lesion and re-establish connections with downstream targets and is not simply due to a reorganization of circuits within the spinal cord \\[[@B1]\\]. This finding led us to ask the question: why do axons regenerate across an injury site in the newt when they do not in mammals?\n\nOne of the main reasons why regeneration fails in mammals is because the environment of the injured spinal cord is inhibitory for axon regeneration \\[[@B2]\\]. After an SCI, a variety of cell types, including astrocytes and meningeal fibroblasts, react in ways that prevent axons from regenerating across the injury site. These reactive cells create physical barriers to regeneration, such as a glial scar and a glia limitans at the border between the cord and the injury site. They also create an extracellular matrix (ECM) that is inhibitory or repulsive for axon growth cone migration.\n\nTherefore, axon regeneration may be enabled in the newt, in part, because the environment of the injury site is not inhibitory. Cells may respond in ways that help rather than hinder axon regeneration such that physical barriers are not created and the ECM is not inhibitory.\n\nMuch of what is known about spinal cord regeneration in salamanders comes from studies of tail regeneration. After tail amputation, a blastema forms and ependymoglia (EG) lining the central canal of the spinal cord elongate an ependymal tube that precedes and serves as scaffold for axon regeneration \\[[@B3]\\]. Regeneration in this context is thought to proceed as a recapitulation of developmental processes, and axons grow into newly developing tissues. Surprisingly, little is known about how axons regenerate after an SCI in the newt. In this context, axons must re-grow through an injury site having mature tissues on both sides of the lesion. This context is more relevant to the problem of spinal cord injury in humans. Older studies of SCI in the newt have noted that a blastema and glial scar do not appear to form \\[[@B4]\\], that axons can bridge large gaps in the cord before ependymal tubes elongate \\[[@B5]\\], and that, if left intact, the meninges can serve as a scaffold for axon regeneration \\[[@B6]\\]. A more recent study of SCI in the axolotl, a neotenic larval salamander, found that EG appear to undergo an epithelial to mesenchymal transition, migrate into the injury site to form a solid mass, and then undergo a mesenchymal to epithelial transition to re-form an ependymal tube that serves as a scaffold for axon regeneration \\[[@B7],[@B8]\\]. In summary, previous studies suggest that physical barriers do not appear to form and EG and the meninges may help axons regenerate. Although O\\'Hara *et al*. \\[[@B7]\\] demonstrated that mesenchymal cells in the axolotl injury site were associated with fibronectin (FN), a permissive ECM protein, little else is known about the nature of the ECM of the injured newt spinal cord.\n\nA focused and detailed study of axon regeneration and the cellular and extracellular environment axons encounter after an SCI in the newt has not been conducted. We, therefore, took advantage of modern labeling and imaging techniques to define, for the first time, stages of newt axon regeneration after a spinal cord transection injury and find that meningeal cells and glia, instead of interacting to form barriers to axon regeneration as they do in mammals \\[[@B9],[@B10]\\], appear to interact to form a permissive environment for axon regeneration in the newt. Our study examines many aspects of newt spinal cord regeneration, establishes the required foundation for further investigation into the cellular and molecular mechanisms controlling this naturally occurring process, and may inform new therapeutic approaches in regenerative medicine.\n\nResults and discussion\n======================\n\nTo understand how newt axons regenerate after an SCI we carefully observed axons regenerating across a complete transection injury, which severed all innervation to the tail and hindlimbs. Animals were allowed to regenerate for 1 day, 3 days, 1 week, 2 weeks, 3 weeks, 4 weeks, 6 weeks or 9 weeks, and prior to tissue harvest an axon tracer was applied rostral or caudal to the original injury site in order to label descending or ascending regenerating axons, respectively. The tracer and nuclei were labeled fluorescently, and the injury site was analyzed in whole-mount preparation on a confocal microscope. At least three animals were observed per tracer application site and time point (Table [1](#T1){ref-type=\"table\"}).\n\n###### \n\nThe number of animals observed at each stage and time point\n\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Time point   Total            Retraction      Growth initiation   Wrapping      Wrapping/\\    Wisping       Wisping/\\     Spiking (no ET)   Spiking (with ET)   Contact       Growth beyond injury (not\\   Growth beyond injury\\\n                                                                                  wisping                     spiking\\                                                          recovered)                   (recovered)\n                                                                                                              (no ET)                                                                                        \n  ------------ ---------------- --------------- ------------------- ------------- ------------- ------------- ------------- ----------------- ------------------- ------------- ---------------------------- -----------------------\n  1 day        **6 (3/3)**      6 (3/3)                                                                                                                                                                      \n\n  3 day        **8 (4/4)**      6 (4/2)         2 (-/2)                                                                                                                                                      \n\n  1 week       **15 (11/4)**    3 (3/-)         12 (8/4)                                                                                                                                                     \n\n  2 week       **11 (6/5)**                     2 (1/1)             3 (1/2)       2 (1/1)       4 (3/1)                                                                                                      \n\n  3 week       **10 (5/5)**                                         2 (2/-)                     4 (-/4)                                       1 (1/-)             1 (1/-)       2 (1/1)                      \n\n  4 week       **8 (4/4)**                                          4 (3/1)                                                                   1 (1/-)                           3 (-/3)                      \n\n  6 week       **7 (4/3)**                                                                      1 (-/1)       1 (-/1)       1 (-/1)           2 (2/-)             1 (1/-)                                    1 (1/-)\n\n  9 week       **6 (4/2)**                                                                                                                                                      2 (1/1)                      4 (3/1)\n\n  **Totals**   **71 (41/30)**   **15 (10/5)**   **16 (9/7)**        **9 (6/3)**   **2 (1/1)**   **9 (3/6)**   **1 (-/1)**   **1 (-/1)**       **4 (4/-)**         **2 (2/-)**   **7 (2/5)**                  **5 (4/1)**\n  ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nThe numbers in parentheses break the totals for each category into rostral and caudal tracer application, respectively. For example, (4/2) indicates 4 animals received rostral tracer and 2 received caudal tracer. Animals that had recovered function were chosen preferentially for rostral tracer application. ET, ependymal tube.\n\nSpinal cord transection injury and axon tracer application\n----------------------------------------------------------\n\nThe spinal column in *Notophthalmus viridescens*contains one cervical (the atlas), about 13 trunk, 1 to 2 sacral, and 22 to 25 caudal vertebrae (Figure [1A](#F1){ref-type=\"fig\"}) \\[[@B11]\\]. The numbers of trunk, sacral and caudal vertebrae appear to be variable. The hindlimbs are innervated primarily by the last two trunk (T-1 and T-2) and first sacral (S1) spinal nerves, though there can be contributions from the third to last trunk (T-3) vertebra and the vertebra after S1 (S2/C1) (Figure [1D](#F1){ref-type=\"fig\"}) \\[[@B11]\\]. This is supported by the fact that the largest spinal nerves are associated with T-2, T-1 and S (Additional file [2](#S2){ref-type=\"supplementary-material\"}) and axon tracer applied to the sciatic nerve labels primarily neurons in these spinal ganglia. To be sure all innervation to the hindlimbs was severed, we aimed to transect the spinal cord between T-4 and T-3 (Figure [1A, D](#F1){ref-type=\"fig\"}), about 1 cm rostral to the hindlimbs. Upon dissection, it was often found that the injury was actually one or two segments rostral to the targeted site (Additional file [2B](#S2){ref-type=\"supplementary-material\"}). Figure [1B](#F1){ref-type=\"fig\"} shows what the intact spinal cord looks like when exposed from the dorsal side. After a complete transection injury, the two ends of the cut cord spring away from each other leaving a gap in the cord (Figure [1C](#F1){ref-type=\"fig\"}). This type of injury was chosen because it is simple to perform consistently, spares no axons, severs the meninges as well as axons, and allows our studies to more seamlessly complement those done by previous investigators \\[[@B1],[@B4],[@B12]\\]. All injured animals were paralyzed and unable to swim for at least 3 weeks (Additional file [1](#S1){ref-type=\"supplementary-material\"}), though the typical reflexive movements of the hindlimbs and tail were observed. Recovery of swimming function was scored and filmed in 6- and 9-week regenerates. One out of seven (14%) 6-week regenerates and four out of six (67%) 9-week regenerates recovered swimming function (Table [1](#T1){ref-type=\"table\"}). This recovery rate is consistent with that reported in Davis *et al*. \\[[@B1]\\].\n\n![**Spinal cord transection injury and axon tracer application**. **(A)**Newt skeleton showing targeted location of injury. Note, the sacroiliac (SI) joints are not associated with the same vertebra in this animal, thus there are two sacral (S) vertebrae where there is usually one. T, trunk; C, caudal. **(B)**Intact and **(C)**completely transected spinal cord viewed from dorsal side. **(D)**Cartoon showing the location of the targeted injury, the spinal ganglia that supply innervation to the hindlimbs (solid line, primary contribution; dotted line, occasional contribution), and rostral tracer application site. **(E)**Cross-section of spinal cord about 500 \u03bcm away from a tracer application site showing the tracer labels axons and neurons specifically. **(F, G)**Longitudinal sections through 1-week regenerates. Asterisk, injury site. The tracer was applied to the animal in (F) and not to the animal in (G). Axons were also labeled with the 3A10 antibody. Application of the tracer 6 hours prior to tissue harvest did not alter the overall appearance of the axons at the injury site. cc, central canal; eg, ependymoglia; gm, grey matter; wm, white matter; m, meninges; D, dorsal; V, ventral; R, rostral; C, caudal. Scale bars: 200 \u03bcm ((F) and (G) are the same scale).](1749-8104-6-1-1){#F1}\n\nTo label descending regenerating axons, a piece of gel foam saturated with the axon tracer biotinylated dextran amine (BDA) was inserted into a transection injury 1 segment (about 2 mm) rostral to the original injury site (Figure [1D](#F1){ref-type=\"fig\"}), and the tracer was allowed to travel for 6 hours. Ascending axons were similarly labeled by applying the tracer one segment caudal to the original injury. BDA travels 3 to 4 mm in 6 hours (Figure [2K, L](#F2){ref-type=\"fig\"}), produces a strong signal, and labels only neurons and their processes in the nervous system (Figure [1E](#F1){ref-type=\"fig\"}). It does not diffuse across an injury site to label axons that have not regenerated across it (Figure [2A-D](#F2){ref-type=\"fig\"}). Transecting the spinal cord to apply the tracer rostral or caudal to the original injury site 6 to 24 hours prior to tissue harvest does not appear to alter the general appearance of axons in the injury site (compare Figure [1F](#F1){ref-type=\"fig\"} and [1G](#F1){ref-type=\"fig\"}).\n\n![**Stages of axon regeneration**. Images are montages of single confocal planes (except where indicated) of longitudinal thick sections. Rostral is to the left. Descending (A, C, E, G, I, K) or ascending (B, D, F, H, J, L) axons were labeled with the axon tracer (magenta), and nuclei are shown in green. Time after injury is indicated in the upper right corner of each panel. **(A, B)**Retraction and **(C, D)**growth initiation stages. Arrowheads, end of cord; asterisk, injury site; dotted line, unlabeled cord opposite the injury site. (B) is a z-projection of four planes to highlight growth cones. **(B^\\'^, B^\\\"^, D^\\'^, D^\\\"^)**Enlargements of dotted boxes in (B) and (D) showing that growth cones (primed boxes) and dystrophic axons (double primed boxes) are present at both stages. **(E, F)**Wrapping stage. **(G, H)**Wisping stage. **(I, J)**Spiking stage. (H) and (I) are z-projections of two confocal planes to show wrapping axons with wisping (H) or spiking (I) axons. **(K, L)**Contact and growth beyond the injury site stage. The animal in (K) had recovered swimming function, while the one in (L) had not. Open arrowheads, residual wrapping axons; arrows, meninges. TV, terminal vesicle; R, rostral; C, caudal. Scale bars: 200 \u03bcm (A-L); 50 \u03bcm (B\\', B\\\", D\\', D\\\").](1749-8104-6-1-2){#F2}\n\nStages of axon regeneration\n---------------------------\n\nOur tracer analyses uncovered six stages of axon regeneration that appeared to be typical and sequential (Figure [2](#F2){ref-type=\"fig\"}). The timing for all but the first two stages was highly variable (Table [1](#T1){ref-type=\"table\"}). This variability is likely due to differences in the age, size, health and activity level of the animals as well as the size of the gap in the cord and not due to differences in the completeness of the transection injury. It was quite easy to completely transect the newt spinal cord and visually confirm that the transection was complete (Figure [1C](#F1){ref-type=\"fig\"}). The adult newts used in this study were collected from the wild and their ages are unknown. As noted below, the size of the gap in the cord tended to increase before it decreased, perhaps because of body movements of the animal. The times in parentheses after each stage indicate the earliest and latest time points at which the stage was observed.\n\n### Retraction stage (1 day to 1 week, most often before 1 week)\n\nAxons appear to have retracted away from the end of the cut cord (Figure [2A, B](#F2){ref-type=\"fig\"}), and many axon tips have a balled-up morphology that is reminiscent of the dystrophic end-bulbs seen on axons that are unable to regenerate (Figure [2B\\\"](#F2){ref-type=\"fig\"}) \\[[@B13],[@B14]\\]. Axon growth cones, however, can also be identified at this stage (Figure [2B\\'](#F2){ref-type=\"fig\"}).\n\n### Growth initiation stage (3 days to 2 weeks, most often at 1 week)\n\nAxons appear to have initiated growth and grown back to the end of the cut cord (Figure [2C, D](#F2){ref-type=\"fig\"}). As in the retraction stage, dystrophic axons and growth cones can be identified (Figure [2D\\', D\\\"](#F2){ref-type=\"fig\"}). The main difference between the first two stages is that many more axons extend to the end of the cut cord in the growth initiation stage. This difference is apparent even when all axons from the spinal cord are viewed in z-projections (Additional file [3](#S3){ref-type=\"supplementary-material\"}). Also in this stage, the central canal, especially on the rostral side, begins to enlarge to form the terminal vesicle (TV).\n\n### Wrapping stage (2 weeks to 4 weeks)\n\nAxons continue to grow and wrap around the end of the TV (Figure [2E, F](#F2){ref-type=\"fig\"}). Residual numbers of dystrophic axons can still be seen at this and all subsequent stages. The meninges also wrap around, or bound, the end of the cord (Figure [2E, F](#F2){ref-type=\"fig\"}) and often begin regenerating across the injury site ahead of the axons and TV (Figure [2F](#F2){ref-type=\"fig\"}). The low density of cells within the injury site suggests that a scar is not forming. The size of the gap in the cord is also larger at this stage than at all other stages (Additional file [4](#S4){ref-type=\"supplementary-material\"}). This could indicate that the wrapping stage is an abortive stage seen only in animals that do not recover function. We do not favor this interpretation, however, because wrapping axons are still evident at all subsequent stages, including the contact and growth beyond stage. This suggests that wrapping is a typical stage through which all regenerating spinal cords progress. The gap in the cord may get bigger before it gets smaller, simply as a result of body movements of the animal.\n\n### Wisping stage (2 weeks to 6 weeks)\n\nAxons begin growing, or wisping, into the injury site ahead of the EG-lined TV (Figure [2G, H](#F2){ref-type=\"fig\"}). Thus, as seen in previous studies of SCI in the newt, a pre-formed ependymal tube is not required for axon regeneration into the injury site \\[[@B5],[@B6]\\]. Wisping axons, however, do not grow into the ECM of the injury site alone but rather remain associated with cells. Many of these cells appear to be continuous with the meninges. At this stage, the meninges do not bound the end of the cord as they did in the wrapping stage but continue regenerating across the injury site, preceding the regenerating axons. Some wisping axons even appear to follow the regenerating meninges (Additional file [5](#S5){ref-type=\"supplementary-material\"}).\n\n### Spiking stage (3 weeks to 6 weeks)\n\nAs wisping axons continue to grow, they may fasciculate to form a spike (Figure [2I, J](#F2){ref-type=\"fig\"}; Additional files [6](#S6){ref-type=\"supplementary-material\"} and [7](#S7){ref-type=\"supplementary-material\"}). Spiking axons are still associated with cells in the injury site and cells that are continuous with the meninges. Some spikes do not contain an ependymal tube (Figure [2J](#F2){ref-type=\"fig\"}), while others do (Figure [2I](#F2){ref-type=\"fig\"}). In spikes with an ependymal tube, the ependymal tube appears to have elongated from the end of the cut cord because the largest part of the TV and residual wrapping axons are seen well behind the tip of the spike (Figure [2I](#F2){ref-type=\"fig\"}; Additional file [7](#S7){ref-type=\"supplementary-material\"}). It is conceivable that a spike may not need to form. For example, if the gap in the cord is not large, wisping axons may be able to grow through the lesion and enter the cord on the opposite side without forming a spike.\n\n### Contact and growth beyond the injury site stage (3 weeks to 9 weeks)\n\nIn this stage, the two ends of the cord make contact with each other, and axons begin growing through the cord on the opposite side of the injury to re-establish functional connections (Figure [2K, L](#F2){ref-type=\"fig\"}). All animals that had recovered swimming ability were in this stage, although not all animals in this stage had recovered function. Also, in all animals that had recovered, the ependymal tubes on each side of the injury had fused together to re-establish a continuous central canal. It was not possible to correlate the distance axons had regenerated beyond the injury site with functional recovery in this study because the axon tracer appears to fade out after 3 to 4 mm (Figure [2K, L](#F2){ref-type=\"fig\"}). This is true even in the intact spinal cord (data not shown). There is some tolerance for sloppiness in how axons regenerate, for even in animals that had recovered function, wrapping axons were still evident (Additional files [8](#S8){ref-type=\"supplementary-material\"}A and [9](#S9){ref-type=\"supplementary-material\"}), some axons appeared to decussate (Additional files [8](#S8){ref-type=\"supplementary-material\"}A\\' and [9](#S9){ref-type=\"supplementary-material\"}), and separate fiber tracts could be seen traveling along the meninges (Figure [2K](#F2){ref-type=\"fig\"}). It would be interesting to determine whether these aberrant axon trajectories are pruned or refined over time in animals that have been allowed to regenerate for longer than 9 weeks. Excessive sloppiness or gross misalignments may be the cause of a delay in or a complete failure of functional recovery (Additional files [8](#S8){ref-type=\"supplementary-material\"}B-F and [10](#S10){ref-type=\"supplementary-material\"}).\n\nSensory axons do not appear to regenerate\n-----------------------------------------\n\nDescending and ascending regenerating axons appear to progress through similar stages (Table [1](#T1){ref-type=\"table\"}; Additional file [11](#S11){ref-type=\"supplementary-material\"}). This was surprising given that sensory axons are not thought to regenerate after an SCI in newts or zebrafish even though these animals recover function \\[[@B6],[@B15]\\]. To determine if ascending regenerating axons are true sensory axons, rather than axons arising from neurons within the spinal cord, an axon tracer was applied to the sciatic nerve 1 or 4 days prior to tissue harvest (Figure [3A](#F3){ref-type=\"fig\"}). In 6-week regenerates, descending axons were also labeled with a different tracer. Our goal was to identify tracer-labeled sensory axons wisping into and growing through and beyond the lesion.\n\n![**Sensory axons may not regenerate**. (**A**) Cartoon showing tracer application sites in relationship to the SCI. Tracer was applied to the sciatic nerve in the hindlimb to label sensory axons (green) and, in 6-week regenerates, rostral to the injury to label descending axons (magenta). Tracer applied to the sciatic nerve labels primarily neurons in spinal ganglia S1 and T-1 (solid lines) and occasionally neurons in T-2 (dotted lines). **(B)**Cross-section through spinal cord and spinal ganglia at T-1, 1 day after SCI and tracer application to the sciatic nerve. The tracer labels sensory neurons in the spinal ganglia (sg), motor neurons (mn), dorsal column axons (dc), and motor neuron axons (mna) exiting to form the ventral root. **(C-E)**Sections through a 3-week regenerate. Sensory axons were labeled with biotinylated dextran amine (BDA; green), and all axons were labeled with 3A10 (magenta in (D, E)). (C) Cross-section at T-1. Inset, spinal ganglia on labeled side. (D, E) Longitudinal sections through the SCI at the level of the dorsal columns (D), and a more ventral level where there are wisping axons (E). Schematic cross-sections of the spinal cord in the upper right corner show where the section is along the D-V axis. **(F-H)**Cross-sections through a 6-week regenerate that had recovered swimming function at T-1 (F), at T-4, about 500 \u03bcm caudal to the SCI (G), and through the center of the SCI at T-4/T-5 (H). Sensory axons were labeled with rhodamine dextran amine (RDA; green), and descending axons were labeled with BDA (magenta). Inset in (F), spinal ganglia on labeled side. Inset in (H), phase image of injury site. Nuclei are shown in blue. D, dorsal; V, ventral; R, rostral; C, caudal. All scale bars: 200 \u03bcm ((C-H) are the same scale; insets in (C, F) are the same scale).](1749-8104-6-1-3){#F3}\n\nTracer application to the sciatic nerve effectively labels, on the ipsilateral side only, sensory neurons in the spinal ganglia, motor neurons, dorsal column axons, and occasionally axons in a ventral location, which are presumably motor neuron axons exiting the cord to form the ventral root (Figure [3B](#F3){ref-type=\"fig\"}). Labeled sensory and motor neurons were found mostly at the level of S1 and T-1 and occasionally at T-2 (Figure [3A](#F3){ref-type=\"fig\"}).\n\nTo determine if sensory axons are among axons wisping into the injury site, 3-week regenerates received tracer application to the sciatic nerve 4 days prior to tissue harvest (n = 3). The spinal cord caudal to the injury site (T-2 to S1) was analyzed in cross-section to determine the effectiveness of the tracer application. Sensory and motor neurons and dorsal column axons were labeled in all three animals (Figure [3C](#F3){ref-type=\"fig\"}). The injury site was analyzed in longitudinal sections to best view wisping axons. In sections through the dorsal columns, tracer-labeled sensory axons could be seen all the way up to the end of the cut cord (n = 2 out of 3; Figure [3D](#F3){ref-type=\"fig\"}). In ventral sections through the region of wisping axons, however, none of the wisping axons were labeled with the tracer (Figure [3E](#F3){ref-type=\"fig\"}).\n\nTo determine if sensory axons grow through and beyond an injury site, 6-week regenerates received rhodamine dextran amine (RDA) application to the sciatic nerve 4 days prior to tissue harvest. BDA was also used to label descending axons. The spinal cord caudal to the lesion and the lesion itself were analyzed in cross-section (n = 4). Sensory and motor neurons and dorsal column axons were effectively labeled with RDA in the spinal cord caudal to the lesion in three out of the four animals (Figure [3F](#F3){ref-type=\"fig\"}). Two of the animals in which RDA was effectively applied had recovered function, and BDA-labeled descending axons could be seen growing through and at least 500 \u03bcm caudal to the injury site (Figure [3G, H](#F3){ref-type=\"fig\"}). Despite this, RDA labeling in dorsal column axons got progressively weaker towards the injury, and RDA-labeled dorsal column axons were never seen in the injury site in any animal (Figure [3G, H](#F3){ref-type=\"fig\"}).\n\nWe were not able to detect tracer-labeled sensory axons wisping into the lesion and growing through or beyond the lesion. This could indicate that sensory axons are either delayed or not regenerating or that the tracer faded out before the injury site. The tracer appears to travel a sufficient distance in four days to label wisping axons because it was seen robustly in dorsal column axons all the way up to the injury site in 3-week regenerates (Figure [3D](#F3){ref-type=\"fig\"}). If the axons had regenerated, however, it is possible that the tracer was not detected in the injury site because of a dilution effect. The most likely explanation for why the tracer faded in the dorsal columns of the 6-week regenerates is because the dorsal column axons were degenerating \\[[@B6]\\]. Therefore, sensory information may not be needed for functional recovery in the newt, or sensory information may be transmitted to the brain via a relay circuit that is formed by ascending propriospinal axons that do regenerate across the injury site.\n\nAxons regenerate through white matter to reach functional targets\n-----------------------------------------------------------------\n\nAxons regenerating beyond the injury site in zebrafish travel preferentially through the grey matter rather than white matter to reach functional targets \\[[@B16]\\]. Our initial study of axon regeneration in whole-mount preparations suggested that this is not true in newts (Figure [2K, L](#F2){ref-type=\"fig\"}). To investigate this further, descending axons in three 6-week regenerates in the growth beyond the injury site stage were analyzed in cross-section about 500 \u03bcm caudal to the injury site. All axons in all three animals traveled in the white matter only (Figure [3G](#F3){ref-type=\"fig\"}). Thus, axons do not re-route through the grey matter to reach functional targets in the newt. This suggests that white matter inhibitors \\[[@B17]\\] either are not present in the newt spinal cord during axon regeneration or that newt axon regrowth is not inhibited by these molecules.\n\nExtracellular matrix of the injured newt spinal cord\n----------------------------------------------------\n\nThe ECM associated with an SCI in mammals is inhibitory or repulsive for axon growth cone migration \\[[@B2]\\]. Chondroitin sulfate proteoglycans (CSPGs) are a major component of this inhibitory ECM. After injury, they are expressed by a variety of cell types, including astrocytes and meningeal cells, and they are expressed in the glial scar and glia limitans that forms at the border between the injury site and the spinal cord. They are also upregulated in the spinal cord itself in a gradient that increases as axons approach the injury site \\[[@B18],[@B19]\\]. ECM proteins that are canonically permissive for growth cone migration, such as FN \\[[@B10],[@B20]\\], laminin (LM) \\[[@B21]\\] and collagen (Col) \\[[@B22]\\], as well as ECM proteins with more ambiguous effects, such as tenascin C (TN-C) \\[[@B18]\\], are also expressed. Overall, however, the ECM is dense and does not support growth cone motility \\[[@B23],[@B24]\\].\n\nAxon regeneration in the newt, then, may be enabled, in part, because the ECM of the injured newt spinal cord is not inhibitory. Consistent with previous studies, our initial axon regeneration experiments suggest that a dense scar does not form in the lesion (Figure [2E-J](#F2){ref-type=\"fig\"}). To confirm this observation and determine if the ECM in the newt contains permissive rather than inhibitory proteins, we used immunohistofluorescence to analyze the expression of CSPGs, TN-C, FN, LM, and Col in wisping and spiking stage regenerates, the stages in which axons grow across the injury site. We used 2.5- to 3-week regenerates to target these stages (Additional file [11](#S11){ref-type=\"supplementary-material\"}), and axons were labeled either with the axon tracer or an antibody against neurofilament associated protein (3A10).\n\nIn the intact spinal cord, CSPGs are not expressed, though they are expressed in the vertebral body in association with chondrocytes (Figure [4A](#F4){ref-type=\"fig\"}). Surprisingly, CSPGs are expressed in the injured newt spinal cord. They are associated with the meninges and blood vessels (Figure [4B, C](#F4){ref-type=\"fig\"}) and are expressed near wisping axons (Figure [4B, D](#F4){ref-type=\"fig\"}). CSPGs are not found in the grey or white matter of the injured spinal cord and do not form a barrier between the cord and injury site. Thus, astrocytes in the spinal cord do not appear to express CSPGs. CSPGs also do not form a dense scar within the lesion.\n\n![**Wisping axons are associated with loose ECM made up of canonically inhibitory and permissive proteins**. Axons were labeled with 3A10 (B-D, G-H, S-T, V) or the axon tracer (F, J-L, N-P, R) and are shown in magenta. Each ECM protein is shown in green, and nuclei are blue. **(A-D)**Chondroitin sulfate proteoglycan (CSPG) expression in the intact spinal cord (A, cross-section) and in wisping stage regenerates (B-D). (B) is a longitudinal section. (C) and (D) are cross-sections through the terminal vesicle (C), and axons wisping into the injury site (D) from the same animal. Schematic longitudinal sections of the spinal cord in the upper right corner of (C) and (D) show where the section is in relationship to the injury site. **(E-V)**Similarly, the expression of tenascin-C (E-H), FN (I-L), Collagen XII (M-P), laminin (Q-T), and pigment (U, V) is shown. **(W)**A section adjacent to the one shown in (B) treated with chondroitinase ABC (chABC) before incubation with the CS-56 antibody. **(X)**A section adjacent to the one shown in (E) treated only with secondary antibody; the primary antibody was omitted. D, dorsal; V, ventral; R, rostral; C, caudal. Scale bar: 200 \u03bcm (A-X).](1749-8104-6-1-4){#F4}\n\nUnlike CSPGs, TN-C, FN, Col, and LM are expressed in the intact cord in association with the meninges and blood vessels (Figure [4E, I, M, Q](#F4){ref-type=\"fig\"}). After injury all of these proteins are expressed in a pattern similar to CSPGs. They are associated with the meninges and blood vessels (Figure [4F, G, J, K, N, O, R, S](#F4){ref-type=\"fig\"}) and are expressed near wisping axons (Figure [4F, H, J, L, N, P, R, T](#F4){ref-type=\"fig\"}). These proteins are not found in the grey or white matter of the injured spinal cord, do not form a barrier between the cord and injury site, and do not form a dense scar within the injury site. We occasionally detected a weak TN-C signal within the white matter of the spinal cord, but this signal was not significantly above background noise (autofluorescence and non-specific staining). A previous study detected a stronger TN-C signal in the white matter of the intact and regenerating newt tail spinal cord \\[[@B25]\\]. We were unable to reproduce these results, perhaps because we were analyzing a different region of the spinal cord, a different regenerative context, or TN-C expression with different TN-C antibodies.\n\nIn summary, similar inhibitory and permissive ECM proteins are expressed in the injured newt and mammalian spinal cords, but the pattern of expression is different. In the newt, the ECM associated with wisping axons is loose and open, and a dense scar does not form. Also, ECM proteins are not present within the grey or white matter of the spinal cord, and the ECM does not form a permanent barrier, or glia limitans, between the cord and the injury site, though we do have preliminary evidence that such a barrier may exist transiently during the wrapping stage (Additional file [12](#S12){ref-type=\"supplementary-material\"}). Overall, even though inhibitory CSPGs are expressed, the ECM of the injured newt spinal cord does appear to be more permissive for axon growth cone migration than that of the injured mammalian spinal cord.\n\nMeningeal and endothelial cells are associated with wisping axons\n-----------------------------------------------------------------\n\nGiven that cells closely associated with axons growing across the injury site and cells that create the permissive ECM are likely to be playing an important role in enabling axon regeneration, we sought to identify these cells. We hypothesized that they may be meningeal and/or endothelial cells because, in the intact and regenerating spinal cord, ECM appears to be expressed by cells of the meninges and blood vessels, not by cells in the grey or white matter (Figure [4](#F4){ref-type=\"fig\"}). Thus, astrocytes and EG in the spinal cord do not appear to express ECM, at least not around their cell bodies. Our hypothesis is also supported by the fact that in the intact spinal cord, pigmented cells are only found in the meninges and blood vessels (Figure [4U](#F4){ref-type=\"fig\"}), and pigmented cells are associated with wisping axons (Figure [4V](#F4){ref-type=\"fig\"}). Additionally, many cells associated with wisping axons appear to be meningeal cells because they are continuous with the meninges (Figure [2](#F2){ref-type=\"fig\"}). The meninges appear to regenerate ahead of regrowing axons, and some regrowing axons even appear to follow the regenerating meninges (Figure [5D](#F5){ref-type=\"fig\"}; Additional file [5](#S5){ref-type=\"supplementary-material\"}).\n\n![**Meningeal and endothelial cells are associated with wisping axons**. **(A-D)**Axons were labeled with the axon tracer (B, C) or 3A10 (D) and are shown in magenta, von Willeband factor (vWF) is green (A-C) and nuclei are blue. (A) Cross-section through intact spinal cord. (B) Longitudinal section through the SCI of a wisping stage regenerate labeled with anti-FN (white, shown separately in (B\\')) and anti-vWF (green, shown separately in (B\\\")). Arrowheads, cells double-labeled with FN and vWF. (C) Cross-section through axons wisping into the injury site of a wisping stage regenerate. Schematic longitudinal section of the spinal cord in the upper right corner shows where the section is in relationship to the injury site. (D) Single confocal plane of a longitudinal thick section of a wisping/spiking stage regenerate. Arrowhead, meninges (m). **(E-I)**Longitudinal section through a wisping stage regenerate imaged with electron microscopy. (E) Region containing axons wisping ahead of the TV. Asterisks, phagocytic cells. (F) Enlargement of meningeal cells in box F of (E). (G) Enlargement of meningeal-like cell in box G of (E) that is associated with dura mater ECM (ecm). Arrowhead, closed loop formed by meningeal-like cell processes; arrow, process resembling dura mater cell process. (H) Enlargement of meningeal-like cell in box H of (E), rotated 90\u00b0. This cell\\'s processes (arrowheads) wrap around a bundle of axons (ax) cut in cross-section. (I) Enlargement of box I in (H) showing that this cell\\'s processes (arrowheads) are also associated with ECM (ecm). D, dorsal; V, ventral; R, rostral; C, caudal. Scale bars: 200 \u03bcm (A-D); 50 \u03bcm (E); 10 \u03bcm (F-H); 1 \u03bcm (I).](1749-8104-6-1-5){#F5}\n\nTo determine if cells associated with wisping axons and ECM are endothelial cells, we used the endothelial cell marker von Willebrand Factor (vWF), which appears to specifically label blood vessels in the newt spinal cord (Figure [5A](#F5){ref-type=\"fig\"}). In wisping stage regenerates, vWF+ cells can be found in the injury site and near (Figure [5B, B\\\"](#F5){ref-type=\"fig\"}) and ahead of (not shown) regenerating axons. These cells are also associated with FN, though not all of the FN is associated with vWF+ cells (Figure [5B, B\\', B\\\"](#F5){ref-type=\"fig\"}). It has been hypothesized that blood vessels and nerves, because of their close proximity and parallel tracts, guide each other during development \\[[@B26]\\]. To determine whether they might guide one another during newt spinal cord regeneration, we analyzed injury sites in cross-sections. We found that though vWF^+^cells and axons are in the injury site at the same time, they do not need to be in direct contact with each other (Figure [5C](#F5){ref-type=\"fig\"}). Therefore, regenerating blood vessels and axons do not appear to guide one another across the injury site, at least not through a contact-mediated mechanism.\n\nMany cells in the injury site were not vWF^+^, and much of the FN was not associated with vWF^+^cells. We, therefore, sought to determine, more conclusively, if any of the cells closely associated with wisping axons and ECM were meningeal cells. Meningeal cell markers such as retinaldehyde dehydrogenase 2 (RALDH2) and NG2 \\[[@B27]\\] did not produce a signal in our preparations (Additional file [13](#S13){ref-type=\"supplementary-material\"}) so we used electron microscopy (EM) to analyze cell types based on ultrastructure. In the intact and regenerating spinal cords, meningeal cells, in general, have an electron-dense cytoplasm filled with mitochondria, an elongated nucleus with some clumped chromatin, and thin dark processes that are often associated with ECM (Figure [5E, F](#F5){ref-type=\"fig\"}; Additional file [14](#S14){ref-type=\"supplementary-material\"}). None of the other cell types in the intact cord shared these characteristics of meningeal cells (compare Additional files [14](#S14){ref-type=\"supplementary-material\"} and [15](#S15){ref-type=\"supplementary-material\"}). In regenerating spinal cords, we observed cells with meningeal cell characteristics closely associated with regrowing axons. Figure [5H](#F5){ref-type=\"fig\"} shows an enlargement of one of these cells. It has an electron dense cytoplasm filled with mitochondria, an elongated nucleus with some clumped chromatin, and thin dark processes that appear to wrap around a bundle of axons that has been cut in cross-section. Processes from this cell are also associated with ECM (Figure [5I](#F5){ref-type=\"fig\"}). Another meningeal-like cell (Figure [5G](#F5){ref-type=\"fig\"}) was associated with dura mater-like elements, namely ECM and fine, dark cell processes (Additional file [14B](#S14){ref-type=\"supplementary-material\"}). Processes from the cell in Figure [5G](#F5){ref-type=\"fig\"} form a closed loop, which suggests they may be attempting to wrap or cup around structures in their environment.\n\nIt is difficult to know with certainty the identity of the meningeal-like cells. Stensaas \\[[@B6]\\] suggested that cells with these characteristics were dark astrocytes derived from EG. Morphologically, however, these cells and their processes resemble meningeal cells far more than they do EG and astrocytes (Additional files [15](#S15){ref-type=\"supplementary-material\"} and [16](#S16){ref-type=\"supplementary-material\"}). Although these cells appear to wrap around axons, we and Stensaas \\[[@B6]\\] saw no evidence of them forming myelin, and therefore we do not think they are oligodendrocytes or Schwann cells. Further support for the idea that these cells may be meningeal cells comes from studies of meningeal cells that are transplanted into injured mammalian spinal cords along with olfactory ensheathing cells (OECs). Such meningeal cells appear to wrap axons into fascicles much like fibroblasts in the perineurium of peripheral nerves wrap axons \\[[@B28],[@B29]\\]. These meningeal cells were also associated with ECM and had morphological features similar to our meningeal-like cells.\n\nEpendymoglial and astrocytic processes are associated with wisping axons\n------------------------------------------------------------------------\n\nAs mentioned above, axons are often seen wisping ahead of the central canal, and thus, a pre-formed ependymal tube is not required for axon regeneration after an SCI. To determine if EG still might be playing a role in axon regrowth, we used an antibody against glial fibrillary acid protein (GFAP). In the intact newt spinal cord, GFAP weakly labels EG cell bodies, strongly labels astrocytic cell bodies and strongly labels radial processes from both these cell types (Figure [6A](#F6){ref-type=\"fig\"}). Therefore, this cell marker enabled us to observe the astrocytic as well as EG response.\n\n![**Glial processes are associated with wisping axons and do not form inhibitory barriers**. Axons were labeled with 3A10 (B, D) or the axon tracer (E) and are shown in magenta. Glia were labeled with anti-GFAP (green), and nuclei are blue (A-D). **(A)**Cross-section through intact spinal cord. GFAP labels EG and astrocytes and all of their processes. **(B)**Longitudinal section through the SCI of a wisping stage regenerate. Astrocytes do not become hypertrophic (compare with (A)). **(C)**Section adjacent to (B) treated only with secondary antibodies; primary antibodies were omitted. **(D)**Longitudinal section through the SCI of a wisping stage regenerate. **(D\\')**Enlargement of dotted box in (D). GFAP^+^processes (arrows) are associated with wisping axons and do not form a glia limitans. Only a few weakly GFAP^+^cells (arrowhead) are in the injury site. **(E)**Single confocal plane of a longitudinal thick section of a growth beyond stage regenerate. GFAP^+^processes run parallel to axons across the injury site. D, dorsal; V, ventral; R, rostral; C, caudal. Scale bars: 200 \u03bcm (A-D; (B) and (C) are the same scale; (D) and (E) are the same scale); 50 \u03bcm (D\\').](1749-8104-6-1-6){#F6}\n\nIn the injured spinal cord, GFAP expression is maintained in astrocytes and EG around the central canal. It does not appear to be upregulated, and GFAP^+^processes do not appear to be thicker than they are in the intact cord (compare Figure [6A](#F6){ref-type=\"fig\"} and [6B](#F6){ref-type=\"fig\"}). Therefore, astrocytes do not appear to become hypertrophic. In the injury site, only a few GFAP^+^cell bodies can be found (Figure [6D, D\\'](#F6){ref-type=\"fig\"}). These cells are more likely to be EG than astrocytes because they express GFAP only weakly. Thus, astrocytes do not migrate into the lesion. In wisping stage regenerates, GFAP^+^processes are closely associated with wisping axons (Figure [6D, D\\'](#F6){ref-type=\"fig\"}). These processes likely originate from EG and astrocytes that surround the central canal. Rather than being aligned perpendicular to axons, as they are during tail regeneration, however, the GFAP^+^processes are aligned more parallel to the direction in which axons travel. This is especially evident in later-stage regenerates (Figure [6E](#F6){ref-type=\"fig\"}). Note also that the GFAP^+^processes do not form a glia limitans between the cord and the injury site as they do in mammals (Figure [6B, D](#F6){ref-type=\"fig\"}).\n\nThese data demonstrate that, while the EG response after SCI appears to be different to that after tail amputation, EG are still likely to be playing an important role in enabling axon regeneration because GFAP^+^processes are closely associated with axons growing across the injury site. Also, while EG do not appear to migrate into the injury site in large numbers, there are a few that appear to do so.\n\nFurthermore, the astrocytic response in the newt is very much different to that seen in mammals. Newt astrocytes do not appear to become hypertrophic, migrate into the injury site, express ECM, or form a glia limitans or scar. Instead of inhibiting axon regeneration as they do in mammals, newt astrocytes may, like EG, play a supportive role because many of the GFAP^+^processes associated with axons likely arise from astrocytes as well as EG.\n\nAn inflammatory response is present but does not appear to be detrimental to regeneration\n-----------------------------------------------------------------------------------------\n\nIt has been hypothesized that inflammation leads to scarring and non-regenerative wound healing and that it is not present in systems that regenerate \\[[@B30]\\]. In support of this, macrophages in the injured mammalian spinal cord contribute to the formation of the scar \\[[@B31]\\] and have even been shown to physically interact with axons to cause retraction \\[[@B32]\\].\n\nTo determine if an inflammatory response might be present in the newt SCI, we used H&E staining. With H&E staining, an inflammatory response is characterized by the appearance of many small round cells that have dark blue nuclei and very little cytoplasm. Lymphocytes have almost no cytoplasm, whereas monocytes have a small rim of cytoplasm. Based on these characteristics, an inflammatory response does appear to be present in wisping stage regenerates (n = 5; Figure [7A, B, B\\'](#F7){ref-type=\"fig\"}). Figure [7B\\'](#F7){ref-type=\"fig\"} highlights some of the more obvious examples of lymphocytes and monocytes that are present at this stage. These types of cells are associated with the fibrin clot that forms in the injury site and can be identified as early as one week after injury (Additional file [17A, B-B\\\"](#S17){ref-type=\"supplementary-material\"}).\n\n![**An inflammatory response is present but not detrimental to axon regeneration**. **(A)**Longitudinal section through a wisping stage regenerate. Axons were labeled with 3A10 (magenta), nuclei are in blue, and fluorescent channels were laid over the phase image. **(B)**A section adjacent to (A) stained with H&E. Many small cells with dark blue nuclei are present near wisping axons. **(B\\')**Enlargement of blue box in (B). Lymphocytes (arrowhead) and monocytes (arrows) can be identified near wisping axons. **(C-E)**Longitudinal section through a wisping stage regenerate imaged with EM. (C) Region containing axons regenerating into the injury site. Many phagocytic cells (asterisks), or macrophages, can be identified near regenerating axons. (D) Enlargement of box D in (C). Processes from this macrophage (m) are in direct contact with regenerating axons (ax). (E) Enlargement of box E in (C). Another example of a meningeal cell-like process (p) that is mingled and closely associated with regenerating axons (ax). ecm, ECM; R, rostral; C, caudal. Scale bars: 200 \u03bcm (A, B); 50 \u03bcm (B\\',C); 5 \u03bcm (D); 1 \u03bcm (E).](1749-8104-6-1-7){#F7}\n\nTo confirm the presence of various types of inflammatory cells, we again tried several cell-type specific antibodies, but these did not produce a reliable signal in our preparations (Additional file [13](#S13){ref-type=\"supplementary-material\"}). We were, however, able to identify many phagocytic cells, or functional macrophages, in the region of regenerating axons in our EM images. Asterisks in Figures [5E](#F5){ref-type=\"fig\"} and [7C](#F7){ref-type=\"fig\"} mark obvious macrophages. These macrophages could even be seen to be in direct contact with axons (Figure [7D](#F7){ref-type=\"fig\"}). Figure [7E](#F7){ref-type=\"fig\"} shows another example of a meningeal-like cell process that is closely associated with regenerating axons as well as ECM.\n\nThese data demonstrate that an inflammatory response does appear to be present in the injured newt spinal cord, but does not appear to be detrimental to the regenerative response. In fact, one of the most severe cases of inflammation was seen in a contact stage regenerate (Additional file [17C,C](#S17){ref-type=\"supplementary-material\"}\\'). The inflammation, which was likely present at earlier stages as well, did not prevent regeneration from progressing to this advanced stage. The interaction of newt macrophages with axons, if it also causes axon retraction in newts, does not prevent the net response to be that of axon growth. Perhaps the beneficial effects the immune system can have on regeneration \\[[@B33]\\] outweigh any detrimental effects in the newt.\n\nConclusions\n===========\n\nAxons appear to be able to regenerate after an SCI in the newt in part because the environment of the injury site is not inhibitory. Instead of forming a dense inhibitory scar and interacting to form a glia limitans and as they do in mammals \\[[@B2],[@B9],[@B10]\\], newt meningeal cells and glia appear to create a permissive environment for axon regeneration. Figure [8](#F8){ref-type=\"fig\"} summarizes the stages of axon regeneration and the roles these cell types play. After an SCI, axons initially retract from the end of the cut cord and appear dystrophic, but then initiate growth by about 1 week. Axons wrap around the end of the TV before growing into the lesion, perhaps because meningeal cells create a transient basal lamina. Meningeal and endothelial cells lead the way across the lesion and are associated with a loose extracellular matrix that permits growth cone migration. Axons grow into the injury site next and are closely associated with meningeal cells and glial processes that extend from EG and astrocytic cell bodies surrounding the central canal. Later in the process, ependymal tubes lined with glia extend into the lesion as well. Then, as a unit, meningeal cells, axons, and glia close the gap in the spinal cord. Finally, axons enter the cord on the opposite side of the injury and travel through white matter to reach functional targets. We also noted that though ascending axons do regenerate, sensory axons do not appear to be among them and that, overall, this regenerative process can occur in the presence of an inflammatory response.\n\n![**Cellular model of newt axon regeneration**. Schematics of longitudinal sections through the spinal cord are shown. See text for details. et, ependymal tube.](1749-8104-6-1-8){#F8}\n\nModel\n-----\n\nBased on our observations, we propose a glia-meningeal cell interaction model of spinal cord regeneration: spinal cord regeneration is enabled when meningeal cells and glia (astrocytes and EG) interact to form a conduit for axon regeneration rather than an inhibitory barrier (scar, glia limitans, and dense ECM). Simpson \\[[@B34]\\] hypothesized that successful spinal cord regeneration depends upon an interaction between mesenchymal cells and the ependymal epithelium that stimulates the ependyma to proliferate and form a scaffold, which guides axons across the lesion. Since astrocytes also appear to be an important component in this interaction, we have included astrocytes with the ependymal cells and refer to them as glia. Simpson postulated that blastemal cells are the important mesenchymal cell after tail amputation, but was not specific about which cells play that role after SCI. We suggest that meningeal cells are an important mesenchymal cell after SCI and that they, too, can provide a scaffold for axon growth and guidance.\n\nGlia\n----\n\nOur model of spinal cord regeneration after an SCI is similar to that after tail amputation in that EG play a central role. Their processes are closely associated with regenerating axons and may provide a substratum for growth cone migration and guidance via cell-cell adhesion. Our data also suggest that astrocytes may play a similar role. These astrocytes may be immediate descendants of EG and not the fully mature astrocytes seen in the mammalian spinal cord \\[[@B35]\\], and this may explain, in part, why they respond so differently to SCI. The arrangement of the glial processes, however, is different from the tail regeneration model, at least initially. Instead of extending radially from a pre-formed ependymal tube to form channels through which axons can grow \\[[@B3]\\], they extend longitudinally ahead of the TV and run roughly parallel to the regenerating axons. Such glial support may allow the axons to grow into the injury site in the absence of a pre-formed tube. They may also help guide the EG lining the central canal, and thus the ependymal tube, into the region of wisping axons, similar to the way in which radial processes guide newborn neurons to their proper locations in the developing cortex. Soma may translocate as their processes shorten, or they may migrate along the processes of other EG \\[[@B36]\\]. Given that very few cells in the injury site express GFAP, we do not think many EG undergo an epithelial to mesenchymal transition to migrate freely into the lesion as proposed for axolotls after SCI \\[[@B7]\\]. This does not rule out the possibility, however, that undifferentiated progeny of EG might migrate into the injury site. Alternatively, given that the mesenchymal cells in the axolotl injury site were also GFAP^-^and were associated with FN, such cells could be meningeal and/or endothelial cells.\n\nMeningeal cells\n---------------\n\nUnlike most previous models of salamander spinal cord regeneration, our model assigns an important role to meningeal cells. Although they are thought to be detrimental to axon regeneration in the injured mammalian spinal cord, meningeal cells are able to stimulate, support, and guide developmental and regenerative processes in other contexts. Meningeal cells are fibroblast-like, and fibroblasts in flesh wounds are responsible for remodeling the ECM into either scar or normal tissue. The difference appears to depend on whether the fibroblasts are stimulated by transforming growth factor (TGF)\u03b21 or TGF\u03b23 \\[[@B37]\\]. As mentioned previously, meningeal cells harvested from the olfactory bulb, a part of the nervous system that does regenerate, and transplanted into a mammalian SCI along with OECs appear to fasciculate axons into perineurial-like structures and enhance the ability of OECs to remyelinate axons \\[[@B28],[@B29]\\]. During development, the meninges serve as substratum for and secrete a chemoattractant, CXCL12, which regulates the migration of Cajal-Retzius cells \\[[@B38]\\]. In the developing forebrain, the meninges secrete retinoic acid, which stimulates radial glia to begin producing neurons \\[[@B39]\\]. Meningeal cells in mammalian SCI express retinoic acid and permissive ECM proteins, such as LM and FN \\[[@B27],[@B40]\\]. Retinoic acid has been shown to stimulate and possibly be an attractant for neurite outgrowth \\[[@B41],[@B42]\\]. While the permissive ECM proteins expressed by meningeal cells *in vitro*tend to form a basal lamina that axons cannot cross, axons appear to grow quite well along it \\[[@B40]\\]. Thus, the basal lamina can either guide or inhibit axon regeneration depending on its orientation. In fact, following a large ablation of the spinal cord in which the meninges were left intact, axons were seen regenerating along the basal lamina of the meninges, which was oriented in a productive direction parallel to axon regrowth \\[[@B6]\\].\n\nAxons\n-----\n\nOur model does not exclude the possibility that axons and axon intrinsic properties also play an important role in newt spinal cord regeneration. The 1 week required for growth initiation suggests that an intrinsic growth potential is activated. This is supported by the fact that a 1-week conditioning lesion encouraged neurite outgrowth from newt retinal explants \\[[@B43]\\]. Activation of intrinsic growth could also enable axons to overcome any inhibition CSPGs might inflict, similar to the way in which embryonic neurons adapt to CSPGs \\[[@B44],[@B45]\\]. Furthermore, matrix metalloproteinases are upregulated during newt spinal cord regeneration (unpublished observations), and if they are expressed in growth cones, they could breakdown inhibitory ECM molecules and allow the axons to grow through regions that would normally be non-permissive. Axons may also directly influence the behavior of the cells in their environment. Nerves are required for newt limb regeneration, and they stimulate Schwann cells to produce anterior gradient protein, a factor that is sufficient to rescue regeneration in de-nervated limbs \\[[@B46]\\]. Regenerating optic nerve axons in the frog appear to stimulate a breakdown of the blood brain barrier at the front of regenerating axons \\[[@B47]\\]. Finally, when degenerated optic nerve fragments containing hypertrophic astrocytes are implanted into an optic nerve injury site in frogs, axons grow into the grafted tissue and appear to stimulate the reactive astrocytes to adopt a more permissive architecture: the astrocytes extend radial processes and form compartments for axon elongation \\[[@B48]\\]. Thus, newt axons may play a role in stimulating appropriate glial and meningeal cell responses.\n\nIf meningeal cells and glia support the development and regeneration of the nervous system in some contexts, why do they not do so in all contexts? Although no definitive answer can be put forward, we suspect that the state of differentiation of the cells may play a key role. In the newt spinal cord, EG maintain characteristics of radial glia of the developing cortex \\[[@B49]\\], astrocytes may be recent progeny of EG \\[[@B35]\\], and many differentiated cells have the ability to dedifferentiate and become more progenitor-like \\[[@B50]\\]. An exciting new study in mouse suggests that mammalian glia, and possibly mammalian meningeal cells, may retain the capacity to respond to an SCI in ways that support axon regeneration. When the tumor suppressor gene *PTEN*(phosphatase and tensin homolog) is deleted in mouse corticospinal neurons, the neurons are able to regenerate axons across an SCI, and their axons cross the injury site in close association with GFAP^+^cells and cell processes \\[[@B51]\\]. These GFAP^+^bridges tended to appear around the edges of the spinal cord, the regions that are closest to the meninges. Learning more about how newt cell types are intrinsically different from their mammalian counterparts and what factors stimulate newt cell types to respond and interact in ways that promote axon regeneration may help us develop methods to cultivate similar behaviors in mammalian cells. Identifying effective treatments for spinal cord injury in mammals has been so intractable that it is imperative we take full advantage of the opportunity to learn how nature has already solved this problem in newts.\n\nMaterials and methods\n=====================\n\nAnimals\n-------\n\nAdult newts, *N. viridescens*, were purchased from Charles D Sullivan Co. Inc. (Nashville, TN, USA), housed at 22\u00b0C in glass aquariums equipped with water filters, and fed live blackworms, *Lumbriculus varietgatus*(Eastern Aquatics, Lancaster, PA, USA). All animal protocols were approved by the University of Utah Institutional Animal Care and Use Committee.\n\nSpinal cord injury\n------------------\n\nNewts were anesthetized by submersion in 0.1% Tricaine (ethyl 3-aminobenzoate methanesulfonate salt; Sigma, St. Louis, MO, USA) in 20 mM Tris-HCl, pH 7.5 for 10 minutes and placed in ice for 30 minutes prior to surgery. Complete transection injuries were performed similar to those described in Davis *et al*. \\[[@B1],[@B12]\\]. After swabbing the skin with 10% providone iodine, a deep incision was made between vertebral bones about 1 cm rostral to the hindlimbs. Using forceps, the vertebrae were gently and slightly separated and the dorsal lamina of the caudal vertebra was removed to expose the spinal cord. The spinal cord was cut completely with fine spring scissors. The completeness of the transection was easily verified by visual inspection (Figure [1C](#F1){ref-type=\"fig\"}). Blood was irrigated with sterile 85% PBS when necessary. Animals were kept moist during the procedure, placed on ice for 15 minutes after the surgery, and allowed to recover in bins that were tilted to produce a shallow area of water. Bins were cleaned and filled with fresh de-chlorinated water every day or every other day throughout the recovery period.\n\nAxon tracer application\n-----------------------\n\nOur procedure was modified from that used in zebrafish \\[[@B15],[@B16]\\]. Three axon tracers were tried: BDA (3 kDa, lysine fixable, Invitrogen, Carlsbad, CA, USA, D7135), RDA (3 kDa, lysine fixable, Invitrogen, D3308), and biocytin (Pierce, Rockford, IL, USA, \\#28022). RDA appeared to work as well as BDA, but was only used when two tracers were needed in the same animal. Biocytin, which was reported to reveal thin processes better than a 10 kDa RDA \\[[@B15]\\], unfortunately did not produce a signal in our preparations, perhaps because of the decalcification step in tissue processing (see below). BDA and RDA were prepared similarly: small pieces of gel foam (Pharmacia and Upjohn Co., Kalamazoo, MI, USA) were soaked in 1.5 \u03bcl of a 10% solution and allowed to air dry 40 minutes in a tissue culture hood to further concentrate the solution. Fast green dye (Fisher, Waltham, MA, USA, F-99) was added to the BDA solution at 0.05 mg/ml to make it visible *in vivo*. Biocytin did not readily go into solution at 10%, but a slurry was made. Small pieces of gel foam were soaked in 10 \u03bcl of the slurry, air dried completely, and stored at -20\u00b0C. To apply tracer to the spinal cord, the cord was transected one vertebral segment (about 2 mm) rostral or caudal to the original injury site using the same procedure described above and a piece of gel foam saturated with tracer was inserted into the gap. To apply tracer to the sciatic nerve, an incision was made along the course of the sciatic blood vessel on the posterior aspect of the hindlimb where the hindlimb joins with the body. Fine forceps were used to tease the nerve from the vessel, and the nerve was transected with fine spring scissors. A piece of gel foam saturated with tracer was applied and the wound was sealed closed with Vetbond tissue adhesive (3M, St. Paul, MN, USA). Excessive bleeding was controlled before tracer application, if necessary, by inserting a piece of gel foam into the wound and allowing the animal to sit on ice for several minutes. Use of this technique was minimized, however, because it decreased the efficiency of tracer uptake.\n\nTissue harvest, fixation, and decalcification\n---------------------------------------------\n\nAnimals were anesthetized as above, placed on ice for at least 15 minutes, then perfused with about 1.5 ml 85% PBS followed by 3 ml periodate-lysine-paraformaldehyde (PFA) fixative (PLP: 75 mM lysine, 10 mM sodium periodate, 0.5% PFA in 85% PBS, pH 7.4; 0.1 M lysine and 2% PFA stocks were stored at -20\u00b0C and ingredients were mixed and used within 2 hours) \\[[@B52]\\]. All incubations were performed rocking at room temperature (RT) and rinses were for 20 to 30 minutes, three times, unless otherwise specified. The spinal column was harvested in about 4 mm long segments, post-fixed in PLP for 2 hours, rinsed with PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na~2~HPO~4~, 1.15 mM KH~2~PO~4~, pH 7.4), rinsed in fresh PBS overnight at 4\u00b0C, decalcified with Morse\\'s solution (22.5% formic acid, 10% sodium citrate) for 24 hours, and rinsed with PBS. Some of the spinal cords prepared for paraffin sections were not perfusion fixed and this did not seem to affect the results, though perfusion did make tissue prepared for thick and frozen sections easier to section. An alternative fixative, 4% PFA, was tried, but the quality of sections was not improved, there was more autofluorescence, and some antibodies did not work as well.\n\nThick sections and fluorescent labeling (spinal cord whole-mounts)\n------------------------------------------------------------------\n\nMethods were adapted from previous studies \\[[@B53],[@B54]\\] and described in Zukor *et al*. \\[[@B55]\\]. All incubations were performed rocking at RT unless otherwise specified. Decalcified tissue was bleached with formamide bleach (6% hydrogen peroxide, 3% formamide, 0.1% Triton X-100 in PBS) for about 2 hours; rinsed with PBS for 20 to 30 minutes, three times; and embedded in 4% agarose. Longitudinal sections containing the whole spinal cord (about 350 \u03bcm thick) were cut on a vibratome and collected in ice cold PBS. Sections were transferred to 0.67 ml centrifuge tubes, permeabilized with PBS-TxD (PBS containing 0.5 to 2% Triton X-100 and 20% DMSO; 0.5% Triton X-100 was sufficient for streptavidin, 2% was necessary for antibodies) overnight and incubated in streptavidin or primary antibody diluted in PBS-TxDB (PBS-TxD containing 1% BSA (EMD, Gibbstown, NJ, USA, \\#2960) and 0.1% fish skin gelatin (Sigma, G7765) as blocking reagents) for 7 days for streptavidin or 10 days for antibody penetration. Sections needing incubation with secondary antibodies were rinsed with PBS-Tx (PBS with 2% Triton X-100) for 30 minutes, five times and incubated in secondary antibody diluted in PBS-TxDB for 10 days. Sections were then rinsed in PBS for 30 minutes, four times; rinsed in TNT (100 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween-20) for 30 minutes; incubated in SYTOX green (Invitrogen, S7020) diluted 1/1,000 in DMSO, then 1/25 in TNT for 3 hours to stain nuclei; and rinsed in TN (TNT without Tween-20) for 10 minutes, three times. Sections were transferred to glass vials, dehydrated through a methanol (MeOH)/TN series (20%, 50%, 75%, 100% MeOH; 15 minutes each); and cleared in BABB (1 part benzyl alcohol, 2 parts benzyl benzoate) overnight at 4\u00b0C (not rocking).\n\nFrozen sections\n---------------\n\nFrozen sections of the injured spinal cord were difficult to obtain presumably because it contains more water than the intact cord and because the spinal column contains tissues of heterogeneous density: the muscle and bone are rigid and dense, while the spinal cord is quite soft. Therefore, methods were adapted from Tokuyasu \\[[@B56]\\]. Perfusion-fixed, decalcified tissue was cryoprotected to 2.3 M sucrose in PBS (3 hours in 25%, 3 hours in 50%, overnight in 75%, 24 hours in 100% 2.3 M sucrose; rocking and at RT) and frozen quickly in OCT (Ted Pella, Redding, CA, USA, \\#27050) with liquid nitrogen to minimize freezing artifacts. Sections 16 to 20 \u03bcm thick were obtained on a cryostat with the chamber temperature set to -35\u00b0C and the object temperature set to -34\u00b0C. For intact spinal cord, variables could be somewhat relaxed. Tissue could be cryoprotected to 1.6 M sucrose and perfusion was not critical. Sections were stored at -20\u00b0C and, prior to fluorescent labeling, thawed, baked onto the slides for 1 hour at 50\u00b0C, and rinsed in PBS for 5 minutes.\n\nParaffin sections\n-----------------\n\nAll incubations were performed with rocking at RT. Decalcified tissue was dehydrated with ethanol (EtOH; 50% for 30 minutes, 70% for 30 minutes, 95% for 30 minutes, 100% for 30 minutes, 100% for 1 hour), infiltrated (75% Hemo-De in EtOH for 20 minutes, 100% Hemo-De for 1 hour, paraffin for 1 hour at 60\u00b0C and 15 Hg vacuum), and embedded in paraffin. Sections 10 to 20 \u03bcm thick were mounted onto slides, left on a 37\u00b0C slide warmer overnight, and stored at 4\u00b0C. Prior to fluorescent labeling or staining, sections were dewaxed (Hemo-De for 10 minutes, 2 times; 75% Hemo-De in EtOH for 5 minutes; 100% EtOH briefly and then for 3 minutes) and rehydrated (95% EtOH for 3 minutes, 70% EtOH for 3 minutes, PBS or water for 5 minutes; PBS was used for fluorescent labeling and water was used for staining).\n\nH&E staining\n------------\n\nParaffin sections rehydrated in water were stained: hematoxylin (Harris modified, with acetic acid; Fisher, Waltham, MA, USA, SH26-500D) 4 minutes, running tap water 5 minutes, 70% EtOH briefly, acidified 70% EtOH (10 drops of concentrated HCl to 200 ml 70% EtOH) 1.5 minutes, bluing solution (10 drops of NH~4~OH in 200 ml EtOH) 1 minute, Scott\\'s solution (1 g sodium bicarbonate, 10 g magnesium sulfate in 500 ml water) 3 minutes, running tap water 5 minutes, 70% EtOH briefly, 0.1% Eosin in 95% EtOH (made from 0.5% Eosin Y Solution, Sigma HT110-1-32) 1 minute, 95% EtOH briefly, 100% EtOH 5 minutes, 75% Hemo-De in EtOH 5 minutes, 100% Hemo-De 5 minutes. Coverslips were mounted with Cytoseal 60 (Fisher, 23-244-256).\n\nFluorescent labeling in paraffin or frozen sections\n---------------------------------------------------\n\nSections were blocked in PBS-TxB (PBS with 0.2% Triton X-100, 1% BSA and 0.1% fish skin gelatin) for 1 hour; incubated with primary antibody diluted in PBS-Tx (PBS with 0.2% Triton X-100) overnight at RT; rinsed with PBS for 10 minutes, three times; incubated with secondary antibody or streptavidin diluted in PBS-Tx for 4 hours; rinsed with PBS for 10 minutes, two times; rinsed with TNT or TN for 10 minutes; incubated with SYTOX green (Invitrogen, S7020) diluted 1/1,000 in DMSO, then 1/50 in TNT or Hoechst 33342 (Invitrogen, A10027) diluted 1/2,000 in TNT for 30 minutes to stain nuclei; rinsed in TN for 10 minutes, 50% TN for 10 minutes, and water for 10 minutes. Coverslips were mounted with Fluoromount G (Fisher, OB100-01) and sealed with clear nail polish. For some experiments on paraffin sections, incubation with primary antibody was extended to 2 or 3 days at RT. This improved signal without adding noise. Sections treated with chondroitinase ABC (chABC) were treated before the blocking step. They were equilibrated in tris acetate buffer (50 mM Tris-HCl, pH 8.5, 60 mM sodium acetate, 0.02% BSA, 0.5% Triton X-100; made fresh, pH 8.0) for 5 minutes, treated with chABC solution (7.5 U/ml chABC stock diluted 1:50 in tris acetate buffer; stock was chABC, Sigma C2905, EC 4.2.2.4, reconstituted in 0.01% BSA) for 1 hour at 37\u00b0C, and rinsed with PBS for 5 minutes, three times.\n\nAntibodies and streptavidins\n----------------------------\n\nPrimary antibodies and dilutions used are listed in Table [2](#T2){ref-type=\"table\"}. Additional primary antibodies that were tried are listed in Additional file [13](#S13){ref-type=\"supplementary-material\"}. Secondary antibodies (Invitrogen, whole antibodies, 2 mg/ml, conjugated to Alexa 488, 568, 594, or 633) were diluted 1/100. Streptavidin-Cy5 (Invitrogen, SA1011) and Streptavidin-Alexa 633 (Invitrogen, S21375) were diluted to 4 \u03bcg/ml. The FN and Col XII antibodies and one TN-C antibody are newt-specific and have been characterized \\[[@B57]-[@B59]\\]. The non-newt-specific antibodies appear to be specific in that they produced appropriate expression patterns and are specific in western blots (not shown, not done for vWF and CSPG antibodies). The specificity of the CSPG antibody was verified by treating adjacent sections with chABC and incubating them with the CSPG antibody. The specificity of the secondary antibodies was verified by treating adjacent sections with secondary antibodies in the absence of primary antibodies. Very little non-specific signal was seen in either case (Figure [4W,X](#F4){ref-type=\"fig\"}), and what little was seen tended to appear in the white matter. The chick-specific TN-C antibody produced an expression pattern similar to the newt-specific antibody, but its signal was much stronger and so it was used instead. The antibody against Col XII was used to approximate the expression of Col I since it is found in association with Col I and may enhance the stability of the ECM by bridging collagen fibrils \\[[@B60],[@B61]\\]. Since paraffin processing may affect CS-56 labeling, CSPGs were analyzed in frozen sections. TN-C, FN, LM, and Col were analyzed in paraffin and frozen sections. Expression patterns were also verified in whole-mount preparations (not shown).\n\n###### \n\nTable of antibodies\n\n  Antigen                                    Antibody type    Company, catalogue number   Dilution, format\n  ------------------------------------------ ---------------- --------------------------- --------------------\n  CSPGs (chick)                              Mouse IgM mAb    Sigma, C8035 (CS-56)        1/100, ascites\n  Tenascin-C (chick)                         Rabbit pAb       Chemicon, AB19013           1/100, concentrate\n  Tenascin-C (newt)                          Mouse IgM mAb    DSHB, MT1                   1/25, concentrate\n  Fibronectin (newt)                         Mouse IgG mAb    DSHB, MT4                   1/50, concentrate\n  Collagen XII (newt)                        Mouse IgG1 mAb   DSHB, MT2                   1/50, concentrate\n  Laminin (mouse)                            Rabbit pAb       Sigma, L9393                1/25, concentrate\n  Fibrin (human)                             Mouse IgG1 mAb   ADI, 350                    1/50, concentrate\n  Neurofilament associated protein (chick)   Mouse IgG1 mAb   DSHB, 3A10                  1/50, supernatant\n  GFAP (cow)                                 Rabbit pAb       Dako, Z0334                 1/200, concentrate\n  von Willebrand factor (human)              Rabbit pAb       Dako, A0082                 1/400, concentrate\n\nADI, American Diagnostica Inc.; DHSB, Developmental Studies Hybridoma Bank; mAb, monoclonal antibody; pAb, polyclonal antibody.\n\nConfocal imaging and image processing\n-------------------------------------\n\nSections labeled with fluorescent dyes were imaged on an Olympus FV300 or FV1000 laser scanning confocal microscope using a 10\u00d7 or 20\u00d7 air objective. Images were processed with Image J, version 1.40 g \\[[@B62]\\] and Adobe Photoshop CS2. Levels were adjusted in Photoshop to maximize the signal to noise ratio but the same adjustments were made to experimental and control images.\n\nPreparation of tissue for electron microscopy\n---------------------------------------------\n\nMethods were adapted from Anderson *et al*. \\[[@B63]\\]. Animals were anesthetized as above, then perfused with about 1.5 ml 85% PBS followed by 3 ml EM fixative (2.5% glutaraldehyde, 1% formaldehyde, 3% sucrose, 1 mM MgSO~4~in 0.1 M cacodylate buffer (CB), pH 7.4). The spinal column was harvested in about 4 mm long segments; post-fixed in EM fixative overnight at RT; decalcified in cold EDTA solution (0.1 M EDTA, pH 7.2 to 7.4 with 4% glutaraldehyde) for 4 days at 4\u00b0C; rinsed in 0.1 M CB for 20 minutes, three times; rinsed in PBS briefly; and embedded in 4% agarose. Longitudinal sections containing the whole spinal cord (about 350 \u03bcm thick) were cut on a vibratome and collected in ice cold PBS. Sections were returned to 0.1 M CB; osmicated (1% OsO~4~in 0.1 M CB) for 1 hour; rinsed in 0.1 M CB for 10 minutes, three times; partially dehydrated (50%, 75% MeOH; 10 minutes each); treated with 1% uranyl acetate in 75% MeOH for 1 hour; fully dehydrated (75%, 85%, 95%, 100%, 100% MeOH; 10 minutes each); infiltrated with resin (100% acetone 10 minutes, two times; 75% resin in acetone overnight; 100% resin 1 hour); embedded in epoxy resin (2 parts dodecenyl succinic anhydride (DDSA, Ted Pella, 18022), 1 part eponate 12 resin (Ted Pella, 18005), 0.06 parts DMP30 accelerator (Ted Pella, 18042)) and cured at 60\u00b0C for 2 days. Sections were re-embedded for longitudinal sectioning, and 80-nm sections were collected onto carbon-coated Formvar films in gold single-hole grids, stained with 5% uranyl acetate for 1 hour and 1% lead citrate for 25 minutes, and imaged at 80 KeV in a JEOL JEM 1400 electron microscope at 5,000\u00d7 magnification. Images were captured digitally on a GATAN Ultrascan 4000 16 megapixel 16-bit camera and mosaicked with *ir-tools*\\[[@B63]\\]. Mosaics were viewed and selections were captured with Viking software.\n\nAbbreviations\n=============\n\nBDA: biotinylated dextran amine; BSA: bovine serum albumin; CB: cacodylate buffer; chABC: chondroitinase ABC; Col: collagen; CSPG: chondroitin sulfate proteoglycan; DMSO: dimethylsulfoxide; ECM: extracellular matrix; EG: ependymoglia; EM: electron microscopy; EtOH: ethanol; FN: fibronectin; GFAP: glial fibrillary acidic protein; H&E: hematoxylin and eosin; LM: laminin; MeOH: methanol; OEC: olfactory ensheathing cell; PBS: phosphate buffered saline; PFA: paraformaldehyde; PLP: periodate-lysine-PFA; RDA: rhodamine dextran amine; RT: room temperature; SCI: spinal cord injury; TN-C: tenascin C; TV: terminal vesicle; Tx: Triton X-100; vWF: von Willebrand factor.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nKAZ conceived of the study with the assistance of SJO, designed and conducted the experiments, analyzed the data, prepared the figures and wrote the manuscript. DTK and SJO helped design the study and experiments, helped interpret the data, and provided comments on the manuscript. SJO is the laboratory principal investigator and oversaw the project. All authors read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\n**Movie 1: recovery of swimming function after spinal cord transection injury**. When an uninjured newt swims, it propels itself forward by pressing its legs close to its side and undulating its body in an S-shaped motion. When it is finished swimming, it brings its legs forward again, perpendicular to its body axis. One day after a complete transection injury, the hindlimbs are completely paralyzed. Four weeks after injury, this same newt is still paralyzed. When it attempts to swim, it does not use its hindlimbs at all. By 7 weeks, this newt has recovered swimming function and swims similarly to an uninjured newt.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**Figure S1: the largest spinal nerves are associated with the S1 and T-1 vertebrae**. **(A)**Close-up view of S1, dorsal aspect. The rib associated with this vertebra articulates with the ilium of the pelvis to form the SI joint. **(B)**Ventral side of spinal column showing vertebrae T-4 to S2/C1. The largest spinal nerves (arrowheads) are associated with T-1 and S1. T-2 is intermediate in size. T-4, T-3 and S2/C1 are small. Dotted circles, approximate location of the spinal ganglia; dotted lines, approximate course spinal nerves take to spinal ganglia. Note that the actual location of the SCI in this animal is one segment rostral (between T-5 and T-4) to the targeted site (between T-4 and T-3). **(C)**Close-up of T-1 and S1 shown in (B). More flesh has been removed to demonstrate that the vertebra associated with the caudal-most large spinal nerve is indeed S1. The rib associated with it articulated with the pelvis. R, rostral; C, caudal.\n\n###### \n\nClick here for file\n\n###### Additional file 3\n\n**Figure S2: retraction and growth initiation stage images seen in z-projections**. The difference between the retraction and growth initiation stages is apparent throughout the whole spinal cord and is not just a function of which z-plane was chosen for presentation. **(A-D)**Images of retraction (A, B) and growth initiation (C, D) shown in Figure [2](#F2){ref-type=\"fig\"}. All are single confocal planes, except (B), which is a z-projection of four planes. **(A\\'-D\\')**Z-projections of all planes through the spinal cord, showing just the axon tracer channel, for the same animals in (A-D). In the retraction stage, fewer axons extend to the end of the cut cord (arrowheads). R, rostral; C, caudal. Scale bars: 200 \u03bcm ((A-D) are the same scale; (A\\'-D\\') are the same scale).\n\n###### \n\nClick here for file\n\n###### Additional file 4\n\n**Figure S3: size of gap in spinal cord at each stage**. Gap size increases before it decreases and is largest during the wrapping stage. Dark lines, median; box, interquartile range (IQR, 25% to 75%); whiskers, most extreme data point that is no more than 1.5 IQR from the box; small circle, outlier. The size of the gap during the wrapping stage is statistically different (asterisk) from that during the retraction stage, with *P*\\< 0.001 (using Bonferroni correction for multiple *t*-tests).\n\n###### \n\nClick here for file\n\n###### Additional file 5\n\n**Movie 2: movie through confocal z-stack of animal shown in**Figure [5D](#F5){ref-type=\"fig\"}**, a wisping/spiking stage regenerate**. The movie begins on the ventral side of the cord and moves in 2-\u03bcm increments through to the dorsal side. Rostral is up. Axons are labeled with 3A10 in magenta, and nuclei are in green. Note the meninges appear to have regenerated across the gap first and that some axons appear to have followed these meninges.\n\n###### \n\nClick here for file\n\n###### Additional file 6\n\n**Movie 3: movie through confocal z-stack of animal shown in**Figure [2J](#F2){ref-type=\"fig\"}**, a spiking stage regenerate**. The spike does not contain an ependymal tube in this animal. The movie begins on the ventral side of the cord and moves in 2-\u03bcm increments through to the dorsal side. Rostral is up. Descending axons are labeled with the axon tracer in magenta and nuclei are in green.\n\n###### \n\nClick here for file\n\n###### Additional file 7\n\n**Movie 4: movie through confocal z-stack of animal shown in**Figure [2I](#F2){ref-type=\"fig\"}**, a spiking stage regenerate**. The spike does contain an ependymal tube in this animal. The movie begins on the ventral side of the cord and moves in 2-\u03bcm increments through to the dorsal side. Rostral is up. Descending axons are labeled with the axon tracer in magenta and nuclei are in green.\n\n###### \n\nClick here for file\n\n###### Additional file 8\n\n**Figure S4: sloppiness and misalignments in axon regeneration**. Axon regeneration has some tolerance for sloppiness, but gross misalignments may result in a delay or failure to recover. All images are montages of single confocal planes of longitudinal thick sections, except (F) which is a cross-section. Axons in all images were labeled with the axon tracer (magenta), and nuclei are green. **(A-A\\\")**A 9-week regenerate in the growth beyond stage that had recovered swimming function. The ependymal tubes have connected (A), some axons appear to decussate (arrowhead in (A\\')) and wrapping axons are still evident (open arrowhead in (A\\\")). **(B)**A 4-week regenerate in the growth beyond stage that had not recovered function. Similar to (A), axons appear to wrap (open arrowhead) and decussate (arrowhead). **(C)**A 9-week regenerate in the growth beyond stage that had not recovered function. A long spike appears to have elongated from the caudal side (dotted lines), and this may indicate that the gap in the cord was large. **(D)**A 6-week regenerate in the spiking stage that had not recovered function. The whole spike appears to be veering off in the wrong direction and does not line up properly with the cord on the other side (dotted lines). Arrow, dorsal root to the spinal ganglia. **(E)**A 9-week regenerate in the growth beyond stage that had not recovered function. Two ependymal tubes (fat arrows) appear to have formed on the rostral side. **(F)**Cross-section through a 9-week regenerate in the growth beyond stage that had not recovered function. Axon regeneration appears to be sloppy in that axons are wisping dorsally towards the tissue wound that was created to gain access to the spinal cord. R, rostral; C, caudal; D, dorsal; V, ventral. Scale bars: 500 \u03bcm ((A-E) are the same scale); 200 \u03bcm (F).\n\n###### \n\nClick here for file\n\n###### Additional file 9\n\n**Movie 5: movie through confocal z-stack of animal shown in Additional file**[8](#S8){ref-type=\"supplementary-material\"}**A-A\\\", a recovered animal**. This is a 9-week regenerate in the growth beyond stage that had recovered function. The movie begins on the ventral side of the cord and moves in 2-\u03bcm increments through to the dorsal side. Rostral is up. Descending axons are labeled with the axon tracer in magenta and nuclei are in green.\n\n###### \n\nClick here for file\n\n###### Additional file 10\n\n**Movie 6: movie through confocal z-stack of animal shown in Additional file**[8](#S8){ref-type=\"supplementary-material\"}E, **a non-recovered animal**. This is a 9-week regenerate in the growth beyond stage that had not recovered function. It begins on the ventral side of the cord and moves in 2-\u03bcm increments through to the dorsal side. Rostral is up. Descending axons are labeled with the axon tracer in magenta and nuclei are in green. Note there are two ependymal tubes on the rostral side.\n\n###### \n\nClick here for file\n\n###### Additional file 11\n\nTable S1: Stages of 2.5- to 3-week regenerates analyzed after initial study.\n\n###### \n\nClick here for file\n\n###### Additional file 12\n\n**Figure S5: a glia limitans-like structure may be present during the wrapping stage**. Longitudinal sections through wrapping stage regenerates. Axons were labeled with the axon tracer in (C, D) and are shown in magenta. Each ECM protein is shown in green, and nuclei are blue. **(A-D)**Col XII (A), FN (B), and TN-C (C) expression wraps around the end of the cord, while LM (D) expression does not. R, rostral; C, caudal. Scale bar: 200 \u03bcm (A-D).\n\n###### \n\nClick here for file\n\n###### Additional file 13\n\nTable S2: Table of other antibodies tested.\n\n###### \n\nClick here for file\n\n###### Additional file 14\n\n**Figure S6: meninges of the intact spinal cord**. Longitudinal section through the intact spinal cord imaged with EM. **(A)**Region containing the meninges. p, pia mater; a, arachnoid mater; d, dura mater; sas, subarachnoid space; ecm, ECM; ef, glial end feet; o, oligodendrocyte. **(B)**Enlargement of box B in (A) showing skinny, dark processes of dura mater cells (arrows) associated with collagen fibrils (c) of the dura mater. **(C)**Enlargement of box C in (A) showing layers of arachnoid cell processes (a), a skinny process from a dura mater cell (arrow), and collagen (c). **(D)**Enlargement of box D in (A) showing a pia mater cell. **(E)**Enlargement of box E in (A) showing neural collagen (nc). **(F)**Enlargement of box F in (A) showing the basement membrane (arrowhead) at the glia limitans of the spinal cord, collagen (c) and neural collagen (nc). R, rostral; C, caudal. Scale bars: 15 \u03bcm (A); 3 \u03bcm (D); 1.5 \u03bcm (B, C, E, F).\n\n###### \n\nClick here for file\n\n###### Additional file 15\n\n**Figure S7: other cell types in the intact spinal cord**. Longitudinal section through the intact spinal cord imaged with EM. **(A)**Region containing the central canal (cc), EG layer (EG) and a portion of the grey mater (gm). **(B)**Enlargement of box B in (A) showing astrocytes. **(C)**Enlargement of box C in (A) showing the cytoplasm of light EG. **(D)**Enlargement of box D in (A) showing the cytoplasm of dark EG. **(E)**A microglial cell. **(F)**An oligodendrocyte. R, rostral; C, caudal. Scale bars: 15 \u03bcm (A, B, F); 5 \u03bcm (E); 3 \u03bcm (D); 1.5 \u03bcm (C).\n\n###### \n\nClick here for file\n\n###### Additional file 16\n\n**Figure S8: glia of the regenerating spinal cord**. (**A-D**) Longitudinal section through a wisping stage regenerate imaged with EM. **(A)**Region containing axons wisping ahead of the TV. **(B)**Enlargement of box B in (A) showing a light glial process (EG~L~) that is associated with regenerating axons (ax). **(C)**Enlargement of box C in (A) showing light (EG~L~) and dark (EG~D~) EG lining the TV. **(D)**Enlargement of box D in (A) showing a dark EG process (EG~D~) associated with axons (ax). R, rostral; C, caudal. Scale bars: 50 \u03bcm (A); 15 \u03bcm (C); 5 \u03bcm (B); 3 \u03bcm (D).\n\n###### \n\nClick here for file\n\n###### Additional file 17\n\n**Figure S9: the inflammatory response in early and late stage regenerates**. **(A,B**) Adjacent longitudinal sections through a 1-week regenerate labeled with an anti-fibrin antibody (A) and stained with H&E (B). A fibrin clot is formed in the injury site (A), and inflammatory cells can already be identified in this clot (B). **(B\\',B\\\")**Enlargement of primed and double primed boxes in (B). Lymphocytes (arrowheads in (B\\')) and monocytes (arrow in (B\\\")) can be identified. Asterisk, fibrin clot. **(C)**Longitudinal section through a contact stage regenerate. A relatively strong inflammatory response has not prevented this animal from progressing to this late stage. **(C\\')**Enlargement of box in (C). Lymphocytes (arrowhead) and monocytes (arrows) can be identified. R, rostral; C, caudal. Scale bars: 200 \u03bcm (A; (B,C) are the same scale); 50 \u03bcm ((B\\',B\\\") are the same scale; C\\').\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nWe thank the following individuals for their valuable contributions to this study: Alejandro S\u00e1nchez Alvarado for a critical reading of the manuscript; Robert Marc, Bryan Jones, Felix Vasquez-Chona, Drew Ferrell, Jia-Hui Yang, Carl Watt, James Anderson for assistance with EM; Larry Stensaas for his expertise in interpreting the EM data; Chris Rodesch and the University of Utah Cell Imaging Core for help with confocal microscopy; Gerald Spangrude for his expertise with the analysis of the inflammatory response; and Namakkai and Vasanthi Rajasekaran for technical assistance with the western blots. Several antibodies were provided by the Developmental Studies Hybridoma Bank, developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. This work was supported by a grant from the Craig H Neilsen Foundation to SJO. KAZ was supported by a NIH Developmental Biology Training Grant (5T32 HD07491). DTK was supported by a NIH Cardiology Fellowship Training Grant (5T32 HL07576).\n"}
+{"text": "**(See the Editorial Commentary by van der Werf on pages 835--6.)**\n\nBuruli ulcer (BU), a neglected tropical disease caused by *Mycobacterium ulcerans,* is a chronic skin and soft tissue infection that can lead to permanent disfigurement and disability. It is a poorly understood mycobacterial disease, mostly affecting rural populations in West and Central Africa \\[[@CIT0001], [@CIT0002]\\].\n\nIn most BU-endemic settings, a diagnosis is made on clinical and epidemiological grounds, after which treatment with an 8-week course of rifampicin plus streptomycin or clarithromycin is initiated empirically \\[[@CIT0003]\\], pending microbiological confirmation where available. BU presents with a diverse range of clinical signs ranging from nodules or edematous plaques to (painless) ulcerations and, rarely, bone involvement. Every clinical expression of BU can be mistaken for another skin condition \\[[@CIT0004]\\]. Moreover, patients who are not microbiologically confirmed respond better to treatment with antimycobacterial antibiotics, which also cover other bacteria \\[[@CIT0005]\\].\n\nBuruli ulcer incidence is currently declining in several endemic countries such as Benin, Ivory Coast, and Democratic Republic of Congo. In 2016, 1864 BU patients were reported to the World Health Organization (WHO) by 10 countries, whereas \\>5000 cases were reported in 2009 \\[[@CIT0006]\\]. It has been suggested that this decline is, at least in part, due to the introduction of control strategies \\[[@CIT0006]\\]. The decline in incidence resulted in a decrease in BU lesions relative to non-BU lesions treated in BU facilities \\[[@CIT0007]\\]. Clinical expertise in the recognition of BU is thus likely to wane, potentially resulting in more diagnostic misclassification. Therefore, laboratory confirmation of BU becomes essential as this would allow improved patient management by initiating more specific therapy in non-BU patients and limiting the prolonged course of rifampicin and streptomycin/clarithromycin to those who truly have BU.\n\nCurrently, laboratory diagnostics for BU include culture, direct smear examination (DSE) for acid-fast bacilli, histopathology, and polymerase chain reaction (PCR) targeting the insertion element IS*2404* \\[[@CIT0008], [@CIT0009]\\]. Studies on the accuracy of diagnostic tests of BU have used varying reference standards including histopathology \\[[@CIT0010]\\], PCR \\[[@CIT0011], [@CIT0012]\\], at least 1 positive test result \\[[@CIT0013]\\], a composite reference standard of several tests \\[[@CIT0012], [@CIT0014]\\], and latent class analysis (LCA) \\[[@CIT0015]\\]. This diversity in reference standards makes it difficult to summarize findings on diagnostic test accuracy \\[[@CIT0008]\\].\n\nThe WHO recommends that clinically suspected patients are confirmed by at least 1 of the above-mentioned laboratory tests \\[[@CIT0016]\\]. Others have proposed a stepwise approach starting with DSE, followed by PCR only in DSE-negative patients to reduce costs \\[[@CIT0017]\\]. Another alternative is a Buruli score based on clinical signs and demographic patient characteristics \\[[@CIT0015]\\], where only patients with an intermediate score are assessed by PCR.\n\nAn accurate diagnostic algorithm that is cost-effective, reduces time to diagnosis, and is feasible in the remote and resource-limited settings where BU is endemic has great potential to reduce misclassification and improve the management of patients presenting with BU-like skin lesions. We therefore aimed to determine the diagnostic accuracy of clinical and microbiological signs in consecutively recruited patients presenting with lesions clinically compatible with BU in a BU-endemic, low-income setting.\n\nMETHODS {#s5}\n=======\n\nStudy Setting and Study Population {#s6}\n----------------------------------\n\nConsecutive patients with lesions compatible with BU (ulcers, nodules, edema or plaques) presenting between March 2012 and March 2015 at the Centres de D\u00e9pistage et de Traitement de l**'**Ulc\u00e8re de Buruli (CDTUB) of Allada and Lalo and in 10 health posts of the commune of Z\u00e8 in southern Benin, living in BU endemic villages, were eligible for study participation. Eligible patients identified at the health posts were referred to the CDTUB for further assessment. Patients presenting with a recent (\\<2 weeks) wound of obvious noninfectious origin (eg, trauma) or with a wound of \\>2 weeks with normal healing were excluded from the study.\n\nData Collection {#s7}\n---------------\n\nAll patients were assessed clinically by a triage nurse at the participating CDTUBs or the supervising nurse at the health posts. The nurses' classification as BU or non-BU based on the WHO diagnostic criteria (young age, residence in an endemic area, undermined edges, location on limbs, necrosis, absence of pain, adenopathy or fever, and hyperpigmented edges) was verified by a clinician, based on a combination of the WHO criteria and clinical experience \\[[@CIT0003]\\].\n\nSamples for mycobacterial analysis were collected at the CDTUB sites. From each ulcerated lesion, 2 swabs were taken from the undermined edges. From nonulcerative lesions, 2 fine-needle aspirates were taken from the center of the lesion. One sample was processed immediately for DSE after Ziehl-Neelsen (ZN) staining. The second sample was placed in a semisolid transport medium \\[[@CIT0011]\\], and stored at 4\u00b0C until weekly shipment to the Mycobacteriology Reference Laboratory (LRM) in Cotonou, Benin. At the LRM, technicians performed DSE after auramine staining using fluorescence microscopy \\[[@CIT0020]\\], IS*2404* real-time PCR \\[[@CIT0021]\\], and in vitro culture for *M. ulcerans* \\[[@CIT0011], [@CIT0016]\\].\n\nOne 4-mm punch biopsy was taken from every lesion and stored in 10% formalin until embedding in paraffin. One section was stained by hematoxylin-eosin and 1 section by ZN. Histopathological reading was done at the University of Parakou (Benin) and the hospital of Chamb\u00e9ry (France). Both histopathologists were blinded to clinical information (except for age, gender, and type and localization of the lesion) and the results of other diagnostic tests. A standardized reading form was used to score histological changes ([Supplementary Data 1](#sup1){ref-type=\"supplementary-material\"}). Based on the score, specimens were classified as probable BU (score 7**--**15), compatible with BU (score 4--6), or not compatible with BU (score \u22643). Discordant classifications between the 2 histopathologists were reread and discussed during face-to-face meetings to reach consensus scores that were used in the analysis.\n\nData Management {#s8}\n---------------\n\nAll data were coded and registered in a Microsoft Access database. A dedicated nurse registered all clinical data while a dedicated laboratory technician registered all laboratory data.\n\nA random selection of the DSE readings at both CDTUBs was controlled quarterly by the LRM, which participated in the external quality assurance program for *M. ulcerans* PCR organized by the Antwerp Institute of Tropical Medicine (ITM) \\[[@CIT0022]\\].\n\nData Analysis {#s9}\n-------------\n\n### Index Tests {#s10}\n\nThe accuracy of 6 index tests was evaluated: clinical diagnosis verified by the treating clinician, DSE after ZN staining, DSE after auramine staining, PCR, culture, and histopathology.\n\n### Reference Standard {#s11}\n\nBecause a gold standard without error or uncertainty is not available for BU \\[[@CIT0023], [@CIT0024]\\], the accuracy of each test was estimated using an expert panel approach in the primary analysis \\[[@CIT0024]\\], and PCR, the best currently available diagnostic test, in a secondary analysis \\[[@CIT0015]\\].\n\nThe expert panel approach was based on a stepwise evaluation of all clinical information and laboratory results. The expert panel consisted of 8 study team members and 3 independent dermatologists. The expert panel determined the final diagnosis (status) of each patient in 3 steps. In the first step, patients were classified as confirmed BU (positive by PCR and/or histopathology \\[score \u2265 7\\]), possible BU (PCR-negative cases that were positive by DSE, clinically suspected of BU, or had a histopathology score between 4 and 6), or non-BU (negative by all microbiological tests, clinically not suspected of BU and with a histopathology score \u2264 3). In the second step, the expert panel reviewed all patient files classified as possible or non-BU in step 1. Each expert panel member independently made a differential diagnosis for every patient based on all available clinical, demographic, epidemiological, histological and mycobacteriological information ([Supplementary Data 2](#sup1){ref-type=\"supplementary-material\"}) as well as clinical photographs (missing for 31 patients). In the third step, all cases for whom there was disagreement regarding the classification in step 2 were discussed during 1 of 3 face-to-face meetings until consensus was reached. During these discussions, more weight was given to the opinion of the independent dermatologists compared to those panel members who were also part of the study team. Participants for whom the expert panel failed to reach a consensus were classified as BU.\n\n### Accuracy of Diagnostic Indicators {#s12}\n\nThe accuracy estimates were calculated as sensitivities, specificities, positive predictive value (PPV), and negative predictive value (NPV). Combined accuracy estimates were also calculated. For histopathology, the accuracy estimates were determined at the cutoff scores for compatible with BU (score 4) and probable BU (score 7) and at the score with an optimal combination of sensitivity and specificity on the receiver operating characteristic (ROC) curve.\n\nThe accuracy of 3 diagnostic approaches was then estimated: the WHO recommendation to consider patients who are positive by at least 1 laboratory test as confirmed BU; the stepwise approach to reserve PCR to DSE-negative patients; and the clinical Buruli score that is followed by PCR only when patients have an intermediate score \\[[@CIT0015]\\] ([Supplementary Data 3](#sup1){ref-type=\"supplementary-material\"}). Three of the characteristics that make up the Buruli score (yellow and green color, and lesion hyposensitivity) were not available in our dataset. The Buruli score was only tested on ulcerative lesions as 2 of its components (characteristic smell and undermined edges) are unavailable for nonulcerative lesions.\n\nNext, since diagnostic accuracy estimates can vary across patient subgroups, we evaluated effect modification by study site, type of lesion, recruitment type, time since start of study, availability of photographs, patient delay before consultation, transport time of samples to the LRM, and HIV status.\n\nStatistical Analysis {#s13}\n--------------------\n\nAll analyses were performed in R version 3.3.2. through RStudio version 1.0.136.\n\nThe standards for reporting diagnostic studies (STARD \\[[@CIT0027]\\]) were followed while writing this manuscript ([Supplementary Data 14](#sup1){ref-type=\"supplementary-material\"}).\n\nEthical Considerations {#s14}\n----------------------\n\nThe study protocol was approved by the Provisional National Committee for Ethics in Health Research in Benin, the Institutional Review Board of ITM, and the Ethical Committee of the University Hospital of Antwerp. The study also received administrative authorization of the Benin Ministry of Health Ethics Board. All participants gave written informed consent.\n\nRESULTS {#s15}\n=======\n\nParticipants {#s16}\n------------\n\nBetween March 2012 and March 2015, 260 eligible patients presented at the study sites: 166 with symptoms or signs clinically compatible with BU and 94 with lesions clinically not compatible with BU ([Figure 1](#F1){ref-type=\"fig\"}).\n\n![Flow of recruited and included patients, process flow of the participating patients in the 3 steps of the expert panel approach, and the regrouping of confirmed and probable Buruli ulcer (BU) patients during the analysis. In the first step of the expert panel approach, 97 (47%) participants were classified as confirmed BU, 77 (38%) as possible BU, and 31 (15%) as non-BU. The expert panel review (step 2) of the 77 participants with possible BU classified 14 participants as confirmed BU, 20 as non-BU, and 43 with a discordant classification. In the third step, 9 of these 43 participants were classified as confirmed BU and 23 as non-BU. For the remaining 11 participants, the expert panel failed to reach a consensus, and these 11 participants were classified as BU. Abbreviations: BU, Buruli ulcer; PCR, polymerase chain reaction.](ciy19701){#F1}\n\nOf the 166 patients clinically suspected to have BU, 133 were enrolled. The 33 not recruited did not differ demographically nor clinically from the study participants, suggesting that, at least among those clinically suspected to have BU, there was no overt selection bias ([Supplementary Data 4](#sup1){ref-type=\"supplementary-material\"}). The total number of eligible patients with lesions clinically not compatible with BU was not documented.\n\nAmong the 227 participants recruited, 205 had complete test results and were included in the analysis ([Figure 1](#F1){ref-type=\"fig\"}). Their demographic and clinical characteristics are summarized in [Table 1](#T1){ref-type=\"table\"}. The patients included in the analysis did not differ demographically nor clinically from the ones with incomplete data ([Supplementary Data 5](#sup1){ref-type=\"supplementary-material\"}).\n\n###### \n\nDemographic and Clinical Characteristics of the 205 Study Participants\n\n  Characteristic                All Patients   BU Patients (n = 131)   Non-BU Patients (n = 74)   OR        (95% CI)   *P* Value                              \n  ----------------------------- -------------- ----------------------- -------------------------- --------- ---------- ----------- -------- ----------------- ----------\n  Female sex                    89             (43.41)                 62                         (47.33)   27         (36.49)     1.56     (.87--2.81)       .13\n  Age, y, median (IQR)          19             (9--42)                 12                         (9--42)   40         (9--42)                                \\<.00001\n  CDTUB                         113            (55.12)                 51                         (38.93)   62         (83.78)     0.09     (.04--.18)        \\<.00001\n  \u2003Allada                       100                                    50                                   50                                                \n  \u2003Lalo                         13                                     1                                    12                                                \n  HIV status                                                                                                                       0.74     (.17--3.23)       .69\n  \u2003Infected                     8              (5.52)                  5                          (5.00)    3          (6.67)                                 \n  \u2003Uninfected                   137            (94.48)                 95                         (95.00)   42         (93.33)                                \n  \u2003Not tested                   60                                     31                                   29                                                \n  Clinically BU, WHO category   127            (61.95)                 120                        (91.60)   7          (9.46)      104.42   (38.66--282.04)   \\<.00001\n  \u20031                            25             (19.69)                 22                         (18.33)   3          (42.86)                                \n  \u20032                            69             (54.33)                 65                         (54.17)   4          (57.14)                                \n  \u20033                            33             (25.98)                 33                         (27.50)   0                                                 \n  Clinically non-BU             78             (38.05)                 11                         (8.40)    67         (90.54)                                \n  \u2003Necrotizing fasciitis        36             (46.15)                 5                          (45.45)   31         (46.27)                                \n  \u2003Chronic ulcer                22             (28.21)                 2                          (18.18)   20         (29.85)                                \n  \u2003Abscess                      5              (6.41)                                                       5          (7.46)                                 \n  \u2003Infected wound               3              (3.85)                  1                          (9.09)    2          (2.99)                                 \n  \u2003Chronic traumatic wound      2              (2.56)                                                       2          (2.99)                                 \n  \u2003Tumor                        2              (2.56)                  1                          (9.09)    1          (1.49)                                 \n  \u2003Erysipelas                   1              (1.28)                                                       1          (1.49)                                 \n  \u2003Osteomyelitis                1              (1.28)                  1                          (9.09)                                                      \n  \u2003Kaposi sarcoma               1              (1.28)                                                       1          (1.49)                                 \n  \u2003Varicose ulcer               1              (1.28)                                                       1          (1.49)                                 \n  \u2003Necrotic wound               1              (1.28)                  1                          (9.09)                                                      \n  \u2003Suppuration                  1              (1.28)                                                       1          (1.49)                                 \n  \u2003Cervical adenopathy          1              (1.28)                                                       1          (1.49)                                 \n  \u2003Ganglionary ulcer            1              (1.28)                                                       1          (1.49)                                 \n\nData are presented as No. (%) unless otherwise indicated.\n\nAbbreviations: BU, Buruli ulcer; CDTUB, Centres de D\u00e9pistage et de Traitement de l**'**Ulc\u00e8re de Buruli; CI, confidence interval; HIV, human immunodeficiency virus; IQR, interquartile range; OR, odds ratio; WHO, World Health Organization.\n\nExpert Panel Approach {#s17}\n---------------------\n\nThe expert panel approach classified 131 (64%) patients as BU and 74 (36%) as non-BU ([Figure 1](#F1){ref-type=\"fig\"}). Among the 74 non-BU patients, the expert panel classified 30 as non-BU with confirmed differential diagnoses and 44 as non-BU with unclear differential diagnoses ([Supplementary Data 6](#sup1){ref-type=\"supplementary-material\"}). Among the 11 patients for whom the expert panel failed to reach a consensus and who were classified as BU, 10 were treated with an 8-week course of rifampicin and streptomycin. One patient was not treated as BU and received ciprofloxacin and cloxacillin.\n\nAccuracy of Diagnostic Indicators {#s18}\n---------------------------------\n\nThe results of the index tests are summarized in [Supplementary Materials 7](#sup1){ref-type=\"supplementary-material\"} and [8](#sup1){ref-type=\"supplementary-material\"}.\n\nUsing the expert panel as the reference, clinical diagnosis by the treating clinician had the highest sensitivity (0.92 \\[95% confidence interval {CI}, .85--.96\\]), followed by PCR (0.65 \\[95% CI, .56--.73\\]), Ziehl-Neelsen DSE (0.47 \\[95% CI, .38--.55\\]), auramine DSE (0.28 \\[95% CI, .21--.37\\]), and culture (0.28 \\[95% CI, .20--.36\\]) ([Figure 2](#F2){ref-type=\"fig\"}). The specificities and PPV of all diagnostics were high (\u22650.91 and \u22650.92, respectively). The NPV of the clinical diagnosis was higher (0.86 \\[95% CI, .76--.93\\]) than that of any of the other tests (\u22640.62 for all other tests). Using PCR as the reference in a secondary analysis, clinical diagnosis remained the diagnostic method with the highest sensitivity (0.92 \\[95% CI, .84--.97\\]) ([Figure 2](#F2){ref-type=\"fig\"}).\n\n![The accuracy estimates of diagnostic indicators with an expert panel and polymerase chain reaction (PCR) as a reference standard. The specificity and positive predictive value of PCR could not be estimated as the expert panel protocol always classified a positive PCR result as a confirmed Buruli ulcer. Abbreviations: CI, confidence interval; DSE, direct smear examination; NPV, negative predictive value; PCR, polymerase chain reaction; PPV, positive predictive value; ZN, Ziehl-Neelsen.](ciy19702){#F2}\n\nFor histopathology, the sensitivity and specificity at a cutoff score of 4 (compatible with BU) were 0.54 (95% CI, .45--.63) and 0.73 (95% CI, .61--.83), respectively. At cutoff, 7 (probable BU), the sensitivity and specificity were 0.18 (95% CI, .12--.26) and 1.00 (95% CI, .95--1.00). Using ROC curve analysis, we could not identify a better cutoff given the low combination of sensitivity and specificity across the entire range of scores, with a maximum performance (optimal combination of sensitivity and specificity) at a score of 3.5 (sensitivity: 0.80 \\[95% CI, .73--.86\\]; specificity: 0.57 \\[95% CI, .45--.68\\]). The area under the curve was 0.73 (95% CI, .67--.80; [Supplementary Data 9](#sup1){ref-type=\"supplementary-material\"}).\n\nThe accuracy of different combinations of diagnostic tests resulted in an increased sensitivity only when PCR was included without reducing specificity nor PPV ([Figure 3](#F3){ref-type=\"fig\"}).\n\n![Incremental accuracy estimates. Abbreviations: CI, confidence interval; DSE, direct smear examination; NPV, negative predictive value; PCR, polymerase chain reaction; PPV, positive predictive value.](ciy19703){#F3}\n\nWhen evaluating effect modification by important covariates, only the performance of clinical diagnosis differed according to study site and type of lesion, with a lower specificity in decentralized settings, and a higher specificity in ulcerative lesions ([Supplementary Data 10](#sup1){ref-type=\"supplementary-material\"}). The NPV of both DSE assays, PCR, and culture was lower among patients recruited in decentralized sites, reflecting the higher BU prevalence in decentralized sites.\n\nIn a sensitivity analysis, no significant differences were observed in accuracy estimates when classifying the 11 participants for whom the expert panel could not reach a consensus as non-BU ([Supplementary Data 11](#sup1){ref-type=\"supplementary-material\"} and [12](#sup1){ref-type=\"supplementary-material\"}).\n\nAssuming that sensitivity and specificity of clinical diagnosis and microbiological tests remain constant, a further reduction of the proportion of BU patients would impact the predictive values, with increases in NPV and decreases in PPV ([Supplementary Data 13](#sup1){ref-type=\"supplementary-material\"}).\n\nAccuracy of Published Diagnostic Algorithms {#s19}\n-------------------------------------------\n\nAll evaluated diagnostic algorithms had similar performances ([Figure 4](#F4){ref-type=\"fig\"}).\n\n![Accuracy estimates of various diagnostic approaches. Abbreviations: CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value; WHO, World Health Organization.](ciy19704){#F4}\n\nDISCUSSION {#s20}\n==========\n\nIn a BU-endemic setting, trained clinicians clinically diagnose BU with a sensitivity of 92% (95% CI, 85%--96%) and specificity of 91% (95% CI, 81%--96%). The sensitivity of a clinical diagnosis was higher than the sensitivity of any laboratory test. Typical ulcerated BU lesions have indeed been reported to be easily diagnosed clinically in BU-endemic areas by experienced clinicians \\[[@CIT0016]\\]. Despite the excellent performance of clinical diagnosis, 14% of the study participants clinically not suspected to have BU were reclassified as BU patients by the expert panel procedure, suggesting that there may be a nonnegligible level of underdiagnosis. The majority (64%) of these missed diagnoses were positive by PCR. In clinical practice, these patients do not receive any laboratory test and are not treated for BU unless their clinical evolution suggests BU, resulting in a considerable delay to treatment. However, if the burden of BU would continue reducing, it would virtually never be missed by the clinicians, but a considerable proportion of clinical suspects would not have BU ([Supplementary Data 13](#sup1){ref-type=\"supplementary-material\"}).\n\nThe sensitivity of PCR (65% \\[95% CI, 56%--73%\\]) in the present study was lower than generally reported in studies using a variety of reference standards (85.4% \\[[@CIT0013]\\], 86.0% \\[[@CIT0028]\\], 87.5% \\[[@CIT0014]\\], 100% \\[[@CIT0015]\\]). Clinical characteristics were included into the expert panel approach which resulted in a considerable proportion of PCR-negative patients being classified as BU and therefore more false-negative PCR results than in studies using only microbiological assays as reference standards. However, the addition of PCR to DSE resulted in an increased sensitivity (69% \\[95% CI, 60%--77%\\]) while the combined specificity remained high (93% \\[95% CI, 85%--98%\\]).\n\nThe performance of published diagnostic algorithms for BU in endemic settings was similar. However, the Buruli score---based on clinical and demographic patient characteristics, reserving PCR for patients with an intermediate score \\[[@CIT0015]\\]---can only be used for the diagnosis of ulcerative lesions. The stepwise approach reserving PCR to DSE-negative patients would therefore be the most cost-effective diagnostic algorithm with the shortest time to results in both ulcerative and nonulcerative lesions.\n\nThe performance of histopathology was poor in this study. Its accuracy estimates were low at both cutoff scores of 4 and 7. The ROC analysis of the histopathology consensus scores did not allow the identification of a more optimal cutoff, possibly due to a poor performance of the standardized reading form. Moreover, the pathologists received biopsies in batches blinded to other test results and clinical information. While blinding of histopathologists allows assessing the value of the test per se, it is in contrast with clinical practice where histopathology reading is interpreted taking into account clinical information. Blinding may thus have underestimated the true added value of histopathology, resulting in test review bias \\[[@CIT0029]\\]. An additional limitation of histopathology in the present analysis is the use of consensus scores. Moreover, the 4-mm punch biopsies recommended by the WHO may sometimes be too small to contain the typical histological changes. Also, the localization of the biopsy in the lesion determines the histological changes captured by the sample. Since in clinical practice histopathology is used when all microbiological tests are negative but there is still a clinical suspicion of BU, a cutoff should be selected at a high sensitivity. However, in the present study the sensitivity of histopathology was disappointingly low at every cutoff.\n\nBecause a gold standard without error or uncertainty is not available for BU, we used an expert panel approach. However, the expert panel had access to all available diagnostic indicators, resulting in incorporation bias \\[[@CIT0023]\\] and possibly overestimating the specificity and PPV of the diagnostic indicators while underestimating their sensitivity and NPV. The true accuracy estimates are probably situated between those measured by PCR and the expert panel. An alternative for diagnostic studies where a good reference standard is lacking is LCA \\[[@CIT0030]\\], which relates observed patterns of test results to unknown or latent categories of patients---those with and those without the condition. For BU, LCA has only been used in 1 study \\[[@CIT0015]\\].\n\nInclusion of multiple sites, clinicians, and histopathologists may have increased interobserver variability. The effect of interobserver variability is likely to be larger for the more subjective tests such as clinical diagnosis and histopathology. We deliberately choose a pragmatic approach, as recommended by Grobbee and Hoes \\[[@CIT0031]\\], where all diagnostic determinants are assessed as much as possible according to daily practice and by the practicing physician, with some effort to standardize measurements. However, generalizing the study results should be done with care. Health workers in other BU-endemic settings may be less experienced than those working in southern Benin.\n\nCONCLUSIONS {#s21}\n===========\n\nDespite the excellent performance of clinical diagnosis, there may be an important level of underdiagnosis. A broader clinical suspicion is therefore recommended to reduce missed BU diagnoses allowing improved patient management.\n\nTaking into consideration diagnostic accuracy, time to results, cost-effectiveness, and clinical generalizability, the stepwise diagnostic approach reserving PCR to microscopy-negative patients performed best and would be suitable in the remote and resource-limited settings where BU is endemic.\n\nSupplementary Data {#s22}\n==================\n\nSupplementary materials are available at *Clinical Infectious Diseases* online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.\n\n###### \n\nClick here for additional data file.\n\n***Acknowledgments.***\u2003We thank Jim Zeegelaar and Felix Atadokpede for their contributions to the clinical expert panel, and Achilleas Tsoumanis, Joris Menten, and Meryam Krit for help with coding in R and advice on statistical analyses.\n\n***Disclaimer.***\u2003The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n***Financial support.***\u2003This work was supported by the Medicor Foundation ([www.medicor.li](http://www.medicor.li)) and UBS Optimus Foundation ([www.ubs.com/microsites/optimus-foundation](http://www.ubs.com/microsites/optimus-foundation)) through the BU differential diagnosis project in Benin. Fondation Follereau Luxembourg ([www.ffl.lu/](http://www.ffl.lu/)) gives core support to the CDTUB in Allada. M. E. was supported by the Department of Economy, Science and Innovation ([www.ewi-vlaanderen.be/](http://www.ewi-vlaanderen.be/)) of the Flemish government (Belgium). B. d. J. was supported by the European Research Council-INTERRUPTB (starting grant number 311725).\n\n***Potential conflicts of interest.***\u2003All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.\n"}
+{"text": "![](medhist00089-0026.tif \"scanned-page\"){.363}\n\n![](medhist00089-0027.tif \"scanned-page\"){.364}\n\n![](medhist00089-0028.tif \"scanned-page\"){.365}\n\n![](medhist00089-0029.tif \"scanned-page\"){.366}\n\n![](medhist00089-0030.tif \"scanned-page\"){.367}\n\n![](medhist00089-0031.tif \"scanned-page\"){.368}\n\n![](medhist00089-0032.tif \"scanned-page\"){.369}\n\n![](medhist00089-0033.tif \"scanned-page\"){.370}\n\n![](medhist00089-0034.tif \"scanned-page\"){.370-1}\n\n![](medhist00089-0035.tif \"scanned-page\"){.370-2}\n\n![](medhist00089-0036.tif \"scanned-page\"){.371}\n\n![](medhist00089-0037.tif \"scanned-page\"){.372}\n\n![](medhist00089-0038.tif \"scanned-page\"){.373}\n\n![](medhist00089-0039.tif \"scanned-page\"){.374}\n\n![](medhist00089-0040.tif \"scanned-page\"){.375}\n\n![](medhist00089-0041.tif \"scanned-page\"){.376}\n\n![](medhist00089-0042.tif \"scanned-page\"){.377}\n\n![](medhist00089-0043.tif \"scanned-page\"){.378}\n\n![](medhist00089-0044.tif \"scanned-page\"){.379}\n\n![](medhist00089-0045.tif \"scanned-page\"){.380}\n\n![](medhist00089-0046.tif \"scanned-page\"){.381}\n\n![](medhist00089-0047.tif \"scanned-page\"){.382}\n\n![](medhist00089-0048.tif \"scanned-page\"){.383}\n\n![](medhist00089-0049.tif \"scanned-page\"){.384}\n"}
+{"text": "Introduction {#s01}\n============\n\nPancreatic ductal adenocarcinoma (PDA) has one of the worst outcomes among all cancers, with a median survival of \u223c6 mo and a 5-yr survival rate of \\<8% ([@bib43]). Patients are often diagnosed late during disease progression, when curative surgical approaches are not feasible. Indeed, the current systemic therapies for patients with advanced PDA provide only temporary benefits, highlighting the need for new therapeutic strategies.\n\nPDA is characterized by abundant desmoplasia that constitutes up to 90% of the total tumor volume and contains extracellular matrix (ECM), immune cells, vasculature, and cancer-associated fibroblasts (CAFs; [@bib31]). CAFs secrete ECM and soluble factors that stimulate cancer progression, and are believed to be derived from mesenchymal cells of different origins that are resident or recruited to the pancreas by neoplastic cells ([@bib33]; [@bib30]; [@bib24]). A major source of CAFs in PDA is pancreatic stellate cells (PSCs), which are resident mesenchymal cells of the pancreas that store lipid droplets and express fibroblast-activation protein \u03b1 (FAP; [@bib6]; [@bib13]; [@bib4]; [@bib31]). Upon activation, PSCs express the myofibroblast protein \u03b1-smooth muscle actin (\u03b1SMA, gene name *Acta2*) and secrete factors that stimulate tumor growth, cell survival, and metastasis ([@bib6]; [@bib21]; [@bib53]; [@bib55]). PSCs are also reported to produce the majority of ECM in PDA ([@bib3]), which acts as a physical barrier that impairs drug delivery ([@bib34]; [@bib22]; [@bib36]), and may also biochemically contribute to drug resistance ([@bib17]; [@bib46]).\n\nAttempts to target the stroma, either by directly targeting CAFs ([@bib34]; [@bib16]; [@bib41]), or by enzymatically digesting the ECM ([@bib22]; [@bib36]), have resulted in reduced tumor growth and improved response to chemotherapy in mouse models and patients ([@bib19]). These preclinical and clinical findings have nominated the PDA stroma, particularly CAFs, as attractive targets for drug development. However, several recent reports have questioned the role of CAFs in PDA maintenance ([@bib7]). Genetic disruption or prolonged pharmacological inhibition of sonic hedgehog, a ligand that stimulates CAFs ([@bib50]; [@bib26]; [@bib40]), or depletion of \u03b1SMA-expressing cells ([@bib35]), resulted in undifferentiated PDA tumors and decreased survival in mice. Furthermore, clinical trials of Smoothened inhibitors, which targeted the G protein--coupled receptor downstream of hedgehog signaling, have failed to demonstrate benefits for PDA patients ([@bib25]), with one randomized trial reporting adverse effects ([@bib58]). Although these findings demonstrate the need to use caution when targeting CAFs, they also highlight the need to systematically determine the composition and function of the PDA stroma to improve the development of effective stroma-targeting drugs. Indeed, based on the expression patterns of various fibroblast markers in vivo, evidence is emerging on the existence of different subtypes of CAFs ([@bib47]; [@bib33]; [@bib24]). However, no precise characterization of CAF subtypes has been performed. Here, we investigate CAF heterogeneity in a novel three-dimensional co-culture system that recapitulates the in vivo symbiotic interactions of CAFs and cancer cells. Our study reveals two spatially separated, mutually exclusive, dynamic, and phenotypically distinct CAF subtypes, underscoring the stromal heterogeneity in PDA and providing an opportunity to develop agents that target specific CAF populations.\n\nResults and discussion {#s02}\n======================\n\nHeterogeneous distribution of myofibroblastic CAFs in PDA {#s03}\n---------------------------------------------------------\n\nTo investigate the inherent heterogeneity of fibroblasts in pancreatic cancer, we evaluated the spatial distribution of \u03b1SMA, a hallmark of myofibroblasts ([@bib11]), in human pancreatic tumors. Immunofluorescence analysis of FAP, a PSC marker, and \u03b1SMA expression in human PDA tissues revealed that the majority of fibroblasts expressed FAP and low levels of \u03b1SMA, whereas a subpopulation of FAP^+^ cells showed substantially elevated expression levels of \u03b1SMA ([Fig. 1 A](#fig1){ref-type=\"fig\"}). These FAP^+^ \u03b1SMA^high^ cells could also be delineated by RNA in situ hybridization (ISH), and were located in direct proximity to neoplastic cells, forming a periglandular ring surrounding cancer cell clusters ([Fig. 1 B](#fig1){ref-type=\"fig\"}). A gradient of \u03b1SMA expression was similarly observed in tumors from KPC (*Kras^LSL-G12D/+^; Trp53^LSL-R172H/+^; Pdx-1-Cre*) mice ([Fig. 1, C and D](#fig1){ref-type=\"fig\"}), a mouse model that recapitulates the human disease ([@bib18]). As a result of the selective high expression of \u03b1SMA, we refer to these periglandular FAP^+^ \u03b1SMA^high^ fibroblasts as myofibroblastic CAFs (myCAFs).\n\n![**High expression of \u03b1SMA is a distinctive property of periglandular CAFs in mouse and human PDA.** (A, left) Representative immunofluorescence (IF) co-staining of FAP (green) and \u03b1SMA (red) in a well-differentiated human PDA (*n* = 4). Counterstain, DAPI (blue). (right) Higher magnification illustrating the distribution and co-localization of FAP and \u03b1SMA. Bars, 50 \u00b5m. T, tumor glands. (B) Representative image of RNA ISH for Cytokeratin 18 (*KRT18,* blue) and \u03b1SMA (*ACTA2*, red) transcripts in a well-differentiated human PDA (*n* = 3). Bar, 50 \u00b5m. T, tumor gland. (C, left) Representative image of RNA ISH for *Fap* (blue) and *Acta2* (red) in a KPC mouse tumor (*n* = 3). (right) Higher magnification. Bars, 25 \u00b5m. T, tumor glands. (D, top) Representative image of fluorescent RNA ISH for *Fap* (green) and *Acta2* (red) in a KPC mouse tumor (*n* = 3), showing transcript distribution across three cell layers of the stroma, starting from the first layer adjacent to the tumor gland (T) and moving outwards. Counterstain, DAPI (blue). Bar, 50 \u00b5m. (bottom) Quantification of *Fap* and *Acta2* fluorescence intensity in the three cell layers. Results show mean \u00b1 SD of three tumor glands. Data are normalized to layer 1. \\*\\*\\*, P \\< 0.001, unpaired Student's *t* test. (E) Representative images of IHC of \u03b1SMA and YFP in sequential tissue sections from KPCY mice, with either preinvasive Pancreatic Intraepithelial Neoplasia (PanIN) or invasive cancer (*n* = 2). Arrows indicate areas of myCAFs. Bar, 50 \u00b5m.](JEM_20162024_Fig1){#fig1}\n\nTo exclude the possibility that myCAFs are neoplastic cells that have undergone epithelial-to-mesenchymal transition (EMT), we analyzed tissues of a KPCY mouse model, in which all neoplastic pancreatic cells express yellow fluorescent protein (YFP; [@bib39]). Periglandular cells expressing high \u03b1SMA levels did not coexpress YFP, confirming that myCAFs are of a nonneoplastic origin ([Fig. 1 E](#fig1){ref-type=\"fig\"}). These findings identify myCAFs as a distinct subpopulation of CAFs with a unique spatial distribution pattern in PDA.\n\nA novel three-dimensional co-culture platform recapitulates in vivo CAF heterogeneity {#s04}\n-------------------------------------------------------------------------------------\n\nTo further characterize CAFs in PDA, we studied PSCs, which are believed to be a major source of FAP^+^ CAFs in PDA stroma ([@bib3]; [@bib13]). Quiescent and lipid-storing PSCs were isolated from WT C57BL/6J mice pancreata (Fig. S1, A and B), and cultured as primary cells or after immortalization with the SV40 large T Antigen. When PSCs are grown in monolayers, they lose their lipid droplets and assume a myofibroblastic phenotype indicated by \u03b1SMA expression, but can be reversed back to a quiescence and lipid-storing phenotype if embedded in Matrigel ([@bib23]). To verify that the characteristic phenotypes of PSCs remained after isolation and immortalization, we cultured PSCs in Matrigel and used Oil Red-O staining to confirm that they reacquired lipid droplets (Fig. S1 C). Additionally, we found that the PSCs, when cultured as a monolayer, showed an adequate response to the vitamin D analogue Calcipotriol ([@bib41]; Fig. S1 D). Moreover, addition of recombinant TGF\u03b2 to quiescent PSCs cultured in Matrigel induced the expression of TGF\u03b2 target genes, such as *Ctgf* and *Col1a1* (Fig. S1 E), demonstrating that the isolated PSCs still respond to common stromal cues.\n\nTo investigate the interactions between cancer cells and PSCs, we established a three-dimensional organotypic co-culture system that combined GFP-labeled tumor-derived murine pancreatic organoids ([@bib20]; [@bib8]) and mCherry-labeled murine PSCs ([Fig. 2 A](#fig2){ref-type=\"fig\"}). Pancreatic organoids are routinely cultured in a defined, mitogen-rich media. However, many factors that are present in this organoid media, including Noggin, B27 supplement, and TGB\u03b2 inhibitor, are known to be potent inhibitors of fibroblast proliferation. To avoid inhibition of PSC proliferation in co-cultures, we used a reduced media without these components, based on DMEM supplemented with 5% FBS. Although cancer-naive PSCs embedded alone in Matrigel remained quiescent, PSCs in co-culture with tumor organoids acquired a CAF phenotype, demonstrated by morphological activation with cellular extensions in close contact with tumor organoids ([Fig. 2 B](#fig2){ref-type=\"fig\"}). Importantly, freshly isolated and nonimmortalized PSCs seeded directly in Matrigel showed similar morphological changes when co-cultured with organoids ([Fig. 2 C](#fig2){ref-type=\"fig\"}), demonstrating that the activated phenotype in co-culture is an inherent feature of PSCs. Electron microscopy of co-cultures revealed the close relationship between cancer cells and PSCs, with only a thin gap filled with matrix separating the two cell types ([Fig. 2 D](#fig2){ref-type=\"fig\"}). Activated PSCs deposited ECM ([Fig. 2 E](#fig2){ref-type=\"fig\"}) containing collagen I ([Fig. 2 F](#fig2){ref-type=\"fig\"}) that was organized into collagen fibrils ([Fig. 2 G](#fig2){ref-type=\"fig\"}). Thus, our co-culture system is the first to recapitulate the desmoplastic reaction of PDA in vitro, with PSCs converting from a resting quiescent state to activated, stroma-producing CAFs.\n\n![**Co-cultures of mouse PSCs and pancreatic cancer organoids recapitulate properties of PDA desmoplasia.** (A) Schematic illustration of the co-culture platform. (B) Representative images of mCherry-labeled PSCs (red) cultured alone or in co-culture with GFP-labeled tumor-derived organoids (green), and imaged by confocal microscopy after 4 or 7 d in bright field and by fluorescent microscopy (*n* = 3). Arrows point to close interactions between organoids and PSCs. Bars, 100 \u00b5m. (C) Bright field images of primary PSCs plated in Matrigel directly after isolation, and either cultured alone or co-cultured with tumor organoids for 5 d (*n* = 3). Bar, 100 \u00b5m. (D) Representative electron microscopy image showing the proximity between organoids and PSCs in co-culture (*n* = 2). Bar, 5 \u00b5m. (E) H&E staining and Masson's trichrome (MT) staining of fixed and paraffin-embedded organoids cultured alone or in co-culture with PSCs (*n* = 2). Bar, 50 \u00b5m. (F) Representative bright field and IF images of collagen I deposition (red) in organoid cultured alone or in co-culture with PSCs (*n* = 2). Bar, 200 \u00b5m. (G) Representative electron microscopy image of banded collagen fibrils (arrow), with fibril diameters ranging between 24 and 35 nm, in the extracellular space between organoids and PSCs (*n* = 2). Bar, 1 \u00b5m. (H) PSC proliferation curves plotting changes in mCherry intensity over time. Results show mean \u00b1 SD of two biological replicates. \\*\\*, P \\< 0.01, unpaired Student's *t* test. (I) Organoid proliferation curves plotting changes in GFP intensity over time. Results show mean \u00b1 SD of three biological replicates. \\*, P \\< 0.05, unpaired Student's *t* test. (J) Passaging of organoids in different culture conditions in the presence or absence of PSCs. Complete media, DMEM/F12 supplemented with mitogens and growth factors. Reduced media, DMEM + 5% FBS. Red dot indicates the passage number when all organoids were found dead. Green dot indicates surviving organoids when the experiment was terminated. Each dot represents one biological replicate. Bars indicate the average number of passages for each condition. (K) RNA ISH of fixed and sectioned co-cultures for \u03b1SMA (*Acta2*, red) and *Krt18* (green) illustrating the spatial distribution of \u03b1SMA^high^ PSCs in comparison to Krt18^+^ (green) tumor organoids (*n* = 2). Counterstain, DAPI. Higher magnification on the right. Bars, 50 \u00b5m.](JEM_20162024_Fig2){#fig2}\n\nPSCs were nonproliferative until organoids were included in co-culture ([Fig. 2 H](#fig2){ref-type=\"fig\"}), and the co-cultures also promoted the proliferation of organoids ([Fig. 2 I](#fig2){ref-type=\"fig\"}). Moreover, whereas organoid monocultures could be passaged indefinitely in complete media ([@bib8]), reduced media conditions precluded prolonged passaging of organoids unless PSCs were present in co-culture ([Fig. 2 J](#fig2){ref-type=\"fig\"}). The mutual proliferative benefits for both epithelial cells and PSCs are consistent with prior reports of similar symbiotic interactions that promote PDA ([@bib6]; [@bib21]; [@bib53]; [@bib44]).\n\nAnalysis by ISH identified a clear distribution of \u03b1SMA^high^ PSCs surrounding the organoids in co-culture ([Fig. 2 K](#fig2){ref-type=\"fig\"}), recapitulating the in vivo finding of a subpopulation of fibroblasts expressing high levels of \u03b1SMA (myCAFs) in close proximity to neoplastic cells ([Fig. 1, A--D](#fig1){ref-type=\"fig\"}). This result also supports the premise that PSCs are a source of myofibroblastic CAFs in addition to BM-derived mesenchymal stem cells (BM-MSCs), which have previously been reported to give rise to myofibroblasts in cancer ([@bib38]). Moreover, this novel co-culture platform confirms the existence of cooperative interactions between cancer cells and PSCs, and provides a way for their systematic characterization.\n\nA population of CAFs secretes inflammatory cytokines and loses myofibroblastic features {#s05}\n---------------------------------------------------------------------------------------\n\nTo identify soluble factors that may mediate the symbiosis observed in co-culture, we analyzed the secretome of PSC and organoid co-cultures. We found that several secreted proteins, including inflammatory cytokines and chemokines, matrix remodeling proteins, and growth factors, were elevated specifically in co-culture ([Fig. 3 A](#fig3){ref-type=\"fig\"}). One of the induced cytokines was interleukin 6 (IL-6), which has been reported to be up-regulated in PDA ([@bib49]) and to promote cancer progression ([@bib27]; [@bib57]; [@bib32]), cachexia, and immune suppression ([@bib15]).\n\n![**Secretion of inflammatory cytokines from CAFs activates STAT3 in PDA organoids.** (A) Quantification of secretome dot blots of conditioned media from mouse tumor organoid monocultures, PSC monocultures, co-cultures, or Matrigel-only controls (MG). Results show normalized mean \u00b1 SD of three biological replicates. (B) Schematic illustration of the trans-well culture platform. (C and D) qPCR analysis of GP130 signaling ligands and receptors in mouse PSCs (C) or tumor organoids (D) cultured in monoculture or trans-well culture. Results show mean \u00b1 SD of four biological replicates. n.d., not detected. (E) ELISA of IL-6, IL-11, and LIF from conditioned media of PSC monocultures, organoid monocultures, or co-cultures. Results show mean \u00b1 SD of three biological replicates. (F) Quantification of secretome dot blots of conditioned media from human primary CAF monocultures, patient-matched tumor organoid monocultures, or co-cultures (*n* = 2). Results show mean \u00b1 SD of two technical replicates for each condition. (G) Representative IF image of KPC mouse tumor stained for phosphorylated STAT3 (Tyr705; green, pSTAT3) and the epithelial marker Cytokeratin 19 (Krt19, red; *n* = 2). Counterstain, DAPI (blue). Bar, 75 \u00b5m. (H) Western blot analysis of pSTAT3 in organoids treated with either 10 ng/ml recombinant IL-6, 10 ng/ml recombinant IL-11, or 50 ng/ml recombinant LIF, in the presence or absence of neutralizing antibodies or isotype controls (*n* = 2). Loading control, Actin. Molecular weights in kilodaltons. (I) Western blot analysis of pSTAT3 in organoids treated with co-culture conditioned media in the presence or absence of neutralizing antibodies against IL-6, IL-11, or LIF (*n* = 3). Loading control, Actin. Molecular weights in kilodaltons. (J) Passaging of organoids in reduced media conditions in monoculture or co-culture with WT (PSC WT) or IL-6 KO PSCs (PSC IL-6 KO). Red dot indicates the passage number when all organoids were found dead. Green dot indicates the passage number of surviving organoids when the experiment was terminated. Each dot represents one biological replicate. Bars indicate the average number of passages for each condition. (K) qPCR analysis of *Il6*, *Il11, Lif*, and *Acta2* transcript levels in PSCs cultured with control media (Matrigel-only conditioned media) or tumor organoid conditioned media. Results show mean \u00b1 SD of five biological replicates for *Il6, Lif*, and *Acta2*, and three biological replicates for *Il11*. (L) Western blot analysis of PSCs cultured with control media or tumor organoid conditioned media (*n* = 3). Loading control, Hsp90\u03b1. Molecular weights in kilodaltons. (M and N) qPCR analysis of *Il6* and *Acta2* in three primary PSC lines (M) and two KPC mouse CAFs (N) cultured with control media or tumor organoid conditioned media. Results show mean \u00b1 SD of two technical replicates for each line. (O) qPCR analysis for *IL6* and *ACTA2* transcript levels in human primary CAFs cultured with control media or conditioned media from the corresponding patient-matched tumor organoids. Results show mean \u00b1 SD of 2 technical replicates. \\*, P \\< 0.05; \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001, unpaired Student's *t* test.](JEM_20162024_Fig3){#fig3}\n\nThe increased cytokine production observed when tumor organoids and PSCs are co-cultured could be the result of a direct physical contact between these populations or alternatively, it could be caused by paracrine signaling between them. To disentangle these two possibilities and to determine which cell type was responsible for the increased cytokine production, we made use of a Transwell system, which allows paracrine interactions between tumor organoids and PSCs, but prevents direct contact between the two cell types ([Fig. 3 B](#fig3){ref-type=\"fig\"}). IL-6 acts by binding to the IL-6 receptor subunit \u03b1 (IL6RA), which in turn facilitates the dimerization and signaling of the GP130 signaling complex ([@bib48]). Because GP130 is known to interact with other ligand-receptor pairs, we measured the expression of additional GP130 signaling partners in both organoids and PSCs in trans-well cultures. Of the GP130 ligands found to be expressed in PSCs, *Il6*, *Il11*, and *leukemia inhibitory factor* (*Lif*) were the most highly up-regulated in trans-well cultures when compared with PSC monocultures ([Fig. 3 C](#fig3){ref-type=\"fig\"}). Both IL-11 and LIF are reported to have roles in cancer progression ([@bib37]; [@bib29]), and LIF expression has been shown to be elevated in human PDA ([@bib10]). In trans-well cultures, organoids did not express detectable levels of *Il6* but expressed *Il6ra*, *Il11ra1*, *Lifr*, and *Gp130* ([Fig. 3 D](#fig3){ref-type=\"fig\"}), indicating that activated PSCs are the sole source of IL-6 in co-culture and that tumor organoids express the receptors needed to respond to PSC-secreted ligands. These results also confirm that direct contact with neoplastic cells is not required for PSCs to initiate cytokine secretion. Enzyme-linked immunosorbent assay (ELISA) was performed to confirm and quantify the elevated secretion of IL-6, IL-11, and LIF in co-culture ([Fig. 3 E](#fig3){ref-type=\"fig\"}).\n\nTo evaluate these findings in human PDA fibroblasts, we isolated CAFs from primary (hT1) and metastatic (hM1) tumors of two PDA patients ([@bib8]; Fig. S1, F--J). We then performed secretome analysis of conditioned media from co-cultures of human PDA CAFs with patient-matched tumor organoids. Secretion of IL-6 was induced in both primary tumor and metastatic co-cultures, whereas secretion of IL-11 and LIF was only induced in the co-culture derived from the primary tumor (hT1; [Fig. 3 F](#fig3){ref-type=\"fig\"}).\n\nOnce dimerized with IL6RA, GP130 is phosphorylated and forms a complex with tyrosine kinases such as Janus kinases (JAKs). JAKs in turn phosphorylate and activate signal transducer and activator of transcription (STAT) factors, most notably STAT3, which plays key roles in cell growth and proliferation ([@bib48]). Immunofluorescence staining revealed detectable activation of STAT3 in KPC tumors, both in cancer cells and the surrounding stroma ([Fig. 3 G](#fig3){ref-type=\"fig\"}). To confirm that tumor organoids activate STAT3 in response to paracrine stimuli, we measured STAT3 phosphorylation in organoids after the addition of recombinant IL-6, IL-11, and LIF to the media. All three ligands robustly activated STAT3 in organoids, and this effect was prevented when neutralizing antibodies against each ligand were added ([Fig. 3 H](#fig3){ref-type=\"fig\"}). Interestingly, IL-6 was able to strongly activate STAT3 in tumor organoids, in contrast to studies using acinar cells isolated from *Kras^G12D^* mice that concluded that IL-6 trans-signaling is necessary to mediate robust STAT3 activation in pancreatic tumor cells ([@bib27]). STAT3 was also activated by the addition of co-culture conditioned media to tumor organoids in monoculture, and this activation was again blocked by addition of neutralizing antibodies, with the anti--IL-6 neutralizing antibody having the most prominent effect ([Fig. 3 I](#fig3){ref-type=\"fig\"}).\n\nAs STAT3 activation is known to control cell survival and proliferation, we investigated its role in co-culture by using CRISPR/Cas9 gene editing to knockout IL-6 in two PSC lines. We confirmed loss of IL-6 secretion in PSCs by ELISA of conditioned media from trans-well cultures (Fig. S1 K). Interestingly, whereas organoids co-cultured with WT PSCs showed, as expected, prolonged passaging ability in reduced media conditions compared with monocultured organoids, continued passaging of organoids was impaired in co-cultures with IL-6--deficient PSCs ([Fig. 3 J](#fig3){ref-type=\"fig\"}). These results demonstrate that tumor organoids in co-culture activate PSCs to secrete multiple factors, which in turn activate signaling pathways in organoids that sustain survival.\n\nTo further investigate the secretory phenotype of organoid-activated PSCs, we exposed quiescent PSCs to organoid-conditioned media (Fig. S2 A). In these conditions, PSCs acquired a CAF phenotype, indicated by morphological activation (Fig. S2 B), proliferation (Fig. S2 C), and up-regulation of *Il6*, *Il11*, and *Lif* mRNA levels ([Fig. 3 K](#fig3){ref-type=\"fig\"}). Unexpectedly, we detected a simultaneous reduction in *Acta2* transcript and \u03b1SMA protein levels ([Fig. 3, K and L](#fig3){ref-type=\"fig\"}). Although \u03b1SMA expression dropped in PSCs activated by tumor organoid-conditioned media, additional fibroblast surface markers, such as FAP and platelet-derived growth factor receptors (PDGFRs; [@bib56]) remained unchanged (Fig. S2, D and E), suggesting a specific loss of myofibroblastic features in cytokine-secreting PSCs. This result was reproduced in three mouse primary PSC lines ([Fig. 3 M](#fig3){ref-type=\"fig\"}). Conditioned media from KPC cancer cells cultured in monolayer also induced this \u03b1SMA^low^ IL-6^high^ phenotype in PSCs, whereas conditioned media of NIH-3T3 fibroblasts did not (Fig. S2 F), suggesting that the ability to induce this secretory phenotype in PSCs is absent in cells of mesenchymal origin.\n\nTo extend our finding to mouse and human CAFs that already have been reprogrammed in vivo and not necessarily derived from PSCs, we isolated CAF lines from KPC mouse tumors, and validated their mesenchymal origin (Fig. S2, G--I). Mouse CAFs cultured with tumor organoid-conditioned media also increased *Il6* and concomitantly lost *Acta2* expression ([Fig. 3 N](#fig3){ref-type=\"fig\"}). Although to a lesser extent, this pattern was also present in primary CAFs isolated from human primary and metastatic PDA samples when cultured with conditioned media from patient-matched tumor organoids ([Fig. 3 O](#fig3){ref-type=\"fig\"}).\n\nPrevious studies have shown that pancreatic cancer cells induce IL-6 secretion in fibroblasts and PSCs ([@bib12]; [@bib14]; [@bib57]; [@bib54]), and that PSCs are a major source of IL-6 in PDA ([@bib14]; [@bib28]). Our data confirm this trait in cancer-naive PSCs that are allowed to interact with cancer cells, as well as in mouse and human PDA-derived CAFs. Importantly, for the first time we couple this phenotype with loss of myofibroblastic features. Such heterogeneity has previously been unappreciated, likely due to the fact that monolayers of fibroblasts uniformly express high levels of \u03b1SMA in culture.\n\nNotably, the marked drop in bulk \u03b1SMA expression induced by organoid-conditioned media ([Fig. 3 K](#fig3){ref-type=\"fig\"}) contrasted with the presence of CAFs expressing high levels of \u03b1SMA in proximity to neoplastic cells observed in co-cultures ([Fig. 2 K](#fig2){ref-type=\"fig\"}) and in vivo ([Fig. 1, A--D](#fig1){ref-type=\"fig\"}). We therefore hypothesized the coexistence of two distinct subtypes of CAFs in co-culture and PDA tissue, one being \u03b1SMA^high^ IL-6^low^ and proximal to the neoplastic cells (myCAFs), and the other being \u03b1SMA^low^ IL-6^high^ induced by paracrine signaling from the tumor compartment and more distantly distributed throughout the tumor. We termed these \u03b1SMA^low^ IL-6^high^ CAFs inflammatory CAFs (iCAFs) for their cytokine-secreting properties.\n\nCAFs coexist as two mutually exclusive and reversible subtypes {#s06}\n--------------------------------------------------------------\n\nTo determine whether distinct CAF subtypes coexist in co-culture, flow cytometry was used to detect \u03b1SMA and IL-6 expression in PSCs co-cultured with tumor organoids (Fig. S2 J). As our hypothesis predicted, co-culture triggered high \u03b1SMA expression only in a subset of PSCs (myCAFs), whereas a separate population showed low \u03b1SMA expression and concomitantly elevated IL-6 expression (iCAFs; [Fig. 4 A](#fig4){ref-type=\"fig\"}). Moreover, we found that trans-well cultures of PSCs and organoids ([Fig. 3 B](#fig3){ref-type=\"fig\"}) markedly induced iCAF formation, but did not cause an increase in myCAFs ([Fig. 4 B](#fig4){ref-type=\"fig\"}). This demonstrates that the formation of myCAFs is contact-dependent, whereas iCAFs may be enriched in the absence of contact with PDA cells. The presence of iCAFs in co-culture was evident by ISH of *Il6* ([Fig. 4 C](#fig4){ref-type=\"fig\"}), which also confirmed that these CAFs were spatially separated from organoids.\n\n![**Two mutually exclusive subpopulations of CAFs with reversible features coexist in pancreatic cancer.** (A and B) Flow cytometric analysis of \u03b1SMA and IL-6 in PSCs cultured alone or in either co-culture (A) or trans-well culture (B) with tumor organoids. Red frame indicates the gate defining myCAFs (\u03b1SMA^high^ IL-6^low^) and black frame indicates the gate defining iCAFs (\u03b1SMA^low^ IL-6^high^). Numbers indicate percentage of cells within marked gate. Graphs on the right are showing the fold change induction of myCAFs and iCAFs in co-culture, normalized to PSCs in monoculture. Results show mean \u00b1 SD of four (A) or two (B) biological replicates. \\*, P \\< 0.05; \\*\\*, P \\< 0.01, unpaired Student's *t* test. (C) Fluorescent RNA ISH of fixed and sectioned co-cultures for *Il6* (red) and *Krt18* (green), illustrating the spatial distribution of IL-6^+^ PSCs (iCAFs) with respect to KRT18^+^ tumor organoids (*n* = 2). Higher magnification on the right. Counterstain, DAPI (blue). Arrow indicates example of an iCAF. Bars, 50 \u00b5m. (D) qPCR analysis of interleukins (*Il6* and *Il11)* and markers of fibroblast (*Pdgfra, Pdgfrb, Acta2,* and *Fap*), epithelial (*Krt19*) and macrophage (*CD11b*) lineages in samples of primary cells sorted from KPC mouse tumors. Sorting was performed using three markers: PDGFR\u03b1 (CD140a) for fibroblasts (*n* = 3), EpCAM for epithelial cells (*n* = 3) and CD45 for immune cells (*n* = 2). Results show mean \u00b1 SD of two to three biological replicates. All gene expression changes are statistically significant when compared with the reference population, P \\< 0.01, unpaired Student's *t* test. (E) Representative image of sequential IHC for IL-6 (purple) and PDGFR\u03b2 (brown) in a KPC mouse tumor (*n* = 3). Arrows indicate double positive cells. T, tumor gland. Bars, 50 \u00b5m. (F, left) Representative image of sequential IHC for PDGFR\u03b2 (gray), IL-6 (brown), and Ki67 (purple) in a KPC mouse tumor (*n* = 3). Arrows indicate examples of triple positive cells. Bar, 50 \u00b5m. (right) Quantification of Ki67 staining in PDGFR\u03b2^+^/IL-6^+^ cells (iCAFs), total of 593 cells were counted. (G) Representative image of sequential IHC for IL-6 (brown) and PDGFR\u03b2 (purple) in a human PDA (*n* = 6). Arrowheads indicate double positive cells. T, tumor gland. Bars, 50 \u00b5m. (H) Representative image of RNA ISH for *Acta2* (blue) and *Il6* (red) in KPC mouse tumors (*n* = 4). Arrows indicate examples of *Acta2*-positive cells in the periglandular area, arrowheads indicate examples of *Il6*-positive cells further away from neoplastic cells. Bar, 50 \u00b5m. T, tumor glands. (I) Representative IF image for \u03b1SMA (green) and IL-6 (red) in a KPC mouse tumor (*n* = 3). Counterstain, DAPI (blue). Arrowheads indicate examples of \u03b1SMA-positive cells in the periglandular area; \\* indicates examples of IL-6--positive cells further away from neoplastic cells. Bar, 50 \u00b5m. T, tumor glands. (J) qPCR analysis of *Il6*, *Il11*, *Lif*, and *Acta2* transcript levels in two PSC lines (PSC4 and PSC5) first grown as monocultures in Matrigel (quiescent PSCs), then transferred to trans-well cultures with tumor organoids (iCAFs), and finally plated as monolayer cultures (myofibroblasts). Results show mean \u00b1 SD of two technical replicates for each PSC line. \\*\\*, P \\< 0.01; \\*\\*\\*, P \\< 0.001, unpaired Student's *t* test.](JEM_20162024_Fig4){#fig4}\n\nTo evaluate whether CAFs produce interleukins in vivo, we dissociated KPC mouse tumors into single cells, isolated CAFs by flow cytometry, and characterized their gene expression. We sorted CAFs based on PDGFR\u03b1 expression because it has been shown to be a specific surface marker for CAFs ([@bib12]). As a comparison, epithelial and immune cells were isolated, using the EpCAM and CD45 surface markers, respectively (Fig. S2 K). No overlap was detected between the CD45^+^ population and the PDGFR\u03b1^+^ population (Fig. S2 L), demonstrating the mutual exclusivity of these two markers. We further confirmed the purity of the three sorted cell populations by qPCR for additional fibroblast, epithelial and immune markers, and demonstrated that PDGFR\u03b1**^+^** CAFs contain the majority of IL-6 transcripts in PDA ([Fig. 4 D](#fig4){ref-type=\"fig\"}). Additionally, we found that IL-11 mRNA is also predominantly present in CAFs, further supporting the presence of an inflammatory CAF phenotype ([Fig. 4 D](#fig4){ref-type=\"fig\"}).\n\nTo investigate the spatial distribution of IL-6^high^ CAFs in vivo, immunohistochemical analysis of KPC tumor tissue was performed. Consistent with the co-cultures, we detected IL-6 expression in cells that were located further away from neoplastic cells in the desmoplastic stroma ([Fig. 4 E](#fig4){ref-type=\"fig\"}). We confirmed that these were iCAFs by co-staining for PDGFR\u03b2, another fibroblast marker ([@bib56]; [Fig. 4 E](#fig4){ref-type=\"fig\"}). Importantly, a fraction of these iCAFs proliferate in vivo, as indicated by Ki67 coexpression ([Fig. 4 F](#fig4){ref-type=\"fig\"}), demonstrating the lack of a senescence-associated secretory phenotype ([@bib9]). Furthermore, we could identify IL-6--expressing CAFs in human PDA by co-staining human tumor tissues for IL-6 and PDGFR\u03b2, confirming that iCAFs are also an inherent trait of human PDA ([Fig. 4 G](#fig4){ref-type=\"fig\"}). To confirm the spatial distribution of myCAFs and iCAFs in KPC tumors, we used ISH and immunofluorescence staining, and demonstrated that periglandular \u03b1SMA^high^ cells were spatially separated from the more distant IL-6^high^ cells ([Fig. 4, H and I](#fig4){ref-type=\"fig\"}), further supporting our observation of two distinct CAF populations in vivo.\n\nTo determine if the iCAF phenotype was permanent or transient, we plated myofibroblastic PSCs from monolayers into Matrigel to obtain quiescent PSCs, and then induced the iCAF phenotype by culturing these cells in trans-well with PDA organoids. As PSCs grown in monolayers are known to obtain myofibroblastic features, we subsequently plated the iCAFs in monolayer, and found that iCAFs rapidly reverted to a myofibroblastic state, down-regulating *Il6*, *Il11*, and *Lif*, and simultaneously up-regulating *Acta2* expression levels ([Fig. 4 J](#fig4){ref-type=\"fig\"}). This observation suggests that CAFs are dynamic and can assume different phenotypes based on their spatial and biochemical niche within the PDA microenvironment.\n\nOverall, our results, both in vivo and in co-culture, show a spatial separation between iCAFs and myCAFs. MyCAFs are located in the periglandular region, suggesting that direct juxtacrine interactions with cancer cells are required for myCAF formation. iCAFs, on the other hand, are induced by secreted factors from cancer cells and are located more distantly from neoplastic cells and myCAFs in PDA. Although we have demonstrated the existence and spatial separation of these two CAF populations, other CAF populations with unique genetic signatures may exist. The existence of these diverse populations and their functions in tumors remain to be elucidated.\n\nInflammatory CAFs and myofibroblasts show distinct transcriptional profiles {#s07}\n---------------------------------------------------------------------------\n\nTo better understand the differences between these subtypes, we compared the transcriptomes of quiescent PSCs (PSCs embedded alone in Matrigel), iCAFs (SMA^low^ IL-6^high^ PSCs cultured in trans-well with tumor organoids), and myofibroblastic PSCs (SMA^high^ IL-6^low^ PSCs grown in monolayer) as a proxy for myCAFs, because obtaining myCAFs from tumors or co-cultures presents a technical challenge (Fig. S2 M). By RNA sequencing, we found clusters of genes uniquely up-regulated in either myofibroblasts or iCAFs ([Fig. 5, A and B](#fig5){ref-type=\"fig\"}; Fig. S2 N; and Table S1). In particular, *Acta2* and TGF\u03b2 response genes, such as *Ctgf* and *Col1a1*, were up-regulated in myofibroblasts compared with quiescent PSCs and iCAFs. On the other hand, cytokines, such as *Il6*, *Il11*, and *Lif*, and chemokines, such as *Cxcl1* and *Cxcl2*, were uniquely up-regulated in iCAFs ([Fig. 5, A and B](#fig5){ref-type=\"fig\"}; and Table S1). Furthermore, Gene Set Enrichment Analysis (GSEA) of iCAFs compared with quiescent PSCs confirmed the up-regulation of cytokine signaling pathways, and identified JAK/STAT signaling as one of the most significantly up-regulated pathways in iCAFs ([Fig. 5 C](#fig5){ref-type=\"fig\"}). Additionally, as expected, basement membranes and smooth muscle contraction pathways were the most significantly down-regulated ones in iCAFs ([Fig. 5 C](#fig5){ref-type=\"fig\"}). The identification of unique transcriptional signatures of myofibroblasts and iCAFs further supports the model that CAFs within the PDA microenvironment acquire distinct phenotypes. In particular, the transcriptomic profiling implies that myofibroblasts are contractile and stroma remodeling, whereas iCAFs are characterized by a secretory phenotype, with the ability to modulate in a paracrine manner cancer cells and other cell types present in the tumor. Importantly, secreted factors from iCAFs, such as IL-6, likely contribute to systemic effects in PDA patients, such as cachexia and immune suppression ([@bib14]; [@bib15]; [@bib28]).\n\n![**Inflammatory CAFs and myofibroblasts have distinct transcriptional profiles.** (A) RNA sequencing analysis of quiescent PSCs (PSCs embedded alone in Matrigel; *n* = 2), iCAFs (PSCs grown in trans-well culture with tumor organoids; *n* = 4) and myofibroblasts (PSCs grown in monolayer; *n* = 2). The heat map shows differentially expressed genes between the three cell states. Uniquely expressed genes for iCAFs and myofibroblasts are indicated in the boxes. Adjusted P \\< 0.01. (B) Lists of the 25 most up-regulated genes in iCAFs and myofibroblasts compared with quiescent PSCs. Adjusted P \\< 0.05. (C) GSEA of most up-regulated and down-regulated pathways in iCAFs compared with quiescent PSCs. (D) Working model illustrating the dynamic relationship between quiescent PSCs, myCAFs and iCAFs.](JEM_20162024_Fig5){#fig5}\n\nIn summary, we have identified two spatially separated, reversible, and mutually exclusive subtypes of CAFs ([Fig. 5 D](#fig5){ref-type=\"fig\"}). iCAFs, which are activated by paracrine factors secreted from cancer cells, are located more distantly from neoplastic cells within the dense tumor stroma. Although iCAFs still retain expression of \u03b1SMA, they are characterized by significantly lower \u03b1SMA levels compared with myCAFs, and instead intensely elevate expression of cytokines and chemokines with known roles in cancer progression and disease pathophysiology. Indeed, iCAFs are a significant source of IL-6 and IL-11 in the PDA microenvironment ([Fig. 4 D](#fig4){ref-type=\"fig\"}), with the ability to stimulate the STAT3 pathway in cancer cells ([Fig. 3 I](#fig3){ref-type=\"fig\"}). However, myCAFs are defined by high \u03b1SMA expression and periglandular location ([Fig. 1, A--D](#fig1){ref-type=\"fig\"}), and their formation appears dependent on juxtacrine interactions with cancer cells ([Fig. 4 B](#fig4){ref-type=\"fig\"}). Furthermore, myCAFs lack the expression of inflammatory cytokines, distinguishing them from iCAFs. Importantly, both inflammatory and myofibroblastic CAFs can be generated from PSCs, and can dynamically reverse from one cell state to the other ([Fig. 4 J](#fig4){ref-type=\"fig\"}). The detailed mechanisms that govern the formation and transition of these cell states will require further investigation.\n\nAlthough we have identified two subtypes of PSC-derived CAFs in PDA, it is likely that additional CAF subtypes with distinct roles in the pathophysiology of this disease exist. Indeed, our data already reveal a large population of CAFs with low expression of both \u03b1SMA and IL-6 ([Fig. 4, A and B](#fig4){ref-type=\"fig\"}). Furthermore, a more thorough definition of stromal subtypes may involve additional factors and more complex genetic signatures. Importantly, although other groups have identified different subtypes of stroma in PDA across patients ([@bib30]), our work is the first to characterize the intratumoral CAF heterogeneity in PDA.\n\nThe concept of intratumoral CAF heterogeneity may address the conflicting reports that have emerged in the field in regard to CAF functions. Indeed, in recent years, different approaches to target the stroma have given contradicting results and, at times, promoted worse outcomes ([@bib26]; [@bib35]; [@bib40]). For instance, attempts to deplete CAFs based on their \u03b1SMA expression have led to decreased survival in tumor-bearing mice ([@bib35]). However, given our results, this approach may have preferentially eliminated myCAFs, while leaving other CAF populations intact. Therefore, the traditional view of the tumor stroma as a uniformly protumorigenic niche calls for reconsideration, as certain CAF subtypes might have protumorigenic properties, whereas others might have antitumorigenic features. Therapeutic development must consider this possibility to provide optimal benefits to PDA patients.\n\nMaterials and methods {#s08}\n=====================\n\nPSC isolation {#s09}\n-------------\n\nPSCs were isolated from WT C57BL/6J mice as previously described ([@bib2]) with minor modifications. In brief, pancreata were minced and digested for 30 min at 37\u00b0C in a dissociation buffer containing 0.05% collagenase P (Sigma-Aldrich) and 0.1% DNase I (Sigma-Aldrich) in Grey's balanced salt solution (GBSS; Sigma-Aldrich). Digested pancreata were filtered through a 100-\u00b5m nylon mesh and washed in GBSS with 0.3% BSA and 0.1% DNase I. After spinning, the cell pellet was resuspended in 9.5 ml GBSS with 0.3% BSA and 43.75% Histodenz (Sigma-Aldrich). 6 ml of GBSS with 0.3% BSA was layered on top of the cell suspension, and the gradient was centrifuged for 20 min at 1,400 RCF (with break switched off). The cells in the fuzzy band just above the interface between the Histodenz and GBSS were harvested, washed in PBS, and plated.\n\nMouse models {#s10}\n------------\n\nKPC mice (*Kras^LSL-G12D/+^; Trp53^LSL-R172H/+^; Pdx-1-Cre*) have previously been described ([@bib18]). Each of the three alleles in the KPC mouse strain (Kras; Trp53 and Pdx1-Cre) were backcrossed individually onto the C57BL/6J mouse strain obtained from The Jackson Laboratory (stock number 000664) for at least 20 generations. To generate the KPC mouse model, mice carrying single alleles were crossed onto each other. The Rosa26-LSL-YFP allele ([@bib45]) was backcrossed onto the C57BL/6J mouse strain obtained from The Jackson Laboratory and introduced into the KPC C57BL/6J strain for a total of at least 20 generations to generate KPC; *Rosa26^LSL-YFP/+^* (KPCY). NOD *scid* gamma (NSG) mice were purchased from The Jackson Laboratory (stock number 005557). All animal procedures and studies were conducted in accordance with the Institutional Animal Care and Use Committee (IACUC) at CSHL.\n\nMouse organoid isolation and culture {#s11}\n------------------------------------\n\nMouse tumor organoids were isolated from KPC mice with histologically verified PDA. Tumor tissue was minced and digested at 37\u00b0C for 12 h in a dissociation buffer containing 0.012% (wt/vol) collagenase XI (Sigma-Aldrich) and 0.012% (wt/vol) dispase (Gibco) in DMEM (Gibco) containing 1% FBS (Gibco). The tissue debris was allowed to settle, and the dissociated cells were pelleted and washed in Advanced DMEM/F12 (Invitrogen) and seeded in growth factor-reduced Matrigel (BD). Organoids were cultured in complete organoid media ([@bib8]); Advanced DMEM/F12 supplemented with 1x GlutaMAX (Gibco), 1x Hepes (Gibco), 1x B27 (Invitrogen), 1.25 mM N-Acetylcysteine (Sigma-Aldrich), 10 nM gastrin (Sigma-Aldrich), 50 ng/ml EGF (PeproTech), 10% RSPO1-conditioned media, 100 ng/ml Noggin (PeproTech), 100 ng/ml FGF10 (PeproTech), and 10 mM Nicotinamide (Sigma-Aldrich).\n\nTo passage, organoids were washed out from the Matrigel using cold Advanced DMEM/F12 supplemented with 1x GlutaMAX (Gibco) and 1x Hepes (Gibco), and mechanically dissociated into small fragments using fire-polished glass pipettes, and then seeded into fresh Matrigel. Passaging was performed at a 1:8 split ratio roughly twice per week. To create frozen stocks, organoids were passaged and mixed with Recovery Cell Culture Freezing Medium (Gibco) and cryopreserved using standard procedures. Cultures were thawed using standard thawing procedures, washed once with Advanced DMEM/F12 supplemented with 1x GlutaMAX (Gibco) and 1x HEPES (Gibco), and seeded in Matrigel with organoid media supplemented with 10.5 \u00b5M Y-27632 (Sigma-Aldrich) for the first passage.\n\nMouse CAF isolation {#s12}\n-------------------\n\nMouse CAFs were isolated from KPC mice with histologically verified PDA. CAFs were isolated from tumors using a combination of outgrowth and clonal isolation. The edge of the tumor mass was minced and dissociated in DMEM containing 1% FBS, 0.125 mg/ml collagenase (Sigma-Aldrich), and 0.125 mg/ml dispase (Life Technologies) for 1.5 h at 37\u00b0C in a thermomixer. The sample was trypsinized for 10 min then quenched in 10% FBS/DMEM. The pellet containing tumor pieces was plated on a 6-cm dish, and fibroblasts were allowed to grow out and attach to the plastic. To avoid cancer cell contamination, after the cell culture was established and passaged at least 10 times, cells were plated in high dilution in a 96-well plate to obtain one cell per well, and single clones were expanded. Validation of fibroblast identity was performed at the protein level by flow cytometry analysis for fibroblast surface markers, and at the genomic level by PCR to verify presence of WT and mutant *Kras* alleles. The primers used are: 5\\': GGGTAGGTGTTGGGATAGCTG, and 3\\': TCCGAATTCAGTGACTACAGATGTACAGAG, giving a 285-bp band for WT and 325-bp band for mutant Kras.\n\nHuman organoid and CAF isolation {#s13}\n--------------------------------\n\nAll human organoid experiments were approved by the IRBs of MSKCC, MDACC, and CSHL, and all subjects taking part in the study provided written informed consent. Tumor tissue was minced and digested in a rotating shaker with collagenase II (5 mg/ml; Gibco) for 12 h at 37\u00b0C in human complete media, i.e., complete media ([@bib8]), supplemented with 50% Wnt3a-conditioned media and 1 \u00b5M Prostaglandin E2 (Tocris; [@bib8]). The tissue was further digested for 15 min at 37\u00b0C in TrypLE (Gibco) in a rotating shaker. The digested tissue was washed repeatedly after digestion. 10% of the tissue suspension, including larger undigested pieces, was plated in 10% FBS/RPMI on 10-cm dishes to allow fibroblast outgrowth. The remaining tissue suspension was seeded in Matrigel and cultured in human complete medium for establishment of organoids. Fibroblast outgrowth was observed in the 10-cm dishes within 48 h. Fibroblast identity was confirmed at the DNA level by Sanger sequencing to verify the WT status of *KRAS* at codon 12 (forward primer; 5\u2032-CTGGTGGAGTATTTGATAGTG-3\u2032, reverse primer; 5\u2032-CTGTATCAAAGAATGGTCCTG-3\u2032). To confirm that the established CAF lines had not become transformed in the isolation process, CAFs (2.0--2.5 \u00d7 10^5^ cells) were resuspended in 100 \u00b5l Matrigel and injected subcutaneously in flanks of 6--8-wk-old NOD SCID gamma mice (The Jackson Laboratory).\n\nMouse KPC tumor cell isolation for monolayer cultures {#s14}\n-----------------------------------------------------\n\nThe primary tumor cells used in monolayer experiments were isolated from KPC tumors by an outgrowth method. In brief, 30--50 mg pieces of tumors were minced in DMEM. The minced tissue was then transferred into 5 ml of DMEM with 2 mg/ml Collagenase V (Sigma-Aldrich) and incubated for 45 min at 37\u00b0C in a thermomixer, with vortexing every 5--10 min. After digestion, 10 ml of 10% FBS/DMEM was added, and samples were pelleted, resuspended in 10% FBS/DMEM, and plated on a 10 cm dish. Cultures were passaged at a high split ratio (1:80) for several passages to favor for the growth of neoplastic cells.\n\nCell culture conditions for monolayer cultures {#s15}\n----------------------------------------------\n\nKPC mouse CAFs, mouse PSCs, and KPC primary tumor cells were cultured in DMEM containing 5% FBS, 1% [l]{.smallcaps}-glutamine, and 1% Penicillin/Streptomycin. NIH-3T3 fibroblasts were cultured in DMEM containing 10% FBS, 1% [l]{.smallcaps}-glutamine, and 1% Penicillin/Streptomycin. Human CAFs were cultured in RPMI 1640 (Gibco) containing 10% FBS, 1% [l]{.smallcaps}-glutamine, and 1% Penicillin/Streptomycin. All cells were cultured at 37\u00b0C with 5% CO~2~.\n\nTransfection of fluorophores and immortalization {#s16}\n------------------------------------------------\n\nTo immortalize primary PSCs and human CAFs, the SV40 large T-antigen was cloned from pBABE-puro SV40 LT (Addgene) into pLVX-IRES-tdTomato (Takara Bio Inc.) using the BamHI site. 293T cells were transfected with the pLVX-SV40 LT-IRES-tdTomato plasmid together with psPAX2 and pMD2.G packaging plasmids using X-treme GENE 9 DNA Transfection Reagent (Roche). Media was changed 24 h later, and lentivirus supernatant was collected after an additional 24 h. The supernatant was filtered through a 0.45-\u00b5m filter and aliquots were kept at --80\u00b0C. Primary PSCs and human CAFs were grown to 70% confluency, and infected with the virus supernatant for 24 h. Virus was aspirated and fresh media was added. After passaging, tdTomato was visualized by fluorescent microscopy to verify infection efficiency.\n\nTo obtain GFP-labeled organoids, PGK-Neo-IRES-EGFP retrovirus was produced in ecotropic Phoenix cells, concentrated with RetroX concentrator (Takara Bio Inc.), and resuspended in Advanced DMEM/F12 supplemented with 10 \u00b5M Y-27632 (Sigma-Aldrich). 5 \u00d7 10^4^ cells were resuspended in concentrated retrovirus and spinoculated at 600 RCF for 1 h at room temperature, washed, and then seeded into Matrigel. To obtain mCherry-labeled PSCs or PGK-Neo-IRES-mCherry, retrovirus was produced in ecotropic Phoenix cells and incubated with PSCs grown in monolayer for 24 h. 2 d after infection, cells were treated with 1 mg/ml G418 (Gibco) for selection.\n\nIL-6 CRISPR/Cas9-mediated knockout {#s17}\n----------------------------------\n\nTo knock out IL-6, lenti-Cas9-Blast plasmids (Addgene Plasmid \\#52962) were transfected into 293T cells to produce lentivirus. The virus was concentrated using Lenti-X Concentrator (Takara Bio Inc.), and resuspended in DMEM with 5% FBS. Immortalized PSCs were infected with the resuspended virus and selected using 2 \u00b5g/ml blasticidin (Thermo Fisher Scientific) to create PSC lines stably expressing Cas9. Short guide RNAs (sgRNAs) against IL-6 (5\u2032-CACCTATACCACTTCACAAGTCGG-3\u2032 and 5\u2032-CACCTAAGCCTCCGACTTGTGAAG-3\u2032) were designed using the CRISPR GRNA Design Tool (Atum) and cloned into the LRG plasmid (Lenti-sgRNA-EFS-GFP; Addgene; [@bib42]). Next, LRG lentivirus was produced in 293T cells, concentrated using Lenti-X Concentrator (Takara Bio Inc.), and resuspended in DMEM with 5% FBS. Cas9-expressing PSCs were infected with resuspended LRG virus and efficient cleavage by the sgRNAs was confirmed using the SURVEYOR assay (Transgenomic), following the manufacturer\\'s protocol using a PCR primer set specific to the sgRNA targeted region of IL-6 (forward primer; 5\u2032-CCTCTGGCGGAGCTATTGAG-3\u2032, reverse primer; 5\u2032-CCAGACAGGAAAGGAACCCC-3\u2032). PSCs were then plated as single clones in 96-well plates, and the clones were allowed to expand. Clones that were found to be GFP positive were then analyzed for loss of IL-6. Deletion of IL-6 was confirmed by Sanger sequencing, and by IL-6 ELISA (R&D Systems) of conditioned media from PSCs grown in trans-well cultures with mouse tumor organoids. IL-6--negative clones were combined into IL-6 KO lines, and clones positive for IL-6 secretion were combined into control lines. In total, two IL-6 KO lines and two control lines were made from two parental PSC lines.\n\nCo-culture and passaging of tumor organoids and CAFs {#s18}\n----------------------------------------------------\n\nFor co-cultures, mouse or human organoids were split at a 1:6 ratio and mixed with 1x10^4^ PSCs or CAFs, seeded in Matrigel, and cultured in DMEM containing 5% FBS and 1% Penicillin/Streptomycin. For passaging experiments, established co-cultures and monocultures were split every 7 d at a 1:4 ratio and fresh medium was added. For trans-well cultures, 2 \u00d7 10^4^ PSCs or human CAFs were seeded in Matrigel on top of the trans-well membrane (1 \u00b5m pore size, Greiner Bio-One) with organoids growing in the lower compartment in 24-well plates.\n\nConditioned media experiments {#s19}\n-----------------------------\n\nFor conditioned media experiments, organoids or monolayer cells were cultured for 3--4 d in DMEM containing 5% FBS, 1% [l]{.smallcaps}-glutamine, and 1% Penicillin/Streptomycin. Media was collected, spun down to remove debris, and added to PSCs plated in Matrigel.\n\nFlow cytometry and cell sorting {#s20}\n-------------------------------\n\nFor flow cytometric analysis of mouse CAF and PSC lines, cells were grown in DMEM containing 5% FBS, trypsinized, and subjected to surface staining with anti-FAP (R&D Systems; 1:40 for PSCs; Abcam; 1:50 for mouse CAFs) or anti--mouse CD326 (EpCAM) APC (eBioscience clone G8.8; 1:160). Subsequent antibodies used for FAP staining are anti--goat/sheep IgG--biotin (Sigma-Aldrich; 1:50) and streptavidin APC-Cy7 (BD; 1:200), Goat anti--rabbit Alexa Fluor 488 and goat anti--rabbit Alexa Fluor 568 (Thermo Fisher Scientific; 1:500). Human CAFs were grown in RPMI containing 10% FBS, and stained and analyzed similarly, using the FAP antibody obtained from R&D Systems.\n\nFor intracellular flow cytometric staining, 5 \u00d7 10^5^ cells were treated with GolgiPlug (BD) for 6 h before fixation and permeabilization with Cytofix/Cytoperm kit (BD) according to the manufacturer's instructions. Cells were stained with the appropriate antibodies in FACS buffer (PBS containing 0.2% BSA, 2 mM EDTA, and 0.02% sodium azide) at 4\u00b0C for 30 min in the dark. The following conjugated antibodies were used: APC-Cy7-anti-\u03b1SMA (1A4; Abcore) and PE-anti--IL-6 (MP5-20F3; BioLegend). Fixable viability dye eFluor450 (eBioscience) was used to differentiate between dead and live cells. The stained populations were analyzed using an LSR-II flow cytometer (BD) and FlowJo software (Tree Star; Version 10). At least 20,000 events were recorded per condition.\n\nFor sorting of CAFs from KPC tumors, tumors (\\>8 \u00d7 8 mm in diam) were minced and digested in DMEM containing 10% FBS, 2.5 mg/ml Collagenase D (Roche), 0.5 mg/ml Liberase (Roche), and 0.2 mg/ml DNase I (Sigma-Aldrich), for 45 min at 37\u00b0C in a rotating shaker. Tumor pieces were then strained through a 40-\u00b5m cell strainer, and red blood cells were lysed using ACK lysis buffer for 4 min on ice. After neutralization with flow buffer (5% FBS/PBS), cells were blocked with rat anti--mouse CD16/CD32 (Fc Block; BD; 1:100 for 20 min), and then subjected to staining with anti--mouse CD45-BV510 (BioLegend; clone 30-F11; 1:100), anti--mouse CD326 (Ep-CAM)-Alexa Fluor 488 (BioLegend; clone G8.8; 1:100), anti-mouse CD31-PE/Vio770 (Miltenyi Biotec; clone 390; 1:10), and anti--mouse CD140a (PDGFR\u03b1)-APC (BioLegend; clone APA5; 1:100) for 30 min on ice. Cells were sorted on the FACSAria cell sorter (BD) for CD45^+^, EpCAM^+^, and PDGFR\u03b1^+^ populations.\n\nqPCR {#s21}\n----\n\n1 \u00b5g RNA was used to reverse transcribe cDNA using the TaqMan reverse transcription reagents (Thermo Fisher Scientific). qPCR was performed using gene-specific TaqMan probes (Applied Biosystems) and master mix (Thermo Fisher Scientific), following the manufacturer's instructions. Gene expression was normalized to Hprt. The following TaqMan Probes were used (Mm, mouse probes; Hs, Human probes): Acta2, Mm01546133_m1 and Hs00426835_g1; Cd11b, Mm00434455_m1; Clcf1, Mm01236492_m1; Cntf, Mm04213924_s1; Cntfr, Mm00516693_m1; Col1a1, Mm00801666_g1; Crlf1, Mm00517026_m1; Ctf1, Mm00432772_m1; Ctgf, Mm01192932_g1; Fap, Mm01329177_m1; Gp130, Mm00439665_m1; Has2, Mm00515089_m1; Hprt, Mm00446968_m1, Hs02800695_m1; Il11, Mm00434162_m1; Il11ra1, Mm01218402_m1; Il6, Mm00446190_m1, Hs00985639_m1; Il6ra, Mm00439653_m1; Il27, Mm00461162_m1; Il27ra, Mm00497259_m1; Krt19, Mm00492980_m1; Lif, Mm00434761_m1; Lifr, Mm00442942_m1; Mmp2, Mm00439498_m1; Osm, Mm01193966_m1; Osmr, Mm01307326_m1; Pdgfra, Mm00440701_m1; Pdgfrb, Mm00435546_m1; Vdr, Mm00437297_m1.\n\nOil Red-O staining {#s22}\n------------------\n\nImmortalized PSCs were cultured on plastic or embedded in Matrigel for 4 d and fixed in 10% formalin for 10 min at room temperature, followed by 5-min incubation in 60% isopropanol. After drying, the cells were incubated with filtered Oil Red-O working solution for 10 min. The working solution was prepared by dissolving 0.35 g Oil Red-O (Sigma-Aldrich) in 100 ml isopropanol, which was then diluted in water at a 3:2 ratio. After staining, the cells were washed four times with water and imaged.\n\nTGF\u03b2 stimulation {#s23}\n----------------\n\nCells were treated with 20 ng/ml of recombinant human TGF\u03b21 (Sigma-Aldrich) for 4--5 d before RNA isolation and qPCR analysis.\n\nImmunofluorescence staining of monocultures and co-cultures {#s24}\n-----------------------------------------------------------\n\nCultures were fixed with 2% PFA for 20 min at room temperature, and washed three times for 10 min in 1x PBS/Glycine solution (10\u00d7 stock: 38.0 g NaCl, 9.38 g Na~2~HPO~4~, 2.07 g NaH~2~PO~4~, and 37.5 g glycine, in 500 ml PBS). Three washes in 1\u00d7 wash solution were performed (10\u00d7 stock: 38.0 g NaCl, 9.38 g Na~2~HPO~4~, 2.07 g NaH~2~PO~4~, 2.5 g NaN~3~, 5.0 g BSA, 10 ml Triton X-100, and 2.5 ml Tween-20, in 500 ml PBS), followed by 1 h of blocking in 1\u00d7 wash solution with 10% horse serum. The cultures were then incubated over night at 4\u00b0C with rabbit anti-collagen I antibody (ab34710, Abcam) diluted 1:200 in 1x blocking solution. After three washes, the cells were incubated for 1 h with Alexa-568 Goat anti-Rabbit IgG antibody (Invitrogen) diluted 1:1000 in 1\u00d7 blocking solution. Counterstain, DAPI (Sigma-Aldrich). After three washes, the slides were mounted and imaged.\n\nISH {#s25}\n---\n\nRNA ISH was performed on freshly prepared formalin-fixed paraffin-embedded (FFPE) tissue sections (6 \u00b5m thickness) using the ViewRNA ISH Tissue 2-Plex Assay (Affymetrix), according to the manufacturer\\'s instructions using 10-min pretreatment and protease treatment incubation times. ISH probes used were as follows: mouse \u03b1SMA, VB1-16010-01; human \u03b1SMA, VA1-10300; mouse FAP, VB1-16010-01; mouse IL-6, VB1-10012; mouse Krt18, VB6-11059; human Krt18, VA6-11561.\n\nImmunohistochemical/immunofluorescent staining of tissues {#s26}\n---------------------------------------------------------\n\nAll stainings were performed on 5-\u00b5m sections of mouse and human tissues. Hematoxylin and Eosin (H&E) staining was performed according to standard protocols. For immunohistochemistry (IHC), sections were deparaffinized, and antigen retrieval was performed in a pressure cooker in 10 mM sodium citrate buffer, pH 6.0. 3% H~2~O~2~ was used to block endogenous peroxidases. Primary antibodies used for IHC were as follows: IL-6 (1:50 for mouse; Cell Signaling Technology; 1:100 for human; Abcam), GFP (to detect YFP; 1:1,000; Abcam), Ki67 (1:250; Thermo Fisher Scientific), PDGFR\u03b2 (1:400 for mouse and 1:200 for human; Abcam), and \u03b1SMA (1:200; Abcam). Hematoxylin was used as nuclear counterstain. For double and triple IHC on mouse tissue, a sequential IHC protocol was applied (Vector Laboratories) and the substrates used were as follows: ImmPACT DAB peroxidase (brown), ImmPACT VIP peroxidase (purple), and ImmPACT SG peroxidase (gray). The same protocol was used for double IHC staining of IL-6 and PDGFR\u03b2 on human tissue, using an automated Ventana Benchmark staining machine (Ventana Medical Systems).\n\nFor IF of pSTAT3 in KPC tumors, frozen sections were first fixed in 3% formaldehyde for 15 min at room temperature, and then fixed in methanol for 10 min at --20\u00b0C. Slides were immediately washed in PBST (1% Tween-20, PBS) after fixation and blocked with 10% goat serum in PBST and 0.3% Triton X-100 for 30 min at room temperature. Slides were then stained with Phospho-Stat3 (1:200; Cell Signaling Technology) and Krt19 (1:200; DSHB TROMA-III) antibodies for 1 h at room temperature. Slides were washed in PBST, and then stained with anti--rabbit Alexa Fluor 488 (1:500; Thermo Fisher Scientific), and anti--rat Alexa Fluor 568 (1:500; Thermo Fisher Scientific) secondary antibodies for 45 min at room temperature. After washing in PBST, sections were stained with DAPI for 5 min at room temperature and coverslipped with Histomount Mounting Solution (Thermo Fisher Scientific).\n\nFor IF of \u03b1SMA and IL-6, FFPE slides were used, and deparaffinization and antigen retrieval were performed according to the IHC protocol. Slides were incubated with primary antibodies for \u03b1SMA (1:100; Dako) and IL-6 (1:50; Cell Signaling Technology) overnight at 4\u00b0C. Slides were washed with TBS (Tris-buffered saline) and stained with anti--mouse Alexa Fluor 488 and anti--rabbit Alexa Fluor 568 (1:1,000; Invitrogen) secondary antibodies for 1 h at room temperature. DAPI was used as counterstain.\n\nFor IF on human tissue, frozen tissue sections were fixed in acetone for 10 min. Slides were blocked for 1 h in 3% BSA/PBS at room temperature, and incubated overnight at 4\u00b0C with primary antibodies for FAP (1:50; R&D Systems) and \u03b1SMA (1:200; Sigma-Aldrich). After washes, the slides were incubated for 1 h at room temperature with anti--mouse Alexa Fluor 488 (1:500; Invitrogen) and anti--sheep Alexa Fluor 555 (1:500; Invitrogen) secondary antibodies, then washed and mounted with Vectashield Hardset containing DAPI (Vector Laboratories). The use of human tissues in this study was approved by The Ethics Review Board (EPN) of Northern Sweden.\n\nImaging {#s27}\n-------\n\nLive cell and immunofluorescence imaging of monocultures and co-cultures was done on a Perkin Elmer Ultraview Vox spinning disc confocal system (Waltham MA) using the Velocity 6.3 software. The system consisted of a Nikon Ti Eclipse inverted microscope (Morrell Instruments Melville NY), Yokogawa CSU-Xi spinning disk, Perkin Elmer laser module 2.0 (6 lines), Hamamatsu R2 CCD, motorized peizo Z stage from Applied Scientific Imaging (Eugene OR), environmental chamber from In Vivo Scientific, temperature regulation with the Smart Air-Therm heater from World Precision Instruments (Sarasota FL), atmosphere was provided using premixed hematology gas with 5% CO~2~.\n\nFluorescence imaging of fixed tissue was done with a Leica TCS SP8 laser scanning confocal (Boulder Grove Il), controlled by the LAS AF 3.3.10134 software. This confocal was mounted on a DMI 6000 CS inverted microscope, equipped with 4 laser lines and 2 PMTs. Bright field imaging of tissue slides were obtained with an Axio Imager.A2 (ZEISS).\n\nQuantification of fluorescence intensity was performed with the Velocity 6.3 software.\n\nElectron microscopy {#s28}\n-------------------\n\nCo-cultures were processed for electron microscopy as previously described ([@bib5]). In brief, co-cultures were fixed overnight with 2% glutaraldehyde and 2% PFA in PBS, rinsed in distilled water and post-fixed with 1% osmium tetroxide in 1.5% potassium ferrocyanide for 1 h. After dehydration, samples were resin infiltrated and polymerized overnight at 60\u00b0C. 100 nm thin sections were collected on 100 mesh grids with and without a supporting film. Following counterstaining with lead citrate, sections were examined in a Hitachi H-7000 transmission electron microscope operated at 75 kV. Images were recorded on Kodak 4489 film and scanned at 2400 dpi.\n\nWestern blot {#s29}\n------------\n\nPSCs or organoids were harvested in Cell Recovery Solution (Corning) and incubated rotating for 1 h at 4\u00b0C. Cells were then pelleted, and lysed in 0.1% Triton X-100, 15 mM NaCl, 0.5 mM EDTA, 5 mM Tris, pH 7.5 supplemented with protease Mini-complete protease inhibitors (Roche) and a phosphatase inhibitor cocktail (PhosSTOP; Roche). Cells were incubated on ice for 30 min before clarification. Standard procedures were used for Western blot. In brief, protein lysates were separated by SDS-PAGE, transferred to a polyvinylidene difluoride (PVDF) membrane, blocked with 5% BSA in TBST (1% Tween 20, tris-buffered saline), and incubated with primary antibodies overnight at 4\u00b0C. Proteins were detected using HRP-conjugated secondary antibodies. Primary antibodies used were: PDGFR\u03b1 (Cell Signaling Technology), PDGFR\u03b2 (Cell Signaling Technology), \u03b1SMA (Dako), Hsp90\u03b1 (EMD Millipore), Actin (Cell Signaling Technology), Phospho-STAT3 (Cell Signaling Technology), STAT3 (Cell Signaling Technology), Pan-Actin (Cell Signaling Technology).\n\nSecretome analysis {#s30}\n------------------\n\nConditioned media was collected after incubation with monocultures of PSCs, CAFs, mouse or human organoids, co-cultures, or Matrigel-only controls for 3--4 d, filtered through a 0.45-\u00b5m filter to remove cell debris, and frozen at --20\u00b0C. Frozen conditioned media was then thawed and used for cytokine dot blots (R&D Systems) according to the manufacturer\\'s instructions.\n\nELISA {#s31}\n-----\n\nFor ELISA of conditioned media of co-cultures or monocultures of PSCs, human CAFs, and mouse and human tumor organoids, cultures were grown in DMEM containing 5% FBS for 3--5 d. Media was then collected, filtered through a 0.45-\u00b5m filter to remove cell debris, aliquoted, and frozen at --20\u00b0C. Thawed media was assayed using the manufacturer\\'s protocol. The following ELISA assays used were: IL-6 (R&D Systems), IL-11 (R&D Systems), and LIF (R&D Systems).\n\nDot blot and Western blot quantification {#s32}\n----------------------------------------\n\nDot blots and Western blot films were scanned at 600 dpi and then quantified in ImageJ (National Institutes of Health). Scanned blots were background subtracted using a rolling ball radius of 50 pixels, and then inverted before being quantified using the integrated pixel density function. For Western blots, a rectangular region was used to quantify each band. Quantification is presented as pixel intensity of pSTAT3 normalized to total STAT3, loading control, and the control condition. For dot blots, a circular region was used to quantify each pair of dots, which was then averaged.\n\nProliferation assays {#s33}\n--------------------\n\nGFP-expressing organoids and mCherry-expressing PSCs were dissociated into single cells and counted. Both cell types were mixed in a 1:1 ratio in 50% Matrigel and 50% media (DMEM with 5% FBS). 70 \u00b5l of a cell suspension containing 7,000 cells of each cell type was plated per well on black 96-well plates with clear bottoms (Corning) on ice. After 30-min incubation in a cell incubator, 200 \u00b5l prewarmed media containing 10.5 \u00b5M RhoK-inhibitor (Sigma-Aldrich) was added on top of the cells. Fluorescence intensity was measured for both GFP and mCherry once a day for up to 7 consecutive days, on a SpectaMax I3 (Molecular Devices) by scanning each well at 52 points and averaging the intensity.\n\nWhen proliferation of nonfluorescent PSCs was measured, 5,000 cells were seeded in 96-well plates in 50% Matrigel/PBS and cultured in 150 \u00b5l control media or conditioned media. Cell proliferation was followed for 5 d with CellTiter-Glo (Promega), with measurements every 24 h.\n\nRNA seq {#s34}\n-------\n\nSamples were collected in 1 ml of TRIzol Reagent (Thermo Fisher Scientific) and stored at \u221280\u00b0C. RNA was extracted using the PureLink RNA mini kit (Thermo Fisher Scientific). RNA quality was assessed on a bioanalyzer using the Agilent RNA 600 Nano kit. We used poly-A pull-down to enrich for mRNAs from total RNA samples (0.2--1 \u00b5g per sample, RIN \\> 8) and proceeded to library preparation using Illumina TruSeq RNA prep kit. Libraries were then sequenced using Illumina HiSeq2000 at the Columbia Genome Center. We multiplexed samples in each lane, which yielded targeted number of single-end 100-bp reads for each sample, as a fraction of 180 million reads for the whole lane. We used RTA (Illumina) for base calling and bcl2fastq (version 1.8.4) for converting BCL to fastq format, coupled with adaptor trimming. We mapped the reads to a reference genome (Human, NCBI/build37.2; Mouse, UCSC/mm9) using Tophat ([@bib51]; version 2.0.4) with 4 mismatches (--read-mismatches = 4) and 10 maximum multiple hits (--max-multihits = 10). To tackle the mapping issue of reads that are from exon--exon junctions, Tophat infers novel exon--exon junctions ab initio, and combined them with junctions from known mRNA sequences (refgenes) as the reference annotation. We estimated the relative abundance (i.e., expression level) of genes and splice isoforms using cufflinks ([@bib52]; version 2.0.2) with default settings. All RNA-seq data are available at Gene Expression Omnibus (GEO) under the accession no. [GSE93313](GSE93313).\n\nDifferential expression analysis and GSEA {#s35}\n-----------------------------------------\n\nGenes expressed in fewer than two libraries were filtered out before differential expression testing. The principle component analysis was calculated using the prcomp function available in R and plotted using a customized R script. Expression normalization and differential expression testing were performed using DESeq ([@bib1]), with dispersion estimation parameters set as: 'method = \"per-condition\", sharingMode = \"maximum\", fitType = \"parametric.\"' Genes with adjusted P \\< 0.01 were selected as significantly differentially expressed between conditions. All plots were produced using customized R scripts.\n\nGSEA on the RNA-seq data were performed by entering fold change data from the differential expression analysis into the GSEA software (Broad Institute) using the Gene sets database c2.cp.v5.1.symbols.\n\nStatistics {#s36}\n----------\n\nFor graphical representation of data and statistical analysis, GraphPad Prism was used. Statistical analysis was performed using Student's *t* test. If multiple *t* tests of the same dataset were performed, correction for multiple comparisons was made using the Sidak-Bonferroni method. Data are presented as mean of biological replicates \u00b1 SD unless otherwise indicated.\n\nOnline supplemental material {#s37}\n----------------------------\n\nFig. S1 shows the isolation process and validation data for the PSCs and human CAFs used in this study. Fig. S2 shows supplementary data from investigating fibroblast heterogeneity in the co-culture model system and the isolation process and validation data for the mouse KPC CAFs used in this study. Table S1, provided as an Excel file, contains the RNA expression analysis (DeSEQ) and pathway analysis (GSEA) comparing quiescent PSCs, iCAFs, and myofibroblasts.\n\nThe authors would like to thank the Cold Spring Harbor Cancer Center Support Grant (CCSG) shared resources: Bioinformatics Shared Resource, Flow Cytometry Facility, the St. Giles Foundation Microscopy Center, Animal & tissue imaging, and the Animal Facility. The CCSG is funded by the National Cancer Institute. We also thank Anette Berglund for technical assistance, and Maya Ridinger, Marina Pasca di Magliano, Yaqing Zhang, Greg Beatty, Anil Rustgi, and Valerie Weaver for constructive advice and discussion.\n\nThis work was supported by the Lustgarten Foundation, where D.A. Tuveson is a distinguished scholar and Director of the Lustgarten Foundation--designated Laboratory of Pancreatic Cancer Research. D.A. Tuveson is also supported by the Cold Spring Harbor Laboratory Association, the National Institutes of Health (NIH; 5P30CA45508-26, 5P50CA101955-07, 1U10CA180944-02, 5U01CA168409-5, 1R01CA188134, and 1R01CA190092-03). In addition, we are grateful for support from the following: Stand Up to Cancer/KWF (D.A. Tuveson), the STARR Foundation (I7-A718 for D.A. Tuveson), DOD (W81XWH-13-PRCRP-IA for D.A. Tuveson), the Precision Medicine Research Associates (PMRA; H. Tiriac and D.A. Tuveson), the Swedish Research Council (537-2013-7277 for D. \u00d6hlund), the Kempe Foundations (JCK-1301 for D. \u00d6hlund) and the Swedish Society of Medicine (SLS-326921, SLS-250831, SLS-175991, and SLS-591551 for D. \u00d6hlund), federal funds through the county council of V\u00e4sterbotten (ALFVLL269081, VLL242121, VLL322391, VLL400421, VLL493371, and VLL582681 for D. \u00d6hlund), the Cancer Research Foundation in Northern Sweden (AMP15-793 and LP11-1927 for D. \u00d6hlund), the Human Frontiers Science Program (LT000403/2014 for E. Elyada; LT000195/2015-L for G. Biffi; LT000190/2013 for I.I.C. Chio), The Weizmann Institute of Science Women in Science award (for E. Elyada), EMBO (ALTF 1203-2014 for G. Biffi), the Italian Ministry of Health (FIRB - RBAP10AHJ for V. Corbo), Associazione Italiana Ricerca Cancro (AIRC n. 18718 for V. Corbo), NIH (R50CA311506-01 for Y. Park), NIH awards (CA101955 UAB/UMN SPORE, 5T32CA148056, and F32CA192904 for L.A. Baker), the Damon Runyon Cancer Research Foundation (Shirley Stein fellow, DRG-2165-13, for I.I.C. Chio), the National Cancer Institute (NCI 1K99CA204725-01A1 for D.D. Engle), and Nancy Gay fellowship (for A.S. Almeida).\n\nThe authors declare no competing financial interests.\n\nAbbreviations used:\u03b1SMA\u03b1-smooth muscle actinBM-MSCBM-derived mesenchymal stem cellCAFcancer-associated fibroblastECMextracellular matrixELISAenzyme-linked immunosorbent assayEMTepithelial-to-mesenchymal transitionFAPfibroblast-activation proteinGSEAGene Set Enrichment AnalysisiCAFinflammatory CAFIFimmunofluorescenceIHCimmunohistochemistryISHin situ hybridizationmyCAFmyofibroblastic CAFPDApancreatic ductal adenocarcinomaPSCpancreatic stellate cell\n\n[^1]: D. \u00d6hlund, A. Handly-Santana, G. Biffi, and E. Elyada contributed equally to this paper.\n"}
+{"text": "1. Background\n=============\n\nHypertension is a common chronic disease that affects a great proportion of the population worldwide. It is considered one of the major risk factors for the development of cardiovascular diseases such as stroke, ischemic heart disease, and heart failure.^\\[[@R1]--[@R4]\\]^ Despite the recent advances in the development of antihypertensive therapeutic treatments and in the pathophysiology of hypertension, the prevalence of hypertension in China remains considerably high, whereas the rates of blood pressure (BP) control are notably low. The outline of the Report on Cardiovascular Disease in China (2014) demonstrated that the hypertension control (BP \\<140/90\u200amm\u200aHg) rate was approximately 9.3% of all patients, and/or 27.4% of treated patients, although 42.6% of hypertensive patients in China were aware of their condition and 34.1% were receiving treatment.^\\[[@R5]\\]^ Furthermore, the BP control rate was lower in female subjects and rural patients compared with that in male subjects and urban patients, respectively. The electrocardiogram (ECG) is a routine, accessible, cost-effective, and recommended diagnostic tool for the initial evaluation and follow-up of hypertensive patients. In the standard surface ECG, nonspecific ST-segment and T-wave (ST-T) changes are a common finding. Recently, a majority of studies have indicated that nonspecific ST-T abnormalities are significantly associated with cardiovascular morbidity and mortality.^\\[[@R6]--[@R9]\\]^ However, the prognostic significance of nonspecific ST-T abnormalities for BP control has not been fully investigated.\n\nThe goal of the present study was to determine whether nonspecific baseline ST-T ECG abnormalities are associated with unsatisfactory BP control in hypertensive patients. The study aimed to provide information on the potential improvement of the management of hypertension.\n\n2. Methods\n==========\n\n2.1. Study population\n---------------------\n\nThe present study population was a subset of the China Stroke Primary Prevention Trial (CSPPT).^\\[[@R2]\\]^ The detailed description of the stage designed and the methodology have been previously described.^\\[[@R2]\\]^ Eligible participants were men and women 45 to 75 years of age who had hypertension. The major exclusion criteria included the history of stroke, myocardial infarction, heart failure, coronary revascularization, and/or congenital heart disease. Individuals with a missing ECG measurement, an illegible ECG, and/or ECG abnormalities inconsistent with nonspecific ST-T changes (abnormal Q wave, atrioventricular block, left ventricle high voltage, and arrhythmia) at baseline (Fig. [1](#F1){ref-type=\"fig\"}) were further excluded. A total of 15,038 participants with hypertension were selected for the present analysis.\n\n![Flow of nonspecific ST-T changes in the China Stroke Primary Prevention Trial. ST-T = ST-segment and T wave.](medi-96-e6423-g001){#F1}\n\nThe CSPPT protocol was approved by the ethics committee of the Institute of Biomedicine, Anhui Medical University, Hefei, China (FWA assurance number FWA00001263). All participants provided written informed consent for their participation in the study protocol. The study protocol was conducted according to the principles of the Declaration of Helsinki that ensure that the safety and well-being of the patients is protected and that the integrity of the data is preserved.\n\n2.2. Study protocol and evaluation criteria\n-------------------------------------------\n\nThe CSPPT was a multicommunity, randomized, double-blinded clinical trial conducted between May 19, 2008, and August 24, 2013, in 32 communities in the Jiangsu and Anhui provinces of China. The present study was a prospective cohort study. The 15,038 participants with baseline ECG measures were stratified according to the presence and/or absence of nonspecific ST-T abnormalities using the Minnesota coding criteria. Eligible participants were subsequently divided into 2 groups: ST-T abnormal and ST-T normal. All ECGs were recorded and analyzed by 2 trained medical professionals. In the case of a disagreement regarding the interpretation of an ECG, a third reading was conducted jointly until a final interpretation of the ECG was achieved.\n\n2.3. Criteria for ECG Definitions\n---------------------------------\n\nThe Minnesota code (MC) criteria for nonspecific ST-T abnormalities were used, as described by the MC ECG classifications 4-3, 4-4, 5-3, and 5-4. The criteria were defined as follows: no ST-J depression \u22650.5\u200amm but ST-segment downward sloping and ST-segment or T-wave nadir at least 0.5\u200amm below the P-R baseline, in any of leads I, II, aVL, or V~2~ to V~6~ (MC 4-3); ST-J depression \u22651.0\u200amm and ST-segment upward sloping or U-shaped, in any of leads I, II, aVL, or V~1~ to V~6~ (MC 4-4); T-wave amplitude zero (flat), negative, or diphasic (negative--positive type only) with \\<1.0-mm negative phase in leads I, II, V~3~ to V~6~, aVL when R-wave amplitude is \u22655.0\u200amm (MC 5-3); and T-wave amplitude positive and T- to R-wave amplitude ratio of \\<1:20 in any of leads I, II, aVL, or V~3~ to V~6~ when R-wave amplitude in the corresponding leads was \u226510.0\u200amm (MC 5-4).\n\nHypertension was defined as seated resting systolic blood pressure (SBP) of \u2265140\u200amm\u200aHg or diastolic blood pressure (DBP) of \u226590\u200amm\u200aHg at both the screening and recruitment visits and/or use of antihypertensive medication. The BP of the participants was measured during the follow-up period every 3 months from 2008 to 2013. During each visit, the participants were required to rest in a seated position for at least 5 minutes prior to the measurement. BP was measured 3 times at 5-minute intervals by a trained physician with an electronic sphygmomanometer (Omron; Dalian, China). The mean of the 3 readings was calculated for the measurement of the BP. The average of 20 measurements at 20 different visits was used as the final BP value. Overall unsatisfactory BP control was defined as SBP of \u2265140\u200amm\u200aHg or DBP of \u226590\u200amm\u200aHg following antihypertensive treatment.\n\nDiabetes mellitus was defined as self-reported clinically diagnosed diabetes or use of hypoglycemic agents or a fasting blood glucose concentration of \u22657.0\u200ammol/L (\u22657.0\u200ammol/L). Body mass index (BMI) was calculated as weight (in kilograms) divided by height (in square meters).\n\nSerum folate and vitamin B~12~ at both the baseline and the exit visits were measured by a commercial laboratory kit using a chemiluminescent immunoassay (New Industrial; Shenzhen, China). Serum homocysteine, fasting lipids, and glucose levels at the baseline and the exit visit were measured using automatic clinical analyzers (Beckman Coulter; California, America) at the core laboratory of the National Clinical Research Center for Kidney Disease, in Nanfang Hospital, Guangzhou, China.\n\n2.4. Statistical analysis\n-------------------------\n\nData were analyzed using the Empower Stats software. The results are presented as mean\u200a\u00b1\u200astandard deviation for continuous variables and frequence(percentage) for categorical variables. Baseline characteristics were compared between participants in the presence and/or absence of nonspecific ST-T abnormalities using \u03c7^2^ tests for categorical variables and *t* tests and/or Wilcoxon rank-sum tests for continuous variables, as appropriate. The prospective association between nonspecific ST-T changes and unsatisfactory BP control was examined using Cox proportional hazard regression models following adjustment for the covariates. Multiple linear regression was used to assess the association between ST-T abnormality and change of blood pressure under treatment. Furthermore, multivariate logistic regression analysis was used to evaluate the impact of electrocardiographic nonspecific ST-T abnormalities on unsatisfactory BP control in the subgroup analyses. A 2-tailed *P*-value \\<0.05 was considered statistically significant.\n\n3. Results\n==========\n\nThe study population included 15,038 participants (62.1% women) from the CSPPT, which is representative of a low- to medium-risk Chinese hypertensive population. The BP control rate for the total population was 53.0%. The prevalence of nonspecific ST-T abnormalities in the population investigated was 8.5%, whereas prevalence was greater in women (10.3%) and diabetics (13.9%).\n\nThe baseline characteristics are shown in Table [1](#T1){ref-type=\"table\"}, according to the presence and/or absence of nonspecific ST-T abnormalities for the entire study population. The baseline characteristics were further stratified according to gender (Table [1](#T1){ref-type=\"table\"}). Following antihypertensive treatment, the ST-T abnormal group exhibited significantly greater SBP, greater change from baseline to final measurement in SBP and DBP, and a greater relative percentage decrease in both SBP and DBP from baseline to the last measurement. In addition, the unsatisfactory BP control rate was significantly greater (54.5%) in the total population, compared with that in individuals with an absence of nonspecific ST-T abnormalities (46.3%).\n\n###### \n\nBaseline and follow-up characteristics of the study participants.\n\n![](medi-96-e6423-g002)\n\nThe presence and absence of nonspecific ST-T abnormalities significantly differed among related demographic and clinical variables (Table [1](#T1){ref-type=\"table\"}). This trend potentially affected the outcome (failure to achieve BP treatment goals) (Table [1](#T1){ref-type=\"table\"}). The independent relationship between the outcome and the presence and/or absence of nonspecific ST-T abnormalities was examined following adjustment for the possible effects of the following parameters: center, treatment group, self-reported arrhythmia, antihypertensive drug use, \u03b2-blocker use, smoking status and alcohol consumption, baseline SBP and DBP, gender, age, pulse rate, glucose, homocysteine, folate, vitamin B~12~, total cholesterol level, high-density lipoprotein cholesterol, and BMI (Table [2](#T2){ref-type=\"table\"}). The unsatisfactory BP control rate of the total population was significantly higher in the ST-T abnormal group (odds ratio \\[OR\\] 1.20, 95% confidence interval \\[CI\\] \\[1.06, 1.36\\], *P*\u200a=\u200a0.005\\]), whereas with regard to the stratified categories significantly different results were obtained in male (OR 1.51, 95% CI \\[1.17, 1.94\\], *P*\u200a=\u200a0.002) and not in female subjects (OR 1.11, 95% CI \\[0.96, 1.28\\], *P*\u200a=\u200a0.160).\n\n###### \n\nAssociation between baseline electrocardiographic ST-T abnormality and failure to achieve blood pressure treatment goals.\n\n![](medi-96-e6423-g003)\n\nThe associations between baseline electrocardiographic ST-T abnormalities and changes in BP following treatment are shown in Table [3](#T3){ref-type=\"table\"}. Following adjustment for the covariables described in Table [2](#T2){ref-type=\"table\"}, the patients with nonspecific ST-T abnormalities exhibited significantly lower changes in SBP and DBP (\u03b2 1.39, 95% CI \\[0.82, 1.96\\], *P*\u200a\\<\u200a0.001, and \u03b2 0.50, 95% CI \\[0.19, 0.82\\], *P*\u200a=\u200a0.002, respectively). In addition, these patients revealed lower values in the relative percentage decrease in SBP and DBP (\u03b2 0.80, 95% CI \\[0.45, 1.15\\], *P*\u200a\\<\u200a0.001, and \u03b2 0.54, 95% CI \\[0.20, 0.88\\], *P*\u200a=\u200a0.002, respectively). Importantly, the aforementioned differences were noted in male and not female subjects.\n\n###### \n\nAssociation between baseline electrocardiographic ST-T abnormality and change of blood pressure under treatment.\n\n![](medi-96-e6423-g004)\n\nTable [4](#T4){ref-type=\"table\"} indicates the results of the multivariate logistic regression models regarding the assessment of the impact of electrocardiographic ST-T abnormalities on unsatisfactory BP control in the subgroup analyses. Following adjustment for the confounders described in Table [2](#T2){ref-type=\"table\"}, the differences were further observed between the following subgroups: patients with enalapril and folic acid treatment, patients from the Jiangsu center, male subjects, older patients at an age \u226560 years, middle and high baseline SBP tertiles, high baseline DBP tertile, low BMI, and comorbid diabetes. The unsatisfactory BP control rate for the aforementioned subjects was significantly greater in the presence of nonspecific ST-T abnormalities. Among these patients notable differences were noted in male subjects (OR 1.51, 95% CI \\[1.17, 1.94\\], *P*\u200a=\u200a0.002) and subjects with comorbid diabetes (OR 1.47, 95% CI \\[1.04, 2.07\\], *P*\u200a=\u200a0.029).\n\n###### \n\nMultivariate logistic regression^\u2217^ evaluating the impact of electrocardiographic ST-T abnormality on unsatisfactory BP control in subgroup analyses.\n\n![](medi-96-e6423-g005)\n\n4. Discussion\n=============\n\nThe major advantage of the present study was the large sample size derived from the CSPPT. The CSPPT was a large randomized trial conducted in adult subjects with hypertension in China without a history of stroke or myocardial infarction. The results of the trial demonstrated that enalapril--folic acid therapy significantly reduced the relative risk of an initial stroke incident by 21% compared with enalapril monotherapy. Given the well-characterized population, the standardized assessment of ECGs using the MC criteria, and the 4.5-year longitudinal follow-up period, the CSPPT study provided a unique opportunity to assess the prevalence of nonspecific ST-T changes in order to identify differences between genders and to explore whether nonspecific ST-T changes were associated with increased risk of unsatisfactory BP control. The study further provided insight regarding the potential predictive value of unsatisfactory BP control.\n\nIn the present study, the rate for the achievement of target BP levels of the total population was 53.0% and that was considerably greater compared with that previously reported in a national survey among adult participants in China (27.4% among treated hypertensive participants).^\\[[@R5]\\]^ The possible explanations of the aforementioned findings may be the rigorous design of the CSPPT and the strict follow-up schedule applied by trained research staff and physicians, which ensured high adherence of patients to the prescribed treatment regimen. The control rate of the present study conducted in China was similar to that noted in 2 studies conducted in America (55.8%, result from the HATT study)^\\[[@R10]\\]^ and Italy (55.6%--66.3%, analysis of a large database).^\\[[@R11]\\]^\n\nThe current study demonstrated that nonspecific ST-T changes in the ECG of hypertensive patients can be used for the identification of a greater risk of unsatisfactory BP control, notably in male participants and patients with diabetes mellitus. The prevalence of nonspecific ST-T abnormalities was 8.5%, and the prevalence was greater in women (10.3%) and in patients with diabetes (13.9%). This finding is consistent with previous studies.^\\[[@R7],[@R12]--[@R16]\\]^ The greater prevalence of nonspecific ST-T changes in women has been attributed to the interplay of a variety of factors of anatomic, structural, hormonal, autonomic, and genetic origin. High blood glucose is associated with comorbidities such as hypertension, hyperlipidemia, and a prothrombotic state that interact synergistically to promote cardiac changes. The latter changes in turn result in ECG abnormalities, including nonspecific ST-T changes.\n\nAs depicted in Table [1](#T1){ref-type=\"table\"}, SBP was significantly greater in hypertensive patients with baseline nonspecific ST-T abnormalities, which was in agreement with a previously reported study by Vinyoles et al.^\\[[@R17],[@R18]\\]^ As a result, nonspecific ST-T changes could be considered an early indicator of poor BP management. Following adjustment for the covariables, the current study indicated that hypertensive patients with baseline nonspecific ST-T abnormalities exhibited a higher rate of unsatisfactory BP control (OR 1.20, 95% CI \\[1.06, 1.36\\], *P*\u200a=\u200a0.005) (Table [2](#T2){ref-type=\"table\"}), and significantly lower changes in SBP and DBP (Table [3](#T3){ref-type=\"table\"}). Lower degrees in relative percentage decreases of SBP and DBP following antihypertensive treatment were further noted (Table [3](#T3){ref-type=\"table\"}).\n\nTo date, there are no conclusive data regarding the association of nonspecific ST-T abnormalities with specific pathophysiologic mechanisms. It has been suggested that nonspecific ST-T changes might represent subclinical coronary artery disease, early left ventricular hypertrophy, increased left ventricular mass, or autonomic imbalance.^\\[[@R7],[@R19]\\]^ The aforementioned parameters may increase the risk of unsatisfactory BP control, although further studies are required to clarify this hypothesis.\n\nRecently, the majority of studies have demonstrated that nonspecific ST-T abnormalities are significantly associated with cardiovascular events and cerebrovascular accidents in hypertensive patients.^\\[[@R20],[@R21]\\]^ In the present study, it was speculated that unsatisfactory BP control might play an intermediary role in this process. However, the exact mechanism of action remains unclear and further research work is required to address this issue.\n\nIn subgroup analyses, following adjustment for confounders, male subjects and the patients with comorbid diabetes with baseline nonspecific ST-T abnormalities exhibited notably greater rates of unsatisfactory BP control compared with the patients who exhibited no abnormalities. The greater risk of unsatisfactory BP control in male compared with female subjects may be related to the lack of the cardioprotective effect of estrogen, which has vasodilating and antioxidant properties, and is considered to influence cardiac natriuretic peptides via the renin--angiotensin system.^\\[[@R22]\\]^ In addition, high blood glucose increased the risk of unsatisfactory BP control, presumably via the association with cardiovascular target organ damage. Further studies are required to demonstrate these theories.\n\nIn conclusion, the identification of nonspecific ST-T abnormalities in the ECG of hypertensive patients is of considerable significance due to their potential application as markers for unsatisfactory BP control. Furthermore, the study is of particular importance as regards the majority of the Chinese population where BP is inadequately controlled. The findings may provide the basis for more intensive management of hypertensive patients who display electrocardiographic nonspecific ST-T abnormalities, notably for male subjects and/or patients with comorbid diabetes. In addition, it may be necessary for clinicians to prescribe more potent antihypertensive drugs during the course of hypertension treatment.\n\nECG machines are readily available in the majority of healthcare facilities and clinics in China. Consequently, ECG screening comprises a potentially accessible and affordable risk assessment tool in primary care settings in order to aid the management of BP control.\n\n5. Limitations\n==============\n\nOne of the limitations encountered was the measurement of the cardiac activity by a single ECG during the baseline physical examination that prevented the exclusion of subsequent abnormalities during the follow-up period. Second, it is well established that nonspecific ST-T abnormalities have been associated with transient physiologic phenomena, namely ingestion of food, change in posture, and/or emotional distress. Additional postulated explanations regarding the nonspecific ST-T abnormalities include central nervous system lesions, abnormalities in the left ventricular wall motion in the absence of coronary artery disease, persistent juvenile pattern, electrolyte disturbances, use of drugs (i.e., digitalis, antiarrhythmic, and psychotropic drugs), and/or athletic ability.^\\[[@R7]\\]^ These processes limit the reproducibility of ST-T segment changes in successive ECG measurements. Third, BP measurements were not carried out at trough for the patients who received antihypertensive drugs and consequently the assessment of BP control could have been influenced to a certain extent. In the future, ambulatory BP monitoring may be a favorable selection for BP management and evaluation.\n\nThe authors appreciate the participants who volunteered to participate in the study, and the data collection staff of the CSPPT team.\n\nAbbreviations: BMI = body mass index, BP = blood pressure, CSPPT = China Stroke Primary Prevention Trial, DBP = diastolic blood pressure, ECG = electrocardiogram, MC = Minnesota code, SBP = systolic blood pressure, ST-T = ST-segment and T wave.\n\nHB and HC contributed equally to this work.\n\nThe CSPPT study was jointly supported by Shenzhen AUSA Pharmed Co Ltd and national, municipal, and private funding bodies, including the National Science and Technology Major Projects Specialized for \"Major New Drugs Innovation and Development\" during the 12th Five-Year Plan Period: China Stroke Primary Prevention Trial (grant zx09101105); the Major State Basic Research Development Program of China (973 program) (2012 CB517703); Clinical Center (grant zx09401013); Projects of National Natural Science Foundation of China (grants 81473052, 81441091, and 81402735); National Clinical Research Center for Kidney Disease, Nanfang Hospital, Nanfang Medical University, Guangzhou, China; State Key Laboratory for Organ Failure Research, Nanfang Hospital, Nanfang Medical University, Guangzhou, China; the Special Project on the Integration of Industry, Education and Research of Guangdong Province (2011A091000031); the Science and Technology Planning Project of Guangdong Province, China (grant no. 2014B090904040); the Science, Technology and Innovation Committee of Shenzhen (JCYL20130401162636527); and research grants from the Department of Development and Reform, Shenzhen Municipal Government (grant SFG 20201744).\n\nThe authors have no conflicts of interest to disclose.\n"}
+{"text": "Objective\n=========\n\nPhysical activity is an important contributor to health and well-being. People with more education consistently report greater participation in physical activity ([@R1]). It is also well established that neighborhood walkability and safety influence participation in physical activity ([@R2],[@R3]). However, the mediating role of the neighborhood environment, in particular walkability and safety, on the association between education level and physical activity level has not been quantified. We hypothesized that perceived neighborhood walkability and safety are important mediators of the relationship between education level and physical activity level. If true, this suggests that efforts to make environments more amenable to physical activity in neighborhoods characterized by low education levels may help to reduce disparities in physical activity ([@R4],[@R5]).\n\nMethods\n=======\n\nWe analyzed aggregated data collected in 2010 and 2012 from 2 modified Behavioral Risk Factor Surveillance System (BRFSS) surveys ([@R6]) in the 39 \"Communities Putting Prevention to Work\" (CPPW) communities ([@R7]) that included questions on perceived neighborhood walkability and safety in addition to the usual BRFSS questions on education and physical activity. CPPW was a 2-year initiative that funded 50 communities to implement policy, systems, and environmental interventions to reduce obesity and tobacco consumption ([@R7]). The median response rate based on CASRO (Council of American Survey Research Organizations \\[<http://www.casro.org/>\\]) was 55% in 2012 (information not available for all communities in 2010). The final analytic sample consisted of 104,084 adults aged 18 years or older after exclusion of those for whom data were missing. Education was classified into 4 levels (less than high school, high school, some college, and college graduate). Physical activity was dichotomized to meeting or exceeding the 2008 US physical activity guidelines (ie, doing at least 150 minutes per week of moderate-intensity, or 75 minutes per week of vigorous-intensity aerobic physical activity, or an equivalent combination of the two) versus not meeting those guidelines ([@R1]). Perceived neighborhood environment for walking was categorized as very pleasant, somewhat pleasant, not very pleasant, and not at all pleasant. Perceived neighborhood safety was categorized as extremely safe, quite safe, slightly safe, and not at all safe. We conducted bivariate analyses to examine the inter-relationship between education level (exposure) and physical activity level (response), education and the mediators (neighborhood walkability and safety), and the mediators and physical activity. Then, we used structural equation modeling (SEM) ([@R8]) ([Figure](#F1){ref-type=\"fig\"}) to select the best-fit model and to estimate the direct effect of exposure on response, indirect effect of exposure on response via mediator, and the total effect (sum of direct and indirect effect). We included the following covariates in the SEM model: age (5 age groups), sex (male vs female), race/ethnicity (non-Hispanic white vs any other race), geographic location (urban area/large city vs rural area/small city), and intervention focus of the community (obesity only or obesity and tobacco vs tobacco only). Lastly, we calculated the mediation proportion as the ratio (percentage) of indirect effect over total effect ([@R9]). Descriptive and bivariate analyses were conducted using SAS-callable \\[SUDAAN, version 9.3, Research Triangle Institute\\] SUDAAN to account for the complex sample design. SEM analysis accounting for complex sampling features was conducted using Mplus, version 6.0 (Muth\u00e9n and Muth\u00e9n).\n\n![Mediation models from education level to meeting physical activity guidelines showing the direct effects of education, the indirect (mediated) effects acting through perceived neighborhood safety and perceived neighborhood walkability, and the proportion of the overall effect due to mediation, Communities Putting Prevention to Work: Behavioral Risk Factor Surveillance System 2010 and 2012. Numbers in the figure are standardized regression coefficients in the structural equation models.\\\nA pair of flow charts shows the relationships between education level and meeting physical activity guidelines, taking into account the covariates of age, sex, race/ethnicity, geography, and community focus and the mediating effects of perceived neighborhood safety from crime and perceived neighborhood walkability. The influence of education level and each significant covariate on meeting physical activity guidelines and perceived neighborhood safety from crime and perceived neighborhood walkability are reported as standardized regression coefficients from the structural equation models. The effects of perceived neighborhood safety from crime and perceived neighborhood walkability on meeting physical activity guidelines are also reported as standardized regression coefficients from the structural equation models in the following tables.**Flow Chart A. Mediation Model From Education to Meeting Physical Activity Guidelines Showing the Direct Effects Of Education and the Indirect (Mediated) Effect Acting Through Perceived Neighborhood Safety^a^Covariates and Mediating EffectsStandardized Regression CoefficientsNeighborhood SafetyPhysical Activity GuidelinesAge\u22120.039**---**Sex0.0590.049Race0.1860.049Education0.1970.123Geography\u22120.0350.045Community focus0.029**---**Neighborhood safety**---**0.048Flow Chart B. Mediation Model From Education to Meeting Physical Activity Guidelines Showing the Direct Effects of Education Level and the Indirect (Mediated) Effect Acting Through Perceived Neighborhood Walkability^b^Covariates and Mediating EffectsNeighborhood WalkabilityPhysical Activity GuidelinesAge\u22120.0780.052Sex**---**0.046Race0.1290.118Education0.1640.044Geography**------**Community focus0.0390.091Neighborhood walkability**------SEM showed that level of education had significant direct and indirect effects on meeting physical activity guidelines. The indirect effects are partially mediated by both perceived neighborhood walkability and perceived neighborhood safety from crime. Neighborhood walkability mediates 11.3% of the total effect of education on meeting physical activity guidelines. Neighborhood safety mediates 6.8% of the total effect of education on meeting physical activity guidelines.Abbreviation: ---, not applicable.^a^ Effects of education level with mediation by perceived neighborhood safety from crime: direct effect = 0.123; indirect effect = 0.197 \u00d7 0.048 = 0.009; total effect = 0.123 + 0.009 = 0.132; mediation proportion = 0.009 \u00f7 0.132 \u00d7 100 = 6.8%.^b^ Effects of education with mediation by perceived neighborhood walkability: direct effect = 0.118; indirect effect = 0.164 \u00d7 0.091 = 0.015; total effect t =0.118 + 0.015 = 0.133; mediation proportion = 0.015 \u00f7 0.133 \u00d7 100 = 11.3%.](PCD-12-E46s01){#F1}\n\nResults\n=======\n\nThe weighted overall sample was 51.8% female (95% confidence interval \\[CI\\], 50.6%--52.9%), and 53.2% (95% CI, 52.0%--54.3%) met or exceeded the 2008 US guidelines for weekly physical activity. Among respondents 16.9% (95% CI, 15.9%--18.0%) reported less than high school education, 24.2% (95% CI, 23.3%--25.2%) high school, 28.5% (95% CI, 27.5%--29.6%) some college, and 30.3% (95% CI, 29.4%--31.3%) were college graduates. Walking in the neighborhood was rated as very pleasant by 54.7% (95% CI, 53.5%--55.9%), somewhat pleasant by 35.7% (95% CI, 34.6%--36.8%), not very pleasant by 6.8% (95% CI, 6.1%--7.5%), and not at all pleasant by 2.8% (95% CI, 2.5%--3.2%) of respondents. Neighborhood safety from crime was perceived as extremely safe by 20.3% (95% CI, 19.4%--21.1%), quite safe by 50.9% (95% CI, 49.7%--52.0%), slightly safe by 24.0% (95% CI, 23.0%--25.1%), and not at all safe by 4.9% (95% CI, 4.4%--5.4%) of respondents. Bivariate analyses showed that with each increasing level of education the proportion of respondents reporting they met physical activity guidelines increased, and perceived neighborhood walkability and safety was greater ([Table](#T1){ref-type=\"table\"}). The 2 perceived environmental measures were also associated with level of physical activity. SEM showed that level of education had significant direct and indirect effects on meeting physical activity guidelines ([Table](#T1){ref-type=\"table\"}) (Figure). The indirect effects are partially mediated by both perceived neighborhood walkability and perceived neighborhood safety from crime. Neighborhood walkability mediates 11.3% of the total effect of education on meeting physical activity guidelines. Neighborhood safety mediates 6.8% of the total effect of education on meeting physical activity guidelines.\n\n###### Bivariate Relationships[a](#T1FN1){ref-type=\"table-fn\"} and Estimates of Effects Among Education Level, Perceived Neighborhood Environment, and Meeting Physical Activity Guidelines, Communities Putting Prevention to Work Behavioral Risk Factor Surveillance System, 2010 and 2012\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Characteristics                                                            Neighborhood Safe from Crime[b](#T1FN2){ref-type=\"table-fn\"}   Rate Walking in Neighborhood[b](#T1FN2){ref-type=\"table-fn\"}   Met or Exceeded Physical Activity\\                                            \n                                                                                                                                                                                                           Guidelines[c](#T1FN3){ref-type=\"table-fn\"}                                    \n  -------------------------------------------------------------------------- -------------------------------------------------------------- -------------------------------------------------------------- -------------------------------------------- ----- ------ ------ ------ ----- ------\n  **Education level**                                                                                                                                                                                                                                                                    \n\n  Less than high school                                                      13.6                                                           35.6                                                           41.4                                         9.4   41.9   40.8   12.2   5.2   42.0\n\n  High school                                                                17.7                                                           50.8                                                           25.8                                         5.7   50.8   39.5   6.4    3.2   50.8\n\n  Some college                                                               19.3                                                           53.0                                                           23.1                                         4.6   54.5   36.3   6.6    2.6   54.8\n\n  College graduate                                                           26.9                                                           57.5                                                           13.7                                         1.8   65.2   29.2   4.2    1.5   59.7\n\n  **Neighborhood safe from crime**                                                                                                                                                                                                                                                       \n\n  Extremely safe                                                             ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   57.3\n\n  Quite safe                                                                 ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   54.2\n\n  Slightly safe                                                              ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   49.1\n\n  Not at all safe                                                            ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   44.6\n\n  **Walking in neighborhood**                                                                                                                                                                                                                                                            \n\n  Very pleasant                                                              ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   56.9\n\n  Somewhat pleasant                                                          ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   49.7\n\n  Not very pleasant                                                          ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   45.5\n\n  Not at all pleasant                                                        ---                                                            ---                                                            ---                                          ---   ---    ---    ---    ---   43.7\n\n  **Estimate of effects**                                                                                                                                                                                                                                                                \n\n  Education \u2192 outcome, total[d](#T1FN4){ref-type=\"table-fn\"}                 0.132 (*P* \\< .001)                                            0.133 (*P* \\< .001)                                            ---                                                                           \n\n  Education \u2192 outcome, direct[d](#T1FN4){ref-type=\"table-fn\"}                0.123 (*P* \\< .001)                                            0.118 (*P* \\< .001)                                            ---                                                                           \n\n  Education \u2192 mediator \u2192 outcome, indirect[d](#T1FN4){ref-type=\"table-fn\"}   0.009 (*P* = .004)                                             0.015 (*P* \\< .001)                                            ---                                                                           \n\n  Mediation proportion                                                       6.80                                                           11.30                                                          ---                                                                           \n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAbbreviation: ---, not applicable.\n\nThe Cochran-Mantel-Haenszel trend test was used in the bivariate analysis. All *P* values are less than .001.\n\nValues are percentages unless otherwise indicated.\n\nMet or exceeded the US 2008 guidelines for weekly physical activity ([@R1]).\n\nValues are standardized regression coefficients.\n\nDiscussion\n==========\n\nOur analysis of the 2010 and 2012 CPPW BRFSS demonstrates that a modest proportion of the important association between education level and meeting physical activity guidelines is mediated by perceived neighborhood walkability and safety. Although the mediating effects are relatively small, these results suggest that interventions focused on enhancing walkability and safety in communities that face challenges in these areas might cut into the consistently observed disparity in meeting physical activity guidelines associated with education level. Given the difficulty of directly intervening on the underlying socioeconomic determinants of health (eg, income, poverty, employment, education), public health strategies focused on improving the environment so that healthy options become more feasible appear promising ([@R4],[@R10]). Improving the environment has been the basis for a series of programs supported by the Centers for Disease Control and Prevention (CPPW, Community Transformation Grants, Racial and Ethnic Approaches to Community Health) that provided funds to states and communities for sustainable policy, system, and environmental interventions ([@R7]).\n\nOur results are limited by the cross sectional design of the CPPW BRFSS. We cannot infer causality between exposures and outcomes. Although the overall sample was large and similar in key characteristics (including physical activity level \\[[@R11]\\]) to national samples, our sample was drawn from only 39 communities and is not nationally representative. A strength of the BRFSS and these analyses is the information available on a large number of potential confounders and the ability to control for them in the analyses. Physical activity, education, walkability, and safety were all assessed by self-report and are thus subject to recall and social desirability bias. It is possible that with an objective measure of physical activity, such as by accelerometer and objective measures of the environment, such as GIS-based walkability indices and actual crime reports, the mediating effects of neighborhood walkability and safety on the education--physical activity relationship might be better elucidated. Studies of this type should become a priority. However, despite these limitations our results do suggest that part of the observed association between education level and meeting physical activity guidelines is mediated by the neighborhood environment, and that interventions designed to improve community environments have the potential to close equity-based gaps in physical activity.\n\nThe findings and conclusions of this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. The authors declare no conflicts of interest and have not received funding to support this work. The authors acknowledge valuable input and support from staff and leadership of the National Center for Chronic Disease Prevention and Health Promotion, and Divisions of Nutrition, Physical Activity and Obesity, and Community Health. Special thanks to Ursula Bauer, Dana Shelton, Leonard Jack, Nicole Flowers, Joan Dorn, Thomas Schmid, and Janet Fulton.\n\nThe opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors\\' affiliated institutions.\n\n*Suggested citation for this article:* Pratt M, Yin S, Soler R, Njai R, Siegel PZ, Liao Y. Does Perceived Neighborhood Walkability and Safety Mediate the Association Between Education and Meeting Physical Activity Guidelines? Prev Chronic Dis 2015;12:140570. DOI: <http://dx.doi.org/10.5888/pcd12.140570>.\n"}
+{"text": "Introduction {#s1}\n============\n\nIn plants, sugars are the key carbon source for energy production and macromolecule biosynthesis, and they are important for carbohydrate accumulation in crop organs consumed as food by humans ([@CIT0007]). Thus, sugar allocation between source leaves (net exporters) and sink organs (net importers), such as young leaves, flowers, and storage organs, is of primary importance for organ development and plant growth, and as a sustainable food supply ([@CIT0001]; [@CIT0055]). For long-distance allocation, sugars are primarily synthesized by photosynthetic carbon fixation in source leaves and are loaded, mostly as sucrose, into phloem-specific companion cell (CC)--sieve element (SE) complexes. This loading occurs either symplasmically via plasmodesmata or via the apoplasmic route, with fine-tuned regulation of sugar partitioning ([@CIT0078]; [@CIT0053]; [@CIT0086]). Regarding the apoplasmic route, plasma membrane-localized sugar transporters facilitate sucrose efflux and import between adjunct CCs and parenchyma cells of minor veins, promoting mass flow in the phloem for continuous sugar allocation ([@CIT0059]; [@CIT0013]; [@CIT0034]).\n\nDuring sucrose loading, SWEET ([s]{.ul}ugar [w]{.ul}ill [e]{.ul}ventually be [e]{.ul}xported [t]{.ul}ransporters) proteins, especially clade III SWEET members localized in the plasma membrane of phloem parenchyma cells, are probably the main efflux mechanism to facilitate sucrose export into the apoplast in both dicot ([@CIT0016]; [@CIT0020]) and monocot plants ([@CIT0005]). An absence of SWEET exporters significantly suppresses sucrose loading and plant growth, with accumulation of sugars and starch in source leaves ([@CIT0016]; [@CIT0040]; [@CIT0005]). The final loading step is mediated by SUT1/SUC2-type ([su]{.ul}crose [t]{.ul}ransporter; H^+^/[suc]{.ul}rose symporter) in the plasma membrane of phloem cells, that actively imports apoplasmic sucrose into CC--SE complexes against a concentration gradient ([@CIT0059]; [@CIT0034]). Plant species with decreased SUT1 uptake ability were stunted, due to lack of sugar loading from source leaves ([@CIT0025]; [@CIT0069]; [@CIT0067]).\n\nSimilarly, sucrose unloading from phloem complexes to sink organs can be symplasmic and apoplasmic depending on organ types and developmental stages ([@CIT0050]; [@CIT0002]; [@CIT0034]). The initial phloem unloading includes sucrose fluxes across SE, CC, and adjacent phloem parenchyma cells. Along the post SE--CC unloading pathway, a fraction of the sucrose could be unloaded from the CCs to the apoplast, probably by simple or facilitated diffusion ([@CIT0050]; [@CIT0034]). The apoplasmic sucrose can be either directly imported into nearby parenchyma cells by a sucrose transporter, or hydrolyzed by a cell wall invertase into hexoses that are subsequently imported by monosaccharide transporters ([@CIT0002]; [@CIT0048]; [@CIT0034]). Although many sugar transporter families have been characterized in Arabidopsis and other crops, transport mechanisms involved in initial sugar unloading are largely unknown. Some SUT-type sucrose transporters were implicated in sucrose export from CCs ([@CIT0006]). Monosaccharide transporters (MSTs) ([@CIT0063]), such as STP ([s]{.ul}ugar [t]{.ul}ransport [p]{.ul}rotein) transporters ([@CIT0064], [@CIT0063]), may be involved in hexose import; however, experimental evidence for their physiological roles is lacking. SWEET transporters not only mediate sucrose efflux for phloem loading, but are also major sugar unloaders mediating sugar efflux, as well as import, from maternal connecting tissues to reproductive sinks ([@CIT0041]; [@CIT0015]; [@CIT0085]). SWEETs have been characterized as low-affinity hexose and sucrose bidirectional transporters (exporters or importers), depending on physiological demands. *At*SWEET9 specifically localizes in the plasma membrane of inflorescence parenchyma cells to facilitate sucrose efflux for nectar formation ([@CIT0041]). Clade III SWEETs expressed in maternal cells of developing seeds catalyzed sucrose efflux and import, supporting embryo development in Arabidopsis ([@CIT0015]) and rice ([@CIT0085]). Moreover, the hexose export activity of *Zm*SWEET4c in corn and *Os*SWEET4, its ortholog in rice, is of primary importance for seed filling and appears to be linked to seed size ([@CIT0068]). These studies highlight the possibility that SWEETs could have a fundamental role in sugar unloading in sink organs as early as the seedling development stage ([@CIT0020]).\n\nDuring vegetative stages, young leaves and roots are major sink organs ([@CIT0079]), with up to 60% of newly synthesized sucrose exported to sink leaves. Unlike simple rosette leaves in Arabidopsis, tomato (*Solanum lycopersicum*) has sequentially developed compound leaves, with multiple leaflets that require an extended morphogenetic phase ([@CIT0065]), suggesting that a controlled, fine-gated sugar unloading mechanism is needed through the phloem to developing leaves. Tomato is an important vegetable crop worldwide (United Nations Food and Agriculture Organization; <http://www.fao.org/>), with many efforts focused on sugar allocation to improve fruit size, sweetness, and quality ([@CIT0057]; [@CIT0023]). However, vigorous vegetative growth is the primary determinant of tomato biomass and fruit yield ([@CIT0077]). Thus, adequate leaf production could increase flower numbers and fruit weight up to 60% ([@CIT0032]; [@CIT0077]). However, transport mechanisms regulating initial sugar unloading to sink leaves are largely unknown in tomato. Early studies using either radiotracers, a membrane-impermeable fluorescent carboxyfluorescein, or green fluorescent protein (GFP) fusion proteins have demonstrated that initial phloem unloading in sink leaves is largely via a symplasmic pathway ([@CIT0060]; [@CIT0074]; [@CIT0054]). However, these data do not exclude an apoplasmic component. For instance, developing maize leaves have been found to switch dynamically from symplasmic to apoplasmic phloem unloading ([@CIT0003]). These results imply that putative, as yet unidentified, membrane carriers which reside in unloading cells may participate in the apoplasmic sugar unloading into tomato sink leaves.\n\nTo date, only two sugar transporters, *Le*SUT1 and *Le*SUT2 sucrose transporters, have been implicated in source to sink sugar distribution during vegetative growth of tomato plants ([@CIT0004]; [@CIT0081]; [@CIT0048]). *Le*SUT1 proteins are localized in the SEs and appear to have a major role in sucrose loading in source leaves ([@CIT0004]; [@CIT0081]), as impaired *LeSUT1* expression increased accumulation of soluble sugars and starch in mature leaves ([@CIT0029]). Although SE-localized *Le*SUT2 has been implicated in sucrose unloading to sink leaves, *LeSUT2* antisense mutants reduced fruit fertility, with no apparent changes in sugar concentrations in leaves or vegetative tissues ([@CIT0029]). Therefore, *Le*SUT2 is probably not the major sucrose unloader required for sugar unloading in young leaves. Based on the insensitivity of sucrose unloading to a metabolic inhibitor ([@CIT0060]), and a steep sucrose concentration gradient between SE--CC complexes and the apoplast in most sink organs performing apoplasmic unloading ([@CIT0050]), a passive sucrose facilitator, such as SWEET sucrose carriers, has been suggested to mediate the efflux ([@CIT0048]; [@CIT0034]). Yet, the genetic evidence is still missing.\n\nOn the other hand, a hexose facilitator may also be required to act in series to re-load apoplasmically hydrolyzed sucrose into the downstream phloem parenchyma cells ([@CIT0050]). For instance, radiotracer analysis indicated that glucose can be retrieved from the apoplasm of sink leaves by phloem parenchyma cells ([@CIT0060]). Thus, a glucose facilitator (e.g. SWEETs) on the plasma membrane of parenchyma cells could be a possible candidate to perform this function driven by the relatively low glucose concentrations in the cytoplasm, compared with the apoplasm ([@CIT0066]; [@CIT0056]).\n\nWe deduced that the sink leaf-expressed *Sl*SWEET transporter could be a promising sugar unloader for vegetative growth. Although the complete SWEET gene family has been identified in tomato 'Heinz 1706' and its expression pattern characterized by RNA sequencing (RNA-seq) ([@CIT0021]; [@CIT0062]), the role of SWEET proteins in phloem unloading during the vegetative stage in tomato is unknown. The objective in this study was to characterize transport properties of the most highly expressed *Sl*SWEET1a (Solyc04g064610) in young developing leaves of tomato plants. Dominant expression of *Sl*SWEET1a--\u03b2-glucuronidase (GUS) fusion proteins in unloading veins and high correlation between *SlSWEET1a* transcripts and carbon importing activity in sink leaves implicated that it had a specific role in sugar unloading. Furthermore, reduced hexose accumulation in young leaves consistently occurred in *SlSWEET1a*-silenced tomato plants. Thus, a putative role for *Sl*SWEET1a in apoplasmic sugar unloading to sink leaves is discussed.\n\nMaterials and methods {#s2}\n=====================\n\nPlant materials and growth conditions {#s3}\n-------------------------------------\n\nThe Micro-tom cultivar of tomato (*Solanum lycopersicum*) was used, except where indicated. Tomato seeds were surface-sterilized for 8 min using diluted bleach solution (30% CLOROX bleach and 0.1% Triton X-100, v/v) and then washed three times with sterilized water. To propagate tomato plants, sterilized seeds were directly germinated in water at room temperature for 2 d and then transferred to a hydroponic system or pots (diameter, 11.5 cm) containing a mixture of soil and peat moss (1:1). All plants were grown in a controlled chamber (25 \u00b0C and 16/8 h light/dark regime with \\~100 \u03bcmol m^\u22122^ s^\u22121^ illumination) before harvest or experiments. To collect various organs (except fruits), tomato plants were grown hydroponically for 3 weeks (for vegetative organs) to 5 weeks (for reproductive organs), then roots, stems, mature leaves (the terminal leaflet \\~4 cm), young leaves (the terminal leaflet \\<2 cm), flower buds (developing green buds), and flowers (first day of anthesis) were collected ([Supplementary Fig. S1](#sup1){ref-type=\"supplementary-material\"} at *JXB* online). For leaves, all leaflets from one leaf were pooled together as one sample. Various stages of fruits were collected from soil-grown plants after 14, 21, 35, or 42 d of flowering. To analyze expression in different regions of leaves, the terminal leaflets of young leaves that were 80% of the final size (\\~3--3.5 cm) were harvested from 5-week-old plants. Then, segments of 1 cm in length from the leaf base to tip were separated for mRNA analysis. For mesophyll protoplast isolation, *Arabidopsis thaliana* ecotype Col-0 plants were grown in peat moss, vermiculite, and perlite (6:1:1) under short-day conditions (22 \u00b0C, 10/14 h light/dark regime, \\~80 \u03bcmol m^\u22122^ s^\u22121^ illumination) for 4--5 weeks before isolation.\n\nAnalysis of mRNA transcripts {#s4}\n----------------------------\n\nTotal RNA was isolated from tomato organs with TRIsure reagent (Bioline, <http://www.bioline.com/>) with an RNA purification column (GeneMark, <http://www.genemarkbio.com/>), whereas RNA samples from fruits were prepared using the cetyltrimethylammonium bromide (CTAB) method ([@CIT0049]). Total RNA transcripts were reverse-transcribed, and gene-specific primers for all 31 *SlSWEET* genes (primer sequences are listed in [Supplementary Table S1](#sup1){ref-type=\"supplementary-material\"}) were used for real-time quantitative PCR (qRT-PCR) as described ([@CIT0011]). Expression of the reference gene *SlActin7* (Solyc11g005330) was used to determine relative expression levels according to the following equation: 1000\u00d72^\u2212^(^Ct^~*SlSWEET*~^\u2212Ct^~*SlActin7*~) (Ct=threshold cycle; [@CIT0011]).\n\nLocalization of GUS fusion proteins {#s5}\n-----------------------------------\n\nTo generate *Sl*SWEET1a--GUS fusion proteins for observation, the entire DNA fragment of *SlSWEET1a* (1467 bp), including all introns and deletion of the stop codon, was amplified from genomic DNA prepared from Micro-tom leaves with specific primers (forward, 5\\'-CCGCGGATGGGTGTTGTTCATACTCTG-3\\'; and reverse, 5\\'-CTGCAGTCCAGCGCCAACTT GCTCAAGCCTTGACTT-3\\'). The resulting fragments were first cloned into a pGM-T vector using AT cloning technology and then cloned into pUTKan (including the GUS fragment) binary vector by *Sac*II and *Pst*I restriction sites (pUTKan-gSlSWEET1a). Subsequently, the native promoter of *SlSWEET1a* (2000 bp) was amplified with specific primers (forward, 5\\'-TTAAACCTGATCGTCACATAAA-3\\'; and reverse, 5\\'-CAGAGTATGAACAACACCCATGTC-3\\') and cloned into pGM-T, and then cloned into a pUTKan-gSlSWEET1a binary vector, using *Sac*I and *Kpn*I restriction sites. Then, Micro-tom plants were transformed with the resulting pUTKan-PSlSWET1a::gSlSWEET1a--GUS construct in the Academia Sinica Transgenic Plant Core Lab (<http://transplant.sinica.edu.tw/en/aboutus/intro/index3.htm>). In total, five T~1~ transgenic tomato plants were obtained. Due to limited numbers of T~1~ transgenic tomato plant leaves, three young leaves were collected from three independent T~1~ transgenic lines and three mature leaves were collected from other independent lines. Flowers were collected from T~2~ lines. All samples were examined by histochemical staining, 16 h for young leaves and mature leaves, and 9 h for flowers, as described ([@CIT0028]). Stained organs were imaged with a LEICA Z16 APO microscope.\n\nTranslocation of carboxyfluorescein {#s6}\n-----------------------------------\n\nTo explore the phloem unloading fluxes in tomato sink leaves, the abaxial surface of the terminal leaflet of mature source leaves from 2-week-old tomato plants was gently abraded with fine sandpaper. A drop of 200 \u00b5l of CFDA \\[5(6)-carboxyfluorescein diacetate, Sigma-Aldrich, USA\\] solution (300 \u00b5g ml^\u22121^) was applied to each leaf, which was then covered with plastic wrap to prevent evaporation. After 3 h, fluorescence in the terminal leaflets (\\<2 cm) of attached sink leaves on intact plants was observed and imaged using a Leica fluorescence microscope equipped with a SOLA light engine (Lumencor, USA) and a Nikon EOS700d digital camera. Images were taken under a GFP filter set and exported by the EOS utility software.\n\nSubcellular localization of GFP fusion proteins {#s7}\n-----------------------------------------------\n\nTo generate *Sl*SWEET1a--GFP fusion proteins, the cDNA fragment of *SlSWEET1a* (741 bp) without the stop codon was amplified from the pDRf1-GW-SlSWEET1a plasmid (see 'Expression of clade I *Sl*SWEET1a in yeast') with specific primers (forward, 5\\'-CACCATGGGTGTTGTTCATACT-3\\'; and reverse, 5\\'-GGCGCCAACTTGCTCAAGCCTTGACTTGC-3\\') and cloned into a pENTR-D-TOPO vector using TOPO cloning technology (Invitrogen, USA). Subsequently, the cDNA was transferred into pUBN-GFP and pUBC-GFP via an LR reaction using Gateway technology ([@CIT0026]). Arabidopsis and tobacco protoplasts were then transformed with the resulting pUBN-SlSWEET1a and pUBC-SlSWEET1a constructs to express N- and C-terminal translational GFP fusion proteins, respectively.\n\nArabidopsis mesophyll protoplasts were isolated as described ([@CIT0083]). A portion of the protoplast preparation (200 \u00b5l at a cell density of 2\u00d710^5^ ml^\u22121^) was mixed with 10--20 \u03bcg of plasmids and 220 \u03bcl of polyethylene glycol (PEG) solution prior to incubation for 15 min, at 25 \u00b0C in the dark. Protoplasts were then washed and re-suspended in 1 ml of W5 solution (154 mM NaCl, 125 mM CaCl~2~, 5 mM KCl, and 2 mM MES, pH 5.7) and incubated for 24--48 h in the dark before observation. To determine the position of the plasma membrane, protoplasts were either co-transformed with the membrane marker pRT101-AtPIP2A--red fluorescent protein (RFP) in a 1:1 ratio ([@CIT0046]) or stained with 0.5 \u03bcl of 10 \u03bcM FM4-64 dye just before observation ([@CIT0011]). Fluorescence was observed on a Carl Zeiss LSM780 confocal microscope (Instrument Development Center, National Cheng Kung University) as described ([@CIT0011]).\n\nExpression of clade I *Sl*SWEET in yeast {#s8}\n----------------------------------------\n\nTo generate expression constructs for yeast, cDNA sequences of clade I *SlSWEET* were amplified using Phusion polymerase (New England Biolabs, USA) with gene-specific primers (primer sequences are given in [Supplementary Table S2](#sup1){ref-type=\"supplementary-material\"}) from leaf cDNA derived from tomato cultivar 'TS19'. The cDNA sequences were confirmed by sequencing and had 100% identity with those in the Sol Genomics Network (<https://solgenomics.net/>). Then, cDNA was first cloned into the pDONR221-f1 vector and subsequently transferred to the pDRf1-GW vector using Gateway technology. The yeast strains YSL2-1 ([@CIT0011]) and EBY4000 were then transformed with the resulting pDRf1-GW-SlSWEET plasmids using the lithium acetate (LiAC) method ([@CIT0024]). Transformants were selected on synthetic deficient medium without uracil (SDM-U medium, containing 1.7 g of yeast nitrogen base without amino acids, 5 g of ammonium sulfate, 2% maltose, 2% agar, and 0.01% histidine, leucine, and tryptophan in 1 liter). The pDRf1 empty vector with a Gateway cassette deleted and a plasmid pDRf1-Hxt5 that harbored a yeast hexose transporter were also transformed as a negative and positive control, respectively, of transport activity ([@CIT0011]).\n\nYeast growth assay {#s9}\n------------------\n\nYeast cells from single colonies were grown at 30 \u00b0C overnight and then refreshed with new SDM-U medium until they reached early mid-log phase growth (OD~600~ of 0.8--1.0). Cells were collected and diluted with SD-U medium (SDM-U medium without 2% maltose) to an OD of 0.2. Then, serial dilutions (1, 10^--1^, 100^--1^, and 1000^--1^) of all desired cells were prepared and 5 \u03bcl from each dilution was spotted on SDM-U medium or SD-U medium supplemented with 2% glucose, galactose, mannose, fructose, or sucrose. To maintain the media at pH 5 or 7, Tris-MES buffers (6.054 g of Tris and 9.76 g of MES in 1 liter) were included. Plates were incubated at 30 \u00b0C for 3--6 d and then scanned.\n\nRadiotracer uptake assay in yeast {#s10}\n---------------------------------\n\nOne portion of overnight yeast culture was diluted with 50 ml of fresh SDM-U medium to an OD~600~ of 0.2. After incubation for 4 h at 30 \u00b0C, yeast cells (\\~0.5 OD~600~) were collected, washed, and re-suspended in 50 mM sodium phosphate buffer (112 mg of Na~2~HPO~4~ and 6.8 g of NaH~2~PO~4~ in 1 liter, pH 5) to an OD~600~ of 5 before assay. For a time-dependent uptake assay, uptake buffer containing 2 mM glucose and 1.5 \u03bcCi of \\[^14^C\\]glucose ml^\u22121^ in 50 mM sodium phosphate buffer (pH 5) was prepared. For a concentration-dependent uptake assay, uptake buffers with 11iter glucose concentrations (0.2, 1, 2, 5, 10, 20, 40, 100, 200, and 500 mM glucose in 50 mM sodium phosphate buffer, pH 5) with same molar ratio of \u03bcCi of \\[^14^C\\]glucose were prepared. For competition assays, uptake buffers (2 mM glucose with 2 \u03bcCi of \\[^14^C\\]glucose ml^\u22121^) without other sugars (control) or supplemented with 10-fold concentrations of various cold sugars (20 mM glucose, galactose, fructose, mannose, sucrose, and maltose, respectively) were prepared. For analysis under various pH values, uptake buffer medium with various pH values were prepared using 50 mM sodium phosphate containing 2 mM glucose with 2 \u03bcCi of \\[^14^C\\]glucose ml^\u22121^. Before the uptake assay, cells were washed once with corresponding cold pH buffers. For treatment with the protonophore NH~4~Cl, yeast cells were pre-treated with 10 mM NH~4~Cl for 10 min at 30 \u00b0C, then uptake buffer of pH 5 (2 mM glucose with 2 \u03bcCi of \\[^14^C\\]glucose ml^\u22121^) was added. To start the uptake assay, 110 \u03bcl of corresponding uptake buffers were always added into an equal volume of cells that was incubated at 30 \u00b0C. At the indicated time points in the time-dependent assay or at 10 min in all other assays, cells were collected on filter paper (MCE Membrane Filter, 0.45 \u03bcm) and washed three times with ice-cold 50 mM sodium phosphate buffer by vacuum filtration. Filter paper (with cells) was transferred into scintillation vials containing 4 ml of Rotiszint^\u00ae^ eco plus (Carl Roth, Germany) and incubated at room temperature for 16--18 h. Radioactivity of lysed cells was quantified in a Tri-Carb 4810TR scintillation counter (Perkin Elmer, USA) and a kinetic curve determined using single rectangular regression (Sigmaplot Version 13).\n\nInduction of VIGS (virus-induced gene silencing) in tomato leaves {#s11}\n-----------------------------------------------------------------\n\nVIGS was carried out as described ([@CIT0044]). In general, the 300 bp fragment of *SlSWEET1a* for gene silencing was selected via the VIGS tool on the Sol Genomics Network (<https://solgenomics.net/>) and amplified from the pDONR221-f1-SlSWEET1a plasmid using Phusion polymerase with specific primers (forward, 5\\'-CACCATGGGTGTTGTTCATACT-3\\'; reverse, 5\\'-CTTCTTCTCTTTAGTTGG-3\\'). The resulting fragment was first cloned into pENTR-D-TOPO (Invitrogen, USA) via TOPO cloning technology, then transferred into pTRV2 via an LR reaction ([@CIT0044]). Then, *Agrobacterium tumefaciens* GV3101 was transformed with the resulting pTRV2-SlSWEET1a construct via the freeze--thaw method ([@CIT0030]). Positive transformants were selected on LB medium (5 g of yeast extract, 10 g of tryptone, 10 g of NaCl in 1 liter) supplemented with kanamycin (50 mg l^\u22121^) and gentamycin (50 mg l^\u22121^). A portion of overnight culture (2.5 ml) derived from an *Agrobacterium* transformant was added to 25 ml of LB fresh medium containing antibiotic, 10 mM MES, and 200 \u03bcM acetosyringone (Sigma-Aldrich, USA) and grown overnight at 30 \u00b0C. Cells were collected by centrifugation at 4000 *g* for 5 min and re-suspended to an OD~600~ of 1.5 with infiltration medium (10 mM MgCl~2~, 10 mM MES, and 200 \u03bcM acetosyringone). An aliquot (1 ml of culture) was injected into cotyledons of 2-week-old tomato plants with a 1 ml syringe. To estimate silencing efficiency, *Agrobacterium* cells carrying the silencing construct, pTRV2-SlPDS-VIGS that would result in reduced expression of the phytoene desaturase gene so as to inhibit chlorophyll biosynthesis and cause etiolation ([@CIT0043]), was also injected into other plants.\n\nAnalysis of sugar contents {#s12}\n--------------------------\n\nYoung leaves (the terminal leaflet \\<2 cm, \\<50% of the final size) and mature leaves (the terminal leaflet \\~4 cm), including all three leaflets, from VIGS-silenced tomato plants were harvested 2 weeks after injection and stored at \u221270 \u00b0C before extraction. Frozen leaves (0.2--0.5 g) were ground into fine powder and 480 \u03bcl of extraction buffer medium \\[containing 460 \u03bcl of methanol (HPLC grade) and 20 \u03bcl of 0.2 mg ml^\u22121^ ribitol\\] was added. Mixtures were incubated at 70 \u00b0C for 10 min and centrifuged at 14 000 *g* for 10 min at room temperature. The supernatant was transferred into a glass vial, with 250 \u03bcl of cold chloroform (HPLC grade) and 500 \u03bcl of cold MQ water added in the given order. Mixtures were vortexed for 15 s and then centrifuged at 4000 *g* for 15 min at 20 \u00b0C. An aliquot of 200 \u00b5l of supernatant was transferred into a new 1.5 ml Eppendorf tube and freeze-dried in an LyoLab 5 LT freeze dryer (LSI, USA). Dry samples were derivatized with 20 \u03bcl of methoxyamination solution (0.02 g of methoxyamine hydrochloride in 1 ml of pyridine) at 30 \u00b0C for 90 min and then the mixture was added to 100 \u03bcl of 99:1 BSTFA+TMCS solution (Macherey-Nagel, Germany) and incubated at 70 \u00b0C for another 120 min. All extracts were then transferred into 250 \u03bcl pulled conical glass insert vials (Agilent, Germany) and analyzed with an Agilent 7890A/5975C GC-MS System with a 60 m\u00d70.250 mm column (Agilent, Germany), as described ([@CIT0058]; [@CIT0075]). Sugars were quantified relative to the internal ribitol standard, using the following retention times: glucose, 23.286 min and 23.521 min; fructose, 23.001 min and 23.093 min; galactose, 23.235 min and 23.487 min; mannose, 23.168 min and 23.378 min; and sucrose, 30.869 min.\n\nResults {#s13}\n=======\n\n*Sl*SWEET1a was highly expressed in aerial sink organs {#s14}\n------------------------------------------------------\n\nThe genome of tomato cultivar 'Heinz 1706' was sequenced in 2012 ([@CIT0072]) and 31 tomato *SlSWEET* genes have been identified ([@CIT0021]; [@CIT0062]). However, these genes were not well annotated to the SWEET family, which makes it difficult to compare orthologous genes. Therefore, the phylogenic tree of all 31 *SlSWEET* genes was generated and classified into four clades ([Supplementary Fig. S2](#sup1){ref-type=\"supplementary-material\"}). To identify the key *SlSWEET* member responsible for sugar unloading into sink tomato leaves, qRT-PCR was performed for all 31 *SlSWEET* genes using cDNA isolated from whole new developing leaves (the terminal leaflet was smaller than 2 cm) of 4-week-old hydroponic-grown tomato plants. Based on relative expression levels in young tomato leaves, transcript levels of *SlSWEET1a* (Solyc04g064610) were markedly higher compared with other genes ([Fig. 1A](#F1){ref-type=\"fig\"}). Interestingly, *SlSWEET1a* was also highly expressed in reproductive sink organs, such as flower buds and flowers ([Fig. 1B](#F1){ref-type=\"fig\"}), but only weakly expressed during the maturation stage of tomato fruits ([Fig. 1C](#F1){ref-type=\"fig\"}). Nevertheless, expression of *SlSWEET1a* was increased \\>10-fold during the maturation stage (35--42 d after flowering) of fruit development.\n\n![Expression of *SlSWEET* genes in tomato organs. (A) Stable mRNA transcripts of 31 *SlSWEET* genes in young leaves. (B) Developmental expression of *SlSWEET1a* in young leaves (YL), mature leaves (ML), roots (R), stems (S), flowers (F), and flower buds (FB) from 3- (vegetative organs) to 5-week- (reproductive organs) old hydroponically grown tomato plants. (C) Gene expression of *SlSWEET1a* in several stages of tomato fruits from soil-grown plants. Total RNA was isolated from various organs, and cDNA was used for qRT-PCR with gene-specific primers. Relative expression level by normalizing to an internal control, *SlActin7*, is shown. Results are means \u00b1SE from three independent biological repeats. DAF=days after flowering.](erz154f0001){#F1}\n\n*Sl*SWEET1a proteins are highly abundant in unloading tissues {#s15}\n-------------------------------------------------------------\n\nTo explore tissue-specific localization of *Sl*SWEET1a proteins, we generated transgenic tomato plants expressing a C-terminal translational GUS gene fusion with the complete genomic *SlSWEET1a* DNA sequence, containing all introns driven by the native *SlSWEET1a* promoter (*Sl*SWEET1a--GUS). After shoot regeneration and transfer to potting mix for 1.5 months, mature and new leaves from T~1~ tomato plants were stained to detect GUS activity. Predominant blue staining was observed in the first- to third-order veins in both young ([Fig. 2A, B](#F2){ref-type=\"fig\"}) and mature leaves ([Fig. 2C, D](#F2){ref-type=\"fig\"}). Dominant expression of the *Sl*SWEET1a--GUS fusion in leaf veins was also observed in the other four independent T~1~ transgenic tomato plants ([Supplementary Fig. S3A](#sup1){ref-type=\"supplementary-material\"}). The same localization to leaf veins was also observed in the three T~2~ transgenic plants, although with less staining (data not shown). Low expression of *Sl*SWEET1a--GUS in T~2~ plants may be the result of an unknown silencing effect, occurring in the tomato Micro-tom cultivar ([@CIT0037]). In flowers derived from T~2~ plants, distinct staining was obvious in stamens ([Fig. 2E, F](#F2){ref-type=\"fig\"}). To determine whether the expression of *Sl*SWEET1a--GUS fusion coincides with sugar unloading, we exposed a nearby mature terminal leaflet to CFDA that, on accumulation in the phloem, is de-esterfied into the membrane-impermeant and fluorescent carboxyfluorescein (CF), to trace carbohydrate unloading in developing leaves ([@CIT0054]). After 3 h of CF loading in the nearby mature leaves, fluorescence was clearly observed in the first- (midrib), second-, and third-order veins in a young leaflet (\\<2 cm) ([Fig. 2G, H](#F2){ref-type=\"fig\"}), but not in mature leaves ([Supplementary Fig. S3B](#sup1){ref-type=\"supplementary-material\"}). The overall fluorescence intensities exhibited a high to low gradient from leaflet base to tip, indicating that the highest sugar unloading activity was located in the base of the young leaflet. Interestingly, relatively higher mRNA levels of *SlSWEET1a* were measured in the leaflet base compared with the middle and tip that had reached 80% of their full size ([Fig. 2I](#F2){ref-type=\"fig\"}) and probably undergone sink--source transition ([@CIT0054]).\n\n![Tissue-specific expression of *Sl*SWEET1a proteins. (A--F) Histochemical staining of GUS activity in transgenic tomato plants expressing *Sl*SWEET1a--GUS fusion proteins driven by the *SlSWEET1a* native promoter. Independent transgenic plants were grown in soil for \\>4 weeks, and selected organs were harvested for staining for 16 h in (A, B) young leaves and (C, D) mature leaves, and 9 h for (E, F) flowers. Leaves and flowers were harvested from T~1~ and T~2~ plants, respectively. Representative terminal leaflets are shown in (A--D). Images in (B), (D), and (F) are magnified images of (A), (C), (E), respectively. (G and H) Fluorescence images of (H) CF dye unloaded into the terminal leaflet of a young leaf (G). The same pattern had been observed in three independent plants. (I)Transcript levels (by qRT-PCR) of *SlSWEET1a* in terminal leaflets of tomato young leaves. Relative expression levels by normalizing to an internal control, *SlActin7*, is shown. B, M, and T indicated samples harvested from the leaflet base, middle, and tip, as shown in (G). Results are means \u00b1SE from four independent biological repeats. Significant differences from expression in the leaflet base (B) were determined by Student's *t*-test and indicated by one (*P*\\<0.05) or two (*P*\\<0.01) asterisks. Scale bars are 1 mm in (B), (D), and (F); 2 mm in (A), (C), (E), (G), and (H).](erz154f0002){#F2}\n\nPlasma membrane-specific localization of *Sl*SWEET1a proteins {#s16}\n-------------------------------------------------------------\n\nTo observe the subcellular localization of *Sl*SWEET1a proteins at a single-cell resolution, the C-terminal *Sl*SWEET1a--GFP was transiently expressed in isolated Arabidopsis mesophyll protoplasts. Green fluorescence derived from *Sl*SWEET1a--GFP enclosed cytosolic chloroplasts indicated by red autofluorescence, and largely overlapped with red fluorescence derived from the plasma membrane dye, FM4-64 ([Fig. 3A--C](#F3){ref-type=\"fig\"}; [@CIT0080]). Furthermore, when Arabidopsis protoplasts were co-transformed with *Sl*SWEET1a--GFP and a plasma membrane marker, AtPIP2--RFP, based on path-analysis of fluorescence intensities, there was co-localization of green and red fluorescence ([Fig. 3D--H](#F3){ref-type=\"fig\"}). Similarly, plasma membrane-specific localization of *Sl*SWEET1--GFP proteins was present in isolated tobacco BY2 cells ([Supplementary Fig. S4A](#sup1){ref-type=\"supplementary-material\"}). However, N-terminal fusion of GFP to *Sl*SWEET1a resulted in mislocalization and aggregation of the fusion proteins in the cytosol ([Supplementary Fig. S4B](#sup1){ref-type=\"supplementary-material\"}), suggesting the localization signal was at the N-terminus of *Sl*SWEET1a.\n\n![*Sl*SWEET1a proteins co-localized with plasma membrane proteins. (A--C) The *Sl*SWEET1a--GFP fusion proteins were transiently expressed in Arabidopsis mesophyll protoplasts driven by the 35S promoter. Fluorescence derived from (A) *Sl*SWEET1a--GFP fusion proteins, (B) FM4-64 staining, and (C) merged signals from the same focal plane are shown. (D--F) Fluorescence images of mesophyll protoplast expressing both *Sl*SWEET1a--GFP fusion proteins and the Arabidopsis membrane marker, AtPIP2A--RFP fusion proteins, are shown. (G and H) Intensity analysis of the indicated line path in the protoplast shown in (F). Scale bar=10 \u00b5m.](erz154f0003){#F3}\n\n*Sl*SWEET1a activity complements glucose uptake deficiency {#s17}\n----------------------------------------------------------\n\nBased on genome structure analysis, the four clade I genes, *SlSWEET1a*, *1b*, *1c*, and *1d* were clustered in a tandem repeat, implying that these clade I homologs could have been derived from gene duplication ([@CIT0062]). Furthermore, due to high identity (31--81%) between amino acid sequences of clade I *SlSWEET* genes ([Supplementary Fig. S5](#sup1){ref-type=\"supplementary-material\"}), perhaps all clade I members have similar transport properties. Therefore, cDNAs of clade I *SlSWEET* genes, 10 genes in total, were isolated and expressed in the baker's yeast (*Saccharomyces cerevisiae*) mutant YSL2-1, which lacks all endogenous hexose transporters but expresses a cytosolic invertase and can grow only on medium containing maltose, but not monosaccharides or sucrose ([@CIT0028]). Expressing the yeast hexose transporter HXT5 allows YSL2-1 to grow on media containing glucose ([Fig. 4A](#F4){ref-type=\"fig\"}). Under acidic growth conditions (pH 5), expression of only *Sl*SWEET1a, 1b, and 3 complemented growth deficiency of YSL2-1 on glucose- and galactose-containing media. However, there was no growth on media with fructose, mannose, or sucrose ([Fig. 4A](#F4){ref-type=\"fig\"}; [Supplementary Fig. S6](#sup1){ref-type=\"supplementary-material\"}). There were similar trends on medium buffered at neutral pH (pH 7), suggesting that the transport activity of clade I *Sl*SWEETs was not dependent on a pH gradient.\n\n![Transport activity of *Sl*SWEET1a in yeast. (A) Complementary growth assay in the YSL2-1 yeast mutant line. Yeast transformants expressing clade I *SlSWEET* genes, yeast hexose transporter (Hxt5), or the empty vector (Vector) were cultured on solid media supplemented with 2% maltose, 2% glucose, or 1% maltose together with 0.2% 2-deoxyglucose (2-DG) under pH 5 and 7 conditions. Images were taken after incubation at 30 \u00b0C for 4--6 d. (B) Time-dependent \\[^14^C\\]glucose uptake activity was measured in YSL2-1 yeast cells expressing *SlSWEET1a* or the empty vector. Result are means \u00b1SE (*n*=3). Significant differences from cells expressing the empty vector were determined. (C) Kinetics of \\[^14^C\\]glucose uptake activity of *Sl*SWEET1a are shown. Results are means \u00b1SE (*n*=4--5). (D) Competition binding ability of other sugars was measured by incubating cells with both \\[^14^C\\]glucose and 10-fold concentrations of the indicated sugars (Glu, glucose; Gal, galactose; Fru, fructose; Man, mannose; Suc, sucrose; Mal, maltose). Relative uptake rate was calculated compared with cells incubated with only 1 mM hot glucose (control). The background uptake of non-transformed cells (empty) was also shown. Results are means \u00b1SE (*n*=5). Significant differences from the control were determined. (E) Effect of various pH conditions and treatment of NH~4~Cl on *Sl*SWEET1a uptake activity in EBY4000 cells. Yeast cells expressing the empty vector, *At*SWEET1, or *Sl*SWEET1a were analyzed. Results are means \u00b1SE (*n*=3). Significant differences from the pH 5 condition were determined. All statistics were performed by Student's *t*-test and indicated by one (*P*\\<0.05) or two (*P*\\<0.01) asterisks.](erz154f0004){#F4}\n\nYSL2-1 cells transformed with an empty vector grew on 2-deoxyglucose (2-DG) medium. In contrast, cells expressing HXT5 had no detectable growth due to uptake of 2-DG that is a poorly metabolized glucose analog and causes cell death in low amounts ([Fig. 4A](#F4){ref-type=\"fig\"}; [@CIT0011]). Similarly, only yeast cells expressing *Sl*SWEET1a, 1b, 2b, and 3 had a significant 2-DG-dependent growth inhibition, especially on medium buffered at neutral pH (pH 7). Furthermore low uptake activity of *Sl*SWEET1c and 1e was only observed at neutral pH. These results indicated that *Sl*SWEET1a, as well as some homologs in clade I, probably functions as a glucose-specific passive transporter.\n\nSWEET1 functions as a glucoses-specific low-affinity carrier {#s18}\n------------------------------------------------------------\n\nTo confirm the glucose-specific transport activity of *Sl*SWEET1a, indicated by the growth assay in YSL2-1 yeast cells ([Fig. 4A](#F4){ref-type=\"fig\"}), a time-dependent glucose uptake assay was done using \\[^14^C\\]glucose in YSL2-1 yeast cells expressing *Sl*SWEET1a. Radioactivity had accumulated in transformed yeast cells after 5 min, with a linear increase until 30 min, whereas there was limited glucose uptake in cells expressing the empty vector ([Fig. 4B](#F4){ref-type=\"fig\"}). The kinetics of the glucose transport activity of *Sl*SWEET1a were determined by exposure to glucose concentrations of 0.1--400 mM for 10 min. Significant uptake of \\[^14^C\\]glucose occurred with exposure to 0.1 mM, with saturation of glucose uptake at a concentration close to 200 mM ([Fig. 4C](#F4){ref-type=\"fig\"}). Therefore, the estimated *K*~m~ and *V*~max~ of *Sl*SWEET1a activities for glucose were 119.3 mM and 28.9 nmol 10^7^ cells^\u22121^ h^\u22121^, respectively, indicating a low-affinity transport mode.\n\nTo address further the substrate specificity of *Sl*SWEET1a, a 10-fold excess of unlabeled sugars was used to compete with \\[^14^C\\]glucose. Only a 10-fold excess of unlabeled glucose, but no other hexoses or sucrose, competed the binding and significantly reduced the amount of \\[^14^C\\]glucose imported ([Fig 4D](#F4){ref-type=\"fig\"}). The lower uptake rate to \\[^14^C\\]glucose under 10-fold higher maltose was attributed to background uptake activity in the YSL2-1 yeast strain ([@CIT0014]). The specific competition ability of unlabeled glucose was also measured by treatment with a 30-fold excess of sugar substrates ([Supplementary Fig. S7](#sup1){ref-type=\"supplementary-material\"}). To verify whether *Sl*SWEET1a proteins functioned as other known SWEET transporters, the EBY4000 yeast strain was used. Similar to *At*SWEET1 ([@CIT0014]), glucose uptake activity of *Sl*SWEET1a was not repressed by alkaline pH ([Fig. 4E](#F4){ref-type=\"fig\"}) and only slightly decreased by NH~4~Cl (10 mM) ([Fig. 4E](#F4){ref-type=\"fig\"}), perhaps due to indirect toxicity.\n\n*Sl*SWEET1a contributes to sugar distribution to sink leaves {#s19}\n------------------------------------------------------------\n\nTo explore the physiological role of *Sl*SWEET1a in leaves, VIGS was used to transiently silence expression of *SlSWEET1a* in leaves. The efficiency of VIGS was confirmed by expressing VIGS-*Sl*PDS in parallel plants with leaf etiolation due to silenced expression of the phytoene desaturase gene ([Supplementary Fig. S8](#sup1){ref-type=\"supplementary-material\"}). Samples collected from non-silenced and control plants (injected with infiltration buffers only, Control/NS) were compared with silenced plants in which *SlSWEET1a* expression in mature leaves was reduced by 50% ([Fig. 5A](#F5){ref-type=\"fig\"}). In silenced young leaves (VIGS-YL), glucose contents were significantly reduced (\\>50%) compared with control leaves ([Fig. 5B](#F5){ref-type=\"fig\"}). The same reduction was also observed for fructose, but not for sucrose. In contrast, glucose and fructose contents in silenced mature leaves were increased \\>2-fold compared with controls ([Fig. 5C](#F5){ref-type=\"fig\"}).\n\n![Effect of *Sl*SWEET1a deficiency on sugar allocation ability in tomato leaves. (A) Gene expression of *SlSWEET1a* in matures leaves of tomato plants injected with buffer (Control) or *Agrobacterium* cells containing the VIGS--*Sl*SWEET1a construct (non-silencing and silencing). Total RNA were isolated and cDNA was used for qRT-PCR with gene-specific primers. Relative expression by normalizing to an internal control, *SlActin7*. Numbers indicate fold changes of *SlSWEET1a* expression compared with control leaves. Results are means \u00b1SE (*n*=5--11). (B and C) Sugar contents in young (B) and mature (C) leaves of tomato plants subjected to the VIGS-*Sl*SWEET1a construct. Two samples from control plants and non-silenced plants in (A) were used as the control/NS for comparison. Results are means \u00b1SE (*n*=3--4). Significant differences from the control (A) or control/NS leaves (B and C) were determined by Student's *t*-test and indicated by one (*P*\\<0.05) or two (*P*\\<0.01) asterisks.](erz154f0005){#F5}\n\nDiscussion {#s20}\n==========\n\nFor decades, possible membrane carriers involved in apoplasmic sugar unloading in developing leaves have been the subject of debate. In our study, distinct high expression of *SlSWEET1a* was detected (qRT-PCR) in young leaves of vegetative tomato plants ([Fig. 1](#F1){ref-type=\"fig\"}), suggesting that the *Sl*SWEET1a proteins may mainly function in young leaves compared with source (mature) leaves. In addition, expression of a whole-gene *Sl*SWEET1a--GUS translational fusion further demonstrated that *Sl*SWEET1a proteins were co-localized in leaf vein tissues ([Fig. 2](#F2){ref-type=\"fig\"}; [Supplementary Fig. S3](#sup1){ref-type=\"supplementary-material\"}). The GUS staining was most evident in the first- to the third-order veins, representing major sites for unloading of photoassimilates from the transport phloem in both tomato sink leaves ([Fig. 2G](#F2){ref-type=\"fig\"}, H) and developing tobacco leaves ([@CIT0074]; [@CIT0054]). Moreover, the carbohydrate unloading activity in a sink leaflet co-occurred with relatively high levels of *SlSWEET1a* transcripts in the leaflet base, compared with the tip ([Fig. 2G--I](#F2){ref-type=\"fig\"}), which is the first region to switch to sugar export during sink/source transition ([@CIT0074]; [@CIT0054]). Based on these results, we suspect that *Sl*SWEET1a probably functions in mediating sugar flows along the phloem unloading pathway in sink leaves. The phloem unit is composed of SE--CC complexes and phloem parenchyma, with central control of sugar loading, transport, and unloading in plants ([@CIT0047]; [@CIT0039]). Before post-phloem unloading to neighboring sink cells, sucrose needs to be exported out of SE--CC complexes and into phloem parenchyma ([@CIT0050]). Sucrose movement between SEs and CCs occurs mostly symplasmically ([@CIT0074]; [@CIT0036]; [@CIT0050]; [@CIT0039]). In contrast, the sucrose efflux between CCs and phloem parenchyma can be mediated in parallel via sym- and apoplasmic routes (with the latter requiring sugar-transporting proteins, such as *Sl*SWEET1a).\n\nTo confer apoplasmic sugar unloading to sink leaves, the plasma membrane would be the primary site to control sugar fluxes. Co-localization of *Sl*SWEET1a--GFP fusion protein with a plasma membrane protein, namely *At*PIP2--RFP ([@CIT0046]), at single-cell resolution was evidence that *Sl*SWEET1a was located at the plasma membrane ([Fig. 3](#F3){ref-type=\"fig\"}; [Supplementary Fig. S4](#sup1){ref-type=\"supplementary-material\"}). The specific subcellular localization of *Sl*SWEET1a in single cells was consistent with transient expression of *Sl*SWEET1a--GFP in leaves of tobacco and Arabidopsis ([@CIT0062]). All major sugar loaders, SUCs/SUTs and STPs, are located in the plasma membrane ([@CIT0029]; [@CIT0002]; [@CIT0017]; [@CIT0003]; [@CIT0051]). In contrast, SWEETs expressed in sink organs were present in plasma or vacuolar membranes ([@CIT0014]; [@CIT0010]; [@CIT0012]; [@CIT0011]). Similar to key SWEET exporters to supply sugars in reproductive organs, *Sl*SWEET1a locates at the correct cellular site, the plasma membrane, to be involved in sugar unloading. Taken together, dominant accumulation of *Sl*SWEET1a proteins in the plasma membrane of vein tissues in young leaves supported its putative role to mediate sugar fluxes across the apoplasm between CCs and phloem parenchyma during sugar unloading to sink leaves.\n\nAlthough strong *SlSWEET1a* RNA accumulation also occurred in flowers during the reproductive stage, most *Sl*SWEET1a protein accumulated in stamens ([Fig. 2E](#F2){ref-type=\"fig\"}, F), suggesting that *Sl*SWEET1a is probably not involved in phloem sugar unloading during the reproductive stage. Furthermore, *AtSWEET8* (also named *AtRPG1*) has been demonstrated to be strongly expressed in Arabidopsis pollen, possibly to mediate import of apoplasmic hexoses required for microspore development ([@CIT0027]). Therefore, whether *Sl*SWEET1a participates in post-phloem unloading required for microsporogenesis in flowers requires further investigation. Interestingly, induced expression of *SlSWEET1a* during fruit maturation implies that *Sl*SWEET1a is involved in sugar homeostasis during fruit maturation. This conclusion correlates with the previous observation that variation of *SlSWEET1a* expression correlates with hexose composition of ripening tomato fruits ([@CIT0062]). Taken together, these results suggest that *Sl*SWEET1a probably has multiple physiological functions, depending on tissue localization and developmental stages.\n\nTransport analyses on yeast cells using complementary growth assays or a radiotracer confirmed that *Sl*SWEET1a was a glucose-specific transporter, consistent with its closest Arabidopsis homolog, *At*SWEET1 ([Fig. 4](#F4){ref-type=\"fig\"}; [Supplementary Fig. S6, S7](#sup1){ref-type=\"supplementary-material\"}; [@CIT0014]) and previous speculations ([@CIT0062]). In contrast, many hexose transporters, such as TMT/TST (tonoplast monosaccharide transporter/tonoplast sugar transporter) or some STPs, have a broad substrate range ([@CIT0066]). For example, Arabidopsis *At*TST1 transports glucose, fructose, and sucrose ([@CIT0082]; [@CIT0061]). Sugar beet *Bv*TST1 also transports both glucose and sucrose ([@CIT0035]), whereas *Cs*HT1 (hexose transporter from *Cucumis sativus*) transports glucose, galactose, mannose, and xylose ([@CIT0017]). Functional redundancy between MSTs would be consistent with MST proteins not having a confirmed physiological role, despite their dominant expression in heterotrophic sink tissues ([@CIT0064]; [@CIT0066]; [@CIT0084]). Thus, high specificity of *Sl*SWEET1a to glucose reflects its putative importance in hexose transport during sugar unloading in young leaves.\n\nIn contrast to most plasma membrane-localized MSTs exhibiting a low *K*~m~ (2--10 mM) ([@CIT0008]; [@CIT0009]; [@CIT0038]; [@CIT0052]), *Sl*SWEET1a import activity exhibited a relatively high *K*~m~ (119 mM) for glucose ([Fig. 4C](#F4){ref-type=\"fig\"}). Although this low-affinity feature may be due to a high turnover rate of *Sl*SWEET1a proteins ([@CIT0013]), it also highlights the possibility that *Sl*SWEET1a actively operates under high physiological concentrations of apoplasmic glucose. Due to significantly higher expression and activities of cell wall invertases observed in young leaves compared with mature leaves ([@CIT0063]; [@CIT0033]), sucrose unloaded to the leaf apoplasm could be hydrolyzed and glucose taken up via the hexose carrier. These observations further support the need for a low-affinity glucose importer, such as *Sl*SWEET1a, operating along the initial sugar unloading path from CCs to parenchyma cells in developing leaves.\n\nMoreover, *Sl*SWEET1a was a proton-independent bidirectional uniporter ([Fig. 4A](#F4){ref-type=\"fig\"}, E), in agreement with all other *At*SWEET members ([@CIT0016]; [@CIT0028]). In that case, a passive glucose carrier, such as *Sl*SWEET1a, would possess a more efficient transport system in response to instant sugar fluxes to assist continuous sugar unloading without the delay for energy biosynthesis. This was nicely in line with passive sugar fluxes observed in young leaves, where the sucrose unloading is insensitive to anoxia and PCMBS, a metabolite inhibitor ([@CIT0060]; [@CIT0074]; [@CIT0022]). Furthermore, proton ATPase is not expressed in the protophloem of growing sinks, such as root tips and expanding region of leaves ([@CIT0070]; [@CIT0073]), indicating a lack of energized transport. Taken together, all transport features of *Sl*SWEET1a supported its putative role in participating in initial sugar unloading from SE--CC complexes to phloem parenchyma in sink leaves.\n\nOur physiological evidence supported the assertion that *Sl*SWEET1a import activity regulated hexose accumulation in young leaves. During the vegetative growth stage of tomato plants, suppressed expression of *Sl*SWEET1a by VIGS significantly reduced accumulation of glucose and fructose in young leaves, with concurrent increased concentrations of these sugars in source leaves ([Fig. 5](#F5){ref-type=\"fig\"}). Since sucrose is the dominant type of phloem-transported sugar ([@CIT0071]; [@CIT0042]), it follows that apoplasmic glucose in sink leaves is mainly derived from hydrolysis of sucrose unloaded into the sink leaf apoplasm. Based on its low affinity, vein localization, and glucose-specific uniport feature, our results were best explained by the model that some *Sl*SWEET1a on the plasma membrane of phloem parenchyma significantly contributes to hexose import in sink leaves. The high glucose concentrations derived from sucrose hydrolysis during phloem unloading drive the import activity of *Sl*SWEET1a to enable glucose uptake. Once in the cytosol, glucose can be quickly distributed to sink cells via symplasmic connections ([@CIT0060]), or it can be compartmentalized into corresponding vacuoles ([@CIT0051]). The finding that knock out of *Sl*SWEET1a resulted in hexose concentrations being reduced in sink leaves is consistent with *Sl*SWEET1a playing a key role in the phloem unloading of hexoses. Similar conclusions have been reached for knock down of the hexose symporter that drives hexose uptake from the apoplasm into storage parenchyma cells of developing tomato fruit ([@CIT0045]) and of *Zm*SWEET4, located on the membrane of basal endosperm transfer cells, mediating hexose uptake into the endosperm of developing maize seeds ([@CIT0068]).\n\nThese results also concur nicely with the report that reduced sink strength by removing active sinks, such as developing leaves and fruits, stimulated sucrose accumulation in mature leaves of a citrus tree ([@CIT0031]), and with the observation that removal of young axillary leaves increased the size of mature leaves and the number of nodes in tomato plants ([@CIT0019]). Furthermore, as observed in wheat grains, once sugar import into sink efflux was diminished by detaching the sink, sucrose efflux out of the SE--CC complex declined immediately ([@CIT0022]). As a consequence, the resulting high sucrose contents in the phloem will suppress expression of the phloem loader, SUC/SUT2 symporter ([@CIT0018]; [@CIT0076]), and thus reduce whole-plant sugar loading and lead to increased carbohydrates in sucrose mesophyll cells ([@CIT0076]). We inferred that the import activity of *Sl*SWEET1a may be an indicator for sink strength and is thus a novel element in whole-plant carbon partitioning from source leaves to sink leaves.\n\nSummary {#s21}\n=======\n\nHerein, we uncovered important functions of the plasma membrane-located uniporter *Sl*SWEET1a in glucose uptake during apoplasmic phloem unloading to young leaves ([Supplementary Fig. S9](#sup1){ref-type=\"supplementary-material\"}). We propose that in vegetative tomato plants, once sucrose is allocated to sink leaves via phloem pressure flow, it will passively exit out of the SE--CC complexes (symplasmically or apoplasmically). In the apoplasmic route, some sucrose may be exported by an unknown sucrose exporter and be directly taken up through an unidentified sucrose carrier, possibly a SUT member. Yet, a significant portion of sucrose is hydrolyzed to glucose and fructose by extracellular invertase in the apoplasm around phloem parenchyma, partially contributing to formation of a sucrose gradient from SE--CC complexes to facilitate efflux ([@CIT0048]). Then, *Sl*SWEET1a, partially located in the plasma membrane of unloading cells, such as phloem parenchyma cells, promotes glucose import to increase apoplasmic solute potential so as to maintain lower turgor pressure toward the SE--CC symplast for continuous sucrose unloading. The fructose-specific carrier remains to be identified. Further studies will be needed to dissect how *Sl*SWEET1a import activity is coordinated with sink strengths during transition of leaf development and transition from the vegetative stage to the reproductive stage. We also speculated that efficient unloading to sink leaves was important for whole-plant growth. Thus, engineering-efficient unloading of hexoses in phloem parenchyma, for example by increasing *Sl*SWEET1a import activity, coupling symport with higher source import capacity ([@CIT0001]), may be key for future crop enhancement.\n\nSupplementary data {#s22}\n==================\n\nSupplementary data are available at *JXB* online.\n\nTable S1. Primers used for qRT-PCR.\n\nTable S2. Primers used to generate yeast-expressing constructs for clade I *Sl*SWEET genes.\n\nFig. S1. Images of the tomato organs analyzed.\n\nFig. S2. Phylogenetic tree of 31 tomato *SlSWEET* genes.\n\nFig. S3. Tissue-specific expression of *Sl*SWEET1a proteins in five T~1~ plants.\n\nFig. S4. Subcellular localization of *Sl*SWEET1a in tobacco and Arabidopsis protoplasts.\n\nFig. S5. Amino acid identity between clade I *Sl*SWEET genes.\n\nFig. S6. Transport activities of clade I *Sl*SWEETs for various sugars in yeast.\n\nFig. S7. Competition binding ability of sugars to *Sl*SWEET1a transport.\n\nFig. S8. Images of tomato plants subject to VIGS.\n\nFig. S9. Proposed model of *Sl*SWEET1a function.\n\n###### \n\nClick here for additional data file.\n\nWe thank Dr Yi-Min Chen at the National Cheng Kung University for his assistance in the GC-MS analysis. We acknowledge the confocal core facility at the National Cheng Kung University for fluorescence observations, and the transgenic plant core lab in Academia Sinica for generating transgenic tomato plants. This work was financially supported by grants from the Ministry of Science and Technology, Taiwan (MOST 105-2628-B-006-001-MY3) and from the DAAD/NSC PPP program (106-2911-I-006-506; 107-2911-I-006-506) to WJG.\n"}
+{"text": "![](hosplond72491-0039){#sp1 .117}\n"}
+{"text": "1. Introduction {#se0010}\n===============\n\nJupiter\\'s moon, Europa, and Saturn\\'s moon, Enceladus, are two of only a few moons in the solar system that bear the possibility of extraterrestrial life (e.g., [@br0090], [@br0270]). Europa has a deep (\u223c100 km) ocean that underlies an icy shell, more than several kilometers deep (e.g., [@br0070], [@br0060], [@br0280], [@br0170], [@br0160], [@br0250], [@br0390], [@br0320], [@br0420]), where chemical interactions at the rocky bottom of the ocean may enable the existence of a habitable environment (e.g., [@br0090], [@br0100], [@br0130], [@br0200]). The Voyager and Galileo (and to a lesser extent, the Cassini-Huygens and New Horizons) spacecrafts/missions discovered many interesting features of Europa including chaos terrains ([@br0330], [@br0410]) and craters ([@br0190], [@br0230], [@br0120], [@br0340]). More recently, based on the Hubble telescope observations, scientists raised the possibility of water vapor plumes at Europa\\'s south pole ([@br0310], [@br0350]). Europa is one of the youngest, largest, and brightest moons in the solar system ([@br0290]). A basic property of Europa is its surface temperature; surface temperature is needed to calculate the properties of its icy shell and ocean dynamics. Thus, an accurate estimation of Europa\\'s surface temperature is required.\n\nThe annual mean surface temperature of Europa was previously estimated by [@br0260]. These authors took into account the obliquity of Europa with respect to the plane of rotation of Jupiter around the Sun. Using a surface albedo of 0.5, they found that the annual mean temperature varies from \u223c110 K at the equator to \u223c52 K at the poles; the global (and annual) mean surface temperature was found to be \u223c100 K (see the gray curve in [Fig. 2](#fg0020){ref-type=\"fig\"}d below). However, the following factors were not taken into account when calculating the annual mean surface temperature of Europa: the eccentricity of Jupiter\\'s orbit around the Sun, the effect of the emissivity of Europa\\'s ice, the heat capacity of the surface ice, the effect of Europa\\'s eclipse (i.e., the time that Europa is within the shadow of Jupiter), the longwave radiation of Jupiter that is absorbed by Europa, and Europa\\'s internal heating. The eccentricity \\[see Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"} and [(21)](#fm0210){ref-type=\"disp-formula\"} below\\], emissivity, Jupiter\\'s radiation, and internal heating factors increase the incoming radiation to Europa\\'s surface from above and below, while the eclipse factor reduces the absorbed incoming solar radiation. In addition, an updated surface Bond albedo of $0.68 \\pm 0.05$ ([@br0140]) that is based on New Horizons measurements should be considered; this by itself reduces the absorbed incoming solar radiation by more than 20%. Moreover, to our knowledge, the seasonal cycle of the surface temperature of Europa has not been previously estimated. The goal of this study is to develop a more accurate estimation of the diurnal, seasonal and annual mean surface temperature of Europa, taking all the above factors into account; here, we systematically investigate the role of the different parameters on the surface temperature of Europa. In addition, the analytic approximation developed here may be used to estimate the surface temperature of other moons in the solar system.\n\nThe brightness temperature of Europa was measured by [@br0370] and [@br0300], based on measurements performed by the Galileo spacecraft. They measured the diurnal temperature cycle at the low latitudes to be between 86 K and 132 K and also provided spatial snapshots of surface temperatures up to, roughly, 70^\u2218^ latitude. Still, the diurnal mean temperature for latitudes poleward of 15^\u2218^ and the temperatures at the high latitudes were not measured. The polar region temperatures are important as in these regions, the internal heating significantly affects the surface temperatures, and measurements of surface temperatures in these regions may help to estimate the internal heating rate and, consequently, the thickness of the ice (see, e.g., [@br0020]). The dependence of surface temperature on the internal heating is studied here.\n\nBelow we first discuss the diurnal, seasonal, and annual mean incoming solar radiation to Europa (Section [2.1](#se0020){ref-type=\"sec\"}). Some of Jupiter\\'s longwave radiation is absorbed by Europa, and this effect is quantified in Section [2.4](#se0070){ref-type=\"sec\"}. We then quantify the effect of Europa\\'s eclipse, i.e., the relative time that Europa is within the shadow of Jupiter (Section [2.5](#se0080){ref-type=\"sec\"}). Next we calculate the surface temperature of Europa (Section [3.1](#se0100){ref-type=\"sec\"}). A summary and discussion close the paper (Section [4](#se0120){ref-type=\"sec\"}). A very rough estimation of the mean thickness of Europa\\'s icy shell as a function of the internal heating rate is detailed in Section [3.2](#se0110){ref-type=\"sec\"}. The parameters that are used in this study are listed in [Table 1](#tbl0010){ref-type=\"table\"}.Table 1List of parameters.Table 1ParameterDescriptionValue*S*~0~Jupiter solar constant51 W\u202fm^\u22122^*e*eccentricity of Jupiter0.048*\u03b5*obliquity of Europa3^\u2218^*\u03b1*~*p*~bolometric Bond albedo of Europa0.68\u202f\u00b1\u202f0.05*p*Europa eclipse relative time0.033*\u03c3*Stefan-Boltzmann constant5.67\u202f\u00d7\u202f10^\u22128^ W\u202fm^\u22122^\u202fK^\u22124^*\u03f5*emissivity of Europa0.94*\u03c1*~*I*~density of ice917 kg\u202fm^\u22123^*c*~*p*,*I*~heat capacity of ice2000 J\u202fkg^\u22121^\u202fK^\u22121^*\u03ba*deep ice heat diffusion constant1.54\u202f\u00d7\u202f10^\u22126^ m^2^\u202fs^\u22121^*\u03ba*~*s*~surface ice heat diffusion constant7.7\u202f\u00d7\u202f10^\u221210^ m^2^\u202fs^\u22121^*g*surface gravity of Europa1.314 m\u202fs^\u22122^*r*~*j*~mean Jupiter-Sun distance7.785\u202f\u00d7\u202f10^11^ m*r*~*e*~mean Europa-Jupiter distance6.71\u202f\u00d7\u202f10^8^ m*a*~*s*~radius of Sun6.96\u202f\u00d7\u202f10^8^ m*a*~*j*~radius of Jupiter6.99\u202f\u00d7\u202f10^7^ m*a*~*e*~radius of Europa1.561\u202f\u00d7\u202f10^6^ m\u03a9~*e*~rotational frequency of Europa2.05\u202f\u00d7\u202f10^\u22125^ s^\u22121^*\u03c9*Jupiter\\'s longitude of the perihelion14.7285^\u2218^*J*~0~Jupiter\\'s radiation constant0.176 W\u202fm^\u22122^\n\n2. Theory {#se0180}\n=========\n\n2.1. The incoming solar radiation {#se0020}\n---------------------------------\n\nBelow, we present results regarding the diurnal cycle, seasonal cycle, and annual mean solar radiation reaching the surface of Europa. Details regarding the derivations of the daily mean insolation can be found in Section [2.2](#se0030){ref-type=\"sec\"} and regarding the seasonal and annual mean insolation in Section [2.3](#se0040){ref-type=\"sec\"}.\n\n[Fig. 1](#fg0010){ref-type=\"fig\"}a depicts the diurnal cycle of the incoming solar radiation at the equator during the northern hemisphere (NH) vernal (spring) equinox and NH summer solstice and at 89^\u2218^S during the southern hemisphere (SH) summer solstice. The incoming solar radiation peaks at the equator during the NH summer solstice and does not exceed the value of the solar constant of Europa/Jupiter, $S_{0} = 51\\ \\text{W}\\ \\text{m}^{- 2}$; the maximal diurnal solar radiation is $\\sim S_{0}$ when the eccentricity is small \\[see Eq. [(1)](#fm0010){ref-type=\"disp-formula\"}\\], as for Europa. The NH summer solstice radiation is higher than the NH vernal equinox radiation as Jupiter (and hence Europa) is closer to the Sun at the NH summer solstice such that the eccentricity effect overcomes the obliquity effect. In other words, at the present time, the solstices are close to the aphelion and perihelion (see [Fig. 2](#fg0020){ref-type=\"fig\"}b below); this situation may be different at different times due to the long (many thousands of years) precession cycle of Jupiter. Close to the pole, even during the NH summer solstice (blue curve), the radiation is much smaller, as is the radiation range.Figure 1(a) Diurnal cycle of the incoming solar radiation at the equator during the equinox (*\u03bb*\u202f=\u202f0, NH vernal equinox, green) and solstice (*\u03bb*\u202f=\u202f*\u03c0*/2, NH summer solstice, red), and at 89^\u2218^S during the solstice (*\u03bb*\u202f=\u202f3*\u03c0*/2, NH winter solstice, blue). (b) Diurnal cycle of the surface temperature at the equator during the equinox (*\u03bb*\u202f=\u202f0, NH vernal equinox, green) and the solstice (*\u03bb*\u202f=\u202f*\u03c0*/2, NH summer solstice, red). The dotted and dashed horizontal lines indicate the daily mean surface temperature calculated using the depth-dependent model \\[Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}\\] (dotted) and the daily mean radiation \\[Eq. [(32)](#fm0320){ref-type=\"disp-formula\"}\\] (dashed). (c) The temperature versus depth at the equator at the equinox (*\u03bb*\u202f=\u202f0, NH vernal equinox) at several times during the diurnal cycle. Note the fast decline in temperature oscillations with depth. (d) Same as b but for 89^\u2218^S, at the solstice (*\u03bb*\u202f=\u202f3*\u03c0*/2, NH winter solstice). The temperature was obtained by integrating the depth-dependent model \\[Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}\\] at a specific latitude.Figure 1Figure 2(a) Annual mean (blue and red) and NH summer (SH winter) solstice (yellow) incoming solar radiation as a function of latitude. Both the numerically integrated (blue) and the analytic (red) approximation \\[Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"}, [(21)](#fm0210){ref-type=\"disp-formula\"}\\] are shown although the two are almost indistinguishable (the maximal difference between the two is 0.01 W\u202fm^\u22122^). (b) Seasonal variations of the surface temperature versus time (in Jupiter\\'s years) for the equator (blue) and for the NH pole (red). The internal heating rate is *Q*\u202f=\u202f0.05 W\u202fm^\u22122^ and the curves are the solution of Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}. The dashed horizontal lines indicate the annual mean temperature based on the annual mean insolation \\[Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"}, [(21)](#fm0210){ref-type=\"disp-formula\"}\\] while the dotted lines indicate the annual mean temperature of the solid lines. *t*\u202f=\u202f0 is set at the spring (vernal) equinox and the dotted vertical lines indicate the time of the aphelion and perihelion. (c) Surface temperature at the NH summer solstice for different internal heating rates. (d) Annual mean surface temperature for different internal heating rates. The gray line indicates the surface temperature as estimated by [@br0260] based on [@br0240].Figure 2\n\nIn [Fig. 2](#fg0020){ref-type=\"fig\"}a, we plot the annual mean and NH summer solstice incoming solar radiation as a function of latitude. Both the numerically integrated and the analytically approximated annual mean curves are presented, and the two are almost indistinguishable. The numerical integration was based on the daily mean insolation \\[Eqs. [(6)](#fm0060){ref-type=\"disp-formula\"}, [(7)](#fm0070){ref-type=\"disp-formula\"}\\] using Eq. [(12)](#fm0120){ref-type=\"disp-formula\"} and $\\langle W\\rangle = \\frac{1}{\\tau}\\int_{0}^{2\\pi}W\\frac{dt}{d\\lambda}d\\lambda$ with $d\\lambda = 0.01\\ \\text{rad}$. Note the very small level of incoming solar radiation ($\\sim 1\\ \\text{W}\\ \\text{m}^{- 2}$) that reaches the poles, suggesting that the internal heating cannot be ignored at the high latitudes. No solar radiation reaches the winter pole during the corresponding winter solstice; thus, we expect the internal heating to have a larger influence on polar region surface temperature during this time.\n\n2.2. Derivation of the daily mean insolation {#se0030}\n--------------------------------------------\n\nBelow we provide the mathematical details regarding the derivations of the daily mean insolation. The reader that is not interested in the mathematical details may skip this section.\n\nThe incoming insolation *W* at a certain latitude *\u03d5* and a certain time of the day on Europa ([@br0150]) is$$W = S_{0}\\left( \\frac{\\overline{d}}{d} \\right)^{2}\\cos\\theta_{s},$$ where $\\theta_{s}$ is the solar zenith angle, $\\overline{d}$ is the mean distance of Jupiter (Europa) from the Sun, *d* is the distance from the Sun, and $S_{0}$ is the solar constant of Jupiter. $\\theta_{s}$ is the solar zenith angle that depends on latitude, season, and time of day and$$\\frac{\\overline{d}}{d} = \\frac{1 + e\\cos(\\lambda - \\omega - \\pi)}{1 - e^{2}},$$ where *e* is the eccentricity of Jupiter, *\u03c9* is the precession (longitude of the perihelion), and *\u03bb* is the longitude of Jupiter with respect to its orbit around the Sun. Seasons are expressed through the declination angle *\u03b4*, which is the latitude of the line connecting the center of Europa and the Sun during noontime. The hour angle *H* indicates the longitude of the subsolar point relative to its position at noon. Then,$$\\cos\\theta_{s} = \\sin\\phi\\sin\\delta + \\cos\\phi\\cos\\delta\\cos H.$$ At sunrise and sunset, the zenith angle is $90^{\\circ}$ such that $\\cos\\theta_{s} = 0$ and$$\\cos H_{0} = - \\tan\\phi\\tan\\delta.$$ The declination angle *\u03b4* is$$\\sin\\delta = \\sin\\varepsilon\\sin\\lambda,$$ where *\u03b5* is the obliquity of Europa with respect to the plane of rotation of Jupiter around the Sun; the maximal obliquity can be calculated as the sum of the axial tilt of Europa, 0.1^\u2218^, the inclination angle of Europa, 0.47^\u2218^, and the axial tilt of Jupiter, 3.13^\u2218^, yielding an angle of 3.7^\u2218^. Yet, since the axes of rotation of both Europa and Jupiter exhibit precession, the above angles can either add up or subtract. We thus estimated the mean present day obliquity as the subsolar latitude at the solstices based on the JPL Horizons web-interface (<https://ssd.jpl.nasa.gov/horizons.cgi#top>); it is $\\varepsilon \\approx 3^{\\circ}$. We performed the calculations presented in this paper also using $\\varepsilon = 3.7^{\\circ}$ and obtained almost identical results for the low latitudes and global mean values, while at the polar regions, the temperature was higher by a few degrees for $\\varepsilon = 3.7^{\\circ}$; see [Table 2](#tbl0020){ref-type=\"table\"}. At the solstice, $\\sin\\lambda = \\pm 1$ such that $\\delta = \\pm \\varepsilon$, while at the equinox, $\\delta = 0$.Table 2Sensitivity of surface temperature (K) of Europa to the different parameters when using the numerical model \\[Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}\\]. Control run parameters: eccentricity (*e*\u202f=\u202f0.048), obliquity (*\u03b5*\u202f=\u202f3^\u2218^), emissivity (*\u03f5*\u202f=\u202f0.94), albedo (*\u03b1*~*p*~\u202f=\u202f0.68), Europa eclipse relative time (*p*\u202f=\u202f0.033), surface ice heat diffusion constant (*\u03ba*~*s*~\u202f=\u202f7.7\u202f\u00d7\u202f10^\u221210^ m^2^\u202fs^\u22121^), Jupiter\\'s longwave radiation constant of Europa (*J*~0~\u202f=\u202f0.176 W\u202fm^\u22122^), and internal heating (*Q*\u202f=\u202f0.05 W\u202fm^\u22122^).Table 2ParameterEquatorPoleGlobal mean15^\u2218^S--15^\u2218^N meancontrol96.146.190.195.9*e*\u202f=\u202f095.845.989.995.6*\u03b5*\u202f=\u202f096.231.290.195.9*\u03b5*\u202f=\u202f3.7^\u2218^96.147.990.195.9*\u03f5*\u202f=\u202f194.645.489.495.1*\u03b1*~*p*~\u202f=\u202f0.5105.849.794.8100.9*p*\u202f=\u202f096.846.490.596.2*\u03ba*~*s*~\u202f=\u202f7.7\u202f\u00d7\u202f10^\u22128^ m^2^\u202fs^\u22121^99.349.091.497.5*J*~0~\u202f=\u202f095.846.190.095.7*Q*\u202f=\u202f095.842.889.995.7*Q*\u202f=\u202f0.2 W\u202fm^\u22122^97.052.890.796.3\n\n2.3. Derivation of the annual mean insolation {#se0040}\n---------------------------------------------\n\nBelow we provide the mathematical details regarding the derivations of the annual mean insolation. The reader that is not interested in the mathematical details may skip this section.\n\n### 2.3.1. Small obliquity {#se0050}\n\nThe daily mean insolation ([@br0220], [@br0040], [@br0050], [@br0150], [@br0010]) can be found by integrating Eq. [(1)](#fm0010){ref-type=\"disp-formula\"} from $- H_{0}$ to $H_{0}$ and dividing by 2*\u03c0* and is given by$$W = \\frac{S_{0}}{\\pi}\\frac{{\\lbrack 1 + e\\cos(\\lambda - \\omega - \\pi)\\rbrack}^{2}}{{(1 - e^{2})}^{2}}(H_{0}\\sin\\phi\\sin\\delta + \\cos\\phi\\cos\\delta\\sin H_{0}),$$ where $H_{0}$ is the hour angle at sunrise and sunset \\[Eq. [(4)](#fm0040){ref-type=\"disp-formula\"}\\]; see [Table 1](#tbl0010){ref-type=\"table\"}. Equation [(6)](#fm0060){ref-type=\"disp-formula\"} is relevant for latitudes $|\\phi| < |\\pi/2 - |\\delta||$, while outside this latitude range, the insolation during the polar night is zero and during the polar day is equal to$$W = S_{0}\\frac{{\\lbrack 1 + e\\cos(\\lambda - \\omega - \\pi)\\rbrack}^{2}}{{(1 - e^{2})}^{2}}\\sin\\phi\\sin\\delta.$$\n\nBelow we will use the following mathematical relations:$$\\tan\\delta = \\frac{\\sin\\delta}{\\sqrt{1 - \\sin^{2}\\delta}} = \\frac{\\sin\\varepsilon\\sin\\lambda}{\\sqrt{1 - \\sin^{2}\\varepsilon\\sin^{2}\\lambda}}.$$ Since the obliquity is small, one can perform the following approximations$$\\tan\\delta \\approx \\sin\\delta \\approx \\delta \\approx \\varepsilon\\sin\\lambda$$$$\\sin H_{0} = \\sqrt{1 - \\tan^{2}\\phi\\tan^{2}\\delta} \\approx \\sqrt{1 - \\varepsilon^{2}\\tan^{2}\\phi\\sin^{2}\\lambda}$$ and, based on Eqs. [(4)](#fm0040){ref-type=\"disp-formula\"} and [(9)](#fm0090){ref-type=\"disp-formula\"}$$H_{0} = \\arccos( - \\varepsilon\\tan\\phi\\sin\\lambda).$$\n\nTo calculate the annual mean insolation, one must take into account the derivative of the time, *t*, with respect to the longitude, *\u03bb* ([@br0220], [@br0150], [@br0010])$$\\frac{dt}{d\\lambda} = \\frac{{(1 - e^{2})}^{3/2}}{{\\lbrack 1 + e\\cos(\\lambda - \\omega - \\pi)\\rbrack}^{2}}.$$ Then, the relative time, *\u03c4*, of one cycle of Jupiter around the Sun may be approximated as$$\\tau = \\int\\limits_{0}^{2\\pi}\\frac{dt}{d\\lambda}d\\lambda \\approx 2\\pi{(1 - e^{2})}^{3/2},$$ as the eccentricity is much smaller than one. The relative time, *\u03c4*, is the ratio of the time passed since the NH summer solstice ($\\lambda = 0$) to the time it takes for Jupiter to complete one cycle around the Sun.\n\nThe total insolation during the year is then$$\\int\\limits_{0}^{2\\pi}W\\frac{dt}{d\\lambda}d\\lambda = \\frac{S_{0}}{\\pi\\sqrt{1 - e^{2}}}\\int\\limits_{0}^{2\\pi}(H_{0}\\sin\\phi\\sin\\delta + \\cos\\phi\\cos\\delta\\sin H_{0})d\\lambda.$$ We solve each part of the integral separately. The first part is$$\\sin\\phi\\int\\limits_{0}^{2\\pi}H_{0}\\sin\\delta d\\lambda = \\varepsilon\\sin\\phi\\int\\limits_{0}^{2\\pi}\\arccos( - \\varepsilon\\tan\\phi\\sin\\lambda)\\sin\\lambda d\\lambda\\quad \\approx \\varepsilon\\sin\\phi\\int\\limits_{0}^{2\\pi}\\left( \\frac{\\pi}{2} + \\varepsilon\\tan\\phi\\sin\\lambda \\right)\\sin\\lambda d\\lambda = \\varepsilon^{2}\\tan\\phi\\sin\\phi\\int\\limits_{0}^{2\\pi}\\sin^{2}\\lambda d\\lambda\\quad = \\pi\\varepsilon^{2}\\sin\\phi\\tan\\phi.$$\n\nThe second part of the integral of Eq. [(14)](#fm0140){ref-type=\"disp-formula\"} is$$\\cos\\phi\\int\\limits_{0}^{2\\pi}\\cos\\delta\\sin H_{0}d\\lambda = \\cos\\phi\\int\\limits_{0}^{2\\pi}\\sqrt{1 - \\sin^{2}\\varepsilon\\sin^{2}\\lambda}\\sqrt{1 - \\varepsilon^{2}\\tan^{2}\\phi\\sin^{2}\\lambda}d\\lambda \\approx \\cos\\phi\\int\\limits_{0}^{2\\pi}\\left( 1 - \\frac{1}{2}\\sin^{2}\\varepsilon\\sin^{2}\\lambda \\right)\\left( 1 - \\frac{1}{2}\\varepsilon^{2}\\tan^{2}\\phi\\sin^{2}\\lambda \\right)d\\lambda \\approx \\cos\\phi\\int\\limits_{0}^{2\\pi}\\left( 1 - \\frac{1}{2}\\sin^{2}\\lambda(\\sin^{2}\\varepsilon + \\varepsilon^{2}\\tan^{2}\\phi) \\right)d\\lambda = \\frac{1}{2}\\pi\\cos\\phi(4 - \\sin^{2}\\varepsilon - \\varepsilon^{2}\\tan^{2}\\phi).$$\n\nFollowing the above, the total insolation during the year can be approximated as:$$\\int\\limits_{0}^{2\\pi}W\\frac{dt}{d\\lambda}d\\lambda \\approx \\frac{S_{0}}{2\\sqrt{1 - e^{2}}}(4\\cos\\phi + \\varepsilon^{2}\\sin\\phi\\tan\\phi - \\cos\\phi\\sin^{2}\\varepsilon).$$ Finally, the annual mean insolation outside the polar regions is$$\\langle W\\rangle = \\frac{1}{\\tau}\\int\\limits_{0}^{2\\pi}W\\frac{dt}{d\\lambda}d\\lambda \\approx \\frac{S_{0}}{4\\pi{(1 - e^{2})}^{2}}\\left( 4\\cos\\phi + \\varepsilon^{2}\\sin\\phi\\tan\\phi - \\cos\\phi\\sin^{2}\\varepsilon \\right).$$\n\nTo find the annual mean insolation in the polar region, we first find the total annual insolation at the pole ($\\phi = \\pi/2$),$$\\int\\limits_{0}^{2\\pi}W\\frac{dt}{d\\lambda}d\\lambda = \\frac{S_{0}\\sin\\varepsilon}{\\sqrt{1 - e^{2}}}\\int\\limits_{0}^{\\pi}\\sin\\lambda d\\lambda = \\frac{2S_{0}\\sin\\varepsilon}{\\sqrt{1 - e^{2}}}.$$ Then, using Eq. [(13)](#fm0130){ref-type=\"disp-formula\"}, we find the annual mean isolation at the poles:$$\\langle W(\\phi = \\pm \\pi/2)\\rangle = \\frac{S_{0}\\sin\\varepsilon}{\\pi{(1 - e^{2})}^{2}}.$$ Then we match the above polar annual mean insolation to the annual mean insolation at the edge of the polar region (i.e., $|\\phi| = \\frac{\\pi}{2} - \\varepsilon$) to obtain the annual mean insolation at the polar regions$${\\langle W\\rangle}_{p} \\approx \\frac{S_{0}}{\\pi{(1 - e^{2})}^{2}}\\left( \\sin\\varepsilon + \\frac{\\cos^{2}\\phi}{4\\sin\\varepsilon} \\right).$$\n\n### 2.3.2. Large obliquity -- the case of Enceladus {#se0060}\n\nThe approximation we developed above is based on the assumption that the obliquity, *\u03b5*, is small. This situation, however, is not the case, for example, for Enceladus, the moon of Saturn, whose obliquity with respect to the orbit around the Sun is 27^\u2218^. Enceladus is also a moon with an underlying ocean ([@br0380]), and, like Europa, is one of the most probable places in the solar system to find extraterrestrial life. Surprisingly, our approximation for the low and mid-latitudes \\[Eq. [(18)](#fm0180){ref-type=\"disp-formula\"}\\] holds also for the relatively large obliquity of Enceladus, while the solar radiation in the polar regions is better approximated using$${\\langle W\\rangle}_{p} \\approx \\frac{S_{0}}{\\pi{(1 - e^{2})}^{2}}\\left( \\sin\\varepsilon + \\frac{\\cos^{2}\\phi}{2\\pi\\sin\\varepsilon} \\right).$$ [Fig. 3](#fg0030){ref-type=\"fig\"} depicts the numerically integrated and analytically approximated annual mean \\[Eq. [(22)](#fm0220){ref-type=\"disp-formula\"}\\] solar radiation for Enceladus; for reference, we also include the solar radiation during the solstice.Figure 3The incoming solar radiation as a function of latitude for Enceladus, the moon of Saturn. Both the numerically integrated and the analytic approximation are shown, and the two are almost indistinguishable except for the region around latitudes \u2212*\u03c0*/2\u202f+\u202f*\u03b5* and *\u03c0*/2\u202f\u2212\u202f*\u03b5* (63^\u2218^S and 63^\u2218^N), at the transition to the polar regions. The parameter values we used to calculate the incoming solar radiation to Enceladus are: *S*~0~\u202f=\u202f15.14 W\u202fm^\u22122^, *e*\u202f=\u202f0.055, *\u03b5*\u202f=\u202f27^\u2218^, and *\u03b1*~*p*~\u202f=\u202f0.8. For simplicity, to estimate the incoming solar radiation during the solstice, we assume that *e*\u202f=\u202f0.Figure 3\n\n2.4. The effect of Jupiter\\'s longwave radiation {#se0070}\n------------------------------------------------\n\nJupiter emits longwave radiation that may affect the surface temperature of Europa. Europa is phased-locked to Jupiter such that one side of Europa always absorbs the longwave radiation of Jupiter while the other side never absorbs this radiation. Below, we show that Jupiter\\'s longwave radiation only slightly affects the surface temperature of Europa, mainly at the low latitudes -- the effect of this radiation is very small at the high latitudes of Europa. For simplicity, we assume that: (i) the axes of rotation of Jupiter and Europa are parallel (in reality, the difference between the two is about $0.5^{\\circ}$), (ii) the eccentricity of Europa around Jupiter is 0 (where in reality is 0.009), (iii) Europa absorbs all longwave radiation of Jupiter reaching its surface without any reflection, (iv) the longwave radiation of Jupiter is distributed evenly during Europa\\'s diurnal cycle (i.e., the Jupiter-facing side of Europa receives as much radiation as the side that never faces Jupiter), and (v) ignore the spatial and temporal variations in the emission temperature (and hence longwave radiation) of Jupiter. Given the assumptions above (especially assumption iv), the analysis below is not exact. We note that the justification for assumption (iv) is our aim to calculate the zonal mean temperature.\n\nConsider [Fig. 4](#fg0040){ref-type=\"fig\"} below. One can view the total radiation emitted from Jupiter as emitted from a point source concentrated at the center of Jupiter; this assumption is valid if one assumes that the radiation emitted from Jupiter\\'s surface is perpendicular to the surface. This \"source\" emits radiation that propagates spherically. The radiation decreases as $1/r^{2}$ where *r* is the distance from the center of Jupiter. This radiation at Europa is $J_{0} = (4\\pi a_{j}^{2}\\sigma T_{j}^{4})/(4\\pi r_{e}^{2}) = 0.176\\ \\text{W}\\ \\text{m}^{- 2}$. Here, $a_{j}$ is the radius of Jupiter, *\u03c3* is the Stephan-Boltzmann constant, $T_{j} = 130\\ \\text{K}$ is the emission temperature of Jupiter ([@br0210]), and $r_{e}$ is the distance between Jupiter and Europa. The maximum of Jupiter\\'s longwave radiation that is absorbed by Europa at a certain latitude $\\phi_{e}$ is $J_{0}\\cos(\\phi_{e} + \\beta)$, where *\u03b2* is indicated in [Fig. 4](#fg0040){ref-type=\"fig\"}. Then, by calculating the average Jovian radiation over all longitudes at a given latitude, the mean Jovian radiation over a diurnal cycle is $J_{0}\\cos(\\phi_{e} + \\beta)/\\pi$ where *\u03b2* can be approximated as $\\tan\\beta = \\frac{a_{e}\\sin\\phi_{e}}{r_{e} - a_{e}\\cos\\phi_{e}} \\approx \\frac{a_{e}}{r_{e}}\\sin\\phi_{e} \\approx 0$ since $r_{e} \\approx 430a_{e}$. Thus, the daily mean radiation absorbed by Europa at a certain latitude is, to a good approximation,$$W_{j} = \\frac{J_{0}}{\\pi}\\cos\\phi_{e}.$$ It follows that this radiation is zero at the poles and maximal at the equator where it is comparable ($0.056\\ \\text{W}\\ \\text{m}^{- 2}$) to the contribution of the internal heating *Q*. Since the daily mean absorbed solar radiation at the equator is $W(1 - \\alpha_{p})(1 - p) \\approx 5\\ \\text{W}\\ \\text{m}^{- 2}$, the effect of Jupiter\\'s longwave radiation on the surface temperature of Europa is small.Figure 4A drawing showing the longwave radiation emitted from a point source of Jupiter, reaching Europa. The relative dimensions are not realistic as the distance between Jupiter and Europa is 10 times the radius of Jupiter and the radius of Europa is 45 times smaller than the radius of Jupiter. The inset shows a scaled configuration of Jupiter (black circle) and Europa (red dot that is indicated by the arrow).Figure 4\n\n2.5. The effect of Europa\\'s eclipse {#se0080}\n------------------------------------\n\nSince the obliquity of both Jupiter and Europa is very small and since Europa is close to Jupiter (the distance between Europa and Jupiter is only \u223c10 times the radius of Jupiter) but much smaller than Jupiter (the radius of Europa is about 45 times smaller than that of Jupiter), Europa passes in the shadow of Jupiter during each of Europa\\'s days. Since Europa is phased-locked to Jupiter, only the side of Europa that faces Jupiter will experience the eclipse -- below, when simulating the diurnal cycle in temperature, we ignore this fact and assume that the eclipse reduces the overall absorbed daily solar radiation by some factor. The relative time that Europa is within the shadow of Jupiter (eclipse conditions) can be approximated as the ratio between the diameter of Jupiter and the perimeter of Europa\\'s orbit around Jupiter; it is $p = 2a_{j}/2\\pi r_{e} \\approx 0.033$, where $a_{j}$ is the radius of Jupiter and $r_{e}$ is the distance of Europa from Jupiter. Here we consider only the umbra effect and ignore the penumbra effect since (i) Jupiter is relatively very far from the Sun and since (ii) Europa is relatively close to Jupiter. See Section [2.6](#se0090){ref-type=\"sec\"} for a more accurate estimation of the eclipse effect.\n\nIt is possible to approximate the decrease in surface temperature at the end of the eclipse by considering only the internal heating and outgoing longwave radiation, using a time-dependent energy balance equation. In this case, the decrease in temperature is primarily controlled by the heat capacity of the surface ice and much less by the internal heating. The decrease in temperature is then less than 0.3 K.\n\n2.6. More accurate estimation for the effect of the eclipse {#se0090}\n-----------------------------------------------------------\n\nBelow we provide the details regarding the more accurate estimation for the effect of the eclipse. The reader that is not interested in the mathematical details may skip this section.\n\nA more accurate estimation is based on [Fig. 5](#fg0050){ref-type=\"fig\"}. The term $a_{s,j,e}$ indicates the radius of the Sun, Jupiter, and Europa, $r_{p}$ is the distance between the Sun and point *p*, $r_{j}$ is the distance of Jupiter from the Sun, $r_{e}$ is the distance between Europa and Jupiter, *d* is the segment of the perimeter of Europa around Jupiter that is in Jupiter\\'s shadow, and 2*\u03b1* is the angle between two rays from the Sun that intercept at point *p*. Thus,$$\\frac{d}{2} = a_{s}\\left( 1 - \\frac{r_{j}}{r_{p}} - \\frac{r_{e}}{r_{p}} \\right)$$$$\\tan\\alpha = \\frac{a_{j}}{r_{p} - r_{j}}$$$$r_{p} = \\frac{r_{j}a_{s}}{a_{s} - a_{j}},$$ which implies that$$d = 2a_{j} - 2r_{e}(a_{s} - a_{j})/r_{j}.$$ The relative time that Europa is in the shadow of Jupiter is approximately$$p = \\frac{d - a_{e}}{2\\pi r_{e}} = \\frac{a_{j} - r_{e}(a_{s} - a_{j})/r_{j} - a_{e}/2}{\\pi r_{e}}.$$ For the parameter values listed in [Table 1](#tbl0010){ref-type=\"table\"}, $d \\approx 2a_{j}$ and $p \\approx a_{j}/(\\pi r_{e}) = 0.033$, as indicated in Section [2.5](#se0080){ref-type=\"sec\"}.Figure 5A drawing depicting the different measures that are used to calculate the time that Europa spends in Jupiter\\'s shadow. The different measures do not reflect the real relative measures.Figure 5\n\n3. Model {#se0190}\n========\n\n3.1. Calculation of the surface temperature of Europa {#se0100}\n-----------------------------------------------------\n\nWe first focus on the diurnal cycle of the surface temperature. Here one has to take into account the heat capacity of the surface of Europa that leads to absorption of solar radiation only during the day and emission of longwave radiation during both day and night. For this purpose, it is possible to construct a simple diffusion equation for ice temperature, *T*, that is forced from below by the internal heat of Europa and from above by the incoming solar (and Jupiter) radiation and outgoing longwave radiation:$$\\frac{\\partial T}{\\partial t} = \\kappa_{s}\\frac{\\partial^{2}T}{\\partial z^{2}}$$ where $\\kappa_{s}$ is the surface ice temperature diffusion coefficient. We note that the surface ice temperature diffusion constant $\\kappa_{s}$ is 2--3 orders of magnitude smaller than the value of Earth; this follows previous studies ([@br0370], [@br0300]) that estimated the thermal inertia of the surface ice of Europa to be tens of times smaller than the value of Earth; the thermal inertia is $\\Gamma = \\sqrt{k_{s}\\rho_{I}c_{p,I}}$ where $k_{s}$ is the thermal conductivity constant which is related to the ice surface diffusion coefficient $\\kappa_{s}$ through $k_{s} = \\rho_{I}c_{p,I}\\kappa_{s}$. The value of $\\kappa_{s}$ was chosen to fit the diurnal variations of surface temperature reported in [@br0300]. The surface boundary condition is$$\\rho_{I}c_{p,I}\\kappa_{s}\\frac{\\partial T_{s}}{\\partial z} = W(1 - \\alpha_{p})(1 - p) + W_{j} - \\epsilon\\sigma T_{s}^{4},$$ where $\\rho_{I}$ is the density of the ice, $c_{p,I}$ is the heat capacity of the ice, $T_{s}(t) = T(z = 0,t)$ is Europa\\'s surface temperature, *W* is the diurnal cycle of the insolation \\[Eq. [(1)](#fm0010){ref-type=\"disp-formula\"}\\], $W_{j}$ is the daily mean absorbed Jupiter\\'s longwave radiation \\[Eq. [(23)](#fm0230){ref-type=\"disp-formula\"}\\], $\\alpha_{p}$ is the planetary albedo of Europa, *p* is the relative time that Europa passes through Jupiter\\'s shadow, and *\u03f5* is the emissivity of Europa ([@br0360]). The boundary condition at the bottom of the surface ice layer is$$\\rho_{I}c_{p,I}\\kappa_{s}\\frac{\\partial T}{\\partial z} = - Q,$$ meaning that the internal heating rate *Q* is proportional to the temperature gradient deep enough within the ice.\n\nUnder the assumptions that Europa has no atmosphere and that there is no heat capacity at Europa\\'s surface (i.e., $c_{p,I} = 0$), it is possible to calculate the daily mean surface temperature of Europa based on the energy balance between the incoming heat fluxes (i.e., shortwave solar radiation, incoming longwave radiation of Jupiter, and internal heating) and the outgoing longwave radiation. More specifically,$$W(1 - \\alpha_{p})(1 - p) + W_{j} + Q = \\epsilon\\sigma T_{s}^{4},$$ where *W* is the daily mean insolation \\[Eqs. [(6)](#fm0060){ref-type=\"disp-formula\"}, [(7)](#fm0070){ref-type=\"disp-formula\"}\\], and the other parameters are explained above. Consequently, Europa\\'s surface temperature can be expressed as$$T_{s}(\\phi) = \\left\\lbrack \\frac{W(1 - \\alpha_{p})(1 - p) + W_{j} + Q}{\\epsilon\\sigma} \\right\\rbrack^{\\frac{1}{4}}.$$ The annual mean surface temperature, $\\langle T_{s}\\rangle$, can be calculated using Eqs. [(33)](#fm0330){ref-type=\"disp-formula\"}, [(12)](#fm0120){ref-type=\"disp-formula\"}, [(13)](#fm0130){ref-type=\"disp-formula\"} and is$$\\langle T_{s}\\rangle = \\frac{1}{\\tau}\\int\\limits_{0}^{2\\pi}T_{s}\\frac{dt}{d\\lambda}d\\lambda.$$ All the parameters in Eq. [(33)](#fm0330){ref-type=\"disp-formula\"} are well constrained except the internal heating *Q*. The internal heating, most probably, has spatial dependence, due to, for example, tidal heating within the ice. However, due to a lack of knowledge and large uncertainties, we assume that it is spatially constant. In addition, [@br0390] indicated that the heat flux at Europa\\'s surface is almost constant. \\[We note that it is easily possible to use spatially variable internal heating and surface albedo in Eq. [(33)](#fm0330){ref-type=\"disp-formula\"}.\\]\n\nA more common way to estimate the surface temperature is using the annual mean insolation, i.e.,$${\\langle T_{s}^{4}\\rangle}^{1/4} = \\left\\lbrack \\frac{\\langle W\\rangle(1 - \\alpha_{p})(1 - p) + W_{j} + Q}{\\epsilon\\sigma} \\right\\rbrack^{\\frac{1}{4}},$$ where $\\langle W\\rangle$ is the annual mean insolation given in Eq. [(18)](#fm0180){ref-type=\"disp-formula\"}. Such an approach was taken by [@br0260]. Still, this is a non-trivial assumption as $\\langle T_{s}\\rangle \\neq {\\langle T_{s}^{4}\\rangle}^{1/4}$. We discuss the validity of Eq. [(35)](#fm0350){ref-type=\"disp-formula\"} below.\n\nTo find the surface temperature, it is necessary to numerically integrate Eq. [(29)](#fm0290){ref-type=\"disp-formula\"} with respect to depth and time using the surface and bottom boundary conditions \\[Eqs. [(30)](#fm0300){ref-type=\"disp-formula\"}, [(31)](#fm0310){ref-type=\"disp-formula\"}\\]. The surface layer is subjected to diurnally and seasonally varying incoming solar radiation where a Jupiter year is equivalent to \u22481220 Europa days. It is thus necessary to use a sufficiently fine temporal and vertical resolution with a long enough integration time.\n\nWe used 3500 vertical levels with a resolution of 0.87 mm, covering a depth of 3.04 m. The effect of the diurnal surface oscillations (due to the diurnal variations of solar radiation) decay within the upper 5 cm while the seasonal surface variations decay within the upper 1 m. The integration time step (in terms of hour angle) is $10^{- 3}\\ \\text{rad}$. The initial conditions were uniform temperature (60 K) with depth, and the model \\[Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}\\] was integrated for 150 Jupiter years -- this integration time was long enough to achieve convergence.\n\nWe first discuss the diurnal cycle of the surface temperature of Europa. The results are summarized in [Fig. 1](#fg0010){ref-type=\"fig\"}. In [Fig. 1](#fg0010){ref-type=\"fig\"}b, we present the diurnal surface temperature versus time at the equator, at the NH vernal equinox and NH summer solstice. Consistent with the incoming solar radiation presented in [Fig. 1](#fg0010){ref-type=\"fig\"}a, the temperature during the NH summer solstice is higher than the NH vernal equinox temperature by a few degrees. The temperature peaks a few hours after the maximum in radiation due to the heat capacity of the ice. The difference between the minimal and maximal temperatures spans about 40 K, consistent with previous studies ([@br0370], [@br0300]). The ice surface temperature diffusion constant, $\\kappa_{s}$, plays an important role in the range of variations -- a smaller constant would permit larger variations while a larger value that is similar to Earth\\'s ice diffusion constant would yield much smaller variations. The relatively fast warming followed by slower cooling is also apparent and similar to that of [@br0370] and [@br0300]. \\[Yet, we note that these studies did not use a temperature diffusion equation to estimate the surface, temperature, and did not include in their calculations all the factors considered here; they only concentrated on the diurnal cycle of the surface temperature and did not calculate the seasonal cycle and annual mean temperature.\\] The daily mean values are also plotted in [Fig. 1](#fg0010){ref-type=\"fig\"}b where the dotted horizontal lines indicate the mean of the solid (numerically calculated) curves while the dashed horizontal lines depict the daily mean temperature calculated based on the daily mean incoming solar radiation \\[Eqs. [(6)](#fm0060){ref-type=\"disp-formula\"}, [(33)](#fm0330){ref-type=\"disp-formula\"}\\]. The difference between these two daily mean temperature estimations is about 3 K, small compared to the diurnal variations in temperature.\n\nThe temperature profile within the surface layer of the ice is presented in [Fig. 1](#fg0010){ref-type=\"fig\"}c. The temperature fluctuations decay very fast with depth, and the different temporal temperature profiles converge to a single profile at a depth of about 5 cm. This is due to the small surface ice temperature diffusion constant, $\\kappa_{s}$, where higher values of $\\kappa_{s}$ would allow deeper convergence together with smaller diurnal fluctuations. The temperature gradient below the fluctuation level is influenced by both the seasonal cycle and the endogenic heating.\n\nWe repeated the surface temperature estimation at the SH summer solstice close to the pole (89^\u2218^S) in [Fig. 1](#fg0010){ref-type=\"fig\"}d. Here the temperature variations are much smaller than the equatorial ones presented in [Fig. 1](#fg0010){ref-type=\"fig\"}b, consistent with the much smaller incoming solar radiation presented in [Fig. 1](#fg0010){ref-type=\"fig\"}a. The difference between the daily mean temperature calculated based on the numerically calculated temperature (dotted line) and the daily mean radiation energy-balance-based temperature (dashed line) is about 1.5 K.\n\nWe now switch to the seasonal cycle of the surface temperature. In [Fig. 2](#fg0020){ref-type=\"fig\"}b, we plot the equatorial (blue) and polar (red) daily mean surface temperature as a function of time. These curves are based on the numerically calculated daily mean temperatures (dotted line in [Fig. 1](#fg0010){ref-type=\"fig\"}b,d). At the equator, the difference between the maximal and minimal temperatures is about 4 K, much lower than the maximum-minimum difference at the pole (\u223c30 K). This is expected due to the relatively large variations in the incoming solar radiation at the high latitudes across the year. The minimum and maximal equatorial temperatures occur at the aphelion and perihelion, close to the equinox, and consistent with [1](#fg0010){ref-type=\"fig\"}b, as at the present day, Jupiter is farther from (or closer to) the Sun during the NH vernal (NH autumnal) equinox, and this eccentricity effect overcomes the effect of obliquity, which is relatively small. Following the above, the equatorial solstice temperature is close to the annual mean equatorial temperature; see also [Fig. 2](#fg0020){ref-type=\"fig\"}c,d. At the poles, as expected, the temperature peaks close to the summer solstice and is minimal at the vernal equinox. There is a change in the rate of cooling starting at the autumnal equinox, when sunlight does not reach the winter pole.\n\nThe NH summer solstice temperature as a function of latitude for several internal heating rates *Q* is plotted in [Fig. 2](#fg0020){ref-type=\"fig\"}c. It is clear that the internal heating rate more drastically affects the surface temperature at the poles, and especially at the winter pole where the incoming solar radiation is minimal; the difference between the temperature of the different heating rates exceeds 2 K. Naturally, as the internal heating is the only source of radiation at the winter pole, the temperature difference due to the internal heating is larger than that of the summer pole. The solstice temperatures at the winter/summer poles vary by $\\sim 3\\ \\text{K}/ \\sim 1\\ \\text{K}$ for a internal heat difference of $0.05\\ \\text{W}\\ \\text{m}^{- 2}$. We note that even in the absence of internal heating ($Q = 0$) and the absence of solar radiation, the solstice winter polar temperature is much greater than zero, due to the heat capacity of the ice. Yet, if the heat capacity of the ice is neglected (as in [@br0260]), the temperature should drop to zero at the winter pole region; still, in this case, the annual mean polar temperature is not zero as the polar regions absorb solar radiation during the summer.\n\nThe numerically calculated annual mean temperature for several internal heating rates *Q* is plotted in [Fig. 2](#fg0020){ref-type=\"fig\"}d. The equator to pole temperature difference is about 50 K, depending on the internal heat rate. As expected, the poles are more drastically affected by the internal heating rate, about 2 K for a difference in internal heating of $0.05\\ \\text{W}\\ \\text{m}^{- 2}$.\n\nThe internal heating in Europa is not well constrained. In addition, it may vary spatially due to tidal heating within the ice, in both the meridional and longitudinal directions (e.g., [@br0260], [@br0390], [@br0240]). There are several sources that contribute to this heating: rocky mantle (metallic core and silicate mantle) radiogenic heating ($6\\text{\u2013}8\\ \\text{mW}\\ \\text{m}^{- 2}$, [@br0030]), tidal heating of Europa\\'s core (e.g., $30\\text{\u2013}230\\ \\text{mW}\\ \\text{m}^{- 2}$, [@br0080]), and tidal heating of the icy shell ([@br0390]). The tidal heating of Europa\\'s ocean is negligible ([@br0080]). In addition, when the ice is sufficiently shallow (typically, shallower than 10 km), it has only one conductive layer without a bottom convective layer, and the tidal heating of the icy shell is relatively small ([@br0240], [@br0030]). Here we assume that the icy shell is only conductive such that the internal heating originates in Europa\\'s rocky mantle. Yet, it is easy to calculate the surface temperature at least when the latitudinal dependence of the tidal heating is specified.\n\nIn [Fig. 6](#fg0060){ref-type=\"fig\"} (solid lines), we plot the global mean (weighted by the cosine of the latitude), minimum (NH polar), and maximum (equatorial) annual mean surface temperatures as a function of the internal heating, *Q*. The mean and maximal (equator) surface temperatures are not drastically affected by the internal heating as the main heating source is the incoming solar radiation; the mean and maximal surface temperatures increase by \u223c1 K for an increase of internal heating from 0 to $0.2\\ \\text{W}\\ \\text{m}^{- 2}$. However, the polar temperature is more drastically affected by the internal heating as the solar radiation is very weak in these locations. As shown in [Fig. 2](#fg0020){ref-type=\"fig\"}c, this increase is much larger at the poles during the (winter) solstice. [Fig. 6](#fg0060){ref-type=\"fig\"} also depicts the temperatures that are based on the annual mean radiation \\[Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"}, [(21)](#fm0210){ref-type=\"disp-formula\"}, [(35)](#fm0350){ref-type=\"disp-formula\"}\\] (dashed lines). These annual mean temperature estimations that are based on an energy balance consideration are higher by a few degrees than the numerically calculated temperatures, probably since the heat capacity of the ice moderates the seasonal variations in temperatures and since the outgoing radiation is proportional to $T_{s}^{4}$, giving more weight to higher temperatures.Figure 6(a) The annual mean global mean (blue) and maximum (equator, yellow) of Europa\\'s surface temperature (in K) as a function of the internal heating rate, *Q* (in W\u202fm^\u22122^). The solid lines indicate temperatures that are based on the numerical solution of Eq. [(29)](#fm0290){ref-type=\"disp-formula\"}, while the dashed lines indicate those that are based on the annual mean radiation and energy balance equation \\[Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"}, [(21)](#fm0210){ref-type=\"disp-formula\"}, [(35)](#fm0350){ref-type=\"disp-formula\"}\\]. (b) Same as (a) for the minimum (NH polar) surface temperature.Figure 6\n\nThe sensitivity of the annual mean surface temperature with regards to the different parameters is summarized in [Table 2](#tbl0020){ref-type=\"table\"}. The equator, NH pole, global mean, and low latitude (15^\u2218^S--15^\u2218^N) annual mean temperatures are given for control values and for specific parameter values that are different from the control values. The parameters include the eccentricity, *e*, obliquity, *\u03b5*, emissivity, *\u03f5*, surface albedo, $\\alpha_{p}$, relative time of Europa\\'s eclipse, *p* (as estimated in Section [2.5](#se0080){ref-type=\"sec\"}), surface ice heat diffusion, $\\kappa_{s}$, longwave radiation from Jupiter $J_{0}$, and internal heating, *Q*.\n\nAs shown in [Figs. 2](#fg0020){ref-type=\"fig\"} and [6](#fg0060){ref-type=\"fig\"}, the internal heating more drastically affects the polar regions, as there the incoming solar radiation heating is relatively small compared to the low latitudes. The effect of the eccentricity is small. This also follows from Eqs. [(18)](#fm0180){ref-type=\"disp-formula\"}, [(21)](#fm0210){ref-type=\"disp-formula\"} from which one can see that the annual mean incoming solar radiation is approximately proportional to $1 + 2e^{2}$ since $e \\ll 1$. Thus, the eccentricity increases the annual mean incoming solar radiation, but since for Europa, $e \\approx 0.05$, this increase does not exceed 0.5%. The obliquity, *\u03b5*, has a significant effect on the high latitudes and thus must be taken into account. We also study the case of maximum obliquity ($\\varepsilon = 3.7^{\\circ}$) in comparison to present day obliquity of the control case ($\\varepsilon = 3^{\\circ}$), and it is apparent that only the polar region is affected by this parameter, by about $3^{\\circ}$K. When ignoring the effect of ice emissivity, i.e., taking $\\epsilon = 1$, the surface temperature decreases by more than 1 K. The effect of Europa\\'s eclipse on surface temperatures reduces the incoming solar radiation by up to 3.3%, and the associated drop in surface temperature does not exceed 1 K.\n\nThe effect of the surface ice heat diffusion, $\\kappa_{s}$, on the annual mean temperatures is large as a much larger value of $\\kappa_{s}$ yielded surface temperatures that are higher by \u223c3 K compared to the control temperature values. Moreover, the diurnal cycle and, at the high latitudes, also the seasonal cycle in the surface temperatures are drastically affected by this parameter, exhibiting much smaller variations. More specifically, at the equator, the range of the diurnal variation reduces to about 7 K (compared to 40 K for the control case) while the range of seasonal variations remains 4 K. At the 89^\u2218^, the diurnal cycle variation is less than 0.2 K (compared to 2 K for the control case) and the seasonal cycle variation is 12 K (compare to 30 K for the control case); see [Figs. 1](#fg0010){ref-type=\"fig\"}, [2](#fg0020){ref-type=\"fig\"}.\n\nWe also examine the effect of the longwave radiation of Jupiter that is absorbed by Europa by setting $J_{0} = 0$. It is clear that the contribution of Jupiter\\'s longwave radiation on Europa\\'s surface temperature is small as the drop in temperature compared to the control case is less than 0.3 K. Thus, this effect can be neglected. [Table 2](#tbl0020){ref-type=\"table\"} also includes cases of no ($Q = 0$) and increased internal heating ($Q = 0.2\\ \\text{W}\\ \\text{m}^{- 2}$). Consistent with the results presented above, the internal heating mainly affects the high latitudes.\n\nIn [Table 3](#tbl0030){ref-type=\"table\"}, we present the annual mean surface temperatures under the assumption of an energy balance between the incoming solar radiation, internal heating, incoming longwave radiation of Jupiter, and emitted longwave radiation of Europa \\[Eq. [(35)](#fm0350){ref-type=\"disp-formula\"}\\]. The setup of this table is similar to [Table 2](#tbl0020){ref-type=\"table\"}. Generally speaking, the surface temperatures that are based on the energy balance assumption are warmer than the numerically calculated ones by several degrees, especially at the poles. It is interesting to compare the case of the increased heat diffusion in ice coefficient (increased $\\kappa_{s}$ in [Table 2](#tbl0020){ref-type=\"table\"}) to the control case of the energy balance calculation ([Table 3](#tbl0030){ref-type=\"table\"}), as very high $\\kappa_{s}$ leads to much smaller diurnal and seasonal variations, which leads to a temporally constant temperature that is equivalent to the temperature calculated based on energy balance assumptions. Indeed the surface temperature estimations of these two cases are very similar. Thus, when the heat diffusion in ice coefficient (or the thermal inertia) becomes larger, the energy-balance-based temperature estimation becomes better.Table 3Sensitivity of surface temperature (K) of Europa to the different parameters using the energy balance assumption \\[Eq. [(35)](#fm0350){ref-type=\"disp-formula\"}\\]. Control run parameters are as in [Table 2](#tbl0020){ref-type=\"table\"}.Table 3ParameterEquatorPoleGlobal mean15^\u2218^S--15^\u2218^N meancontrol98.949.392.298.6*e*\u202f=\u202f098.849.292.198.5*\u03b5*\u202f=\u202f098.931.192.298.6*\u03b5*\u202f=\u202f3.7^\u2218^99.151.592.498.8*\u03f5*\u202f=\u202f197.448.590.897.1*\u03b1*~*p*~\u202f=\u202f0.5110.454.3102.9110.1*p*\u202f=\u202f099.749.692.999.4*J*~0~\u202f=\u202f098.949.392.298.6*Q*\u202f=\u202f098.647.291.998.3*Q*\u202f=\u202f0.2 W\u202fm^\u22122^99.654.393.199.3\n\nGiven the above, we conclude that an accurate albedo map is essential in estimating the low and mid-latitude surface temperatures of Europa while the obliquity and internal heating rate are essential in accurately determining the temperature at the high latitudes. The surface ice heat diffusion has a profound effect on the range of the variability of the diurnal and seasonal cycles of the surface temperatures. To a first approximation, the effects of the eccentricity, Europa\\'s eclipse, and the effect of longwave radiation of Jupiter may be neglected. It is plausible that the polar regions\\' temperatures will be estimated/measured in the future, similar to the measurement of the low and mid-latitude temperatures that followed the Galileo spacecraft\\'s measurements ([@br0370], [@br0300]). If this occurs, for example during future missions (like Europa Clipper of NASA) that will examine the water plumes over the south pole of Europa ([@br0310], [@br0350]), it will be possible to roughly estimate the internal heating rate based on the surface temperature, especially over the winter pole.\n\nIt is possible to provide a very rough estimate of Europa\\'s ice thickness based on the above; see Section [3.2](#se0110){ref-type=\"sec\"}. This estimate may be a reference thickness for studying the effect of both the vertical and horizontal ice flows due to both tidal heating within the ice and ice flow due to pressure gradients that are associated with variations in ice thickness (see, e.g., [@br0020]).\n\n3.2. Ice thickness {#se0110}\n------------------\n\nBelow we provide the details regarding the rough estimation of the ice thickness of Europa. The reader that is not interested in the details of the ice thickness may skip this section.\n\nIt is possible to obtain a lower limit estimate for Europa\\'s ice thickness based on the energy balance between the incoming solar radiation, the outgoing longwave radiation, and the internal heating. This rough estimate may be a reference thickness for studying the effect of both vertical and horizontal ice flow due to both tidal heating within the ice and the ice flow due to pressure gradients that are associated with variations in ice thickness. Here we assume that (i) the internal heating, *Q*, is uniform in space and time, (ii) the icy shell is conductive but not convective such that the temperature within the ice varies linearly with depth, and (iii) the tidal heating within the ice is negligible. These assumptions are probably valid for relatively shallow ice (shallower than 10 km) when there is only one conductive layer ([@br0390]); when there are two layers, an upper conductive layer and a lower convective layer, the temperature is approximately uniform within the convective layer (where the temperature is close to the melting temperature), and tidal heating is not negligible within this layer.\n\nUnder the above assumptions, it is possible to calculate the mean thickness of the icy shell as follows:$$h = \\rho_{I}c_{p,I}\\kappa\\frac{T_{f} - T_{s}}{Q},$$ where $\\rho_{I}$ is the ice density, $c_{p,I}$ is the heat capacity of the ice, *\u03ba* is the deep ice temperature diffusion constant, $T_{f}$ is the freezing (or melting) temperature of the ice, and *h* is the ice thickness. $T_{s}$ (annual mean energy-balance-based) and *h* depend on latitude and the freezing temperature, and $T_{f}$, depends on the thickness of the ice (through the pressure at the bottom of the ice) and on the salinity of the water as follows ([@br0110], [@br0180]):$$T_{f} = 273.16 + 0.0901 - 0.0575 \\times S - 7.61 \\times 10^{- 8} \\times P_{b},$$ where *S* is the salinity of the ocean water, $P_{b}$ is the pressure (in Pa) at the bottom of the ice (i.e., $P_{b} = g\\rho_{I}h$), and $T_{f}$ is given in K. Thus, $T_{f}$ depends linearly on ice thickness.\n\nThe dependence of the thickness of the ice on latitude is not trivial as ice may flow due to gradients in ice thickness (see, for example, [@br0400], [@br0020]). To bypass this complexity, we only estimate the mean ice thickness by performing a spatial mean on Eq. [(36)](#fm0360){ref-type=\"disp-formula\"}. In addition, since $T_{f}$ depends on *h*, the global mean thickness of the ice can be estimated iteratively as follows:$${\\langle h\\rangle}_{j + 1} = \\rho_{I}c_{p,I}\\kappa\\frac{{\\langle T_{f}\\rangle}_{j} - \\langle T_{s}\\rangle}{Q},$$ where *j* is a counter that indicates the number of the iteration. We start the process from a typical mean freezing temperature (e.g., $T_{f} \\approx 270\\ \\text{K}$), find the mean thickness, and then use it to estimate the new mean freezing temperature. Convergence is achieved after a few iterations.\n\nThe global mean ice thickness and freezing temperature as a function of internal heating are shown in [Fig. 7](#fg0070){ref-type=\"fig\"}. First, as predicted by Eq. [(38)](#fm0380){ref-type=\"disp-formula\"}, the mean ice thickness, $\\langle h\\rangle$, inversely depends on the internal heating, *Q*. Second, the thickness hardly depends on the salinity of the water. Third, the mean freezing temperature converges to a constant value for realistic internal heating values (i.e., $Q > 40\\ \\text{mW}\\ \\text{m}^{- 2}$). Fourth, an increase of salinity by 100 ppt decreases the freezing temperature by 5.75 K; this is a direct consequence of Eq. [(37)](#fm0370){ref-type=\"disp-formula\"}.Figure 7(a) Global mean ice thickness as a function of the internal heating rate, *Q*, when the underlying ocean is fresh (blue) and salty (100 and 200 ppt, green and red). The ice thickness is hardly affected by the salinity of the ocean water. The salinity unit is \"parts per thousand\" (ppt), or grams of salt per kilogram of seawater. (b) Global mean freezing temperature of seawater as a function of the internal heating rate, *Q*, for fresh (blue) and salty (100 and 200 ppt, green and red) water. (c) Ice thickness (in km), *h*, and freezing temperature (in ^\u2218^C), *T*~*f*~, as a function of latitude for internal heating of *Q*\u202f=\u202f0.05,0.1 W\u202fm^\u22122^. As mentioned in the text, this is a very rough estimate of the ice thickness as our underlying assumption is that the ice is stagnant and that the tidal heating is negligible.Figure 7\n\nA mean ice thickness of $\\langle h\\rangle = 10\\ \\text{km}$ was obtained for an internal heating of $Q \\approx 0.05\\ \\text{W}\\ \\text{m}^{- 2}$. Below this value, the estimated mean ice thickness should be regarded as a lower limit, as an almost uniform temperature bottom convective layer may be formed, violating our assumption of a linear increase of temperature within the ice with depth. For a mean ice thickness shallower than $\\sim 10\\ \\text{km}$, we expect one conductive layer, and the estimated mean ice thickness is more accurate.\n\nIt is possible to calculate the ice thickness and the freezing temperature as a function of latitude under the assumption that the ice does not flow and is stagnant. In this case, there is a simple energy balance between the internal heating, the incoming solar radiation and the outgoing longwave radiation at each latitude. [Fig. 7](#fg0070){ref-type=\"fig\"}c depicts the ice thickness as a function of latitude for different internal heating rates, using Eq. [(36)](#fm0360){ref-type=\"disp-formula\"}. As expected, the ice thickness increases poleward and the equator to pole gradient becomes smaller as the internal heating increases. The gradient in thickness is on the order of a few kilometers. The freezing temperature as a function of latitude is shown in [Fig. 7](#fg0070){ref-type=\"fig\"}c. As expected, the freezing temperature is higher for shallower ice, and the thickness increases toward the poles. The typical equator to pole freezing temperature gradient is about 0.2^\u2218^C and is larger for thicker ice and a smaller internal heating rate.\n\n4. Discussion and conclusions {#se0120}\n=============================\n\nThe Galileo mission triggered many studies regarding the moon Europa (see, [@br0290]). One of the observations made by the Galileo spacecraft (Photopolarimeter-Radiometer) was used to measure Europa\\'s surface temperature ([@br0370], [@br0300]). [@br0370] concentrated on low latitude temperatures, which were also relatively high, and consequently suggested either a low local albedo (0.5 compared with $0.68 \\pm 0.14$ of [@br0140]) or a very high local endogenic heating (of $1\\ \\text{W}\\ \\text{m}^{- 2}$). The mean surface temperature measurement of [@br0370] was limited to latitudes equatorward of $\\sim \\pm 70^{\\circ}$, and these approximately correspond to an internal heating of $\\sim 0.05\\ \\text{W}\\ \\text{m}^{- 2}$. The temperature measurements of [@br0370] and [@br0300] are diurnal temperatures -- these helped us to tune the range of diurnal variations through the heat diffusion in ice coefficient. We performed the sensitivity tests of the heat diffusion in ice parameter since this parameter is not well constrained in the upper part of the ice and since these tests helped us to understand the role of the surface heat capacity of the ice. The other parameters are fairly constrained except the internal heating. We further approximated the seasonal and annual mean temperature for a range of internal heating.\n\nWe used two ways to estimate the surface temperature of Europa. The first method is by using a diffusion equation for the top (several meters) layer of the ice. This diffusion equation is subject to top and bottom boundary conditions. The second method is based on an energy balance between the incoming and outgoing radiation. The first method is more accurate as it is able to simulate, fairly reasonably, the diurnal cycle while the second method may be suitable to estimate the annual mean temperature. We have also examined an alternative way, similar to the approach used in previous studies ([@br0370], [@br0300]), where we use a one layer simple ordinary differential equation to estimate the surface temperature of Europa; i.e., we solved the following differential equation: $\\rho_{I}c_{p,I}d\\frac{\\partial T_{s}}{\\partial t} = W(1 - \\alpha_{p})(1 - p) + W_{j} - \\epsilon\\sigma T_{s}{(t)}^{4}$ where $T_{s}$ is the surface temperature and the scale depth is $d = \\sqrt{2\\kappa_{s}/\\Omega_{e}}$. However, this approach yielded much colder polar temperatures in comparison to the more fundamental diffusion equation model. We thus refrained from presenting the results of the single layer model.\n\nFuture measurements of Europa\\'s surface temperature may include the polar regions. Exact polar temperatures may help to better estimate the rate of the internal heating as these regions are only partly affected by the solar radiation, and the contribution of the internal heating in these regions is large, especially during the winter solstice. Polar temperatures have hardly any diurnal variations, thus simplifying the estimation of the diurnal mean temperature. Large uncertainties are associated with the internal heating and the ice depth of Europa, and an exact estimation of the polar region temperatures may significantly reduce these uncertainties ([@br0020]).\n\nIn summary, we discussed the diurnal and seasonal variations of the incoming solar radiation to Europa and developed a mathematical approximation for the global mean incoming solar radiation at Europa\\'s surface (Section [2.3](#se0040){ref-type=\"sec\"}). Based on the incoming solar radiation, we estimated the diurnal, seasonal and annual mean surface temperature in two ways: 1) based on the numerical integration of a temperature diffusion equation and 2) based on the energy balance between the incoming solar radiation, internal heating, Jupiter\\'s radiation, and outgoing longwave radiation. Our estimates take into account the eccentricity of Jupiter, as well as Europa\\'s obliquity, emissivity, eclipse, surface ice heat diffusion, Jupiter\\'s longwave radiation, and internal heating. We showed that the temperature varies moderately at the low latitudes and much more drastically at the high latitudes and that the high latitudes are more drastically affected by the internal heating, especially during the winter solstice. The diurnal and seasonal variations are controlled by the heat diffusion in ice coefficient. For a typical internal heating rate of $0.05\\ \\text{W}\\ \\text{m}^{- 2}$ (e.g., [@br0390], [@br0240]), the equator, pole, and global mean annual surface temperatures are 96 K, 46 K, and 90 K, respectively. These values are not far from the, much simpler, energy-balance-based temperatures (99 K, 49 K, and 92 K, respectively) such that energy balance based temperatures can be used when studying the ice flow and ocean dynamics of Europa. Based on the internal heating rate, we provide a very rough estimate for the mean thickness of Europa\\'s icy shell (Section [3.2](#se0110){ref-type=\"sec\"}). We also estimate the incoming solar radiation to Enceladus, the moon of Saturn (Section [2.3.2](#se0060){ref-type=\"sec\"}). The approach we developed here may be applicable to other moons in the solar system.\n\nDeclarations {#se0130}\n============\n\nAuthor contribution statement {#se0140}\n-----------------------------\n\nYosef Ashkenazy: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.\n\nFunding statement {#se0150}\n-----------------\n\nThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\n\nCompeting interest statement {#se0160}\n----------------------------\n\nThe authors declare no conflict of interest.\n\nAdditional information {#se0170}\n----------------------\n\nNo additional information is available for this paper.\n\nWe thank Roiy Sayag and Eli Tziperman for helpful discussions.\n"}
+{"text": "Introduction\n============\n\nSubstance Use Among Indigenous Adolescents\n------------------------------------------\n\nIndigenous people have the oldest continuing cultures in the world \\[[@ref1],[@ref2]\\]. A common experience among Indigenous people is the lasting impact of colonization, which continues to impact the health and well-being of many Indigenous people today \\[[@ref3],[@ref4]\\]. Combined with lower outcomes in many social determinants of health such as education, poorer access to health services, ongoing racism, and housing and employment opportunities, one of the most visible consequences is the increased susceptibility to substance use and related harms experienced by Indigenous adolescents aged 10 to 19 years. This is evident in the lower age of initiation and higher rates of use; For instance, in Australia, substance use initiation among Aboriginal and Torres Strait Islander adolescents is reported to be 2 to 6 years earlier than the national average \\[[@ref5]\\], with some adolescents trying tobacco and petrol sniffing as young as 8 to 10 years old \\[[@ref6],[@ref7]\\]. Additionally, Aboriginal and Torres Strait Islander adolescents are 3 times more likely to report injecting drugs \\[[@ref8]\\]. American Indian and Alaskan Native adolescents, aged 14 to 15 years, are 5 times more likely to report cannabis use and more than twice as likely to report excess alcohol use, compared with their non-Indigenous counterparts \\[[@ref9]\\]. A total 21% of American Indian adolescents living on reserves have tried cannabis in their lifetime, compared with 5% of non-Indigenous adolescents \\[[@ref10]\\]. In Canada, Indigenous adolescents aged 12 to 18 years have been estimated to be twice as likely to report being a current smoker, compared with the non-Indigenous population \\[[@ref11]\\]. Early onset of substance use among Indigenous adolescents has been identified as a risk factor for problematic substance use later in life, as well as other adverse health, social, and family outcomes \\[[@ref12]-[@ref18]\\]. Prevention of adolescent substance use has therefore been identified as a key strategy to improve Indigenous health \\[[@ref3],[@ref19]\\].\n\nPotential of Web-Based Substance Use Prevention Programs\n--------------------------------------------------------\n\nA number of prevention strategies have been developed and evaluated with the aim of delaying and reducing adolescent substance use and preventing associated harms. For mainstream populations, school-based prevention programs have been found to be highly effective in reducing the onset and escalation of substance use \\[[@ref20]-[@ref22]\\]. Community-based and family-based approaches show considerable promise of effectiveness \\[[@ref23],[@ref24]\\], whereas mass media campaigns are deemed not effective in improving drug-related knowledge or reducing substance use \\[[@ref24]\\]. Despite the availability of effective prevention strategies, these programs are not widely implemented, with time and lack of resources commonly cited as barriers to implementation \\[[@ref24],[@ref25]\\]. To address these barriers, a number of programs facilitated by computers (including other electronic devices such as tablets or mobile phones) or the Internet have been developed, with promising results in mainstream populations \\[[@ref26]-[@ref30]\\]. Champion et al \\[[@ref31]\\] systematically reviewed 9 randomized controlled trials (RCTs) of computer- and Internet-delivered prevention programs, of which 6 achieved significant benefits for drug and alcohol outcomes. Advantages of computer- and Internet-delivered prevention programs include reduced implementation costs, higher degrees of implementation fidelity, and less need for personnel to deliver the program \\[[@ref30],[@ref31]\\]. Computer- and Internet-delivered prevention programs may be particularly beneficial for disenfranchised populations, such as Indigenous adolescents, because these programs can overcome issues with access, provide engagement opportunities, and have been found to be culturally compatible for Indigenous adolescents \\[[@ref32]-[@ref35]\\]. Moreover, recent research has shown that Internet and technology use is higher or just as high among Indigenous people, compared with non-Indigenous people, and that Indigenous adolescents feel comfortable using technology and expressing themselves on the Internet \\[[@ref35],[@ref36]\\].\n\nNeed for Evidence-Based Prevention for Indigenous Adolescents\n-------------------------------------------------------------\n\nAlthough there is evidence to support computer- and Internet-delivered substance use prevention approaches in mainstream populations, the effectiveness of these programs cannot be assumed for Indigenous populations. Indigenous populations may require a cross-cultural translation of these approaches, mapped against situational contexts including different communication styles, languages, and different understandings of health and identity \\[[@ref37]-[@ref39]\\]. This may involve adaptation of an existing mainstream program to align with cultural identity and practices (culturally adapted programs), or development of programs specifically for the local Indigenous cultural context (culture-based programs) \\[[@ref40]\\]. Although it is generally accepted that prevention programs should have a good cultural fit with the local cultural context, no studies have systematically assessed whether culture-based, culturally adapted, or culturally unadapted programs are most effective. Furthermore, no conclusive evidence currently exists for the effectiveness of substance use prevention approaches for Indigenous adolescents, including evidence about the most effective type, setting, or delivery method. A recent Australian systematic review of substance use prevention for Aboriginal and Torres Strait Islander youth found limited evidence for the effectiveness of the 8 reviewed programs, primarily due to poor evaluation designs \\[[@ref41]\\]. Other previous reviews have not been able to provide a comprehensive synthesis of international evidence regarding effective prevention approaches for Indigenous populations, because they have focused on one substance \\[[@ref42],[@ref43]\\], one program setting \\[[@ref44],[@ref45]\\], or were primarily focused on one country \\[[@ref43]-[@ref46]\\]. This systematic review will address this gap by reviewing the evidence regarding the effectiveness of prevention programs in reducing substance use and related outcomes for Indigenous adolescents in the United States, Canada, Australia, and New Zealand. These four countries were chosen because Indigenous people have a comparable history of colonization and dispossession by English settlers, resulting in predominantly English-speaking culture in which Indigenous people are a minority. In all the four countries, there is an unequal distribution between Indigenous and non-Indigenous people in terms of the distribution of economic, social, and health care resources. Indigenous people are more likely to live under the poverty line and are over-represented in measures of low socioeconomic position \\[[@ref3]\\]. Consequently, Indigenous people in these four countries experience poorer health and social outcomes compared with their non-Indigenous counterparts \\[[@ref3],[@ref4],[@ref47]\\]. Although there are differences between Indigenous Peoples' culture between these countries and within these countries, similarities exist including an ongoing occupation of the ancestral lands; common ancestry of the occupied land; and cultural norms and values such as ancestors, connection to ancestors, country, family and community, the concept of health as being holistic, and spirituality \\[[@ref1]\\].\n\nAims of Literature Review\n-------------------------\n\nFor Indigenous adolescents, this review will investigate the following: (1) the effectiveness of culturally adapted substance use prevention programs compared with culture-based or culturally unadapted programs in reducing substance use and related outcomes; (2) the effectiveness of prevention programs delivered in a school setting, compared with community, family, clinical, or multisetting (ie, school, community and/or family) in reducing substance use and related outcomes; (3) the effectiveness of computer- and Internet-delivered programs, compared with traditional delivery; (4) the elements of effective substance use prevention programs; and (5) the methodological quality of evaluations of substance use prevention programs.\n\nMethods\n=======\n\nProtocol Registration\n---------------------\n\nThe protocol for this systematic review has been registered in the PROSPERO registry of the University of York (registration number: CRD42017081885) and has followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P) guidelines; see [Multimedia Appendix 1](#app1){ref-type=\"supplementary-material\"} \\[[@ref48]\\].\n\nSearch Strategy\n---------------\n\nPeer-reviewed and gray literature databases will be searched. Searches in the following 8 electronic databases of peer-reviewed journals will be conducted: DRUG, Cochrane, Embase, PsycINFO, Medline, ProQuest, Informit, and CINAHL. Searches of the gray literature will be conducted in the 20 databases listed in [Textbox 1](#box1){ref-type=\"boxed-text\"}. These databases were based on searches conducted in previous literature reviews on topics related to the health of Indigenous Peoples \\[[@ref49],[@ref50]\\] and recommendations from experts and University libraries in Australia and the United States. The reference lists of selected studies will be assessed for further relevant publications. The researchers will also solicit publications from researchers in the field.\n\nSearch terms are based on previous systematic literature reviews about Indigenous substance use programs \\[[@ref40],[@ref46]\\], and computer- and Internet-delivered substance use prevention \\[[@ref21],[@ref31]\\]. For the peer-reviewed databases, the search strategy will consist of combinations of keywords related to the participants (\"Aboriginal,\" \"Torres Strait Islander,\" \"Indigenous,\" \"Australia,\" \"New Zealand,\" \"Canada,\" \"United States of America,\" \"Maori,\" \"First Nation,\" \"Inuit,\" \"American Indians,\" \"Alaskan Indians\" OR \"Alaska Native\" and \"youth,\" \"young,\" \"adolescen\\*\" OR \"teen\"), type of intervention (\"evaluat\\*,\" \"effect\\*,\" \"efficacy,\" \"review,\" OR \"trial\" and \"prevention,\" \"intervention\" \"program\" \"educat\\*\"), and substance-related outcomes (\"substance,\" \"drug,\" \"alcohol,\" \"tobacco,\" \"petrol,\" \"cannabis,\" \"kava,\" \"methamphetamine,\" \"MDMA,\" \"inhalant,\" \"marijuana,\" \"amphetamine,\" \"psycho stimulant,\" \"smok\\*,\" \"illicit drug\" OR \"volatile drug\"). [Textbox 2](#box2){ref-type=\"boxed-text\"} outlines the detailed proposed search strategy to be used in Medline. For the gray literature databases, the search strategy will consist of combinations of keywords and/or topic headings related to the participants and the substance-related outcomes.\n\nEligibility Criteria\n--------------------\n\nStudies will be included if they are published in English language and evaluate a substance use prevention program with Indigenous adolescents from the United States, Canada, Australia, and New Zealand as the primary participant group. Studies will be included if they are published between January 1, 1990 and August 31, 2017. This will capture studies conducted in the early days of substance prevention as well as the most recent studies. Studies will be included in the review provided the participants are Indigenous adolescents, or a mixed sample of adolescents and adults, but with adolescents as the primary target group of the program. The World Health Organization defines adolescents as people aged 10 to 19 years \\[[@ref12],[@ref51]\\]. To be eligible, at least 50% of the sample must identify as Indigenous or the study must report a separate analysis for Indigenous participants.\n\nStudies will be included if they evaluated a prevention program. The search will not be limited to randomized controlled trials (RCTs), as previous reviews in Indigenous substance use evaluations have recorded a lack of RCTs being conducted within this population \\[[@ref52]\\]. Evaluation is therefore defined as either comparing an experimental group with a control group (eg, no intervention, education as usual, or an alternate intervention) and/or comparing change in outcomes across two or more time points. Following a previous review in substance use prevention among Indigenous adolescents \\[[@ref46]\\], studies will be included if the evaluated prevention program has one or more of the following aims: (1) reduce substance use; (2) increase knowledge of substances and their effects; (3) change attitudes toward substances; (4) increase substance use resistance skills; (5) delay substance use initiation; and/or (6) reduce intention to use substances. This review will include both computer- and Internet-delivered and traditional (face-to-face) prevention programs. It will include universal (everyone in the population), selective (members who are at risk of alcohol and other drug use), and indicated (individuals experiencing early signs of alcohol and other drug use) prevention programs.\n\nStudy Record Management\n-----------------------\n\nAll publications identified in the search of peer-reviewed databases and relevant publications from the gray literature will be exported into a bibliographic software Endnote (Clarivate Analytics, Philadelphia, PA, USA), including the citation and abstract. Duplicate publications will be removed.\n\n###### Gray literature databases included in search strategy (n=20).\n\nUnited States\n\n-   American Indian Health\n\n-   Arctic Health\n\n-   One Sky Center\n\n-   Turtle Island Native Network\n\n-   SAMHSA's National Registry of Evidence-based Programs and Practices\n\nCanada\n\n-   National Collaborating Centre for Aboriginal Health\n\n-   National Aboriginal Health Organization\n\n-   Indigenous Studies Portal\n\nAustralia\n\n-   The Australian Indigenous Health InfoNet\n\n-   Closing the Gap Clearinghouse\n\n-   Analysis and Policy Observatory\n\nNew Zealand\n\n-   Maori Health\n\nInternational\n\n-   Global Health\n\n-   Prevention Information & Evidence\n\n-   Gray Literature Report\n\n-   PAIS Index\n\n-   Campbell Library\n\n-   Minority Health and Health Equity Archive\n\n-   CrimeSolutions\n\n-   Native Health Databases\n\n###### Search strategy for systematic review of substance use prevention programs for Indigenous adolescents (example: Medline search).\n\n1.  ((substance OR drug OR alcohol OR tobacco OR petrol OR cannabis OR kava OR methamphetamine OR MDMA OR inhalant OR marijuana OR amphetamine OR \"psycho stimulant\" OR smok\\* OR \"illicit drug\" OR \"volatile drug\") AND (evaluat\\* OR effect\\* OR efficacy OR review OR trial) AND ((Indigenous OR Aborigin\\* OR \"Torres Strait\\*\" OR Maor\\* OR \"First Nation\" OR Inuit OR \"American Indian\\*\" OR \"Alaskan Indian\\*\") AND (Austral\\* OR \"New Zealand\\*\" OR Canad\\* OR Americ\\*)) AND (youth OR young OR adolescen\\* OR teen\\*)).mp. AND (educat\\* OR prevent\\* OR interven\\* OR program).m_titl.\n\n2.  limit 1 to yr=\"1990 - 2017\"\n\n\\[mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier\\]\n\n![Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA) flow diagram of search and selection strategy.](resprot_v7i2e38_fig1){#figure1}\n\nSelection Processes\n-------------------\n\n[Figure 1](#figure1){ref-type=\"fig\"} illustrates the steps to be taken in the study selection process. One reviewer (BL) will screen all titles and abstracts of papers identified in the searches and assess their eligibility against the inclusion criteria. A second reviewer (MS) will screen a random selection of 25% of the publications to ensure accuracy in the study selection. Agreement between the two reviewers will be assessed, and where there is disagreement, this will be reconciled in consultation between the two authors.\n\nAfter initial screening, full text copies of the publications identified as potentially relevant will be downloaded and further assessed for their eligibility by the two reviewers (BL and MS). Cohen kappa will be calculated to evaluate the inter-rater agreement between the two reviewers at the full text screening stage. Where there is disagreement, this will be reconciled in discussion between the two authors. If there is no consensus, expert consultation will be sought from the other, more senior authors. The first reviewer (BL) will screen the reference lists of the eligible studies for further publications to be added into the systematic review.\n\nData Extraction\n---------------\n\nOne reviewer (BL) will extract the following data from the publications into Microsoft Excel: authors, year of publication, country, name of the evaluated program, study design, sample size, targeted age, Indigenous status of participants, target substance, geographical area, setting, type of program, intervention strategy, duration and frequency of program, whether booster sessions were provided, follow-up time points, mode of delivery, community and stakeholder involvement in development, program facilitation, language consideration, control group, outcome measures used, overall quantitative and overall qualitative outcomes related to substance use, and harms. Detailed qualitative and quantitative outcomes (substance-use related or other outcomes measured) for each study will be extracted.\n\nSetting of the programs will be classified into school-based, community-based, family-based, clinical, or multi-setting. School-based programs are defined as those implemented either in a classroom setting during school hours, or as an out-of-school activity delivered by the school. Community-based programs are those implemented with groups within the community (ie, adolescents, parents, or whole community). Family-based are the programs targeting the family unit. Clinical settings are based within a health service, such as a community health service or a general practice. Multi-setting programs have a combination of any of the aforementioned settings.\n\nType of programs will be classified following recommendations by Leske et al \\[[@ref40]\\] into culture-based, culturally adapted, or culturally unadapted programs. Culture-based programs are developed to reflect and incorporate the unique cultural values and beliefs of the Indigenous participants. Culturally adapted programs are modified from existing non-Indigenous programs to be more culturally appropriate to the Indigenous participants. Culturally unadapted programs are developed for other cultural groups (eg, European, African American, or Mexican) and delivered to Indigenous participants without modification.\n\nProgram delivery will be identified as computer- and Internet-delivered or traditional. Computer- and Internet-delivered programs are those that are delivered completely or partially using computers or other Web-based technology. Traditional programs are those in which no part of the program is delivered using technology.\n\nAccording to the most commonly measured substance-related outcomes following Lee et al \\[[@ref46]\\], this study will group outcomes as follows: substance-related knowledge, substance use, and attitudes toward substance use. These 3 outcomes capture most substance-related outcomes studied in substance use prevention programs. To identify the elements of beneficial substance use prevention programs for Indigenous adolescents, two reviewers (BL and MS) will extract the program elements from all studies to identify the key elements of prevention programs that lead to beneficial substance-related outcomes for Indigenous adolescents.\n\nData Synthesis\n--------------\n\nA narrative summary will be provided of the outcomes of the included studies. On the basis of previous reviews of evaluation studies in Indigenous populations, we expect the number of studies to be too low and the quality of studies too varied to warrant a quantitative synthesis of the data \\[[@ref52]-[@ref54]\\]. The narrative summary will include a discussion of whether programs are beneficial to prevention of substances among Indigenous adolescents. Programs will be identified as beneficial if there are beneficial effects on more than 50% of substance-related outcomes, studies reporting positive findings on 50% or less of the evaluated outcomes will be classified as mixed, studies reporting negative outcomes will be classified as iatrogenic, and studies without significant outcomes will be classified as null. The narrative summary will discuss the number of beneficial programs for each program type (culture-based, culturally adapted, and culturally unadapted), type of delivery (computer- and Internet-delivered and traditional), and program setting (school, community, family, clinical, and multi-setting). It will also detail the elements used in the beneficial programs and summarize the most commonly implemented elements.\n\nCritical Appraisal of Risk of Bias in Individual Studies\n--------------------------------------------------------\n\nThe methodological quality of both quantitative and qualitative elements of the studies will be assessed. The quality assessment will be conducted by one reviewer (BL), with a second reviewer (MS) appraising a random selection of 25% of the publications to ensure reliable coding. Following previous systematic reviews of prevention programs for Indigenous people \\[[@ref49],[@ref55],[@ref56]\\], this review will assess the quality of quantitative studies using the Dictionary for Effective Public Health Practice Project Quality Assessment Tool for Quantitative Studies \\[[@ref57]\\]. Sections A (selection bias), B (study design), C (confounders), E (data collection), and F (withdrawals and dropouts) of this tool are rated as strong, moderate, or weak to assess possible bias. Section D (blinding) will be excluded from this study because double-blinding is not feasible in school-based or community-based studies \\[[@ref23]\\]. As prescribed, sections G (intervention integrity) and H (analysis appropriateness) will receive a narrative description rather than categorical ratings. Following standard procedures of this tool, each study will receive a summary rating defined as weak (two or more weak scores), moderate (one weak score is given), or strong (no weak scores are given).\n\nThe methodological quality of qualitative study components will be assessed using a modified version of the qualitative tool by Long and Godfrey \\[[@ref58]\\], which has also previously been used in a systematic review of programs for Aboriginal and Torres Strait Islander people in Australia \\[[@ref56]\\]. The adapted version assesses quality in 3 domains related to evaluation: (1) data collection, the need for clear descriptions of the data collection process; (2) analysis and potential research bias, the transparency of the description of data analyses processes, description of researchers' positioning in the study and the interpretation of findings in line with the literature ; and (3) policy and practical implications, assessment of the populations to which the findings are generalizable and implications for policy and practice.\n\nResults\n=======\n\nData analysis is underway and the results of this systematic review are expected to be submitted for publication in 2018.\n\nDiscussion\n==========\n\nThis paper summarizes the protocol for a systematic review of substance use prevention programs for Indigenous adolescents in the United States, Canada, Australia, and New Zealand. The purpose of this review is to synthesize international evidence regarding the effectiveness of substance use prevention programs for Indigenous populations. It will identify the setting in which prevention programs are most effective, the most beneficial delivery and types of programs, and the elements of effective substance use prevention for Indigenous adolescents.\n\nThis review will provide researchers, policy makers, and program developers with up-to-date information about the strength of the international evidence to support the use of substance use prevention approaches among Indigenous adolescents. It will evaluate whether mainstream programs are effective when implemented in culturally unadapted form among Indigenous adolescents, and will assess evidence supporting the effectiveness of culturally adapted mainstream programs and specific culture-based programs. Finally, this review will provide evidence about the potential to use computer- and Internet-delivered prevention approaches among Indigenous populations.\n\nGiven the high rates of technology and Internet use amongst Indigenous adolescents \\[[@ref34],[@ref36]\\] and effectiveness of computer- and Internet-delivered substance use prevention in mainstream populations \\[[@ref30]\\], there is considerable potential for the use of computers and Web-based technology in the delivery of substance use prevention with Indigenous adolescents \\[[@ref32]\\]. This review will inform the development of future computer- and Internet-delivered prevention programs for Indigenous adolescents, which have the potential to be highly advantageous for Indigenous adolescents, due to the sustainable, low cost, and engaging format that is well-aligned to the preferences of adolescents \\[[@ref34],[@ref59]\\].\n\nThis study was funded by the Australian Government Department of Health. The funder was not involved in the development of this protocol.\n\nConflicts of Interest: None declared.\n\nAuthors\\' Contributions: LS, MT, and NN conceptualized the study. MS, BL, and LS developed the study design and protocol. MS wrote the first draft of the manuscript. All authors read, revised, and approved the final manuscript.\n\nPreferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P) checklist.\n\nPRISMA-P\n\n:   Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol\n"}
+{"text": "Background {#Sec1}\n==========\n\nGlioblastoma multiforme (GBM) is one of the most aggressive primary brain tumor. It has been demonstrated that mutations of the IDH1 gene correlate with outcome in patients with malignant glioma and are considered as independent factors for predicting longer overall survival and progression free survival in patients with GBM \\[[@CR1]\\]. In GBM, cancer cells with stem cell-like features, called cancer stem cells (CSCs), were identified \\[[@CR2], [@CR3]\\]. GBM stem-like cells (GSCs) have a high capacity to resist or to adapt to standard therapies which include surgery followed by radiotherapy and chemotherapy \\[[@CR4]--[@CR6]\\], resulting in a poor prognosis with a median survival time of about 14\u00a0months \\[[@CR7], [@CR8]\\]. Thus, the development of efficient strategies targeting these cells \u00a0is urgently needed.\n\nIt has been demonstrated that CSC distinct pools reside within different regions of the same GBM \\[[@CR9]--[@CR11]\\]. More recently, the presence of two GSC populations, one derived from the GBM area called enhanced lesion (GCSCs) and the other one from the brain area adjacent to the tumor margin (PCSCs) that greatly differ in their growth properties and tumor-initiating ability, \u00a0was identified. Indeed, GCSCs and PCSCs possess key neural stem cell features, such as multipotency, clonogenic ability and extensive self-renewal, together with aberrant growth properties \\[[@CR12]\\]. Moreover, the area adjacent to the tumor shows edema, vascular alterations \\[[@CR13], [@CR14]\\], reactive astrocytes and microglia \\[[@CR15], [@CR16]\\] in addition to\u00a0an abnormal gene expression\u00a0 \\[[@CR17], [@CR18]\\]. It has been demonstrated that tumor recurrence occurs in tissue neighboring GBM in approximately 90% of patients, suggesting a growing relevance for this area in translational research \\[[@CR16], [@CR19]\\].\n\nGSCs can be isolated from surgical specimens through mechanical dissociation of the tumor tissue and culture in a serum-free medium. In this conditions GCSC- and PCSC-derived cell clones are able to grow in vitro in aggregates called neurospheres and maintain an undifferentiated state as demonstrated by the expression of stem cell markers. Moreover, when injected in immunosuppressed mice, these cells are able to generate a tumor identical to the original one in terms of antigen expression and histological features, although GCSCs exhibit a higher tumor-initiating ability and clonogenicity when compared with PCSCs \\[[@CR12]\\]. For all the above, GCSCs and PCSCs represent a good model to study glioblastoma response to treatments and, in particular, to highlight the role of neighboring microenvironment in tumor progression. To date the most effective chemotherapies to treat GBM are alkylating agents such as Temozolomide (TMZ), with a good penetration into the blood--brain barrier \\[[@CR20]\\]. Alkylating agents damage DNA by formation of different small and\u00a0bulky adducts with nucleic acid bases. In particular, TMZ acts by delivering a methyl group to purine bases of DNA leading to cell cycle arrest and, eventually, to apoptosis \\[[@CR21]\\]. O6-Methylguanine-DNA-methyltransferase (MGMT) repairs the most cytotoxic lesions generated by TMZ, by removing the methyl adducts from DNA. MGMT promoter methylation, leading to a transcriptional silencing, correlates with improved survival in GBM patients exposed to alkylating agent treatment. Accordingly, expression of MGMT is one of the most robust predictors of the TMZ response in malignant glioma cells \\[[@CR22]--[@CR24]\\]. Epigenetic mechanisms are increasingly recognized as a major factor contributing to pathogenesis of cancer including glioblastoma. Enzymatic modification of histone proteins regulating gene expression are recently being exploited for therapeutic drug targeting. In particular, histone acetylation and deacetylation have been demonstrated to regulate several physiological and pathological cellular processes. Histone acetylation is mediated by histone acetyltransferases (HATs) and generally allows for active gene transcription. Conversely, histone deacetylation is catalyzed by histone deacetylases (HDACs) and favors gene repression. Eighteen distinct HDACs have been identified so far and they are classified into four classes based on their sequence and catalytic activity. HDAC4 is a member of class II and it has been demonstrated to be involved in progression of GBM \\[[@CR25]\\]. In addition, it has been recently \u00a0reported that a great number of non-histone proteins can undergo reversible acetylation by HATs and HDACs. Modifications in this dynamic equilibrium can disturb cell homeostasis and result in a pathological state \\[[@CR26]\\]. Nevertheless, HDAC inhibitors cause acetylation of both histone and non-histone proteins and exert multiple anti-tumoral effects by inducing differentiation, apoptosis, cell cycle arrest, susceptibility to chemotherapy and inhibition of migration and angiogenesis \\[[@CR27]\\].\n\nEmerging evidences demonstrate that some antiepileptic drugs (AEDs) have a transcriptional regulatory activity via HDAC modulation \\[[@CR28]\\]. In particular, it has been shown that Levetiracetam (LEV), a relatively new AED, increases the transcription of HDACs and recruits corepressor complex on MGMT promoter, thus silencing its activity \\[[@CR22]\\].\n\nIn this work, we used the GCSC and the PCSC neurospheres derived from primary GBM of six patients, according to the WHO 2016 classification, and we evaluated the effect of several antineoplastic drugs such as TMZ, Etoposide (ETO), Irinotecan (IRI) and Carboplatin (CARB) on their proliferation. Since our results demonstrated that the chemotherapeutic agent with less efficacy was TMZ, we subsequently investigated the possibility to increase its cytotoxic activity by the concomitant treatment with LEV.\n\nHere we show that LEV enhances the effect of TMZ on GCSCs proliferation (being less effective on PCSCs) by decreasing MGMT expression, promoting HDAC4 nuclear translocation and activating apoptotic pathway. Although further studies are needed to determine the exact mechanism by which LEV sensitizes GBM stem cells to TMZ, these results suggest that the clinical therapeutic efficacy of TMZ in GBM might be enhanced by the combination treatment with LEV.\n\nMaterials and methods {#Sec2}\n=====================\n\nNeurosphere culture {#Sec3}\n-------------------\n\nA total of six pairs of neurospheres derived from both GBM and peritumoral tissue (at a distance\u2009\\<\u20091\u00a0cm from macroscopic tumor border), called Glioblastoma Cancer Stem Cells (GCSCs) and Peritumoral Cancer Stem Cells (PCSCs), respectively, were kindly provided by Vescovi and Binda. According to the classification of human gliomas by WHO \\[[@CR29]\\], we analyzed the GCSC and PCSC neurospheres for the mutational status of IDH1 \\[[@CR30], [@CR31]\\]. Mutational IDH1/2 status and MGMT promoter methylation status were also investigated in all the available tissue samples from which the neurospheres were derived \\[[@CR32]\\].\n\nThe neurospheres were cultured in NeuroCult\u2122 NS-A Proliferation Kit (StemCell Technologies Inc, Vancouver, BC, Canada) supplemented with 20\u00a0ng/mL human recombinant EGF, 10\u00a0ng/ml human recombinant bFGF and 2\u00a0\u00b5g/mL heparin (all from StemCell Technologies Inc.), as described previously \\[[@CR33]\\]. All cell cultures were maintained at 37\u00a0\u00b0C in a 5% CO2 humidified atmosphere.\n\nIDH analysis {#Sec4}\n------------\n\nIDH1 was amplified from 20\u00a0ng genomic DNA with forward primer 5\u2032- ACCAAATGGCACCATACGA-3\u2032 and reverse primer 5\u2032- TTCATACCTTGCTTAATGGGTGT-3\u2032 using conditions described by Balss J. et al. \\[[@CR34]\\]. The resulting PCR products were sequenced using the ABI Prism BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, 4337455). Thirty cycles were performed employing 10\u00a0mM of the sense primers with denaturing at 95\u00a0\u00b0C for 10\u00a0s, annealing at 56\u00a0\u00b0C for 5\u00a0s and extension at 60\u00a0\u00b0C for 240\u00a0s. A second round of sequencing analysis was performed using the antisense primer. Sequences were determined using the automated AB3130xl (Applied Biosystems, CA, USA). Results were analyzed with Chromas Lite software (Technelysium) and Mutation Surveyor software (SoftGenetics).\n\nO6-Methylguanine-DNA methyltransferase (MGMT) promoter methylation analysis {#Sec5}\n---------------------------------------------------------------------------\n\nO6-methylguanine-DNA methyltransferase (MGMT) promoter methylation patterns were studied by methylation-specific PCR using primers specific for methylated and unmethylated DNA \\[[@CR35]\\] on genomic DNA extracted from paraffin-embedded tissue using QIAamp DNA Mini kit (Qiagen). The annealing temperature was 60\u00a0\u00b0C. DNA from normal lymphocytes treated with SssI methyltransferase (New England Biolabs) was used as a positive control for methylated and unmethylated alleles of MGMT. PCR products were separated onto 3% agarose gel, stained with ethidium bromide, and visualized under UV illumination.\n\nChemotherapeutic drugs and adjuvant molecules {#Sec6}\n---------------------------------------------\n\nTemozolomide, ETO and IRI (Sigma-Aldrich, Saint Louis, MO, USA) were dissolved in dimethyl sulfoxide (DMSO) (Sigma-Aldrich) at the concentration of 100\u00a0mM, 30\u00a0mg/mL and 10\u00a0mg/mL, respectively, while CARB and LEV (Sigma-Aldrich) were dissolved in deionized water at the concentration of 10\u00a0mg/mL and 5\u00a0mg/mL, respectively. All the drugs were stored as stock solutions at -80\u00a0\u00b0C and diluted in stem cells culture medium just before use. The final drug concentrations used in the experiments were: TMZ: 250\u00a0\u00b5M; ETO: 10\u00a0\u00b5M; IRI: 10\u00a0\u00b5g/mL and CARB: 10\u00a0\u00b5g/mL; LEV: 40\u00a0\u00b5g/mL.\n\nWhen the cells were treated with both TMZ and LEV, LEV was added 2\u00a0h before TMZ addition and the cultures were stopped after 48\u00a0h. The chemotherapeutic drug concentrations utilized in our experiments are in the higher range of serum peak levels transiently reached in vivo during high-dose chemotherapy.\n\nBrdU cell proliferation assay {#Sec7}\n-----------------------------\n\nGCSCs and PCSCs were seeded at the density of 50.000 cells/well in 96-well plates and incubated overnight. The cells were then treated with the appropriate chemotherapeutic agents, as indicated in the Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} for 48\u00a0h. Finally 10\u00a0\u00b5M BrdU was added to the plates and the cells were incubated overnight. BrdU proliferation assay was performed according to the manufacturer's instructions (Cell Signaling, \\#6813 Danvers, MA, USA).Fig.\u00a01Effects of chemotherapeutic treatments on patient-derived GSC proliferation. GCSC and PCSC neurospheres derived from GBM of six different patients were treated with TMZ (250\u00a0\u00b5M), ETO (10\u00a0\u00b5M), IRI (10\u00a0\u00b5g/mL) and CARB (10\u00a0\u00b5g/mL) for 48\u00a0h. GCSCs and PCSCs proliferation rate was evaluated by BrdU incorporation. **a** The graph represents the distribution values of BrdU incorporation of the all neurospheres analyzed. **b** The graph represents BrdU incorporation of the GCSCs and PCSCs derived from patients \\#1, \\#2 and \\#3, that exhibit a higher capacity of BrdU incorporation. **c** The graph represents BrdU incorporation of the GCSCs and PCSCs derived from patients \\#4, \\#5 and \\#6, that exhibit a lower capacity of BrdU incorporation. K-means algorithm was used to cluster into 2 groups the BrDU incorporation values. **d** Percentage of decreased BrdU incorporation versus CTR. The results shown are representative of three independent experiments. **a**--**c** \\*p\u2009\\<\u20090.05 vs CTR by Mann--Whitney test. **d** \\*\u2009p\\<\u20090.05, \\*\\*\u2009p\\<\u20090.01, \\*\\*\\*p\u2009\\<\u20090.001 vs CTR by Student's *t*-test\n\nWestern blot analysis {#Sec8}\n---------------------\n\nFor immunoblotting analysis, cells were washed in 1\u2009\u00d7\u2009PBS, harvested and lysed in 1x Cell Lysis Buffer (Cell Signaling \\#9803) containing 1\u00a0mM PMSF (Cell Signaling \\#8553) and a complete protease inhibitor cocktail (Cell Signaling \\#5872) for 30\u00a0min at 4\u00a0\u00b0C. Then the cells were sonicated briefly and the extracts were centrifuged 10\u00a0min at 14,000\u2009\u00d7\u2009*g* in a cold microfuge. Protein concentration was determined by Bradford Protein Assay (Bio-Rad Laboratories Inc, Hercules, CA, USA) according to the manufacturer's instructions. Equal amounts of proteins were then separated by SDS/PAGE (Mini-PROTEAN^\u00ae^ TGX\u2122 Precast Protein Gels, or Mini-PROTEAN TGX stain-free precast PAGE gels, Bio-Rad Laboratories Inc.) and transferred to a nitrocellulose membrane (GE Healthcare, Piscataway, NJ, USA). Membranes were blocked with Tris-buffered saline (TBS) 1X (Bio-Rad Laboratories Inc.) supplemented with 0.1% Tween-20 and containing 5% nonfat milk for 1\u00a0h at room temperature (RT). The primary antibodies used in this work \u00a0were: anti-MGMT (1:500, mouse monoclonal antibody, clone MT3.1, MAB16200, Merk Millipore, Darmstadt, Germany); anti-HDAC4 (1:100, rabbit monoclonal antibody, sc-46672 Santa Cruz Biotechnology, Dallas, Tx, USA); anti-PCNA (1:1000, mouse monoclonal antibody, M0879, Dako, Santa Clara, CA, USA); anti-cleaved Caspase-3 (1:1000, polyclonal antibody, \\#9665, Cell Signaling); anti-\u03b2-actin, (1:10000 mouse monoclonal antibody, Sigma-Aldrich). Blots were then incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000, Vector Laboratories, Burlingame, CA, USA) for 1\u00a0h RT. Signals were captured by ChemiDoc\u2122 Imaging System (Bio-Rad Laboratories, Hercules, CA, USA) using an enhanced chemiluminescence system (SuperSignal Chemoluminescent substrate, Thermo Fisher Scientific Inc. Waltham, MA, USA) and densitometric analyses were performed with Image Lab\u2122 Touch Software (Bio-Rad Laboratories). Nuclear and cytosolic fractions were normalized using stain free technology (Bio-Rad Laboratories Inc.). All experiments were carried out in triplicate and representative results are shown.\n\nImmunofluorescence and confocal microscopy analysis {#Sec9}\n---------------------------------------------------\n\nImmunofluorescence analysis was performed on GCSCs and PCSCs collected onto a glass slide using a Cytospin centrifuge (Shandon Centrifuge, Model Cytospin 3, Marshall Scientific, Hampton, NH, USA), fixed with 4% paraformaldehyde for 20\u00a0min, incubated with 0,01% Triton X-100 for 7\u00a0min and blocked with Super Block solution (UCS Diagnostic S.r.l., Morlupo, Italy) for 5\u00a0min. The slides were incubated overnight at 4\u00a0\u00b0C with the primary antibodies against: MGMT (1:100, Merk Millipore), HDAC4 (1:100; Santa Cruz Biotechnology, INC.) and cleaved Caspase-3 (1:400, Cell Signaling). The next day, the slides were incubated with the following secondary antibodies for 1\u00a0h at RT: Alexa Fluor 584 (1:1000, Invitrogen Molecular Probes, Eugene, OR, USA) and Alexa Fluor 488 (1:1000, Invitrogen Molecular Probes). The cells\u00a0were cover-slipped with ProLong Gold antifade reagent with DAPI (Life Technologies) and examined with a confocal laser scanning microscope (TCS-SP2, Leica Microsystems, GmbH, Wetzlar, Germany) equipped with an Ar/ArKr laser and a HeNe lasers. The images were recovered utilizing the Leica Confocal software. Laser line was at 488\u00a0nm and 543 for alexafluor 488 and alexafluor 568 excitation, respectively. For each analyzed field, optical spatial series each composed of about 10 optical sections with a step size of 1\u00a0\u03bcm were obtained. The images were scanned under a 40\u00d7 oil. In each experiment, negative controls without the primary antibody were included to check for nonspecific staining.\n\nStatistical analysis {#Sec10}\n--------------------\n\nEach experiment was repeated three times. Data are presented as the mean\u2009\u00b1\u2009SD. Statistical analysis was\u00a0generally performed using Student's *t*-test, assuming equal variance, and p-values were calculated based on the 2-tailed test. A p-value of\u2009\\<\u20090.05 was considered statistically significant. The Mann-Whitney test was used for the analysis of distribution values of BrdU incorporation in neurospheres.\n\nResults {#Sec11}\n=======\n\nTMZ was the less effective chemotherapeutic agent in decreasing the GCSC and PCSC proliferation {#Sec12}\n-----------------------------------------------------------------------------------------------\n\nThe analysis of the IDH1 status in all the GCSC and PCSC pairs derived from six patients revealed that all of them are IDH1-wildtype. Moreover, the analysis of mutational IDH1/2 status and MGMT promoter methylation status in all the available tissue samples revealed that all of them were IDH1/2 wildtype and showed an un-methylated MGMT status.\n\nWe then investigated the effects of different chemotherapeutic agents on the proliferation of the neurospheres isolated from the GBM derived from six different patients. For this purpose, GCSCs and PCSCs were treated with TMZ, ETO, IRI and CARB for 48\u00a0h and then the rate of proliferating cells was evaluated by BrdU incorporation. At first, we noted that when the data regarding GCSCs and PCSCs derived from all the patients were analyzed together they were distributed in two clusters: one with the higher and one with lower values of BrdU incorporation (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a). Interestingly, all the chemotherapeutic agents exerted significant effects only in the cells with higher values of BrdU incorporation (GCSCs and PCSCs derived from patients \\#1, \\#2 and \\#3) (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b), whereas none of the drugs affected significantly the cells with a lower proliferation rate (GCSCs and PCSCs derived from patients \\#4, \\#5 and \\#6) (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}c). We then evaluated, more specifically, the percentage of proliferative reduction induced by the chemotherapeutic agents when the GCSCs and the PCSCs derived from the GBM of each patient were considered separately (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}d). As expected, both GCSCs and PCSCs obtained from the GBM of all the patients exhibited a higher resistance in terms of proliferation to the chemotherapeutic drug treatments when compared to Jurkat cells that, as hematopoietic cells, represent a good general model to test chemotherapeutic agent activity \\[[@CR36]\\] (Additional file [1](#MOESM1){ref-type=\"media\"} and Additional file [2](#MOESM2){ref-type=\"media\"}: Figure S1). Although no significant differences were observed in the response between GCSCs and PCSCs, our results showed that, similarly to what happens in Jurkat cells, TMZ had the lower effect, as demonstrated by the percentage of decreased proliferative range (8--23% versus CTR), while CARB had the stronger effect since the percentage of decreased proliferative range was 21--45% versus CTR (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}d).\n\nLevetiracetam sensitized GSCs to TMZ treatment {#Sec13}\n----------------------------------------------\n\nSince our previous results demonstrated that, among the tested antineoplastic drugs, TMZ was the less efficient to decrease the proliferation of both GCSCs and PCSCs, we then investigated whether its anti-proliferative effect might be strengthened when combined with the LEV. LEV concentration used in this study (40\u00a0\u03bcg/mL) is included in the clinical serum therapeutic concentration range achieved in patients at oral doses of 500-1000\u00a0mg twice daily \\[[@CR28]\\]. GCSCs and PCSCs isolated from the GBM derived from six patients were treated with TMZ, LEV or with a combination of TMZ and LEV, for 48\u00a0h. Figure\u00a0[2](#Fig2){ref-type=\"fig\"}a shows that, when analyzed together, the GCSCs derived from all the patients displayed a significant reduction of BrdU incorporation only with the combined treatment with TMZ and LEV. This result is confirmed by the analysis of BrdU incorporation evaluated in the GCSCs derived from each patient (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}b), demonstrating that the slight antitumor effect exerted by the treatment with\u00a0TMZ or LEV alone was strongly enhanced when TMZ and LEV were added in combination. Only the GCSCs derived from patient \\#6 did not display a significant reduction in cell proliferation (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}b). Interestingly, the PCSCs subjected to the same treatments (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a, b) seemed to be more resistant than the GCSCs since only the PCSCs derived from patients \\#4 and \\#5 (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}b) showed a significant decrease of proliferation when exposed to the combination of TMZ and LEV compared to either untreated cells or treated with these drugs alone. These results demonstrate that LEV sensitized GBM stem cells to TMZ and that this effect is stronger in GCSCs than in PCSCs.Fig.\u00a02Effect of TMZ and LEV combined treatment on GCSC proliferation. GCSC neurospheres derived from GBM of six different patients were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) or with a combination of TMZ and LEV for 48\u00a0h. Cell proliferation was evaluated by BrdU incorporation. The results were representative of three independent experiments. Statistical analysis was performed using Mann--Whitney test \\*p\u2009\\<\u20090.05 (**a**) and Student's *t*-test (**b**). P value\u2009\\<\u20090.05 was considered statistically significant Fig.\u00a03Effect of TMZ and LEV combined treatment on PCSC proliferation. PCSC neurospheres derived from GBM of six different patients were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) or with a combination of TMZ and LEV for 48\u00a0h. Cell proliferation was evaluated by BrdU incorporation. The results were representative of three independent experiments. Statistical analysis was performed using Mann--Whitney test \\*p\u2009\\<\u20090.05 (**a**) and Student's *t*-test (**b**). P value\u2009\\<\u20090.05 was considered statistically significant\n\nMGMT expression was downregulated in GCSCs treated with LEV\u2009+\u2009TMZ {#Sec14}\n-----------------------------------------------------------------\n\nTo test the hypothesis that LEV-induced sensitization to TMZ might result from the inhibition of MGMT-mediated DNA repair, we investigated MGMT protein levels. This analysis was performed in the GCSCs derived from patient \\#4, which displayed a high reduction of the proliferative rate after exposure to the combined treatment with LEV\u2009+\u2009TMZ and in PCSCs derived from patient \\#6, which showed resistance to all the treatments. The data obtained by BrdU cell proliferation assay were confirmed by Western blot analysis shown in Figs.\u00a0[4](#Fig4){ref-type=\"fig\"}c, d and [5](#Fig5){ref-type=\"fig\"}c, d that demonstrated a decreased expression of the proliferation marker PCNA in the GCSCs of patient \\#4 when LEV and TMZ were added together, while its expression was not affected in the PCSCs derived from patient \\#6. Western blot analysis revealed a high level of MGMT expression in untreated GCSCs derived from patient \\#4; this expression was slightly decreased after treatment with TMZ and LEV singularly but it was dramatically decreased after the combined treatment with TMZ and LEV (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}a, b). In contrast, none of the treatments seemed to modify the MGMT expression level in the PCSCs derived from patient \\#6 when compared with untreated cells (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}a, b). Immunofluorescence analysis confirmed that the high expression of MGMT in control GCSCs derived from patient \\#4 (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}e) was partially decreased in the presence of either LEV or TMZ, while it was strongly reduced by the combined treatment with TMZ and LEV (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}e). In contrast, none of the treatments modified MGMT expression in PCSCs derived from the GBM of patient \\#6 (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}e). The PCNA and MGMT expression levels were also investigated in the PCSCs derived from patient \\#2 that similarly to the PCSCs of patient \\#6 did not show a reduction of the proliferative rate after exposure to all the treatments. Western blot analysis revealed that none of the treatments significantly affected the MGMT expression level, and the slight effect of TMZ and LEV on PCNA expression was not significantly modified by the combined treatment with TMZ\u2009+\u2009LEV (data not shown).Fig.\u00a04LEV effect on MGMT expression in GCSCs. GCSCs derived from patient \\#4 were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) and with a combination of TMZ and LEV for 48\u00a0h. **a**, **c** Western blot analysis of MGMT and PCNA expression levels; representative immunoblots are shown. **b**, **d** Densitometric analysis of MGMT and PCNA expression levels (normalized to \u03b2-actin levels) of three independent experiments. Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant. **e** Confocal microscopy micrographs of GCSCs from patient \\#4 showing MGMT expression. Nuclei were stained with 4\u2032,6-diamidino-2-phenylindole (DAPI). Original magnification: \u00d7400 Fig.\u00a05LEV effect on MGMT expression in PCSCs. PCSCs derived from patient \\#6 were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) and with a combination of TMZ and LEV for 48\u00a0h. **a**, **c** Western blot analysis of MGMT and PCNA expression levels; representative immunoblots are shown. **b**, **d** Densitometric analysis of MGMT and PCNA expression levels (normalized to \u03b2-actin levels) of three independent experiments. Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant. **e** Confocal microscopy micrographs of PCSCs from patient \\#4 showing MGMT expression. Nuclei were stained with 4\u2032,6-diamidino-2-phenylindole (DAPI). Original magnification: X 400\n\nLEV induced HDAC4 nuclear translocation in TMZ-treated GCSCs but not in PCSCs {#Sec15}\n-----------------------------------------------------------------------------\n\nIt is has been demonstrated that some antiepileptic drugs, such as LEV, are able to regulate the HDACs activity. In particular, HDACs inhibitors can influence TMZ efficacy by modulating the expression of MGMT. Hence, we verified whether in our system, the LEV-induced sensitization to TMZ might be the result of a HDACs-dependent mechanism.\n\nTo this purpose, we first analyzed the expression levels and the cellular localization of HDAC4 in the GCSCs derived from patient \\#4, that, as we previously demonstrated, greatly reduced their proliferative rate following the combined treatment with TMZ and LEV, in comparison with the PCSCs derived from patient \\#6, that, by contrary, showed resistance to the same treatment. Immunofluorescence analysis (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}A) and Western blotting of nuclear and cytosolic fractions (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}B, C) revealed that the combination of TMZ and LEV induced the accumulation of HDAC4 into the nucleus of GCSCs derived from patient \\#4. Indeed HDAC4 expression was low and diffuse in the control and in correspondence of the separate treatments (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}A: panel a, b and c) while a strong nuclear expression was detected in TMZ\u2009+\u2009LEV treated cells (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}: panel d).Fig.\u00a06LEV effect on HDAC4 expression in GCSCs and PCSCs. GCSCs from patient \\#4 and PCSCs from patient \\#6 were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) and a combination of TMZ and LEV for 48\u00a0h. **A** panels a--d HDAC4 expression was evaluated in GCSCs from patient \\#4 by immunofluorescence analysis. Confocal microscopy micrographs showing HDAC4 (green) in untreated control cells (a), TMZ treated cells (b), LEV treated cells (c) and LEV\u2009+\u2009TMZ treated cells (d). **B** Cytoplasmic and nuclear extracts from GCSCs (patient \\#4) were evaluated for HDAC4 expression by Western blot analysis. A representative immunoblot is shown.\u00a0**C**\u00a0Densitometric analysis of HDAC4 expression levels of three indipendent experiments normalized using Stain free technology.\u00a0Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant. **D** panels a--d HDAC4 expression was evaluated in PCSCs from patient \\#6 by immunofluorescence analysis. Confocal microscopy micrographs showing HDAC4 (green) in untreated control cells (a), TMZ treated cells (b), LEV treated cells (c) and LEV\u2009+\u2009TMZ treated cells (d). **E** Cytoplasmic and nuclear extracts from PCSCs (patient \\#6), were evaluated for HDAC4 expression by Western blot analysis. A representative immunoblot is shown. **F**\u00a0Densitometric analysis of HDAC4 expression levels of three indipendent experiments normalized using Stain free technology. Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant\n\nInterestingly, none of the treatments induced HDAC4 nuclear translocation in PCSCs derived from patient \\#6 as demonstrated by the immunofluorescence and Western Blotting of nuclear and cytosolic fractions, as shown in Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}D--F. We also performed Western blot analysis of HDAC4 expression in the total lysate of the PCSCs derived from patient \\#2 subjected to the same treatments described above, demonstrating that the low increase in HDAC4 expression observed after TMZ treatment was not significantly modified by the combined treatment with TMZ and LEV (data not shown).\n\nLEV induced apoptosis in TMZ-treated GCSCs but not in PCSCs {#Sec16}\n-----------------------------------------------------------\n\nIt has been demonstrated that one of the mechanisms involved in the cytotoxic effect of chemotherapeutic agents is apoptosis. Indeed, in Jurkat cells, all the chemotherapeutic treatments significantly induced high levels of the activity of some of the pro-caspases that are known to act as initiators (such as Caspases-2, -8 and -9) and some of the caspases (such as Caspase-3 and -6) that are known to act as effectors of apoptosis (IRI treatment only did not significantly induce the activity of Caspase-6) (Additional file [1](#MOESM1){ref-type=\"media\"} and Additional file [3](#MOESM3){ref-type=\"media\"}: Figure S2). To explore whether the anti-proliferative effect of the combined treatment with TMZ and LEV was associated with apoptotic death, we tested for the presence and cellular localization of cleaved Caspase-3 under the same condition described above. By immunofluorescence analysis, we observed that in the GCSCs derived from patient \\#4, Caspase- 3 is present in the cytoplasm of the control and of the cells that received the individual treatments with LEV or TMZ (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}A: panels a--c), while its expression was strongly increased in the nuclei when the cells received the combined treatment LEV\u2009+\u2009TMZ (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}A: panel d). To confirm this result, we assessed the levels and the intracellular localization of cleaved Caspase-3 by Western blot analysis of nuclear and cytosolic fraction. Figure\u00a0[7](#Fig7){ref-type=\"fig\"}B and C show that Caspase-3 expression was significantly increased in the nuclei of GCSCs of patient \\#4 when the cells were treated with TMZ and LEV together, while none of the treatments exerted Caspase-3 significant modifications in cytosolic extracts. In contrast, immunofluorescence and Western blot analysis performed on PCSCs derived from patient \\#6 demonstrated that although low levels of Caspase-3 were detected in both nuclei and cytosolic fractions, none of the treatments induced changes in its expression (Fig.\u00a0[7](#Fig7){ref-type=\"fig\"}D--F). Caspase- 3 expression was also evaluated by Western blot analysis in the total lysate of the PCSCs derived from patient \\#2. Although a slight but significant increase in the Caspase-3 expression was detectable after TMZ treatment, this effect was not modified by the combined treatment with TMZ\u2009+\u2009LEV (data not shown).Fig.\u00a07LEV effect on Caspase-3 expression in GCSCs and PCSCs. GCSCs from patient \\#4 and PCSCs from patient \\#6 were treated with TMZ (250\u00a0\u00b5M), LEV (40\u00a0\u03bcg/mL) and a combination of TMZ and LEV for 48\u00a0h. **A** panels a--d Caspase-3 expression was evaluated in GCSCs from patient \\#4 by immunofluorescence analysis. Confocal microscopy micrographs showing Caspase-3 (green) in untreated control cells (a), TMZ treated cells (b), LEV treated cells (c) and LEV\u2009+\u2009TMZ treated cells (d). **B**\u00a0Cytoplasmic and nuclear extracts from GCSCs (patient \\#4) were evaluated for Caspase-3 expression by Western blot analysis. A representative immunoblot is shown.\u00a0**C**\u00a0Densitometric analysis of Caspase-3 expression levels\u00a0of three indipendent experiments normalized using Stain free technology. Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant. **D** panels a--d Caspase-3 expression was evaluated in PCSCs from patient \\#6 by immunofluorescence analysis. Confocal microscopy micrographs showing Caspase -3 (green) in untreated control cells (a), TMZ treated cells (b), LEV treated cells (c) and LEV\u2009+\u2009TMZ treated cells (d). **E** Cytoplasmic and nuclear extracts from PCSCs (patient \\#6) were evaluated for Caspase-3 expression by Western blot analysis.\u00a0A representative immunoblot is shown.\u00a0**F** Densitometric analysis of Caspase-3 expression levels of three indipendent experiments normalized using Stain free technology.\u00a0Statistical analysis was performed using Student's *t*-test. P value\u2009\\<\u20090.05 was considered statistically significant\n\nDiscussion {#Sec17}\n==========\n\nGlioblastoma is a very aggressive form of brain tumor particularly resistant to the standard therapies which include maximal surgical resection, followed by combined treatment with radiotherapy and chemotherapy \\[[@CR6]\\]. Alkylating agents, such as TMZ and CARB, and topoisomerase inhibitors, such as ETO and IRI, although effectively improve clinical outcomes when used alone or in combination with radiotherapy, display several adverse effects \\[[@CR6]\\] and their administration has not significantly changed the survival for GBM over the last years, making chemoresistance one of the biggest problems for GBM therapy \\[[@CR37]\\]. Thus, it is evident the need to identify new therapeutic strategies that can increase the life spans of patients affected by GBM.\n\nGlioblastoma stem-like cells (GSCs) represent a subpopulation within the heterogeneous tumor mass of GBM, with\u00a0a high similarity with neural stem cells \\[[@CR6], [@CR38]\\]. They are characterized by elevated proliferative rate and tumorigenic capability in vivo \\[[@CR39], [@CR40]\\] and are thought to be responsible for the resistance to standard therapies \\[[@CR41], [@CR42]\\]. It has been demonstrated that the peritumor tissue is the site of tumor recurrence in 90% of the patients \\[[@CR19]\\]. This area shows complex changes such as edema, increased vascularization, abnormal gene expression\u00a0and presence of numerous specialized cell types\u00a0\\[[@CR13]--[@CR18]\\], and is also the site in which a subpopulation of cancer stem-like cells (PCSCs) has been found \\[[@CR12]\\]. Although both\u00a0GCSCs and PCSCs express stem cell markers, they have different characteristics in terms of self-renewal and tumorigenicity, suggesting that PCSCs may have high relevance in translational research \\[[@CR12]\\].\n\nIn this study, we investigated the response of IDH1-wildtype GCSCs and PCSCs derived from six patients affected by GBM to different chemotherapeutic drugs and verified the possibility to enhance their effect through the combined treatment with adjuvant molecules. The comparison of GCSCs and PCSCs behavior and the identification of the molecules involved in their differential response to the treatments may provide a further insight in the complexity of GBM-neighboring microenvironment, which plays a crucial role in tumor progression.\n\nSince hematopoietic cells represent the primary target of chemotherapy-related adverse effects, here we first analyzed the efficacy of different chemotherapeutic drugs (TMZ, ETO, IRI and CARB) in Jurkat cells (Additional file [2](#MOESM2){ref-type=\"media\"}: Figure S1A). As expected, all the tested antineoplastic agents dramatically decrease the proliferation of these cells. In particular, BrdU assay demonstrates that IRI and CARB have the stronger effect in decreasing proliferation while TMZ is less efficient. Western blot analysis of PCNA expression confirms that among the used compounds, TMZ exerts the lower anti-proliferative activity (Additional file [2](#MOESM2){ref-type=\"media\"}: Figure S1B and C). This effect seems to be related to the activation of apoptotic pathway since all the chemotherapeutic drugs increased the activity of both pro- and effector-caspases (Additional file [3](#MOESM3){ref-type=\"media\"}: Figure S2). We then investigated the effects of the different chemotherapeutic treatments on the proliferation rate of GCSCs and PCSCs. Both cell populations resulted more resistant than Jurkat cells to the different treatments and we noted that all the chemotherapeutic agents exerted significant effects only in the cells with higher values of BrdU incorporation (GCSCs and PCSCs derived from patients \\#1, \\#2 and \\#3), whereas none of the drugs affected significantly the cells with a lower proliferation rate (GCSCs and PCSCs derived from patients \\#4, \\#5 and \\#6). Moreover, although no significant differences \u00a0have been found between GCSCs and PCSCs, TMZ is the drug with the lower efficacy in decreasing the proliferation of neurosphere clones derived from GBM of all the six patients. TMZ is a small molecule that is readily absorbed in the digestive tract and, because of its lipophilia, it is able to cross the blood--brain barrier. TMZ is the most widely chemotherapeutic drug used in patients with GBM, although the majority of the patients demonstrate de novo or acquired resistance, with subsequent tumor progression \\[[@CR5], [@CR43]\\]. Thus, the identification of the mechanisms of resistance and the attempting to enhance its effect can represent a good therapeutic strategy. MGMT repairs cytotoxic DNA lesions generated by TMZ. Many studies have shown a mechanistic link between MGMT activity and TMZ resistance, with suppression of MGMT activity resulting in increased cytotoxicity and MGMT protein overexpression which lead to resistance \\[[@CR31]\\]. However, the role of MGMT in the evolution of acquired resistance is not well established as demonstrated by several papers revealing that MGMT expression is not always related to resistance to TMZ treatment \\[[@CR44]--[@CR46]\\]. Since our results demonstrate that TMZ is the less efficient chemotherapeutic agent in decreasing the proliferation of both GCSCs and PCSCs, we have investigated the possibility of increasing its anti-tumor activity by means of the combined use with LEV. LEV is a relatively new non-enzyme inducing AED strongly recommended as a first line drug for patients with brain tumors \\[[@CR47]\\]. Recently, a growing body of evidence suggests that selected AEDs could lead to significant pharmaco-epigenetic interactions. It has been reported that LEV could act by favoring the recruitment of inhibitory complex, including HDAC, on the MGMT promoter, thus reducing its transcription \\[[@CR25]\\]. Here we show that in the GCSCs the combined treatment with LEV and TMZ decreases the MGMT expression levels and induces the nuclear translocation of HDAC4, suggesting, in agreement to what reported by Bobustuc et al. \\[[@CR25]\\], an HDAC4-dependent inhibitory role for MGMT transcription, thus increasing the sensitivity of these cells to TMZ treatment. In contrast, HDAC4 is expressed, although at low levels, in PCSCs, but none of the treatments modulated its expression and cellular localization. As GSCs show intrinsic deregulation in apoptotic cell death, we investigated whether in our system, the decreased proliferation rate observed in the presence of the combined treatment with LEV and TMZ could be associated with the activation of apoptotic pathway. Here we show an increased caspase 3 nuclear accumulation in GCSCs, while no change of its expression was observed in PCSCs. This result is supported by the analysis of pro-and effector-caspase expression evaluated in Jurkat cells after exposure to different chemotherapeutic agents showing that all the treatments significantly induced high levels of the activity of both classes of caspase compared to untreated cells (Additional file [3](#MOESM3){ref-type=\"media\"}: Figure S2). All these data suggest that activation of apoptotic pathway is involved in the strong anti-proliferative effect exerted by the used antineoplastic drugs.\n\nTaken together our results demonstrate that LEV enhances the TMZ effect on GCSCs by HDAC4-dependent downregulation of MGMT and by the activation of apoptotic pathways. PCSCs seem to be more resistant to the treatment, suggesting that the peritumoral microenvironment can favor the activation of survival mechanisms that make this therapeutic approach less effective. Our results are supported by data reported by Kim et al. demonstrating that the median progression-free survival and overall survival for patients who received LEV in combination with TMZ were significative longer than those for patients who did not receive LEV \\[[@CR48]\\]. In addition, a case report was published by Peddi et al. where a continuous regression of GBM was noted in a patient who received LEV and Dexamethasone without any cancer-targeted therapy, suggesting that the response may be related to Dexamethasone and/or LEV treatment \\[[@CR49]\\]. Although more studies are needed to better evaluate the role of LEV and its molecular mechanism of action, these papers together with our results strongly suggest the beneficial effect of LEV as a chemosensitizer agent.\n\nConclusions {#Sec18}\n===========\n\nAlthough further studies are necessary to better characterize the GSC environment, our results suggest that the clinical therapeutic efficacy of TMZ in GBM might be potentiated by the combination treatment with LEV and that the enhancement of apoptotic pathways may represent a primary goal in the development of new and more effective strategies.\n\nAdditional files\n================\n\n {#Sec19}\n\n**Additional file 1.** Description (Results, Material and Methods) of additional figures S1 and S2. **Additional file 2: Figure S1.** Effects of different chemotherapeutic drugs on Jurkat cell proliferation. (A) BrdU cell proliferation assay of Jurkat cells treated for 48\u00a0h with TMZ (250\u00a0\u00b5M), ETO (10\u00a0\u00b5M), IRI (10\u00a0\u00b5g/ml) and CARB (10\u00a0\u00b5g/ml). (B) Western blot analysis of total lysates from Jurkat cells, treated as described above was performed to detect PCNA expression levels; \u03b2-actin was used as a loading control. (C) Densitometric analysis of three independent experiments on PCNA expression levels. \\*\\*\u2009p\\<\u20090.01, \\*\\*\\*p\u2009\\<\u20090001 vs control by Student's *t*-test. **Additional file 3: Figure S2.** Effects of different chemotherapeutic agents on apoptosis induction in Jurkat cells. Jurkat cells were treated with the same concentrations of the antineoplastic drugs described in additional Fig.\u00a01. After 48\u00a0h, the activity of the pro-caspases-2, -8 and -9 and of the effector caspases -3 and -6 was measured by using *ApoTarget* Caspase Colorimetric Protease Assay. The results are representative of three independent experiments. \\*p\u2009\\<\u20090.05, \\*\\*p\u2009\\<\u20090.01, \\*\\*\\* \u2009p\\<\u20090.001 vs control by Student's *t*-test.\n\nGBM\n\n:   gliobastoma multiforme\n\nCSCs\n\n:   cancer stem cells\n\nGSCs\n\n:   glioblastoma stem-like cells\n\nGCSCs\n\n:   glioblastoma cancer stem cells\n\nPCSCs\n\n:   peritumoral cancer stem cells\n\nTMZ\n\n:   temozolomide\n\nHDACs\n\n:   histone deacetylases\n\nHATs\n\n:   histone acetyltransferases\n\nMGMT\n\n:   O6-methylguanine-DNA methyltransferase\n\nCTR\n\n:   control\n\nLEV\n\n:   levetiracetam\n\nAEDs\n\n:   antiepileptic drugs\n\nPCNA\n\n:   proliferating cell nuclear antigen\n\nETO\n\n:   ETO\n\nIRI\n\n:   irinotecan\n\nCARB\n\n:   CARB\n\nDAPI\n\n:   4\u2032,6-diamidino-2-phenylindole\n\nBMS designed all the experiments, performed BrdU and Apoptosis assays and wrote the manuscript. GD and SS performed the Western Blotting assays. SS and GP collected the neurospheres and performed Immunofluorescence analysis. BMS and AC performed confocal analysis. EB kindly provided the neurospheres. EB performed the analysis of IDH1 status in the neurospheres. LML performed the analysis of IDH1/2 status and MGMT methylation status on tissue samples. GS and GL critically revised the paper. All authors read and approved the final manuscript.\n\nAcknowledgements {#FPar1}\n================\n\nThe Authors sincerely thank Prof. A.Vescovi for kindly providing the neurospheres, Dr Nadia Trivieri for her expert technical assistance and Dr Cristina Giacinti for her support in the BrdU assays.\n\nCompeting interests {#FPar2}\n===================\n\nThe authors declare that they have no competing interests.\n\nAvailability of data and materials {#FPar3}\n==================================\n\nAll original data are available upon request.\n\nConsent for publication {#FPar4}\n=======================\n\nNo parts of this manuscript are being considered for publication elsewhere.\n\nEthics approval and consent to participate {#FPar5}\n==========================================\n\nTissue samples from where neuropsheres were derived were provided by the Istituto Neurologico \"C. Besta\", Milano and by the Istituto di Neurochirurgia, Universit\u00e0 Cattolica del Sacro Cuore, Roma, according to the ethical requirements of the institutional committees on human experimentation. Informed consent of the patients was obtained according to the Helsinki declaration of ethical requirements.\n\nFunding {#FPar6}\n=======\n\nThis work was supported from Ministero dell'Istruzione, dell'Universit\u00e0 e della Ricerca-FIRB \"Accordi di Programma\" 2010 Project RBAP10KJC5_003.\n\nPublisher's Note {#FPar7}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "![](medchirjrev78002-0088){#sp1 .368}\n"}
+{"text": "1. Introduction {#s0005}\n===============\n\nLymphocytes are considered the mainstay of anticancer immunity. As a result, there is significant interest in defining and enhancing their role. While the benefits of radiation therapy are well established, the interplay of radiation and immunity are less well understood. Lymphocytes are considered one of the most radiosensitive mammalian cells. The initial concerns for the impact of immune suppression have more recently been balanced by evidence that radiation therapy may enhance cellular immunity. There is much to be learned about those interactions, especially regarding lymphocyte subgroups. Almost all the studies of the effect of radiation on the various subgroups contain chemotherapy, which has its own profound effects on lymphocytes. To better understand the direct effect of radiation therapy on lymphocytes, we collected data in a well-defined patient group with fairly uniform radiation treatment without the confounding effect of systemic therapy.\n\n2. Materials and methods {#s0010}\n========================\n\nWith institutional review board (IRB) approval and with informed consent, a prospective study was undertaken in consecutive male patients receiving radiation for prostate cancer. One patient was found to have metastatic disease shortly after consent and was excluded, but a total of 15 patients were recruited. No patient received chemo or hormonal therapy. All patients had the pelvic lymphatics treated to 54\u00a0Gy. The lymphatics were defined by the major blood vessels, including the internal and external iliac up to the common iliac vessels. The volume of the pelvis (including the pelvis proper, the proximal femurs to the level of this ischium, the sacrum and L4/5) was drawn and the volume receiving \\>20\u00a0Gy (well above the known sensitivity of lymphocytes) was calculated.\n\nAfter the initial pelvic radiation, most of the patients (n\u00a0=\u00a011) received 70\u00a0Gy to the prostate fossa, 3 received 78\u00a0Gy to the prostate and one underwent a brachytherapy boost. Blood for complete blood counts (CBC) and flow cytometry were collected right before the start of treatment (within 2\u00a0weeks), end of pelvis treatment (within a week) and 3\u00a0months post therapy. ([Table 1](#t0005){ref-type=\"table\"} for characteristics). Cluster of differentiation (CD) markers were used to determine lymphocyte subgroups via flow cytometry. CD3+ cells are considered T lymphocytes, with the CD3+ subgroups of CD4+ (T helper) and CD8+ (T cytotoxic). CD 19+ are considered B cells and CD56+ are natural killer (NK) cells. For brevity, going forward, the CD3+ CD4+ and CD3+ CD8+ will be referred to as CD4+ and CD8+, respectively. Testing was accomplished with the three color TriTest monocolonal antibody panel for CD3/CD8/CD45, CD3/CD4/CD45, CD3/CD19/CD45 and CD3/CD16+ CD56/CD45 with analysis on the Becton Dickinson Macintosh FACSCalibur system (Becton Dickinson and Company, Franklin Lakes NJ, USA). Each specific reagent (20\u00a0\u00b5L) is added to separate tubes containing 50\u00a0\u00b5L of whole blood and vortexed and incubated at room temperature for 15--30\u00a0min. Then 0.5\u00a0ml of BD Lyse Solution is added and incubated for 15\u00a0min. The specimens are then loaded on the flow cytometry analyzer and run as per standard and data collected.Table 1Baseline patient characteristics.Mean/medianrangeStandard deviationAge (years)64/6452--79Total lymphocyte/\u00b5L1998/19221113--3192617CD3+/\u00b5L1378/1388603--246524CD4+/\u00b5L856/924298--1319309CD8+/\u00b5L501/428190--1280283CD19+/\u00b5L318/241130--1054231CD56+/\u00b5L255/192111--753169\n\n3. Statistics {#s0015}\n=============\n\nSample characteristics were described using descriptive statistics. Frequencies and percentages were used to describe categorical variables. Means and standard deviations (or medians and ranges where appropriate) were used to describe continuous variables. A one-sample *t*-test (or Wilcoxon signed-rank test when appropriate) was used to test if specific lymphocyte components experiences a significant change over time. A two-sample *t*-test (or Wilcoxon rank-sum test when appropriate) was used to test for differences in lymphocyte counts between dose groups. A generalized estimating equation (GEE) was used to assess the difference each specific lymphocyte component experienced between time points, while adjusting for intra-patient correlations. Post-hoc Tukey-Kramer adjustments were made for multiple comparisons within each GEE model. Statistical significance was set to p\u00a0\\<\u00a00.05.\n\n4. Results {#s0020}\n==========\n\nAn average of 61% (range 53--69%) of the pelvic bone volume receiving \u226520\u00a0Gy. There was no difference between whether the patient was post prostatectomy versus intact prostate in the degree of lymphocyte decline.\n\nThe overall leukocyte (white blood cell \\[WBC\\] count) declined from a mean of 6370 cells/\u00b5L to 4070\u00a0cells/\u00b5L (34% decline) by the end of treatment, and remained relatively stable (4510\u00a0cells/\u00b5L) at 3\u00a0months, primarily due to the slow recovery of the lymphocytes. ([Table 2](#t0010){ref-type=\"table\"}).Table 2Change in blood counts from pretreatment to end of treatment and at 3\u00a0months post treatment.Mean/medianbaseMean/median end% Decline from baseMean/median 3\u00a0months% Decline from baseTotal leukocytes6370/61004070/420034%/34%4510/450028%/31%Total lymphocytes1998/1922524/47073%/73%832/84857%/56%CD3+1378/1388383/36772%/75%531/46760%/62%CD4+856/924236/21173%/75%305/29463%/65%CD8+501/428138/13072%/76%159/27057%/58%CD19+318/24125/2091%/91%113/9359%/62%Cd56+255/19295/6762%/68%155/12136%/36%\n\nOverall, the total lymphocyte count declined by 73% at the end of treatment (to 27% of baseline) ([Table 2](#t0010){ref-type=\"table\"}). At 3\u00a0months, the lymphocytes had recovered, but were still \\<50% of the starting value. The CD3+, CD4+, CD8+, and CD56+ lymphocyte subgroups were equally sensitive when compared to each other. The CD19+ was significantly more sensitive, declining to just 9% of initial levels. While not significantly different (p\u00a0\\>\u00a00.1), the CD56+ cells showed a trend to being more resistant, with less of a decline (about 10% less) than the CD3+, CD4+ and CD8+ subgroups. None of the subgroups were back to normal levels at 3\u00a0months, but the recovery of the CD56+ cells (P\u00a0\\<\u00a00.0001) was more complete than the other groups. Also, although they were more sensitive, the recovery of the CD19+ cells was more robust (P\u00a0\\<\u00a00.0001) than the other subgroups, so that there was no difference in the per cent decline from baseline at 3\u00a0months compared to the other subgroups.\n\nAs a percentage of the total lymphocytes, because of their greater sensitivity, at the end of treatment the per cent of CD19+ cells had declined and, with their relative resistance, the percent of CD56% cells increased. ([Table 3](#t0015){ref-type=\"table\"}) At three months as cells recovered, the per cent contribution of CD4+ cells had declined some from the end of treatment and the CD56+ increased.Table 3Percent of each subgroup as a constituent of the total lymphocyte population. Note that CD3+ includes both CD3+ CD4+ and CD3+ CD8+ lymphocytes.Baseline % Mean/medianEnd of treatment % mean/median3\u00a0Months % mean/medianCD3+69%/69%71%/72%63%/61%CD4+43%/43%45%/48%36%/37%CD8+24%/25%25%/24%25%/22%CD19+16%/14%6%/4%15%/14%CD56+13%/12%18%/21%19/17%\n\n5. Discussion {#s0025}\n=============\n\nLymphocytes are extremely sensitive to radiation. In our patients, the total lymphocyte count dropped 73%. At 3\u00a0months, recovery occurs, but is incomplete. We showed that as a whole the CD3+ subgroup and specifically the CD3+ CD4+, the CD3+ CD8+ and the CD56+ subgroups exhibited the same radiosensitivity. CD19+ cells were statistically more radiosensitive, but recovered more rapidly so that at 3\u00a0months the overall decline was no different than the other subgroups. NK (CD56+) cells were somewhat more resistant and at 3\u00a0months exhibited a statistically significant smaller decline from baseline than the other lymphocyte sub types. Overall, at the end of treatment, CD4+ and CD8+ had declined about 72% and at 3\u00a0months were about 60% below baseline. CD19+ declined 91%, but at 3\u00a0months was about 60% below baseline and CD56+ cells initially declined 62% but were only 36% below baseline at 3\u00a0months ([Table 2](#t0010){ref-type=\"table\"}).\n\nThere are few studies that have evaluated the effect of radiation therapy alone on lymphocyte subgroups and reported the actual lymphocyte counts ([Table 4](#t0020){ref-type=\"table\"}). Two were in cervix cancer. The first (1) reported on 39 cervix cancer patients with radiation alone. From the figures, CD8+ counts dropped from a mean of 670\u00a0cells/\u00b5L to 280\u00a0cells/\u00b5L (59% decline) and CD56+ from 350\u00a0cells/\u00b5L to 190\u00a0cells/\u00b5L (46% decline). They did not measure CD4 cells. As a percentage of the total lymphocyte count, CD8+ count increased from 27% to 32% and the CD56 from 13% to 20%. They did not report recovery data. In the second study (2) there were 14 patients with pelvic radiation alone and the total lymphocyte count declined from a mean of 1137\u00a0cells/\u00b5L to 357 cells/\u00b5L (69% decline), which improved to 639\u00a0cells/\u00b5L (overall 44% decline) by 6\u00a0months (a 25% recovery). For each of the subgroups, CD4+ declined from 566 cells/\u00b5L to 202 cells/\u00b5L (64% decline) and at 6\u00a0months 223\u00a0cells/\u00b5L (61% decline overall, 3% improvement from end of treatment); CD8+ from 323\u00a0cells/\u00b5L to 196 cells/\u00b5L (39% decline), at 6\u00a0months 298\u00a0cells/\u00b5L (8% decline, 31% recovery); CD19+ from 132\u00a0cells/\u00b5L to 27\u00a0cells/\u00b5L (80% decline), at 6\u00a0months 88\u00a0cells/\u00b5L (33% decline, 47% improvement) and CD56+ from 251 cells/\u00b5L to 129\u00a0cells/\u00b5L (49% decline, at 6\u00a0months 222\u00a0cells/\u00b5L (12% decline, 37% improvement). Both of these studies had patients with whole pelvis treatment, so volumes were similar to ours. There was a big difference in the baseline counts, probably reflecting the different patient populations, labs and methodology, but the cell count declines and recovery were similar.Table 4Comparative data.Total lymphocyte (\u00b5L)CD3+ (\u00b5L)CurrentEric [@b0005]Bach-tiary [@b0010]Belka [@b0015]Maehata [@b0020]CurrentEric [@b0005]Bach-tiary [@b0010]Belka [@b0015]Maehata [@b0020]Base29981137140513631378872952872End524357462917383374323636Decline73%69%67%33%72%67%66%27%f/u3\u00a0mos6\u00a0mos4\u00a0mos3\u00a0mos6\u00a0mos4\u00a0mos832639858531484593Decline57%44%39%60%44%38%\u2028\u2028CD4+ (\u00b5L)CD8+ (\u00b5L)CurrentEricBach-tiaryBelkaMaehtaCurrentEricBach-tiaryBelkaMaehataBase856566560566501670232421299End236202202424138280196130209Decline73%64%64%25%72%59%39%69%30%f/u3\u00a0mos6\u00a0mos4\u00a0mos3\u00a0mos6\u00a0mos4\u00a0mos305223333159298260Decline63%61%41%57%8%38%\u2028\u2028CD19+ (\u00b5L)CD56+ (\u00b5L)CurrentEricBach-tiaryBelkaMaehtaCurrentEricBach-tiaryBelkaMaehataBase318132153163255350251116306End252714809519012929188Decline91%80%91%51%62%46%49%75%49%f/u3\u00a0mos6\u00a0mos4\u00a0mos3\u00a0mos6\u00a0mos4\u00a0mos1138864155222102Decline59%33%58%36%12%12%\n\nThere were two additional studies with dissimilar treatment volumes. The first consisted of 10 seminoma patients given 26\u00a0Gy to the para aortic lymphatics (3). They noted no differences in the baseline counts of the patients and healthy volunteers. The total lymphocyte count dropped from 1405\u00a0cells/\u00b5L to 562 cells/\u00b5L at the end of treatment (67% decline) and was 858\u00a0cells/\u00b5L (39% decline, 28% improvement) at 4\u00a0months. For CD4+ cells baseline was 560 cells/\u00b5L, declining to 202\u00a0cells/\u00b5L (64% decline) and 333 cells/\u00b5L at 4\u00a0months (41% decline, 23% improvement); CD8+ 421\u00a0cells/\u00b5L, to 130 cells/\u00b5L (69% decline) and 260 cells/\u00b5L at 4\u00a0months (38% decline, 31% improvement); CD19+ 153\u00a0cells/\u00b5L to 14\u00a0cells/\u00b5L (91% decline) and 64 cells/\u00b5L at 4\u00a0months (58% decline, 33% improvement): and CD56+ 116\u00a0cells/\u00b5L to 29 cells/\u00b5L (75% decline) and 102 cells/\u00b5L at 4\u00a0months (12% decline, 63% improvement).\n\nIn the second study, 36 lung cancer patients (4) were treated with focal (small field) therapy with either 40--60\u00a0Gy in 4 fraction or 60--70\u00a0Gy in 10 fractions. The total lymphocyte count declined from a mean of 1363\u00a0cells/\u00b5L to 927\u00a0cells/\u00b5L (69% decline); CD4+ declined from 566\u00a0cells/\u00b5L to 424 cells/\u00b5L (64% decline); CD8+ from 299 cells/\u00b5L to 209 cells/\u00b5L (39% decline); CD19+ from 162 cells/\u00b5L to 80 cells/\u00b5L (80% decline), and CD56+ from 306 cells/\u00b5L to 188\u00a0cells/\u00b5L (49% decline). They did not report follow up levels. In general, with the smaller volumes they used, when compared to ours and the other studies, the decline was not as profound, except in the B cell (CD 19+) population. This would suggest that since we are well above the dose sensitivity of lymphocytes that volume treated becomes more important. A study using stereotactic body radiation therapy (SBRT) to small volumes in pancreatic cancer confirmed a more modest decline (35%) in total lymphocyte counts [@b0025] than we see with larger field therapy.\n\nIn spite of the disparities in the populations, what is apparent is that CD3^+^ and the CD3^+^ CD4^+^ subgroups have similar radiosensitivity. CD3^+^CD8^+^ is also similar, but more variable between the studies. CD19^+^ cell are the most radiosensitive, but recover more quickly. CD56^+^ are more resistant and recover towards baseline more quickly. This information should be helpful in trying to understand the effect of radiation therapy on anti-cancer immunity.\n\nFunding {#s0030}\n=======\n\nProvided through a philanthropic gift to the Baylor Scott & White Central Texas Foundation, 2401 S. 31st St. Temple, TX 76508.\n\nDeclaration of Competing Interest\n=================================\n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\n"}
+{"text": "INTRODUCTION\n============\n\nIn multi-cellular organisms, the control of gene expression is key not only for development, but also for adult cellular homeostasis, and deregulation of gene expression correlates with aging ([@b1-bmb-52-086]--[@b4-bmb-52-086]). Aging is the main risk factor for many chronic diseases, such as neurodegeneration, cardiovascular disease, type II diabetes, osteoporosis, and cancer. Several conserved and interconnected pathways are deregulated during aging, delineating \"hallmarks\" or \"pillars\" of aging ([@b5-bmb-52-086], [@b6-bmb-52-086]). However, how these hallmarks are achieved throughout the life of an organism is still poorly understood.\n\nIn the laboratory, several models are used to study aging: (i) chronological aging (*i.e*. the time that passed in the organism's life), which most closely corresponds to the most intuitive definition of aging, (ii) replicative aging (*i.e*. the number of times a cell can divide; most used to study aging in yeast) ([@b7-bmb-52-086]), (iii) cellular senescence, which corresponds to the replicative limit of cells before they enter irreversible cell cycle arrest, also dubbed the 'Hayflick limit' ([@b8-bmb-52-086]), and (iv) progeria or premature aging syndromes (*i.e*. Hutchinson-Gilford or Werner Syndrome) ([@b9-bmb-52-086]). Although aging is a common feature of many taxa, the rate of aging differs between species, and accumulating evidence suggests that the process can be fine-tuned within a species based on environmental cues, inter-individual genetic variations, and even sex ([@b10-bmb-52-086]). In this review, we will focus primarily on transcriptional landscape alterations in chronological aging and pro-longevity models.\n\nREMODELING OF THE PROTEIN CODING TRANSCRIPTIONAL LANDSCAPE WITH AGING AND IN LONGEVITY MODELS\n=============================================================================================\n\nIn this review, we chose to compare age-related and longevity-related gene expression changes in different vertebrate species (*e.g*. mouse, human, rat, African turquoise killifish, *etc*.). Although this enabled us to draw more general comparisons, we understand that each species may have specific aging features, which limits our ability to draw a more general conclusion. Another general limitation in the field of aging genomics is the ill-defined age criteria for young and old animals that are profiled. For example, the age of \"young\" mice used in discussed studies ranges from age 1 to 4 months, while \"old\" mice range from age 18 to 30 months ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}). The lack of consensus age criteria makes reaching definitive conclusions on the effect of aging more difficult. Future studies should strive to use stricter age criteria for the study of aging, as well as include more time points along the life of animals to identify transcriptional trajectories with aging.\n\nGene expression remodeling across cells and tissues in vertebrates\n------------------------------------------------------------------\n\n### Transcriptional remodeling aging signatures across tissues\n\nWe summarized reported age-related transcriptional changes across key tissues that are known to undergo age-related functional decline (*i.e*. liver, heart, pancreatic islets, brain, kidney) ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}). For simplicity and exhaustiveness' sake, both changes in gene expression from microarrays and RNA-sequencing datasets are reported. As illustrated in [Table 1](#t1-bmb-52-086){ref-type=\"table\"}, three common aging transcriptional signatures have been recurrently identified across different tissues and species. A prominent aging signature that affects all categories of tissues is increased expression of inflammatory pathways. More broadly, Gene Ontology (GO) categories of, 'immune response', 'stress response', 'chemokines', 'lysosome', 's100 family', 'caspase family', 'wound healing', 'fibrosis', and 'complement system' encompass genes that if left unattenuated, propagate an environment of oxidative stress and vascular damage ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}) ([@b11-bmb-52-086]--[@b16-bmb-52-086]). Importantly, chronic activation of the immune system is the defining characteristic of 'inflamm-aging' (*i.e*. a low-level, sterile, chronic state of inflammation with age), and the global increase of pro-inflammatory gene expression may thus be a driving factor in the development of age-related diseases ([@b17-bmb-52-086]--[@b19-bmb-52-086]). For example, the inappropriate activation of immune cells can exacerbate cancer progression ([@b20-bmb-52-086]), pathogenesis of type II diabetes ([@b21-bmb-52-086]), and cardiovascular disease susceptibility ([@b22-bmb-52-086], [@b23-bmb-52-086]).\n\nAnother notable general aging signature is an overall decrease in expression of metabolic genes ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}). Disrupted metabolic pathways in the mitochondrion (*i.e*. TCA cycle, electron transport chain), protein synthesis (*i.e*. cysteinyl-tRNA-synthase, phenylalanine-tRNA-synthase), and nutrient sensing (*i.e. Cartpt*, *Npy*, *Ppy*) were observed in the liver, heart, pancreatic islets, cerebellum, and whole brain tissues ([@b24-bmb-52-086]--[@b26-bmb-52-086]). The reduced efficiency and ability to process nutrients may limit nutrient availability and decreased xenobiotic metabolism may increase the event of drug toxicity in aging animals. Such a combination of metabolic decline may lay the foundation for the development of age-related diseases, such as cancer or Alzheimer's disease ([@b27-bmb-52-086]).\n\nA common signature of decreased membrane integrity also seems to be a common theme in the liver, pancreatic islets, and kidney through the increase of 'fibrosis', 'extracellular matrix', and 'wound healing' gene expression ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}) ([@b25-bmb-52-086], [@b28-bmb-52-086], [@b29-bmb-52-086]). The extracellular matrix is a lattice of polysaccharides and proteins that stem from the cell membrane, which provides structural support and scaffolding for physical communication between cells ([@b30-bmb-52-086]). In fibrosis and wound healing, the extracellular matrix expands and remodels to build upon or reestablish the cell barrier ([@b30-bmb-52-086]). However, an overgrowth of the extracellular matrix can hinder signaling between cells. This phenomenon is especially relevant for the liver, pancreatic islets, and kidney, as cells in these tissues heavily rely on membrane accessibility for metabolism, hormone release, and filtration. In the aging liver, the fibrotic extracellular matrix may contribute to the susceptibility to age-related diseases, since liver diseases involving inflammation, such as nonalcoholic fatty liver disease, are worsened with fibrosis ([@b31-bmb-52-086]). Concerning the pancreas, type II diabetes is a great model to understand the consequences of fibrotic islets, as fibrosis has been found to promote beta cell destruction, ultimately compromising the delivery of insulin ([@b32-bmb-52-086]). In the aging kidney, obstruction to the basement membrane may partially explain the age-related decline in glomerular filtration rate ([@b28-bmb-52-086]). Thus, the aging-associated loss in membrane integrity, predicted by the observed transcriptional remodeling, should be ultimately detrimental to the ability of cells to properly interact with the rest of the organism, and lead to functional decline.\n\n### Tissue-specific transcriptional remodeling with aging\n\nAlthough common dysregulated pathways can be identified across tissues, many aging transcriptional defects seem to be tissue-specific ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}). For example, the aging liver presents a characteristic signature of decreased oxidative demethylation gene expression ([@b33-bmb-52-086]) ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}), a process that reverses the effects of regulatory or mutagenic methylation by removing methyl adducts from DNA and RNA ([@b34-bmb-52-086]). Deterioration of this ability to regulate and repair the methylome can induce improper gene expression. The pancreatic islets also reveal a unique aging signature of decreased 'cell-cell signaling', 'nutrient sensing', and 'neuronal recognition' gene expression ([@b25-bmb-52-086]). The decreased capacity to perform these functions creates juxtaposing outcomes. On one hand, decreased expression of these genes can create issues because the pancreatic islets depend on nearby cell communication, glucose sensing, and vagal nerve stimulation to assess energy status ([@b35-bmb-52-086]). Inaccurate sensing and transduction of energy status can harm the aging metabolism by disrupting the timing of pancreatic islet hormone release ([@b35-bmb-52-086]). Contrary to this, Riera *et al*. demonstrates an advantage with the decline of cell communication. Through the ablation of TRPV1 receptors and inhibition of calcitonin gene-related peptide, an insulin secretion blocker, the aging pancreas can improve its insulin release, yielding a more youthful metabolic profile ([@b36-bmb-52-086]). In light of these conflicting responses, pancreatic islet cell communication genes are evidently pleiotropic and are perhaps yet to be fully understood.\n\n### Age-related transcriptional remodeling: what about stem cells?\n\nA hallmark of aging is the depletion and loss of function in adult stem cell pools ([@b6-bmb-52-086]). The ability of stem cells to preserve their quiescent state is central to maintaining a long-term functional stem cell pool for cell regeneration and tissue homeostasis. However, with aging, stem cell proliferation and the consequent rate of cell turnover have been observed to decline ([@b37-bmb-52-086]). The adult somatic stem cells that have been best studied at the genomic level during aging are: neural stem cells (NSCs), muscle stem cells or \"satellite\" cells (MuSCs), and hematopoietic stem cells (HSCs) ([@b38-bmb-52-086]--[@b44-bmb-52-086]) ([Table 2](#t2-bmb-52-086){ref-type=\"table\"}). Interestingly, both HSCs and MuSCs display an upregulation of immune-related genes. This signature is akin to the one observed in bulk aging tissues ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}). Interestingly, a pro-inflammatory state resulting from an inappropriate immune response can both diminish and protect the viability and differentiation potential of the stem cell pool ([@b45-bmb-52-086]). This is also observed in NSCs and may be a product of an inflammatory microenvironment in the brain. In addition, aging HSCs and MuSCs share a transcriptional signature of decreased genome stability. For example, the decline in aging quiescent MuSCs may be explained by the silencing of genes through an increase in H3K27me3 ([@b40-bmb-52-086]). A decrease in epigenetic regulators and cell cycle genes was also observed with age in HSCs ([@b43-bmb-52-086], [@b44-bmb-52-086]), which is accompanied by a loss in self-renewal and differentiation capability ([@b43-bmb-52-086], [@b44-bmb-52-086]). Aging NSCs and HSCs share an increased expression of genes involved in cell adhesion ([@b39-bmb-52-086], [@b43-bmb-52-086]), which may have consequences on migration capacity and survival. Specific to NSCs, decreased lysosomal gene expression was observed, which may explain a reduced ability of aged NSCs to transition out of the quiescent state and into the active stem cell pool ([@b38-bmb-52-086]).\n\n### Lessons from bulk vs. single cell transcriptome profiling\n\nWe have discussed both bulk microarray and RNA-sequencing (RNA-seq) datasets, as well as published emerging single-cell RNA sequencing (scRNA-seq) ([Table 1](#t1-bmb-52-086){ref-type=\"table\"}, [2](#t2-bmb-52-086){ref-type=\"table\"}). Thus far, a major confounding factor of aging 'omics' studies (*i.e*. genome-wide analyses such as transcriptomics and epigenomics) lies in the heterogeneous cellular make-up of mammalian tissues, which makes it difficult to disentangle the significance of transcriptomic changes (*i.e*. resulting from changes in the underlying cellular makeup, or from cell type-autonomous changes). Computational deconvolution algorithms can be used to evaluate the relative contribution of cell type composition changes in transcriptomes of heterogeneous cell population (*i.e*. as is generally the case for tissues) ([@b46-bmb-52-086]), although they require reference purified transcriptional profiles of the cell types present in the mixture. In contrast, cell sorting may be used to analyze the fate of specific cell types. However, sorting relies on the existence of well-established cellular markers (which may not always be readily available for all cell types of interest or in emerging model organisms), and expression of such markers may itself be regulated with age, thus complicating data interpretation. Despite these limitations, bulk level RNA-seq is the more mature technology and enables the study of various aspects of transcription (*e.g*. strand-specificity, splicing isoforms, circRNAs, non-polyAdenylated RNAs, etc.).\n\nIn contrast to bulk methods, advances in single-cell profiling techniques now allow direct efficient and high-resolution transcriptional mapping of the transcriptomes of thousands of cells from heterogeneous samples by harnessing the power of microfluidics (*e.g*. Drop-seq, 10xGenomics Chromium, *etc*.) ([@b47-bmb-52-086]). These advances have enabled profiling at an unprecedented resolution in term of cell-to-cell variations and cell type identification ([@b48-bmb-52-086]). However, few studies have harnessed these single cell technologies in comparative aging transcriptomics studies yet. In a pioneering study of the aging pancreas at the single cell level, a transcriptional signature of loss of cell identity captured in scRNA-seq could not be detected in previous bulk RNA-seq studies ([@b49-bmb-52-086], [@b50-bmb-52-086]). However, it is important to note that most single-cell profiling technologies do not provide information on transcription beyond gene expression level, and they cannot robustly detect lowly expressed genes. Despite these shortcomings, scRNA-seq provides important insights for understanding the heterogeneity of gene expression within a cell population, or heterogeneity in underlying cell types within a mixed population. Both bulk RNA-seq and scRNA-seq still share the limitations of amplification bias and poor cDNA synthesis efficiency ([@b51-bmb-52-086]).\n\nAnother important point to note is that current studies use either ribosomal RNA depletion (ribodepletion) or poly(A) selection to enrich for RNA species of interest prior to library construction and subsequent sequencing. Each method is likely to result in differences in both the amount, representation and quality of the profiled RNA. With ribodepletion, ribosomal-RNA (rRNA) is specifically depleted from an RNA sample so other RNA species (*e.g*. pre-mRNA and mature mRNA, circRNAs, lncRNAs, *etc*.) remain for downstream profiling. In contrast, poly(A) enrichment specifically isolates RNA species with (long) polyadenylated tails, a characteristic specific only to mature mRNAs. Thus, specific changes reported by various studies may be affected by the experimenter's choices in target RNA selection. It will be important to determine the impact of RNA selection on observed signatures of the aging transcriptome.\n\nImpact of longevity interventions on the aging transcriptome\n------------------------------------------------------------\n\nAnalysis of known longevity-extending dietary, genetic, and drug interventions suggest that key pathways that decline transcriptionally with aging may be rescued at least partly by these interventions ([@b52-bmb-52-086], [@b53-bmb-52-086]) ([Table 3](#t3-bmb-52-086){ref-type=\"table\"}). In this review, we focus on the transcriptional effects on common aging signatures of four well-characterized pro-longevity interventions: dietary restriction, *Ames* dwarfism, rapamycin supplementation, and metformin supplementation.\n\nDietary restriction (DR) refers to the limitation of either (i) the total caloric intake or (ii) specific nutrients, in each case without malnutrition ([@b54-bmb-52-086]). Typically, caloric restriction (CR) is achieved by a 25 to 60% reduction of calories from *ad libitum* controls in model organisms ([@b54-bmb-52-086], [@b55-bmb-52-086]). By altering nutrient sensing pathways, DR has been proposed to modulate downstream gene expression to extend longevity ([@b56-bmb-52-086]). CR-specific modulations may partly rescue transcriptional aging through upregulation of DNA methyltransferase activity, histone methylation, and histone deacetylation via HDAC1 and SIRT1 ([@b57-bmb-52-086]). These transcriptional changes have been observed to affect the development of cancer, diabetes, cardiovascular diseases, neurodegenerative diseases, and immune deficiencies in rodents, nonhuman primates, and humans ([@b57-bmb-52-086]). In the case of specific nutrient restriction, limitation of dietary protein or specific amino acids (*i.e*. asparagine, glutamate, methionine, tryptophan) yield the most robust longevity extension across metazoans ([@b55-bmb-52-086], [@b58-bmb-52-086]--[@b62-bmb-52-086]). Particularly with methionine restriction, the decrease in methionine-derivatives, such as methyl donor S-adenosylmethionine, may modulate DNA methylation in a protective manner against age-related neoplastic disease ([@b63-bmb-52-086], [@b64-bmb-52-086]). Notably, key downstream pathways reversed by these interventions at the gene expression level include a decrease in inflammatory processes, and the prevention of age-associated epigenetic changes ([Table 3](#t3-bmb-52-086){ref-type=\"table\"}). These observations are compatible with the notion that dietary restriction can rescue aspects of aberrant age-related transcriptional remodeling.\n\nThe *Ames* dwarf mouse is a well-established longevity model ([@b65-bmb-52-086]). Because of a single nucleotide mutation in the *Prop1* gene, *Ames* dwarf mice lack the transcription factor responsible for pituitary gland cell differentiation ([@b65-bmb-52-086]). Thus, *Ames* dwarf mice exhibit reduced levels of circulating growth hormone, prolactin, and thyroid-stimulating hormone ([@b66-bmb-52-086]). These altered hormone levels can lead to non-autonomous changes in the transcriptional profile, potentially promoting longevity through increased insulin sensitivity and reduced oxidative stress ([@b65-bmb-52-086]). Most notably, these changes include DNA methylation and microRNA regulation ([@b53-bmb-52-086], [@b66-bmb-52-086]--[@b68-bmb-52-086]). Analogous to the effect of dietary restriction, the *Ames* dwarf mouse also displays a more stable epigenome throughout life ([@b52-bmb-52-086]).\n\nRapamycin and metformin supplementation are two of the most widely studied pharmaceutical pro-longevity interventions ([@b69-bmb-52-086]). These two drugs are thought to increase animal longevity by acting as CR mimetics ([@b70-bmb-52-086]). Rapamycin is an inhibitor of the mammalian target of rapamycin (mTOR), a kinase that regulates cell growth in response to nutrients, growth factors, cellular energy, and stress ([@b71-bmb-52-086]). In a fed state, mTOR is activated to initiate protein synthesis, whereas mTOR inhibition with rapamycin mimics a fasting state ([@b70-bmb-52-086]). Halting protein synthesis arrests cell growth, which may explain why rapamycin has been shown to slow aging and neoplastic proliferation ([@b72-bmb-52-086]). At the transcriptional level, rapamycin-induced mTOR inhibition slows the aging methylome ([@b52-bmb-52-086], [@b53-bmb-52-086]). Metformin is a prevalent anti-hyperglycemic drug that primarily works by uncoupling the electron transport chain, thereby mimicking a fasted/low-energy state and stimulating adenosine monophosphate-activated protein kinase (AMPK) ([@b73-bmb-52-086]). When activated, AMPK phosphorylates key nuclear proteins, thereby regulating metabolic gene expression at the transcriptional level to make energy more available through catabolism in response to the fasted state ([@b74-bmb-52-086]). To note, AMPK activation is just one of the molecular effects of metformin, and it is thought that it may also act through other not fully understood pathways as well ([@b70-bmb-52-086]). In essence, rapamycin and metformin seem to mimic aspects of DR at both the translational and transcriptional level.\n\nLimitations of creating a translational therapeutic derived from these animal interventions include difficulty in diet accountability, ethics of gene editing, pharmaceutical toxicity, and potential side effects. However, understanding the transduction pathways of longevity promoting interventions in animals will be key to ultimately apply and translate these interventions to humans.\n\nTranscriptional variability in aging and longevity\n--------------------------------------------------\n\nAccumulating evidence supports a model where the transcriptome becomes less tightly reagulated throughout the aging process. Indeed, a progressive degradation of transcriptional networks robustness and integrity has been observed during aging in *C. elegans* ([@b75-bmb-52-086]) and in mouse tissues ([@b76-bmb-52-086], [@b77-bmb-52-086]). There is still a debate on the prevalence of increased cell-to-cell transcriptional noise in aging cells. Pioneering studies examined the impact of aging on the cell-to-cell levels of expression of a handful of genes ([@b78-bmb-52-086], [@b79-bmb-52-086]). Whereas increased transcriptional noise was observed in aging mouse cardiomyocytes (11 out of 15 tested genes) ([@b78-bmb-52-086]), no changes in transcriptional noise were detected in hematopoietic stem cells isolated from old mice (6 assayed genes) ([@b79-bmb-52-086]). Importantly, existing technical limitations limited the reach of these studies to few genes and cell types, thus making generalizations difficult.\n\nAs discussed above, recent advances in microfluidics have enabled genome-wide single-cell profiling across diverse cell types at high-resolution ([@b47-bmb-52-086]), and will be key to understand the biological impact of transcriptional noise regulation with age. Indeed, two recent studies have leveraged single-cell RNA-seq to query potential changes in cell-to-cell transcriptional noise with aging. Enge *et al*. profiled cells from the human pancreas, revealing that islet endocrine cells from older donors presented increased transcriptional noise, as well as signs of cell identity drift ([@b50-bmb-52-086]). Modulation of cell-to-cell noise with aging may also be context-specific. Indeed, a recent study in mice found that cell-to-cell transcriptional noise was increased upon immune stimulation in naive and effector memory CD4+ T cells from two subspecies of aging mice ([@b80-bmb-52-086]), whereas the opposite was observed in young cells. Together, these emerging discoveries are consistent with the notion that increased cell-to-cell transcriptional noise may be an important feature of mammalian aging.\n\nREMODELING OF THE TRANSCRIPTIONAL SPLICING LANDSCAPE WITH AGING AND IN LONGEVITY\n================================================================================\n\nAlternative splicing in the regulation of aging and longevity\n-------------------------------------------------------------\n\nA large majority of metazoan genes are transcribed into precursor mRNAs (pre-mRNAs), which are constituted of exons separated by introns, and are thus known as \"multi-exonic genes\". Intronic regions are usually spliced to yield a mature mRNA transcript through the action of a nuclear enzymatic ribonucleoproteic complex, known as the \"spliceosome\", which recognizes specific target sequences found at exon/intron boundaries ([@b81-bmb-52-086]). It is generally thought that most introns are spliced out constitutively. However, many splicing signals can be used in a context-dependent manner, which can lead to the emergence of alternatively spliced isoforms ([@b81-bmb-52-086]).\n\nImportantly, nearly all multi-exonic genes are subject to some degree of alternative splicing in one or more cell types (\\>95% of multi-exon transcripts in human cells), which increases the coding potential and regulatory complexity of metazoan genomes ([@b77-bmb-52-086], [@b82-bmb-52-086]). This mode of regulation depends on regulation by the core slicing machinery, which relies on sequence signals in *cis* and *trans* to either promote or repress splicing at such alternative sites ([@b81-bmb-52-086]). The coupling between multiple regulatory layers and context-driven selection of relevant splice sites results from exquisite regulation by input signals at multiple levels. Recent research advances have revealed that core components of the constitutive and alternative splicing machinery (*i.e*. the spliceosome) are actually involved in fine-tuning of the splicing landscape (reviewed in ([@b81-bmb-52-086])).\n\nThough splicing control may be thought of as a core essential process, emerging evidence suggests that its modulation may have key impacts on health and longevity. Indeed, splicing deregulation has been associated with a number of diseases, including age-related diseases (*e.g*. progeria, Alzheimer's, etc.) ([@b83-bmb-52-086], [@b84-bmb-52-086]). In addition, functional enrichment analyses on age-related human transcriptome remodeling have revealed significant changes in the expression of genes associated to genesets 'mRNA binding', 'mRNA processing', and 'RNA splicing' ([@b85-bmb-52-086]). Profound changes in the alternative splicing landscape with aging have been observed in mouse and human brain regions ([@b86-bmb-52-086]--[@b88-bmb-52-086]), with significant splicing deregulation in \\~40% of genes expressed in the human brain ([@b88-bmb-52-086]). Interestingly, Mazin *et al*. found that aging was associated to increased exon retention in human prefrontal and cerebellar cortex, which may result in increased targeting of the affected transcripts to Nonsense Mediated Decay (NMD) ([@b88-bmb-52-086]). To note, there seems to be little correlation between genes affected at the level of transcription and those affected at the level of slicing in aging ([@b86-bmb-52-086]). Age-related splicing defects are also largely detected in brains from patients with neurodegenerative disorders compared to healthy controls ([@b84-bmb-52-086], [@b87-bmb-52-086]), which is consistent with the observation that these diseases have increased prevalence with age. Changes in the alternative splicing landscape with mammalian brain aging have been associated with carbohydrate metabolism and DNA repair ([@b87-bmb-52-086]), or neuronal function ([@b86-bmb-52-086]). Importantly, a recent study using hepatic tissue from rhesus monkeys subjected to 30% CR showed that alternative splicing was central to the response to nutrient restriction and integral to the associated metabolic re-programming ([@b89-bmb-52-086]).\n\nSupporting the hypothesis that splicing dysfunction may directly impact the aging process, a recent study found that differences in splicing factor spleen expression and changes to alternative splicing were associated with strain lifespan across 6 genetically distinct mouse strains ([@b90-bmb-52-086]). This finding is consistent with the fact that expression levels of splicing regulators hnRNPA1 and A2 were positively associated to increased parental longevity in the InCHIANTI human aging study ([@b90-bmb-52-086]). Importantly, the rewiring of the splicing machinery and potential imbalance of splicing factors may have a direct impact on organismal aging and longevity. Indeed, a recent study has found that, in the worm *C. elegans*, Splicing Factor 1 (SFA-1) plays a key role in modulating longevity upon DR ([@b91-bmb-52-086]). Importantly, stable overexpression of SFA-1 led to increased worm longevity, whereas SFA-1 depletion impacted the activity of components of the TORC1 pathway that are normally subjected to alternative splicing ([@b91-bmb-52-086]).\n\nThus, accumulating evidence suggests that splicing alterations may be a hallmark of aging transcriptome. This progressive loss of the ability to fine-tune gene expression in aging cells is likely to have a profound effect on the physiology of aging cells. Alternative splicing, as regulated by splicing factors, may represent an important regulator or early marker of longevity, which deserves further systematic study.\n\nLong non-coding RNAs in the regulation of aging and in longevity\n----------------------------------------------------------------\n\nThousands of long noncoding RNA (lncRNA) transcripts have now been identified, most of which with yet unknown biological functions ([@b92-bmb-52-086]). Although these transcripts were initially thought to be non-coding (hence their name), accumulating evidence has shown that they in fact harbor a wealth of small open reading frames that can give rise to small functional peptides ([@b93-bmb-52-086], [@b94-bmb-52-086]). In addition, emerging evidence suggests that long non-coding RNAs (lncRNAs) may act as aging biomarkers, and perhaps even influence the aging process ([@b95-bmb-52-086]). Emerging studies have focused on exploring the roles of specific lncRNAs in aging of different cell types. Specifically, the *H19* lncRNA was recently demonstrated to be involved in vascular endothelial cell aging through inhibition of the Stat3 signaling pathway ([@b96-bmb-52-086]). In addition, control of *H19* levels is key to maintain adult HSC quiescence ([@b97-bmb-52-086]), thus preventing a premature depletion of the stem cell pool. Moreover, the lncRNA *FLJ46906* may regulate the expression of various aging-associated genes (*e.g. TGFB1*, *CXCL8*, *IL1B*, *IL6*, and *ELN*) through modulation of the NF-\u03baB and AP-1 transcription factors in human fibroblasts ([@b98-bmb-52-086]). Thus, because of the abundance of lncRNAs in vertebrate genomes, it is likely that many of them may play important roles in adult tissue homeostasis and regulation of the aging process. Future studies are needed to shed light on lncRNAs as candidate regulators of the aging transcriptional landscape.\n\nCircular RNAs: exciting new players of the aging transcriptome?\n---------------------------------------------------------------\n\nCircular RNAs (circRNAs), which are covalently closed circular RNAs without a free 5\u2032- or 3\u2032-end, were first identified in viruses by electron microscopy ([@b99-bmb-52-086]). A few years later, they were found in eukaryotic cells, although they were then dismissed as likely insignificant byproducts ([@b100-bmb-52-086]). With the recent emergence of ultra-high-throughput Next Generation Sequencing (NGS), further investigation into transcriptomic landscapes has become possible on an unprecedented scale. Indeed, deep RNA-seq in human cells revealed that circRNAs may even sometimes represent the major transcript isoforms of many genes ([@b101-bmb-52-086]). Indeed, circRNAs are abundant, and largely expressed in a tissue-specific manner ([@b102-bmb-52-086], [@b103-bmb-52-086]). Moreover, recent studies have identified circRNAs across tissues and species (*e.g*. human, mouse, Drosophila, *C. elegans*, Tree Shrew, *etc*.) ([@b101-bmb-52-086], [@b104-bmb-52-086]--[@b106-bmb-52-086]). Circular RNAs are thought to be generated through an original alternative splicing mechanism, known as \"back-splicing\", where the spliceosome covalently bonds the downstream 5\u2032 donor site of an exon to the upstream 3\u2032 acceptor site of another (upstream) exon ([@b107-bmb-52-086]), which may be a promoter by the proximity of transposable element sequences ([@b102-bmb-52-086]). Since the rediscovery of circRNAs in mammalian cells, the cellular roles of these RNAs have been under scrutiny. Studies have found evidence that circRNAs can act as 'microRNA sponges' ([@b108-bmb-52-086]), promoting microRNA-directed RNA degradation ([@b109-bmb-52-086]), trapping RNA-binding proteins ([@b110-bmb-52-086]), or even act as templates for protein translation through the presence of Internal Ribosome Entry Sites (IRES) ([@b111-bmb-52-086]--[@b113-bmb-52-086]). For example, one of the best-studied circRNA species to date is derived from antisense transcription from the *CDR1* locus, and known as *CDR1as*, which harbors \\~70 conserved miR-7 target sites, and can function as a miRNA sponge ([@b108-bmb-52-086]). Because of the ability of some circRNAs to act as miRNA sponges, they are now widely thought to constitute a class of competing endogenous RNA ([@b108-bmb-52-086]).\n\nThis emerging class of RNAs has recently been gaining traction in the aging field, notably because many circRNAs have been observed to be differentially expressed in aging neural tissue across species ([@b104-bmb-52-086], [@b106-bmb-52-086], [@b114-bmb-52-086], [@b115-bmb-52-086]). In *Drosophila*, 262 circRNAs expressed in the Central Nervous Systems (CNS) were found to be upregulated between 1 and 20 days of life ([@b104-bmb-52-086]). Accumulation of circRNA in neural tissue was also observed in mice ([@b115-bmb-52-086]), with \\~5% of brain circRNAs significantly up-regulated with age whereas only \\~1% were significantly downregulated with age (1 *vs*. 22 months) ([@b115-bmb-52-086]), although the young control of that study may have still been showing transcriptional signatures of a developing brain. Those results are consistent previous observations did in *Drosophila*, but, the young mice used can be considered as pre-developmental mice. All the changes in circRNA levels may be due to development and not aging at this age. Future studies would require young post-developmental mice for any comparisons. A recent study found that a global, age-associated, accumulation of circRNAs can also be observed in *C. elegans* ([@b105-bmb-52-086]). About 90% (1052/1166) of measured circRNAs were observed to be up-regulated at day 10 of life compared the L4 larval stage, whereas 37 circRNAs were observed to be significantly down-regulated ([@b105-bmb-52-086]). Cortes-Lopez *et al*. proposed that this accumulation of circRNAs may be the result of their increased stability and the fact that most of *C. elegans*' cells are post-mitotic cells ([@b105-bmb-52-086]). It is important to note that, so far, circRNAs have only been observed to accumulate in adult *C. elegans* and mammalian brains, both of which are enriched in post-mitotic cells. However, those studies do not include spike-in to control the analysis and to conclude accumulation of circRNAs with aging ([@b116-bmb-52-086]).\n\nWhile functional studies of the roles of circRNAs are still lacking in the context of aging research, the downregulation of circRNA *circPVT1* was found to be critical for the establishment of cell senescence in human fibroblasts, through its interaction with microRNA let-7 ([@b117-bmb-52-086]). Although this is not strictly representative of organismal aging, senescent cells do accumulate over the course of lifespan ([@b118-bmb-52-086]), and clearance of these cells leads to improved health and lifespan ([@b119-bmb-52-086]). More generally, the functional impact of circRNA regulation on cells and organisms is still very much an open field of study. Importantly, emerging work suggest that they may be key for normal brain function, since the specific depletion of *Cdr1as* in mice was associated with neuropsychiatric dysfunction ([@b120-bmb-52-086]). The study of circRNA regulation and its impact on organismal physiology is still very much in its infancy, and additional gain and loss-of-function studies for circRNA genes will be needed in adult organisms to shed light on their potential role in health and lifespan.\n\nTRANSPOSABLE ELEMENTS AND ENDOGENOUS RETROVIRUSES: THE SILENT KILLERS WITHIN THE EUKARYOTIC GENOME?\n===================================================================================================\n\nGenomic instability is characterized by an accumulation of somatic mutations, including *de novo* insertions of transposable elements (TEs) ([@b121-bmb-52-086]). TEs, which have also been dubbed \"jumping genes,\" are a type of repetitive DNA that has the capacity to transpose from one genomic location to another ([@b122-bmb-52-086]). TEs constitute anywhere from 3% in *Saccharomyces cerevisiae* to 77%, in Rana esculenta, of eukaryotic genomes ([@b123-bmb-52-086]). In vertebrates specifically, the proportion ranges from 6% in the green spotted pufferfish to 55% in the zebrafish ([@b124-bmb-52-086]). These elements are highly diverse in terms of DNA sequences, mobilization mechanisms, and regulatory mechanisms (reviewed in ([@b125-bmb-52-086], [@b126-bmb-52-086]). Two main classes of transposons have been described: (i) Class I transposons, or RNA-mediated transposons, and (ii) Class II transposons, or DNA-mediated transposons ([@b127-bmb-52-086]). Class I TEs transpose through a \"copy and paste\" mechanism, generating RNA intermediates that can be reverse transcribed and integrated into the genome at a new site ([@b127-bmb-52-086]). In contrast, Class II transposons rely on enzymes that \"cut\" the transposon DNA sequences out of their current loci, and \"paste\" them elsewhere in the genome ([@b127-bmb-52-086]). Because of their mobility, TEs can act as mutagens and have the potential to disrupt genes by inserting themselves into coding sequences and to change gene expression through insertion into regulatory sequences ([@b128-bmb-52-086]). From a bioinformatics perspective, the repetitive nature of TEs have complicated genomic analyses, as short-read NGS technologies have limited ability to distinguish different insertions of the same element family ([@b129-bmb-52-086]). Though \\> 44,000 sequences have been catalogued on Repbase, a database for repetitive elements ([@b130-bmb-52-086]), it is likely that many more remain to be characterized. Fortunately, the emerging availability and reduced cost of long-read sequencing technologies (*e.g*. Pacific Biosciences SMRT sequencing ([@b131-bmb-52-086]), Oxford Nanopore ([@b132-bmb-52-086])) should help overcome these technical hurdles in the coming years ([@b133-bmb-52-086], [@b134-bmb-52-086]). Ultimately, the high sequence and functional variability of transposons means that much remains to be discovered in terms of their mechanisms, roles, and biological consequences, and accumulating evidence suggests that they play a key role during aging and in the pathogenesis of age-related diseases ([Table 4](#t4-bmb-52-086){ref-type=\"table\"}).\n\nTransposon landscapes in traditional and emerging model systems\n---------------------------------------------------------------\n\nA number of invertebrate animal models, including *C. elegans* and *D. melanogaster*, and some vertebrate models, such as the mouse and the zebrafish, have been used in aging and transposable element research ([@b135-bmb-52-086]). Noteworthy, TEs vary in type and in frequency across different organisms. For example, 12% of the *C. elegans* genome is composed of TEs ([@b136-bmb-52-086]), most of which (\\~95%) are DNA transposons ([@b124-bmb-52-086]). In contrast, TEs encompass \\~20% of the *D. melanogaster* genome ([@b137-bmb-52-086]), and LTRs from the gypsy, copia, and Pao families are the most abundant ([@b138-bmb-52-086]). In humans, \\~50% of the genome is repetitive, with the majority of human TEs being LINE-1 elements (\\~17% of the genome) ([@b139-bmb-52-086]). Although invertebrate models have been widely used in studies on transposon and aging processes due to, among other factors, their short lifespans, this has come at the expense of knowledge concerning the relationship between these processes and vertebrate-specific genes, organs and tissues, and physiological processes. For example, methylation status of transposon loci in human blood has been identified as a predictor of menopausal age ([@b140-bmb-52-086]), and transposons have also been shown to induce responses mediated by interferons ([@b141-bmb-52-086]), which are modulators of innate and adaptive immunity.\n\nThough aging studies can be, and have been, carried out in vertebrate models such as the mouse and the zebrafish, their relatively long lifespans in the laboratory (3--4 and 5 years, respectively) make aging studies costly and time-consuming. Consequently, there has been a need for a tractable short-lived vertebrate model to address these limits. The African turquoise killifish (*Nothobranchius furzeri*) is a naturally short-lived vertebrate (\\~0.5 year), which was first found and collected from ephemeral ponds in Mozambique and Zimbabwe ([@b142-bmb-52-086]--[@b145-bmb-52-086]). Other advantages of this model are the availability of several wild-derived strains ([@b146-bmb-52-086]--[@b148-bmb-52-086]), an annotated draft genome sequence ([@b142-bmb-52-086], [@b149-bmb-52-086]) and a rapid genome-editing toolkit ([@b143-bmb-52-086]). In addition, despite a compressed lifespan, the turquoise killifish shows classical phenotypes of vertebrate aging (*e.g*. cognitive impairment, histological changes, telomere shortening, and mitochondrial dysfunction) ([@b150-bmb-52-086]--[@b153-bmb-52-086]). Relevant to the study of TEs in vertebrate aging, 45--65% of the turquoise killifish genome is composed of TEs, a proportion which is similar to that of the human genome ([@b154-bmb-52-086]). Moreover, the elements seem to be transcriptionally active, compatible with the notion of active transposition across tissues in this species ([@b142-bmb-52-086]). Given all of these characteristics, we anticipate that the African turquoise killifish will be an important model to understand the mechanisms and impact of TE regulation in aging and age-related diseases.\n\nMechanisms and breakdown of TE regulation throughout lifespan\n-------------------------------------------------------------\n\nGiven the mutagenic potential of unfettered transposition, cells have evolved layers of pre- and post-transcriptional mechanisms to restrain transposon activity in the germline and somatic tissues ([@b125-bmb-52-086], [@b126-bmb-52-086]). Known pre-transcriptional mechanisms mainly include DNA methylation and TE heterochromatinization ([@b125-bmb-52-086], [@b155-bmb-52-086]). For instance, in female *Drosophila* head tissues and fat bodies, Wood *et al*. found that overexpression of genes encoding chromatin modifying enzymes *Sir2* and *Su(var)3-9* led to repression of the age-related increase in TE expression ([@b156-bmb-52-086]). *Sir2* is an NAD-dependent deacetylase involved in gene silencing at heterochromatic regions in yeast and flies ([@b157-bmb-52-086]--[@b159-bmb-52-086]), while *Su(var)3-9* catalyzes the methylation of Lysine-9 of histone H3 (H3K9) to repress transcription ([@b160-bmb-52-086]). The study also found that overexpression of *Su(var)3-9* was associated with increased lifespan ([@b156-bmb-52-086]). Another study noted a global increase in heterochromatin in the livers of old mice (based on DNase I sensitivity assay and comet tail assay), but associated increased expression of TEs with relaxed heterochromatic regions ([@b161-bmb-52-086]). Although the global increase in heterochromatin in this specific context runs counter to the \"loss of heterochromatin\" model of aging ([@b162-bmb-52-086]), TE activation has, in contrast, been shown to be recurrently associated with heterochromatin loss in *C. elegans*, *D. melanogaster*, mice, and mammalian cells ([@b163-bmb-52-086], [@b164-bmb-52-086]). Additionally, in senescent human fibroblasts, chromatin accessibility was shown to increase at TE insertion loci, thus promoting TE transcription and active transposition ([@b161-bmb-52-086]).\n\nCurrent research is partially focused on identifying novel genetic regulators of transposon activity. A genome-wide CRISPR-Cas9 screen in the K562 and HeLa cancer cell lines was recently performed in order to identify novel regulators of LINE-1 retrotransposition unbiasedly ([@b165-bmb-52-086]). However, how these new TE regulators respond to aging and longevity intervention remains unknown. Accumulating evidence also highlights the interaction between the interferon response and TE activity. Type I interferons (IFNs) are a family of cytokines that modulate immunity against viruses and intracellular pathogens, notably by inducing genes that promote an \"antiviral state\", sensitizing these cells to apoptosis, and by activating immune cells that clear infected cells ([@b166-bmb-52-086]). Importantly, Yu *et al*. showed that LINE-1 stimulates IFN-\u03b2 expression in *in vitro* human and mouse cell models ([@b141-bmb-52-086]). Conversely, treatment with IFN-\u03b2 could repress endogenous LINE-1 retrotransposition, with endogenous IFN-\u03b2 signaling counteracting replication of LINE-1 elements ([@b141-bmb-52-086]). An additional study looking for regulators of cell survival in human acute myeloid leukemia cell lines found that loss of heterochromatin regulator SETDB1 associated with lower cell viability, induction of TEs, and induction of viral response genes, including interferon stimulated genes ([@b167-bmb-52-086]).\n\nPost-transcriptional mechanisms of regulation rely on the degradation of TE RNAs or inhibition of their translation. These rely on siRNA, miRNA, and piRNA pathways ([@b125-bmb-52-086], [@b155-bmb-52-086]). Work by Wood *et al*. also assessed the relationship between siRNA activity and TE levels in *Drosophila*. They observed a repression of the age-associated TE upregulation in mutant flies overexpressing *Dicer-2*, an enzyme necessary for siRNA processing, or in flies hypomorphic for the *Adar* gene, which encodes the main RNA-editing enzyme and functions as a negative regulator of the siRNA pathway ([@b156-bmb-52-086]). Importantly, the presence of Alu elements is a major factor for selecting RNA-editing targets by ADAR in humans cells ([@b168-bmb-52-086]). In another *Drosophila* study focusing on the central nervous system, increased TE expression and transposition events were observed with aging using a *gypsy* GFP reporter (*gypsy*-TRAP reporter) ([@b169-bmb-52-086]). Interestingly, mutation of *Ago2*, another gene involved in RNAi pathways, led to increased TE activity and shortened lifespan ([@b169-bmb-52-086]). To note, although associations between RNAi pathways and TE activity have been noted, the results should be interpreted with caution, as dysregulation of RNAi pathways are likely to lead to a general misregulation of many downstream processes. Canonically, siRNA pathways and piRNA pathways have been thought to be primarily restricted to regulatory roles in somatic and germ line cells, respectively ([@b138-bmb-52-086]). Emerging evidence, however, suggests that piRNA mechanisms may also be active in somatic cells ([@b170-bmb-52-086]). In one type of mushroom body neuron in *Drosophila*, suppression of the piRNA proteins Aubergine and Argonaute3 was associated with elevated TE expression ([@b171-bmb-52-086]), and another study proposed that piRNA components may play a role in alternative splicing of transposable elements in both soma and germ line cells of the fly ([@b172-bmb-52-086]). A final form of regulation which remains loosely characterized involves tRNA-derived fragments which have been shown to target retrotransposon primer binding sites in mouse embryonic and trophoblast stem cells and may, more generally, inhibit transcription or induce RNAi pathways ([@b173-bmb-52-086]). Though multiple layers of native regulatory mechanisms for transposon control exist, research has also been focused on identifying dietary and pharmaceutical interventions that can supplement cellular mechanisms.\n\nTransposon activity modulation by dietary and chemical interventions\n--------------------------------------------------------------------\n\nOne of the focuses of ongoing research is to identify dietary interventions or novel drugs that inhibit transposon activity and determining what impact these have on lifespan and health. In female *Drosophila* heads and fat bodies, DR suppressed the age-related increase in TE expression that was observed in the high calorie (HC) diet control group ([@b156-bmb-52-086]). Additionally, DR seemed to counteract the increase in expression of genes located in constitutive heterochromatin regions that occurred with age in HC-fed flies ([@b156-bmb-52-086]). A high-fat diet similarly promoted TE expression in the livers of middle-aged male mice, and the authors suggest that CR may inhibit these effects ([@b174-bmb-52-086]). Interestingly, knockdown of the chromatin-remodeler *Chd1* led to similar transposon activation as the high-fat diet and aging conditions, supporting a key role for chromatin-level regulation of TE activity ([@b174-bmb-52-086]). Pharmacologically, Wood *et al*. demonstrated that treatment of flies with the reverse transcriptase inhibitor lamivudine repressed retrotransposition in *Drosophila* fat bodies and increased the lifespan of *Dicer-2* mutants ([@b156-bmb-52-086]). Looking forward, the effects of this drug, and other anti-retroviral drugs, should be assessed when administered at different times in an organism's life to determine which dosing periods contribute most towards improving organism function and lifespan. In particular, embryogenesis may be an informative period to test, given that transposition events occurring during this time are likely to lead to large bulks of mutated cells in the body, thus leading to somatic chimerism ([@b138-bmb-52-086]). Importantly, since both avenues for TE regulation (*i.e*. inhibiting expression and inhibiting transposition) may not necessarily lead to the same outcome, both avenues should be explored in the treatment of aging and age-related diseases. For example, inhibition of LINE-1 with a reverse transcriptase inhibitor does not have the same developmental consequences for mouse embryonic stem cells as does a genetic knockdown ([@b175-bmb-52-086]). Since accumulating evidence suggests that transposon reactivation occurs with and in various age-related diseases ([Table 4](#t4-bmb-52-086){ref-type=\"table\"}), it will ultimately be important to identify robust methods to modulate TE activation *in vivo* throughout life.\n\nTransposons: a blessing and a curse?\n------------------------------------\n\nThe idea that TEs play both positive and negative roles, based on factors such as the time of activation and the integration site within the host genome, has been gaining traction ([Table 4](#t4-bmb-52-086){ref-type=\"table\"}). Historically, TEs have been seen as 'genetic parasites' and 'selfish DNA' elements ([@b176-bmb-52-086]). Though some TEs can be described as such, others are non-functional and thus neutral, or have helped shape new functions at the cellular and organismal level ([@b129-bmb-52-086]). Since TEs contain regulatory elements that differentially respond to environmental and stress cues, TE insertions may help dynamic rewiring of gene regulatory networks, and generate novel spatiotemporal gene expression programs ([@b126-bmb-52-086]). Examples to illustrate the potential multi-pronged effects of transposons concern the human Endogenous Retrovirus (HERV) family of transposons. On the one hand, the proteins syncytin-1 and syncytin-2 -- encoded by the HERV-W and HERV-FRD *env* genes -- are required for fusion of trophoblast giant cells to form the syncytiotrophoblast, a key step for placental formation, and contribute to fetal immune tolerance ([@b177-bmb-52-086]). On the other hand, the viral properties of ERVs render them capable of activating the immune system, and this may contribute to the development of autoimmunity and other pathologies ([@b178-bmb-52-086], [@b179-bmb-52-086]).\n\nThus, the ERV families of transposons, and perhaps others, may have antagonistic pleiotropic effects, playing important roles in early development, but contributing to pathogenesis and functional decline post-development. A number of associations between transposons and different age-related disease states have been identified (see [Table 4](#t4-bmb-52-086){ref-type=\"table\"}). In an intriguing example, Guo *et al*. identified an association between global TE levels and Tau pathological burden ([@b180-bmb-52-086]). Moreover, they found that Tau expression induces TE expression in a *Drosophila* model ([@b180-bmb-52-086]), raising the possibility that TEs may be causative agents or drivers of Alzheimer's disease and neurodegeneration. However, the mechanisms underlying this link are still unclear, and will deserve further consideration. Similarly, functionally abnormal TAR DNA-binding protein 43 (TDP-43) is observed in many cases of amyotrophic lateral sclerosis (ALS) and some cases of frontotemporal lobar degeneration (FLTD) ([@b181-bmb-52-086]). Experimentally, TDP-43 has been shown to bind transposable elements in mouse and rat models, differentially bind TEs in healthy and FTLD patients ([@b182-bmb-52-086]), and induce expression of a panel of transposons when human TDP-43 is expressed in Drosophila and mouse models ([@b182-bmb-52-086], [@b183-bmb-52-086]). The data in Drosophila suggests hTDP-43 toxicity may stem from *gypsy* ERV activating DNA damage-mediated programmed cell death ([@b183-bmb-52-086]), but whether this mechanism is conserved and whether it is the sole driving mechanism remains unclear.\n\nFuture work on the role of TE activity in aging and age-related diseases should seek to clarify this relationship by generating precise genomic tools. Ultimately, the field will need to address whether transposon activation is a byproduct or a driver of aging and disease and, if the latter is true, uncover the mechanisms by which this occurs.\n\nTHE EMERGING IMPACT OF SMALL ncRNAs IN AGING AND LONGEVITY\n==========================================================\n\nmiRNAs in the regulation of aging and longevity\n-----------------------------------------------\n\nMicroRNA (miRNA) are a class of endogenous eukaryotic small non-coding RNAs, \\~22 nucleotides long, that play a key role in RNA silencing and post-transcriptional regulation of gene expression ([@b184-bmb-52-086]). miRNAs are abundant and have been proposed to target, and potentially regulate, \\~60% of mammalian genes ([@b185-bmb-52-086]). Accumulating evidence suggests that remodeling of the miRNA expression landscape may be a biomarker, or even a driver, of the aging process ([@b95-bmb-52-086]).\n\nSeveral pioneering studies have identified age-related changes in the miRNA transcriptional landscape across cells and tissues from various evolutionary taxa (*i.e*. worms, mouse, rats, human, macaque, African turquoise killifish, *etc*.) ([@b186-bmb-52-086]--[@b198-bmb-52-086]). Interestingly, miRNA with age-specific changes in *C. elegans* were shown to either promote or repress normal lifespan and stress resistance in the adult organism ([@b186-bmb-52-086], [@b199-bmb-52-086]). Further, expression levels of specific miRNA early in life can predict *C. elegans* longevity ([@b187-bmb-52-086]), supporting the notion that these changes may be responsible for aging phenotypes at least partly. Maturation of miRNA sequences from precursors requires *Dicer* expression ([@b200-bmb-52-086]). Intriguingly, a fat-specific *Dicer* ablation was found to lead to an accelerated aging phenotype in mice, and counteracted positive effect of DR ([@b201-bmb-52-086]). Moreover, transfection of age downregulated miRNAs into 18-month-old mice led to a partial rejuvenation of liver expression profiles, at least in terms of aging markers (*e.g. p16INK4a*) ([@b190-bmb-52-086]).\n\nAn interesting case is that of the miR-29 microRNA family. Indeed, members of this family were found to be upregulated with age in various tissues, including brain, muscle, skin, vascular tissue, heart, liver, lung and kidney ([@b191-bmb-52-086], [@b193-bmb-52-086]--[@b198-bmb-52-086], [@b202-bmb-52-086]). Age-related accumulation of DNA-damage has been proposed to activate miR-29 and p53 ([@b195-bmb-52-086]). It has also been proposed that miR-29 may be involved in a compensatory response limiting an age-related neuronal iron accumulation ([@b196-bmb-52-086]). Although miR-29 upregulation with aging seems ubiquitous, conditional knockdown for miR-29 in mouse brains led to an increase in male lifespan, while female lifespan was decreased ([@b202-bmb-52-086]), suggesting a complex interaction between miR-29 functions in health and lifespan and other factors (*e.g*. sex). Interestingly, a recent study found that TGF-\u03b2 could signal via miR-29 to promote repress the translation of the main H4K20me3 methyltransferases ([@b203-bmb-52-086]). Deficit in cellular H4K20me3, a chromatin mark which plays an important for DNA-damage repair, led to cellular senescence, and might contribute to mouse cardiac aging *in vivo* ([@b203-bmb-52-086]). These observations support the notion that a complex interplay of signaling between non-coding RNAs and the epigenome may be driving aspects of the aging process.\n\nA potential role for tRNAs in aging and longevity regulation?\n-------------------------------------------------------------\n\nThe genetic code is degenerate, with the set of 20 amino acids encoded by 61 base triplets or codons ([@b204-bmb-52-086]). Hence, multiple synonymous codons can be used in coding sequences, giving rise to the inclusion of the same amino acid in the final protein product. Traditional studies of disease genetics and evolutionary selection have focused on non-synonymous variants, which lead to a change in protein sequence and, presumably, protein function. However, accruing evidence now supports the idea that codon choices, while not affecting the protein sequence, can impact translation by interacting with the tRNA pool ([@b205-bmb-52-086]), thus functionally modulating final protein products ([@b206-bmb-52-086], [@b207-bmb-52-086]). Recent examples of the consequences of codon usage biases include modulation of the transcription rate ([@b208-bmb-52-086]), of translation rate ([@b209-bmb-52-086]) or selection against protein misfolding ([@b210-bmb-52-086]). In turn, these processes have been involved in the pathogenesis of neurodegenerative diseases ([@b211-bmb-52-086]--[@b213-bmb-52-086]), and their modulation may be involved in the age-related proteostasis collapse ([@b6-bmb-52-086]). Almost 50 years ago, researchers theorized that changes in the tRNA pool with life or development may result in an otherwise unnoticed role for codon usage bias in aging ([@b214-bmb-52-086]). However, the importance of changes in tRNA pools, and their interactions with codon usage, on aging and lifespan is still largely unknown.\n\nAccumulating evidence showing time and tissue variability of the tRNA pool have led to rekindled interest in this intriguing hypothesis ([@b215-bmb-52-086], [@b216-bmb-52-086]). Because of the generally accepted importance of translation regulation and proteostasis with aging ([@b6-bmb-52-086]), the impact of alterations of the tRNA pool on ribosomal protein synthesis and co-translational folding, interacting with differences in codon usage, may have profound impact on the proteome throughout life. Indeed, proliferation-enriched and differentiation-enriched tRNAs preferentially have anticodons that are enriched to be represented respectively among cell-autonomous vs. intercellular communication genes, suggesting exquisite coordination between transcriptional and translational programs ([@b217-bmb-52-086]). Interestingly, tRNA-Leu overexpression can enhance phosphorylation of S6-kinase under amino acids starvation condition independent of mTOR signaling ([@b218-bmb-52-086]). This observation is consistent with the notion that tRNAs may also play more general regulatory functions.\n\nRecent evidence has revealed the existence of a new class of small non-coding regulatory RNAs derived from tRNAs, known as tRFs (*i.e*. tRNA-derived fragments) ([@b219-bmb-52-086]). Although the biological roles of tRFs are still poorly understood, they have been suggested to play roles in regulating cell proliferation, gene expression and RNA processing, priming viral reverse transcriptases, and modulating the DNA damage response, tumor suppression, and neurodegeneration ([@b219-bmb-52-086]). For instance, inhibition of a tRF species derived from the 3\u2032 half of tRNA-Leu (CAG) can induce apoptosis and enhance the translation of two ribosomal protein (RPS28 and RPS15) ([@b220-bmb-52-086]).\n\nThe expression of tRNA genes is under the control of a specialized polymerase, RNA polymerase III (RNA Pol III) ([@b221-bmb-52-086]). Transcription of specific tRNA genes is thus likely to be influenced by remodeling of the chromatin landscape, notably during aging ([@b1-bmb-52-086]), which may disrupt the homeostasis of tRNA gene pools. Although RNA Pol III genes are generally thought of as 'housekeeping genes', recent evidence suggests that RNA Pol III target selection must be regulated at least in part (reviewed in ([@b222-bmb-52-086])). Indeed, while a large majority of RNA Pol III target genes are shared across cell types and tissues, private targets exist and could be involved in tissue- or contextspecific responses ([@b223-bmb-52-086]). Interestingly, a recent study has shown that a reduction in RNA Pol III levels is associated to increased longevity downstream of TOR activity in yeast, worms and flies ([@b224-bmb-52-086]), supporting the idea that RNA Pol III (and its targets, including tRNA genes) may be a crucial mediator of the nutrient-signaling longevity pathway.\n\nA systematic assessment of tRNA pools, derived tRFs, as well as RNA Pol III targets with mammalian aging will be crucial to understand the role of these phenomena on health and lifespan.\n\nGENERAL DISCUSSION AND IMPLICATIONS\n===================================\n\nThis review has painted a broad picture of the current knowledge in various aspects of transcriptional regulation in aging and longevity models. Understanding gene regulation and misregulation with aging will be of great use in order to help mitigate or reverse aging-related functional decline.\n\nIn addition to humans, sex differences in lifespan have been observed across taxa ([@b10-bmb-52-086]), and mammalian females across species are also generally longer-lived than males, including pilot whales, lions, red deer, prairie dogs, monkeys and apes ([@b10-bmb-52-086]). In most reported cases, female laboratory rats live longer than males ([@b10-bmb-52-086]). Although the existence of a similar female advantage for lifespan in laboratory mice is still debated ([@b10-bmb-52-086]), in standardized husbandry conditions developed for the NIA Interventions Testing Program (ITP), female individuals consistently outlive males at 3 independent sites ([@b10-bmb-52-086]), suggesting that the human female advantage may be recapitulated in laboratory mice in controlled conditions. Many health parameters differ between male and female mice with aging ([@b225-bmb-52-086]). Several reports have indicated that adult stem cell populations (*i.e*. HSCs, NSCs, MuSCs) show increased self-renewal and regenerative capacity in females compared to males ([@b226-bmb-52-086]). Moreover, females exhibit increased abilities for wound healing and liver regeneration ([@b226-bmb-52-086]). Surprisingly, interventions that successfully extend the health- and lifespan of laboratory mice often do so in a sex-dimorphic manner ([@b10-bmb-52-086], [@b55-bmb-52-086], [@b227-bmb-52-086]). For instance, treatment with rapamycin preferentially extends female lifespan, whereas treatment with 17-\u03b1-estradiol only extends male lifespan ([@b55-bmb-52-086]). More surprisingly, genetic manipulations can also extend lifespan in a sex-dimorphic manner, leading to either greater lifespan extension in females vs. males (*e.g. Igf1r* haploinsufficiency), or to lifespan extension exclusively in males (*e.g. Sirt6* overexpression) ([@b10-bmb-52-086], [@b55-bmb-52-086]). Though it has been proposed that differences in the bioavailability or metabolism of longevity-promoting compounds may be responsible for sex-dimorphic longevity effects, the existence of sex-dimorphic longevity effects upon genetic manipulations suggests that other underlying mechanisms are at work. Remarkably, thousands of genes can be regulated in a sex-dimorphic manner across a range of youthful healthy mouse and human somatic tissues (*e.g*. brain, liver, heart, muscle) ([@b228-bmb-52-086]--[@b231-bmb-52-086]). Genes expressed in a sex-dimorphic manner are located on autosomes as well as on sex chromosomes ([@b228-bmb-52-086], [@b230-bmb-52-086]--[@b232-bmb-52-086]), and many of these genes are not directly targeted by sex hormones ([@b233-bmb-52-086]). Functional enrichment analyses of sex-dimorphic gene regulation have identified differentially enriched pathways between male and female tissue transcriptomes, including immune response, oxidoreductase activity, and lipid metabolism ([@b232-bmb-52-086]). In addition, genes expressed in a sex-dimorphic manner in the human brain were proposed to act as mediators of stress susceptibility and depressive symptoms ([@b234-bmb-52-086]), consistent with the idea that sex-dimorphic expression can broadly impact human health and physiology. Studies in *Drosophila* have revealed that the sex-dimorphism in the expression of the mitochondrial Lon protease mediates sex- and age-specific adaptation to oxidative stress, and the sex-dimorphic expression of this protease may be conserved in mice ([@b235-bmb-52-086]). Interestingly, DR leads to a feminization of the gene expression profile of male mouse livers (*i.e*. rendering the gene expression profile of the organ more similar to that of females) ([@b236-bmb-52-086]). These observations raise the intriguing possibility that sex-dimorphic gene expression may play a key role in aging and response to longevity interventions. However, despite accumulating evidence of widespread sex-dimorphism in aging and lifespan across species as well as in mechanisms of gene regulation, few studies have investigated age-related sex-dimorphisms at the molecular level. Thus, systematically understanding the transcriptional underpinnings of sex differences in aging and longevity will be crucial to develop therapeutic strategies to slow down age-related functional decline and diseases which work in both males and females.\n\nAn important caveat to note for all studies of aging and longevity-associated transcriptional remodeling is that the penetrance of such transcriptional changes to cellular and organismal phenotypes may be limited ([@b237-bmb-52-086]). However, proteomics still has a lower sensitivity threshold and higher technical noise than RNA profiling. Thus, it will be important to systematically combine transcriptional mapping approaches to \"proteome\" and \"translatome\" mapping, notably through ribosome-profiling and mass spectrometry, to understand the relationship between mRNA levels and translated proteins ([@b238-bmb-52-086]).\n\nFinally, transcriptional changes are the end-result of changes in upstream regulators, including chromatin remodeling ([@b1-bmb-52-086], [@b2-bmb-52-086]) and modulation of transcription factor activity (*e.g*. FOXOs, NRF2, *etc*.) ([@b239-bmb-52-086], [@b240-bmb-52-086]). The accumulating evidence showing that longevity-interventions can slow-down or reverse aspects of transcriptional aging suggests that aging can be modulated at the molecular level by environmental cues and genetic factors. Understanding the regulatory mechanisms leading to transcriptional defects with aging will be key in the design of therapeutic strategies to slow-down age-related decine, and delay the onset of age-related diseases.\n\nThe authors apologize to those whose work could not be included due to space constraints. J.B. is supported by NIA T32AG05237. B.A.B is supported by NIA R00AG049934 and the Hanson-Thorell family scholarship.\n\n**CONFLICTS OF INTEREST**\n\nThe authors have no conflicting interests.\n\n###### \n\nAging signatures in vertebrate tissues\n\n  Tissue              Upregulated pathways    Downregulated pathways                    Species     Profiling method   Reference\n  ------------------- ----------------------- ----------------------------------------- ----------- ------------------ -------------------------------------------------------------\n  Liver               Fibrosis                Cell Cycle                                Mouse       Microarray         ([@b33-bmb-52-086], [@b241-bmb-52-086])\n                      Immune Response         DNA Replication                                       RNA-seq            ([@b29-bmb-52-086], [@b242-bmb-52-086], [@b243-bmb-52-086])\n                      Inflammation            Drug Catabolic Processes                  Killifish   RNA-seq            ([@b244-bmb-52-086])\n                      Lipid Metabolism        Metabolic Pathways                                                       \n                      Protein Turnover        Oxidative Demethylation                   Rat         RNA-seq            ([@b238-bmb-52-086])\n                      Stress Response         Protein Synthesis                                                        \n                                              RNA Transport                                                            \n  Heart               Caspase Family          Electron Transport Chain                  Human       RNA-seq            ([@b26-bmb-52-086])\n                      Chemokines              Mitochondrion                             Mouse       Microarray         ([@b245-bmb-52-086])\n                      Immune Response         TCA Cycle                                                                \n                      Lysosome                Ubiquitin-dependent Catabolic Processes                                  \n                      s100 Family                                                                                      \n  Pancreatic Islets   Extracellular Matrix    Cell-cell Signaling                       Rat         Microarray         ([@b25-bmb-52-086])\n                      Global Methylation      Metabolic Pathways                        Mouse       WGBS               ([@b246-bmb-52-086])\n                      Immune Response         Neuron Recognition                                    scRNA-seq          ([@b247-bmb-52-086])\n                      Inflammatory Response   Nutrient Sensing                          Human       RNA-seq            ([@b49-bmb-52-086])\n                      Oxidative Stress        Proliferation                                         ChIP-seq           \n                      Wound Healing                                                                 scRNA-seq          ([@b50-bmb-52-086])\n  Brain               *Cerebellum*                                                      Mouse       Microarray         ([@b24-bmb-52-086])\n                      Complement System       Energy Metabolism                                                        \n                      Inflammatory Response   Growth/Trophic Factors                                                   \n                      Stress Response         Inflammatory Response                                                    \n                                              Protein Synthesis                                                        \n                      *Whole Brain*                                                     Killifish   RNA-seq            ([@b248-bmb-52-086], [@b244-bmb-52-086])\n                      Apoptosis               Collagen                                  Rat         RNA-seq            ([@b238-bmb-52-086])\n                      Complement System       Metabolic Pathways                                                       \n                      Immune Response         Mitochondrion                                                            \n                      Inflammation            Protein Synthesis                                                        \n                      Lysosome                Ribosomal Pathways                                                       \n                      Ribosome                RNA Transport                                                            \n                      Stress Response         Signal Transduction                                                      \n  Kidney              Extracellular Matrix    Stress Response                           Human       Microarray         ([@b28-bmb-52-086])\n                      Immune Response                                                                                  \n                      Ribosome                                                                                         \n\n###### \n\nAging signatures in adult mouse stem cells\n\n  Stem Cell                  Upregulated pathways         Downregulated pathways       Profiling method   Reference\n  -------------------------- ---------------------------- ---------------------------- ------------------ -----------------------------------------\n  Neural (SVZ)               Cell Adhesion                Angiogenesis                 RNA-seq            ([@b38-bmb-52-086], [@b39-bmb-52-086])\n                             Cell-cell Signaling          Blood Vessel Development                        \n                             Cell Morphogenesis           Blood Vessel Morphogenesis                      \n                             Nerve Impulse Transmission   Development                                     \n                             Neuron Differentiation       Lysosomes                                       \n                                                          Regulation of Cell Motion                       \n                                                          Response to Growth Factors                      \n                                                          Vasculature                                     \n  Muscle (Satellite Cells)   Apoptosis Regulation         Histone Genes                Microarray         ([@b41-bmb-52-086], [@b249-bmb-52-086])\n                             Immune Response                                                              \n                             Oxidation Reductin                                                           \n                             Vascular Development                                                         \n  Hematopoietic              Cell Adhesion                Cell Cycle                   Microarray         ([@b42-bmb-52-086])\n                             Cell Proliferation           Epigenetic Regulation        RNA-seq            ([@b250-bmb-52-086])\n                             Inflammation                 Genomic Integrity            scRNA-seq          ([@b44-bmb-52-086])\n                             Protein Aggregation                                                          \n                             Repetitive Elements                                                          \n                             Ribosome                                                                     \n                             Stress Response                                                              \n\n###### \n\nEffects of longevity-promoting interventions on aging signatures\n\n  Intervention            Tissue    Effect on aging signatures                                   Species   Profiling method   Reference\n  ----------------------- --------- ------------------------------------------------------------ --------- ------------------ ----------------------\n  Dietary interventions                                                                                                       \n  \u2003Dietary                Liver     *Tissue*                                                     Mouse     RNA-seq            ([@b251-bmb-52-086])\n  \u2003\u2003Restriction                     Decreases cell differentiation and maturation                          RNA-seq            ([@b53-bmb-52-086])\n                                    Decreases protein metabolism                                           ChIP-seq           \n                                    Decreases transcription                                                WGBS               ([@b52-bmb-52-086])\n                                    Prevents age-associated methylation changes                                               \n                                    Slows epigenetic clock                                                                    \n                                    *Tissue*                                                     Rat       Western Blot       ([@b252-bmb-52-086])\n                                    Site specific acetylation at H3K9, K27, and K56                                           \n                                    *Cell (Hepatoma model)*                                      Mouse     RNA-seq            ([@b253-bmb-52-086])\n                                    Alters SIRT gene expression                                                               \n                                    Modifies accessibility of SIRT regulator region                                           \n                          Brain     Hippocampus                                                  Mouse     RNA-seq            ([@b254-bmb-52-086])\n                                    *Tissue*                                                                                  \n                                    Alters calcium signaling                                                                  \n                                    Alters axonal guidance signaling                                                          \n                                    Alters corticotropin-releasing hormone signaling                                          \n                                    Alters synaptic long-term potentiation                                 RNA-seq            ([@b255-bmb-52-086])\n                                    Alters neuronal CREB signaling                                                            \n                                    Alters G-protein coupled receptor signaling                                               \n                                    Decreases senescence-associated secretory phenotype                                       \n                                    Decreases energy regulation                                                               \n                                    Decreases inflammation                                                                    \n                                    Decreases phagocytosis                                                                    \n                                    Prevents age-related methylation changes                                                  \n                                    Whole Brain                                                  Mouse     Western Blot       ([@b256-bmb-52-086])\n                                    *Tissue*                                                                                  \n                                    Attenuates age-associated reduction in histone methylation                                \n                          Adipose   *Tissue*                                                     Mouse     Microarray         ([@b257-bmb-52-086])\n                                    Decreases inflammation                                                 Western Blot       ([@b258-bmb-52-086])\n                                    Decreases necroptosis                                                  RNA-seq            \n                                    Increases stem cell maintenance                                                           \n                                    Increases vascularization                                                                 \n  \u2003Genetic                                                                                                                    \n  \u2003\u2003*Ames* Dwarfism       Liver     *Tissue*                                                     Mouse     RNA-seq            ([@b53-bmb-52-086])\n                                    Alters methionine metabolism                                           ChIP-seq           \n                                    Attenuates age-associated hypermethylation                             WGBS               ([@b52-bmb-52-086])\n                                    Decreases glycine-N-methyltransferase                                                     \n                                    Increases glutathione                                                  Microarray         ([@b68-bmb-52-086])\n                                    Increases mRNA expression of DNMT1 and DNMT3a                          Western Blot       ([@b66-bmb-52-086])\n                                    Increases urea cycle proteins                                                             \n                                    Prevents age-associated miRNA expression changes                                          \n                                    Slows epigenetic clock                                                                    \n                                    Stabilized epigenome according to chronological age                                       \n                                    Suppresses cell development and identity                                                  \n                                    Suppresses polyamine metabolism                                                           \n                          Serum     *Cell*                                                       Mouse     RNA-seq            ([@b67-bmb-52-086])\n                                    Modulates insulin signaling                                                               \n                                    Modulates MAPK signaling                                                                  \n                                    Modulates mTOR signaling                                                                  \n                                    Modulates Wnt signaling                                                                   \n                                    Prevents age-associated miRNA expression changes                                          \n  \u2003Drug                                                                                                                       \n  \u2003\u2003Rapamycin             Liver     *Tissue*                                                     Mouse     WGBS               ([@b52-bmb-52-086])\n                                    Prevents age-associated methylation changes                            RNA-seq            ([@b53-bmb-52-086])\n                                    Slows epigenetic clock                                                 ChIP-seq           \n                          Brain     *Tissue*                                                     Mouse     Western Blot       ([@b256-bmb-52-086])\n                                    Attenuates age-associated histone modifications                                           \n  \u2003\u2003Metformin             Liver     *Tissue*                                                     Mouse     Microarray         ([@b259-bmb-52-086])\n                          Muscle    Attenuates expression of NF-kB                                                            \n                                    Increases anti-inflammatory genes                                                         \n                                    Improves stress response                                                                  \n                                    Increases antioxidant response                                                            \n\n###### \n\nImpact of transposable elements in health and disease\n\n  Organism                                     Cell lines/Cell Types/Tissues                                              Transposon name              How were they measured?                                                                       Changes observed in Transposons                                                                      Phenotypes associated with the listed change                                                                             Reference\n  -------------------------------------------- -------------------------------------------------------------------------- ---------------------------- --------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------ ------------------------------------------\n  Drosophila melanogaster                      Fat body                                                                   Multiple Retro transposons   qPCR, RNA-seq, Immunofluorescence microscopy                                                  Increase in expression with age                                                                      Aging                                                                                                                    ([@b260-bmb-52-086])\n  Drosophila melanogaster                      Heads                                                                      LINE-1, Gypsy                qPCR, Immunochemical staining                                                                 Increase in expression of TEs                                                                        Aging                                                                                                                    ([@b169-bmb-52-086])\n  Mus musculus                                 Tissue (C57BL/6) from Aged Rodent Tissue Bank & NIA Aged Rodent Colonies   LINE-1 & MusD                Nuclease Sensitivity, oligo-dT immuno-FISH staining, RNA-seq, qPCR                            Increase in copy numbers of LINE-1 & MusD retrotransposons in liver and muscle                       Increase in somatic retro-transposition with age                                                                         ([@b261-bmb-52-086])\n  Mus musculus                                 ES cell culture, embryos, 2C-GFP ES cell line                              LINE-1                       RNA-FISH, RNA-seq, RT-qPCR, ChIP Assay, Immunofluorescence analysis, co-immunoprecipitation   LINE-1 inhibition                                                                                    Inhibited embryonic stem cell self-renewal and impaired embryo development                                               ([@b175-bmb-52-086])\n  Homo sapiens                                 Oocyte and embryo                                                          HERVL retrotransposons       RNA-seq, Immunofluorescence Imaging                                                           HERVL and MLT2A1 repeat elements are transiently expressed during the cleavage stage                 DUX4 potentially activating the cleavage stage genes and repetitive elements                                             ([@b262-bmb-52-086])\n  Homo sapiens                                 iPSC culture                                                               HERVL retrotransposons       RNA-seq, Luciferase Assay, ChIP-Seq, RT-qPCR                                                  HERVL and MLT2A1 repeat elements are transiently expressed during the conversion of cleavage stage   DUX4 potentially activating the cleavage stage genes and repetitive elements                                             ([@b262-bmb-52-086])\n  Mus musculus                                 ES cell culture                                                            MERVL retrotransposons       RNA-seq, ATAC-seq, ChIP-Seq and Immunofluorescence and Imaging                                Conversion mESCs to a 2C-like state when MERVL peaks                                                 Murine DUX activating the cleavage stage genes                                                                           ([@b262-bmb-52-086])\n  Homo sapiens                                 HeLa M2cell line                                                           LINE-1                       Mass Spectrometry, Immunofluorescence staining, RNA-FISH, Western                             Peaks during S phase of the cell division                                                            N/A                                                                                                                      ([@b263-bmb-52-086])\n  Homo sapiens                                 Adipose derived stem cells (ADCS)                                          Alu/SINE                     Immunofluorescence, qPCR, ChIP-Seq, Northern Hybridization                                    Knockdown of Alu/SINE in senescent cells                                                             Initiates self-renewal of the cells                                                                                      ([@b264-bmb-52-086], [@b265-bmb-52-086])\n  Drosophila melanogaster                      dAgo2 mutated heads                                                        LINE-1, Gypsy                Western Blots, qPCR Immunochemical staining, GFP Imaging                                      Increase in expression of TEs                                                                        Neuronal decline                                                                                                         ([@b169-bmb-52-086])\n  Homo sapiens                                 H9 ESCs cell line                                                          LINE-1                       RT-qPCR, Retro transposition Assay, TUNEL analysis, Western Blot                              LINE-1 increases when TREX1 decrease                                                                 Neuroinflammation                                                                                                        ([@b266-bmb-52-086])\n  Homo sapiens (ROSMAP project)                Brain                                                                      Multiple TEs                 RNA-seq, PCR, H3KAc ChIP-seq                                                                  Tau levels correlate with transposon levels                                                          Alzheimer's                                                                                                              ([@b180-bmb-52-086])\n  Drosophila melanogaster (Transgenic flies)   Brain                                                                      Multiple TEs                 qPCR, RNA-seq                                                                                 As Tau is expressed, transposon expression increases                                                 Alzheimer's                                                                                                              ([@b180-bmb-52-086])\n  Drosophila melanogaster (Transgenic flies)   Head Tissues                                                               gypsy ERV                    qPCR, RNA-seq                                                                                 Neuronal and glial hTDP-43 expression exhibits elevated expression of gypsy ERV                      Neurodegeneration in diseases such as Amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTDL)   ([@b183-bmb-52-086])\n  Homo sapiens                                 Brain                                                                      Multiple TEs                 RNA-seq, RT-qPCR                                                                              Increased repetitive element accumulation in C9orf72 positive cases                                  Neurodegeneration in diseases such as Amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTDL)   ([@b267-bmb-52-086])\n  Rattus norvegicus                            Cortical neurons                                                           Multiple TEs                 RIP-seq                                                                                       TDP-43 binds to transposons                                                                          N/A                                                                                                                      ([@b182-bmb-52-086], [@b268-bmb-52-086])\n  Mus musculus                                 Brain                                                                      Multiple TEs                 mRNA-seq, CLIP-seq                                                                            Depletion of TDP-43 in mice increases transposon expression, TDP-43 binds to transposons             N/A                                                                                                                      ([@b182-bmb-52-086], [@b269-bmb-52-086])\n  Mus musculus (transgenic)                    Spinal cord                                                                Multiple TEs                 mRNA-seq                                                                                      Increase in expression of transposons when TDP-43 was overexpressed                                  N/A                                                                                                                      ([@b182-bmb-52-086], [@b270-bmb-52-086])\n  Homo sapiens                                 Brain                                                                      Multiple TEs                 CLIP-seq                                                                                      Decreased binding of transposons to TDP-43                                                           Frontotemporal lobar degeneration (FTDL)                                                                                 ([@b182-bmb-52-086])\n  Mus musculus                                 Lymphoma and hepatoc ellular carcinoma tumors                              LINE-1 and MusD              qPCR                                                                                          Copy number increases compared to control tissue                                                     Cancer                                                                                                                   ([@b261-bmb-52-086])\n  Homo sapiens                                 Many tissue types                                                          LINE-1                       Immunohistochemistry                                                                          Present in the cancerous tissue types; mostly absent in healthy tissue                               Cancer                                                                                                                   ([@b271-bmb-52-086])\n  Homo sapiens                                 Blood                                                                      AAlu/SINE and LINE-1         MethyLight PCR                                                                                Transposon methylation status is inversely correlated with menopausal age                            Earlier menopause is associated with earlier onset of age-related disorders                                              ([@b140-bmb-52-086])\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "1. Introduction {#s0005}\n===============\n\nGlioblastoma multiforme (GBM) is a malignant primary brain neoplasm with variable presentation depending on the location of the tumor. GBM is often rapidly progressive with an estimated median survival of 14 to 16\u00a0months [@bb0005]. Symptomatic seizures are common in patients with primary brain tumors [@bb0010]. Tumor-related epilepsy can be the first manifestation of GBM and present in the absence of other neurologic symptoms or focal deficits on examination. Depending on tumor size and location, scalp electroencephalogram (EEG) may not identify epileptiform discharges or electrographic seizures, representing a diagnostic challenge.\n\nTreatment of GBM involves maximal safe surgical resection followed by radiation therapy and adjuvant temozolomide chemotherapy [@bb0015]. In patients with GBM, gross total resection has been associated with two to three month improved survival [@bb0020]. The infiltrative nature of GBM and proximity of tumor to eloquent cortex can make surgical excision challenging. Use of electrocorticography for intraoperative monitoring during an awake craniotomy can guide safe surgical excision. Intracranial EEG can also identify epileptiform discharges undetected on scalp EEG, helping confirm a diagnosis of tumor-related epilepsy.\n\n2. Case description {#s0010}\n===================\n\nA 68-year-old right-handed man presented with a four-week history of paroxysmal neurological episodes characterized by incomprehensible speech lasting several minutes. He also reported episodes dating back to when he was 16\u00a0years old, described as word-finding difficulty lasting several minutes. These were followed by a headache and occurred once or twice a year. At age 23, the events of altered speech became associated with migratory paresthesias in the upper extremities. During his sixth decade of life, the events became accompanied by visual symptoms including jagged lines, floaters, and spiral-shaped light. His recent episodes of dysphasia represented a change in semiology as they did not include paresthesias or visual symptoms and they were not followed by a headache. The episodes involved no abnormal body movements, loss of awareness, or loss of bowel or bladder function. The neurological examination was normal.\n\nGiven new episodes on the background of a long-standing history of spells, a broad differential was considered, including vascular, infectious, inflammatory, and neoplastic etiologies. Outpatient electroencephalogram (EEG) revealed mild left temporal slowing without evidence of epileptiform discharges ([Fig. 1](#f0005){ref-type=\"fig\"}A). Brain MRI demonstrated a T2 hyperintense lesion in the left temporal lobe that enhanced with gadolinium ([Fig. 1](#f0005){ref-type=\"fig\"}B, C). Scattered areas of periventricular T2 hyperintensity were also noted ([Fig. 1](#f0005){ref-type=\"fig\"}D). There were no diffusion weighted imaging (DWI) or apparent diffusion coefficient (ADC) changes.Fig. 1A: Scalp EEG showing mild left temporal slowing without evidence of epileptiform discharges.B: Axial MRI brain T2 FLAIR image indicating a left temporal mass lesion with no significant mass effect or edema.C: Axial and coronal T1 post-gadolinium MRI brain images demonstrating a left temporal mass lesion with contrast enhancement.D: Sagittal MRI brain T2 FLAIR image indicating scattered periventricular hyperintensities.E: MR spectroscopy showing elevated choline to creatine peaks as well as decreased N-acetyl aspartate, consistent with neoplasm.F: Intracranial EEG demonstrating a focal electrographic seizure arising from a single contact of the depth electrodes.G: Pathology indicating cellular atypica, endothelial proliferation, and necrosis consistent with a diagnosis of glioblastoma multiforme.Fig. 1\n\nThe patient was admitted to the epilepsy monitoring unit (EMU) for seizure classification. Despite four days of monitoring, no episodes of speech arrest were captured. Interictal EEG demonstrated left temporal slowing without epileptiform discharges. Cerebrospinal fluid (CSF) analysis revealed 5 white blood cells, normal glucose at 78\u00a0mg/dL, and elevated protein at 56\u00a0mg/dL. Oligoclonal band assay was negative. CSF cytology was negative for malignancy.\n\nMagnetic resonance spectroscopy (MRS) was requested to further characterize the lesion. MRS performed over the left superior temporal gyrus showed elevated choline to creatine peaks with decreased N-acetyl aspartate ([Fig. 1](#f0005){ref-type=\"fig\"}E), a pattern suggestive of neoplastic etiology.\n\nPatient underwent left-sided supratentorial awake craniotomy for resection of the left temporal lobe lesion and histopathological analysis. Intraoperative electrocorticography with electrocortical stimulation for language mapping was performed during the procedure with an 8\u00a0\u00d7\u00a08 high-density grid placed over the left lateral temporal lobe. Functional mapping localized eloquent cortex for language and subsequently excluded from surgical resection. The anterior temporal lobe with the bulk of the tumor was resected. Electrocorticography demonstrated lateralized periodic discharges over the temporal lobe lesion with a focal electrographic seizure arising from a single contact of the depth electrode ([Fig. 1](#f0005){ref-type=\"fig\"}F).\n\nOn histopathological analysis, the surgical specimen was consistent with an isocitrate dehydrogenase (IDH) wild-type GBM ([Fig. 1](#f0005){ref-type=\"fig\"}G). Methylguanine-DNA methyltransferase (MGMT) methylation status was indeterminate. No complications from surgery arose and the patient was referred for stereotactic radiation and temozolomide chemotherapy. He was also discharged on a maintenance dose of levetiracetam.\n\n3. Discussion {#s0015}\n=============\n\nOur patient\\'s spells that started at age 16, described as transient word-finding difficulty, sensory changes, and visual changes followed by headache most consistent with migraine. Reversible visual, sensory, and speech symptoms lasting several minutes followed by a headache are suggestive of migraine with aura. Migraine peaks between ages 15 and 19 in males, which coincides with the onset of our patient\\'s spells at age 16 [@bb0025]. Additionally, patients with migraine may have subclinical white matter lesions on MRI imaging, though similar MRI findings can also represent chronic microvascular ischemia [@bb0030].\n\nHowever, the more recent episodes of speech arrest represented a new type of spell and suggested focal seizures arising from the dominant temporal lobe. Brain MRI identified an enhancing temporal lobe lesion with no significant peri-tumoral edema and little mass effect. The MRI features of the lesion were concerning for a high grade glioma, which are T2 hyperintense contrast-enhancing mass lesions. The temporal lobes are the second most common anatomical location for GBM [@bb0035].\n\nNon-invasive diagnostic measures including MRS also suggested a neoplasm. Conventional in vivo MRS uses protons to detect the concentration of brain metabolites including N-acetyl aspartate, choline, creatine, lactate, myoinositol, and glutamate in order to determine whether a lesion is neoplastic, inflammatory, or ischemic [@bb0040]. In general, astrocytomas have an increase in choline, indicating high glial cellularity and decreased N-acetyl aspartate reflecting neuronal replacement by tumor tissue, which was observed in the left superior temporal gyrus MRS performed for our patient [@bb0040].\n\nA definitive diagnosis was attained upon excisional biopsy. The location of the mass lesion in the dominant temporal lobe represented a surgical challenge, since the presenting symptom of speech arrest indicated that language function could be compromised by resection of the mass. Incorporation of awake craniotomy with electrocortical stimulation and functional language mapping was crucial to ensure maximal resection without adversely affecting eloquent cortex [@bb0045]. The diagnosis of brain tumor-related epilepsy secondary to GBM was made and treatment planning was completed without undue risk for the patient. Electrocorticography also revealed frequent epileptiform discharges and a brief focal seizure, which was not identified despite four days of EMU stay. This underscores the limitations of scalp EEG in identifying seizures arising from deep epileptic foci remote from recording electrodes.\n\nAlthough brain tumors account for 4% of all epilepsies, symptomatic seizures can occur in up to 85% of patients depending on tumor type and location [@bb0010]. TRE occurs in up to 30--50% of patients with GBM [@bb0010]. Molecular markers such IDH mutation and MGMT methylation status are important predictors of tumor behavior and provide prognostic information for GBM cases [@bb0050]. IDH mutations are associated with a better prognosis in gliomas, likely representing an early event in glioma genesis [@bb0050]. In our patient, IDH wild-type status confers a poorer prognosis. MGMT is a DNA repair enzyme whose activity can be silenced by methylation, making tumors more susceptible to alkylator chemotherapy with agents such as temozolomide [@bb0050]. In our patient, the MGMT methylation status was indeterminate, conferring a poorer prognosis. Radiotherapy with concomitant temozolomide and adjuvant temozolomide improves two-year survival in patients with GBM [@bb0015], which was prescribed to our patient.\n\n4. Conclusions {#s0020}\n==============\n\nThis case highlights that even in the apparent setting of a long-standing neurological process, it is important to recognize new symptoms that signify a difference condition and potential pathophysiology. In this case, a repeat evaluation uncovered a new diagnosis of GBM.\n\nThe case also underscores the potential benefits of electrocorticography, which allowed for safe maximal surgical excision of the mass lesion without causing speech deficit or damage to other eloquent cortex. Electrocorticography was also instrumental in identifying epileptiform discharges that could not be detected on scalp EEG, helping confirm the diagnosis of brain tumor-related epilepsy and guiding medical management.\n\nAuthors\\' contributions {#s0025}\n=======================\n\nStudy concept and design: Drs. Sener and Feyissa.\n\nAcquisition of data: All authors.\n\nManuscript preparation and literature review: Drs. Sener and Feyissa.\n\nCritical review of the manuscript: Drs. Tatum, Quinones-Hinojosa, Mahato, and Feyissa.\n\nDisclosures {#s0030}\n===========\n\nDrs. Sener, Mahato, and Tatum report no disclosures relevant to the manuscript.\n\nDrs. Feyissa and Qui\u00f1ones-Hinojosa were attending physicians receiving compensation for care provided to the patient described in this case report.\n\nDr. Tatum is Editor-in-Chief of Epilepsy Behavior Case Reports.\n"}
+{"text": "Introduction\n============\n\nOral cancer is the sixth most prevalent form of cancer, with oral squamous cell carcinomas (OSCCs) exhibiting the highest morbidity of all head and neck cancers.[@b1-ott-11-5657] Although great progress has been made in systemic and individualized treatments of OSCC, the incidence of OSCC has increased significantly over the past few decades[@b2-ott-11-5657] and metastasis remains a great challenge for the treatment of OSCC. Therefore, it is of great significance to investigate the mechanism of OSCC progression and metastasis.\n\nmiRNAs are a group of noncoding RNAs with a length of \\~14--22 nucleotides which are expressed in mammalian cells.[@b3-ott-11-5657],[@b4-ott-11-5657] It has been found that miRNAs play an important role in various pathological and biological processes such as cell proliferation, migration, differentiation, apoptosis, and inflammation.[@b5-ott-11-5657]--[@b7-ott-11-5657] Recently, some researchers found that miRNAs can not only regulate the expression levels of oncogenes and tumor suppressor genes to influence the process of tumor, but also might function as oncogenes or tumor suppressor genes directly.[@b8-ott-11-5657]\n\nMiR-300 has been investigated in some human cancers. In gastric cancer, miR-300 was found to be upregulated and overexpression of miR-300 could promote cell proliferation and cell cycle progression.[@b9-ott-11-5657] In laryngeal squamous cell carcinoma, miR-300 was found to suppress proliferation and invasion of cancer cells by targeting ROS1 and lower miR-300 expression predicts poor prognosis.[@b10-ott-11-5657],[@b11-ott-11-5657] It was also found that miR-300 was upregulated in breast cancer and promoted cell proliferation and cell cycle progression by targeting p53.[@b12-ott-11-5657] This indicated that miR-300 had different functional roles in human cancers. The expression and biological roles of miR-300 in OSCC have not been discovered.\n\nIn this study, we report that miR-300 was downregulated in OSCC tissues compared with normal tissues. Decreased expression of miR-300 was associated with TNM classification and unfavorable prognosis of OSCC patients. Furthermore, miR-300 inhibited the proliferation and invasion of OSCC cells in vitro. Notably, miR-300 may suppress epithelial--mesenchymal transition (EMT) by increasing the expression levels of E-cadherin and decreasing the expression levels of N-cadherin, Vimentin, Snail1 and MMP-2. Moreover, miR-300 mRNA levels were altered due to the differential expression of ET-1. In conclusion, the results of the present study demonstrate that miR-300 may act as a tumor suppressor by inhibiting EMT, suggesting that miR-300 may be a potential biomarker and anticancer therapeutic target for OSCC.\n\nMaterials and methods\n=====================\n\nTissue samples\n--------------\n\nA total of 120 specimens of OSCC were obtained from patients who underwent surgical resection in the Department of Oral and Maxillofacial Surgery, the School of Stomatology, China Medical University between 2008 and 2012. These tumor specimens were stored at \u221280\u00b0C. Experiments using human samples were approved by the ethics committee of the School of Stomatology, China Medical University, and written informed consent was obtained from the donors.\n\nCell culture\n------------\n\nHuman OSCC cell lines Tca8113, Cal-27 and HaCat cells were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). The Tca8113 cells were maintained in 1640 medium and Cal-27 and HaCat cells were maintained in DMEM medium; 10% of fetal bovine serum and 1% of penicillin--streptomycin solution were added to them. All the cell lines were maintained at 37\u00b0C in the presence of 5% CO~2~.\n\nReverse transcription-quantitative polymerase chain reaction (RT-qPCR)\n----------------------------------------------------------------------\n\nTotal RNA was extracted from the frozen tissues and cells using TRIzol (Takara Corporation, Dalian, China) according to the manufacturer's instructions. Using standard spectrophotometric methods, RNA concentration was quantified and its purity was determined. RNA (1 \u00b5g) was reverse transcribed using Harrpin-it\u2122 microRNA and U6 snRNA Normalization RT-PCR Quantitation Kit (GenePharma Co, Shanghai, China). The RT reaction system was as follows: 5\u00d7 MMLV RT buffer 4 \u00b5L, dNTP 0.75 \u00b5L, miRNA and U6 snRNA RT primer mix (1 \u00b5M) 1.2 \u00b5L, MML Reverse Transcriptase (200 U/\u00b5L)[@b2-ott-11-5657] 0.2 \u00b5L, RNA 1 \u00b5g; RNase-free H~2~O was added up to 20 \u00b5L. The following thermal cycling conditions were employed: 25\u00b0C for 30 minutes, 42\u00b0C for 30 minutes and 85\u00b0C for 5 minutes. qPCR reaction system: 2\u00d7 Real-time PCR Master Mix (SYBR) 10 \u00b5L, miRNA and U6 snRNA-specific primer set (10 \u00b5M)[@b1-ott-11-5657] 0.4 \u00b5L, ROX reference dye (50\u00d7)[@b3-ott-11-5657] 0.4 \u00b5L, Taq DNA polymerase (5 U/\u00b5L) 0.2 \u00b5L, miRNA RT product 2 \u00b5L; sterilized H~2~O was added up to 20 \u00b5L. The cycling conditions were as follows: 95\u00b0C for 3 minutes, 95\u00b0C for 12 seconds and 62\u00b0C for 40 seconds.\n\nTo examine the expression of E-cadherin, N-cadherin, Vimentin, Snail1 and MMP-2, cDNA served as the template for amplification of PCR using a cDNA Synthesis kit (PrimeScript\u2122 RT reagent kit, Takara), according to the manufacturer's instructions. The RT reaction system was as follows: 5\u00d7 PrimeScript RT Master Mix 4 \u00b5L, RNA 1 \u00b5g; RNase-free distilled H~2~O was added up to 20 \u00b5L. The following thermal cycling conditions were employed: 37\u00b0C for 15 minutes and 85\u00b0C for 5 minutes. qPCR was conducted using the SYBR Premix Ex Taq II kit (Takara Corporation) with a qPCR reaction system: SYBR Premix Ex Taq II 10 \u00b5L, cDNA 1 \u00b5L, forward primer 0.5 \u00b5L, reverse primer 0.5 \u00b5L and sterile water 8 \u00b5L. The cycling conditions were as follows: 95\u00b0C for 1 minute, 94\u00b0C for 30 seconds, 58\u00b0C for 30 seconds and 72\u00b0C for 10 seconds.\n\nGene expression was normalized to GAPDH as an internal control, and the relative level of miR-300 was calculated using the 2^\u2212\u0394\u0394Ct^ method.\n\nWestern blotting\n----------------\n\nProtein concentrations were determined by the bicinchoninic acid method. Next, samples corresponding to 50 \u00b5g of total protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis in a 10% gel and then transferred onto polyvinylidene difluoride membranes (Sigma, Shanghai, China). After blockage with 5% non-fat milk, the membrane was probed with primary antibodies including E-cadherin (1:1,000, Cell Signaling Technology), N-cadherin (1:500, Cell Signaling Technology), Vimentin (1:1,000, Cell Signaling Technology), Snail1 (1:1,000, Cell Signaling Technology), MMP-2 (1:1,000, Cell Signaling Technology) and GAPDH antibodies (1:2,000, Santa Cruz Biotechnology Inc.) and then probed with secondary antibodies (1:5,000, Santa Cruz Biotechnology Inc.). Proteins were visualized by means of Thermo Pierce ECL (Thermo Fisher Scientific, Waltham, MA, USA).\n\nCell transfection\n-----------------\n\nMiR-300 mimics, miR-300 inhibitor and control were synthesized by GenePharma Co. (GenePharma Co, Shanghai, China). The cells were seeded in six-well plates and were transfected using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific) following the manufacturer's protocol. miR-300 mimics sense: 5\u2032-UAUACAAGGGCAGACUCUCUCU-3\u2032, antisense: 5\u2032-AGAGAGUCUGC CCUUGUAUAUU-3\u2032negative control sense: 5\u2032-UUCUCCGAACGUGUCA CGUTT-3\u2032, antisense: 5\u2032-ACGUGACACGUUCGG AGAATT-3\u2032miR-300 inhibitor: 5\u2032-AGAGAGAGUCUGCCU UGUAUA-3\u2032miR inhibitor NC: 5\u2032-CAGUACUUUUGUGUAG UACAA-3\u2032\n\n3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-etrazolium,inner salt (MTS) assay\n--------------------------------------------------------------------------------------------------------------\n\nProliferation assays were conducted using the Cell Proliferation Assay kit (Promega, Madison, WI, USA). OSCC cells were plated in 96-well plates (2\u00d710^3^ cells/well) and cultured at 37\u00b0C for 18 hours prior to transfection. Subsequently, the cells were transfected with miR-300 mimics and miR-NC, in addition to miR-300 inhibitor and miR-inhibitor NC. The culture plates were taken out at different time periods (24, 48 and 72 hours) and continuously cultured for 2 hours after the addition of 10 \u00b5L MTS in each well. The OD value was measured at a wavelength of 570 nm using enzyme-linked immunosorbent assay.\n\nTranswell assays\n----------------\n\nOSCC cells were transfected for 48 hours. The cells (5\u00d710^4^ cells/well) were resuspended in serum-free medium and plated onto the upper chamber (Corning Costar, Tewksbury, MA, USA). Then, 0.5 mL of medium containing 10% fetal bovine serum (Clark, Beijing, China) was added in the lower chamber to act as chemoattractant. Transwell chambers coated with Matrigel (Corning Inc.) were used. Cells in the Transwell plates were incubated at 37\u00b0C in a humidified atmosphere with 5% CO~2~ for 48 hours. The adherent cells on the upper surface of the insert membrane were carefully removed with cotton tips. Cells on the lower surface of the membrane were fixed with 4% paraformaldehyde for 15 minutes, permeabilized in 0.1% Triton X-100 and stained with 0.1% crystal violet for 30 minutes. The invaded cells were imaged and quantified using a microscope (Olympus, Beijing, China) in five separate fields per membrane.\n\nStatistical analysis\n--------------------\n\nStatistical analysis was carried out in the IBM SPSS 21.0 software (IBM Corporation, Armonk, NY, USA). Student's *t*-test was used to examine statistical differences between two groups. Correlations between the miR-300 expression and clinical characteristics of OSCC patients were assessed by chi-squared test. In addition, overall survival (OS) and disease-free survival (DFS) were calculated by the Kaplan--Meier method. To determine the effect of particular prognostic factors on survival, a multivariate analysis was performed according to the Cox regression model. *P*\\<0.05 was assumed to indicate a statistically significant difference.\n\nResults\n=======\n\nDownregulation of miR-300 in OSCC\n---------------------------------\n\nIn order to examine the oncogenic role of the miR-300, we compared the miRNA expression between the OSCC tissues and normal oral mucosa tissues. qRT-PCR showed that compared with healthy normal tissues, the expression of miR-300 was obviously reduced in OSCC tissues (*P*\\<0.05; [Figure 1A](#f1-ott-11-5657){ref-type=\"fig\"}). Then, we divided the OSCC patients into two groups: patients with metastasis and those without metastasis. Compared with patients without metastasis, patients with metastasis showed significantly decreased level of miR-300 (*P*\\<0.05; [Figure 1B](#f1-ott-11-5657){ref-type=\"fig\"}). Furthermore, compared with HaCat cells, the expression of miR-300 in Tca8113 and Cal-27 cell lines was significantly decreased (*P*\\<0.05; [Figure 1C](#f1-ott-11-5657){ref-type=\"fig\"}).\n\nThe relationship between miR-300 and the clinicopathologic characteristics of the OSCC patients is summarized in [Table 1](#t1-ott-11-5657){ref-type=\"table\"}. Correlations between miR-300 and gender, age and tumor site were not statistically significant. However, the results showed that miR-300 was closely related to TNM classification (*P*\\<0.05; [Table 1](#t1-ott-11-5657){ref-type=\"table\"}).\n\nTo elucidate the prognostic role of miR-300, we examined the relationship between miR-300 expression and patient outcome with long-term follow-up. Kaplan--Meier analysis further showed patients with relatively lower level of miR-300 had significantly decreased OS (*P*\\<0.0001; [Figure 2A](#f2-ott-11-5657){ref-type=\"fig\"}) and DFS (*P*\\<0.0001; [Figure 2B](#f2-ott-11-5657){ref-type=\"fig\"}). Univariate and multivariate analyses were carried out using Cox proportional hazard model to evaluate the impact of miR-300 expression and clinicopathologic factors on the prognosis of OSCC patients. As shown in [Table 2](#t2-ott-11-5657){ref-type=\"table\"}, univariate Cox regression analysis suggested that differentiation (*P*=0.031 for OS, *P*=0.045 for DFS), TNM stage (*P*=0.041 for OS, *P*=0.035 for DFS), miR-300 expression (*P*=0.028 for OS, *P*=0.025 for DFS) as well as tobacco usage (*P*=0.033 for OS, *P*=0.040 for DFS) were significantly associated with poor OS and DFS. Multivariate analyses showed TNM stage (*P*=0.049 for OS, *P*=0.043 for DFS), miR-300 expression (*P*=0.035 for OS, *P*=0.030 for DFS) as well as tobacco usage (*P*=0.045 for OS, *P*=0.049 for DFS) were associated with poor OS and DFS.\n\nmiR-300 inhibits proliferation and invasion of OSCC cells\n---------------------------------------------------------\n\nIn order to analyze the association between miR-300 and OSCC, OSCC cells were transfected with miR-300 mimics or miR-300 inhibitor and the transfection efficiency was detected by RT-qPCR ([Figure 3A and C](#f3-ott-11-5657){ref-type=\"fig\"}). Overexpression of miR-300 significantly inhibited cell proliferation ([Figure 3B](#f3-ott-11-5657){ref-type=\"fig\"}), On the contrary, miR-300 inhibitor significantly promoted cell proliferation ([Figure 3D](#f3-ott-11-5657){ref-type=\"fig\"}). Overexpression of miR-300 significantly inhibited cell invasion ([Figure 3E](#f3-ott-11-5657){ref-type=\"fig\"}) and miR-300 inhibitor promoted cell invasion ([Figure 3F](#f3-ott-11-5657){ref-type=\"fig\"}) of Tca8113 and Cal-27 cell lines.\n\nmiR-300 inhibits the EMT of OSCC cells\n--------------------------------------\n\nTo explore the molecular mechanism of miR-300 in OSCC, we assessed the role of miR-300 in the EMT of OSCC cells. We performed Western blot and RT-qPCR to detect the expression of E-cadherin, N-cadherin, Vimentin, Snail1 and MMP-2 in the OSCC cells transfected with miR-300 mimics and inhibitor. The results of RT-qPCR ([Figure 4A](#f4-ott-11-5657){ref-type=\"fig\"}) and Western blot ([Figure 4B](#f4-ott-11-5657){ref-type=\"fig\"}) demonstrated that overexpression of miR-300 in Tca8113 cells increased the expression of E-cadherin and decreased N-cadherin, Vimentin, Snail1 and MMP-2 expression significantly. The results of RT-qPCR ([Figure 4C](#f4-ott-11-5657){ref-type=\"fig\"}) and Western blot ([Figure 4D](#f4-ott-11-5657){ref-type=\"fig\"}) in Cal-27 cells is the same. On the other hand, RT-qPCR ([Figure 4E](#f4-ott-11-5657){ref-type=\"fig\"}) and Western blot ([Figure 4F](#f4-ott-11-5657){ref-type=\"fig\"}) showed that miR-300 inhibitor in Tca8113 cells reduced E-cadherin expression and increased N-cadherin, Vimentin, Snail1 and MMP-2 expression significantly. The results of RT-qPCR ([Figure 4G](#f4-ott-11-5657){ref-type=\"fig\"}) and Western blot ([Figure 4H](#f4-ott-11-5657){ref-type=\"fig\"}) in Cal-27 cells is the same.\n\nEndothelin-1 (ET-1) inhibiting miR-300 in OSCC\n----------------------------------------------\n\nThe results of cell proliferation and invasion in this study indicated that miR-300 may inhibit the proliferation and invasion OSCC cells. To further understand the mechanisms of miR-300 in OSCC metastasis, ET-1 was used. We found that miR-300 expression was decreased in a dose-dependent manner after ET-1 treatment ([Figure 5A](#f5-ott-11-5657){ref-type=\"fig\"}).\n\nWhen we transfected OSCC with an miR-300 mimic and then treated them with ET-1,we confirmed the role of miR-300 in cell invasion. The data indicate that the miR-300 mimic inhibited ET-1--induced invasion ([Figure 5B](#f5-ott-11-5657){ref-type=\"fig\"}). We also find that the miR-300 mimic but not the control miRNA abolished ET-1--induced EMT ([Figure 5C and D](#f5-ott-11-5657){ref-type=\"fig\"}).\n\nDiscussion\n==========\n\nIn OSCC patients, metastasis is the primary cause of mortality. Thus, it is very important to understand the molecular mechanisms of metastasis. Recently, miRNAs have been reported to promote or suppress tumor metastasis,[@b13-ott-11-5657],[@b14-ott-11-5657] providing a new perspective on the metastatic process. Emerging research suggests that miRNAs play essential roles in the progression of OSCC.[@b15-ott-11-5657]--[@b19-ott-11-5657] Nonetheless, research on the role of miRNAs in OSCC metastasis is lacking. EMT, which enables epithelial cells to acquire invasive mesenchymal phenotype, is attracting increasing attention as an important mechanism for the initial step of metastasis.[@b20-ott-11-5657] In this study, we observed that miR-300 played an important role as a suppressor of EMT in OSCC, which, in turn, inhibited metastasis.\n\nTo gain further insights into the role of miR-300 in OSCC metastasis, the expression of miR-300 was detected in 120 OSCC samples. We found that miR-300 expression was significantly decreased in OSCC tissue. Furthermore, significantly lower levels of miR-300 were found in patients with metastasis than those in patients without metastasis. Decreased expression of miR-300 was associated with adverse clinicopathologic features and poor prognosis of OSCC patients. Functionally, miR-300 was found to inhibit the proliferation and invasion of OSCC cells. Moreover, through inhibiting EMT phenotype (as suggested by increased E-cadherin and decreased N-cadherin, Vimentin, Snail and MMP-2), we further explored whether miR-300 inhibited the metastasis of OSCC cells, and the data of RT-qPCR and Western blot showed that miR-300 could inhibit the EMT of OSCC cells.\n\nConsiderable evidence suggests that tumor cells expressing aberrant levels of ET-1 facilitate tumor development and progression. However, the molecular mechanisms of ET-1--regulated EMT by miRNA in human oral cancers are not well characterized. Our study indicates that miR-300 is downregulated in response to ET-1, and that transfection of cells with miR-300 mimic reduces ET-1--induced cell invasion. The ET-1 signaling has a distinct function in OSCC, namely, regulation of EMT and cell invasion by repressing miR-300.\n\nHowever, there are some limitations of our study. First, the sample size in the subgroup analysis is not big enough; thus, the statistical power is limited. Second, the relevant mechanisms have not been identified. In our future work, we plan to do more in-depth research and analysis. Furthermore, animal experiments need to be devised to investigate the effects of miR-300 on tumor growth in vitro and in vivo.\n\nIn conclusion, the key finding in our study is that miR-300 suppresses metastasis of OSCC by inhibiting EMT. These findings indicate that high miR-300 expression may be a useful therapeutic target for OSCC.\n\nThis study was supported by the Natural Science Foundation of Liaoning Province (20170541012).\n\n**Disclosure**\n\nThe authors report no conflicts of interest in this work.\n\n![The expression of miR-300 in OSCC tissues and cells.\\\n**Notes:** (**A**) The expression of miR-300 was compared between OSCC tissues and normal tissues. (**B**) The expression of miR-300 was compared between OSCC patients with metastasis and those without metastasis. (**C**) The expression of miR-300 was compared between two OSCC cell lines and the HaCat cells. \\**P*\\<0.05.\\\n**Abbreviation:** OSCC, oral squamous cell carcinoma.](ott-11-5657Fig1){#f1-ott-11-5657}\n\n![The prognostic value of miR-300 in OSCC.\\\n**Note:** Kaplan--Meier plots were used to compare the (**A**) overall survival rate and (**B**) disease-free survival rate between patients with high level of miR-300 and those with low miR-300 level.\\\n**Abbreviation:** OSCC, oral squamous cell carcinoma.](ott-11-5657Fig2){#f2-ott-11-5657}\n\n![miR-300 inhibits the proliferation and invasion abilities.\\\n**Notes:** (**A**, **C**) Tca8113 and Cal-27 cells were transfected with miR-300 mimics and inhibitor for 48 hours. The levels of miR-300 were determined by reverse transcription quantitative PCR assay. (**B**, **D**) The cells were treated as indicated above, and the cell proliferation ability was determined by MTS assay. (**E**, **F**) The cells were treated as indicated above, and the cell invasion ability was determined by Transwell. \\**P*\\<0.05, \\*\\**P*\\<0.01.\\\n**Abbreviations:** PCR, polymerase chain reaction; NC, negative control.](ott-11-5657Fig3){#f3-ott-11-5657}\n\n###### \n\nmiR-300 inhibits the EMT of OSCC cells.\n\n**Notes:** (**A**, **C**) miR-300 overexpression significantly increased the mRNA level of E-cadherin and decreased the mRNA level of N-cadherin, Vimentin, Snail1 and MMP-2 in OSCC cells. (**B**, **D**) miR-300 overexpression significantly increased the protein expression of E-cadherin and decreased the expression of N-cadherin, Vimentin, Snail1 and MMP-2 in OSCC cells. (**E**, **G**) miR-300 downregulation significantly decreased the mRNA level of E-cadherin and increased the mRNA level of N-cadherin, Vimentin, Snail1 and MMP-2 in OSCC cells. (**F**, **H**) miR-300 downregulation significantly decreased the protein expression of E-cadherin and increased the expression of N-cadherin, Vimentin, Snail1 and MMP-2 in OSCC cells. \\**P*\\<0.05.\n\n**Abbreviations:** EMT, epithelial--mesenchymal transition; OSCC, oral squamous cell carcinoma; NC, negative control.\n\n![](ott-11-5657Fig4)\n\n![](ott-11-5657Fig4a)\n\n![ET-1 promotes cell migration and EMT expression by downregulating miR-300 expression in OSCC cells.\\\n**Notes:** (**A**) Cells were incubated with ET-1 (10--100 nM) for 24 hours, and miR-300 expression was examined by q-PCR. Cells were transfected with a control miRNA or miR-300 mimic for 24 hours and then stimulated with ET-1 (100 nM) for 24 hours. (**B**) Invasion and (**C**, **D**) EMT marker expression were examined by Transwell invasion assay, RT-qPCR and Western blot. \\**P*\\<0.05.\\\n**Abbreviations:** EMT, epithelial--mesenchymal transition; ET-1, endothelin-1; OSCC, oral squamous cell carcinoma; RT-qPCR, reverse transcription-quantitative polymerase chain reaction.](ott-11-5657Fig5){#f5-ott-11-5657}\n\n###### \n\nThe relationship between miR-300 and the clinicopathologic characteristics in OSCC patients\n\n  Patient characteristics   No of patients   miR-300 low expression (\u2264median) n (%)   miR-300 high expression (\\>median) n (%)   *\u03c7*^2^ value   *P*-value\n  ------------------------- ---------------- ---------------------------------------- ------------------------------------------ -------------- -----------\n  Patients included         120              78 (100)                                 42 (100)                                                  \n  Gender                                                                                                                                        \n  \u2003Male                     65               40 (51.3)                                25 (59.5)                                  0.747          \n  \u2003Female                   55               38 (48.7)                                17 (40.5)                                                 \n  Age (years)                                                                                                                                   \n  \u2003\u226455                      50               36 (46.2)                                14 (33.3)                                  1.846          \n  \u2003\\>55                     70               42 (53.8)                                28 (66.7)                                                 \n  Primary tumor site                                                                                                                            \n  \u2003Buccal                   24               16 (20.5)                                8 (19.0)                                   2.447          \n  \u2003Tongue                   52               37 (47.5)                                15 (35.7)                                                 \n  \u2003Gingival                 25               15 (19.2)                                10 (23.8)                                                 \n  \u2003Other                    19               10 (12.8)                                9 (21.5)                                                  \n  Differentiation                                                                                                                               \n  \u2003Well/moderate            59               38 (48.7)                                21 (50.0)                                  0.018          \n  \u2003Poor                     61               40 (51.3)                                21 (50.0)                                                 \n  T classification                                                                                                                              \n  \u2003T1/2                     52               20 (25.6)                                32 (76.2)                                  28.41          \\<0.01\n  \u2003T3/4                     68               58 (74.4)                                10 (23.8)                                                 \n  N classification                                                                                                                              \n  \u2003N0                       40               19 (24.4)                                21 (50.0)                                  8.077          \\<0.01\n  \u2003N1--3                    80               59 (75.6)                                21 (50.0)                                                 \n  M classification                                                                                                                              \n  \u2003M0                       69               35 (44.9)                                34 (81.0)                                  14.543         \\<0.01\n  \u2003M1                       51               43 (55.1)                                8 (19.0)                                                  \n\n**Abbreviation:** OSCC, oral squamous cell carcinoma.\n\n###### \n\nUnivariate and multivariate Cox proportional hazards model for OS and DFS in OSCC patients\n\n  Variables                                 Univariate   Multivariate                                  \n  ----------------------------------------- ------------ -------------- ------- ------- -------------- -------\n  **OS**                                                                                               \n  Gender (male vs female)                   1.362        0.589--2.003   0.597                          \n  Age (\u226455 vs \\>55) (years)                 1.412        0.687--2.056   0.548                          \n  Differentiation (well/moderate vs poor)   3.783        0.671--3.998   0.031   2.924   0.876--4.221   0.040\n  Tumor size, cm (\u22645 vs \\>5)                1.122        0.812--2.673   0.821                          \n  TNM stage (II+III vs I)                   2.765        0.991--3.629   0.041   2.554   1.222--3.012   0.049\n  miR-300 expression (high vs low)          3.887        1.245--5.283   0.028   3.112   1.675--4.434   0.035\n  Tobacco usage (yes vs no)                 3.221        1.034--4.557   0.033   2.678   1.097--4.553   0.045\n  **DFS**                                                                                              \n  Gender (male vs female)                   1.124        0.518--2.003   0.719                          \n  Age (\u226455 vs \\>55)                         1.301        0.786--2.135   0.613                          \n  Differentiation (well/moderate vs poor)   2.934        0.567--3.786   0.045   2.345   0.433--3.099   0.506\n  Tumor size, cm (\u22645 vs \\>5)                1.024        0.673--2.945   0.839                          \n  TNM stage (II+III vs I)                   3.011        1.034--4.562   0.035   2.997   0.984--3.765   0.043\n  miR-300 expression (high vs low)          3.642        1.345--5.128   0.025   3.423   1.212--4.743   0.030\n  Tobacco usage (yes vs no)                 2.988        1.174--4.632   0.040   2.665   1.112--4.086   0.049\n\n**Abbreviations:** DFS, disease-free survival; HR, hazard ratio; OS, overall survival; OSCC, oral squamous cell carcinoma.\n"}
+{"text": "Background {#Sec1}\n==========\n\nHuman direct participation in the production process has diminished by human's increasing progress in various fields, and instead, the role of the human factor in directing work systems through the control room has increased \\[[@CR1]\\]. In these kinds of tasks, the control of a large difference and sometimes critical parts are taken by human element, and if users cannot process the data quickly and accurately, an error may occur which leads to the occurrence of great events \\[[@CR2]\\]. Human error has long been considered as one of the most important factors in the occurrence of accidents. According to various studies, human errors in complex industries and systems, such as chemical processes, power plants and nuclear power plants, are known as the main causes of accidents. McCafferty (1995) stated that about 80% of the incidents include human error \\[[@CR3]\\]. In Lowe's study (2004), it was found that 64% of the accidents were because of human error. Other studies conducted by Gatchpole et al.(2006) and Krikos and Baker (2007) also showed that human error was involved in most of the events of complex system \\[[@CR4]\\]. Stringfellow (2010) also showed that 30 to 100% of industrial accidents were caused by human factor \\[[@CR5]\\].\n\nThere are various environmental factors which can affect the operators' performance in the control-room and lead to errors. Conditions of work environment such as heat, sound and illumination have an important effect on people's attitudes, behavior and performance \\[[@CR6]\\]. Illumination is one of the most important factors in performance. Because most of the activities in processing the data are done by human's vision system. Therefore, visual performance is very important for control-room operators. The results of the studies show that proper illumination has a positive effect on individual's performance and reduces the occurrence of accidents \\[[@CR7]--[@CR10]\\]. Also, it was found in studies that inappropriate illumination increases the eye fatigue, decreases the performance, and eventually leads to an incident. If environmental conditions, especially illumination, cannot meet the individual's performance requirements, it will reduce the quality of performance and lead to the occurrence of errors. Thus, creating good illumination is very important to individual's proper performance \\[[@CR9]\\]. Various illumination factors (such as intensity and color temperature) affect the individuals' visual and cognitive performance \\[[@CR11]\\]. Various factors such as color contrast, illumination level, viewing time, individual differences, gazing and etc., are effective on visual performance \\[[@CR12]\\]. The individual's efficiency in the control-room is directly affected by visual performance \\[[@CR10]\\]. People need good illumination in the control-rooms to have an appropriate vision \\[[@CR13]\\]. Most studies have investigated the effect of illumination on the individuals' visual performance in laboratories, so the aim of this study was to compare the control room operators' visual performance in two different illumination conditions at a combined cycle power plant through field-work.\n\nMethods {#Sec2}\n=======\n\nParticipants, time and place of the study {#Sec3}\n-----------------------------------------\n\nThis interventional study was carried out in the summer of 2017. The subjects were the male staff of the control-room in \"Parand Power Plant of MAPNA co.\" including 16 operators in 12-h shifts (from 7\u00a0am to 7\u00a0pm and from 7\u00a0pm to 7\u00a0am) in groups of 4 persons. From the 12-h shift, individuals spent 11\u00a0h on visual inspection in the control room. The control-room of the Parand power plant was 10.5\u00a0m \\* 25.5\u00a0m and 3.5\u00a0m high. There were 4 desks in this control-room where each of them were workstations for all the staff. The desk on which the staff focused was selected as the evaluation desk for intervention. There were 100 fluorescent lamps with white light (Power: 40\u00a0W, Color Temperature: 4000 Kelvin, Luminous Flux: 2780 Lumen and Model: Pars Shahab) which were installed in the control-room. They were installed in the ceiling, 50 double-lamps in 10 rows and in each there were 5. The walls were cream color and had a reflection coefficient of 0.7. The floor was gray with a reflection coefficient of 0.5. The roof was matte gray with a reflection coefficient of 0.3. The control-room had 5 windows:two south windows, two western windows and one north window.\n\nTools and method of collecting information {#Sec4}\n------------------------------------------\n\nThe Lux Meter (HAGNER S3) with a precision of 0.01 was used to measure the general illumination. General illumination of the control-room was measured by the network method. The average illumination intensity of the control-room was measured by Lux Meter at a height of 1.2\u00a0m based on the fourth model of the Illumination Engineers Society of North America (IESNA). Hagner S3 which is a combined machine to measure the illumination and luminance was used to measure local illumination and luminance.\n\nThe photocell of the Lux Meter was placed on the table horizontally, and the level of local luminance was measured in front of each display on the work surfaces. There were two displays on each table in the control room. The illumination was measured in front of both displays at work surface. The average local illumination of the work surface for each display was calculated based on the average of three levels of measurement in front of it.Then, the luminance level of the work surface and the display was measured by the Hagner apparatus.\n\nThe Freiburg Visual Acuity test (FrACT) software was used to evaluate the individuals' visual performance. This software was presented by Michael Bach in 1996 to measure the visual acuity \\[[@CR14]\\]. The validity of this software has been proven in the previous study of visual performance \\[[@CR15]\\]. Two tests of this software -Acuity C and Contrast C- was used to evaluate the visual performance.\n\nThe background color is white and the C color is black in the Acuity C test. The direction and size of the letter C is changed in each trial. In order to determine the level of visual performance in this test, the performance indicator will be calculated by Eq. [1](#Equ1){ref-type=\"\"}:\n\nWhere n is the total correct answers in each trial and \u03c4 is the duration of total trials in seconds (The individuals' reaction time is determined for each 18 trials separately).\n\nIn addition to calculating the visual performance indicator, this software also shows visual acuity for each Acuity C test ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\raisebox{1ex}{$1$}\\!\\left/ \\!\\raisebox{-1ex}{$ arcmin$}\\right. $$\\end{document}$). It indicated that the higher the value, the greater the individual's ability to recognize smaller dimensions. Dimension means the opening mouth of the C is in the test.\n\nIn addition to changing the direction, also the background contrast and the letter C will be changed in the Contrast C test. The subject is asked to use chance whenever he did not see the variable. The performance indicator is also calculated based on the total correct responses in the total duration (second) of the trials in this test. In addition, the results of the Contrast C test are expressed in terms of the percentage of contrast (% contrast Weber). The more a person performs the test carefully, the less the obtained number will be. It indicates that the subject detected the smallest contrast. Response time for each trial is 30\u00a0s in the software. If no response is received within this time, the next trial will begin and this non-response will be calculated as the wrong answer. In Fig. [1](#Fig1){ref-type=\"fig\"}, an example of the four main directions of the responses (a), acuity C test (b), contrast C test (c) was shown \\[[@CR16]\\].Fig. 1The example of the four main directions of the responses (**a**), aciuty C test (**b**), contrast C test (**c**)\n\nThe phases of the study {#Sec5}\n-----------------------\n\nThis study was conducted in two phases with interference. It includes the first phase with fluorescent illumination system and the second phase with a combined illumination system includes fluorescent and Light Emission Diodes (LED). The first phase illumination included fluorescent lamps (Power: 40\u00a0W, Color Temperature: 4000 Kelvin, Luminous Flux: 2780 Lumen and Model: Pars Shahab), to which people have been exposed for many years. In the first phase, the intensity of general and local illumination, and luminance were measured in two steps including day shift (9:00 and 15:00) and night shift (21:00 and 3:00) and it was in the beginning and the end of each shift. Then individual's visual performance was evaluated at four times of the day - including morning (at 9:00), afternoon (at 15:00), night (at 21:00), and morning (at 3:00) before and after the intervention. In the next phase, the illumination system was changed and LED lamps (Power: 48\u00a0W, Color Temperature: 4000 Kelvin, Luminous Flux: 5400 Lumen and Model: Mazinoor) with the same color temperature with fluorescent lamps were added to the system over the desk to increase the illumination of 200\u00a0lx at the work surface.The related calculations about local illumination design were used to calculate the height and suitable number of lamps. Then, a light bulb including 2 LED lamps at a height of 1.5\u00a0m above the desk (1\u00a0m below the ceiling) was used. In the second phase, for adapting the staff to the new illumination conditions, 15\u00a0days were considered. After placing the new illumination system above the desk in the second phase --and after 15\u00a0days interval between the evaluations, the intensity of local illumination was measured on this table. Then, the evaluations of the first phase were repeated and recorded (Fig. [2](#Fig2){ref-type=\"fig\"}).Fig. 2View of the room before and after the intervention (**a**. before the intervention, **b**. after the intervention)\n\nData analysis {#Sec6}\n-------------\n\nStatistical analysis was performed using SPSS 22.0 software. Wilcoxon non-parametric statistical test was used to compare the two illumination conditions because the data did not follow normal dispersion.\n\nResults {#Sec7}\n=======\n\nThe mean (Standard Deviation) of age and work experience of the subjects was 36.68 (2.982) years and 10.55 (2.191) years, respectively. The average values of the general and local illumination and the values of day and night luminance were measured before and after the intervention. They are shown in Table [1](#Tab1){ref-type=\"table\"}. The results showed that the illumination level of daylight and night at both work surfaces was less than 300\u00a0lx before the intervention (i.e. the suitable lighting for office work according to the standard). A standard luminance to detect the right color is at least 3 Cd / m^2^. The measurement results showed that the displays luminance and working surfaces was suitable for color recognition in daylight -even before doing the study-, while this amount was below the standard level at night.Table 1Average values of general and local illumination and luminance before and after the interventionParametersAverage of general illumination in control-room (Lux)Illumination of working surface1 (Lux)Luminance of display1 (Cd/m^2^)Luminance of working surface1 (Cd/m^2^)Illumination of working surface2 (Lux)Luminance of display2 (Cd/m^2^)Luminance of working surface2 (Cd/m^2^)Before intervention (first phase)Day3022306.34.22407.15.4Night1801530.160.121350.050.03After intervention (second Phase)Day30241515.212.942616.214.8Night1803500.20.183320.130.1\n\nValues of visual performance indicator and visual acuity of Acuity C Test are shown in Table [2](#Tab2){ref-type=\"table\"} at four times of the day - including morning (at 9:00), afternoon (at 15:00), night (at 21:00), and morning (at 3:00) before and after the intervention.The results of the Wilcoxon test showed that the change of illumination through intervention caused significant changes in performance indicator and visual acuity.Table 2The values of visual performance indicator of Acuity C Test (\u014a- accuracy C) and visual acuity before and after the interventionTimeBefore intervention (Median\u2009\u00b1\u2009IQR)After intervention (Median\u2009\u00b1\u2009IQR)*p*-value^\\*^Parameters9:00\u00a0*N*\u2009=\u200916\u014a0.33\u00a0\u00b1\u00a00.090.5\u00a0\u00b1\u00a00.12\\<\u20090.001Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$)1.62\u00a0\u00b1\u00a00.341.84\u00a0\u00b1\u00a00.38\\<\u20090.00115:00\u00a0*N*\u2009=\u200916\u014a0.22\u00a0\u00b1\u00a00.120.32\u00a0\u00b1\u00a00.11\\<\u20090.001Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$)1.31\u00a0\u00b1\u00a00.561.52\u00a0\u00b1\u00a00.66\\<\u20090.00121:00\u00a0*N*\u2009=\u200916\u014a0.33\u00a0\u00b1\u00a00.120.59\u00a0\u00b1\u00a00.1\\<\u20090.001Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$)1.20\u00a0\u00b1\u00a00.411.51\u00a0\u00b1\u00a00.31\\<\u20090.0013:00\u00a0*N*\u2009=\u200916\u014a0.25\u00a0\u00b1\u00a00.130.34\u00a0\u00b1\u00a00.13\\<\u20090.001Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$)1.02\u00a0\u00b1\u00a00.241.23\u00a0\u00b1\u00a00.23\\<\u20090.001\\*Wilcoxon Signed Ranks test\n\nThe values of visual performance indicator and visual acuity of the contrast C test are shown in Table [3](#Tab3){ref-type=\"table\"} at four times of the day including morning (9 o'clock), afternoon (at 15 o'clock), night (o'clock 21), and morning (at 3 o'clock) before and after the intervention. The results of the Wilcoxon test showed that the change in illumination through intervention caused significant changes in the values of the performance indicator and the percentage of the contrast.Table 3The values of performance indicator of Contrast C test (\u014a- contrast C) and percentage of contrast before and after the interventionTimeBefore intervention (Median\u2009\u00b1\u2009IQR)After intervention (Median\u2009\u00b1\u2009IQR)*p*-value^\\*^Parameters9:00\u00a0*N*\u2009=\u200916\u014a0.2\u00a0\u00b1\u00a00.120.41\u00a0\u00b1\u00a00.12\\<\u20090.001Percentage of Contrast (%)0.75\u00a0\u00b1\u00a00.260.52\u00a0\u00b1\u00a00.25\\<\u20090.00115:00\u00a0*N*\u2009=\u200916\u014a0.14\u00a0\u00b1\u00a00.080.31\u00a0\u00b1\u00a00.09\\<\u20090.001Percentage of Contrast (%)1.51\u00a0\u00b1\u00a00.91.39\u00a0\u00b1\u00a00.88\\<\u20090.00121:00\u00a0*N*\u2009=\u200916\u014a0.22\u00a0\u00b1\u00a00.080.53\u00a0\u00b1\u00a00.08\\<\u20090.001Percentage of Contrast (%)0.55\u00a0\u00b1\u00a00.320.43\u00a0\u00b1\u00a00.31\\<\u20090.0013:00\u00a0*N*\u2009=\u200916\u014a0.11\u00a0\u00b1\u00a00.060.31\u00a0\u00b1\u00a00.07\\<\u20090.001Percentage of Contrast (%)1.13\u00a0\u00b1\u00a00.641.04\u00a0\u00b1\u00a00.64\\<\u20090.001\\*Wilcoxon Signed Ranks test\n\nThe results of comparing the values of visual performance indicators at four different times of the day before and after the intervention are shown in Table [4](#Tab4){ref-type=\"table\"}. The results of the statistical test showed that the differences in the values of visual performance indicators at different times of the day were significant (*p*\u2009\\<\u20090.001).Table 4Comparison of the values of visual performance indicators before and after the interventionBefore Intervention9:0015:00*p*-value21:003:00*p*-value\u014a-acuity C *N*\u2009=\u2009160.33\u2009\u00b1\u20090.090.22\u2009\u00b1\u20090.12\\<\u20090.001^\\*^0.33\u2009\u00b1\u20090.120.25\u2009\u00b1\u20090.13\\<\u20090.001^\\*^Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$) *N*\u2009=\u2009161.62\u2009\u00b1\u20090.341.31\u2009\u00b1\u20090.56\\<\u20090.001^\\*^1.20\u2009\u00b1\u20090.411.02\u2009\u00b1\u20090.24\\<\u20090.001^\\*^\u014a-contrast C *N*\u2009=\u2009160.2\u2009\u00b1\u20090.120.14\u2009\u00b1\u20090.08\\<\u20090.001^\\*^0.22\u2009\u00b1\u20090.080.11\u2009\u00b1\u20090.06\\<\u20090.001^\\*^Percentage of Contrast (%) *N*\u2009=\u2009160.75\u2009\u00b1\u20090.861.51\u2009\u00b1\u20090.9\\<\u20090.001^\\*^0.55\u2009\u00b1\u20090.321.13\u2009\u00b1\u20090.64\\<\u20090.001^\\*^After Intervention9:0015:00*p*-value21:003:00*p*-value\u014a-acuity C *N*\u2009=\u2009160.5\u2009\u00b1\u20090.120.32\u2009\u00b1\u20090.11\\<\u20090.001^\\*^0.59\u2009\u00b1\u20090.10.34\u2009\u00b1\u20090.13\\<\u20090.001^\\*^Visual acuracy ($\\documentclass[12pt]{minimal}\n                \\usepackage{amsmath}\n                \\usepackage{wasysym} \n                \\usepackage{amsfonts} \n                \\usepackage{amssymb} \n                \\usepackage{amsbsy}\n                \\usepackage{mathrsfs}\n                \\usepackage{upgreek}\n                \\setlength{\\oddsidemargin}{-69pt}\n                \\begin{document}$$ \\frac{1}{Arcmin} $$\\end{document}$) *N*\u2009=\u2009161.84\u2009\u00b1\u20090.381.52\u2009\u00b1\u20090.66\\<\u20090.001^\\*^1.51\u2009\u00b1\u20090.311.23\u2009\u00b1\u20090.23\\<\u20090.001^\\*^\u014a-contrast C *N*\u2009=\u2009160.41\u2009\u00b1\u20090.120.31\u2009\u00b1\u20090.09\\<\u20090.001^\\*^0.53\u2009\u00b1\u20090.080.31\u2009\u00b1\u20090.07\\<\u20090.001^\\*^Percentage of Contrast (%) *N*\u2009=\u2009160.52\u2009\u00b1\u20090.851.39\u2009\u00b1\u20090.88\\<\u20090.001^\\*^0.43\u2009\u00b1\u20090.311.04\u2009\u00b1\u20090.64\\<\u20090.001^\\*^\\*Wilcoxon Signed Ranks test\n\nDiscussion {#Sec8}\n==========\n\nThe results of the present study indicated that the lighting system with intensity of 200\u00a0lx resulted in the reduction of the visual performance. On the other hand, the lighting system with intensity of 400 Lux improved visual performance. Therefore, increasing the illumination for administrative work - from the values which are below the permitted level to a higher level- would increase the visual indicators, visual accuracy, and the percentage of the contrast in individuals. Indeed, the increased mean value of the illumination created a actual improvement in the visual performance of the actual working condition. According to the results shown in Table [1](#Tab1){ref-type=\"table\"}, the values of day and night local illumination at working surfaces increased to a level higher than 300\u00a0lx after installation of LED lamps. Unfortunately, the luminance level of the surfaces was less than the permitted level at night -even after the intervention- and it requires a lamp with a higher color temperature. The luminance of the display surface was affected by illumination \\[[@CR17]--[@CR19]\\] and the high intensity of the illumination faded the display images in the user's eyes \\[[@CR20]\\], so we chose an illumination of about 400\u00a0lx for the intervention in illumination. On the other hand, there was no significance difference between the luminance stemmed from the light reflection in the average illumination and the light reflection in the low illumination (about 200\u00a0lx). By increasing the illumination from 200\u00a0lx to 400\u00a0lx, no change in the luminance will occur \\[[@CR21]\\]. However, the results of statistical tests in Tables [2](#Tab2){ref-type=\"table\"} and [3](#Tab3){ref-type=\"table\"} show the effect of illumination on individuals' response rate to visual trials. These results are in line with the previous studies, which showed that illumination has a completely direct effect on visual performance \\[[@CR17]--[@CR19]\\]. According to previous studies, the illumination of 200\u00a0lx causes eye fatigue \\[[@CR22]\\]. Lin in his study concluded that the illumination of about 500\u00a0lx provided a better visual performance than 200\u00a0lx and 1000\u00a0lx \\[[@CR22]\\]. Generally, the results of the present study showed that the visual performance in illumination of 400\u00a0lx was much better than that of 200\u00a0lx. According to these results, it can be argued that the illumination of 400\u00a0lx will improve the eye performance without visual discomfort. These results are in line with those of a study by Shieh et al. in 2000. They investigated the effects of illumination and type of the display on the visual performance of computer users and stated that the visual performance in illumination of 450\u00a0lx was greatly improved compared to 200\u00a0lx, and computer users experienced more visual comfort at the illumination of 450\u00a0lx \\[[@CR23]\\]. Some previous field studies were also consistent with the present study. Juslen et al. investigated the effect of illumination changes on the staffs' visual performance in a food industry. They concluded that increasing the local illumination by adding additional lamps to the general illumination system over the workstations would improve the visual performance and increase the staff's satisfaction in the production line at the workplace \\[[@CR24]\\]. In general, the results of the previous studies show that reducing visual fatigue and improving visual performance, reduce the error rate and improve the quality of the work performance \\[[@CR9]\\]. Therefore, it can be concluded that increasing the intensity of light can reduce the effects of eye fatigue such as headache and eye pain and can also prevent occupational accidents by improving work quality.\n\nThe effect of illumination on the visual performance has been studied in the past \\[[@CR25]\\], but previous studies are not perfect because they did not study the effect of white light in a real environment on individuals and also did not examine the visual performances such as visual acuity. Light improves the visual performance through visual systems. In a study by Lin et al., it was shown that light improved the performances associated with the visual system through the visual system \\[[@CR25]\\]. In addition to the above mentioned points, the effect of two types of lamps with the same color temperature was examined on visual performance in this study. At the same color temperature (4000 Kelvin), the combined LED lighting and fluorescent lighting system improves visual performance and reduces the individuals' reaction time in performing visual tasks. These results are in line with the study of Linhart, which states that using LED lighting systems in industrial environments can have positive effects on visual performance and reduced eye fatigue \\[[@CR16]\\]. The LED system supports optimal visual performance more than the fluorescent illumination system with a fairly low color temperature. The LED-based illumination system does not create gazing, so it leads to more support of optimal visual performance \\[[@CR26]\\].\n\nThe results also showed that the shift time had a significant effect on individuals' working memory performance and caused a decrease in the response speed at the end of the day shift. According to the results, time affects the response time and the number of correct answers in the acuity C duty and contrast C duty, so that response time and visual error increased at the end of the shift compared to the beginning of the shift, and the number of correct answers in both duties decreased. In other words, two visual performances at the end of day shifts decreased significantly compared to the beginning of the shift and also suffered a decline. These variables have not been studied in a field studies, but generally, the results of this study are consistent with those of previous studies about visual performance \\[[@CR27]\\]. Factors influencing visual performance should be mentioned for describing this result. Previous studies showed that the reduction of visual performance in working people was related to the effect of illumination on the eye fatigue and mental fatigue \\[[@CR28], [@CR29]\\]. Shortage and deprivation in sleep are closely related to the reduction of visual performance. As mentioned in numerous studies, long shifts can interfere with the sleep-awakening cycle and reduce the quantity and quality of sleep \\[[@CR30]\\]. One of the main disadvantages of 12-h shift that was mentioned in previous studies is sleepiness which can reduce the visual and vigilance performance and, on the other hand, increase the risk of accidents, and it has been proved that inappropriate illumination increases the level of sleepiness and eye fatigue \\[[@CR31]\\]. It is in line with the study of S.D. Baulk et al., indicating that sleepiness increases significantly at the end of 12-h shifts \\[[@CR32]\\]. This increase is certainly accompanied by a reduction in visual performances.\n\nTherefore, it can be useful to use a lighting system with intensity of 400\u00a0lx and more, because of increased visual performance, reduction of eye fatigue and its complications such as headache, eye pain, dislike of work and etc. In the current study, the effect of the age and gender properties on the visual performance were not been examined. Therefore, it is recommended that these two factors be evaluated in the future studies.\n\nConclusion {#Sec9}\n==========\n\nIn general, it was discovered that the lighting system with intensity of 200\u00a0lx results in the reduction of the visual performance. And, the lighting system with intensity of 400 Lux can significantly improve visual performance. The results of the present study showed that increasing illumination from the values which are below the permitted level -for administrative work- to a higher level would increase the visual performance in individuals. Also, the results showed that time has a significant effect on visual performance, and visual performance changes during time. Considering the importance of visual performance in control-rooms, it is suggested that illumination should be increased in them - at least the final hours of the shift- in order to prevent errors and unexpected accidents.\n\nFrACT\n\n:   The Freiburg Visual Acuity test\n\nIESNA\n\n:   The Illumination Engineers Society of North America\n\nLED\n\n:   Light Emission Diode\n\nSD\n\n:   Standard Deviation\n\nThe authors are grateful for the collaboration of the management and staff of Parand combined cycle power plant of MAPNA to carry out this research. This study was part of an approved research project partially funded by the TUMS (Tehran University of Medical Sciences).\n\nAvailability of data and materials {#FPar1}\n==================================\n\nThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\n\nSAZ managed and planned the project. SY and SGH analyzed the data and made the preliminary information in a state of measurable. KA as a statistician, he re-checked statistical analysis and fixed all the bugs. AZ collected the data in the field, and was a major contributor in writing the manuscript. All authors read and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar2}\n==========================================\n\nEthics approval was obtained by ethics committee of Tehran University of Medical Sciences (TUMS).\n\nConsent for publication {#FPar3}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar4}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar5}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "INTRODUCTION\n============\n\nIt is well known that hospital-acquired pressure injuries (HAPIs) are a significant source of morbidity, pain, patient dissatisfaction, and unnecessary increased cost in the hospital care of children.^[@R1]--[@R10]^ HAPIs can result in permanent damage to the skin and underlying tissues. A single HAPI has been shown to add approximately \\$43,180 to hospital care of a child.^[@R3],[@R4]^ Other estimates indicate that the costs of HAPI in the United States could exceed \\$26.8 billion.^[@R4]^ Despite being a preventable hospital-acquired condition (HAC), HAPIs occur in approximately 1.4% of hospitalized infants and children.^[@R7]--[@R9]^ In children, HAPIs include both those related to immobility and medical devices.^[@R1]--[@R10]^ Some reports show that up to 27% of patients in pediatric intensive care units experience HAPIs.^[@R10]^\n\nFollowing the HAPI prevention bundle (a group of actions which together have been shown to reduce the incidence of HAPI), individual hospitals demonstrated a reduction of serious harm.^[@R1]--[@R10]^ The Solutions for Patient Safety (SPS) network recommends a 3-pronged approach to the identification and prevention of HAPI: (1) conduct active surveillance to detect pressure injuries, (2) implement and measure compliance with the HAPI prevention bundle, and (3) deployment of a wound ostomy nursing team.^[@R1]^ Thirty-three hospitals that participated in the SPS phase 1 initiative to reduce HAPIs saw a statistically significant reduction in SPS-reportable HAPIs (stage 3, 4, and unstageable). These participating hospitals achieved an 80% compliance rate with the established HAPI prevention bundle.^[@R1]^\n\nDespite being a participating SPS children's hospital and having attempted implementation of the SPS HAPI prevention bundle, our hospital remained at a HAPI rate that was more than 3 times the mean for SPS participating children's hospitals. Specifically, the baseline rate of serious pressure injuries was 0.32 per 1,000 patient days when compared with the SPS benchmark mean rate of 0.09. Quality and nursing leaders paused all current unit-specific HAPI activities due to a lack of consistent bundle compliance and standards. We then launched an enterprise-wide HAPI reduction initiative. The purpose of this article is to describe the improvement initiative, the key drivers, and the resulting decrease in the HAPI rate. The goal of this improvement initiative was to reduce the SPS-reportable pressure injuries by 10% in the first year by implementing an action plan with key drivers determined by a causal analysis.\n\nMETHODS\n=======\n\nAn improvement initiative to decrease SPS-reportable pressure injuries was organized around a problem-solving template utilizing a key driver diagram with subactivities placed into 3 groupings: standardization, data transparency, and accountability. This improvement initiative was carried out at Lucile Packard Children's Hospital, Stanford (LPCHS). The system operates approximately 397 licensed beds in a free-standing children's and maternity hospital, associated ambulatory services, and is a wholly owned subsidiary of Stanford University. Of those 397 beds, 36 are beds managed by LPCHS at other sites, and 49 are related to maternity at LPCHS. There are 312 pediatric beds at LPCHS, and these are the beds used in calculating the HAPI rates. Based on institutional guidelines, this project was considered quality improvement and not human subjects research. Therefore, it did not require institutional review board review and approval.\n\nThe improvement initiative team performed a causal analysis to determine contributing factors to inconsistent HAPI bundle compliance and monitoring of practice (Fig. [1](#F1){ref-type=\"fig\"}). We grouped causes into the following categories: practice variation, process variation, data, communication, accountability, and resources. By studying the causes, we created a key driver diagram for the HAPI reduction initiative with activities placed into 3 groupings: standardization, data transparency, and accountability (Fig. [2](#F2){ref-type=\"fig\"}). This structure mirrors previous organizational efforts that showed a significant reduction in central line-associated bloodstream infections.\n\n![Causal analysis shows contributing factors to a high HAPI rate. RT, respiratory therapist; RN, registered nurse; RTW, Return to Ward (duty); WOC, Wound ostomy and continence nurse; IS, information services.](pqs-5-e289-g001){#F1}\n\n![Key driver diagram showing key areas of work in accountability, standardization, and data transparency with contributing actions and dates those actions were activated.](pqs-5-e289-g002){#F2}\n\nBecause different subinitiatives were implemented over time, it is difficult to determine an exact date to differentiate the pre- from the postintervention period. For this review, the baseline data for the preimplementation period were from January 2017 through June 2018. The postimplementation period was from July 2018 through August 2019. We implemented interventions on a rolling basis throughout the postimplementation period (Fig. [2](#F2){ref-type=\"fig\"}). The month of July 2018 was chosen as the pivot date as organizational goals and resources were determined then for the new fiscal year goal supporting HAPI reduction. With the prioritization of HAPI reduction, individual inpatient units developed unit-based HAPI reduction goals in alignment with the organizational initiative. The initiative was led by the leader of our wound care team, which consists of certified wound ostomy nurses in partnership with quality improvement and nursing leadership. The wound care team ensured that the HAPI reduction work was evidence based.\n\nHAPI Definition and Goal\n------------------------\n\nThe purpose of this improvement project was to reduce the rate of SPS-reportable HAPIs per 1,000 patient days. HAPIs included in the SPS-reportable definition are those defined as stage 3, stage 4, or unstageable.^[@R1]--[@R10]^ The original goal was to reduce the outcome measure, the SPS-reportable pressure injury rate per 1,000 patient days, by 10% in the first year. The definition of reportable pressure injuries used by SPS is consistent with definitions of pressure injuries associated with serious harm used by other organizations such as the National Quality Forum, Centers for Medicare and Medicaid Services, SPS, and many U.S. States.^[@R5],[@R6]^\n\nStandardization---Actions\n-------------------------\n\nWithin the key driver grouping of standardization, there were numerous subinitiatives to drive standardization of both process and practice. We assembled a group to evaluate the current state of standardized implementation of the SPS HAPI prevention bundle. An interdisciplinary team, including representatives from our Shared Governance Councils, used a combination of evidence-based practice and lean methodology^[@R11]^ to reach consensus on standard practice. The representatives included the hospital-wide Skin Integrity Prevention and Education (SkIPE) team which is a unit-based skin team and representatives from Clinical Practice Council, Technology and Informatics Council, as well as certified wound and ostomy nurses, respiratory therapists, and information services and quality improvement staff. This team met and reviewed the best practice literature for HAPI prevention, determined and implemented evidence-based interventions for each of the HAPI prevention bundle elements, and evaluated the feasibility of the bundle interventions in a variety of clinical settings (acute care to critical care). After a review of the literature, the group voted to implement the standard SPS HAPI prevention bundle with the addition of nutrition consults as appropriate in high-risk patients. Defined elements of the HAPI prevention bundle included full head to toe skin assessment, device assessment and rotation, patient positioning, appropriate bed/pressure-redistributing surface, moisture management, and nutrition consult for at-risk patients (Table [1](#T1){ref-type=\"table\"}).\n\n###### \n\nDefined Elements of the HAPI Prevention Bundle Used During Postimplementation Phase\n\n![](pqs-5-e289-g003)\n\nAs part of this assessment, the team defined standard processes for HAPI bundle education rounds. Historically, bundle education rounds were conducted variably across inpatient units. The team redefined bundle rounds for HAPI. Specifically, unit-based leadership teams dedicated a minimum of 1\u2009h/d, Monday to Friday, on bundle education rounds. Bundle education rounds focus on compliance with the bundle elements, identification of barriers to practice, and real-time education. These efforts highlight leadership opportunities to remove barriers, create accountability to practice, and demonstrate unit-level support to organizational goals.\n\nAt our institution, there has been inconsistent completion of deep dives historically. An interdisciplinary group convened to look at the current state of deep dives for HAPI and to redesign a standard process for all units. Our institution utilizes deep dives after the occurrence of HAPIs (stage 3, 4, unstageable, and deep tissue injury) to identify modifiable and nonmodifiable factors that resulted in the HAPI occurrence. These deep dives require a team approach and involve the unit clinical nurse specialist, bedside nurses, unit management, and physicians, in coordination with the wound ostomy team. The deep dive process is completed partially by the unit and partly by the wound care team. The unit is responsible for following up with those who provided care for the patient and collecting information about factors that may have contributed to the event. The wound care team identifies the source of the injury through patient activities in the days before the event in conjunction with the location and severity of the injury. The wound team also looks at systemic issues or concerns that may have resulted in a higher likelihood of injury. With the newly designed process, for each SPS-reportable HAPI, there is a deep dive investigation to evaluate potential deviations in care that contributed to the event. The deep dive involves review by both certified wound and ostomy nurses, as well as unit-based leaders. The wound care nurses serve as subject matter experts and identify the suspected cause of the injury, potential contributing factors, and deviations in standard care. The deep dive looks more broadly for additional risk factors and practice deviations, such as identifying what interventions were in use, the overall patient condition, Braden QD score,^[@R7]^ and time in the operating room. Additionally, the deep dive identifies barriers to prevention efforts and steps that may prevent HAPIs in the future.\n\nHistorically, several risk assessment scales were in use within the healthcare system. One of the initial interventions was to standardize the use of a single risk scale for all inpatients, the Braden QD scale.^[@R7]^ This action has components of both data transparency and standardization. The Braden QD Scale is an updated risk scale that accounts for risk from both immobility and use of medical devices. It has been validated for use in preterm infants through age 21 years.^[@R7]^ Medical devices account for a significant percentage of pressure injuries in children, and the Braden QD allows for quantification of that risk related to medical devices.^[@R7]^ To standardize across the hospital and reduce confusion with using multiple scoring tools, we decided to utilize the Braden QD scale for all inpatients, even if over the age of 21 years. The team launched an education campaign for the Braden QD.\n\nThere is also the standardization of HAPI surveillance. Active surveillance occurs twice per month in the areas where the highest percentage of HAPIs occur in our organization, the cardiovascular intensive care unit and the pediatric intensive care unit. Nurses who have been trained by the wound ostomy team perform head to toe skin assessments on each patient within the unit in conjunction with bedside nurses. These assessments are opportunities for just-in-time coaching on HAPI reduction as well as identification of any potential or actual pressure injuries. This process is designed to be educational, so the main focus of active surveillance is coaching and modeling best practices while bringing special attention to pressure injury prevention on a routine basis.\n\nData Transparency---Actions\n---------------------------\n\nWithin the key driver grouping of data transparency, there were numerous subinitiatives to drive the availability of real-time data. Initial efforts focused on bringing clarity to data entry and reports within our electronic health record (Epic, Madison, Wis.) to define and record HAPI bundle compliance better.\n\nThe organization has a patient-level dashboard in Epic that displays each of the HAPI prevention bundle elements and whether they are compliant/noncompliant in real time. The team revised the elements within the HAPI prevention bundle patient dashboard to match the updated HAPI bundle and interventions. A smaller workgroup comprised staff from information services, wound care (consisting of certified wound ostomy nurses), and bedside nursing met several times to define the specifications for the information services Epic build, focusing on the measurement for each of the bundle elements. After this build was complete, a pilot was conducted with the SkIPE and wound care teams to solicit additional feedback and modifications. The outcome was a unit-level dashboard that aggregates the individual bundle compliance for each patient for each shift on a given unit. This dashboard enables the inpatient units to monitor their HAPI bundle compliance by individual bundle element. It allows units to easily review patients who have had HAPIs for any missed bundle elements. The transparency of data from the individual patient level to the aggregate unit level allowed unit leadership and the wound care team to easily identify trends or gaps in the HAPI bundle elements and address them in real time. The Braden QD Scoring was also incorporated into the HAPI bundle round report in Epic for ease of use in the identification of patients at risk for HAPIs for unit leadership to more effectively target staff for bundle education rounds and allocate resources to the patients most at risk.\n\nAccountability---Actions\n------------------------\n\nWithin the key driver grouping of accountability, there were numerous subinitiatives to drive both accountability and communication. For our fiscal year 2019 (September 2018--August 2019), HAPI reduction was a core quality and safety goal for the organization. This goal was a shift from prior years, where a HAC aggregate metric had been the focus of our organization. This shift in focus from a HAC aggregate to HAPI reduction resulted in significantly more attention and engagement in HAPI reduction efforts. Accordingly, inpatient units developed their own HAPI reduction goals, beginning in July 2018. An email was sent to all staff from the executive offices to commend units for their commitment to HAPI reduction, make staff aware of the planned organization-wide efforts to support HAPI reduction, and request that additional unit-based nonenterprise-wide interventions not be deployed without involvement of the enterprise-wide HAPI reduction team.\n\nWe created a HAPI Oversight structure within our quality and safety governance structure. This structure included the creation of the enterprise-wide HAPI reduction team, monthly report outs from that team on the progress in standardization, data transparency, and accountability groups to the hospital HAC Steering Committee. It also included the establishment of a HAC Leadership Meeting in which the leaders of the enterprise-wide HAPI reduction teams reported out monthly on progress and barriers to quality and nursing leadership.By providing visibility to the HAPI reduction activities, the structure helped drive accountability and allow hospital leadership to provide resources or support when the team readily encountered barriers.\n\nAnother area of focus for the accountability key driver was the restructuring of our unit-based SkIPE team. The SkIPE team members, a unit-based resource for skin and wound concerns, ensured the use of appropriate prevention efforts, provided education for areas of individual practice deviation, and facilitated the implementation of nursing practice changes related to pressure injury prevention. As part of our improvement initiative, changes were made to improve the accountability of the SkIPE team by creating and enforcing clear expectations in terms of meeting attendance, activities on their units, and participation in pressure injury prevalence surveys. Additionally, the wound care team worked with individual unit managers to reinforce the accountability of SkIPE team member expectations.\n\nStatistical Analysis\n--------------------\n\nWe calculated the rate of SPS-reportable HAPIs per 1,000 patient days during both the pre- and postimplementation phases and compared mean rates using a 2-sided *t* test assuming unequal variances. Statistical significance was defined as a *P* value \\<0.05. We performed statistical process control analysis with the use of Shewhart control charts. Based on the control limits from the preimplementation period, we looked for patterns in the postintervention period that would permit us to invoke control chart rules that indicate a sustained decrease in the mean of the process. The control limits were modified at the point of implementation, to show process change. The evaluation was performed using Excel (Microsoft, Redmond, Wash.).\n\nRESULTS\n=======\n\nThe results are shown in Figure [3](#F3){ref-type=\"fig\"}. The mean SPS-reportable HAPI rate for the preimplementation phase was 0.3489 and for the postimplementation phase was 0.0609. The difference in rates was statistically significant (*P* \\< 0.00032). This change equates to an 82.5% reduction in HAPI rate. This well exceeded the project goal of a 10% reduction. The number of significant pressure injuries during the preimplementation phase was 49, and the postimplementation phase was 7. This reduced HAPI rate has been sustained thus far through March 2020.\n\n![Control chart showing SPS-reportable HAPI rate at LPCHS. The black vertical dotted line represents the time of process change. The turquoise line represents centerline based on preimplementation data. The dark green line represents centerline postimplementation. The dashed red line is the upper confidence limit and lower confidence limit.](pqs-5-e289-g004){#F3}\n\nThe centerline (mean) displayed on the control chart is calculated based on data from the preimplementation period (mean = 0.3489), and these data determined whether there was statistical process control. The centerline for the postimplementation period (July 2018 to August 2019) is shown in green (mean = 0.0609). In the preimplementation period, 1 point (January 2018) is above the upper confidence limit, signifying special cause variation. In the postimplementation period, all 14 consecutive points fall below the centerline from the preintervention period, satisfying the Shewhart process change rule of at least 8 consecutive points on 1 side of the centerline, signifying a statistically significant shift in the centerline postimplementation.\n\nDISCUSSION\n==========\n\nThe focus on HAPI reduction at our organization has been on SPS-reportable injuries. Because of the associated morbidity and cost of these HAPIs, many organizations target the reduction of HAPIs in these stages. These organizations include the National Quality Forum,^[@R5]^ the Center for Medicare and Medicaid Services,^[@R6]^ SPS,^[@R1]^ as well as many U.S. health departments. Despite our initial efforts to adhere to the SPS 3-pronged approach,^[@R1]^ we were unsuccessful in reducing our baseline SPS-reportable HAPI rate, and it remained well above the SPS average. Our institutional rate significantly decreased after we paused individual unit HAPI reduction efforts and implemented an organization-wide improvement initiative with key drivers in the areas of standardization, data transparency, and accountability with many action items in each of those areas that. Standardization provided 1 standard screening tool, clear alignment between bundle elements and interventions, and real-time education. Data transparency provided bedside clinicians visibility to HAPI bundle compliance rates and unit leadership a modality to compare bundle compliance with HAPI rate. The accountability provided clear strategic alignment across the organization, governance structure for oversight and implementation, and reset of the SkIPE team membership.\n\nThis study has several limitations. First, as the HAPI reduction improvement initiative consisted of many different subinitiative action items which were launched in a staggered overlapping fashion, it is not possible to attribute which action items contributed the most to the HAPI reduction. Such a \"bundle\" approach is not uncommon with improvement initiatives. Related, the roll-out of the subinitiative actions occurred in a busy pediatric healthcare system with many competing other initiatives and projects. Therefore, because of issues with the timing of allocation of resources, such as information services, it was not possible or ever envisioned that all of the action plans would be enacted simultaneously. The actions were rolled out over time in an overlapping fashion. In this retrospective look back to review the effectiveness of the overall initiative, the point in time separating the preinitiative to the postinitiative time frames is somewhat arbitrary. July 1, 2018, was chosen for the reasons described in the Methods.\n\nDespite these limitations, the overall initiative had a substantial impact on decreasing the HAPI rate. The reduction of the HAPI rate by 82.5% well exceeded the initially established goal of a 10% reduction and was also statistically significant. This reduction shows not only a direct improvement in the care that we provide to children but has also had a positive impact on other processes. The HAPI reduction success was used as an example of an improvement project for our hospital's recent successful application for Magnet certification and also will have a positive impact on external evaluation processes such as the U.S. News and World Report process for ranking of children's hospitals. For all of these reasons, the success has been a source of pride for the staff and physicians who work in our healthcare system. We hope that success has also engaged our workforce so that they will be more apt to embrace other improvement initiatives launched in the future openly.\n\nDISCLOSURE\n==========\n\nThe authors have no financial interest to declare in relation to the content of this article.\n\nPublished online April 7, 2020\n\nTo cite: Johnson AK, Kruger JF, Ferrari S, Weisse MB, Hamilton M, Loh L, Chapman AM, Taylor K, Bargmann-Losche J, Donnelly LF. Key Drivers in Reducing Hospital-acquired Pressure Injury at a Quaternary Children's Hospital. Pediatr Qual Saf 2020;2:e289.\n"}
+{"text": "***Cite this article as:*** Sabzghabaei F, Salajeghe M, Soltani Arabshahi SK. Evaluating ambulatory care training in Firoozgar hospital based on Iranian national standards of undergraduate medical education. Med J Islam Repub Iran. 2017 (17 Dec);31:99. https://doi.org/10.14196/mjiri.31.99\n\n\u2191 What is \"already known\" in this topic: {#box1}\n========================================\n\nEvaluating the quality of education is one of concerns in Iranian medical education, specifically in ambulatory section. That highlights the urgency of paying extra attention to the effectiveness of the higher education system.\n\n\u2192 What this article adds: {#box2}\n=========================\n\nThe national standards of ambulatory training in undergraduate medical education are highly useful in evaluating ambulatory training. There is an appropriate approach towards applying the national standards of ambulatory training in undergraduate medical education in Firoozgar Hospital.\n\nIntroduction {#s1}\n============\n\nIranian medical universities need to improve the quality of their medical education and constantly evaluate its status because of their important role in promoting medical education and providing health care and sanitation services.\n\nClinical training mainly concerns training and learning about the patients and their problems, which is the backbone of medical education. Since most of the patients seek ambulatory care services in ambulatory care centers and only a few of them go to specialty hospitals, it is clear that training the medical students in wards only fails to familiarize them with health care and treatment problems of the society ([@R1], [@R2]). According to the rules set by The Ministry of Health and Medical Education (MHME), 50% of the training for the medical students should be in ambulatory care centers ([@R3]). Specific features of medical training in ambulatory centers on the one hand and the fact that the main portion of the general practitioners' activities is performed in ambulatory care centers on the other hand have made it impossible for traditional methods of education to be satisfactory even with massive modifications done to them. Despite their plentiful advantages, reaching efficiency in clinical training in ambulatory care unit is challenging ([@R4]). However, no comprehensive study has been conducted in Iran to cover different aspects of the quality of ambulatory care training. However, some studies such as Alizadeh et al. in Shiraz indicated that most of the learners believed the present ambulatory care training falls short of meeting the forthcoming needs and requirements of general practitioners ([@R5]). Payvandi et al. in Semnan ([@R6]) and Amini et al. in Tabriz ([@R7]) found that more than half of the professors considered the ambulatory care training to be insufficient; moreover, the unsatisfactory findings of studies done by Shaigan et al. in Isfahan University of Medical Sciences and Payvandi et al. in Semnan University of Medical Sciences about ambulatory care training ([@R8], [@R9]) confirmed the insufficiency and the shortcomings of this significant part of medical education in Iran. Such findings indicate the lack of sufficient attention to ambulatory training nationwide and the absence of a proper and binding plan to improve the quality of ambulatory trainings.\n\nOn the other hand, analyzing the trends in higher education in Iran reveals that the utmost concern in Iranian educational issues in the recent years has rested upon the declining trends in the indicators of the quality of education, which highlights the urgency of paying extra attention to the effectiveness of the higher education system ([@R10]). Also, it should be taken into account that evaluating and improving the quality of education is one of the important objectives in higher education worldwide ([@R11]).\n\nAccording to the Fifth Plan of Development and the ratifications of Planning Council of MHME, to implement the general policies of the country, special heed was cast upon inaugurating and streamlining the accreditation and evaluation systems of the institutions and educational courses ([@R12]). Henceforth, considering the society's ethical codes of conduct and its extraordinary conditions concerning education and health care services, the medical universities of Iran are to constantly observe and check the status quo via premeditated and prearranged criteria.\n\nAll these factors stress the need for an appropriate and practical evaluating system in medical universities, specifically in ambulatory section considering its conditions, characteristics, and special whereabouts to evaluate the status quo and improve the quality of education ([@R13]).\n\nTo tackle the paramount issue of evaluating the quality of education in universities, various mechanisms and patterns are being deployed; and in the meanwhile, accreditation as a well-known apparatus is being taken advantage of in top universities and has a substantial place in evaluating and monitoring stages ([@R14]). Accreditation is granting a license to or certifying an educational center, which is being judged by the authorities in a certain field and has observed all specifications ([@R15]). By ambulatory training accreditation we mean that considering all the necessary standards for each constituent element of the program and comparing them with those elements in ambulatory centers, we should evaluate the aforementioned elements to eliminate all the deficiencies and faults and meet the acceptable international criteria in quality for ambulatory centers ([@R16]).\n\nOne of the most creditable guidelines for accreditation of medical education is that of World Federation of Medical Education (WFME), with which different countries have evaluated their medical university education systems ([@R17]-[@R23]).\n\nThe systematized measures to accredit undergraduate medical education in Iran commenced in 3/3/2007, with preparing Iranian national standards of undergraduate medical education in 4/16/2007, which were communicated to the medical universities to take effect to improve the quality of medical education ([@R24],[@R24]).\n\nTherefore, with regard to the obligation for deploying this plan and the need in Firoozgar hospital for evaluating and improving the quality of its ambulatory training, a team consisting of ambulatory care professors of this hospital conjoined with medical education experts in coordination with the national and worldwide movement to evaluate and improve the quality of education in ambulatory centers, instigated the necessary measures to conduct this study.\n\nOn the other hand, deploying standards is a management activity, which is closely interconnected with organizational excellence models. Malcolm Baldrige Excellence Model, which has optimized total quality management (TQM) techniques for education-based health care organizations, is able to create a mutual understanding and conceptual common ground among professors, medical education development centers, and the managers. In this study, to elucidate the management-related obstacles confronting the deployment of the national standards of undergraduate medical education in ambulatory training section, Baldrige framework's scoring matrix was used. Using Baldrige framework's scoring matrix, one can evaluate the deployment of a criterion along 4 dimensions:\n\n***Approach:*** In this dimension, the educational plan (process) of the managers and their attitude towards the demand of deployment of such a criterion is evaluated.\n\n***Deployment:*** In this dimension, the measures taken towards deploying the national standards of ambulatory training throughout the organization or each section and their indicators are evaluated.\n\n***Learning:*** In this dimension, the data collected following the deployment in each section of the standard and its indicators would be systematically assessed and the feedback is offered to the people\n\n***Integration:*** In this dimension, data collected from relevant indicators in one section or other sections of the standard system would be comprehensively and integrally studied and interpreted improvement-wise.\n\nThe present study aimed at determining the degree of the conformity of ambulatory training status quo in Firoozgar hospital with Iranian national standards of undergraduate medical education related to ambulatory training based on Baldrige Excellence Model.\n\nMethods {#s2}\n=======\n\nThis was a descriptive analytic research. Surveying the documents, interviewing, and field observation of ambulatory training processes, with the help of experts of each related specialized field and according to sections and indexes of the national standards of undergraduate medical education for ambulatory training were the techniques adopted in this study to evaluate the status of ambulatory training in Firoozgar hospital. Furthermore, a Ph.D. student of medical education held a number of training sessions for the experts, who were specialists in various but relevant fields in educational ambulatory centers. In these sessions, it was reviewed how to fill out and score index checklists using Baldrige Education Scoring Matrix and their \\[specialists\\'\\] questions were answered and discussed. Through triangulation of data collection methods---interviewing, observation, and surveying the documents---and also triangulation of the sources from which we collected the data---professors, students, and education managers---we tried to increase the reliability and verifiability of the data. The research tool used was the standard checklists released from the secretariat of undergraduate medical education council, which evaluated the 4 domains of educational program, evaluation, training and research resources, and faculty members in the clinics of Firoozgar hospital. Among the indexes of the 7 checklists, which covered the whole course of medical education, the indexes of 4 checklists related to ambulatory training were selected using expert panel method to ensure the face and content validity. To ensure the reliability of the checklists, we used interrater reliability, which was tested using kappa coefficient.\n\nThe education program checklist included 4 subcategories (the development and communication of educational goals,\u00a0training and the duties of the teachers and students, outcomes of educational program, content of the curriculum, and educational strategies), 9 indexes, and 23 indicators.\n\nThe evaluation checklists included 1 subcategory (program evaluation), 4 indexes, and 15 indicators.\n\nThe training and research resources checklist included 2 subcategories (space and facilities, and development of medical education), 3 indexes, and 7 indicators.\n\nThe faculty members' checklist included 1 subcategory (recruitment of faculty), 3 indexes, and 3 indicators.\n\nTo calculate the sample size, since the main body of the education was provided in major clinics, those clinics were selected as samples through census method. Major specialties in this hospital are 3 main groups of internal medicine, surgery, and gynecology, which are practiced in 16 clinics.\n\nThe appropriateness of the current condition of ambulatory training in Firoozgar hospital in each national standard indicator, based on the practical definition of the study, was rated as follows: 1: very appropriate (gaining more than 80% of the score); 2: appropriate (gaining more than 60% up to 80%of the score); 3: medium (gaining more than 40% up to 60%of the score); 4: weak (gaining more than 20% up to 40% of the score); 5: very weak (gaining less than 20% of the score).\n\nFor data analysis, the national standards of undergraduate medical education (NSUME) in ambulatory training and Baldrige Scoring Matrix were used and gap analysis was utilized to determine the gap between status quo and NSUME.\n\nScoring was based on the scale of 0 to100 for answering each question on the checklists, based on Baldrige Excellence Model Scoring Matrix, with score coefficient of 1 for approach, score coefficient of 2 for deployment, 3 for learning, and 4 for integration. Therefore, each indicator in each of the 3 techniques of data collection could receive between 0 to 1000 score, and the average score of the 3 indicators would be considered that indicator\\'s score. Likewise, the average score of all indicators in each domain would be considered the score and the percentage of the gained score for it. Using this matrix, the level of all domains separately and all highlighted clinics of Firoozgar hospital were determined in 4 dimensions of approach, deployment, learning, and integration.\n\nResults {#s3}\n=======\n\nThe 16 studied clinics were divided into 3 groups of internal medicine, surgery, and gynecology. In surgery group, there were 5 clinics of general surgery (2 clinics), neurosurgery, orthopedics, and ENT (ear, nose, and throat). In internal medicine group, there were 9 clinics of hematology, neurology, rheumatology, cardiology, pulmonary, infectious diseases, gastroenterology, and nephrology (2 clinics), and 2 clinics were evaluated in gynecology group.\n\nThe domain of education programs (EP) {#s3-0-4}\n-------------------------------------\n\nBased on the evaluation results of the clinics in the surgery group, it was revealed that this group\\'s performance in the education program (EP) domain was appropriate ([Table 1](#T1){ref-type=\"table\"}). This achievement was due to high scores in the following indexes: preparing and communicating objectives and duties (PCOD), curriculum of science stage (CSS), curriculum of clinical stage (CCS), applying the latest educational strategies (ALES), and professors and assistants' cooperation (PAC). In this domain, the performance in other indexes was evaluated as medium.\n\n###### The status of the education program domain based on national standards of ambulatory training in Firoozgar hospital using Baldrige Excellence Model Matrix\n\n  ----------------------- ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n                          Total score\\                             Clinics of internal medicine group   Clinics of surgery group   Clinics of gynecology group\n                          Triangulation and weighting                                                                              \n\n  Domain                  Dimension of Approach\\                   65.14                                79.69                      68.08\n                          coefficient 1                                                                                            \n\n  Education               Dimension of Deployment coefficient 2    127.8                                133.27                     130.54\n\n  Program                 Dimension of Learning coefficient 3      184.98                               185.15                     187.16\n\n  Domain                  Dimension of Integration coefficient 4   240.43                               243.60                     242.37\n\n  Total score (of 1000)                                            618                                  631                        627\n\n  percentage                                                       61.8%                                63.1%                      62.7%\n\n  Result                                                           Appropriate                          Appropriate                Appropriate\n  ----------------------- ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n\nThe internal medicine group had also an appropriate performance in this domain ([Table 1](#T1){ref-type=\"table\"}). The achievement was mainly due to high scores in the following indexes: PCOD, CCS, ALES, and PAC. In this domain other indexes scored medium.\n\nThe gynecology group showed an appropriate performance in EP domain ([Table 1](#T1){ref-type=\"table\"}). This achievement was mainly due to high score in the following indexes: PCOD, CSS, CCS, PAC, and observing professional values and requirements (OPVR). In this domain other indexes scored medium.\n\nThe domain of evaluation {#s3-0-1}\n------------------------\n\nThe results revealed that the performance of the surgery group clinics in this domain scored medium ([Table 2](#T2){ref-type=\"table\"}). In this domain, preparing and announcing evaluation procedure (PAEP) index was the most appropriate. On the contrary, test validation (TV) and collaboration in assessment (CA) indexes scored very weak and index of commensurability of evaluation with educational objectives (CEEO) scored medium. The main weakness in this index was the low score in the indicator of existence of evaluation system for practical and clinical education in colleagues\\' opinion (EESPCECO).\n\n###### The status of evaluation domain based on national standards of ambulatory training in Firoozgar hospital using Baldrige Excellence Model Matrix\n\n  ------------ ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n               Total score\\                             Clinics of internal medicine group   Clinics of surgery group   Clinics of gynecology group\n               Triangulation and weighting                                                                              \n\n               Dimension of Approach\\                   52.95                                62.91                      49.5\n               coefficient 1                                                                                            \n\n  Evaluation   Dimension of Deployment coefficient 2    97.9                                 114.13                     93.5\n\n  Domain       Dimension of Learning coefficient 3      142.8                                167.75                     137.12\n\n               Dimension of Integration coefficient 4   188.15                               219.33                     178.5\n\n               Total score (of 1000)                    481                                  563                        458\n\n               Percentage                               48.1%                                56.3%                      45.8%\n\n               Result                                   Medium                               Medium                     Medium\n  ------------ ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n\nInternal medicine group of clinics\\' performance scored medium in this domain ([Table 3](#T3){ref-type=\"table\"}). In this domain, the most appropriate index was CEEO, while TV and CA scored medium. In TV and CA indexes, there was considerable discrepancy between the score gained from surveying the documents and interviewing methods and observation method. This discrepancy was due to the difference in individual and organizational status of the participants and the effect this difference had on their attitudes. This means that both clear and fixed methods of evaluation and valid and stable tests in students\\' educational life are highly important; and with regards to cognitive dissonance theory ([@R25]), this importance leads to colossal influence of any unfavorable prior experience on framing their attitudes. On the other hand, it is quite natural for professors who, for years, have assessed their students with relatively stable evaluative methods and considered them as being completely clear, credible, and valid. It is suggested to those in charge to pay special attention to validation of these tests.\n\n###### The status of training and research resources domain based on national standards of ambulatory training in Firoozgar hospital using Baldrige Excellence Model Matrix\n\n  -------------- ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n                 Total score\\                             Clinics of internal medicine group   Clinics of surgery group   Clinics of gynecology group\n                 Triangulation and weighting                                                                              \n\n                 Dimension of Approach\\                   42.55                                54.41                      48.12\n                 coefficient 1                                                                                            \n\n  Research and   Dimension of Deployment coefficient 2    82.5                                 106                        96.12\n\n  education      Dimension of Learning coefficient 3      116                                  157.08                     142.25\n\n  domain         Dimension of Integration coefficient 4   156.3                                203.97                     187.37\n\n                 Total score (of 1000)                    397                                  521                        473\n\n                 Percentage                               39.7%                                52.1%                      47.3%\n\n                 Result                                   Weak                                 Medium                     Medium\n  -------------- ---------------------------------------- ------------------------------------ -------------------------- -----------------------------\n\nThe performance of the gynecology group was medium in evaluation domain ([Table 2](#T2){ref-type=\"table\"}). In this domain, the most appropriate indexes were PAEP, TV, and CEEO, which compared to standards of this study, gained medium scores, and the weakest index was CA that gained weak scores in all the 3 methods of data collection.\n\nThe domain of training and research resources (TRR) {#s3-0-2}\n---------------------------------------------------\n\nThe surgery group clinics showed weak performance in this domain ([Table 3](#T3){ref-type=\"table\"}). The main weakness in this domain was due to weak performance in research specifications and facilities (RSF). In this domain, the most appropriate index was commensurability per capita and with standards (CPCS), and in the meantime, ambulatory education facilities and learning spaces (AEFLS) scored medium. Although the indicators of RFS index in surveying the documents method and interviewing method scored very weak, the experts considered medium scores in their observations for this indicator. Hence, it can be concluded that the ambulatory training learning space is potentially a suitable source for conducting research and that education officials and administrators should take the necessary measures to properly take advantage of this rich source.\n\nThe internal medicine group of clinics\\' performance scored medium in TRR domain ([Table 3](#T3){ref-type=\"table\"}). In this domain, the most appropriate index was CPCS, and AEFLS and RFS were medium.\n\nThe performance of the gynecology group in TRR domain was medium ([Table 3](#T3){ref-type=\"table\"}). In the domain, the most appropriate index was CPCS and the weakest was RSF, while AEFLS index scored medium.\n\nThe domain of faculty members (FM) {#s3-0-3}\n----------------------------------\n\nAll 3 groups of the clinics had an appropriate performance in this domain ([Table 4](#T4){ref-type=\"table\"}), which was due to appropriate performance of all 3 indexes of this domain in the 3 groups of clinics.\n\n###### The status of faculty members domain based on national standards of ambulatory training in Firoozgar hospital using Baldrige Excellence Model Matrixs\n\n  ---------------------------------------- ------------------------------------- ------------------------------------ -------------------------- -----------------------------\n                                           Total score\\                          Clinics of internal medicine group   Clinics of surgery group   Clinics of gynecology group\n                                           Triangulation and weighting                                                                           \n\n  Faculty members domain                   Dimension of Approach coefficient 1   67.26                                71.81                      62.49\n\n  Dimension of Deployment coefficient 2    128.99                                141.88                               130.83                     \n\n  Dimension of Learning coefficient 3      192.19                                211.66                               194.99                     \n\n  Dimension of Integration coefficient 4   254.39                                281.55                               257.16                     \n\n  Total score (of 1000)                    643                                   705                                  645                        \n\n  Percentage                               64.3%                                 70.5%                                64.5%                      \n\n  Result                                   Appropriate                           Appropriate                          Appropriate                \n  ---------------------------------------- ------------------------------------- ------------------------------------ -------------------------- -----------------------------\n\nDiscussion {#s4}\n==========\n\nThe domain of education programs (EP) {#s4-0-1}\n-------------------------------------\n\nAll 3 groups of the clinics observed national standards in this domain at the appropriate level. Baldrige Scoring Matrix also showed a satisfactory status for the 3 groups. Medium performance scores in preparing and communicating educational objectives and professors and students\\' duties (PCEOPSD) subcategory in the clinics of surgery group pinpoints the fact that the educational program that is prepared in the format of a booklet titled, \\\"study guide\\\" is either not distributed properly among students or not taken as seriously as its significance mandates.\n\nIn the national standards of medical education, in EP domain, the emphasis is on the relevance of social and behavioral sciences with clinical activities. In Firoozgar hospital\\'s education program, in compliance with realization of these objectives, there are somewhat planning and some endeavors, but it is hardly sufficient and there are also in-service trainings in the existing trainings of the program to empower the graduates with clinic management and pay due attention to people\\'s physical, spiritual, mental, and social health.\n\nAlso, the opinion polls have found that generally there is no positive attitude towards communicational skills in the patient-physician relationship. Thus, to improve the status of this skill, necessary targeting and trainings should be done and doctors should learn those communicational skills and be obliged to observe them. Moreover, following top-notch methods and approaches of education and modern technology including bolstering IT and utilizing education\\'s scientific principles can be helpful in compensating for the deficits of the program.\n\nFarzianpur et al. study in Tehran University of Medical Sciences, which has been conducted using internal evaluation method based on interviewing managers and faculty members of 9 departments, revealed 73% conformity between EP domain of the university and the basic level of WFME guideline ([@R9]). This endorses the average results of EP domain in the present study (62%/64%). However, Mirzadeh et al. study conducted in Tehran University of Medical Sciences, performed using data collection, observation, and interviewing, and sending questionnaire to all people involved including professors, students, and managers, suggested that the status of medical education in Tehran University of Medical Sciences is medium or higher than national standards\\' recommendations, which themselves are derived from the basic level of WFME guideline, in only 40% of the cases ([@R17]).\n\nShekarchi et al. study in the AJA\\'s Medical University, which was conducted using document review method, showed conformity of 54/37 \u00b1 9.03% between this university\\'s EP and the basic level of WFME guideline ([@R22]), which is commensurate with our present study.\n\nKhaje-Azad et al. study in Baghiat Allah University indicated that the quality of EP based on WFME guidelines is \\\"between basic and quality\\\" level according to professors and \\\"lower than basic\\\" level according to students and is appropriate based on the national standards according to the professors, and weak according to the students ([@R11]), which considering the national standards results, the professors\\' opinion conforms to the results of the present study.\n\nKarimian et al. study in Shiraz University of Medical Sciences suggested that according to head of departments, EP domain in that university has had an appropriate performance in curriculum design index and utilizing new educational approaches (UNEA) index ([@R12]). The results of this study in detecting appropriate indexes conform to the findings of the present study in UNEA.\n\nAshurian et al. study in Isfahan University of Medical Sciences stated that EP domain in that university in medicine was 80% commensurable with national standards of medical education of Iran ([@R7]), which supports the results of the present study.\n\nThe domain of evaluation {#s24-0-3}\n------------------------\n\nIn the evaluation domain, all 3 groups of clinics follow national standards in medium level. One of the strengths of the evaluation domain in Firoozgar hospital\\'s ambulatory training is the existence of a prearranged process and program for assessing the students in all 3 groups of clinics and also in a guidebook titled, \"Clinical Lesson Plan and Log Book\". In the section related to \\\"Recording the Information about Presence in the Hospital\\\", the emphasis is on visiting the patient, how to record patient\\'s history, performing careful examination, registering accurate and correct data, and treating the patient properly; in addition, at the beginning of each note book the objectives of each course are mentioned.\n\nEven though there are various assessments including evaluating professors, students and tests, there is no satisfactory and practical mechanism to give feedback to those who are evaluated, do follow-ups, and monitor the assessments. Also, validity and reliability of the assessment tools are not verified.\n\nShekarchi et al. study in AJA\\'s Medical University, which has been performed applying document review method, demonstrated conformity of 50 \u00b1 4/24% between evaluation domain in this university and the basic level of WFME guidelines ([@R22]), which is commensurate with the findings (48.06) of the present study. However, in Farzianpur et al. study in Tehran University of Medical Sciences, which has been conducted using internal evaluation method based on interviewing managers and faculty members of 9 departments, the percentage of conformity of evaluation domain of that university with the basic level of WFME guideline was 71.5 ([@R26]).\n\nKhaje-Azad et al. study in Baghiat Allah University indicated that the quality of evaluation domain based on WFME guidelines is \\\"between basic and quality\\\" level and based on the national standards is appropriate ([@R11]), which confirms the results of the present study.\n\nAshurian et al. study in Isfahan University of Medical Sciences pinpoints that the evaluation domain of that university in medicine in 61% of the cases conforms to the national standards and is semi-appropriate ([@R7]), which verifies the findings of our study.\n\nKarimian et al. study in Shiraz University of Medical Sciences reveals that more than 50% of the head of the departments in Medical Faculty considered VE index to be in an inappropriate statue ([@R12]), which substantiates the findings of our study in rendering this index as not appropriate.\n\nThe domain of training and research resources (TRR) {#s4-0-4}\n---------------------------------------------------\n\nIn this domain, surgery group of the clinics scored weak in observing the standards, but the other 2 scored somewhat medium. Lack of adequate facilities for pursuing and conducting research projects was the primary reason of poor performance in this domain. Although utilizing richly valuable potentials like research centers in clinics---especially education-based clinics---is an element that can be quite effective and helpful and noting the mechanisms, which have been taken into account in this field; namely, noticing research priorities in dissertations, students\\' active participation in research projects, establishing research centers for students in hospitals, sufficient supporting for exceptional talents\\' researches, and constructive and purposeful interactions with other credible scientific institutes. There is no clear definition of facilities and research activities suitable for ambulatory learning spaces; therefore, it is the responsibility of education officials of the hospitals to ratify policies for the hospital clinics that support research in education.\n\nTo this end, supporting the education development office of the hospital is of utmost importance.\n\nShekarchi et al. study in AJA\\'s Medical University, which has been performed using document review method, demonstrated conformity of 45.83% \u00b1 14.28 between the status of TRR domain of this university and the basic level of WFME guidelines in TRR domain ([@R22]), which corroborates the findings of our study in this domain (46.37%).\n\nFarzianpur et al. study in Tehran University of Medical Sciences, which has been conducted using internal evaluation method based on interviewing managers and faculty members of 9 departments, suggested 62.2% conformity between TRR domain of this university and the basic level of WFME guideline ([@R9]), which endorses the results of the present study.\n\nAshurian et al. study in Isfahan University of Medical Sciences revealed that the TRR domain of that university in medicine in 54% of the cases conforms to the national standards and is semi-appropriate ([@R7]), which substantiates the findings of our study.\n\nThe domain of faculty members (FM) {#s4-0-5}\n----------------------------------\n\nIn FM domain, all the 3 groups of clinics observed the national standards of undergraduate medical education in an appropriate level, but they should endeavor to eradicate such minor defects such as the absence of a plan to develop faculty members, the disproportionateness of faculty members number and activities to their various roles including research, scientific development, promoting community\\'s health, streamlining, and guiding exceptional talents, and providing professional services, which all resulted from defects in planning and guiding specialized and skillful manpower to enable the faculty members to optimally spend all their energy and time on teaching, researching, taking care of the patients, and managing their responsibilities.\n\nShekarchi et al. study in AJA\\'s Medical University, which has been conducted using document review method, suggested conformity of 65% \u00b1 7.07 between the status of FM domain of this university and the basic level of WFME guidelines in FM domain ([@R22]), which validates the findings of our study in this domain (68.98%).\n\nFarzianpur et al. study in Tehran University of Medical Sciences, which has been conducted using internal evaluation method based on interviewing managers and faculty members of 9 departments, revealed 83.8% conformity between FM domain of this university and the basic level of WFME guideline, ([@R9]) in which, similar to our study, the staff were not studied and only the indexes about the faculty members were checked, and likewise the performance was appropriate.\n\nAshurian et al. study in Isfahan University of Medical Sciences revealed that the FM domain of that university in medicine in 74% of the cases conforms to the national standards and is appropriate ([@R7]), which verifies the findings of our study.\n\nKhaje-Azad et al. study in Baghiat Allah University indicated that the quality of FM domain based on WFME guidelines is \\\"between basic and quality\\\" level and based on the national standards, is appropriate according to the professors and is weak according to students ([@R11]), which considering professors\\' opinions in the study based on the national standards approves the results our study.\n\nConclusion {#s5}\n==========\n\nTwo domains of EP and FM were the most successful domains being studied in Firoozgar hospital, based on the national standards of undergraduate medical education in Iran, and the other 2, validation domain and TRR, had a medium performance.\n\nAccording to these findings, ambulatory training in Firoozgar hospital, based on the national standards of undergraduate medical education in Iran, in 4 domains of education program, validation, training and research resources, and faculty members was evaluated to be average. Therefore, it is suggested that considering the national standards of undergraduate medical education, ambulatory training should be provided and implemented in Firoozgar hospital and the necessary measures should be taken to make revisions in domains and indexes, which in this study, it was found that they need some improvement.\n\nOverall, according to Baldrige Matrix, there is an appropriate approach towards applying the national standards of ambulatory training in undergraduate medical education in Firoozgar hospital and most of the indexes have been deployed. In learning dimension, the main challenge was not using the systematic feedback of deploying guideline indexes\\' results in improving the quality of ambulatory training system. In integration dimension, the main challenge was the absence of a holistic view of the national standards of ambulatory training in undergraduate medical education and separate deployment of indexes in different domains of the standard. Since one of the most important sections of an evaluation research is offering suggestions to improve the plan, with regards to the results of this study, the following suggestions are put forward to improve the status of the 4 domains:\n\nIn education program domain {#s5-0-1}\n---------------------------\n\n1.  Using motivational strategies to further involve the faculty members in education programs related to ambulatory training\n\n2.  Paying more attention to professional codes of conduct via merging this category into clinical courses\n\n3.  Enforcing community-oriented education of general practitioners\n\n4.  Paying more attention to teaching ambulatory care by professors and providing them with new training strategies\n\n5.  Paying special attention to teaching appropriate diagnosis and prescription in ambulatory care\n\n6.  Paying attention to facilitating independent involvement of learners in clinics under the supervision of professors\n\nIn validation domain {#s5-0-2}\n--------------------\n\n1.  Conducting supplementary ways of evaluating professors such as evaluating by partners\n\n2.  Providing necessary mechanisms to present the professors with the feedback of evaluation results and reinforcing them\u200f\n\n3.  Placing more emphasis on education-oriented education in periodic evaluations of faculty members\n\nIn training and research resources domain {#s5-0-3}\n-----------------------------------------\n\n1.  Using existing educational spaces and resources in the society, such as private clinics, to conduct research and education projects\n\n2.  Conducting research projects and evaluation programs to more precisely examine the status of education and research resources in ambulatory care training centers\n\n3.  Paying more attention to ambulatory training in programs supporting research projects of professors\n\n4.  Paying attention to improving learning spaces and expanding academic resources in clinics\n\nIn faculty members domain {#s5-0-4}\n-------------------------\n\n1.  Paying more attention to education management and leadership to have education planning, supervising, and objective evaluation of programs\n\n2.  Paying more attention to development of faculty members and consequently the development of scientific ability of the hospital and university\u200f\n\nSuggestions for further studies {#s6}\n===============================\n\nTo evaluate ambulatory training status, future studies can involve more groups, such as graduates, graduate supervisors in health care system, patients referring to hospital clinics, and hospital staff, rather than just students, professors, and education managers.\n\nConflict of Interests {#s7}\n=====================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Introduction {#s1}\n============\n\nThe highly reactive superoxide (O~2~ ^-^), which is produced intracellularly as a common byproduct of aerobic life, is one of the main reactive oxygen species that can damage cellular components. Most of the cells exposed to oxygen possess enzymes that provide protection to superoxide toxicity by catalyzing dismutation of superoxide to hydrogen peroxide and oxygen. Extracellular superoxide is produced for instance by phagocytic leukocytes to inactivate invading microorganisms, or by heterotrophic bacteria, affecting global biogeochemistry [@pone.0091243-Diaz1]. Therefore pathogenic bacteria evolved mechanisms to inactivate superoxide molecules or their production in host macrophages [@pone.0091243-ElBekay1]. *Escherichia coli* has three superoxide dismutases (SodA, SodB, and SodC) which require different metal cofactors and are regulated in different ways [@pone.0091243-Argaman1]--[@pone.0091243-Schrum1].\n\nThe active SodA, or MnSOD, contains manganese. SodA can bind iron and manganese with similar efficiencies [@pone.0091243-Mizuno1] but the iron-substituted SodA is catalytically inactive [@pone.0091243-Vance1]. SodB, or Fe-SOD, contains iron. The cytosolic SodA and SodB proteins have overlapping functions [@pone.0091243-Brown1] but they are functionally not equivalent [@pone.0091243-Hopkin1]. The periplasmic SodC protein contains copper and zinc, and may be important to protect the cell from macrophage killing [@pone.0091243-Pesce1]--[@pone.0091243-Battistoni1].\n\nTranscription of *sodA*, but not of *sodB*, is activated by superoxide stress through the SoxR-SoxS system [@pone.0091243-Compan1], [@pone.0091243-Gaudu1]. Expression of both SodA and SodB is regulated by intracellular iron availability ([Figure 1](#pone-0091243-g001){ref-type=\"fig\"}). Transcription of *sodA* is directly inhibited by the iron-bound form of the Ferric Uptake Regulator (Fe-Fur) [@pone.0091243-Schrum1], while *sodB* transcription is independent of Fe-Fur [@pone.0091243-Niederhoffer1]. However, Fe-Fur levels affect SodB protein production through the small regulatory RNA (sRNA) RyhB, which acts by inhibiting translation initiation and by decreasing mRNA stability [@pone.0091243-Masse1]. SodA expression is also negatively controlled by RyhB [@pone.0091243-Argaman1], therefore SodA expression is inhibited through an incoherent feed-forward loop (FFL). Similar steady-state SodA activities were found in wild type and \u0394*fur* strains [@pone.0091243-Schrum1], suggesting that the FFL results in similar levels of repression at both high and low levels of free intracellular iron ([Figure 1](#pone-0091243-g001){ref-type=\"fig\"}). In order to pinpoint the function of this FFL, I investigated the performance of feed-forward regulation of SodA compared to a control system where SodA expression is independent of Fe-Fur and RyhB.\n\n![Regulation of cytoplasmic superoxide dismutases by intracellular free iron and superoxide levels.\\\nSodA expression is regulated by iron through an incoherent feed-forward loop (shown in red).](pone.0091243.g001){#pone-0091243-g001}\n\nResults {#s2}\n=======\n\nI have developed a mathematical model to study the characteristics of feed-forward regulation of SodA expression. The model is described in detail in the *Models* section. Using this model I compared the steady state and dynamic behaviors of the natural system, where SodA expression is regulated by RyhB and Fe-Fur through a FFL, with a hypothetical control where SodA production does not depend on FeFur or RyhB concentration.\n\nThe incoherent FFL keeps steady state SodA mRNA levels constant {#s2a}\n---------------------------------------------------------------\n\nSimilar SodA activities were reported in wild type cells growing in iron rich conditions and in \u0394*fur* strains [@pone.0091243-Schrum1] where RyhB is expressed at maximal levels [@pone.0091243-Masse1]. The parameter values in the model were chosen to match this experimental observation. Thus the model produced similar SodA mRNA levels at high iron (Fe-Fur-repressed) and low iron (RyhB-repressed) conditions.\n\nFirst I explored the steady-state levels of SodA at intermediate Fe-Fur levels in the absence of superoxide stress. Simulations showed that SodA mRNA levels were kept in a relatively narrow range as Fe-Fur levels were changed from 2 to 2000 nM ([Figure 2](#pone-0091243-g002){ref-type=\"fig\"}). Unlike typical regulatory systems which regulate transitions between a high and a low expression state, i.e. turn a gene on or off, this FFL regulates transitions between similar expression states. Although the SodA mRNA has similar concentrations in these states, it has substantially higher turnover (higher production and degradation rates) at low iron levels ([Figure 2](#pone-0091243-g002){ref-type=\"fig\"}, dotted line).\n\n![Steady state levels (solid black line) and production rate (dotted grey line) of SodA mRNA as a function of Fe-Fur level in the absence of superoxide stress.\\\nThe functional interactions in the feed-forward loop at low and high Fe-Fur levels are shown in bold.](pone.0091243.g002){#pone-0091243-g002}\n\nSimulations of transitions between low and high iron levels {#s2b}\n-----------------------------------------------------------\n\nTo explore how SodA mRNA levels respond to changes in iron availability, I simulated sudden changes in Fe-Fur levels ([Figure 3](#pone-0091243-g003){ref-type=\"fig\"}). SodA expression responded by a transient increase to iron depletion (\u223c2-fold), and returned to its original level after about 1.5 cell generation ([Figure 3A](#pone-0091243-g003){ref-type=\"fig\"}). Such transient expression change, i.e. exhibition of near perfect adaptation, is a typical feature of incoherent FFLs [@pone.0091243-Basu1]--[@pone.0091243-Takeda1]. It results from the delay in RyhB mediated inactivation of SodA mRNA, which depend on the rate of RyhB production relative to its targets production rates and on the efficiency of RyhB pairing with the SodA mRNA. To explore how changes in these parameters affect the shape of the response curve, I performed simulations where RyhB was produced at lower rates than the estimated upper bound used in [Figure 3A](#pone-0091243-g003){ref-type=\"fig\"} (see [*Methods*](#s4){ref-type=\"sec\"}). To obtain the same steady-state SodA mRNA levels at low and high Fe-Fur levels as with the standard parameters used in [Figure 3A](#pone-0091243-g003){ref-type=\"fig\"}, the rate of RyhB pairing with the SodA mRNA was properly increased ([Figure 3B](#pone-0091243-g003){ref-type=\"fig\"}, solid lines). The individual effects of decreased RyhB production rate and increased pairing of RyhB and SodA mRNA were simulated for comparison ([Figure 3B](#pone-0091243-g003){ref-type=\"fig\"}, dashed and dotted curves, respectively). As a result of lower RyhB production rate and more efficient complex formation between RyhB and SodA mRNA, the amplitude of the response decreased and the peak of the response occurred earlier. These effects are due to the improved ability of SodA mRNA to compete with the strong targets for RyhB binding [@pone.0091243-Mitarai1]. Similar simulations with higher sRNA production rates and decreased pairing rates did not change the response curve.\n\n![Simulations of transitions from high to low (A and B) and low to high (C) iron conditions in the absence of superoxide stress.\\\nThe Fe-Fur concentration was changed at zero time from 2000 nM to 1 nM (A and B) or from 1 nM to 2000 nM (C). The solid black curves represent RyhB levels, while the solid red and green curves represent the levels of SodA and SodB mRNAs, respectively. The dashed red line represents the control system where SodA is produced constitutively. (B) Changes in the level of SodA mRNA were simulated using different sRNA production rates (*\u03b1~S~*) and RyhB pairing rates (*\u03b4~A~*). The rate pairs used were chosen to obtain the same SodA mRNA levels at low and high iron conditions. The solid red curve represents simulations with the standard parameters (*\u03b1~S~*\u200a=\u200a280/cell gen; *\u03b4~A~*\u200a=\u200a0.0019/min). The solid blue (*\u03b1~S~*\u200a=\u200a105/cell gen; *\u03b4~A~*\u200a=\u200a0.008/min) and cyan (*\u03b1~S~*\u200a=\u200a70/cell gen; *\u03b4~A~*\u200a=\u200a0.016/min) lines show simulations where the sRNA was produced at lower rates but formed a complex with the SodA mRNA at a higher rate. The dashed and dotted lines represent corresponding simulations when only the sRNA production rate was decreased or the complex formation rate was increased, respectively.](pone.0091243.g003){#pone-0091243-g003}\n\nThe opposite effect was observed when the Fe-Fur concentration was suddenly increased ([Figure 3C](#pone-0091243-g003){ref-type=\"fig\"}). In this case SodA mRNA levels decreased because Fe-Fur blocks only the production of SodA and RyhB mRNA, and it takes more than one cell generation to clear the existing RyhB sRNA molecules from the system.\n\nSimulation of superoxide stress at low and high iron levels {#s2c}\n-----------------------------------------------------------\n\nIntracellular superoxide levels are sensed by the SoxR protein, which contains two \\[2Fe--2S\\] clusters. In its oxidized form, SoxR activates transcription of the SoxS protein, which regulates transcription of about 40 promoters. Induction of superoxide stress by paraquat results in about six-fold increase in sodA transcription [@pone.0091243-Compan1]. Although sodA transcription is increased in \u0394*fur* cells, similar fold activation was observed in wild type and \u0394*fur* cells in the presence of paraquat [@pone.0091243-Compan1]. Therefore in the model I assumed that SoxS and Fur act independently on the *sodA* promoter. Because SoxS in intrinsically unstable with an in vitro half-life of about 2 minutes [@pone.0091243-Shah1], I simulated the effect of superoxide stress simply by increasing the maximal *sodA* transcription rate ([Figure 4](#pone-0091243-g004){ref-type=\"fig\"}). I performed simulations at both low (1 nM) and high (2000 nM) Fe-Fur levels. At low Fe-Fur levels the FFL mediated system responded substantially faster than the constitutive system, both to the appearance and to the removal of superoxide stress ([Figure 4](#pone-0091243-g004){ref-type=\"fig\"} A and C). However, in the simulations at high Fe--Fur levels the response dynamics of SodA mRNA was indistinguishable in the FFL mediated and constitutive systems ([Figure 4](#pone-0091243-g004){ref-type=\"fig\"} B and D).\n\n![Simulations of changes in superoxide levels at low (left panels) and high (right panels) iron conditions.\\\nThe dominant regulatory interactions are indicated on the top. At zero time the maximal transcription rate of SodA (*\u03b1~A~*) was increased 6-fold (top panels) or decreased from this induced level to the uninduced level (bottom panels). The solid black curves represent RyhB levels, while the solid red and green curves represent the levels of SodA and SodB mRNAs, respectively. The dashed red line, which overlaps with the solid red curves in panels B and C, represents the control system where SodA is produced constitutively.](pone.0091243.g004){#pone-0091243-g004}\n\nDiscussion {#s3}\n==========\n\nIn bacteria, small regulatory RNAs are often part of feed-forward motifs which contain both protein and RNA regulators (mixed FFLs) [@pone.0091243-Beisel1]--[@pone.0091243-Mank1]. Unlike in pure transcriptional FFL motifs, where incoming regulatory signals must be integrated at the promoter of the target gene using a certain logic [@pone.0091243-Mangan1], [@pone.0091243-Hunziker1], [@pone.0091243-Buchler1], in mixed motifs the actions of the protein and sRNA regulators on target production are spatially separated and independent of each other. The function of incoherent feed-forward loops in genetic regulation has been addressed both experimentally and theoretically. For example, incoherent FFLs were shown to accelerate response times [@pone.0091243-Mangan2], provide fold change detection [@pone.0091243-Goentoro1], and generate non-monotonic input functions [@pone.0091243-Kaplan1].\n\nRegulatory effects of the mixed FFL {#s3a}\n-----------------------------------\n\nIn this work I studied the function of a mixed FFL embedded into a small regulatory network that controls expression of cytoplasmic superoxide dismutases in *E. coli* ([Figure 1](#pone-0091243-g001){ref-type=\"fig\"}). This system responds to two intracellular input signals, iron and superoxide concentrations. The system maintains SodA mRNA levels at a narrow concentration range at a wide range of iron concentrations, although the SodA mRNA turnover is higher at low iron levels ([Figure 2](#pone-0091243-g002){ref-type=\"fig\"}). Similar to dominant negative autoregulatory systems [@pone.0091243-Semsey1], this FFL is mostly responsible for regulation of expression dynamics during transitions between different environmental conditions.\n\nIn the absence of superoxide, the system responds to changes in iron levels with similar pulse dynamics as was previously predicted for incoherent FFLs where the master regulator was an activator [@pone.0091243-Goentoro1]. In the iron rich LB medium SodA and SodB transcript levels are similar [@pone.0091243-Bernstein1]. However, iron levels greatly exceed manganese levels in *E. coli* grown in iron rich conditions [@pone.0091243-Outten1], and a fraction of SodA proteins is inactive because being substituted by iron instead of manganese. When intracellular iron becomes scarce, the cell stops production of several non-essential iron using proteins, such as SodB, thus allowing essential proteins to utilize the available limited free iron pool [@pone.0091243-Semsey2].\n\nThe loss of SodB activity upon iron depletion is compensated on one hand by a decrease in the level of iron-substituted SodA, and on the other hand by a transiently increased SodA production ([Figure 3A](#pone-0091243-g003){ref-type=\"fig\"}). Simulations of the transition from an iron-depleted to an iron rich environment predict the opposite effect ([Figure 3C](#pone-0091243-g003){ref-type=\"fig\"}). In this case, SodA activity decreases because of the slower production and of the higher iron substitution rates as well.\n\nSuperimposed global controls {#s3b}\n----------------------------\n\nSodA expression is regulated by global regulators responding to intracellular iron and superoxide concentrations. The iron response system is acting through a mixed FFL, involving direct transcriptional and indirect translational regulation, while the superoxide response system acts directly at the transcriptional level ([Figure 1](#pone-0091243-g001){ref-type=\"fig\"}). Because the FFL regulating SodA consists of global regulators which are used for regulation of many other genes, it does not generate an extra cost for the cell. Our simulations suggest that the major advantage of this FFL is that it can conditionally modulate the response time of a superimposed transcriptional control mechanism, allowing faster response of SodA when SodB function is limited. At low iron levels the superoxide stress response is predicted to be about three times faster compared to the control system, while at high iron levels the responses are identical ([Figure 4](#pone-0091243-g004){ref-type=\"fig\"}). This effect is due to the mixed nature of the FFL, which allows differential regulation of SodA mRNA production and degradation rates by controlling transcription initiation (by Fe-Fur) and translation initiation/mRNA stability (by RyhB) separately. In the FFL regulated system the steady state SodA mRNA level is resulted from equilibrium of production, RyhB mediated degradation, natural degradation, and dilution. Compared to the FFL-regulated system, the constitutive system operates with a constant low SodA mRNA degradation rate, and therefore requires a smaller production rate to obtain the same steady state mRNA level. At low Fe-Fur levels, the FFL regulated system allows faster initial SodA mRNA production rate upon superoxide stress compared to the constitutive system, and reach a steady state quickly because of the increased RyhB mediated degradation, which is directly proportional to the level of SodA mRNA. The FFL also allows faster recovery from stress because of the active degradation of the existing SodA mRNAs.\n\nIn conclusion, the FFL allows faster superoxide stress response and better adaptation in iron restricted environments, such as mammalian hosts. The same regulatory system is present in pathogenic strains (e.g. *E. coli* O104:H4, Shigella flexneri 2a str 301), suggesting that the fast response may help in inactivating superoxide molecules generated by the immune system.\n\nMethods {#s4}\n=======\n\nThe dynamical variables I keep track of in our model are the concentrations of RyhB (*S*), SodA mRNA (*m~A~*), and SodB mRNA (*m~B~*). Similar to previous models of sRNA regulation [@pone.0091243-Mitarai1], [@pone.0091243-Semsey2]--[@pone.0091243-Sneppen1], I assume that: (i) the degradation of the sRNA-mRNA complex is faster than the dissociation of the same complex, so that the binding is effectively irreversible; (ii) both the sRNA and the mRNA are inactivated [@pone.0091243-Masse2]; (iii) translation of the mRNA is not possible after the complex with the sRNA is formed. I also assume that 1 nM corresponds to one molecule per cell [@pone.0091243-Halford1], [@pone.0091243-Semsey3].\n\nThe deterministic differential equation that model RyhB dynamics is:\n\nThe first term represents the production of RyhB, which is repressed by Fe-Fur (*F*). In this term, *\u03b1~S~* is the maximal production rate of RyhB per cell generation time [@pone.0091243-Mitarai2]. *K~D~* (\u200a=\u200a0.02 \u03bcM) is the binding constant of Fe-Fur to its operator site [@pone.0091243-Mills1]. In the second and third terms *\u03c4~S~* (\u200a=\u200a30/ln2 min) [@pone.0091243-Masse2] and *\u03c4~G~* (\u200a=\u200a40/ln2 min) represents the passive degradation and dilution (by cell devision) of RyhB, respectively. The last two terms represent the active degradation of the RyhB-mRNA complexes by RNaseE [@pone.0091243-Masse2]. SodA dynamics was modeled by the following equation\n\nBased on *sodA* promoter activities reported in wild type and \u0394*fur* strains [@pone.0091243-Schrum1], I assume that that the *sodA* promoter has a basal activity even in the presence of Fe-Fur. The first term represents this basal activity, while the second term represents the Fe-Fur regulated activity. The effect of superoxide stress through the SoxR-SoxS system is not modeled explicitly; instead, it is simulated by increasing the maximal *sodA* promoter activity (*\u03b1~A~*). The next two terms represent dilution and passive degradation (*\u03c4~A~*\u200a=\u200a11.8/ln2 min) [@pone.0091243-Bernstein1] of the SodA mRNA, while the last term represents the sRNA mediated degradation. The parameter for RyhB pairing with the SodA mRNA, *\u03b4~A~*, was chosen to be 0.0019/min to obtain similar SodA mRNA levels in the absence and in the presence of Fe-Fur [@pone.0091243-Schrum1].\n\nThe dynamics of SodB is modeled by the following equation:\n\nThe first term represents the production of SodB. The next two terms represent dilution and passive degradation (*\u03c4~B~*\u200a=\u200a8.8/ln2 min) [@pone.0091243-Bernstein1] of the SodB mRNA, while the last term represents the sRNA mediated degradation. The parameter for RyhB pairing with the SodB mRNA, *\u03b4~B~*, was chosen to be 0.014/min to reproduce the experimental observation that the SodB level in *\u0394fur* cells is 13% of the wild type level [@pone.0091243-Dubrac1].\n\nThere are several other RyhB target mRNAs exist in the cell, which, similar to SodB, are not regulated directly by Fur. Therefore *m~B~* represents all these mRNAs. The value for *\u03b1~B~* (63 molecules per cell generation) was chosen based on the reported level of SodB mRNA (about 5 molecules/cell) [@pone.0091243-ArbelGoren1] and on the estimate that SodB mRNA production constitutes about 40% of the production of strong targets [@pone.0091243-Mitarai2].\n\nThe values of *\u03b1~A~*, and *\u03b1~S~* were chosen in such a way that *\u03b1~A~*+*\u03b1~B~*\u200a=\u200a*\u03b1~S~*/4 (representing the upper bound for *\u03b1~S~* [@pone.0091243-Mitarai2]), and SodA production was assumed to represent 10% of the production of all targets in the absence of superoxide stress (*\u03b1~A~*\u200a=\u200a7/cell generation). I compare the above natural system to a hypothetical one where SodA levels depend only on superoxide concentration and not on Fe-Fur or RyhB levels. In this case, [equations I](#pone.0091243.e001){ref-type=\"disp-formula\"} and III remain the same, and [equation II](#pone.0091243.e002){ref-type=\"disp-formula\"} becomeswhere \u03b1~AC~\u200a=\u200a\u03b1~A~/3.3 to match the Fur-Fe regulated SodA level.\n\nI thank Kim Sneppen and Namiko Mitarai for stimulating discussions and critical reading of the manuscript.\n\n[^1]: **Competing Interests:**S. Semsey is an Associate Editor at PLOS ONE. This does not alter the author\\'s adherence to all the PLOS ONE policies on sharing data and materials.\n\n[^2]: Conceived and designed the experiments: SS. Performed the experiments: SS. Analyzed the data: SS. Wrote the paper: SS.\n"}
+{"text": "REFER TO THE PAGE 30-36\n\nThere are a number of systemic diseases with cardiovascular manifestations in which echocardiography plays a valuable role for surveillance and follow-up of anticipated cardiovascular abnormalities. Systemic autoimmune disorders are frequently associated to cardiac involvement such as pericarditis, myocarditis, valvular abnormalities or ischemic coronary disease.\n\nAnkylosing spondylitis (AS) is also a chronic systemic inflammatory rheumatic disorder that primarily affects the axial joints and mainly involves young male. Extra-articular manifestations vary widely in terms of both frequency and severity. The most common extra-articular manifestations are uveitis, bowel disease, skin, lung and kidney involvement; less frequently cardiovascular involvement occurs.[@B1] Screening for extra-articular manifestations in patients with AS is important for appropriate management as the presence of extra-articular manifestations may be a consequence of uncontrolled systemic inflammation and may influence treatment decisions.[@B2] The prevalence of cardiac pathologies in patients with AS has been reported to be 10% to 30%.[@B3] It is related to a sclerosing inflammatory process that primarily involves the aortic root and the aortic valve, and the chronic inflammation may extend into the ventricular septum, atrioventricular node, proximal bundle of His and bundle branches or fascicles. Thus, various studies indicate a higher rate of conduction disturbances, valvular heart disease and ascending aortic involvement in patients with AS compared with the normal population.[@B4]\n\nIn this issue of the journal, Park et al.[@B5] reported the early valvular and aortic involvement in young male patients with AS using transesophageal echocardiography. A total of 70 AS patients were divided into AS group I (\\< 10 years of disease duration) and AS group II (\u2265 10 years) depending on their disease duration after the first diagnosis. The thickness of aortic valve was increased in AS patients than in control (control vs. AS group I vs. AS group II, 1.2 \u00b1 0.3 vs. 1.8 \u00b1 0.7 vs. 2.1 \u00b1 0.8 mm, *p* \\< 0.01). The thickness of mitral valve was also increased in AS patients (control vs. AS group I vs. AS group II, 0.9 \u00b1 0.1 vs. 1.2 \u00b1 0.3 vs. 1.2 \u00b1 0.4 mm, *p* \\< 0.01). Aortic root diameter at the sinus of Valsalva was slightly increased in AS group II. Furthermore, aortic strain and distensibility were decreased and aortic stiffness beta index was increased in AS group II compared with control and AS group I. Mitral annulus early diastolic velocity (E\\') and systolic velocity (S\\') by tissue Doppler imaging were slightly decreased in AS group II. However there was no increased rate for valvular regurgitation (aortic and mitral valve) and for conduction disturbances.\n\nThere are some limitations in this study. As they described, the design of this study was observational and cross-sectional, and conducted in a single center. The clinical significance of the results is unclear in this design. Thus a large long-term prospective study is needed to elucidate the clinical significance of early changes in valve, aorta and myocardial function. There was no age-matched control group for AS group II. Mean ages of both control and AS group I were 27 years. However mean age of AS group II was 34 years. Thus it is unclear whether the slight differences in parameters of tissue Doppler and aortic stiffness were from difference of disease duration or from difference of age. Finally the effects of disease activities and drug therapy for AS on cardiac and aortic abnormalities were not fully evaluated in this study.\n\nDespite the limitations, Park et al\\'s study[@B5] give us useful informations about early subclinical changes of aorto-mitral valve and aortic stiffness in AS patients in the era of advanced management of AS with immunosuppressive and/or immune-modulating drugs.\n"}
+{"text": "Dear Editor,\n\nPostoperative femoral neuropathy is an uncommon complication associated with pelvic/abdominal surgery. The main mechanism underlying this neuropathy is stretching and/or prolonged compression of the nerve.[@B1][@B2] The nerve compression can be caused by the self-retaining retractors that are sometimes used during surgery directly constricting the nerve against the pelvic sidewall and inducing ischemia.[@B1]\n\nSymptoms of femoral neuropathy are weakness of ipsilateral hip flexion and knee extension, and sensory deficit on the anteromedial thigh. The prognosis is generally good, with partial or complete recovery being common. Postoperative femoral neuropathy is generally unilateral[@B2][@B3][@B4]; bilateral postoperative femoral neuropathy appears to be very rare.[@B5][@B6] Herein we report a case of bilateral femoral nerve neuropathy that occurred in a patient undergoing abdominal surgery.\n\nA thin 44-year-old woman was diagnosed with rectosigmoid cancer and underwent a Hartmann surgical procedure, which involves resection of the rectosigmoid colon with creation of a colostomy bag. The presurgery general and neurological visits did not disclose any unexpected abnormalities. She had previously been physically healthy, with no underlying diseases such as hypertension, diabetes mellitus, or peripheral neuropathy. During the surgery, she was placed in a lithotomy position and a bilateral self-retaining retractor was used to facilitate the exploration of the surgical field. The surgery took about 2 hours. During recovery from the anesthesia, the patient complained of hypoesthesia over the anteromedial side of both thighs and proximal weakness in both lower limbs; in the immediate postoperative period she could neither stand nor get out of bed.\n\nA neurological examination revealed bilateral weakness of the knee extensor and hip flexor, hypoesthesia in the anteromedial thigh, and bilateral patellar areflexia. Muscle strength was assessed on both legs on the Medical Research Council scale: the iliopsoas and quadriceps muscles scored 2/5, while the adductor, hamstrings, and tibialis anterior muscles scored 5/5. An extensive diagnostic work-up was negative for abnormalities. Neurophysiological studies performed a few days after the onset of symptoms revealed bilateral femoral neuropathy ([Table 1](#T1){ref-type=\"table\"}). The patient underwent physical therapy.\n\nAt a 6-month follow-up the patient had only partially regained sensory and motor functions; she was able to walk with a double cane. Electromyography performed 1 year after symptom onset demonstrated persistent bilateral femoral nerve neuropathy.\n\nThe femoral nerve originates from the posterior division of the ventral branch of the second to fourth lumbar roots behind the psoas muscle; it passes in the groove between the psoas and iliac muscles and descends into the thigh beneath the inguinal ligament. The pathophysiology of the nerve injury in abdominal/pelvic surgery includes compression, stretch, ischemia, and ileopsoas hematoma.[@B1][@B2] The injury is often caused by the use of a self-retraining retractor directly compressing the nerve against the abdominal wall as it courses through the body of the psoas muscle, or compression of the iliac vessels, causing ischemia to the nerve.[@B1] Individual patient factors such as comorbid conditions (e.g., diabetes or peripheral neuropathy), and factors related to the surgery (e.g., prolonged lithotomy position) might contribute to postoperative femoral neuropathy.\n\nThis case represents a rare case of postsurgery bilateral femoral neuropathy. A review of the patient\\'s clinical chart revealed that the duration of the intervention was about 2 hours, which is within the normal range for the Hartmann procedure, and that there were no surgery-related complications. This suggests that the bilateral nerve damage in this patient was most probably caused by improper retractor use.\n\nNeurologists and surgeons should be aware of this rare and potentially treacherous complication of the pelvic/abdominal surgery. Diagnosis should be prompt so that physical therapy is commenced as soon as is practicable.\n\n**Conflicts of Interest:** The authors have no financial conflicts of interest.\n\n###### Electromyography (EMG) summary of lower limb muscles in the patient with bilateral femoral neuropathy\n\n![](jcn-11-398-i001)\n\n  1st EMG (three days after onset)                                              \n  ---------------------------------- --- --- --- --- -------- -------- -------- --------\n  Quadriceps R                       N   N   N   N   Absent   Absent   Absent   Absent\n  Quadriceps L                       N   N   N   N   Absent   Absent   Absent   Absent\n  Adductor longus R                  N   N   N   N   Normal   Normal   Normal   Normal\n  Adductor longus L                  N   N   N   N   Normal   Normal   Normal   Normal\n  Gastrocn (med) R                   N   N   N   N   Normal   Normal   Normal   Normal\n  Gastrocn (med) L                   N   N   N   N   Normal   Normal   Normal   Normal\n  Tibialis anter R                   N   N   N   N   Normal   Normal   Normal   Normal\n  Tibialis anter L                   N   N   N   N   Normal   Normal   Normal   Normal\n\n  2nd EMG (one year after onset)                                              \n  -------------------------------- --- --- --- --- -------- -------- -------- ---------\n  Quadriceps R                     N   N   N   N   2+       2+       2+       Reduced\n  Quadriceps L                     N   N   N   N   2+       2+       2+       Reduced\n  Adductor longus R                N   N   N   N   Normal   Normal   Normal   Normal\n  Adductor longus L                N   N   N   N   Normal   Normal   Normal   Normal\n  Gastrocn (med) R                 N   N   N   N   Normal   Normal   Normal   Normal\n  Gastrocn (med) L                 N   N   N   N   Normal   Normal   Normal   Normal\n  Tibialis anter R                 N   N   N   N   Normal   Normal   Normal   Normal\n  Tibialis anter L                 N   N   N   N   Normal   Normal   Normal   Normal\n\nAmp: amplitude, Dur: duration, Fasc: fasciculation potentials, Fib: fibrillation potentials, IA: insertion activity, L: left, PPP: polyphasic potentials, PSW: positive sharp waves, R: right.\n"}
+{"text": "Introduction {#s1}\n============\n\nInvestigations related to the so called complex systems are widely spread among different scientific communities, ranging from physics and biology to economy and psychology. A considerable part of these works deals with empirical data aiming to extract patterns, regularities or laws that rule the dynamics of the system. In this direction, the concept of complexity measures often emerges. Complexity measures can compare empirical data such as time series and classify them in somewhere between regular, chaotic or random [@pone.0040689-Rosso1], while other complexity measures can differentiate between degrees of correlations [@pone.0040689-Rosso2]. Examples of these measures include algorithmic complexity [@pone.0040689-Kolmogorov1], entropies [@pone.0040689-Shannon1], relative entropies [@pone.0040689-Kullback1], fractal dimensions [@pone.0040689-Mandelbrot1], and Lyapunov exponents [@pone.0040689-Lyapunov1]. These seminal works are still motivating new definitions, and today there are numerous definitions of complexity, which have been successful applied to different areas such as medicine [@pone.0040689-Maes1], [@pone.0040689-Khader1], ecology [@pone.0040689-Parrott1]--[@pone.0040689-Parrott2], astrophysics [@pone.0040689-Schwarz1]--[@pone.0040689-Lovallo1], and music [@pone.0040689-Boon1], [@pone.0040689-Su1].\n\nIt is surprising that this large number of complexity measures is mainly focused on one-dimensional data, while much less attention has been paid to two and higher-dimensional structures such as images. Naturally, there are few exceptions such as the work of Grassberger [@pone.0040689-Grassberger1] and more recent Refs. [@pone.0040689-Andrienko1]--[@pone.0040689-Cai1], though some of the authors of these papers agree that a higher-dimensional approach still represents an open and subtle problem. Furthermore, as it was stated by Bandt and Pompe [@pone.0040689-Bandt1], most of the complexity measures depend on specific algorithms or recipes for processing the data which may also depend on tuning parameters. As a direct consequence, there are huge difficulties for reproducing previous results without the knowledge of details of the methods.\n\nBandt and Pompe not only raised this problem, but they also proposed an alternative method that tries to overcome the previous problems, introducing what they call *permutation entropy* -- a *natural* complexity measure for time series. There are many recent applications of this new technique that confirm its usefulness [@pone.0040689-Ouyang1]--[@pone.0040689-Ribeiro1]. In particular, Rosso et al. [@pone.0040689-Rosso1] have successful applied the Bandt and Pompe ideas together with a relative entropic measure [@pone.0040689-Lamberti1] to differentiate chaotic time series from stochastic ones. They have constructed a diagram, which was first proposed by L\u00f3pez-Ruiz et al. [@pone.0040689-LpezRuiz1], (called as complexity-entropy causality plane) by plotting the relative entropic measure versus the permutation entropy. Intriguingly, chaotic and stochastic series are located in different regions of this representation space.\n\nHere, we show that the complexity-entropy causality plane can be extended for higher-dimensional patterns. We apply this new approach in different scenarios related to two-dimensional structures and the results indicate that the method is very promising for distinguishing between two-dimensional patterns. The following sections are organized as follows. Section II is devoted to review briefly the properties of the permutation information-theory-derived quantifiers and the complexity-entropy causality plane, and also to define an appropriate way to generalize these definitions to higher-dimensional data. In Section III, we work out several applications based on numerical and empirical data. Section IV presents a summary of our results.\n\nMethods {#s2}\n=======\n\nThe ingenious idea of Bandt and Pompe [@pone.0040689-Bandt1] was to define a measure that may be easily applied to any type of time series. The method lies on associating symbolic sequences to the segments of the time series based on the existence of local order, and next, by using probability distribution associated to these symbols, to estimate the complexity quantifier. For purpose of definition, let us consider a time series composed by elements and also -dimensional vectors () defined bywhere . Next, for all the vectors, we evaluate the permutations of defined by . The possible permutations of will be the accessible states of the system, and for each state we estimate the ordinal pattern probability given by\n\nwhere the symbol \\# stands for the number of occurrences of the permutation . Now, we can apply the ordinal patterns probability distribution, , to estimate a complexity measure based on some entropic formulation.\n\nBefore advancing, we note that the previous method may be extended to higher-dimensional data structures such as images. In order to do this, we consider that the system is now represented by a two-dimensional array of size . In analogy to the vector , we define matrices () given bywhere and . Next, for all these matrices, we evaluate the permutations of defined by . The system can now access states for which we calculate the probability distribution through the relative frequencies given by\n\nFor easier understanding, we illustrate this procedure for a small array in [Fig. 1](#pone-0040689-g001){ref-type=\"fig\"}.\n\n![Schematic representation of the construction of the accessible states.\\\nIn this example we have a array (left panel) and we choose the embedding dimensions and . In the right panel we illustrate the construction of the states. We first obtain the sub-matrix corresponding to and that have as elements and, after sorting, this sub-matrix leads to the state \"0132\". We thus move to next sub-matrix and which have the elements and that, after sorting, leads to the state \"1023\". The last two remaining matrices lead to the states \"1230\" and \"0132\". Finally, we estimate the probabilities , that are, , and which are then used in the [equations (1](#pone.0040689.e064){ref-type=\"disp-formula\"}) and (2), leading to and .](pone.0040689.g001){#pone-0040689-g001}\n\nNaturally, the order procedure that defines the permutation is no longer unique as in the one-dimensional case. For instance, instead of ordering the elements of row-by-row, we could also order column-by-column. However, these other definitions will only change the \"name\" of the states in such a way that the set will remain unchanged. Thus, there is no lost of generalization in assuming a given order recipe for defining .\n\nWe note that this procedure is straightforward generalized to accomplish higher-dimensional structures (e.g., the volumetric brain images obtained via functional magnetic resonance imaging), and that it recovers the one-dimensional case by setting and . Here, for simplicity, we focus our analysis on two-dimensional structures.\n\nThe parameters and (known as embedding dimensions) play an important role in the estimation of the permutation probability distribution , since they determine the number of accessible states. In the one-dimensional case, it is usual to choose in order to obtain reliable statistics in the one-dimensional case (for practical purposes, Bandt and Pompe recommend [@pone.0040689-Bandt1]). For the two-dimensional case a similar relationship must hold, i.e., . To go further, we need to rewrite the entropic measures used in Refs. [@pone.0040689-Rosso1], [@pone.0040689-Bandt1]. The first one is called normalized permutation entropy [@pone.0040689-Bandt1] and it is obtained by applying the Shannon's entropy to the probabilities , i.e.,where and . The value of is obtained by considering all the accessible states to be equiprobable, i.e., . By definition, , where the upper bound occurs for a completely random array. We expect for arrays that exhibit some kind of correlated dynamics.\n\nThe other measure [@pone.0040689-Rosso1] is defined by.where is a relative entropic metric between the empirical ordinal probability and the equiprobable state . The quantity is known as disequilibrium and it is defined in terms of the Jensen-Shannon divergence [@pone.0040689-Grosse1] (or also in terms of a symmetrized Kullback-Leibler divergence [@pone.0040689-Lin1]) and can be written aswhere\n\nis the maximum possible value of , obtained when one of the components of is equal to one and all the other vanish.\n\nThe disequilibrium quantifies the degree of correlational structures providing important additional information that may not be carried only by the permutation entropy. In addition, for a given value there exists a range of possible values for [@pone.0040689-Martin1]. This is the main reason why Rosso et al. [@pone.0040689-Rosso1] proposed to employ a diagram of versus as a diagnostic tool, building up the complexity-entropy causality plane.\n\nResults and Discussion {#s3}\n======================\n\nIn the following, we will calculate the diagram of versus to measure the complexity and to distinguish among different two-dimensional patterns.\n\nFractal Surfaces {#s3a}\n----------------\n\nWe generate fractal surfaces through the random midpoint displacement algorithm [@pone.0040689-Fournier1]. This algorithm starts with a square. For each vertex, we assign a random value representing the surface height. Next, we add a new point located at the center of the initial square. We set the height of this point equal to the average height of the previous four vertex plus a Gaussian random number with zero mean and standard-deviation . We also add four points located at the middle segments which connects each initial vertex. For these four points, the heights are equal to the average value between the two closest vertex and the middle point plus a Gaussian random number with zero mean and standard-deviation . Now, we imagine that these 9 points represent four new squares and, for each one, we apply the previous procedure using . By repeating this process times and using , we should obtain a square surface of side with fractal properties. Here, is the Hurst exponent and is the surface fractal dimension. [Figure 2](#pone-0040689-g002){ref-type=\"fig\"} shows several surfaces generated through this procedure for different values of .\n\n![Examples of fractal surfaces obtained through the random midpoint displacement method.\\\nThese are surfaces () for different values of the Hurst exponent . For easier visualization, we have scaled the height of the surfaces in order to stay between and . We note that for small values of the surfaces display an alternation of peaks and valleys (anti-persistent behavior) much more frequent than those one obtained for larger values of . For larger values of , the surfaces are smoother reflecting the persistent behavior induced by the value of .](pone.0040689.g002){#pone-0040689-g002}\n\nWe apply our method for these surfaces aiming to verify how the permutation quantifiers and change with the Hurst exponent , as it is shown in [Fig. 3](#pone-0040689-g003){ref-type=\"fig\"}. In these 3d plots, we show the localization in the causality plane obtained for different values of evaluated from surfaces (). In [Fig. 3a](#pone-0040689-g003){ref-type=\"fig\"}, we use and (circles), and and (squares) as embedding dimensions. Note that the values of and are practically invariant under the rotation and . This invariance is related to the fact that in these fractal surfaces there is not preferential direction. In [Fig. 3b](#pone-0040689-g003){ref-type=\"fig\"}, we employ and . We note basically the same dependence but a different range for and , since this change increases the number of accessible states. These results show that our method properly differentiates fractal surfaces concerning the Hurst exponent. Moreover, we investigate the robustness of the permutation quantifiers under several realizations of the random midpoint displacement algorithm and the results show that both indexes are very stable. For example, the standard-deviation in the values of and are usually smaller than when considering .\n\n![Dependence of the complexity-entropy causality plane on Hurst exponent *h* .\\\nWe have employed fractal surfaces of size (). In (a) we plot and versus for the embedding dimensions and (circles) and also for and (squares). We note the invariance of the index against the rotation and . In (b) we plot the diagram for . We observe changes in the scale of and caused by the increasing number of states. In both cases, as increases the complexity also increases while the permutation entropy decreases. This behavior reflects the differences in the roughness shown in [Fig. 2](#pone-0040689-g002){ref-type=\"fig\"}. For values of the surface is anti-persistent which generates a flatter distribution for the values of leading to values of and closer to the aleatory limit ( and ). For values of there is a persistent behavior in the surfaces heights which generates a more intricate distribution of and, consequently, values of and that are closer to the middle of the causality plane (region of higher complexity).](pone.0040689.g003){#pone-0040689-g003}\n\nLiquid Crystal Textures {#s3b}\n-----------------------\n\nAnother interesting application is related to different patterns that a thin film of a liquid crystal exhibits. These textures are obtained by observing a thin sample of liquid crystal placed between two crossed polarizers in a microscope. The textures give useful information about the macroscopic structure of the liquid crystal. For instance, different phases have different typical textures, and by tracking their evolution one can properly identify the phase transition.\n\nWe first study a lyotropic liquid crystal under isotropic-nematic-isotropic phase transition. [Figure 4](#pone-0040689-g004){ref-type=\"fig\"} shows three snapshots of the texture at different temperatures. In this case, we clearly note the differences in the textures. The leftmost and rightmost textures are at the isotropic phase while the middle one is at the nematic phase. We observe that the pattern is very complex for the nematic phase, while for the isotropic one it is basically random.\n\n![Characteristic textures of a lyotropic liquid crystal at different temperatures and phases.\\\nThe lyotropic system used here is a mixture of potassium laurate , decanol and deuterium oxide -- suitable concentrations in order to get a isotropic nematic isotropic phase sequence [@pone.0040689-Yu1]. These images were constructed by observing the optical microscopy of a flat capillary which contains the mixture at different temperatures. Here, we have used the average value of the pixels of the three layers (RGB) of the original image and a rescaled temperature.](pone.0040689.g004){#pone-0040689-g004}\n\nWe calculate and as a function of the temperature for different values of the embedding dimensions, as it is shown in [Fig. 5](#pone-0040689-g005){ref-type=\"fig\"}. In these plots, the different shaded regions represent the different liquid crystal phases. We note that the phase transitions are successful identified independently of and . However, [Fig. 5c and 5d](#pone-0040689-g005){ref-type=\"fig\"} show a slight different dependence of and versus the temperature when considering and or and . Because the liquid crystal sample is placed in elongated capillary tube, there is a surface effect that act on the liquid crystal molecules. This effect is usually amplified at the phase transition and it is also the reason for differences between the embedding dimensions.\n\n![Dependence of the entropic indexes on the temperature of a lyotropic liquid crystal.\\\nWe plot versus the temperature in (a) and versus the temperature in (b), where we employ . Figures (c) and (d) present the results for and , and also for and . The different shaded areas represent the different liquid crystal phases. Note that the phase transitions are properly identified in all cases. Due to the asymmetry of the elongated capillary tube where the liquid crystal sample is placed, and present slight differences under the rotation and .](pone.0040689.g005){#pone-0040689-g005}\n\nIn this particular phase transition, the difference between the textures are large enough that it can be identified just by visual inspection. However, this is not the usual case and many phase transitions are very difficult to identify. In this context, an interesting question is whether our method can help to distinguish different phases. To address this question, we evaluate and for twelve characteristic textures of different liquid crystals. We download these textures from the webpage of the Liquid Crystal Institute at Kent State University [@pone.0040689-Kent1] and [Fig. 6](#pone-0040689-g006){ref-type=\"fig\"} shows the value of and for each texture in the causality plane. The results allow to conclude that the method ranks the textures in a kind of complexity order where each characteristic texture occupies a different place in this representation space. Moreover, the different values of and indicate that the permutation quantifiers can also identify smooth phase transitions.\n\nNaturally, the location of each texture in the causality plane should be related to physical properties of the liquid crystals. A better understanding of the relation between the permutation quantifiers and these physical attributes may deserves a more careful investigation since some properties of liquid crystals such as the order parameter can be quite hard to empirically measure. In this context, the existence of a clear relation between, for example, the order parameter and or will be experimentally handy. Here, we just have the pictures of the textures in such a way that is very hard to point out these relationships. However, a visual inspection of [Fig. 6](#pone-0040689-g006){ref-type=\"fig\"} suggests that some of the more ordered phases, such as the blue phase (this phase display a cubic structure of defects), are located in the central part of the causality plane (region of higher complexity), while other textures which present a large number of non-ordered defects, such as the Smectic B and C, are positioned closer to the aleatory limit ( and ). Thus, it seems that the permutation quantifiers are capturing in somehow the competition between the orientational order of the phase and, also, the number of defects present in the textures.\n\n![Complexity-entropy causality plane evaluated for several liquid crystal textures [@pone.0040689-Kent1].\\\nHere, we have used the averaged pixel values of the three layers (RGB) of the original image and and . The image sizes are about pixels. We note that each texture has a unique position in the causality plane which indicates that the permutation quantifiers are capable of differentiate not only transitions involving the isotropic phase, but also smoother phase transitions. We further observe that some high ordered phase such as the blue phase are located at the central part of the causality plane (region of higher complexity), while other phases which present a large number of defects such as the Smectic B and C are closer to the aleatory limit ( and ).](pone.0040689.g006){#pone-0040689-g006}\n\nIsing Surfaces {#s3c}\n--------------\n\nAs a last application, we study the permutation measures and applied to Ising surfaces [@pone.0040689-Brito1], [@pone.0040689-Brito2]. These surfaces are obtained by accumulating the lattice spin values of the Ising model defined by the Hamiltonian.where the sum is over all the pairs of first neighbor sites in the lattice. We numerically solve this spin- Ising model on a lattice using the Monte Carlo method with periodic boundary conditions. By using the spin values, we define the surface height for each lattice site aswhere represents the number of Monte Carlo steps. In [Fig. 7](#pone-0040689-g007){ref-type=\"fig\"}, we show three surfaces obtained though this procedure for different values of the reduced temperature , where is the critical temperature of the model. We note the complex pattern exhibited by the surface for , and the almost random patterns for and .\n\n![Examples of Ising surfaces for three different temperatures.\\\nThese surfaces were obtained after Monte Carlo steps for three different temperatures: below , at and above . In these plots, the height values were scaled to stay between and . We note that for temperatures higher or lower than , the surfaces exhibit an almost random pattern. For values of the temperature closer to the surfaces exhibit a more complex pattern, reflecting the long-range correlations that appear among the spin sites during the phase transition.](pone.0040689.g007){#pone-0040689-g007}\n\nWe first investigate the dependence of and on the reduced temperature after a large number of Monte Carlo steps () and for . [Figures 8a and 8b](#pone-0040689-g008){ref-type=\"fig\"} show and for and , and for the rotation and . We note that, at the critical temperature, both indexes display a sharp peak and that they are invariant under the rotation. Moreover, [Fig. 8c](#pone-0040689-g008){ref-type=\"fig\"} presents a 3d visualization of the phase transition for . This higher-dimensional representation can be useful when investigating more complex phase transitions, since a greater number of degrees of freedom allows the critical point to be more visible.\n\n![Dependence of the entropic indexes on the reduced temperature for Ising surfaces.\\\n(a) The permutation entropy and (b) the complexity measure versus the reduced temperature for and , and also for and . We note invariance of indexes under the rotation and . (c) A 3d visualization of the Ising model phase transition when considering . The gray shadows represent the dependences of on and of on .](pone.0040689.g008){#pone-0040689-g008}\n\nWe further study the temporal evolution of and for different reduced temperatures, as it is shown in [Fig. 9](#pone-0040689-g009){ref-type=\"fig\"}. The initial values of the spins were chosen equal to and, as we can see, the values for and are different just after one Monte Carlo step. For , the value of increases over time and around it reaches a plateau. For , the value of increases up to a maximum value around and then starts to approach a lower plateau value. A striking behavior is observed for , where for all temperatures the complexity displays a maximum value before it begins to approach a plateau value. It is worth noting that both quantifiers are very stable after Monte Carlo steps.\n\n![Dependence of the entropic indexes on the number of Monte Carlo steps.\\\nHere, denotes the number of Monte Carlo steps and the reduced temperatures are indicated in the plots. In (a) we show versus and in (b) versus for . We note the stability of both indexes after Monte Carlo steps.](pone.0040689.g009){#pone-0040689-g009}\n\nConclusions {#s3d}\n-----------\n\nWe have proposed a generalization of the complexity-entropy causality plane to higher-dimensional patterns. We applied this approach to fractal surfaces, liquid crystal textures and Ising surfaces. It was shown that the indexes and performed very well for distinguishing between the different roughness of the fractal surfaces. The indexes properly identified the phase transitions of a lyotropic liquid crystal and sorted different characteristic textures in a kind of complexity order. Finally, concerning the Ising surfaces, the indexes not only had identified the critical temperature, but also proved to be stable after Monte Carlo steps. The method also has a very fast and simple numerical evaluation. Taking into account all these findings, we are very optimist that our method can reduce the gap between one-dimensional complexity measures and the higher-dimensional ones.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: HVR. Performed the experiments: HVR PAS EKL. Analyzed the data: HVR PAS. Contributed reagents/materials/analysis tools: LZ RSM EKL. Wrote the paper: HVR LZ RSM.\n"}
+{"text": "INTRODUCTION\n============\n\nThe cough reflex causes modifications of normal breathing patterns \\[[@R1]\\]. This reflex is a complex and precisely timed neuromuscular phenomenon that involves the diaphragm, chest wall muscles, cervical muscles, abdominal muscles, laryngeal abductor and adductor muscles, and both medullary and cortical brain regions \\[[@R2]-[@R4]\\]. Throughout childhood, various developmental phenomena influence the cough reflex. Among these are modifications in the anatomy and functions of the respiratory airways and the central and peripheral nervous systems. Furthermore, because of the immature immunological response, infection is the main cause for cough in infancy. In general, children are more vulnerable to environmental factors. The impairment or absence of the cough reflex can be harmful and even fatal in some disease conditions. However, a cough may be the first symptom of many diseases or conditions affecting the respiratory tract, as a cough can represent more than a defense mechanism. Through its persistence, it becomes a helpful indicator of potential disease conditions for both patients and physicians. Almost all diseases of the respiratory tract, and in some cases of the extra-respiratory tract, can cause chronic cough. For the physician, it is most important to identify the serious disorders that require prompt management \\[[@R5]\\].\n\nThe aim of this article is to provide thorough research and analysis of the existing medical literature up to 2014 on chronic cough in children as a disease entity, including its epidemiologic, etiologic, diagnostic, prognostic, and therapeutic aspects. Given that chronic cough is not commonly considered a concrete pathological entity, but rather a persistent symptom of a chronic disease, reports in the literature addressing a single disorder causing chronic cough were not included in our present research.\n\nRESEARCH METHODOLOGY\n====================\n\nThe following sources were selected to identify the primary studies, secondary sources, and guideline searches:\n\n-   MedLine (through the PubMed search engine)\n\n-   EMBASE\n\n-   Cochrane Database of Systematic Reviews (CDSR; The Cochrane Library)\n\n-   Database of Abstract of Reviews of Effects (DARE)\n\n-   The principal Guideline Banks (LG) on the website of Gruppo Italiano di Medicina Basata sulle Evidenze (GIMBE) \\[[@R6]\\].\n\nFrom the keyword registers in Embase (Emtree) and PubMed (MeSH database), we employed the following terms for data collection: 'cough', 'coughing', and 'chronic', matched case-by-case in separate search strings to the keywords 'epidemiology', 'prevalence', 'incidence', 'etiology', 'diagnosis', 'prognosis', and 'therapy'. In PubMed, we used the pre-defined phrase 'chronic cough children' and the search tool for specific clinical areas known as clinical queries.\n\nWith these, it was possible to separately explore the fields of etiology, diagnosis, prognosis and therapy, as well as any 'clinical prediction guides' that were available. Our literature search was conducted without time limits and was concluded on September 8, 2014. The selection was further restricted to reports written in English or Italian and involving humans within the age range of 0 to 18 years. For articles regarding therapy, we limited our search to randomized controlled trials (RCTs). In our investigation, we excluded non-systematic reviews, editorials, letters, works of pure research, and grey literature.\n\nRESULTS AND DISCUSSION\n======================\n\nPrevalence Studies\n------------------\n\nThere are limited data reporting the prevalence of chronic cough in children. ln a study conducted in 2002 on a population of 2275 children aged 1-15 years living in rural areas of India, chronic recurrent cough was present in 1.06% of cases \\[[@R7]\\]. Whereas in a study conducted in China in 2010 on children living in six different urban areas, a 21-28% increase in chronic cough was observed for each interquartile increase in the concentration of atmospheric pollutants, such as total suspended particles, nitrogen dioxide, and sulphur dioxide \\[[@R8]\\].\n\nData from an Australian prevalence study \\[[@R9]\\] that was conducted on a cohort of children enlisted following urgent admittance to a tertiary center for acute respiratory disease reported that 20% of the children (aged 1-14 years old) had chronic cough (cough for more than 4 weeks). However, this percentage represents the prevalence in a selected population that was admitted to a tertiary center following acute respiratory illness, which is quite different from that in the general population.\n\nMore pertinent to Italian settings was the SIDRIA-2 observational study \\[[@R10]\\] that was conducted on 33,000 children and adolescents in various parts of the country. This study showed a significant increase in cough and chronic catarrh with an increase in prevalence from 2.2% in traffic-free areas to 3.2% in areas with intense traffic. Moreover, this study demonstrated differences between areas with rare truck passage (prevalence 2.0%) and areas with frequent (2.9%) or continuous (3.9%) truck passage. The increase in the prevalence of chronic cough associated with low automobile traffic was not significant.\n\nDiagnosis\n---------\n\nBibliographic research in terms of EBM revealed only one strictly diagnostic publication, which was an Australian study conducted in 2006 \\[[@R11]\\]. The prospective cohort study included 100 children with an average age of 2.8 years who coughed significantly and persistently for more than 3 weeks. Although the trial's overall methodological quality was good, it is worth keeping in mind that these children had a median cough duration of 6 months, suggesting a possible high level of disease severity.\n\nThe authors identified wet cough as a good indicator of a specific cough (a cough with a specific cause), which was the primary aim of the study. In reality, however, with a positive likelihood ratio (LR+) barely above 1 (1.29), this parameter seems to provide an irrelevant diagnostic performance in both populations. As expected, better positive predictability was provided by hemoptysis and chronic dyspnea. In addition, the study revealed diagnostic tools with a positive predictability for a specific etiology, which included the following: objective chest examination with an LR+ of 2.4, radiography anomalies with an LR+ of 2.92, and spirometry with an LR+ of 2.33. This latter examination was performed in only 32 children who were older than 6 years of age. Wet cough was shown to be the lone parameter that was effective mainly in excluding a specific cause in a low-prevalence population (LR- of 0.15). However, factors including history (mainly pointers), clinical evaluation, and diagnostic tools each had an LR- \\>0.3, thus, demonstrating the scarcity of the efficacy in an exclusion diagnosis.\n\nBased on these results, the authors subsequently recommended in their guidelines chest radiography and spirometry as diagnostic tools that should be used in children who have experienced cough for longer than 4 weeks.\n\nCough Questionnaire\n-------------------\n\nA recent systematic review \\[[@R12]\\] determined that only the pediatric cough quality of life (PC-QoL) questionnaire can be considered a valid means for assessing the severity and impact of a cough on a child's health and that evidence on the validity of cough diaries or visual analogue scale scores is currently insufficient. Four relevant publications were obtained using PubMed clinical queries. Two of these publications focused on the validation of the PC-QoL \\[[@R13], [@R14]\\], one was dedicated to the validation of a questionnaire on the quality of the cough \\[[@R15]\\], and the remaining publication regarded the validation of a constructed questionnaire on cough hypersensitivity syndrome \\[[@R16]\\] that was not pertinent to the scope of this review.\n\nEtiology and Algorithms for Management of Chronic Cough\n-------------------------------------------------------\n\nThe earliest publication we found in this field was a systematic review (SR) on the causal relationship between parental smoking and various respiratory symptoms in children, including cough \\[[@R17]\\]. The review, which included 34 studies, demonstrated that the odds ratio (OR) for exposure to parental smoking and cough was 1.40 (95% confidence interval \\[CI\\] 1.27-1.53). The authors concluded that the relationship between parental smoking and respiratory symptoms \"seems very likely\" to be causal given the statistical significance, robustness to adjustment for confounding factors, consistency of the findings in different countries, and evidence of dose response.\n\nIn the above-cited observational study that was conducted in a rural Indian population of 2275 children, Singh *et al.* reported bronchial asthma as the most common cause of chronic cough in 66.7% of the studied group, followed by post-nasal drip in 25%. A family history of smoking was recognized in 16.7% of cases, which is in contrast to 6.4% in the control group (p=0.05). There was no significant association with overcrowding, presence of pets, or the category of fuel employed for home cooking. However, this study cannot be compared to case studies in Western nations because of the large differences in health services, hygiene, and socioeconomic conditions.\n\nThe second study by the Italian SIDRIA-2 Group was conducted in 2005 and focused on indoor exposure and respiratory and allergic disorders \\[[@R18]\\]. The study revealed that pollution was a significant causal link between chronic cough, exposure to domestic molds, and the presence of dogs during the first year of life (OR 1.89; 95%CI 1.31-2.71 and OR 1.87; 95%CI 1.12-3.11, respectively). Indoor exposure to dogs in the first year of life apparently mantains its effects even throughout adolescence (OR 2.03; 95%CI 1.12-3.67). The last four publications found during our research were conducted in different international settings and with more rigorous bases. The earliest of these was a study performed by an Australian group over a 2-year period. In that study, 108 children with an average age of 2.8 years and chronic cough were subjected to a standardized sequence of a diagnostic pathway approach in which chest radiography and spirometry were primarily performed when possible. Bronchoscopic investigation, computed tomography scanning, and oesophageal pH-metric monitoring were included in the second phase of the pathway \\[[@R19]\\].\n\nThe second report is a Turkish study published in 2008 that was conducted on 108 children with an average age of 8.5 years that applied the recommendations encompassed in the guidelines of the American College of Chest Physicians (ACCP) \\[[@R20], [@R21]\\]. In that study, children with specific cough alarm indicators (pointers) were excluded from the algorithm at the beginning. In a third prospective study carried out in the USA that lasted 4 years, a fairly small number of children (n=40) with an average age of 7.8 years and persistent cough for more than 8 weeks were subjected to a rather intensive and detailed work-up that is typical of a third-level center \\[[@R22]\\]. The fourth publication involved a larger cohort than those in the previous three studies. It included 346 children from six Australian centers (five centers were located in cities, and one center was located in a rural area) \\[[@R23]\\]. Moreover, children were separated into four age groups (0-2, 2-6, 6-12, and \\>12 years) and were managed in accordance with a pathway similar to the one utilized by the Turkish authors that followed the ACCP guidelines in which chest radiography and spirometry were conducted during an early phase. In the absence of specific pointers, the children underwent a diagnostic-therapeutic trial with inhaled corticosteroids when the chronic cough was dry/unproductive or with antibiotics when the cough was wet/productive.\n\nAs can be observed in a summary of these studies in Table **[1](#T1){ref-type=\"table\"}**, large differences are evident, including differences in the definition of the chronic cough duration (4 versus 8 weeks), age groups, number of subjects, and duration of the cough prior to enrolment (16 weeks in the studies by Asilsoy and Chang, 18 weeks in the study by Khoshoo, and 6 months in the study by Marchant). Moreover, the study by Khoshoo differs from the other 3 studies by the lack of an a priori definition of the ascribed etiology and how much change was considered to be important. Differences were observed in the-characteristics of the adopted diagnostic procedures, the settings in which the studies were conducted, the initial inclusion or exclusion criteria (children presenting alarm signals indicating a cough with specific etiology were excluded in Asilsoy's study) and in the methodological quality of statistical analyses.\n\nThe four above-mentioned studies present noteworthy differences in the frequencies of various etiologic diagnoses, which were highest for asthma, gastroesophageal reflux (GER), and upper airway disorders (post-nasal drip) in the older populations (Khoshoo and Asilsoy studies). In the younger populations studied, the diagnosis with the highest frequency was protracted bacterial bronchitis. The significance of this latter correlation was demonstrated in the Australian multi-center study in which protracted bacterial bronchitis was observed in approximately 54% of children less than 2 years old (n=124), in 40% of children aged between 2 and 4 years (n=126), in 27% of children aged between 6 and 12 years (n=82), and in 29% of those older than 12 years (n=14). No differences or trends were evident in the frequencies of the other four most often-reported diagnoses. However, a significantly higher frequency of tracheomalacia was observed in the aborigine population than in the white population (29.4% vs. 6.7%; p=0.001).\n\nIntervention Studies\n--------------------\n\nTo assess the validity of management according to the standardized clinical management, we examined a single recent Australian trial study \\[[@R24], [@R25]\\]. That study is the only study dedicated to the use of algorithms subsequent to the 2009 Cochrane SR that failed to identify any pertinent RCT \\[[@R26]\\]. The multi-center study used a pragmatic design to test whether management of chronic cough in children in accordance with an evidence-based pathway was feasible, reliable, and improved clinical outcomes. The principal objective was to test the efficacy of the pathway. Because an RCT comparing use versus non-use of the pathway was not possible, the authors designed their study to evaluate delayed versus early employment of the pathway. Approximately 75% of the group of children enrolled in an Australian trial \\[[@R23]\\] had previously been members of a population examined in a randomized study \\[[@R24], [@R25]\\]. The intervention consisted of the use of a diagnostic-therapeutic algorithm at different timings: 2-3 weeks after the request for admission (early-arm) or 5-7 weeks after the request for admission (delayed-arm). One hundred and forty children with an average age of 4.6 years were randomly selected to undergo intervention within 2-3 weeks, whereas 132 received the intervention within 5-7 weeks, which represented the standard caregiving time at the various centers that were involved in the study. The aim of the study was to demonstrate, with an 82% power and a 5% significance level, the statistical significance of a difference of at least 20% in the \"resolution of the cough\" within 6 weeks of acceptance for admission (this parameter constituted one of the two primary outcomes of the randomized controlled trial). The other primary outcome was the difference in the parent-proxy cough-specific quality of life (PC-QoL) completed within 6 weeks of acceptance for admission. At the end of the study, the intention-to-treat analysis demonstrated a statistically significant difference for the resolution of the cough within 6 weeks (54.3% vs. 29.5%; p=0.01; number needed to treat \\[NNT\\]=4). The authors also affirmed that a significant improvement in PC-QoL was observed in both groups and was greater in the early treated group (the p-value was not specified). That study, in which blindness was obviously not applicable, had good methodological quality in terms of other risks of potential bias. The period of the cough duration in randomized patients was fairly long (4 months), which must be noted particularly in comparison with the Italian children, for whom access to pediatric care is easier, thus, allowing earlier initiation of treatment.\n\nChronic Cough Treatment\n-----------------------\n\nAn examination of the secondary sources not included in the Cochrane Collaboration revealed a review lacking the minimum systematic criteria that we decided not to include \\[[@R27]\\] and an SR that was conducted in 2013 \\[[@R28]\\] in which childhood studies constituted only a scarcely relevant part of the overall study. The CDSR instead presents an \"umbrella review\" composed of up to 15 SRs that are continuously updated on the website, which provides an important, complete, and methodologically irreprehensible body of evidence. However, it currently lacks wide clinical significance because of the general scarcity of well-structured studies on pharmacological and non-pharmacological therapies for chronic cough in children. One of these 15 SRs examined the management algorithms that were mentioned earlier in this review \\[[@R26]\\]. Five SRs \\[[@R29]-[@R33]\\] did not include any pertinent work, nor did our research reveal more recent pertinent trials. These include trials with chromones, anticholinergic agents, obstructive sleep apnea syndrome therapies, modification of indoor environments, and honey or cough drop usage. Regarding \u03b2-2 agonists, no efficacy was demonstrated by the only RCT study that was included in a dedicated SR \\[[@R34]\\]. The same applies to the two trials included in an SR examining the use of leukotriene receptor antagonists \\[[@R35]\\]. In an SR examining the use of methylxanthine, some cohort studies appeared to have demonstrated an efficacy of dubious clinical relevance on the symptoms of chronic cough despite the well-known undesired side-effects of the drug \\[[@R36]\\]. An SR on the use of antibiotics in a chronic cough post-bronchiolitis indicated only one RCT in which the efficacy of clarithromycin was not significant \\[[@R37]\\].\n\nAn SR on antihistamines included three therapeutic studies with conflicting results. The data from the two larger therapeutic studies that described no beneficial effect are in contrast to a smaller study that demonstrated that the second-generation antihistamine cetirizine was beneficial in reducing coughs associated with seasonal allergic rhinitis \\[[@R38]\\].\n\nIn regard to these latter Cochrane SR references \\[[@R29]-[@R33]\\], our research failed to discover any subsequent noteworthy trials. There are two SRs regarding the use of inhaled corticosteroids \\[[@R39], [@R40]\\]. Among the various selected works, only one trial (which contained serious methodological defects) demonstrated a relatively modest efficacy of fluticasone dipropionate for nocturnal cough \\[[@R41]\\]. The drug was administered at a high dosage of 2 mg/day for 3 days, then 500 \u03bcg/day for the following 11 days. None of the dietetic or pharmacological treatments studied in the six RCTs of an SR examining the treatment of GER were shown to be effective \\[[@R42]\\]. However, it is also worth noting that in the work conducted with proton pump inhibitors, only 1 patient of the 11 that were treated experienced an undesired side effect (number needed to harm = 11) \\[[@R43]\\]. Only one work was published after the SR on GER by Chang *et al.*; this work was a Chinese RCT conducted on 19 children assigned to four groups in which the omeprazole-bethanecol combination appeared to reduce daytime coughing in children with GER \\[[@R44]\\].\n\nThe SR on the use of antibiotic therapy \\[[@R45]\\] instead selected only two randomized trials of modest quality: one conducted with erythromycin and the other with an under-dose of amoxicillin clavulanate. In the authors' opinion, on the basis of only these two studies, it would be necessary to treat three children (NNT=3) with an antibiotic in order to obtain a significant improvement in persistent cough. These conclusions were conditioned by the methodological weakness of the study. One particularly relevant study was published after the final update of the latter SR, which was an RCT conducted in Australia in which 50 children with an average age of 1.9 years and with wet cough lasting for more than 3 weeks were randomized to 2 weeks of twice daily oral amoxicillin clavulanate (22.5 mg/kg/dose) or placebo \\[[@R46]\\]. In this selected population, which was treated with low antibiotic doses (higher doses are commonly used in Italy), the NNT was only three children (95%CI 1.8-13.9).\n\nGuidelines\n----------\n\nWe consider that there are four main international guidelines (American, British, Belgian, and Australian) on chronic cough in children that are most relevant \\[[@R21], [@R47]-[@R49]\\], all of which are important documents because they were produced by research groups of cough experts. Most of these groups have studied the phenomenon in great detail over the past two decades. Although numerous articles published in recent years have modified our knowledge of the etiology and management (algorithms and therapy) of chronic cough, until now (September 2014), none of the guidelines have been updated or revised. Moreover, the qualitative analysis of the three fundamental characteristics according to which guidelines should be produced (multidisciplinarity, systematic literature searches, and recommendation grading) leads to appraisals that often cannot be judged positively.\n\nThe ACCP guidelines were published in 2006; this version was not multidisciplinary (only physicians were involved), and the bibliographic assessment of the methodological criteria is not easy to consult \\[[@R21]\\].\n\nThe Belgian guidelines are not multidisciplinary, and they employed a non-conventional method of recommendations \\[[@R47]\\]. The British guidelines lack multidisciplinarity, and bibliographic research was performed by the first author alone and was restricted to articles in English \\[[@R48]\\].\n\nThe Australian guidelines are the most recently updated guidelines \\[[@R49]\\]. The authors state their recommendations according to the GRADE approach. Nonetheless, the authors referred to an inaccessible online document for details regarding the methodology. In addition to this, the authors declare that the research source was limited to PubMed. Only the British Thoracic Society guidelines offer a clear correlation between single recommendations and the sources of evidence supporting them. Although all these guidelines contain practical flow-charts for a diagnostic and/or therapeutic approach for children with chronic cough, they all had a restricted panel of reviewers. Moreover, the limited number of studies and investigations in this field has the potential to introduce bias. An updated cough guideline and an expert panel is expected to be published in the coming months by an international group \\[[@R50]\\].\n\nCONCLUSION\n==========\n\nChronic cough is common in the pediatric population, yet the true prevalence of this condition remains difficult to define. Before receiving appropriate management, many children with chronic cough and their families continue to experience unnecessary or recurring medical consultation with a great impact on their quality of life. Physiologically, there are similarities and also significant differences between adults and children \\[[@R51], [@R52]\\]. As expected, the etiologies and management of cough in children differ from those used for adults. A chronic cough in a child should be treated on the basis of etiology; no evidence supports the use of medication for symptomatic cough relief. The benefits to the child of making an accurate diagnosis and prescribing adequate treatment, therefore, extends well beyond merely eliminating the cough.\n\nRegardless of the setting, whether pediatric or non-pediatric, children with chronic cough should be carefully evaluated using child-specific protocols and algorithms. Those based on good evidence are more likely to improve the clinical outcomes. Knowledge of the pathophysiology of the different conditions that cause chronic cough is essential for disease management. Distinguishing the different phenotypes is helpful in achieving the correct diagnosis. There is a continuing search for the distinctive phenotypes of chronic cough patients who will prove responsive to one treatment rather than another \\[[@R46]\\]. The prevalence of each etiology in a medical setting depends on various issues. Among these are the population under consideration and the variety in the ages of its constituents (infants, preschool, school-children), infectious disease control and prevention, diagnostic procedures employed, disease definition criteria, and the health system. Clinical algorithms for the management of children with chronic cough should consider these factors. Despite the apparently high prevalence of cough in children, the topic has been poorly researched. Further clinical and basic research studies are still needed for better diagnosis, treatment, and prevention of chronic cough in children.\n\nDeclared none.\n\nCONFLICT OF INTEREST\n====================\n\nThe authors confirm that this article content has no conflict of interest.\n\n###### \n\nPrincipal causes of chronic cough in children.\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                                     PBB   Asthma   UACS   GER    Bronchiectasis   Tracheo-malacia   Habitual-psychogenic cough   Spontaneous resolution   Other\n  -------------------------------------------------- ----- -------- ------ ------ ---------------- ----------------- ---------------------------- ------------------------ -------\n  *Marchant 2006* \\[[@R19]\\] (2.6 yrs average age)   40%   4%       3%     3%     6%               \\-                1%                           22%                      21%\n\n  *Asilsoy*\\                                         23%   25%      20%    5%     3%               \\-                4%                           6%                       3%\n  *2008* \\[[@R20]\\] (8.4 yrs average age)                                                                                                                                  \n\n  *Khoshoo 2009* \\[[@R22]\\]\\                         \\-    13%      23%    28%    \\-               \\-                10%                          \\-                       25%\n  (7.8 yrs average age)                                                                                                                                                    \n\n  *Chang*\\                                           41%   15.8%    1.4%   2.3%   9%               6.1%              4.3%                         13.9%                    6.1%\n  *2012* \\[[@R23]\\] (4.5 yrs average age)                                                                                                                                  \n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAbbreviations: PBB: protracted bacterial bronchitis; GER: gastroesophageal reflux; UACS: upper airway cough syndrome\n"}
+{"text": "Introduction\n============\n\nThe morbidity associated with gastric carcinoma (GC) has declined in recent decades; however, GC remains the fourth most common carcinoma and second highest cause of cancer-associated mortality worldwide ([@b1-ol-0-0-6974]). In 2012, there werean estimated 951,600 new cases and 723,100 GC-associated mortalities ([@b1-ol-0-0-6974]). Despite progression in diagnosis and treatment methods, the prognosis for patients with GC remains poor. Due toinconspicuous symptoms in the early stages, the vast majority of patients with GC are already in the advanced stages at the time of first diagnosis, resulting in a poor prognosis ([@b2-ol-0-0-6974],[@b3-ol-0-0-6974]). Therefore, the early diagnosis and treatment of GC are of critical importance for improving the clinical outcome.\n\nMena (also referred to as ENAH-enabled homolog) is a member of the Ena/vasodilator-stimulated phosphoprotein (VASP) family of actin-binding proteins, which function in diverse types of cell ([@b4-ol-0-0-6974],[@b5-ol-0-0-6974]). Ena/VASP proteins are key regulatory molecules that control the cell shape, movement and actin organization on cadherin adhesion contacts, which are frequently affected following malignant transformation ([@b4-ol-0-0-6974],[@b6-ol-0-0-6974]). Mena is a key mediator of cytoskeletal arrangement ([@b7-ol-0-0-6974]). It regulates cell movement by protecting actin filaments from capping proteins during polymerization ([@b8-ol-0-0-6974]). An upregulated Mena expression level was previously reported in mouse and rat invasive breast carcinoma ([@b9-ol-0-0-6974]), as well as in human breast cancer cell lines and tissues ([@b10-ol-0-0-6974]). Similarly, Mena expression was observed to beupregulated in human hepatocellular carcinoma ([@b11-ol-0-0-6974]), colorectal carcinoma ([@b12-ol-0-0-6974]), cervix precursor lesions ([@b13-ol-0-0-6974]) and pancreatic tumor cell lines as well as in primary and metastatic pancreatic tumors ([@b14-ol-0-0-6974]); in normal tissue, Mena expression level was reported at low or non-detectable levels ([@b11-ol-0-0-6974]). However, the clinical significance of Mena in GC remains indistinct. The present study investigated the expression level of Mena in GC to reveal its clinicopathological significance.\n\nMaterials and methods\n=====================\n\n### Patients and tissue samples\n\nThe present study was performed with 106 GC paraffin-embedded tissue samples collected during resection from the Third Affiliated Hospital of Sun Yat-sen University (Guangzhou, China) between January 2001 and December 2004 for immunohistochemical (IHC) analysis. The median age of the patients was 54 years old (range, 29--72 years) and the median tumor size was 6.0 cm (range, 0.8--15.0 cm); the group included 67 male and 39 female patients. From these 106 patients, 32 samples of adjacent non-cancerous tissues were additionally collected as control samples. All patients were pathologically diagnosed with gastric adenocarcinoma. None of the patients had received any type of neoadjuvant therapy and all underwent a radical excision. The clinical information for these samples is summarized in [Table I](#tI-ol-0-0-6974){ref-type=\"table\"}. The date of patient surgery was defined as the initial event of survival analysis, and the date of patient mortality or the censoring of the patient at the last follow-up date was defined as the end time. The interval was defined as the overall survival time for patients.\n\nIn addition, 10 paired GC and adjacent normal tissues (the adjacent normal tissue was defined as at least 5cm from the tumor edge) were collected from the Third Affiliated Hospital of Sun Yat-sen University between June 2013 and February 2015 for reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis. The group included 7 male and 3 female patients, and the median age of the patients was 51 years old (range, 32--69 years). Tissues were collected immediately after surgery.\n\nThe clinicopathological classification and staging were determined according to the American Joint Committee on Cancer criteria ([@b15-ol-0-0-6974]). Written informed consent was obtained from all patients prior to enrollment in the present study. The present study was approved by the Institutional Research Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University.\n\n### RT-qPCR analysis\n\nTotal RNA samples were extracted from 10 paired GC and adjacent normal tissues using TRIzol^\u00ae^ reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer\\'s protocol. Extracted RNA was pretreated with RNase-free DNase. For cDNA synthesis, 2 \u00b5g RNA from each sample was used, according to the RevertAid\u2122 First Strand cDNA Synthesis kit instructions (K1622; Thermo Fisher Scientific, Inc.).\n\nFor the PCR amplification of Mena cDNA, SYBR-Green 2X master mixture (170-8882AP; Bio-Rad Laboratories, Inc., Hercules, CA, USA) was used in a total volume of 20 \u00b5l, according to the manufacturer\\'s instructions, an initial amplification step using Mena-specific primers was performed with a denaturation at 95\u00b0C for 10 min, which was followed by 28 denaturation cycles at 95\u00b0C for 60 sec, primer annealing at 58\u00b0C for 30 sec and primer extension at 72\u00b0C for 30 sec. Upon completion of the cycling steps, a final extension at 72\u00b0C for 5 min was performed prior to the storage of the reaction mixture at 4\u00b0C. The primer sequences were as follows: Mena sense, 5\u2032-GTGCCATTCCTAAAGGGTTGA-3\u2032 and antisense, 5\u2032-GCTGCCAAAGTTGAGACCATAC-3\u2032; GAPDH sense, 5\u2032-TGTTGCCATCAATGACCCC-3\u2032 and antisense, 5\u2032-CTCCACGACGTACTCAGC-3\u2032. The primers were designed with Primer Express version 2.0 software (Applied Biosystems; Thermo Fisher Scientific, Inc.).\n\nGAPDH was used as an internal control, the relative expression level of Mena was calculated using the 2^\u2212\u0394\u0394CT^ method ([@b16-ol-0-0-6974]); all experiments were performed in triplicate.\n\n### IHC analysis\n\nIHC staining was performed to investigate the alteration to protein expression levels in 106 human GC tissues and 32 paired adjacent non-cancerous tissues. Briefly, 4-\u00b5m-thick paraffin sections of the tissue were deparaffinized with xylene and rehydrated in a descending alcohol series. Antigenic retrieval was performed by submerging the slides in EDTA antigenic retrieval buffer and microwave heating for 3 min at 650 W and thentwice more at 350 W for 3 min. To quench endogenous peroxidase activity, the slides were treated with 3% hydrogen peroxide in methanol and then incubated with 1% bovine serum albumin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) at room temperature for 60 min to block nonspecific binding. Subsequently, tissue sections were incubated with a rabbit polyclonal anti-Mena antibody (dilution, 1:100; BD Biosciences, Franklin Lakes, NJ, USA, catalog number: 5111-1) at 4\u00b0C overnight. Normal goat serum (Santa Cruz Biotechnology, Inc.) was used at 4\u00b0C overnight as a negative control. The tissue sections were incubated with a biotinylated anti-rabbit secondary antibody (no dilution; Santa Cruz Biotechnology, Inc.; catalog number: sc-2040) at room temperature for 30 min following 3 washes in PBS, followed by further incubation with a streptavidin-horseradish peroxidase complex (dilution, 1:1500; Abcam, Cambridge, UK; catalog number: ab7403) at room temperature for 30 min. Slides were immersed in 3-amino-9-ethyl carbazole room temperature for 3 min and then counterstained with 10% Mayer\\'s hematoxylin at room temperature for 30 sec. Finally, they were dehydrated and mounted with Crystal Mount.\n\nSlides were imaged at magnification \u00d720 (0.5\u00d70.5 \u00b5m^2^ pixel resolution) using a WSI instrument (ScanScope CS, Aperio, Vista, CA, USA) fitted with a 20\u00d7/0.75 Plan Apo objective lens (Olympus, Center Valley, PA, USA). For the evaluation of immunostaining, the degree of immunostaining was viewed and scored independently by two pathologists, who were blinded to the histopathological characteristics and patient information for the samples. The mean value of the scores provided by the two independent pathologists was used for the comparative evaluation of Mena expression.\n\nThe intensity of Mena staining was graded according to the following criteria: 0, no staining; 1, weak staining (light yellow); 2, moderate staining (yellow brown); and 3, strong staining (brown). The percentage of stained tumor cells was scored as follows: 0, no positive tumor cells; 1, 1--25% positive tumor cells; 2, 26--50% positive tumor cells; 3, 51--75% positive tumor cells; and 4, \\>75% positive tumor cells.\n\nThe staining score was evaluated as the product of the proportion of positive tumor cells and the staining intensity score. The expression level of Mena was defined as follows: '\u2212' (score 0, negative), '+' (score 1--4, weakly positive), '++' (score 5--8, positive) and '+++' (score 9--12, strongly positive). Optimal cut-off values for Mena expression were selected based on the analysis of overall survival (OS) data with the log-rank test. A staining index score of \u22654 was used to define tumors with high Mena expression level whereas \\<4 indicated a low Mena expression level.\n\n### Statistical analysis\n\nAll statistical analyses were performed using SPSS 20.0 software (IBM Corp., Armonk, NY, USA). The difference in Mena expression levels between GC tissues and adjacent non-cancer tissues were analyzed using the \u03c7^2^ test. Survival curves were plotted using the Kaplan-Meier method and compared using the log-rank test. The association between Mena expression level and other clinicopathological characteristics was analyzed usingthe \u03c7^2^ and Fisher\\'s exact tests. Bivariate correlations between the clinicopathological characteristics were determined using Spearman\\'s rank correlation coefficients. Clinicopathological characteristics used to predict the prognosis in clinical practice were evaluated by univariate and multivariate Cox regression analyses. The selected type of Cox model for the univariate analysis was the 'enter' method, and for the multivariate analysis, the 'forward' method. P\\<0.05 was considered to indicate a statistically significant difference.\n\nResults\n=======\n\n### Mena is overexpressed in GC tissues\n\nTo determine whether Mena expression isupregulated in human GC, RT-qPCR was performed on 10 paired GC and adjacent normal tissues. As presented in [Fig. 1](#f1-ol-0-0-6974){ref-type=\"fig\"}, the expression level of Mena mRNA was higher in all 10 GC tissue samples compared with in adjacent normal tissues, with the difference in expression level ranging from 1.5 to 84-fold. In IHC results, the high expression of Mena was observed in 52.83% (56/106) of the patients with GC, whereas weak or no staining was observed in 47.17% of the tumor samples ([Table I](#tI-ol-0-0-6974){ref-type=\"table\"}). In the adjacent non-tumor tissues, Mena protein staining was largely weak or absent; there was a 6.25% (2/32) positive expressionrate detected. The difference between these two groups was statistically significant (\u03c7^2^=18.910; P\\<0.001). As presented in [Fig. 2](#f2-ol-0-0-6974){ref-type=\"fig\"}, Mena staining occurred predominantly in the cytoplasm.\n\n### Mena overexpression is associated with GC clinical characteristics\n\nTo better understand the potential role of Mena in the development and progression of GC, the association of Mena expression level with other clinicopathological indexes in 106 paraffin-embedded archived GC tissues, including 10 stage I tumors, 10 stage II tumors, 84 stage III tumors and 2 stage IVa tumors, was investigated.\n\nAs summarized in [Table I](#tI-ol-0-0-6974){ref-type=\"table\"}, there were no significant associations between Mena expression level and the gender, age, node (N) or metastasis (M) stage, tumor size, grade and the infiltration of adjacent organs in the patients; however, the expression level of Mena was significantly associated with the tumor (T; P=0.033) and TNM stages (P\\<0.001).\n\n### Association between Mena expression level and overall patient survival time\n\nSurvival analysis revealed a clear negative association between the expression level of Mena protein and the OS time of patients with GC (P\\<0.001; [Fig. 3A](#f3-ol-0-0-6974){ref-type=\"fig\"}). In addition, Cox regression analysis revealed that Mena expression level, T stage and N stage were independent prognostic factors for OS time ([Table II](#tII-ol-0-0-6974){ref-type=\"table\"}).\n\nThe prognostic significance of Mena expression status in selective subgroups stratified by the T stage and TNM stage was analyzed. For patients with late-stage tumors (stage III--IVa), the expression level of Mena was strongly associated with the OS duration ([Fig. 3B](#f3-ol-0-0-6974){ref-type=\"fig\"}; P=0.001), which was not the case for patients with early-stage tumors (stages I--II; [Fig. 3C](#f3-ol-0-0-6974){ref-type=\"fig\"}; P=0.181). Similarly, when it was evaluated according to T stage, the effect on outcome associated with the expression level of Mena was significant only in the T3-4 subgroup ([Fig. 3D](#f3-ol-0-0-6974){ref-type=\"fig\"}; P=0.004), and not in the T1-2 subgroup ([Fig. 3E](#f3-ol-0-0-6974){ref-type=\"fig\"}, P=0.200).\n\nDiscussion\n==========\n\nGC is the fourth most common type of cancer and the second leading cause for cancer-associated mortality worldwide, although it exhibits a decreasing trend of incidence ([@b1-ol-0-0-6974],[@b17-ol-0-0-6974]). There has been significant clinical progress in the early diagnosis and treatment of GC during recent decades; however, it is usually diagnosed at a late stage, resulting in a high treatment cost and decreasing the rate of successful curative surgery ([@b18-ol-0-0-6974]). The 5-year OS rate for GC is closely associated with the tumor stage. Patients diagnosed at stage I exhibit a 5-year OS rate of \\>90%, whereas patients diagnosed at stage IV exhibit a 5-year OS rate of \\<5% ([@b19-ol-0-0-6974]). Therefore, there is currently a great clinical demand for early diagnosis and treatment, which are pivotal for improving the outcome of GC.\n\nClassical serum tumor markers, including carcinoembryonic antigen and carbohydrate antigen 19-9 have definite implications for GC diagnosis and monitoring, but the lack of specificity and sensitivity impaired their function ([@b20-ol-0-0-6974]). In recent years, there have been multiple novel tissue-based biomarkers for GC identified, including vascular endothelial growth factor receptor 2 ([@b21-ol-0-0-6974]), excision repair cross-complementation group 1 ([@b22-ol-0-0-6974]), human epidermal growth factor receptor-2 ([@b23-ol-0-0-6974],[@b24-ol-0-0-6974]), Bcl-2 and Ki-67 ([@b25-ol-0-0-6974]). However, most of these molecular markers are not conventionally used in the clinical setting as they do not accurately and efficiently predict the clinical outcome or curative effect. Novel tumor molecular markers are thus required to improve the detection, diagnosis and prognosis of GC.\n\nHuman ortholog of murine Mena, a member of the Ena/VASP protein family that includes Mena, VASP and Evl in mammals, is a key actin polymerization regulatory protein involved in the assembly and dynamics of cytoplasmic actin networks ([@b26-ol-0-0-6974]). The Ena/VASP family is an important regulator of actin cytoskeleton dynamics involved in cell motility. Additionally, alterations to the cellular actin network serve an important role in malignant transformation and tumor progression. Members of the Ena/VASP family that are localized at the tips of protruding filopodia and lamellipodia and adhesion foci function in the control of cell movement, shape and adhesion, which are important biological processesin the development of metastatic potential ([@b27-ol-0-0-6974]). Located on chromosome 1, the Mena gene encodes the 570-residue Mena protein and alternative splicing-derived isoforms ([@b28-ol-0-0-6974]). As a member of the Ena/VASP family, Mena regulates membrane protrusion and cell movement in various types of cells and contexts by influencing the geometry and assembly of actin filament networks through the binding of G-actin and F-actin ([@b26-ol-0-0-6974],[@b29-ol-0-0-6974]--[@b32-ol-0-0-6974]).\n\nMena enhances tumor cell migration toward epidermal growth factor (EGF) in part by interfering with the activity of the inhibitory capping proteins and increasing actin filament elongation, promoting actin polymerization ([@b26-ol-0-0-6974],[@b33-ol-0-0-6974],[@b34-ol-0-0-6974]). The anti-capping activity of Mena is proposed to amplify the barbed end output of the cofilin and Arp2/3 complex pathways, particularly in response to EGF, which is important in the metastatic potential of mammary tumors ([@b26-ol-0-0-6974],[@b29-ol-0-0-6974],[@b35-ol-0-0-6974]). Di Modugno *et al* ([@b28-ol-0-0-6974]) revealed that Mena is overexpressed in 75% of primary mammary carcinomas; consistent with this observation, high expression levels of Mena in breast cancer patients have been associated with poor prognosis ([@b28-ol-0-0-6974],[@b36-ol-0-0-6974]). Similarly, in precancerous lesions of the cervix and colon, the expression of Mena was upregulated with progressive transformation ([@b13-ol-0-0-6974],[@b37-ol-0-0-6974]). It was also detected in pancreatic carcinoma cell lines and in primary and metastatic pancreatic tumor tissues ([@b28-ol-0-0-6974],[@b36-ol-0-0-6974],[@b38-ol-0-0-6974]). Mena maintains the stability of invadopodia, actin-rich protrusions that contain proteases, increasing the matrix degradation activity of tumor cells. Mena activity potentiates EGF-induced tumor cell invasion and membrane protrusion. These previous studies demonstrate that the overexpression of Mena in cancer enables the invasion and metastasis of tumor cells in response to otherwise benign EGF stimulus levels, increasingthe responsiveness to macrophage signaling ([@b26-ol-0-0-6974]).\n\nThe present study presented novel evidence that the upregulation of Mena was associated with poor clinical outcomes in patients with GC, particularly for those with late-stage disease. It was clearly demonstrated that in GC tissues, the expression of Mena at the mRNA and protein levels was markedly higher compared with in the adjacent normal tissues. Therefore, Mena may be a biomarker for GC, which may aid precise diagnoses. However, at present, the precise functions of Mena in human cancer remain unclear. The overexpression of Mena in GC may reflect the aberrant regulation of actin dynamics. However, understanding of the precise mechanism underlying Mena in GC requires further investigation.\n\nThe present study additional lyinvestigated the association between Mena expression level and other clinical features of patients with GC. There was a significant association between Mena expression level and the T and TNM stages, which revealed that Mena may be used as an independent biomarker for the recognition of a subpopulation of GC patients with more aggressive disease. However, the associations between Mena and the gender, age, N stage, M stage, tumor size, grade, and infiltration in patients with GC were not significant.\n\nPrevious studies have reported the prognostic value of Mena in human cancer. For example, numerous studies have observed that high expression levels of Mena areassociated with a poor prognosis in patients with breast cancer ([@b28-ol-0-0-6974],[@b36-ol-0-0-6974]). However, to the best of our knowledge, the prognostic value of Mena in GC has not previously been explored. In the present study, patients with high Mena expression levels had a 1.79% cumulative 10-year survival rate, which was significantly lower than patients in the low Mena expression group (24.0%). Multivariate analysis revealed that the expression level of Mena may be an independent prognostic factor for OS time in GC patients ([Table II](#tII-ol-0-0-6974){ref-type=\"table\"}). Of note, a sub-group analysis demonstrated that patients with high Mena expressionand poor clinical out comes also demonstrated the features of late TNM and T stages.\n\nIn conclusion, to the best of our knowledge, this is the first study to investigate Mena expression level and its clinicopathological and prognostic significance in GC. The results of the present study suggested that Mena was upregulated in GC tissues and associated with the T and TNM stages. Multivariate analysis revealed that Mena may be an independent molecular marker for the prediction of GC prognosis and survival. Therefore, detecting the Mena protein expression level may aid the stratification of patients as a novel therapeutic strategy and establish a rational treatment selection criterion for patients with GC. Further, in-depth study will berequired to investigate the molecular mechanism underlying Mena involvement in the development and progression of GC.\n\nThe present study was supported by the National Natural Science Foundation of China (grant nos. 81672661 and 81502268) and grants from the Guangdong Province Natural Science Foundation (grant nos. 2015A030310126 and 2015A030313182).\n\n![Expression level of Mena mRNA in GC and adjacent non-cancerous tissues. The expression levels of Mena mRNA relative to GAPDH in ten paired GC and adjacent non-cancerous tissues were evaluated by reverse transcription-quantitative polymerase chain reaction analysis. GC, gastric carcinoma.](ol-14-05-6024-g00){#f1-ol-0-0-6974}\n\n![Analysis of Mena protein expression by immunohistochemistry. Mena expression was predominantly localized in the cytoplasm of gastric tumor cells. Representative images of (A) negative, (B) positivestaining for Mena in normal gastric tissues; representative images of (C) negative, (D) positivestaining of Mena in gastric carcinoma tissues. Magnification, \u00d7200.](ol-14-05-6024-g01){#f2-ol-0-0-6974}\n\n![Kaplan-Meier OS curves for patients with gastric carcinoma tumors withhigh and low Mena expression. (A) OS curves for patients with high vs. low Mena expression levels. (B) OS curves for patients with late TNM stages (III--IVa) with high vs. low Mena expression levels. (C) OS curves for patients with early TNM stages (I--II) with high vs. low Mena expression levels. (D) OS curves for patients with T stage 3--4 gastric tumors with high vs. low Mena expression levels. (E) OS curves for patientswith T stage 1--2 gastric tumors with high vs. low Mena expression levels in patients. P-values were produced with a log-rank test. OS, overall survival; TNM, tumor-node-metastasis.](ol-14-05-6024-g02){#f3-ol-0-0-6974}\n\n###### \n\nAssociation between Mena expression level and clinicopathological characteristics.\n\n                                 Mena expression status               \n  ----------------- ------------ ------------------------ ----------- ---------\n  Total             106          50                       56          \n  Gender                                                              0.573\n  \u00a0\u00a0Male            67 (63.2)    33 (49.3)                34 (50.7)   \n  \u00a0\u00a0Female          39 (36.8)    17 (43.6)                22 (56.4)   \n  Age (years)                                                         0.065\n  \u00a0\u00a0\u226560             46 (43.4)    17 (37.0)                29 (63.0)   \n  \\<60              60 (56.6)    33 (55.0)                27 (45.0)   \n  T stage                                                             0.033\n  \u00a0\u00a01               10 (9.4)     8 (80)                   2 (20)      \n  \u00a0\u00a02               10 (9.4)     7 (70)                   3 (30)      \n  \u00a0\u00a03               84 (79.2)    34 (40.5)                50 (59.5)   \n  \u00a0\u00a04a              2 (1.9)      1 (50)                   1 (50)      \n  N stage                                                             0.313\n  \u00a0\u00a00               21 (19.8)    13 (61.9)                8 (38.1)    \n  \u00a0\u00a01               38 (35.8)    17 (44.7)                21 (55.3)   \n  \u00a0\u00a03               47 (44.3)    20 (42.6)                27 (57.4)   \n  M stage                                                             0.813\n  \u00a0\u00a00               99 (93.4)    47 (47.5)                52 (52.3)   \n  \u00a0\u00a01               7 (6.6)      3 (42.9)                 4 (57.1)    \n  TNM stage                                                           \\<0.001\n  \u00a0\u00a0I               13 (12.3)    12 (92.3)                1 (7.7)     \n  \u00a0\u00a0II              18 (17.0)    14 (77.8)                4 (22.2)    \n  \u00a0\u00a0III             68 (64.2)    23 (33.8)                45 (66.2)   \n  \u00a0\u00a0IV              7 (6.6)      1 (14.3)                 6 (85.7)    \n  Tumor size (cm)                                                     0.419\n  \u00a0\u00a0\u22655              74 (69.8)    33 (44.6)                41 (55.4)   \n  \u00a0\u00a0\\<5             32 (30.2)    17 (53.1)                15 (46.9)   \n  Grade                                                               0.570\n  \u00a0\u00a01               4 (3.8)      3 (75)                   1 (25)      \n  \u00a0\u00a02               25 (23.6)    11 (44.0)                14 (56.0)   \n  \u00a0\u00a03               76 (71.7)    36 (47.4)                40 (52.6)   \n  \u00a0\u00a04               1 (9)        0 (0)                    1 (100)     \n  Infiltration                                                        0.742\n  \u00a0\u00a00               101 (95.3)   48 (47.5)                53 (52.5)   \n  \u00a0\u00a01               5 (4.7)      2 (40.0)                 3 (60.0)    \n\nT, tumor; N, node; M, metastasis.\n\n###### \n\nCox-regression analysis of various prognostic parameters in patients.\n\n                           Univariate                Multivariate                             \n  ------------------------ ------------------------- -------------- ------------------------- -------\n  N stage                                            \\<0.001                                  0.002\n  \u00a0\u00a00                      Reference                                Reference                 \n  \u00a0\u00a01                      4.022 (1.955--8.274)                     1.718 (0.784--3.766)      \n  \u00a0\u00a03                      7.015 (3.421--14.386)                    3.273 (1.533--6.988)      \n  Age                                                0.001                                    \n  \u00a0\u00a0\u226560                    Reference                                                          \n  \u00a0\u00a0\\<60                   0.481 (0.319--0.727)                                               \n  Tumor size (cm)                                    0.001                                    \n  \u00a0\u00a0\\<5                    Reference                                                          \n  \u00a0\u00a0\u22655                     0.439 (0.272--0.709)                                               \n  Mena expression status                             \\<0.001                                  0.010\n  \u00a0\u00a0Negative               Reference                                                          \n  \u00a0\u00a0Positive               0.433 (0.284--0.661)                     0.463(0.296--0.724)       \n  T stage                                            \u00a0\u00a00.001                                  0.005\n  \u00a0\u00a01                      Reference                                Reference                 \n  \u00a0\u00a02                      17.539 (2.207--139.398)                  9.680 (1.142--82.080)     \n  \u00a0\u00a03                      36.233 (4.970--264.173)                  16.096 (1.974--133.049)   \n  \u00a0\u00a04                      16.855 (1.516--188.064)                  2.845 (0.217--37.316)     \n\nHR, hazard ratio; CI, confidence interval; N, node; T, tumor.\n\n[^1]: Contributed equally\n"}
+{"text": "1. Introduction {#sec1-sensors-18-02203}\n===============\n\nHuman Activity Recognition (HAR) is a major area of study and, with the growing diffusion of mobile devices, the past few years have witnessed the use of smartphone-based approaches to HAR becoming increasingly popular. Although sensor based HAR in general is an advanced field of investigation, there are still open challenges to be addressed. In addition, smartphone-based solutions, while representing an opportunity, also introduce new challenges. Some of the assumptions generally valid in sensor based HAR may be false when approaching the problem using the smartphone, for instance, the assumption that location and orientation of the sensor are known a priori (as in the case of wearable sensors) \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\]. Moreover, different users may have different ways of carrying the smartphone with them.\n\nBesides these new set of challenges, all obstacles common to any HAR approach have to be faced. Specifically, solutions are (in most cases) trained offline, and often using data exclusively collected in a controlled environment \\[[@B2-sensors-18-02203]\\]. An analysis of performance of these solutions has shown how accuracy levels can drop significantly when moving to free-living environments \\[[@B3-sensors-18-02203],[@B4-sensors-18-02203]\\]. Another problem resides in the lack of personalization. Solutions trained offline on a subset of users' data have been shown to perform with a lower accuracy rate on unknown users (i.e., where their data has not been used to train the classifier) \\[[@B1-sensors-18-02203]\\]. Consequently, research interest towards *online* approaches to HAR (i.e., methods able to implement a training mechanism that can be performed locally and in real time) has recently been growing \\[[@B2-sensors-18-02203]\\]. The biggest obstacle to online methods, however, is the need for labelled data to be used both as data points for training models, and as a ground truth dataset to validate their performance \\[[@B5-sensors-18-02203],[@B6-sensors-18-02203]\\]. This is a general problem for all supervised learning methods, which not only require the presence of a large dataset, but also require human supervision to annotate the dataset. Despite this limitation, supervised methods represent the most common approach to HAR, since achieving HAR in a completely unsupervised fashion may be a difficult task to be accomplished \\[[@B7-sensors-18-02203]\\].\n\nMany solutions have been developed to support and facilitate the process of data annotation \\[[@B5-sensors-18-02203],[@B8-sensors-18-02203],[@B9-sensors-18-02203]\\]. Similar systems can significantly reduce the time required by the data labeling process. Nevertheless, these mechanisms are viable only for offline training methods, where models are trained beforehand on a limited number of subjects and are not viable for online and personalized approaches. In other studies, such as \\[[@B10-sensors-18-02203]\\], crowd labeling solutions of prompting users in order to generate labels have been proposed. This facilitates the collection of data in a semi-automatic fashion where the user either confirms/corrects a label generated automatically or directly provides a new annotation. Crowd labeling approaches facilitate the development of online and personalized training; however, user interaction is still required to label the data, making the process time-consuming and potentially invasive from the user's perspective. Moreover, crowd labeling approaches can introduce labeling errors either due to mistakes or by the user misinterpreting the meaning of a label \\[[@B11-sensors-18-02203]\\].\n\nAn alternative approach to reduce the need for labelled data points is to consider semi-supervised methods that apply methods such as label propagation allowing a classifier to be trained using a dataset of which only a small percentage data points are labelled \\[[@B12-sensors-18-02203]\\]. In either case, human interaction is still required to produce a first level of annotation.\n\nAs an alternative to data-driven methods (both supervised and unsupervised), knowledge-driven methods have also been proposed. These approaches aim to build activity models based on a-priori information, thus not requiring the presence of labelled dataset \\[[@B13-sensors-18-02203]\\]. Despite the main advantage of not requiring annotation, these models in general are not robust to data uncertainty and/or to inter-person variability \\[[@B13-sensors-18-02203]\\]. To overcome either methods' limitations, hybrid approaches (combining knowledge and data-driven) have subsequently been proposed \\[[@B13-sensors-18-02203]\\].\n\nIn this work, we target the data labeling problem by proposing a knowledge-driven automatic labeling method that enables an online supervised training approach. The training phase is encompassed using *weak* labels generated automatically by a heuristic function combining step count and GPS information. The final classification is performed online and locally using only data from the accelerometer. In ref. \\[[@B14-sensors-18-02203]\\], we explored the viability of a supervised approach relying solely on automatic labeling of the training dataset in the form of weak labels to train an accelerometer-based classifier. Although encouraging, preliminary results have shown how label noise (produced by the heuristic) affects the final recognition accuracy. In particular, a larger number of samples is required to reach acceptable levels of accuracy, and, even when collecting a larger dataset results obtained, are less accurate when compared to the same classification approach using a manually labeled dataset. We present here a new heuristic function that is able to generate (along with weak labels) a probability estimation of the label quality. This information can be beneficial when balancing the dataset across the list of activities of interest, allowing the removal of a labeled set of samples believed to be less reliable, while keeping data points more likely to contain higher quality annotations. Using the improved heuristic version, an overall accuracy of 84% has been observed against the initial 74% accuracy obtained in ref. \\[[@B14-sensors-18-02203]\\].\n\nThis work presents a novel method for the automatic labeling of a dataset that reduces the burden of generating a training dataset. In addition, results present a comparison of accuracy performances, examining robustness of conventional Machine Learning (ML) approaches to label noise, which may be relevant not only for HAR using automatic labeling.\n\nThe remainder of this article is structured as follows: [Section 2](#sec2-sensors-18-02203){ref-type=\"sec\"} summarizes relevant literature, the state of the art and limitations of current approaches. In [Section 3](#sec3-sensors-18-02203){ref-type=\"sec\"}, we present a heuristic function used to automatically label acquired data points. The methodology is explained in [Section 4](#sec4-sensors-18-02203){ref-type=\"sec\"}. [Section 5](#sec5-sensors-18-02203){ref-type=\"sec\"} describes the dataset used to evaluate the experiment. Results are presented in [Section 6](#sec6-sensors-18-02203){ref-type=\"sec\"}. Finally, [Section 7](#sec7-sensors-18-02203){ref-type=\"sec\"} and [Section 8](#sec8-sensors-18-02203){ref-type=\"sec\"} provide discussion of results obtained, and possible directions for future work, respectively.\n\n2. Background {#sec2-sensors-18-02203}\n=============\n\nThis section provides an overview of sensor based HAR and related challenges, providing particular emphasis on smartphone-based and more specifically accelerometer-based approaches. Moreover, while presenting this summary of more common approaches, we wish to highlight the advantages and disadvantages of each approach with respect to the dataset annotation problem, and to online and personalized training. This section summarizes the most common sensor modalities, feature extraction techniques and classification approaches. In addition, we will refer to the most common classes of target activities, known limitations and open challenges for HAR.\n\n2.1. Smartphone-Based Activity Recognition {#sec2dot1-sensors-18-02203}\n------------------------------------------\n\nIn HAR, the most common sets of target activities include 'sitting', 'standing', 'walking', 'running', 'cycling' or 'using some means of transportation' \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\]. In some cases, subclasses are also considered such as 'walking up stairs' and 'walking down stairs'. Some solutions can also consider walking with the phone in pocket, rather than keeping it in the hand as two different classes \\[[@B2-sensors-18-02203]\\]. In other cases, classes can be representative of a wider spectrum of activities, as in the case of considering a single class 'standing/walking', or distinguishing macro classes such as 'active' and 'inactive' (for instance to monitor sedentary behaviors) \\[[@B4-sensors-18-02203],[@B15-sensors-18-02203]\\]. Some HAR solutions consider an extended set of target activities that also include some Activities of Daily Living (ADLs) (e.g., cooking, eating or cleaning) \\[[@B2-sensors-18-02203],[@B12-sensors-18-02203]\\]. This is, however, more common in solutions involving the use of environmental sensors (as for example within a smart home scenario), and only a small percentage of smartphone-based systems have been targeting ADLs \\[[@B2-sensors-18-02203],[@B4-sensors-18-02203]\\].\n\nThe classification task can be performed *offline* (when data processing is performed asynchronously and at a later stage with respect to data collection) or *online* (recognition performed in real-time) \\[[@B6-sensors-18-02203]\\]. Similarly, the training phase of models can be performed offline or online \\[[@B2-sensors-18-02203],[@B4-sensors-18-02203]\\]. The combination of offline training and online classification represent the most common approach \\[[@B2-sensors-18-02203],[@B4-sensors-18-02203]\\]. Offline training is often performed in controlled environments, although moving the solution to a free-living setting has been shown to affect the system causing a significant decrease in the accuracy performance \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203],[@B3-sensors-18-02203]\\]. Moreover, offline trained solutions have been observed to perform with lower accuracy rates for unknown subjects (i.e., not used for training purpose), and generally solutions present a significant inter-person variability affecting the recognition rate \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203],[@B6-sensors-18-02203]\\]. On the other hand, *online* training opens the way to personalized solutions. In this case, a pre-trained classifier can be updated when new data (subject specific) become available, in order to reduce the effect of inter-person variability \\[[@B2-sensors-18-02203]\\]. Personalized training, however, exacerbates the need for labeled datasets for training purposes making the approach not viable using conventional supervised methods.\n\nA generic Activity Recognition (AR) process can be divided into a finite set of stages sometimes referred to as Activity Recognition Chain (ARC) \\[[@B6-sensors-18-02203]\\]. A typical ARC will contain the following stages: sensing (data collection), pre-processing, feature extraction and classification.\n\nIn the following subsections, the most common methods for each of these steps are described. Finally, open challenges for HAR are evaluated.\n\n2.2. Sensing Modalities {#sec2dot2-sensors-18-02203}\n-----------------------\n\nThe range of sensors that has been used in sensor-based HAR is rather wide, including motion, contact and inertial sensors, RFID based systems and audio-based solutions. Similarly, many types of sensors have been considered also in the case of smartphone-based HAR. In this case, however, the choice is limited by the range of sensors commonly on board. Smartphone-based solutions rely, in most cases, on inertial sensors. Inertial sensors (accelerometer and gyroscope) are, by far, among the most common used sensors for a number of reasons: they are present in almost all modern smartphone models independently from the model price range;they have proved to provide good recognition accuracy rates;the data provided do not pose privacy issues (e.g., as in the case of microphones);and they are less power demanding compared to other sensors such as GPS.\n\nAccelerometers have been more widely adopted, sometimes in combination with gyroscopes and/or magnetometers \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\]. Accelerometer extracted features have been shown to be more informative, and the combination with data from gyroscopes and magnetometers has been demonstrated to improve the accuracy by 5% \\[[@B1-sensors-18-02203]\\]. In addition, the combination of inertial sensors with other sources of information, such as GPS, barometers, and (more rarely) microphones, has been shown to improve accuracy \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\]. For instance, the combination of accelerometer data with barometer data can provide better accuracy in distinguishing between walking down and up the stairs, compared to an accelerometer-only based solution \\[[@B1-sensors-18-02203],[@B16-sensors-18-02203]\\]. Even if the combination of multiple sensors can be considered as beneficial in terms of accuracy performance, accelerometer-only based approaches can be considered as optimal in terms of the tradeoff between accuracy rates and resource consumption (required computational complexity and power consumption) \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\], a fact of particular relevance for smartphone embedded solutions.\n\n2.3. Preprocessing {#sec2dot3-sensors-18-02203}\n------------------\n\nInertial sensor signals are often preprocessed before feature extraction \\[[@B1-sensors-18-02203]\\]. In most cases, preprocessing includes a filtering and a segmentation step. In the case of HAR, for instance, a low-pass filter can be used to exclude high-frequency signal components. In fact, for more common target activities, 98% and 99% percent of the information are contained in the 1--10 Hz, and in the 1-15 Hz frequency bands, respectively \\[[@B1-sensors-18-02203]\\]. Consequently, most common sampling rates for inertial sensors are in the range of 20--30 Hz to avoid undersampling while avoiding unnecessary battery consumption that higher sample rates would require \\[[@B1-sensors-18-02203]\\]. Segmentation is generally performed using a sliding window approach (either with overlapping or non-overlapping techniques) \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\]. Window sizes, in most cases, vary in the range of 1--10 s intervals. The optimal window size for feature extraction depends on the set of target activities, however, for common target activities of HAR, a window size of 1 s has been often identified as optimal \\[[@B1-sensors-18-02203],[@B17-sensors-18-02203]\\].\n\n2.4. Feature Extraction {#sec2dot4-sensors-18-02203}\n-----------------------\n\nExtracted features can be divided into two main categories: time domain and frequency domain features. Common time domain features include statistical moments of the signal such as mean, variance, skewness and kurtosis \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203],[@B6-sensors-18-02203]\\]. Other common time domain features are minimum and maximum values in the interval, and distance between peaks \\[[@B1-sensors-18-02203],[@B6-sensors-18-02203]\\]. Frequency domain features require Power Spectral Density (PSD) calculation (using Discrete Fourier Transform (DFT)), and are therefore more computationally demanding with respect to time domain features \\[[@B1-sensors-18-02203]\\]. Common frequency domain features are PSD energy and entropy or number and locations of peaks in the PSD \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203]\\].\n\nIn the case of inertial sensors, features can be extracted out of the three channels that three-axis accelerometers and gyroscope provide. Information extracted from the three axes can lead to better accuracy results, especially when location and orientation can be assumed to be known a priori. However, this is not the case for smartphone-based HAR where this particular assumption is not valid, since users may have different ways to carry the smartphone. To deal with this problem, three main alternatives have been explored \\[[@B1-sensors-18-02203]\\]: to adopt a hierarchical strategy in which a classification has the goal of detecting location and orientation of the sensor, and then an appropriately trained classifier can be used for recognition (this solution, however, exacerbates the complexity of generating the training dataset for all possible combinations) \\[[@B18-sensors-18-02203]\\]a much simpler approach is to use the 3D magnitude of the signal \\[[@B19-sensors-18-02203]\\]finally, sign-invariant features (as the absolute value of features on the three axes) can be extracted providing robustness to some orientations and considering only the axis from the top to the bottom of the screen and the one pointing out of the screen \\[[@B15-sensors-18-02203]\\].\n\nIn ref. \\[[@B6-sensors-18-02203]\\], a comparison of results obtained using different sets of features showed how time-domain features (particularly mean and variance) were the most informative. Nonetheless, in many cases, a combination of time and frequency domain features have been used \\[[@B1-sensors-18-02203],[@B2-sensors-18-02203],[@B20-sensors-18-02203]\\].\n\nIn some cases, the classification process does not involve feature extraction. For example, Deep Learning (DL) approaches can make direct use of raw data for classification \\[[@B1-sensors-18-02203]\\].\n\n2.5. Classification Approaches {#sec2dot5-sensors-18-02203}\n------------------------------\n\nSeveral classification approaches have been used for HAR using supervised methods. In smartphone-based HAR, Decision Trees (DTs), Support Vector Machines (SVMs) and k-Nearest Neighbors (kNN) are among the most common \\[[@B2-sensors-18-02203]\\]. DT approaches have the advantage of being among the computationally less expensive models, and for this reason their use in smartphone-based applications is relatively common; however, they may not be the best choice for online methods (since the entire model needs to be recalculated when new samples are acquired) \\[[@B1-sensors-18-02203]\\]. kNN have the advantage of being among the most convenient instance-based methods for online learning, since the adaptation of the model to new data points is simply resolved by storing new calculated features vectors \\[[@B1-sensors-18-02203]\\]. SVM methods have shown good accuracy results and they share with DT the advantage of not being too computationally expensive, at least at the prediction stage \\[[@B1-sensors-18-02203]\\]. Probabilistic models such as Hidden Markov Models (HMMs), Conditional Random Fields (CRFs) and Skip Chain CRFs (SCCRFs) have also been used \\[[@B21-sensors-18-02203],[@B22-sensors-18-02203],[@B23-sensors-18-02203]\\]. Finally, some studies have investigated the use Neural Networks (NNs) for HAR purposes. Although NN models may be computationally demanding (computational complexity can rapidly grow with the number of features in use), they are still viable for online learning approaches, where NN weights can be updated when new data points are available.\n\nAs previously mentioned, more recent works have explored DL approaches \\[[@B1-sensors-18-02203]\\], specifically the use of Convolutional Neural Networks (CNNs) \\[[@B24-sensors-18-02203],[@B25-sensors-18-02203]\\] and Deep Belief Networks (DBNs) \\[[@B26-sensors-18-02203]\\]. Although the use of DLs appears to be more accurate, their use may not represent an optimal solution for smartphone embedded solutions due to the required computational complexity. Some studies (such as as \\[[@B26-sensors-18-02203]\\]), however, have been using DL on the smartphone, at least at prediction stage, for real-time classification.\n\n2.6. Classification Accuracy {#sec2dot6-sensors-18-02203}\n----------------------------\n\nOverall evaluation of performance is generally presented using precision, recall and f-score \\[[@B1-sensors-18-02203],[@B4-sensors-18-02203]\\]. Confusion matrices are more often used to show detailed inter-class variability and distribution of false positives across the target classes \\[[@B1-sensors-18-02203],[@B4-sensors-18-02203]\\].\n\nPast studies in smartphone-based HAR exhibit quite a diverse range of accuracy values between 80% and 97% \\[[@B4-sensors-18-02203]\\]. These variations can be due both to the dataset used, the classification method, and the specific set of target activities being considered. These variations make the comparison of performances a complex task \\[[@B1-sensors-18-02203],[@B4-sensors-18-02203]\\]. For instance, unreliable accuracy levels can be measured when the sample size used for testing is too small. On the other hand, increasing the number of samples used for testing can affect performances reducing the size of training data \\[[@B1-sensors-18-02203]\\]. In a controlled environment, accuracy can reach 97%, although accuracy has been observed to be in the 63--86% for unknown subjects \\[[@B4-sensors-18-02203]\\]. Similarly, performance variations have been observed according the position/s of the smartphone taken into account. This variation can be avoided by restricting the examined position to the most common case (for example, trouser pocket) \\[[@B4-sensors-18-02203]\\].\n\nThe set of target activities can also impact performance. For instance, in some cases, certain classes have been reported as conflicting with others, thus resulting in lower accuracy levels \\[[@B1-sensors-18-02203],[@B4-sensors-18-02203]\\]. An example is the biking (or cycling) activity that has been reported in some cases as conflicting with the walking class \\[[@B4-sensors-18-02203]\\]. Overall accuracy has been observed to vary between 92% excluding cycling and 73% including the cycling activity \\[[@B4-sensors-18-02203]\\].\n\n2.7. Challenges in Smartphone-Based HAR {#sec2dot7-sensors-18-02203}\n---------------------------------------\n\nAlthough HAR has been an active research field for many years, there are still many challenges to address. One of the main challenges is the need for labelled datasets for ground truth annotation \\[[@B6-sensors-18-02203]\\]. This challenge is particularly relevant for both online and personalized approaches. Online methods require a way to obtain data labels directly in free-living conditions, while personalized methods further exacerbate the need for data requiring user specific datasets. Another major issue to be faced is class imbalance \\[[@B6-sensors-18-02203]\\], particularly when collecting data in free-living situations where samples corresponding to some of the activities may be less likely to occur. The process of training using highly imbalanced datasets may alter the final recognition rate significantly, while approaches to balance the training dataset, as in the case of undersampling \\[[@B27-sensors-18-02203]\\] (i.e., random elimination of data points in the majority class) cause a significant decrease in data available for training.\n\nData labelling for ground truth generation is a time-consuming task posing a major obstacle to all supervised methods. To deal with the problem, many approaches have been proposed. An example is the development of annotation tools that facilitate manual annotation attempting (at least partially) to automate the process \\[[@B8-sensors-18-02203],[@B28-sensors-18-02203]\\]. These tools can drastically reduce the effort required to annotate large amounts of data; however, they can only be used in offline conditions. Moreover, annotation tools often rely on video footage for annotation \\[[@B8-sensors-18-02203]\\], making their use less frequent in free-living settings due to inherent privacy issues.\n\nCrowd labelling approaches have been proposed, particularly in the case of the smartphone, where final users can directly annotate data fragments \\[[@B10-sensors-18-02203]\\]. Nevertheless, human interaction is still required at some stage to produce the labelled dataset. This results in the process being potentially time-consuming from the user perspective and poses the problem of label accuracy \\[[@B12-sensors-18-02203]\\]. Semi-supervised methods have been proposed to deal with this problem, however, mostly in other domains \\[[@B4-sensors-18-02203]\\]. Semi-supervised methods require only a subset of samples to be labelled. With this, they can still make use of unlabeled data points in the training process \\[[@B12-sensors-18-02203],[@B29-sensors-18-02203]\\]. Common approaches are label propagation or Multi-Instance Learning (MIL), where labels are assigned to sets of instances rather than to single data points \\[[@B29-sensors-18-02203],[@B30-sensors-18-02203]\\]. The semi-supervised approach has, however, rarely been applied to HAR \\[[@B4-sensors-18-02203]\\], as in \\[[@B12-sensors-18-02203]\\]. Moreover, although these methods can reduce the required time effort to produce ground truth data, partial manual annotation is still required. [Table 1](#sensors-18-02203-t001){ref-type=\"table\"} summarizes some recent HAR approaches using inertial sensors. The comparison highlights how the problems of data annotation and label noise have been rarely addressed.\n\nIn most cases, studies assume availability of a fully labeled dataset. Only refs. \\[[@B12-sensors-18-02203],[@B36-sensors-18-02203]\\] have tackled the annotation problem by means of semi-supervised approaches requiring partial annotation. In ref. \\[[@B10-sensors-18-02203]\\], user solicitation is proposed to generate labels online; however, in ref. \\[[@B12-sensors-18-02203]\\], the use of prompting has been combined with label propagation to minimize the required amount of annotation. In ref. \\[[@B36-sensors-18-02203]\\], an alternative approach has been proposed to reduce the required amount of labeled data, aiming to exploit activity models trained on other users. Similarly, only \\[[@B12-sensors-18-02203],[@B34-sensors-18-02203]\\] have examined accuracy performance considering label noise.\n\n3. Implementation {#sec3-sensors-18-02203}\n=================\n\nThe goal of our research is to investigate novel methods to manage the data annotation problem, thus reducing the burden of ground truth generation and opening the way to online and personalized training methods. More specifically, we present an online annotation method aiming to solve the problem of ground truth generation by means of automatic labeling. Activity monitoring is performed using the smartphone and aiming to generate labels outdoor and indoor combining accelerometer and GPS information.\n\nOur study explores the use of *weak labels*, i.e., noisy labels generated automatically using a heuristic function. The heuristic function combines GPS and step count information in a knowledge-driven model, in order to generate the weak labels that are then used to build the training dataset, thus enabling an online learning approach in a free-living setting.\n\n3.1. Online Training Architecture {#sec3dot1-sensors-18-02203}\n---------------------------------\n\nAs in ref. \\[[@B14-sensors-18-02203]\\], we propose an online training solution that relies on part of the process to be performed on a centralized server application. Specifically, while data collection and final classification are performed locally and in real time on the smartphone, the training phase is performed remotely (server-side). As depicted in [Figure 1](#sensors-18-02203-f001){ref-type=\"fig\"}, the smartphone collects and send raw data to the server along with heuristic generated weak labels. On the server side, the training phase starts whenever a sufficient number of data points have been collected. Finally, parameters of the trained model (in Predictive Model Markup Language (PMML) format) can be sent back to the smartphone to instantiate the trained classifier to perform HAR locally.\n\nThis solution provides two main advantages: it reduces the burden of performing the training on the smartphone, performing this computationally expensive task remotely,it allows for comparing performance (offline) with different classification methods and feature sets at a later experimental stage.\n\nA more efficient scheme following the same architecture can be obtained, however, by sending only extracted features, instead of raw data signals as we did in the current work, to leave the possibility of experimenting using different features sets.\n\n3.2. The Heuristic Function for Automatic Labeling {#sec3dot2-sensors-18-02203}\n--------------------------------------------------\n\nThe heuristic function combines step count and GPS information in order to label data samples. Step detectors are often directly available in modern smartphones. In cases where a step detector is not present on board, an estimation of the step count can be obtained using a peak detection approach based on the magnitude of the 3 axes of the accelerometer signal \\[[@B14-sensors-18-02203]\\]. The GPS is also among the most common on-board sensors in smartphones, and major mobile operating systems allow easy retrieval of the GPS coordinates along with an estimation of the location accuracy.\n\nIn ref. \\[[@B14-sensors-18-02203]\\], we presented an initial version of the heuristic function combining GPS data and step count information in order to automatically label data fragments. This previous experiment highlighted how a supervised approach (trained using weak labels) affects the overall accuracy in comparison with a fully supervised approach trained using a manually labelled dataset. The difference is related to the fact that the heuristic generates an annotation at a coarse level. This is due to different acquisition rates (GPS info update less frequently) that result in higher uncertainty of weak labels compared to the manual labelled ground truth. Preliminary results in ref. \\[[@B14-sensors-18-02203]\\] have also shown different learning curves between the manually labeled and the automatically generated labels. The fully supervised approach (using manually labeled dataset) converges more rapidly to an acceptable accuracy rate. This is reflected in a significative gap in the overall accuracy, particularly when the training dataset is small; however, the gap is partially reduced using a larger number of samples for training.\n\nIn this work, we present an improved version of the heuristic that allows an estimate of the reliability of the generated weak labels that can be used to optimize data points selected for training. This section describes the heuristic function, and how it estimates the *quality* of labels. We then describe selected features, and the classification approach. Moreover, in this paper, we present results calculated over a larger dataset and targeting a different set of activities. In this case, we define the target activities as 'sitting', 'walking', 'running', 'cycling' and 'transportation' (while the study in ref. \\[[@B14-sensors-18-02203]\\] was excluding 'running' and 'cycling').\n\nAs in ref. \\[[@B14-sensors-18-02203]\\], a manually labeled ground truth dataset has been collected for validation. The comparison between the ground truth (manually labeled) and the weak ground truth (labeled automatically) provides information on the quality of labeling of the heuristic, which constitutes essential information in order to evaluate the final accuracy rate.\n\nStep count is used as metric for a first heuristic rule, which can detect a number of activities based on the steps/minute (spm) rate as in \\[[@B14-sensors-18-02203]\\]. A walking pattern corresponds typically to a pace of 90--110 spm \\[[@B39-sensors-18-02203],[@B40-sensors-18-02203]\\]. Similarly, a running stride normally resides in the 160--180 spm range \\[[@B41-sensors-18-02203]\\]. Having the step count information, the spm rate is sampled on a 1 min fragment. This window allows the first level heuristic to deduct, for instance, that a consistent 102 spm rate in a 1 min window will most likely correspond to a 'walking' pattern. The second step is a merging phase. Consecutive time windows are merged when the difference between the two is less then 10 spm. This allows detection of prolonged walking sessions (e.g., longer than 10 min) and prolonged time with spm 0 or close to zero, which the heuristic will label as sedentary. The merging phase is particularly useful when combining the step count information with the GPS, since the GPS is sampled at a lower sampling rate (polled at variable intervals between 1--3 min) to minimize the impact on battery consumption. The second step of the heuristic considers GPS locations within the merged interval. Consecutive timestamped locations are used to estimate user speed. If multiple speed estimations are present in the merged interval (obtained from the step count), the heuristic will calculate the weighted average speed on the merged interval. As in the case of the step count, estimated speed can be used to recognize the most probable locomotion pattern in outdoor activities (\u22431.4 m/s walking, 3--4 m/s running, cycling 4--8 m/s and using some mean of transportation \u226520 m/s) \\[[@B41-sensors-18-02203],[@B42-sensors-18-02203]\\]. The combination of the two information sources can reduce the presence of wrongly assigned labels. For instance, a measured speed comparable with riding some means of transportation allows the algorithm to ignore a false positive in a step count. Similarly, a consistent running spm rate can identify a running on a treadmill activity where the GPS would not produce any valuable information.\n\nIn ref. \\[[@B14-sensors-18-02203]\\], we directly assigned labels merging the step count and GPS information. In this enhanced version of the heuristic, a probability estimation is introduced aiming to assess the *quality* of produced weak labels using a probability model.\n\n### Assessing Quality of Labels {#sec3dot2dot1-sensors-18-02203}\n\nThe probability assigned to a weak label is calculated by merging probability functions corresponding to the target activities, modeled as *membership functions* of a fuzzy set. Fuzzification of values provides a convenient way to map values (steps/minute rate and estimated speed) into a probability model. The set of target activities is modeled into two fuzzy sets: one for the step rate and one for the GPS as depicted in [Figure 2](#sensors-18-02203-f002){ref-type=\"fig\"}. The set of membership functions for the step count is modeled using Gaussian distributions (with mean and standard deviation centered in the normal walking and running rate). The choice of modeling step count probability functions using a Gaussian distribution is based on the fact that walking and running rates are assumed to be normally distributed in the population. The spm rates for walking and running intervals are taken from \\[[@B39-sensors-18-02203],[@B41-sensors-18-02203]\\]. The set of membership functions for the GPS estimated speed is built using trapezoidal functions shaped considering estimated speed. The choice of trapezoidal functions is based on the fact that the assumption of estimated speed being normally distributed cannot be made in this case, since speed is affected by multiple factors (e.g., vehicles in use and traffic conditions).\n\nThe two fuzzy sets are used to generate two labels predicting the most probable ongoing activity independently using GPS and step information with a probability estimation for the two predicted labels $P_{GPS}\\left( label \\right)$ and $P_{step}\\left( label \\right)$. The membership functions used to estimate probability, are designed to produce low probability results (below 50%) for a range of values that could be ambiguous. For instance, a rate of 50 spm over a five-minute interval could either be produced by continuous walking at a very low pace, or it could be the case that the interval includes some walking and some sitting. Similarly, a rate of 130 spm could correspond to an interval including some walking and some running.\n\nThe sets of activities that the two information sources are able to detect do not coincide; i.e., the step count will be able to detect walking, running or to estimate sedentary intervals (as prolonged interval without an increase in the step count); however, it will not be able to provide valuable information to discriminate activities such as cycling or using some means of transportation where the GPS will be more informative. The estimation of final probability, therefore, distinguishes the cases of labels that can be detected by both information sources (sitting, walking, running) and activities that can be detected only by using the GPS. In the first case, the final probability will be estimated by combining the probability obtained from the two set fuzzy sets performing the weighted average of the two probability values: $$P\\left( Walk \\right) = \\frac{P_{step}\\left( walk \\right) + w_{GPS}P_{GPS}\\left( walk \\right)}{1 + w_{GPS}}.$$\n\nThe weighted average allows for using another piece information which is the accuracy of the GPS location. The weight of the GPS prediction ($w_{GPS}$) will be proportional to the accuracy rate of the two locations used to calculate the estimated speed. In this way, the GPS will be more informative in outdoor situations where the accuracy can be expected to be higher. For other cases, as discriminating between transportation and cycling, the produced label will be assigned based on the probability obtained from the GPS. Discriminating cycling and driving activities only using GPS may not be a reliable method in cases such as driving in traffic where the two activities can have comparable speeds. To reduce the presence of mislabeling of this type, after the annotation is generated, the final step of the heuristic is to consider the label generated at the previous interval, converting a consecutive sequence of the type 'Cycling'-'Transportation', or 'Transportation-Cycling' as 'Transportation'.\n\n[Figure 3](#sensors-18-02203-f003){ref-type=\"fig\"} presents an example of the annotation produced by the heuristic, in comparison with the manual annotation.\n\nThe heuristic can be expected to provide reasonably good precision (ratio of correctly assigned weak labels) under certain conditions, such as good GPS signal and occurrence of prolonged consistent activities. On the other hand, the approach is expected to have low recall performance (missing labels). This is because, in other cases, the heuristic will not be able to predict labels with high enough reliability (e.g., presence of high variability as alternating different activities, and/or poor GPS signal). Therefore, our approach is to use the heuristic information for data mining of a weakly labeled dataset, rather than use the heuristic directly to classify the ongoing activity. The automatically labeled dataset can then be used to train a classifier based using only inertial sensors' features, which will be able to classify the on-going activity continuously and also in cases where the heuristic would not be informative.\n\n3.3. Segmentation and Feature Extraction {#sec3dot3-sensors-18-02203}\n----------------------------------------\n\nAccelerometer data is sampled at 30 Hz. Segmentation was performed using a sliding window approach with a 50% overlap and 1 second as window size. Our approach mostly relies on time domain features; however, it includes some frequency domain features. The complete set of features as in [Table 2](#sensors-18-02203-t002){ref-type=\"table\"} includes statistical moments (mean, variance, skewness and kurtosis) extracted from the 3D magnitude (square root sum of the three axes) of the accelerometer signal, and mean, variance and range (max-min difference) at a single axis level. Frequency domain features are computed using the Fast Fourier Transform (FFT) to compute the PSD of the signal; then, the features extracted consist of the number of peaks in the PSD and location of the dominant peak.\n\n4. Methods {#sec4-sensors-18-02203}\n==========\n\nIn this experiment, we compared common classification approaches, prioritizing models that are less computationally demanding at the prediction stage. This comparison has the goal of identifying the best candidate that can be used to perform classification in real time and locally on a smartphone, when dealing with a training set that may contain noise.\n\n4.1. Experimental Protocol {#sec4dot1-sensors-18-02203}\n--------------------------\n\nAs in our previous experiment, the dataset has been manually labeled using a mobile app. The app has been installed on the user's personal smartphone. The user has been told to carry the smartphone as he would normally do. The app starts collecting data from GPS and the accelerometer in the background after first launch, with no further interaction needed for data collection purposes. The user was asked to use the app to annotate activities while performing his normal routine, allowing to have a manually labeled groundtruth to assess both the accuracy of the automatic labeling, and the final detection accuracy of a trained classifier. The user interface of the app has been designed to be intuitive, and no further training was required. The annotation process is done in real time while activities are being performed. The annotation is performed in the following way. The user selects the start of an activity clicking a label, and this activates the selected label on the interface. Labels are considered as mutually exclusive; therefore, the start of an activity would result in closing the one previously activated (if present). As depicted in [Figure 4](#sensors-18-02203-f004){ref-type=\"fig\"}, the interface provides icon buttons for target activities and an error button to notify the presence of an error in the labeling (e.g., user forgetting to activate the new activity in the transition). The error label is a special label in the dataset. This method allows the algorithm to ignore the last label preceding the error notification when generating the final ground truth. Moreover, to reduce the local uncertainty of the labeling process, a window of 4 s (2 before and 2 after the manually labeled transition) has been ignored between two consecutive activities.\n\n4.2. Data Analysis and Validation {#sec4dot2-sensors-18-02203}\n---------------------------------\n\nGiven that data collection is performed in free-living, the data collected is expected to produce a highly imbalanced dataset, where the sitting class can be expected to be the majority. We address the problem of balancing the training dataset using an undersampling approach \\[[@B27-sensors-18-02203]\\]; however, instead of performing random elimination, the estimated quality of labels is used to eliminate those samples believed to be less reliable. The balancing algorithm divides the data points following the partition set that the heuristic generates. All samples in each group are then sorted by the estimated probability that the heuristic provides, leaving at the top samples believed to be more representative. Finally, only the first *n*-samples are considered where *n* is defined considering the class with the minimum number of samples. The canonical undersampling approach would define $n = n_{min}$, where $n_{min} = {min{(|}}C_{1}\\left| , \\right|C_{2}\\left| , \\right|C_{3}\\left| , \\right|C_{4}\\left| , \\right|C_{5}{|)}$, i.e., *n* is equal to the minimum cardinality of the five classes corresponding to the target activities ('sitting, 'walking',' running','cycling' and 'transportation'). Instead, we allowed up to 30% more samples with respect to the minimum class. Therefore, for each set, the number of samples used for training would be: $$n_{i} = {min{(|}}C_{i}\\left| ,n_{min} \\ast 1.3 \\right).$$\n\nOnce the number of samples has been balanced across classes, we can generate the final dataset, where each sample will have the form depicted in [Table 3](#sensors-18-02203-t003){ref-type=\"table\"}. The first column is defined as the label manually collected, the second column contains the weak label (generated by the heuristic), followed by the features defining that sample. Although the data has been collected continuously, we restricted the dataset here considering only data portions for which manual labeling was available.\n\nThe validation routine repeats the following steps in a 10-fold cross validation: balanced samples sets are randomised and divided into a train and test set (10%),the training set is obtained by merging and shuffling the sets and using the weak labels for training,scoring of the classifier is performed on the test set using the labels manually collected.\n\n5. Dataset {#sec5-sensors-18-02203}\n==========\n\nThe dataset consists of 38 days of continuous recording of smartphone sensors in free-living, including raw data samples of accelerometer and GPS. The heuristic function was able to generate 200 h of raw data with weak labels. The manually labeled data portions consist of a total of approximately 100 h. Annotation has been performed in free-living over the 38 day period, whilst performing normal ADLs such as preparing/having breakfast, commuting to work and recording physical activity during working hours. [Table 4](#sensors-18-02203-t004){ref-type=\"table\"} summarizes the total labeled time and number of samples (with a 1 s window size) for all target activities.\n\nThe accelerometer has been sampled at 30 Hz, while GPS sampling was scheduled on a 1--3 min periodic sampling to reduce the impact on battery consumption. Data have been collected extending the dataset in ref. \\[[@B14-sensors-18-02203]\\] on one subject researcher. The position of the smartphone has been constrained to be the most common (trouser pocket). The training dataset obtained after balancing samples across the target activities consisted of about 2 h and 50 min (about 30 min per class). Considering the size of such a dataset, in this specific case, the last 10 days of recording are sufficient to generate a training set of equivalent size. The number of required days of observation, however, depends on how often the specific user performs the target activities or, in other words, on how activities are evenly distributed.\n\n6. Results {#sec6-sensors-18-02203}\n==========\n\nThis section describes results obtained in our experiment, providing first an analysis of the accuracy measured on labels generated by the heuristic function; then, we provide results obtained by comparing different classification approaches to evaluate how the final accuracy is affected by the initial uncertainty of labels.\n\n6.1. Heuristic Accuracy {#sec6dot1-sensors-18-02203}\n-----------------------\n\nAnalyzing the agreement between the manually labeled ground truth and the one generated using the heuristic, it is possible to measure the precision of the heuristic, as a measure of quality of the produced labeling. [Figure 5](#sensors-18-02203-f005){ref-type=\"fig\"} presents the performance of the heuristic evaluated over the dataset, highlighting the percentage of correct labels, the percentage of missing labels (in which the heuristic could not provide a label) and finally the number of candidate samples for data training. The accuracy of the heuristic (percentage of correctly assigned labels) is quite variable. This is because we can expect the heuristic to work better in situations with lower variability (e.g., presence of prolonged activities such as a long walk), rather than high variability situations alternating small fragments of different activities.\n\nThis accuracy fluctuation, however, is (at least partially) mitigated using the information on probability, resulting in a smaller number of samples actually considered as good training data points, when the overall accuracy decreases. The average precision rate of the heuristic despite fluctuations has been observed to be around 85%. On the other hand, the percentage of missing samples provides an idea of the heuristic recall, confirming that, in some cases, the heuristic is not able to generate a label in more than 50% of the cases.\n\n6.2. Performance Evaluation {#sec6dot2-sensors-18-02203}\n---------------------------\n\nThe evaluation of accuracy performance has been performed by comparing different training algorithms including: DTs, Random Forests, kNN, Nearest Centroid, multi-class SVM and NNs. NNs have been observed using different topologies (with one and two hidden layers) and using different *alpha* values (that can help to avoid or reduce overfitting in some cases). [Table 5](#sensors-18-02203-t005){ref-type=\"table\"} presents performance measured in terms of precision, recall and f-score, values averaged on the set of target activities.\n\n[Table 6](#sensors-18-02203-t006){ref-type=\"table\"} depicts the confusion matrix for the classifier with the highest f-score (an NN model using two hidden layers with network topology 18 \u00d7 36 \u00d7 12 \u00d7 6). The confusion matrix provides a more detailed visualization of how error rates are distributed over the target activities, and highlights which activities are more often confused.\n\n7. Discussion {#sec7-sensors-18-02203}\n=============\n\nThe analysis of the quality of the automatic labeling confirms a high variability, directly observable in the precision score of the function heuristic itself. Days with high variability (i.e., with frequent alternations between different activities in a short time) exhibit a lower percentage of correctly assigned labels. The probability estimation, however, allows the algorithm to detect the phenomenon under certain conditions, resulting in a smaller number of samples considered as good candidates for the training set. The observed average precision is between 80--90% meaning that label noise can be expected to affect 10--20% of the labels (considering the entire dataset), while label noise measured on the balanced dataset (used for training) appeared to be reduced by approximately 10%. The results highlight how some methods appear to be more robust to this label noise. It can be expected that instance-based methods can be particularly sensitive to mislabeled samples. In particular, for Nearest Centroid, mislabeled samples appear to decrease the accuracy, an effect that is partially mitigated with kNN. Similarly, DT presents high variability in f-score examining the 10-fold cross validation iterations, while Random Forests appear to be more robust. NNs (together with Random Forests) scored the highest f-score values, obtained using higher values of alpha, which reduces likelihood of overfitting, encouraging smaller weights in the network. Overall, the two methods showing the best results are NNs and Random Forests. These two methods appear to be more robust to label noise and show similar values in both precision and recall. These values are also in line with the average precision of the heuristic as could be expected. Although the two methods have similar performance, NNs may be identified as a better option because of their ability to support online training though partial training (i.e., updating the weights when new samples are available). With both NNs and Random Forests, the trained model appears to preserve the initial heuristic precision, while at the same time augmenting the overall recall, since the model (in contrast with the heuristic) will be able to generate predictions continuously. In other cases (Multi-class SVM and Nearest Centroid), label noise appears to impact finally accuracy, resulting in a decrease in precision between 3% and 15%. Moreover, label noise is not evenly distributed across classes. Samples in the 'transportation' class are sometimes labeled as 'cycling' because of the speed measured on the GPS (as in the case of driving in traffic). Examining the confusion matrix, it can be observed how this puts into conflict the transportation and the cycling class. This is a direct drawback of the choice of not sampling the GPS continuously to reduce the battery drain. Nevertheless, a more accurate labeling can be achieved using continuous GPS monitoring. Similarly, in some cases, the 'walking' class is in conflict with the 'cycling' class. In this case, the difference can be only partially attributed to some incorrect labeling; however, it can also be attributed to a features set generating similar samples for the two classes in some cases.\n\nOverall, this enhanced version of the heuristic appears to improve significantly the final classification accuracy. In particular, the substitution of random elimination (as performed in a previous experiment \\[[@B14-sensors-18-02203]\\]) with the elimination of samples, considered to be less reliable, improves the quality of labeling of the training set by reducing label noise.\n\nResults obtained appear to be in a similar range of accuracy of other related works, as can be observed comparing accuracy with similar studies reported in [Table 1](#sensors-18-02203-t001){ref-type=\"table\"}. The comparison in terms of performance, however, is a hard task, due the fact that other studies may consider a different set of target activities and/or do not consider label noise. On the other hand, our approach requiring the presence of GPS data together with inertial data did not allow for comparing results on publicly available datasets because, in most cases, GPS data is missing and/or data have been collected in a controlled environment. Another limitation of this study is that the dataset (being single user) does not allow to assess inter-person variability or to assess accuracy of personalized methods against the approach using one classifier for all users. Instead, the focus of this study has been limited to assessing viability of an online training approach reducing the burden of generating an annotated dataset by means of automatic labeling. This represents a necessary step on the way towards personalized approaches and will also facilitate the collection of a multiple-user dataset in the future.\n\n8. Conclusions {#sec8-sensors-18-02203}\n==============\n\nWith respect to the previous study \\[[@B14-sensors-18-02203]\\], the estimation of a label's quality appears to increase the overall accuracy significantly. In particular, replacing random elimination of samples with a heuristic based approach appears to reduce the noise introduced by the automatic labeling. The results obtained demonstrate how an online training approach can be achieved avoiding the burden of manual labeling of data, potentially facilitating the development of personalized systems. In ambiguous cases, it could be useful to prompt the user to verify the correctness of labels. This functionality could be restricted only to some specific cases (e.g., ambiguity between 'cycling' and 'transportation') to avoid unnecessary user solicitations. In a similar way, some optimization mechanism can be developed to adapt the probability estimation of labels to the specific user. This adaptation can help to distinguish users exhibiting different walking patterns. A simple way to realize the adaptation could be performed by centering the Gaussian membership function of the fuzzy set, around the normal walking rate of the specific user.\n\nIn this paper, we proposed an automatic labeling approach to deal with the data annotation problem. An estimation of the quality of the labeling produced has also been proposed, and the use of such information has proven to be useful when balancing the dataset, allowing to reduce presence of label noise in the training set. Another interesting outcome of this experiment has been the comparison of conventional classification in presence of label noise. The comparison highlighted how some methods such as DTs or SVM (which have been considered as viable classification methods for smartphone-based HAR \\[[@B1-sensors-18-02203]\\]) may not represent an optimal solution when in the presence of label noise.\n\nA multiple-user dataset is, however, required to measure the advantage of personalization against the one-classifier for an all users approach. Similarly, a larger dataset would allow for comparing users with different ways of carrying the smartphone, while, in this experiment, it has been restricted to the most common (trouser pocket).\n\nThe next steps will include the collection of a multiple-user dataset to further investigate personalized approaches to HAR. In conclusion, this study provided two main outputs. With respect to the preliminary investigation in \\[[@B14-sensors-18-02203]\\], it highlights how information on the quality of produced labeling can be beneficial. Finally, the experiment has identified some viable classification methods as being more robust to a dataset subject to label noise.\n\nF.C., I.C. and C.N. conceived the experiments; F.C. performed the experiment and analyzed the data; P.M., K.S. and J.H. contributed to defining the methodology and the design of the experiment; F.C. wrote the paper. All coauthors contributed to the final paper and reviewing the manuscript.\n\nThis work has been funded by the European Union Horizon 2020 Marie Sk\u0142odowska Curie Action Innovatice Training Networks (MSCA-ITN) ACROSSING project (GA no. 676157).\n\nThe authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.\n\nThe following abbreviations are used in this manuscript: ADLActivity of Daily LivingARActivity RecognitionARCActivity Recognition ChainCRFConditional Random FieldCNNConvolutional Neural NetworkDBFDeep Belief NetworkDFTDiscrete Fourier TransformDLDeep LearningDTDecision TreeEPEmerging PatternsFFTFast Fourier TransformHARHuman Activity RecognitionHMMHidden Markov ModelkNNk-Nearest NeighborsMILMulti-Instance LearningMLMachine LearningNNNeural NetworksPMMLPredictive Model Markup LanguagePSDPower Spectral DensitySCCRFSkip-Chain CRFSVMSupport Vector Machine\n\n![The smartphone app collects raw data samples from the on-board accelerometer and the GPS. The GPS is combined with the step count to generate weak labels using the heuristic function. Extracted features are sent to the server along with the labels in order to train the model. Parameters of the classifier are sent back to the app in Predictive Model Markup Language (PMML) format. Finally, the classification can be performed locally on the smartphone instantiating the appropriate classifier based on the parameters.](sensors-18-02203-g001){#sensors-18-02203-f001}\n\n![Fuzzification of probabilities values for the heuristic function. Step count heuristic (**a**) is modeled using Gaussian membership functions with average and standard deviation based on common steps/minute rates for walking and running. Similarly, (**b**) shows the trapezoidal membership functions used to estimate probability based on measured speed in m/s, for walking, running, cycling and using some means of transportation.](sensors-18-02203-g002){#sensors-18-02203-f002}\n\n![Comparison of ground truth (**a**) with heuristic generated annotation (**b**).](sensors-18-02203-g003){#sensors-18-02203-f003}\n\n![Screenshot of the labeling app showing the buttons to label activities with the 'Walking' activity selected as currently on-going. The error button allows the user to signal whenever an error occurs in the labeling, allowing for ignoring incorrect annotations for the final evaluation.](sensors-18-02203-g004){#sensors-18-02203-f004}\n\n![Accuracy of heuristic function over the 38 days composing the dataset. Label precision has been calculated as the number of correct labels divided by the total number of weak labels. Percentage of missing labels has been calculated as the number of missing labels divided by the number of total labels manually annotated.](sensors-18-02203-g005){#sensors-18-02203-f005}\n\nsensors-18-02203-t001_Table 1\n\n###### \n\nComparison of recent studies on HAR adopting supervised or semi-supervised approach. In most cases dataset is assumed to be fully labeled, while only some studies explored ways to reduce the burden of data annotation. Target Activities (Ly = Lying, Si = Sitting, St = Standing, Wa = Walking, WU = Walking Upstairs, WD = Walking Downstairs, Ru = Running, Be = Bend, Fa = Fall, Da = Dancing, Cy = Cycling, Tr = Transportation, StS = Sit-to-Stand, StL = Sit-to-Lie, LtS = Lie-to-Stand, Tu = Turning, ShT = Sharp Turning).\n\n  Ref                                                    Sensors                    Target activities                                   Labeling                      Accuracy\n  ------------------------------------------------------ -------------------------- --------------------------------------------------- ----------------------------- ----------\n  Stikic et al. 2011 \\[[@B12-sensors-18-02203]\\]         accelerometer              Si, St, Wa, Tr, some ADLs                           Partial (label propagation)   76%\n  Siirtola et al. 2012 \\[[@B15-sensors-18-02203]\\]       accelerometer              Idle (Si/St), Wa, Ru, Cy, Tr                        Full                          95%\n  Anguita et al. 2013 \\[[@B31-sensors-18-02203]\\]        accelerometer              Si, St, Wa, WU, WD                                  Full                          89%\n  Pei et al. 2013 \\[[@B32-sensors-18-02203]\\]            accelerometer, gyroscope   Si, St, Wa, Wa (fast), Tu, ShT                      Full                          92%\n  Bayat et al. 2014 \\[[@B33-sensors-18-02203]\\]          accelerometer              Wa (fast), Wa (slow), Ru WD, WU, Da                 Full                          91%\n  Cleland et al. 2014 \\[[@B10-sensors-18-02203]\\]        accelerometer              St, Wa, Ru, Tr (bus)                                Partial (user prompt)         85%\n  Bhattacharya et al. 2014 \\[[@B34-sensors-18-02203]\\]   accelerometer, gyroscope   Idle (Si/St), Wa, Tr (bus), Tr (tram), Tr (train)   Partial                       79%\n  Reyes-Ortiz et al. 2016 \\[[@B35-sensors-18-02203]\\]    accelerometer, gyroscope   Ly, Si, St, Wa, WU, WD, StS, StL, LtS               Full                          96%\n  Ronao et al. 2016 \\[[@B24-sensors-18-02203]\\]          accelerometer, gyroscope   Ly, Si, St, Wa, WU, WD                              Full                          94%\n  Hong et al. 2016 \\[[@B36-sensors-18-02203]\\]           accelerometer              Ly, Si, St, Wa, Cy, Be, Fa                          Partial (label propagation)   83%\n  Hassan et al. 2018 \\[[@B26-sensors-18-02203]\\]         accelerometer, gyroscope   Ly, Si, St, Wa, WU, WD, StS, StL, LtS               Full                          89%\n  Cao et al. 2018 \\[[@B37-sensors-18-02203]\\]            accelerometer, gyrsocope   Si, St, Wa, WU, WD                                  Full                          94%\n  San-Segundo et al. 2018 \\[[@B38-sensors-18-02203]\\]    accelerometer              Si, St, Wa, WU, WD, Cy                              Full                          91%\n\nsensors-18-02203-t002_Table 2\n\n###### \n\nThe set of features used in the experiment.\n\n  Domain      Signal                                    Features\n  ----------- ----------------------------------------- -----------------------------------------------------\n  Time        3D Magnitude of Acceleration              Mean, Variance, Min, Max, Range, Skewness, Kurtosis\n  Time        *x*-, *y*- and *z*-axes of Acceleration   Absolute value of Mean, Variance, Range\n  Frequency   3D Magnitude of Acceleration              Number of peaks in PSD, Location of highest peak\n\nsensors-18-02203-t003_Table 3\n\n###### \n\nStructure of the dataset. For each data point, the first column contains the manually annotated labels while the second column contains the label generated automatically.\n\n  Label            Weak Label       Feature 1     \u22ef   Feature n\n  ---------------- ---------------- ------------- --- -------------\n  TRANSPORTATION   SITTING          $f_{1}^{1}$   \u22ef   $f_{n}^{1}$\n  TRANSPORTATION   TRANSPORTATION   $f_{1}^{2}$   \u22ef   $f_{n}^{2}$\n  \u22ef                \u22ef                \u22ef             \u22ef   \n  WALKING          WALKING          $f_{1}^{m}$   \u22ef   $f_{n}^{m}$\n\nsensors-18-02203-t004_Table 4\n\n###### \n\nNumber of samples with manual labeling composing the ground truth.\n\n  Class            Samples      Hours:Minutes:Seconds\n  ---------------- ------------ -----------------------\n  Sitting          395004       56:01:51\n  Walking          169908       26:09:26\n  Running          8250         01:20:09\n  Cycling          19132        02:40:35\n  Transportation   75244        11:47:21\n  **Total**        **667538**   **97:59:22**\n\nsensors-18-02203-t005_Table 5\n\n###### \n\nMean values for precision, recall and f-score obtained using 10-fold cross validation for all classifiers.\n\n  Algorithm             Precision   Recall   F-Score\n  --------------------- ----------- -------- ---------\n  Nearest Centroid      0.6816      0.6334   0.6418\n  DT                    0.8249      0.7878   0.7979\n  Random Forests        0.8666      0.8299   0.8394\n  kNN                   0.8355      0.8079   0.81405\n  Multi-class SVM       0.7630      0.7414   0.7424\n  NN 18 \u00d7 12 \u00d7 6        0.8394      0.8127   0.8188\n  NN 18 \u00d7 36 \u00d7 12 \u00d7 6   0.8585      0.8345   0.8410\n\nsensors-18-02203-t006_Table 6\n\n###### \n\nConfusion Matrix obtained using the NN that measured the highest f-score.\n\n                       Sitting       Walking       Running       Cycling       Transportation\n  -------------------- ------------- ------------- ------------- ------------- ----------------\n  **Sitting**          **0.88462**   0.03365       0.00481       0.0625        0.01442\n  **Walking**          0.03967       **0.79332**   0.04802       0.09395       0.02505\n  **Running**          0.00785       0.02356       **0.93979**   0.00524       0.02356\n  **Cycling**          0.01734       0.07514       0.01156       **0.85549**   0.04046\n  **Transportation**   0.02588       0.03512       0.01109       0.14787       **0.78004**\n"}
+{"text": "All data are within the paper.\n\nIntroduction {#sec001}\n============\n\nFilamentous fungi can produce a broad range of (often) toxic secondary metabolites, called mycotoxins. Already at concentrations in the lower micromolar range these mycotoxins, can pose a potential health risk to humans and animals \\[[@pone.0197406.ref001]\\].\n\nOne group of mycotoxins which has drawn more and more attention over the last few years, is summarized as \"emerging\" mycotoxins. This group of toxins includes compounds which (not yet) are neither legally regulated nor routinely analyzed. In general, when talking about \"emerging\" mycotoxins, this relates mainly to secondary fungal metabolites produced by various *Fusarium* species, like beauvericin (BEA), moniliformin (MON), fusaproliferin (FUS) and enniatins (ENNs) \\[[@pone.0197406.ref001], [@pone.0197406.ref002]\\].\n\nThe most prevalent toxins of the \"emerging\" mycotoxins are the ENNs. Up to now, 29 structural analogues, which are mainly produced by *Fusarium* species, are described in literature. Regarding their structural properties, ENNs are cyclic hexadepsipeptides, which are produced non-ribosomally via the enniatin synthetase \\[[@pone.0197406.ref003]\\]. They consist of alternating [d]{.smallcaps}-2-hydroxyisovaleric acids and *N*-methyl-[l]{.smallcaps}-amino acids. Peptide bonds and intramolecular ester (lactone) bonds link the subunits forming an 18-membered ring \\[[@pone.0197406.ref004]\\]. In particular, enniatin B (ENN B) and enniatin B1 (ENN B1) are of central interest as they are the most prevalent compounds of this group of mycotoxins and occur in relatively high concentrations in *Fusarium* contaminated food and feed. For example, in grain concentrations up to 5.8 mg/kg ENN B in Norwegian wheat and 18.3 mg/kg ENN B1 in Finnish spring wheat were found \\[[@pone.0197406.ref005]\\]. Thus, this study focusses on the compounds ENN B and ENN B1 ([Fig 1](#pone.0197406.g001){ref-type=\"fig\"}), as representatives of the group of enniatins. They only differ in one methyl group as highlighted in [Fig 1](#pone.0197406.g001){ref-type=\"fig\"}.\n\n![Structures of ENN B and ENN B1.](pone.0197406.g001){#pone.0197406.g001}\n\nRemarkably, due to their pharmacologic properties a mixture of the enniatins ENN B, B1, A and A1, found application as local antibiotic (marketed under the name \"fusafungine\") to treat upper respiratory tract infections \\[[@pone.0197406.ref006]\\]. Though, because of reported severe allergic reactions to these remedies, the admission to the European market is currently under revision \\[[@pone.0197406.ref007]\\].\n\nSystemic exposure of humans to ENN B and ENN B1 is described as these mycotoxins are detected in blood, urine and breast milk \\[[@pone.0197406.ref008]--[@pone.0197406.ref010]\\]. In various *in vivo*, *in vitro* as well as *ex vivo-*studies, it was shown, that cyclic hexadepsipeptides, like ENNs, are able to cross barriers of mammalian organisms very fast and to a high extent. Therefore, they can easily reach systemic circulation. ENNs are able to cross an *ex vivo* human skin barrier model to a great amount with ENN B showing the highest permeation (k~p,v~ = 9.44 \u00d7 10^\u22126^ cm/h) \\[[@pone.0197406.ref011]\\]. Taevenier et al. \\[[@pone.0197406.ref012]\\] showed that ENNs could penetrate porcine buccal mucosa with steady-state plasma concentrations up to 1.3 mg/L \\[[@pone.0197406.ref012]\\]. In a Caco-2 barrier experiment the absorption after 4 h exposure was \\> 65% for ENN B and ENN B1, also the duodenal bioavailability *in vitro* (Caco-2 model) was \\> 50% after 48 h for both toxins \\[[@pone.0197406.ref013]\\].\n\nSo far, one single study on the permeation of ENNs and BEA across the blood-brain barrier (BBB) in mice was performed. This study was conducted over the course of 100 minutes. The results of this short time *in vivo-*mouse study, indicate a high and rapid influx into the brain with distribution to the brain parenchyma \\[[@pone.0197406.ref014]\\].Our study aims to add more insights to the kinetics of the transfer of ENNs with a longer exposure time (48 h) and eight sampling times over the course of the experiment. Thus, a better kinetic profile can be shown.\n\nIt is known, that several other fungal toxins, like ochratoxin A, T-2 and HT-2 toxins or deoxinivalenol, are able to cross the BBB and affect the viability and functions of brain cells, including astrocytes, microglia and endothelial cells \\[[@pone.0197406.ref015]--[@pone.0197406.ref018]\\]. Considered that ENNs reach systemic circulation, the permeation properties through the BBB are of crucial importance. The passage of ENNs through the BBB could cause neurotoxic effects in cells of the central nervous system.\n\nThe BBB is a safety feature protecting highly sensitive neuronal cells of the brain from adverse effects. In general, the BBB consists of three different cell types, which together form a neurovascular unit: Endothelial cells, astrocytes and pericytes \\[[@pone.0197406.ref019]\\]. The BBB is a highly selective interface which limits the unhindered interaction of substances (endogenous or exogenous) from systemic circulation with the brain. To estimate the probability of a compound to reach the brain, different approaches are applicable \\[[@pone.0197406.ref020], [@pone.0197406.ref021]\\]. In this study, a well-established cell culture model based on primary porcine brain capillary endothelial cells (PBCEC) was used. This model combines many advantages compared to other systems. In contrast to *in vivo-* and *ex vivo-*experiments, there is no need for invasive procedures and interindividual differences can be reduced. Compared to other *in vitro*-models relying on continuous cell lines, the handling and costs might be higher, but morphological features of the brain endothelium are much better reflected in primary cells \\[[@pone.0197406.ref020]\\]. Even in monoculture, with PBCECs, high TEER (transendothelial electrical resistance) values of 1000 \u2126\u00b7cm^2^ and higher can be reached. Such TEER values are often not attained by other *in vitro-*cell culture models that are applied for transport studies. This property of the PBCEC is close to the *in vivo*-situation and ensures the tightness and reliability of a barrier system to study transport characteristics \\[[@pone.0197406.ref016], [@pone.0197406.ref020], [@pone.0197406.ref022]\\]. In order to assess the possibility of ENNs reaching the brain, the transport kinetics of ENN B and B1 were evaluated using this *in vitro*-cell culture system based on PBCEC mimicking the BBB.\n\nAs mentioned above, the BBB consists of different cell types. Thus, as soon as ENNs are able to cross the BBB it becomes relevant to investigate possible (adverse) effects of ENNs on different cells of the BBB. For that matter an astrocytoma cell line (CCF-STTG1 cells) and, besides the primary cells (PBCEC), which are not suited for all of the applied test systems, human brain microvascular endothelial cells (HBMEC) were included in the evaluations. Different endpoints concerning cell death like cytotoxicity, caspase-3 activity and LDH leakage were considered.\n\nMaterial and methods {#sec002}\n====================\n\nChemicals and reagents {#sec003}\n----------------------\n\nAll chemicals for ENN quantitation were purchased at VWR International GmbH (Darmstadt, Germany), Gr\u00fcssing GmbH Analytica (Filsum, Germany) and Sigma-Aldrich GmbH (Steinheim, Germany). Cell culture media and supplements were purchased at Biochrom AG (Berlin, Germany) and PAN-Biotech GmbH (Aidenbach, Germany). Hydrocortisone, bicinchoninic acid (BCA) and bovine serum albumin (BSA) were obtained from Sigma-Aldrich (Steinheim, Germany). Purified water was retrieved from a Purelab Flex 2 purification system (Veolia Water Technologies, Celle, Germany). ENN B (E4511) and ENN B1 (E5286) with purities \u2265 95% were purchased from Sigma-Aldrich GmbH (Steinheim, Germany). For experiments, separate stock solutions of 1 mM ENN B and B1 in acetonitrile (ACN) were prepared and stored at --20\u00b0C.\n\nEffects of ENN B and ENN B1 on cell death {#sec004}\n-----------------------------------------\n\n### Viability assay in cell lines (HBMEC and CCF-STTG1) {#sec005}\n\nTo test the viability and to define the working range for later transport studies ENN B/ENN B1 stock solutions in ACN were diluted to the respective concentrations with serum free culture medium (1% ACN). An astrocytoma cell line, CCF-STTG1 (ATCC, Manassas, USA), and HBMEC cell line (human brain microvascular endothelial cells, kindly provided by Prof. Dr. Karch, Institute of Hygiene, University of Muenster (WWU Muenster)) were cultivated in RPMI 1640 medium supplemented with 100 U/mL penicillin, 100 \u03bcg/mL streptomycin, 2 mM [l]{.smallcaps}-glutamine and 10% (v/v) fetal calf serum (FCS) (medium for HBMEC additionally supplemented with 1 mM Na-pyruvate) using standardized culture conditions (37\u00b0C, 5% CO~2~, saturated humidified atmosphere). Culture medium was changed every 2--3 d, and the cells were subcultivated after trypsination when they reached a microscopic confluence of approximately 80%. For the evaluation of cytotoxic effects of ENN B and B1 in CCF-STTG1 and HBMEC, Cell Counting Kit-8 (CCK-8, Dojindo Laboratories, Tokyo Japan) was applied. The assay was performed according to the manufacturer's instructions with slight modifications.\n\nBriefly, the cells were seeded in 96-well tissue culture plates in a density of 1 \u00d7 10^4^ cells/well (CCF-STTG1) and 4 \u00d7 10^3^ cells/well (HBMEC), respectively. The cells were allowed to grow for 24 h, then the culture medium was replaced by serum free medium (100 \u03bcL/well), and the cells were cultivated for another 24 h (37\u00b0C, 5% CO~2~, saturated humidified atmosphere) before treatment. The cells were treated with ENN B and ENN B1 in a concentration range from 0.1 \u03bcM to 10 \u03bcM and were incubated for 48 h. After toxin exposure, the toxin solution was exchanged with a dye solution WST-8 (CCK-8 solution, water soluble tetrazolium salt-8, 2-(2-methoxy-4-nitrophenyl)-3-)4-nitrophenyl)-5-(2,4-disulfophenyl)-2*H*-tetrazolium monosodium salt), 1:10 diluted with fresh serum free medium, which was added to the cells followed by incubation for 60 min at 37\u00b0C. WST-8 dye is reduced to a water-soluble formazan by cellular dehydrogenases of viable cells, which increases the absorbance at \u03bb = 457 nm and was measured with an Infinite M200 PRO microplate reader with Tecan i-control software version 1.7.1.12 (Tecan, Crailsheim, Germany). The amount of formazan generated by the activity of dehydrogenases in cells is directly proportional to the number of viable cells per well. The results are normalized to a solvent treated negative control (1% ACN).\n\n### Viability assay in primary porcine brain capillary endothelial cells (PBCEC) {#sec006}\n\nPrimary porcine brain capillary endothelial cells (PBCEC) were kindly provided by the group of Prof. Dr. Langer, Institute of Pharmaceutical Technology, WWU M\u00fcnster. Cellular viability was tested after treatment for 48 h using the CCK-8 assay, like applied for the cell lines but with minor modifications. Based on earlier protocols \\[[@pone.0197406.ref016], [@pone.0197406.ref022]\\] for the cultivation of PBCEC, the cryopreserved PBCEC were thawed at 37\u00b0C in a water bath, gently resuspended in culture medium (Medium 199 Earle's with 100 U/mL penicillin, 100 \u03bcg/mL streptomycin, 100 \u03bcg/mL gentamycin, 4.1 mM [l]{.smallcaps}-glutamine and 10% fetal calf serum) and centrifugated at 220 \u00d7 *g* at 20\u00b0C for 10 min. After removal of the supernatant, fresh complete medium was added and the cell pellet was resuspended twice and diluted to desired cell density. The cells were seeded on rat tail collagen coated 96-well tissue culture plates with 100 \u03bcL of the cell suspension per well. PBCEC were allowed to grow for 48 h at 37\u00b0C, 5% CO~2~ under saturated humidified conditions. Complete medium was exchanged for serum free medium (DMEM/Ham's F-12 1:1 with 100 U/mL penicillin, 100 \u03bcg/mL streptomycin, 100 \u03bcg/mL gentamycin, 0.7 mM [l]{.smallcaps}-glutamine and 550 nM hydrocortisone (Sigma-Aldrich, Steinheim, Germany)). On day 5 in culture (96 h after seeding) half of the medium (50 \u03bcL) was removed and exchanged with 50 \u03bcL double concentrated ENN solution to reach the desired final concentrations in a concentration range from 0.1 \u03bcM to 10 \u03bcM. The cells were treated with the test substances in concentrations from 0.1 to 10 \u03bcM for 48 h. Due to the sensitivity of the primary cells, after completion of 48 h, the incubation medium was not exchanged but each well was supplemented with 10 \u03bcL of the CCK-8 solution and incubated for 70 min at 37\u00b0C, 5% CO~2~ in a saturated humidified atmosphere. Then the absorbance of the generated formazan dye was measured at 457 nm with 650 nm as reference wavelength. After subtracting the reference absorption as well as a blank absorption without cells, the viabilities were normalized to a solvent treated negative control (1% ACN).\n\n### Apoptosis: Caspase-3 activity {#sec007}\n\nFor studying caspase-3 activity CCF-STTG1 cells were seeded in 12 well tissue culture plates in a cell density of 2.25 \u00d7 10^5^ cells/well. The cells were grown for 24 h until a confluence of at least 70% was reached, then the medium was changed to serum free medium (RPMI 1640 medium supplemented with 100 U/mL penicillin, 100 \u03bcg/mL streptomycin and 2 mM [l]{.smallcaps}-glutamine). Concentrations from 0.1 \u03bcM to 2.5 \u03bcM were chosen to prevent strong cytotoxic effects, which might interfere with the endpoint of interest. 24 h after adaption to serum free conditions, cells were incubated with ENN B and ENN B1. Caspase-3 activity was determined in lysates. To prepare the lysates, the incubation medium was removed (used in LDH leakage assay) and the cell layer was rinsed with 200 \u03bcL phosphate-buffered saline without magnesium and calcium (PBS^-/-^) twice. Followed by addition of 120 \u03bcL lysis buffer (ice cold) (10 mM TRIS, 100 mM NaCl, 1 mM EDTA and 1% Triton-X-100 in water) and lysis for 15 min on ice. Lysed cells were scraped off, transferred to microreaction tubes and centrifuged (10 000 \u00d7 *g*, 4\u00b0C, 10 min). 30 \u03bcL of the supernatant were subsequently used for the measurement. 33 \u03bcL reaction solution were added to 30 \u03bcL of the supernatant and incubated for 1 and 2 h at 37\u00b0C, 5% CO~2~ and saturated humidified atmosphere and then measured at \u03bb~ex~ = 405 nm and \u03bb~em~ = 520 nm (Infinite M200 PRO microplate reader). Reaction solution comprised 2 parts reaction buffer (50 mM PIPES, 12.7 mM EDTA, 8.1 mM CHAPS ad 100 mL with water, pH 7.4), 3 parts water, 0.5 parts caspase-substrate (1 mM Ac-DEVD-AFC, Sigma-Aldrich GmbH, Seelze, Germany) and 0.05 parts DTT solution (1 M). The results were calculated as \u03bcmol AFC/\u03bcg protein and normalized to the solvent treated control (ACN 0.25%).\n\nFor quantitation of the total protein content a bicinchoninic acid (BCA) assay kit (Sigma-Aldrich GmbH, Seelze, Germany) was applied. Bovine serum albumin (BSA) served as standard for external calibration. 15 \u03bcL of cell lysates were diluted with 200 \u03bcL BCA reagent (BCA solution/4% copper sulfate, 50+1, v/v) in a 96-well tissue culture plate and incubated for 30 min at 37\u00b0C. After that the absorbance was measured at 560 nm with a Infinite M200 PRO microplate reader (Tecan, Crailsheim, Germany).\n\n### Necrotic cell death: Lactate dehydrogenase (LDH) leakage {#sec008}\n\nFor measurement of lactate dehydrogenase (LDH) leakage, cell lysates and treatment solutions (serum free medium with respective ENN B/B1 content) after incubation for 48 h were used. The cell lysates were prepared as described for caspase-3 activation assay. To determine the LDH activity, 15 \u03bcL of the lysate and 40 \u03bcL of the incubation medium were each added to 200 \u03bcL substrate buffer (100 mM HEPES, 10 mM sodium pyruvate and 0.21 mM NADH, adjusted to pH 7.0), incubated over the course of 30 min at 37\u00b0C and absorption was measured every 2 min at 355 nm in a Infinite M200 PRO microplate reader (Tecan, Crailsheim, Germany).\n\nTransfer studies {#sec009}\n----------------\n\nIn accordance with the thawing and seeding procedure for PBCEC, described for the viability test, 500 \u03bcL PBCEC suspension (2.5 \u00d7 10^5^ cells/well) in culture medium were seeded on rat tail collagen coated 12-well Transwell^\u00ae^ microporous polycarbonate filter inserts (Corning, Wiesbaden, Germany; growth area 1.12 cm^2^, pore size 0.4 \u03bcm) in the upper (apical) compartment of the filter system. The lower (basolateral) compartment was filled with 1500 \u03bcL complete medium. This two-compartment system was used as model system for the BBB with the apical compartment reflecting the \"blood\"-side and the basolateral compartment serving as the \"brain\"-side. After 48 h, the complete medium was replaced by serum free medium. Additionally, 550 nM hydrocortisone (Sigma-Aldrich, Steinheim, Germany) was supplemented to urge differentiation of cells. After another 48 h, experiments studying barrier integrity and permeability were performed.\n\nBefore starting each experiment, the TEER values and the electrical capacitance of all cell layers were measured via cellular impedance spectroscopy (cellZscope^\u00ae^, nanoAnalytics, M\u00fcnster, Germany). Only Transwell^\u00ae^ filters with cell layers possessing TEER values of more than 300 \u2126\u00b7cm^2^ and electrical capacitances (C~Cl~) of 0.4 to 0.6 \u03bcF/cm^2^ were used for transfer studies. TEER values relate to the constitution of the tight junctions limiting paracellular transfer of compounds. In addition to the TEER, electrical capacitances (C~Cl~) were monitored for 48 h. Cellular phospholipid bilayers act like an electrical capacitor and also contribute to the cellular impedance. Changes of the C~Cl~ value give information about the integrity of the cell monolayer \\[[@pone.0197406.ref023]\\]. In this experiment, a parallel set up was applied. For each treatment, one triplicate of filters was monitored via cellular impedance spectroscopy over the course of 48 h using a cellZscope^\u00ae^ device--here, the basolateral compartment was prepared with 1650 \u03bcL serum free medium and the volume in the apical compartment was filled to 760 \u03bcL (ratio 1:2.17). Another triplicate for the same incubation conditions was kept in Transwell^\u00ae^ filter cell culture multiple well plates (12 well format) over the course of 48 h. In the multiple well plates the ratio between the apical and the basolateral compartment was 1:3, therefore the starting volume in the apical compartment is 500 \u03bcL and 1500 \u03bcL in the basolateral compartment. Here, samples for quantitation of transfer of enniatins across the barrier were drawn after 1, 2.5, 6.5, 18, 24, 28, 42 and 48 hours in this given ratio i.e. 30 \u03bcL apical and 90 \u03bcL basolateral. After 48 h, the final TEER values of the filters from the multiple well plate were measured to ensure the stability of the cell layer throughout the experiment. To double check the integrity of the barrier, lucifer yellow (LY, 50 \u03bcM), as a paracellular transfer marker, was applied. The amount of LY found in the basolateral compartment after 1 h was measured at \u03bb~ex~ = 430 nm and \u03bb ~em~ = 540 nm.\n\nAfter the completion of the experiment, the polycarbonate filter membranes were extracted. The filters were cut out and extracted with ACN/H~2~O (8+2, v/v) via ultrasonication for 1 h. The extract was evaporated (vacuum concentrator BA-VC-200 H, Bachofer GmbH + Co KG, Reutlingen, Germany) and reconstituted in 200 \u03bcL of ACN/H~2~O (8+2, v/v). The obtained samples and filter extracts were analyzed and quantitated with high performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS).\n\nQuantitation of ENN B and ENN B1 {#sec010}\n--------------------------------\n\nThe quantitation of ENN B and ENN B1 by HPLC-MS/MS was carried out using an 1100 series (Agilent Technologies, Santa Clara, USA) high performance liquid chromatography (HPLC) system and API 3200 (AB SCIEX Germany GmbH, Darmstadt, Germany) mass spectrometer. Both devices were operated with Analyst Version 1.6.2 software (AB SCIEX Germany GmbH, Darmstadt, Germany). A Synergi 4 \u03bcm Fusion-RP 80 \u00c5 50 \u00d7 2.0 mm (Phenomenex) equipped with a Synergy 4 \u03bcm Fusion-RP 4 \u00d7 2.0 mm (Phenomenex) precolumn was used and maintained at 40\u00b0C. A binary gradient consisting of ACN (A) and water (B) (both containing 0.1% formic acid) was applied. The concentration of the analytes was determined directly in cell culture medium and quantitated using a matrix-matched external calibration in serum free cell culture medium (0.1% hydrocortisone), incubated with PBCEC for 48 h. 10 \u03bcL of sample solution were injected. With a flow rate of 500 \u03bcL/min the starting condition was 20% A, which was held for 2 min. The percentage of A was increased to 100% within 3 min and kept constant for further 1.5 min. Starting at 6.5 min the gradient was decreased from 100% A to the starting conditions (20% A) and held for 1.5 min. The parameters of the mass spectrometer (API 3200) were set to: Curtain gas 30 psi N~2~, source temperature 400\u00b0C, nebulizer gas (GS1) 35 psi N~2~ and heater gas (GS2) and 45 psi respectively, and a dwell time of 25 ms per MRM transition was used. Electrospray ionization in positive ionization mode was applied with the ion spray voltage set to 5500 V. The two multiple reaction monitoring (MRM) transitions, which possessed the best signal-to-noise ratios were used for analysis of the two compounds (ENN B and ENN B1), the one with the highest signal intensity was chosen as quantifier. [Table 1](#pone.0197406.t001){ref-type=\"table\"} lists the corresponding HPLC-MS/MS parameters for both ENNs.\n\n10.1371/journal.pone.0197406.t001\n\n###### Parameters for multiple reaction monitoring (MRM) for the two analytes.\n\n![](pone.0197406.t001){#pone.0197406.t001g}\n\n  -------------------------------------------------------------------------------------------------------------\n  Analyte   Q1\u00a0mass \\[*m/z*\\]    Q3\u00a0mass \\[*m/z*\\]                            DP \\[V\\]   CE \\[V\\]   CXP \\[V\\]\n  --------- -------------------- -------------------------------------------- ---------- ---------- -----------\n  ENN\u00a0B     640.5 \\[M\u00a0+\u00a0H\\]^+^   214[^a^](#t001fn001){ref-type=\"table-fn\"}\\   71\\        55\\        5\\\n                                 196[^b^](#t001fn002){ref-type=\"table-fn\"}    71         31         5\n\n  ENN\u00a0B1    654.4 \\[M\u00a0+\u00a0H\\]^+^   210[^a^](#t001fn001){ref-type=\"table-fn\"}\\   58\\        30\\        17\\\n                                 196[^b^](#t001fn002){ref-type=\"table-fn\"}    58         33         16\n  -------------------------------------------------------------------------------------------------------------\n\n^a^ quantifier\n\n^b^ qualifier\n\nPermeability calculations {#sec011}\n-------------------------\n\nTime independent permeability coefficients were derived, to enable the comparison of permeation of different compounds through the BBB. The permeability coefficients *P* were calculated according to [Eq 1](#pone.0197406.e001){ref-type=\"disp-formula\"} \\[[@pone.0197406.ref016]\\]: $$P\\ \\left\\lbrack \\frac{cm}{s} \\right\\rbrack = \\frac{c_{bas}\\left\\lbrack \\% \\right\\rbrack}{c_{0h,\\ ap}\\left\\lbrack \\% \\right\\rbrack} \\bullet \\frac{V_{ap}\\left\\lbrack {cm^{3}} \\right\\rbrack}{A\\ \\left\\lbrack {cm^{2}} \\right\\rbrack \\bullet t\\ \\left\\lbrack s \\right\\rbrack}$$ *c*~*bas*~ represents the amount of test substance, which is found in the basolateral compartment at a distinct timepoint *t*. *V*~*ap*~ with *c*~*0h*,\\ *ap*~ describe the initial volume of the apical (\"blood\") compartment and applied concentration in the apical compartment at the start of the experiment, respectively. The filter surface of the Transwell^\u00ae^ filter insert is presented as *A*. To exclude hindrance of passage by the polycarbonate filter membrane, a cell free experiment was carried out. ENN B and ENN B1 were applied at 10 \u03bcM concentrations in the apical compartment of a rat tail collagen coated filter insert. *P*~*e*,~ the permeability of the mycotoxin across the cell monolayer, corrected for the filter permeability, was calculated according to [Eq 2](#pone.0197406.e002){ref-type=\"disp-formula\"}.\n\n![](pone.0197406.e002.jpg){#pone.0197406.e002g}\n\n1\n\nP\n\ne\n\n\\[\n\ncm\n\ns\n\n\\]\n\n=\n\n1\n\nP\n\nc\n\n\\+\n\nf\n\n\\[\n\ncm\n\ns\n\n\\]\n\n\u2011\n\n1\n\nP\n\nf\n\n\\[\n\ncm\n\ns\n\n\\]\n\nThe permeability of the mycotoxins in the cell free experiment *P*~*f*~ is included in the calculation of the permeability coefficient of the study (*P*~*c+f*~) as shown in [Eq 2](#pone.0197406.e002){ref-type=\"disp-formula\"}.\n\nStatistical analysis {#sec012}\n--------------------\n\nCellular viability studies were performed as a minimum in triplicates with cells from at least three independent passages/preparations (*n* \u2265 9, technical replicates). Caspase-3 assay and LDH leakage assay were performed in duplicates from at least three independent experiments (*n* \u2265 6, technical replicates). The data are presented as the mean \u00b1 standard deviation (SD). The significance indicated refers to the significance level as compared to the solvent treated control calculated with one-way ANOVA and Dunnett's multiple-comparison test via OriginPro 2016G (64-bit) Sr2 b9.3.2.303 (Significance levels: \\* low significant (p \u2264 0.05); \\*\\* medium significant (p \u2264 0.01); \\*\\*\\*, highly significant (p \u2264 0.001)). For transfer experiments PBCEC from at least three different preparations were used and the transport experiments were performed in triplicates (*n* = 9). The data are presented as the mean \u00b1 standard deviation (SD).\n\nResults and discussion {#sec013}\n======================\n\nIn this study, effects of ENN B and ENN B1 on cell death in different cells types, involved in the formation of the BBB, were evaluated. In a further step, transport properties of ENNs across a primary porcine BBB model were investigated.\n\nEffects of ENN B and ENN B1 on cell death {#sec014}\n-----------------------------------------\n\n### Cell viability {#sec015}\n\nTo determine and compare possible effects of ENN B and ENN B1 on cellular viability in different cells of the BBB, a cytotoxicity assay (CCK-8 assay), was carried out using primary porcine brain capillary endothelial cells (PBCEC), human brain microvascular capillary cells (HBMEC, passages 21, 22, 35 and 44) and CCF-STTG1 (passages 58 -- 61), an astrocytoma cell line. The CCK-8 assay is based on the reduction of a water-soluble tetrazolium salt-8 (WST-8) dye by cellular dehydrogenases of viable cells to the corresponding formazan which is directly proportional to the number of metabolically active cells. Cells were incubated with the test compounds in concentrations ranging from 0.1 \u03bcM to 10 \u03bcM for 48 h. The concentrations in the experimental setup were chosen to cover the broad variations of contamination levels with ENNs found in food and feed. Annual contamination levels vary from \\~ 10 \u03bcg/kg up to concentrations of 5.8 mg/kg ENN B and 18.3 mg/kg ENN B1 \\[[@pone.0197406.ref005]\\]. In a pig-feeding study, after oral intake of 0.05 mg/kg b.w. ENNs, reflecting a 1 mg/kg contamination of feed, concentrations in pig plasma were 0.073 \u03bcg/mL (ENN B, 0.114 \u03bcM) and 0.035 \u03bcg/mL (ENN B1, 0.054 \u03bcM) \\[[@pone.0197406.ref024]\\]. In a worst case scenario with 5.8 mg/kg ENN B \\[[@pone.0197406.ref005]\\] this would correlate to theoretical plasma levels of 0.423 \u03bcg/mL (0.662 \u03bcM) and 18.3 mg/kg ENN B1 \\[[@pone.0197406.ref005]\\] would result in 0.641 \u03bcg/mL (0.980 \u03bcM), respectively. In a worst case acute intoxication our chosen concentration range may be relevant. In [Fig 2](#pone.0197406.g002){ref-type=\"fig\"} the results of the viability assay are summarized.\n\n![**Viability of PBCEC (*n* \u2265 9) (A), HBMEC (*n* \u2265 9) (B) and CCF-STTG1 cells (*n* \u2265 9) (C) after treatment with ENN B and ENN B1 for 48 h.** As positive control T-2 toxin (T2, 10 \u03bcM) was applied. Data are presented as the mean \u00b1 standard deviation (SD). The significance indicated refers to the significance level as compared to the solvent treated control (1% ACN) calculated with one-way ANOVA and Dunnett's multiple-comparison (Significance levels: \\* low significant (p \u2264 0.05); \\*\\* moderately significant (p \u2264 0.01); \\*\\*\\* highly significant (p \u2264 0.001)).](pone.0197406.g002){#pone.0197406.g002}\n\nThe treatment of the porcine and human brain capillary endothelial cells lead to different results. While in the continuous cell line, HBMEC, no effect on viability compared to the solvent control, was observed at the concentration range from 0.1 \u03bcM to 10 \u03bcM ([Fig 2B](#pone.0197406.g002){ref-type=\"fig\"}), viability of the primary cells, PBCEC ([Fig 2A](#pone.0197406.g002){ref-type=\"fig\"}), was decreased after treatment with ENN B and ENN B1. ENN B was statistically highly significantly cytotoxic at concentrations above 5 \u03bcM with a reduction of cellular viability to a minimum of 70%. The exposure of PBCEC to ENN B1 resulted in a slightly higher cytotoxic effect. Cell viability was highly significantly inhibited at concentrations of 2.5 \u03bcM and higher, with reduction to 64% relative viability at 10 \u03bcM.\n\nThe effects on viability caused by ENN B and ENN B1 in endothelial cells of the brain, were less pronounced than the effects observed in the astrocytoma cell line, CCF-STTG1 ([Fig 2C](#pone.0197406.g002){ref-type=\"fig\"}). In CCF-STTG1 cells, ENN B and ENN B1 caused considerably stronger effects on cellular viability. Here, ENN B presented highly significant cytotoxic effects at 5 \u03bcM and 10 \u03bcM. After incubation with ENN B1 weak but significant effects were observed already at 0.1 \u03bcM and 1 \u03bcM, strong and highly significant cytotoxic effects were detected at 2.5 \u03bcM and higher. From these results IC~50~ values for ENN B of 8.9 \u03bcM and ENN B1 of 4.4 \u03bcM in CCF-STTG1 cells were calculated.\n\nIn general, a higher cytotoxic potential was observed for ENN B1 in all three cell types. These results are in good accordance to most results found in literature, concerning the cytotoxicity of ENNs.\n\nW\u00e4tjen et al. \\[[@pone.0197406.ref025]\\] describe cytotoxic effects of ENNs in different tumor cell lines (H4IIE rat hepatoma, HepG2 human hepatoma and C6 rat glioma cells) applying the MTT-assay. In C6 glioma and HepG2 hepatoma cells IC~50~ values range between 2.5 and 10 \u03bcM, respectively. The strongest cytotoxic effect is evident in H4IIE cells with IC~50~ values between 1 and 1.5 \u03bcM \\[[@pone.0197406.ref019]\\]. Meca et al. \\[[@pone.0197406.ref026]\\] discuss that almost all ENNs they tested (A, A1, A2, B, B1, B2, B4 and J3) exert cytotoxic effects in HepG2, HT-29 (human colorectal adenocarcinoma cell line) and Caco-2 cells (human colorectal adenocarcinoma cell line) in a dose-dependent manner. After a 48 h incubation ENN B leads to an IC~50~ value of 2.8 \u03bcM and ENN B1 to 3.7 \u03bcM in HT-29 cells. For the other cell lines, HepG2 and Caco-2, treatment with ENN B1 lead to IC~50~ values of 8.5 \u03bcM and 11.5 \u03bcM, respectively, whereas no IC~50~ could be calculated for ENN B \\[[@pone.0197406.ref026]\\]. Ivanova et al. \\[[@pone.0197406.ref027]\\] describe IC~50~ values of ENN B and B1 after 24 h incubation (Alamar Blue^\u2122^ assay). For MRC-5 cells (human lung fibroblast cell line) and HepG2 cells, ENN B shows IC~50~ values of 9.8 \u03bcM and 435.9 \u03bcM, respectively. For ENN B1 IC~50~ values of 4.7 \u03bcM (MRC-5) and 36 \u03bcM (HepG2) are reported \\[[@pone.0197406.ref027]\\]. In general, ENN B1 exhibits higher cytotoxic properties. The comparison of the cytotoxic potential (MTT-assay) of a mixture of ENNs in 27 different cell models, after a longer exposure time (72 h), concludes that the IC~50~ for most tumor cell models is lower (\\<5 \u03bcM) than for normal cell lines (e.g. endothelial cell model (HUVEC)) \\[[@pone.0197406.ref028]\\]. This also supports the observation in our experiments, where CCF-STTG1 cells, an astrocytoma cell line, was considerably more sensitive to ENN B and B1 than normal cell types, like PBCEC or HBMEC. Tumor cells in general are faster dividing cells with higher proliferation than other cell types. Taking cell cycle arrests in the G0/G1 phase caused by ENNs into account \\[[@pone.0197406.ref028]\\], the assumption that ENNs might have an impact on apoptotic signaling pathways associated with cell proliferation can be made. This considered, it is likely that the impact of ENNs on proliferation and cellular viability is higher in tumor cell types. Therefore, the cytotoxicity observed in our study can be ranked as follows ENN B1 \\> ENN B and the sensitivity of cell types: CCF-STTG1 \\> PBCEC \\> HBMEC.\n\nAs the astrocytoma cell line (CCF-STTG1) was substantially more sensitive to the exposure to ENN B and ENN B1, further experiments concerning the type of cell death were carried out with these cells.\n\n### Apoptosis and necrosis: Caspase-3 activation and LDH leakage {#sec016}\n\nTo distinguish whether the cell death in CCF-STTG1 cells was of apoptotic or necrotic nature, the caspase-3 activation, as indicator for the programmed cell death, and LDH leakage, as marker for necrotic cell death, were considered. For evaluations concerning the programmed cell death, CCF-STTG1 cells were incubated with ENN B and ENN B1 for 48 h with concentrations from 0.1 \u03bcM to 2.5 \u03bcM, where first cytotoxic effects were observed. For ENN B a 2.7-fold, highly significant increase of caspase-3 activity compared to the solvent control was shown for the treatment with 2.5 \u03bcM ENN B for 48 h. Only minor differences in comparison to the solvent control were measured for ENN B1 with a maximum increase of caspase-3 activity to 160% at 1 \u03bcM ([Fig 3A](#pone.0197406.g003){ref-type=\"fig\"}).\n\n![**Caspase-3 activation in CCF-STTG1 after incubation with ENN B and ENN B1 for 48 h (*n* \u2265 6) (A). LDH leakage in CCF-STTG1 after incubation with ENN B and ENN B1 for 48 h (*n* \u2265 6) (B).** Positive controls are T-2 toxin (T2, 10 \u03bcM), camptothecin (C, 10 \u03bcM) and Triton X-100 (Tri, 0.1%). Data are presented as the mean \u00b1 standard deviation (SD). The significance indicated refers to the significance level as compared to the solvent treated control (0.25% ACN) calculated with one-way ANOVA and Dunnett's multiple-comparison (Significance levels: \\* low significant (p \u2264 0.05); \\*\\* moderately significant (p \u2264 0.01); \\*\\*\\* highly significant (p \u2264 0.001)).](pone.0197406.g003){#pone.0197406.g003}\n\nConcerning LDH release as a marker for necrotic cell death, slightly higher LDH leakage to the medium was observed for the cells treated with ENN B1 ([Fig 3B](#pone.0197406.g003){ref-type=\"fig\"}), but overall no significant increase of LDH release was detectable for ENN B and ENN B1.\n\nAs reported by W\u00e4tjen et al. \\[[@pone.0197406.ref025]\\] ENN B and ENN B1 induces apoptosis in hepatoma cells in low micromolar concentrations of 0.5 and 1 \u03bcM after 24 h, measured via the caspase-3/7 activation and nuclear fragmentation. In contrast to this, after 48 h incubation, our results suggest a stronger caspase-3 activation caused by ENN B. A significant increase of caspase-3/7 activity after treatment with both, ENN B and ENN B1, with a higher caspase-3 activity for cells treated with ENN B1 is described \\[[@pone.0197406.ref025]\\]. Dornetshuber et al. \\[[@pone.0197406.ref022]\\] also confirm no signs of LDH leakage caused by a mixture of ENNs, whereas already the minimally active concentration of 2.5 \u03bcM ENNs prompts cell cycle arrests and at higher concentration (10 \u03bcM ENNs) a high activation of caspase-3/7 is resulting \\[[@pone.0197406.ref028]\\].\n\nTo sum this up, no strong differences of the effects of ENN B and ENN B1 were expected because their chemical structures and properties are very similar. ENN B induces lower cytotoxicity but an induction of caspase-3 activity. This could indicate different modes of action of ENN B and ENN B1, where ENN B is rather initiating signalling pathways and does not necessarily exert high acute cytotoxicity, whereas ENN B1 is more cytotoxic and will earlier lead to necrotic cell death. But further investigations in terms of different time points, concentrations and considered endpoints are required to investigate differences in ENNs toxicodynamics.\n\nTransfer of ENN B and ENN B1 across a porcine BBB model {#sec017}\n-------------------------------------------------------\n\nTo evaluate whether ENN B and ENN B1 could cross the BBB and subsequently cause neurotoxic effects in the brain parenchyma, an *in vitro-*model using PBCEC mimicking the BBB was applied.\n\nFor passive transport (apical \u2192 basolateral) studies PBCEC cells were incubated with 1 \u03bcM concentrations (= 0.5 nmol) of either ENN B or ENN B1 in the apical (upper) compartment. 1 \u03bcM was chosen, because this concentration did not show any cytotoxic effects in PBCEC over the course of 48 h ([Fig 2A](#pone.0197406.g002){ref-type=\"fig\"}). Furthermore, it was shown that this concentration does not lead to a breakdown of the TEER value ([Fig 4A](#pone.0197406.g004){ref-type=\"fig\"}) and does not impair PBCEC confluency of the monolayer of the barrier-forming endothelial cells reflected by capacitance *C*~*Cl*~ ([Fig 4B](#pone.0197406.g004){ref-type=\"fig\"}).\n\n![**TEER values of PBCEC, treated with 1 \u03bcM ENN B and ENN B1 for 48 h (*n* \\> 6, negative control ACN 0.1%) (A) and C**~**Cl**~**-measurement (B) of PBCEC incubated with 1 \u03bcM ENN B and ENN B1 (*n* \\> 6, negative control ACN 0.1%) over the course of 48 h.** Data are presented as the mean \u00b1 standard deviation (SD). The significance indicated refers to the significance level as compared to the solvent treated control (0.1% ACN) calculated with one-way ANOVA and Dunnett's multiple-comparison (Significance levels: \\* significant (p \u2264 0.05); \\*\\* highly significant (p \u2264 0.01); \\*\\*\\* very highly significant (p \u2264 0.001)).](pone.0197406.g004){#pone.0197406.g004}\n\nSamples were drawn from the apical (\"blood\"-side) and basolateral (\"brain\"-side) compartment of the PBCEC two-compartment model after 1, 2.5, 6.5, 18, 24, 28, 42 and 48 h and quantitated via LC-MS/MS. The results are summarized in [Fig 5](#pone.0197406.g005){ref-type=\"fig\"}. Already after 6.5 h 53% of the apically applied ENN B were found in the basolateral compartment soon reaching a plateau at 70 \u00b1 4% ([Fig 5A](#pone.0197406.g005){ref-type=\"fig\"}). Along with the quantitation in the apical and basolateral cell culture medium, the polycarbonate filter membranes with the attached PBCEC monolayers were extracted and analyzed after 48 h of incubation (data not shown). Even though ENN B was detectable in the filter membranes, the amount in relation to the initially applied amounts was below 1% and therefore negligible in terms of contributing to the overall distribution of ENN B in the system. After completion of the experiment (48 h), the amount of ENN B recovered from the system, i.e. amount of compound in the apical and basolateral compartment summed up, is 92,8 \u00b1 23,1% of the initially introduced amount of ENN B (0.5 nmol). The permeability coefficient *p*~*c*~ was calculated according to Eqs [1](#pone.0197406.e001){ref-type=\"disp-formula\"} and [2](#pone.0197406.e002){ref-type=\"disp-formula\"}. For ENN B the obtained *P*~*e*~ (ENN B) is (14.9 \u00d7 10^\u22126^ cm/s). For ENN B1, a very similar transport kinetic was observed. ENN B1 also showed a very fast and high transfer from the apical to the basolateral compartment with a very similar derived permeability coefficient of *P*~*e*~ (ENN B1) = 14.6 \u00d7 10^-6^ cm/s. The permeability of PBCEC for ENN B and ENN B1 ranges in the same order of magnitude as the high apparent permeability compounds caffeine (*P*~*app*~ = 15.5 \u00d7 10^\u22126^ cm/s) and diazepam (*P*~*app*~ = 12.7 \u00d7 10^\u22126^ cm/s) \\[[@pone.0197406.ref029]\\], which are known to reach the brain. After 6.5 h 44% of the apically applied ENN B1 was detected in the basolateral compartment ([Fig 5B](#pone.0197406.g005){ref-type=\"fig\"}). From 6.5 h to the completion of the experiment after 48 h, the applied 0.5 nmol ENN B1 in the apical compartment were transferred to the basolateral compartment up to 55%. Analysis of the polycarbonate filters including the attached PBCEC revealed that for ENN B1 2.7% of the initially applied amount of compound could be recovered. This could give a hint why the sum of ENN B1 in both compartments did not completely add up to 100% of the originally introduced amount of ENN B1 (0.5 nmol) but only 73,7 \u00b1 20,2% after 48 h. To exclude the involvement of metabolization, a screening was carried out via high resolution mass spectrometry, but no metabolites were detected (data not shown).\n\n![**Passive transfer kinetics of 0.5 nmol (1 \u03bcM apical application) of ENN B (A) and ENN B1 (B) through the PBCEC monolayer over the course of 48 h (*n* = 9)**. Concentrations recovered in the apical and basolateral compartment at every sampling time-point are displayed as relative amount of the initially applied amount of test substance (here: 0.5 nmol ENN B (A) and ENN B1 (B)). Data are presented as the mean \u00b1 standard deviation (SD).](pone.0197406.g005){#pone.0197406.g005}\n\nTo study whether there was any active transport, ENN B and ENN B1 were applied on both sides of the PBCEC barrier in equimolar concentrations of 200 nM for each toxin. This concentration did not have any barrier impairing effect on the PBCEC monolayer. For both compounds, ENN B and ENN B1, no conclusive results concerning active transport could be obtained (data not shown). Still, a minor enrichment in the apical compartment for both mycotoxins could be suggested.\n\nAfter completion of each transport experiment, both active and passive, lucifer yellow (LY) was applied in 50 \u03bcM concentration in the apical compartments. Samples were drawn after 1 h and the amount of LY in the basolateral compartment was quantitated via external calibration. The permeability coefficients for LY through cells treated with the solvent control (ACN 0.1% and 0.02%) and through cells treated with the test compounds for the past 48 h were calculated ([Table 2](#pone.0197406.t002){ref-type=\"table\"}). The permeability coefficients range in the same order of magnitude for test substance treated cells and cells incubated with solvent, which is considerably lower than the permeability coefficients derived for ENN B and ENN B1 and indicates an intact and tight monolayer with low paracellular transfer of compounds.\n\n10.1371/journal.pone.0197406.t002\n\n###### Permeability coefficients of the PBCEC monolayer to ENN B, ENN B1 and LY.\n\n![](pone.0197406.t002){#pone.0197406.t002g}\n\n  Analyte                  Permeability coefficient \\[10^\u22126^ cm/s\\]   Analyte                 Permeability coefficient \\[10^\u22126^ cm/s\\]\n  ------------------------ ------------------------------------------ ----------------------- ------------------------------------------\n  ENN\u00a0B                    14.9                                       ENN\u00a0B1                  14.6\n  LY~(ENN\u00a0B,\\ passive)~    1.9                                        LY~(ENN\u00a0B,\\ active)~    0.7\n  LY~(ENN\u00a0B1,\\ passive)~   2.2                                        LY~(ENN\u00a0B1,\\ active)~   0.9\n  LY~(ACN\\ 0.1%)~          3.2                                        LY~(ACN\\ 0.02%)~        0.2\n\nDerived permeability coefficients (*n* \\> 3) for the permeation of ENN B and ENN B1 from the apical to the basolateral compartment through the PBCEC monolayer for 1 \u03bcM apical application (passive transport). Permeability coefficients for LY from the apical to the basolateral compartment through the PBCEC monolayer after treatment with 1 \u03bcM ENN B and ENN B1 apically (passive transport), 200 nM ENN B and ENN B1 equimolar in both compartments (active transport) and the corresponding negative controls (ACN 0.1% and ACN 0.02%), respectively.\n\nENNs are able to penetrate different barriers in higher organisms. Keeping this in mind, in pigs, ENN B1 is quickly absorbed after oral uptake (*k*~a~ = 4.66 h^-1^~,~ *T*~max~ = 0.24 h) with a very high oral bioavailability of up to 91%. This is even higher than expected from a previous *in vitro-*study, performed via an intestinal uptake experiment with Caco-2 cells. The bioavailabilities for ENN B1 are suggested to be around 55 to 66% \\[[@pone.0197406.ref013], [@pone.0197406.ref030]\\]. In the applied BBB model, a very fast transfer of ENN B and B1 from the \"blood\"-side (apical) to the \"brain\"-side (basolateral) is shown. In mice, the influx of ENN B and ENN B1 shows rapid initial influx rates with a brain tissue distribution of 95% in the parenchyma and only 5% of the ENNs remaining in the capillaries \\[[@pone.0197406.ref014]\\]. No metabolism was observed in the Caco-2 model (*in vitro*), an *in vivo-*study in mice and the applied porcine BBB model described in this study \\[[@pone.0197406.ref013], [@pone.0197406.ref014], [@pone.0197406.ref030]\\]. In general, metabolization of these mycotoxins is described in other models. In liver microsomes of various origins (human, dog and rat), several phase I metabolites of ENN B are formed, hence phase I enzymes like CYP3A and CYP1A play a key role in the metabolism of this mycotoxin \\[[@pone.0197406.ref031]\\]. In pigs an assortment of ENN B and ENN B1 phase I-metabolites was detected and also a correlation between administration route and formation of certain metabolites was found, indicating a pre-systemic metabolism of ENN B1 after oral administration \\[[@pone.0197406.ref032]\\].\n\nConclusion {#sec018}\n==========\n\nIn this study, the transfer of ENN B and ENN B1 across the BBB and effects on cells of the BBB were evaluated. PBCEC present a reliable and well-established two-compartment *in vitro-*model mimicking the BBB. Although this model system cannot completely reflect the conditions in the complex and dynamic milieu of a human brain *in vivo*, it serves as good screening tool. The obtained findings, in combination with an array of toxicologic (*in vitro*) test systems, are crucial for providing valuable information for risk assessment concerning potential neurotoxic effects of food contaminants like mycotoxins.\n\nThe passive transport studies showed a high permeation of ENNs from the \"blood\"-side to the \"brain\"-side, whereas the active transport suggested only a weak efflux of ENNs from the basolateral to the apical compartment. As soon as ENNs come in contact with the BBB or even cross it, effects on different cells of the BBB have to be considered. ENN B and ENN B1 induced high cytotoxicity in the astrocytoma cell line CCF-STTG1. Within the evaluated concentration range and time of exposure, ENNs, especially ENN B, caused the induction of apoptosis rather than necrosis in CCF-STTG1 cells. While the effects on viability in microvascular endothelial cells of the brain were less pronounced than in the astrocytoma cells (CCF-STTG1).\n\nHuman exposure and the lack of toxicity data for full risk assessment of these mycotoxins, support the need for further evaluation of their underlying mechanism of action. Furthermore, the discrepancy between the high activity of cyclohexadepsipeptides *in vitro* and the \"missing\" effects *in vivo* has to be clarified.\n\nWe thank our colleagues from the group of Prof. Dr. Langer, Institute of Pharmaceutical Technology, WWU Muenster, for providing PBCEC cells and Prof. Dr. Karch, Institute of Hygiene, WWU Muenster, for providing HBMEC cells.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n"}
+{"text": "![](edinbmedj74236-0085){#sp1 .565}\n\n![](edinbmedj74236-0086){#sp2 .566}\n\n![](edinbmedj74236-0087){#sp3 .567}\n\n![](edinbmedj74236-0088){#sp4 .568}\n\n![](edinbmedj74236-0089){#sp5 .569}\n"}
+{"text": "Introduction {#s1}\n============\n\nMulticellular development relies on the formation of cell-type-specific gene expression programmes that support differentiation. At the most basic level, these expression programmes are defined by cell signaling pathways that control how transcription factors bind DNA sequences in gene regulatory elements and shape RNA polymerase II (RNAPolII)-based transcription from the core gene promoter (reviewed in \\[[@bib74]\\]). In eukaryotes, the activity of RNAPolII is also regulated by a large multi-subunit complex, Mediator, which can directly interact with transcription factors at gene regulatory elements and with RNAPolII at the gene promoter to modulate transcriptional activation. The Mediator complex functions through regulating pre-initiation complex formation and controlling how RNAPolII initiates, pauses, and elongates. Therefore, Mediator is central to achieving appropriate transcription from gene promoters (reviewed in \\[[@bib4]; [@bib56]\\]).\n\nIn mammalian genomes, CpG dinucleotides are pervasively methylated and this epigenetically maintained DNA modification is generally associated with transcriptional inhibition, playing a central role in silencing of repetitive and parasitic DNA elements ([@bib44]; [@bib70]). Most gene promoters, however, are embedded in short elements with elevated CpG dinucleotide content, called CpG islands, which remain free of DNA methylation ([@bib10]; [@bib38]; [@bib48]). Interestingly, mammalian cells have evolved a DNA-binding domain, called a ZF-CxxC domain, which can recognize CpG dinucleotides when they are non-methylated ([@bib50]; [@bib82]). This endows ZF-CxxC domain-containing proteins with the capacity to recognize and bind CpG islands throughout the genome ([@bib15]; [@bib78]). There are 12 mammalian proteins encoding a ZF-CxxC domain, most of which are found in large chromatin modifying complexes that post-translationally modify histone proteins to regulate gene expression from CpG islands ([@bib53]). This has led to the proposal that CpG islands may function through chromatin modification to affect transcription ([@bib14]).\n\nOne of the first characterised ZF-CxxC domain-containing proteins was lysine-specific demethylase 2A (KDM2A) ([@bib15]). KDM2A is a JmjC-domain-containing histone lysine demethylase that removes histone H3 lysine 36 mono- and di- methylation (H3K36me1/2) from CpG islands ([@bib15]; [@bib80]). Like DNA methylation, H3K36me1/me2 is found throughout the genome and is thought to be repressive to gene transcription ([@bib19]; [@bib42]; [@bib63]). Therefore, it was proposed that KDM2A counteracts H3K36me2-dependent transcriptional inhibition at CpG island-associated gene promoters ([@bib15]). Sequence-based homology searches revealed that KDM2A has two paralogues in vertebrates, KDM2B and FBXL19 ([@bib41]). KDM2B, like KDM2A, binds to CpG islands and can function as a histone demethylase for H3K36me1/2 via its JmjC domain ([@bib28]; [@bib33]). However, unlike KDM2A, it physically associates with the polycomb repressive complex 1 (PRC1) to control how transcriptionally repressive polycomb chromatin domains form at a subset of CpG island-associated genes ([@bib12]; [@bib28]; [@bib34]; [@bib85]). These observations have suggested that, despite extensive similarity between KDM2A and KDM2B and some functional redundancy in histone demethylation, individual KDM2 paralogues have evolved unique functions. FBXL19 remains the least well characterized and most divergent of the KDM2 paralogues. Unlike KDM2A and KDM2B, it lacks a JmjC domain and, therefore, does not have demethylase activity ([@bib53]). Previous work on FBXL19 has suggested that it plays a role in a variety of cytoplasmic processes that affect cell proliferation, migration, apoptosis, TGF\u03b2 signalling, and regulation of the innate immune response ([@bib24]; [@bib83]; [@bib88]; [@bib89]). More recently, it has also been proposed to have a nuclear function as a CpG island-binding protein ([@bib49]), but its role in the nucleus still remains poorly defined.\n\nHere, we investigate the function of FBXL19 in mouse embryonic stem (ES) cells and show that it is predominantly found in the nucleus where it localizes in a ZF-CxxC-dependent manner to CpG island promoters. Biochemical purification of FBXL19 revealed an association with the CDK-containing Mediator complex and we discover that FBXL19 can target CDK-Mediator to chromatin. Conditional removal of the CpG island-binding domain of FBXL19 in ES cells leads to a reduction in the occupancy of CDK8 at CpG islands associated with inactive developmental genes. FBXL19 and CDK-Mediator appear to play an important role in priming genes for future expression, as these genes are not appropriately activated during ES cell differentiation when the CpG island-binding capacity of FBXL19 is abolished or CDK-Mediator is disrupted. Consistent with an important role for FBXL19 in supporting normal developmental gene expression, removal of the ZF-CxxC domain of FBXL19 leads to perturbed development and embryonic lethality in mice. Together, our findings uncover an interesting new mechanism by which CDK-Mediator is recruited to gene promoters in ES cells and demonstrate a requirement for FBXL19 and CDK-Mediator in priming developmental gene expression during cell lineage commitment.\n\nResults {#s2}\n=======\n\nFBXL19 is enriched in the nucleus and binds CpG islands via its ZF-CxxC domain {#s2-1}\n------------------------------------------------------------------------------\n\nFBXL19 shares extensive sequence similarity and domain architecture with the other KDM2 paralogues which are predominantly nuclear proteins ([Figure 1A](#fig1){ref-type=\"fig\"}, \\[[@bib15]; [@bib28]\\]), yet the function of FBXL19 has largely been described in the context of cytoplasmic processes ([@bib24]; [@bib83]; [@bib88]; [@bib89]). Therefore, we first examined the subcellular distribution of FBXL19 to understand whether it was unique amongst KDM2 paralogues in localizing to the cytoplasm. To achieve this, we generated a mouse ES cell line expressing epitope-tagged FBXL19 (FS2-FBXL19) and carried out immuno-fluorescent staining. This revealed that FBXL19 was almost exclusively nuclear ([Figure 1B](#fig1){ref-type=\"fig\"}). When we carried out subcellular biochemical fractionation, FBXL19 was also enriched in the nuclear fractions in agreement with the immuno-fluorescent staining ([Figure 1C](#fig1){ref-type=\"fig\"}). Importantly, FBXL19 was found not only in the soluble nuclear extract but also in the insoluble nuclear pellet, which contains chromatin-bound factors ([Figure 1C](#fig1){ref-type=\"fig\"}). This suggested that FBXL19 may associate with chromatin, like other ZF-CxxC domain-containing proteins.\n\n![FBXL19 binds CpG islands genome-wide via its ZF-CxxC domain.\\\n(**A**) A schematic illustrating KDM2 protein domain architecture. (**B**) Immuno-fluorescent staining for FS2-FBXL19 in ES cells. (**C**) ES cell fractionation and Western blot for factors enriched in the nucleus (TBP), the cytoplasm (CUL1), and FS2-FBXL19. Ponceau S staining indicates protein loading. Cyto -- cytoplasmic fraction, NE -- soluble nuclear extract, NP -- insoluble nuclear pellet. The asterisk indicates a non-specific band. (**D**) Screen shots showing ChIP-seq traces for KDM2B, KDM2A and FS2-FBXL19. BioCAP is included to indicate the location of non-methylated DNA and computationally predicted CpG islands (CGIs) are illustrated above. (**E**) Heatmaps showing enrichment of KDM2 proteins over a 5 kb region centred on non-methylated islands identified by BioCAP (NMIs) (n\u00a0=\u00a027698), sorted by decreasing BioCAP signal. BioCAP is shown for comparison. (**F**) A scatter plot showing the Spearman correlation between FS2-FBXL19 ChIP-seq signal and BioCAP signal at NMIs. (**G**) A schematic illustrating the FS2-FBXL19 transgenes. (**H**) A screen shot showing ChIP-seq traces for WT FS2-FBXL19 and ZF-CxxC FS2-FBXL19 mutants as in G. EV indicates empty vector control. (**I**) A metaplot analysis of WT FS2-FBXL19 and ZF-CxxC FS2-FBXL19 mutants over FBXL19 peaks.](elife-37084-fig1){#fig1}\n\nBased on the nuclear localization of FBXL19 and the fact that it encodes a highly conserved ZF-CxxC domain ([Figure 1---figure supplement 1A](#fig1s1){ref-type=\"fig\"}), we set out to determine whether FBXL19 is a CpG island-binding protein. To achieve this, we carried out chromatin immunoprecipitation followed by massively parallel sequencing (ChIP-seq) for epitope-tagged FBXL19 and compared its binding profile with those we have previously generated for KDM2A and KDM2B in mouse ES cells ([@bib12]; [@bib28]). A visual examination of FBXL19 ChIP-seq signal revealed that it was highly enriched at both computationally predicted CpG islands and genomic regions that contain BioCAP signal, an experimental measure of non-methylated DNA ([@bib13]) ([Figure 1D](#fig1){ref-type=\"fig\"}). We then identified all non-methylated CpG islands (NMIs) using BioCAP data (n\u00a0=\u00a027698) ([@bib54]) and extended our analysis across the genome. We observed that FBXL19 ChIP-seq signal at NMIs was highly similar to that of KDM2A and KDM2B ([Figure 1E](#fig1){ref-type=\"fig\"}). Furthermore, FBXL19 ChIP-seq signal correlated well with the density of non-methylated CpG dinucleotides in NMIs, similarly to KDM2A and KDM2B, ([Figure 1F](#fig1){ref-type=\"fig\"} and [Figure 1---figure supplement 1B,C](#fig1s1){ref-type=\"fig\"}), and almost all FBXL19-occupied sites fell within NMIs ([Figure 1---figure supplement 1D](#fig1s1){ref-type=\"fig\"}). Together these findings demonstrate that FBXL19 is a nuclear protein that binds to CpG islands, an observation supported by a recent independent study ([@bib49]).\n\nKDM2A and KDM2B rely on defined residues in their ZF-CxxC domain to recognize non-methylated cytosine and bind CpG islands ([@bib15]; [@bib28]). To determine whether the association of FBXL19 with CpG islands is dependent on this domain, we generated ES cell lines expressing either a mutant version of FBXL19, in which a key lysine residue was substituted to alanine (K49A, [Figure 1---figure supplement 1A](#fig1s1){ref-type=\"fig\"}) to disrupt the recognition of non-methylated CpGs, or a truncated version of FBXL19, where the ZF-CxxC domain was deleted (\u0394CXXC) ([Figure 1G](#fig1){ref-type=\"fig\"}). Importantly, the expression levels of wild type (WT) and mutant FBXL19 transgenes were highly similar ([Figure 1---figure supplement 1E](#fig1s1){ref-type=\"fig\"}). We then carried out ChIP-seq for epitope-tagged FBXL19 in these cell lines and compared the binding profiles of the mutant FBXL19 to that of WT FBXL19 ([Figure 1H](#fig1){ref-type=\"fig\"}). This revealed a near complete loss of FBXL19 binding to chromatin when the ZF-CxxC domain was deleted and a slightly less dramatic effect in the K49A mutant ([Figure 1H,I](#fig1){ref-type=\"fig\"}, and [Figure 1---figure supplement 1F,G](#fig1s1){ref-type=\"fig\"}). Together, these observations demonstrate that binding of FBXL19 to CpG islands relies on an intact and functional ZF-CxxC domain.\n\nFBXL19 interacts with the CDK-Mediator complex in ES cells {#s2-2}\n----------------------------------------------------------\n\nOur ChIP-seq analyses demonstrated that FBXL19 is targeted to CpG islands in a manner that is highly similar to KDM2A and KDM2B ([Figure 1](#fig1){ref-type=\"fig\"}). Although KDM2A and KDM2B localise to the same genomic regions and show a high degree of sequence conservation, they associate with different proteins ([@bib28]). This raised the interesting possibility that FBXL19 might also have unique interaction partners. To investigate this, we affinity-purified epitope-tagged FBXL19 from ES cell nuclear extract and identified associated proteins by mass spectrometry (AP-MS) ([Figure 2A](#fig2){ref-type=\"fig\"}, [Figure 2---figure supplement 1A](#fig2s1){ref-type=\"fig\"}, and [Figure 2---source data 1](#fig2sdata1){ref-type=\"supplementary-material\"}). This revealed that FBXL19 interacts with SKP1, a known F-box-binding protein that also associates with KDM2A and KDM2B ([@bib6]; [@bib28]), and the nuclear proteasome activator PSME3 ([@bib86]). Interestingly, we also identified multiple subunits of the Mediator complex that appeared to interact with FBXL19 in a sub-stoichiometric manner ([Figure 2A,B](#fig2){ref-type=\"fig\"}). Biochemical purifications of Mediator have identified two distinct assemblies ([@bib52]; [@bib58]; [@bib75]). The first is characterized by the presence of the MED26 subunit which associates with the middle region of Mediator and this form of the complex interacts with RNAPolII ([@bib60]; [@bib61]; [@bib69]; [@bib76]). Alternatively, a kinase module, composed of CDK8/CDK19, MED12/12L, MED13/13L, and CCNC, can bind to Mediator in a manner which is mutually exclusive with MED26 ([@bib75]). Interestingly, our FBXL19 purification identified subunits of the kinase-containing Mediator complex (CDK-Mediator), but not MED26 or RNAPolII ([Figure 2](#fig2){ref-type=\"fig\"}). This suggests that FBXL19 interacts preferentially with CDK-Mediator, however, we cannot exclude the possibly it may also interact with MED26-Mediator. CDK8 and its paralogue CDK19 share 77% amino acid identity (89% similarity) ([@bib5]; [@bib32]; [@bib69]; [@bib81]) and four out of the five peptides identified by mass spectrometry were common between the two proteins (data not shown). Therefore, it is likely that FBXL19 is able to interact with both CDK8- and CDK19-containing Mediator complexes. Importantly, interaction with CDK-Mediator was also evident when we performed affinity-purification of endogenous FBXL19 ([Figure 2---figure supplement 1B--D](#fig2s1){ref-type=\"fig\"}). However, reciprocal immunopurifications of MED12 and CDK8 failed to yield detectable FBXL19 by western blot ([Figure 2---figure supplement 1D](#fig2s1){ref-type=\"fig\"}). This is in agreement with our mass spectrometry analysis, which indicated that the association of FBXL19 with CDK-Mediator is sub-stoichiometric, and suggests that this interaction is likely weak, as opposed to stable, inside\u00a0cells.\n\n![FBXL19 interacts with\u00a0the CDK-Mediator complex in ES cells.\\\n(**A**) A heatmap representing the empirically modified protein abundance index (emPAI) of identified FBXL19-interacting proteins by affinity purification mass spectrometry (AP-MS). Data shown is an average of two biological replicates. The location of the identified Mediator components within the holocomplex is summarised on the right and in (**B**). (**B**) A schematic representation of the identified FBXL19-interacting proteins. Subunits of the CDK-Mediator complex not identified by AP-MS are shown in white. (**C**) Western blot analysis of FS2-FBXL19 and control purifications (EV) probed with the indicated antibodies.\\\n10.7554/eLife.37084.006Figure 2---source data 1.Mass spectrometry data.](elife-37084-fig2){#fig2}\n\nWe next wanted to determine which region of FBXL19 is required for interaction with CDK-Mediator. To do so, we transiently expressed full length FBXL19 or versions of FBXL19 with individual domains removed and performed affinity purification followed by western blot analysis. Intact FBXL19 and a version with the ZF-CxxC domain removed interacted with CDK-Mediator, whereas removing the F-box domain resulted in a loss of this interaction ([Figure 2---figure supplement 1E](#fig2s1){ref-type=\"fig\"}). Therefore, FBXL19 relies on its F-box, and not its capacity to bind non-methylated DNA, for its association with CDK-Mediator.\n\nBased on a candidate approach, it was recently reported that FBXL19 could interact the RNF20/40 E3 ubiquitin ligase in ES cells and regulate histone H2B lysine 120 ubiquitylation (H2BK120ub1) ([@bib49]). In our unbiased biochemical purification of FBXL19, we did not identify an interaction with RNF20/40 by AP-MS or by western blot analysis ([Figure 2A](#fig2){ref-type=\"fig\"} and [Figure 2---figure supplement 1F](#fig2s1){ref-type=\"fig\"}). Furthermore, we failed to observe any relationship between the ability of FBXL19 to associate with CpG islands and the levels of H2BK120ub1 ([Figure 2---figure supplement 1G](#fig2s1){ref-type=\"fig\"}). Therefore, the relevance of this proposed interaction remains unclear.\n\nFBXL19 contains conserved F-box and leucine-rich repeat (LRR) domains. F-box proteins are known to function as scaffolds and substrate recognition modules for the SKP1-Cullin-F-box (SCF) protein complexes that ubiquitylate proteins for degradation by the proteasome ([@bib72]; [@bib73]). Based on the association of FBXL19 with SKP1 and PSME3, we speculated that FBXL19 might function as a SCF substrate-selector for CDK-Mediator, as has previously been observed for the related F-box-containing protein FBW7 ([@bib22]). Interestingly, however, in our FBXL19 purifications, we did not detect the SCF complex components CUL1 and RBX1/2, which are required for ubiquitin E3 ligase activity ([@bib72]) ([Figure 2A](#fig2){ref-type=\"fig\"}). Nevertheless, we investigated in more detail whether FBXL19 might regulate CDK-Mediator protein levels. Treatment of ES cells with the proteasome inhibitor MG132, which sequesters SCF substrates on their substrate selector, did not lead to elevated levels of Mediator subunits in FBXL19 purifications as identified by AP-MS (unpublished observation). In addition, transient overexpression of FBXL19 did not cause an appreciable reduction in CDK-Mediator protein ([Figure 2---figure supplement 1H](#fig2s1){ref-type=\"fig\"}). Based on these observations, we conclude that FBXL19 does not function as a SCF substrate selector for CDK-Mediator. This raised the interesting possibility that FBXL19 may function in a proteasome-independent manner with CDK-Mediator at CpG islands.\n\nFBXL19 recruits CDK-Mediator to chromatin {#s2-3}\n-----------------------------------------\n\nIt has previously been shown that transcription factors can recruit the Mediator complex to enhancers and gene promoters ([@bib65]), yet the complement of mechanisms by which Mediator is targeted to chromatin remains very poorly defined. Given that FBXL19 does not appear to regulate CDK-Mediator protein levels, we hypothesized that it might instead function to recruit CDK-Mediator to chromatin. To test this possibility, we took advantage of a synthetic system we have developed to nucleate proteins *de novo* on chromatin and test their capacity to recruit additional factors ([@bib12]). Fusion of the Tet repressor DNA-binding domain (TetR) to FBXL19 allowed the recruitment of FBXL19 to a short array of Tet repressor DNA binding sites (TetO), engineered into a single site on mouse chromosome 8 (\\[[@bib12]\\], [Figure 3A](#fig3){ref-type=\"fig\"}). To determine whether FBXL19 was sufficient to recruit CDK-Mediator to chromatin, we stably expressed TetR or the TetR-FBXL19 fusion protein in the TetO array-containing ES cell line ([Figure 3---figure supplement 1A](#fig3s1){ref-type=\"fig\"}). We then carried out ChIP for CDK-Mediator subunits ([Figure 3B](#fig3){ref-type=\"fig\"}). Consistent with our biochemical observations ([Figure 2](#fig2){ref-type=\"fig\"}), when we examined the binding of CDK8 and MED12 over the TetO array, we observed an enrichment in the TetR-FBXL19 line. Furthermore, binding of MED4, which is part of the core Mediator complex, was also evident at the TetO array ([Figure 3B](#fig3){ref-type=\"fig\"}). This suggests that FBXL19 may be sufficient to recruit a holo-CDK-Mediator complex, in keeping with its biochemical co-purification with multiple components of both the CDK module and core Mediator complex ([Figure 2](#fig2){ref-type=\"fig\"}). Importantly, stable expression of TetR-KDM2A and TetR-KDM2B did not lead to CDK8 recruitment, indicating that this activity is unique to FBXL19 ([Figure 3---figure supplement 1B](#fig3s1){ref-type=\"fig\"}). Further work will be required to determine the dynamics with which individual Mediator components are recruited to chromatin by FBXL19 as well as the precise composition of such complexes. Together these observations demonstrate that FBXL19 interacts specifically with the CDK-Mediator and can recruit this complex *de novo* to chromatin.\n\n![FBXL19 recruits CDK-Mediator to chromatin.\\\n(**A**) A schematic representation of the integration site of the TetO array. (**B**) ChIP-qPCR analysis of the binding of the indicated proteins across the TetO array in TetR only (top) and TetR-FBXL19 (bottom) ES cell lines. The x-axis indicates the spatial arrangement of the qPCR amplicons with respect to the center of the TetO array (in kb). Error bars represent SEM of three biological replicates.](elife-37084-fig3){#fig3}\n\nFBXL19 is required for appropriate CDK8 occupancy at a subset of CpG island-associated promoters {#s2-4}\n------------------------------------------------------------------------------------------------\n\nFBXL19 binds specifically to CpG islands ([Figure 1](#fig1){ref-type=\"fig\"}) and can recruit CDK-Mediator to an artificial binding site on chromatin ([Figure 3](#fig3){ref-type=\"fig\"}). This raised the interesting possibility that FBXL19 may help to recruit CDK-Mediator to CpG islands in ES cells. To examine this, we first performed CDK8 ChIP-seq in ES cells and compared it to FBXL19 ChIP-seq. Unlike FBXL19, CDK8 occupancy was not restricted to CpG islands ([Figure 4---figure supplement 1A](#fig4s1){ref-type=\"fig\"}), with only 67.5% of CDK8 peaks overlapping with NMIs and CDK8 binding showing a limited correlation with BioCAP signal (Spearman correlation -- 0.48, [Figure 4---figure supplement 1B](#fig4s1){ref-type=\"fig\"}). This is in line with previous studies demonstrating that the Mediator complex is recruited to both enhancers and gene promoters ([@bib40]; [@bib56]). Interestingly, however, we observed enrichment of CDK8 at FBXL19 peaks ([Figure 4---figure supplement 1C](#fig4s1){ref-type=\"fig\"}) with 89.4% of FBXL19 peaks overlapping with CDK8 peaks and NMIs ([Figure 4---figure supplement 1D](#fig4s1){ref-type=\"fig\"}) raising the possibility that FBXL19 may contribute to its occupancy at these sites. To directly test this, we developed an ES cell system in which the exon encoding the ZF-CxxC domain of FBXL19 is flanked by loxP sites (*Fbxl19^fl/fl^)* ([Figure 4---figure supplement 1E](#fig4s1){ref-type=\"fig\"}) and which expresses tamoxifen-inducible ERT2-Cre recombinase. Upon addition of tamoxifen (OHT), the ZF-CxxC-encoding exon is excised, yielding a form of FBXL19 that lacks the ZF-CxxC domain (FBXL19^\u0394CXXC^) ([Figure 4A,B](#fig4){ref-type=\"fig\"}, and [Figure 4---figure supplement 1F](#fig4s1){ref-type=\"fig\"}) and can, therefore, no longer bind CpG islands ([Figure 1](#fig1){ref-type=\"fig\"}). This model cell system allows us to specifically examine the CpG island-associated functions of FBXL19 without affecting its other proposed roles ([@bib24]; [@bib83]; [@bib88]; [@bib89]). Following removal of the ZF-CxxC domain of FBXL19, we observed some reductions in FBXL19 protein levels ([Figure 4B](#fig4){ref-type=\"fig\"}), but importantly CDK8 levels were unaffected ([Figure 4---figure supplement 1G](#fig4s1){ref-type=\"fig\"}). Using this conditional *Fbxl19^\u0394CXXC^* ES cell line, we then examined CDK8 occupancy on chromatin by ChIP-seq before and after OHT treatment. Genome-wide profiling of CDK8 in *Fbxll19^\u0394CXXC^* ES cells did not reveal widespread alterations in CDK8 binding ([Figure 4C](#fig4){ref-type=\"fig\"} left and [Figure 4---figure supplement 1H](#fig4s1){ref-type=\"fig\"}), indicating that FBXL19 is not the central determinant driving CDK8 recruitment to most of its binding sites in the genome. Intriguingly, however, a more detailed site-specific analysis revealed a subset of CDK8 target sites that displayed significant alteration in CDK8 occupancy ([Figure 4C,D](#fig4){ref-type=\"fig\"}, and [Figure 4---figure supplement 1H](#fig4s1){ref-type=\"fig\"}). In keeping with a potential role for FBXL19 in CDK8 recruitment, the majority of the affected sites showed reduced CDK8 binding (n\u00a0=\u00a0783) ([Figure 4C](#fig4){ref-type=\"fig\"}).\n\n![FBXL19 is required for appropriate CDK8 occupancy at a subset of CpG island promoters.\\\n(**A**) A schematic illustrating how addition of OHT (4-hydroxytamoxifen) leads to the generation of *Fbxl19^\u0394CXXC^* ES cells. (**B**) Western blot analysis of an OHT-treatment time course in the *Fbxl19^fl/fl^* ES cell line. Hours of treatment are shown. FBXL19 protein was C-terminally tagged with a T7 epitope ([Figure 2---figure supplement 1B, C](#fig2s1){ref-type=\"fig\"}) to allow for Western blot detection with an anti-T7 antibody. FBXL19 runs as a doublet that shifts in size following deletion of the CxxC domain. TBP was probed as a loading control. (**C**) Metaplots showing CDK8 enrichment in *Fbxl19^fl/fl^* ES cells (WT) and *Fbxl19^\u0394CXXC^* ES cells (OHT) at all CDK8 peaks (left), peaks showing decreased CDK8 occupancy (\u2191, middle) and peaks showing increased CDK8 occupancy (\u2193, right). The number of total peaks in each group is indicated. p-values denote statistical significance calculated by Wilcoxon rank sum test comparing ChIP-seq read counts between WT and OHT samples across a 1.5-kbp interval flanking the center of CDK8 peaks. (**D**) Screen shots showing ChIP-seq traces for CDK8 in wild type *Fbxl19^fl/fl^* ES cells (WT) and *Fbxl19^\u0394CXXC^* ES cells (OHT). BioCAP and FS2-FBXL19 tracks are given for comparison. Left -- CDK8 peaks showing reduced CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells; Right -- CDK8 peaks showing increased CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells (indicated with rectangles). (**E**) Boxplots showing log2 fold change (log2FC) of CDK8 ChIP-seq signal (RPKM) at all CDK8 peaks (n\u00a0=\u00a024273), and those with reduced CDK8 (n\u00a0=\u00a0783, \u2193) and increased CDK8 (n\u00a0=\u00a0379, \u2191). p-values were calculated using a Wilcoxon rank sum test. (**F**) Boxplots showing the\u00a0size of CDK8 peaks as in (**E**). p-values were calculated using a Wilcoxon rank sum test. (**G**) Venn diagrams representing the overlap between CDK8 peaks and super enhancers (SE) at CDK8 peaks as in (**E**). Percent overlap of all SEs is indicated. (**H**) A metaplot (left) showing BioCAP enrichment at CDK8 peaks as in (**E**) and boxplot quantification (right) of BioCAP RPKM levels. p-Values calculated using Wilcoxon rank sum test are indicated. (**I**) A metalplot (left) showing FS2-FBXL19 enrichment at CDK8 peaks as in (**E**) and boxplot quantification (right) of FS2-FBXL19 RPKM levels. p-values calculated using Wilcoxon rank sum test are indicated.](elife-37084-fig4){#fig4}\n\nUpon closer examination of sites with reduced CDK8 occupancy in *Fbxll19^\u0394CXXC^* ES cells, we discovered that they tended to coincide with broad regions of CDK8 enrichment ([Figure 4E,F](#fig4){ref-type=\"fig\"}), a feature often associated with super-enhancers ([@bib84]). However, a comparison of these sites to the location of super-enhancers in mouse ES cells indicated that these were distinct ([Figure 4G](#fig4){ref-type=\"fig\"}). Instead, sites displaying reduction in CDK8 occupancy coincided with broad regions of non-methylated DNA ([Figure 4D,H](#fig4){ref-type=\"fig\"}, and [Figure 4---figure supplement 1H](#fig4s1){ref-type=\"fig\"}), tended to be associated with gene promoters ([Figure 4---figure supplement 1I](#fig4s1){ref-type=\"fig\"}), and had elevated levels of FBXL19 ([Figure 4I](#fig4){ref-type=\"fig\"} and [Figure 4---figure supplement 1H,I,J](#fig4s1){ref-type=\"fig\"}). Together, these observations suggest that although binding of CDK8 to most of its target sites in the genome is achieved independently of FBXL19, a subset of broad CpG island-associated gene promoters appear to rely on FBXL19 for appropriate CDK8 occupancy.\n\nFBXL19 targets CDK8 to promoters of silent developmental genes in ES cells {#s2-5}\n--------------------------------------------------------------------------\n\nDespite associating widely with CpG island gene promoters, FBXL19 appears to play a very specific role in maintaining appropriate CDK8 occupancy at a particular subset of broad CpG island-associated promoters ([Figure 4](#fig4){ref-type=\"fig\"}). We were therefore curious to know whether something distinguishes these gene promoters from other FBXL19-bound promoters, where FBXL19 does not contribute appreciably to CDK8 occupancy ([Figure 4---figure supplement 1H](#fig4s1){ref-type=\"fig\"}). To achieve this, we initially performed gene ontology (GO) analysis ([@bib37]) on the genes with reduced CDK8 levels at their promoters in *Fbxl19^\u0394CXXC^* ES cells. Interestingly, this revealed that these genes were strongly enriched for developmental processes ([Figure 5A](#fig5){ref-type=\"fig\"}) in agreement with our previous discovery that broad CpG islands are an evolutionary conserved feature of developmentally regulated genes ([@bib54]). In contrast, genes associated with an increased binding of CDK8 did not show any significant GO term enrichment (unpublished observation). GO analysis of the genes associated with unchanged CDK8 levels revealed terms for a broad range of basic molecular processes in line with a general role for Mediator in transcriptional regulation ([Figure 5---figure supplement 1E](#fig5s1){ref-type=\"fig\"}).\n\n![FBXL19 targets CDK8 to promoters of silent developmental genes in ES cells.\\\n(**A**) Gene ontology analysis of genes associated with a decrease in CDK8 binding (n\u00a0=\u00a0673).(**B**) A boxplot showing expression levels (log2FPKM) of CDK8-bound genes in wild-type ES cells. CDK8-associated genes are divided based on CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells: all CDK8-bound (n\u00a0=\u00a015161), reduced CDK8 (n\u00a0=\u00a0673, \u2193), and increased CDK8 binding (n\u00a0=\u00a0255, \u2191). p-values were calculated using a Wilcoxon rank sum test. (**C**) A boxplot showing the change in gene expression (log2 fold change) observed by 4sU RNA-seq of CDK8-associated genes (as in B) following RA treatment. p-values were calculated using a Wilcoxon rank sum test. (**D**) A boxplot comparing gene expression levels (log2FPKM) of all (n\u00a0=\u00a019310) and FBXL19-bound (FBXL19+, n\u00a0=\u00a011368) genes separated by low (all genes n\u00a0=\u00a07417; FBXl19+\u00a0genes n\u00a0=\u00a02031) and high expression levels (all genes n\u00a0=\u00a011893, FBXl19+\u00a0genes n\u00a0=\u00a09337) in ES cells (based on [Figure 5---figure supplement 1B](#fig5s1){ref-type=\"fig\"}). (**E**) A boxplot showing CDK8 enrichment at all and FBXL19-bound genes separated by expression level as in (**D**). (**F**) A boxplot showing change in CDK8 binding at the TSSs of all and FBXL19-bound genes divided by expression level as in (**D**). p-values were calculated using a Wilcoxon rank sum test.](elife-37084-fig5){#fig5}\n\nIn pluripotent mouse ES cells, many developmental genes are inactive and only become expressed as cells commit to more differentiated lineages ([@bib17]). Importantly, the genes that exhibited reductions in CDK8 binding in *Fbxl19^\u0394CXXC^* cells were expressed at significantly lower levels than most other genes in ES cells ([Figure 5B](#fig5){ref-type=\"fig\"}). Previous work has identified a subset of poised but inactive developmental genes in ES cells that are proposed to exist in a bivalent chromatin state characterized by the co-occurrence of histone H3 lysine 4 and 27 methylation (H3K4/K27me) ([@bib9]; [@bib57]). Therefore, we examined whether sites that rely on FBXL19 for normal CDK8 binding also corresponded to bivalent regions in ES cells. We found that these regions are enriched for H3K27me3 ([Figure 5---figure supplement 1A](#fig5s1){ref-type=\"fig\"}) and have elevated H3K4me3 compared to other H3K27me3-modified sites that lack CDK8 ([Figure 5---figure supplement 1B](#fig5s1){ref-type=\"fig\"}). Therefore, sites that rely on FBXL19 for normal CDK8 binding and correspond to silent developmental gene promoters are also bivalent. To ask whether these genes are induced during cell lineage commitment, we compared their expression levels in ES cells and following retinoic acid (RA) treatment which induces differentiation ([Figure 6---figure supplement 1B,C](#fig6s1){ref-type=\"fig\"}). This clearly demonstrated that the genes which rely on FBXL19 for appropriate CDK8 binding in the ES cell state can become transcriptionally activated during stem cell lineage commitment ([Figure 5C](#fig5){ref-type=\"fig\"}).\n\nThe observation that FBXL19 appears to be important for CDK8 occupancy at largely inactive genes was intriguing given that previous work characterizing Mediator function has usually focussed on its activity at actively transcribed or induced genes (reviewed in \\[[@bib65]\\]). We therefore set out to examine the relationship between FBXL19, CDK8 and gene expression in more detail. We first separated all promoters based on expression of their\u00a0associated genes ([Figure 5D](#fig5){ref-type=\"fig\"} and [Figure 5---figure supplement 1D](#fig5s1){ref-type=\"fig\"}). We observed that CDK8 occupancy was in general linked to gene activity ([Figure 5E](#fig5){ref-type=\"fig\"}), in agreement with previous reports ([@bib3]; [@bib7]; [@bib25]; [@bib26]; [@bib31]; [@bib32]). However, at FBXL19-bound gene promoters, CDK8 occupancy was similar in both the lowly and highly expressed subsets of genes ([Figure 5E](#fig5){ref-type=\"fig\"}). Importantly, CDK8 occupancy at promoters of lowly expressed genes was reduced in *Fbxl19^\u0394CXXC^* ES cells ([Figure 5F](#fig5){ref-type=\"fig\"}). Therefore, these observations reveal that CDK8 is enriched at promoters of inactive or lowly expressed genes in ES cells and its binding is dependent on recognition of CpG islands by FBXL19.\n\nRemoving the CpG island-binding domain of FBXL19 results in a failure to induce developmental genes during ES cell differentiation {#s2-6}\n----------------------------------------------------------------------------------------------------------------------------------\n\nFBXL19 appears to play a role in recruiting CDK8 to a class of genes that are repressed in the ES cell state and become activated during cell linage commitment ([Figure 5](#fig5){ref-type=\"fig\"}). However, it remained unclear whether FBXL19 is required for the activation of these developmental genes. To address this question, we induced differentiation of *Fbxl19^fl/fl^* and *Fbxl19^\u0394CXXC^* ES cells with RA ([Figure 6A](#fig6){ref-type=\"fig\"}) and compared the expression of several genes which showed reduced CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells ([Figure 6B](#fig6){ref-type=\"fig\"}). We observed no significant differences in the expression of these genes in wild-type *Fbxl19^fl/fl^* and *Fbxl19^\u0394CXXC^* ES cells where these genes are silent. Strikingly, however, following RA treatment, *Fbxl19^\u0394CXXC^* cells failed to induce these genes appropriately ([Figure 6B](#fig6){ref-type=\"fig\"}). Similarly, developmental genes were not appropriately induced during embryoid body differentiation of *Fbxl19^\u0394CXXC^* ES cells ([Figure 6---figure supplement 1A](#fig6s1){ref-type=\"fig\"}). Importantly, this demonstrates that FBXL19 is required for the appropriate activation of developmental gene expression during cell lineage commitment.\n\n![Removing the CpG island-binding domain of FBXL19 results in a failure to induce developmental genes during ES cell differentiation.\\\n(**A**) A schematic illustrating the OHT treatment and differentiation approach. (**B**) RT-qPCR gene expression analysis of genes showing decreased CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells before (ESC) and after RA induction. Expression is relative to the average of two house-keeping genes. Error bars show SEM of three biological replicates, asterisks represent statistical significance calculated by Student T-test: \\*p\\<0.05, \\*\\*p\\<0.01, \\*\\*\\*p\\<0.001. (**C**) Volcano plots showing differential expression (log2 fold change) comparing WT and *Fbxl19^\u0394CXXC^* cells in the ES cell (top) and RA-induced state (bottom). Differentially expressed genes (log2FC\u00a0\\<\u00a0\u22120.5 or log2FC\u00a0\\>\u00a00.5, padj\u00a0\\<0.1) are shown in red. The number of genes considered to be significantly altered in expression are indicated. (**D**) Gene ontology analysis of genes with decreased expression in *Fbxl19^\u0394CXXC^* ES cells following RA treatment (n\u00a0=\u00a0552). (**E**) A boxplot indicating the expression level (FPKM) based on 4sU RNA-seq in wild-type ES cells of the gene groupings in (**C**): n.c. genes n\u00a0=\u00a017869 (no significant\u00a0change), up genes\u00a0=\u00a0889, down genes\u00a0=\u00a0552. p-values calculated using a Wilcoxon rank sum test are indicated. (**F**) A boxplot indicating the log2 fold change in gene expression of the gene groupings in (**E**) upon RA differentiation of ES cells. p-values calculated using a Wilcoxon rank sum test are indicated. (**G**) A boxplot indicating the change in CDK8 binding (log2FC) at the promoters of the gene groupings in (**E**) in ES cells. p-values calculated using a Wilcoxon rank sum test are indicated.\\\n10.7554/eLife.37084.015Figure 6---source data 1.Differential gene expression analysis.](elife-37084-fig6){#fig6}\n\nTo understand the extent to which genes are not appropriately induced during differentiation of *Fbxl19^\u0394CXXC^* ES cells, we examined ongoing transcription in both the ES cell state and after RA-mediated differentiation using short time-scale 4-thiouridine-based labeling and RNA sequencing (4sU RNA-seq) ([Figure 6A](#fig6){ref-type=\"fig\"}). We observed a significant number of genes that become induced upon RA treatment in wild-type cells (n\u00a0=\u00a04051) ([Figure 6---figure supplement 1B](#fig6s1){ref-type=\"fig\"} and [Figure 6---source data 1](#fig6sdata1){ref-type=\"supplementary-material\"}). GO term analysis confirmed that these genes are associated with ES cell differentiation and early embryonic development ([Figure 6---figure supplement 1C](#fig6s1){ref-type=\"fig\"}). We then used differential gene expression analysis to compare transcription in *Fbxl19^fl/fl^* and *Fbxl19^\u0394CXXC^* cells ([Figure 6C](#fig6){ref-type=\"fig\"} and [Figure 6---source data 1](#fig6sdata1){ref-type=\"supplementary-material\"}). While very few significant changes in gene expression were observed in the ES cell state, following RA-induced differentiation we identified a large number of genes (n\u00a0=\u00a0552) that had significantly lower expression levels in *Fbxl19^\u0394CXXC^* cells ([Figure 6C](#fig6){ref-type=\"fig\"}). This is consistent with our results when examining individual genes ([Figure 6B](#fig6){ref-type=\"fig\"} and [Figure 6---figure supplement 1A](#fig6s1){ref-type=\"fig\"}) and with a recent study where the expression of a selection of genes was examined following FBXL19 knock-down ([@bib49]). GO analysis revealed that the set of genes that were not appropriately activated were associated with genes involved in developmental processes ([Figure 6D](#fig6){ref-type=\"fig\"} and [Figure 6---source data 1](#fig6sdata1){ref-type=\"supplementary-material\"}), unlike genes with increased expression (n\u00a0=\u00a0889) which were not associated with these processes ([Figure 6---figure supplement 1D](#fig6s1){ref-type=\"fig\"} and [Figure 6---source data 1](#fig6sdata1){ref-type=\"supplementary-material\"}). It is possible that the observed increases in gene expression in *Fbxl19^\u0394CXXC^* cells following RA induction result from secondary effects or as of yet unidentified roles of FBXL19 in inhibiting gene expression. Nevertheless, consistent with our analysis of FBXL19-dependent CDK8 occupancy at promoters of developmental genes ([Figure 5B,C](#fig5){ref-type=\"fig\"}), genes not appropriately induced were lowly expressed in wild-type ES cells ([Figure 6E](#fig6){ref-type=\"fig\"}) and tended to become activated following RA induction ([Figure 6F](#fig6){ref-type=\"fig\"}). Importantly, these genes exhibited a reduction in CDK8 binding in *Fbxl19^\u0394CXXC^* ES cells ([Figure 6G](#fig6){ref-type=\"fig\"} and [Figure 6---figure supplement 1E](#fig6s1){ref-type=\"fig\"}), supporting the idea that FBXL19 recruits CDK8 to this subset of CpG island-associated developmental genes and that this may prime these genes for activation during differentiation.\n\nFBXL19 target genes rely on CDK-Mediator for activation during differentiation {#s2-7}\n------------------------------------------------------------------------------\n\nGiven that FBXL19 is required for CDK8 occupancy at the promoters of a series of silent developmental genes and for their activation during differentiation ([Figures 5](#fig5){ref-type=\"fig\"} and [6](#fig6){ref-type=\"fig\"}), we hypothesized that FBXL19 may prime these genes for future activation through the activity of CDK-Mediator. To address this interesting possibility, we developed a system to conditionally remove MED13 and its closely related paralogue MED13L in ES cells (*Med13/13l^fl/fl^* ERT2-Cre ES cells) ([Figure 7A](#fig7){ref-type=\"fig\"} and [Figure 7---figure supplement 1A](#fig7s1){ref-type=\"fig\"}). We chose to inactivate MED13/13L as it has previously been shown to physically link the CDK-kinase module to the core Mediator complex and underpin the formation of a functional CDK-Mediator ([@bib46]; [@bib79]). Treatment of the *Med13/13l^fl/fl^* ES cells with OHT resulted in a loss of MED13 and MED13L protein (MED13/13L KO) and reduced levels of the other subunits of the CDK-kinase module ([Figure 7B](#fig7){ref-type=\"fig\"}). Importantly, removal of MED13/13L also led to a loss of CDK8 binding to chromatin ([Figure 7C](#fig7){ref-type=\"fig\"}) but did not have an appreciable effect on the expression of FBXL19 target genes in the ES cell state ([Figure 7D](#fig7){ref-type=\"fig\"}). We next induced differentiation of the MED13/13L KO cells with RA ([Figure 7A](#fig7){ref-type=\"fig\"}). Importantly, when we then analysed the expression of a series of genes that rely on FBXL19 for activation during differentiation ([Figure 7D](#fig7){ref-type=\"fig\"}), we observed that these genes also failed to appropriately induce in MED13/13L KO cells ([Figure 7D](#fig7){ref-type=\"fig\"}). Together this suggests that FBXL19, via its association with CpG islands, recognises a subset of silent developmental genes in ES cells in order to recruit the CDK-Mediator complex and prime these genes for future activation.\n\n![FBXL19 target genes rely on CDK-Mediator for activation during differentiation.\\\n(**A**) A schematic illustrating the OHT treatment to generate MED13/13L KO ES cells and differentiation approach. (**B**) Western blot analysis showing the efficiency of MED13/13 knock-out upon 96 hr OHT treatment of *Med13/13l^fl/fl^* ES cells. Suz12 was blotted as a loading control. (**C**) ChIP-qPCR showing CDK8 enrichment in WT and MED13/13L KO ES cells. Enrichment is relative to gene desert control region. Error bars show standard deviation of two biological replicates. (**D**) RT-qPCR gene expression analysis in WT and MED13/13L KO ES cells before and after RA induction. Expression was normalized to the expression of the PolIII-transcribed gene *tRNA-Lys* and is represented as relative to WT ES cells (for pluripotency genes) or RA-treated WT cells (for differentiation markers). Error bars show SEM of three biological replicates, asterisks represent statistical significance calculated by Student T-test: \\*p\\<0.05, \\*\\*p\\<0.01, \\*\\*\\*p\\<0.001.](elife-37084-fig7){#fig7}\n\nAblating the capacity of FBXL19 to bind CpG islands causes embryonic lethality {#s2-8}\n------------------------------------------------------------------------------\n\nOur results suggest that FBXL19 contributes to gene activation during cell lineage commitment via recognition of CpG islands and recruitment of the CDK-Mediator complex. Given that *Fbxl19^\u0394CXXC^* cells display impaired gene activation in our *in vitro* differentiation model, we asked whether the inability of FBXL19 to bind and function at CpG islands could also affect mouse development. In order to address this, we generated *Fbxl19^\u0394CXXC^* mutant mice by crossing *Fbxl19^fl/fl^* animals with animals constitutively expressing Cre recombinase. From these crosses, we failed to obtain any viable *Fbxl19^\u0394CXXC^* homozygous offspring, indicating that removal of the ZF-CxxC domain of FBXL19 leads to embryonic lethality ([Figure 8A](#fig8){ref-type=\"fig\"}). We then investigated at which stage *Fbxl19^\u0394CXXC^* homozygous embryos were affected and found homozygous embryos were observed at normal Mendelian ratios until 10.5 days postcoitum (dpc) ([Figure 8A](#fig8){ref-type=\"fig\"}). At 9.5 dpc, gross embryonic morphology appeared to be intact but at 10.5 dpc ([Figure 8B](#fig8){ref-type=\"fig\"}) the embryos exhibited a clear growth retardation and showed, to varying extents, reduced elongation of the trunk, hypomorphic limb buds and cephalic region. This included undeveloped facial mesenchyme, including defects in maxillary and mandibular components of first branchial arches (indicated by yellow and red arrows in [Figure 8B](#fig8){ref-type=\"fig\"}, respectively), and hypomorphic cardiac mesenchyme. Some living *Fbxl19^\u0394CXXC^* homozygous embryos, characterized by a beating heart, were observed at 12.5 dpc but they were of similar size and external appearance to 10.5 dpc mutant embryos suggesting that normal development had ceased by this point. Together our findings demonstrate that FBXL19, and its ability to recognize CpG islands, is essential for normal mouse embryonic development.\n\n![Deletion of the ZF-CxxC domain of FBXL19 leads to mouse embryonic lethality.\\\n(**A**) After crossing heterozygotes, the number of live embryos (9.5 dpc or 10.5 dpc) or newborn pups with the indicated genotypes is indicated with the percentages shown in parentheses. (**B**) The morphological changes in 10.5 dpc *Fbxl19^CXXC\u0394/\u0394^* embryos are indicated. Lateral views of wild-type (wt), together with heterozygous (*Fbxl19^CXXC\u0394/+^*) and homozygous (*Fbxl19^CXXC\u0394/\u0394^*) mutants are shown. Maxillary and mandibular components of first branchial arches are indicated by yellow and red arrows, respectively.](elife-37084-fig8){#fig8}\n\nDiscussion {#s3}\n==========\n\nHere, we discover that FBXL19 recognizes CpG islands throughout the genome in a ZF-CxxC-dependent manner ([Figure 1](#fig1){ref-type=\"fig\"}). Unlike other ZF-CxxC proteins which associate with chromatin-modifying complexes, we show that FBXL19 interacts with the CDK-Mediator complex ([Figure 2](#fig2){ref-type=\"fig\"}). This uncovers an unexpected link between CpG islands and a complex that regulates gene expression through interfacing with the transcriptional machinery. We demonstrate that FBXL19 can recruit CDK-Mediator to chromatin ([Figure 3](#fig3){ref-type=\"fig\"}) and, via recognition of CpG islands, plays an interesting role in supporting CDK8 occupancy at a subset of promoters associated with developmental genes, which are inactive in ES cells ([Figure 4](#fig4){ref-type=\"fig\"} and [Figure 5](#fig5){ref-type=\"fig\"}). At these CpG islands, FBXL19 and CDK-Mediator function to prime the associated genes for activation during ES cell differentiation ([Figure 6](#fig6){ref-type=\"fig\"} and [Figure 7](#fig7){ref-type=\"fig\"}). Consistent with an important role of FBXL19 in supporting normal developmental gene expression, removal of the ZF-CxxC domain of FBXL19 leads to perturbed development and embryonic lethality in mice ([Figure 8](#fig8){ref-type=\"fig\"}). Together these new discoveries reveal that CpG islands and FBXL19 can interface with CDK-Mediator to orchestrate normal gene expression during lineage commitment.\n\nPreviously, another F-box protein, FBW7, was shown to function as an E3 ubiquitin ligase substrate selector for the MED13/13L subunit of CDK-Mediator and to regulate its stability through proteasomal degradation ([@bib22]). By controlling CDK-Mediator abundance, one could envisage how this might shape Mediator function in gene expression. FBXL19 also encodes an F-box domain and associates with SKP1, a central component of SCF E3 ubiquitin ligase complexes. However, we do not find evidence that FBXL19 regulates CDK-Mediator stability via the proteasome ([Figure 2---figure supplement 1H](#fig2s1){ref-type=\"fig\"}). Instead, we discover that FBXL19 can recruit CDK-Mediator to chromatin and, more specifically, to CpG islands via its ZF-CxxC domain to support gene activation. Although we currently do not know the defined subunits and surfaces in CDK-Mediator that FBXL19 interacts with, our biochemical experiments indicate that this relies on an intact F-box domain in FBXL19 ([Figure 2---figure supplement 1E](#fig2s1){ref-type=\"fig\"}). It is tempting to speculate that FBXL19 could interact directly with MED13/13L given their recognition by the related F-box protein, Fbw7 ([@bib22]). Nevertheless, a key observation in our work is that FBXL19 appears to have functionally diverged from other F-box proteins during vertebrate evolution by acquiring a DNA-binding domain that allows it to recruit proteins to chromatin, instead of targeting them for ubiquitylation. This general feature is shared with the FBXL19 paralogue, KDM2B, which associates with and recruits the PRC1 complex to CpG islands ([@bib28]; [@bib34]; [@bib85]). In agreement with our observations, a large-scale proteomics screen previously suggested that an interaction between FBXL19 and Mediator may also exist in cancer cells ([@bib77]). Given that CDK8 can function as an oncogene ([@bib1]; [@bib29]; [@bib59]), it will be interesting to understand whether FBXL19 plays a role in targeting CDK-Mediator to CpG islands in non-embryonic tissues, and whether this activity may promote CDK8-driven tumorigenesis.\n\nIndividual Mediator subunits have been shown to interact with DNA-binding transcription factors that recruit the Mediator complex to specific DNA sequences in regulatory elements and support gene expression ([@bib11]; [@bib16]; [@bib30]; [@bib56]). In part, Mediator is thought to achieve this by bridging enhancer elements to the core gene promoter and RNAPolII ([@bib4]; [@bib39]; [@bib64]). However, in mammals these proposed mechanisms are extrapolated from studying only a subset of transcription factors and genes. Therefore, the defined mechanisms that shape Mediator occupancy on chromatin remain very poorly understood. Here, we provide evidence for a completely new gene promoter-associated CDK-Mediator targeting mechanism that relies on FBXL19 and CpG island recognition. Surprisingly, although FBXL19 localizes broadly to CpG islands throughout the genome ([Figure 1](#fig1){ref-type=\"fig\"}), we only observed a reliance on FBXL19 for CDK8 binding at a subset of sites ([Figure 4](#fig4){ref-type=\"fig\"}). Similarly, KDM2B binds broadly to CpG islands, yet has a specificity in shaping PRC1 recruitment and polycomb repressive chromatin domain formation at a subset of developmental genes in mouse ES cells ([@bib28]). We have previously suggested that this could result from ZF-CxxC domain-containing proteins broadly sampling CpG islands, with their activity or affinity for certain sites being shaped by local gene activity or chromatin environment ([@bib45]). In keeping with these general ideas, FBXL19 is required for appropriate CDK8 binding at bivalent genes in ES cells ([Figure 5---figure supplement 1A,B](#fig5s1){ref-type=\"fig\"}). In future work, it will be important to understand whether this unique chromatin state and/or the absence of transcriptional activity at these sites define the requirement for FBXL19 in CDK8 recruitment, or whether FBXL19 targeting occurs at most CpG islands sites but is masked through redundancy with transcription factor and gene activity-dependent targeting modalities.\n\nHistorically, CDK-Mediator has been associated with gene repression ([@bib2]; [@bib20]; [@bib27]; [@bib35]; [@bib46]; [@bib62]). Therefore, we were not surprised to observe that CDK8 occupied the promoters of repressed developmental genes in ES cells. While abrogating the CpG island-binding activity of FBXL19 or deletion of MED13/13L resulted in reduced CDK8 binding at these sites, it did not lead to an appreciable effect on gene expression in the ES cell state ([Figure 6](#fig6){ref-type=\"fig\"} and [Figure 6---figure supplement 1C,D](#fig6s1){ref-type=\"fig\"}). This suggests that CDK-Mediator is not required for the repression of these genes in ES cells. Instead, these genes fail to properly activate during differentiation ([Figure 6](#fig6){ref-type=\"fig\"}, [Figure 6---figure supplement 1](#fig6s1){ref-type=\"fig\"}, and [Figure 7D](#fig7){ref-type=\"fig\"}). These observations are in line with several reports that support the idea that CDK-Mediator is also involved in gene activation ([@bib3]; [@bib26]; [@bib36]; [@bib25]; [@bib32]). Therefore, in the context of mouse ES cells, we propose that CDK-Mediator may be required to prime silent developmental genes for future gene activation through a mechanism that relies on the recognition of CpG island promoters by FBXL19. This priming could contribute to gene activation during differentiation through one of the several mechanisms by which CDK-Mediator has been proposed to affect gene transcription, including regulating RNAPolII pre-initiation complex assembly, polymerase pausing/elongation, or through mediating long range interactions with distal regulatory elements ([@bib4]; [@bib8]; [@bib25]; [@bib32]; [@bib40]). Clearly, understanding the defined mechanism by which FBXL19-dependent CDK-Mediator recruitment supports normal gene activation during differentiation remains an important question for future work.\n\nOur finding that FBXL19 can target CDK8 to gene promoters and that this is required to prime the expression of developmental genes during ES cell differentiation is conceptually reminiscent of recent work in yeast and human systems which suggested that CDK8 is required for appropriate re-induction of inducible genes following an initial activation stimulus ([@bib21]). In this study, CDK8 was found to associate with the promoters of inducible genes following initial activation, even in the absence of appreciable ongoing gene transcription, thereby acting as a form of transcriptional memory ([@bib21]). Interestingly, this requirement for CDK8 in normal gene activation appears shared in the context of our developmental gene induction paradigm. However, we find that, in the case of developmental genes, previous gene activation may not be required for the recruitment of CDK8. Instead, this could be achieved by FBXL19 directly recognising and targeting CDK8 to CpG islands. Therefore, we prefer to view this as priming as opposed to memory. Nevertheless, collectively these observations point to an important role for CDK-Mediator binding to gene promoters prior to activation and as a way of supporting subsequent gene induction. This provides new evidence that the roles of CDK-Mediator in gene regulation are more complicated than simply conveying activation signals to RNAPolII through recruitment by transcription factors and suggests the\u00a0complex\u00a0may have evolved to play a unique role in regulating gene expression in stem cells and during development.\n\nMaterials and methods {#s4}\n=====================\n\nCell culture {#s4-1}\n------------\n\nMouse ES cells were cultured on gelatine-coated dishes in DMEM (Thermo Fisher\u00a0scientific) supplemented with 15% fetal bovine serum (BioSera), L-Glutamine, beta-mercaptoethanol, non-essential amino acids, penicillin/streptomycin (Thermo Fisher\u00a0scientific) and 10 ng/mL leukemia-inhibitory factor. *Fbxl19^fl/fl^* ES cells were treated with 800 nM 4-hydroxytamoxifen (Sigma) for 96 hr in order to delete the ZF-CxxC domain. For RA differentiation of ES cells, 2.5 \u00d7 10^4^ cells/cm^2^ were allowed to attach to gelatinised dishes (\\~12 hr) and treated with 1 \u00b5M retinoic acid (Sigma-Aldrich) in EC-10 medium (DMEM supplemented with 10% fetal bovine serum, L-Glutamine, beta-mercaptoethanol, non-essential amino acids and penicillin/streptomycin) for 72 hr. For embryonic body differentiation, 2 \u00d7 10^6^ cells were plated on non-adhesive 10 cm dishes in EC-10 medium and cultured for the indicated days. For generation of stable cell lines, E14 ES cells were transfected using Lipofectamine 2000 (Thermo Fisher\u00a0scientific) following manufacturer's instructions. Stably transfected cells were selected for 10 days using 1 \u03bcg/ml puromycin and individual clones were isolated and expanded in the presence of puromycin to maintain the transgene expression. TOT2N E14 cells used for TetR targeting experiments were previously described ([@bib12]). 293 T cells were cultured in EC-10 media. Transient overexpression of FBXL19 was performed by transfecting 293 T cells using Lipofectamine 2000 (Thermo Fisher\u00a0scientific) followed by selection with 400 ng/\u03bcl G418 for 48 hr. Proteasome inhibition was performed for 4 hr using 10 \u00b5M MG132 inhibitor (Sigma-Aldrich). All cell lines generated and grown in the Klose and Koseki labs were routinely tested for mycoplasma infection.\n\nGeneration of *Fbxl19^\u0394CXXC^* conditional knock-out mouse {#s4-2}\n---------------------------------------------------------\n\nThe targeting construct was generated from C57BL6/J mouse genomic sequence spanning mm9 chromosome 7: 134,888,487--134,895,358 containing the exons 1 to 6 of *Fbxl19* genomic region. Recombination was carried out by Gateway system (Life Technologies). One of the loxP sequences was inserted at mm9 chr7:134,891,537, and FRT flanked PGK-neo was inserted at mm9 chr7:134,892,000 together with an additional loxP sequence. The targeting construct was electroporated into M1 ES cells to obtain targeted insertion. Clones of targeted ES cells were aggregated with eight-cell embryos to generate the targeted mouse line. The *Fbxl19^fl/fl^* line was generated by removal of the PGK-neo marker gene by mating the targeted mice with mice expressing FLP recombinase. These *Fbxl19^fl/fl^* mice were further mated with mice harboring the *ROSA26-CreErt2* locus to generate *Fbxl19*^fl/fl^:*ROSA26-CreErt2*^+/-^ mice, from which the *Fbxl19^fl/fl^* ES cells used in this study were derived.\n\nGeneration of *Med13/13l^fl/fl^* conditional knock-out\u00a0ES cells line {#s4-3}\n--------------------------------------------------------------------\n\nIn order to generate a conditional *Med13*/*13l^fl/fl^* ES cell line, we inserted loxP sites downstream of exons 7\u00a0and upstream of exons 8 of the *Med13* and *Med13l* genes. The targeting constructs for the insertion of each loxP site were designed to have 150 bp homology arms flanking the loxP site and to carry a mutated PAM sequence to prevent retargeting by the Cas9 enzyme. The targeting constructs were purchased from GeneArt Gene Synthesis (Thermo\u00a0Fisher scientific). The pSpCas9(BB)\u22122A-Puro(PX459)-V2.0 vector was obtained by Addgene (\\#62988). sgRNAs were designed to specifically target the desired genomic region for each loxP insertion (<http://crispor.tefor.net/crispor.py>) and were cloned into the Cas9 vector as previously described ([@bib66]). ES cells that express the ERT2-Cre recombinase from the *ROSA26* locus (*ROSA26:ERT2-Cre*) were used. First, the downstream loxP sites for *Med13* and *Med13l* were targeted. *ROSA26:ERT2-Cre* ES cells were transiently co-transfected with 1 \u03bcg of each Cas9-sgRNA plasmid and 3.5 \u03bcg of each targeting construct using Lipofectamine 3000 (Thermo\u00a0Fischer Scientific). Successfully transfected ES cells were selected for 48 hr with 1 \u03bcg/ml puromycin. Individual clones were screened by genotyping PCR to identify correctly targeted homozygous clones. A clone homozygous for both *Med13* and *Med13l* LoxP1 sites was then used to target the upstream loxP sites using the same transfection protocol and screening strategy. Correct loxP targeting was verified by sequencing of the genomic region surround the loxP sites.\n\nDNA constructs {#s4-4}\n--------------\n\nFor generation of FBXL19 expression constructs, the full length, \u0394CxxC or \u0394F-box cDNA of mouse *Fbxl19* (IMAGE ID 6401846, Source Bioscience) was PCR amplified and inserted into a pCAG-IRES-FS2 vector ([@bib28]) via ligation-independent cloning (LIC). Mutation of the ZF-CxxC domain of FBXL19 (K49A) was generated via site-directed mutagenesis using the Quikchange mutagenesis XL kit (Stratagene). To generate TetR-FBXL19 fusion expression plasmid, *Fbxl19* cDNA was cloned into pCAG-FS2-TetR ([@bib12]) via LIC. For transient overexpression in 293 T cells, full length FBXL19 cDNA was cloned into pcDNA3-2xFlag vector by conventional cloning. All plasmids were sequence-verified by sequencing.\n\nNuclear extract preparation and immunoprecipitation {#s4-5}\n---------------------------------------------------\n\nCells were harvested by scraping in PBS at 4\u00b0C, resuspended in 10x pellet volume (PV) of Buffer A (10 mM Hepes pH 7.9, 1.5 mM MgCl~2~, 10 mM KCl, 0.5 mM DTT, 0.5 mM PMSF, cOmplete protease inhibitor cocktail (Roche)) and incubated for 10 min at 4\u00b0C with slight agitation. After centrifugation, the cell pellet was resuspended in 3x PV Buffer A containing 0.1% NP-40 and incubated for 10 min at 4\u00b0C with slight agitation. Nuclei were recovered by centrifugation and the soluble nuclear fraction was extracted for 1 hr at 4\u00b0C with slight agitation using 1x PV Buffer C (10 mM Hepes pH 7.9, 400 mM NaCl, 1.5 mM MgCl~2~, 26% glycerol, 0.2 mM EDTA, cOmplete protease inhibitor cocktail). Protein concentration was measured using Bradford assay.\n\nFor small-scale co-immunoprecipitation, 600 \u00b5g of nuclear extract was diluted in BC150 buffer (50 mM Hepes pH 7.9, 150 mM KCl, 0.5 mM EDTA, 0.5 mM DTT, cOmplete protease inhibitor cocktail). Samples were incubated with the respective antibodies and 25U benzonase nuclease overnight at 4\u00b0C. Protein A agarose beads (RepliGen) were blocked for 1 hr\u00a0at 4\u00b0C in Buffer BC150 containing 1% fish skin gelatine (Sigma) and 0.2 mg/ml BSA (NEB). The blocked beads were added to the samples and incubated for 4 hr at 4\u00b0C. Washes were performed using BC150 containing 0.02% NP-40. The beads were resuspended in 2x SDS loading buffer and boiled for 5 min to elute the immunoprecipitated complexes.\n\nPurification of FBXL19-FS2 was performed using StrepTactin resin (IBA) as previously described with the exception of the wash buffer used (20 mM Tris pH 8.0, 150 mM NaCl, 0.2% NP-40, 1 mM DTT, 5% glycerol, cOmplete protease inhibitor cocktail) ([@bib28]). Between 10\u00a0and\u00a015 mg of nuclear extract was used for each large scale purification. The samples were treated with 75 U/mL benzonase nuclease (Novagen) in order to disrupt nucleic-acid-mediated interactions.\n\nMass spectrometry {#s4-6}\n-----------------\n\nSamples from FBXL19-FS2 affinity purifications were subjected to in-solution trypsin digestion and mass spectrometry analysis was performed as described previously ([@bib28]). Two biological replicates were performed. A control EV purification was included in each replicate. An interaction with identified proteins was only considered significant if absent from the EV data.\n\nAntibodies {#s4-7}\n----------\n\nAntibodies used for IP were rabbit anti-MED12 (A300-774A, Bethyl laboratories), rabbit anti-CDK8 (A302-500A, Bethyl laboratories), rabbit anti-HA (3724, Cell Signaling Technology). A polyclonal antibody against FBXL19 was prepared in-house by rabbit immunisation (PTU/BS Scottish National Blood Transfusion Service) with a recombinant peptide encoding for amino acids 137--336 of mouse FBXL19 protein. The FBXL19 peptide antigen was coupled to Affigel 10 resin (BioRad) and the antibody was affinity-purified and concentrated.\n\nAntibodies used for Western blot analysis were mouse anti-Flag M2 (F1804, Sigma-Aldrich), mouse anti-SKP1 (sc-5281, Santa Cruz), goat anti-CDK8 (sc-1521, Santa Cruz), rabbit anti-MED12 (A300-774A, Bethyl laboratories), rabbit anti-MED13L (A302-420A, Bethyl laboratories), rabbit anti-MED13 (GTX129674, Genetex), rabbit anti-MED1, (A300-793A, Bethyl laboratories), rabbit anti-MED26 (A302-370A, Bethyl laboratories), rabbit anti-T7 (13246, Cell Signaling Technology), rabbit anti-TBP (ab818, Abcam), rabbit anti-RNF20 (11974, Cell Signaling Technology), rabbit anti-ubiquityl-Histone H2B (Lys120) (5546, Cell Signaling Technology), and mouse anti-ubiquityl-Histone H2B (Lys120) (05--1312, Millipore).\n\nGeneration of T7-FBXL19 ES cell line {#s4-8}\n------------------------------------\n\nAs the FBXL19 antibody failed to work reliably for Western blot analysis, we generated an *Fbxl19^fl^*^/*fl*^ ES cell line in which the endogenous *Fbxl19* gene is tagged with a 3xT7-2xStrepII tag by CRISPR/Cas9 knock-in ([Figure 2---figure supplement 1B](#fig2s1){ref-type=\"fig\"}).\u00a0This allowed us to determine the efficiency of the OHT treatment and endogenous IPs. The tag was synthesised by GeneArt (ThermoFischer Scientific) and the targeting construct was generated by PCR amplification to introduce roughly 150 bp homology arms flanking the 3xT7-2xStrepII tag. The PCR product was cloned into pGEM-T Easy Vector (Promega). The pSpCas9(BB)\u22122A-Puro vector was obtained by Addgene (\\#48139). A sgRNA was designed to overlap with the stop codon of the *Fbxl19* gene (<http://crispr.mit.edu/>) and cloned into the Cas9 vector as previously described ([@bib66]). *Fbxl19^fl/fl^* ES cells were transiently co-transfected with 1 \u03bcg of Cas9-sgRNA plasmid and 3.5 \u03bcg of targeting construct using Lipofectamine 3000 (ThermoFischer Scientific). Successfully transfected ES cells were selected for 48 hr with 1 \u03bcg/ml puromycin. Individual clones were screened by Western blot and genotyping PCR to identify homozygous targeted clones.\n\nChromatin immunoprecipitation {#s4-9}\n-----------------------------\n\nChromatin immunoprecipitation was performed as described previously ([@bib28]) with slight modifications. Cells were fixed for 45 min with 2 mM DSG (Thermo scientific) in PBS and 12.5 min with 1% formaldehyde (methanol-free, Thermo scientific). Reactions were quenched by the addition of glycine to a final concentration of 125 \u00b5M. After cell lysis and chromatin extraction, chromatin was sonicated using a BioRuptor Pico sonicator (Diagenode), followed by centrifugation at 16,000\u00a0\u00d7\u00a0*g* for 20 min at 4\u00b0C. 200 \u00b5g chromatin diluted in ChIP dilution buffer (1% Triton-X100, 1 mM EDTA, 20 mM Tris-HCl (pH 8.0), 150 mM NaCl) was used per IP. Chromatin was precleared with protein A Dynabeads blocked with 0.2 mg/ml BSA and 50 \u00b5g/ml yeast tRNA and incubated with the respective antibodies overnight at 4\u00b0C. Antibody-bound chromatin was purified using blocked protein A Dynabeads for 3 hr at 4\u00b0C. ChIP washes were performed as described previously ([@bib28]). ChIP DNA was eluted in ChIP elution buffer (1% SDS, 100 mM NaHCO~3~) and reversed cross-linked overnight at 65\u00b0C with 200 mM NaCl and RNase A (Sigma). The reverse cross-linked samples were treated with 20 \u03bcg/ml Proteinase K and purified using ChIP DNA Clean and Concentrator kit (Zymo research).\n\nThe antibodies used for ChIP experiments were rabbit anti-FS2 ([@bib28]), rabbit anti-CDK8 (A302-500A, Bethyl laboratories), rabbit anti-MED12 (A300-774A, Bethyl laboratories), mouse anti-MED4 (PRC-MED4-16, DSHB), rabbit anti-SKP1 (12248, Cell Signaling Technology), rabbit anti-KDM2A ([@bib28]), rabbit anti-KDM2B ([@bib28]).\n\nChIP-sequencing {#s4-10}\n---------------\n\nAll ChIP-seq experiments were carried out in biological duplicates. ChIP-seq libraries for FS2-FBXL19 ChIP were prepared using the NEBNext Fast DNA fragmentation and library preparation kit for Ion Torrent (NEB, E6285S) following manufacturer's instructions. Briefly, 30--40 ng ChIP or input DNA material was used. Libraries were size-selected for 250 bp fragments using 2% E-gel SizeSelect gel (Thermo scientific) and amplified with 6 PCR cycles or used without PCR amplification. Templates were generated with Ion PI Template OT2 200 kit v3 and Ion PI Sequencing 200 kit v3, or with Ion PI IC 200 kit (Thermo scientrific). Libraries were sequenced on the Ion Proton Sequencer using Ion PI chips v2 (Thermo scientific).\n\nChIP-seq libraries for CDK8 ChIP were prepared using the NEBNext Ultra DNA Library Prep Kit, and sequenced as 40 bp paired-end reads on Illumina NextSeq500 platform using NextSeq 500/550 (75 cycles).\n\nReverse transcription and gene expression analysis {#s4-11}\n--------------------------------------------------\n\nTotal RNA was isolated using TRIzol reagent (Thermo scientific) following manufacturer's instructions and cDNA was synthesized from 400 ng RNA using random primers and ImProm-II Reverse Transcription system kit (Promega). RT-qPCR was performed using SensiMix SYBR mix (Bioline). Idh1 and Atp6IP1 genes were used as house-keeping controls.\n\n4sU-labeling of nascent RNA transcripts {#s4-12}\n---------------------------------------\n\nFor labeling of nascent RNA transcripts, cells were grown in 15 cm culture dishes. Labeling was performed for 20 min at 37\u00b0C by the addition of 50 mM 4-thiouridine (T4509, Sigma) to the culture medium. After the incubation, the medium was removed and RNA was isolated using TRIzol reagent (Thermo scientific) following manufacturer's instructions. The total RNA was treated with TURBO DNA-free kit (Ambion, Thermo scientific) in order to remove contaminating genomic DNA. 300 \u00b5g RNA was biotinylated using 600 \u00b5g Biotin-HPDP (21341, Pierce, Thermo scientific) in Biotinylation buffer (100 mM Tris-HCl pH 7.4, 10 mM EDTA). The reaction was carried out for one and half hour on a rotor at room temperature. Unincorporated Biotin-HPDP was removed by chloroform/isoamylalcohol (24:1, Sigma) wash followed by isopropanol precipitation of the biotinylated RNA. Labeled biotinylated RNA was isolated using \u00b5Macs Streptavidin Kit (130-074-101, Miltenyi) following manufacturer's instructions and purified using RNeasy MinElute Cleanup kit (Qiagen). The quality of the RNA was confirmed using the RNA pico Bioanalyser kit (Agilent).\n\n4sU RNA sequencing {#s4-13}\n------------------\n\nUp to 1 \u00b5g purified 4sU-labelled RNA was used to prepare libraries for 4sU-RNA-seq. Ribosomal RNA was removed using NEBNext rRNA Depletion Kit and libraries were prepared using NEBNext Ultra Directional RNA Library Prep Kit for Illumina (NEB) following manufacturer's instructions. Library quality was assessed using the High-sensitivity DS DNA Bioanalyser kit (Agilent) and the concentration was measured by quantitative PCR using KAPA Library quantification standards for Illumina (KAPA Biosystems). All 4sU RNA-seq experiments were carried out in biological triplicates. 4sU RNA-seq libraries were sequenced as 80 paired-end reads on Illumina NextSeq500 platform using NextSeq 500/550 High Output Kit v2 (150 cycles).\n\nAnalysis of high-throughput sequencing {#s4-14}\n--------------------------------------\n\nSequencing reads were aligned to the mouse genome (mm10) using bowtie2 ([@bib47]) with the '\\--no-mixed' and '\\--no-discordant' options. Reads that mapped more than once to the genome were discarded. For 4sU RNA-seq analysis, rRNA reads were initially removed by aligning the data to mouse rRNA genomic sequences (GenBank: BK000964.3) using bowtie2. The rRNA-depleted reads were next aligned against mm10 genome using the STAR RNA-seq aligner ([@bib23]). To improve mapping of nascent, intronic sequences, a second alignment step with bowtie2 was included using the reads which failed to map using STAR. PCR duplicates were removed using samtools ([@bib51]). Biological replicates were randomly downsampled to the same number of reads for each individual replicate and merged for visualisation. Bigwig files were generated using MACS2 ([@bib87]) and visualised using the using the UCSC Genome Browser ([@bib67]).\n\nPeak calling was performed using the MACS2 function with the '-broad' option using the biological replicates with matched input ([@bib87]). Peaks mapping to a custom 'blacklist' of artificially high genomic regions were discarded. Only peaks called in both replicates were considered. Differential analysis of CDK8 binding was done using DiffReps ([@bib71]) and the called differential regions were overlapped with CDK8 peaks. Changes in binding of log2FC less than \u22120.5 or more than 0.5 with adjusted p-value below 0.05 were considered significant.\n\nDifferential expression analysis was performed using the DESeq2 package ([@bib55]), version 1.6.3, with R version 3.1.1. Counts were quantified using the summarizeOverlaps() function in R in the mode 'Union' and inter.feature\u00a0=\u00a0FALSE. Genes with an adjusted p-value of below 0.1 and a fold change of at least 1.5 were considered differentially expressed. Statistical analysis was performed using two-sample Wilcoxon rank sum test. Gene ontology analysis was done using DAVID 6.8 ([@bib37]). For CDK8 differentially\u00a0bound peaks, all CDK8 peaks were provided as background. For misregulated genes analysis, all RefSeq genes were provided as background. ATAC peaks were obtained from [@bib43]).\n\nAccession numbers {#s4-15}\n-----------------\n\nChIP-seq and RNA-seq data from the present study are available to download at GSE98756. Previously published studies used for analysis include mouse ES cell BioCAP (GSE43512, \\[[@bib54]\\]), Kdm2B ChIP-seq (GSE55698, \\[[@bib12]\\]), Kdm2A ChIP-seq (GSE41267, \\[[@bib28]\\]), H3K27me3 ChIP-seq (GSE83135, \\[[@bib68]\\]), H3K4me3 ChIP-seq (GSE93538, \\[[@bib18]\\]).\n\nFunding Information\n===================\n\nThis paper was supported by the following grants:\n\n-   Sir Henry Wellcome Postdoctoral Fellowship 110286/Z/15/Z to Angelika Feldmann.\n\n-   http://dx.doi.org/10.13039/100009619Japan Agency for Medical Research and Development to Haruhiko Koseki.\n\n-   http://dx.doi.org/10.13039/100004440Wellcome 098024/Z/11/Z to Robert J Klose.\n\n-   http://dx.doi.org/10.13039/501100000781European Research Council 681440 to Robert J Klose.\n\n-   http://dx.doi.org/10.13039/501100001255Lister Institute of Preventive Medicine to Robert J Klose.\n\nWe thank Anne Turberfield for help with the 4sU-RNAseq protocol and useful discussion, and Hamish King for assistance with computational analysis and critical comments and suggestions. We thank David Brown for initiating the FBXL19 experiments, and Anne Turberfield and Neil Blackledge for critical reading of the manuscript. We thank Ed Hookway and Udo Oppermann for sequencing support on the NextSeq500 at the Botnar sequencing facility in Oxford. Work in the Klose laboratory is supported by the Wellcome Trust, the Lister Institute of Preventive Medicine and the European Research Council. TK and HK are supported by the AMED-CREST programme from the Japan Agency for Medical Research and Development. AF is supported by a Sir Henry Wellcome Postdoctoral Fellowship.\n\nAdditional information {#s5}\n======================\n\nNo competing interests declared.\n\nConceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing---original draft, Writing---review and editing.\n\nResources, Data curation, Formal analysis, Methodology.\n\nResources, Software, Formal analysis.\n\nResources, Contributed to transgenic mouse generation and embryonic stem cell derivation and characterisation.\n\nResources, Contributed to transgenic mouse generation and embryonic stem cell derivation and characterisation.\n\nResources, Contributed to mass spectrometry sample processing and analysis.\n\nResources, Contributed to mass spectrometry sample processing and analysis.\n\nSupervision, Funding acquisition, Investigation, Project administration.\n\nConceptualization, Supervision, Funding acquisition, Investigation, Writing---original draft, Project administration, Writing---review and editing.\n\nAdditional files {#s6}\n================\n\n10.7554/eLife.37084.019\n\nData availability {#s7}\n-----------------\n\nSequencing data generated in this study have been deposited in GEO under accession code GSE98756.\n\nThe following dataset was generated:\n\nDimitrova E2017FBXL19 recruits CDK8-Mediator to CpG islands and primes developmental genes for activation during lineage commitment<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE98756>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE98756)\n\nThe following previously published datasets were used:\n\nBlackledge NPFarcas AMKondo TKing HWMcGouran JFHanssen LLIto SCooper SKondo KKoseki YIshikura TLong HKSheahan TWBrockdorff NKessler BMKoseki HKlose RJ2014Variant PRC1 complex dependent H2A ubiquitylation drives PRC2 recruitment and polycomb domain formation<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE55698>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE55698)\n\nFarcas AMBlackledge NPSudbery ILong HKMcGouran JFRose NRLee SSims DCerase ASheahan TWKoseki HBrockdorff NPonting CPKessler BMKlose RJ2012KDM2B links the Polycomb Repressive Complex 1 (PRC1) to recognition of CpG islands<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE41267>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE41267)\n\nLong HKSims DHeger ABlackledge NPKutter CWright MLGr\u00fctzner FOdom DTPatient RPonting CPKlose RJ2013Epigenetic conservation at gene regulatory elements revealed by non-methylated DNA profiling in seven vertebrates<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE43512>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE43512)\n\nRose NRKing HWBlackledge NPFursova NAEmber KJFischer RKessler BMKlose RJ2016RBYP stimulates PRC1 to shape chromatin-based communication between polycomb repressive complexes<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE83135>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE83135)\n\nBrown DADi Cerbo VFeldmann AAhn JIto SBlackledge NPNakayama MMcClellan MDimitrova ETurberfield AHLong HKKing HWKriaucionis SSchermelleh LKutateladze TGKoseki HKlose RJ2017CFP1 engages in multivalent interactions with CpG island chromatin to recruit the SET1 complex and regulate gene expression.<https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE93538>Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE93538)\n\n10.7554/eLife.37084.036\n\nDecision letter\n\nEspinosa\n\nJoaqu\u00edn M\n\nReviewing Editor\n\nUniversity of Colorado School of Medicine\n\nUnited States\n\nIn the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.\n\n\\[Editors' note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.\\]\n\nThank you for submitting your work entitled \\\"FBXL19 recruits CDK8-Mediator to CpG islands and primes developmental genes for activation during lineage commitment\\\" for consideration by *eLife*. Your article has been reviewed by four peer reviewers, one of whom is a member of our Board of Reviewing Editor and the evaluation has been overseen by a Senior Editor.\n\nThe reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this letter to crystallize our concerns going forward. The work presented was deemed important and interesting but key issues reduced enthusiasm for the manuscript considerably. Given that these revisions are substantial and likely to take a significant amount of time, we are currently declining the present version of the paper. That said, if you feel you can successfully address all the concerns raised, we will be open to consider a revised submission, which would be treated as a de novo submission but sent to the same editors and, if possible, referees for re-assessment.\n\nSummary:\n\nThe reviewers considered the manuscript to be strong, reporting a thorough characterization of FBXL19, the least-well studied of three proteins that contain a ZF-CxxC domain. The authors demonstrate that over-expressed FBXL19 accumulates in the nucleus, and that a majority of its genome-wide binding is at sites associated with CpG islands. Affinity purification of FS2-tagged FBXL19 indicated a physical association with CDK8-containing Mediator complex, and a TetR-FBXL19 fusion protein was able to recruit CDK8 to a chromosomally-integrated artificial TetO array, suggesting that FBXL19 can recruit CDK8 to chromatin. Conditional deletion of the ZF-CxxC domain of FBXL19 led to reduction of CDK8 binding at CpG islands in ES cells. Many of these sites are associated with silent developmental genes that are no longer activated during ES cell differentiation in the FBXL19 mutants. Finally, deletion of the ZF-CxxC domain was shown to be embryonic lethal in mice. Together, these findings suggest that FBXL19 may prime developmental genes for subsequent activation during differentiation, possibly involving recruitment of CDK8. Overall, this represents an interesting finding and potentially a novel role for CDK8. However, concerns about the degree to which this represents a significant advance over previous FBXL19 studies, some missing controls, and a lack of a direct test of CDK8 requirement for expression of these genes resulted in lessened enthusiasm for this manuscript.\n\nThe reviewers discussed these concerns and summarized their critiques into the following points.\n\nEssential revisions:\n\n1\\) A key conclusion of the manuscript is that CDK8-Mediator is required for induction of developmental genes regulated by FBXL19, but this is actually not proven in the manuscript. The authors should deplete CDK8 (and/or CDK19, see comment \\#2 below) and test for the impact on induction of FBXL19-dependent genes.\n\n2\\) Throughout the paper, the authors refer to CDK8-Mediator. Given that there are two paralog kinases associated with the Mediator complex, CDK8 and CDK19, the authors should test for the presence of CDK19 and clarify whether the interaction is exclusive to CDK8-containing complexes or not. In Figure 2A, are the CDK8 peptides identified unique to CDK8 or could they map to CDK19 as well? In Figure 2C, please add western blot for CDK19. As for point \\#1 above, is CDK19 required for control of FBXL19-regulated genes? Additionally, the authors should demonstrate that in the \\\"EV\\\" control lane, at the similar molecular weight, these candidate hits didn\\'t show up in the cut gel, or at least with significantly less numbers of peptides.\n\n3\\) If the ZF-CxxC mediates the interaction between FBXL19 and chromatin, then which functional domain(s) on FBXL19 bind to CDK8-Mediator complex? Lee et al., claimed that the F-box is critical for the interaction between FBXL19 and RNF20. Does the same protein structure bind to CDK8?\n\n4\\) Most of the findings are based on overexpression of FBXL19 -- what is the level relative to endogenous FBXL19? High-level over-expression could drive nuclear accumulation and other observations (including Figure 3). Authors should provide data demonstrating the degree of overexpression produced in the key experiments.\n\n5\\) FBXL19 binding to CpG islands and its requirement in ES differentiation has been described previously: (https://www.ncbi.nlm.nih.gov/pubmed/28453857) \\\"Our genome-wide target mapping unveils the preferential occupancy of Fbxl19 on CpG island-containing promoters \\[...\\]\\\" \\\"Moreover, we reveal that Fbxl19 is critical for proper differentiation of ES cells \\[...\\]\\\" With respect to this, the authors should clarify the novelty of their results presented in the current manuscript.\n\n6\\) The conclusion that FBXL19 does not appear to regulate CDK8-Mediator levels is not fully supported by the data. Figure 2---figure supplement 1H needs to be repeated so that CDK8 bands are not over-exposed as there does appear to be a change in amount/mobility when overexpressed -/+ MG132. Furthermore, there seems to be a noticeable increase of MED12 protein upon FBXL19 overexpression, which was even more obvious when cells were treated with MG132.\n\n7\\) In Figure 4, CDK8 recruitment has only been tested for the FBXL19 ZF-CxxC deletion -is this recapitulated for depletion of the whole protein? Given that Figure 4B appears to show a decrease in the levels of the deletion protein this is an important point to verify.\n\n8\\) The authors need to show that OHT treatment of the delCxxC ES line leads to reduced binding of FBXL19 in conditions used for CDK8 ChIPseq -several genes would be ok -these same genes should be used to validate CDK8 CHIPseq signal.\n\n9\\) Supplemental tables with differential gene expression results should be provided -or at least working gene lists used to reach the conclusions in the manuscript. Full GO enrichment results should be included as supplementary tables so that specific gene lists can be examined.\n\n10\\) Various figure panels are missing statistical tests: Figure 4C needs statistical tests, e.g. KS/Wilcoxon -See Figure 4E; Figure 4G needs statistical tests; Figure 4H and 4I need boxplots and need statistical tests eg KS/Wilcoxon as for E/F; Figure 4H-J should include FBXL19-bound groups; Figure 6B,E,F need statistical tests.\n\n11\\) It is very nice that the authors showed that genome-wide chromatin binding of FBXL19 was completely lost when the ZF-CxxC domain was deleted or mutated. To further confirm that it is indeed this ZF-CxxC domain mediating interaction with the non-methylated CpG islands, some controls are necessary. For example, in Figure 1---figure supplement 1D, the authors showed that almost all of the FBXL19 binding sites are overlapped with NMIs locations, with only 93 exceptions. Then are the signals at these 93 sites also changed in the FBXL19 mutant ChIPSeq data? When the cell DNA methylation levels are modulated, will the binding pattern of FBXL19 changed accordingly?\n\n\\[Editors' note: what now follows is the decision letter after the authors submitted for further consideration.\\]\n\nThank you for submitting your article \\\"FBXL19 recruits CDK-Mediator to CpG islands of developmental genes priming them for activation during lineage commitment\\\" for consideration by *eLife*. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors and the evaluation has been overseen by James Manley as the Senior Editor. The reviewers have opted to remain anonymous.\n\nThe reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.\n\nThe revised manuscript by Dimitrova et al., is an excellent report describing the characterization of FBXL19, a member of the ZF-CxxC domain-containing family of proteins, and its role in recruiting CDK-Mediator to developmental genes in stem cells.\n\nAs for other members of this family, the authors show that FBXL19 recognizes unmethylated CpGs in a ZF-CxxC-dependent manner. Unlike the lysine demethylases KDM2A and KDM2B in this protein family, FBXL19 lacks a JMJ-C domain. Instead, the authors report that FBXL19 interacts with the CDK-Mediator complex. The authors then embark on an experimental tour-de-force to demonstrate that FBXL19 is required for recruitment of CDK-Mediator to a subset of genes containing CpG islands in ES cells, mostly genes involved in lineage commitment. Interestingly, these genes are transcribed at low levels in ES cells when bound by FBXL19 and CDK-Mediator, but their induction is compromised in the absence of FBXL19. This indicates that FBXL19-CDK8-Mediator is required to \\'prime\\' these genes in ES cells and their subsequent induction upon differentiation stimuli. Finally, the authors demonstrate that FBXL19 depletion is embryonic lethal. Overall, the manuscript is very well written, and the data is of excellent quality.\n\nThis is a new submission of a paper that was originally reviewed by the same referees and found to be lacking data to support some of the key conclusions. The Senior Editor opened up the possibility of a new submission without a precise timeline and the authors have taken the opportunity to address the major concerns raised in the original review. The referees are now supportive of publication of a revised manuscript addressing the following issues:\n\nEssential revisions:\n\n1\\) In AP-MS (Figure 2A), what is the background value of emPAI? Are there any other FBXL19-interacting proteins that show significantly high levels of emPAI in the list? How significant are the components of CDK-Mediator complex represented in all the hits?\n\n2\\) In Figure 3, the authors demonstrated that FBXL19 was able to recruit CDK-Mediator complex, such as CDK8 and MED12, to artificially designed chromatin regions. What about CXXC-domain or F-box deletion mutants in this assay?\n\n3\\) In Figure 4B, it is not very clear which band(s) are exactly indicated as full-length FBXL19 and which one(s) are CXXC deletion? Why are there double bands? Did the author use T7 antibody to detect the endogenously introduced protein?\n\n4\\) In Figure 4C, what proportions of each group of CDK8 binding sites overlap with FBXL19? In theory, the author expects to see that CDK8 sites showing binding intensity changes upon tamoxifen treatment would tend to overlap more with FBXL19.\n\n5\\) In Figure 4---figure supplement 1C, please use a Venn Diagram to show directly the number of peaks of CDK8, FBXI19 and NMI overlapping with each other. This would be more straightforward.\n\n6 In Figure 5, please perform the same GSEA or functional annotation analysis on genes that are associated with unchanged CDK8 binding sites upon FBXL19 CXXC-domain deletion.\n\n7\\) In Figure 6B, what is the result for genes that are associated with increased CDK8 binding signals in FBXL19 CXXC-domain deletion mutant upon RA-induced differentiation?\n\n8\\) In Figure 6C, even though the authors claimed that those upregulated genes (889) in FBXL19 CXXC-domain deletion mutant upon RA treatment were not involved in developmental processes, it is still interesting to find that significantly more genes were activated upon CXXC-domain deletion. What explanation can the authors provide for this finding?\n\n9\\) In Figure 7, the authors claim that FBXL19 relies on CDK-mediator complex to activate the specific set of developmental genes. Is it possible to test this hypothesis in the FBXL19 CXXC-deletion system they established in Figure 6? In wild-type ESC, depletion of MED13 will have similar effects as FBXL19 CXXC deletion, whereas in FBXL19 mutant, the effects would be much alleviated?\n\n10\\) In Figure 8, why did heterozygous deletion of FBXL19 CXXC-domain show even better developmental status than the wild type?\n\n11\\) The authors demonstrated that the CXXC-domain mediates the binding of FBXL19 at NMIs, where F-Box mediates its interaction with the CDK-Mediator complex. What about CDK8 chromatin binding or gene expression patterns upon F-Box deletion of FBXL19? Will the authors see the same results as what they found in CXXC-domain deletion mutant?\n\n10.7554/eLife.37084.037\n\nAuthor response\n\n\\[Editors' note: the author responses to the first round of peer review follow.\\]\n\n> Essential revisions:\n>\n> 1\\) A key conclusion of the manuscript is that CDK8-Mediator is required for induction of developmental genes regulated by FBXL19, but this is actually not proven in the manuscript. The authors should deplete CDK8 (and/or CDK19, see comment \\#2 below) and test for the impact on induction of FBXL19-dependent genes.\n\nWe agree with the reviewers that testing the impact that removal of CDK-Mediator has on the induction of FBXL19 target genes is an important experiment. To disrupt CDK-Mediator we have developed a new cell system to remove MED13 and its closely related paralogue MED13L (both of which are expressed in mouse ES cells) in an inducible manner. We chose to delete both MED13 and MED13L because they are thought to play an interchangeable role in physically linking the CDK subunits to the core Mediator and to be required to form an intact CDK-Mediator complex (Kuesel et al., 2009; Tsai et al., 2013). To achieve this, we exploited gene targeting to insert loxP sites around essential exons in both MED13 and MED13L. We also engineered these cells to express a tamoxifen inducible form of Cre recombinase allowing inducible deletion of MED13/13L (Figure 7---figure supplement 1). In new Figure 7, we demonstrate that the addition of tamoxifen leads to complete loss of MED13/13L protein. Furthermore, we show by ChIP that CDK8 no longer occupies its target sites. We then examined FBXL19 target genes after loss of CDK-Mediator binding and found that they are no longer appropriately induced during differentiation (Figure 7D), in keeping with a requirement for CDK-Mediator in FBXL19dependent gene activation. We thank the reviewers for suggesting this important experiment as it now robustly supports and strengthens our key argument that FBXL19 recruits the CDK-Mediator complex to the CpG islands of inactive developmental genes to prime them for appropriate activation during differentiation. We have now edited the text throughout the manuscript to reflect these important new observations.\n\n> 2)Throughout the paper, the authors refer to CDK8-Mediator. Given that there are two paralog kinases associated with the Mediator complex, CDK8 and CDK19, the authors should test for the presence of CDK19 and clarify whether the interaction is exclusive to CDK8-containing complexes or not. In Figure 2A, are the CDK8 peptides identified unique to CDK8 or could they map to CDK19 as well? In Figure 2C, please add western blot for CDK19. As for point \\#1 above, is CDK19 required for control of FBXL19-regulated genes?\n\nAs indicated by the reviewer, CDK8 has a highly similar paralogue, CDK19. As suggested, we have now examined our mass spectrometry results in more detail to determine if we find any peptides that also map to CDK19. From this analysis, we have identified five different peptides covering CDK8/CDK19. Four are common to both CDK8 and CDK19, and one is unique for CDK8 ([Author response image 1](#respfig1){ref-type=\"fig\"}). In our original submission, we had referred to an interaction between FBXL19 and the CDK8-Mediator complex given the identification of a peptide unique to CDK8. However, we agree that this was an oversight as one cannot exclude the possibility that FBXL19 also associates with CDK19-containing Mediator complexes. In fact, given that CDK19 is also expressed in mouse ES cells and is thought to function interchangeably with CDK8, it seems likely that FBXL19 will also interact with CDK19Mediator. Our CKD8 antibody uniquely recognises CDK8, but we have been unable to source an antibody that specifically recognises CDK19, limiting our ability to test whether CDK19 also associates with FBXL19. An antibody with specificity to CDK19 (Atlas antibody HPA007053) has previously been published, but this is currently unavailable, and the company informed us that this is due to a loss of specificity in new batches. Given that we cannot exclude the possibility that FBXL19 also associates with Mediator complexes containing CDK19, we have now added a sentence describing this possibly in the text as follows (subsection \"FBXL19 interacts with the CDK-Mediator complex in ES cells\"):\n\n*'*CDK8 and its paralogue CDK19 share 77% amino acid identity (89% similarity) (Audetat et al., 2017) and four out of the five peptides identified by mass spectrometry were common between the two proteins (data not shown). Therefore, it is likely that FBXL19 is able to interact with both CDK8- and CDK19- containing Mediator complexes.'\n\nFurthermore, because the FBXL19-associated Mediator complex could contain either CDK8 or CDK19, we have edited the text throughout the manuscript to replace 'CDK8-Mediator' with 'CDK-Mediator'. In future work it will be interesting to understand in more detail if CDK8- and CDK19-containing Mediator complexes have redundant or unique roles in gene activation during differentiation.\n\n![CDK8/19 peptide sequences identified by mass spectrometry.\\\nA CDK8 unique peptide is indicated in red. The other four peptides, indicated in yellow, are shared between CDK8 and CDK19.](elife-37084-resp-fig1){#respfig1}\n\n> Additionally, the authors should demonstrate that in the \\\"EV\\\" control lane, at the similar molecular weight, these candidate hits didn\\'t show up in the cut gel, or at least with significantly less numbers of peptides.\n\nWe carried out mass spectrometry analysis on affinity-purified FBXL19 by in-solution digestion followed by LC-MS. This was to overcome the limitations inherent to cutting bands from SDS-PAGE gels and the isolation of tryptic peptides from gel slices. Importantly, we only considered a protein to be an FBXL19 interactor in our mass spectrometry analysis if it was absent from matched EV control purifications analysed using the same in-solution approach. It would have been beneficial to have described this approach in more detail in the initial submission, so we have now expanded our description of these experiments in the revised Material and Methods. Furthermore, we validated that CDK-Mediator components were not found in the 'EV' control purifications by Western blot analysis (Figure 2C).\n\n> 3\\) If the ZF-CxxC mediates the interaction between FBXL19 and chromatin, then which functional domain(s) on FBXL19 bind to CDK8-Mediator complex? Lee et al. claimed that the F-box is critical for the interaction between FBXL19 and RNF20. Does the same protein structure bind to CDK8?\n\nWe thank the reviewers for this suggestion and agree this is an important point. To address this, we performed transient overexpression of FBXL19 followed by co-immunoprecipitation comparing wildtype FBXL19 with versions of FBXL19 in which individual domains had been removed. We found that deletion of the F-box largely disrupted the interaction with CDK-Mediator, while removing the ZF-CxxC domain had little effect on this interaction. We have included these new observations in the revised manuscript (subsection \"FBXL19 interacts with the CDK-Mediator complex in ES cells\", Figure 2---figure supplement 1E). We also attempted to determine whether the LRR domain of FBXL19 was required for this interaction but unfortunately removing this portion of the protein seemed to render it unstable (unpublished observations) limiting our capacity to draw any meaningful interpretation.\n\n> 4\\) Most of the findings are based on overexpression of FBXL19 -- what is the level relative to endogenous FBXL19? High-level over-expression could drive nuclear accumulation and other observations (including Figure 3). Authors should provide data demonstrating the degree of overexpression produced in the key experiments.\n\nWe have examined FBXL19 transcript levels in the transgene lines and they are roughly eight times that of wild type ([Author response image 2](#respfig2){ref-type=\"fig\"}). Despite exhaustive efforts we have been unable to generate an antibody that recognizes the endogenous protein in cellular extracts by Western blot. Therefore, we have been unable to directly compare endogenous FBXL19 protein levels to the FBXL19 levels in transgene containing lines. Nevertheless, we are confident in the conclusions from our key experiments. We have validated that endogenous FBXL19 is in the nucleus and interacts with CDKMediator. This was demonstrated by immunoprecipitating endogenous FBXL19 protein from nuclear extracts and using Western blot to show that FBXL19 associates with CDK8 and MED12 (Figure 2---figure supplement 1D), consistent with our transgene experiments (Figure 2).\n\nGiven that we know FBXL19 is in the nucleus and interacts with CDK-Mediator, tethering experiments in Figure 3 were used to test if de novo targeting of FBLX19 to na\u00efve chromatin would result in recruitment of CDK-Mediator. This showed that FBXL19 binding was sufficient to recruit CDK-Mediator (Figure 3B), in agreement with our observations that when the capacity of FBXL19 to bind CpG island chromatin is disrupted, CDK8 is no longer appropriately recruited to a subset of CpG island-associated target genes (Figure 4).\n\n![Fbxl19 overexpression levels.\\\nRT-PCR showing the expression levels of FS2-Fbxl19 relative to endogenous Fbxl19 in the empty vector (EV) control line. Data is the average of three biological replicates. Error bars represent standard deviation.](elife-37084-resp-fig2){#respfig2}\n\n> 5\\) FBXL19 binding to CpG islands and its requirement in ES differentiation has been described previously: (https://www.ncbi.nlm.nih.gov/pubmed/28453857) \\\"Our genome-wide target mapping unveils the preferential occupancy of Fbxl19 on CpG island-containing promoters \\[...\\]\\\" \\\"Moreover, we reveal that Fbxl19 is critical for proper differentiation of ES cells \\[...\\]\\\" With respect to this, the authors should clarify the novelty of their results presented in the current manuscript.\n\nDuring the preparation of our manuscript, the manuscript cited by the reviewer was published in NAR. It reported that FBXL19 occupied CpG islands and appeared to play a role in cellular differentiation. However, we would like to point out that, although our observations were similar with respect to these two points, the main discoveries resulting from our careful and systematic study are completely distinct from that of the published study. In direct contrast to the published work, we find no evidence for any connection between FBXL19 and RNF20 using unbiased biochemistry and mass spectrometry (Figure 2 and Figure 2---figure supplement 1F). Furthermore, using our conditional genetic ablation strategies we do not observe any effect of FBXL19 loss on H2BK120ub1 (Figure 2---figure supplement 1G). Therefore, we are unable to reproduce the central findings of the published study. In our opinion this severely limits and brings into question the validity and, therefore, the novelty of the published work. Our observations likely differ from those of the NAR study because we use unbiased as opposed to candidate-based (guess work) biochemistry and conditional knockout strategies as opposed to stable knockdown cell lines that are prone to indirect and secondary effects due to the long-term culture necessary for their generation. Nevertheless, we agree with the reviewer that the wording of the indicated sentences could have been more inclusive and contrasting of the published work and have clarified this in the revised manuscript (Introduction, Results section, subsection \"FBXL19 interacts with the CDK-Mediator complex in ES cells\" and subsection \"Removing the CpG island-binding domain of FBXL19 results in a failure to induce developmental genes during ES cell differentiation\").\n\nWe would also like to take this opportunity to clarify the novelty of our study and the importance of its central conclusions. (1) By using unbiased biochemical approaches, we discover a link between FBXL19 and CDK-Mediator in mouse ES cells. (2) Using conditional genetic ablation, de novo targeting strategies, and genomics, we go on to show that FBLX19 can recruit CDK-Mediator to chromatin and the CpG island promoters of silent developmental genes in ES cells. (3) In the absence of FBXL19, these genes are no longer appropriately activated upon differentiation, providing an explanation for why normal differentiation cannot be achieved in the absence of FBXL19. (4) In the revised study, based on the reviewer's suggestions (point 1), we now demonstrate that FBXL19 target genes also rely on CDKMediator for their activation, validating our biochemistry and genomics. (5) Finally, we discover that loss of FBXL19 leads to abnormal development and early embryonic lethality, in agreement with the inability of cells to appropriately regulate gene expression during development.\n\nTherefore, our discoveries are novel in describing a completely new modality by which FBXL19 and CpG islands prime genes for future expression. Furthermore, we provide compelling new evidence that CDKMediator can be recruited to gene promoters independently of classical transcription factors and transcription itself to prime CpG island-associated genes for activation. These new discoveries force us, and the field, to think differently about how CpG islands integrate and regulate gene expression, while also provide new insight into the poorly understood mechanisms by which CDK-Mediator is targeted to genes and functions to control gene expression. We believe these important points are now more clearly framed in the revised manuscript.\n\n> 6\\) The conclusion that FBXL19 does not appear to regulate CDK8-Mediator levels is not fully supported by the data. Figure 2---figure supplement 2H needs to be repeated so that CDK8 bands are not over-exposed as there does appear to be a change in amount/mobility when overexpressed -/+ MG132. Furthermore, there seems to be a noticeable increase of MED12 protein upon FBXL19 overexpression, which was even more obvious when cells were treated with MG132.\n\nWe have repeated these experiments based on the reviewer's suggestions. We again failed to observe any significant difference in CDK8 protein levels following FBXL19 over-expression at lower exposure levels. We agree it appeared that MED12 levels may have increased slightly when FBXL19 was overexpressed in original Figure 2---figure supplement 1G. However, we believe this is an artefact resulting from inefficient transfer on the left had side of the gel. MED12 is a large protein (250KDa) and we previously have encountered difficulties efficiently transferring MED12 to membrane for Western blot analysis. We have now optimised our transfer of large proteins and fail to see any significant or reproducible alterations in MED12 levels (Figure 2---figure supplement 1H). MED12 is known to be regulated by the proteasome in ES cells (e.g. Buckley et al., 2012). In agreement with this, we observe an increase in MED12 levels after MG132 treatment in both FBXL19 overexpressing cells and in the control cells transfected with an empty vector. This effect is therefore not related to FBXL19. Finally, we have also included in the revised manuscript a Western blot analysis of MED13 (which is also part of the CDK module) and again we observe no change in protein levels following FBLX19 overexpression (Figure 2---figure supplement 1H). We conclude that FBXL19 does not regulate CDK module protein levels.\n\n> 7\\) In Figure 4, CDK8 recruitment has only been tested for the FBXL19 ZF-CxxC deletion -is this recapitulated for depletion of the whole protein? Given that Figure 4B appears to show a decrease in the levels of the deletion protein this is an important point to verify.\n\nWe reasoned that conditionally removing the ZF-CxxC domain would be the most surgical way of disrupting binding of FBXL19 to CpG islands, without interfering other potential functions of FBXL19. We agree that deletion of the ZF-CxxC domain appears to cause some reduction in FBXL19 protein levels. We speculate that FBXL19 may be more rapidly degraded if not bound to chromatin (something we have observed for other ZF-CxxC domain containing proteins). Importantly, CDK8 proteins levels are unaffected in the \u0394CxxC-FBXL19 cells (Figure 4---figure supplement 1F). We have now carried out a biochemical purification of \u0394CxxC-FBXL19 and shown that it still interacts with CDK8-Mediator (Figure 2---figure supplement 1E). Given that \u0394CxxC-FBXL19 does not bind to CpG islands (Figure 1) yet still interacts with CDK8, we conclude that reductions in CDK8 recruitment at FBXL19 targets genes is most likely explained by the loss of FBXL19 binding to CpG islands. Nevertheless, we now draw attention to the reduced protein levels of the ZF-CxxC deleted protein as follows (subsection \"FBXL19 is required for appropriate CDK8 occupancy at a subset of CpG island-associated promoters\"):\n\n'Following removal of the ZF-CxxC domain of FBXL19, we observed some reductions in FBXL19 protein levels (Figure 4B), but importantly CDK8 levels were unaffected (Figure 4---figure supplement 1F).'\n\n> 8\\) The authors need to show that OHT treatment of the delCxxC ES line leads to reduced binding of FBXL19 in conditions used for CDK8 ChIPseq -several genes would be ok -these same genes should be used to validate CDK8 CHIPseq signal.\n\nAll commercially available antibodies for FBXL19 that we have tried fail to ChIP FBXL19 and our exhaustive attempts to generate our own FBXL19-specific antibody that works for ChIP have been unsuccessful (two separate epitopes and 4 rabbits). This has precluded us from carrying out the proposed experiment. However, we do observe a uniform loss of FBXL19 binding when the ZF-CxxC domain is removed in epitope-tagged transgene experiments (Figure 1) suggesting that FBXL19 binding to CpG islands relies on an intact ZF-CxxC domain.\n\n> 9\\) Supplemental tables with differential gene expression results should be provided -or at least working gene lists used to reach the conclusions in the manuscript. Full GO enrichment results should be included as supplementary tables so that specific gene lists can be examined.\n\nThis information is now included as Supplementary file 1.\n\n> 10\\) Various figure panels are missing statistical tests: Figure 4C needs statistical tests, e.g. KS/Wilcoxon -See Figure 4E; Figure 4G needs statistical tests; Figure 4H and 4I need boxplots and need statistical tests eg KS/Wilcoxon as for E/F;\n\nWe have now included the necessary statistical tests and plots as requested by the reviewer.\n\n> Figure 4H-J should include FBXL19-bound groups;\n\nAs suggested by the reviewer, we divided all CDK8 peaks into those which have FBXL19 (FBXL19+) and those that do not have FBXL19 (Fbxl19-). We then examined FBXL19, BioCAP and CDK8 enrichment in these groups (Figure 4---figure supplement 1I). As observed when we analysed all CDK8 peaks (Figure 4H-J), we found that CDK8 peaks that are bound by FBXL19 have increased levels of FBXL19, CDK8, and BioCAP signal when compared to sites not bound by FBXL19. Furthermore, sites that rely on FBXL19 for CDK8 binding have broader peaks of FBXL19 enrichment. This is consistent with the fact that these are large CpG islands with broad BioCAP signal (Figure 4F, 4H). The majority peaks that show reduction of CDK8 binding in FBXL19^\u0394CXXC^ ES cells with FBXL19 peaks (519/783) while there is only a small overlap between FBXL19 binding and sites with increased CDK8 binding (90/379). We have now included this information in Figure 4---figure supplement 1I.\n\n> Figure 6B,E, F need statistical tests.\n\nWe have now included the necessary statistical tests.\n\n> 11\\) It is very nice that the authors showed that genome-wide chromatin binding of FBXL19 was completely lost when the ZF-CxxC domain was deleted or mutated. To further confirm that it is indeed this ZF-CxxC domain mediating interaction with the non-methylated CpG islands, some controls are necessary. For example, in Figure 1---figure supplement 1D, the authors showed that almost all of the FBXL19 binding sites are overlapped with NMIs locations, with only 93 exceptions. Then are the signals at these 93 sites also changed in the FBXL19 mutant ChIPSeq data? When the cell DNA methylation levels are modulated, will the binding pattern of FBXL19 changed accordingly?\n\nWe have now compared the enrichment of WT FBXL19 and \u0394CxxC FBXL19 at these 93 sites ([Author response image 3A](#respfig3){ref-type=\"fig\"}), which corresponded to regions that were not associated with NMIs and observed a loss of binding of \u0394CxxC FBXL19 ([Author response image 3B](#respfig3){ref-type=\"fig\"}). Importantly, these same 93 sites also show some BioCAP signal, albeit at a level much lower than is observed at NMIs. This suggests that these sites are very weak CpG islands which were not captured in our NMI peak set due to their low Bio-CAP signal.\n\nFurthermore, to illustrate the relationship between FBXL19 binding and CpG-rich non-methylated DNA, we have examined BioCAP and FBXL19 signal across all ATAC-seq peaks in the mouse ES cell genome ([Author response image 3B](#respfig3){ref-type=\"fig\"}). ATAC-seq peaks represent a combination of regulatory elements that are CpGrich (CpG islands) or CpG-poor (usually distal regulatory elements). This provides the necessary contrast to illustrate the correlation of FBXL19 tracks with sites that contain CpG rich non-methylated DNA (BioCAP), consistent with a FBXL19 binding modality that is dependent on its ZF-CxxC domain.\n\nWe have not been able to examine FBXL19 binding in cells with altered DNA methylation levels as this would require recreation of a DNMT1 (or DNMT3A/B) null situation in our transgene containing cell line (we don't use 5aza-C for this purpose as it causes extensive DNA damage due to covalent adducts which are formed between DNA and DNMT enzymes). However, when we have previously examined DNA methylation-deficient cells and the behaviour of the closely related KDM2A/B proteins, which have nearly identical CpG island binding patterns to FBXL19 (Figure 1D and 1E), we demonstrated that KDM2A/B occupancy is completely dependent on recognition of CpG-rich non-methylated DNA (Blackledge et al., 2010; Zhou et al., 2012).\n\n![(A) A Venn diagram showing the overlap between FBXL19 peaks and NMIs; (B) Metaplot analysis of enrichment of WT FS2-FBXL19, \u0394CxxC FS2-FBXL19, and EV control at FBXL19-occupied NMIs (solid line) and nonNMI (dashed line) regions; (C) Heatmaps showing enrichment of BioCAP, WT FS2-FBXL19, \u0394CxxC FS2FBXL19 and EV control at DNA accessible regions measured by ATAC-seq.](elife-37084-resp-fig3){#respfig3}\n\n\\[Editors\\' note: the author responses to the re-review follow.\\]\n\n> Essential revisions:\n>\n> 1\\) In AP-MS (Figure 2A), what is the background value of emPAI? Are there any other FBXL19-interacting proteins that show significantly high levels of emPAI in the list? How significant are the components of CDK-Mediator complex represented in all the hits?\n\nTo clarify these points, we have now included a table in the supplementary information (Figure 2---source data 1) detailing the AP-MS results. This includes the biological replicate FS2-FBXL19 purifications and matched empty vector (EV) controls. We limited analysis of our AP-MS data to proteins that were uniquely identified in the FBXL19 purifications (i.e. not the EV control). We have included a column highlighting the proteins that were identified in this screen and their associated emPAI scores in Supplementary file 1. Importantly, Mediator components feature heavily in this table, which is what prompted us to examine CDK-Mediator as a possible interaction of functional relevance. We have yet to validate whether other proteins in this table are bona fide interactors and, therefore, their relationship with FBXL19 and its function is not known.\n\n> 2\\) In Figure 3, the authors demonstrated that FBXL19 was able to recruit CDK-Mediator complex, such as CDK8 and MED12, to artificially designed chromatin regions. What about CXXC-domain or F-box deletion mutants in this assay?\n\nIn these experiments we purposely used a version of FBXL19 in which the CxxC domain has a single substitution mutation (K49A) that prevents CpG island binding. We designed the experiment in this way to prevent the TetR-fusion protein from binding to CpG islands throughout the genome. This allowed us to focus our analysis on the artificial nucleation site. Although we have not tested a version of FBXL19 with its CxxC domain deleted, we believe that it would function in a very similar way to the K49A mutant as both proteins are deficient in CpG island binding and we have shown that \u0394CxxC-FBXL19 retains its interaction with CDK-Mediator (Figure 2---figure supplement 1E).\n\nIn all of our artificial tethering experiments we generate clonal stable cell lines expressing the TetR-fusion proteins to ensure uniform expression within lines and comparable expression across test lines. Despite exhaustive efforts, we have been unable to generate stable cell lines expressing FBXL19 lacking the F-box. We are unsure why this is the case but speculate that FBXL19 without the F-box may be unstable or lead to a dominant negative influence that affects cell growth (negative selection). Therefore, we have not been able to examine how the F-box deletion behaves in our tethering experiments. Given that \u0394F-box-FBXL19 does not interact with CDK-Mediator in transient overexpression IP experiments (Figure 2---figure supplement 1E), we assume that CDK8 and MED12 would not be recruited to chromatin by TetR-\u0394F-box-FBXL19. However, we have been unable to directly test this possibility.\n\n> 3\\) In Figure 4B, it is not very clear which band(s) are exactly indicated as full-length FBXL19 and which one(s) are CXXC deletion? Why are there double bands? Did the author use T7 antibody to detect the endogenously introduced protein?\n\nWe generated the T7 knock-in FBXL19 ES cell line (Figure 2---figure supplement 1B, C) in order to detect FBXL19 by Western blot. We observe 'double bands' for FBXL19 and speculate that this may be due to post-translational modification. The closely related KDM2A and KDM2B proteins are also extensively post-translationally modified leading to doublets when examined by Western blot. Following tamoxifen treatment of *Fbxl19^CxxCfl/fl^* cells to remove the CxxC domain, we observe a reduction in mobility for these double bands (with the top band shifting to the previous position of the bottom band). We have now clearly indicated which bands represent WT and \u0394CxxC FBXL19 in Figure 4B and have described this in the figure legend.\n\n> 4\\) In Figure 4C, what proportions of each group of CDK8 binding sites overlap with FBXL19? In theory, the author expects to see that CDK8 sites showing binding intensity changes upon tamoxifen treatment would tend to overlap more with FBXL19.\n\nThe overlap of each group of CDK8 peaks with FBXL19 peaks is shown in Figure 4---figure supplement 1H. While roughly 41% of all CDK8 peaks overlap with FBXL19 peaks, the overlap of down-regulated CDK8 peaks is about 66%. This is highlighted in the text in subsection \"FBXL19 is required for appropriate CDK8 occupancy at a subset of CpG island-associated promoters\". On the other hand, we observed that only 23% of up-regulated CDK8 peaks overlap with FBXL19 peaks.\n\n> 5\\) In Figure 4---figure supplement 1C, please use a Venn Diagram to show directly the number of peaks of CDK8, FBXI19 and NMI overlapping with each other. This would be more straightforward.\n\nWe chose to represent CDK8 enrichment at FBXL19 peaks as a heatmap in order to circumvent the limitation of peak-calling. However, as suggested by the reviewer, we have now also included a Venn diagram that shows that 89.4% of FBXL19 peaks overlap with CDK8 peaks and NMIs (Figure 4---figure supplement 1D) and mentioned this in the text (subsection \"FBXL19 is required for appropriate CDK8 occupancy at a subset of CpG island-associated promoters\").\n\n> 6 In Figure 5, please perform the same GSEA or functional annotation analysis on genes that are associated with unchanged CDK8 binding sites upon FBXL19 CXXC-domain deletion.\n\nWe have now included the GO analysis for genes associated with unchanged CDK8 binding in Figure 5---figure supplement 1. This revealed a broad range of terms associated with basic molecular processes in line with Mediator being implicated as a key regulator of transcription. We now highlight this observation in the text in subsection \"FBXL19 targets CDK8 to promoters of silent developmental genes in ES cells\".\n\n> 7\\) In Figure 6B, what is the result for genes that are associated with increased CDK8 binding signals in FBXL19 CXXC-domain deletion mutant upon RA-induced differentiation?\n\nWe have compared the expression of genes associated with an increase in CDK8 binding in Figure 6---figure supplement 1E. While we observe no overall difference in the expression of these genes in ES cells, there is a slight increase upon RA-induced differentiation (p-value 0.04). This is consistent with Mediator playing a role in gene transcription.\n\n> 8\\) In Figure 6C, even though the authors claimed that those upregulated genes (889) in FBXL19 CXXC-domain deletion mutant upon RA treatment were not involved in developmental processes, it is still interesting to find that significantly more genes were activated upon CXXC-domain deletion. What explanation can the authors provide for this finding?\n\nWe agree with the reviewer that a significant number of genes were more activated in the *Fbxl19^\u0394CxxC^* cells compared to wild type cells upon RA-induced differentiation and that this is potentially interesting. We have yet to find a compelling molecular explanation to link FBXL19 directly to these effects on gene expression. It remains possible that these gene expression changes are due to secondary transcriptional defects that manifest when cells attempt to execute a new expression programme in the absence of normal activation of silent, lineage commitment-associated genes when FBXL19 is not functional, or that FBXL19 plays other, as yet unidentified, roles in inhibiting gene expression. We have now added a sentence drawing attention to these points in the revised manuscript in subsection \"Removing the CpG island-binding domain of FBXL19 results in a failure to induce developmental genes during ES cell differentiation\".\n\n> 9\\) In Figure 7, the authors claim that FBXL19 relies on CDK-mediator complex to activate the specific set of developmental genes. Is it possible to test this hypothesis in the FBXL19 CXXC-deletion system they established in Figure 6? In wild-type ESC, depletion of MED13 will have similar effects as FBXL19 CXXC deletion, whereas in FBXL19 mutant, the effects would be much alleviated?\n\nOne could test whether inducible MED13/13L deletion in the *Fbxl19^CxxCfl/fl^* background would not lead to any further effect on gene activation. However, this would require a significant amount of time to generate and characterise a new *Med13^fl/fl^, Med13l^fl/fl^*, and *Fbxl19^CxxCfl/fl^*cell line. While this is an interesting experiment, we believe that the robust and extensive series of observations in Figure 6 and Figure 7 more than adequately support our conclusion that FBXL19 target genes rely on CDK-Mediator for activation during differentiation.\n\n> 10\\) In Figure 8, why did heterozygous deletion of FBXL19 CXXC-domain show even better developmental status than the wild type?\n\nWe have not observed any significant difference between WT and heterozygous mutant embryos. It is common to have some stage variations until around 12 dpc, even within embryos of the same litter. Therefore, we often find small differences in size and developmental status among embryos regardless of genotypes. Both the WT and heterozygous mutant embryos presented in Figure 8 appear healthy with normal development for this stage. On the other hand, the defects observed in the homozygous *Fbxl19^\u0394CxxC^* embryos cannot be explained by such stage variations.\n\n> 11\\) The authors demonstrated that the CXXC-domain mediates the binding of FBXL19 at NMIs, where F-Box mediates its interaction with the CDK-Mediator complex. What about CDK8 chromatin binding or gene expression patterns upon F-Box deletion of FBXL19? Will the authors see the same results as what they found in CXXC-domain deletion mutant?\n\nThis is an interesting experiment and we hypothesize that disrupting the interaction between CDK-Mediator and FBXL19 via deleting the F-box domain would yield similar results to removing the CxxC domain. However, as described above in response to reviewer point 2, despite exhaustive efforts we have been unable to generate cell lines stably expressing FBXL19 with a deletion of the F-box, precluding us from addressing this point. It will be interesting in future work to see if we can engineer a cell line where we are able to remove the F-box in an inducible manner and examine the effects on CDK8 binding and induction of gene expression.\n\n[^1]: Agilent Technologies, Waldbronn, Germany.\n\n[^2]: These authors contributed equally to this work.\n"}
+{"text": "\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files. The data about the manuscript has been uploaded by using SPSS 17.0 format in a supplemental file.\n\nBackground {#sec005}\n==========\n\nThe prevalence of antenatal depression varies from 7.4% to 50.0% worldwide, and from 5.5% to 23.1% in China \\[[@pone.0167597.ref001]--[@pone.0167597.ref007]\\]. The prevalence of depression among pregnant women is higher than that among postpartum women, and it is higher among pregnant women in middle and late pregnancy \\[[@pone.0167597.ref005], [@pone.0167597.ref008]\\]. Antenatal depression is a debilitating experience which can lead to many problems and sequelae. For example, depressed pregnant women may experience multiple conflicting roles, insufficient social support, uncertainty about future life, instability of emotion, and the discontent of body image. Moreover, there can be a risk of preterm birth and obstetric complications. And the newborns' and husbands' mental and physical health can also be threatened by their mothers' and wives' depression as well \\[[@pone.0167597.ref009]--[@pone.0167597.ref016]\\].\n\nThe predictors of antenatal depression include several socio-demographic and health behavior factors such as young or old age, low educational background, and low socio-economic status. Besides, antenatal depression is always related with threatening life events, such as housing problems, financial difficulties, and marital problems \\[[@pone.0167597.ref017]\\]. And negative life events contribute more to antenatal depressive symptoms than other socio-demographic factors do \\[[@pone.0167597.ref018], [@pone.0167597.ref019]\\]. Moreover, there are psychosocial factors such as stress, low social support, and low optimism level that can lead to an increased antenatal depressive level \\[[@pone.0167597.ref007], [@pone.0167597.ref020]--[@pone.0167597.ref022]\\].\n\nApart from those listed above, automatic thoughts are also believed to have a reciprocal relationship with depression, indicating that automatic thoughts can be the outcome of depression, and it can also have an impact on depressive level \\[[@pone.0167597.ref023]\\]. Automatic thoughts reflect one's underlying core beliefs. If some events, which challenge one's core beliefs, take place, the stream of negative automatic thoughts may run through one's mind, and it will cast a negative interpretation of the events. The negative interpretation may increase the stress level and cause depressive symptoms \\[[@pone.0167597.ref024]--[@pone.0167597.ref029]\\]. That is consistent with the vulnerability model, which postulates that negative automatic thoughts play a mediating role between negative life events and depression \\[[@pone.0167597.ref023]\\]. This mediating effect has been examined in the study among adolescents \\[[@pone.0167597.ref030]\\]. It has also been applied in practice. Some interventions, which employed cognitive-behavioral therapy principles, used this mediating effect with their design of targeting and altering negative automatic thoughts, and building more adaptive automatic thoughts among pregnant women \\[[@pone.0167597.ref024], [@pone.0167597.ref031]\\].\n\nRural pregnant women used to be believed to have a lower depressive level in some studies, because they are supposed to have stronger family support and community connections, which can protect them from the risk of perinatal depression\\[[@pone.0167597.ref032], [@pone.0167597.ref033]\\]. Conversely, some studies showed conflicting findings. According to a quantitative study conducted in four provinces of China from 2001 to 2005, which included 63004 adult participants from both rural and urban sampling sites, there was a higher prevalence of depressive disorders among rural residents than that among urban residents \\[[@pone.0167597.ref034], [@pone.0167597.ref035]\\]. Further, some studies illustrated the difficulties encountered by rural pregnant woman. Pregnant women in rural areas could suffer from the isolation from their spouses, and the mental health services in rural areas were limited \\[[@pone.0167597.ref036]--[@pone.0167597.ref038]\\].\n\nIn present rural China, this issue is more complex. There has been a trend that many male laborers prefer to go to urban areas as migrant workers since the 1980s. Since the dual division system of city and countryside did not allow migrant workers to relocate the whole family with them to urban areas, many wives of the migrant workers had to be left behind in their rural residences, doing farming work and taking care of families. That created a burden on the daily life of the left-behind wives, and led to their worse mental health status \\[[@pone.0167597.ref039], [@pone.0167597.ref040]\\].\n\nThis is the first study that aimed to assess the mediating effect of negative automatic thoughts between negative life events and antenatal depressive symptoms among rural pregnant women in China. We postulate that among rural pregnant women in China, negative life events and antenatal depression are mediated by negative automatic thoughts.\n\nMaterials and Methods {#sec006}\n=====================\n\nStudy design and participants {#sec007}\n-----------------------------\n\nThe study is a cross-sectional hospital-based survey, and it was conducted in Mianzhu County in Sichuan Province and in Gaobeidian County in Hebei Province, China. Ethical approval and consent processes were obtained from the institutional review board of the School of Social Development and Public Policy at Beijing Normal University.\n\nTen volunteers, who were graduate students majoring in psychology, received a two-day training to be qualified interviewers, and they investigated the participants. In Mianzhu County, the data were collected twice, once in June 2012 and once in October 2012. The data collection in October interviewed a new population of pregnant women. In Gaobeidian County, the data were collected only in October. Each collection lasted for approximate eight days. A self-rating questionnaire was distributed to pregnant women who were receiving routine prenatal care at the hospitals. The women completed the self-reporting questionnaires in approximate 30 minutes during their waiting for a routine antenatal check-up. A signature on the consent form was regarded as a sensitive issue in local cultural context, and participants could reject the investigation because of it. Moreover, this anonymous study was little risk to participants. Thus, an informed oral consent was obtained from each study participant instead of a written consent. The study had been explained to each participant by volunteers, and the volunteers had answered the participants' questions prior to asking for their permission to conduct the investigation. Participation was entirely voluntary, and each participant had the right to withdraw or to refuse to provide information at any time during the study.\n\nThe participants from Mianzhu County got interviewed at Mianzhu People\\'s Hospital and Mianzhu Maternal and Child Health Hospital, where around 70% of the pregnant women in this county received antenatal care and post-delivery services. The participants from Gaobeidian County got interviewed at Gaobeidian County's Hospital, where nearly 50% of the pregnant women in this county received antenatal care and post-delivery services. The two hospitals where this study was conducted are the top ranked hospitals in the two counties.\n\nMeasurements {#sec008}\n------------\n\nAntenatal depression was rated by the Edinburgh postnatal depression scale (EPDS).EPDS is a 10-item self-reporting instrument, used to screen for antenatal depressive symptoms, and it is most frequently used in epidemiological research about antenatal depression \\[[@pone.0167597.ref041], [@pone.0167597.ref042]\\]. The items are rated on a 0 to 3 scale, yielding a total score range of 0 to 30. The items focus on the cognitive and affective features of depression. The EPDS has been employed in studies of China, and it was reliable in measuring prenatal and postnatal women\\'s depression \\[[@pone.0167597.ref043]--[@pone.0167597.ref046]\\]. The Cronbach's alpha coefficient of the Chinese version of the EPDS in this study was 0.689. A cut off point score of \u226513 was used to determine depressive symptoms in the bivariate correlation analysis, as was recommended in a previous study \\[[@pone.0167597.ref047], [@pone.0167597.ref048]\\].\n\nThe perceived pressures of pregnancy were rated by an 11-item self-assessment scale which is a short-form of the pregnancy pressure scale (PPS) developed by Zhanghui Chen et al. The Cronbach\\'s alpha coefficient was reported to be 0.84 \\[[@pone.0167597.ref049]\\]. This scale contains three subscales: pressure from identification of the parent's role, pressure from the concerns of maternal and child health, and pressure from the change of body shape or physical activities. The scale measures the perceived stress of major pregnancy-related events by using a 4-point Likert scale from 1 (none or little) to 4 (high). The total score is used as an index of perceived stress. Possible stressful experiences include: fears about important people disliking the baby, concerns about reduced leisure time with a baby, fears about the safe delivery of the baby, anxiety about birth defects, fears about complications during delivery, fears about pain during delivery, concerns about changes in body shape, fears about competence as a mother, fears about the negative impact of the baby on the marital relationship, concerns about providing a healthy living environment to the baby, and other pregnancy-related stresses.\n\nThe negative automatic thoughts were measured by the automatic thoughts questionnaire (ATQ) \\[[@pone.0167597.ref043], [@pone.0167597.ref050]\\]. This 30-item questionnaire is devised to measure the frequency of the occurrence of automatic negative thoughts associated with depression. Four aspects of automatic thoughts are measured. They are personal maladjustment and the desire for change, negative self-concepts and negative expectations, low self-esteem, and helplessness. The responses of every question are provided on a 5-point scale from not at all (1) to all of the time (5). A high total score indicates more frequent negative thoughts. The questionnaire has been used in many studies about depression in China and elsewhere. The studies in China found that a change of the scores of the ATQ was significantly consistent with the change of the scores of the Self-Rating Depression Scale \\[[@pone.0167597.ref051]--[@pone.0167597.ref053]\\]. In the current study, the examination of the internal consistency of the ATQ yielded a coefficient alpha of 0.93, which suggested a high level of internal consistency. The ATQ-30 appears to be a reliable measure of automatic thoughts in depression.\n\nThe negative life events were measured by the life events scale for pregnant women (LESPW) \\[[@pone.0167597.ref054]\\]. This scale is self-rating, and it is designed specifically for pregnant women. It includes life events which lead to different levels of sensation, and the events concern family life, work and study, social relationships, etc. There are 53 events, and they are divided into two groups which involve subjective events (SE) and objective events (OE) separately. The OE are also divided into three levels by the extent they impact the emotion of the pregnant women, and they are grouped into OE1, OE2, OE3. The current study used the OE, because the SE are more related with subjective feelings, while we have used other special tools to measure them, which are negative automatic thoughts, stress, and antenatal depression. Thus, we only need the events of in the OE as the events that can lead to stress and depression. Besides, the current study only employed OE2 and OE3, because the events included in OE2 and OE3 have greater significance on sensation to people, and they can have a greater and long- lasting impact on pregnant women's depressive level. Thus, in order to decrease the influence brought by the temporality of events, we only chose events in OE2 and OE3. The events are all negative life events, which fit well with the objective of this study. A total of 34 life events in OE2 and OE3 were used. In the current study, the scale showed a good internal consistency (Cronbach\\'s Alpha = 0.60).\n\nThe socio-demographic and health behavior factors include the participants' age (18--24 years; 25--29 years; \u226530 years), region (Mianzhu; Gaobeidian); level of education (middle school or lower; high school; college or above), parity (primigravida; others), monthly family income (USD \\<326; USD 326--816; USD \u2265816), living site (village; city), planned pregnancy (yes; no), body mass index (BMI) (overweight is defined as BMI\u226523 among Asian population by WTO), the stage of gestation (0--28 weeks; \\>28 weeks), employment (unemployed; part-time job; full-time job), sleep quality (good; fair; poor), smoking history (yes; no), alcohol use history (yes; no), whether the husband is a migrant worker (yes; no).\n\nStatistical analysis {#sec009}\n--------------------\n\nSPSS 17.0 (SPSS Inc, Chicago, IL) was used for statistical analysis. A descriptive analysis was performed for the socio-demographic and health behavior characteristics (region, age, level of education, monthly household income, living site, employment, planned pregnancy, BMI, gestation stage, parity, sleep quality, smoking history, alcohol use history, whether the husband is a migrant worker). Chi-square tests and a bivariate correlate analysis were performed to examine the correlation between the outcome variable and the independent variables, and the socio-demographic and health behavior factors. Logistic regression was used to test the socio-demographic and health behavior factors, negative life events, stresses of pregnancy, and negative automatic thoughts as predictors of antenatal depression. All of the estimates were accompanied by odds ratios (OR) and 95% confidence intervals.\n\nAMOS 17.0 was used for the path analysis. Negative life events, negative automatic thoughts, stresses of pregnancy, and antenatal depression were included in the models. The comparative fit index (CFI), the incremental fit index (IFI), and the root mean square error of approximation (RMSEA) with 90% confidence intervals were used to estimate the model fit.\n\nResults and Discussion {#sec010}\n======================\n\nResults {#sec011}\n-------\n\nThe mean age of the participants was 25.49 (min = 18.0, max = 42.0; standard deviation \\[SD\\] = 3.85) and nearly half (47.3%) of the participants were 18--24 years old. Nearly half of the pregnant women (47.8%) completed middle school or less. Most of the pregnant women (82.0%) had a family monthly income lower than 816 USD. Most of the pregnant women (86.9%) were unemployed during their pregnancy. Approximately half (50.5%) of the pregnancies were planned. More than half (58.7%) of the women were primigravida, and most of them (70.8%) had been pregnant for more than 28 weeks. The BMI of most of them (74.5%) were over 23. More than half (52.7%) of the pregnant women had a fair quality of sleep, and there were smaller percentages of pregnant women with a good quality of sleep (39.2%) or a poor quality of sleep (8.1%). Almost all (98.2%) of the pregnant women had no smoking history, and almost all (94.3%) of the pregnant women had no history of alcohol use. Over half (54.5%) of the pregnant women's husbands were migrant workers ([Table 1](#pone.0167597.t001){ref-type=\"table\"}).\n\n10.1371/journal.pone.0167597.t001\n\n###### Description of socio-demographic and health behavior characteristics, scores of AND, stresses of pregnancy, NLE, and NAT of the participants (n = 495).\n\n![](pone.0167597.t001){#pone.0167597.t001g}\n\n  Socio-demographic and Health Behavior characteristics   N     \\%\n  ------------------------------------------------------- ----- ------\n  Region                                                        \n  Mianzhu                                                 249   50.3\n  Gaobeidian                                              246   49.7\n  Age                                                           \n  18--24                                                  236   47.3\n  25--29                                                  175   35.1\n  \u226530                                                     88    17.6\n  Education                                                     \n  Middle school or lower                                  239   47.8\n  High school                                             165   33.0\n  College or above                                        96    19.2\n  Monthly family income(USD)                                    \n  \\<326                                                   164   33.1\n  326--816                                                242   48.9\n  \u2265816                                                    89    18.0\n  Living site                                                   \n  Village                                                 396   80.0\n  City                                                    98    20.0\n  Employment                                                    \n  Unemployed                                              424   86.9\n  Part-time job                                           18    3.7\n  Full-time job                                           46    9.4\n  Planned pregnancy                                             \n  Planned                                                 244   50.5\n  Unplanned                                               239   49.5\n  BMI                                                           \n  \\<23                                                    124   25.5\n  \u226523                                                     363   74.5\n  Parity                                                        \n  Primigravida                                            293   58.7\n  Others                                                  206   41.3\n  Gestation stage                                               \n  \u226428weeks                                                130   29.2\n  \\>28weeks                                               323   70.8\n  Sleep quality                                                 \n  Poor                                                    40    8.1\n  Fair                                                    261   52.7\n  Good                                                    194   39.2\n  Smoking history                                               \n  Non                                                     486   98.2\n  Yes                                                     9     1.8\n  Alcohol use history                                           \n  Non                                                     466   94.3\n  Yes                                                     28    5.7\n  Husband is a migrant worker                                   \n  Yes                                                     261   54.5\n  Non                                                     218   45.5\n  Depression                                                    \n  \\<13                                                    427   86.3\n  \u226513                                                     68    13.7\n\nNote: some variables have missing values.\n\nThe mean score of negative automatic thoughts was 39.07 (min = 30, max = 99; SD = 10.71), and the median score was 35.00. The mean score of negative life events was 105.31 (min = 0, max = 515; SD = 100.27), and the median score was 84.00. The mean score of stresses of pregnancy was17.11 (min = 10, max = 40; SD = 4.70).\n\nThe mean score of antenatal depression was 8.71 (min = 0, max = 25; SD = 3.91), and the median score was 9.00. The EPDS cutoff point recommended by Rubertsson's study is a score of \u226513. According to this criterion, the prevalence of major depressive symptoms in the current study was 13.7% \\[[@pone.0167597.ref048]\\].\n\n[Table 2](#pone.0167597.t002){ref-type=\"table\"} displays the bivariate correlations between antenatal depression and the socio-demographic and health behavior factors in the study. Antenatal depression was only significantly correlated with the sleep quality (p = 0.001), while other variables didn't have a significant relationship with antenatal depression.\n\n10.1371/journal.pone.0167597.t002\n\n###### The prevalence of antenatal depression by socio-demographic and health behavior factors.\n\n![](pone.0167597.t002){#pone.0167597.t002g}\n\n                                Depression   Depression   P\n  ----------------------------- ------------ ------------ -------\n  Region                                                  \n  Mianzhu                       219(88.0)    30(12.0)     0.272\n  Gaobeidian                    208(84.6)    38(15.4)     \n  Age                                                     \n  18--24                        196(83.8)    38(16.2)     0.097\n  25--29                        158(90.8)    16(9.2)      \n  \u226530                           73(83.9)     14(16.1)     \n  Education                                               \n  Middle school or lower        203(86.4)    32(13.6)     0.692\n  High school                   139(84.8)    25(15.2)     \n  College or above              85(88.5)     11(11.5)     \n  Monthly family income(USD)                              \n  \\<326                         136(84.0)    26(16.0)     0.516\n  326--816                      208(87.0)    31(13.0)     \n  \u2265816                          79(88.8)     10(11.2)     \n  Living site                                             \n  Village                       341(86.1)    55(13.9)     0.873\n  City                          85(86.7)     13(13.3)     \n  Employment                                              \n  Unemployed                    364(86.7)    56(13.3)     0.449\n  Part-time job                 14(77.8)     4(22.2)      \n  Full-time job                 38(82.6)     8(17.4)      \n  Planned pregnancy                                       \n  Planned                       213(88.4)    28(11.6)     0.162\n  Unplanned                     199(84.0)    38(16.0)     \n  BMI                                                     \n  \\<23                          102(83.6)    20(16.4)     0.401\n  \u226523                           312(86.7)    48(13.3)     \n  Gestation stage                                         \n  \u226428weeks                      111(85.4)    19(14.6)     0.971\n  \\>28weeks                     278(86.1)    45(13.9)     \n  Parity                                                  \n  Primigravida                  255(87.6)    36(12.4)     0.282\n  Others                        171(84.2)    32(15.8)     \n  Sleep quality                                           \n  Poor                          28(70.0)     12(30.0)     0.001\n  Fair                          222(85.1)    39(14.9)     \n  Good                          177(91.2)    17(8.8)      \n  Smoking history                                         \n  Non                           418(86.0)    68(14.0)     0.227\n  Yes                           9(100.0)     0(0.0)       \n  Alcohol use history                                     \n  Non                           405(86.9)    61(13.1)     0.076\n  Yes                           21(75.0)     7(25.0)      \n  Husband is a migrant worker                             \n  Yes                           230(88.1)    31(11.9)     0.145\n  Non                           182(83.5)    36(16.5)     \n\n[Table 3](#pone.0167597.t003){ref-type=\"table\"} displays the results of the bivariate correlations. Antenatal depression was significantly and positively correlated with negative life events (r = 0.33, p\\<0.01), stresses of pregnancy (r = 0.44, p\\<0.01), and negative automatic thoughts (r = 0.54, p\\<0.01). And the results show each of the four variables was significantly correlated with the other three.\n\n10.1371/journal.pone.0167597.t003\n\n###### Bivariate correlate analysis of the scores of antenatal depression, stresses of pregnancy, negative life events and negative automatic thoughts.\n\n![](pone.0167597.t003){#pone.0167597.t003g}\n\n  ----------------------------------------------- ---------------------- --------------------------------------------- --------------------------------------------- ---------------------------------------------\n                                                  Antenatal depression   Stresses of pregnancy                         Life events scale for pregnant women          Negative automatic thoughts\n  Antenatal depression(n = 495)                   1                      0.44[\\*\\*](#t003fn001){ref-type=\"table-fn\"}   0.33[\\*\\*](#t003fn001){ref-type=\"table-fn\"}   0.54[\\*\\*](#t003fn001){ref-type=\"table-fn\"}\n  Stresses of pregnancy(n = 481)                                         1                                             0.27[\\*\\*](#t003fn001){ref-type=\"table-fn\"}   0.49[\\*\\*](#t003fn001){ref-type=\"table-fn\"}\n  Life events scale for pregnant women(n = 491)                                                                        1                                             0.37[\\*\\*](#t003fn001){ref-type=\"table-fn\"}\n  Negative automatic thoughts (n = 495)                                                                                                                              1\n  ----------------------------------------------- ---------------------- --------------------------------------------- --------------------------------------------- ---------------------------------------------\n\n\\*\\*P\\<0.01\n\n[Table 4](#pone.0167597.t004){ref-type=\"table\"} shows the results of logistic regression for predicting the variables of antenatal depression. According to the results, significant predictive variables included sleep quality, the stresses of pregnancy, and negative automatic thoughts. Among them, higher scores of negative automatic thoughts (OR, 1.08; 95% CI, 1.04--1.11) and a higher level of stress of pregnancy (OR, 1.12; 95% CI, 1.04--1.21) were significantly related with a higher level of antenatal depression.\n\n10.1371/journal.pone.0167597.t004\n\n###### Logistic regression analysis of socio-demographic and health behavior factors and life events of pregnancy, stresses of pregnancy, and negative automatic thoughts for antenatal depression.\n\n![](pone.0167597.t004){#pone.0167597.t004g}\n\n                                Independent variable    OR(95% CI)\n  ----------------------------- ----------------------- ---------------------\n  Region                        Mianzhu: Gaobeidian     0.51(0.24--1.09)\n  Age                           Ref: \u226530                \n  18--24                        1.22(0.39--3.76)        \n  25--29                        0.51(0.16--1.61)        \n  Education                     Ref: College or above   \n  Middle school or lower        0.42(0.13--1.32)        \n  High school                   0.59(0.19--1.82)        \n  Monthly family income(USD)    Ref: \u2265816 USD           \n  \\<326                         1.51(0.53--4.32)        \n  326--816                      1.00(0.39--2.59)        \n  Living site                   Village: City           1.17(0.42--3.28)\n  Employment                    Ref: full-time job      \n  Unemployed                    0.90(0.29--2.76)        \n  Part-time job                 1.79(0.29--11.20)       \n  Planned pregnancy             Planned: unplanned      0.89(0.44--1.78)\n  BMI                           \\<23:\u226523                1.42(0.62--3.26)\n  Parity                        Primigravida: Others    0.49(0.21--1.14)\n  Gestation stage               0-28weeks:\\>28weeks     1.65(0.74--3.68)\n  Sleep quality                 Ref: Poor               \n  Good                          0.25(0.08--0.84)        \n  Fair                          0.43(0.14--1.30)        \n  Smoking history               Non: Yes                5.273E8(0.00--0.00)\n  Alcohol use history           Non: Yes                0.82(0.21--3.26)\n  Husband is a migrant worker   Non: Yes                1.38(0.66--2.85)\n  Life events of pregnancy                              1.00(1.00--1.01)\n  Stresses of pregnancy                                 1.12(1.04--1.22)\n  Negative automatic thoughts                           1.08(1.04--1.11)\n\nBased on the regressive model above, we tested three path analytic models that predicted antenatal depression ([Fig 1](#pone.0167597.g001){ref-type=\"fig\"}). In Model 1, negative life events and stresses of pregnancy were used as the predictors, and antenatal depression was the parameter estimate. In Model 2 and 3, negative automatic thoughts were included as a predictor of antenatal depression, and this was also a parameter estimate of negative life events in Model 2 and 3.\n\n![Path models examining associations of negative life events, stresses of pregnancy, negative automatic thoughts and antenatal depression.](pone.0167597.g001){#pone.0167597.g001}\n\nIn Model 1, there are three hypotheses. The first one was that negative life events had a direct effect on antenatal depression. The second one was that stress had a direct effect on antenatal depression. The third one was that stress played a mediating role between negative life events and antenatal depression. The fit indexes were good for this model (\u03c7^2^ (206, n = 275) = 608.265, p\\<0.001; \u03c7^2^/df = 2.95; RMSEA = 0.063, IFI = 0.827, CFI = 0.824).\n\nIn Model 2, there are four hypotheses. The first hypothesis was that negative life events had a direct effect on antenatal depression as important predictors. The second hypothesis was that negative automatic thoughts and stresses of pregnancy had direct effects on antenatal depression separately. The third hypothesis was that stress had a mediating effect between negative life events and antenatal depression, and it had a mediating effect between negative automatic thoughts and antenatal depression. The fourth hypothesis was that negative automatic thoughts had a mediating effect between negative life events and antenatal depression. So we included negative automatic thoughts into Model 2. This Model revealed significant direct and indirect paths from negative life events to antenatal depression, and the fit of the model was better than it of Model 1 (\u03c7^2^ (344, n = 434) = 955.859,p\\<0.001; \u03c7^2^/df = 2.78; RMSEA = 0.060, IFI = 0.859, CFI = 0.857).\n\nWe moved away the path between negative life events and stress, and created Model 3. The significant paths of this model are illustrated in [Fig 1](#pone.0167597.g001){ref-type=\"fig\"}. The model fit was good (\u03c7^2^ (345, n = 434) = 960.604, p\\<0.010; \u03c7^2^/df = 2.78; RMSEA = 0.060, IFI = 0.858, CFI = 0.856).\n\nThe path models demonstrate that antenatal depression was influenced by negative life events, stresses of pregnancy and negative automatic thoughts. In Model 1, the model fit indices indicated that this model was good. In Model 2, after adding negative automatic thoughts, the indices were improved. In Model 3, after moving away the path between negative life events and stresses of pregnancy, the indices did not change much, and the final model was simplified, indicating that the final model was better.\n\nWe assessed how well the three variables could predict antenatal depression in Model 3. The results showed that negative automatic thoughts had a greater direct effect (0.39) than negative life events (0.23) and stresses of pregnancy (0.24) did. The general effect of negative automatic thoughts (0.51) was greater than the effects of negative life events (0.42) and stresses of pregnancy (0.24).\n\nDiscussion {#sec012}\n----------\n\nTo our knowledge, this is the first study examining the mediating effect of automatic thoughts between negative life events and antenatal depression in rural China. The results of this study proved the hypothesis that negative automatic thoughts play an important mediating role between negative life events and antenatal depression.\n\nThe results indicate that there is more than a direct effect between negative life events and antenatal depression \\[[@pone.0167597.ref055]\\]. This result is in agreement with Beck's cognitive behavioral theory \\[[@pone.0167597.ref026]\\]. Previous studies have proved the significant mediating role of automatic thoughts among migrant populations, children and undergraduate students \\[[@pone.0167597.ref023], [@pone.0167597.ref052], [@pone.0167597.ref056]\\]. Moreover, a culturally sensitive cognitive-behavioral therapy was implemented among Chinese pregnant women. It also declared that the key to that intervention was to target the irrational thoughts which were relevant to the life stories told by the pregnant women. And then replace the irrational thoughts with more positive ones \\[[@pone.0167597.ref057]\\]. This current study provided further evidence that for pregnant women in rural China, negative automatic thoughts were also significantly and positively associated with their antenatal depression. From the results of the path analysis, this study also found that by comparison with negative life events, negative automatic thoughts had larger direct and general effects on antenatal depression.\n\nExcept for sleep quality, the current study did not find any other socio-demographic and health behavior factor which was associated significantly with antenatal depression. The result contradicts some previous studies, which found that the risk factors of antenatal depressive symptoms included lower income, lower education, smoking, alcohol abuse, and single status. Some studies in East Asia found that younger age and unemployment could also contribute to antenatal depression \\[[@pone.0167597.ref007], [@pone.0167597.ref010], [@pone.0167597.ref020]--[@pone.0167597.ref022], [@pone.0167597.ref058]\\]. This may be because our sample included a large proportion of pregnant women who were of young ages between 18--29 years old (82.4%). Most of them had educational backgrounds lower than college (80.8%), had monthly family incomes of less than 816 USD (82.0%), and most of them were unemployed (86.9%). The similar characteristics impair the powerfulness of socio-demographic and health behavior factors as predictors.\n\nPrevious studies found that the lack of partner support was among the strongest predictors of depression during pregnancy \\[[@pone.0167597.ref059], [@pone.0167597.ref060]\\]. However, in the current study, there was no significant difference in the depressive level between normal pregnant women who had their husbands around and those whose husbands were migrant workers. It is also known that in this study, over half of the migrant workers (53.2%) visited home at least once a week, and 76.9% of migrant workers contacted their wives every day. It can be drawn that most migrant workers would have a frequent contact with their wives, and this would allow them to provide a good emotional support to their wives. Besides, take financial difficulty, which is another factor associated with perinatal depression, into consideration, since migrant workers could bring the family with more financial income, that could help to decrease their wives' financial stress and antenatal depressive level \\[[@pone.0167597.ref061]\\].\n\nIt is still inconvenient for people in rural China to seek for medical treatment for psychological illness, as a result of a severe lack of medical resources \\[[@pone.0167597.ref062]\\]. What makes it worse, since the stigma towards people with psychological problems still exists, some depressed pregnant women might hesitate to seek for treatment and assistance. Additionally, in order to avoid the risk of medication, some Asian pregnant women are more hesitant to take antidepressants such as SSRIs \\[[@pone.0167597.ref063]\\]. Thus, involving the medical service and socio-cultural background in future studies will help to explore the feature of cognition style of the pregnant women in rural China.\n\nThere are several limitations of this study. First, this study mainly focused on the variable of automatic thoughts. However, as the traditional society in rural China is experiencing great changes in many aspects, not much is known about the situation of community support, family support, marital conflicts, and other external environmental factors regarding pregnant women \\[[@pone.0167597.ref064], [@pone.0167597.ref065]\\]. Thus, without having a deeper consideration about those important issues, our study couldn't uncover more of the characteristics related with the cognition style of Chinese rural pregnant women. Second, as a cross-sectional study, this current study couldn't draw a causal relationship between automatic thoughts and antenatal depression among the participants.\n\nConclusions {#sec013}\n===========\n\nIn conclusion, negative automatic thoughts have a potential important mediating effect between negative life events and antenatal depression.\n\nSupporting Information {#sec014}\n======================\n\n###### This file provides the minimal data set used for the analysis presented in this article.\n\n(SAV)\n\n###### \n\nClick here for additional data file.\n\nIn the process of data collection, Tingting Wang, Min Xu, Xiaoyu Sun, Guannan Jia and Cuiping Li of Sichuan Normal University helped to organize the interviews.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: **Conceptualization:** ZQ XW FL DT.**Data curation:** ZQ YW.**Formal analysis:** YW.**Funding acquisition:** ZQ.**Investigation:** YX XJ XY XD XK.**Methodology:** YW XW.**Project administration:** ZQ.**Resources:** ZQ.**Supervision:** ZQ.**Validation:** ZQ XW.**Writing -- original draft:** YW.\n"}
+{"text": "In the original article, there was a mistake in the legend for Figure 5 as published. The legend for Figure 6 was accidently used instead. The correct legend appears below.\n\n**Figure 5 \\|** The male (blue, *n* = 5) and female (red, *n* = 5) basal serotonin levels were collected for 30 min. A saline injection (marked by a yellow bar) preceded another 30 min of basal serotonin data was collected. At the gray bar, escitalopram (ESCIT, 10 mg kg^\u22121^) was administered and 60 min of basal serotonin data was collection. SEM is shown as a lighter bar behind.\n\nThe authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated.\n\n[^1]: Approved by: Frontiers Editorial Office, Frontiers Media SA, Switzerland\n\n[^2]: This article was submitted to Neuropharmacology, a section of the journal Frontiers in Neuroscience\n\n[^3]: \u2020These authors have contributed equally to this work\n"}
+{"text": "Introduction {#s1}\n============\n\nMonoclonal B-cell lymphocytosis (MBL) is defined by the presence of \\<5\u00d710^9^ clonal B-cells/L in peripheral blood (PB) of healthy individuals [@pone.0052808-Hallek1], [@pone.0052808-Marti1]. Two entities can be distinguished within MBL, based on the absolute count of clonal B-cells: a) those diagnosed in clinical settings and associated with lymphocytosis with a median absolute count of clonal B-cells \\>1.5\u00d710^9^/L; and b) population-screened MBL, the so-called \"low-count\" MBL, with very low median absolute count of clonal B-cells of about 0.05\u00d710^9^/L, identified in population-screening studies of healthy individuals using high-sensitive flow cytometry approaches [@pone.0052808-Dagklis1], [@pone.0052808-Fazi1]. Based on immunophenotypic grounds, MBL can be classified as chronic lymphocytic leukaemia (CLL)-like MBL, with a CD5^+^, CD23^+^ and CD20^low^ phenotype representing the most common subgroup (\u223c 75/80% of MBL), atypical-CLL (CD5^+^, CD20^bright^) and CD5^\u2212^ MBL [@pone.0052808-Marti1]. With advanced flow cytometry techniques, low count \"CLL-like\" MBL is detected in 12%--14% [@pone.0052808-Nieto1], [@pone.0052808-Almeida1] of healthy adults in population-screening studies. Recent research suggests systematic occurrence of clinical CLL-like MBL prior to CLL [@pone.0052808-Landgren1]. However, most CLL-like MBL patients never develop clinical complications and the estimated yearly rate of progression of clinical CLL-like MBL to CLL with treatment requirement is 1--2% [@pone.0052808-Rawstron1]; in turn, the rate of progression of \"low count\" CLL-like MBL is still unknown.\n\nThe aetiology of MBL and CLL remains unknown and few studies have been reported on potential risk factors for MBL. Unambiguous risk factors associated to both CLL and MBL are increasing age [@pone.0052808-Rawstron2] and genetic susceptibility [@pone.0052808-Slager1]. In turn, male predominance is also recurrently reported in CLL but results on MBL are controversial [@pone.0052808-Shim1]. Exposure to pesticide, herbicides and chemical agents has also been associated with CLL [@pone.0052808-Marwick1]. Caucasian ethnicity has long stood as a risk factor for CLL with lower incidence rates among Asian than Caucasian Americans. However, recent studies reported higher CLL incidence rates among Asian US born than Asian foreign born [@pone.0052808-Clarke1] subjects and increasing trends in CLL incidence rates in Taiwan [@pone.0052808-Wu1]. Altogether, these results suggest a potential role for some strong but unidentified environmental factors in the aetiology of CLL. Recently, Moreira et al. (2012) reported that the risk of hospitalisation for infections was more common in newly diagnosed clinical MBL and CLL patients than controls [@pone.0052808-Moreira1]. Other risk factors associated with MBL might include living near a hazardous waste site [@pone.0052808-Shim1].\n\nAs part of a study examining the prevalence of MBL in the general population [@pone.0052808-Nieto1], [@pone.0052808-Almeida1], we investigated potential risk factors associated with \"low count\" MBL using a cross-sectional study design among 452 healthy subjects randomly selected from the Primary Health Care system of the region of Salamanca (Spain). This exploratory study provides initial insights and directions for future research related to MBL, a potential precursor of chronic lymphocytic leukaemia. In particular our findings suggest that exposure to infectious agents leading to serious clinical manifestations in the patient or its surroundings may trigger immune events leading to MBL.\n\nMethods {#s2}\n=======\n\nDesign and Subjects {#s2a}\n-------------------\n\nAs part of a study examining the prevalence of MBL through highly sensitive multicolor flow cytometry in a cohort of 639 healthy adults with normal PB lymphocyte counts from the general population of the Primary Health Care system region of Salamanca (northwest-central Spain) [@pone.0052808-Nieto1], we investigated risk factors associated with \"low-count\" MBL. All these 639 cases have been described previously in clinical, as well as phenotypic/genetic and molecular terms [@pone.0052808-Nieto1], [@pone.0052808-Almeida1]. Among the 639 subjects older than 40 years, 452 (71%) completed a questionnaire with their general practitioner, both blind to the MBL status of the subject. Individuals with MBL were denoted as cases and the remainder classified as controls. MBL cases were further classified as either \"CLL-like\" or non-\"CLL-like\" MBL, based on the presence *versus* absence of a CD5+, CD23+ and CD20^dim^ immunophenotype, respectively. The research protocol was approved by the Ethics Committee of the Cancer Research Center of Salamanca and all participants gave their written informed consent in accordance with the Declaration of Helsinki.\n\nImmunophenotypic Analyses {#s2b}\n-------------------------\n\nThe flow cytometry approach has been described in detail in the previous study [@pone.0052808-Nieto1]. In brief, per case, between 1 and 4 mL of EDTA-anticoagulated PB was immunophenotyped using a highly sensitive 8-color flow cytometry. The minimum number of clustered cellular events required to define the presence of a clonal B-cell population was 50.\n\nStatistical Analyses {#s2c}\n--------------------\n\nAll statistical analyses were performed using STATA10.1 (Statacorp, USA). We used unconditional logistic regression adjusted for age (\\<50, 50--59, 60--69, 70 or more) and sex to calculate odds ratios (OR) and 95% confidence intervals (CI) for MBL in relation to different risk factors from the questionnaire. Further adjustment for family size, number of children or number of siblings was performed as appropriate (for instance, the variable self-reported history of infections among children was further adjusted for number of children). Since pneumococcal and influenza vaccinations are generally provided from the age of 60 years in Spain, further analyses stratified by age (\u226560 years) were performed. Self-reported current drug use was grouped into 14 major groups of drugs as per the Anatomical Therapeutic Chemical Classification System (2003). Sensitivity analysis restricting the outcome to CLL-like MBL was performed.\n\nResults {#s3}\n=======\n\nOverall, 72/452 subjects (16%) were diagnosed with \"low-count\" MBL (mean absolute B-cell count: 0.055; standard deviation: 0.216; maximum: 1.172\u00d710^9^ B-cells/L) ([Table 1](#pone-0052808-t001){ref-type=\"table\"}). Most cases (60/72; 83%) were classified as CLL-like MBL. About half of the cases (48%) and the controls (49%) were males (P\u200a=\u200a0.9). OR of \"low-count\" MBL cases increased with increasing age (P\\<0.001). Cases and controls did not differ in terms of area of recruitment, residence of birth, tobacco and alcohol consumption, body mass index, height, weight and women reproductive history (data not shown).\n\n10.1371/journal.pone.0052808.t001\n\n###### Socio-demographic and descriptive characteristics of \"low-count\" monoclonal B-cell lymphocytosis (MBL) cases and non-MBL subjects (controls).\n\n![](pone.0052808.t001){#pone-0052808-t001-1}\n\n                                                Controls N\u200a=\u200a380   MBL cases N\u200a=\u200a72      OR^1^ & 95% CI\n  -------------------------------------------- ------------------ ------------------ ----------------------\n  **SEX**                                                                            \n  Male                                             181 (48%)           35 (49%)               Ref\n  Female                                           199 (53%)           37 (51%)       1.04 (0.62 to 1.75)\n                                                                                          P-het\u200a=\u200a0.9\n  **AGE**                                                                            \n  \\<50                                              92 (24%)            6 (8%)                Ref\n  50--59                                            99 (26%)           7 (10%)        1.09 (0.35 to 3.35)\n  60--69                                            78 (21%)           19 (26%)       3.74 (1.42 to 9.84)\n  70--79                                            85 (22%)           26 (36%)       4.70 (1.84 to 12.00)\n  80 or more                                        26 (7%)            14 (19%)       8.26 (2.89 to 23.62)\n                                                                                        P-trend\\<0.0001\n  mean (SD)                                         60 (13)            70 (11)       \n  range                                             40 to 97           43 to 95      \n  **AREA OF RECRUITMENT**                                                            \n  Salamanca suburb                                  64 (17%)           8 (11%)                Ref\n  Salamanca centre                                  95 (25%)           19 (26%)       1.50 (0.60 to 3.73)\n  Rural                                            221 (58%)           45 (63%)       1.46 (0.64 to 3.34)\n                                                                                         *P-het\u200a=\u200a0.6*\n  **RESIDENCE AT BIRTH**                                                             \n  Salamanca City                                   221 (59%)           40 (57%)               Ref\n  Salamanca county, excluding Salamanca city        97 (26%)           23 (33%)       1.11 (0.62 to 1.99)\n  Other Spanish counties                            45 (12%)           7 (10%)        0.95 (0.39 to 2.31)\n  Other country                                      9 (2%)             0 (0%)                 NA\n                                                                                         *P-het\u200a=\u200a0.9*\n  **TOBACCO CONSUMPTION**                                                            \n  Never                                            220 (58%)           48 (67%)               Ref\n  Past                                              90 (24%)           15 (21%)       0.86 (0.40 to 1.85)\n  Current                                           69 (18%)           9 (13%)        1.08 (0.44 to 2.65)\n                                                                                         *P-het\u200a=\u200a0.9*\n  **ALCOHOL CONSUMPTION**                                                            \n  Never                                            212 (56%)           42 (60%)               Ref\n  2--4 times/week                                   50 (13%)           7 (10%)        0.92 (0.37 to 2.32)\n  Week-end                                          56 (15%)            4 (6%)        0.57 (0.18 to 1.77)\n  Everyday                                          58 (15%)           17 (24%)       1.57 (0.72 to 3.43)\n                                                                                         *P-het\u200a=\u200a0.2*\n  **BODY MASS INDEX, kg/cm^2^**                                                      \n  \\<25                                              84 (29%)           15 (27%)               Ref\n  25--29                                           129 (45%)           23 (42%)       0.88 (0.42 to 1.85)\n  \u226530                                               75 (26%)           17 (31%)       1.00 (0.45 to 2.22)\n                                                                                        *P-trend\u200a=\u200a1.0*\n  mean (SD)                                        27.6 (4.6)         27.8 (4.0)     \n  *\u2003missing*                                       *92 (24%)*         *17 (24%)*     \n\nAdjusted for age (\\<50, 50--59, 60--69, 70+) and sex.\n\nRef: reference group; N: number; SD: standard deviation; het: heterogeneity.\n\nOR: Odds ratio; CI: confidence interval.\n\nConversely, a clear association with transmission and exposure to infection agents was found ([Figure 1](#pone-0052808-g001){ref-type=\"fig\"}). In detail, \"low-count\" MBL cases were less likely to have reported having pneumococcal (OR: 0.49; 95% CI: 0.25 to 0.95; P\u200a=\u200a0.03) and influenza (OR: 0.52; 95% CI: 0.29 to 0.93; P\u200a=\u200a0.03) vaccination and more likely to have had pneumonia (OR: 3.26; 95% CI: 1.03 to 10.27; P\u200a=\u200a0.04), meningitis (OR: 11.73; 95%CI: 1.45 to 95.13; P\u200a=\u200a0.02) or influenza (OR: 6.72; 95% CI: 0.31 to 146.70; P\u200a=\u200a0.2). Albeit based on small numbers, such association was also supported by an increased number of reported infectious diseases in the children of cases (OR: 2.14; 95%: 0.92 to 5.01; P\u200a=\u200a0.08) and of respiratory diseases among their siblings (OR: 4.35; 95% CI: 1.23 to 15.34; P\u200a=\u200a0.02). Furthermore, the OR for \"low-count\" MBL increased with increasing number of children among cases with children (P\\<0.001), such trend being observed separately in men and women (P\\<0.001). However, childless individuals were also three times more likely to have been diagnosed with MBL compared to individuals with only 1 child and no potential confounders could explain this association.\n\n![Odds ratios for \"low-count\" monoclonal B-cell lymphocytosis by selected potential variables related to infectious agents.\\\nRef: reference group; N: number; het: heterogeneity; OR: Odds ratio; CI: confidence interval ^1^: Adjusted for age (\\<50, 50--59, 60--69, 70+) and sex ^2^: Further adjusted for family size (using the total number of children and siblings; categories: \\<6; 6 or more; missing). ^3^: Further adjusted for number of siblings (categories: \\<2; 3 or more; missing) *^4^:* Further adjusted for number of children (0/2; 3or more; missing) *^5^:* Adjusted for age (\\<70, 70+) and sex, N~controls~\u200a=\u200a189; N~cases~\u200a=\u200a59 P-values were calculated using Wald-test. Black squares indicate OR, the area of each square being proportional to the amount of statistical information contributed. Horizontal lines represent 95% CI.](pone.0052808.g001){#pone-0052808-g001}\n\nMBL cases were also more likely to report haematological cancer and solid cancer among their first-degree relatives, compared to controls ([Table 2](#pone-0052808-t002){ref-type=\"table\"}). Conversely, cases and controls did not differ in terms of the current use of any of the 14 different groups of drugs. However, 6% (N\u200a=\u200a4) of \"low-count\" MBL cases versus 12% (N\u200a=\u200a44) of controls (OR: 0.30; 95% CI: 0.10 to 0.87; P\u200a=\u200a0.03), reported treatment for diabetes (Table S1 in [File S1](#pone.0052808.s001){ref-type=\"supplementary-material\"}). Noteworthy, all associations described above were retained when the analyses were restricted to CLL-like MBL ([Table 3](#pone-0052808-t003){ref-type=\"table\"}). MBL load was not associated with any of the exposure variables, whereas normal B-cell count was decreasing with increasing age and was higher in women than men (data not shown).\n\n10.1371/journal.pone.0052808.t002\n\n###### Odds ratios (OR) estimates, with 95% confidence intervals (CI), for \"low-count\" monoclonal B-cell lymphocytosis by self-reported family history of cancer.\n\n![](pone.0052808.t002){#pone-0052808-t002-2}\n\n                                                                                                         Controls N\u200a=\u200a380   MBL cases N\u200a=\u200a72             OR^1^ & 95% CI\n  ----------------------------------------------------------------------------------------------------- ------------------ ------------------ ------------------------------------\n  **Ever had family history of haematological cancer**                                                                                        \n  None                                                                                                      356 (94%)           63 (87%)                      ref\n  1 family member affected[2](#nt107){ref-type=\"table-fn\"}                                                   22 (12%)           8 (28%)               1.96 (0.79 to 4.86)\n  2 family members affected[2](#nt107){ref-type=\"table-fn\"}                                                   2 (1%)             1 (3%)               3.70 (0.28 to 49.35)\n                                                                                                                                                        *P-trend\u200a=\u200a0.07*\n  Ever (any family members)[2](#nt107){ref-type=\"table-fn\"}                                                  24 (6%)            9 (13%)          2.07 (0.87 to 4.93); *P\u200a=\u200a0.1*\n  \u2003Participant                                                                                               1 (\\<1%)            0 (0%)                        NA\n  \u20031^st^ degree relatives[2](#nt107){ref-type=\"table-fn\"}                                                    22 (6%)            9 (13%)         2.23 (0.93 to 5.38); *P\u200a=\u200a0.07*\n  *\u2003\u2003\u2003Father*                                                                                                *4 (1%)*           *4 (6%)*       11.49 (2.42 to 54.55); *P\u200a=\u200a0.002*\n  *\u2003\u2003\u2003Mother*                                                                                                *4 (1%)*           *0 (0%)*                      *NA*\n  *\u2003\u2003\u2003Sibling* [3](#nt108){ref-type=\"table-fn\"}                                                              *11 (3)*           *6 (8)*          2.46 (0.84 to 7.17); *P\u200a=\u200a0.1*\n  *\u2003\u2003\u2003Children* [4](#nt109){ref-type=\"table-fn\"}                                                             *4 (1%)*           *0 (0%)*                      *NA*\n  \u20032^nd^ degree relatives                                                                                    2 (\\<1%)            0 (0%)                       *NA*\n  **Ever had family history of solid cancer**                                                                                                 \n  None                                                                                                      168 (44%)           26 (36%)      \n  1 family member affected[2](#nt107){ref-type=\"table-fn\"}                                                  145 (38%)           32 (44%)              1.57 (0.87 to 2.84)\n  2 family members affected[2](#nt107){ref-type=\"table-fn\"}                                                  48 (13%)           9 (13%)               1.57 (0.87 to 2.84)\n  3 family members affected[2](#nt107){ref-type=\"table-fn\"}                                                  18 (5%)             4 (6%)               1.66 (0.49 to 5.65)\n  4 family members affected[2](#nt107){ref-type=\"table-fn\"}                                                  1 (\\<1%)            1 (2%)              9.87 (0.42 to 232.60)\n                                                                                                                                                        *P-trend\u200a=\u200a 0.2*\n  Ever (any family members)[2](#nt107){ref-type=\"table-fn\"}                                                 212 (56%)                            1.54 (0.89 to 2.66); *P\u200a=\u200a0.1*\n  \u2003Participant                                                                                               27 (7%)                             0.83 (0.32 to 2.15); *P\u200a=\u200a0.7*\n  \u20031^st^ degree relatives[2](#nt107){ref-type=\"table-fn\"}                                                   164 (43%)                           2.02 (1.17 to 3.47); *P\u200a=\u200a0.01*\n  *\u2003\u2003\u2003Father*                                                                                                87 (23%)                            1.03 (0.54 to 1.97); *P\u200a=\u200a0.9*\n  *\u2003\u2003\u2003Mother*                                                                                                57 (15%)                            0.86 (0.38 to 1.96); *P\u200a=\u200a0.7*\n  *\u2003\u2003\u2003Sibling* [3](#nt108){ref-type=\"table-fn\"}                                                              60 (16%)                          2.92 (1.59 to 5.36); *P\u200a=\u200a0.001\\**\n  *\u2003\u2003\u2003\u2003Prostate* ^,^ [5](#nt110){ref-type=\"table-fn\"}                                                        *6 (2%)*                          *6.31 (1.99 to 20.00); P\u200a=\u200a0.002*\n  *\u2003\u2003\u2003Children* [4](#nt109){ref-type=\"table-fn\"}                                                              6 (2%)                             1.56 (0.28 to 8.59); *P\u200a=\u200a0.6*\n  \u20032^nd^ degree relatives                                                                                    62 (16%)                            0.95 (0.39 to 2.32); *P\u200a=\u200a0.9*\n  **Ever had family history of solid and/or haematological cancers** [2](#nt107){ref-type=\"table-fn\"}                                         \n  Never                                                                                                     156 (41%)           20 (28%)                      ref\n  With family history of solid cancer only                                                                  200 (53%)           43 (60%)              1.89 (1.04 to 3.45)\n  With family history of haematological cancer only                                                          12 (3%)             6 (8%)               4.23 (1.33 to 13.50)\n  With family history of both type of cancers                                                                12 (3%)             3 (4%)               1.93 (0.46 to 8.09)\n                                                                                                                                                    *P-heterogeneity\u200a=\u200a0.02*\n\nRef: reference group; N: number; NA: not estimated; het: heterogeneity.\n\nOR: Odds ratio; CI: confidence interval; 1^st^ degree relatives: parents, siblings and children; 2^nd^ degree relatives: grand-parents.\n\nAdjusted for age (\\<50, 50--59, 60--69, 70+) and sex.\n\nFurther adjusted for family size (using the total number of children and siblings; categories: \\<6; 6 or more; missing).\n\nFurther adjusted for number of siblings (categories: \\<2; 3 or more; missing).\n\nFurther adjusted for number of children (\\<2; 3 or more; missing).\n\nAdjusted for number of brothers (\\<3; 3 or more; missing).\n\n10.1371/journal.pone.0052808.t003\n\n###### Odds ratios (OR) estimates, with 95% confidence intervals (CI), for \"low-count\" CLL-like monoclonal B-cell lymphocytosis (60 out of 72 cases) by previous associated factors.\n\n![](pone.0052808.t003){#pone-0052808-t003-3}\n\n                                                               ControlsN\u200a=\u200a380   MBL casesN\u200a=\u200a60             OR^1^ & 95% CI\n  ----------------------------------------------------------- ----------------- ----------------- ------------------------------------\n  **Age**                                                                                         \n  \\<50                                                            92 (24%)           6 (10%)                      Ref\n  50--59                                                          99 (26%)           5 (8%)               0.78 (0.23 to 2.66)\n  60--69                                                          78 (21%)          16 (27%)              3.19 (1.19 to 8.57)\n  70--79                                                          85 (22%)          22 (37%)              4.06 (1.56 to 10.52)\n  80 or more                                                       26 (7%)          11 (18%)              6.56 (2.21 to 19.45)\n  mean (SD)                                                        60 (13)           69 (12)      \n                                                                                                           *P-trend\\<0.0001*\n  range                                                           40 to 97          43 to 93      \n  **Self-reported infectious diseases among**                                                     \n  Children[2](#nt112){ref-type=\"table-fn\"}                         32 (8%)           8 (13%)         2.33 (0.95 to 5.69); P\u200a=\u200a0.06\n  **Self-reported respiratory diseases among**                                                    \n  Sibling[3](#nt113){ref-type=\"table-fn\"}                          7 (2%)            4 (7%)          4.32 (1.13 to 16.56); P\u200a=\u200a0.03\n  **Self-reported vaccination against**                                                           \n  Pneumococcus[4](#nt114){ref-type=\"table-fn\"}                    78 (41%)          14 (29%)         0.48 (0.24 to 0.99); P\u200a=\u200a0.05\n  Influenza                                                       153 (40%)         26 (43%)         0.58 (0.31 to 1.07); P\u200a=\u200a0.08\n  **Self-reported respiratory tract infections**                                                  \n  Pneumonia                                                        8 (2%)            6 (10%)         4.18 (1.31 to 13.27); P\u200a=\u200a0.02\n  Influenza                                                       1 (\\<1%)           1 (2%)          7.49 (0.35 to 162.48); P\u200a=\u200a0.2\n  Meningitis                                                      3 (\\<1%)           1 (2%)         9.79 (0.80 to 119.99); P\u200a=\u200a0.07\n  **Number of children**                                                                          \n  None                                                            57 (15%)          14 (23%)              3.24 (0.96 to 10.85)\n  1                                                               67 (18%)           4 (7%)                       Ref\n  2                                                               141 (37%)         11 (18%)              1.07 (0.32 to 3.61)\n  3                                                               66 (18%)          11 (18%)              2.07 (0.59 to 7.32)\n  4 or more                                                       46 (12%)          20 (33%)              4.23 (1.24 to 14.43)\n                                                                                                    **P-trend (in parous)\\<0.0001**\n  **Diabetes treatment**                                          44 (12%)           3 (5%)         0.27 (0.08 to 0.91); *P\u200a=\u200a0.04*\n  **Ever had family history of haematological cancer**                                            \n  Ever (any family members)[5](#nt115){ref-type=\"table-fn\"}        24 (6%)           8 (13%)        2.18 (0.88 to 5.40); *P\u200a=\u200a0.09*\n  1^st^ degree relatives[5](#nt115){ref-type=\"table-fn\"}           22 (6%)           8 (13%)        2.36 (0.94 to 5.94); *P\u200a=\u200a0.07*\n  *\u2003\u2003Father*                                                       4 (1%)            4 (7%)        16.30 (3.30 to 80.58); *P\u200a=\u200a0.001*\n  **Ever had family history of solid cancer**                                                     \n  1^st^ degree relatives[5](#nt115){ref-type=\"table-fn\"}          164 (43%)         34 (57%)        1.72 (0.97 to 3.08); *P\u200a=\u200a0.07*\n  *\u2003Sibling* [4](#nt114){ref-type=\"table-fn\"}                     60 (16%)          21 (35%)        *2.43 (1.25 to 4.73); P\u200a=\u200a0.009*\n  *\u2003\u2003Prostate* [6](#nt116){ref-type=\"table-fn\"}                    6 (2%)            6 (10%)       *5.95 (1.71 to 20.73); P\u200a=\u200a0.005*\n\n^1^Adjusted for age (\\<50, 50--59, 60--69, 70+) and sex.\n\nFurther adjusted for number of children (\\<2; 3 or more; missing).\n\nFurther adjusted for number of siblings (categories: \\<2; 3 or more; missing).\n\n*Among 60 years old and older* ^,^ *N~controls~\u200a=\u200a189; N~cases~\u200a=\u200a49.*\n\nFurther adjusted for family size (using the total number of children and siblings; categories: \\<6; 6 or more; missing).\n\nFurther adjusted for number of brothers (\\<3; 3 or more; missing).\n\nRef: reference group; N: number; OR: Odds ratio; CI: confidence interval; SD: standard deviation.\n\nDiscussion {#s4}\n==========\n\nTo our knowledge, this is the first epidemiological study investigating risk factors associated with MBL in the general population. In contrast to CLL, the male predominance was not observed for \"low-count\" MBL in our data whereas, as expected, OR of cases increased with increasing age. Our main findings were that lifetime exposure to several infectious agents might be associated with MBL aetiology.\n\nAssuming that MBL, in particular CLL-like subtype, is a precursor of CLL [@pone.0052808-Landgren1], our findings are consistent with previous studies on CLL reporting that a personal history of pneumonia was associated with subsequent development of CLL [@pone.0052808-Landgren2]--[@pone.0052808-Landgren3]. Recent data also showed an increased risk of hospitalisation for infections among newly diagnosed clinical MBL than controls [@pone.0052808-Moreira1]. Since general population \"low-count\" CLL-like MBL is 100 times more common than CLL among the elderly [@pone.0052808-Marti2] and persists over time without clinical progression [@pone.0052808-Fazi1], it was therefore postulated that \"low-count\" CLL-like MBL are less likely than clinical CLL-like MBL to progress to CLL: among 76 examined patients (median age: 66 years; range: 25--92 years) with low-count population-screening CLL-like MBL, none developed CLL after a 3-year median follow-up, even if they were found to carry 13q deletion with the same frequency to that observed in patients with newly diagnosed CLL and clinical MBL [@pone.0052808-Fazi1]. Hence, \"low-count\" CLL-like MBL might even be a normal stage of the immunosenescence process [@pone.0052808-Fazi1]. The observed link with infections for \"low-count\" CLL-like MBL might appear of limited value if the rate of progression of \"low-count\" CLL-like MBL to CLL is very low; however this progression rate is still unknown and large prospective studies are needed to evaluate which patients with MBL will advance to CLL. The progression from CLL-like MBL of \"low-count\" or clinical populations to CLL might then be due to the occurrence in MBL cells of specific biological and molecular profiles [@pone.0052808-Scarfo1], potentially combined with exposure to some unknown environmental risk factors. Several hypotheses have been advanced to explain an infectious route in CLL. Accordingly, recent reports [@pone.0052808-Ghia1] suggest that chronic and persistent antigenic stimulation may have a role in MBL and CLL aetiology and, respiratory tract infections could particularly be triggers for CLL development [@pone.0052808-Landgren2]. Hepatitis C virus has also been suggested as a potential candidate for dysregulation of the immune system with recent data showing that the three MBL subtypes are more frequent in HCV infected individuals than in the general population [@pone.0052808-Fazi2]. In our data, we could not distinguish between the different hepatitis and no statistical significant association was observed with overall self-reported history of hepatitis or vaccination against hepatitis.\n\nWe observed that family history of cancer in particular haematological cancers was more common in patients with MBL than controls. In the published literature, MBL is indeed more frequent in families with CLL cases, in particular among first-degree relatives of patients with familial CLL [@pone.0052808-Matos1]. Unfortunately, we could not differentiate here the association by haematological subtypes as this information was unavailable. In relation to the use of diabetes treatment and chronic lymphocytic leukaemia, data in the literature is inconsistent; while some authors report an increased risk of CLL among diabetic patients [@pone.0052808-Ekstrom1], others suggest a lower risk of lymphoma [@pone.0052808-Fortuny1] and other cancers [@pone.0052808-LaVecchia1] particularly when using metformin. Further studies are needed to clarify this finding.There are several limitations of this study. The questionnaire provided limited information on occupation (either job title or company type/name was reported). Hence, occupational exposure to contaminants, organic solvents, herbicides or infections could not be examined. Furthermore the origin and type (viral/bacterial and chronic/acute) of infections that occurred among the children and other family members were not specified. However, this type of data is clearly difficult to collect in epidemiological settings. Other potential limitations of our study are the lack of validation of self-reports of family history even though questionnaires were filled in with general practitioners, as well as the limited information on timing of infection relative to date of recruitment. Despite the fact that the number of cases in each stratum was relatively small for some categories and that the number of comparisons performed was high, our results point out a potential role of infectious agents in the development of \"low-count\" MBL in the general population, particularly of those involved in respiratory infections. However, a reverse causality effect that would result on detecting a more suppressed immune system among subjects with MBL or already in the pathway of CLL cannot be excluded, as decreased numbers of normal PB B-cells as well as CD4+CD8+ double-positive T-cells have been specifically reported in \\\"low-count\" MBL [@pone.0052808-Hauswirth1]. Recent findings showing that regulatory T-cells increase gradually from controls to \"clinical\" MBL to CLL [@pone.0052808-DArena1], and that most \"low-count\" MBL subjects show T-cell clones especially among CD4+CD8+ T-cells [@pone.0052808-Fazi1] further support the hypothesis of an altered immune system of MBL patients. Further studies analyzing dysregulations of the immune system in MBL compared to controls are required. Although selection biases cannot be ruled out, the robustness of the study relies in that neither the interviewers nor the study subjects were aware of the MBL status. This small exploratory study provides initial insights and directions for future research. Further studies are needed to evaluate the association between MBL and CLL and to examine the role of infectious agents in the development and progression of both entities.\n\nSupporting Information {#s5}\n======================\n\n###### \n\nTable S1 and list of members of the Primary Health Care Group of Salamanca for the Study of MBL (List S1)\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\nWe thank all the subjects who participated in the study, staff from the Primary Health Care system of the region of Salamanca. We thank Teresa Alonso, Elisabet Guin\u00f3, Laura Costas and Claudia Robles for their work on the classification of diseases and drugs.\n\n[^1]: **Competing Interests:**Dr. Silvia de Sanjos\u00e9 is a co-author and she serves as an Academic Editor for this journal. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.\n\n[^2]: Conceived and designed the experiments: JA AO DC SS. Performed the experiments: JA AO WGN AR-C. Analyzed the data: DC YB. Contributed reagents/materials/analysis tools: DC YB JA WGN AR-C SM-C. Wrote the paper: DC JA AO MG SS YB. Selected the cohort of individuals and performed the coordination with the Primary Health Care Group of Salamanca of the study of MBL (List S1 in File S1): AR. Collected samples and data from healthy individuals and contributed to coordinate sample collection: PF-N.\n"}
+{"text": "INTRODUCTION\n============\n\nAlthough root canal treatments have a high rate (90%) of success when properly conducted \\[[@B1][@B2]\\], failures may occur and are often associated with poorly treated canals \\[[@B2]\\]. In these situations, nonsurgical root canal retreatment is indicated to improve root canal disinfection and debridement, and to place a consistent and homogeneous root canal filling \\[[@B2]\\]. However, several reports have shown that considerable amounts of root filling materials remain after re-instrumentation \\[[@B3][@B4][@B5]\\]. This residual material should be removed because it may harbor bacteria, and weaken the quality of seal of the new root canal filling \\[[@B6][@B7]\\].\n\nMany techniques have been proposed for removing gutta-percha and sealer during root canal retreatment procedures using various instruments, including hand files, burs, nickel-titanium (NiTi) rotary systems, adjunctive solvents, and/or ultrasonics. Some endodontic systems are specially designed for use in retreatment procedures, such as the Mtwo retreatment system (VDW, Munich, Germany). This system comprises 2 instruments (R15/0.05 and R25/0.05) that have a 2-fluted file with an S-shaped cross-sectional design and a cutting tip in order to facilitate penetration into the obturation mass. Nevertheless, all retreatment systems leave residual debris in the canal walls after endodontic retreatment \\[[@B2][@B7][@B8][@B9]\\]. Reciprocating systems, such as Reciproc (VDW) and WaveOne (Dentsply Maillefer, Ballaigues, Switzerland), that were initially developed for root canal preparation can also be used for retreatment purposes. Reciproc instruments have the same cross-sectional design as the Mtwo instruments and are available in 3 different sizes: R25 (25/0.08) for narrow canals, R40 (40/0.06) for medium-volume canals and R50 (50/0.05) for large canals. There is growing evidence of the safety and benefits of using these instruments during retreatment cases \\[[@B8][@B9][@B10]\\]. The reciprocating motion relieves stress on the instrument by alternating counterclockwise and clockwise movements, which extends the life span of the NiTi instrument when compared to continuous rotation \\[[@B11]\\]. Moreover, the idea of a reduced number of NiTi instruments needed to remove root canal obturation materials and to enlarge the canal to a minimum acceptable taper size is indeed appealing due to its safety and technique simplification. From an educational point of view, such procedural simplification and safety can lead to improved outcomes when performed by novice operators \\[[@B12][@B13]\\].\n\nThe aim of this study was to evaluate the amount of remaining root filling materials after retreatment procedures performed by undergraduate students using manual, rotary, and reciprocating instruments using micro-computed tomographic (micro-CT) analysis. The incidence of instrument fracture and the instrumentation time were also evaluated. The null hypothesis tested was that there would be no differences in the quality of root canal filling material removal performed by undergraduate students using different techniques.\n\nMATERIALS AND METHODS\n=====================\n\nTeeth selection\n---------------\n\nThis study protocol was approved by the Grande Rio University Ethics Committee (ID: CAAE 47448315.2.0000.5283). Fifty-three maxillary single rooted teeth presenting one canal and similar root length were selected from a pool of teeth. Buccolingual and mesiodistal radiographs were taken to select teeth with oval canals according to the technique previously described \\[[@B8]\\]. Canal curvature was determined using Schneider\\'s method \\[[@B14]\\], and the angle of curvature of each specimen was measured through an image analysis program (AxioVision 4.5, Carl Zeiss Vision, Hallbergmoos, Germany). Access openings were performed using diamond burs and a size 10 K-file (Dentsply Maillefer) was used to initially negotiate the canal. Only straight teeth with root curvature \\< 10\u00b0, approximately 22 \u00b1 1 mm in length, and an initial apical size equivalent to a size 10 K-file were included (*n* = 30).\n\nRoot canal preparation and filling procedures\n---------------------------------------------\n\nWorking length (WL) was established introducing a size 10 K-file into the root canal until its tip was visible at the apical foramen under an operating microscope (DFV, Valen\u00e7a, RJ, Brazil) with \u00d720 magnification, and reducing 1 mm short of this measurement. After WL determination, the apical foramen diameters of all teeth were standardized to a size 15 K-file (Dentsply Maillefer).\n\nThe same experienced operator performed all canal shaping and filling procedures. Teeth were instrumented with Reciproc R25 files (VDW) as recommended by the manufacturer. Irrigation was performed using a total volume of 25 mL of 2.5% sodium hypochlorite (NaOCl). Smear layer was removed using 3 mL of 17% ethylenediaminetetraacetic acid (EDTA) for 3 minutes. Then, canals were irrigated with 2 mL of bi-distilled water, dried with R25 paper points (VDW), and finally filled with R25 gutta-percha cones (VDW) and AH Plus sealer (Dentsply De Trey, Konstanz, Germany). Condensation of the root filling material was achieved by using the continuous wave of condensation technique, and the down pack (EQ-V Endodontic Obturation System, MetaBiomed, Cheongju, Korea) was established at 5 mm from the WL. Then, the remainders of the canals were filled with the aid of a McSpadden condenser (Dentsply Maillefer). The access cavities were sealed with Cavit G (3M ESPE, Seefeld, Germany). After that, the specimens were radiographed in mesiodistal and buccolingual directions to confirm the quality of the fillings. A lack of voids in all samples was noted in the radiographs; therefore, none of the samples was discarded. The specimens were kept at 100% humidity and 37\u00b0C for 4 weeks to allow complete setting of the sealer.\n\nRoot canal retreatment\n----------------------\n\nTen undergraduate dental students of the same semester had 3 fifty-minute lectures presenting and comparing manual, rotary and reciprocating techniques, their main systems and information regarding their properties, and constructional features. After that, demonstrations of endodontic retreatments using the aforementioned techniques were performed. Each participant received a handout defining the sequence of instruments to be used in each technique. Then, each undergraduate student received 3 teeth and retreated them using each of the 3 retreatment techniques: manual, rotary, and reciprocating groups, which used K-file, Mtwo retreatment file, and Reciproc file, respectively. Thus, each retreatment group included ten different specimens.\n\n1\\. Retreatment with manual K-files (manual K-file group). A size 3 Gates-Glidden drill (Dentsply Maillefer) was used at 5,000 rpm to remove part of the coronal filling material in order to create a reservoir for the solvent. An increment of eucalyptol (0.1 mL) was introduced into the root canal to soften the gutta-percha for 30 seconds before further instrumentation. Then, all root canals were re-instrumented to the original WL using manual K-files (Dentsply Maillefer) up to size 40 with a push-pull movement alternated with a rotary motion. After that, manual K-files in sizes from 45 to 60 were used in a step-back technique.\n\n2\\. Retreatment with Mtwo retreatment files (Mtwo retreatment group). A size 3 Gates-Glidden drill and an increment of eucalyptol (0.1 mL) were used at the same manner as described in the manual K-file group. Then, Mtwo retreatment files R15/0.05 and R25/0.05, and Mtwo rotary files in sizes of 30/0.05, 35/0.04, and 40/0.04 (VDW) were used with an electric motor (VDW Silver, VDW). Speed setting and torque for each instrument were used as recommended by the manufacturer. The instruments were applied using in-and-out movement, associated to a short stroking/brushing motion in a coronal direction to the full original WL. After 3 gentle in-and-out motion strokes, the instruments were removed from the canal, cleaned off by inserting into a clean stand with sponge until the WL was reached.\n\n3\\. Retreatment with Reciproc files (Reciproc group). A size 3 Gates-Glidden drill and an increment of eucalyptol (0.1 mL) were used at the same manner as described in the manual K-file group. Reciproc files were moved in an apical direction in a reciprocating motion using in-and-out pecking motion of about 3 mm in amplitude with a light apical pressure. Brushing motion against the lateral canal walls was also used. After 3 pecking motions, the instrument was removed from the canal, and its flutes were cleaned off. An R25 file was used to remove the filling materials until WL was reached. Afterward, the R40 file (VDW) was used with the same protocol.\n\nEach set of instruments was used in one tooth and then discarded. The criteria for the completion of retreatment procedures were smooth canal walls and no evident filling material on the files. The same irrigation protocol was applied in all groups. A 2.5% NaOCl was delivered using a 30-gauge side-vented needle (Max-i-Probe, Dentsply Rinn, Elgin, IL, USA) between each instrument change, and a total of 25 mL of 2.5% NaOCl was delivered per canal. After final instrumentation, all canals were dried with paper points (Dentsply Maillefer) and the access openings were sealed with Cavit G.\n\nMicro-CT scanning procedures and evaluation\n-------------------------------------------\n\nThe sample was scanned using a micro-CT device (SkyScan 1174v2, Bruker microCT, Kontich, Belgium) after root canal filling and retreatment procedures using the following parameters: 50 kV, 800 mA, isotropic resolution of 21.8 \u03bcm, 360\u00b0 rotation around the vertical axis, rotation step of 0.5\u00b0, and frame averaging of 3, using a 1-mm thick aluminum filter. Then, the NRecon v.1.6.9 software (Bruker microCT) was used to reconstruct the acquired images into cross-sectional slices, and the volume of interest was selected to extend from the cemento-enamel junction to the apex of the root, resulting in the acquisition of 700--800 transverse cross sections per tooth. CTAn v.1.16.1 software (Bruker microCT) was used for measuring initial and residual volumes of filling materials (mm^3^). The percentage of residual volume relative to the initial volume of the filling material was calculated.\n\nStatistical analysis\n--------------------\n\nThe preliminary analysis of the raw pooled data revealed a bell-shaped distribution (D\\'Agostino and Person omnibus normality test), and statistical analysis was performed using parametric methods. One-way analysis of variance (ANOVA) was followed by *post hoc* pair-wise comparisons, which were performed using Tukey multiple comparisons. The alpha-type error was set at 0.05, and SPSS 11.0 (SPSS Inc., Chicago, IL, USA) was used as the analytic tool.\n\nRESULTS\n=======\n\nRemaining filling materials were observed in all specimens. The mean volume (%) of the remaining filling material is shown in [Table 1](#T1){ref-type=\"table\"}. All techniques left an average of 3.67%--16.21% of root filling materials in the root canal after retreatment procedures. The mean volume of remaining material was significantly lower in the Reciproc group when compared to manual K-file and Mtwo retreatment groups ([Figure 1](#F1){ref-type=\"fig\"}, *p* \\< 0.05).\n\n###### Remaining filling material (%) and the time required to complete endodontic retreatments (minutes)\n\n![](rde-43-e5-i001)\n\n  Group                   Remaining filling material (%)   Time (min)\n  ----------------------- -------------------------------- ------------------\n  Manual K-file           13.86 \u00b1 17.83^a^                 23.33 \u00b1 15.70^a^\n  Mtwo retreatment file   16.21 \u00b1 17.08^a^                 10.11 \u00b1 6.15^b^\n  Reciproc file           3.62 \u00b1 5.52^b^                   10.11 \u00b1 5.51^b^\n\nThe values are shown as mean \u00b1 standard deviation. Different superscript letters demonstrate significant differences in the column (*p* \\< 0.05).\n\n![Reconstructed 3-dimensional micro-computed tomographic (micro-CT) images of a given specimen of each group showing the presence of remaining root canal fillings in (A) bucco-lingual and (B) mesio-distal views. (C) Cross-sections at the (c) cervical, (m) middle, and (a) apical thirds showing the presence of remaining root canal fillings.](rde-43-e5-g001){#F1}\n\nThe mean operating time is also presented in [Table 1](#T1){ref-type=\"table\"}. The time required to achieve a satisfactory removal of root canal filling material and refinement was significantly lower in the Mtwo retreatment and Reciproc groups (*p* \\< 0.05). No instrument fracture or accident was observed in any of the groups.\n\nDISCUSSION\n==========\n\nAlthough it has not been shown that there is a correlation between the presence of remaining material and retreatment failure \\[[@B15]\\], removing as much filling material as possible would minimize pulp tissue or any bacteria that may remain hidden in the canal from prior treatment \\[[@B16][@B17]\\]. Therefore, the present study evaluated the amount of remaining root filling materials after retreatment procedures performed by undergraduate students using manual K-files, Mtwo retreatment files, and Reciproc files. Although several techniques have been used to assess the remaining filling material inside the canal walls \\[[@B3][@B4][@B9][@B10][@B17]\\], micro-CT imaging was used because this highly accurate, noninvasive, and reproducible technology enables a precise 3-dimensional quantitative evaluation of residual filling material \\[[@B7][@B8]\\]. Moreover, this technique not only overcomes the limitations of 2-dimensinal images, but it may also limit potential operator bias in the interpretation of the results \\[[@B18]\\].\n\nPrevious studies have suggested that undergraduate dental students can successfully prepare root canals with rotary instruments, and these instruments are increasingly being integrated into clinical courses \\[[@B13][@B19][@B20]\\]. Up to now, no study evaluated the quality of endodontic retreatment performed with manual, rotary, and reciprocating techniques by undergraduate students. Students enrolled in this study received a standardized introduction to the use of the 3 different instrumentation systems that were used in retreatment cases. Prior to treating the samples in this study, retreatment demonstrations using manual K-files, Mtwo retreatment, and Reciproc instruments were shown to each student.\n\nIn the current study, residual filling materials were seen in all specimens, regardless of the retreatment technique used. Similar findings have also been reported in previous studies \\[[@B3][@B4][@B7][@B8][@B9][@B10]\\]; however, the Reciproc technique exhibited the lowest percentage of remaining obturation material. Thus, the null hypothesis tested was rejected. Two previous studies presented controversial results regarding filling material removal when comparing Mtwo retreatment and Reciproc systems \\[[@B9][@B21]\\]. Zuolo *et al*. \\[[@B9]\\] showed that Reciproc files left the lowest mean percentage (4.57%) of remaining filling material, whereas Alves *et al*. \\[[@B21]\\] showed that the percentage of filling material removed with Mtwo instruments (96%) was significantly higher than Reciproc instruments (89%). This discrepancy in the results may be explained by differences in the methodologic design. While maxillary central incisors were analyzed based on photographs taken after splitting the teeth in the former study \\[[@B9]\\], a micro-CT tool was used to evaluate mesial canals of mandibular molars in the latter one \\[[@B21]\\]. Mtwo and Reciproc systems were chosen because they are similar in design, therefore allowing a well-controlled comparison between a single-file reciprocating system and a multi-file rotary system. Although these systems are similar in design, it was not possible to standardize the tapers because Reciproc has a variable taper along the shaft. This feature combined with the reciprocating movement can partly explain the improved performance of Reciproc instruments in the removal of root fillings herein. Moreover, contrary to the previous studies where retreatments were always performed by only one operator with experience in all tested techniques, all retreatment procedures were performed by undergraduate students in the present study. None of them had any type of experience using the 3 retreatment techniques. Thus, the present results showed the reduced learning curve of Reciproc instruments due to its simple usability.\n\nRegarding the operating time, both Mtwo retreatment and Reciproc systems required less time to remove root filling material than manual instruments. This result is consistent with others who reported that mechanical instrumentation was significantly more rapid than hand files \\[[@B22][@B23][@B24]\\]. It can be hypothesized that the active tip and the cutting blades of Mtwo retreatment system and the special design of Reciproc instruments as well as its reciprocating motion positively influenced the time required for the retreatment procedures \\[[@B24][@B25]\\].\n\nIt is well known that solvents can be used to soften and dissolve gutta-percha in the root canals to facilitate its penetration and removal \\[[@B4][@B26]\\]. In the present study, a small (0.1 mL) increment of eucalyptol was used only at the beginning of the retreatment procedures in all groups to soften the coronal filling material, and to improve the penetration of the files.\n\nAccording to the results of the scans after canal obturation, no statistical difference was observed between the mean percentage volumes of the filling materials in all groups before retreatment procedures. Therefore, it allowed a better comparison of the amount of remaining filling material after the second scans because of standardized samples at the beginning of the retreatment procedures. In this study, maxillary single rooted teeth were used; in fact, in most of the studies evaluating retreatment techniques, straight root canals have been used to simplify the standardization of the specimens \\[[@B3][@B4][@B9][@B10][@B17][@B19]\\]. Some previous retreatment studies have filled teeth with lateral condensation; however, the continuous wave of condensation technique was used in this study because it produces a better root canal filling adaptation to the root canal walls than lateral condensation \\[[@B27]\\].\n\nTwo important methodological aspects should be discussed in the present study. First, since the students enrolled were not experienced, the order of the retreatment (manual, rotary or reciprocating) could affect the results. Moreover, depending on the student, the results could be affected, *e.g.*, one skillful student could show better results. Nevertheless, it is important to emphasize that even an unskilled student would take benefits from a procedural simplification.\n\nCONCLUSIONS\n===========\n\nUnder the conditions of this study, it can be concluded that Reciproc was the most effective instrument in the removal of canal filling materials during retreatments performed by undergraduate students. Mtwo retreatment and Reciproc techniques required less time to perform the retreatment procedures. No instrument fracture was observed in any of the groups.\n\n**Conflict of Interest:** No potential conflict of interest relevant to this article was reported.\n\n**Author Contributions:** **Conceptualization:** Silva EJNL, Belladonna FG.**Data curation:** Silva EJNL, Belladonna FG, Moreira EJL.**Formal analysis:** Silva EJNL, Belladonna FG.**Funding acquisition:** Silva EJNL, Moreira EJL.**Investigation:** Silva EJNL, Belladonna FG, Carapi\u00e1 MF, Muniz BL, Rocha MS.**Methodology:** Silva EJNL, Belladonna FG, Carapi\u00e1 MF, Muniz BL, Rocha MS.**Project administration:** Silva EJNL, Belladonna FG.**Resources:** Silva EJNL, Moreira EJL.**Software:** Silva EJNL, Moreira EJL.**Supervision:** Silva EJNL.**Validation:** Silva EJNL.**Visualization:** Silva EJNL, Belladonna FG.**Writing - original draft:** Silva EJNL, Belladonna FG, Moreira EJL.**Writing - review & editing:** Silva EJNL, Belladonna FG.\n"}
+{"text": "The content published in Cureus is the result of clinical experience and/or research by independent individuals or organizations. Cureus is not responsible for the scientific accuracy or reliability of data or conclusions published herein. All content published within Cureus is intended only for educational, research and reference purposes. Additionally, articles published within Cureus should not be deemed a suitable substitute for the advice of a qualified health care professional. Do not disregard or avoid professional medical advice due to content published within Cureus.\n\nIntroduction\n============\n\nLance-Adams syndrome (LAS) is chronic post-hypoxic myoclonus with onset days to weeks after a\u00a0hypoxic episode, when the patient has regained consciousness \\[[@REF1]\\]. The myoclonus is typically multifocal, worse with action, and associated with sudden lapses in muscle tone in the legs, causing a \\\"bouncy gait\\\" \\[[@REF1]\\]. Additional features may include cognitive impairment, seizures, dysarthria, and ataxia \\[[@REF1]\\]. Given its relative rarity, there are no controlled treatment studies of LAS. The majority of cases require polypharmacy management, with an incomplete response. \\\"Bouncy gait,\\\"\u00a0in particular, is notoriously medication-refractory \\[[@REF2]\\]. Here, we report a patient with long-standing LAS who improved markedly when low-dose perampanel was added to his existing treatment regime.\n\nCase presentation\n=================\n\nA 63-year-old Malaysian man of Chinese ethnicity underwent cardiac bypass surgery in 2013 for triple-vessel disease, which was complicated postoperatively by cardiac arrest resulting in hypoxic coma. He was hospitalized for three months, of which one-and-a-half months was in the intensive care unit. During his recovery, the patient experienced jerks involving the whole body and difficulty walking. There was also significant cognitive impairment. Epileptic seizures occurred during his hospitalization but not after discharge. He was treated with twice-daily clonazepam 0.5 mg, levetiracetam 500 mg, and sodium valproate 200 mg.\n\nHe was referred to our center\u00a0in November 2017 from the rehabilitation physician with a diagnosis of \\\"ataxia\\\" causing gait difficulty as his main problem. On examination, there was an obvious \\\"bouncy gait,\\\"\u00a0requiring close supervision and, frequently, assistance to walk just several feet within the consulting room; gait was very hesitant and wide-based with an obvious tendency to falls. There was positive action myoclonus in the arms when performing the finger-to-nose test, more obvious on the left (mild-to-moderate amplitude)\u00a0but no significant myoclonus at rest\u00a0nor with the arms outstretched (and no asterixis). There was no significant stimulus-sensitivity to touch. Eye movements were normal, with mild-to-moderately slurred speech. There were no other features of cerebellar ataxia, Parkinsonism, or dystonia. The Montreal Cognitive Assessment (MoCA) score was 17/30. Brain MRI was normal. Lance-Adams syndrome (LAS) was diagnosed. Acetazolamide 250 mg bid was added, without significant improvement. We have in our patient medical records three video recordings taken, with the patient\\'s written permission, over approximately two years, showing the relative stability of the patient\\'s condition. These are available upon request from the corresponding author.\n\nDue to ongoing gait difficulty\u00a0and based on emerging reports of perampanel use in myoclonic disorders \\[[@REF3]\\], we added perampanel in August 2019. Improvement occurred within the first week on 2 mg/d, with further improvement on 4 mg/d. A home video sent to us by the patient\\'s family showed that his gait was obviously improved. He was able to walk independently, without a walking aid, at home and in a shopping center, with no \\\"bouncy gait\\\" evident. There were no adverse effects, and the benefit has been sustained at least until the time of writing of this report (October 2019).\n\nDiscussion\n==========\n\nMyoclonus in LAS is thought to be most commonly cortical in origin and may relate to abnormal gamma-aminobutyric acid (GABA) and serotonin neurotransmission in the brain \\[[@REF1]\\]. There are no well-defined guidelines for treatment in LAS. Anti-epileptic drugs (AED) such as valproic acid, clonazepam, and/or piracetam/levetiracetam are commonly used, but treatment response is often unsatisfactory \\[[@REF1]-[@REF2]\\]; thus, there is need for new therapeutic options.\n\nPerampanel is a recently introduced AED that antagonizes post-synaptic alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors. It is effective for seizures and myoclonus in different types of epilepsies, including progressive myoclonic epilepsies \\[[@REF3]-[@REF4]\\], with some patients regaining ambulatory capacity post-treatment \\[[@REF2]\\]. Sedation, dizziness, anxiety, and irritability appear to be the most common adverse effects \\[[@REF3]-[@REF4]\\]. To our knowledge, there has been only one fully published case report \\[[@REF3]\\] describing an improvement of myoclonus in LAS using perampanel (Table [1](#TAB1){ref-type=\"table\"}) \\[[@REF3],[@REF5]-[@REF7]\\]. A recent electrophysiological study reported that improvement in cortical myoclonus with perampanel correlated with improvements in the parameters of giant somatosensory-evoked potentials \\[[@REF7]\\].\n\n###### Case reports of patients with Lance-Adams syndrome treated with perampanel\n\nACET=Acetazolamide; CBZ=Carbamazepine; CLON=Clonazepam; F=Female; LAC=Lacosamide; LEV=Levetiracetam; M=Male; NA=Not Available; PER=Perampanel; PIR=Piracetam; PRIM=Primidone; UK=United Kingdom; VAL=Sodium valproate; ZNS=Zonisamide; 5-HT= 5-hydroxytryptophan\n\n  ------------------------------------------ ------------------------------------------------------------------------------------------------------------ ---------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------\n  Variable                                   Steinhoff et al. Epilepsy Behav Case Rep 2016 (Germany) \\[[@REF3]\\]                                          Lazaro Lopez et al. Eur J Hospital Pharmacy 2017 (Spain) (Abstract only) \\[[@REF5]\\]           Yelden et al. Brain Injury 2019 (UK) (Abstract only) \\[[@REF6]\\]                                                 Oi et al. Clin Neurophysiol 2019 (Japan) \\[[@REF7]\\]                                                                         Present Case Report (Malaysia)\n  Age of patient (years)                     36                                                                                                           35                                                                                             \\#1: 69; \\#2: 37                                                                                                 \\#1: 47; \\#2: 31                                                                                                             63\u00a0\n  Gender                                     M                                                                                                            M                                                                                              \\#1: M, \\#2: F                                                                                                   Both M                                                                                                                       M\n  Antecedent event                           Cardiac arrest due to Brugada syndrome                                                                       3 consecutive cardiac arrests                                                                  \\#1: Severe pneumonia; \\#2: Accidental decannulation of the tracheostomy tube                                    NA                                                                                                                           Cardiac arrest in the postoperative period following cardiac surgery\n  Duration of LAS prior to PER (years)       1                                                                                                            NA                                                                                             NA                                                                                                               NA                                                                                                                           6\u00a0\n  Medication treatment prior to PER (mg/d)   LEV (2000); VAL (1500); CLON (2); PIR (7,600); LAC (100), treatments were \\\"in vain\\\"                        LEV; VAL; Propofol; Sodium thiopental; PIR; ZNS; Clonidine; Sodium oxybate; 5-HT; Gabapentin   LEV; VAL; CLON at \\\"high\\\" doses - patients said to be \\\"resistant\\\"                                             \\#1: LEV; CLON; PRIM; CBZ; PIR \\#2: CLON; PIR                                                                                LEV (1000); VAL (400); CLON (1); ACET (250)\n  Perampanel dose (mg/d)                     2mg/d for 1^st^ 3 days, then 4mg/d                                                                           24\u00a0                                                                                            NA                                                                                                               \\#1: 10; \\#2: 4                                                                                                              2mg /d for 1^st^ week, then 4mg/d\n  Clinical response                          Almost complete cessation of myoclonic jerks at 4mg, but no mention of gait improvement (wheelchair-bound)   \\\"Controlled\\\" the myoclonus                                                                   \\#1: Myoclonus greatly improved with improved function; \\#2: Function improved including ambulation and speech   \\#1: Myoclonus improved from marked to severe; \\#2: No improvement in myoclonus (remained moderate); Both improved in ADLs   \u00a0\n  Drugs able to be reduced or discontinued   PIR and LAC stopped                                                                                          All other medications besides LEV, gabapentin, PER, and risperidone were stopped               \\#1: CLON reduced; \\#2: CLON and VAL stopped                                                                     NA                                                                                                                           Not yet\n  Follow-up period                           \\>4 weeks                                                                                                    NA                                                                                             NA                                                                                                               NA                                                                                                                           6 weeks\n  Adverse effects                            Somnolence, but this improved over time                                                                      \\\"Behavioural disorders\\\", requiring risperidone treatment                                     NA                                                                                                               \\#1: None; \\#2: Dizziness and palpitation                                                                                    No\n  ------------------------------------------ ------------------------------------------------------------------------------------------------------------ ---------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------\n\nConclusions\n===========\n\nIn conclusion, we observed a remarkable improvement of LAS with low-dose perampanel treatment, which was also well-tolerated. This agent warrants further study in larger numbers of patients and over longer durations of follow-up, to confirm long-term efficacy and safety for treating LAS. All cases reported thus far have been treated with multiple medications before the introduction of perampanel, and further research should also determine if perampanel can be effectively employed earlier in the treatment course of LAS.\n\nThe authors have declared that no competing interests exist.\n\nConsent was obtained by all participants in this study\n"}
+{"text": "Introduction {#s1}\n============\n\nIn the sports injury aetiology and prevention research field, the use of the 'complex systems approach' has been recently promoted.[@R1] Inspired by previous work that questioned the routine application of reductionist scientific methodologies and statistical techniques,[@R2] Bittencourt and colleagues[@R1] argued for an alternative causal approach. This approach recognises that sports injury is a 'complex emergent phenomenon', resulting from the interactions among different factors (ie, a web of determinants), which may produce regularities (ie, a risk profile) that are antecedent to the emerging pattern (ie, sports injury). In brief, the theoretical assumptions underpinning the complex systems approach can be traced back to general systems theory,[@R5] which identifies several characteristics of complexity as a general philosophical precept. These include, but are not limited to, adaptation and learning, tight coupling, causal feedback, non-linear relations, sensitivity on initial conditions, threshold effects, stochasticity and historical dependency. Those characteristics have featured across multiple discussions in the sports injury scientific literature[@R1]; however, no study has yet applied a recognised method that has the capability to dynamically simulate and better understand complex systems causal patterns and processes. One computational modelling method that has been suggested as a suitable approach for sports injury research[@R1] is agent-based modelling (ABM).\n\nIn this paper, we develop a first-of-its-kind ABM in the field of sports science and apply it in the context of distance running-related injury (RRI). Distance running is considered a pertinent example to use from a technical point of view given that the main participatory-related exposure (ie, workload) can be readily defined. There are three aspects to the model that underpin its concept, design and operation: (1) its development is based on the complex systems approach,[@R1] and so investigating the application of a novel complex systems method in sports injury research represents an important line of inquiry; (2) the ABM is informed by contemporary sports injury and RRI causal theory[@R10]; and (3) it incorporates the acute:chronic workload ratio (ACWR),[@R13] an approach that calculates relative changes in training load. The primary aim of the ABM is to simulate the dynamic relationship between the absolute weekly running distance and RRI, as well as the relative change to weekly running distance and RRI, through the manipulation of various model parameters (see section titled, 'athlete management tools').\n\nUsing ABM to simulate the relationship between workload and sports injury development is a considerable step forward in terms of bringing complexity science and systems thinking to the sports injury literature.[@R1] Indeed, during the past decade, there have been a series of articles that share a number of progressive methodological features. Initially, Quatman *et al* [@R2] proposed a conceptual-methodological framework encompassing the integration of in vivo, in vitro and in silico techniques to better understand the development of anterior cruciate ligament injury. In particular, the authors[@R2] stated that the greatest advances in sports injury research were likely to come from a new methodological paradigm that enabled scientists and clinicians to think, theorise and locate appropriate applications that consider the nature of the complex relationships among different exposures. Shortly afterwards, Mendiguchia *et al* [@R3] argued for the same paradigm shift, advocating that sports injury research was required to move beyond the process of wanting to isolate risk factors to a conceptual model encompassing 'dynamic simulations' and the possibility to 'modify different parameters'.[@R3] More recently, Hulme and Finch[@R4] and Bittencourt *et al* [@R1] have suggested the use of ABM for the explication and testing of theoretical causal assumptions in relation to injury development, as well as for the simulation of complex sports injury aetiologic mechanism(s). Further systems-based work has since applied a method from the human factors and ergonomics domain[@R17] and developed a more holistic, 'complex systems model' of RRI development and prevention.[@R15] Notwithstanding the evolution of systems thinking applications in sports injury research, most scholarly contributions have been descriptive in nature[@R1] or have involved the development of static frameworks and models.[@R15] As such, to advance this body of work, it is necessary to apply a computational modelling approach that can simulate dynamic behaviours within complex sports systems and/or understand how systems change over time.\n\nWith regard to scientific theory and clinical practice, the use of simulation and computational modelling techniques has the potential to generate new insights about sports injury aetiology, which could be used to support clinical decision making. Healthcare practitioners rely on a wide range of study designs and different forms of evidence in which to prescribe the most efficacious therapeutic or preventive interventions to athletes.[@R18] For that reason, there is a need to investigate how, as a proven complementary method to routine epidemiological inquiry,[@R21] ABM can dynamically simulate known mechanisms of sports injury (ie, the relationship between workload and RRI), so that it is possible to develop more sophisticated and clinically\u00a0relevant complex systems models. In taking the next formal step, this study represents a transition away from theoretical description and static modelling approaches,[@R1] and examines the feasibility of simulations for studying the complex and dynamic nature of sports injury moving forwards. Therefore, the purpose of this study is to introduce computational modelling to sports injury research, using ABM as one example of a viable method for studying complex injury dynamics in future theoretical and practical applications.\n\nMethods {#s2}\n=======\n\nAgent-based modelling {#s2a}\n---------------------\n\nAs a computational method, ABM simulates the actions and interactions of heterogeneous, autonomous 'agents', to assess the effects of their behaviour on the system as a whole.[@R25] Agents in an ABM can constitute any self-contained and goal-directed entity, including but not limited to, molecules, cells, pathogens, people (eg, athletes, runners\u00a0and sports teams), animals, automated vehicles, organisations, and/or entire synthetic populations.[@R27] In the case that the agents are representative of individual people, the model operator can assign demographic and lifestyle-related characteristics such as age, sex, diet, medical history and injury susceptibility, as well as cognitive rules pertaining to memory, personality, behaviour and/or intelligence.[@R29] This means that agents can learn over time based on past experiences, update their internal states, adapt to changing environmental circumstances and demonstrate any other characteristic or behaviour that has been explicitly defined. Based on its 'ground-up' modelling approach, ABM can be used to explain how populations self-organise and/or create patterns of global behaviour that are not predictable or programmed into each agent type a priori. For this reason, ABM is a powerful tool when wanting to explore the mechanism(s) by which collective behaviour among individual agents gives rise to emergent-level phenomena (eg, rates of sports injury).\n\nMany different health-related contexts have applied ABM. A notable example is the Global-Scale Agent-Based Model, which simulated 6.5\u2009billion persons and explored how various behaviours and contact points shaped the transmission rate and distribution of the H1N1 swine influenza virus.[@R27] Other studies have integrated ABM with geographic information systems science to improve comprehension of how the measles disease propagates through an urban environment.[@R30] In the non-communicable health context, ABM has been used for multiple purposes,[@R31] including the evaluation of policy-level and environmental intervention strategies for improving diet and promoting exercise.[@R32] Specifically, Yang and coworkers[@R32] used ABM to examine the impact of certain policies aimed to change population-level attitudes towards walking among individuals from different socioeconomic backgrounds. In the medical and healthcare context, ABM has emulated a real-world lifestyle modification programme for individuals with diabetes and estimated the morbidity and economic outcomes associated with the modification of certain parameters (eg, pharmacological delivery options) over a 30-year period.[@R35] Since initial applications in the mid-1990s, the use of ABM has continued to gain popularity in parallel with the evolution of information technology and computing power.[@R36] For further information pertaining to the origins, purpose and general use of ABM, the reader is referred to other more comprehensive sources.[@R25]\n\nThe distance running agent-based model {#s2b}\n--------------------------------------\n\nThe distance running ABM was constructed using the NetLogo toolkit (V.6.0.1), a cross-platform, open-source, programmable modelling environment for simulating natural and social phenomena (<https://ccl.northwestern.edu/netlogo/>).[@R37] The simulation environment was representative of a track and field overlay with dimensions of 70\u00d730 patches (arbitrarily scaled distance units) (online [supplementary material figure 1](#SP1){ref-type=\"supplementary-material\"}). To guide the reader through the different stages of ABM development, the following four phases are described: (1) defining the personal characteristics of the synthetic agent population (hereby referred to as 'runners', or the 'running population'); (2) initialising the model and establishing baseline procedures; (3) implementing four distinct 'athlete management tools' (ie, these tools drive the dynamics of the model); and (4) establishing the conditions surrounding the execution of the simulation procedure itself.\n\n10.1136/bjsports-2017-098871.supp1\n\n![Proportional increase in RRI risk with an increasing ACWR. Higher ratios resulted in an exponentially increasing risk of RRI.\u00a0ACWR,\u00a0acute:chronic workload ratio; RRI, running-related injury.](bjsports-2017-098871f01){#F1}\n\n### Phase one: running population characteristics {#s2b1}\n\nThe distance running ABM contained 1000 runners. This number of runners was chosen to capture aggregate, population-level dynamics and patterns. Each runner possessed personal characteristics that could affect their physical capacity to tolerate an applied external running workload, defined as the number of kilometres (km) undertaken in any given training week (p/w). These characteristics included body mass index (BMI), biomechanics, footwear, sleep, diet, recovery and genetics and were selected based on a recent framework of RRI aetiology.[@R11] To support a comparison of those factors across runners, the relative 'quality' of each characteristic was standardised and drawn from a random-normal distribution with a mean of zero (eg, an 'average' diet or sleep quality), with an SD of 0.3 (ie, this maintained most runners between a range of\u00a0+1\u2009and \u22121).\n\nAn additional characteristic that each runner possessed was a maximum workload potential (MWP) state. In accordance with contemporary sports injury and RRI causal theory,[@R10] surpassing the MWP state was equivalent to the absolute external running workload exceeding a specific musculoskeletal structure's physical capacity. Safely reaching the MWP state without surpassing it assumes a perfect environment, training and management regimen. For each runner in the ABM, their initial MWP state was set to a random-normal mean of 65.0\u2009km p/w, with an SD of 10.0\u2009km p/w. This produced a population-based MWP distribution that acknowledged not all runners had an equal upper workload limit. It is worth noting that the selection of runners' characteristics and the MWP state values are not necessarily integral to the operation of the simulation or the validity of its outputs. Rather, this model was focused on demonstrating how ABM can be both programmed and used to simulate the relationship between workload and RRI risk---and by extension---overall population-level athletic performance.\n\n### The ACWR {#s2b2}\n\nCentral to the distance running ABM is the ACWR.[@R13] As a means of facilitating sports performance optimisation, the ACWR can be used to guide the prescription of future workloads. There are two components to this metric: (1) the 'acute' phase, which represents the training load undertaken in the most recent week (ie, a 1-week block); and (2) the 'chronic' phase, which signifies the average training load undertaken in the month prior (ie, a 4-week block).[@R13] Calculating the ACWR involves dividing the acute phase (eg, 60.0\u2009km of running), by the chronic average (eg, 50.0\u2009km), giving in this case a ratio of 1.2 (ie, 20.0% workload increase).\n\nThe ACWR is theoretically driven and practically appealing. Well-developed physical qualities and musculoskeletal adaptations produced during chronic training phases may build athletic resilience and protect against injury.[@R13] Gradually increasing workloads, and closely tracking week-to-week changes to training regimens, is more important than the absolute applied workload exposure at any given time.[@R39] Prospective epidemiological investigations have found that when the acute training load is equal to, or less than, the chronic phase (ie, ACWR ratio\u00a0\u22641.0), the risk of non-contact, soft-tissue injury is significantly lower than ratios of\u00a0\u22651.5.[@R41] Although further research is yet to be conducted to strengthen existing evidence for example,[@R42] an ACWR between 0.8 and 1.3 has been coined the training 'sweet spot', whereas a ratio between 1.3 and 1.4, and\u00a0\u22651.5, represents a moderate and high-risk injury zone, respectively.[@R13]\n\nIn response to the growing interest around the ACWR, concerns have been raised about the use of rolling averages to assess workload and sports-related injury risk.[@R49] The two main limitations with the traditional ACWR calculation are: (1) averages fail to account for variation over time such that day-to-day patterns and 'spikes' in the applied workload are smoothed; and (2) rolling averages neglect the decaying nature of stimuli over time.[@R49] As such, a non-linear training model that places increasing weighting on the daily workloads undertaken towards the end of a chronic training phase has been proposed.[@R51] The exponentially\u00a0weighted ACWR (EW-ACWR) was found to be significantly more sensitive than the traditional ACWR at identifying injury likelihoods at upper training load ratio ranges (ie,\u00a0\u22651.5).[@R50] Along with the traditional ACWR, the EW-ACWR was incorporated into the distance running ABM as an option for calculating the relative variation in runners' workload.\n\n### Phase two: model initialisation {#s2b3}\n\nAt the start of the simulation, each runner was assigned a standard running history spanning the previous 28-day period. This history allocated a total of 20.0\u2009km p/w in each 7-day block preceding each day in the prior 28-day period. Therefore, at ABM initialisation, each runner had 20.0\u2009km p/w history of running in the previous training week and had a rolling average of 20.0\u2009km p/w for the past month. This produced both an initial ACWR and an EW-ACWR of 1.0 (ie, each runner had a consistent workload in relation to a previously recorded workload over the past 28 days as calculated under each regimen).\n\n### Phase three: athlete management tools {#s2b4}\n\nThe distance running ABM incorporated four athlete management tools that were manipulable by the model operator. The first of these tools is the 'ramp-up rate slider', which dictated the rate at which runners applied and increased workload over time. Specifically, the goal of each runner in the system was to maximise the absolute distance they were able to run per week without sustaining RRI, that is, safely reach the MWP state and remain there. To achieve this, the simulation started with runners' gradually increasing their weekly kilometres at a rate determined by the user-defined ramp-up rate. The lower the ramp-up rate, the longer the time frame before the running population reached a MWP state. Conversely, higher ramp-up rates resulted in runners rapidly ascending to their upper workload limits. For the purposes of experimentation, runners' workloads were increased within the reported ACWR sweet spot of between 5.0% and 30.0%, in increments of 5.0%.[@R13] This resulted in a total of six possible ramp-up rate conditions.\n\nThe second athlete management tool was an ability to approximate a runner's individual adherence, misrepresentation or miscalculation of the advice provided by, for example, a coach or healthcare professional as to how much training should be undertaken per week. The 'random variation slider' introduced noise into each runner's planned workload, adjusting the ramp-up rate by a mean of 0.0%, but with increasing SD of 0.0% (ie, perfect training adherence), 1.0% (ie, moderate training adherence) or 2.5% (ie, poor training adherence). The formula for the calculation of workload in the current week is shown in equation 1, where *cw*=workload in the current week, *pw*=workload in the previous week, *r*=ramp-up rate and *error*=random\u2009variation.\n\n$$cw = pw \\ast \\left( {1 + \\left( {r + error} \\right)} \\right)$$\n\nThe third athlete management tool that was manipulated by the model operator was the ability to adjust the way in which the ACWR was calculated. Although differences between the traditional ACWR and the EW-ACWR metrics correspond to approaches that calculate changes in workload, the distance running ABM was constructed so that altering the estimation of the ratio (ie, either non-weighted or weighted) could affect RRI risk. Therefore, the risk of RRI in each week was based on either the ACWR or EW-ACWR calculation, of which both dynamically responded to the user-defined ramp-up rate and random variation condition. Accordingly, if a given runner's calculated workload ratio was\u00a0\u22651.1, then the likelihood of RRI was proportional to the cubed value of their allocated ACWR or EW-ACWR condition. This produced an exponentially increasing risk of RRI that approximated the observed likelihood of sports injury development as found in empirical studies ([figure 1](#F1){ref-type=\"fig\"}).[@R13] To provide a visual indication to the model operator of the health of the running population at any given time, runners who incurred an RRI doubled in size, turned red in colour and were transferred to the centre of the simulation environment. On sustaining RRI, a given runner's workload dropped to 5.0\u2009km per week.\n\nThe fourth and final athlete management tool under manipulation was a binary condition relating to the runner's individual MWP state. In the latter condition, whereby an individuals' MWP was unknown (ie, 'off'), the calculation of runners' future training was based on the workload in the most current week, multiplied by the ramp-up rate and adjusted for random variation (equation 1). A second condition was constructed whereby runners' workloads were further adjusted based on how close the current workload was to their MWP state. This calculation recognised that the running population had a randomly distributed MWP that was guaranteed to be reached under conditions of continuous, compounding growth (ie, MWP state 'on'). The formula for this calculation is shown in equation 2.\n\n$$cw = pw \\ast \\left( {1 + \\left( {\\left( {r \\ast \\left( {1 - \\left( \\frac{pw}{mwp} \\right)} \\right)} \\right) + error} \\right)} \\right)$$\n\n### Phase four: establishing the conditions of the simulation {#s2b5}\n\nThe set combination of the six ramp-up rates, three random variation conditions, two ACWR calculations and two MWP states produced a 72-condition matrix. Given stochastic elements within features of the ABM, modelling under the 72 different possible conditions was repeated 10 times for 1000 model time steps, or days (\\~143 weeks). This produced a total computational model encompassing 720\u2009000 individual simulated runners monitored over a total of 720\u2009000 days (\\~102, 800 weeks). On completion of the simulation, data were exported from the NetLogo[@R37] software into spreadsheet processing software (Microsoft Excel for Windows).\n\nResults {#s3}\n=======\n\nThere were no differences between the ACWR and the EW-ACWR calculations in relation to changes to workloads or RRI incidence proportions across the six ramp-up rates and the three random variation conditions under both MWP states (online [supplementary material table 1](#SP2){ref-type=\"supplementary-material\"}). The EW-ACWR was, however, more sensitive than the traditional ACWR at responding to individual-level workload fluctuations ([figure 2](#F2){ref-type=\"fig\"}). The differences between the 0.0% and 1.0%, and 1.0% and 2.5% random variation conditions did not considerably affect workloads or RRI incidence proportions (online [supplementary material table 2](#SP1){ref-type=\"supplementary-material\"}). As such, we examined the interaction between the six ramp-up rates and the two most extreme random variation conditions, that is, 0.0% (perfect training adherence) and 2.5% (poor training adherence) under both MWP states.\n\n10.1136/bjsports-2017-098871.supp2\n\n![Dynamic plot visualising the relationship between workload and RRI over 80\u2009weeks for a single runner. Ramp-up rate set to 30.0%, random variation was\u00a0set to 0.0%\u00a0and MWP was\u00a0set to off. The black, grey, orange and red lines represent the EW-ACWR, the traditional ACWR, workload and RRI incidence proportion, respectively. Unlike the ACWR that exhibited smoothing behaviour due to the calculation of rolling averages, the EW-ACWR closely tracked workloads and rapidly dropped off when RRI was sustained. This demonstrates the flexibility of computer coding and ABM in being able to emulate sophisticated training load calculation approaches. Note that at three separate time points, the absolute workload fell short of the runner's MWP state, indicating that RRI was sustained by chance before the runner could reach their absolute upper workload limit. The use of a single runner and 80\u2009weeks (570 elapsed model time steps) is for illustrative purposes.\u00a0ABM,\u00a0agent-based modelling; ACWR, acute:chronic workload ratio; EW-ACWR, exponentially weighted ACWR; MWP, maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f02){#F2}\n\nPerfect training adherence (random variation 0.0%) {#s3a}\n--------------------------------------------------\n\nWhen the random variation was set to 0.0%, and the MWP state was set to on, the running population maintained the highest workloads relative to when the MWP was set to off ([figure 3](#F3){ref-type=\"fig\"}). Similarly, higher ramp-up rates over the simulated time\u00a0frame also resulted in higher maximum workloads. Specifically, at a 5.0% and 30.0% ramp-up rate, the distance performed by the runners was 53.8\u2009km p/w and 62.5\u2009km p/w, respectively. Under the same set of conditions, the RRI incidence proportion was 0.0%. Conversely, with the MWP state set to off (ie, runners could overshoot their MWP state), the RRI incidence proportion climbed from 4.2% to 30.1% across the six ramp-up rates.\n\n![A comparison of workloads and RRI incidence proportions under the two MWP states across six ramp-up rate conditions. Random variation was set to 0.0% over 143\u2009weeks.\u00a0MWP,\u00a0maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f03){#F3}\n\nThe variability around workloads and RRI incidence proportions under the two different MWP states can be viewed in the dynamic ABM output plots ([figures 4 and 5](#F4 F5){ref-type=\"fig\"}). With the MWP set to on, the running population consistently increased their workload, and aware of the threshold over which they would sustain RRI, safely reached a performance ceiling ([figure 4](#F4){ref-type=\"fig\"}). With the MWP state set to off, a given runner invariably surpassed their physical capacity and sustained an RRI ([figure 5](#F5){ref-type=\"fig\"}).\n\n![Dynamic plot visualising the relationship between workload and RRI over 143\u2009weeks for 1000 runners. Ramp-up rate set to 10.0%, random variation set to 0.0%, and MWP set to on. The black, grey, orange and red lines represent the EW-ACWR, the traditional ACWR, workload and RRI incidence proportion, respectively. As runners neared a MWP state, the upper bound of the ACWR inwardly collapsed to 1.0.\u00a0ABM,\u00a0agent-based modelling; ACWR, acute:chronic workload ratio; EW-ACWR, exponentially weighted ACWR; MWP, maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f04){#F4}\n\n![Dynamic plot visualising the relationship between workload and RRI over 143\u2009weeks for 1000 runners. Ramp-up rate set to 10.0%, random variation set to 0.0% and MWP set to off. The black, grey, orange and red lines represent the EW-ACWR, the traditional ACWR, workload and RRI incidence proportion, respectively. The initial oscillatory pattern stabilised as runners sustained RRI at different times due to a randomly distributed MWP state.\u00a0ABM,\u00a0agent-based modelling; ACWR, acute:chronic workload ratio; EW-ACWR, exponentially weighted ACWR; MWP, maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f05){#F5}\n\nThe workload across the six ramp-up rates remained relatively stable with the MWP state set to off ([figure 3](#F3){ref-type=\"fig\"}). Accordingly, a 5.0% ramp-up rate resulted in runners spending a proportionately greater amount of time performing lower weekly running distances to the benefit of fewer RRIs ([figure 6](#F6){ref-type=\"fig\"}). However, a 30.0% ramp-up rate reduced the length of time that runners spent at lower workloads but equally resulted in a higher RRI incidence proportion. Across the population, MWP spikes stabilised with relatively longer periods of workload growth.\n\n![Dynamic plot visualising the relationship between workload and RRI over 428\u2009weeks for a single runner. Random variation set to 0.0%, and MWP state set to off. The black, grey, orange and red lines represent the EW-ACWR, the traditional ACWR, workload and RRI incidence proportion, respectively. The output plot visualises the changes to workload and the RRI incidence proportion when the ramp-up rate is switched, on the fly, between 30.0% and 5.0% every\u00a0\\~71\u2009weeks. Note the greater workload oscillation at a 30.0% relative to a 5.0% ramp-up rate. The use of a single runner and 428\u2009weeks (3000 elapsed model time steps) is for illustrative purposes.\u00a0ABM,\u00a0agent-based modelling; ACWR, acute:chronic workload ratio; EW-ACWR, exponentially weighted ACWR; MWP, maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f06){#F6}\n\nPoor training adherence (random variation 2.5%) {#s3b}\n-----------------------------------------------\n\nIntroducing training error into the runners' ramp-up rates resulted in changes to both workloads and RRI incidence proportions ([figure 7](#F7){ref-type=\"fig\"}). This condition simulated a scenario whereby runners were aware that a MWP state existed, but they could only estimate the value within a 2.5% random variation. At a 5.0% and 30.0% ramp-up rate, the mean distance performed by the runners was 29.9\u2009km p/w and 35.3\u2009km p/w, respectively. The RRI incidence proportion was higher across the six ramp-up rates relative to the 0.0% random variation condition.\n\n![A comparison of workloads and RRI incidence proportions under the two MWP states across six ramp-up rate conditions. Random variation was set to 2.5% over 143\u2009weeks.\u00a0MWP,\u00a0maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f07){#F7}\n\nWith the random variation set to 2.5%, there was an initial upwards workload trajectory as runners climbed towards their MWP state ([figure 8](#F8){ref-type=\"fig\"}). As a given runner approached and misjudged their MWP state due to training error, a higher RRI incidence proportion across the population brought the workload down over the 143 weeks.\n\n![The relationship between workload and RRI over 143\u2009weeks for 1000 runners. Ramp-up rate set to 10.0%, random variation set to 2.5% and MWP set to on. The black, grey, orange and red lines represent the EW-ACWR, the traditional ACWR, workload and RRI incidence proportion, respectively.\u00a0ABM,\u00a0agent-based modelling; ACWR, acute:chronic workload ratio; EW-ACWR, exponentially weighted ACWR; MWP, maximum workload potential; RRI, running-related injury.](bjsports-2017-098871f08){#F8}\n\nDiscussion {#s4}\n==========\n\nThe purpose of this study was to introduce computational modelling to sports injury research, using ABM as one example of a viable method for studying complex injury dynamics in future theoretical and practical applications. To achieve that purpose, an ABM was developed with the aim of simulating the dynamic relationship between the absolute weekly running distance and RRI, as well as the relative change to weekly running distance and RRI, through the manipulation of four athlete management tools (ie, six ramp-up rates, three random variation conditions, two ACWR calculations\u00a0and two MWP states). This was an important step for the integration of complexity science and systems thinking in the sports injury literature,[@R1] particularly given that no study has yet formally demonstrated the use of computational modelling in this context. Previous attempts to describe and/or apply the complex systems approach have resulted in the development of static frameworks or models that are not capable of simulating dynamic behaviours within complex sports systems and/or understanding how systems change over time.[@R15] In addition to the findings related to the optimal management of RRI within running populations, this paper also reiterates the long-standing need for an alternative paradigm involving 'dynamic simulations' and 'complex modelling' as has previously been called for.[@R2] Although the distance running ABM has effectively simulated the occurrence of sports injury in a population of runners, there remains a need to highlight what computational modelling can offer to the field of sports injury research more broadly. Therefore, the following discussion is structured around the main take-home messages and subsequently outlines important considerations when aiming to use simulations in future research-based applications. Our intention is to clarify the potential contribution of ABM\u00a0and to inspire researchers and clinicians to continue to explore computational modelling and further develop applications in the sports injury context.\n\nWith regard to the presented simulation dynamics, a useful way of conceptualising the distance running ABM is to consider the rate at which the running population climbed towards a MWP state. Depending on the predefined ramp-up rate plus error condition, the goal of the agents was to run safely toward their maximum performance level. With the MWP state set to on, and the random variation condition set to 0.0% (ie, perfect adherence to instruction), the running population appropriately identified their workload limits, and sustained the lowest number of RRIs. Conversely, increasing the random variation to 2.5% (ie, poor adherence to instruction) while leaving all other conditions equal adversely affected workloads and RRI incidence proportions over the course of the simulation. The maximum workload across the six ramp-up rates was comparatively lower when training error was higher because runners were misjudging the applied workload, and therefore sustaining RRI despite being aware of their MWP state.\n\nSetting the MWP state to on and increasing the level of random variation in the model is representative of a real-world sports training situation. Distance runners, coaches and qualified healthcare professionals may be aware that an MWP state for each athlete exists, but knowing precisely where that upper individual limit is, and how to get there safely, is arguably one of the greatest challenges in prescribing future training loads. Despite the utility of workload calculation approaches such as the ACWR,[@R41] the distance running ABM has demonstrated that building weekly running distances over time, even within reported sweet spot guidelines of up to 1.3,[@R13] will eventually result in RRI as athletes reach their upper physical limits. This was indicated in the simulation, as the programming code was set so that varying workloads would occasionally surpass the MWP state for some individuals in the absence of extreme relative changes to the applied running distance. Exceeding a physical capacity to tolerate workload is, however, not a new concept, and a 'ceiling effect of safety' has been described in the literature.[@R52] The findings of present study support the view that the calculation and prescription of athletic workloads should not be performed in isolation (ie, with a single metric) and requires a comprehensive, individualised and flexible approach.[@R12]\n\nAnother insight offered by the simulation relates to the trade-off between a lower versus higher weekly ramp-up rate under the 2.5% random variation condition at different stages of recovery following RRI. Results showed that, in general, a higher ramp-up rate after returning from RRI resulted in greater maximum running distances across the population over the course of the simulation as runners quickly returned to their preinjury workload levels. For those athletes in the process of returning from RRI, this may be seen as positive. However, this relationship was offset by a higher RRI incidence proportion. Specifically, in the early stages of recovery and prior to reaching their MWP state, runners experienced a greater margin of error when either miscalculating workload or not observing the recommended ramp-up rate. However, when runners' physical capacity to tolerate workload had been reached, any error to the applied running distance, irrespective of its magnitude, resulted in RRI. This 'hard ceiling' dynamic shows that it may be advantageous for runners who wish to maintain high distances over extended periods of time to think long\u00a0term about their training, or perhaps even refrain slightly from regularly operating at their perceived level of peak performance. While this simulation was modelled on understanding RRI occurrence in a 'general population', the implications of this dynamic indicated fragility at the extremities of performance for the more serious runner who might aim to participate in competitive events. Athletes, running coaches and healthcare practitioners are reminded that although it is necessary to progressively and systematically increase external workloads over time, it is as equally important to continuously monitor and measure internal physiological and psychological responses to that load.[@R40]\n\nThe conceptual basis and development of the distance running ABM was based on the complex systems approach.[@R1] For this reason, it is worthwhile to briefly outline the main distinction between computational modelling methods and traditional statistical models for studying health-related phenomena. First and foremost, simulation techniques (like the ABM here) are best used for exploring and understanding mechanisms and theories in complex systems that are potentially unknown or contested.[@R55] Conversely, statistical models such as regression analyses, are useful for testing a priori hypotheses and analysing already collected data. Simulation methods are, therefore, better suited at generating hypotheses that can be empirically tested with statistical modelling. One theoretical advantage of ABM over other computational modelling methods (eg, systems dynamics modelling, machine learning) is that it allows the programmer to more easily study systems in which the causal mechanisms for one person can change depending on their connection to other individuals.[@R56] A typical example would be in the infectious disease context whereby the probability of contracting a given disease is dependent on whether there are only isolated cases or an epidemic. Although our simulation did not model agent interactions, there is a need for future computational studies to explore this social dimension in relation to sports injury development. As such, the application of simulation methods to a practical sports injury problem could be highly beneficial if used to supplement routine epidemiological inquiry.\n\nWhen studying complex systems phenomena with ABM, it is expected that a greater reliance is placed on theory relative to data.[@R57] That is, ABM cannot offer the same level of external quantitative credibility that traditional statistical modelling can provide.[@R58] However, ABM enables the analyst to establish a balance between realism (ie, face validity), generality (ie, qualitative abstraction) and numerical precision (ie, fineness of model specification). This balance can be achieved by triangulating different forms of evidence and using empirical data to parameterise models when assigning agent characteristics and environmental rules at baseline.[@R59] Notwithstanding the reported guidelines around the development, calibration and validation of computational modelling,[@R55] ABM is effectively an in silico laboratory that can provide scientists and clinicians with a means of understanding what results they might expect if current theories are true, although without undue financial, ethical or logistical implications that are associated with real-world pilot studies.[@R56] For instance, it is possible to situate agents within a social network and broader spatial context, duplicate baseline conditions and subsequently change only one aspect of the model. Thus, a range of experimental scenarios can be trialled repeatedly, providing scientists with an agent-based counterfactual simulation that predicts the impact of different health-related strategies or policies based on the simulation settings provided.[@R56] Consequently, ABM has been used to emulate the randomised controlled trial for patients with diabetic retinopathy, allowing for the examination of hypothetical interventions targeting vision loss.[@R62] The next step might be to follow-up such an ABM with a 'real' experiment by drawing on the theoretical insights generated from the simulation exercise.\n\nGiven the flexibility of computational modelling, it is not surprising to find that ABM has been used to simulate cyclical, self-reinforcing feedback loops among individual microunits, including people, cells and molecules, to identify emergent patterns of behaviour, such as disease transmission dynamics,[@R64] wound healing processes[@R65] and adaptive immunity.[@R67] Indeed, one advantage of ABM is that the complex interactions between agents can (and ideally should) be explicitly modelled. If the theoretical causal mechanisms that are encoded into the simulation programme are accurate, it will provide useful predictions. These causal mechanisms are developed through a synthesis of all evidence, preferably through a causal inference approach that makes assumptions explicit. One challenge when hypothesising these causal mechanisms from observational data is that, traditionally, they require the stable unit value treatment assumption. In brief, this assumption states that the outcome of one individual should not affect the outcome of another individual, irrespective of whether they were exposed or not.[@R36] As explained previously in relation to infectious diseases, this assumption is often violated when there is interference among units leading to biased causal effects[@R68] (eg, exposed athletes might influence the behaviours and potential outcomes of their unexposed peers). More recent work in the causal inference scientific thematic does not require this assumption; however, the suggested methods are not yet widely implemented and so the hypothesised causal mechanisms encoded into simulation models may not be accurate despite best intentions.[@R69]\n\nA final consideration when using regression models (eg, logistic\u00a0and survival analyses), and something to which computational modelling can potentially circumnavigate, is the events-per-variable (EPV)\u00a0requirement.[@R73] In short, the EPV requirement explains that in order to conduct a robust statistical sports injury analysis and avoid 'sparse data bias',[@R74] the number of explanatory variables modelled should be large enough in relation to the number of events (ie, injuries) observed. Not only does the EPV requirement necessitate considerably large sample sizes, but it is also further compounded when the researcher wishes to stratify samples to understand how the applied workload and other time-varying exposures change status over time during follow-up.[@R75] Accordingly, in the absence of large-scale data sets, computational modelling methods such as ABM might prove useful for simulating hundreds or thousands of athletes, each of whom can be assigned characteristics and decision-making rules corresponding to real-world values and observed behaviours. Doing so might lead to a sufficient number of injurious events per explanatory variable modelled, affording insight into the likely mechanisms that generate certain outcomes. In light of the above considerations, computational modelling methods are by no means superior to traditional scientific approaches and statistical models; rather, simulation techniques could be considered (and might prove to be) a useful methodological adjunct. Although beyond the scope of this paper, a more complete list of the advantages and disadvantages of computational modelling, including ABM, should be provided and contextualised within the sports science field.\n\nOn the whole, and in terms of sports injury research, computational modelling has the potential to predict the extent to which different factors and their interactions influence the onset of injury given the settings of the model. Evaluating new hypothetical injury prevention strategies is also possible, provided that these strategies do not change or require additional underlying causal assumptions that are not coded within the model. Irrespective of why simulations and computational modelling techniques are to be used, their development should be carefully planned, debated and scrutinised over a series of iterative stages that starts with a verified working model, similar to the distance running ABM presented. There is now a need for future computational modelling applications to explore how ABM can be used to simulate more advanced complex systems characteristics in relation to sports injury aetiology and prevention.\n\nLimitations and research-based considerations {#s4a}\n---------------------------------------------\n\nThis proof-of-concept study is not without limitation. First and foremost, the intention of the ABM was not to offer original data, nor provide new or practical knowledge about how to safely increase workloads for running performance optimisation. For example, runners' personal characteristics such as BMI were not explanatory in the sense of impacting on the results, and this should motivate future computational applications to either build on the distance running ABM directly (annotated NetLogo Code provided) or draw on its premise to guide the development of dynamic simulations in other sports domains. For that reason, the model was not instantiated with data, and the selection of runners' personal characteristics, as well as numerical values pertaining to workload and RRI risk, were based on subject matter expertise (AH and RON), contemporary RRI causal theory[@R10] and evidence around the ACWR.[@R13] We contend that this provides a practical position in which to further explore computational modelling.\n\nAnother limitation relates to the different classes of agent-based simulations that can be developed. For example, the distance running ABM is more representative of a 'multi-agent' system, as runners responded to their environment (ie, direction from the coach) but operated independently from one another. These conditions are easily modelled with traditional simulation techniques as well. Conversely, a typical ABM in the social sciences aims to understand how the mechanism of interaction between boundedly rational agents leads to the emergence of global patterns and collective behaviour. Extending the current model to include local level interactions between agents and factors could be a means of providing further insight into the role of specific social mechanisms that drive behaviour and injury incidence, as well as recovery among the running (or other sports) populations. Although there are many different ways to advance the distance running ABM, it is essential that extensions are biologically plausible, theoretically reasonable and numerically precise where relevant.[@R58]\n\nUnder the assumption that computational modelling is to gain further traction in the field of sports science, there will be an ongoing requirement for scholarly research to ask important questions such as: when is the use of computational modelling appropriate and for what types of problems should it be applied? What should be included in a given simulation, including its scope, properties, agents, behaviours, environment, inputs and modelled time steps? How can the modelled outputs best be verified, and what type of validation is to be sought (eg, empirical output or macro-face validation)? Finally, what are the technical options for addressing model misspecification? In terms of ABM, detailed resources such as the one provided by Rand and Rust[@R55] represent a useful starting point for sports injury scientists and clinicians who wish to continue to explore what computational modelling can offer as an alternative method.\n\nIn terms of research-based considerations, ABM requires the expert use of an object-oriented programming language (eg, Java, Python\u00a0and C++), and the wide range of available computer-based software packages could be disconcerting. Overcoming these hurdles and transitioning to computational modelling requires a multidisciplinary team comprising clinicians and healthcare professionals, sports injury experts, social scientists, epidemiologists, biostatisticians and, crucially\u00a0individuals with proven level of expertise in coding and programming.\n\nConclusions {#s5}\n===========\n\nIt has long been argued that the adoption of a complex systems approach in sports injury research and practice will help to better understand and prevent injury. However, beyond theoretical description and static models of complexity, little progress has been made towards formalising this approach in a way that is practical to sports injury scientists and clinicians. In short, a true complex systems approach has not yet been applied. However, in demonstrating for the first time the use of ABM to simulate RRI, this study illustrates that computational modelling can be helpful in sports injury research. We used current theory to guide the choices for causal relationships in the simulations of RRI, including variation in the population MWP state, the ACWR calculation and individual differences in adherence to training (or lack thereof). Our results are therefore necessarily consistent with current theory. Specifically, when runners were aware of their upper workload limits and accurately adhered to training-related instruction, the agent population maintained the highest level of performance while sustaining the lowest number of RRIs. Poor adherence to training-related instruction adversely affected workloads and RRI incidence proportions over the course of the simulation. The ABM confirmed that the calculation and prescription of athletic workloads should not rely on the use of single metric and require a comprehensive, personalised and adaptable approach. This is especially true for runners who are operating close to their physiological potential.\n\nAdditional implications extend beyond the presented model. In particular, scientists and clinicians interested in the philosophy of complex systems should start to explore what computational modelling can offer to a sports injury problem of interest. Likewise, descriptions of the specific advantages and disadvantages of the different computational modelling strategies (including ABM) in different contexts (eg, sports injury prediction, prevention\u00a0and treatment) would be helpful. Alongside the continuing use of traditional epidemiological and clinical research-based applications, the use of computational modelling methods should be considered as a complementary methodological approach in sports injury research.\n\n###### What are the findings?\n\n-   Agent-based modelling (ABM) is best used for exploring and understanding mechanisms and theories in complex systems that are potentially unknown or contested. Conversely, formal statistical analyses are useful for testing a priori hypotheses and analysing already collected data.\n\n-   This proof-of-concept study shows that the distance running workload and running-related injury (RRI) dynamics can be simulated using an ABM\u00a0approach.\n\n-   Based on our simulation settings, attempting to maintain workloads within the ACWR sweet spot will still result in athletes surpassing their own 'ceiling of safety' resulting in RRI. This dynamic is exacerbated with greater workload error.\n\n###### How might it impact on clinical practice in the future?\n\n-   Computational modelling methods such as ABM are primarily used to understand how local-level behaviours and interactions among individual 'agents' (eg, molecules, cells\u00a0and athletes) may theoretically lead to the emergence of complex systems patterns (eg, sports injury development).\n\n-   Sports injury scientists and clinicians might want to familiarise themselves with ABM to determine new ways of using this theoretically\u00a0driven method so that it can be effectively applied to a specific problem of interest.\n\n-   In the absence of large-scale data, scientists and clinicians interested in the aetiology and prevention of sports injurry are invited to consider computational modelling as an alternative and complementary method to traditional epidemiological and clinical research-based applications.\n\nWe would like to thank Professor Rod McClure for his insightful and constructive feedback during the early stages of this research. We also thank Mr Tony Carden whose knowledge on the origins and history of complexity theory was most valuable. Dr Simon Feros, Dr Scott Mclean and Ms Natalie Selever provided useful comments pertaining to the flow and readability of the manuscript. We are grateful to Mr Nicholas Patorniti, who helped with technical formatting.\n\n**Contributors:** AH was responsible for the concept, ABM development, methods, results interpretation and write-up. JT was primarily responsible for developing the ABM, contributed to the methods write-up and had editorial input into the manuscript. RON, GJMR and PMS had editorial input into the manuscript and contributed to the write-up. PMS's contribution to this work was funded by the Australian Research Council (FT140100681).\n\n**Funding:** Australian Research Council (grant number: FT140100681).\n\n**Competing interests:** None declared.\n\n**Patient consent:** Not required.\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n"}
+{"text": "Introduction\n============\n\nEarly pregnancy loss is the loss of a pregnancy before 20 gestational weeks. Although the exact mechanisms responsible for abortion are not always clear in the first trimester, fetal factors including aneuploid are the predominant etiology and account for 80% to 90% of early miscarriage. Among fetal factor in terms of the first trimester, approximately half of miscarriages are embryonic causes and half of those is due to fetal chromosomal aberrations.\n\nFetal aneuploidy seems to occur as a result of increased chromosomal nondisjunction during the long arrest period in meiosis I, before ovulation \\[[@B1][@B2][@B3][@B4]\\]. There is well-known association between fetal aneuploidy and the maternal chronologic aging. Maternal chronologic aging is correlated to decreased ovarian reserve which is predisposed to a greater risk for fetal aneuploidy \\[[@B5][@B6]\\]. Based on such an association between maternal ovarian reserve and fetal aneuploidy, multiple predictive markers in terms of ovarian reserve including follicle stimulating hormone (FSH) and antimullerian hormone (AMH) level have been studied to predict fetal aneuploidy \\[[@B7][@B8]\\].\n\nAMH, synthesized by human granulosa cells in the ovary, has exhibited significant promise as a potential marker of ovarian reserve \\[[@B2][@B9][@B10][@B11]\\]. Serum AMH levels decline with age. Several studies have shown that AMH is a better predictor of ovarian reserve than age \\[[@B12][@B13]\\]. It offers several advantages over other traditional markers to predict ovarian reserve. Oocyte quality as well as oocyte quantity affect embryo quality and seem to be associated with aging. AMH is postulated to be associated with both ovarian reserve and function. Another advantage of AMH is its stability. AMH level is constant throughout the menstrual cycle and is not influenced by gonadotropin-releasing hormone agonists, pregnancy, or oral contraceptives \\[[@B11][@B14][@B15][@B16][@B17]\\].\n\nDespite of such an advantages and expected close association with ovarian reserves of AMH, limited number of the study regarding the association between ovarian reserve and fetal aneuploidy in early spontaneous miscarriage was reported.\n\nTherefore, the objective of the present study was to examine the direct relationship between parameters that represent ovarian reserve including AMH and the incidence of fetal aneuploidy in early spontaneous miscarriage.\n\nMaterials and methods\n=====================\n\nA multicenter retrospective study was performed with patients who were diagnosed with the first trimester pregnancy loss between January 2011 and December 2012 in CHA Gangnam Medical Center, CHA Bundang Medical Center, and CHA Gumi Medical Center. This study was approved by the institutional review board of CHA Gangnam Medical Center, CHA University, Korea. Only singletone pregnancies were included in the study. A dilatation and curettage was offered to all patients after the diagnosis of early pregnancy loss before 13 gestational weeks. Pregnancy loss was confirmed by repeat ultrasound examination. Conceptus which had been spontaneously expelled also processed karyotype testing.\n\nKaryotyping was conducted by the Genetic Laboratory of the Fertility Center of CHA Gangnam Medical Center. Karyotype analysis was performed by using classical karyotype cytogenetics (standard tissue culture) or multiplex ligation-dependent probe amplification (MLPA) which remained as consistent techniques over the 10-year period.\n\nPatient characteristics, such as maternal age, gestational age, parity, reproductive history, method of conception, maternal body mass index (BMI), AMH level and FSH levels were reviewed. As conceptional method, natural pregnancy was defined when conceived naturally or with method of intrauterine insemination (IUI) and timed intercourse without controlled ovarian hyperstimulation (COH). In case of COH, the other categories were divided as *in vitro* fertilization (IFV) embryo transfer (ET), intracytoplasmic sperm injection (ICSI), IVF-Freezing ET and IUI according to the procedures.\n\nEach patient\\'s serum AMH and FSH level, serving as a marker of ovarian reserve, was measured at especially infertility clinic. The blood sample for hormone assay were obtained on menstrual cycle day 3 to 5. Serum AMH was quantified with AMH Gen ll ELISA (Beckman-Coulter, Inc., Brea, CA, USA). This assay is standard monoclonal antibody sandwich enzyme immunoassay which is specific for AMH and do not exhibit any significant cross-reactivity with related molecules. All AMH level of the patients were analyzed by this sole assay, which minimize the error caused by using diverse assay. If sample assay was not be completed within 48 hours, the samples were freezed at -20\u2103.\n\nStatistical analyses were performed using the IBM SPSS ver. 20 (IBM Corp., Armonk, NY, USA). Chi-square test and Student\\'s t-test were used for the calculation of significance. The descriptive data were expressed as the mean\u00b1standard deviation. Univariate and logistic regression analyses were performed to evaluate the association between parameters of ovarian reserve and fetal aneuploidy. For multivariate analysis, advanced maternal age was defined with patient age (\u226535 or \\<35 years) because increased rates of fetal aneuploidy have been documented in this age group. A *P*-value of \\<0.05 was considered statistically significant.\n\nResults\n=======\n\nTotal 462 patients were enrolled for the study. The mean age of the women was 35.31\u00b14.12 years and the mean BMI was 21.05 kg/m^2^. A total of 195 patients were tested for AMH level, mean level of which was 3.88\u00b13.50 ng/mL. A total of 264 patients were tested for FSH. The mean level of FSH was 8.53\u00b14.96 mIU/mL. Mean gestational age when the patient was diagnosed with early pregnancy loss was 8.0\u00b11.52 weeks. The number of subjects diagnosed by classical cytogenetic karyotype was 413 (89.4%). Forty-nine (10.6%) subjects were diagnosed by MLPA. Among 462 patients, 211 conceptuses (45.7%) of patients showed the euploid and 251 conceptuses (54.3%) showed the aneuploid. The number of patients who were pregnant naturally was 128 (27.7%) and number of patients who were conceived by timed intercourse was 61 (13.2%). One hundred twenty one patients underwent IVF-ET (26.2%) and 73 received ICSI (15.8%). The most common abnormal karyotype encountered was autosomal trisomy followed by monosomy and double trisomy ([Table 1](#T1){ref-type=\"table\"}).\n\n[Table 2](#T2){ref-type=\"table\"} compares patient demographics including AMH, FSH, and gestational age between euploid and aneuploid group. The euploid group was younger than aneuploid group (34.46\u00b14.36 vs. 36.04\u00b13.78, *P*\\<0.001, respectively). There was no difference in gravidity, parity, BMI and previous abortion history. AMH level is significantly low in fetal aneuploid group than fetal euploid group (3.43\u00b13.18 vs. 4.60\u00b13.86, *P*=0.022). However, there was no significant difference in FSH level between euploid and aneuploid group (8.09\u00b15.09 vs. 8.86\u00b14.85, *P*=0.214). Early pregnancy loss was diagnosed earlier in aneuploid group than euploid group (7.67\u00b11.54 vs. 8.27\u00b11.46, *P*\\<0.001). By classic karyotype cytogenetics methods, 45.7% and 54.3% karyotype of patients was confirmed as euploid and aneuploid, respectively. By MLPA, 44.9% and 55.1% karyotype of patients was confirmed as euploid and aneuploid, respectively. There was no difference according to mode of karyotyping. In natural pregnancy group, 46.1% showed euploid and 53.9% was aneuploid. In subjects who were pregnant by COH, 45.5% and 54.4% of patients showed euploid and aneuploid, respectively. COH did not affect fetal karyotyping (*P*=0.910).\n\nUnivariate analysis revealed that maternal age, gestational weeks at abortion and AMH level were associated with aneuploid of conceptus. With these three markers, we conducted logistic regression analysis to evaluate whether maternal age affects the relationship of AMH and gestational weeks to abnormal karyotyping ([Table 3](#T3){ref-type=\"table\"}). When advanced maternal age was analyzed as a categorical variable (age \u226535 years), the association between fetal aneuploidy and AMH level and gestational weeks reached statistical significance. AMH level was higher and gestational weeks was significantly lower in aneuploid group regardless of maternal age (*P*=0.045 and *P*\\<0.001, respectively).\n\nDiscussion\n==========\n\nThe present study was performed to determine whether a serum AMH level could be an independent predicting marker for aneuploid. The result of the study shows that significant age-independent relationship between elevated AMH level and fetal aneuploid. However, basal FSH is not associated with the fetal aneuploid.\n\nAneuploidy is the most commonly identified chromosome abnormality in humans, occurring in at least 5% of all clinically recognized pregnancies. Despite the high frequency and devastating clinical results of aneuploidy, we know relatively little about factors that modulate and predict the risk of aneuploidy. Although several risk factors except genetic one have been suggested including oral contraceptives, fertility drugs, thyroid antibodies, alcohol consumption, parity and maternal diabetes, none of these have been proved to have any other associations \\[[@B18][@B19][@B20][@B21]\\]. The only one indisputable factor associated with fetal chromosomal aberrations is advanced maternal age. Chronologic ovarian aging influence diminished ovarian reserve. Various endocrinological markers have been used to assess the ovarian reserve such as AMH and FSH. There are conflicting data regarding whether compromised ovarian reserve markers are associated with an increase in aneuploid pregnancies \\[[@B22][@B23][@B24][@B25][@B26]\\].\n\nAMH is the dimeric glycoprotein and the member of the transforming growth factor-\u03b2 superfamily. In the ovary, AMH is synthesized by granulosa cells only after birth or at the end of fetal life. AMH regulates steroidogenesis of ovary and influences folliculogenesis. Recent studies indicate AMH as an important novel measure of ovarian reserve. AMH appears to offer several potential advantages over other tests of ovarian reserve. The circulating AMH seems to be solely of ovarian origin because one study found AMH to be undetectable 3 to 5 days after bilateral ovariectomy \\[[@B27]\\]. Moreover AMH may be a unique endocrine parameter of ovarian function, since several studies have demonstrated that, in contrast to sex steroids, gonadotrophins and peptides, AMH serum levels do not significantly change throughout the menstrual cycle \\[[@B15][@B28][@B29][@B30]\\].\n\nThere are limited numbers of the study that have examined the potential of AMH as predictor of fetal aneuploidy. Seifer and Maclaughlin \\[[@B11]\\] conducted case-control study using stored serum samples which were obtained for antenatal chromosome screening. The study included 25 Down\\'s syndrome sample and 125 unaffected controls. The authors could not find any difference in maternal serum AMH level between Down\\'s syndrome pregnancies and controls. However, the study only focused on trisomy 21 alone and did not take into account the pregnancies that were terminated in the first trimester.\n\nPlante et al. \\[[@B24]\\] reported that AMH value declined significantly with advancing maternal age and AMH level did not differ between women with an aneuploidy fetus and women with a euploid fetus. The authors concluded that AMH level did not predict fetal aneuploidy. However, only 18 patients carrying an aneuploidy fetus served as cases and serum AMH level was measured by two different methods which could resulted in a bias.\n\nIn this study, AMH, as it was related to age, could not be significant predicting factor of fetal aneuploid when multivariate analysis was done by using age as continuous variable. Authors noticed that clinical predicting factor of fetal aneuploid that was usually used to counsel about infertility or normal pregnancy, was whether patient\\'s age is over 35 or not, and above result was obtained by analyzing using categorical variable appointing age 35 as standard. Using a cutoff of 35 years, significant association between elevated AMH and fetal aneuploidy rate was found in the present study (*P*=0.045). Compared with previous results, our data reported conflicting result. The reason of conflicting result might be due to the differences in study population. Previous study included samples which were obtained from not patients who were diagnosed early pregnancy loss but subjects who were in on-going pregnancy. In our study, the serum was obtained before pregnancy. On the other hand, previous study analyzed AMH level during pregnancy. Lastly, our study focused on early pregnancy loss and the mean gestational weeks at abortion was around 8 gestational weeks. However, in the other studies, the pregnancy was maintained more than 13 gestational weeks. Those differences might results in conflicting result among studies \\[[@B11][@B24]\\].\n\nIn comparison to previous study, our mean AMH level was relatively high as 3.43 ng/mL. It is thought to be influenced by involving the whole group not excluding 18 polycystic ovary syndrome patients typically showing high level of AMH.\n\nElevated basal FSH concentration is a known marker for decreased ovarian reserve. There are conflicting results regarding whether basal FSH level may be directly associated with an increase in aneuploid pregnancies. Several studies suggested that elevated basal FSH concentrations were related to aneuploid pregnancies \\[[@B23][@B31]\\]. However, recent studies contradict this conclusion. van Montfrans et al. \\[[@B32]\\] observed no significant effect of basal FSH concentrations on the incidence of early pregnancy loss or abortion of clinically recognized pregnanices in their prospective study. Massie et al. \\[[@B33]\\] evaluated whether a basal FSH level could be an independent predictor of fetal aneuploidy. The authors performed the study with 177 spontaneous miscarriage samples which included 70 euploid and 107 aneuploid. The result demonstrated that an elevated basal FSH was not associated with an increase in fetal aneuploid. The result of our study does not support the findings of earlier studies. Although FSH level in fetal aneuploid group was slightly elevated than in fetal euploid group, there was no statistically significant difference between euploid and aneuploid groups. When make a conclusion by these comprehensive studies, FSH concentration most probably cannot be used to predict the risk of fetal aneuploid in the pregnant woman.\n\nBoth abortion rates and chromosomal anomalies decrease with advancing gestational age by first, second and third trimester. Percentage of chromosomal anomalies of abortus or stillbirth on each trimester are 55%, 35%, 5% respectively. But in the present study consisting of the only first trimester pregnancy, the miscarriage was diagnosed earlier in aneuploid group than euploid group (7.67\u00b11.54 vs. 8.27\u00b11.46, *P*\\<0.001). As our knowledge, no previous study was examined focusing on the association between gestational age and chromosomal anomalies in the early pregnancy loss. Our result suggests that another nonnegligible factor such as infection, medical disorders, immunologic factor, inherited thrombophilias and uterine defects might affect the early spontaneous miscarriage besides fetal factors. Especially, abortion due to immunologic factor happens by miscarriage at implantation stage, and it is thought to be aborted at 1st trimester although fetal karyotype is normal.\n\nThere are several limitations in the present study. The first is that there are some missing values of AMH level in spite of the fact that AMH is significant factor to predict fetal aneuploidy. Initial purpose of this study, however, was to look over general ovarian reserve markers that affect fetal aneuploidy. Thus, all patients were included in study although there were missing values of AMH. To check whether this missing value is random missing value or not, independent t-test was used in both euploid group and anuploid group to compare patient characteristics between AMH-known patient group and AMH-missed value patient group, and no difference between two patient groups in both euploid and aneuploid group was found. Second we used two different methods, classical karyotyping and MLPA to examine fetal karyotyping. Lastly, study and control groups were heterogenous. Although infertility treatment did not make any differences in the result of fetal karyotyping, more than half of the subjects were received controlled ovarian hyperstimulation followed by IVF-ET, ICSI and IUI.\n\nOur study offers several strengths compared to previous reports. The Large sample size in this study relative to others increases the reliability of our findings. We definitely identified fetal chromosome by classical karyotyping which was used for precise results.\n\nWe tried to make the cutoff point of AMH level to be the standard judgement of fetal aneuploidy and to define early pregnancy loss. However receiver operating characteristic curve shows 0.607 which is somewhat insufficient numerical value to expect the obvious result. To deduct the clinically practical cutoff value in spite of ambiguous supposition, larger sample group will be needed.\n\nIn conclusion, low maternal AMH level does appear to be a clinical age-independent marker of fetal aneuploidy in early pregnancy losses but FSH level was not predicting factor of expecting fetal aneuploidy in our study. In near future, further study with prospective manner will be needed to confirm the relationship between serum AMH level and aneuploidy risk in early pregnancy loss.\n\n**Conflict of interest:** No potential conflict of interest relevant to this article was reported.\n\n###### The incidence of chromosomal aberrations in abortus\n\n![](ogs-58-494-i001)\n\n###### Characteristics of patients with euploid versus aneuploid fetal karyotypes\n\n![](ogs-58-494-i002)\n\nValues are presented as mean\u00b1standard deviation or number (%).\n\nAMH, antimullerian hormone; FSH, follicle stimulating hormone; MLPA, multiplex ligation-dependent probe amplification; COH, controlled ovarian hyperstimulation.\n\n###### Logistic regression analysis of potential marker for aneuploidy in early pregnancy loss\n\n![](ogs-58-494-i003)\n\n^a)^Analyzed as a categorical variable (\\<35, \u226535 years).\n"}
+{"text": "BoNT\n\n:   botulinum neurotoxin\n\nGBS\n\n:   ganglioside\u2010binding site\n\nH~C~\n\n:   cell\u2010binding domain\n\nH~N~\n\n:   translocation domain\n\nLC\n\n:   light chain\n\n*Clostridial botulinum* neurotoxins (BoNTs) are responsible for causing the deadly condition, botulism, in vertebrates \\[[1](#feb412931-bib-0001){ref-type=\"ref\"}, [2](#feb412931-bib-0002){ref-type=\"ref\"}, [3](#feb412931-bib-0003){ref-type=\"ref\"}, [4](#feb412931-bib-0004){ref-type=\"ref\"}\\]. There are seven distinct serotypes termed BoNT/A through BoNT/G, of which serotypes /A, /B, /E and /F[5](#feb412931-bib-0005){ref-type=\"ref\"}\\]. Each BoNT serotype can be further categorised into subtypes based on amino acid sequence identity. For example, there are currently eight known subtypes of BoNT/A (/A1\u2010/A8), which share between 84% and 97% sequence identity \\[[6](#feb412931-bib-0006){ref-type=\"ref\"}\\]. While BoNTs are the most toxic biological molecules known to science, they are used in human therapy, especially BoNT/A1 \\[[7](#feb412931-bib-0007){ref-type=\"ref\"}\\].\n\nThe BoNTs contain three major functional domains, a binding domain located in the C\u2010terminal half of the heavy chain (H~C~), a translocation domain located in the N\u2010terminal half of the heavy chain (H~N~) and a Zn^2+^\u2010dependent protease domain located in the light chain (LC). The H~C~ is responsible for targeting the BoNT to the neuronal cell membrane by binding to specific gangliosides and protein receptors on the neuronal cell surface. The H~N~ facilitates entry of the LC into the cytosol where it cleaves a target SNARE protein(s), which inhibits exocytosis. Although there are currently more than 46 different BoNT subtypes, there is limited structural information available for the majority of these natural variants. Many of these subtypes have been found to contain beneficial properties when compared to the commercially available toxins.\n\nThe BoNT subtypes from within the same serotype display a high degree of amino acid sequence identity and similarity; however, several studies have found distinct differences in their properties \\[[8](#feb412931-bib-0008){ref-type=\"ref\"}, [9](#feb412931-bib-0009){ref-type=\"ref\"}, [10](#feb412931-bib-0010){ref-type=\"ref\"}, [11](#feb412931-bib-0011){ref-type=\"ref\"}, [12](#feb412931-bib-0012){ref-type=\"ref\"}\\] (Fig.\u00a0[1](#feb412931-fig-0001){ref-type=\"fig\"}). Although the molecular basis of intoxication is not yet fully understood, the LC appears to define the length of intoxication (duration of action), while both H~N~ and H~C~ appear to be responsible for the spread and speed of cellular entry (onset of action). Considering the toxic nature of BoNTs, they are classed as tier 1 select agents due to their potential misuse in bioterrorism or as a bioweapon. From this perspective, structural details of each subtype may aid the design of broadly BoNT\u2010neutralising antibodies.\n\n![Alignment of the binding domain sequences from BoNT/A1 to A8. BoNT/A1 numbering and secondary structure used for annotation. Figure generated using ESPript \\[[34](#feb412931-bib-0034){ref-type=\"ref\"}\\]](FEB4-10-1474-g001){#feb412931-fig-0001}\n\nPreviously, we and others have determined the crystal structures of the binding domains from BoNT subtypes /A1, /A2, /A3 and /A4, and the related /HA alone \\[[13](#feb412931-bib-0013){ref-type=\"ref\"}, [14](#feb412931-bib-0014){ref-type=\"ref\"}, [15](#feb412931-bib-0015){ref-type=\"ref\"}, [16](#feb412931-bib-0016){ref-type=\"ref\"}, [17](#feb412931-bib-0017){ref-type=\"ref\"}\\], and in complex with various receptors: H~C~/A1\u2010GT1b \\[[18](#feb412931-bib-0018){ref-type=\"ref\"}\\], H~C~/A1\u2010SV2C \\[[19](#feb412931-bib-0019){ref-type=\"ref\"}, [20](#feb412931-bib-0020){ref-type=\"ref\"}\\], H~C~/A2\u2010SV2C \\[[14](#feb412931-bib-0014){ref-type=\"ref\"}, [21](#feb412931-bib-0021){ref-type=\"ref\"}\\] and H~C~/A3\u2010GD1a \\[[22](#feb412931-bib-0022){ref-type=\"ref\"}\\]. Here, we report the crystal structures of the BoNT/A5 and BoNT/A6 receptor\u2010binding domain and compare the binding sites with previous crystal structures of the BoNT/A subtype.\n\nMaterials and methods {#feb412931-sec-0002}\n=====================\n\nAll reagents used were purchased from Sigma\u2010Aldrich (Dorset, UK) or Fisher Scientific (Leicestershire, UK) unless otherwise specified.\n\nProtein expression and purification {#feb412931-sec-0003}\n-----------------------------------\n\nThe binding domain (residues 871--1296) of BoNT/A5 and BoNT/A6 was cloned into the pJ401 vector (Atum Bio, California, USA) from their respective full\u2010length sequences (UniProtKB: C7BEA8 and C9WWY7) with an N\u2010terminal 6xHis tag. Constructs were expressed and purified as described previously \\[[17](#feb412931-bib-0017){ref-type=\"ref\"}\\]. The N\u2010terminal 6xHis tag was not removed from the proteins prior to crystallisation.\n\nProtein crystallisation {#feb412931-sec-0004}\n-----------------------\n\nCrystallisation conditions were screened using commercially available 96\u2010well screens from Molecular Dimensions (Sheffield, UK) at 16\u00a0\u00b0C. H~C~/A5 (4\u00a0mg\u00b7mL^\u22121^) and H~C~/A6 (6\u00a0mg\u00b7mL^\u22121^) were dispensed using an Art Robbins Phoenix crystal screening nano\u2010dispenser into 96\u2010well 3\u2010drop Intelliwell plates (Molecular Dimensions, UK). Multiple screening kits from Molecular Dimensions were used. Crystals of H~C~/A5 were obtained using the sitting\u2010drop vapour diffusion method with 0.1\u00a0[m]{.smallcaps} sodium formate, 0.1\u00a0[m]{.smallcaps} ammonium acetate, 0.1\u00a0[m]{.smallcaps} sodium citrate tribasic dihydrate, 0.1\u00a0[m]{.smallcaps} sodium potassium tartrate tetrahydrate, 0.1\u00a0[m]{.smallcaps} sodium oxamate, 0.1\u00a0[m]{.smallcaps} imidazole, 0.1\u00a0[m]{.smallcaps} 2\u2010\\[N\u2010morpholino\\]ethanesulfonic acid, pH 6.5, 10%(v/v) ethylene glycol, 10%(w/v) PEG 8000 from (MORPHEUS screen, condition G2) and flash\u2010cooled in liquid nitrogen. Crystals of H~C~/A6, however, were obtained using the hanging\u2010drop vapour diffusion method and 0.2\u00a0[m]{.smallcaps} sodium acetate trihydrate, 0.1\u00a0[m]{.smallcaps} Bis\u2010Tris propane\u2010HCl pH 7.5, 22%(w/v) PEG 3350 (based on condition G7 of the PACT Premier screen) and flash\u2010cooled in liquid nitrogen after cryoprotection with 1\u00a0:\u00a01 50%(v/v) glycerol in reservoir solution.\n\nX\u2010ray data collection and structure determination {#feb412931-sec-0005}\n-------------------------------------------------\n\nComplete X\u2010ray diffraction data sets were collected from single crystals of H~C~/A5 and H~C~/A6 (3600 images each) using 0.1\u00b0 oscillations and a wavelength of 0.98\u00a0\u00c5 at beamlines IO3 and IO4 (Diamond Light Source, Didcot, UK). Raw images were processed using DIALS \\[[23](#feb412931-bib-0023){ref-type=\"ref\"}\\], and integrated data were scaled and merged using Aimless \\[[24](#feb412931-bib-0024){ref-type=\"ref\"}\\] from the CCP4 suite \\[[25](#feb412931-bib-0025){ref-type=\"ref\"}\\]. The 3D structures of both proteins were solved by molecular replacement with PHASER \\[[26](#feb412931-bib-0026){ref-type=\"ref\"}\\] using the coordinates from Phyre2 web server homology models \\[[27](#feb412931-bib-0027){ref-type=\"ref\"}\\] as search models. Both models were manually built [coot]{.smallcaps} \\[[28](#feb412931-bib-0028){ref-type=\"ref\"}\\] and refined with [refmac]{.smallcaps} \\[[29](#feb412931-bib-0029){ref-type=\"ref\"}\\] in the CCP4 suite of programs \\[[25](#feb412931-bib-0025){ref-type=\"ref\"}\\]. The structures were validated with [pdb_redo]{.smallcaps} \\[[30](#feb412931-bib-0030){ref-type=\"ref\"}\\], [molprobity]{.smallcaps} \\[[31](#feb412931-bib-0031){ref-type=\"ref\"}\\] and [wwpdb validation]{.smallcaps} \\[[32](#feb412931-bib-0032){ref-type=\"ref\"}\\]. Crystallographic data processing and refinement statistics are given in Table\u00a0[1](#feb412931-tbl-0001){ref-type=\"table\"}. Structure\u2010based figures were generated with either PyMOL (Schr\u00f6dinger, LLC, New York, NY, USA) or MOE (Chemical Computing Group, Quebec).\n\n###### \n\nX\u2010ray data collection and refinement parameters. Outer shell statistics are shown in parentheses\n\n                                H~C~/A5                                       H~C~/A6\n  ----------------------------- --------------------------------------------- ---------------------------------------------\n  Beamline                      I03, DLS                                      I04\u20101, DLS\n  Wavelength (\u00c5)                0.9763                                        0.9159\n  Space group                   P2~1~2~1~2~1~                                 P2~1~2~1~2~1~\n  Unit\u2010cell parameters                                                        \n  a,b,c (\u00c5)                     43.55, 60.27, 185.15                          39.54, 105.59, 112.41\n  \u03b1\u00a0=\u00a0\u03b2 = \u03b3 (\u25e6)                 90, 90, 90                                    90, 90, 90\n  Resolution (\u00c5)                92.57--1.15 (1.17--1.15)                      112.41--1.35 (1.37--1.35)\n  *R* ~merge~ (%)               0.087 (1.098)                                 0.094 (1.725)\n  *R* ~meas~ (%)                0.090 (1.158)                                 0.099 (1.839)\n  *R* ~pim~ (%)                 0.026 (0.363)                                 0.030 (0.631)\n  CC~1/2~ (%)                   0.998 (0.447)                                 0.999 (0.463)\n  Mean\u00a0\\<\u00a0I/\u03c3(I)\\>              11.1 (1.9)                                    10.7 (0.9)\n  Completeness (%)              100 (99.5)                                    100 (99.7)\n  No. of observed reflections   2,107,443 (84,519)                            1,099,418 (43,053)\n  No. of unique reflections     173,797 (8,450)                               104,434 (5,129)\n  Multiplicity                  12.1 (10.0)                                   10.5 (8.4)\n  Refinement statistics                                                       \n  *R* ~work~/*R* ~free~         0.135/0.161                                   0.147/0.168\n  RMSD bond lengths (\u00c5)         0.02                                          0.01\n  RMSD bond angles (\u25e6)          2.20                                          1.66\n  Ramachandran statistics (%)                                                 \n  Favoured                      96.3                                          97.0\n  Allowed                       3.7                                           3.0\n  Outliers                      0                                             0\n  Wilson B\u2010factor (\u00c5^2^)        11.4                                          13.5\n  Average B\u2010factors (\u00c5^2^)                                                    \n  Protein                       16.8                                          20.3\n  Water                         32.6                                          31.5\n  No. of atoms                                                                \n  Protein                       3827                                          3633\n  Water                         522                                           423\n  PDB code                      [6TWP](https://doi.org/10.2210/pdb6TWP/pdb)   [6TWO](https://doi.org/10.2210/pdb6TWO/pdb)\n\nJohn Wiley & Sons, Ltd\n\nResults and Discussion {#feb412931-sec-0006}\n======================\n\nStructure of the BoNT/A5\u2010binding domain (H~C~/A5) {#feb412931-sec-0007}\n-------------------------------------------------\n\nThe crystal of H~C~/A5 belonged to the orthorhombic space group *P*2~1~2~1~2~1~, and it diffracted to a resolution of 1.15\u00a0\u00c5 (Table\u00a0[1](#feb412931-tbl-0001){ref-type=\"table\"}). Electron density was excellent throughout, with all H~C~/A5 residues (except the N\u2010terminal 6xHis tag and Lys871) being easily observed. The structure closely resembles the structures of other BoNT\u2010binding domains \\[[6](#feb412931-bib-0006){ref-type=\"ref\"}\\] with an N\u2010terminal jelly roll\u2010like fold and C\u2010terminal modified \u03b2\u2010trefoil fold containing a conserved ganglioside\u2010binding site (SxWY) (Fig.\u00a0[2](#feb412931-fig-0002){ref-type=\"fig\"}). However, compared to the structure of BoNT/A1 in complex with GT1b (PDB: [2VU9](https://doi.org/10.2210/pdb2VU9/pdb)), the loop of residues 1260--1280, which contains ganglioside\u2010interacting residues, adopts a different arrangement (Fig.\u00a0[3a,b](#feb412931-fig-0003){ref-type=\"fig\"}). It is possible that upon ganglioside binding, the loop changes conformation to allow S1275 and R1276 to take part in the interaction. In comparison with the unbound GD1a\u2010binding site of BoNT/A1 (PDB: [3BTA](https://doi.org/10.2210/pdb3BTA/pdb)) and BoNT/A3 (PDB: [6F0O](https://doi.org/10.2210/pdb6F0O/pdb)), the corresponding site in H~C~/A5 perhaps more resembles that of the latter rather than the former, which is consistent with a higher sequence identity between the sites at residues corresponding to positions 1117, 1254 and 1278 (Figs\u00a0[1](#feb412931-fig-0001){ref-type=\"fig\"} and [4](#feb412931-fig-0004){ref-type=\"fig\"}). Either way, considering that both BoNT/A1 and BoNT/A3 are able to bind to GD1a (PDB: [5TPB](https://doi.org/10.2210/pdb5TPB/pdb) and [6THY](https://doi.org/10.2210/pdb6THY/pdb), respectively), this suggests that BoNT/A5 is able to do so too.\n\n![Crystal structures of H~C~/A5 and H~C~/A6. Overall structure of the binding domain (H~C~) of BoNT/A5 (top) and BoNT/A6 (bottom). Putative ganglioside\u2010 and SV2\u2010binding sites are indicated by a dashed ellipse labelled GBS and SV2BS, respectively.](FEB4-10-1474-g002){#feb412931-fig-0002}\n\n![Comparison of receptor\u2010binding sites of BoNT/A5 and BoNT/A6. (A\u2010C) Comparison of the ganglioside\u2010binding site of H~C~/A1 (PDB: [2VU9](https://doi.org/10.2210/pdb2VU9/pdb)), H~C~/A5 and H~C~/A6. Note that the ganglioside, GT1b (yellow), has been superposed into the putative site of H~C~/A5 and H~C~/A6. (D\u2010F) Comparison of the SV2\u2010binding site of H~C~/A1 (PDB: [5JLV](https://doi.org/10.2210/pdb5JLV/pdb)), H~C~/A5 and H~C~/A6. Note that the protein receptor, SV2C (purple), has been superposed into the putative site of H~C~/A5 and H~C~/A6. (G--I) Comparison of the SV2 glycan\u2010binding site of H~C~/A1 (PDB: [5JLV](https://doi.org/10.2210/pdb5JLV/pdb)), H~C~/A5 and H~C~/A6. Note that the SV2C glycans (purple) have been superposed into the putative site of H~C~/A5 and H~C~/A6. Residues involved in receptor binding are indicated and those which differ from BoNT/A1 are highlighted in bold.](FEB4-10-1474-g003){#feb412931-fig-0003}\n\n![Comparison of the GD1a\u2010binding site. (A) The GD1a\u2010binding site of BoNT/A1 is shown in the bound (red; PDB: [5TPB](https://doi.org/10.2210/pdb5TPB/pdb)) and unbound (yellow; PDB: [3BTA](https://doi.org/10.2210/pdb3BTA/pdb)) conformation (the interacting sugar moieties are shown in white). (B) The GD1a\u2010binding site of BoNT/A3 is shown in the bound (orange; PDB: [6THY](https://doi.org/10.2210/pdb6THY/pdb)) and unbound (cyan; PDB: [6F0O](https://doi.org/10.2210/pdb6F0O/pdb)) conformation (the interacting sugar moieties are shown in grey). (C) The putative GD1a\u2010binding site of BoNT/A5 (pink) and BoNT/A6 (green) is shown in the unbound conformation. Note that residues at positions 1117, 1254 and 1278 either match the corresponding residue in BoNT/A1, BoNT/A3 or neither.](FEB4-10-1474-g004){#feb412931-fig-0004}\n\nIn close proximity to the ganglioside binding loop was observed an unusual feature -- a methylene bridge between the S~\u03b3~ of Cys1280 and N~\u03b6~ of Lys1236 (Fig.\u00a0[5a](#feb412931-fig-0005){ref-type=\"fig\"}), rather than a disulfide bond with a nearby cysteine residue (Cys1235). During refinement of the H~C~/A5 structure, clear electron density was observed between the side chains of Cys1280 and Lys1236, into which a methylene group could be fitted. Weak anomalous data recorded at the start of data collection were used to generate a low\u2010resolution anomalous difference map. Despite the noise, large peaks were observed at the location of sulfur atoms, which confirmed the location of each cysteine residue (Fig.\u00a0[5b](#feb412931-fig-0005){ref-type=\"fig\"}). This specific methylene bond between a lysine and cysteine side chain is unusual, and the mechanism surrounding the formation of a methylene\u2010bridged lysine and cysteine is not fully understood \\[[33](#feb412931-bib-0033){ref-type=\"ref\"}\\]. Whether this bond is biologically relevant remains to be established. While there are indications of this bond in the electron density maps of other BoNT crystal structures, it is possible that this may be an artefact of exposure to synchrotron radiation.\n\n![Electron density around Cys1280 and Lys1236 in H~C~/A5. (A) The electron density around Cys1280 and Lys1236 is clear and continuous, indicating the presence of a covalent bond between the two side chains. The 2mFo\u2010Fc map (blue) is contoured to 0.5 e/\u00c5^3^, and the Fo\u2010Fc (negative: red; positive: green) map is contoured to 0.46 e/\u00c5^3^. (B) A weak sulfur\u2010SAD signal within the diffraction data enabled the calculation of an anomalous map (contoured to 0.06 e/\u00c5^3^; magenta), which indicates the precise position of each sulfur atom for each Cys residue.](FEB4-10-1474-g005){#feb412931-fig-0005}\n\nInspection of the H~C~/A5 structure corresponding to the BoNT/A1 SV2C\u2010binding site (^1139^PRGSVMTT^1146^\u00a0+\u00a0Arg1156) reveals the presence of perhaps a slightly shortened \u03b2\u2010hairpin (Fig.\u00a0[3d,e](#feb412931-fig-0003){ref-type=\"fig\"}). The three different residues at positions 1143, 1144 and 1156 (V\u00a0\u2192\u00a0I, M\u00a0\u2192\u00a0V and R\u00a0\u2192\u00a0M, respectively) do not appear to preclude the possibility of SV2C binding. Indeed, the related binding domain of BoNT/HA possesses the same residues at the corresponding location and is still able to bind to SV2C \\[[20](#feb412931-bib-0020){ref-type=\"ref\"}\\]. However, inspection of the accompanying SV2C glycan\u2010binding site reveals one residue (Gln1064) potentially hindering the binding of glycan (Fig.\u00a0[3g](#feb412931-fig-0003){ref-type=\"fig\"},h). This residue, corresponding to His1064 in BoNT/A1, has been shown to drastically decrease the binding affinity to SV2C \\[[20](#feb412931-bib-0020){ref-type=\"ref\"}\\]. Although this suggests that SV2C may not be the protein receptor for BoNT/A5, it should be noted that there exists a second BoNT/A5 sequence that differs by this one residue (UniProtKB: C1IPK2).\n\nStructure of the BoNT/A6\u2010binding domain (H~C~/A6) {#feb412931-sec-0008}\n-------------------------------------------------\n\nThe crystals of H~C~/A6 belong to orthorhombic space group *P*2~1~2~1~2~1~ and diffracted to a resolution of 1.35\u00a0\u00c5. Electron density was excellent, with all but the first six residues of H~C~/A6 being clearly observed, and like H~C~/A5, the overall protein fold was highly similar to other BoNT H~C~ structures. The ganglioside\u2010binding site was identical to that of H~C~/A5 except for residue 1117, which was a Phe rather than a Tyr (Fig.\u00a0[3b,c](#feb412931-fig-0003){ref-type=\"fig\"}). Although the absence of the hydroxyl group would result in the loss of hydrogen bonding with the terminal sialic acid of GT1b, the side chain can still continue to interact with the carbon ring. Compared to the unbound GD1a\u2010binding site of BoNT/A1 (PDB: [3BTA](https://doi.org/10.2210/pdb3BTA/pdb)) and BoNT/A3 (PDB: [6F0O](https://doi.org/10.2210/pdb6F0O/pdb)), the corresponding site in H~C~/A6 also more resembles that latter rather than the former, even though there is no greater sequence identity between the sites at residues corresponding to positions 1117, 1254 and 1278 (Figs\u00a0[1](#feb412931-fig-0001){ref-type=\"fig\"} and [4](#feb412931-fig-0004){ref-type=\"fig\"}). Like that for the H~C~/A5 structure, BoNT/A6 is predicted to be able to bind to GD1a as well.\n\nFor the corresponding BoNT/A1\u2010SV2C\u2010binding site in H~C~/A6, a larger sequence variation is observed: ^1139^SRSTLLTT^1146^\u00a0+\u00a0Met1156 rather than ^1139^PRGSVMTT^1146^\u00a0+\u00a0Arg1156 for BoNT/A1. Despite these differences, the \u03b2\u2010hairpin remains available to bind to SV2C via mostly backbone--backbone hydrogen bonding (Fig.\u00a0[3d,f](#feb412931-fig-0003){ref-type=\"fig\"}). Like H~C~/A5, H~C~/A6 possesses a different residue in the glycan\u2010binding site at position 1064 (Arg) compared to that of BoNT/A1 (His, Fig.\u00a0[3g,i](#feb412931-fig-0003){ref-type=\"fig\"}), and this has also been reported to significantly reduce binding of glycosylated SV2C \\[[20](#feb412931-bib-0020){ref-type=\"ref\"}\\]. This would suggest that BoNT/A6 may have a lower affinity for SV2C than BoNT/A1. Interestingly, BoNT/A2 also contains an Arg at position 1064 and it has previously been reported that both BoNT/A2 and BoNT/A6 are capable of entering hiPSC\u2010derived neurons faster than BoNT/A1 \\[[8](#feb412931-bib-0008){ref-type=\"ref\"}\\].\n\nConclusions {#feb412931-sec-0009}\n===========\n\nThe BoNT/A subtypes are believed to bind to the target cell surface via a dual\u2010receptor complex involving a ganglioside and protein receptor. For BoNT/A1, they are GT1b (preferentially) and SV2C, respectively, but for most of the others, the exact identities of these receptors have not yet been determined. Structural analysis of the expected binding sites has revealed some differences with that of BoNT/A1, suggesting either an altered binding affinity to each receptor or a different receptor specificity altogether. Our high\u2010resolution structures further add to the body of knowledge around BoNT receptor binding and enhance the available molecular information for engineering novel therapeutic BoNTs and BoNT\u2010binding moieties.\n\nAuthor contributions {#feb412931-sec-0011}\n====================\n\nJRD performed purification and crystallisation of H~C~/A5, supervised AB, processed and analysed the data, and drafted the manuscript. AB performed purification and crystallisation of H~C~/A6. SML analysed the data and edited the manuscript. KRA supervised the study, analysed the data, and edited the manuscript.\n\nConflict of interest {#feb412931-sec-0012}\n====================\n\nThe authors JRD, AB and KRA from the University of Bath declare no competing financial interests. SML is an employee of Ipsen Bioinnovation Limited.\n\nWe thank Diamond Light Source (UK) for the use of beamlines IO3 and IO4\u20101 (proposal mx17212). J.R.D. was supported by a joint postgraduate studentship between University of Bath (UK) and Ipsen Bioinnovation Limited.\n\nData accessibility {#feb412931-sec-0014}\n==================\n\nAccession codes: The atomic coordinates and structure factors (codes [6TWP](https://doi.org/10.2210/pdb6TWP/pdb) and [6TWO](https://doi.org/10.2210/pdb6TWO/pdb)) have been deposited in the Protein Data Bank (<http://www.pdb.org>).\n"}
+{"text": "Introduction\n============\n\nLeishmaniasis refers to a variety of diseases caused by more than 20 species of intracellular protozoan parasites belonging to the genus *Leishmania*. The clinical spectrum of the disease ranges from simple self-limiting cutaneous ulcers to severe disfiguring mucocutaneous, and even to a fatal visceral disease known as Kala Azar. About 98 tropical and subtropical countries are known to be endemic for this disease, with 350 million people at risk and overall prevalence of 12 million worldwide ([@B1]). Iran is endemic to both cutaneous and visceral leishmaniasis. Cutaneous Leishmaniasis (CL) is commonly caused by *L. major* and *L. tropica* ([@B2]). Visceral leishmaniasis (VL), the most life threatening form, is caused by *L. infantum* and in very rare occasions by *L. tropica* ([@B2], [@B3]).\n\nIn VL fever and hepato-splenomegaly are the main clinical signs in which *Leishmania* parasite is dispersed to the internal viscera like spleen, liver and bone marrow ([@B4]). Based on the leishmaniasis clinical symptoms, it is evident that the host immunity factors, *Leishmania* species, and in some cases the *Leishmania* strain, determines the measure of pathogenecity ([@B5]). *Leishmania* spp. has about 8000 genes among only 78 genes are restricted to individual species ([@B6]). In spite of a few species parasite genes implicated in pathogenesis and clinical presentation, the parasite gene expression rates differ greatly among species ([@B6]).\n\nIn leishmaniasis, parasites are challenged by the host immune conditions throughout their life cycle such as temperature increase of visceral tissues (liver, spleen or bone marrow). Such challenges causes *Leishmania* experience biochemical changes in which post transcriptional modification are activated and may eventuate into the emergence of the leishmaniasis pathogencity ([@B7]--[@B14]). Proteomics is an invaluable tool for systematic analysis of the proteome. Analysis of proteome is most commonly performed by a combination of 2-DE and mass spectrometry (MS). 2-DE method could separate proteins in first and second dimensions according to their isoelectric and molecular weight points. With the help of 2-DE and the MS, a variable mixture of proteins is separated, visualized and then identified ([@B15]--[@B16]).\n\nIn this preliminary study, we compared the proteome mapping, in three Iranian isolates of *Leishmania* species including *L. tropica, L. major* and *L. infantum,* with immobilized pH gradient stripes with linear pH 4--7. Moreover, Liquid Chromatography (LC) - mass spectrometry was used for identification of a number of differentially expressed proteins among the three species.\n\nMaterials and Methods\n=====================\n\nLeishmania isolates and cell culture\n------------------------------------\n\nThe proteome of three *Leishmania* species including *L. tropica* (GenBank accession nos. **EF653267**, *L. major* (**JN860745**) and *L. infantum* (**JX289853**) compared and were analyzed. promastigote forms recovered from the \"Iranian *Leishmania* parasite bank\" located in Leishmaniasis lab, School of Public Heath, Tehran University of Medical Sciences (TUMS). The identity of these strains was already obtained by other molecular DNA based methods ([@B2], [@B17]).\n\n***Cell culture --*** Promastigotes recovered from liquid nitrogen (\u2212196 \u00b0C), were mass cultured in RPMI1640 medium (Gibco, Life technologies GmbH, Frankfurt, Germany) supplemented with 15% heat inactivated fetal bovine serum (Gibco, Germany) and 100U/ml penicillin and 100ug/ml streptomycin (Gibco, Germany) and incubated at 24\u00b0C. Promastigotes harvested in the stationary phase.\n\nParasites were harvested washed in sterile Phosphate Buffered Saline (PBS, pH: 7.2--7.4) and were used for protein extraction.\n\nProtein Extraction\n------------------\n\nProteomics analysis was performed on *L. tropica, L. major* and *L. infantum*, three species at the time of study. Promastigotes were harvested by centrifugation at 3000rpm, 4\u00b0 C and 20 minutes and washed three times in sterile PBS (pH: 7.2--7.4) in the same condition for 10 minutes. The cells were resuspended in 5 mM Tris--HCl, pH 7.8, containing 1 mM phenylmethylsulfonyl fluoride (PMSF (Merck, Germany)). Proteins were precipitated by 10% (w/v) trichloroacetic acid (Merck, Germany) in acetone (Merck) with 0.07 %(w/v) dithiothreitol (DTT) (Merck) for 1 hour at \u221220\u00b0C. The samples were then centrifuged at 17500 g (Hettich, Germany) for 15 minutes at 4\u00b0 C and the pellets were washed with ice-cold acetone containing 0.07% DTT, incubated at \u221220\u00b0C for 1 h and centrifuged at 4 \u00b0C. The samples were then solubilized in lysis buffer (9.5 M urea (Merck), 2% (w/v) CHAPS (Merck), 0.8% (w/v) Ampholyte (Bio-Rad, USA) pH 3--10, 1% (w/v) DTT) ([@B18], [@B19]). The concentration of protein was measured by the Bradford assay with bovine serum albumin BSA (Merck) as the standard ([@B20]).\n\nTwo-Dimensional electrophoresis (2-DE)\n--------------------------------------\n\nFor analytical and preparative gels, 120 \u03bcg and 1.2 mg of extracted promastigotes proteins were loaded respectively. IEF was carried out on the 18 cm immobilized pH gradient (IPG) strips (pH 4--7) (Bio-Rad, USA). IPG strips were rehydrated overnight by loading the samples diluted with rehydration buffer containing 8 M urea, 4% CHAPS, 2% ampholyte, 50 mM DTT, and traces of bromophenol blue (Merck). Isoelectric focusing was conducted at 20 \u00b0C with Mutiphor II and a DryStrip kit (GE Healthcare, Germany). The running condition was as follows: 300 V for 90 minute, followed by 500 V for 90 min, 1000 V for 3 h and finally 3500 V for 16 h. The focused strips were equilibrated twice in equilibration solution. The first equilibration was performed in a solution containing 6 M urea, 20% (w/v) glycerol, 2% (w/v) SDS (Merck), 1% (w/v) DTT, and 50 mM Tris-HCl (Merck) buffer, pH 8.8. The second equilibration was performed in a solution with 2.5% (w/v) iodoacetamide (Merck). Separation in the second dimension was performed by SDS-PAGE in a vertical slab of acrylamide (Merck) (12% total monomer, with 2.6% cross-linker) using a PROTEAN II Multi Cell (BioRad). The protein spots in analytical and preparative gels were visualized by silver nitrate (Merck, Germany) and Coomassie brilliant blue CBB/ G-250 (Sigma, Germany) respectively (19-21-22).\n\nGel image Analysis\n------------------\n\nGS-800 densitometer (Bio-Rad) was used for scanning of silver stain gels. Gels were analyzed using the Melanie 6 software (GeneBio, Geneva, Switzerland). The molecular masses of protein on gels were determined by co electrophoresis of standard protein markers (GE Heathcare) and *pI* of the proteins were determined by migration of the protein spots on 18 cm IPG (pH 4--7, linear) strips. 2-DE per sample (each species) was run for three biologically independent replicates, percent volume of each spot was estimated and analyzed by one-way analysis of variance (ANOVA) SAS software, and means were compared by the LSD test at P \u2264 0.01. Spots were only considered to be significantly different in abundance at least between two *Leishmania* species when/at *P* \u2264 0.01.\n\nPeptide extraction and mass analysis\n------------------------------------\n\nThe protein spots of interest were excised from coomassie brilliant blue (CBB) stained gels and analyzed using an Amazon ion trab MS/MS (Bruker Daltonics) Mass spectrometer. Briefly, peptides were solubilized in 0.5 % formic acid and fractionated on a nano flow uHPLC system (Thermo RSLCnano) before online analysis by electrospray ionisation (ESI) mass spectrometry on an Amazon ion trap MS/MS (Bruker Daltonics). Peptide separation was performed on a Pepmap C18 reversed phase column (LC Packings), using a 5 -- 85% v/v acetonitrile gradient (in 0.5% v/v formic acid) run over 45 min. at a flow rate of 0.2 \u03bcl / min. Mass spectrometric (MS) analysis was performed using a continuous duty cycle of survey MS scan followed by up to ten MS/MS analyses of the most abundant peptides, choosing the most intense multiply charged ions with dynamic exclusion for 120s. MS data was processed using Data Analysis software (Bruker) and the automated Matrix Science Mascot Daemon server (v2.1.06) ([@B23]). Protein identifications were assigned using the Mascot search engine to interrogate in house databases of protein sequences for *L. major*.\n\nResults\n=======\n\nComparison of proteome patterns\n-------------------------------\n\nProtein extracts from the three *Leishmania* species including *L. tropica, L. major* and *L. infantum* were separated by 2-DE gel electrophoresis. Multiple gels from the three independent replications were run to ensure reproducibility of the protein homogenates on the 2-DE gels. The analytical gels were visualized by silver staining. [Fig. 1](#F1){ref-type=\"fig\"} show a representative example of the proteins separated/detected on a 2-DE gel in *L. tropica*, *L. major* and *L. infantum*, where a total weight of 120 \u03bcg of proteins had been applied. The gel images were analyzed by Melanie software, and the percent volume of the spots was estimated and compared across the gels. We succeeded in detecting 600 \u00b1 100 spots on the 2-DE gels, from which 638, 590, and 546 spots were statistically analyzed across the replicates in *L. tropica*, *L. major*, and *L. infantum* respectively. A number of 478 spots could be paired in all of the three species (Supplementary Table 1).\n\n![2-DE gels analysis of proteins extracted from different *Leishmania* species: *L. tropica*, *L. major*, and *L. infantum* . In first dimension (IEF), protein was loaded on a 18-cm IPG strip with a linear gradient of pH 4--7. In the second dimension, 12% SDS-PAGE gels were used, with a well for molecular weight standards. Proteins were visualized by silver staining. Arrows represent spots identified by LC Mass spectrometry. All spot numbers correspond to the spot numbers of [Table 1](#T1){ref-type=\"table\"}](IJPA-10-530-g001){#F1}\n\n###### Proteins identified in all three Iranian isolates of *Leishmania* species using LC/MS analysis\n\n  **Spot No. [^a^](#TFN1){ref-type=\"table-fn\"}**   **Protein name & biological process**                                                                                         **Access. No**   **(PI/MW) Exp. [^c^](#TFN3){ref-type=\"table-fn\"}**   **(PI/MW) Theo. [^d^](#TFN4){ref-type=\"table-fn\"}**   **% Volume**          **Score. [^e^](#TFN5){ref-type=\"table-fn\"}**   \n  ------------------------------------------------ ----------------------------------------------------------------------------------------------------------------------------- ---------------- ---------------------------------------------------- ----------------------------------------------------- -------------- ------ ---------------------------------------------- -----\n                                                   **Cell motility and cytoskeleton**                                                                                                                                                                                                                                                                                             \n  382                                              *beta tubulin*                                                                                                                A4HTR1           4.79/45                                              4.74/49                                               0.05           0.02   0.02                                           407\n  457                                              *hypothetical protein, conserved (Contains Myosin_tail_1 domain) 1 HECT (E6AP-type E3 ubiquitin-protein ligase) domain.)*     Q4QCC9           6.08/14                                              4.95/347                                              0.02           ND     ND                                             43\n  510                                              *ADF/Cofilin*                                                                                                                 A4I4A3           6.35/17                                              6.61/15                                               0.17           0.14   0.10                                           540\n                                                   **Cell signaling and vesicular trafficking**                                                                                                                                                                                                                                                                                   \n  309                                              *hypothetical protein, conserved (Contains Ras_like_GTPase superfamily)*                                                      A4I8E6           6.46/42                                              5.92/24                                               0.02           0.07   0.05                                           820\n  501                                              *calmodulin, putative*                                                                                                        A4HU13           4.46/15                                              4.10/16                                               0.05           0.03   0.15                                           773\n                                                   **Intracellular survival / proteolysis**                                                                                                                                                                                                                                                                                       \n  31                                               *calpain-like cysteine peptidase, putative,cysteine peptidase, Clan CA, family C2, putative*                                  E9AD27           6.14/15                                              5.14/700                                              0.06           0.12   0.07                                           83\n  35                                               *calpain-like cysteine peptidase, putative,cysteine peptidase, Clan CA, family C2, putative*                                  E9AD27           5.96/14                                              5.14/700                                              0.04           ND     ND                                             57\n  507                                              *calpain-like cysteine peptidase, putative,cysteine peptidase, Clan CA, family C2, putative*                                  E9AD27           6.54/14                                              5.14/700                                              ND             0.10   ND                                             42\n  774                                              *calpain-like cysteine peptidase, putative,cysteine peptidase, Clan CA, family C2, putative*                                  E9AD27           5.68/66                                              5.14/700                                              ND             0.09   ND                                             45\n  121                                              *hs1vu complex proteolytic subunit-like,hs1vu complex proteolytic subunit-like, threonine peptidase, Clan T(1), family T1B*   Q8I496           5.38/42                                              6.05/24                                               0.07           ND     ND                                             501\n                                                   **Lipid metabolism**                                                                                                                                                                                                                                                                                                           \n  70                                               *hypothetical protein, conserved (Contains Acyl transferase domain ;Acyl transferase domain. S-malonyltransferase* ***)***    E9AFH1           6.68/19                                              6.3/59                                                0.12           0.34   0.25                                           41\n                                                   **microtubule motor movement**                                                                                                                                                                                                                                                                                                 \n  53                                               *hypothetical protein, conserved (Contains Ankarin repeat domain)*                                                            A4IC52           5.42/16                                              5.63/38                                               0.52           0.48   0.28                                           587\n                                                   **Nucleotid metabolism**                                                                                                                                                                                                                                                                                                       \n  369                                              *hypothetical protein, conserved (Contains Nucleoside 2-deoxyribosyltransferase domain)*                                      E9AH20           5.1/17                                               4.95/16                                               0.03           0.04   0.30                                           351\n                                                   **oxidative stress defense**                                                                                                                                                                                                                                                                                                   \n  54                                               *tryparedoxin*                                                                                                                Q6RYT3           5.22/18                                              5.20/16                                               0.35           0.06   0.01                                           322\n                                                   **Protein synthesis**                                                                                                                                                                                                                                                                                                          \n  142                                              *translation elongation factor 1-beta, putative*                                                                              A4IDB2           4.88/51                                              4.61/23                                               0.16           0.01   0.06                                           742\n  503                                              *60S acidic ribosomal protein P2*                                                                                             E9AGM1           4.61/16                                              4.18/10                                               ND             0.15   0.18                                           408\n                                                   **Amino-acid biosynthesis**                                                                                                                                                                                                                                                                                                    \n  527                                              *pyrroline-5-carboxylate reductase*                                                                                           E9AGK4           6.32/43                                              6.21/28                                               0.12           0.03   0.14                                           529\n                                                   **Protein ubiquitination / proteolysis**                                                                                                                                                                                                                                                                                       \n  214                                              *hypothetical protein, conserved (Contains 1 HECT (E6AP-type E3 ubiquitin-protein ligase) domain.)*                           Q4Q4M9           5.88/19                                              6.12/454                                              0.03           0.51   ND                                             45\n                                                   **Transport**                                                                                                                                                                                                                                                                                                                  \n  438                                              *calcium channel protein, putative,ion transporter, putative*                                                                 Q4Q3G0           5.36/16                                              8.6/291                                               0.06           0.34   0.05                                           50\n                                                   **Stress related proteins/Protein folding**                                                                                                                                                                                                                                                                                    \n  6                                                *fk506-binding protein 1-like protein*                                                                                        Q4QBK4           5.02/14                                                                                                    0.23           0.35   ND                                             204\n  222                                              *hypothetical protein, conserved (Contains alphacrystallin-Hsps_p23-like Superfamily)*                                        A4IBY7           4.48/39                                              4.15/21                                               0.52           0.07   0.17                                           441\n                                                   **Unknown**                                                                                                                                                                                                                                                                                                                    \n  402                                              *hypothetical protein, conserved*                                                                                             Q4Q8Y7           4.52/54                                              6.27/95                                               0.02           0.06   0.02                                           45\n  419                                              *hypothetical protein, conserved*                                                                                             Q4Q2Z8           5.42/72                                              9.93/129                                              0.02           0.10   0.01                                           32\n  475                                              *hypothetical protein*                                                                                                        Q4QE59           4.87/26                                              10.88/58                                              0.12           0.01   ND                                             35\n\n**LC/MS analysis./**\n\nThe numbering corresponds to the 2-DE gel in [Figure 1](#F1){ref-type=\"fig\"}./\n\nAccession number in Swiss-Prot./\n\nExperimental p*I* and molecular weight./\n\nTheoretical p*I* and molecular weight./\n\nMascot score./ **ND**: Not Detected (Spot)\n\nHowever 34, 33, and 4 protein spots of *L. tropica*, *L. major*, and *L. infantum* were identified as *Leishmania* species-specific spots ([Fig. 2](#F2){ref-type=\"fig\"}). Altogether, 265 protein spots exhibited reproducible quantitative (*P* \u2264 0.01) changes across the three samples in *Leishmania* species (Supplementary Table 2). Among them, 35, 22 and 11 protein spots were different between the *L. tropica* and *L. major* (presented by **a**), *L. tropica and L. infantum* (presented by **b**), *L. major* and *L. infantum* (presented by **c**) respectively. Seven protein spots were different between all of the three species (presented by **abc**).\n\n![Venn diagram showing shared and unique protein expression between different *Leishmania* species: *L. tropica, L. major*, and *L. infantum*. Regions of overlap between circles indicate gene expression which paired in all the three species/two species. Regions that do not overlap between circles indicate unique gene expression of particular/each species as species-specific spots](IJPA-10-530-g002){#F2}\n\nProtein identification\n----------------------\n\nWithin differentially expressed proteins, which were detected on the analytical gels, we could reliably detect and excite a total number of 28 protein spots on CBB-stained preparative gels. Due to the lack of the protein amount, the remaining proteins could not be detected. The excised protein spots were then analyzed by LC/MS leading to the identification of 24 proteins ([Table 1](#T1){ref-type=\"table\"}). The identified protein spots by mass spectrometry have been labeled in [Fig. 1](#F1){ref-type=\"fig\"}. These proteins were classified in multiple categories according to their species, functions, and biological processes: Cell motility and cytoskeleton, Cell signaling and vesicular trafficking, Intracellular survival / Nucleotid metabolism, Protein synthesis, oxidative stress defense, microtubule motor movement proteolysis, Lipid metabolism, Amino-acid biosynthesis, Protein ubiquitination / proteolysis, Transport, Stress related proteins/Protein folding and hypothetical proteins (Unknown) ([Table 1](#T1){ref-type=\"table\"}). It is worth noting that some of these proteins are hypothetical and their functions in *Leishmania* still remain to be elucidated.\n\nThe clustering of protein expression pattern of differentially expressed proteins in *L. tropica, L. major* and *L. infantum* is presented in [Fig. 3](#F3){ref-type=\"fig\"}. All quantitative information is showed using a color scale in which the color ranges from green for the highest down-regulation to red for the highest up-regulation. Black color indicates no changes in expression pattern of the 3 *Leishmania* species. A comparison among the 3 species revealed that the changes in expression pattern were more pronounced in *L. infantum* compared with *L. tropica* and *L. major.* In addition, the number of up regulated proteins was higher than that of down-regulated proteins. The functional annotation of the 3 species identified proteins in *L. tropica, L. major* and *L. infantum* classified by biological function and processes described in [Table 1](#T1){ref-type=\"table\"}. The biological function pie charts of the cutaneous species (*L. tropica* and *L. major*) are highly similar whereas, the pie chart for the biological function of the visceral species (*L. infantum*) is different from those of cutaneous species ([Fig. 4](#F4){ref-type=\"fig\"}). Among the cutaneous species, most of the proteins functionalities are observed in the following categories: intracellular survival / nucleotid metabolism and cell motility/ cytoskeleton (20--22%). Moreover, the least functionalities are observed among cell signaling/ vesicular trafficking and lipid metabolism (5%). Moreover, in the *L. infantum* species, most functionality are observed among protein synthesis and cell motility/ cytoskeleton (15%) and least of them are observed among amino-acid biosynthesis and oxidative stress defense (7%).\n\n![Clustering pattern of differential proteins expressed in *L. tropica, L. major* and *L. infantum.* All quantitative information is showed using a color scale in which the color ranges from green for the highest down-regulation to red for the highest up-regulation. Black color indicates no changes in expression pattern of 3 *Leishmania* species](IJPA-10-530-g003){#F3}\n\n![Functional annotation of the identified proteins of 3 species *L. tropica, L. major* and *L. infantum* classified by biological function and processes in described in [Table 1](#T1){ref-type=\"table\"}](IJPA-10-530-g004){#F4}\n\nDiscussion\n==========\n\nThe aim of this preliminary study was to apply 2-DE, which is a valuable method in the proteomics arena to analyze the protein profile patterns of the three Iranian cutaneous and visceral *leishmania* species including *L. tropica, L. major* and *L. infantum* to search for species-specific *Leishmania* proteins. In recent years by completion of genome sequencing *Leishmania* parasites coupled with protein separation techniques analysis has given us an insight in better understanding the mechanisms of pathogenesis of *Leishmania* species. Proteomics approaches at the protein expression level provide additional information for further analysis with a biological function. Moreover, comparative proteomics has been successful in determining the virulence biomarkers ([@B24]).\n\nOverall, the proteome 2-DE maps of *L. tropica, L. major* and *L. infantum* were strikingly similar in terms of protein distribution and positioning. By using 18-cm IPG strips at the p*I*4--7 range in 2-DE comparative analysis, we successfully identified more than 700 protein spots for all the three species. These numbers represent about 9% of total proteins in *Leishmania* genome projects ([@B25]). The analysis and discussion about the detected proteins in this study could be categorized into three parts. First, we discuss comprehensively about the biological function of proteins whose expression abundances were common among each of the three species. In the second and third part, we represent the biological function of the proteins which are different and absent/present in the three different species. As it is described in [Table 1](#T1){ref-type=\"table\"} and Venn diagram ([Fig. 2](#F2){ref-type=\"fig\"}), in the first part we have studied a total number of 478 proteins among which we will discuss the common ones. Among the common proteins of three *Leishmania* species, which were surveyed, we would like to mention significant proteins such as beta tubulin and ADF/Coflin. These proteins could be categorized in the cell mobility and cytoskeleton group. ADF/cofilin is existent in all eukaryotic organisms and has been involved in cell motility and cytokinesis. *Leishmania* parasites express only one isoform of ADF/cofilin, which is essential for flagellar assembly and motility ([@B26]). Beta tubulin is known as one of the members of distinct microtubule networks in *Leishmania* and is implicated in locomotion, cell shape and division ([@B27]).\n\nWithin other common proteins we could name tryparedoxin, calpain-like cysteine peptidase, and calmodulin putative from the groups of oxidative stress defense, Intercellular survival/proteolysis, cell signalling, and vesicular trafficking, respectively. Tryparedoxins are special thiol disulfide oxidoreductases related to thioredoxins, which play a crucial role in hydroperoxide detoxification cascades of Kinetoplastida ([@B28]).\n\nOne of the cell signaling proteins was calmodulin-like preotein. Calmodulin is a kind of calcium binding protein. It participates in calcium signaling pathways that regulate multiple critical processes such as growth and proliferation ([@B29], [@B30]). The biological mechanism of elongation factor puts it in the protein synthesis group. Elongation factor1alpha plays a role in protein synthesis and assembly. It is a highly conserved GTP-binding protein involved in protein translation. In addition, it is an actin/microtubule-binding protein which interacts with the cytoskeleton ([@B31]--[@B33]). Another common protein could be calcium channel protein that is a member of the transporter group. A huge number of proteins exist under the group of hypothetical proteins whose biological mechanisms are still not well discovered.\n\nThe second group of proteins, which were studied, was the ones, which were different in all three *Leishmania* species. The differences in proteins mainly occur between *L. tropica* and *L. major* and the least differences occur between *L. major* and *L. infantum* ([Fig. 2](#F2){ref-type=\"fig\"}). It is an interesting point that the proteins, which are different between *L. tropica* and *L. infantum* are less than those of *L. major* and *L. tropica*. *L. tropica* is the main cause of dry cutaneous leishmaniasis lesions; whereas, *L. major* is the main cause of wet CL lesion and *L. infantum* causes visceral leishmaniasis. Recently there have been reports of viscerotropic forms caused by *L. tropica* in either humans or dogs infected to visceral leishmaniasis in Iran and different parts of the world ([@B34], [@B2]). Among significant proteins which are different in *L. tropica* and *L. major* we could mention calcium channel proteins, tryparedoxin and Elongation factor from the groups of oxidative stress defense and protein synthesis respectively whose mechanism have been described previously. Recently, it was demonstrated that *Leishmania* EF-1alpha acts as a virulence factor ([@B35]). This protein could diffuse into the cytosol of infected macrophages, where it is able to activate tyrosine phosphatase-1 leading to macrophage deactivation ([@B31]). Among the proteins different in *L. tropica* and *L. infantum* we could mention proteins such as calmadolin and trypardoxin in which the former role is to transfer the material into the cells and the latter role is defending the host cell against oxidative factors and preventing its death. Other proteins are hypothetical and have an unknown mechanism. Nevertheless, some of these proteins have a specific domain and have a different mechanism such as hypothetical protein, conserved contains nucleoside 2-deoxyribosyltransferase domain, that have a role in nucleotide metabolism.\n\nAmong the most significant proteins between *L. major* and *L. infantum* we could mention Calmadolin. The remaining proteins are partially hypothetical. Another group of proteins, found in this study, was too rare and introduced as absent/present.\n\nHowever, further studies are needed to define precise biological function for the mentioned proteins in the process of pathogenesis. Moreover, the examination of these species must be repeated with other strains in order to sanction the results. In addition, such differences must be evaluated and approved by other methods such as real time PCR and western blotting with monoclonal antibodies.\n\nConclusion\n==========\n\nThe analysis of proteome mapping of 3 *Leishmania* species including, *L. tropica, L. major* and *L. infantum* by 2-DE and mass spectrometry demonstrated that the vast majority of *Leishmania* proteins are commonly expressed among 3 species. Therefore, differentiation, virulence and pathogenesis may be related not only to the immunity situation of the host but also to the differentiation expression of a number of proteins like Stress related proteins/Protein folding and protein ubiquitination/proteolysis. It must be pointed out that further studies must be undertaken using western blotting or real time PCR in order to support the results of the current study.\n\nThe authors would like to express their gratitude to Prof. Ghasem Hosseini Salekdeh from Agricultural Biotechnology Research Institute of Iran, (ABRI), for their kind cooperation in proteomics laboratory. We are grateful to Elham Sarhadi from ABRI for valuable technical assistant. Also special thanks from Prof. Burchmore and Ms Suzzane Eideman from Institute of Infection, Immunity and Inflammation, College of Medical, Veterinary and Life Science, University of Glasgow, United Kingdom for Mass spectrometry analysis. In addition, we are grateful to Dr S.R. Naddaf from Pasteur Institute of Iran for valuable assistant. The authors would like to express their gratitude to Dr. B. Akhoundi and Mrs. S. Charedar for their kind cooperation. This study was financially supported by the Vice Chancellor for Research, School of Public Health, and Tehran University of Medical Sciences, Iran (NO.10695). The authors declare that there is no conflict of interests.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nApproximately 100 million of adult Americans suffer from chronic pain (CP), i.e., pain present daily or repeatedly for more than three months, with an estimated national economic cost of 560--635 billions annually \\[[@B1]\\]. In the United States (US), it is generally recognized that access to care for CP patients is particularly scarce \\[[@B2], [@B3]\\]. While waiting, patients often get discouraged and need to find some relief. Visiting the emergency department (ED) becomes their sole alternative \\[[@B4], [@B5]\\]. Utilization of the ED by patients seeking relief from CP has increased and is estimated to be up to 16% of all ED visits \\[[@B6], [@B7]\\]. In a fast-paced environment such as the ED, patients with CP are challenging to healthcare providers (HCP) \\[[@B5], [@B8]\\]. In fact, repercussions of CP such as depressive symptoms and pain interference with activities are not properly addressed \\[[@B5], [@B9]\\]. Although the ED is not the best clinical environment to assess and treat these patients, CP is more and more prevalent and HCP in acute care settings have to adjust to this new clinical reality and offer some support to these patients \\[[@B4], [@B5], [@B8]\\].\n\nThe Institute of Medicine identified CP as a care priority in the US and highlighted the need to delineate barriers and strategies to improve pain management practice \\[[@B1]\\]. One of the main barriers remains stigmatization of people living with CP by the community, including HCP \\[[@B10], [@B11]\\]. Unfortunately, HCP in the ED are no exception \\[[@B5], [@B8], [@B12]\\]. Several scales have been developed to assess knowledge, beliefs, and attitudes of HCP towards pain and its treatment \\[[@B13]--[@B20]\\], but they are specific to a particular expertise (e.g., diagnosis and prescription) or a pain condition (e.g., back pain) and seem to occult some of the common misbeliefs and myths towards CP and people living with CP \\[[@B21]\\]. As a preliminary step towards improved awareness and knowledge about chronic pain, a scale was then developed to measure knowledge, beliefs, and attitudes towards CP in the community \\[[@B21]\\]. This scale was meant to be more generic and to be used in the general population including HCP. The Chronic Pain Myth Scale (CPMS) was originally developed in French-Canadian. To the best of our knowledge, no scales are available to evaluate common societal misconceptions towards CP in English. Considering the increase in ED frequentations by people suffering from CP and the risks of undertreatment due to stigmatization in this clinical setting, this study thus aimed at determining to what extent a new English adaptation of the CPMS could provide valid scores among US emergency nurses. The internal consistency, construct validity, and internal structure of the translated scale were thus examined.\n\n2. Methods {#sec2}\n==========\n\n2.1. The French-Canadian CPMS {#sec2.1}\n-----------------------------\n\nThe original instrument includes 26 items and was found to measure three dimensions: knowledge, beliefs, and attitudes towards (1) people suffering from CP, (2) biopsychosocial impacts of CP, and (3) treatment of CP \\[[@B21]\\]. Items are scored on a 5-point Likert scale ranging from 1 (completely disagree) to 5 (completely agree). Negatively formulated items (2, 3, 4, 5, 7, 8, 9, 14, 20, 21, 24, 25, and 26) need to be reversed, so a higher score reflects better knowledge, attitudes, and beliefs. Total scores are calculated for each of the three subscales. Among French-speaking individuals from the general population, internal consistency of the CPMS was achieved for all three subscales (*\u03b1*\u2009=\u20090.72--0.82), in addition to its construct validity \\[[@B21]\\].\n\n2.2. Adaptation and Translation Process {#sec2.2}\n---------------------------------------\n\n### 2.2.1. Step 1: English Translation and Adaptation {#sec2.2.1}\n\nThe first step was the forward translation of all items, which was conducted by one of the bilingual investigators (AL) from French-Canadian to English. An adaptation of the translated version by a unilingual English-speaking nurse (MK) without specific expertise regarding chronic pain was done. Some suggestions were made such as using \"really hurt\" instead of \"really have pain\" or \"may worsen\" instead of \"may aggravate.\" Some other modifications were related to the context. For instance, the French-Canadian version included the description of winter sports such as snowmobile; it was then proposed to replace it by motorized sports to be more generic to all sorts of North American areas.\n\n### 2.2.2. Step 2: Synthesis Meeting {#sec2.2.2}\n\nA synthesis meeting was then held using a web-based screen-sharing application to reconcile discrepancies and agree on a preliminary English version.\n\n### 2.2.3. Step 3: Backward Translation {#sec2.2.3}\n\nThe third step consisted in the backward translation of the English items in French-Canadian by one of the bilingual investigators (GM) and without consulting the original version.\n\n### 2.2.4. Step 4: Review Committee {#sec2.2.4}\n\nThe fourth step consisted in the reviewing of the different versions by the two co-investigators (AL and GM). A chart reporting the original French-Canadian version, the translated and adapted English version, the backward French-Canadian translation, and comments from Steps 1, 2, and 3 were used. Consensus was reached after any discrepancies were resolved.\n\n### 2.2.5. Step 5: Pretest {#sec2.2.5}\n\nGuidelines recommend that the prefinal version of a questionnaire should be pretested among 30--40 people from the target setting \\[[@B22]\\]. Therefore, the prefinal translated version of the CPMS was pilot-tested in a convenience sample of 30 graduate nursing students. In July 2017, after obtaining ethical approval, the pilot study was advertised through the Florida State University College of Nursing\\'s website and through the graduate nursing students\\' mailing list, where a link to an online survey (Qualtrics\u00ae) was posted. Once they clicked on the link, they were taken to the first page of the survey where information was given about the study to ensure informed consent and where it was underlined that by taking the survey, they were giving their consent to participate. The instructions included to complete the questionnaire and to annotate any suggestions regarding the clarity. Each of the 26 items was accompanied by a check box (clear vs needs improvement) and a text box where they could include comments. No further modifications were necessary as only 2 participants made some minor comments. The final version of the English CPMS is presented in [Table 1](#tab1){ref-type=\"table\"}.\n\n2.3. Validation Study Design and Setting {#sec2.3}\n----------------------------------------\n\nThe validity of the English CPMS was evaluated through a web-based cross-sectional study. A voluntary, convenience sample was used. Inclusion criteria included nurses who have worked in an ED for at least 6\u2009months, who can read English, have access to the internet, and have the ability to fill out an electronic-based survey. A definition of CP as pain being present daily or repeatedly for more than 3\u2009months was provided at the beginning of the survey.\n\nThe recruitment process included a link to the survey on the Emergency Nurses Association\\'s (ENA) website that was located under the External Research Opportunities tab. No additional information on the study was provided on the ENA website. Facebook, a popular social media platform, was used to recruit participants from the different ENA\\'s state chapters. A participant information clip of 35\u2009seconds featuring a nurse (MK) explaining the focus of the study and eligibility criteria was posted. A link to the survey was provided at the end of the PIC. By clicking on the link, potential participants were taken to the introduction page where information about the study was provided to ensure informed consent. Before beginning the questionnaire, they were informed that by taking the survey they were agreeing to participate and giving their consent for the use of their responses. The study was approved through the Florida States University\\'s Institutional Review Board.\n\nData were collected using a Qualtrics\u00ae electronic survey which allows the direct transfer of data into an SPSS database. The survey was approximately 20\u2009minutes in length to complete and was available from October 1, 2017, to November 1, 2017. The following sociodemographic characteristics were collected at the beginning of the survey: age, sex, ethnicity, highest degree, clinical role, current setting of practice, years of experience, and suffering from chronic pain, i.e., pain present daily or repeatedly for more than 3\u2009months (y/n). Upon completion of the survey, participants were provided with the option to enter a drawing for one of the four 25\\$ Visa gift cards by providing an email address. Responses were limited to one IP address, and a screening for duplicates was done during the database cleaning process.\n\n2.4. Data Analysis {#sec2.4}\n------------------\n\nSociodemographic characteristics (i.e., age, sex, years of experience, clinical setting, and role) and CPMS scores were analyzed descriptively. Frequency tables, means, and standard deviations were used to summarize data for each subscale. Reliability was assessed via internal consistency of the three CPMS subscales reflected by unstandardized Cronbach\\'s alpha coefficients. These coefficients range between 0 (weak reliability) and 1 (perfect reliability) with a cutoff \\>0.7, usually reflecting adequate internal consistency/reliability \\[[@B23]\\]. Floor or ceiling effects were considered present if more than 15% of participants achieved the lowest or highest scores in each of the subscales \\[[@B24]\\]. Convergent construct validity can be determined by comparing scores between groups that are expected to score higher or lower on the scale (known-groups or extreme-groups technique) \\[[@B25]\\]. As it was done to validate the original French-Canadian instrument, the following subgroups were compared using a *T*-test: nurses suffering versus nurses not suffering from CP (i.e., CP\u2009\\>\u20093\u2009months). It was hypothesized that nurses who suffer from chronic pain would score higher in each subscale than nurses who do not. Regarding the internal structure, since the original questionnaire is at an early stage of development and this is the first iteration of the translated tool, confirmatory factor analysis seemed premature and restrictive \\[[@B26]\\]. Hence, exploratory factor analysis was performed as it would allow to identify items that might be problematic. The factors were extracted using principal axis factoring, and orthogonal varimax rotation was used in order to remain consistent with the previous study. According to the literature, loadings greater than 0.32 can be considered acceptable in a sample of at least 300 \\[[@B27]\\]. There are no general rules to determine the required sample size for validation studies. However, rules of thumb (e.g., 3--10 participants for each item contained in the scale to be validated) and absolute minimum sample size can be used for specific tests such as factor analysis \\[[@B25], [@B28]--[@B30]\\]. It was then determined that at least 260 participants (10 participants\u2009\u00d7\u200926 items) were required to validate the internal structure of the CPMS. Data were analyzed using IBM SPSS version 25.\n\n3. Results {#sec3}\n==========\n\n3.1. Participants\\' Characteristics {#sec3.1}\n-----------------------------------\n\nA total of 571 participants were recruited among whom 482 completed the CPMS. The sample comprised primarily females (90.9%), Caucasians (90.2%), and Bachelor\\'s prepared nurses (55.6%). Most nurses worked in an emergency department (92.3%). Sixty-seven percent of the sample reported that they were not suffering from chronic pain (i.e., pain present daily or repeatedly for more than 3\u2009months). Characteristics of the sample are presented in [Table 2](#tab2){ref-type=\"table\"}.\n\n3.2. Internal Consistency {#sec3.2}\n-------------------------\n\nAs shown in [Table 3](#tab3){ref-type=\"table\"}, Cronbach\\'s *\u03b1* coefficients reached the 0.7 cutoff, except for the second subscale (knowledge, beliefs, and attitudes towards biopsychosocial impact of CP). Moreover, internal consistency was maintained on the first subscale among nurses suffering from CP (*\u03b1*=0.88) and nurses not suffering from CP (*\u03b1*=0.88), the same was true for the third subscale (*\u03b1*=0.71 and *\u03b1*=0.70). For the second subscale, Cronbach\\'s alpha was calculated as 0.67 with *\u03b1*=0.64 for nurses suffering from CP and *\u03b1*=0.68 for nurses not suffering from CP. No floor or ceiling effects were found to be present in the data for any of the three subscales.\n\n3.3. Construct Validity of the English CPMS {#sec3.3}\n-------------------------------------------\n\nStatistically significant differences (*p* \\< 0.05) were found between subgroups for the three subscales. Nurses who suffer from CP showed better knowledge, beliefs, and more positive attitudes regarding people suffering from CP (*M*\u2009=\u200932.70, SD\u2009=\u20096.22) than nurses who did not suffer from CP (*M*\u2009=\u200930.36, SD\u2009=\u20095.12; *p* \\< 0.001). Additionally, nurses who suffer from CP had better knowledge, beliefs, and more positive attitudes regarding biopsychosocial impacts of CP (*M*\u2009=\u200940.10, SD\u2009=\u20093.92) than nurses who did not suffer from CP (*M*\u2009=\u200939.26, SD\u2009=\u20093.87; *p*=0.025). However, nurses who suffer from CP had lower knowledge, beliefs, and less positive attitudes regarding treatment of CP (*M*\u2009=\u200926.63, SD\u2009=\u20093.95) than nurses who did not suffer from CP (*M*\u2009=\u200928.02, SD\u2009=\u20093.19; *p* \\< 0.001).\n\n3.4. Internal Structure of the English CPMS {#sec3.4}\n-------------------------------------------\n\nThe Keiser--Meryer--Olkin measure of sampling adequacy (KMO\u2009=\u20090.84) and Bartlett\\'s test of sphericity (*p* value \\<0.001) both indicated that factor analysis was appropriate for our data \\[[@B27]\\]. Seven different factors had eigenvalues greater than Kaiser\\'s criterion of 1. However, based on the scree plot ([Figure 1](#fig1){ref-type=\"fig\"}), a reliable criterion for large samples \\[[@B31]\\], three factors were retained such as it was the case for the original French-Canadian CPMS. Three factors explaining up to 41% of the variance were then extracted. After rotation, the three factors explained 34.09% of the variance. As it was the case for the original CPMS, items 1 through 9 had the highest loadings for factor 1, items 10 through 19 had the highest loadings for factor 2, and items 20 through 26 had the highest loadings for factor 3 (see [Table 1](#tab1){ref-type=\"table\"} for rotated factor structure). The majority of items (92.3%) showed no crossloadings (items \\>0.32 on two or more factors \\[[@B32]\\]).\n\n4. Discussion {#sec4}\n=============\n\nThe CPMS is a generic tool to measure knowledge, attitudes, and beliefs towards people suffering from CP, the biopsychosocial impacts of CP, and its treatment that was never adapted in English nor tested specifically among nurses. This study aimed at examining preliminary validity of the English version of the CPMS according to some psychometric qualities in a sample of emergency nurses. Our study suggests that the English CPMS was reliable and valid among such a population.\n\nAcceptable reliability (*\u03b1* \\> 0.7) for research purposes in large populations \\[[@B23]\\] was reported for the first and third subscales. The second subscale\\'s reliability coefficient was below the cutoff (*\u03b1*=0.67) but is still considered adequate \\[[@B33]\\]. Moreover, whether nurses were suffering or not from CP did not interfere with reliability. After applying the known-groups technique for construct validity, better knowledge and more positive attitudes towards people suffering from CP and its biopsychosocial impacts were found among nurses suffering from CP (as expected). This finding converges with results from the validation studies of the French-Canadian CPMS \\[[@B21]\\] and other scales that compared pain and pain-free participants \\[[@B34]--[@B36]\\]. Interestingly, the only exception was for knowledge, beliefs, and attitudes regarding treatment of CP, where CP sufferers scored lower. This result was also found in the validation study of the French-Canadian CPMS in the general population \\[[@B21]\\]. A logical explanation could be related to the fact that people suffering from CP have trouble finding some relief as well as being believed regarding their pain experience. As a result, they feel discouraged which could lead to negative attitudes towards the treatment of CP. It is also noteworthy that nurses\\' scores on the first subscale (knowledge, beliefs, and attitudes towards people suffering from CP) were lower than scores in HCP in general (*M*\u2009=\u200931.10, SD\u2009=\u20095.64 vs *M*\u2009=\u200939.95, SD\u2009=\u20094.32) \\[[@B21]\\]. This observation corroborates the issue of stigmatization of CP patients from HCP \\[[@B10], [@B11]\\] and more specifically from ED nurses \\[[@B5], [@B8], [@B12]\\]. Lastly, similar internal structure was found supporting the 3-factorial nature of the original CPMS. Considering that four items of the second CPMS subscale presented factor loadings \\<0.32, their relevance could be questionable. However, removing these items from the CPMS did not change our conclusions regarding its internal consistency of construct validity. All 26 items were thus kept. Hence, the vast majority of items demonstrated robust loadings on their respective factors (\\>0.4 or \\>0.5 \\[[@B32]\\]), which provides preliminary support for the cross-cultural equivalence of the English CPMS \\[[@B37]\\].\n\n4.1. Strengths and Limitations {#sec4.1}\n------------------------------\n\nTo our knowledge, this is the first study to adapt and validate the use of the English CPMS. Before validating several psychometric properties, careful attention was given to the cross-cultural adaptation including a pretest with 30 participants \\[[@B22]\\]. Although our sample was mainly composed of Caucasian women, these demographics are representative of the nursing population in the US \\[[@B38]\\]. Moreover, the participants reported various levels of clinical experience and were from at least 20 different states across the US. Another strength of this study is the sample size. While no a priori sample size calculation was conducted, the sample was much larger than rules of thumb (*n*=78--260) \\[[@B25]\\].\n\nConsidering that psychometric properties are not fixed and can vary according to a given population \\[[@B25]\\], we cannot presume that these will be the same in a different sample and/or sample size. The English version of the CPMS should be validated in a heterogeneous sample from the general population, which could lead to confirmatory factor analysis and support of translation invariance \\[[@B37]\\]. Of note, the response rate could not be calculated, which keeps us from evaluating the risk of a nonresponse bias. Additionally, participants in the sample were self-selected which could lead to a participation bias.\n\n5. Conclusion {#sec5}\n=============\n\nGiven that the intent in developing the CPMS was to offer a \"one fits all\" tool, it is encouraging to observe similar psychometrics in terms of reliability, construct validity, and internal structure in a sample composed solely of HCP. The English CPMS is a promising tool to measure knowledge, beliefs, and attitudes towards CP. Its adaptation and preliminary validation in a sample of nurses is a first step towards the improvement of chronic pain management. Indeed, this tool could be used to measure educational needs, design awareness, and educational programs, as well as evaluating their outcomes.\n\nThe publication of this manuscript was made possible through funding by the Florida State University College of Nursing. The authors would like to thank the Emergency Nurses Association (ENA) for their help with the recruitment of participants.\n\nData Availability\n=================\n\nThe data used to support the findings of this study are available from the corresponding author upon request and approval from the Florida State University\\'s Institutional Review Board.\n\nConflicts of Interest\n=====================\n\nThe authors declare that they have no conflicts of interest.\n\n![Scree plot for the exploratory factor analysis. The scree plot indicates the retention of 3 factors.](PRM2019-1926987.001){#fig1}\n\n###### \n\nRotated factor matrix of the English CPMS (*n*=482).\n\n  Items                                                                                                                                                Factor 1    Factor 2    Factor 3\n  ---------------------------------------------------------------------------------------------------------------------------------------------------- ----------- ----------- -----------\n  \\(1\\) Really have pain, it is not in their head                                                                                                      **0.409**   0.263       0.056\n  \\(2\\) Just want to be prescribed drugs                                                                                                               **0.717**   0.081       0.008\n  \\(3\\) Try to obtain sick leave to stop working                                                                                                       **0.746**   0.006       \u22120.091\n  \\(4\\) Just want to be lazy and not accomplish their daily tasks                                                                                      **0.716**   0.203       0.064\n  \\(5\\) Complain of pain to get attention from others                                                                                                  **0.778**   0.089       \u22120.011\n  \\(6\\) Really want to get better                                                                                                                      **0.594**   0.236       \u22120.041\n  \\(7\\) Complain about their pain, but continue their activities (e.g., sports, motorized sports, and watercraft). Their pain should not be that bad   **0.543**   0.095       0.008\n  \\(8\\) Become dependent on their medications, like drug addicts                                                                                       **0.467**   0.028       \u22120.200\n  \\(9\\) Often tend to exaggerate the severity of their condition                                                                                       **0.745**   0.144       \u22120.077\n  \\(10\\) Chronic pain causes several physical symptoms (e.g., muscle tension, change in appetite, reduced mobility, and fatigue)                       0.220       **0.515**   0.122\n  \\(11\\) Chronic pain can have a direct impact on sex life                                                                                             0.262       **0.651**   0.025\n  \\(12\\) People with chronic pain are sometimes not accepted by their relatives                                                                        0.126       **0.494**   \u22120.026\n  \\(13\\) Chronic pain may be associated with negative emotions (e.g., fear, anger, or sadness)                                                         0.184       **0.591**   0.150\n  \\(14\\) People with chronic pain do not tend to isolate themselves                                                                                    0.208       **0.281**   0.072\n  \\(15\\) People with chronic pain usually have more difficulty resisting stressful events of daily life                                                \u22120.128      **0.274**   0.083\n  \\(16\\) The risk of death by suicide is higher among people with chronic pain than in the general population                                          0.144       **0.500**   \u22120.015\n  \\(17\\) Chronic pain costs billions of dollars to our society                                                                                         \u22120.247      **0.270**   0.209\n  \\(18\\) People with chronic pain do not always have access to healthcare services to treat their condition                                            0.190       **0.356**   \u22120.125\n  \\(19\\) Doctors lack time to treat chronic pain                                                                                                       \u22120.011      **0.265**   0.005\n  \\(20\\) Consulting a psychologist is useless unless the person with chronic pain is depressed                                                         \u22120.048      0.129       **0.443**\n  \\(21\\) There is not much to do to improve chronic pain                                                                                               0.113       0.060       **0.388**\n  \\(22\\) Good sleeping habits help reduce chronic pain                                                                                                 \u22120.185      0.286       **0.526**\n  \\(23\\) A balanced diet helps reduce chronic pain                                                                                                     \u22120.193      0.271       **0.616**\n  \\(24\\) Doing physical exercise may worsen chronic pain                                                                                               \u22120.003      \u22120.175      **0.700**\n  \\(25\\) Working may worsen chronic pain                                                                                                               \u22120.132      \u22120.255      **0.640**\n  \\(26\\) The treatment of chronic pain is in the hands of healthcare professionals and not those of the patient                                        0.040       0.010       **0.350**\n  Eigenvalues before rotation                                                                                                                          5.38        3.16        2.12\n  Percentage of variance explained after rotation                                                                                                      16.26       9.38        8.45\n\nExtraction method: principal axis factoring. Rotation method: varimax with Kaiser normalization. Bold type indicates factor loading for each item.\n\n###### \n\nSample description.\n\n  Variable                               \u2009                      *n* (482)   \\%\n  -------------------------------------- ---------------------- ----------- ------\n  Sex                                    Male                   43          8.9\n  Female                                 438                    91.1        \n                                                                            \n  Ethnicity                              Caucasian              435         91.0\n  African American                       10                     2.1         \n  Latino/Hispanic                        14                     2.9         \n  Asian                                  3                      0.6         \n  Native American                        4                      0.8         \n  Other                                  12                     2.5         \n                                                                            \n  Age                                    18--35                 207         43\n  36--55                                 219                    45.5        \n  56--75                                 55                     11.5        \n                                                                            \n  Highest level of education completed   Professional degree    133         27.7\n  Bachelor                               267                    55.6        \n  Masters                                50                     10.4        \n  Doctorate                              5                      1           \n  Other                                  25                     5.2         \n                                                                            \n  Clinical role                          Registered nurse       444         92.3\n  Clinical nurse specialist              4                      0.8         \n  Nurse practitioner                     20                     4.2         \n  Other                                  13                     2.7         \n                                                                            \n  Current clinical practice              Inpatient/acute care   17          3.5\n  Outpatient                             4                      0.8         \n  Primary care                           3                      0.6         \n  Emergency department                   445                    92.5        \n  Other                                  12                     2.5         \n                                                                            \n  Years of experience                    0--5                   164         34.1\n  6--10                                  117                    24.3        \n  11+                                    200                    41.6        \n                                                                            \n  Suffer from chronic pain               Yes                    158         32.8\n  No                                     323                    67.2        \n\n###### \n\nCronbach\\'s alpha coefficients and descriptive statistics of the English CPMS.\n\n  \u2009                 Subscale 1     Subscale 2     Subscale 3\n  ----------------- -------------- -------------- --------------\n  Cronbach\\'s *\u03b1*   0.86           0.67           0.72\n  Mean score\u2009\u00b1\u2009SD   31.10\u2009\u00b1\u20095.64   39.55\u2009\u00b1\u20093.91   27.53\u2009\u00b1\u20093.56\n  Median (range)    31 (9--45)     39 (23--50)    28 (15--35)\n  Possible scores   9--45          10--50         7--35\n\n[^1]: Academic Editor: Federica Galli\n"}
+{"text": "1. Introduction {#sec1-materials-10-00771}\n===============\n\nConcrete is one of the most important construction materials due to its well-recognized advantages, such as low cost, easy moldability and high strength under compression. However, unreinforced concrete exhibits a quasi-brittle behavior and, therefore, needs to be strengthened under tension-dominant loading conditions. As compared to conventional steel-reinforced concrete, fiber-reinforced concrete (FRC), characterized by high post-cracking ductility or even strain-hardening behavior accompanied by distributed cracks at a small crack width, can be employed to control the development of localized cracks particularly in local areas such as concrete cover and corner regions. Modern FRC has experienced a fast development since the 1960s, leading to a variety of FRC material designs with different fiber materials, geometries and performance (see, e.g., \\[[@B1-materials-10-00771]\\] for an overview). FRC is playing an increasingly important role in structural engineering as it partially or fully replaces traditional reinforced concrete. An example for a complete replacement of reinforcing steel is the design of segmented tunnel linings made of FRC \\[[@B2-materials-10-00771]\\].\n\nIn the last few decades, the material properties, as well as the structural performance of plain and fiber-reinforced concrete have been extensively investigated in laboratory environments. However, experiments are in general expensive and are limited to specific test configurations. Therefore, a variety of numerical models for concrete cracking, aiming at reliable prognoses of the fracture processes of concrete structures with or without reinforcement, have been proposed (see, e.g., \\[[@B3-materials-10-00771],[@B4-materials-10-00771],[@B5-materials-10-00771],[@B6-materials-10-00771],[@B7-materials-10-00771],[@B8-materials-10-00771]\\] for an overview). The majority of models for structural analyses of FRC are conceptually based on cracking models for plain concrete, modifying the post-peak regime of the constitutive law in terms of an increase of the residual stress and the fracture energy, so as to represent the enhanced ductility of FRC at a phenomenological level \\[[@B9-materials-10-00771],[@B10-materials-10-00771],[@B11-materials-10-00771],[@B12-materials-10-00771]\\]. To enable the analysis of the influence of specific fiber cocktails on the macroscopic material behavior of FRC, computational meso-scale models for FRC have been proposed, which include the explicit description of individual fibers within representative elementary volumes of FRC samples \\[[@B13-materials-10-00771],[@B14-materials-10-00771],[@B15-materials-10-00771]\\]. However, for computational analyses on a structural level, a multiscale-oriented approach allowing one to formulate the behavior of fiber and matrix and their mutual interactions at different length scales as proposed in \\[[@B16-materials-10-00771]\\] is required. Recently, the authors proposed a multilevel modeling framework, in which, at the lowest scale, the pullout behavior of different fiber types interacting with the concrete matrix at different inclination angles is considered by appropriate sub-models with the model information being appropriately transmitted across the scales depending on the specific fiber cocktail \\[[@B17-materials-10-00771]\\].\n\nAs an essential component of the modeling framework presented in \\[[@B17-materials-10-00771]\\], interface solid elements (ISE), i.e., degenerated solid finite elements with almost zero thickness proposed by Manzoli et al. \\[[@B18-materials-10-00771],[@B19-materials-10-00771],[@B20-materials-10-00771],[@B21-materials-10-00771]\\], have been adopted and successfully applied to the failure analysis of plain and fiber-reinforced concrete structures. As compared with classical zero-thickness interface elements, ISE can be easily implemented based on standard finite element codes by using solid finite elements for the bulk material and for the interfaces. Employing a continuum damage model to approximate the interface degradation, it allows one to describe the interface behavior completely in the continuum framework. Consequently, those specific variational formulations, discrete constitutive relations and integration rules to obtain the internal forces associated with classical interface elements are not required. The artificial initial stiffness that is normally required in zero-thickness interface elements is automatically included in the elastic stiffness of ISE \\[[@B18-materials-10-00771],[@B20-materials-10-00771],[@B21-materials-10-00771]\\]. It is recognized that the interface solid elements share similar features with zero-thickness interface elements. The most notable advantage of this class of models is the fact that no special procedure for the tracking of evolving cracks is necessary. This contributes to its robustness and allows for 3D fracture simulations characterized by complex fracture patterns (see, e.g., \\[[@B22-materials-10-00771]\\]). The crack pattern obtained via discrete representations along prescribed element edges evidently suffers from a certain dependence on the mesh topology. However, the influence on the overall macroscopic material response is tolerable if unstructured meshes with reasonable resolution are used \\[[@B20-materials-10-00771],[@B23-materials-10-00771]\\]. Furthermore, for analyses of heterogeneous materials on the mesoscale level, it was shown that the mesh-dependence of interface elements becomes less of a concern once the mesoscale heterogeneity is modeled \\[[@B24-materials-10-00771],[@B25-materials-10-00771],[@B26-materials-10-00771]\\]. This drawback can be alleviated, e.g., by continuously modifying the local finite element topology at the crack tip to enforce the alignment between the element edges and the crack propagation direction \\[[@B27-materials-10-00771],[@B28-materials-10-00771],[@B29-materials-10-00771]\\]. Alternatively, mesh refinement, at the cost of increased computational expense, can be applied to resolve the large elements along the crack path \\[[@B30-materials-10-00771],[@B31-materials-10-00771],[@B32-materials-10-00771]\\]. The increased computational demand resulting from the duplication of finite element nodes can be controlled by pre-defining the interface elements only in vulnerable regions or applying an adaptive algorithm for the mesh processing during computation \\[[@B33-materials-10-00771],[@B34-materials-10-00771],[@B35-materials-10-00771],[@B36-materials-10-00771],[@B37-materials-10-00771]\\].\n\nAs unstructured, high density finite element discretizations are required in general to reduce the influence of the mesh on the computed crack topology, we propose an adaptive crack model for plain and fiber-reinforced concrete structures, in which the mesh is progressively modified during the analysis as fracture proceeds. It is based on a previous implementation of the ISE technology, where the complete finite element mesh was preprocessed for the insertion of interface solid elements \\[[@B17-materials-10-00771]\\]. Solid elements are generated within interfacial gaps only in the vicinity of the crack path to allow for structural analysis with significantly reduced computational costs. The proposed numerical methods are designed to provide a platform for physically-sound and computationally-efficient and robust numerical analyses of concrete and (fiber or steel) reinforced concrete structures. However, the focus of this paper is specifically on the numerical modeling and analysis of fiber-reinforced materials and structures using adaptive strategies for inserting interface solid elements.\n\n2. Multilevel Model for Fiber-Reinforced Concrete {#sec2-materials-10-00771}\n=================================================\n\nA multilevel model for steel fiber-reinforced concrete materials and structures, which allows one to follow the influence of various design parameters through the different length scales \\[[@B17-materials-10-00771]\\], is briefly summarized in this section. The model consists of a series of model components associated with three different scales involved in the numerical analysis of FRC:Level 1: Modeling of the pullout behavior of single fibers;Level 2: Modeling of the crack bridging effect of fiber cocktails;Level 3: Structural simulation including the opening and propagation of cracks considering the fiber crack bridging effect.\n\n2.1. Level 1: Single Fiber Pullout Model {#sec2dot1-materials-10-00771}\n----------------------------------------\n\nAt the level of the interaction between individual fibers and the matrix (Level 1), the pullout behavior of a single fiber is controlled by interface conditions, the fiber shape and the fiber inclination with respect to a crack. An analytical model for single fiber pullout has been proposed \\[[@B38-materials-10-00771]\\]. This model is able to capture the major mechanisms involved during the pullout of a single steel fiber embedded in a concrete matrix, accounting for different configurations of fiber type and strength, concrete strength and fiber inclination, as demonstrated in [Figure 1](#materials-10-00771-f001){ref-type=\"fig\"}.\n\n2.2. Level 2: Crack Bridging Model {#sec2dot2-materials-10-00771}\n----------------------------------\n\nAt the level of an opening crack within the fiber-concrete composite (Level 2), using the analytical model for single fiber pullout, the bridging effect, i.e., the traction (*t*) vs. the separation (*w*) law, is obtained via the integration of the pullout response of all fibers intercepting the crack.\n\nA unit area of an opening crack and the corresponding volume element of the composite material, containing a number of distributed fibers crossing the crack, is considered ([Figure 2](#materials-10-00771-f002){ref-type=\"fig\"}b). The pullout response of each fiber is dependent on the position $\\widetilde{x}$ and the inclination *\u03b8* of the fiber with respect to the crack plane ([Figure 2](#materials-10-00771-f002){ref-type=\"fig\"}a). As the crack opens (*w* increases), the integration of the individual fiber pullout forces with respect to the position and inclination results in the fiber bridging stress \\[[@B39-materials-10-00771]\\]:$$t^{fib}\\left( w \\right) = \\frac{c_{f}}{A_{f}}\\int_{\\widetilde{x} = 0}^{L_{f}/2}\\left\\lbrack {\\int_{\\theta = 0}^{{arccos}(2\\widetilde{x}/L_{f})}F\\left( \\widetilde{x},\\theta,w \\right)\\, p\\left( \\theta \\right)\\, d\\theta} \\right\\rbrack p\\left( \\widetilde{x} \\right)\\, d\\widetilde{x}.$$\n\nIn Equation ([1](#FD1-materials-10-00771){ref-type=\"disp-formula\"}), $c_{f}$ is the volume fraction of the fibers, and $A_{f}$ is the cross-section area of one fiber. The single fiber pullout force $F\\left( \\widetilde{x},\\mspace{600mu}\\theta,\\mspace{600mu} w \\right)$ is dependent on $\\widetilde{x}$ and *\u03b8*. The spatial dispersion characteristics $p\\left( \\theta,\\mspace{600mu}\\widetilde{x} \\right)$ of the fiber cocktail in the composite are represented by the probability density *p* as a function of the inclination angle *\u03b8* and the position $\\widetilde{x}$ of the fiber. While performing the integration in the present model, the following parameters influencing the evaluation of $p\\left( \\theta \\right)$ and $p\\left( \\widetilde{x} \\right)$ and, consequently, the bridging effect are particularly taken into account \\[[@B17-materials-10-00771]\\]:Fiber orientation: Unlike the usual assumption of the isotropic spatial orientation of fibers (see, e.g., \\[[@B39-materials-10-00771]\\]), an anisotropic fiber orientation as a general result of the casting process, graphically represented as an ellipsoid, is assumed to represent the spatial preference of the fiber cocktail in the global coordinate system. From this ellipsoidal fiber orientation, in association with a given (or potential) crack plane, the probability density $p\\left( \\theta \\right)$ can be obtained by computing the differential volume of the ellipsoid corresponding to a specific value of *\u03b8*.Boundary effect: In the vicinity of boundaries, the fibers tend to orient parallel to the boundary surfaces \\[[@B40-materials-10-00771]\\]. This \"boundary effect\", dependent on the fiber length and the dimension of the mold, is considered by means of \"scanning\" the potential crack plane and excluding the impossible fiber orientations according to the distance to the boundary. As a result, an average $\\overline{p}\\left( \\theta \\right)$ relation is obtained and used in Equation ([1](#FD1-materials-10-00771){ref-type=\"disp-formula\"}) for the calculation of the crack bridging tractions.The fiber distribution $p\\left( \\widetilde{x} \\right)$ with respect to the position is assumed to be homogeneous \\[[@B39-materials-10-00771]\\].Group effect: An additional aspect to be considered is the experimental observations that only 50--90% of hooked-end steel fibers are active due to the \"group effect\" of fibers in the composite material. In the present model, the activity factor is generally assumed to be $70\\%$.\n\nThe bridging effect for an opening crack is obtained as the total contribution of the cohesive traction of concrete and the contribution of the fiber cocktail. As the determination of $t^{fib}\\left( w \\right)$ according to Equation ([1](#FD1-materials-10-00771){ref-type=\"disp-formula\"}) requires a separate computational evaluation, which would slow down the finite element analysis if directly incorporated, the numerically-obtained results for the $t - w$ relationship are replaced by an analytical surrogate function form to be employed in the finite element structural model:$$t\\left( w \\right) = \\left( f_{t}^{*} - t_{1} \\right)\\,\\exp\\left( {- \\frac{w}{w_{ref}}} \\right) + t_{1}\\frac{w_{u} - w}{w_{u}} + t_{2}w\\,\\exp\\left( c_{1} - c_{2}w \\right),$$ where $t_{1}$, $t_{2}$, $c_{1}$ and $c_{2}$ are coefficients determined from a least-square curve fitting to the original numerical evaluation of Equation ([1](#FD1-materials-10-00771){ref-type=\"disp-formula\"}); $w_{ref}$ is a reference value of crack opening; $w_{u} = L_{f}/2$ represents the ultimate crack opening; and $f_{t}^{*}$ is the tensile strength of the FRC composite. This function form can be easily used to represent traction-separation laws for different fiber-concrete composites ([Figure 2](#materials-10-00771-f002){ref-type=\"fig\"}c). [Figure 3](#materials-10-00771-f003){ref-type=\"fig\"} contains the validation examples demonstrating the quality of the model in describing the bridging effect of fibers across the horizontal crack in a notched FRC prism cast in a \"standing\" mold \\[[@B41-materials-10-00771]\\] and the vertical crack in a \"dog bone\" specimen cast in a \"lying\" mold \\[[@B42-materials-10-00771]\\].\n\n2.3. Level 3: Failure Analysis of FRC Structures {#sec2dot3-materials-10-00771}\n------------------------------------------------\n\nLevel 3 of the model is concerned with a finite element implementation employing \"interface solid elements (ISEs)\" for simulating propagating cracks. Instead of the classical zero-thickness interface elements, degenerated 2D and 3D solid elements with almost zero thickness as proposed by Manzoli et al. \\[[@B18-materials-10-00771]\\] and Sanchez et al. \\[[@B20-materials-10-00771]\\] are used. Through a pre-processing algorithm, the original mesh is entirely fragmented, and interface solid elements are generated and placed along all element edges as illustrated in [Figure 4](#materials-10-00771-f004){ref-type=\"fig\"}a. Using degenerated solid elements to represent cracks, the strain in these interface elements is assumed to be exclusively related to the (regularized) unbounded term $\\hat{\\mathbf{\\varepsilon}}$ according to the continuum strong discontinuity concept \\[[@B18-materials-10-00771],[@B20-materials-10-00771]\\]:$$\\mathbf{\\varepsilon} \\approx \\hat{\\mathbf{\\varepsilon}} = \\frac{1}{h}{(\u301a\\mathbf{u}\u301b \\otimes \\mathbf{n}})^{s}.$$\n\nIn 3D, the strain tensor in an ISE can be expressed in the local coordinate system (*N*-*R*-*S*) as:$$\\mathbf{\\varepsilon} \\approx \\frac{1}{h}\\begin{bmatrix}\n{\u301au\u301b_{n}} & {\u301au\u301b_{r}/2} & {\u301au\u301b_{s}/2} \\\\\n{\u301au\u301b_{r}/2} & 0 & 0 \\\\\n{\u301au\u301b_{s}/2} & 0 & 0 \\\\\n\\end{bmatrix},$$ with $\u301au\u301b_{n}$, $\u301au\u301b_{r}$ and $\u301au\u301b_{s}$ as the normal and tangential components of the displacement jump $\u301a\\mathbf{u}\u301b$, respectively, which are determined from the relative displacement of the apex node with respect to its projection on the base ([Figure 4](#materials-10-00771-f004){ref-type=\"fig\"}b). Note that in the present work, only constant strain elements, i.e., three-node-triangular (in 2D analyses) and four-node-tetrahedral (in 3D analyses), are employed both for bulk and interface elements.\n\nAll bulk elements are considered to be linear elastic. The constitutive behavior of the degenerated solid elements is cast in a continuum form equipped with a damage law, which allows one to approximate the behavior of interfacial degradation mechanisms involved during the cracking in FRC materials:$$\\mathbf{\\sigma} = \\left( 1 - d \\right)\\mathbb{C}_{e}:\\mathbf{\\varepsilon} \\approx \\left( 1 - d \\right)\\frac{1}{h}\\mathbb{C}_{e}:{(\u301a\\mathbf{u}\u301b \\otimes \\mathbf{n}})^{s},$$ with *d* as the scalar damage variable and $\\mathbb{C}_{e}$ denoting the elastic stiffness tensor with Poisson's ratio $\\nu = 0$. The loading criterion is defined in terms of the equivalent stress $\\widetilde{\\sigma}$ and the displacement-like internal parameter *\u03b1*:$$f\\left( \\mathbf{\\sigma},\\alpha \\right) = \\widetilde{\\sigma} - t\\left( \\alpha \\right) \\leq 0.$$\n\nHere, $\\widetilde{\\sigma} = \\sqrt{\\sigma_{n}^{2} + \\left( \\sigma_{r}/\\beta \\right)^{2} + \\left( \\sigma_{s}/\\beta \\right)^{2}}$ is considered, with the factor $\\beta$ representing a weighting of normal and shear contributions to account for mixed mode fracture; in the present work, $\\beta = 2.0$ is used \\[[@B33-materials-10-00771],[@B43-materials-10-00771],[@B44-materials-10-00771]\\]. The softening behavior of interface $t\\left( \\alpha \\right)$ is controlled by the crack bridging law Equation ([2](#FD2-materials-10-00771){ref-type=\"disp-formula\"}) presented in the previous section, replacing the crack width *w* with the \"internal parameter\" $\\alpha$.\n\nThe internal parameter *\u03b1* is defined as the maximum value of equivalent separation that the interface experiences during the loading history:$$\\alpha = \\max\\left( \\widetilde{u} \\right) - {\\widetilde{u}}_{0}.$$\n\nHere, the equivalent crack separation is defined as:$$\\widetilde{u} = \\sqrt{\u301au\u301b_{n}^{2} + \\left( \\frac{\u301au\u301b_{r}}{2\\beta} \\right)^{2} + \\left( \\frac{\u301au\u301b_{s}}{2\\beta} \\right)^{2}},$$ and ${\\widetilde{u}}_{0}$ corresponds to the elastic limit state of the interface:$${\\widetilde{u}}_{0} = \\frac{f_{t}^{*}}{K_{0}} = \\frac{hf_{t}^{*}}{E^{*}} \\approx 0,$$ with $K_{0}$ representing the \"rigid\" elastic stiffness of the equivalent interface behavior ([Figure 5](#materials-10-00771-f005){ref-type=\"fig\"}).\n\nThe scalar damage variable $d\\left( \\alpha \\right)$ is obtained by comparing the secant stiffness $K^{\\sec}$ with the elastic stiffness of the equivalent interface behavior ([Figure 5](#materials-10-00771-f005){ref-type=\"fig\"}c) as:$$d\\left( \\alpha \\right) = 1 - \\frac{K^{\\sec}}{K_{0}} = 1 - \\frac{h\\, t}{E^{*}\\left( \\alpha + {\\widetilde{u}}_{0} \\right)}.$$\n\nIt is noticed that the post-cracking behavior of the FRC material is highly nonlinear; such nonlinearity frequently results in numerical difficulties while performing structural simulations. In the present work, the IMPL-EXintegration scheme \\[[@B45-materials-10-00771]\\] is implemented in the context of the interface solid element for FRC. Consequently, due to the explicit nature of damage models, the computation does not require any iteration, neither on the structural level, nor on the constitutive level. This ensures the robustness and efficiency of the computational model in failure analyses of FRC structures even in the case of complex crack configurations.\n\n[Figure 6](#materials-10-00771-f006){ref-type=\"fig\"} demonstrates the model performance while analyzing the behavior of a notched FRC beam subjected to three-point bending. It is observed that the proposed multilevel model captures the fracturing processes in FRC and the load-bearing responses of the structure both qualitatively and quantitatively very well (see Zhan and Meschke \\[[@B17-materials-10-00771]\\] for more details).\n\nIt is worth mentioning that the present work focuses on monotonic loading situations. In the case of cyclic loading, the un- and re-loading branch follow a secant in the traction-separation curve (see [Figure 5](#materials-10-00771-f005){ref-type=\"fig\"}c). In fact, however, residual deformations due to imperfect closing of the rough crack faces after unloading are observed already in cyclic loading of plain concrete. To appropriately represent this mechanism, it will be necessary to consider a damage-plasticity model, e.g., in a similar fashion as was proposed in \\[[@B46-materials-10-00771],[@B47-materials-10-00771]\\], where a smooth transition from open to fully-closed cracks is considered. The irreversible deformations associated with aggregate interlocking and fiber resistance at the crack surfaces are even more pronounced in the case of FRC (e.g., \\[[@B48-materials-10-00771]\\]). However, this issue is beyond the scope of the present paper.\n\n3. Mesh-Processing Techniques {#sec3-materials-10-00771}\n=============================\n\n3.1. Full Insertion of Interface Solid Elements via Preprocessing {#sec3dot1-materials-10-00771}\n-----------------------------------------------------------------\n\nUsing interface solid elements to capture the cracking phenomena in concrete structures, it is necessary to include in the numerical implementation a code block for placing degenerated solid elements into the interfacial gaps prior to the initiation of cracks. For that purpose, a preprocessing algorithm must be executed before applying the loading conditions \\[[@B18-materials-10-00771],[@B20-materials-10-00771]\\].\n\nAs illustrated in [Figure 7](#materials-10-00771-f007){ref-type=\"fig\"}, in a first step, the topology of the original finite element mesh is investigated. Based on the topological information (e.g., the neighboring bulk elements, the common edges and the node connectivity), the original mesh is fragmented. All common edges are duplicated; every bulk element shrinks via offsetting the edges by $h/2$; a \"phantom mesh\" is obtained ([Figure 7](#materials-10-00771-f007){ref-type=\"fig\"}b). Afterwards, every gap between the neighboring bulk elements is filled with solid elements characterized by a very high aspect ratio ([Figure 7](#materials-10-00771-f007){ref-type=\"fig\"}c), resulting in the actual discretization of the domain used for the computation. It is worth emphasizing that three-node-triangular (in 2D analyses) and four-node-tetrahedral (in 3D) are used both for bulk and interface elements.\n\nIn analogy to the classical zero-thickness interface elements, interface solid elements are well suited for capturing complex cracking phenomena (such as crack initiation and opening, crack propagation, kinking or even branching) in quasi-brittle materials without applying any crack path tracking technique \\[[@B19-materials-10-00771]\\]. However, it is observed from a number of structural simulations that the problem sizes (i.e., the system degrees of freedom (DOFs)) of the fully-processed mesh can reach approximately six-times (in 2D) and 20-times (in 3D analysis) the number of DOFs as compared to the original mesh. In order to limit the computational costs arising when using interface elements, different means of numerical implementation, such as performing the mesh fragmentation only in the vulnerable regions \\[[@B37-materials-10-00771]\\] and applying an adaptive algorithm for the insertion of interface elements \\[[@B35-materials-10-00771]\\], are proposed.\n\n3.2. Adaptive Insertion of Interface Solid Elements {#sec3dot2-materials-10-00771}\n---------------------------------------------------\n\nIn the present work, a specific algorithm has been developed and implemented for the adaptive insertion of ISEs to consider potential cracks during the structural failure analysis. The general concept of the proposed algorithm is depicted in [Figure 8](#materials-10-00771-f008){ref-type=\"fig\"}. The adaptive insertion procedure consists of the following major steps:Prior to the application of loading, use the original discretization to generate the phantom mesh.Start the structural simulation with the original mesh.In every load increment (except the first):(a)Modify the mesh by splitting the critical interfaces which were recorded at the end of the previous increment; fill the interfacial gaps with degenerated solid elements.(b)Solve the structural equation system based on the modified mesh.(c)According to the new solution, inspect the stress state in the bulk elements to identify the critical interfaces.Proceed to the next increment.\n\nIt is worth pointing out that similar effort has been made for the classical zero-thickness interface elements \\[[@B35-materials-10-00771]\\]. Nevertheless, the adaptive technique has not yet been considered for the recently-proposed continuum interface elements. The details of the adaptation process are described in the following.\n\n### 3.2.1. Generation of the Phantom Mesh {#sec3dot2dot1-materials-10-00771}\n\nAs illustrated in [Figure 7](#materials-10-00771-f007){ref-type=\"fig\"}, the \"phantom mesh\" is generated before starting the structural analysis. This process is identical for both full and adaptive insertion options. The difference lies in the subsequent steps: For the full insertion, the phantom mesh is directly used as the basis, and subsequently, solid elements are generated by correctly connecting the nodes on the two sides of the interface. However, in the adaptive algorithm, the phantom mesh provides a collection of topological information that is only partially used during simulation. As depicted in [Figure 9](#materials-10-00771-f009){ref-type=\"fig\"}a for a 2D configuration, every \"joint\" (such as joint-m or joint-n) corresponds to a node in the original mesh, and an \"interface\" (e.g., interface-i) connects two joints. In the phantom mesh ([Figure 9](#materials-10-00771-f009){ref-type=\"fig\"}b), every joint controls a cluster of phantom nodes. During the loading procedure, whenever a joint is activated, the original node (e.g., node-n) is replaced by the corresponding cluster of phantom nodes ([Figure 9](#materials-10-00771-f009){ref-type=\"fig\"}c).\n\n### 3.2.2. Determination of \"Critical\" Interfaces {#sec3dot2dot2-materials-10-00771}\n\nThe goal of the algorithm is to generate as few interface solid elements as possible; therefore, only the interfaces subjected to a stress level that is close to the material strength are considered as \"critical\". More specifically, the traction on an interface can be calculated from the stresses (***\u03c3***) in any of the two bulk elements:$$\\mathbf{t} = \\mathbf{\\sigma} \\cdot \\mathbf{n}.$$\n\nConsidering an interface under tension (indicated by a positive normal component of the interface stress, i.e., $t_{n} > 0$), if the magnitude of the traction vector exceeds a given threshold, this interface is denoted as critical:$$\\left| \\mathbf{t} \\middle| > \\right.t^{thr},{with} t^{thr} = \\gamma\\, f_{t}.$$ It should be noted that also other criteria based on, e.g., linear fracture mechanics, are alternatively used in the literature \\[[@B33-materials-10-00771]\\]. The factor $0 \\leq \\gamma \\leq 1$ controls the level of threshold and is generally set as 0.8, according to the authors' experience. Setting a small value of $\\gamma$, too many unnecessary interfaces may be identified as critical and too many ISEs may be generated. On the other hand, a large value ($\\left. \\gamma\\rightarrow 1.0 \\right.$) may lead to the over-stressing of interfaces, since the ISE should be prepared prior to the crack initiation. With the tolerance, the value of 0.8 includes a banded region along the crack path. Consequently, the influence of the insufficient accuracy of the resolution of the stress field at the crack tip inherently obtained from using linear finite elements is eliminated.\n\n### 3.2.3. Mesh Adaptation: Splitting of Interfaces and the Generation of IS Elements {#sec3dot2dot3-materials-10-00771}\n\nIn order to prepare the discretized model to consider potential cracks during the analysis, the critical interfaces identified according to the criterion Equation ([12](#FD12-materials-10-00771){ref-type=\"disp-formula\"}) are fully separated. Since (in 2D) an interface controls two joints, the clusters of phantom nodes corresponding to both joints are activated and used. For the adjacent bulk elements, the node connectivity is updated by using the activated phantom nodes. Two ISEs filling the interfacial gap are generated based on the phantom nodes. In the meantime, the affected interfaces adjacent to this critical interface are partially split and filled with one ISE ([Figure 8](#materials-10-00771-f008){ref-type=\"fig\"}).\n\nThe algorithm for the insertion of ISE in a specific interface in a 2D configuration is depicted in [Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}:Initially, the interface connecting Node-1 and Node-3 is shared by two bulk elements, i.e., Element-T and Element-B. (For convenience, Element-T refers to the bulk element located on the positive side of the interface, and Element-B is on the other side; see [Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}a).As shown in [Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}b, when Joint-1 is split, the corresponding cluster of phantom nodes (including Node-1T and Node-1B) is activated. The connectivity of bulk Element-T and Element-B is updated by replacing the current Node-1 with Node-1T and Node-1B, respectively. Meanwhile, the first interface solid element (ISE-I) is created by using Node-1T and Node-1B.When Node-3 is split, activate the phantom nodes (including Node-3T and Node-3B). The nodes of bulk elements are renewed in the same way; one of the nodes of the existing ISE-I is updated, as well (by using the new Node-3B instead of the current Node-3). In addition, the second ISE (ISE-II) is generated.\n\nNote that the procedure described above is not limited to the processing of a critical interface within one load increment. In general, there are three possible situations for the opening of a 2D interface; the other two cases are as follows: (1) An interface is partially open due to the activation of an adjacent interface; this process can be graphically represented by [Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}a,b. (2) After a certain number of load increments, a partially split interface ([Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}b) becomes fully separated ([Figure 10](#materials-10-00771-f010){ref-type=\"fig\"}c), because the interface itself or a neighboring interface reaches the critical state (as will be illustrated in the numerical examples contained in [Section 4](#sec4-materials-10-00771){ref-type=\"sec\"}).\n\nIn [Figure 11](#materials-10-00771-f011){ref-type=\"fig\"}, the insertion algorithm at a 3D interface is illustrated:Initially, the interface using Nodes-1, 2 and 3 is shared by two bulk elements (Element-T and Element-B).When Node-1 is separated, activate the corresponding cluster of phantom nodes, including Node-1T and Node-1B. Update the nodes of bulk Element-T and Element-B. Create the first interface solid element (ISE-I).When Node-2 is split, activate the phantom nodes (2T and 2B). Update the nodes of both bulk elements, as well as the existing ISE-I. Generate the second ISE (ISE-II).Node-3 is activated. Similarly, the new Node-3T and Node-3B are used; the nodes of the bulk elements and the existing two ISEs (I and II) are renewed. Finally, ISE-III is created.\n\nIn analogy to the 2D situation, the scheme described above is valid for any interface in the mesh; depending on the number (equal to 0, 1 or 2) of already activated joints and the number of nodes that should be separated (1, 2 or 3), there are in total six cases possible in 3D.\n\nFinally, it is worth emphasizing that the systematic description and maintenance of the topological and geometrical data structure of the finite element model play an important role in the implementation of an adaptive remeshing algorithm (see also \\[[@B35-materials-10-00771],[@B36-materials-10-00771],[@B49-materials-10-00771]\\]). The major difference between the proposed algorithm and the existing adaptation techniques for classical interface elements is that solid elements are created in the interfacial gaps, which makes it necessary to update the nodes of the existing bulk elements and, in particular, of the interface solid elements.\n\n4. Numerical Examples {#sec4-materials-10-00771}\n=====================\n\nIn this section, the performance of the proposed adaptive insertion technique for interface solid elements is demonstrated through selected numerical simulation of the crack propagation in plain and fiber-reinforced concrete structures both in 2D and 3D configurations.\n\n4.1. 2D Verification Examples {#sec4dot1-materials-10-00771}\n-----------------------------\n\nThe first examples are concerned with the 2D numerical simulation of plain concrete structures, conducted mainly for illustration and verification purposes.\n\n### 4.1.1. Illustration of Mesh Adaption Process {#sec4dot1dot1-materials-10-00771}\n\nThe \"L-shaped\" panel test is documented in \\[[@B50-materials-10-00771]\\] and has been selected by a few researchers to demonstrate the quality of concrete cracking models (e.g., \\[[@B51-materials-10-00771],[@B52-materials-10-00771]\\]). The 100-mm-thick panel, made of plain concrete, is fixed at the bottom and subjected to a concentrated load, which eventually causes the failure of the panel, as shown in [Figure 12](#materials-10-00771-f012){ref-type=\"fig\"}a. During the test, the macroscopic crack initiates at the inner corner as a result of the concentrated tensile stresses and propagates first with a small angle with respect to the horizontal direction and proceeds almost horizontally through the panel. This test is re-analyzed via the proposed computational model using interface solid elements in a 2D configuration. The material parameters used for plain concrete are: the Young's modulus $E_{c} =$ 25,850 MPa, the tensile strength $f_{t} =$ 2.7 MPa and the tensile fracture energy $G_{f} =$ 0.09 MPa\u00b7mm \\[[@B51-materials-10-00771]\\]. For illustration purposes, a coarse and homogeneous FE-discretization is considered for the original mesh containing 658 constant-strain-triangular elements ([Figure 12](#materials-10-00771-f012){ref-type=\"fig\"}b); in addition, a large value of the width of interfacial gap $h = 4$ mm is used while separating the edges. Note that, in principle, *h* should be small, so that the model topology is not modified excessively by interfacial gaps and holes after mesh processing. On the other hand, as experience shows, a too small value of *h* may lead to numerical problems. In the present work, *h* is generally set to approximately 1/1000 of the bulk element size, based on the studies and suggestions in \\[[@B18-materials-10-00771],[@B53-materials-10-00771]\\].\n\nAs can be observed in [Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}, as the load increases, the concentrated tensile stresses at the inner corner cause the tensile stress (traction) at the respective interfaces to exceed the threshold (Equation ([12](#FD12-materials-10-00771){ref-type=\"disp-formula\"})). The interfaces are activated and split according to the sequence given in [Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}a:With the activation (full splitting) of Interface-1, two nodes belonging to this edge are disintegrated; the corresponding two clusters of phantom nodes are activated, as indicated by the red dashed circles in [Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}b. In the interfacial gap, two ISEs (triangular elements in red color) are created, while in each of the adjacent interfaces, one ISE (triangular element in blue) is inserted.Interface-2 is activated, where similarly, two ISEs are placed in the gap and several ISEs are generated in the vicinity ([Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}c).The activation of Interface-3 leads to the splitting of only one node because the other node is already activated. At Interface-3, in addition to the existing ISE generated when Interface-1 was activated, the second ISE is created ([Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}d).After the split of Edge-4 ([Figure 13](#materials-10-00771-f013){ref-type=\"fig\"}e), Edge-5 and Edge-6 are activated almost simultaneously.\n\nNote that in the mesh adaptation process described above, every single interface follows the scheme introduced in [Section 3.2.3](#sec3dot2dot3-materials-10-00771){ref-type=\"sec\"}.\n\n### 4.1.2. Results from the Adaptive Crack Model {#sec4dot1dot2-materials-10-00771}\n\nA simulation of the L-shaped slab test is conducted by setting $h = 0.01$ mm and $\\gamma = 0.8$. The results obtained from the complete simulation are contained in [Figure 14](#materials-10-00771-f014){ref-type=\"fig\"}. It is observed that the computed force-displacement relation, as well as the crack pattern agrees, despite the relative coarse discretization, well with the experimental data. Nevertheless, as shown in [Figure 15](#materials-10-00771-f015){ref-type=\"fig\"}, the quality of the crack path is improved when using particularly unstructured refined meshes \\[[@B54-materials-10-00771]\\].\n\nIn the following, the efficiency of the adaptive algorithm is discussed based on the results shown in [Figure 14](#materials-10-00771-f014){ref-type=\"fig\"}. The simulation starts with the original mesh, which contains approximately 19% of the number of nodes and 26% of the number of elements as compared to the case of full fragmentation; in other words, at the start of the analysis at increment $i = 1$, $$\\omega_{{node},1} = \\frac{N_{{node},1}^{adp}}{N_{node}^{full}} \\approx 0.19,\\omega_{{elem},1} = \\frac{N_{{elem},1}^{adp}}{N_{elem}^{full}} \\approx 0.26.$$\n\nHere, $\\omega$ denotes the relative problem size while using the adaptive algorithm as compared with the case of full fragmentation. With the growth of the macroscopic crack, ISEs are gradually generated and located along the potential crack path. When the applied displacement reaches $0.6$ mm, the propagating macroscopic crack almost penetrates the sample. Afterwards, the major crack continues to open, and the structure fails rapidly; the relative problem sizes change marginally, approaching approximately 34% for the number of nodes and 40% for the number of elements, respectively, at the end of simulation when $u = 1$ mm at increment $i = 1000$:$$\\omega_{{node},n_{final}} = \\frac{N_{{node},n_{final}}^{adp}}{N_{node}^{full}} \\approx 0.34,\\omega_{{elem},n_{final}} = \\frac{N_{{elem},n_{final}}^{adp}}{N_{elem}^{full}} \\approx 0.40.$$\n\nHere, $n_{final}$ stands for the last loading increment ($n_{final} = 1000$) applied during the computation. In order to provide a practical assessment of the overall efficiency of the proposed cracking model, we define the \"reduction factor\" $\\Omega$ for a specific problem as the average relative problem size using adaptive insertion as compared to that using full fragmentation.\n\nIn this context, the reduction factors $\\Omega_{elem}$ and $\\Omega_{node}$ are computed as:$$\\Omega_{elem} = \\frac{1}{n_{inc}}\\sum\\limits_{i = 1}^{n_{inc}}\\frac{N_{{elem},i}^{adp}}{N_{elem}^{full}} = \\frac{1}{n_{inc}}\\sum\\limits_{i = 1}^{n_{inc}}\\omega_{{node},i}\\,,\\Omega_{node} = \\frac{1}{n_{inc}}\\sum\\limits_{i = 1}^{n_{inc}}\\frac{N_{{node},i}^{adp}}{N_{node}^{full}} = \\frac{1}{n_{inc}}\\sum\\limits_{i = 1}^{n_{inc}}\\omega_{{elem},i}\\,.$$\n\nReferring to the diagram in [Figure 14](#materials-10-00771-f014){ref-type=\"fig\"}b, the reduction factor can be graphically understood as the ratio of the area under the curve to the full area (with the height = 1), which gives $\\Omega_{node} \\approx 0.30$ and $\\Omega_{elem} \\approx 0.36$. From the documented results, it becomes clear that a significant reduction of the total computational cost is achieved when using the proposed adaptive ISE insertion strategy as compared to inserting interface elements a priori in the entire mesh.\n\nThis simulation, coded in MATLAB${}^{\\mathsf{\u00ae}}$ (2016a, MathWorks, Natick, USA) and run on a desktop PC equipped with a quad-core i7 CPU allowing parallelization and 16 GB memory, took approximately 36 s, as compared to 107 s required for the analysis with full insertion of interface solid elements.\n\n### 4.1.3. Factors Influencing the Efficiency of the Adaptive Strategy {#sec4dot1dot3-materials-10-00771}\n\nEvidently, the efficiency of the adaptive insertion technique highly depends on the specific problem and the computational settings. [Figure 16](#materials-10-00771-f016){ref-type=\"fig\"} illustrates the influence of the stress threshold factor ($\\gamma$) for the determination of critical interfaces (see Equation ([12](#FD12-materials-10-00771){ref-type=\"disp-formula\"})). Adopting a small value of $\\gamma = 0.5$ yields the same simulation results as $\\gamma = 0.8$ used in the previous calculation. However, a significantly larger number of interface solid elements is created. With a high value ($\\gamma = 1.0$), although less ISEs are generated, the over-stressing phenomenon due to the delayed processing of interfaces is observed in [Figure 16](#materials-10-00771-f016){ref-type=\"fig\"}a. It is noted that the choice of $\\gamma$ depends on the load incrementation; larger increments require a smaller value of $\\gamma$.\n\nAnother aspect discussed next is the influence of stress distribution. A test is performed, characterized by a square panel subjected to uniaxial tension ([Figure 17](#materials-10-00771-f017){ref-type=\"fig\"}a). This simple case is often considered for the verification of cracking models for quasi-brittle materials and for the assessment of the influence of different model features, such as the representation of the microstructure and the heterogeneous fracture characteristics of concrete \\[[@B55-materials-10-00771],[@B56-materials-10-00771]\\]. The same material parameters as for the \"L-shaped\" panel are assumed. As expected, a localized crack splitting the panel develops and leads to the failure of the specimen ([Figure 17](#materials-10-00771-f017){ref-type=\"fig\"}a). In [Figure 17](#materials-10-00771-f017){ref-type=\"fig\"}b, one can clearly see that in such an extreme situation characterized by a uniform distribution of tensile stresses in the domain, all interfaces reach the critical state simultaneously at the level of applied displacement $0.08$ mm; consequently, the complete mesh is fragmented and filled with interface solid elements. Hence, in the case of a homogeneous stress distribution, the advantage of the adaptive algorithm as compared to a full insertion of ISEs vanishes. In practical situations, however, crack propagation is accompanied with high stress concentrations, where the adaptive algorithm performs excellently as shown in the previous example of the \"L-shape\" test.\n\n4.2. Application to 3D Structural Analyses {#sec4dot2-materials-10-00771}\n------------------------------------------\n\nThe following examples are concerned with the performance of the implemented adaptive insertion technique for interface solid elements in 3D analyses. In all of these examples, $\\gamma = 0.8$ and $h = 0.01$ mm are used for generating the ISEs.\n\n### 4.2.1. Plain Concrete Notched Beam under Bending {#sec4dot2dot1-materials-10-00771}\n\nA notched beam subjected to three-point bending ([Figure 6](#materials-10-00771-f006){ref-type=\"fig\"}) is re-analyzed. The beam has a length of 600 mm, a height of 150 mm and a width of 150 mm. A notch with a depth of 30 mm is located at the bottom side at the mid-span of the beam. A total displacement of 0.5 mm is applied incrementally at the mid-span on the top surface, which leads to the failure of the beam induced by a vertical crack initiating at the notch tip and propagating towards the top surface ([Figure 18](#materials-10-00771-f018){ref-type=\"fig\"}).\n\nThe original 3D FE-discretization contains 6743 nodes and 34,537 linear tetrahedral elements. For comparison, the simulation is first performed with full insertion of ISEs, which requires the completely fragmented mesh filled with three solid elements at every interfacial gap; consequently, the preprocessed mesh includes 138,148 nodes (approximately 20-times that of the original mesh) and 235,939 elements (approximately seven-times that compared to the original mesh), among which, 201,402 are ISEs.\n\nIn a first analysis, the beam is assumed to be made of plain concrete, and in a second analysis, a fiber-reinforced concrete beam is re-analyzed and compared with experimental results. This has been analyzed by the authors as a 2D problem without using an adaptive strategy for the ISE insertion in \\[[@B17-materials-10-00771]\\]. The simulation results for the plain concrete beam, including the load-displacement curve and the crack pattern, are shown in [Figure 18](#materials-10-00771-f018){ref-type=\"fig\"}.\n\nBy employing the adaptive insertion technique, simulation results are obtained without any noticeable difference. However, the computation starts with the original mesh with interface solid elements being first inserted in the vicinity of the notch-tip and then continuously added in the vicinity of the front of the macroscopic crack, which propagates vertically into the beam until the structure fails. At the end of simulation ($u = 0.5$ mm), only 77,538 ISEs are created and distributed along the structural crack ([Figure 19](#materials-10-00771-f019){ref-type=\"fig\"}b). As compared to the case of full insertion, the efficiency of adaptive insertion is clearly seen in [Figure 19](#materials-10-00771-f019){ref-type=\"fig\"}a. At the beginning, only $5\\%$ of the system degrees of freedom (DOF) and $15\\%$ of the number of elements are used for the computation; these values increase with the loading procedure. When the structural crack approaches the top surface ($u >$ 0.3 mm), the insertion of new ISEs in every load step becomes negligible, and the problem size remains more or less constant at the level of approximately $41\\%$ in regard to the system DOFs and $48\\%$ in regard to the number of elements. As a general assessment, the reduction factors are evaluated according to Equation ([15](#FD15-materials-10-00771){ref-type=\"disp-formula\"}) as $\\Omega_{elem} = 0.44$ and $\\Omega_{DOF} = 0.37$ (identical to $\\Omega_{node}$).\n\n### 4.2.2. Fiber-Reinforced Concrete Notched Beam under Bending {#sec4dot2dot2-materials-10-00771}\n\nIn the second analysis, the same beam geometry is adopted, however, now assumed to be made of fiber-reinforced concrete with 60 kg/m${}^{3}$ long hooked-end fibers. [Figure 20](#materials-10-00771-f020){ref-type=\"fig\"}a shows the load-displacement relation obtained from the numerical simulation (full line). The comparison with the experimental result (dotted line) shows a good agreement over the complete loading range. The figure also contains the crack pattern as the contour plot of the crack opening magnitude in the deformed configuration (with 10-fold magnification of displacement). The efficiency of the adaptive insertion technique is clearly shown in [Figure 20](#materials-10-00771-f020){ref-type=\"fig\"}b. In the early stages of load ($u <$ 0.3 mm), due to the quick advance of the macroscopic crack, the relative problem size in regards to the number of degrees of freedom and the number of elements grows rapidly. After the major crack surface approaches the top surface of the beam, the structure exhibits a rather ductile behavior with considerable residual load-bearing capacity as the crack continues to open and the hooked-end steel fibers intercepting the crack are successively activated and eventually pulled out from the concrete matrix. In this loading stage, the relative problem size remains at a certain level without significant change. The simulation proceeds until $u =$ 6 mm due to the fairly ductile behavior of the FRC specimen. The reduction factors are evaluated in this case as $\\Omega_{elem} = 0.46$ and $\\Omega_{DOF} = 0.40$.\n\n### 4.2.3. Plain and Fiber-Reinforced Concrete Notched Beam Subjected to Torsion {#sec4dot2dot3-materials-10-00771}\n\nThe last example qualitatively demonstrates the model performance in the full-3D situation that cannot be simplified to 2D. To this end, the so-called \"Brokenshire test\", where, by imposing torsion onto a concrete prism with an inclined notch, a non-planar crack surface develops \\[[@B57-materials-10-00771]\\], is re-analyzed numerically. The original test on the plain concrete specimen has been replicated by the previous version of the model (see \\[[@B17-materials-10-00771]\\] for more details); similar validations have been performed applying different cracking models including the smeared, embedded and discrete approaches \\[[@B37-materials-10-00771],[@B58-materials-10-00771],[@B59-materials-10-00771]\\]. Here, the computational simulation of the plain concrete beam test is re-conducted using the adaptive crack model. In a second analysis, we assume, that the beam is made of fiber-reinforced concrete, with 60 kg/m${}^{3}$ short straight fibers being added into the concrete matrix.\n\nThe original mesh contains 8119 nodes and 42,427 linear tetrahedral elements. [Figure 21](#materials-10-00771-f021){ref-type=\"fig\"}a shows the load-displacement relation obtained for the FRC prism (full red line). Comparing this result with the result from the plain concrete specimen (full blue line), a considerable increase of the ductility is observed. Experimental data are only available for the plain concrete case and not for the FRC prism. [Figure 21](#materials-10-00771-f021){ref-type=\"fig\"}a contains the range of the experimental load-displacement curves for plain concrete documented in \\[[@B57-materials-10-00771]\\] as dotted lines. An excellent agreement between the computed and the measured load-displacement curves is observed. It is noted that according to the numerical simulations, the crack pattern does not differ significantly between the plain concrete and the FRC prisms. In both cases, a non-planar crack develops (see also the inlet in [Figure 21](#materials-10-00771-f021){ref-type=\"fig\"}b). In regard to the efficiency of the adaptive insertion technique ([Figure 21](#materials-10-00771-f021){ref-type=\"fig\"}b), a similar tendency as found in the previous example of the 3D analysis of FRC beam is noticed (see [Figure 20](#materials-10-00771-f020){ref-type=\"fig\"}b): in the early stage, the relative problem size grows rapidly and becomes almost stationary after the first progressive growth of the crack. The simulation ends with a mesh possessing 106,875 nodes and 194,923 elements. The reduction factors for this simulation are evaluated as $\\Omega_{elem} = 0.56$ and $\\Omega_{DOF} = 0.50$. It is noticed that, as compared to the previous example, the efficiency of the adaptive strategy is less pronounced in this example; this is due to the a priori refined mesh in the local region of the notch. Nevertheless, in regard to the computation time, this simulation took 6.5 h, which is only 23% of the time (28.1 h) required for the analysis using full ISE insertion!\n\n5. Concluding Remarks {#sec5-materials-10-00771}\n=====================\n\nA multilevel modeling framework previously developed for the analysis of fiber-reinforced concrete was implemented in an adaptive computational crack model using continuum interface solid elements (ISEs) for 2D and 3D simulations of fiber-reinforced concrete structures. The ISE formulation allows one to capture complex fracture processes without the need for a crack tracking algorithm. Since the numerical simulation of crack propagation using interface elements usually involves a considerably larger number of degrees of freedom as compared to other classes of concrete cracking models, such as smeared crack models or embedded crack models, an adaptive procedure was proposed to perform structural crack propagation analyses with an acceptable computational effort. In contrast to the naive implementation, where the entire FE mesh is fragmented a priori and all available interfaces are filled with degenerated solid elements prior to the computation, the interface solid elements are now inserted only along those interfaces during the incremental analysis, which are identified as locations of the potential crack path according to a stress criterion. Hence, ISEs are adaptively created in the vicinity of the advancing crack front. The structural simulation starts with the original discretization of the domain using standard solid finite elements. During the failure analysis, depending on the fracture advancement, the problem size increases gradually as ISEs are inserted. The numerical examples have shown that in spite of this increase in computational costs due to mesh adaptation as the loading increases, the adaptive algorithm allows one to significantly reduce the total computational cost without affecting the simulation accuracy. In selected 2D, as well as in 3D applications, the maximum number of degrees of freedom, required only when the fracture process has advanced almost to the state of structural failure, was reduced to approximately 30--40% as compared to an analysis, in which ISEs are a priori inserted along all element edges of the discretized structure. As a consequence, the computation consumes only 15--30% of the time (according to the selected cases for comparison). In large-scale analyses, these savings in time, in association with the considerable reduction in memory requirements, may be the pre-requisite to be able to perform this type of failure analysis, if no high performance hardware environment is available. It was also noticed that the performance of the proposed adaptation technique highly depends on the characteristics of specific problem. The efficiency of the algorithm is higher in problems characterized by high stress concentrations, which is usually the case during the evolution of bending cracks.\n\nFinancial support was provided by the German Research Foundation (DFG) in the framework of Project B2 of the Collaborative Research Center SFB837 \"Interaction modeling in mechanized tunneling\". This support is gratefully acknowledged.\n\nYijian Zhan and G\u00fcnther Meschke developed the model, performed the analysis and wrote the paper.\n\nThe authors declare no conflict of interest.\n\n![Selected results of the single fiber pullout model \\[[@B38-materials-10-00771]\\]: (**a**) straight steel fiber with a 30${}^{\\circ}$ inclination angle with respect to the loading direction; (**b**) hooked-end steel fiber with a 60${}^{\\circ}$ inclination angle.](materials-10-00771-g001){#materials-10-00771-f001}\n\n![Crack bridging model: (**a**) position and inclination of a fiber with respect to the crack; (**b**) unit area of an opening crack in FRC intercepted by fibers with length $L_{f}$; (**c**) sketch of the obtained traction-separation relations for different FRC composites.](materials-10-00771-g002){#materials-10-00771-f002}\n\n![Results of crack-bridging effect in FRC obtained from the investigation of uniaxial tension tests on (**a**) a notched prism cast in a \"standing\" mold and (**b**) a \"dog bone\" specimen cast in a \"lying\" mold.](materials-10-00771-g003){#materials-10-00771-f003}\n\n![(**a**) Processed mesh for computation, obtained after the insertion of solid elements into all interfacial gaps; (**b**) degenerated 3D solid element characterized by the \"base\", the \"apex\" Node 4 and its projection on the base (Point 4'). ISE, interface solid element.](materials-10-00771-g004){#materials-10-00771-f004}\n\n![Interface in (**a**) undeformed and (**b**) deformed configuration. (**c**) Equivalent traction-separation relation for the cohesive interface model.](materials-10-00771-g005){#materials-10-00771-f005}\n\n![Analysis of a three-point bending test on a notched FRC beam: (**a**) photo of the failure state of the specimen and the contour plot of the crack-opening magnitude in the deformed configuration; (**b**) crack patterns represented by the activated interface solid elements at different loading states; (**c**) comparison between the force-displacement relations predicted by the proposed model and from the experiments \\[[@B41-materials-10-00771]\\] for three different fiber cocktails.](materials-10-00771-g006){#materials-10-00771-f006}\n\n![Pre-processing (insertion of interface solid elements in the complete domain): (**a**) original finite element mesh; (**b**) phantom mesh obtained by duplicating the edges and shrinking the bulk elements; (**c**) actual mesh for computation, obtained after insertion of solid elements into all interfacial gaps.](materials-10-00771-g007){#materials-10-00771-f007}\n\n![Concept of adaptive insertion of ISEs.](materials-10-00771-g008){#materials-10-00771-f008}\n\n![Generation and use of the phantom mesh: (**a**) original mesh; (**b**) phantom mesh; (**c**) mesh being processed (interface solid element not yet generated).](materials-10-00771-g009){#materials-10-00771-f009}\n\n![Insertion of solid elements at a specific interface in 2D: (**a**) original mesh with two bulk elements; (**b**) Node-1 split, the first interface solid element inserted (ISE-I, in gray); (**c**) Node-3 split, the second interface solid element inserted (ISE-II, in gray).](materials-10-00771-g010){#materials-10-00771-f010}\n\n![Insertion of solid elements at one interface in 3D: (**a**) original mesh with two bulk elements; (**b**) Node-1 split; ISE-I inserted (gray colored); (**c**) Node-2 split; ISE-II inserted (gray); (**d**) Node-3 split; ISE-III inserted (gray).](materials-10-00771-g011){#materials-10-00771-f011}\n\n![\"L-shaped\" panel test: (**a**) experimental setup (dimensions in mm) and crack pattern \\[[@B50-materials-10-00771]\\]; (**b**) original finite element discretization.](materials-10-00771-g012){#materials-10-00771-f012}\n\n![Progressive insertion of ISEs during the 2D simulation of the \"L-shape\" test: (**a**) original mesh (with the interfaces activated according to the labeled sequence; (**b**--**e**) splitting of Edge-1--4; (**f**) splitting of Edge-5 and -6. The red triangles represent the solid elements created at a new fully-separated interface; the blue triangles are the solid elements at partially open interfaces; the red dashed circles indicate the newly-activated joints.](materials-10-00771-g013){#materials-10-00771-f013}\n\n![Results from the 2D simulation of the \"L-shape\" test using the proposed adaptive crack model: (**a**) comparison between the structural responses obtained from simulation and experiment; (**b**) evolution of the relative problem sizes ($\\omega$) regarding the system degree of freedom and number of elements. The generated interface solid elements are inserted in color in the mesh).](materials-10-00771-g014){#materials-10-00771-f014}\n\n![Results from the 2D simulation of the \"L-shape\" test using the proposed adaptive crack model: (**a**) crack pattern obtained with a structured refined mesh; (**b**) crack pattern obtained with an unstructured refined mesh.](materials-10-00771-g015){#materials-10-00771-f015}\n\n![Results of the 2D simulation of \"L-shape\" test using the proposed adaptive crack model: (**a**) comparison between the structural responses obtained from the simulations with different values of $\\gamma$; (**b**) all of the created interface solid elements (blue lines) for the case of $\\gamma = 0.5$; (**c**) all of the created interface solid elements (blue lines) for the case of $\\gamma = 1.0$.](materials-10-00771-g016){#materials-10-00771-f016}\n\n![Results of the 2D simulation of uniaxial tension on a square specimen: (**a**) computed structural responses and crack pattern; (**b**) evolution of the relative problem sizes regarding the system degree of freedom and number of elements, as well as all of the created interface solid elements (blue lines in the mesh).](materials-10-00771-g017){#materials-10-00771-f017}\n\n![Results from a 3D simulation of the notched FRC beam made of plain concrete: (**a**) load-displacement diagram; (**b**) crack pattern (side view and bottom-side view with the contours representing the crack opening magnitude in the deformed configuration).](materials-10-00771-g018){#materials-10-00771-f018}\n\n![Results from a 3D simulation of the notched FRC beam with adaptive insertion of ISEs: (**a**) evolution of the relative problem sizes in terms of the number of degrees of freedom and the number of elements, respectively; (**b**) generated ISEs at the end of the simulation ($u =$ 0.5 mm).](materials-10-00771-g019){#materials-10-00771-f019}\n\n![Results of a 3D simulation of the fiber-reinforced notched beam test: (**a**) comparison of load-displacement curves obtained from the adaptive finite element analysis and from the experiment \\[[@B41-materials-10-00771]\\] and crack pattern; (**b**) evolution of the relative problem sizes in terms of the number of degrees of freedom and elements, respectively, and generated ISEs at the end of simulation ($u =$ 6 mm).](materials-10-00771-g020){#materials-10-00771-f020}\n\n![Results of a 3D simulation of a plain- and a fiber-reinforced notched prism subjected to torsion: (**a**) load-displacement curves obtained for plain- and fiber-reinforced concrete, respectively, and the crack pattern for the FRC prism; (**b**) evolution of the relative problem sizes in terms of the number of degrees of freedom and elements, respectively. The inlet contains the twisted crack surface obtained from the analysis of the FRC prism.](materials-10-00771-g021){#materials-10-00771-f021}\n"}
+{"text": "INTRODUCTION\n============\n\nBurn is still a devastating emergency with many physical and psychological disabilities^[@B1]^ and mortality and morbidity.^[@B2]^ Bacteria are important causes of nosocomial infection leading to septicemia and death in burn patients denoting to its public health importance.^[@B3]^ For survivors, scarring is still a problem posing psychological effects to the burn patient.^[@B4]^ Resistance to common therapies for several bacteria was noted in burn injuries complicating the situation more,^[@B5]^ even there have been efforts for treatment of burn wounds with medications having less adverse effects, such as medicinal herbals.^[@B6]-[@B10]^\n\nAnnually many individuals in low and middle income countries suffer from burn injuries.^[@B11]^ In the classic manner, the treatment of burn is via daily washing of the wound, removal of the dead tissue and antibiotic dressing till the formation of granulation tissue and later grafting.^[@B12]^ Different methods have been used for the care and dressing of the donor site due to the facilitation in the improvement and reduction of wound symptoms and therefore it implies the further possibility of donor site from the same locale.^[@B6]-[@B10],[@B13],[@B14]^ Among these the use of dry gas, gas dripping with an antibiotic, two or multilayer dressing, the vaseline gauze which is presently utilized in the scald department and amniotic membrane as biological dressing for treatment or care of some wounds can be mentioned.^[@B15]-[@B18]^ But yet, no any study described it's using and efficacy for donor site of skin grafting. Therefore the present study was carried out with an aim to reveal the efficiency of the biological dressing with amniotic membrane for the reduction of pain in the donor site, improvement of the body movement in the donor site limb and evaluation of duration of dressing sloughing from the wound and risk of infection.\n\nMATERIALS AND METHODS\n=====================\n\nThis study was a clinical trial performed on 32 patients. These patients were hospitalized in the burn ward of the Besat Hospital due to burns and among the patients, the individuals aged over 18 years were enrolled. Each patient was considered as an intervention group and also as a control group. The one side of an organ was considered as an intervention area and the other side as a control region.\n\nThe amniotic membrane in a fresh form was provided from the Department of Obstetrics and Gynaecology of Fatimane Hospital in an elective caesarean section. Infections to TORCH, hepatitis B, hepatitis C, AIDS, and syphilis before childbirth was not included. For hepatitis B and C, AIDS and syphilis infections, serological tests of HIVAb, HBS Ag, HCVAb and VDRL were carried out.\n\nAfter the cesarean section and under sterile conditions, the amniotic membrane was separated and placed in the normal saline solution and was immediately transferred to the burn department of the Apostolate Hospital in less than 6 hours from its isolation to be used for the study. The donor from the two limbs and or two harmonious body sections was selected using an electric dermatome with the thickness of 0.014 inch from each side. The dressing of the donor site in one side was done with vaseline gauze and was bandaged.\n\nFor the dressing of the donor site on the opposite, the amniotic membrane was placed on the donor site and then the vaseline gauze on the membrane and then the dry gas was introduced and bandaged. The limbs that were subjected to study (i.e. the amniotic membrane placed on the donor site) was coded A, and the limb which was regarded as the control site (i.e. regular dressing or non-biological) was coded B.\n\nAfter the 24 hours of surgery, the bandage and the dry gases of both sites (case study and control) were removed till the level of vaseline gauze and the two sites were exposed to open air (hospitalized room) and later, the pain score, easiness in the move of organs, the spontaneous separation time of the last layer of dressing on the donor site and the condition of sepsis according to the protocol was provided.\n\nThe pain score and easiness in the movement of the limbs that considered as the donor sites were determined by patient with the 1-5 and 1-10 score respectively in a daily examination from the patients and recorded in a questionnaire. Score 1 represented the lowest pain level and organ movement and 5 showed the highest organ movement level and 10 demonstrated the highest pain level and the validity of the infections in the two sites of case and control based on the daily clinical examination in the case of clinical symptoms based on the probability of sepsis (pain, discharge and inflammation, etc.), the cultivation was procured and in the case of bacterial growth over 100,000 colonies, it was considered positive.\n\nThe patients who were discharged early post operatively, they were followed in the Sheikholraeis clinic and the required information about pain, organ movement, separation of dressings and sepsis were recorded in a questionnaire. The collected information from patients was eventually entered in the SPSS software (version 16, Chicago, IL, USA) and was analyzed using Mann Whitney test. In the context of investigating the presence of pain difference in various days in the donor sites, t-test was used till the fifth day and Mann Whitney test from the sixth day for the comparison of pain and easiness of the limbs movement scores.\n\nRESULTS\n=======\n\nThirty two patients with range of 18-88 years were enrolled while 18 patients (56%) were male. The pain status in the intervention and control areas were shown in [Tables 1](#T1){ref-type=\"table\"} and [2](#T2){ref-type=\"table\"}. The average pain score in the intervention areas was less than the control areas. From the 5^th^ to 14^th^ day, there was no significant difference in the pain score between two sides, although in all these cases, the pain score in the intervention area was less than the control area. The status of movement range in the intervention and control areas were demonstrated in [Tables 3](#T3){ref-type=\"table\"} and [4](#T4){ref-type=\"table\"}.\n\n###### \n\nComparing the average pain score in the two case area of intervention (dressing with amniotic membrane) and control (general dressing) in the varied research days\n\n  **Day**   **Average of intervention area**   **Number**   **Average of control area**   **Number**   **P value**\n  --------- ---------------------------------- ------------ ----------------------------- ------------ -------------\n  1         5.21                               32           7.2                           32           0.0001\n  2         4.15                               32           5.9                           32           0.0001\n  3         3.56                               32           5.12                          32           0.001\n  4         3.06                               32           4.15                          32           0.016\n  5         2.68                               29           3.51                          29           0.057\n\n###### \n\nComparing the average pain ranks in the two case area of intervention (dressing with amniotic membrane) and control (general dressing) in the varied research days\n\n  **Day**   **Average of intervention area**   **Number**   **Average of control area**   **Number**   **P value**\n  --------- ---------------------------------- ------------ ----------------------------- ------------ -------------\n  6         19.69                              21           23.31                         21           0.32\n  7         16.17                              18           20.83                         18           0.17\n  8         9.55                               10           11.45                         10           0.44\n  9         6.25                               6            6.75                          6            0.79\n  10        5.3                                5            5.7                           5            0.81\n  11        3.5                                3            3.5                           3            1.00\n  12        1.75                               2            3.25                          2            PV0.22\n  13        1.75                               2            3.25                          2            PV0.22\n  14        1.5                                1            1.5                           1            1.00\n\n###### \n\nComparing the average move score in the two case organs of intervention (dressing with amniotic membrane) and control (general dressing) in varied research days\n\n  **Day**   **Average of intervention area**   **Number**   **Average of control area**   **Number**   **P value**\n  --------- ---------------------------------- ------------ ----------------------------- ------------ -------------\n  1         3.06                               32           2.03                          32           0.0001\n  2         3.28                               32           2.53                          32           0.007\n  3         3.59                               32           2.78                          32           0.002\n  4         3.81                               32           3.2                           32           0.027\n  5         2.68                               29           3.24                          29           0.012\n\n###### \n\nComparing the average move score in the two case organs of intervention (dressing with amniotic membrane) and control (general dressing in varied research days).\n\n  **Day**   **Average of intervention area**   **Number**   **Average of control area**   **Number**   **P value**\n  --------- ---------------------------------- ------------ ----------------------------- ------------ -------------\n  6         24.12                              21           18.88                         21           0.135\n  7         19.44                              18           17.56                         18           0.135\n  8         11                                 10           10                            10           0.553\n  9         6.42                               6            6.58                          6            0.656\n  10        5.5                                5            5.5                           5            1.000\n  11        3.33                               3            3.67                          3            0.796\n  12        2.5                                2            2.5                           2            1.000\n  13        2.5                                2            2.5                           2            1.000\n  14        1.5                                1            1.5                           1            1.000\n\nThe movement condition in the intervention limbs till the fifth day of hospitalization in comparison to the control limb exhibited a significant difference (*p*=0.01) showing that the intervention cases which were dressed by the amniotic membrane had a better movement status in comparison to the routine dressing. From the 6^th^ day to the 14^th^ day, there was no significant difference between the movement of limbs in both sides, although in most of the cases, the intervention patients had a better movement status in relation to the control organ.\n\nThe dressing separation time in the intervention cases from the days 3 to 18 after the surgery was about 8.15\u00b13.4 days and in the control cases, the dressing separation in the days 5 to 21 was about 9.2\u00b13.7 days while the difference of the dressing separation in intervention and control organs was not significant (*p*=0.29). Regarding the infection of intervention and control group, it was clear that two infection cases had appeared on the site dressed with amniotic membrane where in the general dressing site (control area), no infection was observed. The difference for infection in the dressings of intervention and control organs was not significant (*p*=0.49).\n\nDISCUSSION\n==========\n\nAfter skin grafts, the donor site was painful and caused immobility of patients and also prone them to infection, hypertrophic scar formation and changes in color^[@B19]^ that can increase hospitalization period or even can be the indication for later cosmetic surgery. Immobility can cause deep vein thrombosis (DVT), and respiratory, gastrointestinal, endocrine and electrolyte disorders.^[@B20]^ Based on these complications after surgery especially skin transplantation, it is of great importance to prevent and reduce the complications few days after surgery. The coverage of donor site requires materials that preserve the epidermal function and integrate itself into the process of healing.^[@B16]^\n\nBiological dressing was introduced as a gold standard for temporary covering of wounds.^[@B21]^ All biologic dressings are susceptible to early reaction and the only exception is the amniotic membrane.^[@B21]^ The use of amniotic membrane as biological dressing in the treatment of extensive burn wounds has been described in order to early recovery of patient, improvement of wound healing and its quality.^[@B15]^ Studies demonstrated that use of amnion in burn wounds can lead to reduction of pain intensity and prevent water and electrolyte disturbances and also can help for early preparation of the wound bed for grafting.^[@B20],[@B21]^\n\nThe preparation and application of amniotic membrane according to our proposed method performed with simplicity and efficiency, lead to a significant difference for 3 variables between donor site with routine dressing and dressing with amniotic membrane. Biologic dressing lead to significant reduction of pain score at the site of skin graft in the first few days after surgery. In fact with covering of donor site with amnion, we introduced an occlusive dressing. Growth factors extravasated and accumulated in the interface of biologic dressing and abraded epidermis. This moisture and suitable interface could hasten the wound healing process and epithelialization.^[@B16]^\n\nAmnion prevents the wound from irritation, evaporation, dryness and also nerve ending stimuli on the wound surface. These characters could not be seen in routine dressings. In the first 24-48 hours after any skin abrasion, skin harvesting or even primary repair in surgical wounds, basal epithelial cells can cause a water tight closure in wound layer together with migration and mitosis processes.^[@B4]^\n\nThis layer does not have enough strength against external mechanical stresses and is fragile. With disruption of this fine neo-epithelium, underlying nerve endings are exposed and suffered with contact to non-biological dressings. But after fifth day of graft taking, epithelialization reaches to optimal thickness and resistance in order to preserve its integrity against external stress. Therefore, pain score of both studied and control site did not show any difference after this time. Less pain at the first few days of skin harvesting is important. Patients could be mobilized soon and the need for analgesia during hospitalization is decreased and subsequent complications in respiratory and coagulation system are prevented.^[@B16]^\n\nEarly and feasible mobility of patients with application of amnion over the donor site seems to be a secondary effect of its pain subsiding. Although in this study, there was not statistical difference about the time of dressing separation for two types of dressing, but clinically the mean time of dressing sloughing in studied limb was 8.5 and for routine dressing was 9.2 days. This clinical difference is due to improvement of epithelialisation with amnion dressing while amniotic membrane has low antigenicity.^[@B16]^\n\nApplication of amnion in the first few days adheres tightly to wound surface and has positive modulation for wound healing process either in quality or in rapidity. Accordingly in this condition, possible morbidities of wound healing such as color change or hypertrophic scar would formation can be decreased. Dermal substitutes as biologic dressings can be used for wound care in plastic surgeries for this purpose but may not be cost-benefit.^[@B16]^ In this study, we did not have significant wound infection following to dressing by amnion. Amniotic membrane can prevent bacterial invasion to the donor site of wound by use of amniotic membrane. Similar conclusions could be seen in other studies.^[@B17],[@B22]^\n\nReduction of pain severity and improvement of mobility with application of amniotic membrane over the donor site can cause earlier re-harvesting of graft from previous donor site specially in extensive burns and also decrease the duration of hospitalization, need for analgesia and complications of immobilization. Dressing of donor site of skin graft with amniotic membrane can reduce pain in the early post-operative days and facilitate the patient mobility and early sloughing of dressing from the wound bed.\n\nDue to ease of preparation and use of amniotic membrane as well as absence of adverse effects of its application, it is recommended to use amniotic membrane in the site of skin graft in order to be harvested in burn patients specially with high percentage of burn lesions for pain reduction and improvement of patient mobility and also acceleration of recovery of donor site for graft re-harvesting without any fear of infection.\n\nThe authors wish to appreciate the Hamadan University of Medical Sciences for their support.\n\nCONFLICT OF INTEREST\n====================\n\nThe authors declare no conflict of interest.\n"}
+{"text": "Introduction {#S0001}\n============\n\nPsoriasis is a chronic inflammatory skin disease clinically featured by erythematous plaques covered with silvery scales.[@CIT0001],[@CIT0002] Psoriasis would cause high morbidity duo pain, itching, functional and cosmetic impairments, and even high mortality due to depression and suicidal contemplations. The prevalence of psoriasis is currently estimated to be as high as 2--3% worldwide, becoming a serious global problem.[@CIT0003]--[@CIT0005] Moreover, it is also associated with many comorbidities such as psoriasis arthritis,[@CIT0006] metabolic syndrome[@CIT0007] and cardiovascular disease,[@CIT0008] which brings huge health and economic burden to patients. Although various immune abnormalities have been proposed to be involved in the pathogenesis of psoriasis,[@CIT0009] oxidative stress is also believed to play a pivotal role in the pathophysiological mechanism. Increased production of ROS would induce a vast number of biological responses to the initiation of psoriasis pathogenesis.[@CIT0010]--[@CIT0012] ROS including superoxide anion (O~2~\u2022^\u2212^), \u2022OH free radicals and nonradical molecules such as H~2~O~2~ would induce oxidative damage, such as lipid peroxidation, DNA modification, and secretion of inflammatory cytokines in psoriatic derma.[@CIT0011],[@CIT0012] Oxidative damage markers including malondialdehyde, lipid hydroperoxides, thiobarbituric acid reactive substances, protein carbonyl, and nitric oxide have been detected in patients with psoriasis.[@CIT0011],[@CIT0013] Therefore, antioxidative strategies eradicating ROS may serve as effective and easy treatment options for psoriasis.[@CIT0014] Antioxidants, such as epigallocatechin-3-gallate,[@CIT0015] glabridin,[@CIT0016] proanthocyanidins,[@CIT0017] polyandric acidA[@CIT0018] and other natural compounds[@CIT0019],[@CIT0020] with beneficial effects on cutaneous psoriasis have been reported.\n\nRecently, nanomaterials with enzyme-like activity named nanozymes,[@CIT0021],[@CIT0022] have been exploited as potential therapeutics in various diseases, including Parkinson's disease,[@CIT0023] Alzheimer's disease,[@CIT0024] cancer,[@CIT0025]--[@CIT0027] ischemic stroke,[@CIT0028],[@CIT0029] and ischemia reperfusion injury,[@CIT0030] through mainly eliminating ROS levels in cells. For instance, Mn~3~O~4~ nanozymes have been used as a promising therapeutic agent for treating inflammation because of their excellent ROS scavenging activity.[@CIT0023] Ceria nanoparticles (CeNPs) exhibit tremendous potential as effective antioxidant enzymes, such as peroxidase, oxidase, catalase, and SOD.[@CIT0031],[@CIT0032] These high-performance ROS reduction capacities originate from the dual oxidation states (Ce^3+^/Ce^4+^) on the surface of these particles in which Ce^3+^ is responsible for eliminating O~2~\u2022^\u2212^and \u2022OH, while Ce^4+^ eradicating H~2~O~2~.[@CIT0032] CeNPs have been applied to treat various ROS-associated diseases, including ischemic stroke,[@CIT0033] rheumatoid arthritis,[@CIT0034] autoimmune degenerative disease.[@CIT0035] Nowadays, nanodermatology is an emerging field that uses nanotechnology to facilitate the diagnosis and treatment of skin disease.[@CIT0036],[@CIT0037] However, most of them are inorganic nanomaterials lack of multi-functional and have not been explored for psoriasis treatment.\n\nTo fill this research gap, we designed a multifunctional drug delivery system based on CeNPs capped with \u03b2-CDs for psoriasis treatment ([Scheme 1](#SCH0001)). \u03b2-CDs/CeO~2~ NPs exhibit high mimetic enzymatic activity to eliminate intracellular ROS, rendering them ideal antioxidants for the treatment of oxidative stress-induced damage in psoriasis. Moreover, the introduction of \u03b2-CDs, a family of cyclic oligosaccharides,[@CIT0038],[@CIT0039] on the surface of CeNPs increases their water solubility, biocompatibility, and antioxidant property. Furthermore, the porous nanostructures with unique hydrophobic \u03b2-CDs cavity could be used as a promising drug carrier for hydrophobic molecules by supra-molecular inclusion.[@CIT0040],[@CIT0041] In the present study, we investigated the drug-loading ability and anti-psoriasis activity of \u03b2-CDs/CeO~2~ NPs on IMQ-induced psoriasis-like mouse model by using DIT as the lipophilic model drug. Scheme 1Schematic interpretation of the design of \u03b2-cyclodextrin capped ceria nanoparticles as a nanozyme loaded with dithranol for the combinational therapy of psoriasis.\n\nMaterials and Methods {#S0002}\n=====================\n\nMaterials {#S0002-S2001}\n---------\n\nDIT was obtained from Yuanye Biotechnology Co. Ltd. (Shanghai, China). \u03b2-CDs were purchased from Kelong Chemical Reagent Factory (Chengdu, China). Cerium nitrate was obtained from Bodi Chemical Company Co. Ltd. (Tianjin, China). Carbopol 940 was supplied by Tianliyuan Biotechnology Co. Ltd. (Qingdao, China). IMQ was purchased from Mingxin Pharmaceutical Co. Ltd. as a topical cream (5% imiquimod; Sichuan, China). Halometasone cream (0.05%) was purchased from Huabang Pharmaceutical Co. Ltd. (Chongqing, China). Paraformaldehyde (4%) was obtained from Biosharp Biological Technology (Anhui, China). Methanol was attained from Tedia Compang (high purity, US). Sodium sulfide was purchased from Hengxing Chemical Reagent Company Co. Ltd. (Chengdu, China).\n\nAnimals {#S0002-S2002}\n-------\n\nMale BALB/c mice (6--8 weeks old) were purchased from Qingdao Darenfucheng Animal Husbandry Co. Ltd. (Qingdao, China) and housed under specific pathogen-free conditions by the Experimental Animal Center, Weifang Medical University (Weifang, China). All of the animal experiments performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals published by the Weifang Medical University. The protocol was approved by the Committee on the Ethics of Animal Experiments of Weifang Medical University (Permit Number: 2019SDL096). The treatment of experimental animals followed the 3Rs principles. All experiments were abided by the ethical principles of experimental animal welfare, and every effort was made to minimize suffering.\n\nCells Line and Cells Culture {#S0002-S2003}\n----------------------------\n\nThe HaCaT human keratinocyte cells line was obtained from Procell Life Science & Technology Co. Ltd. (Wuhan, China) and cultured in RPMI 1640 medium (Gibco, NY, USA) supplemented with 15% fetal bovine serum (Beijing, China), 100 U/mL penicillin, and 100 mg/mL streptomycin. The cells were maintained at 37\u00b0C with 5% CO~2~ in a humidified incubator. The growth medium was replaced every 2--3 days.\n\nPreparation of DIT-Loaded \u03b2-CDs/CeO~2~ NPs {#S0002-S2004}\n------------------------------------------\n\nThe synthesis of \u03b2-CDs capped CeO~2~ NPs were performed according to previous literature.[@CIT0042] Cerium nitrate solid particles were added to 30 mL of distilled water, and ultrasonic cleaner was used to dissolve it completely. \u03b2-CDs solid was added and stirred with a magnetic stirrer for 10 min. The colorless liquid became a white mixture. Then, NaOH solid was added to the mixed solution slowly and stirred for another 15 min. The white mixture changes to a reddish brown liquid. The prepared mixture liquid was placed at 125\u00b0C for 6 h using a hydrothermal method, cooled to room temperature, and then left at room temperature for 12 h. The solution of blank NPs was slowly dropped into the dissolved DIT in different concentrations, and the optimal particle size was selected as the dose concentration.\n\nCharacterization of DIT@ \u03b2-CDs/CeO~2~ NPs {#S0002-S2005}\n-----------------------------------------\n\nThe particle size distribution and polydispersity index (PDI) of the blank NPs and drug-loaded NPs were determined in triplicate at 25\u00b0C using a Zeta Sizer Nano ZS90 (Malvern Instruments, Malvern, UK). The morphologies of the blank NPs and drug-loaded NPs were observed by TEM (Tecnai 20, FEI, USA). Measurements were performed in triplicate for each sample. A standard curve was obtained for DIT using a UV-visible light spectrophotometer at 289 nm, and the drug loading (DL) and entrapment efficiency (EE) were measured. The DL (%) and EE (%) were calculated as follows: $$\\documentclass[12pt]{minimal}\n\\usepackage{wasysym}\n\\usepackage[substack]{amsmath}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage[mathscr]{eucal}\n\\usepackage{mathrsfs}\n\\DeclareFontFamily{T1}{linotext}{}\n\\DeclareFontShape{T1}{linotext}{m}{n} {linotext }{}\n\\DeclareSymbolFont{linotext}{T1}{linotext}{m}{n}\n\\DeclareSymbolFontAlphabet{\\mathLINOTEXT}{linotext}\n\\begin{document}\n$${\\rm{Drug}}\\,\\,{\\rm{loading}}\\,(\\% ) = {{{\\rm{Actual}}\\,\\,{\\rm{amount}}\\,\\,{\\rm{of}}\\,\\,{\\rm{drug}}\\,\\,{\\rm{encapsulated}}\\,\\,{\\rm{in}}\\,\\,{\\rm{NPs}}} \\over {{\\rm{Amount}}\\,\\,{\\rm{of}}\\,\\,{\\rm{NPs}}}} \\times 100\\% $$\n\\end{document}$$$$\\documentclass[12pt]{minimal}\n\\usepackage{wasysym}\n\\usepackage[substack]{amsmath}\n\\usepackage{amsfonts}\n\\usepackage{amssymb}\n\\usepackage{amsbsy}\n\\usepackage[mathscr]{eucal}\n\\usepackage{mathrsfs}\n\\DeclareFontFamily{T1}{linotext}{}\n\\DeclareFontShape{T1}{linotext}{m}{n} {linotext }{}\n\\DeclareSymbolFont{linotext}{T1}{linotext}{m}{n}\n\\DeclareSymbolFontAlphabet{\\mathLINOTEXT}{linotext}\n\\begin{document}\n$${\\rm{}}\\matrix{\n \\matrix{\n {\\rm{Entrapment}}\\, \\hfill \\cr \n {\\rm{efficiency}}\\,{\\rm{(\\% )}} \\hfill \\cr} \\cr \n } \\,{\\rm{ = }}\\,{\\matrix{\n {\\rm{Actual}}\\,{\\rm{amount}}\\,{\\rm{of}}\\,{\\rm{drug}}\\, \\hfill \\cr \n {\\rm{encapsulated}}\\,{\\rm{in}}\\,{\\rm{NPs}} \\hfill \\cr} \\over \\matrix{\n {\\rm{Intial}}\\,{\\rm{of}}\\,{\\rm{amount}}\\, \\hfill \\cr \n {\\rm{of}}\\,{\\rm{drug}}\\,{\\rm{used}} \\hfill \\cr} } \\times 100\\%$$\n\\end{document}$$\n\nThe chemical structure of the prepared NPs was characterized using Fourier transform infrared spectrometry (FT-IR; VERTEX 70; Bruker, Bremen, Germany). The spectra of the samples of \u03b2-CDs, cerium nitrate, DIT, \u03b2-CDs/CeO~2~ NPs, and DIT\\@\u03b2-CDs/CeO~2~ NPs were obtained separately. X-ray photoelectron spectroscopy (Ultima IV; Rigaku Corporation, Japan) was used to investigate the microstructure of the \u03b2-CDs/CeO~2~ NPs.\n\nIn vitro Release of DIT@ \u03b2-CDs/CeO~2~ NPs {#S0002-S2006}\n-----------------------------------------\n\nIn vitro drug release study of DIT\\@\u03b2-CDs/CeO~2~ NPs and free DIT solution was conducted by the dialysis bag method (molecular weight cut-off 12,000 Da).[@CIT0043] Release media consisted of acetate buffer pH 3.3 and Milli-Q water in the proportion 50:50 containing 0.5% (v/v) Tween-80. Free DIT solution (2 mg in 2 mL 1% DMSO) and DIT@ \u03b2-CDs/CeO~2~ NPs (corresponding DIT amount was 2 mg, 2 mL) were added separately into the dialysis bags and then immersed in 10 mL of release medium at 37\u00b0C with a shaking speed of 100 rpm. At predetermined time points, 500 \u00b5L of sample was collected, and fresh media replacement was done after sampling. The quantity of DIT was measured using the Shimadzu Prominence high-performance liquid chromatography system and C 18 analytic column (Luna C18(2) 25 cm\u00d74.6 cm,5 mm, Phenomenex Inc, Torrance, CA). The mobile phase is a mixture of methanol and water (80:20, volume/volume ratio). The UV absorbance was measured at a wavelength of 289 nm, with a flow rate of 1.0 mL/min and a 25-\u00b5L injection volume. Standard curve has been measured by applying HPLC. All samples are prepared in triplicate.\n\nSOD- and Catalase-Mimetic Activity Assay {#S0002-S2007}\n----------------------------------------\n\nThe SOD-mimetic activity of the \u03b2-CDs/CeO~2~ NPs was investigated through its O~2~\u2022^\u2212^ scavenging activity with the Total Superoxide Dismutase Assay Kit with nitro-blue tetrazolium (NBT, Biotech, China) as described in the previous literature.[@CIT0044] In brief, 20 \u00b5L of the \u03b2-CDs/CeO~2~ NPs dispersion at different concentrations (0, 5, 10, 20, 40, 80, 160, and 200 \u00b5g/mL) was incubated with assay reagent containing xanthine, xanthine oxidase, and NBT into a 96-well plate at 37\u00b0C for 30 min. The absorbance at 560 nm was then recorded using a Multiskan GO microplate reader (Thermo Scientific, USA), and the O~2~\u2022^\u2212^ inhibition rate was calculated in accordance with the manufacturer's formula.\n\nCatalase-mimicking activity was measured with the catalase assay kit (Biotech, China). In brief, a H~2~O~2~ standard curve was drawn by different concentrations (0, 0.0025, 0.005, 0.01, 0.015, and 0.02 \u00b5M).[@CIT0045] Then, we prepared a blank control and a sample solution (0, 5, 10, 20, 40, 80, 160, 200 \u00b5g/mL) and reacted at 25\u00b0C for 5 min. A stop solution and color-developing solution were added, and then the solution was incubated at 25\u00b0C for 15 min. Then, the absorbance at 520 nm was recorded using a Multiskan GO microplate reader (Thermo Scientific, USA), and the catalase activity was calculated in accordance with the manufacturer's formula.\n\n3-(4,5-Dimethylthiazolyl)-2,5-Diphenyltetrazolium Bromide (MTT) Assay {#S0002-S2008}\n---------------------------------------------------------------------\n\nThe potential cytotoxicity of \u03b2-CDs/CeO~2~ NPs was investigated in MTT assay as we reported earlier.[@CIT0046] The HaCaT human keratinocyte cells line was selected to analyze cell viability in MTT assay (Sigma, MO, USA). HaCaT human keratinocyte cells were seeded into 96-well plates at a concentration of 1\u00d710^4^ cells/well and then incubated at 37\u00b0C and 5% CO~2~ for 24 h to ensure proper stability and adherence. Then, the culture medium was removed. The cells were incubated with fresh medium containing 100 \u00b5g/mL \u03b2-CDs/CeO~2~ NPs for 24 h before 20 \u00b5M H~2~O~2~ was added. After 24 h of cell culture, 10 \u00b5L of MTT (5 mg/mL) was added into each well. Incubation was continued for 4 h, and then 100 \u00b5L of DMSO was appended to each well to dissolve MTT crystals. The reactions were monitored using an iMark microplate reader (Bio-rad, CA, USA) at a wavelength of 490 nm. The experiment was repeated three times.\n\nCellular ROS-Scavenging Activity {#S0002-S2009}\n--------------------------------\n\nThe role of nanocarriers in scavenging ROS was investigated using 2',7'-dichlorodihydrofluorescein diacetate.[@CIT0046] The HaCaT human keratinocyte cells line was selected and seeded into 12-well plates at a concentration of 4\u00d710^4^ cells/well. Cells were incubated at 37\u00b0C and 5% CO~2~ for 24 h to ensure proper stability and adherence, and then the culture medium was removed. Then, \u03b2-CDs/CeO~2~ NPs (100 \u00b5g/mL) were added in the different wells, except in the positive control wells. After 24 h, serum-free medium (1 mL) with 20 \u00b5M H~2~O~2~ was added to each well, except the blank hole. After 6 h, the wells were added with 2',7'-dichlorodihydrofluorescein diacetate (ROS fluorescence probe), incubated for another 20 min, and then observed under fluorescence microscopy.\n\nExperiments on BALB/c Mice {#S0002-S2010}\n--------------------------\n\n### IMQ-Induced Plaque-Like Mouse Model of Psoriasis {#S0002-S2010-S3001}\n\nIMQ-induced psoriasis-like mouse model was established as previously described.[@CIT0047] Male BALB/c mice (6--8 weeks old) were fed a standard diet and provided with free water at room temperature. IMQ is a TLR7/8 ligand and a potent immune activator. A psoriasis model was established as previously described with slight modifications. In specific, IMQ was administered locally to induce and aggravate psoriasis lesions. After 1 week of adaptive feeding, the back shaving area was 2 cm\u00d73 cm. IMQ cream (5%) was applied to the shaved area at the back of the mouse every 24 h for 8 consecutive days (50 mg per mouse), the skin lesions were observed by taking photos every day, and the PASI score was obtained. Animals were randomly grouped as follows: Group 1 (normal mice), Group 2 (mice with IMQ treatment only and induced psoriasis), Group 3 (mice treated with 1 mg of DIT-loaded \u03b2-CDs/CeO~2~ NPs), Group 4 (mice treated with 0.04 mg of DIT), Group 5 (normal mice treated with 1 mg of DIT-loaded \u03b2-CDs/CeO~2~ NPs), and Group 6 (halometasone cream as positive control, 0.1 mg per mouse daily). The second group served as a negative control group, and the sixth group served as a positive control group.\n\n### Evaluation of PASI {#S0002-S2010-S3002}\n\nTo assess the severity of lesions at the back of the psoriatic mice, we followed the clinical criteria of PASI and established an objective scoring system to evaluate the degree of inflammation on the back lesions of the mice.[@CIT0047] Erythema, scaling, and thickening were scored independently by two people on a scale of 0 to 4 (0, none; 1, slight; 2, moderate; 3, marked; and 4, very marked). The scoring was performed every day for 8 days.\n\n### Histopathology of Mouse Back Skin Lesion {#S0002-S2010-S3003}\n\nThe induced dorsal skin lesions were collected at the end of the 8th day, soaked, and then fixed with 4% paraformaldehyde. The fixed samples were dehydrated and embedded in paraffin. Then, 4 \u00b5m microtome sections of the skin were deparaffinized, rehydrated, and stained with hematoxylin and eosin (H&E) as described previously.[@CIT0047] Histopathological sections were observed under a microscope (BN-DC-RGB500, Nanjing, China).\n\n### Weight Ratio of Spleen to Body (Spleen/Body Wt%) {#S0002-S2010-S3004}\n\nThe spleen is the largest immune organ in the human immune system. The weight of the spleen is a sensitive indicator of the body's immune status. The increase in spleen/body mass fraction may reflect the increase in the number of immune-related cells in the spleen, which is associated with diseases related to inflammation and immune activation.[@CIT0043] The body weight of the mice was recorded, and then the spleen of the mice was dissected at the end of the eighth day. Finally, the spleen/body wt % was obtained.\n\n### Immunofluorescence Staining {#S0002-S2010-S3005}\n\nTumor necrosis factor (TNF)-\u03b1 plays a key role in the pathogenesis of psoriasis. The levels of TNF-\u03b1 were detected by immunofluorescence staining of lesions to explore the therapeutic mechanism of DIT-loaded \u03b2-CDs/CeO~2~ NPs in the psoriatic mice. Immunofluorescence staining was performed by a routine staining method.[@CIT0020] Rabbit polyclonal antibody to TNF-\u03b1 (1:150, Abcam) was used as the primary antibody, and anti-rabbit antibody (1:250, Proteintech) was used as the secondary antibody.\n\nStatistical Analysis {#S0002-S2011}\n--------------------\n\nData were analyzed with Statistical Package for the Social Sciences version 16.0 software (SPSS Inc., Chicago, IL, USA). The experiments were carried out in triplicates, and values are expressed as mean \u00b1 standard deviation (SD). Student's *t*-test was conducted to determine significance. Statistical significance was considered at a probability level of p \\< 0.05.\n\nResults and Discussion {#S0003}\n======================\n\nSynthesis and Characterization of \u03b2-CDs-Capped CeO~2~ NPs {#S0003-S2001}\n---------------------------------------------------------\n\nGenerally, CeNPs are synthesized in the organic phase and require further surface modification with hydrophilic ligands for biomedical applications.[@CIT0048] In the present study, the \u03b2-CDs/CeO~2~ NPs were synthesized by a hydrothermal method using unmodified \u03b2-CDs as a protecting agent.[@CIT0042] Several recent studies used \u03b2-CDs as a metal NP stabilizer and size-control agent due to its strong hydroxyl binding to the surface of metal NPs.[@CIT0049]--[@CIT0051] The lipophilic drug (DIT) was loaded in the hydrophobic inner cavity of \u03b2-CDs via host--guest interactions.[@CIT0040],[@CIT0041] The drug EE of DIT was 94.7%, indicating that DIT was highly exhibited in these NPs and the corresponding DL reached to 3.48%.\n\nThe particle size distribution and PDI of the blank and drug-loaded NPs were determined using dynamic light scattering. The average particle size of the blank \u03b2-CDs/CeO~2~ NPs was 60.89\u00b10.32 nm with a PDI of 0.12 ([Figure 1A](#F0001){ref-type=\"fig\"}), whereas that of the DIT-loaded NPs was 79.38\u00b11.06 nm with a PDI of 0.27 ([Figure 1B](#F0001){ref-type=\"fig\"}). TEM results showed that the as-prepared NPs formed a uniform quasi-spherical shape with low polydispersity, and the diameter was consistent with the above particle size measurement results and without obviously visible aggregation among the particles ([Figure 1](#F0001){ref-type=\"fig\"} C and D). The enzyme-mimicking activities of nano-cerium are due to the presence of mixed oxidation states.[@CIT0032],[@CIT0048] Thus, XPS was performed to determine the mixed oxidation states. In the XPS spectrum ([Figure 1E](#F0001){ref-type=\"fig\"}), the peaks at 883.5, 900.1, and 917.5 eV corresponded to Ce ^4+^, whereas the peaks at 889.5, 903.1, and 908.1 eV were related to Ce^3+^. The results indicate that the synthesized NPs have a mixed-valence state (Ce^3+^/Ce^4+^).[@CIT0021],[@CIT0031],[@CIT0032]Figure 1Hydrodynamic diameters of \u03b2-CDs/CeO~2~NPs (**A**) and DIT\\@\u03b2-CDs/CeO~2~ NPs (**B**) measured by DLS; TEM (scale bar = 100 nm) images of \u03b2-CDs/CeO~2~NPs (**C**) and DIT\\@\u03b2-CDs/CeO~2~ NPs (**D**) dispersed in water; (**E**) XPS analysis of \u03b2-CDs/CeO~2~ NPs for Ce^3+^ and Ce^4+^ showing the binding energy levels; (**F**) Fourier transform infrared spectra of \u03b2-CDs, ceric nitrate, DIT, \u03b2-CDs/CeO~2~ NPs and DIT@ \u03b2-CDs/CeO~2~ NPs.\n\nThe bio-composite complex formation of the synthesized particles was further attested through Fourier transform infrared spectroscopy by analyzing the characteristic peaks involved. As shown in [Figure 1F](#F0001){ref-type=\"fig\"}, the presence of \u03b2-CDs was confirmed on the nanoceria particle surface. The main peak at 3419 cm^\u22121^ was associated with OH stretching, and the peak at 3448 cm^\u22121^ in the \u03b2-CDs/CeO~2~ NPs involved the interaction between nanoceria and OH.[@CIT0049] Bands appeared at 1368 cm^\u22121^ (bending mode of CH~2~), 1157 cm^\u22121^ (asymmetric C-O-C stretching), and 1029 cm^\u22121^ (C-O stretching) in free \u03b2-CD and in the \u03b2-CDs/CeO~2~ NPs. Compared with free \u03b2-CDs, the decreased relative intensity of the bands at 937 cm^\u22121^ (skeletal vibration of 1.4 link bond), 755 cm^\u22121^ (ring vibration), and 707 cm^\u22121^ (pyranose ring vibration) in the \u03b2-CDs/CeO~2~ NPs were attributed to the \"fixed\" nature of the \u03b2-CDs on the nanoceria surface that prevented the pyranose ring and skeletal vibration.[@CIT0050] Because of the existence of \u03b2-CDs, the NPs can be dispersed efficiently into the water solutions even at a high concentration of 15 mg\u2219mL^\u22121^, the sediments are observed clearly in the samples without of \u03b2-CDs ([Figure 2A](#F0002){ref-type=\"fig\"}). This result implies that the introduction of \u03b2-CDs on the surface of CeNPs increased their water solubility. After DIT was loaded, the characteristic peaks of DIT at 1614 and 1446 cm^\u22121^ (aromatic ring skeletal vibrations) and 1073, 929, and 747 cm^\u22121^ (C-H in- and out-plane bending) were found in the DIT\\@\u03b2-CDs/CeO~2~ NPs. This finding confirmed the presence of DIT in the NPs.Figure 2(**A**) Photographs of 15 mg/mL CeO~2~ with or without \u03b2-CDs; (**B**) The in vitro release profile of dithranol from \u03b2-CDs/CeO~2~ NPs; suspension of superoxide anions (**C**) and scavenging activities of H~2~O~2~ (**D**) by \u03b2-CDs/CeO~2~ NPs.\n\nThe drug release behavior of DIT@ \u03b2-CDs/CeO~2~ NPs in vitro was studied and release profiles are shown in [Figure 2B](#F0002){ref-type=\"fig\"}. The figure clearly indicates almost 85.3\u00b11.1% DIT were released from free DIT solution in 8 h, whereas DIT release from the NPs was comparatively slow, it showed 73.3 \u00b1 1.6% in 60 h. It is evident that DIT@ \u03b2-CDs/CeO~2~ NPs exhibited an obvious sustained-release profile with no evident burst effects. The pronged-release of DIT from NPs can be explained on the basis that the drug is encapsulated into the hydrophobic \u03b2-CDs cavity by supra-molecular inclusion.\n\nSOD- and Catalase-Mimicking Activity of \u03b2-CDs/CeO~2~ NPs {#S0003-S2002}\n--------------------------------------------------------\n\nAs major antioxidant enzymes, SOD and catalase can protect organisms involved in the neutralization of ROS.[@CIT0052] The SOD- and catalase-mimetic activities of the \u03b2-CDs/CeO~2~ NPs were investigated. SOD, an important antioxidant enzyme in living organisms, can catalyze the deuteration of O~2~\u2022^\u2212^s to H~2~O~2~ and oxygen. SOD-mimicking activity was studied by the classical NBT chromogenic method. In brief, the O~2~\u2022^\u2212^ was produced by the xanthine and xanthine oxidase reaction systems, which further reduced to be detectable blue formazan at 560 nm.[@CIT0044] In the presence of \u03b2-CDs/CeO~2~ NPs, the absorbance of formazan considerably decreased due to its SOD-mimicking activity. As shown in [Figure 2C](#F0002){ref-type=\"fig\"}, the \u03b2-CDs/CeO~2~ NPs possessed catalase-mimicking activity in a dose-dependent manner. The O~2~\u2022^\u2212^ inhibition rate reached to 79.4% in the presence of 200 \u00b5g/mL dispersion.\n\nConsidering that H~2~O~2~ is a common form of ROS,[@CIT0052] we studied the H~2~O~2~-scavenging activity of the \u03b2-CDs/CeO~2~ NPs. Such H~2~O~2~ scavenging activity can be attributed to the catalase-mimicking activity of the nanoparticles.[@CIT0045] When H~2~O~2~ is sufficient, catalase can catalyze H~2~O~2~ to produce water and oxygen. The remaining H~2~O~2~ coupled with a substrate and catalyzed by peroxidase to generate a red product, (N-(4-antipyryl)-3-chloro-5-sulfonate-*p*-benzoquinone monoimine), would be detected at 520 nm.[@CIT0045] As expected, the concentration of H~2~O~2~ decreased with increasing concentration of \u03b2-CDs/CeO~2~ NPs (0--200 \u00b5g/mL), as shown in [Figure 2D](#F0002){ref-type=\"fig\"}. The H~2~O~2~ eliminated efficiency reached about 50% in the presence of 40 \u00b5g/mL \u03b2-CDs/CeO~2~ NPs. All of the above results clearly demonstrated that the \u03b2-CDs/CeO~2~ NPs could play as SOD and catalase mimics. CeNPs possess SOD- and catalase-mimicking activities due to the mixed valence states of Ce^3+^ and Ce^4+^.[@CIT0021],[@CIT0031],[@CIT0032]\n\nIntracellular ROS Scavenging Detection {#S0003-S2003}\n--------------------------------------\n\nConsidering their potential SOD- and catalase-mimetic activities, we investigated the in vitro antioxidant activity of the \u03b2-CDs/CeO~2~ NPs for protecting cells from oxidative damage using the HaCaT cells line as a model. First, the possible toxicity of the \u03b2-CDs/CeO~2~ NPs was evaluated through MTT assay. The viability of the cells treated with \u03b2-CDs/CeO~2~ NPs (0--250 \u00b5g/mL) in the experimental conditions was not obviously altered compared with the control groups ([Figure 3A](#F0003){ref-type=\"fig\"}). Then, we examined the effectiveness of the \u03b2-CDs/CeO~2~ NPs for protecting cells from oxidative damage. As shown in [Figure 3B](#F0003){ref-type=\"fig\"} the viability of the cells incubated with 50 \u00b5M H~2~O~2~ for 24 h was reduced to about 60%, whereas pretreatment with \u03b2-CDs/CeO~2~ NPs (100 and 200 \u00b5g/mL) can prevent cellular damage triggered by H~2~O~2~. Furthermore, the intracellular ROS level was monitored using 2',7'-dichlorofluorescein diacetate as the fluorescence probe.[@CIT0046] As shown in [Figure 3C](#F0003){ref-type=\"fig\"}, a negligible fluorescence signal was observed when the cells were incubated with \u03b2-CDs/CeO~2~ NPs compared with the control experiments. By contrast, high fluorescence signals were observed when the cells were incubated with H~2~O~2~. When the cells were pretreated with \u03b2-CDs/CeO~2~ NPs the fluorescence intensity of the H~2~O~2~-treated cells was significantly reduced, indicating the effective intracellular ROS scavenging activity of \u03b2-CDs/CeO~2~ NPs. Oxidative stress plays a pivotal role in the pathogenesis of psoriasis.[@CIT0013] Previous studies also found that antioxidants can improve the symptoms of psoriasis.[@CIT0010],[@CIT0014]-[@CIT0019] These results suggest that our system has a potential application in psoriasis therapy by suppressing ROS and protecting cells from oxidative stress.Figure 3(**A**) Viabilities of HaCaT incubated with varied concentrations of \u03b2-CDs/CeO~2~ NPs for 48 h; (**B**) HaCaT viabilities for the protective capabilities of \u03b2-CDs/CeO~2~ NPs by different doses after being treated with 50 \u00b5M H~2~O~2~; (**C**) the fluorescence images of HaCaT cells after various treatments with H~2~O~2~ (20 \u00b5M) only, \u03b2-CDs/CeO~2~ NPs (100 \u00b5g/mL) only, H~2~O~2~ and \u03b2-CDs/CeO~2~ NPs stained with DCFH-DA.\n\nAnti-Psoriatic Efficacy in BALB/c Mice {#S0003-S2004}\n--------------------------------------\n\nConsidering that the \u03b2-CDs on the CeO~2~ NPs surface can be used as hydrophobic drug carriers, we loaded DIT, a classic psoriasis treatment drug, and investigated its anti-psoriatic efficacy in the IMQ-induced mouse model. For the convenience of administration, carbopol gel was chosen as the matrix of the DIT\\@\u03b2-CDs/CeO~2~ NPs. After initiation of IMQ and DIT\\@\u03b2-CDs/CeO~2~ NPs treatment, we calculated the PASI, performed H&E staining, and then calculated the weight ratio of spleen to body to evaluate the anti-psoriatic effect. The erythema, scales, and thickness of PASI scored from 0 to 4 are shown in [Figure 4A](#F0004){ref-type=\"fig\"}-[D](#F0004){ref-type=\"fig\"}. The mice treated with DIT\\@\u03b2-CDs/CeO~2~ NPs had lower mean scores for erythema, scales, and skin thickness than the mice treated with IMQ or DIT alone at the end of day 8. This result suggests that the DIT\\@\u03b2-CDs/CeO~2~ NPs exerted a therapeutic effect on psoriasis and IMQ-induced skin inflammation.Figure 4Psoriasis Area and Severity Index (PASI) scoring of psoriatic dorsal regions of mice in different groups (n=5) were evaluated for 8 days, including erythema (**A**), scaling (**B**), infiltration of the mouse skin (**C**), with a scale from 0 to 4. The total score (**D**) was from 0 to 12.\\*p \\< 0.05, compared with IMQ only group (n=5); The ratio of spleen weight to body weight (**E**) were recorded of IMQ and different treatment in different groups.\n\nThe histopathology of the skin samples was performed to further confirm the anti-psoriatic effect of the DIT\\@\u03b2-CDs/CeO~2~ NPs. The phenotypic and the H&E-stained images of skin from different groups are shown in [Figure 5](#F0005){ref-type=\"fig\"}. Severe erythema covered with white scales and marked inflammatory infiltration can be observed in the mice treated with IMQ compared with the normal mice. After treatment with different prescriptions, especially the IMQ+DIT\\@NPs group and IMQ+Halometasone (positive control) group, white scale and erythema of inflammatory skin were significantly reduced and consistent with the visual observation. These observations showed that the IMQ+ DIT\\@NPs were highly effective in alleviating the symptoms of IMQ-induced psoriasis in mice.Figure 5Topical application of IMQ induced psoriatic like change on the dorsal skin of mice. At day 8, photographs were taken from the mice (**A**) and H&E staining (**B**) was performed to observe the differences of the treatment; (**C**) Enlarged view of normal mice group, the IMQ only group and IMQ+ DIT\\@\u03b2-CDs/CeO~2~ NPs were also presented for annotation.\n\nThe spleen is the largest organ in the body's immune system, and the increased spleen/body wt% is an indicator reflecting the enhancement of immune activation-related diseases.[@CIT0043] After 8 days of application with IMQ and different prescription, the mice were sacrificed and the weight ratio of spleen to body was calculated ([Figure 4E](#F0004){ref-type=\"fig\"}). As expected, the weight ratio of spleen to body was the lowest in the normal mouse group and the highest in the IMQ-only group. Those of the IMQ+DIT\\@NPs group and IMQ+Halometasone (positive control) group were between the normal mouse and IMQ-only groups. The calculated spleen/body weight% of the IMQ group (1.02 \u00b1 0.26) was approximately three times larger than that of the normal mouse group (0.31 \u00b1 0.05), and the results showed that the number of cells in the spleen significantly increased. The spleen/body weight% of the IMQ+DIT\\@NPs group was significantly lower than that of the IMQ-only group (p\\<0.01). In addition, the PASI score and spleen/weight % results indicated that the DIT\\@NPs was more effective than the DIT treatment. Overall, these results indicate that the dermal administration of the DIT\\@\u03b2-CDs/CeO~2~ NPs effectively alleviated psoriasis, suggesting that the NPs may be used as a potential therapy for psoriasis.\n\nPsoriasis is a multifactorial disease with myriads of inflammatory mediators.[@CIT0001] For instance, tumor necrosis factor (TNF)-\u03b1 is overexpressed in psoriatic skins. We observed the effects of the NPs on the expression of TNF-\u03b1 using immunofluorescence. As shown in [Figure 6](#F0006){ref-type=\"fig\"}, TNF-\u03b1 expression obviously increased in the IMQ-treated mice compared with the control groups. However, the NPs significantly alleviated the TNF-\u03b1 upregulation induced by IMQ. These findings suggest that the NPs display anti-psoriatic activities in vivo.Figure 6The representative staining of TNF-\u03b1 in psoriatic skins after different treatments including the healthy skin of normal mice, the diseased skin of psoriasis mice, the skin of psoriatic mice treated with DIT, DIT\\@NPs and Halometasone.\n\nConclusion {#S0004}\n==========\n\nIn summary, we fabricated DIT\\@\u03b2-CDs/CeO~2~ NPs exhibiting multi-enzyme mimic activity and functionally drug-loaded activities, which would rescue cells under oxidative stress and provide synergistic anti-psoriatic effects. The DIT could effectively be encapsulated in the inner cavity of \u03b2-CDs with a DL capacity of 3.48% and an EE of 94.7%. The \u03b2-CDs/CeO~2~ NPs could effectively scavenge O~2~\u2022^\u2212^ and H~2~O~2~ and provide remarkable cryoprotection against ROS-mediated damage. More importantly, the DIT\\@\u03b2-CDs/CeO~2~ NPs provide an excellent therapeutic effect in IMQ-induced psoriatic model on the basis of morphological evaluation, PASI calculation, and inflammatory cytokine (TNF-\u03b1) expression. This study paves the way toward the application of nanozyme \u03b2-CDs/CeO~2~ NPs as a powerful tool for psoriasis therapy.\n\nThe authors are grateful for the generous financial support from the National Natural Science Foundation of China (No. 81973671, 21901186), the Natural Science Foundation of Shandong Province, China (ZR2019BB032), Project of Shandong Province Higher Educational Science and Technology Program (No. J18KA279).\n\nAbbreviations {#S0005}\n=============\n\nROS, Reactive oxygen species; \u03b2-CD, \u03b2-cyclodextrin; NPs, nanoparticles; TEM, transmission electron microscopy; XPS, X-ray photoelectron spectroscopy; FT-IR, Fourier transform infrared spectrometry; SOD, superoxide dismutase; DIT, dithranol; IMQ, imiquimod; PASI, Psoriasis Area and Severity Index; PDI, polydispersity index (PDI); DL, drug loading; EE, entrapment efficiency; NBT, nitro-blue tetrazolium; MTT, 3-(4,5-dimethylthiazolyl)-2,5-diphenyltetrazolium bromide (MTT); TNF-\u03b1, tumor necrosis factor-\u03b1; H&E, hematoxylin and eosin; SD, standard deviation.\n\nDisclosure {#S0006}\n==========\n\nThe authors report no conflicts of interest in this work.\n\n[^1]: These authors contributed equally to this work\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nHIV is a retrovirus, characterized by inserting its genomic DNA into the human genome, followed by the phenotypes of acute, chronic, or latent infection based on the interactions of the viral DNA with a host DNA. Once an HIV DNA is inserted into a host genome, no known immune mechanism so far eliminates the viral DNA from the host genome. Several cellular mechanisms, however, govern the HIV DNA expression after the integration, which regulate retroviral replication and thereby control the disease phenotypes or symptoms of an acute, chronic, or latent infection, including the cellular mechanisms that silence the replication of ancient human endogenous retroviruses (HERVs) \\[[@B1]--[@B10]\\]. Development of highly active antiretroviral therapy (HAART) or combination antiretroviral therapy (cART) has changed the natural course of HIV infection. HAART effectively controls the HIV entry, reverse transcription, integration, package, and even release, except for a direct control of the HIV DNA expression \\[[@B10]\\].\n\nCD4 T-cell is the target cell of HIV infection. The status of CD4 T-cells, specifically memory CD4 T-cells after HAART, determines the patient anti-HIV immunity, clinical status, and prognosis. HIV DNA expression in memory CD4 T-cells directly governs the activities of an HIV reservoir or the kinetics of the viral reservoir. Moreover, recent studies reveal that memory CD4 T-cells have stem cell properties and preferentially reside and rest in the bone marrow niche \\[[@B11]--[@B20]\\]. Bone marrow, plus stromal cells, and immune cells comprise a niche where hematopoietic stem cells (HSC) reside. Bone marrow is also a niche of hematopoietic progenitor cells (HPC) and now a niche of memory CD4 T-cells and other immune cells \\[[@B11]--[@B20]\\]. Furthermore, the effect of HIV infection on the niche, HSC, HPC, or memory CD4 T-cells has been addressed repeatedly since 1980s. Therefore, we now give this topic a new meaning in line with the functions of the niche and residing cells in a chronic HIV infection after HAART, specifically on their roles in the eradication of HIV and the cure of AIDS.\n\n2. Chronic HIV Infection {#sec2}\n========================\n\nChronic viral infection, by definition, belongs to the category of persistent infection, involves stages of both insidious and productive infection without rapidly killing or even producing excessive damage of the host cells. The other two types of persistent viral infections or persistent virus-host interactions are latent infection and slow infection. The natural course of HIV infection has been identified by using an antiviral drug \\[[@B69], [@B70]\\]. Without HAART, HIV develops an acute infection in a host and destroys millions of cells per day, specifically CD4 T-cells, among them, the memory CD4 T-cells \\[[@B69], [@B70]\\]. Memory CD4 T-cells have stem cell properties, which supply millions of cells per day via their clonal expansion to fight the invading pathogens. Same as in other viral infections but unlike the others, memory CD4 T-cells dutifully and diligently conduct their clonal expansion and replenish millions of effector cells to fight with the HIV per day. Nonetheless, all these cells turn into fuel to speed up the HIV replication until the memory CD4 T-cell pool is exhausted, by which a chronic infection follows.\n\nWith the inception of HAART, the rapid HIV replication in CD4 T-cells is curbed in multiple steps of the viral lifecycle, except on the viral DNA expression \\[[@B10]\\]. Moreover, the application of HAART pushes the kinetics of HIV infection further into a chronic infection. This not only saves and increases the memory CD4 T-cell pool but also leaves an HIV reservoir based on the feature of a retroviral infection. This viral reservoir is further consolidated when the main stimuli of HIV replication in CD4 T-cells are subdued by HAART, coincidently followed by a deceased clonal expansion of memory CD4 T-cells and a decreased differentiation of effector cells due to the greatly deceased secretions of growth/clonal factors, cytokines, and chemokines, which allow the memory CD4 T-cell to go back to its resting stage \\[[@B10]--[@B20], [@B38], [@B44]\\].\n\nIt is well known that the essence of adaptive immunity rests on its memory function, manifested mainly by memory CD4 T-cells. In HIV infection, one single memory CD4 T-cell against HIV expands to an anti-HIV clone, supplying millions of effector cells to regulate both cellular and humeral even innate immunities against the HIV infection. In contrast to the natural course of chronic HIV infection in quiescent cells, including memory CD4 T-cells and macrophages, the HAART resulted chronic HIV infection may allow a larger pool of memory CD4 T-cells to harbor the HIV DNA than in a natural HIV chronic infection occurring after the CD4 T-cell exhaustion. The potential pool of memory CD4 T-cells harbor HIV DNA, however, is various and dependent on when, how, and whom HAART is applied to, as well as the genetic derivations of an individual in his/her memory CD4 T-cell clonal formation during the HIV infection. A further elucidating the molecular mechanism of the interplay among memory CD4 T-cell clonal expansion, effector cell differentiation, and HAART application leads to a gateway towards reconstitution of patient anti-HIV immunity, eradication of HIV, and a cure of AIDS.\n\n3. Is HSC an HIV Reservoir? {#sec3}\n===========================\n\nTwo points are here crucial for answering this question: what is HSC and what is an HIV reservoir? First, hematopoietic stem cell (HSC) is an adult or tissue stem cell, embodying multipotentiality and self-renewal function. A single HSC can give rise to all lineages of blood and immune cells, reconstituting not only an entire blood system but also the bone marrow niche. Studies in mouse models and bone marrow transplantations in patients have demonstrated this for four decades \\[[@B71]--[@B79]\\]. The methods and techniques that prove HSC multipotentiality have been used to study other types of stem cells, either totipotent or other multipotent stem cells \\[[@B71]--[@B77], [@B21]--[@B66]\\]. One of the methods and technologies is to utilize the DNA marker to define the daughter or progeny cells derived from HSC, which is now used in gene and cell therapies of diseases. Investigators have used different DNA markers or vector transductions of HSC for gene therapy, including but not limited to the treatment of HIV/AIDS. Such studies are represented by the treatment of the Berlin patient and others. The bone marrow transplantations of CCR5-\u039432 stem cells replenish all of the patient CD4 T-cells and bring a cure \\[[@B73], [@B74], [@B79]\\]. The vector transduced HSC has shown lineage differentiation and exhibited anti-HIV effects in its progeny cells detected in the peripheral blood of all experimental subjects \\[[@B75]--[@B78]\\]. Based on the same technologies and principles, on the other hand, if a HIV DNA has been detected in HSC, the HIV DNA is also to be detected in all lineage blood cells or endpoint cells*in vitro* by differentiation experiments, or to be detected in multiple blood and immune cells*in vivo* through the patients. No studies, however, have shown such a result, in contrast to the fact that vector transduced HSC has shown anti-HIV effects in terminal differentiated CD4 T-cells and CCR5 stem cells replenish the entire patient lineage immune cells with a transduced unique DNA marker, CCR5-\u039432 \\[[@B73]--[@B79]\\].\n\nNext, what is an HIV reservoir? Viral reservoir is an anatomical site in which viruses accumulate and persist. HIV reservoir is defined as a cell type or anatomical site where a replication-competent form of the virus accumulates and persists, with more stable kinetic properties than the main pool of actively replicating virus. The same as the other viral reservoirs, HIV reservoir shows the feature of a persistent or chronic infection, specifically under cART or HAART. In other words, an HIV reservoir is a cell type that allows persistence of replication-competent HIV-1 on a timescale of years in patients on optimal antiretroviral therapy \\[[@B44], [@B40]--[@B43]\\]. Since 1980s, investigators have been studying the relationship of cell types and HIV infection. Whether HSC is an HIV reservoir, however, has only been addressed recently. Unlike the reports on whether or not HIV causes AIDS, the reports on whether or not HSC is a reservoir are regarding two concepts. There appears to be a different standard on assay cells and on definition of an HIV reservoir. It is scientifically important to use one standard on what is HSC and what is progenitor cell, in both experimentation and conclusion. Specifically, studies on progenitor cells are not suitable to reach a conclusion for stem cells. Second, there is a fine definition of the retroviral DNA or a retrovirus and a quiescent host cell and a viral reservoir \\[[@B38], [@B44], [@B40]--[@B52]\\]. The key point here, in the context of HIV cure, is clearly that HSC is not a major HIV reservoir based on the present studies \\[[@B38], [@B44], [@B40]--[@B52]\\].\n\nUpon reviewing the literature, like other investigators, we have found out that, up to date, there is a lack of experimental data to show that HIV actively replicates in HSC. Second, there is a lack of studies showing that HSC as an HIV reservoir stably and kinetically provides replication competent virus more than the main pool of actively replicating virus. Third, there is a lack of scientific evidence that HIV DNA in HSC is detected in its multilineage endpoint differentiated cells. Certain reports show that bone marrow cells expressing CD34 phenotypic marker contain viral particle or CD133 hematopoietic progenitor cells harbor HIV DNA. The same studies, nonetheless, conclude that bone marrow, not HSC, may serve as a potentially important reservoir of HIV-1, or CD133 hematopoietic progenitor cells (HPC) harbor HIV genomes, in sharp contrast to concluding that HSC is an HIV reservoir \\[[@B44], [@B45], [@B48], [@B49], [@B52]\\]. Moreover, other studies have shown that, in addition to HSC, many lineage progenitor cells including but not limited to CD4 progenitor cells reside in bone marrow \\[[@B11]--[@B20]\\]. CD133, also dubbed AC133, may be a good marker for the selection of human placental cord blood stem cells*in vitro.*Nevertheless, the same study shows that the freshly isolated cord blood CD34+AC133+ stem cells are not susceptible to HIV-1 infection and may not be a viral reservoir \\[[@B52]\\]. Currently, only latently infected resting CD4+ T-cells fit the proposed definition of a reservoir, and more evidence is necessary to demonstrate that other cell types, including hematopoietic stem cells and macrophages, fit this definition. Aiming at an HIV eradication and AIDS cure, we and other investigators have proposed and now insist that the techniques and methodologies for studying of HSC should be utilized to study memory CD4 T-cells and other immune memory cells, peculiarly for translational research and collaborating multidiscipline study \\[[@B38], [@B44], [@B40]--[@B52]\\]. Through the well-established models and methods that identify HIV reservoirs, further research is urgently required on potential reservoirs in the central nervous system and the gut-associated lymphoid tissue \\[[@B40]--[@B43]\\].\n\n4. Antiviral Therapy and Anti-Inflammatory Drugs on HSPC {#sec4}\n========================================================\n\nHSC is a self-replenishing source of all blood and immune cells and in the highest hierarchy of blood cell differentiation. Next to HSC are multipotent progenitor cells, named MPP, or sometimes HPC. HSC gives rise to HPC. HPC have limited self-renewal ability and limited multipotentiality for differentiation into different blood cells compared to HSC. HSC and HPC have been dubbed HSPC. This is a more experimental or bench research term than HSC or HPC, respectively, representing a group of experimentally purified blood stem cells (HSC) and progenitor cells (HPC), due to technical and physiological limitations on separation of stem cells from blood cells and the amount of cells that can be used to study and repeat the experiments. Importantly, the HPC in the HSPC means the primitive progenitor cells, which are technically indispensable in separating HSC from blood cells by current bench purification techniques. The HPC in HSPC is defined as and means the preliminary primary progenitor cell and is definitely not the lineage progenitor cell that has no self-renewal ability and is at the much lower position in the hematopoietic differentiation hierarchy than HSC or MPP \\[[@B10], [@B44], [@B21]--[@B27], [@B28]\\].\n\nIt is generally recognized that HIV infection affects bone marrow stromal cells as well as the immune cells that reside in the bone marrow niche \\[[@B11]--[@B20], [@B38], [@B21]--[@B27], [@B28]--[@B39]\\]. Since bone marrow is not only the niche of HSPC, but also the niche of memory CD4 T-cells and other immune cells \\[[@B11]--[@B20], [@B38], [@B21]--[@B27], [@B28]--[@B39]\\], we will briefly address the restorative effects of antiviral therapy and anti-inflammatory drugs on HSC, memory CD4 T-cells, and the bone marrow niche, elucidating our point of view on how to utilize anti-inflammatory drug, immunotherapy, and multidiscipline approaches towards an HIV eradication and AIDS cure.\n\nHSC has a unique function on resisting the effects of many drugs. In other words, HSC is refractory to drug effects. This unique function of the blood stem cells is based on a protective mechanism that consisted of cell organelles functioning as a pump, which quickly pumps drugs out of HSC or out of niche, whereby no drugs can affect HSC on its function or cell-cycle status easily \\[[@B24]--[@B27]\\]. HAART or anti-inflammatory drugs plus immunotherapies that are effective on cells dwelling in the niche, specifically on memory CD4 T-cells, will benefit the niche and thereby improve the function of HSC in general, but not on HSC directly. Another intrinsic mechanism that protects the genome of HSC from damages caused by stress, radiation, and so forth, is the quiescent status of HSC, through an enhanced prosurvival gene expression and a strong activation of p53-mediated DNA damage repair responses, in which p21^CipWaf1Sdi1^ (p21) plays an important role \\[[@B24], [@B25], [@B64]--[@B66]\\].\n\nAntiretroviral therapy, anti-inflammatory therapy, and immunotherapy have shown synergistic effects in the treatment of HIV. Although the detailed mechanisms remain to be elucidated, administration of drugs synergistically will create a new type of combination therapy, which will speed the process in restoring the function of memory CD4 T-cells, HSC, and the niche, and directly contribute to reconstituting or reprogramming the patient immunity. cART remains under development. Many compounds in cART or HAART now have new forms or new members, such as Tivicay (dolutegravir), a new integrase inhibitor, and the new versions of Truvada and tenofovir, which are more effective on certain compartments but with less side effects compared to their old versions. New compound or biologics are now targeting HIV transcription, which was off target for cART but now a target of antiviral drugs specifically of biologics, including but not limited to a class of nucleobase-amino acid conjugates, Tanshinone II A, or cyclin T1 splice variant, targeting or binding to TAR and specifically inhibiting HIV genomic RNA transcription or exportation \\[[@B53]--[@B55]\\]. Antiviral biologics are a new type of drugs in the field, working on either genetic or epigenetic regulations. Biologics have been used in cancer treatment and now applied in anti-HIV therapy, such as anti-PDL1 MPDL3280A, MK-3475, BMS-936559, or broadly neutralizing antibody 3BNC117, reverberating that certain anticancer drugs are used for anti-HIV therapy since the beginning of AIDS epidemic in early 1980s \\[[@B56]--[@B58]\\]. Immunotherapy of AIDS is not new either, which was started before the discovery of HIV. In the beginning, immunotherapy is used to relieve the AIDS symptoms. After the development of HAART, immunotherapy combined with anti-inflammation drugs is applied to treat immune reconstitution inflammatory syndrome (IRIS). Now, immunotherapy combined with anti-inflammation drugs is under study for synergizing with HAART to reconstitute patient immunity.\n\nIRIS is an inflammatory reaction in HIV infected patients after the initiation of antiretroviral therapy, resulting from the restored immunity to specific infectious or noninfectious antigens. The anti-inflammatory drugs have been applied to improve or synergize the effects of cART. Although some anti-inflammatory drugs are now used in immunotherapy in treatment of both AIDS and cancer, the effects of these drugs on immunotherapy, specifically in reconstitution of patient immunity, remain to be further studied. The pivotal issues here are the following: Why does IRIS consist of only the restored immunities to pathogens other than HIV? What is the immune mechanism of a much weaker reaction to HIV combined with a retroviral rebound in the scheduled treatment interruption even after a prolonged HAART? Despite studies showing the synergy of anti-inflammatory drugs with cART, what is the underpinning mechanism? Importantly, what is the role of anti-inflammatory drugs and cART in reconstitution of patient anti-HIV immunity in which cART alone has already failed? We believe these questions are in the minds of many investigators, which cannot be resolved by solely applying the treatments to patients without deciphering the molecular mechanism, regardless of cART or HAART, immunotherapy or gene therapy or cell therapy, unless we perform basic study on the memory CD4 T-cells to pursue the answer, by virtue of the fact that memory CD4 T-cells are not only the target cells of HIV, but also the commander in chief of immune functions. Dysfunction of these cells causes acquired immune deficiency, whereas restoration of these cell functions lays down a foundation of immune reconstitution.\n\n5. Antiviral Therapy and Anti-Inflammatory Drugs on Memory CD4 T-Cell Function {#sec5}\n==============================================================================\n\nSimilar to defining how HIV destroys CD4 T-cells and causes AIDS via using an antiviral drug \\[[@B69], [@B70]\\], one can decipher the molecular pathway that restores memory CD4 T-cell function against HIV infection, reconstituting patient anti-HIV immunity*in vivo* via an anti-inflammatory drug or immunotherapy.\n\nThe hectic clonal expansion of memory CD4 T-cells in HIV infection is a double-edged sword in both AIDS pathology and cure, which not only allows cells to be killed by HIV but also allows cells to die evil via pyroptosis, in sharp contrast to the programmed cell death of immune replenishment via apoptosis ([Figure 1](#fig1){ref-type=\"fig\"}) \\[[@B10], [@B38], [@B64]--[@B66], [@B59]--[@B68]\\]. How to utilize anti-inflammatory drug or immunotherapy to reprogram memory CD4 T-cell function whereby to reconstitute patient immunity is an imperative task in both the development of AIDS vaccine and cure. Additionally, to elucidate the molecular effects of anti-inflammatory drug and therapy on the function of memory CD4 T-cell, specifically the organelles such as apoptotic body and exosome, will open an avenue to manufacture new immunotherapy drugs and remedies, which are applied not only to the cure of AIDS but also to the other immune diseases such as cancer.\n\nWe address this from three aspects briefly. First, we aim to define how to expand the memory CD4 T-cell clone specifically against HIV. Second, we define how big the memory CD4 T-cell repertoire is to execute the HIV specific immunity*in vivo*. Third, we define when and how to apply anti-inflammatory drugs to treat the AIDS symptoms, since, besides AIDS, these symptoms are also observed in patients with cancers and in other immune diseases. Additionally, we believe two issues need to be focused in a near future. One is how to utilize anti-inflammatory drugs to increase memory CD4 T-cell clonal expansion synergizing with HAART. Another is how to use anti-inflammatory drug and immunotherapy to reprogram patient antiretroviral immunity towards a cure of group of immune diseases such as AIDS, cancers, or other immune maladies \\[[@B10], [@B44], [@B65], [@B28], [@B80], [@B81]\\].\n\n6. Conclusion {#sec6}\n=============\n\nAligned with the progress in stem cell research, chronic HIV infection, and the treatment of AIDS, we have proposed our view on whether or not HSC is a major HIV reservoir. We address this based on the established scientific standards, concepts, published data, our own experimental results, and conclusions entailed. Consistent with many other investigators, here we again urge allocating resources to resolve the pressing challenges in HIV/AIDS eradication and cure and to reconstitute host anti-HIV immunity by harnessing well-established techniques and methodologies that have been applied in stem cell research and in viral reservoir study, specifically in defining the reservoirs in an HIV infection.\n\nFurthermore, little is known on the underpinning mechanisms of the currently applied treatments for HIV patients, specifically on immune cell renewal, clonal expansion, and differentiation, including but not limited to cART, immunotherapy, and gene therapy. Resources channeled into such studies will not only unveil the synergy, but also decipher the molecular cellular \"synaptic conjunctions\" among the varieties of treatment, opening a gateway to reconstitute patient anti-HIV immunity in a sharp contrast to IRIS.\n\nFinally, we propose aiming to resolve the key steps in memory CD4 T-cell clonal expansion. We have focused on applying multidisciplinary expertise and knowledge of stem cell biology, retrovirology, and immunology, from bench to bedside, to elucidate the interplay of clonal expansion and programed cell death in patients with AIDS, with or without HAART. We will address the mechanisms of formation of apoptosis bodies and exosomes in individuals with AIDS, under placebo or anti-inflammatory drugs or immunotherapies. Ultimately, we will bridge the eradication of HIV and cure of AIDS together, fulfilling a cause that reconstitutes or more accurately reprograms patient immunity toward a cure of not only AIDS but also cancer \\[[@B10], [@B38], [@B44], [@B40]--[@B39]\\].\n\nThe authors thank Dr. Stephen P. Goff for the valuable suggestions on this paper.\n\nConflict of Interests\n=====================\n\nThe authors declare that they have no conflict of interests regarding the publication of this paper.\n\n![A trinity of cure by CD4 T-cell turnover. P21, cytokine, and immunotherapy can affect all three processes indicated by A, B, and C. Through cell renewal and turnover, one can regulate CD4 T-cell pool size, reprogram T-cell immune memory repertoire, and modulate immune activation and function thereby to reconstitute patient immunity towards an HIV cure.](SCI2015-148064.001){#fig1}\n\n[^1]: Academic Editor: Laura Lasagni\n"}
+{"text": "Background {#Sec1}\n==========\n\nFatigue is among the most frequent complaints reported by patients with chronic illnesses \\[[@CR1]--[@CR4]\\] and has far-ranging, often debilitating consequences on their wellbeing and physical, emotional and social functioning \\[[@CR5]\\]. Although there is no consensus definition of fatigue, it is often described as 'a persistent, overwhelming sense of tiredness, weakness or exhaustion resulting in a decreased capacity for physical and\u2044or mental work \\[[@CR6]\\]. Fatigue is a subjective experience and is commonly assessed by means of patient-reported outcome measures (PROMs). PROMs are widely used today in evaluating the effects of illness and treatment on symptoms, functioning, and other outcomes from the patient's perspective \\[[@CR7]\\].\n\nCurrently, there are some 40 generic and disease-specific PROMs for assessing fatigue in use today \\[[@CR8]\\]. Most of these fatigue measures have been evaluated regarding various aspects of validity and reliability. Although these are important psychometric properties reflecting the quality of the measure, they are of little value in interpreting the meaning of scores derived from that measure \\[[@CR9]\\]. Nonetheless, interpretation of scores, in particular changes in scores, is of critical concern in trials evaluating effects of treatments aimed at reducing fatigue, as well as in routine clinical practice in monitoring and managing fatigue in individual patients. In clinical trials, it has long been recognized that conventional statistical significance testing provides information regarding the probability that an effect exists, not about the meaningfulness of the size of the effect \\[[@CR10]\\]. In clinical practice, difficulties in evaluating and interpreting changes in PROM scores often impinge on their usefulness in informing clinical decision-making \\[[@CR11]\\].\n\nThe interpretation of changes in PROM scores may be enhanced by estimates of the so-called minimal important difference (MID). MID was originally defined over 25\u00a0years ago as \"the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient's management\" \\[[@CR12]\\]. During the past decades considerable research attention has been directed towards deriving MIDs for PROMS. In this pursuit a variety of methods have been developed and applied, but no clear consensus exists regarding which method or methods are most suitable.\n\nTo date, two main methods have been applied, namely anchor-based approaches and distribution-based approaches. Descriptions of these methods are beyond the scope of this paper and are summarized in detail elsewhere \\[[@CR13]\\]*.* Briefly, anchor-based approaches use various external criteria (patient-reported, physician-reported, or clinical anchors) to interpret whether a particular magnitude of change is important. For example, a common anchor-based method involves the use of global rating scales (GRS) where MIDs are derived by comparing patients' self-ratings of change (e.g., \"much worse\"---\"much better\") to change in PROM scores. The MID is often defined as lying within the range of \"slightly worse/better\" on the GRS \\[[@CR9]\\]. Distribution-based approaches rely on the statistical characteristics of the distribution of scores in the sample, in which the magnitude of change is generally expressed as a function of the standard deviation (SD) of scores alone or in combination with the reliability of the PROM (standard error of the measurement (SEM)) \\[[@CR14]\\]. Various SD and SEM cut-off values have been proposed for estimating MIDs, including \u00bd or 1/3 SD and 1--2 SEM. Another commonly applied method is the use of effect sizes (ES) or standardized response means (SRM), where change scores are divided by the SD at baseline or the SD of change, respectively. The MID is often defined as change values lying within the range of 0.2-0.5. A disadvantage to distribution-based approaches is that they do not address the clinical importance of the change. Recent recommendations have proposed that as a first-line method multiple anchor-based approaches should be used, which, supported by distribution-based methods, may be triangulated to a single MID value or smaller range of values \\[[@CR14]--[@CR17]\\].\n\nAlthough appealing for its simplicity, the idea of a single, universal MID value for any particular PROM remains elusive for a number of reasons. Firstly, different MID estimation approaches have been shown to yield highly disparate MIDs and hence triangulation (combining different methods to estimate a MID) may be problematic \\[[@CR18]\\]. Secondly, MIDs have also been shown to differ by population and context \\[[@CR14]\\]. For example, MIDs vary by diagnostic group, characteristics of the study sample, e.g., demographics and baseline levels; disease severity; treatment; choice of anchors \\[[@CR18], [@CR19]\\] as well as if MIDs gauge improvement versus deterioration \\[[@CR20]\\]. This variability suggests the need to understand how a particular MID value was determined in order to judge its appropriateness for use in research for interpreting change and/or computing sample sizes, or in clinical practice for monitoring fatigue in specific patient groups \\[[@CR21]\\].\n\nThe purpose of this paper is to compile published MIDs for fatigue PROMs, spanning diagnostic/patient groups and estimation methods, and to provide information relevant for appraising their appropriateness for use in specific clinical trials and in monitoring fatigue in defined patient groups in routine clinical practice.\n\nMethods {#Sec2}\n=======\n\nA systematic literature review where three databases (Scopus, CINAHL and Cochrane) were searched from January 2000 to April 2015 to identify studies with calculated MIDs in fatigue scales, subscales and single item measures. The searches were limited to English language (search string: \"minimal clinical important difference\\*\" OR \"minimal important difference\\*\" OR \"minimal clinically important difference\\*\" OR \"minimally important difference\\*\" OR \"clinical important improvement\\*\" OR \"clinically important improvement\\*\" OR \"minimal important clinical difference\\*\" OR \"minimally important clinical difference\\*\" OR \"responder definition\") AND Fatigue). The search was augmented with screening of article reference lists. All expressions including \"difference/change/improvement\" or equivalent, \"important\" as well as \"minimal\" or \"clinical\", or \"responder definition\" were defined as MIDs. To facilitate the reading all minimally important changes are called MIDs in this paper.\n\nSelection of articles {#Sec3}\n---------------------\n\nInclusion criteria were reporting MIDs in text and/or tables for a fatigue scale, subscale or single item measurement of fatigue. Exclusion criteria were: reported MID was not derived directly in the study; insufficient information supplied about the study sample, study design and/or method for determining the MID; study sample\u2009\\<\u200918\u00a0years, not separate reporting of MIDs for a fatigue subscale and conference abstracts. Exclusion on title/abstract and on full-text levels were done independently by two researchers (\u00c5N and AD), see Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}.Fig. 1Flowchart of selection of articles to include\n\nData extraction {#Sec4}\n---------------\n\nTwo authors (\u00c5N and AD) extracted data regarding MIDs and methods used, including anchors used. The last author (AD) checked all data extraction and prepared the tables. To facilitate interpretation all MIDs are shown as absolute values and decimals are restricted to one significant number only, except for effect sizes. Some studies reported standard deviation (SD) and confidence intervals but these are not shown in our tables or text. The fatigue measurements were identified as multidimensional scales, unidimensional scales or subscales, single item measurement or item bank scales.\n\nResults {#Sec5}\n=======\n\nThe literature search generated 177 articles (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}), of which 41 met the inclusion criteria \\[[@CR22]--[@CR62]\\]. The main reasons for exclusion were: reported MID was not derived in the study; and inadequate information was supplied about the study sample, study design and/or method for determining the MID. Many different expressions were used to name a small but important change in fatigue \\[[@CR13]\\]. In this review we included studies using different phrases for MID (see Table\u00a0[1](#Tab1){ref-type=\"table\"}), e.g. \"MID\", \"MCID\", \"MCII\" or an equivalent expression, all referred to as MID in this paper. Most of these expressions used some variation of \"difference/change/improvement\" or equivalent, \"important\" as well as \"minimal\". Some phrases also included \"clinical\". Two studies used \"responder definition\" \\[[@CR43], [@CR55]\\], see Table\u00a0[1](#Tab1){ref-type=\"table\"}. In two systematic reviews a phrase without \"minimal\" was used \\[[@CR59], [@CR60]\\] but the authors defined values for a small or minimal change.Table 1Included articles with reported fatigue MIDs, *n\u2009=*\u200941First author/countryInstrument/scaleTerm used^a^Bar\u00f3 et al. 2011 \\[[@CR22]\\]\\\nSpainPerform QuestionnaireMinimally important difference, MIDBedard et al. 2013a \\[[@CR23]\\]\\\nCanadaESAS fatigue itemMinimal clinically important difference, MCIDBedard et al. 2013b \\[[@CR24]\\]\\\nCanadaEORTC QLQ-30---Fatigue ScaleMinimal important difference, MIDBedard et al. 2014 \\[[@CR25]\\]\\\nCanadaEORTC QLQ-30---Fatigue ScaleMinimal important difference, MIDBjorner et al. 2007 \\[[@CR26]\\]\\\nUSA, MOS-studyVT/SF-36Minimally important difference, MIDBorghs et al. 2012 \\[[@CR27]\\]\\\nBelgiumQOLIE-3 Energy/fatigue subscaleMinimally important change, MICCella et al. 2002 \\[[@CR28]\\]\\\nUSAFACIT-Fatigue TOI-FMinimal clinically important difference, CID (MCID in short title)Cella et al. 2005 \\[[@CR29]\\]\\\nUSAFACT-FatigueMinimally important difference, MIDColangelo et al. 2009 \\[[@CR30]\\]\\\nCanadaFatigue VAS (0--100)Minimally important difference, MIDde Kleijn et al. 2011 \\[[@CR31]\\]\\\nNetherlandsFASMinimal (clinically) important difference, MCIDGeorge & Pope 2011 \\[[@CR32]\\]\\\nCanadaVAS fatigue (0--100)Minimal important difference, MIDGoligher et al. 2008 \\[[@CR33]\\]\\\nCanadaMFI, FSS, MAF, CFS, FACIT-F, VT/SF-36, GRSMinimal important difference, MIDKhanna et al. 2008 \\[[@CR34]\\]\\\nCanadaFatigue VAS (0--10)Minimally important difference, MIDKosinski et al. 2000 \\[[@CR35]\\]\\\nUSAVT/SF-36Minimally important change, MICKvam et al. 2010 \\[[@CR36]\\]\\\nNorwayEORTC QLQ-C30 Fatigue SubscaleMinimal important difference, MIDKwok and Pope 2010 \\[[@CR37]\\]\\\nCanadaFatigue VAS (0--100)Minimally important difference, MIDLai et al. 2011 \\[[@CR38]\\]\\\nUSAFACIT-Fatigue subscaleMinimally important difference, MIDLasch et al. 2009 \\[[@CR39]\\]\\\nUSASIS Energy/Fatigue and Mental Fatigue subscalesMinimum important difference, MIDMaringwa et al. 2011a \\[[@CR40]\\]\\\n17 countriesEORTC QLQ-C30 Fatigue SubscaleMinimal clinically important difference, MCIDMaringwa et al. 2011b \\[[@CR41]\\]\\\n12 countriesEORTC QLQ-C30 Fatigue SubscaleMinimal important difference, MIDMathias et al. 2009 \\[[@CR42]\\]\\\nUSA and EuropeITP-PACMinimally important difference, MIDMatza et al. 2013 \\[[@CR43]\\]\\\nUSAFAsDResponder definitionMills et al. 2012 \\[[@CR44]\\]\\\nUKNFI-MSMinimum clinically important difference, MCIDPatrick et al. 2003 \\[[@CR45]\\]\\\nUSAFACT-An fatigue subscaleMinimally important difference, MIDPouchot et al. 2008 \\[[@CR46]\\]\\\nCanadaMFI, FSS, MAF, CFS, FACIT-F, VT/SF-36, GRSMinimal clinically important difference, MCIDPurcell et al. 2010 \\[[@CR47]\\]\\\nAustraliaMFI subscalesMinimal clinically importantdifference, MCIDReddy et al. 2007 \\[[@CR48]\\]\\\nUSAFACIT-Fatigue subscale ESAS fatigue itemClinically important improvementRendas-Baum et al. 2010 \\[[@CR49]\\]\\\nCanadaFISMinimally important difference, MIDRobinson et al. 2009 \\[[@CR50]\\]\\\nUSAFSSMinimally important difference, MIDSchwartz et al. 2002 \\[[@CR51]\\]\\\nUSASCFS, POMS-F, single itemMinimally important clinical difference, MICDSch\u00fcnemann et al. 2005 \\[[@CR52]\\]CRQ/Fatigue subscaleMinimal important difference, MIDSekhon et al. 2010 \\[[@CR53]\\]\\\nCanadaFatigue VAS (0--100)Minimally important difference, MIDSpiegel et al. 2005 \\[[@CR54]\\]VT/SF-36Minimally clinically important difference, MCIDTwiss et al. 2010 \\[[@CR55]\\]\\\n8 countriesU-FISResponder definition, RDWard et al. 2015 \\[[@CR56]\\]\\\nUSAVT/SF-36Minimal clinically important improvement, MCIIWells et al. 2007 \\[[@CR57]\\]\\\nUSAFatigue VAS (0--100)Minimal clinically important difference, MCIDWheaton & Pope 2010 \\[[@CR58]\\]\\\nCanadaFatigue VAS (0--100)Minimal important difference, MIDWyrwich et al. 2003 \\[[@CR59]\\]VT/SF-36 CRQ/Fatigue subscaleClinically important difference, CIDWyrwich et al. 2004 \\[[@CR60]\\]VT/SF-36 CHQ/Fatigue subscaleClinically important difference, CIDYost et al. 2011 \\[[@CR61]\\]\\\nUSAPROMIS Fatigue (Fatigue-17, Fatigue-7)Minimally important difference, MIDZeng et al. 2012 \\[[@CR62]\\]\\\n7 countriesEORTC QLQ-C30 Fatigue SubscaleMinimal clinically important difference, MCID^a^All expressions in this column are referred to as \"MID\" in the current study\n\nThe included articles (*n\u2009=*\u200941) reported MIDs for 28 fatigue PROMs (characteristics shown in Table\u00a0[2](#Tab2){ref-type=\"table\"}), resulting in 60 studies/substudies of MIDs. The studies varied in sample size, diagnostic group, MID estimation approach, study design, type of intervention and length of follow up. Sample sizes ranged from *n\u2009=*\u200940 to *n\u2009=*\u20092,583. Sixteen different diagnoses were included in the reviewed studies. Twenty-seven of the studies in the 41 articles were longitudinal and follow-up periods ranged from two days after intervention to one year after baseline. An anchor-based approach alone was used in 39 of the 60 studies or substudies estimating MID, while the rest also used a distribution-based approach. Seventeen of these also included a method of triangulation to define MIDs. Two cross-sectional studies \\[[@CR33], [@CR46]\\] reported MIDs for seven fatigue or vitality scales (MFI, FSS, MAF, CFS, VT/SF-36, FACIT-F and GRS). Other studies determined MIDs for two or more fatigue measures or subscales \\[[@CR28], [@CR47], [@CR48], [@CR51], [@CR59]--[@CR61]\\]. Several PROMs had MIDs determined in a number of different studies and several studies reported MIDs for up to seven PROMs. Nevertheless, most MIDs were derived in single studies, with one study per PROM \\[[@CR22]--[@CR27], [@CR29]--[@CR32], [@CR34]--[@CR43], [@CR45], [@CR49], [@CR50], [@CR52]--[@CR58], [@CR62]\\], see Table\u00a0[3](#Tab3){ref-type=\"table\"}. Altogether, 60 studies or substudies estimating MIDs for global change (not specified direction of change), improvement and/or deterioration are described in Table\u00a0[3](#Tab3){ref-type=\"table\"}. In Table\u00a0[3](#Tab3){ref-type=\"table\"} all score changes are presented as positive values, regardless of the direction of change. Confidence intervals and SDs (if derived in study) are not shown. Numbers are rounded to one decimal place.Table 2Overview of reviewed fatigue scales, subscales and single fatigue item with published MIDs, *n\u2009=*\u200928Name of PROMAbbreviated nameNumber of itemsScore rangeCharacteristics of scaleChalder Fatigue ScaleCFS140--33MultidimensionalChronic Heart Failure Questionnaire, Fatigue subscaleCHQ41--7Unidimensional subscale\\\nNegative score^a^Chronic Respiratory Questionnaire, Fatigue subscaleCRQ41--7Unidimensional subscale\\\nNegative score^a^Edmonton Symptom Assessment System, Fatigue itemESAS Fatigue item10--10Single itemEuropean Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30, Fatigue SubscaleEORTC QLQ-30 Fatigue Subscale30--100Unidimensional subscaleFACIT-Fatigue ScaleFACIT-Fatigue130--52Unidimensional scale/subscale, Negative score^a^FACT-An, Fatigue subscaleFACT-An Fatigue200--80Unidimensional subscaleFatigue Assessment ScaleFAS1010--50UnidimensionalFatigue Associated with Depression QuestionnaireFAsD131--5MultidimensionalFatigue Impact ScaleFIS400--160MultidimensionalFatigue Severity ScaleFSS91--7MultidimensionalGlobal RSGRS10--10Single itemImmune thrombocytopenic Purpura --Patient Assessment Questionnaire, Fatigue subscaleITP-PAC10--100Single itemMultidimensional Assessment of FatigueMAF161--50MultidimensionalMultidimensional Fatigue InventoryMFI2020--100 (4--20 in each subscale)Multidimensional, (5 subscales: GF, PF, RA, RM and MF)^b^Neurological Fatigue Index for multiple sclerosisNFI-MS120--30 in SS, 0--24 in PS and 0--12 in CSMultidimensional,\\\n(3 scales: SS, PS and CS)^c^Perform QuestionnairePQ1212--60Multidimensional, Negative score^a^Profile of Mood States-FatiguePOMS-F70--28Unidimensional subscalePROMIS Fatigue-17Fatigue-171717--85UnidimensionalPROMIS Fatigue-7Fatigue-777--35UnidimensionalQuality of Life Inventory in Epilepsy, Energy/Fatigue subscaleQOLIE-3140--100Unidimensional subscale\\\nNegative score^a^Schwartz Cancer Fatigue ScaleSCFS66--30MultidimensionalSF-36 Vitality scaleSF-VT40--100Unidimensional subscaleSleep Impact Scale, Energy/Fatigue (E/F), Mental Fatigue (MF) subscalesSIS5 (E/F)\\\n3 (MF)0--100Unidimensional subscales\\\nNegative scores^a^Trial Outcome Index-FatigueTOI-F270--108Multidimensional, Negative score^a^Unidimensional Fatigue Impact ScaleU-FIS220--66UnidimensionalVisual Analogue Scale 0--10VAS 0--1010--10Single itemVisual Analogue Scale 0--100VAS 0--10010--100Single item^a^Negative score\u2009=\u2009high values indicate low level of fatigue^b^ *GF* General fatigue, *PF* Physical fatigue, *RA* Reduced activity, *RM* Reduced motivation. *MF* Mental fatigue^c^ *SS* Summary scale, *PS* Physical scale and *CS* Cognitive scaleTable 3Studies/substudies (*n\u2009=*\u200960) with reported MIDs for fatigue scales, subscales or single fatigue items. Study design and population are shown along with estimation methodInstrument/scale (score range)Citation/PopulationDesignAnchor-based MIDDistribution-based MIDTriangulation MIDAnchorGlobal change (% of scale range)ImprovedWorsenedMultidimensional scalesMFI\\\n(20--100)Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons11.5 (14\u00a0%)9.612.8Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons13.3 (17\u00a0%)6.89.5FSS\\\n(1--7)Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons0.6 (10\u00a0%)0.081.2Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons1.2 (20\u00a0%)0.41.0Robinson et al. 2009 MS (*n\u2009=*\u2009249)Cross-sectionalDisease duration, Expanded Disability Status Scale, Patient Assessment of MS Impact, MS Functional Composite0.5--1.1 (8--18\u00a0%)ES 0.3--0.81MAF\\\n(1--50)Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons5.0 (10\u00a0%)1.48.9Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons9.2 (19\u00a0%)5.48.3CFS\\\n(0--33)Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons2.3 (7\u00a0%)0.73.2Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons3.3 (10\u00a0%)1.43.5FIS\\\n(0--160)Rendas-Baum et al. 2010 MS (*n\u2009=*\u2009184)Cross-sectionalExpanded Disability Status Scale (EDSS), SF-36, EQ-5D9--24\\\n15.5 (10\u00a0%)4.8 (1 SEM), 9.6 (2 SEM)\\\n11.6 (1/3 SD), 17.3 (\u00bd SD)10--20TOI-F\\\n(0--108)Cella et al. 2002 Cancer (*n\u2009=*\u20092,583)Cross-sectional\\\nLongitudinal\\\n3 studies, follow-up: 3 d - 12\u00a0mPerformance status, haemoglobin level, response to treatment4.8--26.6 (4--25\u00a0%)4.2 (1 SEM), 10.5 (\u00bdSD)5.0PQ\\\n(12--60)Bar\u00f3 et al. 2011 Cancer (*n\u2009=*\u2009437)Longitudinal Follow-up: 3\u00a0mHaemoglobin level3.73.5SCFS\\\n(6--30)Schwartz et al. 2001 Cancer (*n\u2009=*\u2009103)Longitudinal Follow-up: 2 dGRS (7-step)5.0 (21\u00a0%)2.15.7FAsD\\\n(1--5)Matza et al. 2013 Depression (*n\u2009=*\u200996)Longitudinal Follow-up: 6 wBFI, ESS, CGI-S, Patient's perception of change0.3--0.60.2--0.3NFI-MS\\\n(0--30, 0--24 and 0--12 resp.)Mills et al. 2012 MS (*n\u2009=*\u2009208)Longitudinal Follow-up: 6--8 wGlobal perceived change item (5-step)2.5 (SS) (8\u00a0%)\\\n2.4 (PS) (10\u00a0%)\\\n0.8 (CS) (7\u00a0%)Unidimensional scales or subscalesMFI sbscales: GF, PF, RA, RM and MF\\\n(4--20)Purcell et al. 2010 Cancer (*n\u2009=*\u2009210)Longitudinal Follow-up: 6 w post treatmentScore change pre- and post-radiotherapyGF: 2.1 (13\u00a0%)\\\nPF: 2.0 (13\u00a0%)\\\nRA: 2.4 (15\u00a0%)\\\nRM: 1.6 (10\u00a0%)\\\nMF: 1.4 (9\u00a0%)2 for each subscaleU-FIS\\\n(0--66)Twiss et al. 2010 MS (*n\u2009=*\u2009911)Longitudinal Follow-up: 12\u00a0mEQ-5D6.54.74.2--7.0 (ES 0.3--0.5)\\\n2.4 (1 SEM)FAS\\\n(10--50)de Kleijn et al. 2011 Sarcoidosis (*n\u2009=*\u2009321)Longitudinal Follow-up: 12\u00a0mWHOQOL-BREF/Physical health domain, ROC3.5 (9\u00a0%)3.03.84.2 (ES 0.5)\\\n3.6 (1 SEM)4SF-36 VT\\\n(0--100)Bjorner et al. 2007 Several disease conditions (*n\u2009=*\u20093,445)Cross-sectionalRegression analyses using age, gender, race, disease condition and functional outcomes5/group level 10/individual levelGoligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons10.7 (11\u00a0%)7.318.3Kosinski et al. 2000 RA (*n\u2009=*\u2009693)Longitudinal Follow-up: 6 wPatient global assessment, Physician global assessment, pain, swelling, tenderness4.9--11.1 (5--11\u00a0%)Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons14.8 (15\u00a0%)11.311.9Spiegel et al. 2005 Hepatitis C virusSystematic review Delphi methodES data from included studiesES of 0.2\\\nRange 0.15--0.254.2\\\nRange 3--5Ward et al. 2015 RA (*n\u2009=*\u2009249)Longitudinal Follow-up: 1--4\u00a0mHAQ, CES-D, Health transition item of SF36, Global transition item11.0--20.0 (11--20\u00a0%)Wyrwich et al. 2003 Chronic obstructive pulmonary diseaseDelphi methodPatient change scenarios and\\\nSF36 data12.5Wyrwich et al. 2004 Coronary artery disease/congestive heart failureDelphi methodPatient change scenarios and\\\nSF36 data18.8\\\nRange 18.8--25FACIT-Fatigue\\\n(0--52)Cella et al. 2002 Cancer (*n\u2009=*\u20092,583)Cross-sectional Longitudinal 3 d - 12\u00a0mPerformance status, haemoglobin level, response to treatment3 (6\u00a0%)3Cella et al. 2005 RA (*n\u2009=*\u2009271)Longitudinal Follow-up: 24 wVT/SF-36, MAF3--4 (6--8\u00a0%)4.10 (1 SEM)\\\n2.2--5.5 (ES 0.2--0.5)3--4Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons5.9 (11\u00a0%)2.89.1Lai et al. 2011 SLE (*n\u2009=*\u2009254)Longitudinal Follow-up: 12--52 wPhysician-reported anchors (Physician GA)3--7 (6--13\u00a0%)2.7 (1 SEM)\\\n4.6 (1/3 SD), 6.8 (\u00bdSD)3--6Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons8.3 (16\u00a0%)6.85.2Reddy et al. 2007 Cancer (*n\u2009=*\u2009194)Longitudinal Follow-up: 8 dGlobal Benefit Score (7 step)10FACT-An Fatigue\\\n(0--80)Patrick et al. 2003 Cancer (*n\u2009=*\u2009375)Longitudinal, Follow-up: pre and post chemotherapyHaemoglobin level\\\nRegression analysis4.2POMS-F\\\n(0--28)Schwartz et al. 2002 Cancer (*n\u2009=*\u2009103)Longitudinal Follow-up: 2 dGRS (7-step)5.6 (20\u00a0%)2.15.7EORTC QLQ-C30 Fatigue Subscale\\\n(0--100)Bedard et al. 2013b Cancer (*n\u2009=*\u2009276)Longitudinal Follow-up: 1\u00a0mOverall QoL24.519.7 (1 SEM)\\\n6--15 (0.2--0.5 SD)Bedard et al. 2014 Cancer (*n\u2009=*\u2009369)Longitudinal Follow-up: 1\u00a0mOverall QoL 1--7\\\nOverall health anchor13.6--17.31.8 (1 SEM)\\\n6.7--16.8 (0.2--0.5 SD)Kvam et al. 2010 Multiple myeloma (*n\u2009=*\u2009239)Longitudinal Follow-up: 3\u00a0mGlobal rating of change (7-step but categorized into 3)13.58.6Maringwa et al. 2011a Brain cancer (*n\u2009=*\u2009941)Cross-sectional and longitudinalWHO Performance Status and MMSE12.48.910.0 (1 SEM)Maringwa et al. 2011b Lung cancer (*n\u2009=*\u2009812)Cross-sectional and longitudinalPhysician-rated WHO\\\nPS and weight change14.15.711 (1 SEM)Zeng et al. 2012 Cancer (*n\u2009=*\u200993)Longitudinal Follow-up: 1\u00a0mKPS clinical marker11.47.83.0--3.1 (1 SEM)\\\n5.8--14--7 (0.2--0.5 SD)SIS\\\n(0--100)Lasch et al. 2009 MDD (*n\u2009=*\u2009379)Longitudinal Follow-up: 8 wClinician rated tool (7-step) on severity and improvementE/F: 11.9 (12\u00a0%)\\\nMF: 13.3 (13\u00a0%)8.7 (\u00bdSD)\\\n10.6 (\u00bdSD)CRQ\\\n(1--7)Sch\u00fcnemann et al. 2005 Chronic obstructive pulmonary diseaseSystematic reviewCRQ data from5 studies, patient global ratings anchors and distributions based MIDs0.5--0.6 (8--10\u00a0%)0.47--0.54 (1 SEM)0.5Wyrwich et al. 2003 Chronic obstructive pulmonary diseaseDelphi methodPatient change scenarios and CHQ data2CHQ\\\n(4--28)Wyrwich et al. 2004 Coronary artery disease/congestive heart failureDelphi methodPatient change scenarios and\\\nCHQ data3\\\nRange 3--4QOLIE-31 Energy/fatigue subscale\\\n(0--100)Borghs et al. 2012 Epilepsy (*n\u2009=*\u20091,035)Longitudinal study of 3 RCTs Follow-up: 12 wPatient global impression of change (PGIC)\\\nRegression analysis7.5 (8\u00a0%)5.8 (0.3 ES) 9.4 (1 SEM)Single item measureVAS Fatigue Single item\\\n(0--100)Colangelo et al. 2009 SLE (*n\u2009=*\u2009202)Longitudinal Follow-up: 7.5\u00a0mSelf-rated health (5 step)13.99.1George & Pope 2011 Sj\u00f6gren's syndrome (*n\u2009=*\u200940)Longitudinal Follow-up: \u2264 16\u00a0mSelf-rated health (5 step)6.215.2Kwok & Pope 2010 PsA (*n\u2009=*\u2009200)Longitudinal Follow-up: \u226412\u00a0mSelf-rated health (5 step)8.23.6Sekhon et al. 2010 Systematic sclerosis (*n\u2009=*\u2009109)Longitudinal Follow-up: 7.5\u00a0mSelf-rated health (5 step)10.03.8Wells et al. 2007 RA (*n\u2009=*\u20091,043)Longitudinal Follow-up: 6--12\u00a0m Delphi methodHAQ, Patient Global assessment of disease and pain6.7--17 (7--17\u00a0%)10Wheaton & Pope 2010 SpA (*n\u2009=*\u2009140)Longitudinal Follow-up: 5\u00a0mSelf-rated health (5 step)1.414.4VAS Fatigue\\\nSingle item\\\n(0--10)Khanna et al. 2008 RA (*n\u2009=*\u2009307)Longitudinal Follow-up: 5.9\u00a0mRetrospective anchor (5-step)0.8--1.11.1--1.3Improved: ES\u2009=\u20090.39\\\nWorsened: ES\u2009=\u20090.44GRS Single item (0--10)Goligher et al. 2008 SLE (*n\u2009=*\u200980)Cross-sectionalGRS (7-step)\\\nPaired comparisons1.3 (13\u00a0%)0.31.5Pouchot et al. 2008 RA (*n\u2009=*\u200961)Cross-sectionalGRS (7-step)\\\nPaired comparisons2.0 (20\u00a0%)0.91.5Schwartz et al. 2001 Cancer (*n\u2009=*\u2009103)Longitudinal Follow-up: 2 dGRS (7-step)1.1 (11\u00a0%)ESAS fatigue Single item (0--10)Bedard et al. 2013a Cancer (*n\u2009=*\u2009421)Longitudinal Follow-up: 4--12 wWell-being0.1--1.31.0--1.80.1 (1 SEM)\\\n0.5--1.4 (0.2--0.5 SD)Reddy et al. 2007 Cancer (*n\u2009=*\u2009194)Longitudinal Follow-up: 8 dGlobal Benefit Score (7 step)4ITP-PAC\\\n(0--100)Mathias et al. 2009 ITP (*n\u2009=*\u2009125)Longitudinal Follow-up: 4 wGlobal assessment of change items (15-step)15.0 (15\u00a0%)ES\u2009=\u20090.57PROMIS fatigue item bank scalesPROMIS Fatigue-17 (17--85)Yost et al. 2011 Cancer (*n\u2009=*\u2009101)Cross-sectional and longitudinal Follow-up: 6--12 w23 anchor measuresT-score MID: 2.5--4.5\\\nRaw-score MID: 4.0--8.0 (6--12\u00a0%)Cross-sectional: ES 0.34--0.79\\\nLongitudinal: ES 0.27--0.52PROMIS Fatigue-7 (7--35)Yost et al. 2011 Cancer (*n\u2009=*\u2009101)Cross-sectional and longitudinal Follow-up: 6--12 w23 anchor measuresT-score MID: 3.0--5.0\\\nRaw-score MID: 2.0--3.0 (7--11\u00a0%)Cross-sectional: ES 0.24--0.76 Longitudinal: ES 0.24--0.51GRS (7-step)\u2009=\u2009Global rating scale with 7 response categories: Much more fatigue, Somewhat more fatigue, A little bit more fatigue, About the same fatigue, A little bit less fatigue, Somewhat less fatigue, and Much less fatiguePaired comparisons\u2009=\u2009Participants rated their fatigue in relation to another participant*PsA* Psoriatic arthritis, *SpA* Spondyloarthropathy\n\nMultidimensional scales {#Sec6}\n-----------------------\n\n### Multidimensional fatigue inventory (MFI), score 20--100 {#Sec7}\n\nTwo cross-sectional studies \\[[@CR33], [@CR46]\\] derived MIDs for systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) populations for the MFI total scale, using a patient global rating scale and interviews as anchors. MIDs ranged from 11.5 to 13.3 for global change and 6.8 to 9.6 for improvement and 9.5 and 12.8 for deterioration.\n\n### Fatigue severity scale (FSS), score 1--7 {#Sec8}\n\nThree cross-sectional studies reporting MIDs for the FSS were identified \\[[@CR33], [@CR46], [@CR50]\\]. Diagnostic groups included SLE, RA and multiple sclerosis (MS). Anchor-based approaches were applied in all the three studies and a distribution-based approach (viz. effect size, ES, of at least 0.25) was also applied in one \\[[@CR50]\\]. Two used a patient global rating scale as an anchor \\[[@CR33], [@CR46]\\] whereas the third used clinical anchors and baseline data from a clinical trial to establish MIDs \\[[@CR50]\\] MIDs ranged from 0.5 to 1.2 for global change, 0.08 to 0.4 for improvement and 1.0 to 1.2 for deterioration.\n\n### Multidimensional assessment of fatigue (MAF), score 1--50 {#Sec9}\n\nMID-estimates for the MAF in two cross-sectional studies with SLE and RA patients \\[[@CR33], [@CR46]\\] were estimated to 5.0 and 9.2 for global change, 1.4 to 5.4 for improvement and 8.3 to 8.9 for worsening, using a patient global rating scale.\n\n### Chalder fatigue scale (CFS), score 0--33 {#Sec10}\n\nThe same two cross-sectional studies \\[[@CR33], [@CR46]\\] reported MIDs for the CFS where MIDs for global change were 2.3--3.3; for improvement 0.7--1.4; and for deterioration 3.2--3.5.\n\n### Fatigue impact scale (FIS), score 0--160 {#Sec11}\n\nOne cross-sectional study with MS patients \\[[@CR49]\\] reported MIDs for the FIS ranging from 9--24 points for the different patient and clinician rating anchors, with a mean of 15.5 and SD 4.9. Distribution-based methods yielded MIDs ranging between 4.8--17.3 (1--2 SEM; \u00b1 1/3-1/2 SD). Triangulation of anchor and distribution-based methods gave a MID range of 10--20 points.\n\n### Trial outcome index-fatigue (TOI-F), score 0--108 {#Sec12}\n\nOne study \\[[@CR28]\\] reported TOI-F MIDs using data from three separate cancer trials. Triangulation was used to determine a MID, combining a patient-reported anchor, two physician-reported anchors (including response to treatment ratings), and one clinical anchor (haemoglobin level). MID estimates ranged from 4.8 to 26.6, and a single triangulated MID of 5.0 was recommended.\n\n### Perform questionnaire (PQ), score 12--60 {#Sec13}\n\nOne longitudinal study \\[[@CR22]\\] estimated the PQ MID in cancer patients to be 3.7 for improvement. Triangulation was used to estimate a recommended MID of 3.5.\n\n### Schwartz cancer fatigue scale (SCFS), score 3--30 {#Sec14}\n\nA longitudinal study of the SCFS using a patient-rated anchor \\[[@CR51]\\] reported MIDs for global change was 5.0; for improvement 2.1; and for deterioration 5.7 after a two days follow-up.\n\n### Fatigue associated with depression questionnaire (FAsD), score 1--5 {#Sec15}\n\nMIDs for the FAsD were estimated in one longitudinal study \\[[@CR43]\\] of patients with a clinical diagnosis of depression ranging from 0.3 to 0.6 for improvement and 0.2--0.3 for worsening after 6\u00a0weeks follow-up.\n\n### Neurological fatigue index for multiple sclerosis (NFI-MS), summary score 0--30 {#Sec16}\n\nOne longitudinal study \\[[@CR44]\\] using a patient global assessment of change reported MIDs for the NFI-MS; 2.5 for the ten-item Summary scale, 2.4 for the Physical scale (score range 0--24) and 0.8 for the Cognitive scale (score range 0--12).\n\nUnidimensional scales or subscales {#Sec17}\n----------------------------------\n\n### Multidimensional fatigue inventory (MFI) subscales, score 4--20 {#Sec18}\n\nA longitudinal study \\[[@CR47]\\] derived MIDs in a cancer population (pre and post radiotherapy) for the MFI five subscales. MIDs ranged between 1.4 to 2.4 depending on subscale. A general MID for all MFI subscales was recommended corresponding to 2 points.\n\n### Unidimensional fatigue impact scale (U-FIS), score 0--66 {#Sec19}\n\nOne longitudinal study using EQ5D as an anchor \\[[@CR55]\\] derived MIDs in an MS sample. U-FIS MIDs corresponded to 6.5 for improvement and 4.7 for deterioration, and distribution-based MIDs between 2.4 and 7.0.\n\n### Fatigue assessment scale (FAS), score 10--50 {#Sec20}\n\nMIDs for the FAS were reported in one longitudinal study of sarcoidosis patients using WHOQOL-BREF/Physical health domain and a ROC-curve as anchors as well as distribution based methods \\[[@CR31]\\]. MID ranged between 3.0 and 4.2 and a triangulated MID-value of 4 was suggested.\n\n### Vitality scale (VT) of the medical outcome study SF-36 health survey (SF-36), score 0--100 {#Sec21}\n\nEight studies \\[[@CR26], [@CR33], [@CR35], [@CR46], [@CR54], [@CR56], [@CR59], [@CR60]\\] determined MIDs for the VT scale of the SF-36 using different designs and diagnostic groups; longitudinal with patient- and/or clinician rated anchors, cross-sectional using patient-rated anchors and systematic reviews using combined study data and expert panels. The MIDs ranged from 7.3 to 11.3 for improvement, 11.9 to 18.3 for worsening and 3.5 to 20, for all those with a global change and 4.2 to 18.8 for a triangulated MID.\n\n### FACIT fatigue scale (FACIT-Fatigue), score 0--52 {#Sec22}\n\nSix cross-sectional or longitudinal studies \\[[@CR28], [@CR29], [@CR33], [@CR38], [@CR46], [@CR48]\\] reported MID estimates derived in patients with cancer, SLE, or RA using patient or clinician-rated anchors. In these studies, MIDs varied from 3 to 8.3 irrespective of direction of change, 2.8 to 6.8 for improvement and 5.2 to 9.1 for deterioration. Two of the studies \\[[@CR29], [@CR38]\\] combined various distribution-based approaches (SEM, SD and ES), resulting in MIDs ranging between 2.2 and 6.8, and presented triangulated MIDs ranging between 3 and 6.\n\n### FACT-an fatigue subscale (FACT-An Fatigue), score 0--80 {#Sec23}\n\nOne longitudinal study \\[[@CR45]\\] estimated a MID for improvement of 4.2 in cancer patients using haemoglobin level as a clinical anchor and regression analysis to calculate MID.\n\n### Profile of mood states short form fatigue subscale (POMS-F), score 0--28 {#Sec24}\n\nOne longitudinal study reported MIDs for the POMS-F using a sample of cancer patients undergoing chemotherapy \\[[@CR51]\\]. A global MID of 5.6 points was determined as well as separate MIDs for improvement (2.1 points) and deterioration (5.7 points).\n\n### European organization for research and treatment of cancer quality of life questionnaire core 30 (EORTC QLQ-30)---fatigue scale, score 0--100 {#Sec25}\n\nSix cross-sectional and longitudinal studies \\[[@CR24], [@CR25], [@CR36], [@CR40], [@CR41], [@CR62]\\] reported MIDs derived in a variety of cancer diagnoses. MIDs were reported as 11.4 to 17.3 points for improvement and 5.7--24.5 points for deterioration. Distribution-based MIDs ranged from 3.0 to 19.7.\n\n### Sleep impact scale (SIS), energy/fatigue and mental fatigue subscales, score 0--100 {#Sec26}\n\nOne longitudinal study \\[[@CR39]\\] using a clinician-rated anchor and a distribution-based method to assess change at 8-week follow-up, reported MIDs derived in patients with major depressive disorder (MDD). The anchor-based approach yielded a MID of 11.9 for the Energy/Fatigue subscale, whereas the distribution-based MID was 8.7. The corresponding MIDs for the Mental Fatigue subscales were 13.3 and 10.6, respectively.\n\n### Chronic respiratory questionnaire (CRQ), score 1--7 {#Sec27}\n\nTwo systematic reviews \\[[@CR52], [@CR59]\\] used CRQ data from earlier studies to determine MIDs for the CRQ/Fatigue subscale and triangulated MIDs of 0.5 and 2 were proposed. One of the reviews estimated MIDs between 0.5--0.6 for global change and distribution-based MIDs of 0.47--0.54 \\[[@CR52]\\].\n\n### Chronic heart failure questionnaire (CHQ), score 4--28 {#Sec28}\n\nOne systematic review using CHQ data and an expert panel proposed a MID for the CHQ/Fatigue subscale of 3--4 irrespective of direction and a triangulated MID of 3 \\[[@CR60]\\].\n\n### Quality of life inventory in Epilepsy (QOLIE-31), energy/fatigue subscale, score 0--100 {#Sec29}\n\nOne longitudinal study used 3 randomised controlled trials to examine MID for the QOLIE-31/Energy/fatigue subscale \\[[@CR27]\\]. A MID of 7.5 was defined using a patient rating of change and regression analysis. Distribution-based MIDs ranged between 5.4 and 9.4.\n\n### Visual analogue scale (VAS), score 0--100 or 0--10 {#Sec30}\n\nSix longitudinal studies \\[[@CR30], [@CR32], [@CR37], [@CR53], [@CR57], [@CR58]\\] derived MIDs for the VAS 0--100 and one \\[[@CR34]\\] for the VAS 0--10 in a variety of diagnostic groups. MIDs for the VAS-100 ranged from 1.4 to 13.9 for improvement and 3.6 to 15.2 for deterioration, while the global change varied between 6.7 and 17. One study \\[[@CR57]\\] determined a triangulated MID of 10 using the Delphi method. MIDs for the VAS-10 ranged between 0.8 to 1.1 for improvement and 1.1 to 1.3 for worsening, and were derived from three different anchors and at different follow-up times in three different diagnostic groups (RA, SLE and cancer) \\[[@CR34]\\].\n\n### Global rating scale (GRS), score 0--10 {#Sec31}\n\nMIDs for the single item GRS scale were determined in SLE, RA and cancer patients in two cross-sectional studies \\[[@CR33], [@CR46]\\] and one longitudinal study \\[[@CR51]\\], all using a patient global rating scale as an anchor. Global MIDs ranged from 1.1 to 2.0, while MIDs for improvement were 0.3 to 0.9 and for deterioration 1.5.\n\n### Edmonton symptom assessment system (ESAS) fatigue item, score 0--10 {#Sec32}\n\nTwo longitudinal cancer studies \\[[@CR23], [@CR48]\\] identified MIDs for the fatigue item in the ESAS scale. MIDs for improvement ranged from 0.1 to 4 and between 1.0 and 1.8 for worsening of fatigue. Distribution-based MIDs ranged from 0.1 to 1.4.\n\n### Immune thrombocytopenic Purpura---Patient assessment questionnaire, (ITP-PAC) fatigue subscale, score 0--100 {#Sec33}\n\nOne longitudinal study \\[[@CR42]\\] assessed MIDs using patient impression of change for the ITP-PAC/Fatigue subscale. Global change was defined as 15.0 or as an effect size of 0.57.\n\nPROMIS fatigue item bank scales {#Sec34}\n-------------------------------\n\n### 17-item PROMIS fatigue (fatigue-17) and 7-item PROMIS Fatigue (Fatigue-7), score 17--85 and 7--35 {#Sec35}\n\nOne study \\[[@CR61]\\] derived MIDs for both the PROMIS Fatigue-17 and Fatigue-7 in patients with cancer. The study used both cross-sectional and longitudinal data as well as anchor-based and distribution-based methods. Distribution-based MIDs were reported as effect sizes. For the Fatigue-17, the ES ranged from 0.34--0.79 and 0.27--0.52 for cross-sectional and longitudinal designs, respectively. Corresponding effect sizes for the *Fatigue-7 were* 0.24--0.76 and 0.24--0.51. Triangulated raw score MIDs ranged from 4.0 to 8.0 for the Fatigue-17 and 2.0 to 3.0 for the Fatigue-7 while t-score MIDs varied between 2.5 to 4.5 for the Fatigue-17 and 3.0 to 5.0 for the Fatigue-7.\n\nDiscussion {#Sec36}\n==========\n\nThis systematic review identified 41 studies reporting MIDs for 28 fatigue PROMs or subscales measuring fatigue, yielding a total of 60 studies or substudies estimating MID. It is important to note that there are many more fatigue PROMs available today than the 28 reported here. For example, a critical review of fatigue PROMs from 2009 \\[[@CR8]\\] identified 39 such PROMs; however, only 11 of these overlapped with PROMs in our review. This suggests that there are roughly 56 or more fatigue PROMs currently represented in the literature. Considering the importance attributed to MIDs for interpreting the meaningfulness of change in PROM scores \\[[@CR9], [@CR63]\\], it is somewhat surprising that MIDs are available for only about half of all published fatigue PROMs. Moreover, few PROMs had MIDs that were determined in more than two studies and diagnostic groups, and more than half of the PROMs had MIDs that were derived in only one diagnostic group. Important exceptions were the SF-36 Vitality scale (\\>8 diagnostic groups/8 studies; the FACIT-Fatigue scale (4 diagnoses/6 studies); the EORTC QLQ-C30 Fatigue subscale (6 cancer diagnoses/6 studies); and the VAS-100 Fatigue (6 diagnoses/6 studies). Considering that these scales are some of the most widely used and oldest PROMs in use today it is unsurprising that greater research attention has focused on determining MIDs for these scales; however, it is noteworthy that so few separate studies reported MIDs for commonly used generic fatigue PROMs, such as the MFI, FSS, FIS and FAS.\n\nPrevious research has highlighted considerable variability in MID values as a function of estimation method, population and context \\[[@CR14], [@CR18], [@CR19]\\], suggesting the importance of considering such factors when appraising the appropriateness of published MIDs for use in clinical research and practice. In line with this, substantial variation was observed in MID values for individual fatigue PROMs in this review. For example, MIDs for the SF-36 Fatigue scale ranged from as low as 4.2 to as high as 20.0 points (0--100 point scale) in studies varying in methodologies, anchors, diagnostic groups and direction of change assessed. Similarly, MIDs for the VAS-100 Fatigue scale ranged from 1.4 to 17. MIDs for the cancer-specific EORTC QLQ-C30 fatigue scale also varied between 1.8 and 24.5 points (0--100 scale) and those for the FACIT-Fatigue scale ranged between 6 and 16 (converted to percent), see Table\u00a0[3](#Tab3){ref-type=\"table\"}. This wide variation in MIDs for individual fatigue scales suggests the importance of understanding how any particular MID was derived and of applying this knowledge when appraising its appropriateness for interpreting changes in fatigue scores.\n\nMID estimation methods varied considerably in the identified studies and substudies. However, in accordance with recent recommendations regarding methods for MID estimation \\[[@CR14]\\], nearly all studies applied an anchor-based approach, where at least one anchor was used. Patient global change ratings were by far the most common anchor, but even clinician-reported and clinical anchors were implemented. Where more than one anchor was applied either a range of values was generally reported or, as recommended \\[[@CR14], [@CR63], [@CR64]\\], values were often triangulated to a single or smaller range of MIDs. Distribution-based methods were used in about a third of the studies and only in conjunction with anchor-based approaches. A few studies used a Delphi method (Table\u00a0[3](#Tab3){ref-type=\"table\"}).\n\nIn the studies using several anchors to determine MID values, global MID ranges varied within single studies from as little as two points (percent scores), in relation to the FACIT-Fatigue scale using patient-based anchors \\[[@CR29]\\], to about 20 points for the TOI-F \\[[@CR28]\\] using patient, clinician and clinical anchors. Interestingly, two studies reporting MIDs for the SF-36 Vitality scale, using the same diagnostic group (RA) but different anchors, yielded two distinct ranges of MIDs. In the study by Kosinski et al. \\[[@CR35]\\], using patient and physician global assessments as anchors, MIDs ranged from 4.9--11.1, whereas a range of 11.0--20.0 was reported by Ward et al. \\[[@CR56]\\] using the HAQ, CES-D and the SF-36 health transition item. Neither of these studies triangulated the range of values to a single MID or smaller range of values and hence these wide ranges of MIDs are arguably of questionable practical value for interpreting change in fatigue in RA patients as measured with the SF-36 Vitality.\n\nTriangulation was used in 17 substudies, of which 10 used more than two anchors. This method has been recommended for consolidating MIDs derived from different methods to a single or small range of MID values \\[[@CR14]\\]. However, it has been criticized \\[[@CR19]\\] since it may in practice involve the need to converge widely disparate MIDs derived using different estimation methods and diverse anchors, which often represent very different stakeholder perspectives. An example of a MID triangulated from a wide range of MIDs is the TOI-F \\[[@CR28]\\] where a MID range of 4.4--24.6 (percent scores) was triangulated to 5.0. Where MID ranges are smaller, the value and applicability of the triangulated MID may be more immediately apparent. For example, Sch\u00fcnemann et al. \\[[@CR52]\\] reported a MID range for the CRQ of 6.7--8.5 (percent scores), derived from patient anchors, a systematic review and distribution-based methods, which was triangulated to a MID of 6.7.\n\nA second factor known to influence variation in MID values is the patient population in which the MID is determined. Variation by diagnostic group is exemplified by comparing MIDS from two studies, each using the same estimation method (7-step global rating scale) and study design (cross-sectional) but different diagnostic groups \\[[@CR33], [@CR46]\\]. One of the studies \\[[@CR33]\\] determined MIDs for seven different fatigue PROMs in patients with SLE and the other \\[[@CR46]\\] did the same in patients with RA. Comparison of the global MIDs for the SLE and RA patients, shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}, shows consistently smaller MIDs for SLE versus RA across all seven PROMs. It is noteworthy that most PROMs had MIDs that were determined in only one patient population and the relevance of these MIDs for use in other patient groups thus remains unclear.\n\nA third factor influencing variation in MID values is the context within which the MID is determined. Context issues concern, for example, characteristics of the patient population, e.g., such as baseline state \\[[@CR65]\\], disease severity \\[[@CR66]\\], and direction of change \\[[@CR13], [@CR20]\\], as well as study design and intervention. For example, patients with baseline scores indicating more severe fatigue may value magnitudes of change in fatigue differently than those with less severe fatigue. Corroborating previous research finding, MIDs for improvement differed from those for deterioration in all identified studies. MIDs tended to be larger for deterioration than improvement, except in the EORTC QLQ-30 and VAS Fatigue item. MIDs for improvement were consistently smaller than global MIDs.\n\nA strength of this study is that reported MIDs for fatigue scales or subscales were systematically compiled and described. Assessment for inclusion or exclusion and data extraction from included studies was done independently by two authors (\u00c5N and AD). A limitation is that the search period was restricted to studies from 2000 onwards and search strings for the many variations on MID was also limited and therefore some studies reporting MIDs for fatigue scales may not have been captured in the literature searches. Another limitation is that the description of the study designs and results had to be summarized and simplified in tables and information could be lost. Therefore, when evaluating MIDs the original study/studies should be consulted.\n\nConclusions {#Sec37}\n===========\n\nMIDs vary substantially by estimation method, patient population and context both across and within fatigue PROMs. In light of this variation, published MIDs should be applied judiciously, after carefully considering their applicability to characteristics of the study in question. The information provided in this paper may serve to aid researchers and clinicians in making informed decisions regarding the appropriateness of published MIDs for their particular study and patients.\n\nAbbreviations {#Sec38}\n=============\n\nES, effect size; GRS, global rating scale; ITP, immune thrombocytopenic purpura; MCID, minimal clinical important difference; MCII, minimal clinically important improvement; MDD, major depressive disorder; MID, minimal important difference; MS, multiple sclerosis; PRO, patient reported outcome; PROM, patient reported outcome measure; PROMIS, patient-reported outcomes measurement information system; RA, rheumatoid arthritis; SD, standard deviation; SEM, standard error of measurement; SLE, systemic lupus erythematosus; QoL, quality of life.\n\nFunding {#FPar1}\n=======\n\nThis study was funded by the University of Gothenburg Centre for Person-Centred Care (GPCC).\n\nAvailability of data and materials {#FPar2}\n==================================\n\nNot applicable.\n\nAuthors' contribution {#FPar3}\n=====================\n\n\u00c5N, CT and \u00c5LN planned the study. \u00c5N and AD performed the searches, screened all research hits and extracted the data. All authors contributed to the manuscript writing. AD made the tables. All authors agreed on the final version of the manuscript.\n\nCompeting interests {#FPar4}\n===================\n\nThe authors declare that they have no competing interests.\n\nConsent for publication {#FPar5}\n=======================\n\nNot applicable.\n\nEthics approval {#FPar6}\n===============\n\nNot applicable.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nThe scientific basis for an alternate way of processing the raw data constituting the output of Fourier transform spectrometers is expected to improve substantially the precision of both concentration and energy measurements. Good frequency resolution is attained in a standard traditional dispersive spectrometer at the cost of blocking at the slit of the monochromator most of the photons that would otherwise reach the detector. This reduces considerably the signal to noise ratio. A Fourier transform spectrometer replaces the monochromator by a Michelson interferometer, which yields the cosine Fourier transform of the spectrum as the output. The procedure collects information at all frequencies simultaneously, improving dramatically both speed and the signal to noise ratio because the detector captures the full intensity of the light coming from the sample. The output is an interferogram consisting of a graph giving the radiation intensity as a function of the difference in the optical path length of the two arms of the interferometer, which is accurately measured from the interference fringes of a reference laser. This is referred to as raw data and usually exhibits a complex oscillating structure that must be Fourier transformed to bring out the frequency spectrum in a conventional way. The spectrometer then has to be associated with a numerical processor to display the spectrum. Infrared (IR) absorption spectroscopy went through a major advance when Fourier transform spectroscopy came to the fore, and practically no dispersive IR equipment is in the market today. However, the procedure is also practical in optical spectroscopy, nuclear magnetic resonance spectroscopy, and magnetic resonance spectroscopic imaging.^[@ref1]\u2212[@ref3]^ The new technique came on the scene associated with the development of the fast Fourier transform (FFT) algorithm, a mathematical method for the fast calculation of Fourier transforms of functions, fast enough to process the highly structured interferograms with acceptable precision. The discovery of the FFT algorithm by Cooley and Tukey initiated a new generation of IR instruments and techniques.^[@ref4]^\n\nHence, the mathematical processing of the raw data for displaying them as a standard spectrum in the frequency domain is adequately solved by the FFT algorithm, together with the apodization techniques that control spurious spectral features created by the truncation of the interferogram constituting the raw data. Experimental scans are necessarily finite, and sudden cutoffs at the boundaries have broad unphysical Fourier representations that are discarded by apodization. However, precise quantitative chemical analysis and the accurate determination of the excitation energies of molecular bonds demand additional mathematical processing of the spectra, further to the Fourier transformation of the raw data. The concentration of a chemical species is determined by the area under the peak identifying a characteristic bond of it, whose evaluation demands curve fitting of the data, particularly in the presence of heavy overlapping or structured background. Standard Gauss, Lorentz, and Voigt distributions in practice are the analytical expressions used for fitting the shape of the spectral peaks and determining their areas by integration. However, these distributions are merely interpolation functions because, in rigor, they do not follow from solving a real physical model for the processes causing the peak broadenings.^[@ref5]^ Gauss and Lorentz curves are analytically very different, especially concerning the contribution of the tails to the peak area, which is much more significant for the latter. Tails immerse in the noisy background and their effect owns to the domain of experimental uncertainty when using a tentative model for the lineshape. It has been proven that the assigned peak intensities may show substantial variations with the choice of the lineshape model.^[@ref6]^ The symmetry of these standard distributions evidences their limited ability to describe spectral profiles. It has been demonstrated on a general basis that the lineshapes for photon absorption and emission by atomic or molecular species in a condensed environment are always asymmetric with respect to the net energy of the electronic transition.^[@ref7],[@ref8]^ Hence, a significant source of error is the adoption of a standard distribution not well grounded on the physics of the target to describe the spectral features. The fit of the experimental data by the mathematical curves given by a realistic model for the profiles of the spectral features greatly improves precision. If the theoretical curve reproduces well the physics of the radiation field interacting with the target, then the error is given by the statistical dispersion of the experimental points along the curve that bests the fit, instead of its whole breadth.^[@ref7]\u2212[@ref11]^\n\n2. Theory {#sec2}\n=========\n\n2.1. Lineshape Functions of Electromagnetic Spectra {#sec2.1}\n---------------------------------------------------\n\nLine broadenings and energy shifts are produced mainly by multiphonon processes involving the extended acoustic modes of vibration of the condensed medium embedding the photosensitive orbital. They are activated by the local distortion that follows the sudden excitation or de-excitation of the electronic bonding orbitals and can be calculated analytically, yielding a closed-form mathematical expression for the lineshape function.^[@ref7],[@ref8]^ It is fortunate that this theoretical expression for the spectral distribution in the standard frequency domain has the general form of the Fourier transform of a function in the time domain. This way, the theory gives directly what experimentalists call the raw data, and there is no need in principle of calculating the Fourier transforms of the theoretical and experimental results to compare them and make them fit. Then, the theoretical analysis of the spectra is made in the time domain, working directly with the raw data.\n\nThe lineshape function has been proven in the recent literature to be given by the integral expression^[@ref7]\u2212[@ref9],[@ref11]^where \u210f*ck* is the photon energy, *E* is the energy difference of the two electronic states involved in the transition, *a* is essentially the bond length, and *v*~s~ is the mean speed of sound of the acoustic modes of vibration of the medium. The adimensional constants \u03b1 and \u03b2, and the adimensional dummy time \u03c4 are given bywhere \u0394*F* is the bond mean force variation upon excitation, \u03c1 is the density, *k*~B~ is the Boltzmann constant, and *T* is the temperature. The auxiliary functions *J*(\u03c4; *T*) and *I*(\u03c4) are dependent on the symmetry of the surroundings of the orbital undergoing the transition. For the simplest case of octahedral coordination (OC) of the optically active orbital, they readwith *q*~D~ being the Debye wavevector of the acoustic waves, *aq*~D~ = (12\u03c0^2^)^1/3^, andThe second term on the right-hand side of [eq [1](#eq1){ref-type=\"disp-formula\"}](#eq1){ref-type=\"disp-formula\"} for *F*(\u210f*ck*; *T*), containing the delta function, is the zero-phonon line, and the first one is the phonon-broadened distribution. The lineshape function *F*(\u210f*ck*; *T*) is normalized as^[@ref7]\u2212[@ref9]^and hence the relative contribution of zero-phonon processes to the total is *I*~ZPL~ = exp\\[\u2212\u03b1*J*(\u221e; *T*)\\].\n\nOther symmetries may give more complicated functional forms for *J* and *I*. For example, for tetrahedral coordination (TC) of the optically active orbitals, one hasThe output *f*(*x*, *T*) of the Michelson interferometer, where *x* is the difference in the optical path lengths of the two arms of the interferometer, is the cosine Fourier transform of the spectrum given by the lineshape function [1](#eq1){ref-type=\"disp-formula\"}, i.e.where *k\u0305* = 1/\u03bb is the wave number and *h* is the Planck constant. Substituting [eq [1](#eq1){ref-type=\"disp-formula\"}](#eq1){ref-type=\"disp-formula\"} and solving the integral, this giveswhere *x*, in fact, plays the role of the virtual time *t* = *x*/*c*, conjugated to the angular frequency \u03c9 = *E*/\u210f. Function *f*(*x*; *T*) is actually the interferogram, which Fourier transform infrared (FTIR) spectroscopists call the raw data. Therefore, replacing in [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"}, the pairs of auxiliary functions [3](#eq3){ref-type=\"disp-formula\"} and [4](#eq4){ref-type=\"disp-formula\"} for octahedral symmetry of the optically active orbitals, or [7](#eq7){ref-type=\"disp-formula\"} and [8](#eq8){ref-type=\"disp-formula\"} for tetrahedral coordination of them, one obtains explicit closed-form mathematical expressions for the interferograms. Auxiliary functions for other symmetries can be derived from the general expressions for the electron--phonon coefficients, given in ref ([@ref8]).\n\n2.2. Proposed Method {#sec2.2}\n--------------------\n\nAs both the experimental technique and the general theory, which is well grounded on the physics of the energy transfers between the radiation field and the charges in a condensed system, arrive both to the interferogram expressed in [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"}, in principle, there is no need to perform any Fourier transform of the data to grasp the physical information from the experimental results. Analytical closed-form [expression [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"} depends on only a few parameters, \u03b1, \u03b2, and the net transition energy *E*, per spectral line. Hence, the most practical way to proceed is to find the constants \u03b1, \u03b2, and *E* by means of a best-fit analysis of [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"} to the experimental interferograms. Fourier analysis then becomes just an optional alternative for people who like to identify spectra in the conventional frequency domain.\n\nThe method seems to be highly convenient because it retains all the advantages of Fourier spectrometry avoiding the numerical errors associated with the Fourier integration of rapidly oscillating functions. An important example of this is given by the area under the spectral line, which in agreement with [eq [9](#eq9){ref-type=\"disp-formula\"}](#eq9){ref-type=\"disp-formula\"} is given by half the intensity *f*(0, *T*)/2 of the central maximum of the interferogram. This magnitude is unity under the hypothesis of a single emission or absorption center of our theory, but in empirical grounds is proportional to the number of optically active orbitals and gives the concentration of them. However, the implementation of practical procedures for interpreting the measured interferograms directly with [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"} may be not immediate, particularly when dealing with narrow spectral lines, precisely because their interferograms oscillate strongly with *x*. In particular, the central maximum may be very narrow and its intensity may be strongly affected by the experimental uncertainty of *x*.\n\n3. Results and Discussion {#sec3}\n=========================\n\nThe general procedure can be applied to both wide spectral features, like those displayed by fluorescent compounds, or narrow ones, as the sharp minima observed in the absorption spectra of infrared light passing through many materials. The physical process is essentially the same. [Figures [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}--[3](#fig3){ref-type=\"fig\"} show the mathematical steps of the calculation of the spectrum of YAG:Ce^3+^ (yttrium aluminum garnet, Y~3~Al~5~O~12~, doped with Ce^3+^), a phosphor having many technical applications. This system is particularly interesting because its emission spectrum has been measured at a temperature close to *T* = 0 with a resolution large enough to clearly observe at \u03bb = 489 nm (*k\u0305* = 20\u2009450 cm^--1^) the zero-phonon line corresponding to the main of the two emission bands.^[@ref12]^ The relative intensity of the zero-phonon line of YAG:Ce^3+^ at temperature *T* = 4 K is observed to be 0.27% of the total intensity of the main emission band. The emission has been attributed to competing transitions^[@ref11],[@ref12]^ of the AlO~6~ groups of quasioctahedral coordination inside the complex unit cell of YAG.^[@ref13]^\n\n![Auxiliary functions *J*(*x*/*c*; *T*) and *I*(*x*/*c*) defined in [eqs [3](#eq3){ref-type=\"disp-formula\"}](#eq3){ref-type=\"disp-formula\"} and [4](#eq4){ref-type=\"disp-formula\"} for octahedral symmetry with \u03c4 = *x*/*c*.](ao-2018-02914p_0001){#fig1}\n\n[Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"} shows the auxiliary functions *J*(*x*/*c*; *T*) and *I*(*x*/*c*) for octahedral symmetry, as given by [eqs [3](#eq3){ref-type=\"disp-formula\"}](#eq3){ref-type=\"disp-formula\"} and [4](#eq4){ref-type=\"disp-formula\"}. [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"} represents the theoretically predicted interferogramof two competing emissions with weights 2/3 and 1/3. The weighting factors follow from assuming that the quasioctahedrally coordinated emission center has degenerated *x* and *y* lobes and slightly different *z* lobes. Both *f*~1~ and *f*~2~ have the functional form [10](#eq10){ref-type=\"disp-formula\"} with constants \u03b1, \u03b2, and *E* chosen to fit the measured spectrum. The transition energies are *E*~1~ = 2.535 eV and *E*~2~ = 2.343 eV (corresponding to the wavenumbers *k\u0305*~1~ = *E*~1~/(*hc*) = 20\u2009450 cm^--1^ and *k\u0305*~2~ = *E*~2~/(*hc*) = 18\u2009900 cm^--1^). The other constants are \u03b1~1~ = 10.00, \u03b1~2~ = 12.25, and \u03b2 = \u221e because *T* \u2248 0. The value of *E*~1~ is given by the zero-phonon line, and hence is not an adjustable parameter; the other three constants were chosen to fit the experimental data.\n\n![Filled black circles represent the function *f*(*x*; *T*), given by [eq [11](#eq11){ref-type=\"disp-formula\"}](#eq11){ref-type=\"disp-formula\"}, where *f*~1~ and *f*~2~ have the general form [10](#eq10){ref-type=\"disp-formula\"} with slightly different parameters \u03b1 and *E*. It is expected that *f*(*x*; *T*) will reproduce the experimentally registered Michelson interferogram of two partially resolved spectral features emitted by centers that have concentrations in the ratio of 2:1. The broken line is only a guide to the eye.](ao-2018-02914p_0002){#fig2}\n\nNotice in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"} that the number of calculated points in the interval of *x*, where *f*(*x*/*c*; *T*) is appreciable, seems insufficient to represent properly the too-structured function. This is not really a problem to construct the spectrum because the FFT algorithm is an analytical procedure that finds out the function whose Fourier transform (or anti-transform) passes by the given points. The main aspect is the accuracy of these points, but care must be taken also to avoid aliasing of the frequency by poor sampling of the data.\n\n[Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"} shows the cosine Fourier transform of the function [11](#eq11){ref-type=\"disp-formula\"}, calculated by the FFT algorithm included as a standard tool in the Excel spreadsheet. The agreement of the theoretical curve with the experimental spectrum is quite impressive. As the Fourier transform is unique, this indicates that function [11](#eq11){ref-type=\"disp-formula\"} with the assumed values for the constants should represent with good accuracy the output of the Michelson interferometer. The FFT algorithm was run with 1024 points for *x*, which runs over an interval of optical path differences 0 \u2264 *x* \u2264 0.02 mm.\n\n![Open circles represent the fluorescence emission spectrum of YAG:Ce^3+^ at *T* = 4 K, as measured by Bachmann et al.^[@ref12]^ The solid line represents the Fourier transform of the function represented in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"} by the discrete set of filled black circles.](ao-2018-02914p_0003){#fig3}\n\n[Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"} shows a calculated interferogram of the general formwhere both *f*~1~ and *f*~2~ are given by [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"} with the auxiliary functions [7](#eq7){ref-type=\"disp-formula\"} and [8](#eq8){ref-type=\"disp-formula\"} for tetrahedral symmetry. The weighting factors 3/4 and 1/4 assume a single distorted orbital in the subgroup of coordination 4 inserted in the more complex unit cell. The constants \u03b1~1~ = 18 and \u03b1~2~ = 22, and the wavenumbers *k\u0305*~1~ = 20\u2009510 cm^--1^ and *k\u0305*~2~ = 19\u2009700 cm^--1^ (*E*~1~ = 2.543 eV and *E*~1~ = 2.442 eV) are chosen to give the fit of the experimental spectrum of rhodamine 6G shown in [Figure [5](#fig5){ref-type=\"fig\"}](#fig5){ref-type=\"fig\"}.\n\n![Black dots represent function *f*(*x*; *T*) given by [eq [12](#eq12){ref-type=\"disp-formula\"}](#eq12){ref-type=\"disp-formula\"}, where *f*~1~ and *f*~2~ have the general form [10](#eq10){ref-type=\"disp-formula\"} with parameters \u03b1 and *E* chosen to fit the Michelson interferogram of rhodamine 6G. The broken line is only a guide to the eye.](ao-2018-02914p_0004){#fig4}\n\n![Open circles represent the fluorescence emission spectrum of rhodamine 6G (Thermo Fisher Scientific, Fluorescence SpectraViewer), and the solid line corresponds to the theoretical lineshape function obtained from the Fourier transform of the function passing by the black dots in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}.](ao-2018-02914p_0005){#fig5}\n\n4. Conclusions {#sec4}\n==============\n\nIn essence, the method put forward in the previous sections replaces the Fourier analysis of the interferogram ensuing as the output of a Fourier transform spectroscope by best-fit analysis of [eq [10](#eq10){ref-type=\"disp-formula\"}](#eq10){ref-type=\"disp-formula\"} to it. Apodization becomes unnecessary, and all the advantages of traditional Fourier transform spectroscopy, Jacquinot, Fellgett, and Connes, are conserved. However, the main result is the addition of three new advantages:(i)The procedure gives directly the net energy released or absorbed by the electronic transition, free of Stokes shifts.(ii)The lineshapes are firmly grounded on the physics of the interaction between the electromagnetic radiation field and the target. Hence, the intensities are obtained with better precision because no standard interpolation functions are necessary to fit the data.(iii)The theory holds equally well for transitions between vibronic states, in the infrared region, and states of molecular orbitals, associated with the much wider peaks of luminescence spectra.\n\nThe situation described above should be quite exact in the presence of a single species. However, in the presence of multiple species, the task should be considerably more computationally intensive and error prone because of the overlap of multiple interferograms. This is not a very important drawback because the difficulty can be solved by combining the two methods. An initial Fourier transform can separate the signal into spectral regions corresponding to the various species present and provide numeric conditions for the subsequent fitting task. The Fourier transform may be useful also to identify the contribution of impurities and noise.\n\nThe authors declare no competing financial interest.\n"}
+{"text": "Soft tissue sarcomas are a group of rare malignant tumors originating from connective tissues. They can occur at any age and in any anatomic site.^[@bib1]^ These tumors are challenging to diagnose and treat because of the high degree of malignancy, multiple pathological subtypes, and heterogeneity in biological properties; the disease may also relapse easily or show distant metastasis.^[@bib1]^ Regardless of the pathological type and grade, surgery is still the most effective method for the treatment of soft tissue sarcomas (STSs) except gastrointestinal stromal tumors and desmoid.^[@bib2]^ The presence of osseous, joint, nerve, and vascular invasion are largely determinant of the feasibility of limb sparing surgery. When STSs, especially deep tumors, involve the major vessels, limb salvage may not be possible for main surgical solutions.^[@bib3]^ The choice of the treatment strategy for STSs is often difficult, but standardized surgical resection can increase the local control rate of the STS and improve the limb salvage rate. Therefore, accurate delineation of the local extent of the neoplasm in bones, muscles, and joints is essential, especially by determining the presence and level of major vessel involvement by imaging before surgical treatment, which can contribute to the choice of treatment strategy.\n\nComputed tomography (CT) plays the same critical role as the magnetic resonance imaging (MRI) in staging of STSs. There was no difference between CT and MRI in determining tumor involvement of muscle, bone, joints, neurovascular structures, or major vessel, and the combined interpretation of CT and MRIs did not improve accuracy for staging of STSs.^[@bib4],[@bib5]^ Contrast-enhanced axial images (CEAIs) in the arterial phase and traditional postprocessing techniques used with CEAI data, such as volume rendering (VR), can provide important information for preoperative planning of STSs. Verga et al^[@bib6]^ showed that contrast-enhanced CT and VR are effective in depicting adjacent major tumoral vascular involvement. Li et al^[@bib7]^ reported that postcontrast CT images had 100% sensitivity, 83.3% specificity, 87.5% positive predictive value (PPV), 100% negative predictive value (NPV), and 92.3% accuracy in the detection of vascular invasion, and VR had 71.4% sensitivity, 100% specificity, 100% PPV, 75% NPV, and 84.6% accuracy in the detection of vascular invasion. However, the previously mentioned methods are insufficient for clinical decision-making because these methods cannot show the spatial relationship between the lesions and the adjacent major vessels.\n\nCinematic rendering (CR), a new three-dimensional (3D) technique for CT image postprocessing, presents a photorealistic appearance of CT image data, with the potential to more accurately depict anatomic details and allow perception of shapes and soft tissue structures.^[@bib8]--[@bib10]^ It allows for a more intuitive understanding of the spatial anatomical structures around tumors, especially the spatial relationship between the lesions and the adjacent major vessels.^[@bib11],[@bib12]^ Thus, CR may play a crucial role in preoperative planning and clinical decision-making for STSs of the extremities.\n\nThe purpose of this study was to assess the value of CR from CT data in comparison with CEAI and VR in evaluating the relationship between deep STSs of the extremities and the adjacent major vessels.\n\nMATERIALS AND METHODS\n=====================\n\nThis was a single-center retrospective study that followed the Declaration of Helsinki and good clinical practice guidelines. This retrospective study was approved by the local medical ethics committee. The requirement for written informed consent was waived by the institutional review board.\n\nPatients\n--------\n\nA total of 55 consecutive patients with deep soft tissue tumors of the extremities were retrospectively included from January 2013 to June 2016. The inclusion criteria were as follows:\n\n1.  \\(1\\) Age 18 years or older.\n\n2.  \\(2\\) Presence of deep soft tissue tumors, which were defined as lesions located in the subfascial space, muscle, or muscle space, as well as adjacent to the bone.\n\n3.  \\(3\\) Completion of preoperative CT examinations, including unenhanced and enhanced scans, before fine-needle biopsy.\n\nThe exclusion criteria were as follows:\n\n1.  \\(1\\) An extended interval (more than 4 weeks) between preoperative CT examination and surgery (n = 2).\n\n2.  \\(2\\) Incomplete CT image information (n = 7).\n\n3.  \\(3\\) Surgical histopathological findings showing that the lesions were not STSs (n = 2).\n\n4.  \\(4\\) Diffusely distributed lesions involving multiple tissue layers, which made it difficult to determine the boundary of lesions and the relationships with large blood vessels and bones (n = 1).\n\nBased on the previous inclusion criteria and exclusion criteria, a total of 43 patients were finally included (Fig. [1](#F1){ref-type=\"fig\"}).\n\n![Flowchart of the study population.](rct-43-386-g001){#F1}\n\nComputed Tomography Protocols\n-----------------------------\n\nCT angiography examinations were performed using a 128-slice CT scanner (SOMATOM Definition AS+; Siemens Medical Solutions, Germany) with a standardized protocol. Patients were placed in the supine position on a CT scan table. An unenhanced spiral acquisition of the extremities was planned to include the tumors and at least one adjacent joint with the following parameters: tube voltage, 120 kV; tube current, 100 mAs; collimator width, 128 \u00d7 0.6 mm; slice thickness, 5 mm; slice spacing, 5 mm; and matrix size, 512 \u00d7 512. Contrast bolus tracking was used with a trigger threshold of 100 HU over the region of interest. The trigger point was located in the aortic arch, abdominal aorta, or iliac artery (depending on the tumor location) with the following parameters: tube voltage, 120 kV; tube current, 100 mAs; collimator width, 128 \u00d7 0.6 mm; pitch, 0.8; gantry rotation speed, 0.5 seconds; and variable field of view (depending on the size of the limb). Reconstruction and additional postprocessing were performed by a radiologist on a reformatting workstation (Syngo Via VB10; Siemens Healthcare, Forchheim, Germany). A nonionic contrast agent (Ultravist 300; Bayer Schering Pharma AG, Berlin, Germany) was injected with antecubital venous access at a flow rate of 4.0 mL/s. A total of 80 to 120 mL (1.5 mL per kg of body weight) was injected with a CT-compatible power injector (Bracco ACIST EZEM; Empower CT angiography) followed by injection of 30 mL of saline solution at the same rate. The scan ranges of unenhanced and enhancement images were consistent.\n\nThree-Dimensional Image Postprocessing Methods\n----------------------------------------------\n\nThe raw imaging data of the CT-enhanced arterial phase scan were transferred for 3D postprocessing techniques (Syngo Via VB10; Siemens Healthcare). Volume rendering and CR images were reconstructed by an experienced skeletal radiologist (Reader 1, with 9 years of experience).\n\nThe VR reconstruction parameters, including regions of interest, were selected automatically based on the software presettings and were customized for extremity and CT angiographies. Segmental evaluation was performed by including only the regions of interest instead of evaluating the whole data set if the coverage was long, because it was in run-off studies. In addition to displaying vascular anatomy, VR reconstruction of extremities also showed osseous anatomy, which was essential for surgical planning. The CR reconstruction parameters were also selected automatically by the software presetting named Abdshaded B, which included the following settings: diffuse, 0.5; specular, 0.5; lightmap eucalyptus grove, grey; aperture, tiny; focal plane, center; albedo, 0.85; surface definition, 10; and resolution, 512 \u00d7 512.\n\nImage Review\n------------\n\nAll images were prospectively evaluated by 2 experienced musculoskeletal radiologists (reader 1 with 9 years of experience and reader 2 with 17 years of experience). At the beginning of the study, the 2 readers were provided with hands-on instruction that clearly explained the relationship between the tumors and adjacent major vessels on the 3 types of CT images (CEAI, VR, and CR) used in this study. The definitions of the relationships on CEAI, VR, and CR were provided by referring to the literature in combination with insights from clinical practice.^[@bib6],[@bib13],[@bib14]^ Contrast-enhanced axial images were used to classify the relationship between the mass and the major vessels: (*a*) type 1: distance of the mass from the major vessel of greater than 1 cm; (*b*) type 2: mass adjacent to the major vessel with evidence of a thin adipose film interposition; (*c*) type 3: mass adjacent to vessel without adipose film interposition; and (*d*) type 4a: vessel partially encased and type 4b: totally encased in the lesion.^[@bib4]^ Volume rendering and CR were used to classify the relationship between tumors and major vessels as follows: type 1: normal route of the artery, without stenosis or occlusion interruption; type 2: vascular displacement; type 3: vascular stenosis; and type 4: vascular occlusion.^[@bib13],[@bib14]^ For each observation, each reader was blinded to the findings recorded by the other reader. Reader 1 repeated the image review of the relationship within a month of the previous assessment by following the same procedure. At the end of this process, most disagreements in the definitions of each type of relationship were discussed until a consensus was reached. On a few occasions when a consensus could not be reached by the 2 readers, a third reader, who was a more experienced radiologist and independent from the first two in the image review section, was invited to make the final judgment.\n\nReference Standard\n------------------\n\nComputed tomography evaluations and biopsies were performed for all patients, and the surgeons were aware of the results of the CT imaging evaluations before the surgery because of ethical reasons. During the surgery, the surgeons had to determine the relationship between the tumors and their adjacent major vessels, and their findings were considered as the reference standard. If the surgeons failed to dissect the adjacent major vessels from the tumors, further vascular surgery or amputation was necessitated, which was considered to indicate surgically positive findings for major vessel invasion. Surgically positive findings were also considered if the adjacent vessels adhered to or encased by the tumors could be separated. Surgically negative findings were considered if normal tissue could be found between the vessels and the tumors and the adjacent major vessels could be removed from the tumors.\n\nStatistical Analysis\n--------------------\n\nThe statistical analyses were performed with R 3.4.0 (<http://www.R-project.org>). Cohen \u03ba statistics and 95% confidence intervals (CIs) were used to assess the intrareader and interreader agreement for the type of relationship between tumors and adjacent major vessels on the 3 types of CT images (CEAI, VR, and CR). The \u03ba values were interpreted as follows: poor agreement, 0.01 to 0.20; fair agreement, 0.21 to 0.40; moderate agreement, 0.41 to 0.60; good agreement, 0.61 to 0.80; and excellent agreement, 0.81 to 0.99.\n\nMean values \u00b1 standard deviations were provided for normally distributed variables. Frequencies were provided for categorical data. The \u03c7^2^ test was used to compare the categorical data in the 2 groups.\n\nThe intraoperative findings for vascular invasion were considered as the reference standard. The sensitivity, specificity, PPV, NPV, and accuracy of CEAI, VR, and CR were calculated for the presence of tumoral invasion of adjacent major vessels. The receiver operating characteristic (ROC) curve was plotted and the area under the curve (AUC) was calculated to discriminate the presence of tumoral invasion. The consistency of different ROC values was examined by the DeLong test. The test level was \u03b1 = 0.05, and *P* \\< 0.05 was considered statistically significant.\n\nRESULTS\n=======\n\nPatient Characteristics\n-----------------------\n\nA total of 43 patients (age, 49.51 \u00b1 18.22 years) were involved in the study. Among these patients, 20.93% (9/43) showed surgically positive findings for tumor invasion of adjacent major vessels (Fig. [2](#F2){ref-type=\"fig\"}), and 79.07% (34/43) showed surgically negative findings (Fig. [3](#F3){ref-type=\"fig\"}). The detailed clinical characteristics are listed in Table [1](#T1){ref-type=\"table\"}.\n\n![Contrast-enhanced CT image showing an intramuscular mass in the lower segment of the right thigh in a 25-year-old woman. The histological diagnosis was myofibroblastoma of soft tissue. A, Contrast-enhanced axial imaging: the mass has a well-defined margin and a maximum diameter of 6.5 cm, and the femoral artery (arrow) is totally enclosed by the mass (type 4b). B, Optimization of mass display and longitudinal extent with VR. The femoral artery is narrow, and the shape is irregular but still continuous (type 3). C, Cinematic rendering shows more natural and photorealistic anatomic details, depicting the narrowing of the femoral artery due to complete entrapment by the mass (type 3). Figure [2](#F2){ref-type=\"fig\"} can be viewed online in color at [www.jcat.org](http://www.jcat.org).](rct-43-386-g002){#F2}\n\n![Contrast-enhanced CT image showing a large intramuscular mass in the upper segment of the right thigh in a 51-year-old man. The histological diagnosis was undifferentiated sarcoma. A, Contrast-enhanced axial imaging: the mass grows in the vastus medialis muscle with a maximum diameter of 12.4 cm, and the contrast enhancement is mildly heterogeneous and regular; the femoral artery (arrow) is dislocated by the mass, and normal tissues are interposed with a spacing \\>1 cm (type 1). B, Volume rendering: femoral artery is running in a natural and well-filled manner, with no abnormalities (type 1). C, Cinematic rendering shows the relationship between the mass and the femoral artery with exquisite detail, and the femoral artery is normal with a regular shape (type 1). Figure [3](#F3){ref-type=\"fig\"} can be viewed online in color at [www.jcat.org](http://www.jcat.org).](rct-43-386-g003){#F3}\n\n###### \n\nClinical Characteristics of the 43 Patients\n\n![](rct-43-386-g004)\n\nRepeatability for the 3 Types of CT Images\n------------------------------------------\n\nThe intrareader and interreader agreement values for assessment of the relationship between tumors and adjacent major vessels on the 3 types of CT images are shown in Table [2](#T2){ref-type=\"table\"}. Intrareader agreement values on CEAI, VR, and CR were all excellent (0.984 \\[95% CI = 0.953--1.000\\], 0.934 \\[95% CI = 0.846--1.000\\], and 0.914 \\[95% CI = 0.796--1.000\\]). However, CEAI showed greater interreader agreement than VR and CR (0.969 \\[95% CI = 0.927--1.000\\] vs 0.804 \\[95% CI = 0.682--0.926\\] and 0.761 \\[95% CI = 0.590--0.932\\]).\n\n###### \n\nIntrareader and Interreader Agreement for the Relationship Between Tumors and Adjacent Major Vessels on 3 Types of CT Images\n\n![](rct-43-386-g005)\n\nComparison of the Diagnostic Performance of the 3 Types of CT Images\n--------------------------------------------------------------------\n\nThe relationship between tumors and adjacent major vessels of 43 patients with deep STSs of the extremities on the 3 types of CT images is summarized in Supplemental Digital Content 1, <http://links.lww.com/RCT/A77>. There were all statistically significant difference in the relationship on the 3 types of CT images between the major vessel invasion surgically negative group and positive group (all *P* \\< 0.05). Receiver operating characteristic analysis demonstrated that for CEAI, the optimal criterion was a type 3 or higher. For both VR and CR, the optimal criterion was type 2 or higher. The performance with each optical scale stratified by the 3 types of CT images for determination of tumoral invasion in adjacent major vessels is demonstrated in Table [3](#T3){ref-type=\"table\"} and Figure [4](#F4){ref-type=\"fig\"}.\n\n###### \n\nPerformance of the 3 Types of CT Images for Diagnosis of Major Vessel Invasion\n\n![](rct-43-386-g006)\n\n![The ROC curves of the 3 types of CT images in determining major vessel invasion. Figure [4](#F4){ref-type=\"fig\"} can be viewed online in color at [www.jcat.org](http://www.jcat.org).](rct-43-386-g007){#F4}\n\nIn comparison with CEAI and VR, CR showed lower AUC (0.770 \\[95% CI = 0.582--0.957\\]), accuracy (0.698 \\[95% CI = 0.539--0.828\\]), sensitivity (0.778 \\[95% CI = 0.402--0.960\\]), specificity (0.676 \\[95% CI = 0.494--0.820\\]), PPVs (0.389), and NPVs (0.920) for vascular invasion diagnosis; the accuracy, sensitivity, specificity, PPVs, and NPVs increased to 0.767 \\[95% CI = 0.614--0.882\\], 0.889 \\[95% CI = 0.504--0.994\\], 0.735 \\[95% CI = 0.553--0.864\\], 0.471, and 0.962 for both CEAI and VR, whereas AUC increased to 0.922 \\[95% CI = 0.830--1.000\\] for CEAI and 0.858 \\[95% CI = 0.711--1.000\\] for VR. The results were not statistically significant (all *P* \\> 0.05).\n\nDISCUSSION\n==========\n\nThe main finding of this study was that CR yielded a similar accuracy (69.8%), sensitivity (77.8%), specificity (67.6%), PPV (38.9%), and NPV (92.0%) as CEAI and VR in the detection of the adjacent major vascular invasion in patients with deep STSs of the extremities. To our knowledge, this is a pioneering finding indicating the value of CR in evaluating the relationship between the deep STSs of the extremities and adjacent major vessels.\n\nThe accuracy of CEAI and VR in this study was less than in previous studies.^[@bib7],[@bib13],[@bib15]^ This difference might be due to the location of the lesions assessed in our study. These lesions were located in the deep soft tissues of the extremities rather than superficial soft tissues, which were more prone to cause adjacent vascular displacement and even reduction of the arterial lumen, or result in disappearance of adipose film interposition. These factors can enhance the relationship between tumors and adjacent major vessels, leading to a lower PPV for adjacent major vascular invasion. Our results also showed that the diagnostic accuracies of CR were lower than that of CEAI and VR, which would be caused by the lack of training of these radiologists.\n\nInterestingly, our data demonstrated that the intrareader agreement for CR was excellent (91.4%), similar to that for CEAI (98.4%) and VR (93.4%), where the interreader agreement for CR was good (76.1%), similar to that for VR (80.4%) but lower than that for CEAI (96.9%). This may be related to variations in the understanding of the types of relationships between tumors and adjacent major vessels on CR and VR among different radiologists, because the 2 readers were only provided with hands-on instruction regarding the types of relationships on the 3 types of CT images (CEAI, VR, and CR) but not trained together systematically. Therefore, radiologists should receive training for identification of the relationship between tumors and adjacent major vessels on CR before the technique can be used in clinical practice.\n\nOur data demonstrated that CR offered no statistically significant disadvantages over CEAI and VR in the detection of adjacent major vascular invasion in cases of deep STSs of the extremities. Cinematic rendering is a recently introduced novel 3D technique for postprocessing CT image data. The existing literature claimed that CR shows the best visualization for high-density and high-contrast structures such as contrast-enhanced vessels and simultaneously provides a more natural depiction of the rendered data.^[@bib8]--[@bib11],[@bib16]--[@bib25]^ More importantly, CR allows for a more intuitive understanding of the spatial anatomical structures around tumors, especially the spatial relationship between the lesions and the adjacent major vessels.^[@bib11],[@bib12]^ However, to our knowledge, there is no literature on the advantages and disadvantages of CR over conventional 2D images and VR for CT imaging data, which is necessary to assess a new technology. As a preliminary study, our study results provide initial evidence that CR may have some possible disadvantages such as the lower PPV for adjacent major vascular invasion, which were not statistically significant but clinically significant, but further studies are required to validate these findings.\n\nOur study also has some limitations. First, the sample size was small. Twelve patients were excluded from our initial population because of the exclusion criteria, yielding a small but homogeneous population. Second, caution should be exercised in interpreting the absolute statistical values because of comparatively few cases of the adjacent major vascular invasion.\n\nIn conclusion, although not statistically significant, CR was less accurate and showed lower interreader agreement for vascular invasion of STSs compared with traditional contrast-enhanced axial CT images. Although cinematic rendering may allow for more detailed visualization and surgical planning, it should be combined with traditional imaging modalities for evaluation of vascular invasion.\n\nSupplementary Material\n======================\n\n###### SUPPLEMENTARY MATERIAL\n\nThis study was supported by the Applied Basic Research Projects of Yunnan Province, China (No. 2018FE001-065 and No. 2018FE001-251).\n\nThe authors declare no conflict of interest.\n\nK.L. and R.Y. contributed equally to this study.\n\nSupplemental digital contents are available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site ([www.jcat.org](http://www.jcat.org)).\n"}
+{"text": "introduction {#mdv482_s1}\n============\n\nAntiemetic prophylaxis for chemotherapy-induced nausea and vomiting (CINV) is an important component of cancer treatment management. In the absence of antiemetic prophylaxis, the risk of emesis with antineoplastic agents classified as moderately emetogenic chemotherapy (MEC) is 30%--90%, and \\>90% with highly emetogenic chemotherapy (HEC). CINV can occur within the acute (0--24 h) or delayed (25--120 h) phases of chemotherapy, with increased severity in the delayed setting \\[[@MDV482C1]\\].\n\nAntiemetic treatment guidelines indicate that 5-HT~3~ receptor antagonists (RAs) effectively prevent and control CINV during the acute phase in subjects receiving MEC or HEC \\[[@MDV482C2]--[@MDV482C5]\\], but are generally less effective in preventing CINV in the delayed phase \\[[@MDV482C6], [@MDV482C7]\\]. Aprepitant and fosaprepitant dimeglumine, the water-soluble prodrug that is rapidly converted to aprepitant after intravenous (i.v.) administration, are potent and selective neurokinin-1 (NK~1~) RAs that are effective against CINV in both acute and delayed phases when added to a standard antiemetic regimen (a 5-HT~3~ RA and dexamethasone) in subjects receiving HEC and AC-based chemotherapy \\[[@MDV482C8]--[@MDV482C10]\\]. Although approved for the prevention of MEC- or HEC-associated CINV \\[[@MDV482C11], [@MDV482C12]\\], NK~1~ RA utility in non-AC MEC recipients has been debated, emphasizing the need for well-designed randomized studies to better define their role in this setting \\[[@MDV482C13]\\].\n\nThis study is the first to directly assess the efficacy and safety of a single 150-mg i.v. dose of fosaprepitant combined with a 5-HT~3~ RA and a corticosteroid versus a standard regimen \\[[@MDV482C5]\\] of 5-HT~3~ RA plus corticosteroid-alone for the prevention of CINV in a well-defined non-AC MEC population.\n\nmethods {#mdv482_s2}\n=======\n\nstudy design {#mdv482_s2a}\n------------\n\nThis international, phase III, randomized, double-blind, active-comparator, parallel-group, multicenter, superiority trial (PN031) was conducted at 125 sites across 30 countries. The trial was conducted in accordance with the Declaration of Helsinki and the Good Clinical Practice guidelines. Approval from the independent ethics committee or institutional review board was obtained for each participating center. All subjects provided written informed consent before enrollment.\n\npatients {#mdv482_s2b}\n--------\n\nSubjects aged \u226518 years with confirmed malignant disease, who were treatment naive to MEC and HEC (as defined in the Hesketh classification of emetogenic chemotherapy agents \\[[@MDV482C14]\\]), were eligible. All subjects had to be scheduled to receive \u22651 i.v. dose of MEC on day 1. Combinations of MEC \u00b1 a low emetogenic chemotherapy (LEC) were permitted from days 1--3 when part of an overall MEC regimen and were in accordance with current emetogenicity classification guidelines \\[[@MDV482C2], [@MDV482C3]\\]. Although AC regimens have been considered MEC in previous clinical trials, updated treatment guidelines now consider AC regimens to be HEC \\[[@MDV482C2]--[@MDV482C5]\\]. As a result, AC regimens were not allowed. Minimally emetogenic chemotherapy was permitted throughout the treatment period. Additional cycles of chemotherapy were permitted after the efficacy period.\n\nThe major exclusion criteria for this study were vomiting in the 24-h period before day 1, antiemetic use within 48 h of day 1, symptomatic primary or metastatic central nervous system malignancy causing nausea and/or vomiting, and the use of any dose of cisplatin or other HEC.\n\nrandomization and blinding {#mdv482_s2c}\n--------------------------\n\nSubjects were randomized (1 : 1) to the single-dose fosaprepitant or control regimen via an interactive voice response system/interactive web response system, and stratified based on sex. Study medications were supplied in a blinded manner as fosaprepitant/placebo i.v. bags, ondansetron/placebo capsules, and dexamethasone/placebo capsules.\n\nstudy treatments {#mdv482_s2d}\n----------------\n\nSubjects received fosaprepitant (fosaprepitant regimen) or placebo (control regimen) on top of ondansetron plus dexamethasone in accordance with current antiemetic guidelines \\[[@MDV482C2], [@MDV482C3], [@MDV482C5]\\] and in agreement with regulatory guidance on the study design ([supplementary Table S1, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1)). For the fosaprepitant regimen, i.v. fosaprepitant as a single 150-mg dose was administered \u223c30 min before MEC initiation on day 1. For both regimens, oral ondansetron and oral dexamethasone were taken before MEC on day 1, followed by oral ondansetron 8 h after the first dose. On days 2 and 3, subjects in the control group received ondansetron every 12 h, whereas those in the fosaprepitant group received matching placebo.\n\nThe use of investigator-prescribed rescue medication (e.g. 5-HT~3~ RA, phenothiazines, butyrophenones, benzamides, corticosteroids, benzodiazepines, and domperidone) was permitted throughout the study to alleviate symptoms of established nausea or vomiting.\n\nend points {#mdv482_s2e}\n----------\n\nThe primary efficacy end point was the proportion of subjects who achieved complete response (CR; no vomiting and no use of rescue medication) during the delayed phase (25--120 h following initiation of the first MEC dose). Secondary efficacy end points included the proportions of subjects who achieved CR during the overall and acute phases (0--120 and 0--24 h after MEC initiation, respectively) and the proportion of subjects with no vomiting (regardless of rescue medication use) during the overall phase. A list of exploratory end points assessed in this study is provided in [supplementary Materials, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1).\n\nClinical adverse events (AEs) were graded according to National Cancer Institute Common Toxicity Criteria, version 4.0. Infusion-related reactions, including infusion-site thrombophlebitis and severe infusion-site pain, erythema, and induration, were prespecified as events of clinical interest. Additionally, vomiting was reported as an AE if the vomiting episode occurred outside of the efficacy assessment period or met criteria for a serious adverse event (SAE).\n\nstatistical analysis {#mdv482_s2f}\n--------------------\n\nStatistical methods are described in [supplementary Materials, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1).\n\nThe intent-to-treat (ITT) population (subjects receiving \u22651 dose of study drug and analyzed in their randomized treatment group) was used for the primary efficacy analysis, whereas the all-subjects-as-treated (ASaT) population (all subjects receiving \u22651 dose of study drug and analyzed in the treatment group based on the drug actually received) was used for the safety analysis.\n\nTreatment comparisons for the primary and secondary efficacy analyses included formal tests for superiority using the Cochran--Mantel--Haenszel test, stratified by sex. Superiority of the fosaprepitant versus the control regimen was evaluated using a two-tailed test with *P* \u2264 0.05 indicating a significant difference. No formal test of superiority was carried out for exploratory end points.\n\nDescriptive statistics were provided for demographic variables, baseline characteristics, and AEs. Details on the analyses of exploratory end points and AEs are provided in [supplementary Methods and Results, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1).\n\nresults {#mdv482_s3}\n=======\n\nOverall, 1015 of the 1150 screened subjects were randomized between 30 October 2012 and 03 November 2014. Reasons for nonrandomization are shown in Figure [1](#MDV482F1){ref-type=\"fig\"}. The ITT and ASaT populations comprised 1000 and 1001 subjects, respectively (Figure [1](#MDV482F1){ref-type=\"fig\"}). The overall mean age was 59.6 years. Baseline demographics of the ITT population (Table [1](#MDV482TB1){ref-type=\"table\"}) were similar between treatment groups and balanced with respect to the malignancy type being treated, which were representative of a MEC population. Treatment groups were also balanced regarding the types of chemotherapy regimens used, with most subjects receiving single-day regimens (71.3% and 69.9% for fosaprepitant and control regimens, respectively; [supplementary Table S2, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1)). Chemotherapeutic agents administered were balanced, with carboplatin (\u223c53%) and oxaliplatin (\u223c22%) most commonly used in both treatment groups ([supplementary Table S3, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1)). Overall, 96.6% and 98.4% of subjects completed the study in the fosaprepitant and control groups, respectively. Table 1.Baseline demographics and specific clinical characteristics (ITT population)Fosaprepitant regimen (*N* = 502)Control regimen (*N* = 498)Age \\[mean (SD)\\], years60.0 (11.8)59.1 (12.3)Age \\<50 years, *n* (%)97 (19.3)108 (21.7)Age \u226550 years, *n* (%)405 (80.7)390 (78.3)Sex, *n* (%)\u2003Male204 (40.6)205 (41.2)\u2003Female298 (59.4)293 (58.8)Race, *n* (%)\u2003White424 (84.5)414 (83.1)\u2003Asian21 (4.2)14 (2.8)\u2003Black or African American13 (2.6)8 (1.6)\u2003Other44 (8.8)62 (12.4)Ethnicity\u2003Hispanic or Latino89 (17.7)102 (20.5)Type of malignancy, *n* (%)\u2003Lung129 (25.7)125 (25.1)\u2003Breast110 (21.9)121 (24.3)\u2003Colorectal102 (20.3)91 (18.3)\u2003Gynecologic81 (16.1)71 (14.3)\u2003Gastrointestinal33 (6.6)41 (8.2)\u2003Head and neck12 (2.4)9 (1.8)\u2003Other35 (7.0)40 (8.0)History of motion sickness, *n* (%)28 (5.6)30 (6.0)History of emesis during pregnancy, *n* (%)60 (12.0)61 (12.2)History of alcohol use, *n* (%)224 (44.6)213 (42.8)[^2] Figure 1.CONSORT diagram. ASaT, all subjects as treated; ITT, intent-to-treat.\n\nefficacy {#mdv482_s3a}\n--------\n\nCR in the delayed phase (primary end point) was superior in the fosaprepitant versus the control regimen (treatment difference 10.4%; *P* \\< 0.001) (Figure [2](#MDV482F2){ref-type=\"fig\"}A); this was also consistent in the full analysis set and per-protocol populations (defined in [supplementary Materials, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1)). CR during the overall phase was also superior in the fosaprepitant regimen (treatment difference 10.2%; *P* \\< 0.001). Both regimens had a high CR in the acute phase (treatment difference 2.3%; *P* = 0.184). Figure 2.Proportion of subjects with (A) complete response (CR) and (B) no vomiting in the acute (0--24 h), delayed (25--120 h), and overall (0--120 h) phases following initiation of a first dose of moderately emetogenic chemotherapy.\n\nFosaprepitant regimen was superior to the control regimen for no vomiting in the overall phase (treatment difference 9.8%; *P* \\< 0.001) (Figure [2](#MDV482F2){ref-type=\"fig\"}B). The exploratory end point of no vomiting in the delayed phase was also superior in the fosaprepitant group (treatment difference 8.8%; *P* \\< 0.001). Furthermore, the estimated time-to-first vomiting episode in the overall phase, regardless of rescue medication use, was longer in the fosaprepitant group, compared with the control group (nominal *P* \\< 0.001) ([supplementary Figure S1, available at *Annals of Oncology* online](http://annonc.oxfordjournals.org/lookup/suppl/doi:10.1093/annonc/mdv482/-/DC1)).\n\nThe proportion of subjects with no significant nausea was significantly higher in the fosaprepitant group in the overall phase (83.1% versus 78.3%, *P* = 0.026), but between-group differences in the proportions of subjects with no nausea or no rescue medication use did not reach statistical significance (65.3% versus 61.6%, *P* = 0.156 and 83.9% versus 79.5%, *P* = 0.069, respectively).\n\nThe proportion of subjects with no impact of CINV on daily life \\[Functional Living Index-Emesis (FLIE) total score \\>108\\] was significantly greater for the fosaprepitant versus the control group \\[81% versus 75.5%; odds ratio 1.39; 95% confidence interval (CI) 1.01--1.91; *P* = 0.043\\]. Similar results were observed for individual FLIE domains, with the exception of the vomiting-specific domain (nominal *P* = 0.068).\n\nsafety and tolerability {#mdv482_s3b}\n-----------------------\n\nAEs were reported by 61.3% of subjects in the ASaT population and were comparable between the treatment regimens (Table [2](#MDV482TB2){ref-type=\"table\"}). The AE profile was generally typical of a population receiving emetogenic chemotherapy. The most commonly reported all-grade AEs for the fosaprepitant and control groups were fatigue, diarrhea, and constipation (Table [2](#MDV482TB2){ref-type=\"table\"}). No cases of severe infusion-site pain, erythema, or induration were reported. Three cases (0.6%) of infusion-site thrombophlebitis were reported in the fosaprepitant group, compared with no cases in the control group (treatment difference 0.6%; 95% CI \u22120.2 to 1.7; *P* = 0.085); none of these infusion-site reactions were considered by the investigator to be severe or related to study medication. Table 2.Summary of adverse events (all-subjects-as-treated population)^a^*N* (%)Fosaprepitant regimen (*N* = 504)Control regimen (*N* = 497)Difference versus control regimen, % (95% CI)^b^\u22651 AE312 (61.9)302 (60.8)1.1 (\u22124.9 to 7.2)Drug-related AEs^c^43 (8.5)45 (9.1)\u22120.5 (\u22124.1 to 3.0)Serious AEs39 (7.7)35 (7.0)0.7 (\u22122.6 to 4.0)Serious drug-related AEs^c^1 (0.2)2 (0.4)\u22120.2 (\u22121.3 to 0.7)Death^d^8 (1.6)2 (0.4)1.2 (\u22120.1 to 2.7)Discontinuation^e^ due to AE2 (0.4)2 (0.4)\u22120.0 (\u22121.1 to 1.1)Commonly reported AEs (\u22655% of subjects)\u2003Fatigue76 (15.1)64 (12.9)2.2 (\u22122.1 to 6.5)\u2003Diarrhea64 (12.7)56 (11.3)1.4 (\u22122.6 to 5.5)\u2003Constipation47 (9.3)52 (10.5)\u22121.1 (\u22124.9 to 2.6)\u2003Neutropenia41 (8.1)37 (7.4)0.7 (\u22122.7 to 4.1)\u2003Headache30 (6.0)35 (7.0)\u22121.1 (\u22124.2 to 2.0)\u2003Decreased appetite27 (5.4)32 (6.4)\u22121.1 (\u22124.1 to 1.9)\u2003Alopecia11 (2.2)26 (5.2)\u22123.0 (\u22125.6 to \u22120.7)[^3][^4][^5][^6][^7][^8]\n\nMost AEs were grade 1--2 in severity, with neutropenia being the most commonly reported grade 3--4 AE in both the fosaprepitant and control groups (5.2% versus 4.8%).\n\nOverall, SAEs were reported in 74 subjects (7.4%); the most commonly reported was febrile neutropenia (1.6% and 1.0% of fosaprepitant and control regimen recipients, respectively). Drug-related SAEs were reported in one and two subjects receiving the fosaprepitant (hypersensitivity reaction) and control (one worsening constipation; one allergic reaction) regimens, respectively. Sixteen randomized subjects died, with the day of death relative to treatment day 1 ranging from day 4 to 90. Of these subjects, 10 had AEs with an onset date that occurred between day 1 (baseline) and day 17 (14 days after the last dose of study medication), inclusive, that resulted in death (Table [2](#MDV482TB2){ref-type=\"table\"}). All cases appeared to be attributable to subjects\\' underlying malignancy, other pre-existing conditions, and/or effects of chemotherapy; no deaths were considered by investigators to be related to the study drug.\n\ndiscussion {#mdv482_s4}\n==========\n\nThis is the first study to provide efficacy and safety data on a single 150-mg i.v. dose of fosaprepitant added to a 5-HT~3~ RA and a corticosteroid for the prevention of CINV in adults receiving non-AC MEC. Overall, this single-day, triple-antiemetic fosaprepitant regimen proved to be superior to a standard 3-day control antiemetic regimen for the prevention of CINV associated with MEC.\n\nThe superior results in the delayed and overall phases were similar to those of previous NK~1~ RA clinical trials in subjects receiving HEC \\[[@MDV482C8]--[@MDV482C10], [@MDV482C15]--[@MDV482C18]\\]. Furthermore, consistent with antiemetic guidelines, treatment differences of \\>10% in favor of the fosaprepitant regimen for CR (delayed and overall phases) in the current study are considered to be clinically meaningful for patients \\[[@MDV482C1], [@MDV482C3]\\].\n\nIn contrast to previous studies, the fosaprepitant regimen did not improve CR during the acute phase. This may be attributed in part to a higher than expected CR rate (92%) in the acute phase of the control group compared with that of previous studies (49%--85%) \\[[@MDV482C8]--[@MDV482C10], [@MDV482C15], [@MDV482C16], [@MDV482C18]\\]. Differences in study populations may have been a factor; subjects in previous studies primarily received cisplatin or AC-based MEC regimens \\[[@MDV482C8]--[@MDV482C10], [@MDV482C15], [@MDV482C16], [@MDV482C18]\\], which are now classified as HEC under current treatment guidelines \\[[@MDV482C2], [@MDV482C5]\\]. *Post hoc* analysis of a phase III trial consisting of a broad range of MEC (AC and non-AC regimens) reported CR improvements with an aprepitant regimen in subjects receiving non-AC MEC \\[[@MDV482C17]\\]. Another phase III trial also reported CR improvements with the NK~1~ RA, rolapitant, in a large MEC population stratified based on AC and non-AC regimens; \\>50% of subjects in the study received AC-based regimens \\[[@MDV482C19]\\].\n\nTo our knowledge, this is the first large, global, randomized, controlled superiority trial to prospectively evaluate treatment with a single-dose i.v. NK~1~ RA in a well-characterized non-AC MEC population.\n\nThe fosaprepitant regimen also provided significant improvements in the secondary and exploratory efficacy end points, such as no vomiting in the overall and delayed phases, no significant nausea in the overall phase, and improved quality of life, consistent with previous NK~1~ RA trials \\[[@MDV482C8]--[@MDV482C10], [@MDV482C15], [@MDV482C17], [@MDV482C18]\\]. Taken together, the overall efficacy findings support the clinical benefit provided by the fosaprepitant regimen in a non-AC MEC population.\n\nThe fosaprepitant regimen was generally well tolerated in the current study, and no new safety signals were noted compared with previous fosaprepitant studies \\[[@MDV482C15], [@MDV482C20]\\]. AE profiles for the two treatment regimens were similar and consistent with those in subjects receiving emetogenic chemotherapy \\[[@MDV482C21]\\]. An imbalance in deaths between the treatment arms was observed, but a careful review of each case suggested that these were likely to reflect progression of the underlying disease process, and no deaths were assessed as likely to be related to the study drug by the investigators. Finally, although higher infusion-site AE rates have been previously observed with fosaprepitant \\[[@MDV482C15], [@MDV482C20], [@MDV482C22]\\], only three cases of infusion-site thrombophlebitis were reported in the fosaprepitant group for this study.\n\nThis study included a large, well-balanced non-AC MEC population regarding the types of chemotherapy regimens and agents administered. While these findings indicate that a single-dose fosaprepitant regimen can potentially offer the convenience of completing all antiemetic treatment before MEC initiation in a single day, the study was not designed to evaluate treatment differences among subgroups across the full heterogeneity of the MEC population. Unlike HEC regimens, which are generally homogeneous in terms of emetogenic potential (\\>90%), the risk of emesis in MEC populations ranges from 30% to 90% and is complicated by factors such as chemotherapy sequence and dosing, as well as multiday treatment regimens \\[[@MDV482C2]--[@MDV482C5]\\]. However, in this trial, the majority of subjects received a single-day, non-AC MEC regimen by design, with carboplatin being the most commonly used MEC agent. Although less emetogenic than the classic HEC agent, cisplatin, carboplatin is likely to be in the 'upper end' of the current MEC emetogenicity classification. Additionally, MEC--LEC combination regimens may increase the risk of emesis versus MEC or LEC alone. Antiemetic guidelines advise that additional antiemetic treatment may be needed for chemotherapy regimens that extend beyond day 1 because the risk periods for acute and delayed emesis overlap after the first day of chemotherapy \\[[@MDV482C2], [@MDV482C3]\\]. The role of NK~1~ RA, such as i.v. fosaprepitant, in the non-AC MEC setting may be better defined through additional subgroup analyses of this study population to identify specific MEC subpopulations that are more likely to respond to antiemetic treatment and to determine whether additional prophylactic antiemetic treatment is warranted for multiday chemotherapy regimens. While it is recognized that additional studies may be needed to fine-tune future antiemetic guidelines within the heterogeneous MEC population (e.g. exploring differences between carboplatin and noncarboplatin-containing regimens and between single-day versus multiday antiemetic regimens), our findings clearly demonstrate the benefit of fosaprepitant for the prevention of CINV in a well-defined MEC population. This adds to the available evidence for aprepitant in non-AC-based MEC regimens \\[[@MDV482C17]\\], and suggests that the role of adding an NK~1~ RA in the overall MEC setting, such as aprepitant/fosaprepitant, in all antiemetic guidelines warrants further discussion.\n\nfunding {#mdv482_s5}\n=======\n\nThis study was not supported by a grant. This work was supported by Merck & Co., Inc., Kenilworth, NJ, USA.\n\ndisclosure {#mdv482_s6}\n==========\n\nCW is an employee and stockholder of Merck & Co., Inc. KJ has received honoraria for consultancy from Merck Sharp & Dohme (MSD), Merck & Co., Inc., Helsinn, and Pro-Strakan. SG, EBB, and WV are employees and stockholders of Merck & Co., Inc., and LWL is an employee of Merck & Co., Inc. SN has received honoraria from Millenium and served as a consultant or in an advisory role for Amgen and Millenium (now Takeda Oncology). BLR has served as a consultant or in an advisory role for and has received payment for travel, accommodations, or expenses from MSD and Tesaro. BLR has also received payments for participation on speakers bureaus for MSD and Roche.\n\nSupplementary Material\n======================\n\n###### Supplementary Data\n\nThe authors thank Lydia Kevill for assisting with study supervision and data collection. Medical writing and editorial assistance were provided by Maxwell Chang and Traci Stuve of ApotheCom, Yardley, PA, USA, with funding provided by Merck & Co., Inc., Kenilworth, NJ, USA.\n\n[^1]: Results from this study have been previously presented in part at the American Society of Clinical Oncology Annual Meeting; 29 May--2 June 2015; Chicago, IL, USA.\n\n[^2]: ITT, intent to treat; SD, standard deviation.\n\n[^3]: ^a^One cross-treated subject was randomized to the control group but received fosaprepitant in error; this subject was included in the fosaprepitant group for safety and the control group for efficacy. One subject with missing informed consent received fosaprepitant and was included in the fosaprepitant safety group.\n\n[^4]: ^b^Based on Miettinen and Nurminen method.\n\n[^5]: ^c^Determined by the investigator to be related to any of the study drugs.\n\n[^6]: ^d^Deaths resulting from AEs that had an onset date between the baseline and safety follow-up period (i.e. days 1--17, inclusive).\n\n[^7]: ^e^Study medication withdrawn.\n\n[^8]: AE: adverse event; CI: confidence interval.\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nHuman pluripotent stem cells (hPSCs), such as human embryonic stem cells (hESCs) \\[[@bib1]\\] and human induced pluripotent stem cells (hiPSCs) \\[[@bib2]\\], are expected to be applied to cell therapy, disease modelling, and drug development. It is possible to create hiPSCs from various types of somatic cells by the introduction of reprogramming factors, and various research projects using hiPSCs are ongoing.\n\nClinical trials using hESC-derived retinal pigment epithelium (RPE) and autologous hiPSC-derived RPE have been reported \\[[@bib3], [@bib4], [@bib5]\\], and a clinical trial using allogeneic iPSC-derived RPE is underway. Another has been initiated for banked allogeneic hiPSCs to treat Parkinson\\'s disease \\[[@bib6]\\]. Further clinical trials have been planned to treat heart failure, spinal cord injury, Parkinson\\'s disease, and aplastic anaemia, using cells created from hiPSCs \\[[@bib7], [@bib8], [@bib9]\\]. Although hiPSC-based clinical and industrial applications are becoming realistic, it is still a major safety concern that residual hiPSCs in products for cell therapy could form tumours in transplanted sites \\[[@bib10], [@bib11], [@bib12]\\]. The risk of teratoma formation by hPSCs has been reported in various animal studies \\[[@bib13], [@bib14], [@bib15], [@bib16], [@bib17]\\]. It is reported that only 100 iPS cells are sufficient to produce a teratoma in a mouse model \\[[@bib13],[@bib18]\\]. Therefore, the establishment of a method to detect and eliminate residual hPSCs from cell products without impairing the survival rate and functionality of the cells to be used for cell therapy, disease modelling, and drug development, is required.\n\nSeveral strategies have been reported for selectively eliminating residual hPSCs from differentiated cells, including expressing a suicide gene into hPSCs \\[[@bib19]\\], or using cytotoxic antibodies \\[[@bib20], [@bib21], [@bib22]\\], chemical inhibitors \\[[@bib23], [@bib24], [@bib25], [@bib26]\\], and synthetic RNA or peptide \\[[@bib27],[@bib28]\\]. Cell sorting methods using hPSC-specific antibodies \\[[@bib20],[@bib29],[@bib30]\\] and lectins \\[[@bib31]\\] have also been proposed. However, all of these methods have limitations with respect to specificity, throughput, and other issues, and therefore it is still necessary to develop alternative or additional technologies based on different mechanisms.\n\nWe have previously reported that a lectin designated recombinant N-terminal domain of BC2L-C lectin derived from *Burkholderia cenocepacia* (rBC2LCN) binds to various types of hiPSCs and hESCs, but not to differentiated somatic cells \\[[@bib32],[@bib33]\\]. This lectin binds specifically to the Fuc1-2Gal1-3 motif that is highly expressed on hiPSCs \\[[@bib32],[@bib34]\\]. In addition, podocalyxin, a type1 transmembrane protein, was identified as a predominant glycoprotein ligand of rBC2LCN \\[[@bib35]\\]. As its main practical applications, fluorescence-labelled rBC2LCN allows live staining of hESCs/hiPSCs following its addition to the culture medium and is capable of separating live hPSCs by flow cytometry \\[[@bib33]\\]. The staining is specific to undifferentiated cells and rapidly diminishes depending on their differentiation. Furthermore, based on the finding that rBC2LCN was internalised inside hPSCs after binding to the surface of these cells, recombinant lectin-toxin fusion proteins in which rBC2LCN was fused to several domains of *Pseudomonas aeruginosa* exotoxin A was developed for selective elimination of hPSCs \\[[@bib36],[@bib37]\\].\n\nIn this study, we demonstrate an additional application of rBC2LCN, namely its potential in magnetic bead-based cell separation for reduction of tumourigenic hPSCs from differentiated cell populations. We evaluated cell separation efficiency by flow cytometry and digital PCR analyses. Effective elimination of hPSCs was also verified in a teratoma formation assay in a mouse model.\n\n2. Materials and methods {#sec2}\n========================\n\n2.1. Cell culture {#sec2.1}\n-----------------\n\nThe human ES cell line H9 hNanog-pGZ \\[[@bib1]\\] was maintained in mTeSR1 (STEMCELL Technologies, Vancouver, BC, Canada) on a BD Matrigel growth factor reduced (GFR) matrix (BD Biosciences, San Jose, CA, USA) with zeocin, according to the WiCell feeder independent pluripotent stem cell protocols provided by the WiCell Research Institute ([www.wicell.org](http://www.wicell.org){#intref0010}). The human iPS cell line 201B7 \\[[@bib2]\\] was maintained in mTeSR1 (STEMCELL Technologies) on the BD Matrigel hESC-qualified matrix (BD Biosciences), according to the manufacturer\\'s instructions (STEMCELL Technologies). HDF (ATCC PCS-201-012) was maintained in 10% FBS containing DMEM (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). HDF cells were treated with 10\u00a0\u03bcg/ml of Mitomycin C (Kyowa Hakko Kirin Co., Ltd., Tokyo, Japan) for 120\u00a0min to prevent proliferation. The experiments using hiPSCs and hESCs were approved by the National Institute of Advanced Industrial Science and Technology (AIST) (accreditation numbers and hi2016-099).\n\n2.2. Lectin labelling and magnetic cell separation {#sec2.2}\n--------------------------------------------------\n\nRecombinant BC2LCN lectin (rBC2LCN) (FUJIFILM Wako Pure Chemical Corporation) was labelled with a Biotin Labeling Kit-NH2 (Dojindo). Biotin-conjugated rBC2LCN (1--100\u00a0\u03bcg) or biotin-conjugated BSA were incubated with 50\u00a0\u03bcL of Dynabeads M\u2212280 streptavidin (Thermo Fisher Scientific, Waltham, MA, USA) in 1\u00a0ml of MACS buffer \\[0.5% bovine serum albumin (BSA) and 2\u00a0mM EDTA in PBS\\] on a rotator for 30\u00a0min\u00a0at room temperature. After incubation, the beads were rinsed twice with MACS buffer (rBC2LCN-magnetic bead and BSA-magnetic bead).\n\nCells (hESCs and hiPSCs) were dissociated with ESGRO Complete Accutase (Merck Millipore, Billerica, MA, USA) and mixed with HDF in a ratio of 1:1. HDF cells were pre-marked with a CellTrace Violet cell proliferation kit according to the manufacturer\\'s protocol (Thermo Fisher Scientific) or with mitomycin C treated for proliferation inhibition, depending on the following analysis. A total of 2\u00a0\u00d7\u00a010^6^ mixed cells were incubated with 50\u00a0\u03bcL of the rBC2LCN-magnetic bead, BSA-magnetic bead or magnetic bead alone for 30\u00a0min\u00a0at 4\u00a0\u00b0C in 1\u00a0ml of MACS buffer. The suspensions were placed in a DynaMag magnet (Thermo Fisher Scientific) for 2\u00a0min, and the supernatant with untouched cells was collected for flow cytometry, gene expression analysis and teratoma formation assay.\n\n2.3. Flow cytometry {#sec2.3}\n-------------------\n\nFlow cytometry was performed as described previously \\[[@bib33]\\]. The cells were resuspended at approximately 1\u00a0\u00d7\u00a010^6^\u00a0cells/mL in MACS buffer and incubated with anti--TRA-1-60 antibodies (1:300 dilution; clone TRA-1-60, Merck Millipore) for 1\u00a0h\u00a0at 4\u00a0\u00b0C. Normal mouse IgM (Merck Millipore) was used as an isotype control. The cells were rinsed with MACS buffer and then incubated with Alexa Fluor 488 goat anti-mouse IgM (1:300 dilution; Thermo Fisher Scientific). After further rinsing, cells were stained with propidium iodide (PI) (Thermo Fisher Scientific), and 20,000\u00a0cells were analysed using a Cell Sorter SH800Z (Sony Corporation, Tokyo, Japan). The data were analysed with FlowJo software (BD Biosciences).\n\n2.4. Digital droplet polymerase chain reaction (ddPCR) analysis {#sec2.4}\n---------------------------------------------------------------\n\nMagnetically sorted or unsorted cells were stained with propidium iodide (PI), and PI-negative live cells were collected using a cell sorter SH800Z (Sony Corporation). Total RNA was extracted from frozen cell samples using ISOGEN (Nippon Gene Co., Ltd., Tokyo, Japan) and ddPCR was carried out using a QX200 droplet digital PCR system (Bio-Rad, Hercules, CA, USA), according to the manufacturer\\'s instructions. Samples were analysed using 2x One-Step RT-ddPCR Supermix (Bio-Rad). The primers and TaqMan probe sequences, their concentrations, and thermal cycling conditions used in the ddPCR methods were according to a previous publication \\[[@bib38]\\]. The probe and primer sequences are as follows: *LIN28* probe, FAM-CGCATGGGGTTCGGCTTCCTGTCC-BHQ1, *LIN28* forward primer, CACGGTGCGGGCATCTG, *LIN28* reverse primer, CCTTCCATGTGCAGCTTACTC; *NANOG* probe, FAM-TGCTGAGGCCTTCTGCGTCACACC-BHQ1, *NANOG* forward primer, CTCAGCTACAAACAGGTGAAGAC, *NANOG* reverse primer, TCCCTGGTGGTAGGAAGAGTAAA. As the internal control, TaqMan GAPDH Control Reagent was used (Thermo Fisher Scientific). Each reaction was performed in duplicate. Total RNA quantities used for the reactions were 0.05, 5, and 50\u00a0ng for *GAPDH*, *LIN28*, *NANOG*, respectively. Fluorescence intensities of each droplet in samples were measured using a QX200 droplet reader (Bio-Rad). Positive and negative droplets were discriminated and counted by applying a threshold determined manually in QuantaSoft software. The same threshold was applied to all the wells for each gene on one PCR plate. We adopted the larger measurement result of accepted droplets among the duplicated wells. The number of adopted droplets was \\>10,000. The copy number concentration in the sample was calculated using the numbers of positive and accepted droplets. The concentration results in terms of target copies per microliter were provided by QuantaSoft software. The number of target copies per template RNA (ex, copies/50\u00a0ng of RNA in the case of *NANOG*) was calculated as concentration (copies/\u03bcl)\u00a0\u00d7\u00a020 (\u03bcl)\u00a0\u00d7\u00a0dilution factor of template RNA. The dilution factors were 1 for *NANOG*, 10 for *LIN28* and 1000 for *GAPDH*, respectively. Absolute counts of *NANOG* or *LIN28* were normalised to *GAPDH*. Experiments were performed in triplicate and repeated three times with similar results.\n\n2.5. Teratoma formation assay {#sec2.5}\n-----------------------------\n\nEight-week-old immune-deficient NOD/ShiJic-*scid*Jcl mice (CLEA Japan, Inc., Tokyo, Japan) were used for transplantation. NOD/ShiJic-*scid*Jcl mice were anaesthetised using 2% isoflurane and an animal anesthetizer device (MK-AT210D, Muromachi Kikai Co., Ltd., Tokyo, Japan). The surgical area was disinfected with 70% ethanol. After cutting the centre of the scrotum, a testis was carefully pulled out and injected with 20\u00a0\u03bcL (\\~1.0\u00a0\u00d7\u00a010^6^\u00a0cells) of cell suspension with BD Matrigel hESC-qualified matrix (BD Biosciences). The same treatment was applied to the other testis. Cell-injected testes were returned to their original location, and the wound was sutured. The transplanted animals and tumour growth were observed routinely about once a week. They were sacrificed after the development of tumours larger than 2\u00a0cm in diameter or following an observation period of about 3 months. Tumours were fixed with 4% paraformaldehyde phosphate buffer Solution (FUJIFILM Wako Pure Chemical Corporation). Paraffin embedding, sectioning, and H&E staining of tumours were performed by UNITECH Co., Ltd (Chiba, Japan). Images were obtained using BZ-X710 microscope (Keyence, Osaka, Japan). This experiment was approved by the National Institute of Advanced Industrial Science and Technology (AIST) (accreditation number A2018-290).\n\n2.6. Statistical analysis {#sec2.6}\n-------------------------\n\nA one-way ANOVA followed by Tukey\\'s honest significant difference (HSD) test was used for analysis of ddPCR and teratoma size. Statistical significance was inferred where p\u00a0\\<\u00a00.05. The data were analysed with KaleidaGraph v.4.5.2 (Synergy Software, Reading, PA, USA).\n\n3. Results and discussion {#sec3}\n=========================\n\n3.1. rBC2LCN-bound magnetic beads eliminated hESCs from a heterogeneous cell mixture {#sec3.1}\n------------------------------------------------------------------------------------\n\nTo facilitate the application of hPSC technology, it is important to detect and eliminate residual tumourigenic cells with high sensitivity. As a method to eliminate tumourigenic hPSCs from differentiated cell populations, we applied a magnetic bead-based cell separation system using rBC2LCN. We used hESCs (H9 hNanog-pGZ) \\[[@bib1]\\] and normal human adult dermal fibroblasts (HDF) pre-labelled with CellTrace Violet. Biotin-labelled rBC2LCN was incubated with Dynabeads M\u2212280 Streptavidin (Thermo Fisher Scientific). In this process, rBC2LCN was bound to magnetic beads. We first tested the capability of rBC2LCN-magnetic beads to separate undifferentiated hESCs from a mixed cell population containing hESCs (H9 hNanog-pGZ) expressing *GFP* and HDF pre-labelled with CellTrace Violet at a ratio of 1:1 (each 1\u00a0\u00d7\u00a010^6^\u00a0cells). The mixed cell populations were incubated with rBC2LCN-magnetic beads. The suspensions were treated with a magnet, and the supernatant containing unbound cells was collected for flow cytometry ([Fig.\u00a01](#fig1){ref-type=\"fig\"}). The flow of the experiment and details of each sample are shown in [Fig.\u00a01](#fig1){ref-type=\"fig\"}a and [b](#fig1){ref-type=\"fig\"}, respectively. We tested three concentrations of rBC2LCN (0.1, 1, and 10\u00a0\u03bcg/ml) ([Fig.\u00a01](#fig1){ref-type=\"fig\"}b--h). Residual hESC levels evaluated by the percentage of GFP-positive and CellTrace Violet-negative cells after magnetic sorting were reduced up to 0.092% ([Fig.\u00a01](#fig1){ref-type=\"fig\"}h). The reduction in hESCs was correlated with increased rBC2LCN concentration ([Fig.\u00a01](#fig1){ref-type=\"fig\"}c--h). Magnetic beads alone as a negative control did not affect hESCs dislodged from the HDF/hESC mixture ([Fig.\u00a01](#fig1){ref-type=\"fig\"}e). These results suggest that rBC2LCN-bound magnetic beads efficiently eliminate live pluripotent stem cells from heterogeneous cell mixtures.Fig.\u00a01**Magnetic cell separation using rBC2LCN eliminated hESCs from an hESC/HDF mixture.** (a) Experimental design to remove hESCs using magnetic beads. HDF labelled with a CellTrace Violet, and hESC line H9 hNanog-pGZ (H9) were mixed in a ratio of 1 to 1. The cells incubated with Dynabeads M\u2212280 Streptavidin (M-280-SA) and biotinylated-rBC2LCN or M-280-SA alone were separated by a magnet and then analysed by flow cytometry. (b) Details of sample preparation (c--h) Flow cytometry of cells selected negatively by magnetic beads. (c) HDF alone showed 0% of GFP-positive and CellTrace Violet-negative cells (sample 1). (d) hESC alone showed 99.8% GFP-positive and CellTrace Violet-negative cells (sample 2). (e) 42.0% of GFP-positive and CellTrace Violet-negative cells were detected in the prepared cells using M-280-SA alone (sample 3; control). Selection using (f) 0.1\u00a0\u03bcg/ml (sample 4), (g) 1\u00a0\u03bcg/ml (sample 5), or (h) 10\u00a0\u03bcg/ml (sample 6) of biotinylated-rBC2LCN reduced the ratio of GFP-positive and CellTrace Violet-negative cells in a rBC2LCN concentration (9.72, 1.04, and 0.092%, respectively)-dependent manner.Fig.\u00a01\n\n3.2. rBC2LCN-bound magnetic beads eliminated hiPSCs from a heterogeneous cell mixture {#sec3.2}\n-------------------------------------------------------------------------------------\n\nIn order to test more realistic conditions for cell sorting, we next used Mitomycin C-treated HDF (MMC-HDF) and unlabelled hiPSCs (201B7) \\[[@bib2]\\]. The hiPSC/MMC-HDF mixtures were incubated with rBC2LCN-magnetic beads. The subsequent analysis is the same as in [Fig.\u00a01](#fig1){ref-type=\"fig\"}, and the collected supernatant was used not only for flow cytometry ([Fig.\u00a02](#fig2){ref-type=\"fig\"}) but also for gene expression analysis by ddPCR ([Fig.\u00a03](#fig3){ref-type=\"fig\"}) and teratoma formation assay ([Fig.\u00a04](#fig4){ref-type=\"fig\"}). We detected hiPSCs as tumour rejection antigens-1--60 (TRA-1--60)-positive cells \\[[@bib39]\\] ([Fig.\u00a02](#fig2){ref-type=\"fig\"}). The flow of the experiment and details of each sample are shown in [Fig.\u00a02](#fig2){ref-type=\"fig\"}a and [b](#fig2){ref-type=\"fig\"}, respectively. The percentage of TRA-1-60-positive cells after magnetic sorting by rBC2LCN (0.19--0.29%) closely matched that of an MMC-HDF sample without iPSCs (0.23%) ([Fig.\u00a02](#fig2){ref-type=\"fig\"}c--h). We examined three concentrations of rBC2LCN (1, 10, and 100\u00a0\u03bcg/ml) ([Fig.\u00a02](#fig2){ref-type=\"fig\"}b), which were 10-fold higher than those used in [Fig.\u00a01](#fig1){ref-type=\"fig\"}, and no substantial difference in hiPSC removal efficiency was seen among them. The biotin-labelled BSA-magnetic beads as a negative control had no effect on eliminating hiPSCs from an MMC-HDF/hiPSC mixture ([Fig.\u00a02](#fig2){ref-type=\"fig\"}e). This result indicated that magnetic sorting using rBC2LCN efficiently functioned to eliminate hiPSCs from cell mixtures, and the separation efficiency was comparable to the limit of detection of flow cytometry with TRA-1-60. In subsequent experiments, we fixed the concentration of rBC2LCN to 10\u00a0\u03bcg/ml.Fig.\u00a02**Magnetic cell separation using rBC2LCN eliminated hiPSCs from an hiPSC/MMC-HDF mixture.** (a) Experimental design to remove hiPSCs using magnetic beads. HDF and hiPSCs were mixed in a ratio of 1 to 1. The cells incubated with Dynabeads M\u2212280 Streptavidin (M-280-SA) and either biotinylated-rBC2LCN or biotinylated-BSA, were separated by a magnet, and then analysed by flow cytometry. (b) Details of sample preparation (c--h) Flow cytometry of cells selected negatively by magnetic beads. (c) HDF alone showed 0.23% of TRA-1-60-positive cells (sample 1). (d) hiPSC alone showed 95.2% TRA-1-60-positive cells (sample 2). (e) TRA-1-60-positive cells (51.8%) were detected in the prepared cells using 100\u00a0\u03bcg/ml of biotinylated-BSA (sample 5; control). Selection using (f) 1\u00a0\u03bcg/ml (sample 4), (g) 10\u00a0\u03bcg/ml (sample 5) or (h) 100\u00a0\u03bcg/ml (sample 6) of biotinylated-rBC2LCN reduced the ratio of TRA-1-60-positive cells to levels approximately equal to that of HDF alone (0.19--0.29% vs 0.23%, respectively).Fig.\u00a02Fig.\u00a03**Detection of hiPSCs by droplet digital RT-PCR.** Droplet digital PCR analysis to estimate residual hiPSC frequency after magnetic cell separation by rBC2LCN was performed by evaluating the expression of pluripotent stem cell marker genes, *NANOG* (a, c) and *LIN28* (b, d). After mixing 1\u00a0\u00d7\u00a010^6^ of iPSCs and HDF at a ratio of 1: 1 respectively, negative-sorted cells by rBC2LCN-magnetic beads were analysed (a--d) Absolute counts were normalised to *GAPDH* (per 1000 copies of *GAPDH*). Results are presented as mean\u00a0\u00b1\u00a0standard deviation of independent triplicate experiments (n\u00a0=\u00a03). Data were analysed by one-way ANOVA followed by Tukey\\'s HSD test (a, b) The relative copy numbers of *NANOG* (a) and *LIN28* (b) mRNAs in 50\u00a0ng of total RNA derived from magnetically sorted or unsorted cell mixtures. Significant differences are represented by different letters ([Fig.\u00a03](#fig3){ref-type=\"fig\"}a: p\u00a0\\<\u00a00.001; [Fig.\u00a03](#fig3){ref-type=\"fig\"}b: p\u00a0\\<\u00a00.005) (c, d) Relative copy number comparison between *NANOG* (c) and *LIN28* (d) mRNAs in 50\u00a0ng of total RNA derived from magnetically sorted cell mixtures, and comparison samples in which HDF-derived total RNA was spiked with hiPSC-derived total RNA at a ratio of 0, 0.025%, or 0.25% by weight. There are significant differences between different letters (p\u00a0\\<\u00a00.001). B7: human iPS cell line 201B7. BC2-mag: rBC2LCN-magnetic beads. BSA-mag: BSA-magnetic beads.Fig.\u00a03Fig.\u00a04**Teratoma formation assay**. (a) Schematic illustration summarising the experimental design. Magnetically sorted cells were resuspended in Matrigel and transplanted by injection into a testis (20\u00a0\u03bcL volume with \u22641.0\u00a0\u00d7\u00a010^6^\u00a0cells). (b) Box plot analysis of the size of testes in mice transplanted with Matrigel alone (negative control; n\u00a0=\u00a08), unsorted (positive control; n\u00a0=\u00a016), or sorted (n\u00a0=\u00a016) cells. The number of days from transplantation to sampling of testes varied depending on teratoma growth. The sizes of testes transplanted with unsorted cell mixture were significantly larger than those transplanted with sorted cells or Matrigel alone. Data were analysed by one-way ANOVA followed by Tukey HSD test. \u2217\u2217\u2217: p\u00a0\\<\u00a00.0001. (c) The appearance of the testes taken from NOD/ShiJic-*scid*Jcl mice transplanted with unsorted (upper left) or sorted (upper right) cells. Representative H&E-stained sections of testes with teratomas transplanted with unsorted cell mixtures of iPSCs and MMC-HDF, showing the generation of all three germ layers. Teratoma sections contained immature neuroepithelium (ectoderm), cartilage (mesoderm), and glandular epithelium (endoderm). Scale bar: 50\u00a0\u03bcm.Fig.\u00a04\n\n3.3. Detection of hiPSCs by droplet digital RT-PCR after magnetic cell separation {#sec3.3}\n---------------------------------------------------------------------------------\n\nWe also performed ddPCR analysis, allowing evaluation of the frequency of hiPSCs remaining in the cell suspension after magnetic cell separation using rBC2LCN. For discrimination of dead cells, magnetically sorted or unsorted cells were stained with propidium iodide (PI). The population of PI-negative live cells was collected and analysed ([Fig.\u00a03](#fig3){ref-type=\"fig\"}a--d and see [Supplementary Tables\u00a0S1--3](#appsec1){ref-type=\"sec\"} online). Expression levels of *NANOG* ([Fig.\u00a03](#fig3){ref-type=\"fig\"}a) and *LIN28* ([Fig.\u00a03](#fig3){ref-type=\"fig\"}b) relative to *GAPDH* were dramatically and significantly reduced in the sample sorted by rBC2LCN-magnetic beads, as compared to the sample sorted by BSA-magnetic beads, which was equivalent to the unsorted control sample (*NANOG*: p\u00a0\\<\u00a00.0001, *LIN28*: p\u00a0=\u00a00.0037). Furthermore, the levels of *NANOG* ([Fig.\u00a03](#fig3){ref-type=\"fig\"}c) and *LIN28* ([Fig.\u00a03](#fig3){ref-type=\"fig\"}d) mRNAs in sorted cell mixtures were evaluated by comparing to the samples in which total RNA extracted from HDF (HDF RNA) was spiked with total RNA extracted from hiPSCs (hiPSCs RNA) at a ratio of 0, 0.025, or 0.25%. No statistically significant differences in expression levels of *NANOG* or *LIN 28* were found among HDF RNA, HDF RNA mixed with 0.025% of hiPSCs RNA, or total RNA extracted from the cells sorted by rBC2LCN-magnetic beads. That is, the rBC2LCN magnetic beads removed iPSCs from a 1:1 mixture of iPSCs and HDF to a degree that was not significantly different from HDF alone in *NANOG* and *LIN28* expressions. These results suggested that magnetic cell separation using rBC2LCN is quite efficient for eliminating hPSCs from mixed cell populations.\n\n3.4. Teratomas were not produced by transplantation of negative-sorted cells by rBC2LCN-magnetic beads {#sec3.4}\n------------------------------------------------------------------------------------------------------\n\nThe standard method to define pluripotent stem cells capable of generating tumoural structures containing tissues representing the three germ layers is to perform a teratoma formation assay \\[[@bib40], [@bib41], [@bib42], [@bib43], [@bib44]\\]. To ensure effective removal of teratoma-forming cells, *in\u00a0vivo*, we performed a teratoma formation assay in immunodeficient mice. We transplanted either cells magnetically sorted using rBC2LCN-magnetic beads or unsorted hiPSC/MMC-HDF mixtures into the testes of NOD/ShiJic-*scid*Jcl mice and left them to engraft for 3 months ([Fig.\u00a04](#fig4){ref-type=\"fig\"}a). If any residual hiPSCs remained after sorting, these cells would form teratomas. In the animals transplanted with sorted cells or Matrigel alone, the rate of teratoma formation was 0% (n\u00a0=\u00a08 or four mice, n\u00a0=\u00a016 or eight testes for the injection of unsorted cells or Matrigel alone, respectively; [Fig.\u00a04](#fig4){ref-type=\"fig\"}b and c, and see [Supplementary Table\u00a0S4](#appsec1){ref-type=\"sec\"} online). In contrast, most testes transplanted with unsorted cell mixtures developed teratomas; 14 out of 16 transplanted testes formed large teratomas (\u22651.7\u00a0cm in maximal diameter) (n\u00a0=\u00a08 mice; n\u00a0=\u00a016 testes) ([Fig.\u00a04](#fig4){ref-type=\"fig\"}b and c, and see [Supplementary Table\u00a0S4](#appsec1){ref-type=\"sec\"} online). Statistical analysis of maximal diameter of testes revealed significant size differences between unsorted and sorted cell transplantation (p\u00a0\\<\u00a00.0001, [Fig.\u00a04](#fig4){ref-type=\"fig\"}b and see [Supplementary Table\u00a0S5](#appsec1){ref-type=\"sec\"} online). Further microscopic observation by Haematoxylin and Eosin (H&E) staining showed these teratomas contained tissues derived from the three germ layers: ectoderm (immature neuroepithelium), mesoderm (cartilage), and endoderm (glandular epithelium) ([Fig.\u00a04](#fig4){ref-type=\"fig\"}c). In agreement with the results from flow cytometry ([Fig.\u00a02](#fig2){ref-type=\"fig\"}) and ddPCR ([Fig.\u00a03](#fig3){ref-type=\"fig\"}), magnetic-cell separation using rBC2LCN effectively removed hiPSCs and could exclude potential teratoma formation in NOD/ShiJic-*scid*Jcl mice as no teratomas were observed in any of the tested animals ([Fig.\u00a04](#fig4){ref-type=\"fig\"}b and c).\n\n3.5. Cell sorting efficiency using rBC2LCN-magnetic beads {#sec3.5}\n---------------------------------------------------------\n\nTo evaluate an elimination efficiency of rBC2LCN magnetic beads, we predicted number of residual iPSCs in cell mixture used in teratoma formation assay from the ddPCR data. By calculating based on *NANOG* and *LIN28* expression, the mean \u00b1 2SD value of residual iPSCs in the transplanted cells in the teratoma assay was about 213\u00a0\u00b1\u00a0615 and 65\u00a0\u00b1\u00a0128\u00a0cells, respectively ([Supplementary Fig.\u00a0S1](#appsec1){ref-type=\"sec\"}). These data indicated that approximately 99.8--100% of iPSCs were eliminated by rBC2LCN-magnetic beads. When using a mixture of differentiated cells and undifferentiated hPSCs in a ratio of 1:1, the separation efficiency of hPSCs by magnetic beads with SSEA4 or 57-C11 antibodies was less than 80% \\[[@bib30],[@bib45]\\]. The separation efficiency was up to 99.5% using UEA-1 lectin \\[[@bib31]\\]. Although the experimental conditions were not the same, simple comparisons should be avoided, but this technology showed superior performance compared to existing magnetic bead technology using antibodies and lectins. Our results indicate that even with this technique, trace amounts of hPSCs can remain in the cell products. When used as a quality control technology for cell therapy products, the removal efficiency of this technology is not sufficient for products with a high number of transplanted cells. Previous study reported that an antibody against SSEA-5 glycan expressed on hPSCs is a useful tool to removal of teratoma-forming cells. However, an anti-SSEA-5 antibody alone was insufficient to completely remove teratoma potential and complete removal was achieved only after combining SSEA-5 with two additional pluripotent surface markers (SSEA-5/CD9/CD90 or SSEA-5/CD50/CD200) \\[[@bib29]\\]. Cell sorting by combination of rBC2LCN with other pluripotent surface markers may also improve separation efficiency. Proper combinations of various techniques, including rBC2LCN, to remove residual hPSCs in the manufacturing process are important to ensure the safety of cell therapy products.\n\n4. Conclusions {#sec4}\n==============\n\nWe optimized and validated a method to remove hPSCs from a mixture with human fibroblasts using rBC2LCN-conjugated magnetic beads ([Fig.\u00a01](#fig1){ref-type=\"fig\"}, [Fig.\u00a02](#fig2){ref-type=\"fig\"}, [Fig.\u00a03](#fig3){ref-type=\"fig\"}). Our results suggested that magnetic-cell separation using rBC2LCN effectively removed teratoma-forming pluripotent stem cells from the cell mixture ([Fig.\u00a04](#fig4){ref-type=\"fig\"}). We propose that rBC2LCN is a suitable lectin marker for magnetic beads-based cell separation and that this application will contribute to the development of a practical and feasible method for removal of residual undifferentiated cells. As previously reported, the current magnetic activated cell sorting (MACS) technology is insufficient to meet the purification needs of cell therapy \\[[@bib46]\\]. The relevance and validation of this technique for cellular transplantation will require more detailed experiments. In this study, we eliminated pluripotent stem cells from a fibroblast cell line acutely mixed together. For example, it will be more convincing to demonstrate that pluripotent cells can be separated from pluripotent stem cell-derived cell products or other cellular lineages that have been cultured together in the same conditions over time. It is crucial to properly combine the respective removal techniques according to the application scope. Further experiments are demanded to develop an optimised protocol for removal of residual undifferentiated cells.\n\nData availability statement {#sec5}\n===========================\n\nAll data analysed during this study are included in this published article and its Supplementary Information files.\n\nEthics statement {#sec6}\n================\n\nThis study was carried out in strict accordance with the National Institute of Advanced Industrial Science and Technology (AIST) guidelines for life science experiments (accreditation numbers hi2016--099 and A2018-290). Human embryonic stem cell experiment was approved by the Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT). Studies with mice were performed in accordance with the Guidelines for Proper Conduct of Animal Experiments stipulated by the Science Council of Japan.\n\nAuthor contributions statement {#sec7}\n==============================\n\nY.H., Y.O., and Y.I. designed the research. Y.H., Y.O., Y.I., S.M., Y.N., Y.A., K.H., and M.S. carried out the experiments, Y.H., Y.O., and S.M. analysed the data. Y.H. and Y.O. wrote the paper. H.T. and J.H. supervised the research.\n\nDeclaration of competing interest\n=================================\n\nThe authors declare no competing interests.\n\nAppendix A. Supplementary data {#appsec1}\n==============================\n\nThe following are the Supplementary data to this article:Multimedia component 1Multimedia component 1Multimedia component 2Multimedia component 2\n\nHuman ES cell line H9 hNanog-pGZ was obtained from the WiCell International Stem Cell (WISC) Bank. Human iPS cell line 201B7 (HPS0063) was provided by the RIKEN BRC through the National Bio-Resource Project of the MEXT, Japan. This research was supported by the 10.13039/100009619Japan Agency for Medical Research and Development (10.13039/100009619AMED) under Grant Number JP18be0204429. We would like to thank Editage ([www.editage.jp](http://www.editage.jp){#intref0015}) for English language editing.\n\nPeer review under responsibility of the Japanese Society for Regenerative Medicine.\n\nSupplementary data to this article can be found online at <https://doi.org/10.1016/j.reth.2020.03.017>.\n\n[^1]: Yoshikazu Haramoto and Yasuko Onuma contributed equally.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nApproximately 3% of the world's population carries the hepatitis C virus (HCV), putting more than 200 million people at risk of developing severe liver diseases^[@CR1]--[@CR3]^. HCV displays high genetic heterogeneity and is classified into eight genotypes (gt 1--8) and more than a hundred subtypes^[@CR4]--[@CR8]^. The polyprotein precursor is expressed from a 9.6\u2009kb positive-sense, single-stranded RNA genome ((+) ssRNA)) and is co- and post-translationally cleaved by cellular and viral proteases to produce at least ten viral proteins^[@CR9]--[@CR11]^. In the HCV polyprotein precursor, p7 lies between the structural and non-structural proteins. It is essential for the assembly and secretion of infectious viral particles *in vitro*^[@CR12]--[@CR16]^ and for virus propagation *in vivo*^[@CR17]^, making it an attractive therapeutic target^[@CR18]^.\n\nHCV\u00a0p7 is a small, hydrophobic protein comprising 63 amino acids^[@CR19]^. The structural properties of p7 constructs derived from 1b genotypes have been widely investigated using a range of approaches^[@CR18]^, including electron microscopy (EM)^[@CR20]--[@CR22]^, nuclear magnetic resonance spectroscopy (NMR)^[@CR23]--[@CR31]^, and molecular modeling^[@CR22],[@CR24],[@CR32]--[@CR36]^. The various NMR studies of monomeric p7 suggest a similar architecture that is in broad agreement with secondary structure prediction from its amino-acid sequence, namely two hydrophobic transmembrane (TM) regions separated by a conserved basic loop region^[@CR9],[@CR19]^. Analytical ultracentrifugation measurements suggest that p7 self-assembly results predominantly in hexameric and heptameric viroporins, shown in electrophysiology experiments to be cation selective^[@CR23]^. Computational modeling studies of p7 predict that the monomeric protein adopts hairpin-like structures, which then associate side-by-side with the N-terminal TM helix forming the channel pore^[@CR22],[@CR24],[@CR32],[@CR34]--[@CR36]^.\n\nA strikingly different subunit conformation and packing arrangement was reported for a p7 construct based on a 5a genotype^[@CR29]^. Several features of the hexameric structure determined by NMR were unexpected based on previous structural and functional studies of p7, but recent experiments indicate that the protein is not oligomeric in the detergent *n*-dodecylphosphocholine (DPC) used to derive the structure^[@CR37]^. The secondary structure of the genotype 5a construct, which should be unaffected by the introduction of tertiary restraints in the structure determination, also indicated key differences compared to monomeric structures published hitherto (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"} and references therein).Figure 1Comparison of p7 secondary structures. Sample conditions (top) and secondary structure determinations (bottom) reported for p7 monomers and the hexamer of OuYang *et al*. The sample conditions used for p7(5a/EUH1480) in this work are the same as previously reported^[@CR29]^. For comparison, the PSIPRED secondary structure predictions based on amino acid sequences for p7(1b/J4) and both wildtype and mutated (\\*) p7(5a/EUH1480) are also shown^[@CR82]^. The PSIPRED prediction for J4 with the C27S substitution is identical to wildtype J4 (not shown). For the subunit structure reported by OuYang *et al*., the helices are color-coded according to the original paper^[@CR29]^. The basic loop region \\~33--37, which differs notably between monomeric genotype 1b p7 and oligomeric 5a p7, is indicated by green shading^[@CR23],[@CR24],[@CR26],[@CR28],[@CR29]^. The secondary structures were estimated directly from published data (secondary chemical shifts, NOEs, and/or dipolar waves). In the case of p7(5a/EUH1480)^[@CR29]^, the helical regions were determined from the structure deposited in the PDB (2M6X). Plotted helical regions by residue number: Montserret *et al*. 2010: 3--14, 20--34, 38--46, 48--55 and 59--61; Cook *et al*. 2013: 3--13, 18--35, 42--47 and 50--56; Foster *et al*. 2014: 1--21, 23--31, 38--55 and 59--63; LA Dawson: 7--13, 15--30, 38--46 and 48--57; Ouyang *et al*. 2013: 5--16, 20--41 and 48--58; Oestringer *et al*. 2018 J4: 3--15, 18--34, 40--45, 47--56 and 59--60; Chemical shift-based re-analysis of EUH1480 using published^[@CR29]^ chemical shifts: 3--14, 18--34, 42--43 and 48--56; PSIPRED J4: 4--17, 19--33 and 39--56; PSIPRED EUH1480: 3--32, 39--45 and 47--56; PSIPRED EUH1480 mt5: 3--9 (beta strand), 10--32, 39--45 and 47--56.\n\nHCV has one of the highest mutation rates and genetic variability amongst RNA viruses^[@CR5],[@CR38]^, and the unique genotype 5a p7 construct used to determine an oligomeric structure is very different from the genotype 1b constructs used to derive all other structures (Supplementary Fig.\u00a0[1](#MOESM1){ref-type=\"media\"}). Whereas the function of several genotype 1 and 2 constructs have been studied^[@CR20],[@CR23],[@CR39]--[@CR45]^, attempts at recording electrophysiological measurements of genotype 5a p7 were not successful, making it unclear how this p7 relates to the more commonly studied homologs^[@CR29]^.\n\nWe sought to understand the extent to which sequence divergence has resulted in differences in structural preferences for p7. The structure, dynamics, and solution properties of p7 from genotype 1b (isolate J4) were studied alongside that of genotype 5a (isolate EUH1480) in phosphocholine-based detergents, using a combination of solution NMR spectroscopy and molecular dynamics (MD). The largest differences were detected in residues 40--45, which appear to adopt an unstable helical structure that may be sensitive to solution conditions. The two constructs were otherwise very similar in secondary structure and dynamics, indicating that key features of p7 structure and dynamics are conserved between distantly related HCV.\n\nResults {#Sec2}\n=======\n\nSample preparation and spectral characterization {#Sec3}\n------------------------------------------------\n\nHCV p7 from genotype 1b isolate J4 containing a C27S mutation to prevent disulfide bond formation during preparation^[@CR25]--[@CR27]^ (referred to as p7(1b/J4)) was expressed and purified from *E*. *coli* and solubilized directly into the detergent 6-cyclohexyl-1-hexylphosphocholine (Cyclofos-6) enabling solution NMR analyses of the protein at pH 6.5 and 37\u2009\u00b0C. The spectral quality was unaffected by changes in pH between 6 and 7 or detergent concentrations between 26.8\u2009mM and 268\u2009mM, indicating no change in the structure or oligomeric state over these ranges.\n\nThe p7 construct based on genotype 5a isolate EUH1480 was expressed in *E*. *coli*, purified, and refolded into the detergent DPC as described in reference^[@CR29]^. The 5a construct contained five substitutions (T1G, C2A, A12S, C27T, and C44S) to enable a direct comparison with a previously published structure^[@CR29]^ and is referred to as p7(5a/EUH1480). The resulting ^1^H,^15^N-based spectra collected on p7(5a/EUH1480) were similar to those previously reported, confirming that the protein conformation and environment were similar to those samples used previously^[@CR29]^ and enabling analyses of additional NMR experiments based on published chemical shift values. High quality spectra could be recorded on both p7(1b/J4) and p7(5a/EUH1480) at 37\u2009\u00b0C using conventional, non-TROSY approaches on fully protonated protein (Fig.\u00a0[2A](#Fig2){ref-type=\"fig\"}), consistent with previously published SEC-MALS experiments indicating that the protein is monomeric under these conditions^[@CR37]^. The spectral properties of p7(1b/J4) were similar to those of p7(5a/EUH1480) and the backbone resonances (^1^H~N~, ^15^N, ^13^C\u03b1, ^13^C\u2032) of p7(1b/J4) were assigned using conventional triple-resonance experiments.Figure 2Spectral characterization of HCV p7 isolates. (**A**) 2D ^1^H-^15^N HSQC spectra of p7(1b/J4) in Cyclofos-6 (left), and p7(5a/EUH1480)\u00a0in DPC (right), with amino acid assignments indicated. The amino acid assignments of the p7(5a/EUH1480) spectrum were transferred from^[@CR29]^. The amino acid sequences of the constructs are shown below the spectra. Below the sequence for p7(5a/EUH1480) are indicated the native amino acids that were substituted in this construct used for experiments here and in^[@CR29]^. (**B**) Top: ^13^C\u03b1 secondary chemical shift index for p7(1b/J4) (filled bars) and p7(5a/EUH1480) (open bars), with positive values indicative of \u03b1-helical conformation. Bottom: TALOS-N^[@CR46]^ chemical shift-based secondary structure prediction for p7(1b/J4)(\u26aa) and p7(5a/EUH1480) (X). Chemical shift data for p7(5a/EUH1480) were taken from the BioMagResBank entry 19162^[@CR29]^. Shown above the plot and indicated by unfilled and filled bars are the helical residues determined for p7(5a/EUH1480) from PDB 2M6X and those predicted from chemical shifts for p7(1b/J4), respectively. (**C**) Left: Membrane CS-Rosetta-based structure of p7(1b/J4). Right: A single subunit of p7(5a/EUH1480) showing the horseshoe-like conformation in the oligomeric model 2M6X^[@CR29]^ is shown for comparison. The structures are shaded from blue (N-terminus) to red (C-terminus) and the residues 34--37 are shown as sticks. The residues at the beginning and end of each helix are indicated.\n\nSecondary structure probabilities were determined from chemical shifts of p7(1b/J4) and p7(5a/EUH1480), with chemical shift data for p7(5a/EUH1480) taken from the BioMagResBank (entry 19162)^[@CR29]^. The chemical shift-based predictions from TALOS-N^[@CR46]^ were similar for the two constructs in helix 1 (residues \\~3--15), helix 2 (\\~18--34), and the C-terminal half of helix 3 (residues \\~48--56). In addition to helical breaks centered at residues 16 and 35--37, a discontinuity in helix 3 was observed in both constructs at G46, which is three residues before a proline (P49). Differences between the two constructs were apparent in the N-terminal residues of helix 3 (residues \\~40--45) in p7(5a/EUH1480), which were significantly destabilized compared with p7(1b/J4) (Fig.\u00a0[2B](#Fig2){ref-type=\"fig\"}). In addition, a weak propensity for a C-terminal helix beginning at residue P59 was observed for p7(1b/J4). Surprisingly, the helical residues present in the subunits of the hexameric model of p7(5a/EUH1480) from OuYang *et al*. (PDB 2M6X)^[@CR29]^ are markedly different from the secondary structure derived here from their chemical shift data (Fig.\u00a0[2B](#Fig2){ref-type=\"fig\"}).\n\nA structural model of p7(1b/J4) was generated using chemical shift Rosetta (CS-Rosetta)^[@CR47]^ with a membrane force field^[@CR48]^. The structure, shown in Fig.\u00a0[2C](#Fig2){ref-type=\"fig\"} (left), is most similar to the structures determined previously by Cook *et al*.^[@CR26]^, using a similar approach, and by Montserret *et al*.^[@CR23]^. In the chemical shift-based Rosetta structure of p7(1b/J4), residues 16--33 (helix 2) correspond to the first TM domain, residues 34--37 form a loop, and residues 38--57 (helix 3) correspond to the second TM domain. By contrast, in the subunit structure taken from the hexameric model of p7(5a/EUH1480)^[@CR29]^ (shown at right in Fig.\u00a0[2C](#Fig2){ref-type=\"fig\"}) residues 34--37 are helical and residues \\~42--46 form a loop.\n\nAmide backbone ^15^N relaxation and proton exchange {#Sec4}\n---------------------------------------------------\n\nBackbone dynamics of p7(1b/J4) and p7(5a/EUH1480) were characterized by ^15^N R~1~ and R~2~ relaxation rates and {^1^H-^15^N} heteronuclear Overhauser effects (hetNOEs). Trimmed mean averages of the backbone amide ^15^N R~1~ and R~2~ values were used to calculate apparent molecular weights of the protein and detergent complex (see Experimental Methods). The derived rotational correlation times, \u03c4~c~, were 10.6\u2009ns and 10.1\u2009ns at 37\u2009\u00b0C for p7(1b/J4) and p7(5a/EUH1480), respectively. These values correspond to apparent molecular weights of \\~40\u2009kDa, and are consistent with monomeric p7 embedded in micelles of \\~30\u2009kDa and consistent with SEC-MALS studies of p7(5a/EUH1480)^[@CR37]^. Elevated R~1~ values, and depressed R~2~ and hetNOE values, are indicative of increased flexibility on the picosecond-nanosecond (ps-ns) timescale, as can be seen for the N- and C-termini (Fig.\u00a0[3A](#Fig3){ref-type=\"fig\"}). For both constructs, increased ps-ns timescale flexibility was apparent in helix 1, whereas the least flexible regions, as indicated by lower R~1~ values and higher R~2~ and hetNOE values, were in the middle of helices 2 and 3. Increased ps-ns flexibility is observed for both constructs in loop residues G39, A40, and A41 (T41 in p7(5a/EUH1480)). The relatively high R~2~ values observed for G34 of p7(1b/J4) and H31 of p7(5a/EUH1480) indicate local dynamics on the microsecond-millisecond timescale. The nonhelical residues 35--37 exhibited little or no increase in dynamics compared with the TM domain residues of helix 2 indicating they form a structured loop.Figure 3HCV p7 backbone amide dynamics and amide-water proton exchange. (**A**) ^15^N R~1~, ^15^N R~2~, and ^1^H-^15^N heteronuclear NOEs as a function of residue number for p7(1b/J4) (filled black circles) and p7(5a/EUH1480) (open red circles). The backbone amide heteronuclear NOE value for residue A63 of p7(5a/EUH1480) was \u22122.1 (data point not shown). (**B**) Backbone amide hydrogen exchange data for p7(1b/J4) (top) and p7(5a/EUH1480) (bottom) as a function of residue number. NMR CLEAN chemical exchange (CLEANEX) experiments^[@CR83]^ were recorded using a 50\u2009ms (\u25fc) mixing time. For comparison are shown the empirically-based predictions of the magnitude of amide exchange from primary structure alone (solid line with grey shading)^[@CR84]^. Experiments for (**A**,**B**) were recorded at 600\u2009MHz (^1^H) and 37\u2009\u00b0C. The relaxation data were collected using conventional HSQC-based pulse sequence experiments. Shown above the plots in (**A**,**B**) by unfilled and filled bars are the helical regions determined for p7(5a/EUH1480) from PDB 2M6X and those predicted from chemical shifts for p7(1b/J4), respectively. The light shading for residues 35--39 indicate helical residues in the 2M6X structural model that were found in this study to be nonhelical. The crosshatching for residues 40--45 indicate residues predicted to form an unstable helix.\n\nWhile the per-residue profiles for backbone dynamics of p7(1b/J4) and p7(5a/EUH1480) were broadly similar, differences were seen in residues 39--47 (Fig.\u00a0[3A](#Fig3){ref-type=\"fig\"}). These residues are typically assigned to the start of the second TM domain (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}) but exhibited low helical propensities in both constructs (Fig.\u00a0[2B](#Fig2){ref-type=\"fig\"}), and increased ^15^N R~1~ and decreased ^15^N R~2~ in p7(5a/EUH1480) indicate that these residues in p7(5a/EUH1480) are significantly more dynamic than those in p7(1b/J4).\n\nThe rates of backbone amide proton exchange with water provide additional insights into secondary structure and dynamics since the proton exchange rates reflect hydrogen bond participation in addition to sidechain-dependent intrinsic exchange rates and solvent accessibility. Similar patterns of exchange were observed for the two constructs (Fig.\u00a0[3B](#Fig3){ref-type=\"fig\"}). In particular, H17, which is in a break between helices 1 and 2 exhibited elevated exchange rates, as well as residues 39--41. As expected, amides in the N- and C-termini also showed increased proton exchange rates for both constructs, but the magnitudes in p7(5a/EUH1480) were much higher than that of p7(1b/J4) and approached values expected for the same sequence fully unstructured in water^[@CR49],[@CR50]^ (Fig.\u00a0[3B](#Fig3){ref-type=\"fig\"}). Several residues in helix 1 of both constructs also exhibit amide proton exchange consistent with this helix being solvent exposed or only transiently bound to the micelle surface.\n\nMD-simulation of the p7(1b/J4) structural model in POPC bilayers {#Sec5}\n----------------------------------------------------------------\n\nPrevious atomistic MD simulations have indicated that monomeric forms of the p7(5a/EUH1480) sequence and the p7 sequence from genotype 1b isolate J (p7(1b/J)) could form stable hairpins in 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) bilayers over a 200\u2009ns timescale^[@CR51]^. Identical simulations were carried out for the p7(1b/J4) Rosetta-derived structure to examine whether the hairpin conformation of p7(1b/J4) is also stable in a POPC bilayer. The results are compared with those from p7(1b/J) and p7(5a/EUH1480)^[@CR51]^ in Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}. Visual inspection of snapshots from the simulation did not reveal any large-scale alteration of the structures over the timescale of the simulation. To quantitatively analyze the conformational evolution of the monomeric protein, the structures were separated into four \u03b1-helical segments conserved throughout the simulations as observed previously^[@CR51]^. Helix A (residues 4--14) corresponds approximately to helix 1 identified by secondary chemical shifts, helix B (20--31) corresponds to \\~helix 2, and helix C (39--47) and D (48--54) correspond to \\~helix 3 broken by the kink around G46 (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). Over the 200\u2009ns simulation the maximum variation in the angles between helices B and C, and between C and D were \\~30\u00b0 and \\~40\u00b0, respectively, indicating that the hairpin conformations are stable over the duration of the simulations. By contrast, the variations in the angle between helices A and B were larger, and up to 80\u00b0 for p7(5a/EUH1480). Variation in the angle between helices A and B arises from movements of helix A, which was not embedded in the hydrophobic portion of the membrane in the models of p7(1b/J) and p7(5a/EUH1480). Helix A in the Rosetta-derived structure of p7(1b/J4) adopts a membrane surface-bound conformation that is more stable over the duration of the simulation than the detached conformation of p7(1b/J) and p7(5a/EUH1480). The increased flexibility and solvent accessibility observed by NMR for this helix (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}) is consistent with it being outside of the micelle or loosely associated with the micelle surface.Figure 4MD simulations of p7 structures. Structural evolution of p7 hairpin monomer structures in POPC bilayers measured over 200\u2009ns MD trajectories by means of the angles formed by contiguous \u03b1-helical segments for (**A**) the p7(1b/J4) Rosetta-derived hairpin structure, and (**B**) the hairpin structure of Montserret *et al*. (architecture used for both p7(1b/J) and p7(5a/EUH1480) sequences). (**C**--**E**) Show the time traces of the angles between the helices for p7(1b/J4) (this work; red), p7(1b/J) from Montserret *et al*. (orange), and the sequence of p7(5a/EUH1480) made to adopt the hairpin conformation of Montserret *et al*. (cyan)^[@CR51]^. Horizontal dashed lines correspond to the starting angles of structures based on that of Montserret *et al*.^[@CR23]^ (dashed black) or of p7(1b/J4) from this study (dashed red).\n\nDiscussion {#Sec6}\n==========\n\nHCV p7 is a viroporin essential for infectious virus production and therefore a potential drug target^[@CR52],[@CR53]^, with the search for a truly 'pan-genotypic' HCV treatment ongoing. Current standard treatment of care using the nucleotide analog sofosbuvir targeting the HCV protein NS5B is genotype-dependent in its ability to overcome innate viral resistances and generally requires combinations with other direct acting antivirals to prevent escape mutations^[@CR54],[@CR55]^, hence several compounds targeting p7 have been investigated and could serve as anti-HCV drugs. Some of these compounds are postulated to inhibit the assembled channel whereas others are hypothesized to disassemble the oligomeric channel^[@CR56]^ (reviewed in references^[@CR18],[@CR53]^). Thus, detailed knowledge of the p7 monomer structure and how it assembles to form functional channels remains of great interest. However, p7 poses a challenge for structural studies because most of the protein is buried in the membrane. In addition, it forms oligomers with variable stoichiometry and may only weakly oligomerize in detergent^[@CR20]--[@CR23],[@CR34]^. Thus despite its small size only one atomic-resolution, experimental model for an assembled p7 channel has been published^[@CR29]^. However, that structure included several unexpected features which could not be reconciled with many published structural, biochemical, and functional data and it was subsequently shown that the detergent used to solubilize the protein for those structural studies does not support p7 oligomerization^[@CR37]^. Further questions remain about the secondary structure in that p7 model, which is not expected to be strongly affected by introducing tertiary contacts in the structure calculations. While phosphocholine detergents are strongly denaturing and can introduce artefacts^[@CR57]^, the persistence of structural features across different solubilizing reagents, solution conditions, and sequences can provide an indication of whether those features are likely to be present in a biological membrane.\n\nThe structures of two closely related isolates from 1b (p7 from isolate J and J4; 94% identical) have been studied previously (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). For these 1b-derived p7 constructs studied under different conditions, the largest differences were seen for an N-terminally FLAG-tagged construct in methanol that has a long N-terminal helix extending from the FLAG-tag to residue \\~21 and lacks the break at position \\~15 seen in other studies. By contrast, the 1b- and 5a- derived p7 sequences are two of the most distantly related among known p7 sequences (Supplementary Fig.\u00a0[1](#MOESM1){ref-type=\"media\"}). The sequences of the p7(1b/J4) and p7(5a/EUH1480) constructs studied here differ in 46% (29 of 63) of the positions, providing a good test of whether structural features of p7 are conserved.\n\nThe structure in the region around the strictly conserved dibasic motif (residues K/R33 and K/R35) is of particular interest from a functional point of view since charge neutralizing substitutions result in nonviable virus^[@CR17]^. The dibasic motif is known to be important in ion channel activity^[@CR58]^, and double glutamine or double alanine substitutions in the JFH-1 isolate of genotype 2a result in 100- and 1,000-fold decreases in total infectivity, respectively^[@CR16]^. The oligomer model of the p7(5a/EUH1480) construct indicated that helix 2 extends to residue T41 with a kink of \\~45\u00b0 at G34, whereas the results here indicate that helix 2 ends at \\~G34 in both p7(1b/J4) and p7(5a/EUH1480). Comparison with previously published work on monomeric p7 indicates a consensus that the cytosolic loop is in residues 35--37 (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}), such that R35 is within the loop and K33 is close enough to the end of the TM domain to \"snorkel\" into the lipid headgroup region^[@CR59]^.\n\nThe largest structural and dynamic differences between the two distantly related constructs studied here are in residues \\~40--45, which are immediately C-terminal to the cytosolic loop. The differences seem to arise from increased sequence variability and structural instability in this region. The second TM domain of p7 is typically assigned to residues \\~40--56, with a discontinuity at 46 or 47. Residues 48--56 exhibit strong helicity in most studies, whereas residues 40--45 tend to be more variable. Structural instability was observed previously in this region for the p7(1b/J4) construct in different conditions^[@CR26]^. In the work reported here, residues 40--45 exhibit moderate to high helicity in p7(1b/J4) but are mostly nonhelical in p7(5a/EUH1480). Sequence conservation in residues 40--45 is generally low (Supplementary Fig.\u00a0[2](#MOESM1){ref-type=\"media\"}), and only two of the positions in this region are conserved between 1b/J4 and 5a/EUH1480. Structural variation is also seen here between the 1b-derived p7 constructs having no or very small sequence differences but studied in different membrane mimetics (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). The composition of residues 40--45 tends to be relatively hydrophilic for a TM domain (Supplementary Fig.\u00a0[3](#MOESM1){ref-type=\"media\"}), which can make it particularly sensitive to the membrane mimetic, and the helix here may be more stable in lipid membranes^[@CR28]^.\n\nA helix discontinuity at, or near, G46 was seen in both constructs and also observed in other studies^[@CR23],[@CR24],[@CR26]^. The discontinuity is likely due in part to the proline at position 49. Prolines are often C-terminal to helical kinks and are present at higher frequency within TM helices than in helices of water-soluble domains^[@CR60]^. Kinks can be exaggerated by the detergent environment^[@CR61]^, but MD simulations in lipid membranes are also consistent with at least a small (\\~15--30\u00b0) deflection of the helix (Fig.\u00a0[4E](#Fig4){ref-type=\"fig\"}). Although G46 and P49 are not strictly conserved, one or the other is always present and substitutions in these positions are for more hydrophilic amino acids (Supplementary Fig.\u00a0[2](#MOESM1){ref-type=\"media\"}).\n\nThe published subunit structure of p7(5a/EUH1480) pointed to a helix extending from 20--41 with a kink at G34^[@CR29]^, whereas secondary chemical shifts indicate that residues 35--41 are not helical, and 39--41 are particularly flexible (reference^[@CR62]^ and Fig.\u00a0[2B](#Fig2){ref-type=\"fig\"}). The helical boundaries determined for the subunit structure were based on observation of characteristic NOEs, which is expected to yield similar results compared with chemical shift-based approaches. However, spectral overlap in the ^1^H\u03b1 region can result in information gaps with the NOE approach and chemical shift information tends to be more complete since ^13^C shift information (^13^C\u2032, ^13^C\u03b1, ^13^C\u03b2) is partially redundant. Consistent with a non-helical state, these amides exhibit proton exchange rates comparable to what would be expected for the same sequence unstructured in water (Fig.\u00a0[3B](#Fig3){ref-type=\"fig\"}).\n\nThe pattern of amide proton exchange of p7(5a/EUH1480) in DPC was similar to that of p7(1b/J4) in Cyclofos-6, and the results are consistent with the exchange rates previously reported for p7(5a/EUH1480) in similar conditions^[@CR30]^. Some regions, including the N- and C-termini, helix 1, and H17, exhibited higher exchange rates in p7(5a/EUH1480), which may be due to differences in hydrophobicity between the constructs: The regions including residues 13--20 and 53--58 are more hydrophobic in p7(1b/J4) than p7(5a/EUH1480) (Supplementary Fig.\u00a0[3](#MOESM1){ref-type=\"media\"}), which may result in remodeling of the detergent micelle and more water exposure. Although all substitutions introduced to create p7(5a/EUH1480) decrease its hydrophobicity, the wildtype sequence is also less hydrophobic than p7(1b/J4) in the regions with large differences in amide proton exchange rates. The increased exchange rates and high flexibility in helix 1 (residues \\~5--15) that have been attributed either to exposure to a hydrophilic channel pore in the oligomer^[@CR30]^ or to membrane thinning^[@CR62]^ may be explained more simply by the hairpin models showing that helix 2 is the TM helix, and helix 1 extends out of the membrane^[@CR23],[@CR24]^ or possibly lies on the membrane surface^[@CR26]^.\n\nFinally, we provide further evidence that the hairpin conformation can be stably adopted by different p7 sequences. Most p7 modelling studies predict the monomer to adopt a closely packed helical hairpin conformation, which then assembles side-by-side into oligomers. Most studies have used p7 constructs from genotypes 1 or 2, but MD simulations have shown that p7(5a/EUH1480) can also adopt a hairpin conformation in membranes^[@CR63]^, and that the stability is not strongly sequence dependent^[@CR51]^. Here, we expand those results to show that the Rosetta-derived hairpin conformation of p7(1b/J4) is also stable in a lipid bilayer.\n\nIn conclusion, we find that the genetically distant genotype 1b and 5a p7 constructs behave similarly, as assessed by NMR spectroscopy in detergent and computer modelling in lipid membranes. The results for 5a p7 are largely consistent with previous reports on monomeric p7's from other genotypes and in a range of solubilizing conditions, demonstrating that key structural and dynamic features are conserved among distantly related p7 isolates. Together, these findings provide a foundation for future studies of how p7 monomers assemble to form an oligomeric viroporin.\n\nExperimental Section {#Sec7}\n====================\n\nProtein expression and purification {#Sec8}\n-----------------------------------\n\nHCV p7 (strain J4, C27S, genotype 1b, \"p7(1b/J4)\", and strain EUH1480, T1G/C2A/A12S/C27T/C44S, genotype 5a, \"p7(5a/EUH1480)\") were expressed into inclusion bodies as a fusion to His~9~--trp\u0394LE using the vector pMM-LR6^[@CR64]^. The p7 genes (synthetic DNA strings) were ordered from GeneArt (Life Technologies) inserted into vector pMM-LR6 and transformed into XL10 Gold Competent Cells (NEB) for plasmid propagation and storage. The plasmid with the p7 insert was purified from high optical density (600\u2009nm) cultures (QIAprep Spin Miniprep Kit, QIAGEN) and transformed into BL21(DE3) cells (NEB) for protein expression. Genes were confirmed by DNA sequence analysis (Source Bioscience, Oxford). A large culture in LB was grown overnight at 37\u2009\u00b0C, followed by condensation into a smaller culture of minimal media the next morning (adapted from^[@CR65]^). The isotopically labeled p7 peptides were purified via immobilized metal ion affinity chromatography (IMAC) and released from the fusion protein by cyanogen bromide cleavage in 70% formic acid (1\u2009hour, 0.2\u2009g/ml\u00a0cyanogen bromide). The digest reaction was stopped with 1\u2009N NaOH and dialyzed to water and lyophilized. The dried, cleaved HCV p7(1b/J4) was taken up in 10% SDS, sonicated for 15\u2009minutes, mixed with the same volume of running buffer and filtered through a 0.22\u2009\u03bcm membrane (Millex-GS, Millipore) and loaded on a Sephacryl column (HiPrep 26/60 Sephacryl S-200) and separated using an \u00c4kta Pure FPLC system (GE Healthcare)^[@CR66]^. An isocratic gradient eluted the protein (20\u2009mM sodium phosphate, 1\u2009mM EDTA, 1\u2009mM NaN~3~ and 4\u2009mM SDS, pH 8.2) and the eluate was dialyzed against water and lyophilized^[@CR66],[@CR67]^. The lyophilized, cleaved p7(5a/EUH1480) was purified as previously described^[@CR29],[@CR30]^ using a C18 preparative column (Proto\u00a0300 5\u2009\u03bcm, Higgins Analytical) on a Gilson PrepLC system (321 Pump, UV/VIS-155 detector, UniPoint software). The pure eluate was freeze-dried.\n\nThe lyophilized p7(1b/J4) protein was solubilized directly by addition of Cyclofos-6 (Anatrace, Anagrade) in water at 10\u00d7\u2009(26.8\u2009mM) or 100\u00d7\u2009(268\u2009mM) the CMC. The efficiency of solubilization increased with detergent concentration from 26.8 to 268\u2009mM. Insolubilized material was discarded after bench top centrifugation (5\u2009minutes, 16,000 x *g*). Critically, 40\u2009mM sodium phosphate was added after solubilization from a 200\u2009mM stock of pH between 6.0 and 7.0 (as required), and the pH was readjusted between 6.0 and 7.0 with NaOH. Final protein monomeric concentrations were between 200 and 400\u2009\u03bcM. All NMR samples contained 5% D~2~O and 0.1\u2009mM 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS). P7(5a/EUH1480) protein was prepared in a manner similar to that described previously^[@CR29]^. The lyophilized protein was solubilized in 200\u2009mM DPC (Anatrace, Anagrade) and 6\u2009M guanidine hydrochloride (GuHCl), in the absence of buffer. The protein (\\~300\u2009\u03bcl sample with \\~300\u2009\u03bcM polypeptide concentration) was refolded upon dialysis against two liters 25\u2009mM MES pH 6.5 in the absence of DPC with two buffer changes at two-hour intervals. Crucially, dialysis was performed using 2\u2009kDa MWCO slide-A-Lyzer 100\u2009\u00b5l capacity cups (Thermo Scientific), which provide a surface area that allows full elimination of GuHCl, but retains enough DPC to maintain the protein in solution. The sample was re-equilibrated by dialysis against 30\u2009mL of 200\u2009mM DPC, 25\u2009mM MES pH 6.5 for 24\u2009hours using 10--14\u2009kDa MWCO membranes (Pur-A-Lyzer Mini Dialysis) to allow for detergent exchange. Any insoluble material was removed by centrifugation (5\u2009minutes, 16,000 x *g*). The final protein concentration was \\~300\u2009\u00b5M.\n\nNMR spectroscopy and data analysis {#Sec9}\n----------------------------------\n\nAll NMR spectra were recorded on NMR spectrometers with Oxford Instrument magnets with ^1^H frequencies between 500 to 950\u2009MHz, equipped with home-built triple resonance probes with triple axis gradients^[@CR68]^ or with Bruker TCI CryoProbes with single Z-axis gradients (500 and 600\u2009MHz). Experiments were performed at 30\u2009\u00b0C or 37\u2009\u00b0C as indicated and pH 6.5. NMR spectra were referenced in the direct dimension against DSS at 0 ppm. NMR data were processed using NMRPipe and analyzed using NMRDraw^[@CR69]^, Analysis^[@CR70]^ or CARA^[@CR71]^. The resonances of p7(J4/1b) in Cyclofos-6 could be tentatively assigned using ^15^N-edited NOESY-HSQCs (90\u2009ms and 140\u2009ms mixing times) and a ^15^N-edited TOCSY-HSQC (55\u2009ms mixing time). Assignments were subsequently confirmed using a set of HSQC-based triple-resonance experiments (HNCA, HNCACB/HNCOCACB and HNCO/HNCACO). For characterization of the hydrodynamic properties of the NMR samples, the apparent molecular weights were calculated from the ^15^N R~1~ and R~2~ values by first estimating the rotational correlation time, \u03c4~c~, from the 20% trimmed means of the relaxation rates^[@CR72]^. Trimmed means were used to exclude residues with internal motions faster or slower than the overall tumbling time. Trimmed means for R~1~ and R~2~ for p7(1b/J4) and p7(5a/EUH1480) were 1.32\u2009s^\u22121^ and 15.92\u2009s^\u22121^, and 1.34\u2009s^\u22121^ and 14.80\u2009s^\u22121^, respectively. The molecular weight was then calculated from \u03c4~c~ using Stoke's law^[@CR73]^ assuming a hydration shell of 1.5 water molecules, a solution viscosity of 0.702 centipoise at 37\u2009\u00b0C, and a protein partial specific volume of 0.73\u2009cm^3^/g. To calculate the number of detergent molecules consistent with the scenario of a monomeric p7 embedded in a detergent micelle, the value of the partial specific volume was weighted according to the fraction of the complex that was protein (0.73\u2009cm^3^/g) and detergent (0.94\u2009cm^3^/g).\n\nSecondary structure calculations {#Sec10}\n--------------------------------\n\nTALOS-N calculations of secondary structure were based on chemical shift data for the following nuclei: ^1^H~N~, ^1^H\u03b1, ^15^N, ^13^C\u2032, and ^13^C\u03b1. Chemical shifts for p7(5a/EUH1480) were obtained from BioMagResBank^[@CR74]^ entry 19162^[@CR29]^ and corrected for deuterium isotope effects^[@CR75]^.\n\nMolecular dynamics {#Sec11}\n------------------\n\nStructures for MD simulations were immersed in a fully hydrated POPC lipid bilayer of initial dimensions equal to 75\u2009\u00d7\u200975\u2009\u00d7\u200990 \u00c5^3^. Except for the POPC aliphatic tails, which were modeled by means of a united-atom potential energy function^[@CR76]^, use was made of an all-atom representation. The different components of the assays were described by the macromolecular CHARMM36 (chemistry at Harvard macromolecular mechanics) force field^[@CR77]^. All the simulations reported herein were performed in the isobaric-isothermal ensemble using the NAMD simulation package^[@CR78]^. The temperature and the pressure were maintained at 300\u2009K and 1\u2009atm employing, respectively, softly damped Langevin dynamics and the Langevin piston algorithm^[@CR79]^. Periodic boundary conditions were enforced. The equations of motion were integrated using the r--RESPA multiple-time stepping scheme^[@CR80]^ with a time step of 2 and 4\u2009fs for short- and long-range interactions, respectively. Non-bonded van der Waals interactions were smoothly switched to zero between 10 and 12\u2009\u00c5. The PME algorithm^[@CR81]^ was utilized to account for long-range electrostatic interactions. During equilibration of the membrane, soft harmonic restraints were applied to the heavy atoms of the protein, prior to their slow release after appropriate relaxation of the surroundings. Evolution of the three-dimensional structures was monitored over a simulation time of 200\u2009ns. Each production trajectory was prefaced by equilibration of up to 50\u2009ns. For quantitative analysis of conformational changes, the p7 helical structure was decomposed into four regions, wherein TM1 and TM2 were both split into two domains, A (residues 4 to 14), B (residues 20 to 31), C (residues 39 to 47) and D (residues 48 to 54).\n\nSupplementary information\n=========================\n\n {#Sec12}\n\nSupplementary information\n\n**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary information\n=========================\n\n**Supplementary information** accompanies this paper at 10.1038/s41598-019-44413-x.\n\nThis work was supported by the Oxford Glycobiology Endowment. J.R.S. and J.K.C. were funded by the MRC (L018578). N.Z. is a Fellow of Merton College.\n\nB.P.O. performed protein expression, sample preparation, NMR data collection and analysis, and wrote the paper; J.H.B. performed sample preparation and wrote the paper; J.K.C. performed NMR data collection and analysis; L.A. performed Rosetta modeling; C.C. and F.D. performed MD simulations and wrote the paper; N.H. performed MD simulations; J.R.S. performed NMR data collection and analysis; J.R.S. and N.Z. conceived the study and wrote the paper.\n\nThe authors declare no competing interests.\n"}
+{"text": "![](medphysj68885-0086){#sp1 .84}\n\n![](medphysj68885-0087){#sp2 .85}\n"}
+{"text": "INTRODUCTION\n============\n\nPulmonary vein isolation (PVI) is the main treatment for atrial fibrillation (AF) ([@B01]). Cryoballoon ablation as an efficient and relatively safe procedure that is increasingly being used to perform PVI in patients with AF in recent years ([@B02],[@B03]). Cryoballoon ablation is not inferior to radiofrequency ablation with respect to efficacy for the treatment of AF and there is a lower re-hospitalization rate and re-ablation rate in cryoablation ablation ([@B04],[@B05]). Previous studies have shown that the second-generation cryoballoon is an effective and safe technique in achieving both favorable clinical outcomes and acute PVI ([@B06] [@B07] [@B08] [@B09]). However, recently published data indicates that approximately 90% of PVs are still isolated 3 months after second-generation cryoballoon ablation ([@B10]). Despite encouraging results, recurrence of AF is still relatively frequent after surgery, which may be related to PV reconnection ([@B10] [@B11] [@B12]). To date, limited information is available regarding the factors determining PV reconnection following cryoballoon ablation.\n\nPrevious studies have demonstrated that cryoablation temperature, cryoballoon freeze duration, cryoablation duration, achievement of \u221240\u00b0C within 60s, and balloon warming time are predictors of late PV reconnections after cryoballoon ablation for the treatment of AF ([@B13] [@B14] [@B15] [@B16]). However, the thawing rate, which is related to time and temperature, lacks further research. At the same time, basic experiments have proved that freezing and thawing rates are related to freezing injury, which can cause PVI; therefore, it has practical research significance and clinical value ([@B17],[@B18]). In addition, previous studies have proved that vagus reflex could independently predict long-term PVI ([@B19]). However, there was yet no systematic study based on thawing rate and factors ensuring long-term PVI. In this study, we aimed to evaluate whether thawing rate could be a novel predictor of acute PVI and explore the predictive value of the thawing rate for vagus reflex.\n\nMATERIALS AND METHODS\n=====================\n\nPatients\n--------\n\nWe retrospectively reviewed 151 patients (21 cases of persistent AF and 130 cases of paroxysmal AF) who received cryoballoon ablation for AF using the single big (28 mm) second-generation cryoballoon technique from January 2017 to June 2018. The inclusion criteria were as follows: 1) patients with symptomatic AF who were not responding to class I and III antiarrhythmic drugs; 2) patients with age \\<80 years; 3) patients who underwent cryoballoon ablation for AF using the single big (28 mm) second-generation cryoballoon technique. The exclusion criteria were as follows: 1) history of atrial fibrillation ablation; 2) presence of atrial thrombosis; 3) presence of valvular disease (moderate and severe valvular stenosis, severe valvular regurgitation); 4) history of prosthetic heart valve replacement; 5) pregnant women; 6) patients with severe liver and kidney dysfunction; 7) patients with malignant tumors or hematological diseases. This study was approved by the ethics committee of the Second Hospital of Hebei Medical University. Written informed consent was obtained from each patient.\n\nCryoballoon ablation procedure\n------------------------------\n\nCryoballoon ablation was carried out by the conventional method, as described previously ([@B20]). A 3-dimensional computed tomography (Aquilion ONETM, TSX-301C, Toshiba Medical Systems, Tokyo, Japan) was performed in the week prior to the procedure. Patients were placed under general anesthesia. The goal of the cryoablation was electrical PV isolation of the four major PVs (left inferior pulmonary vein \\[LIPV\\], left superior pulmonary vein \\[LSPV\\], right inferior pulmonary vein \\[RIPV\\], right superior pulmonary vein \\[RSPV\\]) as confirmed by entrance and/or exit blocks. A transseptal puncture was performed using an RF needle (Baylis Medical, Inc., Montreal, QC, Canada). After passing through the transseptal sheath, the cryoballoon catheter entered into the left atrium and it was continuously infused with heparin saline through a 12F oriented sheath. Following all PV venography during raid ventricular pacing, the second-generation cryoballoon system (Arctic Front Advance\u2122 Cardiac Cryoablation Catheter, Medtronic, Minneapolis, MN) was advanced and placed on the ostia of each PV using an inner circular mapping catheter (Achieve, Medtronic). The size of the balloon was selected as 28 mm according to the measurement of the PV ostia from the venograms and computed tomographic images. The cryoballoon was delivered and moved along the Achieve catheter until the pulmonary vein ostia was completely blocked (blocking was determined by pulmonary venography).\n\nIn general, each pulmonary vein was cryoablated at least twice for 180 seconds each time. After cryoablation, the Achieve catheter was used to test whether the PVI was completed. If two cryoablations failed to complete the PVI, cryoablation would continue until PVI was successful. In the implementation of right PVI, a secondary catheter was required to be placed at the phrenic potential of the superior vena cava. A cycle length of 999 ms phrenic nerve pacing was taken to continuously monitor the right phrenic nerve function. If the balloon nadir temperatures reached a temperature below -55\u00b0C or if the diaphragm movement were weakened or disappeared, the cryoablation would be immediately stopped. A total of 604 cryoablations were performed. Twelve of the cryoablations were excluded, including five cases of hypothermia, six cases of inability to determine isolation, and one case of sacral nerve movement disappearance. The remaining 592 cryoablations were analyzed.\n\nCryoablation curve\n------------------\n\nTo avoid the cumulative effect of ablation, we only analyzed the first ablation of each pulmonary vein. The Cryoconsole software (Medtronic CryoCathLP) recorded the cryoablation curves in detail and we only analyzed the thawing phases of the cryoablation curves. The thawing phase was from the end of freezing to the end of thawing, in which the thawing rate was generated. Cryoablation curves from 592 cryoablations were reconstructed using these data ([Figure 1A](#f01){ref-type=\"fig\"}). Our aim was to quantify the thawing rate by fitting a regression line to a specific portion of the cryoablation curve ([@B21]). The slope of the regression line ([Figure 1B](#f01){ref-type=\"fig\"}) provided an immediate measure of the rapidity of the temperature rise. Furthermore, this method had the advantage of averaging the information of several samples, thus compensating for a certain amount of imprecision.\n\nStatistical analysis\n--------------------\n\nAll statistical analyses were performed by using SPSS version 22.0 (SPSS Institute. IL.USA). Quantitative data were expressed as means\u00b1standard deviations (SD) and were compared using Student's t-test. Qualitative data were expressed as number and percentage. Diagnostic performance of the thawing rate as a predictor of PVI was assessed by constructing a receiver operating characteristic (ROC) curve. By calculating the area under curve (AUC) of different thawing rates, the thawing rate interval with the greatest predictive value for acute PVI was obtained. Vagus reflex was defined as sinus bradycardia (\\<40 bpm), asystole, atrioventricular (AV) block, or hypotension that occurred within cryoapplication ([@B22]). ROC curve was used to analyze the predictive value of the thawing rate for vagus reflex. Statistical significance was set at *p*\\<0.05.\n\nRESULTS\n=======\n\nBaseline characteristics\n------------------------\n\nA total of 151 patients (81 males, 70 females; mean age: 61.32\u00b110.22 years; 21 cases of persistent AF and 130 cases of paroxysmal AF) who received cryoballoon ablation for AF were enrolled in this retrospective study from January 2017 to June 2018. The baseline characteristics of participants are showed in [Table 1](#t01){ref-type=\"table\"}. Among the 151 patients, 80 (52.98%) patients had hypertension and 26 (17.22%) had diabetes. Otherwise, 23 (15.23%) patients had a history of smoking, 19 (12.58%) patients had a history of drinking, and 34 (22.52%) patients had coronary heart disease.\n\nThawing rate predicts acute PVI\n-------------------------------\n\nWe first evaluated whether thawing rate could be a novel predictor for acute PVI. The ROC curve analysis results showed that thawing rate was a predictor (AUC=0.596) of acute PVI. The best cut-off value of thawing rate for acute PVI was\u22642.21\u00b0C/S and its sensitivity and specificity were 31.76% and 84.85%, respectively ([Figure 2](#f02){ref-type=\"fig\"}).\n\nITR15 as the most valuable predictor of acute PVI\n-------------------------------------------------\n\nIn order to obtain the thawing rate interval with the greatest predictive value for acute PVI, we selected the interval from the start of thawing to -10\u00b0C, -5\u00b0C, 0\u00b0C, 5\u00b0C, 10\u00b0C, 15\u00b0C, 20\u00b0C ([Figure 3A](#f03){ref-type=\"fig\"}). By calculating the AUC of different thawing rate, it was found that the maximum AUC was for interval thawing rate at 15\u00b0C (ITR15) ([Figure 3B](#f03){ref-type=\"fig\"}, [Table 2](#t02){ref-type=\"table\"}). It showed that the ITR15 (thawing rate\u22642.14\u00b0C/s; AUC, 0.823; 95% CI, 0.788-0.859) was the most valuable predictor of PVI. ITR15 presented 88.62% sensitivity and 67.18% specificity, with a 77.2% of positive predictive value (PPV) and 82.5% negative predictive value (NPV).\n\nTo further verify our results, we compared the ITR15 of the left superior, left inferior, right superior, right inferior pulmonary veins and total pulmonary veins in the successful PVI group and the failed PVI group, respectively. The results showed that the ITR15 of the successful PVI group after cryoballoon ablation was significantly slower than that of the failed PVI group (*p*\\<0.05) in different PV ([Figure 4](#f04){ref-type=\"fig\"}).\n\nITR15 predict vagus reflex\n--------------------------\n\nA total of 43 vagal reflexes were observed during cryoablation. Most vagus reflexes (93.02%) occurred in thawing phase and within 1-3 minutes after the end of thawing. The highest occurrence of vagus reflex was in LSPV (60.47%) and the vagus reflexes occurred in LIPV, RSPV, and RIPV occupying 32.56%, 2.32% and 4.65%, respectively. The ROC curve analysis results showed that ITR15 was a predictor (AUC, 0.64; best cut-off point\u22641.79\u00b0C/s; sensitivity, 67.78%) of vagus reflex ([Figure 5A](#f05){ref-type=\"fig\"}). Meanwhile, the occurrence of vagus reflex group had a slower ITR15 compared to the non-occurrence group (1.77\u00b10.47\u00b0C/S *vs* 2.22\u00b11.01\u00b0C/s, *p*\\<0.01) ([Figure 5B](#f05){ref-type=\"fig\"}).\n\nDISCUSSION\n==========\n\nCryoballoon ablation is increasingly being used to perform PVI in patients with AF ([@B23]). However, there is still a lack of effective indicators to assess the effectiveness of cryoballoon ablation. Previous studies had shown that cryoablation temperature and cryoablation duration could predict acute PVI, but the time and temperature were natural and not artificial results in the process of cryoablation ([@B13],[@B15]). Therefore, although sometimes the cryoablation temperature and time could not meet the expectation of the operator, PVI had already been completed, which made the operator unable to judge the results, leading to excessive increase in the number and time of cryoablation. However, little research has been done with regard to the thawing rate, which was related to time and temperature. In this study, we found that thawing rate was a novel predictor of acute PVI, which could provide a certain reference value when factors such as cryoablation temperature and time cannot effectively predict PVI. Deubner et al. found that a 20-second time frame with a freezing rate of \\<17\u00b0C could best predict PVI ([@B21]). Meanwhile, our study showed that the ITR15 was the most valuable predictor of acute PVI.\n\nHow could freezing and thawing rates predict PVI? From a cellular perspective, experimental studies had demonstrated that freezing and thawing rates could affect cell death. The faster the freezing rate, the more cell death could occur. With rapid cooling, the cell fluid had not yet been removed from the cells under the action of osmotic pressure, which aggravated the degree of intracellular freezing, leading to excessive cell death ([@B17],[@B18],[@B24]). The slower the thawing rate, the more cell death could occur. Slow thawing may promote the entry of extracellular fluid into cells, resulting in cell and tissue damage ([@B25]). Overall, thawing and freezing rates could cause myocardial cell damage in the PV or vestibule of the PV, which was essential for PVI. From the cryoballoon ablation application perspective, rapid freezing and thawing rates were predictors of better PV occlusion; therefore, there were better circumferential contact area and better cryoablation effects. The application of complete occlusive cryoballoon ablation could shorten the contact distance between the balloon and the frozen surface, resulting in a lower temperature on the contact surface, which was essential for long-term PVI.\n\nA previous study had proved that increased vagal reflex could reduce the recurrence rate of AF ([@B19]). Our study showed that ITR15 was a predictor of vagus reflex, although it had a lower predictive value (AUC=0.64), which probably due to vagal reflex may be more closely related to the distribution of vagus nerve at the junction of the atrium and PV ([@B19]). However, the occurrence of vagus reflex group had a slower thawing rate compared to the non-occurrence group. That means, even if this ablation occurred on the vagus nerve, vagus reflex would not occur without a slow thawing rate. In summary, thawing rate was associated with the prognostic factor that may affect long-term PVI.\n\nThere were several limitations in the current study. First, this study was a retrospective study conducted at a single center, which had certain limitations in clinical application. Second, the number of patients included in this study was small. Third, the patients with recurrent AF were less likely to opt for a second surgery because of different state systems, different national conditions and other factors. We were not sure which PV was reconnected; therefore, we were not sure if the thawing rate really had a predictive value for long-term PVI. Fourth, we used a single big (28 mm) second-generation cryoballoon technique. Due to the fact that temperature characteristics may vary with balloon size, our analysis only applied to the 28-mm balloon. Fifth, the complex anatomical structure and size of the pulmonary vein were not considered, which may affect the ablation effect. Lastly, we did not compare the thawing time of the ITR15 in the successful and failed PVI groups with the reported predictors.\n\nCONCLUSION\n==========\n\nThawing rate was a novel predictor of acute PVI and the ITR15 was the most valuable predictor of acute PVI. In addition, ITR15 was a predictive factor ensuring long-term PVI (vagus reflex). Our study showed that thawing rate may serve in the early identification of useless cryoballoon ablation.\n\nAUTHOR CONTRIBUTIONS\n====================\n\nZhang CF and Xie RQ were responsible for the study conception and design, and manuscript critically review. Zhang CF and Yang Y were responsible for the data collection. Zhang CF, Wu JL and You L were responsible for the data analysis and interpretation. Zhang CF was responsible for the manuscript drafting article. Zhang CF, Ma BF and Wu JL were responsible for the statistics. Xie RQ was responsible for the administrative support. All of the authors have read and approved the final version of the manuscript.\n\nThanks to professor Ruiqin Xie, and other professors for their continual guidance and support in the course of our research. This work was supported by Major Research and Development Projects in Hebei Province \\[number 17277729D\\].\n\nNo potential conflict of interest was reported.\n\n![Examples of cryoablation curves and their fitted regression line A, Example of a complete cryoablation curve. B, Examples of a portion of the cryoablation curve (the time frame from the first value \\>-50\u00b0C, with 60s duration) and their fitted regression line.](cln-75-e1672-g001){#f01}\n\n![ROC curve analysis. ROC, receiver operating characteristic.](cln-75-e1672-g002){#f02}\n\n![ROC curve of different thawing rate intervals. A, Thawing curves were divided into seven intervals. ITR-10, interval thawing rate at -10\u00b0C; ITR-5, interval thawing rate at -5\u00b0C; ITR0, interval thawing rate at 0\u00b0C; ITR5, interval thawing rate at 5\u00b0C; ITR10, interval thawing rate at 10\u00b0C; ITR15, interval thawing rate at 15\u00b0C; ITR20, interval thawing rate at 20\u00b0C. B, ROC curve of different thawing rate intervals. ROC, receiver operating characteristic.](cln-75-e1672-g003){#f03}\n\n![ITR15 of different PVs during cryoballoon ablation PV, pulmonary vein; LSPV, left superior pulmonary vein; LIPV, left inferior pulmonary vein; RSPV, right superior pulmonary vein; RIPV, right inferior pulmonary vein; PVI, pulmonary vein isolation; ITR15, interval thawing rate at 15\u00b0C; \\**p*\\<0.05.](cln-75-e1672-g004){#f04}\n\n![Thawing rate and vagus reflex A, ROC curve analysis. B, Comparing ITR15 of the occurrence of vagus reflex group and the non-occurrence group. ITR15, interval thawing rate at 15\u00b0C; ROC, receiver operating characteristic.](cln-75-e1672-g005){#f05}\n\n###### Baseline characteristics of the included patients.\n\n  Characteristics                Number (n,%)\n  ------------------------------ --------------\n  Age (years), Mean\u00b1SD           61.32\u00b110.22\n  Sex(n)                         \n  \u00a0\u00a0\u00a0\u00a0Male                       81 (53.64)\n  \u00a0\u00a0\u00a0\u00a0Female                     70 (46.36)\n  Paroxysmal AF                  130 (86.09)\n  Smoking history                23 (15.23)\n  Drinking history               19 (12.58)\n  Hypertension                   80 (52.98)\n  Diabetes                       26 (17.22)\n  Coronary heart disease         34 (22.52)\n  CHA2DS2-Vasc                   2.04\u00b11.56\n  Ultrasound parameters          \n  \u00a0\u00a0\u00a0\u00a0LAAEV (cm/s)               62.55\u00b123.16\n  \u00a0\u00a0\u00a0\u00a0LAD (mm)                   35.65\u00b13.78\n  \u00a0\u00a0\u00a0\u00a0LVEF (%)                   61.19\u00b17.93\n  Pulmonary vein diameter (cm)   1.66\u00b10.37\n  Procedural variables           \n  \u00a0\u00a0\u00a0\u00a0Procedure time (min)       66.68\u00b136.03\n  \u00a0\u00a0\u00a0\u00a0Fluoroscopic time (min)    10.64\u00b15.85\n  Cryoablation variables         \n  \u00a0\u00a0\u00a0\u00a0Nadir temperature (\u00b0C)     -44.12\u00b17.94\n  \u00a0\u00a0\u00a0\u00a0Freeze-time (s)            2.21\u00b11.01\n  \u00a0\u00a0\u00a0\u00a0Thaw time (s)              39.97\u00b116.62\n\nAF, atrial fibrillation; LAAEV, emptying velocity of left atrial appendage; LAD, left atrial diameter; LVEF, left ventricular ejection fraction.\n\n###### AUC of different thawing rate intervals.\n\n  Interval   AUC     95% CI        *p*-value\n  ---------- ------- ------------- -----------\n  ITR-10     0.736   0.692-0.781   \\<0.01\n  ITR-5      0.754   0.710-0.799   \\<0.01\n  ITR0       0.733   0.689-0.778   \\<0.01\n  ITR5       0.709   0.663-0.754   \\<0.01\n  ITR10      0.783   0.742-0.824   \\<0.01\n  ITR15      0.823   0.788-0.859   \\<0.01\n  ITR20      0.81    0.772-0.849   \\<0.01\n\nAUC, area under curve; 95% CI, 95% confidence interval; lITR-10, interval thawing rate at -10\u00b0C; ITR-5, interval thawing rate at -5\u00b0C; ITR0, interval thawing rate at 0\u00b0C; ITR5, interval thawing rate at 5\u00b0C; ITR10, interval thawing rate at 10\u00b0C; ITR15, interval thawing rate at 15\u00b0C; ITR20, interval thawing rate at 20\u00b0C.\n"}
+{"text": "INTRODUCTION\n============\n\nIrreversible bronchial dilatation, whether localized or diffuse. is known as bronchiectasis. Typically, it is the result of chronic infection, obstruction of nearby airways, or congenital bronchial abnormalities that lead to infection, such as cystic fibrosis or ciliary dyskinesia^([@r1])^. Bronchiectasis affects between 1/1,000 and 1/5,000 people in the general population^([@r2])^.\n\nWhen there is clinical suspicion of bronchiectasis, the investigatory algorithm uses imaging methods for diagnosis, including simple thoracic radiography and high-resolution computed tomography (HRCT) of the lungs. HRCT is currently considered the method of choice for diagnosis. The protocol of acquiring images with slices of 1.0-1.5 mm in thickness, at 10 mm intervals, shows a sensitivity and specificity of 98% and 93-99%, respectively^([@r3]-[@r5])^.\n\nThe use of quantifying systems for bronchiectasis, using imaging examinations---also known as radiological scores---allows abnormal findings to be standardized and correlates well with functional test results and with quality of life^([@r6]-[@r8])^, such systems being used as tools for clinical monitoring of patients and their therapeutic response^([@r9])^.\n\nThe first bronchiectasis quantification system devised for simple radiography was presented in 1958^([@r10])^. Since then, a series of quantification systems using HRCT have been developed^([@r11]-[@r18])^, and others have recently been proposed for tomosynthesis^([@r19])^ and magnetic resonance imaging^([@r20])^. However, the majority of those systems are based on tomography findings in patients with cystic fibrosis.\n\nTo our knowledge, there have been no studies using HRCT to quantify bronchiectasis among adults in developing nations such as Brazil. The objective of the present study was to evaluate radiologist agreement for the quantification of bronchiectasis using HRCT scans and for the identification of the findings associated with the disease, such as mucous plugging, bronchial wall thickening, and pulmonary involvement, using the scoring system proposed by Brody et al.^([@r18])^.\n\nMATERIALS AND METHODS\n=====================\n\nThis was a prospective observational study. We selected 43 HRCT scans of adult patients diagnosed with bronchiectasis at the pulmonology outpatient clinic of a tertiary reference center between March and June of 2008. In all of the patients, the diagnosis of bronchiectasis was based on clinical data and imaging findings. The patients were included consecutively at the time of their outpatient visit. The study was approved by the research ethics committee of the institution, and all participating patients gave written informed consent.\n\nDuring the clinical follow-up, the patients underwent HRCT scans of the chest, as requested by the attending physician. The tomography slices were obtained according to the following criteria: in the supine position; without intravenous contrast; 1-2 mm in thickness; an increase of 10 mm; a high-resolution filter; a matrix of 512 \u00d7 512; and an exposure time of 1-2 s per slice. The images obtained were documented with windows appropriate for the study of the lung parenchyma (window width: 1000 to 1500; window level: -600 to -800) and of the mediastinum (window width: 300 to 400; window level: 50 to 80). The expiratory phase was carried out in 18 patients, but in only 10 of these was it considered viable for analysis by radiologists. Cases in which there was convexity of the posterior wall and no significant reduction of the tracheal lumen during the expiratory phase were excluded, as were those in which there were respiratory movement artifacts that hampered the correct analysis of the images.\n\nThe HRCT scan images were analyzed by two radiologists, both with less than five years\\' experience in thoracic radiology, who had received standard training for HRCT during their academic studies and prior guidance regarding the evaluation and classification of examinations from another radiologist, with more than five years\\' experience in thoracic imaging. The readings were carried out in an independent and blind manner. The two evaluating radiologists were blinded to the patient history and clinical data, which could influence the evaluation of the imaging findings. They evaluated the examinations on tomography films, on a light box. One of the radiologists carried out a subsequent analysis of the HRCT scans 30-90 days after, with the aim of evaluating intraobserver agreement.\n\nUsing the Brody et al. score^([@r18])^, each lung lobe (considering the lingular segment of the right upper lobe to be a separate lobe) was evaluated separately in terms of the presence, size, and extension of the bronchiectasis, as well as the related findings, such as mucous plugging, bronchial wall thickening, parenchymal opacities, areas of ground-glass attenuation, cysts/bullae, and air trapping.\n\nThe final score was calculated by summing the points of each specific finding of all six lobes, multiplied by the severity of the involvement. We standardized the scores on a scale that ranged from 0 to 100, higher values indicating greater disease severity.\n\nSubsequently, the interobserver and intraobserver agreement were evaluated for the quantification of the bronchiectasis and its associated findings on the HRCT scans. For the analysis of agreement, we used the kappa statistic (\u03ba), with 95% confidence intervals, which is useful for the categorization of the variability obtained by the interpretation of the two data sets. For the categorization of bronchiectasis, agreement was classified as follows: \u03ba \\< 0.00. no agreement; \u03ba of 0.00-0.20, slight; \u03ba of 0.21-0.40, fair; \u03ba of 0.41-0.60, moderate; \u03ba of 0.61-0.80, substantial; \u03ba of 0.81 to 1,00, near perfect^([@r21])^.\n\nIn cases in which we did not analyze all of the categories in either of the evaluations, the \u03ba test could not be applied. On those occasions, the calculation of general agreement was used: higher values indicating greater closeness between the evaluations of the observers.\n\nThe interobserver agreement for the final bronchiectasis score of each patient was determined in two ways. The interclass correlation coefficient shows the level of agreement and reproducibility between the two evaluations. Conover^([@r22])^ established the following criteria for the interclass correlation coefficient: \u2265 0.75, excellent correlation; \u2265 0.40 and \\< 0.75, moderate correlation; \\< 0.40, weak correlation.\n\nA Bland-Altman graph^([@r23])^ determines the agreement between two measuring methods, representing the difference between each pair of values by their means. To determine the limit of agreement on the graphs, the standard deviation of the mean of the difference was considered twice.\n\nRESULTS\n=======\n\nWe evaluated the HRCT scans of 43 patients (24 women and 19 men). Ages ranged from 15 to 78 years (mean, 46.5 years). The causative factors of bronchiectasis most often identified were repeat infections and tuberculosis sequelae, both occurring in 14 patients. In nine cases, it was not possible to determine the cause of the bronchiectasis, and those cases were classified as idiopathic. In one patient each, there was a history of primary ciliary dyskinesia, Kartagener syndrome, cystic fibrosis, IgM deficiency, atypical mycobacteriosis, and post-bullectomy bronchiectasis in the setting of chronic obstructive pulmonary disease.\n\nThere was a symmetrical distribution among the lung lobes of the alterations reported by the two observers. Of the findings reported by observer 1, 18% were in the middle or right upper lobes, 20% were in the lower right lobe, 12% were in the left upper lobe, 15% were in the lingula, and 17% were in the left lower lobe, comparable to the 17%, 19%, 19%, 10%, 18%, and 17%, respectively, for observer 2 ([Table 1](#t1){ref-type=\"table\"}).\n\n###### \n\nResults of agreement for each item evaluated.\n\n  ---------------------------------------------------------------------------\n  HRCT findings                             Interobserver\\   Intraobserver\\\n                                            agreement        agreement\n  ----------------------------------------- ---------------- ----------------\n  Bronchiectasis                            \u00a0                \u00a0\n\n  \u00a0\u00a0\u00a0\u00a0Maximum size                          \u03ba = 0.45         \u03ba = 0.54\n\n  \u00a0\u00a0\u00a0\u00a0Mean size                             \u03ba = 0.48         \u03ba = 0.5\n\n  \u00a0\u00a0\u00a0\u00a0Appearance                            \u03ba = 0.43         \u03ba = 0.47\n\n  \u00a0\u00a0\u00a0\u00a0Central distribution                  \u03ba = 0.48         \u03ba = 0.46\n\n  \u00a0\u00a0\u00a0\u00a0Peripheral distribution               71.70%           \u03ba = 0.51\n\n  Mucous plugging                           \u00a0                \u00a0\n\n  \u00a0\u00a0\u00a0\u00a0Central distribution                  68.30%           \u03ba = 0.39\n\n  \u00a0\u00a0\u00a0\u00a0Peripheral distribution               \u03ba = 0.34         \u03ba = 0.35\n\n  Bronchial wall thickening                 \u00a0                \u00a0\n\n  \u00a0\u00a0\u00a0\u00a0Severity                              \u03ba = 0.21         \u03ba = 0.3\n\n  \u00a0\u00a0\u00a0\u00a0Central distribution                  \u03ba = 0.16         \u03ba = 0.29\n\n  \u00a0\u00a0\u00a0\u00a0Peripheral distribution               61.30%           \u03ba = 0.29\n\n  Involvement of the pulmonary parenchyma   \u00a0                \u00a0\n\n  \u00a0\u00a0\u00a0\u00a0Opacity                               \u03ba = 0.39         71.90%\n\n  \u00a0\u00a0\u00a0\u00a0Ground glass attenuation              64.30%           \u03ba = 0.24\n\n  \u00a0\u00a0\u00a0\u00a0Cists/bullae                          \u03ba = 0.47         \u03ba = 0.44\n\n  \u00a0\u00a0\u00a0\u00a0Air trapping                          47.80%           53.10%\n\n  Final score                               \u00a0                \u00a0\n\n  \u00a0\u00a0\u00a0\u00a0Intraclass correlation coefficient    71.90%           \u03ba = 0.24\n\n  \u00a0\u00a0\u00a0\u00a095% confidence interval               \u03ba = 0.44         53.10%\n  ---------------------------------------------------------------------------\n\nIn the evaluation of the maximum size and mean size of the bronchiectasis ([Figure 1](#f1){ref-type=\"fig\"}), interobserver agreement was moderate, with \u03ba values of 0.45 (*p* \\< 0.001) and 0.48 (*p* \\< 0.001), respectively. The intraobserver agreement was also moderate, with \u03ba values of 0.54 (*p* \\< 0.001) and 0.50 (*p* \\< 0.001), respectively.\n\nFigure 1Chest HRCT scans of patients with cylindrical bronchiectasis (arrows). **A:** Bronchiectasis in the left lung, with an estimated bronchial caliber twice that of the adjacent artery. **B:** Bronchiectasis in the right lower lobe, with an estimated bronchial caliber three times greater than normal. **C:** Bronchiectasis in the upper left lobe, the largest with an estimated bronchial caliber four times greater than normal.\n\nIn the classification of the appearance of bronchiectasis (cylindrical, saccular, or varicose) ([Figure 2](#f2){ref-type=\"fig\"}), there was also moderate agreement: \u03ba = 0.43 (*p* \\< 0.001) for interobserver agreement; and \u03ba = 0.47 (*p* \\< 0.001) for intraobserver agreement. For central distribution, the interobserver agreement was moderate (\u03ba = 0.48; *p* \\< 0.001), as was the intraobserver agreement (\u03ba = 0.46; *p* \\< 0.001). For peripheral distribution, the general interobserver agreement was 71.7% and the intraobserver agreement was moderate (\u03ba = 0.51; *p* \\< 0.001). Central bronchiectasis was found by observer 1 in 64% of the cases and by observer 2 in 66%. However, peripheral bronchiectasis was found in 36% of the cases by observer 1 and in 34% by observer 2.\n\nFigure 2CT of the chest in patients with different patterns of bronchiectasis (arrows). **A:** Cylindrical central bronchiectasis. **B:** Varicose central bronchiectasis. **C:** Saccular central bronchiectasis.\n\nFor the presence of mucous plugging, agreement was better when the location was central, with a general interobserver agreement of 68.3% and fair intraobserver agreement (\u03ba = 0.39; *p* \\< 0.001) in the evaluation. For peripheral mucous plugging, interobserver agreement and intraobserver agreement were both fair---\u03ba = 0.34 (*p* \\< 0.001) and \u03ba = 0.35 (*p* \\< 0.001), respectively. Central mucous plugging was found in 42% of the cases by observer 1 and in 46% by observer 2, whereas peripheral mucous plugging was found in 58% of the cases by observer 1 and in 54% by observer 2.\n\nThe levels of interobserver and intraobserver agreement were lowest for the finding of severe bronchial wall thickening (\u03ba = 0.21 and \u03ba = 0.30, respectively; *p* \\< 0.001), especially when the thickening was centralized (\u03ba = 0.16 and \u03ba = 0.29, respectively; *p* \\< 0.001). In the analysis of peripherally distributed bronchial wall thickening, the interobserver agreement was better, with a general agreement of 61.3% and fair intraobserver agreement (\u03ba = 0.29; *p* \\< 0.001). Central bronchial wall thickening was found in 78% of the cases by observer 1 and in 71% by observer 2. In contrast, peripheral bronchial wall thickening was found in 22% of the cases by observer 1 and in 29% by observer 2.\n\nFor the findings of opacity in the pulmonary parenchyma, ground-glass attenuation, and cysts/bullae, the interobserver agreement was fair (\u03ba = 0.39; *p* \\< 0.001), good (general agreement of 64.3%), and moderate (\u03ba = 0.47; *p* \\< 0.001), respectively, comparable to the intraobserver agreement, which was good (general agreement of 71.9%), fair (\u03ba = 0.24; *p* \\< 0.001), and moderate (\u03ba = 0.44; *p* \\< 0.001), respectively.\n\nFor the patients in whom good-quality expiratory phase images were available (*n* = 10), the general interobserver agreement for the finding of air trapping was good (general agreement of 47.8%), although the intraobserver agreement for that finding was slightly better (general agreement of 53.1%).\n\nFor interobserver agreement, the interclass correlation coefficient was 0.85 (95% CI: 0.74-0.91), compared with 0.81 (95% CI: 0.68-0.89) for intraobserver agreement ([Figure 3](#f3){ref-type=\"fig\"}.\n\nFigure 3Interobserver agreement (**A**) and intraobserver agreement (**B**) for the final bronchiectasis score of each patient. The degree of agreement by the intraclass correlation coefficient can be considered excellent and is confirmed by the Bland-Altman graph, in which only a few points were outside the upper and lower limits established.\n\nDISCUSSION\n==========\n\nThe use of HRCT scans to evaluate individuals with lung diseases has been the reason for a series of recent publications in the radiology literature of Brazil^([@r24]-[@r32])^. The results of the present study demonstrate that, in the HRCT evaluation of patients with bronchiectasis, interobserver and intraobserver agreement was generally moderate for each tomography finding in isolation. However, in obtaining the total score for each patient, the degree of agreement and reproducibility between the two evaluations was excellent.\n\nCylindrical bronchiectasis was the most common morphological pattern observed in the present study, a finding that is in agreement with data in the literature^([@r33])^. Varicose bronchiectasis can be confused with other morphological patterns if it is analyzed in a cross-sectional plane, which can mimic cylindrical and cystic bronchiectasis. These artifacts can explain the moderate interobserver and intraobserver agreement (\u03ba = 0.43 and \u03ba = 0.47, respectively) in the characterization of the morphological type of bronchiectasis.\n\nInterobserver agreement and intraobserver agreement were both moderate for the detection of central mucous plugging, whereas they were both only fair for the detection of peripheral mucous plugging. This was probably due to the presence of other findings that could mimic the presence of \\\"tree-in-bud\\\" nodules, such as centrilobular nodules classified as areas of ground-glass attenuation or movement artifacts.\n\nThe thickening of the bronchial wall is a common finding in patients with bronchiectasis. One recent longitudinal study, in an adult population, showed that the severity of bronchial wall thickening was the main determinant of functional decline, implying that HRCT scans could be useful in monitoring the progress of the disease^([@r34])^. Numerous authors have proposed methods of qualitatively evaluating bronchial wall thickening^([@r35])^. However, none of those methods have gained widespread acceptance, and the definition of what constitutes abnormal thickening of the bronchial wall remains uncertain^([@r36])^. With the advent of multi-detector computed tomography (MDCT) devices, it became possible to identify bronchial wall thickening through the use of dedicated software^([@r37])^. In the present study, bronchial wall thickening was the finding for which the level of interobserver and intraobserver agreement was the lowest, principally when the thickening was centralized.\n\nSome studies have shown that the interobserver agreement for the evaluation of bronchial wall thickening is worse than is that for the evaluation of bronchial dilation. Roberts et al.^([@r38])^ demonstrated near perfect interobserver agreement for the severity and extent of bronchiectasis (\u03ba = 0.87 and \u03ba = 0.82, respectively), and moderate interobserver agreement for bronchial wall thickening. Diederich et al.^([@r26])^ found poorer interobserver agreement for those same three parameters (\u03ba = 0.78, \u03ba = 0.76, and \u03ba = 0.64, respectively).\n\nIn the present study, we found moderate interobserver and intraobserver agreement for the presence of pulmonary opacities, ground-glass attenuation, and cysts/bullae. The agreement for such findings could be better if the observers were more experienced in chest radiology, allowing greater distinction of the lesion patterns in the HRCT scans. However, we sought to reproduce the use of the quantification model for bronchiectasis by general radiologists, not thoracic imaging specialists, mirroring what would more commonly take place in daily practice.\n\nIn a study involving 70 individuals with bronchiectasis diagnosed by HRCT, air trapping was observed in 34% of the images obtained during expiration^([@r39])^. In the present study, the sample of patient in whom appropriate expiratory HRCT scans were available was small (*n* = 10), due to the technical difficulties for the acquisition of examinations of good quality. We found moderate interobserver and intraobserver agreement for the determination of air trapping, reflecting what is found in the literature^([@r35])^.\n\nSome studies have demonstrated that MDCT, especially when performed with a 16-slice scanner, is superior to HRCT in the evaluation of bronchiectasis^([@r2],[@r40],[@r41])^. However, HRCT still plays an important role in Brazil, because it is more accessible. HRCT can be carried out on helical tomography devices, with a lower level of technological advancement than MDCT, without diminishing quality. Another advantage of HRCT scans is the acquisition of the image during a short breath-hold (with a duration of approximately 1 s) per tomography slice, whereas the acquisition of the fine MDCT slices of the whole chest requires a breath-hold of 10-15 s^([@r41])^. For dyspneic patients, this difference can be decisive in the quality of the image. In addition, there is a difference between the two methods in terms of the radiation dose, which is much higher in MDCT^([@r42],[@r43])^.\n\nAs limitations of our study, we can cite the evaluation of images on tomography films, which limited the evaluation of the findings. The analysis of the examinations on a workstation would facilitate the differentiation between bronchiectasis and cystic lesions, for example, increasing interobserver agreement. However, there was no significant disadvantage for the final result by score.\n\nIn conclusion, we observed only moderate interobserver and intraobserver agreement in the evaluation of the tomography findings of bronchiectasis, when evaluated in isolation. However, in the final analysis of the findings, there were no significant changes in the classification of the severity of bronchiectasis for each patient. This shows the good reproducibility of the method for the quantification of bronchiectasis, with the aim of determining the severity of the disease.\n\nThe results of the present study, demonstrating an excellent correlation for the tomography score, according to the quantification model proposed by Brody et al.^([@r18])^, illuminate the possibility of correlating the quantification results and the tomography score with the clinical and functional data of the patients. This allows the HRCT of the chest to be considered a prognostic factor and serve as a tool for the non-invasive monitoring of adult patients with bronchiectasis.\n\nStudy conducted in the Department of Diagnostic Imaging of the Escola Paulista de Medicina da Universidade Federal de S\u00e3o Paulo (EPM-Unifesp), S\u00e3o Paulo, SP, Brazil.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nPyle disease (metaphyseal dysplasia) was first described by Edwin Pyle in 1931, as \"a case of unusual bone development\" \\[[@CR1]\\]. Since then only few cases have been reported worldwide with a prevalence \\<1 per million \\[[@CR2]\\]. It is usually mild and incidentally diagnosed, despite the impressive radiological findings of gross metaphyseal widening and thinning of cortical bone \\[[@CR3]\\]. There is a considerable overlap with craniometaphyseal dysplasia (CMD), which is far more common and has more severe phenotypic features, including deafness and visual impairment \\[[@CR4]\\].\n\nWe report an eight-year-old girl, who presented with chronic facial nerve palsy, dental malocclusion, and incidental skeletal dysplasia on imaging studies. To the best of our knowledge, this is the first time that metaphyseal dysplasia is reported in association with cranial nerve compression. The clinical and radiological evidence for and against metaphyseal dysplasia and craniometaphyseal dysplasia is considered.\n\nCase report {#Sec2}\n===========\n\nAn eight-year-old girl was referred to our Pediatric Neurology service with a 6-year history of right-sided facial palsy. Her speech became acutely slurred but subsequently improved. There was no history of preceding trauma, recent infection, immunizations, or a rash. Parents declined to receive oral steroids. Brain magnetic resonance imaging (MRI) was initially obtained a year after the onset of her symptoms and was reported as normal, but it was unavailable for us to review.\n\nAt first evaluation, she was making good academic progress and she had been healthy over the last few years with normal development and physical growth. However, she had long-standing dental malocclusion for which she required extraction of five teeth.\n\nThe child was born at term by elective Caesarean section to healthy non-consanguineous parents. The mother had amniocentesis during this pregnancy as both she and father were carriers of beta thalassemia trait but without perinatal complications. She had a 16-year-old brother, who was fit and well, while family history was negative.\n\nShe had no features of facial dysmorphism such as wide nasal bridge, prominent forehead, hypertelorism, and prominent jaw. There was obvious tooth crowding with malocclusion. The rest of the neurological examination was normal. Laboratory evaluation with autoimmune and infectious screens was negative, while beta thalassemia trait was confirmed. Vitamin D deficiency was noted and replacement therapy was started.\n\nA repeat brain MRI revealed significant bone abnormalities, with thickening of skull vault and skeletal skull base, but normal intracranial appearances. In view of these findings, a brain computed tomography (CT) scan was obtained (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}) and the findings were consistent with a grossly abnormal skull base and a narrowed right facial canal, which led to the right facial nerve palsy. The visualized bones showed generalized expansion, with cancellous bone demonstrating ground glass appearance. In addition, there was abnormal pneumatization of the paranasal and mastoid air cells, with just a few small pneumatized foci in the mastoids, and there was no pneumatization of the sphenoid or maxillary antra. The clivus was expanded and it was abnormal.Fig. 1Diffuse hyperostosis of the bones of the skull. The facial canal is narrowed on the right (*arrow*)\n\nA skeletal survey was also performed, which revealed abnormal thickening of the skull base and calvarium, dental malocclusion and mild deformities of the long bones with abnormal tubulation (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). There was widening of the lesser diaphyseal region and bowing of the humeri and radial bones (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}). Widening of the metacarpals and phalanges (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}) and also widening of the metatarsals and phalanges of both feet, of the clavicles, and of the ribs were also noted. The vertebral bodies were normal in appearance. Typical Erlenmeyer flask deformities were noted to both femurs with grossly widened metaphyses of long bones and marked cortical thinning and osteopenia (Fig.\u00a0[2d, e](#Fig2){ref-type=\"fig\"}). Based on the above radiological findings, the diagnosis of Pyle's disease with facial nerve compression leading to chronic facial nerve palsy was set.Fig. 2**a**--**e** There is thickening of the calvarium and mid-skull base sclerosis. The long bones are under modeled. Erlenmayer flask deformity of the femora and distal tapering of the humeri is noted. The tubular bones of the hand are under tabulated. There is distal broadening of the radius\n\nDiscussion {#Sec3}\n==========\n\nPyle disease is a rare autosomal recessive disease with impressive and characteristic radiological findings but relatively mild clinical features. The clinically asymptomatic heterozygote may have minor disturbances of modeling of the tubular bones \\[[@CR5]\\]. The condition is due to a defect in metaphyseal remodeling that results in gross widening of the metaphyses of long bones, marked thinning of the cortex, and osteoporosis \\[[@CR6]\\]. The characteristic appearance of the long bones is described as Erlenmeyer flask deformity (EFD) and is usually more prominent in distal femur and proximal tibia. Spinal and jaw involvement have been described, while dental malocclusion is also a consistent finding \\[[@CR7], [@CR8]\\]. Patients with Pyle disease are often asymptomatic although genu valgus deformity may be a feature \\[[@CR9]\\]. Dental anomalies may require orthodontic interventions, while skeletal anomalies may need orthopedic surgery \\[[@CR8]\\].\n\nThe diagnosis of Pyle disease is usually made on the basis of clinical and radiological features, although the clinical signs are usually subtle and the diagnosis is made incidentally. The differential diagnosis of metaphyseal dysplasia from craniometaphyseal dysplasia is difficult, but important for the prognosis \\[[@CR10]\\]. Table [1](#Tab1){ref-type=\"table\"} shows the clinical and radiological characteristics of craniometaphyseal disease (autosomal dominant and recessive) and Pyle's disease. In particular, Pyle disease has been confused mainly with the autosomal dominant form of craniometaphyseal dysplasia, which is milder than the recessive type \\[[@CR9]\\]. Pyle disease differs from craniometaphyseal dysplasia clinically by the absence of any dysmorphic facial feature, milder skull involvement, lack of cranial nerve compression, and more striking long bone modeling defects leading to typical EFD \\[[@CR11], [@CR12]\\]. Our patient had no facial dysmorphic features, while she had a typical Erlenmeyer flask deformity in the long bones. The interesting finding is that this patient had an intense thickening of the skull and the calvarium, which led to narrowing of the right facial canal and to the compression of the facial nerve. Cranial nerve compression in Pyle's disease has never been previously reported, since the limited cases described so far in the literature had either none or mild involvement of the skull.Table 1Craniometaphyseal dysplasias vs Pyle's diseaseCMD AR typeCMD AD typePyle diseaseDysmorphic features++++++--Cranial sclerosis++++++Cranial nerve palsy+++--Dental malocclusion+++Bone fragility----+Modeling defects in long bones++++Erlenmayer flask+++Club-shaped configuration++--*CMD AR type* craniometaphyseal dysplasia autosomal recessive type, *CMD AD type* craniometaphyseal dysplasia autosomal dominant type\n\nAs far as the management is concerned, surgical reconstruction of the facial nerve could have been an option \\[[@CR4]\\]. However, as there was such a significant improvement surgical intervention was not deemed to be appropriate. Dental malocclusion was another significant problem for our patient and teeth extractions were necessary. Otherwise, she is able-bodied and lives a normal life.\n\nChronic facial nerve palsy due to compression of the facial nerve in a patient with Pyle's disease represents an unusual novelty, with a strong clinical impact, especially regarding the management and the prognosis. Furthermore, this case delineates the clinical spectrum and phenotype of such a rare clinical entity.\n\nDr Maria Kinali is the senior author.\n\nConflict of interest {#FPar1}\n====================\n\nThe authors declare that they have no conflict of interest.\n"}
+{"text": "![](hosplond72208-0010){#sp1 .166}\n"}
+{"text": "1. Introduction {#sec1}\n===============\n\nThe human nicotinic acetylcholine receptor (AChR) composed of five subunits (2 *\u03b1*, 1 *\u03b2*, 1 *\u03b4*, and either 1 *\u03b3* or 1 *\u03b5* subunit) is well characterized as the target antigen in myasthenia gravis (MG) \\[[@B1]\\]. The *\u03b3* subunit of the fetal receptor is replaced by an *\u03b5* subunit in adult muscle; both subunits share about 53% homology at the amino acid level \\[[@B2]\\]. Pathogenic antibodies are predominantly directed against the *\u03b1* subunit of the AChR. Both antibody responses as well as B-lymphocyte activity have been investigated extensively in MG and are of great diagnostic and prognostic value.\n\nImmunoglobulin G (IgG) autoantibody production is T helper cell-dependent. Although MG is considered a prototypic paradigm for an antibody-driven autoimmune disorder, the pathogenetic importance of T-helper cells is well appreciated. Several studies have been performed comparing T-cell responses involving the *\u03b1* subunit versus the developmentally regulated *\u03b5* subunit using recombinant fragments and purified polypeptides of the human AChR \\[[@B2]--[@B8]\\]. The *\u03b5* subunit is of particular interest as its expression in the adult muscle differs from the fetal *\u03b3* subunit, a fact that may contribute to the escape of clonal deletion and the development of autoreactive T lymphocytes in MG, especially in the myasthenic thymus \\[[@B9]\\]. In accordance with this hypothesis, two reports describe that, in comparison to healthy subjects, only MG patients responded to synthetic peptides of the *\u03b5* subunit by T-cell proliferation \\[[@B2], [@B10]\\]. Consistently, in MG patients with thymomas the *\u03b5* subunit is preferentially expressed \\[[@B11]\\].\n\nWe used an enzyme-linked immunospot (ELISPOT) assay to determine T-cell responses against recombinant fragments and synthetic peptides of the human AChR. In accordance with previous observations we found prominent T-cell responses against the *\u03b5* subunit while no significant differences were notable against alpha subunit epitopes.\n\n2. Patients and Methods {#sec2}\n=======================\n\n2.1. Patients and Controls {#sec2.1}\n--------------------------\n\nPeripheral blood lymphocytes (PBL) were obtained with informed consent from patients with generalized or ocular MG (n=15) or healthy donors (HD, n=9). PBL were isolated by density centrifugation and were either frozen immediately and thawed for analysis or used directly.\n\n2.2. Synthesis of Recombinant Fragments {#sec2.2}\n---------------------------------------\n\nHuman *\u03b1* and *\u03b5* subunit polypeptides were synthesized by PCR on cDNA prepared from total RNA of human calf muscle as described elsewhere \\[[@B12]\\]. Recombinant protein fragments were kindly provided by Wolfgang Wienhold and Arthur Melms \\[[@B13]\\]. Fragments were expressed in*E. coli* and purified by SDS/PAGE with a standard protocol \\[[@B5], [@B12]\\]. Alpha 1-103 and Alpha 1-209 are fragments of the extracellular domain, Alpha 327-298 of the intracellular domain of the *\u03b1* subunit \\[[@B1]\\]. Epsilon 1-221 is a fragment of the extracellular domain of the *\u03b5* subunit.\n\n2.3. Synthesis of Peptides {#sec2.3}\n--------------------------\n\nPeptides were synthesized by solid-phase Fmoc-chemistry on an automated peptide synthesizer for multiple peptide synthesis as described previously \\[[@B14]\\]. Epsilon 116-130, IDGQFGVAYDANVLV, is an HLA-DR3-binding peptide. Epsilon 201-215, ENGEWAIDFCPGVIR, contains a dominant epitope restricted by HLA-DR52a \\[[@B7]\\]. Epsilon 236-250, IRRKPLFYVINIIVP, contains a dominant T-cell epitope that is not HLA-DR3-restricted. As a specificity control peptide, we used the class II-associated invariant chain peptides, CLIP 97-120, LPKPPKPVSKMRMATPLLMQALPM, and CLIP 105-117, SKMRMATPLLMQA. The tetanus toxoid peptide TT 1272-1284 is a promiscuitive HLA-DR3/DR52a binder.\n\n2.4. ELISPOT Assay {#sec2.4}\n------------------\n\nWe measured frequencies of interferon (IFN)-*\u03b3*-secreting T-cells using an ELISPOT (enzyme-linked immunospot) assay as described previously \\[[@B15]\\]. Microtiter filter plates (Millipore) were coated overnight with an anti-human IFN-*\u03b3* monoclonal antibody (mAb) (clone 1-D1K; 10\u2009*\u03bc*g/ml, Mabtech, Sweden). After washing and blocking the plates with culture medium (RPMI 1640 supplemented with 5% fetal bovine serum and antibiotics, all from Gibco), fresh or freshly thawed PBL from MG patients and HD were incubated for 20\u2009h in duplicate in the presence or absence of human AChR antigens or controls (1\u2009*\u03bc*g/ml). Concanavalin A (5\u2009*\u03bc*g/ml; Sigma) was used as positive control. Using biotinylated anti-human IFN-*\u03b3* mAb (clone 7-B61; Mabtech), streptavidin-alkaline phosphatase (Mabtech), and BCIP/NBT as substrate (Sigma), antigen-specific IFN-*\u03b3* secreting T lymphocytes were visualized and counted on a dissecting microscope. Results are calculated and assigned as the numbers of IFN-*\u03b3*-secreting events among 10^6^ PBL minus the corresponding numbers of events per 10^6^ PBL without antigen.\n\n2.5. Statistical Analysis {#sec2.5}\n-------------------------\n\nStudent\\'s*t*-test was performed for statistical analysis. A p value of \\< 0.05 was accepted to be significant.\n\n3. Results and Discussion {#sec3}\n=========================\n\nThe T-cell responses were heterogeneous throughout MG patients. While single individuals did not show any detectable IFN-*\u03b3* secreting T-cells after stimulation with AChR fragments ([Table 1](#tab1){ref-type=\"table\"}), others exhibited marked responses (Figures [1](#fig1){ref-type=\"fig\"} and [2](#fig2){ref-type=\"fig\"}). The responses did not correlate with the AChR-antibody status. For example, patient MG-8 was AChR-antibody negative but exhibited a T-cell response to the AChR protein fragments. Accordingly, some HD presumably AChR-antibody negative gave positive T-cell responses. Analyzing the mean responses of all donors, the mean of detectable (positive) responses only, or the median positive responses only, recombinant fragments of both the *\u03b1* and the *\u03b5* subunit induced a higher T-cell response in MG than in HD ([Table 1](#tab1){ref-type=\"table\"}). However, statistical analysis could only determine a trend and not statistical significance.\n\nThe fragment Epsilon 1-221 showed the highest response. Synthetic peptides of the *\u03b5* subunit induced a lower response. The most remarkable difference between MG and HD was observed with Epsilon 201-215 containing a dominant T-cell epitope ([Table 1](#tab1){ref-type=\"table\"}) \\[[@B7]\\]. Consistent with this finding, Ragheb and colleagues have demonstrated proliferative T-cell responses upon stimulation with synthetic peptides of the *\u03b5* subunit in up to 15% of MG patients including Epsilon 194-209 \\[[@B2]\\].\n\nCorrelations between AChR-specific T-cell responses and paraclinical data (sex, age, or anti-AChR antibody serum titer) were not observed (see [Table 1](#tab1){ref-type=\"table\"}). In an animal model of MG, experimental autoimmune myasthenia gravis (EAMG), Gaertner et al. investigated the pathogenicity of T-cell determinants of the *\u03b5* subunit \\[[@B8]\\]. Although IFN-*\u03b3* secretion by T-cells reactive to *\u03b5* subunit peptides was observed, these cells failed to induce EAMG upon transfer. Hence, these T lymphocytes were demonstrated to be nonpathogenic. It remains to be determined if this observation reflects a peculiarity of the EAMG model or if nonpathogenic, *\u03b5* subunit-specific T lymphocytes are present in MG patients. If so, it is speculated that they may contribute to the integrity of the neuromuscular junction \\[[@B8]\\].\n\nWe conclude that the IFN-*\u03b3* ELISPOT method may provide a valuable tool to measure AChR-specific T-cell responses in MG. In MG patients who tested positive, T lymphocytes specific for epitopes of the AChR *\u03b5* subunit may be a target for therapeutical intervention in MG.\n\nWe thank Bettina Heiling for excellent technical assistance. We thank Wolfgang Wienhold and Arthur Melms for providing reagents.\n\nData Availability\n=================\n\nData supporting the results of this study can be provided by the corresponding author.\n\nConflicts of Interest\n=====================\n\nThe authors declare that they have no conflicts of interest.\n\n![Frequency of IFN-*\u03b3*-secreting events per 10^6^ PBL in single donors. ELISPOT assay was performed as described in the text. Results are assigned as the numbers of IFN-*\u03b3*-secreting events among 10^6^ PBL minus the corresponding numbers of events per 10^6^ PBL without antigen. Negative results (spot number without antigen exceeding spot number with antigen) were defined zero. Grey bars, mean response \u00b1 SD; black bars, mean positive response \u00b1 SD; crosses, median positive response; MG, myasthenia gravis patients; HD, healthy donors. Note the high standard deviations due to heterogeneous responses of single individuals (see [Table 1](#tab1){ref-type=\"table\"}). Note the different Y axis scales using four different recombinant fragments. (a) fragment Alpha 1-103; (b) Alpha 1-209; (c) Alpha 327-398; (d) Epsilon 1-221.](NRI2019-1969068.001){#fig1}\n\n![(a) Mean response, (b) mean positive response, and (c) median response of the numbers of IFN-*\u03b3*-secreting events among 10^6^ PBL minus the corresponding numbers of events per 10^6^ PBL without antigen. Negative results (spot number without antigen exceeding spot number with antigen) were defined zero. MG, myasthenia gravis patients; HD, healthy donors.](NRI2019-1969068.002){#fig2}\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Group                      Donor   Sex\u2009\u2009\\     Age\u2009\u2009\\     AChR-Ab\u2009\u2009\\   \u2009           Spots per 10^6^ PBL                                                                                              \n                                     (f/m)      (years)    (nmol/l)                                                                                                                                  \n  -------------------------- ------- ---------- ---------- ------------ ----------- --------------------- ----------- ----------- ----------- ---------- ---------- ---------- ---------- ---------- -----------\n  MG                         MG-1    f          70         135.6        1           0                     0           0           0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-2    f          62         0.5          11          0                     0           0           0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-3    m          48         10.2         2           160                   560         416         680         n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-4    m          66         0.0          10          0                     0           36          0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-5    m          50         2.2          5           151                   160         320         409         n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-6    m          64         0.4          12          2                     0           28          0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-7    f          66         0.0          38          39                    0           5           0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-8    m          71         0.0          2           88                    20          10          158         n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          MG-9    m          58         2.8          13          5                     5           0           13          0          0          0          0          0          0\n\n  \u2009                          MG-10   f          82         8.1          6           18                    5           0           0           0          3          5          0          0          0\n\n  \u2009                          MG-11   f          25         76.9         20          18                    0           13          0           0          0          0          0          0          0\n\n  \u2009                          MG-12   f          36         1.7          39          0                     0           0           0           0          0          0          0          0          0\n\n  \u2009                          MG-13   f          53         1.5          1           13                    20          20          26          33         20         20         39         0          26\n\n  \u2009                          MG-14   f          38         46.0         3           0                     0           0           11          43         0          0          0          0          0\n\n  \u2009                          MG-15   f          32         22.2         134         111                   0           156         0           33         78         0          0          0          0\n\n                                                                                                                                                                                                     \n\n  Mean response              \u2009       **54.7**   **20.5**   **19.8**     **40.2**    **51.3**              **66.8**    **86.4**    **15.5**    **14.3**   **3.5**    **5.6**    **0.0**    **3.7**    \n\n  SD                         \u2009       \u2009          **16.4**   **38.5**     **33.9**    **57.5**              **146.5**   **129.7**   **197.0**   **19.6**   **29.0**   **7.3**    **14.7**   **0.0**    **9.8**\n\n  Mean positive response     \u2009       \u2009          \u2009          **62.7**     **128.3**   **118.0**             **216.0**   **36.1**    **33.3**    **12.3**   **39.0**   **0.0**    **26.0**              \n\n  SD                         \u2009       \u2009          \u2009          \u2009            \u2009           **60.7**              **219.7**   **150.9**   **274.1**   **6.2**    **39.6**   **10.3**   \u2009          \u2009          \u2009\n\n  Median positive response   \u2009       \u2009          \u2009          **38.0**     **19.8**    **41.5**              **91.8**    **32.5**    **19.5**    **12.3**   **39.0**   **0.0**    **26.0**              \n\n                                                                                                                                                                                                     \n\n  HD                         HD-1    m          47         n.d.         2           0                     5           10          13          n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          HD-2    f          26         n.d.         3           0                     10          0           3           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          HD-3    f          30         n.d.         1           0                     0           0           0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          HD-4    f          32         n.d.         19          56                    0           39          0           n.d.       n.d.       n.d.       n.d.       n.d.       n.d.\n\n  \u2009                          HD-5    m          79         n.d.         4           5                     3           0           13          0          0          3          0          0          0\n\n  \u2009                          HD-6    f          64         n.d.         49          0                     43          0           135         50         10         0          35         0          165\n\n  \u2009                          HD-7    f          53         n.d.         1           4                     11          0           11          11         0          11         0          7          7\n\n  \u2009                          HD-8    f          48         n.d.         6           0                     0           0           0           0          0          0          0          6          0\n\n  \u2009                          HD-9    m          60         n.d.         95          0                     0           0           0           0          0          0          0          63         0\n\n                                                                                                                                                                                                     \n\n  Mean response              \u2009       **48.8**   \u2009          **20.0**     **7.1**     **7.9**               **5.5**     **19.3**    **12.1**    **2.0**    **2.6**    **7.0**    **15.2**   **34.4**   \n\n  SD                         \u2009       \u2009          **17.5**   \u2009            **32.1**    **18.3**              **13.7**    **13.1**    **43.8**    **21.7**   **4.5**    **4.5**    **15.7**   **26.7**   **73.1**\n\n  Mean positive response     \u2009       \u2009          \u2009          **21.4**     **14.1**    **24.8**              **34.7**    **30.3**    **10.0**    **6.5**    **35.0**   **25.3**   **86.0**              \n\n  SD                         \u2009       \u2009          \u2009          \u2009            \u2009           **29.8**              **16.2**    **20.3**    **56.2**    **27.9**   \u2009          **5.7**    \u2009          **32.2**   **111.7**\n\n  Median positive response   \u2009       \u2009          \u2009          **5.0**      **10.0**    **24.8**              **12.5**    **30.3**    **10.0**    **6.5**    **35.0**   **7.0**    **86.0**              \n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nELISPOT assay was performed as described in the text. Results are assigned as the numbers of IFN-gamma-secreting events among 10^6^ PBL minus the corresponding numbers of events per 10^6^ PBL without antigen. Negative results (spot number without antigen exceeding spot number with antigen) were defined zero. MG, myasthenia gravis patients; HD, healthy donors; n.d., not done.\n\n[^1]: Academic Editor: Herbert Brok\n"}
+{"text": "Nasopharyngeal carcinoma (NPC) occurs sporadically in the West but constitute a major health problem in several parts of Asia, including Southern China, Hong Kong and Singapore ([Parkin, 2001](#bib28){ref-type=\"other\"}). In Singapore, the incidence rate in Chinese is among the highest in Asia and the disease ranks as the fifth most common cancer among Chinese males ([Chia *et al*, 2000](#bib10){ref-type=\"other\"}). The histological pattern of NPC among the Chinese population comprise mainly the World Health Organization (WHO) types II (nonkeratinising) and III (undifferentiated) ([Chan *et al*, 1998](#bib8){ref-type=\"other\"}). Both histological types are often considered together as they share similar epidemiological, clinical and serologic characteristics, as distinct from the epidermoid carcinoma of the head and neck region seen in the western population ([Shanmugaratnam and Sobin, 1993](#bib31){ref-type=\"other\"}; [Altun *et al*, 1995](#bib3){ref-type=\"other\"}).\n\nThere is a high incidence of distant metastases with nonkeratinising or undifferentiated NPC, as compared to other epidermoid carcinoma arising in the head and neck region ([Reddy *et al*, 1995](#bib30){ref-type=\"other\"}). This is especially true for patients who present with locally advanced disease ([Ahmad and Stefani, 1986](#bib1){ref-type=\"other\"}). A prospective study showed a high rate of subclinical distant metastasis, with a distinctive feature of bone marrow invasion ([Micheau *et al*, 1987](#bib23){ref-type=\"other\"}). In addition, the incidence of distant failure after radiotherapy treatment for locally advanced disease can be as high as 57% in N~3~ disease ([Lee *et al*, 1992](#bib19){ref-type=\"other\"}). In view of the high rate of systemic relapse, chemotherapy has been incorporated into the primary treatment of locally advanced disease in order to improve the outcome. Concurrent chemoradiotherapy was shown to improve overall and progression-free survival for locally advanced NPC in the Intergroup study ([Al-Sarraf *et al*, 1998](#bib2){ref-type=\"other\"}) and this was replicated in studies carried out in the endemic areas of Singapore ([Wee *et al*, 2004](#bib33){ref-type=\"other\"}), Hong Kong ([Lee *et al*, 2004](#bib18){ref-type=\"other\"}) and Taiwan ([Lin *et al*, 2003](#bib20){ref-type=\"other\"}). However, a significant proportion of patients would still relapse systemically despite combined modality therapy and many of these patients would ultimately succumb to disseminated disease.\n\nPatients with disseminated disease do not behave in a uniform manner. It is hence not surprising to see significantly variable results between studies of similar therapeutic manoeuvres in patients with metastatic NPC. We have shown in our previous study that by using several clinical and laboratory parameters, we were able to define three prognostically distinct groups of patients with disseminated NPC ([Ong *et al*, 2003](#bib26){ref-type=\"other\"}). We proposed that a prognostic index scoring system using these parameters could be used for a more accurate prognostic evaluation of a patient. However more importantly, it can also be used for a more accurate stratification of patients in prospective clinical studies and hopefully help to standardise reporting results of any therapeutic interventions.\n\nWe now report a follow-on study that reanalysed our previous findings to define a new prognostic index score and to validate this new score in a separate cohort of patients with disseminated NPC.\n\nPATIENTS AND METHODS\n====================\n\nPatients\n--------\n\nAll patients were treated at the Department of Medical Oncology, National Cancer Centre between January 1994 and January 2003. There were two different cohorts of patients: the first (Cohort 1) was the group on which the new prognostic index score was derived and included 220 patients treated between January 1994 and December 1999, while the second cohort (Cohort 2) was the group on which the new score was validated. Cohort 2 included 99 patients treated between January 2002 and January 2003 and 21 patients (not included in our previous analysis for Cohort 1) from a previous Institutional Review Board-approved phase II clinical trial conducted in 1996 ([Au *et al*, 1998](#bib4){ref-type=\"other\"}).\n\nAll patients had a histological confirmation of NPC and had computerised tomography (CT) scan of the posterior nasal space, chest X-ray and/or CT scan of the thorax, ultrasound or CT scan of the abdomen and bone radionuclide scan to identify the extent of systemic disease. Patients were classified into the International Union Against Cancer/American Joint Committee on Cancer (UICC/AJCC) stages using the clinical and radiological data.\n\nPretreatment patient and disease characteristics, disease-free interval (DFI) (time from the onset of primary radiotherapy or chemoradiotherapy to the time of distant relapse), type of chemotherapy given, best response to chemotherapy, type of salvage chemotherapy given and date of death were recorded for all patients.\n\nSurvival data\n-------------\n\nThe primary end point of interest was metastatic survival. Metastatic survival was defined as survival subsequent to the development of distant relapse, that is, from the first diagnosis of distant metastases to the time of death. Locoregional recurrence was not considered a distant relapse. The survival status of all patients was verified with Singapore\\'s national death registry for Cohorts 1 and 2 as on 30 June 2000 and 31 December 2003, respectively. The cohorts excluded patients who were nonresidents of Singapore and as notification is mandatory in the event of death, the mortality data obtained from the death registry was complete and exhaustive.\n\nStatistical analysis\n--------------------\n\nThe analysis to derive a new prognostic score was performed on Cohort 1 and focused only on 172 patients who had received chemotherapy. The original prognostic index score was based on all patients including patients who were not treated with chemotherapy ([Ong *et al*, 2003](#bib26){ref-type=\"other\"}). As chemotherapy has an impact on survival outcome in patients with disseminated disease, there was a need to rederive a new prognostic index score using only patients from Cohort 1 who received palliative chemotherapy as all the patients from Cohort 2 were treated with chemotherapy. Univariate and multivariable analyses were performed using the Cox proportion hazards model. The multivariable analyses were undertaken with both forward and backward stepwise procedures for identifying the independent prognostic variables. Factors that were considered in the derivation of the new prognostic index score included age, gender, performance status according to Eastern Co-operative Oncology Group (ECOG) criteria, specific metastatic sites, number of metastatic sites, metastasis at presentation, stage at first diagnosis, DFI, leucocyte count, haemoglobin (Hb) level, and albumin level, on the basis of our previous analysis. In order to construct the new prognostic index score, factors were entered as categorical values as far as possible to keep the computations simple, although categorisation inevitably will result in some loss of information. *P*\u2a7d0.05 was used as the cutoff value of statistical significance for variable selection in the multivariable modelling. The regression coefficient of each independent prognostic variable (the *\u03b2* in the Cox regression equation hazard ratio (HR)=e^*\u03b2*^) is then modified into an integer numerical value to construct the new prognostic index score. The patients were stratified, based on the new prognostic index score, into three different risk groups with significantly different median metastatic survivals.\n\nThe validation of the new prognostic index score was subsequently performed on Cohort 2 patients. Overall and median metastatic survival estimates and curves were obtained using the Kaplan--Meier method and logrank test was used to compare among the three prognostic groups stratified by the new score.\n\nRESULTS\n=======\n\nPatient and disease characteristics ([Table 1](#tbl1){ref-type=\"table\"})\n------------------------------------------------------------------------\n\n[Table 1](#tbl1){ref-type=\"table\"} shows the characteristics of patients in both cohorts who had received at least one line of palliative chemotherapy at diagnosis of distant metastases. All patients had nonkeratinising or undifferentiated NPC. There was a male predominance in both cohorts. Most characteristics were similar between both cohorts. The median age at diagnosis of metastases was 47 years for Cohort 1 and 48 years for Cohort 2. In all, 35 (20.3%) patients in Cohort 1 and 27 (22.5%) patients in Cohort 2 had metastases at diagnosis. The majority had distant relapse after treatment for locally advanced disease previously. Bone was the most common site of metastasis, followed by liver, lung and distal nodes. Most of the patients had multiple sites involved at diagnosis of metastases.\n\nThe proportion of patients with good ECOG status of 0 and 1 was higher in Cohort 2 (95.8%) compared to Cohort 1 (87.8%) (*P*=0.07). Several chemotherapy regimens were used in the first-line setting and these included Cisplatin and Fluorouracil, Paclitaxel and Gemcitabine either alone or in combination with Carboplatin.\n\nSurvival distribution\n---------------------\n\nPatients from Cohort 2 appeared to have better survival, although this was not statistically significant (logrank test *P*-value=0.07). See [Figure 1](#fig1){ref-type=\"fig\"}. The median metastatic survival for patients in Cohort 1 was 12.9 months (95% CI 10.5,15.3) and that for patients in Cohort 2 was 15.6 months (95% CI 13.2, 18.0). The 1-, 2- and 3-year survival proportions were 52, 25 and 10% for Cohort 1 and 66, 33 and 15% for Cohort 2, respectively.\n\nUnivariate and multivariate analysis\n------------------------------------\n\nThe univariate analysis was performed on the 172 patients who were treated with chemotherapy in Cohort 1. Factors considered in the univariate analyses were based on our previous study and included age, gender, performance status, laboratory parameters such as Hb, leucocyte count and albumin level, stage at first diagnosis, sites and number of metastases and DFI. The factors associated with an adverse prognosis were anaemia (Hb\\<12g\u2009dl^\u22121^), ECOG \u2a7e2, DFI\u2a7d6 months, presence of multiple metastases and liver metastases. See [Table 2](#tbl2){ref-type=\"table\"}.\n\nThe variables included in the multivariate analysis were similar to that used in the univariate analysis. Using a significance level of 0.05, the significant independent variables were Hb level (Hb\\<12g\u2009dl^\u22121^ with HR: 2.1), performance status (ECOG\u2a7e2 with HR: 2.6), DFI (metastasis at diagnosis and short DFI with HR: 1.2 and 7.7 respectively). See [Table 3](#tbl3){ref-type=\"table\"}. These three variables predicted negatively for metastatic survival.\n\nNew prognostic index score and risk groups\n------------------------------------------\n\nA numerical score was derived from the regression coefficient of each of the three independent prognostic variables derived above, namely, anaemia (Hb\\<12\u2009g\u2009dl^\u22121^), ECOG\u2a7e2 and DFI. A score of 0 was assigned if the factor was absent or 1, 4, 5 or 10 according to the factor present. See [Table 4](#tbl4){ref-type=\"table\"}. The new prognostic index score for each individual patient was calculated by adding up the scores of each independent factor. The maximum score obtainable was 19, instead of 20 as metastasis at diagnosis and short DFI were mutually exclusive. In view of the gaps between the scores, the possible scores were 0--1, 4--6, 9--11, 14--16 and 19. The patients were stratified into three prognostic groups based on the new prognostic index score: 57 patients in the low-risk group (score: 0--3), 86 patients in intermediate-risk group (score: 4--8) and 25 patients in high-risk group (score: \u2a7e9). The median metastatic survivals for the three different risk groups were 25.3 (95% CI 17.7--33.9), 11.7 (95% CI 9.9--13.6) and 5.8 (95% CI 5.0--6.5) months, respectively. (logrank test, *P*\\<0.0001). See [Figure 2](#fig2){ref-type=\"fig\"}.\n\nValidation of new prognostic index score\n----------------------------------------\n\nThe new prognostic index score was applied to the patients in Cohort 2. The proportion of patients in the low-risk, intermediate-risk and high-risk groups were 49, 36 and 15%, respectively. The median metastatic survivals of these three different groups were as follows: 19.6 (95% CI 16.1--23.1), 14.3 (95% CI 12.3--16.2) and 7.9 (95% CI 6.6--9.2) months respectively (logrank, *P*=0.003). See [Figure 3](#fig3){ref-type=\"fig\"}.\n\nDISCUSSION\n==========\n\nWe previously presented our results on derivation of a prognostic index score derived from all patients in Cohort 1 that included those not treated with chemotherapy ([Ong *et al*, 2003](#bib26){ref-type=\"other\"}). However, as Cohort 2 was used for validation of the score and all the patients in Cohort 2 were treated, we elected to reanalyse Cohort 1 to restrict the derivation of a new prognostic index score to only patients given chemotherapy. The reason for the reanalysis is that chemotherapy has an impact on survival outcome in patients with disseminated NPC, although there are no randomised comparison studies between chemotherapy and best supportive care ([Fandi *et al*, 1994](#bib13){ref-type=\"other\"}; [Hong *et al*, 1999](#bib15){ref-type=\"other\"}). This is also supported by the fact that incorporation of chemotherapy in locally advanced disease has been shown to improve progression-free and overall survival ([Al-Sarraf *et al*, 1998](#bib2){ref-type=\"other\"}; [Lin *et al*, 2003](#bib20){ref-type=\"other\"}; [Wee *et al*, 2004](#bib33){ref-type=\"other\"}). Thus, with reanalysis, we have eliminated the bias with the use of chemotherapy and make the two cohorts more comparable.\n\nWe found that DFI, Hb level (\\<12g\u2009dl^\u22121^) and poor performance status (ECOG\u2a7e2) were significant negative prognostic factors in patients treated with chemotherapy and these three independent variables were used in the derivation of the new prognostic index score. Metastasis at diagnosis and short DFI were regarded as two separate categories within the same DFI variable and thus were mutually exclusive. The natural history of a disease is dependent on the interaction of both patient and disease factors. Anaemia can be related to the disease process itself, host-related factors or treatment given and has been shown to be associated with significant reduction in survival in various cancers, other than NPC ([Caro *et al*, 2001](#bib6){ref-type=\"other\"}; [Bokemeyer *et al*, 2002](#bib5){ref-type=\"other\"}). Anaemia may reflect not only a biologically more aggressive tumour but may be a mediating factor to resistance to treatment and this has been demonstrated in retrospective studies on cervical cancer treated with chemoradiation ([Obermair *et al*, 2001](#bib25){ref-type=\"other\"}). The negative prognostic factor of a poor performance status has been shown in many other tumour types as well ([Paesmans *et al*, 1995](#bib27){ref-type=\"other\"}; [Polee *et al*, 2003](#bib29){ref-type=\"other\"}; [Motzer *et al*, 2004](#bib24){ref-type=\"other\"}). Patients with poor performance status do not tolerate treatment well and it may also be a reflection of the more advanced state of the cancer.\n\nPatients who present with metastasis at diagnosis can be a result of delay in seeking medical attention or more likely, a reflection of the aggressive nature of the disease such that it is widespread by the time the patient becomes symptomatic. Having a short DFI is also reflective of an aggressive disease and has been convincingly shown to portend a poorer outcome in other tumours such as ovarian cancer ([Markman *et al*, 1991](#bib22){ref-type=\"other\"}). Our analysis showed that a short DFI after initial radiotherapy or chemoradiotherapy for the primary tumour had a worse prognostic score compared to metastasis at diagnosis and this is likely due to emerging chemoresistant clones within the tumour of those patients who relapse shortly after treatment.\n\nA study by [Teo *et al* (1996)](#bib32){ref-type=\"other\"} showed that short DFI, the presence of liver metastasis and age at diagnosis were prognostic for metastatic survival in metastatic NPC. Short DFI is included as a factor in our new prognostic index score but the other two variables were found to be not statistically significant on multivariate analysis. This may be a result of our population of patients being limited to only those treated with chemotherapy and the treatment likely negated the prognostic significance of age and site of metastasis. Statistical issues, such as variations in modelling procedure in small to moderate size data sets between the different studies, could be an alternative explanation.\n\nThe metastatic survival of patients with disseminated NPC treated with chemotherapy is highly variable and our findings provide further supportive evidence for that. In Cohort 1, at the time of analysis in June 2000, the metastatic survival varied from 1 to 72 months. Similarly, for Cohort 2, the metastatic survival duration varied from 1 month to 103 months: 10 patients died within 3 months of diagnosis of metastatic disease despite chemotherapy, while 13 patients are still alive after 2 years. In fact, one patient survived 8 years 7 months while another three lived beyond 4 years after diagnosis of metastatic disease. A number of studies have reported similar findings, with several long-term survivors with the use of chemotherapy in metastatic NPC ([Chan *et al*, 1997](#bib9){ref-type=\"other\"}; [Fandi *et al*, 2000](#bib14){ref-type=\"other\"}). Based on the results of our studies, it is possible for single-arm studies or randomised studies that are inadequately powered or stratified to report spuriously improved survival outcome with a new therapeutic intervention because of biased patient selection alone. As a result of this heterogeneity, it is imperative that a method be developed to better prognosticate and stratify patients for more accurate assessment of efficacy of any new therapeutic interventions.\n\nIt is important to note that the new prognostic index score resulted in clear demarcation of three prognostically different groups with nonoverlapping confidence intervals of the median survival durations and this demarcation was replicated in the separate cohort of patients. Another important point to note is that the new prognostic index score incorporates clinical data that are used in routine patient care in most institutions.\n\nThere is accumulating evidence for the use of biomarkers in the detection and prognostication of NPC. The biomarkers include circulating markers in the blood, such as Epstein--Barr virus (EBV) DNA ([Chien *et al*, 2001](#bib11){ref-type=\"other\"}) and CYFRA 21-1 ([Ma *et al*, 2004](#bib21){ref-type=\"other\"}), as well as tumour immunohistochemical markers such as the expression of multidrug-resistance protein ([Hsu *et al*, 2002](#bib17){ref-type=\"other\"}), epidermal growth factor receptors ([Chua *et al*, 2004](#bib12){ref-type=\"other\"}) and signal transducers and activators of transcription factors ([Hsiao *et al*, 2003](#bib16){ref-type=\"other\"}). Among them, plasma EBV DNA has been the most validated and has been shown to be a predictor of poor survival after radiotherapy for local disease ([Chan *et al*, 2002](#bib7){ref-type=\"other\"}). However, these biomarkers are not readily available in most institutions and they are yet to be validated as prognostic factors in the metastatic setting.\n\nIt is conceivable that similar prognostic classifications can be designed and used for other disseminated solid tumours as well, for instance breast or colorectal cancers. Such a prognostic classification based on combination of patient and laboratory factors may be able to categorise patients with disseminated disease more accurately as shown in NPC in this study and hence will be useful for stratifying patients in randomised studies.\n\nWe have validated a new prognostic index score in metastatic NPC patients treated with chemotherapy. This new prognostic index score can stratify patients into three different prognostic groups with significantly different median metastatic survivals. This prognostic classification system will be useful for more accurate prognostication of patients with disseminated NPC. In addition, it can prove useful in the design of clinical trials for metastatic NPC as it can more accurately stratify patients into groups with fairly consistent outcome and thus make the results more comparable and interpretable.\n\nThis work was presented in part at the 40th annual meeting of the American Society of Clinical Oncology, 5--8 June 2004, New Orleans, LA, USA. The authors have no disclosures of financial interests to declare.\n\n![Survival curves for Cohorts 1 and 2. There is no statistically significant difference in metastatic survival between the two cohorts.](92-6602525f1){#fig1}\n\n![Survival by new prognostic index grouping derived from patients in Cohort 1. There is a clear demarcation of survival differences between the three risk groups.](92-6602525f2){#fig2}\n\n![Survival by new prognostic index grouping applied to patients in Cohort 2. There is a statistically significant difference between the three risk groups.](92-6602525f3){#fig3}\n\n###### \n\nPatient and disease characteristics\n\n  \u00a0                                       **Cohort 1**   **Cohort 2**   \u00a0                \n  --------------------------------------- -------------- -------------- --------- ------ ------------------\n  No. of patients                         172            \u00a0              120       \u00a0      \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Survival status*                       \u00a0              \u00a0              \u00a0         \u00a0      See logrank test\n  \u2003Dead                                   130            75.6           91        75.8   \u00a0\n  \u2003Alive                                  42             24.4           29        24.2   \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Age, in years*                         \u00a0              \u00a0              \u00a0         \u00a0      \\<0.001\n  \u2003Median                                 47             \u00a0              48        \u00a0      \u00a0\n  \u2003Interquartile range                    (40,54)        \u00a0              (42,54)   \u00a0      \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Gender*                                \u00a0              \u00a0              \u00a0         \u00a0      0.32\n  \u2003Male                                   144            83.7           95        79.2   \u00a0\n  \u2003Female                                 28             16.3           25        20.8   \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *ECOG status*^a^                        \u00a0              \u00a0              \u00a0         \u00a0      0.07\n  \u20030                                      18             10.5           19        15.8   \u00a0\n  \u20031                                      133            77.3           96        80     \u00a0\n  \u20032                                      14             8.1            2         1.7    \u00a0\n  \u20033                                      6              3.5            3         2.5    \u00a0\n  \u20034                                      1              0.6            0         0      \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Laboratory parameters*                 \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  \u2003Albumin (g\u2009l^\u22121^)                      \u00a0              \u00a0              \u00a0         \u00a0      0.38\n  \u2003\u2003\\<40                                  128            74.4           90        75     \u00a0\n  \u2003\u2003\u2a7e40                                   32             18.6           29        24.1   \u00a0\n  Hemoglobin (g\u2009l^\u22121^)                    \u00a0              \u00a0              \u00a0         \u00a0      0.003\n  \u2003\u2003\\<12                                  106            61.6           54        45     \u00a0\n  \u2003\u2003\u2a7e12                                   63             36.6           66        55     \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  UICC/AJCC stage at first diagnosis^b^   \u00a0              \u00a0              \u00a0         \u00a0      0.002\n  \u2003I                                      5              2.9            2         1.7    \u00a0\n  \u2003IIA                                    1              0.6            4         3.3    \u00a0\n  \u2003IIB                                    31             18             4         3.3    \u00a0\n  \u2003III                                    39             22.7           39        32.5   \u00a0\n  \u2003IVA                                    20             11.6           16        13.3   \u00a0\n  \u2003IVB                                    27             15.7           16        13.3   \u00a0\n  \u2003IVC                                    33             19.2           27        22.5   \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Disease-free interval*                 \u00a0              \u00a0              \u00a0         \u00a0      0.16\n  \u2003Mets at diagnosis                      35             20.3           27        22.5   \u00a0\n  \u2003\u2a7d6 months                              11             6.4            15        12.5   \u00a0\n  \u2003\\>6 month                              125            72.7           78        65     \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Sites of metastasis*                   \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  \u2003Bone                                   118            68.6           79        65.8   0.26\n  \u2003Liver                                  74             43             63        52.5   0.12\n  \u2003Lung                                   65             37.8           51        42.5   0.41\n  \u2003Distant nodes                          63             36.6           39        32.5   0.43\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *No. of metastatic sites*               \u00a0              \u00a0              \u00a0         \u00a0      0.3\n  \u2003Single                                 69             40.1           41        34.2   \u00a0\n  \u2003Multiple                               103            59.9           79        65.8   \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Prior chemotherapy before mets*        \u00a0              \u00a0              \u00a0         \u00a0      0.58\n  \u2003Yes                                    14             8.1            12        10     \u00a0\n  \u2003No                                     158            91.9           108       90     \u00a0\n  \u00a0                                       \u00a0              \u00a0              \u00a0         \u00a0      \u00a0\n  *Salvage chemotherapy*                  \u00a0              \u00a0              \u00a0         \u00a0      0.02\n  \u2003Yes                                    91             52.9           68        61.8   \u00a0\n  \u2003No                                     81             47.1           42        38.2   \u00a0\n\nECOG status refers to performance status as defined by the Eastern Cooperative Oncology Group.\n\nUICC/AJCC refers to International Union against Cancer/American Joint Committee on Cancer.\n\n###### \n\nUnivariate analysis of patients treated with chemotherapy in Cohort 1 (*N*=172)\n\n  **Factor**                               **No. of patients**   **No. alive**   **Hazard ratio (95% CI)**   ***P*-value**\n  ---------------------------------------- --------------------- --------------- --------------------------- ---------------\n  *Gender*                                 \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003Female                                  28                    6               Baseline                    \u00a0\n  \u2003Male                                    144                   36              0.66 (0.41--1.04)           0.075\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Age, in years*                          \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003\u2a7d54                                     78                    23              Baseline                    \u00a0\n  \u200346--65                                  84                    16              1.31 (0.92--1.88)           0.14\n  \u2003\\>65                                    10                    3               1.79 (0.81--3.96)           0.15\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Albumin, g\u2009l*^\u2212*1*^                     \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003\u2a7e04                                     32                    10              Baseline                    \u00a0\n  \u2003\\<40                                    128                   31              1.43 (0.89--1.88)           0.13\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Haemoglobin, g\u2009dl*^\u2212*1*^                \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003\u2a7e12                                     63                    28              Baseline                    \u00a0\n  \u2003\\<12                                    106                   14              2.61 (1.76--3.87)           \\<0.001\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *ECOG status*^a^                         \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u20030--1                                    151                   41              Baseline                    \u00a0\n  \u20032                                       14                    1               2.24 (1.26--4.00)           0.006\n  \u20033--4                                    7                     0               4.19 (1.93--9.18)           \\<0.001\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Leucocyte count, \u00d7 10*^*9*^ *l*^\u2212*1*^   \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003\\<4                                     26                    4               Baseline                    \u00a0\n  \u20034--11                                   112                   35              0.77 (0.48--1.24)           0.29\n  \u2003\\>11                                    31                    3               1.46 (0.83--2.56)           0.19\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  UICC/AJCC stage at first diagnosis^b^    \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003I--II                                   37                    8               Baseline                    \u00a0\n  \u2003III--IVB                                86                    23              1.21 (0.78--1.89)           0.38\n  \u2003IVC                                     33                    9               1.41 (0.81--2.45)           0.22\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Bone metastasis*                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003No                                      42                    10              Baseline                    \u00a0\n  \u2003Yes                                     118                   27              1.24 (0.82--1.85)           0.3\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Liver metastasis*                       \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003No                                      97                    26              Baseline                    \u00a0\n  \u2003Yes                                     74                    16              1.60 (1.13--2.28)           0.008\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Lung metastasis*                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003No                                      106                   27              Baseline                    \u00a0\n  \u2003Yes                                     65                    15              1.19 (0.83--1.70)           0.35\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Distal node metastasis*                 \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003No                                      107                   19              Baseline                    \u00a0\n  \u2003Yes                                     63                    22              0.81 (0.56--1.18)           0.27\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Number of metastatic sites*             \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003Single                                  69                    17              Baseline                    \u00a0\n  \u2003Multiple                                103                   25              1.55 (1.08--2.21)           0.017\n  \u00a0                                        \u00a0                     \u00a0               \u00a0                           \u00a0\n  *Disease-free interval*                  \u00a0                     \u00a0               \u00a0                           \u00a0\n  \u2003\\>6 months                              125                   32              Baseline                    \u00a0\n  \u2003\u2a7d6 months                               11                    1               4.03 (2.02--8.02)           \\<0.001\n  \u2003Metastases at diagnosis                 35                    9               1.25 (0.81--1.94)           0.31\n\nECOG status refers to performance status as defined by the Eastern Cooperative Oncology Group.\n\nUICC/AJCC refers to International Union against Cancer/American Joint Committee on Cancer.\n\n###### \n\nSignificant independent variables from multivariate analysis for patients treated with chemotherapy in Cohort 1 (*N*=172)\n\n  **Factor**                           **Hazard ratio (95% CI)**  ***P*-value**\n  ----------------------------------- --------------------------- ---------------\n  Haemoglobin level (\\<12\u2009g\u2009dl^\u22121^)       2.067 (1.34--3.17)      0.001\n  Performance status (ECOG \u2a7e2^a^)         2.585 (1.45--4.59)      0.001\n  Disease-free interval                            \u00a0              \u00a0\n  \u2003Metastasis at diagnosis                 1.21 (0.75--1.95)      0.031\n  \u2003\u2a7d6 months                              7.656 (2.23--26.25)     \u00a0\n\nECOG refers to performance status as defined by the Eastern Cooperative Oncology Group.\n\n###### \n\nNew prognostic index score and risk groups\n\n  **Factor**                          **Score**   ***\u03b2* (hazard ratio=e**^***\u03b2***^**)^a^**\n  ----------------------------------- ----------- ------------------------------------------\n  Haemoglobin level (\\<12\u2009g\u2009dl^\u22121^)   4           0.73\n  Performance status (ECOG \u2a7e2^b^)     5           0.95\n  Disease-free interval               \u00a0           \u00a0\n  \u2003Metastasis at diagnosis            1           0.19\n  \u2003\u2a7d6 months                          10          2.04\n  Maximum score                       19          \u00a0\n\nLow risk: score 0--3; intermediate risk: score 4--8; high risk: score \u2a7e9.\n\n*\u03b2*=Regression Coefficient.\n\nECOG refers to performance status as defined by the Eastern Cooperative Oncology Group.\n"}
+{"text": "1. Introduction {#sec1-cancers-11-00606}\n===============\n\nPancreatic cancer (PC) is expected to be the second cause of cancer deaths in Western countries by 2030. \\[[@B1-cancers-11-00606],[@B2-cancers-11-00606]\\] Surgical resection remains the only potentially curative treatment, but only 10--20% of cases are resectable at diagnosis. Even in this setting, long-term outcomes remain dismal, with a 5-year overall survival rate of 10%. \\[[@B3-cancers-11-00606]\\] This poor prognosis has led to the use of multimodal approaches, including adjuvant chemotherapy \\[[@B3-cancers-11-00606],[@B4-cancers-11-00606],[@B5-cancers-11-00606],[@B6-cancers-11-00606],[@B7-cancers-11-00606],[@B8-cancers-11-00606]\\], with a clear impact on 5-year overall survival, \\[[@B6-cancers-11-00606]\\] and adjuvant chemoradiation just in case of microscopically positive margins and/or lymph node involvement \\[[@B5-cancers-11-00606],[@B9-cancers-11-00606],[@B10-cancers-11-00606],[@B11-cancers-11-00606]\\].\n\nThe high rate of disease relapse, coupled with a low compliance (only 51% of patients outside the context of a clinical trial receive adjuvant therapy \\[[@B12-cancers-11-00606]\\]) remain major drawbacks of an adjuvant strategy. This has led some authors to uphold the use of preoperative treatment, with potential advantages such as an increased R0 resection rate, better compliance, a reduction in the risk of intraoperative tumor spillage and avoidance of unnecessary surgery, with its related morbidity and mortality, in patients with unfavorable tumor biology. Preliminary data with this approach, either preoperative chemotherapy or chemoradiation, seem promising. \\[[@B13-cancers-11-00606],[@B14-cancers-11-00606]\\]\n\nThe development of predictive tools for individual relapse-risk assessment after upfront multimodal therapy may help to further optimize treatment decision-making. While nomograms \\[[@B15-cancers-11-00606]\\] and molecular prognostic signatures have been validated in other solid tumors, their role in PC seems limited. \\[[@B16-cancers-11-00606],[@B17-cancers-11-00606],[@B18-cancers-11-00606]\\] The use of population models and machine learning algorithms to predict disease evolution might be considered as a potential alternative for this tumor type. \\[[@B19-cancers-11-00606],[@B20-cancers-11-00606],[@B21-cancers-11-00606],[@B22-cancers-11-00606],[@B23-cancers-11-00606],[@B24-cancers-11-00606],[@B25-cancers-11-00606],[@B26-cancers-11-00606]\\]\n\nThe aim of the present retrospective analysis is two-fold. First, to describe the clinical results achieved with the use of induction polychemotherapy (IPCT) followed by chemoradiation (CRT) in patients with resectable PC. Second, to evaluate the individual 2-year relapse-risk on a supervised machine-learning algorithm basis, among a group of patients with PC resected after preoperative therapy.\n\n2. Materials and Methods {#sec2-cancers-11-00606}\n========================\n\n2.1. Patient Elegibility {#sec2dot1-cancers-11-00606}\n------------------------\n\nAll patients diagnosed of potentially curable PC from September 2005 to November 2016 were evaluated by a multidisciplinary team composed of hepatobiliary surgeons, endoscopists, interventional radiologists, medical and radiation oncologists. Initial workup included: clinical examination, laboratory tests including a serum CA-19.9 level, endoscopic ultrasound (EUS) with guided fine needle aspiration biopsy (FNA) of the pancreatic lesion and a CT-scan to define the extent of the disease. When needed, biliary decompression was performed with an endobiliary stent prior to neoadjuvant treatment. Staging laparoscopy was considered in stage IV suspicious cases.\n\nPatients were considered for neoadjuvant therapy if they had a good performance status (\u22642 according to Eastern Cooperative Oncology Group (ECOG)), an adequate hematological, renal and liver function, and a histologically confirmed resectable/borderline-resectable PC. The criteria applied to determine resectability were based on the National Comprehensive Cancer Network, according to the consensus statement of the Society of Abdominal Radiology and the American Pancreatic Association ([Table S1](#app1-cancers-11-00606){ref-type=\"app\"}).\n\n2.2. Neoadjuvant Therapy {#sec2dot2-cancers-11-00606}\n------------------------\n\nNeoadjuvant treatment consisted of 2--4 months of IPCT followed, in the case of stable or responding disease, by CRT.\n\nMost patients received IPCT on an outpatient basis. Before each chemotherapy cycle all patients underwent routine work-up including physical examination, blood tests and treatment-induced adverse effects assessment. IPCT regimens included modified FOLFOXIRI \\[[@B27-cancers-11-00606]\\], GEMOX or GEMOXEL \\[[@B28-cancers-11-00606],[@B29-cancers-11-00606]\\].\n\nIn the case of stable or responding disease after IPCT, three-dimensional conformal external beam radiotherapy (3D-RT) or an intensity-modulated technique (IMRT) were planned, given that both techniques seem to be equally effective and have a similar toxicity profile in the neoadjuvant setting \\[[@B30-cancers-11-00606]\\]. The target volumes and organs at risk were contoured on each of the axial CT slices in the Helax-TMS treatment planning system (Nucletron Scandinavia, Uppsala, Sweden) or in the ADAC Pinnacle treatment planning system (Philips Radiation Oncology Systems, Fitchburg, WI, USA). The clinical target volume included the gross tumor volume of pancreas and the draining locoregional lymph nodes (peripancreatic and retroperitoneal). Conformation and field arrangement ensured that the organs at risk (kidneys, heart, liver, stomach, duodenum and spinal cord) received tolerable doses. Treatment planning followed International Commission on Radiation Units and Measurements recommendations. Patients were immobilized in the supine position. In general, four fields with 15-MV photons were employed to deliver 50 Gy over 4--5 weeks with conventional daily fractions of 1.8--2 Gy, 5 days per week. Five or seven coplanar, equally spaced beams were applied in a variable number of segments in IMRT plans. All patients received concurrent Capecitabine, 850 mg/m^2^ CORRECT bid during the days of radiation. Physical examination, blood test monitoring and therapy-induced toxicity were assessed on a weekly basis.\n\nSurgery was scheduled 4--6 weeks after the end of the neoadjuvant protocol, once progressive disease was ruled out by preoperative CT-scan \u00b1 EUS.\n\n2.3. Histological Data {#sec2dot3-cancers-11-00606}\n----------------------\n\nA standardized histologic evaluation of the surgical specimen was performed in all resected patients. It included: a pathologic stage, vascular and perineural invasion assessment, lymph node status, lymph node ratio (LNR), resection margins (an R0 resection was defined as no tumor within 1 mm of the resection margin), a tumor regression grade according to the College of American Pathologists (CAP) grading system \\[[@B31-cancers-11-00606]\\] ([Table S2](#app1-cancers-11-00606){ref-type=\"app\"}), and the degree of nodal response to treatment, evaluated with a 4-point scale adapted from the Miller & Payne grading system \\[[@B32-cancers-11-00606]\\] ([Table S3](#app1-cancers-11-00606){ref-type=\"app\"}).\n\n2.4. Postoperative Therapy and Follow Up {#sec2dot4-cancers-11-00606}\n----------------------------------------\n\nAfter surgery, adjuvant therapy was administered on a risk-adopted basis, according to pathological findings, patient characteristics and comorbidities. Patient follow-up was initially performed every 3--4 months for the first two years, every 6 months during years 3 and 4, and then annually. The surveillance protocol included physical examination, serum CA-19.9 level and CT-scan.\n\n2.5. Toxicity {#sec2dot5-cancers-11-00606}\n-------------\n\nToxicity during IPCT and CRT was evaluated and graded according to the National Cancer Institute Common Terminology for Adverse Events version 4.03 (NCI CTAE) scale.\n\n2.6. Statistical Analysis {#sec2dot6-cancers-11-00606}\n-------------------------\n\nFor our first aim, descriptive and comparative statistical analyses were performed using SPSS statistical software (IBM SPSS Statistics, version 20, for Windows, Chicago Illinois, USA). Overall survival (OS) was defined as the time elapsed from diagnosis until death (all causes) or last contact when still alive. Progression-free survival (PFS) was calculated from the date of diagnosis to the date of progression (local and/or distant), death (all causes) or last contact when not relapsed. Overall survival and progression-free survival were determined by the Kaplan-Meier method and log-rank test. Two versions of the log-rank test were used (the Mantel-Cox and the Breslow method). All statistical tests were conducted at a two-sided significance level of 0.05.\n\nThis study was approved as an observational post-authorized study for medicines for human use, according to the Ethics and Clinical Investigation Committee of Navarra and the Drugs for Human Use Department of the Spanish Agency for Drugs and Health Products (AEMPS) (ethic code: CUN-QUI-2016-01, 23 March 2016).\n\n2.7. Machine-Learning Algorithms {#sec2dot7-cancers-11-00606}\n--------------------------------\n\nFor our second aim, different machine-learning techniques were used to perform a predictive population model, including: Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM) and K-Nearest Neighbors Algorithm (KNN). A detailed description of the different techniques used is provided on [Appendix A](#app2-cancers-11-00606){ref-type=\"app\"}.\n\nThe goal of these algorithms is to provide an accurate prognostic information to resected PC patients treated with preoperative therapy. In order to avoid the learning process focusing excessively on the particular characteristics of our training data collected (overfitting), among approximately 140 clinical, pathological and analytical features from each patient, only those that were considered most influential in the individual risk of relapse after surgery were taken into account. A univariant exploratory analysis was performed with each of these initial features, and only 8 were finally selected. These selected variables were used for the training of each of the different machine-learning algorithms.\n\nAll the prediction models, except Random Forest, were validated using the 5-fold cross-validation technique, where the sample is divided into k sub-sets, in this case 5. This division is randomly performed, but always keeping the proportion of patients of each class in each of the subgroups. Once the sample is divided into subsets, k-1 subsets are taken as the training set and the remaining subgroup as the test set, in order to validate the algorithm. This process is repeated k times, allowing all possible combinations within the subsets. The result is the arithmetic mean obtained from the k repetitions. Random Forest was validated using a technique named bagging, which consists of bootstrapping the data and training each tree with one subset (bag). Then, each tree is validated with the instances out of its bag (OOB).\n\nAn external validation of the model was planned with a cohort of patients (20% of the global cohort) with potentially curable PC resected after a neoadjuvant approach at the Memorial Sloan Kettering Cancer Center (New York, NY, USA) between December 2008 and April 2016.\n\n3. Results {#sec3-cancers-11-00606}\n==========\n\n3.1. Clinical Data {#sec3dot1-cancers-11-00606}\n------------------\n\n### 3.1.1. Patients Characteristics {#sec3dot1dot1-cancers-11-00606}\n\nBaseline characteristics of the 40 resectable PC patients included in the retrospective analysis are summarized in [Table 1](#cancers-11-00606-t001){ref-type=\"table\"}. The median age at diagnosis was 63 (range 35 to 82) and the male/female ratio was 23/17. ECOG 0, 1 and 2 was found in 5 (12.5%), 33 (82.5%) and 2 (5%) patients, respectively. Most tumors were located in the head-isthmus of the pancreas (80%). According to EUS findings, T-stage was T2 in 3 patients (7.5%) and T3 in 37 patients (92.5%). Twenty-one patients (52.5%) were EUS-N0. Twenty-one of 40 patients (52.5%) required biliary stenting before receiving preoperative therapy.\n\n### 3.1.2. Multimodality Therapy Completion {#sec3dot1dot2-cancers-11-00606}\n\n[Figure 1](#cancers-11-00606-f001){ref-type=\"fig\"} shows the patients flowchart through the therapeutic algorithm.\n\nInduction polychemotherapy included mFOLFOXIRI (*n* = 14; 35%), GEMOXEL (*n* = 21, 52.5%) and GEMOX (*n* = 5, 12.5%). The median number of cycles administered was 4 (range 2 to 9). Three patients (7.5%) did not receive preoperative radiation therapy and were scheduled to surgery after IPCT due to recurrent episodes of cholangitis (one patient), necrotizing pancreatitis after 3 cycles of IPCT (one patient), and a suspicious resectable liver node detected in the post-IPCT re-staging (one patient). An R0 resection was achieved in all of them, including the patient with a new liver node that turned to be a gallbladder adenomyosis.\n\nThirty-seven patients (92.5%) received CRT after IPCT, 54% of them with 3D-RT and 46% with an IMRT technique. Median treatment length was 32 days (range 10 to 45).\n\nFour patients (10%) could not be operated on after CRT. Three of them had a systemic relapse at the time of preoperative re-staging (development of liver metastases and peritoneal carcinomatosis in 2 and 1 patient, respectively). The remaining patient died 16 days after CRT due to a massive hematemesis.\n\nOverall, 36 patients (90%) underwent surgery: 33 patients after receiving the complete neoadjuvant schedule (IPCT + CRT), and 3 patients after being treated only with IPCT, as detailed above. The time frame from last neoadjuvant therapy administered (IPCT + CRT, or IPCT alone) to surgery ranged from 3.5 to 12.8 weeks.\n\nAfter surgery, 9 patients (22.5%) received adjuvant treatment, most of them (77.8%) with the same protocol used in the neoadjuvant scenario. The median number of adjuvant cycles was 3 (range 1 to 4). One patient received adjuvant chemoradiation, because it had been omitted in the neoadjuvant setting. In this case, the patological report confirmed an R0 resection (minimal margin of 6 mm) and a ypT3N1 pancreatic adenocarcinoma with perineural invasion.\n\n### 3.1.3. Surgical Outcome and Pathological Results {#sec3dot1dot3-cancers-11-00606}\n\nAmong the 36 patients who underwent surgery, 29 (80.6%) had a cephalic duodenopancreatectomy and 7 (19.4%) had a distal pancreatectomy. On an intent-to-treat basis, a R0 resection was achieved in 90% of the patients. CAP grade 0 was achieved in 5 patients (13.9%). Vascular and perineural invasion were observed in 3 (8.3%) and 10 (27.8%) patients, respectively. The median number of resected lymph nodes was 11 (range 2 to 22). Five patients (13.9%) were ypN+ (median number of affected nodes was 1). Grade A, B, and C nodal response was observed in 32 (88.9%), 1 (2.8%) and 3 (8.3%) patients. One patient had nodal involvement due to locorregional invasion, and was not classified according to a modified Miller & Payne lymph node grading system. No grade D response (complete nodal pathological response) was described in the pathological review of the resected specimens.\n\n### 3.1.4. Toxicity Profile {#sec3dot1dot4-cancers-11-00606}\n\nToxicities associated with IPCT and CRT are described in [Table 2](#cancers-11-00606-t002){ref-type=\"table\"}.\n\nDuring IPCT, all patients experienced at least one grade 1 or 2 adverse event. Sixteen patients (40%) had one or more grade 3--4 toxicities. Grade 3--4 toxicities included: neutropenia (10 patients, 25%), leukopenia (5 patients, 12.5%), cholangitis (5 patients, 12.5%), diarrhea (3 patients, 7.5%), asthenia (2 patients, 5%) and gastritis (1 patient, 2.5%). Up to 50% of grade 3--4 neutropenias were febrile neutropenias. Granulocyte colony-stimulating factors were used in 7 patients (17.5%). A dose reduction was required in 10 patients (25%), 9 of them in the group treated with mFOLFOXIRI and 1 in the Gemcitabine-based group. Ten patients (25%) had a treatment delay, 8 of them in the group treated with mFOLFOXIRI and 2 in the group treated with Gemcitabine-based chemotherapy. Thirteen patients (32.5%) required at least one hospital admission due to cholangitis (*n* = 5; 12.5%), diarrhea (*n* = 5; 12.5%: grade 2: *n* = 2; grade 3: *n* = 3)**,** febrile neutropenia (*n* = 4; 10%: grade 3: *n* = 3; grade 4: *n* = 1), atrial fibrillation (*n* = 2; 5%), pulmonary embolism (*n* = 1; 2.5%), perianal abscess (*n* = 1; 2.5%), and necrotizing pancreatitis (*n* = 1; 2.5%).\n\nGrade 3-4 leukopenia, neutropenia and febrile neutropenia were more common in the group of patients treated with mFOLFOXIRI (35.7% vs. 0%; 57.1% vs. 7.7%; and 28.6% vs. 7.7%, respectively), whereas cholangitis was more frequent in the Gemcitabine-based group (15.4% vs. 7.1%). Cholangitis was only observed in patients with an endobiliary stent (23.8% vs. 0%), with a ratio of plastic vs. metallic stent in this group of patients of 3:2.\n\nOverall, 11 patients (29.7%) had any grade 3 toxicity during CRT, including: cholangitis (5 patients, 13.5%), thrombocytopenia (4 patients, 10.8%) and asthenia (2 patients, 5.4%). No grade 4 toxicity associated to CRT was reported. Six patients (16.2%) required at least one hospital admission due to cholangitis (*n* = 5; 13.5%), grade 1 fever (*n* = 1; 2.7%), grade 2 nausea (*n* = 1; 2.7%), and grade 3 asthenia (*n* = 1; 2.7%). One patient required CRT discontinuation due to cholangitis. Besides, 6 patients (16.2%) required concomitant capecitabine dose reduction.\n\nThere were no perioperative or in-hospital deaths related to the surgical procedure. Eight patients (22.2%) developed one or more early postoperative complications, including: paralytic ileus (*n* = 4; 11%), infected pancreatic fistula (*n* = 2; 5.6%), respiratory distress syndrome (*n* = 1; 2.8%), gastro-jejunal anastomosis bleeding (*n* = 1; 2.8%), hemoperitoneum secondary to pancreatic-duodenal artery bleeding (*n* = 1; 2.8%), left renal fossa hematoma (*n* = 1; 2.8%) and abdominal sepsis (*n* = 1; 2.8%). In the last 3 cases, an urgent exploratory laparotomy was needed, with a later ICU observation period. The median length of hospitalization was 8.36 days (range 4--35). Ten patients (27.8%) required a red blood cell transfusion.\n\nNine patients (25%) had late surgical complications, including: lymphocele (*n* = 2; 5.6%), peripancreatic collection, (*n* = 2; 5.6%), perihepatic collection (*n* = 1; 2.8%), gastric erosion bleeding (*n* = 1; 2.8%), splenectomy surgical site hematoma (*n* = 1; 2.8%), paralytic ileus (*n* = 1; 2.8%), inflammatory-infectious process (*n* = 1; 2.8%), antrum-pyloric stenosis, that required a surgical derivation (*n* = 1; 2.8%), and pancreatectomy site pseudo-cyst (*n* = 1; 2.8%). All these complications were uneventfully managed.\n\n### 3.1.5. Patients Long-Term Outcome {#sec3dot1dot5-cancers-11-00606}\n\nAfter a median follow-up of 33.5 months (range 3 to 133 months), median progression-free survival (PFS) was 18 months and median overall survival (OS) was 39 months ([Figure 2](#cancers-11-00606-f002){ref-type=\"fig\"}). The 1, 2, 3 and 5-year actuarial PFS were 71.3%, 46.9%, 43.8% and 32.3% respectively. The 1, 2, 3 and 5-year actuarial OS were 89.9%, 71.4%, 51.5% and 34.8% respectively. Median PFS (37 months vs. 18 months; *p* = 0.026) and OS (47 months vs. 8 months; *p* = 0.003) were significantly longer in those patients able to complete the whole therapeutic program (IPCT, CRT and surgery), compared to those who did not receive CRT or surgery. The discrepancy between OS and PFS in this last group of patients may be explained by the reported death due to massive hematemesis before surgery, without evidence of relapse.\n\nSurgical margins status and the presence of vascular invasion among ypN0 patients significantly correlated with survival outcomes. Among those patients who completed the whole therapeutic program, the median PFS was 37 months for those receiving FOLFOXIRI compared to 17 months for those receiving gemcitabine-based IPCT, with a 3-year PFS of 62.3% and 45.5%, respectively.\n\nTwenty-four patients (60%) have relapsed. The pattern of relapse was distant in 18 cases (75%), locoregional in 1 case (4.2%), and both, local and distant in 5 cases (20.8%). Liver was the most common site for distant progression (27.5%). Twenty of the relapsed patients (83.3%) underwent a second-line treatment. Among them, 17 died due to disease progression, 1 died due to cardiological comorbidity and 2 patients were alive at the end of the follow-up.\n\n3.2. Predictive Population Model {#sec3dot2-cancers-11-00606}\n--------------------------------\n\n### 3.2.1. Model Development {#sec3dot2dot1-cancers-11-00606}\n\nBaseline characteristics of the 45 operated patients (resectable/borderline resectable; 37/8) included in the machine-learning algorithms are summarized in [Table 3](#cancers-11-00606-t003){ref-type=\"table\"}. Up to 42 (93.3%) of resected PC received multimodal neoadjuvant therapy (IPCT + CRT). Thirty-seven patients (82.2%) had a cephalic duodenopancreatectomy and 8 patients (17.8%) had a distal pancreatectomy. Only 28.9% of patients (*n* = 13) received adjuvant treatment. Thirty patients (66.7%) relapsed, most of them at distant sites (*n* = 21; 72.4%).\n\nSeveral known prognostic features related to the risk of relapse were collected and are summarized in [Table 4](#cancers-11-00606-t004){ref-type=\"table\"}. By univariate analysis, the most relevant for the model were selected, including: ECOG, the type of IPCT employed, the use of granulocyte-colony-stimulating factors, the type of surgery, the number of resected lymph nodes, the modified LNR, the presence of perineural invasion and the surgical margins status.\n\nEach of the different machine-learning algorithms previously described were trained with the 8 selected variables ([Table 5](#cancers-11-00606-t005){ref-type=\"table\"}). After applying a 5-fold cross-validation technique, it was concluded that Logistic Regression was the best predictive algorithm. According to the three intermediate steps (step Akaike criterion (AIC), non-linear trends and interactions), the appropriate components that should be retained in the model were: the use of granulocyte colony-stimulating factors (yes/no), the number of resected lymph nodes, the presence of perineural invasion (yes/no) and the surgical margins status (R0/R1). Taking into account these features, the model predicts the probability of relapse at 2 years after surgery for an individual patient with an accuracy of 0.71 (95% IC 0.56--0.84, *p* = 0.005), a sensitivity of 0.70, a specificity of 0.73 and a mean area under the curve (AUC) of 0.75 ([Table 5](#cancers-11-00606-t005){ref-type=\"table\"} and [Figure 3](#cancers-11-00606-f003){ref-type=\"fig\"}).\n\n### 3.2.2. External Validation of the Model {#sec3dot2dot2-cancers-11-00606}\n\nAn external validation of the model was performed with a cohort of PC patients from a USA institution.\n\nThe external cohort comprised 49 PC patients, but only 33 of them were considered resectable or borderline-resectable at diagnosis. Among them, in only 27 patients were the required components of the model available. After a descriptive analysis of the cohort, a higher median number of resected lymph nodes was reported in the validation cohort (mean 24.37; median 21; minimum 9; maximum 44) compared to the training dataset (mean 10.93; median 9; minimum 2; maximum 27). In order to minimize differences related to the extent of the lymphadenectomy, only 11 patients from the validation dataset with a number of resected lymph nodes up to the 90th percentile of the training dataset were included in the validation cohort. The descriptive analysis of the four components in the validation dataset is summarized in [Table 6](#cancers-11-00606-t006){ref-type=\"table\"}. Among the 11 patients from the validation cohort, the model-predicted outcome matched with the observed outcome in 7 patients. The predictive accuracy of the model at the individual level was 64%.\n\n4. Discussion {#sec4-cancers-11-00606}\n=============\n\nSeveral studies have shown the feasibility and activity of a preoperative treatment in potentially curable PC, \\[[@B14-cancers-11-00606],[@B33-cancers-11-00606],[@B34-cancers-11-00606],[@B35-cancers-11-00606],[@B36-cancers-11-00606],[@B37-cancers-11-00606],[@B38-cancers-11-00606],[@B39-cancers-11-00606],[@B40-cancers-11-00606],[@B41-cancers-11-00606],[@B42-cancers-11-00606],[@B43-cancers-11-00606],[@B44-cancers-11-00606]\\] ([Table 7](#cancers-11-00606-t007){ref-type=\"table\"}), with R0 resection rates ranging from 25% \\[[@B36-cancers-11-00606]\\] to 76% \\[[@B14-cancers-11-00606]\\], and median overall survival times between 9.4 \\[[@B40-cancers-11-00606]\\] and 34 \\[[@B14-cancers-11-00606]\\] months, with no increase in the rates of postoperative complications \\[[@B45-cancers-11-00606]\\]. Recently a propensity score-matched analysis has suggested a survival benefit in favor of neoadjuvant therapy compared to upfront surgery alone (HR 0.72) or followed by adjuvant treatment (HR 0.83) \\[[@B46-cancers-11-00606]\\], although no randomized phase III trials have been reported to date.\n\nIn the present study we combine the use of intensive IPCT and preoperative CRT in a subset of patients with potentially curable PC. In the metastatic setting, triplet regimens have usually correlated with improved efficacy, \\[[@B47-cancers-11-00606],[@B48-cancers-11-00606],[@B49-cancers-11-00606],[@B50-cancers-11-00606]\\] and this has encouraged its application in locally advanced disease. A preliminary pilot trial with neoadjuvant mFOLFOXIRI in resectable PC has also shown that this regimen is feasible and tolerable in this setting \\[[@B14-cancers-11-00606]\\]. The preoperative use of radiotherapy in resectable PC aims at improving local control \\[[@B43-cancers-11-00606],[@B44-cancers-11-00606]\\], R0 resection rates \\[[@B51-cancers-11-00606]\\] and reducing the incidence of postsurgical complications, due to the induction of fibrosis in pancreatic tissue and surgical bed. Indeed, preoperative radiotherapy improves suitability of pancreatic tissue for anastomosis and reduces the risk of developing a pancreatic fistula or anastomotic leak \\[[@B43-cancers-11-00606],[@B52-cancers-11-00606],[@B53-cancers-11-00606]\\]. Preliminary data have shown that preoperative chemoradiotherapy significantly improves outcomes in resectable PC compared to immediate surgery \\[[@B13-cancers-11-00606]\\].\n\nWith a R0 resection rate of 90% and a median overall survival of 39 months, our results overlap with those achieved with similar strategies \\[[@B14-cancers-11-00606],[@B34-cancers-11-00606],[@B44-cancers-11-00606]\\] ([Table 6](#cancers-11-00606-t006){ref-type=\"table\"}). Our data seem especially noteworthy in the subgroup of resected patients who completed the whole neoadjuvant program, with a median PFS of 37 months and a median OS of 47 months. Local recurrence as the only pattern of relapse was identified in 4.2%, which is a low result for outcomes reported with preoperative strategies.\n\nThe use of an intensified neoadjuvant approach is not without cost. In our series, up to 40% of patients developed grade 3--4 toxicities due to IPCT, 25% required dose reductions or treatment delays (most of them in the mFOLFOXIRI group), and the hospitalization rate was 32.5%. An additional 30% of patients had grade 3 adverse events related to CRT. Our cholangitis rate (12.5% with IPCT and 13.5% with CRT) is in the range of that observed in other studies \\[[@B42-cancers-11-00606]\\], and clearly related to biliary stents, as described previously \\[[@B54-cancers-11-00606],[@B55-cancers-11-00606]\\].\n\nConsidering the high rate of distant relapse in our series, and in an attempt to identify the subset of operated patients with a higher likelihood of relapse, we aim to build up a predictive model based on supervised machine-learning algorithms. The prognostic impact of the four features included in the final prediction model has been previously described on an individual basis. \\[[@B56-cancers-11-00606],[@B57-cancers-11-00606],[@B58-cancers-11-00606],[@B59-cancers-11-00606],[@B60-cancers-11-00606],[@B61-cancers-11-00606]\\] Of note is the role of granulocyte colony-stimulating factors, whose impact on cancer survival has been previously reported in solid tumors. \\[[@B62-cancers-11-00606]\\] The importance of increasing chemotherapy relative dose intensity has also been described in PC patients, with preliminary encouraging results. \\[[@B63-cancers-11-00606],[@B64-cancers-11-00606]\\] With an accuracy above 60% for a 2-year relapse-risk after surgery, this model may be a useful tool, with clinical practice decision-making implications. Among at high risk patients, more intensive surveillance, the use of adjuvant treatment, or even the inclusion of these patients into clinical trials may be considered.\n\nOur study has several limitations, including: its retrospective nature (with the inherent biases related to this type of studies), the small sample size, the heterogeneity in the IPCT and radiotherapy techniques applied, and the fact of being a single institutional experience. Additionally, the limited number of patients included in the dataset of our second aim reduce the accuracy for quantifying interpatient variability effects. In addition, no pharmacodynamic information such as tumor size was included in the prediction model, due to the difficult examination of tumor downstaging by radiologic procedures because of intense inflammation and fibrosis induced by neoadjuvant treatment.\n\n5. Conclusions {#sec5-cancers-11-00606}\n==============\n\nIn summary, the use of an intensified preoperative program with IPCT followed by CRT offers encouraging results in terms of R0 resection rates and survival times at an expense of manageable toxicity. Implementation of machine-learning algorithms may help to identify at-risk patients and tailor adjuvant strategies.\n\nThe authors thank all patients who participated in the study and their families.\n\nThe following are available online at <https://www.mdpi.com/2072-6694/11/5/606/s1>, Table S1: Resectability criteria by National Comprehensive Cancer Network, Table S2: Tumor regression grade according to the College of American Pathologists (CAP) grading system, Table S3: Grade of nodal treatment response scale adapted from the Miller & Payne grading system.\n\n###### \n\nClick here for additional data file.\n\nConceptualization, methodology, and investigation, P.S.E. and J.R.R. Data curation P.S.E., K.H.Y., V.B., O.M., L.D.M.S and J.R.R. Formal analysis, E.O.-I. and P.S.E. Validation, K.H.Y. All authors (P.S.E., E.O.-I., K.H.Y., V.B., L.A.M., O.C., A.C.O., M.P.-S., L.D.M.S., F.R.S., B.L.L., Y.I.B., J.C.S.I., A.B.B., F.P. and J.R.R.) were responsible for writing and review of the manuscript.\n\nThis research received no external funding.\n\nThe authors declare no conflict of interest.\n\n1\\. Logistic Regression:\n\nIt predicts the presence or absence of a characteristic or result according to the values of a set of predictors. It is similar to a linear regression model but is adapted for models in which the dependent variable is dichotomous. The logistic regression coefficients can be used to estimate the odds ratio of each variable independent of the model. To obtain the best classifier we perform three steps. First, we apply Step AIC criterion to obtain the best model with the optimal combination of features. Second, we analyse if any of the selected features present non-linear tendencies by a visual inspection. Last, we consider interactions between the features and we implement Step AIC criterion again to obtain the best predictor.\n\n2\\. Decision Tree:\n\nIt is a map of the possible outcomes of a series of related decisions. It starts with a single node and then branches into possible outcomes. Each of those results creates additional nodes, which branch out into other possibilities. This gives it a shape similar to that of a tree. When a new instance is to be classified, its characteristics are analysed according to the decision tree structure, which indicates the outcome to that instance with the probability to belonging to that category.\n\n3\\. Random Forest:\n\nIt is a decision tree ensemble that combines the knowledge generated by a collection of individual tree making use of randomness in the process. In order to classify a new patient, each tree provides a classification, which its corresponding probability, and the category with the highest probability is chosen.\n\n4\\. Support Vector Machine (SVM):\n\nThis algorithm learns the decision boundary of two different classes of entry points. When the points are non-linearly separable, the algorithm transforms the domain to a higher dimensional space where they can be separated by a linear hyperplane. Moreover, the decision boundary is optimized so that the hyperplane is equidistant to the nearest instances of each category.\n\n5\\. K-Nearest Neighbours Algorithm (KNN):\n\nThis algorithm consists in predicting the category of an instance based on the categories of the closets instances. The proximity of two instances is measured by the similarity of each of their features using for example the Euclidean distance.\n\n![Patients flowchart through the therapeutic algorithm. IPCT: Induction polychemotherapy; CRT: Chemoradiation.](cancers-11-00606-g001){#cancers-11-00606-f001}\n\n![Kaplan-Meier estimates of progression-free survival (**a**) and overall survival (**b**) among patient with resectable PC treated with the neoadyuvant approach.](cancers-11-00606-g002){#cancers-11-00606-f002}\n\n![Receiver Operating Characteristic (ROC) curves of Logistic Regression predictive population model for each of the 5-fold cross validation. Each ROC curve is represented in a different color (black, red, green, blue and purple).](cancers-11-00606-g003){#cancers-11-00606-f003}\n\ncancers-11-00606-t001_Table 1\n\n###### \n\nBaseline characteristics of patients with resectable PC treated with a multimodality neoadjuvant approach, *n* = 40.\n\n  Variables                                *n* (%)\n  ---------------------------------------- -----------\n  **Age-years**                            \n  Median                                   63\n  Range                                    35--82\n  **Gender**                               \n  Male                                     23 (57.5)\n  Female                                   17 (42.5)\n  **ECOG**                                 \n  0                                        5 (12.5)\n  1                                        33 (82.5)\n  2                                        2 (5)\n  **Location**                             \n  Head-Isthmus                             32 (80)\n  Body-Tail                                8 (20)\n  **Baseline EUS-T stage**                 \n  T1-T2                                    3 (7.5)\n  T3                                       37 (92.5)\n  **Baseline EUS-N stage**                 \n  N0                                       21 (52.5)\n  N+                                       11 (27.5)\n  Nx                                       8 (20)\n  **Biliary stent**                        \n  No                                       19 (47.5)\n  Yes                                      21 (52.5)\n  **Neoadjuvant chemotherapy regimen**     \n  Gemcitabine-based                        26 (65)\n  mFOLFOXIRI                               14 (35)\n  **Neoadjuvant radiotherapy technique**   \n  3D-RT                                    20 (54.1)\n  IMRT                                     17 (45.9)\n\ncancers-11-00606-t002_Table 2\n\n###### \n\nAdverse events during neoadjuvant regimen.\n\n  Adverse Event           IPCT ^1^ (*n* = 40)   CRT ^2^ (*n* = 37)                                    \n  ----------------------- --------------------- -------------------- --------- ----------- ---------- ----\n  **Hematological**                                                                                   \n  Anemia                  37 (92.5)             \\-                   \\-        33 (89.2)   \\-         \\-\n  Leukopenia              19 (47.5)             2 (5)                3 (7.5)   29 (78.4)   \\-         \\-\n  Neutropenia             7 (17.5)              9 (22.5)             1 (2.5)   8 (21.6)    \\-         \\-\n  Febrile neutropenia     1 (2.5)               4 (10)               1 (2.5)   1 (2.7)     \\-         \\-\n  Thrombocytopenia        26 (65)               \\-                   \\-        27 (73)     4 (10.8)   \\-\n  **Non-hematological**                                                                               \n  Nausea/Vomiting         12 (30)               \\-                   \\-        15 (40.5)   \\-         \\-\n  Anorexia                18 (45)               \\-                   \\-        20 (54)     \\-         \\-\n  Diarrhea                15 (37.5)             3 (7.5)              \\-        6 (16.2)    \\-         \\-\n  Gastritis               7 (17.5)              1 (2.5)              \\-        13 (35.1)   \\-         \\-\n  Mucositis               10 (25)               \\-                   \\-        2 (5.4)     \\-         \\-\n  Asthenia                32 (80)               2 (5)                \\-        21 (56.7)   2 (5.4)    \\-\n  Peripheral neuropathy   22 (55)               \\-                   \\-        6 (16.2)    \\-         \\-\n  Hand-foot syndrome      1 (2.5)               \\-                   \\-        1 (2.7)     \\-         \\-\n  Cholangitis             \\-                    5 (12.5)             \\-        \\-          5 (13.5)   \\-\n\nAbbreviations: ^1^ IPCT (induction polychemotherapy); ^2^ CRT (Chemoradiotherapy).\n\ncancers-11-00606-t003_Table 3\n\n###### \n\nBaseline characteristics of patients included in 2-year relapse risk prediction model, *n* = 45.\n\n  Variables                        *n* (%)\n  -------------------------------- -----------\n  Age-years                        \n  Median                           64\n  Range                            44-80\n  Gender                           \n  Male                             23 (51.1)\n  Female                           22 (48.9)\n  ECOG                             \n  0                                8 (17.8)\n  1                                35 (77.8)\n  2                                2 (4.4)\n  Location                         \n  Head-Isthmus                     34 (75.6)\n  Body-Tail                        10 (22.2)\n  Multifocal                       1 (2.2)\n  Baseline EUS-T stage             \n  T1-T2                            4 (8.9)\n  T3                               40 (88.9)\n  T4                               1 (2.2)\n  Baseline EUS-N stage             \n  N0                               27 (60)\n  N+                               11 (24.4)\n  Nx                               7 (15.6)\n  Resectability                    \n  Resectable                       37 (82.2)\n  Borderline-resectable            8 (17.8)\n  Neoadjuvant approach             \n  IPCT + CRT                       42 (93.3)\n  IPCT                             3 (6.7)\n  Duration of IPCT-days            \n  Median                           53\n  Range                            40-125\n  Number of CRT session            \n  Median                           25\n  Range                            18-30\n  Type of surgery                  \n  Cephalic duodenopancreatectomy   37 (82.2)\n  Distal pancreatectomy            8 (17.8)\n  Adjuvant treatment               \n  Yes                              13 (28.9)\n  No                               32 (71.1)\n  Relapse                          \n  Yes                              30 (66.7)\n  No                               15 (33.3)\n  Relapse at 2 years               \n  Yes                              22 (48.9)\n  No                               23 (51.1)\n  Type of relapse                  \n  Local                            2 (6.7)\n  Distant                          22 (73.3)\n  Local and distant                6 (20)\n\nECOG: Eastern Cooperative Oncology Group; EUS: Endoscopic ultrasound\n\ncancers-11-00606-t004_Table 4\n\n###### \n\nFeatures included in the preliminary analysis for the machine learning algorithms.\n\n  Variables                                               *n* (%)\n  ------------------------------------------------------- -----------\n  Gender                                                  \n  Male                                                    23 (51.1)\n  Female                                                  22 (48.9)\n  Age-years                                               \n  Min.                                                    44\n  Median                                                  64\n  Mean                                                    63\n  Max.                                                    80\n  Resectability                                           \n  Resectable                                              37 (82.2)\n  Borderline-resectable                                   8 (17.8)\n  ECOG                                                    \n  0                                                       8 (17.8)\n  1                                                       35 (77.8)\n  2                                                       2 (4.4)\n  Neoadjuvant chemotherapy regimen                        \n  mFOLFOXIRI                                              18 (40)\n  Gemcitabine-based                                       27 (60)\n  Granulocyte colony-stimulating factors                  \n  No                                                      37 (82.2)\n  Yes                                                     8 (17.8)\n  Neoadjuvant radiotherapy technique                      \n  3D-RT                                                   18 (40)\n  IMRT                                                    21 (46.7)\n  Not reported                                            6 (13.3)\n  Type of surgery                                         \n  Cephalic duodenopancreatectomy                          37 (82.2)\n  Distal pancreatectomy                                   8 (17.8)\n  ypT                                                     \n  ypT0                                                    6 (13.3)\n  ypT1                                                    19 (42.2)\n  ypT2                                                    6 (13.3)\n  ypT3                                                    12 (26.7)\n  ypTx                                                    2 (4.4)\n  ypN                                                     \n  ypN0                                                    41 (91.1)\n  ypN1                                                    4 (8.9)\n  CAP                                                     \n  0                                                       6 (13.3)\n  1                                                       23 (51.1)\n  2                                                       10 (22.2)\n  3                                                       6 (13.3)\n  CAP Group                                               \n  CAP 0--1                                                29 (64.4)\n  CAP 2--3                                                16 (35.6)\n  Pathological complete response                          \n  No                                                      37 (82.2)\n  Yes                                                     8 (17.8)\n  Resected lymph nodes                                    \n  Min.                                                    2\n  Median                                                  9\n  Mean                                                    10.93\n  Max.                                                    27\n  Pathological lymph nodes                                \n  Min.                                                    0.00\n  Median                                                  0.00\n  Mean                                                    0.16\n  Max.                                                    3.00\n  Modified LNR: 1 + *n*\u00b0 pathological/1 + *n*\u00b0 resected   \n  Min.                                                    0.036\n  Median                                                  0.100\n  Mean                                                    0.125\n  Max.                                                    0.333\n  Miller & Payne nodal response                           \n  A                                                       41 (91.1)\n  B                                                       1 (2.2)\n  C                                                       3 (6.7)\n  D                                                       0 (0)\n  Vascular invasion                                       \n  No                                                      40 (88.9)\n  Yes                                                     5 (11.1%)\n  Perineural invasion                                     \n  No                                                      31 (68.9)\n  Yes                                                     14 (31.1)\n  Surgical margins                                        \n  R0 (Not involved, \\>1 mm)                               43 (95.6)\n  R1 (Involved, \\<1 mm)                                   2 (4.4)\n  SCORE ^1^                                               \n  1                                                       6 (13.3)\n  2                                                       25 (55.6)\n  3                                                       14 (31.1)\n  Dose intensity \\> 80% ^2^                               \n  No                                                      12 (27.3)\n  Yes                                                     32 (72.7)\n  Progression disease at 2 years                          \n  No                                                      23 (51.1)\n  Yes                                                     22 (48.9)\n\n^1^ SCORE include 6 variables (CAP = 0, ypT0-N0, A or D nodal response according to Miller & Payne grading system, absence of vascular invasion, absence of perineural invasion, and R0 resection). SCORE 1 include 6/6 variables, SCORE 2 include 4--5/6 variables, and SCORE 3 include \u22643/6 variables. ^2^ Data not collected in one patient, who received some of neoadjuvant chemotherapy cycles in a different institution.\n\ncancers-11-00606-t005_Table 5\n\n###### \n\nResults from different machine-learning algorithms.\n\n  Variable          Logistic Regression   Decision Tree   Random Forest   Support Vector Machine   K-Nearest Neighbours\n  ----------------- --------------------- --------------- --------------- ------------------------ ----------------------\n  **Accuracy**      0.71                  0.60            0.67            0.60                     0.58\n  **Sensitivity**   0.70                  0.83            0.74            0.65                     0.65\n  **Specificity**   0.73                  0.36            0.59            0.55                     0.50\n  **PPV ^1^**       0.73                  0.58            0.65            0.60                     0.58\n  **NPV ^2^**       0.70                  0.67            0.68            0.60                     0.58\n  **AUC**           0.75                  0.61            0.67            0.61                     0.58\n\nAbbreviations: ^1^ PPV (Positive predictive value); ^2^ NPV (Negative predictive value).\n\ncancers-11-00606-t006_Table 6\n\n###### \n\nBaseline features of the validation dataset that were included in the predictive model.\n\n  Variables                                *n* (%)\n  ---------------------------------------- ----------\n  Granulocyte colony-stimulating factors   \n  No                                       4 (36.4)\n  Yes                                      7 (63.6)\n  Perineural invasion                      \n  No                                       2 (18.2)\n  Yes                                      9 (81.8)\n  Surgical margins                         \n  R0 (Not involved)                        8 (72.7)\n  R1 (Involved)                            3 (27.3)\n  Resected lymph nodes                     \n  Min.                                     9\n  Median                                   17\n  Mean                                     15.55\n  Max.                                     19\n  Progression disease at 2 years           \n  No                                       4 (36.4)\n  Yes                                      7 (63.6)\n\ncancers-11-00606-t007_Table 7\n\n###### \n\nNeoadjuvant studies in resectable pancreatic cancer.\n\n  Study                                             NA        ChT                                           RT      CRT                                    *n*   Stage   Resection     R0 Rate G/Res         Median OS (Months)   Survival Rate (%)                                                   \n  ------------------------------------------------- --------- --------------------------------------------- ------- -------------------------------------- ----- ------- ------------- --------------------- -------------------- ------------------- ------ --------------------- ------------------ ------------------\n  Ishikawa (1994) \\[[@B33-cancers-11-00606]\\]       RT        \\-                                            50 Gy   \\-                                     54    R       74%           \\-                    15                   \\-                  9      30 (2y), 22 (5y)      28 (3y), 22 (5y)   17 (1y), 0 (2y)\n  Evans (1992) \\[[@B37-cancers-11-00606]\\]          CRT       \\-                                            \\-      5-FU, 50.4Gy                           28    R       61%           50%/82%               \\-                   \\-                  \\-     \\-                    \\-                 \\-\n  Evans (2008) \\[[@B38-cancers-11-00606]\\]          CRT       \\-                                            \\-      Gemcitabine 30 Gy                      86    R       74%           66%/95%               22.7                 34                  7      27 (5y)               36 (5y)            0 (5y)\n  BR                                                                                                                                                                                                                                                                                                  \n  Turrini (2009) \\[[@B39-cancers-11-00606]\\]        CRT       \\-                                            \\-      5-FU Cisplatin 45 Gy                   102   R       61%           56%/92%               17                   23                  11     10 (5y)               18 (5y)            0 (5y)\n  Le Scodan (2009) \\[[@B40-cancers-11-00606]\\]      CRT       \\-                                            \\-      5-FU Cisplatin 50 Gy                   41    R       63%           51%/80.7%             9.4                  11.7                5.7    41 (1y), 20 (2y)      48 (1y), 32 (2y)   40 (1y), 0 (2y)\n  Kim (2013) \\[[@B41-cancers-11-00606]\\]            CRT       \\-                                            \\-      Gemcitabine                            68    R       63%           53%/84%               18.2                 27.1                10.9   62 (1y), 44 (2y)      82 (1y), 62 (2y)   33 (1y), 17 (2y)\n  Oxaliplatin                                       BR                                                                                                                                                                                                                                                \n  30Gy                                              I                                                                                                                                                                                                                                                 \n  Golcher (2015) \\[[@B42-cancers-11-00606]\\]        CRT       \\-                                            \\-      Gemcitabine Cisplatin 55.8 to 50.4Gy   66    R       58%           51%/89%               17.4                 25                  \\-     39 (2y), 12 (3y)      \\-                 \\-\n  BR                                                                                                                                                                                                                                                                                                  \n  Casadei (2015) \\[[@B43-cancers-11-00606]\\]        ChT-CRT   Gemcitabine                                   \\-      Gemcitabine 54 Gy                      38    R       61.1%         38.9%/64%             22.4                 \\-                  \\-     \\-                    \\-                 \\-\n  Varadhachary (2008) \\[[@B44-cancers-11-00606]\\]   ChT-CRT   Gemcitabine Cisplatin                         \\-      Gemcitabine 30 Gy                      90    R       58%           55%/96%               17.4                 31                  10.5   37 (2y), 19 (4y)      60 (2y), 36 (4y)   \\-\n  BR                                                                                                                                                                                                                                                                                                  \n  O'Reilly (2014) \\[[@B34-cancers-11-00606]\\]       ChT       Gemcitabine Oxaliplatin                       \\-      \\-                                     38    BR      71%           53%/74%               27.2                 NR                  15     63 (18m)              78 (18m)           25 (18 m)\n  Heinrich (2008) \\[[@B35-cancers-11-00606]\\]       ChT       Gemcitabine Cisplatin                         \\-      \\-                                     28    R       89%           71%/80%               26.5                 19.1                \\-     \\-                    \\-                 \\-\n  Palmer (2007) \\[[@B36-cancers-11-00606]\\]         ChT       Gemcitabine Cisplatin Vs. Gemcitabine alone   \\-      \\-                                     50    R       70% Vs. 38%   46%/75% Vs. 25%/75%   15.6 Vs. 9.9         28.4 (global)       \\-     62 (1y) Vs. 42 (1y)   77.8 (global)      \\-\n  De W Marsh (2017) \\[[@B14-cancers-11-00606]\\]     ChT       mFOLFIRINOX                                   \\-      \\-                                     21    R       81%           76%/94%               34                   35.5                10.1   80 (1y), 60(2y)       81 (1y), 71 (2y)   33 (1y), 0 (2y)\n\nAbbreviations: NA (Neoadjuvant Approach); ChT (Chemotherapy); RT (Radiotherapy); CRT (Chemoradiotherapy); ChT-CRT (Chemotherapy followed by Chemoradiotherapy); N (number of patients included in the study); R (resectable PC); BR (borderline-resectable PC); I (Irresectable PC); NR (Not Reached); G (Global); Res (Resected); NRes (Non-resected).\n"}
+{"text": "Introduction {#s0001}\n============\n\nAlzheimer\\'s disease (AD) is the most frequent case of age-related neurodegenerative dementia, characterized by progressive loss of memory and other cognitive functions[@CIT0001] ^,^ [@CIT0002]. AD is a heterogeneous disease, driven by the interaction between multiple deleterious factors. However, the exact mode of how these factors contribute to impair neuronal functions and neuronal survival still remains undetermined. One of the main markers of AD is the accumulation of \u03b2-amyloid plaques (A\u03b2) in nerve cells. In healthy brain, these aggregates of proteins are degraded and eliminated[@CIT0003]. However, in AD the aggregates accumulate to form insoluble plaques[@CIT0003]. Another characteristic is the presence of insoluble neurofibrillary filaments that is associated with tau protein (PTau)[@CIT0004]. In AD, however, PTau becomes hyperphosphorylated, denaturing and resulting in its dissociation of microtubules, followed by formation of neurofibrillary filaments that aggregate, acting as physical barriers to microtubules[@CIT0004]. In addition, the occurrence of glial cell neuroinflammation, synaptic loss, and specific neuronal death is common in AD[@CIT0005] and can be aggravated by oxidative stress[@CIT0006].\n\nThe knowledge of neurotransmitter disorders in AD has led to the approval of drugs with symptomatic effects[@CIT0007]. The cholinergic hypothesis of AD states that the degeneration of cholinergic neurons in basal forebrain nuclei causes disorders in the presynaptic cholinergic terminals in the hippocampus and neocortex, which are regions of extreme importance for memory disorders and other cognitive symptoms[@CIT0008]. Because of neurodegeneration, the activity of cholinergic neurons and levels of neurotransmitter ACh are reduced. One approach to improve cholinergic neurotransmission is to increase the availability of ACh by inhibition of acetylcholinesterase[@CIT0009].\n\nAcetyl (AChE) and butyrylcholinesterase (BuChE) inhibitors are the main drugs for the clinical treatment of AD in the initial to moderate stage[@CIT0010]. Galantamine and donepezil are selective inhibitors of AChE, whereas rivastigmine inhibits AChE and BuChE with similar affinities. Selective AChE inhibitors have demonstrated better therapeutic effects when compared to nonselective inhibitors[@CIT0011] since BuChE is also associated with drug metabolism and detoxification, lipoprotein metabolism and diseases[@CIT0012]. Thus, our objectives herein were the design, synthesis and pharmacological evaluation of novel 3-substituted-7-aminoalcoxy-coumarins as selective inhibitors of AChE and antioxidant, based on a previously described indanone series[@CIT0013].\n\nMaterials and methods {#s0002}\n=====================\n\nGeneral procedure for the synthesis of 2a--d {#s0003}\n--------------------------------------------\n\nIn a reactional borosilicate tube, 10--15\u2009mmol of dibromoalkanes (4--6 eq.) and 5\u2009mmol (2 eq.) of K~2~CO~3~ were solubilized in 2\u2009ml of acetone ([Scheme 2](#SCH0002){ref-type=\"scheme\"}). To this stirred suspension a solution of 2.5\u2009mmol of 7-hydroxycoumarin (**1**) in 8\u2009ml of acetone was added dropwise. Thereafter, the reactional tube was sealed and the reaction was kept at 60\u2009\u00b0C and stirred for 6--12\u2009h. After reaction completion, acetone was evaporated and the crude reaction partitioned with distilled water and ethyl acetate. The final slurry was precipitated in hexanes under ultrasound irradiation and filtered off.\n\nGeneral procedure for the synthesis of 3a--d {#s0004}\n--------------------------------------------\n\nTo a stirred solution of 1.7\u2009mmol of the respective *O*-alkyl coumarin derivative (**2a--d**), 5\u2009mmol (3 eq.) of sodium acetate in 8\u2009ml of glacial acetic acid and 2.1\u2009mmol (1.3 eq.) of Br~2~ were slowly added ([Scheme 2](#SCH0002){ref-type=\"scheme\"}). The reaction was stirred at room temperature for 2\u2009h. After reagent consumption, the reaction mixture was poured to a beaker containing crushed ice. The formed precipitate was filtered off under vacuum and purified by silica gel column chromatography (hexanes: dichloromethane mixture, 50--90% gradient elution).\n\nGeneral procedure for the synthesis of 4a--d {#s0005}\n--------------------------------------------\n\nIn a reactional vessel, 1.3\u2009mmol of the respective 3-bromo-7-(bromoalkoxy)coumarin derivatives (**3a--d**) and 3.9\u2009mmol (3 eq.) of piperidine were dissolved in 8\u2009ml CH~3~CN ([Scheme 2](#SCH0002){ref-type=\"scheme\"}). The reaction was kept under stirring at 60\u2009\u00b0C for 3--8\u2009h. Acetonitrile was evaporated in a rotary evaporator and the respective products purified by silica gel column chromatography (dichloromethane: methanol, 0--25% gradient elution).\n\nGeneral procedure for the synthesis of 5a--c and 6a--c {#s0006}\n------------------------------------------------------\n\nIn a reaction borosilicate tube, 0.14\u2009mmol of the corresponding derivative (**4a, 4b and 4d**), 0.20\u2009mmol (1.4 eq.) of appropriate phenylboronic acid and 0.42\u2009mmol (3 eq.) of K~2~CO~3~ were solubilized in 4\u2009ml of a solvent mixture (water: ethanol: toluene (2:1:1)) ([Scheme 2](#SCH0002){ref-type=\"scheme\"}). The reaction was degassed with N~2~ then 0.01\u2009mmol (7\u2009mol%) of Pd(PPh~3~)~4~ catalyst added. The reaction tubes were sealed and the mixtures were subjected to magnetic stirring and heating at 65\u2009\u00b0C for 3--5\u2009h. At the end of the reaction, the solvent mixture was evaporated in a rotary evaporator and the respective products purified by silica gel column chromatography (dichloromethane: methanol, 0--25% mixture gradient elution).\n\nCholinesterase inhibition and kinetics assays {#s0007}\n---------------------------------------------\n\nActivity of enzymes and inhibition kinetics were determined using a Bio-Rad iMark microplate reader based on a modification of the Ellman method.[@CIT0014] ^,^ [@CIT0015] Compounds were dissolved in DMSO. The assay solution which contained 60\u2009\u00b5L 5,5\u2032-Dithiobis(2-nitrobenzoic acid) (DTNB) at 1.1\u2009mM, 30\u2009\u00b5L AChE/BuChE at 0.20\u2009U/mL (initial concentration) and 150\u2009\u00b5L tested compound solution with different concentrations. Absorbance was then recorded at *\u03bb*\u2009=\u2009415\u2009nm. After 10-min incubation at 30\u2009\u00b0C, 24\u2009\u00b5L acetylthiocholine iodide/*S*-butyrylthiocholine iodide (at 2.75\u2009mM for activity inhibition assay and 2.75--0.44\u2009mM for kinetic study assay) were added and the absorbance recorded after a 10-min incubation (for activity inhibition assay) or after 0--20\u2009min incubation (for kinetic study assay) at 30\u2009\u00b0C.\n\nMolecular modeling {#s0008}\n------------------\n\nFor EeAChE (*Electrophorus electricus*), the PDB structure 1C2O was used; for EqBuChE (*Equus caballus*), a 3\u2009D homology model was necessarily built from a sequence available in the UniProtKB/Swiss-Prot (entry Q9N1N9) with the automated mode of the protein structure homology-modeling server, Swiss-Model[@CIT0016], using as template the human BuChE (PDB 4TPK)[@CIT0017]. Spartan'14 program \\[Wavefunction, Inc.\\] was utilized to construct and optimize the inhibitors with the PM6 method[@CIT0018]. The program GOLD 5.6 (CCDC Software Ltd., Cambridge, UK) was used to for the docking study with the GoldScore scoring function[@CIT0019].\n\nEvaluation of the antioxidant activity by the ferric reducing ability of plasma (FRAP) method {#s0009}\n---------------------------------------------------------------------------------------------\n\nA 0.5\u2009ml solution of coumarin compounds in methanol (50\u2009\u00b5M final concentration) was mixed with 4.5\u2009ml of the FRAP reagent. After 10\u2009min of incubation at 37\u2009\u00b0C, absorbance at 593\u2009nm was measured using methanol as blank.[@CIT0020] ^,^ [@CIT0021] The calibration curve was prepared with quercetin and the results expressed as: antioxidant index based on quercetine (Q) (mmol Q/mol). The analyses were performed in triplicate.\n\nMurine neuroblastoma cell (N2a) culture and cell viability assay {#s0010}\n----------------------------------------------------------------\n\nN2a cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 0.1% gentamicin in a 5% CO~2~ atmosphere. N2a cells were transferred to a 96-well plate (\u223c10,000 cells/cm^2^) and incubated for 24\u2009h, before treatment with the compounds at 10 or 50\u2009\u00b5M. Cell viability was evaluated by MTT (3-\\[4,5-dimethylthiazol-2-yl\\]-2,5-diphenyl tetrazolium bromide) assay.\n\nResults and discussion {#s0011}\n======================\n\nCompounds design and synthesis {#s0012}\n------------------------------\n\nThe design of the novel alkylamino-coumarin derivatives ([Scheme 1](#SCH0001){ref-type=\"scheme\"}) was based on the structural requirements for mixed-type selective AChE inhibition present in alkylamino-indanone inhibitor recently described[@CIT0013], as well as on the widespread use of coumarins for this pharmacological activity[@CIT0022] ^,^ [@CIT0023]. The coumarin series was based on: 1- the maintenance of the cyclic alkylamino group, which is responsible for the interaction with the cationic catalytic site (CAS) of AChE, exploring different lengths of methylene linkers (2--6); 2-exchange of the indanone nucleus by the coumarin through non-classical isosterism of ring expansion[@CIT0024]; 3- use of hydrophobic groups at position 3 of coumarin, targeting interactions with the peripheral anionic site (PAS) of AChE.\n\n![Design of alkylamino-coumarin cholinesterase inhibitors series.](IENZ_A_1571270_SCH0001_C){#SCH0001}\n\nThe synthesis of the desired compounds started with the 7-OH-coumarin (**1**). In the first step, through an *O*-alkylation reaction of **1** with diverse dibromo-alkanes, the 7-bromoalkoxy-coumarin products (**2a--d**) were obtained in yields of 68--78%. The second step consisted of a bromination reaction of the bromoalkoxy-coumarins (**2a--d**) using Br~2~ in buffered medium of sodium acetate/acetic acid at room temperature, furnishing the brominated derivatives (**3a--d**) in yields ranging between 79--84%. These intermediates (**3a--d**) were then subjected to amination reactions with piperidine in acetonitrile, leading to the formation of the desired 7-amino-alkoxy-3-bromo-coumarin derivatives (**4a--d**) as yellow solids in 95--99% yields after purification by flash chromatography. From the 7-amino-alkoxy-3-bromo-coumarin derivatives with 2, 3 and 5 methylene spacers (**4a, 4b, 4d**), Suzuki cross coupling reactions were then carried out using Pd(PPh~3~)~4~ catalyst and phenyl and 4-dimethylamino-phenyl boronic acids to obtain the final arylated 3-substituted coumarins (**5a--c** and **6a--c**) in yields ranging between 70--75% after purification by flash chromatography ([Scheme 2](#SCH0002){ref-type=\"scheme\"}).\n\n![Reagents and conditions: (i) Br(CH~2~)~n~Br (*n*\u2009=\u20092--5), K~2~CO~3~, acetone, 60\u00b0 C, 2-8h, 68--78%; (ii) Br~2~, AcOH, NaOAc, r.t., 2h, 79--84%; (iii) piperidine, acetonitrile, 60\u00b0 C, 2-5h, 95--99%; (iv) Ph-B(OH)~2~ or 4-(Me)~2~N-Ph-B(OH)~2~, Na~2~CO~3~, Pd(PPh~3~)~4~, H~2~O, EtOH, PhMe, 80\u00b0 C, 3h, 70--75%.](IENZ_A_1571270_SCH0002_B){#SCH0002}\n\nCholinesterase inhibitory activity, biological profile, and SAR analyses {#s0013}\n------------------------------------------------------------------------\n\nThe inhibitory activities of the coumarin compounds (**4a--d**, **5a--c** and **6a--c**) on AChE and BuChE were determined by the Ellman's method[@CIT0014] ^,^ [@CIT0015] using donepezil as the reference compound. As depicted in [Table 1](#t0001){ref-type=\"table\"}, compounds presented potent inhibitory activities against AChE with IC~50~ values varying from 0.02 to 0.92\u2009\u00b5M for compounds **6a** and **6c** respectively. On the other hand, the tested coumarins were not so efficient in inhibiting BuChE with IC~50~ ranging from 0.90 to 15.87\u2009\u00b5M, demonstrating a good selectivity for AChE. The inhibition behavior of the simplest bromo-coumarins (**4a--d**) was quite similar of that related in the literature for AChE[@CIT0013], the bigger the methylene chain the lower the activity. However, we were surprised by compound **4d** with a five-methylene spacer link that was equipotent to **4a** (IC~50~=0.14\u2009\u00b5M for **4d** and IC~50~=0.18\u2009\u00b5M for **4a**). Conversely, the inhibitions of BuChE was in general inverse to those of AChE, and compounds with longest linker chains were more potent in inhibiting BuChE (IC~50~=8.37\u2009\u00b5M for **4a** and IC~50~=5.00\u2009\u00b5M for **4d**). By this way, we decided to evaluate the 3-aryl substituted coumarins with 2, 3 and 5 methylene spacers in the 7-amino-alkoxy group. In general, the substitution of bromine for phenyl (**5a--c**) or 4-dimethylamine-phenyl (**6a--c**) led to compounds with better potencies on the inhibition of both AChE and BuChE, and a reduction in the selectivity index (IC~50~ BuChE/AChE). However, one compound behavior itself differently and presented an interesting profile, the 4-dimethylamine-phenyl substituted coumarin (**6a**) with the best inhibition of AChE (IC~50~=0.02\u2009\u00b5M) and selectivity (IC~50~ BuChE/AChE = 354), quite similar to the reference drug donepezil (IC~50~ AChE = 0.007\u2009\u00b5M and selectivity = 365) ([Table 1](#t0001){ref-type=\"table\"}).\n\n###### \n\nAChE and BuChE inhibitory activities of coumarin compounds.\n\n  ![](IENZ_A_1571270_ILG0001.gif)                                                                                                                                                                                          \n  --------------------------------- ------------------ ----------------------------------------------- -------------------------------------- --------------------------------------- ------------------------------------ ---------------------------------------------------------\n  \u00a0                                 \u00a0                  IC~50~(\u03bcM)\u00b1SD[^a^](#TF1){ref-type=\"table-fn\"}   \u00a0                                      \u00a0                                                                            \n  Compound                          R                  n                                               AChE[^b^](#TF1){ref-type=\"table-fn\"}   BuChE[^c^](#TF1){ref-type=\"table-fn\"}   SI[^d^](#TF1){ref-type=\"table-fn\"}   FRAP value (mmol Q/mol)[^e^](#TF1){ref-type=\"table-fn\"}\n  **4a**                            Br                 1                                               0.18\u2009\u00b1\u20090.009                           8.37\u2009\u00b1\u20090.167                            47                                   NA\n  **4b**                            Br                 2                                               0.37\u2009\u00b1\u20090.008                           15.87\u2009\u00b1\u20090.007                           42                                   NA\n  **4c**                            Br                 3                                               0.55\u2009\u00b1\u20090.010                           4.92\u2009\u00b1\u20090.095                            9                                    NA\n  **4d**                            Br                 4                                               0.15\u2009\u00b1\u20090.005                           5.01\u2009\u00b1\u20090.253                            33                                   NA\n  **5a**                            Ph                 1                                               0.14\u2009\u00b1\u20090.009                           2.50\u2009\u00b1\u20090.177                            18                                   NA\n  **5b**                            Ph                 2                                               0.24\u2009\u00b1\u20090.014                           1.86\u2009\u00b1\u20090.024                            8                                    NA\n  **5c**                            Ph                 4                                               0.45\u2009\u00b1\u20090.036                           0.90\u2009\u00b1\u20090.001                            2                                    NA\n  **6a**                            4-(CH~3~)~2~N-Ph   1                                               0.02\u2009\u00b1\u20090.001                           6.73\u2009\u00b1\u20090.040                            354                                  7.49\u2009\u00b1\u20090.61\n  **6b**                            4-(CH~3~)~2~N-Ph   2                                               0.33\u2009\u00b1\u20090.011                           7.27\u2009\u00b1\u20090.273                            22                                   2.42\u2009\u00b1\u20090.19\n  **6c**                            4-(CH~3~)~2~N-Ph   4                                               0.96\u2009\u00b1\u20090.036                           3.85\u2009\u00b1\u20090.190                            4                                    2.77\u2009\u00b1\u20090.00\n  **Donepezil**                     --                 --                                              0.007\u2009\u00b1\u20090.0002                         2.39\u2009\u00b1\u20090.105                            365                                  --\n\n^a^Concentration required for 50% inhibition of ChEs, data were shown in mean\u2003\u00b1\u2003SD of triplicate independent experiments; ^b^AChE from electric eel; ^c^BuChE from horse serum; ^d^Selectivity index (SI) is defined as BuChE IC~50~/AChE IC~50~. ^e^Antioxidant index based on quercetine (Q); FRAP value (mmol Q/mol). 7,8-dimethoxy-coumarin (NA)[@CIT0025] and ethyl 2--(7,8-dimethoxy-2-oxo-2H-chromen-3-yl)acetate (1.2\u2003\u00b1\u20030.1)[@CIT0025].\n\nThe antioxidant evaluation of coumarin compounds showed that only **6a--c** presented activity in Ferric Ion Reduction Method (FRAP) with values from 2.42 to 7.49\u2009mmol Q/mol ([Table 1](#t0001){ref-type=\"table\"}). Series **4a--d** and **5a--c** did not demonstrate any considerable result, similar to other 7-alkoxy coumarins described in the literature[@CIT0025]. Probably, the antioxidant effect is coming from dimethylamino-phenyl moiety and this feature could be explored in a forthcoming series.\n\nAiming at discovery the mode of action of coumarins described herein, the most potent compounds from the bromo and aryl 3-substituted coumarins, i.e. **4d** and **6a**, were selected for kinetic studies. The linear Lineweaver--Burk equation of the Michaelis--Menten was applied to evaluate the inhibition profile. Increasing concentrations of both compounds were able to increase K~m~ and decrease V~max~, presenting a mixed-type inhibition in AChE as well as in BuChE, as exemplified in [Figure 1](#F0001){ref-type=\"fig\"} for compound **6a** (complete analysis in [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270)). The competitive inhibitory constant (Ki) and the noncompetitive constant (Ki') for **6a** and **4d** are described in Table 3 at [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270). As example, the best Ki values against AChE were obtained for compound **6a**: Ki = 0.001\u2009\u00b5M (competitive) and Ki'=0.010\u2009\u00b5M (noncompetitive).\n\n![Lineweaver-Burk plots of EeAChE inhibition kinetics of compound **6a**. Inset: concentrations used for **6a** are depicted with \\[I\\] graphic symbol.](IENZ_A_1571270_F0001_C){#F0001}\n\nWith the complete inhibitory profile of the target compounds, we proceeded with a molecular modeling evaluation to understand the importance of changing the methylene size spacer and nature of substituents in position 3 of coumarins. Thus, we selected compounds **4a** and **6a**, the strongest 2-methylene spacer bromo and aryl substituted coumarins inhibitors of AChE; and **6c**, the weakest inhibitor. Docking results in EeAChE and EqBuChE are presented in Table 4 at [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270). All inhibitors were generally predicted as better ligands of EeAChE, being **6a** (Goldscore = 78.1) better than **6c** (Goldscore = 71.4) and **4a** (Goldscore = 64.9), whereas **6c** (Goldscore = 67.9) was better than **6a** (Goldscore = 63.0) and **4a** (Goldscore = 57.6) as a ligand of EqBuChE, in qualitative accordance to our experimental results. The molecular docking results of compound **4a**, **6a** and **6c** showed that all were able to occupy the peripheral (PAS) and the catalytic (CAS) sites simultaneously in the EeAChE ([Figure 2](#F0002){ref-type=\"fig\"}) (and Figure 5 at [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270)), as previewed by kinetic evaluations. In the CAS, they interact similarly by means of their protonated piperidinyl group with Trp86 (a cation-\u03c0 interaction). In the PAS, both **6a** and **6c** molecules were involved in \u03c0-stacking interactions with Trp286, which was more effective for **6c**, involving its coumarin ring ([Figure 2](#F0002){ref-type=\"fig\"}). On the other hand, **4a** was only capable of doing weak hydrophobic interactions with Trp286 (Figure 5 at [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270)). The presence of a narrower spacer in **6a** makes its coumarin ring to be best located in the gorge, where it is involved in H-bonds with Tyr337 and the peptide group of Phe295. These H-bonds, that had no counterparts in the **6c**/enzyme complex, were probably the reason for the most effective interaction between compounds with short spacers and EeAChE, which could be related to their greater inhibitory action over the enzyme.\n\n![Superposition of the interaction poses of compounds **6a** (A, carbon atoms in yellow) and **6c** (B, carbon atoms cyan) with EeAChE obtained by molecular docking (Goldscore function). H-bond distances (\u00c5) are shown in yellow. Figure generated with PyMol 0.99 (DeLano Scientific LLC).](IENZ_A_1571270_F0002_C){#F0002}\n\nCell cytotoxicity and in silico ADMET physico-chemical profile analysis {#s0014}\n-----------------------------------------------------------------------\n\nIn order to accede the drugability of tested coumarins, we first proceeded with the cytotoxicity evaluation against N2a cells (neuroblastoma), after 48\u2009h incubation at concentrations of 10 and 50\u2009\u00b5M (Figure 3 and [Supplementary material](https://doi.org/10.1080/14756366.2019.1571270)). The most potent compounds in inhibiting AChE, i.e. **4a**, **5a**, and **6a**, were not cytotoxic at the maximum tested concentration (50\u2009\u00b5M) ([Figure 3](#F0003){ref-type=\"fig\"}). As a rule, long methylene chains (three or five spacers) in phenyl-substituted coumarins (**5b**, **5c**, **6b** and **6c**) could not be useful for further developments due to increase in toxicity.\n\n![Neuroblastoma cell viability after compound treatment. Samples containing compounds were added to the culture 48\u2009h before MTT addition. The compounds were tested at the final concentration of 50\u2009\u00b5M. MTT reduction was evaluated as described in Experimental Procedures. Data are expressed as the percentage of MTT reduction relative to the value for control cells (cells without treatment). Error bars represent standard deviations. \\*\\**p*\u2009\\<\u2009.01; \\*\\*\\**p*\u2009\\<\u2009.001; \\*\\*\\*\\**p*\u2009\\<\u2009.0001.](IENZ_A_1571270_F0003_C){#F0003}\n\nFinally, in silico evaluations showed a good ADMET profile for coumarin compounds. Parameters as topological polar surface area (TPSA), consensus Log P, Log S, human intestinal absorption (HIA), blood--brain barrier permeation (BBB), and P\u2010glycoprotein (P-gP) substrate and drug-likeness profile ([Supplementary material](https://doi.org/10.1080/14756366.2019.1571270))[@CIT0026]. TPSA values and consensus Log P ranged from 42.68 to 45.92 and 3.35 to 5.01, respectively. The moderate polarity (PSA \\< 79 \u00c5^2^) and relative lipophilicic characteristics put our compounds in the yellow compartment of BOILED\u2010Egg model ([Figure 4](#F0004){ref-type=\"fig\"}), having a high probability to access the CNS[@CIT0027], which is fundamental for the distribution of central\u2010acting molecules. Additionally, the most potent compounds **4a**, **5a** and **6a** were not considered as P-gP substrate and having a good drug-likeness profile with no one violation on the Lipinski[@CIT0028], Ghose[@CIT0029], Veber[@CIT0030], Egan[@CIT0031] and Muegge[@CIT0032] rules.\n\n![BOILED-Egg ADMET model[@CIT0027] for coumarin compounds **4a--d**, **5a--c,** and **6a--c**. (HIA) gastrointestinal absorption; (BBB) brain penetration; (PGP+) substrate for P\u2010glycoprotein; (PGP-) Not a substrate for P\u2010glycoprotein.](IENZ_A_1571270_F0004_C){#F0004}\n\nConclusions {#s0015}\n===========\n\nThe designed and synthesized coumarin compounds were able to potently inhibit cholinesterases in the nanomolar range. In general, compounds with narrow methylene linkers were more potent and selective for AChE (with IC~50~ and selectivity of up to 20\u2009nM and 354 times, respectively), and less toxic as well. The introduction of aromatic substituents in position 3 of coumarins led to compounds with better potencies on the inhibition of both AChE and BuChE. As highlighted, compound **6a** could be elected as a hit for *in vivo* studies, showing good AChE inhibition potency and selectivity (IC~50~=20\u2009nM and 354 times), antioxidant properties, low cytotoxicity and good predict ADMET profile.\n\nSupplementary Material\n======================\n\n###### Supplemental Material\n\nAcknowledgments\n===============\n\nThe authors would like to thank Marco E. F. de Lima for furnishing donepezil standard and Marina Amaral from Laborat\u00f3rio de Apoio ao Desenvolvimento Tecnol\u00f3gico (LADETEC, UFRRJ) for the mass spectrometry analysis.\n\nDisclosure statement {#s0016}\n====================\n\nNo potential conflict of interest was reported by the authors.\n\n[^1]: Supplemental data for this article can be accessed [[here]{.ul}](https://doi.org/10.1080/14756366.2019.1571270).\n"}
+{"text": "World Heart Day will be observed September 29, 2014. The focus of World Heart Day this year is creating heart-healthy environments in which persons are able to make heart-healthy choices wherever they live, learn, work, and play. Heart disease and stroke are the world's leading causes of death, claiming an estimated 17.3 million lives in 2008, and representing 30% of all deaths worldwide ([@b1-843]). A heart-healthy environment can help persons make healthy choices to reduce their risk for heart disease. World Heart Day 2014 encourages persons to reduce their risk for cardiovascular disease by promoting smoke-free environments, environments that encourage physical activity, access to healthy food choices, and a heart-healthy planet for all.\n\nCDC is working to help create heart-healthy environments in multiple ways, including community-based approaches, such as the Sodium Reduction in Communities Program (SRCP), and community-clinical linkages, such as the Million Hearts Initiative. SRCP aims to increase access to and accessibility of lower-sodium food options while building the evidence base on population approaches to reduce sodium consumption at the community level. Million Hearts aims to prevent 1 million heart attacks and strokes by 2017 by bringing together communities, health systems, nonprofit organizations, federal agencies, and private-sector partners from across the country to fight heart disease and stroke and their risk factors.\n\nAdditional information about World Heart Day is available at <http://www.world-heart-federation.org/?id=123>. Additional information about Million Hearts, SRCP, and CDC's Healthy Community Programs is available at <http://millionhearts.hhs.gov> and <http://www.cdc.gov/nccdphp/dch/programs/healthycommunitiesprogram/index.htm>.\n"}
+{"text": "Introduction\n============\n\nPistachio is an edible seed from the pistachio tree grown broadly in hot--dry regions of the Middle East, Mediterranean, and Americas (Garcia et\u00a0al. [@b8]). Nutritional value, palatability, easily digestion, high calorie value, vitamins, and minerals are among the characteristics that make pistachio superior to other nuts. However, factors such as harvest, transportation, and processing enhance the risk of fungal infection in pistachio (Salek Zamani [@b21]). Favorable conditions in terms of temperature, relative humidity, and water activity result in fungal growth and production of secondary metabolites such as mycotoxin. Aflatoxins are toxic substances from mycotoxin group produced by fungi including *Aspergillus flavus*, *Aspergillus parasiticus*, and *Aspergillus nomius* (Sarhang Pour et\u00a0al. [@b22]; Daraei Garmakhany et\u00a0al., [@b22]). These toxins cause dangerous diseases such as hepatic cancer in humans (Sarhang Pour et\u00a0al. [@b22]). In order to inhibit from aflatoxin production, prevention methods should be in use (Shantha and Decker [@b25]).\n\nAntimicrobial is a kind of active packaging which can enhance storage life of the food product and protect its microbial health (Seydim and Sarkis [@b23]). Today, studies on packaging of food products are concentrated on biodegradable films such as prepared films from edible proteins of animal and plant origin (zein, gluten, soy, peanut, albumin, gelatin, collagen, casein, and whey protein). Edible coating is a thin layer of natural substances that covers the food product surface and prevents undesirable changes in flavor, texture, and appearance of the food product (Robertson [@b20]). Methyl cellulose is a chemical substance which cannot be found in nature. It is obtained by methylation of 30% of hydroxyl groups which is soluble in cold water and gives a clear solution. However, it is insoluble in hot water and changes to saturated solution. Due to its hydrophobic nature, methyl cellulose has a high potential for inhibition of water and oxygen penetration and so increases microbial and chemical resistance compared to other cellulose derivatives (Baker [@b4]; Shahidi et\u00a0al. [@b24]). Also, different concentrations of methyl cellulose lead to formation of different film with different oxygen barrier abilities during coating of pistachio nuts. Pistachio is produced in large quantities and exported from Iran; hence, controlling infection in this product is necessary. However, no attempt has been done on coating pistachio by hydrocolloids. Therefore, the aim of this research was to investigate the effects of different concentrations of methyl cellulose on chemical changes in pistachio nuts during 4\u00a0months of storage.\n\nMaterials and Methods\n=====================\n\nPreparation of methyl cellulose solution\n----------------------------------------\n\nAbbasali pistachio was obtained from Damghan city. Coating was performed by immersion method. To this end, pistachio kernel was randomly divided into 1\u00a0kg packages. In a beaker containing distilled water, methyl cellulose at concentrations of 0.1%, 0.5%, 1%, and 2% was gradually added, and after 30--45\u00a0min methyl cellulose was completely dissolved in water. After removing air bubbles, methyl cellulose solution was ready for coating (Turhan and Sahbas [@b26]).\n\nCoating process\n---------------\n\nFor coating at each concentration, 1\u00a0kg of pistachio kernels was immersed in degasified methyl cellulose solution, stirred for 5\u00a0min and coated samples were then stored in an oven at 25\u00b0C for 3\u00a0days in order to dry. One kilogram of coated dried samples was packed in polyethylene bags and stored in an incubator at 25\u00b0C for 4\u00a0months (Maftoonazad and Ramaswamy [@b12]).\n\nSample assay\n------------\n\nIn order to evaluate the effect of methyl cellulose concentration on chemical and quality attributes of pistachio nuts, moisture content, peroxide, acidic, and iodine values were measured. Assays were carried out at the beginning of storage time (immediately after coating and drying), in the middle (after 2\u00a0months of storage) and at the end of storage (after 4\u00a0months of storage).\n\nStatistical analysis\n--------------------\n\nThis study was conducted in a completely randomized design (CRD) and Tukey\\'s follow-up test was used to determine the difference between treatments (*P*\u00a0\\<\u00a00.05). Statistical analysis was done using IBM- SPSS (version 19, Armonk, NY, USA). All experiments were performed in three replications.\n\nResults and Discussion\n======================\n\nMoisture content\n----------------\n\nBased on the results obtained from measurement of pistachio moisture content (Fig.[1](#fig01){ref-type=\"fig\"}), the effect of concentration and storage time on moisture content was found to be significant (*P*\u00a0\\<\u00a00.05). The highest moisture content was related to pistachio kernel coated with 2% of methyl cellulose solution and the lowest amount was related to control, 0.1% and 0.5% samples, respectively. The reason for this difference is that when pistachio kernels are immersed in methyl cellulose solution, they absorb some parts of the solution through dispersion phenomenon and pistachio kernels coated with 2% of the solution have more density, lose lower moisture content during drying in an oven as well as during the storage period. The moisture content of the food product is decreased with storage time due to moisture exchange with the environment depending on the environment condition (whether free space or inside a package). It is clear that moisture loss in open-packaged food products is lower, depending on the type of packaging material, and other factors such as temperature, exposure to direct sunlight, and nature of the food product. Results showed that moisture loss was decreased due to type of packaging, unique characteristics of methyl cellulose coating, and particular conditions of storage. Several studies showed reduction in moisture loss due to coating and packaging in food materials. Achour et\u00a0al. ([@b2]), Ur- Rehman ([@b27]), and Babarinde and Fabunmi ([@b3]) reported reduction in moisture loss in packaged dates. Razmkhah ([@b18]) observed that coated pistachio with chitosan had lower moisture loss in comparison with uncoated samples.\n\n![Moisture content of different pistachio-coated samples during storage period.](fsn30003-0355-f1){#fig01}\n\nPeroxide value\n--------------\n\nPeroxide value is the initial product of oxidation of fat substances which does not lead to undesirable flavor and aroma directly but shows the degree of oxidation progress. Peroxide formation proceeds slowly in the initial stages. However, in the secondary stages, it acts as a catalyzer in oil oxidation. More than 89% of pistachio fatty acids are unsaturated. This amount of unsaturated fatty acid leads to enhancement of nutritional value of pistachio but makes it susceptible to autoxidation (Maskan and Karatas [@b13]). Peroxide measurement was done according to the Shantha and Decker ([@b25]) method. The beginning of this research coincided with the harvest season in order to use fresh pistachio. Moreover, heavy polyethylene packages which were oxygen barriers were used and stored in an incubator in order that storage temperature is constant and pistachio packages are away from direct sunlight. Rate of rancidity and browning reactions resulted from lipid oxidation were reduced in the food products that coated with methyl cellulose (Brenji Ardestani et\u00a0al. [@b5]).\n\nAs can be seen from Figure[2](#fig02){ref-type=\"fig\"}, after 4\u00a0months of storage, a significant difference in peroxide value was observed between samples coated with different concentrations of cellulose (*P*\u00a0\\<\u00a00.05). The highest peroxide value was related to control and the lowest was related to the sample coated with a concentration of 2% methyl cellulose. It can be explained that in coated nuts methyl cellulose acts as a barrier agent and inhibits oxidation. In addition, by increasing coating concentration barrier efficiency is increased. On the other hand, pistachio oil has a high oxidative stability and peroxide enhancement in this oil proceeds in a lower rate (Mazinani et\u00a0al. [@b14]). Yen and Liu ([@b28]) found that substances extracted by ethanol (ethanolic extract) can reduce oxidation of linoleic acid to 80%. This study showed that antioxidant activity of these substances is higher than tocopherol but lower than butylated hydroxyanisole (Yen and Liu [@b28]). In spite of studies done on the antioxidant property of these products, there is little information about the role of dietary antioxidants. Bressa et\u00a0al. ([@b6]) separated different compounds from butter-containing cookies which were able to stop autoxidation reactions in the initial stages. In their studies on peanut packaging, Duan et\u00a0al. ([@b7]) reported a positive role of this process in inhibition of oxidation of peanut oil. Razmkhah ([@b18]) showed an increase in peroxide value in chitosan-coated pistachio samples.\n\n![Peroxide value of different pistachio-coated samples during storage period.](fsn30003-0355-f2){#fig02}\n\nAcidic value\n------------\n\nAcidic value is an indicative index for oil quality (Hill and Hanna [@b10]). Pistachio acidity was measured according to the AOCS standard (1993). Results showed that there was a significant difference in terms of acidic value between different concentrations of coating solution during storage time (*P*\u00a0\\<\u00a00.05). As shown in Figure[3](#fig03){ref-type=\"fig\"}, the highest acidic value was related to pistachio samples coated with 2% methyl cellulose and the lowest amount was related to control sample, pistachio samples coated with 0.1% methyl cellulose and pistachio samples coated with 0.5% methyl cellulose, respectively. Increase in the amounts of free fatty acids levels the hydrolysis process in pistachio oil. Lipase and strase cause oxidative reactions of catalase enzyme. These two enzymes separate fatty acids from fat and produce free fatty acids. Therefore, free fatty acids can act as substrates of oxidation reactions. During storage periods the amounts of free fatty acids in pistachio are lower than other nuts; it may be due to the presence of antioxidants in this fruit that inhibit oxidation process propagation (Grosch et\u00a0al. [@b9]). At a given temperature, increase in moisture content enhances hydrolysis of fats and production of fatty acids. As described in the previous section, pistachio nuts coated with higher concentrations of methyl cellulose retain a higher amount of moisture content during storage. It can be expected that the lipid oxidation rate in coated samples with a higher concentration of methyl cellulose would be higher than that of uncoated and coated samples with lower concentration of methyl cellulose. It means that moisture has a positive effect on lipolysis and induces this reaction (Reid [@b19]).\n\n![Acidic value for different pistachio-coated samples during storage period.](fsn30003-0355-f3){#fig03}\n\nIodine value\n------------\n\nIodine index indicates the amount of absorbed iodine for 100\u00a0g oil sample. It is regarded as an index for unsaturation degree of unsaturated fatty acids like oleic acid (Mirnezami Ziabari and Saneii [@b15]; Kirk & Sawyer, [@b15]). Iodine value was measured according to AOAC standard ([@b15]).\n\nResults showed that there was a significant difference in iodine value between coated and uncoated samples. As can be seen from Figure[4](#fig04){ref-type=\"fig\"}, during 4\u00a0months of storage, coated pistachio samples with 0.1% methyl cellulose had the highest loss in iodine value while control sample had the lowest iodine value loss. Moreover, a significant difference was observed between treatments in iodine value during storage time. Iodine value loss during storage periods may be due to reduction in unsaturated fatty acids of oil and consequently destruction of double bands resulting from isomerization, oxidation, and polymerization. As shown in Figure[4](#fig04){ref-type=\"fig\"}, a slight increase in the amount of oleic acid led to a decrease in the amounts of linoleic and linolenic acids, thus reducing the iodine value (Qavami et\u00a0al. [@b17]).\n\n![Results of iodine value test for different pistachio-coated samples during storage period.](fsn30003-0355-f4){#fig04}\n\nTaking image of pistachio nuts under electron microscope by SEM method\n----------------------------------------------------------------------\n\n### Coating diameter\n\nFigure[5A](#fig05){ref-type=\"fig\"}--E indicates that all of the coatings have a good density. Based on mean comparison between different samples, coated sample with 1% methyl cellulose has lower diameter than other samples, but it shows the highest density and viscosity. The reason for this low diameter is high viscosity and density of layers at this concentration.\n\n![Diameter of pistachio nut samples: (A) control; (B) coated with 0.1% of methyl cellulose; (C) coated with 0.5% of methyl cellulose; (D) coated with 1% of methyl cellulose; (E) coated with 2% of methyl cellulose.](fsn30003-0355-f5){#fig05}\n\n### Coating uniformity\n\nCoating uniformity can be evaluated through investigation of coating surface roughness. It should be mentioned that to obtain electron microscopic images, different parts of the surface and cross sections of samples were investigated. Investigation of cross section surface indicates that sample surfaces are thoroughly coated and there is no fracture. As shown in Figure[6](#fig06){ref-type=\"fig\"}, the surface roughness of samples is different. In general, coated samples in D and E are more uniform than those in B and C while sample in E was more uniform than that in D (Fig.[6](#fig06){ref-type=\"fig\"}).\n\n![Coating uniformity: (A) control sample; (B) 0.1% sample; (C) 0.5% sample; (D) 1% sample; and (E) 2% sample.](fsn30003-0355-f6){#fig06}\n\n### Size of coating porosities\n\nAs porosity determination in imaging software is based on resolution degree, images should be selected in a way that only porosity can be measured. Therefore, inverted electron images (BSE) as well as secondary electron images should be taken from a certain point in order to identify real porosities by comparing them. On the other hand, color change in inverted electron images is due to phase change and presence of porosity. Therefore, to obtain inverted electron images, sample surface should not be uneven. As the cross section surface of pistachio sample is not even (to obtain an even sample, its surface should be polished) and as porosities are well detectable in secondary electron images, images were obtained by a secondary electron detector. Results obtained from images of Figure[7](#fig07){ref-type=\"fig\"} are indicated in Table[1](#tbl1){ref-type=\"table\"}. These images were taken by image tool software. As shown in Figure[7](#fig07){ref-type=\"fig\"} and Table[1](#tbl1){ref-type=\"table\"}, a coated sample is more porous than other samples and E coated sample has the lowest porosity.\n\n###### \n\nSize of coating porosity\n\n  Sample code          Porosity percent\n  -------------------- ------------------\n  Control sample       4.3\n  0.1% coated sample   0.6\n  0.5% coated sample   1.9\n  1% coated sample     1\n  2% coated sample     0.6\n\n![Coating porosity: (A) control sample; (B) 0.1% sample; (C) 0.5% sample; (D) 1% sample; (E) 2% sample.](fsn30003-0355-f7){#fig07}\n\nConclusion\n==========\n\nWith respect to images taken from sample surface, D and E coated samples are more uniform than other coated samples.Investigation of cross section surface of samples indicates that E coated sample has higher density and viscosity than B, C, and D coated samples. (B\u00a0=\u00a00.1% sample; C\u00a0=\u00a00.5% sample; and E\u00a0=\u00a02% sample). Appropriate coating is a coating with higher density and viscosity, has a surface of minimum roughness and higher uniformity, lower porosities, and smaller mean size of porosity. With respect to these facts and based on chemical tests, images taken by electron microscope and tables related to properties of pistachio samples (surface, cross section surface, diameter, uniformity, and porosity size), it can be concluded that pistachio sample coated with 1% methyl cellulose is the best sample.\n\nConflict of Interest\n====================\n\nNone declared.\n\n[^1]: **Funding information** No funding information provided.\n"}
+{"text": "Background {#Sec1}\n==========\n\nChronic diseases disproportionately burden Aboriginal and Torres Strait Islander people in Australia, contributing to 80% of the mortality gap between Aboriginal and Torres Strait Islander and non-Indigenous people \\[[@CR1]--[@CR4]\\]. The greatest contributors to this mortality gap are cardiovascular (CV) diseases, with Aboriginal and Torres Strait Islander people three times more likely to suffer a major coronary event than other Australians \\[[@CR3], [@CR5]\\]. Since Aboriginal and Torres Strait Islander health does not just impact the physical wellbeing of an individual but also the social, emotional and cultural well-being of the whole community \\[[@CR6]\\], the implications of CV diseases may extend from individuals through to their families and communities. Effective prevention and management of these conditions is elemental to reach the goal of closing the life expectancy gap between the Aboriginal and Torres Strait Islander population and non-Indigenous Australians within a generation \\[[@CR7]\\].\n\nIn order to improve quality of life and life expectancy for Aboriginal and Torres Strait Islander people living with CV disease, secondary prevention strategies such as cardiac rehabilitation and self-management are critical. Published evidence for best practice CV rehabilitation and self-management indicate key features of effective programs. These include approaches that understand health and wellness through an Aboriginal and Torres Strait Islander lens and ensure an holistic approach with care delivered by a multidisciplinary team that includes Aboriginal and Torres Strait Islander Health Workers (AHWs) or other Aboriginal and Torres Strait Islander care providers \\[[@CR8]--[@CR12]\\]. Further, secondary prevention for CV disease should include physical activity and education that is delivered, where possible, in a group or family setting \\[[@CR13]--[@CR16]\\] and within a culturally responsive environment, such as an Aboriginal and Torres Strait Islander Community Controlled Health Service (ATSICCHS) setting \\[[@CR11], [@CR12]\\] and implement strategies that address a range of chronic conditions and comorbidities where appropriate \\[[@CR10]--[@CR12], [@CR17]\\]. Evaluated CV disease specific, culturally appropriate group programs for Aboriginal and Torres Strait Islander people that incorporate at least some of these evidence-based strategies, show some short term improvements in CV disease risk factor measures \\[[@CR11], [@CR12]\\]. However, documented CV specific group programs for Aboriginal and Torres Strait Islander peoples are rare.\n\nThe implementation and evaluation of broad chronic condition self-management (CCSM) programs are more common in the literature, and have been delivered in Aboriginal and Torres Strait Islander settings. Although not CV specific programs, many people who participate present with CV disease and other comorbidities and/or multiple risk factors for CV disease. Correspondingly, such programs tend to share many strategies recognised as important for CV prevention and management. The *Stanford Chronic Disease Self-Management Program* \\[[@CR18], [@CR19]\\] is a common CCSM strategy made available through the Australian healthcare system in 2004. In a study evaluating the model in an Aboriginal and Torres Strait Islander community, improvements were noted in some CV risk factor measures including BMI, cholesterol levels and glycosylated haemoglobin \\[[@CR20]\\]. The *Flinders Program* is also a widely adopted CCSM program in clinical settings throughout Australia \\[[@CR21]--[@CR24]\\]. When implemented and evaluated at three community controlled health settings in rural, regional and metropolitan South Australia, the *Flinders Program* along with other coordinated systemic CCSM strategies showed small but statistically significant improvements over time in CV risk factor indicators of BMI, total cholesterol, triglyceride, low-density lipoprotein, and glycosylated haemoglobin levels \\[[@CR24]\\] .\n\nThere is undoubtedly a place for cardiac rehabilitation and self-management to occur within broader group programs for CCSM, as these adopt similar CV-specific strategies, address similar risk factors, and have indicated some improvements across CV risk factor indices. However, there is no published evidence examining the effect of such broader CCSM group programs for urban Aboriginal and Torres Strait Islander people who have, or are at risk of, CV disease specifically. The *Work It Out* program is a novel approach to CCSM underpinned by the aforementioned evidenced-based strategies that has been developed and implemented by the *Institute for Urban Indigenous Health (IUIH).* This paper aims to evaluate the program for its effect on clinical outcome measures in a population of Aboriginal and Torres Strait Islander participants with or at risk of CV disease in urban south-east Queensland.\n\nProgram overview {#Sec2}\n----------------\n\nThe *Work It Out* Program is a culturally responsive CCSM group program that has been implemented since 2011 across eight ATSICCHSs in south-east Queensland. Adopting the holistic view of Aboriginal and Torres Strait Islander health, the *Work It Out* program combines an inter-professional allied-health partnership approach based on published evidence of best practice CCSM and rehabilitation strategies with Aboriginal and Torres Strait Islander people.\n\nEntry to the program is by General Practitioner referral at participating ATSICCHSs, with many participants referred for the prevention or management of CV conditions. The progam runs on a successive 12\u00a0week cycle, with clients able to participate in two or more *Work It Out* sessions per week (as dictated by participant and location demands). The program has flexible entry and exit points and allows participants to be absent from several sessions due to responsibilities within family and community, and then return to the program at a later time. A *Work It Out* session consists of a 45\u00a0min 'yarning' (education) session delivered by a variety of health professionals in a culturally-safe environment. This is followed by a one hour exercise program undertaken in a group setting, developed by an accredited exercise physiologist or physiotherapist and indivudally tailored to meet the needs of participant's unique chronic conditions. An AHW or other Aboriginal and Torres Strait Islander staff member is an integral part of the team working closely with the accredited exercise physiologist/physiotherapist at each site.\n\nMethods {#Sec3}\n=======\n\nThis study was underpinned by a conceptual framework based on the principle of Aboriginal and Torres Strait Islander community control. This conceptual framework began with the research design, priorities and direction being set and monitored by a community controlled health organisation, through a relationship of trust developed over time. Two authors (KM, RM) are Aboriginal researchers, ensuring the research was culturally and ethically responsive, and driven by Aboriginal and Torres Strait Islander priorities. The lead author was employed with the *Work It Out* program and developed relationships with participants and colleagues over a 12\u00a0month period, before this study took place. Additionally, an upper-management member of the IUIH staff was part of the research team for this project (AN). This ensured ongoing support of the project that aligned with the IUIH core values. Persistent guidance throughout the entire research process, including data collection, analysis and dissemination was provided by *Work It Out* staff. Significantly, *Work It Out* participants became part of the research process, through the attendance at yarning sessions by the researcher to informally share research outcomes. Frequently, this would involve a reciprocal learning process whereby participants would ask questions of the research that would require further investigation, and regular reporting back to their *Work It Out* group. It is this framework that shaped the direction of the research project which became imperative for conducting ethical and culturally competent quantitative research with the population group. This conceptual framework aligned to the *National Health and Medical Research Council's* guidelines for ethical conduct in Aboriginal and Torres Strait Islander research \\[[@CR25]\\] and shaped ethical research practice, ensuring that Aboriginal and Torres Strait Islander people and perspectives were included in the collection, analysis and dissemination of the research \\[[@CR25]--[@CR29]\\].\n\nA quasi-experimental, pre-post test design was used to test short-term changes in routinely collected clinical outcome measures, following participation in the *Work It Out* program, from baseline to first follow-up assessment, for all participants with or at risk of CV disease. The research hypothesis was that participation in the WIO program is associated with short-term stable or improved clinical outcome measures for participants with or at risk of CV disease.\n\nA secondary outcome of this research was to explore factors that may have influenced positive changes in these clinical outcome measures, to provide a quantitative understanding about elements which may contribute to the *Work It Out* participant's overall wellbeing. This method was specifically chosen as there is a pressing identified need to ensure research questions are developed in a culturally competent manner, considering methods that move away from a narrative of deficit and disadvantage, and instead, seek to empower wellness within Aboriginal and Torres Strait Islander communities \\[[@CR26], [@CR30], [@CR31]\\].\n\nParticipants {#Sec4}\n------------\n\nParticipants were recruited purposively from a population of urban Aboriginal and Torres Strait Islander patients (18\u00a0years or over) with or at risk of chronic disease, who attended one of six participating ATSICCHSs in South-east Queensland between 2012 and 2014. Participants were referred by their general practitioner into the program, after completion of an *Aboriginal and Torres Strait Islander Health Assessment* \\[[@CR32]\\]*, General Practitioner Management Plan* and/or *Team Care Arrangement* \\[[@CR33]\\]*.*\n\nInclusion criteria for this study specified that participants have at least one diagnosed CV disease, or at least one CV disease risk factor. Participants were included in the study if they had any existing CV disease affecting the heart and/or blood vessels \\[[@CR34], [@CR35]\\]. Participants existing CV diseases were diverse and included: congestive heart failure/chronic heart failure; cardiomyopathy; angina; myocardial infarct/acute myocardial infarct; arrhythmia; valvular disease; ischemic heart disease; cerebrovascular disease/stroke; deep vein thrombosis; pulmonary embolism; and, peripheral vascular disease*.* Additionally, participants were included in the study if they had at least one of the following risk factors for CV disease: the presence of diabetes mellitus or insulin dependent diabetes mellitus; high cholesterol; hypertension; overweight/obesity; the presence of a social and emotional wellbeing condition (SEWB); chronic kidney disease; having a heavy alcohol intake; being a current smoker \\[[@CR5], [@CR36], [@CR37]\\]; and, being over 74\u00a0years of age at baseline \\[[@CR38]\\].\n\nData {#Sec5}\n----\n\nUpon entry to the *Work It Out* program, an initial assessment is routinely undertaken by an accredited exercise physiologist or physiotherapist where basic baseline demographic and medical history are collected, as well as baseline outcome measures. At this initial assessment, participants were counselled in plain English about the research project and only those who gave signed, informed consent were included.\n\nThe collection of medical history at baseline determined participants existing CV conditions and risk factors for CV disease which allowed for the decision of inclusion into to the study using the aforementioned inclusion criteria. This medical history was cross-referenced with data from the participants most recent *Aboriginal and Torres Strait Islander Health Assessment*, within patient medical files. Descriptive data, such as the number and presence of other co-morbidities were also collected from these sources. Baseline clinical outcome measures included systolic and diastolic blood pressure (BP) (mmHg); weight (kg) and body mass index (BMI) (kg/m^2^); waist, hip circumference (cm) and Waist Hip Ratio (WHR) (waist cm/hip cm); and, six minute walk test (6MWT). Systolic and diastolic BP was measured using an electronic sphygmomanometer on participants in the seated position. Weight (kg) was measured without shoes, using an electronic body mass scale. Height (cm) was measured using a tape measure from floor to crown of head, and corresponding BMI calculated (BMI\u00a0=\u00a0kg/m2). Waist circumference (cm) was taken using a tape measure at midpoint between the iliac crest and lower ribs. Likewise, hip circumference (cm) was measured with a tape measure at the maximum point of prominence of the buttocks. WHR was calculated using weight and hip circumference measurements (WHR\u00a0=\u00a0Waist circumference/hips circumference). The 6MWT measures the distance an individual can comfortably walk in 6\u00a0min and was undertaken following the standard *American Thoracic Society Guidelines* \\[[@CR39]\\].\n\nThese clinical outcome measures were also collected at six week intervals throughout the program. Whilst the aim was to collect follow-up clinical outcome data after a 12\u00a0week period, the rolling program entry and flexibility in attendance meant the timing of follow up physical assessments varied between participants. Missing follow-up data occurred if participants were absent from the program on the day the assessments were scheduled, or, if they had discontinued the program. Furthermore, due to resourcing and/or time limitations, often not all clinical outcome measures were recorded for each participant at each follow-up session. Only participants who recorded at least one clinical outcome measure at their initial and first follow-up (approx. 12\u00a0weeks) assessment were included for analysis. All data were de-identified onsite at *IUIH* and provided for the purposes of secondary data analysis.\n\nStatistical analysis {#Sec6}\n--------------------\n\nPaired t-tests were used to determine if there was a statistically significant change in clinical outcome measures between baseline and follow-up assessment. In addition to the total cohort, participant sub-groups were allocated during secondary data analysis using the following published evidence-based criteria which was determined a priori:BMI groups -- *healthy*, *overweight*, *obese I*, *obese II* and *extreme obesity* \\[[@CR40]\\];BP categories -- *normal*, *high/normal*, *high*, *very high* \\[[@CR41]\\];Waist and WHR groups -- *no increased risk of metabolic complications, increased risk* and *substantially increased risk* \\[[@CR1], [@CR40]\\].\n\nClinical groups were assigned based on the participant's baseline measurement.\n\nChange in each clinical outcome measure, and potential factors influencing the change were explored using general linear models. Models were built by examining the significance of factors and covariates individually, or in combination. Factors considered were BMI group, presence of CV disease, number of CV conditions, presence of respiratory disease, presence of diabetes mellitus/insulin dependent diabetes mellitus, number of SEWB conditions, presence of a musculoskeletal condition, number of chronic conditions, number of CV risk factors and gender. Presence of DM/IDDM, musculoskeletal, respiratory disease and number of SEWB conditions were chosen because of their overrepresentation within this population. SEWB is often used interchangeably with 'mental health', however, is a much more multifaceted concept with particular meaning for Aboriginal and Torres Strait Islander people \\[[@CR42]\\]. While the term SEWB includes mental health conditions, it also encompasses broader aspects of health and wellbeing such as connection to land or Country, culture, family, kinship and community \\[[@CR42]\\]. Covariates considered included number of days between baseline and follow-up measurement, number of visits between baseline and follow-up, average sessions attended per week, baseline measurement and age. A type I error of less than 0.05 was considered to represent statistical significance for all tests. Due to the small sample size, a maximum of 2 independent variables were included in each model, with those selected having the highest explanatory power. Co-linearity between variables was assessed using the Variance Inflation Factor (VIF); all results had a VIF\u00a0\\<\u00a01.4 indicating co-linearity is not a problem in the models. All analysis were performed using *IBM SPSS for Windows Release 22 (SPSS)*.\n\nAs this is a small and potentially recognisable cohort of participants, confidentiality of study participants must be upheld. Thus, the actual sample size of two sub-groups in each clinical outcome measure have been replaced with the labels \"*n*\u00a0\u2264\u00a010\" or \"*n*\u00a0\u2265\u00a010\" when reporting outcomes. Statistical tests were not conducted on sub-groups where *n*\u00a0\u2264\u00a010.\n\nResults {#Sec7}\n=======\n\nA total of 315 participants had baseline clinical outcome measurements recorded upon initial assessment by an accredited exercise physiologist or physiotherapist. Of this sample, 85 people with (*n*\u00a0=\u00a024) or at risk (*n*\u00a0=\u00a061) of CV disease had at least one paired clinical outcome baseline and follow-up measurement recorded, thus excluding 230 participants from analysis.\n\nParticipant characteristics for the sample who had only a baseline measurement recorded (*n*\u00a0=\u00a0230) were compared to those for participants with both baseline and follow-up measurement (*n*\u00a0=\u00a085). Those that had a paired baseline and follow-up measurement were an average of 5.14\u00a0years older (95%CI: 1.56 to 8.73) than those that had a baseline measurement only (*t*312\u00a0=\u00a02.82, *p*\u00a0=\u00a00.005). Participants did not differ between groups in gender, marital status, employment status or baseline CV disease risk factor characteristics.\n\nFor the sample included in this study (n\u00a0=\u00a085), presenting with more than one CV condition was common, with 11 (45.85%) participants having 2--3 CV conditions. The median length of time between baseline and follow-up assessment was 11.85\u00a0weeks (IQR\u00a0=\u00a06.57--26.64\u00a0weeks) and the median number of sessions attended per person during this period 11.00 (IQR\u00a0=\u00a06.00--19.00). The mean age of participants was 55.26\u00a0years (SD\u00a0=\u00a013.86\u00a0years) and the majority (*n*\u00a0=\u00a061, 71.80%) were female. SEWB conditions were common amongst the 85 participants, with 35 (41.20%) presenting with one SEWB condition and 13 (15.30%) having two-three SEWB conditions.\n\nAt baseline 76.54% of the participants with both a pre and post measurement were classified as obese or extremely obese. The mean BMI was 36.89 (SD\u00a0=\u00a09.65). Almost one third of participants (*n*\u00a0=\u00a018) had high systolic BP and 91.14% had a waist circumference categorised as having substantially increased risk of metabolic complications. There was high variability in the distance walked in the 6MWT, with participants covering an average of 0.38\u00a0km and individual values ranging from 0.07 to 0.95\u00a0km.\n\nAt follow-up participants with matched baseline line data showed small, non-significant decreases in mean weight, BMI, waist circumference, hip circumference, systolic BP, and diastolic BP, and a slight increase in WHR between baseline and follow-up (Table\u00a0[1](#Tab1){ref-type=\"table\"}). The distance walked in the 6MWT significantly increased between baseline and follow-up with a mean change of 0.053\u00a0km (95% CI: 0.01 to 0.07, *p*\u00a0=\u00a00.023) (Table [1](#Tab1){ref-type=\"table\"}). Significant changes were identified for specific participant sub-groups. The *extreme obesity* BMI group demonstrated a significant reduction in weight of 1.57\u00a0kg (95% CI: 0.10 to 3.03, *p*\u00a0=\u00a00.037)*.* For both systolic and diastolic BP, participants with a baseline measurement recorded in the *normal* range, had a significantly higher measurement at follow-up. For systolic BP only, participants in the *high* range at baseline demonstrated a significant decrease at follow-up of 11\u00a0mmHg (95% CI: 3.18 to 18.82, *p*\u00a0=\u00a00.009) (Table [1](#Tab1){ref-type=\"table\"}).Table 1Summary of changes in clinical outcome measures for individuals with baseline and first follow-up assessmentClinical Outcome Measure*n*Mean Change (95% CI)Baseline Measurement (SD)Follow-up Measurement (SD)*p-value\u2020*Weight (kg)65\u22120.57 (\u22121.28, 0.14)98.90 (27.97)98.32 (27.65)0.113Weight within BMI Groups (kg)\u2003 *Healthy*\u2264100.26 (\u22123.30, 3.82)66.53 (7.84)66.80 (8.13)--\u2003 *Overweight*\u226510\u22120.02 (\u22121.12, 1.08)76.35 (9.76)76.33 (10.40)0.974\u2003 *Obese I*15\u22121.21 (\u22123.11, 0.69)92.39 (12.37)91.17 (11.11)0.194\u2003 *Obese II*120.85 (\u22120.46, 2.15)95.64 (9.67)96.48 (9.67)0.181\u2003 *Extreme Obesity*19\u22121.57 (\u22123.03, \u22120.10)130.55 (27.23)128.98 (28.09)0.037\\*BMI (kg/m^2^)64\u22120.20 (\u22120.47, 0.07)35.85 (8.92)35.65 (8.80)0.15Systolic BP (mmHg)58\u22120.71 (\u22124.85, 3.43)128.21 (16.84)127.50 (16.31)0.734Systolic BP within BP Categories (mmHg)\u2003 *Normal*\u2265108.19 (3.45, 12.93)110.29 (6.41)118.48 (10.06)0.002\\*\u2003 *High/Normal*\u226510\u22120.79 (\u22128.17, 6.59)129.21 (4.88)128.42 (14.82)0.825\u2003 *High*18\u221211.00 (\u221218.82, \u22123.18)148.06 (7.87)137.06 (18.49)0.009\\*Diastolic BP (mmHg)58\u22120.12, (\u22123.49, 3.25)80.50 (10.17)80.38 (11.29)0.943Diastolic BP within BP Categories (mmHg)\u2003 *Normal*\u2265105.31 (1.59, 9.04)73.63 (6.65)78.94 (9.55)0.007\\*\u2003 *High/Normal*\u226510\u22125.87 (\u221213.48, 1.74)85.13 (2.87)79.25 (13.88)0.121\u2003 *High*9\u22128.11 (\u221216.08, \u22120.14)93.33 (4.06)85.22 (10.12)--Waist circumference (cm)60\u22120.58 (\u22123.06, 1.89)116.32 (19.86)115.73 (18.98)0.63Waist circumference within Waist circumference groups (cm)\u2003 *No increased risk*\u2264105.33 (\u22125.01, 15.68)93.66 (10.06)88.33 (8.08)--\u2003 *Increased risk*\u2264101.77 (\u221214.15, 17.69)84.07 (3.00)85.83 (4.54)--\u2003 *Substantially increased risk*54\u22121.05 (\u22123.74, 1.65)119.7 (17.94)118.62 (17.61)0.44Hip circumference (cm)57\u22120.65 (\u22122.55, 1.25)123.57 (18.39)122.9 (19.80)0.49WHR (waist/hip)570.0007 (\u22120.024, 0.025)0.94 (0.093)0.95 (0.091)0.95WHR within WHR Groups (waist/hip)\u2003 *No increased risk*\u2264100.108 (\u22120.001, 0.219)0.81 (0.022)0.92 (0.12)--\u2003 *Substantially increased risk*\u226510\u22120.016 (\u22120.038, 0.004)0.97 (0.080)0.95 (0.086)0.126MWT (km)470.053 (0.01, 0.07)0.38 (0.18)0.43 (0.16)0.023\\**\u2020 Paired t-test used to compare baseline and follow-up measurement. Statistical testing only conducted when n\u00a0\\>\u00a010 in each group*\\**Significant at p\u00a0\\<\u00a00.05*\n\nAfter adjusting for baseline measurements, statistically significant models predicting positive changes in 6MWT and systolic and diastolic BP were developed. Participants who walked further in their 6MWT at baseline, saw smaller changes in their follow-up 6MWT distance. The best fitting model for change in 6MWT adjusted for baseline 6MWT, included number of SEWB conditions as a factor (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). This model predicted that after adjusting for baseline 6MWT measurement, those individuals who had two or three SEWB conditions walked an average of an additional 0.143\u00a0km (95%CI: 0.05 to 0.23, *p*\u00a0=\u00a00.002) at follow-up, compared to those with no SEWB conditions (Fig. [1](#Fig1){ref-type=\"fig\"}).Fig. 1Relationship between baseline 6MWT and change in 6MWT (follow-up -- baseline), and predicted regression equations. Fig. 1 legend: blue circle 0 SEWB Conditions, green circle 1 SEWB Condition, red circle 2--3 SEWB Conditions\n\nIt was noted that participants with two or three SEWB conditions had a lower mean weight and walked substantially further at baseline (mean weight\u00a0=\u00a087.74\u00a0kg, mean baseline 6MWT distance\u00a0=\u00a00.55\u00a0km), compared to those that had one SEWB condition (mean weight\u00a0=\u00a0104.33\u00a0kg, mean baseline 6MWT distance\u00a0=\u00a00.38\u00a0km) and those that presented with no SEWB conditions (mean weight\u00a0=\u00a0100.83\u00a0kg, mean baseline 6MWT distance\u00a0=\u00a00.39\u00a0km).\n\nLike the 6MWT, the largest decreases in systolic and diastolic BP between baseline and follow-up occurred in those participants with the highest baseline readings. After adjusting for baseline measurement, change in diastolic BP was found to be influenced by the number of CV conditions and age (Fig. [2](#Fig2){ref-type=\"fig\"}). Participants with one CV condition experienced an average decrease in diastolic BP which was 10.6\u00a0mmHg larger (95% CI: 3.3 to 18.0\u00a0mmHg, *p*\u00a0=\u00a00.004) than participants with no CV conditions, after adjusting for baseline diastolic BP and age (Fig. [2](#Fig2){ref-type=\"fig\"}). Similarly, the model for change in systolic BP showed that after adjusting for the baseline reading, participants with one CV condition experienced a 10.77\u00a0mmHg (95% CI:0.40 to 21.13, *p*\u00a0=\u00a00.042) larger decrease in systolic BP, when compared to those who didn't present with a CV condition. Further, the model for diastolic BP showed that for every one year increase in age, the change in diastolic BP is expected to decrease by 0.26\u00a0mmHg (95% CI: 0.09 to 0.43\u00a0mmHg, *p*\u00a0=\u00a00.002) after adjusting for number of CV conditions and baseline diastolic BP (Fig. [2](#Fig2){ref-type=\"fig\"}).Fig. 2Relationship between age and change in diastolic blood pressure (follow-up -- baseline), and predicted regression equations. Fig. 2 legend: blue circle 0 CV Conditions, green circle 1 CV Condition, red circle 2+ CV Conditions\n\nDiscussion {#Sec8}\n==========\n\nThe participants in the Work it Out program are purposely selected to have at least one chronic condition or risk factor, thereby automatically selecting a portion of the larger Aboriginal and Torres Strait Islander population with poorer health. However the baseline clinical outcome measures indicate a population that has a substantial burden of ill health with multiple conditions and/or risk factors and a high level of complex health needs. It appears this urban south-east Queensland Aboriginal and Torres Strait Islander cohort has similar characteristics to cohorts accessing CV rehabilitation, prevention and CCSM programs in rural, regional and urban areas of other Australian states, specifically Tasmania, Western Australia and South Australia \\[[@CR11], [@CR12], [@CR34], [@CR35]\\].\n\nThe analysis undertaken during this study demonstrated that statistically significant changes in some clinical outcome measures were detected after approximately 12\u00a0weeks in the *Work It Out* program for people with or at risk of CV disease. Results indicated that changes in clinical outcome measures were seen in measures that were more responsive to exercise within a shorter time period, and in participants with a worse starting condition. This is emulated in a Tasmanian cardiopulmonary rehabilitation and secondary prevention program \\[[@CR12]\\] which recognised the difficulty in achieving losses in weight, BMI and waist circumference over an eight week period, noting small effect sizes. Correspondingly, the 6MWT is proven to be more responsive to change over a shorter time period, with more modest changes seen over longer periods \\[[@CR43]\\].\n\nThe reduction in systolic BP for participants in the *high* category of 11\u00a0mmHg is notable and clinically relevant, and is greater than that experienced in both a similar cohort \\[[@CR21]\\] and a meta-analysis of conventional cardiac rehabilitation programs \\[[@CR44]\\]. It is echoed however, in results from a similar cardiac rehabilitation program \\[[@CR11]\\] which indicated that along with lifestyle changes and medication compliance, a higher measure of BP at baseline may correspond to a greater loss in the short term, as demonstrated in this study. A reduction in systolic BP may have particular clinical importance for this group, as clinical trials have shown that control of systolic hypertension specifically, reduces total mortality, CV mortality, stroke and the occurrence of other heart failure events \\[[@CR45], [@CR46]\\].\n\nThe statistical models that were created for both systolic and diastolic BP showed that both appeared to be influenced by the number of CV conditions, after adjusting for their baseline measurement. It is possible that this points towards prescribed intensity of exercise, with participants who presented with two or more CV conditions more likely prescribed a reduced exercise intensity than those who presented with one CV condition, thus potentially explaining the smaller changes at follow-up assessment. The effect of medication could also be important, with participants who presented with multiple CV conditions more likely to have their BP regulated by medication and thus show a smaller effect from exercise and lifestyle modification. Lastly, this outcome could be influenced by motivation. That is, participants with one CV condition may be more motivated to make lifestyle changes, compared to those who have numerous CV conditions, who may find it more challenging. The exact reason for the influence of number of CV conditions on change in systolic and diastolic BP in this population is an area for further research.\n\nThe statistical model also demonstrates the impact of age on diastolic BP. It is likely this is due to the natural progression of diastolic BP, which rises until approximately age 50 and tends to level off or fall in later life \\[[@CR47], [@CR48]\\].\n\nThe 15% improvement in 6MWT is consistent with what is noted in conventional short-term outpatient cardiac rehabilitation programs \\[[@CR49]\\] and points to the effect of regular exercise on every day physical functioning. The model indicated that change in 6MWT appeared to be influenced by number of SEWB conditions, with the overall pattern of participants who walked further at baseline seeing the smallest improvement at follow-up. However, this trend did not hold for those experiencing several SEWB conditions who also had a smaller mean weight compared to those with one or no SEWB conditions. This may suggest that those participants with two or three SEWB had the ability to walk further at baseline because they were leaner and thus more physically able. However, while weight is likely to influence the 6MWT result at baseline, it cannot explain the differential improvements in 6MWT at follow-up. What this may reveal then, is the social benefit gained from attending the program and/or the physiological improvements in depressive symptoms that are often associated with exercise.\n\nThe presence of psychosocial stress is linked to a lack of motivation in carrying out work or normal activities for Aboriginal and Torres Strait Islander people \\[[@CR50]\\]. Hence, we propose that participants who began the program with two or three SEWB conditions, could have been significantly lacking in motivation, and that their superior performance in the 6MWT at baseline compared to other participants was due to their better physical condition (i.e. lighter body weight). A separate qualitative study involving a small sample of *Work It Out* participants (*n*\u00a0=\u00a022), identified social connectedness as a strong theme, with reports of \"having a place to meet others,\" connecting and re-connecting with family and friends and a shared responsibility to help each other emphasised \\[[@CR51]\\]. So, we hypothesise that as participants with numerous SEWB conditions continued to attend the program, they benefited from this important social support, thus increasing motivation levels. This increase in motivation may have manifested as a greater change in the 6MWT, as this is proven to be highly responsive to exercise in a short term period \\[[@CR43]\\]. Moreover, qualitative analysis revealed overwhelming reports of decreases in anxiety and depression in participants due to participation in the program \\[[@CR51]\\]. This is supported by literature which shows that for participants with diagnosed depression, physical exercise in a group setting may substantially increase 6MWT distance, improve quality of life and depressive symptoms, and increase perceptions of social support when compared to an intervention of traditional cognitive behavioural therapy \\[[@CR52], [@CR53]\\]. This highlights the potential physiological effects of exercise for participants with numerous SEWB conditions.\n\nAlthough the distance walked in six minutes may be seen as an objective measure of physical capacity, the potential flow-on effects for individuals with two or three SEWB could be considerable. It is important too, to examine the meaning of this physical increase in 6MWT in the context of Aboriginal and Torres Strait Islander people participating in a community based health program. For Aboriginal and Torres Strait Islander people, walking an extra 143\u00a0m may be reflected in an increase in the ability to physically participate more wholly in community, interact with family and carry out cultural activities. More research determining the exact effects of the *Work It Out* program on participants with SEWB conditions is warranted.\n\nThis study has several limitations which need to be considered. First, only changes that occurred over an approximate 12\u00a0week period were assessed. Further analysis is required over a longer time period in order to see if current benefits are sustained and/or positive trends in changes to clinical outcome measures continue. Second, the bivariate results of this study are likely impacted by regression to the mean and should be viewed with caution. However, the impact of regression to the mean was reduced in the modelling but the inclusion of the baseline measurement in the model. Hence the modelling results reported are more likely representative of true covariate effects compared to the simpler bivariate analysis. Further, the statistical analysis and power was restricted due to the small sample size, particularly in certain sub-groups. This is a result of the *Work It Out* program being relatively new and also a lack of follow-up data. It is recommended that a similar analysis be conducted once more participants have progressed through the program. Furthermore, the exclusion of participants without a follow-up assessment may offer some bias to results. However, as age is the only characteristic identified that differs between the included and excluded cohorts, this is unlikely to be significant. As an observational study, this project is subject to the potential limitations of this study design. These include a lack of control group and an inability to establish definitive cause and effect relationships. However, the overall outcome is the impact of the *Work It Out* program and how this has influenced behaviour change in participants' everyday lives, so an observed positive change is deemed reflective of the program's success. Lastly, the diversity of Aboriginal and Torres Strait Islander communities must be considered, and thus, caution must be exercised when generalising these outcomes beyond urban south-east Queensland.\n\nConclusions {#Sec9}\n===========\n\nEvaluation of the *Work It Out* program over an approximate 12\u00a0week period found statistically significant improvements for those participants who began the program with high systolic BP; those whose BMI was classified as extremely obese; and, in the 6MWT. These findings suggest that short-term improvements were seen in those clinical outcome measures that were more responsive to exercise and behaviour change within shorter time periods. This could indicate a trend for improvements in other clinical outcome indicators over the longer term. It was noted that there were increased benefits with particular to those participants suffering numerous SEWB conditions and the reasons for this require further investigation. Whilst evaluation of other aspects of the *Work It Out* program are required, this study shows that the program could prove a useful model for CV rehabilitation and prevention for other urban Aboriginal and Torres Strait Islander populations.\n\n6MWT\n\n:   Six Minute Walk Test\n\nAHW\n\n:   Aboriginal and Torres Strait Islander Health Worker\n\nATSICCHS\n\n:   Aboriginal and Torres Strait Islander Community Controlled Health Service\n\nBMI\n\n:   Body Mass Index\n\nBP\n\n:   Blood Pressure\n\nCCSM\n\n:   Chronic Condition Self-Management\n\nCV\n\n:   Cardiovascular\n\nIQR\n\n:   Interquartile Range\n\nIUIH\n\n:   The Institute for Urban Indigenous Health\n\nSEWB\n\n:   Social and Emotional Wellbeing\n\nWHR\n\n:   Waist to Hip Ratio\n\nThe researchers would like to acknowledge the Traditional Owners and Custodians of the land where this research has taken place and thank all participants for their involvement in the *Work it Out* program and this research.\n\nFunding {#FPar1}\n=======\n\nThis research was supported by the National Heart Foundation, through a National Heart Foundation Australian Indigenous Scholarship. The funders had no role in the design of the study, or the analysis and interpretation of the results.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nConsent for publication of raw data not obtained and dataset could in theory pose a threat to confidentiality. The data were extracted from routine measurements as part of program evaluation. As the *Work It Out* program encompasses a relatively small and recognisable population, it is not possible to completely de-identify data. To maintain confidentiality of study participants and ensure that participants cannot be identified raw data is not available.\n\nAuthor's contributions {#FPar3}\n======================\n\nKM designed the study, prepared the data, conducted the data analysis and interpretation and wrote the manuscript. MLG participated in the design of the study and supervised the data analysis and interpretation, and reviewed the manuscript. RM participated in the design of the study and reviewed the manuscript. AN participated in the design of the study and reviewed the manuscript. All authors read and approved the final version of the manuscript.\n\nEthics approval and consent to participate {#FPar4}\n==========================================\n\nEthical approval for research of the *Work It Out* program was obtained from the *University of Queensland's Human and Behavioural Research Ethics Committee*, approval number: *2,011,001,283*. Approval from the Institute for Urban Indigenous Health's Board of Directors, comprised of community leaders as well as medical, community health, Aboriginal and Torres Strait Islander health and academic professionals, was gained for use of de-identified data for the purposes of secondary data analysis for this project. Ethics approval for the analyses of this secondary data was then cleared by the *Human Research Ethics Committee* at *Queensland University of Technology*, approval number: *130,000,692*.\n\nUpon entry to the *Work It Out* program participants were counselled in plain English about the research project and only those who gave signed, informed consent were included.\n\nConsent for publication {#FPar5}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar6}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar7}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Background\n==========\n\nHuman milk oligosaccharides (HMOs) are known to be the most relevant factor for the development of intestinal microbiota in breast-fed infants \\[[@B1]\\]. Also, HMOs have been reported to play important roles in preventing adhesion of pathogens and toxins to epithelial surfaces \\[[@B2]\\]. Fucosyloligosaccharides, such as 2\u2032-fucosyllactose, lacto-*N*-fucopentaose and lacto- *N*-difucohexaose, are common HMOs. Fucosylated oligosaccharides act as growth stimulating factors for select Bifidobacteria and soluble analogs of receptors for pathogenic bacteria, thereby protecting infants against infection from enteric pathogens and binding of toxins \\[[@B3],[@B4]\\]. Specifically, \u03b1-1,2-linked fucosylated oligosaccharides are reported to exhibit protective activity against several pathogens including *Campylobacter jejuni*\\[[@B3],[@B5]\\], *Salmonella* enteric serotype *Typhimurium*\\[[@B6]\\], Enterotoxigenic *E. coli*\\[[@B7]\\], *Helicobacter pylori*\\[[@B8]\\] and noroviruses \\[[@B9]\\]. Among them, 2\u2032-fucosyllactose (2-FL) is the most abundant fucosyloligosaccharide in human milk and accounts for more than 30% of total HMOs \\[[@B3],[@B5]\\]. Low levels of 2-FL in the milk of sore mothers have been reported to be associated with a higher rate of diarrhea in breast-fed infants \\[[@B3]\\]. Hence, 2-FL is a promising oligosaccharide for nutraceutical and pharmaceutical purposes.\n\n2-FL can be synthesized through the enzymatic fucosylation of lactose by \u03b1-1,2 fucosyltransferase (FucT2), which requires guanosine 5\u2032-diphosphate (GDP)-[l]{.smallcaps}-fucose as a donor of [l]{.smallcaps}-fucose \\[[@B10]\\]. *Escherichia coli* is known to be able to synthesize GDP- [l]{.smallcaps}-fucose since GDP- [l]{.smallcaps}-fucose is used for biosynthesis of colanic acid, one of the main components of the cell wall \\[[@B11]\\]. Therefore, 2-FL can be produced via engineering of the GDP-[l]{.smallcaps}-fucose biosynthetic pathway and overexpression of the fucosyltransferase gene in metabolically engineered *E*. *coli*. Figure [1](#F1){ref-type=\"fig\"} shows the metabolic pathway for biosysnthesis of GDP-[l]{.smallcaps}-fucose and 2-FL in recombinant *E*. *coli*.\n\n![**The metabolic pathway for GDP--fucose and 2\u2032-fucosyllactose (2-FL) biosynthesis in recombinant*E. coli*.** The names of enzymes are abbreviated as follows; ManA, mannose 6-phosphate isomerase; ManB, phosphomannomutase; ManC, mannose 1-phosphate guanylyltransferase; Gmd, GDP- [d]{.smallcaps}-mannose-4,6-dehydratase; WcaG, GDP-4-keto-6-deoxymannose 3,5-epimerase 4-reductase; FucT2, \u03b1-1,2-fucosyltransferase. Pi, GDP and GTP denote phosphate, guanosine 5\u2032-diphosphate and guanosine 5\u2032-triphosphate.](1475-2859-11-48-1){#F1}\n\nPreviously, biosynthesis of fucosyloligosaccharides using a recombinant microorganism and fucosyltransferase has been reported. The enzymatic synthesis of 2-FL was examined by using purified FucT2, GDP-[l]{.smallcaps}-fucose and lactose \\[[@B10]\\], however, the high cost of GDP-[l]{.smallcaps}-fucose and FucT2 purification may be a limiting factor for large-scale production of fucosyloligosaccharides. Production of several fucose-containing lacto-oligosaccharides in recombinant *E. coli* was also reported through simultaneous overexpression of fucosyltransferase and the regulatory protein for colanic acid biosynthesis \\[[@B12],[@B13]\\], which suggested that whole cell synthesis of fucosyloligosaccharides through direct amplification of the GDP-[l]{.smallcaps}-fucose biosynthesis might be feasible.\n\nTo construct an efficient 2-FL production system by metabolic engineering, an understanding and detailed analysis of a cellular metabolic network involved in the 2-FL biosynthesis is important. Elementary flux mode (EFM) analysis has emerged as a powerful tool for metabolic pathway analysis. EFM analysis is a useful mathematical tool for defining and describing all metabolic routes that are both stoichiometrically and thermodynamically feasible for a group of enzymes. The EFM analysis can decompose a complex metabolic network of many highly interconnected reactions into uniquely organized pathways that support steady state of metabolism. EFM analysis can provide identification of all genetically independent pathways, determination of the most efficient physiological state of a cell, and analysis of metabolic network properties such as robustness and regulation \\[[@B14]-[@B16]\\]. Hence, it can be a useful tool for understanding dynamics of cellular metabolism and rational design of the host strain's metabolism for 2-FL production.\n\nWe have previously developed a recombinant *E*. *coli* system for efficient production of GDP- [l]{.smallcaps}-fucose by metabolic engineering. An enhancement of GDP- [l]{.smallcaps}-fucose production was achieved by modulation of several factors for biosynthesis of GDP- [l]{.smallcaps}-fucose such as amplification of GDP- [d]{.smallcaps}-mannose biosynthesis, regeneration of NADPH and manipulation of the guanosine nucleotides biosynthetic pathway \\[[@B17]-[@B19]\\].\n\nIn the present study, the GDP-[l]{.smallcaps}-fucose production system was applied for efficient production of 2-FL by introduction of the FucT2 gene from *Helicobacter pylori* into the recombinant *E. coli* able to overproduce GDP- [l]{.smallcaps}-fucose. Whole cell biosynthesis of 2-FL from lactose was assessed in a series of batch fermentations for recombinant *E. coli* overexpressing the necessary genes for GDP- [l]{.smallcaps}-fucose production and the FucT2. An EFM analysis for 2-FL production in the recombinant *E. coli* was used to compare and evaluate experimental results.\n\nMethods\n=======\n\nStrains and plasmids\n--------------------\n\n*E. coli* TOP10 \\[F- *mcr*A \u0394( *mrrhsd*RMS- *mcr*BC) \u03c680 *lac*Z\u0394M15 \u0394 *lac*X74 *rec*A1 *ara*D139 \u0394( *araleu*) 7697 *gal*U *gal*K *rps*L (Str^R^) *end*A1 *nup*G\\] was used for genetic manipulation. *E*. *coli* BL21star(DE3) \\[F^\u2212^*omp*T, *hsd*SB(r~B~^\u2212^m~B~^\u2212^), *galdcm rne131* (DE3)\\] (Invitrogen, Carlsbad, CA, USA) and JM109(DE3) *end*A1 *gln*V44 *thi*-1 *rel*A1 *gyr*A96 *rec*A1 *mcr*B^+^ \u0394(*lac-pro*AB) e14- \\[F\\' *tra*D36 *pro*AB^+^*lacI*^q^*lac*Z\u0394M15\\] *hsd*R17(r~K~^-^m~K~^+^) (DE3)\\] (NEB, Ipswich, MA, USA) were used for production of GDP-[l]{.smallcaps}-fucose and 2-FL. Plasmid pmBCGW containing the polycistronic *gmdwcaG* gene cluster and *manBmanC* gene cluster was previously constructed using plasmid pETDuet-1 \\[[@B18]\\]. The gene encoding FucT2 was obtained by the polymerase chain reactions (PCR) using the genomic DNA of the *Helicobacter pylori* 26695 strain (ATCC 700392) as template \\[[@B20]\\]. Two PCR primers of fucT2_F and fucT2_R were used for the amplification of the *fucT2* gene. After digestion of PCR fragments of the *fucT2* gene and pCOLADuet-1 (Merck Biosciences, Darmstadt, Germany) with *Nco*I and *Sac*I, the DNA fragments were ligated to construct plasmid pHfucT2. Plasmids and primers used in this work are listed in Table [1](#T1){ref-type=\"table\"}. The constructed plasmid was confirmed by DNA sequencing. The conditions for PCR reaction, DNA manipulation and bacterial transformation followed the descriptions in the previous study \\[[@B21]\\].\n\n###### \n\nList of primers and plasmids used in this study\n\n  **Name**           **Sequence of PCR primers and description for plasmids**                                                       **Source**\n  ------------------ -------------------------------------------------------------------------------------------------------------- -----------------------------------\n  PCR primers        \u00a0                                                                                                              \u00a0\n  fucT2_F (*Nco*I)   5\u2032-ACATG[CCATGG]{.ul}CTTTTAAGGTGGTGCAA-3\u2032                                                                      *H*. *pylori* 26695 (ATCC 700392)\n  fucT2_R (*Sac*I)   5\u2032-AGTCC[GAGCTC]{.ul}TTAAGCGTTATACTTTTGGGA-3\u2032                                                                  \n  Plasmids           \u00a0                                                                                                              \u00a0\n  pETDuet-1          two T7 promoters with two MCS, pBR322 replicon (copy number \\~40), Amp^r^                                      Merck Biosciences\n  pCOLADuet-1        two T7 promoters with two MCS, ColA replicon (copy number 10\u2009\\~\u200912), Kan^r^                                    Merck Biosciences\n  pmBCGW             derived from pETDuet-1, P~T7~-*manB-manC* ( *Nco*I/ *Sac*I)-P~T7~-*gmd-wcaG* ( *Nde*I/ *Xho*I)-T~T7~, Amp^r^   Lee et al., 2009\n  pHfucT2            derived from pCOLADuet-1, P~T7~- *fucT2* ( *Nco*I/ *Sac*I)-P~T7~-MCS2-T~T7~, Kan^r^                            this study\n\nBatch fermentation\n------------------\n\nBatch fermentation was carried out in a 250\u2009ml flask containing 50\u2009ml of LB medium at 25\u00b0C and pH 6.8. Agitation speed was maintained at 250\u2009rpm. When dry cell mass reached 0.3\u2009g/l, 0.1\u2009mM isopropyl-\u03b2-[d]{.smallcaps}-thiogalactopyranoside (IPTG) was added to culture broth. After 3\u2009h of additional cultivation, 2.6\u2009g/l (or 14.5\u2009g/l) lactose was added for 2-FL production.\n\nAnalytical methods\n------------------\n\nCell concentration was measured by optical density (OD) at 600\u2009nm using a spectrophotometer (Biomate 5, Thermo, NY, USA). Overexpression of FucT2 inside the cell was analyzed by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, 12% polyacrylamide). After 3\u2009h of 0.1\u2009mM IPTG induction, cells were collected and the concentration was adjusted to around 7.2\u2009g/l. They were resuspended in 50\u2009mM potassium phosphate buffer (pH 7.0) and disrupted by an ultrasonic processor. After centrifugation at 15,000\u2009\u00d7\u2009g for 20\u2009min, the supernatant (soluble fraction) and debris (insoluble fraction) were separated. Ten microliters of the soluble protein fraction (approximately 0.04\u2009mg) and the same volume of the total and insoluble protein fractions were subjected to SDS-PAGE. Gels were stained with Coomassie brilliant blue solution and images were analyzed using a densitometer.\n\nConcentrations of lactose, 2-FL and acetate in batch fermentations were determined by using a high performance liquid chromatography (HPLC) system (Agilent Technologies 1200 Series) equipped with a Rezex ROA Organic Acid H^+^ column (Phenomenex, Torrance, CA, USA) and a refractive index (RI) detector (Agilent, Palo Alto, CA, USA). The column was eluted with 0.01\u2009N H~2~SO~4~ at a flow rate of 0.6\u2009ml/min at 50\u00b0C.\n\nIn order to confirm 2-FL biosynthesis, culture broth at the end of the batch fermentation was collected and analyzed using a liquid chromatography/mass spectrometry (LC/MS) system. The LC (Agilent Technologies 1100 Series) was equipped with an Agilent Zorbax Eclipse ZDB-C8 (4.6x150 mm, 5 \u03bcm) column and an Agilent LC/MSD Trap XCT Plus detector. The column was eluted at a flow rate of 0.4\u2009ml/min by the following gradient program: 95% (v/v) eluent A (15\u2009mM ammonium acetate) and 5% eluent B (acetonitrile) for 1\u2009min; 5% to 95% eluent B over 6\u2009min; 95% eluent B over 10\u2009min. The scan range for MS was 70--600 mass-to-charge ratio (m/z).\n\nConstruction of metabolic network model for *E. coli* producing 2-FL from lactose\n---------------------------------------------------------------------------------\n\nA metabolic network model was constructed for 2-FL producing *E*. *coli* that grows on lactose. The *E. coli* network was based on a model that was introduced by Stelling et al. \\[[@B22]\\] to examine the relationship between structure and function in metabolic networks. Furthermore, the model has been used for calculating elementary flux modes in previous reports \\[[@B23],[@B24]\\]. The metabolic network was composed of 108 reactions, which were involved in carbon central metabolism, amino acid synthesis, fatty acid synthesis and biomass production (Additional file [1](#S1){ref-type=\"supplementary-material\"}). The catabolic part of the model included substrate uptake reactions, glycolysis, pentose phosphate pathway, TCA cycle, and excretion of by-products (e.g. acetate, formate, lactate, and ethanol). Previous networks were extended to include the anaplerotic reactions (e.g. malic enzyme and pyruvate oxidase) in addition to parallel pathways for initial acetate metabolism. The anabolic part of the model covers the conversion of precursors into building blocks like macromolecules and biomass. The core *E. coli* model from Stelling et al. \\[[@B22]\\] was modified in this research to account for lactose consumption and synthesis of 2-FL. Among the reactions added for 2-FL synthesis, some minor adjustments were made to simplify the model. Lactose was assumed to break down to 2 moles of glucose because galactose can be easily converted into glucose-6-phosphate. ATP was used in place of GTP for energy transfer. As for the mass balance, it should be noted that ADP is formed whenever ATP is consumed for all the metabolic reactions in the network. The mass balance equation on ATP is therefore the negative of the mass balance on ADP and thus the two equations are linearly dependent. Therefore, ADP can be excluded from the model in order to simplify the subsequent EFM calculation. The same is true for other cofactor pairs like NADP/NADPH and NAD/NADH. The EFM pathways in the model were estimated using METATOOL 5.1 \\[[@B14],[@B16]\\] with Matlab.\n\nResults\n=======\n\nExpression of \u03b1-1,2-fucosyltransferase (FucT2) in recombinant *E. coli*\n-----------------------------------------------------------------------\n\nThe expression pattern of FucT2 was investigated during a batch fermentation of recombinant *E*. *coli* harboring plasmid pHfucT2. In order to maximize the expression of the soluble form of FucT2 in the recombinant *E. coli*, 0.1\u2009mM of IPTG was used. As shown in Figure [2](#F2){ref-type=\"fig\"}, a 33\u2009kDa protein (consistent with FucT2, \\[[@B20]\\]) was found in both soluble and insoluble fraction. While a significant amount of FucT2 was expressed in inclusion bodies, biosynthesis of 2-FL was expected because a soluble form of FucT2 was available as well.\n\n![**SDS-PAGE analysis of the cell crude extract of recombinant*E. coli*BL21star(DE3) strains harboring pCOLADuet-1 and pHfucT2, respectively.** Cells were harvested after 3\u2009h of 0.1\u2009mM IPTG induction. T, S and I denote total, soluble and insoluble protein fractions, respectively. The arrow indicates the corresponding protein band with the estimated molecular weight of FucT2. Lane M indicates size marker.](1475-2859-11-48-2){#F2}\n\nBatch fermentations\n-------------------\n\nFrom the preliminary experiments (whole cell bioconversion of 2\u2009g/l lactose with *E*. *coli* BL21star(DE3) strain), we concluded that the *E*. *coli* BL21star(DE3) strain is not beneficial for 2-FL production because it consumed lactose for growth and maintenance instead of converting to 2-FL (data not shown). Some *E*. *coli* strains, such as DH5\u03b1 and JM series, are known to be unable to assimilate lactose or utilize lactose extremely inefficiently due to partial deletion of the *lac*Z gene, which codes for \u03b2-galactosidase. As such, these *E. coli* strains might be useful for 2-FL production. Hence, *E*. *coli* JM109(DE3) enabling overexpressing proteins under the control of *T7* promoter was chosen as an alternative host stain for 2-FL production.\n\n2-FL production for BL21star(DE3) and JM109(DE3) was compared under batch fermentation conditions. In order to allow sufficient production of both GDP-[l]{.smallcaps}-fucose biosynthetic enzymes and FucT2 inside the cells, the cells were cultivated for 3 h after 0.1\u2009mM IPTG induction. Then, 2.6\u2009g/l of lactose was added to initiate 2-FL production without addition of additional sugar because GDP- [l]{.smallcaps}-fucose can be produced from LB media \\[[@B17],[@B18]\\]. During the fermentations, extracellular 2-FL production (in the medium) was monitored by HPLC analysis. As a result, a small amount of 2-FL (10\u2009mg/l) was produced in the batch fermentation of recombinant *E*. *coli* BL21star(DE3). Meanwhile, much higher amount of 2-FL was produced in the batch fermentation of recombinant *E. coli* JM109(DE3). About 140\u2009mg/l of 2-FL was produced from 2.6\u2009g/l of lactose while 0.4\u2009g/l of lactose remained unused at the end of the fermentation (data not shown). These results suggest that the lactose concentration should be controlled at more than 0.5\u2009g/l to maintain 2-FL production. Consequently, a yield of 60\u2009mg 2-FL/g lactose was obtained from the batch fermentation of *E*. *coli* JM109(DE3) when 2.6\u2009g/l of lactose was used. In order to obtain a higher amount of 2-FL, a batch fermentation with a higher concentration of lactose was carried out. Figure [3](#F3){ref-type=\"fig\"} shows the profiles of lactose consumption and 2-FL production in the batch fermentation of recombinant *E*. *coli* JM109(DE3) with 14.5\u2009g/l lactose. The cells consumed lactose slowly but produced 2-FL constantly for 96\u2009h. After 96\u2009h of fermentation, the 2-FL concentration did not increase any further and lactose consumption stopped. As a result, a maximum 2-FL concentration of 1.23\u2009g/l was obtained, which corresponded to a nine-fold (1.23\u2009g/l vs. 140\u2009mg/l) increase as compared with the previous fermentation with 2.6\u2009g/l lactose. 2- FL yield increased to 90\u2009mg 2-FL/g lactose when 14.5\u2009g/l of lactose was used. This improvement might be caused from increased lactose availability inside the cell. The results of the batch fermentations are summarized in Table [2](#T2){ref-type=\"table\"}.\n\n![**Profile of 2-FL production in the batch fermentation of recombinant*E. coli*JM109(DE3) strain harboring plasmids pmBCGW and pHfucT2.** After 3\u2009h of 0.1\u2009mM IPTG induction, 14.5\u2009g/l of lactose was added for 2-FL production. Symbols denote as follows; *grey circle*, dry cell mass; *green triangle*, 2-FL concentration; *blue square*, lactose concentration; *red square*, acetate concentration. Measurement of cell, lactose, acetate and 2-FL concentrations were done by three independent experiments. Symbols in the figure show the representative values of the batch fermentations.](1475-2859-11-48-3){#F3}\n\n###### \n\n**Summary of batch fermentations of*E*.*coli*strains producing 2-FL from lactose**\n\n  **Strains**     **Plasmids**       **Initial lactose concentration (g/l)**   **Maximum dry cell mass (g/l)**   **Maximum 2-FL concentration (g/l)**   **Yield (g 2-FL/g lactose)**\n  --------------- ------------------ ----------------------------------------- --------------------------------- -------------------------------------- ------------------------------\n  BL21star(DE3)   pmBCGW\u2009+\u2009pHfucT2   2.56\u2009\u00b1\u20090.04                               1.84\u2009\u00b1\u20090.05                       0.01\u2009\u00b1\u20090.001                           0.005\u2009\u00b1\u20090.001\n  JM109(DE3)      pmBCGW\u2009+\u2009pHfucT2   2.55\u2009\u00b1\u20090.02                               1.17\u2009\u00b1\u20090.05                       0.14\u2009\u00b1\u20090.015                           0.06\u2009\u00b1\u20090.005\n  \u00a0               \u00a0                  14.54\u2009\u00b1\u20090.67                              1.70\u2009\u00b1\u20090.28                       1.23\u2009\u00b1\u20090.011                           0.09\u2009\u00b1\u20090.004\n\nConfirmation of 2-FL biosynthesis by recombinant *E. coli* overexpressing GDP- [l]{.smallcaps}-fucose biosynthetic enzymes and FucT2\n------------------------------------------------------------------------------------------------------------------------------------\n\nLC/MS analysis was performed to confirm the biosynthesis of 2-FL in the recombinant *E*. *coli* JM109(DE3) strain overexpressing ManB, ManC, Gmd, WcaG and FucT2. HPLC data showed that a compound with the identical retention time to 2-FL was detected in the culture broth (Figure [4](#F4){ref-type=\"fig\"}B). MS scanning data (compound with RT\u2009=\u20096.6\u2009min) showed ion fragment of m/z 487.1, which is compatible with 2-FL (Figure [4](#F4){ref-type=\"fig\"}C).\n\n![**LC/MS analysis of 2-FL biosynthesis in the batch fermentation of the recombinant*E*.*coli*JM109(DE3) overexpressing ManB, ManC, Gmd, WcaG and FucT2.** At the end of batch fermentation, culture broth was collected for confirmation of extracellular 2-FL production. HPLC analysis of 100\u2009mg/l 2-FL standard solution ( **A**), HPLC analysis of culture broth of *E*. *coli* JM109(DE3) harboring pmBCGW\u2009+\u2009pHfucT2 ( **B**) and MS analysis of the compound with the retention time\u2009=\u20096.6\u2009min in the culture broth of *E*. *coli* JM109(DE3) harboring pmBCGW\u2009+\u2009pHfucT2 ( **C**).](1475-2859-11-48-4){#F4}\n\nEvaluation of 2-FL yield using EFM analysis for 2-FL producing *E. coli* from lactose\n-------------------------------------------------------------------------------------\n\nIn order to evaluate the efficiency of 2-FL production from lactose using the recombinant *E*. *coli* JM109(DE3) strain, elementary flux mode (EFM) analysis was employed to estimate a maximum theoretical yield of 2-FL from lactose. Figure [5](#F5){ref-type=\"fig\"} shows the prediction of theoretical 2-FL yield versus biomass yield for *E*. *coli* growing on lactose. Our experimental result from a batch fermentation of 14.5\u2009g/l of lactose resulted in a biomass yield of 0.1\u2009g biomass/g lactose. This suggests that 2-FL production from lactose by the engineered *E. coli* reached 20% of the maximum 2-FL production capacity.\n\n![**Calculation of the theoretical maximum yield of 2-FL from lactose.** Elementary flux mode (EFM) analysis was carried out for 2-FL producing *E*. *coli.*](1475-2859-11-48-5){#F5}\n\nDiscussion\n==========\n\nIn human milk, \u03b1-1,2-fucosylated oligosaccharides such as 2-FL, are known to have protective activity against pathogenic bacteria and their toxins \\[[@B3],[@B5],[@B9]\\]. Hence, the availability of large amounts of \u03b1-1,2-fucosylated oligosaccharides would be useful in ready-to-use materials, drugs for fundamental investigations or therapeutic purposes. As biosynthesis of GDP-[l]{.smallcaps}-fucose, a key compound for biosynthesis of \u03b1-1,2-fucosylated oligosaccharides, requires a number of enzymes and cofactors such as NADPH and GTP, a whole-cell conversion approach might be more realistic for industrial production than other chemical or enzymatic approaches \\[[@B25]\\].\n\nIn previous studies, improvements of GDP-[l]{.smallcaps}-fucose production in recombinant *E. coli* were attempted by manipulating the pathways enabling GDP- [l]{.smallcaps}-fucose biosynthesis from glucose \\[[@B17]-[@B19]\\]. This study was undertaken to upgrade the GDP-[l]{.smallcaps}-fucose production system for 2-FL production through additional overexpression of FucT2. To this end, the *fucT2* gene from *H*. *pylori* was cloned and overexpressed in the *E*. *coli* BL21star(DE3) strain. While an insoluble form of FucT2 was a major expression form, an active FucT2 was also observed in the recombinant *E*. *coli* as shown in Figure [2](#F2){ref-type=\"fig\"}. With the aid of FucT2 overexpression, recombinant *E*. *coli* BL21star(DE3) could produce 2-FL in the batch fermentation with 2.6\u2009g/l lactose. However, 2-FL yield was fairly low (5\u2009mg 2-FL/g lactose). *E*. *coli* BL21star(DE3) seemed to assimilate lactose instead of converting to 2-FL. Most of the initially added lactose was consumed within 12\u2009h of fermentation with a marginal growth during the fermentation (data not shown). These results suggested that *E*. *coli* BL21star(DE3) is not appropriate for 2-FL production.\n\nPreviously, several attempts for production of fucosylated (or sialylated) oligosaccharides from lactose have been made using the derivative of *E*. *coli* JM107 and JM109 since these strains are unable to produce an active \u03b2-galactosidase due to the insertion of the M15 single strand DNA into the *lacZ* gene \\[[@B12],[@B13],[@B26],[@B27]\\]. In these cases, glucose (or glycerol) was used as another carbon source for GDP-fucose production (or CMP-*N*-acetylneuraminic acid). These nucleotide sugars are subsequently used for fucosylation (or sialylation). According to the previous reports, *E*. *coli* JM109(DE3) was chosen as an alternative host strain for 2-FL production. As expected, the use of *E*. *coli* JM109(DE3) allowed production of a considerable amount of 2-FL in the batch fermentation, corresponding to a 14-fold increase in 2FL concentration compared with the value obtained in BL21star(DE3) strain. However, a theoretical yield of 2-FL from lactose was predicted to be 1.4\u2009g 2-FL/g lactose when cells cannot utilize lactose for growth. Meanwhile, our experimental yield of 2-FL was only about 0.1\u2009g 2-FL/g lactose. This result indicates that more than 90% of lactose consumed was used for other purposes such as biomass production and endogeneous metabolism. Slow consumption of lactose was also observed in the batch fermentation with mixed sugars (2\u2009g/l of lactose and 5\u2009g/l of mannose), where we expected that lactose could be mainly used for 2-FL production as mannose might be used for cell growth. Although an enhancement of 2-FL yield (0.13\u2009g 2-FL/g lactose) was obtained, most of the consumed lactose was not used for 2-FL production (data not shown). It is probable that incompletely inactivated \u03b2-galactosidase or cryptic \u03b2-galactosidase inside the cell might cause the reduction of 2-FL yield from lactose. Insertion of \u03bb(DE3) into the genome might cause the slow assimilation of lactose, which was supported by a previous report that \u03b1-complementation of LacZ was observed with the insertion of the DE3 cassette into the genomic DNA of *E*. *coli* JM101 \\[[@B26]\\].\n\nSince slow consumption of lactose and production of biomass were observed in the batch fermentation with 2.6\u2009g/l lactose, another batch fermentation with a high concentration of lactose (14.5\u2009g/l) was conducted in order to obtain a higher amount of 2-FL. Generally, *E*. *coli* is known to produce acetate when growing on an excessive sugar even under aerobic conditions. However, JM109(DE3) strain showed no acetate production even when excessive amounts (14.5\u2009g/l) of lactose were added as displayed in Figure [3](#F3){ref-type=\"fig\"}. It is generally known that acetate formation is accelerated when the metabolic fluxes to pyruvate exceed the capacity of the respiratory metabolism \\[[@B28],[@B29]\\]. Slow consumption of lactose might not lead to acetate formation, suggesting that the lactose utilization rate by *E. coli* JM109(DE3) is not fast enough to cause acetate formation. This result could be beneficial for designing a 2-FL production process such as a fed-batch type operation since acetate formation is known to be one of the main problems occurring in fed-batch type operations of *E*. *coli*\\[[@B17]\\]. However, low yield and productivity of 2-FL from lactose will need to be improved for industrial applications. The 2-FL production system in this study was verified by comparing the experimental 2-FL yield with the theoretical maximum yield estimated by the EFM analysis. 2-FL yield corresponded to about 20% of the theoretical maximum yield, which suggests further modifications via metabolic engineering of a host strain or optimization of fermentation processes should be carried out for improvement of 2-FL yield. Increased FucT2 solubility and intracellular lactose availability may be considered as primary approaches for improvement of 2-FL yield.\n\nConclusions\n===========\n\nIn this study, construction of efficient 2-FL production system was attempted. The *fucT2* gene from *H. pylori* was introduced into the recombinant *E. coli* able to overproduce GDP- [l]{.smallcaps}-fucose and biosynthesis of 2-FL was observed in the batch fermentation with lactose. The 2-FL production system using the *E*. *coli* JM109(DE3) strain showed a low rate of lactose assimilation and produced a considerable amount of 2-FL in the simple batch fermentation without acetate formation. The experimental 2-FL yield corresponded to 20% of the theoretical maximum yield, which indicates that more research should be conducted. Efficient microbial 2-FL production may be useful for utilizing 2-FL as a nutraceutical compound for various applications.\n\n**Competing interests**\n=======================\n\nThe authors declare that they have no competing interests.\n\n**Authors' contributions**\n==========================\n\nWHL, YSJ and JHS designed research. WHL, PP, JQ and JHJ performed the experiments. NSH and MJM contributed general advice and materials. WHL, PP, JQ, YSJ and JHS analyzed data and wrote the manuscript. YSJ and JHS supervised all works. All authors read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\nMETATOOL input file.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis research was supported by World Class University (WCU) program (R32-2008-000-10183-0) and the Advanced Biomass R&D Center (ABC) (2011--0031359) both funded by the Korean Ministry of Education, Science and Technology. Won-Heong Lee was supported by National Research Foundation of Korea Grant funded by the Korean Government (Ministry of Education, Science and Technology) (NRF-2011-357-F00041). Panchalee Pathanibul was funded by the Royal Thai government scholarship. LC/MS analysis was assisted by Metabolomics Center, Roy J. Carver Biotechnology Center of University of Illinois at Urbana-Champaign.\n"}
+{"text": "1. Introduction\n===============\n\nSystemic lupus erythematosus (SLE) is an autoimmune disease where the dysregulated T-cells activation, over-production of auto-antibodies and accumulation of immune complexes contribute to its phenotypes (e.g., photosensitivity, skin rashes, prolonged fatigue) and complications (e.g., lupus nephritis, coronary artery disease and osteoporosis) \\[[@B1-ijms-16-09794]\\]. Worldwide, the SLE incidence is more prevalent among the Asians, Afro-Americans, Afro-Caribbeans and Hispanic Americans, than the Caucasians \\[[@B2-ijms-16-09794]\\]. There also exists a sex difference in the prevalence of SLE where a ratio of seven women to men was reported \\[[@B3-ijms-16-09794]\\]. In Malaysia, the prevalence of SLE varies among the three main ethnic groups, *i.e.*, 57/100,000 for Chinese, 33/100,000 for Malays and 14/100,000 for Indians \\[[@B4-ijms-16-09794]\\]. Over the past decades, the underlying genetic basis in the development of SLE was unraveled by numerous genetic linkage analyses and association studies \\[[@B5-ijms-16-09794]\\].\n\nIn Malaysia, several potential SLE-predisposing candidate genes involved in immune functions, *i.e.*, genes encode for chemokines, interleukins and their receptors, tumour necrosis factors, T-cell surface receptors, components of the classical complement cascade, as well as major histocompatibility complex have been previously studied \\[[@B6-ijms-16-09794],[@B7-ijms-16-09794],[@B8-ijms-16-09794],[@B9-ijms-16-09794],[@B10-ijms-16-09794],[@B11-ijms-16-09794],[@B12-ijms-16-09794],[@B13-ijms-16-09794],[@B14-ijms-16-09794],[@B15-ijms-16-09794],[@B16-ijms-16-09794],[@B17-ijms-16-09794]\\]. Adding to this expanding list, our present study investigated the relationship between the programmed cell death 1 (*PDCD1*) gene and the development of SLE, which has not yet been reported in our Malaysian population. *PDCD1* is mapped to chromosome 2q37.3 and encodes for PD-1 molecule, which is an immunoinhibitory receptor of the CD28/B7 family. This PD-1 receptor is inducibly expressed on the activated T- and B cells, as well as myeloid cells. It plays a crucial role in the regulation of peripheral tolerance which involves the inactivation or suppression of self-reactive T- and B-cells through the activation-induced cell death or allergy in preventing autoimmunity \\[[@B18-ijms-16-09794],[@B19-ijms-16-09794],[@B20-ijms-16-09794]\\]. This was proven when the C57BL/6-*PDCD1*^\u2212/\u2212^ mice presented with autoimmune features resembling to SLE, e.g., lupus-like glomerulonephritis with the deposition of IgG3 and C3, arthritis and splenomegaly \\[[@B21-ijms-16-09794],[@B22-ijms-16-09794]\\].\n\nGenetic association studies were performed in different populations, *i.e.*, European, Caucasian, Mexican, European American, African American and Han Chinese, to examine the correlation between *PDCD1* gene polymorphisms and SLE susceptibility \\[[@B23-ijms-16-09794],[@B24-ijms-16-09794],[@B25-ijms-16-09794],[@B26-ijms-16-09794],[@B27-ijms-16-09794]\\]. Nevertheless, the causative role of *PDCD1* gene remains inconclusive owing to inter-study heterogeneity in the SNP studied or allele involved \\[[@B28-ijms-16-09794]\\]. We therefore, aimed to study four identified *PDCD1* SNPs, *i.e.*, PD1.1 (G\\>A; promoter region), PD1.3 (G\\>A; intron 4), PD1.5 (C\\>T; exon 5) and PD1.6 (G\\>A; 3\\'-UTR), in our Malaysian SLE cohort.\n\n2. Results\n==========\n\nAll the genotypes of these biallelic polymorphisms were reported in our population, except PD1.3 which AA genotype was absent ([Table 1](#ijms-16-09794-t001){ref-type=\"table\"}). The observed genotype frequencies among the unaffected controls were conformed to the Hardy-Weinberg Equilibrium (*p* \\> 0.05). In this study, the allele and genotype frequencies of all *PDCD1* SNPs were not significantly different between the SLE patient and control groups (*p* \\> 0.05) ([Table 1](#ijms-16-09794-t001){ref-type=\"table\"}). We also observed the absence of significant association between SLE and PD1.1, PD1.3, PD1.5 and PD1.6 SNPs, irrespective of the model of inheritance used.\n\nijms-16-09794-t001_Table 1\n\n###### \n\nAllele and genotype frequencies of programmed cell death 1 (*PDCD1*) single nucleotide polymorphisms (SNPs) among systemic lupus erythematosus (SLE) patients and healthy controls.\n\n  Allele/Genotype   Frequency, *n* (%)   Fisher's (*p* Value)   OR (95% CI)   \n  ----------------- -------------------- ---------------------- ------------- -------------------------\n  **PD1.1**                                                                   \n  A                 156 (39.0)           157 (39.2)             1.0000        0.9896 (0.7449--1.3145)\n  G                 244 (61.0)           243 (60.8)                           \n  AA                37 (18.5)            36 (18.0)              1.0000        1.0341 (0.623--1.718)\n  AG                82 (41.0)            85 (42.5)              0.8393        0.9402 (0.632--1.399)\n  GG                81 (40.5)            79 (39.5)              0.9187        1.0425 (0.699--1.555)\n  **PD1.3**                                                                   \n  A                 2 (0.5)              2 (0.5)                1.0000        1.0000 (0.1402--7.1343)\n  G                 398 (99.5)           398 (99.5)                           \n  AA                0 (0)                0 (0)                  1.0000        \\-\n  AG                2 (1.0)              2 (1.0)                1.0000        1.0000 (0.140--7.170)\n  GG                198 (99.0)           198 (99.0)             1.0000        1.0000 (0.140--7.170)\n  **PD1.5**                                                                   \n  C                 273 (68.2)           276 (69.0)             0.8789        0.9658 (0.7164--1.3019)\n  T                 127 (31.8)           124 (31.0)                           \n  CC                88 (44.0)            90 (45.0)              0.9199        0.9603 (0.647--1.425)\n  CT                97 (48.5)            96 (48.0)              1.0000        1.0202 (0.689--1.510)\n  TT                15 (7.5)             14 (7.0)               1.0000        1.0772 (0.506--2.295)\n  **PD1.6**                                                                   \n  A                 228 (57.0)           216 (54.0)             0.4339        1.1292 (0.8543--1.4926)\n  G                 172 (43.0)           184 (46.0)                           \n  AA                72 (36.0)            61 (30.5)              0.2885        1.2818 (0.845--1.945)\n  AG                84 (42.0)            94 (47.0)              0.3652        0.8166 (0.550--1.212)\n  GG                44 (22.0)            45 (22.5)              1.0000        0.9715 (0.606--1.556)\n\nFurther analysis was performed by stratifying our study cohort into three ethnic groups, *i.e.*, Chinese, Malays and Indians. No relationship was observed between PD1.1, PD1.3 and PD1.6 polymorphisms, and SLE susceptibility in either of the ethnic groups (*p* \\> 0.05) ([Table 2](#ijms-16-09794-t002){ref-type=\"table\"}). However, statistically significant difference in genotype frequency of PD1.5 between SLE patients and healthy controls was observed in both Malay and Indian groups (*p* \\< 0.01) ([Table 2](#ijms-16-09794-t002){ref-type=\"table\"}). [Table 3](#ijms-16-09794-t003){ref-type=\"table\"} shows the association analysis of different models of inheritance for PD1.5 and SLE among the Malays and Indians. The association between PD1.5 and SLE was statistically significant with a *p* value of 0.0003 and OR of 3.50 (95% CI = 1.74--7.05) under the best-fitting over-dominant model among the Malays ([Table 3](#ijms-16-09794-t003){ref-type=\"table\"}). On the other hand, the Indian group showed a significant inverse association between PD1.5 and SLE under the best-fitting over-dominant model (*p* \\< 0.0001) ([Table 3](#ijms-16-09794-t003){ref-type=\"table\"}).\n\nijms-16-09794-t002_Table 2\n\n###### \n\nGenotype frequencies of *PDCD1* SNPs among SLE patients and healthy controls of different ethnics.\n\n  SNP         Genotype Frequency, *n*                                                               \n  ----------- ------------------------- ----- ---- ---- ------------ ------------ ---- ------------ ------------\n  **PD1.1**                                                                                         \n  **AA**      28                        28    ns   8    7            ns           1    0            ns\n  **AG**      53                        53    27   33   2            0                              \n  **GG**      34                        34    35   30   12           15                             \n  **PD1.3**                                                                                         \n  **AA**      0                         0     ns   0    0            ns           0    0            ns\n  **AG**      0                         0     1    1    1            1                              \n  **GG**      115                       115   69   69   14           14                             \n  **PD1.5**                                                                                         \n  **CC**      56                        46    ns   25   44           **0.0022**   7    0            **0.0063**\n  **CT**      49                        60    42   21   **0.0006**   6            15   **0.0007**   \n  **TT**      10                        9     3    5    ns           2            0    ns           \n  **PD1.6**                                                                                         \n  **AA**      52                        47    ns   18   14           ns           2    0            ns\n  **AG**      49                        55    33   35   2            4                              \n  **GG**      14                        13    19   21   11           11                             \n\nns: Non-significant; *p* value in bold: significant.\n\nijms-16-09794-t003_Table 3\n\n###### \n\nInheritance model analysis of PD1.5 and SLE in Malay and Indian groups.\n\n  Inheritance Model   Genotype            Malay               Indian                                                  \n  ------------------- ------------------- ------------------- ------------ ------- ------------------- -------------- ------\n  Dominant            C/C                 1.00                **0.0012**   187.6   1.00                **0.0006**     33.7\n  C/T-T/T             3.05 (1.53--6.06)   0.00 (0.00--NA)                                                             \n  Recessive           C/C-C/T             1.00                0.46         197.5   1.00                0.088          42.7\n  T/T                 0.58 (0.13--2.54)   NA (0.00--NA)                                                               \n  Co-dominant         C/C                 1.00                **0.0015**   187.1   1.00                **0.0003**     31.1\n  C/T                 3.52 (1.72--7.22)   0.00 (0.00--NA)                                                             \n  T/T                 1.06 (0.23--4.80)   1.00 (0.00--NA)                                                             \n  Over-dominant       C/C-T/T             1.00                **0.0003**   185.1   1.00                **\\<0.0001**   29.1\n  C/T                 3.50 (1.74--7.05)   0.00 (0.00--NA)                                                             \n  Log-additive        \\-                  2.01 (1.12--3.58)   **0.016**    192.3   0.25 (0.05--1.31)   0.071          42.3\n\n*p* Value in bold: significant; AIC: Akaike Information Criterion; NA: not applicable.\n\n3. Discussion\n=============\n\nGenome wide linkage scanning for SLE susceptibility loci in Icelandic and Swedish families had revealed a favourable linkage to the long arm of chromosome 2 at 2q37 \\[[@B29-ijms-16-09794]\\]. The new susceptibility locus for SLE, *SLEB2*, was then identified via further mapping in a combined set of Nordic families (Icelandic, Swedish and Norwegian) \\[[@B30-ijms-16-09794]\\]. In view of the biological role of PD-1 in the regulation of peripheral tolerance, the *PDCD1* gene that is located within this locus might serve as a potential predisposing gene to SLE. The interaction between PD-1 and its ligands, PD-L, was found to negatively co-stimulate and attenuate the auto-reactive T- and B-cells through the inhibition of cytokine production and cell cycle arrest in G~0~/G~1~ phase \\[[@B31-ijms-16-09794]\\]. The deficiency of *PDCD1* gene expression had been proven *in vivo* to result in inadequate auto-reactive lymphocytes removal and auto-antibodies production \\[[@B22-ijms-16-09794],[@B31-ijms-16-09794],[@B32-ijms-16-09794]\\]. It was thus evidenced that the PD-1-PD-L pathway is pivotal in maintaining peripheral self-tolerance and preventing autoimmune diseases, e.g., SLE \\[[@B20-ijms-16-09794]\\].\n\nIn our multiethnic sample cohort, the similar distribution of allele and genotype frequencies for *PDCD1* polymorphisms were observed in both SLE and control groups. Our findings corresponded well with three earlier reports on Taiwanese and Caucasian SLE patients, where the efforts to demonstrate the association between *PDCD1* gene and SLE susceptibility were failed \\[[@B33-ijms-16-09794],[@B34-ijms-16-09794],[@B35-ijms-16-09794]\\]. However, statistically significant association was reported for PD1.5 following further sample stratification of our Malaysian cohort into different ethnic groups. Our association analysis revealed that the PD1.5C/T genotype was significantly associated to SLE susceptibility in Malays with a 3.5-fold increased risk compared to the homozygotes when the over-dominant model was assumed. On the other hand, the heterozygous genotype of this exonic polymorphism was significantly correlated to the Indian healthy controls under the same inheritance model. These differences in SNP association to disease in different ethnics revealed a possible interaction with other genetic or environmental factor in determining the predisposition to SLE \\[[@B36-ijms-16-09794]\\]. In a study by Lee *et al.* (2009) \\[[@B28-ijms-16-09794]\\], the relationship between PD1.5C allele and SLE susceptibility was also established among the Europeans.\n\nOf all, the PD1.3 is the most extensively investigated. The PD1.3 G to A substitution was demonstrated to abolish the binding of the runt-related transcription factor 1 (RUNX-1) by altering its binding site. This disruption in *PDCD1* transcription will lead to the hyperactivity of self-reactive lymphocytes and subsequent breakdown of peripheral tolerance. Further evidences had shown the correlation between PD1.3A allele and a decreased expression of PD-1 receptor in SLE patients \\[[@B23-ijms-16-09794],[@B37-ijms-16-09794]\\]. In addition, the relationship of PD1.3 variant and the surface level of PD-1 receptor on activated CD4+CD25+ T cells also grounded its functional involvement in SLE. These regulatory T cells are crucial in maintaining our body's self-tolerance by curbing the hazardous activities of self-reactive T cells \\[[@B37-ijms-16-09794],[@B38-ijms-16-09794]\\].\n\nDespite its strong functional grounds, the association between this intronic polymorphism and SLE was failed to address with sufficient consistency. Prokunina *et al.* and Sanchez *et al.* had demonstrated the association between PD1.3A allele and SLE susceptibility in the European, Mexican, Swedish and Argentine cohorts \\[[@B23-ijms-16-09794],[@B26-ijms-16-09794],[@B39-ijms-16-09794]\\]. On the contrary, a protective effect of PD1.3A allele on SLE was reported among the Spanish cohort \\[[@B19-ijms-16-09794]\\]. This variant however, was not associated to SLE in our Malaysian population, as well as previously reported northern Sweden, U.S. and Polish cohorts \\[[@B40-ijms-16-09794],[@B41-ijms-16-09794],[@B42-ijms-16-09794]\\]. Our failure in replicating this relationship might be due to the relatively low minor allele frequency of PD1.3 variant in our population (0.5%) compared to the Europeans (13%) \\[[@B43-ijms-16-09794]\\]. Hence, the possibility for the minor allele of PD1.3 to be an important predisposing variant in SLE was excluded owing to its extreme rarity in populations of non-European descent, e.g., 0.008% in Asians \\[[@B24-ijms-16-09794]\\].\n\n4. Experimental Section\n=======================\n\n4.1. Study Subjects\n-------------------\n\nOur study cohort was comprised of three main ethnics of the Malaysian population, *i.e.*, Malays, Chinese and Indians. A total of 200 SLE patients admitted to the University Malaya Medical Centre (UMMC), Kuala Lumpur, Malaysia were volunteered for this case-control association study. Two hundred age- and race-matched healthy individuals were also recruited in the control group. The study protocol was approved by the Medical Ethics Committee (Ref. No. 380.1), and written informed consent was also obtained from the study participants.\n\n4.2. Genotyping of PDCD1 Polymorphisms\n--------------------------------------\n\nVenous blood samples were collected and genomic DNA was extracted with a similar DNA extraction approach as reported previously \\[[@B44-ijms-16-09794],[@B45-ijms-16-09794],[@B46-ijms-16-09794]\\]. Next, the genotyping of *PDCD1* variants was performed via real-time polymerase chain reaction (PCR) using pre-designed TaqMan SNP Genotyping Assays on an Applied Biosystems (Foster City, CA, USA) 7500 Fast Real-Time PCR System ([Table 4](#ijms-16-09794-t004){ref-type=\"table\"}). The real-time PCR was conducted under universal thermal cycling conditions: denaturation step at 95 \u00b0C for 20 s, followed by 40 cycles of denaturation step at 95 \u00b0C for 3 s and annealing/extension step at 60 \u00b0C for 30 s. The singleplex PCR reaction mixture consists of 5 \u03bcL of 2\u00d7 TaqMan GTXpress master mix, 0.5 \u03bcL of 20\u00d7 pre-designed TaqMan SNP Genotyping Assay, 1 \u03bcL of 20 ng/\u03bcL DNA template, and deionized water in a total reaction volume of 10 \u03bcL.\n\nijms-16-09794-t004_Table 4\n\n###### \n\nPre-designed TaqMan SNP Genotyping Assays for genotyping of *PDCD1* genetic variants.\n\n  SNP     Assay ID for Pre-Designed TaqMan SNP Genotyping Assay\n  ------- -------------------------------------------------------\n  PD1.1   C_57931321_10 (\\[V\\]: Allele G; \\[F\\]: Allele A)\n  PD1.3   C_57931290_10 (\\[V\\]: Allele G; \\[F\\]: Allele A)\n  PD1.5   C_57931286_20 (\\[V\\]: Allele T; \\[F\\]: Allele C)\n  PD1.6   C_172862_10 (\\[V\\]: Allele G; \\[F\\]: Allele A)\n\n\\[V\\]: VIC-Probe; \\[F\\]: 6-FAM-Probe.\n\n4.3. Statistical Analysis\n-------------------------\n\nAllele and genotype frequencies of the studied *PDCD1* SNPs were determined in both SLE and healthy control groups, and subgroup analyses were performed following ethnic stratification. The observed genotype frequencies were tested for concordance to Hardy-Weinberg Equilibrium (HWE) by using a Chi-squared (\u03c7^2^) test. The Fisher's Exact Test was used to assess the significance of the differences in observed allele and genotype frequencies between the diseased and control group. Odds ratio (OR) with 95% confidence interval (CI) was also determined. The SNP association to SLE was measured on (i) dominant model; (ii) recessive model; (iii) co-dominant model; (iv) over-dominant model; and (v) log-additive model. The Akaike Information Criterion (AIC) was calculated to determine the best-fit model of inheritance.\n\n5. Conclusions\n==============\n\nOf all the investigated *PDCD1* variants, only PD1.5 was significantly associated to the genetic predisposition to SLE in our Malaysian population. The remarkable ethnic and geographic variations in the distribution of *PDCD1* variants might explain the discordant findings related to its contribution in SLE susceptibility. On the other hand, the lack of consistent association between *PDCD1* gene and SLE in different population background might rule out its direct involvement in disease predisposition, implying instead a possible gene-environment interaction.\n\nThis research was supported by Fundamental Research Grant Scheme, FP073/2010B.\n\nKek Heng Chua, Lay Hoong Lian and Tien Eang Cheah conceived and designed the experiments; Xiu Jia Sim and Tze Pheng Lau performed the experiments and analysed the data; All authors were involved in drafting and approving the final manuscript.\n\nThe authors declare no conflict of interest.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction {#s1}\n============\n\nAmphetamine (AMPH) is both addictive, with several notable episodes of widespread abuse worldwide, and therapeutic, for treating narcolepsy, attention deficit hyperactivity disorder, obesity, and traumatic brain injury [@pone.0060763-Bales1], [@pone.0060763-Howell1]. While there is little debate that behavioral effects of this important psychostimulant are associated with a hyperdopamine state [@pone.0060763-Carboni1]--[@pone.0060763-Kuczenski2], the underlying mechanisms by which this condition manifests have been the subject of intense study. Two, what ostensibly appear to be mutually exclusive, views have emerged. On the one hand, AMPH enhances tonic dopamine signaling by reversing dopamine transporter (DAT) direction, leading to a non-exocytotic, action potential-independent type of release or \"efflux\" that is driven by vesicular depletion and the redistribution of dopamine to the cytosol [@pone.0060763-Fleckenstein1], [@pone.0060763-Sulzer1]. On the other hand, AMPH enhances phasic dopamine signaling by promoting burst firing of dopamine neurons [@pone.0060763-Paladini1], [@pone.0060763-Shi1], inhibiting dopamine uptake [@pone.0060763-May1], [@pone.0060763-Ramsson1], and up-regulating vesicular dopamine release [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2]. How AMPH concurrently activates tonic and phasic dopamine signaling, the two fundamental modes of communication used by dopamine neurons [@pone.0060763-Schultz1], yet elicits opposing actions on vesicular dopamine stores is perplexing and unresolved.\n\nPresynaptic neurotransmitter vesicles are functionally and anatomically segregated into at least three distinct pools, readily releasable, recycling, and reserve, that are interrogated by electrical stimulation of short, intermediate, and long duration, respectively [@pone.0060763-Rizzoli1]. Distinct vesicular stores have also been proposed to contribute to exocytotic dopamine release in a stimulus-dependent manner [@pone.0060763-Ewing1]--[@pone.0060763-Yavich1]. At the cellular level, AMPH exerts differential actions on dopamine vesicle populations [@pone.0060763-Anderson1]--[@pone.0060763-Riddle1]. Moreover, although not systematically evaluated to assess distinct vesicular stores, AMPH effects on electrically evoked levels of extracellular dopamine in the striatum *in vivo* are stimulus-dependent, with increases revealed by short trains and decreases by long trains [@pone.0060763-Dugast1], [@pone.0060763-Stamford1]. It is thus interesting to speculate that AMPH depleting the reserve pool drives tonic dopamine signaling by providing a source of cytosolic dopamine for efflux, but enhancing the readily releasable pool drives phasic dopamine signaling by augmenting vesicular dopamine release.\n\nHere we use *in vivo* voltammetry and vary stimulus duration to test the novel hypothesis that AMPH elicits opposing actions on dopamine stores. In support of this hypothesis, we show in the dorsal striatum that AMPH increased exocytotic dopamine release evoked by a short train, which interrogates the readily releasable pool, but decreased release evoked by a long train, which interrogates the reserve pool. A concurrent augmentation of tonic and phasic dopamine signaling was also observed. Vesicular depletion and enhanced tonic signaling appear to be linked because these effects were specific to AMPH and not cocaine, and to the dorsal but not ventral striatum, whereas activation of vesicular release and phasic signaling generalized across psychostimulants and striatal sub-regions. Our results thus support a model of AMPH differentially targeting vesicular stores to reconcile its paradoxical effects on dopamine neurons and identify regionally distinct actions of this psychostimulant in the striatum that may relate to its addictive and therapeutic properties.\n\nMethods {#s2}\n=======\n\nExperimental Design {#s2a}\n-------------------\n\nThe experimental design is shown in [Figure 1](#pone-0060763-g001){ref-type=\"fig\"}. Three durations of stimulus trains, short (0.4 s), intermediate (2 s), and long (10 s), were applied to each animal and repeated after administration of the saline control or drug treatment. A frequency of 60 Hz was used for all stimulations. Stimulus current was \u00b1300 \u00b5A for long and intermediate trains, and \u00b1125 \u00b5A for the short train. The lower current intensity was selected for the short train to elicit evoked responses mirroring the amplitude and dynamics of naturally occurring phasic dopamine transients [@pone.0060763-Robinson1]. As such, we refer to these responses as \"phasic-like\". This short train is also reinforcing in the operant paradigm of intracranial self-stimulation [@pone.0060763-Cheer1]. Sufficient time was allowed between trains for evoked responses to recover (5 s per pulse; [@pone.0060763-Wightman1]). Extracellular dopamine was measured in urethane-anesthetized rats by fast-scan cyclic voltammetry (FSCV) at a carbon fiber microelectrode (CFM) implanted in the dorsal and ventral striatum, as described previously [@pone.0060763-Ramsson1]. Vesicular dopamine release was resolved from dopamine uptake for all evoked responses [@pone.0060763-Wightman1], [@pone.0060763-Wu1]. A low (1 mg/kg, i.p.) and high (10 mg/kg, i.p.) dose of AMPH was evaluated to assess dose-dependent effects. A high dose of cocaine (40 mg/kg i.p.) was evaluated for comparison.\n\n![Experimental timeline.\\\nThree stimulation trains with different durations (0.4 s, 2 s, and 10 s), indicated by the horizontal line under each evoked response, were applied before and after psychostimulant administration at time 0 min. Note that evoked responses are on a second timescale, while the overall design is shown in minutes.](pone.0060763.g001){#pone-0060763-g001}\n\nAnimals {#s2b}\n-------\n\nAdult male Sprague-Dawley rats (\u223c350--400 g), purchased from Harlan (Indianapolis, IN, USA), were housed under standard conditions of lighting and temperature. Food and water were provided *ad libitum*. Protocols were approved by the Institutional Animal Care and Use Committee of Illinois State University. Care was in accordance with NIH guidelines (publication 86--23).\n\nSurgery {#s2c}\n-------\n\nRats were anesthetized with urethane (1.6 g/kg, i.p.) and immobilized in a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA). Deltaphase Isothermal Pads (Braintree Scientific, Braintree, MA, USA) maintained core temperature throughout surgery. Burr holes were drilled overlying targeted regions, dura was removed, and electrodes lowered along a vertical trajectory using stereotactic coordinates obtained from a brain atlas based on a flat-skull position [@pone.0060763-Paxinos1] and utilizing bregma and dura as reference points. All coordinates, anteroposterior (AP), mediolateral (ML) and dorsoventral (DV) are given in mm. The stimulating electrode targeted the medial forebrain bundle (AP: \u22124.6, ML: +1.4, DV: \u22127.0), and a CFM targeted the dorsal (AP: +1.2, ML: +3.0, DV: \u22124.5 to 5.0) and ventral (AP: +1.2, ML: +2.0, DV: \u22126.5 to 7.5) striatum. The reference electrode, a chloridized silver wire, was placed in the contralateral superficial cortex.\n\nElectrochemistry {#s2d}\n----------------\n\nFSCV was performed by a Universal Electrochemistry Instrument (UEI; Department of Chemistry Electronic Shop, University of North Carolina, Chapel Hill, NC, USA), which was computer controlled by commercially available software (ESA Bioscience, Chelmsford, MA, USA). The potential of the CFM was linearly scanned at 10 Hz from a resting value of \u22120.4 V to 1.3 V (versus the reference electrode) and back again at a rate of 400 V/s. The peak oxidation current for dopamine recorded during each scan was converted to a concentration based on post-calibration of the CFM using flow-injection analysis in a buffer consisting of 150 mM sodium chloride with 15 mM TRIS and adjusted to a pH of 7.4 [@pone.0060763-Wu2]. Dopamine was identified from the background subtracted voltammogram [@pone.0060763-Michael3].\n\nElectrical Stimulation {#s2e}\n----------------------\n\nElectrical stimulation was computer generated and consisted of biphasic pulses (2 ms each phase). Stimulus trains were applied to a twisted bipolar stimulating electrode (Plastics One, Roanoke, VA, USA) through a constant-current generator and optical isolator (NL 80, Neurolog, Medical Systems, Great Neck, NY, USA).\n\nData Analysis {#s2f}\n-------------\n\nDopamine responses electrically evoked by short and medium stimulations were analyzed for maximal concentration (\\[DA\\]~max~) and parameters described vesicular dopamine release and dopamine uptake according to [@pone.0060763-Wightman1]:where \\[DA\\]~p~ is the concentration of dopamine released per stimulus pulse, f is the frequency of stimulation, and k is the first-order term describing dopamine uptake. Data were best fit to [Equation 1](#pone.0060763.e001){ref-type=\"disp-formula\"} using non-linear regression with a simplex algorithm [@pone.0060763-Wu1]. First-order, as opposed to Michaelis-Menten, kinetics was selected to characterize dopamine uptake because of concern that AMPH alters both K~m~ and V~max~, which is difficult to resolve with *in vivo* voltammetry [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2]. However, similar AMPH-induced changes in \\[DA\\]~p~, the focus of the present study, have been reported using both kinetic models [@pone.0060763-Daberkow1]. Dopamine responses evoked by long trains were analyzed for vesicular dopamine release using single curve analysis [@pone.0060763-Wu1]. The reason is that [Equation 1](#pone.0060763.e001){ref-type=\"disp-formula\"} assumes that vesicular dopamine release is constant, and AMPH clearly caused time-dependent changes in recordings evoked by long trains as evident by the pronounced slowing of the upward slope during the train, especially in the dorsal striatum. In single curve analysis, which does not assume a kinetic mechanism for dopamine uptake, the slope of the downward portion of the evoked signal (i.e., uptake) is subtracted from the upward portion (i.e., release - uptake) to calculate vesicular dopamine release:The only assumption of single curve analysis regarding uptake is that rates governing up- and downward portions are identical at the same dopamine concentration, which is also the same assumption as in [Equation 1](#pone.0060763.e001){ref-type=\"disp-formula\"}. It should be emphasized that because of DAT reversal, uptake measured in the presence of AMPH more faithfully represents net dopamine clearance, i.e., the difference between extracellular removal by uptake and addition by efflux [@pone.0060763-Ramsson1], [@pone.0060763-John1]. Nevertheless, the combination of these effects is accounted for in the analysis, which permits a direct determination of vesicular dopamine release (i.e., \\[DA\\]~p~).\n\nNon-electrically evoked changes in extracellular dopamine representing tonic and phasic dopamine signaling were chemically resolved from the FSCV recordings with principal component regression (PCR) using dopamine, pH and background drift as analytes [@pone.0060763-Keithley1], [@pone.0060763-Hermans1]. For training sets, dopamine and pH changes were obtained from the electrically evoked responses, whereas background drift was obtained during baseline recording in the time between stimulations. PCR was performed sequentially on 5-min epochs. Spontaneously occurring dopamine transients were identified and characterized with peak-finding software (Mini-Analysis, Synaptosoft, Decatur, GA, USA).\n\nStatistical Analysis {#s2g}\n--------------------\n\nWhen appropriate, data are presented as the mean \u00b1 SEM. \\[DA\\]~max~ and \\[DA\\]~p~ were statistically analyzed using a two-way ANOVA with drug treatment and stimulus duration as independent variables, followed by sequential Bonferroni *post hoc* tests. Effects of drug treatment on k were analyzed using a one-way ANOVA with a Tukey\\'s *post hoc* test. Tonic dopamine levels were statistically analyzed using a one-way ANOVA with repeated measures. Statistical analysis was performed using SPSS Version 18 for Windows (SPSS). Significance was set at *p*\\<0.05.\n\nDrugs {#s2h}\n-----\n\nUrethane, cocaine hydrochloride, and d-amphetamine sulfate were purchased from Sigma (St. Louis, MO, USA). All drugs were dissolved in 150 mM NaCl prior to injection. *d*-amphetamine and cocaine doses were determined by base weight.\n\nResults {#s3}\n=======\n\nPsychostimulant effects on evoked dopamine levels {#s3a}\n-------------------------------------------------\n\nIndividual recordings of electrically evoked dopamine levels collected during the four treatments are shown in [Figure 2](#pone-0060763-g002){ref-type=\"fig\"} for the dorsal striatum and [Figure 3](#pone-0060763-g003){ref-type=\"fig\"} for the ventral striatum. Average results for \\[DA\\]~max~, the maximal concentration of the evoked signal, and obtained from these recordings are shown in [Figure 4A](#pone-0060763-g004){ref-type=\"fig\"} (left, dorsal striatum; right, ventral striatum). Both individual responses and averaged results demonstrate drug-, dose-, stimulus-, and region-dependent effects, and four general observations can be made. First, psychostimulant effects were inversely related to stimulus duration in both striatal regions. Second, AMPH but not cocaine decreased \\[DA\\]~max~ evoked by the long train, and this only occurred in the dorsal striatum. Third, AMPH was more proficient in increasing \\[DA\\]~max~ evoked by the short train in the ventral striatum, whereas cocaine elicited greater effects in the dorsal striatum. And fourth, the high dose of AMPH was more proficient at increasing \\[DA\\]~max~ during short trains in both striatal regions compared to the low dose. Statistical analysis of \\[DA\\]~max~ revealed a significant effect of drug treatment in the dorsal (F~3,75~\u200a=\u200a13.45, *p*\u200a=\u200a\\<0.001) and ventral (F~3,74~\u200a=\u200a8.81, *p*\\<0.001) striatum, a significant effect of stimulus duration in the dorsal (F~2,75~\u200a=\u200a47.94, *p*\\<0.001) and ventral (F~2,74~\u200a=\u200a13.96, *p*\\<0.001) striatum, and a significant interaction in the dorsal (F~6,75~\u200a=\u200a8.45, *p*\\<0.001) and ventral (F~6,74~\u200a=\u200a3.08, *p*\\<0.01) striatum. In the dorsal striatum, 10 mg/kg AMPH and 40 mg/kg cocaine significantly (*p*\\<0.002) increased \\[DA\\]~max~ evoked by the short train, but only cocaine was effective at the intermediate train (*p*\\<0.001). Both doses of AMPH (1 and 10 mg/kg) significantly (*p*\\<0.001) decreased \\[DA\\]~max~ evoked by the long train, whereas cocaine was without effect. In the ventral striatum, both doses of AMPH and cocaine significantly (*p*\\<0.01) increased \\[DA\\]~max~ evoked by short and intermediate trains, but were without effect with the long train.\n\n![Representative psychostimulant- and stimulation-dependent effects on evoked dopamine dynamics in the dorsal striatum.\\\n**A.** Saline. **B.** 1 mg/kg AMPH. **C.** 10 mg/kg AMPH. **D.** 40 mg/kg cocaine (COC). AMPH and cocaine altered the amplitude of evoked dopamine signals in the dorsal striatum, while saline had no effect. In contrast to cocaine, there was an inverse relationship between stimulus duration and evoked dopamine amplitude following AMPH. Application of the stimulus train is indicated by the solid line underneath each representative response for short (left), intermediate (middle) and long (right) durations.](pone.0060763.g002){#pone-0060763-g002}\n\n![Representative psychostimulant- and stimulation-dependent effects on evoked dopamine dynamics in the ventral striatum.\\\n**A.** Saline. **B.** 1 mg/kg AMPH. **C.** 10 mg/kg AMPH. **D.** 40 mg/kg cocaine (COC). AMPH and cocaine increased evoked dopamine amplitude for at stimulus durations in the ventral striatum, while saline had no effect. Application of the stimulus train is indicated by the solid line underneath each representative response for short (left), intermediate (middle) and long (right) durations.](pone.0060763.g003){#pone-0060763-g003}\n\n![Averaged psychostimulant- and stimulation-dependent effects.\\\n**A.** The maximal concentration of electrically evoked dopamine (\\[DA\\]~max~). **B.** Vesicular release (\\[DA\\]~p~). Stimulus duration is shown along the x axis. Psychostimulants differentially elicited stimulus-dependent effects on \\[DA\\]~max~ and \\[DA\\]~p~. Data are the ratio of post-drug over pre-drug response (Post/Pre) for the dorsal (left) and ventral (right) striatum and are expressed as mean \u00b1 SEM. \\*, significantly different from saline (*p*\\<0.05).](pone.0060763.g004){#pone-0060763-g004}\n\nPsychostimulant effects on vesicular dopamine release and dopamine uptake {#s3b}\n-------------------------------------------------------------------------\n\nObserved psychostimulant-induced changes in \\[DA\\]~max~ could arise from altered vesicular dopamine release and/or dopamine uptake, because both mechanisms regulating extracellular dopamine in the striatum operate concurrently during the stimulus train [@pone.0060763-Wightman1]. Evoked responses were therefore analyzed to determine the respective contributions of these presynaptic mechanisms to \\[DA\\]~max~. [Figure 4B](#pone-0060763-g004){ref-type=\"fig\"} shows vesicular dopamine release (\\[DA\\]~p~). Overall, \\[DA\\]~p~ and \\[DA\\]~max~ ([Fig. 4A](#pone-0060763-g004){ref-type=\"fig\"}) tracked each other well. Statistical analysis of \\[DA\\]~p~ revealed a significant effect of drug treatment in the dorsal (F~3,73~\u200a=\u200a8.36, *p*\\<0.001) and ventral (F~3,72~\u200a=\u200a6.79, *p*\\<0.001) striatum, a significant effect of train duration in the dorsal (F~2,73~\u200a=\u200a30.45, *p*\\<0.001) and ventral (F~2,72~\u200a=\u200a19.53, *p*\\<0.001) striatum, and a significant interaction in the dorsal (F~6,73~\u200a=\u200a6.33, *p*\\<0.001) and ventral (F~6,72~\u200a=\u200a4.26, *p*\\<0.001) striatum. In the dorsal striatum, 10 mg/kg AMPH and cocaine significantly increased \\[DA\\]~p~ for the short train (*p*\\<0.02). 1 mg/kg AMPH was without effect, and no treatment had significant effects for the intermediate train. Both doses of AMPH significantly decreased \\[DA\\]~p~ for the long train (*p*\\<0.01), while cocaine had no effect. All drug treatments significantly increased \\[DA\\]~p~ in the ventral striatum for both short and intermediate trains (*p*\\<0.03) but were without effect for the long stimulation.\n\nPsychostimulant effects on dopamine uptake are shown in [Table 1](#pone-0060763-t001){ref-type=\"table\"}. Low- and high-dose AMPH and cocaine robustly decreased dopamine uptake (k) to a similar degree in both striatal regions. Statistical analysis of k revealed a significant effect of drug treatment in both the dorsal (F~3,54~\u200a=\u200a10.53, *p*\\<0.001) and ventral (F~3,54~\u200a=\u200a15.80, *p*\\<0.001) striatum. Each drug treatment significantly decreased dopamine uptake compared to saline control in both striatal regions (*p*\\<0.01). AMPH- and cocaine-mediated uptake inhibition is consistent with our previous work using Michaelis-Menten kinetics [@pone.0060763-Ramsson1], [@pone.0060763-Wu1], [@pone.0060763-Wu2], and the degree of inhibition was similar to our previous work using first-order kinetics [@pone.0060763-Ramsson2], as is used here. This result, indicating no distinct effects of drug treatment or striatal region on dopamine uptake, and the excellent correspondence between \\[DA\\]~max~ and \\[DA\\]~p~ shown in [Figure 3](#pone-0060763-g003){ref-type=\"fig\"}, suggest that psychostimulant-induced changes in \\[DA\\]~max~ evoked by the trains used in this study are dominated by changes in vesicular dopamine release. The one overt exception is the intermediate train in the dorsal striatum, where cocaine increased \\[DA\\]~max~ without a corresponding change in \\[DA\\]~p~. In this case, reduced dopamine uptake dominates the increase in \\[DA\\]~max~. Overall, these results demonstrate that AMPH and cocaine increase vesicular dopamine release in both striatal regions with the short train but that AMPH decreases vesicular dopamine release in the dorsal striatum with the long train.\n\n10.1371/journal.pone.0060763.t001\n\n###### Psychostimulant effects on dopamine uptake.\n\n![](pone.0060763.t001){#pone-0060763-t001-1}\n\n                  Saline                    AMPH (1 mg/kg)                                AMPH (10 mg/kg)                               Cocaine (40 mg/kg)\n  ------------- ----------- ---------------------------------------------- ---------------------------------------------- ----------------------------------------------\n  **Dorsal**     0.96\u00b10.05   0.57\u00b10.04[\\*\\*](#nt102){ref-type=\"table-fn\"}   0.53\u00b10.08[\\*\\*](#nt102){ref-type=\"table-fn\"}   0.66\u00b10.06[\\*\\*](#nt102){ref-type=\"table-fn\"}\n  **Ventral**    0.95\u00b10.04   0.62\u00b10.05[\\*\\*](#nt102){ref-type=\"table-fn\"}   0.50\u00b10.08[\\*\\*](#nt102){ref-type=\"table-fn\"}   0.60\u00b10.05[\\*\\*](#nt102){ref-type=\"table-fn\"}\n\nData are the mean \u00b1 SEM.\n\n, significantly different from saline (*p*\\<0.01).\n\nPsychostimulant effects on tonic dopamine signaling {#s3c}\n---------------------------------------------------\n\n[Figure 5](#pone-0060763-g005){ref-type=\"fig\"} shows a representative background-subtracted FSCV recording (black) collected immediately surrounding the time of injecting high-dose AMPH. This non-electrically evoked trace, representing current measured at the peak oxidative potential for dopamine (i.e., along the horizontal white line of the pseudocolor plot below), gradually increases across the 5-min epoch. Individual voltammograms collected along the two vertical white lines of the pseudocolor plot (blue) are overlaid with a dopamine voltammogram collected during electrical stimulation (black) earlier in this recording (data not shown). While there is evidence for dopamine in the individual voltammograms and in the sequential voltammograms displayed in the pseudocolor plot for this non-electrically evoked trace, other analytes obscure its selective measurement with FSCV alone. However, PCR (red) resolves the dopamine component of this FSCV recording, demonstrating an activation of tonic dopamine signaling by AMPH.\n\n![Representative effects of AMPH on tonic dopamine levels in the dorsal striatum.\\\nThe black line (left y axis) in the top panel shows background-subtracted current, and pseudocolor plot underneath displays all background-subtracted cyclic voltammograms immediately following administration of the high dose (10 mg/kg) of AMPH. Current, which was measured at the peak oxidative potential for dopamine (horizontal white line on the pseudocolor plot), was converted to dopamine concentration (red line, right y axis) using PCR. **INSET.** Background-subtracted cyclic voltammograms taken at 150 s and 250 s (blue arrows, blue line) and from the post-drug electrically evoked (60 Hz, 0.4 s) dopamine signal (black line).](pone.0060763.g005){#pone-0060763-g005}\n\n[Figure 6](#pone-0060763-g006){ref-type=\"fig\"} shows the average effects of the four treatments on tonic dopamine signaling as determined by PCR analysis for the first 10 min of the FSCV recording after drug injection, which is just prior to the first stimulation of the post-drug period (see [Fig. 1](#pone-0060763-g001){ref-type=\"fig\"}). This initial recording period was selected for analysis to avoid interactions between stimulation, psychostimulants, and tonic dopamine signaling. In the dorsal striatum ([Fig. 6A](#pone-0060763-g006){ref-type=\"fig\"}), AMPH (10 mg/kg) elicited the fastest and largest increase in tonic dopamine levels. Statistical analysis revealed a significant effect of treatment (F~3,22~\u200a=\u200a3.38, *p*\u200a=\u200a0.04), time (F~4,22~\u200a=\u200a11.99, *p*\\<0.001), and interaction (F~12,22~\u200a=\u200a2.13, *p*\u200a=\u200a0.03). A *post hoc* comparison of the average change across the last two minutes of the time course (INSET) revealed that only 10 mg/kg AMPH significantly increased tonic dopamine levels compared to saline (*p*\\<0.01). In the ventral striatum ([Fig. 6B](#pone-0060763-g006){ref-type=\"fig\"}) region, the effects of each psychostimulant were largely indistinguishable from each other and only slightly different than the saline control. Statistical analysis revealed a significant effect of only time (F~4,22~\u200a=\u200a3.90, *p*\u200a=\u200a0.02). Overall, these results suggested that AMPH is more effective at increasing tonic dopamine signaling than cocaine and in the dorsal compared to the ventral striatum initially after drug injection.\n\n![Averaged psychostimulant-induced increases in tonic dopamine levels.\\\n**A.** Time course of the effects of AMPH and cocaine (COC) on tonic dopamine levels. Dopamine concentrations were determined using PCR and averaged across 10-s bins. The time period is the epoch immediately following drug injection and prior to the first post-drug. **B.** Dopamine levels from **A.** above but only shown at two-minute intervals. These data were used for statistical analysis. Data in the dorsal (left) and ventral (right) striatum are expressed as mean \u00b1 SEM. \\*, significantly different from other treatments (*p*\\<0.05).](pone.0060763.g006){#pone-0060763-g006}\n\nPsychostimulant effects on phasic dopamine signaling {#s3d}\n----------------------------------------------------\n\nIncreased \\[DA\\]~max~ of phasic-like dopamine responses evoked by the short train ([Figs. 2](#pone-0060763-g002){ref-type=\"fig\"}, [3](#pone-0060763-g003){ref-type=\"fig\"}, [4](#pone-0060763-g004){ref-type=\"fig\"}) suggests that both amphetamine and cocaine activate phasic dopamine signaling. These results are thus consistent with the two psychostimulants augmenting naturally occurring dopamine transients in awake, freely behaving animals [@pone.0060763-Daberkow1], [@pone.0060763-Aragona1], [@pone.0060763-Wightman2]. While psychostimulant-induced burst firing of dopamine neurons is typically blunted under anesthesia [@pone.0060763-Koulchitsky1] unless revealed by D2 antagonists [@pone.0060763-Paladini1], [@pone.0060763-Shi1], dopamine transients are elicited by AMPH in a subset of animals in this preparation [@pone.0060763-Ramsson2]. An example of this activation is shown in [Figure 7](#pone-0060763-g007){ref-type=\"fig\"}. Before drug injection, the dopamine response evoked by the short train was small and no dopamine transients were observed ([Fig. 7A](#pone-0060763-g007){ref-type=\"fig\"}). In sharp contrast, high-dose AMPH dramatically increased this evoked phasic-like signal, mediated by augmented vesicular dopamine release and inhibited dopamine uptake ([Fig. 4](#pone-0060763-g004){ref-type=\"fig\"} and [Table 1](#pone-0060763-t001){ref-type=\"table\"}), and transient frequency ([Fig. 7B](#pone-0060763-g007){ref-type=\"fig\"}). To better view the presence or absence of dopamine transients, FSCV recordings are expanded in the INSET. These short-lived, non-electrically evoked deflections were identified as dopamine by the sequential voltammograms displayed in the pseudocolor plot below each trace and by the overlay of the individual voltammogram for the transients (black) with that obtained from the evoked signal established to be dopamine (red) to the left in the INSET.\n\n![Representative effects of AMPH on phasic dopamine signaling in the ventral striatum.\\\n**A.** Pre-drug. **B.** Post-AMPH. Traces show 90 s of a recording with a short-duration (0.4 s) stimulation applied at 5 s (see line underneath). The color plot serially displaying all background-subtracted cyclic voltammograms is shown underneath. **INSET.** Time-expanded view. Individual background-subtracted voltammograms are shown at the top left and compare dopamine collected during the evoked phasic-like response (black line) to pre-drug baseline (**A.**) or a dopamine transient collected post-drug (**B.**) as indicated by vertical white line in the pseudocolot plot (red line).](pone.0060763.g007){#pone-0060763-g007}\n\nTo complement evoked phasic-like responses, we thus analyzed these dopamine transients to obtain a more physiological assessment of psychostimulant effects on phasic dopamine signaling. [Figure 8](#pone-0060763-g008){ref-type=\"fig\"} shows the time course of dopamine transients for high-dose AMPH (Panel A) and cocaine (Panel B) in the dorsal and ventral striatum (top and bottom, respectively) in the subset of animals where this phasic activity was observed (see legend for details). Transients were analyzed for frequency (left), amplitude (middle), and duration (right). Time 0 min is drug injection. The time when short, intermediate and long trains were applied during the post-drug period is demarcated by vertical dashed lines at 10, 12 and 22 min, respectively. High-dose AMPH and cocaine activated dopamine transients in both striatal subregions. Transients were rarely observed during pre-drug recording and were not observed after saline or low-dose AMPH. Both psychostimulants increased the frequency of dopamine transients to a greater extent in the ventral compared to the dorsal striatum, and AMPH was more effective than cocaine in both striatal subregions. The onset of dopamine transient activation was also slower for cocaine. A clear inhibition and rebound in transient frequency was observed following the long train in both the dorsal and ventral striatum after AMPH. This effect is most likely related to feedback inhibition by released dopamine [@pone.0060763-Kuhr1], with the additional combination of AMPH and the long train depleting vesicular dopamine release in the dorsal striatum ([Fig. 4](#pone-0060763-g004){ref-type=\"fig\"}). Overall, results for dopamine transients are consistent with those for evoked phasic-like responses and suggest that AMPH and cocaine activate phasic dopamine signaling.\n\n![Averaged effects of psychostimulants on dopamine transients.\\\n**A.** AMPH. **B.** Cocaine. Dopamine transients were analyzed in terms of frequency (left), amplitude (middle) and duration (right) in both the dorsal and ventral striatum (DS and VS, respectively). Data are transient characteristics compiled into 60-s bins and express as the mean \u00b1 SEM. Each histogram shows transient characteristics for the 10 minutes before, and the 65 minutes after drug injection (at time 0 min). Phasic dopamine transients were observed following AMPH in 3 of 7 animals in the dorsal striatum and 5 of 7 in the ventral striatum and following cocaine in 3 of 7 animals in the dorsal striatum and 1 of 7 animals in the ventral striatum.](pone.0060763.g008){#pone-0060763-g008}\n\nDiscussion {#s4}\n==========\n\nThe goal of the present study was to reconcile the paradoxical effects of AMPH on dopamine neurons. To this end, we tested the novel hypothesis that AMPH depletes the reserve pool but up-regulates the readily releasable pool. This hypothesis was formulated based on three key observations reported in the literature. First, dopamine neurons contain distinct vesicular storage pools. Second, different train durations interrogate different vesicular storage pools. And third, AMPH effects on electrically evoked dopamine levels in the dorsal striatum appear inversely related to train duration. We tested this hypothesis using a novel experimental design. When taken together, our results support a model of AMPH activating tonic dopamine signaling by depleting the reserve pool to drive non-exocytotic efflux, but activating phasic dopamine signaling by up-regulating the readily releasable pool to drive vesicular dopamine release.\n\nExperimental Design {#s4a}\n-------------------\n\nFour features highlight the utility of the experimental design. First, different train durations, selected to demonstrate stimulus-dependent AMPH effects, were applied to the same animal. Although this strategy fosters inter-animal comparisons, it also risks train interactions because dopamine release depends upon stimulation history [@pone.0060763-Montague1]. However, stability of the saline control and replicating stimulus-dependent AMPH effects demonstrated previously in separate animals indicated that judicial spacing of trains was sufficient to minimize interaction. Second, evoked dopamine dynamics were resolved into the respective contributions of vesicular release and uptake. Most previous studies examining stimulus-dependent AMPH effects report dopamine levels only and therefore do not directly assess release. Third, the status of dopamine storage pools was related to tonic and phasic dopamine signaling. Such an integrated view of AMPH action has not been available. And fourth, we compared AMPH to cocaine, which is recognized to inhibit DAT and increase vesicular release, but not to deplete vesicular stores *in vivo*.\n\nAMPH enhances tonic and phasic dopamine signaling {#s4b}\n-------------------------------------------------\n\nTonic dopamine signaling, which is characterized by a steady-state basal level of dopamine and controlled by slow irregular firing of dopamine neurons and presynaptic input [@pone.0060763-Venton1], enables movement, cognition and motivation [@pone.0060763-Schultz1]. AMPH robustly increases tonic dopamine levels measured by microdialysis [@pone.0060763-Kuczenski2], but comparatively greater elevations in dialysate dopamine relative to other DAT-inhibiting psychostimulants such as cocaine [@pone.0060763-Kuczenski1] are attributed to the unique action of AMPH eliciting non-exocytotic efflux [@pone.0060763-Fleckenstein1], [@pone.0060763-Sulzer1]. We show here that only high-dose AMPH increased tonic dopamine levels and this only occurred in the dorsal striatum. Analytical differences between measurement techniques may have contributed to discrepancies between the present measures with FSCV and microdialysis studies [@pone.0060763-Borland1]. While FSCV excels at fast measurements with a small probe, inherent limitations in selectivity require the use of statistical methods such as PCR to resolve the dopamine component of tonic changes [@pone.0060763-Keithley1]. Microdialysis exhibits superior selectivity but suffers from implantation damage due to the considerably larger probe that overestimates the increase in tonic dopamine levels with dopamine uptake inhibitors [@pone.0060763-Borland1]. Thus, measurements of tonic dopamine levels using both approaches should be carefully scrutinized. We should also emphasize that a conservative approach with FSCV was used to minimize error, by only characterizing the first 10-min post-drug epoch and by incorporating background drift as a PCR component [@pone.0060763-Hermans1]. Increases in tonic dopamine levels may thus have occurred after this time. Another consideration when comparing the present and microdialysis studies is anesthesia, which inhibits dopamine neuron firing [@pone.0060763-Kelland1]. However, observed effects of saline and low- and high-dose AMPH on tonic dopamine levels are consistent with un-anesthetized recordings [@pone.0060763-Daberkow1]. The contribution of efflux to AMPH-induced increases in tonic dopamine levels measured by FSCV and observed here and elsewhere [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2] has not been determined. However, efflux is implicated using the present experimental design because increased tonic levels are associated exclusively with vesicular depletion.\n\nPhasic dopamine signaling, in which burst firing of dopamine neurons generates sub-second changes in extracellular dopamine called transients [@pone.0060763-OwessonWhite1], is important for goal-directed behavior and reinforcement learning [@pone.0060763-Schultz1]. Cocaine activates burst firing [@pone.0060763-Koulchitsky1], the amplitude, frequency and duration of naturally occurring dopamine transients [@pone.0060763-Aragona1], [@pone.0060763-Wightman2], and evoked phasic-like dopamine responses [@pone.0060763-Ramsson2], [@pone.0060763-Oleson1]. AMPH has also been shown to augment evoked phasic-like dopamine responses, as well as spontaneously occurring and cue-evoked dopamine transients [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2]. Consistent with these previous studies, we show here that both AMPH and cocaine activated evoked phasic-like dopamine responses and dopamine transients. Anesthesia likely attenuated these effects by inhibiting burst firing [@pone.0060763-Kelland1] and phasic activation by psychostimulants [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2], [@pone.0060763-Wightman2], [@pone.0060763-Koulchitsky1], [@pone.0060763-Stuber1]. However, awake, freely behaving animals do not tolerate intermediate and long stimulus trains, so anesthesia is required to assess recycling and reserve pools.\n\nStimulus-dependent effects of AMPH on \\[DA\\]~max~ {#s4c}\n-------------------------------------------------\n\nThe present results, obtained by applying different train durations to the same animal, are consistent with previous work applying these same trains individually in separate animals. For example, in the presence of AMPH and in the dorsal striatum, the long train decreased \\[DA\\]~max~ [@pone.0060763-Stamford1], [@pone.0060763-Kuhr2], [@pone.0060763-Kuhr3], the intermediate train elicited minimal to no effect [@pone.0060763-May1], [@pone.0060763-Ramsson1], and the short train increased \\[DA\\]~max~ [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2]. Similar results were obtained in the ventral striatum, except that the long train did not decrease \\[DA\\]~max~, which is also consistent with previous work [@pone.0060763-Kuhr3]. We additionally extend these studies by comparing AMPH effects to cocaine, which only elicited increases or no change in \\[DA\\]~max~, and by determining the underlying change in vesicular dopamine release, which permits analysis of storage pools. Indeed, because both AMPH and cocaine robustly inhibit dopamine uptake ([Table 1](#pone-0060763-t001){ref-type=\"table\"}, [@pone.0060763-Ramsson1]--[@pone.0060763-Ramsson2], [@pone.0060763-Wu2], [@pone.0060763-John1], [@pone.0060763-Jones1]), observed alterations in \\[DA\\]~max~ have a complex origin.\n\nAMPH elicits opposing actions on readily releasable and reserve pools for dopamine {#s4d}\n----------------------------------------------------------------------------------\n\nWork with model synapses indicates that readily releasable, recycling, and reserve pools of neurotransmitters are interrogated by short, intermediate, and long duration trains, respectively [@pone.0060763-Rizzoli1]. We used this approach to investigate the effects of AMPH on dopamine stores. In the dorsal striatum, each stimulus train elicited a distinct action on vesicular dopamine release in the presence of high-dose AMPH: increase, no change, and decrease for short, intermediate, and long trains, respectively. Taken together, these results suggest that AMPH augments the readily releasable pool, exerts no effect on the recycling pool, and depletes the reserve pool in the dorsal striatum. By contrast, the readily releasable pool and to a lesser extent the recycling pool were up-regulated without depletion of the reserve pool by AMPH in the ventral striatum and cocaine in both striatal sub-regions. As a psychostimulant with multiple actions, AMPH could augment vesicular dopamine release by several mechanisms, such as: (1) inhibiting monoamine oxidase [@pone.0060763-Scorza1] and activating tyrosine hydroxylase [@pone.0060763-Kuczenski3], leading to greater cytosolic dopamine levels, vesicular packaging, and ultimately quantal size; (2) increasing membrane excitability as a DAT substrate [@pone.0060763-Ingram1]; and (3) enhancing exocytosis by liberating vesicular Ca^2+^ stores [@pone.0060763-Mundorf1]. Depleting the reserve pool suggests another mechanism, re-distributed cytosolic dopamine being re-packaged by the readily releasable pool. This latter postulate is supported by the greater capacity of this vesicle population to sequester cytosolic dopamine [@pone.0060763-Fleckenstein2], [@pone.0060763-Volz1]. Moreover, robust depletion of vesicular dopamine stores by AMPH, well established using reduced preparations [@pone.0060763-Anderson1], [@pone.0060763-Chen1], [@pone.0060763-Mundorf1], [@pone.0060763-Bowyer1]--[@pone.0060763-Schmitz1], appears to occur independently in separate classes of dopamine vesicles [@pone.0060763-Anderson1], [@pone.0060763-Chen1]. Depletion involves AMPH acting as a weak base to destabilize the proton gradient across vesicles and as a substrate of the vesicular monoamine transporter to inhibit and/or reverse its action [@pone.0060763-Fleckenstein1], [@pone.0060763-Sulzer1]. How these mechanisms might differ across dopamine storage pools, as our results would suggest, remains to be determined.\n\nWe also do not know why AMPH depleted vesicular dopamine stores in the dorsal but not ventral striatum. One possible origin is regional differences in DAT. For example, DAT binding and V~max~ for dopamine uptake are higher in the dorsal striatum [@pone.0060763-Cass1], [@pone.0060763-Marshall1], and DAT is more glycosylated with a higher molecular weight in the ventral striatum [@pone.0060763-Lew1]. Although K~m~ for dopamine uptake is similar in the two regions [@pone.0060763-Wu2], [@pone.0060763-Marshall1], AMPH is a more potent competitive inhibitor of dopamine uptake in the dorsal compared to the ventral straitum [@pone.0060763-Ramsson1]. We are not aware of comparable regional differences in the vesicular monoamine transporter. Another possible origin is regional differences in vesicular dopamine stores. As mentioned above, different classes of dopamine vesicles exhibit different sensitivities to the depleting actions of AMPH [@pone.0060763-Anderson1], [@pone.0060763-Chen1]. Consistent with region-specific actions of AMPH on vesicular dopamine stores, we have recently shown that AMPH may up-regulate vesicular dopamine release in the ventral striatum by mobilizing the reserve pool but by activating dopamine synthesis and inhibiting dopamine degradation in the dorsal striatum [@pone.0060763-Avelar1]. Different distributions of small, clear and large, dense-core vesicles in the two striatal sub-regions [@pone.0060763-Hondebrink1] may also contribute to the differential response to AMPH. Clearly, more work needs to be done to resolve the differential depleting effects of AMPH on dopamine vesicles in the dorsal and ventral striatum.\n\nNew model of amphetamine action {#s4e}\n-------------------------------\n\nWe propose a new model of AMPH action: activating tonic dopamine signaling by depleting the reserve pool, which elevates cytosolic dopamine and drives reverse transport through DAT, while concurrently activating phasic dopamine signaling by up-regulating the readily releasable pool, which drives vesicular dopamine release. This model is supported here by the first report of a selective coupling between tonic activation and vesicular depletion coincident with phasic activation and up-regulated vesicular release. Revealing this unique combination of AMPH effects underscores the utility of the experimental design employed. Indeed, slice voltammetry has demonstrated a parallel between robust vesicular depletion and micromolar dopamine efflux, but no measures of phasic signaling or its release component were examined [@pone.0060763-Jones2], [@pone.0060763-Patel1], [@pone.0060763-Schmitz1]. Moreover, *in vivo* voltammetry has demonstrated concurrent activation of tonic and phasic dopamine signaling and up-regulated vesicular release, but effects on the reserve pool were not assessed [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2]. Further supporting our proposed model is that, in contrast to AMPH in the dorsal striatum, AMPH in the ventral striatum and cocaine in both striatal sub-regions did not deplete vesicular stores or elevate tonic dopamine levels, despite phasic activation and up-regulated vesicular release.\n\nTwo confounds need addressing. First, coupling between tonic activation and vesicular depletion was not observed for low-dose AMPH in the dorsal striatum. It could be that, while cytosolic dopamine increased as a result of vesicular depletion, low-dose AMPH was insufficient to inhibit monoamine oxidase and prevent its intracellular degradation and/or to reverse DAT direction and cause efflux. Both AMPH effects are dose-dependent [@pone.0060763-Scorza1], [@pone.0060763-Sitte1]. Also consistent with this interpretation is that vesicular depletion alone does not elicit efflux [@pone.0060763-Jones2] and that both vesicular depletion and blockade of monoamine oxidase are required for cytosolic levels to increase [@pone.0060763-Mosharov1]. In contrast, there are other reports demonstrating that increases in cytosolic dopamine alone are sufficient to induce efflux [@pone.0060763-OwessonWhite1], [@pone.0060763-Sulzer2]. Second, low-dose AMPH also did not activate phasic dopamine signaling or vesicular dopamine release in the dorsal striatum. However, this lack of response is an anesthesia artifact, because both are enhanced in awake, freely behaving animals [@pone.0060763-Daberkow1].\n\nImplications for psychostimulant neurobiology {#s4f}\n---------------------------------------------\n\nWe demonstrate fundamentally similar and distinct mechanisms for two major classes of psychostimulants, AMPH representing the so-called dopamine \"releasers\" (i.e., eliciting non-exocytotic efflux) and cocaine representing the DAT \"inhibitors\" [@pone.0060763-John1]. While AMPH and cocaine share phasic activation through augmented vesicular dopamine release ([Fig. 4](#pone-0060763-g004){ref-type=\"fig\"}, [@pone.0060763-Daberkow1], [@pone.0060763-Ramsson2], [@pone.0060763-Oleson1], [@pone.0060763-Jones1], [@pone.0060763-Kile1]--[@pone.0060763-Venton2]) and enhanced burst firing [@pone.0060763-Paladini1], [@pone.0060763-Shi1], [@pone.0060763-Koulchitsky1], they differ in tonic activation. In particular, cocaine requires action potential-dependent mechanisms whereas AMPH does not [@pone.0060763-Benwell1]--[@pone.0060763-Westerink1]. Inhibition of dopamine uptake ([Table 1](#pone-0060763-t001){ref-type=\"table\"}, [@pone.0060763-Ramsson1]--[@pone.0060763-Ramsson2], [@pone.0060763-Wu2], [@pone.0060763-John1], [@pone.0060763-Jones1]) would contribute to augmented tonic and phasic signaling by both psychostimulants. However, activation of vesicular dopamine release may be more important than uptake inhibition, especially for phasic signaling, because release better tracks \\[DA\\]~max~ ([Fig. 4](#pone-0060763-g004){ref-type=\"fig\"}, [@pone.0060763-Daberkow1]).\n\nThe neurobiological implications of these psychostimulant actions are not presently known, but they could be profound. Several drugs of abuse have now been demonstrated to augment dopamine transients, including amphetamine, cocaine, nicotine and ethanol [@pone.0060763-Daberkow1], [@pone.0060763-Wightman2], [@pone.0060763-Stuber1], [@pone.0060763-Cheer2], [@pone.0060763-Robinson2]. The greater activation of phasic dopamine signaling by abused drugs compared to natural rewards and the subsequent usurpation of normal reward processing to promote addiction [@pone.0060763-Hyman1] may thus represent a unifying mechanism. While both classes of psychostimulants would promote reinforcement learning by activating the direct (\"Go\") pathway in the basal ganglia via enhanced phasic signaling and D1 receptor binding, AMPH would more robustly inhibit the indirect (\"No Go\") pathway (i.e., disinhibition of behavior) via enhanced tonic signaling and D2 receptor binding [@pone.0060763-Wiecki1], because of the added contribution of non-exocytotic efflux. Future directions should also investigate how intrastriatal differences in AMPH action relate to the diverse roles of dopamine signaling in this region for promoting drug reinforcement and addiction [@pone.0060763-Everitt1], [@pone.0060763-Volkow1].\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: PAG DPC. Performed the experiments: DPC. Analyzed the data: DPC. Wrote the paper: PAG DPC SAJ. Assisted with statistical analysis: SAJ.\n"}
+{"text": "Background\n==========\n\nDilated cardiomyopathy (DCM) is a myocardial disease characterized by dilatation of the left ventricle, reduced systolic function and increased sphericity of the left ventricle. This disease has been described in different species and multiple genes have been found in the human \\[[@B1]\\], mouse \\[[@B2]\\] and hamster \\[[@B3]\\] causing DCM. These genes mainly encode cyto-skeletal components of the cardiac myocytes and can be divided into sarcomeric and extra-sarcomeric proteins. The identified sarcomeric proteins involved in DCM include \u03b1-cardiac actin, encoded by *ACTC*\\[[@B4]\\], cysteine-rich protein 3 (*CSRP3*) \\[[@B5]\\], LIM-domain binding factor 3 (*LDB3*, also known as *Cypher*or *ZASP*) \\[[@B6]\\], myosin heavy polypeptide 7 (*MYH7*) \\[[@B7]\\], titin cap (*TCAP*) \\[[@B8]\\], \u03b1-tropomyosin (*TPM1*), troponin I (*TNNI3*) \\[[@B9]\\], troponin T (*TNNT2*) \\[[@B7]\\], titin (*TTN*) \\[[@B10]\\] and vinculin (*VCL*) \\[[@B11]\\]. The extra-sarcomeric proteins implicated in DCM are encoded by the genes including caveolin 1 (*CAV1*) \\[[@B2]\\], desmin (*DES*) \\[[@B12]\\], lamin A/C (*LMNA*) \\[[@B13]\\], phospholamban (*PLN*) \\[[@B14]\\] and sarcoglycan \u03b4 (*SGCD*) \\[[@B3]\\]. The genes encoding all of the above proteins are located on the autosomal chromosomes. X-linked genes implicated in DCM include dystrophin (*DYS*) \\[[@B15]\\] and tafazzin (*TAZ*) \\[[@B16]\\]. In addition, mitochondrial dysfunction and mitochondrial DNA (mtDNA) mutations have been associated with maternally inherited DCM \\[[@B17]\\]. Furthermore, DCM has also been described with arrhythmias, with mutations in genes encoding sodium \\[[@B18]\\] and potassium channels \\[[@B19]\\].\n\nDCM has been described in many different breeds of mostly giant and large dogs, including the Dobermann \\[[@B20]\\], Great Dane \\[[@B21]\\], Newfoundland \\[[@B22]\\] and Irish Wolfhound \\[[@B23]\\]. Clinical variation exists in the presentation and progression of DCM between different dog breeds and breed specific variation has also been found in histological findings in DCM-affected hearts tissue \\[[@B24]\\]. Since clinical DCM may be a late onset disease, following a long pre-symptomatic course, dogs are often used for breeding before the disease becomes apparent \\[[@B25]\\]. So far, no causative mutation has been discovered in canine DCM. The phenotype of the adult onset forms of canine DCM in most breeds is consistent with a defect in components of the cytoskeleton.\n\nOf the 14 autosomal DCM candidate genes for the dog, *ACTC*, *CAV1*, *CSRP3*, *DES*, *LDB3*, *LMNA*, *MYH7*, *PLN*, *SGCD*, *TCAP*, *TNNI3*, *TNNT2*, *TPM1*and *VCL*, genomic information and/or polymorphic markers were already available for *ACTC*\\[[@B26],[@B27]\\], *DES*\\[[@B28]\\], *PLN*\\[[@B29]\\], *SGCD*\\[[@B30]\\] and *TPM1*\\[[@B31]\\]. In this article, we describe a complete set of polymorphic markers for these 14 candidate genes for canine DCM. The markers, both microsatellites and Single Nucleotide Polymorphisms (SNPs), provide a useful tool to perform linkage and association studies between each of these genes and DCM in the different dog breeds. Furthermore, we present the annotation of 14 candidate genes in the canine genome, which will facilitate mutation screening of these genes.\n\nGenomic Annotation\n==================\n\nThe 14 canine DCM candidate genes were identified on the canine genome by means of a BLAST analysis \\[[@B32]\\], using available canine and human DNA sequences as a query (Table [1](#T1){ref-type=\"table\"}). The exons were identified based on the corresponding human exon sequence (retrieved from \\[[@B33]\\], Table [1](#T1){ref-type=\"table\"}). Each gene was found to be covered by 1 to 5 contigs of the *Canis familiaris*genome build 1.1. (Additional file [1](#S1){ref-type=\"supplementary-material\"} and Table [1](#T1){ref-type=\"table\"}). *CAV1*was covered by 2 neighbouring contigs and the 3 coding exons matched the human ones. Exon 1 of the dog seemed to have an extra nucleotide (T, position 336 of \\[Genbank: [AAEX01048547](AAEX01048547)\\]) compared to human exon 1 of *CAV1*. However, this nucleotide was not present in the single trace file of the *Canis familiaris*Trace Archive \\[[@B34]\\] covering this sequence. Canine DES had 1 amino acid less than the human protein. The canine LDB3 protein is 67 amino acids shorter than human LDB3. Canine LMNA had 1 amino acid extra compared to the human protein. Exon 24 of canine *MYH7*seemed to have 1 bp extra (G, bp 7,902 of \\[Genbank: [AAEX01041100](AAEX01041100)\\]), however, this nucleotide was not present in any of 11 *Canis familiaris*trace sequences covering this position. Without this extra nucleotide, canine exon 24 matched the human exon. Canine TNNI3 had 1 amino acid extra compared to the human protein. For *TNNT2*, coding exons 1, 15 and 16 could not be recognized in canine genomic contigs. *TNNT2*exon 6 showed 1 extra bp compared to human (G, bp 5622 of \\[Genbank: [AAEX01013360](AAEX01013360)\\]), however, this nucleotide was not found in the 2 traces covering this DNA sequence. Without this additional bp, exon 6 matched the corresponding human exon exactly in length. Exon 12 had 1 codon less than the human gene. Exon 13 was located at the end of genomic contig \\[Genbank: [AAEX01013360](AAEX01013360)\\] and although its terminal 2 putative bp were not included in this contig, exon 12 seemed to match the human exon. For the remaining candidate genes, *ACTC*, *CSRP3*, *PLN*, *SGCD*, *TCAP*, *TPM1*and *VCL*, the annotated canine exons matched the corresponding human exons exactly. We could not identify non-coding exons. Apparently, the conservation of these exons is too low for identification purposes. Complementary DNA sequencing is necessary to identify these non-coding exons. All of the predicted introns of the 14 candidate genes started and ended with the canonical GT and AG dinucleotides, respectively \\[[@B35]\\]. Even though a high quality DNA sequence of the canine genome has recently become available, it has not yet been fully annotated.\n\n###### \n\nAssignment, genomic location and the degree of sequence conservation compared to human of the canine DCM candidate genes.\n\n  ------------------------------------------------------------------------------------------------------------------------------------\n  **Gene**         **Annotation of canine gene**   **CFA**   **Similarity to human ^4^**                               \n  ---------------- ------------------------------- --------- --------------------------------------------- ------ ---- ---------------\n  ***ACTC***       AF203019 (C),                   290378    13478                                         377    30   100% AAB59619\n\n                   AF203020 (C)                                                                                        \n\n  ***CAV1***       U47060 (C)                      341049    48546, 48547                                  178    14   96% NP_001744\n\n  ***CSRP3***      BC024010 (H)                    265968    17412                                         194    21   99% AAH24010\n\n  ***DES***        BK005142 (C)                    273074    55032                                         469    37   97% NP_001918\n\n  ***LDB3***       NM_007078 (H)                   361816    16582, 16583^(5)^, 16584                      660    4    79% AB014513\n\n  ***LMNA***       AF427092 (C)                    310777    12733, 12734                                  665    7    98% CAI15522\n\n  ***MYH7***       NM_000257 (H)                   355349    41099, 41100                                  1935   8    98% NP_000248\n\n  ***PLN***        Y00399 (C)                      357525    14037                                         52     1    96% CAI21610\n\n  ***SGCD***       NM_000337 (H)                   303006    16848, 16849^(5)^, 16850^(5)^, 16851, 16852   289    4    98% NP_000328\n\n  ***TCAP***       NM_003673 (H)                   309889    22011                                         167    9    90% CAA09479\n\n  ***TNNI3***      AF506750 (C)                    344887    53923                                         211    1    95% CAG46782\n\n  ***TNNT2***\\     NM_000364 (H)                   367317    13359, 13360                                  254    7    90% NP_000355\n  **(ex 2--14)**                                                                                                       \n\n  ***TPM1***       NM_000366 (H)                   288398    08742                                         284    30   99% AAH07433\n\n  ***VCL***        NM_003373 (H)                   211998    16404                                         1134   4    99% NP_054706\n  ------------------------------------------------------------------------------------------------------------------------------------\n\n^1^Sequence used to identify the canine gene in the dog genome, Genbank accession numbers; C = canine sequence, H = human sequence; ^2^Transcript ID numbers of human annotation \\[33\\] used to annotate the canine gene; ^3^canine genomic contig in which the gene\\'s coding exons were identified; ^4^the percentage identity of each canine protein compared to the human protein (Genbank accession number is listed); ^5^canine genomic contig containing only intronic sequence.\n\nThe conservation of the coding region of each gene was assessed by BLAST comparison of the cDNA and derived amino acid sequences with those of human (at the website of NCBI \\[[@B36]\\], BLASTN and TBLASTX analysis, respectively). The percentages of identity at the nucleotide level varied between 88 and 95% (Table [1](#T1){ref-type=\"table\"}). At the amino acid level, the percentages of identity varied in general between 90--100%, except for the canine LDB3 protein, that was 79% identical to the human protein. The canine ACTC protein appeared to be identical to the human protein. In *LDB3*, a relatively low percentage of identity was found between the canine and human gene, both at the cDNA and the protein level. This was caused by the large (inframe) loss of part of exons (i.e. 4, 7, 8 and 9) compared to the human gene: the canine gene had 660 codons, the human gene had 734 codons.\n\nThe chromosomal position of the 14 canine candidate genes can be found in Table [1](#T1){ref-type=\"table\"}.\n\nWhen analysing the location of the genes in the dog genome (Table [1](#T1){ref-type=\"table\"}), using the canine-human comparative map of Guyon et al. \\[[@B37]\\], each was found to be syntenic to the human location.\n\nPolymorphisms\n=============\n\nSingle Nucleotide Polymorphism detection\n----------------------------------------\n\nWe used denaturing high-performance liquid chromatography (DHPLC) analysis for the detection of SNPs in amplified genomic canine DNA fragments. Polymorphisms were assessed in DNA from Newfoundland dogs. For each gene, several DNA fragments of approximately 500 bp were selected based on melting profile (analyzed with WAVEMAKER\u2122 software from Transgenomic) with a maximum of 2 melting temperatures covering each product. The melting behaviour of a fragment depends on the fragment\\'s DNA sequence. Primers were designed using Primer3 \\[[@B38]\\] and annealing temperatures of the PCRs were optimized (Table [2](#T2){ref-type=\"table\"}). Touchdown PCR amplification of these fragments was performed with DNA of Newfoundland dogs (n = 16; 8 unrelated founders of a pedigree of Newfoundland dogs and 8 family members), using HotStartTaq DNA Polymerase (Qiagen). The Touchdown (TD) PCR program consisted of a denaturing step of 5 min at 95\u00b0C, followed by 14 cycles of 95\u00b0C 30 sec, Ta +7\u00b0C 30 sec, 72\u00b0C 20 sec, with a Ta decrease of 0.5\u00b0C/cycle, followed by 25 cycles of 30 sec at 94\u00b0C, 30 sec at Ta\u00b0C, 30 sec at 72\u00b0C, followed by a final extension at 72\u00b0C for 2 min (Ta in Table [2](#T2){ref-type=\"table\"}). Subsequently, a heteroduplex formation step was carried out to allow formation of hetero- and homo-duplex products; the PCR products were heated 5 min at 95\u00b0C, after which the temperature was decreased gradually (38 cycles of 1 min, temperature decreasing 1.5\u00b0C/cycle), followed by a final step of 5 min at 10\u00b0C. Mutation analysis of the PCR products, based on the presence of heteroduplexes, followed on a WAVE instrument (WAVE Nucleic Acid Fragment Analysis System, Transgenomic). Multiple WAVE patterns of a single PCR fragment in different dogs pointed at existence of both homoduplexes and heteroduplexes and, therefore, indicated potential presence of SNPs in the fragment. In that case, the PCR fragment (of at least of 2 dogs per WAVE pattern) was cleaned (Shrimp Alkaline Phosphatase/ExoI) and the DNA sequence was obtained to determine the identity of the SNPs, by a commercial company (Lark Technologies\u2122, UK).\n\n###### \n\nSingle Nucleotide Polymorphisms in the DCM candidate genes. For each SNP its origin, its primers and the PCR conditions, and its informativity are listed.\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------\n  **Gene**      **dbSNP**\\         **SNP**              **Primers (5\\'-3\\')**\\      **Ta (\u00b0C)^1^**   **Prod. size (bp)**   **Informativeness^2^**   \n                **access. no.**\\                        **Forward;**\\                                                                               \n                **ss4985\\...**                          **Reverse**                                                                                 \n  ------------- ------------------ -------------------- --------------------------- ---------------- --------------------- ------------------------ ----\n  ***ACTC***    2973               5,452 G/A^a,3,4^     gccctggattttgagaatgagat\\    62.0 ^1^         1067                  0.14                     12\n                                                        acgatcagcaataccagggtaca                                                                     \n\n  ***CAV1***    2978               30,312 A/G ^b,5^     tgagtgccttgcttgtgg\\         62.0             565                   0.28                     24\n                                                        gcatcattggaacttgttgg                                                                        \n\n                2979               30,088 G/A^5^        tgagtgccttgcttgtgg\\         62.0             565                   0.24                     24\n                                                        gcatcattggaacttgttgg                                                                        \n\n  ***CSPR3***   2980               31,216 A/G ^6^       ggaggccaggatgagaac\\         62.0             507                   0.15                     22\n                                                        gtttattgtactgaatgatggtcag                                                                   \n\n                2981               25,753 T/C ^6^       aatcatcctcccattgttcc\\       58.0             510                   0.37                     24\n                                                        cagaagtgctcatagtctttaccc                                                                    \n\n                2982               25,446 A/G ^6^       aatcatcctcccattgttcc\\       58.0             510                   0.24                     24\n                                                        cagaagtgctcatagtctttaccc                                                                    \n\n                2983               28,779 A/G ^6^       atggacctttgtatctccag\\       58.0             455                   0.19                     24\n                                                        tctgtaggtttcattcattgg                                                                       \n\n                2984               28,742 C/A ^6^       atggacctttgtatctccag\\       58.0             455                   0.19                     24\n                                                        tctgtaggtttcattcattgg                                                                       \n\n                2985               28,737 G/A ^6^       atggacctttgtatctccag\\       58.0             455                   0.19                     24\n                                                        tctgtaggtttcattcattgg                                                                       \n\n                2986               28,642 T/A ^6^       atggacctttgtatctccag\\       58.0             455                   0.19                     24\n                                                        tctgtaggtttcattcattgg                                                                       \n\n  ***DES***     2989               15,228 C/T ^7^       cgtcacaacccccacaag\\         67.0             530                   0.30                     8\n                                                        gctgggtgccatgaggtc                                                                          \n\n                2990               15,224 C/G ^7^       cgtcacaacccccacaag\\         67.0             530                   0.19                     8\n                                                        gctgggtgccatgaggtc                                                                          \n\n                2991               15,166 G/A ^7^       cgtcacaacccccacaag\\         67.0             530                   0.19                     8\n                                                        gctgggtgccatgaggtc                                                                          \n\n                2992               15,006 C/T ^7^       cgtcacaacccccacaag\\         67.0             530                   0.19                     8\n                                                        gctgggtgccatgaggtc                                                                          \n\n                2993               19,903 T/C ^7^       agggcagagggagaccag\\         66.0             575                   0.30                     8\n                                                        gacctaatggtgggctttacc                                                                       \n\n                2975               19,196 C/T ^c,7,8^   ttgcttgaccactaccagga^9^\\    57.0 ^1^         402                   0.35                     12\n                                                        agatgttcttagccgcgatg^10^                                                                    \n\n                2976               19,105 G/A ^7,8^     ttgcttgaccactaccagga^9^\\    57.0 ^1^         402                   0.30                     12\n                                                        agatgttcttagccgcgatg^10^                                                                    \n\n  ***LDB3***    2987               14,090 C/T ^11^      tgttaatcacctctgcggatagt\\    58.0             540                   0.33                     24\n                                                        ggctccctacacgttgatg                                                                         \n\n                2988               25,205 T/C ^12,d^    gcctcctccatcctgacc\\         66.0             566                   0.19                     24\n                                                        cctcccagtaccctgtaggc                                                                        \n\n                s2974              25,452 A/G ^12^      gcctcctccatcctgacc\\         66.0             566                   0.38                     24\n                                                        cctcccagtaccctgtaggc                                                                        \n\n  ***PLN***     2994               51,818 A/G ^13^      tggtttgccttcatacactacaac\\   64.0             573                   0.21                     14\n                                                        tgtcttcatctgtgggattttg                                                                      \n\n  ***SGCD***    2995               30,703 G/C ^14^      ccttcagacccccatctagg\\       66.0             521                   0.36                     8\n                                                        ccacctgacataatcccactttag                                                                    \n\n                2996               151,312 A/G ^14^     ggaggtagcaaagtatagtgctc\\    62.0             558                   0.30                     8\n                                                        atgttcatgccaacaagc                                                                          \n\n                2997               29,656 C/G ^14^      ttccagccaactgagaagc\\        58.0             525                   0.30                     8\n                                                        cactgtcatttccatgtcaacc                                                                      \n\n                2998               116,470 A/G ^14^     gcaatctcctcctccagacc\\       58.0             529                   0.38                     8\n                                                        tcatggcctcactctgatctc                                                                       \n\n  ***TCAP***    2999               28,606 C/T ^15,e^    gctgcttcccttgaatgc\\         64.0             588                   0.28                     24\n                                                        cagacagtggcaggaatcg                                                                         \n\n                2977               29,957 T/C ^15,f^    gtagagggtagcagatttcagg\\     69.0             555                   0.26                     16\n                                                        ctctgggcaaactacaaagc                                                                        \n\n                3000               30,330 A/G ^15,g^    tgctttgtagtttgcccagag\\      64.0             557                   0.30                     8\n                                                        agccagccaccctgtttac                                                                         \n\n                3001               30,687 C/T ^15,h^    tgctttgtagtttgcccagag\\      64.0             557                   0.30                     8\n                                                        agccagccaccctgtttac                                                                         \n\n  ***TNNT2***   3002               10,466 C/T ^16^      tgaccctcacttggggaac\\        58.0             519                   0.38                     24\n                                                        cgcagggctcttccagac                                                                          \n\n                3003               10,577 T/C ^16^      tgaccctcacttggggaac\\        58.0             519                   0.38                     24\n                                                        cgcagggctcttccagac                                                                          \n\n                3004               10,671 T/C ^16^      tgaccctcacttggggaac\\        58.0             519                   0.38                     24\n                                                        cgcagggctcttccagac                                                                          \n\n  ***VCL***     3005               177,743 G/A^17^      tgcaggccacagagatgc\\         62.0             491                   0.30                     8\n                                                        ggaatgagggcggagcag                                                                          \n  ------------------------------------------------------------------------------------------------------------------------------------------------------\n\n^1^All PCR program were Touchdown (TD) at the listed Ta, accept for the *ACTC*SNP: 94\u00b0C 5 min, 35\u00d7 (94\u00b0C 30 sec, 62\u00b0C 1 min, 72\u00b0C 1 min), 72\u00b0C 10 min, 20\u00b0C \u221e; and for the *DES*SNPs 19,196 C/T and 19,105 G/A: 94\u00b0C 10 min, 35\u00d7 (94\u00b0C 30 sec, 57\u00b0C 30 sec, 72\u00b030 sec), 72\u00b0C 10 min, 20\u00b0\u221e; ^2^The informativeness of each SNP was described by its polymorphism information content (PIC), based on the number of genotyped chromosomes (\\#chr) listed; ^3^SNP detected while sequencing available *ACTC*SNPs (166C/T and 38C/T of \\[Genbank: [AF203019](AF203019)\\] and 289T/A of \\[Genbank: [AF203020](AF203020)\\], these were in the dog genome, respectively, 4,871G/A, 5,000 G/A and 5,454 A/T in \\[Genbank: [AAEX01013478](AAEX01013478))\\] \\[26\\]; ^4^in genomic contig AAEX01013478; ^5^AAEX01048546; ^6^AAEX01017412; ^7^AAEX01055032; ^8^SNP detected while sequencing available *DES*SNPs (1,808C/T and 1,851G/C of \\[Genbank: [BK005142](BK005142)\\], these were in the dog genome, respectively, 19,262 G/A and 19,218C/G in \\[Genbank: [AAEX01055032](AAEX01055032)\\]) \\[28\\]; ^9^M13-tailed F-primer: 5\\'- GTTTTCCCAGTCACGAC\\-\\-\\-- 3\\'; ^10^M13-tailed R-primer: 5\\'- CAGGAAACAGCTATGAC\\-\\-\\--3\\'; ^11^in genomic contig AAEX01016584; ^12^AAEX01016582; ^13^AAEX01014037; ^14^AAEX01016848; ^15^AAEX01022011; ^16^AAEX01013360; ^17^AAEX01016404; ^a^is identical to SNP BICF237J37997 (Broad, at \\[39\\]); ^b^identical to BICFPJ1220038; ^c^identical to BICFPJ152241; ^d^identical to BICF231J18538; ^e^identical to BICFG630J165218; ^f^identical to BIFG630J165217; ^g^identical to BICFG630J165215; ^h^identical to BICFG630J165213.\n\nTwenty-eight SNPs were discovered by WAVE analysis (Table [2](#T2){ref-type=\"table\"}). No indication of the presence of a SNP was found in WAVE fragments of *LMNA*, *MYH7*and *TNNI3*(3, 5 and 3 fragments analyzed, respectively). One new SNP, TCAP SNP 29,957 T/C in genomic contig \\[Genbank: [AAEX01022011](AAEX01022011)\\], was found when we resequenced a *TCAP*fragment in a group of Newfoundland dogs. WAVE analysis of this fragment had not indicated presence of a potential SNP -- although the obtained DNA sequences showed that both homozygous and heterozygous animals were among the dogs used for WAVE analysis. Conversely, sometimes WAVE analysis indicated potential presence of SNPs, yet sequencing of dogs with different WAVE patterns did not confirm these. This could be due to the sequencing procedure used.\n\nIn search of additional SNPs for canine *ACTC*and *DES*, genomic DNA fragments containing SNPs annotated by others (Table [2](#T2){ref-type=\"table\"}) were resequenced. After PCR amplification of these fragments, 1 \u03bcl of 1:15 diluted PCR product was used in a Tercycle big dye reaction with the F-PCR-primer for the ACTC SNP and a HPLC-purified M13 F-primer (5\\'-GTTTTCCCAGTCACGAC-3\\') for the DES SNPs. The Tercycle consisted of 25 cycles of 30 sec at 96\u00b0C, 15 sec at 55\u00b0C and 2 min at 60\u00b0C. After purification (Sephadex TM G50 Superfine, Amersham Biosciences), each product was processed with an ABI PRISM^\u00ae^3100 Genetic Analyzer (Applied Biosystems). Five SNPs (*ACTC*5,452G/A; *DES*19,196C/T and 19,105G/A; *LDB3*25,452A/G and *TCAP*29.957 T/C) were identified by resequencing areas of earlier described SNPs (Table [2](#T2){ref-type=\"table\"}).\n\nOf the total of 33 identified SNPs, 4 were in coding regions (*DES*15,006C/T, *LDB3*14,090C/T, *TCAP*29,957T/C and *TNNT2*10,466C/T). These exonic SNPs, however, did not cause polymorphisms at the amino acid level. Comparing the 33 newly discovered SNPs to the dog SNP database of the Broad Institute \\[[@B39]\\] showed 25 of our SNPs to be new, the remaining 8 SNPs matched SNPs present in the Broad database (see Table [2](#T2){ref-type=\"table\"}). This indicates that, in addition to the many SNPs that have become available by random sequencing of the dog genome, many more canine SNPs exist. Our limited search for SNPs in 14 DCM candidate genes took place in a single breed, the Newfoundland dog. However, a high percentage of SNPs found in one breed can be expected to be polymorphic in other breeds too \\[[@B40]\\]. All identified SNPs were submitted to dbSNP and the respective accession numbers are listed in Table [2](#T2){ref-type=\"table\"}.\n\nDetection of microsatellite polymorphisms\n-----------------------------------------\n\nSimple DNA sequences composed of CA, GAAA or GA repeats were identified in the genomic contigs that contain the candidate genes or in neighbouring contigs. For *VCL*, a polymorphic microsatellite became available through personal communication with P.Stabej (Table [3](#T3){ref-type=\"table\"}; a repeat was obtained from BAC RP81-251B5, isolated using methods as described in \\[[@B28]\\] with an overgo probe based on murine *VCL*exon 17, F-overgo CCAAGGTCAGAGAAGCCTTCCAAC, R-overgo AAGTCAGGCTCCTGAGGTTGGAAG). Primers were designed from the DNA sequence flanking the repeats and the forward primer was fluorescently labelled with 6-FAM or HEX. For some microsatellites, a 3-primer protocol was used for the PCR amplification (Table [3](#T3){ref-type=\"table\"}), using an M13-tailed (GTTTTCCCAGTCACGAC\\-\\-\\-\\-- (5\\'-3\\')) F-primer, a 6-FAM-labelled M13 primer (GTTTTCCCAGTCACGAC (5\\'-3\\')) and a R-primer. Genotyping PCR reactions were incubated 12 min at 94\u00b0C, followed by 35 cycles of 10 sec at 94\u00b0C, 15 sec at Ta\u00b0C and 30 sec at 72\u00b0C, and a final step of 20 min at 72\u00b0C (Ta in Table [3](#T3){ref-type=\"table\"}). An ABI PRISM ^\u00ae^3100 Genetic Analyzer (Applied Biosystems) was used for genotyping and allele sizes were determined with Genescan Analysis 3.7 and Genotyper 3.7 software (Applied Biosystems). Eleven polymorphic microsatellites were developed for *ACTC*(2 markers), *CAV1*(1), *CSRP3*(1), *LMNA*(2), *MYH7*(1), *TNNI3*(3) and *VCL*(1) (Table [3](#T3){ref-type=\"table\"}). The markers, mostly CA-repeats, showed multiple allele sizes (2--6 alleles/marker) in a group of 16 Newfoundland dogs (Table [3](#T3){ref-type=\"table\"}). To describe the informativeness of our microsatellite markers, the polymorphism information content (PIC) was obtained based on the genotypes of unrelated founders of a family of Newfoundland dogs (Table [3](#T3){ref-type=\"table\"}). According to \\[[@B41]\\], 2 of the 11 newly designed microsatellites were considered highly informative (PIC\\>0.50), 7 reasonable informative (0.25\\<PIC\\<0.50) and 2 slightly informative (PIC\\<0.25) in the Newfoundland founder dogs. Besides the 11 polymorphic microsatellites, 2 other markers were found to be monomorphic in the group of Newfoundland dogs, but might be polymorphic in other breeds. This was a *MYH7*CA-repeat (position 11,730 of \\[Genbank: [AAEX010141100](AAEX010141100)\\]) and a *TNNI3*CA-repeat (position 17,739 of \\[Genbank: [AAEX01053915](AAEX01053915)\\]. An already available microsatellite for TPM1 \\[[@B31]\\] was shown to be highly informative in our group (Table [3](#T3){ref-type=\"table\"}).\n\n###### \n\nPolymorphic microsatellite markers for canine DCM candidate genes^1^\n\n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  **Gene**   **Repeat**   **Primers (5\\'-3\\')**\\                            **Ta (\u00b0C**)   **Detected alleles**   **Informativeness^2^**   **Origin (bp \\... of contig AAEX 010\\...)**   **Distance to gene^3^**                         \n                          **Forward;**\\                                                                                                                                                                                                 \n                          **Reverse**                                                                                                                                                                                                   \n  ---------- ------------ ------------------------------------------------- ------------- ---------------------- ------------------------ --------------------------------------------- ------------------------- --------------------- ------------------------\n  *ACTC*     15CA         actccgaagaaggaagtcaac^4^\\                         57.0          234--238               2                        0.17                                          10                        bp 37,720/..13479     69.2 kb downstr. Stop\n                          gttcccatctatgagggctat                                                                                                                                                                                         \n\n             20CA         ggaacaaggtgctgttagacc^5^\\                         59.0          338--356               5                        0.42                                          24                        bp 5,945/..13478      intragenic (intron 4)\n                          cacattccaccgagtaggc^7^                                                                                                                                                                                        \n\n  *CAV1*     13CA         ccacagagctagaaagctacg^4^\\                         54.5          240--242               2                        0.28                                          24                        bp 39,315/\\...48546   8.2 kb downstr. Stop\n                          tgttgcaaacaccctatgat                                                                                                                                                                                          \n\n  *CSRP3*    15CA         catgtcctgcaagttaatggt^4^\\                         53.0          237--245               3                        0.41                                          24                        bp 38,367/..17412     2.8 kb upstr. Start\n                          ggatttctattctgggtttcc                                                                                                                                                                                         \n\n  *LMNA*     16CA         gggtggtagatgagcatttc^6^\\                          54.5          204--212               4                        0.36                                          12                        bp 5,128/..12734      9.3 kb downstr. Stop\n                          gaagagaacaagtgggcaag                                                                                                                                                                                          \n\n             18GAAA       ggaagatgagactgttagaatgc^5^\\                       57.0          321--344               6                        0.67                                          24                        bp 26,754/..12735     22.6 kb downstr. Stop\n                          caggccatgattacttttcc^7^                                                                                                                                                                                       \n\n  *MYH7*     21CA         gatatcctgggattaaagactgg^5^\\                       58.0          351--363               4                        0.37                                          24                        bp 1,418/..41098      36.6 kb downstr. Stop\n                          ctattttgccctcttcatgg^7^                                                                                                                                                                                       \n\n  *TNNI3*    20CAa        tcaaacagggaaacctgaac^6^\\                          57.0          297--301               3                        0.38                                          24                        bp 597/..53929        119.3 kb upstr. Start\n                          gattattcagctcccagaacc^7^                                                                                                                                                                                      \n\n             20CAb        ttccagttgattgtttctctgc^5^gcggtttagcactgcattc^7^   59.0          302--306               2                        0.08                                          24                        bp 13,248/..53916     110.2 kb downstr. Stop\n\n             17GA         tccaacctcagggtactgg^5^catgccatggagctatgc^7^       59.0          304--312               3                        0.37                                          24                        bp 48,910/..53930     179.8 kb upstr. Start\n\n  *TPM1*     19CA^8^      actgtgtccagagtgcagcta^4^\\                         60.0          467--483               4                        0.67                                          12                        bp 88,113/..08742     6.5 kb downstr. Stop\n                          gattgctagactggc                                                                                                                                                                                               \n\n  *VCL*      15GAAA^9^    caatttcttttccaatcacattag ^10^\\                    54.0          150--170               6                        0.69                                          24                        bp 12,680/..16406     88.6 kb downstr. Stop\n                          gccattttgcattctcttcaa                                                                                                                                                                                         \n  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n^1^Microsatellites for *DES*and *SGCD*were demonstrated in, respectively, \\[28\\] and \\[30\\]; ^2^The informativeness of each SNP was described by its polymorphism information content (PIC), based on the number of genotyped chromosomes (\\#chr) listed; ^3^Based on genomic build 1.1; ^4^F-primer fluorescently labelled with 6FAM; ^5^Three-primer protocol used; ^6^F-primer fluorescently labelled with HEX; ^7^extra tail on R-primer: GTGTCTT\\-\\-\\-- (5\\'-3\\') to promote addition of an Adenosine residue at the 3\\'-end of the complementary DNA strand; ^8^Microsatellite demonstrated in \\[31\\]; ^9^Personal communication P.Stabej; ^10^F-primer fluorescently labelled with TET.\n\nThe distance between the microsatellite and the corresponding gene was derived from the dog genome build 1.1 \\[[@B42]\\] and can be found in Table [3](#T3){ref-type=\"table\"}. This distance varied from zero for an intragenic microsatellite to 179.8 kb. The genomic locations of polymorphic microsatellites, already available for *DES*, *SGCD*, *TPM1*and *VCL*, were determined. For *DES*a CA-repeat \\[[@B28]\\] was located at position 5,688 of \\[Genbank: [AAEX01055032](AAEX01055032)\\], 9.0 kb downstream of the stop codon. For *SGCD*both a GAAA-repeat and a CA-repeat were available \\[[@B30]\\]. The first was located at position 76,364 of \\[Genbank: [AAEX4801016848](AAEX4801016848)\\], the second at position 42,047 of the same genomic contig and both markers are in intron 7 of SGCD. For *TPM1*a GA-repeat \\[[@B31]\\] was located at position 88,113 of \\[Genbank: [AAEX01008742](AAEX01008742)\\], 6.5 kb downstream of the stop codon. A polymorphic GAAA-repeat for *VCL*showed to be located at position 12,680 of \\[Genbank: [AAEX01016406](AAEX01016406)\\] in the dog genome, 88.6 kb downstream of the stop codon.\n\nConclusion\n==========\n\nWith the annotation of these 14 candidate genes for DCM and the identification of polymorphic markers, the genes can be evaluated for the involvement in breed specific DCM. The SNPs and microsatellites presented in this paper are a powerful tool to analyse linkage between the fourteen candidate genes encoding cytoskeletal proteins and DCM in the dog. The annotation of each gene facilitates screening of these genes for mutations in naturally occurring canine DCM in specific breeds, potential models for forms of human DCM.\n\nAuthors\\' contributions\n=======================\n\nACW carried out the molecular genetic studies and drafted the manuscript. PAJL participated in the design of the study, helped to draft the manuscript. BAvO participated in the design of the study and was co-applicant for funding. WEO participated in the coordination of the study. JDMcE conceived of the study and was main applicant for funding. She phenotyped the dogs, collected the samples and extracted the genomic DNA.\n\nAll authors had read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\nOverview of the genomic organization of the canine *ACTC*(A), *CAV1*(B), *CSRP3*(C), *DES*(D), *LDB3*(E), *LMNA*(F), *MYH7*(G), *PLN*(H), *SGCD*(J), *TCAP*(K), *TNN-I3*(L), *TNN-T2*(M), *TPM1*(N) and *VCL*(P) gene, in build 1.1 of the canine genome. The size of each coding exon, its actual location in bp in the respective genomic contig, 10 bp of DNA sequence at the 5\\'end and 3\\'end of the exon, 10 bp of the flanking intron and the intron sizes are listed. In case the coding sequence of a gene was covered by multiple *Canis familiaris*genomic contigs, the size of the intron covered by more than one contig was based on information of the respective chromosome. For the exons containing the start and the stop codon, the number of coding bp is listed as ORF (open reading frame); the location of the respective codon is listed between brackets.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThis study and A.C. Wiersma were supported by a grant from the Kennel Club Charitable Trust Canine Health Foundation Fund, United Kingdom, and by the Faculty of Veterinary Science, University of Liverpool, United Kingdom. Special thanks to Francine Jury of the Centre for Integrated Genomic Medical Research (CIGMR, University of Manchester, United Kingdom) for her help with the WAVE analyses. Thanks to Polona Stabej (Faculty of Veterinary Medicine, University of Utrecht, The Netherlands) for the *VCL*microsatellite data.\n"}
+{"text": "Background {#Sec1}\n==========\n\nYoung adults need good work ability to sustain them through a long working life as they replace the aging population \\[[@CR1]\\]. Opportunities for recovery at work may be very important to this group as work environments for young workers, especially young women, are reported to be declining \\[[@CR2]\\]. To our knowledge, however, associations between recovery opportunities at work and work ability have not previously been studied among young workers.\n\nThe definition of *work ability* can vary in different fields. In occupational health it is commonly defined as the balance between individual resources (e.g., health, knowledge, and attitudes) and working conditions (e.g., content, demands, and organization) \\[[@CR1]\\]. In the present study work ability was measured by the work ability score (WAS) assessed in the work ability index (WAI) \\[[@CR3]\\].\n\nSeveral definitions for the concept of *recovery* include situational characteristics that diminish load effects \\[[@CR4]\\], a need to recuperate from work-induced fatigue experienced primarily after a day of work \\[[@CR5]\\], and a desire to be temporary relieved from demands in order to restore one's resources \\[[@CR6]\\]. As the last two definitions are related more to personal needs for immediate recovery, than to recovery opportunities afforded by structural aspects of work, the first definition was used in the current study.\n\nFlexible working conditions have been found to offer recovery opportunities associated with health and well-being \\[[@CR7]\\]. Worktime control has also been shown to be a promising tool to maintain health, well-being, and job-related outcomes, including performance \\[[@CR8]\\]. The possibility of taking breaks at work and influencing other aspects of one's working hours has been shown to have positive effects on work-related fatigue, sleep, and health complaints, but not on future absenteeism \\[[@CR4]\\]. Rest breaks were shown in one review to have a positive effect on performance and productivity \\[[@CR9]\\], while another review found that flexible and compressed workweek schedules also offered recovery opportunities that had a positive effect on performance and productivity \\[[@CR10]\\]. Although several recovery opportunities at work have been studied and seem to have an effect on health, well-being, performance, and productivity, these have not been studied specifically in young workers.\n\nThough no studies of associations between recovery opportunities at work and work ability have been found for young workers, there are some closely related studies. In an interview study on work ability, recovery opportunities at work were included in a nuanced picture of experiences of work ability among young workers \\[[@CR11]\\]. Further, young women have been found to strive for balance between stress and recovery at work to maintain their health and work ability \\[[@CR12]\\]. One interesting hypothesis suggests that lack of recovery associated with prolonged work schedules and overtime can generate pressure on work ability among young workers \\[[@CR13]\\], but specific recovery opportunities at work are not discussed.\n\nBecause young workers (aged 16--29 years), especially young women, report more strenuous work and experience more fatigue after work than older workers \\[[@CR2]\\], recovery opportunities during the working day could be important in promoting excellent work ability. According to this report by the Swedish Work Environment Authority, young workers had less opportunity to decide when to do a work task and had more monotonous work than older workers; consequently, the work environment of young Swedish workers is poorer and includes fewer recovery opportunities that that of workers aged 30\u00a0years and older. Young workers with high work demands most likely have a greater need of recovery at work and hence a work environment with recovery opportunities than those with fewer demands \\[[@CR14]\\].\n\nThe concept of recovery is complex. It includes both recovery during the work day (internal) and recovery between work days (external) \\[[@CR14]\\]. While we recognize this complexity, the present study focused solely on recovery opportunities at work since this topic appears to be unexplored among young workers.\n\nThe aim of this study was to investigate the association between opportunities for recovery at work and excellent work ability among young workers, especially for young workers with high work demands. Specific research questions were: i) is the probability higher to report excellent work ability when a higher degree of recovery opportunities at work are present?, ii) does the association in i) for the total sample hold when adjusted for high work demands and educational level? and iii) is the association between recovery opportunities at work and excellent work ability stronger among those with high work demands?\n\nMethods {#Sec2}\n=======\n\nStudy design and data collection {#Sec3}\n--------------------------------\n\nThis is a cross-sectional population study using the Work Environment Survey by Statistics Sweden on behalf of the Swedish Work Environment Authority. This survey is based on the Labour Force Survey and includes additional questions asked in a telephone interview and a follow-up postal questionnaire. The purpose of the survey is to describe the physical and psychosocial work environment of the employed population. The employed population was defined as those aged 16--64 years who worked at least one hour during the measurement week. The Work Environment Survey has been conducted every second year since 1989. For each of the years 2009, 2011, and 2013 the randomized selected sample was about 10 000--16 000, of whom approximately 8100-- 12 400 answered the telephone interview and about 4800--7800 also answered the postal questionnaire. The telephone interview asks participants background questions on employment, work strain, leadership, and work ability. A week later the postal questionnaire asks about 121 specific occupational and health items. The development and validation of the method is described in a report from Statistics Sweden \\[[@CR15]\\].\n\nStudy sample {#Sec4}\n------------\n\nIn 2009--2013 about 19 000 individuals responded to the full telephone interview. Of these 4949 were young workers, 18--29 years of age. The study sample consists of 2351 young workers who responded to the telephone interview, the postal questionnaire, and the WAS question in the telephone interview (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}). There was no overlap of individuals in the three years chosen for the study. Inclusion criteria for the study sample were age range 18--29 years, and having answered the single WAS question. For the subsample of individuals with high work demands, *N*\u2009=\u2009788, the criterion was to have reported having either a job that was generally physically strenuous to a large extent *or* a high workload with far too much to do to a large extent, *or* both. The selection of this subsample was chosen as women and men with high work demands could have an increased need of recovery opportunities at work \\[[@CR14]\\].Fig. 1Selection of the study group. The study group included young workers who answered the telephone interview, the questionnaire, and also the question about work ability ranked according to Work Ability Score, WAS\n\nIndividual characteristics {#Sec5}\n--------------------------\n\nIndividual characteristics drawn from the survey were used to describe the sample: sex, age, education, years in present occupation, trade, work ability, and health. The variable educational level was dichotomized into compulsory school or high school versus post-gymnasium, college, or university. Three health variables were identified: pain in at least one location including upper back or neck, lower back, shoulder or arms, wrists or hands, and hips, legs, knees or feet; feeling very tired after work; and sleep difficulties due to thoughts of work. These health variables were ranked by frequency: every day, a few days per week, a day per week, a few days per month, seldom or not at all. The three first alternatives implied impaired health.\n\nOutcome {#Sec6}\n-------\n\n### Excellent work ability {#Sec7}\n\nWork ability can be measured by using part of the work ability index (WAI), a Finnish self-report instrument with seven dimensions measured on 10 items, by which individuals assess their own work ability \\[[@CR1]\\]. In the present survey work ability was measured by using the WAS \\[[@CR3]\\], taken from, and shown to correspond with, the WAI \\[[@CR16], [@CR17]\\]. The WAS measures \"current work ability compared with the lifetime best\" on a scale ranging from 0 \"cannot work at all right now\" to 10 \"my work ability is at its best right now\". Excellent work ability was ranked 10 on the scale, as in other studies \\[[@CR3], [@CR16]\\].\n\nExplanatory variables {#Sec8}\n---------------------\n\n### Recovery opportunities at work {#Sec9}\n\nBecause of the lack of knowledge of possible associations between recovery opportunities at work and excellent work ability, the five items measuring recovery opportunities at work in this study were chosen mainly from research about associations between recovery opportunities at work and health, well-being, performance, and productivity. Factors in accord with the Recovery Opportunity Scale \\[[@CR4]\\], which measures recovery opportunities in relation to health, were also used. Overall, the items in the questionnaire were used to assess opportunities for recovery at work, not lack of recovery at work.\n\nThe first question, on the possibility of deciding one's own working hours \\[[@CR4], [@CR10], [@CR18]\\], had three answer alternatives: \"Yes, I can be on flexitime\", \"Yes, I have in other ways relatively free working hours\", and \"No, I cannot influence my working hours\". The first two alternatives were considered recovery opportunities.\n\nThe second question asked about the possibility of deciding the work pace. Although this was not previously considered a recovery factor \\[[@CR19]\\], it was included in the hypothesis because it occurs in regulations on workers' opportunities to influence their own work \\[[@CR20]\\]. A third question asked about the possibility of taking short breaks at any time to talk, \\[[@CR4], [@CR9], [@CR18]\\]. These two questions had the same response alternatives:\" nearly all the time\" and \"about three quarters of the time\" represented recovery opportunities at work, in contrast to the alternatives of \"half the time\", \"about a quarter of the time\", \"a little\" and \"no, not at all\".\n\nThe fourth question, about the possibility of deciding when to perform a work task \\[[@CR8]\\], was answered by \"always\", \"mostly\", \"mostly not\", or \"never\", with the two first alternatives signifying recovery opportunities at work. The fifth and last question asked about variation in the respondents' work. On a scale of 0--5, responses ran from \"monotonous work\" to \"varied work\", and ratings of 4 and 5 indicated an opportunity for recovery at work. This question was chosen because the opportunity to take spontaneous breaks increases with varied work \\[[@CR14]\\].\n\nPossible confounders {#Sec10}\n--------------------\n\nHigh work demand was seen as a possible confounder for the analyses of the total sample to answer the first research question. This variable was, however, also used for creating a subsample, as described in the study sample section, to answer the third research question. For the second research question, two possible confounders were considered: high work demands and educational level. High work demands can influence work ability \\[[@CR21]\\] and have a negative association with recovery opportunities at work \\[[@CR14]\\] and was defined in terms of physically strenuous work to a large extent *and/or* high workload with far too much to do to a large extent. Poor basic education has been shown to relate to poor work ability among young adults \\[[@CR13]\\] and was therefore considered as a confounder.\n\nStatistical analysis {#Sec11}\n--------------------\n\nSAS version 7.11 (SAS Institute, Cary, NC, USA) was used for the analyses in this study. In all analyses the data were stratified for women and men \\[[@CR22]\\]. A subsample of those with high work demands was also analyzed. Descriptive data were obtained through frequency analyses.\n\nThe associations between recovery opportunities at work and excellent work ability was analyzed with logistic regressions (proc genmod in SAS). The effect measures retrieved from these regressions were prevalence ratios (PRs) instead of the commonly used odd ratios (ORs). This is due to the outcome being common and hence ORs cannot be interpreted as approximations of PRs. From the logistic regressions PRs were calculated as the effect of the exposure around the prevalence of 50\u00a0%. In the presented PRs the group with the lower degree of recovery opportunities at work was used as the reference category. The 95\u00a0% confidence intervals (CIs) for PRs were calculated using the delta method \\[[@CR23]\\].\n\nA guideline of applied logistic regression was followed \\[[@CR24]\\] to build models for testing associations between recovery opportunities at work and excellent work ability. Co-linearity was checked by examining pairwise correlations by cross tables considering\u2009\\>\u200980\u00a0% in a diagonal or any cells with no answers. This was performed between all exposure variables, between all exposure variables and the possible confounders of high work demands and educational level, and between the two confounders. *\"Being involved in planning the work\"* showed co-linearity with the variable *\"Possibility of deciding when to perform a work task*\". *\"Being involved in planning the work\"* was therefore excluded from further analyses due to more missing compared to *\"Possibility of deciding when to perform a work task*\". Univariate logistic regression analyses were then performed to select variables for the multiple regression with the criteria of *p*\u2009\\<\u20090.25.\n\nThe steps in the guideline for creating multiple models set criteria for variable inclusion at *p*\u2009\\<\u20090.25, changes in parameter estimates for exposure variables at\u2009\\<\u200915\u00a0%, and likelihood ratio tests between nested models. If an explanatory variable had *p*\u2009\u2265\u20090.25, but changed parameter estimates for other explanatory variables when the models were considered with and without it, it was included as a confounder. Possible interactions between exposure factors were also tested.\n\nEthics {#Sec12}\n------\n\nThis study was approved by the Regional Ethical Review Board at the University of Gothenburg, Sweden (Reg.no. 221--15).\n\nResults {#Sec13}\n=======\n\nCharacteristics of the study sample {#Sec14}\n-----------------------------------\n\nThe proportion of women was 55\u00a0% in the study sample and 56\u00a0% in the subsample (Table\u00a0[1](#Tab1){ref-type=\"table\"}). High work demands were reported by 34\u00a0% of the women (*N*\u2009=\u2009439) and 33\u00a0% of the men (*N*\u2009=\u2009349). Self-reported work ability was high for women and men (9.2) overall and in the group with high work demands (9.0--9.1). The young workers, especially women, generally had few years of work experience. The men seemed to have lower educational levels than the women, especially among those with high work demands.Table 1Characteristics of the study sample and the subsampleYoung workers without/with high work demandsWomenWomen with high work demandsMenMen with high work demandsN\u2009=\u20092351/N\u2009=\u2009788N\u2009=\u20091295 (55\u00a0%)N\u2009=\u2009439 (56\u00a0%)N\u2009=\u20091056 (45\u00a0%)N\u2009=\u2009349 (44\u00a0%)*Individual factors*Work ability: mean9.29.09.29.1\u2003range1--101--101--101--10\u2003SD1.31.41.21.4Age in years: mean24.324.324.824.7\u2003range18--2918--2918--2918--29\u2003SD3.33.23.23.1Years in present occupation: mean2.52.83.13.7\u2003range1/12--141/12--141/12--151/12--15\u2003SD2.32.52.52.8N%N%N%N%Excellent work ability81463252576736420458Educational level\u2003Compulsory school/high school70555261606926626175\u2003Post--gymnasium/college/university5744517340356348525Pain in at least one location of the body at least one day per week the last three months79162318754524419858Tired out after work at least one day per week the last three months69955324755154924571Sleep difficulties due to thoughts of the work keeping you awake at least one day per week the last three months2461911326155157020*Recovery opportunities*Possibility of deciding working hours\u2003Yes, flexible working hours or free working hours74258208486165916447\u2003No, cannot influence working hours54442229524334118453Possibility of deciding the work pace\u2003At least 3/4 of the time48438114264954714542\u2003At most half of the time80362322745575320458Possibility of taking short breaks\u2003At least 3/4 of the time49539106245775515545\u2003At most half of the time79061330764664518955Possibility of deciding when to perform a work task\u2003Mostly or always58145152356045817250\u2003Mostly not or never70155280654454217450Having work that is mostly\u2003Varied54843172404113912636\u2003Monotonous74057263606356122264(*N* number of workers, *SD* standard deviation)\n\nThe young women and men tended to work in different trades (not shown in the tables). The largest group of women, both in the whole sample and in the subsample with high work demands, was found in the service, health care, and retail sector. Men were most often employed in construction and manufacturing, and then in the service, health care, and retail sector, though men with high work demands worked mainly in construction and manufacturing.\n\nWomen with high work demands reported the poorest health (pain, fatigue, and sleep difficulties), although pain and fatigue were common in the whole sample (Table\u00a0[1](#Tab1){ref-type=\"table\"}).\n\nRecovery opportunities at work were reported at various levels, as about half of the study sample experienced some recovery at work. Those with high work demands, especially women, seemed to report fewer of several of the investigated recovery variables than the sample as whole. Possibilities of deciding one's own working hours and to have varied work appeared to be quite similar for women and men in general, but at least the first variable seemed to be less common among those with high work demands.\n\nAssociations between recovery opportunities at work and excellent work ability {#Sec15}\n------------------------------------------------------------------------------\n\nUnivariate associations between recovery opportunities at work and excellent work ability for both young women and men are presented in Table\u00a0[2](#Tab2){ref-type=\"table\"}. For men only, *possibility of deciding the work pace*, *possibility of taking short breaks, possibility of deciding when to perform a work task,* and *varied work* showed unadjusted associations with excellent work ability. For men with high work demands, all of these variables except for *possibility of taking short breaks* also showed unadjusted associations with excellent work ability.Table 2Univariate associations between recovery opportunities at work and excellent work abilityWomenWomen with high work demandsN\u2009=\u20091282--1289N\u2009=\u2009435--437Exposed (n)Cases (n)Univariate modelExposed (n)Cases (n)Univariate model*Recovery opportunities at work*PR(95\u00a0% CI)PR(95\u00a0% CI)Possibility of deciding working hours7424851.1(0.998--1.26)2081301.2(0.98--1.44)Possibility of deciding the work pace4843161.1(0.96--1.22)114731.2(0.94--1.48)Possibility of taking short breaks4953221.1(0.95--1.20)106631.1(0.82--1.29)Possibility of deciding when to perform a work task5813741.1(0.94--1.18)152921.1(0.88--1.33)Having mostly varied work5483561.1(0.95--1.20)1721011.0(0.83--1.24)MenMen with high work demandsN\u2009=\u20091043--1053N\u2009=\u2009344--349Exposed (n)Cases (n)Univariate modelExposed (n)Cases (n)Univariate model*Recovery opportunities at work*PR(95\u00a0% CI)PR(95\u00a0% CI)Possibility of deciding working hours6164011.1(0.93--1.21)164981.0(0.82--1.27)Possibility of deciding the work pace4953431.2(1.09--1.41)145961.3(1.03--1.62)Possibility of taking short breaks5773941.2(1.09--1.41)155991.3(0.98--1.53)Possibility of deciding when to perform a work task6044141.3(1.11--1.43)1721141.4(1.06--1.66)Having mostly varied work4112961.4(1.17--1.54)126861.4(1.06--1.71)(*N* number of workers included in the univariate analyses, *n* number of workers exposed for the variable respectively being a case, *PR* prevalence ratios, 95\u00a0% CI\u2009=\u200995\u00a0% confidence interval)\n\nThe variables included in the multiple models are shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}. For women, the exposure variable *possibility of deciding working hours* was included in the multiple model I with *p*\u2009=\u20090.08. In the multiple model II the same exposure variable was included; however, *p* fell to 0.1. *Possibility of deciding the work pace* was included in both these models only to adjust the models and thus should be viewed as a confounder. In the multiple model III for women with high work demands, both *possibility of deciding working hours* and *possibility of deciding the work pace* were included, however, with *p*\u2009=\u20090.1 and *p*\u2009=\u20090.2, respectively. In the multiple model IV for women only *possibility of deciding working hours* was included with *p\u2009=*\u20090.046.Table 3Multiple regression model. Parameter estimates (b coefficient) and p (*p*-values) for associations between recovery opportunities at work and excellent work ability for the study sample and the subsampleWomenWomen with high work demandsN\u2009=\u20091279N\u2009=\u2009434Multiple model I^a^Multiple model II^b^Multiple model III^c^Multiple model IV^d^Parameter estimatesPParameter estimatesPParameter estimatesPParameter estimatesP*Intercept*0.363\\<0.00010.504\\<0.00010.0820.556\u22120.0860.644*Recovery opportunities at work*Possibility of deciding working hours0.2120.0770.1800.1370.3200.1090.395**0.046** ^f^Possibility of deciding the work pace0.1330.283^e^0.0900.472^e^0.2910.207..Possibility of taking short breaks\\...\\.....Possibility of deciding when to perform a work task\\...\\.....Having mostly varied work\\...\\.....*Confounder*High work demands..\u22120.3060.013\\...**.**Educational level\\...\\...0.3360.094MenMen with high work demandsN\u2009=\u20091025N\u2009=\u2009345Multiple model I^a^Multiple model II^b^Multiple model III^c^Multiple model IV^d^Parameter estimatesPParameter estimatesPParameter estimatesPParameter estimatesP*Intercept*0.0430.7150.1620.223\u22120.1500.391\u22120.4160.139*Recovery opportunities at work*Possibility of deciding working hours\\...\\.....Possibility of deciding the work pace0.2400.1030.2360.1100.3510.1620.2560.317^e^Possibility of taking short breaks0.1970.1730.1660.256^e^\\....Possibility of deciding when to perform a work task0.2460.0970.2320.1180.4030.0990.491**0.049** ^f^Having mostly varied work0.488**0.0005** ^f^0.484**0.0006** ^f^0.526**0.027** ^f^0.579**0.019** ^f^*Confounder*High work demands..\u22120.2680.056\\....Educational level\\...\\...0.3340.213^a^Multiple model with no confounders. ^b^Multiple model with two confounders: high work demands and educational level. ^c^Multiple model for the subsample with no confounders. ^d^Multiple model for the subsample with one confounder: educational level. ^e^Variable included only to adjust the multiple model. ^f^The bold figures representing *p*-values\u2009\\<\u20090.05. (N\u2009=\u2009number of workers included in the multiple analyses, . = the variable was not included in the multiple model)\n\nModel I for the study sample of women included neither of the specified confounders (high work demands and educational level), but model II for the study sample of women included high work demands, Table\u00a0[3](#Tab3){ref-type=\"table\"}. Further, model III for the subsample of women did not include the educational confounder, but model IV for the subsample of women did.\n\nOther exposure variables were included in the multiple models for men. Having reported *varied work* was included in the two multiple models for all men, I and II, and also for men with high work demands, models III and IV, all with *p*\u2009\\<\u20090.05. *Possibility of deciding when to perform a work task* was included in both models I and III, with *p*\u2009=\u20090.097 and *p*\u2009=\u20090.099, respectively. In model II, this exposure variable was included with *p*\u2009=\u20090.1, and in model IV, *p* for this variable was 0.049. *Possibility of deciding the work pace* was included in models I, II, and III, with *p*\u2009=\u20090.1, *p*\u2009=\u20090.1, and *p*\u2009=\u20090.2, respectively. Finally, the variable *possibility of taking short breaks* was included, with *p*\u2009=\u20090.2 for all men in model I. In model II this exposure variable was only included to adjust the multiple model, as was *possibility of deciding the work pace* in model IV, and these should therefore be seen as confounders.\n\nSimilar to women, shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}, model I for the study sample of men included neither of the specified confounders (high work demands and educational level), and model II for the study sample of men included high work demands. Further, model III for the subsample of men did not include the educational confounder, but model IV for the subsample of men did, also in similarity with women.\n\nThe size of the effect of recovery opportunities at work on excellent work ability for young women is shown in Table\u00a0[4](#Tab4){ref-type=\"table\"} and for young men in Table\u00a0[5](#Tab5){ref-type=\"table\"}. The largest effects were found for men who reported *varied work* (PR\u2009=\u20091.3 in all models), for men with high work demands with *possibility of deciding when to perform a work task* (PR\u2009=\u20091.3, model IV), and for women with high work demands who reported *possibility of deciding working hours* (PR\u2009=\u20091.2, models III and IV).Table 4Prevalence ratios. Prevalence ratios based on multiple regression models for associations between recovery opportunities at work and excellent work ability for young womenWomenWomen with high work demandsN\u2009=\u20091279N\u2009=\u2009434Multiple model I^a^Multiple model II^b^Multiple model III^c^Multiple model IV^d^*Recovery opportunities at work*PR(95\u00a0% CI)PR(95\u00a0% CI)PR(95\u00a0% CI)PR(95\u00a0% CI)Possibility of deciding working hours1.1(0.98--1.24)1.1(0.96--1.22)1.2(0.94--1.41)1.2(0.98--1.46)Possibility of deciding the work pace1.1^e^(0.94--1.20)1.0^e^(0.92--1.17)1.2(0.89--1.42)..Possibility of taking short breaks\\...\\.....Possibility of deciding when to perform a work task\\...\\.....Having mostly varied work\\...\\.....^a^Multiple model with no confounders. ^b^Multiple model with two confounders: high work demands and educational level. ^c^Multiple model for the subsample with no confounders. ^d^Multiple model for the subsample with one confounder: educational level. ^e^Variable included only to adjust the multiple model(*N* number of workers included in the multiple analyses, *PR* prevalence ratios, 95\u00a0% CI\u2009=\u200995\u00a0% confidence interval, . = the variable was not included in the multiple model)Table 5Prevalence ratios. Prevalence ratios based on multiple regression models for associations between recovery opportunities at work and excellent work ability for young menMenMen with high work demandsN\u2009=\u20091025N\u2009=\u2009345Multiple model I^a^Multiple model II^b^Multiple model III^c^Multiple model IV^d^*Recovery opportunities at work*PR(95\u00a0% CI)PR(95\u00a0% CI)PR(95\u00a0% CI)PR(95\u00a0% CI)Possibility of deciding working hours\\...\\.....Possibility of deciding the work pace1.1(0.96--1.29)1.1(0.96--1.29)1.2(0.90--1.49)1.1^e^(0.85--1.42)Possibility of taking short breaks1.1(0.95--1.26)1.1^e^(0.93--1.24)\\....Possibility of deciding when to perform a work task1.1(0.97--1.30)1.1(0.96--1.29)1.2(0.93--1.52)1.3(0.96--1.60)Having mostly varied work1.3(1.10--1.46)1.3(1.10--1.45)1.3(0.99--1.61)1.3(1.01--1.67)^a^Multiple model with no confounders. ^b^Multiple model with two confounders: high work demands and educational level. ^c^Multiple model for the subsample with no confounders. ^d^Multiple model for the subsample with one confounder: educational level. ^e^Variable included only to adjust the multiple model. (*N* number of workers included in the multiple analyses, *PR* prevalence ratios, 95\u00a0% CI\u2009=\u200995\u00a0% confidence interval, . = the variable was not included in the multiple model)\n\nDiscussion {#Sec16}\n==========\n\nTo have varied work was found to be a recovery opportunity associated with excellent work ability for young men. In addition, for men with high work demands, the possibility of deciding when to perform a work task was also associated with excellent work ability. For young women with high work demands, the recovery opportunity to decide one's own working hours was associated with excellent work ability.\n\nAssociations between recovery opportunities at work and excellent work ability {#Sec17}\n------------------------------------------------------------------------------\n\nThe most distinct finding in the current study was that young men with varied work seemed most likely to experience excellent work ability. The recovery obtained from varied work among the men can probably include both the possibility to take breaks, which has been shown to maintain performance among adult workers \\[[@CR9]\\], and the possibility of deciding the work pace, both of which are in line with a theoretical framework of recovery in relation to variety in the job setting \\[[@CR14]\\]. A review of studies in work physiology has clearly shown that variation in physical workload has a significant influence on recovery in relation to musculoskeletal disorders \\[[@CR25]\\]. Creating varied work content could therefore contribute to promoting excellent work ability in young men, especially since younger workers tend to have more monotonous work than older workers \\[[@CR2]\\].\n\nThe association between varied work and excellent work ability was not, however, any stronger for young men with high work demands. This indicates that varied work is important for excellent work ability, independent of the level of work demands. Further, the different results for varied work in women and men might be explained by gender segregation of different work tasks in the labour market \\[[@CR26]\\], with different contributions to variation of work postures and work movements. Even in the same occupation it is known that work tasks often differ between women and men \\[[@CR27]\\], and men are therefore more likely able to create variation in their work.\n\nFor men with high work demands, the possibility of deciding when to perform a work tasks was also found to be associated with excellent work ability. This recovery opportunity is included in the concept \"global worktime control\", associated with job satisfaction, but not with performance or productivity \\[[@CR8]\\]. Hence, the result of work ability in the current study has not, to our knowledge, been presented earlier.\n\nEnabling women with high work demands to decide their own working hours could be a promising way to promote excellent work ability. This result is in line with earlier studies among adult workers. Having flextime has been shown to have an impact on job satisfaction \\[[@CR8]\\] and a positive effect on productivity \\[[@CR10]\\], although previously reported results were not specific for young women. Adult women with poor health and the opportunity to adjust their worktime were prospectively associated with increased work ability and return to work \\[[@CR28]\\]; however, a large part of that study's sample had poor work ability at baseline, which casts doubt on the appropriateness of a direct comparison with results of the current study. As no other recovery opportunities at work were found in the present study to be associated with excellent work ability among young women, other external factors such as duration and/or quality of sleep and/or relaxation between work-days may, in our opinion, be more important to their maintaining excellent work ability. Consequently, the situation could be more complex for young women, and possible important recovery opportunities were not taken into account in this study.\n\nMethodological considerations {#Sec18}\n-----------------------------\n\nObvious strengths of this study include its large population sample of a group not earlier investigated, but it also has limitations. The high proportion of young workers reporting excellent work ability and reporting high work ability in general might have muted the contrasts between exposed and unexposed groups and at least in part account for the small effects that were found. Furthermore, the limited scope of taking only internal recovery into account could make interpretations of the results uncertain, but probably contributes to a clearer focus.\n\nThe possible cohort effects in the current study may be an additional weakness as the sample was selected from different years. This sampling method of using surveys from three subsequent years was selected to obtain large groups for a study sample stratified for gender and further divided into subsamples with high work demands, while retaining enough power in the statistical analyses. Although the general work environment did actually change over the six years of the study \\[[@CR2]\\], there is no reason to assume that possible associations between recovery opportunities at work and excellent work ability should also have changed substantially.\n\nExcellent work ability can be measured in different ways. In the present study, as in earlier studies \\[[@CR3], [@CR16]\\], the cutoff for excellent work ability was 10 on the WAS. However, scores of 9--10 on the WAS have also been used for young workers \\[[@CR13], [@CR29]\\], in circumstances where choosing a score of 10 would have resulted in too few cases. Because self-rated work ability was high in the study sample, 10 on the WAS was considered appropriate for excellent work ability.\n\nThe questions in the telephone interview and the questionnaire were validated by Statistics Sweden, as occupational demands can be difficult to assess by self-reported exposure \\[[@CR30]\\]. The validation procedure for these questions has been carefully described \\[[@CR15]\\], and the questions used in the self-reported questionnaire were found to give reliable descriptions of actual work environments and conditions. Following that validation study, further work to increase the validity and reliability of the questions was also undertaken by Statistics Sweden; however, this work has not been carried out to a large extent since 2009, so the questions set for the three cohorts in the present study were nearly identical in formulation.\n\nThe choice of age group in the study sample warrants discussion. The upper age limit of 29 follows the limit for young workers used by Statistics Sweden, although their group starts at 16\u00a0years old. The lower limit of 18\u00a0years was chosen because most young adults aged 16--18 years continue in high school, and we wanted to examine young people at work.\n\nThe main limitation of the present study is the cross-sectional study design, which hinders interpretation of possible causal associations. Despite this limitation, however, the study has some obvious strengths. The design of a population-based register study is a broad attempt to capture the topic among a large group of young employees, and the well-described method selected for building the multiple models can be seen as an advantage. Furthermore, the research topic is, to our knowledge, poorly investigated in young workers, despite its possible importance to a sustainable working life for this young group.\n\nApplications {#Sec19}\n------------\n\nThe results of this study could be used when planning organizational actions to promote excellent work ability among young workers. Varied work might contribute to excellent work ability in men and plausibly a healthier workplace for all employees \\[[@CR31], [@CR32]\\]. Also, facilitating worktime control could be one way to promote excellent work ability, especially for women with high work demands, though the awareness of other opportunities for recovery at work and in leisure may be more important. Further studies, longitudinal or qualitative, are greatly needed to explore how recovery opportunities through the workplace can contribute to excellent work ability among young working women and men.\n\nConclusions {#Sec20}\n===========\n\nFor young men, having varied work can contribute to excellent work ability. In addition, for men with high work demands, the possibility of deciding when to perform a work task may be favourable for excellent work ability. For young women with high work demands, the possibility of deciding one's working hours can contribute to excellent work ability. Employers could use these opportunities for recovery in promoting work ability among young workers.\n\nWe wish to thank everyone involved in this project.\n\nFunding {#FPar1}\n=======\n\nFORTE; Swedish Research Council for Health, Working life and Welfare contributed with funding to this study. This organization has no role in the design of the study and collection, analysis, and interpretation of data and writing the manuscript.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nData and materials are stored at Occupational and Environmental Medicine, Gothenburg University. The data will not be shared due to other ongoing research projects using the material. In addition, all use of original data is legally limited by Statistics Sweden.\n\nAuthors' contributions {#FPar3}\n======================\n\nMB, JS, MH, and AE. designed the study. MB. carried out the data analysis and the interpretation of the data with help from AE. MB. wrote the manuscript, and JS, MH, and AE. discussed the manuscript with MB. and contributed to its final form. All researchers have read and approved the manuscript.\n\nCompeting interests {#FPar4}\n===================\n\nThe authors declare that they have no competing interests.\n\nConsent for publication {#FPar5}\n=======================\n\nNot applicable.\n\nEthics approval and consent to participate {#FPar6}\n==========================================\n\nAll procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or National Research Committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the cohort by Statistics Sweden when agreed to participate in the telephone interview and the questionnaire. This study was approved by the Regional Ethical Review Board at the University of Gothenburg, Sweden (Reg.no. 221--15).\n"}
+{"text": "Since its discovery in 1991, *Porcine circovirus type 2* (PCV2) has been one of the most present swine virus within the domestic population, reaching seroprevalence levels near to 100%[@b1][@b2]. Even though PCV2 is a small non-enveloped ssDNA virus with a circular 1766--1768\u2009bp long genome[@b3], the complexity of Porcine circovirus diseases (PCVD)[@b4] lies in every other trait, from the variety of subclinical and clinical syndromes[@b5], to the multifactorial nature of the disease onset[@b6], passing through the plethora of transmission routes which help the dissemination and the persistence of the pathogen[@b7]. Other fascinating aspects of this small virus are the remarkable genetic variability and peculiar evolutionary pathways that have showed to be particularly relevant also in relation to the changes in clinical manifestations occurring through time.\n\nAt least five Open Reading Frames (ORF) have been reported to be effectively transcribed, but the most studied and widely sequenced region is the ORF2, which encodes for the Cap protein[@b3][@b8][@b9][@b10]. This protein represents the only component of the viral capsid and has been proven to be the major target of the host immune response[@b11][@b12][@b13]. ORF2-based classification criteria[@b14][@b15] have been collectively adopted to define PCV2 genotypes because of its higher phylogenetic signal and lower tendency to recombine.\n\nPCV2 geographical distribution and its widespread presence have been undoubtedly linked to livestock movements and trade routes leading to the rapid spread of new strains in various countries[@b16].\n\nPCV2 genotype 2a is considered to have been the most prevalent until 2003, when a change in genotype prevalence (commonly known as genotype shift) occurred in favour of genotype 2b[@b2], with a parallel enhancement of the outbreak severity[@b17][@b18]. A similar situation have relapsed since the appearance of the genotype 2d in 2010[@b19][@b20], which is rapidly spreading with detriment to PCV2b prevalence[@b13]. The aforementioned high evolutionary rate (about 10^\u22123^--10^\u22124^ substitutions/site/year)[@b21] and the huge viral population size provide optimal conditions for natural selection to act. One of the suggested reasons for its wide acquired genetic variability could reside in selective pressures promoted also by vaccination-induced immunity escape[@b22].\n\nSince commercial vaccines against PCV2 became commonly available in 2006, their use has increased rapidly thanks to their high efficiency in controlling clinical signs and economical losses[@b23]. All currently available vaccines, independently from the production techniques (i.e. inactivation, chimeric PCV1-2 virus, Cap subunit expressed in baculovirus system), share the immunogenic capsid protein of the PCV2a genotype[@b24]. However, in the last decades some concerns have been raised about protection achieved against recently emerged genotypes (i.e. PCV2d). Some published works and many field reports claim episodes of vaccine failure in association with outbreaks of PCV2d[@b25][@b26]. Further experimental and field trials have disavowed this hypothesis, demonstrating the full protection induced by PCV2a based vaccine against all other genotypes[@b27][@b28]. Nevertheless, this cannot rule out a certain differential efficacy of current vaccines, which remains irrelevant when animals are vaccinated and raised in optimal conditions, but that can become evident in everyday veterinary practice (i.e. sub-optimal conditions). The role of vaccination in shaping viral evolution has been reported for different diseases affecting both animals and human beings. When immunity is not sterilizing, wild strains are able to circulate in a new \"challenging\" environment, made of less susceptible-immune hosts. Numerous examples are available of viruses that adapted to this new scenario by immuno-escape (*Hepatitis B virus*[@b29], *avian Metapneumovirus*[@b30]), increase in virulence (Marek's disease[@b31]), or both (*Infectious bursal disease virus*[@b31]). The aforementioned requirements for vaccine evolution and vaccine-induced pathogen replacement are present also for PCV2[@b32]. Unfortunately, this hypothesis is hard to be demonstrated due to widespread application of vaccines in domestic populations. Kekarainen *et al*.[@b33] provided a first evidence of the vaccination role in affecting the genetic variability of PCV2 strains collected from few vaccinating and non-vaccinating farms[@b33]. However, it is extremely challenging to recruit a significant number of farms where vaccination has not been applied for a long time despite PCV2 circulation. Additionally, contact network between farms is so dense that would be impossible to exclude the recent introduction of strains previously circulating in vaccinated farms[@b34].\n\nThe present study circumvents these limitation by investigating and comparing selective pressures acting on the capsid protein of PCV2 strains collected from farming animals before and after the vaccine introduction, and on PCV2 strains originating from populations of the same species (*Sus scrofa*) with different vaccination status: the vaccinated domestic pig population and the non-vaccinated wild boar one. Many epidemiological studies have been conducted on PCV2 presence among wild boars, revealing an overall and genotype-specific prevalence comparable to what is commonly found in domestic pigs[@b35][@b36][@b37][@b38]. Consequently, the wild swine population represents a major source of genetic variability and/or simply of viral exchange, but it is not concerned by the vaccination burden. Therefore the aim of this study was to bring together the available genomic information on wild and domestic pig PCV2 strains to explore their evolutionary pathways in populations with different immune status, to evaluate the potential contribution of genotype 2a-based vaccines in conditioning PCV2 epidemiology and evolution.\n\nResults\n=======\n\nSix databases were obtained by matching and disposing the sequences into six different selected categories, using an imposed time-related cut-off to divide samples collected from non-vaccinated animals (before 2005) and from a period when vaccination was widely adopted (after 2008). Hence we included 112 PCV2a sequences before vaccination (dataset 1), 145 PCV2a sequences after vaccination (dataset 2), 163 PCV2b sequences before vaccination (dataset 3), 84 PCV2b sequences after vaccination (dataset 4), 127 sequences covering the three main PCV2 genotypes (2a, 2b and 2d) from domestic pigs after 2008 (dataset 5), 127 sequences accounting for the same genotypes from wild boars after 2008 (dataset 6). Pairwise genetic distance was evaluated within all datasets and it demonstrated a substantially overlapping pattern between database pairs compared during the study, thus assuring that different dN-dS values were not biased by an unbalanced between-population diversity ([Supplementary Figure 1](#S1){ref-type=\"supplementary-material\"}).\n\nEstimated substitution rate was within the typical range of ssDNA viruses and comparable with what reported by other authors for PCV2[@b13][@b21] ([Fig. 1](#f1){ref-type=\"fig\"}). Despite different mean values reported for different dataset, the overlap of 95% Highest Posterior Density (95HPD) supports that the evolutionary rate was the same for all considered populations.\n\nThe three main analytic approaches allowed a systematic exploration of selective pressures in action on all the matched databases, with the exception of the domestic and wild population comparison, which did not undergo the directional selection study due to the presence of different \"only wild\" clades ([Supplementary Figure 2](#S1){ref-type=\"supplementary-material\"}). In fact, MEDS was designed with HIV-1 drug resistance in mind and, even if applicable wherever episodic directional selection occurs along multiple lineages, some conditions must be met. One of those is the presence of a rich collection of different background sequences well interspersed in the tree topology. If all the background sequences were so closely related that many foreground and background regions were separated by a single branch, it would be difficult to separate directional selection from founder effects, which would result in a loss of power.\n\nWithin the first group (databases 1 and 2), represented by the 2a genotype sequences before and after the vaccination introduction, many sites have been identified as undergoing diversifying selection by either MEME or at least by two other methods among SLAC, FEL and FUBAR ([Fig. 2a](#f2){ref-type=\"fig\"} and [Table 1](#t1){ref-type=\"table\"}): sites 59, 63, 136, 169 and 191 resulted statistically significant among the before-vaccination PCV2a sequences, while sites 47, 59 and 134 emerged from the after-vaccination PCV2a database. Other sites were proposed as targets of differential selection by site-by-site comparison within the same analyzed data: 10, 57, 169, 171, 176, 186 and 210 ([Fig. 2a](#f2){ref-type=\"fig\"} and [Table 2](#t2){ref-type=\"table\"}). The study of directional selection by MEDS highlighted the following sites toward the relative amino acids: 59-K, 206-I, 210-E, 232-K ([Table 3](#t3){ref-type=\"table\"}). The evaluation of the same residues from the Circovac^\u00ae^ vaccine strain picked up different amino acids at position 59 (R), 206 (K) and 210 (D), but not at 232 (K) residue ([Fig. 3](#f3){ref-type=\"fig\"}).\n\nIn the comparison of databases 3 and 4, two sites (59, 80) from database 3 and one from database 4 (228) appeared under episodic diversifying selection ([Fig. 2b](#f2){ref-type=\"fig\"} and [Table 1](#t1){ref-type=\"table\"}), whereas four sites (112, 119, 137, 191) showed to be under differential selective pressure ([Fig. 2b](#f2){ref-type=\"fig\"} and [Table 2](#t2){ref-type=\"table\"}). The investigation of directional pressures revealed two sites, 131 and 191, directed toward P and T respectively ([Table 3](#t3){ref-type=\"table\"}). Again, strains collected post-vaccination tend to differentiate from Circovac^\u00ae^ strain (131-T and 191-D) ([Fig. 3](#f3){ref-type=\"fig\"}).\n\nWild and domestic populations (database 5 and 6) turned out to share few sites (68, 88, 134, 169) under episodic diversifying selection. Two others (21, 39) were identified only for the wild boar group, while many more (59, 80, 86, 225, 228, 230) were recognised as significant for the domestic pig viral population ([Fig. 2c](#f2){ref-type=\"fig\"} and [Table 1](#t1){ref-type=\"table\"}). Differential selection appeared to act on sites 26, 29, 60, 130, 166 and 225 ([Fig. 2c](#f2){ref-type=\"fig\"} and [Table 2](#t2){ref-type=\"table\"}).\n\nDiscussion\n==========\n\nVaccines have been widely used for disease control in both human and animal medicine. Besides the obvious advantages, it is undeniable that vaccination strategies can alter the equilibrium between host and pathogen and modify the competitive hierarchy among viral strains[@b39]. Vaccine-induced immunity differs from the natural scenario in different ways. At first, it provides a pre-existing immunity that enhances the host resistance to pathogens, potentially increasing the fitness of more virulent strains[@b40]. Due to the limited number of available vaccine strains, the population immunity is expected to be more homogeneous than in natural conditions, potentially facilitating the emergence of vaccine-escape mutants, not differently from what happens in the case of drug-resistant mutants. This is particularly relevant for high-turnover animal populations, where vaccine-induced immunity plays a pivotal role due to the limited time for natural immune response to develop and act on viral evolution before the subject is removed from the population[@b31].\n\nEven if PCV2 is effectively controlled by vaccination, recent episodes of vaccine failure have been reported[@b25][@b26][@b41]. Although further studies have confirmed the efficacy of current vaccines, at least when applied optimally, some evidences suggesting a differential cross-protection and virulence among strains raised concerns about the threat of PCV2 evolution in response to widespread vaccination implementation[@b22][@b25][@b26][@b42]. If the vaccination has an effect on viral evolution and/or on the selection of different strains, it is expected that 1) it has an impact on viral population size, 2) it determines an increase in selective pressure, in particular on epitopic regions, 3) it promotes a tendency to evolve distancing from the vaccine sequence itself. While the change over time of different genotypes population dynamics has been reported by several authors[@b13][@b20], the other two points have never been examined in detail. A recent study (Reiner *et al*.)[@b43] reported the possible effect of vaccination on different genotype prevalence and on PCV2 sequence variation, involving especially an epitopic region located in the viral capsid, by comparing strains obtained from vaccinated and non-vaccinated herds in Germany. Even if an association between some SNP and vaccination status was proven, it is hard to determine if it is actually due to selection or if it is related to the association with particular genotype/clades, to the different genotype prevalence in herds with different status or to the strain temporal evolution. Additionally, the exclusion of correlated SNPs hampered the definition of the genomic positions eventually involved in the escape process[@b43]. Finally, a widespread between farms PCV2 circulation has been reported in countries with a high pig production[@b34] and consequently, even if the precise vaccination status was known in Reiner *et al*.[@b43], this paradoxically did not imply that the detected strains had not previously evolved in a vaccinated population.\n\nOn the other hand, the present study compares, using a bioinformatics and biostatistics approach, the strength of selective forces acting on viral population collected from different animal populations and time periods, briefly categorizable as \"vaccinated\" and \"non-vaccinated\". Differently from Reiner *et al*.[@b43], no epidemiological information were available about the vaccination status of animals included in the study. Nevertheless, it must be stressed that PCV2 vaccines are currently the most sold product in swine farming[@b24]. Since their introduction, almost if not all pigs are vaccinated directly or indirectly through maternal immunity[@b24]. Moreover, it has been estimated that 99% of all growing pigs in the USA are vaccinated against PCV2 at/or around the time of weaning[@b22]. Finally, considering the aforementioned dense PCV2 spreading network among farms and countries, it can be realistically assumed that since the introduction of vaccines, PCV2 strains have been under a vaccine-induced selective pressure.\n\nAt first, the effect and strength of evolutionary pressure was evaluated comparing PCV2a field strains collected before and after the introduction of commercial vaccines, also based on PCV2a genotype strains. The evidence that the evolutionary rate was the same for all considered populations suggests that an uneven variability between groups should be attributed to the differential action of selective forces.\n\nAccordingly, site by site comparison reported different sites, located (with the only exception of amino acid 10) in experimentally determined epitopic regions[@b44][@b45][@b46][@b47], where selective strength varied between pre and post-vaccination period, being the second one characterized by an increased rate of non-synonymous substitutions. Both populations revealed sites subjected to either pervasive and/or episodic diversifying selection, although the absolute number was lower in the post-vaccination period. Two phenomena can be advocated: sequences belonging to dataset 1 were collected from an expanding and potentially diversifying viral population, while those sampled in the post-vaccination period originated from an already shrinking population, as demonstrated by Shen *et al*.[@b22] and Franzo *et al*.[@b13][@b22]. Another non-conflicting hypothesis suggests that the presence of a population immunity ascribable to an extremely limited number of vaccine strains (and consequently epitopes) could have reduced the tendency to diversification, while prompting a directional change in the viral capsid away from the vaccine specific antigenic determinants. Accordingly to this hypothesis, three out of four amino acids detected under directional episodic selection (all located in epitopes) showed a tendency to mutate toward different amino acids from those of the Circovac^\u00ae^ vaccine, the first one to be licensed and the only whose capsid sequence was available. Most interestingly, changes in each of these three amino acids (59-206-210) have been experimentally demonstrated to impair the binding of monoclonal antibodies[@b44][@b48]. Additionally, in these amino acid positions PCV2a strains tend to mutate toward the amino acid profile of PCV2d (data not showed), for which episodes of vaccine failure have been reported[@b25][@b26]. Interestingly, when Shen *et al*.[@b22] evaluated the prevalence of PCV2a and PCV2b in samples originated from 61 USA sites (both performing or not PCV2 vaccination programs) 5 years after the vaccination introduction, a remarkable difference was observed compared with the pre-vaccination era[@b22]. Particularly, not only PCV2a was not detected in vaccinated farms but its prevalence in non-vaccinated ones was also lower, with respect to the one estimated before the vaccination introduction. Additionally, by partial sequencing of the capsid gene, some amino-acidic mutations considered potentially favourable for the evasion of the immunity induced by currently used vaccines were observed. It is consequently possible that the introduction of PCV2a- based vaccines dramatically changed the environment where PCV2 circulated, leading to a remarkable shift in the PCV2a population dynamics[@b13] and selective forces acting on it. Thus, this and other studies[@b22][@b43], provide preliminary but concordant evidences that some of the new PCV2a strains could have the potential to, at least partially, escape from the vaccine strain induced protection. Based on these results, it could be concluded that the three criteria for vaccine-driven selection have so been met for PCV2a.\n\nAs expected, PCV2b displayed a less clear, even if comparable, pattern that supports a less marked change in evolutionary dynamics after vaccination introduction. Interestingly, the herein reported mutation in position 131 was also described by Shen *et al*.[@b22] in PCV2b strains sampled after vaccination introduction. We decided to evaluate this phenomenon also on the whole PCV2 population by selecting strains of different genotypes collected from different animal populations belonging to the same specie but subjected to different \"vaccine pressure\" (i.e. wild boars and domestic pigs). The wild boar genotype distribution mirrored the one of the domestic pig, confirming the circulation of each genotype and the rise of PCV2d in the wild population too ([Supplementary Figure 2](#S1){ref-type=\"supplementary-material\"}). Also in this case different sites were proven to be under different selective pressure by site-by-site comparison. Nevertheless, just two out of six positions were under a stronger positive selection in domestic pigs, even if it is still suggestive that one of these two (amino acid 130) has been reported to be involved in immuno-escape by Saha *et al*.[@b48]. Additionally, more sites were estimated to be under diversifying selection in the domestic pig alignment than in the wild one, allowing the speculation of a stronger action of selective pressures in this population. Remarkably, when we evaluated the strength of selective pressures acting on strains collected from 1) wild boars before and after vaccination introduction and 2) wild boars and domestic pigs before vaccination introduction, no significant difference was demonstrated by site-by-site comparison (data not shown). Thus, the genuine effect of vaccine introduction was confirmed to predominate over time and/or host related factors in affecting viral evolution. Unfortunately, the limited amount of data, the unavoidable presence of some methodological assumptions in the used statistical methods, not always perfectly fit by our databases, prevented to conclusively support the role of vaccination in driving non-PCV2a strains evolution.\n\nNevertheless the results of the present study, combined with the rise of PCV2b and then PCV2d genotypes in a vaccinated population specularly to PCV2a decline, support that the wide application of vaccination could have determined a differential fitness among genotypes, affecting their epidemiology and evolution.\n\nConclusion\n==========\n\nThe present study represents the first in depth analysis and comparison of selective forces acting on PCV2 strains circulating in animal populations with different vaccination status. The results evidence a change between the evolutionary patterns and selective force strength acting on viral populations circulating in vaccinated and non-vaccinated animal populations.\n\nConsequently, it is possible to speculate that, as reported for other pathogens, vaccination strategies could be actually conditioning the evolution and epidemiology of PCV2. The information herein reported provides a strong substratum for further studies and prompts the implementation of complementary researches directed to the understanding of the mechanisms that drive this phenomenon at single animal/local scale level. Additional studies, based on extensive deep sequencing of viral sub-populations sampled over time from animals of known vaccination status, could provide a deeper insight into the basis of PCV2 evolution and confirm the role of vaccination in driving its evolution. Besides providing an interesting model that could be applied to other human and animal diseases, this stresses the compulsoriness of a continuous monitoring of viral epidemiology, particularly for rapidly evolving viruses like PCV2, and the necessity to share the related information to prevent or promptly act in response to the potential emergence of actual vaccine-immunity escape mutants.\n\nMaterials and Methods\n=====================\n\nDataset\n-------\n\nA collection of PCV2 ORF2 sequences for which sampling data and host species were available was downloaded from Genbank. The ORF2 gene was selected because it represents the major immunogen of PCV2 and the more variable region of its genome. Additionally, this segment is also the most commonly used for diagnostic and classification purpose and consequently it is the one for which the highest number of sequences, originating from several countries and different host species, is available. The presence of recombination was evaluated using RDP4[@b49] and recombinant sequences were excluded from further analysis. Briefly, scan settings were adjusted according to RDP manual and recombination events were accepted if detected by at least 3 method with a p-value \\>10^\u22125^ (with Bonferroni correction). Genotypization was then performed using a phylogenetic approach according to Franzo *et al*.[@b15] and different databases were created according to genotype, host species and collection data. Considering that vaccination against PCV2 was commonly introduced from 2006 and different vaccines were licensed all over the world by 2008[@b50], sequences were divided according to two conservative time points; sequences collected before 2005 were assumed to originate from non-vaccinated animals, while sequences collected after 2008 were considered to originate from vaccinated animals. Although there is no guarantee that all sequences collected after 2008 originated actually from vaccinated animals, the widespread use of vaccination let support the idea that, at least on a regional scale, viruses collected since 2008 have been evolving in a \"vaccinated environment\".\n\nPCV2 strains originating from domestic pigs were divided into: 1) PCV2a collected before 2005, 2) PCV2a collected after 2008, 3) PCV2b collected before 2005, 4) PCV2b collected after 2008. Due to the limited number of PCV2d sequences collected before 2005, PCV2d strains were not considered in the present analysis. Similarly, PCV2c was excluded from the study because of its negligible relevance from an epidemiological prospective and the minimum sequence availability. Finally, two other databases (5 and 6 respectively), one for domestic pigs and one for wild boars were set up including sequences of genotypes 2a, 2b and 2d collected after 2008. Unfortunately the limited sequence availability hampered the creation of a fully geographically matched datasets. Even if this was theoretically possible, many sequences collected from specific locations and time periods originated from a limited number of studies investigating substantially homogeneous viral populations. Such artificially unbalanced genetic diversity in different datasets would made difficult to differentiate the phenotypic diversity due to selective pressure action from the one simply caused by genomic variability (i.e. random mutations). For this reasons, efforts were made to stratify sequences in different datasets according to a combination of genotype, geographic origin and category while maintaining a comparable within-dataset genetic distance ([Supplementary Tables 1](#S1){ref-type=\"supplementary-material\"} and [2](#S1){ref-type=\"supplementary-material\"} and [Supplementary Figure 1](#S1){ref-type=\"supplementary-material\"}).\n\nAll sequences were aligned at the amino acid level and then the nucleotide sequences were back-translated using the MAFFT algorithm[@b51] implemented in TranslatorX[@b52].\n\nEvolutionary rate and analysis of selective pressures\n-----------------------------------------------------\n\nEvolutionary rates were estimated independently for the each dataset using the Bayesian serial-coalescent based method implemented in BEAST 1.8.2[@b53]. Substitution models were selected based on the Bayesian information criteria (BIC) calculated using Jmodeltest2.1.2[@b54] while the best molecular clock was selected based on the Bayesian factor (BF) scores, calculated by estimating the marginal likelihood of different models using both Path Sampling (PS) and Stepping Stone (SS) methods as proposed by Baele *et al*.[@b55].\n\nThe selective pressure on the viral proteins was evaluated separately for each database (1--6) using different methods based on the estimation of the difference between non-synonymous and synonymous substitution rates (dN-dS). Pervasive and episodic diversifying/purifying selection was estimated using SLAC, FEL, FUBAR and MEME[@b56][@b57][@b58]. The significance value was set to p\u2009\\<\u20090.05 for the FEL and MEME methods and to p\u2009\\<\u20090.1 for SLAC, which claims to be more conservative[@b59]. The results of FUBAR, were accepted when the posterior probability was greater than 0.9. Sites were assumed to be under diversifying selection when detected by MEME, able to model also episodic diversifying selection, or by at least two other methods. Differences in the site-by-site selection patterns among different databases were investigated for each dataset pair using the batch files *CompareSelectivePressure*.*bf* implemented in HyPhy[@b60]. To this purposes databases 1--2, 3--4, 5--6 were compared. The presence of episodic directional selection (i.e. when substitutions toward a small number of target amino acids are selected for in at least some part of the tree) was also investigated using MEDS[@b61], merging databases 1--2 and 3--4 and setting sequences collected after 2008 as foreground branches. All trees required for the analysis were reconstructed using FastTree[@b62].\n\nAdditional Information\n======================\n\n**How to cite this article**: Franzo, G. *et al*. Porcine circovirus type 2 (PCV2) evolution before and after the vaccination introduction: A large scale epidemiological study. *Sci. Rep.* **6**, 39458; doi: 10.1038/srep39458 (2016).\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Figure 1\n\n###### Supplementary Figure 2\n\n###### Supplementary Table 1\n\n###### Supplementary Table 2\n\n**Author Contributions** G.F. and C.M.T. planned the study. C.M.T. prepared the datasets. G.F. performed the analysis. G.F., C.M.T., wrote the manuscript. All authors have revised the manuscript and cooperate for its improvement. All authors read and approved the final manuscript.\n\n![Evolutionary rate (substitution/site/year) calculated for different datasets (i.e. PCV2a before and after vaccination introduction, PCV2b before and after vaccination introduction and PCV2 strains detected in domestic pigs and wild boars).\\\nMean substitution rate and 95HPD interval are reported for each group.](srep39458-f1){#f1}\n\n![Line graph reporting dN-dS values estimated for each codon position in database 1--2 (i.e. PCV2a before and after vaccination introduction) (**a**), 3--4 (i.e. PCV2b before and after vaccination introduction) (**b**) and 5--6 (PCV2 strains detected in domestic pigs and wild boars) (**c**). For clarity reasons only FUBAR results are reported. Sites detected to be under statistically significant positive selection by MEME or by at least two other methods are reported as colour coded circle (i.e. pre-vaccination groups in a and b and wild boars in c) or triangle (i.e. post-vaccination groups in a and b and domestic pigs in c), while sites detected under different diversifying selection through site-by-site comparison are reported as asterisks. Epitopes experimentally discovered by Trible *et al*.[@b46] (blue), Lekcharoensuk *et al*.[@b47] (red), Mahe *et al*.[@b44] (green), Ge *et al*.[@b63] (black) are symbolized by segments.](srep39458-f2){#f2}\n\n![Quaternary and tertiary structure of the PCV2 capsid protein reconstructed by homology modelling.\\\nThe surface of the capsid and of the single capsomer is represented as a mesh. Amino acid positions detected to be under a directional selection promoting a change in the viral capsid away from the vaccine specific antigenic determinants have been highlighted in red (PCV2a dataset) or blue (PCV2b dataset).](srep39458-f3){#f3}\n\n###### Table reporting sites detected under statistically significant (in bold) diversifying selection by different methods.\n\n  \u00a0                   Codon        SLAC         FEL     MEME        FUBAR                                                                                                                                                                       \n  ------------------ -------- --------------- -------- ------- --------------- --------------- -------- --------------- --------------- --------------- --------------- --------------- --------------- --------------- ------- --------------- ---------------\n  PCV2a                 47           \u00a0           \u00a0      5.041   **0**.**086**         \u00a0           \u00a0          3.998       **0**.**037**         \u00a0               \u00a0             \\>100       **0**.**035**         \u00a0           \u00a0         0.555       **0**.**964**\n  59                  0.595        0.552       2.896    0.339       0.023           0.990       2.995        0.300          46.621       **0**.**000**      37.679       **0**.**000**       0.198           0.622       0.608       0.858      \n  *63*                5.514    **0**.**062**   3.850    0.281       4.445       **0**.**031**   3.303        0.324           7.091           0.054           2.536           0.322           0.781       **0**.**978**   0.652   **0**.**926**  \n  72                    \u00a0            \u00a0         3.975    0.201         \u00a0               \u00a0         3.304        0.056             \u00a0               \u00a0             \\>100           0.077             \u00a0               \u00a0         0.435   **0**.**934**  \n  130                 1.211        0.460         \u00a0        \u00a0         2.390           0.262         \u00a0            \u00a0            16.670           0.086             \u00a0               \u00a0             0.528       **0**.**911**     \u00a0           \u00a0        \n  133                 3.312        0.146         \u00a0        \u00a0         2.052           0.261         \u00a0            \u00a0             5.327           0.216             \u00a0               \u00a0             0.494       **0**.**921**     \u00a0           \u00a0        \n  134                   \u00a0            \u00a0         2.550    0.385         \u00a0               \u00a0         4.026        0.237             \u00a0               \u00a0             \\>100       **0**.**000**         \u00a0               \u00a0         0.705   **0**.**924**  \n  136                 3.719    **0**.**051**     \u00a0        \u00a0         2.949       **0**.**009**     \u00a0            \u00a0             \\>100       **0**.**020**         \u00a0               \u00a0             0.492       **0**.**971**     \u00a0           \u00a0        \n  169                 0.839        0.515         \u00a0        \u00a0         0.479           0.804         \u00a0            \u00a0             \\>100       **0**.**000**         \u00a0               \u00a0             0.109           0.634         \u00a0           \u00a0        \n  191                 4.474    **0**.**057**   4.441    0.166       2.861       **0**.**020**   3.226        0.078           \\>100       **0**.**011**       \\>100           0.102           0.521       **0**.**971**   0.383   **0**.**918**  \n  206                 3.685        0.293       4.521    0.281       2.398           0.151       3.745        0.119           \\>100           0.190           \\>100           0.148           0.358       **0**.**905**   0.476   **0**.**931**  \n  PCV2b                 59        \u22125.116       0.954      \u00a0           \u00a0            \u221217.425      0.080          \u00a0               \u00a0            45.823       **0**.**028**         \u00a0               \u00a0               \u00a0           \u00a0           \u00a0               \u00a0\n  80                  1.957        0.444         \u00a0        \u00a0         7.287           0.169         \u00a0            \u00a0             \\>100       **0**.**004**         \u00a0               \u00a0               \u00a0               \u00a0           \u00a0           \u00a0        \n  228                   \u00a0            \u00a0         1.712    0.760         \u00a0               \u00a0         5.593        0.487             \u00a0               \u00a0             \\>100       **0**.**036**         \u00a0               \u00a0           \u00a0           \u00a0        \n  Domestic vs Wild      21         1.020       0.646      \u00a0           \u00a0             0.186       0.949          \u00a0               \u00a0             \\>100       **0**.**012**         \u00a0               \u00a0               \u00a0           \u00a0           \u00a0               \u00a0\n  39                  1.692        0.553         \u00a0        \u00a0         0.269           0.891         \u00a0            \u00a0             \\>100       **0**.**000**         \u00a0               \u00a0               \u00a0               \u00a0           \u00a0           \u00a0        \n  59                    \u00a0            \u00a0         0.545    0.388         \u00a0               \u00a0         1.360        0.625             \u00a0               \u00a0            66.010       **0**.**001**         \u00a0               \u00a0           \u00a0           \u00a0        \n  68                  \u22120.726       0.757       \u22121.377   0.983      \u22120.407           0.816       \u22123.661       0.055          99.854       **0**.**006**       \\>100       **0**.**002**         \u00a0               \u00a0           \u00a0           \u00a0        \n  80                    \u00a0            \u00a0         0.909    0.132         \u00a0               \u00a0         3.213    **0**.**032**         \u00a0               \u00a0             \\>100       **0**.**033**         \u00a0               \u00a0         0.616   **0**.**903**  \n  86                    \u00a0            \u00a0         0.653    0.318         \u00a0               \u00a0         2.257        0.244             \u00a0               \u00a0             \\>100       **0**.**042**         \u00a0               \u00a0           \u00a0           \u00a0        \n  88                  1.560        0.538       0.260    0.533       1.749           0.535       1.385        0.592           \\>100       **0**.**004**       \\>100       **0**.**001**         \u00a0               \u00a0           \u00a0           \u00a0        \n  134                 \u22121.547       0.772       0.228    0.566       4.650           0.220       1.421        0.336           \\>100       **0**.**000**       \\>100       **0**.**015**         \u00a0               \u00a0           \u00a0           \u00a0        \n  169                 0.496        0.610       0.641    0.292      \u22121.096           0.772       2.645        0.154          68.979       **0**.**001**       \\>100       **0**.**000**         \u00a0               \u00a0           \u00a0           \u00a0        \n  190                   \u00a0            \u00a0           \u00a0        \u00a0           \u00a0               \u00a0           \u00a0            \u00a0               \u00a0               \u00a0               \u00a0               \u00a0             2.018       **0**.**958**     \u00a0           \u00a0        \n  225                   \u00a0            \u00a0         0.294    0.680         \u00a0               \u00a0         0.835        0.326             \u00a0               \u00a0             \\>100       **0**.**022**         \u00a0               \u00a0           \u00a0           \u00a0        \n  228                   \u00a0            \u00a0         0.278    0.759         \u00a0               \u00a0         0.712        0.496             \u00a0               \u00a0             \\>100       **0**.**010**         \u00a0               \u00a0           \u00a0           \u00a0        \n  230                   \u00a0            \u00a0         0.272    0.519         \u00a0               \u00a0         1.564        0.251             \u00a0               \u00a0             \\>100       **0**.**019**         \u00a0               \u00a0           \u00a0           \u00a0        \n\nFor each method dN-dS values or \u03c9+ (dN/dS) and the respective p-values are reported.\n\n^a^Without v.\u2009=\u2009Value obtained for datasets including sequences of PCV2 strains collected from non-vaccinated populations.\n\n^b^With v.\u2009=\u2009Value obtained for datasets including sequences of PCV2 strains collected from vaccinated populations.\n\n###### Table reporting sites where dN/dS difference between the codon positions of the considered alignments was statistically significant (p-value \\<0.05).\n\n  \u00a0                       Codon       Site by site\n  ------------------ --------------- ---------------\n  PCV2a                    10         **0**.**026**\n  57                  **0**.**051**  \n  169                 **0**.**055**  \n  171                 **0**.**037**  \n  176                 **0**.**055**  \n  186                 **0**.**024**  \n  210                 **0**.**017**  \n  PCV2b                    112        **0**.**044**\n  119                 **0**.**044**  \n  137                 **0**.**054**  \n  191                 **0**.**040**  \n  Domestic vs Wild         26         **0**.**048**\n  29                  **0**.**053**  \n  60                  **0**.**049**  \n  130                 **0**.**005**  \n  166                 **0**.**042**  \n  225                 **0**.**055**  \n\n###### Table reporting sites under episodic directional selection, setting sequences collected from vaccinated animals as foreground branches.\n\n  \u00a0        Codon        MEDS        \n  ------- ------- ----------------- -----------------\n  PCV2a     59            K          **\\<0**.**001**\n  206        I     **\\<0**.**001**  \n  210        E     **\\<0**.**001**  \n  232        K     **\\<0**.**001**  \n  PCV2b     131           P          **\\<0**.**001**\n  228        T      **0**.**001**   \n\n^a^Amino acid toward directional selection points.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "I. Introduction\n===============\n\nMost of the medical records we have today are either in unstructured or semi-structured form. Documents from the domain of laboratory reports, for example, consist of attributes from a closed set of attribute types and their respective measurements, and they are already in the desired structured form. However, data in documents like doctor\\'s notes, discharge statements, etc. is mostly unstructured, and very difficult to analyse.\n\nHealthcare Data Extraction and Analysis (HEDEA) is a system to extract information from clinical reports, including prescriptions, blood and urine test reports, and medical notes. The system stores this information and retrieves it as a single-page anamnesis of the patient, with all past records displayed in a comprehensive tabular form. This can help a physician to thoroughly analyse the patient\\'s medical history without sifting through years of paperwork, some of which go missing.\n\nAnother area in which a structured information system can be useful is biomedical research. Epidemiological research is highly dependent on the statistical evaluation of medical data; hence, it requires data available in a structured form \\[[@B1]\\]. Obtaining such structured medical data through study-specific tests on a carefully selected subgroup of patients is both time and cost-intensive. Another way to obtain the necessary data is to manually examine archives filled with clinical reports and extract the required information, which is also an immensely time-consuming process. It would be desirable to perform analysis of those clinical records automatically or at least semi-automatically.\n\nOther advantages of such a system include easier knowledge transfer between doctors, accountability and compliance of hospital procedures, systematic tracking of patient\\'s health, and preservation of knowledge of current treatments and medical procedures for future reference.\n\n1. Related Work\n---------------\n\nHarkema et al. \\[[@B2]\\] designed a system to extract key information from clinical and biomedical text. Fette et al. \\[[@B3]\\] described techniques to process semi-structured clinical data and storage in a data warehouse. Atzmueller et al. \\[[@B4]\\] showed that a structured information system can help augment the data of an existing medical study or help create completely new complex medical hypotheses by studying patterns in what may seem to be independent factors. Black \\[[@B5]\\] explained how the data from such a system can be used to check a medical hypothesis. The system can also help determine groups of patients suitable for becoming part of a new study based on selection criteria, as shown by Kamal et al. \\[[@B6]\\]. Concerns related to doctor--patient confidentiality breaches were addressed to some extent by the systems designed by Aberdeen et al. \\[[@B7]\\] and Yang and Garibaldi \\[[@B8]\\]. Information extraction using various machine learning models has also been explored by Sondhi et al. \\[[@B9]\\], Uzuner et al. \\[[@B10]\\], Bae and Kim \\[[@B11]\\], Park et al. \\[[@B12]\\] and Glavas \\[[@B13]\\], and further analysed by Kraus et al. \\[[@B14]\\]. Chang et al. \\[[@B15]\\] proposed an approach for auto-assessment of health quality using classification and identification techniques.\n\nWhile similar systems do exist for specific applications in various parts of the world, there is a need for system tailored for use in Indian healthcare. Athavale and Zodpey \\[[@B16]\\] discussed the importance of informatics in the context of Indian healthcare and highlighted the lack of effective Electronic Health Records (EHR) management and analytics, while suggesting potential benefits of EHR in India. Another shortcoming of previously proposed systems is that they mostly cater to one specific requirement. Since most of the data is the same, multiple systems for different use cases and users (patients, physicians, and researchers), cause unnecessary duplication, making maintenance of databases difficult and cumbersome. There is a need to design a comprehensive system that caters to all kinds of users through different access modes, while maintaining data consistency and privacy. We tried to solve this problem through HEDEA, as proposed in this paper.\n\nII. Case Description\n====================\n\n1. HEDEA System Design\n----------------------\n\nThe complete system involves three primary units---an information extraction unit, an information storage unit, and an analysis unit. The information extraction unit extracts information from standard semi-structured text files using regular expressions, and passes it to the information storage unit, which stores the information under relevant attributes in an encrypted centralized database using the patient ID information as the primary key. This database is then queried by the analysis unit to generate results.\n\nFor identification of a patient across healthcare facilities, we use the national identification system in India called Aadhaar, a 12-digit number issued to citizens of the country. Using this identification number coupled with biometric authorisation, the patient can allow access to his/her medical history to the physician as and when needed, or it is possible for the patient to lock it to prevent any unauthorised access of information. If the information is locked by the patient, it cannot even be used for anonymous analytics.\n\n### 1) Functionality architecture\n\nThe functionality architecture of HEDEA is shown in [Figure 1](#F1){ref-type=\"fig\"}. The complete data is stored in a centralized database, accessible over the internet via appropriate user access keys.\n\nUsers upload medical reports to the system, where relevant information is extracted by the information extraction unit using regular expressions, tagged and stored by the information storage unit in the patient accounts, and based on access rights assigned by the patient, a comprehensive analytics database (CAD) for analysis and research without identifying information. The system uses regular expressions because most of the information to be extracted follows a precise format and resides in the vicinity of identifying information tags. Regular expressions make it easier to identify such information, and they vastly reduce the prospect of noise in the extracted information.\n\nFor any query related to a patient\\'s medical history, the patient\\'s Aadhaar number is fed to the system\\'s analysis & output (A&O) unit, authenticated by the patient using a password, and the patient\\'s complete (or partial, controlled by 'from' and 'to' dates) medical record is presented. For any query in the category of research & analysis, the query is fed into the A&O unit, and the results for the category are presented accordingly in raw format or in the form of charts or graphs, as needed. The data in this case is sourced from the CAD, which does not contain any identifying information.\n\n### 2) Information extraction using regular expressions\n\nThe information extraction unit was written in Python 3, and it uses regular expressions to extract information from text with a specified format. Data corresponding to attributes such as date of examination, weight, height, symptoms, and prescribed medicine are extracted from the file and stored along with the patient\\'s ID number in a file for each visit. Similarly, blood and urine test results with the date and patient ID are also extracted and stored in the database.\n\nThe information set used to design the information extraction unit was created by collecting a set of 90 healthcare reports from different patients with their consent. It contains a mix of doctors\\' notes, blood and urine test reports, and prescription files, from a number of healthcare institutes.\n\nThe input text files are scanned for matching patterns, and the relevant information is extracted. This is done with the help of regular expressions. For semi-structured data, the regular expressions directly extract the required data because the relevant data is expected to be labelled to a reference keyword, albeit in different formats. The data value should follow a specified format or type, or it should be in the vicinity of a reference keyword, as defined in the list of regular expressions.\n\nFor unstructured data, generally found in discharge notes and the like, another approach using regular expressions is used. A data value in the unstructured text should follow a specified format or type, or it should be a part of a list of keywords. As an example, for the reference keyword 'blood pressure', keywords such as 'high', 'low', or 'normal' are allowed, or numerical strings such as '110/90' or '120/80'. The data value should lie in the vicinity of the reference keyword. Each probable data value is assigned a score based on the distance from the keyword, which is the difference between the total number of words in the sentence with the keyword and the number of words occurring between the two. The score is reduced by a large factor if the keyword and data value occur in different sentences. The score is augmented slightly for numerical data values. For example, as shown in [Figure 2](#F2){ref-type=\"fig\"}, '120/80' will have a higher score than the word 'normal'. Thus, the attribute 'blood pressure' is assigned a value '120/80' in the database for the particular tuple.\n\nSome extracted information, such as date and ID information, are processed before storage to maintain consistency in data formats, thus making operations like searching and sorting of data easier.\n\n### 3) Analysis & output unit\n\n#### (1) Patient history\n\nThe output is obtained in an HTML page generated by the HEDEA A&O unit, which was developed using common gateway interface (CGI) scripting. The output, formatted in tables, has attributes as columns and corresponding records with dates as rows. Thus, any fluctuation in a patient\\'s vital signs and test results can be easily detected, and relevant medical action can be taken. The page shows results in three categories, namely, prescriptions, blood reports, and urine reports. An example output has been shown in [Figure 3](#F3){ref-type=\"fig\"}. This is now a completely structured transformation of the unstructured data available in the reports.\n\nPatient history can also be analysed using charts that can be generated on any numerical attribute. This has been demonstrated using the blood sugar levels of a patient over four weeks in [Figure 4](#F4){ref-type=\"fig\"}. HEDEA enables data aggregation and visualization even if the tests were performed at different pathological laboratories.\n\n#### (2) Analytics for medical research\n\nHEDEA can be used to generate analysis reports and charts using information in the central database without identifying information of the patient. This information is only provided if the patient allows the information to be used for medical research purposes by allowing the said information to be stored into the CAD. An example chart is shown in [Figure 5](#F5){ref-type=\"fig\"}. This chart uses the calculated body mass index (BMI) of all patients with relevant details in the CAD using the latest weight and height information, and presents the count of people falling in each segment.\n\nIII. Discussion\n===============\n\nIn this paper, we introduced an information extraction and presentation system that was designed to recognize and classify basic attributes present in medical records. We proposed a natural language processing model involving keyword-based and rule-based approaches to cope with the inherent complexity and structure of these records. A rich set of features are extracted using regular expression template patterns. At the retrieval step, only the necessary information is displayed for the relevant user, with all personal information removed from data for medical research, and only patient-authorised information available to the medical practitioner, with authentication based on Aadhaar ID.\n\nFor data extraction from unstructured text, an additional layer of model-based search using a convolutional neural network can be used to verify obtained results. This has been demonstrated by Li and Huang \\[[@B17]\\] for this exact use case.\n\nA mobile application can be designed for easier access to patients in comparison to the current web-based solution, which will help encourage adoption of the system.\n\n**Conflict of Interest:** No potential conflict of interest relevant to this article was reported.\n\n![System architecture diagram.](hir-24-148-g001){#F1}\n\n![Unstructured extraction using regular expressions and distance scoring.](hir-24-148-g002){#F2}\n\n![Sample output.](hir-24-148-g003){#F3}\n\n![Blood sugar levels of a patient.](hir-24-148-g004){#F4}\n\n![Body mass index (BMI) of patients in the comprehensive analytics database.](hir-24-148-g005){#F5}\n"}
+{"text": "Fire remains one of the largest reasons for non-natural fatalities and causes dramatic loss of property and gross national product in the range of billion euros every year. Cotton, a natural textile fiber, is extensively used in everyday products, such as clothing, furniture, and industrial products. Yet, cotton exhibits a low limiting oxygen index, ease of ignition, and high flammability[@b1]. Many attempts have been made to make fire-retardant cotton fabric[@b2][@b3]. Most effective present-day fire-retardant materials use heavy metals[@b4][@b5][@b6][@b7] or bromine[@b8][@b9]-containing species to efficiently arrest fires. In the light of environmental and health concerns, and upcoming legislation restrictions regarding brominated fire retardants[@b9][@b10][@b11], it is thus very important to develop new procedures for intrinsically fire retardant bulk materials, as well as, for fire-retardant and fire-blocking coatings, at best using environmentally benign materials and processes[@b12][@b13]. The combination of tailored polymers with inorganic nanoparticles, e.g. nanoclays[@b14][@b15][@b16][@b17][@b18][@b19][@b20][@b21][@b22][@b23] or carbon nanotubes[@b24][@b25][@b26][@b27], glass fabric[@b28], polyhedral silsesquioxanes (POSS)[@b29][@b30], layered double hydroxides[@b31][@b32][@b33][@b34], providing oxygen barrier properties and mechanical strength, preventing dripping of hot, molten plastics, and enabling an intumescent, self-expanding foam-like heat barrier behavior is a most promising approach for fire-retardant coating materials[@b35]. Aiming at a quantitative understanding in such materials requires approaches towards structurally well-defined surface coatings, in which benefits in the fire behavior can explicitly be linked to defined composition and architecture.\n\nInteresting progress towards nanostructured, fire-resistant coatings has been reported for ultrathin coatings based on step-wise, layer-by-layer (LbL) deposition of different polymers and nanoclay[@b22][@b36][@b37][@b38][@b39][@b40][@b41][@b42], in which substrates are alternatingly dipped into polymer and nanoclay to build up multilayers of organic and inorganic phases. Even though attractive fire barrier properties could be shown, the process requires several deposition steps needed to reach the desired properties (e.g., 20--40 deposition steps)[@b43][@b44][@b45][@b46][@b47][@b48]. Recently, a high efficiency LbL coating based on starch showed self-extinguishing behavior of cotton only after 4--8 deposition steps[@b48]. However, it is yet difficult to reach thicknesses beyond tenth or hundreds of nanometer via LbL, and the technique remains sequential, and laborious. Hence, there is still a tremendous need to develop simple, fast and scalable procedures to prepare thick coatings with tailored nanostructure based on readily available constituents and presenting excellent fire barrier properties.\n\nHere, we show a one-pot, single-step self-assembly approach to prepare thick, bioinspired, layered, hybrid brick-wall coatings with well-defined nanostructure on cotton textiles, as formed through self-assembly of tailored mixtures of sodium carboxymethyl cellulose (CMC) and montmorillonite (MTM) nanoclay at high fractions of the inorganic component. The coating thickness on the textile can be tailored using different CMC/MTM concentrations in the coating bath. We study the fire barrier and fire retardancy properties as a function of coating thickness by themogravimetric analysis (TGA), fire break-through test, vertical flame test (VFT), thermal imaging using forward looking infrared (FLIR) camera, and cone calorimetry. In addition, we modify the surface of the fire-retardant coating to impart amphiphobic and self-cleaning features. Hence, this study demonstrates a scalable process, that can be introduced into existing coating technologies to provide bioinspired, multifunctional, fire barrier nanocomposite coatings on textile, using benign water-based room temperature processing.\n\nResults\n=======\n\nSelf-assembled fire-barrier coatings on textile and structural characterization\n-------------------------------------------------------------------------------\n\nThe general strategy for the preparation of self-assembled, layered, bioinspired nanocomposite coatings on cotton textiles is shown in [Fig. 1a](#f1){ref-type=\"fig\"}. We start with the preparation of CMC/MTM dispersions by kneading a paste at CMC/MTM\u2009=\u200960/40 (w/w) at ca. 8\u2009wt% solid content in water. This allows the formation of homogeneous, concentrated slurries of both components due to well-balanced polymer/nanoclay interactions -- even at high solid content. Well-defined core/shell nanoplatelets are formed during this process by adsorption of CMC onto MTM nanoplatelets via hydrogen bonding and the hydrophobic effect[@b49]. During water evaporation, these slurries self-assemble into well-defined, hybrid, nanostructured brick-walls, with well-ordered alternating hard and soft layers as pre-encoded by the core/shell structure of the polymer-coated nanoclay[@b49][@b50][@b51]. For applying the coating, the cotton textiles are passed through a bath containing different concentrations of the CMC/MTM (60/40 w/w) hydrocolloid dispersion (1, 1.75, 2.5, 5\u2009wt%, after dilution from the initial paste) at 10\u2009mm/min, and subsequent drying. We denote the specimens as CMC~60~MTM~40~-x%, where the subscripts denote the weight fractions of both components and 'x%' denotes the concentration of the coating dispersion in wt%. Reasonably, we can expect that the coating thickness (weight/area) depends on the concentration of the coating dispersions and that a morphological transition may occur starting from coating of individual fibers at low concentration/viscosity to the presence of a fully coated film at high concentration/viscosity. For comparison, we also prepare a nacre-mimetic nanocomposite film by solution casting of a CMC/MTM dispersion in a petri dish. This allows characterizing the nanocomposite structure in a straightforward manner and serves as a comparison to structures created on textiles. The structural characterizations of the nanocomposite film, as well as, CMC~60~MTM~40~-1%, and CMC~60~MTM~40~-5% coated fabrics are displayed in [Fig. 1b--f](#f1){ref-type=\"fig\"}. X-ray diffraction (XRD) of the nanocomposite film shows a primary diffraction peak (**q**\\*) in the range of 1.7--2.3\u2009nm^\u22121^, corresponding to gallery spacing in the range of 3.7--2.7\u2009nm. Although the diffractogram of the coated fabric does not present very sharp peaks due to the undulated surface of the textile, it still clearly shows a hump (arrow in [Fig. 1b](#f1){ref-type=\"fig\"}) in the same range, thus confirming the successful formation of nacre-mimetic structures on the fabric. Weak and broad humps in both diffractograms appear around 3.5--5.3\u2009nm^\u22121^, which is the 2**q**\\* peak corresponding to the higher-order peaks characteristic of a lamellar morphology. The dominant peak of pure MTM is strongly diminished, indicating the achievement of high levels of delamination in the kneading process.\n\nFurther investigation of the morphology by SEM confirms the desired well-aligned layered structure in the films and coatings. Notably, the high magnification SEM image of CMC~60~MTM~40~-5% ([Fig. 1f](#f1){ref-type=\"fig\"}, inset) clearly shows that in addition to coating of individual fibers, a continuous film with 8.5\u2009\u00b1\u20093\u2009\u03bcm thickness remains on the surface of the whole fabric when coating using high concentrations ([Fig. 1e](#f1){ref-type=\"fig\"}).\n\nSuch a film is absent for lower slurry concentrations in the coating bath (e.g. 1\u2009wt%), where only a conformal coating of the fibers exists ([Fig. 1d](#f1){ref-type=\"fig\"}, inset and [Supplementary Fig. S1](#S1){ref-type=\"supplementary-material\"}). Formation of such a film is related to the high viscosity of the slurry and to the coating procedure. It is an important difference to polymer/nanoclay multilayers formed by LbL, which can only form conformal coatings on fibers[@b38][@b52][@b53].\n\nThe thermal and thermo-oxidative stability of the coated fabrics and control are examined by thermogravimetric analysis (TGA) in both N~2~ and air atmosphere from 25 to 850\u2009\u00b0C at 10\u2009\u00b0C/min as shown in [Fig. 2](#f2){ref-type=\"fig\"}.\n\nThe coated fabrics exhibit similar decomposition curves like the control, but they undergo a slightly earlier onset of degradation ([Fig. 2a,b](#f2){ref-type=\"fig\"}). Differential thermogravimetric analysis (DTGA) plots in [Fig. 2b](#f2){ref-type=\"fig\"} show that pure cotton (control) begins to degrade around 280 and 275\u2009\u00b0C in N~2~ and air, respectively. The initial minute weight loss at ca. 110\u2009\u00b0C is due to the removal of moisture. The major weight loss occurs between 280 to 400\u2009\u00b0C in both atmospheres resulting from the decomposition and dehydration of the cellulose leading to thermally stable residues. The second weight loss between 400 to 530\u2009\u00b0C in air corresponds to the further oxidation of the char to produce CO and CO~2~[@b54]. Despite having an earlier onset, CMC~60~MTM~40~-coated fabrics undergo slower degradation with increasing the slurry concentration as seen from the smaller and wider peak in the DTGA plots when going from CMC~60~MTM~40~-1% to CMC~60~MTM~40~-5% ([Fig. 2b](#f2){ref-type=\"fig\"})[@b55]. This slower rate of thermal degradation arises from the highly ordered, layered arrangement of the large MTM nanoplatelets (aspect ratio\u2009=\u2009260[@b56]) inside the CMC/MTM coating. On one hand, this layered nanocomposite structure provides thermal insulation and reduces the heat transfer into the cotton fabric, and on the other hand, it lowers oxygen diffusion due to the tortuous pathways (oxygen barrier property\u2009=\u20090.022\u2009cm^3^\u2009\u00b7\u2009mm\u00b7m^\u22122^\u2009\u00b7\u2009day^\u22121^\u00b7atm^\u22121^ at 50%RH and 0.115\u2009cm^3^\u2009\u00b7\u2009mm\u00b7m^\u22122^\u00b7day^\u22121^\u00b7atm^\u22121^ at 80%RH)[@b54][@b57]. The residual mass of the coated fabrics at 500\u2009\u00b0C (after the first step of degradation), and at 800\u2009\u00b0C (almost at the final stages of degradation) in N~2~ increases with increasing slurry concentrations ([Fig. 2c](#f2){ref-type=\"fig\"}). This indicates a higher inorganic content originating from thicker coatings. The average weight per area added to the fabric ranges between 0.6 and 3.8\u2009mg/cm^2^ with increasing the concentration from 1 to 5\u2009wt% in the coating bath. The values reflect the tendency found in SEM, where thinner conformal coatings on individual fiber are observed at lower slurry concentrations, while a continuous thicker film is formed in addition to coating of individual fibers at higher concentrations ([Fig. 1d--f](#f1){ref-type=\"fig\"}). Moreover, such mechanically robust CMC/MTM coating on cotton also provides better mechanical integrity to coated fabrics as compared to the uncoated ones[@b49][@b50]. We observed a stiffening of the coated cotton fabrics with increasing the coating concentration from 1 to 5\u2009wt% due to higher inorganic content and coating thickness.\n\nFire-barrier and fire-retardancy test\n-------------------------------------\n\nNext, we study the ability of the textiles to withstand direct high intensity flames from a short distance (130\u2009mm) using a fire break-through test. All samples were fixed in a circular holder with ca. 7\u2009cm diameter and then exposed to a high temperature torch at nearly 90\u00b0 angle. [Figure 3](#f3){ref-type=\"fig\"} displays time-lapse photographs taken during the test. The control catches fire within 3\u2009s, and is consumed completely after 18\u2009s without leaving any residue in the holder. In contrast, all coated fabrics char into a solid barrier material resisting long term exposure to direct flames. Most coated fabrics undergo shrinkage during charring in the direct flame, and get loose from the sample holder (see inset of [Fig. 3b--d](#f3){ref-type=\"fig\"}). We believe that contraction and distortion should be less relevant for larger specimens.\n\nAn increase of the coating thickness limits this behaviour, and CMC~60~MTM~40~-5% chars into a shape-persistent, solid barrier, which completely restricts the fire from breaking through ([Fig. 3e](#f3){ref-type=\"fig\"}). [Figure 3](#f3){ref-type=\"fig\"} further shows that the time to catch fire gradually increases with increasing slurry concentration, and the best fire barrier is observed for CMC~60~MTM~40~-5%, which chars but does not support any flame even after direct exposure for 60\u2009s.\n\nTo study flame retardancy and self-extinguishing behavior, we performed bench-scale vertical flame test (VFT) on specimens with dimensions of 160\u2009mm\u2009\u00d7\u200912.5\u2009mm (L\u2009\u00d7\u2009W) in analogy to EN ISO 11925-2. The flame was kept at the bottom of the sample and ignited for 5\u2009s. A forward looking infrared camera (FLIR) allows monitoring the spatiotemporal temperature profiles and standard video monitors the general behavior. [Figure 4](#f4){ref-type=\"fig\"} shows the results of bench-scale vertical flame test. The optical photographs and FLIR images after 5\u2009s of ignition, and optical photographs at the end of burning are shown in [Fig. 4a--c](#f4){ref-type=\"fig\"}. The photographs after 5\u2009s of ignition show that the flame is much brighter and stronger in case of the control, but becomes less vigorous for fabrics coated in higher slurry concentrations and thus having higher thicknesses. [Figure 4d](#f4){ref-type=\"fig\"} further depicts that the maximum burning height of the samples reached by either flame propagation or smoldering decreases with increasing coating thickness.\n\nThe control burns immediately and is consumed completely within 19\u2009s after ignition ([Fig. 5a](#f5){ref-type=\"fig\"}). CMC~60~MTM~40~-1% still allows flame propagation and complete burning, yet it shows a significant amount of coherent char with some slits at the end. The behavior of CMC~60~MTM~40~-1.75% and CMC~60~MTM~40~-2.5% is different. Those are not consumed completely, but burn up to half of the sample height. The flame reaches a maximum height of ca. 85 and 80\u2009mm in 8 and 7\u2009s after ignition for CMC~60~MTM~40~-1.75% and CMC~60~MTM~40~-2.5%, respectively, leaving behind a broken solid char due to the presence of higher coating thickness. Although they self-extinguish right after removal of the direct flame, a glowing frontier persists, which propagates very slowly within the burning height (ca. 50\u2009mm, [Fig. 5e](#f5){ref-type=\"fig\"}) and vanishes completely at 50 and 90\u2009s after ignition for CMC~60~MTM~40~-1.75%, and CMC~60~MTM~40~-2.5%, respectively, ([Fig. 5d](#f5){ref-type=\"fig\"}). CMC~60~MTM~40~-5% performs extremely well in terms of fire retardancy. It self-extinguishes once the direct flame is removed, and smolders up to a maximum height of 30\u2009mm in 9\u2009s after ignition. It does not support either flame or glow propagation, and forms a continuous and intact solid charred residue due to the highest coating thickness. Dripping is not observed in any sample.\n\nThe spatiotemporal FLIR imaging allows deeper insights into the thermal behavior and flame propagation. While stills from the FLIR videos after 5\u2009s of ignition are shown in [Fig. 4b](#f4){ref-type=\"fig\"}, the full range temperature vs. time profiles of the samples are shown in [Fig. 5](#f5){ref-type=\"fig\"}. First, we discuss the temperature profiles at different sample heights (from 10 to 150\u2009mm) of the control and the CMC~60~MTM~40~-coated fabrics from two intermediate slurry concentrations, serving as the most instructive examples (CMC~60~MTM~40~-1.75% and CMC~60~MTM~40~-2.5%; [Fig. 5a--c](#f5){ref-type=\"fig\"}). Afterwards, we focus on the comparison of the temperature profiles at two intermediate sample heights (50 and 100\u2009mm) for all samples ([Fig. 5d--f](#f5){ref-type=\"fig\"}).\n\nThe temperature profiles of the control in [Fig. 5a](#f5){ref-type=\"fig\"} reveal a consistent shift of the traces with increasing sample height. They exhibit temperatures around 400--500\u2009\u00b0C during passage of the flame front. The behavior is different for CMC~60~MTM~40~-1.75% and CMC~60~MTM~40~-2.5% samples. Here, the flame stops before reaching 100\u2009mm (max. flame height ca. 85\u2009mm), and therefore, the temperature profiles at 100, 120 and 150\u2009mm are essentially flat against time, except for the initial 5--15\u2009s, where some hot gases from the flame and the sample pass the markings. At 10\u2009mm, both samples show the initial burning process initiated by the exposure to the direct flame. At 50\u2009mm, interestingly, two maxima can be identified ([Fig. 5b--d](#f5){ref-type=\"fig\"}). The first maximum corresponds to the initial flame exposure and burning of the sample, and originates from the ascent of hot gases. While, the second maximum at comparatively lower temperature arises from a delayed glow frontier, which smolders slowly across the sample until the glow vanishes. The glow frontier stops just below the 50\u2009mm marking (lower temperature increase) in CMC~60~MTM~40~-1.75%, while it stops just on or above the marking in CMC~60~MTM~40~-2.5%, causing higher temperature rise in the second maxima as compared to CMC~60~MTM~40~-1.75%. The photographs in [Fig. 5e](#f5){ref-type=\"fig\"} serve to illustrate the two maxima at different times. At 100\u2009mm the two maxima are absent as the glowing frontiers vanish before reaching this height ([Fig. 5f](#f5){ref-type=\"fig\"}). The comparison of all coated samples also shows that all coated fabrics show a significant drop in temperature at a sample height 100\u2009mm or higher, except for CMC~60~MTM~40~-1%, which does not differ much from the control. Similar to the control, CMC~60~MTM~40~-1% is consumed completely (yet with higher char), due to the comparably thin coating.\n\nTo better predict the combustion behavior of the materials in real fire situation, we further estimate the heat release rate (HRR) and flammability of the coated fabrics and the control using cone calorimetry under a constant heat flux of 50\u2009kW/m^2^ ([Table 1](#t1){ref-type=\"table\"} and [Fig. 6a--f](#f6){ref-type=\"fig\"})[@b58]. Although the time to ignition (TTI) remains similar for all CMC~60~MTM~40~-coated fabrics and the control, increasing the coating thickness for the coated fabrics results in continuous reduction in peak heat release rate (pkHRR), total heat release (THR), and total smoke release (TSR). Furthermore, the heat profiles are more extended in the time axis for increasing coating thickness. This demonstrates that all CMC~60~MTM~40~-coated fabrics exhibit significantly higher fire-retardancy as compared to the control[@b55]. The CMC~60~MTM~40~-5% reduces the pkHRR and THR to 37% and, 44%, respectively, compared to the control. This improved fire-retardancy for CMC~60~MTM~40~-5% stems from the thicker MTM- containing nanocomposite coating, which also leads to largely increased protective char formation on the coated textile surface ([Fig. 6b--f](#f6){ref-type=\"fig\"}) in accordance with TGA[@b54]. This thick protective solid char acts as a thermal barrier and lowers the formation of flammable volatiles, therefore, leading to the reduction in pkHRR and THR[@b36][@b54][@b59]. Lower pkHRR is a desired parameter to exhibit efficient fire-retardancy in real fire situation as it retards or diminishes the self-propagation or spreading of the flame to other materials when the external flame or ignition source is removed[@b39]. Similarly, reduction in TSR is important as smoke can cause serious problems like choking, unconsciousness, lack of breathing during evacuation[@b39][@b54][@b60]. A 53% reduction in TSR for CMC~60~MTM~40~-5% as compared to the control is noteworthy, indicating that not much dense smoke is released during burning of the coated fabric. Therefore, CMC~60~MTM~40~-5% appears as a very effective fire retardant nanocomposite coating.\n\nSurface morphology characterization\n-----------------------------------\n\nTo evaluate the surface morphology before and after burning, we further characterized selected samples by SEM ([Fig. 7](#f7){ref-type=\"fig\"}). The SEM images show that CMC~60~MTM~40~ coatings on textiles retain a similar weave structure as in the control ([Fig. 7a,b](#f7){ref-type=\"fig\"}, and [Supplementary Fig. S1a](#S1){ref-type=\"supplementary-material\"}). Under high magnification, the individual fibers of the control show smooth and clean surfaces ([Fig. 7g](#f7){ref-type=\"fig\"}) but become rougher after applying the coating ([Fig. 7h](#f7){ref-type=\"fig\"}, and [Supplementary Fig. S1b](#S1){ref-type=\"supplementary-material\"}). Moreover, for CMC~60~MTM~40~-5%, the individual fibers are linked together, and the gap between them, which may serve as sites for the penetration of heat/flames, are diminished due to a thicker coating ([Fig. 7h](#f7){ref-type=\"fig\"})[@b36][@b52][@b53]. The SEM images after burning show that the thinner conformal coating of CMC~60~MTM~40~-1% is enough to preserve the weave structure, although considerable shrinkage is observed, creating gaps between yarns ([Supplementary Fig. S1c,d](#S1){ref-type=\"supplementary-material\"}). In contrast, due to thicker coating, CMC~60~MTM~40~-5% shows an intumescent effect by forming an expanded foam-like char with a lot of bubbles on its surface ([Fig. 7c,f](#f7){ref-type=\"fig\"}). These bubbles are formed due to evolution of gas while burning the intercalated polymer during exposure to flames. As a result, a porous foam-like structure develops that provides good heat insulation to the underlying materials[@b57][@b61]. A significant expansion of the coating from the initial 8.5\u2009\u00b1\u20093 to 25--50\u2009\u03bcm enables CMC~60~MTM~40~-5% to self-extinguish quickly once the direct flame is removed and to stop the flame propagation ([Fig. 7i](#f7){ref-type=\"fig\"}).\n\nFurthermore, dehydroxylation of the aluminosilicates results in self-cooling via water formation and also improves mechanical integrity due to char formation[@b49]. Interestingly, after burning, we also observe some hollow fibers with diameters between 7--15\u2009\u03bcm, and hence similar to the average diameters of the original fibers. This confirms that the CMC/MTM dispersion penetrates through the textile during the coating procedure and coats the individual fibers.\n\nAmphiphobic and self-cleaning features\n--------------------------------------\n\nGiven the diverse requirements of materials in general, especially in a real application context, it is important to develop multifunctional property profiles. One of the major bottlenecks in the application of water-borne polymer/clay nanocomposite coatings, no matter whether prepared by LbL strategy or by the faster self-assembly approach herein, is the inherent susceptibility to water[@b50][@b62]. Apart from fire and heat barrier properties, to promote functionality and applicability of the coated fabrics in harsh or humid environment, we sought to impart them with an amphiphobic coating, providing water (and oil) repellency and combining it with self-cleaning features. The latter is generally obtained by coatings of low surface energy and hierarchical roughness[@b30][@b63][@b64][@b65][@b66].\n\nWe choose a fabric coated from intermediate slurry concentration (CMC~60~MTM~40~-2.5%) and performed a straightforward surface modification with trichloro(*1H,1H,2H,2H*--perfluorooctyl)silane ([Fig. 8a](#f8){ref-type=\"fig\"}), which binds to the hydroxyl groups available at the CMC. We emphasize that other amphiphobic coatings without fluorine may also be deposited given the availability of hydroxyl groups[@b67][@b68][@b69][@b70][@b71]. Due to the high polarity of the CMC, the unmodified coated fabric rapidly absorbs water and bends under the weight of the absorbed water. Yet, once modified with the low surface energy fluorosilane, the fabric shows hydrophobic behavior with high repellency and high contact angles towards water ([Fig. 8b,d](#f8){ref-type=\"fig\"}, water contact angle ca. 100\u00b0, not shown here).\n\nAt the same time, the modification also provides protection against non-polar liquids such as salad oil, a clear advantage of using fluorinated coatings over only hydrophobic coatings ([Fig. 8c](#f8){ref-type=\"fig\"}). The snap shot series in [Fig. 8d](#f8){ref-type=\"fig\"} further shows that the water droplets bounce on the surface of the modified coated fabric, and easily roll down with a low sliding angle (\\<25\u00b0). This is due to the low adhesiveness resulting from the cooperation of the surface hierarchical structure of the fabric and the low surface energy of fluorosilane[@b30][@b64][@b72]. This is a beneficial feature for self-cleaning surfaces, and prompted us to also investigate the potential for self-cleaning. Using deposited activated carbon, we find that a series of water droplets easily removes the deposited dirt and recovers the pristine clean surface ([Fig. 8e](#f8){ref-type=\"fig\"}). To check repeatability, we perform four cycles (1 cycle\u2009=\u2009dirt deposition/cleaning/drying) of the self-cleaning test on the same coated fabric sample and obtain pristine clean surface each time without any macroscopic damage. Surface modification of the coated fabric can further be extended to more eco-friendly and biocompatible silicon based compounds in the future[@b67][@b73][@b74]. Overall this demonstrates that simple surface coatings can be applied to the nacre-mimetics CMC/MTM fire barrier coatings, and amphiphobic and self-cleaning features can be imparted as desired for application needs.\n\nDiscussion\n==========\n\nWe demonstrated a single step, large-scale, self-assembly approach to prepare thick, bioinspired, well-defined, hybrid brick-walls as fire-retardant coatings on textiles. The coating weight/thickness is tuned by changing the slurry concentrations, and the morphology can be changed from a conformal fiber coating (low concentration) to a micrometer-thick continuous coating (high concentration) on the textile. Hence, this process makes a stark difference to polymer/nanoclay multilayers formed by LbL, which could provide only conformal and ultrathin coatings (\\<1\u2009\u03bcm) on fibers in alternating deposition from dilute suspensions[@b38][@b43][@b44][@b45][@b52][@b53].\n\nThe fire barrier properties of the coated fabrics increase with increasing slurry concentration, making CMC~60~MTM~40~-5% the best fire barrier coated fabric. It not only withstands long-term high temperature direct flame exposure, but also prevents the penetration of flame through it by forming a shape persistent solid char during fire break-through test. It exhibits notably lower temperature build-up during VFT, and significantly decreased pkHRR (37%), THR (44%), and TSR (53%) compared to the control as seen from cone calorimetry, suggesting the generation of fewer flammable volatiles, and smoke, respectively. Furthermore, the SEM images of the post-burn chars of all coated fabrics in VFT showed that the fabric weave structure is well maintained even for CMC~60~MTM~40~-1%. CMC~60~MTM~40~-5% exhibits an intumescent effect and forms a thick and expanded foam-like char with lot of bubbles on its surface due to gas evolution. This nanoporous char provides good heat insulation to the underlying materials and improves fire retardancy.\n\nIn terms of functional benefits, we imparted amphiphobic and self-cleaning features by a simple surface modification with a fluorosilane. This furnishes a water- and oil-resistant surface based on water-borne CMC/MTM coatings, imparts self-cleaning, and thereby increases the usefulness for applications in harsher conditions.\n\nImportantly, the presented strategy opens a generic platform towards sustainable, high-performance and non-flammable thick nanocomposite coatings. As a major advantage, these fire-retardant coatings are devoid of any halogen atoms or heavy metals, often required in present day fire-barrier materials, and processed from water using sustainable building blocks. We expect that this continuous roll-to-roll coating process can be scaled to large quantities and implemented into existing coating technology. While we showcased the applicability on textiles, we believe it to be applicable on diverse substrates, and that further tuning of type of polymer system will allow an optimization of the properties and increased intumescent behavior. Further studies on these systems are currently under way.\n\nMethods\n=======\n\nMaterials\n---------\n\nSodium carboxymethyl cellulose (CMC, degree of substitution\u2009=\u2009carboxymethyl groups per anhydroglucose unit\u2009=\u20090.7, M~w~\u2009=\u200990\u2009kg/mol, Aldrich), Na-Cloisite (MTM, Rockwood), Trichloro(*1H,1H,2H,2H*--perfluorooctyl)silane (97%, Aldrich), MilliQ water, and pure cotton textile (ISO 105-F, 150\u2009mm width, 110\u2009g/m^2^ density) were used for all experiments.\n\nPreparation of bioinspired hybrid fire-retardant nanocomposite coating on textile\n---------------------------------------------------------------------------------\n\nCMC and MTM were added as premixed (60/40 w/w) powder into a IKA HKD-T-0.6 high-performance kneader with two wide-bladed kneading elements containing the appropriate amount of water to reach a total concentration of ca. 8\u2009wt%. After homogenization for 12\u2009h, the slurry was diluted to different concentrations (1, 1.75, 2.5 and 5\u2009wt%), and degassed. Then a custom-made continuous coating machine was used to guide the cotton textile (width 150\u2009mm, density 110\u2009g/m^2^) from a roll through the coating bath at 10\u2009mm/min for each slurry concentration. Afterwards, the fabrics were dried at ambient conditions. We use the following nomenclature for the coating compositions. Textile coated with CMC/MTM\u2009=\u200960/40 w/w at various concentrations are abbreviated as CMC~60~MTM~40~-x%, where x\u2009=\u20091, 1.75, 2.5 or 5\u2009wt%.\n\nSurface modification of bioinspired hybrid fire-retardant nanocomposite coating on textile\n------------------------------------------------------------------------------------------\n\nA CMC~60~MTM~40~-2.5% coated textile was dipped into a 5\u2009mM solution of trichloro(*1H,1H,2H,2H*--perfluorooctyl)silane in heptane, and left for 15\u2009min. Afterwards, the coated textile was rinsed with heptane and water few times and dried.\n\n**Field emission-scanning electron microscopy (SEM)** was performed on a Hitachi S-4800 field emission microscope using 1.5\u2009kV acceleration voltage.\n\n**Thermogravimetric analysis (TGA)** was done using a NETZSCH TG 209\u2009C instrument in both air and N~2~ atmosphere from 25 to 850\u2009\u00b0C at a heating rate of 10\u2009\u00b0C/min.\n\n**Wide-angle x-ray diffraction (WAXD)** was performed using an Empyrean setup from PANalytical using the Bragg Brentano parallel-beam geometry. An Empyrean Cu x-ray tube LFF HR (line source of 12\u2009\u00d7\u20090.04\u2009mm^2^) provided CuK\u03b1~\u03b1~ radiation with *\u03bb*\u2009=\u20091.542\u2009\u00c5 at 40\u2009kV voltage and 40\u2009mA current.\n\n**Fire break-through tests** were performed on the coated fabrics and the uncoated control using a Soudogaz^\u00ae^ X 2000 PZ gas torch with maximum flame temperature of ca. 1750\u2009\u00b0C. All samples were conditioned for 1--2 days at 23\u2009\u00b1\u20092\u2009\u00b0C and 50\u2009\u00b1\u20095% relative humidity (RH) prior to test. Then samples were cut into circular shape having ca. 7\u2009cm in diameter and fixed in a circular metal holder keeping a 130\u2009mm distance between the samples and the origin of the flame. The samples were exposed to the direct flame at an angle close to 90\u00b0 and the fire break-through process was monitored using a video camera.\n\n**Bench-scale vertical flame tests (VFT)** were performed on the coated fabrics and the uncoated control with dimensions of 160\u2009mm\u2009\u00d7\u200912.5\u2009mm (L\u2009\u00d7\u2009W) in analogy to EN ISO 11925-2. The samples were conditioned for 1--2 days in normal climate at 23\u2009\u00b1\u20092\u2009\u00b0C temperature and 50\u2009\u00b1\u20095% RH prior to test. A yellow flame was kept at the bottom of the sample for 5\u2009s. The burning was monitored using a forward looking infrared camera A655sc camera (FLIR) in the temperature range of 20--660\u2009\u00b0C, and standard video.\n\n**Cone calorimeter experiments** were conducted on a FTT cone calorimeter at 50\u2009kW/m^2^ heat flux, with an exhaust flow of 24\u2009L/s, according to ISO 5660, on sample dimensions of 100\u2009mm\u2009\u00d7\u2009100\u2009mm (L\u2009\u00d7\u2009W) conditioned for 24\u2009h at 23\u2009\u00b0C and 50% RH prior to test. The heat release rate was determined by the measurement of the oxygen consumption derived from the oxygen concentration and the flow rate in the combustion product stream. Parameters such as time to ignition (TTI), peak heat release rate (pkHRR), total heat release (THR), and total smoke release (TSR) were evaluated.\n\nAdditional Information\n======================\n\n**How to cite this article**: Das, P. *et al*. Large-scale, thick, self-assembled, nacre-mimetic brick-walls as fire barrier coatings on textiles. *Sci. Rep.* **7**, 39910; doi: 10.1038/srep39910 (2017).\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nThe authors acknowledge financial support from the IGF 17502N grant agreement. This work was performed in part at the Center for Chemical Polymer Technology CPT, which is supported by the EU and the federal state of North Rhine-Westphalia (grant no. EFRE 30 00 883 02). We thank Franz-Josef Steffens and Anja Pelzer for experimental help.\n\n**Author Contributions** P.D. carried out all the experiments. A.W. and M.M. conceived the project. A.W., P.D., and H.T. analyzed the data and wrote the manuscript. All authors discussed results and commented on the manuscript.\n\n![Preparation of bioinspired nanocomposite coatings and structural characterization.\\\n(**a**) Schematic of the self-assembled nanocomposite films based on CMC polymer, (R\u2009=\u2009H or CH~2~-COONa) and montmorillonite (MTM) natural nanoclay via concentration-induced self-assembly of polymer-coated nanoclay platelets with intrinsic hard/soft architecture, and coating of textiles from different concentrations of CMC/MTM\u2009=\u200960/40 w/w dispersions at 10\u2009mm/min. (**b**) XRD of the pure MTM film, CMC~60~MTM~40~ film, and CMC~60~MTM~40~-5% coated fabric. The arrow shows the appearance of a diffraction hump for the coated textile. (**c**--**f**) SEM images of the cross-sections of (**c**) CMC~60~MTM~40~ film (scale bar 5\u2009\u03bcm), (**d**) CMC~60~MTM~40~-1% (scale bar 200\u2009\u03bcm, inset: scale bar 10\u2009\u03bcm), (**e**) CMC~60~MTM~40~-5% (scale bar 200\u2009\u03bcm), and (**f**) high magnification image of the CMC~60~MTM~40~-5% (scale bar 20\u2009\u03bcm). CMC~60~MTM~40~-1% exhibits only a conformal coating of the fiber (**d**, inset, and [Supplementary Fig. S1](#S1){ref-type=\"supplementary-material\"}), and CMC~60~MTM~40~-5% shows the ordered layered structure of the nacre-mimetic coating (**f**, inset: scale bar 2\u2009\u03bcm).](srep39910-f1){#f1}\n\n![Thermal and thermo-oxidative stability of CMC~60~MTM~40~-coated fabrics and the control.\\\n(**a**) TGA plot of the control and CMC~60~MTM~40~-1%, 1.75%, 2.5%, and 5% coated fabrics in N~2~. (**b**) DTGA plots (top) in N~2~ and (bottom) in air. (**c**) The average coating weight added on the textile from different slurry concentrations (calculated gravimetrically), and the residual mass calculated from TGA at 500 and 800\u2009\u00b0C in N~2~ atmosphere for all samples.](srep39910-f2){#f2}\n\n![Fire break-through test.\\\nPhotographs of (**a**) the uncoated control, (**b**) CMC~60~MTM~40~-1%, (**c**) CMC~60~MTM~40~-1.75%, (**d**) CMC~60~MTM~40~-2.5% and (**e**) CMC~60~MTM~40~-5% at times indicated during exposure to a high temperature torch flame (ca. 1750\u2009\u00b0C) at nearly 90\u00b0 angle and 130\u2009mm away from the sample. Right column: Shrinkage of the coated fabrics during charring. Sample edges are marked in dotted white line.](srep39910-f3){#f3}\n\n![Bench-scale vertical flame test (VFT) analysis.\\\n(**a**) Photographs from VFT of CMC~60~MTM~40~-coated fabrics and the control after 5\u2009s of ignition. (**b**) Corresponding FLIR images after 5\u2009s of ignition with a color scale from 20 to 660\u2009\u00b0C. (**c**) Photographs of VFTs at the end of burning. (**d**) Maximum burning height and the minimum time needed to reach that height for CMC~60~MTM~40~-coated fabrics and the control.](srep39910-f4){#f4}\n\n![Temperature vs. time plots of the CMC~60~MTM~40~-coated fabrics and the control at different sample height during bench-scale VFT.\\\n(**a**--**c**) Temperature profiles of the control, CMC~60~MTM~40~-1.75%, and CMC~60~MTM~40~-2.5% at 10, 50, 100, 120 and 150\u2009mm sample height. (**d**,**f**) Comparative study of the temperature profiles of all CMC~60~MTM~40~-coated fabrics at (**d**) 50 and (**f**) 100\u2009mm. (**e**) FLIR images of CMC~60~MTM~40~-1.75% and CMC~60~MTM~40~-2.5% corresponding to the first temperature peak at ca. 5\u2009s (hot gas) and second temperature peak at 50\u2009s and 65\u2009s, respectively, (glowing frontier), (temperature scale bar same as in [Fig. 4b](#f4){ref-type=\"fig\"}).](srep39910-f5){#f5}\n\n![Cone calorimetry measurements of the CMC~60~MTM~40~-coated fabrics and the control.\\\n(**a**) Heat release rate of the CMC~60~MTM~40~-coated fabrics for all slurry concentrations and the control. (**b**--**f**) Photos of the residues of the coated fabric and the control after burning in cone calorimetry, (**b**) control, and (**c**--**f**) CMC~60~MTM~40~-1%, 1.75%, 2.5% and 5%, respectively.](srep39910-f6){#f6}\n\n![SEM images of the CMC~60~MTM~40~-5% coated fabric and the control.\\\n(**a**,**d**,**g**) Surface and cross-section of the control, and (**b**,**e**,**h**) those of CMC~60~MTM~40~-5%, before burning. (**c**,**f**,**i**) CMC~60~MTM~40~-5% after burning in VFT showing (**c**) the surface, and (**f**,**i**) the cross-section. (**i**) Individual hollow fibers are seen after burning of CMC~60~MTM~40~-5%, encircled in red. (The control was completely consumed during burning, so no image can be shown after burning) (**a**,**b**,**c**,**d**, scale bar 1\u2009mm; (**e**,**f**) scale bar 300\u2009\u03bcm; (**g**,**h**) scale bar 100\u2009\u03bcm; (**i**) scale bar 50\u2009\u03bcm).](srep39910-f7){#f7}\n\n![Surface modification and photographs showing amphiphobicity, water roll off and self-cleaning of the coated fabrics.\\\n(**a**) Schematic representation of surface modification of CMC~60~MTM~40~-coated fabrics with trichloro(*1H,1H,2H,2H*--perfluorooctyl)silane (not in scale). Photographs of (**b**) water droplets, and (**c**) salad oil droplets, on both unmodified and modified coated fabrics. (**d**) Water droplet is soaked on the surface of the unmodified coated fabric, whereas it can easily roll off and bounce off from the modified coated fabric. (**e**) Self-cleaning behavior of modified coated fabric.](srep39910-f8){#f8}\n\n###### Overview from cone calorimetry measurement.\n\n  Sample                    TTI (s)   pkHRR (kW/m^2^)   avg. HRR (kW/m^2^)   THR (MJ/m^2^)   TSR (m^2^/m^2^)\n  ------------------------ --------- ----------------- -------------------- --------------- -----------------\n  control                    8\u2009\u00b1\u20092        88\u2009\u00b1\u20091             28\u2009\u00b1\u20092.5           1.8\u2009\u00b1\u20091          4.5\u2009\u00b1\u20091\n  CMC~60~MTM~40~ \u2500 1%        8\u2009\u00b1\u20092        73\u2009\u00b1\u20093             28\u2009\u00b1\u20091.5          1.7\u2009\u00b1\u20090.1         3.6\u2009\u00b1\u20090\n  CMC~60~MTM~40~ \u2500 1.75%     9\u2009\u00b1\u20093        70\u2009\u00b1\u20092             25\u2009\u00b1\u20092.1          1.6\u2009\u00b1\u20090.1        3.8\u2009\u00b1\u20090.6\n  CMC~60~MTM~40~ \u2500 2.5%      5\u2009\u00b1\u20091        69\u2009\u00b1\u20093             27\u2009\u00b1\u20090.6          1.7\u2009\u00b1\u20090.1        1.0\u2009\u00b1\u20090.5\n  CMC~60~MTM~40~ \u2500 5%        7\u2009\u00b1\u20092        55\u2009\u00b1\u20097             17\u2009\u00b1\u20092.2           1\u2009\u00b1\u20090.1         2.1\u2009\u00b1\u20091.1\n\nTTI\u2009=\u2009time to ignition, pkHRR\u2009=\u2009peak heat release rate, avg. HRR\u2009=\u2009average heat release rate, THR\u2009=\u2009total heat release rate, TSR\u2009=\u2009total smoke release.\n"}
+{"text": "Introduction {#S0001}\n============\n\nThe pathogenesis of viral myocarditis (VMC) can be divided into three stages. In the first stage, viral infection directly injures the cardiomyocytes while inducing the innate immune response of the host to eliminate pathogens. In the second stage, myocardial necrosis from the first stage induces inflammatory cells to attack normal myocardium, augmenting the injury. In the third stage, there is a wide range of immune injury to cardiomyocytes and myocardial fibrosis, resulting in progression to dilated cardiomyopathy \\[[1](#CIT0001)\\]. Many studies have indicated that CD4^+^ T cells and their cytokines played a critical role in the last two stages \\[[2](#CIT0002)\\]. Recently, many researchers found that CD4^+^ T cell subsets, including Th1, Th2, Th17, and Treg, play an important role in viral myocarditis, and the functions of each subgroup varied in myocarditis. Huber and others discovered that male mice who mostly had a Th1 cell-mediated immune response were more susceptible to the CVB3 virus than female mice who had a Th2-mediated immune response \\[[3](#CIT0003)\\]. Rangachari et al. reported that Th17 cells and its cytokine, IL-17, increased the severity of viral myocarditis and autoimmune myocarditis \\[[4](#CIT0004)\\]. Treg cells could inhibit the expression of inflammatory cytokines and attenuate the severity of viral myocarditis \\[[5](#CIT0005)\\]. Therefore, it is important to regulate Th cell differentiation in viral myocarditis.\n\nTwo elements determines the differentiation of naive T cells, the specific transcription factors and cytokines in the local microenvironment. A variety of cytokines are activated in the pathogenesis of viral myocarditis. These cytokines are closely related to Th cell differentiation. Interestingly, these cytokines also have major overlap with the cytokines regulated by the cholinergic anti-inflammatory pathway (CAP). The CAP is a recently proposed immunoregulatory pathway that inhibits the release of inflammatory cytokines by connecting the nervous system to the immune system, ameliorating the inflammatory response of many diseases, such as sepsis, ulcerative colitis, and rheumatoid arthritis. Our previous studies have indicated that the CAP significantly decreases the level of Th17 cell-related IL-17A and IL-6 as well as Th1 cell-associated TNF-\u03b1 \\[[6](#CIT0006)--[10](#CIT0010)\\]. Galitovskiy et al. also found that \u03b17-nicotinic acetylcholine receptor (\u03b17-nAChR) agonist, nicotine, increased the ratio of Treg cells and reduced the ratio of Th17 cells, improving the prognosis of ulcerative colitis \\[[11](#CIT0011)\\]. These results suggest that CAP may regulate the differentiation of CD4^+^ T cells in VMC mice.\n\nAlthough the anti-inflammatory effect of CAP on VMC has been demonstrated in previous studies from our group and others \\[[6](#CIT0006)--[8](#CIT0008),[12](#CIT0012)\\], it has not been reported whether CAP can affect the differentiation of CD4^+^ T cell subsets. However, based on previous studies, we hypothesized that CAP could inhibit the inflammatory response by modulating the differentiation of Th cells subsets in VMC, and reducing myocardium lesions. Therefore, this study will focus on the regulation effect of CAP on Th cell subsets.\n\nResults {#S0002}\n=======\n\nEffects of nicotine or methyllycaconinitine on the regulation of spleen CD4^+^ T cell differentiation in vitro {#S0002-S2001}\n--------------------------------------------------------------------------------------------------------------\n\nT-bet, GATA3, ROR-\u03b3 and Foxp3 are the specific transcription factors of Th1, Th2, Th17 and Treg cells, respectively; therefore, we analyzed the distribution of Th1, Th2, Th17 and Treg cells in the spleens of control and VMC mice \\[[13](#CIT0013)\\]. Nicotine treatment could upregulate GATA3 and Foxp3 expression while downregulating T-bet and ROR-\u03b3 expression compared to the PBS group. (P\u00a0\\<\u00a00.05) However, the expression levels of GATA3 and Foxp3 were reduced, and the expression levels of T-bet and ROR-\u03b3 were elevated in the methyllycaconitine (MLA) group in the separate CD4^+^ T cells. (P\u00a0\\<\u00a00.05) ([Figure 1](#F0001))10.1080/21505594.2018.1482179-F0001Figure 1.The expression of Th cell-associated specific transcription factors in the spleen cells from CVB3-infected mice (n\u00a0=\u00a05 in each group). Nicotine upregulated GATA3 and Foxp3 expression, while it downregulated T-bet and ROR-\u03b3 expression. A. Representative levels of the specific transcription factors. B. Absolute intensity ratio of the specific transcription factors to Gapdh. \\*P\u00a0\\<\u00a00.05 versus the PBS group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nEffects of nicotine or methyllycaconinitine treatment on the percentage of CD4^+^ T cells in the spleen in vitro {#S0002-S2002}\n----------------------------------------------------------------------------------------------------------------\n\nThen, we used flow cytometry to detect the percentage of the Th subgroup in separated CD4^+^ T cells from CVB3-induced mice. The percentages of IFN-\u03b3^+^ CD4^+^ Th1 and IL-17^+^ CD4^+^ Th17 cells were increased in the MLA group, but they were decreased in the nicotine group. (P\u00a0\\<\u00a00.05) However, the percentages of IL-4^+^ CD4^+^ Th2 and CD4^+^ CD25^+^ Foxp3^+^ Treg cells were increased in the nicotine group and decreased in the MLA group. (P\u00a0\\<\u00a00.05) ([Figure 2](#F0002))10.1080/21505594.2018.1482179-F0002Figure 2.The percentage of Th cell subsets in the spleen from CVB3-infected mice (n\u00a0=\u00a05 in each group). The percentages of IFN-\u03b3^+^ CD4^+^ Th1 and IL-17^+^ CD4^+^ Th17 cells were increased in the MLA group, but they were decreased in the nicotine group. CD4, IFN-\u03b3, IL-4, IL-17, CD25 and Foxp3 were labeled by FITC, PE-CY5.5, APC, PE, APC, and PE, respectively. A, B, C and D are the representative results for each group and the percentage of double positive cells. \\*P\u00a0\\<\u00a00.05 versus the PBS group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nSurvival rate {#S0002-S2003}\n-------------\n\nBALB/C mice were injected with 10^5^ TCID~50~ of CVB3 virus by intraperitoneal injection to generate a viral myocarditis model. The mice injected with CVB3 had weakness, weight loss, back arching and irritability. There was no death in the control group after 14\u00a0days; 45% (18 of 40) mice survived in the VMC group. The survival rate of the nicotine group was 60% (24 of 40), and it was 30% (12 of 40) for the MLA group. The survival rate was significantly decreased in the VMC and MLA groups compared with the control group (each P\u00a0\\<\u00a00.05 versus the control group), but no significant change in the survival rate was seen in the nicotine group compared with the control group (P\u00a0=\u00a00.062 versus the control group). Compared with the VMC and MLA groups, the survival rate was slightly increased by nicotine, but this effect did not reach statistical significance (P\u00a0=\u00a00.282 versus the VMC group; P\u00a0=\u00a00.054 versus the MLA group). ([Figure 3](#F0003))10.1080/21505594.2018.1482179-F0003Figure 3.Survival rate (n\u00a0=\u00a040 in each group) and Viral replication in the myocardium (n\u00a0=\u00a05 in each group). A. The survival curve of each group. B. Viral replication in the myocardium of each group.\n\nMyocardial histopathology {#S0002-S2004}\n-------------------------\n\nThe mice that survived were sacrificed on days 7 and 14. The myocardium of the VMC group showed a severe injury with cellular infiltration and necrosis. Nicotine alleviated the severity of cellular infiltration and necrosis compared with the VMC group. (P\u00a0\\<\u00a00.05) ([Figure 4](#F0004)). There was no significant difference in histological score between the VMC group and the MLA group.10.1080/21505594.2018.1482179-F0004Figure 4.Myocardial histopathology changes for each group (n\u00a0=\u00a05 in each group). A and E show the lack of histopathology changes for the control group on days 7 and 14; B and F show representative histopathology changes for the myocarditis group on days 7 and 14. Several small foci of cellular infiltrations are shown; C and G show representative histopathological changes for the nicotine group on days 7 and 14. The extent of cellular infiltration was less severe in the nicotine group than in the methyllycaconitine and myocarditis groups; small and limited foci of cellular infiltrations were obtained. D and H show representative histopathological changes for the methyllycaconitine group on days 7 and 14. Several large foci of cellular infiltrations are obtained. I and H are the Box-plots of the histopathology scores for each group on days 7 and 14. The score of the nicotine group was significantly lower than those of the myocarditis and methyllycaconitine groups. ^&^P\u00a0\\<\u00a00.05 versus the myocarditis group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nViral replication in the myocardium {#S0002-S2005}\n-----------------------------------\n\nWe detected the CVB3-RNA abundance in the myocardium of the VMC, nicotine and MLA groups by fluorescent quantitative PCR-analysis. There were no significant differences between the treatment and VMC groups on days 7 and 14. ([Figure 3](#F0003))\n\nThe levels of the pro-inflammatory cytokines in the heart on days 7 and 14 {#S0002-S2006}\n--------------------------------------------------------------------------\n\nThen, we examined the expression of IL-1, IL-6 and TNF-\u03b1 for each group on days 7 and 14. On day 7, the levels of IL-1, IL-6 and TNF-\u03b1 were significantly lower in the nicotine group than in the VMC group. (P\u00a0\\<\u00a00.05) However, the expression of those pro-inflammatory cytokines was elevated in the MLA group compared to the VMC group. (P\u00a0\\<\u00a00.05) On day 14, no significant differences were found in the level of pro-inflammatory cytokines among the Nicotine group, VMC group and MLA group. (P\u00a0\\>\u00a00.05) ([Figure 5](#F0005))10.1080/21505594.2018.1482179-F0005Figure 5.Expression of pro-inflammatory cytokines in the heart (n\u00a0=\u00a05 in each group). A. Nicotine treatment significantly deceased the expression of IL-1, IL-6 and TNF-\u03b1 compared with the myocarditis group on day 7. B. On day 14, no significant differences were found in the level of pro-inflammatory cytokines among the nicotine group, myocarditis group and methyllycaconitine group. ^&^P\u00a0\\<\u00a00.05 versus the myocarditis group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nThe expression of specific transcription factors and cytokines in a Th cell subset in the heart on days 7 and 14 {#S0002-S2007}\n----------------------------------------------------------------------------------------------------------------\n\nIFN-\u03b3, IL-4, IL-17 are specific cytokines for Th1, Th2 and Th17 cells, respectively; therefore, we could observe the percentage variations of Th1, Th2, Th17, and Treg cells with changes in the levels of specific transcription factors and cytokines of the heart for each group. On day 7, the levels of IFN-\u03b3, IL-17, T-bet and ROR-\u03b3 were lower in the nicotine group than in the VMC group. Otherwise, the levels of GATA3, Foxp3 and IL-4 were higher in the nicotine group than in the VMC group. (P\u00a0\\<\u00a00.05) However, the MLA group had the opposite result. Compared with the VMC group, the expression levels of IFN-\u03b3, IL-17, T-bet and ROR-\u03b3 were elevated, and the expression levels of GATA3, Foxp3 and IL-4 were decreased. (P\u00a0\\<\u00a00.05) ([Figure 6](#F0006)) On day 14, there was no significant difference between the treatment and VMC groups. ([Figure 7](#F0007))10.1080/21505594.2018.1482179-F0006Figure 6.Expression of specific transcription factors and cytokines for Th cell subsets in the heart on day 7 (n\u00a0=\u00a05 in each group). The levels of IFN-\u03b3, IL-17, T-bet and ROR-\u03b3 were lower in the nicotine group than in the myocarditis group. Otherwise, the levels of GATA3, Foxp3 and IL-4 were higher in the nicotine group than in the myocarditis group. A is the absolute intensity ratio of the cytokines for the Th cell subset. B is the absolute intensity ratio of the specific transcription factors. ^&^P\u00a0\\<\u00a00.05 versus the myocarditis group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.10.1080/21505594.2018.1482179-F0007Figure 7.Expression of specific transcription factors and cytokines of the Th cell subset in the heart on day 14 (n\u00a0=\u00a05 in each group). There was no significant difference between the treatment and myocarditis groups. A is the absolute intensity ratio of cytokines for the Th cell subset. B is the absolute intensity ratio of the specific transcription factors. ^&^P\u00a0\\<\u00a00.05 versus the myocarditis group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nThe percentage of th subsets in the spleen in vivo {#S0002-S2008}\n--------------------------------------------------\n\nThe percentage of Th subsets in the spleens of control and VMC mice was analyzed. Compared with the VMC group, the percentages of Th1 and Th17 cells decreased, and the percentages of Th2 and Treg cells increased in the nicotine group on day 7.(P\u00a0\\<\u00a00.05) Compared with the VMC group, the percentages of Th1 and Th17 cells were elevated and the percentages of Th2 and Treg were decreased in the MLA group. (P\u00a0\\<\u00a00.05) ([Figure 8](#F0008))10.1080/21505594.2018.1482179-F0008Figure 8.The percentages of Th subsets in the spleen (n\u00a0=\u00a05 in each group). On day 7, the percentages of Th1 and Th17 cells were deceased, while the percentages of Th2 and Treg cells were increased in the nicotine group. ^&^P\u00a0\\<\u00a00.05 versus the myocarditis group and ^\\#^P\u00a0\\<\u00a00.05 versus the methyllycaconitine group.\n\nDiscussion {#S0003}\n==========\n\nThe results of this study indicated the regulatory effect of the CAP on the differentiation of CD4^+^ T cells in viral myocarditis. First, stimulation of spleen cells from VMC mice with nicotine in vitro altered the percentage of Th cell subsets by increasing the percentage of Th2 and Treg cells and decreasing the percentage of Th1 and Th17 cells. Second, nicotine treatment in vivo altered the percentages of Th cell subsets in the spleen in a similar manner. The change in the spleen resulted in variations in the specific transcription factors and cytokines expression in the myocardium. The expression levels of Th2-specific transcription factor GATA3 and Treg cells-specific transcription factor Foxp3 were elevated, and the expression levels of Th1-specific transcription factor T-bet and Th17-specific transcription factor ROR-\u03b3 were reduced in the myocardium. Finally, nicotine treatment reduced inflammatory cytokines (IL-1\u03b2, IL-6, and TNF-\u03b1) in the myocardium, ameliorated the necrosis of cardiomyocytes and cellular infiltration in mice with viral myocarditis.\n\nThe CAP, which is composed of the vagus nerve, acetylcholine secreted by the vagus nerve terminal and receptor \u03b17 acetylcholine receptors, is important. Infection and ischemia release cytokines, activate signals, and transfer factors into the nucleus tractus solitarius via the vagus nerve and through the cranial nerve network, activating the release of acetylcholine on the vagus nerve. Acetylcholine activates \u03b17 nicotinic acetylcholine receptors on macrophages, lymphocytes and other inflammatory cells, regulating the synthesis and release of inflammatory cytokines and attenuating the systemic inflammatory response \\[[14](#CIT0014)\\]. Compared with traditional humoral immunity, the CAP is extremely sensitive and rapid, with accurate localization of the inflammatory response to the infected tissue. The CAP had an anti-inflammatory effect at the early phase of the inflammatory response \\[[12](#CIT0012)\\]. However, the protective effect disappeared after splenectomy \\[[15](#CIT0015)\\]. The spleen was the intermediate link for the CAP in which CD4^+^ T cells and their subsets played an important role in the anti-inflammatory function. Many studies have shown that the CAP could regulate the differentiation of CD4^+^ T cells by inhibiting the release of inflammatory cytokines, improving the prognosis of inflammatory diseases. Wu et al. found that the CAP reduced the percentage of Th17 cells, increased the percentage of Th2 cells and attenuated the inflammatory response of arthritis in mice \\[[16](#CIT0016)\\]. Nizri et al. also demonstrated that nicotine treatment inhibited the response of Th1 and Th17 cells and attenuated neuroinflammation in autoimmunity encephalomyelitis \\[[17](#CIT0017)\\]. Nicotine treatment could ameliorate inflammatory bowel disease severity through increasing the proportion of Treg cells and decreasing the proportion of Th17 cells \\[[11](#CIT0011)\\]. In this research, we proposed that the CAP could regulate the differentiation of CD4^+^ T cells in viral myocarditis. The results revealed that nicotine could downregulate the expression of T-bet and ROR-\u03b3 as well as upregulate the expression of GATA3 and Foxp3 in CD4^+^ T cells. The results of flow cytometry were also consistent with those of WB. Nicotine treatment decreased the percentages of Th1 and Th17 cells and increased the percentages of Th2 and Treg cells in the spleen. The proportions of Th1 and Th17 cells were increased and the proportions of Th2 and Treg cells were decreased in the MLA group. These results demonstrated that the CAP could regulate the differentiation of spleen CD4^+^ T cell subsets, decrease the percentages of Th1 and Th17 cells and increase the proportions of Th2 and Treg cells.\n\nIn the first two stages of the process of viral myocarditis, there are many mechanisms involved in injury to the myocardium, including the following four features: 1. direct injury with virus infection and replication; 2. virus-specific immunity; 3. adaptive immunity induced by viral replication; and 4. autoimmunity \\[[18](#CIT0018)\\]. CD4^+^ T cells and their subgroups mediate adaptive immunity, participating in the development and progression of viral myocarditis. For Th17 cells, substantial evidence had demonstrated that they are involved in the pathogenesis of inflammatory diseases. It had been found that anti-IL-6 treatment could reduce the proportion of Th17 cells, reducing the inflammatory reaction of autoimmune myocarditis \\[[19](#CIT0019)\\]. In viral myocarditis, Yang et al. found that administration of recombinant IL-23 reduced the percentage of Th17 cells and inflammation in mice \\[[20](#CIT0020)\\]. Similarly, IL-27 inhibited the differentiation of Th17 cells, ameliorated the symptoms and improved of the survival rate of viral myocarditis in mice \\[[21](#CIT0021)\\]. However, the effect of Th2 and Treg cells is unclear, although many studies have shown that both Th2 and Treg cells could attenuate the inflammatory response and play a protective role in inflammatory diseases. It had been discovered that female mice, which predominantly have Th2 cell immune responses, are less likely to develop viral myocarditis \\[[22](#CIT0022)\\]. This is reported that Treg cells were also involved in the susceptibility differences. The monocytic myeloid-derived suppressor cells upregulated the percentages of Treg and CD4^+^ IL-10^+^ T cells to reduce the inflammatory response in viral myocarditis \\[[23](#CIT0023)\\]. Th2 cell-associated cytokine IL-33 could alleviate the severity of viral myocarditis in TLR3 knock-out mice, increase the percentage of Th2 cells, and prevent the progression from acute myocarditis to chronic myocarditis \\[[24](#CIT0024)\\]. Papageorgiou et al. found that type-2 thrombospondin could activate Treg cells, reduce myocardium lesions and improve cardiac dysfunction in CVB3-induced myocarditis \\[[25](#CIT0025)\\]. These studies indicated that Th2 and Treg cells play a protective role in viral myocarditis, improving the disease prognosis. However, unlike the Th cell subsets described above, the role of Th1 in viral myocarditis is not fully understood. Jian et al. reported that IFN-\u03b1, a Th1 cell-related cytokine, could induce differentiation of splenic T cells to Th1 cells in viral myocarditis mice, which could reduce virus replication and protect the mice \\[[26](#CIT0026)\\]. Similarly, IFN-\u03b2, aTh1 cell-related inflammatory cytokine, could prevent viral replication and improve the prognosis of viral myocarditis \\[[27](#CIT0027)\\]. However, Yue et al. found that mutations in the IP-10 gene could suppress the level of IFN-\u03b3, reducing myocardial lesions and inflammatory infiltration \\[[28](#CIT0028)\\]. In addition, they reported that mutations of MCP-1 reduced the mortality, serum markers, and histopathological changes of viral myocarditis mice \\[[29](#CIT0029)\\]. There was a divergence of the effect of Th1 cells. IFN-\u03b1/\u03b2 could prevent viral replication, but IFN-\u03b3 would increase the inflammatory response. Further research is needed to determine the role Th1 cells play in inflammatory diseases. Based on the above results and theoretical basis, after clarifying the regulation effect of the CAP on the differentiation of splenic CD4^+^ T cells, we further found that nicotine had the effect of the percentage alteration of CD4^+^ T cells subsets on viral myocarditis. In the acute and subacute phases of viral myocarditis, high levels of viral replication led to cardiomyocytes necrosis, activated the acquired immune system of the host, and resulted in infiltration of the myocardial interstitium. The results of WB showed that nicotine treatment could change the expression of Th cell-specific cytokines and transcription factors in the myocardium of mice with viral myocarditis. The levels of Th1 cell-associated T-bet and IFN-\u03b3 as well as Th17 cell-associated ROR-\u03b3 and IL-17 expression were upregulated. Th2 cell-related GATA3 and IL-4 and Treg cell-related Foxp3 expression levels were downregulated. Thus, we suggested that the CAP could alter the proportion of CD4\u00a0+\u00a0T-lymphocyte subsets in the spleens of mice, and possibly alter the percentage of T cell subgroups in the myocardium interstitium. This change in the proportion of Th cells ultimately reduced the systemic inflammatory response, alleviated myocardial injury, and protected the mice from viral myocarditis.\n\nInterestingly, there were no significant differences in those groups in terms of the abundance of CVB3 RNA in the myocardium. This might be related to inhibition of Th1 cell differentiation. Previous studies indicated that although Th1 cells inhibited viral replication, they increase the systemic inflammatory response. Our study revealed that activation of the CAP decreased the proportion of Th1 cells and IFN-\u03b3, reducing the anti-virus effect and systemic inflammatory response, which resulted in a net a protective effect. In addition, inflammatory cytokines and Th cell-associated transcription factors were not significantly different at 14\u00a0days, which might be related to the progression of viral myocarditis. The virus is nearly eliminated in the animal model of viral myocarditis at 14\u00a0days; by that time, it is progressing into the chronic phase \\[[30](#CIT0030)\\]. Although our previous studies and this experiment demonstrated that the cholinergic anti-inflammatory pathway could have an anti-inflammatory effect in the early phase of inflammation, evaluation of the effect of the CAP in the chronic phase requires further study.\n\nThe analyses in vivo at day 7 and day 14 were performed on mice that survived after CVB3 infection in the study. Although the analyzed mice in each group were selected randomly, the study might have the potential of survivor bias, due to excluding mice that die prior to these experimental endpoints from the analysis.\n\nConclusion {#S0003-S2001}\n----------\n\nIn this study, we preliminarily found that the cholinergic anti-inflammatory pathway could regulate CD4^+^ T cell differentiation, increase the percentages of Th2 and Treg cells, and decrease the percentages of Th1 and Th17 cells in the spleen. Those effects possible changed the levels of specific transcription factors of CD4^+^ T cell subsets in the myocardium, reduced myocardial lesions and inflammatory infiltration, attenuated the expression of pro-inflammatory mediators in the acute and subacute phases of the viral myocarditis model. This research provides a new understanding of the development and progression of viral myocarditis as well as provides a potential new target for myocarditis treatment.\n\nMaterials and methods {#S0004}\n=====================\n\nMice {#S0004-S2001}\n----\n\nMale, specific pathogen-free, 4-week-old BALB/c mice were obtained from the Laboratory Animal Center of Shanghai (Shanghai, China) and were kept in a pathogen-free facility in the experimental animal center of the Wenzhou Medical University. All animals received humane care according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication, 8th Edition, 2011). The study protocol was approved by the Wenzhou Medical University Committee on Ethics in the Care and Use of Laboratory Animals.\n\nVirus {#S0004-S2002}\n-----\n\nCVB3 (Nancy strain) was maintained in Hep2 cells. A 50% tissue culture infectious dose (TCID50) assay was used to determine the viral titer. BALB/c mice were intraperitoneally inoculated with 0.2\u00a0mL of normal saline containing 10^5^ TCID50 of the virus to generate a viral myocarditis model. The control group as intraperitoneally injected with the same dose of normal saline.\n\nIn vitro CD4^+^ T cell culture {#S0004-S2003}\n------------------------------\n\nCD4^+^ T cells were purified by positive selection using MACS (Miltenyi Biotec, \\# 130--049-201) from the spleen of 7-day infected VMC mice. The day of cells separation was defined as day 0. All cells were cultured in six-well plates in the presence of nicotine (Sigma-Aldrich Co, \\#N3876) or methyllycaconitine (Sigma-Aldrich Co, M168). They were randomly divided into three groups, the phosphate buffer solution (PBS) group, nicotine group (100 \u03bcmol/L), and MLA group (100 \u03bcmol/L). All drugs were administered once daily for 5 consecutive days, starting on day 1. Cells were harvested and subjected to flow cytometry and Western blot analyses.\n\nIn vivo mouse experiment {#S0004-S2004}\n------------------------\n\nThe day of virus inoculation was defined as day 0. Nicotine (1.2\u00a0mg/kg per day) and methyllycaconitine (2.4\u00a0mg/kg per day) were intraperitoneally given to VMC mice for consecutive 14\u00a0days, starting 24\u00a0h after viral inoculation. Meanwhile, the control and VMC groups were intraperitoneally given the same dose of normal saline solution. Eight surviving mice from each group were sacrificed to extract the heart and spleen on days 7 and 14. The spleens were used for flow cytometry analysis and the hearts were divided into three parts. One part was used for Western blot analyses, another for studying myocardial histopathological and the last for q-PCR analysis.\n\nSurvival analysis {#S0004-S2005}\n-----------------\n\nWe randomly chose 40 mice from each group for a 14-day survival analysis.\n\nHistopathological analysis and myocarditis scoring {#S0004-S2006}\n--------------------------------------------------\n\nThe hearts collected from the CVB3 infected mice on days 7 and 14 were fixed in 10% formalin overnight. After fixation, the heart tissues were placed in a series of alcohol for dehydration and then embedded in paraffin. Sections (5 \u03bcm thick) of hearts tissues were cut and stained with hematoxylin and eosin (H&E). Pathological scores were blindly graded by two independent observers based on the following semi-quantitative scale: 0\u00a0=\u00a0no lesion; 1\u00a0=\u00a0lesion involving 25% of the myocardium; 2\u00a0=\u00a0lesions involving 25 to 50% of the myocardium; 3\u00a0=\u00a0lesions involving 50 to 75% of the myocardium; and 4\u00a0=\u00a0lesions involving 75 to 100% of the myocardium, as previously described \\[[31](#CIT0031)\\]. The scores for every section were averaged.\n\nFlow cytometry {#S0004-S2007}\n--------------\n\nPurified CD4^+^ T cells were counted and suspended in RPMI 1640 containing 10% fetal bovine serum at a density of 1\u00a0\u00d7\u00a010^6^ cell/ml. To analyze the percentages of Th1, Th2 and Th17 cells, the cells were stimulated for 6\u00a0h with leukocyte activation cocktail (BD Pharmingen, \\#550,583) at 37\u00b0C, 5% CO~2~ in a 24-well culture plate. After 6\u00a0h of incubation, the cells were harvested and stained with FITC-conjugated anti-mouse CD4 antibody (BD Pharmingen, \\#553,046). After the cells were washed, the cells were fixed and permeabilized by BD cytofix/cytoperm plus fixation/permeabilization kit (BD Pharminge, \\#554,715). The cells were intracellularly stained with PE-Cy5.5-conjugated anti-mouse IFN-\u03b3 antibody (BD Pharmingen, \\# 560,660), PE-conjugated anti-mouse IL-17A antibody (BD Pharmingen, \\# 560,436) and APC-conjugated anti-mouse IL-4 antibody (BD Pharmingen, \\# 554,436). For the Tregs, suspended unstimulated CD4^+^ T cells were stained with FITC-conjugated anti-mouse CD4 antibody and APC-conjugated anti-mouse CD25 antibody (BD Pharmingen, \\#561,048). After washing, fixing, and permeabilizing, the cells were intracellularly stained with PE-conjugated anti-mouse Foxp3 antibody (BD Pharmingen, \\#560,414). Then, the cells were measured on a FACS-Calibur flow cytometer. The data were analyzed by FlowJo software. For gating, we used isotype staining. The isotype control we used are as followed: PerCP-Cy\u21225.5 Rat IgG1, \u03ba Isotype Control(BD Pharmingen, \\# 554,436), PE Rat IgG1, \u03ba Isotype Control (BD Pharmingen, \\# 554,685) and APC Rat IgG1, \u03ba Isotype Control (BD Pharmingen, \\# 554,686). For Th1, Th2 and Th17 cells, we selected lymphocytes based on their forward and side scatter properties. Then we gated CD4^+^ IFN-\u03b3^+^ Th1 cells, CD4^+^ IL-4^+^ Th2 cells and CD4^+^ IL-17^+^ Th17 cells based on their fluorescence intensity, respectively. For Treg cells, we selected lymphocytes and CD4^+^ Th cells. Then we gated CD25^+^ Foxp3^+^ Treg cell based on their fluorescence intensity. The experiment was repeated 3 times for each sample.\n\nReal-time PCR {#S0004-S2008}\n-------------\n\nTotal RNA was obtained with TRIzol Reagent (Invitrogen, \\# 10,296,010) and then converted into cDNA with a RevertAid RT Reverse Transcription Kit (Thermo fisher, \\# K1691) according to the manufacturer's instructions. Real time-polymerase chain reaction (RT-PCR) was performed with a LightCycler\u00ae 480 System (Roche) using SYBR Green I Master (Roche, \\# 04707516001). The primer sequences are as follows: CVB3: F-GTCTGCCTGCGTTTATTTC, R-ACTCAGCGTATCGTTTGGA and GAPDH: F-AGGGAAATCGTGCGTGACAT, R-CATCTGCTGGAAGGTGGACA. The relative gene expressions were normalized to the level of GAPDH mRNA transcripts and quantified by the 2-\u25b3\u25b3CT method. The experiment was repeated 3 times for each sample.\n\nWestern blot {#S0004-S2009}\n------------\n\nThe total proteins of heart and spleen were lysed in Radio Immunoprecipitation Assay (RIPA) Lysis Buffer (Beyotime Biotechnology, \\#P0013B), which contained PMSF (Beyotime Biotechnology, \\#ST506). A bicinchoninic (BCA) protein assay kit (Thermo fisher, \\#23,225) was used to detect the protein concentration. Samples containing 40 \u03bcg of protein were separated on a 10 to 15% SDS-PAGE gel and transferred to a polyvinylidene (PVDF) membrane. The membranes were blocked for 2\u00a0h with 5% nonfat milk powder at room temperature. Then, the membranes were incubated with primary antibody at 4\u2103 overnight, and they were incubated with a secondary HRP-conjugated antibody (1:5000, biosharp, \\#BL002A) for 2\u00a0h on the second day. Then, the membranes were detected with ECL. The protein bands were scanned, and the band density was calculated using AlphaEaseFC software. GAPDH was used as a control to calculate the expression of all proteins. The antibodies used included the following: mouse anti-T-bet monoclonal antibody (1:1000, Abcam, \\#ab91109), rabbit anti-GATA3 polyclonal antibody (1:1000; Abcam, \\#ab106625), rabbit anti-ROR\u03b3 polyclonal antibody (1:1000; Abcam, \\# ab207082), goat anti-IL-1 monoclonal antibody (1:1000; Cell Signaling Technology, \\#50,794), goat anti-IL-6 monoclonal antibody (1:1000; Cell Signaling Technology, \\#12,912), rabbit anti-TNF polyclonal antibody (1:1000; Abcam, \\#ab6671), rat anti-IL-4 polyclonal antibody (1:1000; Abcam, \\#ab11524), rabbit anti-IL-17 polyclonal antibody (1:1000; Abcam, \\# ab79056), rabbit anti-Foxp3 polyclonal antibody (1:1000; Abcam, \\# ab54501) and rabbit anti-GAPDH polyclonal antibody (1:1000; Cell Signaling Technology, \\# 5174). The experiment was repeated 3 times for each sample.\n\nStatistics {#S0005}\n==========\n\nData are expressed as the mean\u00b1S.D. The Kaplan-Meier method was used to analyze the survival rate. Comparisons of each group were performed by one-way analysis of variance (ANOVA) and the Least -- Significant Difference (LSD). The differences in the pathological scores were evaluated using a Mann--Whitney U-test. Analysis was performed with SPSS 17.0 statistical software for Windows. A value of p\u00a0\\<\u00a00.05 was considered significant.\n\nDisclosure statement {#S0006}\n====================\n\nNo potential conflict of interest was reported by the authors.\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "INTRODUCTION\n============\n\nAtopic dermatitis (AD) has been considered as a T helper (Th) cell type 2 disease characterized by predominant Th2-mediated cytokine production, including IL-4, IL-5, and IL-13, elevated serum IgE, and eosinophilia ([@B1][@B2]). Thymic stromal lymphopoietin (TSLP), IL-25, and IL-33 are mainly produced by epithelial cells, and have important functions of inducing Th2-type adaptive responses and group 2 innate lymphoid cells, which contribute to AD phenotype ([@B3][@B4]). However, Th1 cells and innate inflammatory cytokines are also involved in the pathogenesis of AD. Interferon-\u03b3 (IFN-\u03b3) and IL-12 are also expressed in chronic AD lesions. The expression of IFN-\u03b3, but not IL-4, in skin correlates with the clinical course of AD and is known to be downregulated with AD improvement ([@B5]). AD patients have elevated levels of CC chemokine receptor 4 ligands, including thymus and activation-regulated chemokine (TARC) and macrophage-derived chemokine (MDC), which mediate preferential Th2 recruitment ([@B6]). Additionally, CXC chemokine receptor 3 ligands, which induce Th1 polarization, such as monokine induced by IFN-\u03b3, are also elevated in AD patients compared to normal controls ([@B6]). In addition to inflammatory microenvironment imbalance, important barrier-related proteins including filaggrin (FLG) and loricrin (LOR) are decreased in AD ([@B7]). In primary keratinocytes, tumor necrosis factor (TNF)-\u03b1 downregulates FLG and LOR through c-Jun N-terminal kinase ([@B8]). These results suggest that complex inflammatory networks, involving innate, Th1, and Th2 responses, orchestrate AD pathogenesis. The aim of this study was to investigate the effects of innate, Th1, and Th2 inflammatory cytokines on the expression of genes implicated in the pathogenesis of AD using human keratinocyte cell line HaCaT.\n\nMATERIALS AND METHODS\n=====================\n\nHaCaT cell culture and cytokine stimulation\n-------------------------------------------\n\nHaCaT cells were grown to 100% confluence and starved for 24 hours in DMEM without fetal bovine serum, followed by their stimulation with various combinations of innate (TNF-\u03b1, 10 ng/ml; PeptroTech, Rocky Hill, NJ, USA), Th1 (IFN-\u03b3, 10 ng/ml; PeptroTech), and Th2 (IL-4, 50 ng/ml; PeptroTech) cytokines for 24 hours ([@B9][@B10][@B11]). The morphological changes in cells were examined using a phase-contrast microscope (Olympus CX41; Olympus, Tokyo, Japan).\n\nCell viability assay\n--------------------\n\nThe effects of cytokine stimulation on HaCaT cell growth were determined using a water-soluble tetrazolium salt assay kit according to the manufacturer\\'s instructions (EZ-Cytox cell viability assay kit; ITSBio, Seoul, Korea).\n\nHuman sample\n------------\n\nSkin specimens from six patients with AD and 12 healthy controls were collected under the approval of the Institutional Review Board at Gachon University Gil Medical Center (GBIRB2016-082). All 6 AD patients received no treatment for at least 4 weeks and had chronic disease (average duration 6.83\u00b14.40 years) with a recent exacerbation in 1--3 months AD (SCORAD between 20 and 70, all with elevated IgE) ([Table 1](#T1){ref-type=\"table\"}).\n\n###### Characteristics of patients with atopic dermatitis\n\n![](in-18-e9-i001)\n\n  Patient No.   Sex/age   Duration (yr)   Aggravation (mon)   Co-morbidities\n  ------------- --------- --------------- ------------------- -------------------\n  1             M/20      10              3                   Asthma\n  2             M/19      12              2                   Allergic rhinitis\n  3             M/21      2               2                   Allergic rhinitis\n  4             M/13      5               2                   Urticaria\n  5             M/25      10              1                   \\-\n  6             M/46      2               3                   \\-\n\nReal-time PCR validation\n------------------------\n\nTotal RNA from skin specimens and HaCaT cells was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany). The viable HaCaT cells were harvested with density centrifugation using 40% Percoll (Sigma-Aldrich, St. Louis, MO, USA) and used for RNA extraction. The mRNA level of *CCL22* (encoding MDC), *CCL17* (encoding TARC), *IL5*, *IL13*, *FLG*, *LOR*, *IL25*, *TSLP*, and *IL33* in cytokine-stimulated HaCaT cells was evaluated by real-time PCR and compared with the expression profiles examined in the lesional skin of AD patients. Primer sequences are listed in [Table 2](#T2){ref-type=\"table\"}.\n\n###### Primer sequences for real-time PCR\n\n![](in-18-e9-i002)\n\n  Target gene                                    Primer sequence\n  ---------------------------------------------- ------------------------------------------------\n  *CCL22*                                        Forward: 5\u2032-GAA GCC TGT GCC AAC TCT CT-3\u2032\n  Reverse: 5\u2032-GGG AAT CGC TGA TGG GAA CA-3\u2032      \n  *CCL17*                                        Forward: 5\u2032-CGG ACC CCA ACA ACA AGA GA-3\u2032\n  Reverse: 5\u2032-CTC CCT CAC TGT GGC TCT TC-3\u2032      \n  *IL5*                                          Forward: 5\u2032-TCT ACT CAT CGA ACT CTG CTG A-3\u2032\n  Reverse: 5\u2032-CCC TTG CAC AGT TTG ACT CTC-3\u2032     \n  *IL13*                                         Forward: 5\u2032-TGT TTG TCA CCG TTG GGG AT-3\u2032\n  Reverse: 5\u2032-TGA GTC TCT GAA CCC TTG GC-3\u2032      \n  *FLG*                                          Forward: 5\u2032-TGA AGC CTA TGA CAC CAC TGA-3\u2032\n  Reverse: 5\u2032-TCC CCT ACG CTT TCT TGT CCT-3\u2032     \n  *LOR*                                          Forward: 5\u2032-GAG GTG TTT TCC AGG GGC A-3\u2032\n  Reverse: 5\u2032-TGG GGT TGG GAG GTA GTT GTA-3\u2032     \n  *IL33*                                         Forward: 5\u2032-TGT CAC ATT GGG CAA AGT T-3\u2032\n  Reverse: 5\u2032-CAG TAA GCA GTG TTA TCA GGA A-3\u2032   \n  *IL25*                                         Forward: 5\u2032- TTG TTT GTT TAC TCA TCA CTC AG-3\u2032\n  Reverse: 5\u2032- TCC TCC TCA GAA TCA TCC A-3\u2032      \n  *TSLP*                                         Forward: 5\u2032-TCC TCT GAA GAC CTG ACC-3\u2032\n  Reverse: 5\u2032-TCT CCT TTC TCC CTA ATC CTC-3\u2032     \n  *GAPDH*                                        Forward: 5\u2032-CTG GGC TAC ACT GAG CAC C-3\u2032\n  Reverse: 5\u2032-AAG TGG TCG TTG AGG GCA ATG-3\u2032     \n\nStatistical analysis\n--------------------\n\nThe data are presented as the mean\u00b1standard error of the mean (SEM). All data are representatives of 2 or more independent experiments. When necessary, a 2-group comparison was performed using a Student\\'s *t*-test. A p-value \\<0.05 was considered statistically significant.\n\nRESULTS\n=======\n\nThe viability of HaCaT cells was unaltered following treatment with TNF-\u03b1, IFN-\u03b3, or IL-4 alone or the combination of TNF-\u03b1 and IL-4 ([Fig. 1](#F1){ref-type=\"fig\"}). Moreover, combinatorial treatment of IFN-\u03b3 and IL-4 did not affect the viability of HaCaT cells (data now shown). However, treatment of HaCaT cells with the combination of TNF-\u03b1 and IFN-\u03b3 resulted in a significant reduction in the cell viability ([Fig. 1](#F1){ref-type=\"fig\"}). This decrease in the viability of cells was likely attributable to the synergistic cytotoxic effects of IFN-\u03b3 and TNF-\u03b1. Next, we compared the gene expression profiles of HaCaT cells stimulated with various combinations of cytokines with those observed in human AD skin samples ([Fig. 2](#F2){ref-type=\"fig\"}). Consistent with previous reports ([@B1][@B2][@B5][@B7]), the expression of genes encoding Th2 chemokines (*CCL22* and *CCL17*) and Th2 cytokines (*IL5*, *IL13*, *IL25*, and *TSLP*) was increased in human AD skin lesions, while the expression of genes related to the cornified cell envelope (*FLG* and *LOR*) was reduced. The expression of IL-33, a recently described Th2-linked cytokine ([@B12]), was significantly decreased in AD skin samples. The gene expression profiles observed for HaCaT cells were similar upon their stimulation with TNF-\u03b1 and/or IFN-\u03b3 but not IL-4. Treatment with IFN-\u03b3 alone or TNF-\u03b1 and IFN-\u03b3 combination augmented the expression of *CCL22* and *CCL17* and decreased the expression of *FLG* and *LOR*. Expression of IL-25 and TSLP was increased by combined stimulation with IL-4 and TNF-\u03b1. IL-4 treatment inhibited the expression of *CCL22* and *CCL17* induced by TNF-\u03b1 and IFN-\u03b3 in HaCaT cells, as previously reported ([@B13][@B14]). Furthermore, *FLG* expression increased upon treatment of HaCaT cells with IL-4 alone but decreased in the presence of the combination of IL-4 and TNF-\u03b1. These results imply that the innate and Th1 inflammatory cytokines are required to reproduce AD-like features in HaCaT cells. Stimulation of HaCaT cells with TNF-\u03b1 and IFN-\u03b3 significantly upregulated the expression of *IL33* gene. The expression of classical Th2 cytokines IL-5, IL-13, and IL-25 was markedly upregulated in HaCaT cells following treatment with IL-4.\n\n![Assessment of HaCaT cell viability following their stimulation with various combinations of Th1 (IFN-\u03b3, 10 ng/ml or TNF-\u03b1, 10 ng/ml) and Th2 (IL-4, 50 ng/ml) cytokines for 24 hours. (A) Evaluation of the morphological changes in cytokine-stimulated HaCaT cells using phase-contrast microscopy. Original magnification \u00d720. (B) The effects of cytokine stimulation on the growth of HaCaT cells were measured using a water-soluble tetrazolium salt assay and the cell viability was determined by measuring the absorbance at 450 nm wavelength. Graphs show the mean\u00b1standard error of the mean.\\\n^\\*\\*\\*^p\\<0.001 (Student\\'s *t*-test).](in-18-e9-g001){#F1}\n\n![Correlation between AD-related gene expression in lesional skin and that observed in cytokine-stimulated HaCaT cells. HaCaT cells were cultured in the presence of IFN-\u03b3 (10 ng/ml), TNF-\u03b1 (10 ng/ml), and/or IL-4 (50 ng/ml) for 24 hours. The expression of *CCL22* (encoding macrophage-derived chemokine), *CCL17* (encoding thymus and activation-regulated chemokine), *IL5*, *IL13*, *FLG* (filaggrin), *LOR* (loricrin), *IL33, IL25*, and *TSLP* was evaluated by real-time PCR as a fold change normalized to the expression of *GAPDH* in skin samples from AD patients (left) and cytokine-stimulated HaCaT cells (right). mRNA expression of each gene in cytokine-stimulated HaCaT cells was compared with non-stimulated control. Graphs show the mean\u00b1standard error of the mean.\\\nMDC, macrophage-derived chemokine.\\\n^\\*^p\\<0.05, ^\\*\\*^p\\<0.01, ^\\*\\*\\*^p\\<0.001 (Student\\'s *t*-test).](in-18-e9-g002){#F2}\n\nDISCUSSION\n==========\n\nAD is a complex inflammatory skin condition, which is incompletely understood. Epidermal barrier defects and dysregulated Th2 immune responses have crucial roles in the pathogenesis of AD. However, it has been reported that human keratinocytes are more sensitive to Th1-derived lymphocytes than those derived from Th2 in terms of chemokine release ([@B15]). In addition, IL-4 influences Th1-type responses, including antigen-induced arthritis, autoimmune uveoretinitis, and T cell transfer model of colitis ([@B16][@B17][@B18]). TNF-\u03b1 along with Th2 cytokines plays an important role to induce AD-like features in epidermal differentiation proteins ([@B19]). Our observations have indicated that innate and Th1-type cytokines induce pathogenic changes implicated in the development of AD in HaCaT cells. Although primary human epidermal keratinocytes could be a primary choice to study molecular characteristics of skin, immortalized keratinocytes such as HaCaT cells may be an alternative model, considering their usefulness for *in vitro* assays ([@B20]). HaCaT cells stimulated with cytokines successfully reproduced inflammatory changes, except for IL-33 expression, observed in skin lesions of AD patients ([Fig. 2](#F2){ref-type=\"fig\"}). IL-33, IL-25, and TSLP are known to drive type 2 innate lymphoid cell response ([@B3][@B4]). Despite their similarities, IL-25 and IL-33 have distinct features. IL-25 is constitutively expressed in cellular compartments of epithelial cells and released upon exposure to allergen proteases, whereas IL-33 is a nuclear protein, which has regulatory proteins in its nuclear localization ([@B21]), and acts as an alarming signal once it is released by damaged epithelial cells ([@B22]). Because the number of mast cells is also increased in AD, overproduction of mast cell chymase could degrade IL-33 and affect its balance in AD ([@B23]). In addition, IFN-\u03b3 is another key molecule which is reported to upregulate IL-33 levels in keratinocytes of AD patients ([@B9][@B24]). Since most AD patients in this study exhibited chronic disease with recent exacerbation and decreased *IFNG* expression (data not shown), it is plausible that disease status affects the cytokine expression profiles in the lesion of AD. Therefore, cytokine-stimulated HaCaT cell line could be an appropriate tool for the demonstration of mixed chronic and aggravated status of AD. This hypothesis is further supported by the remarkable decrease in cell viability of HaCaT cells upon stimulation with TNF-\u03b1 and IFN-\u03b3 ([Fig. 1](#F1){ref-type=\"fig\"}) and increased expression of IL-33 from damaged HaCaT cells. Taken together, our data suggest that Th1 and Th2 cytokines do not function dichotomously and that a complicated inflammatory network drives AD-like changes. Therefore, future *in vitro* experiments using HaCaT cells should employ various cytokine combinations beyond Th2 polarization to observe optimal expression of AD-related genes.\n\nThe authors thank Eun-Hui Lee and Jinsun Jang (Gachon University, Korea) for their technical assistance. This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2016R1D1A1A09916492).\n\n**Conflict of Interest:** The authors have no conflict of interest to declare.\n\n**Author Contributions:** Conceptualization: Kim HJ, Jung Y; Data curation: Kim HJ, Baek J, Lee JR; Formal analysis: Kim HJ, Roh JY, Jung Y; Supervision: Roh JY, Jung Y; Writing - original draft: Kim HJ; Writing - review & editing: Roh JY, Jung Y.\n\nAD\n\n:   atopic dermatitis\n\nFLG\n\n:   filaggrin\n\nIFN\n\n:   interferon\n\nLOR\n\n:   loricrin\n\nTh\n\n:   T helper\n\nTNF\n\n:   tumor necrosis factor\n\nTSLP\n\n:   thymic stromal lymphopoietin\n\n[^1]: ^\u2020^These authors contributed equally to this work.\n\n[^2]: ^\u2021^Present address: Lee Dermatology Clinic, Seoul, Korea.\n"}
+{"text": "Background {#Sec1}\n==========\n\nWorldwide prevalence of obesity almost doubled between 1980 and 2008 \\[[@CR1]\\] and a similar trend has been observed in Switzerland \\[[@CR2]\\]. Several socio-demographic and behavioural factors have been shown to influence weight gain. A consistent positive association between marital status \\[[@CR3]\\], occupational position \\[[@CR4]\\], low educational level, economic difficulties \\[[@CR5]\\] and weight gain has been reported. Still, the impact of SES on weight gain might differ according to gender \\[[@CR6]\\] or to the country's level of socioeconomic development -- while in high income countries a high socioeconomic status (SES) is generally related to a lower prevalence of obesity, the opposite association is found in low income countries \\[[@CR7]\\].\n\nIn Switzerland, several cross-sectional studies have shown an inverse association between obesity and socio-demographic and behavioural factors \\[[@CR8],[@CR9]\\], but whether socio-demographic and behavioural factors have an impact on weight gain has never been investigated prospectively. Indeed, one of the main objectives of the Swiss national programme on healthy eating and physical activity (PNAAP) is achieving a health weight \\[[@CR10]\\] and such data are important for adequately designing health promotion policies and to evaluate their impact in the target population.\n\nThus, we aimed to assess the socio-demographic and behavioural determinants of weight gain, using prospective data from the Swiss population-based CoLaus study.\n\nMethods {#Sec2}\n=======\n\nThe Cohorte Lausannoise (CoLaus) study {#Sec3}\n--------------------------------------\n\nThe CoLaus study is a population-based study assessing the clinical, biological and genetic determinants of cardiovascular disease in the city of Lausanne, Switzerland \\[[@CR11]\\]. The initial recruitment took place between June 2003 and May 2006 and enrolled 6,733 participants (3,544 women) aged 35--75 years; participation rate was 41%.\n\nFollow-up was conducted between April 2009 and September 2012 and included all participants of the baseline study willing to participate. At follow-up, participants attended a single visit which included, as in the baseline assessment, an interview, a physical exam, and blood and urine collections in the fasting state. Average follow-up time was 5.5\u00a0years.\n\nSocio-economic data {#Sec4}\n-------------------\n\nEducational level was categorized as primary, secondary school, apprenticeship and university. Nationality was defined by the country of birth and categorized into Swiss and the most frequent nationalities in the canton (French, Italian, Portuguese and Spanish) and other. Marital status was categorized into living in couple (married and other relationship) or living alone (single, divorced or widowed).\n\nReceiving social help was assessed with the question: \"Do you receive social help?\". Because all individuals residing in Switzerland receive financial compensation when they retire, the response to this variable is not informative beyond the retirement age. Therefore, men older than 65\u00a0years and women older than 63 or 64\u00a0years were not considered as receiving social help (N\u2009=\u2009638). In Switzerland, social help is provided as financial support to people with disabilities or whose income is insufficient to support themselves or their family, and can thus be considered as an indicator of financial adversity.\n\nClinical and anthropometric data {#Sec5}\n--------------------------------\n\nSmoking status was defined as never, former and current. Physical activity was self-reported and participants were considered as physically active if they reported practicing leisure time physical activity at least twice per week. Body weight, height and waist circumference (WC) were measured using standard procedures \\[[@CR11]\\]. The same procedure was used in the baseline and follow-up examinations. BMI was defined as weight(kg)/height(m)^2^. Normal BMI was defined as BMI\u2009\\<\u200925\u00a0kg/m^2^, overweight as 25\u2264BMI\\<30 kg/m^2^ and obesity as BMI\u226530 kg/m^2^. Abdominal obesity was defined as WC\u2265102 cm for men and WC\u2009\u2265\u200988\u00a0cm for women.\n\nWeight gain was considered in two ways. First, as a continuous variable representing the rate of weight gain, computed as the difference between the follow-up and baseline weight divided by the number of years of the follow-up (Kg/year). Second, as having or not having gained 5\u2009Kg or more during the follow-up; the 5\u2009Kg cut-off was chosen according to the recommendations of the World Health Organization that weight gain in adulthood should not exceed 5\u2009Kg over the entire adult life \\[[@CR12]\\].\n\nEthics {#Sec6}\n------\n\nThe study was approved by the Institutional Ethics Committee of the University of Lausanne and all participants provided written informed consent.\n\nStatistical analysis {#Sec7}\n--------------------\n\nAnalyses were conducted after excluding participants without BMI at baseline or follow-up, without SES data at baseline. As illness could lead to involuntary weight loss, participants who reported involuntary weight loss during the last 12\u00a0months at follow-up were also excluded.\n\nStatistical analyses were performed using Stata version 12.0 for Windows (Stata Corp, College Station, Texas, USA). Descriptive results were expressed as number of participants (percentage) or as mean\u2009\u00b1\u2009standard deviation. Bivariate analyses were performed using chi-square or Fisher\\'s exact test for qualitative variables and one-way analysis of variance (ANOVA) for quantitative variables. Multivariable analyses were performed using ANOVA or logistic regression and results were expressed as multivariable adjusted mean\u2009\u00b1\u2009standard error for ANOVA or as Odds ratio (OR) and 95% confidence interval (CI) for logistic regression. In multivariable analyses, several models were considered -- adjusted only for age and gender and adjusted for all the variables significantly associated with weight gain. As BMI and waist were correlated, two separate multivariable models were tested (one with BMI and another with waist). Statistical significance was considered for p\u2009\\<\u20090.05.\n\nResults {#Sec8}\n=======\n\nCharacteristics of the participants {#Sec9}\n-----------------------------------\n\nOf the initial 5,064 participants with follow-up data, 67 (1.3%) were excluded because of missing values on BMI at baseline or follow-up and 528 (10.4%) because of involuntary weight loss during the last 12\u00a0months. The socio-demographic and anthropometric characteristics of the remaining 4,469 participants are summarized in Table\u00a0[1](#Tab1){ref-type=\"table\"}.Table 1**Sample characteristics (N\u2009=\u20094,469) at baseline and factors associated with weight gain, excluding participants reporting involuntary weight lossAll participantsWeight gain (kg/year)Gain\u22655 kg (%)N4,469868**Age groupN (%)Mean\u2009\u00b1\u2009SDN (%)\u2003\\[35-45\\]1 390 (31.1)0.46\u2009\u00b1\u20090.92334 (24.0)\u2003\\[45-55\\]1 344 (30.1)0.33\u2009\u00b1\u20090.88272 (20.2)\u2003\\[55-65\\]1 178 (26.4)0.21\u2009\u00b1\u20090.86202 (17.2)\u2003\\[65--75\\]557 (12.5)0.06\u2009\u00b1\u20090.7460 (10.8)\u2003p-value comparing groups\\<0.001\\<0.001Gender\u2003Women2 369 (53.0)0.31\u2009\u00b1\u20090.92461 (19.5)\u2003Men2 100 (47.0)0.31\u2009\u00b1\u20090.84407 (19.4)\u2003p-value comparing groups0.900.95Born in Switzerland\u2003No1 660 (37.1)0.34\u2009\u00b1\u20090.85340 (20.5)\u2003Yes2 809 (62.9)0.29\u2009\u00b1\u20090.90528 (18.8)\u2003p-value comparing groups\\<0.050.17Nationality\u2003Swiss2 809 (62.9)0.29\u2009\u00b1\u20090.90528 (18.8)\u2003French285 (6.4)0.35\u2009\u00b1\u20090.7754 (19.0)\u2003Italian238 (5.3)0.41\u2009\u00b1\u20090.8756 (23.5)\u2003Portuguese226 (5.1)0.24\u2009\u00b1\u20090.9136 (15.9)\u2003Spanish157 (3.5)0.36\u2009\u00b1\u20090.8034 (21.7)\u2003Other754 (16.9)0.35\u2009\u00b1\u20090.86160 (21.2)\u2003p-value comparing groups0.120.21Receiving social help\u2003No3 930 (87.9)0.30\u2009\u00b1\u20090.86728 (18.5)\u2003Yes539 (12.1)0.37\u2009\u00b1\u20091.00140 (26.0)\u2003p-value comparing groups0.06\\<0.001Marital status\u2003Living alone1 414 (31.7)0.35\u2009\u00b1\u20090.96316 (22.4)\u2003Living in couple3 052 (68.3)0.29\u2009\u00b1\u20090.84552 (18.1)\u2003p-value comparing groups\\<0.05\\<0.001Educational level\u2003Primary755 (16.9)0.30\u2009\u00b1\u20090.98150 (19.9)\u2003Apprenticeship1 579 (35.3)0.29\u2009\u00b1\u20090.92324 (20.5)\u2003Secondary school1 158 (25.9)0.31\u2009\u00b1\u20090.86217 (18.7)\u2003University977 (21.9)0.32\u2009\u00b1\u20090.76177 (18.1)\u2003p-value comparing groups0.840.44Smoking status\u2003Never1 848 (41.4)0.29\u2009\u00b1\u20090.85312 (16.9)\u2003Former1 513 (33.9)0.26\u2009\u00b1\u20090.84274 (18.1)\u2003Current1 108 (24.8)0.41\u2009\u00b1\u20090.97282 (25.5)\u2003p-value comparing groups\\<0.001\\<0.001Physical activity\u2003No2 018 (45.2)0.35\u2009\u00b1\u20090.95457 (22.7)\u2003Yes2 451 (54.8)0.27\u2009\u00b1\u20090.82411 (16.8)\u2003p-value comparing groups0.005\\<0.001Body mass index categories\u2003Normal2 192 (49.1)0.35\u2009\u00b1\u20090.70366 (16.7)\u2003Overweight1 651 (36.9)0.29\u2009\u00b1\u20090.85343 (20.8)\u2003Obese626 (14.0)0.22\u2009\u00b1\u20091.37159 (25.4)\u2003p-value comparing groups0.002\\<0.001Abdominal obesity\u2003No3 220 (72.1)0.34\u2009\u00b1\u20090.74583 (18.1)\u2003Yes1 249 (28.0)0.22\u2009\u00b1\u20091.16285 (22.8)\u2003p-value comparing groups\\<0.001\\<0.001Results are expressed as average\u2009\u00b1\u2009standard deviation or as number of participants (percentage). Statistical analysis by chi-square or analysis of variance.\n\nDeterminants associated with weight gain {#Sec10}\n----------------------------------------\n\nBivariate analysis of the baseline factors associated with rate of weight gain or weight gain \u22655\u2009Kg are summarized in Table\u00a0[1](#Tab1){ref-type=\"table\"}. Rate of weight gain decreased with age, and was lower among participants born in Switzerland, living in couple or physically active. Obese (BMI or abdominal) participants had a lower rate of weight gain, while current smokers had a higher rate. Receiving social help tended to be associated with weight gain (p\u2009=\u20090.06). The factors associated with a weight gain \u22655\u2009Kg were similar, although the association with being born in Switzerland was no longer significant, while receiving social help became significant.\n\nResults of multivariable analyses adjusting for age and gender (model 1) and for the main confounders (model 2) are summarized in Table\u00a0[2](#Tab2){ref-type=\"table\"}. Weight gain rate was negatively associated with age, living in couple and being physically active, and positively associated with current smoking. Weight gain rate was also associated with nationality (other than Portuguese), while no significant relationship was found with being Swiss born or BMI status. Multivariable logistic regression on gaining \u22655\u2009Kg provided similar results, i.e. that older vs. younger participants and those living in couple vs. alone had, respectively, a 62% (95% CI: 0.28-0.51) and a 27% (95% CI: 0.62-0.86) decreased risk of weight gain, while current vs. never smokers had a 63% increased risk (95% CI: 1.35-1.95). Participants receiving social help had a 43% higher likelihood (95% CI: 1.16-1.77) of gaining \u22655\u00a0kg.Table 2**Multivariable analysis of the factors associated with weight gain (N\u2009=\u20094,469), excluding participants reporting involuntary weight loss at follow-upWeight gain (kg/year)Gain \u22655\u00a0kgModel 1Model 2Model 2Model 1Model 2Model 2**Age group\u2003\\[35-45\\]0.46\u2009\u00b1\u20090.020.45\u2009\u00b1\u20090.020.47\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003\\[45-55\\]0.33\u2009\u00b1\u20090.020.33\u2009\u00b1\u20090.020.33\u2009\u00b1\u20090.02*0.80 (0.67 - 0.96)0.76 (0.63 - 0.92)0.74 (0.62 - 0.90)*\u2003\\[55-65\\]0.21\u2009\u00b1\u20090.030.22\u2009\u00b1\u20090.030.21\u2009\u00b1\u20090.03*0.65 (0.54 - 0.79)0.60 (0.49 - 0.73)0.56 (0.46 - 0.69)*\u2003\\[65--75\\]0.06\u2009\u00b1\u20090.040.08\u2009\u00b1\u20090.040.08\u2009\u00b1\u20090.04*0.38 (0.28 - 0.51)0.37 (0.27 - 0.50)0.35 (0.25 - 0.47)*p-value between groups\\<0.001^a^\\<0.001\\<0.001\\<0.001^a,d^\\<0.001^d^\\<0.001^d^Gender\u2003Women0.31\u2009\u00b1\u20090.020.31\u2009\u00b1\u20090.020.31\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003Men0.30\u2009\u00b1\u20090.020.31\u2009\u00b1\u20090.020.30\u2009\u00b1\u20090.020.97 (0.84 - 1.13)0.93 (0.79 - 1.08)1.02 (0.87 - 1.19)p-value between groups0.55^b^0.880.730.72^b^0.340.80Born in Switzerland\u2003No0.32\u2009\u00b1\u20090.020.32\u2009\u00b1\u20090.02-1 (ref.)1 (ref.)-\u2003Yes0.30\u2009\u00b1\u20090.020.30\u2009\u00b1\u20090.02-0.97 (0.83 - 1.13)1.04 (0.89 - 1.22)-p-value between groups0.470.52-0.660.61Nationality\u2003Swiss0.30\u2009\u00b1\u20090.02-0.30\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003French0.34\u2009\u00b1\u20090.05-0.33\u2009\u00b1\u20090.050.96 (0.70 - 1.32)-0.98 (0.71 - 1.34)\u2003Italian0.45\u2009\u00b1\u20090.06-0.45\u2009\u00b1\u20090.06*1.41 (1.02 - 1.93)*-1.36 (0.98 - 1.88)\u2003Portuguese0.13\u2009\u00b1\u20090.06-0.12\u2009\u00b1\u20090.06*0.65 (0.45 - 0.95)*-*0.60 (0.41 - 0.88)*\u2003Spanish0.35\u2009\u00b1\u20090.07-0.35\u2009\u00b1\u20090.071.12 (0.75 - 1.66)-1.10 (0.74 - 1.65)\u2003Other0.32\u2009\u00b1\u20090.03-0.31\u2009\u00b1\u20090.031.07 (0.88 - 1.31)-1.01 (0.82 - 1.25)p-value between groups0.006-0.005\\-\\--Receiving social help\u2003No0.30\u2009\u00b1\u20090.010.30\u2009\u00b1\u20090.010.30\u2009\u00b1\u20090.011 (ref.)1 (ref.)1 (ref.)\u2003Yes0.34\u2009\u00b1\u20090.040.33\u2009\u00b1\u20090.040.34\u2009\u00b1\u20090.04*1.43 (1.16 - 1.77)1.28 (1.03 - 1.58)1.32 (1.07 - 1.64)*p-value between groups0.320.490.400.0010.030.02Marital status\u2003Living alone0.36\u2009\u00b1\u20090.020.36\u2009\u00b1\u20090.020.35\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003Living in couple0.28\u2009\u00b1\u20090.020.29\u2009\u00b1\u20090.020.29\u2009\u00b1\u20090.02*0.73 (0.62 - 0.86)0.76 (0.65 - 0.90)0.77 (0.66 - 0.91)*p-value between groups0.007\\<0.05\\<0.05\\<0.0010.0010.002\u2003Educational level\u2003Primary0.32\u2009\u00b1\u20090.031 (ref.)1 (ref.)1 (ref.)\u2003Apprenticeship0.31\u2009\u00b1\u20090.02\\--1.05 (0.84 - 1.30)\\--\u2003Secondary school0.30\u2009\u00b1\u20090.03\\--0.88 (0.70 - 1.11)\\--\u2003University0.29\u2009\u00b1\u20090.03\\--0.81 (0.63 - 1.03)\\--p-value between groups0.95\\--0.02^d^Smoking status\u2003Never0.29\u2009\u00b1\u20090.020.29\u2009\u00b1\u20090.020.29\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003Former0.28\u2009\u00b1\u20090.020.28\u2009\u00b1\u20090.020.28\u2009\u00b1\u20090.021.14 (0.95 - 1.36)1.14 (0.95 - 1.36)1.13 (0.94 - 1.36)\u2003Current0.39\u2009\u00b1\u20090.030.37\u2009\u00b1\u20090.030.37\u2009\u00b1\u20090.03*1.63 (1.35 - 1.95)1.59 (1.32 - 1.91)1.56 (1.29 - 1.88)*p-value between groups0.003\\<0.05\\<0.05\\<0.001^d^\\<0.001^d^\\<0.001^d^Physical activity\u2003No0.34\u2009\u00b1\u20090.020.34\u2009\u00b1\u20090.020.34\u2009\u00b1\u20090.021 (ref.)1 (ref.)1 (ref.)\u2003Yes0.28\u2009\u00b1\u20090.020.28\u2009\u00b1\u20090.020.28\u2009\u00b1\u20090.02*0.72 (0.62 - 0.83)0.79 (0.68 - 0.92)0.77 (0.66 - 0.90)*p-value between groups\\<0.05\\<0.05\\<0.05\\<0.0010.0030.001Body mass index categories-\u2003Normal0.33\u2009\u00b1\u20090.020.33\u2009\u00b1\u20090.02-1 (ref.)1 (ref.)-\u2003Overweight0.30\u2009\u00b1\u20090.020.30\u2009\u00b1\u20090.02-*1.46 (1.23 - 1.73)1.43 (1.21 - 1.70)*-\u2003Obese0.25\u2009\u00b1\u20090.030.24\u2009\u00b1\u20090.04-*1.95 (1.57 - 2.43)1.89 (1.51 - 2.37)*-p-value between groups0.130.08-\\<0.001^d^\\<0.001^d^Abdominal obesity\u2003No0.32\u2009\u00b1\u20090.02-0.32\u2009\u00b1\u20090.021 (ref.)-1 (ref.)\u2003Yes0.27\u2009\u00b1\u20090.03-0.26\u2009\u00b1\u20090.03*1.58 (1.34 - 1.87)*-*1.52 (1.28 - 1.80)*p-value between groups0.08-\\<0.05\\<0.001\\<0.001OR: Odds Ratio; BMI: Body Mass Index. Results are expressed as multivariable adjusted mean\u2009\u00b1\u2009standard error of the mean (sem) or as odds-ratio (OR) and (95% CI). Statistical analysis by analysis of variance or logistic regression. Model 1, adjusting for age and gender, except ^a^adjusted for gender only; ^b^adjusted for age only; Model 2, adjusted for all the variables in the model (indicated in the column); ^d^p-value of the test for trend; \u2212, not included in the model. Statistically significant (p\u2009\\<\u20090.05) ORs are indicated in italic.\n\nRepeating the same analyses on the whole sample (i.e. including participants who were initially excluded from the analysis) led to similar findings, except that the difference in rate of weight gain between physically active and non-active participants was no longer significant, and that the increased likelihood of gaining \u22655\u00a0kg was borderline significant among participants receiving social help (Additional file [1](#MOESM1){ref-type=\"media\"}: Table S1).\n\nDiscussion {#Sec11}\n==========\n\nTo our knowledge, this is the first study that assessed the socio-demographic determinants of weight gain in the general Swiss adult population. Our results suggest that weight gain is negatively associated with age, Portuguese nationality, living in couple and physical activity, and positively associated with current smoking, receiving social help and obesity. Conversely, no association between weight gain and educational level was found.\n\nThe fact that younger people tend to gain more weight is in accordance with other studies \\[[@CR13]\\]. This can probably be explained by the ceiling phenomenon of weight gain through life, which means that because older people are heavier, they are less likely to gain weight. No difference in weight gain was found between normal weight, overweight and obese participants; conversely, obese participants had an increased likelihood of gaining \u22655\u00a0kg. A possible explanation is that besides the substantial group of obese patients with weight gain \u22655\u00a0kg, there was also a large group of obese patients with weight loss, most likely due to medical reasons. Indeed, the prevalence of obese subjects with weight loss (40%) was higher than in overweight (36%) or normal weight (29%) participants (\\<0.001). Our results thus suggest that many obese participants tend to lose weight, but that this trend is overcompensated by a significant fraction of obese participants who gained more than 5\u00a0kg during the study period. Also, many normal weight or overweight participants tend to gain weight, but less than 5\u00a0kg (Figure\u00a0[1](#Fig1){ref-type=\"fig\"}).Figure 1**Proportion of normal weight, overweight and obese participants who had a weight gain \u22655\u00a0kg, 0 \u2264 . \\<5\u00a0kg and a weight loss between baseline and follow-up (N\u2009=\u20094469).**\n\nBeing married (and particularly the marital transition to being married) has been associated with an increase in body weight in most \\[[@CR3],[@CR14],[@CR15]\\] but not all \\[[@CR16]\\] studies. In this study, living in couple was strongly protective against weight gain, and adjusting for other family covariates such as having children did not change the results (adjusted mean\u2009\u00b1\u2009sem: 0.36\u2009\u00b1\u20090.02 vs. 0.29\u2009\u00b1\u20090.02\u00a0kg/year for living alone vs. living in couple, p\u2009\\<\u20090.05). A possible explanation is the positive effect of living in couple on health-related behaviours \\[[@CR17],[@CR18]\\], but further studies are needed to better assess this point.\n\nSmoking at baseline was positively associated with weight gain, and this association persisted after multivariable adjustment. A possible explanation is that some smokers quit during follow-up, which led to increased weight \\[[@CR19]\\], although this statement has been challenged \\[[@CR20]\\]. Other explanations include a less healthy lifestyle of smokers (i.e. less healthy eating and less physical activity) and also the positive association between number of cigarettes smoked and central fat accumulation \\[[@CR21]\\] and a J- or U-shaped association has also been found between waist circumference and visceral fat area with total lifetime smoking amount \\[[@CR22]\\]. Hence, smoking should not be considered as a method for weight maintenance, and the beneficial effects of smoking cessation on health largely overcome the effects of post-cessation weight gain \\[[@CR23]\\].\n\nReceiving social help was associated with an increased likelihood of gaining \u22655\u00a0kg over the follow up. In Switzerland, social help is provided as financial support to people with disabilities or whose income is insufficient to support themselves or their family, and can thus be considered as an indicator of financial adversity. The association between receiving social help and weight gain was only partially attenuated after multivariable adjustment. Similar to other studies \\[[@CR5]\\], our results suggest that even in a wealthy country like Switzerland, financial difficulties are positively associated with weight gain, and that this association is independent from educational level. One possible explanation is that financial difficulties might prompt a decrease in diet quality \\[[@CR24]\\], but this aspect requires to be further investigated.\n\nLow educational level has been shown to be associated with increased weight gain \\[[@CR5],[@CR25],[@CR26]\\]. In this study, no significant association between education and weight gain was found, although participants with university education tended to be at lower risk of weight gain. A possible explanation might be a low statistical power due to a small sample size. Also, given the large educational differences in obesity at baseline \\[[@CR9]\\], ceiling effects may be at play, with people in the lowest educational category already having reached a plateau in obesity prevalence.\n\nStrengths and limitations {#Sec12}\n-------------------------\n\nThe main strengths of this study are the use of prospective data and of objective anthropometric measurements. This study also has limitations. First, 25% of participants at baseline were not followed-up and were thus not included in this analysis. Participants who accepted to be followed were significantly younger, more educated, more frequently born in Switzerland, received social help less frequently, were less frequently smokers and more frequently active than participants who refused follow-up (Additional file [2](#MOESM2){ref-type=\"media\"}: Table S2). Second, no information was available regarding slimming diets at baseline, so it was not possible to adjust for this covariate. Third, a considerable number of participants were foreigners, and education categorization might differ according to the educational system. However, to reduce this issue, we used broad categorizations of education, and almost all foreigners came from European countries, where the educational system is quite comparable. Finally, changes in some sociodemographic or behavioural factors during follow-up might have influenced weight gain. Still, taking into account the multiple possibilities (for instance, physical activity would be split in four groups depending on baseline and follow-up status) would considerably complicate the model and increase the risk of small sized groups, leading to nonsignificant associations due to large confidence intervals of the estimators. Thus, and also considering the relative short follow-up period, we chose to apply a classic analytical method, taking into account only the baseline data.\n\nPolicy implications {#Sec13}\n-------------------\n\nOur results are important for public health professionals and policy makers for several reasons. First, being physically active was negatively associated with weight gain, supporting the importance of promoting physical activity, also through environments favouring the practice of physical activity \\[[@CR27],[@CR28]\\]. Second, participants with financial difficulties had a higher tendency to gain weight over the follow-up, probably due to their intake of caloric-dense, less expensive foods \\[[@CR29]-[@CR31]\\]. Importantly, several randomized controlled trials have shown that education alone does not impact the purchase of healthy foods, and that cost reduction and/or promotions are needed to increase fruit and vegetable intake \\[[@CR32]-[@CR34]\\]. Thus, efforts should be made in the promotion of healthy eating, namely by decreasing the costs of healthy foods rather than just implementing food education campaigns. Finally, the fact that current smokers have an increased risk of gaining over 5\u00a0kg could be used as an additional argument for prompting smoking cessation.\n\nConclusions {#Sec14}\n===========\n\nIn Switzerland, financial difficulties and current smoking are positively associated with weight gain and living in couple, being older or physically active are negatively associated with weight gain.\n\nAdditional files {#Sec15}\n================\n\nAdditional file 1: Table S1.Multivariate analysis of the factors associated with weight gain (N = 4,997), including participants reporting involuntary weight loss.Additional file 2: Table S2.Comparison of the baseline characteristics between participants who refused (N=1573) and who accepted (N = 4968) follow-up.\n\n**Competing interests**\n\nThe authors declare that they have no competing interest.\n\n**Authors' contributions**\n\nFG made most of the statistical analyses and wrote most of the manuscript; PMV made part of the statistical analysis and wrote part of the manuscript; SS provided guidance for the statistical analyses and wrote part of the manuscript. PV and GW revised the manuscript for important intellectual content. PMV had full access to the data and is the guarantor of the study. All authors read and approved the final manuscript.\n\nThe CoLaus study was supported by grants from the Swiss National Science Foundation \\[grant no: 33CSCO-122661 and FN 33CSC0-139468\\]; GlaxoSmithKline and the Faculty of Biology and Medicine of Lausanne, Switzerland. FG was supported by a Scientific Mobility Grant from the Lisbon Faculty of Medicine/Calouste Gulbenkian Foundation. SS is supported by an Ambizione Grant (n\u00b0 PZ00P3_147998) from the Swiss National Science Foundation (SNSF).\n"}
+{"text": "![Credit to Stanford School of Medicine.](autopsy-03-01001-g01){#g01}\n\nRonald F. Dorfman (1923-2012), an Emeritus Professor of Pathology at the Stanford University School of Medicine, had significant input in advancing the study of diseases of hematopoietic cells and the lymph nodes, which included identifying a disease that has been named after him. Dr. Dorfman was one of the founders of the field of Hematopathology, and, with Dr. Costan Berard, founded the Society of Hematopathology in 1981, serving as its president from 1982 through 1984, and published a classification of non-Hodgkin lymphomas in 1974,[@B001] which was a variant of the Rappaport classification.[@B002]\n\nDr. Dorfman was born in Johannesburg, South Africa, on March 14, 1923 and studied medicine at the University of the Witwatersrand. From 1944 to 1946 his medical studies were suspended due to his military service during World War II. In 1948, he received his medical degree and completed medical and surgical internships and residencies at the Johannesburg General Hospital. After that, he went to England, to obtain his MD in Internal Medicine at the Royal Postgraduate Medical School and Hammersmith Hospital. At one point, he concluded that clinical medicine was not his passion and switched to pathology, initially in London and subsequently at the South African Institute for Medical Research (SAIMR) in Johannesburg. During his nine years of working there, his interest in the lymphoreticular system was stimulated by Dr. George Oettl\u00e9, who was the head of the South African National Cancer Institute at that time.\n\nDuring the early 1960s at the SAIMR, Dr. Dorfman came across two very unusual cases. The patients were African boys, aged 9 and 19, presenting enlarged lymph nodes in the neck, producing a \"bull-neck\" appearance. The oldest patient's disease had started 10 years earlier, and the disease followed a stable course. The histological appearance was so uncommon that the slides were submitted to several consultants in England, including Dr. Alistair Robb-Smith at Oxford University and Professor Donald Harrison at the Postgraduate Medical School of London. Varied opinions included \"an unusual reaction to an undetermined infectious agent,\" a cholesterol-type lipidosis, and even a form of Letterer-Siwe disease. In a fascinating article published recently by Dr. Dorfman in a periodical of the Adler Museum of Medicine, University of the Witwatersrand, Johannesburg, he described in detail the identification of this rare and intriguing disease.[@B003]\n\nIn 1963, due to his opposition to a society that established the superiority of white races over the black people of South Africa, he immigrated with his family to St. Louis, Missouri, USA, as Assistant Professor in the Surgical Pathology Division of the Department of Pathology at the Washington University School of Medicine, invited by Dr. Lauren V. Ackerman. While working there, one day a surgical pathology fellow named Juan Rosai, born in Poppi, Italy, came to his office with slides of two peculiar cases, which he identified among a series of cases he had been studying, including lymph node biopsies with the diagnosis of \"malignant reticuloendotheliosis\" -- a generic term embracing malignant disorders of the so-called reticuloendothelial system. The patients were a boy and a girl, both 7 months old presenting enlarged painless cervical lymph nodes. The two doctors studied the slides of these cases as well as those from Dr. Dorfman's two patients from Johannesburg, and immediately agreed that they were dealing with the same disease. Reviewing the literature, Dr. Rosai found two more cases with great clinical and histological similarities to theirs.[@B004]^,^[@B005] They prepared and submitted the first description of the entity they named \"sinus histiocytosis with massive lymphadenopathy (SHML): a newly recognized benign clinicopathologic entity\".[@B006]\n\nThe publication stimulated a great deal of interest and by the end of 1971 they had collected 30 new cases received in consultation, motivating their second paper with detailed analysis of their 34 cases, establishing the validity of SHML as a definitive clinicopathologic entity.[@B007] They described the histological features in early and advanced phases of the disease, and called attention to the most conspicuous feature, seen in greater or lesser degree in every case: the presence of apparently viable hematopoietic cells (mostly lymphocytes) within extremely large cells, with abundant, clear, granular or finely vacuolated cytoplasm and large, round, vesicular nuclei containing distinct nucleoli. Later, histochemical and immunophenotypic studies were able to support these large cells as having the features of histiocytes and activated monocytes/macrophages.[@B008] They differed from reactive sinus histiocytes by virtue of their strong staining for S100 protein. In addition, they differed from Langerhans cell histiocytes in that they failed to express CD1a antigen, and they do not have Birbeck\u00b4s granules at the electron microscopy. The other peculiar finding was the presence of dozens of lymphocytes located within cytoplasmic vacuoles, which most probably represented the ability of these cells to enter and leave the histiocytic cytoplasm without undergoing degenerative changes, a phenomenon known as emperipolesis (from the Greek: em = inside, peri = around, polesis = going about). They further emphasized the essentially benign nature of the disease and the capacity of this disorder to clinically simulate a malignant process.\n\nIt was in 1973 that a group of Spaniards published a paper using the eponym Rosai-Dorfman for the first time, and from that time this name was adopted by many others, particularly for those cases where the disease was exclusively extranodal.[@B009]\n\nDr. Rosai, who had been appointed Head of Anatomic Pathology at Yale University in New Haven, Connecticut, USA, created an archive of all the cases of SHML he was aware of, first at Minnesota and after at Yale. As an increasing number of cases were sent to Dr. Rosai and Dr. Dorfman in consultation over the ensuing years, they began to appreciate that SHML affected not only lymph nodes but also many different extranodal sites, at times without lymph node involvement. The most prominent extranodal manifestations were observed in the skin and soft tissues, upper respiratory tract, bone, genitourinary system, lower respiratory tract, and oral cavity, but nowadays the literature has shown that this disease can occur anywhere.\n\nIn 1990, the editor of *Seminars in Diagnostic Pathology* decided to devote an entire issue of the journal to a review of the clinical and pathological features of 423 cases of extranodal SHML.[@B010]\n\nNowadays, the cause of Rosai-Dorfman disease is still unknown. It is suspected that it may be of viral etiology and that minor upper respiratory tract infections may stimulate the characteristic interaction between lymphocytes, the unique histiocytes and plasma cells, resulting in massive lymphadenopathy and the widespread extranodal lesions of this disorder. There is no specific treatment for Rosai-Dorfman disease and the lesions are not responsive to antibiotics. Treatment is not necessary in most cases and enlarged lymph nodes usually resolve spontaneously; however, some patients may require surgery, radiation therapy and/or chemotherapy because of severe disease manifestations.\n\nA month before his death, Dr. Dorfman went to the Stanford Department of Pathology to give his opinion on an unusually difficult example of Rosai-Dorfman disease. I wondered what comments he would make about the challenging case of SHML presented in this issue of *Autopsy and Case Reports*. As in this case, a series of 14 deaths occurring in the SHML registry were described, and it was noted that many of these patients had had immunologic and/or hematologic abnormalities.[@B011]^,^[@B012]\n\n\"Ron Dorfman was considered a gifted diagnostician, teacher, and researcher, but those who knew him well were equally impressed by his warmth. He was considered a consummate gentleman and scholar.\" These were the words of Dr. Stacey E. Mills, Editor-in-Chief of *The American Journal of Surgical Pathology* at the time of Dr. Dorfman's death last June.\n\nBeyond the enigmatic disease of Rosai-Dorfman, we hope that this brief overview of Dr. Dorfman's life can inspire many young pathologists around the world to be good observers, to seek the answers to their questions, and to practice medicine with humanity and passion.\n\nThe author thanks to Prof. Juan Rosai for reviewing the text.\n\nZerbini MCN. Dr. Dorfman legacy to pathology: beyond Rosai-Dorfman disease \\[editorial\\]. Autopsy Case Rep \\[Internet\\]. 2013;3(1): 1-4. <http://dx.doi.org/10.4322/acr.2013.001>\n"}
+{"text": "I[NTRODUCTION]{.smallcaps} {#sec1-1}\n==========================\n\nInfectious diseases in type 2 diabetes can complicate diabetic ketoacidosis, derange hyperglycemia, or precipitate new onset diabetes. Among infections, tuberculosis being the most common infectious agent.\\[[@ref1]\\] Novel pandemic influenza virus H1N1\\[[@ref2]\\] strain needs to be kept in mind when dealing with new onset diabetes with co-existing infectious agent, which, if present, is associated with high mortality.\n\nA 63-year-old female, with no known chronic illness, was hospitalized in month of Aug 2010 with fatigue, breathlessness on exertion, productive cough with scanty sputum, and high-grade fever of 04 days duration. Quick emergency assessment had shown temp 101\u00b0F, tachycardia, tachypnea, p02: 90% at room air, and normotensive. Clinical chest examination was unremarkable. Further evaluation revealed NHO in both lung fields on chest X-ray, hyperglycemia 325 mg/dl, detected for first time. There was no ketonuria, no leukocytosis. Her symptoms and oxygen desaturation were out of proportion to clinical finding and chest X-ray findings. Patient was managed with insulin infusion and empirical broad-spectrum antibiotic coverage in ICU. As her condition worsened over next 12 hours, infection with novel influenza virus was thought and empirically started on oseltamivir after taking throat swab for H1N1 test and later, the sample was tested positive for H1N1 influenza by RT-PCR \\[Figures [1](#F1){ref-type=\"fig\"} and [2](#F2){ref-type=\"fig\"}\\]. Clinical course in the hospital was complicated by oxygen dependence requiring 10-12 ltr/hour of oxygen administration by nasal mask. Oseltamivir was continued for 02 weeks with escalation of antibiotic coverage for any methicillin-resistant *staphylococcus aureus* organisms. Her blood sugar was well controlled with continuous I.V. insulin infusion and gradually changed over to subcutaneous biphasic insulin. Appropriate prophylaxis against H1N1 was given to all nurses and paramedical staff and proper disposal of waste as per WHO guidelines. She made an uneventful recovery and was discharged on biphasic insulin @ 35 units/day with well-controlled blood sugar levels.\n\n![Chest X-ray showing pluffy opacities in both lung fields](IJEM-16-438-g001){#F1}\n\n![Computed tomography scan chest: Done at 48 hrs of admission, showing confluence of opacities with consolidation](IJEM-16-438-g002){#F2}\n\nThe first pandemic of 21^st^ century, witnessed with detection of the first case in Mexico 2009, became evident that human-to-human transmission is caused by novel influenza virus, which is a re-assortment of 4 different strains of influenza virus genes causing novel H1N1 2009 pandemic strain. Though it behaved like seasonal flu virus, the morbidity and mortality were more with risk factors like diabetes, pregnancy, cardiovascular status, obesity; chronic illness suffered more from this pandemic flu. Presently, we are in post-pandemic phase, and there are sporadic cases reported with mortality in high-risk cases \\[[Figure 3](#F3){ref-type=\"fig\"}\\]. The non-specific symptoms like fever, running nose, sore throat, body aches in flu make it difficult to identify and isolate cases initially \\[[Table 1](#T1){ref-type=\"table\"}\\]. WHO has listed diabetes as silent-epidemic, which is prevalent worldwide, taking toll of deaths in the age group of 35-65 years; 1 out of 10 deaths are attributed to diabetes, a ratio that increases to 1 out of 4 in certain vulnerable populations. People with diabetes and influenza are 3 times more likely to die from complications than without diabetes. Immunological research from the Hospital Clinico Universitario de Valladolid in Spain and the University Health Network found high levels of a molecule called interleukin 17 in the blood of severe H1N1 patients and low levels in patients with the mild form of the disease. Interleukin 17 can produce inflammation and autoimmune diseases.\\[[@ref3]\\] There were only few reported cases of H1N1 infection precipitating diabetes. Two cases from Iran\\[[@ref4]\\] with precipitation of diabetic ketoacidosis and H1N1 infection had fatal outcome. In a known diabetic, after H1N1 infection, hospitalization is known to triple and quadruple risk of ICU admission once hospitalized.\\[[@ref5]\\] Our case was complicated by precipitation of new onset hyperglycemia and occurrence of pneumonia with rapid decline in general condition. High flow oxygen was required to be delivered by facemask to maintain oxygen saturation. To best of our opinion, there were no reported cases of precipitation of diabetes with influenza A (H1N1) pdm09 virus with positive outcome from India.\n\n![Real time Influenza data in India. \\[Source: WHO Influenza Laboratory Surveillance Information by the Global Influenza Surveillance and Response System (GISRS). India, Generated on 14/09/2012 06: 14:47 UTC\\]](IJEM-16-438-g003){#F3}\n\n###### \n\nDistinguishing clinical features from common Cold and Flu\n\n![](IJEM-16-438-g004)\n\nThis case lays emphasis on clinical acumen in recognition of a co-existing infectious process in a new onset diabetes mellitus, H1N1 infection, and prompt institution of oseltamivir to retard the progression of disease and decrease the mortality associated with it. Vaccination offers better advantage for diabetes patients.\n\n**Source of Support:** Nil\n\n**Conflict of Interest:** None declared\n"}
+{"text": "Introduction {#Sec1}\n============\n\nClassically, the only experimental tools available to achieve targeted cell death were either slow-acting, non-specific pharmacological manipulation of cell types, or the 'brute force' killing of cells using either a cryoprobe or direct damage by a focused high-power laser, frequently operating in the ultra violet. Another widely used technique in zebrafish, which uses a joint genetic and pharmacological approach, is the nitroreductase (NTR) system. In the presence of the genetically-encoded bacterial NTR enzyme expressed specifically in cells of interest, the pro-drug metronidazole (MET) is catalytically converted to a cytotoxic DNA cross-linking agent that induces cell death by apoptosis^[@CR1],\\ [@CR2]^. This protocol has been used to great effect in the zebrafish\u00a0in systems such as kidney podocytes, the liver and the heart^[@CR2],\\ [@CR3]^. However, a distinct disadvantage to this approach is the time required for the pro-drug to act, often requiring from 24--72 hrs to induce apoptosis, and a washout of MET can be required to mitigate side-effects^[@CR2],\\ [@CR3]^.\n\nTo remedy this, advances in transgenic techniques now allow the expression of inducible phototoxic proteins in specific cells of interest, such that locally-applied low-power light of the correct wavelength induces apoptosis^[@CR4]^. The KillerRed protein is one such genetically-encoded photosensitizer where excessive quantities of reactive oxygen species (ROS) are generated in membrane-bound KillerRed-expressing cells upon illumination with 520--590\u2009nm light in the presence of oxygen, causing apoptosis within an hour^[@CR5],\\ [@CR6]^.\n\nThe KillerRed protein induces strong phototoxicity in cells due to the presence of a water channel between the intracellular solvent and the protein's methylene bridge, facilitating transport of molecular oxygen to the excited chromophore^[@CR7]^. Generation of radicals within the photosensitizing protein occurs through the interaction between the excited triplet state chromophore electrons and molecular oxygen. It has been shown that for KillerRed, type I energy transfer occurs and that superoxides are the most common cytotoxic agent^[@CR8]^. A fraction of the generated ROS bleaches the KillerRed protein and the remainder reacts with the cell in which the photosensitizer is expressed, ultimately resulting in oxidative damage to the plasma membrane and leading to cell death. This is a standard indicator of tissue damage in photodynamic therapy, a clinical light-activated procedure that uses localised oxidative damage through photochemical processes to selectively cause cell death^[@CR9],\\ [@CR10]^.\n\nThis optogenetic ablation tool has been successfully employed in a number of different cellular systems. Transient transfection of KillerRed into cultured 293\u2009T human kidney cells followed by illumination for 10\u2009min (535--575\u2009nm excitation filter, 5.8\u2009W/cm^2^) triggered a 40--60% cell death rate with no phototoxic effect observed in DSRed2 controls^[@CR5]^. When KillerRed was targeted to mitochondria in B16 melanoma cells, almost\u00a0all cells were killed after illumination for 45\u2009min (535--575\u2009nm excitation filter, 3.3\u2009W/cm^2^)^[@CR5]^. Mitochondrial-targeted KillerRed-expressing HeLa Kyoto cells were used as photosensitizing agents in a tumour xenograft^[@CR11]^, wherein irradiation with 593\u2009nm laser light increased tumour apoptosis from 6% to 14%.\n\nTransgenic zebrafish expressing membrane-bound KillerRed have also been developed for *in vivo* studies demonstrating a loss of cell viability correlated with impaired function. Optical illumination of KillerRed expressed in hindbrain rhombomeres 3 & 5 led to significant cell death within the hindbrain, whilst optogenetic ablation of heart cells expressing membrane-bound KillerRed caused reduced pumping efficiency and pericardial edema^[@CR4]^.\n\nAlthough this approach has been successfully used *in vivo*, the localisation, activation, quantification and control of this process through light illumination have not been optimised to maximise local cell optoablation whilst maintaining the health of the sample. We therefore developed and characterised a system to generate localised phototoxicity in a tightly-controlled spatiotemporal manner. This is based around a selective plane illumination microscope (SPIM) wherein the combined abilities for single cell targeting and high-speed treatment minimise collateral damage. We also used the inherent perpendicular light arms of the SPIM methodology to introduce a secondary activating beam through the imaging arm. We chose a Bessel intensity profile for this beam\u00a0to maximise the depth penetration and use the\u00a0\"self healing\" qualities of the beam as it penetrates through tissue.\n\nThe development and recent advances in the SPIM have revolutionised long-term time-lapse imaging with minimal perturbation to development in larval zebrafish. Good lateral and axial resolution, as well as significantly minimised phototoxic effects, are achieved by optically sectioning the sample using a light-sheet and collecting the resulting fluorescent signal perpendicularly. The light-sheet is typically formed using a cylindrical lens, restricting illumination, and hence fluorescence excitation, to the focal plane of the detection optics. This minimises out-of-focus fluorescence and photo-damage^[@CR12]^.\n\nLarval zebrafish are ideally suited for investigating fundamental developmental and repair processes *in vivo* since they are optically transparent and can be genetically manipulated with relative ease. The kidney is integral to normal blood pressure control and regulation; co-ordinated physiological systems, notably the renin-angiotensin system, act to maintain homeostasis. At 3 days post fertilisation (dpf) the only cells that express renin are located at the distal point of the anterior mesenteric artery (AMA)^[@CR13]^ at a tissue depth of approximately 200 \u03bcm. These perivascular cells tightly associated with endothelial cells prove a specific, yet anatomically challenging, system in which to demonstrate the ability to precisely control cell death in small subsets of cells with minimum off-target effects. A stable transgenic line with renin cell-specific expression of membrane-tethered KillerRed (*ren*:mem-KillerRed) was therefore established and used to optimise cell-specific *in vivo* optogenetic ablation.\n\nTo optimise targeted cell death, we compared the fluorescence loss between three different illumination methods. We used real-time dosimetry as a predictor of light administration efficiency and cellular apoptosis whilst minimising unwanted off-target effects, and confirmed cell death in photobleached KillerRed-expressing cells. The successful integration of a SPIM-independent excitation beam with a Bessel profile into the imaging path of the SPIM enabled unparalleled precision in targeting phototoxicity in a tightly controlled spatiotemporal manner at depth *in vivo*. We demonstrate that these specific targeting strategies significantly increase the speed of optoablation whilst simultaneously increasing the fish survival rate. This has not previously been possible for anatomically deep tissues such as the kidney, and provides new opportunities for biomedical research including developmental biology and pathophysiology.\n\nResults {#Sec2}\n=======\n\nTo evaluate the optical targeting efficiency of the SPIM light-sheet, KillerRed^+^ kidney cells in *ren*:mem-KillerRed;*kdrl*:GFP fish (schematic, Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}) were illuminated using either standard epifluorescence microscopy or our bespoke SPIM system (see Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). For epifluorescence experiments, we used a mCherry filter set that was well-matched to the excitation spectrum of KillerRed. A light power of 35\u2009mW was measured at the focal plane, corresponding to an approximate intensity of 4.5\u2009W/cm^2^, which was applied to the same region of the fish over the entire light treatment.Figure 1Evaluation of KillerRed optical targeting *in vivo* using epifluorescence and SPIM. (**a**) Schematic showing perivascular renin cell location in 3 dpf zebrafish. Mural renin cells are shown in red, arterial vessels in green, veins in blue, and the glomerulus (GL), nephron and cloaca in yellow. The dorsal aorta (DA) is distinguished from the anterior mesenteric artery (AMA) by using dark and light green to represent them, respectively. (**b**) *ren:*mem-KillerRed cells along the AMA budding off the DA in 3 dpf *ren:*mem-KillerRed;*kdrl*:GFP fish were illuminated using either an epifluorescence microscope (top panel, mCherry filter, 1hr), or the SPIM light-sheet at 3\u2009mW or 4\u2009mW (bottom panels, maximum intensity projections, 561\u2009nm laser, 1hr). Representative images from 0, 10, 20 and 30\u2009min illumination are shown. (**c**) Graphs showing the fluorescence intensity as a percentage of maximum intensity (arbitrary fluorescence, a.u.), plotted against time and averaged over n\u2009=\u20096 fish (with 95% confidence intervals, indicated by the filled dashed lines). Arrows and rotated times at foot of arrows indicate the 1/e value. Inset: percentage of fish surviving following illumination by epifluorescence or light-sheet microscopy for 1\u2009hr (n\u2009=\u200918). (**d**) Morphological changes of *ren:*mem-KillerRed cells were assessed using multiphoton microscopy (MPM). MPM images of 3 dpf *ren*:mem-KillerRed;*\u03b1SMA*:GFP renin-expressing cells were taken prior to- and post-treatment using the SPIM light-sheet. Images are presented as single depth slices. Red arrows indicate renin cells. (**e**) Endothelial cell structure prior to- and post-light treatment using the SPIM light-sheet was assessed using MPM in 3 dpf *ren*:mem-KillerRed;*kdrl*:GFP fish. Images are presented as maximum intensity projections. (**f**) Apoptosis of *ren:*mem-KillerRed cells was confirmed using an Apop-Tag *in-situ* TUNEL staining kit on 15\u2009\u03bcm sections of 3 dpf *ren:*mem-KillerRed;*kdrl*:GFP larvae illuminated with the SPIM and fixed immediately. White arrows designate apoptosis-positive post-treatment KillerRed^+^ cells. Images show a maximum intensity projection over 5\u2009\u03bcm using confocal light scanning microscopy. All scale bars represent 30\u2009\u03bcm. Figure 2Introduction of a Bessel beam into the imaging path of the SPIM set up allows rapid, localised targeting of KillerRed-expressing cells *in vivo*. (**a**) Schematic diagram of the imaging path of the SPIM system, for fluorescence imaging, with the integrated Bessel beam for cell targeting. OBJ1, OBJ2 are objective lenses; LED, light emitting diode; DC1, DC2, DC3, DC4, dichroic mirrors; L1, L2, L3, L4, L5, L6, lenses; SLM, spatial light modulator; BE1, beam expander; P1, polarizer; Ph1, pin hole; M1, mirror. Coloured lines shown in light blue, blue, green, red, grey and yellow represent light paths for the incident SPIM light-sheet, excited blue fluorescence, excited green fluorescence, excited red fluorescence, imaged transmission light and incident Bessel beam illumination, respectively. (**b**) KillerRed cells in 3 dpf *ren*:mem-KillerRed larvae (indicated by the red box) were located using low power, low frame-rate 488\u2009nm light-sheet excitation (top left panel). The 561\u2009nm Bessel beam was translated to the correct position and superimposed with brightfield illumination for a single frame (top right panel). Brightfield illumination was then turned off and a fluorescence emission image acquired every 10\u2009s using the Bessel illumination until the fluorescence intensity had decreased to the background level approximately 10\u2009min later (representative images shown in bottom panel). (**c**) Graphs showing the fluorescence intensity as a percentage of the maximum intensity (arbitrary fluorescence units, a.u.), plotted against time and averaged (with 95% confidence intervals, indicated by the filled dashed lines). Cells were targeted using the Bessel beam with incident powers of 1\u2009mW, 0.5\u2009mW or 0.25\u2009mW (as measured at the back aperture of OBJ2). The fluorescence intensity measured from acquired images and represented as a percentage of the maximum value, averaged over between n\u2009=\u200910 and n\u2009=\u200912 fish. This is represented individually for each power, and as a comparison between the three different targeting methods: epifluorescence illumination, SPIM light sheet illumination and Bessel beam illumination. Data has been truncated at 600\u2009s to clearly show the initial rapid decay when using Bessel illumination. (**d**) SPIM image showing subsets of KillerRed-expressing cells, which were targeted using the Bessel beam. A representative example is given where the correct location of cells was confirmed using the 488\u2009nm light-sheet, and the 561\u2009nm Bessel beam translated to the dorsal end of the AMA. Relative location within the fish is demonstrated by co-localising fluorescence signal with a brightfield image. Light treatment was performed, targeting a subset of cells (shown before 1^st^ Treatment in the top panel) along the AMA and the decrease in fluorescence intensity confirmed using the light-sheet (shown before 2^nd^ Treatment in the second panel). The remaining cells, which showed no decrease in fluorescence intensity after the first treatment, were then targeted for a second treatment. No evidence of KillerRed expression remained after the second light treatment (shown in the bottom Post-2^nd^ Treatment panel). All scale bars represent 30\u2009\u03bcm.\n\nThe decrease in KillerRed fluorescence during illumination with epifluorescent light was measured from images collected over time (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}, top panel), and compared to that of the targeted illumination performed using the SPIM light-sheet (\u03bb\u2009=\u2009561\u2009nm) at 3\u2009mW and 4\u2009mW (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}, bottom panels). This corresponds to an approximate intensity of 0.37\u2009kW/cm^2^ and 0.5\u2009kW/cm^2^ for the light-sheet used at 3\u2009mW and 4\u2009mW, respectively, though the plane was continually scanning across the volume of interest. The improvement in image quality when using the SPIM was evident, allowing imaging in two spectrally separate channels with good lateral and axial resolution. The cell boundaries could clearly be resolved and surrounding cells observed during photobleaching (Supplementary video\u00a0[1](#MOESM2){ref-type=\"media\"}). The decrease in fluorescence intensity was quantified and plotted using 95% confidence intervals error bars (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}). Photobleaching efficiency - the time for fluorescence emission to fall to 1/e of its initial value - improved significantly using light-sheet compared with epifluorescence illumination. On average, epifluorescence photobleaching fell to 1/e at 760\u2009+/\u2212\u2009250\u2009s compared to the 234\u2009+/\u2212\u200950\u2009s and 188\u2009+/\u2212\u200940\u2009s taken to reach this value when using 3\u2009mW and 4\u2009mW SPIM illumination, respectively. No significant power dependence on the rate of the light-sheet photobleaching effect was observed over the power ranges used. Importantly, the survival rate for fish illuminated using the light-sheet was 100% compared to only 10% (n\u2009=\u200918) of fish surviving the epifluorescence illumination protocol (Fig.\u00a0[1c](#Fig1){ref-type=\"fig\"}, inset).\n\nTo confirm that the reduction in fluorescence intensity was linked to cell death, cellular morphology was assessed using multiphoton microscopy (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}). Multiphoton images of KillerRed^+^ renin cells were acquired in 3 dpf *ren*:mem-KillerRed;*\u03b1SMA*:GFP fish before phototoxicity-inducing SPIM illumination (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}, PreAblation panels). The fish were transferred to the SPIM for KillerRed activation and cell optogenetic ablation, followed by 1 hr incubation at 28\u2009\u00b0C before multiphoton images were retaken (Fig.\u00a0[1d](#Fig1){ref-type=\"fig\"}, Post-Treatment panels). Alterations in cellular morphology and a decrease in *\u03b1SMA*:GFP signal were evident (red arrow), and the *ren*:mem-KillerRed signal was eliminated. Perivascular renin cells are tightly associated with the underlying endothelium, therefore a similar experiment was performed on 3 dpf *ren*:mem-KillerRed;*kdrl*:GFP fish to assess endothelial cell viability post-light treatment (Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"}). Endothelial cells and the AMA remained viable and patent post-illumination, confirming the specificity of the light-induced damage to the renin-expressing cells. Apoptotic cell death within renin cells along the AMA of 3 dpf *ren*:mem-KillerRed;*kdrl*:GFP fish after 1 hr of light treatment was confirmed using an anti-digoxigenin (anti-dig)-based Apop-tag fluorescence TUNEL staining (Fig.\u00a0[1f](#Fig1){ref-type=\"fig\"}). Anti-dig-rhodamine^+^ cells were present along the AMA (indicated with arrows), with little off-target staining present. Additional evidence is provided in Supplementary Figure\u00a0[1](#MOESM1){ref-type=\"media\"}; acridine orange is an intravital dye that stains all cells, but accumulates within early apoptotic cells to stain the nuclei bright green due to chromatin condensation^[@CR14]^. At the end of light-sheet treatment, bright signal accumulation can be seen within cells that were KillerRed^+^ prior to treatment (Supplementary Figure\u00a0[1a](#MOESM1){ref-type=\"media\"}). Furthermore, whole mount anti-Dig^+^ staining visualised with Nitro Blue Tetrazolium showed appropriate signal in the correct location (Supplementary Figure\u00a0[1b](#MOESM1){ref-type=\"media\"}).\n\nTo improve the targeting of KillerRed^+^ cells *in vivo*, a Bessel beam was integrated into the imaging light path of the SPIM (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}). In this configuration, a separate 488\u2009nm laser was available to form the light-sheet, which was used at low power to locate KillerRed^+^ cells without photobleaching and enabled accurate targeting of the 561\u2009nm Bessel beam to the region of interest (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}). Images were acquired every 10\u2009s until no KillerRed^+^ fluorescence was detected (Fig.\u00a0[2b](#Fig2){ref-type=\"fig\"}). The decrease in fluorescence intensity was measured over time following illumination with the Bessel beam at 0.25\u2009mW, 0.5\u2009mW and 1\u2009mW (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}), corresponding to approximately 1.5\u2009kW/cm^2^, 3\u2009kW/cm^2^ and 6\u2009kW/cm^2^ intensity at the sample, respectively. Plotting this against the decay curves of the epifluorescence- and SPIM-illuminated embryos shows that using the Bessel beam allows rapid and reproducible targeting between fish. Photobleaching efficiency showed no significant difference between incident laser powers; the time taken to reach the 1/e value was 81\u2009s, 65\u2009s and 88\u2009s for the 1\u2009mW, 0.5\u2009mW and 0.25\u2009mW incident laser powers of the Bessel beam, respectively (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). As with illumination using the light-sheet, the fluorescence intensity change was not dependent on laser power, with the fluorescence decrease curves falling within one standard error of each other.\n\nTargeting subsets of KillerRed^+^ cells along the AMA was successfully accomplished for the first time (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}), allowing targeted cell optogenetic ablation within vessels of approximately 50\u2009\u00b5m in length. When the Bessel beam was situated at the distal end of the AMA segment containing KillerRed^+^ cells (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}, 1^st^ Treatment panel), a small area was photobleached in under 10\u2009min. Comparison with the remaining fluorescence, targeted in the second treatment (Fig.\u00a0[2d](#Fig2){ref-type=\"fig\"}, 2^nd^ Treatment panel), strongly suggests that a single cell was targeted in the first treatment since no bleaching of the lower end of the AMA occurred. Again, total bleaching of the second region was performed in fewer than 10\u2009min. The bleaching of all the fluorescence signal was confirmed after completion of the illumination protocol (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}, Post-2^nd^ Treatment panel).\n\nFitting the averaged decay curves to exponential functions showed the different targeting characteristics of the three illumination methods (Fig.\u00a0[3a](#Fig3){ref-type=\"fig\"}); epifluorescence was slow and non-specific, SPIM light-sheet illumination was faster and allowed a volume of cells to be targeted and the Bessel beam targeted subsets of cells extremely rapidly. To quantify this, each curve was fitted with a two-term exponential, and the half-life of the fast term calculated, averaged for each group and compared between groups (Fig.\u00a0[3b](#Fig3){ref-type=\"fig\"}). Epifluorescence illumination provided the slowest decay, shown by the longest half-life of the three methods, 223\u2009s on average. It also showed the widest spread, indicating the lack of reproducibility of the targeting. SPIM light-sheet illumination tightened the reproducibility and significantly decreased the half-life of the decay terms to 130\u2009s for 3\u2009mW illumination and 110\u2009s for 4\u2009mW (p\u2009=\u20090.0023), the different powers of which were not statistically different. However, the Bessel beam illumination was by far the most reproducible and rapid targeting method. The average half-life of the fast term was 24\u2009s for 0.25\u2009mW illumination, 26\u2009s for 0.5\u2009mW illumination and 31\u2009s for 1\u2009mW illumination, which were not significantly different from each other. This was significantly faster than the light-sheet illumination (p\u2009=\u20090.0002) and epifluorescence illumination (p\u2009\\<\u20090.0001), and was highly reproducible.Figure 3Fitting of decay curves highlights distinction between targeting speed. (**a**) Averaged fluorescence intensity profiles for each group were fitted to two-term exponential functions. The calculated fast and slow decay curve half-lives were calculated and are show, along with the measured time for the curve to reach 1/e. The curve fitting precision is represented by the adjusted R2\u00a0value. (**b**) Individual fluorescence intensity decay curves were fitted to two-term exponential functions for each sample. KillerRed photobleaching follows this two-term decay curve extremely well, comprising a fast and a slow decay component. The half-life was calculated from the fast term of each fitted curve, and averaged for each group. Individual points are plotted, with error bar designating the 95% confidence intervals. \\*\\*p\u2009\\<\u20090.01, \\*\\*\\*\\*p\u2009\\<\u20090.0001.\u00a0\n\nDiscussion {#Sec3}\n==========\n\nThe application of our novel SPIM system allows the rapid generation of phototoxicity in a tightly controlled spatiotemporal manner. Through the use of the genetically-encoded photosensitizer fluorescent protein KillerRed, this phototoxicity response can be manipulated to cause targeted cell death. By incorporating a Bessel beam, we are able to induce rapid and highly-localised cell death, coupled with time-lapse imaging capabilities separating imaging illumination from cell targeting.\n\nUsing our innovative set up, we demonstrate two independent methodologies for targeted cell death with minimal off-target damage to the sample. The first involved continuous scanning of the light-sheet over the volume of KillerRed^+^ cells for 60\u2009min, providing excellent imaging quality and axial information whilst inducing targeted cell death. This method has the advantage that it can be achieved on a standard SPIM microscope without any modification to the optics. Whilst the power required to illuminate the deep-set renin cells using epifluorescence microscopy was such that the fish did not survive the illumination protocol, the use of the SPIM allowed specific and sensitive targeting. This was particularly highlighted by the fact that the underlying endothelium remained intact. Renin cells are perivascular-like and are intimately linked with the underlying endothelium through peg and socket junctions; if any off-target effects were to be seen, it would be expected to be in such closely associated cells. Evidence from Fig.\u00a0[1e](#Fig1){ref-type=\"fig\"} and Supplementary video\u00a0[1](#MOESM2){ref-type=\"media\"} show that this is not the case, indicating that the targeting is restricted to the KillerRed-expressing cells.\n\nThe second method involved the application of a\u00a0Bessel beam, in combination with a light-sheet, to locate regions of interest in order to target individual subsets of cells. A Bessel beam is a 'self-healing', non-diffracting focal line of light with a central maximum that is resistant to diffractive spreading^[@CR15]^. These distinctive properties of Bessel beams have been exploited with great success; for example, the long focal depth over large axial distances has been used for cellular photoporation^[@CR16]^, and the beam\u00a0has been scanned to create light sheets sufficiently thin to achieve isotropic 3D resolution^[@CR17]^.\n\nThe use of the Bessel beam not only improved the ability to target much more precisely than with the light-sheet, but also increased the speed of targeting by approximately an order of magnitude when compared to epfiluorescence illumination. By integrating the Bessel beam through the imaging arm of the SPIM and decoupling it from the light-sheet illumination, we were able to use a high NA objective (0.8, compared to the 0.3 NA of the illumination objective) to launch the Bessel beam, producing a highly localised targeting beam that is more precise and a set up that is more flexible.\n\nIn 3 dpf larvae, renin is localised to the AMA, which branches off the dorsal aorta^[@CR18]^. The distal end of the AMA is approximately 200\u2009\u03bcm from the dorsal surface of the larvae, and is therefore extremely difficult to image and target. The successful targeting of subsets of renin cells along the AMA demonstrates the power of using the Bessel beam in conjunction with genetically-encoded photosensitizing proteins. This is particularly apparent when comparing the targeted illumination of the SPIM and Bessel beam with 'bulk' epifluorescence illumination, where the untargeted high-power illumination required to illuminate the deep-seated renin cells led to a low survival rate. Using our powerful, targeted cell-death technique, we will be able to explore the functional roles of renin cells with respect\u00a0to the surrounding vasculature, blood pressure and kidney physiology. Our accurate targeting and quantification methodology leading to localised damage and high fish survival rates indicates that the method is suitable for other organs and cell types within the developing embryo as appropriate transgenic animals become available.\n\nTaking the results from the curve fitting together with the qualitative information gained, our results show that both methods are more effective at photobleaching, and less harmful to the fish, than epifluorescence illumination. It is interesting that all curves fitted more accurately to two-term rather than single-term exponential functions. There is evidence to suggest that\u00a0the two-term exponential decay of photosensitizing proteins may correspond to distinctive contributions of the oxygen concentration variation during light administration^[@CR9]^, which could be used to probe the oxygen dynamics during photodynamic activity.\n\nThe cellular localisation of KillerRed expression is important for effective phototoxicity generation. Mitochondrial-, histone H2A- and membrane-bound KillerRed, as in the present study, have been shown to effectively mediate rapid cell death, possibly due to direct oxidation of membrane lipids^[@CR5],\\ [@CR6],\\ [@CR19]^. Conversely, cytosol-bound KillerRed produces a weak phototoxic effect^[@CR5]^.\n\nImplicit dosimetry was used throughout the study, since the decrease in fluorescence intensity acts as a proxy measurement for ROS generation and hence cellular apoptotic response. Although a dependence on power was expected for the photobleaching characteristics, this was not seen when illuminating with the Bessel beam. A slight but non-significant decrease in the 1/e value was seen when illuminating with a 3\u2009mW compared to 4\u2009mW beam. This would suggest that we might be saturating the photosynthesizing photobleaching reaction, with either the local oxygen or photosensitizing protein acting as the limiting factor.\n\nIn conclusion, we have developed a method to generate phototoxicity *in vivo* in a rapid and precisely controlled spatiotemporal manner with minimal off-target perturbation to closely-associated endothelial cells. These advances will transform the landscape of targeted cell ablation, enabling the study of discrete cells (e.g. renin progenitor cells, pericytes and immune cells) in the development and pathophysiology of previously inaccessible tissues with the absence of any detectable off-target damage. The ability to modulate the level of induced damage in a dose-dependent manner within minutes has wide-ranging applications across models used within biomedical research, particularly when coupled with long term, low perturbation imaging capabilities.\n\nMethods {#Sec4}\n=======\n\nFish lines and husbandry {#Sec5}\n------------------------\n\nExperiments were approved by the University of Edinburgh Animal Welfare and Ethical Review body (AWERB) and conducted in accordance with the UK Home Office\u00a0Animals (Scientific Procedures) Act 1986. Fish (*Danio rerio*) were maintained at 28.5\u2009\u00b0C, as previously described^[@CR20]^. Embryos were staged according to Kimmel *et al*.^[@CR21]^ and anesthetised with 40\u2009\u00b5g/ml tricane methanesulfonate. The established *\u03b1SMA*:GFP^[@CR22]^ and *kdrl*:GFP^[@CR23]^ transgenic lines were used to label smooth muscle cells and endothelial cells, respectively.\n\nGeneration of Tg(ren:mem-KillerRed) fish {#Sec6}\n----------------------------------------\n\nTg(*ren*:mem-KillerRed) was created using Three-way Gateway cloning (Invitrogen) and the Tol2kit^[@CR24]^. Membrane bound KillerRed (Evrogen, Cat.\\# FP966) was driven by a previously characterised 6.46-kb DNA sequence^[@CR13]^ immediately upstream of the *ren* translational initiation site^[@CR18]^. An entry clone containing the *ren* promoter sequence was recombined with mem-KillerRed and a simian virus 40 polyA into the Tol2kit destination vector pDestTol2CG2 (containing tol2 ends and cardiac myosin light chain:GFP). Plasmid DNA was co-injected with transposase mRNA and a single founder fish was used to establish a stable line. The expression of *ren*:mem-KillerRed was comparable to that of *ren*:LifeAct-RFP^[@CR13]^.\n\nEpifluorescence Microscopy {#Sec7}\n--------------------------\n\nA Leica MZ16 F stereomicroscope with top lighting was used. The light source was a 100\u2009W high-intensity mercury burner lamp. Standard mCherry (578\u2009nm/46\u2009nm) and RFP (549\u2009nm/18\u2009nm) filters were used.\n\nEpifluorescence optoablation process: *ren:*mem-KillerRed larvae were oriented laterally in a petri dish on their left hand side to provide the best imaging angle for renin, in 0.5% agar covered with system water. The mCherry filter was applied, and fluorescence emission images were collected for monitoring purposes every 15\u2009s for 60\u2009min (100 ms exposure). Given a measured light power of 35\u2009mW at the focal plane, this corresponds to an approximate intensity of 4.5\u2009W/cm^2^.\n\nSelective Plane Illumination Microscope Apparatus {#Sec8}\n-------------------------------------------------\n\nAn in-house SPIM system was built, using a Vortran Versalase multiple wavelength system as the light source, consisting of 3 lasers (405\u2009nm, 488\u2009nm and 561\u2009nm). A single mode optical fibre coupled the laser light to the SPIM illumination path, which is built as previously described^[@CR25]^. The appropriate laser power was set using Stradus Versalase software, and a power meter (Thorlabs, PM100D) was used to verify collimated beam power, immediately prior to the light-sheet focussing using a cylindrical lens. The resulting beam was then focused onto the sample (10X 0.3NA Nikon CFI Fluor water dipping objective) to produce a light-sheet 400\u2009\u00b5m high and approximately 2\u2009\u00b5m thick at the beam waist. This corresponds to an approximate intensity of 0.37\u2009kW/cm^2^ and 0.5\u2009kW/cm^2^ for the light-sheet used at 3\u2009mW and 4\u2009mW, respectively. Larvae were embedded in 0.5% agar (in standard system water) inside an FEP tube (fluorinated ethylene propylene, Adtech), which is refractive-index-matched to water. The tube was then suspended between the imaging and illumination objectives, and orientated using a translation and rotation stage (Thorlabs). Imaging of the sample was through a 16X 0.8 NA Nikon CFI LWD Plan Fluor water dipping objective (N16LWD-PF). The resulting GFP and KillerRed emission signals were separated by a 490--550\u2009nm dichroic beamsplitter (T550lpxr-UF2, Chroma Technology) and passed through 525/39\u2009nm and 630/69\u2009nm emission filters respectively (MF525-39, MF630-60, Thor Labs) onto two QI-Click Mono CCD cameras (Q-Imaging Inc). Brightfield images were separated using a 700--850\u2009nm dichroic beamsplitter (T700spxr-UF2, Chroma Technology) and recorded on a Prosilica GS650 camera at 100\u2009Hz. The entire system was controlled through a custom interface written in the Objective C language.\n\nLight-sheet optoablation process: 3 dpf *ren:*mem-KillerRed;*kdrI:*GFP fish were embedded in 0.5% agar (in system water) and suspended in the FEP tubing, tail up, and oriented such that the light-sheet was directly incident upon the cerebellum, exiting via the heart. This is the optimal orientation for imaging renin cells. The 561\u2009nm light-sheet was set to illuminate continuously, and scan across the entire volume of AMA renin expressing cells every 10\u2009s for 1 hr. When used, the 488\u2009nm laser was only pulsed during the camera exposure, to minimise the light dose to the specimen.\n\nThe Bessel beam (a power-adjustable Vortran Versalase multiwavelength system used at 561\u2009nm) used as the activation laser was integrated into the SPIM through the imaging path of the system (Fig.\u00a0[2a](#Fig2){ref-type=\"fig\"}) by means of a 488/561\u2009nm dichroic beamsplitter (Di01-R488/561-25\u2009\u00d7\u200936, Semrock) placed behind the imaging objective. The beam of the activation laser was expanded (BE1) to fill the free aperture of a spatial light modulator (SLM, Holoeye LC-R 720 Spatial Light Modulator, reflective). The SLM was used to generate an axicon phase profile to produce a Bessel beam. The SLM surface was then re-imaged onto the back aperture of the imaging objective using a 4-f relay system (L1, L2). An additional phase gradient was generated using the SLM in order to direct the desired diffracted order along optical axis. An adjustable iris (Ph1) placed at the focus of the first relay lens (L1) blocked all other diffracted orders from reaching the imaging objective (OBJ2). The beam was aligned to the centre of the field-of-view of OBJ2, and small adjustments could be made by moving the mirror (M1) in order to target a specific region in the fish. The Bessel beam area within the imaging plane corresponded to intensities of approximately 1.5\u2009kW/cm^2^, 3\u2009kW/cm^2^ and 6\u2009kW/cm^2^ for the 0.25\u2009mW, 0.5\u2009mW and 1\u2009mW beams, respectively.\n\nBessel beam optoablation process: 3 dpf *ren:*mem-KillerRed fish were similarly embedded in 0.5% agar (in system water) and suspended in the FEP tubing, tail up, and optimally oriented as described for the light-sheet treatment. *ren*:mem-KillerRed^+^ cells were located using the 488\u2009nm light-sheet, which is at the lower end of the KillerRed excitation spectrum. The Bessel beam was then translated to the region of interest with the brightest point closest to the DA. The cells of interest were continuously exposed to the laser beam and an image was acquired every 10\u2009s for 10--20\u2009min, until all the fluorescence had reduced to the background level.\n\nMultiphoton Imaging {#Sec9}\n-------------------\n\nTwo-photon imaging was performed on a TriM Scope II 2-photon inverted microscope (LaVision BioTec, Germany). A pulsed Ti:Sapphire laser (Chameleon-Ultra II; Coherent, USA) was used to excite GFP at 930\u2009nm. The KillerRed protein was excited at 1100\u2009nm with an optical parametric oscillator laser (Chameleon Compact OPO, Coherent) pumped at 800\u2009nm by a separate Ti:Sapphire laser source. Depth stacks of the AMA and surrounding vasculature were acquired using a 40x water immersion objective (NA\u2009=\u20091.15, MRD77410 Apo LWD; Nikon, Japan) with a step-size of 0.37\u2009\u00b5m. The GFP and KillerRed emission signals were separated by a 590\u2009nm short-pass dichroic filter (Chroma Technology), and collected through 525/70\u2009nm and 620/60\u2009nm bandpass filters (Chroma Technology) respectively onto a pair of photomultiplier detectors (H7422 GaAsP, Hamamatsu Photonics, Japan).\n\nImage Analysis {#Sec10}\n--------------\n\nImage registration was performed using the Fiji StackReg plugin for rigid bodies^[@CR26]^. During epifluorescence and Bessel beam illumination, a single image was taken per timepoint. During SPIM illumination, depth stacks were acquired across the KillerRed-expressing volume at each timepoint, and maximum intensity projections were calculated from these raw image stacks by our custom-written acquisition software. A region of interest was selected around the KillerRed-expressing cells, and background fluorescence was measured in an equivalently-sized dark region in which there were no KillerRed-expressing cells. The average intensity of the background was subtracted from the region of interest, the maximum value determined and all subsequent values calculated as a percentage of that maximum value.\n\nCurve fitting and Statistics {#Sec11}\n----------------------------\n\nCurve fitting was performed using the curve fitting toolbox in MATLAB and the Statistics Toolbox (Release 2015b, The Mathworks, Inc., Natick, Massachusetts, USA). Double exponential fits were either used on every individual dataset and used to calculate the fast decay term half-life, or the average of each dataset within a single group compared. The parameters and R-squared value were calculated in each instance. Statistical analyses were performed with Graphpad Prism 5 (La Jolla, CA) and 95% confidence intervals are reported on each graph. A one-way ANOVA was performed on the fast decay curve half-life for each group, followed by a Tukey test for multiple comparisons.\n\nTUNEL staining {#Sec12}\n--------------\n\nAn ApopTag Red *In Situ* Apoptosis detection kit (Millipore, S7165) was used in conjunction with an anti-Dig-AP antibody. After optogenetic ablation procedures, *ren:*mem-KillerRed;*kdrl:*GFP fish were immediately fixed for 20\u2009min in 4% PFA, embedded in optimal cutting temperature (OCT) compound, cryofrozen in dry-ice-cooled isopentane and sectioned sagitally in 15 \u03bcm sections. Slides were kept at \u221220\u2009\u00b0C until required, then thawed to room temperature and the appropriate kidney-containing sections determined. Samples were rapidly re-fixed in 4% PFA (20\u2009min, room temperature), permeabilised with 0.1% Triton-X and washed three times for 5\u2009min in PBT (PBS, 0.1% tween20). Equilibrium buffer was added (45\u2009min, room temperature), then 16\u2009\u00b5l TdT enzyme and 30\u2009\u00b5l reaction buffer (90\u2009min, 37\u2009\u00b0C, kit) were applied to the slide. The reaction was stopped using stop buffer (20\u2009min, room temperature, kit), then incubated in blocking solution and an alkaline phosphatase-conjugated Anti-Digoxigenin (anti-Dig) (11093274910, Roche) antibody (2\u2009hours, room temperature) and washed four times for 15\u2009min. Sections were mounted with ProLong Gold anti-fade mountant.\n\nAcridine orange live staining {#Sec13}\n-----------------------------\n\n3 dpf *ren:*mem-KillerRed embryos were imaged incubated for 30\u2009min in the dark in system water containing 2\u2009\u03bcg/ml acridine orange (Sigma), followed by three 5-min washes in conditioned medium, and imaged live immediately afterwards. The light-sheet optoablation protocol was then followed, with the volume of KillerRed+ cells illuminated with the lightsheet at 4\u2009mW every 10\u2009s for 1 hr. Acridine orange signal was then imaged using the 488\u2009nm lightsheet at a power of 4\u2009mW.\n\nElectronic supplementary material\n=================================\n\n {#Sec14}\n\nSupplementary Information Supplementary Video 1\n\n**Electronic supplementary material**\n\n**Supplementary information** accompanies this paper at doi:10.1038/s41598-017-05028-2\n\n**Publisher\\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nWe gratefully acknowledge funding from BHF Centre for Research Excellence, the EPSRC, the MRC and Kidney Research UK. MTC was supported by the Conselho Nacional de Desenvolvimento Cient\u00edfico e Tecnol\u00f3gico (Brazil). We thank Dr Linda Mullins for critical reading of the manuscript and the Wellcome Trust for infrastructure support for the zebrafish aquarium. Multiphoton imaging was carried out at the CALM_Live imaging facility at the University of Edinburgh.\n\nJ.J.M. and J.M.G. were responsible for funding and overall direction of the project. C.B.B. designed and performed the experiments, prepared the figures and wrote the manuscript. M.T.C. and C.B. designed and built the Bessel beam integration into the SPIM L.K.Y. and J.T. built the SPIM setup. C.M. and S.R. designed and created the mem-KillerRed zebrafish. R.V. performed multiphoton imaging. All authors reviewed the manuscript.\n\nCompeting Interests {#FPar1}\n===================\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Introduction {#sec1}\n============\n\nMethods for converting mechanical energy into light emission are promising for use in illumination,^[@ref1]^ displays,^[@ref2]^ and stress sensors.^[@ref3],[@ref4]^ Conventionally, strain-induced luminescence, which is also known as mechanoluminescence (ML), is the most common method for achieving such energy conversion.^[@ref5],[@ref6]^ For example, Jha and Chandra reported the impulsive excitation of the ML in SrAl~2~O~4~:Eu^2+^ and Dy^3+^ phosphors.^[@ref7]^ However, this mechanism has several limitations that affect its practical applications. First, the high-intensity ML materials are usually inorganic materials,^[@ref8]^ such as quartz,^[@ref9]^ rare-earth ion-doped aluminates,^[@ref10]^ and doped zinc sulfide.^[@ref11]^ To generate ML, their large Young's modulus results in a high threshold pressure at the scale of several MPa.^[@ref12],[@ref13]^ Second, the ML is usually accompanied by the material damage and the decay of the luminescence intensity, which affect ML's reproducibility.^[@ref14],[@ref15]^ Recently, novel triboelectrification-induced luminescence (TIEL) that could convert kinetic energy into light emission was reported.^[@ref16]^ The TIEL exhibited an exceptionally low threshold that was 2--3 orders of magnitude lower than the conventional ML. It also possessed high luminescence intensity. If the TIEL and the ML could be achieved in a single material, significant luminescence intensity from the mechanical--optical conversion can be achieved.\n\nIn this work, we report a stretchable hybrid luminescent composite (HLC) that could generate both ML and TIEL under different mechanical stimuli. The HLC possessed two luminescence modes. The luminescence could be excited either by externally applied mechanical strain or by contacting objects that interact with the HLC. On the one hand, when it was deformed, such as being twisted or folded, the ZnS/Cu phosphors experienced mechanical strain that triggered the ML of the phosphors. On the other hand, as the multilayered composite luminescent material contacted with an external moving object, the triboelectrification at the contact interface induced electroluminescence of the phosphors. A series of internal and external parameters were investigated on how they influenced luminescent intensity for each of the modes above. It was found that the luminescent intensity from the two modes could be superposed. We used the HLC material to fabricate a fiber-based light source that could be driven by air flow. The overall luminescent intensity was enhanced by over 72% compared to that obtained solely from the ML. The HLC reported in this work has such potential applications as self-powered light sources and sensors as means of detecting dynamic motions and interaction.\n\nResults and Discussion {#sec2}\n======================\n\nThe schematic diagram of the HLC is shown in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}a. The HLC basically consisted of a phosphor and a polydimethylsiloxane (PDMS) matrix; and polytetrafluoroethylene (PTFE) nanoparticles were added into the top layer of the HLC. The cross-sectional scanning electron microscope (SEM) view of the HLC is presented in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}b, which clearly shows a two-layered structure. The ZnS/Cu phosphor was well dispersed in the PDMS matrix that was used for transferring externally applied stress to the particles. The magnified view in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}c presents the distribution of the PTFE nanoparticles within the HLC. These particles play two roles. In addition to reducing the friction coefficient between the HLC and a contact object, they also enhanced the triboelectrification between the contact surfaces because the PTFE strongly attracts negative triboelectric charges.^[@ref17],[@ref18]^ The magnified SEM image of the phosphor is shown in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}d. The ZnS/Cu phosphor has a diameter of \u223c26 \u03bcm. It is a commonly used material that could be excited by either mechanical stress^[@ref19]\u2212[@ref22]^ or a charging electric field.^[@ref23]\u2212[@ref25]^ The X-ray diffraction (XRD) patterns show that the phosphors have a wurtzite structure ([Figure S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf)). A photograph of the as-fabricated HLC is shown in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}e. A diagram of the experimental setup used for spectral measurement is presented in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}f. An optical fiber-based probe collected the optical emission from the back of the HLC. The detailed fabrication process is discussed in [Methods](#sec4){ref-type=\"other\"}.\n\n![Structure of the stretchable HLC. (a) Schematic of the HLC. (b) Cross-sectional SEM image of the HLC. (c) Magnified image of the PTFE nanoparticles distributed in the PDMS matrix. (d) Magnified image of phosphor. (e) Photograph of an as-fabricated HLC. (f) Diagram of the experimental setup used for the optical measurement. Schematic illustration of the external stimuli from (g) stretching and (i) sliding with a contact object. Photographs of the ML (h) and the TIEL (j) from the HLC.](ao9b01717_0001){#fig1}\n\nThe HLC has two basic luminescence modes for the mechanical--optical conversion, as illustrated in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}g--j. The first mode is based on the ML. The luminescence can be induced by applying a mechanical force to the HLC. When stretched, the ZnS/Cu phosphor produces ML ([Video S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_001.mp4)). The fundamental mechanism of the ML process has been extensively studied in the literature.^[@ref26],[@ref27]^ Basically, the strain-induced electric field of the charged dislocations causes band bending of the ZnS crystal. As a result, the electrons trapped in the shallow donor level tunnel to the conduction band. When the excited electrons of the Cu^2+^ ions fall back to the *t*~2~ level of the Cu^2+^, photons are emitted ([Figure S2](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf)).^[@ref26]^ A photograph of the ML as the HLC was being stretched is presented in [Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}h. For the second mode, if a contact object slides on the top surface of the HLC, the phosphor underneath the sliding trajectory can also produce luminescence \\[[Figure [1](#fig1){ref-type=\"fig\"}](#fig1){ref-type=\"fig\"}i,j, [Video S2](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_002.mp4)\\]. This type of luminescence is referred to as TIEL, as previously reported.^[@ref16]^ Essentially, the TIEL relies on the coupling of triboelectrification and electroluminescence. When a relative sliding occurs between two dissimilar materials, surface charge transfer takes place due to the triboelectrification effect. The surface triboelectric charges generate a transient electric field along the sliding trajectory, exciting the phosphor underneath. As indicated by the discussion above, the HLC can produce light emission when triggered by multiple types of stimuli.\n\nIn the following sections, the two luminescence modes are investigated in detail, especially on respective influencing factors. First, the major factors that influence the TIEL are investigated in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}. On the one hand, the two parameters regarding the design of the material are shown in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}a,b, that is, the PTFE concentration and the overall thickness, respectively. As the content of the PTFE nanoparticles increased, the luminescence intensity initially increased and then decreased, resulting in an optimal weight percentage at 7.0%. The luminescent intensity was enhanced by over 300% compared to that obtained from the luminescent composite without PTFE nanoparticles. This finding shows that the addition of the PTFE nanoparticles do promote the triboelectrification between the two contacting surfaces. However, it was found that excessive PTFE nanoparticles reduced the transparency of the HLC, which blocked the luminescence from transmitting through the HLC. We also fabricated one-layered HLC and bilayered HLC. The overall thickness of the one and two layer structure is the same, which is about 0.33 mm. The content of PTFE nanoparticles in the one-layered HLC and the first layer of bilayered HLC was 6.98 wt %. There were no PTFE particles in the second layer of the bilayered HLC. The TIEL of the one-layered HLC and bilayered HLC was measured at a velocity of 10 mm/s and a stress of 25 kPa, as shown in [Figure S3](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf). The TIEL intensity of the bilayered HLC is over twice that of the one-layered HLC. Soon, Jeong et al. reported one-layered ML devices.^[@ref20],[@ref28]^ Wang et al. reported a multilayered ML device using the polyvinylidene fluoride film to enhance ML intensity.^[@ref29]^ Wei et al. reported a multilayered TIEL device using ZnS/Cu particles as a luminescent layer and fluorinated ethylene propylene (FEP) film as an electrification layer.^[@ref16]^ However, the device is not stretchable. The advantage of the bilayered HLC is that the bilayered HLC can integrate ML and TIEL. The bilayered HLC also show higher luminescence intensity than the one-layered HLC when they were driven by the same mechanical stimulus. In addition, the overall thickness of the HLC could also affect the luminescence intensity with constant thickness of the first layer. As the thickness increased from 0.28 to 1.53 mm, the luminescence intensity decreased substantially by over 90% of the original value. This can be intuitively explained by the fact that increasing the thickness prevents the emitted light from escaping through the HLC. On the other hand, another two parameters regarding how the contact object interacted with the HLC were also examined, as shown in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}c,d. The luminescence intensity was found to be very sensitive to the applied pressure at the contact interface ([Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}c). Indeed, as the pressure increased from 5 to 40 kPa (velocity, 80 mm/s), the luminescence intensity was enhanced by a factor of 6.5. This enhancement is attributed to the fact that high contact pressure favors the generation of high-density triboelectric charges.^[@ref30]^ As a result, a high electric field can be obtained to excite the luminescence. As the pressure surpassed 40 kPa, the luminescence intensity tended to saturate because the triboelectric charges density reached a maximum limit ([Figure S4](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf)).^[@ref31]^ Besides, as shown in [Figure [2](#fig2){ref-type=\"fig\"}](#fig2){ref-type=\"fig\"}d, the velocity of the sliding object exerted a profound effect on the luminescence intensity (stress, 20 kPa). Similar observation was also reported in previous research.^[@ref16]^ Higher sliding velocity essentially increased the equivalent frequency of the changing electric field,^[@ref32]^ which benefited the electroluminescence of the phosphor.\n\n![Optical measurement results of the TIEL from the HLC. (a) Luminescence spectra with increasing contents of the PTFE nanoparticle. (b) Luminescence spectra with increasing the HLC thickness. (c) Luminescence spectra as the stress at the contact surfaces varies from 5 to 40 kPa. (d) Luminescence spectra as the velocity of the sliding object varies.](ao9b01717_0002){#fig2}\n\nSecond, a number of factors influencing the ML were investigated with experimental results shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}. A linear motor was adopted to apply a periodic stretching strain to the HLC. The strain rate substantially influenced the luminescence intensity. The ML intensity had a positive linear correlation with the strain rate, as shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}a. This observation was consistent with previous reports.^[@ref20]^ When the average strain rate increased from 139 to 451%/s, the ML intensity increases by approximately 377% (the initial length of the HLC is 7 mm). Detailed observations on the peak wavelength revealed that the ML spectrum shifted slightly to shorter wavelength as the strain rate increased. This is attributed to the activation of different energy levels of the ZnS/Cu phosphor, as reported previously.^[@ref28]^ Shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}b, the peak wavelength shifted from 519 to 513 nm. Besides, the applied strain was also positively correlated with the luminescence intensity, as exhibited in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}c. The luminescence intensity increased by about 3.5 times when the strain increased from 28.6 to 128.6%. The maximum principal stress is proportional to the strain of the film.^[@ref33]^ In addition, larger HLC thicknesses resulted in higher luminescence intensities. The luminescent intensity increased by about 5.6 times when the thickness of the HLC changed from 0.28 to 1.53 mm, as shown in [Figure [3](#fig3){ref-type=\"fig\"}](#fig3){ref-type=\"fig\"}d. This is because that more phosphors were involved in the ML process when the thickness of the HLC increased. Then, more generated emissions could be obtained during the same detection time of the spectrometer.\n\n![Optical characteristics of the ML from the HLC. (a) ML spectra and (b) peak wavelength with increasing the externally applied strain rate. (c) ML spectra under different strains of the HLC. (d) Dependence of the ML spectra on the thickness of the HLC.](ao9b01717_0003){#fig3}\n\nBased on the two luminescence modes discussed above, a wind-driven luminescent device was fabricated, in which the TIEL and the ML could be superimposed to promote the overall luminescence intensity. We fabricated two kinds of HLC with and without embedded PTFE particles on the surface. The HLC were sliced into fibers with a width of \u223c1 mm and a length of 3.5 cm. The fibers were rolled up around a gas tube at one end, leaving the other end of the fibers free-standing. There is no significant difference in the luminous intensity between two kinds of HLCs, as shown in [Figure S5](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf). The devices showed little TIEL. It is because PTFE and PDMS are very close in the triboelectric series.^[@ref34]^ There were little transferred surface triboelectric charges between the layer with PTFE particles and the layer without PTFE particles. Then, we fabricated two types of fibers. As illustrated in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}a, the device was composed of two types of fibers bundled together. The first type had a core--shell structure. The core was based on the HLC, and the shell was made of parylene coating with a thickness of 2 \u03bcm \\[[Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}b(i)\\]. This shell material not only served as a protective coating that promoted the mechanical robustness of the HLC-based fiber against external stimuli but also acted as an electrification material that generated and retained the triboelectric charges. Another type of the fibers was made of plain FEP, as shown in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}b(ii). FEP is transparent and then will not prevent the emitted light from being transmitted to the spectrometer. It is a good triboelectrically negative material.^[@ref35]^ Much surface triboelectric charges could be generated on the parylene-coated HLC, which could generate TIEL. The device was fabricated through assembling the two types of fibers together at one end and leaving the other end of the fibers free-standing. The photograph of the mixed fibers is shown in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}b(iii). To trigger the luminescent device, air flow was introduced from the assembled end along the length of the fibers, as revealed in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}c. Excited by the air flow, the mixed fibers fluttered violently, producing high-intensity luminescence, as shown in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}d. The luminescence comes from two sources. First, the strain incurred within the HLC-based fibers generates ML of the phosphor. Second, the contact between the FEP fibers and the HLC-based fibers induced the TIEL. As the air flow pressure increased, the luminescence intensity became significantly enhanced ([Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}e). This observation was supported by the experimental data presented in [Figure [4](#fig4){ref-type=\"fig\"}](#fig4){ref-type=\"fig\"}f. If the FEP fibers were removed, the HLC-based fibers alone produced a luminescence intensity that was 42% lower than that produced by the hybrid fibers. This is because that the HLC-based fibers mainly generated the ML with few TIEL.\n\n![Demonstration of the HLC used in an air-driven luminescence device. (a) Schematic diagram of the core--shell-structured HLC fibers, with a magnified view of the core--shell structure. (b) Photographs of the HLC-based fibers, FEP fibers, and mixed fibers with an average width of \u223c1 mm. (c) Static image of the wind-driven luminescence device. (d) Photograph of the wind-driven device with flowing air. (e) Dependence of the luminescence spectra on the releasing pressure of the air gun. (f) Luminescence spectra obtained from the mixed fibers and the HLC-based fibers, respectively.](ao9b01717_0004){#fig4}\n\nConclusions {#sec3}\n===========\n\nIn summary, in this work, a stretchable bilayered luminescent composite is developed based on combination of strain-induced luminescence and TIEL. The composite film embedded with green light-emitting doped ZnS particles exhibits two luminescence modes. The luminescence can be excited either by externally applied mechanical strain or by contacting objects that interact with the HLC. The surface triboelectric charges resulting from the sliding motion generate a charging electric field along the sliding trajectory, thereby exciting the underlying phosphors. The low-pressure threshold (5 kPa) of the electroluminescence implies that these devices have wide applications. Furthermore, a gas flow system was presented to simulate the effect of wind, and increasing the air compressor discharge pressure was found to increase the luminescence of the composite film. The luminescent intensity collected from the device decreased after elimination the TIEL of the device. This study demonstrates the potential applications of this HLC in self-powered light sources.\n\nMethods {#sec4}\n=======\n\nFabrication of the Stretchable HLC {#sec4.1}\n----------------------------------\n\nThe stretchable HLC studied here had a two-layered structure. In the first layer, ZnS/Cu particles (Shandong Wanji, M-G-AAA) were mixed with the mixture of PDMS and curing agent (10:1 weight ratio, Dow Corning 184) at a weight ratio of 2:1. The mixed solution was spread over an acrylic plate covered with scotch tape and smoothed with a plastic card. After curing at 80 \u00b0C for 1 h, the second layer was spun onto the mixture. The spun speed for the prepared second layer was 2000 rpm (SETCAS Electronics, KW-4B). This second layer consisted of ZnS/Cu particles, PDMS matrix, and PTFE nanoparticles (DuPont). The ZnS/Cu particles and the PDMS matrix were mixed at a weight ratio of 1:1. The ratio between ZnS/Cu and PTFE particles changed from 10:0 to 10:3.\n\nFabrication of the Wind-Driven Luminescent Device {#sec4.2}\n-------------------------------------------------\n\nThe wind-driven luminescent device consisted of the FEP fibers and the HLC-based composite fibers. An FEP film of 100 \u03bcm thickness was cut into 5 cm \u00d7 3.5 cm pieces. The FEP fibers were obtained from partially slicing a FEP piece in which each slice had a width of \u223c1 mm and a length of 3.5 cm. The FEP piece was cut into fibers by laser (Universal VLS 2.3). ZnS/Cu particles were mixed at a weight ratio of 2:1 with the PDMS matrix. After curing at 80 \u00b0C for 1 h, the composite material was cut into 5 cm \u00d7 3.5 cm pieces. To obtain the mixed fibers, the HLC was partially sliced the same as the FEP fibers. Then, the HLC-based fibers were deposited with a layer of parylene-N coating by vacuum vapor deposition (PDS-2010 Specialty Coating Systems), forming a core--shell structure. The weight of parylene-N used in depositing was 2 g. The two types of fibers were bundled together and rolled up around a gas tube that had an external diameter of 5 mm at one end, leaving the other end of the fibers free-standing.\n\nOptical Characterization {#sec4.3}\n------------------------\n\nTIEL measurement: a sponge (3M 9448A) was cut into pieces with a size of 0.4 cm \u00d7 0.5 cm. A piece of the sponge covered with a layer of nitrile rubber was fixed on a linear motor (Zolix LMA-TR-200-E10). The covered sponge slid on the HLC via a linear motor. The applied forces between the moving object and the HLC were measured by a dynamometer (Mark-10 M5-20). ML measurement: the HLC was stretched and then released by a linear motor. A piece of the HLC (7 mm length) was used to measure the ML spectra of the HLC at different strain rates. The luminescence spectra of the HLC samples were collected by a spectrometer (Idea Optics NOVA). The spectrometer probe was located at a constant position. The ML and TIEL images were captured by a digital camera (Canon EOS 5D Mark III).\n\nThe Supporting Information is available free of charge at [https://pubs.acs.org/doi/10.1021/acsomega.9b01717](https://pubs.acs.org/doi/10.1021/acsomega.9b01717?goto=supporting-info). ML produced by stretching ([MP4](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_001.mp4))TIEL produced by finger sliding with nitrile gloves ([MP4](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_002.mp4))XRD pattern of phosphor, showing a wurtzite structure; band diagram of the ZnS/Cu phosphors; TIEL of the one-layered HLC and bilayered HLC; luminescence spectra as the stress at the contact surfaces varies from 5 to 50 kPa; luminous intensity of the HLC with and without embedded PTFE particles in the surface under air flow ([PDF](http://pubs.acs.org/doi/suppl/10.1021/acsomega.9b01717/suppl_file/ao9b01717_si_003.pdf))\n\nSupplementary Material\n======================\n\n###### \n\nao9b01717_si_001.mp4\n\n###### \n\nao9b01717_si_002.mp4\n\n###### \n\nao9b01717_si_003.pdf\n\nThe authors declare no competing financial interest.\n\nThis research was supported by the National Key R & D Project from Ministry of Science and Technology, China (grant nos. 2016YFA0202701 and 2016YFA0202703), National Science Foundation of China (grant no. 51572030), and Natural Science Foundation of Beijing Municipality (grant no. 2162047).\n"}
+{"text": "Plain language summary\n======================\n\nThis study examined whether the 24-hour pattern of motor activity, ie, information that can be collected easily from wrist-worn activity monitors (such as Fitbit or other activity trackers), can be used to identify whether the effects of a promising new treatment for depression are both rapid and durable. The therapeutic drug ketamine produces rapid (within minutes to hours) antidepressant effects in up to 70% of patients with depression, including those with treatment-resistant depression, the more severe form of the disease. However, in patients who respond to ketamine, the duration of clinical benefit varies, typically ranging from a day to a week. Researchers are eager to identify easily obtained biomarkers, such as information obtained from activity trackers, that could predict which subjects are likely to respond to ketamine, as well as which responders are likely to have a continued antidepressant effect vs those who will quickly relapse. This study used easily worn activity watches to collect activity data from subjects with depression. We found differences between baseline patterns of 24-hour activity and initial response to ketamine (ie, response vs nonresponse) and differences between baseline patterns and continuing response to ketamine. The findings show that prior to treatment, 24-hour patterns of activity might be used to predict both response and relapse patterns in certain depressed individuals. While the results are preliminary, they offer a glimpse into the potential benefits of using simple, noninvasive technologies to identify and individualize personalized therapies for mood disorders.\n\nIntroduction\n============\n\nThe molecular and cellular targets of rapid antidepressant interventions---such as sleep deprivation (SD) or the glutamatergic modulator ketamine---are the focus of intense research because they may provide clues for developing novel and improved rapidly acting antidepressant treatments. In addition to identifying the mechanisms underlying rapid antidepressant response, research is needed to understand the neurobiology of relapse and the durability of any such response. Studies have found that the timing of relapse following rapid interventions differs, suggesting that these therapies have different mediators. For instance, in SD, relapse rapidly follows recovery sleep. In contrast, relapse may extend for days (or weeks) after ketamine treatment.[@b1-ndt-14-2739]\n\nNumerous studies have indicated that chronotherapeutic interventions, such as SD, partial SD, sleep-phase advance, and bright-light therapy (BLT), can initiate an antidepressant response. For instance, previous studies found that 24-hour patterns of motor activity are advanced after SD and BLT and that different antidepressant treatment interventions are often associated with increased and decreased day and night motor activity.[@b2-ndt-14-2739] Furthermore, these rapid antidepressant responses can be extended using conventional drug therapies or used to augment antidepressant response to conventional therapies.[@b2-ndt-14-2739] The ability of electroconvulsive therapy and repetitive transcranial magnetic stimulation to extend the SD response have also been explored, with limited success.[@b3-ndt-14-2739],[@b4-ndt-14-2739] While this evidence suggests that the sleep--wake system can affect and prolong antidepressant response, its role in rapid antidepressant response to ketamine is complex and not well understood. For example, while lithium extended response to SD in a mixed population of subjects with either major depressive disorder (MDD) or bipolar depression,[@b5-ndt-14-2739] it did not extend duration of response to ketamine in subjects with bipolar depression vs MDD.[@b1-ndt-14-2739],[@b6-ndt-14-2739]\n\nAt a system or cellular/molecular level, several factors have also been implicated in extending antidepressant response to SD. For instance, naps may shorten the duration of rapid antidepressant response,[@b2-ndt-14-2739],[@b7-ndt-14-2739] while repeated sleep-phase-advance episodes[@b8-ndt-14-2739] extend SD response rather than shorten response duration. Although pharmacological interventions that reduce microsleeps and naps (ie, modafinil, caffeine, and flumazenil) do not predictably enhance response to SD,[@b9-ndt-14-2739],[@b10-ndt-14-2739] the benzodiazepine receptor antagonist flumazenil has been found to extend response to SD,[@b10-ndt-14-2739] suggesting that sleep and circadian interactions could be involved. At the molecular level, a prolonged SD response is associated with the interaction between a variant of the circadian clock-associated GSK promoter (rs334558^\\*^C) and the long/short form of the serotonin transporter (5-HTTLPR) 5HT allele.[@b11-ndt-14-2739] Diurnal variation in brain derived neurotrophic factor (BDNF) levels predict mood response to SD,[@b12-ndt-14-2739] and increased BDNF levels at 240 minutes post-ketamine infusion predict continued mood response.[@b13-ndt-14-2739] The relationship between sleep, BDNF, and mood has a genetic basis that affects mood response to ketamine[@b14-ndt-14-2739] as well as slow-wave sleep (SWS) production.[@b15-ndt-14-2739]\n\nUnlike monoaminergic antidepressants, which require weeks to months to achieve antidepressant efficacy, ketamine acts within hours to produce a clinically significant antidepressant response.[@b16-ndt-14-2739],[@b17-ndt-14-2739] Indeed, numerous placebo-controlled, double-blind, and open-label studies have demonstrated that ketamine has rapid antidepressant effects; while often transient, these can nevertheless also be sustained in a significant proportion of responders.[@b1-ndt-14-2739],[@b6-ndt-14-2739],[@b18-ndt-14-2739]--[@b20-ndt-14-2739] Ketamine's rapid antidepressant effects have been associated with increased synaptic strength and plasticity following altered glutamatergic signaling and with associated changes in dendritic spines and protein synthesis, including BDNF.[@b21-ndt-14-2739] Enhanced synaptic plasticity and neuronal synchronization, especially in areas involved in mood and behavior, have rapid antidepressant effects[@b22-ndt-14-2739],[@b23-ndt-14-2739] that (along with BDNF) correlate with SWS.[@b24-ndt-14-2739] Ketamine also acts on mTOR and clock genes (*Bmal1*, *Per2*, *Cry1*) to reset circadian timing,[@b25-ndt-14-2739],[@b26-ndt-14-2739] and on BDNF and SWS to affect sleep quality.[@b24-ndt-14-2739] Taken together, the evidence suggests that numerous interactions between sleep homeostatic and circadian systems are possible.\n\nActigraphy-monitored activity--rest rhythms were recently examined as a paradigm to evaluate circadian rhythms as a core construct within the Arousal and Regulatory System domain of the National Institute of Mental Health (NIMH) Research Domain Criteria (<https://www.nimh.nih.gov/research-priorities/rdoc/index.shtml>). Wrist activity has been used as an adjunct measure of central timekeeping[@b27-ndt-14-2739],[@b28-ndt-14-2739] and may have diagnostic and treatment utility,[@b29-ndt-14-2739] especially combined with other measures (eg, dim light melatonin onset).[@b30-ndt-14-2739] A prior analysis from our laboratory examined how initial mood response to ketamine (occurring within 24 hours) altered motor activity patterns over the next few days.[@b29-ndt-14-2739] We found that an initial mood response to ketamine increased the amplitude of motor activity in responders vs nonresponders. In contrast, the current analysis examines how continued response is related to baseline motor activity patterns. The result may help identify at pretreatment which individuals may have short or extended treatment response to ketamine.\n\nMethods\n=======\n\nThis study was conducted as part of several investigations exploring ketamine's antidepressant mechanism of action in patients with treatment-resistant depression. Specifically, the study was part of protocols investigating the clinical effects of ketamine in MDD patients who subsequently received riluzole (another glutamatergic modulator) and the effects of ketamine in a double-blind study of MDD and bipolar depression patients, some of whom received maintenance mood stabilizers. The studies were conducted at the NIMH Clinical Research Center Mood Disorders Research Unit in Bethesda, Maryland (NCT00088699). Specific details have been reported previously.[@b1-ndt-14-2739],[@b24-ndt-14-2739] The studies were approved by the Combined Neuroscience Institutional Review Board of the National Institutes of Health. All subjects provided written informed consent before entry into the study and were assigned a clinical research advocate from the NIMH Human Subjects Protection Unit to monitor the consent process and research participation throughout the study.\n\nParticipants\n------------\n\nA total of 51 inpatients (29 females) aged 20--65 (mean 42.6\u00b11.6 years) with a DSM-IV-confirmed clinical diagnosis of MDD (n=30) or bipolar depression (n=21) were included in the analysis ([Table S1](#ts1-ndt-14-2739){ref-type=\"table\"}). All patients were experiencing a current major depressive episode lasting at least 4 weeks and were required to have previously failed to respond to at least one adequate antidepressant trial (as assessed by the modified Antidepressant Treatment History Form).[@b31-ndt-14-2739] All patients had a Montgomery--\u00c5sberg Depression Rating Scale (MADRS) score \u226520. Patients with bipolar depression were required to have not responded to a prospective open trial of a mood stabilizer while at the NIMH (lithium or valproate for at least 4 weeks at therapeutic levels \\[serum lithium 0.6--1.2 mEq/L or valproic acid 50--125 \u00b5g/mL\\]). Exclusion criteria included psychotic features, a DSM-IV diagnosis of drug or alcohol abuse or dependence in the last 3 months, or the presence of an unstable serious medical illness. Female subjects could not be pregnant or nursing. All participants had been free of psychotropic medications for 2--5 weeks prior to the study, with the exception of mood stabilizers among some patients with bipolar depression (lithium, n=14; valproate, n=5). Cigarette use was permitted during the clinical trial but alcohol use was not. Participants were not allowed to nap during the 3 days prior to and after the infusion procedure and selected a preferred sleep schedule between 10 pm and 7 am.\n\nExperimental design\n-------------------\n\nThe 51 participants wore an Actiwatch (AW64; Philips, Amsterdam, the Netherlands) for 2--3 days before and 3 days after a scheduled ketamine infusion. The watch was removed during testing procedures and bathing. A diary was used to track watch removal and replacement. Ketamine infusion was conducted as previously described.[@b19-ndt-14-2739] Briefly, at about 10 am, MDD or bipolar depression patients received a single intravenous infusion of 0.5 mg/kg racemic ketamine hydrochloride over the course of 40 minutes. Depressive symptoms were examined via MADRS ratings conducted at baseline (60 minutes pre-ketamine infusion), 230 minutes postinfusion (D0), 1 day postinfusion (D1), and 3 days postinfusion (D3). At all time points, change in depressive symptoms was expressed as change in score from baseline.\n\nDay 1 ketamine response\n-----------------------\n\nPatients showing \\<50% improvement on D0 or D1 were classified as D1 nonresponders (NRs). Patients exhibiting a 50% reduction in MADRS scores at 230 minutes postinfusion (D0) or at 9 am on D1 were classified as ketamine responders. This group of responders was subdivided into brief responders (R~B~) and continued responders (R~C~) based on the duration of their clinical response. Specifically, responders who did not meet the 50%-reduction response criteria on D3 were classified as R~B~. Patients who responded on D0 and/or D1 and maintained their response through D3 were classified as R~C~. The days defining the R~B~ sample are of interest because, during this 2-day interval, patients typically exhibit the highest relapse rate during the 4-week postinfusion period.[@b1-ndt-14-2739]\n\nData collection and analysis\n----------------------------\n\nData collection methods have been previously described.[@b29-ndt-14-2739] Briefly, activity counts were collected in 1-minute bins and edited for activity data outliers and intervals of watch removal. One-minute averages were then calculated for each hour. Fitted parameter estimates of 24-hour activity pattern differences (amplitude, timing, and midline statistic of rhythm \\[MESOR\\]) were compared for the NR vs responder (R~B~, R~C~) groups. The clock time of the estimated peak activity (acrophase) was used as a reference marker of timing.\n\nStatistics\n----------\n\nGraphPad Prism (v.6.04) was used to compare 24-hour patterns of wrist-activity for baseline vs treatment on D1 (first day after infusion) and D3. Specifically, differences in 24-hour activity patterns were compared in the ketamine R~B~, R~C~, and NR cohorts at baseline, D1, and D3. Motor activity amplitude and timing were the primary focus of this analysis. [Table 1](#t1-ndt-14-2739){ref-type=\"table\"} reports the clock time of the estimated peak activity time (acrophase).\n\nThe analytical methods have been described previously.[@b29-ndt-14-2739] Briefly, the best-fit 24-hour curves for day (baseline, D1, and D3) and cohort differences (NR, R~B~, and R~C~) were characterized using a least squares sine-wave fit to activity scores. Period, amplitude, and phase estimates of wrist activity were derived using a minimal least squares algorithm to fit the 24-hour time series to a sinusoidal curve of the form (*y*\\[*t*\\] = M + A~sin~\\[2\u03c0\u22c5ft + P\\]) where M is MESOR (a \"baseline\" central value of the sinusoid), A amplitude, f frequency, and P phase. Frequency was constrained to be 6.28 (24 hours). Best-fit parameters for baseline, amplitude, and phase were derived for the aforementioned daily contrasts (eg, R~B~ and R~C~ vs NR). Group and treatment differences among 24-hour activity curves were evaluated using an *F*-test \"loss of fit\" when applying shared vs independently derived parameter estimates to group data. Bonferroni correction for multiple within-day curve comparisons was used with a corrected \u03b1-criterion significance level of *P*\\<0.0167, two-tailed.\n\nResults\n=======\n\nA post hoc analysis examined multiple demographic factors, including age, age of onset, and rating scale scores, and revealed no significant differences in these measures across the three cohorts ([Table S1](#ts1-ndt-14-2739){ref-type=\"table\"}). D1 and D3 MADRS scores were used to classify the baseline patient sample as NR (MDD 17, bipolar depression 13), R~B~ (MDD 7, bipolar depression 3), or R~C~ (MDD 6, bipolar depression 5). Amplitude and peak activity estimates for the NR, R~B~, and R~C~ cohorts at baseline, D1, and D3 are displayed in [Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}.\n\nRating scale scores at baseline, D1, and D3\n-------------------------------------------\n\nAt baseline, no significant difference between MADRS scores was observed across the three cohorts ([Figure 2](#f2-ndt-14-2739){ref-type=\"fig\"}; [Table S1](#ts1-ndt-14-2739){ref-type=\"table\"}).\n\nOn D1, MADRS scores varied among the three cohorts (*F*=25.69, *df* 2,48; *P*\\<0.0001). MADRS scores for the R~B~ (*t*=4.64, *df* 38; *P*\\<0.001) and R~C~ cohorts (*t*=6.88, *df* 39; *P*\\<0.0001) were significantly lower than for the NR cohort. No significant difference in MADRS scores was observed between the R~B~ and R~C~ cohorts ([Figure 2](#f2-ndt-14-2739){ref-type=\"fig\"}). On D3, MADRS scores also varied between the cohorts (*F*=15.51, *df* 2,46; *P*\\<0.0001). MADRS scores for the R~B~ cohort were higher than for the R~C~ cohort (*t*=3.648, *df* 19; *P*\\<0.0017), and MADRS scores for the R~C~ cohort were lower than for the NR cohort (*t*=4.05, *df* 7; *P*\\<0.0001). No significant difference in MADRS scores was noted between the R~B~ and NR cohorts ([Figure 2](#f2-ndt-14-2739){ref-type=\"fig\"}).\n\n24-Hour activity patterns in the NR, R~B~, and R~C~ cohorts\n-----------------------------------------------------------\n\nThe overall 24-hour baseline pattern of activity for the three cohorts was characterized by low values at night (mean \\<25 counts/minute) that began to rise from nighttime lows at about 6 am. Values continued to rise to peak values (200--300 counts/minute) until about 2 pm, then decreased. The lowest levels were observed at about 3 am. Mood-associated 24-hour activity pattern differences were found both pre- and postketamine ([Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}).\n\nWithin-day cohort contrasts\n---------------------------\n\nPre- and postketamine treatment-parameter estimates of 24-hour activity patterns are summarized in [Table 1](#t1-ndt-14-2739){ref-type=\"table\"} and [Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}. On each day, constraining the three cohorts to share MESOR, amplitude, and phase (timing) parameter values resulted in a significant loss of fit (baseline, *F*=3.996, *df* 6,1207, *P*=0.0006; D1, *F*=6.409, *df* 6,1185, *P*\\<0.0001; D3, *F*=4.375, *df* 6,1108, *P*\\<0.0002). Cohort differences between individual parameter tests were examined each day. At baseline, trends were observed for the amplitude of the R~B~ cohort (but not the R~C~ cohort) to be blunted relative to the NR cohort ([Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"} top panel; [Table 1](#t1-ndt-14-2739){ref-type=\"table\"}) and for the timing (acrophase) of the R~C~ cohort to be advanced relative to the NR cohort.\n\nOn D1, the timing of the R~B~ cohort was advanced relative to the NR cohort (*F*=9.074, *df* 1,927; *P*=0.0027; [Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}, middle panel) and trended toward an advance relative to the R~C~ cohort (*F*=3.939, *df* 1,495; *P*=0.0477). Also on D1, a trend was observed for the amplitude of the R~C~ cohort to be increased relative to the NR cohort (*F*=5.210, *df* 1,948; *P*=0.0227). Timing did not differ between the R~C~ and NR cohorts. On D3, the amplitude of the R~C~ cohort was greater than that of the R~B~ cohort (*F*=10.02, *df* 1,462; *P*=0.0017; [Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}, top panel), with a trend to be greater than the NR cohort (*F*=4.464, *df* 1,883; *P*=0.0349). No amplitude or timing differences were seen between the R~B~ and NR cohorts. Timing also did not differ between the R~C~ and NR cohorts.\n\nDiscussion\n==========\n\nThe current study sought to identify baseline and post-treatment indicators of rapid vs continuing antidepressant response to the glutamatergic modulator ketamine. We found that ketamine's rapid antidepressant effects were associated with advanced timing and amplitude differences that varied with duration of antidepressant response. These effects were most pronounced when comparing brief (D1) and continued (D3) response activity patterns with the patterns of ketamine NRs. Notably, trends were observed for baseline 24-hour activity amplitude and timing that were associated with subsequent rapid antidepressant effects. Specifically, the timing of 24-hour activity (phase) in R~B~ and R~C~ subjects was earlier than in NR subjects. In addition, baseline activity amplitude trended lower in R~B~ than NR subjects and in R~B~ than R~C~ subjects. At D3, the amplitude for R~C~ subjects was increased relative to R~B~ subjects, with a trend to increase relative to NR subjects. The current findings, which are specific to treatment-resistant depression, underscore the motor activity amplitude and timing effects of ketamine on rapid antidepressant response. On the one hand, this motor activity result may implicate central or noncentral circadian factors in moderating the durability of mood response to ketamine. On the other hand, homeostatic or environmental factors may contribute, and additional measures would need to be monitored to resolve the contribution of these factors to measured response. The fact that clinical response to ketamine was related to baseline differences in motor activity and, furthermore, that ketamine infusion altered 24-hour patterns of motor activity, suggests that underlying mechanisms linked to the generation of 24-hour activity patterns predict and contribute to the durable features of ketamine-mediated mood effects.\n\nBaseline differences between responders and nonresponders\n---------------------------------------------------------\n\nBaseline differences between 24-hour activity patterns for the NR, R~B~, and R~C~ cohorts indicated that differences in circadian-related circuitry were linked to the acute and durable features of the mood response. These differences may occur at the genetic, cellular, systems, or behavioral level. Interestingly, baseline differences between ketamine responders and NRs with regard to both sleep--wake patterns and motor activity patterns have previously been described.[@b24-ndt-14-2739],[@b32-ndt-14-2739] These findings, as well as ketamine's preclinical effects on altered circadian gene expression,[@b26-ndt-14-2739] suggest possible ketamine-associated interactions between the sleep homeostatic and circadian systems. However, it is beyond the scope of the current study to clearly separate their individual contributions.\n\nThis study extends the earlier finding that biomarkers of motor activity timing are important in examining the underlying mechanisms of ketamine's rapid antidepressant effects. Specifically, such markers might be used to identify clinical interventions that could potentially augment, extend, or diminish antidepressant response, particularly because the present results suggest that ketamine responders tend to have advanced activity timing, and NRs tend to have both delayed and elevated activity patterns during pretreatment. While our findings do not provide evidence of ketamine's effects on circadian clock-related molecules, if ketamine acts on central timekeeping, the results may suggest that ketamine-mediated mood effects are linked to markers of circadian timekeeping (specifically, phase and amplitude). Such markers have previously been shown to be altered by clock gene-related variants. It should be noted that while clock gene variants have been associated with increased motor activity[@b33-ndt-14-2739] -- as was also observed here with ketamine NRs at baseline -- the sample size in this study was not adequate to determine whether such variants contributed to cohort differences in activity patterns.\n\nDifferences in rapid antidepressant response for the NR, R~B~, and R~C~ cohorts\n-------------------------------------------------------------------------------\n\nThe relationship between ketamine's pharmacological effects and its effects on mood and activity level is complex. Activity amplitude and level often increase during effective antidepressant treatment,[@b34-ndt-14-2739] and in the current study amplitude effects were indeed greater for the R~C~ cohort (on both D1 and D3) than for the R~B~ cohort. Interestingly, rapid antidepressant response on D1 for the R~B~ cohort was associated with a phase advance of activity but not with change in amplitude ([Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}). The fact that the amplitude and timing parameters for the R~B~, R~C~, and NR cohorts differed at baseline, D1, and D3 ([Figure 1](#f1-ndt-14-2739){ref-type=\"fig\"}) suggests biological differences in the organization and function of the sleep and circadian timekeeping systems. Delayed circadian phase, which was present in NRs relative to responder cohorts, has been linked to elevated glutamatergic levels in emerging depression.[@b35-ndt-14-2739] It is interesting to note that both delayed phase and increased amplitude may be markers of nonresponse in individuals with treatment-resistant depression. Delayed-phase phenotypes have previously been associated with increased circadian amplitude of body temperature[@b36-ndt-14-2739] as well as with increased risk of anxiety, depressive, or substance-use disorders.[@b37-ndt-14-2739]\n\nBaseline versus postinfusion effects between the NR, R~B~, and R~C~ cohorts\n---------------------------------------------------------------------------\n\nComparison of response maintenance features of the NR, R~B~, and R~C~ cohorts demonstrated differences in motor activity amplitude, timing, and level of activity at baseline, with some effects persisting on D1 and D3. These differences indicate that timekeeping differences both mediate and moderate the duration of antidepressant response to ketamine. Specifically, blunted amplitude at baseline was associated with brief antidepressant response and elevated amplitude with more persistent response or nonresponse to ketamine. The observation that amplitude did not vary from baseline to D3 in the R~B~ cohort suggests that low amplitude is a trait-like feature of brief response and rapid relapse. One might speculate that low amplitude could be used to identify which subjects would benefit from particular future treatment interventions, but this remains to be determined given that amplitude estimates varied across both the NR and responder groups. The fact that some interventions, such as BLT or conventional antidepressant treatments, increase 24-hour rhythm amplitude[@b38-ndt-14-2739]--[@b40-ndt-14-2739] suggests that pretreatment or supplemental BLT might be used to prolong ketamine's antidepressant effects.\n\nIn the R~B~ and R~C~ cohorts, but not in NRs, ketamine was associated with advanced timing (R~B~) and increased amplitude (R~C~) of 24-hour activity. This suggests that advanced timing and increased amplitude are specifically related to brief vs continued antidepressant response to ketamine. Interestingly, the advanced motor activity timing observed on D1 was not present in the R~B~ cohort on D3, suggesting that the advance in the R~B~ cohort was associated with positive mood response on D1, and that the delay was associated with relapse on D3. Altered timing and increased amplitude are associated with antidepressant response;[@b40-ndt-14-2739]--[@b44-ndt-14-2739] therefore, the current results indicate possible linkage to duration of antidepressant response.\n\nOne might speculate that the relationship between 24-hour amplitude, clock gene variants, and circadian plasticity (eg, increased range of synchronization to environmental cues) might be associated with relapse and continued response, such that low amplitude (weak) circadian rhythms would allow for greater phase instability and larger phase shifts.[@b45-ndt-14-2739],[@b46-ndt-14-2739] These novel and as yet unexplored possibilities are discussed in the following section.\n\nCircadian clock- and sleep-associated factors\n---------------------------------------------\n\nMutations in such canonical clock genes as *CLOCK* and *PER* and associated clock-related mutations may underlie the temporal variation in activity. These have been described in clinical and preclinical mood-disorder studies.[@b33-ndt-14-2739],[@b47-ndt-14-2739]--[@b51-ndt-14-2739] Clock gene variants are often associated with diurnal preference, and are thus useful targets for exploring the pathogenesis of mood disorders.[@b52-ndt-14-2739] *CLOCK* gene variants affect timekeeping markers,[@b53-ndt-14-2739],[@b54-ndt-14-2739] and specific markers such as amplitude and phase vary with morning--evening chronotypes,[@b36-ndt-14-2739] suggesting both are useful for exploring rapid antidepressant response to ketamine. The extent to which response and relapse might be related to *CLOCK* gene- or sleep-related molecules remains to be determined.\n\nMarkers of circadian timekeeping---for instance, circadian amplitude and phase change---have been interpreted as markers of the intrinsic strength of the central circadian clock.[@b55-ndt-14-2739],[@b56-ndt-14-2739] In this regard, high-amplitude rhythms are often viewed as more stable and resistant to change by external perturbations, whereas those with low amplitude are more labile. Notably, other factors also influence amplitude and phase.[@b36-ndt-14-2739],[@b57-ndt-14-2739],[@b58-ndt-14-2739] At the gene level, circadian systems with mutant clock gene homologues have both lower activity amplitude and greater phase response to external challenges.[@b56-ndt-14-2739] Both may be related to continued ketamine mood response.\n\nIf increased circadian amplitude is related to a strong central clock, then clock strength would be expected to be related to continued mood response and greater resistance to external perturbations. Conversely, decreased amplitude and a weak clock would be consistent with D3 relapse in the R~B~ cohort and less resistance to change from external cues and vulnerability to depressogenic factors. Further, because circadian amplitude is related to the genetic redundancy of clock genes,[@b56-ndt-14-2739] these results may indicate that the low-amplitude cohort that relapses rapidly has less clock gene redundancy than the high-amplitude cohort that exhibits continued antidepressant response. Differences in circadian amplitude and incidence of clock gene variants may thus contribute to the transient vs preserved mood-response patterns in these cohorts.\n\nConclusion\n==========\n\nThis clinical study found that markers of motor activity associated with timing and amplitude were important indicators and mediators of durable antidepressant response to ketamine. The current findings suggest that, prior to treatment: 1) a higher amplitude delayed 24-hour motor activity pattern is associated with nonresponse; 2) a low amplitude 24-hour activity pattern is associated with rapid relapse; and 3) a high amplitude advanced activity pattern is associated with continued response. Together, these observations suggest that biological elements underlying the expression of 24-hour motor activity rhythms contribute to durable antidepressant response and to relapse. In addition, postinfusion results suggested that increased 24-hour activity amplitude was associated with prolonged response. In contrast, in the R~B~ cohort, rapid changes in motor activity timing were associated with rapid response followed by rapid relapse. Overall, the findings indicate that the amplitude and timing of 24-hour motor activity contribute to the durability of ketamine's rapid antidepressant effects. It is important to recognize that while 24-hour patterns of motor activity might be associated with underlying central timekeeping effects, activity patterns are also influenced by other behavioral and environmental effects that mask the central effect. Therefore, studies using more established markers of central circadian timing (eg, dim light melatonin onset, 24-hour body temperature) are warranted to determine whether the strength and organization of the circadian system specifically contribute to preserving ketamine's rapidly acting antidepressant effects.\n\nSupplementary material\n======================\n\n###### \n\nDemographics of brief responder (R~B~) versus continued responder (R~C~) and nonresponder (NR) cohorts\n\n                             Total                                                  Ketamine response   *\u03c7*^2^     *df*       *P*-value                                           \n  -------------------------- ------------------------------------------------------ ------------------- ---------- ---------- ----------- ---- ------ ------ --------- ---------- ---------------\n                                                                                                                                                                                  \n  Sex                                                                                                                                                        0.06      2          0.97\n  \u2003Female                    29                                                     56.9                6          54.5       6           60   17     57                          \n  \u2003Male                      22                                                     43.1                5          45.5       4           40   13     43                          \n  Bipolar (total)            21                                                     41.2                5          45.5       3           30   13     43.3   4.48      2          0.11\n  Mood stabilizers (total)   19                                                     37.3                5          45.5       3           30   11     37     3.16      2          0.21\n  \u2003Lithium                   14                                                     27.5                4          36.4       3           30   7      23                          \n  Valproic acid              5                                                      9.8                 1          9.1        0           0    4      13                          \n                                                                                                                                                                                  \n                                                                                                                                                             ***F***   ***df***   ***P*-value**\n                                                                                                                                                                                  \n  Age (years)                42.6\u00b11.6[\\*](#tfn4-ndt-14-2739){ref-type=\"table-fn\"}   40.4\u00b13.4            43.1\u00b14.2   43.3\u00b12.1   0.25        2    0.78                               \n  Age of onset (years)       17.7\u00b11.2                                               17.7\u00b12.1            17.1\u00b12.2   17.9\u00b11.8   0.04        2    0.97                               \n  MADRS baseline             33.2\u00b10.7                                               32.5\u00b11.4            33.2\u00b11.5   33.4\u00b11.0   0.11        2    0.89                               \n\n**Notes:** *\u03c7*^2^ test comparing mood stabilizers was pooled.\n\nMean \u00b1 SEM.\n\n**Abbreviation:** MADRS, Montgomery--\u00c5sberg Depression Rating Scale.\n\nThe authors thank the 7SE research unit and staff for their support, and the patients for their invaluable contributions. This work was supported by the Intramural Research Program at the National Institute of Mental Health, National Institutes of Health (IRP-NIMH-NIH; ZIA MH002927; NCT00088699), by a NARSAD Independent Investigator Award to CAZ, and by a Brain and Behavior Mood Disorders Research Award to CAZ. These organizations had no further role in study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.\n\n**Disclosure**\n\nCAZ is listed as a coinventor on a patent for the use of ketamine in major depression and suicidal ideation; as a coinventor on a patent for the use of (2*R*,6*R*)-hydroxynor-ketamine, (*S*)-dehydronorketamine, and other stereoisomeric \"dehydro-\" and hydroxylated metabolites of (*R,S*)-ketamine metabolites in the treatment of depression and neuropathic pain; and as a coinventor on a patent application for the use of (2*R*,6*R*)-hydroxynorketamine and (2*S*,6*S*)-hydroxynor-ketamine in the treatment of depression, anxiety, anhedonia, suicidal ideation, and posttraumatic stress disorders. He has assigned his patent rights to the US government, but will share a percentage of any royalties that may be received by the government. The other authors report no conflicts of interest in this work.\n\n![Baseline and ketamine treatment 24-hour motor activity markers of rapid mood response.\\\n**Notes:** Mean \u00b1 SEM estimated amplitude (left) and timing (phase; right) parameters for a 24-hour sinusoidal curve fit to wrist activity shown for three ketamine-treated patient cohorts: brief responders (R~B~), continued responders (R~C~), and nonresponders (NR). Parameter values shown for 3 days: baseline (BL), day 1 (D1) postketamine, and day 3 (D3) postketamine. Amplitude is expressed in mean activity counts/minute/hour, and timing is the clock time of estimated peak activity. Bonferroni-corrected \u03b1-criterion was set to *P*\\<0.0167. Letters above bars indicate the results comparing the amplitude and phase parameters of the NR cohort with the R~C~ and R~B~ cohorts. Numerals above bars indicate the results comparing the amplitude and timing parameters between the R~C~ and R~B~ cohorts. Group contrasts: a, b, *P*\\<0.05 (NS) and 0.005, respectively; 1, 2, *P*\\<0.05 (NS) and 0.005, respectively, vs R~C~.\\\n**Abbreviation:** NS, not significant.](ndt-14-2739Fig1){#f1-ndt-14-2739}\n\n![Baseline (BL), day 1 (D1), and day 3 (D3) Montgomery--\u00c5sberg Depression Rating Scale (MADRS) score.\\\n**Notes:** Baseline (BL), day 1 (D1), and day 3 (D3) Montgomery-Asberg Depression Rating Scale (MADRS) scores in three patient cohorts: brief responders (R~B~), continued responders (R~C~), and nonresponders (NR). No differences among group scores were observed at BL. On D1, scores for the NR cohort were higher than for both the R~B~ and R~C~ cohorts, which did not differ from each other. On D3, scores for the NR and R~B~ cohorts were both higher than for the R~C~ cohort.](ndt-14-2739Fig2){#f2-ndt-14-2739}\n\n###### \n\nBaseline circadian markers of wrist activity in brief (R~B~) vs continued (R~C~) ketamine responder and nonresponder (NR) groups\n\n  -----------------------------------------------------------------------------------------------------------------------------\n                     MESOR (counts)                      Amplitude (counts)                 Acrophase (clock time)\n  ------------------ ----------------------------------- ---------------------------------- -----------------------------------\n  NR (n=30)          167.4\u00b15.3                           127.0\u00b17.5                          14:59\u00b10:14\n\n  R~B~ (n=10)        132.1\u00b16.4                           98.1\u00b19.1                           14:12\u00b10:21\n\n  R~C~ (n=11)        160.3\u00b16.9                           119.1\u00b19.7                          14:08\u00b10:19\n\n  Group contrasts\\   \\                                   \\                                  \\\n  \u2003NR vs R~C~\\       NS\\                                 NS\\                                *F*=5.1, *df* 1,1186; *P*=0.0247\\\n  \u2003NR vs R~B~\\       *F*=12.8, *df* 1,946; *P*=0.0004\\   *F*=4.29, *df* 1,946; *P*=0.039\\   *F*=2.29, *df* 1,946; *P*=0.13\\\n  \u2003R~C~ vs R~B~      *F*=8.87, *df* 1,498; *P*=0.003     *P*=0.12                           NS\n  -----------------------------------------------------------------------------------------------------------------------------\n\n**Note:** Mean estimates \u00b1 SEM.\n\n**Abbreviations:** MESOR, midline statistic of rhythm; NS, not significant.\n"}
+{"text": "Introduction {#s1}\n============\n\n*\"...it appears that the true history of the colonization of the land, now Bermuda, is lost forever in oblivion\"* [@pone.0011375-Taylor1]\n\nStudies of island biodiversity have focused largely on adaptive radiations associated with neoendemism (i.e., \"cradles\" of diversity) [@pone.0011375-Schluter1]--[@pone.0011375-Losos1]. There are myriad factors that promote spectacular biodiversity on islands, but the factors that contribute to neoendemism ultimately derive from the fact that the islands formed *de novo* with no connection to a larger landmass and thus have abundant \"empty\" ecological niche space. However, the isolation that defines islands can also preserve genetic diversity of relict lineages, a pattern known as paleoendemism [@pone.0011375-Gillespie1], [@pone.0011375-Stebbins1]. The most prominent example of this phenomenon among vertebrates is the tuatara (*Sphenodon*) that represents a clade of reptiles once widespread, but now restricted to two remaining species found only on the offshore islands of New Zealand. Moreover, given numerous threats such as climate change, introduced species, and habitat destruction, insular fauna are subject to increased risk of extinction [@pone.0011375-Jamieson1]. This vulnerability is of particular concern for paleoendemics, as these taxa represent a disproportionately high amount of phylogenetic diversity [@pone.0011375-Faith1]--[@pone.0011375-Isaac1].\n\nThe islands of Bermuda (32.33\u00b0N, 64.75\u00b0W) are an isolated, 54 km^2^ archipelago (referred to as \"island\", hereafter) approximately 1000 km east of the United States ([Fig. 1a](#pone-0011375-g001){ref-type=\"fig\"}). The island is currently home to a single endemic terrestrial vertebrate, the scincid lizard *Plestiodon longirostris* (formerly *Eumeces longirostris* [@pone.0011375-Brandley1], [@pone.0011375-Smith1]), although several other terrestrial vertebrates, including a tortoise (*Hesperotestudo*) and multiple species of birds, inhabited the island until from the Middle to Late Pleistocene [@pone.0011375-Sterrer1]--[@pone.0011375-Olson6]. *Plestiodon longriostris* is currently considered critically endangered [@pone.0011375-IUCN1] and faces continuing threats of extinction through human-caused habitat loss, competition with and predation from introduced species, and entrapment in discarded bottles [@pone.0011375-Davenport1]. Once abundant [@pone.0011375-Garman1], [@pone.0011375-Wingate1], the species\\' range has dwindled to several small sub-populations, the largest of which occurs on Southampton Island and contains only an estimated 400 individuals [@pone.0011375-Davenport1]. Preliminary microsatellite analysis indicates that the current genetic diversity of *P. longirostris* is low [@pone.0011375-Coughlan1].\n\n![The location of Bermuda, phylogeny of *Plestodon*, and molecular estimates of divergence times.\\\na., Map showing the location of Bermuda relative to North America. b., Phylogeny of the genus *Plestiodon* (outgroups not shown; see Brandley et al, 2010). Branch lengths are in units of time and represent the means of the posterior distribution. Numbers above or below the nodes indicate posterior probabilities. Triangles indicate groups for which multiple species were sampled, but are not shown. Taxa in blue are those that inhabit eastern North America. Green bars indicate the 95% credible interval for estimated divergence dates for that node. c., Posterior probability distribution of the age of the divergence between *Plestiodon longirostris* and its sister lineage. Areas shaded in red are values that exceed the 95% credible interval.](pone.0011375.g001){#pone-0011375-g001}\n\nPre-plate tectonics biogeographic hypotheses assumed that *P. longirostris* was a relict of an ancient lineage that became isolated on Bermuda after the closing of a land bridge [@pone.0011375-Taylor1]. This \"relict\" hypothesis was also influenced by the species\\' unique morphological characteristics (including an elongated snout). However, recent geologic studies have demonstrated that, although part of an ancient, volcanic sea-mount, most of the exposed, terrestrially habitable portion of Bermuda has never been connected to a larger landmass and instead consists of limestone deposited in the Pleistocene [@pone.0011375-Hearty1]--[@pone.0011375-Hearty4], the maximum age of which is no more than two million years [@pone.0011375-Vacher1]. Moreover, all known Bermudan vertebrate fossils are limited to sediments from the latter half of the Pleistocene [@pone.0011375-Meylan1]--[@pone.0011375-Olson6]. A more reasonable alternative hypothesis is that the species is quite young, at most as old as the maximum age of Bermuda (1--2 million years), and is a descendent of one of the several species of *Plestiodon* inhabiting the eastern United States that dispersed over water to Bermuda following the emergence of the island. Another reptile (indeed, the only other potentially native reptile), the diamondback terrapin (*Malaclemys terrapin*), is a likely very recent immigrant that descended from populations of the same species that currently inhabit the eastern United States [@pone.0011375-Parham1], [@pone.0011375-Davenport2].\n\nTo evaluate the evolutionary history of this unique skink, and specifically to test the \"relict\" and \"recent immigrant\" hypotheses, we conducted Bayesian relaxed molecular clock divergence dating analyses of an extensive, multi-locus DNA data set including all *Plestiodon* species from the eastern United States and the Bermudan *P. longirostris*. We demonstrate that the oceanic island of Bermuda, despite its young age (no more than two million years) and history of extreme changes in available habitat, harbors the last representative of one of the earliest North American lineages of *Plestiodon* that diverged \u223c16 million years ago.\n\nResults and Discussion {#s2}\n======================\n\nThe results of the phylogenetic analysis (pruned to show relevant taxa only; [Fig. 1b](#pone-0011375-g001){ref-type=\"fig\"}) show that the lineage that includes modern *P. longirostris* does not descend from any of the extant *Plestiodon* lineages that inhabit eastern North America. In fact, Bayesian divergence dating analyses using a relaxed molecular clock indicate that it is one of the earliest diverging lineages of the entire North American clade ([Figs. 1b,c](#pone-0011375-g001){ref-type=\"fig\"}). Furthermore, this divergence occurred \u223c16 million years ago (Ma) (95% credible interval\u200a=\u200a11.5 to 19.8 Ma), well before any modern species, and the lineage contains no extant representative other than *P. longirostris*. We are therefore left with the remarkable conclusion that a two million-year-old island contains the sole survivor of an ancient lineage that predates the existence of Bermuda by well over 10 million years.\n\nAlthough no fossils of the *P. longirostris* lineage are known from mainland North America [@pone.0011375-Taylor1], the phylogenetic and divergence date data clearly indicate that ancestors once existed on the mainland. However, *P. longirostris* is present in Bermudan fossil beds dated to approximately 400,000 years ago. Given the young age of Bermuda, these results support the hypothesis that within the past 400,000 to two million years, individuals from the North American mainland *P. longirostris* lineage dispersed to a presumably recently emerged Bermuda 1000 km offshore, yet subsequently became extinct on the mainland. Although we can only speculate how these colonizing individuals dispersed over water, we note that both hurricanes and ocean currents are known to transport living lizards and debris to and from islands [@pone.0011375-Calsbeek1]--[@pone.0011375-Censky1], and that the powerful Gulf Stream ocean current runs along eastern North America to the mid-Atlantic Ocean [@pone.0011375-Davenport2].\n\nThus, despite its young age, the island preserves the last representative of one of the oldest lineages of mainland North American *Plestiodon* -- it is essentially an evolutionary \"life raft\". This \"life raft\" role is remarkable considering that extreme fluctuations in sea level during the Pleistocene have intermittently decreased the available terrestrial habitat on Bermuda by orders of magnitude [@pone.0011375-Olson5]. This extreme contraction in habitat was the likely cause of the extinction of several endemic birds [@pone.0011375-Meylan1]--[@pone.0011375-Olson6]. Bermuda\\'s only other native reptile, *Malaclemys terrapin*, only colonized the island in the past 3000 to 400 years from populations that currently inhabit the Eastern United States [@pone.0011375-Parham1].\n\nThese results are extremely unlikely to be the result of error in divergence date estimates or phylogenetic uncertainty. One advantage of Bayesian methods of divergence dating is their ability to incorporate error in both the calibration age constraint and phylogeny, and to infer posterior probability distributions of estimated ages. The 95% credible interval of the date of divergence between *P. longirostris* and other North American *Plestiodon* species ranges from 11.5 to 19.8 Ma and excludes the earliest date that Bermuda may have emerged (2 Ma). Although the relationship between *P. longirostris* and its sister lineage is not well supported (posterior probability\\<0.95), this species is nonetheless excluded from any other younger clades with strong statistical support ([Fig. 1b](#pone-0011375-g001){ref-type=\"fig\"}). In other words, the lineage cannot be any younger than the other major clades that all predate the emergence of Bermuda. Furthermore, if this lack of resolution represents a rapid radiation at the base of the North American *Plestiodon* phylogeny, then the age of the *P. longirostris* lineage is even older (\u223c13--21 Ma; [Fig. 1b](#pone-0011375-g001){ref-type=\"fig\"}). Finally, extensive analyses of this data set have demonstrated that the use of models that account for rate heterogeneity among subsets of the DNA dramatically improve divergence date estimates and help mitigate potential problems inherent in using distantly related age calibrations [@pone.0011375-Brandley2].\n\nWe also note that the history of *P. longirostris* somewhat parallels that of the extinct Bermudan turtle, *Hesperotestudo bermudae*. Fossil evidence indicates this species became extinct on Bermuda 300,000 years ago, yet was the last representative of a genus of tortoise that inhabited North America from the Oligocene to Pleistocene [@pone.0011375-Meylan1], [@pone.0011375-Olson5]. With the caveat that more recent fossils of North American *Hesperotestudo* may yet to be discovered, these data currently suggest that Bermuda also served as an evolutionary life raft for this genus after extinction on the mainland.\n\nAlthough we certainly do not discount the profoundly important role of islands in generating biodiversity, our results highlight the frequently overlooked role of islands in preserving diversity (acting as evolutionary \"museums\"). This role is of particular importance given that preservation of phylogenetic diversity has been an increasingly important goal of conservation biology as the extinction of \"old\" species would result in a greater loss of genetic diversity than that of a \"young\" species with close phylogenetic relatives [@pone.0011375-Faith1]--[@pone.0011375-Isaac1]. Therefore, because the Bermuda skink, *Plestiodon longirostris*, represents the sole representative of one of the earliest diverging lineages among North American *Plestiodon*, efforts to preserve this species are also preserving \u223c12 to 20 millions of years of unique evolutionary history at the risk of extinction.\n\nMaterials and Methods {#s3}\n=====================\n\nTaxon and character sampling {#s3a}\n----------------------------\n\nDNA for 62 individuals representing 37 of \u223c43 recognized species of *Plestiodon* and 25 outgroups was isolated from tissue using Qiagen DNeasy\u2122 columns (see [@pone.0011375-Brandley2] for detailed specimen information and methods). We amplified BDNF, MKL, mtDNA \\[ND1+tRNAs\\], PRLR, PTGER4, R35, RAG1, and SNCAIP loci using standard PCR techniques (Genbank numbers upon acceptance). In few cases, we were unable to obtain reliable sequences for some species; in this case, another species from the same family was used. PCR products were cleaned using ExoSap-IT (USB Corp.). Purified templates were dye-labeled using BigDye\u2122 (ABI) and sequenced on an ABI 3077\u2122 automated DNA sequencer. Nucleotide sequences were examined and aligned by eye. This process was relatively straightforward for the protein-coding genes (BDNF, MKL, mtDNA ND1, PRLR, PTGER4, R35, RAG1, and SNCAIP) due to their codon reading frames. MtDNA tRNAs were aligned according to their secondary structure, and regions in which homology was uncertain due to multiple insertions and deletions were excluded from subsequent analysis. The size of the final data set for phylogenetic analysis was 7667 bp.\n\nPhylogenetic analyses {#s3b}\n---------------------\n\nBrandley et al. [@pone.0011375-Brandley2] demonstrated that accommodating different rates of evolution among subsets of DNA data may improve divergence time estimation, especially when different subsets of the data evolve at different rates. We therefore partitioned the data *a priori* by locus and codon position (and a single partition for the tRNAs) for a total of 28 partitions. For each partition, we determined the appropriate model of nucleotide substitution using the Bayesian information criterion (BIC) [@pone.0011375-Schwarz1].\n\nAll phylogenetic analyses of the combined data set were conducted using BEAST v1.4.8 [@pone.0011375-Drummond1] assuming an uncorrelated lognormal relaxed molecular clock [@pone.0011375-Drummond2]. A total of seven analyses were performed. Each analysis used a coalescent starting tree and was run for 10^8^ generations, sampled every 10,000^th^ generation. We used the program\\'s default prior distributions with the exception of GTR substitution rates in which we used a uniform (0,100) distribution, and the date distributions of the most recent common ancestor of the three clades used for calibration (see below). To determine convergence, we constructed cumulative posterior probability plots for each clade using the *cumulative* and *compare* function in AWTY [@pone.0011375-Nylander1]. Posterior probabilities\u22650.95 are considered statistically significant clade support [@pone.0011375-Huelsenbeck1].\n\nBecause the *Plestiodon* fossil record, and the record of fossil skinks in general, are extremely poor, we used three fossil calibration age prior distributions from non-scincid fossil taxa whose phylogenetic placement in the squamate tree was recently inferred [@pone.0011375-Conrad1]. The age of crown Episquamata (represented here as *Anniella, Aspidoscelis*, *Basiliscus*, and *Bipes*) was calibrated using the age of the earliest stem \"anguimorph\" fossils, *Becklesius, Dorsetisaurus*, *Paramacellodus*, and *Pseudosaurilius* [@pone.0011375-Brandley2], [@pone.0011375-Conrad1]. We chose a lognormal distribution so that the earliest possible sampled age corresponds to 148 Ma and the older 97.5% credible interval (CI) encompasses the earliest age of crown Squamata (180 Ma: standard deviation\u200a=\u200a1.769; [@pone.0011375-Conrad1], [@pone.0011375-Wiens1]). The age of the divergence between Amphisbaenia (*Bipes biporus*) and Teiidae (*Aspidoscelis*) was calibrated using the age (Albian - Cenomanian boundary) of the earliest teiioid (Polyglyphanodontidae) fossils, (e.g., *Bicuspidon* [@pone.0011375-Conrad1], [@pone.0011375-Nydam1]). We chose a lognormal distribution so that the earliest possible sampled age corresponds to 96 Ma and the older 97.5% credible interval (CI) encompasses the earliest age of crown Episquamata (148 Ma; standard deviation\u200a=\u200a2.016). The age of Scinciformata (represented here by skinks, Gerrhosauridae, and Xantusiidae) was calibrated using the age (Berriasian) of the fossil *Sakurasaurus* [@pone.0011375-Conrad1], [@pone.0011375-Evans1]. We chose a lognormal distribution so that the earliest possible sampled age corresponds to 138 Ma and the older 97.5% credible interval (CI) encompasses the earliest age of the root (151 Ma; standard deviation\u200a=\u200a1.309). We therefore enforced the monophyly of these clades in accordance with recent phylogenetic analyses that have inferred these relationships [@pone.0011375-Townsend1], [@pone.0011375-Hugall1]. The full phylogeny, including all outgroups, is provided in [Fig. S1](#pone.0011375.s001){ref-type=\"supplementary-material\"}.\n\nSupporting Information {#s4}\n======================\n\n###### \n\nFull phylogeny including fossil calibrations (in red). Boxes indicate 95%CIs of node ages.\n\n(1.48 MB EPS)\n\n###### \n\nClick here for additional data file.\n\nWe thank J. Davenport, A. Dornburg, J. McGuire, A. Seago, and one anonymous reviewer for advice and/or comments on the manuscript; J. Aguilar, K. Ashton, J. Campbell, S. Daniels, U. Garc\u00eda, L. Kitson, J. Lazell, I. L\u00f3pez, R. Macey, A. Mendoza, M. Moci\u00f1o, R. Murphy, E. P\u00e9rez, T. Reeder, J. Richmond, CAS, FMNH, KUZ, LSUMZ, MVZ, UMMZ, and YPM for providing tissue samples; and J. Ichikawa, I. Katsube, A. Mendoza Hern\u00e1ndez, E. P\u00e9rez Ramos, A. Seago, and Y. Yamamoto for fieldwork assistance.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This project was funded the National Natural Science Foundation of China (30700062) awarded to X.G.; and the the National Science Foundation East Asia and Pacific Program grants (OISE 0513295 and OISE 0611646, in conjunction with the National Natural Science Foundation of China and the Japan Society for the Promotion of Science), National Science Foundation Doctoral Dissertation Improvement grant (DEB 0709885), the Linnean Society Systematics Fund, and multiple grants from the Museum of Vertebrate Zoology awarded to M.C.B. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: MB. Analyzed the data: MB. Contributed reagents/materials/analysis tools: YW XG ANMdO MFO TH HO. Wrote the paper: MB.\n\n[^2]: Current address: Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut, United States of America\n"}
+{"text": "Conflicts of interest\n=====================\n\nThe authors declare that there is no con\ufb02ict of interest.\n\nIntroduction\n============\n\nThe helicarionid land snail genus *Otesiopsis* Habe, 1946 is known from Japan and Taiwan ([@B5428703]). There are five recognised species: *O. japonica* (M\u00f6llendorff, 1885), *O. taiwanica* (Kuroda, 1941), *O. kawaguchii* Habe, 1961, *O. kanmuriyamensis* Azuma, 1973 and *O. yamamotoaizoi* Minato, 1974 ([@B5428663], [@B5428693], [@B5364729], [@B5364719], [@B5364748]). While *O. japonica* and *O. taiwanica* have relatively wide distribution areas in western Japan and on the entire island of Taiwan, respectively ([@B5364758]; [@B5364739]), the remaining three species seem to inhabit only limited mountain areas or small islets ([@B5364729]; [@B5364719]; [@B5364748]). The population densities of *Otesiopsis* are considered low (e.g. [@B5364768]). At present, there are only five occurrence records in GBIF (<https://www.gbif.org/species/4599174>). Due to the low population densities and/or restricted distributions, all species except *O. taiwanica* are listed in the national level red data book ([@B5364780]).\n\nGageodo Island is a small island in South Korea, with an area of about 9.1 km^2^ and situated about 120 km from the Korean peninsula (Fig. [1](#F5364791){ref-type=\"fig\"}). The highest mountain on the island is Mt. Doksilsan with an elevation of 639 m. The non-marine gastropods have so far not been investigated on Gageodo Island.\n\nMaterials and Methods\n=====================\n\nOn 12-14 June 2019, during a field survey, three specimens of the genus *Otesiopsis* were collected on Gageodo Island, Sinan County, South Jeolla Province (Fig. [1](#F5364791){ref-type=\"fig\"}). The shells were examined under a light microscope (Olympus SZ40). The collected specimens were deposited in the Specimen Repository of Kuroha Shell Museum (Voucher No: MNKS3347; 3405; 3476).\n\nResults and Discussion\n======================\n\nOrder Stylommatophora Schmidt, 1855 (see [@B5428683])\n\nFamily Helicarionidae Bourguignat, 1877 (see [@B5428673])\n\nGenus *Otesiopsis* Habe, 1946\n\n*Otesiopsis* sp.\n\nDescription\n-----------\n\nShell medium size for this genus, very thin, glossy, dull chestnut brown, with weak growth lines. Whorls 5\u00bd increasing regularly in diameter. Protoconch 1\u00bd whorls with a smooth and glossy surface. Spire moderately low. Suture shallow but distinct. Body whorl with a strong peripheral keel. Base strongly glossy, more convex than the upper part and concave in the umbilical region. Aperture wide, semilunate in outline. Outer lip very thin, sharp, non-reflected, angulated at the periphery, gently curved at the base. Columellar lip short, oblique, slightly thickened. Umbilicus narrowly open.\n\nHeight 3.64 mm, width 7.24 mm (Figs [2](#F5364831){ref-type=\"fig\"}, [3](#F5364835){ref-type=\"fig\"}, Voucher No: MNKS3405).\n\nRemarks\n-------\n\nOf the three specimens collected in this study, only one individual was sexually mature, which was confirmed by its egg-laying behaviour. Although intraspecific variation needs to be investigated in the *Otesiopsis* snails on Gageodo Island, the adult specimen examined here exhibited distinct differences from other *Otesiopsis* species. The Gageodo *Otesiopsis* has a smaller body size and thicker shell base than *O. japonica* and *O. taiwanica*. It can be distinguished from *O. kawaguchii* and *O. kanmuriyamensis* by having a perforated umbilicus and from *O. yamamotoaizoi* by a larger body size and carinate periphery. *O. yamamotoaizoi* recorded from the Danjo Islands (Nagasaki prefecture, Japan), situated about 380 km southeast of Gageodo Island, is the geographically closest species of *Otesiopsis*.\n\nThe *Otesiopsis* snails occurred from Gageodo Island are likely to be an undescribed species. However, morphologies of its reproductive system and radula, which are important traits for classification of *Otesiopsis* (e.g. Habe 1946; Azuma 1973), were not obtained in this study. Further anatomical studies using additional adult specimens are needed to clarify this newly recorded land snail from South Korea. In addition, since the *Otesiopsis* species are generally rare and vulnerable, further field surveys on Gageodo Island and its adjacent areas would be of importance in order to confirm its conservation status.\n\nWe express our sincere gratitude to Y. Kim for considerable assistance in the field survey and N. Kawaza for helpful information on Japanese *Otesiopsis*. This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2016R1A6A1A05011910). Finally, we are grateful to A. Jochum, T. S. Liew and T. Hirano for providing us with constructive comments on this manuscript.\n\nConflicts of interest\n=====================\n\nThe authors declare that there is no con\ufb02ict of interest.\n\n![Sample collection locality in Gageodo Island, South Korea (34\u00b003\\'41.2\\\"N, 125\u00b007\\'29.5\\\"E).](bdj-07-e46984-g001){#F5364791}\n\n![Living individual of *Otesiopsis* collected from Gageodo Island, South Korea. Scale bar = 5.0 mm.](bdj-07-e46984-g002){#F5364831}\n\n![Shell morphology of *Otesiopsis* collected from Gageodo Island, South Korea (Voucher No: MNKS3405). The protoconch was broken. Part of the shell was underexposed. Scale bar = 1.0 mm.](bdj-07-e46984-g003){#F5364835}\n\n[^1]: Academic editor: Alexander M. Weigand\n"}
+{"text": "Introduction {#s1}\n============\n\nHand hygiene is widely regarded as the single most important intervention to reduce the burden of health care-associated infections and the transmission of antimicrobial resistance within the hospital setting [@pone.0105866-World1]. The contemporary approach to promotion of hand hygiene amongst healthcare workers involves a multimodal strategy incorporating the use of alcohol-based handrub at the point of care [@pone.0105866-World1], [@pone.0105866-Pittet1]. The WHO 'My 5 Moments for Hand Hygiene' methodology defines *when* healthcare workers should perform hand hygiene during patient care [@pone.0105866-Sax1], [@pone.0105866-Sax2]. Healthcare worker compliance with these indications is part of routine performance feedback, an essential strategy for behaviour change [@pone.0105866-Anderson1]. WHO recommendations also exist for *how* to perform hand hygiene, but these are rarely monitored or included in performance feedback programs. This technique is based on European standards (EN 1500) and involves six distinct steps, or poses [@pone.0105866-World1]. Correct performance of this technique results in increased product coverage and greater reductions of bacterial colony forming units when compared with incomplete actions [@pone.0105866-Widmer1], [@pone.0105866-Widmer2].\n\nSureWash (Glanta Ltd, Dublin, Ireland) is a commercially available device that combines e-learning and patented video measurement technology to teach healthcare workers how to perform a hand hygiene action. It uses interactive on-screen feedback to encourage grounded cognition and reflection on the technique of hand hygiene. The aim of this approach of situated cognition is that the physical act of hand hygiene becomes a prompt to the actions of good technique. The device can be left in a clinical area to be used independently by healthcare workers, and provides immediate and individualised performance feedback.\n\nThe primary objective of this study was to assess the efficacy of SureWash to improve hand hygiene technique amongst healthcare workers in an institution with a long history of hand hygiene promotion [@pone.0105866-Pittet1]. Our secondary objective was to evaluate the ability of SureWash to assess the adequacy of hand hygiene actions performed by healthcare worker staff compared to assessment by trained human observers.\n\nMethods {#s2}\n=======\n\nEthics statement {#s2a}\n----------------\n\nWe followed the principles expressed in the Declaration of Helsinki. This study was approved by the Ethics Commission for Human Research at the University of Geneva (protocol 12--258).\n\nDesign {#s2b}\n------\n\nWe performed a controlled before-after study with blinding of assessors ([**Figure 1**](#pone-0105866-g001){ref-type=\"fig\"}). Allocation was not randomised and there was no placebo intervention. First, baseline assessment (T~0~) of hand hygiene technique was performed in two healthcare worker groups (A and B) to ensure that they did not have significantly different pre-intervention hand hygiene technique. The first follow-up (T~1~) assessment was then performed after Group A (intervention group) had received the intervention but Group B (control group) had not. Subsequently, Group B was exposed to the intervention, and a second follow-up (T~2~) measurement of both groups was performed. The application of the intervention to Group B (the original control group) and T~2~ measurement was performed to 1) examine the persistence post-intervention of any improvement in Group A technique, 2) demonstrate reproducibility of intervention effect in two groups of subjects, and 3) allow Group B to benefit from this quality improvement intervention. This design has been referred to as the \"untreated-control group design that uses dependent pretest and posttest samples and switching replications\" [@pone.0105866-Harris1].\n\n![Study design.](pone.0105866.g001){#pone-0105866-g001}\n\nSetting {#s2c}\n-------\n\nThe University of Geneva Hospitals is a 2200-bed primary and tertiary care hospital in Geneva, Switzerland with a long history of hand hygiene promotion [@pone.0105866-Pittet1], [@pone.0105866-Pittet2]. Healthcare workers are exposed to training in hand hygiene technique during an infection control education session on employment commencement, posters throughout the hospital, and guidelines on the infection control intranet site. This study was performed in four acute care wards within the department of internal medicine, consisting of two pairs of adjacent wards on different hospital floors. In 2013, healthcare worker compliance with indications for when to perform hand hygiene was 75.8% (95% confidence interval \\[CI\\], 70.7--80.4) as measured by direct observation according to WHO methodology [@pone.0105866-Sax1], [@pone.0105866-Sax2].\n\nParticipants {#s2d}\n------------\n\nAll healthcare workers with patient-care responsibilities in the four participating acute-care wards were eligible to participate on a voluntary basis. Subjects were required to provide written, informed consent and were excluded if 1) unlikely to remain in the study wards throughout the study period, or 2) if they currently worked -- or were likely to work during the study period -- in wards in both study groups. Healthcare workers that were not recruited were able to use the SureWash during the intervention phase, but were not monitored for the study.\n\nIntervention {#s2e}\n------------\n\nThe intervention involved self-directed use of the SureWash unit, which was left unsupervised in the staff tea room. Healthcare workers were able to use it in both 'training mode' and 'assessment mode' throughout the intervention phase (four weeks in Group A and 12 weeks in Group B). 'Training mode' consisted of a slideshow with information regarding when and how to perform hand hygiene, and required healthcare workers to practice their own technique in the presence of immediate feedback. 'Assessment mode' involves healthcare workers performing a hand hygiene action and receiving a score (in percentage format) reflecting degree to which each pose was performed correctly and for adequate duration.\n\nProcedure and data collection {#s2f}\n-----------------------------\n\nThe study was implemented from March to September 2013. The study design is presented in [**Figure 1**](#pone-0105866-g001){ref-type=\"fig\"}. At baseline (T~0~), participants completed a brief survey including age, sex, profession, number of years spent working at HUG, and prior participation in the institutional infection control training course. They were then invited by the investigators to perform a hand hygiene action as recommended by hospital guidelines using alcohol-based handrub. This action was recorded by the SureWash device in a purpose-built \"study mode\" whereby it captured video of the hand hygiene action, but provided no feedback other than to indicate to the user that their hands were in the correct position. The action was assessed by the device and the video stored for subsequent assessment by observers. Immediately after recording this action, each healthcare worker was asked to mime the 6 poses by following an on-screen demonstration.\n\nFollowing the one-week recruitment and baseline assessment period (T~0~), Group A was exposed to the intervention for four weeks. During this time, the SureWash unit was left in the staff tea room, available for self-directed hand hygiene technique education and training (as described in the intervention section above) at healthcare workers\\' convenience. Each participant was able to use the device according to their interest and availability: there was no minimum or maximum number of uses required.\n\nSubsequently, during the first follow-up (T~1~), each participant was again asked to perform a hand hygiene action as at T~0~. Group B was then exposed to the intervention for twelve weeks followed by a final assessment of participants in both Group A and B (T~2~).The intervention phase was longer in Group B because whereas the two wards in Group A shared a common tea room, those in Group B did not. In addition, Group B was exposed to the intervention during the summer period when healthcare workers take leave and are therefore frequently absent. Both factors translated to decreased exposure of Group B subjects to the intervention.\n\nFollowing each of the three assessments, two observers (AS and VC) independently reviewed the hand hygiene videos in random order, assessing duration and performance of each pose ([**Figure 2**](#pone-0105866-g002){ref-type=\"fig\"}). A purpose-built interface was developed to facilitate this review process. For bilateral poses, performance of each side was assessed separately. The observers were blinded to study group and the SureWash assessment of the action. Following the review process, data could be exported from the device for analysis. This dataset included the following information for every pose: study group, date, the SureWash unit\\'s automatic assessment of pose performance (measure of \"effort\") and both reviewers\\' binary assessment of whether the pose was performed correctly.\n\n![Poses recommended for hand hygiene actions.\\\nAfter applying a palmful of the product in a cupped hand; 1) rub hands palm to palm; 2) right palm over left dorsum with interlaced fingers and vice versa; 3) palm to palm with fingers interlaced; 4) backs of fingers to opposing palms with fingers interlocked; 5) rotational rubbing, backwards and forwards with clasped fingers of right hand in left palm and vice versa; 6) rotational rubbing of left thumb clasped in right palm and vice versa. Text adapted from reference 1.](pone.0105866.g002){#pone-0105866-g002}\n\nOutcomes {#s2g}\n--------\n\n### Primary objective {#s2g1}\n\nThe predefined outcome used to assess the impact of the SureWash unit on hand hygiene technique was performance of a complete hand hygiene action as rated by both observers. A hand hygiene action was judged as complete if all six recommended poses were performed and the action lasted for 20 seconds or more ([**Figure 2**](#pone-0105866-g002){ref-type=\"fig\"}) [@pone.0105866-World1]. For bilateral poses, both sides had to be performed in order for the posed to be accepted as correctly performed. Poses could be performed in any sequence. The number of times that the SureWash unit was used by each group was recorded as a process measure.\n\n### Secondary objective {#s2g2}\n\nTo evaluate the diagnostic capacity of the SureWash unit we compared the human observer assessment (dichotomous) with the SureWash automated assessment (continuous). The SureWash unit produces a measure of the \"effort\" with which each pose is performed. This \"effort\" measure was a unit-less continuous variable.\n\nStatistical methods {#s2h}\n-------------------\n\nThe sample size calculation was based on the proportion of healthcare workers in each study group performing complete hand hygiene actions at the first follow-up (T~1~). In the absence of prior data, we estimated that 60% of healthcare professionals would perform a complete hand hygiene action at baseline, and proposed that an absolute improvement of 30% following the SureWash intervention would be clinically pertinent. With a two-sided alpha of 0.05 and a power of 0.8, we required 38 participants in each arm. At baseline, however, we noted that no healthcare workers performed a complete hand hygiene action. We had recruited 34 and 29 subjects into the two groups. We therefore performed an estimation of study power based on this new information. Assuming a 10% loss to follow-up (30 and 26 subjects), we had a power of 0.80 to detect a delta of 30% using two sided alpha of 0.05.\n\nCategorical baseline covariates were presented using counts and percentages, with subjects from the two groups compared using Fisher\\'s exact test. Inter-rater agreement between the two blinded observers was computed using Cohen\\'s kappa. These values were interpreted according to Fleiss [@pone.0105866-Fleiss1].\n\n### Primary objective {#s2h1}\n\nThe proportion of healthcare workers performing a complete hand hygiene action in each group was compared at T~0~ to assess the assumption that the two groups had similar baseline hand hygiene technique. We then evaluated change in the proportion of healthcare workers performing a complete hand hygiene action from T~0~ to T~1~ in both Group A (intervention) and Group B (control). The initial control group was then exposed to the intervention, and we evaluated change in the proportion of healthcare workers performing a complete hand hygiene action from T~1~ to T~2~ in both Group B (now intervention) and Group A (now control). For each comparison, the null hypothesis of no difference between the two groups was tested using Fisher\\'s exact test.\n\n### Secondary objective {#s2h2}\n\nWe used the subset of poses for which the two human observers provided the same assessment. We used Receiver Operating Characteristic (ROC) curve analysis to assess the diagnostic performance of this measure, summarised using area under the curve (AUC). The cutoff value for the SureWash \"effort\" measure that best discriminated between adequate and inadequate performance (as determined by the human raters) was determined independently for each pose. These optimal cutoffs were selected as the value that maximised Youden\\'s J statistic. We described performance of the device using these optimal cutoffs by presenting sensitivity, specificity, positive predictive value, negative predictive value and accuracy when compared to the human observer. Accuracy is calculated as the number of poses correctly judged by the SureWash device as either adequate or inadequate divided by the total number of poses. The other parameters were calculated as usual. Confidence intervals (CIs) were computed using the Clopper-Pearson method [@pone.0105866-Clopper1].\n\nStatistical analyses were performed using the R software/environment, version 3.0.1 (R Foundation for Statistical Computing), including 'irr' and 'ROCR' packages [@pone.0105866-R1], [@pone.0105866-Sing1].\n\nResults {#s3}\n=======\n\nSixty-three healthcare workers were recruited, 34 in Group A and 29 in Group B. No eligible healthcare workers refused to participate (due to scheduled rotations only one doctor was eligible), producing a 100% participation rate. Baseline characteristics are presented in [**Table 1**](#pone-0105866-t001){ref-type=\"table\"}. Follow-up was incomplete. Details of follow-up and reasons for missed data are outlined in the flow diagram ([**Figure 3**](#pone-0105866-g003){ref-type=\"fig\"}). Data were missing for six subjects at T~1~ and 14 subjects as T~2~. Two subjects refused to participate on three occasions and were therefore classified as having withdrawn from the trial. The SureWash unit was used 213 and 151 times by healthcare workers in Group A and Group B, respectively, during their intervention phases.\n\n![Study flow diagram.\\\nAll eligible subjects agreed to participate.](pone.0105866.g003){#pone-0105866-g003}\n\n10.1371/journal.pone.0105866.t001\n\n###### Baseline characteristics of study participants.\n\n![](pone.0105866.t001){#pone-0105866-t001-1}\n\n                                                          Group A (n\u200a=\u200a34)   Group B (n\u200a=\u200a29)   p-value\n  ------------------------------------------------------ ------------------ ------------------ ---------\n  Female gender                                               26 (76)            23 (85)         0.522\n  Age category                                                                                   0.022\n  \\<20                                                         0 (0)              0 (0)        \n  20--29                                                       1 (4)              4 (20)       \n  30--39                                                      16 (64)             5 (25)       \n  40--49                                                       8 (32)            11 (55)       \n  Profession                                                                                     0.574\n  Nurse assistant                                             10 (29)             5 (19)       \n  Nurse                                                       21 (62)            19 (70)       \n  Doctor                                                       1 (3)              0 (0)        \n  Other                                                        2 (6)              3 (11)       \n  Years worked at HUG[\\*](#nt102){ref-type=\"table-fn\"}                                           0.338\n  \\<1                                                          1 (3)              0 (0)        \n  1--5                                                         4 (13)             4 (15)       \n  6--10                                                        7 (22)             2 (7)        \n  \\>10                                                        20 (63)            21 (78)       \n  Infection control course completed                                                             0.347\n  No                                                           9 (26)            12 (44)       \n  Yes, in 2013                                                 0 (0)              0 (0)        \n  Yes, in 2012                                                 3 (9)              1 (4)        \n  Yes, before 2012                                            22 (65)            14 (52)       \n\nCounts are presented with percentages in parentheses. Responses to each question may not sum to total number of participants due to unanswered questions.\n\n\\*HUG, University of Geneva Hospitals.\n\nPrimary outcome: Impact of intervention on hand hygiene technique {#s3a}\n-----------------------------------------------------------------\n\nAgreement between the two human raters for each pose is presented in [**Table 2**](#pone-0105866-t002){ref-type=\"table\"}. According to Fleiss\\'s qualitative descriptors for kappa values, agreement was \"fair to good\" for poses 1 and 3, and \"excellent\" for the other four poses.\n\n10.1371/journal.pone.0105866.t002\n\n###### Pass rate and interrater agreement between the two human observers regarding performance of each pose.\n\n![](pone.0105866.t002){#pone-0105866-t002-2}\n\n  Pose                 Kappa          Descriptor\n  ---------------- ------------- ---------------------\n  1                    0.735         Fair to good\n  2 (left/right)    0.974/0.950   Excellent/Excellent\n  3                    0.586         Fair to good\n  4                    0.776           Excellent\n  5 (left/right)    0.817/0.807   Excellent/Excellent\n  6 (left/right)    0.813/0.773   Excellent/Excellent\n\nAll kappa values were computed using 169 subjects, and were significant, with p-values computed as \\<0.001. Poses are illustrated in [Figure 2](#pone-0105866-g002){ref-type=\"fig\"}.\n\nThe primary outcome measure was performance of a complete hand hygiene action. No participants performed a complete hand hygiene action at baseline (T~0~): 0/34 (0.0% \\[95% CI; 0.0%, 10.3%\\]) and 0/29 (0.0% \\[95% CI; 0.0%, 11.9%\\]) in Groups A and B, respectively. The two groups were therefore similar at baseline (p\\>0.99).\n\nBetween T~0~ and T~1~, Group A received the intervention and Group B acted as control ([**Figure 1**](#pone-0105866-g001){ref-type=\"fig\"}). The number of Group A participants that performed a complete action increased from 0/34 (0.0% \\[95% CI; 0.0%, 10.3%\\]) at T~0~ to 1/30 (3.3% \\[95% CI; 0.1%, 17.2%\\]) at T~1~ (p\u200a=\u200a0.47). There was no change in Group B at T~1~, as none of 27 participants (0.0% \\[95% CI; 0.0%, 12.8%\\]) performed a complete action (p\\>0.99).\n\nBetween T~1~ and T~2~, Group B received the intervention and Group A acted as control ([**Figure 1**](#pone-0105866-g001){ref-type=\"fig\"}). No Group B participants performed a complete action post-intervention 0/24 (0.0% \\[95% CI; 0.0%, 14.3%\\]) at T~2~ compared to 0/27 (0.0% \\[95% CI; 0.0%, 12.8%\\]) at T~1~ (p\\>0.99). There was also no change in Group A: 1/30 (3.3% \\[95% CI; 0.1%, 17.2%\\]) at T~1~ to 1/25 (4.0% \\[95% CI; 0.1%, 20.3%\\]) at T~2~ (p\\>0.99).\n\n*Post-hoc* analysis: Impact of intervention on hand hygiene poses per action {#s3b}\n----------------------------------------------------------------------------\n\nGiven the rarity of this primary outcome, we performed a post-hoc assessment of the number of poses performed correctly during each hand hygiene action. The rationale for this post-hoc analysis was that not all incomplete actions are equal: an incomplete action with one pose performed is likely to be less effective in removing organisms from the hand than an incomplete action with five poses, for example. Therefore it is of interest to evaluate the impact of the intervention on the number of poses performed per action, as it is plausible that such an effect may have a positive impact on quality of care. However, as a *post-hoc* analysis, these results should be considered exploratory.\n\nFirst, the proportion of subjects performing each pose, stratified by intervention status, is presented as a descriptive result in [**Table 3**](#pone-0105866-t003){ref-type=\"table\"}. Prior to exposure to the intervention, all poses except pose 2 were performed by less than half of the subject. Pose 4 was performed least frequently. At baseline, healthcare workers generally performed a hand hygiene action comprising of a continuous movement, with infrequent distinct and repeated poses. Improvements were observed in all six poses in the post-intervention period.\n\n10.1371/journal.pone.0105866.t003\n\n###### Number of poses performed correctly according to the two observers, stratified by subject intervention status.\n\n![](pone.0105866.t003){#pone-0105866-t003-3}\n\n  Pose         Pre-intervention   Post-intervention               \n  ----------- ------------------ ------------------- ------------ ------------\n  1               24 (39.3%)         26 (42.6%)       43 (79.6%)   43 (79.6%)\n  2 (Left)        40 (65.6%)         39 (63.9%)       40 (74.1%)   41 (75.9%)\n  2 (Right)       37 (60.7%)         36 (59.0%)       38 (70.4%)   39 (72.2%)\n  3               23 (37.7%)         23 (37.7%)       44 (81.5%)   37 (68.5%)\n  4               9 (14.8%)           4 (6.6%)        18 (33.3%)   11 (20.4%)\n  5 (Left)        21 (34.4%)         16 (26.2%)       31 (57.4%)   30 (55.6%)\n  5 (Right)       21 (34.4%)         17 (27.9%)       31 (57.4%)   31 (57.4%)\n  6 (Left)        11 (18.0%)          7 (11.5%)       22 (40.7%)   23 (42.6%)\n  6 (Right)       11 (18.0%)          5 (8.2%)        22 (40.7%)   25 (46.3%)\n\n\"Pre-intervention\" includes Group A subjects at T~0~ and Group B subjects at T~1~ (n\u200a=\u200a61). \"Post-intervention\" includes Group A subjects at T~1~ and Group B subjects at T~2~ (n\u200a=\u200a54). Poses are illustrated in [Figure 2](#pone-0105866-g002){ref-type=\"fig\"}.\n\nSecond, the number of poses performed per action by the two groups were compared ([**Figure 4**](#pone-0105866-g004){ref-type=\"fig\"} **)**. The two groups performed a similar number of poses correctly at T~0~, before either had been exposed to the intervention: median 2.0 (IQR, 1.5) in Group A and median 1.0 (IQR, 1.5) in Group B (p\u200a=\u200a0.12).\n\n![Number of poses performed correctly per hand hygiene action, by study group and study phase.\\\nGroup A was exposed to the intervention for four weeks between baseline (T~0~) and the first follow-up (T~1~). Group B was exposed to the invention for 12 weeks between the first follow-up (T~1~) and the second follow-up (T~2~). Median and interquartile ranges are represented by the horizontal line and box, respectively. Upper and lower whiskers extend to minimum and maximum values that lie within 1.5 times the interquartile range from the 75th and 25th percentile, respectively. Each p-value relates to the null hypothesis that the two groups perform the same number of poses correctly at that time point.](pone.0105866.g004){#pone-0105866-g004}\n\nThe number of poses performed by Group A (intervention) subjects increased from median 2.0 (IQR, 1.5) at T~0~ to 3.8 (IQR, 2.3) immediately post-intervention at T~1~ (p\\<0.001). Over the same period, there was a lesser absolute increase in the number of poses performed by Group B (control) subjects: median 1.0 (IQR, 1.5) at T~0~ to 2.0 (IQR, 1.8) at T~1~ (p\u200a=\u200a0.03). At T~1~, Group A performed more poses that Group B (p\\<0.001).\n\nGroup B was then exposed to the intervention. The number of poses performed by Group B subjects increased from median 2.0 (IQR, 1.8) at T~1~ to 4.0 (IQR, 2.1) at T~2~ (p\\<0.001). Over the same period, there was no significant change in the number of poses performed by Group A (now control) subjects: median 3.8 (IQR, 2.3) at T~1~ to 3.0 (IQR, 1.5) at T~2~ (p\u200a=\u200a0.49). The number of poses performed by Group A subjects at T~2~ remained significantly higher than baseline (p\\<0.001). At T~2~, Group A and Group B subjects performed a similar number of poses per action (p\u200a=\u200a0.89).\n\nSecondary outcome: Diagnostic capacity of SureWash {#s3c}\n--------------------------------------------------\n\nROC curves for each pose are presented in [**Figure 5**](#pone-0105866-g005){ref-type=\"fig\"}. Performance characteristics when the optimal cutoff (which maximised Youden\\'s J statistic) was employed are presented in [**Table 4**](#pone-0105866-t004){ref-type=\"table\"}.\n\n![Receiver operating characteristic curves for each pose.\\\nGrey points indicate the diagnostic cutoff that maximises Youden\\'s J statistic. AUC, area under the curve.](pone.0105866.g005){#pone-0105866-g005}\n\n10.1371/journal.pone.0105866.t004\n\n###### Performance characteristics of SureWash as a diagnostic test when compared to human observers.\n\n![](pone.0105866.t004){#pone-0105866-t004-4}\n\n  Pose       n       Sensitivity         Specificity          Accuracy               PPV                 NPV\n  -------- ----- ------------------- ------------------- ------------------- ------------------- -------------------\n  Pose 1    147   0.55 (0.44--0.66)   0.84 (0.72--0.92)   0.67 (0.59--0.75)   0.82 (0.7--0.91)    0.58 (0.47--0.68)\n  Pose 2    332   0.81 (0.75--0.86)   0.85 (0.78--0.91)   0.82 (0.78--0.86)   0.90 (0.85--0.94)   0.72 (0.63--0.79)\n  Pose 3    134   0.49 (0.36--0.61)   0.68 (0.56--0.79)   0.58 (0.49--0.67)   0.61 (0.47--0.74)   0.56 (0.45--0.67)\n  Pose 4    157   0.78 (0.58--0.91)   0.70 (0.61--0.78)   0.71 (0.64--0.78)   0.35 (0.23--0.48)   0.94 (0.87--0.98)\n  Pose 5    307   0.66 (0.57--0.74)   0.87 (0.81--0.91)   0.79 (0.73--0.83)   0.77 (0.68--0.85)   0.79 (0.73--0.85)\n  Pose 6    310   0.89 (0.79--0.95)   0.82 (0.77--0.87)   0.84 (0.79--0.88)   0.63 (0.53--0.72)   0.95 (0.92--0.98)\n\nComputed using cutoff values selected to maximise Youden\\'s J statistic. Estimations provided with 95% confidence intervals in parentheses. Poses are illustrated in [Figure 2](#pone-0105866-g002){ref-type=\"fig\"}.\n\nPPV, positive predictive value; NPV, negative predictive value.\n\nDiscussion {#s4}\n==========\n\nThis trial was performed to assess the utility of SureWash in improving hand rubbing technique in a healthcare institution with a long history of hand hygiene promotion [@pone.0105866-Pittet1], [@pone.0105866-Pittet2]. Baseline (T~0~) results demonstrated a need for such an intervention, with no healthcare workers able to perform a hand hygiene action as recommended by WHO [@pone.0105866-World2]. This trial did not demonstrate an impact of the SureWash device on the proportion of healthcare workers able to perform a complete hand hygiene action using strict criteria. However, a *post-hoc* analysis demonstrated that exposure to this device had a significant and (in Group A) durable impact on the number of poses performed correctly per hand hygiene action. Finally, the device demonstrated good diagnostic capacity when compared to human observers.\n\nThese findings are consistent with and extend those of two previous publications using the SureWash unit. Gosh *et al.* used the device in one clinical ward during two six-day phases; the first without feedback (16 subjects) and the second with feedback (34 subjects) [@pone.0105866-Ghosh1]. Inter-rater agreement between two human observers was 0.76 (Krippendof\\'s alpha), with agreement of 0.74 and 0.56 for each observer with the device. Using a less strict definition of complete hand hygiene action (1 second for each pose), the pass rate for hand hygiene actions increased modestly from 62.5% to 64.7% (p\\<0.05). Higgins *et al.* used the SureWash device as part of an institution-wide multimodal hand hygiene promotion campaign [@pone.0105866-Higgins1]. The pass rate for handwashing (rather than hand rubbing) technique using adenosine triphosphate monitoring increased from 52% before implementation of training with SureWash to 79% after (p\\<0.001). Compliance with the six recommended poses was not specifically assessed. Our study confirms the diagnostic capacity of SureWash using a larger sample size that Gosh *et al.* and builds on data from both studies regarding its impact on hand hygiene technique by using a controlled study design implemented in the absence of concurrent interventions, with assessment of each pose, and by using hand rubbing, the preferred technique for routine hand hygiene [@pone.0105866-World1].\n\nThe importance of hand hygiene technique with regard to product coverage and reduction in bacterial counts on hands has been demonstrated previously [@pone.0105866-Widmer1], [@pone.0105866-Widmer2], [@pone.0105866-Kampf1]. The baseline results of this trial suggests that an infection control course on employment commencement, educational posters in clinical areas and availability of guidelines are not sufficient to teach hand hygiene technique. Monitoring and performance feedback is a key strategy to improving healthcare worker hand hygiene behaviour [@pone.0105866-World1], [@pone.0105866-Anderson1], but this traditionally focuses on *when* to perform it rather than *how*. More intensive training can be resource intensive [@pone.0105866-Widmer2]. For example, in a recent study using UV-light technology to assess hand hygiene technique immediately following training, 5200 healthcare workers were exposed to 15-minute education sessions in groups of five to eight [@pone.0105866-Szilagyi1]. This was a major logistic operation and required at least 160 hours work. In contrast, a potential strength of the SureWash unit is that it can be left in clinical areas for independent use by healthcare workers, liberating infection control professionals for other activities. This benefit needs to be counter-weighed against the operating cost of the device.\n\nWe did not demonstrate an impact on the number of healthcare workers performing a complete hand hygiene action. In fact, only two such actions were observed during the study. This may reflect the stringency of the outcome measure definition: six poses (three of which must be repeated bilaterally) performed correctly during at least 20 seconds as judged by two independent human raters. We would consider \"poses per action\" or a microbiologic measure a preferable outcome measure when designing future studies. However this result belies a change in behaviour that occurred nevertheless. At baseline, when asked to perform a hand hygiene action, the overwhelming majority of healthcare workers slid one hand over the other in a continuous, seemingly random movement. Following the intervention, we observed that participants instead made repeated stereotyped poses. This can be appreciated in the *post-hoc* analysis of poses per action. Several approaches could be considered to optimise the impact of this intervention in busy clinical settings: alternative placement or longer exposure to the device; ward based role-models or 'champions' to inspire friendly competition, benchmarking of results against other wards; or a more formal credentialing requirement. Uptake is likely to vary between settings, and a flexible approach involving frontline ownership may be most effective.\n\nThis trial demonstrates that SureWash has good capacity to distinguish between correctly and incorrectly performed poses. However, two issues should be considered when reviewing these data. First, this analysis was performed on recordings made during the three assessment periods (in \"study mode\"), when immediate feedback was not provided to healthcare workers. During standard use, healthcare workers receive immediate feedback in the form of green bars that extend when the pose is being correctly performed. Thus healthcare workers quickly refine their technique by making minor adjustments to hand position or movement, and agreement between the device and human observers could be expected to increase. Second, though good inter-rater agreement between the two observers supports their reliability as a reference diagnostic technique, the human review process was clearly imperfect, involving a degree of subjective judgement.\n\nThese data must be interpreted in the context of the study design. First, this trial was designed to assess the efficacy of SureWash as an educational tool. Consequently, we assessed healthcare workers\\' capacity to perform hand hygiene technique on request, rather than covertly monitoring actual hand hygiene technique during routine patient care. Second, this trial does not provide data regarding the importance of performing hand hygiene as per WHO recommendations. However, the superiority of the WHO technique with regard to product coverage and reduction in bacterial colony forming units has been demonstrated previously [@pone.0105866-Widmer2]. Third, human review of video images was used to assess the primary outcome and also as \"gold-standard\" reference test to evaluate the device\\'s diagnostic performance (secondary objective). Whilst we attempted to quantify reliability of human observers by presenting inter-rater agreement, we acknowledge that this \"gold standard\" is imperfect. Finally, due to anonymity considerations, we were unable to track individual healthcare workers\\' performance through each of the three assessments and correlate improvement with their use of the SureWash unit.\n\nIn summary, no healthcare workers were able to perform a complete hand hygiene action at baseline despite a long institutional history of hand hygiene promotion. While we were unable to demonstrate an increase in complete hand hygiene actions, exploratory *post-hoc* analysis suggested that exposure to SureWash significantly increased the number of poses performed per action, and this effect persisted 12 weeks post intervention. This study identifies a need for further study of hand hygiene technique and demonstrates the potential utility of the SureWash device. Future studies should explore methods to maximise the uptake and effectiveness of this device as well as the impact of improved hand hygiene technique on transmission events or laboratory surrogates.\n\nThe authors gratefully acknowledge Prof Arnaud Perrier and Mme Sandra Merkli for their support, and sincerely thank Mme Sophie Robert, Mme Soraya Hamache, M. Pierre-Andr\u00e9 Berret, Mme Murielle Tarpin-Lyonnet and the healthcare workers of the internal medicine department for their participation. We are grateful to Erik von Elm for his careful reading of this manuscript and thoughtful comments.\n\n[^1]: **Competing Interests:**Trinity College Dublin has been granted patents related to SureWash in USA (US 8090155 B2, Hand washing monitoring system) and EU (EP 2015665 B1, A Hand Washing Monitoring System). These are licensed to Glanta Ltd, a Trinity College Dublin spin-out company that markets the device. DC is a paid employee of Glanta Ltd. GL is co-inventor of SureWash, and founder and chairman of Glanta Ltd. The other authors declare no conflicts of interest. This does not alter the authors\\' adherence to all the PLoS ONE policies on sharing data and materials. AS and DP had full access to all data and vouch for the results.\n\n[^2]: Conceived and designed the experiments: AS AI VC AGA DC GL DP. Performed the experiments: AS AI VC DP. Analyzed the data: AS. Wrote the paper: AS. Developed trial-specific software: DC. Critical review and revision of the manuscript: AI VC AGA DC GL DP.\n"}
+{"text": "1. Introduction {#sec1-materials-09-00274}\n===============\n\nThe adherence of microorganisms on contact surfaces and its persistence for weeks or even months is a main cause of cross-contaminations and spread of infections in healthcare and food processing facilities \\[[@B1-materials-09-00274],[@B2-materials-09-00274],[@B3-materials-09-00274],[@B4-materials-09-00274]\\]. The problem is emphasized by increasing the presence of microorganisms that have become resistant to a number of different antibiotics. The infection of surgical wounds by *Staphylococcus aureus* is a common problem in hospitals and its ability to develop resistance to antibiotics contributes to the danger to patients \\[[@B5-materials-09-00274]\\]. Despite hand-washing campaigns and routine cleaning, infection rates remain high. This has raised the need to combat pathogenic microorganisms to lower the risk of acquiring infections. A promising and effective strategy to prevent spreading of microorganisms in hygiene-sensitive areas is the refinement of inanimate surfaces with antimicrobial properties. Among different strategies, are contact-mediated killing metals like copper (Cu) and silver gaining in interest as self-sanitizing material. A further advantage in antimicrobial metals is that small concentrations are already sufficient to significantly inhibit the metabolism of bacteria, and thus achieve high anti-microbial efficacies (e.g., copper is recognized by the United States Environmental Protection Agency (EPA) as being able to continuously kill more than 99.9% of bacteria that cause hospital-acquired infections in 2 h \\[[@B6-materials-09-00274]\\]). Since copper-containing proteins are involved in cell metabolism, and, hence, copper in low concentration is an essential trace element in living organisms \\[[@B7-materials-09-00274]\\], there is little risk to patients and consumers. However, the redox properties of copper can cause some damage that has been studied for years \\[[@B8-materials-09-00274],[@B9-materials-09-00274]\\]. The inactivation of microorganisms exposed to Cu has been attributed to the following mechanisms: generation of reactive oxygen species that causes oxidative stress, membrane damage due to the loss of structural integrity, and alteration to the conformational structure of proteins or degradation of the cellular DNA and/or RNA \\[[@B10-materials-09-00274]\\]. Materials with Cu doping are already used in medical instruments and sanitary facilities. Thus far, 500 copper alloys were registered as legally permitted antimicrobial materials. A variety of methods have been developed to deposit copper, as metal or as oxide, on surfaces, such as soaking, sol-gel coating, chemical washing, and atmospheric and vacuum coating. Plasma-based techniques are widely used, wherein the most frequently used processes for thin film deposition are based on low-pressure physical vapor deposition (PVD)---for instance, magnetron sputtering \\[[@B11-materials-09-00274]\\], plasma-immersion ion implantation \\[[@B12-materials-09-00274]\\] and chemical vapor deposition (CVD) using organometallic precursors \\[[@B13-materials-09-00274]\\]. Recent progress in the development of atmospheric pressure cold plasma technology has attracted increasing attention in thin film deposition. Intense research efforts are directed accordingly to identify new promising processes for the development of novel bioactive materials. Although there are some examples with respect to coatings with copper or silver nanoparticles, only a few reports can be found on the deposition of metal-containing thin films by using low temperature atmospheric pressure plasma. For instance, Beier *et al.* showed the antimicrobial effect of silver-containing silicon oxide layers generated by atmospheric pressure plasma chemical vapor deposition by introducing a silver nitrate solution into the gas discharge \\[[@B14-materials-09-00274]\\]. This aerosol-assisted process, in which an inorganic nanoparticle-containing dispersion is atomized and injected in atmospheric pressure plasma, is also reported by Fanelli *et al.* \\[[@B15-materials-09-00274]\\]. Another novel approach is the synthesis of copper particles by evaporating and sputtering bulk material from the powered electrode, which is also subject of the present work \\[[@B16-materials-09-00274]\\]. The advantage of the aforementioned plasma process is the delivery of metallic particles in high purity, which was also studied by Lazea-Stoyanova *et al.* by using an atmospheric radiofrequency plasma jet \\[[@B17-materials-09-00274]\\]. For this study, a non-thermal air operated DC plasma jet was used to deposit copper films on temperature labile acrylonitrile butadiene styrene (ABS) at room temperature \\[[@B18-materials-09-00274]\\]. ABS was selected for this study because of its increasing utilization as medical polymer. Due to their processability, recyclability, and mechanical properties, polymeric materials have become an integral part in the healthcare industry \\[[@B19-materials-09-00274]\\]. Moreover, today's ongoing trend towards replacing conventional materials, such as metals, in medical devices by high performance polymers leads to many innovative materials. An essential contribution for medical polymers are styrenic polymers. Their physical and chemical properties meet the demands for many applications covering medical devices up to pharma packaging. Among these materials, ABS is the most important of the styrene copolymers \\[[@B20-materials-09-00274]\\]. The latest technical developments enable the production or grafting of ABS with high added value required for different applications. By incorporating fillers or nanoparticles, processing properties, and the mechanical resistance of ABS can be improved. For example, the thermal stability and tensile strength can be increased by adding silica nanoparticles \\[[@B21-materials-09-00274]\\], by copolymer composites of metal hydroxide nanorods and graphene nanosheets \\[[@B22-materials-09-00274]\\] or by the incorporation of carbon nanotubes and copper particles \\[[@B23-materials-09-00274]\\] whereas silver-containing powder have been added to achieve antibacterial properties \\[[@B24-materials-09-00274],[@B25-materials-09-00274]\\]. However, plasma treated polymers have been characterized in terms of different operating parameters, including air flow rate and input current. X-ray photoelectron spectroscopy (XPS) and atomic force microscopy (AFM) have been used to investigate the chemical composition and the size, as well as the distribution of deposited particles. The bactericidal properties of Cu-modified samples were evaluated against *Staphylococcus aureus (S. aureus*).\n\n2. Results and Discussion {#sec2-materials-09-00274}\n=========================\n\n2.1. Chemical Characteristics {#sec2dot1-materials-09-00274}\n-----------------------------\n\nPlasma-induced surface modification of ABS was investigated by XPS analysis. Representative survey spectra of pristine ABS and ABS exposed to air plasma are shown in [Figure 1](#materials-09-00274-f001){ref-type=\"fig\"}.\n\n[Figure 1](#materials-09-00274-f001){ref-type=\"fig\"} shows that the surface of pristine ABS is composed of carbon (C 1s at 285 eV), nitrogen (N 1s at 399 eV), and oxygen (O 1s at 532 eV). According to its chemical structure, ABS is an oxygen-free polymer, but the presence of surface-near oxygen has been reported frequently and may be related to additives used for the fabrication of this copolymer and to contaminations from ambient air (e.g., adsorbed water) \\[[@B26-materials-09-00274],[@B27-materials-09-00274],[@B28-materials-09-00274]\\]. However, XPS results reveal changes in the survey spectrum of ABS upon exposure to air plasma. New emission lines have been identified that are attributed to copper. The nature of the Cu in the films can be further analyzed from the Cu 2p region of the X-ray photoelectron (XP) spectrum. The corresponding core level Cu 2p spectrum (representative for all of the surfaces studied) in the 930--970 eV region is shown in [Figure 2](#materials-09-00274-f002){ref-type=\"fig\"}.\n\nIntense Cu 2p~3/2~ and Cu 2p~1/2~ peaks are centered at 934 and 954 eV, respectively, along with shake-up features at \\~942 and \\~962 eV for Cu 2p~3/2~ and 2p~1/2~ core levels, the latter being indicative of Cu(II) state \\[[@B29-materials-09-00274],[@B30-materials-09-00274]\\]. The peak positions and the relative intensities of the satellites from these levels are strongly related to the presence of copper oxides, in particular the formation of CuO and/or Cu(OH)~2~ phase at the surface is suggested \\[[@B31-materials-09-00274],[@B32-materials-09-00274]\\].\n\nIn order to understand the influence of experimental parameters on Cu thin film deposition, a thorough XPS analysis was conducted. [Figure 3](#materials-09-00274-f003){ref-type=\"fig\"} shows the time-dependent variation of the elemental composition of plasma treated ABS specimens for input currents of 8 and 30 mA, respectively. It can readily be seen that the deposition of copper is poor at short treatment times. In detail, when using a low input current of 8 mA, Cu was detected only at treatment times above 218 s. Nevertheless, Cu is only present in trace amounts with a maximum of 0.8 at. %. Substantial changes in Cu content can be seen for input currents of 30 mA. Already after 60 s, a Cu amount of 1.5 at. % was detected. It is also evident that increasing the treatment time leads to an increased deposition of copper with a maximum of almost 8 at. % after 864 s. The Cu content, however, slightly decreased at low input currents after 864 s. In the operation of the plasma jet, different modes can be observed that can be characterized primarily as a self-pulsing operation and a true stable DC operation. The plasma characteristic is strongly dependent on the operating current, which, in turn, primarily determines the deposition of copper. For lower currents, a self-pulsing mode is likely the dominant process, while for higher currents, a DC operation is becoming more prevalent \\[[@B33-materials-09-00274],[@B34-materials-09-00274]\\]. With the different discharge modes, plasma parameters are changing and affecting deposition rates consequently. According to these findings, data presented hereinafter were obtained for treatment times of 432 s while changing different process parameters.\n\nA comparison of the elemental composition of plasma treated ABS specimens at various input currents and air flow rates analyzed by XPS is depicted in [Figure 4](#materials-09-00274-f004){ref-type=\"fig\"}. After plasma treatment, the relative surface composition was altered significantly. Compared with pristine ABS, a notable reduction in carbon content is accompanied by a remarkable increase of copper and oxygen concentrations for all tested surfaces. Furthermore, an increase of the nitrogen concentration from initially 2 at. % up to 5.7 at. % was measured. Among the parameters investigated, the input current is the most influential on the elemental surface composition. The relative surface concentration of Cu was found to increase with higher operating current from 0.4 at. % for 8 mA up to 7.5 at. % Cu at 30 mA. A possible explanation for the present result is the energy provided to the plasma changes with increasing input current. Previous characterizations of the plasma reveal that the gas production rate was enhanced with rising input currents \\[[@B34-materials-09-00274]\\]. Consequently, since Cu species were provided into the plasma by sputtering of the electrode, it is most likely that, for high input currents, electron collision processes are becoming important, which influence the generation of reactive plasma species that favor the deposition of copper. In terms of the air flow rate, the results show that this process parameter only marginally effects the elemental composition when operating at 30 mA.\n\nThe chemical identification of functional groups has been conducted by high resolution XPS analysis. [Figure 4](#materials-09-00274-f004){ref-type=\"fig\"} shows the curve-fitted spectra in the C 1s and N 1s regions for the surface of pristine and plasma treated ABS, respectively. In the case of C 1s spectra, it is apparent that only slight changes in the peak shape were induced. The C 1s spectra have been fitted with components assigned to functional groups according to the literature \\[[@B26-materials-09-00274],[@B35-materials-09-00274]\\]. The C 1s envelope was fitted with three components: the peak at 285.0 eV is attributed to C--C and C--H bonds, the second component at 286.5 eV can be associated with either C--O--H or C\u2261N nitrile groups, and the peak at 289.0 eV is assigned to R--O--C=O carboxylic groups or to imide groups. After treatment, a new peak appears at 287.9 eV, which suggests the formation of C=O carbonyl groups and/or N--C=O amide groups. Hence, it was found that, depending on the process parameters, oxygen and nitrogen added by the air plasma forms new double bonds with carbon. In contrast, the N 1s spectrum was found to change after plasma treatment. In the case of pristine ABS, the N 1s core level spectrum shows a single peak at 399.7 eV, which is consistent with the binding energy value reported for C\u2261N functionalities present in the polymer backbone. As can be seen in [Figure 5](#materials-09-00274-f005){ref-type=\"fig\"}, a new peak appears at 407.0 eV after plasma treatment that may arise from oxidized species of nitrogen (e.g., nitro or nitrate groups) \\[[@B36-materials-09-00274]\\]. The formation of these new surface functionalities results from the reaction of surface radicals with nitric oxide, which is generated in the gas discharge \\[[@B34-materials-09-00274]\\].\n\nOverall, functional groups remain on the surface after plasma treatment, but, additionally, the process slightly alters the surface chemistry due to the modification of the polymer chains and the removal of adsorbed contaminants. Along with surface chemical modification, changes in the wettability have been observed depending on the process parameters. ABS is characterized by a poor wettability with an averaged water contact angle of 81\u00b0. After plasma treatment, static water contact angle measurement revealed contact angles ranging from 62\u00b0 at an input current of 15 mA and an air flow rate of 8 slm up to 29\u00b0 at 30 mA and 16 slm. As described in the literature, the adhesion of a thin metallic layer depends on the surface chemistry and can be improved by surface functionalization \\[[@B37-materials-09-00274],[@B38-materials-09-00274],[@B39-materials-09-00274]\\]. Therefore, an increase in surface functionality and in wettability may contribute to the deposition of copper-containing films.\n\n2.2. Topographical Characteristics {#sec2dot2-materials-09-00274}\n----------------------------------\n\nThe bactericidal effect of metal nanoparticles has been attributed to their small size and high surface to volume ratio, which allows them to interact closely with microbial membranes \\[[@B40-materials-09-00274]\\]. The surface topography as well as the distribution of deposited copper particles of plasma treated surfaces was examined by using AFM. To get a better understanding of CuO thin film growth, AFM analysis of surfaces treated with various air flow rates was conducted according to the different amount of deposited Cu. Representative micrographs of pristine ABS and of ABS exposed to different air flow rates at a given input current of 30 mA are displayed in [Figure 6](#materials-09-00274-f006){ref-type=\"fig\"}. Additionally, in order to evaluate the size of deposited particles, height profiles have been extracted from a scanning area of 2.5 \u00b5m \u00d7 2.5 \u00b5m. However, the topography of pristine ABS is characterized by a well-defined hill-and-valley structure with hills of several \u00b5m in diameter, which can be seen in the corresponding height profile. There are also a few dust particles observed on the surface. After plasma treatment, a more irregular surface topography is evident, which is characterized by the presence of copper nanoparticles with a size of several tens of nanometer. Clearly, the variation of the flow rate affects the copper surface coverage: For surfaces of 4 slm, plasma treatment, nanosheet-shaped particles of different diameters have been identified. It has further been found that particles are deposited in islets with no specific distribution pattern. The approximate particle size was estimated to be in the range from 40 to 100 nm with distances in the order of 100\u2212500 nm, which is much smaller than the size of pathogenic bacteria. For instance, the size of bacterial species ranges from 0.2 \u00b5m to the largest cells of sulfur bacteria with a diameter of 750 \u00b5m, whereas for Staphylococci, diameters of 0.5\u22121.5 \u03bcm are reported \\[[@B41-materials-09-00274],[@B42-materials-09-00274]\\]. However, by applying higher flow rates, the number of deposited particles was increased, which led to a particle distribution in a close-packed arrangement. These findings are even more pronounced at the highest flow rate of 16 slm. The corresponding height profiles for surfaces of 8 and 16 slm reveal that the particles have a size ranging from 10 to 80 nm in diameter. Furthermore, a coalescence of spherical surface features can be seen which initiates the forming of bigger compact particles, a process known as agglomeration.\n\nIn order to estimate the particle size more precisely, histograms (statistical distribution of z-values) of AFM images presented in [Figure 6](#materials-09-00274-f006){ref-type=\"fig\"} are investigated. For ease of comparison, the measured height distributions are also plotted as cumulative distribution, which is the integral of the density. As can be seen in [Figure 7](#materials-09-00274-f007){ref-type=\"fig\"}, the average height of non-treated ABS was approximately 118 nm with values ranging between 60 and 180 nm. These values correspond to the hilly landscape of the ABS substrate as well as to the presence of dust particles that originate from sample handling in ambient air. Therefore, processed data of plasma treated substrates presented in [Figure 7](#materials-09-00274-f007){ref-type=\"fig\"} are obtained from flattened images. Hence, comparing the histograms of plasma treated surfaces depending on the air flow rates, a maximum height of 28, 37, and 42 nm for 4, 8, and 16 slm, respectively, was determined. The slight but significant change to higher z values with increasing flow rate indicates some agglomeration of round-shaped copper nanoparticles.\n\n2.3. Antimicrobial Results {#sec2dot3-materials-09-00274}\n--------------------------\n\nBased on the XPS analysis, the antimicrobial activity of copper-modified surfaces and the impact of different amounts of copper towards the Gram-positive bacteria *Staphylococcus aureus* after two hours of incubation were evaluated. In particular, the percentage reduction of colony forming units for each set of samples was calculated to express the change of the microbial population relative to the pristine ABS. The change was determined as follows: $$\\%\\text{~Reduction} = \\ \\frac{\\left( {N_{0} - N} \\right) \\times 100}{N_{0}}.$$\n\n*N*~0~ is the number of colony forming units on pristine ABS, and *N* is the number of colony forming units on plasma treated samples.\n\nIt can be readily seen from the results depicted in [Figure 8](#materials-09-00274-f008){ref-type=\"fig\"} that the deposited copper oxide films demonstrate biocidal activity. In detail, for a low input current of 8 mA, a reduction of up to 84% was achieved among the different air flow rates investigated. At medium input current of 15 mA, an increase of the reduction with a maximum of 87% can be observed. By increasing the input current to 30 mA, the highest percentage reduction of about 93% was obtained in this study. According to the XPS results presented in [Figure 4](#materials-09-00274-f004){ref-type=\"fig\"}, the following assumption can be made: the inactivation of *S. aureus* after contact with copper-coated ABS surfaces was significantly affected by the concentration of copper. Greater antimicrobial activity was noticed for surfaces exposed to an input current of 30 mA, which led to a surface coverage of Cu concentrations between 5 and 7 at. %. Regarding the inactivation mechanisms, it should be noted that copper is a contact-killing agent by interacting either directly with the cellular membrane or intracellularly. In this way, the mechanism for the antimicrobial behavior of the copper-coated ABS might be directly related to the release rate of copper. The release behavior of copper in turn, is influenced, for instance, by the layer thickness, size of copper particles, and the aqueous environment \\[[@B43-materials-09-00274]\\]. Regarding the thickness of the copper oxide films, it can be estimated that the deposits obtained at input currents of 30 mA are the size of a few nanometers only. Moreover, for the lowest flow rate of 4 slm, AFM images ([Figure 6](#materials-09-00274-f006){ref-type=\"fig\"}) revealed a non-uniform film, which was characterized by single copper particles. Apparently, due to the small amount of copper deposited at an operating current of 8 mA, it is most likely that no copper is left on the surface after two hours, which results in low inactivation rates. Hence, in this study, a maximum reduction of 93% has been achieved for plasma treated surfaces with the highest copper concentration. These results suggest further studies to optimize deposited copper concentrations with respect to an improved bactericidal activity. In addition, further research is required to understand the correlation between antibacterial effect and copper release rate.\n\n3. Materials and Methods {#sec3-materials-09-00274}\n========================\n\n3.1. Polymeric Material {#sec3dot1-materials-09-00274}\n-----------------------\n\nThe experiments were carried out on ABS plates of commercial grade with a size of 10 mm \u00d7 10 mm and a thickness of 1.5 mm, which were purchased from Goodfellow GmbH (Bad Nauheim, Germany).\n\n3.2. Plasma Source and Plasma Treatment {#sec3dot2-materials-09-00274}\n---------------------------------------\n\nThe design of the plasma jet (shown in [Figure 9](#materials-09-00274-f009){ref-type=\"fig\"}) is based on a micro hollow discharge geometry that was presented in earlier studies \\[[@B16-materials-09-00274],[@B44-materials-09-00274],[@B45-materials-09-00274]\\]. The jet consists of an inner electrode made from a copper tube with an inner diameter of 0.8 mm and an outer diameter of 1.6 mm. This cathode is isolated from the outer grounded electrode by an alumina ceramic tube with an inner diameter of 1.6 mm and a wall thickness of 0.8 mm. The ground electrode enclosing the ceramic is sealed by screw cap. In the center of the cover is a small hole with a diameter of 0.8 mm through the 0.4 mm thick lid. The ceramic tube is 0.4 mm longer than the inner high voltage electrode, which keeps it separated from the cover.\n\nPlasma is generated in the gap between the inner and outer electrode when a negative voltage of 2 \u00b1 0.2 kV from a DC high voltage power supply (PS FX06R50, Glassman High Voltage Inc., High Bridge, NJ, USA) is applied. A resistor of 50 k\u2126 connected in series limits the current (see [Figure 10](#materials-09-00274-f010){ref-type=\"fig\"}). For the experiments, the current was adjusted to 30 mA. Voltage across the 50 k\u03a9 resistor was monitored with a high voltage probe (P5100A, Tektronix, Beaverton, OR, USA). From current and voltage drop across the resistor, the power dissipated in the plasma is determined to be about 15 W. Compressed air was used as operating gas flowing through the inner copper tube electrode. The gas flow rate was controlled by a mass flow controller (MF1C01324CMAV0, MKS Instruments, Andover, MA, USA) and a four-channel mass flow and pressure display unit (647C-4-R-1-N, MKS Instruments). With a gas flow of 8 slm, the plasma is expelled from the 0.8 mm orifice in the cap generating an afterglow plasma plume of 5--10 mm. The temperature of the plasma between the electrodes is about 1500 K, which is the approximated value for the gas temperature in a micro hollow cathode discharge in air at atmospheric pressure \\[[@B46-materials-09-00274]\\]. However, due to the air flow, the plasma plume temperature at 10 mm distance in axial direction from the orifice is already close to room temperature. This gas temperature of the expelled plume was measured with an optical interference thermos-sensor (UMI 4, FISO Technologies, Quebec, QC, Canada). For the experiments, the plasma jet device was fixed along the *z*-axis of a stepper motor controlled xyz-stage (FF500/CNC, Proxxon GmbH, F\u00f6hren, Germany) at a distance of 10 mm to the ABS samples. For a uniform exposure, the jet was moving in a meandering way, covering an area of 20 mm \u00d7 20 mm (with the sample sitting in the middle) across the sample twice. The second meandering route was perpendicular to the first. To vary treatment times, the stepper-motor speed and the distance between the passes of the meander can be changed.\n\n3.3. Surface Analysis {#sec3dot3-materials-09-00274}\n---------------------\n\nThe chemical surface composition was determined by high-resolution scanning XPS. Spectra were acquired using an Axis Ultra DLD electron spectrometer (Kratos Analytical, Manchester, UK) with a monochromatic Al K\u03b1 source (1486 eV). Charge neutralization was implemented by low energy electron injected into the magnetic field of the lens from a filament located directly atop the sample. The instrument was set to the medium magnification (field of view 2) lens mode and by selecting the slot mode, providing an analysis area of approximately 250 \u00b5m in diameter. Survey spectra and core level spectra of all identified elements were collected at a pass energy of 80 eV and for the highly resolved measured carbon (C 1s) peak the pass energy was set to 10 eV. Three different spots on each specimen were analyzed twice and averaged. Data acquisition and processing were carried out using CasaXPS software, version 2.14dev29 (Casa Software Ltd., Teignmouth, UK). Due to sample charging, the binding energy scale was corrected for all samples by setting the C 1s binding energy to 285.0 eV for the C--(H, C) component in the C 1s spectra. Concentrations are given in atomic percent (at. %).\n\nAFM was performed with a scanning probe microscope diCP-II (Veeco Instruments, Santa Barbara, CA, USA) in the non-contact mode, especially tapping mode. An area of 5 \u00b5m \u00d7 5 \u00b5m and 2.5 \u00b5m \u00d7 2.5 \u00b5m was scanned using a pyramidal silicon tip doped with n-type phosphorus with a resonance frequency of 273--389 kHz and a force constant of 20--80 N/m (Veeco, RTESPA-CP, Mannheim, Germany). AFM-images were recorded on selected samples and processed using the open-source software Gwyddion 2.39 \\[[@B47-materials-09-00274]\\].\n\n3.4. Antimicrobial Test {#sec3dot4-materials-09-00274}\n-----------------------\n\nFor microbiological experiments, *Staphylococcus aureus* DSM 799 (ATCC 6538) was used in concentrations of 10^7^ cfu/mL in sodium chloride solution (0.85%). *S. aureus* was obtained from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). The investigated specimen was UT RowePaedM SA 200 mm bulk/unsterile (Company RoweMed, Parchim, Germany) from ABS covered with copper as described above. The specimen was inoculated with 50 \u00b5L of the *S. aureus* suspension by pipetting. Afterwards, a drying step of 2 h under aseptic conditions followed. The recovery was done with nutrient solution (tryptic soy, Merck, Darmstadt, Germany). The specimens were given in 10 mL nutrient broth and shaken for 15 min at 350 rpm at room temperature. The recovery of survivors was realized using the surface-spread-plate count method with tryptic soy agar plates, and it was completed with an overnight cultivation in an incubator at 37 \u00b0C. The surface-spread-plate count method is a surface counting method employed for aerobic bacteria. For the test, 100 \u00b5L of all serial dilutions of the broth were plated out on the whole surface-area of the petri dish. Serial dilutions were performed as a 1 in 10 dilution. The detection limit of this procedure was 10 cfu/mL. If the number of microorganisms fell below the detection limit, *i.e.*, no viable microorganisms have been found, these values in the graphs were set at the detection limit.\n\n4. Conclusions {#sec4-materials-09-00274}\n==============\n\nThe objective of this work was to evaluate the bactericidal properties of polymer surfaces modified by copper deposits from a cold atmospheric pressure plasma. XPS analysis revealed the surface functionalization of ABS along with the deposition of copper oxide films. It was shown that the deposited Cu concentration strongly depends on the input current (*i.e.*, energy dissipated in the plasma) and on the air flow rate. AFM images confirmed the deposition of copper particles at the nanoscale size. The described films showed their potential use as biocidal coating, which might be of great importance for the reduction of microbial load on inanimate surfaces to prevent, for instance, the transfer of infections in health care facilities and to protect materials from microbial degradation. Furthermore, since this is an environmentally friendly process for copper thin film deposition combined with low temperatures, it is compatible with various materials, including textiles and weaving, which provides advanced properties for tailored applications. Further studies will be focused on the deposition of copper-rich films on different materials.\n\nThe authors are grateful for financial support from the Ministry of Economics of the Federal State Mecklenburg Vorpommern (Grant No. V-630-00INP-20144/018).\n\nJana Kredl, Katja Fricke and Juergen F. Kolb conceived and designed the experiments; Jana Kredl, Uta Schnabel, and Katja Fricke performed the experiments; Jana Kredl, Uta Schnabel and Katja Fricke analyzed the data; Jana Kredl, Juergen F. Kolb, Uta Schnabel, Martin Polak, Klaus-Dieter Weltmann, and Katja Fricke wrote the paper.\n\nThe authors declare no conflict of interest.\n\nThe following abbreviations are used in this manuscript:\n\nABS\n\nacrylonitrile butadiene styrene\n\nDC\n\ndirect current\n\nXPS\n\nX-ray photoelectron spectroscopy\n\nAFM\n\natomic force microscopy\n\nEPA\n\nEnvironmental Protection Agency\n\nPVD\n\nphysical vapor deposition\n\nCVD\n\nchemical vapor deposition\n\nslm\n\nstandard liter per minute\n\nS. aureus\n\nStaphylococcus aureus\n\n![Survey spectra of pristine acrylonitrile butadiene styrene (ABS) and of ABS exposed to air plasma (operating parameters: *Q*~air~ = 8 slm, *I* = 30 mA, *t* = 432 s).](materials-09-00274-g001){#materials-09-00274-f001}\n\n![Cu 2p core level XP spectrum of ABS exposed to air plasma (operating parameters: *Q*~air~ = 8 slm, *I* = 30 mA, *t* = 432 s).](materials-09-00274-g002){#materials-09-00274-f002}\n\n![X-ray photoelectron spectroscopy (XPS) results: Atomic concentration (at. %) of carbon (C); copper (Cu); nitrogen (N); and oxygen (O) depending on treatment time and applied input current. Results are presented as averaged value from three measurements for each specimen. (Operating parameter: *Q*~air~ = 4 slm).](materials-09-00274-g003){#materials-09-00274-f003}\n\n![XPS results: Atomic concentration (at. %) of carbon (C); copper (Cu); nitrogen (N); and oxygen (O) depending on applied input current and gas flow rate. Results are presented as averaged value from three measurements for each specimen.](materials-09-00274-g004){#materials-09-00274-f004}\n\n![XPS high resolution C 1s spectra and N 1s core level spectra of pristine ABS and of ABS exposed to air plasma (operating parameters: *Q*~air~ = 8 slm, *I* = 30 mA, *t* = 432 s).](materials-09-00274-g005){#materials-09-00274-f005}\n\n![Atomic force microscopy (AFM) images (5 \u00b5m \u00d7 5 \u00b5m and 2.5 \u00b5m \u00d7 2.5 \u00b5m) and height profiles of pristine ABS and of ABS after plasma treatment for different air flow rates (operating parameters: *I* = 30 mA, *t* = 432 s).](materials-09-00274-g006){#materials-09-00274-f006}\n\n![Cumulative height distribution plots (**left**) of the topographical images shown in [Figure 6](#materials-09-00274-f006){ref-type=\"fig\"}. The plots are normalized to show probability values from 0 to 1; (**right**) Corresponding histograms of pristine ABS and of ABS after plasma treatment obtained at different air flow rates (operating parameters: *I* = 30 mA, *t* = 432 s).](materials-09-00274-g007){#materials-09-00274-f007}\n\n![Effect of air flow rate and input current on the inactivation of *Staphylococcus aureus*, which is presented in percentage reduction. Values are the average of triplicate. The number of colony forming units on pristine ABS was *N*~0~ = 1 \u00d7 10^7^, and the detection limit was 99.9999%.](materials-09-00274-g008){#materials-09-00274-f008}\n\n![(**Left**) Schematic design of the micro hollow discharge geometry of the DC plasma jet; (**Right**) Surface temperature of ABS during plasma treatment.](materials-09-00274-g009){#materials-09-00274-f009}\n\n![Equivalent circuit diagram of the DC plasma jet.](materials-09-00274-g010){#materials-09-00274-f010}\n"}
+{"text": "Background\n==========\n\nSuccessful management of chronic disease in routine practice is an elusive task \\[[@B1],[@B2]\\]. As the number of patients with chronic medical illness and the literature regarding their optimal management have grown, attempts have been made to improve their care by implementing new models of care delivery. Implementation of these models involves intervening in specific ways in clinical settings and organizations, and each model has organizational elements that are considered necessary for model implementation. For example, in the chronic care model, delivery system design and information systems are among the elements that are specifically identified \\[[@B3]\\].\n\nThe number of interventions on an organizational level to improve delivery of care and outcomes for patients with chronic disease has grown. However, these care models and organizational strategies have not met with uniform success \\[[@B4]-[@B14]\\]. We believe that an important reason for this variation in outcomes is that interventions do not adequately take into account the characteristics of clinical systems in which patients receive care.\n\nClinical microsystems are the building blocks of healthcare delivery: the individual clinics, units, or other areas where care is delivered. The complex adaptive system (CAS) framework has been applied to clinical microsystems as a theoretical model for better understanding them \\[[@B15]-[@B22]\\]. This framework suggests that clinical settings are environments in which individuals learn, inter-relate, self-organize, and co-evolve in response to changes in their internal and external environments, in turn shaping those environments \\[[@B15],[@B19]\\]. Because inputs and outputs in CAS may not be proportional or predictable, interventions that are successful in one setting may not be successful in another. However, evidence suggests that interventions congruent with the CAS framework and characteristics are in general more likely to be effective \\[[@B21]-[@B25]\\].\n\nThe insight that clinical settings are CASs is important to the field of implementation research, as it provides guidance for how to approach disseminating research findings into routine care. The CAS framework suggests that local contexts and local interactions between individuals are critical considerations in designing interventions, and that leveraging these may lead to improvements in system performance. However, we wanted to expand on this insight by exploring the possibility that interventions must also be congruent with the nature of the disease or diseases of the patients being cared for. Diseases may mediate the way that interventions influence a patient\\'s care. The level of complexity of different diseases, and the ways that chronic diseases impact patients\\' lives, varies greatly depending on the type of disease, leading to the need for different approaches. For example, self-monitoring in diabetes may be more difficult than congestive heart failure (CHF) because it may be more difficult to monitor carbohydrates and calories than salt, and involves the pain of fingersticks versus standing on a scale to check weight. Therefore, the most effective patterns of communication between patients with type 2 diabetes and their providers may be different from those for patients with CHF, which in turn may affect the way that providers and staff interact and structure the delivery of care for those groups of patients. Because of this, we believe that for interventions whose goal is to improve the performance of a clinical system to be most effective, they must take into account not only the nature of the system, but also the nature of the disease.\n\nThe purpose of this paper is twofold. First, we build on the literature suggesting that interventions consistent with CAS are more likely to be effective \\[[@B22]\\] by conducting a systematic review of organizational interventions focused on improving care of patients with CHF. This work builds on our previously published systematic review of interventions to improve outcomes for patients with Type 2 diabetes, expanding the data regarding the importance of considering health care settings as CAS beyond a single chronic disease. Like diabetes, CHF is a common disease whose management is broadly relevant. We also chose CHF because of the growing number of studies of interventions to improve CHF outcomes through changing the way that care is delivered in clinical settings.\n\nOur second purpose is to compare the findings of the specific types of interventions that appear to be most effective for CHF and diabetes, to identify differences in the specific CAS characteristics associated with more effective outcomes for each disease. We hypothesized that there are fundamental differences between CHF and diabetes in terms of their impact on patients\\' day-to-day lives, the behaviors that are required for their successful management, and the structure of care delivery that best supports successful management.\n\nMethods\n=======\n\nSystematic review of organizational interventions for CHF\n---------------------------------------------------------\n\nOur methods mirrored those in our previously published systematic review of organizational interventions to improve care of patients with type 2 diabetes \\[[@B22]\\]. Specific elements are as follows.\n\nSearch strategy\n---------------\n\nWe defined organizational interventions as those that explicitly attempt to affect or change organizational structures or processes to implement evidence-based practice. Our search strategy was based on four components: the strategy developed by the Effective Practice and Organization of Care (EPOC) Group of the Cochrane Collaboration \\[[@B26]\\]; additional search terms for types of organizational interventions not included in EPOC, such as total quality improvement, PDSA (Plan-Do-Study-Act), and practice redesign; additional search terms identified in recent systematic reviews of quality improvement initiatives; and bibliographies and Medline indexing terms of relevant publications.\n\nTo focus the search on CHF, we added disease-specific MeSH and text word terms, ran a preliminary search, and reviewed 2,559 titles and abstracts (determined by saturation, until no further new terms were identified), for additional text word terms. The search terms are listed in Additional File [1](#S1){ref-type=\"supplementary-material\"}. We did not search the management literature, nor did we seek out unpublished data. We searched Medline from 1989 through 17 July 2008.\n\nInclusion and exclusion criteria\n--------------------------------\n\nWe included randomized, quasi-randomized, or controlled clinical trials published in English and conducted in economically developed countries identified as such by the International Monetary Fund or the Organization for Economic Cooperation and Development \\[[@B27]\\]. We excluded non-English articles, with the rationale that non-English studies comprise only 1% of the EPOC registry. Because our goal was to understand interventions in routine outpatient practice, and to have uniformity in the types of settings included, we excluded studies conducted in nursing home or palliative care settings. To focus on the impact of interventions of process of care or patient outcomes, we excluded studies reporting only the following non-clinical outcomes: patient or provider knowledge; self-efficacy; satisfaction; or other attitudes and beliefs. To minimize heterogeneity among study populations, we excluded studies of *cor pulmonale*patients exclusively. Finally, to focus on interventions that attempted to improve care by changing the organizations or settings in which care was delivered, we excluded: care pathway interventions without organizational components (*e.g*., patient or provider education only); work site health interventions; exercise rehabilitation or diet only; and disease prevention or screening only.\n\nFour investigators independently reviewed overlapping groups of differing halves of the citations\\' titles and abstracts generated by the full literature search to assess agreement regarding potentially eligible publications. If eligibility was uncertain after review of the title and abstract, the full article was reviewed. Eligible studies were independently reviewed and jointly abstracted in detail by teams of two investigators. Disagreements were resolved by consensus of the group of investigators.\n\nAssessment of leveraging of characteristics of CAS\n--------------------------------------------------\n\nEligible publications of organizational interventions as defined by the inclusion and exclusion criteria were then independently evaluated by two raters with content expertise in complexity science to assess the extent to which each reported intervention utilized the following four recognized CAS characteristics \\[[@B15],[@B19],[@B22]\\]: individuals\\' capacity/ability to learn; the interconnections between individuals; the ability of participants to self-organize; and the tendency of participants to co-evolve. Each intervention was given a point for each of the characteristics present in the study design, for a possible lowest score of 0 and highest score of 4. If a study contained more than one intervention, each was assessed separately. Definitions of each characteristic, along with examples of specific interventions felt to reflect each characteristic, are summarized in Table [1](#T1){ref-type=\"table\"}. An example of an intervention felt to include all four CAS characteristics included the addition of a nurse practitioner-led clinic (changing the interconnections between patients and providers), protocol development, and patient education (learning). Patients received individualized feedback (self-organization), and the frequency of visits and type of feedback changed depending on the patient\\'s progress or symptoms (self-organization and co-evolution). An intervention that included only one CAS element was one in which data from a one-time patient survey was used to generate standardized care suggestions embedded within an electronic health record (only assessed as changing interconnections among patients and providers by adding a new contact point). Additional File [2](#S2){ref-type=\"supplementary-material\"} contains detail of each reported intervention and its CAS characteristic rating. The raters were blinded to the outcomes of the studies. The kappa for these scores between reviewers was 0.84, with conflicts subsequently resolved by discussion.\n\n###### \n\nCharacteristics of Complex Adaptive Systems Abstracted\n\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Characteristic      Definition                                                                               Example\n  ------------------- ---------------------------------------------------------------------------------------- -----------------------------------------------------------------\n  Agents who Learn    \u2022 People can and will process information, as well as react to changes in information    \u2022 \\'Health Buddy\\' with educational content\\\n                                                                                                               \u2022 Teach guidelines\n\n                                                                                                               \n\n  Interconnections    \u2022 Change in pattern of interactions, including non-verbal communication, among agents\\   \u2022 Letters to patients\\\n                      \u2022 Introducing new agents into the system                                                 \u2022 Nurse-led heart failure group clinic\\\n                                                                                                               \u2022 Clinical reminders\n\n  Self-organization   \u2022 Order is created in a system without explicit hierarchical direction                   \u2022 Flexibility in tailoring intervention to individual patients\n\n  Co-evolution        \u2022 The system and the environment influence each other\\'s development                     \u2022 Individualized \\'HOME\\' treatment plan that changes over time\n  ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nAssessment of reported outcomes\n-------------------------------\n\nBecause of the great heterogeneity among reported outcomes, we did not use effect size as the outcome variable. Instead, we used a rating scale to assess the effectiveness of the intervention. The outcomes of each study were rated by two independent reviewers on a scale of 0 (no effect), 0.5 (mixed results), and 1 (intervention effective) based on the type (process versus outcome), number, and statistical significance of outcomes reported. Table [2](#T2){ref-type=\"table\"} summarizes the criteria for each rating category, as well as provides examples of outcomes felt to reflect each category of effectiveness. Reviewers were blinded to study intervention, author, and title of manuscript, and one outcome rater was different from the intervention raters. The kappa for these scores was 0.86 with conflicts resolved by discussion.\n\n###### \n\nCriteria used to classify intervention effectiveness, with examples of outcomes reflecting each level of effectiveness\n\n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  Outcome Score   Criteria                                                                                                                                       Example\n  --------------- ---------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------\n  0               \u2022 No statistically differences between control and intervention groups, or between intervention and baseline, on process or outcome measures   \u2022 No difference in adherence, NYHA class, \\# visits, or \\# hospitalizations\n\n                                                                                                                                                                 \n\n  0.5             \u2022 Trends without significance\\                                                                                                                 \u2022 Significant improvement in adherence, trends for CHF-related admission and total number of hospital days\n                  \u2022 Mixed outcomes (significant improvement in minority of measures)\\                                                                            \n                  \u2022 Significant improvement compared with baseline, but not with control                                                                         \n\n                                                                                                                                                                 \n\n  1               \u2022 Statistically significant improvement:\\                                                                                                      \u2022 Significant reduction in all-cause mortality and all-cause admissions at 3 months\n                  -all outcomes if \u22643 endpoints\\                                                                                                                 \n                  -majority of outcomes if \\> 3 endpoints                                                                                                        \n  ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nStatistical analysis of association between CAS characteristics and intervention effectiveness\n----------------------------------------------------------------------------------------------\n\nWe used Fisher\\'s exact test to test the significance of the relationship between total number of characteristics of CASs utilized in an intervention and the strength of outcomes reported, as well as between each individual characteristic and the strength of outcomes. Because a mismatch between the unit of allocation and analysis may bias a study towards positive results, we divided studies into two groups based on whether a unit of analysis error was present. A second analysis using Fisher\\'s exact test was performed including only those studies that did not contain a unit of analysis error. Finally, a third analysis using logistic regression was performed to weight studies based on both sample size and duration of intervention, with outcome rating as the independent variable, and CAS score, sample size, and intervention duration as independent predictors.\n\nAll statistical analysis was performed using Stata 9.2 (College Station, Texas).\n\nAnalysis contrasting CAS characteristics associated with effectiveness for CHF and diabetes\n-------------------------------------------------------------------------------------------\n\nThe relationship between each individual CAS characteristic utilized in an intervention and strength of reported outcomes was assessed, using Fisher\\'s exact test with and without intervention with unit of analysis error, as well as logistic regression. The individual characteristics associated with intervention effectiveness for CHF were compared with those previously reported for diabetes \\[[@B22]\\].\n\nResults\n=======\n\nSystematic review of organizational interventions for CHF\n---------------------------------------------------------\n\nOur search identified 2,510 publications for CHF. Of those, 112 were potentially eligible based on review of title and abstract, and were fully reviewed by a team of two investigators. Based on this review, 44 articles were eligible for inclusion \\[[@B28]-[@B71]\\]. Figure [1](#F1){ref-type=\"fig\"} details the numbers of articles eligible and ineligible at each stage of review. Seven studies had unit of allocation error, in which the unit of randomization was either the provider or clinic, but the unit of analysis was the patient. Two studies reported two distinct interventions \\[[@B50],[@B59]\\]; each intervention was analyzed separately for a total of 46 interventions across 44 studies. The CAS scores for each study are contained in Additional File [2](#S2){ref-type=\"supplementary-material\"}.\n\n![**Articles eligible and ineligible at each stage of review**.](1748-5908-5-66-1){#F1}\n\nOnly 13 interventions out of 46 (28%) received a rating of 1 for outcomes through demonstrating significant improvement in most or all reported outcomes; all others were felt to have mixed or negative results. All interventions incorporated at least one CAS characteristic, with 41% utilizing two CAS characteristics, 28% utilizing three, and 24% utilizing four. Ninety-three percent of reported interventions were judged to change the pattern of interconnections between individuals, typically through the introduction of a new person such as a case manager. Thirty-five interventions (76%) impacted learning; sixteen (35%) allowed self-organization of study participants; and in thirty (65%), the intervention evolved over time based on factors such as the patient\\'s status or symptoms.\n\nThe association between number of CAS characteristics leveraged in an intervention and its effectiveness is shown in Table [3](#T3){ref-type=\"table\"}. None of the studies utilizing only one or two characteristics demonstrated significant improvement in most or all outcomes. All studies utilizing three or four CAS characteristics demonstrated at least mixed results, and ninety-one percent of those using all four CAS characteristics received the highest rating of effectiveness based on having demonstrated statistically significant improvement in most or all outcomes. This association between number of CAS characteristics utilized and the effectiveness of an intervention was significant (p \\< 0.001), and remained so after studies with unit of allocation error were excluded (p \\< 0.001). This association also remained significant in logistic regression analysis, adjusting for sample size and intervention duration (p \\< 0.001).\n\n###### \n\nDistribution of CAS and intervention effectiveness for CHF studies\n\n  -------------------------------------------------------------------------------------------------------------\n  Total CAS\\   Rating of Intervention Effectiveness   Total No. Studies with each CAS Score            \n  Score                                                                                                \n  ------------ -------------------------------------- --------------------------------------- -------- --------\n               **0**                                  **0.5**                                 **1**    \n\n  **1**        1                                      2                                       0        3\n\n  **2**        6                                      13                                      0        19\n\n  **3**        0                                      10                                      3        13\n\n  **4**        0                                      1                                       10       11\n\n  **Total**    **7**                                  **26**                                  **13**   **46**\n  -------------------------------------------------------------------------------------------------------------\n\nP \\< 0.001\n\nThree individual CAS characteristics were associated with CHF intervention effectiveness: learning (p = 0.05), self-organization (p \\< 0.001), and co-evolution (p = 0.002). These associations remained significant after excluding studies with unit of analysis error. The association between interconnections and effectiveness was not significant (p = 0.72). The detail of analysis for individual CAS characteristics and intervention effectiveness is shown in Additional File [3](#S3){ref-type=\"supplementary-material\"}.\n\nAnalysis contrasting CHF and diabetes\n-------------------------------------\n\nA systematic review of interventions to improve care of patients with diabetes through changing the way health care organizations delivered care was previously published. The methods for that review were identical to those reported here with the exception of the disease-specific search terms used and 32 studies were identified. In that review, the presence of CAS characteristics and effectiveness of interventions were also assessed, and the association between the two was performed with Fisher\\'s exact test, with and without studies with unit of allocation error. We used those data to compare the differences between the individual CAS characteristics associated with intervention effectiveness for CHF and diabetes.\n\nThe association of individual CAS characteristics with intervention effectiveness between studies of organizational interventions for patients with type 2 diabetes, and those with CHF is shown in Table [4](#T4){ref-type=\"table\"}. Only the CAS characteristic \\'co-evolution\\' was significantly associated with intervention effectiveness in both diabetes and CHF. Learning and self-organization were associated intervention effectiveness for CHF, and interconnections were associated with intervention effectiveness for diabetes.\n\n###### \n\nAssociation between individual CAS characteristic and intervention effectiveness for type 2 diabetes \\[[@B22]\\] and CHF\n\n  CAS characteristic   type 2 diabetes^22^   CHF               \n  -------------------- --------------------- ----------- ----- ------------\n  Learning             80%                   p = 0.07    76%   p = 0.05\n                                                               \n  Interconnections     77%                   p = 0.05    93%   P = 0.72\n                                                               \n  Self-organization    27%                   p = 0.58    35%   p \\< 0.001\n                                                               \n  Co-evolution         70%                   p = 0.003   65%   p = 0.002\n\nDiscussion\n==========\n\nThis systematic review of interventions to improve outcomes of patients with CHF through changing care delivery processes in clinical settings is consistent with literature reporting that interventions that attempt to improve patient outcomes through impacting the organizations in which care is delivered have mixed results \\[[@B2],[@B4]-[@B6]\\]. In this review, the majority (72%) of interventions were not effective in significantly improving outcomes. Our analysis of these interventions through the lens of a CAS perspective again demonstrates that interventions consistent with a CAS perspective are more likely to be effective in improving outcomes. We interpret this as providing further evidence that the clinical settings are CASs. For interventions to be effective in improving patient outcomes, they must take this into account.\n\nThe difference in the individual CAS characteristics associated with intervention effectiveness for patients with type 2 diabetes and CHF brings a new perspective to the consideration of clinical settings as CASs. Specifically, we believe that when implementing interventions to improve outcomes of patients with chronic disease, not only must the characteristics of the organization be considered, but so must the characteristics of the disease and its treatment. Interventions must be appropriately matched to the level of complexity of not only the organization, but also of the disease, as disease and treatment characteristics may influence what interventions are more likely to be effective.\n\nThe interplay of differences between chronic diseases within the context of CAS clinical systems will affect which approaches are more or less likely to be effective for patients with a specific disease. The level of uncertainty inherent in diseases and their treatments may be an important contributor to these differences. To illustrate this point, Table [5](#T5){ref-type=\"table\"} outlines how potential differences between type 2 diabetes and CHF in terms of uncertainty may influence the CAS characteristics that were associated with intervention effectiveness for each disease. For example, the myriad combinations of lifestyle and medication approaches to managing type 2 diabetes may be more complex and nuanced than those for CHF, and the symptoms of worsening glycemic control may be more subtle and insidious than worsening volume status, leading to greater uncertainty in the management of diabetes relative to CHF \\[[@B72]\\]. This greater uncertainty may influence the effectiveness of CAS characteristics in interventions, or the effectiveness of combinations of characteristics. Because one way that individuals can navigate uncertainty is through relationships, interconnections may be particularly important with increasing uncertainty.\n\n###### \n\nPotential differences between type 2 diabetes and CHF with regards to uncertainty, and how they might influence CAS characteristic effectiveness\n\n  CAS characteristic   type 2 diabetes                                                                                                                         CHF\n  -------------------- --------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------\n  Learning             Treatment is nuanced and complex, making efforts to improve outcomes through learning more difficult.                                   Less uncertainty in treatment guidelines allows more prescriptive, algorithmic approaches to management that may be more easily learned.\n                                                                                                                                                               \n  Interconnections     Greater degree of uncertainty in terms of symptoms and management, leading to greater reliance on interconnections to manage disease.   Lesser degree of uncertainty in terms of symptoms and management may lead interventions focused on interconnections less effective.\n                                                                                                                                                               \n  Self-organization    Greater uncertainty in management and symptoms of exacerbation may make efforts to self-organize more difficult.                        Less uncertainty regarding management and symptoms of exacerbation may make efforts to self-organize more effective.\n                                                                                                                                                               \n  Co-evolution         Course and symptoms evolve over time in unique trajectory.                                                                              Course and symptoms evolve over time in unique trajectory.\n\nHowever, for both CHF and type 2 diabetes, patients have a chronic disease that is changing over time, and the recognition of the dynamic nature of the evolution of disease in interventions is important. Also in both cases, the fact that the clinical settings in which patients receive care are CASs is an important contextual consideration, as no two are exactly alike.\n\nThe implication of these findings for implementation research whose goal is to change organizations to improve care of patients with chronic disease is that we must shift our focus in intervention design. While considerations such as cost, ease of implementation, and level of disruption to the clinical setting are important, the levels of complexity of the organization and the disease are even more important. Intervention design for chronic disease requires a greater level of nuance, individualization, flexibility, and assessment over time. Specific implications of this insight include the need to pay attention to or explicitly change the relationships between individuals as a strategy to improve outcomes, the importance of allowing \\'local\\' input or control into the intervention design, and the need to provide feedback regarding the impact of the intervention and the possibility to change the intervention based on this feedback.\n\nThis study has several limitations. The first is the relatively small number of studies of organizational interventions. However, despite these small numbers, the associations found are significant ones. A more overarching limitation is the difficulty in applying the lens of a CASs perspective to traditional intervention design, with specific regard to assigning scores regarding CAS characteristics retrospectively. Our method of independent review of interventions and results using groups of separate reviewers was intended to offset this methodological limitation, and our kappa scores suggest that reviewers did have a consistent ability to make these retrospective assessments.\n\nOther limitations include the possibility of publication bias, which may have led either to negative studies not being published, or to the interventions being described in less detail, making assessment of CAS characteristics more difficult. Negative studies are well represented in the distribution of outcomes in the included studies, and the inter-rater consistency suggests that sufficient information was available to make an assessment. All raters are from the same institution, and the possibility of bias in terms of the CAS characteristics on which we focused is possible, but made less likely by our use of characteristics consistently recognized as key in the CAS literature.\n\nFinally, this analysis is limited to studies of patients with type 2 diabetes and CHF. The applicability of our findings to other chronic diseases, or to acute disease processes, has yet to be demonstrated. However, we believe that the strength of our results across a combined 76 examples of interventions (46 CHF and 32 diabetes) for two distinct chronic diseases is at least suggestive of the range of application of the CAS framework in clinical settings.\n\nSummary\n=======\n\nThe significant association between CAS characteristics and effectiveness of reported outcomes for patients with CHF builds on the idea that for interventions to be effective, they must be consistent with the CAS nature of the clinical systems in which they are applied. However, the fact that different CAS characteristics are associated with intervention effectiveness for CHF and type 2 diabetes suggests that the nature of the diseases being treated may mediate the intervention effectiveness. We suggest that the level of uncertainty associated with a disease and its treatment may be an important characteristic that must be considered in designing interventions to lead to the greatest improvement in patient outcomes.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nLKL conceived this analysis using the database conceived by VL, PN, and JP, rated studies, performed preliminary statistical analysis, interpreted findings, and drafted the manuscript. MP rated studies, interpreted findings, and helped to draft the manuscript. JP participated in the design of the study and helped to draft the manuscript. VL conceived the systematic review and database, rated studies, and helped to draft the manuscript. PN conceived the systematic review and database and helped to draft the manuscript. RMcD participated in the design of the study, provided theoretical expertise, interpreted findings, and helped to draft the manuscript. All authors have read and approved the final manuscript.\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\n**Search strategy to identify studies of organizational interventions to improve outcomes for patients with congestive heart failure**. Search completed 17 July 2008. Additional file [1](#S1){ref-type=\"supplementary-material\"} is a word document detailing the keywords and number of results identified by each keyword used in our search strategy.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\n**Summary of eligible studies of organizational interventions on outcomes of patients with congestive heart failure**. Additional file [2](#S2){ref-type=\"supplementary-material\"} is a word document listing each eligible study, along with details regarding sample size, our ratings of its intervention and the number of CAS characteristics leveraged, follow-up duration, and presence or absence of unit of analysis error.\n\n###### \n\nClick here for file\n\n###### Additional file 3\n\n**Detail of analysis for individual CAS characteristics and intervention effectiveness for CHF**. Additional file [3](#S3){ref-type=\"supplementary-material\"} is a table in word document format that lists each CAS characteristic, the number of studies in which the characteristic was utilized, and the range of intervention effectiveness scores for those studies.\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nThe research reported here was supported by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service (TRX \\# 01-091 & REA 05-129). Investigator salary support is provided through this funding, and through the South Texas Veterans Health Care System. Dr. McDaniel receives support from the IC^2^Institute of the University of Texas at Austin. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs.\n"}
+{"text": "Summary of Chapters {#Sec1}\n===================\n\nThroughout the book chapters, researchers have highlighted the recent advancement in microfluidic areas, particularly those involving microdroplets.\n\nSimon and Lee focused on microfluidics droplet manipulations and applications, including droplet fusion, droplet fission, mixing in droplets, and droplet sorting. By combining these operations, they have shown promising applications in executing chemical reactions and biological assays at the microscale.\n\nDay and Karimiani discussed dropletisation of bio-reactions.\n\nZhang and Liu elaborated the physics involved in multiphase flows and microdroplets dynamics. They emphasized the important dimensionless parameters relating to droplet dynamics with droplet generation process as an example.\n\nBarber and Emerson discussed the fundamental droplet handling operations and the recent advances in electrowetting microdroplet technologies. They also provided an overview of droplet-based electrowetting technologies in biological and chemical applications.\n\nDroplet-based microfluidics as a biomimetic principle in diagnostic and biomolecular information handling were highlighted by K\u00f6hler addressing potential of applying segmented fluid technique to answer to the challenges of information extraction from cellular and biomolecular systems.\n\nUsing the flow rates, applied pressures, and flow rate ratios in a closed feedback system, the active control of droplet size during formation process in microfluidics was achieved by Nguyen and Tan.\n\nVelev, Petsev, and Chang discussed droplet microreactors for materials synthesis. They briefly described microfluidics for droplet generation as well as fabrication technology. They provided detail study of transport in microchannels and droplet microfluidics for mesoporous particle synthesis.\n\nKaminski, Churski, and Garstecki reviewed the recent advances in building modules for automation of handling of droplets in microfluidic channels, including the modules for generation of droplets on demand, aspiration of samples onto chips, splitting and merging of droplets, incubation of the content of the drops, and sorting.\n\nZagnoni and Cooper have demonstrated the use of on-chip biocompatible microdroplets both as a carrier to transport encapsulated particles and cells, and as microreactors to perform parallel single-cell analysis in tens of milliseconds.\n\nGeneral Situation {#Sec2}\n=================\n\nHere we try to explore the technology development cycle and market trend for microfluidics devices. Microfluidic systems were first pioneered by Stanford's research introducing a chromatography chip about 30 years ago \\[[@CR1]\\]. It was probably too ahead of time, yet only 15 years later, an avalanche of microfluidics developments was triggered by Manz's group \\[[@CR2]\\] introduction of on-chip capillary electrophoresis (CE). This technology went through a Gartner hype cycle as illustrated in Fig.\u00a0[10.1](#Fig1){ref-type=\"fig\"}. Manz's CE chip resulted in a technology trigger to lead to inflated expectations in the late nineties for microfluidics, mirroring the Silicon Valley Technology bubble hype. Since then, there have been thousands of researchers developing microfluidic systems for various applications and with different goals. \\[[@CR3]\\] Some were interested in basic research, some in commercial applications. However, very few of them were commercially successful in finding the ground-breaking applications. Microfluidics failed to deliver the initial promises to provide a revolutionary technology platform for life sciences and hence disappointed investors. So far, the most successful droplet microfluidics device is the inkjet printer; the commercialization of other miniaturization technology remains highly attenuated even though some areas have made good progress, such as Caliper's LabChip. Why is it that with such tremendous effort there is so little outcome? Let's analyze the reasons for the slow adoption of this promising enabling platform technology. We will further discuss if this technology is close to finding the \"holy-grail\" of analytics despite the past disappointing track record.Fig. 10.1Gartner Hype Cycle for Microfluidic Technology. The development of the capillary electrophoretic (CE) chip initially triggered the technology development. An example of a device produced during the peak expectation phase is exemplified by the micro polymerase chain reaction (PCR) system. After multiple disappointments currently the technology has now entered the slope of enlightenment\n\nScientific and Technology Origin {#Sec3}\n================================\n\nThe chosen approach to demonstrate the value of microfludic applications contrary most likely is the major problem. Microfluidic systems have not been developed based on industrial or applications demand. These systems are mostly based on \"leftover\" manufacturing equipments and tools from the semiconductor industry. Using a push-pull analogy, microfluidics systems are \"pushed\" by manufacturers rather than \"pulled\" from market demand. The semiconductor industry follows the well known Moore's law, increasing wafer size, and shrinking device dimensions. The industry constantly needs to invest huge amounts of capital equipment with a short technology advancement cycle. In order not to obsolete the costly equipment, device manufacturers found microelectromechanical systems (MEMS) attractive. It is economical to convert the outdated integrated circuits (IC) production lines to produce MEMS devices such as pressure sensors, accelerometers etc. Meanwhile, integrated MEMS devices are also following Moore's law, although somewhat delayed in comparison to the ICs. Therefore, further converting such production lines to make microfluidic devices becomes the next natural option. The critical dimensions of these devices are well within the capability of existing semiconductor equipment and they are relatively simple to make. They need only a few fabrication steps, with contact printing for lithography often proving to perfectly suffice. The only special tool usually required is the wafer bonder, as well as the availability of etching method for glass. Next comes the basic question: who wants these devices and why? One of the fundamental problems of microfluidic devices not being commercially successful is rooted in the simple fact that they were NOT developed based on market demand, but quite contrarily. Such a starting point was risky as microfluidics development was often used to justify longer lifetime of leftover and aged IC facilities. Then the problem became how to find the applications and market demand for those devices. \"Retrofitting\" is well documented to rarely work.\n\nFortunately, there are now researchers who adopted the right approach. A new age of microfluidics devices for heat exchanging, mixing, and subsequent high performance liquid chromatographic (HPLC) separations are offered for example by Agilent based on the application demand for the device with specific performance in the market place, not to just redeplying old fabrication production line. These microfluidic devices are made of six layers of stainless steel cut by laser and glued together (see Fig.\u00a0[10.2a](#Fig2){ref-type=\"fig\"}). The devices are cheap, reliable, and able to withstand high pressure. To make them more user-friendly, the device extensions for connection can be bent to different angles based on application demand. Also previously a version of the HPLC chip that incorporated sample preparation was made from polyimide using printed circuit board (PCB) technology (see Fig.\u00a0[10.2b](#Fig2){ref-type=\"fig\"}). It enabled integration of heaters to locally control temperature.Fig. 10.2Agilent chip device for (**a**) mixer from 6 layers of stainless steel and (**b**) LC-MS from Polyimide. Both devices are used in commercial products for proteomic mass spectrometry and for ultra high pressure liquid chromatography, respectively\n\nOther fabrication techniques for microfluidics started to emerge, such as polymer-based microfluidics using polydimethylsiloxane (PDMS). The PDMS process is simple and it does not even require a well equipped cleanroom. Nevertheless the material itself is permeable to certain molecules which brings other problems making PDMS devices less competitive. Injection molding and hot embossing (imprinting) are other examples of different approaches compared to employing the silicon wafer processing facilities.\n\nAre there any other problems with microfluidics? Firstly there is a scaling law which predicts problems for quantitative molecular detection limits at the nanometer scale. Every technique has a detection limit requiring a certain number of molecules to be presented. This limit is not altered with the sample size, i.e., very small samples have to be highly concentrated to be exceed the limit of detection. This makes theses samples too concentrated to be of any interest. A restriction to pure compounds, or at the percentage level, seems to be interesting for more academic research only. Optimal fluidic dimensions for practical analytical chemistry look like to be from about 5--50\u00a0\u03bcm. That is a problem but still does not explain why the microfluidics devices are not flooding the market and why they have not \"wiped out\" conventional systems.\n\nExample: PCR on Chip {#Sec4}\n====================\n\nPerhaps we can now analyze one popular microfluidics device as an example: miniaturized polymerase chain reaction system (microPCR). This process was first demonstrated by Northrup in 1993 and since then, hundreds of research groups have been designing their own systems in highly innovative approaches. However, none of them has been commercially introduced. The initial incentive seems very simple: the microPCR needs to be small so that it only requires the use of very small amount of reagents making the PCR economical. Surprisingly, in reality that is not always an advantage even though some researchers like to claim so. PCR is so sensitive that it can detect only a few molecules of DNA or RNA. Smaller amount of reagents indeed brings the cost down but the negative effect is that it decreases the risk of detection reliability through lowering the sample volume. Typically, a sample with volume from 5 \u03bcL to 10 \u03bcL can be used to detect one molecule of DNA. Using 10\u00a0\u03bcL for comparison, if the sample is split into 100 units with 100\u00a0nL each, then on the average the DNA concentration has to be increased 200 times to have a single DNA molecule in each sample. In reality that means that we are losing sensitivity by lowering sample volume making it unsuitable for direct diagnoses of infectious diseases. There are two exceptions, one is digital PCR \\[[@CR4]\\] and the other one is PCR with sample pre-concentration \\[[@CR5]\\]. Digital PCR divides one sample into into hundreds or thousands of tiny wells. It is based on exploiting use of sample dilution so extreme that a significant number of wells will intentionally receive no DNA while others gain a single DNA template to seed the PCR. The count of amplified wells determines the absolute number of DNA molecules in the original sample, making this PCR system intrinsically quantitative. That is an excellent approach and one that is specifically enabling through miniaturization. The only drawback is that for many applications quantitative PCR is not always required, and therefore the digital PCR is often overkill. However should quantitative PCR prove to be necessary, digital PCR could provide the answer.\n\nA second case where the sample can be small is shown in Pipper's work as they run a pre-concentration step prior to PCR itself. His starting volume was only 40\u00a0\u03bcL compared to conventional Qiagen protocol requiring volume of 140\u00a0\u03bcL. Nevertheless he was able to run real-time RT-PCR with only a 100\u00a0nL sample volume while achieving two cycles smaller critical threshold, demonstrating that a small volume of PCR sample can be used for diagnostics without sacrificing the limit of detection.\n\nThis brings us to another problem which is working with clinical samples. These assays typically require binding of active component such as protein or DNA/RNA to achieve immobilization, washing off most of unwanted substances, and eventually release of the active component for further processing. A typical 140\u00a0\u03bcL volume of clinical sample as mentioned before is far too large to fit inside a micromachined microfluidic system. Also some reagents have to be stored separately from each other as well as outside the microfabricated device. If the sample as well as the reagents have to be stored separately (most likely in plastic devices), is there any justification of using the microfabricated device itself? Some researchers believe that the whole system can be produced by injection molding, such as GenExpert from Cepheid \\[[@CR6]\\] which is one of the very few commercially available systems performing fully automated sample preparation followed by real-time PCR. A different approach was taken by Veredus Laboratories. They followed a previously described path of using outdated semiconductor process from ST Microelectronics to make advanced PCR systems with in situ hybridization \\[[@CR7]\\]. The system is more labor intensive than sample-to-answer system such as GenExpert, but it is capable of identifying numerous genes simultaneously, offering advantage when screening for a few closely related pathogenic strains or detection of pathogens for homeland security applications.\n\nEconomical {#Sec5}\n==========\n\nFrom a technology development cycle perspective, there are other reasons why adoption of microfluidics technology is so slow.\n\nReason number one is the lack of economy-of-scale. In order for any technology to take off, it has to reach the tipping point in the market place to inflame the \"viral effect\" that triggers a high volume need; in economic terms, the economy-of-scale has to be in place. Without high volume it is hard to reduce manufacturing cost, and without an affordable price, it is hard for the new technology to be widely adopted. It is known as the \"chasm\" in the technology adoption cycle \\[[@CR8]\\]. It becomes a \"chicken-and-egg\" dilemma. So what are the potential high volume markets? Over the years we have seen increasing rate of adoption of biological research helped by droplet microfluidic devices as tools. Examples of such significant progress are HPLC \\[[@CR9]\\], \"fluidic transistors\" by Cytonix \\[[@CR10]\\], and high throughput screening of biological reactions \\[[@CR11]\\]. Digital microfluidics using \"fluidic transistors\" has potentially wide applications in diagnostic, chemical detection, bio-sequencing and synthesis as well as tissue engineering. The strong growing demand for fast, reliable, repeatable, and cost-effective biological analysis and diagnostic systems has driven the development of such systems. Microfluidic systems have been proven to be an enabling technology platform, benefitting through extensive research performed over years of exploration. However, currently, the devices were individually researched and prototyped by many academic research groups or small commercial groups. Each device has individual fabrication steps and choice of materials. It is lack of a \"standard\" manufacturing process which prevents large scale repeatable production, therefore lack of the momentum of building critical mass towards the tipping point. High capital investment and low return on capital becomes the barrier. In our opinion, the commercialization community of microfluidic systems needs to converge on to adopting \"standard\" materials and manufacturing techniques. Interestingly with digital microfluidics we start to see the genesis of such a trend. Digital microfluidics has become a flexible platform for various bioprocessing and bioanalytical applications.\n\nReason number two is the co-development of supporting and companion technologies such as detection systems. Often, microfluidic devices are not stand alone as ready-to-use systems, they need to be integrated with other devices to form the complete system for given applications. If a technology platform is too ahead of its prime time, it will lack the associated supporting infrastructure, thus it would be suppressed until the companion technologies catch up. For example, in the case of microfluidic diagnostic devices such as micro real-time PCR, there is need for miniature reliable optical sensing devices and signal processing. In the past 10 years, CCD imaging and digital signal processing have made tremendous progress to make fast, reliable, and cost-effective diagnostic system possible.\n\nReason number three is the socio-economic environment. In the past 20 years, the bioscience community focused efforts on finding drugs for treating diseases. Now there is a political-social-economical shift towards early disease diagnoses and prevention to reduce the rapid increase in healthcare burden due to expensive treatment. Microfluidic systems have proven to be critical building blocks for bioanalysis and diagnostic instrumentation, and some of the devices have shown potential to be the consumer product for environmental monitoring and pandemic prevention diagnostic tools \\[[@CR5]\\]. Also, for any technology platform, during the early development stage, there is need for enthusiasm from visionaries and investors. In the past 10 years, the venture capital community shifted investment strategy towards emerging markets, which reduced the early stage technology platform survival rate in developed countries due to lack of funding to turn the corner on the s-curve of innovation life cycle (see Fig.\u00a0[10.3](#Fig3){ref-type=\"fig\"}). Nowadays, the situation starts to turn around.Fig. 10.3Innovation life cycle\n\nSuccessful introduction of droplet microfluidics into the market requires scenario analysis in the early stage of the product development cycle, as would be expected for any other product development. The purpose is to identify the key drivers in the application market place and uncertainties, then to come up with several scenarios and corresponding technology trends so that the likelihood for commercial success can be more precisely recommended. Here the key drivers are cost effectiveness, high sensitivity, reliability, high-speed, and portability to perform bioanalysis. Cost effectiveness requires small sample volume and here the microfluidics has its place. The key uncertainties of the product development are convergence of repeatable large scale manufacturing techniques, macro-economic condition, and the emerging and development of competing technologies. As an example we can look at severe acute respiratory syndrome (SARS) \\[[@CR12]\\] pandemic diagnostic market in 2003. At the time of the SARS pandemic the diagnoses was performed at specialized well equipped clinics and hospitals, e.g., in Singapore with its 4.5 million population all SARS testing was conducted only at Tan Tock Seng Hospital using laboratory-sized PCR systems. Luckily the early symptom of SARS is the onset of fever which could be detected by ultra fast infrared (IR) cameras. This mass testing practically eliminated the SARS virus spreading. This pandemic serves as a wakeup call. What would happen if technology such as offered by the IR camera is not effective for future pandemic? It is the perfect opportunity for microfludic technology to be implemented into a product that can penetrate consumer market. From this example we can see the importance of scenario analysis to spot the trend ahead of the market need and the necessary layout corresponding strategy for technology commercialization.\n\nOutlook {#Sec6}\n=======\n\nOur previous discussion and overview may look pessimistic, but in fact we are just trying to identify the reasons why, in spite of a lot of efforts, the results are still evasive. So now comes the question: what kind of future awaits microfluidics? There are many examples of new technologies which looked so promising but soon were forgotten. Will microfluidics follow such a path? We believe that most likely this will not be happening. There are areas where microfluidics will eventually be the dominant if not the only technology. Obvious prime applications are anything with volume or weight limitations, for example, in space program applications \\[[@CR13]\\] where weight limit is the dominate factor that filters out the conventional approaches. We can envision remotely controlled system for Moon or Mars exploration, that in microfluidics-based technology will be top candidate for any diagnostic and analytical tool due to its small volume and corresponding light weight. Besides these rather exotic systems, where else could microfluidics prevail? We have already mentioned digital PCR and surface-based virtual reaction chambers (VCRs). Their advantages are obvious: digital PCR can be used to determine absolute number of DNA copies in the original sample and VCR-based systems cost only a few cents. What else? Of course, capillary electrophoreses is an example, liquid and gas chromatography and heat exchanger/mixers are also available and successfully marketed by Agilent. We can expect further development in these fields.\n\nFurther, we envision three major streams of future development apart from currently existing commercially successful devices.\n\nA first stream could be massively parallel systems for drug or patient screening that are capable of competing with fully automated robotic systems used by big centralized hospital laboratories. Here the cost of the microfluidics is not critical because it is orders of magnitude cheaper than the current robotic approaches. An example of this approach is Steve Quake's massively parallel system \\[[@CR14]\\]. Also Affymetrix's DNA chip \\[[@CR15]\\] probably fits into this category.\n\nA second stream could be simple microfluidic devices for point-of-care applications, where the cost of both capital investment and cost per test are of utmost importance. Here the microfluidics technology will compete with injection molding which naturally brings up a question, if there is even a chance that microfluidics can to win this contest. Injection molded parts are so cheap that their disposability is more economical than any attempt of cleaning the parts and reuse them. From a practical point of view, when it comes to clinical diagnostics, the doctors firmly insist on disposable devices to maintain an absolute sterile environment for the assay, and reduce ambiguity of determining results. This poses serious cost issue to microfluidics because currently they are just too expensive. Even channel free systems such as surface-based microfluidics relying on electrowetting is too costly. They are actually very interesting examples of versatile microfluidics systems due to the fact that they can be easily reprogrammed so that the layout of the microfluidics channel can be quickly changed. However, the reality is that for routine testing/diagnoses we do not need to change the microfluidics layout because there are simple techniques to achieve it so the versatility is not always needed. In this case, the technology can be considered overkill.\n\nThere are other competing techniques such as droplet-based PCR \\[[@CR16]\\] which is based on single step lithography and simple heater. It can be probably further simplified to either use stamping (as shown schematically in Fig.\u00a0[10.4](#Fig4){ref-type=\"fig\"}) or eliminate requirement for lithography.Fig. 10.4(**a**) Satellite picture reconstruction of Moses leading his people across the Red Sea (copyright by The Glue Society, Australia, reprinted with permission), at this scale it is very unlikely, but nevertheless it is \"an incredible story\" showing that hydrostatic forces are dominated by surface tension. Inspired by earlier Takehiko Kitamori's presentation we show here schematically (**b**) device with two regions separated from each other by a hydrophilic/hydrophobic surface patterning. It can be relatively easily achieved at the micron scale where surface tension is much greater than hydrostatic forces. (**c**) Photograph of an actual device based on hydrophilic/hydrophobic concept\n\nAnother example is emulsion PCR \\[[@CR17]\\]. Here the PCR is performed on beads each containing only single template molecule. Each bead is enclosed in a tiny sample droplet with PCR master mix and the thermal cycling is performed inside the droplet. The advantage of this system is that typically there is only a single DNA molecule and single bead within each droplet thus eliminating interference with other DNAs. Once the PCR is completed the emulsion is spread over a picowell plate reader. The size of each well is only 40\u00a0\u03bcm and beads 28\u00a0\u03bcm forming a system where there is only a single bead residing within each well which is enabling of single molecule sequencing.\n\nA third stream of microfluidic devices is used for cell biology and tissue engineering research support \\[[@CR18]\\]. Here the high cost of the microfluidic systems for research is tolerated as long as really novel effects or information can be achieved. So far cell biology is supported by microfluidics in areas of protein crystallization, stem cell sorting and differentiation, embryo handling \\[[@CR19]\\] structured tissue engineering as well as regenerative medicine. One typical example is seeding stem cells on a scaffolding to form a bioartificial microreactor, such as kidney \\[[@CR20]\\] or liver. Also potential patients would clearly benefit from bioartificial organs such as kidney which would eliminate their frequent visits of dialysis centers improving quality of their lives.\n\nThere will always be niche areas where microfluidics could play an important role, such as digital PCR for quantitative molecular testing for routine medical diagnostic.\n\nOverall there is definitely light at the end of the tunnel but it will take some time to get there.\n"}
+{"text": "Introduction {#s1}\n============\n\nCoronary artery disease (CAD) is a leading cause of mortality and morbidity worldwide [@pone.0039574-Lopez1]. Susceptibility to CAD is influenced by combined effects of environmental and inherited genetic factors, and consequently some families are particularly affected [@pone.0039574-Marenberg1]. In 2007, several genome-wide association studies consistently identified a region on chromosome 9p21.3 as being the most strongly associated with CAD [@pone.0039574-1], [@pone.0039574-Helgadottir1], [@pone.0039574-McPherson1], [@pone.0039574-Samani1]. This finding has been replicated in multiple case-control studies in several population groups in numerous ethnicities [@pone.0039574-Abdullah1], [@pone.0039574-Assimes1], [@pone.0039574-Broadbent1], [@pone.0039574-Hinohara1], [@pone.0039574-Larson1], [@pone.0039574-Schunkert1], [@pone.0039574-Shen1], [@pone.0039574-Shen2], [@pone.0039574-Talmud1], making 9p21.3 the most replicated molecular genetic association with coronary heart disease to date. Other variants at chromosome 9p21.3 have been linked with susceptibility to several other complex diseases including type 2 diabetes [@pone.0039574-Scott1], [@pone.0039574-Zeggini1], aortic aneurism [@pone.0039574-Helgadottir2], ischemic stroke [@pone.0039574-Gschwendtner1], [@pone.0039574-Matarin1], several cancers [@pone.0039574-Bishop1], [@pone.0039574-Chen1], [@pone.0039574-Debniak1], [@pone.0039574-Driver1], [@pone.0039574-Gayther1], [@pone.0039574-Healy1], [@pone.0039574-Kumar1], [@pone.0039574-Shete1], [@pone.0039574-Wrensch1] and frailty [@pone.0039574-Melzer1].\n\nWithin the 9p21.3 locus, multiple single nucleotide polymorphisms (SNPs) in strong linkage disequilibrium have been associated with CAD [@pone.0039574-1], [@pone.0039574-Samani1], [@pone.0039574-Schunkert2]. The risk (minor) allele occurs with high frequency among many populations (minor allele frequency \u223c50% in European populations) [@pone.0039574-Assimes1], [@pone.0039574-Hinohara1], [@pone.0039574-Larson1], [@pone.0039574-Schunkert1], [@pone.0039574-Shen1], [@pone.0039574-Shen2] and confers a modest, yet highly reproducible increase in risk of approximately 1.3-fold per copy [@pone.0039574-Schunkert2]. It has been suggested that the 9p21.3 locus may have clinical utility as an early marker for CAD susceptibility [@pone.0039574-Brautbar1].\n\nThe association between the 9p21.3 risk locus and CAD appears to be independent of established risk factors, including elevated lipid levels, high blood pressure, obesity and diabetes [@pone.0039574-McPherson1], [@pone.0039574-Broadbent1], [@pone.0039574-Schunkert1], and the mechanism underlying the association remains enigmatic. The risk locus contains no protein coding genes or known microRNAs. The nearest genes, approximately 100 kb upstream of the risk locus, are a pair of tumor suppressor genes (cyclin dependent kinase inhibitors, *CDKN2A* and *CDKN2B)* that are involved in regulation of the cell cycle and have no demonstrated role in CAD to date. The risk locus overlaps exons 13--20 of a recently identified large, non-coding, antisense RNA of unknown function, named *ANRIL* (antisense noncoding RNA in the INK4 locus, also known as *CDKN2BAS*) [@pone.0039574-Pasmant1]. *ANRIL*, like other non-coding RNAs, is predicted to play a role in regulation of gene expression and is expressed in cells and tissues that are affected by atherosclerosis [@pone.0039574-Broadbent1]. Variants associated with CAD are located within intronic and 3\u2032 flanking sequences and recent data suggest that these, and other variants in close proximity, mediate risk of CAD by altering expression of *ANRIL* (and possibly decreasing expression of *CDKN2A* and *CDKN2B*) via multiple, independent *cis*-regulatory elements [@pone.0039574-Cunnington1], [@pone.0039574-Harismendy1] and, more specifically, by influencing the expression and structure of individual splice variants produced from the *ANRIL* gene [@pone.0039574-Burd1], [@pone.0039574-Jarinova1], [@pone.0039574-Liu1], [@pone.0039574-Holdt1], [@pone.0039574-Folkersen1]. The region contains a dense assembly of gene expression enhancers and two CAD risk SNPs are located in one of these motifs, which disrupts a binding site for the transcription factor, STAT1 [@pone.0039574-Harismendy1]. Less clear are the risk allele-associated changes in expression of *CDKN2A* and *CDKN2B*; two reports find they are both concordantly decreased with *ANRIL* [@pone.0039574-Liu1], [@pone.0039574-Folkersen1], while decreased *CDKN2A* [@pone.0039574-Cunnington1] and increased *CDKN2B* expression has also been reported [@pone.0039574-Cunnington1], as well as strong evidence for direct involvement of *ANRIL* in epigenetic repression of both *CDKN2A* and *CDKN2B* [@pone.0039574-Yap1], [@pone.0039574-Yu1]. The downstream effects of altered activation of these regulators of the cell cycle progression pathway on CAD risk remain unknown.\n\nWe now have a greater understanding of the regulatory events at the genomic region proximal to the risk locus, but to understand the mechanism underlying the association between 9p21.3 and CAD, pathways involved in disease susceptibility in tissues critical to its pathogenesis need to be defined. Further downstream targets of *ANRIL* linking this genomic region to atherosclerotic processes fundamental to CAD have yet to be determined. We hypothesized that variants within the 9p21.3 risk locus may be associated with altered expression of genes in myocardial and vascular tissues, which contributes to the development of cardiovascular pathology. To test this hypothesis and identify pathways that might be influenced by the 9p21.3 variants, we investigated associations between rs1333049, a representative SNP from the 9p21.3 locus, with global gene expression in several key cardiovascular tissues, including heart tissue from donors with no previously diagnosed heart disease (predominant cause of death, cerebral vascular accident) and carotid plaque tissues from carotid endarterectomy patients. Our data suggest altered expression of multiple genes in these tissues and we propose a common transcriptional mechanism that might relate cardiovascular gene expression to the 9p21.3 risk locus.\n\nResults {#s2}\n=======\n\nClinical Characteristics and Genotype Frequencies in Heart Donors and Patients {#s2a}\n------------------------------------------------------------------------------\n\nThe baseline characteristics of heart donors, heart valve patients and carotid endarterectomy patients are listed in [Table 1](#pone-0039574-t001){ref-type=\"table\"}. For all cohorts, the genotype frequencies were in Hardy-Weinberg equilibrium (donors p\u200a=\u200a0.762, heart valve patients p\u200a=\u200a0.701, carotid endarterectomy patients p\u200a=\u200a1.00) and were in concordance with other European populations [@pone.0039574-Samani1]. For heart donors, associations between baseline characteristics and 9p21.3 genotype are reported in [Table 2](#pone-0039574-t002){ref-type=\"table\"}.\n\n10.1371/journal.pone.0039574.t001\n\n###### Baseline characteristics of heart donors, heart valve patients and carotid endarterectomy patients.\n\n![](pone.0039574.t001){#pone-0039574-t001-1}\n\n  Variable                                                                         Heart Donors (n\u200a=\u200a108)   Heart Valve Patients (n\u200a=\u200a104)   Carotid Endarterectomy Patients (n\u200a=\u200a106)\n  ---------------------------------------------- -------------------------------- ------------------------ -------------------------------- -------------------------------------------\n  Age[\\*](#nt101){ref-type=\"table-fn\"} (years)                                             48\u00b113                        62\u00b110                                  70\u00b19\n  Gender                                                     Male (%)                     55 (51)                      78 (75)                                85 (80)\n                                                            Female (%)                    52 (48)                      26 (25)                                21 (20)\n  Ethnicity                                                European (%)                   104 (96)                    104 (100)                              106 (100)\n                                                       African-American (%)                4 (4)                        0 (0)                                  0 (0)\n  LVEF[\\*](#nt101){ref-type=\"table-fn\"} (%)                                                52\u00b115%                         --                                    --\n  Cause of Death                                  Cerebral vascular accident (%)          78 (72)                         --                                    --\n                                                        Gunshot wound (%)                  10 (9)                         --                                    --\n                                                    Motor vehicle accident (%)             9 (8)                          --                                    --\n                                                         Head trauma (%)                   7 (7)                          --                                    --\n                                                            Other (%)                      4 (4)                          --                                    --\n  rs1333049                                                   GG (%)                      32 (30)                      38 (37)                                34 (32)\n                                                              GC (%)                      55 (51)                      48 (46)                                52 (49)\n                                                              CC (%)                      21 (19)                      18 (17)                                20 (19)\n\nmean\u00b1standard deviation.\n\n10.1371/journal.pone.0039574.t002\n\n###### Baseline characteristics of heart donors by 9p21.3 (rs1333049) genotype and allele frequency.\n\n![](pone.0039574.t002){#pone-0039574-t002-2}\n\n  Variable                                                     rs1333049 genotype    allele     p-value(additive model)                \n  ---------------------------------------------- ------------ -------------------- ----------- ------------------------- ------ ------ -------\n  Age[\\*](#nt102){ref-type=\"table-fn\"} (years)                     45.0\u00b113.9        49.4\u00b112.3          48.2\u00b111.2           --     --    0.280\n  Gender                                           Male (%)         18 (33)          27 (49)            10 (18)           0.57   0.43   0.519\n                                                  Female (%)        14 (27)          27 (52)            11 (21)           0.53   0.47  \n  Ethnicity                                        European         29 (28)          52 (51)            21 (21)           0.54   0.46   0.237\n                                                     A-A             2 (50)          2 (50)              0 (0)            0.75   0.25  \n  LVEF[\\*](#nt102){ref-type=\"table-fn\"} (%)                          45\u00b116            57\u00b114              49\u00b114             --     --    0.329\n  Cause of Death                                   CVA (%)          19 (24)          42 (54)            17 (22)           0.51   0.49   0.230\n                                                    GW (%)           3 (30)          6 (60)             1 (10)            0.60   0.40  \n                                                   MVA (%)           4 (44)          4 (44)             1 (11)            0.67   0.33  \n                                                    HT (%)           3 (43)          2 (28)             2 (28)            0.57   0.43  \n                                                  Other (%)          3 (75)          1 (25)              0 (0)            0.88   0.12  \n\nmean\u00b1standard deviation.\n\nA-A\u200a=\u200aAfrican-American; CVA\u200a=\u200a cerebral vascular accident; GW\u200a=\u200agunshot wound; HT\u200a=\u200ahead trauma; LVEF\u200a=\u200aleft ventricular ejection fraction; MVA\u200a=\u200a motor vehicle accident.\n\nGene Expression Profile Associated with 9p21.3 Risk Allele in Myocardium {#s2b}\n------------------------------------------------------------------------\n\nTo investigate associations between 9p21.3 genotype and gene expression with minimal confounding from the influence of advanced coronary artery disease, genome-wide analysis of Affymetrix Human Gene 1.0 ST expression profiles was performed in cardiac tissue from heart donors. Despite having no prior diagnosis of a cardiac event, this group are likely to have a spectrum of subclinical atherosclerosis, consistent with their mean age of 48 years ([Table 1](#pone-0039574-t001){ref-type=\"table\"}). This analysis identified 59 gene transcripts that were putatively differentially expressed in association with the rs1333049 high-risk C allele (fold-change \\>1.1 per copy) at the chosen level of significance (p\\<0.05, not corrected for multiple comparisons). Of these, 46 transcripts remained significantly associated with 9p21.3 genotype after adjustment for age, gender, ethnicity and cause of death. The majority of these genes (70%) were down-regulated in association with the risk allele, and showed a modest fold-change in expression (median fold change \u200a=\u200a1.12 per copy of the risk allele, range 1.10--1.33). Analysis of the chromosomal location of the differentially expressed genes revealed an even distribution throughout the genome with no chromosome or cytoband significantly over-represented ([Figure S1](#pone.0039574.s001){ref-type=\"supplementary-material\"}).\n\nThe 20 most differentially expressed genes in heart donors are shown in [Figure 1](#pone-0039574-g001){ref-type=\"fig\"} (p\\<0.05 adjusted for age, gender, ethnicity and cause of death; additional information provided in [Table S1](#pone.0039574.s005){ref-type=\"supplementary-material\"}). The gene encoding the extracellular matrix protein periostin, *POSTN*, was the most differentially expressed gene associated with rs1333049 genotype (down-regulated 1.33-fold per copy of the risk allele, p\u200a=\u200a0.002). This was confirmed by RT-qPCR in the same samples, which showed that expression of *POSTN* was down-regulated 1.84-fold per copy of the risk allele (95% CI 1.10--3.06, p\u200a=\u200a0.020, [Figure 2](#pone-0039574-g002){ref-type=\"fig\"}) and that 5% of the variance in *POSTN* levels could be attributed to 9p21.3 genotype. Associations between 9p21.3 genotype and the expression of three other transcripts, *CCDC80*, *VCAM1* and *GAP43*, were also validated by RT-qPCR, although statistical significance was only replicated for *CCDC80* and *VCAM1* (p\u200a=\u200a0.024 for both, [Figure 2](#pone-0039574-g002){ref-type=\"fig\"}). While no associations between the risk allele and expression of genes in close proximity to the risk locus, including *ANRIL*, *CDKN2A*, *CDKN2B* and *MTAP,* were detected by array, RT-qPCR analysis showed lower expression of *CDKN2B* in individuals carrying the risk allele (down-regulated 1.35-fold per copy of the risk allele, p\u200a=\u200a0.046, [Figure 2](#pone-0039574-g002){ref-type=\"fig\"}). RT-qPCR corroborated array data for *CDKN2A* and *ANRIL*, which showed no significant difference between genotypes (*CDKN2A* p\u200a=\u200a0.585; *ANRIL* p\u200a=\u200a0.541, [Figure 2](#pone-0039574-g002){ref-type=\"fig\"}). Overall, expression levels of *CDKN2B* and *CDKN2A*, and *CDKN2A* and *ANRIL* were positively correlated (Pearson correlation co-efficient *CDKN2B*/*CDKN2A*\u200a=\u200a0.367, p\u200a=\u200a0.001; *CDKN2A*/*ANRIL*\u200a=\u200a0.252, p\u200a=\u200a0.045). Expression of *ANRIL* and *POSTN* were also strongly positively correlated (Pearson correlation co-efficient 0.658, p\u200a=\u200a4.59\u00d710^\u22121\u2212^).\n\n![Twenty top-ranked genes altered in association with the 9p21.3 risk allele in donor hearts (n\u200a=\u200a108).\\\nAnalysis was performed using a fold-change threshold of \\>1.1 per copy of the risk allele and p\\<0.05 (adjusted for age, gender, ethnicity and cause of death; not corrected for multiple comparisons). Each bar represents an individual gene, as indicated by the gene symbol. Genes are ranked in order of fold-change from greatest to smallest (left to right). Analysis of the top 10 most down-regulated genes (indicated by line below graph) identified a shared combination transcription factor binding sites within the promoter regions of 7 of these genes (genes indicated by asterisks).](pone.0039574.g001){#pone-0039574-g001}\n\n![Comparison of microarray data with real-time PCR for selected genes in donor hearts (n\u200a=\u200a108).\\\nPeriostin, *CCDC80* (coiled-coil domain containing 80) and *VCAM1* (vascular cell adhesion molecule 1), the most differentially expressed genes altered in association with the 9p21.3 risk allele in heart donors, gave an equivalent, statistically significant decrease in expression by microarray and RT-qPCR methods. In contrast to array data, RT-qPCR for *CDKN2B*, a gene in close proximity to the 9p21.3 risk locus, also showed a significant decrease in expression. No significant association between the 9p21.3 risk allele and expression levels of *ANRIL* or *CDKN2A* was detected in these individuals, by either method. The 9p21.3 (rs1333049) low-risk GG genotype is depicted in green; GC heterozygotes (intermediate risk) are shown in orange; the high-risk CC genotype is shown in red. Statistical analysis was performed assuming an additive genetic model and all p-values have been adjusted for age, gender, ethnicity and cause of death.](pone.0039574.g002){#pone-0039574-g002}\n\nUsing a false discovery rate (FDR) of 0.001, analysis of Gene Ontology biological process (GO_biological process) terms associated with the differentially expressed transcripts identified significant enrichment for genes involved in response to wounding (p-value from gene set enrichment analysis (GSEA): 2.65\u00d710^\u221210^), cell migration (p-value from GSEA: 4.89\u00d710^\u22129^) and the inflammatory response (p-value from GSEA: 1.97\u00d710^\u22127^), all processes that contribute to the development and progression of atherosclerosis. Consistent with these findings, analysis of proprietary GeneGo ontology terms (MetaCore database) revealed that the differentially expressed transcripts were enriched for gene biomarkers (genes previously reported to have altered expression in disease) of inflammation (p-value from GSEA: 7.28\u00d710^\u22129^) and myocardial infarction (p-value from GSEA: 1.53\u00d710^\u22129^), despite originating from cardiac tissue with no diagnosed heart disease (FDR\u200a=\u200a0.001). The subset of genes associated with each of these GO_biological process and GeneGo ontology terms (25 genes in total) were predominantly down-regulated in association with the high-risk 9p21.3 allele (88%). Network analysis showed that these genes mapped to a set of receptors through direct interactions and formed part of a network of pathways that regulate a group of transcription factors, including NF-\u03baB (p-value from network analysis: 1.25\u00d710^\u221265^, [Figure S2](#pone.0039574.s002){ref-type=\"supplementary-material\"}).\n\nTo investigate whether the differentially expressed genes may be co-regulated, the transcriptional control of the differentially expressed genes was analysed. This showed that the gene set was most strongly enriched by genes known to be regulated by the transcription factor Sp1 (18 out of 46 genes, p-value from network analysis: 4.83\u00d710^\u221261^). Of the 10 most down-regulated genes, further analysis of the promoter regions revealed that specific combinations of transcription factor binding sites were over-represented among 7 of these genes: *CCDC80*, *VCAM1*, *GAP43*, *EGR1*, *TNFAIP6 CILP, BDNF* ([Figure 1](#pone-0039574-g001){ref-type=\"fig\"}). Consistent with the network of genes associated with GO_biological process and GeneGo ontology terms above, these binding site signatures comprised a NF-\u03baB binding site in close proximity (\\<200 bases) to an HMGB1 site (p-value from one-sided Fisher exact test: 1.14\u00d710^\u22125^) and a RelA (NF-\u03baB subunit) binding site in close proximity to a TATA binding protein site (6 genes, *CILP* not included; p-value from one-sided Fisher exact test: 2.11\u00d710^\u22127^). Transcription factor binding sites were located in evolutionarily conserved regions of DNA that share more than 70% identity with the mouse genome. No combinations of transcription factor binding sites were identified among the promoter regions of the 10 most up-regulated genes in heart donors.\n\nCombined Analysis of Gene Expression Profiles Associated with the 9p21.3 Risk Allele Across Myocardial and Vascular Tissues {#s2c}\n---------------------------------------------------------------------------------------------------------------------------\n\nTo identify common pathways associated with the 9p21.3 risk allele irrespective of differences in tissue type or disease state, an analysis of the genes most significantly differentially expressed in either myocardial, carotid plaque, aorta or mammary artery tissues was performed. An expression fold-change threshold of \\>1.10 per copy of the risk allele at a more conservative level of significance of p\\<0.01 (not corrected for multiple comparisons) was used for this analysis to reduce the potential confounding influence of the presence of coronary artery disease in patient samples and to obtain a feasible number of genes for canonical pathway modeling (computationally intensive). This analysis of expression profiles identified 59 gene transcripts in carotid plaque (56% down-regulated), 20 transcripts in aorta (media/intima, 60% down-regulated), 65 transcripts in mammary artery (media/intima, 60% down-regulated) and 11 transcripts in donor heart (73% down-regulated) tissues that were putatively differentially expressed in association with the high-risk rs1333049 C allele (154 unique transcripts in total, 60% down-regulated, [Table S2](#pone.0039574.s006){ref-type=\"supplementary-material\"}). No transcripts were consistently altered in all tissues, although ubiquitin specific peptidase 15 (*USP15*), a gene shown to induce cardiac hypertrophy in mice [@pone.0039574-Isumi1], showed an equivalent increase in expression in aorta and mammary artery tissues (aorta +1.16-fold per copy of the risk allele, p\u200a=\u200a0.008; mammary artery +1.15-fold, p\u200a=\u200a0.003). No significant associations were detected between the risk allele and overall expression of genes in close proximity to the risk locus (*CDKN2A*, *CDKN2B*, *ANRIL*, *MTAP*) in any of the tissues (an expanded analysis of transcripts spanning a \\>10 Mb region around the risk locus, from *DMRTA1* to *IFNB1*, in heart donors is provided in [Table S3](#pone.0039574.s007){ref-type=\"supplementary-material\"}).\n\nCanonical pathway modelling of the genes most significantly altered in association with the 9p21.3 risk locus in myocardial and vascular tissues combined, identified a highly significant association with the cell cycle G1 phase progression pathway (p-value from canonical pathway modelling: 1.08\u00d710^\u2212258^, [Figure 3](#pone-0039574-g003){ref-type=\"fig\"}), in which proteins encoded by *CDKN2A* and *CDKN2B* (p16 and p15, respectively) play an important regulatory role. Of the 116 out of 154 differentially expressed genes associated with this pathway, 58% were down-regulated in association with the 9p21.3 risk allele ([Figure S3](#pone.0039574.s003){ref-type=\"supplementary-material\"}, [Table S2](#pone.0039574.s006){ref-type=\"supplementary-material\"}). This analysis indicated that the differentially expressed genes were most likely to be transcriptionally regulated by this pathway, via transcription factors including E2F1, E2F4 and Sp1 ([Figure 3](#pone-0039574-g003){ref-type=\"fig\"}). There were no combinations of transcription factor binding sites in common among the 10 most up- or down-regulated genes from each tissue. However, further analysis of the transcriptional regulation of the combined set of differentially expressed genes confirmed that they were most enriched for genes regulated by Sp1 (33 genes, p-value from network analysis: 6.70\u00d710^\u221285^).\n\n![Canonical pathway modeling of the most significantly differentially expressed genes in donor heart (n\u200a=\u200a108), carotid plaque (n\u200a=\u200a106), aorta (n\u200a=\u200a104) and mammary artery (n\u200a=\u200a88) tissues.\\\n(A) A schematic depicting part of the cell cycle G1 phase progression pathway, indicating the regulatory role of p16^INK4^ and p15, which are encoded by two genes located adjacent to the 9p21.3 risk locus (*CDKN2A* and *CDKN2B)*. The majority of the most significantly differentially expressed genes (fold-change \\>1.1 per copy of the risk allele, p\\<0.01 not corrected for multiple comparisons,) associated with the 9p21.3 risk allele in myocardial and vascular tissues were predicted to be transcriptionally regulated by this pathway, predominantly by the transcription factors E2F1, E2F4 and Sp1 (116 out of 154 genes, see [Table S2](#pone.0039574.s006){ref-type=\"supplementary-material\"} for the list of genes associated with this pathway and [Figure S2](#pone.0039574.s002){ref-type=\"supplementary-material\"} for gene network analysis). More than half of these genes (58%) were down-regulated in association with the 9p21.3 risk allele, which would be concordant with lower levels of *ANRIL* expression. Green arrows indicate positive regulation; red arrows indicate negative regulation. (B) Venn diagram indicating the proportion of differentially expressed genes predicted to be regulated by E2F1, E2F4 and Sp1. The genes predicted to be regulated by each of these transcription factors is indicated by the symbols \\*, \u00a7 and \u00b6 respectively, in [Table S2](#pone.0039574.s006){ref-type=\"supplementary-material\"}.](pone.0039574.g003){#pone-0039574-g003}\n\nDiscussion {#s3}\n==========\n\nUsing genome-wide expression profiling, we have identified genes whose expression patterns may be associated with the 9p21.3 CAD risk allele in donor myocardial tissue and in vascular tissues from CAD patients. Within heart, shared combinations of transcription factor binding sites among genes with putatively altered expression suggested co-regulation of these genes by a shared mechanism. Canonical pathway modeling of the most differentially expressed genes across all tissues identified the cell cycle G1 phase progression pathway as the most likely mechanism regulating their expression. This pathway facilitates the activation of genes required for cell cycle progression and cell proliferation [@pone.0039574-Dimova1] and is controlled by proteins encoded by *CDKN2A* and *CDKN2B*, two genes located adjacent to the 9p21.3 risk locus. Hence, we propose that risk variants at 9p21.3 increase the activation of this pathway in cardiovascular tissues, leading to a proliferative phenotype that promotes cardiac hypertrophy and vascular remodeling, and contributes to increased susceptibility to atherosclerosis.\n\nWe report significantly lower expression levels of *CDKN2B*, but not *CDKN2A* or *ANRIL*, in heart donors carrying the 9p21.3 risk allele. Recent studies suggest that genetic variation within the 9p21.3 region may alter the expression of *CDKN2B*, *CDKN2A*, *ANRIL* (the large non-coding RNA that spans the CAD risk locus), and/or other genes on chromosome 9 located up to 1 Mb from the risk locus [@pone.0039574-Harismendy1], [@pone.0039574-Burd1], [@pone.0039574-Jarinova1], [@pone.0039574-Liu1], [@pone.0039574-Holdt1], [@pone.0039574-Folkersen1]. Consistent with these findings, lower protein levels of p15 (*CDKN2B*) and p16 (*CDKN2A*) have been reported in association with the 9p21.3 risk allele in aortic smooth muscle cells [@pone.0039574-Almontashiri1]. These effects may be mediated by disruption of enhancer elements within the risk locus [@pone.0039574-Harismendy1] and/or by altering expression of *ANRIL* splice variants or overall transcript levels [@pone.0039574-Harismendy1], [@pone.0039574-Burd1], [@pone.0039574-Jarinova1], [@pone.0039574-Liu1], [@pone.0039574-Holdt1], [@pone.0039574-Folkersen1]. These findings are concordant with the phenotype of the Chr4^\u039470kb/\u039470kb^ knock-out mouse, in which targeted deletion of the orthologous 9p21.3 risk interval on mouse chromosome 4 resulted in severely reduced expression of nearby genes, including *CDKN2A* and *CDKN2B*, and increased proliferation of vascular cells [@pone.0039574-Visel1].\n\nOur study analyzed genes from myocardial and vascular tissues that were putatively altered in association with the 9p21.3 risk allele to identify common pathways downstream of these modest effects that are relevant to CAD. We found that 75% of the most differentially expressed genes across all tissues, regardless of disease state, were predicted to be regulated by the pathway that controls the cell cycle transition from G1 to S phase ([Figure 3](#pone-0039574-g003){ref-type=\"fig\"}). This pathway activates the E2F family of transcription factors, including E2F1 [@pone.0039574-Frolov1]. In addition to activating genes involved in cell proliferation, E2F1 has been shown to function as an inhibitory regulator for the inflammatory mediator NF-\u03baB and down regulate cytokine-induced expression of adhesion molecules, including *VCAM1* [@pone.0039574-Chen2]. Our data in heart tissue suggests down-regulation of multiple genes associated with inflammation and adhesion processes, including *VCAM1*. Furthermore, the shared transcription factor binding site signatures identified among conserved promoter regions of 7 of the top 10 most down-regulated genes, included either a Nf-\u03baB or RelA (a subunit of NF-\u03baB) binding site. Together, these results suggest that the cell cycle G1 phase progression pathway is activated in individuals with the 9p21.3 risk allele. This may promote a proliferative phenotype that leads to adverse cardiac hypertrophy and vascular remodeling, and an increased susceptibility to CAD. This is consistent with the findings of Jarinova *et al* [@pone.0039574-Jarinova1], who reported altered expression of genes involved in cell proliferation in association with the 9p21.3 risk allele, in whole blood from healthy subjects and stable CAD patients. Activation of this pathway may be one of the mechanisms by which CAD risk variants may increase CAD risk.\n\nThe mechanism underlying increased activation of the cell cycle G1 phase progression pathway in individuals with the 9p21.3 risk allele may center on two genes located adjacent to the risk locus, *CDKN2A* and *CDKN2B*. These genes block this pathway by inhibiting the cyclin-dependent kinases (CDK) 4 and 6. Previous studies suggest that total expression levels of *CDKN2A*, *CDKN2B* and *ANRIL* are positively correlated in blood and tissues [@pone.0039574-Cunnington1], [@pone.0039574-Liu1], [@pone.0039574-Folkersen1], with expression of *ANRIL* and one or both of *CDKN2A* and *CDKN2B* reported to be down-regulated in association with the risk allele [@pone.0039574-Cunnington1], [@pone.0039574-Jarinova1], [@pone.0039574-Liu1], consistent with our study. In support of this, Harismendy *et al* [@pone.0039574-Harismendy1] found that CAD risk SNPs altered the sequence of an enhancer element within the 9p21.3 risk locus and disrupted long-range physical interactions between the enhancer and the *CDKN2A/B* locus. To add another layer of complexity, expression of ANRIL splice variants may also be altered in individuals with the risk allele [@pone.0039574-Burd1], [@pone.0039574-Jarinova1], although splicing varies by tissue type [@pone.0039574-Folkersen1]. In whole blood, Jarinova *et al.* [@pone.0039574-Jarinova1] showed increased expression of short splice variants and decreased expression of a long splice variant of *ANRIL* association with the risk allele. Expression of the long *ANRIL* splice variant was correlated with expression of *CDKN2B*, suggesting that the shift from long to short splice variants may influence the expression of nearby genes. Although *ANRIL* has been shown to repress expression of *CDKN2A* and *CDKN2B* through recruitment and retention of polycomb repressive protein complexes at the 9p21.3 locus [@pone.0039574-Yap1], [@pone.0039574-Yu1], it is unknown whether short or other alternatively spliced *ANRIL* transcripts may increase the efficiency of this epigenetic mechanism. Further research is needed to determine whether protein levels of p16 (*CDKN2A*) and/or p15 (*CDKN2B*) are down-regulated in association with the 9p21.3 risk allele in tissues relevant to the development of CAD. This would be predicted to increase activation of the cell cycle G1 phase progression pathway and promote a proliferative phenotype.\n\nIt is unknown whether the modest putative changes in gene expression observed in the current study would be sufficient to promote the development of CAD in individuals with the 9p21.3 risk allele. Of particular relevance is the observation that deletion of the equivalent 9p21.3 risk region in mice did not cause CAD, despite a marked reduction in *CDKN2A* and *CDKN2B* expression and a proliferative phenotype [@pone.0039574-Visel1]. This suggests that other factors may be required for the development of CAD and that activation of the cell cycle G1 phase progression pathway may represent just one aspect of the mechanism underlying the association between the 9p21.3 risk allele and increased CAD risk. Moreover, none of the genes identified as having altered expression in association with the 9p21.3 risk allele remained significant after correction for multiple comparisons and these findings need to be validated in an independent sample set. It should also be noted that the majority of heart donors were on life-support as a result of head trauma or cerebral vascular accident and thus the myocardial gene expression profile may have been affected by the traumatic injuries, acute drug treatments and underlying sub-clinical cardiac disease prior to the donation of tissue. Similarly, the gene expression profiles of vascular tissues were likely to have been affected by chronic drug treatment.\n\nIn summary, our data suggest that the 9p21.3 CAD risk locus may be associated with an altered pattern of gene expression in myocardial tissue from donors with no diagnosed heart disease and in vascular tissues from heart patients. These expression profiles, while tissue-specific, may be regulated through the cell cycle G1 phase progression pathway, which is inhibited by proteins encoded by *CDKN2A* and *CDKN2B,* two genes located adjacent to the risk locus. We speculate that the CAD risk variants decrease expression of *CDKN2A* and *CDKN2B*, which leads to activation of CDK4, CDK6, and downstream transcription factors, including E2F1, E2F4 and Sp1, and to altered expression of their target genes. Our data suggests that this network of genes may be altered in association with 9p21.3 risk SNPs in cardiovascular tissues, both before and after the onset of overt disease. These findings may help elucidate one of the mechanisms by which the 9p21.3 locus contributes to increased CAD risk.\n\nMethods {#s4}\n=======\n\nHuman Samples {#s4a}\n-------------\n\n### Donor Heart Tissue {#s4a1}\n\nHeart tissue from the left ventricular free wall of organ donors was collected by the Cleveland Clinic Kaufman Center for Heart Failure human heart tissue bank (n\u200a=\u200a108) between August 1993 - May 2005. Heart specimens were rapidly frozen in liquid nitrogen at the time of harvest and stored at \u221280\u00b0C until use. The decision that the heart could not be used for transplantation was made by members of the clinical organ procurement team. Reasons for rejection included histocompatibility mismatch, structural damage to the heart, cardiac disease or other elements of the medical history that made the donor undesirable. The research team was not contacted until donation for transplant had been ruled out and the family provided informed written consent to use the heart for research. The study was approved by the Cleveland Clinic Institutional Review Board (IRB 2378). All procedures were in accordance with institutional guidelines.\n\n### Aorta, Mammary Artery and Carotid Plaque Tissue {#s4a2}\n\nTissue from the media/intima layers of aorta (n\u200a=\u200a104) and mammary artery (n\u200a=\u200a88) were collected from Swedish patients undergoing aortic valve surgery or surgery for aortic aneurysm and carotid plaque tissue was collected from 106 Swedish patients undergoing carotid endarterectomy as previously described [@pone.0039574-Folkersen1]. All patients provided informed written consent. The study was approved by the Karolinkska University ethics committee (01--199, 02--146, 02--147, 2006/784--31/1, 2005/880--31/3) and all procedures were in accordance with institutional guidelines.\n\nSample Preparation and Genotyping {#s4b}\n---------------------------------\n\n### Donor Heart Tissue {#s4b1}\n\nTotal RNA and genomic DNA was simultaneously extracted from frozen tissue after automated grinding (Retsch Mixer Mill MM301, Haan, Germany) in TRIzol\u00ae (Invitrogen, Carlsbad, CA) and chloroform [@pone.0039574-Chomczynski1]. RNA was purified using RNeasy Midi columns (Qiagen, Valencia, CA) according to manufacturer's instructions and tested for quantity and quality (by visual assessment of gel plots) with an Experion (Bio-Rad Laboratories, Hercules, CA). RNA samples were digested with DNase I (Invitrogen). First strand cDNA synthesis was performed from 2 \u00b5g of total RNA with Superscript III (Invitrogen), followed by RNase H digestion (Invitrogen) as previously described [@pone.0039574-Ellmers1]. Genomic DNA was extracted from the remaining organic and interphase layers of the TRIzol\u00ae extract (see protocol at <http://genome-www.stanford.edu/DFSP/materials.shtml>). Individuals were genotyped in duplicate for rs1333049 with Taqman SNP assay C_1754666_10 (Applied Biosystems, Foster City, CA) on a Rotor-Gene 3000 and analyzed with Rotor-Gene version 6.1 software (Corbett Research, Sydney, Australia). Reactions were optimized for 10 \u00b5L volumes with 0.5\u00d7 the recommended probe concentration. Genotypes were validated for a subset of randomly selected samples by re-genotyping (n\u200a=\u200a66) or sequencing (n\u200a=\u200a7, see [Methods S1](#pone.0039574.s009){ref-type=\"supplementary-material\"} and [Table S4](#pone.0039574.s008){ref-type=\"supplementary-material\"} for primer sequences). Both methods gave 100% concordance with original genotypes.\n\n### Aorta, Mammary Artery and Carotid Plaque Tissues {#s4b2}\n\nTotal RNA was isolated from tissues and genomic DNA was isolated from peripheral blood leukocytes as previously described [@pone.0039574-Folkersen1]. Genotyping for rs1333049 was performed for aorta and mammary artery samples with custom printed Cardiometabo chips (Illumina, San Diego, CA) on an iScan instrument (Illumina). Genotypes for the carotid plaque samples were imputed from Human 610W-Quad Bead array (Illumina) data [@pone.0039574-Folkersen2].\n\nMicroarrays {#s4c}\n-----------\n\n### Donor Heart Tissue {#s4c1}\n\nIndividual myocardial gene expression profiles were generated for all samples with Human Gene 1.0 ST arrays (Affymetrix, Santa Clara, CA) according to manufacturer's instructions (GEO accession: GSE22253). These arrays use 25-mer oligonucleotides to measure mRNA transcript abundance and consist of 844,550 probes representing approximately 27,900 transcripts [@pone.0039574-Robinson1]. Probe intensities were estimated using linear models fitted to Robust Multi-Array (RMA)-background corrected and quantile normalized data with R software (<http://www.R-project.org>) and the Aroma.affymetrix (<http://groups.google.com/group/aroma-affymetrix>) and Bioconductor Limma packages [@pone.0039574-Gentleman1], [@pone.0039574-Smyth1]. Transcripts were annotated using the file HuGene-1_0-st-v1.na25.hg18.transcript.csv created on 20 March 2008 available from <http://www.affymetrix.com/support/technical/byproduct.affx?> product\u200a=\u200ahugene-1_0-st-v1. Associations between rs1333049 genotype and myocardial gene expression were tested with R software and the Bioconductor Limma package [@pone.0039574-Gentleman1], [@pone.0039574-Smyth1] using an additive genetic model (see [Figure S4](#pone.0039574.s004){ref-type=\"supplementary-material\"} for QQ plot). None of the associations remained significant after correcting for multiple comparisons using the method of Benjamini and Hochberg [@pone.0039574-Benjamini1], and unadjusted p-values have been reported. Genes altered \\>1.1-fold per copy of the risk allele at an unadjusted p-value \\<0.05 that remained significant after adjusting for age, gender, ethnicity and cause of death (dichotomized as 'cerebral vascular accident' or 'other') were considered to be of potential biological significance. The conservative fold-change threshold (\\>1.1-fold per copy of the risk allele) was applied to limit bioinformatic analyses to those transcripts most likely to have a functional effect, however this may have resulted in some false-negative associations.\n\n### Aorta, Mammary Artery and Carotid Plaque Tissues {#s4c2}\n\nIndividual gene expression profiles were generated for aorta and mammary artery samples with Human Exon 1.0 ST arrays (Affymetrix) and for carotid plaque samples with HG-U133 plus 2.0 arrays (Affymetrix, GEO accession: GSE21545) as previously described [@pone.0039574-Folkersen2]. Gene array analysis of aorta, mammary artery and carotid plaque samples was performed using a additive genetic model, as previously described [@pone.0039574-Folkersen1] (see [Figure S4](#pone.0039574.s004){ref-type=\"supplementary-material\"} for QQ plots).\n\nReal-Time PCR in Donor Heart Tissue {#s4d}\n-----------------------------------\n\nTo validate array data in donor hearts, real-time PCR (RT-qPCR) was performed using Taqman gene expression assays with inventoried probes (Applied Biosystems) for *POSTN* (assay Hs00170815_m1), *CCDC80* (Hs00277341_m1), *VCAM1* (Hs01003372_m1) and *GAP43* (Hs00967138), the most differentially expressed genes putatively associated with rs1333049 genotype, and for *ANRIL* (Hs01390879_m1), *CDKN2A* (Hs00923894_m1) and *CDKN2B* (Hs00793225_m1), genes adjacent to the 9p21.3 risk locus. Reactions (20 \u00b5L or 10 \u00b5L) were performed in triplicate or duplicate on a 7500 Fast Real-Time PCR System (Applied Biosystems) or a Lightcycler 480 Real-Time PCR system (Roche Diagnostics, Indianapolis, IN) according to manufacturer's instructions. Quantification was performed with 7500 software version 2.0 (Applied Biosystems) or Lightcycler 480 software release 1.5.0 (Roche). Expression levels were converted to relative quantities and normalized to signal recognition particle 14 (*SRP14*), tumour protein translationally controlled 1 (*TPT1*) and eukaryotic elongation factor 1A1 (*EEF1A1*), as previously described [@pone.0039574-Pilbrow1]. RT-qPCR gene expression data displayed consistently skewed distributions and were log-transformed prior to analysis. Associations between gene expression and 9p21.3 genotype were tested with linear regression adjusting for potential confounding factors (age, gender, ethnicity and cause of death dichotomized as 'cerebral vascular accident' or 'other') using SPSS Statistics software, version 19 (IBM). Data on post-mortem interval (the time between death and tissue collection) was unavailable and gene expression analyses were not adjusted for this factor. Relative expression levels of *POSTN*, *CCDC80*, *VCAM1*, *GAP43*, *CDKN2B*, *CDKN2A* and *ANRIL* were calculated from RT-qPCR and microarray data using geometric means and standard errors. No RT-qPCR analyses were performed in human vessel samples.\n\nBioinformatic Analysis {#s4e}\n----------------------\n\n### Chromosomal Distribution of Differentially Expressed Genes {#s4e1}\n\nThe chromosomal distribution of differentially expressed genes was analyzed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) version 6.7 [@pone.0039574-Huangda1], [@pone.0039574-Huangda2] functional annotation tool and visualized using the R/Bioconductor package Geneplotter (<http://www.bioconductor.org/packages/2.3/bioc/html/geneplotter.html>) based on annotation from hugene10st.db (<http://www.bioconductor.org/packages/2.3/data/annotation/html/hugene10st.db.html>) originally sourced from Entrez Gene on April 2, 2008.\n\n### Gene Set Enrichment Analysis (GSEA) {#s4e2}\n\nThe biological functions of the differentially expressed genes in heart donors was analyzed with Gene Ontology -- biological process terms (GO_biological process, [www.geneontology.org](http://www.geneontology.org)) and proprietary GeneGo ontology terms using the MetaCore database (GeneGo Inc, St Joseph, MI) functional enrichment by ontology tool. The MetaCore database (GeneGo Inc) includes a manually curated database [of human protein-protein, protein-DNA and protein-compound interactions, metabolic and signaling pathways. GeneGo ontologies are represented by canonical pathway maps, cellular process networks, disease biomarker networks, drug target networks, toxicity networks and metabolic networks. The](http://ofhumanprotein-protein,protein-DNAandprotein-compoundinteractions,metabolicandsignalingpathways.GeneGoontologiesarerepresentedbycanonicalpathwaymaps,cellularprocessnetworks,diseasebiomarkernetworks,drugtargetnetworks,toxicitynetworksandmetabolicnetworks.The) functional enrichment by ontology tool ranks the [relevance of matches between the set of differentially expressed genes and GO_biological process and GeneGo ontology terms on the probability of the match occurring by chance, given the size of the database (p-value from GSEA)](http://relevanceofmatchesbetweenthesetofdifferentiallyexpressedgenesandGO_biologicalprocessandGeneGoontologytermsontheprobabilityofthematchoccurringbychance,giventhesizeofthedatabase(p-valuefromGSEA)). For all gene set enrichment analyses, the false discovery rate (FDR) was set at 0.001 to limit the number of false positive results to 0.1%.\n\n### Network Analysis {#s4e3}\n\nThe relationship between the subset of differentially expressed genes in heart donors associated with GO_biological process and GeneGo ontology terms was analyzed with the MetaCore database (GeneGo Inc) analyze network (receptors) algorithm, using the default settings. Briefly, this algorithm generates a list of receptors and transcription factors closely associated with the differentially expressed genes (initial gene set) based on interactions contained within the MetaCore database, then builds a network for each receptor consisting of all the shortest paths (ie the smallest possible number of direct interactions) from the receptor to the nearest transcription factors. The probability that the intersection between the initial gene set and the genes/proteins in the network occurs by chance, given the size of the initial gene set, the network and the interaction database (p-value from network analysis), follows a hypergeometric distribution and was used to rank the relevance of the networks generated.\n\nThe association between individual transcription factors and differentially expressed genes in heart and vascular tissues was analyzed with the MetaCore database (GeneGo Inc) transcription regulation algorithm. This algorithm identifies the shortest path of direct interactions from the differentially expressed genes (the initial gene set) to the nearest transcription factors based on interactions in the MetaCore database, then builds a network for each transcription factor identified. The transcription factor networks were ranked on their statistical significance as described for the algorithm above.\n\n### Enrichment Analysis for Combinations of Transcription Factor Binding Sites (TFBS) {#s4e4}\n\nPromoter regions (\u22121500 bases to +200 bases from the transcription start site) of the top 10 most up- or down-regulated genes from each tissue were extracted and analyzed computationally for common transcription factor binding sites using oPOSSUM [@pone.0039574-HoSui1]. oPOSSUM determines the over-representation of transcription factor binding sites (TFBS) within a set of co-expressed genes compared with a pre-compiled background set. The background set comprises computationally predicted TFBS within evolutionarily conserved regions +/\u221210,000 bases of the predicted transcription factor start site in genes that are orthologous in human and mouse. To limit the number of false positive binding sites identified, the analysis was restricted to regions within the promoter that had \\>70% sequence identity between human and mouse and the transcription factor inter-binding distance was limited to a maximum of 200 base pairs. A one-tailed Fisher exact test was used to determine probability of a non-random association between the initial set of co-expressed genes and the combination of transcription factor binding sites identified.\n\n### Canonical Pathway Modelling {#s4e5}\n\nTo identify pathways associated with the 9p21.3 risk allele across all tissues, canonical pathway modelling of the differentially expressed genes from all tissues combined was performed with the MetaCore database (GeneGo Inc) canonical pathway modelling algorithm, using the default settings. Genes altered \\>1.1-fold per copy of the risk allele at a p-value \\<0.01 (not corrected for multiple comparisons) from each tissue were considered to be of potential biological relevance and comprised the initial gene set for this analysis. A more stringent p-value threshold was selected to reduce the potential confounding influence of the presence of coronary artery disease in the patient samples and to obtain a feasible number of genes for canonical pathway modeling, which is computationally intensive. Briefly, this algorithm identifies all canonical pathways in the database that include genes in the initial gene set, then builds networks consisting of all genes/proteins and interactions from all such pathways. The networks were ranked on the probability that the intersection between the initial gene set and the genes/proteins in the network occured by chance, which is calculated using the hypergeometric probability distribution and takes into account the size of the initial gene set, the network and the number of genes/proteins associated with canonical pathways in the database (p-value from canonical pathway modelling).\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Chromosomal location of 46 transcripts identified as altered in association with the 9p21.3 risk allele (rs1333049) in the myocardium of donors.** Green bars indicate the location of transcripts that were up-regulated in association with the 9p21.3 risk allele; red bars indicate the location of transcripts that were down-regulated in association with the 9p21.3 risk allele (fold-change \\>1.1 per copy of the risk allele, p\\<0.05 adjusted for age, gender, ethnicity and cause of death; not corrected for multiple comparisons). The differentially expressed transcripts were not significantly clustered within individual chromosomes or within particular chromosomal regions.\n\n(TIFF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Network analysis of 25 differentially expressed genes associated with gene ontology and MetaCore database biological process and biomarker terms in heart donors (n\u200a=\u200a108).** These genes mapped to a single network that integrated several intracellular signalling pathways, potentially leading to regulation of multiple transcription factors. Most of the genes (70%) were down-regulated in association with the high-risk 9p21.3 allele (fold-change \\>1.1 per copy of the risk allele, p\\<0.05 adjusted for age, gender, ethnicity and cause of death; not corrected for multiple comparisons). The differentially expressed genes are indicated by red (up-regulated) or blue (down-regulated) circles adjacent to each gene, with the intensity of the colour representative of the magnitude of the fold-change in expression. Thick green lines indicate relationships between proteins that form part of canonical pathways.\n\n(TIFF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Canonical pathway modeling of the most significantly differentially expressed genes in donor heart (n\u200a=\u200a108), carotid plaque (n\u200a=\u200a106), aorta (n\u200a=\u200a104) and mammary artery (n\u200a=\u200a88) tissues.** The majority of the most significantly differentially expressed genes (116 out of 154 genes, fold-change \\>1.1 per copy of the risk allele, p\\<0.01 not corrected for multiple comparisons) associated with the 9p21.3 risk allele in myocardial and vascular tissues were predicted to be transcriptionally regulated by this gene network, which forms part of the cell cycle G1 phase progression pathway. Of these genes, most were predicted to be regulated by the transcription factors E2F1, E2F4 and Sp1 (see [Table S2](#pone.0039574.s006){ref-type=\"supplementary-material\"} for the complete list of genes associated with this pathway, with those genes predicted to be regulated by E2F1, E2F4 and Sp1 indicated by the symbols \\*, \u00a7 and \u00b6 respectively). The differentially expressed genes are indicated by red (up-regulated) or blue (down-regulated) circles adjacent to each gene, with the intensity of the colour representative of the magnitude of the fold-change in expression. Thick green lines indicate relationships between proteins that form part of canonical pathways.\n\n(TIFF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**QQ Plots for assessing the effect of 9p21.3 genotype on global gene expression in donor heart (n\u200a=\u200a108), carotid plaque (n\u200a=\u200a106), aorta (n\u200a=\u200a104) and mammary artery (n\u200a=\u200a88) tissues.** Each plot compares the distribution of p-values for all associations (y-axis) against a theoretical distribution of p-values assuming no effect (x-axis). For donor heart and vascular tissues, the observed distribution of p-values matched the theoretical null distribution or deviated below the line y\u200a=\u200ax, suggesting that any associations between gene expression and 9p21.3 genotype may have occurred by chance. Consequently, none of the associations remained significant after correction for multiple comparisons.\n\n(TIFF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**The 20 most differentially expressed genes associated with the 9p21.3 risk allele in myocardium from 108 heart donors (fold-change \\>1.1 per copy of the risk allele, unadjusted p\\<0.05).**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**The most significantly differentially expressed genes associated with the 9p21.3 risk allele in donor myocardium, and patient aorta, mammary artery and carotid plaque tissues (fold-change \\>1.1 per copy of the risk allele, unadjusted p\\<0.01).**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Affymetrix microarray analysis of associations between 9p21.3 (rs1333049) genotype and transcripts adjacent to the risk locus in heart donors.**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**PCR Primers for Sequencing rs1333049.**\n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n(DOCX)\n\n###### \n\nClick here for additional data file.\n\nWe gratefully acknowledge the donation of human myocardium, aorta, mammary artery and carotid plaque for research purposes.\n\n**Competing Interests:**The authors have declared that no competing interests exist.\n\n**Funding:**This work was funded by the Cleveland Clinic Foundation ([www.clevelandclinic.org](http://www.clevelandclinic.org)), the Heart Foundation of New Zealand ([www.heartfoundation.org.nz](http://www.heartfoundation.org.nz)), the New Zealand Foundation of Research, Science and Technology (now the Ministry of Science and Innovation, [www.msi.govt.nz](http://www.msi.govt.nz)), the Health Research Council of New Zealand ([www.hrc.govt.nz](http://www.hrc.govt.nz)), the Maurice and Phyllis Paykel Trust ([www.paykeltrust.co.nz](http://www.paykeltrust.co.nz)), Lotteries Health New Zealand ([www.dia.govt.nz/Services-Lottery-Grants-Health-Research](http://www.dia.govt.nz/Services-Lottery-Grants-Health-Research)), the Swedish Research Council ([www.vr.se](http://www.vr.se)), the Swedish Heart-Lung Foundation ([www.hjart-lungfonden.se](http://www.hjart-lungfonden.se)), the European Commission ([www.ec.europa.eu](http://www.ec.europa.eu)), the Danish Agency for Science, Technology and Innovation (en.fi.dk) and a donation by Fredrik Lundberg. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\n\n[^1]: Conceived and designed the experiments: VAC APP RWT CSM WHWT. Performed the experiments: WES APP LM. Analyzed the data: JFP CMB MAB APP. Wrote the paper: APP. Conceived the study: VAC APP RWT CSM WHWT. Provided donor heart tissue for the study: CSM WES. Prepared the heart donor tissue, RNA and DNA samples: WES APP. Performed the array experiments: LM. Provided gene expression and genotyping data for aorta, mammary artery and carotid plaque tissues: LF AFC AG PE. Performed the genotyping and wrote the first draft of the manuscript: APP. Performed the RT-qPCR analysis of the donor hearts: APP NMW. Performed statistical and bioinformatic analysis: APP JFP CMB MAB APP. Provided clinical input: WHWT RWT AMR. Contributed to the interpretation of the data and revision of the manuscript: APP LF JFP CMB LM NMW WES WHWT MAB RWT AMR AFC AG PE CSM VAC. Obtained funding for the study: APP VAC RWT AMR CSM AG PE.\n"}
+{"text": "Background {#Sec1}\n==========\n\nThe phylloplane alone represents the largest biological surface on Earth, outnumbering the cells of the plants themselves \\[[@CR1], [@CR2]\\]. The microorganisms that live in this region multiply and occupy newly formed niches while the leaves are expanding \\[[@CR2], [@CR3]\\]. They are influenced by sunlight and the plants metabolism that have nutrients, including carbohydrates, organic acids and amino acids \\[[@CR4], [@CR5]\\]. Furthermore, the cuticle reduces water evaporation as well as leaching the metabolites in the leaves, resulting in a favorable environment \\[[@CR6]--[@CR8]\\]. These surfaces are an open environment that receive migrants transferred by various mechanisms including rain, animals and deposition of aerial particles, which contributes to a large microbial diversity \\[[@CR2], [@CR9]\\].\n\nThe phylloplane microorganisms can be shared randomly among its neighbors, but their survival and presence is generally regulated by the plant \\[[@CR10]\\]. Furthermore, a theory proposes a possible transfer of microorganisms through generations \\[[@CR11]\\]. However, this whole microbiome can be affected by environmental factors, including radiation \\[[@CR12]\\] and pollution \\[[@CR13]\\], as well as biotic factors such as leaf age and the presence of other microorganisms \\[[@CR14]\\].\n\nThe microbiome present in the phylloplane includes a diversity of bacteria, fungi, yeasts, algae and other microorganisms that have commensal, pathogenic and mutualistic interactions with the plant \\[[@CR2], [@CR15], [@CR16]\\]. Bacteria are numerically dominant in the phylloplane environment, of which a large part are proteobacteria, actinobacteria and bacteroidetes \\[[@CR9], [@CR15], [@CR17]--[@CR22]\\]. This variation is observed in different species of plants that have characteristic communities of bacteria in the phylloplane, varying between genotypes \\[[@CR23], [@CR24]\\], as well as between species and taxonomic classifications \\[[@CR17]\\].\n\nThe microbial activities in the leaves can significantly influence the plant health \\[[@CR25]--[@CR27]\\]. The nitrogen fixation in the phylloplane is the main mechanism for the addition of nitrogen in tropical humid ecosystems \\[[@CR28]\\] and temperate forests \\[[@CR29]\\]. Plants can still be affected by the production of growth hormones \\[[@CR27]\\] and indirect protection against pathogens \\[[@CR3], [@CR25], [@CR30]\\]. In this environment, the cuticle \\[[@CR8]\\] and the trichomes \\[[@CR31], [@CR32]\\] are also considered defense components in the phylloplane that together with the microorganisms, constitute a complex region \\[[@CR33], [@CR34]\\].\n\nNext-generation sequencing technology (NGS) had a great impact in the microbial genomics field \\[[@CR35], [@CR36]\\]. The method provides new insights into the non-cultivable microorganisms and the complex host-microbe interactions \\[[@CR19], [@CR37]\\]. In this approach, the metagenomics used a hypervariable region of the highly conserved 16S rRNA gene as a phylogenetic marker allowing the characterization of the diversity of organisms of the total microbiota found in a given habitat \\[[@CR9], [@CR20], [@CR21], [@CR38]--[@CR40]\\].\n\n*Theobroma cacao* L. is a plant native to the South American rainforest, belonging to the Malvaceae family \\[[@CR41]\\]. It is considered that this plant has two large groups of origin according to their morphological and genetic characteristics and geographic location \\[[@CR42], [@CR43]\\]. The cocoa has great industrial importance since it is the raw material for chocolate \\[[@CR44]\\]. Currently, the genotype CCN51 is the most commercialized clone in several countries due to its great productivity. In contrast, the Catongo genotype is used as a model of sustainability of the fungus *Moniliophthora perniciosa* \\[[@CR31], [@CR45]\\], which causes witches' broom disease in cocoa trees.\n\nThe CCN51 genotype is resistant to the fungus *Moniliophthora perniciosa* \\[[@CR45]\\], and its phylloplane has twice as greater index of short glandular secreting trichomes than the Catongo genotype. A total of 41 proteins from leaf water washes (LWW) of the CCN51 identified by mass spectrometry revealed 28 plant proteins and 13 bacterial proteins \\[[@CR31]\\]. This variation of the short glandular secreting trichomes index between the two genotypes may interfere in the microbial community of the phylloplane.\n\nIn this study, we propose that two contrasting genotypes for resistance to witches' broom have variations in the phylloplane microbiota. Therefore, using an independent culture approach, the total microbiome of the *T. cacao* phyloplane of the genotypes CCN51 and Catongo, were studied and characterized. We show that the differences in the index of glandular trichomes in the contrasting genotypes may affect the variety of bacterial microbioma symbiont of the phylloplane, and that the CCN51 genotype presents an exclusive genera with antagonistic potential against phytopathogens in relation to the catongo genotype, reinforcing its preference of agriculture for the cultivation and commercialization.\n\nMaterials and methods {#Sec2}\n=====================\n\nPlant material and DNA extraction {#Sec3}\n---------------------------------\n\nA total of 300 plants of *Theobroma cacao* L. were cultivated in the greenhouse at CEPEC / CEPLAC (Cacao Research Center, Ilh\u00e9us-BA); 150 plants of the Catongo genotype and 150 of the CCN51 genotype. The plants were kept at room temperature and drip irrigation to avoid leaf washing. Seven pots of plants were randomly selected from each genotype to form the pool of the first biological sample, and another second group, also with seven pots of plants to form the pool of the second biological sample (Additional\u00a0file\u00a0[1](#MOESM1){ref-type=\"media\"}: Figure S1). Therefore, the four biological samples (two from CCN51 and two from Catongo) underwent the extraction of the metagenomic DNA and analyzed in triplicates experimental (Additional\u00a0file\u00a0[2](#MOESM2){ref-type=\"media\"}: Figure S2).\n\nYoung leaves were collected within 15 to 20\u2009days after leaf primordium formation, and the metagenomic DNA extraction from the phylloplane was obtained through leaf water wash according to the method described by Shepherd \\[[@CR46]\\]. Furthermore, other young leaves of the CCN51 and Catongo genotypes were collected with the purpose of observing the phylloplane topography and microbes using the Scanning Electron Microscope (SEM) Quanta 250 model (FEI Company).\n\nFor extraction of total DNA, each leaf was washed by immersion for 15\u2009s in a beaker containing 100\u2009ml of distilled water maintained at temperature 8\u2009\u00b0C. The microbiota was obtained from the LWW by filtering through a 0.22\u2009\u03bcm cellulose membrane to retain the microorganisms. Afterwards, the membrane was distributed in eppendorf tubes, flash-frozen in liquid nitrogen and freeze dried until complete elimination of water. 0.5\u2009g of freeze-dried membrane was weighed and DNA was extracted using the PowerSoil\u00ae DNA Isolation Kit (MoBio Laboratories, USA) according to the manufacturer's instructions. DNA quality was checked on 0.8% (*w*/*v*) agarose gel and concentration and purity measured using the Nanodrop (Thermo Scientific, USA).\n\nLibrary construction and sequencing {#Sec4}\n-----------------------------------\n\nBacterial 16S rRNA gene sequences of the V3-V4 hypervariable region were amplified by PCR using the (341*F*) forward and (805*R*) reverse primer \\[[@CR47]\\]. PCR was performed in a final volume of 25\u2009\u03bcL containing the following: 2\u2009\u03bcL of template DNA, 12.5\u2009\u03bcL of HiFi HotStart ReadyMixPCR Kit (Kapa Biosystems) and 5\u2009\u03bcL of each oligonucleotide. Amplification was performed on the Mx 3005P apparatus (Agilent Technologies) under the following conditions: 95\u2009\u00b0C for 3\u2009min, followed by 25\u2009cycles of 95\u2009\u00b0C for 30\u2009s, 55\u2009\u00b0C for 30\u2009s and 72\u2009\u00b0C for 30\u2009s, and a final elongation step at 72\u2009\u00b0C for 5\u2009min.\n\nAmplicons from each biological replicate (3 amplifications for each of the four DNA extractions) were purified using the Agencourt\u00ae AMPure\u00ae XP system (Beckman Coulter, USA). The quality of the purified amplicons was evaluated in 1.5% agarose gel. A new PCR with Nextera XT Index Kit (FC-131--1002) with final volume of 50\u2009\u03bcL was performed in order to add the barcodes, using dual indexing strategy with two 8-base indices. The new amplification was performed under the same conditions as the previous PCR, except for the number of cycles (8). After quantification of the 12 samples using the Kapa Library Quantification kit (Additional\u00a0file\u00a0[3](#MOESM3){ref-type=\"media\"}: Figure S3), the libraries were sequenced on the Illumina MiSeq\u2122 equipment using the V3 kit (MiSeq\u00ae Reagent - Illumina).\n\nData analysis {#Sec5}\n-------------\n\nRaw bacterial sequence reads were initially subjected to the following preprocessing steps and quality controls: (i)\u2009\u2264\u2009100 nucleotides in length (not including sample barcodes) or more than 600\u2009bp were not considered and (ii) reads were trimmed at the beginning of a poor quality region with 10\u2009bp analyzed in FastQC \\[[@CR48]\\] software with a Phred-score\u2009\u2264\u200920. In subsequent screenings, files were processed using MeFit \\[[@CR49]\\], to identify the best possible overlap region, with the least number of mismatching bases and carry out the merger.\n\nFiles were demultiplexed and end chimeras removed using the Quantitative Insights into Microbial Ecology (QIIME) \\[[@CR50]\\] software package and operational taxonomic units (OTUs) were assigned by clustering the sequences with a threshold of 99% identity against the Greengen database version 13.8 16S rRNA \\[[@CR51]\\]. OTUs, were assigned to \"chloroplasts\" and \"mitochondria\" before it was rarefied and served as input for alpha and beta diversity analysis, were filtered. The Qiime package generated rarefaction curves (richness of population analysis) and the calculation of the population diversity analysis estimator Chao1, as also, Alpha (within-sample richness) and beta diversity (between-sample dissimilarity) estimate. Using the GeanAIEx \\[[@CR52]\\] software, the Principal Coordinates Analysis (PCoA) chart was plotted according to weighted UniFrac metrics (\u03b2-diversity). To test whether there is a significant difference in bacterial community composition among genotypes CCN51 and Catongo, we used the analysis of similarity (ANOSIM) with 999 permutations \\[[@CR53]\\].\n\nResults {#Sec6}\n=======\n\nAfter pre-processing, filtering and rarefaction, the sequencing produced a total of 1.7 million reads of the V3-V4 variable region of the 16S rRNA from leaf water washes of the two contrasting cacao genotypes for resistance of witches' broom disease, caused by *M. perniciosa* fungus. The average number of reads per sample was 95.398, ranging from 42.068 to 346.420. The identified bacteria were classified according to phylum, class, order, family, and genus (Additional\u00a0file\u00a0[4](#MOESM4){ref-type=\"media\"}: Table S1). A total of 10 phyla and 73 genera were identified in the Catongo genotype and a total of 11 phyla and 91 genera in the CCN51 genotype (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a). At the phylum level, proteobacteria is the most abundant phylum in the two genotypes, followed by cyanobacteria also in both genotypes and by bacteroidetes in the genotype Catongo (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b) and actinobacteria in the CCN51 genotype (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}c).Fig. 1Distribution of identified bacteria in leaf water washes of *Theobroma cacao*. **a** Distribution of the number of bacteria identified. Distribution of frequency of reads identified according to phylum: (**b**) Catongo genotype and (**c**) CCN51 genotype. Error bars indicate the standard deviation between the frequencies of three experimental samples of each biological sample\n\nThe three dominant bacterial taxonomic classes in the Catongo and CCN51 genotypes were Gammaproteobacteria, Alphaproteobacteria and Actinobacteria (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a and b). Among the orders identified (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a), oceanospirillales, rickettsiales and enterobacteriales, were the three most abundant orders in both genotypes, 45.9, 21.6 and 9.6% for Catongo and, 43.0, 20.1 and 6.6% for CCN51, respectively.Fig. 2Distribution of the frequency of identified bacteria according to class in leaf water washes from contrasting cacao genotypes for *M. perniciosa* resistance. **a** susceptible Catongo genotype. **b** resistant CCN51 genotype. Only classes that represent \u22651% of the total population in at least one sample, are included. Error bars indicate the standard deviation between the frequencies of three experimental samples of each biological sample. **c** Comparison of the number of identified bacterial phylum in the phylloplane between two genotypes\n\nAt the family level, oceanospirillales (44.3%) is the most abundant order in the CCN51 genotype, followed by rickettsiales (20.7%) and enterobacteriales (6.8%). Rickettsiales (42.0%) is the most abundant order followed by enterobacteriales (18.7%) and sphingomonadales (4.9%), in the Catongo genotype. Comparison at genus level was carried out to reveal bacteria commonly or specifically identified in both genotypes, (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}c). Genotype CCN51 presented 33 genera of bacteria exclusive to its phylloplane in comparison to Catongo. The *Stenotrophomonas* genus was prevalent with 13.5% exclusively found in the CCN51 genotype, whereas the genus *Sphingomonas* prevailed with 75.3%, in the Catongo genotype (Additional\u00a0file\u00a0[5](#MOESM5){ref-type=\"media\"}: Table S2).\n\nIn terms of bacterial diversity, rarefaction curves generated from the library reads (42.000 reads per sample) based on a cutoff 99% sequence identity showed an asymptote for both genotypes, which tended to stabilize indicating sufficient sampling to capture most OTUs within communities (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}b). The graphic analysis showed the differences in biodiversity, because the genotype CCN51 represents the curves of the upper part of the figure, revealing that microbial communities from genotype CCN51 were more diverse than those from microbial communities from the genotype Catongo. The PCoA analysis based on weighted UniFrac metrics showed that the bacterial community were clustered per genotype (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a), indicating the distinct bacterial diversity between CCN51 and Catongo. The ANOSIM results also showed that there is a significant difference in bacterial composition between genotypes CCN51 and Catongo (Global *R*\u00a0=\u20090.996, *P*\u00a0\\<\u20090.05).Fig. 3PCoA plot and rarefaction curves determined for all 12 samples of microorganisms from phylloplane of the two contrasting cacao genotypes for *M. perniciosa* resistance. **a** Principal Coordinates Analysis (PCoA) between bacterial communities. **b** Rarefaction curves demonstrating species richness (Chao1) and diversity (PD entire tree)\n\nElectron microscopy analysis of the leaves revealed that in the phylloplane many epiphytes occur in large bacterial aggregates, and fungi (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). The field images provides spatial view of microbiome locations. Isolated bacterial cells have also been observed and some mixed aggregates can be found. The bacteria were visualized next to stomates, cell junctions and mainly in the foliar veins.Fig. 4Scanning electron microscopy analysis. **a** Adaxial surface of the CCN51 genotype. **b** Abaxial surface of the CCN51 genotype. **c** Adaxial surface of the Catongo genotype. **d** Abaxial surface of the Catongo genotype\n\nDiscussion {#Sec7}\n==========\n\nCocoa is a source of raw material for chocolate production and is cultivated in tropical and subtropical regions around the world \\[[@CR44]\\]. Great losses in cacao production happen due to fungal diseases, such as witches' broom and frosty pod, caused by *M. perniciosa* and *M. roreri* \\[[@CR54]\\]*,* respectively. These pathogens are hemibiobrophyc and start the infection process with spores deposited on the phylloplane from young cacao tissues (leaf and fruits). There is no effective chemical control after the invasion of the apoplast and onset of the parasitic phase of the disease \\[[@CR55], [@CR56]\\]. The mechanisms of spore germination and pre-infection processes in contrasting resistance cacao genotypes has been analyzed \\[[@CR57]\\]. Furthermore, the topography of the epicuticular wax layer \\[[@CR58]\\], the short glandular secreting trichomes and the importance of water-soluble components of the phylloplane \\[[@CR31]\\] have also been analyzed. *M. perniciosa* tends to have a relatively short epiphytic phase \\[[@CR31], [@CR59]\\] and requires few or no exogenous nutrients in this phase \\[[@CR60]\\]. Thus, if biocontrol occurs at the epiphytic phase, antagonists that act as antibiotics (rather than competition) should be the most effective \\[[@CR61]\\]. In this work, we analyze the composition and bacterial diversity of the phylloplane of CCN51 and Catongo cacao genotypes, resistant and susceptible to witches' broom respectively. DNA extracted from microorganisms recovered by filtration from leaf water washes was analyzed by metagenomic approach based on the sequencing of the v3-v4 hypervariable region from the 16S rRNA.\n\nMicrobiome communities contained abundance of genera within the proteobacteria and cyanobacteria phylum, with prevalence of *Candidatus Portiera* in both genotypes (Additional\u00a0file\u00a0[4](#MOESM4){ref-type=\"media\"}: Table S1). This genus presented a single species *Candidatus Portiera aleyrodidarum* sp. which provides amino acids and carotenoids \\[[@CR62]\\] to its host *Bemisia tabaci* \\[[@CR63]\\]. In the phylloplane, this microorganism may be acting symbiotically with *T. cacao*, participating in the organic metabolism with the contribution of amino acids tryptophan and also participating in the photosynthesis providing carotenoids \\[[@CR62], [@CR64]\\]. Tryptophan may be involved in the cellular elongation of young leaves of *T. cacao*, as this amino acid is a precursor of indolylacetic acid, a growth hormone \\[[@CR65], [@CR66]\\].\n\nGammaproteobacteria is dominant in the taxonomic composition at class level of phylloplane communities of *T. cacao* (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). It differs from tropical and temperate community structures already described. Phylloplane communities in Canadian forests were dominated by Alphaproteobacteria (68%) \\[[@CR40]\\], contrasting with 27% in Malaysia \\[[@CR21]\\] and 22.8% in Panama \\[[@CR67]\\] in tropical trees. However, percentages of Alphaproteobacteria in trees of tropical climates were similar to percentages found in *T. cocoa*. Some studies report that phylloplane bacteria vary among plants of different developmental stages and genotypes \\[[@CR10], [@CR29]\\]. The bacterial diversity in the phylloplane appears to be as high as that presented in roots or in the human gut \\[[@CR10]\\]. Others describe that microbial diversity and plant species may change according to the environment, climate and geography \\[[@CR22], [@CR26], [@CR40], [@CR68]\\], revealing patterns of change in the phylloplane microbial communities of each species across geographically separated ecosystems. The large diversity of microbiota at the phylloplane may also influence plant evolution, as described by the hologenome theory. Both host and symbiont genomes can be transmitted from one generation to the next \\[[@CR11]\\].\n\n*Sphingomonas* (75.3%), the predominant genus among the exclusive ones in Catongo, has a pigmentation which confers protection against UV radiation to the phylloplane \\[[@CR2]\\], also assists in transportation of substrates (e.g. sugars, vitamins, siderophore) \\[[@CR22]\\] and acts as regulator of stress-related responses, such as PhyR and EcfG \\[[@CR26]\\]. Several members of the genus *Sphingomonas* isolated from plants (*Arabidopsis thaliana*, *Acacia caven*, *Oryza sativa* and *Nicotiana tabacum*) conferred protection in *A. thaliana* against *Pseudomonas syringae* and *Xanthomonas campestris*, reducing disease symptoms or diminishing pathogen growth in the phylloplane \\[[@CR69]\\]. *Sphingomonas melonis* and *Methylobacterium extorquens* demonstrated a profound impact on the transcriptome of the plant *Arabidopsis thaliana*, researchers found that the expression of nearly 400 genes may be involved in the plant defense responses \\[[@CR70]\\]. Nonetheless, *Stenotrophomonas* is the predominant genus (13.5%) among the exclusive ones in CCN51 and it was reported as being characteristic from plant leaves (*Chlorophytum comosum*, *Olea europaea* and *Dracaena draco*), that grow in cold temperate climates \\[[@CR71]\\]. The regulation system of pathogenicity factors (Rpf) and diffusible signal factor (DSF), are also conserved in this type of genus \\[[@CR72]\\]. Other predominant genera among the exclusive ones in CCN51, were *Lysobacter* (5.89%) and *Paenibacillus* (3.13%). *Lysobacter* spp. has been shown to be important as biological control agents, producing both antibiotics and enzymes capable of degrading the cell walls from host fungi in *Cucumis sativus* and *Solanum lycopersicum* \\[[@CR73], [@CR74]\\]. In contrast, a species of the genus *Paenibacillus* (*P. peoriae*), demonstrated a broad inhibition spectrum in several taxonomic groups of bacteria and fungi \\[[@CR75]\\].\n\nRarefaction analyses (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}b) and PCoA analysis show that the total bacterial diversity in the genotype CCN51 phylloplane was larger in comparison to the Catongo genotype and the bacterial community is clustered as per the genotype type (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a). According to the rarefaction curve, the CCN51 genotype of *T. cacao* showed higher OTUs than the Catongo genotype. The rarefaction curves trends to plateau suggests that a good coverage of the entire community of the phylloplane was achieved. The difference in the curve between the biological samples of the CCN51 genotype can be explained by the interval of 15\u2009days between the collection of the first and second biological samples. The highest index of short glandular secreting trichomes that occur in the witch's broom-resistant CCN51 genotype compared to the susceptible Catongo \\[[@CR31]\\], may affect the amount and variety of proteins and metabolites released into the phylloplane \\[[@CR2]\\]. We believe that it might be the cause of the qualitative and quantitative differences in the microbial community of the two genotypes shown in the results. Furthermore, these phylloplane variations, due to plant metabolites and of the microbial community, together with variations in the topography of the phylosphere between genotypes (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}), strongly suggests that they may contribute to the differences in resistance to disease occurring between the CCN51 and Catongo genotypes \\[[@CR31]\\].\n\nBacterial communities presented distinct colonization patterns in the *T. cacao* phylloplane (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). Some studies have described that penetration of the germinal tube of the fungus *M. perniciosa* \\[[@CR76], [@CR77]\\], and colonization of other microorganisms, can occur at the base of the glandular trichoma, junctions of the cells, sites of lesions, stomata, and in the veins \\[[@CR2], [@CR78]\\], and may undergo changes at different seasons and age of the leaf. The diversity is lower during hot and dry months and higher during rainy and cold seasons \\[[@CR79]\\]. In young leaves, communities are made up of a greater number of microorganisms relative to mature and senescent leaves, as well as at different seasons \\[[@CR79], [@CR80]\\]. The formation of aggregates by bacteria may constitute between 30 and 80% of the total bacterial population in certain species of plants \\[[@CR81]\\].\n\nIn the rice phylloplane, the microbiome presented greater diversity in cultivated and controlled plants in pots than those cultivated in the open field \\[[@CR82]\\]. Bacterial community composition in phylloplane of *Deschampsia antarctica* at different locations in open fields, revealed significant differences \\[[@CR83]\\]. The phylloplane and its microbial communities are interrelated \\[[@CR84]\\] and can provide a structural and functional model microenvironment \\[[@CR3], [@CR85]\\] to understand plant-pathogen interactions and thus to select more resistant plants, which will contribute to the continuity of food production.\n\nConclusions {#Sec8}\n===========\n\nIn this study, to the best of our knowledge, we describe for the first time the phylloplane bacterial communities of *T. cacao*. In addition, we performed the first evaluation of hosts identity and an analysis of diversity in two contrasting genotypes for witch's broom resistance. Proteobacteria is the most abundant phylum in the two genotypes, with prevalence of *Candidatus Portiera* in both. Genotype CCN51, resistant to the witch's broom, has a greater diversity of bacterial microbioma in comparison to Catongo and also greater amount of exclusive microorganisms in the phylloplane with antagonistic action against phytopathogens. The bacterial diversity among phylloplane populations are distinct between the genotypes according the PCoA analysis and validated by statistics ANOSIM that showed a significant difference in bacterial composition between genotypes CCN51 and Catongo. The study revealed the importance of epiphytic microbiome and may be a highly valuable tool in the process of biological control. The findings will be of great value for improving the understanding of the defense and interaction mechanisms that occur in the phylloplane.\n\nAdditional files\n================\n\n {#Sec9}\n\nAdditional file 1:**Figure S1.** Distribution of plants in the greenhouse. (A) Selected plants: green (first biological), red (second biological). (B) Panoramic photo of plants. (DOCX 2745 kb) Additional file 2:**Figure S2.** Extraction of the metagenomic DNA in triplicates experimental. (A) First biological sample - CCN51. (B) First biological sample - Catongo. (C) Second biological sample - CCN51 and (D) Second biological sample - Catongo. (DOCX 151 kb) Additional file 3:**Figure S3.** Quantification of libraries. (A) Electrophoresis on 1% (w / v) agarose gel with the six standards, 12 libraries (quantified in triplicates) and three negative controls -- a, b and c: first biological sample - CCN51; d, e and f: first biological sample - Catongo; g, h and i: second biological sample - CCN51; j, k and l: second biological sample - Catongo; NC: negative control. (B) Dissociation curve -- a: libraries; b: negative control. (DOCX 307 kb) Additional file 4:**Table S1.** Bacteria identified and classified according to phylum, class, order, family, and genus for in the genotypes CCN51 and Catongo, with a threshold of 99% identity against the Greengen database version 13.8 16S rRNA. (DOCX 19 kb) Additional file 5:**Table S2.** Bacterial genera exclusive to the phylloplane of the genotypes CCN51 and Catongo. (DOCX 17 kb)\n\nANOSIM\n\n:   Analysis of similarity\n\nDSF\n\n:   Diffusible signal factor\n\nLWW\n\n:   Leaf water washes\n\nNGS\n\n:   Next-generation sequencing technology\n\nOTUs\n\n:   Operational taxonomic units\n\nPCoA\n\n:   Principal coordinates analysis\n\nPCR\n\n:   Polymerase chair reaction\n\nQIIME\n\n:   Quantitative Insights Into Microbial Ecology\n\nRpf\n\n:   System of regulation of pathogenicity factors\n\nSEM\n\n:   Scanning Electron Microscope\n\nWe thank to Dr. Pedro Mangabeira for her help in electron microscopy and Dra. Cl\u00e1udia Fortes for the revision of the english language.\n\nFunding {#FPar1}\n=======\n\nAuthors would like to acknowledge thank CAPES (PDSE2016--88881.132547 / 2016--01) for the funding of the sandwich bag. This work also was supported by FCT through funding of the DeST: Deep Semantic Tagger project, ref. PTDC/CCI-BIO/28685/2017, LaSIGE Research Unit, ref. UID/CEC/00408/2013 and BioISI, ref. ID/MULTI/04046/2013. The funding bodies had no role in the design of the study, in data collection, analysis or interpretation, or in writing the manuscript.\n\nAvailability of data and materials {#FPar2}\n==================================\n\nThe data that support the findings of this study are available from the corresponding author upon reasonable request.\n\nConceived and designed the experiments: CPP, JOS. Contributed into experimental design: KPG. Preparation of the libraries and sequencing: JOS, KTdSEF. Extraction of metagenomic DNA and reagents: CPP, RPR. Scanning electron microscopy: PAOM. Metagenomic data analysis: FMC, RPMD, JOS. All authors had access to the data, reviewed and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar3}\n==========================================\n\nNot applicable.\n\nConsent for publication {#FPar4}\n=======================\n\nNot applicable.\n\nCompeting interests {#FPar5}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar6}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nGERD is usually characterized by symptoms caused by esophageal mucosal exposure to acidic gastric contents. The pathophysiology is related to failure of the lower esophageal sphincter (LES) to act as a barrier to gastric reflux. The adequacy of LES function is dependent on its pressure, total length and portion exposed to positive intra-abdominal pressure. Defects in one or a combination of the above factors will lead to reflux. ([@R1]) The prevalence of GERD has been increasing for the last 2 decades. ([@R2]) It affects 10-20 % of the adult population in the developed world, with its highest prevalence in North America. An estimated 10 billion dollars is spent treating GERD in the USA every year. ([@R2])\n\nTreatment of GERD begins with lifestyle modifications. Proton pump inhibitors (PPIs) have been shown by multiple randomized controlled trials (RCT) to be the most effective medical therapy for GERD. Several large meta-analysis of multiple RCT showed that PPIs are superior to histamine 2 receptor blockers (H2RB) in short term treatment of esophagitis and long term maintenance therapy to prevent symptomatic relapse and endoscopicaly proven esophagitis. ([@R3],[@R4])\n\nMost patients with GERD are well controlled with medical management. However, surgery should be considered in patients with progressive disease, extraesophageal symptoms, and GERD-related complications such as strictures or Barrett\\'s esophagus. It should also be considered in younger patients intolerant of PPIs or post menopausal female patients at risk for fractures due to PPI-induced calcium malabsorption.\n\nLaparoscopic Nissen fundoplication, which involves a 360 degree wrap of the gastric fundus around the esophagus at the gastro-esophageal junction, is the current surgical procedure of choice for severe, refractory GERD. Studies have shown over 90 % success in symptom resolution, improved quality of life and overall improvement of general health in over 70 % of patients. Reoperation for symptoms of heartburn, dysphagia, bloating and inability to belch due to disruption or slippage of a fundoplication has been reported in up to 7% of patients after laparoscopic Nissen fundoplication. ([@R5])\n\nCASE REPORT {#s2}\n===========\n\nThe patient was a twenty nine years old male with long standing Gastro-Esophageal Reflux Disease (GERD) who had been treated with high dose proton pump inhibitors (PPIs) for over 3 years with little response. He presented for surgical consultation. The patient complained of symptoms of heartburn gradually progressing to throat pain, ear pain and tinnitus. Symptoms would worsen when laying supine, after a large meal and with specific foods such as chocolate, spicy foods and alcohol. The patient had some relief of his symptoms with PPIs initially but then the symptoms recurred.\n\nPre-operative upper endoscopy showed a Hill grade I gastro-esophageal valve, ([Fig. 1](#fig1){ref-type=\"fig\"}) grade A esophagitis and normal gastric and duodenal mucosa. Urease test was negative for Helicobacter pylori. Motility study showed only 25% of swallows produced peristaltic wave throughout the esophagus with a mean distal amplitude of 77mmHg. The remaining swallows produced a mixture of low amplitude synchronous waves, partially propagated and non-propagated waves. The lower esophageal sphincter (LES) measured to 3 cm, with 1 cm above the diaphragm. The LES had a mean resting pressure of 9 mmHg, which relaxed with swallowing. A twenty four hour pH study produced a deMeester score of 20.25 (normal range below 14.5).\n\n![pre-operative EGD showing Hill grade I gastro-esophageal valve](jscr-2010-5-1fig1){#fig1}\n\nThe patient elected to undergo a transoral incision-less fundaplication, in which a 4 cm gastro-esophageal valve was created after the endoscopic application of 14 polypropylene transmural fasteners through the gastric fundus and the distal esophagus (1-3 cm above the Z-line). The procedure also produced a 270 degree endoluminal wrap of the fundus around the intra-abdominal esophagus. ([Fig 2](#fig2){ref-type=\"fig\"}) The total operative time was 75 minutes. There were no major intra-operative complications, minimal intra-operative blood loss and the patient was discharged home the morning after surgery.\n\n![Post operative view of the esophageal valve.](jscr-2010-5-1fig2){#fig2}\n\nEarly after surgery, the patient complained of throat pain and abdominal bloating. These complaints, however, soon resolved. Two months after the procedure, the patient\\'s symptoms of reflux had significantly resolved. He was almost completely off his PPIs, he could eat many of the foods which gave him reflux before surgery without any difficulty and he could now sleep lying flat, something which he had been unable to do for three years.\n\nDISCUSSION {#s3}\n==========\n\nA novel approach in the evolution of the surgical treatment of GERD is the transoral incisionless fundoplication (TIF). TIF employs a tissue plication device called EsophyX (EndoGastric Solutions, Redmond, WA, USA) that endoscopically replicates a 270-320 degree gastric fundoplication to create a 3-4cm neo-valve at the GE junction. The procedure is done under general anesthesia, with either oro-tracheal or naso-tracheal intubation, with the patient positioned in either the left lateral decubitus or supine position on the operating table. The operating team consists of an endoscopist who provides continued retroflex visualization of the GE junction and a second operator, usually a surgeon, who performs the actual tissue manipulation and plication.\n\nAfter a single transoral introduction of the instrument over the endoscope, the gastric fundus is retracted by a helical retractor. This creates a 3-4 cm long sleeve of tissue that is plicated by 6mm H-shaped polypropylene fasteners in a 200 to 300 degree fashion. Full thickness serosa-serosa plication at 1-2 cm above the gastro-esophageal junction begins with deployment of fasteners at the posterior and anterior side as visualized with the endoscope in the retroflexed view. Plication then continues at the greater curvature, 3-4 cm above the gastro-esophageal junction as the tissue is being pulled down. ([Fig 3](#fig3){ref-type=\"fig\"}) By the end of the procedure (which typically takes 45-75 minutes), TIF increases the length of intra-abdominal esophagus, reduces small hiatal hernias, creates a 3-4cm GE valve and recreates the angle of His.\n\n![Full thickness serosa-serosa plication at 1-2 cm and 3-4 above the gastro-esophageal junction.](jscr-2010-5-1fig3){#fig3}\n\nThe performance of TIF with EsophyX has been FDA approved in the USA since September 2007. Over the last 2 and half years, 2000 cases have been performed in the USA. A further 1000 cases have been performed in Europe. The efficacy of the procedure has been reported in several studies. Cadiere and his group were able to show no adverse effects, greater then 50% improvement in GERD-HRQL scores, elimination of heartburn in 93% of patients, PPI use in 71% of patients and esophagitis in 55% of patients at 2 year follow up. ([@R6]) Newer unpublished data from two case series (n = nearly 70 patients) from two independent investigators in United States demonstrate normalization of esophageal pH in 76 % of patients and complete independence from PPI\\'s in approximately 90% of patients. 87% of subjects were satisfied and 82% were asymptomatic based on HRQL and GSRS scores.\n\nComplication rates are low after TIF. Early on in the European experience, three patients sustained non-lethal esophageal perforations during passage of the EsophyX device. No further cases of perforation have been reported since. Intra-operative bleeding requiring blood transfusion is very uncommon.\n\nTIF with EsophyX is effective in decreasing symptom severity, daily PPI requirements and improving quality of life (in the short term) in patients with moderate to severe GERD. TIF can also reduce small (\\<2cm) hiatal hernias. TIF with EsophyX successfully increases LES resting pressure, decreases acid reflux and supports esophageal healing in chronic GERD. TIF with EsophyX is less invasive than its laparoscopic counterpart. Long term studies are needed to establish whether it is will be as durable as more traditional forms of anti-reflux surgery.\n"}
+{"text": "Background {#Sec1}\n==========\n\nOpen-chest cardiopulmonary resuscitation (OCCPR) came into use in the USA in the late 1800s as the salvage maneuver following cardiac arrest. It simultaneously includes control of infra diaphragmatic hemorrhage by cross-clamping of the descending thoracic aorta, in addition to direct cardiac massage, when necessary \\[[@CR1], [@CR2]\\]. Because it is impractical to conduct randomized controlled trials comparing between OCCPR and closed-chest cardiopulmonary resuscitation (CCCPR) in traumatic cardiac arrest cases due to ethical reasons, the therapeutic impact of OCCPR has only been evaluated in observational and cohort studies \\[[@CR3]--[@CR5]\\]. However, the survival benefit of OCCPR compared to CCCPR has not been clearly determined even in a recent meta-analysis \\[[@CR6]\\].\n\nBecause of the poor cost-effectiveness and potential infectious risks to medical staff \\[[@CR7], [@CR8]\\], in addition to the aforementioned evidence, recent guidelines have restricted the indications for the use of resuscitative thoracotomy (RT) \\[[@CR2], [@CR9], [@CR10]\\]. The indications for RT in those guidelines are generally based upon a positive finding of signs of life (SOL: detectable blood pressure, respiratory or motor effort, cardiac electrical activity, or pupillary activity) and the time from onset of cardiac arrest because patient survival is believed to be rare after more than 15\u2009min of cardiopulmonary resuscitation \\[[@CR11]\\]. However, most of the recommendations have been based on descriptive studies of small sample sizes or expert opinions and not based on a large-scale cohort study evaluating the comparative effectiveness of OCCPR to CCCPR.\n\nIn the present study, we evaluated the effectiveness of OCCPR, compared to CCCPR, in trauma patients who had SOL upon emergency department (ED) admission, based on the hypothesis that OCCPR is associated with better survival outcomes than CCCPR in those patients.\n\nMethods {#Sec2}\n=======\n\nStudy design and settings {#Sec3}\n-------------------------\n\nA retrospective cohort study analyzing data of the Trauma Quality Improvement Program (TQIP) database between 2010 and 2016 was conducted. The TQIP database is a subset of the National Trauma Databank of the American College of Surgeons Committee on Trauma, which stores data of patients aged more than 15\u2009years and suffered severe injury, defined by abbreviated injury scale (AIS) \u2265\u20093. At the end of 2016, more than 700 level 1 and level 2 trauma centers participated in the TQIP database. Trained specialists abstracted more than 100 variables for each patient as well as information on the treating hospital. Details in the TQIP database are available at <https://www.facs.org/quality-programs/trauma/tqp/center-programs/tqip>. Survival to hospital discharge of patients who received OCCPR was compared to that of patients who received CCCPR only.\n\nThis study complied with the principles of the 1964 Helsinki Declaration and its later amendments. The study and its protocols were in compliance with the institutional review board of Riverside University Health System---Comparative Effectiveness and Clinical Outcomes Research Center (approval number: 1636962). The requirement for informed consent for each patient was waived based on the use of anonymized patient and hospital data.\n\nStudy population {#Sec4}\n----------------\n\nWe included trauma patients who met the following criteria: (i) presence of SOL upon hospital arrival and (ii) received OCCPR or CCCPR within 6\u2009h of hospital arrival. Due to the inclusion criteria of the TQIP database, patients younger than 16\u2009years were not included. We also excluded patients who had a nonsurvivable injury defined by 6 points in AIS, patients without exact information on injury mechanism, or patients without exact information on SOL upon ED arrival. The patients were divided into the OCCPR and the CCCPR groups, and their outcomes were compared between the two groups.\n\nData collection {#Sec5}\n---------------\n\nThe following variables were collected from the TQIP database: age, gender, insurance type, vital signs (systolic blood pressure, heart rate, respiratory rate), Glasgow Coma Scale (GCS), body temperature, presence or absence of SOL upon ED arrival, year of injury, injury mechanism (i.e., blunt or penetrating), AIS in each body region, Injury Severity Score (ISS), total prehospital transport time, implementation of OCCPR or CCCPR and their timing (recorded on an hour basis after ED arrival), length of hospital stay, and survival status at hospital discharge. Hospital information regarding trauma center level, teaching status, and number of trauma surgeons was also collected.\n\nOutcome measure and cohort definitions {#Sec6}\n--------------------------------------\n\nThe study outcome was defined as survival to hospital discharge. The OCCPR group was defined as patients who received OCCPR within the first 6\u2009h of ED arrival with or without CCCPR prior to OCCPR, considering the clinical importance of the first 6\u2009h after injury \\[[@CR12]\\]. Meanwhile, the CCCPR group was defined as patients who received CCCPR only, within the first 6\u2009h of ED arrival. Patients who had both codes for OCCPR and CCCPR were classified into the OCCPR group because executing CCCPR is common during the preparation for OCCPR, but the reverse cannot be true from a practical clinical perspective. The implementation of OCCPR and CCCPR was identified using the procedure codes of the International Classification of Diseases 9th Revision Clinical Modification (ICD-9-CM) 37.91 and 99.63, respectively.\n\nStatistical analysis {#Sec7}\n--------------------\n\nMissing values were treated by multiple imputation using chained equations with 10 iterations and creation of 15 datasets, based on the assumption of missing at random in missing mechanism as well as previous studies using the TQIP database \\[[@CR13], [@CR14]\\].\n\nThree statistical models were used for analyses: (i) logistic regression analysis, (ii) instrumental variable analysis, and (iii) propensity score matching analysis. Covariates used for case-mix classification, used in the logistic regression model and in the instrumental variable model, included patient age, gender, insurance type, year of injury, injury mechanism (i.e., penetrating or blunt), vital signs upon ED arrival (systolic blood pressure, heart rate, respiratory rate), GCS, and body temperature at ED arrival, maximum AIS by body region, ISS, total prehospital transport time, and hospital characteristics (American College of Surgeons verification level and teaching status). The variables were selected based on clinical perspective. Issues with variable multicollinearity were assessed using variance inflation factor (VIF) analysis, and the tolerance value was set at less than 2. In the instrumental variable analysis, which is an established technique to control unmeasured confounding in non-randomized data \\[[@CR15]\\], the number of trauma surgeons in a hospital (categorized by whether more or less than 8 surgeons) was used as the instrumental variable. The cut-off value of the instrumental variable was determined according to the categorization in the number of trauma surgeons in the TQIP database. This approach was conducted using a two-stage least-squares regression adjusted by the aforementioned variables, based on the null hypothesis that there was no association between the number of trauma surgeons in a hospital and the actual implementation of OCCPR. A partial *F* test was conducted to assess an issue of weak instruments, and a value of *F*-statistic more than 10 was regarded as acceptable.\n\nConsidering the heterogeneity in the characteristics between the OCCPR and CCCPR groups, we also performed a propensity score matching analysis \\[[@CR16]\\]. In this analysis, a logistic regression model was applied to estimate the propensity score to predict OCCPR in each patient using the variables mentioned above. Propensity score matching extracted 1:1 matched pairs from the OCCPR and CCCPR groups. Match balance between the two groups was assessed using the absolute standardized mean difference (ASMD), and values of less than 0.1 were considered acceptable. The caliper width was set as the standard deviation of the logit-transformed propensity score multiplied by 0.002 to achieve well match balance between two groups. The chi-square test was used for intergroup comparison in the propensity score-matched cohort. Furthermore, cumulative incidence curves for the in-hospital mortality in the propensity score-matched subjects were constructed. The Fine and Gray test was used to estimate the subdistribution hazard ratio for in-hospital mortality, considering the competing risk between in-hospital mortality and survival discharge \\[[@CR17]\\].\n\nSince the information on survival outcome was lacking in some patients, in addition to the multiple imputation method, we performed sensitivity analyses in which the outcome was imputed based on the most optimistic and pessimistic scenarios, where all the missing information on survival outcome was assumed as survival or death, respectively. The aforementioned logistic regression model was applied in these sensitivity analyses.\n\nDescriptive statistics were used to display categorical variables as counts and percentages, and numeric or ordered variables as medians and 25th--75th percentiles, after pooling all the imputed datasets into one dataset. Predictive statistics were used to display the estimators as point estimation and 95% confidence intervals (CIs) integrated across the imputed datasets, based on Rubin's rule \\[[@CR18]\\]. The level of significance was defined as *p*\u2009\\<\u2009\u00a00.05 for all statistical analyses. All the analyses were performed using R 3.5.3 (R Foundation for Statistical Computing, Vienna, Austria) with add-on packages of \"mice \\[[@CR19]\\]\" for multiple imputation, \"Matching \\[[@CR20]\\]\" for propensity score matching, \"AER \\[[@CR21]\\]\" for instrumental variable analysis, and \"cmprsk \\[[@CR22]\\]\" for the Fine and Gray test.\n\nResults {#Sec8}\n=======\n\nThe flow diagram of patients is shown in Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}. A total of 2682 patients (OCCPR 1032; CCCPR 1650) were eligible for analysis. The characteristics of na\u00efve data, including the proportion of missing values, are shown in Additional\u00a0file\u00a0[1](#MOESM1){ref-type=\"media\"}. The major baseline characteristics of the patients of the multiply imputed cohort are shown in Table\u00a0[1](#Tab1){ref-type=\"table\"}. The OCCPR group was older than the CCCPR group (median \\[25th--75th percentiles\\]\u2009=\u200945 \\[28, 63\\] and 32 \\[23, 49\\], respectively), and the proportion of blunt trauma was higher in the OCCPR group (71.8%) than in the CCCPR group (42.7%). Fig. 1Flow diagram of patient selection. Abbreviations: SOL, signs of life; OCCPR, open-chest cardiopulmonary resuscitation; CCCPR, closed-chest cardiopulmonary resuscitationTable 1Baseline characteristics of the patients in the multiply imputed dataset (major variables)VariablesOCCPR (*n*\u2009=\u20091032)CCCPR (*n*\u2009=\u20091650)ASMDAge, years old, median \\[IQR\\]45 \\[28, 63\\]32 \\[23, 49\\]0.45Gender, female, *n* (%)267 (25.9)311 (18.8)0.17Type of injury\u2003Blunt741 (71.8)705 (42.7)0.61\u2003Penetrating291 (28.2)945 (57.3)0.61\u2003Total prehospital transport time, min, median \\[IQR\\]48 \\[32, 88\\]35 \\[25. 59\\]0.15\u2003Transfer from another hospital, yes, *n* (%)110 (10.7)117 (7.1)0.13Highest AIS score per body region, median \\[IQR\\]\u2003Head3 \\[0, 5\\]0 \\[0, 1\\]0.79\u2003Face0 \\[0, 1\\]0 \\[0, 0\\]0.28\u2003Neck0 \\[0, 0\\]0 \\[0, 0\\]0.02\u2003Chest3 \\[0, 4\\]3 \\[3, 4\\]0.45\u2003Abdomen0 \\[0, 3\\]3 \\[0, 4\\]0.43\u2003Spine0 \\[0, 2\\]0 \\[0, 0\\]0.22\u2003Upper extremities0 \\[0, 2\\]0 \\[0, 1\\]0.02\u2003Pelvis and lower extremities0 \\[0, 3\\]0 \\[0, 2\\]0.17\u2003Skin/superficial0 \\[0, 0\\]0 \\[0, 0\\]0.00Injury Severity Score26 \\[19, 35\\]26 \\[20, 36\\]0.10Systolic blood pressure, mmHg, median \\[IQR\\]100 \\[72, 132\\]97 \\[69, 127\\]0.09Heart rate, bpm, median \\[IQR\\]103 \\[70, 129\\]110 \\[77, 133\\]0.12Respiratory rate, bpm, median \\[IQR\\]16 \\[0, 22\\]17 \\[8, 24\\]0.11Body temperature, \u00b0C, median \\[IQR\\]36.0 \\[35.1, 36.5\\]36.0 \\[35.2, 36.5\\]0.00Glasgow Coma Scale, median \\[IQR\\]3 \\[3, 9\\]3 \\[3, 13\\]0.22Time from ED arrival to OCCPR, hours, median \\[IQR\\]1 \\[0, 1\\]----Time from ED arrival to CCCPR, hours, median \\[IQR\\]0 \\[0, 1\\]0 \\[0, 1\\]0.25*Abbreviations: OCCPR* open-chest cardiopulmonary resuscitation, *OCCPR* closed-chest cardiopulmonary resuscitation, *ASMD* absolute standardized mean difference, *IQR* interquartile range, *AIS* abbreviated injury scale, *ED* emergency department\n\nDetailed baseline characteristics of the cohort are shown in Additional file [2](#MOESM2){ref-type=\"media\"}. The characteristics of the hospitals of the cohort are shown in Table\u00a0[2](#Tab2){ref-type=\"table\"}. The number of trauma surgeons was higher in the OCCPR group than in the CCCPR group. In total, 157 patients (15.2%) in the OCCPR group and 193 patients (11.7%) in the CCCPR group survived to hospital discharge (crude odds ratio \\[95% CI\\]\u2009=\u20091.35 \\[1.06--1.73\\], *p*\u2009=\u20090.017). The median length of hospital stay \\[25th--75th percentiles\\] among the survivors in the OCCPR group and the CCCPR group was 18\u2009days \\[6--35\\] and 19\u2009days \\[10--32\\], respectively. Table 2Baseline characteristics of the hospitals in the multiply imputed datasetVariablesOCCPR (*n*\u00a0=\u20091032)CCCPR (*n*\u2009=\u20091650)ASMDACS trauma center level, *n* (%)\u2003I807 (78.2)1248 (75.6)0.06\u2003II225 (21.8)402 (24.4)0.06Teaching status, *n* (%)\u2003University719 (69.7)1097 (66.5)0.07\u2003Community226 (21.9)470 (28.5)0.15\u2003Non-teaching87 (8.4)83 (5.0)0.14Number of trauma surgeons, \\>\u20098, *n* (%)400 (38.8)445 (27.0)0.25*Abbreviations: OCCPR* open-chest cardiopulmonary resuscitation, *OCCPR* closed-chest cardiopulmonary resuscitation, *ASMD* absolute standardized mean difference, *ACS* American College of Surgeons\n\nIn the logistic regression analysis, all of the VIFs of used variables were lower than 2, which eliminated the issue of multicollinearity in our model. OCCPR was significantly associated with higher survival at discharge in the logistic regression analysis (adjusted odds ratio \\[95% CI\\]\u2009=\u20091.99 \\[1.42--2.79\\], *p*\u2009\\<\u2009\u00a00.001). In the instrumental variable analysis, the linear regression analysis demonstrated that there was a significant increase in the likelihood of OCCPR implementation according to the number of trauma surgeons (adjusted odds ratio \\[95% CI\\]\u2009=\u20091.11 \\[1.08--1.15\\], *p*\u2009\\<\u2009\u00a00.001, *F*-statistic\u2009=\u200928.9). Therefore, the null hypothesis that there was no association between the number of trauma surgeons and the actual implementation of OCCPR was rejected. However, survival to hospital discharge was not significantly affected by the instrumental variable in the linear regression analysis adjusted by OCCPR (adjusted odds ratio \\[95% CI\\]\u2009=\u20091.01 \\[0.98--1.04\\], *p*\u2009=\u20090.485). Therefore, the variable \"number of trauma surgeons in a hospital\" satisfied the requirements of the instrumental variable. The two-stage least-squares analysis with this instrumental variable also demonstrated significantly higher survival associated with OCCPR (adjusted odds ratio \\[95% CI\\]\u2009=\u20091.16 \\[1.02--1.31\\], *p*\u2009=\u20090.021).\n\nIn the propensity score matching analysis, through the one-to-one matching process, 531 matched pairs were selected. The major baseline characteristics of the patients in the multiply imputed and propensity score-matched cohort are shown in Table\u00a0[3](#Tab3){ref-type=\"table\"}. The detailed baseline characteristics of the cohort are shown in Additional\u00a0file\u00a0[3](#MOESM3){ref-type=\"media\"}. The characteristics of the hospitals of the cohort are shown in Table\u00a0[4](#Tab4){ref-type=\"table\"}. Table 3Baseline characteristics of the patients of the multiply imputed and propensity score-matched dataset (major matched variables)VariablesOCCPR (*n*\u2009=\u2009531)CCCPR (*n*\u2009=\u2009531)ASMDAge, years old, median \\[IQR\\]39 \\[26, 56\\]40 \\[25, 56\\]\\<\u20090.01Gender, female, *n* (%)127 (23.9)127 (23.9)\\<\u20090.01Type of injury\u2003Blunt339 (63.8)342 (64.4)0.01\u2003Penetrating192 (36.2)189 (35.6)0.01Total prehospital transport time, min, median \\[IQR\\]44 \\[30, 73\\]39 \\[28, 69\\]0.02Transfer from another hospital, yes, *n* (%)47 (8.9)47 (8.9)\\<\u20090.01Highest AIS score per body region, median \\[IQR\\]\u2003Head0 \\[0, 3\\]0 \\[0, 4\\]0.07\u2003Face0 \\[0, 1\\]0 \\[0, 1\\]\\<\u2009 0.01\u2003Neck0 \\[0, 0\\]0 \\[0, 0\\]0.04\u2003Chest3 \\[3, 4\\]3 \\[3, 4\\]0.05\u2003Abdomen2 \\[0, 3\\]2 \\[0, 4\\]0.07\u2003Spine0 \\[0, 0\\]0 \\[0, 2\\]0.03\u2003Upper extremities0 \\[0, 2\\]0 \\[0, 2\\]0.03\u2003Pelvis and lower extremities0 \\[0, 3\\]0 \\[0, 3\\]\\<\u20090.01\u2003Skin/superficial0 \\[0, 0\\]0 \\[0, 0\\]0.03Injury Severity Score26 \\[19, 35\\]27 \\[20, 38\\]0.08Systolic blood pressure, mmHg, median \\[IQR\\]92 \\[67, 127\\]96 \\[69, 130\\]0.05Heart rate, bpm, median \\[IQR\\]107 \\[71, 130\\]106 \\[75, 130\\]0.03Respiratory rate, bpm, median \\[IQR\\]16 \\[0, 22\\]16 \\[8, 24\\]0.09Body temperature, \u00b0C, median \\[IQR\\]36.0 \\[35.0, 36.5\\]36.0 \\[35.3, 36.5\\]0.04Glasgow Coma Scale, median \\[IQR\\]3 \\[3, 11\\]3 \\[3, 12\\]0.06*Abbreviations: OCCPR* open-chest cardiopulmonary resuscitation, *OCCPR* closed-chest cardiopulmonary resuscitation, *ASMD* absolute standardized mean difference, *IQR* interquartile range, *AIS* abbreviated injury scale, *ED* emergency department, *ACS* American College of SurgeonsTable 4Baseline characteristics of the hospitals of the multiply imputed and propensity score-matched datasetVariablesOCCPR (*n*\u2009=\u2009531)CCCPR (*n*\u2009=\u2009531)ASMDACS trauma center level, *n* (%)\u2003I424 (79.9)416 (78.3)0.04\u2003II107 (20.1)115 (21.7)0.04Teaching status, *n* (%)\u2003University377 (71.0)362 (68.2)0.06\u2003Community125 (23.5)136 (25.6)0.05\u2003Non-teaching29 (5.5)33 (6.2)0.03*Abbreviations: OCCPR* open-chest cardiopulmonary resuscitation, *OCCPR* closed-chest cardiopulmonary resuscitation, *ASMD* absolute standardized mean difference, *ACS* American College of Surgeons\n\nAll the values of ASMDs in the adjusted variables were less than 0.1, indicating a well-matched balance (Tables\u00a0[3](#Tab3){ref-type=\"table\"} and [4](#Tab4){ref-type=\"table\"} and Additional file [3](#MOESM3){ref-type=\"media\"}). In the propensity score-matched cohort, the median age \\[25th--75th percentiles\\] was 40\u2009years old \\[25--56\\], and 254 patients (23.9%) were female. Penetrating injury was observed in 381 patients (35.9%). In total, 89 patients (16.8%) in the OCCPR group and 58 patients (10.9%) in the CCCPR group survived to hospital discharge. The OCCPR group showed significantly higher survival also in the propensity score-matched cohort (odds ratio \\[95% CI\\]\u2009=\u20091.66 \\[1.13--2.42\\], *p*\u2009=\u20090.009). The results of these three analyses are summarized in Table\u00a0[5](#Tab5){ref-type=\"table\"}. Table 5Comparative effectiveness of OCCPR, compared to CCCPR, for survival to hospital discharge evaluated by the logistic regression analysis, instrumental variable analysis, and propensity score matching analysisModelsNumber of survivorsAdjusted odds ratio \\[95% confidence interval\\]*p* valueOCCPRCCCPRLogistic regression analysis157/1032 (15.2%)293/1650 (11.7%)1.99 \\[1.42--2.79\\]\\<\u20090.001Instrumental variable analysis157/1032 (15.2%)293/1650 (11.7%)1.16 \\[1.02--1.31\\]0.021Propensity score matching analysis89/531 (16.8%)58/531 (10.9%)1.66 \\[1.13--2.42\\]0.009*Abbreviations: OCCPR* open-chest cardiopulmonary resuscitation, *CCCPR* closed-chest cardiopulmonary resuscitation\n\nThe cumulative incidence curve of in-hospital mortality until 30\u2009days after admission in the propensity score-matched subjects is shown in Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}. A marked difference in mortality was observed, particularly in the first day of admission. The Fine and Gray test showed that OCCPR was significantly associated with lower in-hospital mortality (subdistribution hazard ratio \\[95% confidence intervals\\]\u2009=\u20090.92 \\[0.86--0.98\\], *p*\u2009=\u20090.009). Fig. 2Cumulative incidence curves for in-hospital mortality in the propensity score-matched subjects\n\nIn the sensitivity analyses, where all the missing information on the outcome was assumed as survival, as in the original analysis, OCCPR was significantly associated with higher survival at hospital discharge (adjusted odds ratio \\[95% CI\\]\u2009=\u20092.87 \\[2.32--3.55\\], *p*\u2009\\<\u20090.001). The similar result was observed in the analysis where all the missing information on the outcome was assumed as death (adjusted odds ratio \\[95% CI\\]\u2009=\u20091.80 \\[1.31--2.47\\], *p*\u2009\\<\u20090.001).\n\nDiscussion {#Sec9}\n==========\n\nIn the present retrospective observational study, we evaluated the survival benefit of OCCPR compared to CCCPR in trauma patients who had SOL upon ED arrival. All of the three statistical models indicated a significant survival benefit of OCCPR. To the best of our knowledge, this is the first study to validate the indications for OCCPR by using a large-scale dataset and demonstrates a more favorable survival outcome of OCCPR compared to those of CCCPR.\n\nOCCPR has several theoretical advantages over CCCPR in trauma patients' resuscitation efforts \\[[@CR1], [@CR2]\\]. It would be difficult to achieve sufficient systemic blood perfusion by CCCPR in cases of multiple rib fractures and flail chest due to the reduced rib cage compliance. Cross-clamping of the descending thoracic aorta, which is usually combined with OCCPR, enables maintenance of cerebral/coronary blood perfusion and temporal control of infra diaphragmatic hemorrhage. However, previous studies failed to demonstrate the advantages of OCCPR compared to CCCPR \\[[@CR3], [@CR4]\\]. The discrepancy might be mainly explained by the difference in the study populations. The evaluated population in the present study was limited to patients who had SOL upon ED arrival; however, 41.2% of evaluated patients in a previous study \\[[@CR3]\\] had cardiac arrest upon ED arrival. Moreover, 46.1% of the present study cohort were victims of penetrating mechanism, whereas all of the analyzed patients in a previous study were injured by blunt mechanisms \\[[@CR3]\\]. The differences in the prehospital medical system and the distribution of trauma center as well as their coverage area might also have influenced the results of the studies. While the prehospital transport time was longer in the present study than that in a previous Japanese study \\[[@CR3]\\], paramedics in the USA are allowed to provide a variety of medical interventions compared to their Japanese counterparts. The differences in these factors might have largely affected the survival rate observed in the present study (13.0% in the overall study cohort), which was higher than that of previously reported \\[[@CR23]--[@CR26]\\].\n\nJoseph et al. \\[[@CR27]\\] reviewed patients who underwent exploratory thoracotomy (ICD-9-CM, 34.02) within the first hour of hospital admission using the TQIP database and reported 9.6% survival rate. However, their study included patients without SOL on hospital arrival (28.6% of the analyzed cohort). Furthermore, OCCPR is only a part of exploratory thoracotomy and is not always performed during the procedure \\[[@CR28], [@CR29]\\]. In contrast, the present study used more strict definitions for the interventions (i.e., OCCPR or CCCPR), and this might explain some differences in patient background and the results between their research and ours.\n\nThe duration from the time of cardiac arrest to OCCPR implementation is one of the important indicators in the recommendation of the current guidelines \\[[@CR2], [@CR9], [@CR10]\\]. Actually, Yamamoto et al. \\[[@CR30]\\] reported that a short-duration CCCPR time before OCCPR was associated with significantly higher recovery of spontaneous circulation (ROSC) rates. However, in a retrospective database analysis, there remained a major concern that OCCPR was performed only in patients who could not achieve ROSC by CCCPR alone \\[[@CR6]\\]. This bias is generally associated with worse outcomes in patients who received OCCPR. Although the present study could not overcome this bias due to the nature of the TQIP database in which detailed time course data in minutes were not available, considering the direction of this bias, the results suggested that the presence of SOL might be one of the strong indicators for the implementation of OCCPR. However, further studies accounting for the detailed time course are necessary to control this bias.\n\nThe strength of the present study was that a large number of patients were analyzed using several rigorous statistical approaches. The data on the presence or absence of SOL, one of the indications for RT in the current guidelines \\[[@CR2], [@CR9], [@CR10]\\], were available in the TQIP database. Moreover, the procedures of OCCPR and CCCPR could be clearly identified. However, this study also had some limitations. Since this was a retrospective registry-based study, the issue of residual confounding was unavoidable. The indication for OCCPR cannot be fully explained by registry variables. Although we made the best effort to overcome this bias by using an instrumental variable model, specifying an ideal instrumental variable is challenging in a retrospective registry-based study because there is no established method to verify the inexistence of the pathway from the instrumental variable to the outcome. The issue of potential uncertainty in the registry data, including ICD-9-CM procedure codes, was also one of the limitations of this study. There were missing data to some degree, including the study outcome, to which we used a multiple imputation method. Detailed information on time course in minutes was unavailable, as described above. Data on neurological outcomes, the ultimate outcome measure of resuscitation treatments, were also unavailable in the TQIP database. Finally, data on patients younger than 16\u2009years were not available, which prevented us from evaluating this age group.\n\nDespite these limitations, this was the first well-designed large-scale study that corroborated one of the indications in the current recommendations for the use of OCCPR in trauma patients. The results of our study showed a significant association between OCCPR implementation and favorable survival outcome in trauma patients who had SOL on ED arrival. From these results, it would be expected that patient survival increases by maintaining SOL until ED arrival combined with subsequent OCCPR. Further studies from other countries are necessary to confirm the generalizability of our findings. In addition, the influence of time from onset of cardiac arrest, which is another major criterion in the current trauma guideline, should also be statistically evaluated in future studies to identify the best possible candidates for OCCPR.\n\nConclusions {#Sec10}\n===========\n\nOCCPR was associated with significantly higher survival to hospital discharge than CCCPR in trauma patients with SOL upon ED arrival. Further studies to confirm these results and to assess neurological outcomes, taking detailed time course into account, are required.\n\nSupplementary information\n=========================\n\n {#Sec11}\n\n**Additional file 1: Table S1** Baseline characteristics and proportion of missing data in na\u00efve dataset.**Additional file 2: Table S2** Baseline characteristics of the patients in the multiply imputed dataset (all variables).**Additional file 3.** Baseline characteristics of the patients in the multiply imputed and propensity score-matched dataset (all the matched variables).\n\nAIS\n\n:   Abbreviated injury scale\n\nASMD\n\n:   Absolute standardized mean difference\n\nCCCPR\n\n:   Closed-chest cardiopulmonary resuscitation\n\nCI\n\n:   Confidence interval\n\nCPR\n\n:   Cardiopulmonary resuscitation\n\nED\n\n:   Emergency department\n\nGCS\n\n:   Glasgow Coma Scale\n\nICD-9-CM\n\n:   International Classification of Diseases 9th Revision Clinical Modification\n\nISS\n\n:   Injury Severity Score\n\nOCCPR\n\n:   Open-chest cardiopulmonary resuscitation\n\nROSC\n\n:   Recovery of spontaneous circulation\n\nRT\n\n:   Resuscitative thoracotomy\n\nSOL\n\n:   Signs of life\n\nTQIP\n\n:   Trauma Quality Improvement Program\n\nVIF\n\n:   Variance inflation factor\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nSupplementary information\n=========================\n\n**Supplementary information** accompanies this paper at 10.1186/s13054-020-03259-w.\n\nThe authors thank Xiaofei Zhang for his invaluable efforts, including data acquisition, to this study.\n\nA.E., M.K., R.C., and Y. O contributed to the project development and study design. M.K., Z.J.H., and R.C. contributed to the data collection and interpretation of the data. A.E. performed the data analysis. A.E. and R.C. wrote the manuscript. M.K., Z.J.H., R.C., and Y.O. critically edited and revised the manuscript. The authors read and approved the final manuscript.\n\nThe authors have no conflicts of interest and no source of funding to declare.\n\nAll data used in the present study are available from the American College of Surgeons Trauma Quality Improvement Program.\n\nThe study and its protocols were approved by the institutional review committee of Riverside University Health System---Comparative Effectiveness and Clinical Outcomes Research Center. The requirement for informed consent for each patient was waived based on the use of anonymized patient and hospital data.\n\nNot applicable\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "INTRODUCTION\n============\n\nThe ANGPTLs (angiopoietin-like proteins) are a family of secreted proteins which share tertiary structural domains with angiopoietins \\[[@B1]\\], with N-terminal CCD (coiled coil domain) and C-terminal FLD (fibrinogen-like domain) of which the CCD is functionally associated with reduced plasma TAGs (triacylglycerols) \\[[@B2]\\]. *ANGPTL3* is almost exclusively expressed in the liver with minor expression found in kidney \\[[@B3],[@B4]\\]. ANGPTL3 has been shown to inhibit LPL (lipoprotein lipase) *in\u00a0vitro* and *in\u00a0vivo* \\[[@B5]--[@B7]\\], and mice lacking ANGPTL3 protein have increased LPL activity and reduced levels of plasma TAG \\[[@B8],[@B9]\\].\n\nGWAS (genome wide association studies) in humans have shown that SNPs (single nucleotide polymorphisms) at loci near *ANGPTL3* are associated with plasma TAG levels \\[[@B10]--[@B12]\\]. Sequencing of the coding region of *ANGPTL3* revealed additional loss-of-function mutations associated with low levels of plasma TAG \\[[@B4]\\]. Furthermore, Musunuru et al. \\[[@B13]\\] have identified E128X and S17X as two novel nonsense mutations in *ANGPTL3.* Loss of function in *ANGPTL3* results in a condition referred to as FHBL2 (familial combined hypolipidaemia, OMIM \\#605019), featuring reduction in all major plasma lipoproteins VLDL (very-low-density lipoprotein), LDL (low-density lipoprotein) and HDL (high-density lipoprotein). In addition to FHBL2 phenotype also plasma NEFAs (non-esterified fatty acids), insulin, fasting plasma glucose and HOMA-IR (homoeostatic model assessment of insulin resistance) values are significantly lower in individuals homozygous for S17X mutation compared with carriers and non-carriers \\[[@B14],[@B15]\\] linking the hypolipidaemic effect of ANGPTL3-deficiency with insulin sensitivity.\n\nInsulin is the primary hormone controlling whole body glucose and lipid homoeostasis. Activation of the insulin-signalling pathway suppresses hepatic secretion of TAG-enriched VLDL1 particles without affecting the secretion of lipid-poor VLDL2 \\[[@B16]\\]. The number of TAG-enriched VLDL1 particles is elevated during insulin resistance, contributing to hypertriglyceridaemia \\[[@B17]\\]. Despite elevated plasma glucose and TAG levels liver persistently continues glucose production and VLDL secretion during IR \\[[@B18],[@B19]\\]. Insulin inhibits VLDL secretion through PI3K (phosphoinositide 3-kinase)-mediated abruption in the second-step bulky TAG-assembly of VLDL1 particles and apoB-100 availability \\[[@B20]\\]. Unlike in the muscle and adipose tissue, insulin does not directly increase glucose uptake in the liver. Postic et al. \\[[@B21]\\] demonstrated down-regulation of GLUT2 (glucose transporter 2) by insulin in a dose-dependent manner *in\u00a0vivo* and *in\u00a0vitro*. Hepatic GLUT2 expression is induced by high glucose concentration which, during feeding, prevent the inhibitory effect of insulin on GLUT2 expression \\[[@B21]\\].\n\nWe utilized wild-type IHHs (immortalized human hepatocytes) and IHH cells with silenced ANGPTL3 to study how the secretion levels of ANGPTL3 are regulated in hepatocytes and how the silencing of *ANGPTL3* affects VLDL secretion and hepatocyte-specific glucose uptake. The aim of the research is to seek further insight into the molecular mechanisms related to hypolipidaemia and insulin sensitivity reported in humans with ANGPTL3 deficiency \\[[@B14]\\].\n\nMATERIALS AND METHODS\n=====================\n\n*ANGPTL3*-silenced cell-line and cell culture conditions\n--------------------------------------------------------\n\nLiver IHHs immortalized by SV40 large T-Antigen (IHH, ATCC\u00ae PTA-5565\u2122) were transduced with MISSION\u2122 shRNA (small hairpin RNA) Lentiviral Vector particles (TRCN0000242782, Sigma Aldrich) targeting *ANGPTL3* (NM_014495.2) or with non-target shRNA (SHCOO2, Sigma Aldrich) \\[MOI (multiplicity of infection) 1\\]. Positive cells were selected against 5\u00a0\u03bcg/ml puromycin for 12\u00a0days. The transduction was repeated once after the first selection. Cells were cultured in Williams medium E (Gibco by Life Technologies, 22551-022) with added 10% (v/v) FBS and glutamine 0.2\u00a0mg/ml and incubated +37\u00a0\u00b0C. FBS was removed during experiments and total protein from cell lysates was used to normalize the data. Cells were washed with PBS (pH\u00a07.4) and lysed in RIPA buffer. Protein concentration was measured with Bradford protein assay (Bio-Rad). Following compounds were used: insulin (bovine, Sigma-Aldrich), Wortmannin (Sigma-Aldrich), Akt1/2 inhibitor (Sigma-Aldrich), rosiglitazone (Cayman Chemical) and GW9662 (Sigma-Aldrich).\n\nEnzymatic measurements of TAG and PLs (phospholipids)\n-----------------------------------------------------\n\nTAGs from media and cell lysates were measured with an enzymatic method (GPO-PAP 1488872 kit; Roche Diagnostics Gmbh). Phospholipids were measured from the cell culture media with an enzymatic method (FS 60080970; Diasys Diagnostics).\n\nANGPTL3 and apoB-100 ELISA assays\n---------------------------------\n\nConcentration of ANGPTL3 was measured utilizing ELISA assay developed by Robciuc et al. \\[[@B22]\\]. ApoB-100 was measured with Human apolipoprotein B ELISA kit (Mabtech) according to instructions provided by the manufacturer.\n\nLabelled oleic acid and TAG extraction\n--------------------------------------\n\nCells were incubated in the presence of 5.5\u00a0mM glucose and 0.5% (w/v) BSA-complexed with 0.375\u00a0mM oleic acid and 0.1\u00a0\u03bcCi \\[^14^C\\]-oleic acid (PerkinElmer) for 24\u00a0h. Cells were collected in 1\u00a0ml 2% NaCl--PBS buffer. Lipids were extracted by adding 2\u00a0ml of methanol and 1\u00a0ml chloroform and centrifuged (2500 rev/min for 10\u00a0min at room temperature). H~2~O (1\u00a0ml) and chloroform (1\u00a0ml) were added and again centrifuged (2500 rev/min, for 10\u00a0min at room temperature). The upper phase was removed and the lower phase dried under nitrogen and solubilized in 100\u00a0\u03bcl of chloroform and applied on a TLC-plate. The TLC-plates were run in a chamber containing (*N*-hexane, diethylether, acetic acid and H~2~O (65/15/1/0.25, v/v). Iodine vapour stained areas containing the TAGs were scraped out from the TLC-plate and radioactivity was measured by liquid scintillation counting (Wallac LS-Beta-Counter).\n\nMeasurement of glucose uptake\n-----------------------------\n\nCells were incubated in the presence of 5.5 or 20\u00a0mM glucose and 1\u00a0\u03bcCi of deoxy-[D]{.smallcaps}-glucose, 2-\\[1,2-^3^H (N)\\] (1 mCi/ml, lot 638084, Perkin Elmer). Radioactivity from cell lysates and media was measured by liquid scintillation counter (Wallac LS-Beta-Counter).\n\nQPCR (quantitative PCR) analysis\n--------------------------------\n\nRNA was extracted with Qiagen RNA purification kit and synthesized with SuperScript\u00ae VILO\u2122 cDNA Synthesis Kit (Life Technologies). Each QPCR contained primers (see [Table 1](#T1){ref-type=\"table\"}) at 300\u00a0nM concentration and 10\u00a0ng of cDNA. The total volume of each reaction was 10\u00a0\u03bcl containing SYBR\u00ae Green PCR Master Mix (Life Technologies). The QPCR program: 95\u00a0\u00b0C for 10\u00a0min followed by 45 cycles 95\u00a0\u00b0C 15\u00a0s, 60\u00a0\u00b0C 1\u00a0min.\n\n###### QPCR-primers\n\n  Name, gene name         Forward                   Reverse\n  ----------------------- ------------------------- --------------------------\n  *ACTIN*, *ACTB*         GATGTGGATCAGCAAGCAGGA     AGCATTTGCGGTGGACGAT\n  *ANGPTL3*               CCAGAACACCCAGAAGTAACT     TCTGTGGGTTCTTGAATACTAGTC\n  *ANGPTL4*               GCCTATAGCCTGCAGCTCAC      CAAGTGGAGAAGGGTACGGA\n  *ANGPTL8*, *C19ORF80*   AGGTCTTAAAGGCTCACG        TTCCATCCAGGCAGATTC\n  *DGAT2*                 CATCCTCATGTACATATTCTGC    TGGGAAAGTAGTCTCGAAAG\n  *HNF4\u03b1*, *HNF4A*        AGTACATCCCAGCTTTCTG       AATGTAGTCATTGCCTAGGAG\n  *IL-6*                  GGTACATCCTCGACGGCATCT     GTGCCTCTTTGCTGCTTTCAC\n  *IRS-1*                 CAGAATGAAGACCTAAATGACC    ATGCATCGTACCATCTACTG\n  *IRS-2*                 AAGAGTGAAGATCTGTCTGG      ATCTAACAGAGTCCACAGATG\n  *PEPCK*, *PCK1*         GGGCATCCTCAGGCGGCT        GATAACCGTCTTGCTTTCGAT\n  *PGC1A*, *PPARGC1A*     GCAGACCTAGATTCAAACTC      CATCCCTCTGTCATCCTC\n  *PI3K 85*, *PIK3R1*     AAGAAGACTTGAAGAAGCA       TCAACCACATCAAGTATTGG\n  *PPARALPHA*, *PPARA*    CCTAAAAAGCCTAAGGAAACC     GATCTCCACAGCAAATGATAG\n  *PPAR*\u03b3, *PPARG*        GATCCAGTGGTTGCAGATTACAA   GAGGGAGTTGGAAGGCTCTTC\n  *PTEN*                  GGCTAAGTGAAGATGACAATC     GTTACTCCCTTTTTGTCTCTG\n  *TGH*, *TGH/CES1*       GTCTTTCTGGGCATTCCATT      CTCTCCACGTCTTGTAGGCA\n  *TNF\u03b1*, *TNFA*          CCCAGGCAGTCAGATCATCTT     AGCTGCCCCTGAGCTTGA\n  *TRB3*, *TRIB3*         GACCGTGAGAGGAAGAAG        GAGTATCTCAGGTCCCAC\n\nStatistics\n----------\n\nThe experiments were conducted in triplicate (*n*=3) and performed using 2--3 separately transduced IHH cell cultures. Results are expressed as means\u00b1S.D. One-way or two-way ANOVA was used to compare the significance between groups followed by Games--Howell *post hoc* test. The difference between groups was considered statistically significant if *P*\\<0.05 using Games--Howell *post hoc* test (\\*\\<0.05, \\*\\*\\<0.01, \\*\\*\\*\\<0.001).\n\nRESULTS\n=======\n\nWe first tested insulin response in wild-type IHH cells by incubating the cells in the presence of insulin and measured the secretion rate of VLDL by analysing the cell culture media for TAG, apoB-100 and PLs after 24\u00a0h. Insulin treatment resulted in a dose dependent down-regulation of the secreted TAG, PL ([Figure 1](#F1){ref-type=\"fig\"}a) and apoB-100 ([Figure 1](#F1){ref-type=\"fig\"}b). Insulin in a dose-dependent manner reduced also the secretion of ANGPTL3 ([Figure 1](#F1){ref-type=\"fig\"}c). We calculated a TAG/apoB-100 ratio and demonstrated that insulin stimulation results in a significant reduction of hepatocyte TAG secreted per apoB-100 protein ([Figure 1](#F1){ref-type=\"fig\"}d). These data are in line with reports on insulin mediated down-regulation of TAG-enriched VLDL1 particles without affecting the (lipid-poor) VLDL2 secretion \\[[@B16],[@B23]\\].\n\n![The effect of insulin on VLDL and ANGPTL3 secretion\\\nWild-type IHH-cells were incubated in 5.5\u00a0mM glucose for 24\u00a0h at variable concentrations of insulin. (**a**) Secretion of TAG and PL during insulin stimulation (\u03bcg/mg). (**b**) Secretion of apoB-100 (ng/mg). (**c**) Secretion of ANGPTL3 (ng/mg). (**d**) The amount of TAG (ng) for each ng of apoB-100.](bsr2014-0115i001){#F1}\n\nInsulin mediates its functions via IRS (insulin receptor substrate)--PI3K-dependent pathway \\[[@B24]\\]. To test whether VLDL secretion in IHH cells is regulated via PI3K and its downstream target Akt, we incubated cells with a PI3K inhibitor Wortmannin and Akt1/2 inhibitor. Treatment with Wortmannin partly abolished the inhibitory effect of insulin on TAG and PL secretion ([Figure 2](#F2){ref-type=\"fig\"}a). Akt1/2 inhibitor, much like Wortmannin, partially abolished the inhibitory effect of insulin on TAG secretion indicating that insulin regulates the secretion of VLDL via PI3K--Akt pathway in IHH cells ([Figure 2](#F2){ref-type=\"fig\"}b). Wortmannin or Akt1/2 did not abolish the inhibitory effect of insulin on ANGPTL3 secretion suggesting that secretion of ANGPTL3 is not down-regulated via PI3K or Akt ([Figures 2](#F2){ref-type=\"fig\"}c and [2](#F2){ref-type=\"fig\"}d).\n\n![The effect of PI3K Wortmannin, Akt-specific inhibitor Akt1/2 and PPAR\u03b3 antagonist GW9662 on VLDL and ANGPTL3 secretion\\\nWild-type IHH-cells were incubated in 5.5\u00a0mM glucose and 100\u00a0nM insulin (\u2212/+), Wortmannin (0.5\u00a0\u03bcM,10\u00a0\u03bcM), Akt1/2 inhibitor (20\u00a0\u03bcM) or PPAR\u03b3 antagonist GW9662 (10\u00a0\u03bcM) for 24\u00a0h. (**a**) Secretion of TAG and PL (\u03bcg/mg). (**b**) Secretion of TAG (\u03bcg/mg). (**c**) Secretion of ANGPTL3 (ng/mg). (**d**) Secretion of ANGPTL3 (ng/mg).](bsr2014-0115i002){#F2}\n\nSince ANGPTL3 protein secretion is regulated by insulin, we investigated whether nuclear PPARs (peroxisome proliferator-activated receptors) \u03b1 and \u03b3 and their agonists fenofibrate, used to treat high plasma TAG and cholesterol levels, and rosiglitazone, an insulin sensitizer, would affect the secretion of ANGPTL3. We demonstrate that PPAR\u03b3 agonist rosiglitazone, in a dose-dependent manner, down-regulated mRNA expression ([Figure 3](#F3){ref-type=\"fig\"}a) and the secretion of ANGPTL3 ([Figure 3](#F3){ref-type=\"fig\"}b). PPAR\u03b1 agonist fenofibrate did not affect the secretion of ANGPTL3 (results not shown). Rosiglitazone, much like insulin, reduced the secretion of TAG in a dose-dependent manner ([Figure 3](#F3){ref-type=\"fig\"}c). PPAR\u03b3 antagonist (GW9662) abolished the inhibitory effect of insulin on TAG secretion ([Figure 2](#F2){ref-type=\"fig\"}b). These data demonstrate that rosiglitazone mimics insulin action in cultured human hepatocytes by decreasing the secretion of TAG and ANGPTL3 in a dose-dependent manner. These data also support the hypothesis that insulin may down-regulate VLDL secretion via PPAR\u03b3. However, PPAR\u03b3-antagonist (GW9662) did not abolish the inhibitory effect of insulin on ANGPTL3 secretion ([Figure 2](#F2){ref-type=\"fig\"}d) indicating that insulin and PPAR\u03b3 down-regulate the expression of ANGPTL3 via different pathways.\n\n![The effect of rosiglitazone on VLDL and ANGPTL3 secretion\\\nWild-type IHH-cells were incubated in 5.5\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212) for 24\u00a0h following RNA extraction from cell lysates. Secretion of ANGPTL3 PL and TAG was measured in the presence of rosiglitazone. (**a**) Expression of *ANGPTL3* (AU, arbitary unit). (**b**) Secretion of ANGPTL3 (ng/mg). (**c**) Secretion of TAG and PL (\u03bcg/mg).](bsr2014-0115i003){#F3}\n\nTo investigate ANGPTL3-specific functions in hepatocytes, we silenced *ANGPTL3*-gene by transducing wild-type IHH cells with a lentiviral vector carrying *ANGPTL3*-targeting shRNA-fragment. Control cells were transduced with non-targeting shRNA. We achieved 85% silencing in *ANGPTL3* expression ([Figure 4](#F4){ref-type=\"fig\"}a) and a 90--95% decrease in ANGPTL3 protein secretion ([Figure 4](#F4){ref-type=\"fig\"}b). We then measured the concentration of TAG, PL and apoB-100 from the cell culture media after 24\u00a0h incubation. Surprisingly the secretion of apoB-100, PL and TAG were all increased in *ANGPTL3*-silenced cells compared with control cells under non-insulin conditions (\u2212) ([Figures 5](#F5){ref-type=\"fig\"}a--[5](#F5){ref-type=\"fig\"}c). Also a short-term secretion rate of apoB-100 was higher in *ANGPTL3*-silenced cells under non-insulin conditions ([Figure 5](#F5){ref-type=\"fig\"}d). Insulin stimulation (+) reduced the secretion of TAG and PL in silenced cells ([Figures 5](#F5){ref-type=\"fig\"}a--[5](#F5){ref-type=\"fig\"}b) near to the level of TAG secreted by the control cells with a smaller decrease in secreted apoB-100 levels ([Figure 5](#F5){ref-type=\"fig\"}c). We calculated a TAG/apoB-100 ratio for both control and silenced cells and concluded that the ratio was more dramatically reduced in *ANGPTL3*-silenced cells \\[47%, from 60 to 32 (mass/mass)\\] compared with control cells \\[16%, from 45 to 38 (mass/mass)\\] during insulin stimulation (+) ([Figure 5](#F5){ref-type=\"fig\"}e). These findings suggest that the silencing of *ANGPTL3* results in a more pronounced shift from the secretion of large TAG-enriched VLDL1-type particles into small lipid-poor VLDL2-type particles during insulin stimulation. These data are in line with plasma measures from ANGPTL3-deficient subjects who had a 4.3-fold reduction in the TAG/apoB-100 ratio in VLDL when compared with non-carriers \\[[@B14]\\] and suggests that *ANGPTL3*-silenced cells display enhanced insulin sensitivity as VLDL1 particles dominate the VLDL profile in insulin-resistant states.\n\n![Silencing of *ANGPTL3*\\\nIHH-cells were transduced with *ANGPTL3*-targeting or non-targeting shRNA (control). Cells were incubated in 5.5\u00a0mM glucose with 100\u00a0nM insulin (+/\u2212) for 24\u00a0h. RNA was extracted from cell lysates. (**a**) Expression of *ANGPTL3* (AU, arbitrary unit). (**b**) Secretion of ANGPTL3 (ng/mg).](bsr2014-0115i004){#F4}\n\n![Secretion of VLDL from *ANGPTL3*-silenced and control IHH cells\\\n*ANGPTL3*-silenced and control cells were incubated in media containing 5.5\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212). (**a**--**c**) Secretion of TAG, PL (\u03bcg/mg) and apoB-100 (ng/mg). (**d**) Secretion of short-term apoB-100 (ng/mg). (**e**) The amount of TAG (ng) for each ng of apoB-100.](bsr2014-0115i005){#F5}\n\nWe then measured glucose uptake in control and silenced cells by utilizing a non-hydrolysable radiotracer, \\[^3^H\\]-labelled deoxy-[D]{.smallcaps}-glucose. Basal glucose uptake without insulin stimulus was increased by 50% in *ANGPTL3* silenced cells compared with control cells after 2\u00a0h ([Figure 6](#F6){ref-type=\"fig\"}a) and after 24\u00a0h ([Figure 6](#F6){ref-type=\"fig\"}b). Treatment with rosiglitazone improved hepatocyte glucose uptake by 40% in control cells ([Figure 6](#F6){ref-type=\"fig\"}a). Acute insulin stimulus in the presence of 5.5\u00a0mM glucose did not increase hepatocyte glucose uptake in control cells and significantly decreased glucose uptake in *ANGPTL3*-silenced cells after 2\u00a0h ([Figure 6](#F6){ref-type=\"fig\"}a) and 24\u00a0h ([Figure 6](#F6){ref-type=\"fig\"}b). When the glucose concentration was increased up to 20\u00a0mM the inhibitory effect of insulin on glucose uptake was abolished ([Figure 6](#F6){ref-type=\"fig\"}c). We conclude that the silencing of *ANGPTL3* results in increased glucose uptake comparable with rosiglitazone treatment in the absence of insulin stimulus and that insulin inhibits hepatocyte-specific glucose uptake in the presence of low glucose concentrations but not under high glucose concentration.\n\n![Hepatic glucose uptake in *ANGPTL3*-silenced and control cells\\\n*ANGPTL3*-silenced and control cells were incubated in the presence of \\[^3^H\\] labelled deoxy-[D]{.smallcaps}-glucose. Radioactivity (CPM) was measured from the cell lysates. (**a**) Hepatic glucose uptake measured after 2\u00a0h incubation period in media containing 5.5\u00a0mM glucose (CPM/mg). Cells were incubated with 50\u00a0\u03bcM rosiglitazone or 100\u00a0nM insulin (+/\u2212). (**b**) Hepatic glucose uptake measured after 24\u00a0h incubation period in media containing 5.5\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212) (CPM/mg). (**c**) Hepatic glucose uptake measured after 2\u00a0h incubation period in media containing 20\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212) (CPM/mg).](bsr2014-0115i006){#F6}\n\nWe reasoned that the increased glucose uptake in *ANGPTL3*-silenced hepatocytes would cause elevation in glycogen and intracellular TAG levels and measured the intracellular TAG levels after 24\u00a0h incubation period. We observed elevated intracellular TAG levels in *ANGPTL3*-silenced cells as compared with control cells ([Figure 7](#F7){ref-type=\"fig\"}a). To test whether silencing of *ANGPTL3* affected hepatic fatty acid uptake we incubated control and *ANGPTL3*-silenced cells in media containing \\[^14^C\\]-labelled oleic acid bound to 0.5% BSA for 12\u00a0h and measured the uptake of fatty acids and its conversion into intracellular TAG. There were no differences in fatty acid incorporation in intracellular TAG and the secretion rates of TAG between control and *ANGPTL3*-silenced cells ([Figure 7](#F7){ref-type=\"fig\"}b).\n\n![Intracellular TAG and fatty acid uptake in *ANGPTL3*-silenced and control cells\\\n(**a**) Intracellular TAGs were measured in *ANGPTL3*-silenced and control cell lysates after 24\u00a0h incubation with 5.5\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212). (**b**) *ANGPTL3*-silenced and control cells were incubated in media containing 5.5\u00a0mM glucose, 0.375\u00a0mM oleic acid complexed with 0.5% BSA and C^14\u2212^ labelled oleic acid for 12\u00a0h. Radioactivity (DPM) in intracellular TAG and secreted TAG was measured (DPM/mg).](bsr2014-0115i007){#F7}\n\nFinally we investigated whether silencing of *ANGPTL3* changes the expression of insulin responsive genes. QPCR demonstrated that there were no significant changes in gene expression levels of *ANGPTL8* or *ANGPTL4* between control and silenced cells ([Figure 8](#F8){ref-type=\"fig\"}a). Insulin receptor *IRS-2*, *GLUT2* and *PI3K* (regulatory subunit p85) show a moderate (but non-significant) up-regulation in *ANGPTL3*-silenced cells in non-insulin conditions without any differences in *IRS-1* or PI3K inhibitor PTEN (phosphatase and tensin homologue deleted on chromosome 10) ([Figure 8](#F8){ref-type=\"fig\"}b). Expression of *PGC1\u03b1* (peroxisome proliferator-activated receptor \u03b3 co-activator 1-\u03b1) and its downstream targets *PEPCK* (phosphoenolpyruvate carboxy-kinase) and *TRB-3* (tribbles homologue 3) were significantly down-regulated in *ANGPTL3*-silenced cells, whereas expression of *PPAR\u03b1* was significantly up-regulated in *ANGPTL3*-silenced cells ([Figure 8](#F8){ref-type=\"fig\"}c). The mRNA levels of PPAR\u03b3 displayed a moderate down-regulation in *ANGPTL3*-silenced cells upon insulin stimulation ([Figure 8](#F8){ref-type=\"fig\"}c). The expression of TGH (TAG hydrolase) and DGAT2 (diacylgylcerol acyltransferase 2) involved in VLDL assembly were up-regulated in *ANGPTL3*-silenced cells during insulin stimulus while the expression of inflammation marker TNF\u03b1 (tumour necrosis factor \u03b1) was significantly down-regulated in the silenced cells ([Figure 8](#F8){ref-type=\"fig\"}d). There was no significant difference in IL-6 (interleukin-6) or HNF4\u03b1 (hepatocyte nuclear factor \u03b1) expression ([Figure 8](#F8){ref-type=\"fig\"}d). These data suggest a down-regulation of gluconeogenic genes and an increase in genes promoting TAG synthesis in *ANGPTL3*-silenced cells as compared with control.\n\n![Gene expression patterns in *ANGPTL3*-silenced and control IHH cells during non-insulin condition (Ins-) and insulin stimulation (Ins+)\\\n(**a--d**) *ANGPTL3*-silenced cells and control cells were incubated with 5.5\u00a0mM glucose and 100\u00a0nM insulin (+/\u2212) for 24\u00a0h followed by total RNA extraction from cell lysates, cDNA synthesis and q-PCR (AU, arbitrary unit).](bsr2014-0115i008){#F8}\n\nDISCUSSION\n==========\n\nSince ANGPTL3 is primarily expressed in the liver \\[[@B3],[@B4]\\] and affects TAG clearance \\[[@B5]--[@B9]\\] and possibly fatty acid release from adipose tissue \\[[@B25]\\] it is unclear how extracellular signals affect ANGPTL3 secretion rates and how silencing of *ANGPTL3* would affect glucose and lipid metabolism in the liver. We show that activation of PPAR\u03b3 by its agonist rosiglitazone, much like insulin, decreased the expression and secretion of ANGPTL3 accompanied by reduced secretion of TAG. The association of rosiglitazone and reduced VLDL--TAG secretion has been established *in\u00a0vivo* \\[[@B26],[@B27]\\]. Down-regulation of *ANGPTL3* gene expression upon rosiglitazone treatment can also be found in microarray data of primary human hepatocytes \\[[@B28]\\] and is verified by our data including decreased dose-dependent protein secretion. The effect of insulin on *ANGPTL3* expression and secretion was shown in HepG2 cells \\[[@B29]\\] and also verified by our data. These findings indicate that *ANGPTL3* is a target gene for insulin and PPAR\u03b3 action and demonstrate that stimulation of PPAR\u03b3 by rosiglitazone, mediates or mimics insulin action with a concomitant reduction in ANGPTL3 and VLDL secretion.\n\nInsulin stimulation reduces the secretion of TAG-enriched VLDL1-type particles in the liver without affecting the basal secretion of the lipid-poor VLDL2-type particles \\[[@B16]\\]. This balance is disrupted in IR leading to increased secretion of VLDL1 and increased plasma lipid levels \\[[@B16],[@B17]\\]. Interestingly a loss of function mutation in the *ANGPTL3-*gene results in decreased levels of all lipoprotein types\u00a0in plasma \\[[@B14]\\] but whether the hypolipidaemic phenotype is caused by decreased hepatic VLDL secretion or due to enhanced TAG clearance in the peripheral tissue (or both) has not been verified clearly. The work by Musunuru et al. \\[[@B13]\\] suggested lower apoB-100 production rate in homozygous human carriers of ANGPTL3 loss of function mutations. Shimizugawa et al. \\[[@B9]\\] and Ando et al. \\[[@B30]\\] suggest that the low plasma lipid levels (especially TAGs) observed in ANGPTL3-deficient mice were likely caused by increased VLDL--TAG clearance via induced LPL activity rather than decreased VLDL secretion.\n\nWe did not detect decreased apoB-secretion in *ANGPTL3*-silenced cells compared with control; however, silencing of *ANGPTL3* caused a major decrease in secreted TAG during insulin stimulation with a more subtle effect on apoB-100 secretion. This observation is in line with the analysis on VLDL particles isolated from plasma of *ANGPTL3*-deficient subjects showing a dramatic reduction in VLDL particle size and a 4.3-fold reduction in the TAG/apoB-100 ratio \\[[@B14]\\]. We demonstrate that a shift from the secretion of TAG-enriched VLDL1-type particles into the secretion of TAG poor VLDL2-type particles under insulin stimulus was more pronounced in ANGPTL3 silenced cells compared with control cells suggesting elevated insulin sensitivity. Higher insulin sensitivity in homozygous *ANGPTL3* loss-of-function carriers was demonstrated previously by Robciuc et al. \\[[@B14]\\] thus supporting our cell data. Additionally Inukai et al. \\[[@B31]\\] showed elevated ANGPTL3 expression in mice in both insulin-deficient and -resistant diabetic states. Therefore down-regulation of *ANGPTL3* might have a protective role against IR.\n\nLiver uptake accounts for about 30% disposal of plasma glucose \\[[@B32]\\]. Silencing of *ANGPTL3* resulted in 20--50% increase in glucose uptake compared with control cells and probably explaining glucose conversion into intracellular TAGs and their increased secretion in *ANGPTL3*-silenced cells under insulin-depleted conditions. Human data do not support an increased prevalence of hepatic steatosis in *ANGPTL3*-deficient subjects \\[[@B15],[@B33]\\] suggesting that fluctuations in the availability of substrates, plasma glucose and insulin concentrations *in\u00a0vivo* as well as uptake of glucose and fatty acids by other tissues regulate hepatic glucose uptake and VLDL secretion and thereby preventing hepatic TAG accumulation in ANGPTL3-deficient individuals.\n\nOur data on VLDL secretion and glucose uptake suggests that both insulin and PPAR\u03b3 might mediate some of their functions via affecting the expression of *ANGPTL3*. Activation of PPAR\u03b3 decreases the transcription of insulin-responsive genes involved in the control of glucose production (gluconeogenesis) \\[[@B34]\\] and PPAR\u03b3 can directly activate GLUT2 transporter thus influencing hepatic glucose uptake \\[[@B35]\\] which can explain the increased short-term glucose uptake levels upon rosiglitazone treatment in the present study. We show that silencing of *ANGPTL3* resulted in significant down-regulation of transcription factor *PGC1\u03b1* expression and its downstream targets *PEPCK* and *TRB3* (tribbles homologue 3), a protein kinase acting as a negative inhibitor for Akt phosphorylation \\[[@B36],[@B37]\\]. Elevation of *PGC1\u03b1* expression in liver has been linked to IR in several mouse and human studies \\[[@B37]--[@B39]\\] thus the down-regulation of *PGC1\u03b1* and its downstream targets as observed in *ANGPTL3*-silenced cells suggest decreased levels of gluconeogenesis and fatty acid oxidation, processes which display aberrant activation in insulin resistant states (See [Figure 9](#F9){ref-type=\"fig\"} as a summary for the pathways).\n\n![Hypothetic model for ANGPTL3 function in a hepatocyte\\\nSilencing of ANGPTL3 enhances GLUT2-mediated glucose uptake, increases TAG synthesis and induces down-regulation of PGC1\u03b1 and its downstream targets, including TRB3, hypothetically leading to enhanced Akt phosphorylation levels. Secretion of TAG-enriched VLDL-1-type particles is decreased upon insulin stimulus.](bsr2014-0115i009){#F9}\n\nInsulin induces glucose uptake in insulin-responsive tissues (muscle, white adipose tissue) via GLUT4 but is not required for GLUT2-mediated glucose uptake by the liver \\[[@B39]\\]. In fact, insulin dose dependently decreases the expression of hepatocyte *GLUT2* during the first 6--12\u00a0h of insulin stimulation *in\u00a0vitro* and *in\u00a0vivo*. The inhibitory effect of insulin on *GLUT2* expression is abolished after 24\u00a0h despite continuous insulin stimulation and can also be reversed by high (10--20\u00a0mM) glucose concentration \\[[@B21]\\]. Therefore hepatic glucose uptake is driven by extracellular glucose concentration and not directly due to the elevated insulin levels (in other words GLUT2 is not insulin dependent). Increased glucose uptake in *ANGPTL3*-silenced cells might be due to higher production rate of GLUT2 and/or increased plasma membrane translocation and activation of GLUT2. In conclusion, our results give insights into the mechanisms of *ANGPTL3*-silencing in hepatic lipid and glucose metabolism. Since humans with ANGPTL3-deficiency display hypolipidaemia and insulin sensitivity it might be beneficial to target *ANGPTL3 i*n the liver, the major site of *ANGPTL3* expression, to balance lipid and glucose homoeostasis and lower the risk of cardiometabolic disorders.\n\nWe thank Sari Nuutinen from the National Institute for Health and Welfare (Helsinki, Finland) for her technical assistance and Marius Robciuc from the National Institute for Health and Welfare for his previous work with lentiviral shRNA particles and ANGPTL3-protein.\n\nAUTHOR CONTRIBUTION\n===================\n\nMatti Jauhiainen and Christian Ehnholm supervised the study. Jarkko Soronen participated in study design and scientific discussion of the data. Pirkka-Pekka Laurila contributed to the scientific discussion of the data. Jari Metso contributed to the biochemical analysis of the experiments.\n\nFUNDING\n=======\n\nThe study was financially supported by the Academy of Finland \\[grant number 257545\\], Finnish Foundation for Cardiovascular Research, Jenny and Antti Wihuri Foundation, Finska L\u00e4kares\u00e4llskapet, Magnus Ehrnrooth foundation and Novo Nordisk Fonden.\n\n[^1]: ^2^These authors contributed equally to this work.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nUntil recently, the Burch colposuspension was the standard procedure for the treatment of female stress urinary incontinence (SUI) \\[[@CR1], [@CR2]\\]. Since its introduction in 1995, tension-free vaginal tape (TVT) became the new \"gold\" standard in the surgical treatment of SUI. TVT has been proven to be as effective as the Burch colposuspension with success rates ranging from 84% to 95%, but has less side effects and complications \\[[@CR3]-[@CR5]\\]. In 2001, Delorme introduced transobturator tape (TOT) with a high success rate \\[[@CR6]\\]. This sling is also located underneath the mid-urethra, but runs laterally through the obturator membrane to the upper part of the thigh.\n\nThere are two basic techniques for performing TOT: \"outside-in\" described by Delorme \\[[@CR6]\\] and \"inside-out\" (TVT-O) described by de Leval \\[[@CR7]\\]. These techniques differ not only in the way the needle is placed and carried forward but also in the different designs of the introducer. After the introduction of the outside-in technique by Delorme in 2001, several other similar procedures were developed. Among them is Monarc, which consists like TVT-O of a monofilament macroporous polypropylene mesh and which has shown similar and equally good short- to medium-term success rates as TVT-O \\[[@CR8], [@CR9]\\].\n\nHowever, until now, there is little evidence that one technique is superior to the other. Furthermore, no comparative trials are published with follow-up longer than 1\u00a0year. Therefore, the aim of this study is to compare the outcome and the quality of life results of two existing different TOT procedures inside-out (TVT-O) and outside-in (Monarc) for the treatment of SUI after a 2--4-year follow-up.\n\nMaterials and methods {#Sec2}\n=====================\n\nBetween October 2004 and December 2006, 191 consecutive women with an indication for surgical treatment of SUI underwent either the TVT-O or the Monarc procedure. This study was approved by the Medical Ethical Committee of the St. Elisabeth Hospital as part of a multicentre study on the outcome of transobturator tape as treatment for predominant female SUI.\n\nIncluded were 93 women with a TVT-O and 98 women with a Monarc procedure. Women who did not show up for their 2--4-year follow-up visit first received a questionnaire and, if they did not reply, were later contacted by one of the investigators (MH).\n\nExcluded were women with recurrent and difficult to treat urinary tract infections, significant and predominantly symptoms of urge urinary incontinence, post voiding bladder retention of more than 150\u00a0ml, bladder capacity of less than 100\u00a0ml, and physical or mental impairment to adequately register outcome results or complications.\n\nA standardized urogynecological history and physical examination was performed preoperative, after 2\u00a0months, 1\u00a0year, and after 2 to 4\u00a0years. Preoperative multichannel urodynamic investigation was carried in all women according to the International Continence Society standards \\[[@CR10], [@CR11]\\].\n\nAll women were asked to complete the short version of the Incontinence Impact Questionnaire (IIQ-7) and the Urogenital Distress Inventory (UDI-6) before and at the above-mentioned postoperative intervals. These questionnaires are internationally accepted as valid disease-specific quality of life (QoL) instruments and are sensitive to change \\[[@CR12]\\]. They have been translated into Dutch language and validated \\[[@CR13]\\]. The questionnaires were anonymously entered in a database, and researchers were blinded to the individual results of these questionnaires. The total score of the IIQ-7 and UDI-6 was transformed to a scale from 0 to 100 (a higher score indicates more bother).\n\nWomen were assigned to either a TVT-O or a Monarc procedure just before surgery was to take place. The inside-out procedure was performed as described by de Leval \\[[@CR7]\\], and a TVT-O (Gynecare, Sommerville, NJ, USA) tape was inserted. The Monarc (American Medical Systems, Minneapolis, MN, USA) tape was inserted through the outside-in route as using the technique recommended by the manufacturer. Both procedures use a monofilament, macroporous polypropylene tape, which is placed under the urethra without any tension. Surgery was not accompanied by other urogynecological procedures.\n\nAll procedures were carried out under general anesthesia, and hence, no intra-operative cough stress test was performed. Cystoscopy was performed only in case of encountering bloody urine. All women received preoperative antibiotics, and this was continued for 7\u00a0days postoperative. All procedures were performed in the St. Elisabeth Hospital Tilburg, The Netherlands.\n\nWe adhered to the recommendations of IUGA on outcome measures \\[[@CR14]\\]. Cure of SUI was defined as the statement of the woman of not experiencing any loss of urine upon physical exercise. Other definitions in this study are in accordance with the terminology of the International Continence Society \\[[@CR10]\\].\n\nContinuous variables were compared using unpaired *t* test, and dichotomous variables were compared using chi-square test. Fisher's exact test was used if cross tabs had a cell with an expected frequency \\<5. The minimum level of significance was 0.05. Statistical analyses were performed with SPSS (Windows version 16.0).\n\nResults {#Sec3}\n=======\n\nAfter a 2--4-year follow-up, 161 (84%) women were available for evaluation. The TVT-O group consisted of 75 women compared to 86 women in the Monarc group. The mean follow-up of the TVT-O group was 38\u00a0months (range 12--53\u00a0months) and 39\u00a0months (range 12--53\u00a0months) in the Monarc group. For their 2--4-year follow-up control, 76 (47%) women visited our outpatient department, 51 (32%) women answered a questionnaire by mail, and 34 (21%) women answered a telephonic questionnaire by one of the investigators (MH). The response rate of the QoL questionnaires was 92%, 81%, 81%, and 70% respectively prior to, 2\u00a0months and 1\u00a0year postoperatively, and after 2--4\u00a0years. There were no differences in method of follow-up or response rate of QoL questionnaires between both groups.\n\nPreoperative data are listed in Table\u00a0[1](#Tab1){ref-type=\"table\"}. No differences between both groups were observed. In women with mixed urinary incontinence (MUI), anti-cholinergic medical treatment was given before surgical treatment was considered and did not alleviate their symptoms. The majority (89%) of the women underwent preoperative pelvic floor physiotherapy. Table\u00a01Preoperative data\u00a0TVT-OMonarc*p* valueNumber of women (*n*, %)93 (49%)98 (51%)Medical history\u2009Age (years, mean \u00b1 SD)49.2\u2009\u00b1\u20098.949.5\u2009\u00b1\u200910.30.8^a^\u2009Type of incontinence (*n*, %)\u00a0SUI74 (80%)74 (76%)\u00a0MUI19 (20%)23 (24%)0.6^b^\u2009Parity (number of vaginal deliveries; mean, SD)2\u2009\u00b1\u200912\u2009\u00b1\u200910.8^a^\u00a0Multiparous (*n*, %)84 (90%)92 (94%)0.7^c^\u2009Menopausal status (*n*, %)\u00a0Premenopausal51 (55%)58 (59%)\u00a0Climacterial9 (10%)6 (6%)0.5^b^\u00a0Postmenopausal33 (35%)34 (35%)\u2009Severity of incontinence (daily episodes; *n*, %)79 (85%)79 (81%)0.6^b^\u2009Pelvic floor physiotherapy prior to surgery (*n*, %)\u00a0Performed82 (88%)88 (90%)0.5^c^\u2009Previous incontinence surgery8 (9%)9 (9%)1.0^b^\u2009Previous prolapse surgery19 (22%)15 (16%)0.3^b^Physical examination\u2009Urethral hypermobility (*n*, %)80 (86%)90 (92%)0.3^b^\u2009Urinary incontinence on coughing (*n*, %)62 (67%)67 (68%)1.0^b^\u2009Vaginal wall prolapse (grade 1 or more; *n*, %)25 (27%)24 (24%)0.7^b^Urodynamic investigation\u2009Maximum bladder capacity (ml, mean \u00b1 SD)443\u2009\u00b1\u2009121433\u2009\u00b1\u20091300.6^a^\u2009Number of women with urodynamic SUI (*n*, %)61 (66%)64 (65%)0.9^b^\u2009Number of women with detrusor overactivity (*n*, %)5 (6%)7 (7%)0.7^b^\u2009Maximum urethral closure pressure (cm H~2~O, mean\u2009\u00b1\u2009SD)59\u2009\u00b1\u20093164\u2009\u00b1\u2009430.5^a^\u2009Number of women with MUCP\u2009\\<\u200920 cm H~2~O (*n*, %)5 (5%)1 (1%)0.1^ca^Students *t* test (two-sided)^b^Chi-square test^c^Fisher's exact test\n\nSurgical data are depicted in Table\u00a0[2](#Tab2){ref-type=\"table\"}. No concomitant urogynecological surgery was performed, but in 15 women (8%), surgery was combined with non-urogynecological procedures. There are no differences in surgery time, amount of blood loss, and type of anesthesia between TVT-O and Monarc. Table\u00a02Surgery characteristics\u00a0TVT-OMonarc*p* valueLength of surgery (min; mean \u00b1 SD, range)16\u2009\u00b1\u20095 \\[10--34\\]16\u2009\u00b1\u20096 \\[10--30\\]0.9^a^Amount of blood loss (ml; mean \u00b1 SD, range)55\u2009\u00b1\u200953 \\[0--300\\]62\u2009\u00b1\u200942 \\[0--250\\]0.3^a^Type of anesthesia (*n*, %)General85 (91%)84 (86%)0.2^b^Spinal8 (9%)14 (14%)Simultaneous performed secondary surgical procedure\u2009Hysteroscopic endometrial ablation/myoma resection22Introital plasty2Sterilization42LLETZ1Cystoscopic resection of bladder endometiosis1Inguinal hernia repair1^a^Student *t* test (two-sided)^b^Chi-square test\n\nOutcome data are presented in Table\u00a0[3](#Tab3){ref-type=\"table\"}. The cure rate for TVT-O was 72% after 2--4\u00a0years compared to 65% in the Monarc group, while improvement was observed in respectively 12% and 21% (*p*\u2009=\u20090.3). Preexisting urge incontinence resolved in seven women (32%) after TVT-O and eight women (31%) after Monarc (*p*\u2009=\u20091.0). De novo urge incontinence developed in two cases in the TVT-O group versus four cases in the Monarc group (*p*\u2009=\u20090.2). Table\u00a03Outcome of TVT-O and Monarc procedures\u00a0TVT-O*p* value within TVT-O groupMonarc*p* value within Monarc group*p* value between groupsStress urinary incontinence (SUI)2\u00a0months after surgeryNo SUI76 (85%)78 (83%)Improved7 (9%)12 (13%)0.4^a^Unchanged6 (7%)4 (4%)12\u00a0months after surgeryNo SUI66 (77%)73 (77%)Improved13 (15%)16 (17%)0.9^b^Unchanged7 (8%)6 (6%)2--4\u00a0years after surgeryNo SUI54 (72%)56 (65%)Improved9 (12%)18 (21%)0.3^b^Unchanged12 (16%)12 (14%)Quality of life analysisIncontinence Impact Questionnaire (IIQ-7, mean \u00b1 SD)Prior to surgery56\u2009\u00b1\u20092151\u2009\u00b1\u2009220.1^c^2\u00a0months after surgery11\u2009\u00b1\u200920**0.0**^d^13\u2009\u00b1\u200921**0.0**^d^0.6^c^12\u00a0months after surgery13\u2009\u00b1\u200921**0.0**^d^11\u2009\u00b1\u200919**0.0**^d^0.7^c^2--4\u00a0years after surgery11\u2009\u00b1\u200928**0.0**^d^12\u2009\u00b1\u200920**0.0**^d^0.6^c^Urodynamic Distress Inventory (UDI-6, mean \u00b1 SD)Prior to surgery45\u2009\u00b1\u20091745\u2009\u00b1\u2009160.7^c^2\u00a0months after surgery17\u2009\u00b1\u200918**0.0**^d^16\u2009\u00b1\u200917**0.0**^d^0.7^c^12\u00a0months after surgery16\u2009\u00b1\u200920**0.0**^d^16\u2009\u00b1\u200917**0.0**^d^0.9^c^2--4\u00a0years after surgery17\u2009\u00b1\u200918**0.0**^d^17\u2009\u00b1\u200917**0.0**^d^1.0^c^Urge urinary incontinence (UUI)Prior to surgery present22 (24%)26 (26%)0.7^b^UUI resolved, 2\u00a0months after surgery12 (55%)14 (54%)0.7^b^UUI resolved, 12\u00a0months after surgery13 (59%)17 (65%)0.5^b^UUI resolved, 2--4\u00a0years after surgery7 (32%)8 (31%)1.0^b^Prior to surgery absentDe novo UUI, 2\u00a0months after surgery02 (4%)0.5^a^De novo UUI, 12\u00a0months after surgery3 (4%)6 (6%)0.5^a^De novo UUI, 2--4\u00a0years after surgery2 (3%)4 (5%)0.2^aa^Fisher's exact test^b^Chi-square test; statistically significant differences are in bold^c^Student *t* test (two-sided)^d^Wilcoxon signed ranks test, pre- compared to postoperative situation within one group\n\nBoth the IIQ-7 and UDI-6 demonstrated a statistically significant increase in QoL decrease in impairment caused by symptoms of SUI after 2\u00a0months, 1\u00a0year, and 2--4\u00a0years in both TOT groups.\n\nAny complications are presented in Table\u00a0[4](#Tab4){ref-type=\"table\"}. No major intra-operative complications occurred. After 2\u00a0months, voiding difficulty was more observed in the TVT-O group. Tape release was carried out in one woman in the TVT-O group because of vaginal pain and a superficial located and palpable tape. She remained continent. In one other woman, tape release was indicated for voiding difficulty. Afterward, voiding became normal and they remained continent. Table\u00a04Complications and second incontinence procedures\u00a0TVT-OMonarc*p* valueIntra-operative complications (*n*, %)Bleeding\u2009\\>\u2009100\u00a0ml^a^4 (4%)5 (5%)0.9^b^Fausse route (within vagina)1 (1%)1 (1%)0.6^b^Second TOT procedure^c^01 (1%)1.0^b^Postoperative complications 2\u00a0months (*n*, %)Voiding difficulty10 (12%)3 (3%)**0.03**^b^Vaginal erosion01 (1%)1.0^b^Postoperative complications 12\u00a0months (*n*, %)Voiding difficulty4 (5%)3 (4%)0.7^b^Vaginal erosion1 (1%)3 (3%)0.6^b^Tape release1 (1%)1 (1%)1.0^b^Dyspareunia1 (1%)1 (1%)1.0^b^Persisting pain in thigh02 (2%)0.5^b^Postoperative complications 2--4\u00a0years (*n*, %)Voiding difficulty3 (4%)5 (6%)0.7^b^Vaginal erosion^d^1 (1%)4 (5%)0.4^b^Dyspareunia1 (1%)01.0^b^Persisting pain in thigh01 (1%)1.0^b^Second incontinence procedures (*n*, %)\\<1\u00a0year after primary procedure3 (4%)3 (4%)\\>1\u00a0year after primary procedure2 (2%)1 (1%)Total5 (6%)4 (5%)0.5^ba^Range 150--300^b^Fisher's exact test^c^Second procedure performed within the same surgery session due to accidently premature cutting of the tape^d^Total erosions after a 2--4-year follow-up\n\nThere were five cases of vaginal erosion, which were all excised under local anesthesia, and all remained continent afterward. In five women after TVT-O and in four women after Monarc, a second anti-incontinence procedure was performed (*p*\u2009=\u20090.5). As second incontinence procedures, tension-free vaginal tape (seven times), tension-free vaginal tape Secur (once), and Remeex system (once) were performed. The stress incontinence was resolved in all nine women. In one other woman, a stitch was placed under the mid-urethra under local anesthesia after which her already improved SUI disappeared.\n\nDiscussion {#Sec4}\n==========\n\nThis report comprises a large comparative study of both procedures with the longest follow-up until now. The main objective of this study was to compare outcome and complications for an inside-out versus an outside-in transobturator approach. The second objective was to evaluate results of both obturator approaches after a 2--4-year follow-up. We did not find significant differences between both approaches. Approximately 85% of the women found their symptoms to be improved after 2--4\u00a0years, which are satisfying results.\n\nMost studies comparing TVT-O and Monarc have a short follow-up time and differ with respect to definitions of success. Debodinance compared 50 TVT-O procedures to 50 Monarc procedures in a prospective observational study \\[[@CR15]\\]. With a definition of cure as being continent at postoperative urodynamic assessment, 47 (94%) women with TVT-O and 45 (90%) with Monarc were cured after 1\u00a0year. Lee et al. prospectively compared 100 women who underwent either a TVT-O (*n*\u2009=\u200950) or a TOT (Dow Medics, Korea, *n*\u2009=\u200950) procedure and defined cure as a negative cough stress test and no reports of urine leakage during stress \\[[@CR16]\\]. Equal cure rates for TVT-O and TOT were found: 86% versus 92% 1\u00a0year after surgery. Liapsis et al. followed up 114 women who were prospectively randomized to TVT-O or Monarc for 1\u00a0year \\[[@CR17]\\]. Cure was defined as a negative cough stress test during multichannel urodynamic investigation and a 1-h pad test giving a weight of less than 1\u00a0g. Cure rate in the TVT-O group was 87% (53 of 61 women) and 90% (48 of 53 women) in the Monarc group. Subjective cure was remarkably lower: 80% in the TVT-O group versus 77% in the Monarc group.\n\nNot many studies have investigated the medium-term results (\\>1\u00a0year) of obturator tapes. Waltregny et al. followed up 91 women prospectively after a TVT-O procedure with a minimum follow-up of 3\u00a0years \\[[@CR18]\\]. Cure of SUI was defined by the disappearance of subjective SUI, as assessed by a SUI symptom scale score equal to 0. Disappearance and improvement of SUI was observed in 88% and 9% of the women, respectively. Wang et al. found an 83% cure rate 36\u00a0months after a TVT-O procedure. The clinical outcome was regarded as cured when the cough test was negative at the follow-up visit. Their results were based on 30 women with a 36-month follow-up.\n\nThe above-mentioned comparative studies show as presented in this study that both TVT-O and Monarc are equally effective in curing SUI after 1\u00a0year \\[[@CR15]-[@CR17]\\]. On longer follow-up, our results are lower compared to the mentioned studies \\[[@CR18], [@CR19]\\]. This might be due to our definition of cure. We intentionally used the statement of women about losing urine upon physical exercise. Although this is a subjective parameter, it does give information about the experience of women over a longer period of time. This may provide more or different information than objective parameters like a postoperative stress test, pad test, or urodynamic investigation, which just reflect a measurement at one point in time. We realize however that in a small number of women it is quite possible that \"failures\" are actually reporting the symptom of urge incontinence rather than the diagnosis of genuine stress incontinence. To demonstrate this difference, objective testing would be required.\n\nOur strict definition of success, being no loss of urine at all, may be unrealistic compared to the occurrence of incontinence and experience of women in the normal population. Becoming fully continent after TVT is not synonymous with being satisfied. A number of operated women still leaked postoperatively but were satisfied. These were mostly women with severe incontinence who experienced substantial improvement, the incontinence being reduced to an acceptable degree. This corroborates the finding that women can cope with a certain degree of incontinence for which therapy is not warranted \\[[@CR20]\\]. However, for scientific purposes, it is perhaps better and gives accurate insight in cure and improvement rates. For counseling purposes to women, the improvement rate is also of importance. In our study with a long follow-up of 2 to 4\u00a0years, 72% of women with a TVT-O were cured and another 12% improved, while in the Monarc group, these numbers are, respectively, 65% and 21%. This indicates that most women are still much better off after these incontinence procedures.\n\nIn determining success, another important parameter is the improvement in quality of life. In our study, QoL parameters improved significantly in both groups, and no differences were found between the two procedures. In general, the improvement observed after both procedures seems comparable to QoL improvement after a TVT procedure \\[[@CR3]\\].\n\nOther functional changes are related to the development of de novo urge incontinence and voiding difficulty or the resolution of these symptoms. The risk for developing de novo urge incontinence seems to be equal after Monarc and TVT-O procedures. This observation is confirmed by the other comparative studies \\[[@CR15]-[@CR17]\\]. The number of women reporting voiding difficulty after 2\u00a0months is higher in the TVT-O group. There is no explanation for this difference. Latthe's results \\[[@CR21]\\] indicate that voiding difficulty was less for combined TVT-O and TOT tapes compared to the retropubic TVT procedure. On subgroup analysis, they found no differences between the procedures by either obturator route.\n\nAnatomical complications like bladder or urethra injury or bleeding of more than 300\u00a0ml did not occur in this study. Furthermore, the three remaining complications in this study (two accidental fausse routes and one premature cutting of the tape) are relative minor incidents. Both TVT-O and Monarc have substantial lower intra-operative complication rates than TVT, but no differences between both procedures are observed \\[[@CR21]\\].\n\nWhen TVT-O and Monarc became available, there was some evidence that one procedure could be less safe than the other. Achtari et al. \\[[@CR22]\\] showed by cadaveric dissection that TVT-O runs more closely to the obturator canal, making TVT-O more prone to possible injury of the obturator nerve and vessels. We did not find any obturator nerve or vessel injury, and apparently, this theoretical risk does not exist in clinical practice.\n\nAlthough the number of women is too small to draw conclusions, there seems to be a little higher erosion rate in the Monarc group. These results are confirmed by other publications but a significant difference was never reached, mostly due small patient numbers \\[[@CR21]\\]. Nevertheless, the postoperative complication rate in our study population is low, and no differences between both procedures were observed.\n\nAdvantages of this study are the large number of women with only 16% loss to follow-up after 2--4\u00a0years. Other advantages are a strict outcome parameter, a consistent treatment strategy where procedures were not combined with other urogynecological treatments, the standardized urogynecological work-up (including urodynamic investigation), and the fact that all women were treated in one clinic by two urogynecologists and residents (always with supervision and the presence of these urogynecologist during surgery).\n\nThe draw-back of this study may be the bias in follow-up results. Not all women visited our outpatient department for the follow-up procedure. There may be a difference in the results of women visiting our outpatient department compared to women answering questionnaires by mail or telephone. Nevertheless, the use of interviews by telephone allowed us to minimize the loss to follow-up.\n\nThis study shows equal clinical cure rates and improvement in quality of life for both TVT-O and Monarc as a surgical treatment for stress urinary incontinence on the long run. Nevertheless, while longer follow-up data of TVT are now available \\[[@CR23]\\], longer follow-up than 4\u00a0years is still mandatory.\n\n**Conflicts of interest** None.\n\n**Open Access** This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.\n"}
+{"text": "Background\n==========\n\nThe importance of physical activity as a means to lower cardiovascular disease risk is well established \\[[@B1]-[@B4]\\]. Arguably, lifestyle activity or habitual activity associated with daily living has the most public health potential and is of great interest to physical activity researchers at present \\[[@B5]\\]. Traditionally, physical activity interventions have focused on promoting more leisure-time activities (e.g., structured exercise programmes). But, lack of evidence for effectiveness and sustainability of any changes \\[[@B6]\\], has led to growth in interest in designing environments that are conducive to active living, for instance promoting habitual activities in leisure time or active commuting) \\[[@B7]-[@B9]\\].\n\nPrevious research has also established higher risk of CVD, particularly in South Asian or African descent ethnic minority groups compared to the European descent Dutch population \\[[@B10]\\]. These ethnic minority groups from South Asian or African descent constitute about one-fifth of the non-Western population in the Netherlands \\[[@B11]\\], and have lower rates of recommended physical activity with less contribution of active commuting compared to the Dutch majority population \\[[@B12]\\]. This makes them an important target group for prevention. In European-origin populations, preventive interventions often target groups in *lower* socioeconomic positions \\[[@B13]\\], as these groups are at relatively high risk of cardiovascular disease and often report lower levels of LTPA compared to the higher socioeconomic position groups \\[[@B14]-[@B16]\\]. Whether the focus of preventive interventions on persons in lower socioeconomic positions is also relevant among the South Asian and African descent ethnic minority groups is less clear.\n\nFew studies have been carried out on this topic, and they show conflicting results: some have reported that the patterns within ethnic groups were similar to those in groups of European origin, while others found a lack of differences in socioeconomic position and LTPA \\[[@B17]-[@B21]\\]. In addition, the studies were mainly carried out among African-Americans in the United States. The situation in the European ethnic population groups may be different, as differences in national context may affect health behavior \\[[@B22]\\]. In the Netherlands, we previously found that low socioeconomic position was associated with increased risk of metabolic syndrome among European-Dutch people, but not among the South Asian and African origin ethnic minority groups \\[[@B23]\\]. Currently, it is uncertain whether such a difference in association also exists between socioeconomic position and physical activity in these ethnic groups.\n\nGiven the potential public health gain that can be achieved among these high risk populations, the aim of this study, was to assess the relationship of active commuting and LTPA with measures of socioeconomic position across two large ethnic minority groups of South Asian and African origin living in the Netherlands as compared to the European-Dutch population.\n\nMethods\n=======\n\nStudy population\n----------------\n\nWe used data from the population-based SUNSET (Surinamese in the Netherlands: Study on Health and Ethnicity) study, which was set up to gain insight into the cardiovascular risk profile in people 35 to 60\u2009years of age in three ethnic groups (South Asian-Surinamese, African-Surinamese, and European-Dutch). The recruitment and design of the study have been described elsewhere \\[[@B10],[@B24]\\]. In brief, potential participants were randomly sampled (n\u2009=\u20092,975) from the population register of Amsterdam, the Netherlands. Between 2001 and 2003, potential participants were approached at home for a structured face-to-face interview with a trained interviewer. The interview contained questions on lifestyle, migration history, socioeconomic position, and general health status. Ethnicity was confirmed by self-report during the interview. After the interview, participants of South Asian-Surinamese, African-Surinamese, and European-Dutch ethnicity were invited for a physical examination at a local health center. The response to the interview was 60\u2009% in the Surinamese groups and 61\u2009% in the European-Dutch group. All participants from the SUNSET study who provided information on ethnicity and physical activity were included in the present analysis: n\u2009=\u20091,626 eligible participants from the European-Dutch (n\u2009=\u2009567), South Asian-Surinamese (n\u2009=\u2009370), and African-Surinamese (n\u2009=\u2009689) groups. The study was approved by the Institutional Review Board of the Academic Medical Center of the University of Amsterdam. All participants provided written informed consent.\n\nPhysical activity\n-----------------\n\nFor the current analysis we focused on easily modifiable domains of activity (i.e., active commuting and LTPA). These were assessed with an adapted Short Questionnaire to Assess Health-Enhancing Physical Activity (SQUASH). The original questionnaire has been validated in European-Dutch populations, but not in ethnic minority groups \\[[@B25],[@B26]\\]. We adapted the questionnaire, after consulting with researchers familiar with the target group, by adding some extra activities and an open-ended question identifying additional physical activity in leisure time \\[[@B12]\\]. Subsequently, the tailored questionnaire was judged on face validity by minority participants of the same background, but not further formally validated. In brief, we identified frequency, duration, and intensity of different types of major activities in the domains of active commuting (cycling and walking) and LTPA (cycling, walking, gardening, do-it-yourself activities, dancing, yoga, any other physical activity, and up to four sports). Metabolic equivalents of task (MET) values were determined based on the compendium of Ainsworth \\[[@B27]\\]. MET hours per week were calculated for active commuting and LTPA separately. MET hours per week were skewed and therefore square-root-transformed. Only moderate to vigorous activities (MET \\>3, non-age-specific) were included in this study, because light-intensity physical activity may be more susceptible to measurement error and response bias in subjective physical activity assessment \\[[@B28]\\]. Moreover, the focus on moderate to vigorous intensity activities will potentially reflect better the association of physical activity with health benefits.\n\nSocioeconomic position\n----------------------\n\nSocioeconomic position was assessed in the questionnaire by self-reported highest level of education attained and occupational class. Due to the relatively low number of persons in some of the original categories, we categorized level of education as low (secondary school and below) and high (college, polytechnic, or university). Occupational class was defined as high (professionals, managers, employers, and higher grade routine non-manual employees) and low (lower grade routine non-manual, other lower grade, and skilled and unskilled manual employees) \\[[@B29]\\].\n\nAnalysis\n--------\n\nTo assess differences in characteristics between ethnic groups, we used one-way analysis of variance for continuous variables (Kruskal-Wallis in the case of unequal variance between groups) and the chi-square test for categorical variables. Post-hoc tests with Bonferroni adjustment for multiple testing (three comparisons) were used to assess differences in characteristics.\n\nWe plotted absolute levels of physical activity in each ethnic group by socioeconomic position; this was done separately for men and for women. These figures were age-sex standardized in the ethnic groups to the age-sex distribution in the total study population (direct standardization). Linear regression models were used to assess the association between MET-hours per week, education, and occupational class. These analyses were adjusted for age and stratified by sex and ethnicity. Moreover, to calculate formal ethnicity interaction in the relationship with physical activity and socioeconomic position, we ran additional sex-stratified models, including interaction terms for ethnicity and socioeconomic position. An overall *p*-value for interaction was calculated based on the likelihood ratio test, comparing the model with and without interaction terms for socioeconomic position by ethnicity. Values of *p\u2009\\<*\u20090.05 were considered significant. Additionally, an ordinal regression analysis in categories of physical activity was considered to verify our findings from the linear regression analyses. Finally, because missing data for occupation occurred more often in the ethnic minority groups and these cases were more likely to report lower levels of active commuting and lower levels of education, we performed additional sensitivity analyses in which two extremes were considered: a situation in which all missing data for occupation was of low occupational class, and a situation in which all missing data for occupation was considered high occupational class. All analyses were performed in R version 2.15.0 (A Language and Environment for Statistical Computing, Vienna, Austria).\n\nResults\n=======\n\nThe European-Dutch population was older than the South Asian-Surinamese and African-Surinamese populations (Table [1](#T1){ref-type=\"table\"}). This was the case for both men and women. Level of education and occupational class were higher in European-Dutch men compared to South Asian-Surinamese and African-Surinamese men. In women, level of education and occupational class was also higher in European-Dutch women compared to South Asian-Surinamese women. Occupational class was higher in African-Surinamese women compared to South Asian-Surinamese women, but not different from European-Dutch women.\n\n###### \n\nCharacteristics of the European-Dutch, South Asian-Surinamese, and African-Surinamese men and women included in this study\n\n  \u00a0                                               **Men**                **Women**                                                                                        \n  ---------------------------------------- ---------------------- ----------------------- -------------------------- ---------------------- ----------------------------- --------------------------\n  n (%)                                          276 (41.3)             162 (24.2)                231 (34.5)               291 (30.4)                208 (21.7)                   458 (47.9)\n  Age, mean (SD)                                 47.8 (6.7)           44.3 (6.6)^\\*^            43.8 (6.3)^\\*^             47.4 (6.8)              44.8 (6.6)^\\*^              43.5 (5.8)^\\*\u2020^\n  **Education**                                                                                                                                                           \n  Low, n (%)                                   **80 (29.5)**          **112 (70.0)**            **136 (59.6)**           **111 (38.3)**            **139 (68.1)**               **207 (45.8)**\n  High, n (%)                                  **191 (70.5)**      **48 (30.0)**^**\\***^    **92 (40.4)**^**\\***^        **179 (61.7)**         **65 (31.9)**^**\\***^       **245 (54.2)**^**\u2020**^\n  **Occupational class**                                                                                                                                                  \n  Low, n (%)                                   **111 (40.2)**         **101 (62.3)**            **118 (51.1)**           **119 (40.9)**            **117 (56.2)**               **190 (41.5)**\n  High, n (%)                                  **160 (58.0)**      **45 (27.8)**^**\\***^    **83 (35.9)**^**\\***^        **162 (55.7)**         **63 (30.3)**^**\\***^         **211 (46.1)**^\u2020^\n  **Physical activity (MET hours/week)**                                                                                                                                  \n  in active commuting, median (IQR)          **0.0 (0.0-10.0)**     **0.0 (0.0-0.0)\\***    **0.0 (0.0-5.8)**^**\u2020**^    **0.0 (0.0-7.4)**          **0.0 (0.0-5.4)**        **2.4 (0.0-8.2)**^**\u2020**^\n  transformed mean (SD)                          1.6 (1.9)             0.5 (1.1)^\\*^             1.2 (1.5)^\u2020^              1.4 (1.8)                  1.1 (1.5)                1.6 (1.7)^**\u2020**^\n  in LTPA, median (IQR)                     **25.5 (10.5-47.6)**    **20.5 (5.1-39.2)**      **24.5 (9.4-51.2)**      **24.5 (11.7-38.5)**   **16.6 (5.2-30.6)**^**\\***^     **19.1 (6.6-42.0)**\n  transformed mean (SD)                          5.2 (3.0)             4.4 (3.1)^\\*^             5.4 (3.9)^\u2020^              4.8 (2.5)                 4.0 (2.7)\\*                  4.5 (3.0)\n\nMET: metabolic equivalent of task. IQR: interquartile range. SD: standard deviation. LTPA: leisure-time physical activity. Bold values represent an overall group difference (chi-square or rank-sum p\u2009\\<\u20090.05). Bonferroni adjusted group differences (3 comparisons): \\* different from European-Dutch (p\u2009\\<\u20090.05). ^\u2020^ different from South Asian-Surinamese (p\u2009\\<\u20090.05).\n\nEuropean-Dutch and African-Surinamese men engaged more in active commuting and LTPA compared to South Asian-Surinamese men. African-Surinamese women engaged more in active commuting than South Asian-Surinamese women. European-Dutch women engaged more in LTPA compared to South Asian-Surinamese women.\n\nFigure [1](#F1){ref-type=\"fig\"} and [2](#F2){ref-type=\"fig\"} illustrate the differences in absolute level of physical activity by measure of socioeconomic position in the ethnicity-sex groups (absolute values available in Additional file [1](#S1){ref-type=\"supplementary-material\"} and Additional file [2](#S2){ref-type=\"supplementary-material\"}). For active commuting, higher level of education was positively associated with MET hours/week in European-Dutch men (beta: 0.93, 95\u2009% CI: 0.45 1.40), and South Asian-Surinamese men (0.56, 0.19 0.92) but not in African-Surinamese men (\u22120.06, -0.45 0.33, *p* for interaction\u2009=\u20090.002, Table [2](#T2){ref-type=\"table\"}). For occupational class, the pattern of differences in active commuting across ethnic groups was similar compared to education in men, although the interaction between active commuting and ethnicity was not significant. In women, we observed the reverse for the socioeconomic position measures: there was no ethnicity interaction for active commuting by level of education, whereas we found a statistically significant differential association by ethnicity for occupational class. The gradient for active commuting and occupational class was less profound in ethnic minority women than in European-Dutch women. In European-Dutch women, active commuting was associated positively with occupational class (beta: 0.86, 95%CI: 0.44 1.28), but this was not observed in South Asian-Surinamese women (0.20, -0.25 0.66) or in African-Surinamese women (0.18, -0.15 0.50, *p* for interaction\u2009=\u20090.024, Table [2](#T2){ref-type=\"table\"}). The additional sensitivity analyses showed that this interaction for occupational class remained fairly stable in the extreme situation when all missing data were considered of low occupational class (*p* for interaction\u2009=\u20090.07) or high occupational class (*p* for interaction\u2009=\u20090.02, data available on request).\n\n![**Mean absolute levels of active commuting and leisure-time physical activity (square-root-transformed and age-sex standardized MET hours/week) in European-Dutch, South Asian-Surinamese, and African-Surinamese ethnic groups by level of education (low/high).** Statistical significance of the difference of the high compared to the low level of education group is indicated by: \\*\\*\\* (p\u2009\\<\u20090.001), \\*\\* (p\u2009\\<\u20090.01), \\* (p\u2009\\<\u20090.05).](1471-2458-12-815-1){#F1}\n\n![**Mean absolute levels of active commuting and leisure-time physical activity (square-root-transformed and age-sex standardized MET hours/week) in European-Dutch, South Asian-Surinamese, and African-Surinamese ethnic groups by occupational status (manual/non-manual).** Statistical significance of the difference of the non-manual and manual group is indicated by: \\*\\*\\* (p\u2009\\<\u20090.001), \\*\\* (p\u2009\\<\u20090.01), \\* (p\u2009\\<\u20090.05).](1471-2458-12-815-2){#F2}\n\n###### \n\n**Association of*active commuting*in MET hours/week with education and occupational class*in men and women***\n\n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n  \u00a0                        **Men**            ***p*for interaction**   **Women**            ***p*for interaction**                                                              \n  ------------------------ ------------------ ------------------------ -------------------- ------------------------ ------------------ ------------------- ------------------- -------\n  **Education**                                                                                                                                                                 \n\n  Low                      ref\u2009=\u20090            ref\u2009=\u20090                  ref\u2009=\u20090              \u00a0                        ref\u2009=\u20090            ref\u2009=\u20090             ref\u2009=\u20090             \u00a0\n\n  High, \u03b2 (95\u2009% CI)        0.93 (0.45 1.40)   0.56 (0.19 0.92)         \u22120.06 (\u22120.45 0.33)   0.002                    0.53 (0.09 0.96)   0.22 (\u22120.22 0.65)   0.43 (0.12 0.73)    0.593\n\n  **Occupational class**                                                                                                                                                        \n\n  Low                      ref\u2009=\u20090            ref\u2009=\u20090                  ref\u2009=\u20090              \u00a0                        ref\u2009=\u20090            ref\u2009=\u20090             ref\u2009=\u20090             \u00a0\n\n  High, \u03b2\\                 0.62 (0.17 1.07)   0.52 (0.12 0.91)         0.28 (\u22120.14 0.70)    0.45                     0.86 (0.44 1.28)   0.20 (\u22120.25 0.66)   0.18 (\u22120.15 0.50)   0.024\n  (95\u2009% CI)                                                                                                                                                                     \n  -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\nMET: metabolic equivalent of task. Presented betas are based on square-root-transformed MET hours/week from moderate- and vigorous-intensity physical activity. Estimates are modeled separately for education and occupational class and adjusted for age.\n\nThe overall test for interaction is a likelihood ratio test comparing the models with and without interaction terms.\n\nFor LTPA, the association with education did not appear to significantly differ across ethnic groups among men (Table [3](#T3){ref-type=\"table\"}). For women, we also found no differences in the association between LTPA and the measures of socioeconomic position across the ethnic groups.\n\n###### \n\n**Association of*LTPA in MET hours/week*with education and occupational class*in men and women***\n\n  \u00a0                        **Men**              ***p*for interaction**   **Women**            ***p*for interaction**                                                               \n  ------------------------ -------------------- ------------------------ -------------------- ------------------------ ------------------- ------------------- ------------------- -------\n  **Education**                                                                                                                                                                    \n  Low                      ref\u2009=\u20090              ref\u2009=\u20090                  ref\u2009=\u20090              \u00a0                        ref\u2009=\u20090             ref\u2009=\u20090             ref\u2009=\u20090             \u00a0\n  High, \u03b2 (95\u2009% CI)        0.41 (\u22120.35 1.17)    1.03 (\u22120.00 2.06)        \u22120.01 (\u22121.04 1.02)   0.377                    0.13 (\u22120.48 0.73)   0.72 (\u22120.08 1.51)   0.20 (\u22120.35 0.75)   0.536\n  **Occupational class**                                                                                                                                                           \n  Low                      ref\u2009=\u20090              ref\u2009=\u20090                  ref\u2009=\u20090              \u00a0                        ref\u2009=\u20090             ref\u2009=\u20090             ref\u2009=\u20090             \u00a0\n  High, \u03b2 (95\u2009% CI)        \u22120.36 (\u22121.09 0.36)   0.92 (\u22120.17 2.00)        \u22120.27 (\u22121.33 0.79)   0.16                     0.22 (\u22120.36 0.81)   0.13 (\u22120.72 0.97)   0.06 (\u22120.53 0.64)   0.93\n\nLTPA: leisure-time physical activity. MET: metabolic equivalent of task. Presented betas are based on square-root-transformed MET hours/week from moderate- and vigorous-intensity physical activity. Estimates are modeled separately for education and occupational class and adjusted for age.\n\nThe overall test for interaction is a likelihood ratio test comparing the models with and without interaction terms.\n\nThe results from the ordinal regression analysis were consistent with the associations and interactions presented above (Additional file [3](#S3){ref-type=\"supplementary-material\"}).\n\nDiscussion\n==========\n\nMain findings\n-------------\n\nWe found a positive association between socioeconomic position and physical activity, in terms of active commuting and LTPA. However, the association between socioeconomic position and active commuting differed across ethnic groups. In men, we found a positive association with socioeconomic position (in particular for high level of education) in the European-Dutch, which was less strong in South Asian-Surinamese and African-Surinamese men. Among women we observed a similar pattern of interaction for socioeconomic position (but then for occupational class), with a positive association in the European-Dutch women and less strong associations in the South Asian-Surinamese and African-Surinamese women.\n\nFor LTPA, among both men and women, we were unable to observe a clear pattern of differences in the association with socioeconomic position between the ethnic groups.\n\nOur study shows that active commuting appears to be less strongly related to socioeconomic position across ethnic minority groups compared to the positive association in the European-Dutch. This was most apparent among the African-Surinamese. We are unaware of any earlier population studies that focused specifically on the relationship of active commuting and measures of socioeconomic position across multiple ethnic groups.\n\nThe weaker socioeconomic gradient with active commuting in the Surinamese compared to the European-Dutch could be related to the relatively higher number of European-Dutch with high levels of education \\[[@B30]\\], which was not captured by our measure of socioeconomic position.\n\nAnother possible explanation for the observed differential association between active commuting and socioeconomic position across ethnic minority groups is a difference in preferred mode of transportation. We speculate that the higher level of active commuting in the high socioeconomic position European-Dutch compared to their low socioeconomic position counterparts may be related to the frequency of bicycle use in this population compared to other groups \\[[@B12],[@B31]\\], while the observation of a lack of association between low and high socioeconomic position for active commuting with African-Surinamese groups might be related to a preference for walking or using public transportation for their commutes \\[[@B12],[@B31],[@B32]\\]. Environmental factors, such as neighborhood factors, car/bike ownership and distance to employment, could be an important explanation for the observed patterns between the ethnic groups. However, these factors were not assessed in this study. More research is clearly necessary which takes into account the influence of such environmental factors.\n\nFor LTPA, we observed no clear differential association with socioeconomic position across ethnic minority groups. In our study, the low socioeconomic groups from all backgrounds engaged in levels of LTPA that were comparable to their high socioeconomic position counterparts; this diminishes the likelihood of a socioeconomic gradient. A lack of association between LTPA and socioeconomic position has been reported before \\[[@B33]\\]. However, other studies have found an ethnicity interaction in the relationship between LTPA and socioeconomic position \\[[@B34]\\] or associations of LTPA and socioeconomic position, also in ethnic minority groups \\[[@B18],[@B19]\\]. The difference between these studies and our own might be related to the socioeconomic context in which the studies were carried out. In our study, the low socioeconomic position groups engaged in relatively high levels of LTPA (compared to their high socioeconomic position counterparts), which diminishes the likelihood of a socioeconomic gradient. One methodological difference between our study and that of others potentially relating to this finding was the broad measurement of physical activity in our study. The focus of our questionnaire was to measure habitual activity in leisure time, in addition to measure participation in sports and exercise in leisure time. In other studies, and of course depending on the reasons for the inquiry, sports and exercise were often the only group of activities identified from the leisure time domain \\[[@B35]\\]. The habitual activity (e.g., walking, cycling, gardening, dancing, exercising at home) additionally measured in our study in LTPA could be less constrained by barriers (e.g., financial, or facilities within range) often reported in low socioeconomic position groups. Therefore, this could have resulted in higher levels of measured physical activity during leisure time in the low socioeconomic position group in our study; which potentially diminishes the socioeconomic gradient in LTPA. Additionally, results from qualitative studies indicate that those in low socioeconomic position groups and ethnic minority groups are aware of the beneficial health aspects of LTPA and exercise, which could also have mitigated the socioeconomic gradient in our study population \\[[@B36]-[@B38]\\].\n\nFinally, although not as strong as the European-Dutch gradient, the positive gradient in active commuting for the South Asian-Surinamese and the lack of ethnicity interaction in the relationship between socioeconomic position and LTPA fits with the convergence hypothesis that the socioeconomic gradient in the South Asian-Surinamese and African-Surinamese ethnic minority groups may change in the direction of European-Dutch levels. This is in line with the observation by Bos et al. that the overall mortality pattern in Surinamese men and women and cardiovascular mortality in Surinamese women compared to the European-Dutch may indicate convergence of health-behavior with increasing duration of residence \\[[@B39]\\]. Additionally, Nierkens et al. suggested that the Dutch Surinamese population, especially men, have already reversed the socioeconomic gradient in smoking behavior that would normally be expected according to the stages of the tobacco epidemic \\[[@B40]\\]. In the Indian ethnic minority group living in the United Kingdom (UK), convergence towards the majority population of a wide range of health behaviors is also observed \\[[@B41],[@B42]\\], this group can be characterized with an even longer average duration of residence compared to other South Asian ethnic minority groups in the UK, and the South Asian- and African-Surinamese in the Netherlands \\[[@B43]\\].\n\nGiven these examples, it seems probable that in the long term the physical activity patterns (also for active commuting) observed in the South Asian-Surinamese and African-Surinamese populations will converge towards European averages. Especially in the case of socioeconomic advancement in the Surinamese groups, this balance is expected to tip in favor of a more European pattern of socioeconomic gradient. To be able to observe such trends, though, longitudinal data on physical activity and socioeconomic position within the ethnic groups is a prerequisite.\n\nLimitations\n-----------\n\nSocioeconomic position is a complex social construct for which various indicators have been shown to capture different aspects of importance in relation to the outcome under study \\[[@B44]\\]. In the Netherlands, education is identified as a stable indicator of socioeconomic position \\[[@B45]\\]. Because of the small numbers per group in some of the original categories, we were limited to a relatively crude high-low comparison. As a result we might have missed subgroup patterns. Therefore, we presented an additional measure of socioeconomic position, namely based on occupational class. We did not include other measures of socioeconomic position (such as income), which could have provided additional insights into the association of socioeconomic position and physical activity across ethnic minority groups. For example, higher income could enable easier access to facilities \\[[@B44]\\].\n\nWe were unable to take into account processes of acculturation in this study, while acculturation processes in the migrant population are also expected to contribute to the convergence of physical activity behavior towards the majority population in a country \\[[@B46]\\]. Duration of residence was measured and previously used as proxy of acculturation \\[[@B22]\\], but not included in this analyses because of a lack of variation observed in our study population. The majority of the ethnic minority individuals in our study population have a relative long duration of residence, which makes such an analysis less informative.\n\nFurthermore, self-reported levels of physical activity are prone to reporting bias or cognitive bias \\[[@B47]\\], resulting in measurement error that could have influenced the presented findings. Another point relating to our measurement of physical activity is that we focused on moderate to vigorous physical activity, which is expected to capture the relationship with health, but excluded light walking activity, which may also have a beneficial effect on health. Moreover, the focus on LTPA and active commuting meant that we could not take into account potential socio-economic differences in occupational activity \\[[@B48]\\], however in our study population the relationship between leisure-time/active commuting and occupational physical activity was similar across ethnic groups (unpublished). Additionally, there might be cross-cultural differences in reporting levels of physical activity \\[[@B49]\\]. Since we were mainly interested in the association of physical activity *within* each ethnic minority group, we believe that cross-cultural differences in reporting might be less of a problem in this study.\n\nFinally, we acknowledge that the generalizability of the findings presented in this study may be limited to the included ethnic groups and context. The data were collected between 2001 and 2003. Since then, circumstances may have changed. However, the data presented are relevant, as more recent studies, and studies including more information on environmental factors, are unfortunately not yet available for these ethnic minority groups in the Netherlands. Additionally, this study reported on cross-sectional study data, which means that any causal inference from this data should be made with caution. Other patterns in active commuting or LTPA by socioeconomic position might exist in those from ethnic minority groups of similar descent who live elsewhere or with a different migration history.\n\nConclusion\n==========\n\nOur study suggests that the positive socioeconomic gradient in active commuting observed in the European-Dutch may be less strong in the South Asian- and African-Surinamese ethnic groups. This was not observed for LTPA. The independent effect of ethnicity on active commuting was observed for men and women, but this was dependent on the measure of socioeconomic position used.\n\nOur findings imply that, contrary to the typical targeting of physical activity interventions to the lower socioeconomic groups often applied in European-origin populations, a broader focus may be needed for activities or recommendations for physical activity among African- and South Asian-origin groups. Specifically, public health workers should be aware of the current low levels of active commuting in both low and high socioeconomic position groups from these populations. Based on this, it seems relevant to stimulate active commuting among the African-Surinamese and South Asian-Surinamese in the Netherlands in both low and high socioeconomic position groups.\n\nAbbreviations\n=============\n\nLTPA: Leisure-time physical activity; MET: Metabolic equivalents of task; SUNSET: Surinamese in the Netherlands: Study on Health and Ethnicity; SQUASH: Short Questionnaire to Assess Health-Enhancing Physical Activity.\n\nCompeting interests\n===================\n\nThe authors declare that theyhave no competing interests.\n\nAuthors' contributions\n======================\n\nJM designed the study, carried out the data analysis and interpretation, and drafted the first version of the manuscript. LB participated in interpretation of data and critical revisions of the manuscript. CA, KS, IV contributed to the conception and design of the study, participated in interpretation of data and critical revisions of the manuscript. All authors approved the final version of this manuscript.\n\nPre-publication history\n=======================\n\nThe pre-publication history for this paper can be accessed here:\n\n<http://www.biomedcentral.com/1471-2458/12/815/prepub>\n\nSupplementary Material\n======================\n\n###### Additional file 1\n\nActive commuting and LTPA in MET hours/week by level of education in men and women.\n\n###### \n\nClick here for file\n\n###### Additional file 2\n\nActive commuting and LTPA in MET hours/week by occupation in men and women.\n\n###### \n\nClick here for file\n\n###### Additional file 3\n\n**Association of*active commuting*with education and occupational class*in men and women.***\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nWe would like to acknowledge Dr. Anton Kunst for his expertise regarding some of the more complex social class issues in this paper. Additionally, we would like to acknowledge Colleen Higgins for her assistance with the language editing of this manuscript. This project was supported by the Academic Medical Center, University of Amsterdam, and the VENI fellowship (grant number 916.76.130) awarded by the board of the Council for Earth and Life Sciences (ALW) of the Netherlands Organisation for Scientific Research (NWO) as part of its Innovational Research Incentives Scheme. LMB is a recipient of a VENI fellowship (grant number 916.10.156) awarded by the Netherlands Organisation for Scientific Research (NWO) as part of its Innovational Research Incentives Scheme.\n"}
+{"text": "INTRODUCTION\n============\n\nSpermatogenesis is a highly regulated process that involves both mitotic and meiotic divisions, as well as cellular differentiation to yield mature spermatozoa from undifferentiated germinal stem cells. The endocrine regulation of spermatogenesis occurs through the combined action of hormones and growth factors, including the effects of retinoic acids (RAs), which are vitamin A derivatives essential for meiosis initiation \\[[@B1],[@B2]\\]. Epigenetic modifications and chromatin remodelling are also important regulators in spermatogenesis \\[[@B3]\\]. PIWI-interacting RNAs (piRNAs) are small non-coding RNAs synthesized in germ cells, where they are believed to repress the expression of deleterious retrotransposons, thereby ensuring genome integrity during meiosis \\[[@B4]\\]. In mice these small non-coding RNAs interact with the Argonaute family proteins (Ago and PIWI clades of proteins), such as MIWI, MILI and MIWI2 members, which also participate in piRNA biogenesis. Glycerol-3-phosphate acyltransferase-2 (GPAT2) was unexpectedly found to be required for piRNA biogenesis. The knockdown of *Gpat2* in germline stem cells impairs primary piRNA production, and it was shown that the GPAT2 protein physically interacts with MILI in both germline stem cells and mouse testis. The mechanism by which GPAT2 promotes piRNA production remains unknown, but because this function occurs despite the absence of the acyltransferase motif, H(X~4~)D, the regulation of piRNA production by GPAT2 is unlikely to require acyltransferase enzymatic activity \\[[@B5]\\].\n\nGPAT isoforms catalyse the first and rate-limiting step in *de novo* glycerolipid biosynthesis, acylating glycerol 3-phosphate with a long-chain fatty acyl-CoA to form lysophosphatidic acid (LPA). LPA is the substrate for 1-acylglycerol-3-phosphate acyltransferase which catalyses the formation of phosphatidic acid, the precursor for triacylglycerol and glycerophospholipid biosynthesis. In mammals, four GPAT isoforms (GPAT1--GPAT4) have been described which differ in their subcellular locations, tissue expression pattern, substrate preference, transcriptional regulation and sensitivity to sulfhydryl group reagents such as *N*-ethylmaleimide \\[[@B6]\\]. GPAT1 and GPAT2 are mitochondrial isoforms, whereas GPAT3 and GPAT4 are located in the endoplasmic reticulum. GPAT1, GPAT3 and GPAT4 are highly expressed in lipogenic tissues such as liver and adipose tissue where they initiate the *de novo* synthesis of triacylglycerol and glycerophospholipids \\[[@B6]\\]. During adipocyte differentiation of 3T3-L1 fibroblasts, the expression of these three isoforms increases; *Gpat1* and *Gpat4* are moderately induced (10- and 5-fold, respectively) in differentiated adipocytes and *Gpat3* is highly induced (\\>60-fold) \\[[@B7]--[@B9]\\]. In contrast, *Gpat2* mRNA expression is not regulated by nutritional status, and it is at least 50-fold higher in testis than in any other tissue \\[[@B10]\\].\n\nGPAT2 also differs from the other GPAT isoforms in that it belongs to a group of genes named 'cancer-testis genes' (CTs). CTs are important at specific stages during spermatogenesis, but are expressed at low levels or not expressed at all in somatic tissues. CTs are ectopically overexpressed in cancers from different locations, and contribute to the tumour phenotype \\[[@B11],[@B12]\\]. We have shown that human GPAT2 is overexpressed in several types\u00a0of cancer and in cancer-derived human cell lines, and that its expression contributes to the tumour phenotype, because in tumour cells with diminished GPAT2 expression, the proliferation and migration rates are lower, and in mouse xenograft models, cells with diminished GPAT2 expression have lower tumorigenicity \\[[@B13]\\].\n\nTo understand the physiological role of GPAT2\u00a0in the testis, we focused on its developmental regulation during spermatogenesis. Our data show that *Gpat2* is expressed at a specific stage of the spermatogenesis, consistent with meiosis I prophase and that its on--off expression pattern responds predominantly to epigenetic modifications and probably to RA signalling.\n\nMATERIALS AND METHODS\n=====================\n\nChemicals\n---------\n\nAll chemicals were purchased from Sigma unless otherwise indicated.\n\nAnimals, cell lines and culture conditions\n------------------------------------------\n\nAnimal protocols were approved by the Facultad de Ciencias M\u00e9dicas, Universidad Nacional de La Plata Institutional Animal Care and Use Committee (Approval Number T10-02-2013). Male BALB/c mice were housed on a 12-h light/12-h dark cycle with free access to water and Cargill Rodent chow. Before testis dissections, mice were killed\u00a0in a\u00a0CO~2~\u00a0chamber. Cells were purchased from A.T.C.C. and were grown at 37\u00b0C in a 5% CO~2~ atmosphere with a 98% relative humidity. The hamster cell line CHO (Chinese-hamster ovary)-K1 was maintained in Ham\\'s F12 (Gibco) medium, the mouse macrophage cell line RAW 264.7 and the embryonic fibroblast 3T3-L1 mouse cell line in Dulbecco\\'s modified Eagle\\'s medium (DMEM) (Gibco), the human breast cancer cell line derived from metastatic pleural effusion MCF7\u00a0in RPMI 1640 medium (Gibco) and the mouse Sertoli cell line TM4\u00a0in DMEM/Ham\\'s F12 (Gibco). All media were supplemented with 10% FBS and antibiotics (50\u00a0units/ml penicillin and 50\u00a0\u03bcg/ml streptomycin).\n\nAnalysis of mRNA transcription\n------------------------------\n\nTotal RNA from BALB/c mouse testes at 3, 7, 11, 15, 20, 35 and 50\u00a0days post-partum (dpp), RAW 264.7 and 3T3L1 cells was isolated with TRIzol Reagent (Invitrogen) according to the manufacturer\\'s instructions. RNA quality was determined by gel electrophoresis and 260/230 and 260/280\u00a0nm absorbance ratios. A 1\u00a0\u03bcg sample of total RNA was used for cDNA synthesis employing the High Capacity Reverse Transcription Kit (Applied Biosystems). A 1/15 cDNA dilution was used for the qPCR (quantitative real-time PCR) with ABsolute QPCR SYBR Green Mix (Thermo). Primers were designed to amplify a fragment of 187\u00a0bp between exon 16 (forward primer: 5\u2032-ATCCTACTGCTGCTGCACCT-3\u2032) and exon 18 (reverse primer: 5\u2032-ACAGCAGCTTTGCACTCAGA-3\u2032) of the mouse *Gpat2* transcript. The thermal profile was 95\u00b0C for 15\u00a0min, followed by 40 cycles of 95\u00b0C for 30\u00a0s, 55\u00b0C for 1\u00a0min and 72\u00b0C for 30\u00a0s, on a Stratagene Mx3000P apparatus. RNA expression was quantified in duplicate using the \u0394\u0394*C*~T~ method, and each sample was normalized to two different housekeeping genes: *Gapdh* (forward primer: 5\u2032-CTGGAGAAACCTGCCAAGTA-3\u2032; reverse primer: 5\u2032-TGTTGCTGTAGCCCGTATTCA-3\u2032) and *Rpl13a* (forward primer: 5\u2032-ATGACAAGAAAAAGCGGATG-3\u2032; reverse primer: 5\u2032-CTTTTCTGCCTGTTTCCGTA-3\u2032) using qBase software.\n\n*In silico* evaluation of mouse *Gpat2* expression\n--------------------------------------------------\n\nTo validate experimental results regarding the *Gpat2* mouse expression profile during testis development, microarray data from pre- and post-natal ages were collected to construct a box plot graphic. Gene expression information was obtained from GSE4818 (*n=* 21, mouse testis prenatal development) and GSE12769 (*n=* 20, mouse testis postnatal development) datasets, both developed in the Affymetrix genechip mouse genome 430 2.0 platform (Mouse430_2). To generate a homogeneous dataset in fetal and adult mouse testes, the frozen robust multi-array analysis (fRMA) package from Bioconductor (<http://www.bioconductor.org/>) was employed. The fRMA preprocessing algorithm allows the analysis of independent oligo-microarray studies/batches, and then combines the data for further statistical analysis. *Gpat2* mRNA expression level was estimated by using the expression values of the Affymetrix probe 1456208_AT.\n\n*In situ* hybridization\n-----------------------\n\nRiboprobes for *in situ* hybridization were prepared by digesting mouse *Gpat2* cDNA from the pcDNA3.1-Gpat2 construct at the BamHI and EcoRI sites and subcloning the 1456\u00a0bp fragment into pGEM11z(f)+ vector (Promega). The tissue preparation, probe synthesis and hybridization were performed at the University of North Carolina at Chapel Hill Neuroscience Center Molecular Neuroscience Core Facility (<http://www.med.unc.edu/neuroscience/core-facilities/molecular-neuroscience>). Testes from 7, 15 and 30 dpp mice were fixed in 4% paraformaldehyde in 0.1\u00a0M PBS, and processed as previously described \\[[@B14]\\]. Slides were visualized in an Olympus BX52 microscope.\n\nImmunohistochemistry analysis\n-----------------------------\n\nTestes from 7, 15, 30 and 40 dpp mice were fixed in Bouin\\'s solution in the same paraffin block, and 4\u00a0\u03bcm sections were cut. Samples were processed as previously described \\[[@B14]\\]. Slides were counter-stained with haematoxylin to visualize the nuclei and analysed with an Olympus BX52 microscope.\n\n3T3-L1 differentiation assay\n----------------------------\n\n3T3-L1 cells were seeded in 12-well plates and grown in routine medium until confluence, then stimulated with 2\u00a0\u03bcg/ml insulin, 0.25\u00a0\u03bcM dexamethasone and 0.5\u00a0mM isobutylmethylxanthine (IBMX) \\[[@B15]\\]. Full differentiation was achieved on day 8. RNA (triplicate samples) was isolated with TRIzol Reagent at different times: 0, 12, 18, 24, 36, 48, 96, 144 and 192\u00a0h. A 1\u00a0\u03bcg sample was used for cDNA synthesis, and *Gpat2* content was evaluated by qPCR.\n\nCloning of *Gpat2* gene promoter regions\n----------------------------------------\n\nPrimer pairs were used to amplify different regions of the putative *Gpat2* promoter upstream of the ATG start codon from mouse genomic DNA; all included a terminal MluI (forward primer) or HindIII (reverse primer) restriction sites. Forward primers: \u22121324P fragment (\u22121324/+2105) 5\u2032-ACGCGTGGTTCTG-AAACTGGAGGTCAG-3\u2032, \u22121165P fragment (\u22121165/+2105) 5\u2032-ACGCGTGACCAGGAGAGGGTGCTAGA-3\\', \u2212873P frag-ment (\u2212873/+2105) 5\u2032-ACGCGTACAGCTGACCAAAAGC-CACT-3\u2032, \u2212642P fragment (\u2212642/+2105) 5\u2032-ACGCGTTG-GGGTAATTGGTTCTCACC-3\u2032, \u2212156P fragment (\u2212156/+2105) 5\u2032-ACGCGTTGTGGTACAGGCAGCAAGTC-3\u2032 and +165P fragment (+165/+2015) 5\u2032-ACAGTCAGAGGCAAG-CTGGT-3\u2032. The reverse primer was 5\u2032-AAGCTTTCTGTAAG-ATCAGTGAATCAAGCAC-3\u2032. To test for possible downstream regulation, the first non-translated exon of *Gpat2* mRNA as well as the first intron were included in the promoter constructs \\[[@B16]\\]. All five fragments were amplified using iProof High-Fidelity DNA Polymerase (Bio-Rad Laboratories), double digested with MluI and HindIII and ligated into the pGL3-Basic Vector (Promega). All plasmids were prepared using the Plasmid Miniprep kit (Qiagen), quantified by UV spectrophotometry and sequenced.\n\nCell transfection, hormone response evaluation and luciferase assay\n-------------------------------------------------------------------\n\nTo study basal promoter activity, CHO-K1 cells were seeded in a 24-well plate. The next day Lipofectamine 2000 (Invitrogen) was used to co-transfect 0.5\u00a0\u03bcg/well of each pGL3 construct with 0.05\u00a0\u03bcg/well of pRL*-Renilla* Luciferase Reporter Vector (*Renilla*). Thirty-six hours after transfection, cells were harvested and lysed and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega). Relative luciferase activity was normalized by the *Renilla* luciferase activity. To evaluate the hormone response, CHO-K1, MCF7 or TM4 cells were seeded in a 48-well plate 24\u00a0h before Lipofectamine 2000 co-transfection with 1.14\u00a0\u03bcg/well of pGL3 constructs and 0.11\u00a0\u03bcg/well of *Renilla* luciferase plasmid. Five hours after transfection, medium with hormones and lacking FBS, was added to the cells. The next day, the medium was replaced with fresh medium and hormones. Forty-eight hours after transfection, cells were harvested and lysed and luciferase activity was measured as described above. All experiments were performed in duplicate and repeated at least twice. Hormones tested (concentration range is indicated within parentheses): oestradiol (0.1--10\u00a0nM), corticosterone (0.1--0.2\u00a0\u03bcM), all-*trans*-retinoic acid (ATRA) (2--20\u00a0\u03bcM) and 9-*cis*-retinoic acid (9cisRA) (2--8\u00a0\u03bcM).\n\nDAC and TSA treatments\n----------------------\n\nThe effect of the DNA methyltransferase inhibitor 5-aza-2-deoxycitidine (DAC) and the histone deacetylase inhibitor trichostatin A (TSA) on *Gpat2* transcription was studied in RAW 264.7 cells. Cells were seeded in a 12-well plate to reach 20% confluence. The next day, the medium was changed, and cells were treated with 2\u00a0\u03bcM DAC (for 'DAC' and 'TSA + DAC' treatments) or DMSO (control). Media and DAC were replaced every 24\u00a0h for 3\u00a0days. TSA (500\u00a0nM) was added on the fourth day ('TSA' and 'TSA + DAC' treatments) and cells were harvested on day 5. Total RNA was isolated using TRIzol Reagent (Invitrogen); 1\u00a0\u03bcg was used for cDNA synthesis and *Gpat2* mRNA content was measured by qPCR. Each treatment was performed twice in triplicate.\n\n*In\u00a0vitro* methylation\n----------------------\n\nThe \u2212156P *Gpat2* promoter construct (3\u00a0\u03bcg/reaction) was incubated with 160\u00a0\u03bcM *S*-adenosylmethionine (SAM) and 10 units of SssI methylase (CpG methyltransferase, M.SssI, New England Biolabs) ('methylated') or without enzyme ('unmethylated') for 4\u00a0h at 37\u00b0C, followed by a 20\u00a0min inactivation at 60\u00b0C. Plasmids were purified using the commercial kit Illustra GFX PCR DNA and Gel Band Purification Kit (GE Healthcare Life Sciences) and Lipofectamine 2000 (Invitrogen) was used to co-transfect with *Renilla* luciferase plasmid into CHO-K1 cells. The next day, cells were harvested and luminescence was measured using the Dual-Luciferase Reporter Assay System (Promega) and normalized to that of *Renilla*.\n\nGerm cell purification\n----------------------\n\nGerm cells were isolated from testes of mice at 11, 15 and 30 dpp as described in \\[[@B17]\\]. Cells were incubated overnight at 34\u00b0C in 4% CO~2~ to allow Sertoli cells to adhere to the culture plate surface. The germ-cell-enriched supernatant was removed, and centrifuged for 10\u00a0min at 800\u00a0***g*** at room temperature, and then genomic DNA was extracted from the pellet.\n\nBisulfite treatment and sequencing\n----------------------------------\n\nUsing DNeasy Blood & Tissue Kit (Qiagen), total DNA was purified from whole testis from 7 dpp mice and from germ cells isolated from mouse testis at selected ages. DNA purity and concentration were evaluated using a NanoDrop spectrophotometer (Thermo Scientific). For bisulfite treatment, 2\u00a0\u03bcg of genomic DNA was treated employing the EpiTect Bisulfite Kit (Qiagen). PCR was performed using Taq DNA polymerase (Invitrogen); the protocol was similar to a nested PCR, but using the same set of primers. For each reaction, 100\u00a0ng of bisulfite-converted DNA was used, and cycling conditions were: 2\u00a0min at 94\u00b0C for initial denaturation, followed by a touchdown step of eight cycles of 45\u00a0s at 94\u00b0C, 40\u00a0s at 64--56\u00b0C, 1.5\u00a0min at 72\u00b0C; 30 cycles of 45\u00a0s at 94\u00b0C, 40\u00a0s at 56\u00b0C, 1.5\u00a0min at 72\u00b0C and a final extension step of 6\u00a0min at 72\u00b0C. Samples of 5\u00a0\u03bcl of these PCR products were then used as templates with the same reaction conditions. Primers used were: forward: 5\u2032-ATTGGTTGGTTTTTTAGTTGTTGAG-3\u2032; reverse: 5\u2032-ATTCCACATCAATCCCTACCTAAC-3\u2032. The products of interest (197\u00a0bp) were gel purified with the Illustra GFX PCR DNA and Gel Band Purification Kit, ligated into pGEMT-Easy Vector (Promega) and transformed into *Escherichia coli* JM109 competent cells. Insertion was confirmed by restriction enzyme digestion, and eight clones were chosen for sequencing.\n\nMouse testis subcellular fractionation and immunoblotting\n---------------------------------------------------------\n\nTestes from 7, 15, 20 and 35 dpp BALB/c mice were removed, rinsed with ice-cold PBS and submerged in precooled buffer H \\[10\u00a0mM Hepes/KOH, pH\u00a07.4, 0.25\u00a0M sucrose, 1\u00a0mM EDTA, 1\u00a0mM DTT, 0.002% protease inhibitor cocktail (general use) 1:6 (w/v)\\]. Testes were homogenized with ten up-and-down strokes in a motor-driven Teflon-glass vessel. Large debris and nuclei were pelleted by centrifuging twice at 600\u00a0***g*** for 5\u00a0min. The supernatant (post-nuclear homogenate) was centrifuged for 10\u00a0min at 10000\u00a0***g***. The pellet was resuspended in 1\u00a0ml of buffer H/g of testis and homogenized in a Dounce tissue grinder with a glass pestle to obtain a mitochondria-enriched protein fraction.\n\nSamples of 100\u00a0\u03bcg of the mitochondria-enriched protein fractions were separated by SDS/PAGE (8 or 12% gel), transferred to a PVDF membrane (Bio-Rad Laboratories) and probed with a 1:500 dilution of anti-GPAT2 antibody \\[[@B14]\\] or with 1:2000 of the anti-voltage-dependent anion channel (VDAC) antibody (Affinity Bioreagents), as loading control. Membranes were then washed extensively and probed with 1:5000 dilution of horseradish peroxidase-conjugated goat anti-rabbit IgG antibody (Thermo-Pierce). For chemiluminescent detection, the membranes were incubated with Super Signal detection kit (Thermo-Pierce).\n\nRESULTS AND DISCUSSION\n======================\n\n*Gpat2* mRNA and protein are highly expressed in pachytene spermatocytes\n------------------------------------------------------------------------\n\nGPAT2 is expressed in male germ cells, particularly in primary spermatocytes, and its expression changes during rat sexual development \\[[@B14]\\]. The first round of spermatogenesis in mammals is characterized by the synchronized appearance of particular cell types\u00a0at different stages of the spermatogenic cycle among the seminiferous tubules, and this model was used to analyse differential gene expression during spermatogenesis \\[[@B18],[@B19]\\]. In mouse testis, only Sertoli cells and spermatogonia populate the seminiferous tubules until 10 dpp. Subsequently, meiotic cells appear in the tubules: at 11 dpp meiosis advances no further than the zygotene stage, at 13 dpp to the early pachytene, at 15 dpp to the middle pachytene, at 18 dpp to the late pachytene, at 20 dpp to the round spermatid stage and at 30 dpp spermatids have already reached the elongation phase and sperm tail accessory structures are being constructed. After this period, synchrony is lost and all stages of spermatogenesis can be observed; however, apart from mature spermatozoa, pachytene spermatocytes and round spermatids remain the most abundant germ cell types\u00a0in seminiferous tubules \\[[@B20]\\]. In order to study the relative expression of *Gpat2* mRNA during the first wave of spermatogenesis, total RNA was extracted from testis at selected ages and expression was assessed by qPCR ([Figure 1](#F1){ref-type=\"fig\"}A). The mRNA content was low at 3 and 7 dpp, and then increased, reaching a peak at 15 dpp, after which expression fell by 50% to a value that remained unchanged throughout adulthood. To evaluate mRNA expression in the seminiferous tubules*, in situ* mRNA hybridization was performed on slides of mouse testis at 7, 15 and 30 dpp ([Figure 1](#F1){ref-type=\"fig\"}C). The label was undetectable at 7 dpp, but positive staining was observed at 15 and 30 dpp. Strong labelling was observed at 15 dpp and most cells populating the tubules were positive for *Gpat2* mRNA, except for the basal germ cells corresponding to spermatogonia. In contrast, at 30 dpp the label was detected only in primary spermatocytes, and mRNA expression was abruptly extinguished in the subsequent stages of spermatogenesis. These results are consistent with the analysis of microarray data collected from mouse testis, from fetal age to adulthood (duplicates in 11, 12, 14, 16 and 18\u00a0days post-coitum (dpc), 0, 3, 6, 8, 10, 14, 18, 20, 30 and 35 dpp) which shows the time course of gene expression in murine testis development (GEO Datasets: GDS2098 and GSE12769). Postnatal *Gpat2* followed a similar profile to the one described in the present study ([Figure 1](#F1){ref-type=\"fig\"}B), and the microarray data demonstrated that *Gpat2* transcription is transiently active during embryonic development, reaching a peak at 15 dpc ([Figure 1](#F1){ref-type=\"fig\"}B).\n\n![*Gpat2* mRNA expression analysis during the first wave of spermatogenesis\\\n(**A**) *Gpat2* mRNA content was analysed in mouse testis at selected ages and the *Gpat*2 mRNA relative expression level for each sample was normalized to both *Gapdh* and *Rpl13A* mRNA expression level. (**B**) *In silico* analysis of pre- and post-natal *Gpat2* expression was performed using microarray information obtained from the public database GEO Profiles (NCBI). (**C**) mRNA expression was assayed by *in situ* hybridization performed on slides of 7, 15 and 30 dpp mouse testis. Arrows indicate positive reaction. Scale bar=50\u00a0\u03bcm.](bj4710211fig1){#F1}\n\nTo verify GPAT2 protein expression, Western blot analysis and immunohistochemistry (IHC) were performed. Western blotting showed that the protein expression peaked at day 20, consistent with the high transcriptional rate at 15 dpp and the moderate transcription level at 20 dpp ([Figure 2](#F2){ref-type=\"fig\"}A). IHC was also performed on mouse testes at 7, 15, 30 and 40 dpp ([Figure 2](#F2){ref-type=\"fig\"}B). Similar to the mRNA expression pattern, GPAT2 protein was undetectable in 7 dpp testis, whereas a positive cytoplasmic signal was detected in testes from 15, 30 and 40 dpp mice. Again, the distribution of the signal varied among the different germ cells within the seminiferous tubules. Strong staining was present in nearly all spermatocytes at 15 dpp (when pachytene spermatocytes predominate). This stain was localized to pachytene spermatocytes at 30 and 40 dpp and became more diffuse in later stages of spermatogenesis ([Figure 2](#F2){ref-type=\"fig\"}B). Thus, *Gpat2* mRNA expression and maximum GPAT2 protein content within adult testis were restricted to pachytene spermatocytes. This temporospatial profile suggests that GPAT2 plays a specific role in spermatogenesis and that strict transcriptional regulatory mechanisms operate to ensure an on--off expression pattern. The specific role for GPAT2\u00a0in spermatogenesis is suggested by its involvement in the metabolism of a class of small non-coding RNAs called piRNAs that are predominantly synthesized in germ cells; *Gpat2* silencing in germline stem cells severely impaired primary piRNA production. In addition, in both germline stem cells and testis, GPAT2 interacts with MILI protein, a key player in piRNA metabolism \\[[@B5]\\]. The expression of piRNAs in mouse testis occurs at two distinct phases: pre-pachytene piRNAs are present in fetal mice, are enriched in transposon sequences, and are synthesized by the primary and secondary ('ping-pong') pathways, whereas pachytene piRNAs originate from discrete genomic loci (piRNA clusters), and their synthesis is independent of the ping-pong cycle \\[[@B21],[@B22]\\]. This expression profile of *Gpat2* reinforces the idea that GPAT2 functions in the primary piRNA biosynthesis pathway, since *Gpat2* expression peaks not only when both pre-pachytene (embryonic) and pachytene piRNA are synthesized, but also occurs simultaneously with *Mili* transcription (GEO Datasets: GDS2098 and GSE12769).\n\n![GPAT2 protein expression analysis during the first wave of spermatogenesis\\\n(**A**) GPAT2 protein content was analysed in testis mitochondrial fractions extracted at selected ages and normalized to the expression of the mitochondrial VDAC. Results are expressed as the means for three independent experiments \u00b1 S.D. (**B**) IHC was assessed in 7, 15, 30 and 40 dpp mouse testis (\u00d7600 magnification). Arrowheads indicate the germ cells showing a strong signal at 30 and 40 dpp. Scale bar=50\u00a0\u03bcm. Sg, spermatogonia; PS, pachytene spermatocytes; RS, round spermatids; ES, elongating spermatids. Scale bar=50\u00a0\u03bcm.](bj4710211fig2){#F2}\n\nEvaluation of *Gpat2* minimal promoter\n--------------------------------------\n\nEpigenetic modifications operate with *trans*-acting factors to achieve transient gene expression and ensure the progression of spermatogenesis. To identify the promoter region and study the regulatory elements of the *Gpat2* gene, the 5\u2032 sequence flanking the transcription start site (TSS) (\u22121320 to translation start site) was analysed by online programs and applications, such as GPMiner, Alibaba 2.0, MethPrimer and Epigenomics (NCBI) to recognize promoter elements such as the TATA box, CCAAT box and GpC islands. The first intron was included because in other genes, transcription factor-binding sites or CpG islands are present in this region \\[[@B16],[@B23]\\]. The *Gpat2* promoter lacked a canonical TATA box and/or CAAT boxes in the 5\u2032-flanking region, but a CpG island containing ten CpG pairs was detected from \u221292 to +32. These features are characteristic of a 'CpG island promoter' class \\[[@B24]\\]. To establish the region of minimal promoter activity, serial deletions of the putative promoter region were made, cloned into a luciferase reporter vector and transiently transfected into CHO-K1 cells. Basal transcription promoter activity was located between \u2212156 and +1 position ([Figure 3](#F3){ref-type=\"fig\"}), a sequence that contains three potential binding sites for Sp1 (specificity protein 1, which can initiate transcription when the TATA box is absent \\[[@B24]\\]) and the CpG island described above. These constructs were used to evaluate the influence of CpG methylation or hormones on *Gpat2* promoter activity.\n\n![Basal promoter activity of the murine *Gpat2* gene\\\nSerial deletions of the putative promoter region were made, cloned into luciferase vector and transiently transfected into CHO-K1 cells. Relative luciferase activity was normalized to the *Renilla* luciferase internal control. Results are expressed as the means \u00b1 S.D. of replicates from two independent experiments.](bj4710211fig3){#F3}\n\nMethylation status of *Gpat2* promoter changes during the first wave of spermatogenesis\n---------------------------------------------------------------------------------------\n\nqPCR and *in situ* hybridization experiments during the first wave of spermatogenesis demonstrated that *Gpat2* mRNA transcription occurred transiently in meiotic cells; expression was higher in pachytene spermatocytes and abruptly decreased in subsequent stages of spermatogenesis. We previously speculated that ectopic GPAT2 expression in human cancer cells could be the consequence of epigenetic modifications \\[[@B13]\\]. Thus, because epigenetic events are critical in regulating spermatogenesis, we asked whether these mechanisms could regulate *Gpat2* transcription during mouse spermatogenesis. To answer this question, we first assayed the effect of chromatin modification reagents such as the DNA methyltransferase inhibitor DAC and the histone deacetylase inhibitor TSA on *Gpat2* transcription. The murine cell line RAW 264.7 expresses low basal levels of *Gpat2*, but DAC treatment induced *Gpat2* expression 2-fold (*P*\\< 0.01), whereas TSA treatment did not change mRNA expression. The combination of both TSA and DAC treatments enhanced *Gpat2* transcription almost 6-fold ([Figure 4](#F4){ref-type=\"fig\"}A), illustrating synergistic effects on *Gpat2* transcription. These experiments demonstrate that epigenetic changes such as hypomethylation, have a profound effect on the rate of *Gpat2* transcription and that the modifications on histone acetylation become important when methylation is low. This result is consistent with microarray data showing that mice null for histone deacetylase 2 highly up-regulate *Gpat2* expression in the heart, a tissue in which *Gpat2* is not constitutively expressed (GEO profile GDS2624/1456208_at/*Gpat2*). To determine whether methylation has the opposite effect on the *Gpat2* promoter, i.e. methylation is associated with *Gpat2* silencing, the construct \u2212156P containing the minimal promoter of *Gpat2* and the typical CpG island located between \u221292 and +32, was methylated *in\u00a0vitro* and luciferase activity was assayed ([Figure 4](#F4){ref-type=\"fig\"}B). Unmethylated construct activity was 20-fold higher than the internal *Renilla* luciferase control, whereas methylated construct activity was lower than the internal control (10% of the activity), suggesting that CpG methylation of the *Gpat2* minimal promoter prevents transcription. In summary, whereas DNA hypomethylation and histone acetylation are associated with high transcription rates, DNA methylation of the *Gpat2* promoter prevents transcription.\n\n![Epigenetic mechanisms involved in *Gpat2* transcription regulation\\\n(**A**) The effect of the DNA methyltransferase inhibitor DAC and the histone deacetylase inhibitor TSA on *Gpat2* transcription on the murine cell line RAW 264.7 was assessed by qPCR and *Gpat2* mRNA relative expression level for each sample was normalized to its *Gapdh* and *Rpl13A* mRNA expression level. Bars with different letters indicate statistical difference (*P*\\< 0.01, one-way ANOVA and Tukey\\'s test). (**B**) *In\u00a0vitro* methylation of the *Gpat2* minimal promoter construct (\u2212156P) was performed and both methylated and unmethylated constructs were transfected into CHO-K1 cells. Promoter activity was measured as luciferase activity relative to the internal control *Renilla* luciferase (\\*\\* *P*\\< 0.01, Student\\'s *t* test). (**C**) Total DNA from 7 dpp testes or from isolated germ cells from 11, 15 and 30 dpp mouse testes was purified and bisulfite-modified, and the methylation status of the CpG island located on the proximal promoter region was analysed after PCR amplification and sequencing. The top bar shows a schematic representation of the location of CG pairs along the fragment that was amplified: inverted triangles represent the CG pairs and the grey box represents the first exon. Average methylation status of each position among all the sequenced clones is represented by circles: a black filled circle indicates that all cytosines were methylated at that position among all the clones (100% methylated) and an empty circle indicates that all the cytosines were unmethylated at that position among all the clones (0% methylated). Partially black circles indicate intermediate percentages of clones with a methylated cytosine at that position. The average methylation for each position is shown for 7, 11 and 15 dpp, considering all of the clones sequenced in each case (five to ten clones). However, at 30 dpp 37% of the clones were almost fully unmethylated (10% or less methylated cytosines) and 63% were fully methylated (70% or more methylated cytosines). Thus, at this age, the average methylation of each position is shown separately for the unmethylated clones (37% of the clones) or the methylated ones (63% of the clones) to emphasize the coexistence of two populations of germ cells in terms of *Gpat2* promoter methylation status at 30 dpp.](bj4710211fig4){#F4}\n\nTo elucidate the methylation status of the *Gpat2* promoter *in\u00a0vivo*, total genomic DNA was purified from 7 dpp mouse testis and from germ cells isolated from mice at 11, 15 and 30 dpp. At 7 dpp all clones were fully methylated, at 11 dpp a dramatic hypomethylation occurred in all the CpG pairs within the island, and at 15 dpp all clones were remethylated ([Figure 4](#F4){ref-type=\"fig\"}C). Interestingly, at 30 dpp (when primary and secondary spermatocytes populate the tubules) the clones presented an on--off pattern: the CpG islands were either totally unmethylated (37.5% of the clones) or fully methylated (62.5% of the clones). To corroborate that promoter methylation is the mechanism by which *Gpat2* expression is repressed in other tissues, genomic DNA from liver was bisulfite-modified and the *Gpat2* CpG island was sequenced. As expected, this region was totally methylated (results not shown). Thus, hypomethylation occurs at a specific time that is consistent with the initiation of meiosis (11 dpp) and precedes the maximum mRNA content detected at 15 dpp. Taken together, our data demonstrate that hypomethylation of the CpG island present in the *Gpat2* proximal promoter increases gene expression. A similar correlation between methylation status of the proximal promoter and gene expression was also observed for another member of the piRNA pathway, the mouse *Miwi* gene \\[[@B25]\\], whose temporospatial expression from midpachytene spermatocytes to round spermatids was associated with hypomethylation of the CpG island proximal promoter. Besides the spermatogenic cells, PIWI genes are overexpressed in some cancer types\u00a0and this ectopic expression correlates with poorer clinical outcomes, suggesting that PIWI genes play a functional role in cancer biology \\[[@B26]\\]. Epigenetic mechanisms ensure the transient expression of the piRNA pathway genes at specific stages of spermatogenesis and their aberrant expression in somatic cancer cells \\[[@B27]\\]. Whether the expression of piRNAs and PIWI proteins in cancer is a consequence or a cause of malignancy remains unclear. However, it has been confirmed that the expression of piRNA pathway genes is essential for the progression of spermatogenesis, i.e. the disruption of *Miwi2* and *Mili* genes cause a meiotic-progression defect, loss of germ cells and sterility \\[[@B28],[@B29]\\]. In our previous work, we confirmed not only the presence of human *GPAT2* in the cancer-derived cell line MDA-MB-231, but also its epigenetic regulation and relevance for the development of the tumour characteristics, since the GPAT2 knockdown dramatically decreased cellular proliferation, migration and *in\u00a0vivo* tumorigenicity. In addition, on the basis of *in silico* information, high *GPAT2* expression occurs in human breast cancer, melanoma, lung cancer and prostate cancer \\[[@B13]\\].\n\n*Gpat2* gene transcription is up-regulated *in\u00a0vitro* by retinoic acid\n----------------------------------------------------------------------\n\nThe varying expression of GPAT2 protein in mouse testis motivated us to search for hormone-responsive elements in its promoter region, particularly for those related to male sexual development and function. The promoter region used for analysis was the one selected for the \u22121324P construct, and the online program Alibaba 2.1 \\[[@B30]\\] was used to predict transcription factor-binding sites. This program uses the TRANSFAC public database, which contains an extensive compilation of published binding site data for eukaryotic transcription factors, their experimentally proven binding sites and regulated genes. For this region the program predicted seven glucocorticoid-response elements, nine retinoid-response elements \\[both retinoid X receptor (RXR) and RA receptor (RAR)\\], three oestrogen-response elements and two progesterone-response elements. We transfected the \u22121324P construct into cell lines that express the receptor for the hormone that was being evaluated. We tested *Gpat2* promoter activity with oestradiol, corticosterone and ATRA/9cisRA in MCF7, CHO-K1 or TM4 cell lines, respectively. No response was observed with oestradiol or corticosterone at the tested concentrations (results not shown). However, positive regulation was observed with two active vitamin A metabolites, 9cisRA and ATRA ([Figures 5](#F5){ref-type=\"fig\"}A and [5](#F5){ref-type=\"fig\"}B). We observed that 8\u00a0\u03bcM 9cisRA and 10\u00a0\u03bcM ATRA duplicated \u22121324P activity ([Figures 5](#F5){ref-type=\"fig\"}A and [5](#F5){ref-type=\"fig\"}B), but no synergistic effect was observed when both retinoids were present (results not shown). All the *Gpat2* active promoter constructs were up-regulated by both isomers of RA. Because both basal activity and the RA response decreased downstream of the \u2212156P construct and because RA--retinoid receptor complexes usually interact with promoters near the transcription initiating site, it is possible that the predicted binding site for RA at +89\u00a0in the first intron might be involved in the *Gpat2* promoter retinoid response ([Figures 5](#F5){ref-type=\"fig\"}C and [5](#F5){ref-type=\"fig\"}D).\n\n![Retinoic acid up-regulates *Gpat2* promoter activity\\\nThe \u22121324P construct was transfected into TM4 cells and the effect of retinoids on *Gpat2* promoter activity was evaluated with 2, 4, 6 and 8\u00a0\u03bcM 9cisRA (**A**) and 2, 5, 10 and 20\u00a0\u03bcM ATRA (**B**). Both isomers significantly increased promoter activity (one-way ANOVA and Tukey\\'s test), and 8\u00a0\u03bcM 9cisRA and 10\u00a0\u03bcM ATRA doubled it. (**C** and **D**) The effect of 9cisRA and ATRA on *Gpat2* active promoter constructs was tested: all were significantly up-regulated by both isomers of RA (Student\\'s *t* test). Results represent means \u00b1 S.D. of the replicates from two independent experiments. \\*\\* *P*\\< 0.01, \\* *P*\\< 0.05. Luciferase activity is expressed relative to that of *Renilla* internal control.](bj4710211fig5){#F5}\n\nTaking into account that RA synthesized endogenously by premeiotic spermatocytes autonomously induces meiotic initiation by controlling the RAR-dependent expression of *Stra8* \\[[@B31]\\] and that *Gpat2* expression correlates with the onset of prophase I of meiosis \\[at 13.5 dpc in ovaries (GEO profile GDS2223/1456208_at/Gpat2) and at 15 dpp in testes ([Figure 1](#F1){ref-type=\"fig\"})\\] it could be speculated that RA may up-regulate *Gpat2* transcription *in\u00a0vivo*. Increasing evidence demonstrates that ATRA influences the epigenetic environment regulating both DNA demethylation and histone acethylation/deacethylation \\[[@B32]--[@B34]\\]. Thus, the up-regulation of *Gpat2* expression by RA might be a consequence of promoter demethylation, histone deacetylation and/or canonical *trans*-activation. Further experiments are required to validate this hypothesis.\n\n*Gpat2* expression is not induced when 3T3-L1 fibroblasts are differentiated to adipocytes\n------------------------------------------------------------------------------------------\n\nAlthough *Gpat2* was initially cloned by its sequence homology to *Gpat1*, an enzyme that promotes triacylglycerol biosynthesis from endogenous and exogenous acyl-CoAs, primarily in liver \\[[@B6]\\], the role of GPAT2 in lipid metabolism remains controversial. We previously reported that LPA, the product of the GPAT-catalysed reaction, could not be detected when GPAT2 was overexpressed \\[[@B14]\\]. Overexpression of murine *Gpat2* in CHO-K1 cells shows no increase in GPAT activity when glycerol 3-phosphate and palmitoyl-CoA, oleoyl-CoA or linoleoyl-CoA are used as substrates; in contrast, a 2-fold increase in a chloroform-soluble product is observed when arachidonoyl-CoA is used as a substrate, although this increase is due to phosphatidic acid with no accumulation of LPA \\[[@B14]\\]. Moreover, *Gpat2* overexpression increases acylglycerol-3-phosphate acyltransferase activity when both oleoyl-CoA and arachidonoyl-CoA are used as substrates \\[[@B14]\\]. These results strongly suggest that GPAT2 may have an alternative acyltransferase activity. Additionally, GPAT2 expression is low in liver and adipose tissue, and fasting and refeeding do not alter the expression of murine *Gpat2* \\[[@B10]\\]. In order to determine whether GPAT2 might play a role in lipogenesis, its expression was assessed in differentiating 3T3-L1 mouse fibroblasts. No *Gpat2* mRNA induction was detected after cells differentiated into adipocytes ([Figure 6](#F6){ref-type=\"fig\"}), whereas, as expected, cells contained numerous large lipid droplets and a high content of TAG (results not shown). This result is consistent with the fact that the *Gpat2* promoter lacks binding sites for factors that regulate lipogenesis, such as the ones belonging to sterol-regulatory-element-binding protein 1 (SREBP1), cholesterol-regulatory-element-binding protein (ChREBP) or the peroxisome-proliferator-activated receptor (PPAR) family and suggests that the primary function of GPAT2 may be unrelated to the synthesis of triacylglycerol.\n\n![*Gpat2* expression decreases during 3T3-L1 differentiation into adipocytes\\\n*Gpat2* mRNA content was evaluated by qPCR during 3T3-L1 mouse fibroblasts differentiation into adipocytes and the *Gpat2* mRNA relative expression level for each sample was normalized to its *Gapdh* and *Rpl13A* mRNA expression level. The black arrow indicates the day when the differentiation cocktail was added. *Gpat2* mRNA content decreased after cell differentiation.](bj4710211fig6){#F6}\n\nIn conclusion, the expression of GPAT2 during spermatogenesis is consistent with the requirement for piRNA action and its transient expression is determined primarily by epigenetic modifications and possibly by RA, a key regulator of meiosis entry.\n\nWe thank Mario Raul Ramos for the illustrations, Marianela Santana for her technical assistance and Dr Megumi Aita (UNC Molecular Neuroscience Core Facility) for her help with the *in situ* hybridization experiments.\n\nAUTHOR CONTRIBUTION\n===================\n\nMagali Pellon-Maison, Maria Gonzalez-Baro, Mauro Montanaro, Maria Garcia-Fabiani conceived and designed the experiments. Maria Garcia-Fabiani, Mauro Montanaro, Ezequiel Lacunza, Magali Pellon-Maison, Elizabeth Cattaneo performed the experiments. Maria Garcia-Fabiani, Maria Gonzalez-Baro, Magali Pellon-Maison, Mauro Montanaro, Ezequiel Lacunza analysed the data. Maria Gonzalez-Baro, Rosalind Coleman, Magali Pellon-Maison contributed reagents/materials/analysis tools. Magali Pellon-Maison, Maria Garcia-Fabiani, Maria Gonzalez-Baro, Rosalind Coleman, Mauro Montanaro wrote the paper.\n\nFUNDING\n=======\n\nThis work was supported by the National Institutes of Health (NIH) \\[grant numbers 1R03TW008698 (to M.R.G.-B.) and R01DK56598 (to R.A.C.)\\]; the Agencia Nacional de Promoci\u00f3n Cient\u00edfica y Tecnol\u00f3gica (ANPCyT) \\[grant number PICT 2246 (to M.R.G.-B.)\\]; the Consejo Nacional de Investigaciones Cient\u00edficas y T\u00e9cnicas (CONICET) \\[grant number PIP0310 (to M.R.G.-B.)\\]; the UNLP \\[grant number M168 (to M.R.G.-B.)\\]; and the Universidad Nacional de La Plata (UNLP) \\[grant number PPIDM004 (to M.P.-M.)\\]. M.A.M., E.L., M.P.-M. and M.R.G.-B. are members of the Carrera del Investigador Cient\u00edfico y Tecnol\u00f3gico and MBGF is a fellow of CONICET, Argentina. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.\n\nATRA\n\n:   all-*trans*-RA\n\nCHO\n\n:   Chinese-hamster ovary\n\n9cisRA\n\n:   9-*cis*-retinoic acid\n\nCTs\n\n:   cancer-testis genes\n\nDAC\n\n:   5-aza-2-deoxycitidine\n\nDMEM\n\n:   Dulbecco\\'s modified Eagle\\'s medium\n\ndpc\n\n:   days post-coitum\n\ndpp\n\n:   days post-partum\n\nfRMA\n\n:   frozen robust multi-array analysis\n\nGPAT\n\n:   glycerol-3-phosphate acyltransferase\n\nIHC\n\n:   immunohistochemistry\n\nLPA\n\n:   lysophosphatidic acid\n\npiRNA\n\n:   PIWI-interacting RNA\n\nqPCR\n\n:   quantitative real-time PCR\n\nRA\n\n:   retinoic acid\n\nRAR\n\n:   RA receptor\n\nTSA\n\n:   trichostatin A\n\nVDAC\n\n:   voltage-dependent anion channel\n"}
+{"text": "Background {#Sec1}\n==========\n\nAcute suppurative thyroiditis (AST) is a result of bacterial infection and represents a relatively rare condition in the thyroid gland. The thyroid gland is resistant to microbial infection, because of factors such as its encapsulation, iodine content, and rich blood supply \\[[@CR1], [@CR2]\\]. As a result, AST rarely develops in healthy individuals. Typically, AST is more likely to occur in children and in the left side of the neck. In 80% of patients with AST, the age at onset is before 10\u00a0years old (with 30% between birth and 2\u00a0years old), and only 8% occur in adulthood \\[[@CR3]\\]. The presence of a left piriform sinus fistula has been reported as important, as a potential route of infection \\[[@CR4]\\]. AST on the right side of thyroid gland in adults is thus rarely seen. We encountered a case of AST in the right lobe of the thyroid in a healthy woman. Moreover, AST developed against a background of papillary thyroid carcinoma (PTC).\n\nCase presentation {#Sec2}\n=================\n\nA 59-year-old woman was introduced to our hospital with a 4-week history of fever, sore throat, and swollen neck, after first visiting a primary-care physician and receiving antibiotics. She had no chronic diseases and appeared not to be to be in an immunocompromised state (she was well-nourished, had no diabetes and did not use any steroids, and human immunodeficiency virus (HIV) antibody was negative). On examination, a nodule showing pain, erythematous changes, and warmth was palpable in the anterior neck on the right side. The nodule showed limited mobility and no adjacent lymphadenopathy. The patient had no medical history of note, and no infectious symptoms such as cough, headache, or abdominal or joint pain were identified, other than the neck pain. Axillary temperature was 36.9\u00b0C, heart rate was 109 beats/min, and blood pressure was 169/88\u00a0mmHg. Table\u00a0[1](#Tab1){ref-type=\"table\"} shows the results of laboratory examination at the first visit (day X). Hematological tests revealed a high erythrocyte sedimentation rate (102\u00a0mm/hr) and elevated concentrations of C-reactive protein (10.4\u00a0mg/dL). A slight increase in thyroid-stimulating hormone (TSH) was also identified, showing subclinical hypothyroidism. Negative results were obtained for thyroglobulin antibody and serum thyroglobulin level was high (3590\u00a0ng/mL). Neck ultrasonography showed a 47-mm nodule, with ill-defined margins, irregular form, and a hypoechoic, heterogeneous appearance, in the area of the right thyroid, suggesting thyroid malignancy (Fig.\u00a0[1a](#Fig1){ref-type=\"fig\"}). Contrast-enhanced computed tomography (CT) confirmed a 37\u2009\u00d7\u200937\u2009\u00d7\u200942-mm nodule within the right thyroid lobe at the middle and lower pole. The thyroid mass resulted in displacement of the trachea toward the left and showed enhancement in the peripheral area of the nodule (Fig.\u00a0[1b](#Fig1){ref-type=\"fig\"}). No signs of metastasis were apparent, including in the lungs and bone. Fine needle aspiration (FNA) was performed on the first visit. Findings on cytological examination were suggestive of AST, because little nuclear atypia was evident, and numerous number of neutrophils were seen in the background (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). No PTC was apparent at that time. Based on the clinical course and cytology, AST was confirmed to be diagnosed and the patient was admitted to the endocrinology department.Table 1Results of laboratory testing at the time of first visit (day X)WBC7270/\u03bcL\u2003neutro67.3%\u2003eos0.3%\u2003baso0.4%\u2003mono6.6%\u2003lymph25.4%RBC458\u2009\u00d7\u200910^4^/\u03bcLHg14.2g/dLHct41.6%Plt24.5\u2009\u00d7\u200910^4^/\u03bcLESR102mmAlb4.2g/dLT-Bil1.0mg/dLAST22U/LALT17U/L\u03b3-GTP22U/LLDH233U/LBUN9.9mg/dLCrea0.52mg/dLCK71U/LCRP10.4mg/dLNa141mEq/LK4.2mEq/LCl105mEq/LFPG105mg/dLHbA1c6.1%HIV antibodynegativeFree T32.4pg/mLFree T41.0ng/dLTSH4.55\u03bcU/mLTPO Ab\\<\u20092.55IU/mLTg Ab\\<\u20096.12IU/mLTg3590ng/mL*ESR* erythrocyte sedimentation rate, *FPG* fasting plasma glucose, *HbA1c* hemoglobin A1c, *TSH* thyroid-stimulating hormone, *Tg* thyroglobulinFig. 1Ultrasonography on day X reveals a hypoechoic lesion with ill-defined margins and irregular form, appearing avascular and heterogeneous (**a**). Computed tomography of the neck on admission (day X\u2009+\u20095) also reveals a low-density lesion in the right thyroid gland, 37\u2009\u00d7\u200937\u2009\u00d7\u200942\u00a0mm in size with enhancement in the marginal area (**b**)Fig. 2Cytology from FNA shows scant nuclear atypia (**a**), with numerous neutrophils in the background (**b**). Four months later, cytology reveals overlapping cell clusters, high nuclear density, nuclear groove (\u25b3), and intranuclear cytoplasmic inclusion bodies (\u25b2), leading to a diagnosis of papillary thyroid carcinoma (**c**)\n\nThe clinical course is shown in Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}. Antibiotic therapy produced complete response, with rapid improvement within 1\u00a0week. Serum thyroglobulin level tended to decrease (528\u00a0ng/mL on day X\u2009+\u20098; 47.3\u00a0ng/mL on day X\u2009+\u200934). Blood cultures on day X\u2009+\u20096 yielded negative results. CRP concentration was just 0.71\u00a0mg/dL on day X\u2009+\u20098, and the patient was discharged with complete disappearance of symptoms. A barium swallow study was performed after CRP turned negative, but no fistula of the pyriform sinus was detected (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). Neck ultrasonography 4\u00a0months after the onset showed that the nodule in the right thyroid gland had shrunk to 27\u00a0mm in diameter, but still showed like malignancy. FNA was repeated, and cytological examination revealed overlapping cell clusters, high nuclear density, nuclear grooves, and intranuclear cytoplasmic inclusion bodies, leading to a diagnosis of PTC (Fig.\u00a0[2c](#Fig2){ref-type=\"fig\"}). The patient subsequently underwent total thyroidectomy and bilateral level D1 lymph node dissection. Histological examination revealed a 20-mm PTC in the right lobe with invasion to the sternothyroid muscle, and a 16-mm PTC in the left lobe (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). PTC in the left lobe was not found by ultrasonography before surgery. One possible reason is inhomogenous thyroid grand by adenomatous goiter, but it is unclear why left PTC was not found. Histological examination first detected it after the surgery. The right PTC was well-differentiated, but with large necrotic regions. The left PTC was close to the sternothyroid muscle. No lymph nodes contained metastatic PTC. Cord-like tissue considered as fistulous tract was not found to communicate with the right lobe and the hypopharynx. Pathological staging was pT3N0. Postoperatively, the patient received radioactive iodine ablation.Fig. 3Summary of the clinical courseFig. 4Barium swallow study (frontal view) does not show any fistula from the apex of the pyriform recessFig. 5Gross (**a** and **b**) and microscopic (**c** and **d**) appearance of papillary carcinoma of the thyroid\n\nDiscussion and conclusions {#Sec3}\n==========================\n\nThis represents a rare case of right-sided AST concomitant with PTC, found in an adult woman who did not appear immunocompromised and did not have any other foci of infection. In addition, we could not find any evidence of piriform sinus fistula, even after resolution of inflammation.\n\nSome reports have described concomitant AST and thyroid cancer. One case developed AST after FNA of a PTC, and was therefore considered as infection secondary to needle aspiration \\[[@CR5]\\]. Haddad and colleagues reported a case of AST in a patient with ischemic heart failure and type one diabetes mellitus \\[[@CR6]\\]. In another case, a pregnant woman who had given birth by Cesarean section was diagnosed with AST after thyroidectomy \\[[@CR7]\\]. The present case clearly differed from these cases in the lack of a clear cause of AST. The reasons for the AST occurring with malignant tumor in the present case remains unclear, as this woman showed no sign of infectious disease and was not immunocompromised, and no piriform sinus fistula was present. Because the AST and PTC showed identical locations, the PTC could easily be imagined to be infected with bacteria, but the mechanisms were not clear. A previous report \\[[@CR8]\\] suggested that an abnormal blood supply from the PTC could facilitate infection, and an abnormal blood supply from malignant tumor may thus have resulted in infection in our case.\n\nTakai et al. first proposed the presence of a left piriform sinus fistula as a route of infection \\[[@CR4]\\]. In the present case, however, we could not find any sign of a piriform sinus fistula. Repeated inflammation may result in adhesions within the fistula, and potentially masking the fistula in cases with repeated episodes of infection. However, this patient had presented with the first episode of inflammation, so fistula as the route of infection seems unlikely in our case. We attempted to culture bacteria from the thyroid nodule to provide insights into the source of infection, but the results were negative. Cultivation of bacteria probably failed because the patient had been given antibiotics for about 1\u00a0week before the sample was obtained for cultivation.\n\nIn 80% of AST patients, the age at onset is less than 10\u00a0years old, with 30% occurring between birth and 2\u00a0years old, and only 8% of cases occur in adulthood \\[[@CR3]\\]. An overview of 109 cases of AST reported that 85 patients had first experienced AST in childhood \\[[@CR9]\\]. Patients diagnosed with AST are likely to have experienced repeated inflammation of the neck, but our patient had no any past history of neck infections. The same overview also reported that 92% of patients showed left-sided infection, with bilateral infection in only 2% \\[[@CR9]\\]. Kingsbury described poor development of the right branchial arch in the prenatal period \\[[@CR10]\\]. Park reported that piriform fistula is more likely on the left than on the right, because the pharyngeal arch is drawn out to the nasal side during formation of the aortic arch by the left fourth branchial arch \\[[@CR11]\\], which is why AST mostly occurs in the left lobe of the thyroid. AST on the right side of thyroid gland in adults is rarely seen. Those cases of right-sided AST in adults that have been reported have shown backgrounds of infection such as infective endocarditis \\[[@CR8], [@CR12]\\] or miliary tuberculosis \\[[@CR13]\\], or immunocompromise due to steroid use or HIV infection \\[[@CR14], [@CR15]\\]. In our case, the patient was HIV-negative and did not have any other infection or underlying diseases and was thus not considered to be in a compromised condition. The cause of AST remains unknown, but the possibility of some involvement of the PTC must be considered.\n\nThis case was uncommon in terms of age and lesion location, which made differentiation of AST from malignancy difficult in the early phase. A paper by Lin reviewed 30 patients with malignant thyroid cancer who showed clinical features similar to AST \\[[@CR16]\\]. The significant characteristics of malignant thyroid tumor were clearly indicated as follows: 1) higher age at diagnosis (*P*\u2009=\u20090.0155); 2) presence of dysphonia (*P*\u2009=\u20090.0325); 3) right lobe involvement (*P*\u2009=\u20090.0151); 4) larger thyroid mass (*P*\u2009=\u20090.0013); 5) presence of anemia (*P*\u2009=\u20090.0075); and 6) sterile pus culture (*P*\u2009=\u20090.0013). Our case met 4 of these 6 clinical features suggestive of malignancy. The symptoms and signs of malignant thyroid tumor may mimic those of infectious thyroiditis, so we should be careful in diagnosing AST or aggressive malignant tumor. Long-term follow-up using both ultrasonography and FNA is also necessary. We obtained findings indicating PTC 5\u00a0months after identifying the presence of AST, on the third cytological examination. No findings even suggestive of PTC were evident from the first cytological examination, with relatively few variant epithelial cells and numerous leukocytes. One possibility was that we sampled a location where lymphocytes were gathered or that was necrotic.\n\nWe encountered a case of AST in the right lobe of a healthy woman. AST can develop with thyroid malignant tumor, so we have to take it into our account that malignant tumor may exist in the background when AST is identified on the right side of thyroid gland with a healthy subject.\n\nAST\n\n:   Acute suppurative thyroiditis\n\nESR\n\n:   Erythrocyte sedimentation rate\n\nFNA\n\n:   Fine needle aspiration\n\nFPG\n\n:   Fasting plasma glucose\n\nHbA1c\n\n:   Hemoglobin A1c\n\nPTC\n\n:   Papillary thyroid carcinoma\n\nTg\n\n:   Thyrogloblin\n\nTSH\n\n:   Thyroid-stimulating hormone\n\nAll data generated or analysed during this study are included in this published article.\n\nHO and MN mainly examined and determine how to treat the patient and were major contributors in writing the manuscript. SK and MM also examined the patients. MY, MY and TS gave many advices during treating the patients. TF, IM, NA and HK performed the patient's operation. TI, AA, NI and RM performed the histological examinations of the thyroid. All authors read and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar1}\n==========================================\n\nNot applicable.\n\nConsent for publication {#FPar2}\n=======================\n\nWritten informed consent was obtained from subjects for publication of this report. A copy of the written consent is available for review upon requests.\n\nCompeting interests {#FPar3}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar4}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "In recent years, tens of thousands of human long noncoding RNAs (lncRNAs) have been annotated by genome-wide signature analysis or multiple deep-sequencing technology[@b1][@b2][@b3]. With increased reports suggesting their functionality, it has been realized that lncRNAs are more than transactional 'noise\\'[@b4]. LncRNAs can bind to various proteins, DNAs or RNAs to form functional complexes[@b5][@b6], and participate in multiple cellular processes including epigenetic modification[@b7][@b8][@b9][@b10], transcriptional[@b11][@b12] and post-transcriptional regulation[@b13][@b14][@b15][@b16] of gene expression, signal transduction[@b17], transportation[@b18] and many others[@b19]. LncRNAs have been implicated in differential physiological and pathological processes, such as stemness maintenance[@b20], organ development[@b13], cancer progression[@b9], immune regulation[@b21] and pathogen infection[@b22][@b23].\n\nHIV-1 latent infection in resting CD4^+^ T lymphocytes is the major obstacle to eradicate the virus in patients after suppressive combination antiretroviral therapy (cART)[@b24][@b25][@b26]. Because of its presence, the life-long cART is required, otherwise the viremia will rapidly rebound. HIV-1 latency could be due to deficiency of transcriptional factors such as nuclear factor-kappa B (NF-kB) or nuclear factor of activated T-cells (NFAT)[@b27][@b28], lack of the viral accessory protein Tat[@b29] or condensed chromatin structure and epigenetic regulation[@b30][@b31][@b32][@b33]. In addition, a cluster of microRNAs including miR-28, miR-125b, miR-150, miR-223 and miR-382, which are enriched in resting CD4^+^ T lymphocytes, target the 3\u2032-untranslated repeat of HIV-1 mRNA to restrict viral gene expression, and subsequently contribute to HIV-1 latency[@b34].\n\nHere we investigated the role of lncRNAs in HIV-1 latent infection, and have identified a lncRNA NRON, which potentially contributes to HIV-1 latent infection. NRON RNA significantly inhibits the HIV-1 promoter activity and decreased the cellular abundance of viral transactivator protein Tat. Further mechanism studies suggest that Tat protein is linked to the ubiquitin/proteasome components CUL4B and PSMD11 by NRON, and then degraded. Depletion of NRON RNA, especially in combination with a latency-reverse agent SAHA, robustly reactivated the latently infected proviruses both in an *in vitro* latency model and primary resting CD4^+^ T lymphocytes isolated from infected individuals receiving suppressive cART. Our data indicated that NRON plays a role in HIV-1 latency maintenance, and it could be a new target for reversing viral latency.\n\nResults\n=======\n\nLncRNA NRON represses HIV-1 replication\n---------------------------------------\n\nTo determine whether any lncRNA was involved in HIV-1 replication and latency, we selected several lncRNAs that were previously reported to be highly expressed in human immune system including the thymus[@b18], lymph nodes and white blood cells[@b35], and examined their expression in human primary CD4^+^ T lymphocytes with real-time quantitative reverse transcriptase--PCR (qRT--PCR). Some lncRNAs were highly expressed in comparison with the mRNA level of the control gene *T-bet*, which was reported to be expressed in CD4^+^ T lymphocytes[@b36] ([Fig. 1a](#f1){ref-type=\"fig\"}). A few of the abundant lncRNAs showed significantly higher expression levels in resting CD4^+^ T lymphocytes than those in activated cells ([Fig. 1b](#f1){ref-type=\"fig\"}). On screening with a group of lncRNA-specific short interfering RNA (siRNA) pools of two or three siRNAs for one target lncRNA to reduce off-target effects, we found that several lncRNAs including NRON and TUG1 affected the replication of HIV-1 (Fig. 1c and [Supplementary Fig. 1](#S1){ref-type=\"supplementary-material\"}). The depletion of NRON, which exerted a definitely higher expression level in resting CD4^+^ T lymphocytes through Northern blotting detection ([Fig. 1d](#f1){ref-type=\"fig\"}), significantly enhanced the viral replication in a time-course study ([Fig. 1e](#f1){ref-type=\"fig\"}).\n\nNRON specifically represses HIV-1 transcription\n-----------------------------------------------\n\nTo investigate which stage of HIV-1 replication was likely affected by NRON, we employed the defective HIV-1 provirus plasmid pNL4-3-deltaE-EGFP for transfection assay in HEK293T cells[@b37], which also express endogenous NRON RNA at a high level ([Supplementary Fig. 2a](#S1){ref-type=\"supplementary-material\"}). The intracellular Gag proteins expression obviously increased when NRON was depleted by siRNAs ([Fig. 2a](#f2){ref-type=\"fig\"}). Conversely, Gag proteins expression decreased when NRON was overexpressed in HeLa cells, which showed a low expression level of endogenous NRON ([Supplementary Fig. 2a](#S1){ref-type=\"supplementary-material\"}). Furthermore, the expression level of HIV-1 total mRNA in HEK293T cells showed a significant increase when NRON was knocked down ([Fig. 2b](#f2){ref-type=\"fig\"}), implying that the viral transcription or viral RNA stability could be affected by NRON. As many lncRNAs can regulate gene transcription, we employed a reporter luciferase system for HIV-1 promoter activity to examine whether NRON affected HIV-1 transcription. The HIV-1 promoter activity significantly increased after NRON was depleted in HEK293T cells ([Fig. 2c](#f2){ref-type=\"fig\"}), and was greatly reduced when NRON was overexpressed in TZM-bl cells, which also showed a low expression level of endogenous NRON ([Supplementary Fig. 2a,c](#S1){ref-type=\"supplementary-material\"}). To determine whether NRON suppressed the transcription of other viral promoters, we employed reporter luciferase systems for cytomegalovirus (CMV), Moloney murine leukemia virus (MMLV) and Rous sarcoma virus (RSV) promoters. Only slight changes were observed in these promoter reporter systems when NRON was knocked down ([Supplementary Fig. 2d--f](#S1){ref-type=\"supplementary-material\"}), indicating that NRON could specifically suppress HIV-1 promoter activity.\n\nNRON induces Tat protein degradation\n------------------------------------\n\nOn the basis of the above observations, we next sought to identify the targeting site(s) of NRON on HIV-1 promoter. NRON has been reported as a negative regulator of NFAT signalling pathway[@b18]. As HIV-1 harbours two sets of canonical NFAT-binding site in its 5\u2032 long terminal repeat (LTR), which can enhance the viral transcription[@b38][@b39][@b40], we constructed a HIV-1 promoter reporter plasmid with mutations at both these sites. Dual-luciferase reporter assay showed that the activity of NFAT-binding site-mutated HIV-1 promoter still increased on the knockdown of NRON in HEK293T cells ([Supplementary Fig. 3a](#S1){ref-type=\"supplementary-material\"}). Further, it has been shown that NFAT signal activity in HEK293T cells was relatively weak without stimulation by phorbol myristate acetate/ionomycin[@b18], which is consistent with latently infected resting CD4^+^ T lymphocytes. Therefore, our data indicated that NRON could regulate HIV-1 transcription through a NFAT-independent pathway in our HEK293T cell line monitor system.\n\nThe intact HIV-1 5\u2032 LTR promoter contains four functional regions, namely the modulatory region (\u2212454 to \u2212105 nucleotide (nt) relative to the transcription start site), the enhancer region (\u2212104 to \u221279\u2009nt), the basal promoter (\u221278 to \u22121\u2009nt) and TAR element (+1 to + 60\u2009nt)[@b41]. We constructed mutated HIV-1 promoters with deletions of the modulatory region and the enhancer region, or the TAR element ([Supplementary Table 1](#S1){ref-type=\"supplementary-material\"}). Although the promoter activity was significantly decreased because of these mutations, the core HIV-1 promoter with only the basal promoter and the TAR element could still be regulated by NRON ([Supplementary Fig. 3b](#S1){ref-type=\"supplementary-material\"}). However, the TAR element-deficient promoter did not exhibit this feature on modification ([Supplementary Fig. 3c](#S1){ref-type=\"supplementary-material\"}). It is well known that the TAR element recruits Tat, which then exploits CDK9 and Cyclin-T1 and plays a key role in transcription elongation[@b42][@b43][@b44]. The specific U-residue-rich bulge (+23 to +25\u2009nt) in the stem of TAR RNA is required for Tat binding, and mutation of the U residues in the bulge will significantly reduce its affinity to Tat[@b45][@b46][@b47]. When we substituted the U residues in the bulge with G residues in the TAR region to disrupt the Tat--TAR interaction, dual-luciferase assay showed that the mutated promoter was no longer affected by NRON knockdown ([Supplementary Fig. 3d](#S1){ref-type=\"supplementary-material\"}). We then investigated whether NRON was still functional in the absence of Tat. We indeed found that, in the absence of Tat-expressing plasmid, the regulation of NRON on HIV-1 promoter activity was no longer significant ([Supplementary Fig. 3e](#S1){ref-type=\"supplementary-material\"}). All these data indicated that the Tat--TAR axis was involved in NRON function.\n\nSince Tat is a RNA-binding protein, we initially hypothesized that NRON could bind to Tat protein and disrupt its recruitment to TAR RNA, thus functioning as a competing endogenous RNA[@b14]. RNA co-immunoprecipitation (RNA co-IP, RIP) and real-time qRT--PCR data showed that the enrichment of NRON by Tat was more than 15-fold increase compared with the green fluorescent protein (GFP) control in NRON-overexpressed TZM-bl cells ([Fig. 2d](#f2){ref-type=\"fig\"}). In addition, Tat could also potently enrich the endogenously expressed NRON in HEK293T cells ([Supplementary Fig. 4a](#S1){ref-type=\"supplementary-material\"}), indicating that Tat binds to NRON in an HIV-1 promoter-independent manner. We then confirmed that the association between Tat and TAR-Luc RNA reduced on NRON overexpression in TZM-bl cells ([Fig. 2e](#f2){ref-type=\"fig\"}). Unexpectedly and interestingly, Tat protein abundance, initially detected as the input control by western blotting for RNA co-IP assay, significantly decreased on NRON overexpression in TZM-bl cells ([Fig. 2e](#f2){ref-type=\"fig\"}), while the Tat mRNA level remained unchanged ([Supplementary Fig. 4b](#S1){ref-type=\"supplementary-material\"}). Conversely, when NRON was depleted in HEK293T cells, the Tat protein level increased. However, the control GFP protein level showed no significant changes ([Fig. 2f](#f2){ref-type=\"fig\"}). Meanwhile, knockdown of NRON RNA did not affect the protein level of other HIV-1 proteins, such as Nef and Vpr ([Supplementary Fig. 4c](#S1){ref-type=\"supplementary-material\"}). Collectively, these data suggested that NRON RNA specifically induced the reduction of Tat at the protein level.\n\nNRON forms a complex with CUL4B and PSMD11\n------------------------------------------\n\nBecause of the findings thus far, we started to notice that three of the reported NRON-binding proteins, CUL4B, PSMD11 and HUWE1 (UREB1), belong to the ubiquitin/proteasome system[@b18]. We speculated that these proteins were involved in the process of NRON-mediated Tat degradation. To examine this hypothesis, we first investigated the roles of these proteins in HIV-1 transcription by a loss-of-function assay ([Supplementary Fig. 4d](#S1){ref-type=\"supplementary-material\"}). The transcription repression of HIV-1 promoter by NRON overexpression in TZM-bl cells was counteracted by CLU4B or PSMD11 knockdown, but not by HUWE1 ([Fig. 3a](#f3){ref-type=\"fig\"}). Correspondingly, the depletion of these proteins remarkably reversed Tat reduction caused by NRON overexpression in TZM-bl cells ([Fig. 3b](#f3){ref-type=\"fig\"}). Meanwhile, NRON RNA could be enriched by CUL4B ([Fig. 3c](#f3){ref-type=\"fig\"}) or PSMD11 ([Fig. 3d](#f3){ref-type=\"fig\"}), indicating that they were associated with NRON RNA robustly. Further, we found that Tat interacted with CUL4B or PSMD11 by co-IP; however, the interaction significantly decreased in the presence of RNase A ([Supplementary Fig. 4e](#S1){ref-type=\"supplementary-material\"}). When NRON was knocked down in the HEK293T cells, the interactions between Tat and CUL4B or PSMD11 were significantly diminished ([Supplementary Fig. 4f](#S1){ref-type=\"supplementary-material\"}). Moreover, the *in vitro* ubiquitination assay and the subsequent western blotting detection showed that Tat ubiquitination was significantly decreased when NRON was depleted ([Fig. 3e](#f3){ref-type=\"fig\"}). The same result was observed when CUL4B or PSMD11 was knocked down ([Fig. 3f](#f3){ref-type=\"fig\"}). Collectively, these data suggested that NRON mediated Tat degradation by bridging Tat and the components of ubiquitin/proteasome, and by facilitating the ubiquitination of Tat protein.\n\nNRON participates in maintaining HIV-1 latency\n----------------------------------------------\n\nSince NRON showed higher expression level in resting CD4^+^ T lymphocytes than in activated cells, we sought to determine whether this lncRNA functions in HIV-1 latent infection. Initially, we confirmed whether NRON could regulate HIV-1 transcription in primary resting CD4^+^ T lymphocytes. HIV-1 promoter reporter system plasmids, Tat-expressing plasmids, control siRNA or NRON siRNA were nucleofected into resting CD4^+^ T lymphocytes by electroporation. Dual-luciferase reporter assay showed that the activity of HIV-1 promoter significantly increased on NRON knockdown ([Fig. 4a](#f4){ref-type=\"fig\"}), which suggested that NRON RNA suppressed HIV-1 transcription in resting CD4^+^ T lymphocytes. We then examined the effect on Tat protein in resting NRON-depleted CD4^+^ T lymphocytes. Western blotting showed that the Tat protein level significantly increased on NRON knockdown ([Fig. 4b](#f4){ref-type=\"fig\"}). These data indicated that NRON RNA could suppress HIV-1 transcription by inducing Tat protein degradation in resting CD4^+^ T lymphocytes, which were the major population of HIV-1-latently infected cells.\n\nFurther, we generated a modified primary CD4^+^ T lymphocyte-based latency model[@b48][@b49]. A *bcl-2* open reading frame (ORF) was properly inserted into the *nef* ORF in the defective HIV-1 provirus plasmid pNL4-3-deltaE-EGFP to express Bcl-2 protein for promoting cell survival ([Supplementary Fig. 5a,b](#S1){ref-type=\"supplementary-material\"}). The resulting HIV/VSV (vesicular stomatitis virus)-pseudotyped viruses were used to infect the activated human primary CD4^+^ T lymphocytes. The GFP+ cells were enriched and expanded, and then cultured for \u223c1 month to allow the cells to convert to resting status ([Supplementary Fig. 5d,e](#S1){ref-type=\"supplementary-material\"}). When we transfected the cells with NRON-specific siRNAs, the viruses in the latently infected cells were significantly reactivated ([Fig. 4c,d](#f4){ref-type=\"fig\"} and [Supplementary Fig. 5f](#S1){ref-type=\"supplementary-material\"}). Furthermore, after transfection with NRON-specific siRNAs supplemented with the treatment of suberoylanilide hydroxamic acid (SAHA), a well-known histone deacetylase (HDAC) inhibitor and latency-reversing agent (LRA)[@b50], the reactivation level was much higher than that treated with nonspecific siRNA in combination with SAHA ([Fig. 4c,d](#f4){ref-type=\"fig\"}). Moreover, we examined the effect of knockdown of NRON upon HIV-1 latency in resting CD4^+^ T lymphocytes directly isolated from HIV-1-infected individuals receiving suppressive cART. Alu-PCR was employed to confirm that the cells harboured proviral HIV-1 DNA ([Supplementary Fig. 5g](#S1){ref-type=\"supplementary-material\"})[@b34]. After transfection of the resting cells with NRON-specific siRNAs in combination with SAHA treatment, the virion-associated RNA in the culture supernatant was significantly higher than that treated with nonspecific siRNA in combination with SAHA ([Fig. 4e](#f4){ref-type=\"fig\"} and [Supplementary Fig. 5h](#S1){ref-type=\"supplementary-material\"}). No significant change was observed by NRON knockdown only ([Supplementary Fig. 5i](#S1){ref-type=\"supplementary-material\"}), which may be because of the fact that the proviruses were more difficult to be reactivated from the primary resting CD4^+^ T lymphocyte directly isolated from HIV-1-infected individuals receiving suppressive cART than that from the HIV-1-latently infected primary CD4^+^ T lymphocytes, which were experimentally generated.\n\nWe then directly isolated resting CD4^+^ T lymphocytes from HIV-1-infected individuals receiving suppressive cART, and detected the intracellular HIV-1 proviral DNA, the expression level of HIV-1 RNA and NRON RNA. We found that the intracellular HIV-1 RNA expression, after being normalized with the cell-associated viral DNA, was inversely correlated with the NRON RNA expression level ([Fig. 4f](#f4){ref-type=\"fig\"}). Collectively, these data indicated that NRON contributed to HIV-1-latent infection. The abundance of NRON in resting CD4^+^ T lymphocytes efficiently blocks the expression of viral proteins and could, therefore, allow the infected cells to evade from the immune surveillance *in vivo*.\n\nDiscussion\n==========\n\nBecause of their essential regulatory roles in various cellular molecular networks, it is easy to understand why lncRNAs have been implicated in viral infection and immune regulation. A recent report showed that the lncRNA NEAT1 possessed an unspecific effect on the exporting of many mRNAs including HIV-1 RNA[@b51]. The NRON lncRNA was previously implicated in the severity of inflammatory bowel disease through its inhibitory effects on NFAT signalling[@b18][@b52]. As HIV-1 harbours NFAT-binding sites in its 5\u2032 LTR promoter region[@b38][@b39][@b40], it is highly likely that NRON may suppress HIV-1 replication in activated primary CD4^+^ T lymphocytes through inhibiting NFAT signal, which was described in a recent report by others[@b53]. However, in latently infected resting CD4^+^ T lymphocytes, NFAT signal activity is relatively weak without stimulation, which is consistent with our HEK293T and HeLa cell line monitor system[@b18][@b28]. In these systems, the mutation of NFAT-binding sites in the promoter did not affect the inhibitory effect of NRON on HIV-1 transcription ([Supplementary Fig. 3a](#S1){ref-type=\"supplementary-material\"}). In addition, knockdown of NRON could significantly enhance the viral production from latently infected resting CD4^+^ T lymphocytes without reactivating the cells ([Fig. 4](#f4){ref-type=\"fig\"}). These data support our hypothesis that NRON RNA could potently inhibit HIV-1 transcription and play a role in maintaining HIV-1 latency in an NFAT-independent mechanism in resting cells.\n\nOur work further indicates that, by forming a complex with CUL4B and PSMD11, NRON lncRNA recruits the essential HIV-1 regulatory protein Tat to the ubiquitin/proteasome system and induces its protein degradation. In this scenario, NRON functions as an adaptor between the ubiquitin/protease system and its target protein. We have found that the induced degradation of viral protein by NRON is of specificity, as NRON does not affect the stability of Nef and Vpr proteins. Tat contains several lysine amino-acid residues, and the ubiquitination of K71 enhances its transactivation activity, but not degradation of the protein[@b54]. The effect of NRON complex-induced Tat ubiquitination therefore could be due to ubiquitination at other lysine residues. More detailed studies are needed to identify the entire molecular mechanism. Further, it is notable that NRON in human has one spliced transcript, whereas three spliced transcripts of NRON have been found in mice. It remains to be determined whether NRON also induced the degradation of other proteins in human or mouse cells. Conversely, the specificity of the NRON--CUL4B/PSMD11 interaction and whether other components of the ubiquitin/proteasome system are involved in this process remains to be determined in future studies. As lncRNA HOTAIR could also be an adaptor between the ubiquitin system and its target proteins Ataxin-1 and Snurportin-1 (ref. [@b55]), it may be interesting to systematically investigate whether other lncRNAs could mediate the specific degradation of their binding proteins and the molecular mechanisms underlying these processes.\n\nBecause of latency, productive viral replication does not occur in resting primary CD4^+^ T lymphocytes of HIV-1-infected individuals receiving suppressive cART. However, replication-competent proviral DNA and multiply spliced or unspliced viral RNA can easily be found in these cells[@b56][@b57][@b58][@b59][@b60]. The deficiency of Tat protein is the direct consequence of several latency mechanisms such as the deficiency of transcriptional factors such as NF-kB or NFAT, condensed chromatin structure and epigenetic regulation, as well as the enrichment of suppressive cellular microRNAs[@b27][@b28][@b30][@b31][@b32][@b33][@b34]. It has been well known that the Tat--TAR interaction leads to the recruitment of the pTEFb complex including CDK9/cyclin-T1, which hyperphosphorylates of the C-terminal domain of RNA polymerase II and significantly increases transcription efficiency[@b42][@b43][@b44]. Recent studies have demonstrated that this positive-feedback regulation strongly controls HIV-1 latency, and Tat induction could robustly activate latently infected viruses[@b61]. Our work suggests that, by specifically inducing the degradation of Tat protein, NRON at high concentration in the resting CD4^+^ T lymphocytes prevents the accumulation of Tat and therefore acts as a barrier for the transcriptional activation of proviruses. The endogenous NRON expression level in the resting CD4^+^ T lymphocytes of HIV-1-infected individuals is inversely correlated with the intracellular viral RNA expression level, further supporting this hypothesis. Taken together, we propose a novel mechanism contributing to HIV-1 latency maintenance and also a new target for the development of LRAs[@b62].\n\nMethods\n=======\n\nEthics statement\n----------------\n\nThis research was approved by the Ethics Review Board of Sun Yat-Sen University and Guangzhou 8th People\\'s Hospital. Written informed consent was provided by all study participants.\n\nPlasmids and constructs\n-----------------------\n\nHIV-1 infectious clone pNL4-3 and its derivative pNL4-3-deltaE-EGFP were obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH (refs [@b37], [@b63]). The pHIV-Pro-Luc reporter plasmid was constructed by replacing the CMV promoter of the *luciferase* gene in the pMIR-REPORT Luciferase vector (Ambion) with the 5\u2032 LTR sequence of HIV-1~NL4-3~. Mutations or deletions were introduced into the promoter by PCR-based strategies, and the primer pairs were listed in [Supplementary Table 1](#S1){ref-type=\"supplementary-material\"}. The *bcl-2* gene was chemically synthesized and inserted into the *nef* region of pNL4-3-deltaE-EGFP provirus plasmid at the XhoI cutting site. Its reading frame is consistent with the *nef* ORF.\n\nThe two exons of HIV-1 full-length *tat* with an haemagglutinin (HA) epitope tag at its exon-2 3\u2032 terminus were PCR-amplified from pNL4-3 clone. After connection of two fragments with overlapping PCR, it was inserted into the pcDNA3.1 vector. Human full-length 2,735-bp *NRON* was amplified using PCR with total cDNA of human peripheral blood mononuclear cells (PBMCs) as the template. Human *CUL4B* and *PSMD11* with a FLAG epitope tag sequence at their 3\u2032 terminus was amplified using PCR with total cDNA of human PBMCs as the template. *NRON*, tagged *CUL4B*, or *PSMD11* were then inserted into the pcDNA3.1 vector. All constructs were verified by DNA sequencing. The *gfp* coding sequence (CDS) was tagged with HA or FLAG tag at the C terminus and was constructed into the pcDNA3.1 vector[@b64]. The N-terminus FLAG-tagged human ubiquitin B was constructed into the pcDNA3.1 vector[@b65]. The renilla luciferase-expressing plasmid pRL-CMV was obtained from Promega as a transfection normalization reporter for dual-luciferase reporter assay.\n\nCells and transfection\n----------------------\n\nTZM-bl cells, which harbour an HIV-1 promoter-driven luciferase gene, were obtained from AIDS Reference Reagent Program, NIH. Human HEK293T (American Type Culture Collection (ATCC)), HeLa (ATCC) and TZM-bl cells were maintained in DMEM (Hyclone) supplemented with 10% fetal bovine serum (Invitrogen), 100\u2009units\u2009ml^\u22121^ of penicillin and 100\u2009\u03bcg\u2009ml^\u22121^ of streptomycin at 37\u2009\u00b0C. All the cell lines had been tested for mycoplasma using a PCR assay and were mycoplasma-free. The HEK293T, HeLa and TZM-bl cells were transfected using Lipofectamine 2000 (Invitrogen) for plasmids and siRNAs by following the manufacturer\\'s instructions. The cells were collected at 48\u2009h post transfection for dual-luciferase reporter assays and protein detection. Human primary CD4^+^ T lymphocytes were nucleofected with Amaxa Nucleofector using the Amaxa Human T Cell Nucleofector Kit according to the manufacturer\\'s instructions.\n\nPurification of human primary CD4^+^ T lymphocytes\n--------------------------------------------------\n\nThe PBMCs were isolated from healthy human donors through Ficoll gradient centrifugation, followed by culturing in the conditioned RPMI 1640 medium. Human primary CD4^+^ T lymphocytes were then purified with a human CD4^+^ T-cell isolation kit according to the manufacturer\\'s instructions (BD Biosciences). The isolated human primary CD4^+^ T lymphocytes were then maintained in the conditioned RPMI 1640 medium (Hyclone) and stimulated with phytohaemagglutinin (5\u2009ng\u2009ml^\u22121^, Roche Applied Science) and interleukin-2 (IL-2, 10\u2009ng\u2009ml^\u22121^, R&D Systems), or anti-CD3 antibody (1\u2009\u03bcg\u2009ml^\u22121^, BD Biosciences), anti-CD28 antibody (1\u2009\u03bcg\u2009ml^\u22121^, BD Biosciences) and IL-2 (10\u2009ng\u2009ml^\u22121^, R&D Systems) for 48\u2009h. Then, the cells were washed three times with PBS buffer, and were cultured in the presence of IL-2 (10\u2009ng\u2009ml^\u22121^).\n\nHIV-1 infection\n---------------\n\nHEK293T cells at 3 \u00d7 10^6^ per dish were plated on 100-mm cell culture dish. Twenty-four hours later, the cells were transfected with 10\u2009\u03bcg of pNL4-3 plasmids with Lipofectamine 2000 (Invitrogen) according to the manufacturer\\'s instructions. Cell supernatants were harvested at 48\u2009h post transfection and were stored at \u221280\u2009\u00b0C. To normalize viral inputs, the amount of p24 was measured using an HIV-1 p24 ELISA kit according to the manufacturer\\'s instructions (Clonetech). The activated primary CD4^+^ T lymphocytes were seeded into 12-well cell-culture plates (2 \u00d7 10^6\u2009^ml^\u22121^, 1\u2009ml per well) and were transfected with 200\u2009pmol lncRNA-specific siRNAs or nonspecific control with Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer\\'s instructions. Eight to twelve hours later, the cells were infected with the equivalent of 10\u2009ng HIV-1 p24 antigen per well for 3\u2009h at 37\u2009\u00b0C. Then, the supernatants were removed and the cells were washed three times with fresh PBS buffer. The cells were maintained in the conditioned RPMI 1640 medium supplemented with IL-2 (10\u2009ng\u2009ml^\u22121^). The culture supernatants were collected on day 7 post infection. For the time-course study of the viral replication, at 7 days post infection, the cells were transfected with NRON siRNAs or nonspecific control for the second time. The culture supernatants were collected at 12\u2009h (day 0), day 3, day 7, day 10 or day 14 post infection. All of the culture supernatants were detected using the HIV-1 p24 ELISA kit according to the manufacturer\\'s instructions (Clonetech).\n\nThe chemical synthesis of siRNAs\n--------------------------------\n\nTwo or three specific siRNAs for one target lncRNA or protein were pooled together to reduce off-target effects. The chemically synthesized siRNAs and nonspecific control were purchased from RiboBio (Guangzhou). The target sequences of siRNAs were listed in [Supplementary Table 2](#S1){ref-type=\"supplementary-material\"}.\n\nLuciferase assay\n----------------\n\nTwenty-four hours before transfection, HEK293T cells (2 \u00d7 10^4^ cells per well) or TZM-bl cells (2.5 \u00d7 10^4^ cells per well) were seeded into 48-well plates. To perform the reporter assay for wild-type or mutated HIV-1 promoter activities, 2\u2009ng of reporter plasmids, 0.2\u2009ng of pRL-CMV and 2\u2009ng of pcDNA3.1-Tat-HA were co-transfected with 30\u2009pmol of siRNAs or nonspecific control into HEK293T cells. In all, 1\u2009ng of pRL-CMV, 5\u2009ng of pcDNA3.1-Tat-HA, 40\u2009ng of pcDNA3.1-NRON and/or 30\u2009pmol of siRNAs were co-transfected into TZM-bl cells. To perform the reporter assay for CMV, MMLV or RSV promoter activities, 2\u2009ng of pMIR-Reporter or 10\u2009ng of MMLV or RSV promoter reporter plasmids were transfected with 30\u2009pmol of siRNAs or nonspecific control into HEK293T cells. Human primary CD4^+^ T lymphocytes (3--5 \u00d7 10^6^ cells per well) from healthy donors were nucleofected with 2\u2009\u03bcg of HIV-1 promoter reporter plasmid, 200\u2009ng of pRL-CMV, 2\u2009\u03bcg of pcDNA3.1-Tat-HA and 100\u2009pmol of siRNAs. Dual-luciferase reporter assay was performed at 48\u2009h post transfection using the Promega Dual-Luciferase Reporter Assay System according to the manufacturer\\'s instructions[@b66][@b67].\n\nRNA isolation and real-time qRT--PCR\n------------------------------------\n\nRNA was isolated with TRIZOL reagent according to the manufacturer\\'s instructions (Ambion), and was treated with RQ-1 DNase (Promega) before reverse transcription. Reverse transcription reactions were performed with the PrimeScript RT reagent Kit (TaKaRa). Quantitative PCR was performed with the SYBR Premix ExTaq Kit (TaKaRa) on a CFX96 Real-Time System (Bio-Rad) by following the manufacturer\\'s instructions. Primers for real-time qRT--PCR were listed in [Supplementary Table 3](#S1){ref-type=\"supplementary-material\"}. Human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or \u03b2-actin mRNA was measured as endogenous controls. For RIP--qPCRs, the amount of a target RNA was normalized to 10% total input sample RNA level in each RIP sample and was represented as percentage relative to input sample using the delta Ct method[@b68].\n\nCo-IP and western blotting\n--------------------------\n\nHEK293T cells at 1.2 \u00d7 10^6^ per dish were plated onto 60-mm cell culture dish. Twenty-four hours later, the cells were transfected with 1\u2009\u03bcg of pcDNA3.1-Tat-HA and/or 0.8\u2009\u03bcg of pcDNA3.1-CUL4B-FLAG or pcDNA3.1-PSMD11-FLAG. The cells were co-transfected with 400\u2009pmol of specific or control siRNAs when needed. Alternatively, 1.5 \u00d7 10^6^ TZM-bl cells were plated onto 60-mm cell culture dish. Twenty-four hours later, the cells were transfected with 1\u2009\u03bcg of pcDNA3.1-NRON, 1.5\u2009\u03bcg of pcDNA3.1-Tat-HA and/or 1\u2009\u03bcg of pcDNA3.1-CUL4B-FLAG or pcDNA3.1-PSMD11-FLAG. Cells were co-transfected with 400\u2009pmol of specific or control siRNAs when needed. At 48\u2009h post transfection, cells were collected and disrupted with RIPA lysis buffer (150\u2009mM NaCl, 50\u2009mM Tris-HCl (pH 7.5), 1\u2009mM EDTA, 1% NP40 and 0.5% Triton X-100) containing protease inhibitor cocktail (Sigma) and RNaseOut (Invitrogen) for 30\u2009min on ice. The cell lysates were clarified with centrifugation at 12,000\u2009*g* for 10\u2009min at 4\u2009\u00b0C. The supernatants were pre-cleared with agarose beads, and then mixed with anti-HA beads (Sigma A2095) or anti-FLAG beads (Sigma F2426) and incubated at 4\u2009\u00b0C for 4\u2009h to overnight. The beads were washed five times with cold lysis buffer at 4\u2009\u00b0C. RNase A was added to wash buffer (20\u2009ug\u2009ml^\u22121^) when needed. RNA isolation followed by real-time qRT--PCR, or western blotting, was then performed[@b67]. The anti-HA antibody (1:2,000 diluted, mouse monoclonal, MBL International Corporation, M180-3), anti-FALG antibody (1:1,000 diluted, rabbit polyclonal, MBL International Corporation, PM020), anti-\u03b2-actin antibody (1:1,000 diluted, rabbit polyclonal, CST \\#4967), anti-GAPDH antibody (1:5,000 diluted, rabbit polyclonal, Proteintech Group, 10494-1-AP) or anti-p24 antibody (1:500 diluted)[@b64] were used as the primary antibodies. Images were analysed using the Quantity One Software (Bio-Rad). The uncropped scans of the most important blots were supplied in [Supplementary Fig. 6](#S1){ref-type=\"supplementary-material\"}.\n\n*In vitro* ubiquitin assay\n--------------------------\n\nHEK293T cells at 1.2 \u00d7 10^6^ per dish were plated onto 60-mm cell culture dish. Twenty-four hours later, the cells were transfected with 1\u2009\u03bcg of pcDNA3.1-Tat-HA, 2\u2009\u03bcg of pcDNA3.1-Flag-Ub or empty pcDNA3.1 vector, and 400\u2009pmol of specific or control siRNAs. MG132 (5\u2009\u03bcM) was added into the cultures 12\u2009h before the cells were harvested. At 48\u2009h post transfection, cells were collected and disrupted by lysis buffer for anti-HA immunoprecipitation. Western blot analysis was then performed with anti-HA or anti-FLAG primary antibodies.\n\nNorthern blotting\n-----------------\n\nFor NRON lncRNA detection, 20\u2009\u03bcg of total RNA was separated by 1.2% formaldehyde-denature agarose gel, and then transferred to Amersham Hybond-N^+^ membrane (GE Healthcare) and ultraviolet-crosslinked RNA to the membrane. The membrane was prehybridized with salmon sperm DNA (100\u2009\u03bcg\u2009ml^\u22121^, Sigma), and was hybridized with NRON and Actin-specific probes. Then, the membrane was washed three times, dried, exposed on a phosphorimager screen (GE Healthcare) and scanned on a laser scanner (GE Healthcare)[@b69]. The reverse primers of NRON and Actin for real-time qRT--PCR were used as oligonucleotide probes, and were labelled with \u03b1-^32^P (PerkinElmer) by T4 polynucleotide kinase (NEB).\n\nHIV *in vitro* latency model\n----------------------------\n\nThe HIV-1/VSV-pseudotyped viruses were packaged in HEK293T cells by co-transfecting pNL4-3-deltaE-EGFP-Bcl-2 and pVSV-G. The activated primary CD4^+^ T lymphocytes isolated from healthy individuals were infected with the pseudotyped viruses, and the GFP-positive cells were sorted out by FACS and expanded by culturing in the conditioned RPMI 1640 medium containing anti-CD3 antibody (1\u2009\u03bcg\u2009ml^\u22121^, BD Biosciences), anti-CD28 antibody (1\u2009\u03bcg\u2009ml^\u22121^, BD Biosciences) and IL-2 (10\u2009ng\u2009ml^\u22121^, R&D Systems) for 7 days. The cells were cultured for \u223c4 weeks with low concentration of IL-2 (2\u2009ng\u2009ml^\u22121^ R&D Systems), and then the GFP-negative cells were enriched by FACS and subjected to activation by various reagents.\n\nHIV-1-infected individual samples\n---------------------------------\n\nThe HIV-1-infected individuals with the blood plasma viral RNA less than 20 copies per ml and the number of CD4^+^ lymphocytes higher than 200\u2009\u03bcl^\u22121^ were recruited for our study. The PBMCs were isolated through Ficoll gradient centrifugation, followed by culture in the conditioned RPMI 1640 medium. For NRON knockdown assay, the integrated HIV-1 proviruses in the resting CD4^+^ T lymphocytes were confirmed with Alu-PCR[@b34]. The resting CD4^+^ T lymphocytes were transfected with siRNAs (200\u2009nM) and treated in combination with SAHA (300\u2009nM), or with anti-CD3 antibody, anti-CD28 antibody and IL-2 as positive controls. Viral RNAs in culture supernatant were isolated and detected by real-time qRT--PCR at 48\u2009h after transfection[@b34]. For intracellular RNA expression and proviral DNA detection, total RNA and DNA were directly isolated from the primary resting CD4^+^ T lymphocytes. HIV-1 viral RNA and NRON RNA expression levels were detected with real-time qRT--PCR, and then normalized to the endogenous control \u03b2-actin mRNA using the delta Ct method[@b34]. The integrated proviral DNA levels were detected using real-time PCR with a primer set targeting Gag ORF[@b34][@b56], and then normalized to \u03b2-actin DNA using the delta Ct method[@b68].\n\nStatistical analysis\n--------------------\n\nResults of the experiments were presented as mean\u00b1s.d. (error bars). Student\\'s unpaired *t-*test was used to determine significance. \\**P*\\<0.05 and \\*\\**P*\\<0.01 indicated significant difference. Correlation analysis was determined using a simple linear regression analysis in the SPSS software.\n\nData availability\n-----------------\n\nThe data that support the findings of this study are available from the corresponding author upon request.\n\nAdditional information\n======================\n\n**How to cite this article:** Li, J. *et al.* Long noncoding RNA NRON contributes to HIV-1 latency by specifically inducing tat protein degradation. *Nat. Commun.* 7:11730 doi: 10.1038/ncomms11730 (2016).\n\nSupplementary Material {#S1}\n======================\n\n###### Supplementary Information\n\nSupplementary Figures 1-6 and Supplementary Tables 1-3\n\nWe obtained the HIV-1 provirus plasmid pNL4-3-deltaE-EGFP that was constructed in the laboratory of Dr Robert Siliciano from the National Institutes of Health AIDS Reagent Program. This work was funded by National Special Research Program of China for Important Infectious Diseases (No.2013ZX10001004), National Science and Technology Major Project (No.2012ZX10001003-004-002), Guangdong Innovative Research Team Program (No.2009010058), the National Natural Science Foundation of China (NSFC-NIH project) (No.81561128007) and the Joint-innovation Program in Healthcare for Special Scientific Research Projects of Guangzhou, China (201508020256) to H.Z.\n\n**Author contributions** J.L. designed and performed the most experiments and analyses; C.C. and G.G. participated in generating the *in vitro* latency model; C.C. and X.M. contributed to some plasmid constructing, real-time qRT--PCR and western blotting experiments. Y.Z., B.L., S.Z., F.Z., C.L. and Y.Y. participated in some experiments. W.P.C. recruited the HIV-1-infected individuals. H.Z. directed and supervised the project and interpretation of data. J.L. and H.Z. prepared the manuscript.\n\n![LncRNA NRON represses HIV-1 replication.\\\n(**a**) The expression levels of lncRNAs in activated primary CD4^+^ T lymphocytes were detected with real-time qRT--PCR. The expression of *T-bet* was detected as positive control (*n*=3). (**b**) Real-time qRT--PCR detection of the lncRNAs expression level differences between the resting and activated primary CD4^+^ T lymphocytes from a same donor (*n*=3). (**c**) The activated primary CD4^+^ T lymphocytes were transfected with siRNAs against indicated lncRNAs or nonspecific control and were infected with HIV-1~NL4-3~ viruses. HIV-1 productions in the cultures were detected by p24 ELISA at 7 days post infection (*n*=3). (**d**) Northern blotting detection of NRON expression in the resting (R) or activated (A) primary CD4^+^ T lymphocytes from the same donors. Numbers indicated the fold change related to control. (**e**) The activated primary CD4^+^ T lymphocytes were transfected with siRNAs against NRON or nonspecific control and were infected with HIV-1~NL4-3~ viruses. HIV-1 productions in the cultures were detected with p24 ELISA at indicated time points post infection (*n*=3). The results in **a**--**c**,**e** show mean\u00b1s.d. (error bars). \\**P*\\<0.05, Student\\'s unpaired *t-*test.](ncomms11730-f1){#f1}\n\n![NRON represses HIV-1 transcription by inducing Tat degradation.\\\n(**a**) HEK293T cells were co-transfected with pNL4-3-deltaE-EGFP provirus plasmid and siRNAs against NRON or nonspecific control. Gag proteins expression was determined by western blotting with anti-p24 antibody at 72\u2009h after transfection. (**b**) Total HIV-1 mRNAs including spliced and unspliced mRNAs were detected with real-time qRT--PCR in the same HEK293T cells described in **a** (*n*=3). (**c**) HEK293T cells were co-transfected with pHIV-Pro-Luc reporter plasmids, pcDNA3.1-Tat-HA, pRL-CMV (as a transfection normalization reporter) and siRNAs against NRON or nonspecific control. The promoter activity was determined by dual-luciferase reporter assay at 48\u2009h after transfection (*n*=3). (**d**) Overexpressed NRON RNA was co-immunoprecipitated by HA-tagged Tat in TZM-bl cells, and the enriched RNA was determined by real-time qRT--PCR. RNA co-immunoprecipitated by HA-tagged GFP was set as a control (*n*=3), and the precipitated proteins were detected by western blotting. (**e**) The pcDNA3.1-NRON or empty vector was co-transfected with pcDNA3.1-Tat-HA into TZM-bl cells, the TAR-Luc RNA was enriched by RNA co-immunoprecipitation with anti-HA agarose and quantified with real-time qRT--PCR (*n*=3). The precipitated proteins were detected by western blotting. (**f**) Tat and control GFP were detected by western blotting on NRON knockdown in HEK293T cells. Numbers indicated the fold change related to control. Data in **b**--**e** show mean\u00b1s.d. (error bars). Results in **a**,**f** represent three independent experiments. \\**P*\\<0.05, \\*\\**P*\\<0.01, Student\\'s unpaired *t-*test.](ncomms11730-f2){#f2}\n\n![NRON conjugates Tat protein to the ubiquitin/proteasome system and induces its ubiquitination.\\\n(**a**) The indicated siRNAs were co-transfected with HIV-1 promoter reporter plasmids into HEK293T cells, and the promoter activities were measured with dual-luciferase reporter assay (*n*=3). (**b**) siRNAs against CUL4B and PSDM11 were co-transfected with pcDNA3.1-Tat-HA and pcDNA3.1-NRON into TZM-bl cells, and the Tat protein levels were detected by western blotting at 48\u2009h post transfection. Numbers indicated the fold change related to control. The endogenous NRON RNA in HEK293T cells was co-immunoprecipitated by ectopically expressed FLAG-tagged CUL4B (**c**) or PSMD11 (**d**) and was quantified using real-time qRT--PCR (*n*=3). *In vitro* ubiquitination assay was performed, and western blotting showed the ubiquitin-labelled Tat when NRON (**e**), CUL4B or PSMD11 (**f**) was knocked down. Data in **a**,**c**,**d** show mean\u00b1s.d. (error bars). Results in **b**,**e**,**f** represent three independent experiments. \\**P*\\<0.05, \\*\\**P*\\<0.01, Student\\'s unpaired *t-*test.](ncomms11730-f3){#f3}\n\n![Depletion of NRON reactivates HIV-1 viruses in latently infected CD4^+^ T lymphocytes.\\\n(**a**) Primary resting CD4^+^ T lymphocytes were nucleofected with HIV-1 promoter reporter system plasmids, pcDNA3.1-Tat-HA and siRNAs against NRON or nonspecific control. The promoter activity was determined with dual-luciferase reporter assay at 48\u2009h after transfection (*n*=3). (**b**) Tat and control GFP were detected by western blotting on NRON knockdown in nucleofected primary resting CD4^+^ T lymphocytes. Numbers indicated the fold change related to the control. (**c**) The latently infected cells were transfected with NRON siRNAs or nonspecific control, or were transfected with siRNAs in combination with the treatment of SAHA, and detected by FACS at 48--72\u2009h post transfection. The GFP+ ratio indicated the reactivation level (**d**; *n*=3). (**e**) Resting CD4^+^ T lymphocytes isolated from HIV-1-infected individuals on suppressive cART were transfected with siRNAs in combination with the treatment of SAHA. After 48\u2009h, HIV-1 virion-associated RNAs in the supernatants were isolated and detected with real-time qRT--PCR (*n*=3). (**f**) The intracellular HIV-1 RNA and NRON RNA expression levels were detected in resting CD4^+^ T lymphocytes isolated from HIV-1-infected individuals on suppressive cART (*n*=20), and the correlation between the HIV-1 RNA and NRON RNA levels was shown. The simple linear regression analysis was performed and linear regression line was shown. Data in **a**,**d**,**e** show mean\u00b1s.d. (error bars). Results in **b** represent three independent experiments. \\**P*\\<0.05, Student\\'s unpaired *t-*test.](ncomms11730-f4){#f4}\n\n[^1]: These authors contributed equally to this work.\n"}
+{"text": "Introduction {#section1-2050313X19882843}\n============\n\nLupus erythematosus is an autoimmune disease that can be skin limited or systemic. Discoid lupus erythematosus (DLE) is the most common variant of cutaneous lupus. While it has a lower chance (5%--10% life time) compared to other subtypes to progress to systemic forms, it has significant morbidity because of its tendency to involve cosmetically important areas such as the face and scalp and resulting disfiguring scarring alopecia.^[@bibr1-2050313X19882843]^ Similar to other subtypes of lupus, DLE is most commonly seen in women and in the third to fourth decade of life.^[@bibr2-2050313X19882843]^ DLE typically presents as erythematous to violaceous papules or plaques that eventually develop scales, follicular plugging, central hypopigmentation and scarring in photo-distributed areas.^[@bibr3-2050313X19882843]^ A localized form of the condition occurs in 80% of the patients with lesions developing solely on the face, ears, and/or scalp and 20% of patients present a disseminated form of the disease defined by the presence of lesions beyond the head and neck area.^[@bibr2-2050313X19882843]^ Rarely, discoid and other forms of cutaneous lupus may present in linear fashion following the lines of embryologic migration (lines of Blaschko).^[@bibr4-2050313X19882843]^ We report, the first to our knowledge, case of linear lupus with *en coup de saber* presentation clinically mimicking acute morphea.\n\nCase report {#section2-2050313X19882843}\n===========\n\nA 24-year-old non-smoking middle eastern male presented to dermatology clinic with a history of slowly progressive asymptomatic linear skin eruption. On clinical examination, a sharply demarcated pink to violaceous linear patch was noted extending from the root of his frontal scalp to the nasal tip ([Figure 1(a)--(b)](#fig1-2050313X19882843){ref-type=\"fig\"}). Given the clinical presentation, a Doppler ultrasound and a skin biopsy were performed to rule out vascular etiology and early *en coup de saber* morphea. However, the histological picture was diagnostic of discoid lupus showing characteristic features of basal vacuolation, follicular plugging, lichenoid infiltrate and mucin deposition ([Figure 1(c)](#fig1-2050313X19882843){ref-type=\"fig\"}). Full body examination, review of systems and laboratory investigations did not reveal any features of systemic lupus. The patient was prescribed an alternative day regimen of high potency topical steroid (clobetasol dipropionate 0.05% cream) and calcineurin inhibitor (tacrolimus 0.1% ointment) and strict photoprotection to halt inflammation and prevent scarring. He was seen in follow-up 4\u2009months later, at which time his lesions had mostly cleared.\n\n![(a) Erythematous plaque runs from the root of the frontal scalp down to the glabella. Two red and one purple papules are present in the region of the glabella, as well (white arrows). (b) The erythematous plaque appears to continue, after a brief interruption, down the left dorsum of the nose (black arrows). (c) Periodic acid--Schiff (PAS) staining of skin biopsy shows follicular plugging, vacuolar degeneration of the basal layer and lymphocyte predominant periadnexal inflammation.](10.1177_2050313X19882843-fig1){#fig1-2050313X19882843}\n\nDiscussion {#section3-2050313X19882843}\n==========\n\nLinear forms have been described in many morphological subtypes of lupus erythematosus, such as subacute, discoid, panniculitis and profunda. Various attempts have been made at explaining the linearity of these lesions as well as their tendency to follow the lines of Blaschko. For example, linear nevoid and linear inflammatory dermatosis both follow Blaschko lines and both show evidence of genetic mosaicism as a potential cause. Therefore, antigenic effects mediated by T-cell activity targeted at the mosaic cells could produce the observed skin lesions.^[@bibr5-2050313X19882843]^ Others have postulated that the linear lesions may be due to microchimerism that specifically targets skin cells in a mechanism like that of graft-versus-hosts disease.^[@bibr5-2050313X19882843]^ However, the predominate view is that the presence of the linear lesions is caused by genetically unique keratinocytes that, when exposed to ultraviolet light, produce an inappropriate cytokine response.^[@bibr5-2050313X19882843]^\n\nAlthough overlap syndromes that fulfill the diagnostic criteria of both DLE and morphea have been described, the case we present illustrates a case of mimicry.^[@bibr5-2050313X19882843]^ Upon skin biopsy, only classic histological findings of DLE, such as follicular plugging, increased mucin, basement membrane thickening and perivascular lymphocytic inflammation with involvement of adnexal elements were observed.^[@bibr6-2050313X19882843]^ This knowledge had clear implications on our choice of treatment and further emphasizes the importance and near-necessity of performing skin biopsies and subsequent histopathological analysis when using other mainstay methods of diagnosing cutaneous maladies.\n\nCutaneous lupus erythematosus is an autoimmune skin disease favoring sun exposed skin. The most common form is DLE. Similar to other subtypes of lupus, DLE may rarely follow the lines of Blaschko. We report here, for the first time, a case of linear DLE mimicking *en coup de saber* morphea.\n\n**Declaration of conflicting interests:** The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.\n\n**Funding:** The author(s) received no financial support for the research, authorship, and/or publication of this article.\n\n**Informed consent:** Patient consent was obtained for publication of this case report.\n"}
+{"text": "Introduction {#sec1}\n============\n\nCancer is a name used to describe a large and diverse group of diseases most notably characterized by a rapid increase in the growth and proliferation of certain cells (that have acquired genomic mutations) and a resultant tumour mass. This malignant tumour mass is distinguished from benign tumours because of the ability to metastasize, i.e. spread to other distant organs.\n\nHow cancer originates {#sec2}\n=====================\n\nCancer originates from a single mutation within the genome of a cell. Subsequent accumulation of more mutations can turn a normal cell into an aberrant cell.[@bib1] A mutation in which a tumour suppressor such as p53, Rb or p16^INK4a^/p14^ARF^ is knocked down, or an oncogene such as Ras become constitutively active, leads to excessive proliferation of the cell,[@bib2], [@bib3], [@bib4] and with each cell division, the chances of cells acquiring more mutations increases. Eventually, cells accumulate enough mutations to trigger endless growth and tumour formation.[@bib3]\n\nThe difficulty in treating cancer {#sec3}\n=================================\n\nGenerally, cancer is not noticed by patients until it has gotten to a late stage -- metastasis. At this stage, there is not much doctors can do. Doctors may try to remove the tumour mass surgically and administer chemotherapy and radiation, but apart from the side effects these techniques have on normal cells, chemotherapy is not very effective in completely wiping out cancer cells, because in a given tumour, there are different types of cancer cells with different kinds of mutations. Indeed, cancer tumours do not contain homogenous cancer populations, rather they contain heterogeneous cancer populations.[@bib5] This is one reason for the difficulty in treating cancer, as it is not easy to administer a drug combination that targets each of the cancer cells in a tumour, due to their genomic diversity. This concept is known as tumour heterogeneity (diversity in the cancer cells that make up a tumour). Based on literature review, tumour heterogeneity seems to be the main cause of cancer relapse and resistance to chemotherapy -- the biggest challenges in cancer treatment.\n\nOrigin of tumour heterogeneity {#sec4}\n==============================\n\nTumour heterogeneity is as a result of both intrinsic and extrinsic factors. Intrinsic factors include genetic and epigenetic mutations that contribute to tumorigenicity while extrinsic factors include the microenvironment around the tumour that interacts with the tumour to aid its progression.[@bib6]\n\nTheories of tumour heterogeneity {#sec5}\n================================\n\nResearchers have put forward theories to explain the origin of tumour heterogeneity from the intrinsic point of view. In this review, I will discuss the two prominent theories: the clonal evolution theory and the cancer stem cell theory. Both theories have data supporting them, but not without inconsistencies.\n\nClonal evolution theory {#sec6}\n=======================\n\nThis theory is based on Darwin\\'s theory of evolution. According to this theory, an initial mutation in a cell is passed on to each of the daughter cells and with each subsequent division, these daughter cells acquire and accumulate more and more transforming mutations. Over time, the cells with the most mutations that cause a growth and proliferation advantage are clonally expanded, leading to the formation of a tumour mass with clones of different aberrant cells.[@bib7] The evidence for this theory came from observations of different tumours. It was found that all the cancer cells in a given tumour had one or few founder mutations in common, while some mutations were specifically found in individual cancer cells. This implies that the cancer cells evolved from one original mutated cancer cell.[@bib7]\n\nCancer stem cell theory {#sec7}\n=======================\n\nThis is a more recent theory based on the observation that there are a few cells within many, if not all tumour masses that display stem cell-like properties namely: the ability to self-renew, as well as the ability to give rise to all the different types of cancer cells within that tumour.[@bib5] According to this theory, these cancer stem cells (CSCs) can be placed at the apex of a hierarchy, and can undergo either symmetric or asymmetric division ([Fig.\u00a01](#fig1){ref-type=\"fig\"}). When they undergo symmetric division, they either produce two identical daughter cells that are replicas of them (CSCs) -- this is known as 'self-renewal'; or two identical daughter cells that are progenitors (non-CSCs) and can subsequently differentiate to form any of the types of cancer cells within the tumour. There is also asymmetric division in which the stem cell divides to produce two non-identical daughter cells. One is a replica of the stem cell (CSC), while the other is a progenitor cell (non-CSC) which can go on to differentiate into any cancer cell type within the tumour (with contribution from stimuli produced by the microenvironment).[@bib5]Fig.\u00a01Cancer stem cell hierarchy. According to the cancer stem cell theory, a cancer stem cell is at the top of a hierarchy. It has the ability to self-renew (arrow 1), i.e. give rise to a daughter cell that is an exact replica of itself. It also has the ability to give rise to any of the cancer cell types within a tumour. It does this by dividing to form a daughter cell (non-CSC) that can subsequently divide severally before differentiating into diverse cancer cells (arrows 2a--5d, excluding arrow 2b). Tumour cell plasticity also occurs, in which a progenitor cell (non-CSC) spontaneously de-differentiates into a CSC (arrow 2b). It is thus able to initiate a new tumour.\n\nIt is very likely that the two theories of tumour heterogeneity may apply to human cancers of different types, and in some tumours, both models may apply.\n\nEvidence for the existence of cancer stem cells {#sec8}\n===============================================\n\nIn 1997, Bonnet and Dick[@bib8] showed that a specific group of leukaemic cells expressing CD34 on their cell surface and lacking CD38 were capable of initiating new tumours when injected into immunosuppressed NOD/SCID mice. Cancer cells with this type of ability have also been demonstrated to be present in solid human tumours.[@bib9] Importantly, these CSCs capable of initiating new tumours, have been shown to be relatively few within the tumour and this has raised the possibility that the current drugs used in managing cancer, destroy other cells within the tumour but not the CSCs (because they are relatively few), hence these CSCs regenerate the tumour once the therapy is discontinued, leading to relapse as well as resistance to the therapeutics previously used.[@bib5]\n\nThis concept is an important difference between the clonal evolution and the CSC theories, i.e according to the CSC theory, only a subpopulation of the cancer cells within the tumour, are cancer stem cells and are capable of initiating a tumour in a different location (e.g during metastasis or xenotransplantation). Therefore, the CSCs are said to be the drivers of tumour progression, even though they are just a minority. The clonal evolution theory however, claims that all cancer cells within a tumour, have the ability to form a new tumour mass in a different location and each tumour mass consists of numerous clones of different cancer cells harbouring different genetic mutations; hence, the heterogeneous nature of malignant tumours.[@bib5]\n\nChallenges of the cancer stem cell theory {#sec9}\n=========================================\n\nThe cancer stem cell theory, like many other scientific theories has been subjected to critique. One major drawback of the CSC model is the lack of definitive surface markers. The markers seen on CSCs tend to be similar to those on normal adult stem cells, normal cancer cells or normal tissues[@bib10], [@bib11], [@bib12], [@bib13] and also, the markers seen on different CSCs tend to differ from one another.[@bib13] Researchers have argued that there should be specific surface markers on CSCs that can be targeted specifically without the risk of damaging normal adult cells; and also, certain CSC markers should be expressed on many types of CSCs.[@bib13] In reality, it is not unrealistic to find CSCs sharing surface markers with normal adult stem cells, since it has been proposed that CSCs originate from mutations in normal stem cells.[@bib13] Also, tumour heterogeneity is useful in explaining the reason why CSCs of a particular organ express different markers in different locations (for instance in different human hosts).\n\nEffect of selective targeting of cancer stem cells {#sec10}\n==================================================\n\nCurrently, researchers are developing drugs that take tumour heterogeneity into account, as this is obviously a key reason for the mass failure in cancer therapy that has been recorded over the years. Based on the cancer stem cell model, a scan of 16,000 drugs was carried out, and one of them (Salinomycin) was found to be highly effective in killing breast CSCs, while sparing non-CSCs.[@bib14] The idea behind this was that, since it is CSCs that drive tumour progression, then wiping out CSCs will lead to shrinkage of the tumour mass. Indeed, the results were very close to expectations, but not without problems. It has been found that under conditions of genetic manipulation, mammary non-CSCs are able to utilize the mechanism of epithelial-to-mesenchymal transition (EMT) in order to undergo de-differentiation and become mammary CSCs once again.[@bib5], [@bib15], [@bib16] This is known as tumour cell plasticity. Another mechanism by which mammary non-CSCs can spontaneously de-differentiate into mammary CSCs *in\u00a0vivo* without any genetic manipulation, has subsequently been described[@bib17] -- See [Fig.\u00a01](#fig1){ref-type=\"fig\"}. Furthermore, tumour cell plasticity has been demonstrated in JARID1B-negative melanoma cells which spontaneously revert into a JARID1B-positive state, thereby giving rise to an increase in tumour growth.[@bib18] Hence, administering drugs like Salinomycin and Abamectin, (which specifically destroy CSCs, while sparing non-CSCs) might lead to tumour shrinkage initially, but with time the tumour will relapse if one or more of the non-CSCs is able to de-differentiate into a CSC, and worse still, the resulting tumour will most likely be resistant to the chemotherapy previously used.[@bib5]\n\nFuture strategies {#sec11}\n=================\n\nFuture research should be aimed at providing new drug combinations that kill both CSCs and non-CSCs or at least drugs that prevent non-CSC-to-CSC transition can be combined with drugs that selectively destroy CSCs. This may hopefully be more effective in the long run.\n\nConclusion {#sec12}\n==========\n\nCancer is extremely difficult to treat successfully, because there are too many things to be targeted therapeutically at the same time. Targeting one thing while leaving out the other, could lead to an initial tumour shrinkage, with an eventual relapse of the cancer and consequent drug resistance. The cancer stem cell theory is helping to advance the state of research, but not without its inherent challenges and limitations. Currently, research is on-going to develop combination therapies that will destroy cancer cells, prevent relapse, as well as have minimal side effects. This may seem to be ambitious, considering the complex nature of cancer, but I believe that, the remedy for individual cancers is nearer to us now than it ever was before!\n\nPeer review under responsibility of Chongqing Medical University.\n"}
+{"text": "Frozen accidents: evolution of the genetic code {#Sec1}\n===============================================\n\nA key challenge in genetic engineering and synthetic biology is to *change the unchangeable*---i.e., to thaw the so-called frozen metabolic accidents (FMAs). The term FMA refers to processes that are thought to be immutable, because their modification requires altering multiple intertwined components at the same time. Historically, the term \"frozen accident\" was coined to explain certain characteristics of the standard genetic code (SGC), which is outlined in the following. In fact, the assignment of the 20 canonical amino acids to the 64 codons of the SGC is clearly non-random, i.e., related amino acids typically occupy contiguous areas in the codon table \\[[@CR1]\\]. Several possible reasons for this have been proposed. Chemical interactions between amino acids and the tertiary structures of RNA-binding sites of codons (or anti-codons; \"stereochemical theory\") \\[[@CR2]\\], co-evolution of amino-acid biosynthesis and code structure (\"coevolution theory\") \\[[@CR3]\\], and selection for robustness (\"error minimization theory\") \\[[@CR4]\\] could all have contributed to the evolution of the SGC. Furthermore, the \"frozen accident\" perspective introduced by Francis Crick 50\u2009years ago \\[[@CR5]\\] explains the universality of the SGC by its effective fixation in the earliest life forms, such that any major change would be strongly selected against because it would immediately affect large numbers of proteins. This scenario does not require that the original assignment of codons occurred entirely by chance. In fact, the SGC displays clear signs of optimization: it is very robust, albeit not the most robust possible \\[[@CR6], [@CR7]\\], and the 20 canonical amino acids are thought to be virtually ideal for building soluble protein structures with close-packed cores \\[[@CR8]\\]. But the frozen accident perspective also emphasizes that once the SGC assignment had been made, it became essentially immutable.\n\nAmong the deleterious effects of codon reassignment, it is thought that inhibition of horizontal gene transfer (HGT) might have played a major role \\[[@CR1]\\]. In fact, HGT remains a key factor in microbial evolution \\[[@CR9]\\] and, assuming that the capacity to undergo HGT was essential for the early prokaryotes \\[[@CR10]\\], even small alterations of the code would have had detrimental effects on HGT, genetically isolating the species affected and dooming them to extinction. HGT itself also explains the universality of the code. Advanced simulations of evolution starting with several genetic codes and extensive HGT lead to a single universal code \\[[@CR11]--[@CR13]\\] (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a).Fig. 1.Evolution and instances of \"frozen (metabolic) accidents.\" **a** The standard genetic code (SGC) is a frozen genetic accident and became universal during early prokaryote evolution because of the eminent role of horizontal gene transfer (HGT) (*left side*). Each *circle* represents a species and different genetic codes are indicated by different fillings. Species with rare codes became extinct because they were less amenable to stable HGT. The code that remained became the SGC. Once it was universally adopted, it persisted even when HGT became less important. On the *right*, the evolution of photosystem subunits is schematically shown as an example of a FMA. In the thylakoid membrane (*yellow*), the core photosystem subunits (*circles*), having more than one transmembrane domain (TM), multiply interact with each other (*dotted lines*) and with their co-factors. They also interact to a lesser extent with other transmembrane photosystem proteins (with one TM, *rectangles*), as well as with peripheral photosystem proteins and soluble proteins (*triangles*). This leads to different rates and modes of evolution and sequence diversification (highlighted in *gray*, with increased variation being indicated by darker shading). The sequences of photosystem core subunits changed least during evolution and represent FMAs. **b** RuBisCO, PSII, and the molybdenum (Mo)-dependent cyanobacterial nitrogenase. RuBisCO consists of eight large (*light green*) and eight small (*dark green*) subunits and catalyzes carboxylation of ribulose-1,5-bisphosphate during the Calvin cycle. PSII contains a core with several subunits, of which D1 and D2 are shown. Other PSII subunits, including the oxygen-evolving complex (OEC), surround the core. In cyanobacteria, phycobilisomes act as antenna complexes, whereas light-harvesting complex II (LHCII) serves this function in eukaryotes. The Mo-nitrogenase is a two-component system comprising a heterotetrameric dinitrogenase complex with a P-cluster (P; Fe~8~S~7~) and an Fe-Mo cofactor (M), and a homodimeric dinitrogenase reductase with a Fe~4~S~4~ cluster (F)\n\nThe \"frozen metabolic accident\" concept {#Sec2}\n=======================================\n\nThe term FMA was introduced by Paul Falkowski and coworkers to explain the evolutionary dynamics of photosynthetic genes \\[[@CR14]\\]. Maximal co-evolution among photosynthetic genes occurs when their protein products physically interact with each other, and in prokaryotes, such genes are often clustered at the genomic level \\[[@CR14]\\]. Thus, co-evolved photosynthetic proteins are found in thylakoid multiprotein complexes (photosystems I (PSI) and II (PSII), cytochrome *b*~6~*f* complex, and ATPase) and in soluble enzyme complexes like the tetrapyrrole biosynthesis enzyme magnesium protoporphyrin IX chelatase and the Calvin cycle enzyme ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) \\[[@CR14]\\] (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b). The rate of amino acid substitution for photosynthetic proteins in the cores of the photosynthetic multiprotein complexes (with multiple transmembrane domains (TMs)) are indeed markedly lower than those of the small subunits surrounding the cores (with 1 TM), or peripheral or soluble photosynthetic proteins that lack TMs (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a). This implies that interactions of proteins with other proteins, lipids, or cofactors constrain their evolution in the core photosynthetic apparatus of cyanobacteria and photosynthetic eukaryotes \\[[@CR14]\\]. Interestingly, proteins involved in different multiprotein complexes also co-evolved, e.g., core proteins of PSI and PSII, as well as cytochrome *b*~6~*f* complex and NADH dehydrogenase proteins, suggesting that their functional linkage in thylakoid electron transport was responsible for their co-adaptation. Moreover, the stoichiometry of the output of the light reactions in terms of ATP and NADPH almost perfectly matches the needs of the Calvin cycle, indicating substantial co-evolution of entire functionally linked processes.\n\nWhen environmental conditions changed several hundred million years ago, the functions of some of these conserved core components of photosynthesis were compromised. This holds in particular for their performance in the presence of oxygen and high light intensities---conditions that directly (oxygen) or indirectly (the ozone layer that enabled colonization of terrestrial habitats by blocking UV radiation) resulted from oxygenic photosynthesis. The two most prominently affected proteins are the D1 protein of PSII and the RuBisCO enzyme (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b). Reactive oxygen species (ROS) directly damage D1 under high light intensities and inhibit its continuous replacement by newly synthesized copies \\[[@CR15]--[@CR19]\\], while RuBisCO's propensity to employ oxygen instead of CO~2~ as substrate essentially wastes light energy \\[[@CR20]\\]. Falkowski and coworkers estimated that these features of PSII and RuBisCO together reduce the potential overall efficiency of photosynthesis by at least 50% \\[[@CR14]\\].\n\nLike D1 and RuBisCO, nitrogenase (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b)---the enzyme complex that fixes atmospheric nitrogen in prokaryotes---evolved before oxygen was freely available in the atmosphere, and the core proteins of cyanobacterial nitrogenases remained virtually unchanged following the transition to an oxidizing atmosphere \\[[@CR21], [@CR22]\\]. Hence, nitrogenase represents a third instance of a FMA \\[[@CR23]\\]. Because of the oxygen sensitivity of the iron-sulfur clusters in these nitrogenases (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b), an estimated 20 to 30% of marine nitrogenase activity is inhibited by O~2~. It is tempting to speculate that this evolutionary inflexibility might also explain why cyanobacterial nitrogenases were not endosymbiotically acquired by eukaryotes.\n\nRedressing the effects of FMAs {#Sec3}\n==============================\n\nThe functional shortcomings of RuBisCO, nitrogenase, and the D1 subunit of PSII have stimulated attempts to enhance their efficiency, aiming to improve crop yield and biomass production. These efforts have encountered many obstacles, which revealed additional facets of the consequences of FMAs. Exploiting interspecific diversity in RuBisCO activity, improving the activity of its auxiliary protein RuBisCO activase, and altering the levels of regulatory metabolites have been identified as promising ways to enhance the enzyme's function in crop plants \\[[@CR20]\\]. While these approaches are still at an early stage and are complicated by the fact that in crop plants the small and large subunits of the enzyme are encoded in the nucleus and chloroplast, respectively, they have already uncovered additional constraints on the evolution of RuBisCO. For instance, the enzyme's subunits (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b) co-evolve not only with each other \\[[@CR14]\\], but also with its assembly factors \\[[@CR24]\\]. This highlights why the modification of FMAs will often require the exchange of entire modules of co-evolved components, including structural proteins, auxiliary factors, and the genetic elements necessary for their efficient expression \\[[@CR25]\\]. The number of auxiliary proteins implicated can even exceed the number of structural proteins present in the mature complex, in the case of RuBisCO prompting the coining of the term \"rubiscosome,\" describing its evolution from a stand-alone enzyme into an enzyme complex that involves various auxiliary factors \\[[@CR26], [@CR27]\\]. Indeed, it has only recently become possible to express a functional plant RuBisCO in *Escherichia coli* by co-expressing the large and small subunits of RuBisCo together with its five assembly factors \\[[@CR28]\\]. This heterologous plant RuBisCO expression system promises to provide a way to test variants of the enzyme for enhanced function in the genetic workhorse *E. coli*. In fact, a non-native Calvin cycle or parts of it have previously been functionally reconstituted in *E. coli* \\[[@CR29], [@CR30]\\], yeast \\[[@CR31]\\], or *Rhodobacter capsulatus* \\[[@CR32]\\], allowing its optimization by directed evolution \\[[@CR33]\\] or laboratory evolution \\[[@CR30], [@CR34], [@CR35]\\]. Because several natural and synthetic autotrophic pathways besides the Calvin cycle exist, it is possible to design alternative carbon fixation cycles in plants, and such efforts are now underway \\[[@CR36]--[@CR41]\\].\n\nWith regard to nitrogenases, three strategies are available for enhancing biological nitrogen fixation in crops: (i) boosting the process in naturally plant-associated bacteria, (ii) inducing formation of the root nodules that permit symbiosis between crop plants and N~2~-fixing bacteria, and (iii) directly transferring prokaryotic nitrogenase genes into plant genomes \\[[@CR42], [@CR43]\\]. An important initial step was taken more than 40\u2009years ago, when recombinant *E. coli* strains with nitrogenase activity were constructed by genetic engineering \\[[@CR44]\\]. However, these transgenic strains exhibited much lower nitrogenase activity than the original host and could not support diazotrophic growth on nitrogen-free medium. *E. coli*-based systems were also used more recently to combine the nitrogenase with electron transport components from plant organelles as power supplies for future engineering of diazotrophy in cereal crops \\[[@CR45]\\]. However, overcoming the FMA of nitrogenases by mitigating their oxygen sensitivity has not yet been accomplished.\n\nWhile recombinant *E. coli* strains are available for RuBisCO and nitrogenase that provide a platform for modifying these suboptimal enzymes by genetic engineering and laboratory evolution, the situation for the PSII core proteins, including the D1 protein, is less favorable. Reconstitution of the photosynthetic light reactions in *E. coli* has proved to be a challenging task and has not yet been achieved. In fact, owing to their slow tempo of evolution (as expected for products of a FMA), the PSII core proteins are largely conserved between cyanobacteria and plants. Nevertheless, attempts to replace the cyanobacterial PSII core proteins D1, CP43, CP47, and PsbH with their counterparts from flowering plants (with which they share between 78 and 86% identity at the amino acid level) have been only partially successful \\[[@CR25]\\] (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). The resulting strains were either non-photoautotrophic or severely impaired in photosynthesis \\[[@CR46]--[@CR49]\\], demonstrating that the exchange of one component of this module of co-adapted PSII core proteins can markedly perturb the multiple interactions and the functionality of the entire PSII core.Fig. 2.Current approaches to modifying the PSII core. *Left panel*: In PSII from cyanobacteria, algae and plants, the core subunits (*circles*) multiply interact (*dotted lines*) with each other and to a lesser extent with other transmembrane (*rectangles*) and peripheral (*triangles*) proteins. *Middle panel*: If core subunits in cyanobacteria (*blue*) are replaced by their plant pendants (*green*), their sequence differences may weaken or disrupt these interactions (*thinner dotted lines*), severely affecting the function of PSII. *Right panel*: Stephen Mayfield and co-workers \\[[@CR50]\\] replaced an entire green algal PSII core with variants from other green algal species (indicated by a lighter shade of green). This would be expected to maintain the interactions between PSII core subunits, but might have affected interactions with other PSII proteins\n\nIn light of this limitation, Stephen Mayfield and co-workers proposed to replace the entire PSII core with its counterpart from another species, thereby maintaining the multiple intrinsic interactions within the PSII core that have evolved over millions of years in each photosynthetic species \\[[@CR50]\\] (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}). To test this hypothesis, the six original core PSII genes (*psbA*, *B*, *C*, *D*, *E*, and *F*) from the chloroplast genome of the green alga *Chlamydomonas reinhardtii* were deleted and replaced by a single synthetic construct that contained the orthologous genes from *Volvox carteri* or *Scenedesmus obliquus* (two other green algal species) or, as a control, from *C. reinhardtii* \\[[@CR50]\\]. In addition, the effect of replacing only subsets of the six genes was investigated. These experiments showed that (i) the strains reconstituted with the *C. reinhardtii* PSII gene sets showed the best photosynthetic performance, albeit lower than in the starting WT strain, and (ii) in the strains with replacements from *V. carteri* and *S. obliquus*, photosynthetic performance declined with increasing numbers of exchanged genes. These results imply that both the organization of the substitute genes in a synthetic construct (that might lack some *cis*-acting elements and their original operon structures, leading to suboptimal gene expression) and off-target effects of the PSII gene deletions (that might affect other operons and adjacent tRNA genes) in the *C. reinhardtii* chloroplast could decrease the photosynthetic performance of the transgenic strains. Moreover, the experiment could not clarify whether the roadblocks presented by FMAs can be truly removed by exchanging the entire ensemble (or module) of interacting and co-evolved proteins. In fact, the substituted green algal proteins were so similar with respect to their sequence---with average identities of between 93% (*S. obliquus*) and 98% (*V. carteri*) relative to the deleted original *C. reinhardtii* genes---that the strong negative effect on PSII function resulting from perturbations in gene expression in the host system used might have masked any subtle positive effect produced by the retention of the intrinsic interactions between the six co-adapted proteins. In consequence, a more appropriate experimental approach would be to replace entire cyanobacterial photosystem cores by their functional equivalents from higher plants \\[[@CR25]\\], for which single subunit exchanges have proven to be clearly detrimental (see above). In such experiments, the exchange of multiple subunits might alleviate the strong negative effects of single subunit substitution, given that the experimental setup can achieve efficient expression of the introduced genes and efficient assembly of the corresponding proteins---the latter one might even require introduction of plant-specific assembly factors. Moreover, it might be necessary to replace not only the six core subunits, but also additional PSII proteins that physically and/or functionally interact with the PSII core.\n\nNovel approaches to bypassing roadblocks caused by FMAs {#Sec4}\n=======================================================\n\nFunctional resolution of FMAs by replacing the evolutionarily immutable protein (complexes) with more efficient variants is an enormous challenge. But the issue is not only of interest for evolutionary biologists, but would have significant implications for plant productivity, as it might be the only way to enhance photosynthetic performance and nitrogen fixation rates. With respect to photosynthesis, this endeavor involves the realization of an evolutionary development that never actually took place. One needs to create a primordial version of photosynthesis under aerobic conditions at high light intensities---the very conditions produced by the advent of oxygenic photosynthesis (see above)---and let it evolve. The outcome of this evolution might be to reduce the light sensitivity of the PSII reaction center and the current susceptibility of the carbon fixation cycle to the level of oxygen in the atmosphere.\n\nHow can this challenge be tackled? A \"conservative\" solution to create novel proteins would be to fully exploit the potential of the 20 canonical amino acids encoded by the SGC. Indeed, concepts for systematically designing entire libraries of \"non-natural proteins\" (based on the canonical amino acid repertoire) that can be employed in vivo to replace natural proteins have been presented, and proof of their practicability has been provided \\[[@CR51]--[@CR53]\\]. A second possibility is to expand or alter the genetic code. In fact, some organisms have succeeded in co-opting the two non-canonical amino acids selenocysteine and pyrrolysine into the code \\[[@CR1]\\], and nonstandard amino acids with residues that have unusual chemical properties could be employed for the design of novel protein functions hitherto not possible with canonical amino acids. Thus, concepts for recoding genomes by re-assigning or deleting codons have been developed with the aim of enabling multivirus resistance, enhanced incorporation of nonstandard amino acids, or biocontainment by synthetic auxotrophy \\[[@CR54]\\]. Such strains have been designed, generated, and subsequently streamlined by adaptive evolution \\[[@CR55]--[@CR57]\\].\n\nThese two concepts for the design of new proteins to counteract the repercussions of FMAs may be applied either \"conservatively,\" i.e., exploiting the opportunities offered by such scenarios (including the possibility of using non-natural proteins), or \"unconventionally,\" i.e., employing entirely new components and interactions. In addition, the recognition that all interacting/co-evolving components in a module may need to be altered must be taken into account. In consequence, three approaches for redressing the aftereffects of FMAs can be envisioned: exploiting (i) natural or (ii) non-natural variants of the components of the module concerned, and (iii) re-designing the corresponding biological process (almost) from scratch (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}). All three approaches can be expected to require extensive laboratory evolution \\[[@CR58]\\] to streamline the novel processes, i.e., exploiting the high rate of evolution in microbial systems to optimize the expression, interactions, and performance of the introduced proteins for maximal implementation of the desired function. The third approach must follow an \"alternative evolutionary trajectory,\" either beginning with a primordial progenitor version of the process or re-designing the process from scratch. These three general concepts will now be discussed in the context of the re-engineering of the PSII core in order to bypass a FMA. In this instance, the difficulties arise from the special biophysical properties and pigment-binding sites of its components, as well as their multiple interactions and complex biogenesis \\[[@CR59]--[@CR62]\\].Fig. 3.Overview of the three general strategies for the resolution of FMAs. *Left panel*: The module responsible for the FMA is replaced by a natural variant of the module. This might be effective if the alternative module has evolved beneficial characteristics due to high selective pressure. Possible trade-offs and interactions with proteins extraneous to the transferred module might need to be considered, but could be mitigated by laboratory evolution (see Fig.\u00a0[4](#Fig4){ref-type=\"fig\"} for more details). *Middle panel*: The module is replaced by non-natural proteins that have the same interaction potentials but different sequences. This will very likely require a step-wise strategy, with successive replacement of individual proteins via a complementation-based approach, combined with streamlining of protein functions by laboratory evolution (see Fig.\u00a0[5](#Fig5){ref-type=\"fig\"} for more details). *Right panel*: Alternative evolution of a primordial version of the FMA or its entire de novo design might be the only solution if the general design of the FMA is fundamentally flawed and cannot be corrected by the first two approaches (see Fig.\u00a0[6](#Fig6){ref-type=\"fig\"} for more details)\n\nExploiting natural sequence variations to enhance PSII functions {#Sec5}\n================================================================\n\nAlthough FMAs, by definition, are largely refractory to natural evolution, screening of existing biodiversity might nevertheless provide variants that mitigate their negative effects. For PSII, and in particular the D1 protein, the isolation of the non-model green alga *Chlorella ohadii* (from desert soil crusts) and its physiological characterization may lead to a breakthrough in enhancing the light resistance of plant PSII \\[[@CR63]--[@CR66]\\]. Even under extremely high irradiation intensities, growth of *C. ohadii* remains unimpaired and its photosynthetic O~2~ evolution actually increases at high light levels \\[[@CR63]\\]. In contrast to other photosynthetic eukaryotes, very little of its D1 is degraded under such conditions, indicating that (i) PSII in *C. ohadii* might generate fewer ROS than do canonical PSII complexes \\[[@CR63]\\] or (ii) PSII is less accessible to or better protected against ROS. Because green algae and flowering plants employ similar PSII complexes, with light-harvesting proteins as antennas (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}b), transfer of the entire *C. ohadii* PSII core module into a flowering plant should be more straightforward than that between cyanobacteria and flowering plants (see above), assuming that efficient expression of all components can be achieved, and biogenesis of the complex does not require (too many) algal-specific auxiliary proteins. However, although the molecular mechanisms of the high light tolerance of *C. ohadii* are still elusive \\[[@CR63]\\], it appears to be unlikely that a modified PSII core is responsible for the enhanced light tolerance, with a combination of special antenna complexes and auxiliary factors providing a more plausible explanation. Moreover, the specialized ecological niche occupied by *C. ohadii* also raises the possibility of an as-yet unknown trade-off connected with the alga's high light resistance, although enhanced growth of *C. ohadii* is observed even at low light intensities \\[[@CR65]\\], which argues against the explanation that solely special antenna complexes (in terms of a reduced size) are responsible for the high light tolerance of this organism. But even if transfer of the PSII core of *C. ohadii* is not a promising approach to enhance light stress resistance, the question arises whether it is practicable to exchange entire PSII cores between green algae and flowering plants, thereby extending the approach of Gimpel et al. \\[[@CR50]\\] to the next level of complexity. In fact, dysfunctional interactions between the transferred module and extrinsic components might affect the overall function of the process (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). The latter can in principle be rectified by laboratory evolution (although this is not yet practicable in flowering plants) or by replacing the extrinsic components. A third possibility would be to streamline the plant PSII complex with a green algal core in a microbial test system that is amenable to genetic manipulation and laboratory evolution.Fig. 4.Replacing modules resulting from FMAs by natural variants. The original variant (*left*) is replaced by another natural variant with advantageous characteristics (*middle*). While the intrinsic interactions within the transferred module should already be optimal (*thick dotted lines*), perturbation of interactions (*thin dotted lines*) with extrinsic components might affect the overall function of the process. In that case, these extrinsic components must also be exchanged or suitably modified by laboratory evolution (*right*) to accommodate the chimeric multiprotein complex, perhaps also involving adaptive sequence variation (indicated by altered shading) of contact sites in the complex itself. Note that this approach is not likely to succeed for PSII (see main text) but might be feasible for other instances of FMAs\n\nExploiting non-natural sequence variations to enhance PSII functions {#Sec6}\n====================================================================\n\nWhile the exact number of natural protein sequences is unknown, it clearly represents a minuscule fraction of all possible proteins. Indeed, for a protein containing 100 amino acids, 20^100^ possible sequences exist, enough to fill a volume larger than that of Avogadro's number of universes \\[[@CR67]\\]. Certainly, only a small fraction of all possible proteins is compatible with biological systems, because many sequences would simply result in intrinsically disordered proteins. Of the compatible fraction, only a very tiny part has been exposed to evolutionary pressures on Earth, leaving a vast number of non-natural proteins capable of replacing existing natural proteins or mediating entirely novel functions. This concept is not only plausible; it has already been shown to be practicable \\[[@CR51]\\]. Libraries of non-natural proteins with the potential to sustain the growth of living cells can be constructed in a systematic and knowledge-based manner, and non-natural proteins that perform specific functions have been identified by phenotypically complementing mutations in natural proteins \\[[@CR51]\\]. This approach has resulted in the identification of non-natural proteins that rescued deletion mutants lacking natural proteins with various specific activities, including phosphoserine phosphatase, citrate synthase, threonine deaminase, and enterobactin esterase \\[[@CR68]\\]. Moreover, established engineering parameters were used to generate simple non-natural four-\u03b1-helix bundles and build basic oxidoreductase activities into these scaffolds to create completely artificial redox proteins (\"maquettes\") that can plug into natural biochemical pathways and are functionalized in vivo \\[[@CR52], [@CR53], [@CR69]\\].\n\nThe central question in the context of PSII is whether a protein with the same tertiary structure and the same complex interaction pattern and cofactor-binding sites as a natural PSII core protein might be capable of substituting for that PSII protein even though their primary sequences might be entirely different. Indeed, analogously to the first test cases for non-natural proteins in *E. coli* described above, such an approach might employ libraries of non-natural proteins expressly designed to have features similar to those of the target to be replaced, e.g., the same number of transmembrane domains and the same cofactor-binding sites. Rescuing a mutant that lacks a specific natural protein with the aid of a suitable non-natural protein could then be accomplished by a complementation-based strategy, provided that the recipient can also be propagated heterotrophically (as in cyanobacterial and algal model systems). Alternatively, the heterotrophic step can be avoided if the natural protein is replaced by the non-natural one via homologous recombination, which is feasible in cyanobacteria and the chloroplasts of some eukaryotes. The resulting lines might then contain a non-natural protein that has replaced a natural co-adapted protein whose evolution had been constrained by a FMA (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}). The next step would be to optimize the introduced non-natural proteins by laboratory evolution, either by optimizing each protein before the next one is introduced or by introducing and optimizing them as a batch. In the former case, one might create a system comprised of components that display interactions and features similar to those of the original, but differ markedly from them in their primary sequences, whereas in the latter case interactions and features might differ substantially from the original. This concept is intriguing, but will be very difficult to implement. First, it requires detailed knowledge of the structure--function relationships of the proteins concerned. Even for the best-studied protein complexes---like the photosystems, RuBisCO, or ribosomes---our current understanding might not be sufficient. Secondly, a non-natural protein or maquette with similar structure and interactions to its natural \"predecessor\" but completely different in sequence might be refractory to the natural assembly process, making novel assembly factors necessary. Thirdly, as laboratory evolution is feasible only in microbial systems, if specific aspects of plant photosynthesis need to be re-designed, the corresponding components must first be transferred into a suitable microbial host. Furthermore, if a new system consisting of non-natural proteins displays the same intrinsic interactions, it might also represent a soon-to-be FMA. Finally, replacing the original proteins with non-natural proteins or maquettes with similar functions will not enhance the process if the basic structure of the process is inherently suboptimal, for instance if any heterodimerically organized reaction center of PSII inevitably leads to D1-based photoinhibition \\[[@CR70]\\]. In the latter case, its re-design with non-natural proteins with similar characteristics will also be in vain, and the only solution is to redesign the process from scratch as outlined below.Fig. 5.Resolving the consequences of FMAs using non-natural components. A similar workflow as for replacing the original variant by another natural variant (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}) is employed, except that, in the first step, the natural protein (*left*) is replaced by a functional non-natural protein from a library specifically designed for this purpose (*middle*). This can be achieved via homologous recombination (in the case of cyanobacterial or plastid transformation) or by complementation of a mutant that is grown heterotrophically (shown here). The function of the protein can be enhanced by laboratory evolution (*right*), optimizing interactions (from *thin dotted lines* to *thick dotted lines*) and involving adaptive sequence variation (indicated by altered shading) of the introduced non-natural protein and its binding partners\n\nRedesign from (almost) scratch to enhance PSII functions {#Sec7}\n========================================================\n\nHow can the deleterious consequences of a FMA be redressed if the process itself is already intrinsically flawed? Or, with respect to PSII, what if all designs of water-splitting PSII found in extant organisms share the same inevitable flaw of photosensitivity and D1 damage? In fact, it can be argued that the energetics of the reaction mediated by PSII necessitates PSII being susceptible to ROS generation. In this scenario, re-designing PSII such that it circumvents ROS production is impossible, although it might well be engineered so that less ROS are produced, ROS are better scavenged, or that central parts of the complex are less accessible to ROS. Another thought is that a certain level of photosensitivity of PSII has been positively selected for during evolution because this serves to protect PSI against photoinhibitory damage, which is considered to be more dangerous than PSII photodamage \\[[@CR71]--[@CR74]\\]. In this case, high light tolerance of both PSII and PSI has to be enhanced to have a positive impact on the resistance of the photosynthetic light reactions towards high levels of light, and only two enhancement strategies remain open: (1) re-design of the process, starting from a primordial version that does not exhibit the relevant defect(s), or (2) complete replacement with a synthetic process designed from scratch (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}). With regard to the former possibility, all natural light-dependent and water-splitting systems in cyanobacteria, algae, and plants are related to each other, but anoxygenic bacterial relatives of PSII have a less complex structure and might serve as models of primordial versions of PSII \\[[@CR70], [@CR75]--[@CR78]\\]. Starting from these primitive versions of PSII, one needs to recapitulate the evolution from non-oxygenic, bacterial, single-photosystem situations to a cyanobacterial, oxygenic, two-photosystem solution that operates optimally in the presence of oxygen and high light. This option presents various formidable technical and conceptual obstacles, and will require a more detailed understanding of the mechanisms and evolution of photosynthesis than has yet been attained. The second possibility---inventing a synthetic system that is devoid of the flaws of the natural one---is even more challenging. Although it is clear that water splitting capacity will always be associated with the potential to produce ROS and that it is unlikely that a new type of in vivo photosynthesis can be constructed from scratch in the near future, the associated opportunities are fascinating, as one could also combine chlorophylls and bacteriochlorophylls in the same organism, or even incorporate novel pigments that absorb more of the light spectrum than conventional ones can utilize.Fig. 6.Comprehensive redesign of the process that is associated with the FMA. A primordial state of the process might be reconstructed by, for instance, comparing anoxygenic bacterial and oxygenic photosynthesis (*upper panels*). From this primordial process, an alternative version of the process to be replaced can be constructed in a step-wise manner, employing genetic engineering, as well as in vitro approaches using libraries of suitable compounds (*middle panels*). If a suitable primordial process cannot be reconstructed or is already flawed, the process might be re-designed from scratch, requiring deep knowledge of the original process (*bottom panels*)\n\nHarnessing synthetic biology to resolve frozen metabolic accidents {#Sec8}\n==================================================================\n\nDue to their immense significance for the biosphere in terms of nitrogen and carbon fixation, efforts to enhance the processes associated with FMAs are of the utmost interest. Elegant manipulations of the genetic code have demonstrated that it is feasible to get around the frozen *genetic* accident, but frozen *metabolic* accidents involve another level of complexity, owing to the intricate biophysical and structural interdependencies implicated in the frozen state, which extend to auxiliary components. To resolve the roadblocks caused by FMAs, a comprehensive knowledge of the process to be modified is vital, and elegant genetic engineering strategies have to be designed. The most promising and technically feasible approaches of resolving FMAs currently are to exploit the few evolutionary opportunities that they have left open by screening extant biological diversity for possible remedies (e.g., the high-light resistant PSII from *C. ohadii*) and applying laboratory evolution in cases where the process can be transferred to a suitable host (as in the case of nitrogenase and the Calvin cycle in *E. coli*). A much more challenging approach will be to test large sets of non-natural proteins or maquettes for their ability to overcome some of the limitations imposed by FMAs, and this will be only possible in suitable microbial systems that are accessible to efficient genetic engineering and laboratory evolution. The ultimate challenge will be to redesign from scratch a process that has been trapped in an evolutionary blind alley and is in principle suboptimally adapted to current conditions, at least in terrestrial environments. However, replacing this process with a completely different one amounts to discovering an evolutionary path that was never realized in nature, because the appropriate combination of starting conditions and selective forces never materialized. In this context it remains to be shown whether PSII truly qualifies as FMA because its sensitivity to light stress due to ROS production might have evolved regardless of the conditions (aerobic or pre-aerobic) and might therefore be inevitable, although it is clear that at least some organisms apparently have developed mechanisms that can largely overcome this limitation (the *C. ohadii* case). Moreover, the photosensitivity of PSII also protects PSI against photoinhibition, extending the need for re-design of both photosystems. Whichever explanation(s) for the inherent light sensitivity of PSII will finally turn out to be correct, for enhancing the function of a process assumed to be a FMA one might ideally start from a primordial process that is sufficiently flexible to allow it to be re-designed by employing non-natural proteins and laboratory evolution. For such experiments, *E. coli* is the ideal workhorse, and given that fixation of nitrogen \\[[@CR44]\\] and carbon \\[[@CR30]\\] and the biosynthesis of chlorophyll \\[[@CR79]\\] and carotenoids \\[[@CR80]--[@CR82]\\] have already been reconstituted in this host, the light reactions of photosynthesis are an especially attractive target of this approach. Thawing out FMAs will, however, require synthetic biology and laboratory evolution, but also further knowledge of the reactions themselves.\n\nThe author thanks Paul Hardy and Tatjana Kleine for critical reading of the manuscript.\n\nFunding {#FPar1}\n=======\n\nThe author receives funding for his work on synthetic biology and enhancing photosynthesis from the Deutsche Forschungsgemeinschaft (DFG; GRK 2062 and SFB-TRR 175).\n\nAvailability of data and materials {#FPar2}\n==================================\n\nNot applicable.\n\nDL wrote, read, and approved the final manuscript.\n\nCompeting interests {#FPar3}\n===================\n\nThe author declares that he has no competing interests.\n\nPublisher's Note {#FPar4}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "![](hosplond72559-0018){#sp1 .380}\n"}
+{"text": "Introduction\n============\n\nThe marine reptile clade Ichthyosauromorpha comprises two groups, namely Ichthyosauriformes that lasted for about 160 million years and spread worldwide ([@ref-18]; [@ref-28]) and Hupehsuchia, which is a small group of heavily-built reptiles known only from the Spathian (Lower Triassic) of Hubei Province, China ([@ref-45]; [@ref-2]). The two groups faced contrasting fates at the end of the Early Triassic but they were almost equally diverse in the Spathian, when they were relatively new to the history of life ([@ref-29]). Early ichthyosauriforms from that time period had been known since 1929 based on *Grippia longirostris* from Spitsbergen ([@ref-42], [@ref-43]), and three additional species were reported within the same century, namely *Chaohusaurus geishanensis* from China ([@ref-46]), *Utatsusaurus hataii* from Japan ([@ref-40]), and *Parvinatator wapitiensis* from Canada ([@ref-38]). However, none of them was represented by a complete skeleton and it was not until complete body fossils were reported that we learned the bauplan of these animals---at least *Chaohusaurus* and *Utatsusaurus* appeared like lizards with flippers ([@ref-36]; [@ref-33]), unlike later ichthyosaurs that were fish-shaped. Three more species have been recognized since, namely *Gulosaurus helmi* from Canada ([@ref-6]), *Cartorhynchus lenticarpus* from China ([@ref-28]), and *Sclerocormus breviceps* from China ([@ref-11]). *Thaisaurus chonglakmanii* from Thailand ([@ref-16]) may also belong to the list but further study is needed to establish the exact stratigraphy.\n\n*Chaohusaurus* is by far the best-known genus of the Spathian ichthyosauriforms, being represented by dozens of skeletons, and more than 10 scientific papers focusing on the genus have been published ([@ref-46]; [@ref-1]; [@ref-36]; [@ref-34], [@ref-35]; [@ref-14]; [@ref-5]; [@ref-31], [@ref-32], [@ref-26]; [@ref-47]). Three species are currently recognized in the genus. The type species *Chaohusaurus geishanensis* and the second species *Chaohusaurus chaoxianensis* are known from the Chaohu fauna in Anhui Province, China. The third species, *Chaohusaurus zhangjiawanensis*, is from the Nanzhang-Yuan'an fauna in Hubei Province. Of the three species, *Chaohusaurus chaoxianensis* is the most abundant, whereas the other two are known from only a few specimens each. Most of the specimens of *Chaohusaurus chaoxianensis* were collected from former limestone quarries in the Majiashan area, located north of the eastern outlet of Chaohu Lake, Anhui Province, China. Multi-year excavations that started in 2010, by a joint team from the Peking University, University of California, Davis, University of Milan, and Anhui Geological Museum, unearthed about 60 marine vertebrate specimens from the area. About 40 of the specimens were tentatively identified as *Chaohusaurus chaoxianensis* based on the presence of poorly ossified carpals or tarsals, which is among the features that distinguish the species from *Chaohusaurus geishanensis* ([@ref-32]).\n\nA recent examination of morphological variation in these specimens revealed four morphotypes that likely represented females and males of two taxa, which were tentatively referred to as Types A and B ([@ref-26]). The two morphotypes were distinguished based on a suite of both qualitative and quantitative morphological characters. There are at least a dozen specimens for each morphotype---numbers that are far greater than those for the other two species of *Chaohusaurus*. These morphotypes most likely represent different species given that morphological differences do not reflect sexual dimorphism ([@ref-26]). Type A ([Fig. 1](#fig-1){ref-type=\"fig\"}), for which 13 specimens are known, contains the holotype of *Chaohusaurus chaoxianensis* so these specimens are considered to belong to this species. Type B ([Fig. 2](#fig-2){ref-type=\"fig\"}), with 21 specimens, had yet to be named or described, although some of the specimens were previously included in descriptive studies of *Chaohusaurus chaoxianensis* (see \"Systematic Paleontology\"). It is therefore important to clarify the taxonomic confusion in the existing literature. The purpose of this paper is to describe Type B as a new species of *Chaohusaurus*, and clarify its diagnostic differences with *Chaohusaurus chaoxianensis*.\n\n![The most complete specimen of *Chaohusaurus chaoxianensis* (AGB6256).\\\n(A) Photograph. (B) Approximate bone map. See the section \"Osteological abbreviations\" for abbreviations. Scale bar is five cm in total.](peerj-07-7561-g001){#fig-1}\n\n![Holotype of *Chaohusaurus brevifemoralis* sp. nov (AGB7401).\\\n(A) Photograph. (B) Approximate bone map. See the section \"Osteological abbreviations\" for abbreviations. Scale bar is five cm in total.](peerj-07-7561-g002){#fig-2}\n\nMaterials and methods\n=====================\n\nSpecimens\n---------\n\nThe specimens used in this study are listed in Systematic Paleontology. Most specimens are articulated skeletons lacking some parts of the body. All specimens have undergone at least some degree of compressional deformation during preservation, adding biases to the morphological data. The degree of compression depends on the specimen. Thick bones or bony structure are most prone to the bias from such deformation. For example, when comparing specimens of similar sizes, the shaft of long bones may appear wider in one specimen than the other because of preservational deformation; the wider shaft has been flattened and widened more severely through compaction. In extreme cases, a convex surface my appear slightly concave because of compaction. The extremity of the same bones, on the other hand, seems to suffer less from similar flattening and widening, probably because the average density of the bone in these areas are higher than in the shaft. Apart from long bones, cranial structures are often distorted by compactional deformation because of their thickness. It is therefore important to compare many specimens first to grasp the original morphology of bones before compression.\n\nNomenclature\n------------\n\nThe electronic version of this article in portable document format will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank Life science identifiers (LSIDs) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix <http://zoobank.org/>. The LSID for this publication is: urn:lsid:zoobank.org:pub:3FA09089-C940-4499-ABBC-B48F0F70F38E. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.\n\nPhylogenetic analysis\n---------------------\n\nThe phylogenetic position of *Chaohusaurus brevifemoralis* was analyzed based on a modified version of a published taxon-character data matrix ([@ref-29]). Character order has been modified to improve anatomical consistency; the correspondence between the new and old numbers are summarized in [Table S1](#supp-2){ref-type=\"supplementary-material\"}. The following characters states and descriptions were revised. A new character was added as Character 61, Dorsal orbital margin: (0) on the same plane as the rest of orbital margin; (1) with medial excursion away from the main orbital plane. The original character 68, Dentigerous region in adults, was divided into two characters, namely Character 74, Premaxillary teeth presence and Character 75, maxillary teeth presence. A new character state was added to Characters 47 and 95 respectively, to reflect the unique morphology in *Chaohusaurus* and *Cartorhynchus*---Character 47, Basipterygoid process: (0) located antero-laterally; (1) located postero-laterally and short, giving basisphenoid a square outline in dorsal view, (2) located postero-laterally and markedly expanded laterally, being wing-like, giving basisphenoid a marked pentagonal shape in dorsal view; and Character 95, Humerus anterior flange: (0) absent; (1) present but concave or notched; (2) present and complete; (3) present but reduced proximally, leaving leading edge tuberosity. State 0 of Character 108, Shape of the posterior surface of ulna, contained the word \"radius\" that needed to be replaced by \"ulna.\" Character state description of Character 177, Presacral count, was revised to: (0) less than 35; (1) between 35 and 54; (2) more than 55. These new states were incorporated into the matrix through re-coding relevant taxa. Some characters were sorted for anatomical consistency ([Table S1](#supp-2){ref-type=\"supplementary-material\"}), and the original Characters 144, Interdigital separation, and 168, Spatium interosseum between tibia and fibula size, were removed to avoid redundancy and inconsistency.\n\nAfter the character revision described above, *Chaohusaurus brevifemoralis* was added to the data matrix and coded. The following characters in other Early Triassic taxa were recoded to reflect updated information through direct observations and the literature. The new coding are: *Cartorhynchus lenticarpus*, 36(1), 37(1), *Chaohusaurus geishanensis*, 22(1), 36(1), 59(0/1), 63(0); *Chaohusaurus chaoxianensis*, 22(1), 26(1), 28(1), 36(1), 37(1), 42(?), 46(?), 59(0/1), 63(0), 71(?), 93(1), 119(1), 172(?), 173(0), 174(0), 196(0), 197(0); *Chaohusaurus zhangjiawanensis*, 26(1), 28(?), 37(1), 116(0), 135(0); *Hupehsuchus nanchangensis*, 34(0). Also, dental characters were recoded for nasorostrans, and Characters 168, Spatium interosseum between tibia and fibula presence, and 169, Hind fin leading edge element in adults, were recoded to comply with a previous study ([@ref-10]). The revised character matrix is provided in [Data S1](#supp-1){ref-type=\"supplementary-material\"}.\n\nThe resulting character-taxon matrix was analyzed by TNT 1.5 ([@ref-9]) and PRAP2 ([@ref-37]) with PAUP\\* 4b10 ([@ref-41]). Searches in TNT used a combination of options defined by \"xmult = hit 100 replications 100 drift 10 hold 10.\" Searches in PRAP2 were based on default setting of 200 replicates of rachet searches. The Bremer support values were calculated in TNT, using a command \"bsupport !!+0 1.\" Bootstrap values were also calculated in TNT, based on 10,000 replicates.\n\nSexing\n------\n\nWe are not providing new results on sexing of the specimens over what were recently published ([@ref-26]). However, it may be useful to summarize the rationale behind the published sexing results that are adopted in the present paper. [@ref-26] found that two taxa, herein identified as *Chaohusaurus chaoxianensis* and *Chaohusaurus* *brevifemoralis*, each contained two morphotypes with long and short limbs, respectively. Such intraspecific dimorphism in relative limb length to the body is commonly known in multiple marine tetrapods, where males have longer limbs than females but never the other way around. This phenomenon is rationalized by the notion of the \"organ of prehension\" laid out by Darwin. That is, males usually have organs to hold females during courtship, which he called \"organ of prehension,\" and elongated limbs often evolve to serve that purpose. Based on this rationale, [@ref-26] identified the morphs with elongated limbs as males. See [@ref-26] for the details.\n\nStatistical analysis\n--------------------\n\nAs stated earlier, the purpose of the present paper is purely to describe a new species that was previously established as a distinctive morphotype based on both qualitative and quantitative characters, aided by statistical analyses ([@ref-26]). It would be repetitive to report the analyses again, so the readers are advised to refer to that paper for the statistical backgrounds.\n\nSystematic paleontology\n=======================\n\nIchthyosauromorpha [@ref-28]Ichthyosauriformes [@ref-28]*Chaohusaurus* Young and Dong 1972*Chaohusaurus chaoxianensis* ([@ref-1])\n\n*Anhuisaurus chaoxianensis*, Chen, 1985\n\n*Anhuisaurus faciles*, Chen, 1985\n\n*Chensaurus chaoxianensis* Mazin et al. 1991\n\n*Chensaurus faciles* Mazin et al. 1991\n\n*Chensaurus chaoxianensis* Motani and You 1998a, in part\n\n*Chaohusaurus geishanensis* Motani and You 1998b, in part\n\n*Chaohusaurus chaoxianensis* Motani et al. 2015c, in part\n\n**Holotype.** AGM AGB2905 (previously referred to by a field number P45-H85-25).\n\n**Paratype.** P45-H85-24 (whereabouts unknown).\n\n**Referred specimens.** AGM AGB2906 (P45-H85-20, holotype of *A. faciles* [@ref-1]), 5855, 6252, 6259, 6261, 6262, 6608, 6609, 7404, 7409, 7413; NGM P45-H85-21, P45-H85-23 (paratype of *A. faciles* [@ref-1]); IVPP V11362.\n\n**Revised diagnosis.** Narrow notch between two anterior sub-flanges of humerus; antero-proximal flange of radius well differentiated from shaft; ulnar distal fan nearly symmetrical relative to bone axis; only two tarsal ossifications in most individuals except largest; caudal peak neural spine not bifurcated ([Fig. 3](#fig-3){ref-type=\"fig\"}).\n\n![Comparison of diagnostic features between *Chaohusaurus brevifemoralis* sp. nov and *C. chaoxianensis*.\\\n(A) Anterior flange of the humerus, which is poorly developed in *C. brevifemoralis* (AGB 6258) but well-developed in *C. chaoxianensis* (AGB 6256) to leave a notch in the middle. (B) Anterior flange of the radius, which is poorly developed in *C. brevifemoralis* (AGB 7403, paratype) but well-developed in *C. chaoxianensis* (AGB 2905, holotype). (C) Anterior expansion of the distal ulnar shaft, which is present in *C. brevifemoralis* (AGB 6260) and makes the distal fan of the ulna appear asymmetrical relative to the ulnar axis (note: usually smaller than in the specimen figured here) but absent in *C. chaoxianensis* (AGB 6262), leaving the distal fan appear symmetrical relative to the ulnar axis. (D) Distal tarsal, at least one of which is always present in *C. brevifemoralis* (AGB 7401, holotype) but is absent in most specimens of *C. chaoxianensis* (AGB 2905, holotype) except the largest individuals. (E) Caudal peak neural spine (middle one of the five figured), which is bifurcated in *C. brevifemoralis* (AGB 7401, holotype) but remains single in *C. chaoxianensis* (AGB 5855). Red arrows point to apomorphic character states. Orange lines in (E) describe the inclination angles used in identification of the first anticlined neural spine explained in text. Scale bars are one cm.](peerj-07-7561-g003){#fig-3}\n\n**Remarks.** The caudal peak neural spine of the holotype may appear bifurcated due to a damage. However, the impression left on the matrix suggests that it was not bifurcated in life. The morphotype representing this species was referred to as Type A by [@ref-26].\n\n*Chaohusaurus brevifemoralis* sp. nov.\n\n*Chensaurus chaoxianensis* Motani and You 1998a, in part\n\n*Chaohusaurus geishanensis* Motani and You 1998b, in part\n\n*Chaohusaurus chaoxianensis*, Motani et al. 2015c, in part\n\n*Chaohusaurus chaoxianensis*, Zhou et al. 2017, in part\n\n**Etymology.** The specific name refers to the shortness of the femur relative to the body in comparison to other species.\n\n**Holotype.** AGM AGB7401. A complete skeleton of a male individual, lacking only the tip of the tail. From bed 621. See [Figs. 2](#fig-2){ref-type=\"fig\"} and [4](#fig-4){ref-type=\"fig\"}.\n\n![Skull of the holotype of *Chaohusaurus brevifemoralis* sp. nov (AGB7401).\\\n(A) Dorsal view. (B) Lateral and slightly dorsal view. (C) Approximate bone map for (A). (D) Same for (D). See the section \"Osteological abbreviations\" for abbreviations. Scale bar is one cm in total.](peerj-07-7561-g004){#fig-4}\n\n**Paratypes.** GMPKU P-3086. A nearly complete skeleton of a female individual lacking the tip of the snout and most of the forelimb and pedal phalanges. From bed 633. The specimen was recently figured ([@ref-47]). See also [Fig. 5](#fig-5){ref-type=\"fig\"}. AGB7403. A partially disarticulated skull and the upper body down to the forelimb ([Fig. 6](#fig-6){ref-type=\"fig\"}). From bed 628.\n\n![Skull of *Chaohusaurus brevifemoralis* sp. nov in one of the paratypes (GMPKU P-3086).\\\n(A) Ventral view of the mandible and lateral view of the right side of the skull. (B) Right mandibular ramus from dorso-lateral direction. (C) Approximately lateral view of the left side of the skull. (D) Approximate bone map for (A). (E) Same for (B). (F) Same for (C). See the section \"Osteological abbreviations\" for abbreviations. Scale bar is one cm in total.](peerj-07-7561-g005){#fig-5}\n\n![Skull of *Chaohusaurus brevifemoralis* sp. nov in one of the paratypes (AGB7403).\\\n(A) Planar view of the specimen. (B) Approximate bone map for (A). (C) Reconstruction of the parasphenoid-basisphenoid complex. See the section \"Osteological abbreviations\" for abbreviations. Short scale bars are one cm each.](peerj-07-7561-g006){#fig-6}\n\n**Referred specimens.** AGM AGB5846a, 5846b, 5846c, 6253, 6254, 6255, 6258, 6260, 6605, 7402, 7407, 7408, 7410, MT10022; GMPKU P-1101, P-3093; IVPP V11361.\n\n**Ambiguous specimens.** AGB6607 and 7400 are not included in the formal list of the specimens but may belong to *Chaohusaurus brevifemoralis*. As reported elsewhere ([@ref-26]), both share the same quantitative scaling trends and qualitative features with *Chaohusaurus brevifemoralis*. Yet, they appear osteologically immature compared to specimens of the same sizes, that is, their bones are slender, and extremities of long bones are not fully formed. At least one of them (AGB6607) is large enough to be an adult of *Chaohusaurus brevifemoralis*. Based on this inferred immaturity, we are not including them in the list of referred specimens at this point.\n\n**Locality.** Majiashan, Chaohu, Anhui Province, China.\n\n**Horizons.** Ammonoid *Subcolumbites* zone, Spathian, Lower Triassic. Known specimens are from beds 621 to 638 (about 248.53--248.34 Ma) that were previously dated using astrochronology ([@ref-8]).\n\n**Diagnosis.** Humeral anterior flange poorly developed, with weakly concave preaxial margin near midshaft; radial antero-proximal flange poorly developed; ulnar distal fan asymmetrical relative to bone axis, due to anterior expansion of distal preaxial margin; femur short for trunk length in comparison to *Chaohusaurus chaoxianensis*; tibia proximally narrow for trunk length in comparison to *Chaohusaurus chaoxianensis*; bifurcated neural spine near caudal peak; three tarsal ossifications in most individuals except newborns ([Figs. 3](#fig-3){ref-type=\"fig\"} and [7](#fig-7){ref-type=\"fig\"}).\n\n![Femoral length plotted against trunk length.\\\nF, female; M, male. Within each species, males tend to have longer femora for the trunk length than females, while *C. chaoxianensis* tends to have longer femora for the trunk length than *C. brevifemoralis*, especially when compared within the same sex. Polygons represent convex hulls for each species. Based on [@ref-26].](peerj-07-7561-g007){#fig-7}\n\n**Remarks.** The morphotype representing this species was established as Type B by [@ref-26]. Quantitative and qualitative comparisons with *Chaohusaurus chaoxianensis* (Type A) are found in that paper.\n\nResults\n=======\n\nMorphological descriptions\n--------------------------\n\nThe body of *Chaohusaurus brevifemoralis* sp. nov is usually preserved in a curled posture, as seen in the holotype ([Fig. 2](#fig-2){ref-type=\"fig\"}). This tendency for curled posture is shared with *Chaohusaurus chaoxianensis* ([Fig. 1](#fig-1){ref-type=\"fig\"}), and probably with other species of the genus. To the best of our knowledge, there are only two specimens of *Chaohusaurus* that are preserved in approximately straight posture, one of which was previously figured ([@ref-36]).\n\nThe body proportion in adults is similar to those of non-hueneosaur ichthyopterygians---the tail occupies about half of the total length, and the skull is usually about a quarter of the precaudal length (i.e., about one eighth of the total length). The relative size of the skull to the total length is much larger in long-snouted hueneosaur ichthyopterygians, where the values scatter around 20--30%.\n\nThe sizes of the major structures in the holotype and four more specimens are summarized in [Table 1](#table-1){ref-type=\"table\"} and compared to the same metrics for four specimens of *Chaohusaurus chaoxianensis*. To avoid redundancy, absolute sizes are not reported in text.\n\n10.7717/peerj.7561/table-1\n\n###### Measurements of major structures in *Chaohusaurus brevifemoralis* in females and males of different sizes.\n\n![](peerj-07-7561-g013)\n\n  Species                             *C. brevifemoralis*   *C. chaoxianensis*                                                                                    \n  ----------------------------------- --------------------- -------------------- ----------- ---------- ----------- ---------- ---------- ----------- ----------- -----------\n  Length along the vertebral column                         (709+)               (816+)                                                               542         505\n  SVL                                 238.9                 442                  529                    578                                           305         288\n  Trunk length                        168                   334                  408.1       415.3      469         136.6      261.7      294.13      229.38      201\n  Presacral vertebral count                                 36                                          37                                36          35          36\n  Sacral vertebral count                                    3                                                                                                     \n  Preserved caudal vertebral count                          (36+)                                                                                                 \n  Skull length                        70.86                 114.61               120.9                  111.75      58.97                             75.62       87\n  Orbit length                        20.71                 35.07                            25.82      32.57                  25.73                              26.1\n  Scleral ring aperture length                              16.2                                                                                                  \n  UTF maximum diameter                                      15.3                             (18.49\u2212)                                                             \n  External naris length                                     8.63                                        7.53                                          5.18        6.01\n  Humerus length                      7.02                  23.22                            18.7       25.59       6.97       19.56      20.38       14.03       16.5\n  Humerus distal width                6.04                  15.72                            13.4       20.53       5.27                  15.44       11.2        10.04\n  Radius length                       7.74                  21.69                24.02       16.32      20.96       6.64       15.53      18.35       13.94       13.5\n  Radius proximal width               4.76                  14.71                13.34       11.42                  4.54       10.42      13.34       9.18        9.38\n  Ulna length                         7.94                  20.52                24.79                  22.31       6.87       15.38      17.46       14.03       12.46\n  Ulna distal width                   3.68                  15.98                11.88                  16.47       4.77       8.36       13.84       9.51        7.35\n  Intermedium maximum diameter        1.29                  6.38                 6.17                   7.66                   0.89       6.05        4.54        3.39\n  Femur length                        5.22                                       19.38       17.2       20.39                  11.9       14.93       8.68        9.43\n  Femur distal width                  4.75                                       13.88       10.32      15.81                  8.17       10.47       6.73        9.21\n  Tibia length                        6.01                  17.99                            15.93                             10.14      14.38       9.61        9.87\n  Tibia proximal width                3.42                  6.98                             7.06                              5.5        7.28        5.38        4.36\n  Fibula length                       7.38                  19.13                20.12       18.01                             11.42      15.84       10.25       10.02\n  Fibula distal width                 4.4                   12.01                14.22       12.06                             7.61       10.78       7.48        7.01\n  Astragalus maximum diameter         1.49                  6.66                             6.77                              3.92       4.96        4.1         3.62\n  Specimen note                       Smallest              Holotype             Largest M   Paratype   Largest F   Smallest   Holotype   Largest M   Typical F   Typical M\n\n### Cranium\n\nCranial suture patterns are given in [Figs. 4](#fig-4){ref-type=\"fig\"}--[6](#fig-6){ref-type=\"fig\"}, so the contact patterns among bones are minimally reported in the description below to save space. Exceptions are bones surrounding major fenestrae, and ambiguous or polymorphic cases. Unless otherwise stated, descriptions of cranial and mandibular elements below are based on the holotype ([Fig. 4](#fig-4){ref-type=\"fig\"}) and the paratypes ([Figs. 5](#fig-5){ref-type=\"fig\"} and [6](#fig-6){ref-type=\"fig\"}), one of which revealing some hidden parts of cranial elements through partial disarticulation.\n\n**Premaxilla.** The premaxilla is a slender bone that occupies most of the snout in lateral view, where its length is slightly less than half of the skull length. Its extent along the midline is less because the nasals extend anteriorly in-between the two premaxillae. The premaxilla is approximately triangular but has short sub- and supranarial processes emerging from the ventral and dorsal corners of the posterior margin, which forms the anterior margin of the external naris. One of these processes may be missing in some specimens without a clear pattern. The premaxilla bears teeth, as described below.\n\n**Maxilla.** The maxilla is about half as long as the premaxilla along the jaw margin. It has an ascending postnarial process that excludes the lacrimal from the margin of the external naris. This process is robust, with its base width rivaling the antero-posterior length of the external naris. The suture with the lacrimal appears smooth. The maxilla forms the posterior and postero-ventral margins of the external naris, forming a continuous curve that contributes to about a third of the oval perimeter of the opening. The anterior extent of the maxilla in lateral view is slightly anterior to the external naris, while the posterior end of the bone in lateral view overlaps the anteriormost part of the orbit narrowly. The maxilla bears teeth, as described below.\n\n**Nasal.** The nasal is a major bone that forms a part of the skull roof from the anterior orbital to prenarial region, behind the major part of the snout. It is widest near the lacrimal, and tapers to a point both posteriorly and anteriorly. The posterior ends of the right and left nasals are widely separated by the frontals while the anterior ends meet along the sagittal line. In the holotype, the right nasal appears to extend far more anteriorly than the left nasal, but this is likely an artifact of preservation ([Figs. 4A](#fig-4){ref-type=\"fig\"} and [4C](#fig-4){ref-type=\"fig\"}).\n\n**Frontal.** The frontal is a major bone that forms the skull roof in the orbital region. Its shape is approximately an elongated triangle in dorsal view. Its postero-lateral process narrowly participates in the antero-lateral margin of the upper temporal fenestra, forming the most medial segment of the anterior terrace of the fenestra ([Fig. 4](#fig-4){ref-type=\"fig\"}). The bone is excavated in the terrace region, being separated from the roof part by a conspicuous ridge. The ridge and terrace are confluent with the corresponding structures of the postfrontal.\n\n**Parietal.** The parietal has the smallest exposure of the bone of the skull roof, comprising the main corpus forming the posteriormost part of the skull roof and a supratemporal process that extends postero-laterally from the corpus. The corpus appears small compared to those of most ichthyopterygians in its relative size to the skull, in width and length. There is a ridge at the posterior end of the skull roof, followed by a slope that is better seen in occipital view than dorsally. They likely represent the parietal ridge and slope that are present in mixosaurs and parvipelvians ([@ref-23]). The inter-parietal suture is not straight: it is bent once in front of the pineal foramen and deeply interdigitated posterior to the foramen ([Figs. 4](#fig-4){ref-type=\"fig\"} and [6](#fig-6){ref-type=\"fig\"}). It is evident that there is a descending ventral flange along the margin of the upper temporal fenestra, as expected in neodiapsids. The flange is separated from the roof by a conspicuous ridge.\n\n**Postfrontal.** The postfrontal forms a major part of the bar between the orbit and upper temporal fenestra. It is posterior two-thirds are excavated to form most of the anterior terrace of the upper temporal fenestra ([Fig. 4](#fig-4){ref-type=\"fig\"}). This proportion of excavated area is unusually high compared to ichthyopterygians. The excavated terrace is separated from the skull roof by a sharp and conspicuous ridge that continues into the frontal.\n\n**Prefrontal.** The prefrontal is the largest of the bones surrounding the orbit ([Figs. 4](#fig-4){ref-type=\"fig\"}--[6](#fig-6){ref-type=\"fig\"}). It has a structure often referred to as an \"eyebrow,\" a small shelf above the antero-dorsal corner of the orbit. This feature is shared with basal ichthyopterygians, such as *U. hataii* and *Grippia longirostris* ([@ref-15]; [@ref-33]; [@ref-24]; [@ref-7]). The bone is swollen laterally around the base of this structure, making it appear unusually massive for a cranial dermal bone.\n\n**Lacrimal.** The lacrimal appears as a narrow strip of bone between the maxilla and prefrontal in lateral view ([Figs. 4](#fig-4){ref-type=\"fig\"} and [5](#fig-5){ref-type=\"fig\"}). It is wider away from the orbit. See Orbit below for its participation in the structure.\n\n**Jugal.** The jugal is a narrow and J-shaped bone in lateral view ([Figs. 4](#fig-4){ref-type=\"fig\"}--[6](#fig-6){ref-type=\"fig\"}) but there is indeed much three-dimensionality to the morphology of the bone. The bone is most robust near the postero-ventral corner, where it has a thick and rounded cross-section. From there extends a narrow and horizontal plate of bone anteriorly, forming the maxillary process along the antero-ventral margin of the orbit. This plate widens anteriorly along the orbital rim. This ventral margin of the orbit formed by the jugal is continuous with the anterior wall of the orbit formed by the prefrontal, together giving depths to the orbital rim. Postero-dorsal to the postero-ventral corner extends the postorbital process. Unlike the maxillary process that is flat horizontally, the postorbital process is flat vertically. The postorbital process articulates mainly with the postorbital but it also contacts the quadratojugal and may be partially overlapped by the squamosal.\n\n**Postorbital.** The postorbital is best seen in AGB7403, where the bone is disarticulated from the rest of the skull ([Fig. 6](#fig-6){ref-type=\"fig\"}). Its general outline may be compared to the bottom half of the waxing crescent. There are two short processes extending antero-dorsally (postfrontal process) and postero-dorsally (supratemporal process), respectively.\n\n**Supratemporal.** The supratemporal is a large bone that forms the postero-lateral corner of the upper temporal fenestra. It resembles the squamosal of neodiapsids. it is essentially a triradiate bone, with the postorbital process extending anteriorly, parietal process medially, and descending process ventrally. The descending process is excavated to allow acceptance of the dorsal end of the quadrate, as in *Ichthyosaurus* ([@ref-17]).\n\n**Squamosal.** The identity of squamosal and quadratojugal requires future scrutiny because some specimens show just one of them each. The following description is based on the tentative identification given in [Figs. 4](#fig-4){ref-type=\"fig\"}--[6](#fig-6){ref-type=\"fig\"} with question marks. The squamosal in *Chaohusaurus* spp. resembles that described for *Mixosaurs atavus* in that it is slightly swollen laterally and occupies the middle part along the height of the cheek, unlike in most diapsids where it is located dorsally. The bone covers parts of the supratemporal and quadrate superficially, and sometimes also parts of the postorbital and jugal. The bone appearsmissing in some specimens, possibly reflecting its superficial position that makes it easy for the bone to flake off. However, analyses of 3D morphology are necessary to clarify if the bone is truly missing in these specimens. This bone was identified in several different locations and with different shapes depending on the specimen in a previous publication ([@ref-47]). One of them, figured in their [Fig. 4B](#fig-4){ref-type=\"fig\"}, matches the current description (see \"Discussion\").\n\n**Quadratojugal.** Again, the identity of squamosal and quadratojugal requires future scrutiny because one of the bones may not be clearly present in some specimens (see \"Squamosal\"). The quadratojugal is a vertically elongated plate of bone that is slightly broader dorsally than ventrally. It is constricted above the ventral end, where the bone is expanded and thickened to form a cup that sits on the dorso-lateral part of the quadrate condyle, as in *Ichthyosaurus* ([@ref-17]).\n\n**Quadrate.** The outline of the quadrate in postero-medial view may be compared to that of a human auricle, with the quadrate condyle occupying the position of the lobule of the auricle. This shape is most evident in the right element in AGB7403 that has been disarticulated and exposed in postero-medial view ([Fig. 6](#fig-6){ref-type=\"fig\"}). The condyle has two ridges separated by a shallow groove, as in most reptiles ([@ref-39]). When in articulation, the bone is largely concealed by the quadratojugal that covers most of its lateral aspect, and the supratemporal that encloses its dorsal end.\n\n**Supraoccipital.** The supraoccipital is missing in the holotype but exposes its postero-dorsal view in AGB7403, being disarticulated from but associated with the parietals ([Fig. 6](#fig-6){ref-type=\"fig\"}). It is a large bone for the occiput, being slightly larger than the roof part of a single parietal. Its antero-dorsal margin is convex, and postero-ventral margin concave. The lateral margins are approximately straight. The shape of the convex margin approximately fits that of the concave posterior margin of the parietal pair, to which it is supposed to articulate. No articular facets are seen on the exposed side, as expected from the condition in other ichthyosauriforms.\n\n**Exoccipital.** There is a pair of exoccipitals preserved behind the skull of AGB7403, in disarticulation ([Fig. 6](#fig-6){ref-type=\"fig\"}). The exoccipital resemble that described for *Ichthyosaurus* in having a shape somewhat reminiscent of a boot in lateral view; it is wider ventrally than the dorsally, the posterior margin is approximately vertical, and the anterior margin is strongly concave ([@ref-17]). The toe of the boot is conspicuously curled-up in *Ichthyosaurus*, but the curling is not as pronounced in the present form.\n\n**Basioccipital.** The basioccipital of *Chaohusaurus* in general is largely unknown. In AGB7403, there is a bone between the skull and the atlantal pleurocentrum that may represent the basioccipital, being exposed in what appears to be the dorsal view given the presence of a pair of articular facets, probably for the exoccipitals ([Fig. 6](#fig-6){ref-type=\"fig\"}). The bone is longer than the atlantal pleurocentrum antero-posteriorly, and its length approximately matches the distance between the posterior margins of the parasphenoid and basisphenoid that will be mentioned below.\n\n**Basisphenoid.** The basisphenoid is known from only one specimen, AGB7403, which exposes more than half the bone, dislocated to be behind the supraoccipital ([Fig. 6](#fig-6){ref-type=\"fig\"}). There is no sign of a major foramen in the exposed side, so it is most likely the ventral side. [Figure 6C](#fig-6){ref-type=\"fig\"} represents the estimation of the bone outline through mirroring of the exposed half. The reconstructed shape is reminiscent of what was drawn for *Limnoscelis* ([@ref-39]). Unlike in *Ichthyosaurus*, the basisphenoid of *Chaohusaurus* is not fused to the parasphenoid. There is a pair of processes pointing postero-laterally from the posterior corners of the bone, and this process continues ventrally to the main body of the bone where they become side walls of a triangular area that is excavated from the ventral side. This area seems to serve as the socket for the main body of the parasphenoid. The triangular area matches the shape of the main body of the parasphenoid ([Fig. 6C](#fig-6){ref-type=\"fig\"}). Conspicuous striations are seen lateral to the triangular area, resembling those that are reported for the parasphenoid below. The basipterygoid process is short, broad, and directly laterally, not antero-laterally as in most reptiles. The anterior surface of the basisphenoid is slightly concave.\n\n**Parasphenoid.** The parasphenoid is best seen in AGB7403 ([Fig. 6](#fig-6){ref-type=\"fig\"}). It is a large bone that covers the base of the braincase region, comprising a long cultriform process and a fan-shaped main body. The anterior half of the fan fits into a triangular socket at the ventral part of the basisphenoid, while the posterior half extends more posteriorly than the main body of the basisphenoid, almost reaching the tip of the postero-lateral process of the basisphenoid ([Fig. 6C](#fig-6){ref-type=\"fig\"}). The main body is covered by strong striations that radiate from the base of the cultriform process. When articulating the parasphenoid to the basisphenoid, these striations would become a part of a series with the striations reported above for the basisphenoid. There is a pair of swelling at the base of the cultriform process, which are probably basal tubera ([@ref-39]).\n\n**Pterygoid.** The pterygoid is best seen in GMPKU P-3086 ([Fig. 5A](#fig-5){ref-type=\"fig\"}). It is similar to the pterygoid of basal ichthyopterygians, such as *Grippia longirostris*. There is a transverse flange marked by a weak ridge that runs from postero-medial to antero-lateral. Pterygoidal teeth are not present. The part posterior to this ridge is elevated and then becomes slightly twisted toward the posterior end, forming the quadrate ramus. The anterior part tapers to a point along the sagittal plane. As with the pterygoid, there is no sign of the margin for the suborbital fenestra.\n\n**Palatine.** GMPKU P-3086 has a part of the palatine exposed, showing that it is located in its usual position in ichthyopterygians, that is, antero-lateral to the anterior ramus of the pterygoid ([Fig. 5A](#fig-5){ref-type=\"fig\"}). There is no evidence of the suborbital fenestra in the exposed part of this bone. Otherwise, the exposure is too limited to allow morphological descriptions.\n\n**Orbit.** The orbit is large and occupies a substantial part of the postnarial skull, with a maximum diameter that is about twice as large as that of the upper temporal fenestra. The orbital rim does not form a circle, ellipse, or ovoid that would fit onto a plane as in ichthyosaurians. Instead, the rim exhibits a conspicuous medial excursion in the dorsal part, off the lateral plane formed by the remainder of the rim ([Figs. 4A](#fig-4){ref-type=\"fig\"} and [4C](#fig-4){ref-type=\"fig\"}). The excursion corresponds to the constriction of the skull roof near the narrow contact between the prefrontal and postfrontal. Similar constriction of the skull roof and an associated excursion of the orbital rim have been recognized in *Grippia longirostris*, *U. hataii* and *Cartorhynchus lenticarpus* ([@ref-33]; [@ref-24]; [@ref-28]) and seems to be a basal ichthyosauriform feature. The excursion makes the eyeball vulnerable to stresses from the dorsal aspect. This vulnerability is partly remedied by the extensive growth of the scleral ring, as mentioned below.\n\nThe orbital rim is formed by the prefrontal anteriorly and antero-dorsally, postfrontal postero-dorsally, postorbital posteriorly, jugal ventrally, and slightly by the lacrimal antero-ventrally. The participation by the lacrimal may not be evident in [Fig. 4D](#fig-4){ref-type=\"fig\"} but it is largely because of the posterior side of the prefrontal being visible in lateral view, making the orbital rim smaller than it is. It is evident in AGB7403 ([Fig. 6](#fig-6){ref-type=\"fig\"}). The frontal is excluded from the orbital rim by a narrow contact between the pre- and postfrontals.\n\n**External Naris.** The normal axis for the planed formed by the margin of the external naris points dorso-laterally and slightly anteriorly, being directed more dorsally than laterally in the holotype. The opening is elongated, and surrounded by the maxilla ventrally and posteriorly, nasal dorsally, and premaxilla anteriorly. The lacrimal is completely excluded from the margin, as in basal ichthyopterygians.\n\n**Upper Temporal Fenestra.** The upper temporal fenestra extends from antero-medial to postero-lateral direction, with its parasagittal extent being less than the lateral extent ([Fig. 4](#fig-4){ref-type=\"fig\"}). The skull of *Chaohusaurus* does not appear narrow despite the smallness of the parietals thanks to the wide upper temporal fenestra. It is bordered by the parietal medially and posteromedially, supratemporal posterolaterally, postorbital antero-laterally, postfrontal anteriorly and slightly by frontal anteromedially. There is no evidence for the squamosal to participate in the margin of the fenestra.\n\n**Lower Temporal Fenestra.** The lower temporal fenestra is a prominent structure in the cheek that is vertically elongated and approximately wedge-shaped. It is completely open ventrally. This morphology is shared with basal ichthyosauromorphs, including *Cartorhynchus lenticarpus* and Hupehsuchia ([@ref-4]; [@ref-28]; [@ref-44]), but not with ichthyopterygians.\n\n**Pineal Foramen.** The pineal foramen is completely enclosed between the right and left parietals. Its center is located posterior to the midpoint along the inter-parietal suture. Its shape is ovoid, being wider posteriorly than anteriorly. It is located behind the line connecting the posterior margins of the orbits in the holotype, but the location may differ depending on the direction of compaction experienced by relevant specimens.\n\n**Scleral ring.** In parvipelvian ichthyosaurs, the scleral ring has a central part that is shaped like the wall of a low conical frustum, with the aperture corresponding to the top of the frustum, and a peripheral part that continues from the base of the frustum to form a round wall around the axis of the central frustum (which is expected to be nearly parallel to the optical axis of the eyeball) ([@ref-25]). The central frustum and peripheral wall form an angle of about 110\u00b0 to 130\u00b0 ([@ref-25]; [@ref-19]). The center of the frustum part may protrude beyond the plane formed by the orbital margin in life. The central part, however, is preserved flattened in most specimens because of preservational compaction, and appear as a flat ring rather than the wall of a conical frustrum. The scleral ring is very thin and deformable even in extant specimens of reptiles, and can easily change its shape plastically through preservation.\n\nThis basic arrangement is already present in *Chaohusaurus brevifemoralis*, although the peripheral wall appears shorter in the present form than in Jurassic forms. The right scleral ring of the holotype suggests that the central frustum formed an angle of about 120\u00b0 with the peripheral wall, giving depth to the scleral ring. The left ring has a flattened central part, making it appear as if the central part was completely parallel to the plane formed by the orbital margin, but this is most likely an artifact of preservation as in many parvipelvian specimens. The scleral ring is the largest bony structure in the cranial region, apart from the entire skull itself, filling the large orbit ([Fig. 4](#fig-4){ref-type=\"fig\"}). The aperture of the ring is also large. A part of the dorsal aspect of the ring is exposed through the constriction of the skull where the medial excursion of the orbital rim occurs. It is seen there that the peripheral wall of the scleral ring fills the embayment of the skull roof ([Figs. 4A](#fig-4){ref-type=\"fig\"} and [4C](#fig-4){ref-type=\"fig\"}), solidifying the eyeball that would otherwise lack bony support in the dorsal direction. The ring is inclined, with the normal axis pointing mostly laterally but also tilted anteriorly and dorsally to some extent. The left scleral ring of the holotype seems to comprise about 16 plates, although it is difficult to arrive at the exact number because the preservation is not perfect. There seem to be 18 plates in the right scleral ring of GMPKU P-3086 but the number, again, is not conclusive.\n\n### Mandible\n\nThe mandible most likely comprises seven elements, including the coronoid that is lacking in derived ichthyosaurs ([@ref-17]). The mandibular rami are slightly curved and inclined in natural posture, with the anterior part more vertical than the posterior part that is half horizontal. The lateral side of the mandible appears flat in specimens with disarticulated jaws but this is because of flattening by compaction.\n\n**Dentary.** As with ichthyopterygians, the dentary is an elongated bone that is longer than half the length of the mandible. It is the only dentigerous bone in the lower jaw. The presence of a dental groove is not clearly established because the medial side of the dentary is rarely exposed. However, it is seen in AGB7403 that the teeth are medial to a bony wall, as seen in subthecodont ichthyopterygians. See below for its dentition.\n\n**Surangular.** The surangular is the most prominent bone of the mandible in lateral view, occupying most of the posterior half of the structure ([Figs. 4](#fig-4){ref-type=\"fig\"} and [6](#fig-6){ref-type=\"fig\"}). It extends anteriorly beyond the anterior margin of the external naris. Posteriorly, it supports the lateral aspect of the articular, which forms the jaw articulation with the quadrate. There seems to be no direct contact with the quadrate.\n\n**Angular.** The angular is an elongated bone that covers the ventral aspect of the mandible, from the posterior end to about the midpoint. It is visible both laterally and medially. However, its lateral exposure is limited, unlike in *Cartorhynchus lenticarpus* and many basal neodiapsids where it occupies the ventral half or more of the posterior part of the lateral surface ([@ref-28]).\n\n**Articular.** The articular is barely visible in lateral view, although its dorsal margin may be seen as a narrow band surrounding the poster-dorsal corner of the surangular in some specimens ([Fig. 6](#fig-6){ref-type=\"fig\"}). The articulation with the quadrate condyle seems to occur with the postero-medial surface of the articular, unlike in *Ichthyosaurus* where the articulation is through the antero-medial surface ([@ref-17]). Thus, there is no retroarticular process of the articular in *Chaohusaurus brevifemoralis*.\n\n**Splenial.** The splenial is a thin and narrow sheet of bone that lies on the medial side of the dentary. Its complete shape is not seen in any of the specimens available, although GMPKU P-3086 shows most of the bone as previously figured ([@ref-47]). The bone tapers to a point posteriorly, and probably anteriorly. The bone is longer than the dentary and seems to extend over about three quarters of the mandibular length.\n\n**Coronoid.** The coronoid is one of the most enigmatic bones in *Chaohusaurus*. The right side of GMPKU P-3086 and the left side of AGB7403 reveal sutures between the coronoid and surangular that resemble each other closely ([Figs. 5](#fig-5){ref-type=\"fig\"} and [6](#fig-6){ref-type=\"fig\"}). However, this suture seems to be closed on the left side of GMPKU P-3086 and most other specimens. The coronoid occupies the area surrounding the coronoid process, mainly dorsally and medially. The bone is also visible slightly in lateral view. The bone is wide dorsally, forming a shallow basin between the coronoid process and the articular. The width of the bone is evident by comparing the lateral ([Fig. 5A](#fig-5){ref-type=\"fig\"}) and more dorsal ([Fig. 5B](#fig-5){ref-type=\"fig\"}) views of the right coronoid of GMPKU P-3086. The coronoid process is very weakly developed.\n\n**Prearticular.** The prearticular is known from partial exposures only. The right mandibular ramus of AGB7403 is exposed laterally but medial elements have slipped ventrally relative to the lateral elements, exposing their lateral sides ([Fig. 6](#fig-6){ref-type=\"fig\"}). It is seen there that the anterior ramus of the prearticular is long and slender, tapering anteriorly to a point. Its anterior extent overlaps the most posterior part of the dentary.\n\n### Dentition\n\nThere are 35 teeth in the left premaxilla of the holotype, leaving about three vacant tooth positions among them. Nine teeth are recognized in the left maxilla. In total, there are 47 tooth positions in the left upper jaw ramus of the holotype. The count is similar in AGB7403, which has ten teeth in the right maxilla, and 29 teeth and about five vacant tooth positions in the left premaxilla. the tooth counts in the left maxilla and right premaxilla are uncertain. Both specimens are male individuals, so the slight difference in the total counts likely reflects individual variation. The mandibular teeth are not exposed in the holotype, but they can be counted in AGB7403, where the dentary has 42 teeth and at least eight vacant tooth positions, making the total tooth count to be about 50. This number is similar to the mandibular tooth count in *U. hataii*, which was previously estimated to be approximately 50 ([@ref-20]).\n\nThe dentition is heterodont, with slender and conical anterior teeth and robust and rounded posterior teeth. In the left maxilla of the holotype, the posteriormost six, of the nine teeth, are robust and rounded, while the anterior three appear similar to the premaxillary teeth in size and shape. The transition between the two tooth types is abrupt. The robust posterior teeth are two to three times wider in diameter than the anterior teeth. The root seems to expand basally and the pulp cavity is open, as in basal ichthyopterygians.\n\n### Hyoid\n\n**Ceratohyal.** There is a pair of ceratohyal, located in-between the mandibular rami near their posterior ends. They most likely represent the first ceratohyals of reptiles ([@ref-27]). The ceratohyal I is a small and slender rod, which is unsuitable for suction feeding ([@ref-27]).\n\n### Vertebral column\n\n**Atlas.** The neural arch and pleurocentrum of the atlas are exposed in the holotype (AGB7401). The neural arch is divided into the right and left halves, as in some reptiles, including ichthyosaurs ([@ref-19]). The pleurocentrum of the atlas has a convex articular surface ([Fig. 4](#fig-4){ref-type=\"fig\"}), which is most likely the anterior surface given the preserved orientation, and that similar pleurocentra are known for *U. hataii* and *Cymbospondylus piscosus*. These bones are comparable to the corresponding axial pleurocentra in size, and are unlikely to be the intercentra. Another specimen (AGB7403) exhibits the posterior view of the pleurocentrum from a slightly lateral angle, where the posterior articular facet is concave. The bone is about only half as long as the subsequent vertebral centra ([Fig. 6](#fig-6){ref-type=\"fig\"}). There is no bone that can be positively identified as the atlantal rib in any of the specimen, although AGB7043 has a short and slender bone near the atlas, possibly representing the atlantal rib.\n\n**Axis.** The axial neural spine is broader antero-posteriorly by about twice compared to the subsequent neural spines ([Fig. 6](#fig-6){ref-type=\"fig\"}). The neural arch has a rib facet laterally, forming the dorsal half of the synapophysis. The ventral half of the synapophysis is on the dorso-lateral part of the atlantal centrum, appearing almost triangular.\n\n**Neural arches and spines.** The neural spines change their shapes along the vertebral column. In the cervical and anterior dorsal series, the neural spines and arches are about equal in height with each other. Both the neural spine and arch are almost perpendicular to the basal plane of the neural arch, although the axis of the former is shifted posteriorly relative to that of the latter. The neural spines gradually become longer, broader, and inclined posteriorly while the neural arches stay in approximately the same posture and height, with antero-posterior broadening in the mid-dorsal series. The lengthening of the neural spines and steepening of the posterior inclination stop early in the posterior dorsal series, and the neural spine size and angle do not change greatly in the sacral series. Then, the steepening of the neural spines restarts in the caudal series, reaching the extreme inclination, where the neural spines are almost horizontal, by about the 13th caudal vertebra. Then the inclination trend reverses, with the neural spines rapidly becoming less inclined and then start to anticline (i.e., inclined anteriorly as opposed to the normal posterior inclination) at about the 18th vertebra.\n\nThe most notable feature of the neural spine is found in the caudal peak neural spine, which is bifurcated into two branches, one dorsal and the other antero-dorsal (cpn in [Figs. 3E](#fig-3){ref-type=\"fig\"} and [8](#fig-8){ref-type=\"fig\"}). As noted elsewhere ([@ref-26]), structures that are likely homologous with these two branches also exist in *Chaohusaurus chaoxianensis* as thickened bars but the two are connected by a thin flange of bones. That flange is absent *Chaohusaurus brevifemoralis*. The caudal peak vertebra in this context was identified based on the following criteria. First, it is the second vertebra from the last narrow caudal neural spine (lnn in [Figs. 3E](#fig-3){ref-type=\"fig\"} and [8](#fig-8){ref-type=\"fig\"}), which is inclined about 45\u00b0 posteriorly. The caudal neural spines preceding this last narrow one are steeply inclined posteriorly ([Fig. 8](#fig-8){ref-type=\"fig\"}), sometimes appearing nearly horizontal. Second, the caudal peak vertebra is in front of the first caudal neural spine to have a completely rounded distal end (fran in [Figs. 3E](#fig-3){ref-type=\"fig\"} and [8](#fig-8){ref-type=\"fig\"}), with clear anterior inclination of about 30--50\u00b0. Third, the caudal peak vertebra is usually placed where the curvature of the tail reaches its peak, although preservational bias may obscure this feature. Finally, at least in *Chaohusaurus brevifemoralis*, the caudal peak neural spine is the first anticline neural spine, although it is not necessarily the case in some specimens of *Chaohusaurus chaoxianensis*. Anticlination in this case is judged by the angle formed by two lines, one connecting the postero-dorsal corner of the neural spine with the postero-ventral corner of the neural arch, and the other connecting the antero-ventral and postero-ventral corners of the neural arch ([Fig. 3E](#fig-3){ref-type=\"fig\"}, orange lines). This angle is greater than 90\u00b0 (i.e., the neural spines are inclined posteriorly, which is the norm) in most neural spines except in the posterior part of the tail, and the first anticline neural spine is where the angle first becomes less than 90\u00b0 in the caudal series.\n\n![Caudal peak region of *Chaohusaurus brevifemoralis* sp. nov. in one of the paratypes (GMPKU P-3086).\\\nScale bar is one cm.](peerj-07-7561-g008){#fig-8}\n\n**Hemal arches and spines.** The hemal arches and spines are present starting from the third caudal vertebra to the tip of the tail. GMPKU P-3086 reveals the best representation of the hemal arches and spines in the mid-caudal region, where there is a caudal peak associated with anticlination of neural spine (see above). In front of the anticline neural spine, the hemal spines seem to be absent, while the hemal arch is U-shaped in antero-posterior view ([Fig. 8](#fig-8){ref-type=\"fig\"}; see fig. 1g of [@ref-26]). The hemal arches appear absent in the caudal peak region of the holotype but this lack probably reflects a preservational bias.\n\n**Ribs.** The first four ribs of the left side of the holotype have an expanded, fan-shaped head, which most likely represents unified capitulum and tubercle. These ribs are short and do not appear to participate in the formation of the trunk rib cage. The four pairs of ribs most likely represent cervical ribs. If so, there are five cervical vertebrae because the atlas seems to lack associated ribs.\n\nThe dorsal ribs are also single-headed, although their proximal ends typically have eight-shaped cross-section, suggesting the head is a combination of the capitulum and tubercle. Anterior and mid dorsal ribs articulate with both the neural arch and vertebral centra via a synapophysis, at least down to the 28th vertebra in the holotype. The articulation with the neural arch is lost somewhere in the posterior dorsal region, as the articular facet on the centrum moves ventrally to be completely located on the lateral side of the centrum alone. The exact location of this transition is obscure. The dorsal ribs become increasingly longer more posteriorly but they start to shorten rapidly in the last five pairs of the dorsal series, so that the last dorsal rib is only as long as the first sacral rib.\n\nThere are three sacral ribs in the holotype, identified as such based on a combination of the following observations: short; with thickened shafts; sufficiently wide at the tip to allow articulation with the ilium; oriented so that their tips converge, that is, the first one is postero-laterally inclined, the middle one laterally, and the third one antero-laterally; and located in the sacral region. The same number of sacral ribs were previously identified in AGB6253, a female specimen with embryos ([@ref-30]). At least the third sacral rib of the holotype seems to retain the eight-shaped cross-section of the head, seen in the more anterior ribs.\n\nIt appears that the first 10 caudal vertebrae have rib facets in the holotype, although only the first three caudal ribs are actually preserved. In AGB6253, at least the first nine caudal vertebrae have rib facets but only the first two caudal ribs are preserved. The parapophyses on the sacral and caudal centra are located at the antero-ventral corners of the centra. The rib heads are not eight-shaped in caudal ribs unlike in the more anterior ribs.\n\n### Shoulder girdle\n\n**Scapula.** The scapula is a lunate plate of bone that rolls slightly to wrap around the ventro-lateral part of the shoulder rib cage, as in basal ichthyopterygians. The anterior margin is round without any indentation or emargination. The glenoid and coracoid facet forms a thickened postero-proximal corner ([Fig. 9B](#fig-9){ref-type=\"fig\"}).\n\n![Shoulder girdle of *Chaohusaurus brevifemoralis* sp. nov.\\\n(A) AGB7405. (B) AGB6260, (C) AGB7400. (D) Approximate bone map for (A). (E) Same for (B). (F) Same for (C). See the section \"Osteological abbreviations\" for abbreviations. Scale bar is one cm.](peerj-07-7561-g009){#fig-9}\n\n**Coracoid.** The coracoid is a plate of bone that appears like a fan with a short and thickened stem ([Fig. 9B](#fig-9){ref-type=\"fig\"}). The stem ends laterally with the glenoid and scapular facets. The inter-coracoidal suture is straight in the central region only. Both the anterior and posterior margins of the coracoid have concave regions.\n\n**Interclavicle.** The interclavicle is one of the most poorly known bones in *Chaohusaurus* in general. The complete outline of the interclavicle is seen only in two specimens that may or may not belong to *Chaohusaurus brevifemoralis*, namely AGB6607 and 7400 (see \"Systematic Paleontology\"). The bones in these specimens are approximately V-shaped plates ([Fig. 9F](#fig-9){ref-type=\"fig\"}), with a convex anterior margin and a concave posterior margin. The anterior process is shorter than in *Cartorhynchus*, and coarsely and strongly striated in approximately radial orientations. The overall shape of the bone would be similar to that of *U. hataii* if the posterior process of the latter interclavicle was removed. Some other specimens show in-between the two clavicles one or two pieces of bone that may appear to be a part of the interclavicle ([Figs. 9B](#fig-9){ref-type=\"fig\"} and [9C](#fig-9){ref-type=\"fig\"}). However, after examining many specimens, these parts are most likely parts of the clavicle that wraps around the anterior side of the interclavicle and that have been cracked. It appears that the interclavicle is minimally exposed when the dermal girdle is in articulation, leading to the difficulty of seeing the bone in most specimens.\n\n**Clavicle.** The clavicle is the most prominent bone of the shoulder girdle. It is a J-shaped bone that forms a complete U-shape when the right and left elements are articulated ([Fig. 9B](#fig-9){ref-type=\"fig\"}). AGB7400 shows a different posture of the clavicles ([Fig. 9C](#fig-9){ref-type=\"fig\"}) but this is the only specimen where the clavicle has been rotated so that its distal tip points laterally, while virtually all other specimens, including the holotype and GMPKU P-3086, have the clavicular distal tips pointing posteriorly. The condition in AGB7400 is unlikely to reflect the natural posture because the other posture is seen in over a dozen specimens.\n\nThe clavicle is thickest in two places, near the proximal end and in the midshaft region where an angle is formed antero-ventrally ([Fig. 9](#fig-9){ref-type=\"fig\"}). This midshaft angle is located in-between the coracoid and scapular articulations. The part proximal to this angle is the main body of the clavicle, which wraps around the anterior margin of the interclavicle medially and faces the gap between the coracoid and clavicle (see below) laterally. The thick proximal ends of the right and left clavicles seem to overlap slightly in front of the interclavicle, thus concealing the anterior aspect of the latter bone completely.\n\n### Forelimb\n\n**Humerus.** The humerus typically has an anterior flange extending preaxially from the main shaft of the bone, as in basal ichthyosauromorphs. This flange in *Chaohusaurus brevifemoralis* is divided into a proximal and distal sub-flange as in *Chaohusaurus chaoxianensis*, but with a notable gap between the two unlike in the latter species. The relative extent of this gap changes ontogenetically, together with that of the anterior flange. In the smallest individual, the sub-flanges appear virtually absent ([Fig. 9A](#fig-9){ref-type=\"fig\"}), with the anterior margin of the humerus being concave. The concave margin has radial, rather than longitudinal, striations, suggesting the presence of a very narrow anterior flange along the margin. The sub-flanges appear proximally and distally and extend toward the center of the anterior margin with growth. The gap between them therefore narrows with growth, although it never disappears completely. In *Chaohusaurus chaoxianensis*, a notch-like incision exists between the two sub-flanges, instead of an obvious gap as in *Chaohusaurus brevifemoralis*.\n\nThe head of the humerus is located postero-proximally. Its ossification is usually incomplete, with only the largest specimen revealing a humeral head that is partly ossified ([Fig. 10D](#fig-10){ref-type=\"fig\"}). The degree of ossification of the humeral head is used to judge the relative degree of osteological maturity in thunnosaurian ichthyosaurs ([@ref-12]). Similar usage is difficult in *Chaohusaurus brevifemoralis* because of the poor ossification in almost all specimens ([Fig. 10](#fig-10){ref-type=\"fig\"}).\n\n![Forelimb of *Chaohusaurus brevifemoralis* sp. nov.\\\n(A) AGB7405. (B) AGB6260, (C) AGB7400. (D) AGB7408. (E) Approximate bone map for (A). (F) Same for (B). (G) same for (C). (H) Same for (D). See the section \"Osteological abbreviations\" for abbreviations. Scale bar is one cm.](peerj-07-7561-g010){#fig-10}\n\n**Radius.** The radius is a flattened long bone that is wider proximally than distally. The peripheral shaft of the bone seems to be absent because the entire anterior margin of the bone has radial orientation of striation that are perpendicular to the margin, rather than those that are parallel to the main axis of the bone as in the proper shaft region. As a result, there is a band of \"anterior flange\" extending along the entire peripheral margin of the bone, making the bone appear wide. The anterior flange widens proximally, where there is the antero-proximal flange of the radius in *Chaohusaurus chaoxianensis*. In *Chaohusaurus chaoxianensis*, the contrast in widths between this antero-proximal flange and the more distal part of the flange is conspicuous, with an abrupt change. However, in the present species, the difference is much smaller because the flange gradually widens proximally without any abrupt change. As a result, the antero-proximal flange of the radius appears small at the first sight.\n\n**Ulna.** The ulna is an elongated and flattened bone with a proximal shaft and a distal fan, as in basal ichthyopterygians. The distal fan of the ulna typically appears asymmetric relative to the longitudinal axis of the bone, developing more postaxially than preaxially. This morphology differs from that in a typical ulna of *Chaohusaurus chaoxianensis* where the fan is more symmetrical except in one individual where one of the ulnae has a weakly asymmetrical distal fan (AGB6259; compare [@ref-28]: fig. 1C) and [@ref-31]: figs. 2E, F)).\n\n**Carpals.** Much of the carpus region is without hard tissue. There are three disk-shaped carpal ossifications, corresponding to the intermedium, ulnare, and fourth distal carpal, in nearly all individuals regardless of body size. Exceptions are a large specimen which seems to have only two carpal bones (AGB6260; [Fig. 10C](#fig-10){ref-type=\"fig\"}), the terminal embryo described elsewhere that lacks any carpal ossification (AGB6253), and the largest specimen with four carpals, adding the radiale to the list above of three bones (AGB7408). The largest individual of *Chaohusaurus chaoxianensis* also has four carpals ([@ref-32]), but this individual is smaller than large specimens of *Chaohusaurus brevifemoralis* with only three carpals. Carpal ossifications in *Chaohusaurus brevifemoralis* appear smaller than those of *Chaohusaurus chaoxianensis* relative to the carpus length.\n\n**Metacarpals.** There are five metacarpals in all specimens, except in one specimen (AGB6258) where six metacarpals are present in both right and left forelimbs ([Fig. 10B](#fig-10){ref-type=\"fig\"}). The extra metacarpal in this exceptional individual is located preaxially, as in some hupehsuchian ichthyosauromorphs. Unlike in hupehsuchians, however, there is no ossification for an extra preaxial distal carpal that would form the base of the extra metacarpal. Also, there are no phalanges that extends distally from this extra metacarpal. The first and fifth distal carpals lack a complete shaft along their peripheral margins, which are notched. These bones are not completely lunate as in most basal ichthyopterygians because of the notch.\n\n**Manual phalanges.** The forelimb is hypophalangeal, with a phalangeal formula is 1-3-3-3-1. The third phalanx of the second digit is not always present, making the formula 1-2-3-3-1 in some specimens. The phalangeal formula does not seem to depend on size. As in basal ichthyopterygians, most phalanges are flattened cylinders that are approximately symmetrical across the long axis but elements along the peripheral margin of the limb tend to be asymmetrical. The phalanges are well spaced from each other, more so than in *Chaohusaurus chaoxianensis*.\n\n### Pelvic girdle\n\n**Ilium.** The ilium changes its shape with growth. In juveniles, it is essentially a short and flattened rod of even thickness that curves slightly ([Fig. 11A](#fig-11){ref-type=\"fig\"}). In adults, the curvature becomes steeper and a conspicuous iliac blade is formed in the dorsal side of distal part, which is approximately horizontal in natural posture ([Fig. 11C](#fig-11){ref-type=\"fig\"}). The acetabular facet also expands, resulting in a constriction of the bone between the acetabulum and the iliac shaft.\n\n![Pelvic girdle and hind limb of *Chaohusaurus brevifemoralis* sp. nov.\\\n(A) GMPKU P-3086 (one of the paratypes). (B) AGB5846b. (C) AGB6253. (D) AGB7402. (E) Approximate bone map for (A). (F) Same for (B). (G) Same for (C). (H) Same for (D). See the section \"Osteological abbreviations\" for abbreviations. Scale bars are made of one mm squares.](peerj-07-7561-g011){#fig-11}\n\n**Pubis.** The pubis is the largest of the pelvic bones, although it is not remarkably larger than the ischium ([Figs. 11A](#fig-11){ref-type=\"fig\"} and [11B](#fig-11){ref-type=\"fig\"}). The bone is thickened along its antero-lateral margin, toward the acetabular facet that is located at the posterior end of this margin. The acetabular facet is robust and faces postero-laterally. There is an obturator foramen that opens posterolaterally, just medial to the acetabular fact ([Figs. 11A](#fig-11){ref-type=\"fig\"} and [11B](#fig-11){ref-type=\"fig\"}). Because of this location, the medial margin of the foramen is thick while the lateral margin is much thinner. The inter-pubic margin is approximately straight, although the right and left elements are usually preserved at a distance from each other.\n\n**Ischium.** The ischium is a lunate bone ([Fig. 11](#fig-11){ref-type=\"fig\"}). Its antero-lateral corner is thickened to form the acetabular facet, which is thinner than the corresponding facets in the pubis and ilium. The facet is also narrower than the corresponding facet of the pubis. As a result, the acetabulum is not oriented completely laterally but slightly posteriorly ([Fig. 11B](#fig-11){ref-type=\"fig\"}).\n\n### Hind limb\n\n**Femur.** The femur is a robust bone that is wider distally than proximally, with a weakly constricted shaft in-between. The shaft in many specimens may appear wider than in life depending on the degree of compression during preservation. There is a trochanter on the ventral side of the bone, located slightly distal to the head. Its homology with the known ventral femoral trochanters of reptiles, such as the internal trochanter and fourth trochanter ([@ref-39]), is unknown. There is no fossa on the femoral sides. The distal condyle of the femur is flattened, making the knee joint inflexible.\n\n**Tibia.** The tibia is a cylindrical bone in the holotype, unlike other limb elements that are flattened at least to some extent. In most other specimens, however, the bone has been compacted to be a flat element. The tibia is wider proximally than distally, and slightly curved posteriorly toward the distal end.\n\n**Fibula.** The fibula is a flattened long bone with a fan-shaped distal end. The fan develops anteriorly to the shaft, curving the anterior margin of the shaft more strongly than the posterior margin, which in turn is almost straight. This relationship between the shaft and the fan is not obvious in most specimens because of compaction, but is seen at least in better-preserved specimens, including the holotype and AGB6253.\n\n**Tarsals.** Much of the tarsal region is without hard tissue. There are three tarsal ossifications of various sizes, representing the calcaneum, astragalus, and the fourth distal tarsal. This number seems to be constant regardless of body size. They are all disk-shaped and without three-dimensional sculpture. The astragalus is usually the largest, and the fourth distal tarsal the smallest, although the relative size between the astragalus and calcaneum is highly variable depending on the individual.\n\n**Metatarsals.** The metatarsals are similar to the metacarpals in general shape, although the former have shafts that are more slender than the latter, at least in the holotype. This difference is not the result of a preservational bias because the left metacarpals and metatarsals are preserved next to each other in the specimen ([Fig. 2](#fig-2){ref-type=\"fig\"}).\n\n**Pedal phalanges.** The pedal phalangeal formula is 1-2-3-3-1 where countable. As with the metatarsals, the pedal phalanges are more slender than corresponding manual phalanges.\n\nPhylogenetic analysis\n---------------------\n\nBoth software packages yielded an identical length of 713 steps for the shortest trees (CI = 0.363, RI = 0.787). TNT found more than 10,000 equally parsimonious trees, the strict consensus trees of which is given in [Fig. 12](#fig-12){ref-type=\"fig\"}, with Bremer support and bootstrap values. We tried removing *Wumengosaurus* from the outgroup, but it did not affect the consensus topologies. *Chaohusaurus brevifemoralis* appeared as the sister taxon of the sympatric species, *Chaohusaurus chaoxianensis*. This clade has a Bremer support value of 3, and bootstrap support of 82%. The clade is supported by the following unambiguous apomorphies: 98(1), 116(1), 118(1) and 135(1). In contrast, the genus *Chaohusaurus* is weakly supported, without any unambiguous apomorphies---detailed studies of *Chaohusaurus geishanensis* and *Chaohusaurus zhangjiawanensis* are necessary to clarify the nature of the clade.\n\n![Phylogenetic hypothesis of basal Ichthyosauromorpha including *Chaohusaurus brevifemoralis* sp. nov.\\\nNumbers given to nodes are Bremer support value/bootstrap value. See text for how they were computed.](peerj-07-7561-g012){#fig-12}\n\nThe purpose of the present phylogenetic analysis is to locate the new species in the phylogenetic tree of Ichthyosauromorpha, and not to revise the phylogeny of the group. However, it is still worth describing the broader tree briefly. The strict consensus tree topology did not differ drastically from what were published before based on the relevant series of data matrix ([@ref-28], [@ref-29]; [@ref-11]; [@ref-10]). The interrelationships among major clades within Ichthyosauromorpha remained the same throughout these studies. Relationships within some of the major clades have changed through the studies but the present result has minimal difference from the tree given in the last iteration of the series ([@ref-29]), with the only difference found in the breakdown of the clade for the genus *Mixosaurus*. As a result, the distribution of apomorphies along the tree remained similar to what was found in the previous iteration. For example, Ichthyosauriformes is supported by the following character state changes: 11(1), 36(1), 59(1), 100(1), 115(1), 123(1), 144(1), 178(1) and 198(1). Ichthyopterygia shares the following unambiguous apomorphies: 48(1), 89(1), 95(2), 99(0) and 140(1), while Grippidia shares the following two: 34(1) and 113(0).\n\nDiscussion\n==========\n\nThe holotype of *Chaohusaurus brevifemoralis* sp. nov (AGB7401) revealed for the first time the three-dimensional relationship between the orbit and the scleral ring in basal Ichthyosauromorpha. It has been known that the skull roof was constricted in the middle of the orbit in basal ichthyosauriforms ([@ref-15]; [@ref-33]; [@ref-7]) and hupehsuchians ([@ref-2], [@ref-4]) but the ramification of this morphology on the eyeball was never clear because the skulls are usually flattened through postmortem compaction. The scleral rings of the holotype retain their three-dimensional depths ([Fig. 4](#fig-4){ref-type=\"fig\"}). Both of them are exposed dorsally, through the embayment of the skull roof formed by the medial excursion of the dorsal orbital margin, in addition to the exposure through the main part of the orbit. The dorsal exposure reveals the peripheral wall of the scleral ring while the orbital exposure shows the central frustum wall. The function of this dorsal embayment is unclear, while there must be a disadvantage from the additional exposure from the dorsal aspect making the eyeball vulnerable to stresses from that direction. It is unlikely that the embayment allowed the animal to look up better---the limited space inside the eye socket may not have allowed much movement of the eyeballs, even though it appears that looking up to some extent was a part of the lifestyle of *Chaohusaurus brevifemoralis*, with the main part of the orbit facing not completely laterally but slightly dorsally. Despite the ambiguity of function, this embayment and the consequent dorsal exposure of the eyeball was a wide-spread feature found in virtually all ichthyosauromorphs outside of Ichthyosauria.\n\nThe anterior flange of the humerus of *Chaohusaurus* was considered to be notched until recently, based on *Chaohusaurus chaoxianensis* that was the only species in which the flange was completely preserved ([@ref-34], [@ref-35]). However, a new picture has now emerged based on recent findings of new specimens and taxa, including the present species. The margin is indeed concave in three of the four *Chaohusaurus* species, namely *Chaohusaurus geishanensis*, *Chaohusaurus zhangjiawanensis* and *Chaohusaurus brevifemoralis*, although there is a narrow band of flange at least in the latter two species based on the radial orientations of the surface striations perpendicular to the anterior margin. The concave flange morphology is most likely a basal character in Ichthyosauromorpha because it is also present in *Cartorhynchus lenticarpus* ([@ref-28]) and hupehsuchians ([@ref-2], [@ref-3], [@ref-4]; [@ref-44]). Therefore, the notched morphology in *Chaohusaurus chaoxianensis* is an autapomorphy of the species. In the light of this recognition, the complete and convex humeral anterior flanges in grippidians (*Utatsusaurus*, *Grippia* and *Gulosaurus*) ([@ref-21], [@ref-22]; [@ref-6]), represent a shared apomorphy with the rest of the basal ichthyopterygians.\n\nThe position of the pineal foramen relative to the orbit has been used as a character with phylogenetic information content (reproduced as Character 59 in the present study). Typically, the pineal foramen is posterior to the orbit in basal forms and located in-between the orbits in derived forms. *Chaohusaurus brevifemoralis* represents an intermediate state. In specimens with dorso-ventral preservational compaction, the pineal foramen tends to be mostly posterior to the line connecting the posterior limits of the orbits. However, the anterior end of the foramen very slightly crosses the line. More problematically, this positioning varies with specimens. In specimens with oblique preservational compaction, the foramen is often found to be largely in-between the orbits. The same is true in other species of the genus *Chaohusaurus*.\n\nParticipation of the frontal in the orbital rim has been considered an important basal feature of ichthyosauriforms. It is minimally present in *Grippia longirostris* and *U. hataii* because the pre- and postfrontals leave only a narrow gap between them to allow the frontal to be a part of the orbital margin, unlike in *Cartorhynchus lenticarpus* where the frontal forms a substantial portion of the orbital rim as in basal neodiapsids. In *Chaohusaurus brevifemoralis*, this participation is absent because of a narrow contact between the pre- and postfrontals that excludes the frontal from the orbital rim. Despite this difference, it is important to recognize that the overall morphology of the orbital margin in *Grippia* and *Utatsusaurus* is much closer to that of *Chaohusaurus brevifemoralis* than to the same of *Cartorhynchus*. That is, the dorsal part of the orbital rim has a medial excursion corresponding to the narrow lateral embayment of the skull roof, exposing the eyeball dorsally. The only difference is whether the pre- and postfrontals contact each other narrowly or not at the deepest part of the embayment. A similar situation is known in hupehsuchians, where some specimens are preserved with pre- and postfrontals in contact while others have a narrow gap between the two, exposing the frontal to the orbital margin. Therefore, the narrow contact between the pre- and postfrontal may not be a significant phylogenetic character. Based on the tree topology in [Fig. 12](#fig-12){ref-type=\"fig\"}, the contact in *Chaohusaurus* is not necessarily homologous with the robust articulation between the pre- and postfrontals in Ichthyosauria.\n\nParticipation of the frontal in the margin of the upper temporal fenestra was once suggested in *U. hataii* ([@ref-33]) but a later study questioned this morphology ([@ref-7]). Some of the sutures in the specimen in question were not completely clear, leaving room for controversy. Based on well-preserved specimens with unambiguous cranial sutures at hand, it is clear that at least in *Chaohusaurus brevifemoralis* the frontal entered the anterior margin of the upper temporal fenestra, in a manner similar to what was previously described for *U. hataii*. This morphology is unusual among diapsids but it is necessary to recognize that it existed in basal ichthyosauriforms. Reevaluation of the feature in other basal ichthyosauriforms is needed based on well-preserved specimens.\n\nThe squamosal is the most problematic bone in the cranium that warrants a discussion. The bone is interpreted here to be located in-between the supratemporal and quadratojugal, near the center of the height of the cheek, without participating in the upper temporal fenestra. This interpretation differs from a previous one that located the bone along the dorsal margin of the cheek that forms the lateral margin of the upper temporal fenestra. This interpretation was based on two specimens (GMPKU P-3086 and P-3101), but the same study located the squamosal in the same position as in the present work in another specimen (GMPKU P-3188) ([@ref-47]). We reinterpreted GMPKU P-3086 and concluded that the squamosal did not reach the upper temporal fenestra. The apparent contact between the bone and the fenestra is an artifact of damages in the lateral margin of the upper temporal fenestra that expanded the opening laterally, and disarticulation and translation of the bones in the area. If the preserved morphology is original and the squamosal was located along the upper temporal fenestra: (1) the height of the lower temporal fenestra would become about three-quarters of the total cheek height when the proportion is only about half in other specimens; (2) the lateral margin of the upper temporal fenestra would be located too laterally relative to the parietal---the supratemporal process of the parietal extends about halfway through the width of the upper temporal fenestra in other specimens when it is only about a third in the specimen in question; (3) similarly, the upper temporal fenestra would be too large relative to the orbit---the fenestra is less than half of the orbit in maximum diameter in other specimens but the proportion is much larger than half in the specimen in question ([Table 1](#table-1){ref-type=\"table\"}); and (4) the preserved margin of the upper-temporal fenestra would need to be bent unnaturally in three-dimension to reach the squamosal. The problems would disappear if the actual lateral margin of the upper temporal fenestra was more dorsally located than the squamosal, and breakage of the margin led to expansion of the upper temporal fenestra while displacing the margin ventrally. However, scrutiny using 3D reconstruction of CT images would be necessary in the future to reach a solid conclusion.\n\nIt would be useful to compare the morphology of *Chaohusaurus* spp. with those of grippidians (*Utatsusaurus*, *Grippia*, *Gulosaurus*, and *Parvinatator*), from the Lower Triassic. However, detailed comparisons are difficult given the quality of the specimens available for grippidian taxa, and the lack of understanding of the morphology of *Chaohusaurus geishanensis* and *Chaohusaurus zhangjiawanensis*. as well as the lack of three-dimensional reconstruction of the skull in all of the relevant specimens. We therefore consider it premature to present a detailed comparison. However, it is useful to point out some of the unambiguous character state changes. For example, one of the most remarkable differences between grippidians and the more basal ichthyosauriforms is the clear presence of the lower temporal fenestra in the latter as a narrow but deep ventral incision between the postorbital and quadratojugal, while it is lost in the former. Also, the interpterygoid vacuity is open in *Chaohusaurus* and other basal ichthyosauriforms but closed in grippidians and some other ichthyopterygians. Postcranially, *Chaohusaurus* and other basal ichthyosauriforms all lack the pisiform and exhibit delayed ossification of the mesopodium while grippidians and some other ichthyopterygians possess the bone and lack any sign of delay in mesopodial ossification. Additionally, as discussed above, humeral anterior flange morphology differs between the two assemblages.\n\nThe anterior end of the clavicle is almost always preserved in the neck region ([Fig. 8](#fig-8){ref-type=\"fig\"}), and there is a gap between the clavicle and coracoid in most specimens. Given its occurrence in most of the specimens, this gap probably reflects the anatomical positions of the bones in life. The presence of such a gap is unusual among reptiles but a similar gap was identified in *Ophthalmosaurus* ([@ref-19]). Note, however, that the gap is absent in another thunnosaurian ichthyosaur, *Stenopterygius* ([@ref-13]), so it is difficult to establish homology between the gaps in *Chaohusaurus brevifemoralis* and *Ophthalmosaurus* without further studies of this character. Another possible interpretation of the preserved gap in *Chaohusaurus brevifemoralis* is that there was a shared mechanism to anteriorly displace the clavicles postmortem, but the hypothesis currently lacks evidence and a candidate mechanism.\n\nThe co-existence of three species of *Chaohusaurus* in a small geographic region may sound strange. However, as we argued elsewhere, only two of the three, namely *Chaohusaurus chaoxianensis* and *Chaohusaurus brevifemoralis*, are truly sympatric ([@ref-26]). The two species differ in average body size ([@ref-26]), with *Chaohusaurus brevifemoralis* being larger than *Chaohusaurus chaoxianensis*. It is therefore likely that the two utilized different resources in the same geographic region. Phylogenetic analyses suggested a sister-group relationship between the two species. It may be hypothesized that the speciation between the two species involved pursuit of different recourses, at least partly facilitated by the evolution of different body sizes.\n\nConclusions\n===========\n\nThe new ichthyosauriform *Chaohusaurus brevifemoralis* is distinguished from existing species based on a suite of features, including the bifurcation of the neural spines near the caudal peak and short femora relative to the body. These differences are unlikely to be sexual dimorphisms, which are present in the species as an independent suite of characters ([@ref-26]). *Chaohusaurus brevifemoralis* is larger than its sister taxon, *Chaohusaurus chaoxianensis*, in body size. The two species were sympatric but likely pursued different resources from each other's.\n\nSupplemental Information\n========================\n\n10.7717/peerj.7561/supp-1\n\n###### Taxon-character data matrix of Ichthyosauromorpha.\n\nSave the file as a NEXUS file.\n\n###### \n\nClick here for additional data file.\n\n10.7717/peerj.7561/supp-2\n\n###### Character number correspondences between the present study and [@ref-29].\n\n###### \n\nClick here for additional data file.\n\nWe thank Huang Fa-zhong, Hu Yu-fen, and Qi Fei for the preparation of a part of the specimens with AGB numbers used in this study, including the holotype. We also thank Tetsuya Sato and Tian-fen Hu for the preparation of the paratype and other referred specimens deposited at GMPKU, and a part of referred specimens at AGB.\n\na\n\n:   astragalus\n\nan\n\n:   angular\n\nar\n\n:   articular\n\natn\n\n:   atlantal neural spine\n\natp\n\n:   atlantal pleurocentrum\n\naxc\n\n:   axis centrum\n\naxn\n\n:   axis neural spine\n\naxr\n\n:   axis rib\n\nbo\n\n:   basioccipital\n\nbs\n\n:   basisphenoid\n\nc\n\n:   calcaneum\n\nc3\n\n:   3rd centrum\n\nch\n\n:   ceratohyal\n\ncl\n\n:   clavicle\n\nco\n\n:   coronoid\n\ncorac\n\n:   coracoid\n\ncpn\n\n:   caudal peak neural spine\n\ncr\n\n:   caudal rib\n\nd\n\n:   dentary\n\nd.v.f.\n\n:   descending ventral flange of parietal\n\nem\n\n:   extra preaxial metacarpal\n\neo\n\n:   exoccipital\n\nf\n\n:   frontal\n\nfe\n\n:   femur\n\nfi\n\n:   fibula\n\nfran\n\n:   first round anticlined neural spine\n\nH\n\n:   humerus\n\nhe\n\n:   hemal arch\n\nicl\n\n:   interclavicle\n\nil\n\n:   ilium\n\nin\n\n:   intermedium\n\nis\n\n:   ischium\n\nj\n\n:   jugal\n\nl\n\n:   lacrimal\n\nlnn\n\n:   last narrow neural spine\n\nm\n\n:   maxilla\n\nn\n\n:   nasal\n\nn3-5\n\n:   3rd to 5th neural spines\n\npb\n\n:   pubis\n\npl\n\n:   palatine\n\npo\n\n:   postorbital\n\npra\n\n:   prearticular\n\nprf\n\n:   prefrontal\n\nps\n\n:   parasphenoid\n\npt\n\n:   pterygoid\n\nptf\n\n:   postfrontal\n\np.sl.\n\n:   parietal slope\n\nq\n\n:   quadrate\n\nqj\n\n:   quadratojugal\n\nR\n\n:   radius\n\nr\n\n:   radiale\n\nsa\n\n:   surangular\n\nsc\n\n:   scapula\n\nsp\n\n:   splenial\n\nsq\n\n:   squamosal\n\nsr\n\n:   sacral rib\n\nst\n\n:   supratemporal\n\nti\n\n:   tibia\n\nU\n\n:   ulna\n\nu\n\n:   ulnare\n\nv10-v70\n\n:   vertebrae\n\nvc\n\n:   vertebral centrum\n\ni-v\n\n:   metacarpal or metatarsal\n\n4\n\n:   fourth distal carpal.\n\n```{=html}\n<!-- -->\n```\n\nAGM\n\n:   Anhui Geological Museum, Hefei, Anhui, China\n\nGMPKU\n\n:   Geological Museum, Peking University, Beijing, China\n\nIVPP\n\n:   Institute of Vertebrate Paleontology and Paleanthropology, Chinese Academy of Sciences, Beijing, China\n\nNGM\n\n:   Nanjing Geological Museum.\n\nAdditional Information and Declarations\n=======================================\n\nThe authors declare that they have no competing interests.\n\n[Jian-dong Huang](#author-1){ref-type=\"contrib\"} conceived and designed the experiments, contributed reagents/materials/analysis tools, prepared figures and/or tables, approved the final draft.\n\n[Ryosuke Motani](#author-2){ref-type=\"contrib\"} conceived and designed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft, took photographs and measurements.\n\n[Da-yong Jiang](#author-3){ref-type=\"contrib\"} conceived and designed the experiments, contributed reagents/materials/analysis tools, approved the final draft, discussed the specimens on site.\n\n[Andrea Tintori](#author-4){ref-type=\"contrib\"} conceived and designed the experiments, approved the final draft, discussed the specimens on site.\n\n[Olivier Rieppel](#author-5){ref-type=\"contrib\"} conceived and designed the experiments, approved the final draft, discussed the specimens on site.\n\n[Min Zhou](#author-6){ref-type=\"contrib\"} conceived and designed the experiments, contributed reagents/materials/analysis tools, approved the final draft.\n\n[Xin-Xin Ren](#author-7){ref-type=\"contrib\"} conceived and designed the experiments, contributed reagents/materials/analysis tools, approved the final draft.\n\n[Rong Zhang](#author-8){ref-type=\"contrib\"} conceived and designed the experiments, contributed reagents/materials/analysis tools, approved the final draft.\n\nThe following information was supplied regarding data availability:\n\nRaw data is available in the [Supplemental Files](#supplemental-information){ref-type=\"supplementary-material\"}.\n\nThe following information was supplied regarding the registration of a newly described species:\n\nPublication LSID:\n\nurn:lsid:zoobank.org:pub:3FA09089-C940-4499-ABBC-B48F0F70F38E\n\n*Chaohusaurus brevifemoralis* LSID: urn:lsid:zoobank.org:act:B0085E53-C6E4-4C2E-96FF-6D485D3585BA.\n"}
+{"text": "Background\n==========\n\nStress urinary incontinence (SUI) is defined by the International Continence Society as the complaint of involuntary urinary leakage on effort or exertion, or on sneezing or coughing \\[[@B1]\\]. It is a common problem. Prevalence has been estimated at 17-45% of adult women in resource-rich countries \\[[@B2]\\] One cross-sectional study (15,308 women in Norway, aged less than 65 years) found that the prevalence of stress incontinence was 4.7% in women who had not borne a child, 6.9% in women who had had caesarean deliveries only, and 12.2% in women who had had vaginal deliveries only \\[[@B3]\\]. The causes of SUI are urethral hypermobility, in which there is laxity of pelvic floor support, and intrinsic sphincter dysfunction caused by the inability of the urethral sphincter itself to close \\[[@B4]\\]. A multitude of surgical and non-surgical treatment modalities has been described to correct SUI \\[[@B5]\\]. One well-accepted technique for surgical management of urethral hypermobility is the open Burch colposuspension \\[[@B5]-[@B7]\\]. Two to 3 permanent or delayed absorbable sutures are passed through the endopelvic fascia lateral to the midurethra and bladder neck and then through the ipsilateral Cooper\\'s ligament and tied with gentle tension \\[[@B8]\\] A short-term cure rate (defined as the percentage with complete continence) of 73% to 92%, and a success rate (defined as the percentage with cure or improvement) of 81% to 96% have been reported.\\[[@B9]\\] This technique\\'s effectiveness continues for the long term; after 5 to 10 years, approximately 70% of patients are still continent.\\[[@B9],[@B10]\\] Laparoscopic Burch colposuspension was introduced in the early 1990s for the surgical treatment of stress urinary incontinence (SUI) \\[[@B11]\\] Laparoscopic Burch colposuspension (LBC) has been described using the transperitoneal or extraperitoneal approach, using 3 to 5 trocars. The extraperitoneal route is favored by most authors \\[[@B15],[@B21],[@B18],[@B27]\\] and is similar to the technique described by Burch \\[[@B5]\\]. In this approach, the space of Retzius is rapidly dissected using a balloon, or without a balloon by finger and pneumodissection with CO~2~\\[[@B15],[@B23]\\]. The extraperitoneal approach also avoids intraperitoneal pelvic adhesions, minimizes the risk of intra-abdominal injury, and is associated with a shorter learning curve. The main disadvantage of extraperitoneal laparoscopic colposuspension is the risk of increased absorption of CO~2~leading to pneumomediastinum and pneumothorax \\[[@B16],[@B28]\\]. The transperitoneal approach is suitable for patients undergoing concomitant pelvic surgery \\[[@B17],[@B19],[@B25],[@B29],[@B30]\\]. The operative time with this technique may be prolonged due to the need to take down adhesions, mobilize the bladder, and difficulty in retracting intra-abdominal organs \\[[@B29]\\]. Laparoscopic pelvic surgery provides better visualization, shorter hospital stay, better cosmetics, less postoperative pain, and faster recovery to normal daily activity \\[[@B12]\\]. However, despite the renewed interest in the application of laparoscopic technique in the management of SUI, a dichotomy of opinion remains regarding its long-term efficacy.\\[[@B12]\\] Laparoscopic colposuspension is historically regarded as having good, short-term success rate of over 90% \\[[@B12],[@B13]\\] but this rate declines with longer follow-up to 59%-68% \\[[@B13]\\]. The complication rate related to the laparoscopic approach is higher than the open procedure (5-8% vs. 8-22%)\\[[@B14]\\]. The most common intraoperative complication is lower urinary tract injury. Bladder injury, which occurs at an incidence of 2.17-18%, is common in patients with prior pelvic surgery \\[[@B14]-[@B20]\\]. Bladder catheter drainage during surgery and meticulous dissection help prevent most bladder injuries. In the majority of cases, these injuries can be managed laparoscopically obviating the need to convert to an open procedure \\[[@B21]\\]. Conversion rates, especially in the earlier stages of learning, can be as high as 26%\\[[@B21]\\]. Rare cases of partial ureteral obstruction have been reported\\[[@B17],[@B22]\\]. The development of overactive bladder after laparoscopic Burch colposuspension is a well-recognized phenomenon \\[[@B17],[@B18],[@B23]-[@B26]\\]. It occurs at an incidence of 2.8%-8% and has been attributed to extensive dissection of the bladder\\[[@B17],[@B23],[@B26]\\]. The incidence of postoperative permanent or transient urinary retention is low (1.8%) \\[[@B17]\\]. However, there are not many reports on the long-term outcomes of laparoscopic colposuspensions. The purpose of this study was to present the long term results of laparoscopic Burch colposuspension for SUI.\n\nMethods\n=======\n\nA randomized prospective trial was conducted from September 2010 to January 2013. The extraperitoneal laparoscopic Burch colposuspension was performed by an operator on 96 women, mean age was 54,3 \u00b1 3,7 years all of whom suffered from SUI or mixed urinary incontinence. At visit patients were evaluated by means of detailed medical and standradised urogynecological history, clinical examination, cough stress test, urinalysis and urine culture, and instrumental examination like full urodynamic study with urethral closure pressure and voiding cystourethrography. \\[[@B31]\\] All women had urodynamically proven SUI. Inclusion criteria included women with SUI and failed conservative therapy. Exclusion criteria included: previous retropubic continence, intrinsic sphincter dysfunction (abdominal leak point pressure less than 60 cm H~2~O), medically unsuitable for laparoscopic or open surgery, and major degrees of coexisting pelvic organ prolapse, requiring surgery other than a simple rectocele repair. Coexisting idiopathic detrusor overactivity was not an exclusion criterion for entry into the study. Urinary urgency, urgency incontinence, and detrusor overactivity were assessed preoperatively and postoperatively. Women were reviewed at 6 weeks, 6 months, 18 months postoperatively. The Short Form-36 (SF-36) was administered at both baseline and follow up \\[[@B31]\\]. This compares eight scales that can be collapsed into two summary measures assessing physical and mental health, the Physical Component Summary (PCS) and Mental Component Summary (MCS), respectively. In addition, the SF-36 has a general health question (excellent, very good, good, fair, and poor). Only the summary scales and general health question are reported in this study. Summary scores are presented as T-scores with means of 50 and SD of 10 points \\[[@B32]\\]. Lower scores indicate better general, physical, and mental health using the SF-36 survey. Women completed the Short Urinary Distress Inventory (SUDI) and Short Incontinence Impact Questionnaire (SIIQ) at baseline and follow up.\\[[@B33]\\] These instruments assess symptom distress and life impact, respectively, of urinary incontinence. At the 6-week postoperative review, data on resumption of normal activities were also collected via standard questioning. Patient satisfaction was assessed on a visual analogue scale (VAS; 0-100 where 100 represented being completely satisfied and 0 completely unsatisfied). At the 6-month review, urodynamics tests were repeated, and the Genito-Urinary Treatment Satisfaction Scale (GUTSS) was used to assess satisfaction with surgery.\\[[@B34],[@B35]\\]. The scale range is 0-32, with higher scores indicating greater satisfaction. An intraoperative cystoscopy was performed on all the women at the end of the procedure to exclude any urethral and bladder injuries. The Foley catheter was removed 24 hr after the operation, and then intermittent self-catheterizations were performed until the postvoid residual urine was less than 50 mL. Normally distributed continuous variables were analyzed using the unpaired Student\\'s t-test. Categorical variables were compared using the chi-square test or Fisher\\'s exact test. Six-month incidence of stress incontinence or detrusor overactivity symptoms, and the presence of either at urodynamics tests were analysed using logistic regression. Eighteen-month stress incontinence, urgency, and urgency incontinence symptoms were analysed using ordinal logistic regression. For the sensitivity analysis of stress incontinence symptoms at 18 months, we pooled \\'occasional\\' and \\'frequent\\' and added all missing values to this outcome and to the denominator. For the incontinence self-reporting measures (the SUDI, SIIQ) and health status measure, missing scores were imputed using hot deck, where the deck was defined as the treatment group \\[[@B36]\\] For SF-36 items contributing to the PCS and MCS scales, horizontal mean imputation was used.\\[[@B37]\\] Between groups comparisons at baseline and follow up were made using the independent t-teat test; for baseline analyses, the dependent t test was used, and for baseline/follow up by group analyses, analysis of covariance (ANCOVA) was used with the baseline scores entered as the covariate. The statistical significance was set at P \\< 0.05.\n\nResults and discussion\n======================\n\nThe preoperative characteristics of the patients are summarized in Table [1](#T1){ref-type=\"table\"}. After 6 months SUI is 28% with *p*= 0.22, detrusor overactivity is 11% with *p*= 0.88, SUI and/or detrusor overactivity 36% with *p*= 0.41, patient satisfaction 90% with *p*= 0.52 (Table [2](#T2){ref-type=\"table\"}). Five women had detrusor overactivity on urodynamics before but not after surgery. Nine women developed detrusor overactivity on urodynamics after surgery, (*p*= 0,67).\n\n###### \n\nThe preoperative characteristics of population\n\n  VARIABLE                                      LBC           P VALUE\n  --------------------------------------------- ------------- ---------\n  **Age (years)**                                             0.39\n                                                              \n  Mean (SD)                                     51.0 (9.9)    \n                                                              \n  N. of subjects analysed                       96            \n                                                              \n  **Parity**                                                  0.50\n                                                              \n  Mean (SD)                                     2.8 (1.3)     \n                                                              \n  N. of subjects analysed                       94            \n                                                              \n  **Weight (Kg)**                                             0.55\n                                                              \n  Mean (SD)                                     73.3 (14.9)   \n                                                              \n  N. of subjects analysed                       91            \n                                                              \n  **Urinary urgency symptoms (%)**              67            0.50\n                                                              \n  Proportion                                    62/92         \n                                                              \n  **Detrusor overactivity (urodynamics) (%)**   7             0.12\n                                                              \n  Proportion                                    6/87          \n\n(LBC):Laparoscopic Burch colposuspension\n\n###### \n\nUrodynamics and patient satisfaction outcomes at 6 months LBC\n\n  VARIABLE                                      LBC     P VALUE\n  --------------------------------------------- ------- ---------\n  **SUI (%)**                                   28      0.22\n                                                        \n  Proportion                                    23/83   \n                                                        \n  **Detrusor overactivity (urodynamics) (%)**   11      0.88\n                                                        \n  Proportion                                    10/87   \n                                                        \n  **SUI and/or Detrusor overactivity (%)**      36      0.41\n                                                        \n  Proportion                                    30/83   \n                                                        \n  **Patient satisfaction**                      90      0.52\n                                                        \n  Proportion                                    66/73   \n\n(LBC):Laparoscopic Burch colposuspension\n\n(SUI): Stress urinary incontinence\n\nWe asked all patients to complete the Short Form-36 (SF-36). The general health score is improved after surgery (2,60 \u00b1 1.02 versus 2,76 \u00b1 1.06) with *p*= 0.09. The Physical Component Summary (PCS) baseline score is 46.29 \u00b1 10.95 versus 49.54 \u00b1 10.41 after treatment with *p*= 0.01,(Table [3](#T3){ref-type=\"table\"}) so there was a significant baseline to follow up improvement. The Mental Component Summary (MCS) improved also, infact baseline score is 42.19 \u00b1 12.57 versus 42.70 \u00b1 13.03 with *p*= 0.87.(Table [3](#T3){ref-type=\"table\"})\n\n###### \n\nAnalysis of self-reported measures at baseline and 6-month follow up by treatment cohort\n\n  VARIABLE             Baseline        Follow up       P VALUE\n  -------------------- --------------- --------------- ---------\n  **General health**   2.76 \u00b1 1.06     2.60 \u00b1 1.02     0.09\n                                                       \n  **SF-36 PCS**        46.29 \u00b1 10.95   49.54 \u00b1 10.41   0.01\n                                                       \n  **SF-36 MCS**        42.19 \u00b1 12.57   42.70 \u00b1 13.03   0.87\n                                                       \n  **SUDI**             50.22 \u00b1 20.73   23.92 \u00b1 17.90   \\<0.01\n                                                       \n  **SIIQ**             49.98 \u00b1 23.90   31.40 \u00b1 23.83   \\<0.01\n\n(LBC):Laparoscopic Burch colposuspension\n\n(SF-36:MCS): Short Form-36/Mental Component Summary\n\n(SF-36:PCS): Short Form-36/Physical Component Summary\n\n(SIIQ):Short Incontinence Impact Questionnaire\n\n(SUDI): Short Urinary Distress Inventory\n\nTable [3](#T3){ref-type=\"table\"} also shows the SUDI or SIIQ scales. The SUDI baseline score is 50.22 \u00b1 20.73 versus 23.92 \u00b1 17.90, while SIIQ score is 49.98 \u00b1 23.90 versus 31.40 \u00b1 23.83 with p \\< 0.01. In both questionnaires there is an improvement. Satisfaction with treatment outcomes from the GUTSS at 6-month follow up is 29.5 \u00b1 6.3 with *p*= 0.46. (Table [4](#T4){ref-type=\"table\"}). At 18 months after surgery the 31% of patients has ocassionally stress incontinence, while 6% frequently stess incontinence(p = 0.38); ocassionaly urinary urgency is recorded in 40%, and 23% like frequently nurinary urgency (p = 0.40); ocassionaly urge incontinence is presented in 39% patients, while 18% like frequentely (p = 0.21).(Table [5](#T5){ref-type=\"table\"}). The Burch colposuspension appears to be an effective and durable anti-incontinent procedure\\[[@B6]\\]. The same surgery, performed by laparoscopic approach, is gaining popularity because it supposedly presents advantages such as, smaller incisions with better esthetic results, easier access to Retzius space, improved visualization of the surgical field, minimal intraoperative blood loss and lower requirement of analgesics in the postoperative period, in addition to lower cost, shorter hospital stay and rehabilitation period of patients\\[[@B38]-[@B40]\\]. Many authors describe cure rates for laparoscopic Burch surgery similar to those obtained with open technique, however with comparatively shorter follow-up\\[[@B16],[@B41]-[@B43]\\]. LBC has been performed for over a decade with a relatively small number of reported prospective randomised trials \\[[@B41],[@B42]\\]. The role of LBC in the treatment of urinary stress incontinence has changed with the introduction of the tension-free vaginal tape (TVT) procedure. From our data, rate of cure for stress incontinence at 6-month postoperative urodynamics was 72% for LBC. Presently, there are only a few, small randomised controlled trials comparing LBC and TVT, with relatively small numbers and short follow-up times.\\[[@B43],[@B44]\\] TVT is also a minimally invasive procedure that is relatively quick to perform, requiring little equipment, and having a shorter learning curve than LBC. As more evidence is accumulated about the long term success rates of TVT, it may ultimately become the firstline choice for stress incontinence surgery.\\[[@B6],[@B7]\\]. A long term series of Burch colposuspensions have demonstrated excellent durability Since the early 1990s, laparoscopic colposuspension has emerged as a treatment modality in an effort to reduce the surgical morbidity associated with the open Burch colposuspension and to achieve a comparable cure rate\\[[@B17],[@B44]\\].\n\n###### \n\nSatisfaction with treatment outcomes at 6-month follow up: GUTSS scores\n\n  VARIABLE                               Median   IQR   P VALUE\n  -------------------------------------- -------- ----- ---------\n  **Satisfaction with outcome\\***        16.0     4.0   0.47\n                                                        \n  **Satisfaction with care\\*\\***         14.0\u00ec    4.0   0.64\n                                                        \n  **Overall GUTSS satisfaction\\*\\*\\***   29.5     6.3   0.46\n\n\\*Scale range: 0218. The higher the score, the greater the level of satisfaction.\n\n\\*\\*Scale range: 0216. The higher the score, the greater the level of satisfaction.\n\n\\*\\*\\*Scale range: 0234. The higher the score, the greater the level of satisfaction.\n\n(GUTSS):Genito-Urinary Treatment Satisfaction Scale\n\n(IQR): Interquartile range.\n\n###### \n\nSubjective symptoms at 18 months of follow up for LBC\n\n  VARIABLE                             LBC           P VALUE\n  ------------------------------------ ------------- ---------\n  **Stress incontinence (%)**                        0.38\n                                                     \n  Occasionally                         (31), 24/77   \n                                                     \n  Frequently                           (6), 5/77     \n                                                     \n  **Urinary urgency (%)**                            0.40\n                                                     \n  Occasionally                         (40), 30/75   \n                                                     \n  Frequently                           (23), 17/75   \n                                                     \n  **Urge incontinence (%)**                          0.21\n                                                     \n  Occasionally                         (39), 29/74   \n                                                     \n  Frequently                           (18), 13/74   \n                                                     \n  **Patient satisfaction (%)scores**                 0.10\n                                                     \n  Proportion                           (58), 43/74   \n\n(LBC):Laparoscopic Burch colposuspension\n\nLapitan et al. \\[[@B45],[@B46]\\] reviewed 33 trials that involved a total of 2,403 women, who underwent open retropubic colposuspensions and found an overall cure rate between 68.9% and 88.0%. They reported that the overall continence rates were approximately 85-90% within the first year and 70% after five years of treatment.\n\nThere are more than 150 published reports about laparoscopic colposuspensions. However, the long-term outcomes of laparoscopic colposuspension are uncertain, due to the limited duration of follow-up in most series. In 2006, the Cochrane Incontinence Group suggested that the laparoscopic colposuspension may be as good as open colposuspension at two years post surgery according to the currently available data \\[[@B47]-[@B50]\\].\n\nDoret at al. observed that long term results with laparoscopic Burch colposuspension are relatively good but a bit lower than those published with traditional open technique. The effects of the learning curve with an evolving technique are to be considered when analyzing the results \\[[@B51]\\].\n\nConclusion\n==========\n\nThe LBC has significant advantages, without any apparent compromise in short-term and long term outcomes. It determines improvement in objective and subjective measures of disease and in patient satisfaction at 6 months,18 months of follow up. Quality of life significantly improves after laparoscopic Burch colposuspension and reoperations are uncommon.\n\nList of abbreviations\n=====================\n\n(ANCOVA): Analysis of covariance; (GUTSS): Genito-Urinary Treatment Satisfaction Scale; (LBC): Laparoscopic Burch colposuspension; (MCS): Mental Component Summary; (PCS): Physical Component Summary; (SF-36): Short Form-36; (SIIQ): Short Incontinence Impact Questionnaire; (SUDI): Short Urinary Distress Inventory; (SUI): Stress urinary incontinence; (TVT): Tension-free vaginal tape; (VAS): Visual analogue scale.\n\nCompeting interests\n===================\n\nThe authors declare that they have no competing interests.\n\nAuthors\\' contributions\n=======================\n\nDP: conception and design, interpretation of data, given final approval of the version to be published.\n\nFI: conception and design, interpretation of data, given final approval of the version to be published\n\nGDL: acquisition of data, drafting the manuscript, given final approval of the version to be published\n\nEI: acquisition of data, drafting the manuscript, given final approval of the version to be published\n\nGR acquisition of data, drafting the manuscript, given final approval of the version to be published\n\nAR acquisition of data, drafting the manuscript, given final approval of the version to be published\n\nNR acquisition of data, drafting the manuscript, given final approval of the version to be published\n\nBA: critical revision, interpretation of data, given final approval of the version to be published\n\nAuthors\\' information\n=====================\n\nDP:Associate Professor of Urology at University \\\"Federico II\\\" of Naples.\n\nFI: Associate Professor of Urology at University \\\"Federico II\\\" of Naples.\n\nGDL: Consultant, Department of Urology - Hospital Santa Maria delle Grazie\n\nEI: Resident in Urology Training Programme at University Federico II of Naples\n\nGR: Resident in Urology Training Programme at University Federico II of Naples\n\nAR: Resident in Urology Training Programme at University Federico II of Naples\n\nNR: Resident in Urology Training Programme at University Federico II of Naples\n\nBA: Associate Professor of Surgery at University \\\"Federico II\\\" of Naples.\n\nDeclarations\n============\n\nPublication of this article has come from personal funds.\n\nThis article has been published as part of *BMC Surgery*Volume 13 Supplement 2, 2013: Proceedings from the 26th National Congress of the Italian Society of Geriatric Surgery. The full contents of the supplement are available online at <http://www.biomedcentral.com/bmcsurg/supplements/13/S2>\n"}
+{"text": "Introduction\n============\n\nChemokines are primarily known for their roles in inflammatory and immune responses. However, these small cytokines have also been shown to be upregulated during development of the granulation tissue of wounds and to act on endothelial cells and keratinocytes, cells that perform important functions in the development of the healing tissue ([@bib29]; [@bib40]; [@bib27]; [@bib34]). In addition to these cell types, fibroblasts are also critical players during formation of the granulation tissue. These cells are activated by cytokines and growth factors that are released from platelets and produced by macrophages. The activated fibroblasts proliferate and migrate across the provisional matrix formed by the fibrin--plasma fibronectin clot. As the clot is digested by plasmin, fibroblasts replace it with cellular fibronectin, tenascin, and collagen III ([@bib28]; [@bib6]), molecules that are critical for migration of endothelial cells and keratinocytes into the wound and for proper scar formation. Some of the fibroblasts infiltrating the wound differentiate to become myofibroblasts ([@bib16]; [@bib43]), cells that have bundles of \u03b1-smooth muscle actin (SMA)[\\*](#fn1){ref-type=\"fn\"} and can contract, contributing to closure of the wound ([@bib17]; [@bib25]).\n\nChemokines are small, positively charged, secreted proteins that consist of an NH~2~-terminal region of variable conformation followed by a loop, three antiparallel \u03b2 strands, and a COOH-terminal \u03b1 helix ([@bib7]). They can be divided into four families based on the position of the first two cysteines. The two major families are the CXC family in which the two cysteines are separated by any single amino acid (e.g., cCAF, interleukin \\[IL\\]-8, gro\u03b1/melanocyte growth--stimulating activity \\[MGSA\\], SDF-1, PF4, IP-10) and the CC family in which the two cysteines are adjacent (MCPs, RANTES, Eotaxin, MIPs) ([@bib44]; [@bib2]; [@bib55]). These proteins have no modifications other than two disulfide bonds and are multifunctional. Chemokines function in a very tightly regulated dose- and time-dependent manner, strongly suggesting that their actions are affected by the microenvironmental conditions ([@bib13]; [@bib18]; [@bib45]; [@bib53]).\n\nThe first evidence that chemokines are associated with healing was reported in 1990 when it was shown that the chemokine chicken chemotactic and angiogenic factor (cCAF) is overexpressed during wound healing ([@bib29]). This chemokine is highly expressed in the first 24--48 h after injury and remains elevated for at least 16 d after wounding ([@bib29]; [@bib31]; [@bib33]). It is primarily expressed by the fibroblasts of the granulation tissue, especially where interstitial collagen is abundant and by the endothelial cells of microvessels of the granulation tissue ([@bib29]; [@bib32]).\n\nIn the chicken chorioallantoic membrane (CAM) assay, at low doses cCAF is chemotactic for monocyte/macrophages and lymphocytes and after several days of exposure to this chemokine, the ectoderm of the CAM becomes thickened and a granulation-like tissue develops beneath the cCAF-containing pellet. In this granulation-like tissue, there is an increase in the amount of interstitial collagen and the fibroblasts in the mesoderm are consistently aligned with the collagen fibers and appear to cause tissue contraction ([@bib34]). At high concentrations, however, cCAF stimulates blood vessel sprouting from the preexisting vessels of the CAM (angiogenesis) in the absence of leukocyte chemotaxis ([@bib34]).\n\nOther CXC chemokines have also been associated with wound-healing events. For example, in burn wounds, gro\u03b1/MGSA is expressed by keratinocytes as they differentiate after reepithelialization of the wound. Furthermore, CXCR2, the receptor for MGSA, is present in migrating and proliferating keratinocytes ([@bib40]; [@bib45]). In the granulation tissue, MGSA expression is associated with fibroblasts, smooth muscle cells/myofibroblasts, and a subpopulation of macrophages ([@bib40]). This pattern of expression strongly suggests a role of this chemokine in healing of burn wounds. It also has been shown that in transgenic mice expressing IP-10, a CXC chemokine that inhibits angiogenesis, wounds heal poorly and exhibit defects in development of the granulation tissue ([@bib27]).\n\nIt is becoming increasingly more evident that chemokines are expressed at the sites of injury and that they affect processes involving proper development of the granulation tissue. Fibroblasts are critical participants in the development of this healing tissue and they also express high levels of chemokines upon stimulation by stress-inducing agents such as those released upon wounding. Despite this correlative evidence, little is known about how these small cytokines affect wound fibroblast function. cCAF is highly expressed by the fibroblasts of healing tissue, stimulates the formation of granulation-like tissue in the CAM, and is highly homologous to several human chemokines ([@bib50]). Therefore, we investigated the effects of this chemokine on proliferation and differentiation of fibroblasts, both important processes during granulation tissue development. We find that cCAF stimulates fibroblasts to differentiate into myofibroblasts and accelerates wound closure in vivo*.*\n\nResults\n=======\n\nEffects of cCAF on fibroblast growth\n------------------------------------\n\nTo determine the effects of cCAF on fibroblast growth, primary chicken embryonic connective tissue fibroblasts were cultured in the presence of 2% donor calf serum. Embryonic fibroblasts behave much like wound fibroblasts ([@bib4]), and the presence of a small amount of serum in the medium mimics the conditions of the wound, where serum factors are abundant. Because these are primary fibroblasts isolated weekly from different embryos, rather than cell lines, there are genetic variations and therefore differences in the levels of response to the various treatments. As a consequence, internal controls were always included and results from different batches of cells were not averaged. Therefore, all figures depict a representative experiment out of several performed for each type of experiment.\n\nWhen fibroblasts are cultured in the presence of serum, cCAF suppresses proliferation of these cells by 25% compared with untreated cells ([Fig. 1](#fig1){ref-type=\"fig\"}; *P* \\< 0.01). This effect is dose dependent, with the greatest suppression occurring at 500--750 ng cCAF/ml medium ([Fig. 1](#fig1){ref-type=\"fig\"} A), and at 2 d after plating (unpublished data). It is commonly observed that chemokines cause their effects only in a narrow range of concentrations; at concentrations higher than the optimal dose, their receptors are very quickly desensitized and/or downregulated (e.g., [@bib55]; [@bib38]). Although chemokine concentrations of 10^\u22122^--10 ng/ml can chemoattract and activate leukocytes, chemokines acting on other cell types such as endothelial cells, smooth muscle cells and fibroblasts require concentrations in the range of 10^2^ ng/ml ([@bib19]; [@bib18]; [@bib26]). This is within physiological range; wound fluid from burn patients has MGSA concentrations of 10^2^--10^3^ ng/ml at 6--7 d after injury ([@bib45]). Experiments using an antibody specific for cCAF showed that this antibody abrogates the effects of this chemokine on proliferation ([Fig. 1](#fig1){ref-type=\"fig\"} B).\n\n![**The effects of cCAF on fibroblast growth.** (A) Primary connective tissue embryonic fibroblasts were treated for 3 d with increasing doses of cCAF. Treatment was applied every 24 h. cCAF suppresses fibroblast proliferation in a dose-dependent pattern characteristic of most chemokine functions. Maximum suppression was observed at a dose of 750 ng/ml. (B) Inhibition of cCAF function with an antibody specific to this chemokine abrogates this suppression of growth. Results shown are representative of several independent experiments. Double asterisk indicates *P* \\< 0.01; Asterisk indicates *P* \\< 0.05.](0103062f1){#fig1}\n\nAlthough the decrease in growth induced by cCAF is significant, it is not large. Therefore, we examined the possibility that this reduction in cell numbers was due to increased cell death. Cells grown under the same conditions as above were prepared to detect apoptosis by determining whether DNA laddering occurs. Both untreated fibroblasts and those treated with cCAF had intact DNA ([Fig. 2](#fig2){ref-type=\"fig\"} A), showing that this chemokine did not induce cells to undergo apoptosis. Similar studies performed to detect whether cCAF induced necrotic cell death using trypan blue showed that this chemokine does not cause cells to undergo necrosis ([Fig. 2](#fig2){ref-type=\"fig\"} B).\n\n![**The effects of cCAF on cell death.** (A) To detect cCAF-induced apoptosis we used the DNA Laddering Detection system from Roche Biochem. Fibroblasts were plated, treated as described in Materials and methods, the DNA was prepared as suggested by the manufacturer, and agarose gel electrophoresis was performed. No DNA laddering was detected in cells treated with cCAF, N-peptide, or C-peptide. The positive control for DNA laddering provided by the manufacturer consisted of U937 cells treated for 3 h with camptothecin (4 \u03bcg/ml). (B) Cells incubated with trypan blue, which stains necrotic cells blue, showed that cCAF does not stimulate cell death by necrosis. The results shown represent one of three experiments.](0103062f2){#fig2}\n\nWe have shown previously that the COOH-terminal peptide (28 amino acids) of cCAF is by itself angiogenic in vivo ([@bib34]; [@bib35]). Therefore, we tested the possibility that these effects of cCAF on fibroblast growth are also mediated by the COOH terminus of the molecule. Cells were treated with the COOH-terminal peptide following the same regimen of treatment as for cCAF; we found that this peptide did not cause a decrease in cell numbers ([Fig. 3](#fig3){ref-type=\"fig\"}). We then turned to the NH~2~ terminus of the molecule because it has been shown for several chemokines that their chemotactic properties for leukocytes require this portion of the protein ([@bib7]; [@bib51]; [@bib52]; [@bib1]; [@bib34]). Treatment of fibroblasts with the first 15 amino acids of the NH~2~ terminus (N-peptide) resulted in a decrease in growth similar to that induced by cCAF ([Fig. 3](#fig3){ref-type=\"fig\"}). Furthermore, the dose- and time-dependent pattern of suppression of growth induced by the N-peptide was the same as that of the whole cCAF molecule, albeit at a higher molar concentration (unpublished data).\n\n![**Effects of the C- and N-peptides of the cCAF protein on fibroblast growth.** Treatment with the C-peptide (28 amino acids) did not suppress the growth of connective tissue fibroblasts, whereas treatment with the N-peptide (15 amino acids) suppresses the growth of fibroblasts to a level similar to that caused by the whole cCAF molecule. Results shown are representative of three independent experiments. Asterisk indicates *P* \\< 0.05.](0103062f3){#fig3}\n\ncCAF stimulates \u03b1-SMA expression\n--------------------------------\n\nThe results described above show that the decrease in cell numbers induced by cCAF is specific and not due to cell death, but that the effect is small and requires 2--3 d to be detected, even with repeated applications of cCAF. This could mean that cCAF plays a role in triggering fibroblasts to slow down growth and develop a differentiated phenotype with a consequent decrease in cell division. Therefore, we investigated the possibility that cCAF stimulates differentiation of fibroblasts into myofibroblasts, cells that play an important role in wound closure and proliferate at a slower rate than normal fibroblasts ([@bib36]; [@bib41]; [@bib11]; [@bib24]). Immunolabeling of cCAF-treated fibroblasts with an antibody specific for \u03b1-SMA, a marker for myofibroblasts, revealed the presence of many cells showing intense staining for fiber bundles, whereas control cells showed fewer and less brightly stained fibers ([Fig. 4, A and B](#fig4){ref-type=\"fig\"}). As was found for the growth studies, treatment with the NH~2~-terminal peptide stimulated \u03b1-SMA production, whereas the COOH-terminal peptide did not ([Fig. 4, C and D](#fig4){ref-type=\"fig\"}). Immunoblot analysis of cell extracts using this antibody showed increased levels of \u03b1-SMA in cells treated with cCAF or the N-peptide compared with cells treated with the C-peptide and control cells ([Fig. 4](#fig4){ref-type=\"fig\"} E). Inhibition of cCAF or N-peptide function by treatment of the cells in the presence of the cCAF antibody abrogated the increase in \u03b1-SMA expression ([Fig. 4, F and G](#fig4){ref-type=\"fig\"}).\n\n![**Effect of cCAF and the NH** ~2~ **terminus on** \u03b1**-SMA expression.** (A) Untreated embryonic connective tissue fibroblasts were immunolabeled for \u03b1-SMA, a marker for myofibroblast differentiation. Some fibroblasts show a small amount of staining for this protein, which is characteristic of fibroblasts in culture. (B) Cultures treated with cCAF or (C) with the N-peptide for 3 d show that more fibroblasts are stained for \u03b1-SMA, that more fibrils are present in the cells, and that the staining is more intense than for untreated cells or those treated with the C-peptide **(**D). (E) Immunoblot analysis for \u03b1-SMA to quantify the results observed in (A--D). All lanes in the immunoblots contain equal amounts of total protein, as measured by the DC protein assay (Bio-Rad Laboratories). Cells treated with 750 ng/ml of cCAF or of the N-peptide show much higher levels of \u03b1-SMA when compared with untreated or C-peptide--treated cells. *n* = 4 and asterisk indicates *P* \\< 0.05. Inhibition of cCAF (F) or N-peptide (G) with an anti-cCAF antibody blocks the increase in \u03b1-SMA expression. Results shown in E, F, and G are representative of three or more experiments.](0103062f4){#fig4}\n\nAlthough \u03b1-SMA is considered the characteristic marker for myofibroblasts, vimentin, desmin, and myosin heavy chain, as well the muscle proteins myoD and myogenin, are sometimes expressed by myofibroblasts ([@bib49]; [@bib37]). Using antibodies recognizing the chicken proteins, we immunoblotted cell extracts from cCAF and N-peptide--treated fibroblasts. There was no detectable desmin, myosin heavy chain, or myoD present in either control or treated cells (unpublished data). However, vimentin levels increased with cCAF or N-peptide treatment in a manner similar to that observed for \u03b1-SMA ([Fig. 5](#fig5){ref-type=\"fig\"}). Although desmin and the sarcomeric proteins myoD and MHC are expressed in some myofibroblasts, their presence or absence depends on the origin of the myofibroblasts, whereas most myofibroblasts express both vimentin and \u03b1-SMA ([@bib15]).\n\n![**Effect of cCAF and the NH~2~ terminus on vimentin expression.** Immunoblot analysis for vimentin, another marker for myofibroblasts. Cells treated with 750 ng/ml of cCAF or of the N-peptide show higher levels of vimentin than do untreated or C-peptide--treated cells. All lanes contain equal amounts of total protein, as measured by the DC protein assay (Bio-Rad Laboratories).](0103062f5){#fig5}\n\nTo determine whether this increase in \u03b1-SMA and vimentin results in a functional phenotype, we examined the effect of cCAF on contraction of fibroblast-seeded collagen gels. Fibroblasts plated on collagen often grow in clusters, but cellular clusters in plates treated with cCAF or N-peptide were more contracted than clusters in control cells ([Fig. 6](#fig6){ref-type=\"fig\"}, A--C). When collagen gels were released after 4--5 d of treatment, cCAF or N-peptide--treated cells contracted the gels more tightly than did untreated cells ([Fig. 6](#fig6){ref-type=\"fig\"}, D--F). Immunolabeling of sections prepared from collagen gels with the antibody to \u03b1-SMA showed that treated gels contain more cells with fibers staining for \u03b1-SMA than control gels ([Fig. 6](#fig6){ref-type=\"fig\"}, G--I). [Fig. 6](#fig6){ref-type=\"fig\"} J shows a quantitative analysis of cCAF- and of N-peptide--induced collagen gel contraction. To demonstrate that this contraction is the result of an increase in \u03b1-SMA expression, we treated the cells with phosphorothioate antisense oligodeoxyribonucleotides (ODN) to inhibit production of \u03b1SMA. Antisense ODN strongly decreased the ability of the cells treated with cCAF or the N-peptide to contract the collagen gels, but did not have an effect on the contractility of cells treated with the COOH terminus ([Fig. 7](#fig7){ref-type=\"fig\"} A). Detection of \u03b1-SMA in these gels shows that the levels of this protein in the gels treated with antisense oligonucleotides is similar to those of the control, whereas the levels in the gels treated with cCAF or the N-peptide alone or in the presence of sense oligonucleotides, have higher levels of \u03b1SMA ([Fig. 7](#fig7){ref-type=\"fig\"} B).\n\n###### \n\n**Increase in collagen gel contraction by fibroblasts treated with cCAF.** (A--C) Untreated fibroblasts plated on collagen gels form loose clusters (A), whereas fibroblasts treated with cCAF (B) or N-peptide (C) form clusters that are much tighter and generally pull away from the collagen. (D--F) Collagen gels, after release from the edge of the culture dish, show moderate contraction for untreated cells (D) but much greater contraction for cells treated with cCAF (E) or N-peptide (F). (G--I) Sections through the gels depicted in D--F were immunolabeled for \u03b1-SMA to show the presence of cells containing \u03b1-SMA. The staining is much less intense in untreated cells (G) than in cells treated with cCAF (H) or N-peptide (I). (J) Contracted gels were photographed and their areas were measured and analyzed using NIH Image. Treatments with both cCAF and N-peptide stimulated fibroblasts to contract gels more efficiently than did untreated cells. Asterisk indicates *P* \\< 0.05. Bars, 100 \u03bcm.\n\n![](0103062f6ai)\n\n![](0103062f6j)\n\n![**Inhibition of \u03b1-SMA expression using oligonucleotides.** (A) Phosphorothioate antisense ODN were used to inhibit \u03b1SMA expression. The sequence of ODN is provided in Materials and methods. When preparing the collagen gels, 2 \u03bcM of antisense or sense ODN were incorporated into the gel before the seeding of CEFs. The treatments to induce gel contraction were performed as in [Fig. 6, A--F](#fig6){ref-type=\"fig\"}, except that 2 \u03bcM of antisense or sense ODN was added each time the chemokine or peptide was applied. The ODN strongly inhibited gel contraction stimulated by cCAF or the N-peptide. (B) Immunoblot analysis of \u03b1SMA production in the gels from the experiments in A.](0103062f7){#fig7}\n\nT\u03bf determine whether cCAF directly stimulates \u03b1-SMA mRNA expression, we used reverse transcription (RT)-PCR with specific primers to this molecule to examine the effects of cCAF on mRNA levels. This chemokine causes an increase in \u03b1-SMA mRNA levels within 1 h of treatment, and mRNA levels remain elevated for more than 12 h after treatment ([Fig. 8](#fig8){ref-type=\"fig\"}).\n\n![**Stimulation of** \u03b1**-SMA expression in cells treated with cCAF.** Fibroblasts treated with cCAF show rapid increase of \u03b1-SMA mRNA, peaking at 3 h and declining by 24 h. Total RNA extracted using TRIzol reagent and RT-PCR was performed as described in Materials and methods. A fragment of the 18S rRNA was used as control to normalize the amount of \u03b1-SMA.](0103062f8){#fig8}\n\ncCAF-induced effects on wound closure\n-------------------------------------\n\nIn wounded tissue, myofibroblast differentiation contributes to closing of wounds ([@bib41]). Our finding that cCAF stimulates fibroblasts to acquire myofibroblastic phenotype and function in culture led us to determine whether cCAF affects wound closure in vivo. Full thickness and same size excision wounds were made on the underside of wings of 2-wk-old chicks, and for each bird one wing was treated with vehicle (H~2~O) and the other with 1 \u03bcg of cCAF every other day. The bandages were removed at days 3, 5, and 7, the wounds were photographed, dressed with clean bandages, and treatments applied. For each wound, rate of wound closure was examined by taking digital photographs of the same wound through time and then determining the areas of the wounds from the photographs using NIH Image. [Fig. 9, A and C](#fig9){ref-type=\"fig\"}, show a wounded wing before and after 3 d of treatment with vehicle, whereas [Figs. 9, B and D](#fig9){ref-type=\"fig\"}, show the contralateral wing of the same bird, treated with cCAF at the same times. [Fig. 9](#fig9){ref-type=\"fig\"} E shows the results of analyzes and comparison of the wound areas of several birds over time**.** On day 3, the cCAF-treated wounds had closed to a significantly greater degree than the vehicle-treated wounds. At day 5, the closure of the control wounds was still less than that of the experimental wounds, but closure in the controls may have accelerated during this period. Although cCAF-treated wounds close faster, by day 7 both control and treated wounds were completely closed. Wounds treated with the N-peptide also showed accelerated wound closure during days 0--3, albeit less strongly than with the whole cCAF molecule ([Fig. 9](#fig9){ref-type=\"fig\"} F).\n\n![**Accelerated closure of wounds treated with cCAF or with N-peptide.** Excision wounds were made on the underside of chick wings and then treated every other day with vehicle (H~2~O) or 1 \u03bcg cCAF. The dashed lines trace the edges of the wounds. The underlying muscle can be seen within the wounds. Control wings (A and C) show less wound closure at day 3 than cCAF-treated wounds (B and D). Dashed straight lines mark the length of the forewings to facilitate evaluation of the size of the wounds. (E) Average percent wound closure in eight birds for cCAF or vehicle treatment. (F) Average percent wound closure of eight birds/treatment with N-peptide or vehicle. Areas of the wounds were measured on digital photographs using NIH Image software. Percentage wound closure was calculated by comparing the area of the closing wound to the area of the same wound on day 0. At day 3, *P* \\< 0.05 for both treatments.](0103062f9){#fig9}\n\nSections through the wounds show that the granulation tissue of cCAF-treated wounds appears more dense earlier in wound healing (days 3 and 5) than the tissue of the wounds treated with vehicle ([Fig. 10](#fig10){ref-type=\"fig\"}, A--D). In particular, the granulation tissue of the control wounds at day 5 ([Fig. 10](#fig10){ref-type=\"fig\"} C) resembles that of the treated on day 3 ([Fig. 10](#fig10){ref-type=\"fig\"} B). By day 7, the cCAF-treated wounds appear to be already remodeling, whereas those treated with vehicle looked much like the granulation tissue of cCAF-treated wounds at 5 d ([Fig. 10, E and F](#fig10){ref-type=\"fig\"}). It is known that wound contraction causes the healing tissue to become more dense and compact, therefore our findings strongly suggest that cCAF stimulates early wound contraction and closure.\n\n![**Histology of excision wounds treated with cCAF.** Paraffin-embedded sections of excision wounds treated every other day with vehicle or 1 \u03bcg cCAF were stained with Masson Trichrome. At 3 d after wounding, control wounds (A) have much less dense granulation tissue than those treated with cCAF (B). By day 5, the granulation tissue of the control wounds (C) resembles the tissue of the treated wounds at day 3 (B). The cCAF-treated wounds continue to have denser repair tissue than the control wounds on day 5 (D). By day 7, the difference between the control wounds (E) and cCAF-treated wounds (F) is greatly diminished. Bar, 25 \u03bcm.](0103062f10){#fig10}\n\nTo test whether this acceleration in wound closure was due to an increase in the number of myofibroblasts in the granulation tissue, we immunostained vehicle-treated and cCAF-treated wounds for \u03b1-SMA ([Fig. 11](#fig11){ref-type=\"fig\"}). At day 3, the granulation tissue of the cCAF-treated wounds has many more cells that label for \u03b1-SMA ([Fig. 11](#fig11){ref-type=\"fig\"} B) than does the tissue of vehicle-treated wounds ([Fig. 11](#fig11){ref-type=\"fig\"} A). On day 5, there continue to be many more cells staining for \u03b1-SMA in the cCAF-treated wounds ([Fig. 11](#fig11){ref-type=\"fig\"} D) than in the control tissue ([Fig. 11](#fig11){ref-type=\"fig\"} C). By day 7, this difference has greatly decreased; the treated and untreated wounds have similar staining for myofibroblasts ([Fig. 11, E and F](#fig11){ref-type=\"fig\"}). Quantification of the number of cells labeling for \u03b1-SMA in the wounds shows that there are significantly more myofibroblasts in the cCAF-treated wounds at days 3 and 5, but the difference is no longer significant by day 7 ([Fig. 11](#fig11){ref-type=\"fig\"} G). The amount of \u03b1-SMA in the wounds was not quantified by Western blotting because wounds treated with cCAF contain many more new blood vessels that are surrounded by cells that express \u03b1-SMA. Whereas it is easy to distinguish the blood vessel--associated cells that contain \u03b1-SMA from myofibroblasts in the granulation tissue by microscopy, we would not be able to make that distinction by immunoblot analysis. The earlier appearance of \u03b1-SMA--staining cells in wounds treated with cCAF suggests that the acceleration of wound contraction by cCAF is due to an increase in myofibroblasts at the early stages of wound repair.\n\n![**Immunostaining for** \u03b1**-SMA in the granulation tissue of excision wounds treated with cCAF.** Paraffin-embedded sections of excision wounds treated every other day with vehicle or 1 \u03bcg cCAF were immunostained with an antibody against \u03b1-SMA. At 3 d after wounding, control wounds (A) show almost no cells labeling for \u03b1-SMA, whereas those treated with cCAF show already a significant number of cells labeling for \u03b1-SMA (B). At day 5, both controls (C) and cCAF-treated wounds (D) show more cells expressing \u03b1-SMA, with cCAF-treated wounds showing a much greater increase. By day 7, cCAF-treated wounds (F) continue to have more labeled cells than do control wounds (E), but the difference is greatly diminished. (G) Quantitation of the data in A--F (see Materials and methods for quantitation procedures). Triple asterisk indicates *P* \\< 0.001. Bar, 88 \u03bcm.](0103062f11){#fig11}\n\nDiscussion\n==========\n\nThe work presented here shows that the CXC chemokine, cCAF, when applied repeatedly, suppresses fibroblast growth to \u223c75% of untreated cells, in a time- and dose-dependent manner. This decrease in growth is not due to cell death but rather it correlates with cCAF-induced expression of the myofibroblast marker, \u03b1-SMA. Although the decrease in proliferation takes time to be detected, \u03b1-SMA expression levels rise rapidly following cCAF treatment, strongly suggesting that this chemokine stimulates myofibroblast differentiation directly. The expression of \u03b1-SMA in cCAF-treated cells results in an increase in the ability of the fibroblasts to contract collagen gels. Specific inhibition of \u03b1-SMA expression resulted in abrogation of contraction. All of these effects are specific and can be accomplished by the NH~2~-terminal peptide of the cCAF molecule by itself. Furthermore, in vivo, when applied to excision wounds, this chemokine accelerates wound closure and the granulation tissue becomes more dense early in wound healing. This correlates with an increase in the number of myofibroblasts in the tissue. These results indicate that cCAF stimulates fibroblasts to differentiate into myofibroblasts and strongly suggest that this function may be responsible for the more effective wound closure induced by this chemokine.\n\nAlthough little is known about chemokines and wound closure, expression of cCAF, IL-8, and MGSA is elevated until wound closure and then decreases to low but still elevated levels during granulation tissue formation ([@bib29]; [@bib14]). In addition, knockout mice for CXCR2, a receptor for IL-8 and MGSA, exhibit delayed wound closure ([@bib10]). The results presented here shine light into these finding in vivo. Our observations suggesting that cCAF stimulates \u03b1-SMA expression directly, that this elevation in expression leads to increased contraction of collagen gels and to more rapid wound contraction and closure, indicate that chemokines potentially play significant roles in formation of the granulation tissue of wounds. Furthermore, the small (15 amino acids) NH~2~-terminal peptide of the cCAF molecule has the same effects, suggesting that similar behavior in a human chemokine could be a promising target in designing drugs that affect the differentiation of myofibroblasts.\n\nIt has been shown that myofibroblasts can differentiate from fibroblasts when these cells are exposed to TGF\u03b21 in culture ([@bib9]) and that, in vivo, this growth factor directly stimulates myofibroblast differentiation ([@bib48]). Furthermore, most previously known stimulators of myofibroblast differentiation appear to act indirectly through TGF\u03b21 ([@bib48]). Whether cCAF can stimulate \u03b1-SMA independently of TGF\u03b2 is not known. However, the fact that this chemokine stimulates \u03b1-SMA expression shortly after treatment is initiated, and that it stimulates signals through G-protein--coupled receptors, whereas TGF\u03b21 elicits its effects through serine kinase receptors, suggests that they stimulate different signal transduction pathways to activate \u03b1-SMA expression. Nevertheless, it is possible that these signal transduction pathways share common downstream signaling molecules. We are currently investigating the molecular mechanisms of cCAF-induced \u03b1-SMA production, whether the human homologues of cCAF function in the same manner, and how they relate to the signaling mechanisms stimulated by TGF\u03b2.\n\nCells containing \u03b1-SMA (myofibroblasts, smooth muscle cells, and pericytes) play important functions in a variety of processes involved in wound healing, vasculogenesis/angiogenesis, and pathological conditions, especially in diseases that are characterized by excess scarring. In wound healing, myofibroblasts are particularly important in wound closure and contraction. For example, lack of myofibroblasts after corneal surgery leads to corneal flattening and widening of the wound, whereas stimulation of myofibroblast differentiation in noncontractile fetal wounds leads to contraction of the wounds ([@bib21]; [@bib25]). In disease states characterized by an accumulation of myofibroblasts, such as pulmonary fibrosis and scleroderma, myofibroblasts are thought to contribute significantly to the pathology of the disease, primarily because myofibroblasts tend to be highly fibrogenic ([@bib43]). In addition, myofibroblasts are present in the stroma of many tumors and appear to be important for the survival of these tumors ([@bib8]; [@bib47]).\n\nIn addition to stimulating wound closure through the differentiation of myofibroblasts, cCAF may also be acting to increase the stability of new blood vessels in the granulation tissue. Smooth muscle cells of the new vasculature are known to differentiate from mesenchymal cells in response to signals from the endothelial cells (e.g., [@bib5]). These smooth muscle cells are essential for vascular maturation in connective tissue ([@bib5]). CXC chemokines are produced by the endothelial cells and fibroblasts of the connective tissue and promote angiogenesis ([@bib29]; [@bib30], [@bib31]; [@bib34]; [@bib3]). These chemokines are known to affect endothelial cell migration, but part of their role in the formation of new blood vessels may be in stimulating fibroblasts to acquire \u03b1-SMA and become the smooth muscle cells that stabilize the newly formed vasculature (unpublished data).\n\nBecause myofibroblast accumulation is prominent and high levels of chemokines are present in conditions characterized by excessive scarring, such as keloids, scleroderma, and pulmonary fibrosis ([@bib54]; [@bib41]), it is possible that chemokines may participate in such diseases by stimulating myofibroblast differentiation. For example, levels of MCP-1, IL-8, and MIP-1\u03b1 are high in pulmonary fibrosis ([@bib23]; [@bib20]). These chemokines are also elevated in sclerotic tissue ([@bib22]; [@bib20]). Myofibroblasts in keloids express MGSA, whereas the cells of normal scars do not ([@bib42]). Our results suggest that some of the problems in these conditions may be due to the high levels of chemokines, leading to an increase in myofibroblast numbers, excess deposition of matrix molecules, and contraction of the tissue.\n\nControlling the differentiation of myofibroblasts could mitigate the effects of fibrotic and other diseases. CXC chemokines could be suitable for this purpose because they are not constitutively expressed and do not have the broad-ranging effects that TGF\u03b21 does. In addition, chemokines are very small molecules with no modification other than disulfide bonds, therefore they can be produced using recombinant approaches without much difficulty and they bind to 7-transmembrane receptors, which are highly amenable to pharmacological manipulations. As a consequence, once the mode of action of these proteins on \u03b1-SMA expression is deciphered, they or their antagonists could be used to modulate the presence of myofibroblasts in both disease states and in abnormal wound healing. Furthermore, the ability of the 15 amino acid NH~2~-terminal peptide to stimulate effects similar to those stimulated by the whole cCAF molecule strongly suggests that the peptide itself or peptide mimetics could be used for treatment of wounds with impaired closure.\n\nIn conclusion, a major role of cCAF in the granulation tissue development may be the stimulation of proper wound closure through the stimulation of myofibroblast differentiation. This is a previously unknown function for chemokines and it could represent a novel mechanism for the induction of myofibroblast differentiation. In addition, our results may explain why chemokines contribute to the pathology of fibrotic diseases in which myofibroblasts play a significant part.\n\nMaterials and methods\n=====================\n\nMaterials\n---------\n\nAll tissue culture media and materials were purchased from GIBCO BRL. cCAF was either purified as described in [@bib34] or synthesized by Gryphon Sciences. The 16 amino acid NH~2~-terminal peptide was also synthesized by Gryphon Sciences. The 28 amino acid COOH-terminal peptide was synthesized by Milligen Biosearch. Antibodies used were: anti--\u03b1-SMA (Sigma-Aldrich); anti-TGF\u03b2 1,2,3 (R&D Systems); antivimentin, antidesmin, antimyosin heavy chain (Hybridoma Bank); anti-myoD (Santa Cruz Biotechnology, Inc.); anti--mouse horseradish peroxidase, anti--mouse Texas red (Amersham Pharmacia Biotech); anti--mouse Alexa (Molecular Probes); and anti--mouse FITC (Dako). The anti-cCAF rabbit serum was prepared by Robert Sargeant (Ramona, CA). Other materials used include: nuclear staining To-Pro3 (Molecular Probes), ECL reagents (Amersham Pharmacia Biotech), Vectashield mounting medium (Vector Laboratories), DC protein assay kit (Bio-Rad Laboratories), TRIzol reagent (GIBCO BRL), Vitrogen 100 collagen (Cohesion Technologies), Biocclusive bandage (Johnson and Johnson Medical); and Classic 18S rRNA primers (Ambion). TGF\u03b21 was a gift from A. Roberts (National Institutes of Health, Bethesda, MD).\n\nCell cultures\n-------------\n\nPrimary chicken embryonic connective tissue fibroblasts (CEFs) were cultured as described previously in [@bib34]. Briefly, these fibroblasts were isolated from 10-d-old chicken embryo body walls, plated at 6 \u00d7 10^6^ cells/100-mm plate and cultured for 4 d in 199 medium with 5% donor calf serum, 1% chick serum, and 0.3% tryptose phosphate broth. These cultures were passaged once (secondary cultures); the cells were plated at 0.4 \u00d7 10^6^ cells/35-mm plate in 199/tryptose phosphate broth and 2% donor calf serum.\n\nFibroblast growth assay\n-----------------------\n\nSecondary fibroblasts were plated at 0.4 \u00d7 10^6^ cells/35-mm plate and cCAF was added to experimental plates. For each experiment, plating efficiency was determined by trypsinizing plates 4 h after plating. Cells were counted with a Coulter particle counter to confirm even plating of cells. The media was replaced approximately every 16 h with 1 ml of serum-free 199 medium containing 0.3% tryptose phosphate broth and 2% donor calf serum, and 100--1,000 ng cCAF (9--90 nM) or C- or N-peptide (64--640 nM) was added to the experimental plates, whereas control plates contained media only. On day 3, plates were trypsinized to remove all cells and cells were counted using a Coulter counter. To test for specificity, anti-cCAF rabbit serum was preincubated with cCAF (3 \u03bcl serum/750 ng cCAF \\[68 nM\\] or N-peptide \\[480 nM\\]) for 1 h at room temperature before being added to cells.\n\nTrypan blue staining\n--------------------\n\nThe supernatant of treated and untreated CEFs was collected and the cells centrifuged to a pellet and resuspended in a small volume of medium. 1% trypan blue was added to a final concentration of 0.5% trypan blue. Cells were stained for 5 min and then counted in a hemocytometer. Necrotic cells stained blue.\n\nApoptosis assay\n---------------\n\nTo detect cCAF-induced apoptosis we used the DNA Laddering Detection System from Roche Biochem. Fibroblasts were plated, treated as described above, and the DNA was prepared as suggested by the manufacturer. Briefly, at the end of each treatment, cells were trypsinized, pelleted, resuspended in binding buffer, and incubated for 10 min at room temperature. After this incubation period, isopropanol was added and the preparation vortexed and passed through a filter tube with glass binding fleece. The samples were then centrifuged for 1 min at 8,000 rpm, the flow through was discarded, and the DNA retained in the fleece was washed two times with wash buffer and then eluted with prewarmed elution buffer and used for agarose gel electrophoresis.\n\nImmunostaining for \u03b1-SMA\n------------------------\n\nPlates of fibroblasts were treated with cCAF as described previously. After 4 d, the cells were rinsed with PBS, fixed in 4% paraformaldehyde, permeabilized with 0.15% Triton X-100, and incubated with PBS containing 0.1 M glycine for 10 min. Cells were blocked for 30 min with 10% goat serum in PBS, incubated with mouse anti--\u03b1-SMA to a final IgG concentration of 20 \u03bcg/ml in 1% BSA/PBS for 1 h at room temperature, and washed three times with 0.1% BSA/PBS for 10 min each. The cells were then incubated in To-Pro3 (1:1,000) and goat anti--mouse FITC or sheep anti--mouse Texas red (1:100) in 1% BSA/PBS for 1 h at room temperature, washed three times for 10 min with 0.1% BSA in PBS, and mounted with Vectashield. Confocal fluorescence microscopy was performed on a ZEISS LSM510. Collagen gels were rinsed with PBS, fixed in 4% paraformaldehyde for 2 h, washed three times 30 min each with PBS, and incubated for an additional 30 min in PBS containing 0.1 M glycine. This treatment was followed by incubation at 4\u00b0C O/N in 15% sucrose and then incubation under the same conditions in 30% sucrose. After a brief rinse with PBS, the gels were frozen in OCT, sections were prepared and collected on gelatin-coated slides, rinsed with PBS, fixed in 4% paraformaldehyde for 10 min, and then incubated again in PBS containing 0.1 M glycine for 10 min. This was followed by blocking for 30 min with 10% goat serum in PBS, incubation in primary and secondary antibody as described above, and mounting with Vectashield.\n\nImmunoblotting\n--------------\n\nPlates of fibroblasts were treated as described previously with 750 ng/ml cCAF or 2.5 ng/ml TGF\u03b2. To block TGF\u03b2 or cCAF activity, anti-cCAF (3 \u03bcl) or anti-TGF\u03b1 (1 \u03bcl) antibody was preincubated in 1 ml media for 1 h at room temperature before being added to cells. After 4 d of treatment, protein extracts were prepared in 1 ml 150 mM RIPA buffer containing protease inhibitors. Protein concentrations were determined using the DC protein assay kit and samples were adjusted to contain equal amounts of protein. SDS-PAGE was performed on 7.5% separating Doucet gels ([@bib12]). Protein transfer to nitrocellulose was performed using a wet-transfer apparatus (Bio-Rad Laboratories) at 100 V for 45 min. The membranes were blocked for 1 h in 5% milk in TTBS and then incubated overnight at 4\u00b0C in anti--\u03b1-SMA (1:1,500), antivimentin (1:100), antidesmin (1:100), antimyosin heavy chain (1:100), or anti-myoD (1:200) in 1% milk in TTBS. The membranes were washed three times for 20 min each with TTBS, incubated in anti-mouse HRP at 1:10,000 in 1% milk for 1 h, and washed as above; and the bands were visualized using the ECL.\n\nCollagen gel contraction\n------------------------\n\n1.5 ml Vitrogen 100 collagen gels were made in 35-mm plates. Secondary fibroblasts were plated at 0.4 \u00d7 10^6^ cells/35-mm plate on top of the gels. After the cells had adhered to the collagen, cCAF was added to experimental plates. The medium was replaced approximately every 16 h with 1 ml of 199 supplemented with 0.3% tryptose phosphate broth and 2% donor calf serum, and 750 ng cCAF (68 nM) or N-peptide (480 nM) was added to the experimental plates. Control plates contained media only. After 4 d, gels were released. Treatment was continued for two more days and gels were photographed every 12 h. The photographs were used for evaluation of gel contraction by determining the area of the gel using NIH image analysis. For the experiments involving inhibition of \u03b1-SMA expression, we used phosphorothioate antisense ODN, which was synthesized and HPLC purified by Sigma-Aldrich Genosys, to block the \u03b1-SMA production. The sequence of ODN is specific for the 3\u2032-untranslated region of \u03b1-SMA (5\u2032-CACAGTAATATGCTAAAAAGAC-3\u2032) as described previously ([@bib46]; [@bib39]). The sense strand of this ODN was used as control (5\u2032-GTCTTTTTAGCATATTACTGTG-3\u2032). When preparing the collagen gels, 2 \u03bcM of antisense or sense ODN were incorporated into the gel before the seeding of CEFs. The treatments and measurement of gel contraction were performed as described above, except that 2 \u03bcM of antisense or sense ODN were applied each time the chemokine or peptide was applied.\n\nRT-PCR\n------\n\nTotal RNA for \u03b1-SMA was extracted using TRIzol reagent from untreated fibroblasts, fibroblasts treated with 750 ng/ml cCAF, or fibroblasts treated for varying periods of time with cCAF. The RT-PCR procedure was performed using the Promega Access RT-PCR System, which is designed to finish RT and PCR in one tube and following the protocol recommended by Promega, except that 1.5 times the recommended amount of dNTP, reverse transcriptase, and Tfl DNA polymerase were used to ensure strong synthesis of 18S ribosomal RNA. The reaction conditions included: 1 \u03bcg total RNA, first strand synthesis at 48\u00b0C for 45 min, then 95\u00b0C for 5 min to inactivate the reverse transcriptase, followed by DNA amplification at 95\u00b0C for 45 s, 58\u00b0C for 60 s, 68\u00b0C for 90 s for 40 cycles. Finally, 68\u00b0C for 7 min to extend the strands. 3 \u03bcl of Quantum mRNA classic 18S primers (Ambion) were added to the reaction to produce the control band. The primers used for the amplification of \u03b1-SMA were: sense primer 5\u2032-GGAGCACCTGAGGACATTGAC-3\u2032 and antisense primer 5\u2032-GCTTCAGTCAGCAGAGTTGGG-3\u2032. RT-PCR products were analyzed by electrophoresis in 1.5% agarose and the density of the bands was measured by densitometry using Glyko BandScan.\n\nWounding experiment\n-------------------\n\nFull-thickness excision wounds (\u223c0.5 \u00d7 0.5 cm) were made using a scalpel blade on the underside of the wings of 2-wk-old chicks. The left wing was treated with vehicle alone (water) and the right wing with 1 \u03bcg cCAF (90 nM). The wounds were photographed immediately after wounding and then covered with Biocclusive bandage. 50 \u03bcl vehicle (water) or cCAF was deposited through the bandage onto the wound using a 30-gauge needle. This procedure was repeated the next day and every other day thereafter. On days 3, 5, and 7 the bandages were removed and the wounds were photographed before replacing the bandages and applying the treatment again. Wings were placed on a flat surface at a distance of 25 cm from the digital camera and were always photographed with the same camera settings. All images were printed at the same magnification.\n\nPreparation and staining of wing sections\n-----------------------------------------\n\nAt the specified time points, chickens were killed with sodium pentobarbitol. The wounded wings were collected and fixed for 18 h in 4% paraformaldehyde and decalcified for 3 d in 5% formic acid, 2.5% formaldehyde at 4\u00b0C. The tissue was embedded in paraffin and sectioned. Sections were stained with Masson Trichrome to visualize interstitial collagen. Other sections were immunolabeled for \u03b1-SMA. Sections were deparaffinized three times with 15 min washes in Hemo-De and rehydrated in 5 min washes of ethanol (100, 95, 70, 50, and 30%). After rinsing with PBS, sections were fixed in 2% paraformaldehyde 1 h. Autofluoresence and nonspecific staining was blocked by 30 min in 0.1 M glycine in PBS, followed by 30 min in 1% Evans Blue in PBS to quench red blood cell autofluorescence. Sections were incubated with \u03b1-SMA antibody in 1% BSA in PBS (1:50) for 2 h at room temperature. After three 10 min washes in 0.1% BSA in PBS, the sections were incubated for 40 min with anti--mouse Alexa antibody in 1% BSA in PBS (1:200). They were then washed and mounted with Vectashield. To quantify the number of myofibroblasts in the wounds, the number of fluorescently labeled cells in four high power fields of each wound was counted and statistical analysis was applied.\n\nStatistical methods\n-------------------\n\nSignificance was determined using Student\\'s *t* test for comparison between two means and ANOVA for comparison between more than two means. All data were examined to assure homogeneity of variance. Means were considered significantly different when *P* \\< 0.05.\n\nWe thank Dr. Andrew Grosovsky for the use of his Coulter counter and Ibsen Chen for embedding and sectioning the wing tissue. We also thank Shuyan Lu for her suggestions and stimulating discussions.\n\nThis work was supported in part by a Dissertation Research Grant from the University of California, Riverside, and a GAANN fellowship (US Department of Education) to J.E. Feugate and by National Institutes of Health grant GM48436 to M. Martins-Green.\n\nAbbreviations used in this paper: CAM, chicken chorioallantoic membrane; cCAF, chicken chemotactic and angiogenic factor; CEF, embryonic connective tissue fibroblast; IL, interleukin; MGSA, melanocyte growth--stimulating activity; ODN, oligodeoxynucleotide; SMA, smooth muscle actin.\n\n[^1]: Address correspondence to Dr. Manuela Martins-Green, Department of Cell Biology and Neuroscience, University of California, Riverside, CA 92521. Tel.: (909) 787-2585. Fax: (909) 787-4286. E-mail: <mmgreen@ucrac1.ucr.edu>\n"}
+{"text": "Introduction {#s1}\n============\n\nTick-borne encephalitis virus (TBEV) is a flavivirus that belongs to the *Flaviviridae* family. Flaviviruses comprise many human pathogens including the commonly known Dengue virus (DENV), Japanese encephalitis virus (JEV), West Nile virus (WNV), Yellow fever virus (YFV), and Zika virus (ZIKV) ([@B1]). With respect to TBEV, three subtypes of the virus exist: European (TBEV-Eu), Siberian (TBEV-Sib), and Far Eastern (TBEV-FE) ([@B2]).\n\nTBEV is transmitted to humans primarily from infected ticks, mainly from the *Ixodes* family. The virus can also be transmitted from unpasteurized dairy products from infected livestock ([@B3]--[@B5]). Infection with TBEV causes tick-borne encephalitis (TBE), an acute viral infection that affects the central nervous system (CNS) with often severe long-term neurological consequences ([@B3], [@B4], [@B6], [@B7]). The first TBE-like disease was described as early as in the eighteenth century in Scandinavian church records ([@B8]). Traditionally, the disease is described as a syndrome with a biphasic course beginning with an influenza-like illness followed by a second neuroinvasive phase with neurological symptoms of variable severity, ranging from meningitis to severe meningoencephalitis with or without myelitis ([@B3], [@B4], [@B6]) (Figure [1](#F1){ref-type=\"fig\"}). It shall be noted, however, that also monophasic patterns of disease development have been described ([@B9]). Upon infection, virus is detected in serum in the first phase of the disease but rarely in the second phase ([@B10]).\n\n![Overview of the classic biphasic disease-pattern of human TBEV infection. The viremic first phase includes influenza-like symptoms and occurs around 1 week after virus transmission. It is estimated that 65--70% of infected individuals clear the virus after this phase, but for one third of the patients, an asymptomatic disease phase follows before the second phase of disease begins. In this phase, symptoms of meningitis or encephalitis occur, including fever, headache, tremor, nystagmus, altered state of consciousness, cranial nerve paralysis, and spinal nerve paralysis. Classically, no virus is detected in sera or plasma in the second phase of disease. Around 30% of patients that enter the second phase of disease will suffer from long lasting sequeale, with a decreased quality of life. Figure compiled from Lindquist and Vapalahti ([@B3]), Taba et al. ([@B4]), and Haglund and Gunther ([@B6]).](fimmu-09-02174-g0001){#F1}\n\nDue to increased geographic distribution of TBEV as well as a marked increase in morbidity in many areas, TBEV-infection has more recently caught attention as a public health problem. TBE is now observed in large parts of Europe as well as in northern Asia ([@B3], [@B4]). The main risk areas for TBE in Europe are primarily parts of central and eastern Europe as well as the Baltic and Nordic countries. With respect to central Europe, risk areas extend from Switzerland in the west into northern Italy and the Balkan countries ([@B11]). The incidence of TBEV-infection in endemic countries varies from year to year ([@B12]--[@B14]), however, an overall upsurge has been reported in certain parts of Europe, including the borders between Austria, Slovenia, and Italy ([@B15], [@B16]). These changes have been related to climatic, ecological, environmental, and socioeconomic factors that all can lead to an increased risk of human exposure to infected ticks ([@B17]--[@B20]).\n\nThe total number of annual cases has been estimated to be up to 13,000, and as such the infection constitutes the most important tick-borne viral disease ([@B4]). More than 30% of patients with clinical symptoms from TBEV-infection develop prolonged sequelae, some of which may become life-long including neuropsychiatric symptoms, severe headaches, and a general decrease in quality of life ([@B3], [@B4], [@B6], [@B7]). The mortality rates differ between the strains. Infection with the Far Eastern strain (TBEV-FE) has a mortality rate of 5--35%, whereas the other two strains (TBEV-Eu and TBEV-Sib) have mortality rates of 1--3% ([@B3], [@B4]). There is no specific treatment (e.g., antivirals) for TBE; rather, symptomatic treatment is the only available option ([@B3], [@B4], [@B9]).\n\nOf importance, TBE may be prevented by vaccination. There are in total four licensed vaccines to TBE. Two vaccines based on TBE-Eu subtype are licensed in Europe and two are licensed in Russia. Additionally, a TBEV-vaccine based on the Far Eastern subtype is produced and marketed in China. All vaccines are based on formalin-inactivated strains of TBEV ([@B3], [@B4], [@B21], [@B22]). In areas where the disease is highly endemic, WHO recommends that vaccination should be offered to all groups above 1 year of age ([@B4], [@B23]). Primary vaccination against TBE includes three doses of the vaccine within the first year, followed by revaccinations every third to fifth year to maintain immunity. Vaccination is generally considered effective and TBE incidence has decreased substantially in TBEV-endemic regions with successful vaccination-programs ([@B24]). Randomized controlled trials in large populations have shown high immunogenicity with often-strong antibody production and acceptable rates of adverse events following vaccination ([@B25]--[@B28]). Breakthrough TBE after vaccination is generally considered rare ([@B4]). However, over the last years, vaccine failures have been reported, in particular in middle-aged and elderly individuals, who have completed the primary vaccination ([@B29]--[@B31]).\n\nInfection with TBEV triggers humoral and cell-mediated immune responses. A confirmed diagnosis of TBE is established by the detection of specific IgM and IgG in serum. IgM antibodies have been observed in sera very early in symptomatic TBE disease, whereas IgG antibodies peak in the convalescent phase of disease ([@B32]). IgG antibodies can persist over lifetime and prevent TBE ([@B4], [@B33]). Early after clinical disease onset, TBEV-specific antibodies can also be found in the cerebrospinal fluid (CSF) ([@B32], [@B33]). In contrast to the humoral immune response, the cell-mediated immune responses elicited to natural infection have been rather poorly studied until recently. The latter responses may contribute both to host resistance against infection as well as to pathological reactions affecting the target organ of the virus, i.e., primarily the CNS.\n\nHere, we review recent progress in studies of the cell-mediated immune response to human TBEV infection. A particular emphasis is devoted to natural killer (NK) cell- and T cell-mediated responses. Responses to TBEV are discussed in context of cell-mediated immune responses toward other flavivirus infections. We also discuss some immunopathological aspects of TBE with a particular emphasis on cell-mediated immune reactions in the CNS. Cell-mediated immune reactions in the CNS may contribute to neural damage with severe consequences of brain function, and could in the worst cases lead to fatal outcome. First, however, some aspects of the TBEV itself are covered.\n\nTBEV and other flaviviruses {#s2}\n===========================\n\nAll flaviviruses are enveloped and have a positive-sense single stranded RNA genome, which *per se* acts as messenger RNA upon entrance in the host cell. The RNA encodes for a polyprotein, which is co- and post-translationally cleaved by viral and cellular proteases into three structural proteins; capsid (C), precursor membrane (prM) and envelope glycoproteins (E), and seven non-structural proteins including NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 ([@B34]). Flaviviruses enter the cell through clathrin-dependent endocytosis upon attachment of the E protein to a receptor. Heparan sulfate has been identified as such receptor for TBEV ([@B35]); however, there are most likely also other yet not identified receptors for the virus. Following cell entry, the flavivirus is delivered to endosomes ([@B36]), in which the low pH triggers the E protein to fuse with the endosomal membrane and the nucleocapsid is released into the cytosol. The assembly of immature flavivirus virions ([@B36]--[@B38]), including TBEV ([@B39], [@B40]), occurs in the ER, and the viral particles are transported to the Golgi apparatus. The virion particles are immature until the envelope protein is rearranged and prM is cleaved by the host enzyme furin in the acidic environment in the Golgi apparatus. Immature particles are non-infectious and proteolytic cleavage of prM is a prerequisite for viral infectivity. However, studies have shown that complete cleavage of prM is not necessary for viral infectivity ([@B41]--[@B43]).\n\nIn general, species of flaviviruses have many similarities, but their preferred host cells differ. TBEV is shown to replicate 10,000-fold higher in human neuronal cells as compared with epithelial cells ([@B44]). A similar infection pattern has recently been shown for ZIKV ([@B45]).\n\nNK cells {#s3}\n========\n\nNK cells are innate lymphocytes, though recent studies have revealed \"adaptive\" features of these cells ([@B46], [@B47]). They are perhaps best known for their ability to kill virus-infected and tumor cells. NK cell cytotoxicity is regulated by the expression of numerous activating and inhibitory receptors that sense ligands on neighboring healthy and altered cells. Several activating receptors recognize molecules that are up-regulated on cells during conditions of cellular stress, such as viral infection and transformation \\[reviewed in ([@B48], [@B49])\\] whereas many inhibitory receptors, e.g., human killer cell Ig-like receptors (KIR) bind to HLA class I molecules. Additionally, NK cells have an important role in producing cytokines and chemokines, as well as by other means interacting with other immune and non-immune cells.\n\nHuman NK cells are classically defined as CD3^\u2212^ (T cell receptor negative), CD56^+^ cells and represent about 15% of peripheral blood lymphocytes. These cells have for long been divided into two main subsets; CD56^bright^ and CD56^dim^ cells ([@B50]). The CD56^bright^ NK cells are thought to be less mature and are commonly known as primarily cytokine-producing cells with low cytotoxic ability, whereas CD56^dim^ NK cells are best known for their potent cytotoxic activity upon target cell recognition ([@B51]). However, the latter are also ample cytokine producers upon interaction with target cells ([@B51]). Both \"natural\" and antibody-mediated NK cell cytotoxicity is mediated by exocytosis of cytoplasmic granules containing perforin and granzymes ([@B52]). Cytotoxic responses may also to various degrees involve TRAIL- and Fas-ligand-mediated induction of apoptosis ([@B53], [@B54]). CD56^dim^ NK cells frequently express CD16 (Fc\u03b3RIII), KIRs, and CD57, which regulate their function and define distinct stages of NK cell maturation ([@B55]), whereas CD56^bright^ NK cells largely lack expression of these molecules.\n\nThe role of NK cells in human TBEV infection {#s4}\n============================================\n\nDirect evidence for a protective role of NK cells has been found in experimental models of several viral infections, including cytomegalovirus and influenza, and a number of studies have indicated that they play a role also in protection against viral infections in humans. For example, NK cell-deficiencies in humans result in severe herpes virus infections in childhood and adolescence ([@B56]). NK cells may also have a protective role in human TBEV-infection. At the same time, responses mediated by these cells may be associated with development of symptoms in the course of TBEV-infection. Although there is only little known about NK cells in TBE, NK cells have been detected in CSF of patients with TBE ([@B57]), an observation that indicates transmigration through the blood brain barrier (BBB).\n\nTo gain a better understanding of the NK cell response to human TBEV-infection, we recently performed a longitudinal study providing an in-depth analysis of the human NK cell response to acute TBEV-infection in a well-defined cohort of TBE patients. The study had an emphasis on NK cell responses during the second stage of disease from which clinical samples were available. NK cell activation, as measured by expression of the proliferation marker Ki67, was apparent at the time of hospitalization ([@B58]) (illustrated in Figure [2](#F2){ref-type=\"fig\"}). Concomitant with the increase in NK cell activation in the acute stage of disease, augmented levels of IL-12, IL-15, IL-18, IFN-\u03b3, and TNF were detected in patient plasma. In parallel with high levels of activation, the activated NK cells expressed less perforin, granzyme B, and Bcl-2. By 3 weeks after hospitalization, the NK cell activation decreased to levels seen in healthy controls. This TBEV-induced NK cell activation was restricted predominantly to more differentiated CD57^+^CD56^dim^ NK cells. Functionally, CD56^dim^ NK cells responded poorly to target cells at the time of hospitalization, but they recovered functional capacity to healthy control levels during the convalescent phase. The poor functionality of NK cell responses was exclusive for target cell recognition, since NK cell responses induced by IL-18 and IL-12 remained unchanged throughout the disease ([@B58]).\n\n![Overview of the cell-mediated immune response to TBEV-infection. CD56^dim^ NK cells (mainly highly differentiated CD57^+^ cells) are highly activated at the time of hospitalization during the second phase of disease. They express significantly higher levels of Ki67, CD38, and produce less cytokines in response to target cells as compared to the convalescent phase. NK cells then become fully normalized, comparable to healthy control levels, at the convalescent phase as assessed 3 months after hospitalization. CD4 T cells show a similar pattern as the CD56^dim^ NK cells. They are activated at hospitalization and express significantly higher levels of Ki67, CD38, HLA-DR, and Granzyme B as compared to the convalescent phase. CD4 T cells have retracted to normal healthy control levels at the convalescent phase. CD8 T cells show a different pattern and peak in activation at 1 week after hospitalization. They express significantly higher levels of Ki67, CD38, HLA-DR, Granzyme B, Perforin, PD-1, T-bet, and Eomesodermin as compared CD8 T cells at the convalescent phase. The CD8 T cell activation subsequently return to normal healthy control levels at the convalescent phase, just as is the case for CD4 T cells and NK cells. Figure compiled from Blom et al. ([@B58]) and Blom et al. ([@B59]).](fimmu-09-02174-g0002){#F2}\n\nNK cell responses toward other acute flavivirus infections in humans {#s5}\n====================================================================\n\nTo be able to interpret the above-mentioned NK cell responses to acute TBEV infection, it is important to understand NK cell responses to other acute virus infections, including acute flavivirus infections. In this respect, NK cells have to various extent been studied *ex vivo* in other acute flavivirus infections, including DENV ([@B60], [@B61]) and WNV ([@B62], [@B63]) as well as hepatitis C virus (HCV), a distant relative within the *Flaviviridae* family ([@B64], [@B65]). NK cells have also been studied after vaccination with the live attenuated YFV 17D vaccine ([@B66]--[@B68]). They have been suggested to influence disease severity and outcome, and to contribute to viral control in these infections, even though underlying mechanisms are not well studied.\n\nIn this context, it was observed that the absolute number of NK cells in patients with a mild form of infection with DENV was higher as compared to patients with the more severe form of the infection, dengue hemorrhagic fever (DHF) ([@B69]). Reduced numbers of NK cells in the circulation may be indicative of migration toward peripheral target organs. Furthermore, a higher frequency of NK cells expressing CD69 early on during the infection in children developing severe DHF has been reported ([@B70]). In recent studies of DENV-infection, we found NK cells to be robustly activated during the first week after symptom debut. Here, the response seemed to be confined largely to the CD56^bright^ subset of NK cells and less mature CD56^dim^ NK cells (our own unpublished studies). Noteworthy in the context of acute TBEV infection, activation of NK cells may also occur very early, even before the onset of symptoms. This possibility is supported by the observation that the highest levels of NK cell activation in most TBEV infected patients were observed already at the time of hospitalization ([@B58]). This notion is corroborated in studies of YFV vaccinated individuals. An early response by NK cells was observed in study subjects vaccinated with YFV-17D, where expression of both Ki67 and CD69 was increased on NK cells as early as 1 week after vaccination ([@B66]). Accumulation of adaptive-like NK cells expressing the activating receptor CD94/NKG2C has been reported in some human viral infections ([@B71]--[@B73]); however, no expansion of NKG2C^+^ NK cells in blood has been observed in TBEV-infection or any other flavivirus infection ([@B58]). It can, however, not be excluded that this type of expansion could occur locally, e.g., at the site of infection.\n\nIn addition to the observed activation of NK cells *in vivo* in different flavivirus infections, a protective role of NK cells is also supported by *in vitro* data. For example, primary activated human NK cells have been shown to inhibit WNV-infection of Vero cells ([@B63]) and IFN\u03b1-activated NK cells can kill HCV-infected hepatoma cells *in vitro* ([@B65]). In addition, flavivirus-infected target cells have been reported to display virus-mediated up-regulation of MHC class I ([@B74]), and could thereby theoretically evade lysis from NK cells by engaging inhibitory receptors. The dampened NK cell responses to target cells in acute TBEV-infection further support this notion ([@B58]). On the other hand, increased MHC class I expression could result in enhanced T cell responses. In such a scenario, one may speculate that flaviviruses may have been driven more toward escape from innate immunity rather than from adaptive T cell immunity ([@B75]).\n\nT cells {#s6}\n=======\n\nIn contrast to NK cells, CD8 and CD4 T cells recognize specific foreign peptide sequences presented by HLA class I and II molecules, respectively ([@B76]). Like NK cells, major functions of CD8 T cells are to kill infected cells through the release of perforin and granzymes, and to secrete cytokines such as IFN-\u03b3, TNF, and IL-2. The cytotoxic T cell response to acute infection can typically be divided into three phases; priming and expansion, resolution and contraction, and memory formation. During the first phase, na\u00efve CD8 T cells divide and differentiate into effector cells acquiring high cytotoxic ability ([@B77]). Following viral clearance, the effector T cell population contracts and the majority of the pathogen-specific T cells enter apoptosis. A small pool of pathogen specific T cells (5--10%) survives as memory cells in the third stage ([@B78]). Memory T cells are a principal component of immunity against intracellular pathogens such as viruses. They are distinguished by their capacity to survive long-term, and undergo rapid and robust proliferation and acquisition of effector function upon antigen re-exposure ([@B78]). Memory T cells can vary in their phenotype, localization, and function allowing them to protect the host against a broad array of potential insults.\n\nDistinct stages of CD8 T cell differentiation are defined by the expression of specific surface markers such as the isoforms of CD45 and expression of the homing receptor CCR7. These stages of differentiation are useful in the characterization of responses to, e.g., anti-viral responses. The set stages define CD45RA^+^CCR7^+^ as naive (T~N~), CD45RA^\u2212^CCR7^+^ as central memory (T~CM~), CD45RA^\u2212^CCR7^\u2212^ as effector memory (T~EM~), and CD45RA^+^CCR7^\u2212^ as effector memory RA (T~EMRA~) CD8 T cells ([@B79], [@B80]). T~CM~ cells primarily reside in secondary lymphoid organs, possess the greatest proliferative potential among the memory T cell subsets and can rapidly expand and differentiate following re-challenge. T~CM~ cells have higher sensitivity to antigenic stimulation, are less dependent on co-stimulation and provide better feedback to DCs and B cells compared to T~N~ cells. T~EM~ cells can migrate between tissues and secondary lymphoid organs and provide immune surveillance.\n\nThe role of T cells in human TBEV infection {#s7}\n===========================================\n\nDue to difficulties in identifying the acute phase of viral infection in humans, T cell responses to viral infections have to a large extent been addressed in studies of pathogens causing chronic infections such as HIV-1, EBV, HCV, and CMV ([@B81]--[@B85]). Such responses can be very robust, as exemplified by the massive clonal expansion of antigen-specific CD8 T cells seen in many infections ([@B83], [@B85]). Based on these studies, it has also become clear that the resulting populations of human CD8 T cells display striking phenotypic differences, as determined by the expression profiles of surface markers ([@B79]--[@B81]). In contrast to many other infections, including some flavivirus infections ([@B86]--[@B89]), there are only few studies of T cell responses to TBEV-infection in humans. This hold true for acute as well as cases with prolonged TBE disease. Noteworthy, however, one report has shown that TBEV-specific CD4 T cells from naturally infected patients show a higher level of polyfunctionality in response to antigen in the convalescent phase of disease, as compared to TBE-vaccine specific T cells ([@B90]).\n\nThe general lack of studies more systematically characterizing the human T cell response to TBEV-infection prompted us to study the primary T cell-mediated immune response in patients diagnosed with TBE with a particular emphasis of CD8 T cells ([@B59]). Similar to our study on NK cells ([@B58]), the T cell study focused on responses during the second stage of disease from which clinical samples were available. During this phase, CD8 T cells were strongly activated, as detected by increased expression of Ki67, within 1 week of hospitalization (illustrated in Figure [2](#F2){ref-type=\"fig\"}). A large part of these CD8 T cells expressed high levels of perforin and granzyme B, and low levels of the anti-apoptotic protein Bcl-2. In contrast to CD8 T cells, CD4 T cells showed only low or at most moderate levels of activation. The TBEV-antigen specific CD8 T cells had a T~EM~ PD-1^+^ phenotype throughout the course of disease. TBEV-specific CD8 T cells were predominantly Eomes^+^Ki67^+^T-bet^+^ in the acute stage of disease. This pattern was replaced by an Eomes^\u2212^Ki67^\u2212^T-bet^+^ profile in the convalescent phase of disease. TBEV-specific CD8 T cells were mainly monofunctional in the acute stage of disease, and tended to become more polyfunctional in the convalescent phase when clinical symptoms retracted ([@B59]).\n\nT cell responses toward other acute flavivirus infections in humans {#s8}\n===================================================================\n\nTo be able to better interpret the above-mentioned T cell responses to acute TBEV infection, we compared the present results with T cell responses to other acute virus infections, including infections by other flaviviruses. The live attenuated YFV-vaccine strain can replicate after vaccination leading to a detectable viral load similar to a mild infection. Thus, this vaccine can be utilized as a controlled model to study mild acute viral infection in humans. CD8 T cells become activated within 1--2 weeks after vaccination with the YFV vaccine ([@B86], [@B91], [@B92]). YFV antigen-specific CD8 T cells predominantly display a T~EM~ PD-1^+^ phenotype, which transition into a T~EMRA~ PD-1^\u2212^ memory phenotype ([@B86]). With respect to DENV-infection, a high level of functionality of DENV-specific T cells is associated with a better disease outcome ([@B93]). Similarly, patients hospitalized with (severe) TBE show a low level of T cell functionality in the acute stage of disease ([@B59]), indicating the importance of high function among virus-specific T cells for beneficial disease outcome. CD8 T cells have been shown primarily to respond with IFN-\u03b3 to JEV in asymptomatic JEV-exposed donors ([@B87]).\n\nActivation of CD4 T cells with an optimal magnitude, specificity and kinetics may be a requirement for viral clearance and protective immunity. In the immune response induced by the YFV vaccine, activation of CD4 T cells (peak at 10 days after vaccination) precedes that of CD8 T cells, and this may be of importance to elicit strong immunological memory ([@B86]). Furthermore, CD4 T cell release of IFN-\u03b3 may have an impact on disease outcome since CD4 T cells, and not CD8 T cells, were shown to dominate the IFN-\u03b3 response in recovered Japanese encephalitis (JE) patients. In addition, a high quality polyfunctional CD4 T cell response can be associated with better disease outcome in JE patients ([@B87]). In murine models, a perforin-dependent mechanism by the CD8 T cells has been shown to clear WNV from infected neurons, thereby suggesting an immunopathological role of T cells in mice ([@B88]). In this context, it is of interest to note that TBEV-specific T cells have a high content of both perforin and granzyme B ([@B59]), but whether the same perforin-dependent mechanism is causing immunopathogenesis in acute infection with TBEV remains to be investigated.\n\nCross-reactivity within the family of flaviviruses {#s9}\n==================================================\n\nImmunological cross-reactivity between TBEV and other species within the flavivirus family may also contribute to disease ([@B94], [@B95]). Antibody-dependent enhancement (ADE) is a described phenomenon that can occur when non-neutralizing antibodies facilitate virus entry into host cells, leading to increased infectivity in the cells. ADE is commonly observed *in vitro* in cell culture-based models ([@B96]), but it is questioned as to which degree this phenomenon occurs *in vivo*. A recent study from non-human primates *in vivo*, did not observe increased ZIKV titers after prior infection with heterologous flaviviruses ([@B97]). Protective cross-reactivity of flaviviruses has been reported as well, opposing increased pathogenesis upon pre-exposure to other species of flaviviruses ([@B98]). TBEV has been suggested to cause both pathogenic and protective cross-reactivity. Polyclonal sera against members of the TBE serocomplex (including TBEV, Kyasanur Forest disease virus, Omsk hemorrhagic fever virus, and Langat virus) enhance viral replication of TBEV *in vitro* ([@B96]). However, it has recently been shown that antibodies generated from TBEV infection or from the TBEV vaccine can mediate cross neutralization against other, if not most, of the members of the TBE virus complex ([@B99]). Furthermore, sera from some individuals vaccinated against TBEV and JEV neutralized WNV, and the neutralization was enhanced by YFV vaccination in some recipients ([@B95]), altogether indicating that previous flavivirus exposure may sometimes provide a degree of protection to new flaviviruses.\n\nCell-mediated immunity and cross-reactivity caused by TBEV and other flaviviruses has been less well studied. Recently, a study demonstrated that JEV- and JE vaccine-specific T cells cross-react with DENV ([@B87], [@B100]). In line with this, it was also recently shown that vaccination with YFV vaccine could induce ZIKA-specific T cells, thereby suggesting cross-protection of flavivirus-specific T cells ([@B101]). The latter phenomena opened up a discussion as to the possibility of utilizing the YFV vaccine to protect against Zika virus infection ([@B101]). In the present context, it remains to be investigated whether YFV vaccination elicits protective cross-reactive immunity also toward TBEV.\n\nImmunopathological aspects of TBEV infection in the CNS {#s10}\n=======================================================\n\nIn contrast to the significant interest in emerging infections such as the recent Zika pandemic \\[reviewed in ([@B45], [@B102], [@B103])\\], studies of the immune response toward TBEV-infection as such, and TBEV-induced immunopathology in particular, have been rather limited.\n\nImmune and none-immune mechanisms have been proposed contribute to the crossing of TBEV over the BBB and invasion of the CNS \\[reviewed in ([@B104])\\]. Cytokines may facilitate this process. Cytokines such as TNF-a and IL-6 have an impact on endothelial cell permeability that may induce a BBB disruption ([@B105], [@B106]), leading to crossover of the virus into the CNS. A distinct mechanism by which the TBEV could possibly cross the BBB is the *Trojan Horse* mechanism ([@B107]), by which TBEV-infected immune cells such as dendritic cells, neutrophils, monocytes, macrophages, and T cells would migrate into the parenchymal compartment causing infection of neurons or other cells in the brain and the spinal cord. Yet, an alternative route is invasion via the olfactory epithelium ([@B108], [@B109]).\n\nAfter the landmark discovery of the lymphatic system present in the meninges that connects the CNS to the peripheral blood ([@B110], [@B111]), the classical concept of the CNS as an immune privileged site has been replaced by a view of an immune regulated site. Hence, under normal conditions a continuous transmigration of lymphocytes, monocytes, DCs and macrophages occurs. They may serve to detect any kind of infection or injury in the brain \\[reviewed in ([@B112])\\]. Similar to other CNS infections, increased frequencies of T cells have been reported in CSF of TBE patients ([@B57]). Hence, activated T cells are crossing the BBB in TBE; however, the role of T cells at this site not well understood ([@B113], [@B114]). On one hand they could contribute toward clearing viral infection but on the other hand they may mediate immunopathology within the CNS. Corroborating the latter speculations are findings in which granzyme B^+^ CD8 T cell infiltrates have been linked to cell-death in infected human neuronal tissue ([@B113]) and, in parallel, mice with CD8 T cell deficiency have been shown to have prolonged survival upon infection with TBEV compared to mice with adoptively transferred CD8 T cells to immuno-competent mice ([@B114]). Furthermore, studies of post-mortem tissue of TBE patients have shown a predominance of macrophages/microglia and CD3^+^ T cells (both CD4^+^ and CD8^+^) in brain parenchyma ([@B113]). As seen in other flavivirus infections, macrophages and microglia also play a role in tissue destruction in human TBE.\n\nIn relation to NK cells and their possible role in causing immunopathogenesis, it is of interest to note that also these cells have been detected in the cerebrospinal fluid (CSF), albeit in low numbers, in patients with severe TBE meningitis or encephalitis ([@B57]). Activated NK cells may be protective, but they may also, like T cells, take part in immunopathological reactions as they are known to participate in direct killing of infected cells, indirect killing through cytokines or chemokines, or by the recruitment of inflammatory cells into the tissues ([@B115], [@B116]). Although recent results support a role for NK cells in clinical TBEV-infection, more studies are needed to provide a better understanding of the role NK cells play in pathogenic processes of human TBE infection.\n\nKnowledge and experience gained in the field of the immunopathogenesis of other diseases affecting the CNS and its immunological compartments could be helpful in understanding TBE-specific diseases patterns. For example, in multiple sclerosis (MS), an inflammatory disease with pathology affecting the CNS ([@B117]), the concentration of the Sphingosine-1-Phosphate (S1P) in CSF is elevated and S1P-signaling is altered. In MS, binding of S1P to S1P1-receptors expressed on lymphocytes leads to invasion of autoreactive T cells into the CNS, the latter contributing to the hallmarks of the disease including demyelination and neurodegeneration ([@B118]). Interestingly, during the phase of acute infection in TBEV-infected patients, the levels of S1P in blood and CSF are highly elevated ([@B119]). This increase might promote a proinflammatory response. An increased production of extracellular S1P can be regulated by modulators of the S1P pathway, such as fingolimod, which is an immunomodulatory drug used in the treatment of MS ([@B118]). Therefore, therapeutic options used in other CNS diseases that share common immunopathogenic mechanisms with TBE could be used as models to aid in the development of new strategies for TBE treatment.\n\nConcluding remarks {#s11}\n==================\n\nIn the present review, we have focused our attention to recent insights into the cell-mediated immune response to human TBEV infection, with an emphasis on studies of NK cell and CD8 T cell mediated responses. Until recently, the latter have been poorly studied. As yet, however, much more needs to be learnt with respect to these responses and research in this area should be encouraged. We have also addressed some aspects of TBEV CNS pathogenesis, a process still far from understood in detail. Clearly, however, cell-mediated immune responses likely play an important role in this process. As TBE continues to be an increasing global health problem and challenge, much more research is needed into this emerging disease. Several areas of research of the TBEV itself, and the clinical disease TBE merit further studies. Not the least, the specific organ pathogenesis caused by TBEV and the immune response, including infiltrating immune cells, needs more investigation. Furthermore, the possibility of antiviral treatment and other possible treatment modalities needs much more thorough investigation to prevent disease development and the often severe sequeale following infection of humans with TBEV.\n\nAuthor contributions {#s12}\n====================\n\nKB, AC, and H-GL wrote the manuscript. All other authors provided valuable contributions and insights into the manuscript.\n\nConflict of interest statement\n------------------------------\n\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\n\n**Funding.** The authors were founded by the Swedish Research Council, the Swedish Foundation for Strategic Research, Karolinska Institutet and the Stockholm County Council.\n\n[^1]: Edited by: Alan Chen-Yu Hsu, University of Newcastle, Australia\n\n[^2]: Reviewed by: Alessandro Marcello, International Centre for Genetic Engineering and Biotechnology, Italy; Manuela Zlamy, Innsbruck Medical University, Austria\n\n[^3]: This article was submitted to Microbial Immunology, a section of the journal Frontiers in Immunology\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nSeveral flaviviruses, such as the yellow fever virus (YFV), Japanese encephalitis virus (JEV), dengue virus (DENV), Zika virus (ZIKV), West Nile virus (WNV), and tick-borne encephalitis virus, are important human pathogens. Flaviviruses are spread worldwide, though some species show a pronounced restriction to defined regions of endemicity, such as YFV to sub-Saharan Africa and tropical Latin America and JEV to Southeast Asia and the Asia-Pacific. Certain flaviviruses, such as DENV, WNV, and, most recently, ZIKV, are (re-)emerging in new areas ([@B1][@B2][@B3]). Some evidence suggests the first autochthonous transmission of JEV in Africa ([@B4]).\n\nYellow fever (YF) is an acute viral hemorrhagic disease which is currently endemic to \u223c50 countries with \u223c1 billion people living at risk of infection. Despite the availability of a highly efficient vaccine (YFV-17D; e.g., Stamaril), an estimated \u223c0.2 million YFV infections with 29,000 to 60,000 deaths occur annually ([@B5]). Recent YFV outbreaks in Angola (2015-2016), the Democratic Republic of the Congo (2016), Brazil (2017), and Nigeria (2018) and a shortage of the YF vaccine supply raised serious concerns about the preparedness for future outbreaks ([@B6], [@B7]). Since the Aedes aegypti mosquito, the main YFV vector, is omnipresent in (sub)tropical Asia, YFV spillover to Asia and the establishment of epidemics involving urban transmission become increasingly realistic ([@B8], [@B9]).\n\nFor JEV, several licensed inactivated and live attenuated vaccines, including Ixiaro (inactivated vaccine) and JE-CVax (Imojev; YFV-17D-based chimeric live attenuated vaccine \\[c-LAV\\]), are available ([@B10], [@B11]). Another YFV-17D-based tetravalent c-LAV, namely, against dengue (CYD-TDV, Dengvaxia), has reached marketing licensure and is being introduced in some countries/regions. However, there are serious concerns related to the use of this vaccine in dengue virus-seronegative individuals, mainly because of aggravation of dengue disease by preexisting antibodies (antibody-dependent enhancement \\[ADE\\]) of DENV infection ([@B12][@B13][@B14]).\n\nVaccination against flaviviruses generally relies on the strategy to mount protective humoral immunity against structural proteins, in particular neutralizing antibodies (nAbs) elicited against the viral envelope (E) protein ([@B5], [@B15]), though also CD4^+^ T cells seem to contribute to the protective activity of current YFV vaccines ([@B16]). Nonetheless, experimental evidence obtained with mice and nonhuman primates for YFV ([@B17][@B18][@B20]) and, more recently, in mice also for WNV ([@B21]) and ZIKV ([@B22]) clearly shows that also nonstructural (NS) proteins, in particular NS1, can evoke protective humoral and cellular immune responses. Of note, NS1 is not part of the infectious flavivirus particle and thus not a target of nAbs. Likewise, immunization with an adenovirus vector encoding the NS3 protein of YFV-17D elicited strong CD8^+^ T cell responses, which resulted in some degree of protection in mice against subsequent challenge ([@B23]). However, full protection was observed only when the vaccine included the structural proteins of YFV-17D as the antigen as well ([@B23], [@B24]), obviously in line with the accepted role nAbs play in YFV infection. Thus, besides humoral immune responses against the E protein, cellular immune responses against the NS proteins may to some extent also contribute to immunity against flaviviruses. However, no flavivirus vaccines that are based on any of these NS proteins as the target antigen have been developed or licensed yet for human use. Intriguingly, the genome of chimeric flavivirus vaccines (JE-CVax/Imojev or CYD-TDV/Dengvaxia) consists of sequences of antigenically distinct flaviviruses (respectively, JEV and YFV-17D or DENV and YFV-17D) and may therefore exert some dual protective activities. Here we demonstrate that vaccination of mice with a construct similar to JE-CVax/Imojev provides rapidly complete protection against a massively lethal YFV challenge, with a single dose being sufficient for full efficacy. Moreover, we show that this protection is, notwithstanding its unexpected potency, mediated not by nAbs but by multiple complementary and vigorous responses directed against the NS proteins of YFV-17D.\n\nRESULTS {#s2}\n=======\n\nJE-CVax provides full dual protection against lethal JEV and YFV challenge in mice. {#s2.1}\n-----------------------------------------------------------------------------------\n\nJE-CVax is a c-LAV that consists of the YFV-17D genome, of which the prM and E genes have been replaced by the corresponding sequences of JEV SA14-14-2. AG129 mice were vaccinated with either 103 or 104 PFU of JE-CVax and 28\u2009days later challenged with 103 PFU (equivalent of 1,000 50% lethal doses \\[LD50\\]) of YFV. This resulted in, respectively, 80 or 100% survival, while YFV infection was uniformly lethal in all nonvaccinated controls (see [Fig.\u00a0S1A](#figS1){ref-type=\"supplementary-material\"} in the supplemental material). Therefore, throughout the rest of the study, animals were vaccinated and challenged with 10^4^ PFU of JE-CVax (full survival in vaccinated mice) and 10^3^ PFU of YFV (full mortality in nonvaccinated mice). JE-CVax was originally developed as a JEV vaccine. As expected, unlike nonvaccinated animals (*n*\u2009=\u200916), all AG129 mice vaccinated with either JE-CVax (10^2^, 10^3^, or 10^4^ PFU; *n*\u2009\u2264\u20096) or the inactivated JEV vaccine Ixiaro (*n*\u2009=\u200910; 2 times at 1\u2009\u03bcg: twice 1/6th human dose) ([@B25]) were completely protected (*P* \\> 0.0001) against lethal JEV challenge ([Fig.\u00a01A](#fig1){ref-type=\"fig\"}). Remarkably, vaccination with JE-CVax resulted also in 97% survival (*n*\u2009=\u200935/36) against a massively lethal YFV challenge ([Fig.\u00a01B](#fig1){ref-type=\"fig\"}). All placebo-vaccinated (*n*\u2009=\u200938) and Ixiaro-vaccinated (*n*\u2009=\u200912) animals had to be euthanized for humane reasons (mean days to euthanasia \\[MDE\\], 14.6\u2009\u00b1\u20092.8\u2009days and 15.4\u2009\u00b1\u20093.5, *P* \\> 0.0001). Importantly, JE-CVax also conferred similarly vigorous protection against YFV in C57BL/6 wild-type (wt) mice (*n*\u2009=\u200916) against intracranial (i.c.) challenge with 10^4^ PFU of YFV ([Fig.\u00a01D](#fig1){ref-type=\"fig\"}). In AG129 mice, a benefit (60% survival) could already be observed 7\u2009days postvaccination (dpv). At 14 dpv or later, all animals were fully protected against lethal challenge ([Fig.\u00a01C](#fig1){ref-type=\"fig\"}). To establish that JE-CVax-mediated protection against YFV is specific and not the result of some residual cross-reactivity as previously observed for certain flaviviruses in mice ([@B26]), we challenged age-matched nonvaccinated (*n*\u2009=\u20097) or JE-CVax-vaccinated and YFV-17D-challenged (*n*\u2009=\u20096) AG129 mice with 10^4^ PFU of the more distantly related ZIKV (strain MR766). No protective activity was observed (MDE for nonvaccinated and vaccinated mice, 23.5\u2009\u00b1\u20095.4 and 17.4\u2009\u00b1\u20098.8\u2009days; *P* = 0.4831) ([Fig.\u00a0S1B](#figS1){ref-type=\"supplementary-material\"}). Thus, a single-dose immunization with JE-CVax provides fast (\u226414 dpv) and virus-specific protection against lethal YFV exposure in mice.\n\n![*In vivo* evaluation of JE-CVax-mediated dual protection against lethal JEV SA14-14-2 and YFV-17D challenge. (A to D) AG129 and C57BL/6 mice were first vaccinated via the i.p. route with either 10^4^ PFU of JE-CVax (blue), 1/6th of a human dose of Ixiaro (green), or assay medium as a negative control (red). Animals vaccinated with Ixiaro were boosted with another 1/6th of a human dose of Ixiaro 14 dpv. In order to facilitate vaccine virus replication ([@B28]), wild-type C57BL/6 mice receiving JE-CVax vaccination were treated with MAR1-5A3 antibody. AG129 mice were i.p. challenged with 10^3^ of PFU JEV SA14-14-2 at 28 dpv (A) or with 10^3^ PFU of YFV-17D at 28 dpv (B) or at 0, 4, 7, 14, 21, and 28 dpv (C). C57BL/6 mice were i.c. challenged with 10^4^ PFU of YFV-17D at 28 dpv (D). Animals were observed for 5 weeks after challenge and were euthanized when humane endpoints were reached. The data represent cumulative results of at least two independent experiments. Log rank (Mantel-Cox) survival analysis test was performed for statistical significance. \\*\\*, *P* \u2264\u20090.01; \\*\\*\\*\\*, *P*\u2009\u2264\u20090.001 compared to the nonvaccinated group.](mBio.02494-19-f0001){#fig1}\n\n10.1128/mBio.02494-19.1\n\n*In vivo* characterization of protective efficacy of JE-CVax vaccination against lethal YFV-17D and ZIKV-MR766 challenge in AG129 mice. (A) AG129 mice (*n*\u2009=\u20095) were vaccinated i.p. with 10^3^ to 10^4^ PFU of JE-CVax (blue) and at 28 dpv challenged i.p. with 10^3^ PFU of YFV-17D. (B) AG129 mice (*n*\u2009=\u20096) were first vaccinated i.p. with 10^4^ PFU JE-CVax and at 28 dpv challenged i.p. with 10^3^ PFU of YFV-17D (blue). At 28 days post-YFV-17D challenge, animals were challenged a second time with 10^4^ PFU of ZIKV-MR766 and observed for mortality for the following 5 weeks. Age-matched nonvaccinated (red) animals were challenged with 10^3^ PFU of YFV-17D (*n\u2009*=\u20095 \\[A\\]) or 10^4^ PFU of ZIKA-MR766 (*n\u2009*=\u20097 \\[B\\]) as a control. Log rank (Mantel-Cox) survival analysis test was performed for statistical significance. \\*\\*, *P*\u2009\u2264\u20090.01 compared to the nonvaccinated group. Download FIG\u00a0S1, JPG file, 0.2 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nJE-CVax mediates protection against YFV without involvement of nAbs. {#s2.2}\n--------------------------------------------------------------------\n\nTo explore whether humoral immunity is involved in JE-CVax-mediated protection against YFV, serum of AG129 mice (i) at day 0 (prevaccinated), (ii) infected with YFV-17D before euthanasia (terminal serum), (iii) vaccinated with JE-CVax (day 28; postvaccinated), or (iv) vaccinated with JE-CVax and challenged with YFV-17D (day 56; postchallenge) was analyzed for total binding antibodies and nAb. All animals vaccinated with JE-CVax or Ixiaro seroconverted to JEV positivity. Sera of nonvaccinated animals that had been infected with YFV-17D showed only some residual reactivity for JEV (as detected by indirect immune fluorescence assay \\[IIFA\\] \\[[Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"}\\]). In contrast, nAbs against JEV were exclusively detected in serum samples of JE-CVax- or Ixiaro-vaccinated animals (50% neutralizing antibody titer as determined using CPE-based virus neutralization tests \\[log~10~ CPENT~50~\\], 2.48 \u00b1 0.29 or 1.86 \u00b1 0.36, respectively) ([Fig.\u00a02A](#fig2){ref-type=\"fig\"}). Only when JE-CVax-vaccinated mice were challenged at a later stage with YFV-17D were nAbs against the latter virus raised (log~10~ CPENT~50~, 1.66 \u00b1 0.30; determined 28 days after YFV exposure). Also in serum of JE-CVax- or Ixiaro-vaccinated C57BL/6 mice, only nAbs against JEV (log~10~ CPENT~50~, 1.66 \u00b1 0.12 or 1.61 \u00b1 0.09, respectively) were detectable. However, all the C57BL/6 mice intraperitoneally (i.p.) vaccinated with 10^4^ PFU YFV-17D in the presence (*n*\u2009=\u20098) or absence of MAR1 (*n*\u2009=\u20099) developed nAbs against YFV (log~10~ CPENT~50~, 2.11 \u00b1 0.25 or 1.78 \u00b1 0.29, respectively) and survived lethal i.c. challenge of YFV. Thus, neither JE-CVax nor Ixiaro induces YFV-specific nAbs in mice.\n\n![Serological analysis of serum of JE-CVax-vaccinated and YFV-17D- or ZIKV-MR766-challenged animals. (A) Detection of nAbs against JEV and YFV. CPE neutralization tests (CPENT) for JE-CVax (circles) and YFV-17D (squares) were performed on sera day 0 prior to vaccination (preimmune, red), day 28 after vaccination (blue), and after challenge (study endpoint, orange) for samples of JE-CVax-vaccinated AG129 mice, of JE-CVax-vaccinated mice after subsequent YFV-17D-challenge (*n*\u2009=\u200934), of mice hyperimmunized with JE-CVax (*n*\u2009=\u200913; first bleed 2 weeks post last booster immunization, blue), and of mice vaccinated with Ixiaro (green). Limit of detection (LOD) for virus neutralization was log~10~ 20 (1.3). Data are presented as log~10~ CPENT~50~ (mean \u00b1 SD). The data presented are from \u22653 independent experiments. Statistical significance was determined using one-way ANOVA. \\*\\*\\*\\*, *P* \u2264\u20090.0001 for mean log~10~ CPENT~50~ titers against JEV or YFV compared to mean log~10~ CPENT~50~ titers before JE-CVax vaccination and before YFV-17D-challenge, respectively. (B) Quantitation of anti-YFV NS1 binding antibodies by direct ELISA. Serum from naive, nonvaccinated mice (red) or mice that had been vaccinated with 10^3^ to 10^5^ of PFU JE-CVax (blue) or that had been infected with 10^4^ PFU of YFV-17D (orange) or with 10^5^ PFU of ZIKV-MR766 (pink) were collected either 28\u2009days postimmunization or when euthanized at the humane endpoint (*n*\u2009\u2265\u20095). The data are means of two independent analyses. Statistical significance was determined using one-way ANOVA. \\*\\*, *P*\u2009\u2264\u20090.01 compared to YFV-17D. (C) Binding of serum antibodies to NS1-expressing cells. HEK 293 cells were transfected with a plasmid expressing YFV-17D NS1 as a transcriptional fusion to GFP (top) or infected with the YFV-17D-mCherry reporter virus (bottom). Either 48 h after transfection or 72 h after infection, cells were stained with the anti-YFV NS1-specific MAb 1A5 (left), with serum from mice that were vaccinated with JE-CVax (center), or with serum from naive, nonvaccinated mice (right). Graphs show flow cytometric analysis of GFP or mCherry fluorescence and visualization of anti-YFV NS1 antibody binding using a PE-Cy7-conjugated goat anti-mouse IgG secondary antibody. The fraction of NS1-positive cells (GFP or mCherry) stained by MAb 1A5 or serum of JE-CVax-immunized mice (anti-mouse IgG) is given as a percentage in the upper right quadrant. Data from one representative experiment out of four independent experiments are shown.](mBio.02494-19-f0002){#fig2}\n\n10.1128/mBio.02494-19.2\n\nDetection of cross-reactive antibodies against JEV and YFV in sera of YFV-17D-infected and JE-CVax-vaccinated mice. AG129 mice were prebled before either infection with 10^3^ PFU of YFV-17D or vaccination with 10^4^ PFU of JE-CVax/Ixiaro and bled again either at the onset of sickness (YFV-17D) or at 28 dpv (JE-CVax). Preserum (A), serum of YFV-17D infected mice (B), serum of JE-CVax-vaccinated mice (C), and serum of Ixiaro-vaccinated mice (D) were analyzed by both JEV (top) and YFV (bottom) indirect immunofluorescence assay (IIFA; Euroimmun) at a magnification of \u00d720. Download FIG\u00a0S2, JPG file, 0.4 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nJE-CVax and YFV-17D induce comparable levels of anti-NS1 YFV antibodies. {#s2.3}\n------------------------------------------------------------------------\n\nFrom the IIFA analysis ([Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"}), it is obvious that sera from JE-CVax-vaccinated mice contain cross-reactive but nonneutralizing Abs against YFV-17D ([Fig.\u00a02A](#fig2){ref-type=\"fig\"}). These non-nAbs may possibly be attributed to NS1, a strong B cell antigen. To assess the presence of anti-YFV NS1 antibodies in JE-CVax-vaccinated AG129 mice, a direct enzyme-linked immunosorbent assay (ELISA) was performed on sera from mice vaccinated with 10^3^ to 10^5^ PFU of JE-CVax or mice infected with 10^4^ PFU of YFV-17D or 10^5^ of PFU ZIKV-MR766. Serum was obtained either at the onset of disease (YFV-17D, 10^5^ PFU of JE-CVax and ZIKV) or at 28 dpv. Levels of anti-YFV NS1 antibodies in the different JE-CVax-vaccinated groups were statistically not different (*P*\u2009\\>\u20090.05) from those in the YFV-17D-infected groups ([Fig.\u00a02B](#fig2){ref-type=\"fig\"}). Moreover, very low cross-reactivity was noted for samples from ZIKV-infected mice. These findings were also confirmed by flow cytometry analysis ([Fig.\u00a02C](#fig2){ref-type=\"fig\"}). Serum antibodies from JE-CVax-vaccinated mice bound to cells that overexpress YFV NS1 as well as to cells that had been infected with YFV-17D. In fact, serum from JE-CVax-vaccinated mice resulted in staining comparable to that obtained when monoclonal antibody (MAb) 1A5, specifically directed against YFV NS1, was used ([@B27]).\n\nJE-CVax induces YFV-specific antibodies mediating ADCC. {#s2.4}\n-------------------------------------------------------\n\nTo determine the potential mechanism that non-nAbs elicit for protection against YFV, an antibody-dependent cellular cytotoxicity (ADCC) reporter bioassay was carried out using HEK293T cells infected with YFV-17D (expressing mCherry) ([@B28]) as target cells and murine Fc\u03b3 receptor IIIA (Fc\u03b3RIIIa)-expressing Jurkat reporter cells as effector cells ([@B29]). Hyperimmune mouse serum from JE-CVax-vaccinated AG129 mice induced clear ADCC responses and in a dose-dependent manner, whereas serum from nonvaccinated mice failed to do so ([Fig.\u00a03](#fig3){ref-type=\"fig\"} and [Fig.\u00a0S3A](#figS3){ref-type=\"supplementary-material\"} and [B](#figS3){ref-type=\"supplementary-material\"}).\n\n![Role of antibody-dependent cell-mediated cytotoxicity (ADCC) conferred by JE-CVax hyperimmune serum in the protection against YFV. JE-CVax hyperimmune serum was tested for its ability to mediate ADCC activity compared to serum of nonvaccinated mice (normal serum) at 3:1 (A) and 10:1 (B) effector (E)-to-target (T) cell ratios. Experiments were conducted twice, each in triplicate, and data are presented as means \u00b1 SEMs for fold changes compared to control (CC) (i.e., mean reporter signal plus three SDs from E:T in the absence of hyperimmune serum). Values from noninfected target cells incubated with E in the presence of either hyperimmune serum or normal serum at highest antibody concentrations (dilution 1:9) are indicated as Control-9. Statistical significance was determined using two-way ANOVA. \\* and \\*\\*\\*\\*, *P* \u2264\u20090.05 and\u20090.0001 compared to normal serum.](mBio.02494-19-f0003){#fig3}\n\n10.1128/mBio.02494-19.3\n\nEffect of JE-CVax hyperimmune serum in antibody-dependent cellular cytotoxicity (ADCC). JE-CVax hyperimmune serum (circles) was tested for its ability to mediate ADCC activity in comparison to serum of nonvaccinated mice (squares) at 3:1 (A) and 10:1 (B) effector (E)-to-target (T) ratios. Data are from experiments conducted twice, each in triplicate, and presented as means \u00b1 SEMs. The average of relative light unit signal plus three standard deviations from E:T ratios in the absence of hyperimmune serum was considered the background signal (CC). Statistical significance was determined using two-way ANOVA. \\*, \\*\\*, \\*\\*\\*, and \\*\\*\\*\\*, *P* \u2264\u20090.05, 0.01, 0.001, and\u20090.0001 compared to normal serum. Download FIG\u00a0S3, JPG file, 0.2 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nJE-CVax induces polyfunctional T cell responses against both YFV and JEV antigens. {#s2.5}\n----------------------------------------------------------------------------------\n\nTo assess whether also cellular immune response against YFV may contribute to the protective activity, enzyme-linked immunosorbent spot (ELISpot) assays (tumor necrosis factor alpha \\[TNF-\u03b1\\] and/or gamma interferon \\[IFN-\u03b3\\]) and intracellular staining of cytokines (TNF-\u03b1 and IFN-\u03b3) were performed on total splenocytes obtained from AG129 mice (*n*\u2009=\u20095) and C57BL/6 mice (*n*\u2009=\u200910) 18 and 4 weeks after JE-CVax immunization, respectively. Unlike the case with splenocytes of nonvaccinated mice ([Fig.\u00a0S6E](#figS6){ref-type=\"supplementary-material\"} to [H](#figS6){ref-type=\"supplementary-material\"}), robust and specific IFN-\u03b3 and/or TNF-\u03b1 production was observed from splenocytes of either mouse strain vaccinated with JE-CVax ([Fig.\u00a04A](#fig4){ref-type=\"fig\"} to [C](#fig4){ref-type=\"fig\"} and [Fig.\u00a0S4](#figS4){ref-type=\"supplementary-material\"}) after recall with a major histocompatibility complex (MHC) class I-restricted peptide derived from YFV-17D NS3 or YFV-17D total cellular antigen. In line, flow cytometric analysis revealed robust and YFV-specific intracellular cytokine production in CD4^+^ and CD8^+^ T cells from spleens of JE-CVax-vaccinated AG129 and C57BL/6 mice when stimulated *ex vivo* with the YFV NS3 peptide or YFV-17D total cellular antigen ([Fig.\u00a04D](#fig4){ref-type=\"fig\"} and [Fig.\u00a0S4](#figS4){ref-type=\"supplementary-material\"} to [S6](#figS4){ref-type=\"supplementary-material\"}). Overall, JE-CVax vaccination induced specific long-lasting T cell responses against YFV, and cellular immunity against YFV was more vigorous than that elicited against JEV ([Fig.\u00a04A](#fig4){ref-type=\"fig\"} to [C](#fig4){ref-type=\"fig\"} and [Fig.\u00a0S4](#figS4){ref-type=\"supplementary-material\"} to [S6](#figS4){ref-type=\"supplementary-material\"}). This observation is consistent with YFV NS proteins serving as more immunogenic T cell antigens than prM/E of JEV ([@B30], [@B31]).\n\n![Detection of protective T cell responses directed against YFV. (A to C) ELISpot assay data showing TNF-\u03b1 (A) IFN-\u03b3 (B and C) and production by splenocytes of AG129 mice (*n*\u2009=\u20095; A and B) or C57BL/6 (*n*\u2009=\u200910; C) at 18 and 4 weeks, respectively, after vaccination with 10^4^ PFU of JE-CVax, following 16 h *ex vivo* restimulation with either an MHC class I-restricted peptide derived from YFV-17D NS3 ([@B32]) or the lysate of YFV-17D- or JEV SA14-14-2-infected Vero E6 cells. Stimulation using lysate of noninfected Vero E6 cells served as a negative control. The data are derived from two independent experiments. Spot counts were normalized by subtraction of the number of spots in corresponding wells stimulated with uninfected Vero E6 cell lysate. (D) Cytokine expression profile of YFV-specific T cells. Shown are IFN-\u03b3 and TNF-\u03b1 production profiles of YFV-specific CD4^+^ and CD8^+^ T cells from JE-CVax-vaccinated AG129 and C57BL/6 mice 18 and 4 weeks, respectively, postvaccination, as determined by intracellular cytokine staining. Mouse splenocytes were stimulated 16 h *ex vivo* with either an MHC class I-restricted NS3 peptide, cell lysate of YFV-17D-infected Vero E6 cells, or lysate of uninfected Vero E6 cells. The data are derived from two independent experiments and normalized by subtraction of number of cytokine-secreting T cells in corresponding samples in which uninfected Vero E6 cell lysate was used as recall antigen. (E) T cell-mediated *in vivo* protection against YFV. Loss of protection resulting from antibody-mediated T cell depletion ([@B30], [@B52]) suggests a direct functional involvement of CD4^+^ and CD8^+^ T cells in JE-CVax-mediated immunity against YFV in C57BL/6 mice (*n*\u2009\u2265\u20097) that had been vaccinated with 10^4^ PFU of JE-CVax and subsequently challenged intracranially with 10^4^ PFU of YFV-17D. Depletions were performed by administration of 0.5\u2009mg of anti-mouse CD4 and/or anti-mouse CD8 antibodies i.p. on day \u22122 and day 0 each prior to YFV challenge. Log rank (Mantel-Cox) survival analysis test was performed for statistical significance. \\* and \\*\\*, *P* \u2264\u20090.05 and\u20090.01 compared to vaccinated group (*n*\u2009=\u20095); ^+^, *P*\u2009\u2264\u20090.05 compared to CD4-depleted group (*n*\u2009=\u20098).](mBio.02494-19-f0004){#fig4}\n\n10.1128/mBio.02494-19.4\n\nT cell responses directed against YFV or JEV antigens in ELISpot assay. Representative wells of ELISpot assays showing TNF-\u03b1 (A) and IFN-\u03b3 (B and C) production by splenocytes of AG129 mice (A and B) and C57BL/6 mice (C) at 18 and 4 weeks, respectively, after vaccination with 10^4^ PFU JE-CVax, following 16 h of *ex vivo* restimulation with an MHC class I-restricted peptide derived from YFV NS3^32^ or the lysate of YFV-17D- or JEV SA14-14-2-infected Vero E6 cells. Stimulation using lysate of noninfected Vero E6 cells served as a negative control. Download FIG\u00a0S4, JPG file, 0.3 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02494-19.5\n\nDetection of T cell responses in mice by intracellular staining for TNF-\u03b1 and IFN-\u03b3 using flow cytometry. Shown are results from flow cytometric analysis for intracellular TNF-\u03b1 (A, B, E, and F) and IFN-\u03b3 (C, D, G, and H) production by CD4^+^ (A, C, E, and G) or CD8^+^ (B, D, F, and H) T cells from vaccinated AG129 mice (A to D) and C57BL/6 mice (E to H) following stimulation with either NS3 MHC class I peptide or cell lysate of YFV-17D or JEV SA14-14-2-infected Vero E6 cells. The percentage of total CD4^+^ or CD8^+^ TNF-\u03b1- or IFN-\u03b3-secreting T cells analyzed in flow cytometric analysis in AG129 mice (*n*\u2009=\u20095) and C57BL/6 mice (*n*\u2009=\u200910) was determined. The tables in panels I and J represent *P* values between cytokine-secreting populations of antigen- versus noninfected Vero E6 cell-stimulated samples (statistical significance; paired *t* test) in flow cytometric analysis for splenocytes from AG129 and C57BL/6 mice, respectively. The data were compiled from two independent experiments, and dotted lines represent average background in control samples collected from non-vaccinated animals. Download FIG\u00a0S5, JPG file, 0.5 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02494-19.6\n\nDetection of T cell responses through intracellular cytokine staining in vaccinated (A to D) AG129 and C57BL/6 mice versus nonvaccinated baseline controls (E to H). Representative depiction of flow cytometric analysis for intracellular TNF-\u03b1 and IFN-\u03b3 production by CD4^+^ (gated CD3^+^ CD8^--^) (A and C) or CD8^+^ T cells (gated CD3^+^ CD8^+^) (B and D) from JE-CVax-vaccinated (A and B) AG129 mice, 18 weeks postvaccination and (C and D) C57BL/6 mice, 4 weeks postvaccination, following 16h *ex vivo* stimulation with lysate of YFV-17D- or JEV SA14-14-2-infected Vero E6 cells, an MHC class I-restricted NS3 peptide, or uninfected Vero E6 cells. Results of similar analyses of nonvaccinated AG129 (E and F) and C57BL/6 mice (G and H) are shown. Download FIG\u00a0S6, JPG file, 0.7 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nBoth CD4^+^ and CD8^+^ T cells contribute to JE-CVax-mediated protection against YFV. {#s2.6}\n-------------------------------------------------------------------------------------\n\nTo determine whether YFV-specific T cell responses directly contribute to the JE-CVax-mediated protection against YFV, T cell depletion experiments were performed with C57BL/6 mice ([@B32]). Animals at 5 weeks postvaccination were administered with anti-mouse CD4 and/or anti-mouse CD8a T cell-depleting antibodies twice, i.e., 2 days before and immediately prior to YFV challenge. Unlike for vaccinated (but not further treated) animals (that were included as immunization controls \\[*n*\u2009=\u20095\\] and that survived an intracranial YFV challenge), in vaccinated but antibody-treated mice the previously observed full protection against YFV was partially lost by targeting CD4^+^ (*n*\u2009=\u20091/8), CD8^+^ (*n*\u2009=\u20093/7), and CD4^+^ CD8^+^ (*n*\u2009=\u20095/7) T cells for depletion ([Fig.\u00a04E](#fig4){ref-type=\"fig\"}). All nonvaccinated animals (*n*\u2009=\u20095; *P* = 0.0027) succumbed to YFV challenge as before ([Fig.\u00a01D](#fig1){ref-type=\"fig\"}). The mortality resulting from T cell depletion, especially the increased mortality observed in the doubly depleted animal group, suggests that in the absence of nAb both CD4^+^ and CD8^+^ T cells contribute to the JE-CVax-mediated protection against YFV in C57BL/6 mice.\n\nCTL epitopes within the YFV-17D NS proteins have higher binding affinity toward human MHC class I molecules than those of mice. {#s2.7}\n-------------------------------------------------------------------------------------------------------------------------------\n\nCytotoxic T lymphocyte (CTL) epitopes and their binding affinities toward respective MHC class I molecules are different in different species. *In silico* analysis of YFV-17D NS proteins was performed to compare the probable CTL epitopes and their binding efficiencies, using an algorithm that takes peptide processing, presentation, and MHC-peptide complex stability into account. Data presented in [Fig.\u00a0S10](#figS10){ref-type=\"supplementary-material\"} indicate that as a predictor of immunogenicity, peptides of YFV-17D NS proteins have higher binding affinity to human MHC than to mouse MHC.\n\nDISCUSSION {#s3}\n==========\n\nNeutralizing antibodies against the E protein are generally considered a primary correlate of protection against flaviviruses ([@B5], [@B15]). However, some preclinical studies suggested that several NS proteins when used as immunogens for vaccination alone or in combination could induce some degree of protection against viral challenge in mice and nonhuman primates ([@B17][@B18][@B20]). For instance, a recombinant vaccinia virus or replication-deficient adenoviral vectors expressing YFV-17D NS1, NS2a, and NS2b together or NS3 alone, respectively, resulted in some partial protective immunity against a lethal challenge of YFV-17D in mice ([@B17], [@B23]). However, full survival could never be achieved and reached maximally 60 to 80%, versus 100% for YFV-17D-vaccinated controls. Similarly, 80% protection against challenge with the African YFV strain Dakar 1279 was observed in monkeys following repeated immunization with purified NS1 as the sole vaccine antigen, and protection correlated with the levels of non-nAbs against NS1 ([@B19]). Conversely, vaccination of mice with E protein alone (or in combination with NS1 or NS3) resulted in complete protection ([@B23], [@B24]).\n\nLike Imojev (commercially available live attenuated JE vaccine), our live attenuated JE-CVax vaccine is a chimeric flavivirus that consists of the genome of YFV-17D from which the prM/E genes have been replaced by those of JEV SA14-14-2, and hence, it is nearly identical to Imojev (i.e., a retroengineered genetic copy) but not formally the same live attenuated virus. We therefore used JE-CVax to assess whether it can offer, besides protection against JEV, protection against YFV challenge. Since AG129 mice are highly susceptible to lethal JEV SA14-14-2 and YFV-17D infection ([Fig.\u00a0S7](#figS7){ref-type=\"supplementary-material\"}) ([@B28], [@B33]), we used these two vaccine strains as established surrogates for wt JEV ([@B33]) and wt YFV ([@B34], [@B35]), respectively (demanding lower biosafety containment for handling). A single dose of JE-CVax provided, in addition to the expected protection against JEV challenge, nearly complete protection ([@B35], [@B36]) against a massive (1,000 LD~50~\\] challenge with YFV ([Fig.\u00a01B](#fig1){ref-type=\"fig\"}). The protective activity against YFV was raised fast, and a survival benefit could be observed already within 7\u2009days after vaccination ([Fig.\u00a01C](#fig1){ref-type=\"fig\"}). These observations were further corroborated by the fact that vaccination of immunocompetent C57BL/6 mice with a single dose of JE-CVax provided complete protection (21/21) against lethal intracranial challenge ([@B17], [@B23]) with YFV ([Fig.\u00a01D](#fig1){ref-type=\"fig\"} and [Fig.\u00a04E](#fig4){ref-type=\"fig\"}). This activity of JE-CVax was YFV specific, as AG129 mice that had been vaccinated with Ixiaro (the inactivated JEV vaccine) were not protected against lethal YFV challenge. Furthermore, mice that had been vaccinated with JE-CVax and that later survived YFV challenge did not survive a subsequent lethal challenge with the ZIKV ([Fig.\u00a0S1](#figS1){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.02494-19.7\n\n*In vivo* infectivity of YFV-17D and JEV SA14-14-2 in AG129 mice. AG129 mice were inoculated via the i.p. route with different doses of YFV-17D (10^0^ PFU \\[orange solid line; *n*\u2009=\u200916\\], 10^1^ PFU \\[orange dashed line; *n*\u2009=\u20093\\], or 10^2^ PFU \\[orange dotted line; *n*\u2009=\u20093\\]) (A) or JEV SA14-14-2 (10^0^ PFU \\[purple solid line\\], 10^1^ PFU \\[purple dashed line\\], or 10^2^ PFU \\[purple dotted line\\]; *n* =3) (B). Animals were monitored over a period of 5 weeks and were euthanized when humane endpoints were reached. Download FIG\u00a0S7, JPG file, 0.2 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nOthers have shown in mice that cross-reactive antibodies (together with cross-reactive T cells) may confer partial protection against flaviviruses from different serocomplexes, as demonstrated for JEV (both Vero cell-derived JEV-P3 strain-based inactivated vaccine and JEV SA14-14-2)-vaccinated mice challenged with DENV ([@B26]). Likewise, a chimeric Japanese encephalitis/dengue virus 2 experimental vaccine, ChinDENV (originally designed to induce immunity against DENV-2 prM/E), was shown to protect against both JEV and DENV-2 challenge in mice ([@B31], [@B36]). The protection observed could, however, still largely be explained by the induction of considerable levels of E-specific partially cross-reactive nAbs that neutralized both DENV and JEV. Additionally, the study ([@B31]) demonstrated that vaccination also induces JEV antigen-specific T cell responses (T cells producing IFN-\u03b3 and interleukin 2 \\[IL-2\\] following stimulation with JEV antigen), suggesting a possible contribution of the cellular immunity in the defense against JEV challenge in ChinDENV-immunized mice. Another study, in which AG129 and IFNAR mice were used, reported that during heterotypic dengue virus infection, CD8^+^ T cells provide some degree of protection in the absence of detectable levels of nAbs ([@B37]). However, in this case a mild (sublethal) DENV-4 infection was used for priming rather than a true vaccine, and despite this priming, only limited (partial) and short-term protection was observed against DENV-2. Moreover, the E proteins of DENV-2 and DENV-4 share high (\u223c64%) sequence homology, including conserved T cell epitopes ([@B38]), and therefore, the contribution of the E protein to protective T cell responses could obviously not be distinguished from what DENV NS proteins may contribute. In contrast, although we also observed some residual cross-reactivity for YFV in serum samples of JE-CVax-vaccinated mice (binding antibody in IIFA \\[[Fig.\u00a0S2](#figS2){ref-type=\"supplementary-material\"}\\]), JE-CVax failed to induce any detectable nAb titers against YFV even following repeated boosting ([Fig.\u00a02A](#fig2){ref-type=\"fig\"}). This finding is in full accordance with the absence of cross-nAbs in mice, monkeys, and humans after JE-CVax or YFV-17D vaccination ([@B39][@B40][@B41]). Therefore, cross-reactivity of serum from JE-CVax-vaccinated mice could be attributed to (i) induction of E-based broad flavivirus cross-reactive non-nAbs resulting from flavivirus infection/vaccination ([@B26], [@B42]) and (ii) reactivity against YFV-NS1 that is expressed within and on the surface of infected cells and is target of specific binding but non-nAbs ([@B20], [@B43]). In fact, we demonstrate equivalent levels of anti-YFV-NS1 antibodies in JE-CVax-vaccinated and YFV-17D-infected mice ([Fig.\u00a02B](#fig2){ref-type=\"fig\"} and [C](#fig2){ref-type=\"fig\"}).\n\nAlthough we show that JE-CVax immunization resulted in complete protection against YFV-induced disease, there were variations in the actual levels of anti-YFV nAbs after YFV challenge. In some animals, no nAbs were detected against YFV after YFV-17D-challenge. This lack of YFV nAb indicates that JE-CVax possibly conferred sterilizing immunity in these mice. Since such protection cannot be explained by nAb against the YFV-prM/E, it may be accredited to non-nAbs ([@B42][@B43][@B44]) and/or adaptive cellular immunity ([@B16], [@B30], [@B32], [@B45]). Some correlation between anti-NS1 antibody levels and the dose of JE-CVax needed to cause protection was observed ([Fig.\u00a02B](#fig2){ref-type=\"fig\"} and [C](#fig2){ref-type=\"fig\"} and [Fig.\u00a0S1A](#figS1){ref-type=\"supplementary-material\"}), and serum from JE-CVax-vaccinated mice was found to induce an ADCC response against YFV-17D ([Fig.\u00a03](#fig3){ref-type=\"fig\"} and [Fig.\u00a0S3](#figS3){ref-type=\"supplementary-material\"}). In addition, in our model both CD4^+^ and CD8^+^ T cells seem (in a likely association with anti-NS1 antibodies) to be involved in JE-CVax-mediated protection against YFV ([Fig.\u00a04E](#fig4){ref-type=\"fig\"}). Previously, only humoral immunity and CD4^+^ T (but not CD8^+^ T) cells have been implied to be sufficient and required for protection against YFV ([@B16]), with a strong emphasis on nAb as a historically established immunological correlate of protection for YFV ([@B5]) and for flaviviruses (such as JEV) in general ([@B1], [@B15], [@B39]). nAbs possibly block viral spread, whereas cellular immunity efficiently eliminates intracellular viruses either directly or targeted by non-nAbs toward infected cells in an Fc-dependent manner (for example, via ADCC toward cells expressing YFV NS1 on their surfaces) ([@B18], [@B21], [@B43], [@B44]). Indeed, YF-induced CD8^+^ T cells have been shown to act as a \"backup defense\" system in the absence of nAbs and to participate in viral clearance in particular from the central nervous system (CNS) in mice ([@B23], [@B32]). Moreover, strong CD8^+^ T cell responses are also detected in the response to human vaccines ([@B30]). As we show here, immunization of AG129 and C57BL/6 mice with JE-CVax elicited protective polyfunctional YFV antigen-specific CD4^+^ and CD8^+^ T cell responses (expressing the Th1-type cytokines TNF-\u03b1 and IFN-\u03b3), which is in line with a previous study with a chimeric Japanese encephalitis/dengue virus 2 vaccine ([@B31]). Previously, clinical studies with humans also demonstrated that vaccination with CYD-TDV/Dengvaxia elicits cell-mediated immunity directed against YFV nonstructural proteins ([@B46], [@B47]). Our comparative *in silico* analysis indicated that the YFV-17D backbone is more immunogenic in humans than in mice and provides a hint that CTL responses should be at least as effective in humans as in mice. Likewise, several HLA-binding YFV-17D peptides thus predicted have recently also been confirmed experimentally using human-derived T cells and vaccinated human volunteers ([@B48][@B49][@B51]) ([Fig.\u00a0S10](#figS10){ref-type=\"supplementary-material\"}). Collectively, our data suggest that JE-CVax-mediated vigorous protection against lethal YFV challenge depends on the combined effects of several effector principles, including both the humoral and cellular immune responses, yet definitely other than nAbs.\n\nTo assess the efficacy of JE-CVax, we employed mice as the *in vivo* model and the YFV-17D vaccine strain ([@B16], [@B17], [@B23], [@B44], [@B52]) as the challenge virus. This experimental setup implies some constraints. Generally, mice are not susceptible to human flavivirus infection ([@B53]) and the YFV-17D is not virulent in humans ([@B5]). To overcome these limitations, we made use of two established, complementary and stringent mouse infection models that are accepted surrogates for testing of flavivirus vaccines ([@B11]); (i) immunocompromised mice (AG129 mice \\[[@B28], [@B34], [@B35]\\]) that develop fatal neurotropic infection when challenged peripherally with YFV-17D, in particular when inoculated with a highly lethal (1,000 LD~50~) challenge virus dose ([Fig.\u00a0S7](#figS7){ref-type=\"supplementary-material\"}) ([@B28]), and (ii) immunocompetent wt mice (C57BL/6 \\[[@B16], [@B17], [@B23]\\] that can be challenged i.c. with YFV-17D. YFV-17D was originally developed by adapting a viscerotropic clinical isolate (YFV-Asibi) to replication in mouse brains (\"fixed to mouse brain\") ([@B54]). This vaccine virus can hence be considered a genuine mouse-adapted neurotropic and neurovirulent YFV strain. For this reason, it could also be considered in this study as the challenge virus. Besides experimental convenience (YFV-17D does not require biosafety level 3 \\[BSL3\\] containment), YFV-17D is a widely accepted, i.e., best-characterized and hence most widely used, challenge strain in mouse models. Classical i.c. inoculation of YFV-17D consistently causes fatal disease in mice ([@B52]) that cannot be distinguished from that induced by wt-YFV strains ([@B44], [@B52]). Also, because JE-CVax expresses the prM/E protein of JEV, which belongs to a serocomplex other than YFV, vaccination and subsequent challenge with YFV-17D compares to a certain extent to a heterotypic flavivirus vaccination challenge (as described by others \\[[@B37]\\]), however, with a markedly more pronounced vaccine efficacy. Therefore, in conclusion, similar mechanisms should hold when using wild-type YFV as the challenge virus ([@B44], [@B52]). Obviously, before proceeding to clinical use, this principle should be confirmed in JE-Cvax-vaccinated nonhuman primates, demonstrating protection from subsequent challenge with virulent wt YFV strains.\n\nLive attenuated JE SA14-14-2 and inactivated Vero cell vaccine account for the biggest use by volume, while Imojev is licensed and used in only a few Asian countries. It is evident from the literature that Imojev remains safe and immunogenic in individuals with a known history of JE vaccination ([@B55]). However, further studies are required to address whether Imojev will also induce protective immunity against YFV in those that have already been vaccinated with one of the other JE vaccines. In any case, formal demonstration of vaccine efficacy in humans will be challenging considering that YFV-specific nAbs that are established as a correlate of protection for YFV-17D are absent ([@B39][@B40][@B41]), and hence protection against lethal infection may instead need to be shown in nonhuman primate vaccine challenge models. Previous safety and immunogenicity studies of chimeric live attenuated viruses (JE-CVax and ChimeriVax-DEN2) in humans indicated that preexisting immunity to the parental YFV-17D vaccine virus does not interfere with immunization but rather induces long-lasting cross-neutralizing antibody responses ([@B39], [@B56], [@B57]). Importantly, if our data on the dual protection conferred by c-LAV in mice could be translated to other species, including humans, this would imply that the JE vaccine JE-CVax (and likewise Imojev) may provide dual protection, i.e., against both JEV and YFV. A dual protective effect may be of particular relevance in case YFV may one day---as is suspected ([@B8], [@B9])---cause outbreaks in (sub)tropical Asia. Given the capacity problems with the production of the current YFV vaccine, having another licensed vaccine (i.e., JE-CVax/Imojev) available as an alternative means to protect against YFV may at such time help to contain an outbreak. In addition, those already vaccinated with JE-CVax may be expected to be protected against YFV. A similar principle may apply to other chimeric flavivirus vaccines that consist of a YFV-17D backbone (such as CYD-TDV/Dengvaxia) ([@B12]) and others under development (such as our recently proposed chimeric YFV-17D/ZIKV vaccine candidate) ([@B28]). Likewise, c-LAVs could be generated against DENV and other viruses that may cause ADE using the backbone of the parent virus (e.g., DENV) from which the prM/E genes have been replaced by that of antigenically more distantly related viruses or serotypes, in line with experimental evidence that the DENV NS1 protein can also serve as a protective antigen ([@B81]). Such an approach may avoid potentially harmful nAb responses. The same principle may apply to c-LAV for ZIKV using a ZIKV backbone ([@B58]) and prM/E sequences of another flavivirus that has not been shown to cause ADE.\n\nTo conclude, we demonstrate that JE-CVax efficiently and rapidly induces high cross-protective efficacy (\u223c100%) in mice against YFV challenge, even with an exceedingly aggressive challenge inoculum. The study provides evidence that c-LAV flavivirus vaccines may be developed solely based on NS proteins. Moreover, immunization with a chimeric flavivirus, whereby the prM/E genes of the backbone have been replaced by that of yet another flavivirus, may have a dual protective effect. A vaccine such as Imojev/JE-CVax may thus be suitable for off-label use, namely, for protection against YFV, which in this case is not mediated by nAbs.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nCells and medium. {#s4.1}\n-----------------\n\nBHK-21J and Vero E6 cells used in this study were a generous gift from Peter Bredenbeek, Leiden University Medical Center (LUMC), the Netherlands. Cells were maintained in seeding medium containing MEM Rega-3 medium (Gibco, Belgium) supplemented with 10% fetal calf serum (FCS; Gibco), 2\u2009mM glutamine (Gibco), and 0.75% sodium bicarbonate (Gibco). HEK 293 cells (human embryonic kidney 293 cells; ATCC CRL-1573) were cultured in Dulbecco modified Eagle medium (DMEM; Gibco) containing 10% FCS and 100 U/ml of penicillin-streptomycin solution (pen-strep; Gibco). Virus culture and cytopathic effect-based virus neutralization testing (CPENT) were performed in assay medium, which was the seeding medium supplemented with only 2% FCS. HEK 293 cells were transfected with YF-NS1-GFP using TransIT-LT1 transfection reagent (Mirus Bio LLC, Belgium), according to the manufacturer's instructions. Infection of HEK 293 cells with YFV-17D-mCherry (see below) was performed in DMEM supplemented with 2% FCS and 100 U/ml of pen/strep solution. All cultures were maintained at 37\u2009\u00b0C in an atmosphere of 5% CO~2~ and 95% to 99% humidity.\n\nVirus, vaccines, and antigens. {#s4.2}\n------------------------------\n\nStamaril (G5400) and Ixiaro (JEV16F290) were from Sanofi Pasteur (France) and Valneva (Austria), respectively. Stamaril was passaged two times in Vero E6 cells (YFV-17D-G5400P2) and stored at \u221280\u00b0C. YFV-17D-G5400P2 was used throughout the study to challenge mice and is referred to as YFV-17D. YFV-17D-based Japanese encephalitis c-LAV Imojev (Chimerivax-JE, JE-CVax) is not available in Europe and was, therefore, retroengineered according to publicly available sequence information ([@B59]) (patent number WO2006044857A2). To that end, a DNA fragment encoding the prM and E proteins of JEV vaccine strain SA14-14-2 was custom synthetized (IDT Integrated DNA Technologies, Haasrode, Belgium) and subcloned into the YFV-17D expression vector pShuttle-YFV-17D ([@B59]) (patent number WO2014174078 A1) of our plasmid-launched live attenuated vaccine (PLLAV)-YFV-17D platform using standard molecular biology techniques and thereby replacing the YFV-17D prM/E sequences. Two adaptive mutations in NS2A and NS4B genes and an additional Kas1 site at the end of the prM/E coding sequence ([@B59]) were introduced by site-directed mutagenesis. To generate JE-CVax virus, BHK-21J cells were transfected with PLLAV-JE-CVax using TransIT-LT1 transfection reagent, following the manufacturer's instructions. Upon onset of cytopathic effect (CPE), JE-CVax virus was harvested, centrifuged at 4,000\u2009rpm at 4\u00b0C for 10 min, aliquoted, and stored at \u221280\u00b0C. Similarly, the live attenuated Japanese encephalitis virus vaccine JEV SA14-14-2 (GenBank accession no. [AF315119.1](https://www.ncbi.nlm.nih.gov/nuccore/AF315119.1)) was generated fully synthetically from overlapping DNA fragments (IDT Integrated DNA Technologies, Haasrode, Belgium), assembled by overlap extension PCR, and subsequent homologous recombination in yeast. The recombinant JE-CVax and JEV SA14-14-2 viruses were subsequently passaged on Vero E6 cells to generate virus stocks. As an alternative challenge virus, ZIKV strain MR766 was used ([@B60]). Virus titers were determined on BHK-21J cells by plaque assays (PFU per milliliter) and CPE-based assays (50% tissue culture infective doses \\[TCID~50~\\] per milliliter) as described below.\n\nA YFV-17D reporter virus (YFV-17D-mCherry) was generated that expresses the red fluorescent protein mCherry as a translational fusion to the N terminus of YFV-17D C protein. In brief, using standard PCR techniques and homologous recombination in *Saccharomyces cerevisiae* (strain YPH500), a synthetic DNA fragment encoding codons 2 to 236 of mCherry (GenBank accession no. [AY678264.1](https://www.ncbi.nlm.nih.gov/nuccore/AY678264.1)) was inserted in YFV-17D genome ([@B61]) immediately downstream of nucleotide (nt) position 181, flanked (i) at its 5\u2032 terminus by a glycine-serine linker (BamHI site) and (ii) at the 3\u2032 end by a Thosea asigna virus 2A peptide (EGRGSLLTCGDVEENPG/P) ([@B62]) followed by a repeat of codons 2 to 21 of the YFV-17D C gene, yet with an alternative codon usage to avoid RNA recombination during viral replication. YFV-17D-mCherry was rescued by transfection of the resulting PLLAV-YFV-17D-mCherry of BHK 21J as described previously ([@B28]). A full characterization of YFV-17D-mCherry is available upon request (M. A. Schmid, N. Mishra, S. Sharma, R. Boudewijns, J. Neyts, and K. Dallmeier, unpublished data).\n\nPlasmid pCMV-YFV-17D NS1-IRES-EGFP, which drives the mammalian expression of YFV-17D NS1 as a transcriptional fusion to enhanced green fluorescent protein (EGFP) was generated by PCR cloning of YFV-17D nt 2381 to 3508 cDNA (including an E protein-derived N-terminal signal peptide) ([@B63]) flanked by a 5\u2032-terminal Kozak sequence and 3\u2032-terminal stop codon into the NheI and SalI sites of pIRES2-EGFP (Clontech catalog no. 6029-1).\n\nAn MHC class I-restricted peptide from YFV-17D nonstructural protein 3 (NS3) (sequence ATLTYRML) ([@B64]) was synthetized by Eurogentec (Seraing, Belgium). Total cellular antigen for YFV-17D and JEV SA14-14-2 was prepared first by infecting Vero E6 cells with YFV-17D and JEV SA14-14-2, respectively, at a multiplicity of infection (MOI) of 0.1. Noninfected Vero E6 cells were used as a control. Four days postinfection, cells were harvested either by trypsinization or by detaching through pipetting when CPE was visible. Following centrifugation, cell pellets were resuspended in phosphate-buffered saline (PBS) and cell lysates were prepared by four freeze-thaw cycles. Overnight UV irradiation was performed to inactivate the virus in the cell lysate preparations, and large debris was removed via filtering through 70-\u03bcm cell strainers (BD Biosciences).\n\nAnimals, hyperimmune serum, infection, and T cell depletions. {#s4.3}\n-------------------------------------------------------------\n\nHost IFN signaling is the major barrier to viscerotropism and for pathogenicity of neurotropic flaviviruses ([@B65]). In line, wild-type mice are poorly susceptible to infection with flaviviruses ([@B28], [@B66][@B67][@B68]), including YFV-17D infection and vaccination ([@B34], [@B35], [@B65]). Likewise, type I (IFN-\u03b1/\u03b2) and type II (IFN-\u03b3) interferon signaling mutually controls YFV-17D infection. Unlike in humans ([@B69], [@B70]), type I IFN (IFN-\u03b1/\u03b2) can restrict YFV-Asibi as well as YFV-17D infection in mice ([@B34], [@B35], [@B71], [@B72]). Similarly, IFN-\u03b3 exerts restriction on YFV-17D replication, dissemination, and clearance in mice ([@B35], [@B73]). YFV-17D is neurotropic in wt mice when directly injected into the brain ([@B52]). In AG129 mice, IFN-\u03b1/\u03b2 and -\u03b3 receptors are knocked out, which results in neurotropic infection following peripheral inoculation of YFV-17D. Therefore, to allow transient replication (and thus vaccination) of YFV-17D in wt mice, MAR1-5A3 antibodies were coadministered to block type I IFN signaling in C57BL/6 mice.\n\nImmunodeficient IFN-\u03b1/\u03b2 and -\u03b3 receptor knockout mice (AG129; B&K Universal, Marshall BioResources, UK) were bred in-house. AG129 mice have been shown to be highly susceptible to lethal YFV-17D infection, serving as a well-established surrogate rodent challenge model for wt YFV infection ([@B34], [@B35], [@B74]). Six- to 8-week-old male AG129 mice were used for all experiments after random assignment into different groups. Animals were kept in individually ventilated type-2 filter top cages on a 12-h/12-h day/night cycle with water and food *ad libitum*. Housing of animals and procedures involving animal experimentation were conducted in accordance with institutional guidelines approved by the Ethical Committee of the KU Leuven, Belgium (licenses P168/2012, P103/2015, P140/2016, and P005/2018). Throughout the study, animals were vaccinated intraperitoneally (i.p.) with either 10^4^ PFU of JE-CVax, 1/6th human dose of Ixiaro ([@B25]), or 2% assay medium. Animals vaccinated with Ixiaro were boosted on 14 dpv with 1/6th human dose of Ixiaro. All the vaccinated animals were challenged i.p. with 10^3^ PFU of either YFV-17D or JEV SA14-14-2 (both corresponding to 1,000 LD~50~) 28 dpv, if not stated otherwise. An additional 4 weeks after YFV-17D challenge, some animals were rechallenged i.p. with 10^4^ PFU of ZIKV-MR776 ([@B60]). Animals were observed for morbidity (weight loss) and humane endpoint once daily. The humane endpoint is defined as either paresis/difficulty in walking, paralysis (hind legs/soured eyes), moribundity/ataxia/tremors/difficulty in breathing, 20% weight loss, or quick weight loss (15% within 1 or 2\u2009days); animals were immediately euthanized once a humane endpoint was reached. Throughout the study, bleedings were performed through submandibular puncture on day 0 (prevaccination), day 28 (postvaccination), and day 56 (postchallenge).\n\nHyperimmune serum was prepared by vaccinating AG129 mice with 10^4^ PFU of JE-CVax, followed by two boosts with 10^5^ PFU of JE-CVax at 14-day intervals. Another 14\u2009days after the second booster, animals were bled twice per week for the following 4 weeks. All serum batches were then pooled and CPENTs for JE-CVax and YFV-17D were performed. We did not observe any YFV nAbs in the hyperimmune sera but did see an \u223c3.6-fold (log~10~ CPENT~50~ titer, 3.03 \u00b1 0.18) selective increase in neutralizing titers against JE-CVax compared to single vaccination ([Fig.\u00a03A](#fig3){ref-type=\"fig\"}). Normal mouse control serum was prepared by pooling serum from 18 nonvaccinated AG129 mice.\n\nImmunocompetent wt C57BL/6JOIaHsd, i.e., C57BL/6, mice were purchased from ENVIGO Labs, the Netherlands, and were maintained and manipulated as described for AG129 mice, with some modifications ([@B23], [@B28]). Since flaviviruses do not readily replicate in immunocompetent wild-type mice ([@B28]), the mice were immunized with 10^4^ PFU of JE-CVax in the presence of 2.5\u2009mg of an IFN-\u03b1/\u03b2 receptor subunit 1 (IFNAR-1) binding monoclonal antibody, MAR1-5A3, administered i.p. 1 day prior to immunization. MAR1-5A3 antibody (0.5\u2009mg) was also readministered i.p. on day 4 and day 7 postvaccination. Animals were bled 28\u2009days postvaccination and challenged i.c. with 10^4^ PFU of YFV-17D in a volume of 30\u2009\u03bcl. A full characterization of immunogenicity of YFV-17D in various mouse strains is available upon request (J. Ma, N. Mishra, R. Boudewijns, J. Neyts, and K. Dallmeier, unpublished data). For T cell depletion studies, C57BL/6 mice were either sham vaccinated or vaccinated i.p. with 1\u2009\u00d7\u200910^4^ PFU of JE-CVax 35\u2009days prior to i.c. challenge with 1\u2009\u00d7\u200910^4^ PFU of YFV-17D. At day \u22122 and day 0 prior to YFV challenge, 0.5\u2009mg of either anti-mouse CD4 (clone GK1.5; Leinco Technologies, USA) or anti-mouse CD8a (clone 53-6.7; Leinco Technologies) or a combination of both was administered i.p. ([@B32], [@B75]).\n\nIIFA. {#s4.4}\n-----\n\nTo determine the seroconversion of animals, all JEV, YFV, and ZIKV IgG indirect immunofluorescence assays (IIFAs) were performed as per the manufacturer's instructions (Euroimmun, L\u00fcbeck, Germany), except for the use of labeled secondary antibody and the mounting agent glycerine, which were replaced by Alexa Fluor 488 goat anti-mouse IgG (A-11029; Thermo Fisher Scientific) and 4\u2032,6-diamidino-2-phenylindole (DAPI; ProLong antifade reagent with DAPI; Thermo Fisher Scientific), respectively. Serum from nonvaccinated animals served as a naive, negative control. Slides were visualized using a fluorescence microscope (FLoid cell imaging station; Thermo Fisher Scientific).\n\nPlaque assay and plaque reduction neutralization testing (PRNT). {#s4.5}\n----------------------------------------------------------------\n\nViral titers of YFV-17D or JE-CVax preparations were determined using plaque assays on BHK-21J cells. In brief, 10^6^ BHK-21J cells per well were plated in 6-well plates and cultured overnight in seeding medium. Cells were washed with PBS and inoculated with virus of different dilutions prepared in the assay medium for 1 h at room temperature (RT). Culture supernatants of uninfected cells were used as negative controls. Cells were thoroughly washed with the assay medium and overlaid with 2\u00d7 minimal essential medium (MEM; Gibco, Belgium) supplemented with 4% FCS and 0.75% sodium bicarbonate containing 0.5% low-melting-point agarose (Invitrogen, USA). The overlay was allowed to solidify at RT; cells were then cultured for 7\u2009days at 37\u00b0C, fixed with 8% formaldehyde, and stained with methylene blue. Plaques were manually counted and plaque titer was determined as PFU per milliliter.\n\nThroughout the study, all the virus neutralization assays, i.e., PRNT and CPE-based virus neutralization testing (CPENT), were performed with YFV-17D and JE-CVax. JE-CVax has previously been established as a safe substitute for JEV, a BSL3 pathogen, when virus neutralization tests need to be performed at BSL2 ([@B76]). PRNT was performed similarly to the plaque assays for viral titration, with some modifications. Briefly, a step was added, in which different serum dilutions made in the assay medium were first inoculated with YFV-17D (20 to 50 PFU) or JE-CVax (50 to 100 PFU) virus for 1 h at 37\u00b0C and then added to the cells. All sera were assayed in triplicate in serial dilutions 1:20, 1:66, 1:200, 1:660, 1:2,000, and 1:6,600. Plaques were manually counted and PRNT~50~ were calculated using the Reed and Muench method ([@B77]). Culture-derived YFV-17D and JE-CVax were used as positive virus controls, while culture supernatants of uninfected cells were used as a negative cell control. PRNT~50~ values for each sample represent geometric means of three independent repeats, and data are presented as log~10~ PRNT~50~ (mean \u00b1 standard deviation \\[SD\\]).\n\nCPE assay and CPENT. {#s4.6}\n--------------------\n\nViral titers for culture-derived YFV-17D or JE-CVax (TCID~50~) and 50% neutralizing antibody titers (log~10~ CPENT~50~) were determined using CPE-based cell assays and CPE-based virus neutralization tests, respectively, on BHK-21J cells as described previously ([@B78]), with some modifications. In brief, 2\u2009\u00d7\u200910^4^ BHK-21J cells/well were plated in 96-well plates overnight in seeding medium. The medium was then replaced with assay media containing different virus dilutions and cultured for 5\u2009days at 37\u00b0C. Later, assays were first visually scored for CPE and then stained with MTS/phenazine methosulphate (PMS; Sigma-Aldrich) solution for 1.5 h at 37\u00b0C in the dark. Post-MTS/PMS staining, absorbance was measured at 498\u2009nm for each well. All assays were performed in six replicates, and TCID~50~ per milliliter was determined using the Reed and Muench method ([@B77]).\n\nCPENTs were performed similarly to the CPE assays for viral titration, with some modifications. Briefly, a step was added, in which different serum dilutions made in the assay medium were first inoculated with 100 TCID~50~ of YFV-17D or JE-CVax virus for 1 h at 37\u00b0C and then added to the cells. All sera were assayed in triplicate in serial dilutions 1:20, 1:66, 1:200, 1:660, 1:2,000, and 1:6,600. CPE neutralization was calculated with the following formula: percent neutralization activity = percent CPE reduction = (OD~virus+serum~ -- OD~VC~) \u00d7 100/(OD~CC~ -- OD~VC~), where OD is optical density; 50% neutralization titers (CPENT~50~) were calculated using the Reed and Muench method ([@B77]). Culture-derived YFV-17D and JE-CVax were used as positive virus controls (VC), while culture supernatants of uninfected cells were used as a negative cell control (CC). CPENT~50~ values for each sample represent geometric means of three independent repeats, and data are presented as log~10~ CPENT~50~ (mean \u00b1 SD). CPENT for detection of nAbs was validated against a standard PRNT, yielding a strong correlation (*R*^2^ = 0.71; *P*\u2009=\u20090.018) between PRNT~50~ and CPENT~50~ ([Fig.\u00a0S8A](#figS8){ref-type=\"supplementary-material\"}) and similar anti-JEV nAb titers in postvaccination and postchallenge serum samples ([Fig.\u00a0S8B](#figS8){ref-type=\"supplementary-material\"} and [C](#figS8){ref-type=\"supplementary-material\"}).\n\n10.1128/mBio.02494-19.8\n\nCorrelation of nAb titers determined as log~10~ PRNT~50~ and log~10~ CPENT~50~ in matched serum samples of JE-CVax-vaccinated and/or YFV-17D-challenged mice. AG129 mice (*n*\u2009=\u20097) were vaccinated with 10^4^ PFU of JE-CVax and at 28 dpv challenged with 10^3^ PFU YFV-17D. Serum was harvested and neutralization assays, i.e., CPENT and PRNT, were performed as described in Materials and Methods. Data show a good correlation (Pearson correlation \\[*R*^2^\\] = 0.071; *P*\u2009=\u20090.02) between log~10~ YFV PRNT~50~ and log~10~ YFV CPENT~50~ of matched samples (A). There was no marked increase in the log~10~ JE-CVax PRNT~50~ and log~10~ JE-CVax CPENT~50~ (*P* values, 0.156 and 0.062, respectively) found when comparing matched serum samples from before and after challenge with YFV-17D using Wilcoxon matched-pairs signed rank test (B and C). The limit of detection for each assay was log~10~20, i.e., 1.3. Download FIG\u00a0S8, JPG file, 0.2 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nELISA. {#s4.7}\n------\n\nSerum antibodies recognizing YFV NS1 were detected by indirect enzyme-linked immunosorbent assay (ELISA), in essence as previously described ([@B75], [@B76], [@B80]). In brief, ELISA plates (Nunc MaxiSorp; Thermo Fisher Scientific) were coated with 1\u2009\u03bcg/ml of recombinant YFV NS1 (Bio-Rad; catalog no. PIP052A) in 50\u2009mM carbonate buffer (pH 9.6) overnight at 4\u00b0C. After three washes with PBS-T (PBS with 0.05% Tween 80), plates were blocked with 2% bovine serum albumin (BSA) in PBS-T for 1 h at 37\u00b0C or overnight at 4\u00b0C. After three washes with PBS-T, wells were treated with serial dilutions of test sera (2-fold serial dilution in PBS-T) for 2 h at room temperature. Serial dilutions of the YFV NS1-specific mouse IgG2a monoclonal antibody (clone 1A5, kindly provided by J. J. Schlesinger) ([@B27]) starting at 10\u2009\u03bcg/ml served as a standard. After four washes with PBS, plates were incubated with horseradish peroxidase-labeled goat anti-mouse IgG antibody (Sigma-Aldrich; catalog no. AP124P; diluted 1:3,000 in PBS-T) for 1 h. After another four washes with PBS, bound antibodies were detected via conversion of added tetramethylbenzidine (TMB; SureBlue TMB microwell peroxidase; KPL). The reaction was stopped after 10 min by adding equal quantities of 1 M HCl solution, and absorbance was measured at 450\u2009nm. After background subtraction, relative anti-YFV NS1 titers were determined by comparison to the standard curve generated for MAb 1A5 included in each assay plate. To that end, the dilution at which each individual test serum yielded an OD at 450 nm (OD~450~) of 1 was used to calculate an absolute anti-NS1 antibody concentration (equivalent concentration), assuming a binding similar to that of YFV-17D NS1 by MAb 1A5. Only values that exceeded three times the background signal were considered positive.\n\nADCC bioassay. {#s4.8}\n--------------\n\nTo assess the possible role of nonneutralizing YFV antibody-mediated protection against YFV after JE-CVax vaccination, antibody-dependent cell-mediated cytotoxicity (ADCC) bioassays ([@B29]) were performed as prescribed by the manufacturer (ADCC reporter bioassay complete kit; Promega; catalog no. G7010). In brief, target cells were prepared by infecting HEK 293T cells with YFV-17D-mCherry in assay medium. Cells were incubated at 37\u00b0C postinfection and later, upon onset of CPE, harvested by trypsinization. Cells were again plated in white, flat-bottom 96-well assay plates (Viewplate-96; PerkinElmer; catalog no. 6005181) at densities of 7,500 and 25,000 cells per well for 8 h at 37\u00b0C in assay medium. Later, in a separate 96-well plate, JE-CVax hyperimmune and nonimmune heat-inactivated mouse serum samples (starting dilution of 1/9) were serially diluted 3-fold in RPMI 1640 medium (Gibco, Thermo Fisher Scientific) supplemented with 4% low-IgG serum. ADCC bioassay effector cells (Jurkat V variant cells) were diluted to 3\u2009\u00d7\u200910^6^ cells/ml in RPMI 1640 medium. The supernatant from the infected cell plate was replaced with fresh RPMI medium (25\u2009\u03bcl/well), and diluted serum samples (25\u2009\u03bcl/well) and E cells (25\u2009\u03bcl/well) were added to the infection plate. After incubation at 37\u00b0C for 24 h, Bio-Glo luciferase assay reagent (75\u2009\u03bcl/well) was added, and luminescence was measured using a Spark multimode microplate reader (Tecan). The average background plus three SDs was calculated and used as background.\n\nIntracellular staining of NS1 protein in HEK 293 cells. {#s4.9}\n-------------------------------------------------------\n\nHEK 293 cells transfected with pCMV-YFV-17D NS1-IRES-EGFP or infected with YFV-17D-mCherry were detached with trypsin-EDTA (0,05%), centrifuged (at 2,500\u2009rpm and 4\u00b0C for 5 min), and suspended in FACS-B (Dulbecco\\'s phosphate-buffered saline \\[DPBS\\], no Ca^2+^/Mg^2+^, 2% fetal bovine serum \\[FBS\\], 2\u2009mM EDTA). Not more than 5\u2009\u00d7\u200910^6^ cells per well were seeded into round-bottom 96-well plates (Costar, Corning Inc.) and spun down, and the supernatant was removed. Dead cells were stained *in vitro* with ZombieAqua (Biolegend; 1:500 diluted in DPBS) to exclude from further analysis. After washing and fixation with 2% paraformaldehyde (in FACS-B), cells were permeabilized by 0.1% saponin (in FACS-B with 1% normal mouse serum) with streptavidin added (streptavidin/biotin blocking kit; Vector Laboratories) to block endogenous biotin (permeabilizing and blocking solution). The cells were then stained with anti-NS1 primary antibody solution (clone 1A5; 5\u2009\u03bcg/ml), or JE-CVax-vaccinated mouse serum (1:10) in permeabilizing and blocking solution. Anti-NS1 antibody binding was detected by a biotinylated goat anti-mouse IgG secondary antibody solution (Thermo Fisher Scientific; catalog no. A16076; 1:200 dilution). The biotinylated secondary antibody was stained subsequently with streptavidin-phycoerythrin (PE)-Cy7 (Biolegend; 1:200 dilution). After the cells were washed, they were resuspended in FACS-B and filtered through 100-\u03bcm nylon meshes (Sefar; ELKO Filtering; 03-100/44) prior to analysis on a flow cytometer (LSR Fortessa X-20; Becton, Dickinson). The data were analyzed using FlowJo 10 software (TreeStar). The gating strategy for the analysis is depicted in [Fig.\u00a0S9A](#figS9){ref-type=\"supplementary-material\"}.\n\n10.1128/mBio.02494-19.9\n\nGating strategy for flow cytometry analysis. (A and B) Exclusion of debris was achieved by gating out the FSC-low population in a plot of FSC-A versus SSC-A. Then, only single cells were retained by elimination of the high-SSC-W population in a plot of SSC-W versus SSC-H. In a subsequent step, live cells were selected by gating out the Zombie Aqua-positive population as shown in a plot of NS1-eGFP/FSC-A versus ZA. Finally, for the detection of anti-NS1 antibodies (A), cells were gated based on positivity for NS1-eGFP and positivity/negativity of anti-NS1 Ab PE-Cy7. (B) For intracellular cytokine staining, based on positivity for CD3e (eFluor450 conjugated) and negativity (CD4) or positivity for CD8a (APC/Cy7-conjugated), CD4^+^ and CD8^+^ T cell populations were defined as CD3^+^ CD8^\u2212^ and CD3^+^ CD8^+^ populations, respectively. Finally, cells were gated based on positivity/negativity for IFN-\u03b3 and positivity/negativity of TNF-\u03b1. Samples from nonvaccinated mice were used to set the boundaries that define cells positive and negative for intracellular markers. Download FIG\u00a0S9, JPG file, 0.5 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\n10.1128/mBio.02494-19.10\n\nBinding predictions for peptides (8- to 14-mers) comprised in the YF-17D nonstructural proteins for mouse, human, and rhesus macaque MHC classI. For the 20 strongest binding peptides (lowest dissociation constant \\[*K~d~*\\]) for each haplotype, the position of the first amino acid in the NS proteins is given, as well as the affinity of the peptide for the MHCI. HLA-binding peptides indicated in colors are (or at least contain) YFV-17D-derived peptides that have been experimentally confirmed by Lund (O. Lund, PloS One 6:e26494, 2011, <https://doi.org/10.1371/journal.pone.0026494>) (blue), de Melo et al. 2013 (A. B. de Melo, E. J. M. Nascimento, U. Braga-Neto, R. Dhalia, A. M. Silva, M. Oelke, J. P. Schneck, J. Sidney, A. Sette, S. M. L. Montenegro, and E. T. A. Marques, PLoS Negl Trop Dis 7:e1938, 2013, <https://doi.org/10.1371/journal.pntd.0001938>) (yellow), Blom et al. (K. Blom, M. Braun, M. A. Ivarsson, V. D. Gonzalez, K. Falconer, M. Moll, H.-G. Ljunggren, J. Micha\u00eblsson, and J. K. Sandberg, J Immunol 190:2150--2158, 2013, <https://doi.org/10.4049/jimmunol.1202234>) (red), and Kongsgaard et al. (M. Kongsgaard, M. R. Bassi, M. Rasmussen, K. Skj\u00f8dt, S. Thybo, M. Gabriel, M. B. Hansen, J. P. Christensen, A. R. Thomsen, S. Buus, and A. Stryhn, Sci Rep 7:1--14, 2017, <https://doi.org/10.1038/s41598-017-00798-1>) (green). Download FIG\u00a0S10, JPG file, 0.2 MB.\n\nCopyright \u00a9 2020 Mishra et al.\n\n2020\n\nMishra et al.\n\nThis content is distributed under the terms of the\n\nCreative Commons Attribution 4.0 International license\n\n.\n\nProcessing of mouse spleens for the preparation of single-cell suspensions. {#s4.10}\n---------------------------------------------------------------------------\n\nSix- to 8-week-old C57BL/6 or AG129 mice were vaccinated with 10^4^ PFU of JE-CVax and 4 and 18 weeks later, the animals were euthanized for analysis. Spleens were harvested and processed for ELISpot and flow cytometric analyses. To generate single-cell suspensions, spleens were pushed through 70-\u03bcm cell strainers (BD Biosciences) with syringe plungers, digested in 1.0\u2009mg/ml of type 1 collagenase and 10 U/ml DNase for 30 min, vigorously pipetted, and filtered through 100-\u03bcm nylon meshes. Spleen samples were then incubated with red blood cell lysis buffer (eBioscience) for 8 min at room temperature and washed twice with FACS-B.\n\nELISpot assay. {#s4.11}\n--------------\n\nMouse TNF-\u03b1 and IFN-\u03b3 enzyme-linked immunospot (ELISpot) assays were performed with a mouse TNF-\u03b1 ELISpot kit (ImmunoSpot MTNFA-1M/5; CTL Europe GmbH) or a mouse IFN-\u03b3 ELISpot kit (ImmunoSpot MIFNG-1M/5; CTL Europe GmbH) according to the manufacturer's instructions. Assay plates (96-well polyvinylidene difluoride \\[PVDF\\] membrane), antibodies, enzymes, substrate, and diluent were included in the kits. Briefly, 4\u2009\u00d7\u200910^5^ mouse splenocytes/well were plated with either with 5\u2009\u03bcg/ml of YFV-17D NS3 ATLTYRML peptide antigen ([@B60]) or with 50\u2009\u03bcg/ml of total Vero E6 cellular antigen in RPMI 1640 medium (Gibco, Belgium) supplemented with 10% fetal bovine serum, 2\u2009mM [l]{.smallcaps}-glutamine, and 0.75% sodium bicarbonate. After 24 h of incubation at 37\u00b0C, spots of mouse TNF-\u03b1 or IFN-\u03b3 were visualized by subsequent addition of detection antibody, enzyme, and substrate. All plates were scanned using an ImmunoSpot S6 universal reader (CTL Europe GmbH). Spot counts were normalized by subtracting the number of spots from corresponding samples stimulated with noninfected Vero E6.\n\nIntracellular cytokine staining for memory T cells and flow cytometry. {#s4.12}\n----------------------------------------------------------------------\n\nTo restimulate memory T cells, freshly isolated single cell suspensions of splenocytes were seeded at 3\u2009\u00d7\u200910^6^ per well in a round-bottom 96-well plate and incubated with either 5\u2009\u03bcg/ml of the MHC class I- restricted peptide from YFV-17D NS3 (ATLTYRML) ([@B60]) or 50\u2009\u03bcg/ml of total cellular antigen (infected or noninfected Vero E6 cell lysate). Following overnight incubation, the splenocytes were incubated for 2 h with 5\u2009\u03bcg/ml of brefeldin A (Biolegend) for intracellular trapping of cytokines and then stained with Zombie Aqua (1:200) in PBS for 15\u2009min. Splenocytes were then stained for CD3 (4\u2009\u03bcg/ml eFluor 450 \u03b1-mouse CD3 antibody; Thermo Fisher Scientific) and CD8 (2\u2009\u03bcg/ml of allophycocyanin \\[APC\\]/Cy7 \u03b1-mouse CD8a antibody; Biolegend) in PBS for 20\u2009min before fixation in 2% paraformaldehyde (Sigma-Aldrich) and permeabilization and blocking in a mixture of 0.1% saponin and 1% normal mouse serum. Finally, splenocytes were stained intracellularly for TNF-\u03b1 (6.5\u2009\u03bcg/ml PE anti-mouse TNF-\u03b1; Biolegend) and IFN-\u03b3 (2\u2009\u03bcg/ml of APC anti-mouse IFN-\u03b3; Biolegend) prior to analysis on a flow cytometer (LSR Fortessa X-20; Becton, Dickinson). Gating on forward scatter A (FSC-A)/side scatter A (SSC-A) excluded debris; SSC-H/SSC-W and FSC-H/FSC-W excluded doublet cells. The data were analyzed using FlowJo 10 software (TreeStar). To determine the percentage of responding CD4^+^ or CD8^+^ T lymphocytes, the percentages of responders from samples stimulated with noninfected Vero E6 cell lysates were subtracted from the corresponding responses. The gating strategy for the analysis is depicted in [Fig.\u00a0S9B](#figS9){ref-type=\"supplementary-material\"}.\n\nBinding prediction for cytotoxic T lymphocyte (CTL) epitopes within the YFV-17D nonstructural proteins. {#s4.13}\n-------------------------------------------------------------------------------------------------------\n\nBinding to the MHC class I molecules of peptides comprised in the NS proteins of YFV-17D was predicted by NetMHCPan4.0 (<http://www.cbs.dtu.dk/services/NetMHCpan/>). The full amino acid sequence of YFV-17D NS proteins was entered into the server to predict the 20 peptides (8- to 14-mers) with the highest affinity to their respective MHC/HLA class I molecules, such as mouse H2-Kb (C57BL/6 and S129 background), rhesus macaque Mamu-A\\*01, and some commonly found HLA types (HLA-A\\*02, HLA-A\\*24, and HLA-B\\*15) in men ([@B79]).\n\nStatistical analysis. {#s4.14}\n---------------------\n\nGraphPad Prism 7 (GraphPad Software, Inc.) was used for all statistical evaluations. Quantitative data were represented as means \u00b1 standard deviations and obtained from at least three independent experiments. For ADCC assays, flow cytometry analysis and ELISpot assay data were represented as means \u00b1 standard errors of means (SEMs). Statistical significance was determined using survival analysis with log rank (Mantel-Cox) test, one-way analysis of variance (ANOVA) (neutralization titers and ELISA), two-way ANOVA (ADCC), paired *t* test (flow cytometry), and Wilcoxon matched-pairs signed rank test (comparison of paired postvaccinated and postchallenge samples). Correlation studies were performed using linear regression analysis with Pearson's correlation coefficient. Values were considered statistically significantly different at *P* values\u2009of \u22640.05.\n\n**Citation** Mishra N, Boudewijns R, Schmid MA, Marques RE, Sharma S, Neyts J, Dallmeier K. 2020. A chimeric Japanese encephalitis vaccine protects against lethal yellow fever virus infection without inducing neutralizing antibodies. mBio 11:e02494-19. <https://doi.org/10.1128/mBio.02494-19>.\n\nWe thank Katrien Geerts and Sarah Debaveye for their excellent technical assistance, Jef Ceulemans and Jonas Piot for their contribution to assay development, and Robert Vrancken and Joeri Auwerx for assisting in the generation of recombinant JEV SA14-14-2.\n\nThe project received funding from the European Union's Horizon 2020 research and innovation program under RABYD-VAX grant agreement no. 733176. This work was further supported by KU Leuven IOF Hefboom (IOF HB/13/010) and KU Leuven C3 (C32/16/039) grants. R.E.M. was granted a fellowship by the CNPq/FWO (No 52/2012). M.A.S. was granted a senior postdoctoral fellowship by the KU Leuven Rega Foundation.\n\n[^1]: Present address: Niraj Mishra, Gene Therapy Division, Intas Pharmaceuticals Ltd., Biopharma Plant, Ahmedabad, Gujarat, India; Michael Alexander Schmid, Humabs BioMed-Vir Biotechnology, Bellinzona, Switzerland; Rafael Elias Marques, Brazilian Biosciences National Laboratory (LNBio), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, Sao Paulo, Brazil.\n"}
+{"text": "Background {#Sec1}\n==========\n\nCystinuria is one of the most common autosomal recessive genetic disorders with an incidence of 1 in 7000 worldwide \\[[@CR1]\\]. Patients suffer from significant morbidity and few new options have been developed in the past 20\u2009years \\[[@CR2]\\]. Cystinuria is caused by an inherited defect in the transport of cystine and dibasic amino acids (ornithine, lysine, and arginine) in renal tubular cells. Cystine is not soluble in urine so kidney and bladder stones form when the renal tubule fails to reabsorb the amino acid back into the bloodstream. Recurrent cystine kidney stones are associated with pain, frequent urinary tract infections, bleeding, urinary tract obstruction, need for multiple surgical procedures, and kidney failure \\[[@CR2]\\]. Medical treatment to prevent the formation of cystine stones is not very effective and has many unpleasant side effects \\[[@CR3]\\]. Mutations in the gene encoding a renal amino acid transporter (*SLC3A1*, or the rBAT protein) have been identified in most patients with cystinuria establishing the most common cause of the disease termed cystinuria type I \\[[@CR4]\\]. A second gene *SLC7A9,* encoding for b^0,+^type amino acid transporter 1 (b^0,+^AT) accounts for a smaller fraction of cases \\[[@CR5], [@CR6]\\]. The rBAT protein (*Slc3a1* gene product) heterodimerizes with the b^0,+^AT protein (*Slc7a9* gene product) within proximal tubular cells in the nephron, thereby mediating resorption of cystine and dibasic amino acids from the urine \\[[@CR7]\\]. Knowledge of the genetic basis of cystinuria provides opportunities for developing treatments.\n\nMouse models of cystinuria type I have been generated. Peters et al. identified a missense mutation in *Slc3a1* in a N-ethyl-N-nitrosourea (ENU) mutagenesis screen in C3HeB/FeJ mice \\[[@CR8]\\]. The authors demonstrated reduced kidney weight in homozygous males as compared to wild type males and homozygous females. Additionally, plasma urea levels were elevated in homozygous males from 13 to 20\u2009weeks old as compared to wild type animals. Livrozet et al. identified a spontaneous mutation in *Slc3a1* in 129SvPasCrl mice resulting in cystinuria \\[[@CR9]\\]. Renal function was not significantly impaired in the mutant mice as measured by serum creatinine measurements. However, they did observe increased macrophages and interstitial fibrosis in the mutant mice. Ercolani et al. demonstrated bladder outlet obstruction in male cystinuria mice on a mixed C57Bl/6 and 129/SvJ backround \\[[@CR10]\\]. The genetic strategy for generation of this cystinuria mouse line has been reported \\[[@CR11]\\]; however, we sought to more fully characterize this cystinuria mouse strain.\n\nRecently, it has been debated whether or not kidney stone formation is a pathogenic process contributing to kidney disease \\[[@CR12]\\]. Blockages produced by crystals of various composition have been reported to lead to chronic kidney disease, acute kidney injury, renal colic, or nephrocalcinosis \\[[@CR13]\\]. Crystal deposits may be found in the nephron, renal vasculature, and/or urinary tract \\[[@CR13]\\]. Beyond direct blockage, additional molecular mechanisms have been explored including involvement of the inflammasome and crystal granuloma formation \\[[@CR14], [@CR15]\\]. It has been observed that cystinuric patients develop chronic kidney disease (CKD) even more commonly than usual stone formers \\[[@CR16]--[@CR18]\\]. Hypertension has been shown to be associated with CKD in cystinuria patients \\[[@CR19]\\]. Cystinuria has been associated with lower renal function and quality of life compared to other kidney stone formers \\[[@CR20]\\]. Nonetheless, little is known about what contributes to the development of CKD in patients with cystinuria other than stone formation. We sought to more fully characterize the cystinuria mouse strain previously reported by Ercolani et al. by evaluating plasma amino acid levels and to determine if this strain develops CKD over time \\[[@CR10]\\]. Additionally, though one previous report found decreased glutathione levels in leukocytes of patients with cystinuria \\[[@CR21]\\], glutathione levels have not been studied in other tissues or in *Slc3a1*^\u2212/\u2212^ animals.\n\nMethods {#Sec2}\n=======\n\nAnimals {#Sec3}\n-------\n\n*Slc3a1*^\u2212/\u2212^ and wild type mice (male and female) were bred and maintained as described previously and according to the Institutional Animal Care and Use Committee of the Nashville Tennessee Valley Healthcare System VA and Vanderbilt University Medical Center \\[[@CR10], [@CR11]\\]. Normal and breeder chow diets were obtained from Lab Diet (St. Louis, MO) with 5L0D corresponding to normal and 5LJ5 to breeder. Mice were housed at the Nashville Tennessee Valley Healthcare System VA. Measurements of weight and length were performed on mice that were bred and maintained on breeder chow. To measure length, mice were anesthetized with isoflurane and measured from the tip of the nose to the tip of the tail. Animals were randomly assigned to experimental groups. At the conclusion of the study, animals were euthanized in accordance with the American Veterinary Medical Association by CO~2~ gas followed by cervical dislocation.\n\nReal time-PCR {#Sec4}\n-------------\n\nMouse tissues were removed and immersed in RNAlater (Qiagen, Valancia, CA) for mRNA expression studies. RNA was extracted using a Qiagen RNeasy Kit. Primer pairs were designed using Primer Express software (Applied Biosystems, Foster City, CA) and GAPDH expression was used as an endogenous control. The mRNA expression levels of *Slc3a1* and *Slc7a9* were assessed by quantitative real time PCR using SYBR green and an ABI 7900HT Sequence Detection System (Applied Biosystems). Primers are listed in Table\u00a0[1](#Tab1){ref-type=\"table\"}.Table 1Primers for real time PCRPrimer nameSequence 5\u2032 to 3\u2032mGAPDH-FCTCCACTCACGGCAAATTCAAmGAPDH-RGATGACAAGCTTCCCATTCTCG*Slc3a1*-FCCCGGGAACGCCCATCACTT*Slc3a1*-RCCCACTGCATCGGTGACTTGG*Slc7a9*-FTGTGGGTGCCATCAGTCTGGC*Slc7a9*-RTGGCCATGGGCAGGTTTCTGT\n\nWestern blot, immunostaining, and histology {#Sec5}\n-------------------------------------------\n\nWhole cell protein extracts were prepared from \\~\u200930\u2009\u03bcg kidney tissue using a kit for simultaneous isolation of RNA and protein (Macherey-Nagel, Bethlehem, PA) according to the manufacturer's instructions. Proteins (10--15\u2009\u03bcg/lane) were resolved on a 4--12% Bis-Tris gel, transferred to nitrocellulose and analyzed by immunoblot using rabbit anti-SLC3A1 (Proteintech, Rosemont, IL; 16,343--1-AP, 1:800) and mouse anti-\u03b2 actin (Novus, Centennial, CO; NB600--501, 1:10,000) as a loading control with a Licor Odyssey Infrared Imaging System. For immunofluorescent detection of rBAT, kidney samples from 10 to 12\u2009week old mice kept on a regular diet were harvested, stained, and imaged as previously described except that sections were paraffin embedded and sectioned by the Vanderbilt University Medical Center Translational Pathology Shared Resource \\[[@CR22]\\]. The primary rabbit anti-rBAT (16343--1-AP, Proteintech, Chicago, IL) diluted 1:200 was followed by donkey anti-rabbit Alexafluor 594 (Life Technologies, Waltham, MA) diluted 1:500 to detect the rBAT protein in red. Nuclei are stained with 4,6-Diamidino-2-phylindole (DAPI) in blue. For hematoxylin and eosin (H&E) staining of kidney and bladder samples, the bisected kidneys and bladders with stones removed were placed into 4% paraformaldehyde at 4\u2009\u00b0C overnight with rocking. The tissues were paraffin-embedded, sectioned, and stained by the Vanderbilt Translational Pathology Shared Resource Core. Images were acquired on an Olympus BX51 microscope.\n\nX-ray imaging {#Sec6}\n-------------\n\nThe image shown is of fourteen-week-old male *Slc3a1*^\u2212/\u2212^ mice. To monitor stone formation over time, mice were taken to the Vanderbilt University Institute of Imaging Science, anesthetized with isoflourane and oxygen until immobile, then placed in a Faxitron 2000 X-ray machine at setting 35 for an exposure time of 4\u2009s.\n\nBiomarker and amino acid measurements {#Sec7}\n-------------------------------------\n\nFor BUN and creatinine measurements, blood was collected into a microvette CB 300 Z tube with clot activator (Sarstedt, Newton, NC) by submandibular bleed under isoflurane anesthesia with a 5.5\u2009mm lancet. Samples were clotted at room temperature for \\>\u200930\u2009min and centrifuged at 4500\u2009rpm for 10--20\u2009min. Serum was immediately aliquoted and stored at \u2212\u200980\u2009\u00b0C for measurement of the blood urea nitrogen by the Vanderbilt University Medical Center Comparative Pathology Laboratory and measurement of creatinine by the University of Alabama-Birmingham O'Brien Center Core C Biomarkers Laboratory by LC-MS/MS. Plasma amino acid concentrations were determined by reverse phase HPLC using a modified version of the methods of Bidlingmeyer et al. \\[[@CR23]\\].\n\nGlutathione measurements {#Sec8}\n------------------------\n\nGlutathione levels were measured using a GSH-GLO Glutathione Assay (Promega, Madison, WI) according to the manufacturer's instructions. The protein concentration of liver lysates was determined by BCA and concentrations normalized to 400\u2009ng/\u03bcl. Liver lysates were assayed +/\u2212\u2009500\u2009\u03bcM TCEP. TCEP reduces any oxidized glutathione present in the sample. Luminescence was measured using a FLUOStar-Omega microplate reader.\n\nResults {#Sec9}\n=======\n\nWe first confirmed knockout of *Slc3a1* using RT-PCR to evaluate expression levels in various tissues. As expected, *Slc3a1* RNA levels were dramatically reduced in tissues where expression normally occurs, such as the kidney and small intestine (Fig.\u00a0[1](#Fig1){ref-type=\"fig\"}a). Additionally, as expected, *Slc7a9* RNA levels were unaltered by knockout of *Slc3a1* (Fig. [1](#Fig1){ref-type=\"fig\"}b)*.* We next used Western blot and immunofluorescent microscopy to evaluate for loss of rBAT expression in kidneys of *Slc3a1*^\u2212/\u2212^ mice. Western blot demonstrated loss of rBAT expression from kidney lysates (Fig. [1](#Fig1){ref-type=\"fig\"}c). Immunofluorescent microscopy for rBAT demonstrated that knockout of *Slc3a1* resulted in loss of rBAT expression in the proximal tubules of *Slc3a1*^\u2212/\u2212^ mice when compared the wild type mice (Fig. [1](#Fig1){ref-type=\"fig\"}d and e).Fig. 1*Slc3a1* transcripts and rBAT protein expression are lost in male *Slc3a1* knockout mice. *Slc3a1*^\u2212/\u2212^ mice demonstrate loss of *Slc3a1* RNA (**a**) but retention of *Slc7a9* RNA (**b**) in kidney and small intestine. Bladder and liver are provided as negative controls. Shown are the averages of two independent experiments done in triplicate (mean\u2009\u00b1\u2009SD). (**c**) Representative Western blot of rBAT from wild type and *Slc3a1*^\u2212/\u2212^ mice. Immunofluorescence was used to evaluate for rBAT expression in wild type (**d**) and male *Slc3a1*^\u2212/\u2212^ mice (**e**). Knockout animals demonstrate loss of rBAT expression in the proximal tubule\n\nKnockout mice did not develop kidney stones but did develop bladder stones detectable by x-ray as reported previously (Fig.\u00a0[2](#Fig2){ref-type=\"fig\"}a) \\[[@CR10]\\]. Consistent with previous observations, we observed gender differences between male and female *Slc3a1*^*\u2212/\u2212*^ mice with female mice very rarely developing bladder stones \\[[@CR10]\\]. We did not observe a high rate of bladder stone formation in male *Slc3a1*^\u2212/\u2212^ mice on normal chow (0.31% cystine) (Table\u00a0[2](#Tab2){ref-type=\"table\"}). However, when we placed the male *Slc3a1*^\u2212/\u2212^ mice on breeder chow (0.36% cystine), we observed a higher rate of bladder stone formation. Comparing the % of mice with bladder stones at 28\u2009weeks, 100% of *Slc3a1*^\u2212/\u2212^ mice on breeder chow had bladder stones whereas only 42% of *Slc3a1*^\u2212/\u2212^ mice on regular chow exhibited bladder stone formation (Fig. [2](#Fig2){ref-type=\"fig\"}b). The rate of bladder stone formation in mice on the breeder chow is depicted in Fig. [2](#Fig2){ref-type=\"fig\"}c. These results demonstrate that dietary intake can have a major impact on the rate of stone formation in *Slc3a1*^\u2212/\u2212^ mice.Fig. 2Male *Slc3a1*^\u2212/\u2212^ mice demonstrate differing rates of bladder stone formation dependent upon diet. **a** X-rays of mice with typical bladder stone formation. **b** 42% of mice on normal chow had bladder stones by 28\u2009weeks, whereas 100% of mice had bladder stones by 28\u2009weeks on breeder chow (*N*\u2009=\u200912). **c** X-ray of *Slc3a1*^\u2212/\u2212^ animals revealed the development bladder stone formation on breeder chow as depicted by a Kaplan-Meier plot (*N*\u2009=\u20098)Table 2Comparison of normal and breeder chow dietsStandard (5L0D)Breeder (5LJ5)Crude protein not less than23%17%Crude fat not less than4.5%11%Cystine0.31%0.36%Sodium0.40%0.43%\n\nWe observed that male *Slc3a1*^\u2212/\u2212^ mice were consistently lower in weight when compared to *Slc3a1*^+/+^ mice of the same age (Fig.\u00a0[3](#Fig3){ref-type=\"fig\"}a,b). This weight difference was not recapitulated in the female mice. All mice exhibited the same length from their nose to the tip of their tail, indicating an overall metabolic phenotype rather than nutritional deficiency contributed to the difference in weights of the male mice raised and maintained on breeder chow (Fig. [3](#Fig3){ref-type=\"fig\"}c). We evaluated the plasma amino acid levels in *Slc3a1*^\u2212/\u2212^ mice and compared them to wild type mice (Table\u00a0[3](#Tab3){ref-type=\"table\"}). Consistent with the reduced weight in male *Slc3a1*^\u2212/\u2212^ mice, we observed a more severe phenotype with multiple amino acid levels different between wild type and knockout animals. Evaluation of amino acids in both male and female *Slc3a1*^\u2212/\u2212^ mice revealed only ornithine, lysine, and taurine to be reduced in both sexes. Therefore, knockout of rBAT resulted in reduce plasma ornithine and lysine in these animals. Cysteine, which forms cystine when two molecules are joined together, is metabolized to taurine \\[[@CR24]\\] which was also reduced across sexes. Interestingly, plasma cystine was reduced only in male *Slc3a1*^\u2212/\u2212^ despite both male and female mice exhibiting cystinuria \\[[@CR11]\\].Fig. 3Male *Slc3a1* \u2212/\u2212 mice exhibited lower weight compared to wild type mice of the same age. **a** Comparison of weight of 10--12\u2009week old wild type (WT) and *Slc3a1*^\u2212/\u2212^ (KO) animals on breeder chow (\\*, *p*\u2009\\<\u20090.05 Mann Whitney test, *N*\u2009=\u20095--6\u2009\u00b1\u2009SD). **b** Picture of age matched (12\u2009week old) WT (left) and KO (right) male mice. **c** All mice exhibited the same nose-to-tail length (*N*\u2009=\u20095--6\u2009\u00b1\u2009SD)Table 3Plasma amino acid levels (\u03bcmole/liter)Female*p*Male*p*Both*Slc3a1* ^+/+^*Slc3a1* ^\u2212/\u2212^*Slc3a1* ^+/+^*Slc3a1* ^\u2212/\u2212^*p*Aspartate8.4\u00a0\u00b1\u00a01.68.5\u00a0\u00b1\u00a02.90.459.44\u00a0\u00b1\u00a02.949.6\u00a0\u00b1\u00a02.10.45Glutamate23\u00a0\u00b1\u00a02.635\u00a0\u00b1\u00a0190.0950\u00a0\u00b1\u00a010.646\u00a0\u00b1\u00a08.70.3Serine/Asparagine61\u00a0\u00b1\u00a09.346\u00a0\u00b1\u00a0100.012,\\*46.1\u00a0\u00b1\u00a05.2351\u00a0\u00b1\u00a09.70.12Glycine126\u00a0\u00b1\u00a017115\u00a0\u00b1\u00a0220.19104\u00a0\u00b1\u00a011.1127\u00a0\u00b1\u00a0150.007,\\*Histidine/Glutamine383\u00a0\u00b1\u00a026363\u00a0\u00b1\u00a0400.16310\u00a0\u00b1\u00a044416\u00a0\u00b1\u00a0650.004,\\*3-Methyhistidine258\u00a0\u00b1\u00a078275\u00a0\u00b1\u00a0660.3557\u00a0\u00b1\u00a09.9746\u00a0\u00b1\u00a09.20.03,\\*1-Methyhistidine10.6\u00a0\u00b1\u00a01.511\u00a0\u00b1\u00a03.30.296.5\u00a0\u00b1\u00a00.185.9\u00a0\u00b1\u00a00.113.2e^\u22125^,\\*Taurine157\u00a0\u00b1\u00a041101\u00a0\u00b1\u00a020.007,\\*108\u00a0\u00b1\u00a08.3892\u00a0\u00b1\u00a0180.04,\\*\\*Arginine34\u00a0\u00b1\u00a06.932\u00a0\u00b1\u00a07.40.427\u00a0\u00b1\u00a04.1530\u00a0\u00b1\u00a05.20.15Threonine/Citruline240\u00a0\u00b1\u00a025196\u00a0\u00b1\u00a0380.022,\\*179\u00a0\u00b1\u00a038149\u00a0\u00b1\u00a0340.09Alanine271\u00a0\u00b1\u00a037224\u00a0\u00b1\u00a0800.11248\u00a0\u00b1\u00a047196\u00a0\u00b1\u00a0400.03,\\*Proline76\u00a0\u00b1\u00a01854\u00a0\u00b1\u00a0150.027,\\*57\u00a0\u00b1\u00a01545\u00a0\u00b1\u00a0140.1Hydroxyproline10\u00a0\u00b1\u00a02.46.1\u00a0\u00b1\u00a020.006,\\*8.18\u00a0\u00b1\u00a00.8810\u00a0\u00b1\u00a01.20.006,\\*Cystine5.9\u00a0\u00b1\u00a03.66.2\u00a0\u00b1\u00a03.60.4513\u00a0\u00b1\u00a03.47.05\u00a0\u00b1\u00a01.50.0008,\\*Tyrosine45\u00a0\u00b1\u00a01251\u00a0\u00b1\u00a0130.2137\u00a0\u00b1\u00a09.334\u00a0\u00b1\u00a08.70.28Valine129\u00a0\u00b1\u00a021117\u00a0\u00b1\u00a0270.2110\u00a0\u00b1\u00a02588\u00a0\u00b1\u00a0240.07Methionine69\u00a0\u00b1\u00a01553\u00a0\u00b1\u00a0170.0657\u00a0\u00b1\u00a019.346.5\u00a0\u00b1\u00a0160.15Ornithine35\u00a0\u00b1\u00a0611\u00a0\u00b1\u00a023.7e^\u22126^,\\*34\u00a0\u00b1\u00a01010.5\u00a0\u00b1\u00a03.20.0001,\\*\\*Lysine200\u00a0\u00b1\u00a03086\u00a0\u00b1\u00a0154.5e^\u22126^,\\*147\u00a0\u00b1\u00a03177.4\u00a0\u00b1\u00a0160.0003,\\*\\*Isoleucine60\u00a0\u00b1\u00a01159\u00a0\u00b1\u00a0190.4547\u00a0\u00b1\u00a011.635.3\u00a0\u00b1\u00a0110.04,\\*Leucine84\u00a0\u00b1\u00a01484\u00a0\u00b1\u00a0260.4869\u00a0\u00b1\u00a01654.9\u00a0\u00b1\u00a0160.07Phenylalanine36\u00a0\u00b1\u00a0535\u00a0\u00b1\u00a050.3229\u00a0\u00b1\u00a0425.5\u00a0\u00b1\u00a04.40.047,\\*Tryptophan34\u00a0\u00b1\u00a0726\u00a0\u00b1\u00a080.0630\u00a0\u00b1\u00a0328.1\u00a0\u00b1\u00a03.60.11Plasma amino acid levels were determined as described in the Materials and Methods section. Rows where two amino acids are described (i.e. Histidine/Glutamine) indicates that these amino acids are not separable on the chromatogram. \\*, indicates a *p*\u2009\\<\u20090.05 via Student's T test comparing KO to WT of that sex. The p for \"both\" indicates that this particular amino acids was reduced in both sexes comparing KO to WT\n\nPrevious reports showed elevation of BUN in some *Slc3a1*^\u2212/\u2212^ mice strains but no elevation in serum creatinine in others \\[[@CR8], [@CR9]\\]. We therefore sought to evaluate for the development of CKD in this mouse strain with a mixed C57Bl/6 and 129/SvJ background \\[[@CR10]\\]. Interestingly, male mice fed only normal chow showed elevated BUN, whereas male mice fed breeder chow showed both elevated BUN and creatinine (Fig.\u00a0[4](#Fig4){ref-type=\"fig\"}). Mice evaluated were between 40 and 87\u2009weeks old (Additional files [1](#MOESM1){ref-type=\"media\"} and [2](#MOESM2){ref-type=\"media\"}). Histologic evaluation of the kidney tissue from *Slc3a1*^\u2212/\u2212^ mice with higher creatinine on normal chow (Fig.\u00a0[5](#Fig5){ref-type=\"fig\"}b) demonstrated dilated tubules from obstruction and increased fibrosis when compared to those with normal creatinine (Fig. [5](#Fig5){ref-type=\"fig\"}a). The observed dilation of tubules and fibrosis was also observed in mice with elevated creatinine (Fig. [5](#Fig5){ref-type=\"fig\"}d) compared to normal creatinine when placed on the breeder chow (Fig. [5](#Fig5){ref-type=\"fig\"}c). Histologic evaluation of the bladders of male *Slc3a1*^\u2212/\u2212^ mice revealed increased inflammatory infiltrate within the bladder wall and cystine stones in the bladder in mice having a normal serum creatinine (Fig. [5](#Fig5){ref-type=\"fig\"}f) when compared to the bladders of wild type mice (Fig. [5](#Fig5){ref-type=\"fig\"}e). Some *Slc3a1*^\u2212/\u2212^ mice with an elevated serum creatinine showed embedded cystine crystals within the bladder wall in addition to bladder inflammatory infiltrate (Fig. [5](#Fig5){ref-type=\"fig\"}g).It is known that cystinuria results in stone formation leading to obstruction and CKD. However, we sought to determine if there might be other metabolic consequences of cystinuria beyond just stone formation. Glutathione, or \u03b3-l-glutamyl-l-cysteinyl-glycine (GSH), serves as a detoxifier, protector from oxidative damage, and is arguably the most important low molecular weight antioxidant in cells \\[[@CR25]\\]. The availability of cysteine, formed by two molecules of cystine, is a major determinant of the regulation of GSH synthesis \\[[@CR26]\\]. Therefore, cystine deficiency could affect GSH levels. Martensson et al. previously reported decreased leukocyte GSH levels in homozygous cystinuric patients \\[[@CR21]\\]. We sought to evaluate GSH levels in male *Slc3a1*^\u2212/\u2212^ mice which exhibited lower plasma cystine levels. Therefore, we evaluated the levels of reduced (GSH) and oxidized (GSSG) glutathione in the livers of *Slc3a1*^\u2212/\u2212^ mice. We observed lower GSH and lower GSSG in the livers of *Slc3a1*^\u2212/\u2212^ mice compared to WT controls (Fig.\u00a0[6](#Fig6){ref-type=\"fig\"}). The total amount of glutathione (GSH\u2009+\u2009GSSG) was lower (Fig. [6](#Fig6){ref-type=\"fig\"}), implying a possible connection between the lower availability of cysteine and total glutathione reserve. A reduction of the ratio of reduced to oxidized forms of glutathione was also observed (Fig. [6](#Fig6){ref-type=\"fig\"}), suggesting that the ability of the liver to respond to oxidative stress is reduced in *Slc3a1*^\u2212/\u2212^ mice.Fig. 4Male *Slc3a1* \u2212/\u2212 mice have higher blood urea nitrogen (BUN) and creatinine indicative of the development of CKD. **a** Comparison of BUN in wild type (WT) and *Slc3a1*^\u2212/\u2212^ (KO) animals on a regular or breeder chow diet. KO animals exhibited elevated BUN on both regular and breeder chow when compared to WT animals on the same diet (\\*, *p*\u2009\\<\u20090.05 Mann Whitney test, *N*\u2009=\u20096--19\u2009\u00b1\u2009SEM). **b** Comparison of serum creatinine in WT and KO animals on a regular or breeder chow diet. KO animals exhibited elevated creatinine on the breeder chow diet but not the regular chow diet (\\*, *p*\u2009\\<\u20090.05, Mann Whitney test, *N*\u2009=\u20096--19\u2009\u00b1\u2009SEM). Each dot on the graph represents an individual measurement from an individual mouseFig. 5Histological analysis of wild-type and male *Slc3a1*^\u2212/\u2212^ kidneys and bladders at 10--12\u2009months of age. Kidneys of mice with elevated creatinine (**b** and **d**) demonstrate more tubular dilatation and fibrosis when compared to mice with a normal creatinine (**a** and **c**) whether on regular (**a** and **b**) or breeder chow (**c** and **d**). Bladders of wild-type (**e**), knockout with normal creatinine (**f**), and knockout with elevated creatinine (**g**) mice are also shown. The knockouts demonstrate inflammatory infiltrate and cystine crystals. Panel (**g**) demonstrates cystine crystals embedded within the bladder wall (arrows). Shown are representative H&E stains from each group of miceFig. 6Glutathione levels in male *Slc3a1*^\u2212/\u2212^ mice at 8\u2009weeks of age. Reduced (GSH) and oxidized (GSSG) glutathione levels were measured in the livers of wild-type and *Slc3a1*^\u2212/\u2212^ animals. GSH, GSSG, total and the ratio of reduced to oxidized forms (GSH:GSSG) were all lowered in *Slc3a1*^\u2212/\u2212^ mice (*N*\u2009=\u20096\u2009\u00b1\u2009SEM)\n\nDiscussion {#Sec10}\n==========\n\nTreatment for cystinuria has not dramatically changed in the past 20\u2009years despite greater understanding of the genetic basis \\[[@CR1]\\]. Mouse models for cystinuria not only confirm the genetic basis of the disease but also provide models for testing and evaluating new therapies \\[[@CR10]\\]. There are likely genes that affect stone formation in patients with cystinuria that could be derived from analysis of patients and then tested in mouse models. There are likely genes that affect stone formation in patients with cystinuria that could be derived from analysis of patients and then tested in mouse models. Recently, Zee et al. demonstrated that \u03b1-lipoic acid prevents cystine stone formation in the mouse model reported in this article \\[[@CR11]\\].\n\nThree type A cystinuria mouse models have been generated including the one described herein that was generated by exon 1 deletion \\[[@CR10]\\]. The other models consist of spontaneous mutation (E383K) \\[[@CR9]\\] or ENU induced mutagenesis (D140G) \\[[@CR8]\\]. The D140G model on a 129S2/SvPasCrl background demonstrated a 40% reduction in survival by 15\u2009weeks with elevation in serum BUN though no increase in serum creatinine or interstitial kidney fibrosis \\[[@CR9]\\]. Possible gender differences in these observations were not commented on though stones mainly occurred in males \\[[@CR9]\\]. The E383K model on the C3HeB/FeJ background showed elevated serum BUN in 20\u2009week old male mice and decreased kidney weight in 32--48\u2009week old male mice with neither phenotype observed in female mice \\[[@CR8]\\]. The observed gender difference in stone formation between male and female in the context of cystinuria has been observed in patients and reviewed elsewhere \\[[@CR27], [@CR28]\\]. The difference in stone formation does not appear to be due to differences in urinary cystine levels, but it may due to differential cystine aggregation in male versus female urine \\[[@CR27], [@CR29]\\]. Nonetheless, our study demonstrates that CKD does occur in aged male *Slc3a1*^*\u2212/\u2212*^ mice.\n\nLittle is known about what factors affect the phenotype of cystinuria and progression to CKD. \\[[@CR2]\\] We found that the diet fed to the *Slc3a1*^\u2212/\u2212^ mice had a dramatic effect on the rate of stone formation. Common treatment for cystinuria includes limiting dietary sodium and protein intake, increased fluid intake, urinary alkalization, and possibly even thiol drugs and captopril \\[[@CR1]\\]. A previous study demonstrated that a low protein diet with more plant protein sources reduced cystine excretion in cystinuria patients \\[[@CR30]\\]. We have not found any studies on the role of diet in cystine stone formation in mouse models of cystinuria. Human clinical studies on the effects of diet for potential modification of human cystinuria are lacking \\[[@CR2]\\]. Future studies could be directed at evaluating dietary changes in a more controlled manner to tease out modifications which could slow the progression of stone formation in cystinuria.\n\nThe inability to reclaim cystine from the urine likely alters cystine metabolism in the whole animal. Knockout of *Slc3a1* affects its expression not only in the kidney but also the intestine. The intestinal peptide transporter Pept1 (*Slc15a1*) is thought to permit reabsorption of cystine and other amino acids from the gut somewhat compensating for loss of *Slc3a1* in the intestine \\[[@CR31], [@CR32]\\]. How this might contribute to overall gender difference in cystinuria and our observed differences in plasma amino acid levels is currently unknown.\n\nOur results reveal the metabolic phenotype to be more severe in male knockout animals when compared to females. Glutathione levels were lower in the *Slc3a1*^\u2212/\u2212^ mice. Given the importance of glutathione in detoxification and oxidative stress, cystinuria may result in an overall metabolic state of increased sensitivity to injury from a variety of mechanisms \\[[@CR25]\\]. As patients with cystinuria have been reported to develop more CKD than other stone formers, lower levels of GSH could increase the sensitivity of *Slc3a1*^\u2212/\u2212^ mice to development of and recovery from injury resulting in worsened CKD. Previous research has demonstrated that anti-oxidant cystine metabolites, GSH, and taurine exert protection against progression of renal fibrosis \\[[@CR33]--[@CR35]\\]. Additionally, others have found these to be protective against renal ischemia reperfusion injury \\[[@CR36]--[@CR39]\\]. In patients, cystinuria results in kidney stones and obstruction. Reduced glutathione could result in worsened injury or inability to recover from injury in cystinuria when compared to stone formers with non-cystinuric disease pathology. Importantly, new therapies such as \u03b1-lipoic acid may reduce stone formation by increasing solubility of cystine in the urine but have no effect on regaining cystine transport \\[[@CR11]\\]. Therefore, such therapy may reduce stone formation but may not improve cystine or oxidative metabolism overall. Our data suggest that diet manipulation, supplementation, and/or drugs that work to improve the ratio of reduced to oxidized glutathione in vivo should be tested in mouse models for their effect on the development of CKD. Such studies would provide further guidance for care of cystinuric patients and hopefully guide the development of new therapies to help prevent kidney disease progression in this population.\n\nConclusions {#Sec11}\n===========\n\nWe have demonstrated that dietary intake can have an impact on the severity of the phenotype of cystinuria both in regards to stone formation and the development of CKD in a mouse model of cystinuria. Male mice exhibited a more severe phenotype concerning animal weight and plasma amino acid composition. Cystinuria affected glutathione composition in the liver. Future studies will systematically evaluate dietary components on the cystinuria phenotype and what effect the resultant alteration in glutathione metabolism has on physiology and pathophysiology.\n\nAdditional files\n================\n\n {#Sec12}\n\nAdditional file 1:Supplementary figure demonstrating BUN and creatinine over time in WT and KO mice on regular and breeder chow. (PDF 135 kb) Additional file 2:Table of WT and KO mice on regular and breeder chow with BUN, creatinine, gross morphology and stone weight. (XLSX 11 kb)\n\nb^0,+^AT\n\n:   *Slc7a9* gene product\n\nBUN\n\n:   Blood urea nitrogen\n\nCKD\n\n:   Chronic kidney disease\n\nDAPI\n\n:   4,6-Diamidino-2-phylindole\n\nENU\n\n:   N-ethyl-N-nitrosourea\n\nGSH\n\n:   \u03b3-l-glutamyl-l-cysteinyl-glycine, glutathione\n\nGSSG\n\n:   Oxidized glutathione\n\nKO\n\n:   Knockout\n\nrBAT\n\n:   *Slc3a1* gene product\n\nSEM\n\n:   Standard error of the mean\n\nWT\n\n:   Wild type\n\n**Publisher's Note**\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n\nNone.\n\nLEW and MHW conceived and designed the experiments. LEW, RCW, RAV, FS, TAK, and TMB acquired and analyzed the data. LEW, EJW, JAT, AS, and MHW wrote and revised the manuscript. All authors read and approved the final manuscript.\n\nThe funders played no role in the design of the study, collection, analysis, and interpretation of data, or in writing the manuscript. L.E.W. was supported by a Career Development Award from the Department of Veterans Affairs \\[BX002797\\], the Vanderbilt O'Brien Kidney Center \\[P30 DK114809\\], the Vanderbilt Center for Kidney Disease, and a Career Development Award from the American Society of Gene & Cell Therapy. J.A.T. and A.S. were supported by the National Institutes of Health \\[DK11278--01\\]. M.H.W. was supported by the National Institutes of Health \\[DK093660\\], Department of Veterans Affairs \\[BX002190 and BX004258\\], and by the Vanderbilt Center for Kidney Disease. Histology, blood chemistry, and hematology services were provided by the Translational Pathology Shared Resource at Vanderbilt, supported by NCI/NIH Cancer Center Support Grant \\[2P30 CA068485--14\\], the Vanderbilt Mouse Metabolic Phenotyping Center Grant \\[5U24DK059637--13\\], and the Vanderbilt O'Brien Kidney Center \\[P30 DK114809\\].\n\nThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\n\nAll animal studies were approved by the Institutional Animal Care and Use Committee of the Nashville Tennessee Valley Healthcare System VA and Vanderbilt University Medical Center.\n\nNot applicable.\n\nThe authors declare that they have no competing interests.\n"}
+{"text": "Amalberti R, Brami J. 'Tempos' management in primary care: a key factor for classifying adverse events, and improving quality and safety. This paper was published in print with an incorrect doi. The print citation is *BMJ Qual Saf* 2012;21:729--736 and the correct doi is 10.1136/bmjqs.2010.048710.\n"}
+{"text": "![](indmedgaz72455-0017){#sp1 .77}\n\n![](indmedgaz72455-0018){#sp2 .78}\n\n![](indmedgaz72455-0019){#sp3 .79}\n\n![](indmedgaz72455-0020){#sp4 .80}\n\n![](indmedgaz72455-0018-a){#f1 .78}\n"}
+{"text": "Surgeons have applied the principles of tissue engineering for years; they transplant and shift bone and other tissues within a patient to promote regenerative potential. The advent of technology for fabricating structures (eg, matrices, TCP bone fillers, etc.) with geometric fidelity, compositional consistency, and tissue-specific identity offers further promise that regenerative potential of tissue and whole organ systems can be achieved. While attaining scar tissue might be sufficient for soft tissue applications such as skin or in muscle or tendon, achieving an outcome in most orthopedic indications that is not mechanically solid and weight-supporting would be insufficient.\n\nImplicit in the strategy of repairing bone is a need to gain restitution of space, achieve mechanical integrity, and regenerate functional continuity. When the biology or anatomy are insufficient to attain a full repair, therapeutic use of graft material has been used to omit compliance features such as strain tolerance, reduced stiffness, and attenuated strength, and instead promote primary or membranous-type bone formation within the physical approximation of a graft material. In the broadest sense, 3 basic components are required of a graft: osteoprogenitor cells, osteoinductive factors, and an osteoconductive matrix or scaffold.\n\nA more critical assessment would discriminate cell lineage and capacity for differentiation, define metabolic support with regard to defined vascular and immunogenic transparency, and evaluate material properties such as surface area-to-volume ratio, porosity, and modeling and degradation capacity during early integration. Any of these subcategories is subject to further reduction, and in the context of a short overview of appropriate carriers, this discussion has been limited to conductive properties that offer advantages for bone repair and might facilitate the delivery of cells, or mineral, or even consider cytokines for adjunctive intervention. Within the domain of osteoconductivity, shape, composition and matrix turnover are inextricable components.\n\nThe shape of the biomaterial template is critical to the success of an osteoconductive carrier. The essence of shape permeates more than just the 3-dimensional (3-D) form of the material and is part of the molecular domain as well. A central tenet of biomineralization is that nucleation, growth, morphology, and aggregation of the inorganic crystals of bone are regulated by organized assemblies of organic macromolecules. The close spatial relationship of hydroxyapatite crystals with type-I collagen fibrils in the early stage of bone mineralization is a relevant example. Hydroxyapatite is a natural mineral component of hard tissue, composing 60 --70% of bone. It is also evident that combining hydroxyapatite with protein does not render the macroscopic form of bone nor impart its characteristic properties. Unlike fabricated materials that can be developed from components with predictable properties, biological systems control desired properties with an intrinsic rationale that discriminates essential from nonessential factors.[@CIT0001] Living organisms avoid geometric randomness by segregating structures that resonate with function. Anatomical variations that do not result in significant input to the whole organism remain \"neutral\" with regard to selection pressures. Within the context of \"more demand -- more function\" equilibrium is the eventual arbiter of change---biologic systems are not static but in a constant shifting response to their stimulation. To a large extent the symmetry of the stimulus imposes an order of stability.\n\nThe challenge of most conductive materials is that they emerge from a static platform and, in placement, force the living system to adapt to placement, dimension, and different properties, and eventually are only successful in degradation and replacement or in integration. Materials that have been developed for orthopedic applications and made available as grafting substitutes include cancellous and cortical allograft bone, ceramics such as sintered coralline matrices, hydroxyapatite and tri-calcium phosphate, demineralized bone matrix, bone marrow, composite polymer grafts, and recently various combination carriers endowed with growth factors. Complications include availability, cost, variable biological absorption profiles, brittleness, immune stimulation, as well as the economic reality of regulatory hurdle. Polymers, in and of themselves, constitute a nearly uncontainable universe of bone application potentials. Many natural and synthetic polymers have wide use in bone engineering and material development. Among the best known and characterized of the synthetic polymers are polycaprolactone (PCL), polyethylene glycol (PEG), poly(L-lactide) (PLLA), and polyglycolide (PGA) and copolymers such as poly(lactic-*co*-glycolic acid) (PLGA). Natural polymers such as collagen and hyaluronic acid (HA) also have been widely defined for potential use in bone applications. Early use of these polymer matrices was widely considered little more than place holders for cells, in essence armatures for placing cells and maintaining tissue dimension during the healing process.\n\nIn the context of understanding matrix-cell interaction, a demand for new and more sophisticated matrices has been fostered. Rather than merely defining a place for cells to rest, materials now play an active role in guiding tissue development. Although bone can appear de novo, it more often develops from accretion on a scaffold of matrix that contains appropriate vascular and compositional arrangement. As such, both 2-dimensional (2-D) and 3-D patterns have been shown to enhance osteoconductivity.[@CIT0002] Bone has significantly more matrix than cells, and cell regulation through anchorage dependent mechanisms is an established premise.[@CIT0003]--[@CIT0005] Compensatory mechanisms for changing sensitivity to mechanical stimulation have been shown to undergo adaptive or kinetic regulation, likely tied directly to osteoblast attachment to immobilized molecules in the extracellular matrix (ECM). Extracellular matrix molecules promote cell spreading by resisting cell tension, thereby promoting structural rearrangements. Ideally, the evolution of new materials will provide more than a substrate affording tissue compatibility and define a scaffold that will be not only be structurally enhancing but conductively optimum for bone formation.\n\nAn early strategy for enhancing primary binding sites in bone tissue engineering was to include integrin polypeptide sequences in the backbone.[@CIT0006] Integrins are cell-surface glycoprotein receptors which mediate interactions between similar and different cells as well as between cells and extracellular matrix proteins. These interfaces are involved in physiological processes, such as embryogenesis, hemostasis, wound healing, immune response, and formation/maintenance of the tissue architecture.[@CIT0007]--[@CIT0009] As has so often been the case with efforts to structure natural systems, further observation has resulted in better definition of the material needs.[@CIT0010] The ECM serves a dual role to the extent that the provisional matrix must not only serve the foundation for bone repair but also mediate a biophysical barrier to prevent fibroblast invasion and generation of scar tissue.[@CIT0011], [@CIT0012] Given the need for cell migration and cell situation in proximity to the area, one of the important challenges to achieve in optimizing matrix conductivity is understanding the pericellular environment and the regulation of proteases, vascular invasion, receptor specificity, and cell attachment and differentiation. The synergy and interdependency between adhesion, ECM, and proteolysis are important concepts that must be understood to engineer scaffolds capable of holding up to standards that are more than cell decoration. Moreover, the reactive specificity to loading, degradation, therapeutic delivery during absorption remains a key aim of both academic and industrial designs. Achieving conductivity comes with challenges of best-fit integration, delivery, and in integrated modeling. The more liquid the delivery is, the more modular the components are; and the more adaptive the matrix is to meeting the intended application, the more likely that the conductivity will not be excluded by the morphology of the injury site.\n\nThe concept of molecular self-assembly for the development of new biomaterials comes from close observation and modeling of events well known in biology.[@CIT0013] In the most primal of function form, DNA sets a foundation that resonates in microfilament and microtubule assembly, in its self-complementary double-helix annealing, and in lipid membrane development. Self-assembling properties are also found in proteins that are critical to forming the extracellular matrix of connective tissues such as collagens, laminins, and fibronectins.[@CIT0014]--[@CIT0016] Within the assembled matrix, key sequences have been shown to promote cell adhesion, cell migration, endothelial cell monolayer development, and the inhibition of angiogenesis. Enrichment of sequences in defined and specific synthetic approaches has been shown to facilitate desired integration of cells to scaffolding.[@CIT0017] Among the more known extensions of applied self-assembly potential has been the development of Matrigel (BD Biosciences, San Jose, California) by Hynda Kleinman at the NIH; countless studies have used the material as a standard for 3-D matrix studies.\n\nSemino offers a review of future platforms for designer matrices.[@CIT0018] With a strong basis founded in work with self-assembling hydrogels,[@CIT0019]--[@CIT0021] the sentinel element of future application promises the ability to extend structural and biomechanical similarities of matrices that additionally provide instructive capacity for cells as a regulatory template that specifies cell signaling. Most likely, the matrices and gels will need to play complementary roles---first optimizing the conditions for conduction of the proper cells to the repair site and, second, amplifying and possibly enhancing specific intentions in compromised patients.\n\nOne such application might be in consideration of patients who use tobacco. Evidence of nicotine interfering with bone healing in the spine is well known.[@CIT0021]--[@CIT0023] With knowledge of an endothelial nicotinic acetylcholine receptor (nAChR) being instructive to endothelial proliferation, survival, migration and tube formation in vitro, it might be possible to fine-tune the hydrogel or self-assembling matrix to block exogenous nicotine receptors that retard this angiogenic pathway. Traditional drug delivery systems have been based on synthetic polymer materials, or animal-derived collagen, which may contain residual growth and/or viruses from animal tissues. Peptide hydrogels are ideally suited for drug delivery as they are pure, easy to design and use (eg, non-toxic, nonimmunogenic, bio-absorbable), and can be applied locally to a particular tissue.\n\nSince Zhang first discovered peptide hydrogels in the early 1990s, numerous applications have been developed that show promise in regenerative medicine.[@CIT0024] Composed of self-assembling amino acid chains (peptides), the gel is about 99% aqueous by volume and offers a deliverable, low-viscosity solution for reaching difficult anatomy. Coupled to needed flow and deliverability qualities and to the formulation of the lattice hydrogels, it is possible to deliver molecules specific to accentuating conductivity, reducing potential inhibition, and exaggerating the biophysical properties of matrix. These gels can be chemically engineered to release proteins from the gel over hours, days, or even months, and the gel itself is eventually broken down into harmless amino acids, which are the building blocks of proteins. While not offering the initial strength needed for weight bearing, important aspects of conductivity needed to support the repair and integrated regenerative modeling are found.\n\nA potential to exploit the instructive capacity of the hydrogels in conduction with polymerizing, or even physically-static, conductive scaffolds seems to offer opportunity for synergism. Critical to therapeutic adoption for bone repair products are attributes such as immediate mechanical properties, functional biological activity during integration that does not weaken the construct, and a resorption profile that is metabolically benign. Additionally, each therapeutic application must consider the trade-off gained in attaining a tight apposition to the walls of the defect thus achieving mechanical solidarity and avoiding structural gaps in the delivery that might predispose the repair to fibrous interposition with a safe, with off-the-shelf, cost-effective application.\n\nOne interesting material in early stages of development is an osteconductive calcified triglyceride with remarkable bone-like mechanical properties.[@CIT0025] Initially a liquid created by combining fatty acids and calcium carbonate, the material is touted as an isothermal, non-toxic alternative to methyl methacrylate. When used as a bone void filler, carbon dioxide generated during the curing process extends a porosity that supports bone integration, and material adhesive qualities stabilize the now-filled margins of the defect at the interface of the native tissue. Although adhesive, the material is readily molded to any shape. As formulated, the material is conductive, resorbable, has an isothermal curing temperature, and during consolidation achieves porosity similar to that of bone. More critical to the adoption may be the fact that block material achieves similar mechanical characteristics to bone within 24 hours.\n\nFuture matrices would seem to benefit from the synergy gained by incorporating amphiphilic hydrogel moieties that support appropriate cell differentiation within the structural sufficiency provided by the solid material. Secondary domains of improvement could include the addition of stem cells into the matrix. As a first consideration, sourcing autologous adult stem cells to support regeneration and integrate bone voids would provide appropriate cells in conjunction with the osteoconductive matrix. One of the key values in this pairing is to provide inflammatory insulation gained by the use of stem cells during the repair process and hasten the conversion of matrix. Gains made in immune modulation are particularly important in building skeletal material, where the need to sustain weight-bearing support is critical to stabilizing the repair.\n\nConsiderations for osteoconductive materials for bone repair and replacement have developed conceptually and advanced parallel with a better understanding of not only bone biology but of materials science. First models of material replacements utilized a reductionist-constructionist logic: define the constituents of the material in terms of its morphology and chemical composition, and then engineer material with similar content and properties as a means of accommodating a replacement. Unfortunately for biologic systems, empiric formulation is insufficient to promote adequate integration in a timely fashion. Early hydroxyapatite formulations required years to remodel, and rather than providing an embracing source of mechanical similarity, they challenged the body to overcome the insulation of difference instead of integrating the interface.\n\nBone conduction and a better understanding of materials enhancing skeletal repair has evolved from scalar understanding as a science in translational medicine. Knowledge of how cells attach to matrix remains a distinct part of the continuum recognizing how tissues respond to force and have optimized insight into what drives the biological assembly of tissues. That concept, in turn, reflects the cell mechanics of what is required to neutralize strain and necessary for the body to adapt a neutral biology. Commercially successful osteoconductive matrices will still need to provide sufficient mechanical stability for skeletal repair. Moreover, future matrices will need to translate their bio-logical surfaces as more than a scaffold to be decorated with cells. Conductivity will be improved by formulations that enhance function, further extended from understanding what composition best suits cell attachment, and be adopted by conveniences of delivery that meet those criteria.\n"}
+{"text": "###### Strengths and limitations of this study\n\n-   This is the first longitudinal study to compare the effects of adherence and non-adherence to spectacles in children following vision screening at age 4--5\u2009years on both visual acuity and developing literacy.\n\n-   Nesting the study within the Born in Bradford birth cohort allows adjustment for confounding factors.\n\n-   The study is observational in nature reflecting real-life adherence to spectacle wear.\n\n-   The study is not a randomised controlled trial; therefore, allocation to the adherent or non-adherent groups is not exact and may underestimate the effect of non-adherence.\n\nIntroduction {#s1}\n============\n\nVisual development in humans occurs in early\u00a0life[@R1] with the presence of reduced visual acuity (VA) in young children potentially indicating an associated condition such as significant refractive error, strabismus and/or amblyopia.[@R2] The UK National Screening Committee (UK NSC) recommends visual screening for all children at age 4--5\u2009years,[@R3] (first year of school) in order to identify a potential reduction in VA. For those who fail the screening test (\\>0.20\u2009logarithm of the minimum angle of resolution (logMAR) in one or both eyes),[@R3] the follow-up clinical pathway includes referral for a cycloplegic refraction and fundus examination to confirm the VA finding,\u00a0to determine the presence and magnitude of any refractive error and to rule out eye disease.[@R4] In those with reduced VA, treatment generally consists of the wearing of spectacles[@R5] and may be combined with occlusion therapy[@R6] (wearing an eye patch or atropine drops). However, adherence to treatment, both spectacle wear[@R7] and occlusion therapy, is known to be variable.[@R9]\n\nDecreased VA, both near and distance and also the presence of\u00a0refractive error in young children, has been reported to be associated with reduced literacy levels.[@R10] However, there is a paucity of evidence on the impact of non-adherence to spectacle wear on VA and early developing literacy in children. Early literacy skills such as letter recognition,[@R13] word reading and decoding[@R14] taught in the first years of school are indicators of future reading performance and educational attainment, which in turn affect long-term health and social outcomes.[@R15] The initial school years are a crucial time for the development of these key literacy skills[@R17] and it is important to understand the impact of non-adherence to spectacle wear on visual outcome and educational attainment.\n\nLow educational attainment is associated with socioeconomic deprivation,[@R16] which makes the investigation of the relationship between VA and literacy difficult, as in order to account for potential confounding factors, comprehensive epidemiological data are required. Born in Bradford (BiB) is a large birth cohort, which collected maternal and early-life measures from mothers and their children in Bradford and details of recruitment have been previously reported.[@R18] By linking separately\u00a0collected vision and literacy data in children in the BiB cohort, we had the opportunity to explore the association between VA, spectacle wear and literacy development while taking into account the effects of potential confounders. The aim of this study is to examine the impact of adherence to spectacle wear on VA and early developing literacy skills in children during their first 3\u2009years of school.\n\nMethods {#s2}\n=======\n\nThis is a prospective, longitudinal study nested within the BiB cohort following children from the point of their initial vision screening at age 4--5\u2009years. The study took place between 2012 and 2015. Baseline epidemiological data collected from mothers and children of the BiB cohort, literacy measures, vision screening results and repeat measures of vision and literacy were linked in order to evaluate the longitudinal impact of adherence to spectacle wear on VA and early literacy.\n\nPopulation {#s2a}\n----------\n\nAll children invited to join the study were participating in the BiB,[@R18] a longitudinal, multiethnic birth cohort study aiming to examine the impact of environmental, psychological and genetic factors on maternal and child health and well-being. Bradford is an ethnically diverse city (approximately, half of the births are to mothers of South Asian origin) with high levels of socioeconomic deprivation. The cohort is broadly representative of the city's maternal population of childbearing age.\n\nPatient and public involvement {#s2b}\n------------------------------\n\nThe BiB project emphasises the importance of involving parents and ensuring they are central to the research that is prioritised; what is important to the parents, how people find out the results from the research projects and what it means for their families. The participants were asked their views on many research topics including literacy levels, vision and the impact of vision on literacy. The participants suggested that these topics are of high importance and should be prioritised. The preliminary findings have been reported to the parents to provide verification of the data, ensuring that the findings reflect true patient experiences. Their ideas are essential in developing and revising current information provided to parents and carers. Their involvement has allowed the research to be prioritised around the needs and requirements of patients and carers. Finally in the dissemination of the research results, the parents will be central to publicising this study and its findings to local people, schools and the wider community.\n\nRecruitment {#s2c}\n-----------\n\nAs part of a BiB study, children's literacy levels on school entry (termed 'Reception Class' in England, UK and defined as year 1 of this study) were measured between September 2012 and July 2014 in Bradford schools. Two thousand nine hundred and thirty BiB children from 74 of the 123 primary schools (60%) participated. Of the 2930,\u00a0432 (14.7%) failed their vision screening ([figure 1](#F1){ref-type=\"fig\"}) and were referred for follow-up cycloplegic investigation, these children are defined as the treatment group. A further 512 BiB children from the same schools (randomly selected using Excel's random number generator) who had passed vision screening were also invited to participate and were defined as the comparison group, giving a total of 944 participants in the study. Consent was opt-out and parents received a letter via the schools requesting continued participation prior to each annual assessment. Of the 944,\u00a0893 (94.6%) consented to participate in year 2 and 650/944 (68.9%) participated in year 3 ([figure 1](#F1){ref-type=\"fig\"}).\n\n![Flow chart of the study participants. \\*Treatment group=children\u2009who failed vision screening and were referred for cycloplegic assessment. \\*\\*Adherent=prescribed\u2009spectacles worn at each visual acuity assessment. \\*\\*\\*Non-adherent=children\u2009who failed to attend cycloplegic examination and also children who attended but failed to wear prescribed spectacles at each visual acuity assessment.\u00a0\u2020Total\u2009number of eligible BiB children. \u2021All\u2009BiB children who failed vision screening and additionally had a literacy score measured during the same school term. \u00a7Random\u2009sample of BiB children who passed vision screening and additionally had a literacy score measured during the same school term.\u00a0BiB, Born\u2009in Bradford.](bmjopen-2017-021277f01){#F1}\n\nBaseline vision assessments:\u00a0year 1 {#s2d}\n-----------------------------------\n\nThe vision screening programme for children aged 4--5\u2009years in Bradford is conducted in the first year of school by orthoptists with 97% of eligible children being screened.[@R19] The screening includes standard protocols for measurement of monocular distance VA.[@R20] VA was measured at a distance of 3\u00a0m\u00a0using the LogMAR Crowded Test (Keeler, Windsor, UK) which has four letters per line, with each letter having a score of 0.025; the total score for each line thus represents 0.10 log unit (online\u00a0[supplementary information 1](#SP1){ref-type=\"supplementary-material\"}). A matching card was used and knowledge of letters was not therefore necessary to perform the test. VA was measured to threshold (ie, best achievable VA with no defined endpoint). In addition, cover test at 6\u2009m and 1/3\u2009m was performed. The data formed the baseline vision data (year 1). No child in the study was wearing spectacles at the baseline assessment.\n\n10.1136/bmjopen-2017-021277.supp1\n\nChildren failing to achieve the VA pass criterion (\\>0.20\u2009logMAR in one or both eyes) set by the UK NSC[@R3] or who had a strabismus detected on cover testing were referred for follow-up. The standard clinical pathway[@R4] following vision screening entailed referral to either to a community optometrist or the hospital eye service where a cycloplegic refraction (1% cyclopentolate hydrochloride) and fundus examination were undertaken, either by a paediatric ophthalmologist or an optometrist. Spectacles were prescribed based on the result of the cycloplegic refraction and clinical judgement; children were generally prescribed spectacles, including low degrees of hypermetropia (\\>+1.00\u2009DS to\u00a0+3.00\u2009DS), if they had a reduced VA. A follow-up appointment was then arranged with the orthoptist approximately 8\u2009weeks after the cycloplegic examination to repeat the VA measurement, with the child wearing spectacles if they had been prescribed. Children assessed by a community optometrist of their choice had the results of their examination returned to the hospital eye service and also had a follow-up appointment arranged with an orthoptist.\n\nAll VA testing, both at the point of vision screening and at follow-up, was performed using the same method of measurement. The results of the follow-up assessment, including cycloplegic refraction, VA with the prescribed glasses, cover testing and fundus and media examination, were extracted from the medical notes. The ophthalmic staff did not have knowledge of the baseline literacy assessment.\n\nBaseline literacy assessments: year 1 {#s2e}\n-------------------------------------\n\nLiteracy was measured on school entry (year 1) by trained research assistants within the same academic term as the vision screening. The research assistants were unaware of the VA results. An age-appropriate literacy measure, the Woodcock Reading Mastery Tests-Revised subtest: letter identification (ID), a validated reading skill test, was used to assess early literacy.[@R22] Letter-ID measures the child's ability to identify single letters, an essential skill mastered prior to reading and one of the best predictors of future reading achievement.[@R15] The letter-ID test is a test of knowledge of letters (the complete alphabet is used) and the child must verbally identify the name of each letter. This literacy measure specifically uses varied font type; the size of the letters approximate to 1.1 log unit (20/250) at 33\u2009cm, therefore, the performance on this test is not affected by the level of VA. Letter-ID was collected in both raw and age-standardised format. In addition, receptive vocabulary\u00a0was measured using the British Picture Vocabulary Scale (BPVS)[@R23] an indicator of cognitive ability, providing a representation of IQ in young children. This measure is included to adjust for potential confounding due to levels of general cognitive ability.\n\nFollow-up assessments: years 2 and 3 {#s2f}\n------------------------------------\n\nVision and literacy measures were repeated within the same school term approximately 12\u2009months (year 2) and 24\u2009months (year 3) after the baseline measurements. Both the vision and the literacy assessments were administered on the same day by the same personnel who were unaware of previous vision or literacy results. VA and literacy was measured as detailed above. VA found to be\u00a0\u22650.10\u2009logMAR was repeated with a pinhole and near VA was measured using the Bailey-Lovie near-vision chart[@R24] (online\u00a0[supplementary information 1](#SP1){ref-type=\"supplementary-material\"})\u00a0and whether the child was wearing spectacles was recorded. In order to present the real-life impact of adherence to spectacle wear, all VA measures reported are presenting VAs, that\u00a0is, measured with spectacles if worn at the time of the assessment in school. Parents and children were not given prior warning of these assessments.\n\nStatistical analysis {#s2g}\n--------------------\n\nChildren with baseline data for both vision and literacy in year 1 and who had at least one follow-up measure in either year 2 or year 3 were included in the final analysis ([figure 1](#F1){ref-type=\"fig\"}). The statistical model selected for the analyses, using projections over time, takes into account missing data and requires a minimum of measures at two time points. Using this type of statistical analysis allows inclusion of a greater number of participants giving maximum power to the analyses.[@R25] The characteristics of children participating in the study were compared initially using \u03c7^2^ test or two-sided t-tests as appropriate. Children in the treatment group were retrospectively divided into two subgroups, adherent and non-adherent. Adherence was defined as wearing prescribed spectacles at the time of assessment; otherwise, children were defined as non-adherent. Children who were assessed two\u00a0times but only wore the spectacles on one occasion were classed as non-adherent. A sensitivity analysis was conducted to assess the extent to which the results varied by changing the definition of adherence.\n\n### Analysis of VA {#s2g1}\n\nTo investigate the effect of spectacle wear over time on VA, multilevel longitudinal models[@R25] were first constructed with VA as the outcome measure for the child's better and worse eye. The models measure change within the individual and change between individuals over time and allow for individual differences in the rate of change over time.[@R25] A quadratic term was included to model the non-linear trajectory of change. The model also includes an interaction term to compare the relationship between age and group, to test whether differences by group are the same at different ages. Unadjusted analysis was initially undertaken with subsequent adjustment for demographic and socioeconomic factors reported in the literature to be associated with reduced VA: early-life factors[@R26] (gender, gestational age, birth weight, route of birth) and maternal factors[@R27] (ethnicity, mother's age at delivery, mother's level of educational attainment and being in receipt of means-tested benefits). Predicted outcomes were plotted to visualise group differences and change in the outcomes for each group over time.\n\n### Analysis of literacy {#s2g2}\n\nIn order to estimate the association between the letter-ID and VA, the same multilevel and longitudinal modelling approach was adopted, but with the final letter-ID score as the outcome measure. The raw letter-ID scores were used in the analysis in order to explore change over time. After estimating differences between the groups and accounting for the initial letter-ID at baseline (year 1), further adjustment was undertaken for the factors reported in the literature to be associated with educational attainment,[@R28] the early-life factors and maternal factors as stated above. Spherical equivalent refraction (SER) (sphere plus half cylinder) of the better eye was included as was BPVS score in order to account for cognitive ability. The results of these models are presented along with predicted outcomes for each of the groups. Effect sizes are generally reported when appraising educational interventions. To demonstrate group differences at each time point, effect sizes were calculated for the letter-ID scores using Cohen's d.[@R30]\n\n### VA: year 3 {#s2g3}\n\nChildren were unable to accurately perform the near VA (logMAR) test until year 3; we are therefore unable to provide a longitudinal analysis. In year 3, we have measures of both near VA and distance VA and present the correlation between the near and distance VA at this time point only. Additionally, we analysed the\u00a0association between near VA and literacy to examine if the results differed from the association between distance VA and literacy\u00a0in year 3 only.\n\nAll analyses were carried out using Stata V.13 (StataCorp).\n\nResults {#s3}\n=======\n\nData from 801 (85%) children from 67 schools were included in the final analysis ([figure 1](#F1){ref-type=\"fig\"}). Twelve children in the treatment group were excluded from the analysis as they had ocular conditions other than refractive error (eg, nystagmus) confirmed in their medical notes, leaving 368 children in the treatment group and 433 in the comparison group. Of 368, 230 (62.5%) of children in the treatment group had attended for the initial cycloplegic examination and been prescribed spectacles, 3/368 (0.8%) attended but no cycloplegic refraction information was available, 23/368 (6.3%) had been prescribed spectacles but had not returned for follow-up VA assessment and 112/368 (30.4%) had failed to attend any appointment following vision screening. Of the 253 children in the treatment group with cycloplegic refraction results, 157/253 (62.1%) had astigmatism (\\>1.00\u2009DC) either alone (n=19) or in combination with hypermetropia (\\>+3.0\u2009DS) (n=56), low hypermetropia (\\>+1.0\u2009DS to\u00a0+3.0\u2009DS) (n=16) or myopia (\u2264\u22120.50\u2009DS) (n=66). Of 253, 35 (13.8%) had hypermetropia alone, 11 (4.3%) had myopia alone and 50 (19.8%) children had low hypermetropia. Of 253,\u00a055 (21.7%) additionally had anisometropia (\u22651.0\u00a0D difference). For those children with a cycloplegic refraction result ([table 1](#T1){ref-type=\"table\"}) the SER ranged from \u22127.875 to\u00a0+7.50\u00a0D in the better eye and \u22128.25 to\u00a0+7.50\u00a0D in the worse eye. Of the 368,\u00a014 (3.8%) children had a constant or intermittent strabismus, 5 of whom had been prescribed occlusion therapy for amblyopia at follow-up after vision screening. Those children were not excluded from the analysis as they met the initial VA referral criteria and had been prescribed spectacles.\n\n###### \n\nCharacteristics of Born in Bradford children and mothers included in the analyses\n\n                                               Comparison group n=433   Treatment group n=368   P values\\*\n  -------------------------------------------- ------------------------ ----------------------- ------------\n  Children                                                                                      \n  \u2003Age (months) year 1                         60 (4.2)                 60 (4.5)                0.119\n  \u2003Gender                                                                                       \n  \u2003\u2003Male                                       229 (51.1)               183 (49.7)              \n  \u2003\u2003Female                                     219 (48.9)               185 (50.3)              0.693\n  \u2003Ethnicity                                                                                    \n  \u2003\u2003White                                      125 (28.0)               91 (24.9)               \n  \u2003\u2003Pakistani                                  262 (58.7)               232 (63.4)              \n  \u2003\u2003Other                                      59 (13.3)                43 (11.7)               0.403\n  \u2003Route of birth                                                                               \n  \u2003\u2003Vaginal                                    342 (77.0)               291 (79.7)              \n  \u2003\u2003Caesarean                                  102 (23.0)               74 (20.3)               0.355\n  \u2003Gestational age at birth (weeks)            277 (12.0)               276 (13.0)              0.158\n  \u2003Birth weight (g)                            3184 (550.0)             3128 (573.0)            0.155\n  \u2003VA better eye                               0.113 (0.049)            0.271 (0.138)           \\<0.001\n  \u2003VA worse eye                                0.135 (0.046)            0.428 (0.189)           \\<0.001\n  \u2003SER better eye\u2020                             --                       1.19 (0.95)             --\n  \u2003SER worse eye\u2020                              --                       1.98 (1.27)             --\n  Mother                                                                                        \n  \u2003Age (years)                                 27.3 (5.4)               28.1 (5.7)              \\<0.001\n  \u2003Mother's education                                                                           \n  \u2003\u2003\\<A -level                                 227 (64.5)               190 (69.3)              \n  \u2003\u2003A-level or above                           125 (35.5)               84 (31.7)               0.201\n  \u2003In receipt of means-tested benefits (yes)   163 (45.0)               144 (50.1)              0.139\n\nValues are numbers (%) or mean (SD).\n\nVAs are measured in logMAR; therefore, higher values represent poorer VA.\n\n\\*Difference between comparison and treatment groups (\u03c7^2^ or t-test as appropriate).\n\n\u2020Cycloplegic results were available for the treatment group only.\n\nlogMAR,\u00a0logarithm of the minimum angle of resolution; SER, spherical equivalent refraction; VA, visual acuity.\n\nBaseline (year 1) characteristics of the children in the comparison and treatment groups are shown in [table 1](#T1){ref-type=\"table\"}. A small mean difference (\u22120.021 logMAR, 95%\u2009CI \u22120.022 to \u22120.020) in VA between the eyes of the comparison group was found, equating to one letter difference. This is not clinically significant but is statistically significant, therefore, VAs are presented for the better and worse eye separately. Higher levels of VA were found in both eyes of the comparison group compared with the treatment group (\u03c7^2^ p\\<0.001) ([table 1](#T1){ref-type=\"table\"}). The only demographic factor found to differ between the comparison and the treatment group was the average mother's age which was around 10\u2009months more in the treatment group (\u03c7^2^ p\\<0.001).\n\n[Table 2](#T2){ref-type=\"table\"} presents the baseline (year 1) characteristics of those children in the treatment group retrospectively categorised as adherent (173/368, 47.0%) and non-adherent (195/368, 53.0%) ([figure 1](#F1){ref-type=\"fig\"}). In the non-adherent group, no child wore spectacles at their year 2 assessment and 39/195 (20%) wore them in year 3 only. At baseline, the group subsequently classed as adherent had a lower level of VA compared with the non-adherent group in both the better and worse eye ([table 2](#T2){ref-type=\"table\"}). The only other factor that differed between the adherent and the non-adherent groups was the mother's level of education with 50/173 (39.1%) of adherent children having mothers educated to A-level or above compared with only 34/195 (23.3%) of the non-adherent group (\u03c7^2^ p=0.005). BPVS did not differ between the adherent and non-adherent children (p=0.553) suggesting no difference in cognitive ability.\n\n###### \n\nBaseline characteristics of participants in the treatment group retrospectively classed as adherent and non-adherent\n\n                                               Adherent n=173 (47.0%)   Non-adherent n=195 (53.0%)   P values\\*\n  -------------------------------------------- ------------------------ ---------------------------- ------------\n  Children                                                                                           \n  \u2003Age (months) year 1                         59.4 (4.5)               59.6 (4.5)                   0.850\n  \u2003Gender                                                                                            \n  \u2003\u2003Male                                       81 (46.8)                102 (52.3)                   \n  \u2003\u2003Female                                     92 (53.2)                93 (47.7)                    0.293\n  \u2003Ethnicity                                                                                         \n  \u2003\u2003White                                      48 (27.9)                43 (22.2)                    \n  \u2003\u2003Pakistani                                  103 (59.9)               129 (66.5)                   \n  \u2003\u2003Other                                      21 (12.2)                22 (11.3)                    0.387\n  \u2003Route of birth                                                                                    \n  \u2003\u2003Vaginal                                    137 (79.6)               154 (79.8)                   \n  \u2003\u2003Caesarean                                  35 (20.4)                39 (20.2)                    0.973\n  \u2003Gestational age at birth (weeks)            276 (13.0)               275 (14.0)                   0.383\n  \u2003Birth weight (g)                            3121 (569.0)             3134 (579.0)                 0.833\n  \u2003VA better eye\u2020                              0.292 (0.150)            0.256 (0.129)                0.008\n  \u2003VA worse eye\u2020                               0.465 (0.197)            0.399 (0.175)                0.001\n  \u2003SER better eye                              1.18 (0.86)              1.20 (1.02)                  0.960\n  \u2003SER worse eye                               2.02 (1.20)              1.96 (1.33)                  0.657\n  \u2003Language ability scores\u2021                    97.8 (15.6)              96.8 (16.4)                  0.553\n  Mother                                                                                             \n  \u2003Age (years)                                 28.1 (5.8)               28.0 (5.7)                   0.845\n  \u2003Mother's education                                                                                \n  \u2003\u2003\\<A -level                                 78 (60.9)                112 (76.7)                   \n  \u2003\u2003A-level or above                           50 (39.1)                34 (23.3)                    0.005\n  \u2003In receipt of means-tested benefits (yes)   61 (45.5)                83 (55.7)                    0.087\n\nValues are numbers (%) or mean (SD).\n\nVAs are measured in logMAR; therefore, higher values represent poorer VA.\n\n\\*Difference between adherent and non-adherent treatment groups (\u03c7^2^ or t-test as appropriate).\n\n\u2020No child was wearing spectacles at the baseline assessment.\n\n\u2021Age-adjusted language ability measure for British Picture Vocabulary Scale.\n\nlogMAR, logarithm of the minimum angle of resolution; SER, spherical equivalent refraction;\u00a0VA, visual acuity.\n\nVisual acuity {#s3a}\n-------------\n\nAt baseline compared with the comparison group both the adherent (mean difference: 0.337 logMAR; 95%\u2009CI 0.304 to 0.370) and non-adherent groups (mean difference: 0.273 logMAR; 95%\u2009CI 0.241 to 0.305) had lower levels of VA in the worse eye. [Table 3](#T3){ref-type=\"table\"} and [figure 2](#F2){ref-type=\"fig\"} present the VA trajectories over the course of the study. These show that after adjusting for previously described early-life and maternal variables, the VA of both eyes for all three groups; the comparison, the adherent and the non-adherent groups improve over time.\n\n![Projected visual acuity (logMAR) trajectory (with 95% CIs) by group over time (child's age in months) for the better and worse eye, fully adjusted for all early-life and maternal covariates.\u00a0logMAR, logarithm of the minimum angle of resolution.](bmjopen-2017-021277f02){#F2}\n\n###### \n\nChange in visual acuity for the better and worse eye over time by group: comparison, adherent and non-adherent\n\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n                                  Unadjusted (worse eye)\\              Adjusted\u2020 (worse eye)\\               Unadjusted (better eye)\\             Adjusted\u2020 (better eye)\\\n                                  (95%\u2009CI)                             (95%\u2009CI)                             (95%\u2009CI)                             (95%\u2009CI)\n  ------------------------------- ------------------------------------ ------------------------------------ ------------------------------------ ------------------------------------\n  Constant                        0.177 (0.159 to 0.194)\\*             0.386 (0.124 to 0.648)\\*             0.240 (0.026 to 0.454)\\*             0.240 (0.026 to 0.454)\\*\n\n  Age (months)                    \u22120.009 (\u22120.011 to \u22120.008)\\*\\*\\*      \u22120.009 (\u22120.011 to \u22120.007)\\*\\*\\*      \u22120.006 (\u22120.008 to \u22120.005)\\*\\*\\*      \u22120.006 (\u22120.008 to \u22120.005)\\*\\*\\*\n\n  Age (months) squared            0.00016 (0.00012 to 0.00020)\\*\\*\\*   0.00016 (0.00012 to 0.00021)\\*\\*\\*   0.00010 (0.00007 to 0.00013)\\*\\*\\*   0.00010 (0.00006 to 0.00014)\\*\\*\\*\n\n  Group (reference: comparison)                                                                                                                  \n\n  Adherent                        0.337 (0.309 to 0.366)\\*\\*\\*         0.337 (0.304 to 0.370)\\*\\*\\*         0.184 (0.162 to 0.106)\\*\\*\\*         0.170 (0.144 to 0.196)\\*\\*\\*\n\n  Non-adherent                    0.277 (0.250 to 0.305)\\*\\*\\*         0.273 (0.241 to 0.305)\\*\\*\\*         0.150 (0.128 to 0.172)\\*\\*\\*         0.148 (0.123 to 0.174)\\*\\*\\*\n\n  Age\u00a0\u00d7 group interaction                                                                                                                        \n\n  Age \u00d7 adherent                  \u22120.008 (\u22120.009 to \u22120.007)\\*\\*\\*      \u22120.008 (\u22120.009 to \u22120.007)\\*\\*\\*      \u22120.004 (\u22120.005 to \u22120.004)\\*\\*\\*      \u22120.004 (\u22120.005 to \u22120.003)\\*\\*\\*\n\n  Age \u00d7 non-adherent              \u22120.003 (\u22120.003 to \u22120.001)\\*\\*\\*      \u22120.003 (\u22120.004 to \u22120.001)\\*\\*\\*      \u22120.001 (\u22120.002 to 0.000)             \u22120.002 (\u22120.004 to 0.000)\n  -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n\n\u00d7 : interaction between group and age to determine if the effect of being in a particular group changes with age. The total effect for any one group is the coefficient for age plus the additional effect of age for that group.\n\n\\*P\\<0.05, \\*\\*P\\<0.01, \\*\\*\\*P\\<0.001.\n\n\u2020Model adjusted for gender, ethnicity, gestation period, birth weight, birth route, maternal education status, maternal age and means-tested benefit status.\n\nThe VA of all children improved with increasing age, \u22120.009 log units per month (95%\u2009CI \u22120.011 to \u22120.007) (worse eye) and \u22120.006 log units per month (\u22120.008 to \u22120.005) (better eye) ([table 3](#T3){ref-type=\"table\"}).\n\nOver and above this improvement, the adherent group (worse eye) improved by a further \u22120.008 log units per month (95%\u2009CI \u22120.009 to \u22120.007). The adherent children, therefore, improved overall by \u22120.017 (95% CI \u22120.020 to --0.015) log units per month (approximately, two letters every 3\u2009months) and also demonstrated a small amount of improvement in the better eye above that expected from age ([table 3](#T3){ref-type=\"table\"}).\n\nThe non-adherent group (worse eye) improved by \u22120.003 log units per month (95%\u2009CI \u22120.004 to \u22120.001) above that expected from age. The non-adherent children, therefore, improved overall by \u22120.012 log units per month (95%\u2009CI \u22120.014 to \u22120.010). No additional improvement above that expected from age was demonstrated in the better eye ([table 3](#T3){ref-type=\"table\"}).\n\nSensitivity analysis redefining the classification of adherence did not materially affect the results.\n\nLiteracy {#s3b}\n--------\n\nThe unadjusted model shows the final letter-ID score reduces by \u22120.9 units (95%\u2009CI\u00a0\u22121.15 to \u22120.64) for every one line (0.10 logMAR) fall in VA of the better eye ([table 4](#T4){ref-type=\"table\"}). This association persists but is weaker after fully adjusting for the socioeconomic and demographic factors, the letter-ID score declines by \u22120.327 units (95%\u2009CI\u00a0\u22120.540 to \u22120.115) for every one line fall in VA. Separate adjusted analysis of the VA level of the worse eye shows similar results but with weaker association, letter-ID score declines by \u22120.260 units (95%\u2009CI\u00a0\u22120.414 to \u22120.105) for every one line fall in VA.\n\n###### \n\nAssociations between letter-ID score, visual acuity (better eye), maternal and early-life factors\n\n  ---------------------------------------------------------------------------------------------------------------\n  Factor                               Unadjusted model\\        P values   Fully adjusted model\\       P values\n                                       (95%\u2009CI)                            (95%\u2009CI)                    \n  ------------------------------------ ------------------------ ---------- --------------------------- ----------\n  Constant                             18.82 (17.91 to 19.73)   \\<0.001    \u221220.6 (\u221228.2 to \u221213.0)      \\<0.001\n\n  Age                                  1.30 (1.21 to 1.38)      \\<0.001    1.28 (1.19 to 1.37)         \\<0.001\n\n  Age squared                          \u22120.02 (\u22120.02 to \u22120.02)   \\<0.001    \u22120.020 (\u22120.022 to \u22120.017)   \\<0.001\n\n  Visual acuity: change in letter-ID   \u22120.90 (\u22121.15 to \u22120.64)   \\<0.001    \u22120.327 (\u22120.540 to \u22120.115)   0.003\n\n  Per 0.1\u00a0log unit (one line)                                                                          \n\n  Letter-ID baseline (year 1)                                              0.348 (0.326 to 0.371)      \\<0.001\n\n  BPVS                                                                     0.019 (\u22120.001 to 0.039)     0.064\n\n  Ethnicity                                                                                            \n\n  Pakistani heritage                                                       0.668 (\u22120.016 to 1.353)     0.056\n\n  Other                                                                    1.174 (1.159 to 2.189)      0.023\n\n  Gender                                                                                               \n\n  Female                                                                   0.471 (\u22120.093 to 1.035)     0.102\n\n  Birth weight (per 100\u2009g)                                                 0.074 (0.008 to 0.141)      0.029\n\n  Gestational age (weeks)                                                  \u22120.053 (\u22120.257 to 0.151)    0.611\n\n  Receiving benefits                                                       \u22120.086 (\u22120.661 to 0.4990)   0.770\n\n  Mothers level of education           (higher than A-level)               0.765 (0.156 to 1.374)      0.014\n\n  Mothers age at birth (years)                                             \u22120.048 (\u22120.100 to 0.005)    0.075\n  ---------------------------------------------------------------------------------------------------------------\n\nBPVS,\u00a0British\u2009Picture Vocabulary Scale (baseline standardised score); ID, identification.\n\nChildren of mothers educated to A-level or above had increased letter-ID scores (0.765 units; 95%\u2009CI 0.156 to 1.374) compared with those with lower qualifications. Ethnicity other than white British or Pakistani heritage was associated with better letter-ID score, which might reflect the higher number of mothers educated to above A-level in this group. Greater birth weight was also associated with increased letter-ID score ([table 4](#T4){ref-type=\"table\"}). Adjustment for SER made no difference and was not associated with letter-ID (p=0.306). It was therefore not included in the models. Similarly, subsequent analysis replacing SER with refractive error categories did not show an association with letter-ID (online\u00a0[supplementary information 2](#SP2){ref-type=\"supplementary-material\"}).\n\n10.1136/bmjopen-2017-021277.supp2\n\nA predictive model of the letter-ID score over time for children in each group ([figure 3](#F3){ref-type=\"fig\"}) was constructed using both the unadjusted and adjusted data from the VA trajectories ([table 3](#T3){ref-type=\"table\"}) and incorporated into the model reporting letter-ID ([table 4](#T4){ref-type=\"table\"}). The unadjusted trajectory shows both adherent and non-adherent groups at baseline have lower letter-ID scores than the comparison group. The predicted trajectory of improvement in the adherent group is greater than the non-adherent group with the later letter-ID scores of the adherent group converging on those of the comparison group by year 3. The non-adherent group although improving over time does not catch up with the adherent or the comparison groups. After adjusting for socioeconomic and demographic variables, the trend is similar but with a smaller difference between the groups.\n\n![Predicted letter-ID scores over time (child's age in months) based on the trajectories of the visual acuity (adjusted model) of the better eye. The adjusted model includes all early-life and maternal covariates for the comparison, adherent and non-adherent groups.\u00a0ID,\u00a0identification.](bmjopen-2017-021277f03){#F3}\n\n[Table 5](#T5){ref-type=\"table\"} presents the effect size of wearing spectacles on the letter-ID scores between the groups annually over the 3\u2009years of the study. Comparing the letter-ID scores between the adherent and the non-adherent group, a gradual increase in the effect size over time is demonstrated with the greatest effect size (0.11) between the adherent and non-adherent groups shown in year 3.\n\n###### \n\nAnnual Literacy Scores by group\n\n  ------------------------------------------------------------------------------------\n  Year   Group          Letter-ID\\    Comparison groups                Effect size\\\n                        (raw score)                                    (Cohen's d)\\*\n  ------ -------------- ------------- -------------------------------- ---------------\n  1      Comparison     25            Comparison versus\u00a0adherent       0.06\n\n         Adherent       24.3          Comparison versus\u00a0non-adherent   0.06\n\n         Non-adherent   24.3          Adherent versus\u00a0non-Adherent     0.00\u2020\n\n  2      Comparison     34.7          Comparison versus\u00a0adherent       0.05\n\n         Adherent       34.4          Comparison versus\u00a0non-adherent   0.13\n\n         Non-adherent   34.0          Adherent versus\u00a0non-Adherent     0.07\n\n  3      Comparison     39.1          Comparison versus\u00a0adherent       0.08\n\n         Adherent       38.8          Comparison versus\u00a0non-adherent   0.18\n\n         Non-adherent   38.4          Adherent versus\u00a0non-Adherent     0.11\n  ------------------------------------------------------------------------------------\n\nSD 10.9 at year 1, 5.6 at year 2 and 3.8 at year 3.\n\n\\*Based on group difference divided by the pooled SD of letter-ID score.\n\n\u2020In\u00a0year 1, there is no difference as spectacle wear has not commenced.\n\nID,\u00a0identification.\n\n### VA: year 3 {#s3b1}\n\nThe results demonstrate a statistically significant correlation between near and distance VA in year 3 (right eye, r=0.663\u2009and left eye, r=0.642) (online\u00a0[supplementary information 3](#SP3){ref-type=\"supplementary-material\"}). In addition, the associations between the near VA and literacy score and distance VA and literacy score are approximately the same (online\u00a0[supplementary information 4](#SP4){ref-type=\"supplementary-material\"}).\n\n10.1136/bmjopen-2017-021277.supp3\n\n10.1136/bmjopen-2017-021277.supp4\n\nDiscussion {#s4}\n==========\n\nThis is the first longitudinal study to assess the effect of adherence/non-adherence to spectacle wear on VA and literacy in children following vision screening. The VA of children who adhered to spectacle wear was found to improve at a far greater rate compared with those who were non-adherent, with the VA of adherent children reaching similar levels to the VA of the comparison children by the end of the study. Our results further indicate that early developing literacy is affected by the level of VA even after adjusting for socioeconomic and demographic factors associated with educational attainment. The letter-ID score declines by approximately 1.5% for every one line of reduction in VA. In this and similar populations,[@R14] where children have been reported to have reduced VA levels (\\>0.30\u2009logMAR in better eye), there is likely to be an impact on developing literacy skills. The effect size (0.11) of being adherent to spectacle wear compared with non-adherence in year 3 of our study is the same as that reported in a Chinese study providing free spectacles to children[@R32] and is comparable with reported educational interventions.[@R33] Thus, children who fail vision screening and adhere to spectacle wear have the potential to improve their VA, further influencing early literacy development.\n\nAdherence to spectacle wear is highly influenced by socioeconomic and demographic factors, particularly maternal education, a factor that is also known to be associated with educational attainment.[@R34] Children with reduced VA and who are in less educated families are less likely to adhere to treatment, which will further impact on their educational attainment and future life chances. We were, however, able to adjust for the many associated maternal and early-years factors, the value of embedding this study within a birth cohort. A study examining academic performance in US schools reports that failing vision screening was predictive of being in the lowest quartile of academic performance.[@R35] Conversely, a longitudinal study of children aged 9--10\u2009years in Singapore, Dirani *et al* [@R36] found VA did not play a significant role in predicting academic performance. However,\u00a0the children were older, mainly myopic and only a small number of participants had decreased VA which may account for the difference in their findings relative to ours.\n\nThe VA of children in all groups (adherent, non-adherent and comparison group) continued to improve throughout this study. The improvement in VA found in the comparison group is similar to that reported for normal visual development, with optimum VA achieved around 6\u2009years of age.[@R37] The improvement in VA of the worse eye found in adherent children over the time of the study was significantly greater than that expected solely from visual development[@R39] or indeed from retest variability[@R40] and was almost double that of the comparison group. Little additional improvement above that expected from visual development was demonstrated in the worse eye of the non-adherent children, an indication that the improvement in the adherent children is not due to regression to the mean. The longitudinal observation of the children demonstrates improvement in VA and in literacy, with the non-adherent group demonstrating persistently lower literacy scores throughout the study, although the effect is attenuated after adjusting for other factors. Annual improvement in academic achievement is well recognised and is particularly notable in the early years of schooling with the initial improvement thought to be associated with the effect of entering school, combined with rapid early child development followed by a plateau in academic growth as children progress through school grades.[@R20]\n\nEarly literacy development is complex and associated with socioeconomic and demographic factors, in particular maternal education. However, even after taking these factors into account VA continues to be associated with literacy; the poorer the level of VA, the greater the reduction in the literacy score. In a Singaporean study,[@R39] a strong association between paternal level of education and academic school performance was reported. As one might expect, higher levels of maternal education have a positive impact on literacy.[@R41] In addition, mothers with higher educational attainment are more likely to effectively access health services, and are more likely to adhere to prescribed treatment.[@R43]\n\nOur study shows an association between VA and literacy score but no association between SER and literacy. Neither did further analysis by refractive error types indicate an association with literacy, this is most likely related to a lack of power due to the small numbers when refractive error is categorised in our study. Our findings differ from previous studies reporting an association between refractive error and literacy.[@R11]\n\nHypermetropia has been reported to be associated with poor literacy. A large cross-sectional American study (vision in preschoolers - hyperopia in preschoolers (VIP-HIP)) of preschool children aged 4--5\u2009years found that children with uncorrected hypermetropia in conjunction with reduced binocular near VA (worse than 20/40) have poorer literacy than those with hypermetropia and a good level of binocular near VA.[@R12] The VIP-HIP study reports that the level of binocular near VA was predictive of literacy scores; with hypermetropic children with binocular near VA better than 20/40, demonstrating literacy scores similar to those children who were emmetropic. Although the VIP-HIP study does not report distance VA levels of the children, it does state that the analysis of the distance VA resulted in similar findings, an indication that distance VA levels may also influence early literacy scores.\n\nAstigmatism has also been reported to be associated with reduced literacy. In native American children bilateral uncorrected astigmatism (\u22651.00\u2009DC) has been reported to reduce reading fluency, and children with moderate astigmatism are reported to have lower VA and fluency than those with no or low astigmatism.[@R11] The findings reported from both the above studies may indicate that moderate to high degrees of uncorrected hypermetropia or astigmatism which reduce VA is associated with a reduction in literacy scores.\n\nClassroom-based tasks where fixation frequently changes are reported to require high levels of distance VA (0.33\u00a0logMAR) and slightly lesser levels of near VA (0.72\u00a0logMAR),[@R44] this is most probably due to print size for early readers being enlarged. We would suggest therefore that where VA is reduced beyond that required in the learning environment, it will impact on a child's developing literacy and hence the association we report between distance VA and literacy.\n\nThe longitudinal design of this study provides an insight into development of VA and literacy in the early years of schooling, and the use of linked data from the mothers and children participating in the BiB cohort study permitted the many potential confounding factors associated with educational attainment to be accounted for. We include children with a wide range of refractive error and VAs allowing a robust analysis of the influence of both factors on developing literacy. The study does however have some weaknesses. It is not a randomised controlled trial and non-adherence was defined retrospectively by the failure of the child to wear their prescribed glasses at one assessment; it is possible that this was a unique event and is not representative of the child's true adherence to spectacle wear over the course of the study. If this is indeed the case, then the random misclassification is likely to underestimate the difference found between the adherent and non-adherent groups.[@R45] In addition, the sensitivity analysis redefining non-adherence does not demonstrate any material difference in the results.\n\nA cycloplegic examination was not undertaken for all children and there will be some children with reduced vision who were not identified at screening (false negatives). No child who had a cycloplegic refraction was found to be a false positive but a proportion of the children who failed to attend for the cycloplegic examination may be false positives. This misclassification will similarly be random, underestimating the size of estimates of effect and suggests our estimates may be conservative.[@R45]\n\nVA is the sole measure of visual function reported from the study and it is possible other measures of visual function are also associated with academic performance; further research would be required to explore these associations. The VA assessment and the literacy test are both letter based and children who struggle with letter-ID may also demonstrate a poor ability with the VA test. However, all children used a matching technique, a skill that is present in children as young as 3\u2009years[@R46] and no child who failed the screening was classed as false positive.\n\nDuring visual maturation, the presence of neurodevelopmental disorders such as refractive error, and strabismus may contribute to a reduction in VA and early intervention is required. This study demonstrates that wearing spectacles is an effective intervention to improve VA, and that this will impact positively on developing literacy. The children who do not adhere to spectacle wear are likely to be those in families who are less well educated. Further research is required to better understand the reasons for non-adherence and evaluate interventions to promote adherence to spectacle wear. This has the potential to improve vision and\u00a0support future life chances in children who may already face educational disadvantage.\n\nSupplementary Material\n======================\n\n###### Reviewer comments\n\n###### Author\\'s manuscript\n\nWe thank all the families and schools who took part in this study, the orthoptists from Bradford Teaching Hospitals Foundation Trust who conducted the vision screening programme, the researchers from the Starting Schools programme who collected the literacy measures, Patrick Friis, Alexandra Morris and Hannah Farrugia who collected follow-up measures and the Data Support Team from Bradford Institute for Health Research who created and maintain the data linkage system.\n\n**Contributors:** AB initiated the project, designed data collection, monitored data collection for the whole study, wrote the statistical analysis plan, cleaned and analysed the data, and drafted and revised the paper. She is the guarantor. BK wrote the statistical analysis plan, cleaned the data and revised the draft paper. BC initiated the project and revised the draft paper. BTB contributed to the design of the study and revised the draft paper. MB contributed to the design of the study and revised the draft paper. JB contributed to the design of the study and revised the draft paper. TAS initiated the project, wrote the statistical analysis plan and revised the draft paper.\n\n**Funding:** AB is funded by a National Institute for Health Research Post-Doctoral Fellowship Award (PDF-2013-06-050). The Born in Bradford study presents independent research commissioned by the National Institute for Health Research Collaboration for Applied Health Research and Care (NIHR CLAHRC) and the Programme Grants for Applied Research funding scheme (RP-PG-0407-10044).\n\n**Disclaimer:** The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.\n\n**Competing interests:** None declared.\n\n**Patient consent:** Not required.\n\n**Ethics approval:** This study was approved by National Research Ethics Committee Yorkshire and the Humber-South Yorkshire (Ref 13/YH/0379).\n\n**Provenance and peer review:** Not commissioned; externally peer reviewed.\n\n**Data sharing statement:** No additional data are available.\n"}
+{"text": "Introduction\n============\n\nCardiovascular disease (CVD) is one of the main causes of morbidity and mortality worldwide.[@B1] According to 2013 death statistics in Korea, CVD was the second leading cause of death.[@B2] Previous studies have declared that dyslipidemia is an important risk factor for CVD by contributing to the initiation and progression of atherosclerosis; therefore, its management is important for reducing the burden of CVD.[@B3]\n\nMeanwhile, the prevalence of dyslipidemia is steadily rising in Korea, and statistics published by the Korean Society of Lipidology and Atherosclerosis (KSLA) revealed that in 2013 57.6% of men and 38.3% of women, accounting for a total of 47.8% of people over 30 years of age (more than 16 million), had dyslipidemia. When the low-density lipoprotein cholesterol (LDLC) cutoff value was set to 100 mg/dL for diabetic patients, 9 out of every 10 diabetic adults had dyslipidemia. For patients with hypertension, 2 out of every 3 hypertensive adults were diagnosed with dyslipidemia.[@B4]\n\nThe use of lipid-lowering medications such as statin and fibrate are important for the treatment of dyslipidemia, especially in high-risk patients. However, lifestyle interventions are also important for managing dyslipidemia and are considered initially after diagnosis, since dietary factors can influence lipid levels and regular exercise improves lipid profiles, while smoking has been known to have a detrimental effect.[@B5] Villegas et al.[@B6] reported that a combination of protective factors including normal body mass index (BMI), never smoking, light alcohol consumption, prudent diet and regular physical activity was associated with a significantly lower prevalence of dyslipidemia. Pharmacological therapy requires diagnosis by a doctor, and patients may be initially reluctant to take daily medication, making lifestyle interventions an attractive first choice for intervention. In fact, patients may be more compliant to behavioral modification and behavioral factors have the advantage that intervention for prevention of chronic diseases and lowering of cardiovascular risks can be emphasized at the population level. The Nurses\\' Health Study cohort was able to conclude that adherence to lifestyle guidelines for diet, exercise, and smoking was associated with a very low risk for coronary heart disease.[@B7]\n\nPrevious reports have studied the prevalence, awareness, and treatment of dyslipidemia in Korea. Although awareness and treatment rates are slowly rising, their rates are still low (13.7% and 7.4% in 2010 compared to 6.1% and 1.9% in 2005, respectively).[@B8] As for lifestyle interventions, there is a lack of studies examining adherence rates to clinical guidelines for lifestyle therapy in those with dyslipidemia, especially comparing those who are aware and those who are not aware of their diagnosis.\n\nLack of awareness about dyslipidemia may act as an additional barrier to adequate health behavior. However, whether awareness of dyslipidemia affects health behavior is not clear. In a study by Kitagawa et al., patients with high awareness of their health status showed a positive attitude towards diet and exercise as lipid-lowering treatment, and high adherence to drug therapy. However, subjects were limited to high-risk patients on prescription for pravastatin making it difficult to apply the results to the general population.[@B9] Another study conducted in a Chinese province that analyzed dyslipidemia awareness and influencing factors demonstrated that awareness was associated with a lower odds ratio (OR) for drinking (OR=0.78) and physical activity (OR=0.714), and with higher OR for increased BMI (OR=1.547), age (increasing OR for older age groups), education (higher OR for higher level of education), and family history of dyslipidemia (OR=3.62).[@B10] However, criteria for the factors mentioned were looser than clinical guidelines: drinking any kind of alcohol more than once a week, obesity as BMI \u226524 kg/m^2^, and exercising not less than once a week.\n\nNo other studies analyzing awareness of dyslipidemia and associated factors were found by the authors. Therefore, this study compared adherence to health behavior according to clinical guidelines between dyslipidemia subjects who were aware of their diagnosis and those who were not, with higher adherence expected in the awareness group.\n\nMethods\n=======\n\n1. Study Population\n-------------------\n\nThis study was performed using data from the fifth Korean National Health and Nutrition Survey 2010--2012 (KNHANES V). The KNHANES is a nationally representative study conducted regularly by the Korea Centers for Disease Control and Prevention to assess the health and nutritional status of non-institutionalized civilians in Korea. It uses a multi-stage probability sample design and trained interviewers to administer questionnaires to participants. The KNHANES V is based on data from 3,800 households in 576 randomly selected survey areas, with the number selected in proportion to the size of each area.[@B11] For the selection of dyslipidemia patients, from a total of 25,334 individuals that participated in the KNHANES V, we excluded those aged \\<20 years (n=6,140), those who did not provide an answer for the assessment of awareness of dyslipidemia (n=1,518), whose lipid profiles were not assessed (n=959), and those who did not fast for over 9 hours (n=485). Those without available data for adequate physical activity, BMI, self-perceived health status, education, household income, and nutrient intake (n=2,115) were also excluded in the final analysis. The present study received an exemption from informed consent by the institutional review board committee of the Seoul National University Hospital because this study used public data provided by the KNHANES (IRB No. X-1602-333-903).\n\n2. Measurement\n--------------\n\nBlood samples were collected from each participant the morning after fasting for at least 8 hours. Samples were processed, refrigerated immediately, and transported to the central laboratory to be analyzed within 24 hours after transportation. Total cholesterol (TC), high-density lipoprotein cholesterol (HDLC), triglycerides (TG), and fasting glucose levels were assessed using a Hitachi 7600 automatic chemistry analyzer (Hitachi, Tokyo, Japan). Direct LDLC measurements that were assessed by the automatic analyzer were used when available (participants eligible for sampling of heavy metal levels or with TG levels \u2265200 mg/dL), and when unavailable, the Friedwald formula was used if the TG level was less than 400 mg/dL. Blood pressure was measured using a mercury sphygmomanometer (Baumanometer; WA Baum Co. Inc., Copiague, NY, USA) 3 times consecutively on the right arm with the participant in a seated position after at least 5 minutes of rest. The final blood pressure was obtained by averaging the second and third blood pressure measurements.[@B12]\n\n3. Definitions\n--------------\n\nDyslipidemia was defined by levels of TC (\u2265240 mg/dL), HDLC (\\<40 mg/dL), TG (\u2265200 mg/dL), and LDLC (\u2265160 mg/dL, \u2265130 mg/dL, or \u2265100 mg/dL according to risk category) based on the National Cholesterol Education Program-Adult Treatment Panel III (NCEP-ATP III) guidelines,[@B13] and whether participants answered that they were receiving treatment for dyslipidemia or taking medication to lower cholesterol levels.\n\nDyslipidemia awareness was assessed by the questions, \"Have you been diagnosed with dyslipidemia by a doctor?\\\" and \"Are you currently suffering from dyslipidemia?\\\" Patients who answered \"yes\\\" to either question were considered to be aware of their condition and assigned to the awareness group. Those who were diagnosed with dyslipidemia according to assessment of their lipid levels according to the NCEP-ATP III guidelines, but answered \"no\\\" to the aforementioned questions were included in the control group (unawareness group). Controlled dyslipidemia was also defined according to target levels set by NCEP-ATP III guidelines according to risk category.\n\nDiabetes mellitus was identified based on fasting blood glucose levels (\u2265126 mg/dL) or use of insulin or oral hypoglycemic agents.[@B14] Hypertension was diagnosed if subjects were taking medication for hypertension, or if average systolic blood pressure was over 140 mm Hg, or average diastolic blood pressure was over 90 mm Hg.[@B15]\n\nThe study population was divided into two groups according to the definition of obesity (BMI \u226525 kg/m^2^) by the World Health Organization Regional Office for the Western Pacific Region and adopted by the Korean Society for the Study of Obesity.[@B16]\n\n4. Health Behavior for Control of Lipid Levels\n----------------------------------------------\n\nInformation related to health behavior that has been reported as being beneficial for the control of lipid levels was collected from the KNHANES V database, and these were divided into two groups: behavioral risk factors, and nutritional risk factors. KSLA 2015 and International Atherosclerosis Society (IAS) 2013 guidelines were used when applicable.\n\n5. Behavioral Risk Factors\n--------------------------\n\nInformation on current smoking status was collected through a self-reporting questionnaire, and those who answered that they smoked, or occasionally smoked, were considered current smokers, regardless of the amount. Smoking is a major cause of atherosclerotic cardiovascular disease (ASCVD), and the IAS states that high priority must be given to the prevention or cessation of smoking in lifestyle intervention for dyslipidemia.[@B17]\n\nData on excessive alcohol consumption (\u22653 standard drinks per occasion) were also collected through the self-reporting questionnaire. KSLA 2015 guidelines recommend limiting alcohol intake to 1--2 drinks per occasion.[@B18]\n\nAdequate physical activity was also assessed through answers to questionnaires and participants were considered to engage in adequate physical activity if they reported having carried out over 30 minutes of moderate intensity physical activity at least 5 days per week, or over 20 minutes of heavy intensity at least 3 per week. Epidemiological studies have shown that physical inactivity is associated with increased risk for ASCVD and regular physical activity has beneficial effects on lipoproteins.[@B19] The IAS recommends approximately 30 minutes of daily moderate intensity activity, and specifies that the activity should be aerobic and carried out for a minimum of 5 days per week.[@B17] For vigorous intensity activity, at least 20 minutes for 3 days a week was considered as adequate according to previous American Heart Association (AHA) guidelines.[@B20]\n\n6. Nutritional Factors\n----------------------\n\nNutritional factors were measured based on adherence to dietary recommendations that could be assessed using nutritional information provided by the KNHANES V. Therefore, adequate intake of macronutrients that was monitored by performing a 24-hour food recall and analyzed by the CAN-Pro software ver. 3.0 (Korean Nutrition Society, Seoul, Korea) was compared between the awareness groups. Adequate fiber intake (\u226525 g/d), carbohydrate intake (\\<65% of total calories per day), fat intake (\\<30% of total calories per day), and protein intake (\u226515% of total calories per day) were analyzed. The 2015 KSLA guidelines recommend limiting total daily carbohydrate intake and total fat intake to less than 30% of total calories per day and eating food rich in fiber for an intake of over 25 g of fiber per day.[@B18] Diets high in cholesterol and saturated and trans-fat resulted in increased TG, LDLC, and total blood cholesterol. Previous studies have demonstrated that diets high in carbohydrate can increase TG levels, and dietary fiber can help lower blood cholesterol levels. For evaluation of protein and carbohydrate intake which was not specified in the 2015 KSLA guidelines, the American College of Sports Medicine recommendations were used where protein intake of 15% and carbohydrate intake of up to 60%--65% of total calories per day was prescribed for lifestyle management of dyslipidemia.[@B21]\n\n7. Statistical Analysis\n-----------------------\n\nAll statistical analyses were performed using STATA statistical software ver. 14.0 (Stata Corp., College Station, TX, USA). All analyses were weighted to the Korean standard population from the years 2010 to 2012, reflecting the sampling method, response rate, and population structure of the KNHANES study.\n\nUnpaired t-tests and chi-square tests were applied to continuous variables and categorical variables respectively in order to compare mean values and percentages of demographic and clinical characteristics between dyslipidemia patients according to awareness. Logistic regression was used to analyze which variables of health behavior were associated with awareness of dyslipidemia after adjusting for age. Afterwards, multivariable logistic regression was performed for each health behavior adjusting for age, education level, residential area (rural or urban), household income, self-perceived health status, marital status, and obesity (BMI \u226525 kg/m^2^). Adjusted OR, 95% confidence intervals (CI), and P-values were measured for the display of strength of each association. A P-value of \\<0.05 was considered significant.\n\nResults\n=======\n\n1. Baseline Characteristics of the Study Population\n---------------------------------------------------\n\nThe characteristics of the study population are described in [Table 1](#T1){ref-type=\"table\"}. Among the 6,624 dyslipidemia patients (14,415,037 when weighted), 17.1% were aware of their dyslipidemia status. Among the male population, 13.2% were aware of their diagnosis of dyslipidemia, while 22.6% of female dyslipidemia patients were aware of their diagnosis.\n\nThe mean age of the group who were aware of their condition was higher than that of the group that was unaware in both men and women. With respect to self-perceived health status, 14.3% of male patients in the unaware group responded that their health status was poor or very poor while 26.2% of those in the aware group gave the same response. Among men, 13.4% of the unaware and 17.5% of the aware received an elementary school or lower level of education, while 41.1% of women unaware of their dyslipidemia status received an elementary school or lower level of education and 52.2% of aware women reported receiving the same degree of education.\n\nAwareness of dyslipidemia in both male and female patients was associated with a higher prevalence of diabetes mellitus and hypertension. Lipid profiles were also significantly favorable in aware groups with higher HDLC levels and lower TC, TG, and LDLC levels in both male and female dyslipidemia patients, although differences in the level of TG was not statistically significant in men.\n\n2. Health Behavior for Control of Lipid Levels\n----------------------------------------------\n\nBoth men and women who were aware of their diagnosis of dyslipidemia had lower current smoking percentages (38.5% and 3.6%, respectively), and fewer aware women were likely to consume alcohol in excess (23.8% vs. 13.5%). When adjusted for other characteristics, female patients in the awareness group had a lower current smoking rate compared with those in the unaware group (OR, 0.55; 95% CI, 0.32 to 0.94), but when adjusted for the prevalence of diabetes and hypertension, the result was not statistically significant (OR, 0.61; 95% CI, 0.35 to 1.06). As can be observed in [Table 2](#T2){ref-type=\"table\"}, there were no significant differences for alcohol consumption or adequate physical activity.\n\n3. Adequate Macronutrient Intake\n--------------------------------\n\nThe only favorable nutritional factor observed in the aware groups by crude proportions was adequate intake of fat found in women aware of their diagnosis of dyslipidemia (93.7% vs. 96.0%). Although female patients aware of their diagnosis showed a higher OR for adequate protein intake (OR, 1.26; 95% CI, 1.00 to 1.58) when adjusted for covariates, the result was not statistically significant when additionally adjusted for the prevalence of diabetes and hypertension (OR, 1.22; 95% CI, 0.98 to 1.53). Men also showed a higher OR for adequate intake of carbohydrate (OR, 1.33; 95% CI, 1.04 to 1.72), but when adjusted for diabetes and hypertension the result was not significant (OR, 1.28; 95% CI, 0.99 to 1.67). There were no other statistically significant differences as shown in [Table 3](#T3){ref-type=\"table\"}.\n\n4. Subgroup Analysis according to Prevalence of Diabetes and Hypertension\n-------------------------------------------------------------------------\n\nIn subgroup analysis of dyslipidemia patients without either hypertension or diabetes, men aware of their diagnosis of dyslipidemia had a higher OR for adequate carbohydrate intake (OR, 1.70; 95% CI, 1.06 to 2.72) after adjusting for other factors ([Table 4](#T4){ref-type=\"table\"}).\n\nIn the subgroup analysis of dyslipidemia subjects with diabetes mellitus, there were no significant differences in health behavior according to awareness ([Table 5](#T5){ref-type=\"table\"}). In the subgroup analysis of dyslipidemia subjects with hypertension, aware women showed a lower OR for smoking (OR, 0.46; 95% CI, 0.24 to 0.91) and a higher OR for adequate protein intake (OR, 1.37; 95% CI, 1.03 to 1.84), but when adjusted for covariates the results were no longer statistically significant ([Table 6](#T6){ref-type=\"table\"}).\n\nDiscussion\n==========\n\nThis study suggests that there is minimal difference in health behavior between those aware and those unaware of their diagnosis of dyslipidemia among Korean adults over the age of 20 years. There was no statistically significant difference in health behavior and adequate macronutrient intake between the awareness groups after adjusting for demographic variables and the prevalence of diabetes and hypertension, perhaps because diabetes and hypertension are more important factors influencing adherence to health behavior. The only beneficial health behavior was found in the subgroup analysis, where adequate carbohydrate intake was observed in men with neither hypertension nor diabetes.\n\nWhen analyzing crude proportions of adequate health behavior, rates of current smoking (38.5%) and excessive alcohol consumption (66.1%) in men aware of their diagnosis of dyslipidemia were higher than desirable and rates of adequate physical activity in both male and female dyslipidemia patients (19.8% in aware men and 15.9% in aware women) showed rates with ample room for improvement. Although rates of adequate fat intake were high in patients with dyslipidemia (93.4% in aware men and 96.0% in aware women), adequate intake of fiber in all subjects (2.7% in aware men and 1.2% in aware women), and carbohydrate in women (18.5% in aware women) were comparatively low. These results show that there is room for improvement in health behavior in the general population with dyslipidemia, even in those aware of their conditions. Unfortunately, awareness was not significantly associated with the health behaviors mentioned above after adjustment for covariates, implying that merely increasing awareness is not enough to promote health behavior. However, causality cannot be determined due to the cross-sectional design of this study.\n\nAdequate control of lipid levels is important. A 1% reduction in total serum cholesterol levels can lead to a 2% to 3% reduction in risk of coronary disease,[@B22] and a meta-analysis revealed that a 10% reduction in serum cholesterol was linked to a 13% to 14% reduction in CHD mortality and 8% to 10% reduction in total mortality.[@B23] Although treatment of dyslipidemia with medication is important for adequate control of lipid levels, associated health behaviors and education for their implementation are also of importance and several guidelines (KSLA 2015, NCEP-ATP III, and IAS 2013 guidelines) actively recommend lifestyle intervention after diagnosis. In the study performed on dyslipidemia patients under pravastatin treatment in Japan, high awareness of health and positive attitude towards diet and exercise, and high adherence to drug therapy were related with favorable overall lipid levels.[@B9]\n\nResults published by the Korean Ministry of Health and Welfare on data from the KNHANES reveal that awareness rates of dyslipidemia are steadily rising (24.0% in 2005 to 59.0% in 2013); however, it is still low when compared to published results of awareness of other chronic diseases such as hypertension (65.3% in 2013) and diabetes (74.3% in 2013).[@B24] The diagnosis and improved awareness of dyslipidemia are critical first steps for its adequate management; results in this study show that lipid profiles were more favorable for both men and women aware of their condition which may have been the result of lipid-lowering agents. Use of medication is inevitably higher in groups aware of their condition, as can be observed in the baseline characteristics of subjects for this study. Because there were no significant differences in adherence to health behavior recommended by clinical guidelines except for adequate carbohydrate intake in men with neither hypertension nor diabetes and rates of several types of health behavior showed plenty of room for improvement, adherence to health behaviors must be emphasized. Such change in behavior can be expected to have additional beneficial effects on lipid levels.\n\nAlthough additional long-term studies are required to evaluate whether awareness of dyslipidemia can affect health behavior, results of this study suggest that awareness of dyslipidemia alone may not be enough to promote health behavior. One explanation can be that patients may be lacking in proper dyslipidemia education and instruction from their healthcare providers. This may also be the case for other chronic diseases. A study carried out on KNHANES 1998--2012 data analyzing adherence to dietary recommendations among Korean adults with diabetes mellitus concluded that Korean patients with diabetes have poor adherence to dietary recommendations and healthy lifestyle published by the Korean Diabetes Association regardless of awareness of diabetes.[@B25] However, another study by Bardenheier et al.[@B26] showed that both men and women aware of their diagnosis of diabetes differed in their daily intake of sugar and protein when compared to those unaware of their diabetic status. Although the difference in outcome may be due to the varied definitions of factors analyzed for association with awareness, the disparities between these studies may suggest that other factors exist apart from awareness influencing adherence to health behavior, and that the role of healthcare providers and practitioners may be crucial for lifestyle modification. Although well-established guidelines for chronic diseases call for lifestyle change as the initial line of therapy, it has been reported that physicians often do not follow these practices.[@B27] Correct insights for medical professionals with respect to guidelines in the management of dyslipidemia, and monitoring and support systems may be required for better lifestyle behavior. Further studies are needed to provide insight on attitudes of both physicians and patients towards management of dyslipidemia through lifestyle interventions.\n\nThe development and establishment of public health policies to promote adequate behavior for the management of dyslipidemia may also be of help. A project carried out in Finland found that health promotion campaigns at a national level through various activities (such as media campaigns and health fares) can help increase population awareness on prevention of chronic diseases and prompt subjects to make beneficial lifestyle changes.[@B28] Issues such as the perception of severity of risk factors and diseases, benefits of lifestyle modification, and calls to action should be addressed.[@B29] As a further step, a multi-team approach and cooperation between healthcare providers and the community can provide fully integrated care leading to improved awareness and management of dyslipidemia.\n\nIn addition, interventions at community levels may also be helpful. Fritsch et al.[@B30] reported that lipid-based interventions at the worksite (education and coaching on lifestyle and lipid values through classes and phone call interventions) can improve lifestyle behavior including exercise and diet, contribute to continuous health care, and lead to improved lipid values. After a 7-month intervention program, participants had an average of 5.2% reduction in TC.[@B30] However, this was a short-term study performed at companies that were all members of a health insurer in North Carolina. Further large-scale studies carried out for a longer time intervals, or randomized controlled trials may be required to identify intervention methods that can effectively enhance health behavior in dyslipidemia patients.\n\nThis study has several strengths. First, to the knowledge of the authors, this is the first study to analyze adherence to health behavior recommended by clinical guidelines for the management of dyslipidemia according to patients\\' awareness. Furthermore, this study was conducted using data from a nationally representative sample of the Korean population, making the estimates of this study generalizable to the population of dyslipidemia patients in Korea. Additionally, KNHANES provides data collected through standardized laboratory and physical measurements. Finally, data was used from three consecutive years, providing a large sample size and powering the statistical ability to report associations.\n\nThis study also has several limitations. Because it is based on a cross-sectional design, it is difficult to assess any temporal relationship between awareness of dyslipidemia and health behavior and adequate macronutrient intake. Although daily intake of macronutrients was assessed for nutritional factors due to limited information and lack of data on total daily cholesterol intake or level of saturated or trans-fatty acids, evaluation of dietary patterns (such as AHA 2020 ideal healthy diet or Mediterranean diet) may be more appropriate due to inconclusive evidence of an independent effect of macronutrient intake on outcomes.[@B31] Furthermore, it is unclear whether the status of each health behavior that the subjects reported would continue long term. There is also a need for the consideration of recall and social desirability bias, since data collected through self-reporting questionnaires was used for this study.\n\nIn summary, this is the first national-level study to analyze adherence to health behavior stipulated by clinical guidelines (by the KSLA and IAS) in dyslipidemia patients according to awareness. In this study, awareness was not associated with adequate health behavior except for adequate carbohydrate intake in men found in subgroup analysis of dyslipidemia subjects without hypertension or diabetes. Although further studies are required to assess any temporal relationship between awareness and health behavior, development of procedures including counseling and education by healthcare providers could be considered to guide patients according to guidelines for optimal control of lipid levels and prevention of CVD. Furthermore, large-scale prospective cohort studies that can help define reliable and practical plans for the enhancement of healthy behavior for the management of dyslipidemia are needed.\n\n**CONFLICT OF INTEREST:** No potential conflict of interest relevant to this article was reported.\n\n###### Characteristics of dyslipidemia patients according to awareness of dyslipidemia\n\n![](kjfm-38-64-i001)\n\nValues are presented as mean (95% confidence interval) for age and lipid levels and % (95% confidence interval) for the remaining variables. All data were weighted to the Korean standard population. Data and P-values were obtained from t-test for continuous variables and chi-square test for categorical variables. Diagnosis of DM: fasting blood glucose \u2265126 mg/dL, use of insulin or oral hypoglycemic agents. Diagnosis of HTN: average systolic blood pressure \u2265140 mm Hg or diastolic blood pressure \u226590 mm Hg, or taking medication for HTN.\n\nDM, diabetes mellitus; HTN, hypertension; NA, not available.\n\n^\\*^Less participants were available for analysis due to measurement methods (male: n=3,076, female: n=3,432).\n\n###### Behavioral risk factors according to awareness of dyslipidemia\n\n![](kjfm-38-64-i002)\n\nValues are presented as % (95% confidence interval) or adjusted odds ratio (95% confidence interval), weighted to the Korean standard population. Crude proportions with P-values were obtained from chi-square tests, and logistic regression analysis was performed to examine the association between awareness of dyslipidemia and each health behavior with the adjustments specified. Model 1: adjusted for age; model 2: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, and household income; model 3: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, household income, prevalence of diabetes, and prevalence of hypertension.\n\nBMI, body mass index.\n\n^\\*^Statistically significant, P\\<0.05. ^\u2020^Excessive alcohol consumption: \u22653 standard drinks on \u22651 occasion in an average week. ^\u2021^Adequate physical activity: \\>150 minutes per week of moderate-intensity activity or \\>60 minutes per week of vigorous-intensity activity.\n\n###### Nutritional factors according to awareness of dyslipidemia\n\n![](kjfm-38-64-i003)\n\nValues are presented as % (95% confidence interval) or adjusted odds ratio (95% confidence interval), weighted to the Korean standard population. Crude proportions with P-values were obtained from chi-square tests, and logistic regression analysis was performed to examine the association between awareness of dyslipidemia and each health behavior with the adjustments specified. Model 1: adjusted for age; model 2: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, and household income; model 3: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, household income, prevalence of diabetes, and prevalence of hypertension.\n\nBMI, body mass index.\n\n^\\*^Statistically significant, P\\<0.05. ^\u2020^Adequate intake of fiber: \u226525 g/d. ^\u2021^Adequate intake of carbohydrate: \\<65% of total calories per day. ^\u00a7^Adequate intake of fat: \\<30% of total calories per day. ^\u2225^Adequate intake of protein: \u226515% of total calories per day.\n\n###### Subgroup analysis of health behavior according to awareness of dyslipidemia in subjects without hypertension or diabetes mellitus\n\n![](kjfm-38-64-i004)\n\nValues are presented as % (95% confidence interval) or adjusted odds ratio (95% confidence interval), weighted to the Korean standard population. Crude proportions with P-values were obtained from chi-square tests, and logistic regression analysis was performed to examine the association between awareness of dyslipidemia and each health behavior with the adjustments specified. Model 1: adjusted for age; model 2: adjusted for age, obesity (body mass index \u226525 kg/m^2^), self-perceived health status, education, residential area, and household income.\n\n^\\*^Statistically significant, P\\<0.05. ^\u2020^Excessive alcohol consumption: \u22653 standard drinks on \u22651 occasion in an average week. ^\u2021^Adequate physical activity: \\>150 minutes per week of moderate-intensity activity or \\>60 minutes per week of vigorous-intensity activity. ^\u00a7^Adequate intake of fiber: \u226525 g/d. ^\u2225^Adequate intake of carbohydrate: \\<65% of total calories per day. ^\u00b6^Adequate intake of fat: \\<30% of total calories per day. ^\\*\\*^Adequate intake of protein: \u226515% of total calories per day.\n\n###### Subgroup analysis of health behavior according to awareness of dyslipidemia in subjects with diabetes mellitus\n\n![](kjfm-38-64-i005)\n\nValues are presented as % (confidence interval) or adjusted odds ratio (95% confidence interval), weighted to the Korean standard population. Crude proportions with P-values were obtained from chi-square tests, and logistic regression analysis was performed to examine the association between awareness of dyslipidemia and each health behavior with the adjustments specified. Model 1: adjusted for age; model 2: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, and household income; model 3: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, household income, and hypertension.\n\nBMI, body mass index.\n\n^\\*^Excessive alcohol consumption: \u22653 standard drinks on \u22651 occasion in an average week. ^\u2020^Adequate physical activity: \\>150 minutes per week of moderate-intensity activity or \\>60 minutes per week of vigorous-intensity activity. ^\u2021^Adequate intake of fiber: \u226525 g/d. ^\u00a7^Adequate intake of carbohydrate: \\<65% of total calories per day. ^\u2225^Adequate intake of fat: \\<30% of total calories per day. ^\u00b6^Adequate intake of protein: \u226515% of total calories per day.\n\n###### Subgroup analysis of health behavior according to awareness of dyslipidemia in subjects with hypertension\n\n![](kjfm-38-64-i006)\n\nValues are presented as % (confidence interval) or adjusted odds ratio (95% confidence interval), weighted to the Korean standard population. Crude proportions with P-values were obtained from chi-square tests, and logistic regression analysis was performed to examine the association between awareness of dyslipidemia and each health behavior with the adjustments specified. Model 1: adjusted for age; model 2: adjusted for age and obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, and household income; model 3: adjusted for age, obesity (BMI \u226525 kg/m^2^), self-perceived health status, education, residential area, household income, and diabetes mellitus.\n\nBMI, body mass index.\n\n^\\*^Statistically significant: P\\<0.05. ^\u2020^Excessive alcohol consumption: \u22653 standard drinks on \u22651 occasion in an average week. ^\u2021^Adequate physical activity: \\>150 minutes per week of moderate-intensity activity or \\>60 minutes per week of vigorous-intensity activity. ^\u00a7^Adequate intake of fiber: \u226525 g/d. ^\u2225^Adequate intake of carbohydrate: \\<65% of total calories per day. ^\u00b6^Adequate intake of fat: \\<30% of total calories per day. ^\\*\\*^Adequate intake of protein: \u226515% of total calories per day.\n"}
+{"text": "1. Introduction {#sec1-cancers-12-00714}\n===============\n\nRecent clinical immunotherapy (IT) trials have demonstrated that immune checkpoint inhibitors (ICI) can cause immune rejection of a minority of human tumors \\[[@B1-cancers-12-00714]\\]. A prerequisite for ICI efficacy is a state of preexisting tumor immunity that is often correlated with intratumoral CD8+ T lymphocyte infiltration \\[[@B2-cancers-12-00714]\\], modified by many other conditions and factors such as the PD-1/PD-L1 pathway. Cytotoxic CD8+ T cells are almost always a critical element driving tumor immune regression, for which tumor regression antigens as initiators and targets are prerequisites. This also means that generating de novo tumor immunity is of paramount importance if the scope of successful IT is to be broadened. The possibility of integrating ICIs within radiation therapy (RT) of solid cancers is being intensely investigated in what can only be described an unprecedented number of clinical trials \\[[@B3-cancers-12-00714]\\], but the efficacy of this combined approach is still uncertain. Furthermore, there are many features of the interaction of RT with the immune system in cancer therapy that need further elucidation.\n\nWe and others have shown that RT can boost anti-tumor immune responses clinically \\[[@B4-cancers-12-00714],[@B5-cancers-12-00714],[@B6-cancers-12-00714]\\], but rarely does RT convert a non-responder into a responder. Similar observations can be obtained in preclinical models of IT-RT combinations. Immunogenic murine tumors, such as the FSA1 fibrosarcoma used in this study, grow relentlessly in the face of concomitant immunity, defined as a state of effective immunity coupled with progressive tumor growth. Numerous tumor immune escape mechanisms have been proposed. For this tumor model, we have previously described two-zones of tolerance that are generated by small and large tumor burdens, with an intermediate window of concomitant immunity \\[[@B7-cancers-12-00714]\\]. In this model, as tumors grow, tumor-specific Tregs are the first to downregulate immunity and this is followed later by non-specific immune suppression by myeloid cells. Under some circumstances, removal of the primary by RT, or surgery, has been reported to unmask an immune state \\[[@B8-cancers-12-00714]\\]. Since concomitant immunity can decrease the radiation dose required for primary tumor cure \\[[@B9-cancers-12-00714]\\], understanding how immunosuppression functions is essential to improving local and distant radiation effects. Furthermore, if immunity were to be effective against microscopic metastatic disease, a compelling case can be made for treating oligometastatic disease. The importance of tumor burden in driving immunosuppression cannot be overstated, however it remains a largely unrecognized confounding factor when assessing tumor immunity in response to RT, with and without IT.\n\nHere, we use a moderately immunogenic murine model of fibrosarcoma to explore the effect of hypofractionated RT (hRT) with and without PD-1 blockade on antitumor immunity as it relates to tumor load and examine the ability of the concomitant immune state to deal with distant microscopic disease. Clinically, localized high-risk soft tissue sarcoma (STS) is typically treated with wide local resection in combination which RT in the neoadjuvant or adjuvant setting with or without chemotherapy \\[[@B10-cancers-12-00714],[@B11-cancers-12-00714]\\] but distant recurrences remain a major cause of mortality suggesting the presence of micrometastatic disease at the time of initial treatment \\[[@B12-cancers-12-00714]\\]. Neoadjuvant hRT is under investigation as a novel treatment approach to STS, as is anti-PD-1 treatment, which speaks to the relevance of the study.\n\n2. Results {#sec2-cancers-12-00714}\n==========\n\n2.1. Tumor Load Determines the Immune Balance---Locally and Systemically {#sec2dot1-cancers-12-00714}\n------------------------------------------------------------------------\n\nImmunogenic FSA1 tumors grow progressively \\[[@B13-cancers-12-00714]\\] as concomitant immunity becomes suppressed, first by tumor-specific Tregs and later by non-specific activated myeloid cells \\[[@B7-cancers-12-00714],[@B14-cancers-12-00714]\\]. Less is known about how therapy-induced tumor regression influences the immune system locally and systemically \\[[@B8-cancers-12-00714]\\]. As a prelude to such studies, in a series of experiments we examined how immune parameters in untreated T1 and T2 tumors and in spleen were influenced by tumor burden ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}). In this model, a second inoculum (T2) was injected into the contralateral side 7 days after the primary injection. Not surprisingly, not all T2 inocula grew indicating a state of immunity from the T1 injection, but some did yield progressively growing tumors ([Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}B). The data from these experiments were pooled and intratumoral lymphoid and myeloid cell populations in T1 and T2 tumors were plotted against individual tumor size or, for the spleen, against combined size (T1 + T2) in individual linear regression graphs and summarized in heatmaps ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}D--F, [Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}).\n\nIn the lymphoid compartment ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}A), as T1 and T2 tumors grew their content of CD4+ lymphocytes decreased progressively (T1, R = \u22120.493, *p* = 0.027), while the fraction of regulatory CD4+ T cells (Tregs) increased (T1, R = 0.77, *p* \\< 0.001). This scenario was mirrored in the spleen (CD4, R = \u22120.485, *p* = 0.03; Tregs, R = 0.715, *p* \\< 0.001). The CD8+ cell compartment also tended to shrink, but this was significant only in the spleen (T1, n.s.; spleen, R = \u22120.487, *p* = 0.029). Their activation status, assessed by PD-1 expression and intracellular IFN-\u03b3 levels, were little affected.\n\nThe myeloid compartment was more variable between locations ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}B) with CD11b cells increasing in T1, but not T2 tumors and spleen. Polymorphonuclear-myeloid derived suppressor cells (PMN-MDSC) increased only in larger T1 tumors (\\>10 mm diameter, R = 0.674, *p* = 0.001) while M2 macrophages showed more measured increases in T1 tumors (T1, R = 0.695, *p* = 0.001). The percent of M-MDSC and PD-L1+ CD11b+ myeloid cells increased in the spleen (R = 0.564, *p* = 0.01; R = 0.604, *p* = 0.005, respectively), but not in tumors although values in tumors were generally high, perhaps leaving little room for further increases. PD-L1 expression levels did rise in T1 tumors and spleens, suggesting activation. Overall, the myeloid content of T2 tumors was hardly affected, in keeping with myeloid changes being mainly associated with larger tumor burden. Not surprisingly, lymphoid:myeloid ratios decreased dramatically with tumor size/burden in all locations ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}C).\n\nStatistical comparisons of infiltrates in T1 and T2 tumor groups is complicated by their different sizes. Therefore, T1 tumors were grouped as being below and above the median size of 11.5 mm, which allowed large and small T1 tumors to be distinguished with 100% accuracy (95% cross-validated) by the combined differences in CD4/CD11b ratio, CD8/CD11b ratio, and PMN-MDSCs (*p* \\< 0.0001). Small T1 tumors could be compared with T2 tumors as they were the same size, and tests of equality of means showed only the CD4:CD11b ratio to be significantly different (*p* \\< 0.021). However, if the effect of this variable was statistically neutralized, the Treg content became significant (*p* \\< 0.004). These comparisons suggest that tumors self-regulate their immune content depending on their size, but the lymphoid compartment of T2 tumors was influenced by the presence of primary T1 tumors, especially with respect to Tregs. Clearly, tumor burden is a major influence on the intratumoral and systemic immune profiles in tumor-bearing animals, altering lymphoid and myeloid compartments, something that must be kept in mind when interpreting the effects of RT on immune parameters. Interestingly, tumor burden also correlated with the circulating neutrophil count and the circulating neutrophil:lymphocyte ratio, although no correlation was observed between tumor burden and the circulating lymphocyte count ([Figure S3](#app1-cancers-12-00714){ref-type=\"app\"}).\n\n2.2. Radiation Plus Anti-PD-1 Monoclonal Antibody Can Drive Superior Systemic Tumor Control {#sec2dot2-cancers-12-00714}\n-------------------------------------------------------------------------------------------\n\nPilot experiments were conducted to develop the optimal hRT regimen for the treatment of established 6--8 mm diameter FSA1 T1 tumors in the absence of distant, microscopic disease. A single dose of 8 Gy caused tumor growth delay but no cure, 2 fractions of 8 Gy cured a few mice, while 8 Gy \u00d7 3 gave 100% cure ([Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}A). We chose to use the 8 Gy \u00d7 3 regimen to the primary T1 tumors to determine if curative hRT could affect the growth of distant micrometastases, in the context of PD-1 checkpoint blockade ([Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}B). T1 tumors were just palpable when T2 inocula were given and, in the face of the immunity present, only 35% of T2 (6 of 17) grew ([Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}B left). hRT (8 Gy \u00d7 3) as a single modality, again, induced complete regression of T1 tumors, but any immunity that was generated was insufficient to cause abscopal regression of T2 tumors. Anti-PD-1 treatment alone, started on day 9, was ineffective both at controlling 6--8 mm primary tumors and at preventing the appearance of T2 tumors. However, addition of anti-PD-1 to focal hRT systemically controlled T2 as well as T1 tumors, curing 100% of mice (0 of 17 T2 grew, [Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}B right, *p* = 0.018). Of note, radiation-induced lymphopenia was not altered by the addition of anti-PD1 ([Figure S4](#app1-cancers-12-00714){ref-type=\"app\"}).\n\n2.3. hRT Alters the Immune Balance {#sec2dot3-cancers-12-00714}\n----------------------------------\n\nThe immune interplay between hRT-induced primary tumor regression and microscopic unirradiated deposits was examined, with and without anti-PD-1 treatment. Immune profiling was performed 7 days after hRT when irradiated primary tumors were regressing ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}A). In the lymphoid compartment of irradiated T1 tumors, CD8+ T cells dramatically and significantly increased within the CD45+ population (*p* \\< 0.001), which was, to a lesser extent, also true in distant T2 tumors (*p* \\< 0.05) ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}B). Although there was no increase in the overall proportion of CD8+ T cells that were positive for IFN-\u03b3, the actual level of IFN-\u03b3 expression on a per cell basis was enhanced in T1 tumors and spleen, but less in T2 tumors. Tests of equality of group means showed that the increase in the number of CD8+ cells and IFN-\u03b3 expression levels in irradiated tumors were highly significant (*p* \\< 0.0001), and that either measure was 100% accurate in predicting treatment response. CD4+ cells increased only in T2 tumors after hRT, while Tregs increased in all sites, with statistical significance being reached in primary T1 tumors (*p* \\< 0.05). An increase myeloid cells was noticeable in the spleen, but the high level of M2 macrophages characteristic for primary tumors decreased dramatically after hRT (*p* \\< 0.01; [Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}C). Lymphoid:myeloid ratios increased in T1 and T2 tumors, but for different reasons (i.e., CD8 vs. CD4, respectively) ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}D). Tests of equality between means for T2 and irradiated T1 tumors showed that levels of IFN-\u03b3 expression could discriminate between groups with 92% accuracy.\n\nThese profiles were consistent with radiation-induced regression of T1 tumors that was assisted by activated CD8+ cells, but this could not overcome systemic barriers to cause abscopal regression of microscopic lesions. This was confirmed by examining the PD-1/PD-L1 axis. Increases in CD8+ cells in the irradiated primary tumor were accompanied by increases in the number of CD8+ cells expressing PD1 and their levels of PD1 expression ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}B). Similar increases in the number of PD-1+CD8+ T cells occurred in the spleen albeit without a change in expression levels. However, unirradiated T2 tumors if anything, surprisingly, had decreased PD-1+CD8+ numbers and expression levels. Similarly, the fraction of PD-L1+ myeloid cells and their expression levels were upregulated after hRT both in the irradiated primary tumor and in the spleen, but not in T2 tumors ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}C). It should be noted that in this model hRT did not alter the, typically 60--80%, PD-L1 expression by tumor cells defined as CD45- cells in any of the tumor sites ([Figure S5](#app1-cancers-12-00714){ref-type=\"app\"}).\n\nImportantly, since the irradiated T1 tumors were regressing at the time of analysis, i.e., lesser tumor load in these animals, any changes observed in immune status could relate to (1) a direct radiation-immune effect, (2) an indirect effect through the reduction in tumor burden, or (3) both. However, the observation of the highly significant rise in the CD8+ infiltrate and their ability to produce IFN-\u03b3 in irradiated T1 tumors ([Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}B) is independent of the reduction in tumor size ([Figure 1](#cancers-12-00714-f001){ref-type=\"fig\"}A), as is the proportional rise in CD4 Tregs.\n\n2.4. Tumor hRT Plus Systemic PD-1 Blockade Increases the Functionality of T Cells---Locally and Systemically {#sec2dot4-cancers-12-00714}\n------------------------------------------------------------------------------------------------------------\n\nIn contrast to hRT alone, combining systemic anti-PD1 with hRT triggered abscopal regression of microscopic disease (*p* = 0.018; [Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}). Changes in immune cell composition in T1 tumors that were equal in size in these two groups and spleen were therefore examined. Anti-PD-1 alone had no effect on immune cell composition. Interestingly, compared to hRT alone, the combination of ICI and hRT further increased functionality of CD8+ T cells, measured by the intensity of the intracellular IFN-\u03b3 signal in both, the irradiated tumor (*p* = 0.003), and the spleen (*p* \\< 0.001; [Figure 3](#cancers-12-00714-f003){ref-type=\"fig\"}B). T2 tumors obviously could not be evaluated as they did not grow.\n\n3. Discussion {#sec3-cancers-12-00714}\n=============\n\nThe immunogenic FSA1 murine fibrosarcoma used in this study has been the subject of many investigations by ourselves \\[[@B7-cancers-12-00714],[@B14-cancers-12-00714],[@B15-cancers-12-00714]\\] and others \\[[@B8-cancers-12-00714],[@B9-cancers-12-00714],[@B16-cancers-12-00714]\\]. In this study we show it to be PD-1 resistant, recapitulating the clinical response to anti-PD-1 described so far in STS patients. Two phase II studies of anti-PD-1 monotherapy in patients with advanced sarcoma have been published recently. One reported a lack of response to Nivolumab in patients with advanced uterine leiomyosarcoma \\[[@B17-cancers-12-00714]\\], the other put the objective response rate in a STS cohort of various types receiving Pembrolizumab at 18% \\[[@B18-cancers-12-00714]\\]. In our murine model, anti-PD-1 treatment was ineffective unless given with locally ablative hRT that, although it generated strong local CD8+ T cell responses, was itself remarkably inefficient in activating systemic immunity to cause abscopal regression even of microscopic disease. Importantly, this gain in anti-tumor efficacy with the hRT-ICI combo came at no apparent increased toxicity compared to either monotherapy.\n\nIn this study, T2 secondary inocula were given one week after the primary, which decreased the incidence of tumor take. Those that did grow presumably did so only because for them tolerance was induced. hRT (8 Gy \u00d7 3) monotherapy of T1 tumors dramatically enhanced the representation and function of intratumoral CD8+ T cells. This was also evident in the spleen, but T2 tumors seemed less affected. Indeed, in T2 tumors CD4+ cell changes were more marked than those for CD8+ cells. This study therefore suggests that hRT of primary tumors can activate immunity that is likely helpful in causing local regression, but that this is insufficient to break tolerance at other sites. This is not unlike what has been observed in the clinical setting. A comparison of paired of human sarcoma specimen before and after RT showed that radiation might promote an intratumoral immune effector signature and upregulate Major Histocompatibility Complex (MHC) class I expression \\[[@B19-cancers-12-00714]\\]. Similarly, the expression of IFN-associated genes in post-SBRT tumor biopsy specimens correlated with an abscopal response in patients with advanced solid tumors treated with radiotherapy and pembrolizumab \\[[@B20-cancers-12-00714]\\].\n\nThe general concept that hRT can interface with the immune system on multiple levels has been established. This includes tumor neoantigen release \\[[@B21-cancers-12-00714]\\], tumor adjuvant release (DAMPS) \\[[@B22-cancers-12-00714]\\] and enhanced type I IFN signaling \\[[@B23-cancers-12-00714]\\] leading to activation of dendritic cells and the antigen processing machinery, activation and clonal expansion of effector T cells at the tumor site (TILs) and elsewhere \\[[@B24-cancers-12-00714],[@B25-cancers-12-00714]\\], and tumor death receptor upregulation \\[[@B26-cancers-12-00714]\\]. There is a growing body of evidence that ICIs can enhance many of these RT-induced effects, with preclinical and occasional clinical reports of abscopal responses \\[[@B3-cancers-12-00714],[@B27-cancers-12-00714]\\]. The ideal radiation dose and fractionation regimen for stimulating a systemic anti-tumor immune response remains controversial. A dose per fraction of 8 Gy has been shown to best induce antitumor immunity and abscopal response in preclinical models \\[[@B28-cancers-12-00714],[@B29-cancers-12-00714],[@B30-cancers-12-00714]\\] and hRT is being investigated with promising results. Timing is probably critical. In one study, combination immune-radiotherapy with concurrent but not sequential PD-1/PD-L1 signaling blockade improved treatment outcomes \\[[@B31-cancers-12-00714]\\]. The Polish Sarcoma Study Group reported the results of 5 \u00d7 5 Gy followed by surgery three to seven days later. This strategy was found to be safe with an estimated 5-year local failure rate of 19% \\[[@B32-cancers-12-00714]\\]. Furthermore, the preliminary results of an ongoing phase II trial (NCT02701153) on preoperative hRT (5 \u00d7 6 Gy) showed a 19% major wound complication rate and a mean pathologic necrosis score of 57%, which are similar to data from historical controls \\[[@B33-cancers-12-00714]\\].\n\nThe considerable importance of tumor burden in determining the state of immunity has long been recognized, although it is only rarely taken into consideration, clinically or preclinically. Immunosuppression is integral to the proper function of both lymphoid and myeloid systems, with the former tending to show antigen specificity through Tregs, which often precedes a more general immune shutdown led by myeloid cells in patients with advanced disease \\[[@B14-cancers-12-00714]\\]. The latter state may also allow metastases to develop. Experimental evidence over the past 50 years have documented many mechanisms of tumor escape \\[[@B34-cancers-12-00714]\\], but the fact that both high and low sized inocula of FSA1, and other immunogenic tumors, can depress responses to subsequent challenges has been known for many years \\[[@B7-cancers-12-00714],[@B35-cancers-12-00714]\\] and the importance of time between tumor challenges studied \\[[@B36-cancers-12-00714]\\]. These experiments relate strongly to how tumor-induced changes in the immune system might promote the development of secondary tumors and present systemic barriers that have to be overcome before immune regression responses can be systemically generated. Not surprisingly, immunotherapy of FSA1 with C. parvum was only effective against moderate size inocula that could generate immunity \\[[@B16-cancers-12-00714]\\]. Not only were small-size inocula not rejected but tumor take was actually enhanced by immunotherapy. These basic rules will apply to most immunogenic tumors and attempts at immunotherapy.\n\n4. Materials and Methods {#sec4-cancers-12-00714}\n========================\n\n4.1. Mice {#sec4dot1-cancers-12-00714}\n---------\n\nAnimal experiments used immunocompetent 6- to 8-weeks-old C3Hf/Kam female mice bred and maintained in the defined-flora, AALAC-accredited animal facility of the Department of Radiation Oncology at the University of California, Los Angeles. All experimental protocols adhered to local and national animal care guidelines with IACUC approval. The protocol number for these experiments, it is: \\#1999-173-61, the approval date was 14 July 2017.\n\n4.2. Cell Line and Reagents {#sec4dot2-cancers-12-00714}\n---------------------------\n\nFSA1 is a moderately immunogenic undifferentiated fibrosarcoma induced by single injection of methylcholanthrene into the flank of the same strain of C3H mice \\[[@B37-cancers-12-00714]\\]. This tumor model has been used extensively, as quoted above. FSA1 cells were grown from frozen stock in vitro in Dulbecco Modified Eagle Medium (DMEM) (Corning Incorporated, Corning, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Sigma-Aldrich Corporation, St. Louis, MO, USA country) and 1% antibiotic-antimycotic (Corning Incorporated, Corning, NY, USA), at 37 \u00b0C in a humidified atmosphere containing 5% CO2 and harvested by trypsinization for injection. Anti-PD-1 mAb clone RMP1-14 were purchased from BioXCell^\u00ae^ (West Lebanon, NH, USA) and 300 \u00b5g per mouse injected intraperitoneally (i.p.) on days 9, 12, 15 and 18.\n\n4.3. Tumor Model and RT {#sec4dot3-cancers-12-00714}\n-----------------------\n\nTumor inocula of 5 \u00d7 10^5^ FSA1 cells in 100 \u00b5L of Dulbecco's phosphate-buffered saline (DPBS; Corning Incorporated) were injected subcutaneously into the upper thigh region. The right thigh was used to generate primary tumors (T1) and seven days later, the same number of tumor cells was injected in the left side to mimic a metastatic site (T2). Mice were randomly assigned to one of four treatment groups with *n* = 17 mice per group: no treatment, anti-PD-1 monoclonal antibody (mAb) alone, RT alone, and the combination. RT was delivered to T1 tumors when they were 6--8 mm diameter, which was nine days after implantation i.e., two days after implantation of T2, when were not yet palpable, mimicking a micrometastatic site and based an average T1 tumor take rate of 100% ([Figure 2](#cancers-12-00714-f002){ref-type=\"fig\"}). For hRT, mice were anesthetized via a single i.p. injection of ketamine 100 mg/kg and xylazine 6 mg/kg and 3 fractions of 8 Gy were delivered to T1 tumors using a Gulmay RS320 X-ray unit at 300 kV and 10 mA with 1.5 mm Cu and 3 mm Al beam filtration and a dose rate of 1.7 Gy/min (Gulmay Medical LtD., Camberley, Surrey, UK) with the rest of the body, including T2 tumors, shielded. Dosimetry was verified by TLDs and Gafchromic EBT3 film calibrated against NIST-traceable sources. Tumor diameters were measured in two perpendicular dimensions with Vernier calipers every third day for the duration of the experiment. Mice were euthanized if their body-conditioning score reached 2. Separate cohorts of *n* = 6 mice from each treatment group (plus *n* = 20 of untreated mice) were used for immune profiling to analyze treatment-related local and systemic immune changes. The data from the repeated experimental groups were consistent and therefore pooled for presentation. Mice were sacrificed on day 18 and tumor, spleen, and blood harvested for analysis.\n\n4.4. Tissue Harvest and Immune Profiling {#sec4dot4-cancers-12-00714}\n----------------------------------------\n\nComplete differential blood counts were obtained from 100 \u00b5L whole blood taken via cardiac puncture of euthanized mice harvested into heparinized capillary tubes (Fisher Scientific, Pittsburg, PA, USA) using a Hemavet HV950 (Drew Scientific Inc., Miami Lakes, FL, USA). Tumors and spleens were harvested on ice. Tumors were cut into small pieces, enzymatically digested in 15 mL DPBS supplemented with Collagenase D 1 mg/mL (Sigma) and DNAse I 0.1 mg/mL (Sigma) for 30 min under slow rotation at 37 \u00b0C. Spleens were teased open and red blood cells lysed with ACK lysing buffer (Lonza) for 5 min. 1--2 \u00d7 10^6^ cells in DPBS were stained with fixable viability stain (BD Horizon FVS 510) prior to incubation with rat anti-mouse CD16/CD32 Fc blocking mAb (BD Pharmingen, San Jose, CA, USA; Clone 2.4G2) for 5 minutes at 4 \u00b0C followed by 30 mins at 4 \u00b0C with the following fluorochrome-conjugated antibodies in 2 separate panels: AF700 anti-mouse CD45.2 (BioLegend, San Diego, CA, USA; Clone 104), PerCP-Cy5.5 anti-mouse CD8a (BioLegend, Clone 53-6.7), APC/Cy7 anti-mouse CD4 (BioLegend, Clone GK1.5), PE anti-mouse CD25 PE (BioLegend, Clone PC61), APC anti-mouse FoxP3 (Invitrogen, Clone FJK-16s), BV785 anti-mouse PD-1 (BioLegend, Clone 29F.1A12), BV650 anti-mouse IFN-\u03b3 (BioLegend, Clone XMG1.2), FITC anti-mouse CD11b (BioLegend, Clone M1/70), APC/Cy7 anti-mouse Ly-6G (BioLegend, Clone 1A8), PE-CF594 anti-mouse Ly-6C (BD Horizon, Clone AL-21), APC anti-mouse F4/80 (BioLegend, Clone BM8), BV711 anti-mouse CD206 (BioLegend, Clone C068C2), PE anti-mouse PD-L1 (BioLegend, Clone 10F.9G2). For panels containing FoxP3, surface marker staining was performed first, prior to fixation and permeabilization using the BD Pharmingen\u2122 Transcription-Factor Buffer Set. To determine the functionality of CD8+ T cells, single cell suspensions were stimulated ex vivo with PMA (Phorbol 12-Myristate 13-Acetate) and ionomycin (BD Pharmingen, Leukocyte Activation Cocktail with BD GolgiPlug\u2122) for 6 h at 37 \u00b0C, prior to staining for surface markers and intracellular IFN-\u03b3. Data were collected on a 14-color LSRFortessa\u2122 (BD Biosciences, San Jose, CA, USA) and analyzed using FlowJo (FlowJo, LLC, Ashland, OR, USA). Gating strategies and subset definition for lymphoid and myeloid cell populations are shown in [Figure S1](#app1-cancers-12-00714){ref-type=\"app\"} and [Table S1](#app1-cancers-12-00714){ref-type=\"app\"}. Local immune infiltrates referring to individual tumors are plotted against \"tumor size\" while systemic immune profiles (e.g., spleen) are shown versus collective \"tumor burden\" or \"tumor load\", as the sum or T1 and T2.\n\n4.5. Statistics {#sec4dot5-cancers-12-00714}\n---------------\n\nData are presented as means +/\u2212 standard error of the mean (S.E.M.) using GraphPad Prism. Student's *t*-tests were applied to compare differences between treatment groups, considered significant at the 5% probability level. Simple linear regression analysis was used to determine the correlation of each individual immune endpoint with tumor burden according to the Pearson correlation coefficient r, with multivariate regression analysis highlighting the strongest of these immune correlates. The nature of immune endpoints that were necessary and sufficient to distinguish between large and small tumors or between treatment groups (at equal tumor size) was established through multivariate linear discriminant analysis and accuracy measured with and without cross-validation. Differences in the proportion of T2 tumor control was assessed by the Fisher's exact test. IBM SPSS Statistics for Windows, Version 24.0. (Armonk, NY, USA) was used to conduct the multiple regression and multivariate linear discriminant analysis. Heatmaps were constructed using R version 3.6.1.\n\n5. Conclusions {#sec5-cancers-12-00714}\n==============\n\nIn this study, anti-PD-1 ICI monotherapy was ineffective against palpable and microscopic tumors but could synergize if hRT was given to the primary supporting the notion that the PD-1/PD-L1 axis can be a key negative feedback mechanism. In other words, RT requires ICI to overcome the systemic barriers that are generated by primary tumors and that prevent the radiation-induced local immunity from spreading to generate effective abscopal regression of micrometastatic disease. This requirement is similar to the previous findings using anti-CTLA4 plus hRT \\[[@B28-cancers-12-00714]\\]. This hypothesis is currently being tested in the SU2C-SARC032 randomized phase II trial (NCT03092323), comparing neoadjuvant RT with or without pembrolizumab followed by surgery in patients with clinically localized extremity STS at high risk for developing metastatic disease (tumor size \\> 5 cm, intermediate- to high-grade). It cautions that tumor burden and associated immune suppression are major factors in determining the outcome and immunologic correlative studies that examine functional endpoints will be critical for understanding of the such interactions \\[[@B38-cancers-12-00714]\\].\n\nThe following are available online at <https://www.mdpi.com/2072-6694/12/3/714/s1>, Figure S1: Multi-parametric flow cytometry gating strategies for lymphoid and myeloid cell populations, Figure S2: Immune cell profile in the primary tumor T1 depends on the tumor size, Figure S3: Linear regression analysis for circulating lymphocytes and neutrophils versus tumor burden, Figure S4: Impact of radiation treatment alone or in combination with PD-1 blockade on circulating neutrophils and lymphocytes, Figure S5: PD-L1 expression by tumor cells in response to radiation therapy alone or in combination with anti-PD-1, Table S1: Phenotypic markers to define immune cells.\n\n###### \n\nClick here for additional data file.\n\nConceptualization, J.P.N., A.K., W.H.M., and D.S.; methodology, J.P.N., M.-H.L., and C.N., formal analysis, J.P.N., J.W.S., W.H.M., and D.S.; resources, P.N., W.H.W., and D.S.; writing---original draft preparation, J.P.N., W.H.M., and D.S.; writing---review and editing, J.P.N., A.K., W.H.M., and D.S.; visualization, J.P.N., and T.R., supervision, W.H.W., and D.S.; funding acquisition, P.N., W.H.W., and D.S. All authors have read and agreed to the published version of the manuscript.\n\nThis research was funded by the Cancer Foundation Luxembourg (J.P.N.), UCLA's CTSI KL2 Award (A.K.), the SARC Career Enhancement Program (A.K.), and the NIH with 5R01CA191234-05 (D.S.) and 1R01CA226875-01 (D.S.).\n\nThe authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.\n\n![Immune suppressive mechanisms kick in locally and systemically as tumor load increases. The percentages of lymphoid (**A**) and myeloid immune cell subsets (**B**), and the ratios thereof (**C**) in the primary tumor T1, secondary tumor T2 and spleen of untreated mice depend on the tumor size and total tumor burden (*n* = 20). Heatmap showing a rise in myeloid and lymphoid suppressor cells dominance with increase in tumor burden in spleen (**D**), in T1 (**E**), and in T2 (**F**). Heatmap was constructed by normalizing each immune cell subset across all mice. Values are *z*-scores for each immune subset in each mouse. r represents the Pearson correlation coefficient. \\* *p* \\< 0.05; \\*\\* *p* \\< 0.01; \\*\\*\\* *p* \\< 0.001; ns not significant.](cancers-12-00714-g001){#cancers-12-00714-f001}\n\n![Micrometastatic fibrosarcoma is effectively controlled with ablative irradiation combined with anti-PD-1. (**A**) In a preliminary experiment, unilateral tumors were treated with irradiation (RT) alone to analyze the radiation dose response (*n* = 5 per radiation dose). (**B**) Primary (T1) and secondary (T2) tumors diameters were monitored in the no treatment, anti-PD-1, irradiation and combination groups (*n* = 17 per treatment group).](cancers-12-00714-g002){#cancers-12-00714-f002}\n\n![Immune cell profiles and PD-1/PD-L1 expression levels in the primary tumor T1, the secondary tumor T2 and the spleen in response to radiation therapy (RT) with or without anti-PD-1 (RT+aPD1) (*n* = 6 per group). (**A**) Overall mean relative changes in immune cells subsets; (**B**) Lymphoid subsets; (**C**) Myeloid subsets; (**D**) Ratio of lymphoid-to-myeloid subsets. In every tissue sample, there was a proportion of undetermined cells among the CD45+ cells, which did not change between the control and the irradiated groups (not shown). Data are mean +/\u2212 S.E.M. with \\* *p* \\< 0.05; \\*\\* *p* \\< 0.01; \\*\\*\\* *p* \\< 0.001; ns not significant. ND: No Data.](cancers-12-00714-g003){#cancers-12-00714-f003}\n"}
+{"text": "Introduction {#S1}\n============\n\nAdenine phosphoribosyl-transferase (APRT) is the key enzyme that transforms adenine into adenylate monophosphate (AMP). In case of APRT deficiency, xanthine oxidase rapidly oxidizes adenine into 2,8-dihydroxyadenine (2,8-DHA), which is then eliminated by the kidneys through tubular secretion. 2,8-DHA is insoluble at a physiologic range of pH; therefore, it can easily precipitate in the urine. Although subjects with APRT deficiency may have repeated kidney stones, some of these patients may remain asymptomatic throughout life \\[[@R1]\\]. Nevertheless, a handful of reports propose that APRT deficiency can also lead to 2,8-DHA crystal-induced acute or chronic renal failure \\[[@R3]\\].\n\nIn this paper, we report a case of a middle-aged man with APRT deficiency that led to 2,8-DHA crystal-induced chronic active tubulointerstitial nephritis.\n\nCase {#S2}\n====\n\nA 59-year-old man was referred for the evaluation of progressive chronic kidney disease with a serum creatinine level of 5.5 mg/dL (487 \u03bcmol/L). He was on aspirin, furosemide and diltiazem for mild coronary artery disease and hypertension. He had a history of only one kidney stone 10 years before, at which time the creatinine level was normal. He denied any family history of kidney stones and any physical complaints. His blood pressure was 140/75 mmHg and his physical examination was otherwise unremarkable. His records showed that the creatinine levels had increased steadily over the past 3 years from 1.6 mg/dL (141 \u03bcmol/L) to 3.3 mg/dL (290 \u03bcmol/L). At the time of referral 3 months later, the creatinine level had further increased to 5.5 mg/dL (487 \u03bcmol/L). Complete blood count was normal except a haemoglobin level of 9.4 g/dL (94 g/L). Serum electrolyte levels were within the normal range. A renal ultrasound and CT-scan revealed kidneys of normal size without any kidney stones. Urinalysis revealed a specific gravity of 1015 and a protein concentration of 0.3 g/L. Microscopic examination revealed multiple spherical brownish crystals that had a pseudo-Maltese cross appearance under polarized light. These crystals raised suspicion for the diagnosis of 2,8-DHA crystalluria (Figure [1](#F1){ref-type=\"fig\"}). Renal biopsy revealed that 76% of the glomeruli were obsolescent while the remaining glomeruli were unremarkable. An H&E stain on a frozen section of the kidney sample revealed a moderate interstitial fibrosis and tubular atrophy with severe inflammation composed principally of lymphocytes, eosinophils and macrophages that surrounded multiple intra-tubular birefringent crystals (Figure [2](#F2){ref-type=\"fig\"}). Using HPLC, the urine concentration of 2,8-DHA was shown to be very high (111 \u03bcmoL/mmoL of creatinine) whereas it is not detectable in healthy controls, providing further support for the diagnosis of 2,8-DHA crystal-induced interstitial nephritis.\n\n![Urine microscopy. Urinary sediment showing spherical brownish crystals (top panel) that have a birefringent and pseudo-Maltese cross appearance under polarized light (lower panel). Original magnification \u00d7400.](sfn110fig1){#F1}\n\n![Renal biopsy. Haematoxylin and eosin staining of a frozen section reveals interstitial inflammation surrounding multiple intra-tubular round brownish crystals that are birefringent under polarized light. Original magnification \u00d7600.](sfn110fig2){#F2}\n\nMeanwhile, the creatinine concentration peaked to as high as 7.5 mg/dL (665 \u03bcmol/L). The patient was treated with allopurinol 300 mg/day and put on a low-purine diet and a high-fluid regimen. The serum creatinine level decreased gradually over the next 2 months to reach a nadir of 2.5 mg/dL (225 \u03bcmol/L). However, due to side effects, the allopurinol dose had to be reduced to 100 mg/day. During the following 2 years, the serum creatinine levels remained stable at 2.2 mg/dL (200 \u03bcmol/L) and a urine examination remained free of 2,8-DHA crystals.\n\nDiscussion {#S3}\n==========\n\nIn the present case report of 2,8-DHA crystal-induced renal failure, the patient had but one episode of kidney stone, at which point his creatinine level was normal. In this particular case, the lack of familiarity with APRT deficiency and the tardy recognition of these unusual but pathognomonic crystals were, at least partly, responsible for the progression of renal failure and delayed diagnosis. Even in Iceland, where the prevalence is the highest and there is a greater awareness of the disease by the medical community, it is still not uncommon to encounter delayed diagnosis of this condition \\[[@R1]\\]. Fortunately, timely administration of high fluid intake, a low-purine diet and use of allopurinol led to a significant improvement of renal function of our patient over the following months.\n\nAPRT deficiency is classified as a rare inherited stone-forming disease that is transmitted in an autosomal recessive fashion. Type I APRT deficiency is seen in Caucasians and is characterized by a complete deficiency of APRT activity, whereas Type II APRT deficiency is primarily seen in the Japanese population and is characterized by a reduction of APRT activity in lysed erythrocytes, ranging from 10 to 25% of normal \\[[@R1],[@R5]\\]. While the heterozygotic forms of the APRT deficiency do not generally lead to symptoms, the homozygote patients invariably present 2,8-dihydroxyadeninuria and often experience multiple episodes of kidney stones. Nevertheless, up to 15% of the patients may remain asymptomatic throughout life. Even though the enzyme deficiency is present at birth, the symptoms related to APRT deficiency may occur from early childhood to the seventh decade of life.\n\nStone analysis using biochemical methods, including colorimetric reaction and thermogravimetric analysis, can misidentify the 2,8-DHA stones as uric acid nephrolithiasis. Although differentiation of the two disorders can easily be made with infrared and UV spectrophotometry or x-ray crystallography of crystals or stones, recognition of the characteristic 2,8-DHA crystals on urinalysis is invaluable for the proper diagnosis.\n\nMore recently, there have been reports of acute renal failure associated with APRT deficiency and 2,8-DHA crystalluria \\[[@R3],[@R6]\\]. In addition, there have also been reports of progressive renal failure in patients with previous history of repeated kidney stones \\[[@R1]\\]. However, the absence of previous multiple kidney stones does not preclude the diagnosis of 2,8-DHA-induced renal failure. Indeed, in our patient there was but one episode of previous kidney stone. This is in keeping with a reported case of 2,8-DHA-induced graft failure in a renal transplant recipient without any prior repeated history of kidney stones \\[[@R4],[@R7]\\]. In these patients, it has been proposed that significant intra-tubular precipitation of 2,8-DHA crystals may lead to chronic tubulointerstitial injury and hence, progressive renal failure \\[[@R8]\\].\n\nAPRT deficiency is a rare disorder and only few cases of renal failure have been reported in the literature. Because of the lack of familiarity with this condition, it is possible that 2,8-DHA crystal-induced renal failure may be an under-diagnosed aetiology of chronic kidney disease \\[[@R9]\\]. In this regard, we believe that the nephrology community should become familiar with this condition and consider such diagnosis in unusual cases of renal failure, as it is easily treatable if diagnosed in a timely manner. Since the urinary 2,8-DHA crystals are usually abundant in untreated patients with 2,8-dihydroxyadeninuria, a urine microscopy as performed by a skilful observer is the single most important diagnostic procedure \\[[@R5]\\]. The presence of 2,8-DHA in urinary sediment can be confirmed by infrared and UV spectrophotometry. The diagnosis may also be confirmed by determination of purine derivative profile showing increased amounts of adenine in both serum and plasma (HPLC) or by measurement of the level of enzymatic activity.\n\nTeaching points {#S4}\n===============\n\n2,8-Dihydroxyadeninuria is a rare inherited stone-forming disease caused by APRT deficiency.APRT deficiency could lead to acute renal failure or progressive kidney disease in adulthood without previous significant history of kidney stones.Recognition of 2,8-DHA crystals in the urine is of utmost importance in diagnosis and treatment of 2,8-DHA crystal-induced renal injury.A simple treatment based on allopurinol administration, high fluid intake and a low-purine diet can reverse or halt renal damage and avoid premature dialysis.\n\nDr M. Agharazii holds a clinician scientist scholarship from *Le Fonds de la Recherche en Sant\u00e9 du Qu\u00e9bec* (FRSQ).\n\n*Conflict of interest statement.* None declared.\n"}
+{"text": "Introduction\n============\n\nBone's resistance to fracture is determined by its structural and material characteristics, which are determined by life-long remodeling of bone by osteoclastic bone resorption and osteoblastic bone formation.[@b1-cia-4-241] The discovery of the receptor activator of nuclear factor *\u03ba*B ligand (RANKL), osteoprotegerin (OPG) and RANK is a major contribution in our understanding of the balanced coupling in space, time and quantity between bone formation and resorption and its disconnection in several bone diseases characterized by increased bone resorption and destruction, such as osteoporosis, rheumatoid arthritis (RA), Paget's disease of bone and metastatic bone disease.[@b2-cia-4-241]\n\nIn this review I focus on the clinical relevance of the RANKL/RANK/OPG pathway in health and disease and the effect of denosumab, a recombinant fully humanized monoclonal RANKL-specific antibody that inhibits bone resorption, in a wide spectrum of diseases. For an extensive list of references on preclinical studies, the reader is referred to a recent in-depth review.[@b2-cia-4-241]\n\nThe RANKL/OPG/RANK pathway: the discovery of how bone formation and resorption are coupled\n==========================================================================================\n\nFor many decades, it has been known that bone resorption and formation in healthy adults are coupled in space, time and quantity,[@b3-cia-4-241] but the exact mechanisms behind this coupling remained for a long time obscure.[@b4-cia-4-241] In the late 1990s, the first component identified for a novel pathway regulating bone remodeling was OPG.[@b5-cia-4-241],[@b6-cia-4-241] OPG was discovered by sequencing random clones from a cDNA library and its gene was shown to encode a novel member of the tumor necrosing factor (TNF) receptor family. Overexpression of the OPG gene in mice resulted in high bone mass and marked reduction in osteoclast number and activity, while OPG knockout (KO) mice had low BMD, increased numbers of osteoclasts, had more woven bone, and developed spontaneous fractures.[@b7-cia-4-241] OPG is a decoy receptor for RANKL and in this way decreases bone resorption. OPG is produced by osteoblasts, endothelial cells, vascular smooth muscle and other cells.\n\nAlready in 1980 it was suggested that osteoblasts might be involved in osteoclastogenesis.[@b8-cia-4-241] The nature of this hypothesized 'osteoclast activating factor' remained elusive, until 1998, when several laboratories independently identified RANKL as a new member of the TNF family of transmembrane and soluble ligands that could bind to OPG[@b9-cia-4-241],[@b10-cia-4-241] and which was one year earlier identified as a product of activated T cells that promotes dentritic cell survival and co-stimulation.[@b11-cia-4-241] RANKL KO mice have high bone mass and virtually no osteoclasts. RANKL binds to RANK and is in this way involved in numerous aspects of osteoclast differentiation and function (fusion, differentiation, attachment to bone, activation and survival). In most instances RANKL relies on macrophagecolony stimulating factor (M-CSF) as a cofactor for osteoclast differentiation, but RANKL can stimulate osteoclastogenesis and bone resorption in mice lacking functional M-CSF and no factor or combination of factors have been shown to be capable of restoring bone resorption when RANKL is absent, indicating the dominant role of RANKL in the regulation of bone resorption.[@b2-cia-4-241] RANKL is produced by cells of the osteoblastic lineage and by activated T cells.\n\nRANK is a cell membrane receptor of the TNF family to which RANKL is bound and is found on dendritic cells and cells of the osteoclast lineage. RANK KO in mice results in the same phenotype as RANKL KO with high bone mass and virtually no osteoclasts.\n\nThe activation of RANK in osteoclasts results in activation of several intracellular signal transduction pathways, which bind the nuclear factor *\u03ba*B (NF *\u03ba*B). After ubiquitination of signal transducers (which results in their degradation by proteasomes), NF *\u03ba*B is released, so that it can translocate to the nucleus, where it upregulates cofactors that induce osteoclastogenic gene transcription (Figure [1](#f1-cia-4-241){ref-type=\"fig\"}).\n\nThe many available data suggest that the RANKL:OPG ratio represents an important determinant of bone resorption.[@b2-cia-4-241] Numerous growth factors, hormones, cytokines and drugs that influence bone turnover have been shown to influence the expression of RANKL and OPG.[@b2-cia-4-241] The RANKL:OPG ratio is decreased by estrogens and increased by glucocorticoids (GC), parathyroid hormone (PTH), PTH-related protein (PTHrP) and prostaglandins. Other molecules and agents that regulate RANKL and/or OPG and that are proresorptive include interleukin-1 (IL-1), IL-17, TNF-\u03b1, 1,25-dihydroxyvitamin D and bone morphogenetic protein 2 (BMP-2). Those which suppress osteoclastogenesis and regulate RANKL and/or OPG are IL-4, IL-13, interferon-\u03b3 (IFN-\u03b3) and transforming growth factor-\u03b2 (TGF-\u03b2).\n\nOPG prevents the ability of proresorptive agents (IL-1, TNF-\u03b1, PTH, PTHrP, vitamin D) to increase bone resorption in mice.[@b2-cia-4-241] In RANK KO mice, IL-1 and TNF-\u03b1 induced the appearance of few if any osteoclasts with no evidence of bone resorption. TNF-\u03b1 was shown to require very low 'permissive' levels of RANKL to stimulate osteoclasts. Thus, whereas TNF-\u03b1, IL-1 and other molecules have certain independent stimulatory effects on osteoclasts, the very existence of those osteoclasts remains dependent on RANKL/RANK signaling.[@b2-cia-4-241]\n\nThe Wnt (wingless) signaling in osteoblasts is also a source of regulation of OPG. Ablation of the intracellular canonic signaling pathway (\u03b2-catenin) in osteoblasts resulted in reduced OPG expression and bone loss, while its over expression resulted in increased OPG expression and high bone mass.[@b12-cia-4-241] Ablation of osteocytes in mice increased bone resorption and the RANKL:OPG ratio.[@b13-cia-4-241] These observations have enabled to integrate the RANKL/RANK/OPG and Wnt signaling pathways in understanding normal and diseased bone.[@b14-cia-4-241]\n\nThe RANKL/OPG/RANK pathway: its role in healthy bone turnover\n=============================================================\n\nThe dominant physiological role of the RANKL/RANK/OPG pathway in the regulation of bone remodeling has been demonstrated in the above mentioned transgenic and KO mice models.\n\nInjection of recombinant RANKL and OPG in mice induced a rapid and marked effect on bone turnover markers (BTM) and number of osteoclasts. One RANKL injection stimulated bone resorption within 1 hour.[@b15-cia-4-241] One OPG injection resulted in decreased BTM within 2 hours and a 50% to 60% reduction in osteoclast numbers within 12 to 24 hours with a long lasting effect which is fully reversible after 1 to 2 months, and which is the result of the long half life and the time required to generate new osteoclasts once the drug is cleared.[@b16-cia-4-241] In small animals, OPG had positive effects on bone mineral content (BMC) and BMD, trabecular bone volume (TBV) and bone strength.[@b2-cia-4-241] Similar positive effects were found with denosumab in monkeys. Bone strength parameters were strongly correlated with bone mass parameters, suggesting that RANKL inhibition improved bone strength primarily by increasing bone mass and that bone formed during anti-RANKL treatment has normal material properties. RANKL inhibition in animals also prevented bone loss after ovariectomy, orchidectomy, GC use, and disuse.[@b2-cia-4-241]\n\nThe discovery of the RANKL/OPG/RANK pathway and these experiments have elucidated our understanding of cellular and mechanisms of bone remodeling. During normal bone remodeling, osteoclast precursors derived from the circulation or bone marrow develop a RANKL/RANK driven cell contact with osteoblasts. This enables osteoclast precursors to differentiate to mature osteoclasts and activates their bone resorption capacity by attaching to bone, polarizing, becoming mobile, expressing the calcitonin receptor and tartrate-resistant alkaline phosphatase (TRAP), producing H^+^ by carbonic anhydrase II, secretion of H^+^ by the osteoclast-specific vacuolar ATP-ase pump and Cl^\u2212^ by the Cl^\u2212^ channel to form HCL that degrades bone mineral, lysosomal release of cathepsin-K and matrix metalloproteinases (MMPs) that degrade bone matrix proteins, and by contributing to osteoclast mobility and survival.[@b17-cia-4-241]\n\nNon-osseous effects of the RANKL/OPG/RANK pathway\n=================================================\n\nThe above-mentioned OPG, RANKL and RANK KO animal models revealed significant extraskeletal manifestations with regard to vascular calcifications, the immune system and mammary gland development.\n\nOPG KO animals developed medial calcification of the aorta and renal arteries,[@b7-cia-4-241] which could be prevented by transgenic over expression of soluble OPG. RANKL and OPG have been identified in human atherosclerotic plaques.[@b18-cia-4-241] The mechanism by which OPG could serve a protective role in vascular disease is unclear. Experiments in animal models of atherosclerosis indicate that the prevention of vascular calcification is probably related to the suppression of bone resorption, with neutral effects on existing atherosclerotic changes.[@b2-cia-4-241]\n\nRANK and RANKL KO results in lymph node agenesis, indicating the role of the RANKL/RANK/OPG pathway development and function of the immune system.[@b2-cia-4-241] As mentioned earlier, activated T cells produce RANKL and thereby regulate the function and survival of dendritic cells.[@b11-cia-4-241] However, RANKL inhibition did not lead to dendritic cell suppression, probably due to the central role of CD40 cross-talk between T cells and dendritic cells, which compensates for the total absence of RANKL/RANK in KO mice.[@b2-cia-4-241] RANKL and RANK are found in bone erosions of RA, rendering RANKL a possible target to prevent bone erosions in RA.\n\nRANKL/RANK KO mice have a defect in mammary gland development during pregnancy and lactation.[@b2-cia-4-241] However, OPG transgenic mice did not show failure of lactation, while bone resorption was inhibited.[@b2-cia-4-241]\n\nThe RANKL/OPG/RANK pathway in osteolytic and destructive bone diseases\n======================================================================\n\nThe RANKL/OPG/RANK pathway plays a crucial and consistent role in a wide spectrum of diseases of bone characterized by increased bone resorption, disturbed coupling between bone formation and resorption, and bone destruction.\n\nMutations of the genes for RANKL, OPG and intracellular signal transducers of NF *\u03ba*B have been identified in diseases characterized by focal disturbed bone remodeling, such as juvenile and classical Paget's disease of bone, familial expansile osteolysis, expansile skeletal hyperphosphatasia and inclusion body myopathy combined with Paget's disease and frontotemporal demantion (IBMPFD).[@b19-cia-4-241]\n\nRheumatoid arthritis (RA) is characterized by joint inflammation and bone erosions, periarticular and generalized bone loss. Bone erosions were already described in 1878 as 'caries of the joints,'[@b20-cia-4-241] but only during the last decades it has been shown that these erosions contain multinucleated cells which were identified as osteoclasts.[@b21-cia-4-241]--[@b23-cia-4-241] Animal models of RA (collagen-induced arthritis (CIA), adjuvant-induced arthritis (AIA) and TNF-mediated arthritis) indicated early upregulation of RANKL and suppression of OPG in inflamed joints. RANKL KO and OPG prevented formation of erosions, without effect on inflammation.[@b2-cia-4-241] This indicates that bone destruction and inflammation can be dissociated on the basis of specific inhibition of the RANKL/RANK pathway. OPG also prevented periarticular and generalized bone loss in such animal models[@b2-cia-4-241] and preserved cartilage in several arthritic animal models.[@b2-cia-4-241] RANKL, RANK and OPG are expressed in human articular cartilage, but their functional roles have not been clearly demonstrated, and RANK is probably inactive in chondrocytes.[@b2-cia-4-241] A likely mechanism for chondroprotection by OPG is to preserve subchondral bone against subchondral osteoclast invasion.[@b2-cia-4-241]\n\nMore than 100 years ago it had been observed that bone was a fertile soil for the localization and growth of particular types of cancer cells.[@b24-cia-4-241] Recently it has become clear that there is a bidirectional communication between tumor cells and bone cells, resulting in a vicious circle of bone destruction and tumor growth.[@b2-cia-4-241],[@b25-cia-4-241],[@b26-cia-4-241] In metastatic bone disease, cancer cells are able to increase the RANKL: OPG ratio directly or by the help of T-cells, osteoblast/stromal cells and endothelial cells, along with the production of other osteoclast mediators such as PTHrP. This stimulates osteoclasts to remove bone, enabling cancer cells to make space for further growth within bone,[@b25-cia-4-241],[@b26-cia-4-241] and releasing growth factors from bone that further stimulate tumor growth. Multiple myeloma cells are even more destructive for bone, as they release not only RANKL, but also dickkopff (DKK), which suppresses bone formation, ampliflying in this way room for tumor growth.[@b26-cia-4-241]\n\nDenosumab: the effect of RANKL inhibition in osteoporosis, RA and metastatic bone disease\n=========================================================================================\n\nDenosumab is a fully human monoclonal antibody with a high affinity and specificity for RANKL and binds and neutralizes the activity of human RANKL in a similar fashion to the action of OPG.[@b2-cia-4-241] Denosumab does not cross-react with TNF-\u03b1, TNF-\u03b2, TNF-related apoptosis-inducing ligand (TRAIL), or CD40 ligand.\n\nPostmenopausal osteoporosis\n---------------------------\n\nThe first clinical study on RANKL inhibition was performed in postmenopausal women with BMD-osteoporosis using OPG bound to the constant fragment (Fc) of human IgG1 (Fc-OPG), which showed a dose-dependent decrease of markers of bone resorption within 4 days with a maximum decrease of 80% with the highest dose and lasting 45 days, without changes in markers of bone formation.[@b27-cia-4-241] However, Fc-OPG was not further studied because denosumab was superior in reducing bone resorption at lower dose and with longer duration,[@b24-cia-4-241] antibodies to Fc-OPG could be raised[@b24-cia-4-241] and because of the concern that Fc-OPG could bind to TRAIL, which could inhibit its role in tumor surveillance.[@b28-cia-4-241]\n\nDenosumab has been studied in postmenopausal women in the context of its effect on bone remodeling, BMD, parameters of bone strength in the hip region, and safety.\n\nIn a randomized, placebo-controlled phase-1 study in 49 healthy postmenopausal women, a single SC injection of denosumab caused a dose-dependent rapid (within 12 hours), profound (up to 84%) and sustained (up to 6 months) decrease in bone resorption (reflected by urinary N-telopeptide (NTX)) and was well tolerated.[@b28-cia-4-241] Bone formation (as reflected by bone-specific alkaline phosphatase (BALP)) did not decrease remarkably until after 1 month, indicating that the effect of denosumab is primarily antiresorptive. Intact PTH levels increased up to 3-fold after 4 days, but returned towards baseline with follow-up. This study indicated that denosumab effectively decreases bone resorption in humans and that long intervals between doses of denosumab might be possible.\n\nEfficacy and safety of denosumab were evaluated in a randomized, placebo-controlled, dose-ranging phase 2 study in 412 postmenopausal women with low BMD in the spine or femur and included a group of patients treated with open-label alendronate 70 mg/week.[@b29-cia-4-241] Denosumab (at a dose of 6, 14 or 30 mg every 3 months or 14, 60, 100 or 210 mg every 6 months over a period of 12 months) resulted in a rapid and sustained effect on BTM and a rapid increase in BMD in the spine, hip and distal radius, which was superior to placebo and similar or greater (for changes in BMD in the total hip and in the distal radius and in markers of bone resorption) than with weekly 70 mg of alendronate (Figure [2](#f2-cia-4-241){ref-type=\"fig\"}). These changes were sustained in an extension study over 24 months (Figure [3](#f3-cia-4-241){ref-type=\"fig\"}).[@b30-cia-4-241] Adverse events (AE) and serious AE (SAEs) were similar in character and percentage with denosumab compared with placebo. In terms of increase in BMD, the dose of 30 mg/3 months and 60 mg/6 months appeared optimal.\n\nIn a post hoc analysis of this study in a subgroup of 116 patients treated with denosumab 60 mg every 6 months, weekly alendronate 70 mg or placebo, structural geometry of the hip was evaluated on DXA images, and derived strength indices were calculated, including bone cross-sectional area, section modulus and buckling ratio.[@b31-cia-4-241] At 12 and 24 months denosumab and alendronate improved these parameters compared with placebo. Denosumab effects were greater than alendronate at the intertrochanteric and femoral shaft sites. These results suggested that denosumab treatment may lead to improved bone biomechanical properties, similar or greater than with weekly alendronate.\n\nThe phase 2 study was continued for an additional 24 months to study the long-term efficacy and safety of denosumab and the effects of discontinuing and restarting different doses of denosumab treatment.[@b32-cia-4-241] With continuous treatment, denosumab led to further gains in BMD at the lumbar spine, total hip and distal 1/3 radius and sustained reduction in BTM throughout the course of the study. The effects on BTM were fully reversible with discontinuation and restored with subsequent retreatment (Figure [4](#f4-cia-4-241){ref-type=\"fig\"}).\n\nIn a double-blind phase 3 study, including 1180 postmenopausal women with low BMD (T-score \u2264 2.0 in the lumbar spine or total hip), the efficacy and safety of denosumab 60 mg every 6 months and alendronate 70 mg weekly during 12 months treatment were compared.[@b33-cia-4-241] At the total hip, denosumab significantly increased BMD compared to alendronate (2.5% versus 2.6%, p \\< 0.0001). Furthermore, significantly greater increases in BMD were observed with denosumab at all measured skeletal sites (12-month treatment difference: 1.1% in the lumbar spine, 0.6% in the femoral neck, 1.0% in the trochanter and 0.6% in the 1/3 radius; p \\< 0.001 at all sites). Compared to alendronate, treatment with denosumab resulted in greater reductions in BTM (serum C-telopeptide-1 \\[sCTX1\\] for bone resorption and P1NP for bone formation) at each time point assessed through month 9 for sCTX1 and month 12 for P1NP. Maximum reduction of P1NP in the alendronate group was reached at month 3 and remained constant throughout the study (Figure [5](#f5-cia-4-241){ref-type=\"fig\"}). In contrast, maximal reduction of sCTX1 in the denosumab group was observed at 1 month, with attenuated reduction in sCTX1 at the end of the 6-month dosing interval. AEs were similar for denosumab- and alendronate-treated subjects. SAEs were similar between denosumab and alendronate for infections (1.5% versus 1.0%, p = 0.61) and malignant neoplasm (1.0% versus 0.9%, p = 1.00).\n\nSimilar changes in BMD and BTM were reported in a 2-year randomized, double-blind, placebo-controlled study in 332 postmenopausal women with BMD-osteopenia treated with the same dose of denosumab (60 mg every 6 month).[@b34-cia-4-241] SEAs for infections were significantly higher in the denosumab group (4.9%) then with placebo (0.6%, p \\< 0.05) and similar for neoplasm (2.4% with denosumab versus 0.6% with placebo, p = 0.215).\n\nMeanwhile, the pivotal fracture trial (Fracture REduction Evaluation of Denosumab in Osteoporosis every 6 Months, the 'FREEDOM' study) with fracture reduction as a primary endpoint has been accomplished in 7,808 postmenopausal women with BMD-osteoporosis,[@b35-cia-4-241] and a peer-reviewed publication is expected soon.\n\nRheumatoid arthritis\n--------------------\n\nIn animal models of arthritis (CIA and AIA) RANKL expression has been demonstrated within 1 to 14 days of the appearance of joint inflammation.[@b2-cia-4-241] OPG fully prevented the occurrence of bone erosions, without effect on cartilage loss and inflammation. In patients with RA, RANKL expression has been documented in erosions. Furthermore, baseline serum levels of the RANKL:OPG ratio predicted future joint damage independent of other predictors of joint damage in patients with early untreated RA.[@b36-cia-4-241]\n\nIn a placebo-controlled study in 227 patients with RA taking methotrexate, 71 were treated with denosumab 60 mg, 72 with 180 mg denosumab and 75 had placebo injections every 6 months during 12 months.[@b37-cia-4-241] At 6 months, the increase in MRI erosion score from baseline was significant lower in the 180-mg denosumab group than in the placebo group (0.06 versus 1.75, p = 0.007). A significant difference in the modified Sharp erosion score was observed as early as 6 months in the 180-mg denosumab group (0.05) compared with placebo (0.59, p = 0.019), and at 12 months, both the 60-mg (0.33) and 180-mg (0.19) denosumab groups were significantly different from the placebo group (1.34, p = 0.012 and p = 0.007, respectively). Denosumab caused sustained suppression of BTM. There was no evidence of an effect of denosumab on joint space narrowing or measure of disease activity. Rates of AE and SAE were comparable between the denosumab and placebo groups.\n\nMetastatic bone disease\n-----------------------\n\nDenosumab has been studied in patients with breast and prostate cancer and multiple myeloma.[@b38-cia-4-241]--[@b42-cia-4-241] Denosumab was well tolerated and decreased markers of bone resorption, even in patients previously resistant to bisphosphonates (BP) in reducing bone resorption.\n\nIn patients with multiple myeloma (n = 25) or bone metastases from breast cancer (n = 29), a single SC dose of denosumab reduced bone resorption for at least 84 days and was well tolerated.[@b38-cia-4-241] The decrease in bone turnover markers was similar in magnitude but more sustained than with IV pamidronate.\n\nIn 255 women with breast cancer-related bone metastases, subcutaneous denosumab was similar to IV BP in suppressing bone turnover and reducing skeletal-related events risk.[@b39-cia-4-241] The safety profile was consistent with an advanced breast cancer population receiving systemic therapy.\n\nIn women with non-metastatic breast cancer and low bone mass who were receiving adjuvant aromatase inhibitor therapy, twice-yearly administration of denosumab significantly increased BMD over 24 months at trabecular and cortical bone, with overall AE rates similar to those of placebo.[@b40-cia-4-241]\n\nA next study evaluated the effects of denosumab in IV BP-na\u00efve patients with breast cancer-related bone metastases. Denosumab suppressed bone turnover and seemed to reduce SRE risk similarly to IV BPs, with a safety profile consistent with an advanced cancer population receiving systemic therapy.[@b41-cia-4-241]\n\nSafety of denosumab\n===================\n\nFracture repair\n---------------\n\nIn studies of the effect of RANKL inhibition on fracture repair, high doses of OPG given to rats at the time of fracture did not influence the formation of a normal fracture callus, but delayed callus remodeling.[@b42-cia-4-241] In an animal study of transverse femur fracture healing, alendronate and denosumab delayed the removal of cartilage and the remodeling of the fracture callus.[@b43-cia-4-241],[@b44-cia-4-241] However, this did not diminish the mechanical integrity of the healing fractures in mice receiving these treatments. In contrast, strength and stiffness were enhanced in these treatment groups when compared to control bones.[@b44-cia-4-241]\n\nImmunosuppression\n-----------------\n\nAs mentioned earlier, the RANKL/OPG/RANK pathway is involved in the development of the immune system. Therefore, safety issues about infection and neoplasm deserve special attention when therapeutically interfering with this pathway.[@b2-cia-4-241]\n\nOPG did not alter cellular or humoral immunity, or render mice more susceptible to bacterial challenge at doses that inhibited bone resorption.[@b2-cia-4-241] Furthermore, OPG prevented bone erosions and bone loss without affecting the degree of inflammation.[@b2-cia-4-241]\n\nIn the published clinical trials on osteoporosis, no significant differences in infections reported as SAE were found between denosumab, placebo and alendronate,[@b28-cia-4-241]--[@b33-cia-4-241] except in one study in 332 subjects with osteopenia (4.9% with denosumab versus 0.6% with placebo, p \\< 0.05).[@b34-cia-4-241] The six SAEs of infections associated with denosumab in the 24-month phase-2 study were common community-acquired infections that were successfully treated with standard antibiotics during uncomplicated hospital courses.[@b32-cia-4-241] The rate of infections remained unchanged from year 1 to year 2 in the denosumab group.[@b32-cia-4-241]\n\nNo significant differences in neoplasms reported as SAE were found between denosumab, placebo and BP-treated patients.[@b38-cia-4-241]--[@b41-cia-4-241]\n\nOsteonecrosis of the jaw\n------------------------\n\nOsteonecrosis of the jaw (ONJ) is a potential side effect of BPs, especially when given IV in high doses.[@b45-cia-4-241] The incidence of ONJ in patients using oral BPs for osteoporosis is low, and associated with other risks for ONJ, such as infection, and possibly other, not yet well defined risks and co-morbidities.[@b45-cia-4-241] Up until now, ONJ has not been reported in the clinical trials with denosumab.\n\nComparison of denosumab with bisphosphonates\n============================================\n\nThe effect of RANKL inhibition is quite unique among antiresorptive agents. RANKL inhibition differed from the effects of BPs on bone in several aspects.\n\nFirst, OPG, at least in animal models, does not damage the osteoclasts but decreases their number while BPs inhibit the function of osteoclasts by damaging the cytoskeleton, without affecting or even increasing the number of osteoclasts.[@b29-cia-4-241],[@b46-cia-4-241],[@b47-cia-4-241] Second, denosumab resulted in a quicker decrease in markers of bone resorption than alendronate.[@b29-cia-4-241] Third, this effect on bone resorption was more pronounced than with alendronate.[@b29-cia-4-241] Fourth, the effect of denosumab is completely reversible and cleared over a relatively short period of time, in contrast to the long term effects of BPs.[@b28-cia-4-241] The complete reversibility of bone remodeling after denosumab could result in a quicker response of changes in BMD than BPs when anabolic agents such as PTH are subsequently given.[@b48-cia-4-241] Fifth, denosumab effects were greater than alendronate on BMD and on strength indices derived from hip geometry evaluated by DXA at the intertrochanteric and femoral shaft sites.[@b31-cia-4-241] Sixth, patient preference could be different between denosumab 60 mg every 6 months and weekly bisphosphonates.[@b49-cia-4-241] Lastly, denosumab is the first anti-bone resorbing agent that has been shown to halt bone erosions in RA.[@b37-cia-4-241] Although bisphosphonates are successful in preventing erosions in animal models of inflammatory arthritis, support from studies in humans is lacking,[@b50-cia-4-241]--[@b52-cia-4-241] with the exception of some effect of zoledronate that showed slowing in the number of erosions.[@b53-cia-4-241]\n\nConclusions\n===========\n\nThe RANKL/RANK/OPG pathway plays a central role in coupling bone formation and resorption during normal bone turnover and a central and common role in a wide spectrum of diseases characterized by disturbed bone remodeling and increased bone resorption and destruction (osteoporosis, Paget's disease of bone, RA, metastatic bone disease). Clinical trials in postmenopausal women indicate that denosumab, a RANKL-specific recombinant humanized monoclonal antibody, is effective in suppressing bone resorption resulting in increase in BMD in osteoporosis, and has the potential to prevent progression of erosions in RA and of skeletal-related events in metastatic bone disease. The effects on fracture reduction in osteoporosis are awaited from the recently finished FREEDOM study. In clinical trials with denosumab, overall AEs and SAEs were similar to placebo, indicating a favorable safety profile in these diseases, but data on long-term safety will be needed.\n\n**Disclosures**\n\nThe author discloses no conflicts of interest.\n\n![Free RANKL (ie, not bound by osteoprotegerin \\[OPG\\]) binds to the transmembrane RANK receptor, which upregulates intracellular signal transducers which are involved in cytoskeletal organization, cell motility, growth and survival, and some also bind NF *\u03ba*B. After ubiquitination, signal transducers are released from NF *\u03ba*B and degraded by proteasomes. NF *\u03ba*B can than migrate to the nucleus, were it upregulates transcriptional regulators that start osteoclastogenic gene transcription.[@b2-cia-4-241]](cia-4-241f1){#f1-cia-4-241}\n\n![Phase 2 study of denosumab every 6 months in postmenopausal women with low (bone mineral density) BMD: lumbar spine, total hip, and distal 1/3 radius BMD at 12 months. Adapted from *J Bone Miner Res*. 2007;22:1832--1841,[@b30-cia-4-241] with permission of the American Society for Bone and Mineral Research; and from *N Engl J Med*. 2006;354:821--831.[@b29-cia-4-241]](cia-4-241f2){#f2-cia-4-241}\n\n![Effect of 4 years of denosumab every 6 months on lumbar spine (LS) bone mineral density. Adapted from *J Bone Miner Res*. 2007;22:1832--1841,[@b30-cia-4-241] with permission of the American Society for Bone and Mineral Research.](cia-4-241f3){#f3-cia-4-241}\n\n![Effect of denosumab re-treatment and changes to lumbar spine (LS) and total hip bone mineral density. Reproduced with permission from Miller PD, Bolognese MA, Lewiecki EM, et al. Amg Bone Loss Study Group. Effect of denosumab on bone density and turnover in postmenopausal women with low bone mass after long-term continued, discontinued, and restarting of therapy: a randomized blinded phase 2 clinical trial. *Bone*. 2008;43(2):222--229.[@b32-cia-4-241] Copyright \u00a9 2008 Elsevier.](cia-4-241f4){#f4-cia-4-241}\n\n![Effect of denusomab vs alendronate head to head trial on bone markers. Reproduced from *J Bone Miner Res*. 2009;24:153--161,[@b33-cia-4-241] with permission of the American Society for Bone and Mineral Research.\\\n**Notes:** ^a^significantly different from alendronate, p \u2264 0.0001.](cia-4-241f5){#f5-cia-4-241}\n"}
+{"text": "Introduction {#s1}\n============\n\nIntegrins are cell adhesion receptors that regulate cell migration, cytoskeletal remodeling, and gene expression [@pone.0055184-Huttenlocher1], [@pone.0055184-Delon1], [@pone.0055184-Streuli1]. In humans, 24 integrins are formed by specific non-covalent pairing of 18 \u03b1 and 8 \u03b2 subunits [@pone.0055184-Hynes1]. Each subunit has a large extracellular region that is involved in ligand-binding and a single-pass transmembrane segment for the transmission of allostery across the cell's plasma membrane. This is followed by a short cytosolic tail (CT) except that of the integrin \u03b24 subunit [@pone.0055184-Tan1]. Integrin CTs associate with cytoskeletal, adaptor, and signaling proteins, which allow cells to communicate extracellular biochemical and mechanical signals with intracellular signaling pathways [@pone.0055184-Hynes1], [@pone.0055184-PuklinFaucher1], [@pone.0055184-Boettiger1].\n\nIntegrin \u03b14\u03b21 (CD49dCD29; very late activation antigen, VLA-4) is expressed abundantly on leukocytes except neutrophils. The other leukocyte integrin having the same \u03b1 subunit is \u03b14\u03b27. Integrin \u03b14\u03b21 binds to the alternatively spliced connecting segment -1 (CS-1) in fibronectin, activated endothelium-expressed vascular cell adhesion molecule-1 (VCAM-1), and osteopontin [@pone.0055184-Mould1], [@pone.0055184-Elices1], [@pone.0055184-Barry1]. In addition to fibronectin and VCAM-1, integrin \u03b14\u03b27 binds mucosal addressin cell adhesion molecule-1 (MadCAM-1) that is expressed on high endothelial venules of Peyer's patches and in gut-associated lymphoid tissues, allowing the targeting of lymphocyte subsets to these sites [@pone.0055184-Berlin1], [@pone.0055184-Briskin1]. Apart from the widely reported \u03b22 integrins [@pone.0055184-Tan1], both \u03b14 integrins mediate rolling and firm adhesion of lymphocytes on endothelium [@pone.0055184-Berlin2], [@pone.0055184-Alon1]. VCAM-1-engaged integrin \u03b14\u03b21was shown to up-regulate integrin \u03b1L\u03b22-mediated leukocyte adhesion, suggesting crosstalk between integrins [@pone.0055184-Chan1], [@pone.0055184-May1]. The importance of \u03b14 integrins is also underscored by embryonic lethality observed in mice that were homozygous for integrin \u03b14 gene ablation [@pone.0055184-Yang1]. Subsequently, the use of chimeric mice provided evidence that \u03b14 integrins are also essential for the normal development of T and B lymphocytes in the bone marrow [@pone.0055184-Arroyo1]. Hence, \u03b14 integrins are attractive targets for the development of therapeutics for inflammatory diseases. The drug Natalizumab, which is a humanized function-blocking mAb that binds the \u03b14 subunit, has been used for the treatment of autoimmune diseases such as multiple sclerosis and Crohn disease [@pone.0055184-Sandborn1], [@pone.0055184-Rice1].\n\nIntegrin \u03b14\u03b21 mediates chemotactic and haptotatic cell migration on VCAM-1 whereas replacing the \u03b14 CT with that of either integrin \u03b12 or \u03b15 induces focal complex formation with a concomitant increase in the strength of cell adhesion [@pone.0055184-Kassner1]. Hence, intracellular signaling events derived from integrin \u03b14\u03b21and other \u03b21 integrins are different even though they have a common \u03b21 subunit, suggesting the importance of the \u03b1 subunits in integrin signaling. A seminal report by Liu et al., identified \u03b14 CT, but not CTs of \u03b1IIb, \u03b13A, \u03b15, \u03b16 and \u03b21 integrins, as a binding partner of the adaptor protein paxillin [@pone.0055184-Liu1]. Using fragments of integrin \u03b14 CT and paxillin, the interaction sites were mapped to E983-Y991 in \u03b14 and A176-D275 in paxillin [@pone.0055184-Liu2], [@pone.0055184-Liu3]. Interestingly, integrin \u03b19 CT has also been shown to interact with paxillin possibly because of the sequence homology between \u03b14 and \u03b19 CTs [@pone.0055184-Liu4], [@pone.0055184-Young1].\n\nPaxillin is a widely expressed 68-kDa adaptor protein that contains five leucine-rich LD repeats and four LIM domains in its N- and C-terminal halves, respectively. Its LIM3 and LIM4 domains have been shown to interact with protein tyrosine phosphatase (PTP)-PEST [@pone.0055184-Shen1], [@pone.0055184-Cote1]. PTP-PEST regulates the activity of p130Cas (Crk-associated-substrate) that is involved in adhesion mechano-sensing and cell migration [@pone.0055184-Garton1], [@pone.0055184-Sawada1]. The N-terminal region of paxillin that contains the LD repeats supports the binding of many proteins, including Src, Csk, vinculin, focal adhesion kinase (FAK) and proline rich tyrosine kinase 2 (Pyk2) [@pone.0055184-Turner1], [@pone.0055184-Rose1]. The stretch of amino acids A176-D725 in paxillin that binds integrin \u03b14 CT encompasses the LD3 and LD4 repeats [@pone.0055184-Liu3].\n\nPaxillin-integrin \u03b14 and -integrin \u03b19 interactions inhibit cell spreading and lamellipodia formation [@pone.0055184-Liu1], [@pone.0055184-Young1], [@pone.0055184-Nishiya1]. Mutating Tyr991 to Ala in the \u03b14 CT disrupts its interaction with paxillin and Jurkat T cells expressing this mutated integrin showed extensive spreading on VCAM-1 [@pone.0055184-Liu1]. Transgenic mice homozygous for \u03b14 Y991A had reduced number of Peyer's patches and exhibited poor recruitment of leukocytes in thioglycollate-elicited peritonitis compared with wild-type mice [@pone.0055184-Feral1]. Paxillin binding to \u03b14 CT is also regulated by post-translational modification of the latter. Phosphorylation of Ser988 in \u03b14 CT inhibits its binding to paxillin [@pone.0055184-Han1]. Notably, \u03b14 CT with pSer988 was reported to be lacking at the trailing region of migrating cells [@pone.0055184-Goldfinger1]. Taken together, these data establish a role of paxillin in regulating adhesion sites turnover that is critical in cell migration [@pone.0055184-Rose1].\n\nBased on high throughput screening, a small molecule has been identified to disrupt the \u03b14 CT-paxillin interaction and it reduced leukocyte recruitment to sites of inflammation in mice [@pone.0055184-Kummer1]. Conceivably, structural determination of \u03b14 CT-paxillin interaction will provide valuable information to understand how the interaction is regulated and pave the way for the development of novel therapeutics. In this study, we determine the conformation of the full-length integrin \u03b14 CT in aqueous solution and analyze its interactions with recombinant paxillin (residues G139-K277 encompassing LD2-LD4) and synthetic peptides containing LD2, LD3 or LD4. We provide evidence that the C-terminal region of integrin \u03b14 CT adopts a helical conformation and it is involved in binding to the LD3 and LD4 repeats of paxillin.\n\nResults and Discussion {#s2}\n======================\n\nNMR Studies of Integrin \u03b14 CT {#s2a}\n-----------------------------\n\n3-D structures of the CTs of \u03b1IIb, \u03b1M and \u03b1X integrins were determined by NMR spectroscopy in lipid micelles by having a myristoyl chain covalently-linked to the N-terminus of each CT [@pone.0055184-Vinogradova1], [@pone.0055184-Chua1], [@pone.0055184-Chua2]. The micelle anchoring property of the myristoyl group that mimics the transmembrane segment imparts conformational stabilization to the \u03b1 CTs [@pone.0055184-Vinogradova1], [@pone.0055184-Chua1], [@pone.0055184-Chua2]. We have attempted a similar strategy to solve the NMR structure of \u03b14 CT in DPC lipid micelles. However, NMR spectra of the myristoylated \u03b14 CT was found to be extremely broad, precluding structural characterization under such conditions (data not shown). It is noteworthy that apart from the conserved membrane proximal GFFKR motif, the primary structure of \u03b14 CT is unique from other \u03b1 CTs (Supplementary [Figure S1](#pone.0055184.s001){ref-type=\"supplementary-material\"}). Interestingly, \u03b14 CT shows well dispersed NMR spectra in free aqueous solution. [Figure 1](#pone-0055184-g001){ref-type=\"fig\"} shows a section of the 2-D NOESY spectrum of \u03b14 CT, at 278 K, delineating NOE connectivites among the down-field shifted (6.5--9.0 ppm), amide and aromatic, proton resonances with the up-field shifted (0.9--4.5 ppm) aliphatic proton resonances. NOE connectivities were also detected from the N^\u03b5^H proton of the residue W22, resonating at 10.2 ppm, with the aliphatic protons ([Figure 1, left panel](#pone-0055184-g001){ref-type=\"fig\"}). These NOE cross-peaks potentially indicate populated folded conformations of the \u03b14 CT in aqueous solution. The sequence-specific resonance assignments of \u03b14 CT was achieved by combined analyses of 2-D TOCSY and 2-D NOESY spectra. The secondary chemical shifts or deviation from random coil values of C^\u03b1^H resonances of each amino acid of \u03b14 CT are shown in [Figure 2](#pone-0055184-g002){ref-type=\"fig\"}. Helical conformations appear to be populated for the stretch of C-terminus residues, L13-S27, and membrane proximal N-terminal segment, residues G3-R7, of \u03b14 CT as indicated by the negative deviation for C^\u03b1^H chemical shifts. The secondary chemical shifts are less pronounced for other C-terminal residues including K28-D32, indicating a lack of preferred secondary conformations ([Figure 2](#pone-0055184-g002){ref-type=\"fig\"}). Further, analyses of 2-D NOESY spectra of \u03b14 CT revealed sequential and medium range NOEs (C^\u03b1^H/NH: i to i+2, i+3 and i+4) diagnostic of helical conformations for most of the residues, L13-N30, at the C-terminus ([Figure 3](#pone-0055184-g003){ref-type=\"fig\"}). Helical type medium range NOEs were also detected for the N-terminal membrane proximal region.\n\n![Folded conformations of \u03b14 CT.\\\nA section of the two-dimensional ^1^H-^1^H NOESY spectrum of \u03b14 CT in aqueous solution showing NOE contacts among low-field resonances (6.5 ppm-9.0 ppm) with up-field resonances (0.8 ppm- 4.5 ppm) (right panel). NOE contacts from the low-field shifted N^\u03b5^H resonance, 10.2 ppm, of the residue W22 with backbone and sidechain aliphatic proton resonances (left panel). NOESY spectra were acquired in aqueous solution, pH 5.6, 278 K.](pone.0055184.g001){#pone-0055184-g001}\n\n![Secondary structure of the \u03b14 CT from chemical shift deviations.\\\nBar diagrams representing deviation of C\u03b1H chemical shifts from random coil values for amino acid residues of \u03b14 CT in aqueous solution.](pone.0055184.g002){#pone-0055184-g002}\n\n![Summary of NOE contacts of the \u03b14 CT.\\\nBar diagram summarizing type (intra, sequential, medium-range) and number of NOE contacts observed for each amino acid of the \u03b14 CT in aqueous solution.](pone.0055184.g003){#pone-0055184-g003}\n\nThree-dimensional Structure of Integrin \u03b14 CT {#s2b}\n---------------------------------------------\n\nAn ensemble of conformations of \u03b14 CT were obtained from 300 NOE driven distance constraints based on 107 intra-residue, 95 sequential and 98 medium range NOEs ([Table 1](#pone-0055184-t001){ref-type=\"table\"}). [Figure 4](#pone-0055184-g004){ref-type=\"fig\"} shows superposition of the backbone atoms of the structural ensemble of \u03b14 CT for residues A2-D31 (panel A), residues A2-L13 (panel B) and residues Q14-D31 (panels C-D). Higher root mean square deviation (RMSD) values for the backbone and all heavy atoms can be seen for the superposed structures while including residues A2-D31 of \u03b14 CT ([Figure 4A](#pone-0055184-g004){ref-type=\"fig\"}, [Table 1](#pone-0055184-t001){ref-type=\"table\"}). However, RMSD values are found to be lower for the individual segment of the N- and C-termini of the molecule ([Table 1](#pone-0055184-t001){ref-type=\"table\"}). Notably, the C-terminal region, residues Q14-D31, of \u03b14 CT demonstrates a well-defined backbone and side-chain topology in comparison to the N-terminal region ([Table 1](#pone-0055184-t001){ref-type=\"table\"}, [Figure 4](#pone-0055184-g004){ref-type=\"fig\"}). The \u03b14 CT assumes a bend or kinked helical conformation in free solution ([Figure 5](#pone-0055184-g005){ref-type=\"fig\"}). The membrane proximal region demarcates a conserved helical conformation that is connected to the C-terminal helix through a bend formed by residues S11-L13. The present study reveals a different conformation of the \u03b14 CT compared with CTs of other integrins. Three-dimensional structures of the CTs of \u03b1IIb (20-residue), \u03b1M (24-residue) and \u03b1X (35-residue) are characterized by an N-terminus membrane proximal helix followed by a C-terminal loop [@pone.0055184-Vinogradova1], [@pone.0055184-Chua1], [@pone.0055184-Chua2]. The tertiary topologies of these \u03b1 CTs are stabilized by long-range packing between the loop and the N-terminal helix. On the other hand, the longer CT (57-residue) of \u03b1L integrin assumes a packed structure consisting of three helices [@pone.0055184-Bhunia1]. By contrast, the C-terminal region of \u03b14 CT adopts a helical conformation that does not show any long-range packing interactions with the membrane proximal helix ([Figure 5](#pone-0055184-g005){ref-type=\"fig\"}). The C-terminal helix of \u03b14 CT can potentially be stabilized by a number of polar interactions, ionic and/or hydrogen bonds, by the sidechains of residues E16-R19, K28-D31, N17-S21, D20-S23 and N26-S29 ([Figure 5](#pone-0055184-g005){ref-type=\"fig\"}). In addition, the indole ring of residue W22 is in a close proximity with the aliphatic sidechain of residue I25 and guanidinium sidechain of residue R18, implying probable cation-\u03c0 and/or non-polar packing interactions. The helical structure of \u03b14 CT has patches of negatively and positively charged surfaces for its C-terminal region, whereas the N-terminal region is largely positively charged ([Figure 6](#pone-0055184-g006){ref-type=\"fig\"}).\n\n![Three-dimensional structure of the \u03b14 CT.\\\n(A) Superposition of backbone atoms (N, C\u03b1, C\u2032) of twenty lowest energy conformers of the \u03b14 CT for residues A2-D31. (B) Superposition of backbone atoms (N, C\u03b1, C\u2032) of twenty lowest energy conformers of the \u03b14 CT for N-terminal residues A2-L13. (C-D) Superposition of backbone atoms (N, C\u03b1, C\u2032) of twenty lowest energy conformers of the \u03b14 CT for C-terminal residues Q14-D31. Superposition of the sidechains of residues Q14-D31 are also shown (D). Figures were generated by INSIGHT II.](pone.0055184.g004){#pone-0055184-g004}\n\n![Inter side-chain interactions and disposition of the \u03b14 CT helical structure.\\\nRibbon representations of the helical structure of \u03b14 CT. Sidechains are shown as sticks. The helical structure contains a bent at the end of N-terminal half. A number of ionic and/or hydrogen bonding interactions are probable for the C-terminal region of the \u03b14 helix. Figure was generated by INSIGHT II.](pone.0055184.g005){#pone-0055184-g005}\n\n![Electrostatic surface potential of \u03b14 CT.\\\nA representative structure is shown for before and after 180\u00b0 rotation along the z-axis. Surfaces in red, blue and white represent, respectively, negatively charged, positively charged and neutral residues. The figure was generated by PyMOL.](pone.0055184.g006){#pone-0055184-g006}\n\n10.1371/journal.pone.0055184.t001\n\n###### Summary of structural statistics of the twenty lowest energy structures of \u03b14 CT in aqueous solution.\n\n![](pone.0055184.t001){#pone-0055184-t001-1}\n\n  ------------------------------------ ---------------------------------------------- -------------------------------------------\n  **Distance restraints**                      Intra-residue(\\|i -- j\\|\u200a=\u200a0)                              107\n                                                 Sequential(\\|i -- j\\|\u200a=\u200a1)                               95\n                                                 Medium range (2\u2264\\|i-j\\|\u22644)                               98\n                                                           Total                                          300\n  **Dihedral angle constraints (\u03a6)**                                                                      31\n  **Constraints violations**                     Average NOE violation (\u00c5)                               0.24\n                                                 Maximun NOE violation (\u00c5)                               \u22640.24\n  **Deviation from mean structure**            Backbone atoms (N, C\u03b1, C\u2032) (\u00c5)          1.15 (N-terminal: 0.85, C-terminal: 0.45)\n                                                      Heavy atoms (\u00c5)                  1.86 (N-terminal: 1.52, C-terminal: 0.95)\n  **Ramachandran plot analysis**          \\% residues in the most favorable region                        93\n                                          \\% residues additionally allowed region                          7\n                                        \\% residues in the generously allowed region                       0\n                                            \\% residues in the disallowed region                           0\n  ------------------------------------ ---------------------------------------------- -------------------------------------------\n\nMapping Residues of PaxLD2-LD4 that Interact with the Integrin \u03b14 CT by ^15^N-^1^H HSQC {#s2c}\n---------------------------------------------------------------------------------------\n\nFor interactions studies, we have expressed and purified full-length \u03b14 CT and an N-terminus fragment of paxillin (residues G139-K277 or PaxLD2-LD4) that encompassed LD2, LD3 and LD4. The N-terminus region of paxillin that contains the LD repeats has been shown to bind the \u03b14 CT [@pone.0055184-Liu3]. The ^15^N-^1^H HSQC spectra of PaxLD2-LD4 and \u03b14 CT were assigned using standard triple resonance experiments. An overlay of ^15^N-^1^H HSQC spectra of the ^15^N-labeled PaxLD2-LD4 in the absence of (black contour) and in the presence of (red contour) two-fold excess of unlabeled \u03b14 CT is shown ([Figure 7](#pone-0055184-g007){ref-type=\"fig\"}). There are significant perturbations in the HSQC spectra of PaxLD2-LD4 upon addition of \u03b14 CT, which suggest interactions. Notably, a large number of HSQC cross-peaks of PaxLD2-LD4 demonstrate loss in intensity as a consequence of complex formation ([Figure 7](#pone-0055184-g007){ref-type=\"fig\"} panels A-D). In addition, new HSQC peaks are observed close to the ^15^N-^1^H HSQC peaks of residues A16, V45 and L131 (panel B), residue L82 (panel C) and residue S112 (panel D). Because of the extensive resonance overlapping in the ^15^N-^1^H HSQC spectra, arising from a preponderance of similar residues in the amino acid sequence of paxillin fragment, binding induced changes are assessed only for well separated ^15^N-^1^H HSQC cross-peaks. The ^15^N-^1^H HSQC cross-peaks of Gly and Ser/Thr residues are well separated from others as a result of intrinsic upfield shift in ^15^N chemical shift ([Figure 7A](#pone-0055184-g007){ref-type=\"fig\"}). The intensity of the ^15^N-^1^H HSQC peaks are significantly diminished for residues G37, G54, G57, G69, S122, S134 and T125. These Gly residues are situated at the linker region between the LD repeats. Similar perturbations can be seen for residues in the LD repeats of PaxLD2-LD4 ([Figure 7](#pone-0055184-g007){ref-type=\"fig\"}). Residues of PaxLD2-LD4 showing binding induced resonance perturbations are listed in [Table 2](#pone-0055184-t002){ref-type=\"table\"}. Interestingly, resonance perturbation can be seen for residues located in all the three LD repeats and those in the linker regions. In addition, more residues are perturbed in LD3 and LD4 repeats in comparison to the LD2 repeat. The ^15^N-^1^H HSQC cross-peaks of residues E6, L7, L11, L12 of LD2 repeats are not significantly affected in the presence of the \u03b14 CT ([Figures 7B and 7D](#pone-0055184-g007){ref-type=\"fig\"}). However, changes observed for the ^15^N-^1^H HSQC cross-peaks of the linker residues of PaxLD2-LD4 are rather intriguing ([Table 2](#pone-0055184-t002){ref-type=\"table\"}). Collectively, we surmise that binding to the 32-residue \u03b14 CT induces global conformational changes of the entire sequence of PaxLD2-LD4. These are likely to yield ^15^N-^1^H HSQC spectral changes away from the binding interface [@pone.0055184-Zuiderweg1].\n\n![Determination of interactions between ^15^N-labeled PaxLD2-LD4 and \u03b14 CT by ^15^N-^1^H HSQC NMR.\\\n^15^N-^1^H HSQC and selected sections (panels A-D) of ^15^N-^1^H HSQC spectrum of ^15^N-labeled PaxLD2-LD4 in the absence of (black contour) and in the presence of (red contour) unlabelled \u03b14 CT at a ratio of 1\u22362 (PaxLD2-LD4:\u03b14 CT).](pone.0055184.g007){#pone-0055184-g007}\n\n10.1371/journal.pone.0055184.t002\n\n###### Residues of PaxLD2-LD4 showing changes in intensity (\\>70%) and/or chemical shifts in ^15^N-^1^H HSQC spectra after addition of \u03b14 CT.\n\n![](pone.0055184.t002){#pone-0055184-t002-2}\n\n  Structural Location                                         Residue\n  ------------------------ -----------------------------------------------------------------------------\n  LD2                                                  D8, R9, N15, A16, V17\n  Linker between LD2-LD3       D27, E28, A29, S31, S32, G37, A38, S40, T48, G53, G54, G57, N67, G69\n  LD3                                         E73, V75, R76, S78, E80, S81, L82, E85\n  Linker between LD3-LD4    S92, T98, N100, S105, Q108, R109, T111, S112, S120, A121, S122, S123, A124,\n  LD4                                         T125, R126,L131, A133, S134, L135, D137\n\nG-S-N-**L-S^5^-E-L-D-R-L^10^-L-L-E-L**-**N^15^-A-V**-Q-H-N^20^-P-P-G-F-P^25^-A-D-E-A-N^30^-S-S-P-P-L^35^-P-G-A-L-S^40^-P-L-Y-G-V^45^-P-E-T-N-S^50^-P-L-G-G-K^55^-A-G-P-L-T^60^-K-E-K-P-K^65^-R-N-G-G-R^70^-G-**L-E-D-V^75^-R-P-S-V-E^80^-S-L-L-D-E^85^-L-E**-S-S-V^90^-P-S-P-V-P^95^-A-I-T-V-N^100^-Q-G-E-M-S^105^-S-P-Q-R-V^110^-T-S-T-Q-Q^115^-Q-T-R-I-S^120^-A-S-S-A-**T^125^-R-E-L-D-E^130^-L-M-A-S-L^135^-S-D**-F-K.\n\n**a**:Amino acid sequence of PaxLD2-LD4. The LD repeats are bold faced.\n\nMapping Residues of Integrin \u03b14 CT that Interact with PaxLD2-LD4 by ^15^N-^1^H HSQC {#s2d}\n-----------------------------------------------------------------------------------\n\n^15^N-^1^H HSQC spectra of ^15^N-labeled \u03b14 CT in the absence of (black contour) and in the presence of (red contour) unlabeled PaxLD2-LD4 are shown ([Figure 8](#pone-0055184-g008){ref-type=\"fig\"}). Addition of PaxLD2-LD4 caused chemical shift and/or intensity changes for several ^15^N-^1^H HSQC cross-peaks of \u03b14 CT, indicating binding ([Figure 8A](#pone-0055184-g008){ref-type=\"fig\"}). From the combined chemical shift changes of ^15^N and ^1^HN nuclei of \u03b14 CT, residues Q8, S11, E15, E16, S23, Y24, I25, N26, S29, N30 and D32 exhibit higher chemical shift changes ([Figure 8B](#pone-0055184-g008){ref-type=\"fig\"}). In addition, ^15^N-^1^H HSQC cross-peaks of residues K6, S11, S27 and S29 become less intense in the presence of \u03b14 CT ([Figure 8A](#pone-0055184-g008){ref-type=\"fig\"}). This may result from the broadening of resonances as intermediate chemical exchange between the free and bound states of the molecule occurs. Comparing the two sets of ^15^N-^1^H HSQC data we obtained for interactions between PaxLD2-LD4 with \u03b14 CT, there are more resonance perturbations detected in PaxLD2-LD4 than \u03b14 CT. This suggests that PaxLD2-LD4 undergoes larger conformational changes compared with \u03b14 CT when they interact. However, chemical shift changes of \u03b14 CT occurred upon binding with PaxLd2-LD4 were reproducible in repeated measurements. As can be seen, PaxLD2-LD4 induced resonance perturbations mainly from residues in the C-terminal helix of \u03b14 CT ([Figure 8](#pone-0055184-g008){ref-type=\"fig\"}). Limited resonance perturbations of \u03b14 CT membrane proximal residues K1-K10 were detected ([Figure 8B](#pone-0055184-g008){ref-type=\"fig\"}). These results demonstrate that the C-terminal region of \u03b14 CT is primarily responsible for its interactions with PaxLD2-LD3. This is consistent with the finding of a previous study that identified involvement of the C-terminal region of \u03b14 CT for binding to paxillin [@pone.0055184-Liu2].\n\n![Determination of interactions between ^15^N-labeled \u03b14 CT and PaxLD2-LD4 by ^15^N-^1^H HSQC NMR.\\\n(A) ^15^N-^1^H HSQC spectrum of ^15^N-labeled \u03b14 CT in the absence of (black contour) and in the presence of (red contour) unlabelled PaxLD2-LD4 at a ratio of 1\u22362 (\u03b14 CT:PaxLD2-LD4). (B) A bar diagram summarizing combined chemical shift changes of ^15^N and ^1^HN resonances of the \u03b14 CT as a function of amino acids.](pone.0055184.g008){#pone-0055184-g008}\n\nBinding of Integrin \u03b14 CT with LD-containing Peptides of Paxillin {#s2e}\n-----------------------------------------------------------------\n\nAtomic-resolution structures have been determined for LD repeats of paxillin in complex with well-folded FAT domain of FAK and with CH domain of the adaptor protein parvins [@pone.0055184-Hayashi1], [@pone.0055184-Liu5], [@pone.0055184-Hoellerer1], [@pone.0055184-Wang1], [@pone.0055184-Stiegler1]. We therefore examine the binding interactions of three synthetic peptide fragments containing LD2 (NLSELDRLLLELNAVQHN), LD3 (VRPSVESLLDELESSVPSPV) and LD4 (ATRELDELMASLSDFKFMAQ), with the \u03b14 CT. ^15^N-^1^H HSQC spectra overlays of \u03b14 CT in the absence of (black contour) and in the presence of (red contour) LD3 ([Figure 9A](#pone-0055184-g009){ref-type=\"fig\"}) and LD4 ([Figure 9B](#pone-0055184-g009){ref-type=\"fig\"}) are shown. Addition of LD3-containing peptide caused chemical shift changes only for residues Q8, S11, N17, S27, S29, N30, D31 of the \u03b14 CT ([Figures 9A and 9C](#pone-0055184-g009){ref-type=\"fig\"}). Addition of LD4-containing peptide reduced the signal intensity of ^15^N-^1^H cross-peaks of \u03b14 CT residues Q8, S21, S27, S23, Y24, and I25, presumably occurring from conformational exchanges ([Figure 9B](#pone-0055184-g009){ref-type=\"fig\"}). By contrast, ^15^N-^1^H HSQC spectra of the \u03b14 CT were largely unaffected in the presence of LD2-containing peptide (Supplementary [Figure S2](#pone.0055184.s002){ref-type=\"supplementary-material\"}). Indeed, PaxLD2-LD4 experiments have provided insights into residues corresponding to LD3 and LD4 that are affected by \u03b14 CT interactions ([Table 2](#pone-0055184-t002){ref-type=\"table\"}). Collectively, these results suggest that \u03b14 CT binds directly to paxillin LD3 and LD4 repeats.\n\n![Determination of interactions between ^15^N-labeled \u03b14 CT and LD repeats of paxillin by ^15^N-^1^H HSQC NMR.\\\n^15^N-^1^H HSQC spectrum of ^15^N labeled \u03b14 CT in the absence of (black contour) and in the presence of (red contour) unlabelled peptide-containing LD3 (A) or unlabelled peptide-containing LD4 (B) at a ratio of 1\u22364 (\u03b14 CT:peptides). Bar diagrams summarizing combined chemical shift changes of ^15^N and ^1^HN resonances of the \u03b14 CT as a function of amino acids in the presence of LD3 (C) or LD4 (D).](pone.0055184.g009){#pone-0055184-g009}\n\nA Molecular Model of \u03b14 CT in Complex with LD3 and LD4 Repeats of Paxillin {#s2f}\n--------------------------------------------------------------------------\n\nLD repeats adopt helical structures whereby the non-polar face of the helix, containing most of the Leu residues, are inserted into the binding pocket of target proteins whereas the polar face of the helix that contains acidic residues Asp/Glu remains exposed to the solvent [@pone.0055184-Hayashi1], [@pone.0055184-Liu5], [@pone.0055184-Hoellerer1], [@pone.0055184-Wang1], [@pone.0055184-Stiegler1]. Because the C-terminus of the \u03b14 CT is highly polar, we generated a docking model of \u03b14 CT with LD3 and LD4 repeats of paxillin by maximizing potential polar interactions in the complex ([Figure 10](#pone-0055184-g010){ref-type=\"fig\"}). In the docked structure, helices of LD3 and LD4 repeats are arranged in a sequential orientation with the \u03b14 CT helix, whereas LD3 helix is parallel and LD4 helix orients in an anti-parallel fashion ([Figure 10](#pone-0055184-g010){ref-type=\"fig\"}). There are a number of potential ionic, hydrogen bond and non-polar packing interactions that may sustain the \u03b14 CT and paxillin LD repeats complex ([Figure 10, left panel](#pone-0055184-g010){ref-type=\"fig\"}). The interface between LD3 repeat and \u03b14 CT can potentially be stabilized by salt bridges formed by side-chains of residues R2 and E6 of LD3 and residues E15/E16 and R19 of \u03b14 CT, respectively. In addition, sidechains of residues D10 and D13 of the LD3 are in a close proximity with the sidechains of N26 and N30 of \u03b14 CT, suggesting interactions via hydrogen bonds. The non-polar sidechain of residue L9 of LD3 is partially exposed and can make van der Waals' packing with the aromatic sidechain of residue W22 of \u03b14 CT. The LD4 helix docks onto the opposite face of the \u03b14 helix. The helix-helix packing can be maintained by potential ionic and/or hydrogen bonding interactions among the side-chains of residues E4, E7, S11 of LD4 with the side-chains of residues K28, D20 and S21 of \u03b14 CT. There are also packing interactions of residues L8 and F15 of LD4 with Y24 and L13 of \u03b14 CT, respectively. The model that we proposed herein for \u03b14 CT in complex with LD3 and LD4 of paxillin can be supported by experimental findings in which mutating \u03b14 CT residue E16 (or E983) or residue Y24 (or Y991) to Ala disrupted the binding of \u03b14 CT to paxillin [@pone.0055184-Liu2]. Further, residue S21 (or S988) of \u03b14 CT, located at the interface of the complex, was known to modulate paxillin binding due to phosphorylation [@pone.0055184-Han1].\n\n![Docked structure of the \u03b14 CT with LD3 and LD4 repeats of paxillin.\\\nThe C-terminal region of the \u03b14 CT (in green ribbon) may form interfaces with helices of LD3 (in stick) and LD4 (in stick). The interactions could be maintained by ionic and/or hydrogen bonding and van der Waals' packing among non-polar residues (right panel). The probable packing of side-chains among the non-polar and aromatic residues, L8, F15 (in red) of LD4 with Y24, L13 (in green) of \u03b14 CT and L9 (in purple) of LD3 with W24 of \u03b14 CT(in green) represented by space filling. The potential ionic and/or hydrogen bond interactions are marked by broken lines. Figures were generated by INSIGHT II.](pone.0055184.g010){#pone-0055184-g010}\n\nConclusions {#s2g}\n-----------\n\nOur study provides for the first time the atomic structure of integrin \u03b14 CT. It also provides molecular insights into interactions between \u03b14 CT and LD repeats of paxillin. Unlike CTs of \u03b1IIb, \u03b1M and \u03b1X integrins that all have an N-terminal helix followed by a C-terminal loop, the \u03b14 CT adopts a helical C-terminal region that is involved in paxillin binding. Conceivably, sequence and structural variations of \u03b1 CTs of integrins can contribute toward recruiting specific signaling proteins. Finally, our proposed model of \u03b14 CT in complex with LD3 and LD4 of paxillin will be useful for the design and testing of small molecules that can disrupt this interaction and therefore potentially anti-inflammatory.\n\nMaterials and Methods {#s3}\n=====================\n\nSynthetic Peptides {#s3a}\n------------------\n\nThe sequence of \u03b14 cytoplasmic tail from residues Lys968 to Asp999 (^968^KAGFFKRQYKSILQEENRRDSWSYINSKSNDD^999^) is re-numbered from 1--32 for ease of reference. Additionally, the paxillin region Gly139 to Lys277 or PaxLD2-LD4 is also re-numbered from 1--139. All synthetic peptides of \u03b14 CT and those containing paxillin LD repeats (LD2: NLSELDRLLLELNAVQHN; LD3: VRPSVESLLDELESSVPSPV; LD4: ATRELDELMASLSDFKFMAQ) were purchased from GL Biochem (Shanghai, China). They were further purified using a reverse phase HPLC, Waters\u2122 connected to a C18 column (300 \u00c5 pore size, 5 \u00b5M particle size). A linear gradient of acetonitrile/water with a flow rate of 2 ml/min was used to elute the peptides, and the major peak fractions were collected and lyophilized into powder form. Mass spectrometry was used to verify molecular weights of the peptides.\n\nExpression and Purification of \u03b14 CT and PaxLD2-LD4 {#s3b}\n---------------------------------------------------\n\nThe full length \u03b14 CT (Lys968 to Asp999) was cloned into a pET-31b(+) vector (Novagen EMD, San Diego) with N-terminal ketosteroid (KSI) [@pone.0055184-Chua1], [@pone.0055184-Chua2], [@pone.0055184-Bhunia1] which has a Met cleavage site inserted prior to the \u03b14 CT sequence. The recombinant plasmid was transformed into BL21(DE3) cells. Transformed cells were cultured overnight in Luria-Bertani (LB) broth. The culture was seeded in 1\u2236100 volume ratio either in LB for the preparation of unlabeled proteins or in isotope-enriched M9 minimal media, containing ^15^N ammonium chloride without/with ^13^C-glucose for the production of isotope labeled samples at 37\u00b0C in a shaking incubator. IPTG (1 mM) was used to induce protein expression for 18 hours at 25\u00b0C with a shaking speed of 150 rpm. *E. coli* cells were harvested by centrifugation at 5000 rpm for 20 min, and the bacterial pellet re-suspended in a buffer containing 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0. Re-suspended cells were lysed via sonication on ice to release the recombinant fusion proteins. As the KSI recombinant protein is targeted to the inclusion bodies, cell pellets were collected via centrifugation at 14000 rpm for 30 min and re-solubilized in a buffer containing 8 M urea, 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0. The supernatant containing the solubilized KSI-\u03b14 CT was affinity purified using Nickel-NTA acid (QIAGEN) beads making use of the 6-His tag that was attached to the N-terminus of KSI-\u03b14 CT. The fusion protein was then eluted in buffer containing 8 M urea, 0.5 M imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0. The eluted fractions were pooled and dialyzed against water at 4\u00b0C for 2 days to remove the urea, causing the formation of KSI-\u03b14 CT precipitates that were subsequently collected by centrifuging at 5000 rpm for 30 min. The KSI-\u03b14 CT precipitates were dissolved in 70% formic acid. For every 1 mg of KSI-\u03b14 CT, 37.5 mg of cyanogen bromide was used for the cleavage reaction. The reaction was purged by N~2~ gas and left in the dark for 22 hours. Sodium hydroxide was used to neutralize the cyanogen bromide and the solvent was removed using a rotary evaporator leaving behind a thin film of precipitate. The precipitate was dissolved in 10 mM sodium phosphate buffer, pH 6.5 and further purified using HPLC. The identity of the cleaved peptide was verified by mass spectrometry analysis.\n\nThe PaxLD2-LD4 (residues G139 to K277) was cloned into the pET24a(+) vector with an initiation Met introduced before G139. The construct also contained a C-terminal 6-His tag for affinity purification. The plasmid DNA was transformed into BL21(DE3) cells. Protein was produced, unlabeled or isotope (^15^N, ^15^N/^13^C) labeled, by IPTG induction at 18\u00b0C for 18 hours. *E. coli* cells were harvested by centrifugation at 5000 rpm, 4\u00b0C, for 20 min. The cell pellet was re-suspended in buffer containing 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0 and lysed via sonication on ice to release recombinant proteins. The cell lysate was centrifuged at 14000 rpm, 4\u00b0C, for 30 min. The supernatant was collected and affinity purification of PaxLD2-LD4 performed using NTA beads. Washing steps were performed in buffer containing 20 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0. Bound PaxLD2-LD4 protein was eluted in buffer containing 0.3 M imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 8.0. Eluted proteins were dialyzed against buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.0 at room temperature for 1 hour. The protein was further purified using HPLC with a linear gradient of water/acetonitrile solvents.\n\nNMR Experiments {#s3c}\n---------------\n\nAll NMR experiments were recorded on a Bruker DRX 600-MHz instrument equipped with an actively shielded cryo-probe. 10% Deuterium oxide and 2 mM 2,2-dimethyl-2-silapentane-5-sulfonate (DSS) was added to all NMR samples. Chemical shifts were referenced to DSS. 2D TOCSY (mixing time: 50 ms) and 2D NOESY (mixing time: 200 ms) spectra were recorded for 0.5 mM of \u03b14 CT dissolved in water, pH 5.6 at 278 K. Raw NMR data were processed using TOPSPIN 2.1 and analyzed with SPARKY. ^15^N-^1^H HSQC spectra of \u03b14 CT and PaxLD2-LD4 were assigned by triple resonance HNCACB and CBCA(CO)NH experiments. Triple resonance NMR experiments were carried out using doubly labeled (^15^N/^13^C) samples of \u03b14 CT and PaxLD2-LD4 dissolved in 10 mM sodium phosphate buffer, pH 5.6, at 298K. For interactions studies, ^15^N-^1^H HSQC spectra of either ^15^N-labeled \u03b14 CT (100 \u00b5M) or PaxLD2-LD4 (200 \u00b5M) were obtained in the presence of unlabeled binding partners at molar ratio of 1\u22361 and 1\u22362 in 10 mM sodium phosphate buffer, pH 6.5, 298 K.\n\nStructure Calculation and Modeling {#s3d}\n----------------------------------\n\nNOE intensities of \u03b14 CT NOESY spectra were qualitatively categorized into strong, medium and weak and translated to the upper bound distance limit of 2.5 \u00c5, 3.5 \u00c5 and 5.0 \u00c5 respectively. These distance constraints were used for structure calculations using CYANA (Combined assignment and dynamics for NMR applications) 2.1 [@pone.0055184-Guntert1]. For structure calculation backbone dihedral angle (\u03a6) values were restricted to \u221230\u00b0 to 120\u00b0 to limit the conformational search. Of the 100 structures, 20 lowest energy structures were selected for evaluation and analyses. PROCHECK-NMR [@pone.0055184-Laskowski1] was employed to evaluate the stereochemical quality of the structural ensemble and figures were prepared using PyMOL, MOLMOL, Discovery Studio Visualizer 2.0 and Insight II. Docking of \u03b14 CT with LD peptides of paxillin was performed using Insight II software. Helical structures of LD3 and LD4 peptide fragments were constructed for docking with \u03b14 CT. Several round of docking exercises were conducted to achieve optimal sidechain-sidechain packing interactions. The model complex was further energy minimized using discover force field to relieve short inter-atomic contacts.\n\nSupporting Information {#s4}\n======================\n\n###### \n\n**Comparison of primary structures of representative \u03b1 and \u03b2 CTs of integrins.** Alignment of amino acid sequences of CTs of \u03b14, \u03b1X, \u03b1M, \u03b1L, \u03b1D, \u03b1IIb integrins.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Determination of interactions between paxillin LD2 peptide and \u03b14 CT by ^15^N-^1^H HSQC NMR.** ^15^N-^1^H HSQC spectra of \u03b14 CT in the absence (black contour) and in the presence (red contour) of LD2 peptide.\n\n(TIF)\n\n###### \n\nClick here for additional data file.\n\nWe thank Dr. M.L. Tang for generating the \u03b14 integrin and paxillin expression constructs. We also acknowledge Ms. Harini Mohanram for her help with the figure preparation.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: GLC SMT SB. Performed the experiments: GLC ATP. Analyzed the data: GLC SB SMT. Contributed reagents/materials/analysis tools: GLC ATP SMT SB. Wrote the paper: GLC SMT SB.\n"}
+{"text": "![](brmedchirj271532-0095){#sp1 .382}\n"}
+{"text": "\n"}
+{"text": "Background\n==========\n\nModern dance developed from expressionist dance, which rejected the rules of ballet. It originated in Germany in the 1920s and was defined by rules of a specific technique evolved in the USA in the years 1930--1950, employing elements of African and Native American folklore. Modern dance employs all the techniques of ballet; however, a modern dancer is not obliged to uncritically follow the classical rules of execution of these techniques.\n\nClassical style strongly emphasizes esthetics, precision, and height. In modern dance, the distance and motion through space is of utmost importance.\n\nCohan writes that the most important elements of a modern dancer's work are: \"ground work\", \"work with the center of gravity\", and \"motion in space\" \\[[@b1-medscimonit-20-1082]\\]. He mentions that the most important thing in a well performed movement is self-awareness, and that modern dance consists of \"centering, gravitation, balance, posture, gestures, rhythm, motion in space, and breathing\".\n\nAlthough ballet is a form of art, it has much in common with professional sports. Modern style choreographers often propose very dangerous dancing movements from the point of view of biomechanics of the locomotor system. These dangerous situations created by \"movement designers\" are often the result of absence of basic knowledge about the capabilities and limitations of the human locomotor apparatus.\n\nModern dancers in many shows perform falls, using their upper limbs as shock absorbers, they roll around the stage, women lift men, and men not infrequently lift persons heavier than themselves, they wear heavy costumes, they walk on stilts, hang from ropes, etc. In addition, the performances often take place outdoors or in halls that are not prepared for such events.\n\nThe majority of dancers' expressive movements consist of jumps. Research shows that some of them, particularly during the landing phase (during the eccentric phase of muscle work related to shock absorption), generate high values of the vertical component of ground reaction force (GRF), which may reach 7.4 BW \\[[@b1-medscimonit-20-1082]--[@b3-medscimonit-20-1082]\\]. Serious injuries often happen during these phases of jumps ([Table 1](#t1-medscimonit-20-1082){ref-type=\"table\"}). Peak forces of impact phases occur after a few dozen milliseconds \\[[@b2-medscimonit-20-1082]\\].\n\nAccording to Luke, Solomon, Liederbach, and Nicholas \\[[@b4-medscimonit-20-1082]--[@b8-medscimonit-20-1082]\\], the largest percentage of injuries affecting professional dancers are chronic injuries, such as soft-tissue inflammation, injuries resulting from overload, and muscle strains and tears. Fractures represent the lowest percentage of all injuries. Various authors report different locations total number of injuries \\[[@b2-medscimonit-20-1082],[@b9-medscimonit-20-1082]--[@b13-medscimonit-20-1082]\\]. Nevertheless, ankle joint, foot, spine, hip joint, and knee joint are the most frequently listed regions of the body injured by dancers. Research by Gorwa shows that modern dancers are primarily exposed to injuries of the spine, Achilles tendon, knee joint, and hip joint \\[[@b2-medscimonit-20-1082]\\].\n\nStudies performed by Gorwa also show that up to 93% of modern dancers (female and male) suffer from chronic injuries directly related to their profession \\[[@b2-medscimonit-20-1082]\\]. Around 30% of persons who reported these conditions associated them with specific movements. Female modern dancers complained of spine conditions such as neck pain and lumbar or sacral spine region pain; they associated them with \"sudden twists\" and the very expressive movements so common in modern dancing style. Male modern dancers listed pain and injuries of the spine, feet, and knees as chronic conditions and indicated these problems were caused by lifting dancing partners, as well as by \"bizarre\" choreography.\n\nSince correct technique is a factor that significantly decreases the risk of injury \\[[@b14-medscimonit-20-1082],[@b15-medscimonit-20-1082]\\], and the degree of motor habit control (e.g., the proper execution of a given sports technique) determines the values of forces recorded during the landing phase \\[[@b16-medscimonit-20-1082]\\], we observed kinematic values and recorded the technique of execution of a modern dance movement referred to as the \"stag jump\".\n\nThe purpose of this study was to conduct a thorough analysis of kinematic structure of the landing phase in a selected expressive dance movement called the \"stag jump\" performed as a case study of 1 male and 1 female professional modern dancer. Researchers who compared the technique and morphological differences between modern and ballet dancers were Bronner and Ojofetimi \\[[@b17-medscimonit-20-1082]\\], Solomon and Micheli \\[[@b12-medscimonit-20-1082]\\], and Solomon et al. \\[[@b18-medscimonit-20-1082]\\]. The methodology and measurements presented in this work are part of a project whose purpose will be to devise guidelines for new training methods to prevent injuries in dancers.\n\nThe conducted kinematic analysis consisted in establishing such characteristic parameters as trajectories of joint centers and movement sequences of lower limbs, pelvis movement, angle changes in hip, knee, and ankle joints in the sagittal plane, as well as changes of foot position in relation to the ground. Synchronization of kinematic measurements with the measurement of vertical components of ground reaction forces in the landing phase during this stage of the study allowed us to identify impact loads on the locomotor apparatus of the studied dancers and to forecast methods for minimizing such loads.\n\nMaterial and Methods\n====================\n\nMaterial\n--------\n\nThis work is a case study of 2 professional modern dancers (a female and a male) who work full time in dance theatre, 6 days per week on average, and do not perform any other job. They train for 42 hours a week. Daily training sessions last 8 hours on average. During the dance season, however, the number of daily working hours increases to 11--13. There are about 8--9 monthly performances, but during the \"artistic season\" there may be up to 20 monthly performances \\[[@b2-medscimonit-20-1082]\\]. It should be stressed that although artists in ballet dance groups in Poland have identical education (National Ballet Schools: Primary and Secondary \\[[@b2-medscimonit-20-1082]\\]) around 7% of members of modern dance groups begin their dance education after age 18 \\[[@b2-medscimonit-20-1082]\\].\n\nEach participant performed the expressive movement called the \"stag jump\".\n\nLaboratory experiments\n----------------------\n\nThe tests used for this study were carried out in the Biomechanical Laboratory of the Chair of Biomechanics at the University School of Physical Education in Poznan, Poland. Kinematic values and line graphs of GRF were determined during the measurements. Kinematic values were computed with the use of APAS motion analysis system and GRFs were measured with the use of Kistler force plate sampling at 1000 Hz. The motion of test participants was recorded by 4 Basler digital cameras with recording frequency of 200 Hz. The images recorded by the cameras were transferred to a computer, where, with the use of APAS software, the films were processed and the locations of markers positioned on dancers' bodies were determined. At the same time, GRFs acting during the landing phase were recorded. Due to the extensiveness of performed dancing movements, the cameras were arranged to allow us to precisely determine the kinematics of the pelvis and the right lower limb during the landing phase. The number and location of markers ([Figure 1](#f1-medscimonit-20-1082){ref-type=\"fig\"}) made it possible to determine the positions of right lower limb joint centers and, in consequence, the relative angular motion of particular lower limb segments and the pelvis.\n\nResults\n=======\n\nThe measurements allowed the authors to determine values of lower limb joint angles and pelvis position within the coordinate system ([Figure 2](#f2-medscimonit-20-1082){ref-type=\"fig\"}). The following angle *vs.* time graphs were drawn:\n\n-   for hip joint -- flexion/extension, abduction/adduction, and rotation;\n\n-   for knee joint -- flexion/extension and rotation;\n\n-   for talocrural joint -- dorsal/plantar flexion and pronation/supination.\n\nThe usage of force plates allowed us to establish ground reaction forces (GRF) acting during the landing phase that followed the performed movement.\n\nBased on the results of the measurements, it was possible to assess movements performed by the dancers by analyzing the line graphs of GRF and line graphs of angles in lower limb joints. This paper concentrates on the analysis of the \"stag jump\" movement performed by 2 modern dancers -- a woman and a man. The following diagrams ([Figures 3](#f3-medscimonit-20-1082){ref-type=\"fig\"}--[10](#f10-medscimonit-20-1082){ref-type=\"fig\"}) present the results of measurements performed on modern dancers.\n\nThe pelvis rotation in relation to the long axis of the body was 30\u00b0 for the female dancer and 25\u00b0 for the male dancer. Pelvis tilt in the sagittal plane is significant during the entire landing phase and fluctuates between 32\u00b0 and 50\u00b0 for the female dancer and 21--31\u00b0 for the male dancer. Both pelvis rotation in relation to the vertical axis and pelvis tilt in the sagittal plane result from specific positioning of the non-supporting limb ([Figure 6](#f6-medscimonit-20-1082){ref-type=\"fig\"}).\n\nThe line graph of joint angles in the sagittal plane ([Figures 7](#f7-medscimonit-20-1082){ref-type=\"fig\"} and [8](#f8-medscimonit-20-1082){ref-type=\"fig\"}) shows that they are similar in shape but differ in the range of motion (ROM). This is especially true in the hip joint angle ([Figure 7](#f7-medscimonit-20-1082){ref-type=\"fig\"}). The ROM in the hip joint is 50\u00b0 for the female and 36\u00b0 for the male. In the knee joint ([Figure 8](#f8-medscimonit-20-1082){ref-type=\"fig\"}), ROM is 32\u00b0 for the female dancer and 31.7\u00b0 for the male dancer. ROM of ankle joint ([Figure 9](#f9-medscimonit-20-1082){ref-type=\"fig\"}) in the sagittal plane was 60\u00b0 for the female dancer and 59\u00b0 for the male dancer.\n\nDiscussion\n==========\n\nThis study allowed us to describe the way a dancer positions (coordinates) the body during landing that follows a dance movement and the forces that act on the lower limb. The values of the vertical component of ground reaction reached the levels that equaled almost 4 times the body weight of man and a little over 3.2 times the weight of woman. Considering that the values of reaction in the lower limb joints exceed the values of ground reaction \\[[@b19-medscimonit-20-1082],[@b20-medscimonit-20-1082]\\], one can observe the loads that a dancer's musculoskeletal system is exposed to. Such high loads are a frequent cause of injuries to dancers, as well as musculoskeletal problems that dancers already suffer from during their careers.\n\nGround reactions measured during the landings of classical dancers in other studies were even higher than in modern dancers (2.65 BW for females and 3.7 BW for males) \\[[@b2-medscimonit-20-1082],[@b21-medscimonit-20-1082]\\]. It is directly related to the requirements imposed on the technique of dance movements, which are much more rigorous in classical dance.\n\nInterestingly, the landing phase durations in this movement are very short. The impulsive character of the peak vertical component of ground reaction acting on feet during the landing phase that follows a ballet jump poses the greatest risk of joint injury, beginning with the lowest joints -- metatarsal and ankle joints \\[[@b22-medscimonit-20-1082]\\].\n\nThe short time allowed for the performance of a movement and the need to immediately transition to the following one leave dancers extremely little time to coordinate the entire body and prepare for receiving the huge loads observed. The requirement to perform each dance movement strictly according to the rules considerably limits the ability to properly absorb the impact during landing. This is why it is so important to define the best way (from the biomechanical point of view and taking into account the artistic design) to perform individual dance movements by dancers. It is obvious that the maximum value of the vertical component of GRF depends on the way dancers will use their feet as a form of biological shock absorbers.\n\nThe following observations can be formulated on the basis of the analysis of joint angles and GRF peaks.\n\nAt the moment the foot contacts the ground, it remains in plantar flexion in a manner that only the toes are in contact with the ground. When the vertical component of ground reaction reaches its maximum, the foot is resting flat on the ground (parallel to the ground). Therefore, the center of mass is transferred from toes and metatarsus to the vicinity of the ankle joint, which decreases the moment of vertical ground reaction forces (and thus the forces generated by muscles working around the ankle joint). In addition, the flat positioning of the foot lowers the center of the ankle joint, shortening the moment arms of the components acting in the horizontal plane.\n\nWhen the vertical component of GRF reaches its peak, thighs and shins assume almost vertical positions. This leads to the conclusion that when the greatest loads act on the limb, these will be mostly compressive loads, for which the long bones of the lower limb are well prepared (they possess the best strength in the longitudinal direction).\n\nThe analysis of foot angle in relation to the ground shows that after landing on his toes, the male dancer bends his foot dorsally and for a short period of time (about 0.2 s) puts weight on the entire foot and then stands on the toes. The female dancer lands in a different manner -- after touching the ground with the toes, she puts weight on the entire foot and keeps the foot in this position throughout most of the support phase.\n\nConclusions\n===========\n\nModern dance generates many potentially dangerous situations for the locomotor system. One of the most significant factors that lead to injuries is high dynamic loads (forces) of an impact nature that occur particularly during the landing phases of numerous expressive jumps.\n\nPeak forces are generated over a very short time period, which makes coordinating such movement structures very difficult. It is important to realize that these forces are transmitted via the relatively small surface of the dancer's feet that are only weakly protected by characteristic footwear -- the forefoot region and toes.\n\nProperly mastering movement technique is tremendously important in minimizing injuries. Many publications on injury prevention emphasize that proper technique is fundamental in preventing injuries \\[[@b16-medscimonit-20-1082],[@b23-medscimonit-20-1082]\\]. The knowledge of dynamic loads acting on the dancer's body and kinematics of movement may represent a significant contribution in the description of the proper technique of dance movements.\n\nThe following values that best describe the landing phase of the \"stag jump\" were established:\n\n-   loads transmitted to the dancer's foot;\n\n-   line graphs of joint angles in the lower limb;\n\n-   changes in position of the foot in relation to the ground;\n\n-   values of joint angles, as well as the angle between the foot and the ground when the highest value of the vertical component of ground reaction occurs.\n\nThis allowed us to analyze the influence of movement technique on the values of external loads (GRF).\n\nAnalysis of the conducted research reveals significant differences in the technique of the same dance movement. Research by Orishimo et al. \\[[@b24-medscimonit-20-1082]\\], as well as Broner and Ojofeitimi \\[[@b17-medscimonit-20-1082]\\], suggests that these differences in technique probably do are not due to the dancer's sex, professional experience, or educational path of the dancer. The female dancer went through all levels of ballet education in Poland (Primary and Secondary National Ballet School), and the male dancer began his education as an adult at the DanceWeb workshop in Vienna. However, such a conclusion needs to be confirmed by further research with a larger sample.\n\nHamilton et al. observed correctly that certain dance movements can be mastered only until puberty, when the body can still be \"molded\" \\[[@b25-medscimonit-20-1082]\\].\n\nDuring the next phases of research, the results of these measurements will be used to determine loads transmitted by musculoskeletal system when performing dancing movements. For this purpose, a mathematical model of lower limb movement will be prepared, allowing us to determine the forces generated by muscles and the forces transmitted by joint surfaces. The values thus obtained will complement the results of the research.\n\nOn the basis of these results, works are being conducted in cooperation with ballet teams, choreographers, and dance teachers to devise training methods using the methodology and the potential of a biomechanical laboratory. These training methods would be aimed preventing injury and physical overload, as well as increasing dancers' knowledge of biomechanics and kinesiology of the locomotor apparatus in dancing.\n\nThe authors would like to thank their colleagues: Jacek M\u0105czy\u0144ski, MS and Krzysztof Kmiecik, MEng for their help during preparation of manuscript and a number of illustrations.\n\n**Source of support:** This study was financed in stages from the following sources: the research grant from the Polish Ministry of Science and Higher Education No NN404515938, a grant for the statutory activity of the Chair of Biomechanics, USPE in Pozna\u0144\n\n![Location of markers: R MT -- the head of metatarsal bone of second toe of right foot, R HEEL -- right foot calcaneus, R LMAL -- center of right foot lateral malleolus, R TIB -- right limb tibia, R LCON -- left femur lateral epicondyle, R THI -- right thigh, R GTRO -- right femur greater trochanter, L ASIS -- left anterior superior iliac spine, R ASIS -- right anterior superior iliac spine, SACR -- L~5~S~1~.](medscimonit-20-1082-g001){#f1-medscimonit-20-1082}\n\n![Trajectories of joint centers and successive positions of lower limb segments during the landing phase of the \"stag jump\" movement performed by the female dancer (**A**) and male dancer (**B**).](medscimonit-20-1082-g002){#f2-medscimonit-20-1082}\n\n![Line graph of the vertical component for: (**A**) male dancer and (**B**) female dancer.](medscimonit-20-1082-g003){#f3-medscimonit-20-1082}\n\n![(**A, B**) Line graph of pelvis tilt angle in the sagittal plane and pelvis rotation in relation to the vertical axis of a dancer (male dancer results).](medscimonit-20-1082-g004){#f4-medscimonit-20-1082}\n\n![(**A, B**) Line graph of pelvis tilt angle in the sagittal plane and pelvis rotation in relation to the vertical axis of a dancer (female dancer results).](medscimonit-20-1082-g005){#f5-medscimonit-20-1082}\n\n![Distinctive lower limb arrangement during the landing phase of the \"stag jump\" modern dance movement.](medscimonit-20-1082-g006){#f6-medscimonit-20-1082}\n\n![Line graph of hip joint angle in the flexion-extension plane: (**A**) male dancer, (**B**) female dancer. Mark indicates the moment of peak vertical reaction.](medscimonit-20-1082-g007){#f7-medscimonit-20-1082}\n\n![Line graph of knee joint angle in the bending-extending plane: (**A**) male dancer, (**B**) female dancer. Mark indicates the moment of peak vertical reaction.](medscimonit-20-1082-g008){#f8-medscimonit-20-1082}\n\n![Line graph of ankle joint angle in the plantar bending -- dorsal bending plane: (**A**) male dancer, (**B**) female dancer. Mark indicates the moment of peak vertical reaction.](medscimonit-20-1082-g009){#f9-medscimonit-20-1082}\n\n![Line graph of foot angle in relation to the ground: (**A**) male dancer, (**B**) female dancer. Mark indicates the moment of peak vertical reaction.](medscimonit-20-1082-g010){#f10-medscimonit-20-1082}\n\n###### \n\nTopography and frequency of injuries in professional modern dancers \\[[@b2-medscimonit-20-1082]\\].\n\n  Body parts                       Female dancers (n=7) \\[%\\]   Male dancers (n=7) \\[%\\]\n  -------------------------------- ---------------------------- --------------------------\n  Achilles tendon                  42.9                         42.9\n  Ankle joint                      14.3                         28.6\n  Metatarsus                       0.0                          42.9\n  Phalanges                        14.3                         28.6\n  Lumbar and sacral spine region   85.7                         85.7\n  Knee joint                       42.9                         100.0\n  Shank                            14.3                         28.6\n  Thigh                            14.3                         14.3\n  Hip joint                        28.6                         57.1\n\n[^1]: Study Design\n\n[^2]: Data Collection\n\n[^3]: Statistical Analysis\n\n[^4]: Data Interpretation\n\n[^5]: Manuscript Preparation\n\n[^6]: Literature Search\n\n[^7]: Funds Collection\n"}
+{"text": "Background {#Sec1}\n==========\n\nCirrhosis is the end stage of progressive fibrosis that is caused by various reasons and that responds poorly to medical conservative treatment. Chronic damage to the liver leads to the extensive accumulation of extracellular matrix (ECM) among the hepatocytes. Epidemiological data state that 1.03 million cirrhotic patients worldwide die each year from severe associated complications \\[[@CR1]\\].\n\nCurrently, liver transplantation is the most effective therapy for advanced hepatic diseases. Among those fortunate enough to receive liver transplantation, the survival rates at 3, 12, and 36 months are 94%, 88%, and 79%, respectively \\[[@CR2]\\]. However, we should be take into account the lack of donor organs, the high costs, and the long-term use of immunosuppressants after transplantation. Thus, there is an urgent need to find alternative therapeutic strategies. Recent studies have shown that hepatocytes in the cirrhotic liver still have the potential to regenerate, but there is an imbalance between regeneration and necrosis \\[[@CR3]\\]. A potential hypothesis states that a fully functioning part of the liver could be created through the proliferation of the infused cells that will remodel the injured liver. It is doubtful whether increasing the number of hepatocytes alone would be an effective treatment for the patients.\n\nBased on the proof-of-concept, hepatocytes were transplanted to treat liver-related diseases \\[[@CR4]\\]. Because of the limited number of hepatocytes and the lack of their proliferation and stability in vitro, the efficacy of grafted hepatocytes decreased progressively. Hence, it is crucial to find another readily available cell source.\n\nThis review aims to highlight all currently available evidence regarding the use of stem cells for treatment of liver cirrhosis and to determine whether there is any factual basis for their potential.\n\nStem cells in regenerative medicine {#Sec2}\n===================================\n\nStem cells, termed as clonogenic undifferentiated cells, cannot just self-renew indefinitely but can differentiate into a variety of cell lineages, including pluripotent embryonic stem cells (ESCs), induced pluripotent stem cells (iPSCs), hematopoietic stem cells (HSCs), hepatic stem cells, mesenchymal stem cells (MSCs), and so forth (Fig. [1](#Fig1){ref-type=\"fig\"}).Fig. 1The different types of stem cells isolated from different tissues differentiate into hepatocytes. ALD alcoholic liver disease, ESCs embryonic stem cells, HSCs hematopoietic stem cells, iPSCs induced pluripotent stem cells, MSCs mesenchymal stem cells, NAFLD nonalcoholic fatty liver disease, PH partial hepatectomy\n\nSplenic teratomas could be formed after infusion of ESCs \\[[@CR5]\\]. The application of ESCs is therefore limited because of their potential for malignancy. iPSCs are artificially derived from a nonpluripotent cell and thus ethical issues remain the major obstacle to their clinical administration. Furthermore, the only available source of HSCs is the hematopoietic system, and this restricts their clinical application. Hepatic stem cells have been identified in fetal as well as mature liver. During embryonic development, the cells within the liver bud are recognized as hepatoblasts which are bipotent, giving rise to both hepatocytes and bile-duct epithelial cells. Moreover, cells in the ductal plates in fetal and neonatal livers are also hepatic stem cells. Their capacity to repopulate the liver upon transplantation is also well studied in animal models \\[[@CR6], [@CR7]\\]. Hepatic progenitor cells (HPCs), also defined as hepatic stem cells, are rare in normal adult livers (0.01%), located in the Canals of Hering, and all regenerative responses are mainly granted by mature hepatocytes except in certain disease states \\[[@CR7]\\].\n\nHPCs are activated after liver injury, such as alcoholic liver disease (ALD) and nonalcoholic fatty liver disease (NAFLD) \\[[@CR8]\\]. Oxidative stress, which plays the main role in the pathogenesis of ALD and NAFLD, promotes the accumulation and differentiation of HPCs into hepatocytes \\[[@CR9], [@CR10]\\]; furthermore, HPCs can differentiate into hepatocytes in vivo and promote liver regeneration after partial hepatectomy or acute toxic liver injury \\[[@CR11]\\]. This suggests that infusion of the progenitor cells may alleviate the damage of hepatocytes which is caused by long-lasting oxidative stress or partial hepatectomy.\n\nThe proliferation of HPCs as a response to chronic liver damage is minimal \\[[@CR11]\\] and is correlated with the severity and localization of the inflammatory infiltrate \\[[@CR12]\\]. Manipulation of the HPC microenvironment may be used as a therapeutic approach for the alleviation of liver insufficiency \\[[@CR11], [@CR12]\\].\n\nIn addition, evidence has suggested that mesenchymal cells through the processes of mesenchymal-epithelial or epithelial-mesenchymal transition (MET/EMT) may contribute to adult liver regeneration during chronic liver injury \\[[@CR13]\\]. Mesenchymal cells in the liver may be derived not only from their own progenitor cells but also from the bone marrow (BM) by migrating to the injured liver \\[[@CR14], [@CR15]\\], although this statement is controversial. This suggests that not only HPCs but also mesenchymal cells simultaneously contribute to the initiation and development of liver diseases, although the mechanisms remain unclear \\[[@CR16]\\]. This indicates that the interaction between HPCs and mesenchymal cells is important for remodeling of injured liver. The accumulating evidence suggests that HPCs could be the best alternative treatment for hepatic damage; however, HPCs may cause carcinogenesis and fibrogenesis, as has been shown in vitro \\[[@CR6]\\]. Before thorough viewing of their therapeutic potential, a better knowledge of the factors that determine HPC differentiation and their possible malignant transformation is necessary.\n\nThe therapeutic potential of MSCs has been extensively investigated as well as their differentiation, immunoregulatory properties, and secretion of trophic factors. In contrast to ESCs, iPSCs, and HPCs, MSCs do not have any ethical problems and have become the ideal alternative.\n\nDuring the past few years, MSCs have been mainly isolated from the bone marrow (BM-MSCs). Alternative sources of MSCs have been proposed, such as from adipose tissue (AD-MSCs), umbilical cord blood (CB-MSCs), umbilical cord (UC-MSCs), and amniotic fluid.\n\nThe application of MSCs {#Sec3}\n=======================\n\nBM-MSCs are capable of undergoing differentiation into hepatic cells and recovering liver function, indicated by the apoptosis of hepatic stellate cells, decreased transforming growth factor (TGF)-\u03b21, and alpha-smooth muscle actin (\u03b1-SMA) gene expression \\[[@CR17]\\]. AD-MSCs, which are more immunocompatible and easier to isolate than BM-MSCs, have a protective role against liver fibrosis \\[[@CR18]\\]. UC-MSCs show a more beneficial immunogenic profile and stronger overall immunosuppressive potential than BM-MSCs \\[[@CR19]\\].\n\nAlthough MSC differentiation into hepatocytes has been demonstrated in vivo, evidence suggests that various trophic and immunomodulatory factors play a key therapeutic role in the treatment of liver fibrosis. The trophic factors, which are secreted by MSCs, prevent apoptosis of hepatocytes with the help of antiapoptotic factors (hepatocyte growth factor (HGF) and insulin-like growth factor (IGF-1)), angiogenetic factors (vascular endothelial growth factor (VEGF)), mitogenetic factors (epidermal growth factor (EGF), HGF, and nerve growth factor (NGF)), and TGF-\u03b1 \\[[@CR20], [@CR21]\\]. Because of the smaller and less complex immunogenic potency, MSC-free therapy might constitute a better alternative treatment.\n\nFurther clinical trials have evaluated the efficiency of transplanted MSCs for treating patients with liver fibrosis. Several clinical trials have been designed to evaluate their therapeutic potential in hepatic cirrhosis treatment \\[[@CR22]--[@CR26]\\] (Table [1](#Tab1){ref-type=\"table\"}). The results of the studies seem to be promising, with improvements in model for end-stage liver disease (MELD) score and metabolic parameters, but data on histological improvement are weak. Long-term outcomes after UC-MSC treatment would be preferable for patients with liver cirrhosis \\[[@CR22], [@CR23]\\], although the short-term efficacy of infused BM-MSCs is favorable \\[[@CR24]--[@CR26]\\]. It should be noted that the number of infused cells, the delivery route, and the frequency of injection per patient vary in the studies. Different sources of MSCs and various populations of patients may be more convincing for any therapeutic effect. Moreover, AD-MSCs and UC-MSCs have better immunocompatibility, and they are more vitalized and much easier to isolate than BM-MSCs from older patients \\[[@CR18], [@CR19]\\]. The efficacy of autologous BM-MSCs may suffer from aging differentiation and deficiency in vitality \\[[@CR18], [@CR22]\\]. In contrast, allogeneic UC-MSCs are free from these limitations \\[[@CR19], [@CR22]\\]. Furthermore, for prognosis and better analysis on the difference between stem cells, the follow-up time of patients should be prolonged with the creation of time points. The results are also limited because of small sample sizes and absence of control groups \\[[@CR22]--[@CR26]\\]. Currently, there are no standardized protocols for clinical trials and it is not possible to monitor whether the infused MSCs home to the targeted organs or not.Table 1MSCs in clinical trials treating liver fibrosisCell sourceDelivery routeNo. of cellsPatient populationNo. of patientsFollow-up periodEfficacyLimitationsReferenceUC-MSCsIntravenous5 \u00d7 10^5^/kg, three timesChronic hepatitis B30 treatment, 15 control12 monthsImprovement of liver function and MELD score; reduced ascitesNo track of the infused UC-MSCs and the histological evidence in the studied patients\\[[@CR22]\\]UC-MSCsIntravenous5 \u00d7 10^5^/kg, three timesPrimary biliary cirrhosis7 treatment48 weeksDecrease in serum ALP and \u03b3-GGT; alleviation of fatigue and pruritusNo track of the infused UC-MSCs and histological evidence alterations in the studied patients; less detailed follow-up time points\\[[@CR23]\\]BM-MSCsIntravenous infusion1 \u00d7 10^7^/kgLiver cirrhosis due to hepatitis C virus15 treatment,10 control6 monthsImprovement in the frequency of encephalopathy, jaundice, ascites, bleeding tendency, and lower limb edemaLess detailed follow-up time points\\[[@CR24]\\]Autologous BM-MSCsHepatic artery0.75 \u00b1 0.50\u00d7 10^6^/patientHepatitis B virus cirrhosis27 treatment,29 control24 weeksSignificant improvement in liver functionDuring follow-up, patients were lost about 1/3\\[[@CR25]\\]Autologous BM-MSCsPeripheral vein1 \u00d7 10^6^/kgEnd-stage liver disease due to hepatitis C virus20 treatment,20 control6 monthsSignificant improvement in liver functionNo histological evidence; less detailed follow-up time points\\[[@CR26]\\]*ALP* alkaline phosphatase, *BM-MSC* bone marrow-derived mesenchymal stem cell, *\u03b3-GGT* glutamyl transpeptidase, *MELD* model for end-stage liver disease, *MSC* mesenchymal stem cell, *UC-MSC* umbilical cord-derived mesenchymal stem cell\n\nGholamrezanezhad et al. \\[[@CR27]\\] have shown that there was no significant improvement in liver function after a 1-month period of follow-up because the homing ability of BM-MSCs into the liver occurred in just a limited number of infused cells. Peng et al. \\[[@CR28]\\] also mentioned that the homing ability of MSCs is the main cause why autologous MSC transplantation did not achieve acceptable long-term effects on the prognosis of a patient. The lingering problem of cell-based therapies is whether the delivered cells home within the injured sites or not and how to increase their homing ability.\n\nHoming {#Sec4}\n======\n\nMigration or homing within the injured tissues is influenced by multiple factors including the delivery route, the number of infused cells, culture conditions, and others. We review various factors that are related to the migration of MSCs.\n\nAdministration routes of MSCs {#Sec5}\n-----------------------------\n\nThe delivery route for MSCs seems to be crucial for therapeutic efficiency. Traditional administration of MSCs is mainly via intrahepatic injection, intrasplenic injection, and by intravenous infusion. Systemic delivery of cells may cause a large number of rapid losses of cells within the capillaries, especially in the lungs, which creates a short lifespan for remaining MSCs \\[[@CR29]\\]. Furthermore, infusion of cells with heparin significantly decreases the number of entrapped AD-MSCs within the lungs and increases the number of cells which are accumulated in the liver \\[[@CR30]\\]. The vascular patency may be an essential factor for MSCs flowing into the targeted tissue. Intrahepatic injection appeared to be the ideal way to administer stem cells, with less entrapment of cells in the circulation, and more MSCs differentiating into hepatocytes \\[[@CR31]\\]. Furthermore, administration of the MSCs via the portal vein or hepatic artery shows homing efficacy less than 5% and 20--30%, respectively \\[[@CR32], [@CR33]\\]. The hepatic artery thus seems to be the best delivery route and shows better homing efficacy; however, the vascular patency should be checked before infusion.\n\nOptimizing cultivation conditions {#Sec6}\n---------------------------------\n\nDuring expansion, freshly isolated MSCs lose ligands or receptors which respond to migratory signals \\[[@CR34]\\]. Migration is a passage-dependent process; with a higher number of passage there is a decrease in efficacy of homing. Also, high culture confluence impairs the migration of MSCs due to upregulation of tissue inhibitor of metalloproteinase (TIMP)-3 \\[[@CR35]\\]. Moreover, hypoxia induces the expression of leptin which is associated with activation of both the STAT3/hypoxia-inducible factor-1\u03b1 (HIF-1\u03b1)/VEGF and stromal cell-derived factor (SDF)-1/CXCR4 signaling pathways \\[[@CR36]\\]. It is suggested that hypoxic preconditioning augments the recruitment of MSCs.\n\nStimulating the target site to recruit MSC mobilization {#Sec7}\n-------------------------------------------------------\n\nIn the acute phase of injury, inflammatory cytokines which were released from the damaged tissues recruit monocytes for tissues repair. Compared with unirradiated mice, more MSCs homed in mice that received total body irradiation \\[[@CR37]\\], suggesting that infused MSCs are moved first to injured sites. However, in patients with a subchronic or chronic phase of the disease, some indispensable chemokines for homing may be minimal or absent; therefore external stimuli may provide a simple and available novel approach for homing.\n\nPerry et al. \\[[@CR38]\\] used degenerate electrical waveforms for patients with skin scars and showed that electrical stimulation significantly reduced scar scores and may guide cell migration. Furthermore, the physiological electrical field induced MSCs to graft to the anode in vitro, which had no influence on cell senescence and phenotype \\[[@CR39]\\]. Meanwhile, pulsed focused ultrasound noninvasive local pressure waves deposit energy within the targeted tissues that change the level of local chemoattractants and enhance the efficacy of homing \\[[@CR40]\\]. Mechanical stretching could also enhance engrafted MSC homing within injured tissues via hypoxia, vascularization, and proliferation \\[[@CR41]\\]. In summary, external stimuli may be used to control or induce direct migration of MSCs.\n\nGenetically modified MSCs {#Sec8}\n-------------------------\n\nBecause of the presence of specific integration between ligand and receptor, one hypothesis is that changing the level of the receptor/ligand on MSCs may improve the efficiency of homing within the targeted tissues.\n\nIn the acute phase of injury, the damaged tissue releases numerous stromal cell-derived factors (SDF-1\u03b1), but their receptor (CXCR4) is at a low level on the cultured MSCs. MSCs with overexpressed CXCR4 have better migration potential toward SDF-1\u03b1 and secrete more trophic factors, including HGF and VEGF which stimulate hepatocyte regeneration \\[[@CR42]\\]. Ryu et al. \\[[@CR43]\\] further explained that Akt, ERK, and p38 signal pathways are also related to the SDF-1/CXCR4 axis.\n\nHGF is the most effective mitogen in hepatocyte regeneration and, during tissue injury, its biological effects rely on tyrosine kinase receptor and on cellular mesenchymal to epithelial transition factor (c-met) \\[[@CR44]\\]. Genetic loss of c-met compromises the potential of hepatic oval cells, including their proliferation, migration, and differentiation \\[[@CR45]\\]. Liu et al. \\[[@CR46]\\] demonstrated that the HGF/c-met signaling pathway is crucial for MSC homing within the injured liver and that it facilitates the liver repair. Overexpressed receptor or ligands on the MSCs corresponds with the specific cytokines which are released from the injured organs and could induce homing directly within the targeted tissues.\n\nMicroRNAs or noncoding RNAs target mRNA for degradation or inhibition and may determine the migration of MSCs. More than 60 different microRNAs in MSCs have been recently described and some of them are involved in migration, including let7, microRNA-10b, microRNA-27b, microRNA-335, and microRNA-886-3b \\[[@CR47]\\]. Overexpressed microRNA-211 through the STAT3/microRNA-211/STAT5A signal pathway enhanced migration \\[[@CR48]\\]. Upregulation of microRNA-221 and microRNA-26b enhanced MSC migration via the chemotactic response towards HGF through activation of PI3K/Akt signaling \\[[@CR49]\\]. In addition, some other microRNAs suppressed migration of MSCs---microRNA-27b suppressed the directional migration of MSCs by targeting SDF-1\u03b1, and overexpression of microRNA-124 significantly inhibited the chemotactic migration towards HGF by downregulation of Wnt/\u03b2-catenin signaling \\[[@CR50]\\]. It is suggested that microRNAs are involved in MSC potential, including their differentiation, paracrine function, proliferation, survival, and migration. Upregulation or downregulation of microRNAs in MSCs could regulate the migration.\n\nConclusion {#Sec9}\n==========\n\nThe present review demonstrates that stem cell therapy has a favorable therapeutic effect. Currently, the crucial factor that determines the benefit of MSCs is the homing efficacy. The disadvantages of MSC therapy in clinical trials include the risks of iatrogenic tumorigenesis, cellular embolism, and the optimum time for the infusion of cells. Moreover, its safety in clinical trials should be approved by institutional ethics committees. In conclusion, the results on MSCs which were used for the treatment of liver fibrosis are promising, but we need to know the underlying mechanism of their therapeutic effects.\n\nAD-MSC\n\n:   Adipose-derived mesenchymal stem cell\n\nALD\n\n:   Alcoholic liver disease\n\nBM-MSC\n\n:   Bone marrow-derived mesenchymal stem cell\n\nCB-MSC\n\n:   Cord blood-derived mesenchymal stem cell\n\nc-met\n\n:   Cellular mesenchymal to epithelial transition factor\n\nCXCR4\n\n:   Chemokine receptor type 4\n\nECM\n\n:   Extracellular matrix\n\nEGF\n\n:   Epidermal growth factor\n\nESC\n\n:   Embryonic stem cell\n\nHGF\n\n:   Hepatocyte growth factor\n\nHIF-1\u03b1\n\n:   Hypoxia-inducible factor-1\u03b1\n\nHPC\n\n:   Hepatic progenitor cell\n\nHSC\n\n:   Hematopoietic stem cell\n\nIGF-1\n\n:   Insulin-like growth factor-1\n\niPSC\n\n:   Induced pluripotent stem cell\n\nMELD\n\n:   Model for end-stage liver disease\n\nMSC\n\n:   Mesenchymal stem cell\n\nNAFLD\n\n:   Nonalcoholic fatty liver disease\n\nNGF\n\n:   Nerve growth factor\n\nSDF\n\n:   Stromal cell-derived factor\n\nTGF\n\n:   Transforming growth factor\n\nTIMP\n\n:   Tissue inhibitor of metalloproteinase\n\nUC-MSC\n\n:   Umbilical cord-derived mesenchymal stem cell\n\nVEGF\n\n:   Vascular endothelial growth factor\n\nWe thank Dr. Jiaying Liu for supporting us with human UCB-MSCs.\n\nFunding {#FPar1}\n=======\n\nThis work was supported by the National Natural Science Foundation of China (No. 81770591), the Gilead Sciences Research Scholars Program in Liver Disease---Asia, the Key Medical Talents Fund of Jiangsu Province (ZDRCA2016007)\u00a0and\u00a0the Medical Innovation Team Project of Jiangsu Province (CXTDA2017023).\n\nAvailability of data and materials {#FPar2}\n==================================\n\nNot applicable.\n\nYZ, YL, LZ, JL, and CZ designed the manuscript and analyzed the literature. YZ, YL, and CZ wrote the manuscript and prepared the table. All authors reviewed and approved the final manuscript.\n\nEthics approval and consent to participate {#FPar3}\n==========================================\n\nNot applicable.\n\nConsent for publication {#FPar4}\n=======================\n\nAll authors consent to the publication of this manuscript.\n\nCompeting interests {#FPar5}\n===================\n\nThe authors declare that they have no competing interests.\n\nPublisher's Note {#FPar6}\n================\n\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\n"}
+{"text": "Occupational consciousness emerged as a construct from my doctoral work on intergenerational play within families in post-apartheid South Africa. It refers to ongoing awareness about the dynamics of hegemony and recognition that dominant practices are sustained through what people do every day, with implications for personal and collective health (Ramugondo, [@cit0046]). The emergence of the construct in this context signifies an ongoing struggle with negotiating long-standing dynamics of power that were laid down during colonialism, and maintained under black majority rule. Occupational consciousness provides a language through which people can describe how their individual and collective everyday doing can resist and challenge hegemonic practices that sustain all forms of unequal power relations. In providing a theoretical foundation to occupational consciousness as a construct in occupational science, this discussion advances the theorizing practice of the discipline in promoting understandings of human occupation.\n\nThe first three sections of the paper introduce theorizing as a distinct scholarly practice, outline the genesis of occupational consciousness as a construct and provide both theoretical and philosophical foundations to the construct. This is followed by an analysis of synergies between occupational consciousness and other related constructs. Occupational consciousness is then advanced as a critical notion that frames everyday doing as a potentially liberating response to oppressive social structures. The paper concludes by proposing potential avenues for further theorizing and research.\n\nTheorizing as Scientific Practice and Interrogating the Epistemological Gaze {#s0001}\n============================================================================\n\nOccupational science, a discipline concerned with human occupation and its situatedness in context (Whiteford, Townsend, & Hocking, [@cit0053]) and role and function in society (Clark et al., [@cit0010]; Yerxa, [@cit0055]), cannot underplay its central role in theorizing about occupation (Ramugondo & Kronenberg, [@cit0047]). The term occupation has different meanings in professional and general public discourses, with various configurations of occupation illuminating the complex nature of the construct and emphasizing different aspects of what it refers to. Additionally, a number of constructs or concepts emerging from occupational science have different meanings in the everyday English language. Outlining, unpacking and critiquing these constructs has great potential for a generative scholarship or theorizing, which is required to build and sustain occupational science as a discipline.\n\nTheorizing has only recently come to be regarded as a recognizable scientific practice and a critical aspect of growing and deepening scholarship in the social sciences (Swedberg, [@cit0051]). Theorizing refers to the process of developing a system of interconnected ideas that condense and organize knowledge about the social world, \"explaining how some aspect of the social world works and why\" (Neuman, [@cit0045], p. 57). The process of developing a system of interconnected ideas about how human occupation \"is shaped, embedded and negotiated within, as well as how it contributes to the shaping of, social systems and structures\" (Laliberte Rudman, [@cit0037], p. 55) is an important exercise for occupational scientists. How else could we advance occupational science as a recognizable body of knowledge, or develop, deepen and extend understandings of human occupation in order to communicate effectively with one another, and develop or critique empirical evidence?\n\nCentral to theory are constructs or concepts. While these abstractions are often used interchangeably some authors draw a clear distinction, suggesting that concepts are phenomena with specific definitions that are broadly agreed upon while constructs are more complex phenomena, with multiple dimensions and thus possibly contested. Others, however, do not make this distinction. For instance, Neuman ([@cit0045]) suggested that concepts differ in terms of level of abstraction and whether they operate singularly or in clusters, are simple or complex, and narrow or broad in scope. Thus Neuman's definition of abstract concepts corresponds with the definition of a construct, referring to aspects of the world that are not directly observable but can help people organize their thoughts and expand understandings. In this paper, I approach both concepts and constructs as synonymous, using the terms interchangeably.\n\nConcepts in social science emerge in a variety of ways. Some originate from classical theory, while others come out of deep contemplation and reflection, or after examining and synthesizing research findings (Neuman, [@cit0045]; Swedberg, [@cit0051]). Constructs in occupational science have tended to originate from the latter. The power available to the theorist in shaping knowledge by proposing new constructs, however, needs careful examination and interrogation. Theorizing in occupational science, as in many social science disciplines or humanities, involves reconstructing people's stories and languaging their lived experience. Many authors have pointed to the skewed ontological positions of theorists in occupational therapy and occupational science in terms of gender, socio-cultural and socio-economic factors, whereby Western, Caucasian, female, middle-class, heterosexual and ableist constructions of meaningful occupation are favoured (Hammell, [@cit0029]; Hocking, [@cit0031]; Iwama, [@cit0032]; Kantartzis & Molineux, [@cit0033]; Kronenberg, Algado, & Pollard, [@cit0035]). For this reason, Hammell (2011) warned of the potential danger of universalism and theoretical imperialism, arguing that theories from a diversity of cultural perspectives are likely to promote inclusivity and a richer understanding of occupation.\n\nGenesis of the Construct and Definition {#s0002}\n=======================================\n\nIn writing about a construct emerging from research on people who may be viewed as vulnerable, by virtue of their relative distance from the intellectual exercise of theorizing that does not involve them directly yet pertains to their everyday lives, it is imperative that I disclose some aspects of my socio-cultural positioning in relation to them. Occupational consciousness is a construct with potential and particular relevance in occupational science that emerged from my doctoral work. The study explored intergenerational shifts and continuities in children's play within family, and investigated associated factors (Ramugondo, [@cit0046]). The family I researched is black South African, representing a statistically dominant group in the country in terms of race and political affiliation to the ruling elite, but largely marginalized in terms of economic participation and growth (Van der Westhuizen, [@cit0052]). I share both racial and ethnic identities with this family, to whom I will refer as the Gudani family.\n\nHaving grown up during apartheid South Africa, I am a first generation university graduate in my family, only coming into contact with occupational therapy as a profession during my undergraduate studies. Although both parents within the Gudani family held post-matric qualifications, the family had never been exposed to occupational therapy. My presence within their community as a researcher of play sparked considerable interest, prompting a request for inclusion in the study from one family. The other two families that participated were approached in accordance with inclusion criteria for purposive sampling. Recruitment followed ethical processes as approved by the Human Research Ethics Committee of the University of Cape Town. Data were collected from all three families, although my doctoral thesis was based on analysis and synthesis of the extensive data gathered from the Gudani family, who were given a copy of the thesis and audio recordings of narratives from the grandmother that captured previously unknown details about her childhood.\n\nThe study was partly prompted by anxieties about the apparent decline of play in relation to what adults knew of their own childhood play, that were noted in the post-apartheid era. These anxieties were shared with me by various community leaders during my visits across South Africa covering six provinces (Limpopo, Eastern Cape, Kwazulu-Natal, Northern Cape, Free State and Gauteng). Observations in my personal and professional life as an occupational therapist echoed these sentiments. Similar anxieties are reflected in literature about British childhood play at the turn of the 20^th^ century (Barnes & Kehily, [@cit0003]; Bishop & Curtis, [@cit0005]). The study sought to explore whether such anxieties in post-apartheid South Africa could be settled in the same manner as they were in Britain, where researchers concluded that rather than total decline, contemporary British childhood play was in many respects embedded in historical practices (Barnes & Kehily, [@cit0003]; Boyes, [@cit0008]). Given South Africa's colonial and apartheid past, as well as large-scale, on-going, and rapid politico-ideological and macro-economic changes, and an accelerated pace of modernization (Boehnke & Bergs-Winkels, [@cit0007]; Haste, [@cit0030]), a study of change in children's play within family during early 20^th^ to 21^st^ centuries promised to yield interesting lessons about how play evolves as an occupation, and how individuals across generations within family assert agency in constructing a shared play narrative.\n\nThe Gudani family was studied as a case (Flyvbjerg, [@cit0018]; Stake, [@cit0049]), situated in post-apartheid South Africa and reflective of its historical make-up as well as ongoing transitions. The three generations (grandmother, parents and children) were approached as embedded cases, allowing for historically situated play portraits to emerge for each generation. A focus on play as cultural practice necessitated an ethnographic lens to inform how I interrogated events and narratives about play, especially with regards to third generation. Participant observation became the main data collection method during the 24 days that I visited the family, arriving in the morning and departing after the last child had gone to bed. During this time, audio footage was collected mainly via a listening device which I wore around my neck, capturing all conversations, including television audio output whenever I was in the lounge area. Consistent with a critical paradigm that purports power to be discursive, circulated through dominant ideologies and practices (Frisby, [@cit0022]; Habermas, [@cit0028]; Marcuse, [@cit0041]), I held on very lightly to notions of play as presented in academic text, opting to fore-front what emerged as the family's own constructions of what counts as play. Useful in this approach was Sutton-Smith's view of play as elusive, and often framed in literature through theoretical lenses that are already contaminated by what professions and disciplines wish to see in it and use to assert power within the academy (Sutton-Smith, [@cit0050]).\n\nWhat emerged from the study was that while in this particular family play had retained elements found in its institutionalized forms, such as games -- albeit less prominent in the third generation -- unstructured play had become complicated, reflecting the complexification of the play narrative (Ramugondo, [@cit0046]). This complication came as a result of western-led social change characterized by modernization, global awareness and consumerism, and accelerated by technology in the form of television and cell-phones. Interestingly, even as adults in the family bemoaned the seemingly \"lost\" third generation, signaling the fading away of real play as a form of cultural loss, their role in this process was missing in the narratives. Even as the television set went on when the first person got up in the morning and got switched off only when the last person went to bed, no mention was made of how this could compete with formalized or structured games, or influenced unstructured play. Television, particularly local and American soap operas such as *Days of our Lives* and *The Young and the Restless*, not only determined time use and the patterning of occupations in and outside of family life, but also inserted itself into unstructured play and everyday conversation.\n\nThe apparent lack of awareness about the role of television in what became of play within the third generation, and the accompanying anxiety about the loss of collectively shared notions of \"real\" play, led to the emergence of occupational consciousness as a construct with potential relevance in occupational science. The construct requires further conceptualization and clarity about its theoretical underpinnings. This is especially so since it shares both converging and diverging points -- in how it has been used, or with other related concepts -- in other disciplines.\n\nUnderpinning Theoretical and Philosophical Foundations {#s0003}\n======================================================\n\nPotential convergence exists between occupational consciousness and other sociological concepts associated with social inequity and systems of stratification, particularly those grounded on Marxian and Weberian thought, such as class and status consciousness; \"the subjective awareness of class or status location and the implications of such awareness for social action\" (Laumann & Senter, [@cit0036], p. 1307). It is not so much in the notion of class or status location -- the conception of which has been problematized in sociology (Marshall, [@cit0042]; Wright, [@cit0054]) -- that potential synergy lies, but the formulation of consciousness as an emotional and psychic response to social stratification; always in interaction with the makings and consequences of such stratification (Reay, [@cit0048]). This sociological perspective on consciousness has permeated much of post-colonial theory, and provides a central theoretical grounding for how I continue to conceptualize occupational consciousness.\n\nSocial scientists have long concerned themselves with how social stratification positions unequal groups in characteristic patterns of social relationships, but without much theoretical development to explain the link and the nature of such relationships (Laumann & Senter, [@cit0036]; Nash, [@cit0043]; Reay, [@cit0048]). In dealing with oppression and casting a critical theoretical lens to \"intervene in those ideological discourses of modernity that attempt to give a hegemonic 'normality' to the uneven development and the differential\" (Bhabha, [@cit0004], p. 171), post-colonial theorists have provided important insights in the quest to understand these dynamics. Fanon and Biko in particular, emerging from a liberation thought on colonialism and black domination, have been able to illustrate how often times both the oppressed and the oppressors reinforce oppressive systems. This perspective echoes a view expressed by Paulo Freire with reference to the social and political landscape in Latin America during the 1960s and 1970s where he observed that: \"both the metropolitan society and the dependent society, totalities in themselves, are part of a greater whole, the economic, historical, cultural, and political context in which their mutual relationships evolve... the dependent society is by definition a silent society. Its voice is not an authentic voice, but merely an echo of the voice of the metropolis -- in every way, the metropolis speaks, the dependent listens. (Freire, [@cit0021], pp. 503-504)\" Fanon (1952/[@cit0017]), through his treatment of the notion of internalization, contended that the success of colonialism throughout the African continent could not depend solely on the might of colonial states, but also implicit consent of the oppressed. It is in this tacit but unwitting consent that I saw parallels between how oppression can be enacted and sustained, and how the Gudani family could be displaced from the central role of orchestrating its own play narrative across generations. Like colonialism, this shift in control for the occupational trajectory of children's play for the family did not happen overnight. Occupational consciousness is thus concerned with how the things people do every day, individually and collectively, sustain systems and structures that support and promote certain occupations or certain ways of doing, to the exclusion of others. It is also about uncovering the trappings of human occupation that perpetuate these systems and structures. Occupational consciousness, however, is not about framing the oppressed as \"cultural dopes\" (Garfinkel, [@cit0025]) or as perpetually unaware and incapable of resisting dominance. To the contrary, the term unearths and gives language to acts of resistance that may already be in existence in certain communities that have and continue to face marginalisation and oppression, but have not yet been theoretically explored in occupational science.\n\nBoth Fanon and Biko centered their philosophy of liberation on the notion of consciousness. Freire ([@cit0021]) also wrote extensively about the necessity for conscientization in order to break the 'culture of silence' in the dependent society. He noted, however, that even as there may be similarities in how the path to liberation may be advanced in both Latin America and other areas of the Third World, each context will present particular nuances. Consistent with sociological formulations that view it as a response to social stratification (Reay, [@cit0048]), consciousness is also referred to as a mental attitude in continuous interaction with oppressive systems (Biko, [@cit0006]; Fanon, 1961/[@cit0016]) and a commitment to fight all forces that seek to use identity as a stamp that marks out subservience (Biko, [@cit0006]). It is a response that requires that one acknowledges that one is indeed oppressed \"because it is from the 'Other' and her Exteriority that the new truth-claims spring forth and demand explanation\" (Dussel, [@cit0015], p. 21).\n\nFreire ([@cit0021]) also stressed the need for the oppressed to acknowledge this reality, as a first critical step to self-liberation. This stance begins to deal with what social scientists had long evaded -- a theoretical analysis of the patterns of social relationships between unequal groups -- albeit from the position of those that admit to suffer the consequences of such social stratification. While Biko and Fanon's analyses came from liberation thought regarding colonialism and black domination, there are interesting parallels between what can be understood of this analysis and relational patterns between groups with unequal access to meaningful and dignified occupation, and the manner in which systems and structures that sustain such stratification are maintained.\n\nBiko and Fanon consistently argued that the instrumental mode of division is central to the colonial project, with regions within colonies differently marked for privilege across history (Fanon, 1961/1963). Biko ([@cit0006]) observed how in Apartheid South Africa, racial difference was used to stratify people with regards to privilege and resources. Lighter skinned black people could climb the social strata and earn more privileges by successfully proving to the authorities that they were of a different race. Biko ([@cit0006]) saw this as a deliberate design by the system to not only stratify people socially, but also in terms of their aspirations. Stratified this way, it becomes not only logical but legitimate for people in different social strata to view their positions as deserved. Those with sustained privileges and accumulated resources would then fight hard to retain access, while those with less may resign themselves to the limits prescribed for them. The consequences of division within a colonial project -- for those at the lower end of privilege -- is that they not only end with feeling separate from the privileged other, but also reduced. For those that benefit from oppressive systems, *the oppressed other* and his or her ways would need to change substantially, in order to enjoy the same entitlements.\n\nConsciousness, for Biko, therefore begins with a realization that regardless of our positions across social strata, we are all oppressed, and the fact that people are oppressed to varying degrees is a deliberate design of the system. This deliberateness renders individuals and groups within an unequal system able to participate in their own and each other's oppression. It was for this reason that Biko saw a strong interrelationship between a consciousness of the self and the emancipatory project. This realization led to one of the famous quotes Biko is known by: \"The most potent weapon in the hands of the oppressor is the mind of the oppressed\" (Biko, [@cit0006], pp. 101-102). Although it begins with the individual, consciousness cannot end there. It serves to disrupt the cycle through which subjectivities developed through oppressive systems and *othering* are sustained.\n\nWhile Biko wrote about oppression with a focus on race, blackness and whiteness, Fanon signaled very clearly that a central issue of liberation philosophy is not so much the color of one's skin, but how power is used by those with means within an unequal society and the reasons that underlie their continuous hold on power. Fanon's analysis of these issues, however, aligns very closely with Biko's central arguments about what sustains oppressive systems; the legitimization of ongoing access to privilege for those with relative power, and the reduction of the other to a lesser being who can graduate to a place of significance only through the denial of self and concomitant emulation of the elite.\n\nWith reference to the pitfalls of national consciousness amongst the middle-class and political elite at the end of colonial regimes across African and Latin-American States, Fanon (1961/1963) found that the newly developed \"native bourgeoisie\" often failed to govern with attention to new social relations that needed to be understood and developed in the interest of the whole nation. In their haste to replace \"the foreigner\" and to fit the cosmopolitan mold, they paid little attention to the problems faced by peoples of the new nation from the perspectives of all who make up that nation. This, Fanon termed intellectual laziness, which serves to engraft social relations between unequal groups along the lines already laid down by colonialism. Nationalization, in this narrow sense, \"has nothing to do with transforming the nation; it consists, prosaically, of being the transmission line between the nation and a capitalism, rampant though camouflaged, which today puts on the masque of neo-colonialism\" (Fanon, 1961/1963, p. 122). This, for those who have suggested that post-colonial Africa is a myth (Aidoo, [@cit0001]; Grosfoguel, [@cit0027]), is referred to as coloniality, and defined by Maldonado-Torres ([@cit0040]) thus: \"Long-standing patterns of power that emerged as a result of colonialism, but that define culture, labour, intersubjective relations, and knowledge production well beyond the strict limits of colonial administrations. Thus coloniality survives colonialism. It is maintained alive in books, in the criteria for academic performance, in cultural patterns, in common sense, in the self-image of peoples, in aspirations of self, and so many other aspects of our modern experience. In a way, as modern subjects we breathe coloniality all the time and every day. (p. 243)\" Consciousness, as understood by Biko ([@cit0006]) and Fanon (1961/1963), can thus be seen as a mental attitude and commitment to resist coloniality; patterns of power that define and sustain dominant cultural practices and patterns and in so doing reproduce unequal intersubjective relations. It requires particular attention to mechanisms that serve to divide people, denying those in the periphery their voice, or the ability to tell stories that matter to them. It is recognition that to accede to these mechanisms is to deny humanity a chance at collective self-understanding.\n\nIntersections between consciousness and occupation hold potential for theorizing about human occupation within occupational science as a discipline, perhaps even introducing the philosophy of liberation (Dussel, [@cit0013], [@cit0014], [@cit0015]; Fanon, 1961/1963) into the discipline. The philosophy of liberation has been best articulated by Dussel -- renowned as its co-founder amongst a group of Argentinian philosophers who emerged in the 1960s during military dictatorship (Burton & Osorio, [@cit0009]; Dussel, [@cit0015]; Mahvish, [@cit0039]) -- to refer to an explanation of oppression and a critique of dominant accepted truths about the status quo, from the viewpoint of alterity; the perspective of the oppressed other (Dussel, [@cit0015]). Human occupation may very well be central to oppression and coloniality, in that *doing* is the most visible enactment of unequal intersubjective relations. Consistent with a transactional view of occupation (Dickie, Cutchin, & Humphry, [@cit0012]), what people do every day sustains discourses that produce and maintain \"truths\" about self or selves and others. Similarly, Ramugondo and Kronenberg ([@cit0047]) have argued that human beings in fact are occupied, not only in terms of using their time, energy and personal resources, but also in that they are jointly culpable for what becomes of themselves and others, through their action in and on the world. In drawing links between human occupation and consciousness as understood from Fanon and Biko, and grounded within the philosophy of liberation, I highlight the need for a wakefulness or alertness about how the things that human beings do every day intersect with inequality and oppression.\n\nSynergies with Other Related Constructs {#s0004}\n=======================================\n\nThe generation of constructs within occupational science in recent years has been celebrated by many, viewed as critical to advancing theory related to occupation and its situatedness. Some of these constructs can form interconnected groups, which I refer to here as construct clusters, borrowing from Neuman's ([@cit0045]) description of concept clusters; a collection of \"associated concepts that are consistent and mutually reinforcing\" (p. 65). Occupational consciousness, as a construct, enjoys synergy with a number of constructs introduced to occupational science during the past decade, as well as occupational therapy discourse. This paper serves to highlight some of the conceptual links between occupational consciousness and these other constructs as an introduction to what I hope will be an \"ongoing occupational terminology interactive dialogue\" (Laliberte Rudman, [@cit0037], p. 55).\n\nThere may also be other constructs or concepts I have not identified, that fall within the same cluster of construct for theorizing purposes. In outlining some of the conceptual links between occupational consciousness and the four constructs I identify here, I am hoping to illuminate some potential for occupational science terminology to interact in ways that advance theorizing about human occupation. The four constructs I begin to draw from in this occupational terminology interactive dialogue are occupational possibilities, occupational choice, occupational apartheid, and collective occupation.\n\nLaliberte Rudman ([@cit0037]), having analyzed discourses pertaining to aging and occupations that become constructed as ideal or non-ideal, introduced occupational possibilities as a construct in occupational science. Drawing from a Foucauldian concept, governmentality, she was able to demonstrate how transactions between socially contextualized structure and agency shape occupation at individual and collective levels, arguing that this is the mechanism through which occupational possibilities are differently shaped for different categories of persons or social groups. Laliberte Rudman ([@cit0037]) further proposed that a generative way to advance understanding of the interplay between structure and agency in the negotiation and enactment of occupation was not only understanding how occupational possibilities come about within context, but also how they are taken up and at times resisted by individuals and collectives. Occupational consciousness links very strongly with the latter part of this proposition, highlighting that it is through everyday doing at individual and collective levels that systems and structures that support and promote certain occupations to the exclusion of others, are sustained. This point was also made by Angell ([@cit0002]), who argued that human occupation could \"be a site of both resistance to and reproduction of the social order\" (p. 104).\n\nOccupational consciousness, as an occupational science construct, underscores this resistance, suggesting that some level of alertness to how everyday doing intersects with oppression is necessary at individual and collective levels in order to disrupt the cycle through which subjectivities developed through governmentality are sustained. Where occupational possibilities frame occupation \"as an essential object and target of contemporary technologies of government\" (Laliberte Rudman, [@cit0037], p. 58), occupational consciousness frames human occupation as a possible response. This response is essential for dignified living, especially for those who are oppressed by circulating dominant discourses. As Laliberte Rudman ([@cit0037]) argued, the circulation of discourses appears \"to be based on a 'true' understanding of what is to be governed... with some authorities and agents having more power to influence how discourses are shaped and what discourses come to be most pervasive\" (p. 56). Occupational consciousness is concerned with how human occupation sustains discourses that produce and sustain these truths in relation to self as the oppressed other.\n\nAdvancing the notion that dominant discourses may be sustained by individual and collective action in the world, Galvaan ([@cit0023], [@cit0024]) introduced the construct of occupational choice. She argued that rather than just a conscious act to assert individual agency onto the world, making choices about what to do or not to do was a transactional act between person and context; a mechanism through which \"agents and structures define and reproduce each other\" (2012, p. 160). Drawing from Bourdieu's critique of rational choice theory and insight about the relationship between structure and agency in shaping social action, Galvaan ([@cit0023]) noted that the occupational choices of adolescents in a Western Cape community of South Africa \"were contingent upon, and consistent with, those historic socioeconomic and politically asserted patterns of occupational engagement that have developed and been perpetuated in Lavender Hill since apartheid\" (p. 154).\n\nThis patterning occurred even as both the adolescents and adults at times explicitly disapproved of some of these occupations that were historically predicated and a form of occupational injustice, for example dropping out of school or abusing alcohol (Galvaan, [@cit0023]). This restricted mindset in relation to making occupational choices intersects well with the notion of occupational consciousness. Both constructs point to a form of individual and collective culpability to the circulation of dominant discourses, borne by those who are oppressed but cultivated and maintained by oppressive systems that privilege certain subjectivities over others.\n\nHowever, culpability for ongoing oppression and dominance is not only reserved for those that are forced to bear it by virtue of having limited access to structural power. Drawing from Aristotellian and Orwellian perspectives, Kronenbeg and Pollard (2005) introduced the notion of occupational apartheid, arguing that unequal power relations and consequent access or lack thereof to meaningful and dignified occupation results from a reluctance on the part of privileged sectors of society to confront injustice, because \"the world's economies have been benefitting from *it* \\[emphasis added\\] materially and in terms of local political and social stability\" (p. 66). The authors' use of a contentious term \"apartheid\" is deliberate, aiming to excavate the \"systematic segregation of occupational opportunity\" (Kronenberg & Pollard, [@cit0034], p. 59), which advances access to power for some but not all people.\n\nUnlike occupational choice and occupational consciousness, which are constructs that illuminate individual and collective culpability for perpetuating dominant practices and discourses on the part of the oppressed, occupational apartheid exposes individual and collective culpability and complacency on the side of those who benefit from oppressive systems. While both occupational apartheid and occupational consciousness focus on the fact that some agents have more power than others to influence which discourses become pervasive through human occupation, both constructs also bring to light the relational mechanisms and dynamics between unequal agents. Occupational apartheid exposes the illusion held by those privileged by dominant structures that the oppressed other has no capacity to realize power, while occupational consciousness points to the response always available to the oppressed to disrupt occupational apartheid, through everyday doing.\n\nHumans' individual-collective culpability and complacency regarding occupations that perpetuate oppressive systems or disrupt dominant practices that marginalize, is articulated in Ramugondo and Kronenberg's ([@cit0047]) definition of collective occupations, a construct introduced to occupational science by Fogelberg and Frauwirth ([@cit0019]). In their definition, Ramugondo and Kronenberg suggested collective occupations to be those \"that are engaged in by individuals, groups, communities and/or societies in everyday contexts\" that \"may reflect an intention towards social cohesion or dysfunction, and/or advancement of or aversion to a common good\" (p. 10). The authors further asserted that \"these collective occupations may have consequences that benefit some populations and not others\" (p. 10). They grounded this teleological or explanatory approach to collective occupation on the notion of *ubuntu*; an African interactive ethic defined by Cornell and Van Marle ([@cit0011]) as referring to human interconnectedness, or how peoples' humanity is constantly shaped in interaction with each other, assigning responsibility to both the individual and the community for the other's existence.\n\nThrough this perspective, Ramugondo and Kronenberg ([@cit0047]) argued that peoples' shared humanity is \"constantly being shaped by what we are able or unable to do within groups, communities, and society\" (p. 12). In a similar vein, Laliberte Rudman ([@cit0037]) suggested that occupational possibilities could provide a lens to critically appraise the ways and ends to which occupation is being promoted. Ramugondo and Kronenberg proposed that in paying attention to the \"intentional stance\" of individual and collective occupation, humans as agents within society \"may be persuaded to ask of themselves: Whom do my or our occupations serve?\" (p. 6).\n\nIt is the intentionality within collective occupation (Ramugondo & Kronenberg, [@cit0047]) that intersects with occupational consciousness. In raising consciousness about collective occupations, individuals, families, groups and whole communities begin to elevate the work required to build just societies from mere rhetoric, to paying attention to the ways in which human occupation contributes to or detracts from building such societies. Occupational consciousness, however, focusses on those who exist on the margins of unjust societies and calls attention to how, in their respective spaces of occupational influence, individuals can disrupt societal dynamics that perpetuate unjust worlds.\n\nIn paying attention to only four constructs that share conceptual synergies with occupational consciousness, my intention is not to exclude other constructs or concepts that may also share links with these constructs individually, or as another separate construct cluster. The four constructs, along with occupational consciousness, are outlined here because they begin to illuminate the politics of human occupation (Frank, [@cit0020]; Kronenberg & Pollard, [@cit0034]; Laliberte Rudman, [@cit0038]).\n\nOccupational Consciousness: A Critical Notion about Everyday Doing and Living {#s0005}\n=============================================================================\n\nAttention to the politics of human occupation cannot be separate from the philosophy of liberation, and without regard to how coloniality (Aidoo, [@cit0001]; Grosfoguel, [@cit0027]; Maldonado-Torres, [@cit0040]; Ndlovu-Gatsheni, [@cit0044]) impacts those on the margins of the globalized world. As Dussel (2013) argued, it is on the basis of a philosophy of liberation as logic, politics, ethics and interrogating the status quo, that those on the periphery of the dominant world can have a dialogue with modernity and begin to rethink their problems. Rather than grand ideals for a global revolution, a simple attempt by those on the periphery towards dignified living is central to this undertaking; the possibility of authoring the production, reproduction and development of lives in the material, social and cultural sense (Burton & Osorio, [@cit0009]; Dussel, [@cit0015]). In occupational science terms, this refers to those excluded from driving global occupational possibilities (Laliberte Rudman, [@cit0037]), being able, at times, to resist dominant discourses in order to orchestrate their own occupational narratives (Goldstein, Kielhofner, & Paul-Ward, [@cit0026]) in accordance with what informs meaningful and dignified living. This resistance requires alertness to how human occupation intersects with dominance and perpetuates inequality and oppression, what I refer to as occupational consciousness. In this sense, occupational consciousness illuminates the critical perspective of the oppressed other.\n\nOccupational consciousness is thus about adopting transgressive acts to disrupt the cycle of oppression through human occupation. In introducing this construct in occupational science as part of a philosophy of liberation, I contend that everyday doing -- while the most difficult thing to change at individual or collective levels -- is probably the most powerful mechanism through which to resist dominant discourses and practices. This point was made explicit at the two most recent congresses hosted by the Occupational Therapy Africa Regional Group (OTARG) in Zambia, and Zimbabwe. Having introduced occupational consciousness as an emerging occupational science construct at the Zambian OTARG Congress in 2011, I was approached by one of the Zambian delegates. She remarked that what I presented appeared as if I had \"peeped into\" Zambian homes. I asked her to explain, and she replied that it is because she knew many families in Zambia, including her own, where the television set was on during all hours of the day, without much interrogation about why this is so, and with little reflection about what unrestricted television viewing might lead to in the future with regards to local cultural practices, as well as health and well-being.\n\nAt the OTARG Congress in Zimbabwe in 2013, I had the privilege of co-facilitating a workshop with my colleague, Roshan Galvaan, titled \"Occupational Choice and Occupational Consciousness: An African Dialogue\". During discussions, one of the participants remarked that she was not too sure whether she liked the construct occupational consciousness or not, indicating that it is somewhat burdensome. She added however, that she could see how once communities have had their consciousness raised about their everyday occupations, they would conduct their own advocacy on any issue of injustice.\n\nI found that these insights, from delegates at two separate congresses on the African continent, illuminated occupational consciousness as a critical notion and a potentially liberating response to oppressive social structures. Just like the dominant discourses within which they are socio-culturally embedded, patterned occupational engagements are difficult to break owing to their normalizing effect. Unrestricted television watching, for example, regardless of its adverse long-term consequences within families and potentially whole communities, becomes pervasive within particular contexts because *this* is what everyone does. It also fits in very well within an ongoing colonial project posing as modernity, where the oppressed other, having been sufficiently reduced, must change himself or herself substantially by emulating the dominant other in order to rise to a place of significance. Occupational consciousness, as a critical tool for self-advocacy, calls for individuals within families and communities on the margins of dominant worlds to pay attention to their occupational influence; how their everyday doing can begin to disrupt societal dynamics of dominance. The term also alerts occupational scientists to pay attention to acts of resistance in everyday occupations as lived by communities on the margins of society.\n\nConclusions and Future Possibilities {#s0006}\n====================================\n\nIn this paper, building on a previous article which introduced the construct (Ramugondo, [@cit0046]), I have argued that occupational consciousness is a response that is always available to individuals and collectives on the periphery of dominant worlds to disrupt the cycle of oppression and inequality through everyday doing; a form of resistance, and thus a potentially liberating transgressive act. I have made my argument by bringing together human occupation and consciousness, as understood by Fanon (1961/1963) and Biko ([@cit0006]) to mean a mental attitude and commitment to resist coloniality. In doing this, I have introduced the philosophy of liberation (Dussel, [@cit0013], [@cit0014], [@cit0015]) into theorizing about human occupation, illuminating the critical perspective of the oppressed. The philosophy of liberation may enable the political and economic analysis Frank ([@cit0020]) called for as part of defining occupations, and understanding \"meaningful everyday activity (as) political\" (p. 56).\n\nUnwitting consent on the part of the oppressed is central to the ways global dominant cultural practices assert and embed themselves within local contexts, evident in how intergenerational play within a South African context was marked by one family assuming a marginal role in orchestrating its own play narrative. Such mechanisms, which rely on continuing unequal intersubjective relations and thus perpetuate coloniality, could be at play across various forms of human occupation. I hope this paper will serve as an invitation to scholars to take up post-colonial theory, coloniality and liberation philosophy to further explore ways in which human occupation is implicated. This could expand inter or transdisciplinary work within occupational science for critical conceptualizations of human occupation.\n\nBy beginning to draw conceptual links between occupational consciousness and other related constructs that also focus on the politics of human occupation, this paper promotes an ongoing occupational terminology interactive dialogue, and advances theorizing as a scholarly practice within occupational science. It is hoped that further work will outline the complex nature of relationships between the cluster of concepts identified in this paper and relevant others, to offer possible fuller explanations of why people engage in occupations that promote oppression and inequality across contexts.\n\nIn addition, further reflection on occupational consciousness may be advanced by addressing research questions such as: (a) are there specific indicators for or expressions of occupational consciousness within individuals and collectives? (b) in what ways do individuals and communities articulate acts of resistance in everyday occupation as expressed in different languages and lived within local contexts? (c) is occupational consciousness a transferrable disposition across occupations and contexts, and how does it relate to different subjectivities? (d) how does occupational consciousness intersect with an occupational justice approach? (e) what mechanisms and practices are available and effective in raising occupational consciousness? (f) what intersections lie between social justice and occupational consciousness in individuals within communities, and across institutional levels within corporates and the state? These questions, especially when those on the margins are involved as co-researchers, would further illuminate everyday doing as a powerful liberating and transgressive act, to resist dominant discourses and practices that deny dignified living for all. The last question, in particular, underscores the fact that raising consciousness in the oppressed is not enough in and of itself and may in fact bear the neo-liberal danger of \"responsibilizing\" the oppressed to overcome their oppression. The question highlights instead intersections between occupational consciousness and institutional power as worthy of further exploration.\n"}
+{"text": "**What is already known**\n\nKetamine is an anesthetic agent widely used for pediatric sedation. Currently, the initial dose of ketamine is based on the patient's weight, ranging from 1 to 2 mg/kg intravenously.\n\n**What is new in the current study**\n\nThis study identified variables that can affect the induction dose of ketamine required for sedation in pediatric patients undergoing diagnostic tests and derived a formula for calculating the dose using weight, age, and body surface area.\n\nINTRODUCTION\n============\n\nExaminations using diagnostic imaging modalities such as computed tomography and magnetic resonance imaging are becoming increasingly common in pediatric sedation and anesthesia units. Therefore, it is becoming important to ensure adequate sedation of patients undergoing such examinations \\[[@b1-ceem-18-012]\\].\n\nIf the level of sedation is inadequate, a child may move during the imaging procedure, which interferes with the examination, whereas deep sedation may lead to complications, such as apnea and hemodynamic instability. Therefore, appropriate doses of anesthetic agents are essential for successful examination of children undergoing diagnostic imaging tests \\[[@b2-ceem-18-012],[@b3-ceem-18-012]\\].\n\nKetamine is an anesthetic agent that is widely used for pediatric sedation. Currently, the initial dose of ketamine is based on the patient's weight and is in the range of 1.0--2.0 mg/kg intravenously, and a further dose is considered if the duration of examination is extended or if the patient is noticed to be about to wake up or shows movement. However, the optimal dose of ketamine can vary for a number of reasons, including the patient's medical condition, the environment in the examination room, and the type of diagnostic test being performed. Furthermore, anthropometric variables other than weight, such as age, height, and body surface area (BSA), could influence the optimal dose of ketamine. Therefore, a specific guideline for determining the optimal dose of ketamine could help to improve the success rate of diagnostic examinations and decreasing the risk of complications \\[[@b4-ceem-18-012]-[@b6-ceem-18-012]\\].\n\nThe aim of this study was to identify variables that could affect the dose of ketamine required for induction of sedation in pediatric patients undergoing diagnostic tests and to devise a formula for calculating the optimal dose using those variables.\n\nMETHODS\n=======\n\nThis study was approved by the Institutional Ethical Committee (IRB number: 4-2016-0733). Informed consent was waived due to the retrospective nature of the study. This retrospective observational study included children aged 0 to 18 years who were sedated in the pediatric sedation and anesthesia unit at our tertiary medical center from January 2011 to August 2016. The medical records of patients who received ketamine for sedation during an imaging procedure were evaluated.\n\nPediatric inpatients and outpatients who were referred for diagnostic imaging (computed tomography or magnetic resonance imaging) under sedation were included in the study. An anesthesiologist and a nurse monitored each patient's hemodynamic status and depth of sedation during the examination. Monitoring devices for pulse oximetry, noninvasive measurement of blood pressure, electrocardiography, and capnography were connected to the child before the examination. Glycopyrrolate 0.004 mg/kg was injected intravenously and ketamine 1 mg/kg was administered for intravenous sedation. Loss of consciousness was judged by disappearance of the eyelash reflex and no resistance to fitting of the mask. If the patient remained conscious, an additional 1-mg/kg dose of ketamine was administered. The sedation level was maintained at 5 to 6 points on the Modified Ramsay Sedation Scale. When the imaging study was finished, the child was taken to the recovery room. The child was discharged from the recovery room if hemodynamic values were maintained within \u00b120% of the pre-sedation values, the modified Aldrete post anesthesia recovery score remained over 9 points, and if the child was fully awake.\n\nWe recorded the patient characteristics, including weight, age, height, and BSA. The type of imaging test, doses of sedatives administered until unconsciousness, adverse events, respiratory interventions for sedation, and duration of sedation were also recorded. Data for patients who experienced an adverse event, such as desaturation or hemodynamic instability, were excluded from the analysis. BSA was defined as follows: 0.016667\u00d70.5\u00d7body weight (kg)\u00d70.5\u00d7height (cm).\n\nStudy outcomes and data analyses\n--------------------------------\n\nThe primary aim of the study was to identify the factors most closely related to the dose of ketamine required for induction of sedation. The relationships of sex, age, weight, height, and BSA with the dose of ketamine required for induction of sedation were analyzed using linear regression. The regression equation was then used in multivariate analysis. The variance inflation factor was checked for multiple collinearity before performing the multivariate linear regression. Multicollinearity was determined to be present when the variance inflation factor was \\>10. If multiple collinearity exists between independent variables, they cannot be analyzed in a single model; therefore, the study variables were analyzed using three different models. Each model was inspected visually for linearity, heteroscedasticity, and normality of the residuals. All analyses were performed using SAS ver. 9.4 (SAS Institute Inc., Cary, NC, USA).\n\nRESULTS\n=======\n\nTwo hundred and ninety-six of 1,200 pediatric patients who received ketamine for sedation during diagnostic imaging had complete medical records available for analysis. After exclusion of the records for 230 patients who either did not have accurate information available for height and weight or experienced adverse events, data for 66 patients were available for inclusion in the analysis. The demographic characteristics of the patients are shown in [Table 1](#t1-ceem-18-012){ref-type=\"table\"}. The average dose of ketamine required for induction of sedation was 2.29 mg/kg. In the univariate linear regression analyses, age, height, body weight, and BSA were identified to be factors affecting the dose required ([Table 2](#t2-ceem-18-012){ref-type=\"table\"} and [Fig. 1](#f1-ceem-18-012){ref-type=\"fig\"}).\n\nThe variance inflation factor was checked for multiple collinearity before performing the multivariate linear regression. The independent variables of height, weight, and BSA showed multiple collinearity, so it was not possible to analyze them in the same model at the same time ([Table 3](#t3-ceem-18-012){ref-type=\"table\"}). Therefore, the relationship between patient age and each variable was analyzed using three different models ([Table 4](#t4-ceem-18-012){ref-type=\"table\"}). Model 3, in which age and BSA were analyzed together, had the highest adjusted R-squared value at 0.658; therefore, we chose model 3, which suggested the following equation for determining the dose of ketamine required for induction of sedation taking into account age and BSA:\n\nDose of ketamine to induce sedation (mg)=-1.62+0.7\u00d7age (mo)+36.36\u00d7BSA (m^2^)\n\nAdditionally, to verify a normal distribution and iso-dispersion of residuals, a residual analysis was performed using studentized residuals. As shown in [Fig. 2](#f2-ceem-18-012){ref-type=\"fig\"}, the dots appear to be randomly distributed with no particular tendency centering around zero. The trend is centered around zero but the variance around zero is also scattered uniformly and randomly. To evaluate the skewness of the distribution, we constructed a probability-probability plot ([Fig. 3](#f3-ceem-18-012){ref-type=\"fig\"}). The points on this plot form a nearly linear pattern, indicating that the normal distribution is a good model for this dataset.\n\nDISCUSSION\n==========\n\nIn this study, we reviewed the medical records of children who were successfully sedated with ketamine and performed a regression analysis of the relationship between the dose of ketamine and several anthropometric characteristics. The regression analysis showed that the dose of ketamine required for induction of sedation increased with patient age, height, weight, and BSA. The multiple regression analysis indicated that the appropriate dose of ketamine for induction of sedation in children can be calculated accurately using the following formula:\n\n-1.62+0.7\u00d7age (mo)+36.36\u00d7BSA (m^2^).\n\nThe optimal dose of ketamine required for sedation is also associated with other anthropometric characteristics, including age, height, and BSA. According to the current guideline, the required dose of ketamine is determined on the basis of weight; an intravenous loading dose of 1--2 mg/kg is used in children \\[[@b7-ceem-18-012],[@b8-ceem-18-012]\\]. Linear regression analysis in this study identified age, BSA, and height to be significant factors for determining the dose. The most explicable regression equation was used to devise a formula that could determine the dosage of ketamine required for pediatric sedation. If the dosage is determined considering only one factor, the dose range may be too wide. Furthermore, children may be susceptible to small dose changes, which could make it difficult to identify the appropriate dose. Physicians who are inexperienced in pediatric sedation may find it particularly difficult to select the appropriate dose. Consideration of other patient factors, including weight, can increase the accuracy and success rate of sedation.\n\nIn this study, we found that it was useful to include BSA in the formula used to estimate the dosage of ketamine required for pediatric sedation. In children, many physical factors, including weight, can affect the pharmacokinetics and pharmacodynamics of an anesthetic agent. Therefore, these factors need to be considered in combination, rather than considering weight in isolation \\[[@b9-ceem-18-012]-[@b12-ceem-18-012]\\]. An earlier study that calculated the dose of propofol required for sedation also considered weight in addition to age \\[[@b13-ceem-18-012]\\]. However, for ketamine, only weight has been considered until now in research concerning the optimal dose required for sedation.\n\nWe did not include patient age as a variable in the formula. Body composition, function, and degree of protein binding vary according to age, so the response to administration of the same amount of medication is age-dependent. Therefore, the optimal dose should also vary according to age. Calculating the dose considering age and weight together rather than weight alone will help to establish a guideline that takes the clinical response into account. However, it is not easy to reflect age in an objective manner. The existing guideline suggests that children need a greater dose per kilogram than adults, and there is no formula for calculating the dose by substituting a specific age or other factor(s). Unlike in the existing guideline, the range of selectable doses can be reduced by reflecting age as a numeric item, as in the equation presented here.\n\nKetamine is a derivative of cyclohexamine that produces sedation and analgesia by separating the cerebral cortex from the limbic system of the brain while preserving the body's reflexes, including the vomiting and laryngeal reflexes. In the existing guideline, the suggested induction dose is determined on the basis of weight alone. However, there is wide patient-to-patient variability in the response to ketamine. Therefore, it is not appropriate to consider only weight when determining the optimal induction dose for sedation. This is the first study to show that the optimal ketamine dose can be calculated using other factors, including body weight. However, even if the guideline presented in this study is applied, the possibility of an adverse event due to ketamine cannot be ruled out, so precautionary measures are still necessary during sedation using ketamine.\n\nThis study has some limitations. First, the optimal dosage of a drug should be determined with a focus on preventing side effects by minimizing overdose; the dose of ketamine used in this study (1 mg/kg) may be appropriate for some patients, but may be an overdose in some patients despite achieving the desired sedation level. Second, the proposed formula could be cumbersome to use when administering ketamine in routine practice because it requires additional calculation of BSA. Third, the patients had various underlying medical conditions. Fourth, the study relied on accurate documentation of medical records, which may have been inadequate. However, this shortcoming may be offset by the use of objective measurements of sedation depth, such as loss of the eyelash reflex and acceptance of the face mask, during the study period. Fifth, various anesthesiologists performed the measurements. Judgment of patient movement and level of sedation may vary between operators, and the extent to which this variation may have influenced the findings of the study is unknown.\n\nIn summary, the anthropometric variables of age, height, and BSA should be taken into account when estimating the dose of ketamine required for induction of anesthesia in pediatric patients. The equation proposed here may be helpful for determining the dosages required for pediatric sedation. However, further clinical investigations are required to validate its effectiveness.\n\nNo potential conflict of interest relevant to this article was reported.\n\nThis research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI16C0793).\n\n![Scatter plot about linear correlation. (A-D) Correlation between various characteristic and induction dose of ketamine for sedation. This shows a tendency toward increasing slope line with increasing (A) age, (B) height, (C) body weight, and (D) body surface area (BSA).](ceem-18-012f1){#f1-ceem-18-012}\n\n![Scatter plot. Residual plot. Dependent variable: induction dose of ketamine for pediatric sedation. The dots appear randomly distributed with no particular tendency centering around zero.](ceem-18-012f2){#f2-ceem-18-012}\n\n![P-P plot of regression standardized residual. Probability plot. Dependent variable: induction dose of ketamine for pediatric sedation. The points on this plot form a nearly linear pattern, which indicates that the normal distribution is a good model for this dataset.](ceem-18-012f3){#f3-ceem-18-012}\n\n###### \n\nPatient demographics and clinical data\n\n  Characteristics             Value\n  --------------------------- -----------------\n  Sex (male/female)           41/25\n  Median age (mo)             1.45 (0.4--3.3)\n  Height (cm)                 54.6 \u00b1 10.7\n  Body weight (kg)            4.8 \u00b1 2.6\n  Type of imaging procedure   \n  \u2003CT                         24 (36.3)\n  \u2003MRI                        42 (63.7)\n  \u2003CT+MRI                     0\n  Sedation time (min)         21.7 \u00b1 12.1\n\nValue are presented as number, median (interquartile range), mean\u00b1standard deviation, or number (%).\n\nCT, computed tomography; MRI, magnetic resonance imaging.\n\n###### \n\nUnivariate regression analysis to identify independent factors affecting induction dose of ketamine for sedation (n=66)\n\n  Variable           \u03b2 (SE)           P-value\n  ------------------ ---------------- ----------\n  Sex                -2.000 (2.257)   0.379\n  Age (mo)           1.336 (0.128)    \\< 0.001\n  Height (cm)        0.646 (0.066)    \\< 0.001\n  Body weight (kg)   2.657 (0.272)    \\< 0.001\n  BSA (m\u00b2)           71.409 (6.991)   \\< 0.001\n\nSE, standard error; BSA, body surface area.\n\n###### \n\nCalculation of variance inflation factor to check for multicollinearity\n\n  Variable                 Variance inflation factor\n  ------------------------ ---------------------------\n  Age                      4.581\n  Height                   50.954\n  Body weight              442.738\n  Body surface area (m\u00b2)   713.506\n\nMulti-collinearity exists if the variance inflation factor is \\>10.\n\n###### \n\nMultivariate linear regression analysis of age and other variables including height, body weight, and BSA\n\n                      \u03b2 (standard error)               P-value\n  ------------------- -------------------------------- ---------\n  Model 1             Adjusted R-squared value=0.651   \n  \u2003Intercept          -7.982 (5.912)                   0.182\n  \u2003Age (mo)           0.812 (0.246)                    0.002\n  \u2003Height (cm)        0.299 (0.121)                    0.016\n  Model 2             Adjusted R-squared value=0.651   \n  \u2003Intercept          2.494 (1.796)                    0.170\n  \u2003Age (mo)           0.825 (0.241)                    0.001\n  \u2003Body weight (kg)   1.213 (0.492)                    0.016\n  Model 3             Adjusted R-squared value=0.658   \n  \u2003Intercept          -1.620 (3.062)                   0.599\n  \u2003Age (mo)           0.749 (0.247)                    0.004\n  \u2003BSA (m^2^)         36.359 (13.294)                  0.008\n\nBSA, body surface area.\n"}
+{"text": "Introduction {#s1}\n============\n\nThe greatly expanded brain size and highly developed cognitive abilities are the most significant features that set humans apart from other species. In addition, human brain size is also highly variable in the general population, ranging from 981 ml to 1,795 ml (1,462 ml in males and 1,266 ml in females, on average) [@pone.0050375-Rushton1]. Recent imaging studies using MRI techniques have revealed a high heritability (0.82--0.87) of brain volume and its correlation with general intelligence [@pone.0050375-Thompson1], [@pone.0050375-Posthuma1], working memory, perceptual organization and processing speed [@pone.0050375-Posthuma2]. It has been well established that the enlarged brain volume is the basis of our unique cognitive capacity, and a reduction of brain volume has been reported in several brain diseases such as schizophrenia [@pone.0050375-Steen1] and Attention-Deficit/Hyperactivity Disorder (ADHD) [@pone.0050375-Valera1]. As a complex quantitative trait with high heritability, brain volume is likely regulated by many genes. But so far only a handful of such genes have been reported by studying patients with rare brain developmental defects, *e.g.* microcephaly [@pone.0050375-Kumar1]. Though the reported microcephalin genes are important in explaining the enlarged human brain during evolution [@pone.0050375-Woods1], recent studies have indicated that they only account for a small part of brain volume variation in the general population [@pone.0050375-Wang1], [@pone.0050375-Rimol1]. Fortunately, recent genome wide association studies have identified several promising loci significantly associated with intracranial volume and head circumference [@pone.0050375-Taal1]--[@pone.0050375-Ikram1]. Nevertheless, all of these studies were performed only in populations of European ancestry and some of the variants (e.g., rs7890687 and rs9915547) identified in these GWAS were fixed (monomorphic) in Chinese population, suggesting that additional genes/variants may modulate brain volume variation in Chinese population.\n\nAdditionally, the genetic dissection of schizophrenia (SCZ), a common mental disorder with high heritability provides opportunities to identify genes associated with brain volume variation since SCZ patients have decreased total brain volume compared to normal controls [@pone.0050375-Steen1], [@pone.0050375-Honea1], [@pone.0050375-Cahn1]. This is consistent with the hypothesis that the pathogenesis of SCZ is related to abnormal brain development [@pone.0050375-Lewis1]. Though numerous linkage and association studies, especially recent genome wide association studies have identified many loci significantly associated with schizophrenia [@pone.0050375-McCarthy1]--[@pone.0050375-Xu1], the etiology of schizophrenia remains poorly understood. Among the hypotheses that explain the etiology of schizophrenia, the neurodevelopmental hypothesis [@pone.0050375-Fatemi1] has been supported by the majority of the published data. This hypothesis predicts that a disruption of brain development during early life underlies the later emergence of psychosis during adolescence or early adulthood. These evidences indicate that schizophrenia susceptibility genes may regulate the unique features of human brain development and dysfunction of these genes likely disrupted the normal development of brain, which eventually lead to schizophrenia susceptibility. In fact, recent studies demonstrated that some schizophrenia susceptibility genes do regulate brain development [@pone.0050375-Mao1], [@pone.0050375-Duan1]. In light of these findings, we hypothesize that schizophrenia susceptibility genes may regulate brain development and affect the total brain volume.\n\nTo detect the relationship between schizophrenia susceptibility genes and brain volume, we earlier systematically studied the genetic association between schizophrenia susceptibility genes and brain volume variation in a large cohort of healthy subjects. This led to identification of a highly significant chromosomal region, 5q23.2--33.1, a region that has been well studied and shown strong association with SCZ in multiple world populations [@pone.0050375-Sherrington1]--[@pone.0050375-Pimm1]. Recently, Chen *et al*. systematically studied this region by using a large sample (N\u200a=\u200a3,422, including case-control and family-based samples) and dense SNP markers. They found haplotypes spanning SPEC2, PDZ-GEF2, LOC728637, and ACSL6 were significantly associated with schizophrenia in five independent samples [@pone.0050375-Chen1], [@pone.0050375-Chen2]. We further replicated the associations in a Chinese sample [@pone.0050375-Luo1]. Collectively, these consistent results strongly suggested genetic variants near these four genes (SPEC2, PDZ-GEF2, LOC728637, and ACSL6) may contribute to schizophrenia susceptibility and brain development.\n\nResults {#s2}\n=======\n\nInterleukin-3 is Strongly Associated with Brain Volume Variation in Chinese {#s2a}\n---------------------------------------------------------------------------\n\nFor the initial analyses in Chinese population, we performed a genetic screening to detect the association of cranial volume (the approximate of brain volume, which is highly correlated with brain volume [@pone.0050375-Jorgensen1], [@pone.0050375-Ricard1]) with sequence variations located in the 5q23.2--33.1 region. The cranial volumes of 1,013 healthy individuals (460 males and 553 females) were measured (see methods), followed by genotyping of 20 single nucleotide polymorphisms (SNPs) in the 5q23.2--33.1 region spanning about 809 kb. To test whether schizophrenia susceptibility variants in 5q23.2--33.1 are associated with brain volume, we initially genotyped 8 tagging SNPs covering the four genes (SPEC2, PDZ-GEF2, LOC728637, and ACSL6). The single SNP association was conducted using linear regression under an additive model and the p-values were obtained by the Wald test as implemented in PLINK [@pone.0050375-Purcell2]. The results showed that six of these 8 SNPs were significantly associated with cranial volume in females, but not in males (**[Table S1](#pone.0050375.s019){ref-type=\"supplementary-material\"}**). For fine-scale mapping, we genotyped another 12 SNPs and identified a sharp signal in the region containing IL3, showing a strong female-specific association with cranial volume ([**Fig. 1a**](#pone-0050375-g001){ref-type=\"fig\"} **and [Table S1](#pone.0050375.s019){ref-type=\"supplementary-material\"}**). Among the 7 highly significant SNPs (-logP\\>3.3) covering IL3, one was located in exon 1 (rs40401, Ser to Pro), one in intron 2 (rs31481), one in the promoter (rs31480), and four in the upstream region (rs3914025, rs3916441, rs31400 and rs3846726) ([**Fig. 1a**](#pone-0050375-g001){ref-type=\"fig\"}), clearly implicating IL3 as the responsible gene. The associations between these 7 SNPs and brain volume were still highly significant (corrected p\\<0.01) even using the most stringent Bonferroni correction for multiple testing (**[Table S1](#pone.0050375.s019){ref-type=\"supplementary-material\"}**). Further haplotype analysis combining the 7 SNPs indicated strong linkage disequilibrium (LD) among the SNPs (**[Fig. S1a](#pone.0050375.s001){ref-type=\"supplementary-material\"}**) with only two major haplotypes, one showing positive association (P\u200a=\u200a4\u00d710^\u22125^), the other showing negative association (P\u200a=\u200a8\u00d710^\u22124^) with cranial volume in females (**[Table S2](#pone.0050375.s020){ref-type=\"supplementary-material\"}**). None of the described associations in females were observed in males (**[Table S3](#pone.0050375.s021){ref-type=\"supplementary-material\"}**), implying that the association of IL3 with brain volume is sex-specific. To capture missing common SNPs, we re-sequenced the IL3 gene region (4 kb) in 150 randomly selected Chinese individuals and found no additional SNPs.\n\n![Genetic association of the 5q23.2--33 SNPs with brain volume and impacts of promoter SNP (rs31480) on the expression of IL3.\\\n**(a)** The distribution of the --logP of the 20 SNPs tested across the 5q23.3--33.1 region (middle panel). The locations of the six known coding genes are displayed. **(b)** The brain volume distributions of the three genotypes at rs31480, on average, TT genotype carriers have a brain volume of 1257 ml and CC genotypes have 1216 ml (\\*\\*\\*P\\<0.001, two tailed Student's *t-test*). **(c)** C allele of rs31480 is completely conserved across a variety of species. **(d)** The oligonucleotides for testing the binding activity of SP1. The predicted binding sequence is underlined containing the rs31480 variation site (red). **(e)** The result of electrophoretic mobility shift assay, showing that the probe containing T allele can bind SP1 (Lane 2) but the C allele cannot (Lane 3). Similar results were obtained using HeLa or MCF7 nuclear extracts (Lane 6 and 7, 10 and 11). Competition experiments using a 100-fold excess of unlabeled probe (Lane 4 and 5, 8 and 9) confirm the specificity of the probe. Binding to the unknown protein/complex was also observed ([Figure 1e](#pone-0050375-g001){ref-type=\"fig\"}, arrow C and D), again, the probe containing T allele showed stronger binding than C allele. **(f)** Assays of promoter activities by relative luciferase expression in HeLa, construct with T allele has significant higher expression activity than C allele. Values of relative luciferase activity are expressed as mean \u00b1 s.d. (results of three independent experiments, each containing three replicates). \\*\\*\\*P\\<0.001 (one tailed Student's *t-test*).](pone.0050375.g001){#pone-0050375-g001}\n\nReplication of the Association between IL3 and Brain Volume Variation in Europeans {#s2b}\n----------------------------------------------------------------------------------\n\nTo confirm our initial findings from the Chinese population, we conducted a replication analysis in three independent samples of European ancestry, for which total brain volume had been determined based on magnetic resonance imaging (MRI). For the seven SNPs showing strong association in Chinese sample, two SNPs in different LD regions in Europeans (rs3916441 and rs40401, **[Fig. S1](#pone.0050375.s001){ref-type=\"supplementary-material\"}**) were included in the replication analysis. Only the healthy controls of these samples were used. We found that the most significant SNP in Chinese, rs3916441, was also significantly associated with total brain volume in the BIG (Dutch Brain Imaging Genetics study) sample (p\u200a=\u200a3.5\u00d710^\u22124^; n\u200a=\u200a486) ([**Table 1**](#pone-0050375-t001){ref-type=\"table\"} **)**. In CBDB/NIMH (Clinical Brain Disorders Branch/National Institute of Mental Health Sibling Study) sample (n\u200a=\u200a188), rs3916441 also showed a trend of association (p\u200a=\u200a0.0516) ([**Table 1**](#pone-0050375-t001){ref-type=\"table\"}). We noticed the female specific association of rs3916441 with brain volume in Chinese was not the situation in European samples. Interestingly, rs3916441 was also significantly associated with total gray matter volume in the CBDB/NIMH and BIG samples ([**Table 2**](#pone-0050375-t002){ref-type=\"table\"}), which may point towards a mechanism explaining the effects of IL3 on brain structure.\n\n10.1371/journal.pone.0050375.t001\n\n###### Replication of the most significantly associated SNPs in genetically divergent populations.\n\n![](pone.0050375.t001){#pone-0050375-t001-1}\n\n  SNP          Polymorphism   Replication samples                                                                                                      \n  ----------- -------------- --------------------- ------------ ------- ------------ ------------ ------- --------------------- ---------------------- -------\n  rs3916441        C/T               0.697          **0.0467**   0.164   **0.0516**   **0.0416**   0.462   **3.5\u00d710** ^\u2212**4**^   **3.08\u00d710** ^\u2212**4**^   0.130\n  rs40401          G/A               0.147            0.261      0.467     0.703        0.580       NA             NA                     NA             NA\n\nNA: Not available.\n\n10.1371/journal.pone.0050375.t002\n\n###### Association of rs3916441 with gray matter and white matter in genetically divergent populations.\n\n![](pone.0050375.t002){#pone-0050375-t002-2}\n\n  SNP          Polymorphism   Replication samples                                                                          \n  ----------- -------------- --------------------- ------- ------- ----------- ----------- ------- ----------- ----------- -------\n  rs3916441        C/T            **0.0228**        0.128   0.110   **0.001**   **0.001**   0.178   **0.002**   **0.004**   0.138\n  rs40401          G/A               0.171          0.128    NA        NA          NA        NA        NA          NA        NA\n\nNA: Not available.\n\nImpacts of rs31480 on Transcription Factor Binding and IL3 Expression {#s2c}\n---------------------------------------------------------------------\n\nTo capture the causal variants of IL3 in Chinese population, we performed bioinformatics analysis for the 7 highly significant SNPs according to their genomic locations and allelic differences in transcription factor binding affinities, and we found SNP rs31480 showing potential functional effects. rs31480 is located in the IL3 promoter (\u221216 bp upstream of the transcription start site (TSS), **[Fig. S2](#pone.0050375.s002){ref-type=\"supplementary-material\"}**), within a putative binding site of the transcription factor SP1. Interestingly, there is a significant difference of 41 ml in the average cranial volume between individuals carrying the two homozygotes at rs31480 (1,257 ml for TT carriers and 1,216 ml for CC carriers, p\u200a=\u200a4.7 \u00d710^\u22124^, two tailed student t-test) ([**Fig. 1b**](#pone-0050375-g001){ref-type=\"fig\"} **and [Table S4](#pone.0050375.s022){ref-type=\"supplementary-material\"}**), suggesting that rs31480 could regulate brain volume variation. In addition, we noticed the C allele (ancestral allele, determined by comparison with the chimpanzee homologous sequence) of rs31480 is completely conserved across a wide variety of species ([**Fig. 1c**](#pone-0050375-g001){ref-type=\"fig\"} **and [Fig. S2](#pone.0050375.s002){ref-type=\"supplementary-material\"}**), also suggesting a functional conservation of rs31480. The T allele (derived allele) is prevalent in East Asian populations (0.556 in Chinese and 0.568 in Japanese), but relatively rare in Europeans (0.198) and Africans (0.136) (<http://www.hapmap.org>).\n\nThe C to T change at rs31480 could change the binding affinity of SP1 and influence the expression of IL3. Electrophoretic mobility shift assay (EMSA) with purified recombinant human SP1 protein ([**Fig. 1d**](#pone-0050375-g001){ref-type=\"fig\"}) showed that SP1 binds to the sequence containing the T allele ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}, lane 2) but not the C allele ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}, lane 3). Similar results were observed when using HeLa ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}, lane 6 and 7) and MCF-7 ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}, lane 10 and 11) cell nuclear extracts as the source of SP1 protein ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}). Finally, competition experiments using unlabeled oligonucleotides corroborated the SP1 binding specificity to the T allele ([**Fig. 1e**](#pone-0050375-g001){ref-type=\"fig\"}, lane 4 and 5, lane 8 and 9). These data suggest that rs31480 has an \"on\" or \"off\" effect on SP1 binding to the IL3 promoter.\n\nTo test whether rs31480 also influences IL3 promoter activity, we performed transactivation assays using the luciferase reporter gene. The promoter region encompassing nucleotides \u2212436 to +164 (relative to the ATG start codon at +1) of IL3 was amplified by PCR from genomic DNA of two individuals homozygous with respect to the corresponding genotypes (TT and CC) for rs31480. Sequencing analysis of the amplified promoter fragments did not detect other sequence differences except for rs31480. As shown in [**Fig. 1f**](#pone-0050375-g001){ref-type=\"fig\"} and **[Fig. S3a](#pone.0050375.s003){ref-type=\"supplementary-material\"}--c**, the transcriptional activity of the IL3 promoter containing the T allele was indeed significantly higher than that of the C allele in all cell lines tested (Hela, CHO, SK-N-SH, and COS-7). The T allele of rs31480 thus enhances the IL3 promoter activity through the binding of transcription factor SP1. For the most significant SNP rs3916441, since our functional prediction analysis did not give any hint for the functional role of this SNP, whether it plays any functional role for IL3 is yet to be determined.\n\nIt should be noted that, rs31480 was not significantly associated with brain volume in the BIG sample of Europeans, and it was not available in the ANDI and CBDB/NIMH samples. Another SNP rs40401 in high linkage with rs31480 was also not significant in these samples. The differences in association for this SNP (rs31480) are likely due to the genetic heterogeneity between Chinese and Europeans as shown in **[Fig. S1](#pone.0050375.s001){ref-type=\"supplementary-material\"}**.\n\nIL3 and its Receptors are Mainly Expressed in Neural Progenitors and Mature Neurons {#s2d}\n-----------------------------------------------------------------------------------\n\nIL3 exerts its biological effects through a receptor that is composed of a ligand-specific \u03b1 (IL3RA) subunit and a signal transducing \u03b2 subunit (IL3RB) common to IL3/IL5/GM-CSF. The mouse IL3 receptor has two distinct \u03b2 subunits, one that functions only in IL3-mediated cell signaling (\u03b2IL3) and a second that is shared with IL5 and GM-CSF (IL3RB or CSF2RB). We studied the expression of IL3 and its receptors in the developing mouse brain and found that IL3 and its receptors were continuously expressed in mouse brain from embryonic day (E) 12.5 to adult life as revealed by RT-PCR ([**Fig. 2a**](#pone-0050375-g002){ref-type=\"fig\"}), with a peak expression level at postnatal day (P) 1 to 4 ([**Fig. 2b**](#pone-0050375-g002){ref-type=\"fig\"}), a stage with active neural proliferation and neurogenesis. We also noticed that \u03b2IL3 is only expressed from E14.5 to P7 ([**Fig. 2a**](#pone-0050375-g002){ref-type=\"fig\"}), a stage accompanied by the dramatic increase of the neocortex volume [@pone.0050375-Zhang1]. Immunostaining revealed that IL3 and its receptors were mainly expressed in the neocortex region of the mouse brain ([**Fig. 2c--n**](#pone-0050375-g002){ref-type=\"fig\"} and **[Fig. S4](#pone.0050375.s004){ref-type=\"supplementary-material\"}** and **[S5](#pone.0050375.s005){ref-type=\"supplementary-material\"}**). We also detected weak expression of IL3RA in the CA1 and CA3 regions of hippocampus (**[Fig. S6](#pone.0050375.s006){ref-type=\"supplementary-material\"}**), hilus of the dentate gyrus (**[Fig. S7](#pone.0050375.s007){ref-type=\"supplementary-material\"}**), and lateral septal nucleus, dorsal part (LSD) (**[Fig. S8](#pone.0050375.s008){ref-type=\"supplementary-material\"}**). Compared to IL3RA, IL3RB showed higher expression in the mouse brain ([**Fig. 2a, b**](#pone-0050375-g002){ref-type=\"fig\"}). It was extensively expressed in mouse brain including neocortex ([**Fig. 2i--k**](#pone-0050375-g002){ref-type=\"fig\"}) and hippocampus (**[Fig. S6](#pone.0050375.s006){ref-type=\"supplementary-material\"}**). Since all IL3RA positive cells also expressed IL3RB (**[Fig. S6](#pone.0050375.s006){ref-type=\"supplementary-material\"}** and **S9**), we focused on IL3RA hereinafter. To characterize IL3RA-positive cells, we performed co-immunostaining and found IL3RA was expressed in SOX2- and nestin-positive neural progenitors at early developmental stage ([**Fig. 3a--e**](#pone-0050375-g003){ref-type=\"fig\"}, **[Fig. S4](#pone.0050375.s004){ref-type=\"supplementary-material\"}** and **S10**). As development continues, the expression of IL3RA was down-regulated in neural progenitors ([**Fig. 3f--h**](#pone-0050375-g003){ref-type=\"fig\"}). After birth, expression of IL3RA was found in some Tuj1-positive neurons (**[Fig. S11](#pone.0050375.s011){ref-type=\"supplementary-material\"}**). However, we noticed that many IL3RA positive cells were Tuj1-negative ([**Fig. 3i--k**](#pone-0050375-g003){ref-type=\"fig\"} **, [Fig. S11a](#pone.0050375.s011){ref-type=\"supplementary-material\"}--i**). Co-labeling with GFAP excluded their identity as glial cells ([**Fig. 3l--n**](#pone-0050375-g003){ref-type=\"fig\"}). Further double immunostaining showed many IL3RA positive cells also weakly expressed Tbr2, a marker for intermediate progenitor cells (IPCs) ([**Fig. 3o--q**](#pone-0050375-g003){ref-type=\"fig\"} and **[Fig. S12](#pone.0050375.s012){ref-type=\"supplementary-material\"}**). In contrast, in addition to co-expression with IL3RA, IL3RB was expressed in neurons and glial cells (**[Fig. S13](#pone.0050375.s013){ref-type=\"supplementary-material\"}**). Taken together, these results demonstrate that IL3 and its receptors are mainly expressed in neural progenitors and neurons in the developing neocortex.\n\n![Spatiotemporal expression profiling of IL3 and its receptor in the developing mouse brain.\\\n(**a**) RT-PCR revealed the expression of IL3 and its receptor in developing mouse brain from E12.5 to adult. (**b**) Quantitative PCR showed that the expression of IL3 and its receptors peaks at P1--P4, a period with active neural proliferation and neurogenesis. Data are expressed as mean \u00b1 s.e.m. (n\u200a=\u200a3). (**c--n**) Immunohistochemistry analysis indicated that IL3 and its receptor were expressed in the mouse brain. Co-expression of IL3 and IL3RA were detected (arrows in **l--n**), indicating the activation of IL3-mediated signaling pathways in the developing mouse brain. (**o--p**) IL3RA is expressed in radial glia (resides in ventricular zone and characterized by long radial processes, arrowhead in **o**) and migratory neurons (arrowheads in **p**). Ctx, cortex; VZ, ventricular zone. Scale bars, (c, d, e) 25 \u00b5m; O, 10 \u00b5m.](pone.0050375.g002){#pone-0050375-g002}\n\n![IL3RA is mainly expressed in neural progenitors and neurons.\\\n(**a, b**) Expression of IL3RA was detected in SVZ and IZ regions. Co-labeling with SOX2 showed IL3RA expression cells in SVZ are SOX2 positive, indicating these cells are neural progenitors. However, IL3RA positive cells in IZ are SOX2 negative, indicating these cells are not neural progenitors. (**c--e**) In the early stages of brain development, co-expression of IL3RA and SOX2 was found in neocortex region. With the development of the central nervous system, expression of SOX2 was down-regulated or disappeared in IL3RA positive cells (**f--h**). We also found many IL3RA positive cells were not mature neurons (**i--k**) or glial cells (**l--n**). Co-labeling with TBR2 demonstrated IL3RA positive cells are intermediate progenitor cells (IPCs) (**o--q**). VZ, ventricular zone; SVZ, subventricular zone; IZ, intermediate zone; CP, cortex plate. Scale bars, 25 \u00b5m.](pone.0050375.g003){#pone-0050375-g003}\n\nIL3 Promotes the Proliferation of Neural Progenitors {#s2e}\n----------------------------------------------------\n\nIL3 is known to activate three signaling pathways, the JAK/STAT, the MAPK, and the PI3K/AKT pathways, among which the MAPK signal pathway regulates cell proliferation [@pone.0050375-Reddy1]. Since we found IL3RA co-expressed with the cell proliferation markers Ki67 and pH3 (**[Fig. S14](#pone.0050375.s014){ref-type=\"supplementary-material\"}**), IL3 could play a role in promoting the proliferation of neural progenitors. We thus examined the effect of IL3 treatment on the proliferation of neural progenitors isolated from E13.5 cortex. We first verified the expression of IL3RA and IL3RB in cultured neural progenitors (**[Fig. S15a](#pone.0050375.s015){ref-type=\"supplementary-material\"}**). Anti-pH3 immunolabeling showed that the number of proliferating cells was significantly increased in IL3-treated samples compared to the controls ([**Fig. 4a,b**](#pone-0050375-g004){ref-type=\"fig\"}), which is consistent with published observations [@pone.0050375-Cattaneo1]. Western blotting confirmed that IL3 could activate MAPK pathway in both MEG01 cells ([**Fig. 4f**](#pone-0050375-g004){ref-type=\"fig\"}) and cultured neural progenitors ([**Fig. 4g**](#pone-0050375-g004){ref-type=\"fig\"}). The phosphorylation of MAPK1/2 was significantly increased after IL3 treatment, indicating that IL3 can activate proliferation pathway in neural progenitors. We also investigated another proliferation related pathway, the JAK/STAT pathway. We found that JAK2 phosphorylation was increased after IL3 treatment ([**Fig. 4g**](#pone-0050375-g004){ref-type=\"fig\"}), further supporting the involvement of IL3 in the proliferation of neural progenitors.\n\n![IL3 promotes proliferation and survival of neural progenitors.\\\n(**a**) IL3 promotes neural progenitor's proliferation. (**b**) Quantification of proliferating cells (pH3+) after IL3 treatment. ^\\#^ *P*\\<0.05 (n\u200a=\u200a4, one-tailed Student's *t*-test). (**c**) Trophic effect of IL3 on neural progenitors. Neural progenitors from E12.5 mice were grown in the absence of any factor and in the presence of different concentrations of IL3 for 36 hours, then cell viability were determined. Note that 3.0 ng/ml IL3 could promote survival of neural progenitors significantly (n\u200a=\u200a8 for each condition). (**d--e**) Neurotrophic effects of IL3 on neural progenitors and neurons. Neural progenitors were first cultured in neurobasal medium with B27 supplement for about 24 hours, then the medium was replaced with neurobasal medium containing N2 supplement and different concentrations of IL3 were added. The cultures were maintained for 3 days and cell viability was determined. IL3 has significant effects on this culture condition on progenitors (**d**) and neurons (**e**) (n\u200a=\u200a8 for control group and 20 ng/ml group, n\u200a=\u200a16 for other groups). *y*-axis, cell viability (normalized to control), *x*-axis, concentration of IL3 (ng/ml). Data are expressed as mean \u00b1 s.e.m. \\**P*\\<0.05, \\*\\**P*\\<0.01 (two-tailed Student's *t-test*). (**f**) IL3 activates PI3K-AKT, MAPK1/2 and Gsk3\u03b2 signal pathways in MEG01 cell line. (**g**) IL3 activate MAPK, JAK/STAT, and PI3K/AKT pathways in primary cultured neural progenitors. The phosphorylation level of AKT, MAPK1/2, JAK2 and GSK3\u03b2 was increased after IL3 treatment.](pone.0050375.g004){#pone-0050375-g004}\n\nWe next tested whether IL3 could drive neuronal differentiation *in vitro*. The progenitor cells were cultured under differentiation condition, treated with IL3, and the expression level of cell type-specific markers was measured by quantitative PCR. We found that IL3 had no effect on neuronal differentiation *in vitro*. After IL3 (10 ng/ml) treatment, the expression of all tested genes was not changed significantly (**[Fig. S16](#pone.0050375.s016){ref-type=\"supplementary-material\"}**). Collectively, these data suggest that IL3 promotes the proliferation of neural progenitors through the activation of MAPK and JAK/STAT pathways, but has no effects on neural differentiation.\n\nNeurotrophic Effects of IL3 on Neural Progenitors and Neurons {#s2f}\n-------------------------------------------------------------\n\nIL3 is reported to have trophic effects on neurons [@pone.0050375-Kamegai1]. It promotes the survival of sensory neurons and protects against neuronal death induced by FeSO4 and A\u03b2 [@pone.0050375-Wen1], [@pone.0050375-Zambrano1]. We speculated that IL3 might also have similar trophic effects on cortical neural progenitors. To test this, we determined cell viability of cultured neural progenitors and neurons in media with different growth factors and supplements. The results suggest that when the nutrition is deficient, the IL3 pathway could be activated to protect against cell death induced by starvation (**[Fig. S15b](#pone.0050375.s015){ref-type=\"supplementary-material\"},c**, [**Fig. 4c,d**](#pone-0050375-g004){ref-type=\"fig\"}). Similar results were obtained on neurons when cultured using the previously reported culture method [@pone.0050375-Wen1] ([**Fig. 4e**](#pone-0050375-g004){ref-type=\"fig\"}).\n\nThe Bcl-x~L~ was reported has a role in neuronal survival mediated by IL3 pathway [@pone.0050375-Wen1], so we studied the expression of Bcl-x~L~ in neural progenitors and found there was no significant change after treated with IL3 (**[Fig. S15e](#pone.0050375.s015){ref-type=\"supplementary-material\"}**). The signaling through the PI3K/AKT pathway is one of the most potent intracellular mechanisms to promote cell survival. It is well established that IL3 can activate the PI3K/AKT pathway [@pone.0050375-Reddy1]. To further study the mechanism of neural progenitor survival mediated by IL3 and to test if the PI3K/AKT pathway participates in IL3-mediated survival of neural progenitors, we studied the interactions between IL3 and AKT1 in neural progenitors by western blotting. In untreated progenitors, the level of phosphorylated AKT (the active form) is low ([**Fig. 4g**](#pone-0050375-g004){ref-type=\"fig\"}). However, the level of phosphorylated AKT was dramatically increased after IL3 treatment ([**Fig. 4g**](#pone-0050375-g004){ref-type=\"fig\"}). In addition, analysis of three AKT1 SNPs indicated significant association with brain volume (**[Table S5](#pone.0050375.s023){ref-type=\"supplementary-material\"}**). Taken together, these results indicate that IL3 promotes the survival of neural progenitors by activating the PI3K/AKT pathway.\n\nIL3 Activates Estrogen Receptor \u03b1 and \u03b2 *in vitro* {#s2g}\n--------------------------------------------------\n\nAs shown above, the genetic association of the IL3 SNPs with brain volume was female-specific in Chinese. To test whether estrogen could regulate the expression of IL3, we treated the K562 cell line (which expresses both IL3 and estrogen receptor (**[Fig. S17a](#pone.0050375.s017){ref-type=\"supplementary-material\"}**)) with estrogen and found no overt change of IL3 expression (**[Fig. S17c](#pone.0050375.s017){ref-type=\"supplementary-material\"}**--**e**), while the expression of TFF1 as a control was significantly increased after estrogen treatment (P\\<0.001) (**[Fig. S17b](#pone.0050375.s017){ref-type=\"supplementary-material\"}**--**d**). These results suggest that the transcription of IL3 could not be directly regulated by estrogen.\n\nIt was reported that the MAPK and PI3K/AKT pathways could activate the estrogen receptor (ER) [@pone.0050375-Kato1], [@pone.0050375-Campbell1]. To investigate whether IL3 could activate ER genes through the two regulated pathways, we constructed three vectors, 3ERE-PGL3 (contains three repeats of estrogen response element (ERE)), ER~\u03b1~ and ER~\u03b2~ respectively. We first studied the expression of IL3 receptors and estrogen receptors in MEG-01 and HEK293T cell lines and we found IL3 receptors were expressed in both cell lines ([**Fig. 5a,b**](#pone-0050375-g005){ref-type=\"fig\"}). The vectors (3ERE-PGL3 and ER~\u03b1,~ or 3ERE-PGL3 and ER~\u03b2~) were then co-transfected into the IL3 receptor-expressing MEG-01 cell line, followed by IL3 or E2 treatments. As expected, IL3 could activate both ER genes, especially ER~\u03b2~ ([**Fig. 5c,d**](#pone-0050375-g005){ref-type=\"fig\"}). In addition, the activation of ER~\u03b2~ induced by E2 was enhanced by IL3. We confirmed these results in HEK293T cells ([**Fig. 5e,f**](#pone-0050375-g005){ref-type=\"fig\"}). Hence, the activation of ER genes by IL3 and estrogen and the interaction between them may explain the sex-specific functional effects of the sequence polymorphism at rs31480 on brain volume in females. The reported association studies on SCZ patients is consistent with our observation, in which IL3 showed a significant association with schizophrenia only in females. The hypothesized sex-specific regulation of brain volume is illustrated in [**Fig. 6**](#pone-0050375-g006){ref-type=\"fig\"} **and [Fig. S18](#pone.0050375.s018){ref-type=\"supplementary-material\"}.**\n\n![IL-3 activates estrogen receptors in MEG01 and HEK293T cells.\\\nThere are two estrogen receptors, ER~\u03b1~ and ER~\u03b2~. MEG01 and HEK293T cell lines were used to test whether IL3 can activate estrogen receptors. Expression of IL-3 receptors (IL3RA, CSF2RB) were verified by RT-PCR in MEG01(a) and HEK293T (b) cell lines. In HEK293T cell line, we also detected the expression of estrogen receptors (ER~\u03b1~ and ER~\u03b2~) (b). Constructs containing three repeats of estrogen response element (3ERE) and estrogen receptor (ER~\u03b1~ or ER~\u03b2~) were co-transfected into MEG01 and HEK293T cell lines prior to IL-3 or estrogen (E2) treatment. IL-3 can activate ER~\u03b1~ and ER~\u03b2~ in both cell lines (c-d for MEG01 and e-f for HEK293T), and this effect was enhanced by the estrogen, indicating there were interactions between IL-3 and estrogen activation. Data are expressed as mean \u00b1 s.e.m. (three independent assays, each containing 3 replicates).\\\n^\\#^P\\<0.10,\\\n\\*P\\<0.05,\\\n\\*\\*P\\<0.01,\\\n\\*\\*\\*P\\<0.001.](pone.0050375.g005){#pone-0050375-g005}\n\n![Model for regulation of brain volume by IL3 genetic variation and expression level.\\\nIndividuals with different genotypes at rs31480 have differential expression level of IL3 (individuals with TT genotype have higher IL-3 expression than CC carriers), which lead to differential activation of signaling pathways mediated by IL3. The differential activation of signaling pathways further influence the proliferation and survival of neural progenitors, eventually lead to brain volume variation for individuals with different genotypes at rs31480.](pone.0050375.g006){#pone-0050375-g006}\n\nDiscussion {#s3}\n==========\n\nBrain volume is an important quantitative trait that underlies our most complex cognitive abilities and human evolution is characterized by a dramatic expansion in brain size and complexity. Undoubtedly, our unique genetic makeup played a decisive role in our enlarged brain and human brain size must be highly regulated during development. The hypothesis that SCZ is a brain disease unique to humans [@pone.0050375-Dean1] suggests that SCZ susceptibility genes may regulate the unique features of human brain development and dysfunction of these genes may disrupt the normal development of the human brain. In this study, we provide evidence that IL3 may regulate human brain volume variation. First, our genetic association results strongly suggest the association between IL3 and brain volume. IL3 is located in 5q31.1, one of the most successfully replicated regions that may harbor SCZ susceptibility genes [@pone.0050375-Sherrington1]--[@pone.0050375-Luo1]. In fact, 5q23--31 was ranked number 2 of all chromosomal regions implicated to harbor SCZ susceptibility genes in a genome-wide meta-analysis of SCZ [@pone.0050375-Lewis2]. As accumulating data support that SCZ is a neurodevelopmental disorder, it's likely that there are potential genes in 5q23.2--33.1 that regulate brain development. However, though studies repeatedly found association between genes located in 5q23.2--33.1 and SCZ, the detailed expression pattern of these genes and their function during brain development is not known. Here, for the first time, we detailed the spatiotemporal expression pattern of IL3 and its receptors in developing mouse brain and we found IL3RA is mainly expressed in neural progenitors and neurons, which also support the importance of IL3 signaling pathway in brain development. Also, our *in vitro* proliferation and survival assays further validate the pivotal roles of IL3 in the development of central nervous system. Collectively, these results provide novel insights to the involvement of IL3 in brain development, supporting the neurodevelopmental hypothesis of schizophrenia.\n\nIt should be noted that during the initial screening in the Chinese sample, we used cranial volume as proxy of brain volume. Though cranial volume is not exactly equal to brain volume, the correlation between these two variables is very high [@pone.0050375-Ricard1]. More importantly, we have successfully identified the association between cranial volume and MCPH1 gene by applying this method in our previous study [@pone.0050375-Wang1]. In addition, the successful replication of our initial findings in genetically divergent populations further support the reliability of our method.\n\nWe realized that the association data and the functional data did not refer to the same SNP, which could be explained by several possible reasons. First, though rs3916441 has the smallest p value in our screening sample, it is located about 27 kb upstream of IL3, and the likelihood of rs3916441's direct regulation of IL3 expression is relatively small. Second, the p value of rs3916441 and rs31480 is very close in our screening sample, and they are highly linked (r^2^\u200a=\u200a0.89) in Chinese. Hence, the functional data suggests rs31480 is probably the causal SNP in brain volume regulation. Nevertheless, we have successfully replicated the significant associations of IL3 variants with brain volume in BIG sample (rs3916441), and we also observed a marginal significant association in CBDB/NIMH sample (rs3916441). Although the association of these SNPs did not reach genome-wide significance, considering the non-overlap of the studied samples and different genetic backgrounds of Chinese and Europeans, IL3 is likely an authentic gene contributing to brain volume variation in general populations. To date, no genes have been shown significantly associated with brain volume and only a few genes were associated with intracranial volume in recent genome-wide association studies [@pone.0050375-Taal1]--[@pone.0050375-Ikram1], suggesting an extremely complicated genetic regulation of brain volume.\n\nGrowing evidence have suggested that the interaction between immune and nervous systems may play an important role in the pathogenesis of schizophrenia [@pone.0050375-Potvin1]. The immune and nervous systems interact with each other through cytokines, a family of proteins that are secreted by a specific group of cells of the immune system and have pleiotropic effects on many cell types, including proliferation, differentiation, and survival. IL-3 is a cytokine that induces growth and differentiation of hematopoietic stem cells and a variety of cell types originating in the bone marrow. Recent studies have demonstrated the important role of IL3 in the central nervous system (CNS). It is expressed in the hippocampus and cortices of normal mouse brain [@pone.0050375-Tabira1], and it stimulates the growth and proliferation of microglial cells [@pone.0050375-Frei1], [@pone.0050375-Frei2]. Studies also found that IL3 facilitates the survival of sensory neurons significantly and stimulates the formation of the neural network [@pone.0050375-Moroni1]. In addition, IL-3 has been found to be able to promote the process extension of cultured cholinergic [@pone.0050375-Kamegai1] and prevent delayed neuronal death in the hippocampus [@pone.0050375-Wen1]. In fact, rat interleukin 3 receptor \u03b2-subunit was cloned from cultured microglia [@pone.0050375-Appel1], and disruption of IL3 production in brain led to neurologic dysfunction [@pone.0050375-Cockayne1]. All of these studies strongly suggest that IL3 is a pivotal protective factor for CNS. More importantly, Chen *et al.* recently reported that IL3 was significantly associated with brain disease such as schizophrenia in three independent Irish samples [@pone.0050375-Chen2], [@pone.0050375-Chen3], [@pone.0050375-Edwards1]. IL3 receptors, including IL3RA and CSF2RB (or IL3RB), were also found significantly associated with schizophrenia in three different populations [@pone.0050375-Lencz1]--[@pone.0050375-Chen4]. In addition, decreased IL-3 levels in the first-episode and drug-na\u00efve patients with schizophrenia was also reported [@pone.0050375-Xiu1]. These convergent evidences strongly indicate the involvement of the IL3 pathway in schizophrenia. Interestingly, we noticed that rs3916441, which is most significantly associated with brain volume, was also significantly associated with schizophrenia in females [@pone.0050375-Chen2], implying the interaction between IL3 and gender may play vital roles in normal brain development and schizophrenia susceptibility. Though many investigations support the involvement of IL3 in brain function and schizophrenia, the precise expression pattern of IL3 and its receptors in developing brain is not well characterized and it's not clear how genetic variation within IL3 affect brain development and schizophrenia susceptibility.\n\nIn summary, we have demonstrated that IL3 plays crucial roles in the development of the central nervous system. We identified a genetic variant (rs31480) in the promoter of IL3 that is significantly associated with brain volume in the general population. This polymorphism influences the expression of IL3 and the differential IL3 expression of the two alleles at rs31480 can influence neural progenitor pool expansion and maintenance during neurodevelopment. Furthermore, our findings that IL3 can promote proliferation and survival of neural progenitors further support the proposed novel role of IL3 in the central nervous systems.\n\nMaterials and Methods {#s4}\n=====================\n\nSamples {#s4a}\n-------\n\nOur screening samples are from Yunnan province of southwestern China (n\u200a=\u200a1,013). Replication samples included samples from the Alzheimer's Disease Neuroimaging Initiative (ADNI; n\u200a=\u200a204), samples from the US National Institute of Health (CBDB/NIMH; n\u200a=\u200a188) and samples from the Dutch Brain Imaging Genetics study (BIG; n\u200a=\u200a486).\n\nScreening Samples: Chinese Samples {#s4b}\n----------------------------------\n\nThe detailed information of the sample screening was described in our previous study [@pone.0050375-Wang1]. Briefly, a total of 1,013 unrelated healthy individuals including 460 males and 553 females were included. The identities of the subjects were self-declared and confirmed by their written ID profiles. All the sampled individuals are from Yunnan province of southwestern China. Written informed consents for this study were obtained from all the subjects, and the research protocol was approved by the internal review board of Kunming Institute of Zoology, Chinese Academy of Sciences. The ages of the 1,013 individuals range from 19 to 28 years with 98% of them being 21--26 years old.\n\nReplication Samples: ADNI Sample {#s4c}\n--------------------------------\n\nThe MRI and genotyping data in this replication sample were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.ucla.edu). One goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer's disease (AD). For up-to-date information, see [www.adni-info.org](http://www.adni-info.org). The ADNI participants consist of patients with AD, patients with MCI, and elderly healthy individuals. They were aged 55--90 years and recruited from 59 sites across the U.S. and Canada. Written informed consent was obtained from all 822 participants and the study was conducted with prior Institutional Review Board approval. Of 822 participants, 204 unrelated non-Hispanic Caucasian healthy controls were used in this study [@pone.0050375-Saykin1].\n\nReplication Samples: CBDB/NIMH Sample {#s4d}\n-------------------------------------\n\nAll subjects are the healthy control participants of Clinical Brain Disorders Branch/National Institute of Mental Health Sibling Study, a study aimed at identifying schizophrenia susceptibility genes and related intermediate biologic phenotypes [@pone.0050375-Goldman1]. Subjects with good quality of structural data and genotyping were included the study.\n\nReplication Samples: BIG Sample {#s4e}\n-------------------------------\n\nIn this study, a total of 486 healthy control subjects aged 18--35 years from the Brain Imaging Genetics (BIG) study at the Donders Institute for Brain, Cognition and Behaviour of the Radboud University Nijmegen (Medical Centre) were included. The BIG study is a study of self-reported healthy individuals included into earlier imaging studies at the Donders Centre for Cognitive Neuroimaging. Subjects are of European Caucasian descent and generally highly educated [@pone.0050375-Franke1]. The study was approved by the regional medical ethics committee (CMO regio Arnhem/Nijmegen) and all participants provided written informed consent prior to participation.\n\nMeasurement of Cranial Volume: Chinese Screening Sample {#s4f}\n-------------------------------------------------------\n\nThe cranial volume was measured and calculated as described in our previous study [@pone.0050375-Wang1]. Three principal dimensions of the cranium were measured including 1) Maximum antero-posterior length (L, measured between glabella and the inion). 2) Maximum breadth (B, biparietal diameter; measured between two parietal eminences). 3) Cranial height (H, basi-bregmatic height, measured between the internal acoustic meatus to the highest point of the vertex). Then the cranial volumes were computed using the following formula [@pone.0050375-Wang1], [@pone.0050375-Manjunath1]: Male, 0.337 (L-1.1) (B-1.1) (H-1.1) +406.01 cc; Female, 0.400 (L-1.1) (B-1.1) (H-1.1) +206.60 cc.\n\nMeasurement of Brain Volume: Replication Samples (ANDI) {#s4g}\n-------------------------------------------------------\n\n3D T1-weighted brain MRI scans were acquired using a sagittal 3D MP-RAGE sequence following the ADNI MRI protocol [@pone.0050375-Jack1]. Baseline 1.5T MRI scans from 204 participants were downloaded from the ADNI public website (<http://www.loni.ucla.edu/ADNI/>) onto local servers at Indiana University School of Medicine. As detailed in previous studies [@pone.0050375-Risacher1], FreeSurfer V4 software (<http://surfer.nmr.mgh.harvard.edu/>), a widely employed brain segmentation and cortical parcellation tool, was used to label cortical and subcortical tissue classes using an atlas-based Bayesian segmentation procedure and to extract the measure of brain volume.\n\nMeasurement of Brain Volume: Replication Samples (CBDB/NIMH) {#s4h}\n------------------------------------------------------------\n\nAll structural MRI were acquired on a 1.5 Tesla GE scanner (GE Medical Systems, Milwaukee, Wisconsin) using a T1-weighted spoiled gradient recalled (SPGR) sequence (repetition time, 24 ms; echo time, 5 ms; number of excitations, 1; flip angle, 45\u00b0; matrix size 256\u00d7256; FOV 24\u00d724 cm^2^), with 124 sagittal slices (0.94\u00d70.94\u00d71.5 mm^3^ resolution). Images were processed using the FreeSurfer [@pone.0050375-Buckner1] toolbox (version 5). Total Brain Volume (TBV) and Total Grey Matter volume (TGM) measurements were calculated as previously described [@pone.0050375-Goldman1], [@pone.0050375-Goldman2]. TGM was defined as sum of tissue probabilities for the grey matter region. TBV was defined as the sum of total gray matter volume, total white matter volume, and cerebrospinal fluid.\n\nMeasurement of Brain Volume: Replication Samples (BIG Sample) {#s4i}\n-------------------------------------------------------------\n\nSubjects were scanned at 3 Tesla (n\u200a=\u200a486) MRI scanners and T1-weighted structural magnetic resonance imaging data (3D MPRAGE) were acquired (more information on the image acquisition can be found in our previous study [@pone.0050375-Bralten1]). All scans covered the entire brain and had a voxel-size of 1\u00d71\u00d71 mm^3^. To calculate total brain volume, raw DICOM MR imaging data were converted to NIFTI format using the conversion as implemented in SPM5 (<http://www.fil.ion.ucl.ac.uk/spm/software/spm5/>). Normalizing, bias-correcting, and segmenting into gray matter, white matter, and cerebrospinal fluid was performed using the VBM toolbox (VBM5.1 Toolbox version 1.19) in SPM using priors (default settings). This method uses an optimized VBM Protocol [@pone.0050375-Ashburner1], [@pone.0050375-Good1] as well as a model based on Hidden Markov Random Fields (HMRF) developed to increase signal-to-noise ratio [@pone.0050375-Cuadra1]. Total volume of gray matter, white matter, and cerebrospinal fluid was calculated by adding the resulting tissue probabilities. Total brain volume was defined as the sum of white matter and gray matter volume.\n\nSNP Selection, SNP Tagging, Genotyping and Sequencing {#s4j}\n-----------------------------------------------------\n\nSNP selection was based on the previous association studies including our recent data [@pone.0050375-Chen1]--[@pone.0050375-Luo1]. We focused on the four genes that identified by Chen *et al.* recently using systematically mapping in large independent samples. In addition, our recent data and LD in Chinese were also considered. We selected 8 SNPs for the initial screening (rs3756295, rs40396, rs1291602, rs31251, rs1355095, rs2240525, rs3914025, rs31400), additional SNPs were included according to the association significance and whether they are tagging SNPs. Totally, we selected 20 SNPs for fine mapping. The 20 SNPs were genotyped using the SNaPshot method (Applied Biosystems). We sequenced the IL3 gene (including the 500 bp upstream and downstream, respectively) in 150 randomly selected individuals through direct sequencing. The conservation analyses were performed by using UCSC genome browser [@pone.0050375-Fujita1]. (<http://genome.ucsc.edu/>).\n\nPrediction of DNA-binding Motifs {#s4k}\n--------------------------------\n\nWe used Dragon ERE finder [@pone.0050375-Bajic1], a web-based program for identification and interactive analyses of estrogen response elements (EREs) to predict EREs in the upstream region of IL3. AliBaba (<http://www.gene-regulation.com/pub/programs/alibaba2/index.html>) was used to predict and compare DNA-binding motifs in the promoter region with alternative alleles.\n\nCell Culture, Treatment, and RNA Extraction {#s4l}\n-------------------------------------------\n\nK562 cells were routinely cultured in DMEM (Gibco) supplemented with 10% FBS (Hyclone), 100 u/ml penicillin and 100 ug/ml streptomycin. Before treatments, the cells were maintained in phenol red-free DMEM containing 10% dextran-coated charcoal-stripped *fetal bovine serum* (*DCC*-*FBS*) (Hyclone) for a minimum of 3 days with the media changed every day. Cells were treated with 10 nM 17-beta-estradiol (E2) (Sigma) for 2 to 24 hours. Total RNA was harvested and prepared using TRIzol (Invitrogen) following the manufacturer's instructions.\n\nQuantitative Real-time PCR {#s4m}\n--------------------------\n\nReverse transcription PCR (RT-PCR) was performed using the Omniscript RT Kit (Qiagen) following the manufacturer's instructions. We carried out real-time quantitative PCR using gene specific primers, and the fold change in expression was calculated using the \u0394\u0394C~t~ (threshold cycle) method. The GAPDH was used as the internal control.\n\nEMSA {#s4n}\n----\n\nEMSAs were performed with a Lightshift\u2122 chemilumescent EMSA kit (Pierce). The single-strand oligonucleotides were biotinylated with Biotin 3\u2032 End DNA labeling Kit (Pierce) and then annealed to form double strands. The nuclear extracts of MCF-7 and U2OS were prepared by CelLytic\u2122 NuCLEAR\u2122 Extraction kit (Sigma). HeLa nuclear extracts were purchased from Santa Cruz Biotech. The binding reactions were performed for 20 mins at room temperature in 10 mM Tris-HCl (PH 7.5), 1 mM MgCl~2~, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 50 ug/ml poly (dI-dC)(dI-dC) and 4% glycerol, 35 fmol biotin 3\u2032-end -labeled double-stranded oligonucleotides, and purified recombinant SP1 protein (Alexis) or nuclear extracts. After incubation, samples were separated on a native 6% polyacrylamide gel and then transferred to a nylon membrane. The positions of biotin end-labeled oligonucleotides were detected by a chemilumescent reaction with streptavidin-horseradish peroxidase according to the manufacturer's instructions and visualized by autoradiography. For competition assays, we pre-incubated 100-fold excess of unlabeled oligonucleotide probe with SP1 or nuclear extracts before adding the biotin-labeled probe. The nucleotide sequences of the double-stranded oligonucledtides with either T or C allele are:\n\nT-allele: 5\u2032-TGCACATATAAGGCGGGAGG[T]{.ul}TGTTGCCAACGCTTCAGAGC-3\u2032.\n\nC-allele: 5\u2032-TGCACATATAAGGCGGGAGG[C]{.ul}TGTTGCCAACGCTTCAGAGC-3\u2032.\n\nPromoter Cloning and Reporter Gene Assays {#s4o}\n-----------------------------------------\n\nTo construct IL3 promoter, we amplified fragments encompassing nucleotides -436 to +164 (relative to the ATG start codon at +1) of IL3 by PCR from genomic DNA of two individuals homozygous with respect to the corresponding genotypes (TT and CC) for rs31480, using primers tailed with *Xhol* and *HindIII* restriction sites, and directionally subcloned them into the *Xhol* and *HindIII* sites of the pGL3-Basic expression vector (Promega). We verified all recombinant clones by bi-directional DNA sequencing. HeLa, COS-7, CHO, and SK cells were routinely cultured in DMEM supplemented with 10% FBS with antibiotics. The cells were plated at 2.5\u00d710^5^ cells per well in a 24-well plate the day before transfection and incubated overnight at 37\u00b0C in 5% CO~2~. Transient transfection assays were conducted in these cells using the Lipofectamine 2000 transfection reagent (Invitrogen), all assays were performed in at least three independent experiments with minimum of three replicates. The reporters containing either T allele or C allele were transfected into these cells together with a *Renilla* luciferase control vector. After 24h incubation, we collected the cells and measured luciferase activity using the Dual-Luciferase Reporter Assay System (Promega).\n\n3ERE Cloning and IL3 Activation Assays {#s4p}\n--------------------------------------\n\nThree repeats of Estrogen Response Elements (EREs) tailed with *Xhol* and *HindIII* restriction sites were synthesized and annealed to form double-stranded nucleotides. The sequence is 5\u2032-CCG CTCGAG TA GGTCA GCG TGACC TA TA GGTCA GCG TGACC TA TA GGTCA GCG TGACC TA AAGCTT GGG-3\u2032. We directionally cloned it into the pGL3-Basic vector after restriction enzyme digestion. We confirmed the construct by sequencing. The estrogen receptor alpha and beta vectors were kindly provided by professor Sylvie Mader (Facult\u00e9 de M\u00e9decine, Universit\u00e9 de Montr\u00e9al) and Leigh C. Murphy (University of Manitoba). MEG-01 and HEK293T, IL3 receptor expression positive cell lines were cultured in PRMI 1640 and DMEM respectively supplemented with 10% FBS, 2 mM L-glutamine, 1 mM Sodium pyruvate and 1% antibiotics. Co-transfection of 3ERE-pGL3 (2 ug), ER~\u03b1~ (1 ug) or ER~\u03b2~ (1 ug) were performed by using a Nucleofector Device from Amaxa Biosystems (Lonza Cologne AG, Germany) in MEG01 cell line, while co-transfection in HEK293T cell line were used Lipofectamine 2000 transfection reagent (Invitrogen) as previously described, and with pRL-TK as the internal control. After six hours incubation, 17-beta-estradiol and recombinant human IL3 protein (Invitrogen) were added into the medium with a final concentration of 10 nM. Cells were harvested and their luciferase activity was measured after additional 24 h incubation. All assays were performed in at least three independent experiments with a minimum of triplications.\n\nStatistical Analysis {#s4q}\n--------------------\n\nHardy-Weinberg equilibrium of each SNP was assessed by using GENEPOP (v 4.0) [@pone.0050375-Rousset1]. Association of single SNP with total brain volume (TBV) and their additive effects on this quantitative trait were tested by utilizing PLINK or SAS statistical software using the linear regression option, with age, sex and IQ (optional) as covariates; for the analyses on total gray/white matter, TBV was also considered as a covariate [@pone.0050375-Purcell3]. To account for sex-specific effects, we used a statistical model where the mean effect of SNP dose on the phenotype was allowed to differ for the two sexes. The p-value was adjusted by the conservative Bonferroni correction according to the number of independent SNPs and the divided internal samples separated by sex. We used the Haplo Stats [@pone.0050375-Schaid1] to infer the haplotype frequency and to perform the haplotype association test. We used the Haploview [@pone.0050375-Barrett1] to calculate pairwise LD indices r^2^ and D', to define LD blocks and to select the tag SNPs. Haplotypes were inferred with the PHASE program by the Bayesian statistical methods based on the genotype data [@pone.0050375-Stephens1]. Sequence alignment and assembly were conducted by DNASTAR software package. The analysis of quantitative PCR data was based on the \u0394C~t~ values.\n\nImmunohistochemistry {#s4r}\n--------------------\n\nThe C57BL/6J mice were used in this paper and all animal procedures described herein were approved by the University Committee of Animal Resources at the University of Rochester. For the purposes of staging embryos, noon of the day a vaginal plug was detected was taken to be embryonic day 0.5 (E0.5). Mice were deeply anesthetized, perfused transcardially with PBS, followed by 4% PFA in PBS, pH 7.3. Then the brains were dissected, postfixed in 4% PFA in PBS at 4\u00b0C overnight, washed three times in PBS, and cryoprotected in 30% sucrose in PBS before rapid freezing in OCT compound (TissueTek). For antigen retrieval, cryosections (20 \u00b5m) were heated in 10 mM citrate buffer (pH 6.0) at 95\u00b0C for 10 min. The sections were permeabilized and blocked in PBS plus 0.1% Tween-20, 5% horse serum, and incubated with primary antibody overnight at 4\u00b0C, washed in PBS three times and incubated with fluorescently labeled secondary antibody for 1 h at room temperature. The primary antibodies used were mouse monoclonal anti-IL3RA (Santa Cruz, 1\u2236100), rabbit polyclonal anti-IL3RA (Santa Cruz, 1\u2236100), rabbit polyclonal anti-IL3RB (Santa Cruz, 1\u2236100), goat polyclonal anti-IL3 (Santa Cruz, 1\u2236100), rabbit monoclonal anti-\u03b2-III tubulin (Tuj1) (Covance, 1\u22361000), mouse anti-NeuN (Chemicon, 1\u2236300), goat anti-SOX2 (Santa Cruz, 1\u2236500), rabbit anti-Nestin (Abcam, 1\u2236300), rabbit anti-TBR2 (Millipore, 1\u2236200), rabbit anti-pH3 (Santa Cruz, 1\u2236200), rabbit anti-PROX1 (Covance, 1\u22361000), anti-estrogen receptor \u03b2 (Santa Cruz, 1\u2236100), rabbit anti-Ki67 (Novocastra, 1\u22361000), rabbit anti-GFAP (1\u22362000). Secondary antibodies used were Alexa Fluor 488 (Invitrogen, 1\u2236500), 546 (Invitrogen, 1\u2236500) conjugated to donkey anti-mouse, rabbit or anti-goat (Invitrogen). DNA was stained with 4\u2032,6-diamidino-2-phenylindole (DAPI; Molecular Probes). Images were acquired with a Zeiss laser confocal microscope and analysed with LSM 510 software (Carl Zeiss).\n\nPrimary Cultures {#s4s}\n----------------\n\nFor neural progenitor's culture, cerebral cortices from C57BL/6 mice embryos (E12.5--14.5) were dissected in HBSS solution (Invitorgen), the meninges and other parts were removed under dissecting microscope and only the cortices were retained. After several washes with HBSS, the cortices were minced and dissociated mechanically with trituration, filtered through a 70 \u00b5m cell strainer (BD Falcon). Then, the cells were culture in neurobasal medium (Invitrogen) or DMEM/F12 (Millipore) with different supplements and growth factors according to experimental requirements. For neurons culture, the procedures are same with neural progenitors except the age of mice embryos (E14.5--17.5).\n\nProliferation Assays {#s4t}\n--------------------\n\nFor generation of neurospheres, the dissociated cells from E14.5--17.5 mice cortices were cultured for 2--5d in the medium containing neurobasal medium, B27 (Invotrogen) and N2 (Invitrogen) supplements, 100 U/mL penicillin, 100 \u00b5g/mL streptomycin, FGF2 (10 ng/ml), and EGF (10 ng/ml). For monolayer cultures, the generated neurospheres were collected and spun down (200 g for 5 minutes), then triturated with a Pasteur pipette to obtain single cells. The single cell suspensions were replated onto poly-L-lysine (50 \u00b5g/ml)/laminin (10 \u00b5g/ml)--coated Lab-Tek Chamber Slide (Thermo Scientific). Cells were treated with different concentrations of recombinant mouse IL3 (10 and 100 ng/ml) and cultured in above medium. Untreated cells served as control. After 24 or 48 hours culture, cells were fixed with 4% PFA and subjected to mmunohistochemistry. Immunostaining and morphometry were carried out to assess the numbers of proliferating progenitors (Ki67 and pH3 positive). More than 4 random microscopic fields (20\u00d7) were analyzed and about 8000 cells were counted for each condition.\n\nAssessment of Differentiation Markers *in vitro* {#s4u}\n------------------------------------------------\n\nTo investigate whether IL3 can drive neural differentiation, we used quantitative real-time PCR method described previously [@pone.0050375-Schneider1] to evaluate the effects of IL3 on neural differentiation. First, neural progenitors were cultured in neurobasal medium (containing B27 supplements, N2 supplements,10 ng/ml FGF2 and 10 ng/ml EGF) for 2 days, then FGF2 and EGF was removed and replaced by neurobasal medium with 1% serum (vol/vol), B27 supplements, N2 supplements and IL3 (10, 100, and 200 ng/ml). Four days after addition of recombinant mouse IL3 (Invitrogen), cells were harvested for the RNA extraction. Untreated cells served as control. RNA was isolated using Trizol reagent (Invitrogen) according to the manufacturer's instructions and treated with DNase I (Fermentas). cDNA was synthesized from 3 \u00b5g total RNA using oligo-dT primers and Superscript III Reverse Transcriptase (Invitrogen). Quantitative PCR was performed by using the Bio-Rad iCycler & iQ Real-Time PCR Systems. The following primer pairs were used: mouse Sox2-F, GCGGAGTGGAAACTTTTGTCC, mouse Sox2-R, CGGGAAGCGTGTACTTATCCTT; mouse Nestin-F, CCCCTTGCCTAATACCCTTGA, mouse Nestin-R, GCCTCAGACATAGGTGGGATG; mouse \u03b2-III-tub-F, TAGACCCCAGCGGCAACTAT, mouse \u03b2-III-tub-R, GTTCCAGGTTCCAAGTCCACC; mouse GFAP-F: CCCTGGCTCGTGTGGATTT, mouse GFAP-R: GACCGATACCACTCCTCTGTC; mouse ENO2-F: GTCCCTGGCCGTGTGTAAG, mouse ENO2-R: CATCCCGAAAGCTCTCAGC; mouse nestin-F: CCCTGAAGTCGAGGAGCTG, mouse nestin-R: CTGCTGCACCTCTAAGCGA; mouse PLP1-F: TGAGCGCAACGGTAACAGG, mouse PLP1-R: CCCACAAACTTGTCGGGATG. The iCycler PCR analysis was performed using the SYBR Green master mix, according to the manufacturer's recommendations (BioRad). The specificity of product was ensured by melting curve analysis and agrose gel electrophoresis. cDNA content of samples was normalized to the expression of GAPDH.\n\nCell Viability Assays {#s4v}\n---------------------\n\nTo investigate whether IL3 has protective or trophic effects on neural progenitor's survival, cerebral cortices from E12.5 or E13.5 mice embryos were dissociated and the isolated cells (5\u00d710^4^) were plated onto poly-L-lysine coated 96-well plate (Corning). The survival of neural progenitors was investigated by three culture conditions. Firstly, we studied the neurotrophic effects of IL3 on neural progenitors through culturing the progenitors in the absence of any factor (with neurobasal medium only) and in the presence of IL3 (0.02--50 ng/ml). Recombinant mouse IL3 was added into the medium after the culture initiated two hours. After 36 hours incubation, cell viability was determined by measurement of cellular ATP levels (CellTiter-Glo Luminescent Cell Viability Assay, Promega). Secondly, neural progenitors were first cultured in neurobasal medium supplemented with B27 and Glutamax (Invitrogen). On day 2 of culture, the medium was replaced with serum-free neurobasal medium containing N2 supplement and different concentrations of mouse IL3 (0.1--20 ng/ml). The cultures were maintained for 3 days and cell viability was measured. Thirdly, neural progenitors were first cultured in neurobasal medium supplemented with B27 and Glutamax. On day 2 of culture, IL3 was added into the medium and the cultures were maintained for 3 days and cell viability was measured. For studying the effects of IL3 on neurons, cerebral cortices from E16.5--E18.5 mice embryos were dissociated and cultured in DMEM/F12 medium with 5% FBS. On day 2 of culture, the medium was replaced with serum-free DMEM/F12 containing N2 supplement and different concentrations of IL3 (0.1--20 ng/ml). The cultures were maintained for 2 or 3 days and cell viability was determined as described above.\n\nWestern Blotting {#s4w}\n----------------\n\nNeural progenitors from E13.5 mice were first cultured in neurobasal medium under proliferating condition (containing B27 supplement, Glutamax, 10 ng/ml FGF2 and EGF) for one week. To exclude the interference of other factors, the supplements, FGF2 and EGF were removed from the medium for about 16 hours prior to IL3 treatment (3 ng/ml). Proteins from MEG01 cells (IL3 treated) and neural progenitors were homogenized in RIPA lyses buffer (Cell signaling) containing a cocktail of protease inhibitor (Sigma Chemical, MO, USA) and phosphatase inhibitor (Cell Signaling). Proteins were quantified by BCA method (Pierce). Extracted protein (40 \u00b5g) was separated by SDS-polyacrylamide gel electrophoresis and transferred to PVDF or Nitrocellulose membrane by electrophoretic transfer. The membrane was blocked, incubated with primary antibodies for overnight at 4\u00b0C, washed three times with TBST, and then incubated with secondary antibody for 1 hour at room temperature. Antibodies used in western blot are as follows: Rabbit anti-phospho-AKT (Thr308) (Cell Signaling), Rabbit anti-AKT1 (Cell Signaling), Rabbit anti-phospho-JAK2 (Tyr 1007/1008) (Cell Signaling), Rabbit anti-JAK2 (Cell Signaling), Rabbit anti-phospho-GSK3\u03b2 (Ser9) (Cell Signaling), Rabbit anti-phospho-ERK1/2 (Cell Signaling), Rabbit anti-ERK1/2 (Cell Signaling), Rabbit anti-GSK3\u03b2 (BD), and Rabbit anti-actin (Abcam). Immunoreactivity was detected with an enhanced chemiluminescence system (Pierce, IL, USA) with colored markers (Fermentas) as molecular size standards.\n\nSupporting Information {#s5}\n======================\n\n###### \n\n**Linkage disequilibrium (LD) pattern of the studies SNPs in Chinese (CHB) and Europeans (CEU).** (a) In screening sample (CHB), they are four haplotype blocks and all the 7 highly significant association SNPs are located in block 2. (b) In CEU, they are five haplotype blocks and the 7 highly linked SNPs in CHB are disrupted. LD values (r^2^) for each pair of markers were calculated by Haploview (v4.2). Haplotype blocks were defined according to the criteria of Gabriel et al.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**The C allele of rs31480 is highly conserved in vertebrate.** rs31480 (box in red) is locates \u221216 bp upstream of the IL3 promoter, and only 10 bp downstream of the highly conserved TATA binding site (box in blue). The C allele (ancestral allele) is completely conserved in all of the listed species, implying functional importance of rs31480. Note that rs31480 is lies in a primates conserved region (up panel), also suggesting the importance of rs31480 in primates. TSS, transcription start site.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Impacts of rs31480 on promoter activity in CHO, SK-N-SH and COS-7 cell lines.** Promoter activity of the construct with T allele is significantly higher than C allele in CHO (a) and COS-7 (b) cells. In SK-N-SH cells (c), the trend is same as CHO and COS-7 though the differences were not reached significant level. Values of relative luciferase activity are expressed as mean \u00b1 s.d. (results of a triplicate assay). ^\\#^P\u200a=\u200a0.07, \\*\\*P\\<0.005 (Student's *t-test*).\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Expression of IL3RA in embryonic mouse brain (E12.5).** IL3RA is expressed in SOX2 positive cells in Frontal cortex and hippocampus, two regions that associated with higher level cognitive functions. Scale bar, 50 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3RA is mainly expressed in the neocortex region of mouse brain.** (a--f) Co-expression of IL3RA and IL3RB in agranular retrosplenial cortex (RSA), barrel field of the primary somatosensory cortex (S1BF). (g--l) Expression of IL3RA was also found in perirhinal cortex (Prh), primary Motor Cortex (M1) and secondary Motor Cortex (M2). Note that IL3RA positive cells were also expressed Tbr2 weakly, indicating they were not mature neurons. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Expression of IL3RA and IL3RB in hippocampus.** (a--c) Co-expression of IL3RA and IL3RB in CA3 region of hippocampus. (d--i) IL3RA is expressed in CA1 and CA3 of hippocampus, some of IL3RA positive cells were Tuj1 positive, indicating they were mature neurons, whereas others were Tuj1 negative. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3RA is expressed in hilus of dentate gyrus.** Prox1 was used to label the granule cell layer (GCL) of the dentate gyrus, note IL3RA positive cells in hilus were prox1 negative. Scale bar, 50 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Double immunostaining of IL3RA and SOX2 revealed co-localization of IL3RA and SOX2 in developing mouse brain.** (a--c) IL3RA expression cells were SOX2 positive in lateral septal nucleus, dorsal part (LSD), indicating they were neural progenitors. (d--i) In cortex, some IL3RA positive cells still express SOX2, but for many of IL3RA positive cells, expression of SOX2 was down-regulated or turned-off, demonstrating they were converted into immediate progenitors or neurons. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Co-expression of IL3RA and IL3RB in mouse brain.** All IL3RA positive cells were also expression IL3RB. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3RA is mainly expressed in neural progenitors at early embryonic stage.** (a--c) At E12.5, IL3RA is expressed in sox2 positive progenitors in frontal cortex. (b--e) Co-expression of IL3RA and nestin, a marker for neural progenitors. (f--h) At E14.5, IL3RA is expressed in some sox2 positive progenitors in cingulate cortex. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Double immunostaining analysis of IL3RA and Tuj1 in the developing mouse brain.** At early stage of brain development (From E14.5-P1), IL3RA is not expressed in mature neurons (a--i). However, at P2 stage, a proportion of IL3RA positive cells are mature neurons as revealed by co-localization with tuj1 (j--l), a marker for mature neurons. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3RA is expressed in immediate progenitor cells (IPCs).** IL3RA positive cells also express Tbr2 weakly, indicating these cells were immediate progenitors. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3RB is expressed in neurons and glia cells.** (a--c) Co-immunofluorescence of IL3RB and NeuN revealed expression of IL3RB in mature neurons. (d--f) IL3RB also expressed in some glia cells (GFAP positive). Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Expression of IL3RA in proliferating neural progenitors.** Some IL3RA positive cells also expressed Ki67 and pH3, marker for proliferation cells, indicating they were active proliferation. Scale bar, 25 \u00b5m.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Expression of IL3 receptors (IL3RA and IL3RB) in cultured neural progenitors and trophic effects of IL-3 on neural progenitors and neurons.** (a) IL3 receptors (IL3RA and IL3RB) were expressed in cultured neural progenitors (from E13.5 mice) as revealed by RT-PCR, however, \u03b2IL3 was not detected. (b) When there were 5% FBS in neurobasal medium, IL-3 has no trophic effects on neural progenitors. Neural progenitors from E12.5 mice were cultured in neurobasal medium (supplemented with 5% FBS and IL-3) for 36 hours then cell viability was measured. (c) IL-3 (1 ng/ml) significantly promotes survival of neurons when there were no any factors in neurobasal medium. Neurons from E17.5 mice were cultured in neurobasal medium (supplemented with B27 and Glutamax) for 12 days, then B27 and Glutamax were removed from medium and different concentrations of IL-3 were added. Cell viability was determined after IL-3 treatment for 2 days. (d) However, when there were B27 supplement in neurobasal medium, trophic effects of IL-3 on neurons was disappeared. Neurons from E18 mice were first cultured in neurobasal medium (supplemented with B27 and different concentrations of IL-3) for 24 hours, cell viability was then measured. **Y**-axis, cell viability (normalized to control); **X**-axis, concentration of IL-3 (ng/ml). Data are expressed as mean \u00b1 s.e.m. (n\u200a=\u200a8 for each group). \\*P\\<0.05 (Student's *t-test*). (e) The expression of BCL-xL was not regulated by IL3 in neural progenitors.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL3 has no effects on neural differentiation.** Neural progenitors were first cultured in neurobasal medium under proliferation condition (containing 10 ng/ml FGF2 and EGF), after 4 day's culture, FGF2 and EGF were removed. Then 2% FBS and different concentrations of IL3 (10, 100, 200 ng/ml) were added. The cultures were maintained for 4 days and RNA was isolated for quantification. Relative gene expression was not changed for all of the tested genes, indicating IL3 has no effect on neural differentiation. Real-time PCR analysis of beta-III Tubulin (Tubb3) (a) and Enolase 2 (ENO2) (b), two neuron specific markers, GFAP (glia cells marker) (c), sox2 (neural stem cells marker) (d), Tbr2 (immediate progenitors marker) (e), and DCX (new born neurons marker) (f). None of these cell specific markers showed significant change after different concentration IL-3 treatment.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**IL-3 is not regulated by estrogen.** We first confirmed that K562 cell line expressed IL3 and estrogen receptors (ESR1 and ESR2) by RT-PCR (a). (b--c) Expression of TFF1 and IL3 was not changed after treated by vehicle (DMSO) for different times (0 h--24 h). TFF1, an estrogen response gene, showed significantly elevated after treated by estrogen (10 nM) (d), however, expression of IL-3 was not changed after estrogen treatment (e), indicating IL-3 is not regulated by estrogen. Data are expressed as mean \u00b1 s.e.m. (three independent assays, each containing 3 replicates). ^\\*\\*^P\\<0.01.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Model for sex-specific association of IL-3 and brain volume.** Different genotypes at rs31480 (TT vs. CC) influence IL-3\u2032s expression in both males and females. However, since the estrogen receptors (ER) level is low in males, therefore, even the signaling pathways mediated by IL-3 were different in TT and CC carriers, the total activation level of ERs was not significant changed. But in females, the activation level of ERs was different between TT and CC carriers due to high level of ERs. In addition, estrogen can further enhance ER activity in females. As a result, signaling pathways mediated by ER were greatly activated in TT carriers than in CC carriers, which may influence brain development, eventually lead to difference of brain volume in TT and CC carriers at rs31480.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Marker characteristics and association significance in females.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Core haplotype association analysis in females.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Marker characteristics and association significance in males.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Average cranial volumes of female individuals with three different genotypes at each of the seven SNPs covering IL-3.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\n###### \n\n**Marker characteristics of AKT1 and association significance.**\n\n(DOC)\n\n###### \n\nClick here for additional data file.\n\nWe are grateful to all the voluntary donors of DNA samples in this study. We thank Wei Zhang, Hui Zhang, Xiaoling Xie and Yan-jiao Li for their technical help.\n\n[^1]: **Competing Interests:**The authors have declared that no competing interests exist.\n\n[^2]: Conceived and designed the experiments: XJL ML LG BS. Performed the experiments: XJL ML LH MD. Analyzed the data: XJL ML KN QC DRW AAV MR VSM AJS LS GF BF JCC XNC. Contributed reagents/materials/analysis tools: XX JKW XBQ KX YMP XYC YL XDS. Wrote the paper: XJL ML LG BS.\n"}
+{"text": "Introduction {#s1}\n============\n\nThe number of times an article is cited in scientific journals reflects its impact on a specific biomedical field or specialty and reflects the impact of the authors' creativity.[@b01]--[@b03] The Institute for Scientific Information (ISI) has been collecting citation and other academic impact information since 1945, and this information has been electronically available since 1979. ISI (now a subsidiary of Thomson Corp.) has named their newest journal citation system \"Science Citation Index Expanded,\" and it is one of the databases available under the banner of the Web of Science.[@b04] Citation data from peer-reviewed articles are indexed for more than 10,000 high-impact journals in the sciences and social sciences, as well as in the arts and humanities.\n\nIn recent years, several citation analyses of the top-cited articles have become available for various areas of clinical neurology and neurosurgery, including epilepsy,[@b05] multiple sclerosis,[@b06] Parkinson's disease,[@b07] brain metastases research,[@b08] and neurosurgery.[@b09]--[@b11] However, no similar study has been performed in the field of essential tremor (ET). ET is one of the most common involuntary movement disorders, and it is the most common tremor disorder.[@b12]--[@b14] The past few years have witnessed an exponential growth in the number of scientific publications on ET, mainly across the disciplines of epidemiology, clinical features, and pathology.[@b15]--[@b18]\n\nWe used the electronic version of the Science Citation Index Expanded to determine which published articles in ET have been cited most often by ranking the 100 top-cited works. The expectation was to provide references for future studies and clinical practices.\n\nMethods {#s2}\n=======\n\nA search was performed using the bibliometric database ISI Web of Science for articles including the keyword \"Essential tremor\\*\" in the title (the asterisk was included in the search string as a wild card character). In early April 2013, we identified the 100 top-cited articles ([Table\u20051](#t01){ref-type=\"table\"}) published in professional journals since 1960 by the Web of Science. All articles in each given journal were ranked by the number of citations listed on Web of Science. The full texts were mainly selected by PubMed and ScienceDirect.\n\nTABLE 1The 100 Top-Cited ET ArticlesRankArticleNumber of Citations1Benamer HTS, Patterson J, Grosset DG, et al. Accurate differentiation of parkinsonism and essential tremor using visual assessment of \\[123I\\]-FP-CIT SPECT imaging: the \\[123I\\]-FP-CIT study group. *Mov Disord* 2000;15:503--510.2932Limousin P, Speelman JD, Gielen F, Janssens M. Multicentre European study of thalamic stimulation in parkinsonian and essential tremor. *J Neurol Neurosurg Psychiatry* 1999;66:289--296.2643Louis ED, Ottman R, Allen Hauser W. How common is the most common adult movement disorder? Estimates of the prevalence of essential tremor throughout the world. *Mov Disord* 1998;13:5--10.2034Bain PG, Findley LJ, Thompson PD, et al. A study of hereditary essential tremor. *Brain* 1994;117:805--824.2025Lou JS, Jankovic J. Essential tremor: clinical correlates in 350 patients. Neurology 1991;41:234--238.1646Gulcher JR, J\u00f3nsson P, Kong A, et al. Mapping of a familial essential tremor gene, FET1, to chromosome 3q13. *Nat Gen* 1997;17:84--87.1547Jenkins IH, Bain PG, Colebatch JG, et al. A positron emission tomography study of essential tremor: evidence for overactivity of cerebellar connections. *Ann Neurol* 1993;34:82--90.1538Benabid AL, Pollak P, Seigneuret E, Hoffmann D, Gay E, Perret J. Chronic VIM thalamic stimulation in Parkinson's disease, essential tremor and extra-pyramidal dyskinesias. *Acta Neurochir Suppl (Wien)* 1993;58:39--44.1369Ondo W, Jankovic J, Schwartz K, Almaguer M, Simpson RK. Unilateral thalamic deep brain stimulation for refractory essential tremor and Parkinson's disease tremor. *Neurology* 1998;51:1063--1069.13310Louis ED. Essential tremor. *Lancet Neurol* 2005;4:100--110.13211Louis ED, Faust PL, Vonsattel JP, et al. Neuropathological changes in essential tremor: 33 Cases compared with 21 controls. *Brain* 2007;130:3297--3307.13112Rajput AH, Offord, KP, Beard CM, Kurland LT. Essential tremor in Rochester, Minnesota: a 45-year study. *J Neurol Neurosurg Psychiatry* 1984;47:466--470.13013Koller WC, Busenbark K, Miner K, et al. The relationship of essential tremor to other movement disorders: report on 678 patients. *Ann Neurol* 1994;35:717--723.12814Deuschl G, Wenzelburger R, L\u00f6ffler K, Raethjen J, Stolze H. Essential tremor and cerebellar dysfunction. Clinical and kinematic analysis of intention tremor. *Brain* 2000;123:1568--1580.12715Higgins JJ, Pho LT, Nee LE. A gene (ETM) for essential tremor maps to chromosome 2p22--p25. *Mov Disord* 1997;12:859--864.12616Benito-Le\u00f3n J, Bermejo-Pareja F, Morales J-M, Vega S, Molina J-A. Prevalence of essential tremor in three elderly populations of central Spain. *Mov Disord* 2003;18:389--394.11817Dogu, O, Sevim S, Camdeviren H, et al. Prevalence of essential tremor: door-to-door neurologic exams in Mersin Province, Turkey. *Neurology* 2003;61:1804--180611218Stolze, H, Petersen, G, Raethjen J, Wenzelburger R, Deuschl G. The gait disorder of advanced essential tremor. *Brain* 2002;124:2278--2286.11019Gironell A, Kulisevsky J, Barbanoj M, L\u00f3pez-Villegas D, Hern\u00e1ndez G, Pascual-Sedano B. A randomized placebo-controlled comparative trial of gabapentin and propranolol in essential tremor. *Arch Neurol* 1999;56:475--480.11020Elble, RJ. Physiologic and essential tremor. *Neurology* 1986;36:225--231.10921Zesiewicz TA, Elble RJ, Louis ED, et al. Practice parameter: therapies for essential tremor: report of the quality standards subcommittee of the American Academy of Neurology. *Neurology* 2005;64:2008--2020.10522Asenbaum S, Pirker W, Angelberger P, Bencsits G, Pruckmayer M, Br\u00fccke T. \\[123I\\]\u03b2-CIT and SPECT in essential tremor and Parkinson's disease. *J Neural Transm* 1998;105:1213--1228.10023Bucher SF, Seelos KC, Dodel RC, Reiser M, Oertel WH. Activation mapping in essential tremor with functional magnetic resonance imaging. *Ann Neurol* 1997;41:32--40.10024Wills AJ, Jenkins IH, Thompson PD, Findley LJ, Brooks DJ. Red nuclear and cerebellar but no olivary activation associated with essential tremor: a positron emission tomographic study. *Ann Neurol* 1994;36:636--642.10025Alesch F, Pinter MM, Helscher RJ, Fertl L, Benabid AL, Koos WT. Stimulation of the ventral intermediate thalamic nucleus in tremor dominated Parkinson's disease and essential tremor. *Acta Neurochir (Wien)* 1995;136:75--81.9926Deuschl G, Elble RJ. The pathophysiology of essential tremor. *Neurology* 2000;54(11 SUPPL. 4):S14--S20.9827Cohen O, Pullman S, Jurewicz E, Watner D, Louis ED. Rest tremor in patients with essential tremor: prevalence, clinical correlates, and electrophysiologic characteristics. *Arch Neurol* 2003;60:405--410.9728Bain P, Brin M, Deuschl G, et al. Criteria for the diagnosis of essential tremor. *Neurology* 2000;54(11 SUPPL. 4):S79729Laesson T, Sjogren T. Essential tremor: a clinical and genetic population study. *Acta Psychiatr Scand Suppl* 1960;36:1--176.9630Louis ED, Marder K, Cote L, et al. Differences in the prevalence of essential tremor among elderly African Americans, whites, and Hispanics in Northern Manhattan, NY. *Arch Neurol* 1995;52:1201--1205.9531Koller WC, Lyons KE, Wilkinson SB, Troster AI, Pahwa R. Long-term safety and efficacy of unilateral deep brain stimulation of the thalamus in essential tremor. *Mov Disord* 2001;16:464--468.9432Elble RJ. Diagnostic criteria for essential tremor and differential diagnosis. Neurology 2000;54(11 SUPPL. 4):S2--S6.9433Rautakorpi I, Takala J, Marttila RJ, Sievers K, Rinne UK. Essential tremor in a Finnish population. *Acta Neurologica Scand* 1982;66:58--67.9134Louis ED, Ottman R, Ford B, et al. The Washington Heights-Inwood Genetic Study of Essential tremor: methodologic issues in Essential-tremor research. *Neuroepidemiology* 1997;16:124--133.8835Colebatch JG, Findley LJ, Frackowiak RSJ, Marsden CD, Brooks DJ. Preliminary report: activation of the cerebellum in essential tremor. *Lancet* 1990;336:1028--1030.8836Hariz MI, Shamsgovara P, Johansson F, Hariz G-M, Fodstad H. Tolerance and tremor rebound following long-term chronic thalamic stimulation for Parkinsonian and essential tremor. *Stereotact Funct Neurosurg* 1999;72:208--218.8637Singer C, Sanchez-Ramos J, Weiner WJ. Gait abnormality in essential tremor. *Mov Disord* 1994;9:193--196.8638Benito-Le\u00f3n J, Louis ED. Essential tremor: emerging views of a common disorder. *Nat Clin Pract Neurol* 2006;2:666--678.8539Louis ED, Zheng W, Jurewicz EC, et al. Elevation of blood \u0392-carboline alkaloids in essential tremor. *Neurology* 2002;59:1940--1944.8540Louis ED, Ottman R. How familial is familial tremor? The genetic epidemiology of essential tremor. *Neurology* 1996;46:1200--12058541Lombardi WJ, Woolston DJ, Roberts JW, Gross RE. Cognitive deficits in patients with essential tremor. *Neurology* 2001;57:785--790.8442Louis ED, Shungu DC, Chan S, Mao X, Jurewicz EC, Watner D. Metabolic abnormality in the cerebellum in patients with essential tremor: a proton magnetic resonance spectroscopic imaging study. *Neurosci Lett* 2002;333:17--20.8343Louis ED, Barnes L, Albert SM, et al. Correlates of functional disability in essential tremor. *Mov Disord* 2001;16:914--920.8344Leehey MA, Munhoz RP, Lang AE, et al. The fragile X premutation presenting as essential tremor. *Arch Neurol* 2003;60:117--121.8245Boecker H, Wills AJ, Ceballos-Baumann A, et al. The effect of ethanol on alcohol-responsive essential tremor: a position emission tomography study. *Ann Neurol* 1996;39:650--658.8146Benito-Le\u00f3n J, Louis ED, Bermejo-Pareja F. Population-based case-control study of cognitive function in essential tremor *Neurology* 2006;66:69--74.8047Pahwa R, Lyons K, Hubble JP, et al. Double-blind controlled trial of gabapentin in essential tremor. *Mov Disord* 1998;13:465--467.8048Raethjen J, Lindemann M, Schmaljohann H, Wenzelburger R, Pfister G, Deuschl G. Multiple oscillators are causing parkinsonian and essential tremor. *Mov Disord* 2000;15:84--94.7949Busenbark KL, Nash J, Nash S, Hubble JP, Koller WC. Is essential tremor benign? *Neurology* 1991;41:1982--1983.7950Jankovic J. Essential tremor: a heterogeneous disorder. *Mov Disord* 2002;17:638--644.7851Rajput, A, Robinson CA, Rajput AH. Essential tremor course and disability: a clinicopathologic study of 20 cases. *Neurology* 2004;62:932--936.7752Louis ED, Ford B, Frucht S, Barnes LF, X-Tang M, Ottman R. Risk of tremor and impairment from tremor in relatives of patients with essential tremor: a community-based family study. *Ann Neurol* 2001;49:761--769.7753Hellwig B, H\u00e4ussler S, Schelter B, et al. Tremor-correlated cortical activity in essential tremor. *Lancet* 2001;357:519--523.7754Hallett M, Dubinsky RM. Glucose metabolism in the brain of patients with essential tremor. *J Neurol Sci* 1993;114:45--48.7655Benito-Le\u00f3n J, Bermejo-Pareja, F, Louis ED. Incidence of essential tremor in three elderly populations of Central Spain. *Neurology* 2005;64:1721--1725.7556Connor GS. A double-blind placebo-controlled trial of topiramate treatment for essential tremor. *Neurology* 2002;59:132--134.7557Hubble JP, Busenbark KL, Wilkinson S, Penn RD, Lyons K, Koller WC. Deep brain stimulation for essential tremor. *Neurology* 1996;46:1150--1153.7558Sydow O, Thobois S, Alesch F, Speelman JD. Multicentre European study of thalamic stimulation in essential tremor: a six year follow up. *J Neurol Neurosurg Psychiatry* 2003;74:1387--1391.7459Helmchen C, Hagenow A, Miesner J, et al. Eye movement abnormalities in essential tremor may indicate cerebellar dysfunction. *Brain* 2003;126:1319--1332.7460Ondo W, Hunter C, Vuong KD, Schwartz K, Jankovic J. Gabapentin for essential tremor: a multiple-dose, double-blind, placebo-controlled trial. *Mov Disord* 2000;15:678--682.7361Jankovic, J. Essential tremor: clinical characteristics. *Neurology* 2000;54(11 SUPPL. 4):S21--S25.7362Dupuis MJM, Delwaide PJ, Boucquey D, Gonsette RE. Homolateral disappearance of essential tremor after cerebellar stroke. *Mov Disord* 1989;4:183--187.7363Louis ED, Ford B, Lee H, Andrews H. Does a screening questionnaire for essential tremor agree with the physician's examination? *Neurology* 1998;50:1351--1357.7264Salemi G, Savettieri G, Rocca WA, et al. Prevalence of essential tremor: a door-to-door survey in Terrasini, Sicily. *Neurology* 1994;44:61--64.7265Shahed J, Jankovic J. Exploring the relationship between essential tremor and Parkinson's disease. *Parkinsonism Relat Disord* 2007;13:67--76.7166Louis ED, Vonsattel JPG, Honig LS, Ross GW, Lyons KE, Pahwa R. Neuropathologic findings in essential tremor. *Neurology* 206;66:1756--1759.7167Deng H, Le W, Jankovic J. Genetics of essential tremor. *Brain* 2007;130:1456--1464.7068Higgins JJ, Loveless JM, Jankovic J, Patel PI. Evidence that a gene for essential tremor maps to chromosome 2p in four families. *Mov Disord* 1998;13:972--977.7069Benito-Le\u00f3n J, Louis ED, Bermejo-Pareja F. Elderly-onset essential tremor is associated with dementia. *Neurology* 2006;66:1500--1505.6970Jeanneteau F, Funalot B, Jankovic J, et al. A functional variant of the dopamine D3 receptor is associated with risk and age-at-onset of essential tremor. *Proc Natl Acad Sci U S A* 2006;103:10753--10758.6871Hubble JP, Busenbark KL, Pahwa R, Lyons K, Koller WC. Clinical expression of essential tremor: effects of gender and age. *Mov Disord* 1997;12:969--972.6872Rajput AH, Rozdilsky B, Ang L, Rajput A. Clinicopathologic observations in essential tremor: report of six cases. *Neurology* 1991;41:1422--1424.6873Louis ED. Essential tremor. *N Engl J Med* 2001;345:887--891.6774Koller WC, Vetere-Overfield B. Acute and chronic effects of propranolol and primidone in essential tremor. *Neurology* 1989;39:1587--1588.6775Geraghty JJ, Jankovic J, Zetusky WJ. Association between essential tremor and Parkinson's disease. *Ann Neurol* 1985;17:329--333.6776Hornabrook RW, Nagurney JT. Essential tremor in Papua New Guinea. *Brain* 1976;99:659--672.6777Pahwa R, Lyons KL, Wilkinson SB, et al. Bilateral thalamic stimulation for the treatment of essential tremor. *Neurology* 1999;53:1447--1450.6678Louis ED, Ford B, Lee H, Andrews H, Cameron G. Diagnostic criteria for essential tremor: a population perspective. *Arch Neurol* 1998;55:823--828.6679Kralic JE, Criswell HE, Osterman JL, et al. Genetic essential tremor in \u03b3-aminobutyric acidA receptor \u03b11 subunit knockout mice. *J Clin Invest* 2005;115:774--779.6580Tanner CM, Goldman SM, Lyons KE, et al. Essential tremor in twins: an assessment of genetic vs environmental determinants of etiology. *Neurology* 2001;57:1389--1391.6581Findley LJ, Koller WC. Essential tremor: a review. *Neurology* 1987;37:1194--1197.6582Bermejo-Pareja F, Louis ED, Benito-Le\u00f3n J. Risk of incident dementia in essential tremor: a population-based study. *Mov Disord* 2007;22:1573--1580.6483Tr\u00f6ster AI, Woods SP, Fields JA, et al. Neuropsychological deficits in essential tremor: an expression of cerebello-thalamo-cortical pathophysiology? *Eur J Neurol* 2002;9:143--151.6484Pahwa R, Lyons KE, Wilkinson SB, et al. Comparison of thalamotomy to deep brain stimulation of the thalamus in essential tremor. *Mov Disord* 2001;16:140--143.6485Hua SE, Lenz FA, Zirh TA, Reich SG, Dougherty PM. Thalamic neuronal activity correlated with essential tremor. *J Neurol Neurosurg Psychiatry* 1998;64:273--276.6486Louis ED, Honig LS, Vonsattel JPG, Maraganore DM, Borden S, Moskowitz CB. Essential tremor associated with focal nonnigral Lewy bodies: a clinicopathologic study. *Arch Neurol* 2005;62:1004--1007.6387Koller WC, Royse VL. Efficacy of primidone in essential tremor. *Neurology* 1986;36:121--124.6288Louis ED, Ferreira JJ. How common is the most common adult movement disorder? Update on the worldwide prevalence of essential tremor. *Mov Disord* 2010;25:534--541.6189Louis ED, Vonsattel JPG. The emerging neuropathology of essential tremor. *Mov Disord* 2008;23:174--182.6190Findley LJ, Cleeves L, Calzetti S. Primidone in essential tremor of the hands and head: a double blind controlled clinical study. *J Neurol Neurosurg Psychiatry* 1985;48:911--915.6191Axelrad JE, Louis ED, Honig LS, et al. Reduced Purkinje cell number in essential tremor: a postmortem study. *Arch Neurol 2008*;65:101--107.5992Ondo WG, Jankovic J, Connor GS, et al. Topiramate in essential tremor: a double-blind, placebo-controlled trial*Neurology* 2006;66:672--677.5993Gasparini M, Bonifati V, Fabrizio E, et al. Frontal lobe dysfunction in essential tremor: a preliminary study. *J Neurol* 2001;248:399--402.5994Benito-Le\u00f3n J, Louis ED, Bermejo-Pareja F. Risk of incident Parkinson's disease and parkinsonism in essential tremor: a population based study. *J Neurol Neurosurg Psychiatry* 2009;80:423--425.5895Findley LJ. Epidemiology and genetics of essential tremor. *Neurology* 2000;54:S8--S13.5896Halliday DM, Conway BA, Farmer SF, Shahani U, Russell AJC, Rosenberg JR. Coherence between low-frequency activation of the motor cortex and tremor in patients with essential tremor. *Lancet* 2000;355:1149--1153.5897Shill HA, Adler CH, Sabbagh MN, et al. Pathologic findings in prospectively ascertained essential tremor subjects. *Neurology* 2008;70:1452--1455.5798Rajput AH, Rozdilsky B, Ang L, Rajput A. Significance of Parkinsonian manifestations in essential tremor. *Can J Neurol Sci* 1993;20:114--117.5799Goldman MS, Ahlskog JE, Kelly PJ. The symptomatic and functional outcome of stereotactic thalamotomy for medically intractable essential tremor. *J Neurosurg* 1992;76:924--928.57100Stefansson H, Steinberg S, Petursson H, et al. Variant in the sequence of the LINGO1 gene confers risk of essential tremor. *Nat Genet* 2009;41:277--279.56\n\nBasic information concerning country of origin (based on the first author), institution, year of publication, publication name, and citations of the target articles were collected from Web of Science using the analyze tool. Additional analyses were then performed to determine authorship, article type, study design, and level of evidence. For each study, the level of evidence was graded according to Hadorn et al.[@b19]\n\nCited half-life is defined as the number of publication years from the current year, which account for 50% of current citations received.[@b04] This index helps to evaluate the age of the majority of cited articles published in a journal.[@b04] The h-index aims to measure the cumulative impact of a researcher's output; the value of h is equal to the number of papers (n) that have n or more citations.[@b20]\n\nResults {#s3}\n=======\n\nJournals and publication dates of the top-cited articles on ET {#s3a}\n--------------------------------------------------------------\n\nThe 100 top-cited articles ([Table\u20051](#t01){ref-type=\"table\"}) were published in 26 journals. Journal title, impact factor, number of articles, and cited half-life are listed in [Table\u20052](#t02){ref-type=\"table\"}. Almost half of the articles were retrieved from two journals: *Neurology* (n\u200a=\u200a31) and *Movement Disorders* (n\u200a=\u200a18). The 100 top-cited articles on ET were published from 1960 to 2010 ([Figure\u20051](#f01){ref-type=\"fig\"}). The greatest number of top-cited articles (n\u200a=\u200a10) were published in 2000, and eight of the top 10-cited articles were published in the 1990s ([Table\u20051](#t01){ref-type=\"table\"}) ([Figure\u20051](#f01){ref-type=\"fig\"}).[@b18],[@b21]--[@b29] Ninety-one of the articles were published within the past 25 years ([Figure\u20051](#f01){ref-type=\"fig\"}).\n\n![Publication Years for the 100 Top-Cited ET Articles.](tre-03-186-4307-1-g001){#f01}\n\nTABLE 2Journals that Published the 100 Top-Cited ET ArticlesImpact FactorRankJournal20115-Year Impact FactorNumber of ArticlesCited Half-Life1*Neurology*8.3127.634318.72*Movement Disorders*4.5054.449185.33*Annals of Neurology*11.08910.64479.04*Brain*9.45710.54577.85*Archives of Neurology*7.5846.92878.96*Journal of Neurology Neurosurgery and Psychiatry*4.7644.9536\\>10.07*Lancet*38.27833.79738.98*Nature Genetics*35.53233.09626.89*Acta Neurochirurgica*1.5201.83929.810*New England Journal of Medicine*53.29850.07517.811*Lancet Neurology*23.46220.59813.812*Journal of Clinical Investigation*13.06915.43019.413*Proceedings of the National Academy of Sciences of the United States of America*9.68110.47217.814*Nature Clinical Practice Neurology*7.6366.71214.415*Acta Psychiatrica Scandinavica*4.2204.2991\\>10.016*Parkinsonism & Related Disorders*3.7953.17313.917*European Journal of Neurology*3.6923.23914.518*Journal of Neurology*3.4733.20316.519*Journal of Neurosurgery*2.9653.0881\\>10.020*Journal of Neural Transmission*2.7302.66216.121*Acta Neurologica Scandinavica*2.4692.34719.822*Journal Neurological Sciences*2.3532.44118.723*Neuroepidemiology*2.3052.81716.324*Neuroscience Letters*2.1052.16818.325*Stereotactic and Functional Neurosurgery*1.8491.91319.426*Canadian Journal of Neurological Sciences*0.9681.1771112\n\nSixty-two authors contributed two or more articles to the list, and five contributed 10 or more ([Table\u20053](#t03){ref-type=\"table\"}). The number of authors per article ranged from one to 30, with the most common figures being five (16 articles), three (15 articles), two (12 articles), and seven (12 articles). The most frequent first authors of the top-cited articles in ET were Elan D. Louis (n\u200a=\u200a27) and William C. Koller (n\u200a=\u200a15); when combined, they accounted for 42% of the articles on the list ([Table\u20053](#t03){ref-type=\"table\"}).\n\nTABLE 3Authors with Two or More Top-Cited ET ArticlesRankAuthorNumber of Citation ClassicsNumber of Articlesh-indexAs First AuthorAs Co-Author1Louis ED271710482Koller WC15411583Jankovic J12210834Pahwa R1138465Lyons K11011216Findley LJ936367Rajput AH854408Rajput A826379Benito-Le\u00f3n J7612910Elble R6241911Deuschl G6245612Bermejo-Pareja F6152013Busenbark K5141614Ford B5053515Vonsattel JP5053416Ondo W4312017Tr\u00f6ster AI4133718Honig LS4042819Wilkinson S4041720Wenzelburger R4041621Thompson PD4045422Ottman R4044123Hubble JP4042724Brooks DJ4048725Bain PG3212526Raethjen J3121727Speelman JD3033228Ross GW3033929Pullman S3031230Marsden CD30312631Jurewicz EC3031032Hariz MI2203133Connor GS211334Colebatch JG2113835Benabid AL2116036Stolze H2111937Deng H2112238Wills AJ2111539Alesch F2111840Hauser W2022341Weiner WJ2024542Tang MX2025043Stefansson K2022444Schwartz K2021945Rozdilsky B2022946Robinson CA202847Overman J2021048Oertel WH2025549Nash J2021450Moskowitz CB2021451Lee H2023152Kong A202653Jenkins IH2021654Jakobsson F202655Higgins JJ2021856Benedickz J202157Frackowiak RSJ2027658Faust PL2021259Cote L2023360Borden S202461Barnes LF202762Andrews H20226\n\nAccording to their countries of origin (based on the first author), more than half of the articles were generated in the United States (n\u200a=\u200a54), which led the list ([Table\u20054](#t04){ref-type=\"table\"}). The United Kingdom (n\u200a=\u200a11) was the second-most productive country, followed by Spain (n\u200a=\u200a8), Germany (n\u200a=\u200a7), and Canada (n\u200a=\u200a4). France and Sweden each contributed three articles to the list. The 13 leading institutions that provided two or more top-cited ET articles are listed in [Table\u20055](#t05){ref-type=\"table\"}. Columbia University in the United States produced the largest number of top-cited ET articles (n\u200a=\u200a19), followed by the University of Kansas (n\u200a=\u200a18), and Baylor College of Medicine (n\u200a=\u200a9), both in the United States, and M\u00f3stoles University Hospital in Spain (n\u200a=\u200a6).\n\nTABLE 4Country of Origin of the 100 Top-Cited ET ArticlesRankCountry of OriginNumber of Articles1USA542United Kingdom113Spain84Germany75Canada46France37Sweden38Austria29Italy210Iceland211Finland112Belgium113Turkey114Papua New Guinea1\n\nTABLE 5Institution of Origin of Authors with Two or More Top-Cited ET ArticlesRankCountry of OriginNumber of Articles1Columbia University, New York, New York, USA192University of Kansas Medical Center, Kansas City, Kansas, USA183Baylor College of Medicine, Houston, Texas, USA94M\u00f3stoles University Hospital, M\u00f3stoles, Madrid, Spain65University of Saskatchewan, Saskatoon, Saskatchewan, Canada56Hammersmith Hospital, London, United Kingdom47Christian-Albrechts-University of Kiel, Kiel, Germany48Oldchurch Hospital, Romford, United Kingdom39Institute of Neurology, London, United Kingdom210deCODE Genetics, Reykjavik, Iceland211Southern Illinois University School of Medicine, Springfield, Illinois, USA212University of Vienna, Austria213National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland, USA2\n\nAmong the 100 top-cited ET articles, a large number had a case-series design (n\u200a=\u200a34), and the next most common design was the case-control study (n\u200a=\u200a24) ([Table\u20056](#t06){ref-type=\"table\"}). Only six articles were clinical trials ([Table\u20056](#t06){ref-type=\"table\"}). According to the study designs, the level of evidence of articles is presented in [Table\u20057](#t07){ref-type=\"table\"}. Most articles were classified as level B evidence.\n\nTABLE 6Study Design of the Clinical StudiesStudy DesignNumber of Articles***Experimental study***Randomized, controlled trial, open-label trial6***Observational study***Cohort study4Case-control study24***Descriptive study***Cross-sectional study10Case series34Case report2***Methods paper***1***Review articles or expert opinion***18***Animal research***1\n\nTABLE 7Level of Evidence of the Clinical StudiesLevels of EvidenceNumber of Articles*Level A: Well-conducted RCT with 100 patients or more (including multi-center and meta-analyses); well-conducted RCT with fewer than 100 patients (one or more institutions and meta-analysis; well-conducted study)*1*Level B: Well-conducted case-control study, poorly controlled or uncontrolled (including RCT with one or more major or three or more minor methodological flaws), observations studies with high potential for bias (case series with comparison to historical controls), case series or case reports, conflicting evidence with more support*79*Level C: Expert opinion*18\n\nWe also performed an analysis in which we excluded any examples of self-citations (i.e., instances in which authors cited their own work), and the results of this analysis did not change the main findings of the study (e.g., country of origin, institution of origin, top cited authors, etc.).\n\nDiscussion {#s4}\n==========\n\nIn medical literature, the study of the number of times authors reference an article is one measure of the influence of the publication, and this type of citation analysis is widespread.[@b01]--[@b03] The evaluation of specialty-wide citation analysis has been reported in other areas of the neurosciences.[@b05]--[@b11] By ranking the 100 most-cited works, we aimed to determine which published journal articles on ET have exerted the most citation influence. The top article was cited 293 times. This figure is far lower than that found in Parkinson's disease, in which the top-cited article garnered 4,327 citations.[@b07]\n\nIn 2002, Callaham et al.[@b30] found that the impact factor of the original publishing journal, rather than the methodology or quality of the research, was an effective predictor of the number of times an article would be cited per year. However, as shown in the present study ([Table\u20052](#t02){ref-type=\"table\"}), the actual citation value of the individual article did not positively relate to the impact factor of the journal.\n\nIn general, as shown in [Table\u20053](#t03){ref-type=\"table\"}, the 100 top-cited articles in ET were articles that have been available for 10 or more years, and only one target article was published more recently (in 2010). It has been reported that scientific articles begin to be cited 1 or 2 years after publication, and reach a maximum citation rate 7 to 10 years after publication.[@b31] However, an interval of 10 to 20 years is needed for maximal recognition of prominent articles in a field.[@b32],[@b33] This may explain why recently published articles were cited rarely, and few appeared on the list.\n\nAmong the 100 top-cited articles on ET, the most common design was the case series design (n\u200a=\u200a34), followed by the case-control study (n\u200a=\u200a24) ([Table\u20056](#t06){ref-type=\"table\"}), implying that descriptive and observational studies are most frequent for ET. This suggests the relative ease of carrying out simpler study designs in ET. It is known that different study designs could correspond to different levels of evidence. In the hierarchy of research study designs, systematic reviews, meta-analyses, and well-conducted randomized clinical trials (RCTs) provide the highest quality of evidence for most clinical or interventional questions, and the lowest grade is applied to expert opinions. Among the 100 target articles, there were only six RCTs. Moreover, only one RCT was found among the top 20 medical articles. This is a common finding among the different studies that have analyzed the 100 most-cited papers in different disciplines.[@b07],[@b31],[@b34]--[@b36] There are several possible explanations for the low numbers of RCTs. First, RCTs are expensive and time consuming. Second, it is difficult to gather large sample and control groups. Third, it may be that RCTs were published relatively recently, so they had not yet reached a significant number of citations.\n\nCompared to Parkinson's disease, there are very few articles addressing more basic science issues in ET (e.g., laboratory studies including gene discovery, molecular biology, and cellular biology; neuroimaging studies; neuropathological studies; and animal model studies).[@b07] Scientific understanding of ET is in its infancy.\n\nThis study had a number of potential limitations. First and foremost, this survey was restricted to journal articles with the term \"essential tremor\" in the title. In other words, some significant ET articles could have been excluded, and it is possible that these were high-quality articles with a large number of citations. Second, this type of citation analysis does not include citations in textbooks and lectures, and an author's or authors' potential preference to cite articles in the journal in which they seek to publish their work.[@b37] Third, there is a clear time effect in citation analysis, with the most recent articles being at a disadvantage.[@b30] The time from publication played an apparent role here, with late 1990s and early 2000s being the most prominent years of publication for the 100 most-cited articles. Recent articles are clearly disadvantaged in citation analysis. Fourth, the language of publication plays a major role, with an obvious bias for articles published in English-language journals. Fifth, there are biases inherent in the age of the database, as any articles published in the 19th or early 20th century would be excluded, which likely causes some true \"classic\" articles to be excluded. Finally, Kuhnian philosophy[@b38] would tell us that in a scientific community there is a tendency for adherence to a paradigm. In this context, there is \"snowball effect\" for citations, as other authors are more likely to cite an article because of previous citations to that article rather than its content or quality.\n\n**Funding:** Dr. Benito-Le\u00f3n is supported by the National Institutes of Health, Bethesda, MD, USA (R01 NS039422), and the Commission of the European Union (grant ICT-2011-287739, NeuroTREMOR, principal investigator). Dr. Elan D. Louis has received research support from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD, USA (NINDS R01 NS042859, principal investigator; NINDS R01 NS39422, principal investigator; NINDS T32 NS07153-24, principal investigator; NINDS R01 NS073872, principal investigator; NINDS R21 NS077094, co-investigator; and NINDS R01 NS36630, co-investigator) and the Parkinson's Disease Foundation (principal investigator).\n\n**Financial disclosures:** None.\n\n**Conflict of Interests:** The authors report no conflict of interest.\n\n[^1]: J.B-L. and E.D.L. were both responsible for the conception, design, organization, and execution of the research project and manuscript. J.B-L. wrote the first draft of the manuscript.\n"}
+{"text": "All relevant data are within the paper and its Supporting Information files.\n\nIntroduction {#sec001}\n============\n\nMajor depression is a highly disabling medical condition that largely contributes to the global disease burden \\[[@pone.0235046.ref001]\\]. Presently, it is the most significant risk factor for suicides. Roughly one third of patients with major depression do not respond to prescription antidepressants, but for those who do, the therapeutic effects are evident with a delay of weeks or months of medication. Electroconvulsive therapy (ECT) remains among the most potent treatments for pharmacoresistant depression. Reported response rates to ECT are high, especially for melancholic depression \\[[@pone.0235046.ref002]--[@pone.0235046.ref004]\\]. Although relatively safe, ECT may produce adverse effects, such as retrograde amnesia, headache, and nausea \\[[@pone.0235046.ref002]\\].\n\nThe neurobiological basis of the antidepressant effects of ECT is poorly understood. However, the induction of intrinsic neurotrophic mechanisms, such as activation of BDNF (brain-derived neurotrophic factor) signaling, has been proposed to play a significant role \\[[@pone.0235046.ref005]--[@pone.0235046.ref007]\\]. Increase in cortical and hippocampal BDNF mRNA \\[[@pone.0235046.ref008]--[@pone.0235046.ref010]\\] and protein \\[[@pone.0235046.ref011]--[@pone.0235046.ref013]\\] have been consistently reported after electroconvulsive shock (ECS, an animal model of ECT) treatments. BDNF modulates formation and plasticity of neuronal networks \\[[@pone.0235046.ref014]--[@pone.0235046.ref016]\\], and infusions of BDNF into the prefrontal cortex and hippocampus have been shown to mimic the behavioral effects of antidepressants in rodents \\[[@pone.0235046.ref017],[@pone.0235046.ref018]\\]. BDNF has also been implicated in other antidepressant treatments, since the BDNF receptor TrkB (tropomyosin related kinase B) is activated by a variety of pharmacologically diverse antidepressant drugs \\[[@pone.0235046.ref019]--[@pone.0235046.ref021]\\], with animals having decreased BDNF-TrkB signaling showing reduced responses to antidepressant treatments \\[[@pone.0235046.ref019],[@pone.0235046.ref022]--[@pone.0235046.ref025]\\].\n\nRather than mere seizure manifestation or its desired duration, certain post-ictal (i.e. after seizure) events, such as slow wave EEG activity and EEG burst-suppression, have been suggested to predict the efficacy and onset-of-action of ECT \\[[@pone.0235046.ref026]--[@pone.0235046.ref030]\\]. General anesthetics, such as isoflurane, dose-dependently produce slowing of EEG activity. When anesthesia deepens, a burst-suppressing EEG pattern is achieved, characterized by bursts of neural activity interrupted by transient periods of electrocerebral silence \\[[@pone.0235046.ref031]--[@pone.0235046.ref035]\\]. This similarity to the post-ictal effects of ECT and deep anesthesia prompted research on the exciting possibility that burst-suppressing anesthesia (referred to as \"narcotherapy\") would be sufficient to recapitulate the therapeutic effects of ECT in depressed patients. In preliminary clinical studies, isoflurane showed an antidepressant effect comparable to ECT, even after a single dose \\[[@pone.0235046.ref031],[@pone.0235046.ref032],[@pone.0235046.ref036]\\]. However, subsequent findings remained however inconsistent and did not unequivocally support therapeutic effects of anesthesia in depressed patients \\[[@pone.0235046.ref037]--[@pone.0235046.ref040]\\].\n\nRecent clinical and preclinical observations have renewed the interest to investigate the antidepressant effects of deep anesthesia \\[[@pone.0235046.ref041]\\]. Weeks et al. demonstrated that a series of ten burst-suppressing isoflurane anesthesia sessions for 15 minutes was comparable to ECT in antidepressant efficacy in patients with medication-refractory depression, and more tolerable than ECT regarding neurocognitive side effects \\[[@pone.0235046.ref042]\\]. The same group subsequently reported similar findings with repeated propofol anesthesia \\[[@pone.0235046.ref043]\\]. Moreover, we and others have shown that a single isoflurane anesthesia exposure produces antidepressant-like effects in the learned helplessness depression model and in the forced swim test \\[[@pone.0235046.ref044],[@pone.0235046.ref045]\\], while halothane, another anesthetic agent that produces negligible burst-suppression, lacks such effects \\[[@pone.0235046.ref045]\\]. Furthermore, isoflurane activates TrkB receptors in a dose-dependent manner, with the most prominent effects observed when burst-suppression is achieved \\[[@pone.0235046.ref035],[@pone.0235046.ref044]\\]. However, activation of TrkB becomes evident even during slight sedation with agents not shown to possess antidepressant effects, indicating that TrkB activation is not *per se* sufficient for antidepressant effects \\[[@pone.0235046.ref046]\\], and other mechanisms are likely involved.\n\nTo further test \"the cerebral silence hypothesis\" of ECT and the antidepressant effects of isoflurane anesthesia in particular, we investigated whether repeated isoflurane exposures increase BDNF protein, while ameliorating depressive-like symptoms in Wistar outbred rats (Crl:WI(Han)) subjected to chronic mild stress (CMS). We have recently shown that the depressive-like phenotype in these rats is restored by repeated ECS treatments, which also readily increases BDNF synthesis, while the selective serotonin reuptake inhibitor (SSRI) citalopram was ineffective \\[[@pone.0235046.ref047]\\].\n\nMaterial and methods {#sec002}\n====================\n\nAnimals {#sec003}\n-------\n\nA total of 44 adult male Wistar outbred rats (Crl:WI(Han)) were used for the studies (Charles River, Sulzfeld, Germany). Age upon arrival was 9 weeks. Rats were single-housed in Makrolon type III cages on Altromin soft wood granulate. Standard laboratory chow (Altromin 1324 standard diet; Altromin, Lage, Germany) and tap water were provided *ad libitum*, except when CMS procedure required food and/or water deprivation. The controlled 12 h light/12 h dark schedule was only disturbed during stress procedure. All rats were adapted to the laboratory and habituated to handling for at least one week before starting the experiments. Experiments were done in compliance with the European Communities Council Directive of 24 November 1986 (86/609/EEC), and were approved by the animal subjects review board of University of Veterinary Medicine Hannover (LAVES--Lower Saxony State Office for Consumer Protection and Food safety, approval number 12/0871). All efforts were made to minimize pain or discomfort of the animals used. Animals were handled daily, and their general well-being, indicated by grooming behavior and body posture was monitored. Body weight of the animals was measured at least every other day. Two animals that showed more than 20% weight loss over a period of 3 days (humane end point) were excluded from the experiments.\n\nChronic mild stress (CMS) {#sec004}\n-------------------------\n\nRats were exposed to mild stressors at varying time points during light and dark period as described \\[[@pone.0235046.ref047],[@pone.0235046.ref048]\\]. The stressors were delivered daily except when isoflurane/sham treatments were given, and the behavioral performances of the animals were assessed as shown in **[Fig 1](#pone.0235046.g001){ref-type=\"fig\"}**. Stressors included periods of (1) continuous light (24 h/d), (2) food deprivation (24 h), (3) water deprivation (14 h), (5) swim sessions in 40\u00b0C water (10 minutes in a transparent plexiglas cylinder (50 cm deep, 25 cm diameter) containing 20 cm of water), (6) swim sessions in 15\u00b0C water (5 minutes in similar conditions to swim sessions in 40\u00b0C water), (7) wet bedding (16 h, 300 ml of tap water on Altromin soft wood granulate), (8) restraint stress (30 min) and (9) social crowding (four rats in one Makrolon type III cage). No stress was applied on the days of behavioral testing or isoflurane/sham administrations. Control rats were left undisturbed and handled regularly.\n\n![Study timeline showing stress induction, isoflurane anesthesia administration and behavioral testing.\\\nCMS = chronic mild stress, EPM = elevated plus maze, ISO = isoflurane anesthesia, OFT = Open field test, SCT = sucrose consumption test.](pone.0235046.g001){#pone.0235046.g001}\n\nWeight measurement {#sec005}\n------------------\n\nBody weight was monitored during the course of experiments as a measure for general health. A reduction in body weight or a diminished weight gain reflects a reduced well-being of the rats \\[[@pone.0235046.ref049]\\].\n\nSucrose consumption test (SCT) {#sec006}\n------------------------------\n\nHedonic deficits induced by CMS can be measured as a decrease in consumption or preference for sweet solution \\[[@pone.0235046.ref050]\\]. During test sessions rats had free access to a bottle of 1% sucrose solution and a bottle of tap water for 14 h. No food or water deprivation was performed before testing. Animals were habituated to the testing in three habituation trials. The position of the bottles was switched after every test session to avoid possible effects of side preference in drinking behavior. The individualized acquisition of sucrose intake provides the opportunity to select between anhedonic-like rats (stress responders) and hedonic-like rats (stress non-responders). Anhedonic- or hedonic-like behavior is based on the individual amount of sucrose solution intake. Rats showing \\>25% within-subject decrease in sucrose consumption were considered anhedonic while rats showing \\<10% within-subject decrease in sucrose consumption were considered hedonic \\[[@pone.0235046.ref001]\\]. Animals not belonging to either criterion were considered as unclassifiable. Sucrose consumption was measured 3 to 5 times before CMS. After isoflurane anesthesia, alterations in anhedonic-like behavior were assessed by estimating within subject changes in sucrose consumption. According to Christensen et al., positive treatment responders were anhedonic-like animals showing \\>20% within-subject increase in sucrose consumption, whereas non-responders show \\<20% within-subject increase in sucrose consumption. The mean of the sucrose consumption determined in these trials was set as 100% and considered baseline level.\n\nOpen field test {#sec007}\n---------------\n\nThe open field test is a routine method to measure locomotor activity and anxiety-like behavior in rodents \\[[@pone.0235046.ref051]\\]. The test was performed in a round arena made of black PVC (diameter 80 cm) which was divided into three zones (center, inner, outer). The animals were placed individually in the center of the open field. Distance moved and time spent in the center of the open field was recorded for 5 min and analyzed with EthoVision\u00aeXT7 software (Noldus Information Technology, Wagening, Netherlands).\n\nElevated plus maze test {#sec008}\n-----------------------\n\nThe elevated-plus maze measures the level of anxiety in rodents \\[[@pone.0235046.ref052]\\]. The apparatus was constructed with black plastic. It comprises two open arms (50x10 cm), two enclosed arms (50x10x30 cm), and a central platform (10x10 cm). The configuration has the shape of a plus sign, and the apparatus is elevated 80 cm above the floor level. At the beginning of the test, rats were placed on the central platform always facing the same closed arm. The behavior of rats in the test was analyzed for 5 min using the EthoVision\u00aeXT7 software (Noldus Information Technology, Wagening, Netherlands). Time spent in different sections of the maze (open and closed arms) and the frequency of entries into open and closed arms were assessed.\n\nIsoflurane exposure {#sec009}\n-------------------\n\nRats were randomly allocated to the treatment groups: control + sham (N = 8), control + isoflurane (N = 8), CMS + sham (N = 14), CMS + isoflurane (N = 14). Rats were placed into an anesthesia box and exposed to isoflurane (induction: 4% for 2 min; maintenance: 2% for 13 min; airflow of 1.0 l/min). This isoflurane dosing regimen produces a rapid burst-suppression EEG state highly reliably in both rodents and humans \\[[@pone.0235046.ref031],[@pone.0235046.ref034],[@pone.0235046.ref035],[@pone.0235046.ref045],[@pone.0235046.ref053]\\]. Sham animals were kept in the anesthesia boxes for 2 min without isoflurane. To measure the behavioral outcomes in SCT and open field tests in between administrations, and to model the preliminary clinical studies demonstrating isoflurane's antidepressant effect \\[[@pone.0235046.ref032],[@pone.0235046.ref042]\\], a single treatment was given once every third day over 15 days (= 5 total treatments).\n\nBDNF ELISA {#sec010}\n----------\n\nAfter the behavioral experiments, the animals were euthanized by decapitation after a brief exposure to carbon dioxide. Tissue samples from the medial prefrontal cortex and hippocampus were rapidly dissected and snap-freezed. BDNF protein levels were analyzed using a commercial BDNF ELISA kit (Quantikine^\u00ae^ ELISA Kit, catalog \\#DBD00, R&D Systems Europe Ltd., Abingdon, UK). The samples were homogenized in NP++ lysis buffer (137 mM NaCl, 20 mM Tris, 1% NP-40, 10% glycerol, Pierce\u2122 Protease and Phosphatase Inhibitor tablets (Thermo Fisher Scientific, Waltham, MA), 48 mM NaF), incubated on ice for 15 minutes, centrifuged (16,000 g, 15 min, 4\u00b0C), and the supernatants were collected for further processing. The samples were acidified to pH 3 with 1 M HCl, followed by neutralization with 1 M NaOH. The samples were loaded on a pre-coated (with monoclonal BDNF antibody) and pre-blocked 96-well plate containing serial diluted BDNF standards and hippocampal samples from adult male conditional Bdnf^-/-^ knockout \\[[@pone.0235046.ref054],[@pone.0235046.ref055]\\] and wild-type mice (kindly provided by Dr. Maribel Rios), and incubated for 2 hours at RT. The plate was then incubated with HRP-conjugated secondary monoclonal BDNF antibody for 1 hour at RT, followed by three washes with provided wash buffer, and then incubated with color reagents (hydrogen peroxide and chromogen). The reaction was stopped with 2 M H~2~SO~4~ after a 30-minute incubation, and the plate was read for absorbance in 450 nm. The obtained results were normalized to total protein concentrations of each sample.\n\nStatistics {#sec011}\n----------\n\nData are shown as mean \u00b1 SEM (standard error of mean). Two-way analysis of variance (ANOVA) (two categorical independent variables), repeated measures ANOVA followed by Sidak's multiple comparisons test, or Student's unpaired t-test were used for statistical evaluation (Prism 7 software, GraphPad (La Jolla, CA, USA). A *P*\\<0.05 was considered statistically significant. Details of statistical tests are shown in [S1 Table](#pone.0235046.s003){ref-type=\"supplementary-material\"}.\n\nResults {#sec012}\n=======\n\nChronic mild stress induced alterations in sucrose consumption and body weight in rats {#sec013}\n--------------------------------------------------------------------------------------\n\nChronic mild stress (CMS) is considered one of the most valid animal models of depression \\[[@pone.0235046.ref056]\\]. In this model, as the name implies, the animals are repeatedly subjected to various stressors during a course of several weeks, which may induce depression-like phenotypes, most notably anhedonia (e.g. reduced consumption of sweetened solution). The strength of this model is that animals respond to chronic, but not acute, administration of antidepressant drugs (and to ECS), as compared to drugs without clinical antidepressant properties that show no effects \\[[@pone.0235046.ref056]\\]. We recently employed this model using a stress-sensitive substrain of male outbred Wistar rats (Crl:WI(Han)) \\[[@pone.0235046.ref047]\\]. Repeated ECS ameliorated depression-like phenotypes induced by CMS and significantly increased BDNF levels. Additionally, the SSRI citalopram had negligible effects on both phenotype and BDNF levels \\[[@pone.0235046.ref047]\\].\n\nAs shown recently in this rat strain (Neyazi et al. 2018), during three weeks of CMS rats begin to segregate based on behavioral change to anhedonic-like or hedonic-like behavioral groups. According to Christensen et al. (2011), anhedonic-like animals are expected to show a \\>\u200925% within-subject decrease in sucrose intake, whereas hedonic-like rats are expected to show a \\<\u200910% within-subject reduction in sucrose intake. Animals not responding to either criterion are considered unclassifiable. In the present experiments, anhedonic-like behavior was present in 53.8% (14/26) of the animals, whereas hedonic-like behavior was detected in the remaining animals (12/26) (**[Fig 2A](#pone.0235046.g002){ref-type=\"fig\"}**). None of the unstressed control animals showed anhedonic-like behavior. The average weight gain of the rats during three weeks of CMS was significantly lower compared to weight gain of unstressed controls (Repeated measures ANOVA: F~2,39~ = 26.44, P\\<0.0001) (**[Fig 2B](#pone.0235046.g002){ref-type=\"fig\"}**).\n\n![Effects of chronic mild stress on sucrose consumption and body weight.\\\n**A**). Sucrose consumption survey over 3 weeks of CMS demonstrates that 53.8% (14/26) of the animals responded to stress with anhedonic behavior characterized by a \\>25% within-subject decrease in sucrose consumption (small dashed lines). 46.2% (12/26) of the CMS exposed animals were classified stress-resilient, showing a within-subject decrease in sucrose intake of \\<10% (wide dashed lines). Hedonic-like behavior was present in all unstressed control rats. **B**) Stressed rats gained significantly less body weight during three weeks of CMS compared to controls (Repeated measures ANOVA: F~2,39~ = 26.44, P\\<0.0001). CMS = chronic mild stress. Data is shown as mean \u00b1 SEM. \\*\\*\\<0.01, \\*\\*\\*\\<0.001 (control vs. anhedonic-like), \\#\\#\\<0.01, \\#\\#\\#\\<0.001 (control vs. stress-resilient), repeated measures ANOVA followed by Sidak's multiple comparisons test.](pone.0235046.g002){#pone.0235046.g002}\n\nLack of effects of repeated isoflurane anesthesia on behavioral changes induced by CMS {#sec014}\n--------------------------------------------------------------------------------------\n\nThe animals were next subjected to 15-minute burst-suppressing isoflurane anesthesia (induction: 4%; maintenance: 2%) \\[[@pone.0235046.ref034],[@pone.0235046.ref035]\\] or sham anesthesia every third day over a 15-day period for a total of 5 consecutive treatments (**[Fig 1](#pone.0235046.g001){ref-type=\"fig\"}**). Antidepressant and anxiolytic effects of isoflurane were assessed after the first (rapid), third, and fifth drug exposure using the sucrose consumption, the open field, and the elevated plus maze tests. Based on the criteria by Christensen et al., positive treatment responders were considered as anhedonic-like animals showing \\>20% within-subject increase in sucrose intake, whereas non-responders were considered to show \\<20% within-subject increase in sucrose intake \\[[@pone.0235046.ref001]\\]. Exposure to isoflurane exerted no significant effects on sucrose consumption in anhedonic-like (Repeated measures ANOVA: F~1,12~ = 0.06915, P = 0.80), stress-resilient (F~1,10~ = 0.04871, P = 0.83) or sham rats (F~1,14~ = 4.117, P = 0.06) (**[Fig 3A](#pone.0235046.g003){ref-type=\"fig\"}**). If anything, the sucrose consumption observed in non-stressed rats was reduced by isoflurane treatment, although this effect was not significant. Sucrose consumption in anhedonic-like groups remained low throughout the experiments, indicating that a depressive-like phenotype induced by the CMS protocol was sustained throughout the experiments (**[S1 Fig](#pone.0235046.s001){ref-type=\"supplementary-material\"}**). In addition, isoflurane produced only minor behavioral effects in the open field or the elevated plus maze tests (**[Fig 3B and 3C](#pone.0235046.g003){ref-type=\"fig\"}, [S2 Fig](#pone.0235046.s002){ref-type=\"supplementary-material\"}**). Isoflurane exerted contrasting effects in control and anhedonic-like groups on number of entries (Two-way ANOVA, treatment x phenotype: F~2,35~ = 3.579, P = 0.0385) and time spent in open arm (F~2,35~ = 5.845, P = 0.0065) of the elevated plus-maze (**[Fig 3C](#pone.0235046.g003){ref-type=\"fig\"}**), and decreased overall locomotor activity in the open field task after the first (Two-way ANOVA, treatment effect: F~1,36~ = 5.693, P = 0.0224) and third (F~1,36~ = 4.212, P = 0.0475) isoflurane administrations (**[S2 Fig](#pone.0235046.s002){ref-type=\"supplementary-material\"}**).\n\n![Lack of antidepressant effects of isoflurane anesthesia in a chronic mild stress model of depression.\\\n**A)** Changes in sucrose consumption of control, stress-resilient, and anhedonic animals 2 days after isoflurane administrations following CMS exposure. **B)** Distance traveled and time spent at arena center in open field test 24 hours after 5th isoflurane administration. **C)** Results in elevated plus maze test 3 days after 5th isoflurane exposure. ISO = isoflurane anesthesia. Data is shown as mean \u00b1 SEM. \\*\\<0.05, two-way ANOVA.](pone.0235046.g003){#pone.0235046.g003}\n\nBrain BDNF levels remain unaltered after CMS and isoflurane administrations {#sec015}\n---------------------------------------------------------------------------\n\nAfter the behavioral experiments the animals were euthanized and samples collected from the medial prefrontal cortex and hippocampus to determine BDNF protein levels. To test the specificity of the ELISA assay, we also determined BDNF expression in hippocampal homogenates obtained from adult male conditional Bdnf^-/-^ mice and their wild-type littermates. Results from BDNF protein analysis show negligible effects of the animals' response to CMS (Two-way ANOVA, phenotype effect, PFC: F1,15 = 0.3992, P = 0.54; HC: F1,28 = 0.7168, P = 0.40), and isoflurane anesthesia (Two-way ANOVA, treatment effect, PFC: F1,15 = 0.05878, P = 0.81; HC: F1,28 = 0.4036, P = 0.53 (**[Fig 4A](#pone.0235046.g004){ref-type=\"fig\"}**).\n\n![CMS and isoflurane anesthesia have no significant effect on BDNF protein levels in the rat medial prefrontal cortex (PFC) and hippocampus (HC).\\\n(B) Assay specificity was determined with cortical samples of conditional BDNF^-/-^ mice that showed negligible signal in comparison to wild-type littermates. CMS = chronic mild stress, ISO = isoflurane anesthesia. Data is shown as mean \u00b1 SEM. \\*\\*\\*\\<0.001, Student's unpaired t-test.](pone.0235046.g004){#pone.0235046.g004}\n\nDiscussion {#sec016}\n==========\n\nPost-ictal EEG suppression has been proposed to predict the antidepressant effects of ECT \\[[@pone.0235046.ref026]--[@pone.0235046.ref030]\\]. Like ECT, the volatile anesthetic isoflurane causes EEG burst-suppression in humans and rodents when adequate dosing is applied \\[[@pone.0235046.ref032],[@pone.0235046.ref035]\\]. Clinical and preclinical evidence indicates that such burst-suppressing isoflurane anesthesia ameliorates depressive symptoms in patients, and elicits antidepressant-like effects in rodents \\[[@pone.0235046.ref031],[@pone.0235046.ref032],[@pone.0235046.ref036],[@pone.0235046.ref042],[@pone.0235046.ref044],[@pone.0235046.ref045],[@pone.0235046.ref057]\\]. Already a single brief isoflurane anesthesia has demonstrated antidepressant-like effects in the forced swim test and learned helplessness model in rodents \\[[@pone.0235046.ref044],[@pone.0235046.ref045]\\]. Amelioration of anhedonic behavior was also observed after a single isoflurane anesthesia exposure in a mouse model of CMS \\[[@pone.0235046.ref057]\\]. Here, we utilized a CMS model in a stress-sensitive substrain of rats that respond to ECS, but not citalopram \\[[@pone.0235046.ref047]\\] in order to further test the antidepressant-like effects of repeated burst-suppressing isoflurane anesthesia. The dosing of isoflurane was selected based on our earlier data to achieve reliable burst-suppression pattern and TrkB signaling \\[[@pone.0235046.ref035]\\], which is one of the main pathways targeted by antidepressants \\[[@pone.0235046.ref058]\\]. A subset of rats responded to stress by showing reduced sucrose consumption (a marker of anhedonia), while some of the animals remained stress-resilient, a finding previously observed \\[[@pone.0235046.ref001],[@pone.0235046.ref047]\\]. Unexpectedly, we found no significant behavioral changes in any of the treatment groups after isoflurane administrations at any point during the course of the experiments.\n\nRepeated exposures to anesthesia had no impact on BDNF levels, a finding that contrasts numerous studies showing that all other antidepressants increase BDNF synthesis \\[[@pone.0235046.ref058]\\]. During ECT practice, an electric current is delivered onto the scalp of the patient under anesthesia, which leads to transient epileptiform EEG activity. This robust increase in neuronal activity likely underlies the stimulatory effects of ECT on BDNF levels, since various types of neuronal stimuli--especially generalized convulsions--have been shown to increase BDNF synthesis \\[[@pone.0235046.ref014],[@pone.0235046.ref059]--[@pone.0235046.ref061]\\]. Isoflurane shares the capability to induce electrocerebral silence with ECT, but it often brings no preceding convulsions or seizure activity. Indeed, the effects of anesthetics, such as isoflurane, on brain and blood BDNF levels generally remain negligible or even decrease \\[[@pone.0235046.ref044],[@pone.0235046.ref062]--[@pone.0235046.ref070]\\]. Anesthesia also blocks rTMS (repetitive transcranial magnetic stimulation) induced BDNF synthesis \\[[@pone.0235046.ref071]\\].\n\nDespite anesthesia producing a state of widespread depression in the CNS, paradoxical neuronal excitation has been reported with diverse anesthetics, especially when the concentration of anesthetic is low \\[[@pone.0235046.ref072],[@pone.0235046.ref073]\\]. This is particularly well exemplified by ketamine, a rapid-acting antidepressant that is used in subanesthetic dosing to treat depression. Ketamine provokes cortical excitability by increasing glutamatergic neurotransmission \\[[@pone.0235046.ref074],[@pone.0235046.ref075]\\]. This excitatory response has been shown to be required for its antidepressant-like effect in rodents \\[[@pone.0235046.ref076],[@pone.0235046.ref077]\\]. Furthermore, an increase in cortical excitability after ketamine administration has been associated with a positive antidepressant treatment response in patients \\[[@pone.0235046.ref078]\\]. More recently, we have shown that subanesthetic doses of nitrous oxide, a putative rapid-acting antidepressant \\[[@pone.0235046.ref079]\\], readily up-regulates BDNF synthesis and several other markers of neuronal excitation \\[[@pone.0235046.ref046]\\]. Notably, both subanesthetic ketamine and nitrous oxide evoke slow wave activity, as measured by EEG, after the peak of their pharmacological effects, resembling the post-ictal state following ECT. Rapid-acting antidepressants may therefore require both a phase of neuronal excitation, and emergence of slow wave activity to elicit their therapeutic effects \\[[@pone.0235046.ref080]\\]. It's tempting to speculate that isoflurane's antidepressant effects may be dependent on the treatment protocol\u00b4s (unpredictable) capability to induce sufficient neuronal excitation, BDNF synthesis, and EEG silencing. Indeed, isoflurane has been shown to elicit antidepressant-like effects in a mouse model of CMS using an administration protocol that also increased BDNF expression \\[[@pone.0235046.ref057]\\]. Additionally, isoflurane is known to produce occasional excitatory responses and behavioral hyperactivity/agitation particularly during anesthesia induction and emergence \\[[@pone.0235046.ref081]--[@pone.0235046.ref083]\\]. Excitation may also occur during deep burst-suppressing anesthesia, where isoflurane increases cortical excitability in response to various stimuli \\[[@pone.0235046.ref084]--[@pone.0235046.ref086]\\]. Validation of anesthesia treatment regimens capable of producing these effects should therefore be considered in future studies \\[[@pone.0235046.ref041]\\].\n\nSupporting information {#sec017}\n======================\n\n###### Sucrose consumption of the anhedonic rats remain significantly lower than in control group throughout the experiments.\n\nCMS = chronic mild stress. Data is shown as mean \u00b1 SEM. \\*\\*\\*\\<0.001, \\*\\<0.05, Repeated measures ANOVA followed by Sidak's multiple comparisons test.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Effects of isoflurane in open field test 24 hours after 1st and 3rd isoflurane administration.\n\nISO = isoflurane anesthesia. Data is shown as mean \u00b1 SEM.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\n###### Statistical analyses and *n* numbers.\n\n(PDF)\n\n###### \n\nClick here for additional data file.\n\nWe would like to thank Dr. Claudia Brandt for assistance during the experiments, and Dr. Giuseppe P. Cortese for language editing and his constructive comments on the manuscript.\n\n[^1]: **Competing Interests:**T.R. and W.T. are listed as co-inventors on a patent application wherein in new tools enabling the development of rapid-acting antidepressants and the efficacy monitors thereof are disclosed based on the homeostatic emergence of slow wave EEG (Rantam\u00e4ki T, Kohtala S, Theilmann W: Methods for determining the therapeutic efficacy of rapid-acting antidepressants and personalized antidepressant therapy related thereto. National Board of patents and registration of Finland, \\#20176012). T.R. and W.T. have assigned their patent rights to the University of Helsinki but will share a percentage of any royalties that may be received by the University of Helsinki. This does not alter our adherence to PLoS ONE policies on sharing data and materials. Other authors have declared that no competing interests exist.\n"}
+{"text": "A crucial mechanism in the establishment of cellular senescence is the activation of the INK4/ARF locus, which is epigenetically regulated and under tight control of the Polycomb group (PcG) Trithorax group (TrxG) proteins \\[[@R1]\\]. In proliferating cells, the locus is silenced by Polycomb repressive complexes (PRCs), and the chromatin is enriched in H3K27me3 \\[[@R2]\\]. Upon senescence triggers, PRCs are displaced and the repressive H3K27me3 mark is removed. Instead, the Trithorax MLL1 complex is recruited and leads to deposition of H3K4me3 and transcriptional activation of the locus \\[[@R1]\\]. Correspondingly, it was shown that overexpression of PcG proteins delayed the onset of replicative senescence, while deletion of MLL1 bypassed OIS by repressing the INK4/ARF locus \\[[@R1],[@R3]\\]. Interestingly, we recently discovered that transcriptional regulation of p16INK4a can occur independently of H3K4me3 and H3K27me3 and is instead dependent on the proper function of transcription factors \\[[@R4]\\]. While characterizing DPY30, a common member of all H3K4 histone methyltransferase (HMTase) complexes (including MLL1), we found that DPY30 is a crucial regulator of cell proliferation. Depletion of DPY30 in human fibroblasts, as well as in transformed cells, led to severe impairment of cell cycle progression and a senescence-like phenotype \\[[@R4]\\]. This included a flattened and enlarged morphology, elevated levels of reactive oxygen species (ROS), activation of DNA damage response pathways (DDR), increased SA-\u03b2-galactosidase activity, formation of senescence-associated heterochromatin foci (SAHFs) and increased p16INK4a expression levels.\n\nAlthough activation of INK4/ARF was in accordance with the observed senescence-like phenotype, it immediately raised the question how this activation takes place in the absence of an intact DPY30-MLL/Set1 complex. We wondered whether a H3K4-specific HTMase complex might still be active in the absence of DPY30. However, H3K4me3 levels were significantly reduced on the p16INK4a promoter. These results indicate that DPY30 is essential for an active HMTase complex, but that its action is not strictly required to activate p16INK4a expression. Nevertheless, we observed an active chromatin conformation at the p16INK4a promoter in DPY30 knockdown cells that was characterized by hyperacetylation of histones H3 and H4, as well as increased Ets1/2 occupancy, a transcription factor required for activation of the INK4/ARF locus \\[[@R5]\\]. To better understand how the loss of DPY30 can lead to a senescence-like phenotype, we screened for direct DPY30 target genes using a genome-wide approach that combined ChIP sequencing for DPY30 and H3K4me3 with expression arrays in wild-type and DPY30 knockdown cells. Interestingly, DPY30 target genes are predominantly involved in cell cycle and proliferation regulation. For one of the direct DPY30 targets, *i.e.* the ID proteins, we showed that their transcriptional silencing upon DPY30 depletion was partially responsible for the senescence-like phenotype, as reintroduction of ID1 and ID3 led to a partial rescue of the observed phenotype. Accordingly, p16INK4a was de-repressed, however independently of Polycomb activity and H3K27me3. ID proteins can negatively regulate the function of transcription factors, including Rb and Ets1/2. During cell cycle progression, Ets1/2 is phosphorylated and activated by MAP kinases. ID proteins bind to phosphorylated Ets1/2 and counter balance their activity \\[[@R6]\\]. In OIS, Ets1/2 are constitutively phosphorylated by oncogenic Ras/MAPK signaling and therefore overwrite the steady state controlled by ID proteins, resulting in p16INK4a expression \\[[@R7]\\]. In replicative senescence, p16INK4a is activated due to increased expression of Ets1/2 and decreased expression of ID proteins. However, how changes in Ets1/2 and ID protein expression during aging are regulated is still elusive. Preliminary observations in aged mouse keratinocytes and replicative senescent human fibroblasts suggest that decreased ID protein expression could be a direct consequence of lower DPY30 expression levels. Interestingly, the expression of other H3K4 HMTase complex members remained unchanged in aged cells. These findings indicate that INK4/ARF is regulated by the Ets1/2 transcription factor via the fine-tuning of ID protein expression by DPY30, which might be physiologically relevant in aging (Figure [@R1]).\n\n![In aging, expression levels of DPY30 are decreased, which impairs the activity of the MLL/Set1 complex and downregulates ID protein expression. As a consequence, Ets1/2 transcription factors are free to bind to, and activate expression of the INK4/ARF locus. Thereafter, cells exit cell cycle and enter cellular senescence.](aging-05-590-g001){#F1}\n\nWe speculate that the regulation of ID protein expression by DPY30 could also be implicated in tumorigenesis. In addition to a frequent deregulation of PcG and TrxG proteins in cancer, ID protein levels are elevated in some types of tumors, which was associated with disease severity and poor prognosis \\[[@R8]\\]. We wondered if elevated ID protein levels in cancer could result from deregulated DPY30-MLL/Set1 complex action. After DPY30 depletion in several cancer cell lines, we observed not only a decrease in ID protein expression but also that proliferation was compromised, suggesting that indeed ID protein expression is under control of DPY30 and that their overexpression could be beneficial for the proliferation rate in these transformed cells. However, whether upregulation of ID proteins by DPY30 in cancer follows the same axis as in aging, namely, the regulation of Ets1/2 activity and consequently of INK4/ARF expression, has not been determined and requires further studies.\n\nGiven the importance of understanding the molecular mechanisms involved in cellular senescence and tumorigenesis, we believe that these recent findings \\[[@R4]\\] unravel an additional part of this complex mosaic of pathways and suggest an alternative possibility of p16INK4a regulation. Further investigations on the transcriptional regulation and function of the ID proteins will help to further dissect their role in aging and could possibly unravel ID protein overexpression as a tumor marker and, importantly, as a potential target for designing novel anti-cancer treatments.\n"}
+{"text": "INTRODUCTION {#sec1}\n============\n\nThe hypothalamus plays a pivotal role in feed intake and energy balance regulation in birds (Kuenzel, [@bib29]; Kuenzel et\u00a0al., [@bib28]). With few exceptions, such as a lack of response to specific orexigenic peptides (Orexin A and B, galanin, motilin and melanin concentrating hormone) (Furuse, [@bib17]), feed intake and energy homeostasis seem to be regulated in a similar way in chickens as in mammals (Furuse, [@bib17]; Boswell, [@bib4]; Kuenzel et\u00a0al., [@bib28]; Richards and Proszkowiec-Weglarz, [@bib49]). Pro-opiomelanocortin (**POMC**), Agouty-related protein (**AGRP**), and neuropeptide Y (**NPY**) are synthesized in the arcuate nucleus of the hypothalamus, where they play opposing roles in feed intake regulation in mammals (Ollmann et\u00a0al., [@bib42]; Kalra et\u00a0al., [@bib24]; Schwartz et\u00a0al., [@bib56]; Mizuno et\u00a0al., [@bib40]). Through its receptor, insulin contributes to nutrient sensing, regulation of energy expenditure, food intake control (Nandi et\u00a0al., [@bib41]), and regulation of liver glucose production (Inoue, [@bib21]). Foxo1 and Gpr17 have been recently proposed as the targets of the insulin receptor (**IR**) cascade in AGRP neurons of the arcuate nucleus to regulate peripheral metabolism and energy balance and insulin sensitivity (Ren et\u00a0al., [@bib46]; Ren et\u00a0al., [@bib45]; Scherer et\u00a0al., [@bib54]). In birds the infundibular nucleus plays a similar role to that of the mammalian arcuate nucleus (Wang et\u00a0al., [@bib67]).\n\nChickens exhibit several peculiarities in plasma glucose level control, glucose transporters, insulin release, and IR signaling (Hazelwood, [@bib18]; Rideau, [@bib50]; Simon et\u00a0al., [@bib60]; Dupont et\u00a0al., [@bib15]). Despite these peculiarities, a cross-talk exists in birds between insulin and the hypothalamus. Hyperphagia develops and glucose-stimulated insulin levels are enhanced following lesions of the ventromedial hypothalamus (**VMH**) in geese and chickens (Jaccoby et\u00a0al., [@bib22]). Intraventricular injection of insulin inhibits food intake in young chickens via the central melanocortin system (Shiraishi et\u00a0al., [@bib59]). Chicken brain IRs show specific glycosylation, enhanced affinity, and different tyrosine kinase activity in response to nutritional alterations vs. peripheral IRs (Simon and Leroith, [@bib62]). Moreover, IRs have been localized by immunohistochemistry in neurons of several structures of hypothalamus (the paraventricular nucleus, ventromedial and lateral regions, and infundibular nucleus) and in the infundibulum, they co-localize with \u03b1-melanocyte stimulating hormone and NPY neurons (Shiraishi et\u00a0al., [@bib58]). Central insulin injection modified hypothalamic levels of several mRNAs. In one study, *POMC*, cocaine- and amphetamine-regulated transcript (***CART***) and corticotropin-releasing hormone (***CRH***) mRNAs were up-regulated, whereas *NPY* or *AGRP* mRNAs were unaltered (Honda et\u00a0al., [@bib20]). In another study, *POMC* mRNA increased and *NPY* mRNA decreased (Shiraishi et\u00a0al., [@bib59]). In contrast, intraperitoneal insulin injection did not modify *NPY* mRNA levels in the hypothalamus of selected high or selected low weight chickens at 90 days of age (Zhang et\u00a0al., [@bib69]). The lack of insulin effect on hypothalamic *NPY* mRNA in these lines following intraperitoneal injection (vs. central injection in other studies) may rely on the time course for insulin access and/or age of the chickens. In another study however, intraperitoneal insulin injection increased hypothalamic tryptophan hydroxylase 2 mRNA in 5 day-old chickens of selected high or selected low weight chickens (Rice et\u00a0al., [@bib48]). Hypothalamic *NPY* mRNA content was higher in selected high weight chickens than in selected low weight chickens (Zhang et\u00a0al., [@bib69]), though at a younger age opposite results were observed (Rice et\u00a0al., [@bib48]). Changes in hypothalamic *NPY* mRNA have also been observed in response to changes in nutritional status. *NPY* mRNA levels increased following feed restriction (Boswell et\u00a0al., [@bib5],[@bib6]) or prolonged fasting, except in the lateral area (Zhou et\u00a0al., [@bib70]) and in day-old chickens unfed for 48\u00a0h (Higgins et\u00a0al., [@bib19]). In the latter model, a large number of genes (among which a large number of receptors) were differentially expressed in the hypothalamus between the various nutritional states (fed, unfed for 48\u00a0h, or refed).\n\nThe goal of the present study was to gain insight into insulin control of gene expression in the chicken hypothalamus, using 23 selected mRNAs and two experimental models. In the first model, insulin deficiency was induced in fed broiler chickens (16 or 17 days of age) by immune-insulin privation for 5\u00a0h. Insulin deprived chickens exhibited a large decrease in food intake, large increases in plasma glucose, non-esterified fatty acids (NEFA), amino acids (alpha-NH~2~-non protein), and glucagon, and a decrease in plasma T~3~. In 3 peripheral tissues of the same chickens used in the present study, transcriptome studies identified a rather large number of insulin-dependent genes: 1,573 in liver and 1,225 in muscle, but much less in abdominal adipose tissue following immune insulin privation (Ji et\u00a0al., [@bib23]; Simon et\u00a0al., [@bib63]). A short fasting period (5\u00a0h) was included as a second model, as fasting represents another status of \"insulin privation\". Plasma insulin levels were not, however, measured in this experiment because of the presence of anti-insulin antibody in the insulin-deprived group. As compared to fed control chickens, 5\u00a0h fasting slightly decreased plasma glucose (though not significantly) and amino acid levels, increased plasma NEFA and glucagon levels, and decreased plasma T~3~. Therefore, several changes at the plasma level are shared by these two experimental models. The 23 mRNAs assessed included most of those discussed earlier. Others were chosen mainly from our study on effects of prolonged fasting in day-old chickens (48\u00a0h without feed) in the hypothalamus (Higgins et\u00a0al., [@bib19]) or had been identified as differentially expressed in the hypothalamus during development of genetically selected fat or lean chickens (Byerly et\u00a0al., [@bib7]). Fat and lean chickens differ in their glucose-insulin relationship, with a slightly higher insulin sensitivity in fat chickens (Simon and Leclercq, [@bib61]). Finally, other selected mRNAs were identified as insulin-dependent in liver or muscle, after insulin privation (Dupont et\u00a0al., [@bib16]; Simon et\u00a0al., [@bib63]). As a whole, selected mRNAs are involved in feed intake regulation or neuronal, endocrine, or transcriptional control or glucose metabolism and glucose sensing.\n\nMATERIALS AND METHODS {#sec2}\n=====================\n\nAnimals and Experimental Protocol {#sec2-1}\n---------------------------------\n\nThe animals and experimental protocol used were described previously (Dupont et\u00a0al., [@bib16]). Briefly, male broiler chickens (ISA 915, Institute de Selection Animale, Saint Brieuc, France) were fed a conventional balanced diet ad libitum (3,050 kCal or 12.8 mJ/kg metabolizable energy, 22% proteins (N6.25), based on corn, wheat, peas, soybean meal, corn gluten, and rapeseed oil) and exposed to a 14L:10D light regimen. At 16 to 17 d of age, chickens of similar body weight were divided into five experimental groups (n = 7 birds each) as follows: fed groups that received a single intra-venous (**i.v.**) injection (1.5\u00a0mL/kg) of either normal guinea pig serum (**NS**, PromoCell, Heidelberg, Germany) or anti-porcine insulin guinea pig serum (**AI**) (**C_1** and **AI_1** groups, respectively), fed groups that received three i.v. injections (1.5\u00a0mL/kg each) at time 0, 2, and 4\u00a0h of either NS or AI (**C_5** and **AI_5** groups, respectively), and a fasted group that was fasted for 5\u00a0h and received three i.v. injections of NS at the times described above (**F_5** group). The 5\u00a0h time point was based on availability of antiserum. Insufficient anti-serum was available to assess longer time points, while still administering the antiserum at 2\u00a0h intervals. Four replicates were performed at 16 d and three at 17 d of age to allow for collection of all samples within an hour on each day. Immune sera were prepared as described previously (Simon et\u00a0al., [@bib60]). Food intake was measured 1\u00a0h (C_1 and AI_1) or 5\u00a0h (C_5 and AI_5) after the first i.v. injection. At the same time, blood was sampled under EDTA, cooled on ice, and the plasma rapidly collected by centrifugation. Birds were killed by cervical dislocation after blood sample collection. Hypothalamic tissue samples were collected manually using a dissecting microscope as described previously (Byerly et\u00a0al., [@bib8]; Higgins et\u00a0al., [@bib19]) based on published information (Kuenzel and Masson, [@bib31]) and stored at \u221280\u00b0C until RNA extraction. Initial incisions were made just anterior to the occulomotor nerve (*nervus occulomotorius*) and posterior to the *tuberculum olfactorium*. Next, lateral cuts were made 2\u00a0mm from the midline to yield a rectangular piece of tissue. This was placed on its side, and a final cut was made at a depth immediately below the subseptal organ (*organum subseptale*) and parallel to the basal surface of the hypothalamus. Plasma glucose levels were measured by the glucose oxidase method (Glucose Beckman Analyzer 2, Beckman, Palo Alto, CA). All procedures were approved by the Agricultural Agency and the Scientific Research Agency of INRA and conducted in accordance with the guidelines for Care and Use of Agricultural Animals in Agricultural Research and Teaching.\n\nRNA Isolation and Reverse Transcription-Quantitative PCR {#sec2-2}\n--------------------------------------------------------\n\nTotal RNA was extracted using RNeasy Midi kits (Qiagen, Valencia, CA), according to the manufacturer\\'s protocol and quantified using the RiboGreen Quantitation kit (Invitrogen, Carslbad, CA). Two-step reverse transcription-quantitative PCR (**RT-qPCR**) was performed to analyze hypothalamic mRNA levels. The RT reactions (20 \u03bcL) consisted of 1 \u03bcg of RNA, 50 units Superscript III reverse transcriptase (Invitrogen), 40 units of an RNase inhibitor (Invitrogen), 0.5\u00a0mM dNTPs, and 2.5 \u03bcM anchored oligo dT primers. A pool of all RNA (1 \u03bcg) from all treatment groups was used as a negative control for genomic DNA contamination and was processed as the other samples, but with omission of Superscript III enzyme. The RT reactions were diluted to 200 \u03bcL before being subjected to PCR. PCR was performed in 15 \u03bcL reactions containing 1 \u03bcL of diluted RT reaction, 400\u00a0nM of each gene-specific primer, 7.5 \u03bcL of 2 \u00d7 PCR buffer (100\u00a0mM KCl, 20\u00a0mM Tris-HCl, pH 9, 0.2% Triton X-100), 3.8\u00a0mM MgCl~2,~ 0.12\u00a0U/\u03bcL Taq polymerase, 400\u00a0nM dNTPs, 40\u00a0nM fluorescein (Invitrogen), and SYBR Green I Nucleic Acid Gel Stain (Invitrogen) diluted 1:10,000, and the MyIQ Single-Color Real-Time PCR Detection System (Bio-Rad, Hercules, CA). Thermal cycling parameters were: 1 cycle at 95\u00b0C for 5 minutes, followed by 40 cycles of 95\u00b0C for 15 seconds, 60\u00b0C for 30 seconds, and 72\u00b0C for 30 seconds. Dissociation curve analysis and gel electrophoresis were employed to ensure that a single PCR amplicon of appropriate size was amplified in each reaction. PCR products were sequenced to confirm their identities. Primer sequences, designed using Primer3 software (Rozen and Skaletsky, [@bib53]), are listed in Table [1](#tbl1){ref-type=\"table\"}. The data obtained were normalized to the housekeeping gene \u03b2-actin (**ACTB**), and the data were transformed using the equation 2^\u2212Ct^, where Ct represents the fractional cycle number when the amount of amplified product reached a fixed threshold for fluorescence. The data are presented as fold changes relative to the C_1 group for statistical analysis.\n\n###### \n\nGene-specific primers used for the analysis of mRNA levels using quantitative real-time RT-PCR.[^a^](#tb1fn1){ref-type=\"fn\"}\n\n  Gene Name[^b^](#tb1fn2){ref-type=\"fn\"}   GenBank Accession No.[^c^](#tb1fn3){ref-type=\"fn\"}   Forward primers (5'-3')     Reverse primer (5'-3')         PCR size (bp)\n  ---------------------------------------- ---------------------------------------------------- --------------------------- ------------------------------ ---------------\n  AGRP                                     NM_0,010,31457                                       AAGTCTGGCCTGGGAAGAG         CCCCCTGCAGAAGATGAG             122\n  ACTB                                     X0082                                                CAGGATGCAGAAGGAGATCACA      TAGAGCCTCCAATCCAGACAGAGTA      101\n  CHAT                                     NM_204,610                                           TGATGAGGGCAGAGTTGACA        TGCGTCCTTAAATCTCTGCAT          124\n  CRH                                      NM_0,011,23031                                       CACAGCAACAGGAAACTGATGGAAA   AAAGAGGTGACATCAGAGCAGCACTATG   101\n  DIO2                                     NM_204,114                                           ACTGTTTGAGGGCGCTAAACC       AAACACTAGCCCTCCAGAATACCTT      110\n  DPYSL2                                   NM_204,494                                           CTGCCAAGACCCACAACATA        TTCCCTTGGCTGATAACCAC           91\n  EGR1                                     NM_204,136                                           TCAGCACTTTCAGACATGACATCA    AGTACCAGTGGAAGAGGTGAATGC       101\n  ENO2                                     NM_204,876                                           GCCTCCTGCTCAAAGTCAAC        CATTCTCCTGGGCCAACTTA           75\n  GCK                                      AF525739                                             ACGAAGCACCAGATGTACTCC       GGAGATGCACTCCGAGATGT           86\n  GLUT1                                    NM_205,209                                           GGGATCAATGCGGTTTTCTA        CGAAGAGCGAAACAACAGTG           124\n  GLUT2                                    NM_207,178                                           TATCGTCACAGGCATCCTCA        GAAGAACTGCAGCAGAGCAG           118\n  GLUT3                                    NM_205,511                                           CTCAACCAGCTGGGCATAGT        AGTGGCCAAAGTGCTTCAGT           89\n  GLUT8                                    NM_204,375                                           TGCAGGCAAATTGAAGCTAA        TGCCTGTTACTTGCTGGAGA           124\n  GPI                                      NM_0,010,06128                                       CCATGCTGGGAGTGTGGTAT        GCTGGAAGTAGGCAGCAAAG           99\n  HK1                                      NM_204,101                                           CGTTACCTGCATTTCGGACT        CTTGCAAGGGAAGGAGAATG           88\n  LEP                                      LN794246[^d^](#tb1fn4){ref-type=\"fn\"}                GGCGATTCCAACGCCTACT         CCGCCATGGCTTGCA                102\n  LEPR                                     NM_204,323                                           TTCCAAACCCCAAGAATTGCT       CAAATGACATTGCTTCAGGGTG         102\n  MC4R                                     NM_0,010,31514                                       CGGGAGGCTGCTATGAACAA        AGCTGATGATGCCCAGAGTCA          101\n  NPY                                      NM_205,473                                           GGGAAAGCACAGAAAACATTCC      AAATCCCATCACCACATCGAA          101\n  PGAM1                                    NM_0,010,31556                                       AGGCTCAGGTGAAGATCTGG        TGCTGATGGTGCTGAAGAAG           87\n  POMC                                     NM_0,010,31098                                       CGCTACGGCGGCTTCA            TCTTGTAGGCGCTTTTGACGAT         88\n  TAS1R1                                   XM_425,734                                           GCGTGGCAGAAACACCAGAT        GCTTGACCAATATGCGCTTGT          64\n  TH                                       NM_204,805                                           CGGACTGCTGTCATGAGCTA        CAGTTGCTCCCAGAGATGC            102\n  TRH                                      NM_0,010,30383                                       TGGATGACATCCTGCAGAGATC      GGAAAGCCATTGTGGCAGA            116\n\n^a^All primers used for expression analysis were designed using the Primer3 program (<http://fakker.wi.mit.edu/primer3/input.htm;> Rozen and Skaletsky, [@bib53]).\n\n^b^Abbreviation of the gene names are defined in Table [2](#tbl2){ref-type=\"table\"} and text.\n\n^c^Reference chicken gene sequences that contain the corresponding PCR products list.\n\n^d^From Seroussi et\u00a0al. ([@bib57]).\n\nStatistical Analysis {#sec2-3}\n--------------------\n\nAll data were analyzed by one-way ANOVA using the general linear models (**GLM**) procedure of the Statistical Analysis System (**SAS**) software (SAS Institute, Cary, NC, USA). The treatment means for food intake and plasma glucose level were compared by Fisher\\'s test or Student *t* test, while the Duncan\\'s multiple range test option of the GLM procedure for SAS was used to determine significance of mean differences for gene expression analysis. Significance was set at *P*\u00a0\\<\u00a00.05.\n\nRESULTS {#sec3}\n=======\n\nInsulin immune-neutralization for 1 (AI_1) or 5 h (AI_5) significantly (*P* \\< 0.008 and *P* \\< 0.0001, respectively) decreased feed intake in comparison to corresponding control groups (Figure [1](#fig1){ref-type=\"fig\"}). Plasma glucose levels were significantly increased after insulin immune-neutralization at both time points, while chicks fasted for 5\u00a0h had similar (*P* \\> 0.05) plasma glucose levels as chicks fed *ad libitum* (Figure [1](#fig1){ref-type=\"fig\"}). These findings indicate that AI depressed insulin activity which led to increased glucose levels and decreased feed intake. Despite the large and multiple effects observed following insulin immuno-neutralization at the level of plasma parameters, liver or muscle IR signaling, and liver or muscle transcriptome (Dupont et\u00a0al., [@bib16]; Simon et\u00a0al., [@bib63]), levels of none of the selected mRNAs were modified in the hypothalamus at either 1 h or 5 h insulin privation, as compared to the respective fed controls (Table [2](#tbl2){ref-type=\"table\"}). *EGR1* mRNA, which decreased both in liver and leg muscle after insulin-neutralization (Dupont et\u00a0al., [@bib16]; Simon et\u00a0al., [@bib63]), was lower in the hypothalamus in the 5 h insulin neutralization group compared to the 5 h fasted group but not to the 5 h fed group. A similar situation was observed for *CRH* mRNA (Table [2](#tbl2){ref-type=\"table\"}). Origin of the opposite changes in *EGR1* and *CRH* in response to insulin-neutralization and fasting are unknown.\n\n![Effects of insulin immune-neutralization and fasting on (**A**) cumulative feed intake and (**B**) plasma glucose levels. At 16 to 17 d of age, chickens at similar body weight were divided into five experimental groups (n = 7 birds each) as follows: fed groups that received a single i.v. injection (1.5\u00a0mL/kg) of either normal guinea pig serum (NS, PromoCell, Heidelberg, Germany) or anti-porcine insulin guinea pig serum (AI) (C_1 and AI_1 groups, respectively), fed groups that received three i.v. injections (1.5\u00a0mL/kg each) at time 0, 2, and 4\u00a0h of either NS or AIS (C_5 and AI_5 groups, respectively), and a fasted group that was fasted for 5\u00a0h and received three i.v. injections of NS at the times described above (F_5 group). Each value represents the mean \u00b1 SE of seven birds (n = 7). Due to the large differences in feed intake between 1\u00a0h (C_1 and AI_1) and 5\u00a0h (C_5 and AI_5) after the first i.v. injection, results for feed intake at 1\u00a0h and 5\u00a0h were analyzed separately in order to test for differences due specifically to AI. Different letters denote statistically significant differences for mean comparisons (*P* \\< 0.05).](pex247fig1){#fig1}\n\n###### \n\nEffect of insulin immuno-neutralization and fasting on mRNA expression levels in the hypothalamus of chickens.\n\n  Gene Name   Treatment                                                                        \n  ----------- --------------------- ------------------- -------------------- ----------------- -------------------\n  POMC        100.00 \u00b1 12.28        97.57 \u00b1 16.44       86.76 \u00b1 17.87        92.48 \u00b1 22.36     114.46 \u00b1 30.34\n  MC4R        100.00 \u00b1 7.27         90.11 \u00b1 7.74        93.76 \u00b1 4.60         85.92 \u00b1 7.79      107.97 \u00b1 10.46\n  EGR1        100.00 \u00b1 5.98^a^      97.68 \u00b1 8.78^a^     92.64 \u00b1 4.92^a,b^    76.07 \u00b1 5.13^b^   102.96 \u00b1 9.26^a^\n  CRH         100.00 \u00b1 15.19^a,b^   82.32 \u00b1 7.95^a,b^   84.01 \u00b1 12.21^a,b^   74.38 \u00b1 9.60^b^   116.17 \u00b1 18.16^a^\n  TRH         100.00 \u00b1 6.80         95.50 \u00b1 8.08        94.39 \u00b1 4.77         86.01 \u00b1 7.76      108.61 \u00b1 10.47\n  ENO2        100.00 \u00b1 4.97         99.28 \u00b1 1.88        98.05 \u00b1 1.82         97.07 \u00b1 7.60      108.34 \u00b1 7.25\n  PGAM1       100.00 \u00b1 11.84        83.22 \u00b1 3.40        80.06 \u00b1 6.88         85.88 \u00b1 10.78     92.19 \u00b1 14.14\n  GPI         100.00 \u00b1 4.55         98.51 \u00b1 3.44        99.29 \u00b1 1.48         94.82 \u00b1 8.30      109.99 \u00b1 7.47\n  DPYSL2      100.00 \u00b1 5.60         98.65 \u00b1 5.27        97.12 \u00b1 2.94         96.52 \u00b1 7.45      104.13 \u00b1 9.36\n  HK1         100.00 \u00b1 4.92         97.63 \u00b1 2.69        102.17 \u00b1 4.67        96.79 \u00b1 6.78      102.50 \u00b1 6.02\n\nReverse transcription quantitative PCR (RT-qPCR) was used to quantify mRNA levels for pro-opiomelanocortin (*POMC*), melanocortin 4 receptor (*MC4R*), early response gene (*EGR1*), corticotropin-releasing hormone (*CRH*), thyrotropin-releasing hormone (*TRH*), neural enolase 2 (*ENO2*), phosphoglycerate mutase 1 (*PGAM1*), glucose phosphate isomerase (*GPI*), dihydropyrimidinase like protein (*PRYSL2*), and hexokinase 1 (*HK1*) in hypothalami collected from ad libitum fed chickens 1\u00a0h after one normal serum (NS) injection (C_1) or one anti-insulin serum (AI) injection (AI_1) and after 5\u00a0h during which three injection of NS (C_5) or AI (AI_5) were administered. Birds fasted for 5\u00a0h (F_5) were also analyzed for comparison. Each value represents the mean \u00b1 SE of seven birds (n = 7).\n\n^a,b^Different letters denote statistically significant differences for mean comparisons (*P* \\< 0.05).\n\nIn contrast, 5\u00a0h fasting altered the levels of several mRNAs. *NPY*, sweet taste receptor 1 (*TAS1R1*), and iodothyronine deiodinase 2 (*DIO2*) mRNA levels increased as compared to both fed controls or 5 h insulin deprived fed chickens (Figure [2](#fig2){ref-type=\"fig\"}), and *LEPR* mRNA levels were increased by 5\u00a0h fasting as compared to fed controls (Figure [3](#fig3){ref-type=\"fig\"}). *LEP* messenger levels were unchanged (Figure [3](#fig3){ref-type=\"fig\"}). Fasting also increased mRNAs for several glucose transporters (*GLUT1*, *GLUT3*, and *GLUT8*) and *GCK*, also as compared to both fed controls or 5 h insulin deprived chickens (Figure [4](#fig4){ref-type=\"fig\"}). Choline O-acetyltransferase (*CHAT)* and tyrosine hydroxylase (*TH*) mRNA levels were unchanged (Figure [5](#fig5){ref-type=\"fig\"}), suggesting a lack of activation of cholinergic and dopaminergic systems in the hypothalamus, at the mRNA level, in any of the present conditions. Finally, levels of eight other mRNAs were unaffected by any of the treatments (see Table [2](#tbl2){ref-type=\"table\"}).\n\n![Effects of insulin immune-neutralization and fasting on hypothalamic levels of mRNA for (**A**) neuropeptide Y (NPY), (**B**) agouty gene related protein (AGRP), (**C**) sweet taste receptor 1 (TAS1R1), and (**D**) iodothyronine deiodinase 2 (DIO2). See legend to Figure [1](#fig1){ref-type=\"fig\"} for further details.](pex247fig2){#fig2}\n\n![Effects of insulin immune-neutralization and fasting on hypothalamic levels of mRNA for (**A**) leptin (LEP) and (**B**) leptin receptor (LEPR). See legend to Figure [1](#fig1){ref-type=\"fig\"} for further details.](pex247fig3){#fig3}\n\n![Effects of insulin immune-neutralization and fasting on hypothalamic levels of mRNA for genes involved in glucose transport and metabolism. (**A**) PCR products for glucose transporters 1, 2, 3, and 8 (GLUT1, GLUT 2, GLUT3, and GLUT8, respectively). Hypothalamic levels of mRNA for (**B**) GLUT1, (**C**) GLUT2, (**D**) GLUT3, and (**F**) GCK were determined by reverse transcription-quantitative PCR. (**E**) PCR products for glucokinase (GCK). Numbers 1 to 6 in panels A and E refer to different tissues. Samples are identified as follows: 1, C_5; 2, AI_5; 3, no RT sample; 4, brain sample from 3 wk old chicken; 5 liver sample from embryonic day 18 and 6, blank. See legend to Figure [1](#fig1){ref-type=\"fig\"} for further details.](pex247fig4){#fig4}\n\n![Effects of insulin immune-neutralization and fasting on hypothalamic levels of mRNA for (**A**) choline O-acetyltransferase (*CHAT)* and (**B**) tyrosine hydroxylase (*TH*). See legend to Figure [1](#fig1){ref-type=\"fig\"} for further details.](pex247fig5){#fig5}\n\nDISCUSSION {#sec4}\n==========\n\nIn the present study, 5 h insulin immune neutralization in fed chickens had no effect on hypothalamic levels of the mRNAs investigated (only *EGR1* tended to decrease). This was highly surprising, considering the inhibition of food intake, changes in levels of numerous mRNAs in peripheral tissues, known effects of central insulin injection on the chicken hypothalamus, and the multiple and major effects of insulin in the mammalian hypothalamus. Though, to our knowledge, insulin has not been measured in chicken cerebrospinal fluid (**csf**), it would be very unlikely if both csf insulin and glucose levels had not been altered during the 5\u00a0h insulin neutralization, as the half-life of plasma insulin is about 5\u00a0min in chickens and about 7\u00a0min in turkeys (Langslow, [@bib32]; McMurtry et\u00a0al., [@bib38]). Uptake of plasma glucose and insulin through blood-brain barrier and relationships between plasma glucose and insulin levels and csf glucose and insulin levels have been investigated in mammals. Changes in csf are tempered in comparison with peripheral variations of glucose or insulin, but do exist, showing some delay and variations of lower magnitude. The adjustment of csf insulin is disturbed in some pathologies (Begg, [@bib2]). In contrast, changes in csf glucose level have been observed in chickens in various conditions: a decrease in response to peripheral hypoglycemia (induced by intravenous tolbutamide or insulin injection) with a delay of about 20 to 40 minutes (Anderson and Hazelwood, [@bib1]; Simon and Leclercq, [@bib61]) or in response to overnight fasting of 16\u00a0h (Simon and Leclercq, [@bib61]). Since no changes were observed for any of hypothalamic mRNAs presently investigated despite the rather extreme experimental conditions, the inhibition of food intake following insulin neutralization must rely on mechanisms other than control at the transcription level in the hypothalamus within the period of time investigated. Effects on neurotransmitter release within the hypothalamus are a likely possibility. A change in glucose sensing or metabolism and/or a decrease in insulin signaling is another possibility. Up to now, major steps of the insulin-IR cascade have not been characterized in the chicken brain. Only IR tyrosine kinase activity has been measured and shown to be refractory to extreme nutritional changes (Simon and Leroith, [@bib62]). In peripheral tissues, the early steps of insulin receptor signaling appear to be responsive to insulin in the liver but refractory in muscle and adipose tissue (Dupont et\u00a0al., [@bib14]; Dupont et\u00a0al., [@bib15]). So far, only the phosphorylation of Src homology 2 domain containing transforming protein 1 adaptor protein (**Shc**), an IR substrate, appeared insulin sensitive in muscle (Dupont et\u00a0al., [@bib14]).\n\nPlasma glucose levels were not significantly different between 5\u00a0h fasting and fed control chickens. A catabolic state had likely already developed during the 5\u00a0h fast, as suggested by the significant increases in plasma NEFA and glucagon levels and significant decrease in amino acid index reported previously (Dupont et\u00a0al., [@bib16]). The 5\u00a0h fast certainly represents a transition step during which a hunger state develops and orexigenic mechanisms are stimulated. Though some orexigenic gene mRNA levels did not change in the 5\u00a0h fasting group (e.g., *POMC*, *MC4R*), expression of 8 genes was altered out of the 23 measured. These genes are involved in the control of food intake, endocrine control of metabolism, or glucose metabolism.\n\nHypothalamic *NPY* and *DIO2* mRNAs were increased by 5\u00a0h fasting. Similar increases were previously observed for *NPY* and *DIO2* mRNA following fasting in growing chickens or in 48\u00a0h non-fed, day-old chickens (Boswell et\u00a0al., [@bib5],[@bib6]; Zhou et\u00a0al., [@bib70]; Higgins et\u00a0al., [@bib19]). In liver, *DIO2* mRNA was, in contrast, down-regulated by both 5\u00a0h fasting and 5\u00a0h insulin deprivation (Dupont et\u00a0al., [@bib16]). Expression of two other genes of the endocrine system was changed in the hypothalamus following fasting. *CRH* (as compared to insulin-immuno-neutralization) and *LEPR* (as compared to fed controls) mRNA levels were increased following fasting. CRH has been previously shown to decrease feed intake in chickens (Denbow et\u00a0al., [@bib12]), and differential CRH signaling may influence appetite regulatory mechanisms in chickens with different body weight and feed consumption (Cline et\u00a0al., [@bib10]). Moreover, CRH acts as a mediator for a number of anorexic peptides in chickens (Honda et\u00a0al., [@bib20]; Tachibana et\u00a0al., [@bib65]; Tachibana et\u00a0al., [@bib66]). An increase in *CRH* mRNA was previously observed when insulin was administered in chickens (Honda et\u00a0al., [@bib20]). Alternatively, the increase in *CRH* mRNA levels following fasting might be a result of fasting-induced stress. The increase in *LEPR* mRNA is consistent with our previous results (Higgins et\u00a0al., [@bib19]). The chicken *LEP* gene, which has been identified only very recently, is expressed in the hypothalamus (Seroussi et\u00a0al., [@bib57]). In the present study, hypothalamic *LEP* mRNA levels were unchanged by fasting.\n\n*TAS1R1*, which has been proposed in mice as a hypothalamic sensor of glucose (Ren et\u00a0al., [@bib47]), is expressed at a higher level in chickens selected for increased body fat than in their counterparts selected for leanness (Byerly et\u00a0al., [@bib7]). In the present study, hypothalamic *TAS1R1* showed no changes when glucose levels largely increased following insulin neutralization but did increase following the 5 h fasting period, most likely in response to transient alterations in plasma glucose levels. Since glucose is the energetic fuel in brain, the increase in hypothalamic *TAS1R1* mRNA may signal a quick adaptation to transient decreases in glucose levels during fasting. A similar adaptation is also suggested by the changes observed at the level of mRNAs coding for *GLUT1, GLUT3, GLUT8*, and *GCK*.\n\nIn mammals, a large family of glucose transporters facilitates cell glucose uptake. Among these, GLUT4, which largely accounts for insulin sensitive glucose uptake in muscle and adipocytes, is also present in the hypothalamus, where it appears to play a critical role in glucose uptake (Ren et\u00a0al., [@bib45]). The *GLUT4* gene is not present in the yet incompletely sequenced chicken genome. Recently, GLUT12 appeared to act as an insulin-sensitive glucose transporter in chicken muscles and may play a similar role in this species as GLUT4 in mammals (Coudert et\u00a0al., [@bib11]). To our knowledge, the presence of GLUT12 in chicken hypothalamus has not been shown. Herein, other GLUT mRNAs were investigated. GLUT1 is expressed ubiquitously and facilitates basal glucose uptake. GLUT2 has a low affinity to glucose, and its expression is limited to the basolateral membrane of hepatocytes, kidney, small intestine, and pancreatic beta-cells. In mammals GLUT2 is also expressed in the hypothalamus, contributing to the control of food intake (Stolarczyk et\u00a0al., [@bib64]). GLUT3 is a high affinity transporter responsible for glucose uptake in tissues and cells where extracellular substrate concentration is low (e.g., neuronal cells) (Kono et\u00a0al., [@bib27]), while GLUT8 has been shown to be insulin responsive in the blastocyst of mammals (Carayannopoulos et\u00a0al., [@bib9]). The presence of *GLUT1*, *GLUT3*, and *GLUT8* mRNAs was confirmed by PCR in brain and liver tissues of chickens. In contrast, mRNA expression of *GLUT2* was not detectable in the chicken hypothalamus in the present study (data not shown). Similar expression patterns were observed by Kono et\u00a0al. ([@bib27]) in chicken tissues. The 5 h fasting treatment increased hypothalamic mRNA levels for *GLUT1*, *GLUT3*, and *GLUT8* (Figure [4](#fig4){ref-type=\"fig\"}), with *GLUT1* demonstrating the largest effect.\n\nMammalian hexokinase IV (glucokinase, **GCK**) plays a crucial role in glucose homeostasis (Postic et\u00a0al., [@bib44]; Matschinsky, [@bib37]) by phosphorylating hexose into hexose-6-phosphate in the first step of glycolysis. GCK is predominantly located in pancreatic beta-cells and in the liver, but also is present in glucose-sensitive tissues including the hypothalamus (Magnuson, [@bib36]; Schuit et\u00a0al., [@bib55]; Penicaud et\u00a0al., [@bib43]). Within the hypothalamus of mammals, GCK has been shown to be expressed in the arcuate nucleus, lateral nucleus, dorsomedial nucleus, ventromedial nucleus, and paraventricular nucleus (Lynch et\u00a0al., [@bib34]; Maekawa et\u00a0al., [@bib35]; Li et\u00a0al., [@bib33]). Even though the chicken *GCK* gene has not been identified in the chicken genome, a 1,326\u00a0bp liver cDNA fragment that shows high homology with human *GCK* cDNA has been sequenced (Berradi et\u00a0al., [@bib3]). Chicken *GCK* has been shown to be insulin dependent in fed chickens (Dupont et\u00a0al., [@bib16]), *GCK* mRNA increased in liver in response to a meal (Rideau et\u00a0al., [@bib51]) and protein levels decreased during fasting (Dupont et\u00a0al., [@bib16]; Rideau et\u00a0al., [@bib51]). Furthermore, the use of a specific activator of mammalian GCK supported the existence a functional and potent GCK in chicken. This glucokinase activator induced hypoglycemia and inhibited food intake without recruiting insulin, suggesting that liver glucokinase activation by itself accounted for the hypoglycemia (Rideau et\u00a0al., [@bib52]). We detected *GCK* mRNA in the chicken hypothalamus (Figure [4](#fig4){ref-type=\"fig\"}E). To our knowledge, this is the first report showing the presence of the mRNA for this enzyme in the avian hypothalamus. Previously, Berradi et\u00a0al. ([@bib3]) were not able to detect chicken GCK in chicken brain at the protein level. It is, however, possible that the protein level of GCK in whole brain is too low to be detected. In mammals, GCK is recognized as a specific marker for glucose sensing neurons (Yang et\u00a0al., [@bib68]; Dunn-Meynell et\u00a0al., [@bib13]; Kang et\u00a0al., [@bib26], [@bib25]).\n\nIn conclusion, 5\u00a0h insulin immuno-neutralization in fed chickens was without any effect on the various mRNAs presently investigated in the hypothalamus, despite an early decrease in food intake. A change in glucose and/or insulin signaling are likely to be involved in this \"satiety\" mechanism. The fact that all the mRNAs studied remained at a \"normal\" fed level is largely unexpected. By contrast, 5\u00a0h of fasting altered mRNA levels for 8 genes involved in food intake, glucose transport, and glucose metabolism. This short period of fasting was clearly perceived as a signal of the development of a glucose shortage event. Therefore, these \"fast responder\" genes are likely components of a mechanism of emergency adaptation to food deprivation in the chicken.\n"}
+{"text": "Introduction {#sec1-0300060520910891}\n============\n\nBreast tuberculosis (BTB) was first reported by Cooper in 1829. Breast tissue appears to be resistant to tuberculosis (TB) infection because of its unfavorable environment for the survival and proliferation of tuberculosis bacteria.^[@bibr1-0300060520910891],[@bibr2-0300060520910891]^ The incidence of BTB is thus low, ranging from 0.1% to 0.52%.^[@bibr3-0300060520910891]^ However, the incidence has been gradually increasing as a result of increasing numbers of multidrug-resistant tuberculosis bacteria, increasing numbers of patients with immunocompromised disorders such as acquired immunodeficiency syndrome, especially in developing countries, and increasing immigration from regions where TB is prevalent.^[@bibr3-0300060520910891][@bibr4-0300060520910891][@bibr5-0300060520910891][@bibr6-0300060520910891][@bibr7-0300060520910891]--[@bibr8-0300060520910891]^ BTB is a rare form of extrapulmonary TB, accounting for approximately 0.1% of all breast diseases and 3% to 4.5% of surgically-treated breast disorders in developing countries.^[@bibr9-0300060520910891]^ The clinical manifestations of BTB are complex and unspecific, making it difficult to differentiate from other breast disorders such as breast cancer and breast sclerosing adenosis.^[@bibr1-0300060520910891],[@bibr10-0300060520910891]^ Ultrasound (US) imaging is crucial for identifying breast lesions, and characterizing the US imaging features of BTB lesions may thus facilitate its diagnosis. However, few studies to date have reported on the use of US to characterize BTB lesions. In the present study, we retrospectively characterized the US imaging features of 45 patients with BTB to further our understanding of the disease progression and to provide valuable information for its clinical diagnosis. To the best of our knowledge, this is the largest study of US applied in patients with BTB to date.\n\nPatients and methods {#sec2-0300060520910891}\n====================\n\nPatients {#sec3-0300060520910891}\n--------\n\nWe retrospectively reviewed patients diagnosed with BTB who underwent US examinations from January 2010 to December 2017 at Zhejiang Integrated Traditional and Western Medicine Hospital, China. This study was approved by the institutional review board of Zhejiang Integrated Traditional and Western Medicine Hospital, China and all patients provided written informed consent.\n\n### Ultrasound examination {#sec4-0300060520910891}\n\nPatients were examined using a Philips IU22\u00ae Ultrasound instrument (Philips, Amsterdam, The Netherlands) with a L12-5 probe (corresponding frequency 5.0--12.0\u2009MHz) or an Acuson Sequoia 512 instrument (Siemens Medical Solutions, Mountain View, CA, USA) with a 15L8W probe (corresponding frequency 8.0--15.0 MHz). The pulse repetition frequency was 750\u2009Hz, filter M, and gain 250. Sonography-guided core-needle biopsy and fine-needle aspiration cytology were performed using a BARD\u00ae MAGNUM\u00ae instrument with 18G and 20G core needles, respectively (Bard Peripheral Vascular, Inc., Tempe, AZ, USA). To fully expose the bilateral breasts during US examination, patients were kept in a supine or lateral position with their hands placed symmetrically beside a pillow. Enclosed breast masses, nipple discharge, or skin sinus tracts were identified by the naked eye. All patients underwent routine US examination of bilateral breasts and axillary lymph nodes at least twice, in a radial pattern. We performed biopsy via an 18G biopsy needle for nodular BTB lesions, and aspiration for GeneXpert examination (Cepheid, Sunnyvale, CA) via a 20G needle prior to biopsy for lesions with cystic components. The findings of the US examinations were retrospectively reviewed by two radiologists with at least 5 years of experience. The US features of the lesions, including position, shape, margin, orientation, echogenicity pattern, posterior acoustic features, calcification, and vascularity, were analyzed and recorded.\n\n### Blood supply examination {#sec5-0300060520910891}\n\nThe blood supply and blood distribution of the lesions were evaluated using color Doppler flow imaging (CDFI). According to the Adler flow categorization,^[@bibr11-0300060520910891]^ the blood flow of the mass was classified into four categories: 0, avascular; I, one or two pixels containing blood flow; II, one main vessel in the area and/or several small vessels; and III, \u22654 vessels or diffuse reticular flow. In addition, alterations of axillary lymph nodes were also evaluated by US imaging.\n\n### Data analysis {#sec6-0300060520910891}\n\nAll data were presented as mean\u2009\u00b1\u2009standard deviation.\n\nResults {#sec7-0300060520910891}\n=======\n\nWe retrospectively reviewed the breast lesions in 45 female patients diagnosed with BTB. The patients' demographic and clinical data are summarized in [Table 1](#table1-0300060520910891){ref-type=\"table\"}. Of these 45 patients, 17 were diagnosed pathologically, eight by GeneXpert examination, and the other 20 by both pathology and GeneXpert examination. Twenty-six of the 45 cases had palpable masses, 13 had breast skin thickening and ulceration, two had breast pain, and five had fatigue, low fever, and night sweats. Twenty-three patients also had pulmonary TB and two had pelvic TB. In addition, 10 had a history of pulmonary TB, three had a history of lymphatic TB, two had a history of BTB, and the other five had no history of either TB or TB involving other organs.\n\n###### \n\nDemographic and clinical data of patients with breast tuberculosis (n\u2009=\u200945).\n\n![](10.1177_0300060520910891-table1)\n\n  Parameter                   Value\n  --------------------------- ---------------\n  Age range (mean), years     21--56 (32.4)\n  Marital status              \n  \u2003Single                     8\n  \u2003Married                    37\n  Breast-feeding history      \n  \u2003Yes                        30\n  \u2003No                         15\n  Involved breast             \n  \u2003Left                       27\n  \u2003Right                      18\n  Signs in the breast         \n  \u2003Mass                       26\n  \u2003Pain                       2\n  \u2003Sinus                      13\n  \u2003Asymptomatic               4\n  Associated disease          \n  \u2003Axillary lymphadenopathy   3\n  \u2003Pulmonary TB               23\n  \u2003Pelvic TB                  2\n\nTB, tuberculosis.\n\nAll 45 cases had unilateral BTB lesions ([Table 1](#table1-0300060520910891){ref-type=\"table\"}). The locations and extents of the lesions are given in [Table 2](#table2-0300060520910891){ref-type=\"table\"}. The maximal diameters of the BTB masses ranged from 0.5 to 3.8\u2009cm (mean 1.5\u2009\u00b1\u20090.7\u2009cm). Three patients (6.7%, 3/45) had enlarged multiple ipsilateral axillary lymph nodes with a maximum short-axis diameter \\>0.6\u2009cm. These involved lymph nodes were further proven to be lymphatic TB nodes via biopsy.\n\n###### \n\nUltrasound features of breast tuberculosis masses (n\u2009=\u200945).\n\n![](10.1177_0300060520910891-table2)\n\n                                Number   \\%\n  ----------------------------- -------- ------\n  Position                               \n  \u2003Upper outer quadrant         21       46.7\n  \u2003Lower outer quadrant         7        15.5\n  \u2003Upper inner quadrant         6        13.3\n  \u2003Lower inner quadrant         3        6.7\n  \u2003Underneath the nipple        8        17.8\n  Size (maximal diameter)                \n  \u2003\\<1.0\u2009cm                     2        4.4\n  \u20031.0--3.0\u2009cm                  36       80.0\n  \u2003\\>3.0\u2009cm                     7        15.6\n  Shape                                  \n  \u2003Round/oval                   26       57.8\n  \u2003Irregular                    19       42.2\n  Orientation                            \n  \u2003Parallel                     40       88.9\n  \u2003Non-parallel                 5        11.1\n  Margin                                 \n  \u2003Indistinct                   15       33.3\n  \u2003Distinct                     30       66.7\n  Echogenicity                           \n  \u2003Complex echoic               31       68.9\n  \u2003Hypoechoic                   14       31.1\n  Calcifications                         \n  \u2003Yes                          8        17.8\n  \u2003No                           37       82.2\n  Posterior acoustic feature             \n  \u2003No change                    32       71.1\n  \u2003Shadowing                    3        6.7\n  \u2003Enhancement                  10       22.2\n  Blood flow (Adler category)            \n  \u20030                            8        17.8\n  \u2003I                            36       80.0\n  \u2003II                           1        2.2\n  \u2003III                          0        0\n\nThe 45 BTB lesions were classified into one of three types based on their US imaging features. Type I, nodular type ([Figures 1](#fig1-0300060520910891){ref-type=\"fig\"}[](#fig2-0300060520910891){ref-type=\"fig\"}[](#fig3-0300060520910891){ref-type=\"fig\"}[](#fig4-0300060520910891){ref-type=\"fig\"}--[5](#fig5-0300060520910891){ref-type=\"fig\"}): 25 cases (55.5%, 25/45) with internal heterogeneous hypoechogenicity (11/25) or complex echogenicity (14/25), including 18 oval, two round, and five irregular lesions; 16 of the 25 cases (64.0%, 16/25) had a disease course of \u22646 months and nine (36.0%, 9/25) had a disease course \\>6 months. Type II, abscess type ([Figure 6](#fig6-0300060520910891){ref-type=\"fig\"}): 7 cases (15.6%, 7/45) with thick-wall cysts with oval shapes and posterior acoustic enhancement, including two (28.6%, 2/7) with internal punctate hyperechogenicity that were movable under probe pressure; five of the seven cases (71.4%, 5/7) had a disease course of \u22646 months and two (28.6%, 2/7) had a disease course \\>6 months. Type III, sinus type ([Figures 7,8](#fig7-0300060520910891){ref-type=\"fig\"}): 13 cases (28.9%, 13/45) with internal heterogeneous echogenicity, including 11 extending in a parallel pattern and two in a non-parallel pattern, all irregular with a cord-like hypoecho extending to the surface of the breast skin; three (23.1%, 3/13) cases had a disease course of \u22646 months and 10 (76.9%, 10/13) had a disease course \\>6 months. In addition, eight of the 45 cases (17.8%, 8/45) exhibited calcifications, including six macrocalcifications and two microcalcifications. In one case, the calcification increased from 0.4 to 0.8\u2009cm after 3 months.\n\n![Sonographic appearance of a 43-year-old woman with nodular breast tuberculosis. (a) Ultrasound demonstrated a well-defined oval nodule, with internal anechoic area (arrow) and posterior acoustic enhancement. (b) Color Doppler flow imaging demonstrated two pixels containing blood flow in the margin of the nodule (arrows), with blood flow classified as category I based on the Adler categorization.](10.1177_0300060520910891-fig1){#fig1-0300060520910891}\n\n![Sonographic appearance of a 28-year-old woman with nodular breast tuberculosis. (a) Ultrasound imaging of the right breast demonstrated an ill-defined irregular 'starfish'-like nodule in the upper inner quadrant, with a large internal anechoic region and posterior acoustic enhancement (arrow). (b) Color Doppler flow imaging demonstrated one pixel containing blood flow in the margin of the lesion (arrow), with blood flow classified as category I based on the Adler categorization.](10.1177_0300060520910891-fig2){#fig2-0300060520910891}\n\n![Sonographic appearance of a 32-year-old woman with nodular breast tuberculosis. Ultrasound imaging of the left breast demonstrated a hypoechoic nodule extending in a parallel orientation in the upper inner quadrant, with a circumscribed margin and internal heterogeneous echo. A punctate hyperechoic microcalcification was detected inside the nodule (arrow).](10.1177_0300060520910891-fig3){#fig3-0300060520910891}\n\n![Sonographic appearance of a 36-year-old woman with nodular breast tuberculosis. Ultrasound imaging of the left breast demonstrated a round ill-defined hypoechoic nodule extending in a non-parallel orientation in the lower outer quadrant. A hyperechoic macrocalcification was detected inside the nodule (arrow), with 1-mm diameter and posterior acoustic shadowing.](10.1177_0300060520910891-fig4){#fig4-0300060520910891}\n\n![Sonographic appearance of a 44-year-old woman with nodular breast tuberculosis. Ultrasound imaging of the right breast demonstrated an irregular hypoechoic nodule extending in a non-parallel orientation in the lower outer quadrant, with an angular margin (arrows).](10.1177_0300060520910891-fig5){#fig5-0300060520910891}\n\n![Sonographic appearance of a 28-year-old woman with abscess breast tuberculosis. Ultrasound imaging of the right breast demonstrated an ill-defined oval nodule, with posterior acoustic enhancement and disseminated internal punctate hyperecho (arrow), which move with probe pressure.](10.1177_0300060520910891-fig6){#fig6-0300060520910891}\n\n![Sonographic appearance of a 30-year-old woman with sinus breast tuberculosis. Ultrasound imaging demonstrated a complex mass mainly composed of cystic components, with posterior acoustic enhancement extending to the breast skin and forming a sinus (arrow).](10.1177_0300060520910891-fig7){#fig7-0300060520910891}\n\n![A 36-year-old woman with sinus breast tuberculosis. (a) Ultrasound imaging of the left breast demonstrated an irregular hypoechoic mass extending in a non-parallel orientation in the lower outer quadrant, presenting a 'crater'-like sinus connecting the breast mass to the localized breast skin (arrows). (b) Color Doppler flow imaging demonstrated a few pixels containing blood flow in the margin of the mass (arrow), and the blood flow was classified as category I based on the Adler blood flow assessment. (c) The breast skin was thickened and ulcerated with light yellow pus discharging (arrow), as seen by the naked eye.](10.1177_0300060520910891-fig8){#fig8-0300060520910891}\n\nBlood flow was classified according to Adler category by CDFI ([Table 2](#table2-0300060520910891){ref-type=\"table\"}).\n\nDiscussion {#sec8-0300060520910891}\n==========\n\nBTB is classified as primary or secondary, depending on its origin. Primary BTB is extremely rare and results from direct infection via the breast ducts opening to the nipple or from breast skin abrasions.^[@bibr12-0300060520910891]^ Secondary BTB is much more common and has three main routes of spread: 1) direct spread by extension from surrounding TB lesions such as in the sternum, costa, pleural, and mediastinum; 2) retrograde lymphatic spread from TB lesions in the ipsilateral axillary, neck, chest, supraclavicular, and subclavian nodal regions; and 3) hematogenous spread from TB lesions such as mesenteric lymphatic TB via the blood circulation. The direct and retrograde lymphatic routes are the most common spread routes of secondary BTB.^[@bibr12-0300060520910891],[@bibr13-0300060520910891]^ In the current study, five patients (11.1%, 5/45) showed no significant abnormalities in any organs except the breast and had no history of any other TB lesions, indicating probable primary BTB. The susceptibility to BTB has been reported to be dramatically increased in young, married, pregnant, and lactating women.^[@bibr2-0300060520910891],[@bibr14-0300060520910891],[@bibr15-0300060520910891]^ Two of the current patients (2/45) were lactating, and lactation has been reported to increase the susceptibility to TB infection by 7% to 30%,^[@bibr6-0300060520910891],[@bibr14-0300060520910891],[@bibr15-0300060520910891]^ possibly as a result of ectasia of the breast ducts and the increased number of vessels.\n\nPrevious studies demonstrated that 30% to 75% of cases of BTB were accompanied by involvement of the ipsilateral axillary lymph node.^[@bibr15-0300060520910891],[@bibr16-0300060520910891]^ In this study, only three patients (6.7%, 3/45) developed enlarged axillary lymph nodes, which were proven to be axillary lymphatic TB by biopsy. However, whether concomitant lymphatic TB is a primary source or a secondary involvement remains controversial.^[@bibr15-0300060520910891]^ The low incidence of BTB concomitant with axillary lymphatic TB in the current study (6.7%, 3/45) was likely due to the relatively low sensitivity of US for detecting axillary lymphatic lesions or to a lack of structural alterations in the involved lymph nodes, despite remaining invasion of tuberculosis bacteria.\n\nBTB lesions can be classified into nodular, disseminated, and abscess types.^[@bibr16-0300060520910891]^ In the present study, we classified the BTB lesions into nodular, sinus, or abscess type, according to their US features, accounting for 55.5%, 28.9%, and 15.6% of lesions, respectively. No disseminated-type lesions were observed in this study. The pathological hallmarks of BTB are tuberculous granuloma, caseous necrosis, and liquefaction necrosis, and the BTB types classified in the present study were thought to be related to their pathological alterations. Lesions pathologically identified with tuberculous granuloma, caseous necrosis, and little/no liquefied necrosis were generally classified as nodular type by US, and lesions with a mass of liquefied necrosis were generally classified as abscess type. Lesions in which the necrosis in the BTB lesion broke through the skin to form a sinus tract connecting the breast mass to the localized breast skin were generally classified as sinus type by US. Six months of anti-TB therapy combined with suction or drainage of the abscesses is recommended for abscess type BTB, while antibiotic treatment is also recommended for sinus type BTB, and mastectomy is required if the lesion extends to a large, painful, or ulcerated mass.^[@bibr17-0300060520910891]^ Classification of BTB lesions based on their US features would therefore help to identify disease progression and inform the treatment regimen.\n\nSinus formation is a unique finding highly suspicious of BTB, with about 8% to 39% of BTB lesions associated with sinus formation.^[@bibr13-0300060520910891],[@bibr14-0300060520910891],[@bibr16-0300060520910891],[@bibr17-0300060520910891]^ Of the current 45 BTB cases, 13 were classified as sinus type (28.9%, 13/45), accompanied by breast skin thickening and ulceration. The incidence of BTB with associated sinus formation is likely to be related to the disease course. In our study, 10 of the 13 patients with sinus type BTB (76.9%, 10/13) had a disease course longer than 6 months. A long-term disease course may result from low immunity, ultimately inducing the formation of a tuberculous abscess when liquefactive necrosis occurs. Tuberculous abscesses can gradually extend to the skin, causing skin ulceration and suppuration. Patients with plasma cell mastitis can also develop sinus formation. Lipid accumulation in the enlarged breast ducts in patients with plasma cell mastitis may trigger inflammation and further induce the formation of an abscess. About 20% of patients with plasma cell mastitis presented with sinus tracts.^[@bibr9-0300060520910891],[@bibr18-0300060520910891],[@bibr19-0300060520910891]^ Both BTB and plasma cell mastitis can therefore present with internal hypoecho or heterogeneous echo in a 'crater' shape, making it difficult to differentiate between these two diseases using US, and a definite diagnosis requires US-guided fine-needle aspiration cytology or biopsy specimens.\n\nCalcification is not common in BTB lesions. About 20% of BTB cases have been reported to develop calcifications, which is considered as a sign of a healed granuloma.^[@bibr18-0300060520910891][@bibr19-0300060520910891][@bibr20-0300060520910891][@bibr21-0300060520910891]--[@bibr22-0300060520910891]^ However, BTB calcification has rarely been reported in US studies.^[@bibr18-0300060520910891]^ Calcification was detected in 8 of the 45 cases (17.8%, 8/45) in the current study, including six macrocalcifications and two microcalcifications. In addition, 74% of patients with breast ductal carcinoma and 55.8% with breast sclerosing adenosis developed calcifications, though these are usually associated with multiple punctate or clustered microcalcifications as detected by US.^[@bibr4-0300060520910891],[@bibr23-0300060520910891]^ The present study showed a relatively low incidence of calcification (17.8%, 8/45), and most calcifications were \\>5\u2009mm in diameter (75%, 6/8). One or two punctate microcalcifications were observed in two cases and no clustered calcifications were observed in any cases. Both these two cases had a disease course of \\<6 months and the diameter of the calcification increased after 3 months, and we therefore hypothesized that the microcalcification might have occurred in the early stage of BTB. Tuberculous granulomas will enlarge and calcium salts will be gradually deposited as the disease progresses, leading to the formation of macrocalcifications during the late stage of BTB. Further studies are needed to explore the correlation between the size of the calcification and the duration of the disease.\n\nPrevious imaging studies of BTB have mainly focused on gray images and few have used color images. In the current study, we used CDFI to detect and classify the blood supply conditions of the BTB lesions based on the Adler blood flow categorization. The blood flow was classified as category I in 36 cases, with a punctate color blood-flow signal in the margin and no blood-flow signal in the center. This was thought to be associated with the pathology of the BTB. Pathological alterations in the central BTB lesion are likely to involve caseous or liquefaction necrosis, resulting in avascularity at the center of the lesion. However, the margins of the lesion might retain residual normal tissue or form a tuberculous granuloma, thereby leading to vascularity of the marginal regions. The blood flow in eight cases (17.8%, 8/45) was classified as Adler category 0 and in one case (2.2%, 1/45) as Adler category II, possibly because of the large area of tuberculous granuloma with rich capillaries during sinus formation. Overall, the blood-flow signals in 44 cases (97.8%, 44/45) were classified as below category II, suggesting that hypovascularity might be a common US feature of BTB lesions.\n\nThe present study had several limitations. First, no information was obtained for cases of disseminated BTB, and further investigations of these are therefore needed. Second, we used CDFI to assess lesion blood flow rather than more advanced US techniques such as enhanced contrast US. Finally, we did not analyze other imaging data such as mammography or magnetic resonance imaging data because of the limited sample size.\n\nIn conclusion, BTB lesions can be classified as nodular, abscess, and sinus types based on their US features. BTB is characterized by hypovascularity and internal multi-macrocalcifications. US can provide good characterization of BTB, with great potential for contributing to its clinical diagnosis.\n\nDeclaration of conflicting interest {#sec9-0300060520910891}\n===================================\n\nThe authors declare that there is no conflict of interest.\n\nFunding {#sec10-0300060520910891}\n=======\n\nThis research was supported by the Science and Technology Plan of Hangzhou (20180533B68) and Agriculture and Social Development Plan (20190101A09).\n\nORCID iDs\n=========\n\nWenzhi Zhang <https://orcid.org/0000-0002-4369-7448>\n\nGaoyi Yang <https://orcid.org/0000-0002-0759-3684>\n\nTianzhuo Yu <https://orcid.org/0000-0003-3745-6864>\n"}
+{"text": "1. Introduction\n===============\n\nQuercetin is a category in the class of flavonoids, and a sub class of flavonol. Flavonoids are plant polyphenolics found as pigments in fruits, vegetables, seeds, nuts, flowers, barks and leaves. It is also found in medicinal botanicals, such as Ginkgo biloba, Hypericum perforatum (St. John's Wort), and Sambucum canadensis (El-der) \\[[@B001]\\]. The International Union of Pure and Applied Chemistry's (IUPAC's) name for quercetin is 3, 3', 4', 5, 7-pentahydroxyflavone (or its synonym 3, 3', 4', 5, 7-pentahydroxy-2-phenylchromen-4-one). Fig.1 shows the chemical structure of quercetin. The hydroxyl (-OH) groups attached at positions 3, 5, 7, 3', and 4' and the catechol B-ring produce the antioxidant properties of quercetin \\[[@B002], [@B003]\\]. The antioxidant and the free radical scavenging properties of quercetin have been reported to contribute to anti carcinogenic and anti inflammatory effects, and haves been extensively studied by researchers around the world \\[[@B002]\\].\n\nExtensive amounts of *in vitro* and *in vivo* animal research on quercetin's pharmacological activities have been carried out, suggesting that quercetin might be used as a new therapeutic approach to decrease blood pressure \\[[@B004]\\], to inhibit fibronectin production by keloid derived fibroblasts \\[[@B005]\\], to inhibit neointimal hyperplasia in the abdominal aorta of rats \\[[@B006]\\], to treat gout \\[[@B007]\\], to inhibit asthmatic syndrome \\[[@B008]\\] and to promote dermal wound healing \\[[@B009]\\].\n\nCurcumin, commercially available in a mixture of curcumins (curcuminoids), contains \u2500 77% pure curcumin, \u2500 17% demethoxycurcumin and \u2500 3% bisdemethoxycurcumin \\[[@B010]\\] (Fig [1](#F001){ref-type=\"fig\"}). Curcuminoids are derived from Curcuma longa Linn, one of the most popular medicinal herbs, and are a polyphenolic. These compounds are yellow pigments and have been, commonly used as a dietary spices, natural coloring agents in foods, household medicines and insect repellents in South and Southeast Asia for thousands of years \\[[@B011]\\]. Curcumin and its synthetic derivatives (curcuminoids) show an array of pharmacological properties, such as antibacterial \\[[@B012] - [@B014]\\], antioxidant \\[[@B013], [@B015] - [@B016]\\], anti inflammatory \\[[@B017], [@B018]\\], anti tumor \\[[@B019], [@B020]\\] and anti proliferation \\[[@B018], [@B021]\\] properties. Curcumin/curcuminoids also possess potency as medicines for the treatment of diseases, including Alzheimer's disease \\[[@B022], [@B023]\\], cancer \\[[@B024], [@B025], [@B026]\\], diabetes, gastric ulcers \\[[@B027]\\], malaria \\[[@B028], [@B029]\\] and for the treatment of wounds \\[[@B030] - [@B032]\\].\n\nA variety of methods for quantitatively detecting curcumin and quercetin contents have been reported. Among these, spectrophotometric methods are the most commonly used \\[[@B033] - [@B036]\\]. Thin layer chromatography (TLC) or column chromatography was usually used for separation of curcuminoids \\[[@B037] - [@B039]\\]. High performance liquid chromatography (HPLC) \\[[@B040] - [@B045]\\] and, high performance thin layer chromatography (HPTLC) \\[[@B039], [@B046], [@B047]\\] are the commonly used methods for quantitatively detecting the quercetin and curcuminoids contents. Some advanced methods have been developed for the analysis of curcuminoids contents, namely, ultra performance liquid chromatography quadrupole time of flight mass spectrometry (UPLC-qTOF-MS) \\[[@B048]\\], ultra performance liquid chromatography tandem mass spectrometry (UPLC-MS) \\[[@B049]\\], high performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS) \\[[@B050]\\] and electrochemical-HPLC \\[[@B051]\\].\n\nFor the above techniques, spectrophotometric methods are not available to quantify the individual curcuminoids due to the curcumin derivative's being also absorbed at the same wavelength. Furthermore, LC-MS and/or qTOF are complicated and need expensive instrumentation. Even though HPTLC and TLC are widely used to study the fingerprints of plants, these methods are not suitable for analyzing compounds in combinations of herbs products like Chinese medicinal materials (because such products normally contain more than one herb). For simultaneous determination of quercetin and curcuminoids, HPLC method is the recommended technique because it uses separation, identification and quantification of the analytes from plant extracts, foods, pharmaceutical products, and body fluids.\n\nIn the present study, a simple isocratic reversed phase HPLC method was developed according to international conference harmonisation (ICH) guidelines \\[[@B052]\\] for the simultaneous quantitative detection of quercetin and curcuminoids. The method was also validated by using market available traditional Chinese medicine materials such as granules, pills and tablets.\n\n2. Materials and Methods\n========================\n\nCurcumin (mixture of curcumin, demethoxycurcumin, and bisdemethoxycurcumin) was obtained from Acros Organics, USA. Quercetin anhydrous was obtained from Sigma, USA. The HPLC grade acetonitrile and methanol were purchased from J.T. Baker, USA. Analytical grade acetic acid was obtained from QR\u00ebC, Malaysia. Nylon membrane filters 0.45 \u03bcm were purchased from Whatman, England.\n\nHPLC analysis was performed using a Shimadzu-LC system (Shimadzu, Japan) equipped with an CBM-20A controller, LC-20AT pump, DGU-20A5 prominence degasser, SIL-20A auto sampler, SPD-20AV detector and CTO-10ASvp column oven.\n\nChromatographic separations were achieved using a Thermo Hypersil Gold column (250 mm \u00d7 4.6 mm I.D.: 5 \u03bcm). A security guard column (Zorbax Eclipse Plus) packed with a replaceable C-18 cartridge (12.5 mm \u00d7 4.6 mm ID.: 5 mm) was used to protect the analytical column. A reverse phase HPLC assay was carried out using an isocratic elution with a flow rate of 1.3 mL/minutes, a column temperature of 35\u00b0C, a mobile phase of acetonitrile and 2% v/v acetic acid (pH 2.60) (40% : 60% v/v) and a detection wavelength of 370 nm. The injection volume was 20 \u03bcL of each solutions. The total run time was 18.5 minutes for each injection. Data were acquired and processed with LC-Solution Software. Solvents and distilled water were prior filtered through a 0.45-\u03bcm nylon membrane by using a set of glass bottles with the aid of a vacuum pump (Fisherbrand FB 70155, Fisher Scientific, UK).\n\nTwenty mg of a mixture of curcumin (containing mainly curcumin, demethoxycurcumin and bisdemethoxycurcumin) and 20 mg of quercetin were accurately weighed using a microbalance (Sartorius, MC5, Germany) and dissolved in 20 mL of HPLC grade methanol in a 20 mL volumetric flask. The mixtures were diluted to 320 \u03bcg/mL with HPLC grade methanol; and were then serially doubling diluted to 1.22 ng/mL. These solutions were used as calibration standards for the quantitative determinations of the limit of detection (LOD), the limit of quantification (LOQ) and yhe limit of linearity (LOL), and for the linear range analysis. Three quality control (QC) samples at concentrations of 3.75 \u03bcg/mL, 100 \u03bcg/mL and 160 \u03bcg/mL, respectively, were prepared from the stock solution. All solutions were stored in tightened screw cap bottles to avoid evaporation and were protected from light, and were kept in a refrigerator (4\u00b0C) for not more than two weeks.\n\n![Chemical structures of quercetin, and the curcuminoids: curcumin, demethoxycurcumin and bisdemethoxycurcumin.](2093-6966-v17-n04-036-g001){#F001}\n\nStandard solutions with concentrations in the range from of 1.22 ng/mL to 320 \u03bcg/mL were injected in duplicate into the HPLC unit. The LOD and LOQ of quercetin (QUE), bisdemethoxycurcumin (BDMC), demethoxycurcumin (DMC) and curcumin (CUR) were determined in a at the lower concentration range based on the signal to-noise ratio. According to The United Sates Pharmacopeia (USP), the LOD and the LOQ are in terms of 2 or 3 times, and 10 times the noise level respectively. The LOL was determined by plotting a calibration curve (mean value of the peak areas against the concentrations) beginnings with the LOQ concentration and proceeding to the data point that deviated from the regression line. The coefficient of determination (R^2^ \u2265 0.999) was used as a guideline to evaluate the model fit of a regression equation.\n\nLinear ranges for quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin included concentrations of 1.25, 5, 20, 40, 80, 140 and 200 \u03bcg/mL. Separate calibration curves were constructed for quercetin, bisdemethoxycurcumin demethoxycurcumin and curcumin by plotting the peak areas against the concentrations, and the methods were evaluated by determining the coefficient of determination (R^2^). Unknown assay samples were quantified by referencing them to these calibration curves.\n\nQC samples (3.75, 100 and 160 \u03bcg/mL) were used to validate intra day and inter day accuracies and precisions. Intra day precisions and accuracies were determined by using a replicate analysis (n = 6) of the QC samples on the same day under the same analytical conditions. Inter day precisions and accuracies were tested by using a replicate analysis (n = 3) of the same QC samples on six consecutive days. The precision is calculated from the mean of the accuracy and the relative standard deviation (RSD). Accuracy is a measure of how close the experimental value to the true value, and is expressed as a percent. The experimental value was calculated from the calibration curve by using the linear regression equation, y = mx + c. The constant m is the slope of the curve. The constant c is the y intercept and can be determined by extrapolating the straight line to the y axis.\n\nFour variation parameters of robustness were studied: change in organic composition by \u00b1 2.0% (Table [4a](#T04a){ref-type=\"table\"}) , change in acetic acid concentration by \u00b1 1.0% v/v (effect of buffer pH) (Table [4b](#T04b){ref-type=\"table\"}), change in the flow rate of \u00b1 0.1 mL/min (Table [4c](#T04c){ref-type=\"table\"}) and change in the column temperature of \u00b1 5.0\u00b0C (Table [4d](#T04d){ref-type=\"table\"}). The retention time, peak area, resolution, tailing factor, theoretical plate number and capacity factor values obtained from the variation parameters were compared to those obtained for the normal method conditions. The differences were analyzed by using SPSS version 20, and a one way analysis of variance (ANOVA), followed by Tukey's test. *P*-values \\< 0.05 were considered significant.\n\nThe system suitability parameters were assessed by using six replicate analysis of the QC sample at 160 \u03bcg/mL. The acceptance criteria were in accordance with the guidelines of the Centre for Drug Evaluation and Research \\[[@B053]\\].\n\nThe method developed in this study was used to quantitatively determination the quercetin and the curcuminoid contents of extracts, pills and tablets made from Chinese medicinal plants.\n\n3. Results\n==========\n\nThe LOD and the LOQ were determined based on the signal to noise (S/N) ratio, with the S/N \\> 3 and the S/N \\> 10 for the LOD and the LOQ, respectively. The LODs of quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin were 0.00488, 0.62500, 0.07813 and 0.03906 \u03bcg/mL, respectively. The LOQs of quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin were 0.03906, 2.5000, 0.31250 and 0.07813 \u03bcg/mL, respectively (Table [1](#T001){ref-type=\"table\"}) The linearity for detecting quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin was tested against a mixture of calibration standards with concentration ranging from 1.22 ng/mL to 320 \u03bcg/mL. The LOL of each compound was determined from a separate calibration curve. Quercetin was linear up to 200 \u03bcg/mL, while bisdemethoxycurcumin, demethoxycurcumin and curcumin were linear up to 320 \u03bcg/mL.\n\n###### LOD, LOQ, LOL and linear regression analysis parameters for QUE, BDMC, DMC and CUR\n\n  Compounds   LOD (\u03bcg/mL)   LOQ (\u03bcg/mL)   LOL (\u03bcg/mL)   Regression analysis (1.25 --- 200 \u03bcg/mL)                   \n  ----------- ------------- ------------- ------------- ------------------------------------------ --------------- ---------\n  QUE         0.00488       0.03906       200           70055.85913                                1521.41433      0.99993\n  BDMC        0.62500       2.50000       320           1807.72930                                 --- 440.28180   0.99984\n  DMC         0.07813       0.31250       320           10011.55795                                40.13501        0.99985\n  CUR         0.03906       0.07813       320           34176.44088                                3645.08890      0.99993\n\nLOD, limit of detection; LOQ, limit of quantification, LOL, limit of linearity; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.\n\n![Chromatograms of quercetin and curcuminoids. QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.](2093-6966-v17-n04-036-g002){#F002}\n\nLinear calibration curves in the range from 1.25 to 200 \u03bcg/ mL were constructed for each compound by plotting the peak area against the concentration. The retention times and the peak areas are tabulated in (Table [2](#T002){ref-type=\"table\"}) The values of R^2^, the y-intercept and the slope for each compound's calibration plot are shown in (Table [1](#T001){ref-type=\"table\"}) A regression analysis of the data showed a linear relationship for quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin, with excellent R^2^ values of 0.99993, 0.99984, 0.99985 and 0.99993 \u03bcg/mL, respectively.\n\n###### Retention times and responses data for calibration standards of QUE, BDMC, DMC, and CUR\n\n  Concentration(\u03bcg/mL)   Retention time (n = 5)   Peak area (n = 5)              \n  ---------------------- ------------------------ ------------------- ---------- -------\n  QUE                                                                            \n  1.25                   3.970                    0.117               94937      0.676\n  5                      3.972                    0.066               367965     0.739\n  20                     3.972                    0.041               1438240    0.624\n  40                     3.973                    0.055               2781685    0.508\n  80                     3.972                    0.029               5582929    0.437\n  140                    3.972                    0.048               9735618    0.866\n  200                    3.972                    0.053               14073938   0.368\n  BDMC                                                                           \n  1.25                   13.823                   0.308               1859       1.611\n  5                      13.840                   0.095               8843       1.181\n  20                     13.842                   0.093               37086      1.089\n  40                     13.843                   0.087               71560      1.044\n  80                     13.846                   0.117               143659     1.073\n  140                    13.846                   0.134               249462     1.835\n  200                    13.849                   0.060               363457     0.850\n  DMC                                                                            \n  1.25                   15.214                   0.227               14705      0.273\n  5                      15.229                   0.096               52692      0.540\n  20                     15.230                   0.074               204602     0.665\n  40                     15.232                   0.073               398446     0.436\n  80                     15.237                   0.099               798153     0.867\n  140                    15.236                   0.120               1384220    1.416\n  200                    15.242                   0.039               2015583    0.158\n  CUR                                                                            \n  1.25                   16.708                   0.199               46645      0.856\n  5                      16.718                   0.077               182515     0.901\n  20                     16.719                   0.061               701982     0.700\n  40                     16.720                   0.064               1358591    0.299\n  80                     16.725                   0.096               2737751    0.423\n  140                    16.725                   0.108               4749355    0.897\n  200                    16.734                   0.067               6866971    0.313\n\nRSD, relative standard deviation; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.\n\nThe peaks of quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin were well separated at different retention times with resolutions of 32.195, 2.887 and 2.830 for quercetin-bisdemethoxycurcumin, bisdemethoxycurcumin-demethoxycurcumin and demethoxycurcumin-curcumin, respectively. No interferences or excipient peaks co eluted with the analytes were observed, indicating the method is selective and specific in relation to the medium and excipients used in this study (Fig [2](#F002){ref-type=\"fig\"}), (Table [2](#T002){ref-type=\"table\"}).\n\nPrecision and accuracy data for the intraday and the inter-day variations for the three QC samples are summarized in (Table [3](#T003){ref-type=\"table\"}). The RSD values for the intraday and the inter day precisions were \\< 1%. For the accuracy test, the intraday and the inter day accuracies ranges from 98.292% to 103.617%, confirming the accuracy of the method.\n\n###### Precisions and accuracies for intraday and interday repetitions for the quantitative detection of QUE, BDMC, DMC and CUR\n\n  ---------------------------------------------------------------------------------------\n  Concentration\\   Intra day^\\*^   Inter day^\u2020^                                 \n  (\u03bcg/mL)                                                                       \n  ---------------- --------------- -------------- ---------- ---------- ------- ---------\n  QUE                                                                           \n\n  3.75             263151          0.432          99.589     263350     0.323   99.665\n\n  100              7064599         0.717          100.821    7221470    0.646   103.060\n\n  160              11221611        0.806          100.010    11218287   0.968   100.070\n\n  BDMC                                                                          \n\n  3.75             6243            0.576          98.588     6181       0.854   97.669\n\n  100              182293          0.723          101.084    186683     0.878   103.513\n\n  160              286851          0.654          99.32746   288040     0.805   99.738\n\n  DMC                                                                           \n\n  3.75             37700           0.635          100.310    37687      0.466   100.276\n\n  100              1010004         0.752          100.880    1037410    0.078   103.617\n\n  160              1590498         0.651          99.289     1594989    0.844   99.569\n\n  CUR                                                                           \n\n  3.75             129618          0.655          98.292     129152     0.297   97.929\n\n  100              3456218         0.732          101.022    3544535    0.275   103.606\n\n  160              5448675         0.711          99.576     5454012    0.829   99.673\n  ---------------------------------------------------------------------------------------\n\n\\*Intra day repetitions for each concentration were analyzed on the same day. \u2020Inter day repetitions for each concentration, were analyzse on six consecutive days. RSD, relative standard deviation; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.\n\n###### (a). Robustness -- change in organic composition\n\n  System suitability              Compound    Change in the normal organic composition of acetonitrile: 2% acetic acid                                                 \n  ------------------------------- ----------- -------------------------------------------------------------------------- ----------- ---------- ----------- ---------- -------\n  Retention time, t~R~(minutes)   QUE         3.993                                                                      0.690       4.251      0.155       3.761      0.040\n  BDMC                            13.951      0.342                                                                      17.645      0.374      11.280      0.084      \n  DMC                             15.340      0.291                                                                      19.543      0.285      12.330      0.084      \n  CUR                             16.829      0.245                                                                      21.617      0.296      13.464      0.082      \n  Peak area                       QUE         6853044                                                                    0.433       6836934    0.441       6867445    0.117\n  BDMC                            167417      0.647                                                                      161504      0.801      146484      0.578      \n  DMC                             940836      0.404                                                                      903191      0.781      965307      0.071      \n  CUR                             3302593     0.236                                                                      3206134     0.555      3367309     0.114      \n  Resolution, R                   QUE         \\-                                                                         \\-          \\-         \\-          \\-         \\-\n  BDMC                            32.498      0.379                                                                      36.449      0.471      29.120      0.063      \n  DMC                             2.908       0.208                                                                      3.272       2.369      2.736       0.181      \n  CUR                             2.850       0.237                                                                      3.243       1.648      2.666       0.124      \n  Tailing factor, T~f~            QUE         1.371                                                                      0.254       1.347      0.115       1.392      0.074\n  BDMC                            1.533       0.364                                                                      1.283       2.314      1.080       0.200      \n  DMC                             1.160       0.484                                                                      1.083       0.151      1.431       0.082      \n  CUR                             1.094       0.094                                                                      1.076       0.284      1.114       0.037      \n  Theoretical plate, N            QUE         8752.133                                                                   1.463       8857.791   0.312       8520.171   0.238\n  BDMC                            15931.889   1.147                                                                      16311.011   0.058      16303.130   0.103      \n  DMC                             14298.287   1.761                                                                      16569.474   1.029      14210.321   0.233      \n  CUR                             16008.049   1.202                                                                      16543.754   0.535      15157.508   0.340      \n  Capacity factor, k'             QUE         0.680                                                                      0.344       0.777      0.906       0.601      0.327\n  BDMC                            4.878       0.202                                                                      3.800       0.209      3.800       0.209      \n  DMC                             5.463       0.232                                                                      7.214       1.592      4.247       0.206      \n  CUR                             6.097       0.253                                                                      8.038       0.481      4.729       0.209      \n\n###### (b). Robustness -- change in acetic acid concentration\n\n  System suitability              Compound    Change in the acetic acid concentration (% v/v )                                                 \n  ------------------------------- ----------- -------------------------------------------------- ----------- ---------- ----------- ---------- -------\n  Retention time, t~R~(minutes)   QUE         3.972                                              0.175       4.054      0.064       3.893      0.071\n  BDMC                            13.868      0.310                                              14.549      0.086      13.177      0.167      \n  DMC                             15.255      0.265                                              16.017      0.085      14.542      0.153      \n  CUR                             16.743      0.213                                              17.590      0.084      16.028      0.141      \n  Peak area                       QUE         7039483                                            0.562       6966950    0.525       6952833    0.630\n  BDMC                            180475      0.541                                              176885      0.575      152439      0.895      \n  DMC                             1000716     0.736                                              987128      0.551      956266      0.670      \n  CUR                             3433379     0.754                                              3428762     0.533      3428762     0.558      \n  Resolution, R                   QUE         \\-                                                 \\-          \\-         \\-          \\-         \\-\n  BDMC                            32.327      0.172                                              33.254      0.244      31.950      0.268      \n  DMC                             2.900       0.370                                              2.974       0.303      3.033       0.527      \n  CUR                             2.840       0.429                                              2.904       0.339      2.966       0.608      \n  Tailing factor, T~f~            QUE         1.366                                              0.077       1.364      0.215       1.370      0.110\n  BDMC                            1.493       1.377                                              1.463       0.331      1.060       0.139      \n  DMC                             1.160       1.075                                              1.137       0.103      1.325       0.823      \n  CUR                             1.085       0.148                                              1.092       0.050      1.083       0.108      \n  Theoretical plate, N            QUE         8711.993                                           0.267       8877.546   0.460       8548.948   0.269\n  BDMC                            15740.557   0.397                                              16067.808   0.689      16308.146   0.664      \n  DMC                             14041.181   0.701                                              14691.580   0.675      14241.082   1.031      \n  CUR                             15793.019   0.472                                              16098.239   0.701      15531.342   0.811      \n  Capacity factor, k'             QUE         0.656                                              1.783       0.680      1.484       0.610      0.803\n  BDMC                            4.798       1.202                                              5.036       0.658      4.449       0.511      \n  DMC                             5.333       0.988                                              5.637       0.698      5.014       0.478      \n  CUR                             6.016       1.416                                              6.295       0.628      5.629       0.443      \n\n###### (c). Robustness -- change in flow rate\n\n  ---------------------------------------------------------------------------------------------------------------\n  System suitability      Compound    Change in flow rate                                                 \n  ----------------------- ----------- --------------------- ----------- ---------- ----------- ---------- -------\n  Retention time, t~R~\\   QUE         3.972                 0.175       4.291      0.105       3.696      0.130\n  (minutes)                                                                                               \n\n  BDMC                    13.868      0.310                 14.953      0.321      12.909      0.333      \n\n  DMC                     15.255      0.265                 16.442      0.284      14.235      0.279      \n\n  CUR                     16.743      0.213                 18.038      0.262      15.668      0.298      \n\n  Peak area               QUE         7039483               0.562       7606272    0.662       6530571    0.497\n\n  BDMC                    180475      0.541                 194216      0.753      167111      1.593      \n\n  DMC                     1000716     0.736                 1078076     0.714      928707      1.345      \n\n  CUR                     3433379     0.754                 3700134     0.690      3185325     1.198      \n\n  Resolution, R           QUE         \\-                    \\-          \\-         \\-          \\-         \\-\n\n  BDMC                    32.327      0.172                 32.779      0.199      32.047      0.928      \n\n  DMC                     2.900       0.370                 2.921       0.608      2.936       2.014      \n\n  CUR                     2.840       0.429                 2.864       0.723      2.868       1.647      \n\n  Tailing factor, T~f~    QUE         1.366                 0.077       1.360      0.183       1.371      0.287\n\n  BDMC                    1.493       1.377                 1.490       1.891      1.539       1.614      \n\n  DMC                     1.160       1.075                 1.157       1.447      1.181       2.364      \n\n  CUR                     1.085       0.148                 1.081       0.101      1.087       0.207      \n\n  Theoretical plate, N    QUE         8711.993              0.267       9148.347   0.429       8249.430   0.420\n\n  BDMC                    15740.557   0.397                 16035.103   1.342      15696.046   2.851      \n\n  DMC                     14041.181   0.701                 14374.944   1.036      13420.220   0.844      \n\n  CUR                     15793.019   0.472                 16216.013   1.854      15379.165   2.364      \n\n  Capacity factor, k'     QUE         0.656                 1.783       0.661      0.832       0.627      0.762\n\n  BDMC                    4.798       1.202                 4.780       0.942      4.750       3.067      \n\n  DMC                     5.351       0.661                 5.355       0.497      5.350       2.846      \n\n  CUR                     5.966       0.632                 5.985       0.500      5.862       0.427      \n  ---------------------------------------------------------------------------------------------------------------\n\n###### (d). Robustness -- change in column temperature\n\n  System suitability              Compound    Change in column temperature                                                 \n  ------------------------------- ----------- ------------------------------ ----------- ---------- ----------- ---------- -------\n  Retention time, t~R~(minutes)   QUE         3.956                          0.031       4.063      0.074       3.861      0.162\n  BDMC                            13.673      0.070                          14.647      0.172      12.810      0.268      \n  DMC                             15.037      0.064                          15.980      0.153      14.167      0.236      \n  CUR                             16.502      0.064                          17.423      0.143      15.657      0.196      \n  Peak area                       QUE         7628483                        0.252       7620525    0.254       7633341    0.259\n  BDMC                            196493      0.261                          202870      0.253      172397      0.136      \n  DMC                             1091099     0.300                          1124567     0.281      1058404     0.205      \n  CUR                             3738544     0.244                          3836306     0.285      3643910     0.196      \n  Resolution, R                   QUE         \\-                             \\-          \\-         \\-          \\-         \\-\n  BDMC                            31.946      1.437                          32.560      0.233      31.471      0.267      \n  DMC                             2.872       1.359                          2.698       0.334      3.155       0.481      \n  CUR                             2.829       0.575                          2.718       0.305      3.106       0.559      \n  Tailing factor, T~f~            QUE         1.343                          0.056       1.329      0.267       1.351      0.124\n  BDMC                            1.551       0.421                          1.233       0.525      1.089       0.077      \n  DMC                             1.186       0.241                          1.098       0.268      1.491       1.051      \n  CUR                             1.097       0.082                          1.094       0.107      1.098       0.129      \n  Theoretical plate, N            QUE         8734.237                       0.300       8837.810   0.448       8619.473   0.295\n  BDMC                            15779.175   0.595                          15065.104   0.600      16276.545   1.203      \n  DMC                             13866.206   1.175                          15742.765   0.475      15286.704   0.311      \n  CUR                             15846.706   0.791                          15917.987   0.394      15793.349   0.183      \n  Capacity factor, k'             QUE         0.698                          1.598       0.717      1.945       0.616      0.692\n  BDMC                            4.855       0.115                          5.190       0.989      4.366       0.214      \n  DMC                             5.449       0.349                          5.752       0.966      4.962       1.315      \n  CUR                             6.058       0.201                          6.362       0.942      5.581       1.015      \n\nThe normal conditions of HPLC are a mobile phase of acetonitrile: 2% acetic acid (pH 2.60) = 40% : 60 % v/v, flow rate 1.3 mL/min at UV wavelength of 370 nm and column temperature at 35\u00b0C. RSD, relative standard deviation; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.\n\n###### System suitability parameters, calculation formula and recommendations\n\n  Parameter               Formula                                   Recommendation\n  ----------------------- ----------------------------------------- --------------------------\n  Precision               RSD = S/x\u0304^\\*^100                          RSD \u2264 1% for n \u2265 5\n  Resolution, R           R = (t~R2~-- t~R1~)/(1/2)(t~w1~--t~w2~)   \\> 2\n  Tailing factor, T~f~    Tf = Wx/2f                                \u2264 2\n  Theoretical plates, N   N = 16(t~R~/t~w~)^2^                      Column efficiency \u2265 2000\n  Capacity factor, k      k' = (t~R~-- t~0~)/t~0~                   \\> 2\n\nS, standard deviation; x\u0304 , mean of the data; t~R~, retention time of analyte 1; tw, peak width measured to the baseline of the extrapolated straight sides to baseline; W~x~, width of the peak determined at either 5% (0.05) or 10% (0.10) from the baseline of the peak height; f, distance between peak maximum and peak front at W~x~; t~0~, elution time of the void volume or non retained components.\n\n###### System suitability testing\n\n  Parameter              QUE        BDMC    DMC         CUR                                     \n  ---------------------- ---------- ------- ----------- ------- ----------- ------- ----------- -------\n  Retention time, t~R~   3.970      0.021   13.840      0.027   15.230      0.025   16.723      0.021\n  Peak area              11221611   0.806   286851      0.654   1590498     0.651   5448675     0.711\n  Resolution, R          \\-         \\-      32.195      0.321   2.887       0.364   2.830       0.370\n  Tailing factor, T~f~   1.369      0.108   1.501       0.261   1.165       0.144   1.081       0.051\n  Theoretical plate, N   8803.785   0.359   15552.398   0.865   13763.145   0.646   15568.252   0.910\n  Capacity factor, k'    0.684      0.846   4.870       0.415   5.460       0.406   6.093       0.391\n\nRSD, relative standard deviation; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin. N, number of theoretical plates; k', capacity factor; Mean of six replicate injections of quality control (QC) standard of 160 \u03bcg/mL.\n\nRobustness is a measure of the method's capability to remain unaffected by small, but deliberate, variations in the method parameters \\[[@B052]\\]. The robustness parameters tested were the mobile phase's composition, the concentration of acetic acid (pH effect), the flow rate and the column temperature. The results are tabulated in Table 4( [a](#T04a){ref-type=\"table\"} - [d](#T04d){ref-type=\"table\"}). The retention times for all four compounds due to variations in the parameters were significantly different compared to those for the normal parameters. The peak area for curcumin was not significantly different after changing the acetic acid concentration from 2% to 3%, but was significantly different after changing the concentration from 2% to 1%. Quercetin, bisdemethoxycurcumin and demethoxycurcumin were shown to have significant differences in their peak area when the concentration of acetic acid was changed. Changes in the acetonitrile's composition and temperature were shown not to cause significant differences in quecetin's peak areas, however significant differences were seen in curcumin, bisdemethoxycurcumin and demethoxycurcumin peak areas. Increasing or decreasing the flow rate by 0.1 mL/min from normal conditions significantly raised or reduced the values of the peak areas of quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin. Although changes in experimental conditions changed the retention time, the peak area and the values of the system's suitability parameters, the four analyzed peaks were still well resolved from each other and from additional small peaks and showed good resolution in the tested parameters (Fig [3](#F003){ref-type=\"fig\"})\n\nThe system suitability criteria were in accordance with the Centre for Drug Evaluation and Research (CDER) guidelines \\[[@B054]\\] and are summarized in (Table [5](#T005){ref-type=\"table\"}) The mean values of the six replicate injections of 160 \u03bcg/mL QC standards were used to evaluate the retention time, the peak area, the resolutions for the analyte peaks, the tailing factor, the number of theoretical plates and the capacity factor. The results for the system suitability parameters are shown in (Table [6](#T006){ref-type=\"table\"}) The RSD values for the tested parameters were \\< 1%, indicating the precision of the method. The tested parameters passed the criteria under the CDER guidelines except for the capacity factor value for quercetin (\\< 2) \\[[@B053]\\]. This is because the retention time of quercetin was quite fast and just 1 minute behind the solvent peak. However, the quercetin peak was well resolved from the solvent peak and from the front additional small peak.\n\n![Combined chromatograms of quercetin (QUE), bisdemethoxycurcumin (BMDC), demethoxycurcumin (DMC), curcumin (CUR) analyzed at different conditions: (a) acetonitrile: 2% acetic acid at a flow rate of 1.3 mL/minutes, 35\u00b0C (b) acetonitrile: different acetic acid concentrations (40% : 60% v/v) at a flow rate of 1.3 mL/min, 35\u00b0C (c) acetonitrile: 2% acetic acid (40% : 60% v/v) at different flow rates, 35\u00b0C (d) acetonitrile: 2% acetic acid (40% : 60% v/v) at a flow rate of 1.3 mL/min at different temperatures.](2093-6966-v17-n04-036-g003){#F003}\n\nThe proposed method was applied to quantitatively detect the quercetin and curcuminoids in Chinese medicines such as plant granule extracts, tablets and pills. The results of 19 samples are summarized in (Table [7](#T007){ref-type=\"table\"}). In the tested samples, BDMC had the highest concentration compared to the other two curcuminoids tested (DMC and CUR), and was found in the formulations of granule extracts, tablets and pills (such as samples 12, 13, 15, 16, 18 and 19) (Table [7](#T007){ref-type=\"table\"}). The preference of BDMC over CUR in the medicine might be due to its strong biological properties, which its use as a cure for diseases or as a supplement for certain purposes. Quercetin was found in most of the tested samples, indicating that this compound is common and useful for treatment. (Fig [4](#F004){ref-type=\"fig\"}) shows the chromatograms for the quercetin and the curcuminoids found in the tested samples.\n\n###### Concentration of QUE, CUR, DMS and BDMC in Chinese medicines\n\n  No   Chinese medicine                    Type                           Concentration (mean \u00b1 S.D\\*) (\u03bcg/100 mg)                         \n  ---- ----------------------------------- ------------------------------ ------------------------------------------ ---------- ---------- ----------\n  1    Gao liang jiang (\u9ad8\u826f\u59dc)            Single plant granule extract   0.7532                                     N.D        134.8739   0.5270\n  2    Jin qian cao (\u91d1\u94b1\u8349)               Single plant granule extract   4.0618                                     N.D        N.D        0.8263\n  3    Yu jin (\u90c1\u91d1)                       Single plant granule extract   0.3195                                     69.1060    27.2286    27.1020\n  4    E su (\u83aa\u672f)                         Single plant granule extract   0.5983                                     79.5922    42.6982    8.6812\n  5    Jiang huang (\u59dc\u9ec4)                  Single plant granule extract   3.6523                                     N.D        933.8122   796.0621\n  6    Yu xing cao (\u9c7c\u8165\u8349)                Single plant granule extract   1.7930                                     N.D        N.D        1.3424\n  7    Ting li zi (\u8476\u82c8\u5b50)                 Single plant granule extract   1.3604                                     N.D        N.D        N.D\n  8    Tu si zi (\u83df\u4e1d\u5b50)                   Single plant granule extract   3.9300                                     N.D        N.D        N.D\n  9    Di yu (\u5730\u6986)                        Single plant granule extract   0.8962                                     N.D        N.D        N.D\n  10   Kui hua (\u6127\u82b1)                      Single plant granule extract   311.0307                                   N.D        N.D        N.D\n  11   Sang ju yin (\u6851\u83ca\u996e)                Formulation granule extract    0.7402                                     N.D        0.3558     0.2537\n  12   Chai hu su gan san(\u67f4\u80e1\u758f\u809d\u6563)      Formulation granule extract    0.2029                                     126.8843   48.3408    1.6417\n  13   Xiao yao san (\u900d\u9065\u6563)               Formulation granule extract    0.4991                                     97.9203    2.5534     0.4301\n  14   Long dan xie gan tang(\u9f99\u80c6\u6cc4\u809d\u6c64)   Formulation granule extract    11.1482                                    5.2111     1.2817     0.1236\n  15   Sang ju gan mao pian(\u6851\u83ca\u611f\u5192\u7247)    Tablet                         17.3489                                    173.6155   2.8579     N.D\n  16   Dan zhi xiao yao pian(\u4e39\u6800\u900d\u9065\u7247)   Tablet                         7.8101                                     135.1892   1.0883     0.2624\n  17   Long dan xie gan pian(\u9f99\u80c6\u6cc4\u809d\u7247)   Tablet                         N.D                                        5.5352     6.7428     0.2378\n  18   Bu zhong yi qi (\u8865\u4e2d\u76ca\u6c14)           Tablet                         0.9052                                     623.1338   5.9485     0.5964\n  19   Xiao yao wan (\u900d\u9065\u4e38)               Pill                           12.015                                     79.7951    11.7471    1.1516\n\n\\*n = 3; N.D, not detected; QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.\n\n4. Discussion\n=============\n\nThe HPLC method was developed by optimization of the mobile phase conditions so that quercetin, bisdemethoxycurcumin, demethoxycurcumin and curcumin peaks could be simultaneously detected by using the same solvent system and an isocratic method. The flow rate, acetic acid concentration and column temperature were varied to determine the chromatographic conditions giving the best separation and the shortest analysis time. UV visible sperctrophotometry in the wavelength from 200 to 500 nm was used for the detection of quercetin and curcuminoids; 370 nm was chosen as appropriate wavelength for the analysis of quercetin and curcumin derivatives.\n\nThe retention times for quercetin (3.97 minutes), bisdemethoxycurcumin (13.84 minutes), demethoxycurcumin (15.23 minutes) and curcumin (16.72 minutes) were reasonable because the method is simple and general. The chromatograph peaks for mixtures of curcumin were identified based on their percentages in the mixtures. Most of the commercially available curcumin/turmeric products contain mixtures of curcumin, demethoxycurcumin and bisdemethoxycurcumin. Among these, curcumin (46% \u2500 72%) is the major compound, followed by demethoxycurcumin (11% \u2500 28%) and bisdemethoxycurcumin (3% \u2500 14%). All four analyte peaks were well separated from each other and from small additional peaks.\n\nThe linear ranges of quercetin (0.039 \u2500 200 \u03bcg/mL), bisdemethoxycurcumin (2.500 \u2500 320 \u03bcg/mL), demethoxycurcumin (0.313 \u2500 320 \u03bcg/mL) and curcumin (0.078 \u2500 320 \u03bcg/mL) are suitable for the analysis of most the pharmaceutical products, containing the compounds and for the analysis of crude herbs. The low LOD and LOQ values indicate that the method provides adequate sensitivity. The R2 values \\> 0.999 for the regression model for the calibration curves confirm the good linearity of the method.\n\nThe accuracies ranged from 98.292% \u2500 103.617%, and the precisions were less than 1% which indicate that the proposed method is well validated and suitable for quantitatively detecting curcuminoids and quercetin simultaneously in pharmaceutical products, herb materials and various turmeric and quercetin containing products.\n\nSystem suitability testing is important to ensure the performance of the system before and during the analysis. As defined in the United States Pharmacopeia/National Formulary (USP/NF) \\[[@B054]\\] system suitability parameters were established as a direct result of the ruggedness and the robustness of the experiments. The system suitability testing proved that the proposed method will allow the separation of all four anaytes and will produce satisfactory peak shapes.\n\n![Chromatograms for Chinese medicinal plant extracts (a) containing quercetin and (b) containing curcuminoids. QUE, quercetin; BDMC, bisdemethoxycurcumin; DMC, demethoxycurcumin; CUR, curcumin.](2093-6966-v17-n04-036-g004){#F004}\n\n5. Conclusion\n=============\n\nA simple isocratic RP-HPLC method with UV detection has been developed for simultaneous detection of quercetin, curcumin, demethoxycurcumin and bisdemethoxycurcumin. The analytes were well separated and detected within 19 minutes. This method was validated for specificity, linearity, precision, accuracy and robustness as per ICH guidelines. The data showed good selectivity and sensitivity, a wide linear range, precision and accuracy. The method was sensitive to HPLC conditions; that is, changes in the mobile phase's composition, the pH, the column temperature and the flow rate affected the retention time and response, but did not affected the separation of the compounds. In addition, each parameter showed good repeatability of the retention time and response. In conclusion, the proposed method is simple, easy and cost effective, no specific solvent is involved and it utilizes common HPLC instruments with UV detectors. Hence, this UV-HPLC method is suitable for routine analysis of quercetin and curcuminoid formulations or products.\n\n**Conflict of interest** The authors declare that there are no conflict of interest.\n"}
+{"text": "Introduction {#Sec1}\n============\n\nIt is estimated that diabetes affects 422 million adults worldwide \\[[@CR1]\\] and 3.7 million people in the UK (2016--2017 data) \\[[@CR2]\\]. Diabetic macular edema (DMO) is the leading cause of blindness in people of working age in developed countries, and an estimated 12% and 27% of patients with type 1 and type 2 diabetes mellitus, respectively, have DMO \\[[@CR3]\\]. DMO occurs due to impairment in the blood retinal barrier and increased leakage from blood vessels due to the loss of supportive cells, cell dysfunction, and inflammatory changes \\[[@CR4]\\]. One key regulator in this process is vascular endothelial growth factor (VEGF), which has been successfully targeted with anti-VEGF monoclonal antibodies such as ranibizumab, aflibercept and bevacizumab. Corticosteroids provide an alternative therapeutic strategy by blocking leukotriene and prostaglandin synthesis via glucocorticoid receptors and subsequently acting on the arachidonic acid pathway. In addition, corticosteroids inhibit and interfere with other pro-inflammatory molecules, such as VEGF-alpha, interleukin-6 and intercellular adhesion molecule-1, and increase vasoconstriction by nitric oxide inhibition \\[[@CR4]--[@CR8]\\].\n\nThe tradition clinical treatment of DMO was largely dependent on macular laser (either focal or 'grid' distribution) treatment, supplemented with unlicensed short-acting steroid injections. Over the last decade care has been transformed by a shift to intravitreal anti-VEGF injections \\[[@CR9]\\], with or without macular laser treatment, as the primary treatment for DMO. There are however a significant proportion of patients who have insufficiently responsive DMO despite anti-VEGF treatment, for whom other treatment options are required.\n\nThis article does not contain any studies with human participants or animals performed by any of the authors.\n\nBrief Overview of the Iluvien\u2122 Fluocinolone Acetonide Implant {#Sec2}\n=============================================================\n\nThe Iluvien\u2122 fluocinolone acetonide (FAc) 0.19\u00a0mg intravitreal implant (Alimera Sciences Ltd., Alpharetta, Georgia, USA) is a slow-release preparation containing fluocinolone acetonide 0.19\u00a0mg (C24H30F2O6) \\[[@CR8]\\]. The implant is injected intravitreally and maintains a slow release of 0.2 \u03bcg per day for up to 36\u00a0months. The FAMOUS study demonstrated sustained release, as assessed by the concentrations of FAc in the aqueous humor, with levels of slightly more than\u20092\u00a0ng/ml for approximately the first 3\u00a0months followed by a maintained concentration of 0.5--1.0\u00a0ng/ml through a period of 36\u00a0months \\[[@CR10]\\].\n\nOverview of the FAME Trials {#Sec3}\n===========================\n\nEfficacy {#Sec4}\n--------\n\nThe FAc 0.19\u00a0mg implant was approved for treatment of DMO after it was shown to be effective for the treatment of DMO in the Fluocinolone Acetonide in diabetic Macular Edema trials (FAME A and B studies) \\[[@CR11], [@CR12]\\], which were two parallel multicenter, randomized, sham-injection-controlled trials evaluating the safety and efficacy of the FAc implant in DMO.\n\nParticipants with persistent DMO and at least one previous macular laser treatment were randomly assigned to receive high-dose FAc (0.5\u00a0\u00b5g/day) (*n*\u2009=\u2009393), low-dose FAc (0.2\u00a0\u00b5g/day) (*n*\u2009=\u2009375), or sham injection (*n*\u2009=\u2009185; sham group) in the ratio 2:2:1. All participants were eligible for rescue laser after 6\u00a0weeks from baseline injection. The primary efficacy endpoint was percentage of patients with best corrected visual acuity (BCVA)\u2009\u2265\u200915 letter (L) gain from baseline at 24\u00a0months using the early treatment diabetic retinopathy study (ETDRS) chart. Secondary outcomes were visual function and foveal thickness. An additional study drug or sham could be given after 1\u00a0year if study retreatment criteria were met. During the study clinicians were able to give 'off protocol' therapies, such as anti-VEGF, at their own discretion, and these patients were not excluded from the trial or analysis.\n\nIn both FAc implant groups, there was a significantly higher percentage of participants with\u2009\u2265\u200915L gain compared to\u00a0the control at month 24: 28.7% in the low-dose treatment arm, 28.6% in the high-dose treatment arm and 16.2% in the sham treatment arm (*p*\u2009=\u20090.002); at 36\u00a0months the percentage was 28.7% (low dose), 27.8% (high dose) and 18.9% (sham) (*p*\u2009=\u20090.018). When only those left in the trial at month 36 were considered in the analysis, the percentage increased to 33.0, 31.9 and 21.4%, respectively (*p*\u2009=\u20090.030). Benefit was seen in both dosages at week 3 and at all subsequent time points. Mean letter gain at month 24 was 4.4L in the low-dose group, 5.4\u00a0L in the high-dose group and 1.7L in the sham group (*p*\u2009=\u20090.02 and *p*\u2009=\u20090.016 vs. sham). This benefit was maintained at 3\u00a0years, with a mean letter gain at month 36 of 5.3L in both FAc groups and 2.0 in the sham (*p*\u2009\u2264 \u00a00.018).\n\nIn a pre-planned subgroup analysis looking at patients with chronic DMO, the percentage with a\u2009\u2265\u200915L gain at month 36 compared to sham was doubled: 34.0% in the low-dose group (*p*\u2009\\<\u20090.001), 28.8% in the high-dose group (*p*\u2009=\u20090.002) and 13.4% in the sham group, suggesting more significant improvement in the more chronic cases \\[[@CR13]\\].\n\nAdverse Effects {#Sec5}\n---------------\n\nThe main ocular adverse effects revealed by the FAME trials were the development of visually significant cataract and rise in intraocular pressure (IOP) or glaucoma. The development of cataract and the percentage of participants requiring cataract surgery were significantly higher in both the low- and high-dose treatment arms than in the control (sham) group. In the low-dose group 81.7% patients developed cataracts by month 36, with 80.0% undergoing cataract surgery; in the high-dose group, 88.7% developed cataracts, with 87.2% undergoing surgery; in the sham group, 50.7% developed cataracts, with 27.3% undergoing surgery. While there was a higher proportion requiring cataract surgery in the FAc groups than in the sham group, visual improvement post cataract surgery was similar to that of the baseline pseudophakic patients, and BCVA at 36\u00a0months was similar in those phakic and pseudophakic patients at baseline \\[[@CR14]\\].\n\nA rise in IOP is a major concern to clinicians who are considering the use of intraocular corticosteroids and, therefore, the IOP profile of the FAc implant revealed in the FAME trials was of particular interest. Overall IOP-related adverse events were more common in the FAc groups than in the sham group. At 36\u00a0months the proportion of patients who had had elevated IOP requiring at least 7\u00a0days of topical IOP-lowering drops was 38.4% for the low-dose treatment group and 47.3% for the high-dose treatment group, compared to 14.1% in the sham group. An IOP of\u2009\u2265\u200930\u00a0mmHg developed in 16.3% of participants in the treatment groups by month \u00a024 and in 18.4% by month 36. The development of elevated IOP that required incisional surgery at month 24 and 36 was 1.6 and 4.8% in the low-dose group, respectively, 6.4 and 8.1% in the high-dose group, respectively, and 0.5% in the sham group. It should be noted that these rates for incisional surgery at 24\u00a0months are based on the trial data on file; they are lower than the rates published by Campochiaro et al. as their analysis included some cases with a follow-up between 24 and 36\u00a0months due to the trial analysis not being truncated at month 24 \\[[@CR11], [@CR12]\\].\n\nPatients who had been treated with corticosteroid previously were eligible to be included in the study provided that they did not have a prior history of corticosteroid-induced ocular hypertension (sometimes termed 'steroid responders'). A further post hoc analysis of patients who had previously received corticosteroid treatment and were thus known to be 'non-responders' (*n*\u2009=\u200972) found that none of this group of patients required incisional surgery to lower IOP, whereas 6.1% of those who had not received prior corticosteroid, required IOP-lowering surgery (*n*\u2009=\u200918) \\[[@CR15]\\]. Further assessment of changes in the optic nerve head using the fundus photographs used in the FAME trials concluded that treatment with FAc was not associated with significant glaucomatous changes with or without elevated IOP \\[[@CR16]\\]. Endophthalmitis occurred in 0.2% (2 eyes) with the FAc implant compared to none in the sham group.\n\nOverall the trials showed that both the low-dose and high-dose FAc implants produced a significant improvement in BCVA over 3\u00a0years. The benefit-to-risk ratio showed the low-dose implant to be superior to the high-dose implant, and the authors of the trials concluded that the former would provide a substantial visual benefit and provide a valuable addition to treatment options available in managing patients with DMO. Following these pivotal studies, the FAc 0.19\u00a0mg implant (0.2\u00a0\u00b5g per day) was licensed for use in the USA and Europe \\[[@CR17]--[@CR20]\\]. Treatment is the UK is guided by the National Institute for Health and Care Excellence (NICE) (UK) Technology Appraisal (TA301) which restricts its use to the treatment of insufficiently responsive DMO (defined as inadequate response to conventional therapy \\[laser and/or anti-VEGF\\], either no reduction in central retinal thickness or minimal reduction from treatment and a persistence in macular edema) only where the eye is pseudophakic and where the manufacturer provides the FAc implant at the discount agreed in the patient access scheme \\[[@CR17]\\].\n\nReal World Data and Long-Term Safety Outcomes {#Sec6}\n=============================================\n\nTo date there have been numerous papers looking at the real world efficacy and safety profile of the FAc implant at years\u00a01 and 2 and data is\u00a0emerging for 3\u00a0years. Since the FAME trials the face of DMO management has changed, with anti-VEGF becoming the predominant first-line treatment in most areas; consequently most published studies have looked at clinical efficacy, side effect profile and use of the FAc implant when the patient is insufficiently responsive to the current gold standard treatment. An overview of these publications is given in Table\u00a0[1](#Tab1){ref-type=\"table\"}.Table\u00a01Summary of real world data publications for the fluocinolone acetonide 0.19\u00a0mg (FAc 0.19\u00a0mg) implantFirst authorType of studyNumber of eyes (number of participants)Duration of follow-up (months)Mean CRT reduction (\u00b5m)Mean BCVA gain at last visit (ETDRS letters)IOP-related adverse events (*n*))Cataract formation requiring surgery (*n*))^a^Endophthalmitis (*n*))Bertelmann 2015 \\[[@CR21]\\]Case report2 (1)13\u00a0months for eye 16\u00a0months for eye 2270 for eye 1139 for eye 25L for eye 1 and 10L for eye 21/2 (managed with drops)1/20Elaraoud 2016 \\[[@CR22]\\]Retrospective cohort study22 (22)3148.96.4L1/22 (managed with drops)n/a0Elaraoud 2016 \\[[@CR23]\\]Case series12 (6)6296.910L0n/a0Elaraoud 2016 \\[[@CR24]\\]Case series10 (5)12357.910.5L0n/a0Massin 2016 \\[[@CR25]\\]Prospective study comparing prior laser (group 1) vs prior laser and anti-VEGF (group 2)17 (16)Group 1: 7 (6)Group 2: 10 (10)12Group 1: 299Group 2: 251Group 1: 5.6LGroup 2: 0.9L3/17 (managed with drops)1/17 (1/5 phakic eyes)0Alfqawi 2017 \\[[@CR26]\\]Retrospective cohort study28 (23)121988L3/28 (managed with drops)n/a1/28Figueira 2017 \\[[@CR27]\\]Prospective phase 4 study (RESPOND)12 (12)122923.7L5/12 (managed with drops)1/12 (1/4 phakic eyes)0El-Ghrably 2017 \\[[@CR28]\\]Retrospective cohort study57 (57)46 with 3\u00a0months of data, 34 with 6\u00a0months of data, 22 with 12\u00a0months of data3, 6, 123\u00a0months: 1026\u00a0months: 11712\u00a0months: 1263\u00a0months: 5.8L6\u00a0months: 6.7L12\u00a0months: 5.1L6/57 (managed with drops)n/a0Holden/Currie 2017 \\[[@CR29], [@CR31]\\]Retrospective cohort study (ICE-UK)233 (208)12153 (median)4.3L (median)29/233 (managed with drops).6/233 (6/26 phakic eyes)0Quhill 2016 \\[[@CR32]\\]Case report1 (1)2450715L0n/a0Bailey 2017 \\[[@CR9]\\]Retrospective cohort study345 (305)120 with 18\u00a0months of data/53 with 24\u00a0months of data14 (120 eyes had 18 months and 53 had 24\u00a0months)954.5L at 18\u00a0months/5.3L at 24\u00a0months48/345 (managed with drops)1/345 required trabeculectomyn/an/aFusi-Rubiano 2018 \\[[@CR33]\\]Retrospective cohort study29 (27)22 with 24\u00a0months of data/6 with 36\u00a0months of data26 (22 eyes had 24\u00a0months and 6 had 36\u00a0months)114 (103 at 24\u00a0months/65 at 36\u00a0months)6.5L at 24\u00a0months/11L at 36\u00a0months2/29 (managed with drops)n/a0*CRT* Central retinal thickness, *BCVA* best corrected visual acuity, *ETDRS* early treatment for diabetic retinopathy study, *IOP* intraocular pressure^a^n/a: \u2009All eyes are pseudophakic or underwent combined cataract/implant at baseline\n\nBertelmann et al. reported a case of bilateral FAc implant in a phakic patient with refractory DMO of over 20\u00a0years duration (left eye at baseline, right eye at month 6), that resulted in an improvement in retinal thickness and BCVA, with a reduction in central retinal thickness (CRT) in the left eye from 642 to 372\u00a0\u00b5m at month 13, and from 473 to 334\u00a0\u00b5m in the right eye at month 6 \\[[@CR21]\\]. Early outcomes were also described by Elaraoud et al. in 22 eyes across three sites in the UK \\[[@CR22]\\]. These showed a mean reduction in CRT of 148.9\u00a0\u00b5m at 3\u00a0months and mean visual gain of 6.4L. Data at 6 and 12\u00a0months were also reported in another paper by Elaraoud et al. in patients with bilateral implants and DMO, with a mean increase in BCVA of 10L at 6\u00a0months and of 10.5L at 12\u00a0months, as well as sustained CRT reduction at months 6 and 12 \\[[@CR23], [@CR24]\\].\n\nThe results of a prospective phase IV study in France by Massin et al. \\[[@CR25]\\] which enrolled patients with chronic DMO further supported the safety and efficacy data on the FAc implant for up to 12\u00a0months of follow-up. Patients included in the study were given the FAc implant at baseline, and clinical outcomes were monitored for over 1\u00a0year. Study participants were separated into two groups: chronic DMO insufficiently responsive to laser (group 1) with or without anti-VEGF (group 2). The patients were excluded from the study if they had one or more of the following: IOP of \\> 21\u00a0mmHg at screening; history of steroid response; need for \\> 2 IOP-lowering drops at screening; pre-existing glaucoma. Of the 16 patients, six patients (7 eyes) were in group 1 and ten (10 eyes) were in group 2. Twelve eyes were pseudophakic. Eyes of patients in group 1 had a significantly longer duration of DMO than those of patients in group 2 (7.6 vs 3.6\u00a0years, respectively). A reduction in CRT was seen at week 1, and the mean CRT reduction was 239 and 147\u00a0\u00b5m at month 1 for groups 1 and 2, respectively. This reduction was maintained at month 6 (281 \\[group 1\\] and 167\u00a0\u00b5m \\[group 2\\]), month 9 (295 \\[group 1\\] and 172\u00a0\u00b5m \\[group 2\\]) and month 12 (299 \\[group 1\\] and 251\u00a0\u00b5m \\[group 2\\]). There was a mean gain in BCVA at all time points, with a gain of 2.0L and 6.4L in groups 1 and 2, respectively, at month 1, and gain at month 12 of 5.6L and 0.9L, respectively. Four patients required top-up therapy. The IOP increased in three eyes, and the increase was controlled with IOP-lowering drops alone. Better CRT reduction and BCVA gain at month 12 were seen in the patients in group 1, which had the more chronic DMO, thereby supporting the FAME data.\n\nThe results of a study by Alfaqawi et al. supported the 12-month data of Massin et al. \\[[@CR25]\\]. These authors reported outcomes from a UK cohort, looking at 28 eyes treated with the FAc implant at a single center \\[[@CR26]\\]. There was an overall mean gain in BCVA at month 12 of 8L, with 25% (*n*\u2009=\u20097) of patients gaining at least 15L and 36% (*n*\u2009=\u200910) gaining at least 10L. The mean CRT decreased by 198\u00a0\u00b5m at 12\u00a0months. The authors noted that two eyes required additional anti-VEGF after 10\u00a0months and that three eyes developed IOP that was managed with drops alone. One eye developed vitreous hemorrhage and one eye developed endophthalmitis.\n\nThe RESPOND trial in Portugal, an open label, phase IV, multicenter, non-randomized trial involving 12 patients (4 phakic, 8 pseudophakic) showed rapid improvement in CRT and gain in BCVA after 1\u00a0week which were sustained at all time points up to 12\u00a0months. Mean visual gain was 3.7L across all patients, with greater improvement in the pseudophakic eyes (increase of 6.8L) compared to phakic eye (decrease of 2.5L) \\[[@CR27]\\]. Two patients developed a rise in IOP, which was managed with drops alone. A limiting factor of this study is the lack of BCVA analysis post cataract surgery which does not allow full assessment of the efficacy of the FAc implant on vision. In a subsequent UK study, El-Ghrably et al. showed a mean reduction in CRT of 126\u00a0\u00b5m and BCVA gain of 5.1L at month 12 in patients \\[[@CR28]\\]. Eyes that developed an increased IOP were managed with topical IOP-lowering drops alone.\n\nThe main limitation of all these early outcome studies is limited numbers; however clinical efficacy and adverse effects appear to be comparable to those of the FAME trials.\n\nMore recently, the Iluvien Clinical Evidence Study (ICE-UK) study evaluated UK visual outcomes and IOP in eyes treated with the FAc implant between April 2013 and April 2015 at 13 ophthalmology centers in the UK \\[[@CR29]\\]. The study evaluated 12\u00a0months pre- and post-FAc implant results for 208 patients (233 eyes), of which 89% (*n*\u2009=\u2009207 eyes) were pseudophakic at time of implant. Overall, by month 12, the improvement in visual acuity (VA) was at least 5L in 44% of patients, at least 10L in 30% of patients and at least 15L in 18% of patients. A significant improvement in VA was seen at all time points post implant in those with a baseline VA of \\> 55L; however, a non-significant improvement was seen in those with a baseline VA of \\< 55L.\n\nIn the 12\u00a0months prior to implant, 191 (82%) eyes had had at least 1 anti-VEGF injection, 101 eyes (43%) had received steroid treatment and 146 eyes (63%) had at least one macular laser treatment. By month 12, adjunctive treatments had been given to 69 eyes (30%). Five eyes required cataract surgery within the first 3\u00a0months. IOP-lowering drops were required in 15% of patients with no prior history; none required surgical management.\n\nFurther analysis of fellow eyes in the ICE-UK study showed that a letter gain of at least 15L was achieved by more often by treated eyes than by fellow eyes (18 vs. 4%) and that the mean reduction in CRT was 113\u00a0\u00b5m in treated eyes compared to \u2212\u00a013\u00a0\u00b5m in fellow eyes \\[[@CR30]\\]. Comparison of retinal thickness in these patients revealed that eyes with a CRT of at least 400\u00a0\u00b5m or greater were significantly more likely to develop at least a 10, 25 and 50% reduction in CRT compared to those with CRT of \\< 400\u00a0\u00b5m \\[[@CR31]\\].\n\nAlthough this study had large numbers, there were numerous limitations, including the retrospective nature of the study, incomplete data sets and inconsistent documentation in medical records. These were acknowledged by the authors who made attempts to minimize their effect.\n\nLonger Term Real World Data {#Sec7}\n===========================\n\nExtended clinical benefit from the FAc implant from 12 to 24\u00a0months was first reported by Quhill et al. in a pseudophakic patient treated with a FAc implant for DMO unresponsive to laser and anti-VEGF treatments \\[[@CR32]\\]. These authors noted an initial rapid reduction in CRT of 507\u00a0\u00b5m at 1\u00a0week and gain in BCVA of 15L which was maintained up to 24\u00a0months. No top-up treatment was required and there was no recurrence of fluid.\n\nIn their large cohort study, Bailey et al. looked at visual and safety outcomes for up to 24\u00a0months in 345 eyes treated with the FAc implant at 14 sites across the UK using data extraction from a single electronic medical records system \\[[@CR9]\\]. Of the patients, 91% had received prior therapy for DMO, with 28.4% having had prior macular laser therapy and 84.6% having had prior anti-VEGF treatment. The mean gain in BCVA was 4.5L at 18\u00a0months (*n*\u2009=\u2009120) and 5.3L at 24\u00a0months (*n*\u2009=\u200953). A minimum gain of 15.0L had been achieved by 15% of eyes at 12 months, by 15% of eyes at 18\u00a0months and by 20.8% of eyes at month 24. Mean CRT at baseline and last observation was 451.2\u00a0and 355.5\u00a0\u00b5m, respectively. The percentage of eyes requiring treatment post FAc implant was reduced, with the number of eyes requiring anti-VEGF treatment, laser treatment and steroid treatment reduced by 51.3, 2.1 and 2.5%, respectively. Regarding the IOP, 13.9% of eyes required IOP-lowering drops, 7.2% had an IOP of \\> 30\u00a0mmHg, and 0.3% required surgery to manage IOP and had a prior history of IOP event. Prior to the FAc implant, 14.2% of eyes were receiving IOP-lowering medication at baseline. In eyes previously treated with steroid and no IOP problems, there were no new IOP events. Despite large numbers, data extraction from electronic medical record systems relies on complete and accurate data entry and due to the retrospective nature of analysis, missing data could not be obtained, which is a limitation of the study.\n\nFusi-Rubiano et al. looked at real world 2-year outcomes in a UK pseudophakic cohort and found an improvement in BCVA and reduction in CRT with the FAc implant that was sustained at 1 and 2\u00a0years \\[[@CR33]\\]. Additional 3-year data in this publication was broadly similar to the longer term results seen in the FAME trials, with 50% (*n*\u2009=\u20093) improving by 15L. Supplementary treatment was needed in most cases; however, the overall burden of treatment had been reduced. Only two eyes required IOP-lowering drops in this cohort, and none required incisional surgery. Although limited by small numbers, especially for the 36-month outcomes, this study does show some of the first results for a 3-year follow-up.\n\nVitrectomized Eyes {#Sec8}\n==================\n\nSince the advent of intravitreal agents there has been discussion over their use in vitrectomized eyes, particularly in terms of pharmacokinetics and drug clearance. Results from some studies using animal models have suggested that anti-VEGF intravitreal agents have a shorter half-life and are cleared more quickly in vitrectomized eyes \\[[@CR34]\\]. In contrast, analysis of 3-year data in a DRCR.net trial of ranibizumab and prompt or deferred laser treatment found that there was no significant difference in long-term outcome at annual time points in vitrectomized eyes when compared to non-vitrectomized eyes \\[[@CR35]\\]. Although the results of this study are not directly comparable to those of two studies looking at sustained-release dexamethasone \\[[@CR36], [@CR37]\\], it may be worth noting that these studies found no statistically significant difference in the vitreoretinal pharmacokinetics between vitrectomized and non-vitrectomized eyes \\[[@CR36]\\] and showed good clinical efficacy in vitrectomized eyes \\[[@CR37]\\].\n\nIn one of the first publications looking solely at the treatment of vitrectomized eyes with the FAc implant, Kumar et al. described two cases of vitrectomized eyes previously treated with anti-VEGF and triamcinolone that showed complete resolution of DMO up to 1\u00a0year post implant \\[[@CR38]\\]. Meireles et al. looked at the efficacy and safety of the FAc implant in 26 eyes with prior vitrectomy from six European countries \\[[@CR39]\\]. Mean time to insertion was 24.2\u00a0months post vitrectomy, and mean follow-up time was 255\u00a0days. Mean gain in BCVA was 11.7L, and there was a mean reduction in CRT of 233.5\u00a0\u00b5m. There was a mean rise of IOP from baseline of 1.4\u00a0mmHg, and eight eyes required IOP-lowering drops or had been on them prior to FAc implant. The authors concluded that the FAc implant was safe and effective in vitrectomized eyes.\n\nIn a further retrospective analysis of 43 eyes, comparing vitrectomized (*n*\u2009=\u200924) and non-vitrectomized (*n*\u2009=\u200919) eyes treated with FAc implant for DMO, Pessoa et al. found a gain of at least 15L in 37.5% of vitrectomized eyes (group 1) and in 36.8% of non-vitrectomized eyes (group 2) \\[[@CR40]\\]. In addition, there was a mean reduction in CRT from baseline of 217.7 and 155.6\u00a0\u00b5m in groups 1 and 2, respectively; however there was no statistical difference between the two groups. There was no significant difference for IOP and the need for IOP-lowering medications between the groups. Although these results appear to show no significant difference between the two groups, the mean follow-up was 8.1\u00a0months, and thus further long-term follow-ups and studies are needed.\n\nAll of these studies are limited by small numbers and their retrospective nature. There is limited evidence for real world outcomes in vitrectomized eyes, and further studies are needed in these patients.\n\nDespite good outcomes, care should be taken in vitrectomized eyes. The migration of the implant into the anterior chamber in patients who had had prior vitrectomy was seen in two eyes in the series by El-Ghrably et al., where posterior capsular defect had been identified previously \\[[@CR28]\\]. There were no long-term adverse effects for these eyes. Moisseiev et al. also reported a case of 'floater' in a patient with previous vitrectomy where the implant had centralized in the visual axis with vitreous attachment \\[[@CR41]\\]. It should be noted that YAG vitreolysis was required to break the adhesion, but this treatment resolved the patient's symptoms. In another case, Andreatta et al. reported a dislodgment of FAc implant into the infusion cannula during vitrectomy \\[[@CR42]\\].\n\nHow Does the FAc 0.19\u00a0mg Implant (Iluvien\u2122) Fit into Current Treatment Regimens, Patient Management and Guidelines {#Sec9}\n==================================================================================================================\n\nRegulatory Approvals {#Sec10}\n--------------------\n\nIn 17 European countries the FAc 0.19\u00a0mg implant has been approved for the treatment of patients with DMO insufficiently responsive to available therapies \\[[@CR17]--[@CR20]\\]. In some countries, such as the UK, Italy and Spain, reimbursement is restricted for use only in eyes with an intraocular (pseudophakic) lens \\[[@CR17], [@CR19], [@CR20]\\]. In the USA, the implant is approved for patients with DMO who have had prior treatment with corticosteroid and found not to have a clinically significant rise in IOP \\[[@CR18]\\]. It is contraindicated in patients with confirmed or suspected ocular or periocular infections, patients with glaucoma with a CDR\u2009\u2265\u20090.8 and patients with a known hypersensitivity to any component of the implant \\[[@CR18]\\].\n\nDMO is multifactorial disease and as such there should be a variety of treatment options available to tackle it. Anti-VEGF is now the gold standard in the treatment of DMO; however, not all patients respond or are able to maintain monthly anti-VEGF treatments. In the study of Gonzalez et al. \\[[@CR43]\\], almost 40% of patients treated with anti-VEGF had a minimal response (less than 5L gain) in BCVA after 3\u00a0months, and this was associated with worse long-term BCVA. The FAc implant as a corticosteroid is an effective treatment option for the management of DMO and, as suggested by the European Society of Retinal Specialists (EURETINA), should be considered for patients lacking a response to current first-line treatments \\[[@CR44]\\]. These guidelines also advise that retreatment with FAc implant can be considered after 1\u00a0year. Although the FAc 0.19\u00a0mg implant IS particularly appropriate for pseudophakic patients, phakic patients are being treated but should be advised of the high risk of cataract formation.\n\nWill It Be Cost Effective and Beneficial for My Patients? {#Sec11}\n=========================================================\n\nCost-effect analyses have shown the FAc implant to be a cost-effective treatment in insufficiently responsive DMO in both phakic and pseudophakic patients \\[[@CR45], [@CR46]\\] because of its long-acting efficacy, safety profile and tolerability. The requirement for adjunctive treatment post implant within 3\u00a0years is lower than that for average anti-VEGF treatment regimes \\[[@CR9], [@CR33]\\]. A recent comparative study of anti-VEGF treatments in DMO reported that a median of nine injections of aflibercept and ten injections of ranibizumab/bevacizumab were required to control DMO over a 12-month period \\[[@CR47]\\].\n\nIn a UK study, Quhill and Beiderbeck undertook a cost analysis of the FAc implant looking at the overall cost of treatment with the FAc implant and one additional ranibizumab treatment per year, compared to ranibizumab as needed, using an average of 14 lucentis\u00a0treatments over 3\u00a0years \\[[@CR46]\\] (based on the RESTORE extension study \\[[@CR48]\\]). These authors calculated a cost saving of around \u00a36068 in pseudophakic eyes using the FAc implant and \u00a35341 in phakic eyes (taking into account the added cost of cataract surgery). Additionally, reducing the need for treatment is an important benefit for patients, as the number of injections has been shown to affect quality of life, with a higher frequency causing an increase in work leave and anxiety \\[[@CR49]\\]. In addition, vital treatment time and workload are reduced in already overstretched ophthalmology departments.\n\nConclusion {#Sec12}\n==========\n\nThe FAc 0.19\u00a0mg implant (Iluvien\u2122) is a licensed treatment for use in DMO and has been shown to be effective, to have good tolerability and to be a cost-effective option for insufficiently responsive DMO, with a long-acting duration up to 36\u00a0months in both vitrectomized and non-vitrectomized eyes. It should be considered as a second-line option for treatment of DMO in patients whose eyes do not sufficiently respond to first-line treatment (unless otherwise contraindicated). It is also a valuable alternative in cases where anti-VEGF treatment is not a viable option. Clinicians may consider evaluating 'IOP steroid responsiveness' prior to treatment with FAc implant to reduce the risk of IOP-related events in patients. Ongoing studies will help to further define the patients with DMO most likely to benefit from the FAc implant, but this implant already has a valuable place in our treatments for this sight-threatening disease.\n\n**Enhanced digital features**\n\nTo view enhanced digital features for this article go to 10.6084/m9.figshare.7022222.\n\n**Change history**\n\n2/7/2020\n\nThis article was originally published with the copyright: the author(s) 2018, CC-BY-NC. However, the license should be the author(s) 2018, CC-BY.\n\n**Change history**\n\n2/7/2020\n\nThis article was originally published with the copyright: the author(s) 2018, CC-BY-NC. However, the license should be the author(s) 2018, CC-BY.\n\nDuring the peer review process, the manufacturer of the agent under review was offered an opportunity to comment on the article. Changes resulting from the comments received were made by the author based on their scientific and editorial merit.\n\nFunding {#FPar1}\n=======\n\nNo funding or sponsorship was received for this study or publication of this article.\n\nAuthorship {#FPar2}\n==========\n\nAll named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.\n\nDisclosures {#FPar3}\n===========\n\nWilliam Fusi-Rubiano has received sponsorship from Alimera to attend Euretina 2017. Rupal Morjaria has received sponsorship from Alimera to attend Euretina 2017. Mark Lane and Rebecca R. Blow have no disclosures to declare. Alastair K. Denniston receives a proportion of his funding from the Department of Health's NIHR Biomedical Research Centre for Ophthalmology at Moorfields Eye Hospital and UCL Institute of Ophthalmology, and from the Wellcome Trust, through a Health\u00a0Improvement Challenge grant (200141/Z/15/Z). The views expressed in the publication\u00a0are those of the author and not necessarily those of the\u00a0Department of Health.\n\nCompliance with Ethics Guidelines {#FPar4}\n=================================\n\nThis article does not contain any studies with human participants or animals performed by any of the authors.\n\nOpen Access {#d29e1171}\n===========\n\nThis article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (<http://creativecommons.org/licenses/by-nc/4.0/>), which permits any noncommercial use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.\n"}
+{"text": "INTRODUCTION {#s1}\n============\n\nBenthic foraminifera (eukaryotic single-celled organisms) have existed since the Paleozoic and colonize a variety of marine sediments, ranging from shallow water environments to the deep sea. Many species produce an inorganic shell (test), which is preserved in accumulating sediment layers after their death. Based on these attributes, benthic foraminifera serve as important tools for paleoenvironmental reconstructions, including estimations of organic matter fluxes in marine sediments and paleoproductivity studies (e.g. [@BIO030056C1]; [@BIO030056C29]). They mainly feed on microalgae or phytodetritus (e.g. [@BIO030056C7]; [@BIO030056C19]; [@BIO030056C39]) and their response to variations in phytodetritus fluxes can be observed on foraminiferal community compositions ([@BIO030056C27]) or metabolic activities ([@BIO030056C22]; [@BIO030056C41]). However, there is still a lack of studies quantifying foraminiferal carbon and nitrogen metabolism to improve the understanding of the connection between foraminiferal biomass and sediment organic matter fluxes.\n\nHigh organic matter fluxes are characteristic for estuarine environments, where planktonic primary production can reach a gross particulate production of up to 176\u2005g\u2005C\u00a0m^\u22122^\u00a0year^\u22121^ ([@BIO030056C67]). These fluxes show strong seasonal changes in quantity, characterized by high fluxes (500--2000\u2005g\u2005C\u00a0m^\u22122^\u00a0day^\u22121^) in summer months, typically ending with a high productivity peak in autumn, followed by a quick decrease towards winter months where they decline to zero. Microphytobenthos is additionally present throughout the year, with a strong reduction of abundant species in winter ([@BIO030056C62]). In such high productivity environments, foraminifera play an important role in sediment nutrient fluxes and phytodetrital carbon and nitrogen processing ([@BIO030056C45]; [@BIO030056C73]). For example, [@BIO030056C45] have evaluated that the genus *Ammonia* contributes with 1--7% to the processing of newly introduced, fresh phytodetrital carbon in an estuarine environment. Further, [@BIO030056C10] have determined that the foraminiferal aerobic carbon mineralization in intertidal mudflats can reach up to 7% of the total diffusive oxygen uptake.\n\n*Ammonia tepida* is a common and dominant foraminiferal species in shallow water environments, showing seasonal oscillations in population densities and in size fractions ([@BIO030056C2]; [@BIO030056C49]). There is a high genetic variability within the *A.* *tepida* species-complex*,* and several molecular types have been previously identified ([@BIO030056C25]; [@BIO030056C64])*.* Morphological traits resemble those of molecular type T6 ([@BIO030056C25]).\n\n*Ammonia* is an effective grazer of microalgae and its physiology, successful reproduction and survival is driven by factors including food availability, food quality or environmental temperature ([@BIO030056C5], [@BIO030056C6], [@BIO030056C7]; [@BIO030056C39]). Laboratory feeding experiments can provide valuable information on their response to different food availability modes and were applied in this study to investigate phytodetritus intake of *A. tepida* under stable conditions. Specimens of *A. tepida* were fed in two different feeding modes (single, high quantity pulse versus constant feeding) and for the constant feeding at two different temperature cycles (18:20\u00b0C, 23:25\u00b0C, 12:12\u2005h). Labelled food material (^13^C and ^15^N) was used and the cytoplasmic content of phytodetrital derived carbon (pC) and nitrogen (pN), as well as fluctuations in total cytoplasmic organic carbon (TOC) or nitrogen (TN), were analyzed in *A. tepida.* The aim was to investigate the response of this species to variations in phytodetritus flux (quantity or quality) and temperature, which provides fundamental information on the species-specific feeding ecology and fluxes of TOC and TN through foraminiferal communities. Additionally, phytodetritus intake of *A. tepida* individuals in different size fractions (125--250\u2005\u00b5m, 250--355\u2005\u00b5m, \\>355\u2005\u00b5m) was compared. Smaller individuals of some other foraminiferal species take up more food particles than larger individuals and adult individuals show lower nutrient turnover rates ([@BIO030056C37]; [@BIO030056C53]). These intraspecific variations can be an important factor to understand carbon cycling on the seafloor and the role of foraminiferal nutrient cycling ([@BIO030056C52]). To date there exist no data on phytodetritus-derived carbon and nitrogen intake across intraspecific size classes of foraminifera. Therefore, a second feeding experiment was carried out to observe pC and pN intake within three size classes of *A. tepida* (125--250\u2005\u00b5m, 250--355\u2005\u00b5m, \\>355\u2005\u00b5m). For the comparison of the size-specific nutrient intake, pC and pN content were scaled with individual weight. Allometric scaling is widely used in theoretical biology to describe relationships of organismic mass or biovolume (X) with physiological parameters (Y), (here, pC and pN intake) according to the power function where 'a' is the scaling coef\ufb01cient and 'b' is the scaling exponent (log--log slope) ([@BIO030056C35]; [@BIO030056C69]; [@BIO030056C68]). Here, a decrease of weight specific individual nutrient intake, due to higher nutrient demand and increased metabolic rates of smaller individuals was expected (b\\<1). Since *A. tepida* is a dominant protist in intertidal sediments, the impact of size class distributions on carbon and nitrogen cycling within *A. tepida* populations was estimated. This was done by extrapolating carbon and nitrogen data from the feeding experiment to size-specific abundances of *A. tepida* within sediment cores collected in spring 2015 at the Elbe Estuary. With this study, we wanted to extend the knowledge on patterns of phytodetritus intake and size-related pC and pN contents of the abundant foraminiferal species *A. tepida,* to obtain a more detailed understanding of the feeding ecology of intertidal foraminifera and their role in sediment carbon and nitrogen fluxes.\n\nRESULTS {#s2}\n=======\n\nExperiment 1: Comparison of food uptake at a single high feeding pulse versus lower constant feeding {#s2a}\n----------------------------------------------------------------------------------------------------\n\nCulture conditions (O~2~, salinity, pH) were constant over time, but an increasing mortality of specimens was observed with time, so the initial total of 45 samples resulted in 37 elemental and isotope analyses. The individual values of pC or pN showed distinct patterns in the three approaches ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). With constant feed settings, the lowest intake was observed at day 2 at constant feed 2 (25:23\u00b0C). At day 4 values for pC and pN peaked, followed by a steady decrease thereafter. For constant feed 1 (20:18\u00b0C), the highest pC and pN values were observed at the end, on day 28. For the single pulse approach (20:18\u00b0C), a peak for pC and pN appeared on day 7. Cytoplasmic TOC and TN content of the analyzed specimen also showed some variations over the course of the treatments. Noticeable is the increase in TOC and TN at constant feed 2 (25:23\u00b0C) in comparison with the stable patterns at constant feed 1 (20:18\u00b0C) and the generally highest TOC and TN values for the single pulse approach at day 4 and 7. Fig. 1.**Comparison of pC and pN (phytodetritus C and N intake) and cytoplasmic TOC and TN (total) in *A. tepida* for the different experimental approaches.** Arrows on the top indicate food supply, small arrows=2\u2005mg, large arrow=6\u2005mg phytodetritus dry weight. Black=carbon data, grey=nitrogen data, *n*=number of total replicates for the specific sampling point. Temperature settings were 20:18\u00b0C for constant feed 1; 25:23\u00b0C for constant feed 2 and 20:18\u00b0C for single pulse.\n\nIn one of the replicates on day 28 in constant feed 1 (20:18\u00b0C), the number of *A. tepida* individuals increased from the initial 55 specimens to 88 individuals, including 35 individuals of small test size with intense green cytoplasmic coloration, determined as young individuals. Those 35 young individuals were pooled together in an individual sample ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}, triangles).\n\nThe maximum diameter of young individuals was 243\u2005\u00b5m (\u00b123) and for adults 375\u2005\u00b5m (\u00b172). The fractions of pC:TOC and specially pN:TN (percent), were higher in young individuals than in adults ([Fig.\u00a02](#BIO030056F2){ref-type=\"fig\"}). Generally, the fraction of pN:TN was considerably higher than pC:TOC in all samples ([Fig.\u00a03](#BIO030056F3){ref-type=\"fig\"}). Fig. 2.**Relative amounts of pC:TOC and pN:TN in percent for Experiment 1.** Constant feed 1 (20:18\u00b0C) day 2--7, *n*=3; day 14 and 28, *n*=1, constant feed 2 (25:23\u00b0C): day 2--7, *n*=3; day 14 and 28, *n*=1, single pulse (20:18\u00b0C): day 2--7, *n*=3; day 14, *n*=6).Fig. 3.**Individual content of foraminiferal cytoplasmic organic carbon (TOC) and nitrogen (TN) per size fraction analyzed in Experiment 2.** Circles show data for day 2 (*n*=3 for each size fraction and day), triangles show data for day 4 (*n*=3 for each size fraction and day).\n\nExperiment 2: Size specific cytoplasmic TOC and TN and phytodetritus intake {#s2b}\n---------------------------------------------------------------------------\n\nSize class specific contents of cytoplasmic TOC and TN were strongly increasing with size class, but did not show significant differences between day 2 and day 4. Values for TOC and TN data (day 2 and day 4) of Experiment 2 per size fraction are shown in [Fig.\u00a03](#BIO030056F3){ref-type=\"fig\"}. Individual phytodetritus intake as pC and pN per size fraction showed slight but not significant deviations between days ([Table\u00a01](#BIO030056TB1){ref-type=\"table\"}). Distributions of living *A. tepida* specimen as counted from sediment cores collected parallel with specimen for the feeding Experiment 2 showed highest numbers in the 125--250\u2005\u00b5m size at all three stations ([Table\u00a01](#BIO030056TB1){ref-type=\"table\"}). Stations 1 and 2 were characterized by high densities of living *A. tepida*, while Station 3 showed very low total abundances. Due to these very low numbers of individuals per sediment volume, the ratios of foraminiferal size fractions might be inaccurate for the core from Station 3. Station 3 was therefore excluded from extrapolations of carbon and nitrogen budgets. Extrapolated values to total abundances of living *A. tepida* individuals resulted in total population TOC \u223c50\u2005\u00b5g\u00a0cm^\u22123^, pC \u223c6.20\u2005\u00b5g\u00a0cm^\u22123^, TN \u223c8.8\u2005\u00b5g\u00a0cm^\u22123^ and pN \u223c1.10\u2005\u00b5g\u00a0cm^\u22123^ (Station 1) and TOC \u223c55\u2005\u00b5g\u00a0cm^\u22123^, pC \u223c7.90\u2005\u00b5g\u00a0cm^\u22123^, TN 10\u2005\u00b5g\u00a0cm^\u22123^ and pN \u223c1.50\u2005\u00b5g\u00a0cm^\u22123^ (Station 2).Table\u00a01.**Abundance of living *A. tepida* individuals**\n\nAssumptions for the application of general least squares (GLS) regressions for specific weight, TOC, TN, pC and pN were met. Weight-specific TOC content of *A. tepida* was relatively similar across size fractions ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}A, Carbon, b \u223c0.90), while the weight-specific TN content decreased with increasing size with a proportionality of \u223c0.75 ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}A, Nitrogen). Phytodetritus intake scaled similar for pC and pN per individual dry weight (proportionality between 0.55 and 0.57), but the weight-specific intake became relatively lower with increasing foraminiferal weight ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}B). Interestingly, the TN related pN uptake also showed a \u223c0.75 proportionality ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}C, Nitrogen). Fig. 4.**Size-related general least squares (GLS) regressions.** (A) Individual TOC and TN content per individual dry weight, from Experiment 1 (constant feed 1 at 20:18\u00b0C, day 28) and Experiment 2 (day 2 and day 4, *n*=3 per sampling day). (B) Carbon and nitrogen uptake rates per individual dry weight and (C) individual carbon and nitrogen intake per individual TOC and TN content. Equations show log back transformed scaling coefficients. All regressions were significant (*P*\\<0.01). Ferret diameter of adult individuals Exp. 1=375\u2005\u00b5m (\u00b172) and young individuals=243\u2005\u00b5m (\u00b123). Exp. 2 size ranges are mesh widths of sieves.\n\nExtrapolations of individual TOC, TN, pC and pN to natural populations of *A. tepida* of Station 1 and Station 2 (Station 3 was excluded due to low *A. tepida* abundances) show that the gross TOC and TN pool and phytodetritus processing of the *A. tepida* community in April 2015 was dominated by the smallest size fraction (125--250\u2005\u00b5m) at both stations with a specifically high pN intake at Station 2 ([Fig.\u00a05](#BIO030056F5){ref-type=\"fig\"}). Fig. 5.**Ratios of carbon and nitrogen pools in populations of *A. tepida***. Relative contributions of size fractions of *A. tepida* to total population TOC and TN pools and phytodetritus intake (pC, pN) at Station 1 and Station 2, as extrapolated from feeding Experiment 2 to relative abundance data from stained sediment cores.\n\nThe raw isotope and elemental data for both experiments can be found in [Table\u00a0S1](Table\u00a0S1).\n\nDISCUSSION {#s3}\n==========\n\nResponse to different feeding modes in *A. tepida* {#s3a}\n--------------------------------------------------\n\nGenerally, empirical studies show that foraminiferal feeding behavior and phytodetritus intake varies between species according to the primary productivity of the habitat or their adaptations ([@BIO030056C46]; [@BIO030056C72]; [@BIO030056C13], [@BIO030056C14], [@BIO030056C15]). Different strategies include the rapid response to pulses of food supply by high food uptake, or constant feeding at moderate food concentrations ([@BIO030056C1]; [@BIO030056C41]). The results of this study clearly show different metabolic responses of *A. tepida* to different feeding modes and temperatures as expressed in distinct patterns of cytoplasmic pC, pN, TOC and TN levels ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). Constant food supply at 20\u00b0C caused a prolonged uptake of phytodetritus at stable levels of TOC and TN throughout the experiment, indicating that the costs for metabolic maintenance were covered. But there was no increase in individual biomass over the course of time in this approach, therefore individual growth was not highly supported by the constant supply of *D. tertiolecta* detritus, at least at the adult stage. On the other hand, the single, high pulse of phytodetritus caused an increase in individual biomass, particularly in TOC ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). Differences in pC and pN between constant feed 1 (20:18\u00b0C) and the single pulse suggest different rates of food turnover or usage of cellular storage products. The decrease of TOC and TN at the end of the single pulse approach is most likely caused by exhaustion of such storage products and limited subsequent uptake of the aging and degrading food particles. A recent study with *A. tepida* fed with frozen diatoms also showed a comparable pattern of cytoplasmic carbon increase following a single food pulse of 587\u2005mg\u2005C\u00a0m^\u22122^ ([@BIO030056C40]). In that study, the individual carbon content showed a stronger increase than we found in the present study, which was also followed by a decrease in TOC after the seventh day of incubation. In contrast to the present study, the food derived ^13^C signal did not decrease simultaneously with the cytoplasmic carbon content after day seven. This is most likely related to the different breakdown dynamics of soft chlorophyte cells (the food source in this study) in contrast to the diatoms with their surrounding frustule used by [@BIO030056C40] and the conceivably slower turnover of diatom derived ^13^C label.\n\nA similar response -- a strong increase in individual carbon following a high pulse of food (1.1\u2005g\u2005C\u00a0m^\u22122^) with subsequent loss of cytoplasmic carbon -- was shown earlier in *Cribrostomoides subglobosum* from the Norwegian Sea ([@BIO030056C1]). Here, the individual organic carbon in *C. subglobosum* almost doubled within 2\u00a0days after food supply. This was interpreted as a quick response to the food pulse, resulting from the opportunistic feeding strategy of *C. subglobosum*, coherent with the ability of some foraminiferal species to take advantage of high food quantities and build up reserves as a response to fluctuating food availability ([@BIO030056C1]; [@BIO030056C21]; [@BIO030056C26]; [@BIO030056C30]; [@BIO030056C39]).\n\nOpportunistic foraminiferal species have been further identified based on their strong response to pulses of live algae, which triggered reproduction, or increased population densities in laboratory feeding experiments ([@BIO030056C28]). These observations were consistent with findings on population fluctuations in nature, where the increase in population densities coincided with phytoplankton blooms in these species ([@BIO030056C27]; [@BIO030056C34]). Considering the large test size of the young *A. tepida* individuals (\u223c243\u2005\u00b5m) in this study, on the last day of constant feed 1 at 20:18\u00b0C (Experiment 1), the reproduction most likely occurred randomly at the beginning of the incubation period. Laboratory grown juvenile specimens of *A. tepida* showed a test size of \u223c125\u2005\u00b5m (8 chamber stage) about 2\u00a0weeks after release from the parental individual ([@BIO030056C20]). However, the amount of newly released juveniles (35) indicates that more than one adult individual must have reproduced, since it was described that asexual reproduction results in the release of a maximum 12--24 calcified juveniles ([@BIO030056C65]).\n\n*A. tepida* is also considered as an opportunistic feeder with a strong response to food input ([@BIO030056C39]). Our study revealed a pattern of phytodetrital nutrient uptake consistent with such a strategy in the single pulse approach, with increased food uptake and individual biomass until day 7. In contrast, the constant input of lower amounts of food resulted in reduced metabolic activity (reduced gain of biomass, steady and low food intake compared with the high single pulse, [Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). Regulation of metabolic activity in *A.* *tepida* to survive under unfavorable environmental conditions has been reported recently ([@BIO030056C40]). Here, a lowering of metabolic activity to a minimum might have been induced by unfavorable nutrient availability. Typically, strategies including reduced metabolic activity followed by strong responses to high fluxes of food are observed in the deep sea, where food is a limiting factor ([@BIO030056C22]; [@BIO030056C41]). The intertidal is not considered nutrient limited and the results of this experiment show a flexibility in the response of *A. tepida* to different feeding modes.\n\nOther environmental factors, like temperature, can also have a strong effect on the rates of carbon and nitrogen intake and metabolism in *A. tepida*, as demonstrated by the strong variation of pC and pN patterns at constant feed 1 (20:18\u00b0C) and constant feed 2 (25:23\u00b0C) at the same levels of food supply ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). This can be explained by the strong effect of temperature on metabolic activity, specifically on respiration rates controlling foraminiferal carbon processing. Thermal response curves revealed a steep rise of oxygen consumption in *A. tepida* between 20\u00b0C and 30\u00b0C ([@BIO030056C7]) and in respiration of *A. tepida* which was substantially higher at 23\u00b0C compared to 7\u00b0C ([@BIO030056C10]). This effect of temperature has also been observed on the level of carbon processing in *A. tepida* in a former feeding experiment ([@BIO030056C73]). There, a single low food pulse (\u223c220\u2005\u00b5g\u00a0C\u00a0m^\u22122^) at three different temperatures (20\u00b0C, 25\u00b0C, 30\u00b0C) resulted in the highest levels of pC in *A. tepida* at 25\u00b0C. In the current study, the constant food supply caused a similar rapid rise in pC at 25\u00b0C, but only until the fourth day, and it dropped on the following days despite constant food supply. The former study was carried out in vials containing solely phytodetritus ([@BIO030056C73]), while in this study a heat sterilized sediment layer was present. Sediment composition reportedly influences locomotion speed in *A. tepida* ([@BIO030056C33]), while live prey and inorganic matter slow foraminiferal movement and high organic matter content increases locomotion speed. Therefore, the high content of inorganic particles in the sediment in Experiment 1 (in contrast to exclusively organic matter) probably decreased foraminiferal food particle accumulation by decreasing locomotion through the sediment. However, the effect of increased temperature is evident in both studies. Increased environmental temperatures might therefore alter the organic matter processing by *A. tepida* in its natural environment and cause shifts in intertidal food webs. The rapid decrease in pC and pN in constant feed 2 (25:23\u00b0C) in this study might also be explained by bacterial presence in the experimental setup. The test (calcified shell) of the species *Ammonia* is characterized by immersions surrounding the umbilical plug which are often filled with organic material and might therefore host associated bacteria. Despite careful cleaning of the specimens, bacteria can remain in these cavities and reproduce during the course of the experiment, especially at high temperatures. These bacteria can contribute to the mineralization of organic matter, or the phytodetritus in the experiment, and get isotopically enriched. Bacteria can also serve as a food source for *Ammonia* species ([@BIO030056C8]; [@BIO030056C37]; [@BIO030056C57]) and the ingestion of lower isotope enriched bacteria can cause a dilution of the isotope signal or lower the pC and pN content in the foraminifera.\n\nAltogether, this experiment showed clear adaptations of *A. tepida* to variations in food availability at the level of phytodetritus intake and internal carbon and nitrogen reserves of adult individuals. This implies a generalistic adaptation of *A. tepida* to different levels of food supply, with a strong sensitivity of food intake and processing to environmental temperature.\n\nThe mortality of the individuals in Experiment 1 was relatively high (see [Fig.\u00a0S1](Fig.\u00a0S1)), compared to other laboratory experiments with *Ammonia tepida* ([@BIO030056C18]; [@BIO030056C40]; [@BIO030056C50]). The culture conditions in Experiment 1 of this study were set up as close as possible to natural conditions and followed the literature documenting successful cultivation of intertidal foraminifera ([@BIO030056C4]; [@BIO030056C38], [@BIO030056C39]; [@BIO030056C47]). These suggest cultivation in filtered habitat seawater, to ensure the supply of micronutrients and specific associated bacteria, which are necessary for growth and reproduction. It is known that ecological strategies involving the production of juveniles in large quantities implicate a high mortality rate during ontogenesis, caused by environmental factors and genetic variability within the population ([@BIO030056C12]). This effect has also been observed in studies on natural populations of intertidal foraminifera ([@BIO030056C48]). That study documented a survivorship of individuals reaching the reproductive stage of \u223c1.5% and a higher mortality in autumn in populations of *Haynesina depressula*. Individuals in Experiment 1 of this study included all size fractions (125--500\u2005\u00b5m), with a higher amount of small individuals according to the natural population in the sample, and were sampled in autumn which might have effected mortality rates in this experiment. Summarizing, differences in mortality rates in laboratory experiments may occur as a result of different cultivation techniques, different distribution of ontogenetic stages within the experimental setup, the sampling season or the characteristics of the respective populations inhabiting the sampling area.\n\nSize-related carbon and nitrogen content, processing and intraspecific carbon and nitrogen scaling {#s3b}\n--------------------------------------------------------------------------------------------------\n\nForaminifera play an important role in intertidal sediment biogeochemical fluxes, not only because of their high abundance, but also because of their role as primary consumers and detritivores. Foraminiferal nutrient cycling is also controlled by the abundances of intraspecific size classes and their relative contributions to the carbon and nitrogen pool. Larger *A. tepida* specimens contained considerably higher amounts of TOC and TN and showed higher pC and pN intake per individual ([Table\u00a01](#BIO030056TB1){ref-type=\"table\"}). Individual volume-specific carbon content typically shows an exponential relationship, which was also observed experimentally in several marine protists (e.g. [@BIO030056C43]). Here, the results show values for pooled individuals within a size range, which should help to estimate contributions of foraminiferal populations to TOC and TN pools in intertidal environments ([Fig.\u00a03](#BIO030056F3){ref-type=\"fig\"}). Individual pC and pN intake of *A. tepida* was high and comparable with that of abundant foraminiferal species in productive shelf environments (compare with [@BIO030056C14], [@BIO030056C15]). Also, contributions of total populations of *A. tepida* to phytodetritus processing was within the range found for foraminiferal key species in shelf environments (e.g. Uvigerinids, [@BIO030056C14], [@BIO030056C15]). In studies at the northern Wadden Sea ([@BIO030056C3]; [@BIO030056C59]), the biomass of macrofauna in tidal flats has been estimated to reach maximum values of \u223c50--65\u2005g ash-free dry weight\u00a0m^\u22122^ (\u223c25--33\u2005g\u2005TOC\u00a0m^\u22122^, estimated from [@BIO030056C61]), and the biomass of mesofauna to reach \u223c1\u2005g ash-free dry weight\u00a0m^\u22122^ (\u223c500\u2005mg\u2005TOC\u00a0m^\u22122^). In comparison, the maximum biomass of living *A. tepida* on our sampling date in April 2015 was \u223c500\u2005mg\u2005C\u00a0m^\u22122^ (within the 0--1\u2005cm sediment layer). Therefore, the biomass of *A. tepida* can at times approach those of maximum values of mesofauna, confirming a major contribution of *A. tepida* to the intertidal C and N pool at times of high abundances. Analogous, oxygen respiration measurements on intertidal foraminifera, including *A. tepida*, reported a high contribution of aerobic remineralization in intertidal mudflats ([@BIO030056C10]).\n\nOn the sampling day at the end of April 2015, there were high abundances of *A. tepida* in the samples at Station 1 and Station 2, close to the shoreline. The station behind the tideway, Station 3, showed much lower numbers of *A. tepida*, most likely because of stronger exposure to currents at this sampling spot. The remarkably high abundances of living *A. tepida* individuals in the 125--250\u2005\u00b5m size fraction indicates the recent occurrence of a reproductive event close to the sampling date. The size-specific content of and increased specific demand for food in smaller individuals are shown in the results of Experiment 2. This can be explained by the need to invest in growth and structure. The scaling relationships of weight-specific nitrogen content ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}A, Nitrogen) and demand for phytodetrital nitrogen ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}B,C, Nitrogen), show a prominent similarity between weight-specific TN content and demand for phytodetrital nitrogen per cytoplasmic nitrogen content (b \u223c0.75, in all cases). This is possibly related to an orientation of food demand on the limiting nutrient nitrogen. The high cytoplasmic ratios of pN:TN in contrast to lower pC:TOC ratios in *A. tepida* ([Fig.\u00a02](#BIO030056F2){ref-type=\"fig\"}) also demonstrate a high nitrogen demand. A size-related metabolic scaling relationship of 0.75 for interspecific respiration is also present among several foraminiferal species ([@BIO030056C17]). This is consistent with theories explaining relationships between metabolic activities and organismic mass, surface area, or size of structural elements ([@BIO030056C9]; [@BIO030056C35]; [@BIO030056C36]). The scaling of phytodetrital C or N intake per individual weight or per cytoplasmic TOC or TN shows a much higher relative intake and incorporation of phytodetritus at smaller size classes ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}B,C). This might indicate a higher turnover of C and respiratory losses in larger individuals and a stronger C retention in smaller ones, due to an increased incorporation into cytoplasmic components. Additionally, the high pC:TOC and especially high pN:TN in the young individuals of Experiment 1 ([Fig.\u00a02](#BIO030056F2){ref-type=\"fig\"}) underlines a strong demand for food-derived nitrogen sources in young individuals.\n\nThe extrapolation of cytoplasmic TOC, TN, pC and pN in *A. tepida* to the core material demonstrates the strong effect of high abundances of small individuals on the contributions to carbon and nitrogen pools and phytodetritus processing in populations of *A. tepida* ([Fig.\u00a03](#BIO030056F3){ref-type=\"fig\"}, [Table\u00a01](#BIO030056TB1){ref-type=\"table\"}). The relative amounts of TOC, TN, pC and pN at Station 1 and Station 2 are dominated by the 125--250\u2005\u00b5m fraction ([Fig.\u00a05](#BIO030056F5){ref-type=\"fig\"}), expressing a strong effect of the size-specific scaling of phytodetritus intake ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}) at times after reproduction events.\n\nThe high total biomass of *A. tepida* (e.g. 136\u2005ind\u2005cm^\u22123^ at Station 2), the fact that the highest concentrations of *A. tepida* are usually found within the uppermost layer of 0--0.5\u2005cm ([@BIO030056C10]; [@BIO030056C66]) or 0--0.25\u2005cm sediment ([@BIO030056C2]) and the ability of foraminifera to expand their cytoplasm far beyond their test size, can reduce the space available per individual. Examples for pseudopodial expansion in the genus *Ammonia* are shown in [@BIO030056C11]. This cytoplasmatic expansion possibly increases the impact of population-density depending factors (e.g. space, food resources) and may result in intraspecific competition within *A. tepida* populations at times of reproduction and shortly thereafter. This might be a critical factor controlling the cyclicity observed in population densities of intertidal foraminifera. A flexible response to food supply as observed in Experiment 1 in adult individuals might therefore help to withstand such competitive pressures at times of resource shortage. However, further investigations are necessary to support this hypothesis.\n\nGiven the production of particulate carbon in the sampling area (e.g. [@BIO030056C67]), the pC values and abundances of *A. tepida* derived from this study, the *A. tepida* community in the sampling area is capable of processing of a maximum of \u223c11% of phytodetrital TOC in the productive season. With the corresponding nitrogen values from the experiment, the contribution of *A. tepida* to the processing of particulate organic nitrogen derived from phytodetritus can be estimated to max. \u223c6%. There is not much known about foraminiferal nitrogen regeneration from organic matter, but isotope tracer studies in the Arabian Sea have shown a high contribution to the processing of phytodetrital nitrogen by benthic foraminifera ([@BIO030056C15]). Feeding experiments with *A. tepida* have also shown a coupling of carbon and nitrogen intake from *D. tertiolecta* detritus ([@BIO030056C73]), which is also apparent in the present study (see [Fig.\u00a0S2](Fig.\u00a0S2)). The patterns of cytoplasmic nitrogen content in this study are similar to the carbon patterns ([Fig.\u00a01](#BIO030056F1){ref-type=\"fig\"}). Food-derived nitrogen in *A. beccarii* is most probably used to synthesize amino acids as an element of the OM in the foraminiferal test ([@BIO030056C55]). The increased production of such amino acids in the small size fraction of this experiment might therefore explain the higher demand for nitrogen in the younger specimens ([Fig.\u00a04](#BIO030056F4){ref-type=\"fig\"}B,C, Nitrogen). Further, *Ammonia sp.* reportedly stores assimilated nitrate under anoxic and dysoxic conditions ([@BIO030056C54], [@BIO030056C56]), although this was most likely not the case in this study (oxic conditions). An increase in cytoplasmatic nitrogen pools or nitrogen storage and might have resulted in a decoupling of pC and pN under the different treatments (compare with [Fig.\u00a0S2](Fig.\u00a0S2)).\n\nThe switch of the food sources from the period prior to the experiment (fed regularly with living algae) to phytodetritus at the experiments, might have affected the carbon and nitrogen processing in *A. tepida*, probably by lowering the intake to adapt to the new food source. However, natural switches of food availability occur in estuaries due to seasonal changes in algal community structures, changes in transportation rates of riverine OM, or increased availability of dead algae matter following a phytoplankton or microphytobenthos bloom. Generally, there is a broad range of potential algal food sources in the natural environment of the German Wadden Sea ([@BIO030056C62],[@BIO030056C63]; [@BIO030056C67]). The foraminiferal specimens in this study were fed with monospecific cultures of living *D. tertiolecta* for maintenance and the experiments were carried out with phytodetritus of this algae (to track the isotope signal of the food source). These diets do not correspond with the natural variety of food sources for *A. tepida* in the German Wadden Sea. However, *A. tepida* individuals kept in the laboratory have previously shown a good response to *Dunaliella*-derived diets ([@BIO030056C7]; [@BIO030056C42]; [@BIO030056C73]). Considering the abundance of several sources of OM as potential food with different nutritional value and processing rates in nature, results of this experiment might reflect a good estimate of the mean carbon and nitrogen processing of *A. tepida*.\n\nIn summary, adult individuals of *A. tepida* showed flexible patterns of phytodetritus processing as a response to different feeding modes. This could be interpreted as metabolic adaptation to different modes of food availability and represent a generalistic survival strategy in environments with strongly fluctuating food availability throughout the year. The overall content of TOC, TN, and phytodetrital intake (pC and pN) in *A. tepida* were high and comparable with foraminiferal key species in other productive ocean regions, reaching biomass values approaching that of maximum values of intertidal mesofauna. This supports the consideration of *A. tepida* as an important player in intertidal nutrient fluxes in seasons with high *A. tepida* abundances. Large individuals showed substantially higher contents of TOC and TN, and of phytodetrital intake. However, phytodetritus processing (specifically N) relative to individual size (dry mass) increased with decreasing individual size, most likely due to a higher energy demand to cover growth costs in juveniles. This higher relative food demand in young individuals affects the contributions to the total foraminiferal TOC and TN pools and phytodetritus uptake, increasing the impact of the small size fraction on intertidal nutrient cycling.\n\nMATERIALS AND METHODS {#s4}\n=====================\n\nSampling site and preparations for feeding experiments {#s4a}\n------------------------------------------------------\n\nForaminifera were collected at low tide in the intertidal mudflat of the German Wadden Sea near Friedrichskoog, at the northern part of the Elbe Estuary ([Fig.\u00a06](#BIO030056F6){ref-type=\"fig\"}). Samples were taken in April and August 2015. For the feeding experiments, the sediment surface was sampled to collect high numbers of living individuals of *A. tepida*. The sediment was pre-sieved at the sampling sites through 500\u2005\u00b5m and 125\u2005\u00b5m mesh to remove larger meiofauna and organic particles. At the University of Vienna, living *A. tepida* individuals were picked and collected in crystallizing dishes containing a thin layer of Wadden Sea sediment (particle size \\<60\u2005\u00b5m). Traces of movement on this sediment and pseudopodial activity served as vital signs, as well as protoplasmic color (rich yellow) and particle accumulation around the aperture ([@BIO030056C45]; [@BIO030056C51], [@BIO030056C52]). Specimens were repeatedly (every 4--5\u2005days) fed with living *Dunaliella tertiolecta* (Chlorophyta) until the start of the experiments and the green coloration of foraminiferal cytoplasm served as another indicator of vital activity. Additionally in April, three sediment cores were taken (4.5\u2005cm diameter) on the vertices of an acute-angled triangle (Station 1 and Station 2 were \u223c120\u2005m off the shore line, \u223c65\u2005m apart and covered with diatom mats, Station 3 was located behind a tide way \u223c550\u2005m off the shoreline). The uppermost cm of the sediment cores was stained with Rose Bengal and Ethanol to determine the abundances of living *A. tepida* individuals of the size fractions 125--250\u2005\u00b5m, 250--355\u2005\u00b5m and \\>355\u2005\u00b5m. Fig. 6.**Sampling area.** (A) Location of the intertidal mudflats at Friedrichskoog/German Wadden Sea. (B) Position of sampling spots for sediment core samples (St. 1=Station 1, St. 2=Station 2, St. 3=Station 3. One sediment core was taken at each position).\n\nTo produce an isotope-labeled food source, a culture of *D.* *tertiolecta* was grown in f/2 medium ([@BIO030056C23]; [@BIO030056C24]) enriched with 98 atom % ^13^C (NaH^13^CO~3~, Sigma-Aldrich) and 98 atom % ^15^N (Na^15^NO~3~, Sigma-Aldrich). The culture was harvested by centrifugation, rinsed three times in sterile, carbon and nitrogen free artificial seawater, frozen in liquid nitrogen and freeze-dried to produce artificial phytodetritus ([@BIO030056C32]; [@BIO030056C73]).\n\nSample preparation and analysis {#s4b}\n-------------------------------\n\nAfter termination of the respective experiments (for a detailed description see sections 'Experiment 1: Comparison of food uptake from a single high feeding pulse versus lower constant feeding' to 'Experiment 2: Size specific carbon and nitrogen processing'), foraminifera were removed from the experimental dishes and frozen at \u221220\u00b0C to stop metabolic activity. Specimens were carefully cleaned to remove adhering particles, transferred to tin capsules and dried overnight at 50\u00b0C. The tests were decalcified with (10--15\u2005\u00b5l) 4% HCl. The remaining cytoplasm was dried in a final drying step for 3\u2005days at 50\u00b0C. The emerging mass of CaCl~2~ through decalcification was subtracted from the dry weight by using the reaction equilibrium for CaCO~3~ and HCl. All glassware used for preparation was combusted at 500\u00b0C for 5\u2005h, and picking tools and tin capsules were cleaned in a solution of dichloromethane and methanol (CH~2~Cl~2~:CH~4~O; 1:1, *v:v*) to remove organic contaminants.\n\nForaminiferal samples were analyzed at the Stable Isotope Laboratory at the University of Vienna for Environmental Research (SILVER). Cytoplasmic content of organic carbon or nitrogen and ratios of ^13^C/^12^C and ^14^N/^15^N were determined with an Isotope Ratio Mass Spectrometer (DeltaPLUS, Thermo Finnigan, Vienna, Austria) interfaced (ConFlo III, Thermo Finnigan) to an elemental analyzer (EA 1110, CE Instruments, Vienna, Austria). Atom% of the samples were derived from isotope ratio data and were calculated using the Vienna Pee Dee Belemnite standard (*R*) for C (*R*~VPDB~=0.0112372) and atmospheric nitrogen for N (*R*~atmN~=0.0036765), where *X* is ^13^C or ^15^N:\n\nThe excess (E) of heavy stable isotopes in the labelled samples was calculated to determine the individual content of phytodetrital carbon (pC) and nitrogen (pN) within foraminiferal cytoplasm ([@BIO030056C44]) where *X* is ^13^C or ^15^N:\n\nThe product of isotope excess and total foraminiferal carbon and nitrogen content was used to calculate the amount of incorporated isotope *I*~iso~, and following the amount of phytodetrital carbon (pC) and nitrogen (pN) within the foraminiferal cytoplasm *I*~phyto~ ([@BIO030056C32]):\n\nStatistics were calculated using R version 3.3.2 (corresponding packages for analysis of experimental data are listed in the respective sections below), and graphs were done using ggplot2.\n\nExperiment 1: Comparison of food uptake from a single high feeding pulse versus lower constant feeding {#s4c}\n------------------------------------------------------------------------------------------------------\n\nForaminifera from the August 2015 batch were transferred to vials (d=4.5\u2005cm, h=6\u2005cm, 55--60 specimen per vial) containing 0.8\u2005g heat sterilized Wadden Sea sediment (\\<63\u2005\u00b5m) and covered with plankton net squares. The experiment started 2\u00a0weeks after sediment collection. To obtain triplicate samples for 2, 4, 7, 14, 21, and 28\u2005day incubations for the three different culture conditions, 45 vials were prepared and placed in aquaria. Incubations conditions were as follows: constant feed 1=constant feeding at temperature cycles of 20\u00b0C:18\u00b0C; constant feed 2=25\u00b0C:23\u00b0C, single pulse=single high feeding pulse at 20\u00b0C:18\u00b0C=single pulse, each cycle was 12\u2005h:12\u2005h ([Table\u00a02](#BIO030056TB2){ref-type=\"table\"}); cycles were chosen to mimic natural diurnal temperature fluctuations; standard temperature fluctuations of incubation chambers: \u00b10.6 at 20\u00b0C. Each of the three setups was incubated within one aquarium filled with filtered North Sea Water and aerated with ambient air, to support the stability of culture conditions during the experimental period. At the beginning of the experiment, labelled phytodetritus (*D. tertiolecta*) was introduced into the respective vials at day 0 for the single pulse approach (\u223c1670\u2005mg\u2005C\u00a0m^\u22122^) and the constant feeding approach (\u223c560\u2005mg\u2005C\u00a0m^\u22122^). In constant feeding mode, additional food (\u223c560\u2005mg\u2005C\u00a0m^\u22122^) was supplied every third day ([Table\u00a02](#BIO030056TB2){ref-type=\"table\"}). The amount of added carbon was adapted according to planktonic primary production (daily gross particulate production approximately 500--2000\u2005mg\u2005C\u00a0m^\u22122^\u00a0day^\u22121^) in the German Bight ([@BIO030056C67]) during summer months. Oxygen, pH and salinity were recorded on a regular basis during the run of the experiment to retain environmental conditions within the optimum range. Some replicates contained dead individuals at the time of sampling. Dead individuals were excluded from further analysis. Towards the end of the experimental period, some replicates had to be pooled to obtain sufficient amounts of live specimens for further elemental and isotopic analysis.Table\u00a02.**Culture and feeding conditions for different feeding experiments**\n\nExperiment 2: Size-specific carbon and nitrogen processing {#s4d}\n----------------------------------------------------------\n\nSediment samples collected in April 2015 were sieved to obtain three size classes (small: 125-250\u2005\u00b5m; medium: 250-355\u2005\u00b5m; large: \\>355\u2005\u00b5m). Foraminifera were picked out and triplicate samples for each size class (50--150 individuals, depending on size class) were set up in crystallizing dishes, containing 280\u2005ml artificial Seawater (modified after [@BIO030056C16]) and incubated at 20\u00b0C, starting four weeks after field sampling. Samples were taken 2 and 4\u00a0days after the initial feeding pulse of *D. tertiolecta* phytodetritus ([Table\u00a02](#BIO030056TB2){ref-type=\"table\"}) and processed for EA-IRMS analysis as described above.\n\nTotal organic carbon (TOC), total nitrogen (TN), and pC and pN of foraminiferal cytoplasm were compared between size classes and between sampling days (Mann--Whitney *U*-test), to determine the effect of incubation time and size class on cellular element content or phytodetritus intake. Then, the data were log normalized to fit linear regressions (general least squares, GLS) to explore the relationships of size-related intraspecific variables (individual dry weight, TOC, TN) and phytodetritus intake (pC, pN). The back-transformed regression equations were used to interpret the proportionality of size-dependent TOC and TN content and individual food intake, where, e.g. an isometric relationship (slope of the regression, b \u223c1) describes an even rise of the measured variables with foraminiferal size (no difference in size-related cytoplasmic TOC and TN or phytodetritus intake), and a lower slope (b\\<1) describes a negative proportionality, i.e. a decrease of cytoplasmic TOC and TN levels and phytodetritus intake with increasing size.\n\nFinally, foraminiferal TOC, TN, pC and pN were projected on the three size fractions in a natural population of *A. tepida* by using the abundance data from the Rose Bengal stained sediment cores, to estimate the effect of size-dependent phytodetritus intake based on the natural size distribution of an *A. tepida* community and to assess the total contribution of this community to the sediment bulk TOC and TN pool.\n\nData analysis was done using R and related packages ggplot2, dplyr, doBy ([@BIO030056C31]; [@BIO030056C58]; [@BIO030056C60]; [@BIO030056C70]; [@BIO030056C71]).\n\nSupplementary Material\n======================\n\n###### Supplementary information\n\nWe thank Matthias Nagy and Gloria Quell for help with the field sampling.\n\n**Competing interests**\n\nThe authors declare no competing or financial interests.\n\n**Author contributions**\n\nConceptualization: J.W., A.J.E., P.H.; Methodology: J.W., P.H.; Formal analysis: J.W.; Investigation: J.W., P.B., M.W., M.G.; Resources: W.W.; Writing - original draft: J.W.; Writing - review & editing: P.B., A.J.E., W.W., P.H.; Visualization: J.W.; Supervision: P.H.\n\n**Funding**\n\nA part of the isotope analysis in this study was funded by the The Micropaleontological Society (TMS) Grant-In-Aid. Open access funding provided by University of Vienna.\n\n**Supplementary information**\n\nSupplementary information available online at <http://bio.biologists.org/lookup/doi/10.1242/bio.030056.supplemental>\n"}
+{"text": "INTRODUCTION\n============\n\nAromatherapy is one of the alternative medicines using essential oils and has long been used as an herbal medicine. Recently essential oils have been empirically used worldwide for clinical conditions including various kinds of inflammatory diseases, such as allergy, rheumatism, and arthritis. These activities have mainly been recognized through clinical experience, but there has been relatively little evidence about the pharmacological actions of these oils.\n\nSeveral investigators have suggested that tea tree \\[[@B1], [@B2]\\] and lavender \\[[@B3]\\] oils suppressed allergic symptoms through the suppression of histamine release \\[[@B4], [@B5]\\] and cytokine production \\[[@B6]\\] in vitro and in vivo. Several essential oils such as eucalyptus \\[[@B7]\\] and lavender \\[[@B8]\\] oils inhibited carrageenan-induced paw edema. Moreover, in human, skin application of tea tree oil was reported to suppress the edema induced by intradermal injection of histamine \\[[@B9]\\]. However, the chronic effects of essential oils using inflammatory mice model have hardly been investigated.\n\nPreviously we reported that the essential oils such as geranium oil suppressed the adherence response of neutrophils in vitro \\[[@B10]\\], and that the intraperitoneal administration of geranium oil lowered neutrophil recruitment into the peritoneal cavity induced by injection of a chemotactic agent, casein in vivo \\[[@B11]\\]. We also reported that both intraperitoneal and cutaneous applications of the oil suppressed cellular inflammation and neutrophil accumulation to the inflammatory sites which were induced by curdlan, a linear (1 \u2192 3)-\u03b2-D-glucan known as an immunostimulating substance in fungi \\[[@B12]\\]. These results suggested the possibility that geranium oil might effectively suppress symptoms in inflammatory disease associated with neutrophil activities.\n\nIn the present study, we investigated the effects of geranium oil on carrageenan-induced foot edema and collagen-induced arthritis, which are models for acute and chronic inflammation accompanied by neutrophil accumulation.\n\nMATERIALS AND METHODS\n=====================\n\nEssential oils\n--------------\n\nGeranium oil was provided by Pranarom (Kenso-igakusha Ltd, Tokyo, Japan). The oil was diluted to 0.625, 1.25, 2.5% solution by 2.5% dimethyl sulfoxide (DMSO), and 25 \u03bcL of Tween 20 was added to 2 mL of the essential oil solution. Main constituents of the oil based on the company\\'s data were citronellol (22.42%), geraniol (18.25%), linalool (5.59%), citronellyl formate (10.24%), geranyl formate (7.36%), guaiadiene (6.88%), and isomenthone (7.58%).\n\nAgents\n------\n\nPolyoxyethylene (20) sorbitan monolaurate (Tween 20) was purchased from Wako Pure Chemical Industries, Ltd (Osaka, Japan). Bovine collagen II was from Cosmo Bio (Tokyo, Japan), complete Freund\\'s adjuvant (CFA) was from DIFCO (Michigan, USA). Carrageenan \u03bb, Hexadecyltrimethylammonium bromide (HTAB), human myeloperoxidase (MPO), and tetramethylbenzidine (TMB) were purchased from Sigma-Aldrich (Tokyo, Japan).\n\nAnimals\n-------\n\nAll animal experiments were performed according to the guidelines for the care and use of animals approved by Teikyo University. Six-week-old male DBA mice (Charles River Japan, Inc, Kanagawa, Japan) were used for all animal experiments. The photoperiods were adjusted to 12 hours of light and 12 hours of darkness daily, and the environmental temperature was constantly maintained at 21\u00b0C. The mice were kept in cages housing 4--6 animals and were given ad libitum access to food and water.\n\nCarrageenan-induced edema: footpad reaction\n-------------------------------------------\n\nFootpad reaction was based on the method of Abe et al \\[[@B13]\\] and partly modified. Ten mg of carrageenan were dissolved in 1 mL of saline and 0.05 mL of the solution was injected to the mouse left footpad to induce edema. The photos of feet from a lateral view were taken before and 6 and 24 hours after carrageenan injection using a digital camera, and foot thickness was measured from the photo ([Figure 1](#F1){ref-type=\"fig\"}). The edema was calculated by the difference of thickness between 0 and 6 or 24 hours. Ten minutes after carrageenan injection, the mice were intraperitoneally given 0.2 mL of 2.5% geranium solution. A dose of 2.5% solution corresponds to 5 \u03bcL of pure oil. Control mice received 0.2 mL of 2.5% DMSO solution. Mice were sacrificed by carbon dioxide 24 hours after carrageenan injection. The feet were resected 5 mm above their heels, soaked in 2 mL of 80 mM sodium phosphate buffer, pH 5.4, containing 0.5% HTAB (0.5% HTAB solution), weighed and kept at \u221220\u00b0C until the MPO assay. We used a nontreated right foot of the same mice as a reference.\n\nMyeloperoxidase (MPO) assay\n---------------------------\n\nThe MPO assay was based on the method of De Young et al \\[[@B14]\\] and partly modified. Frozen samples were thawed at room temperature and homogenized at 0\u00b0C using a Polytron (Kinematica AG, Lucerne, Switzerland). The homogenates were poured into sampling tubes and centrifuged at 12000\u00d7g at 4\u00b0C for 15 minutes.\n\nTriplicate 30 \u03bcL samples of resulting supernatant were poured into 96 well microtiter plates. For assay, 200 \u03bcL of a mixture containing 100 \u03bcL phosphate buffered saline, 85 \u03bcL of 0.22 M sodium phosphate buffer, pH 5.4, and 15 \u03bcL of 0.017% hydrogen peroxide were added to the wells. The reaction was started by the addition of 20 \u03bcL of 18.4 mM TMB\u00b72HCl in 8% aqueous dimethylformamide. Plates were stirred and incubated at 37\u00b0C for 3 minutes and then placed on ice where the reaction in each well was stopped by addition of 30 \u03bcL of 1.46 M sodium acetate buffer, pH 3.0. The MPO value was calculated by measuring the absorbance of samples at 620 nm (OD value) followed by its conversion into MPO values per foot.\n\nCollagen-induced arthritis\n--------------------------\n\nInduction of type II collagen-induced arthritis was based on the method of Ochi et al \\[[@B15]\\] and partly modified.\n\nCollagen II from bovine articular cartilage was dissolved overnight at 4\u00b0C in 0.1 M acetic acid at a concentration of 2.5 mg/mL. The solution was emulsified with 1.2 times volume of CFA, and 100 \u03bcL of the emulsion were administered subcutaneously at the base of the tail of the mice for immunization on day \u221221. Booster injection of 100 \u03bcL of the emulsion was given on day 0. Mice were intraperitoneally given 0.2 mL of geranium oil solutions from day 0 to 21, 5 days per week (injection period). Control mice were given 0.2 mL of 2.5% DMSO solution. Their weight and paws were measured 2 days each week from day 0 to 39.\n\nMouse paws were scored for arthritis based on the method of Kim et al \\[[@B16]\\] using a macroscopic scoring system ranging from 0 to 4 (0, no swelling; 1, swelling of one joint; 2, two joints involved; 3, more than two joints involved; 4, severe arthritis over the entire paw and joints). The arthritic score for each mouse was the sum of the scores of all four paws.\n\nStatistical analysis\n--------------------\n\nThe results were expressed by the mean \u00b1 standard error. The data were statistically compared using the Student\\'s *t*-test and the \u03c7-square test.\n\nRESULTS\n=======\n\nEffects of geranium oil on carrageenan-induced edema in the hindpaws of mice\n----------------------------------------------------------------------------\n\nCarrageenan injection to the footpad increased the foot thickness of control mice by 1.68 \u00b1 0.08 mm after 6 hours and the swelling continued for 24 hours (1.55 \u00b1 0.16 mm) ([Figure 1](#F1){ref-type=\"fig\"}). Intraperitoneal injection of geranium oil significantly suppressed the increase in foot thickness both 6 and 24 hours after carageenan injection (0.96 \u00b1 0.13 mm and 0.91 \u00b1 0.19 mm, resp). [Figure 2](#F2){ref-type=\"fig\"} shows photos of the typical swelling of control mice and that of geranium-treated mice 24 hours after carageenan injection. These photos clearly indicated that the foot treated with geranium oil was less swollen than the control foot.\n\nIn order to confirm the inflammatory response, foot weights and MPO activity in foot homogenates were measured.\n\nThe weight of the carrageenan-injected control foot was significantly increased compared with the nontreated foot (0.28 \u00b1 0.01g and 0.17 \u00b1 0.00g, resp), as shown in [Figure 3](#F3){ref-type=\"fig\"}. This figure also shows that intraperitoneal injection of geranium oil significantly lowered the weight gain (0.21 \u00b1 0.01g).\n\nThe same feet were used for measurement of MPO activity which represented the number of neutrophils. Carrageenan injection to the footpad induced a marked increase of the MPO value of the foot compared with the nontreated foot (61.51\u00b116.84 units/foot and 4.02\u00b11.96 units/foot, resp) ([Figure 4](#F4){ref-type=\"fig\"}). Geranium oils suppressed the increase of MPO value significantly (44.38 \u00b1 6.30 units/foot). This suggested that intraperitoneal injection of geranium oil lowered neutrophil accumulation to the carrageenan-injected foot.\n\nEffects of geranium oil on collagen-induced arthritis in mice\n-------------------------------------------------------------\n\nNext, we examined the effects of the oil against the collagen-induced arthritis in mice as a chronic inflammation model.\n\nOne of control mice immunized with collagen II (on days \u221221 and 0) developed an edema (arthritis) from day 7, and then most of them elicited edema, 6 of 10 on day 21 and 7 of 10 on day 39 ([Figure 5](#F5){ref-type=\"fig\"}). Their symptoms were aggravated gradually after the second collagen II injection. In mice given 5 \u03bcL of geranium oil, edema of the feet was observed only on one animal with slight swelling. There were statistical differences between control and the 5 \u03bcL geranium oil group, on day 10 and after day 17 using \u03c7-square test. No aggravation of symptoms was observed even after completion of geranium oil injection.\n\nOral administration of indomethacin used as a reference suppressed the edema during the injection period, but 1 week after completion of injections the feet appeared to be swollen (data not shown).\n\nTime course of the score of inflammatory symptoms is depicted in [Figure 6](#F6){ref-type=\"fig\"}. The symptoms of the control mice were exacerbated time-dependently. The score of the mice injected with 5 \u03bcL of geranium oil was clearly lower than that of control mice with statistical significance on days 24, 35, and 39. On the other hand, 1.25 and 2.5 \u03bcL oils seemingly lowered the scores, but the differences were not statistically significant.\n\n[Figure 7](#F7){ref-type=\"fig\"} shows typical pictures of the feet of a control and a 5 \u03bcL geranium injected mouse.\n\n[Figure 8](#F8){ref-type=\"fig\"} indicates the change of body weight during treatment. The weight of mice treated with geranium oil decreased immediately after oil injection. The reduction was largest in the group injected with 5 \u03bcL of geranium oil. Their weight loss was gradually recovered, but the recovery was slow in the groups injected with 2.5 and 5 \u03bcL of geranium oil. In the group injected with 5 \u03bcL of the oil, 2 mice died on days 18 and 21.\n\nDISCUSSION\n==========\n\nIn the present study, we showed that intraperitoneal administration of geranium oil suppressed two types of inflammatory responses, carrageenan-induced edema and collagen-induced arthritis.\n\nWhen mice received intraperitoneal injection of 5 \u03bcL of geranium oil 10 minutes after carrageenan injection, carrageenan-induced edema was significantly suppressed at 6 and 24 hours. This indicates that the suppressive effect of the oil on the acute inflammation continued at least for 24 hours. We also measured the weight and MPO activity as parameters of neutrophil accumulation at 24 hours, and the results suggest that the oil suppressed the acute inflammation accompanied by neutrophil accumulation (Figures [3](#F3){ref-type=\"fig\"} and [4](#F4){ref-type=\"fig\"}).\n\nWe previously reported that intraperitoneal administration of geranium oil suppressed the casein-induced accumulation of neutrophils in the peritoneal cavity \\[[@B11]\\], and both intraperitoneal and cutaneous applications of the oil suppressed cellular inflammation and neutrophil accumulation to the inflammatory sites which were induced by a (1 \u2192 3)-\u03b2-D-glucan, curdlan \\[[@B12]\\]. In this study, geranium oil inhibited the carrageenan-induced edema from 6 hours and neutrophil accumulation for 24 hours. These indicate that geranium oil suppresses the neutrophil accumulation in various inflammatory responses from an early stage.\n\nThe cellular mechanism of the suppression of edema by geranium oil remains to be clarified. It is reported that carrageenan injection induced the production of histamine and cytokine such as TNF-\u03b1 \\[[@B17], [@B18]\\]. TNF-\u03b1 is one of the major inflammatory cytokines with the capacity to prime activation of the neutrophils for their various functions \\[[@B19]\\]. In our previous study, we showed that geranium oil inhibited TNF-\u03b1-induced neutrophil adherence to a plastic plate at very low concentrations (0.00625%) in vitro \\[[@B10]\\], so we can assume that this suppression might be partially caused by inhibiting neutrophil response to TNF-\u03b1 in vivo.\n\nWe also evaluated the long-term effect of the oil using collagen-induced arthritis mouse as the chronic anti-inflammatory disorder model. Collagen-induced arthritis in mice has many characteristics in common with human rheumatoid arthritis such as foot swelling and has been the most used animal model for the disease \\[[@B15], [@B20]\\]. Five \u03bcL of geranium oil suppressed the swelling and the effect continued even after cessation of oil injection. As far as we know, this is the first report indicating that geranium oil suppressed the later phase of inflammatory response as well as the earlier phase in in vivo experiments.\n\nRheumatoid arthritis is considered an autoimmune disease involving joint inflammation associated with TNF-\u03b1 production. Various cells such as Th1 cells, neutrophils and macrophages, and their cytokines such as IL1 and TNF infiltrate the synovial tissues to destroy joints \\[[@B21]\\]. In our experiments, geranium oil was administered to the mice after a booster injection of collagen II. We can therefore assume that the oil suppresses the later phase of autoimmune reaction, or the onset of symptoms after autoimmune reaction, not the early stage of the reaction.\n\nPrevious pathological studies \\[[@B20], [@B22]\\] on experimental murine arthritis showed that there were marked edema of synovium and infiltration of many polymorphonuclear cells such as neutrophils in the early phase of arthritis onset, followed by the chronic destructive phase in which pronounced proliferation of synovium containing mononuclear cells was observed. From these findings and our previous data, we can speculate that geranium oil may suppress the onset of the symptoms at least partially through inhibition of neutrophil infiltration. To check this possibility, we wish to evaluate the MPO value in further study.\n\nIt was noted that 5 \u03bcL of geranium oil suppressed the foot swelling during and after the oil injection period, suggesting a long-lasting effect of this oil. Preliminary study showed that indomethacin inhibited the swelling only during its administration, and 1 week after completion of the injection the feet gradually swelled (data not shown). This indicates that repetitive administration of the oil may elicit a long-lasting effect.\n\nIn aromatherapy, several essential oils can be applied as a help in therapeutic treatments for inflammatory symptoms with lesional neutrophil accumulation, such as arthritis, aphthous stomatitis, lesional bacterial or fungal infections. Their effectiveness is postulated clinically, but little experimental evidence has been obtained. Our two results give basic evidence about the activity of geranium oil for both acute and chronic inflammatory disorders.\n\nIn relation to the application of the essential oil, we must mention its toxicity. In our later experiment, 2 mice of the group which were intraperitoneally given 5 \u03bcL of geranium oil died during the experiment. The body weight of this group was greatly reduced, so the administration protocol might have been too severe for them. In order to develop a less toxic administration procedure, it is our opinion that the selection of administration routes of the essential oil must be critical, since cutaneous application of geranium oil suppressed the curdlan-induced skin inflammation without apparent toxic response \\[[@B12]\\]. By optimizing the dosage and administration route, we hope to propose safer and more effective treatment protocol using essential oil for inflammatory diseases.\n\nThis work was supported in part by a Grant (no 15590401) from the Ministry of Education Culture, Sports, Science and Technology of Japan.\n\nFigures and Tables\n==================\n\n![Effects of intraperitoneal administration of geranium oil on foot swelling induced by carrageenan injection. Carrageenan was injected to left footpad of mice, and 10 minutes after the injection, geranium oil or DMSO was given intraperitoneaouly. Six and 24 hours later, the increase of the foot thickness was measured. Each value represents an average from 5 mice and the standard error. \\**P* \\< .05 difference from control.](MI2006-62537.001){#F1}\n\n![Typical swelling of control mice and that of geranium-treated mice. Nontreated foot (a), carrageenan injected control (b), and geranium-treated (c) feet 24 hours after carrageenan injection were shown. The mouse was administered with geranium oil 10 minutes after carrageenan injection.](MI2006-62537.002){#F2}\n\n![Effects of intraperitoneal administration of geranium oil on foot swelling estimated by their weight. Carrageenan was injected to left footpad of mice, and 10 minutes after the injection, geranium oil or DMSO was given intraperitoneaouly. Twenty four hours later, their feet were resected to measure their weight. Each value represents an average from 5 mice and the standard error. \\*\\* *P* \\< .01 difference from control.](MI2006-62537.003){#F3}\n\n![Effects of intraperitoneal administration of geranium oil on MPO activity in foot homogenates. Carrageenan was injected to left footpad of mice, and 10 minutes after the injection, geranium oil or DMSO was given intraperitoneaouly. Twenty four hours later, their feet were resected to measure their MPO activities. Each value represents an average from 5 mice and the standard error. \\* *P* \\< .05, \\*\\* *P* \\< .01 difference from control.](MI2006-62537.004){#F4}\n\n![Effects of intraperitoneal administration of geranium oil on the ratio of the mice which revealed feet swelling by collagen II induction. Collagen II with CFA was subcutaneously injected to the base of the tail of the mice on days \u221221 and 0. Geranium oils were given from day 0 to 21, 5 days/week. Each value represents percentages of mice with foot swelling. \\* *P* \\< .05 difference from control using \u03c7-square test.](MI2006-62537.005){#F5}\n\n![Effects of intraperitoneal administration of geranium oil on the inflammatory score. Collagen II with CFA was subcutaneously injected to the base of the tail of the mice on days \u221221 and 0. Geranium oils were given from day 0 to 21, 5 days/week. The arthritis was scored as described in material and methods. Each value represents an average from 5--10 mice and the standard error. \\**P* \\< .05 difference from control.](MI2006-62537.006){#F6}\n\n![Macroscopic arthritis of control mice and that of geranium-treated mice. The mice were treated as represented in the legend to [Figure 5](#F5){ref-type=\"fig\"} and their feet were observed on day 39. Arrows indicate the swelling of foot. (a) Control mouse; (b) mouse administered with 5 \u03bcL geranium oil.](MI2006-62537.007){#F7}\n\n![Changes in body weight during and after geranium oil injection to mice. Collagen II with CFA was subcutaneously injected to the base of the tail of the mice on days \u221221 and 0. Geranium oils were injected from day 0 to 21, 5 days/week. Each value represents an average of 5--10 mice and the standard error. \u22c7 : death of mouse.](MI2006-62537.008){#F8}\n"}
+{"text": "Background\n==========\n\nThe availability of multiple genome sequences provides a valuable reference facilitating systematic family-wide or even genome-wide investigation of gene function. Information on gene structure, evolution and family relationships can be drawn and predictions of biochemical function can be made through sequence comparisons. Functional processes in cells, however, are driven by proteins and a deeper understanding of gene function will ultimately require information on protein interactions, protein expression levels, modifications and sites of action. Antibodies provide a valuable means of gaining such information. Several initiatives to generate monoclonal antibodies on a genome-wide scale are under consideration \\[[@B1],[@B2]\\]. Large scale profiling of commercial and newly generated polyclonal antibodies to over 700 antigens has previously been described \\[[@B3]\\]. Panels of monoclonal antibodies, however, would have advantages over polyclonal antibodies by being a renewable resource of defined, homogeneous composition. Potential cross-reactivity will be less than in a complex polyclonal mixture. Furthermore, the availability of multiple independent antibodies, as shown here, allows independent verification of results.\n\nGeneration of antibodies on such a scale presents a range of challenges, spanning the generation of antigen through generation and validation of antibodies to production, tracking and application in a relevant biological read-out. One of the first bottlenecks is the creation of quality recombinant protein in high throughput. This goal requires methods for primer design, cloning, sequence confirmation, protein expression, purification and quality control of the resulting products. In this study, protein products derived from both bacterial and mammalian systems were used as targets for antibody generation. *Escherichia coli*provides an efficient system for protein expression, and generation of soluble product can be aided by addition of solubility enhancing and affinity purification tags \\[[@B4]\\]. In addition, a protein expression system based on transient transfection of mammalian cells \\[[@B5]\\] was used for expression of receptor extracellular domains \\[[@B6]\\].\n\nPhage display is a scalable method of generating antibody reagents, and phage-antibody libraries can provide a rich source of antibody diversity, potentially providing hundreds of unique antibodies per target. The antibody gene, once isolated, can be conveniently shuttled into a variety of expression formats, providing a renewable resource of antibody protein \\[[@B7]\\]. We report here the generation of an antibody phage display library of over 10^10^clones and its application to the selection and screening of over 38,000 antibody clones. DNA sequencing allows redundancy to be removed from the antibody panel and permits a definitive description of the resulting antibody gene and its product. Over 7,200 unique recombinant antibodies to 290 targets were identified. Of these, 4,437 were picked and their specificity determined against a wider panel of antigens. In addition, detection sensitivity was measured for 100 antibodies to 10 antigens using a bead based flow cytometry assay, with sensitivity below 18,000 antigens/bead demonstrated for all 10 antigens. This assay was also predictive of performance in detecting endogenous levels of antigen by flow cytometry. Finally, we illustrate their application in immunohistochemistry using tissue microarrays to produce protein expression profiles.\n\nThus, we demonstrate the potential of high throughput processes for the generation and validation of recombinant proteins and antibodies. We illustrate examples of information, such as cross-reactivity, sequence, and performance data, that may form part of a simple standardized validation protocol. Apart from exemplifying the potential of such large scale approaches, the validated antibody and protein reagents generated in this study will have research and diagnostic potential and have been made available, along with the characterization data, to the scientific community \\[[@B8],[@B9]\\].\n\nResults\n=======\n\nHigh-throughput antibody selection\n----------------------------------\n\nWe report the construction of an antibody phage display library of 1.1 \u00d7 10^10^clones and its utilization for high throughput antibody generation and characterization. The antibody library was created by sequentially cloning a repertoire of light chain variable regions (VL) followed by cloning of heavy chain variable regions (VH). The heavy and light chain repertoires were created by PCR amplification from human lymphocytes from 43 donors, mainly collected from peripheral blood. Human VH genes are grouped into seven families and VL genes into six kappa and ten lambda families, based on sequence homologies. PCR primers were designed with reference to these VH and VL germline families \\[[@B10]\\]. The full set of primers along with detailed protocols for their use are listed in Additional data file 1.\n\nIn a first step, the VL repertoire was cloned into the *Nhe*I/*Not*I sites of pSANG4 (Figure [1](#F1){ref-type=\"fig\"}). Subsequently, plasmid DNA was prepared from this library before cloning the VH repertoire into the *Nco*I/*Xho*I sites. The final format of the antibody is a single chain Fv (scFv) with VH and VL fragments joined by a flexible linker peptide (Gly~4~Ser Gly~4~Ser Gly~3~Ala Ser).\n\n![Schematic representation of the process for the high throughput generation and screening of recombinant antibodies.](gb-2007-8-11-r254-1){#F1}\n\nThe library was generated using B lymphocytes from 42 human peripheral blood donations and 1 tonsil. Diversity was maintained at a maximum by a variety of strategies. All donations were maintained separately and combined to 11 pools at the mRNA stage. V region primers (14 heavy chain, 13 kappa light chain and 15 lambda light chains) were kept separate during primary PCR for each of the 11 mRNA pools, that is, 462 separate primary PCR reactions were performed. Each of the kappa light chain families and three pools of the lambda chain families were transformed separately and only pooled into two master sets (kappa and lambda) once DNA was isolated from each library stock to prepare vector for heavy chain cloning. Transformation of each of the 7 heavy chain families into both the kappa and lambda light chain libraries was done separately creating a total of 14 pools of scFv libraries, supplemented with 2 additional sub-libraries representing VH3/Vkappa1 and VH/kappa3 combinations. Finally, phage were prepared individually from each of the 16 different aliquots and pooled to maintain equivalent representation of each sub-aliquot.\n\nPhage particles were rescued using a trypsin sensitive helper virus \\[[@B11]\\]. Thus, phage particles incorporating the minor coat protein encoded solely from the helper phage are susceptible to trypsin cleavage and will not be infective. In contrast, those displaying an antibody-gene 3 fusion will retain infectivity after trypsin treatment. This reduces background from \\'non-participating\\' phage particles that have been generated without utilizing an antibody-gene 3 product from the library. This increased the efficiency of binder isolation such that we routinely required only two rounds of selection. High throughput selection was further facilitated using a liquid culture amplification/rescue method, eliminating the need for plating out and scraping of bacterial culture plates.\n\nAntibodies were selected to 404 antigen targets (representing 280 genes) that were primarily produced in *E. coli*\\[[@B4]\\] or in mammalian cells \\[[@B6]\\], with a proportion sourced commercially. The majority of genes (214/280) were cell surface receptors, with 30% of the immunoglobulin superfamily represented (139 genes). Receptors were chosen as they comprise a major class of drug targets that provide information on the communication potential of cells, and act as identifiers of different cell types, which can be used in their purification. Most of the genes were murine in origin (81%), with the remainder being human (18%) and rat (1%). A full gene list is provided at the \\'Antibody Atlas db\\' \\[[@B8]\\] and in Additional data file 2.\n\nFollowing selection, resultant antibody populations were sub-cloned into an expression vector (pSANG14-3F), which fuses the antibody scFv to a tri-FLAG tag sequence and the dimeric enzyme bacterial alkaline phosphatase. This provides a convenient enzymatic tag to detect binding as well as driving dimerization, thereby increasing avidity of binding \\[[@B7]\\]. From each selection, 94 clones were screened and sequenced, and a sub-set of up to 22 unique clones identified for further specificity and sequence characterization. The process is outlined in Figure [1](#F1){ref-type=\"fig\"}.\n\nPrimary screening and sequencing of selected antibodies\n-------------------------------------------------------\n\nThe primary screening assay to identify binders was initially performed with bacterial lysates in a 96-well format using direct detection of alkaline phosphatase fused to the antibody fragment. This convenient assay requires no secondary antibody and involves a single cycle of binding and washing of the antibody-alkaline phosphatase fusion, followed by addition of enzymatic substrate. In subsequent work, the assays were performed in a 384-well format using europium labeled secondary reagents, to more closely reflect the system we used for secondary screening (see below).\n\nTo identify unique binders, all primary clones were sequenced using a single primer that annealed upstream of the 5\\' end of the VH gene. In total, 38,164 clones were screened and sequenced. Of these, 9,384 were positive in primary screening ELISA. Sequence analysis of all of these identified 7,236 unique clones. With 292 antigens passing selection, this represents a success rate of 72% with an average of 25 clones for each positive antigen.\n\nFigure [2a](#F2){ref-type=\"fig\"} shows the number of unique scFv binders generated for each target, demonstrating the diversity achieved from the library. This figure compares the success rates of protein generated in both bacterial and mammalian expression systems. It is clear that the performance of targets derived from mammalian expression systems were superior in terms of generating target specific, sequence unique scFvs. For bacterially expressed antigens, 54% of the targets gave rise to one or more unique binders after the primary screening ELISA stage, with 19% giving rise to 22 or more unique clones. For mammalian expressed antigens, 82% of the targets gave rise to one or more unique binders after the primary screening ELISA, with 52% giving rise to 22 or more unique clones. The diversity of antibodies generated from the library across a wide range of antigens clearly illustrates the quality of the library and methods used. Furthermore, it suggests that more screening would generate a greater diversity of antibodies. Generation of hundreds of antibodies to a single antigen using a similar antibody phage display library has previously been described \\[[@B12]\\].\n\n![High throughput generation and characterization of recombinant antibodies. **(a)**For each target antigen, 94 clones were screened in ELISA and sequenced to identify unique clones. The plot shows the number of antigens selected (x-axis) versus the number of unique positive antibodies generated (y-axis) for each antigen. Separate plots are presented for antigens produced in either bacterial or mammalian expression systems, illustrating the improved success rate for mammalian antigens. **(b)**Example specificity data for antibodies selected against Slam f9 (produced in the HEK293 mammalian expression system). All antibodies are screened against target antigen, the relevant fusion partner that was used in selection, keyhole limpet hemocyanin (KLH), thyroglobulin, myoglobin, cytochrome c, human IgG, laminin, fibronectin, \u03b1-glycerol phosphate dehydrogenase, and total protein lysates from zebra fish (*D. rerio*) and yeast (*S. pombe*). Results are shown for 22 different antibodies as well as our routine anti-desmin control (des-D7) and a no antibody control. Detection was via time resolved fluorescence and values are shown on a logarithmic scale. **(c)**Global summary of secondary ELISA data for all antibodies in secondary screening. Signal generated on specific antigen is shown for all 4,437 samples (solid block). Signal achieved on one of the ten irrelevant antigens (cytochrome c) is also shown.](gb-2007-8-11-r254-2){#F2}\n\nTo identify panels of antibodies for further study, sequences were clustered according to the heavy chain complementarity determining regions (CDRs) with a particular focus on CDR3, which is most variable. Clones were also ranked according to signal-to-noise ratio in primary screening ELISA with priority given to clones with the highest signal-to-noise ratio and most divergent HCDR3 amino acid sequences. Up to 22 scFv clones were chosen for each target, yielding 4,437 antibodies to 286 antigens for specificity testing and high quality sequencing.\n\nTo give a definitive description of each antibody, all 4,437 clones undergoing secondary screening were sequenced in depth with 6 primers covering the VH and VL gene segments in both forward and reverse orientation. A consensus sequence was generated and the most closely related VH and VL germ line sequences were identified (Figure [3](#F3){ref-type=\"fig\"}). This was done by comparing the VH and VL segment of each clone against a database of human antibody germline genes. The database uses 51 VH germline genes (across 7 sub-families, VH1-VH7), 40 germline V kappa genes (across the 6 sub-families Vkappa 1-Vkappa 6) and 31 germline V lambda genes (across the 10 V lambda families) \\[[@B10]\\]. Analysis shows that all VH, Vkappa and V lambda sub-families were represented in the selected population and most individual germline genes were identified as the closest hit in at least one selected clone. Thus, the full repertoire of antibody germline genes is being accessed during selection of the library. Figure [3](#F3){ref-type=\"fig\"} shows the frequency of different combinations of the VH, Vkappa and V lambda sub-families in the selected repertoire. Although the relative contribution of each germline family was normalized in the construction of the library, the results show a disproportionate occurrence of certain germline families (VH1, VH3, Vkappa 1, Vlambda 6). Consistent with previous work, this may reflect a greater degree of antibody diversity within these families in the initial library or may reflect emergence of clones during selection or screening as a result of improved expression from these families \\[[@B13]\\]. As well as assigning the closest germline gene for each clone, a detailed analysis was conducted to identify position and sequence of the framework and CDRs and this information is available at the Antibody Atlas db \\[[@B8]\\].\n\n![Frequency of VH and V kappa/V lambda germline gene combinations selected from the library. High quality sequence of all 4,437 clones undergoing secondary screening was generated by sequencing with six primers covering the VH and VL gene segments in both forward and reverse orientations. A consensus sequence was generated and the most closely related antibody germ line genes were identified. **(a)**Frequency of different combinations of VH and V kappa/V lambda germ-line genes occurring in the selected antibodies, represented both numerically and by color coding. **(b)**Frequency of different combinations of VH and V kappa/V lambda germline genes among the selected antibodies.](gb-2007-8-11-r254-3){#F3}\n\nSecondary specificity screening\n-------------------------------\n\nThe selected clones were tested in a secondary screening assay in order to confirm the results of the primary screening ELISA, and to identify antibody clones that were cross-reactive with a panel of irrelevant proteins. Each antibody clone was tested against the target antigen, the fusion partner used, eight irrelevant purified proteins and total protein lysates from zebrafish and yeast, representing more complex non-mammalian protein mixes. All secondary specificity screens took advantage of the sensitivity and dynamic range of time resolved fluorescence using europium labeled secondary antibodies. Figure [2b](#F2){ref-type=\"fig\"} shows a representative cross-reactivity profile on all 12 antigens, for 22 antibodies generated to a single antigen. In a global view of the data, Figure [2c](#F2){ref-type=\"fig\"} summarizes the range of ELISA signals achieved on secondary screening for all 4,437 clones on their specific antigen compared with the signal generated from one of the irrelevant antigens (cytochrome c). This illustrates that the signals achieved span three orders of magnitude. The signal level achieved is dependant on multiple factors, including efficiency of antigen coating, expression level of each antibody and the affinity of that antibody for antigen. Using 1,000 fluorescence units as a threshold, 76.6% of clones undergoing secondary screening were considered positive. Within this group, 80% of clones were specific for target antigen (defined as absence of signal above 1,000 units on all other antigens). Twelve percent of cross-reactivity assay failures at this stage were from antibodies that recognized the antigen fusion partners. Thus, although deselection with the fusion partner reduced the proportion of fusion binders generated, it did not completely prevent such antibodies arising.\n\nThe above analysis shows that 23% of clones deemed positive on primary screening were scored as negative on secondary screening. For many this was due to the increased stringency in the secondary screen since many of these failures were also the poorest performers in the primary screen. For five antigens, all antibody clones were negative on secondary screening, and in these cases the antigen was assumed to have precipitated over time in storage. This was substantiated in one case where the clones were positive on the same antigen in an alternative antigen construct (not shown).\n\nMirroring the primary screening results, antibodies to mammalian-produced antigens had a higher success rate in secondary screening, where 64% of antibodies generated were confirmed as antigen specific and, of those, 69% were scFv with high secondary ELISA scores (defined as 10-fold above the cut off, that is, 10,000 Europium units). In contrast, 57% of scFv generated to bacterially produced antigens were confirmed as antigen specific on secondary screening and, of those, 54% had high secondary ELISA scores. Analyzing the data by antigen target showed that 88% of all targets had at least one antigen-specific binder in the cross-reactivity assay, and 89% of these had at least one binder with high secondary ELISA score. In conclusion, secondary screening has confirmed the results of the initial high throughput screen for the majority of clones and has demonstrated high specificity for the target antigen in 80% of cases.\n\nAssessing performance of antibodies\n-----------------------------------\n\nAlthough affinity is a useful definitive measurement in describing the properties of an antibody, measurement of absolute affinities for thousands of clones is impractical. A relative ranking of clones is possible using inhibition ELISA \\[[@B14]\\] or, more simply, with an end-point ELISA, provided the antibody concentration is normalized and the signal is within the dynamic range of the detection system. The ultimate measure of antibody utility, however, relates to performance and sensitivity in specific assays such as flow cytometry. When using biological samples, however, many factors apart from affinity affect the outcome, for example, level of antigen expression, modification of targets, effect of fixation, and so on. To assess the performance of a panel of antibodies in flow cytometry, independently of these factors, we have employed an assay using beads with defined antigen capture capacity ranging from 0 to \\>600,000 copies per bead. The read out from this assay gives an indication of the potential level of sensitivity a given antibody may have in cell based flow cytometry. Figure [4a](#F4){ref-type=\"fig\"} shows staining of Jagged-1 coated beads, illustrating detection capability below 10,000 copies per bead for 4/9 antibodies. Normalized ELISA provides an alternative means of ranking antibodies and Figure [4c](#F4){ref-type=\"fig\"} demonstrates a correlation between signal achieved in normalized ELISA and performance in the bead based assay.\n\n![Assessing performance of a panel of anti-Jagged-1 antibodies. **(a)**Flow cytometry calibration beads with varying number of anti-human Fc antibodies were coated with Jagged-1-Fc fusion to yield antigen display levels of 29,000, 83,000, 204,000 and 619,000 Jagged-1 molecules/bead. These were labeled with a panel of different recombinant antibodies raised against Jagged-1 and binding was detected with labeled anti-FLAG antibodies. The resulting histograms are shown, giving different levels of sensitivity. In Jag1_D5 for example, five peaks are visible corresponding to uncoated beads and each of the four antigen coated beads. In the case of Jag1_C5, where there is lower sensitivity, only the two beads with highest density are resolved while the others merge with that of the uncoated bead. Where all four beads are clearly resolved, we have calculated the theoretical limit of detection of receptors per bead (Rec. \\#). **(b)**46C mouse embryonic stem cells were stained with the panel of Jagged-1 antibodies and analyzed by flow cytometry. Staining for three antibodies (Jag1_D5, Jag1_D7 and Jag1_A1.2) is shown. The sensitivity of each corresponds to that seen in the bead assay above. **(c)**A plot of normalized ELISA scores with performance in the bead-based flow cytometry performance assay for nine Jagged-1 specific antibodies. ELISA was carried out using 1 \u03bcg/ml of purified antibody. The mean time resolved fluorescence score (expressed as Eu counts) was plotted against median fluorescent intensity determined for an antigen density of 618,888 antigens/bead. The line represents a linear regression analysis of the data (R^2^value = 0.8323). **(d)**Immunohistochemical (Jag1_A6 scFv antibody) and **(e)***in situ*hybridization staining of Jagged-1 in developing kidney of E14.5 mouse embryo, demonstrating localization of staining (dark purple) in developing renal tubules.](gb-2007-8-11-r254-4){#F4}\n\nUsing Jagged-1 as an example, we show that performance in the bead assay in turn correlates with ability to detect endogenous levels of antigen. Jagged-1, which is expressed on embryonic stem cells \\[[@B15]\\] is a ligand for Notch receptors. The Notch signaling system is an ancient and widely used pathway in metazoan development and evolution, where it links the fate of one cell to its neighbors. Notch signaling involves a bi-molecular interaction between receptor and ligand on opposing cells. Mammalian notch signaling involves a choice of four notch receptors (Notch 1-4) and five ligands (Jagged 1,2, delta 1,3,4). Flow cytometry analysis of ES cells stained with three different anti-Jagged-1 antibodies is shown in Figure [4b](#F4){ref-type=\"fig\"}. This shows that signal intensity varies with different antibodies and this, in turn, corresponds to the performance ranking found in the bead based assay.\n\nThe bead assay was applied to 90 other antibodies representing 9 different target genes. In this experiment beads were coated with antigen at a range of densities (18,000, 57,000, 459,000 antigens/bead). To help summarize and present the results for all 90 antibodies, the median fluorescent intensity for the bead coated with 459,000 target molecules was calculated. The signal relative to that of uncoated bead is plotted in Figure [5](#F5){ref-type=\"fig\"} and shows a wide range of signals up to 4,000-fold over background (values and antibody names are given in Additional data file 3). Detection of beads coated with 18,000 copies per bead was achieved with at least one antibody for all 9 genes tested. The sensitivity limit is illustrated in Figure [5](#F5){ref-type=\"fig\"} by using different data point symbols to indicate the lowest antigen density detected for each antibody. In summary, sensitive detection in bead based flow cytometry has been demonstrated for a large panel of the antibodies and detection of endogenous levels in cultured cells exemplified by Jagged-1 in embryonic stem cells.\n\n![Detection sensitivity in a bead-based performance assay for a panel of antigens. The sensitivity limits of 90 antibodies to 9 different antigens were tested using a mix of antigen coated beads of different antigen densities. To summarize these data, the relative median fluorescent intensity on the bead comprising 459,000 antigen copies per bead was calculated (actual values and clone IDs given in Additional data file 2) and the ratio relative to uncoated beads plotted. To further illustrate sensitivity achieved, the identity of the lowest density bead that could be detected is indicated by the data label according to the following guide: 459,000 antigen copies per bead only (diamonds), down to 57,000 antigen copies per bead (square), and down to 18,000 antigen copies per bead (triangles). The antigens were: 1, Efna2; 2, Efna4; 3, Plaur; 4, Alcam; 5, Il13ra1; 6, Sigrr; 7, Ngfr; 8, Cd22; and 9, Vcam1.](gb-2007-8-11-r254-5){#F5}\n\nExpression profiling using tissue microarrays\n---------------------------------------------\n\nWe used 381 antibodies that had passed secondary screening to stain acetone fixed frozen tissue microarrays. From this survey, positive staining was observed for 143 (37%) antibodies representing 46/56 antigens tested. Figure [6a-c](#F6){ref-type=\"fig\"} illustrates an example of an anti-CD5 antibody giving expected staining of scattered human lymphoid cells in lymph node, skin and Peyers patches \\[[@B16]\\]. We also illustrate immunohistochemical staining with some of the antibodies discussed in previous sections. For example, Figure [4d](#F4){ref-type=\"fig\"} illustrates detection of endogenous levels of Jagged-1 in the developing kidney of a 14.5 day embryo. This result was confirmed by *in situ*hybridization as shown in Figure [4e](#F4){ref-type=\"fig\"}. In addition, the three best performing antibodies against nerve growth factor receptor (Ngfr) illustrated in Figure [5](#F5){ref-type=\"fig\"} also detected endogenous levels of Ngfr in immunohistochemistry. With these antibodies, expected peripheral nerve staining \\[[@B17]\\] was present in several adult cores, including large ducts in salivary gland and oesophagus (Figure [6d](#F6){ref-type=\"fig\"}). Occasional clustered staining was seen in sagittal brain and was also present in a 14.5 day murine embryo where the edge of the central nervous system and spinal cord, choroid, dorsal root ganglia and large nerve tracks were stained strongly (Figure [6e](#F6){ref-type=\"fig\"}).\n\n![High throughput immunohistochemistry on tissue microarrays. Example data demonstrating cell surface staining of **(a-c)**CD5 (antibody ant165_155_E04) on lymphocytes in human lymphoid tissue, **(d-e)**nerve growth factor receptor (antibody ant54_71_C07) in nerve bundles attached to the murine esophagus and in developing nerve tracts in E14.5 embryo and **(f-h)**nuclear staining of the transcription factor ELF1 (antibody ant46_62_F12) in cerebrum, liver and uterus. Tissues are: (a) lymph node; (b) skin; (c) Payer\\'s patch in small intestine; (d) murine nerve bundles in esophagus; (e) nerve tracts in 14.5 day embryo with myoblast sheet above; (f) cerebrum; (g) liver; and (h) uterus.](gb-2007-8-11-r254-6){#F6}\n\nPhage display provides the opportunity to generate multiple independent antibodies to provide confirmation of generated expression profiles. Details of overlapping staining profiles with multiple antibodies are given (Additional data file 4). In summary, we demonstrate that it is possible to generate antibodies in high throughput for use in protein expression profiling by immunohistochemistry with an overall success rate consistent with that achieved using hybridoma derived monoclonal antibodies and polyclonal antibodies \\[[@B3],[@B18],[@B19]\\].\n\nDiscussion\n==========\n\nWe describe the creation of an antibody phage display library of over 10^10^clones and its application to high throughput antibody generation and characterization. Many factors affect success rate throughout the individual steps of the process described here. For example, success rate of protein expression is variable and we have previously described the effect of sequence features and choice of fusion partner in *E. coli*expression \\[[@B4]\\]. In *E. coli*, receptor extracellular domains were particularly badly expressed, resulting in insoluble inclusion body formation (not shown). This may reflect a requirement for receptors to be trafficked through a eukaryotic secretion system, to ensure correct disulphide bond formation, native protein folding, good secreted protein expression yields and protein stability during purification and storage. In this study we have therefore utilized both bacterial expression and a transient expression system based in mammalian cells \\[[@B6]\\]. We used the mammalian expression system for receptor extracellular domains, which were the major class of proteins expressed in this system.\n\nTable [1a](#T1){ref-type=\"table\"} summarizes the success rate achieved for this class of protein. Approximately 70% of gene constructs gave rise to a soluble product in culture supernatant, as judged by western blot using an anti-His tag antibody (not shown). For 62% of the positive constructs, 50 \u03bcg of purified protein were purified from a 50 ml culture, providing sufficient material for antibody generation/screening. The success rate was improved further in some cases by adjusting the boundaries of the expressed domain and by using larger culture volumes (unpublished).\n\n###### \n\nSuccess rates in protein expression, antibody isolation and immunohistochemistry\n\n  Step                                                              Number    Percentage\n  ----------------------------------------------------------------- --------- ------------------------\n  **(a) Mammalian expression of receptor extra-cellular domains**             \n  Sequenced confirmed clones                                        157       100%\n  Detected in supernatant by western blot                           108       69%\n  50 \u03bcg purified product generated from 50 ml culture               67        62% (43% cumulative)\n                                                                              \n  **(b) Success rate of antigens in generation of antibodies**                \n  Mammalian antigens for selection                                  205       100%\n  Antigens generating antibody                                      181       88%\n  Bacterial antigens for selection                                  192       100%\n  Antigens generating antibody                                      103       54%\n  **Total antigens for selection**                                  **404**   **100%**\n  **Antigens generating antibody**                                  **292**   **72%**\n                                                                              \n  **(c)Success rate for primary antibody screening**                          \n  Number of antibodies screened                                     38,164    100%\n  Number positive on primary screen                                 9,384     24.6%\n  Unique clones generated                                           7,236     77.1% (19% cumulative)\n                                                                              \n  **(d)Success rate for secondary antibody screening**                        \n  Clones undergoing secondary specificity screening                 4,437     100%\n  Positive in secondary screening                                   3,400     76.6%\n  Specificity demonstrated                                          2,725     80.1% (61% cumulative)\n                                                                              \n  **(e) Success rate in immunohistochemistry**                                \n  Stained on tissue microarrays                                     381       100%\n  Positive staining observed                                        143       37%\n\nThe selection success rate was dependent on the antigen source, with mammalian produced antigens being superior to bacterially produced antigens (Figure [2](#F2){ref-type=\"fig\"}, Table [1b](#T1){ref-type=\"table\"}). In mammalian systems, misfolded proteins are degraded \\[[@B20]\\], and this additional quality control may have led to improved selection results. In contrast, most of the bacterially produced proteins were fused to tags such as maltose binding protein (MBP), which enhanced production of soluble product \\[[@B4]\\]. Although this has improved the yield of soluble product, it is possible that there is a greater proportion of misfolded protein in the bacterially produced antigen preparations depending on the protein target being expressed \\[[@B21]-[@B23]\\], giving higher backgrounds and compromising selections. For example, expression of the HPV oncoprotein E6 in *E. coli*as an MBP fusion resulted in high soluble expression levels but a major fraction was misfolded aggregate with a minor fraction containing properly folded protein with full biological activity \\[[@B24]\\].\n\nWhere antibodies are raised by selection on native like protein (for example, by phage panning on extracellular domains generated by mammalian expression), it might be expected that many of the antibodies produced will recognize the native form of antigen. Subsequent recognition by such antibodies of both native and denatured forms will then be dependant on whether the target epitope is available after denaturation, for example, within fixed tissues. This is in contrast to other approaches where denatured antigens are used to raise antibodies, for example, by immunization \\[[@B3],[@B18],[@B19]\\]. Here, recognition of native antigen by antibodies raised on denatured antigen will depend on whether the linear epitope is found in native conditions. Thus, antibodies raised by different approaches may have different utility in recognizing native protein and different sensitivities to fixation/denaturation.\n\nIn general terms, we observe a success rate of 37% for generating a positive profile (Table [1d](#T1){ref-type=\"table\"}; detailed in Additional data file 3), which is broadly similar to that described previously \\[[@B3],[@B18],[@B19]\\]. It is difficult, however, to compare immunohistochemistry results in a more specific way across projects since different organisms, genes and antibody generation/fixation regimes are used. In addition, specificity of antibody staining is an important consideration and different groups verify staining patterns to different extents. For example, false positive binding to unrelated antigens can occur, not just with polyclonal antibodies, but also with monoclonal antibodies, as exemplified by previous studies with antibodies on protein microarrays \\[[@B25]-[@B27]\\]. In a similar way, exposing hybridoma derived monoclonal antibodies to the majority of tissue in the body has been shown to result in false positive staining \\[[@B18]\\]. Systematic verification of staining patterns from expression databases or literature can provide supporting evidence (for example, widespread nuclear expression of ELF1 (Figure [6g-i](#F6){ref-type=\"fig\"})). *In situ*hybridization can also provide supporting evidence that a given profile is correct (as exemplified for Jagged-1 in Figure [4d,e](#F4){ref-type=\"fig\"}). Finally, the availability of multiple antibodies to the same target can increase confidence in the veracity of a given profile by independently generating similar or identical profiles \\[[@B18]\\]. Among the positive antibodies profiled here, 58 antibodies (15 genes) represented cases where more than one antibody gave the same or overlapping profiles (examples in Figure [6](#F6){ref-type=\"fig\"} and summarized in Additional data file 3).\n\nHere, we show the feasibility of antibody generation in a high throughput manner. The bottlenecks in this process was not antibody generation but the cloning, verification and generation of high quality antigen and the preparation and use of the antibodies in downstream applications, for example, in immunohistochemistry. The library and selection process employed was particularly effective, resulting in 7,236 unique antibodies from screening and sequencing 38,164 clones (Table [1c](#T1){ref-type=\"table\"}). From 292 successful selections, an average of 25 unique binding antibodies was generated for every 94 clones screened. The success was due to steps taken during library construction to maintain diversity to a maximum and also due to optimization of the selection and screening process (discussed previously in Results). Having developed the library and established the system for carrying out this process, 48-96 antigens can be selected and undergo primary screening and sequencing by 2 people over a 6 week period.\n\nIn this study we have sequenced and screened 38,000 antibodies and describe different degrees of characterization for a proportion of these, ranging from specificity ELISA through to detection of endogenous levels of antigens using arrays of adult tissue. A standardized system of validation of antibodies across multiple labs would add to the value of high throughput antibody generation initiatives and the use of antibodies in general. Antibody characterization is complex, however, since the utility of antibodies is application dependent and is subject to many factors, including endogenous levels of antigen expression, availability of epitopes and their sensitivity to fixation or denaturation, antibody affinity/specificities and the sensitivity of detection system used. The affinity of an antibody for a given antigen is a property that can be measured in absolute terms and is related to performance, provided the test antigen accurately represents the target antigen in a biological sample. Affinity determination using surface plasmon resonance provides accurate affinity values but is laborious and requires specialist equipment that is not universally available. Inhibition ELISA can, within limitations, also provide affinity values \\[[@B14]\\]. In this study we have utilized a simple, sensitive end-point ELISA assay with wide dynamic range to measure binding and permit performance ranking of the clones (following normalization of antibody concentration). We have also introduced a surrogate performance assay for flow cytometry, replacing cells with beads with different antigen coating densities, allowing a measurement of sensitivity in terms of target density per bead. In a similar way, a simple surrogate could be envisioned to rank the potential of clones as western blot reagents by measuring detection limits of denatured antigen on support matrices such as PVDF. While such assays do not guarantee success with biological samples, they can facilitate ranking of clones for subsequent investigations.\n\nAntibody specificity is another important property that relates to the relative affinity for other \\'irrelevant\\' epitopes. Cross-reactivity can arise through conserved interactions with related antigens or by \\'polyspecificity\\' involving different modalities of binding to unrelated antigens \\[[@B28],[@B29]\\]. It is difficult to be definitive about measuring cross-reactivity since it depends on what antigens are presented in any application. Immunohistochemistry on microarrays of fixed tissues to determine global protein expression pattern represents a particularly challenging application with a high proportion of the proteome being presented at varying levels. Nonetheless, in this study we demonstrate specific detection of non-abundant proteins and provide supporting data from independent antibodies, *in situ*hybridization or alternative data sources to confirm the resultant staining profiles for a proportion of these (Additional data file 3).\n\nTo identify obviously cross-reactive clones in a simple \\'first pass\\' assay, we have used a panel of 11 irrelevant antigens, including non-mammalian lysates. Protein microarrays or multiplex bead assays would be more informative in providing a wider range of proteins and these may become more widely available in the future. Standardized western blotting of irrelevant tissue lysates (such as yeast/zebrafish lysates used here) may provide an alternative source of \\'arrayed\\' irrelevant proteins using a technique currently available in many labs that could identify binders with relatively low specificity.\n\nConclusion\n==========\n\nHistorically, antibody validation has been *ad hoc*, making it difficult to compare antibodies from different sources, resulting in much wasted effort and cost. Here we demonstrate the high throughput generation and validation of recombinant antibodies on an unprecedented scale. Large scale efforts to generate and validate antibodies will not only provide a wealth of reagents for research and diagnostics, but could also drive the introduction of a more consistent quality control framework to facilitate comparisons of binders from different sources for use in different applications, whatever the source.\n\nMaterials and methods\n=====================\n\nAntigen cloning and expression\n------------------------------\n\nAntigens were obtained either from commercial sources, or were expressed from *E. coli*or transient mammalian expression systems using systems we have described previously \\[[@B30]\\]. For primer design, mRNA sequence was downloaded from the Ensembl entry \\[[@B31]\\] with coding and 5\\' and 3\\' untranslated region (UTR) annotation. Pfam domains \\[[@B30]\\] were annotated onto the nucleic acid sequence and primers designed to amplify either the full-length open reading frame or specific Pfam domains. Inserts were amplified from cDNA using nested PCR, as described previously \\[[@B4]\\], GATEWAY cloned into expression vectors \\[[@B32]\\] and DNA sequence confirmed. All steps in antigen production, including PCR amplification, sequence confirmation, expression and purification, were electronically tracked using a custom designed laboratory information management system (unpublished). In the case of automated primer design of the immunoglobulin superfamily (IgSF), targets were identified from a search of the ENSEMBL database (NCBI m34 mouse assembly) for ENSMUSP (protein) entries that contained annotated IgSF Pfam domains and signal peptides. The transcript sequence with annotated protein features, including coding sequence signal peptides, transmembrane (TM) regions and Pfam domain boundaries, were batch downloaded and used as a template for automated primer design. Outer primers were designed to the UTRs to amplify the full-length open reading frame using Primer3 \\[[@B33]\\]. Three design strategies were employed for the inner primers: EC+SIGP-euk, coordinates 1 to minus 5 amino acids upstream of the TM, minus a stop codon for HEK293E expression with carboxy-terminal tags; EC-SIGP-pro, coordinates plus 1 amino acid downstream of the SIGP to minus 5 amino acids upstream of the TM plus a stop codon for *E. coli*expression with amino-terminal tags; and IC-pro, coordinates plus 5 amino acids downstream of the TM to the end of the coding sequence plus a stop codon for *E. coli*expression with amino-terminal tags.\n\nOur *E. coli*protein expression system was described previously \\[[@B4]\\] and utilizes the T7 RNA polymerase promoter driving cytoplasmic expression in BL21-DE3 cells (Novagen, Nottingham, UK). This line has an inducible RNA polymerase gene under the control of the lac promoter. Expression was performed using either shake flasks or 24-well blocks with the Studier auto-induction protocols \\[[@B34]\\]. Most proteins were expressed with an amino-terminal decahistidine (His10) tag for affinity purification and either a thioredoxin (Trx) or MBP solubility enhancing fusion. In common with many low molecular weight eukaryotic proteins \\[[@B4]\\], the Rab proteins did not require a solubility enhancing fusion for high level soluble expression. Expression in mammalian cells was by transient transfection of HEK293E suspension cells \\[[@B6]\\] with expression vectors designed to add a carboxy-terminal His10 tag or His10-rat Cd4 (domains 3 and 4) fusion \\[[@B35]\\]. Most receptor ectodomains were expressed in HEK293E cells at a 50 ml culture volume, but low expressing clones were scaled up to a 200 ml culture volume. Proteins were affinity purified using a FPLC system or Qiagen8000 robot. All proteins were subject to a quality control procedure before handover for antibody selection, which required a minimum of 50 \u03bcg and greater than 80% purity of full-length target protein. Proteins were stored in a buffer containing 50% or 10% glycerol for -20\u00b0C or -80\u00b0C storage, respectively\n\nConstruction of pSANG4 display vector\n-------------------------------------\n\nA modified version of the phage display vector pHEN1 \\[[@B36]\\] was constructed by inserting a novel/cloning linker region. This was created by annealing primers NcNotlinkS and NcNotlinkA, extending with DNA polymerase and cutting with *Nco*I/*Not*I before cloning into the *Nco*I/*Not*I site of pHEN1 to give pSANG3. This allows sequential cloning of antibody light chains via *Nhe*I/*Not*I and heavy chains via *Nco*I/*Xho*I. The *pelB*signal sequence in pSANG3 was replaced with the signal sequence from M13 gene 3 to create the vector pSANG4. This leader is potentially more useful for ligation independent cloning. (Ligation independent cloning allows direction cloning without restriction enzymes or ligase enzymes, by using the 5\\'-3\\' exonuclease activity of DNA polymerase to create single stranded overhangs for annealing. The presence of a single nucleotide causes the exonuclease activity to arrest where that nucleotide becomes incorporated and so long overhangs can be engineered \\[[@B37]\\].) The M13 leader and 5\\' UTR were created by annealing oligonucleotides G3HindNdeS and G3NcoA (see Additional data file 4), extending with DNA polymerase (Novagen), digesting and cloning into the *Hin*dIII and *Nco*I sites of pSANG3. G3HindNdeS is based on the sequence of M13 but adds a stop codon to stop elongation initiated at the *lac*Z start codon, which is upstream in the vector. The primers also introduce a new *Nde*I restriction site at the start of the leader sequence and introduce a silent mutation in the tenth codon of the leader with a view to increasing the potential overhang in ligation independent cloning. Constructs were confirmed by sequencing with primers LMB3 and fdtseq1. The resultant vector, pSANG4, is represented in Figure [1](#F1){ref-type=\"fig\"} and more fully described in Additional data file 4.\n\nConstruction of display library\n-------------------------------\n\nAn antibody phage display library was created by sequentially cloning a repertoire of light chain variable regions (VL) followed by cloning of heavy chain variable regions (VH) in pSANG4. The VH and VL pools were cloned through primer encoded *Nco*I/*Xho*I and *Nhe*I/*Not*I sites, respectively. The VH and VL domains were initially cloned into an intermediate, out of frame cloning vector, pSANG2. The final format of the antibody is a scFv with VH and VL fragments joined by a flexible linker peptide (Gly~4~Ser Gly~4~Ser Gly~3~Ala Ser).\n\nA more detailed description of the library construction is provided in Additional data file 4. In brief, antibody heavy and light chain variable region repertoires were created by PCR amplification from human lymphocytes. Fourteen heavy and 28 light chain primer pairs were used and maintained separately in 462 separate primary PCRs with cDNA originating from 43 lymphocyte donors. Light chains were cloned as *Nhe*I/*Not*I fragments. To maintain diversity, the VL repertoire was cloned as 9 separate sub-libraries with an average of 10^7^clones per library (6 kappa families and 3 pools of lambda families). Plasmid DNA was prepared from these and pooled into two separate kappa or lambda chain sets. The VH repertoire was sub-cloned via *Nco*I*/Xho*I sites to these two light chain libraries. Fourteen different sub-libraries were prepared, consisting of each of the seven heavy chain families (VH1-VH7) combined with either kappa or lambda light chains. It has been suggested \\[[@B13]\\] that particular combinations of VH and VL families express better and so two additional sub-libraries of VH3/\u03ba3 and VH3/\u03ba1 were generated, giving a combined total of 1.1 \u00d7 10^10^inserts in the \\'McCafferty\\' phage display library.\n\nHigh throughput selection\n-------------------------\n\nRescue of phage particles from the library was carried out essentially as described \\[[@B38]\\] but using trypsin-cleavable helper phage \\[[@B11],[@B39]\\], which increased the efficiency of binder isolation such that only two rounds of selection were required. Selection was as described previously \\[[@B38]\\] with the following modifications (described in more detail in Additional data file 4). Following the rescue of the individual library aliquots, the antibody-phage particles were purified by sequential polyethylene glycol precipitation (Sigma, Poole, UK) and cesium chloride gradient centrifugation and the quality assessed by confirming the presence of full length antibody-gene 3 fusion by western blotting using anti-gene 3 antibody (New England Biolabs, Hitchin, UK; data not shown). Each aliquot was titrated and pooled according to the size of the aliquot to normalize the contribution of each clone. Final concentration was 10^13^colony forming units/ml and 50 \u03bcl was used per selection. For the selections, each antigen was coated at 10 \u03bcg/ml onto a single well of a 96-well microtiter plate. We routinely performed a deselection process before commencing the first round of selection for recombinant protein targets with common fusion partners, for example, human IgG1 Fc, rat CD4 (domains 3 and 4), MBP, and Trx. This procedure helped to deplete binders to fusion partners.\n\nTwo rounds of selections were performed using the \\'McCafferty\\' scFv antibody-phage library. Following binding of the library to immobilized antigen, phage were eluted in 100 \u03bcl of freshly prepared 125 \u03bcg/ml TPCK-trypsin (Sigma) in 50 mM Tris pH 8, 1 mM CaCl~2~at room temperature for 15 minutes. A mid-log culture of *E. coli*TG1 cells (200 \u03bcl) was added to each well containing the eluted phage and shaken slowly (150 rpm) at 37\u00b0C for 1 hour. At this stage aliquots can be taken to titer the output. Bacteria were pelleted at 3,000 \u00d7 g for 10 minutes, resuspended in 500 \u03bcl of 2 \u00d7 TY medium with ampicillin and 2% glucose (2 \u00d7 TY amp/glu) and grown overnight at 30\u00b0C. The next day, 1 ml cultures of 2 \u00d7 TY amp/glu in a deep well plate were inoculated with 20 \u03bcl of the overnight culture and incubated at 37\u00b0C for 1 hour at 600 rpm. Helper phage were added to a MOI of 10 and incubated with gentle shaking at 150 rpm at 37\u00b0C for 1 hour. Bacteria were pelleted at 3,000 \u00d7 g for 10 minutes, resuspended in 500 \u03bcl 2 \u00d7 TY medium with ampicillin and kanamycin and incubated overnight at 30\u00b0C with shaking at 600 rpm. The selected outputs were then evaluated for specific antigen binding using phage ELISA \\[[@B40]\\]. Individual, selected scFv colonies were rescued as scFv-phage particles in 96-well microtitre wells and transferred to specific antigen (5 \u03bcg/ml) coated wells. Anti-M13-horseradish peroxidase conjugate (GE Healthcare, Little Chalfont, UK) and 3,3\\',5,5\\'-tetramethyl benzidine substrate (Sigma) were used to detect bound phage.\n\nSub-cloning selected populations\n--------------------------------\n\nPrior to screening of individual clones, the selected populations were sub-cloned into an expression vector pSANG14-3F \\[[@B7]\\]. This vector fuses the antibody to alkaline phosphatase to drive dimerization and facilitate production of antibody product and subsequent detection of binding. It also introduces a hexahistidine tag for purification and a tri-FLAG tag for detection of binding via secondary anti-FLAG antibodies (Sigma). Sub-cloning of selected populations prior to screening also increased success rate by reducing the proportion of growth-advantaged clones with truncated inserts in the screened population. Selected antibody gene populations were amplified by PCR from the glycerol stocks of the second round of selection using M13 LeadSeq (5\\'-AAATTATTATTCGCAATTCCTTTGGTTGTTCCT-3\\') and NotMycSeq (5\\'-GGCCCCATTCAGATCCTCTTCTGAGATGAG-3\\') primers. The amplified DNA was run on a 1% agarose gel; insert excised and purified using a QiaQuick PCR purification kit (Qiagen, Crawley, UK). Two micrograms of the purified PCR products were digested with *Nco*I and *Not*I restriction enzymes (New England Biolabs) and electrophoresed in a 1% agarose/TBE gel. The DNA was purified from the gel slice using the QiaQuick gel purification kit (Qiagen) and then ligated into the *Nco*I/*Not*I digested pSANG14-3F expression vector using T4 DNA ligase (New England Biolabs). After purification of the DNA from the ligation reaction, the ligated product was electroporated into electrocompetent BL21 (DE3) cells (Novagen). Ninety-four clones per target were picked, grown and stored at -80\u00b0C.\n\nPrimary screening of selected populations\n-----------------------------------------\n\nThe primary and secondary screening ELISAs were performed essentially as described previously \\[[@B7]\\]. All primary screening ELISAs were performed on scFv released from bacterial pellets, following overnight induction, using Bug Buster lysis buffer (Novagen). For the primary screening ELISA, we initially performed the screening assay in a 96-well format using direct detection of scFv binding via the alkaline phosphatase fusion protein and addition of pNPP substrate (Sigma). For each target, one 96-well flat bottomed Maxisorp polystyrene plate (Nunc, Roskilde, DK) was coated with the specific antigen and one ELISA plate with the appropriate control antigen (for example, MBP, Fc, Trx). The scFvs were assayed as single points on each plate and any scFv with a specific antigen signal \u22653-fold higher than the control was scored as a positive. Later assays were performed in 384-well ELISA plates using time resolved fluorescence, which involves using europium labeled secondary detection (Perkin Elmer, Beaconsfield, UK) to mirror the secondary screening assay. Here, for each target, one 384-well flat bottomed black polystyrene plate (Nunc Maxisorp) was coated with both the specific antigen and with the appropriate control antigen. ScFv binding was detected using an anti-FLAG secondary antibody (Sigma) conjugated to europium (Perkin Elmer). Again, an antibody clone was deemed to be antigen specific in the primary screening ELISA where the mean of the two replicate scores on the target antigen was at least three-fold greater than the mean of the two replicate scores for the control protein.\n\nSecondary screening of selected populations\n-------------------------------------------\n\nThe secondary screening ELISA was used to identify cross-reactive antibodies. ScFv was produced by overnight growth in 1 ml of auto-induction media \\[[@B34]\\] in a 96-well deep well microtitre plate (Costar, Lowell, MA, USA). Periplasmic extracts was prepared by centrifugation at 3,000 g for 10 minutes followed by resuspension in 40 \u03bcl TES buffer (50 mM TrisHCl, pH 8.0, 1 mM EDTA, 20% (w/v) sucrose, 2 \u03bcl/ml protease inhibitor cocktail set VII (CalBioChem, Nottingham, UK)). After 5-10 minutes 60 \u03bcl of a one-fifth dilution of TES supplemented with 2.5 mM MgCl, 2 \u03bcl/ml protease inhibitor cocktail set VII, and 25 U/ml Benzonase (VWR, Lutterworth, UK) were added. Recombinant antibody fused to alkaline phosphatase (scFv-AP) was mixed with phosphate-buffered saline (PBS)/1% milk protein then transferred to pre-blocked 384-well assay plates for development of the ELISA. For convenience, the scFv-AP-antigen binding step was performed overnight at 4\u00b0C. Six 384-well ELISA plates were required to screen each plate of clones. There were two antigens per 384-well plate coated in duplicate for the panel of 12 antigens. The antigen panel comprised the specific antigen, a fusion partner control, keyhole limpet hemocyanin, thyroglobulin, myoglobin, cytochrome c, human IgG, laminin, fibronectin, \u03b1-glycerol phosphate dehydrogenase (Sigma) and total protein lysates from zebra fish (*Danio rerio*) and yeast (*Schizosaccharomyces pombe*). Plates were washed and bound scFv-AP detected with an anti-FLAG secondary antibody conjugated to biotin (Sigma) followed by a streptavidin-europium conjugate (Perkin Elmer), which was detected using time resolved fluorescence.\n\nSequence analysis\n-----------------\n\nEach plate of 94 scFv clones, corresponding to a single target antigen, was sequenced using a single forward primer, pSANG Fwd (5\\'-TATGAAATACCTGCTGCCGACC-3\\'). Sequences were analyzed using Blaze 2, software from Cambridge Antibody Technology (Cambridge, UK). Following alignment, duplicate clones were identified and removed. The remaining scFv clones were then aligned according to the sequence of the heavy chain CDR3 region, which is the most variable segment. Up to 22 clones per target were chosen based on both their signal-to-noise ranking from the primary screening ELISA (highest to lowest) and on the diversity of their HCDR3 amino acid sequences. Selected clones were \\'cherry picked\\' for further sequence and specificity analysis, with each derivative plate containing four sets of scFv specific to four different target antigens. The selected scFv clones were sequenced in depth using six sequencing primers, three forward primers (VF1: 5\\'-GGGGAATTGTGAGCGGA-3\\'; VF2: 5\\'-GATCGAGATCTCGATCCCGCGA-3\\'; Hlink: 5\\'-ACCGCCAGAGCCACCTCCGCC-3\\') and three reverse primers (Llink: 5\\'-GGCGGAGGTGGCTCTGGCGGT-3\\'; VR1: 5\\'-CGTGCGGCAGTAATTTCC-3\\'; VR2: 5\\'-TGTAGTAATATCGCCCTGAGCAGCC-3\\'). The reads were assembled using phrap, and a minimum read depth of 3 was required for all bases before any consensus call was accepted. The consensus DNA sequence was subjected to six frame translation and vector encoded sequences flanking the VH and VL genes identified (the *pelB*signal sequence, the scFv linker sequence (Gly~4~Ser Gly~4~Ser Gly~3~Ala Ser) and the alkaline phosphatase gene). The closest germline variable region gene was identified and the various framework and CDRs were identified.\n\nPerformance ranking of scFvs using flow cytometry calibration beads\n-------------------------------------------------------------------\n\nFlow cytometry calibration beads with defined numbers of anti-human IgG receptors (Quantum Simply Cellular anti-Human IgG bead sets, Bangs Laboratories Inc., Fishers, IN, USA)) were used to assess antibody performance in flow cytometry. This allows the generation of beads with defined amounts of human Fc fused antigens, permitting an assessment of sensitivity limits for selected antibodies. For example, the set described in Figure [4](#F4){ref-type=\"fig\"} used beads with 29,588, 83,460, 204,228, and 618,888 antigen copies per bead. Beads were incubated with blocking buffer (PBS/0.1% (w/v) bovine serum albumin (BSA; Sigma)/0.01% (w:v) NaN~3~) prior to incubation with an excess of Fc-tagged antigen. Following binding of the antigen, the beads were washed by centrifugation in blocking buffer (300 \u00d7 g for 4 minutes), and blank (uncoated) beads were then added to the pool of beads. The final bead mixture was dispensed into 96-well plates. Following a further wash step, as above, 100 \u03bcl of scFv at 2 \u03bcg/ml or a suitable control was added to the beads and allowed to bind for 1 hour on ice. After washing, bound scFv was detected using anti-FLAG-biotin (Sigma) at 1 \u03bcg/ml followed by streptavidin-PE (Pharmingen, Oxford, UK) at 0.5 \u03bcg/ml. Antigen binding to the beads was confirmed by staining with an Fc-specific anti-IgG-FITC antibody (Jackson Immunolaboratories, West Grove, PA, USA). This antibody did not bind to the beads in the absence of Fc-tagged antigen. Data were collected using a FC500MPL flow cytometer (Beckman Coulter, High Wycombe, UK), WinMDI \\[[@B41]\\] was used to create the histogram plots shown in Figure [4](#F4){ref-type=\"fig\"}, and median fluorescence intensities of each of the bead peaks were generated using the log/log method within the CXP software (Coulter). The limit of sensitivity for each scFv clone was determined using the QuickCal v2.3 software (Bangs Laboratories).\n\nFlow cytometry with ES cells\n----------------------------\n\n46C mouse embryonic stem cells were grown in LIF supplemented medium in feeder-free adherent culture as described \\[[@B42]\\]. This cell line is known to express Jagged-1 \\[[@B15],[@B42]\\]. Cells were removed from adherent culture using trypsin EDTA and stained in 96 U-well plates in 100 \u03bcl at 3 \u00d7 10^5^/well with all procedures being performed on ice. ScFv were applied for 1 hour at 1 \u03bcg/ml diluted in PBS supplemented with 1% BSA and 0.01% sodium azide. Following washing, anti-FLAG-biotin (Sigma; clone M2) was applied at 1 \u03bcg/ml diluted in PBS supplemented with 0.1% BSA and 0.01% sodium azide. After an hour\\'s incubation, the cells were washed and 100 \u03bcl of streptavidin-phycoerythrin conjugate (Becton Dickinson, Oxford, UK) was applied at 0.5 \u03bcg/ml for 30-60 minutes. After washing, the cells were resuspended in PBS/0.1% BSA/sodium azide and staining was visualized using a FC500-MPL flow cytometer. Analysis was performed using WinMDI with dead cells gated out as determined by forward versus side scatter profiles. No staining was evident in the absence of scFv or in the presence of an irrelevant scFv. Histogram plots showing levels of staining with streptavidin-phycoerythrin for control and test are shown.\n\nImmunohistochemistry\n--------------------\n\nFrozen tissue arrays of normal adult human tissue and developmental and adult murine tissue were prepared by Covance Ltd (Yorkshire, UK) as previously described by Kononen *et al*. \\[[@B43]\\]. These encompassed one slide type of 28 different human adult tissues or a second slide type of 23 different adult murine cores, with murine adult sagittal brain and E14.5 embryo sections added. In each array, two tissue cores were included from each donor sample and two donors were used to represent each tissue type. Sections of the microarrays were fixed by immersion in acetone at ambient temperature for 15 minutes, blocked for endogenous alkaline phosphatase activity (DAKO, Ely, UK) and non-specific protein interaction. Alkaline phosphatase-fused scFv antibodies were incubated for 1 hour in blocking solution then an alkaline phosphatase-labeled anti FLAG tag antibody (Sigma) was applied for 30 minutes to amplify the quantity of label for enzyme detection using NBT/BCIP substrate (DAKO), which was also applied for 30 minutes. Although binding can be detected directly via the alkaline phosphatase fusion partner, we have found improved sensitivity using a secondary alkaline phosphatase labeled antibody \\[[@B7]\\]. The arrays were counterstained with nuclear fast red (Vector Labs, Peterborough, UK), dehydrated in alcohol, cleared in xylene and mounted in Eukitt resinous mountant (ProSciTech, Poway, CA, USA)).\n\n*In situ*hybridization\n----------------------\n\nE14.5 mouse embryos (C57BL/6JTyrC-Brd) were fixed in 10% neutral-buffered formalin for 48 hours and embedded in paraffin wax. Eight micrometer sections on super frost + slides were hybridized with a 1,170 bp digoxygenin-labeled antisense RNA probe to the mouse Jag1 gene (ENSMUSG00000027276) generated from a DNA template. The template was synthesized by RT-PCR with gene specific primers (Jag1_1\\_f: CTAGGCCTGGAGCTTCCACATCTGC; Jag1_1\\_r: TAATACGACTCACTATAGGGAGCAGTCCCGGTGGTGAACCTGGAT).\n\nTranscription was performed using an Ambion Maxi-Script kit with the addition of digoxygenin UTP (Roche, London, UK). Sixty nanograms of probe were used for hybridization at 65\u00b0C for 6 hours, using a Ventana (Illkirch, France) Discovery platform with BlueMap and RiboMap kits, according to manufacturer\\'s guidelines. The final stringency of washing was 3 \u00d7 0.1 SSC at 85\u00b0C. Sections were counterstained with Nuclear Fast Red\n\nAbbreviations\n=============\n\nBSA, bovine serum albumin; CDR, complementarity determining region; ES, embryonic stem; MBP, maltose binding protein; Ngfr, nerve growth factor receptor; PBS, phosphate-buffered saline; scFV, single chain Fv; TM, transmembrane; Trx, thioredoxin; UTR, untranslated region; VH, heavy chain variable region; VL, light chain variable region.\n\nAuthors\\' contributions\n=======================\n\nJMC coordinated and led the overall project, designed and helped construct the library (with MOS and KJV) and prepared the manuscript along with DJS. Protein production from primer design and gene cloning through to protein purification and quality control was led by MRD and carried out by MRD, SDJC, AMC, RLP, SPS. DJS led the selection and screening of the antibody library along with ARP, VC, SK, CDM and WR. WEW prepared antibodies for immunohistochemistry and flow cytometry. JLY/ARP carried out the flow cytometry work and JLY characterized antibodies to Notch and jagged antigens. Immunohistochemistry was carried out by KFC, JSC, GF, YH, WJH, and JNM, led by AW. AKK carried out *in situ*hybridization. Informatics support, including sequence analysis and database/website construction, was provided by SREC, GJG, GLM, and JX.\n\nAdditional data files\n=====================\n\nThe following additional data are available with the online version of this paper. Additional data file [1](#S1){ref-type=\"supplementary-material\"} is a table listing all 406 antigens used in antibody selection and the number of unique antibodies generated. Additional data file [2](#S2){ref-type=\"supplementary-material\"} is a table listing median fluorescent intensities of antigen beads as plotted in Figure [5](#F5){ref-type=\"fig\"}. Additional data file [3](#S3){ref-type=\"supplementary-material\"} provides a description of staining profiles supported by multiple antibodies in immunohistochemistry. Additional data file [4](#S4){ref-type=\"supplementary-material\"} describes the construction and rescue of the \\'McCafferty\\' antibody phage display library.\n\nSupplementary Material\n======================\n\n###### Additional data file 1\n\nAll 406 antigens used in antibody selection and the number of unique antibodies generated.\n\n###### \n\nClick here for file\n\n###### Additional data file 2\n\nMedian fluorescent intensities of antigen beads as plotted in Figure [5](#F5){ref-type=\"fig\"}.\n\n###### \n\nClick here for file\n\n###### Additional data file 3\n\nDescription of staining profiles supported by multiple antibodies in immunohistochemistry.\n\n###### \n\nClick here for file\n\n###### Additional data file 4\n\nConstruction and rescue of the \\'McCafferty\\' antibody phage display library\n\n###### \n\nClick here for file\n\nAcknowledgements\n================\n\nAll work was carried out within the \\'ATLAS of protein expression\\' group at the Sanger Institute, funded by the Wellcome Trust. We thank Derek Stemple and Jurg Bahler, who kindly supplied the zebrafish and yeast, respectively, for lysate production. We thank David Tannahill for access to the Jagged-1 ISH data. We thank the Small Projects Sequencing Team, who performed all of the sequencing for this project (all from Wellcome Trust Sanger Institute). We thank Colin Hardman for supplying the Blaze 2 sequence analysis software package, Sarah Carmen for the trypsin sensitive helper phage and Cambridge Antibody Technology plc for their kind permission in allowing their use. JLY and WR were funded by the EU Eurostemcell consortium <http://www.eurostemcell.org/>. AK-K is funded by the EU EmbryoExpress program at the Wellcome Trust Sanger Institute.\n"}